rxEPA
EPA/600/R-15/068 | January 2016 | www.epa.gov/isa
United States
Environmental Protection
Agency
Integrated Science Assessment for
Oxides of Nitrogen - Health Criteria
Office of Research and Development
National Center for Environmental Assessment, Research Triangle Park, NC
-------
EPA/600/R-15/068
United States T ->ni/:
Environmental Protection January 2016
Agency www.epa. gov/ncea/isa
Integrated Science Assessment
for Oxides of Nitrogen—
Health Criteria
January 2016
National Center for Environmental Assessment-RTF Division
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
DISCLAIMER
This document has been reviewed in accordance with the U.S. Environmental Protection Agency policy
and approved for publication. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
-------
CONTENTS
INTEGRATED SCIENCE ASSESSMENT TEAM FOR OXIDES OF NITROGEN xvi
AUTHORS, CONTRIBUTORS, AND REVIEWERS xx
CLEAN AIR SCIENTIFIC ADVISORY COMMITTEE OXIDES OF NITROGEN NAAQS REVIEW
PANEL xxv
ACRONYMS AND ABBREVIATIONS xxvi
PREAMBLE xxxviii
1. Process of Integrated Science Assessment Development xxxviii
Figure I Schematic of the key steps in the review of National Ambient Air Quality
Standards. xxxix
Figure II Characterization of the general process for developing an Integrated Science
Assessment. xl
2. Literature Search xli
Figure III Illustration of literature search and study selection process used for
developing Integrated Science Assessments. xliii
3. Study Selection xliii
4. Evaluation of Individual Study Quality xliv
a. Atmospheric Science and Exposure Assessment xlv
b. Epidemiology xlvi
c. Controlled Human Exposure and Animal Toxicology xlviii
d. Ecological and Other Welfare Effects xlix
5. Evaluation, Synthesis, and Integration of Evidence across Disciplines and Development of
Scientific Conclusions and Causal Determinations I
a. Evaluation, Synthesis, and Integration of Evidence across Disciplines I
b. Considerations in Developing Scientific Conclusions and Causal Determinations Iv
Table I Aspects to aid in judging causality. Ivii
c. Framework for Causal Determinations lix
Table II Weight of evidence for causal determination. Ix
6. Public Health Impact Ixii
a. Approach to Identifying, Evaluating, and Characterizing At-Risk Factors Ixiii
Table III Characterization of evidence for potential at-risk factors. Ixv
b. Evaluating Adversity of Human Health Effects Ixv
c. Concentration-Response Relationships Ixvi
7. Public Welfare Impact Ixvii
a. Evaluating Adversity of Ecological and Other Welfare Effects Ixviii
b. Quantitative Relationships: Effects on Welfare Ixx
PREFACE Ixxi
Legislative Requirements for the Review of the National Ambient Air Quality Standards Ixxi
Overview and History of the Review of the Primary National Ambient Air Quality Standards for
Nitrogen Dioxide Ixxiii
Table I History of the primary National Ambient Air Quality Standards for nitrogen
dioxide since 1971. Ixxiv
m
-------
CONTENTS (Continued)
EXECUTIVE SUMMARY Ixxviii
Purpose and Scope of the Integrated Science Assessment Ixxviii
Sources and Human Exposure to Nitrogen Dioxide Ixxix
Health Effects of Nitrogen Dioxide Exposure Ixxxi
Table ES-1 Causal determinations for relationships between nitrogen dioxide exposure
and health effects from the 2008 and 2016 Integrated Science Assessment for
Oxides of Nitrogen. Ixxxii
Short-Term Nitrogen Dioxide Exposure and Respiratory Effects Ixxxiii
Figure ES-1 Evidence for relationships of short-term and long-term nitrogen dioxide
exposure with asthma presented as biological pathways. Ixxxiv
Long-Term Nitrogen Dioxide Exposure and Respiratory Effects Ixxxiv
Nitrogen Dioxide Exposure and Other Health Effects Ixxxv
Policy-Relevant Considerations for Health Effects Associated with Nitrogen Dioxide Exposure Ixxxvi
Summary of Major Findings Ixxxvii
CHAPTER 1 INTEGRATED SUMMARY 1-1
1.1 Purpose and Overview of the Integrated Science Assessment 1-1
1.2 Process for Developing Integrated Science Assessments 1-3
1.3 Content ofthe Integrated Science Assessment 1-5
1.4 From Emissions Sources to Exposure to Nitrogen Dioxide 1-6
1.4.1 Emission Sources and Distribution of Ambient Concentrations 1-6
Figure 1-1 Reactions of oxides of nitrogen species in the ambient air. 1-8
1.4.2 Assessment of Nitrogen Dioxide Exposure in Health Effect Studies 1-9
1.4.3 Factors Potentially Correlated with Nitrogen Dioxide Exposure to Consider in
Evaluating Relationships with Health Effects 1-12
1.5 Health Effects of Nitrogen Dioxide Exposure 1-15
1.5.1 Respiratory Effects 1-16
Figure 1-2 Characterization ofthe evidence for health effects related to nitrogen dioxide
exposure in a mode of action framework. 1-17
1.5.2 Health Effects beyond the Respiratory System 1-22
Table 1-1 Key evidence contributing to causal determinations for nitrogen dioxide
exposure and health effects evaluated in the Integrated Science Assessment
for Oxides of Nitrogen. 1-31
1.6 Policy-Relevant Considerations 1-37
1.6.1 Durations of Nitrogen Dioxide Exposure Associated with Health Effects 1-37
1.6.2 Lag Structure of Relationships between Nitrogen Dioxide Exposure and Health Effects 1-39
1.6.3 Concentration-Response Relationships and Thresholds 1-40
1.6.4 Regional Heterogeneity in Effect Estimates 1-42
1.6.5 Public Health Impact 1-44
1.7 Conclusions 1-48
CHAPTER 2 ATMOSPHERIC CHEMISTRY AND AMBIENT CONCENTRATIONS OF OXIDES
OF NITROGEN 2-1
2.1 Introduction
2.2 Atmospheric Chemistry and Fate
Figure 2-1 Schematic diagram ofthe cycle of reactive, oxidized nitrogen species in the
atmosphere.
2.3 Sources
2.3.1 Overview
2-1
2-1
2-3
2-8
2-8
Figure 2-2 U.S. national average NOx (sum of nitrogen dioxide and nitric oxide)
emissions from 1990 to 2013. 2-9
IV
-------
CONTENTS (Continued)
Figure 2-3 Major sources of NOx (sum of nitrogen dioxide and nitric oxide) emissions
averaged over the U.S. from the 2008 and 2011 National Emissions
Inventories. 2-11
Figure 2-4 Percentage contributions from major sources of the annual NOx (sum of
nitrogen dioxide and nitric oxide) emissions averaged over the 21 largest U.S.
core-based statistical areas with populations greater than 2.5 million
compared to the national average. 2-13
Table 2-1 Source distribution of the annual NOx (sum of nitrogen dioxide and nitric
oxide) emissions in the 21 largest U.S. core-based statistical areas with
populations greater than 2.5 million—2011 National Emissions Inventory. 2-14
2.3.2 Highway Vehicles 2-15
2.3.3 Off-Highway Vehicles and Engines 2-16
2.3.4 Fuel Combustion—Utilities and Other 2-18
Figure 2-5 Fuel Combustion-Other emissions of NOx (sum of nitrogen dioxide and nitric
oxide) versus average ambient January temperature for the 21 largest U.S.
core-based statistical areas with populations greater than 2.5 million. 2-19
2.3.5 Other Anthropogenic Sources 2-19
Table 2-2 Relative contributions to Other Anthropogenic NOx (sum of nitrogen dioxide
and nitric oxide) sources in selected U.S. cities.3 2-21
2.3.6 Biogenics and Wildfires 2-22
2.3.7 Emissions Summary 2-23
2.4 Measurement Methods 2-23
2.4.1 Federal Reference and Equivalent Methods 2-23
2.4.2 Other Methods for Measuring Nitrogen Dioxide 2-26
Figure 2-6 Comparison of nitrogen dioxide measured by cavity attenuated phase shift
spectroscopy to nitrogen dioxide measured by
chemiluminescence/molybdenum oxide catalytic converter for 4 days in
October 2007 in Billerica, MA. 2-28
Figure 2-7 Comparison of nitrogen dioxide measured by quantum cascade-tunable
infrared differential absorption spectroscopy to nitrogen dioxide measured by
chemiluminescence with photolytic converter during April and May 2009 in
Houston, TX. 2-29
2.4.3 Satellite Measurements of Nitrogen Dioxide 2-30
Figure 2-8 Seasonal average tropospheric column abundances for nitrogen dioxide
(1015 molecules/cm2) derived by ozone monitoring instrument for winter (upper
panel) and summer (lower panel) for 2005 to 2007. 2-31
Figure 2-9 Seasonal average tropospheric column abundances for nitrogen dioxide
(1015 molecules/cm2) derived by ozone monitoring instrument for winter (upper
panel) and summer (lower panel) for 2010 to 2012. 2-32
2.4.4 Measurements of Total Oxides of Nitrogen in the Atmosphere 2-33
2.4.5 Ambient Sampling Network Design 2-34
Figure 2-10 Map of monitoring sites for oxides of nitrogen in the U.S. from four networks. 2-35
2.5 Ambient Concentrations of Oxides of Nitrogen 2-36
2.5.1 National-Scale Spatial Variability 2-36
Figure 2-11 U.S. 98th percentiles of 1-hour daily maximum nitrogen dioxide concentrations
for 2011 -2013. 2-38
Figure 2-12 U.S. annual average nitrogen dioxide concentrations for 2013. 2-39
Table 2-3 Summary statistics for 1-hour daily maximum nitrogen dioxide concentrations
(ppb) based on state and local air monitoring stations. 2-40
Table 2-4 Summary statistics for nitrogen dioxide, nitric oxide, and sum of nitrogen
dioxide and nitric oxide annual average concentrations (ppb) based on state
and local air monitoring stations. 2-41
Figure 2-13 Seasonal average surface nitrogen dioxide concentrations in ppb for winter
(upper panel) and summer (lower panel) derived by ozone monitoring
instrument/Goddard Earth Observing System-Chem for 2009-2011. 2-43
-------
CONTENTS (Continued)
2.5.2 Urban-Scale Spatial Variability 2-44
Figure 2-14 Coefficient of divergence for ambient nitrogen dioxide concentrations between
monitor pairs in four U.S. cities. 2-45
Figure 2-15 Coefficient of divergence for ambient nitrogen dioxide concentrations among a
subset of five Los Angeles, California monitors. 2-46
Table 2-5A Percent difference in annual average nitrogen dioxide concentration between
monitors in Boston, Massachussets. 2-47
Table 2-5B Percent difference in annual average nitrogen dioxide concentration between
monitors in Los Angeles, California for 2011. 2-48
2.5.3 Microscale to Neighborhood-Scale Spatial Variability, Including near Roads 2-49
Table 2-6 Summary of near-road nitrogen dioxide concentration gradients from studies
with passive samplers and averaging times of 12 hours to 1 month. 2-50
Table 2-7 Summary of near-road nitrogen dioxide concentration gradients from studies
with averaging times of 1 hour or less. 2-52
Table 2-8 Distribution of differences in higher 1-hour nitrogen dioxide concentrations
10-15 m and lower 1-hour nitrogen dioxide concentrations at 80-100 m from
the road at two locations with heavy traffic. 2-56
Table 2-9 Seasonal and diurnal variation of differences in 1-hour nitrogen dioxide
concentrations 10-15 m and 80-100 m from the road at two locations with
heavy traffic. 2-58
Figure 2-16 Diurnal variation of differences in 1-hour nitrogen dioxide concentrations
10-15 m and 80-100 m from the road in Los Angeles, CA and Detroit, Ml. 2-60
Figure 2-17 Absolute difference in 1-hour nitrogen dioxide concentrations 10-15 m and
80-100 m from the road in Los Angeles, CA and Detroit, Ml. 2-61
Figure 2-18 Influence of nitrogen dioxide concentration magnitude on the ratio of nitrogen
dioxide concentrations at <1 m from the road (nearest concentration) to
concentrations at 200-350 m (farthest concentration) for 1-week averaging
times in rural Wales. 2-63
Figure 2-19 Percentage difference in 1-hour nitrogen dioxide concentration between
10-15 m distance and 80-100 m distance from a road with heavy traffic in two
U.S. cities. 2-65
Table 2-10 Comparison of nitrogen dioxide concentrations at U.S. near-road and non-
near-road monitors for 2014. 2-67
Table 2-11ARoadside and urban background nitrogen dioxide concentrations in London,
U.K. 2010-2012. 2-73
Table 2-11B Roadside and urban background nitrogen dioxide concentrations in London,
U.K. 2004-2006. 2-74
Table 2-12 Selected nitrogen dioxide measurements with potential nonhighway source
influences for 2014. 2-79
2.5.4 Seasonal, Weekday/Weekend, and Diurnal Trends 2-82
Figure 2-20 January and July hourly profiles of nitric oxide and nitrogen dioxide (ppb) for
the site in Atlanta, GA with the highest 1-hour nitrogen dioxide concentrations. 2-82
Figure 2-21 Weekend/weekday hourly profiles of nitric oxide and nitrogen dioxide (ppb) for
the site in Atlanta, GA with the highest nitrogen dioxide concentrations. 2-84
2.5.5 Multiyear Trends in Ambient Measurements of Oxides of Nitrogen 2-84
Figure 2-22 U.S. national average of annual 98th percentile of 1-hour daily maximum
nitrogen dioxide concentration at 24 sites, 1980-2012. 2-85
Figure 2-23 Trend in nitrogen dioxide concentrations at Elizabeth Lab monitoring site near
New Jersey Turnpike 1980-2014. 2-86
2.5.6 Background Concentrations 2-86
2.6 Conclusions 2-88
VI
-------
CONTENTS (Continued)
CHAPTER 3 EXPOSURE TO OXIDES OF NITROGEN
3.1 Introduction
3.2 Methodological Considerations for Use of Exposure Data
3.2.1 Measurement
3.2.2 Modeling
3.2.3 Choice of Exposure Metrics in Epidemiologic Studies
Table 3-1 Summary of exposure estimation methods, their typical use in nitrogen dioxide
epidemiologic studies, and related errors and uncertainties.
Figure 3-1 Average nitrogen dioxide concentrations measured in studies using different
monitor siting.
3.3 Characterization of Nitrogen Dioxide Exposures
3.3.1 Nitrogen Dioxide Concentration as an Indicator of Source-Based Mixtures
Figure 3-2 Spatial variability in concentrations of near-road nitrogen dioxide, nitric oxide,
the sum of nitric oxide and nitrogen dioxide, carbon monoxide, and other near-
road pollutants.
Table 3-2
Table 3-3 Summary (mean, range) within 300 m of monitoring sites, by site type, in a
spatially dense monitoring campaign in New York City, NY, based on 2-week
integrated samples per season.
3.3.2 Indoor Dynamics
Table 3-4 Indoor nitrogen dioxide and nitrous acid concentrations in the presence and
absence of combustion.
3.4 Exposure Assessment and Epidemiologic Inference
3.4.1 Conceptual Model of Total Personal Exposure
3.4.2 Personal-Ambient Relationships and Nonambient Exposures
Table 3-5 Ambient, outdoor, transport, indoor, and personal nitrogen dioxide
measurements (ppb) across studies.
Table 3-6
Table 3-7
Correlations between measured nitrogen dioxide concentrations from
personal, outdoor, indoor, and ambient monitors.
Metaregression results from 15 studies examining the relationship between
personal nitrogen dioxide exposure measurements and ambient
concentrations.
3.4.3 Factors Contributing to Error in Estimating Exposure to Ambient Nitrogen Dioxide
Figure 3-3 Distribution of time sample population spends in various environments, from
the U.S. National Human Activity Pattern Survey (all ages).
Table 3-8
Table 3-9
Figure 3-4
Figure 3-5
Mean fraction of time spent in outdoor locations by various age groups in the
National Human Activity Pattern Survey study.
Mean ventilation rates (L/min) at different activity levels for different age
groups.
Regional-scale variability in nitrogen dioxide for urban and rural area data
across the United Kingdom.
Urban-scale variability in nitrogen dioxide and the sum of nitric oxide and
nitrogen dioxide in Atlanta, GA.
3.4.4 Confounding
Table 3-10 Synthesis of nitrogen dioxide ambient-ambient copollutant correlations from
measurements reported in the literature.
Figure 3-6 Summary of temporal nitrogen dioxide-copollutant correlation coefficients from
measurements reported in studies listed in Table 3-10, sorted by temporal
averaging period.
Table 3-11 Pearson correlation coefficients between ambient nitrogen dioxide and
personal copollutants.
Table 3-12 Pearson correlation coefficients between personal nitrogen dioxide and
ambient copollutants.
_3-2
_3-2
_3-5
.3-19
.3-20
.3-22
.3-24
3-24
3-27
Near- and on-road measurements of nitrogen dioxide, nitric oxide, and the
sum of nitric oxide and nitrogen dioxide. 3-29
.3-34
.3-36
.3-37
.3-40
.3-41
.3-43
.3-45
3-52
.3-55
.3-55
.3-56
.3-58
.3-59
.3-61
.3-62
.3-63
.3-67
.3-80
.3-86
3-86
VII
-------
CONTENTS (Continued)
Table 3-13 Pearson correlation coefficients between personal nitrogen dioxide and
personal copollutants. 3-87
Table 3-14 Correlation coefficients between indoor nitrogen dioxide and indoor
copollutants. 3-88
3.4.5 Implications for Epidemiologic Studies of Different Designs 3-93
Table 3-15 The influence of exposure metrics on error in health effect estimates. 3-97
3.5 Conclusions 3-106
CHAPTER 4 DOSIMETRY AND MODES OF ACTION FOR INHALED OXIDES OF NITROGEN
4-1
4.1 Introduction 4-1
4.2 Dosimetry of Inhaled Oxides of Nitrogen 4-2
4.2.1 Introduction 4-2
4.2.2 Dosimetry of Nitrogen Dioxide 4-3
Table 4-1 Small molecular weight antioxidant concentrations in epithelial lining fluid and
predicted penetration distances for nitrogen dioxide. 4-10
4.2.3 Dosimetry of Nitric Oxide 4-22
4.2.4 Summary of Dosimetry 4-25
4.3 Modes of Action for Inhaled Oxides of Nitrogen 4-27
4.3.1 Introduction 4-27
Table 4-2 Chemical properties of nitrogen dioxide and nitric oxide that contribute to
proposed modes of action. 4-28
4.3.2 Nitrogen Dioxide 4-30
4.3.3 Nitric Oxide 4-54
4.3.4 Metabolites of Nitric Oxide and Nitrogen Dioxide 4-58
4.3.5 Mode of Action Framework 4-61
Figure 4-1 Summary of evidence for the mode of action linking short-term exposure to
nitrogen dioxide and respiratory effects. 4-62
Figure 4-2 Summary of evidence for the mode of action linking long-term exposure to
nitrogen dioxide and respiratory effects. 4-64
Figure 4-3 Summary of evidence for the mode of action linking exposure to nitrogen
dioxide with extrapulmonary effects. 4-66
Figure 4-4 Summary of evidence for the mode of action linking exposure to nitric oxide
with extrapulmonary effects. 4-68
4.4 Summary 4-68
CHAPTER 5 INTEGRATED HEALTH EFFECTS OF SHORT-TERM EXPOSURE TO OXIDES
OF NITROGEN 5-1
5.1 Introduction 5-1
5.1.1 Scope of Chapter 5-1
5.1.2 Evidence Evaluation and Integration to Form Causal Determinations 5-2
5.2 Respiratory Effects 5-6
5.2.1 Introduction 5-6
5.2.2 Asthma Exacerbation 5-7
Table 5-1 Resting exposures to nitrogen dioxide and airway responsiveness in
individuals with asthma. 5-14
Table 5-2 Exercising exposures to nitrogen dioxide and airway responsiveness in
individuals with asthma. 5-15
Table 5-3 Fraction of individuals with asthma having nitrogen dioxide-induced increase
in airway responsiveness to a nonspecific challenge. 5-18
Figure 5-1 Change in provocative dose due to exposure to nitrogen dioxide in resting
individuals with asthma. 5-20
Figure 5-2 Log-normal distribution of change in provocative dose due to exposure to
nitrogen dioxide in resting individuals with asthma. 5-22
vm
-------
CONTENTS (Continued)
Table 5-4
Figure 5-3
Table 5-5
Table 5-6
Table 5-7
Figure 5-4
Table 5-8
Table 5-9
Mean and upper percentile concentrations of nitrogen dioxide in epidemiologic
studies of lung function in populations with asthma.
5-29
Associations of nitrogen dioxide ambient concentrations or personal exposure
with percentage change in forced expiratory volume in 1 second (top plot) and
change in percent predicted forced expiratory volume in 1 second (bottom
plot) in children and adults with asthma.
Epidemiologic studies of lung function in children and adults with asthma.
Characteristics of controlled human exposure studies of lung function in
individuals with asthma.
Mean and upper percentile concentrations of nitrogen dioxide in epidemiologic
studies of respiratory symptoms in populations with asthma.
Associations of ambient nitrogen dioxide concentrations with respiratory
symptoms and asthma medication use in children with asthma.
Epidemiologic studies of respiratory symptoms and asthma medication use in
children with asthma.
Characteristics of controlled human exposure studies of respiratory symptoms
in individuals with asthma.
Table 5-10 Mean and upper percentile concentrations of oxides of nitrogen in studies of
asthma hospital admissions and emergency department visits.
Table 5-11 Copollutant model results from Iskandar et al. (2012) for a 20-ppb increase in
24-h average nitrogen dioxide (NCte) concentrations and a 40-ppb increase in
24-h average NOx (sum of NO and NCb) concentrations.
Figure 5-5
Figure 5-6
Figure 5-7
Table 5-12
Table 5-13
Table 5-14
Table 5-15
Figure 5-8
Table 5-16
Concentration-response function for the association between 3-day average
(lag 0-2) nitrogen dioxide concentrations and emergency department visits for
pediatric asthma in the Atlanta, GA area.
Rate ratio and 95% confidence intervals for asthma-related emergency
department visits in single-pollutant and joint effect models for each pollutant
at lag 0-2 days.
Percentage increase in asthma hospital admissions and emergency
department visits in relation to short-term increases in ambient nitrogen
dioxide concentrations.
Corresponding risk estimates for studies presented in Figure 5-7.
Characteristics of controlled human exposure studies of pulmonary
inflammation in populations with asthma.
Characteristics of animal toxicological studies of pulmonary inflammation.
Mean and upper percentile concentrations of nitrogen dioxide in epidemiologic
studies of pulmonary inflammation and oxidative stress in populations with
asthma.
Associations of personal or ambient nitrogen dioxide with exhaled nitric oxide
in populations with asthma.
.5-33
.5-34
.5-49
.5-51
.5-56
.5-57
.5-70
.5-72
.5-81
.5-87
5-90
_5-93
_5-94
_5-98
5-103
Epidemiologic studies of pulmonary inflammation and oxidative stress in
children and adults with asthma.
5.2.3 Allergy Exacerbation
Table 5-17 Epidemiologic studies of allergy exacerbation.
5.2.4 Exacerbation of Chronic Obstructive Pulmonary Disease
Table 5-18 Epidemiologic panel studies of adults with chronic obstructive pulmonary
disease.
Table 5-19 Characteristics of controlled human exposure studies of adults with chronic
obstructive pulmonary disease.
Table 5-20 Mean and upper percentile concentrations of nitrogen dioxide in studies of
hospital admission and emergency department visits for chronic obstructive
pulmonary disease.
Figure 5-9 Percentage increase in chronic obstructive pulmonary disease hospital
admissions and emergency department visits in relation to nitrogen dioxide
concentrations.
Table 5-21 Corresponding risk estimate for studies presented in Figure 5-9.
5-105
5-107
5-108
5-123
5-125
5-127
5-129
5-132
5-134
5-139
5-140
IX
-------
CONTENTS (Continued)
5.2.5 Respiratory Infection 5-142
Table 5-22 Characteristics of experimental studies of susceptibility to infection. 5-143
Table 5-23 Epidemiologic studies of respiratory infections reported or diagnosed in
children. 5-149
Table 5-24 Mean and upper percentile concentrations of nitrogen dioxide in studies of
hospital admissions and emergency department visits for respiratory infection._ 5-153
Figure 5-10 Percentage increase in respiratory infection-related hospital admissions and
emergency department visits in relation to nitrogen dioxide concentrations. 5-160
Table 5-25
Table 5-26
Corresponding risk estimate for studies presented in Figure 5-10.
Characteristics of experimental studies of subclinical lung host defense
effects.
_ 5-161
_ 5-164
_ 5-169
_ 5-171
Figure 5-11 Risk ratio and 95% confidence intervals for associations between various lag
1 day nitrogen dioxide metrics and respiratory emergency department visits. 5-179
Figure 5-12 Spatial correlations for nitrogen dioxide metrics in the Atlanta, GAarea. 5-180
Figure 5-13 Percentage increase in all respiratory disease hospital admissions and
emergency department visits in relation to nitrogen dioxide concentrations. 5-181
5.2.6 Aggregated Respiratory Conditions
Table 5-27 Mean and upper percentile concentrations of nitrogen dioxide in studies of
hospital admissions and emergency department visits for aggregated
respiratory conditions.
Table 5-28 Corresponding risk estimate for studies presented in Figure 5-13.
5.2.7 Respiratory Effects in Healthy Populations
Table 5-29 Mean and upper percentile oxides of nitrogen concentrations in epidemiologic
studies of lung function in the general population.
Table 5-30 Epidemiologic studies of lung function in children and adults in the general
population.
Table 5-31
Characteristics of controlled human exposure studies of lung function and
respiratory symptoms in healthy adults.
Table 5-32 Mean and upper percentile concentrations of nitrogen dioxide in epidemiologic
studies of respiratory symptoms in children in the general population.
Table 5-33 Epidemiologic studies of respiratory symptoms in children in the general
population.
Table 5-34 Mean and upper percentile concentrations of oxides of nitrogen in
epidemiologic studies of pulmonary inflammation and oxidative stress in the
general population.
Figure 5-14 Associations between ambient nitrogen dioxide concentrations and exhaled
nitric oxide among children and adults in the general population.
Table 5-35 Epidemiologic studies of pulmonary inflammation, injury, and oxidative stress
in children and adults in the general population.
Table 5-36 Characteristics of controlled human exposure studies of pulmonary
inflammation, injury, and oxidative stress in healthy adults.
5.2.8
Table 5-37 Characteristics of animal toxicological studies of pulmonary inflammation,
injury, and oxidative stress.
Respiratory Mortality
Figure 5-15 City-specific concentration-response curves of nitrogen dioxide and daily
chronic obstructive pulmonary disease mortality in four Chinese cities.
5.2.9 Summary and Causal Determination
Figure 5-16 Associations of ambient or personal nitrogen dioxide with respiratory effects
adjusted for fine particulate matter, elemental/black carbon, or particle number
concentration/ultrafine particles.
5-182
5-184
5-186
5-190
5-203
5-207
5-209
5-214
5-217
5-218
5-230
5-232
5-236
5-239
5-239
5-248
5-249
Figure 5-17 Associations of ambient nitrogen dioxide with respiratory effects adjusted for a
volatile organic compound or carbon monoxide.
Table 5-38 Corresponding effect estimates for nitrogen dioxide-associated respiratory
effects in single- and copollutant models presented in Figures 5-16 and 5-17. _ 5-250
Table 5-39 Summary of evidence for a causal relationship between short-term nitrogen
dioxide exposure and respiratory effects. 5-253
-------
CONTENTS (Continued)
5.3 Cardiovascular Effects 5-258
5.3.1 Introduction 5-258
5.3.2 Myocardial Infarction 5-259
Figure 5-18 Associations between short-term exposure to oxides of nitrogen and hospital
admissions for ischemic heart disease. 5-261
Table 5-40 Corresponding risk estimates for hospital admissions for ischemic heart
disease for studies presented in Figure 5-18. 5-262
Table 5-41 Epidemiologic studies of ST-segment amplitude. 5-271
5.3.3 Arrhythmia and Cardiac Arrest 5-273
Table 5-42 Epidemiologic studies of arrhythmia and cardiac arrest. 5-274
Table 5-43 Epidemiologic studies of out-of-hospital cardiac arrest. 5-276
5.3.4 Cerebrovascular Disease and Stroke 5-278
Figure 5-19 Associations between short-term exposure to oxides of nitrogen and hospital
admissions for cerebrovascular disease and stroke. 5-280
Table 5-44 Corresponding risk estimates for hospital admissions for cerebrovascular
disease and stroke for studies presented in Figure 5-19. 5-281
5.3.5 Decompensation of Heart Failure 5-287
5.3.6 Increased Blood Pressure and Hypertension 5-287
Table 5-45 Epidemiologic studies of blood pressure. 5-288
5.3.7 Venous Thromboembolism 5-292
5.3.8 Aggregated Cardiovascular Effects 5-292
Figure 5-20 Associations between short-term exposure to oxides of nitrogen and hospital
admissions for all cardiovascular disease. 5-295
Table 5-46 Corresponding effect estimates for hospital admissions for all cardiovascular
disease studies presented in Figure 5-20. 5-296
5.3.9 Cardiovascular Mortality 5-299
Figure 5-21 Pooled concentration-response curve for nitrogen dioxide and daily stroke
mortality in eight Chinese cities for lag 0-1 day. 5-302
5.3.10 Subclinical Effects Underlying Cardiovascular Effects 5-302
Table 5-47 Epidemiologic studies of heart rate/heart rate variability. 5-304
Table 5-48 Epidemiologic studies of QT-interval duration. 5-314
Table 5-49 Epidemiologic studies of biomarkers of cardiovascular effects. 5-317
Table 5-50 Characteristics of controlled human exposure studies of cardiovascular
effects. 5-328
Table 5-51 Characteristics of animal toxicological studies of cardiovascular effects. 5-332
5.3.11 Summary and Causal Determination 5-334
Table 5-52 Summary of evidence, which is suggestive of, but not sufficient to infer, a
causal relationship between short-term nitrogen dioxide exposure and
cardiovascular effects. 5-339
5.4 Total Mortality 5-342
5.4.1 Introduction and Summary of the 2008 Integrated Science Assessment for Oxides of
Nitrogen 5-342
5.4.2 Associations between Short-Term Nitrogen Dioxide Exposure and Mortality 5-343
5.4.3 Associations between Short-term Nitrogen Dioxide Exposure and Mortality in All-Year
Analyses 5-344
Table 5-53 Air quality characteristics of studies evaluated in the 2008 Integrated Science
Assessment for Oxides of Nitrogen and recently published multicity and select
single-city studies. 5-345
Figure 5-22 Summary of multicity studies that examined the association between
short-term nitrogen dioxide exposure and total mortality. 5-348
Table 5-54 Corresponding percentage increase in total mortality for Figure 5-22. 5-349
Figure 5-23 Percentage increase in total, cardiovascular, and respiratory mortality from
multicity studies in relation to ambient nitrogen dioxide concentrations. 5-350
Table 5-55 Corresponding percentage increase for Figure 5-23. 5-351
XI
-------
CONTENTS (Continued)
5.4.4 Potential Confounding of the Nitrogen Dioxide-Mortality Relationship 5-352
Table 5-56 Percentage increase in total and cause-specific mortality for a 20-ppb
increase in 24-hour average nitrogen dioxide concentrations in single- and
copollutant models with particulate matter in all-year analyses or ozone in
summer season analyses. 5-354
5.4.5 Modification of the Nitrogen Dioxide-Mortality Relationship 5-356
5.4.6 Potential Seasonal Differences in the Nitrogen Dioxide-Mortality Relationship 5-357
5.4.7 Nitrogen Dioxide-Mortality Concentration-Response Relationship and Related Issues 5-359
Figure 5-24 Percentage increase in total and cause-specific mortality associated with
short-term increases in ambient nitrogen dioxide concentration at single day
lags, individual lag days of a constrained polynomial distributed lag model,
and multiday lags of an unconstrained distributed lag model. 5-360
Figure 5-25 Percentage increase in total and cause-specific mortality associated with
short-term increases in ambient nitrogen dioxide concentration in single- and
multi-day lag models in a multicity study in China. 5-361
Figure 5-26 Flexible ambient concentration-response relationship between short-term
nitrogen dioxide concentrations and mortality at Lag Day 1. 5-363
Figure 5-27 China Air Pollution and Health Effects Study concentration-response curve for
the association between total and cause-specific mortality and
24-hour average nitrogen dioxide concentrations at lag 0-1 days. 5-364
Figure 5-28 Concentration-response curve for association between total mortality and
24-hour average nitrogen dioxide concentrations at lag 0-1 days in the four
cities of the Public Health and Air Pollution in Asia study. 5-366
5.4.8 Summary and Causal Determination 5-367
Table 5-57 Summary of evidence, which is suggestive of, but not sufficient to infer, a
causal relationship between short-term nitrogen dioxide exposure and total
mortality. 5-369
CHAPTER 6 INTEGRATED HEALTH EFFECTS OF LONG-TERM EXPOSURE TO OXIDES OF
NITROGEN 6-1
6.1 Scope and Issues Considered in Health Effects Assessment 6-1
6.1.1 Scope of Chapter 6-1
6.1.2 Evidence Evaluation and Integration to Form Causal Determinations 6-2
6.2 Respiratory Effects 6-4
6.2.1 Introduction 6-4
6.2.2 Development of Asthma or Chronic Bronchitis 6-5
Figure 6-1 Associations of long-term exposure to oxides of nitrogen with asthma
incidence in longitudinal cohort studies of children. 6-6
Table 6-1 Longitudinal studies of long-term exposure to oxides of nitrogen and asthma
incidence in children. 6-7
Figure 6-2 Overall and age-specific associations between annual average air pollutant
concentrations at the birth residence and asthma during the first 8 years of
life. 6-15
Table 6-2 Characteristics of animal toxicological studies of long-term nitrogen dioxide
exposure and respiratory effects. 6-27
6.2.3 Severity of Asthma, Chronic Bronchitis, and Chronic Obstructive Pulmonary Disease:
Respiratory Symptoms and Hospital Admissions 6-30
Figure 6-3 Concentration-response relationships between asthma-related effects and
indoor nitrogen dioxide illustrated with constrained, natural spline and
threshold functions in hierarchical ordered logistic regression models. 6-32
Table 6-3 Longitudinal studies of long-term nitrogen dioxide exposure and respiratory
symptoms in children. 6-33
Figure 6-4 Within-community odds ratios for bronchitis symptoms associated with
nitrogen dioxide adjusted for a copollutant in the 12 communities of the
Children's Health Study. 6-37
6.2.4 Development of Allergic Disease 6-40
XII
-------
CONTENTS (Continued)
6.2.5 Lung Function and Lung Development 6-42
Figure 6-5 Associations of oxides of nitrogen with lung function or lung development from
longitudinal studies of children. 6-43
Table 6-4 Longitudinal studies of long-term nitogen dioxide exposure and lung function
and lung development in children. 6-44
Figure 6-6 Community-specific average growth in forced expiratory volume in 1 second
(mL) among girls and boys from 1993 to 2001, plotted against average
nitrogen dioxide concentrations from 1994 through 2000. 6-50
Figure 6-7 Community-specific proportion of 18-year-olds with a forced expiratory volume
in 1 second below 80% of the predicted value, plotted against the average
concentrations of nitrogen dioxide from 1994 through 2000. 6-51
6.2.6 Changes in Lung Morphology 6-56
6.2.7 Respiratory Infection 6-58
6.2.8 Chronic Obstructive Pulmonary Disease 6-61
6.2.9 Summary and Causal Determination 6-62
Table 6-5 Summary of key evidence for a likely to be a causal relationship between
long-term nitrogen dioxide exposure and respiratory effects. 6-68
6.3 Cardiovascular Effects and Diabetes 6-72
6.3.1 Introduction 6-72
6.3.2 Heart Disease 6-73
Table 6-6 Epidemiologic studies of long-term exposure to oxides of nitrogen and heart
disease. 6-75
6.3.3 Cerebrovascular Disease and Stroke 6-79
Table 6-7 Epidemiologic studies of long-term exposure to oxides of nitrogen and
cerebrovascular disease or stroke. 6-80
6.3.4 Hypertension 6-83
Table 6-8 Epidemiologic studies of long-term exposure to oxides of nitrogen and
hypertension and blood pressure. 6-85
6.3.5 Cardiovascular Mortality 6-88
6.3.6 Markers of Cardiovascular Disease or Mortality 6-88
Table 6-9 Characteristics of toxicological studies of long-term nitrogen dioxide exposure
and cardiovascular effects. 6-90
6.3.7 Diabetes 6-91
Table 6-10 Epidemiologic studies of long-term exposure to oxides of nitrogen and
diabetes or diabetes-related effects. 6-92
6.3.8 Subclinical Effects Underlying Cardiovascular Disease and Diabetes 6-96
6.3.9 Summary and Causal Determination 6-97
Table 6-11 Summary of evidence, which is suggestive of, but not sufficient to infer, a
causal relationship between long-term nitrogen dioxide exposure and
cardiovascular effects and diabetes. 6-100
6.4 Reproductive and Developmental Effects 6-103
6.4.1 Introduction 6-103
6.4.2 Fertility, Reproduction, and Pregnancy 6-105
Table 6-12 Key epidemiologic studies of oxides of nitrogen and reproductive and
developmental effects. 6-110
6.4.3 Birth Outcomes 6-116
6.4.4 Postnatal Development 6-127
Table 6-13 Characteristics of toxicological studies of nitrogen dioxide exposure and
reproductive and developmental effects. 6-134
6.4.5 Summary and Causal Determination 6-135
Table 6-14 Summary of evidence supporting the causal determinations for relationships
between long-term nitrogen dioxide exposure and reproductive and
developmental effects. 6-137
xm
-------
CONTENTS (Continued)
6.5 Total Mortality 6-140
6.5.1 Introduction and Review of Evidence from 2008 Integrated Science Assessment for
Oxides of Nitrogen 6-140
6.5.2 Recent Evidence for Mortality from Long-Term Exposure to Oxides of Nitrogen 6-142
Figure 6-8 Association between long-term exposure to oxides of nitrogen and total
mortality. 6-147
Table 6-15 Corresponding risk estimates for Figure 6-8. 6-148
Figure 6-9 Associations between long-term exposure to oxides of nitrogen and
cardiovascular mortality. 6-150
Table 6-16 Corresponding risk estimates for Figure 6-9. 6-151
Figure 6-10 Associations between long-term exposure to oxides of nitrogen and
respiratory mortality. 6-154
Table 6-17 Corresponding risk estimates for Figure 6-10. 6-155
6.5.3 Summary and Causal Determination 6-157
Table 6-18 Summary of evidence, which is suggestive of, but not sufficient to infer, a
causal relationship between long-term nitrogen dioxide exposure and total
mortality. 6-159
6.6 Cancer 6-160
6.6.1 Introduction 6-160
6.6.2 Lung Cancer 6-161
Table 6-19 Characteristics of toxicological studies of carcinogenicity and genotoxicity with
exposure to nitrogen dioxide. 6-167
6.6.3 Leukemia Incidence and Mortality 6-169
6.6.4 Bladder Cancer Incidence and Mortality 6-170
6.6.5 Breast Cancer Incidence 6-171
6.6.6 Prostate Cancer Incidence 6-171
6.6.7 Other Cancer Incidences and Mortality 6-172
6.6.8 Genotoxicity 6-172
6.6.9 Summary and Causal Determination 6-174
Table 6-20 Summary of evidence, which is suggestive of, but not sufficient to infer, a
causal relationship between long-term nitrogen dioxide exposure and cancer. _ 6-176
CHAPTER 7 POPULATIONS AND LIFESTAGES POTENTIALLY AT RISK FOR HEALTH
EFFECTS RELATED TO NITROGEN DIOXIDE EXPOSURE 7-1
7.1 Introduction 7-1
7.2 Approach to Evaluating and Characterizing the Evidence for At-Risk Factors 7-2
Table 7-1 Characterization of evidence for factors potentially increasing the risk for
nitrogen dioxide-related health effects. 7-3
7.3 Pre-Existing Disease/Conditions 7-4
Table 7-2 Prevalence of selected respiratory, cardiovascular, and metabolic diseases
and disorders among adults by age and region in the United States in 2012. 7-5
7.3.1 Asthma 7-5
Table 7-3 Controlled human exposure studies evaluating pre-existing asthma. 7-8
Table 7-4 Epidemiologic studies evaluating pre-existing asthma. 7-9
7.3.2 Chronic Obstructive Pulmonary Disease 7-9
Table 7-5 Controlled human exposure studies evaluating pre-existing chronic obstructive
pulmonary disease. 7-11
Table 7-6 Epidemiologic studies evaluating pre-existing chronic obstructive pulmonary
disease. 7-12
7.3.3 Cardiovascular Disease 7-12
Table 7-7 Epidemiologic studies evaluating pre-existing cardiovascular disease. 7-14
Table 7-8 Controlled human exposure and toxicological studies informing risk due to
pre-existing cardiovascular disease. 7-16
xiv
-------
CONTENTS (Continued)
7.3.4 Diabetes 7-16
Table 7-9 Epidemiologic studies evaluating pre-existing diabetes. 7-18
7.3.5 Obesity 7-19
Table 7-10 Toxicological study evaluating pre-existing obesity. 7-20
Table 7-11 Epidemiologic studies evaluating pre-existing obesity. 7-21
7.4 Genetic Factors 7-23
Table 7-12 Epidemiologic studies evaluating genetic factors. 7-26
7.5 Sociodemographic Factors 7-29
7.5.1 Lifestage 7-29
Table 7-13 Epidemiologic studies evaluating childhood lifestage. 7-33
Table 7-14 Toxicological studies evaluating childhood lifestage. 7-35
Table 7-15 Epidemiologic studies evaluating older adult lifestage. 7-38
Table 7-16 Controlled human exposure studies informing risk for older adult lifestage. 7-42
7.5.2 Socioeconomic Status 7-43
Table 7-17 Epidemiologic studies evaluating socioeconomic status. 7-46
7.5.3 Race/Ethnicity 7-50
Table 7-18 Epidemiologic studies evaluating race/ethnicity. 7-52
7.5.4 Sex 7-53
Table 7-19 Epidemiologic studies evaluating sex. 7-55
7.5.5 Residence in Urban Areas 7-61
Table 7-20 Epidemiologic studies evaluating urban residence. 7-62
7.5.6 Proximity to Roadways 7-62
Figure 7-1 Map of population density in Los Angeles, CA, in relation to primary and
secondary roads. 7-64
Table 7-21 Epidemiologic studies evaluating proximity to roadways. 7-67
7.6 Behavioral and Other Factors 7-68
7.6.1 Diet 7-68
Table 7-22 Controlled human exposure and toxicological studies evaluating diet. 7-70
Table 7-23 Epidemiologic studies evaluating diet. 7-71
7.6.2 Smoking 7-71
Table 7-24 Controlled human exposure study evaluating smoking. 7-72
Table 7-25 Epidemiologic studies evaluating smoking status. 7-73
7.6.3 Physical Activity 7-76
Table 7-26 Epidemiologic studies evaluating physical activity. 7-77
7.7 Conclusions 7-77
Table 7-27 Summary of evidence for potential increased nitrogen dioxide exposure and
increased risk of nitrogen dioxide-related health effects. 7-79
APPENDIX: EVALUATION OF STUDIES ON HEALTH EFFECTS OF OXIDES OF NITROGEN A-1
Table A-1 Scientific considerations for evaluating the strength of inference from studies
on the health effects of oxides of nitrogen. A-2
REFERENCES R-1
xv
-------
INTEGRATED SCIENCE ASSESSMENT TEAM FOR
OXIDES OF NITROGEN
Executive Direction
Dr. John Vandenberg (Director)—National Center for Environmental Assessment-RTF
Division, Office of Research and Development, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Ms. Debra Walsh (Deputy Director)—National Center for Environmental Assessment-RTP
Division, Office of Research and Development, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Reeder Sams II (Acting Deputy Director)—National Center for Environmental
Assessment-RTP Division, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Mary Ross (Branch Chief)—National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Steven J. Dutton (Acting Branch Chief)—National Center for Environmental
Assessment, Office of Research and Development, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Ellen Kirrane (Acting Branch Chief)—National Center for Environmental Assessment,
Office of Research and Development, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Scientific Staff
Dr. Molini M. Patel (Team Leader, Integrated Science Assessment for Oxides of
Nitrogen)—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Breanna Alman—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. James Brown—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Barbara Buckley—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Evan Coffman—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Laura Datko-Williams—Oak Ridge Institute for Science and Education, National Center
for Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
xvi
-------
Dr. Erin Hines—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Ellen Kirrane—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Dennis Kotchmar—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Thomas Long—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Thomas Luben—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Stephen McDow—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Connie Meacham—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jennifer Nichols—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Michelle Oakes—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Elizabeth Oesterling Owens—National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Joseph P. Pinto—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Kristen Rappazzo—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Jennifer Richmond-Bryant—National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Mr. Jason Sacks—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Tina Stevens—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. David Svendsgaard—National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Lisa Vinikoor-Imler—National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
XVII
-------
Ms. Brianna Young—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Technical Support Staff
Ms. Marieka Boyd—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Kenneth J. Breito—Senior Environmental Employment Program, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Nathan Ellenfield—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Charlene Finley—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Carolyn Gatling—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. William Griffin—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Gerald Gurevich—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Saturo Ito—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Katie Jelen—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Ryan Jones—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Emily Lau—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Diane LeBlond Cummings—Oak Ridge Institute for Science and Education, National
Center for Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Danielle Moore—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
XVlll
-------
Ms. Candis O'Neal Edwards—Oak Ridge Institute for Science and Education, National
Center for Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Kyle Painter—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Sandy Pham—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Olivia Philpott—Senior Environmental Employment Program, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Richard N. Wilson—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Barbara Wright—Senior Environmental Employment Program, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
xix
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
Authors
Dr. Molini M. Patel (Team Leader, Integrated Science Assessment for Oxides of
Nitrogen)—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Breanna Alman—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. James Brown—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Barbara Buckley—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Evan Coffman—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Laura Datko-Williams—Oak Ridge Institute for Science and Education, National Center
for Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Rachelle Duvall—National Exposure Research Laboratory, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Erin Hines—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Ellen Kirrane—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Dennis Kotchmar—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Thomas Long—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Thomas Luben—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Stephen McDow—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jennifer Nichols—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Michelle Oakes—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Elizabeth Oesterling Owens—National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Cincinnati, Ohio
xx
-------
Dr. Jennifer Peel—Department of Environmental and Radiological Health Sciences,
Colorado School of Public Health, Colorado State University, Fort Collins, CO
Dr. Joseph P. Pinto—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Edward Postlethwait—Department of Environmental Health Sciences, Ryals School for
Public Health, University of Alabama at Birmingham, Birmingham, AL
Dr. Kristen Rappazzo—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Jennifer Richmond-Bryant—National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Mr. Jason Sacks—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Giuseppe Squadrito—Department of Environmental Health Sciences, Ryals School for
Public Health, University of Alabama at Birmingham, Birmingham, AL
Dr. Tina Stevens—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. David Svendsgaard—National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. George Thurston—Department of Environmental Medicine, New York University
School of Medicine, Tuxedo, NY
Dr. Lisa Vinikoor-Imler—National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Gregory Wellenius—Department of Community Health (Epidemiology Section), Brown
University, Providence, RI
Ms. Brianna Young—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Contributors
Mr. Adam Benson—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Rachel Housego—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Meagan Madden—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
xxi
-------
Ms. April Maxwell—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Ihab Mikati—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Lok Namsal—Universities Space Research Association, NASA Goddard Space Flight
Center, Greenbelt, MD
Dr. Havala Pye—National Exposure Research Laboratory, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Adrien Wilkie—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Reviewers
Mr. Chad Bailey—Office of Transportation and Air Quality, Office of Air and Radiation,
U.S. Environmental Protection Agency, Ann Arbor, MI
Dr. Lisa Baxter—National Exposure Research Laboratory, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Melinda Beaver—National Exposure Research Laboratory, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Roger Erode—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Philip Bromberg—School of Medicine, University of North Carolina, Chapel Hill, NC
Mr. Matthew Davis—Office of Children's Health Protection, U.S. Environmental Protection
Agency, Washington, DC
Ms. Laurel Driver—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Steven J. Dutton—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Mark Frampton—Pulmonary and Critical Care Division, Department of Medicine,
University of Rochester School of Medicine, Rochester, NY
Dr. Terry Gordon—Department of Environmental Medicine, New York University School
of Medicine, Tuxedo, NY
Dr. Stephen Graham—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Beth Hassett-Sipple—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Gary Hatch—National Health and Environmental Effects Research Laboratory, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
xxn
-------
Mr. Marc Houyoux—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Scott Jenkins—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Deborah Luecken—National Exposure Research Laboratory, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Joseph McDonald—Office of Transportation and Air Quality and National Risk
Management Research Laboratory, Office of Research and Development,
U.S. Environmental Protection Agency, Cincinnati, OH
Ms. Connie Meacham—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Deirdre Murphy—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. David Orlin—Air and Radiation Law Office, Office of General Counsel, U.S.
Environmental Protection Agency, Washington, DC
Mr. Chris Owen—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jennifer Peel—Department of Environmental and Radiological Health Sciences,
Colorado School of Public Health, Colorado State University, Fort Collins, CO
Dr. Edward Postlethwait—Department of Environmental Health Sciences, Ryals School for
Public Health, University of Alabama at Birmingham, Birmingham, AL
Dr. Havala Pye—National Exposure Research Laboratory, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Venkatesh Rao—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Elise Richman—Office of Children's Health Protection, U.S. Environmental Protection
Agency, Washington, DC
Dr. Mary Ross—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Reeder Sams II—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Stephanie Sarnat—Department of Environmental Health, Rollins School of Public
Health, Emory University, Atlanta, GA
Dr. Matthew Strickland—Department of Environmental Health, Rollins School of Public
Health, Emory University, Atlanta, GA
Dr. John Vandenberg—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Debra Walsh—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Nealson Watkins—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
XXlll
-------
Ms. Larke Williams—Office of Science Policy, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Margaret Zawacki—Office of Transportation and Air Quality, Office of Air and
Radiation, U.S. Environmental Protection Agency, Ann Arbor, MI
xxiv
-------
CLEAN AIR SCIENTIFIC ADVISORY COMMITTEE
OXIDES OF NITROGEN NAAQS REVIEW PANEL
Chair of the Statutory Clean Air Scientific Advisory Committee Appointed by the
Environmental Protection Agency's Administrator
Dr. Ana Diez Roux—Drexel University, Philadelphia, PA
Oxides of Nitrogen Review Panel Members
Dr. H. Christopher Frey (Panel Chair)*—North Carolina State University, Raleigh, NC
Mr. George A. Allen**—Northeast States for Coordinated Air Use Management
(NESCAUM), Boston, MA
Dr. Matthew Campen—University of New Mexico, Albuquerque, NM
Dr. Ronald Cohen—University of California, Berkeley, Berkeley, CA
Dr. Douglas Dockery—Harvard University, Boston, MA
Dr. Philip Fine—South Coast Air Quality Management District, Diamond Bar, CA
Dr. Panos Georgopoulos—Rutgers University, Piscataway, NJ
Dr. Jack Harkema**—Michigan State University, East Lansing, MI
Dr. Michael Jerrett—University of California, Los Angeles, Los Angeles, CA
Dr. Joel Kaufman—University of Washington, Seattle, WA
Dr. Michael T. Kleinman—University of California, Irvine, Irvine, CA
Dr. Timothy V. Larson—University of Washington, Seattle, WA
Dr. Jeremy Sarnat—Emory University, Atlanta, GA
Dr. Richard Schlesinger—Pace University, New York, NY
Dr. Elizabeth A. (Lianne) Sheppard—University of Washington, Seattle, WA
Dr. Helen Suh* *—Northeastern University, Boston, MA
Dr. Ronald Wyzga**—Electric Power Research Institute, Palo Alto, CA
Dr. Junfeng (Jim) Zhang—Duke University, Durham, NC
* Immediate Past statutory CASAC Chair appointed by the EPA Administrator
**Members of the statutory CASAC appointed by the EPA Administrator
Science Advisory Board Staff
Mr. Aaron Yeow (Designated Federal Officer)—U.S. Environmental Protection Agency,
Science Advisory Board (1400R), 1200 Pennsylvania Avenue, NW, Washington, DC
20460-0001, Phone: 202-564-2050, Fax: 202-565-2098, yeow.aaron@epa.gov
XXV
-------
ACRONYMS AND ABBREVIATIONS
Acronym/
Abbreviation
.
8-OhdG
a
a-ATD
A4
AADT
AB
ABI
Abs
ABTS* -
ABTS2-
ACS
ADRB2
AER
AERMOD
AHR
AHSMOG
aj
AIRES
AK
AL
ALKP
ALRI
a.m.
AM
Meaning
radical species
urinary 8-hydroxy-29-
deoxy guano sine
alpha, exposure factor
alpha 1-antitrypsin
deficiency
not classifiable for humans
or animals
annual average daily traffic
Alberta
ankle brachial index
absorbance coefficient
2,2'-azino-bis (3-
ethylbenzothiazoline - 6-
sulfonic acid) radical
2,2'-azino-bis
(3-ethylbenzothiazoline-6-
sulphonic acid)
American Cancer Society
beta-2-adrenergic receptor
air exchange rate
American Meteorological
Society /Environmental
Protection Agency
Regulatory Model
airway hyperresponsiveness
California Seventh-Day
Adventists cohort
air exchange rate
Aerosol Research Inhalation
Epidemiology Study
Alaska
Alabama; alpine
alkaline phosphatase
acute lower respiratory
infection
ante meridiem (before noon)
alveolar macrophages
Acronym/
Abbreviation
ANPR
APEX
APHEA
AQCD
AQI
AQS
AR
AT
ATS
ATSDR
Aug
avg
AW
AZ
R
BAL
BAMSE
BC
EC/EC
BD
BEIS
BHPN
BIR
bkg
BL
BMI
BP
Meaning
advanced notice of public
rulemaking
air pollution exposure
model
Air Pollution and Health: A
European Approach study
air quality criteria document
air quality index
air quality system
airway responsiveness
Atascadero
American Thoracic Society
Agency for Toxic
Substances and Disease
Registry
August
average
area wide
Arizona
beta
bronchoalveolar lavage
Children, Allergy, Milieu,
Stockholm, Epidemiology
Survey
black carbon, British
Colombia
black carbon/elemental
carbon
bronchodilator
Biogenic Emission
Inventory System
N-bis (2-hydroxy-propyl)
nitrosamine
birch
background
bronchial lavage
body mass index
blood pressure
XXVI
-------
Acronym/
Abbreviation
Br
BS
BSA
BTEX
BW
BWHS
CeHe
C
C&RT
Ca2+
CA
Ca
v--a,csm
CAA
CalNex
CAMP
CAMx
CAN
CAP
CAPES
CAPS
CARB
CASAC
CASNET
Cb
Meaning
bromide
black smoke
body surface area
sum of the VOCs benzene,
toluene, ethybenzene,
xylene
body weight; bronchial
wash
Black Women's Health
Study
benzene
degrees Celsius; the product
of microenvironmental
concentration; carbon;
classification and regression
tree
calcium
California; cat allergen
ambient NCh concentration
ambient concentration at a
central site monitor
Clean Air Act
California Research at the
Nexus of Air Quality and
Climate Change
Childhood Asthma
Management Program
Comprehensive Air Quality
Model with Extensions
Canada
concentrated ambient
particle
China Air Pollution and
Health Effects Study
cavity attenuated phase shift
carbachol
Clean Air Scientific
Advisory Committee
Clean Air Status and Trends
Network
NO2 concentration
contribution away from the
influence of the road
Acronym/
Abbreviation
CBV
CBVD
CC16
CDC
Cfar
CFD
CFR
cGMP
CHAD
CHD
Chemilum
CHS
Ci
CI(s)
cIMT
Q
CJ-A
CJ-B
Cf
CL/MC
CL/PC
C1NO
C1N02
cm
CMAQ
^-•near
CO
Coj
CBSA
core-based statistical area
Meaning
cerebro vascular
cerebrovascular disease
club cell protein
Centers for Disease Control
and Prevention
farthest concentration
computational fluid
dynamics
Code of Federal Regulations
cyclic guano sine
monophosphate
Consolidated Human
Activity Database
coronary heart disease
chemiluminescence
Children's Health Study
average NCh concentration
in the z'th microenvironment;
substrate concentrations
confidence interval(s)
carotid intima-media
thickness
average NCh concentration
in the jth microenvironment
Ciudad Juarez—Site A
Ciudad Juarez—Site B
chloride
chemiluminescence
analyzer with a MoOx
catalytic converter
chemiluminescence
analyzer with measurements
from a photolytic converter
nitrosyl chloride
nitryl chloride
centimeter
Community Multiscale Air
Quality
nearest concentration
carbon monoxide; Colorado
ambient exposure to NCh
outdoor concentration
XXVll
-------
Acronym/
Abbreviation
C02
COD
CoH
COLD
COPD
C-R
CRDS
CRP
CS
CT
CTM
CIS
Cu
Cv
cv
CVD
Cx
D
d
D
D.C. Cir
DBF
DC
DEARS
Dec
DEP
DEPcCBP
df
DHA
DJF
Meaning
carbon dioxide
coefficient of divergence
coefficient of haze
cold-dry air
chronic obstructive
pulmonary disease
concentration-response
(relationship)
cavity ring down
spectroscopy
C-reactive protein
central site
Connecticut
chemical transport models
California Teachers Study
copper
NO2 concentration
contribution from vehicles
on a roadway
coefficient of variation
cardiovascular disease
NO2 concentration at a
distance x from a road
molecular diffusion
coefficient of NO2
distance
distance in kilometers, day
District of Columbia Circuit
diastolic blood pressure
District of Columbia
Detroit Exposure and
Aerosol Research Study
December
diesel exhaust particles
diesel exhaust particle
extract-coated carbon black
particles
degrees of freedom
dehydroascorbate
December, January,
February
Acronym/
Abbreviation
DL
DLM
DNA
DNC
DOAS
DOCs
dPD
DPF
DPPC
DVT
e.g.
Ea
Enaj
EEC
EC
ECG
ECP
ECRHS
ED
EGU
Ey
ELF
Ena
eNO
eNOS
Eo
Eoj
EP
Meaning
distributed lag
Polynomial distributed lag
model
deoxyribonucleic acid
Democratic National
Convention
differential optical
absorption spectroscopy
diesel oxidation catalysts
change in provocative dose
diesel particulate filter
dipalmitoyl
phosphatidylcholine
deep vein thrombosis
exempli gratia (for example)
the sum of an individual's
ambient NO2 exposure
indoor exposures from
nonambient sources
exhaled breath condensate
elemental carbon
electrocardiographic
eosinophil cationic protein
European Community
Respiratory Health Survey
emergency department
electric power generating
unit
indoor NO2 exposure in the
jth microenvironment
epithelial lining fluid
the sum of an individual's
nonambient NO2 exposure
exhaled nitric oxide
endothelial nitric oxide
synthase
outdoor microenvironmental
NO2 exposures
outdoor NO2 exposure in the
jth microenvironment
entire pregnancy
XXVlll
-------
Acronym/
Abbreviation
EPA
EP-A
EP-B
ESCAPE
ESR
ET
ET-1
ETS
Exp
F
FE-AADT
Feb
FEF
FEF25-75%
FEF 50%
FEM
FEVi
FL
FR
FRM
FVC
Y
Y'
g
&
g/bhp-h
GA
GAM
GASPII
GCLC
Meaning
U.S. Environmental
Protection Agency
El Paso— Site A
El Paso— Site B
European Study of Cohorts
for Air Pollution Effects
erythrocyte sedimentation
rate
total personal exposure
vasoconstrictor endothelin-1
environmental tobacco
smoke
exposure
female
Fleet equivalent annual
average daily traffic
February
forced expiratory flow
forced expiratory flow at
25-75% of exhaled volume
forced expiratory flow at
50% of forced vital capacity
federal equivalent method
forced expiratory volume in
1 second
Florida
Federal Register
federal reference method
forced vital capacity
gamma; uptake coefficients
semivariogram
gram
grams per brake
horsepower-hour
Georgia
generalized additive models
Gene and Environmental
Prospective Study in Italy
gene that encodes the
catalytic subunit for the
Acronym/
Abbreviation
GCLM
GD
GEE
GEOS
GINI
GINI SOUTH
GINIplus
GIS
GLM
GLMM
GM-CSF
GPS
GPx
GS*
GSD
GSH
GSNOR
GSR
GSS
GST
GSTM1
GSTP1
GSTT1
glutamate-cysteine ligase
Meaning
gene that encodes the
regulatory subunit for the
human enzyme
glutamate-cysteine ligase
gestation day
generalized estimating
equations
Goddard Earth Observing
System
German Infant Nutritional
Intervention
German Infant Nutritional
Intervention covers the
urban city of Munich,
Germany, and its
surrounding areas
(approximately 28,000 km2)
German Infant Nutritional
Intervention plus
environmental and genetic
influences
geographic information
systems
generalized linear model
generalized linear mixed
model
granulocyte
macrophage-co lony
stimulating factor
global positioning system
glutathione peroxidase
glutathione radical
geometric standard
deviation
glutathione
nitrosoglutathione reductase
glutathione reductase
glutathione synthetase
glutathione S-transferase
glutathione S-transferase
Mul
glutathione S-transferase Pi
1
glutathione S-transferase
theta 1
XXIX
-------
Acronym/
Abbreviation
GW
h
H+
H2S04
HC
hCAEC
HC1
HDL
HDM
HERO
HEV
HF
HFE
HFn
HGF
HI
HIST
HMOX
HN02
HN03
HN04
HO-1
H02
H02N02
HONO
HOONO
HR
HRV
HS
HSC
hs-CRP
IA
Meaning
gestational week
hour(s)
hydrogen ion
sulfuric acid
hydrocarbon(s)
human coronary artery
endothelial cell
hydrochloric acid
high-density lipoprotein
house dust mite; house dust
mite allergen
Health and Environmental
Research Online
hold-out evaluation
high frequency; high
frequency component of
HRV
human hemochromatosis
protein
high frequency domain
normalized for heart rate
hepatocyte growth factor
Hawaii
histamine
heme oxygenase
nitrous acid
nitric acid
peroxynitric acid
heme oxygenase- 1
hydroperoxyl radical
peroxynitric acid
nitrous acid
pemitrous acid
hazard ratio(s); heart rate
heart rate variability
hemorrhagic stroke
Harvard Six Cities
high sensitivity C-reactive
protein
Iowa
Acronym/
Abbreviation
I ACID
i.e.
I.V.
ICAM-1
ICAS
ICD
ICS
ID
IDW
IFN-y
IgE
IGM
IHD
IL
IL-6
IL-8
He
IM
IMSI
IMT6seg
IMTcca
IN
INDAIR
INFj
iNOS
IOM
IQR
IRP
IRR
IS
Meaning
inorganic acid
id est (that is)
intravenously
intercellular adhesion
molecule 1
Inner-City Asthma Study
International Classification
of Diseases; implantable
cardioverter defibrillators
inhaled corticosteroids
Idaho
inverse distance weighting
interferon gamma
immunoglobulin E
impaired glucose
metabolism
ischemic heart disease
interleukin; Illinois
interleukin-6
interleukin-8
isoleucine
immediately after exposure
Integrated Mobile Source
T 1 • .
HlQlCdlOI
intima-media thickness of
the left and right common
carotid arteries, internal
carotid arteries, and carotid
bulbs
intima-media thickness of
the common carotid artery
Indiana; isoprene nitrate
probabilistic model for
indoor pollution exposures
infiltration of outdoor NO2
inducible nitric oxide
synthase
Institute of Medicine
interquartile range
Integrated Review Plan
incidence rate ratios
ischemic stroke
XXX
-------
Acronym/
Abbreviation
ISA
IT
IUGR
IVF
j
JE
k
kcal
kg
fe
kj
km
kPa
KS
KY
L
LA
LAT
LB
LEW
LDH
LE
LETO
LF
LF/HF
LIE
LIF
LISA
Meaning
Integrated Science
Assessment
intratracheal
intrauterine growth
restriction
in vitro fertilization
microenvironment
joint model estimate
reaction rate; decay constant
derived from empirical data
kilocalorie(s)
kilogram(s)
second-order rate
constants)
decay rate
kilometer(s)
kilopascal(s)
Kansas
Kentucky
liter(s)
Louisiana; Los Angeles;
Lake Arrowhead
L-type amino acid
transporter
Long Beach
low birth weight
lactate dehydrogenase
Lake Elsinore
Long-Evans Tokushima
low-frequency component
of HRV
\JL 1 JJX V
ratio ofLF andHF
components of HRV
Long Island Expressway
laser induced fluorescence
Lifestyle-Related factors on
the Immune System and the
Development of Alleraies in
Acronym/
Abbreviation
LISAplus
LM
LN
LOESS
LOOCV
LOPAP
LOX-1
Lp-PLA2
LRTI
LUR
V
ug/m3
m
M
MA
Ml
M2
M3
M4
MAAS
max
MCP-1
MD
MDA
ME
MESA-Air
Meaning
Lifestyle-Related factors on
the Immune System and the
Development of Allergies in
Childhood plus the
influence of traffic
emissions and genetics
Lompoc
Lancaster
locally weighted scatterplot
smoothing
leave-one-out
cross-validation
long path absorption
photometer
lectin-like oxidized low
density lipoprotein receptor
lipoprotein-associated
phospholipase A2
lower respiratory tract
infection
land use regression
mu; micro
micrograms per cubic meter
meter
male
Massachusetts
Month 1
Month 2
Months
Month 4
Manchester Asthma and
Allergy Study
maximum
monocyte chemoattractant
protein-1
Maryland
malondialdehyde
Maine
Multi-Ethnic Study of
Childhood
MET
Atherosclerosis and Air
Pollution
MET receptor tyro sine
kinase gene
XXXI
-------
Acronym/
Abbreviation
METH
METS
MI
mm
ML
mL
MLI
MMEF
mmHg
MMP
MMP-3
MMP-7
MMP-9
MN
mo
MO
MOA
mol
MoOx
MPO
mRNA
MS
MT
n
N
N203
N204
N205
NA
Na+
NAAQS
NAB
NaCl
Meaning
methacholine
metabolic equivalents
myocardial infarction
("heart attack"); myocardial
ischemia; Michigan
minimum
Mira Loma
milliliter(s)
mean linear intercept
maximum (or maximal)
midexpiratory flow
millimeters of mercury
matrix metalloproteinase
matrix metalloproteinase-3
matrix metalloproteinase-7
matrix metalloproteinase-9
Minnesota
month(s)
Missouri
mode(s) of action
mole
molybdenum oxide
myeloperoxidase
messenger ribonucleic acid
Mississippi
Montana
sample size; total number of
microenvironments that the
individual has encountered
nitrogen; population number
dinitrogen trioxide
dinitrogen tetroxide
dinitrogen pentoxide
not available
sodium ion
National Ambient Air
Quality Standards
North American
Background
sodium chloride
Acronym/
Abbreviation
NADPH
NAL
NAMS
NAS
NC
NCEA
NCICAS
NCore
ND
NDMA
NE
NEI
NFKB
NH
NH3
(NH4)2SO4
NHAPS
NHS
NJ
NLCS
nm
NM
NMMAPS
NMOR
NO
NO2
NO3
NO3«
Meaning
reduced nicotinamide
adenine dinucleotide
phosphate
nasal lavage
National Air Monitoring
Stations
National Academy of
Sciences
North Carolina
National Center for
Environmental Assessment
National Cooperative Inner-
City Asthma Study
National Core network
North Dakota
N-nitrosodimethylamine
Nebraska
National Emissions
Inventory
nuclear factor kappa-light-
chain-enhancer of activated
B cells
New Hampshire
ammonium sulfate
National Human Activity
Pattern Survey
Nurses Health Study
New Jersey
Netherlands Cohort Study
on Diet and Cancer
nanometer
New Mexico
The National Morbidity
Mortality Air Pollution
Study
N-nitrosomorpholine
nitric oxide
nitrogen dioxide
nitrite
nitrate
nitrate radical
XXXll
-------
Acronym/
Abbreviation
nonHS
NOS
NOx
NOy
NOz
NQ01
NR
NS
NV
NY
OACID
O3
OAQPS
OC
OH
8-OHdG
OK
OLETF
OLM
OMI
OR
OVA
P
Meaning
nonhemhorragic stroke
nitric oxide synthase
the sum of NO and NO2
oxides of nitrogen
reactive oxides of nitrogen
(e.g., HNO3, HONO, PAN,
particulate nitrates)
NADPH-quinone
oxidoreductase (genotype)
not reported; no quantitative
results reported; near road
not statistically significant
Nevada
New York
organic acid
ozone
Office of Air Quality
Planning & Standards
organic carbon
hydroxide; Ohio
8-hydroxy-29-
deoxy guano sine
Oklahoma
Otsuka Long-Evans
Tokushima Fatty
ozone limiting method
ozone monitoring
instrument
odds ratio(s); Oregon
ovalbumin
p-value, probability of
Acronym/
Abbreviation
PAH(s)
PAMS
PAN
PAPA
Pb
PEL
PC
PCA
PCO
PD
PE
PEF
PFK
PIAMA
PiZZ
Pi
-1 J
PK
p.m.
PM
P
Pa
PA
PAARC
obtaining a result equal to or
"more extreme" than what
was actually observed,
assuming that the null
hypothesis is true
Pearson correlation
pascal(s)
policy assessment;
Pennsylvania
air pollution and chronic
respiratory diseases
Meaning
polycyclic aromatic
hydrocarbon(s)
photochemical monitoring
stations
peroxyacetyl nitrate;
peroxyacl nitrate
Public Health and Air
Pollution in Asia
lead
planetary boundary layer
provocative concentration
principal component
analysis
protein carbonyl
provocative dose
pulmonary embolism
peak expiratory flow
phosphofructokinase
prevention and incidence of
asthma and mite allergy
severe alpha-1 antitrypsin
deficiency
air pollutant penetration
pyruvate kinase
post meridiem (after noon)
particulate matter
XXXlll
-------
Acronym/
Abbreviation
PMio
PMlO-2.5
Meaning
In general terms, particulate
matter with a nominal mean
aerodynamic diameter less
than or equal to 10 urn; a
measurement of thoracic
particles (i.e., that subset of
inhalable particles thought
small enough to penetrate
beyond the larynx into the
thoracic region of the
respiratory tract). In
regulatory terms, particles
with an upper 50% cut-point
of 10 ± 0.5 um aerodynamic
diameter (the 50% cut point
diameter is the diameter at
which the sampler collects
50% of the particles and
rejects 50% of the particles)
and a penetration curve as
measured by a reference
method based on Appendix
J of 40 CFR Part 50 and
designated in accordance
with 40 CFR Part 53 or by
an equivalent method
designated in accordance
with 40 CFR Part 53.
In general terms, particulate
matter with a nominal mean
aerodynamic diameter less
than or equal to 10 um and
greater than 2.5 um; a
measurement of thoracic
coarse particulate matter or
the coarse fraction of PMio.
In regulatory terms,
particles with an upper 50%
cut-point of 10 um
aerodynamic diameter and a
lower 50% cut-point of
2.5 um aerodynamic
diameter (the 50% cut point
diameter is the diameter at
which the sampler collects
50% of the particles and
rejects 50% of the particles)
as measured by a reference
method based on Appendix
Oof 40 CFR Part 50 and
designated in accordance
with 40 CFR Part 5 3 or by
an equivalent method
designated in accordance
with 40 CFR Part 53.
Acronym/
Abbreviation
PM2.5
PMA
PMN(s)
PNC
PND
pNNSO
pNO
pN03
PPARy
ppb
ppm
PROtEuS
PTB
Meaning
In general terms, particulate
matter with a nominal mean
aerodynamic diameter less
than or equal to 2.5 urn; a
measurement of fine
particles. In regulatory
terms, particles with an
upper 50% cut-point of
2.5 um aerodynamic
diameter (the 50% cut point
diameter is the diameter at
which the sampler collects
50% of the particles and
rejects 50% of the particles)
and a penetration curve as
measured by a reference
method based on Appendix
L of 40 CFR Part 50 and
designated in accordance
with 40 CFR Part 53, by an
equivalent method
designated in accordance
with 40 CFR Part 53, or by
an approved regional
method designated in
accordance with Appendix
C of 40 CFR Part 58.
phorbol myristate acetate
polymorphonuclear cell(s),
polymorphonuclear
leukocyte
particle number
concentration
postnatal day
Proportion of pairs of
successive normal sinus
intervals exceeds
50 milliseconds divided by
the total number of
successive pairs of normal
sinus intervals
particulate nitrogen species
particulate nitrate
peroxisome proliferator
activated receptor gamma
parts per billion
parts per million
Prostate Cancer and
Environment Study
preterm birth
XXXIV
-------
Acronym/
Abbreviation
PVMRM
Qi
Q2
Q3
Q4
Q5
QC-TILDAS
QT interval
QTc
QTVI
QUIC
R3
RAG
RANCH
RBC
RC(=O)
RC(=0)OON02
REA
REGICOR
RH
RI
RIVM
rMSSD
RNS
RONO2
Meaning
plume volume molar ratio
method
1 st quartile or quintile
2nd quartile or quintile
3rd quartile or quintile
4th quartile or quintile
5th quintile
quantum cascade—tunable
infrared laser differential
absorption spectrometer
time between start of Q
wave and end of T wave in
ECG
corrected QT interval
QT variable index
Quick Urban and Industrial
Complex
Pearson correlation
coefficient; Spearman
correlation coefficient
square of the correlation
coefficient
ragweed
road traffic and aircraft
noise exposure and
children's cognition and
health
red blood cells
acyl group
peroxyacylnitrates
Risk and Exposure
Assessment
Registre Gironi del Cor
relative humidity
Rhode Island
National Air Quality
Monitoring Network of the
National Institute of Public
Health and the Environment
root mean square of
successive differences; a
measure of HRV
reactive nitrogen species
organic nitrates
Acronym/
Abbreviation
ROS
RR
RSNO
RSV
RV
8
sec
S. Rep.
s/L
S/N
SALIA
SA-LUR
SAPALDIA
SAT
SBP
SC
SCR
SD
SDNN
SE
Se
SEARCH
sec
SEI
Se-L
SES
Se-S
Sess.
SFe
SGA
sGaw
Meaning
reactive oxygen species
risk ratio(s), relative risk
S-nitrosothiols
respiratory syncytial virus
Riverside
Sigma, random error
second(s)
Senate Report
seconds per liter
Signal-to-noise ratio
Study on the Influence of
Air Pollution on Lung,
Inflammation, and Aging
source-area land use
regression
Swiss Study on Air
Pollution and Lung Disease
in Adults
switching attention test
systolic blood pressure
South Carolina
selective catalytic reduction
standard deviation; South
Dakota; San Dimas
standard deviation of all
normal-to-normal intervals,
an index of total HRV
standard error
selenium
Southeast Aerosol Research
Characterization
second(s)
socio-economic index
low selenium
socioeconomic status
supplemented selenium
session
sulfur hexafluoride
small for gestational age
specific airway conductance
XXXV
-------
Acronym/
Abbreviation
SHARP
SHEDS
SHEEP
sICAM-1
SLAMS
SM
SNP
SO2
S04
SOA
SOD
SP-D
SPE
sRaw
SRTT
ST segment
sVCAM-1
T
t
TEARS
Tl
T2
T3
TEARS
TCHS
Meaning
Study of Houston
Atmospheric Radical
Precursors
Stochastic Human Exposure
and Dose Simulation
Stockholm Heart
Epidemiology Program
soluble intercellular
adhesion molecule-1
state and local air
monitoring stations
Santa Maria
single nucleotide
polymorphism
sulfur dioxide
sulfate
secondary organic aerosols
superoxide dismutase
surfactant protein D
single-pollutant model
estimate
specific airway resistance
simple reaction time test
segment of the
electrocardiograph between
the end of the S wave and
beginning of the T wave
soluble vascular adhesion
molecule-1
tau, half-time
fraction of time spent in a
microenvironment across an
individual's
microenvironmental
exposures, time
thiobarbituric acid reactive
substances (species)
first trimester
second trimester
third trimester
thiobarbituric acid reactive
substances
Taiwan Children Health
Study
Acronym/
Abbreviation
TEA
Thl7
Th2
TIA
TIM
TIMP-2
tj
TLR
TN
TNF
TNF-a
TSP
TWA
TX
U.S.C.
UCD
UF1
UF2
UFP
UK
U.K.
ULTRA
UP
URI
U.S.
UT
VA
Val
VCAM-1
Meaning
triethanolamine
T helper cell 17
T-derived lymphocyte
helper 2
transient ischemic attack
timothy
tissue inhibitor of matrix
metalloproteinase-2
fraction of total time spent
in the jth microenvironment
Toll-like receptor
Tennessee
tumor necrosis factor
tumor necrosis factor alpha
total suspended solids
time-weighted average
Texas
U.S. Code
University of California,
Davis
ultrafme particle number
beginning at 3 nanometers
ultrafine particle number
beginning at 15 nanometers
ultrafine particle(s)
universal kriging
United Kingdom
The Exposure and Risk
Assessment for Fine and
Ultrafine Particles in
Ambient Air Study
conducted in Europe
Upland
upper respiratory infection
United States of America
Utah
Virginia
valine
vascular adhesion
molecule-1
minute volume
XXXVI
-------
Acronym/
Abbreviation
VEGF
VOC
VPTB
VT
VT
vWF
WBC
WHI
WHO
WI
WV
WY
X
Y
Meaning
vascular endothelial growth
factor
volatile organic compound
very preterm birth
tidal volume
ventricular
tachyarrhythmias; Vermont
von Willebrand factor
white blood cell
Women's Health Initiative
World Health Organization
Wisconsin
West Virginia
Wyoming
distance from the road
health effect of interest
Acronym/
Abbreviation
yr
Z
Z*
Zn
Meaning
fraction of time spent
indoors
fraction of a day spent in
each indoor
micro environment
fraction of all time spent
outdoors
fraction of a day spent in
each outdoor
microenvironment
year(s)
covariate vector; the
measured concentration;
standard normal deviate
the true concentration
XXXVll
-------
PREAMBLE
Process of Integrated Science Assessment Development
This Preamble outlines the general process the United States Environmental Protection
Agency (U.S. EPA) uses to develop an Integrated Science Assessment (ISA), including
the framework for evaluating weight of evidence and drawing scientific conclusions and
causal judgments. The ISA provides a concise review, synthesis, and evaluation of the
most policy-relevant science to serve as a scientific foundation for the review of the
National Ambient Air Quality Standards (NAAQS).1 The NAAQS are established based
on consideration of the air quality criteria (represented by the ISA) for the pollutants
identified by the Administrator using Section 108 of the Clean Air Act (CAA). The
pollutants currently identified are carbon monoxide (CO), lead (Pb), oxides of nitrogen,
photochemical oxidants, particulate matter (PM), and sulfur oxides (CAA. 1990a. b).
Figure I depicts the general NAAQS review process. Information for individual NAAQS
reviews is available online.2
The development of the ISA is preceded by the release of an Integrated Review Plan
(IRP) that discusses the planned scope of the NAAQS review; the planned approaches for
developing the key assessment documents [e.g., ISA, Risk and Exposure Assessment (if
warranted), Policy Assessment]; and the schedule for release and review of the
documents and subsequent rulemaking notices. The key policy-relevant questions
included in the IRP serve to clarify and focus the NAAQS review on the critical scientific
and policy issues, including addressing uncertainties discussed during the previous
review and newly emerging literature. The IRP is informed by a U.S. EPA-hosted public
science and policy issue workshop that "kicks off the review of the NAAQS for a given
criteria pollutant by seeking input on the current state of the science and engaging
stakeholders and experts in discussion of the policy-relevant questions that will frame the
review.
1 The general process for NAAQS reviews is described at http://www.epa.gov/ttn/naaqs/review.html.
2 Information for individual NAAQS reviews is available at www.epa.gov/ttn/naaqs.
xxxvin
-------
Workshop on
science-policy issues
EPA
proposed
decisions on
Integrated Review Plan (IRP): timeline and key
policy-relevant issues and scientific questions
Integrated Science Assessment (ISA): evaluation and
synthesis of most policy-relevant studies
Risk/Exposure Assessment (REA):
quantitative assessment, as warranted, focused
on key results, observations, and uncertainties
Policy Assessment (PA): staff analysis of
policy options based on integration and
interpretation of information in the ISA and REA
Figure I
w
Public hearings
and comments
on proposal
Agency decision
making and draft
final notice
Interagency
review
Clean Air Scientific
Advisory Committee
(CASAC) review
Public comment
Schematic of the key steps in the review of National Ambient Air
Quality Standards.
This Preamble is a general discussion of the basic steps and criteria used in developing an
ISA. Details and considerations specific to an individual ISA are included in the IRP as
well as the Preface and other introductory materials for that assessment. The general
process for ISA development is illustrated in Figure II. An initial step (not shown) is
publication of a call for information in the Federal Register that invites the public to
provide information relevant to the assessment, such as new or recent publications on
health or welfare effects of the pollutant or data from the fields of atmospheric and
exposure science.
The fundamental process for developing an ISA includes:
• Literature searches;
• Study selection;
• Evaluation of individual study quality;
• Evaluation, synthesis, and integration of the evidence; and
• Development of scientific conclusions and causal determinations.
AXX1X
-------
Literature Search and
Study Selection
(See Figure III)
Evaluation of Individual Study Quality
After study selection, the quality of individual studies is evaluated by the U.S. EPAor outside experts in the fields
of atmospheric science, exposure assessment, dosimetry, animal toxicology, controlled human exposure,
epidemiology, biogeochemistry, terrestrial and aquatic ecology, and other welfare effects, considering the design,
methods, conduct, and documentation of each study. Strengths and limitations of individual studies that may affect
the interpretation of the study's results are considered.
Develop Initial Sections
Review and summarize conclusions from
previous assessments and new study results by
discipline and category of outcome/effect (e.g.,
lexicological studies of lung function,
biogeochemistry studies of forests).
Peer Input Consultation
Review of draft materials by scientists from
both outside and within the U.S. EPA in public
meeting or public teleconference.
Evaluation, Synthesis, and Integration of Evidence
Integrate evidence from scientific disciplines. Evaluate evidence for related groups of endpoints or outcomes to
draw conclusions for specific health or welfare effect categories, integrating health or welfare effects evidence with
information on mode of action and exposure assessment.
Development of Scientific Conclusions and Causal Determinations
Characterize weight of evidence and devetopjudgments regarding causality for health or welfare effect categories.
Develop conclusions regarding concentration- or dose-response relationships, potentially at-risk populations,
lifestages, or ecosystems.
Draft Integrated Science Assessment
Evaluation and integration of newly published studies
Final Integrated Science Assessment
Clean Air Scientific Advisory Committee
Independent review of draft documents for scientific
quality and sound implementation of causal
framework at public meetings.
Public Comments
Comments on draft Integrated Science Assessment
solicited by the U.S. EPA
Note: U.S. EPA = United States Environmental Protection Agency.
Figure II
Characterization of the general process for developing an
Integrated Science Assessment.
xl
-------
In developing an ISA, the U.S. EPA reviews and summarizes the evidence from studies
on atmospheric sciences, human exposure, animal toxicology, controlled human
exposure, epidemiology, and/or ecology and other welfare1 effects. In the process of
developing the first draft ISA, the U.S. EPA may convene a peer input meeting in which
the scientific content of preliminary draft materials is reviewed by subject-matter experts
to ensure that the ISA is up-to-date and is focused on the most policy-relevant findings.
This input also assists the U.S. EPA with the integration of evidence within and across
disciplines.
The U.S. EPA integrates the evidence across scientific disciplines or study types and
characterizes the weight of evidence for relationships between the pollutant(s) being
evaluated and various outcomes. Integrating evidence on health or welfare effects
involves collaboration among scientists from various disciplines. For example, an
evaluation of health effects evidence would generally include integrating the results from
epidemiologic, controlled human exposure, and toxicological studies; considering
exposure assessment; and applying the causal framework (described below) to draw
conclusions.
Integration of results on health or welfare effects that are logically or mechanistically
connected (e.g., respiratory symptoms, asthma exacerbation) informs judgments of
causality on a broader health effect category (e.g., effects on the respiratory system).
Using the causal framework described in this Preamble. U.S. EPA scientists consider
aspects, such as strength, consistency, coherence, and biological plausibility of the
evidence, and develop causal determinations on the nature of the relationships with the
pollutant(s) being evaluated. Causal determinations often entail an iterative process of
review and evaluation of the evidence. One or more drafts of the ISA are released for
review by the Clean Air Scientific Advisory Committee (CASAC) and the public, and
comments received on the characterization of the science as well as the implementation
of the causal framework are carefully considered in revising the draft ISA and completing
the ISA.
2. Literature Search
In addition to the call for information in the Federal Register referenced above, the
U.S. EPA maintains an ongoing literature search process to identify relevant scientific
studies published since the last ISA for a given criteria pollutant. Search strategies are
1 Under CAA Section 302(h) [42 U.S.C. 7602(h)], language referring to "effects on welfare" includes, but is not
limited to, "effects on soil, water, crops, vegetation, man-made materials, animals, wildlife, weather, visibility, and
climate, damage to and deterioration of property, and hazards to transportation, as well as effects on economic
values and on personal comfort and well-being."
xli
-------
designed a priori for pollutants and scientific disciplines and iteratively modified to
optimize identification of pertinent publications. Papers are identified for inclusion in
several additional ways: specialized searches on specific topics, identification of new
publications by relational searches conducted using citations from previous assessments,
review of tables of contents for journals in which relevant papers may be published,
identification of relevant literature by expert scientists, review of citations in previous
assessments, and recommendations by the public and CASAC during the call for
information and external review processes. This multipronged search strategy aims to
identify all relevant epidemiologic, controlled human exposure, toxicological, ecological,
and welfare effects studies published since the last ISA as well as studies related to
exposure-response relationships, mode(s) of action, and populations and lifestages at
increased risk of air pollution-related health effects. Also relevant to the ISA are studies
and data analyses on atmospheric chemistry, air quality and emissions, environmental
fate and transport, dosimetry, toxicokinetics, and exposure.
References identified through the multipronged search strategy are then "screened" by
title and abstract. References that are judged to be potentially relevant based on review
beyond the title are "considered" for inclusion in the ISA and are added to the Health and
Environmental Research Online (HERO) database developed by the U.S. EPA.1 These
"considered" references can be found on the HERO project page for the particular ISA.
Studies and reports that have undergone scientific peer review and have been published
(or accepted for publication) are eligible for review in the ISA. Further, only studies that
have been ethically conducted (e.g., approval by an Institutional Review Board or
Institutional Animal Care and Use Committee) are eligible for review in the ISA. Each
"included" reference is cited in the ISA as a hyperlink to the ISA project page in the
HERO database. Additional review steps (described in Section 3^ below) precede a
decision on whether a study will be "included" in the ISA. This literature search and
study selection process, including identification of "screened," "considered," and
"included" references, is depicted in Figure III.
1 The list of "considered" and "cited" references and bibliographic information is accessible to the public through
HERO (http://hero.epa.govX
xlii
-------
Figure
Caltfor
Information
and
Literature
Search
Citations from
Past Assessments
Peer Review
Recommendations
Illustration of literature search and study selection process used
for developing Integrated Science Assessments.
Each ISA builds upon the conclusions of previous assessments for the pollutant under
review. The U.S. EPA focuses on peer-reviewed literature published since the completion
of the previous ISA and on any new interpretations of previous literature, integrating the
results of recent scientific studies with previous findings. Important earlier studies may
be discussed in detail to reinforce key concepts and conclusions or for reinterpretation in
light of newer data. Earlier studies also are the primary focus for some topics covered in
the ISA where research efforts have subsided, or if these earlier studies remain the
definitive works available in the literature.
3. Study Selection
References considered for inclusion in the ISA undergo abstract and full-text review to
determine whether they will be included in the ISA. The selection process is based on the
extent to which the study is informative, pertinent, and policy relevant. Informative,
pertinent, and policy-relevant studies include those that describe or provide a basis for
characterizing the relationship between the criteria pollutant and health or welfare effects,
including studies that offer innovation in method or design and studies that reduce
uncertainty on critical issues. Emphasis is placed on studies that examine effects
xliii
-------
associated with pollutant concentrations and exposure conditions relevant to current
human population and ecosystem exposures, and particularly those pertaining to
concentrations currently found in ambient air. Other studies are included if they contain
unique data, such as a previously unreported effect or mode of action for an observed
effect, or examine multiple concentrations to elucidate exposure-response relationships.
4. Evaluation of Individual Study Quality
After studies are selected for inclusion, individual study quality is evaluated by reviewing
the design, methods, conduct, and documentation of each study, but not the study results.
This uniform approach aims to assess the strengths, limitations, and possible roles of
chance, confounding, and other biases that may affect the interpretation of individual
studies and the strength of inference from the results of the study. Particular aspects or
the absence of some features in a study do not necessarily define a less informative study
or exclude a study from consideration in an ISA. As stated initially, the intent of the ISA
is to provide a concise review, synthesis, and evaluation of the most policy-relevant
science to serve as a scientific foundation for the review of the NAAQS, not extensive
summaries of all health, ecological, and other welfare effects studies for a pollutant. A
primary issue in the decision to include a study is whether it provides useful qualitative or
quantitative information on exposure-response relationships for effects associated with
pollutant exposures at doses or concentrations relevant to ambient conditions that can
inform decisions on whether to retain or revise the standards.
Generally, in assessing the scientific quality of studies on health and welfare effects, the
following considerations are taken into account.
• Were study design, study groups, methods, data, and results clearly presented in
relation to the study objectives to allow for study evaluation? Were limitations and
any underlying assumptions of the design and other aspects of the study stated?
• Were the ecosystems, study site(s), study populations, subjects, or organism
models adequately selected, and are they sufficiently well defined to allow for
meaningful comparisons between study or exposure groups?
• Are the air quality data, exposure, or dose metrics of adequate quality and
sufficiently representative of information regarding ambient conditions?
• Are the health, ecological, or other welfare effect measurements meaningful,
valid, and reliable?
• Were likely covariates or modifying factors adequately controlled or taken into
account in the study design and statistical analysis?
• Do the analytical methods provide adequate sensitivity and precision to support
conclusions?
xliv
-------
• Were the statistical analyses appropriate, properly performed, and properly
interpreted?
Additional study quality considerations specific to particular disciplines are discussed
below.
a. Atmospheric Science and Exposure Assessment
Atmospheric science and exposure assessment studies that are considered for inclusion in
the ISA focus on measurement of, behavior of, and exposure to ambient air pollution
using quality-assured field, experimental, and/or modeling techniques. The most
informative measurement-based studies will include detailed descriptive statistics for
measurements taken at varying spatial and temporal scales. These studies will also
include a clear and comprehensive description of measurement techniques and
quality-control procedures used. Quality-control metrics (e.g., method detection limits)
and quantitative relationships between and within pollutant measurements
(e.g., regression slopes, intercepts, fit statistics) should be provided when appropriate.
Measurements that include contrasting conditions for various time periods
(e.g., weekday/weekend, season), populations, regions, and categories (e.g., urban/rural)
are particularly useful. The most informative modeling-based studies will incorporate
appropriate chemistry, transport, dispersion, and/or exposure modeling techniques with a
clear and comprehensive description of model evaluation procedures, metrics, and
technique strengths and limitations. The ISA also may include analyses of data pertinent
to characterizing air quality or exposure, such as emissions sources and ambient air
pollutant concentrations. Sources of monitoring and modeling data should be clearly
referenced and described to foster transparency and reproducibility of any analysis. In
general, atmospheric science studies and data analyses focusing on locations pertinent to
the U.S. will have maximum value in informing review of the NAAQS.
Exposure measurement error, which refers to inaccuracies in the characterization of the
exposures of study participants, can be an important contributor to uncertainty in air
pollution epidemiologic study results. Exposure measurement error can influence
observed epidemiologic associations between ambient pollutant concentrations and health
outcomes by biasing effect estimates toward or away from the null and/or widening
confidence intervals around those estimates (Zeger etal.. 2000). Factors that could
influence exposure estimates include, but are not limited to: choice of exposure metric,
spatial variability of the pollutant concentration, nonambient sources of exposure,
topography of the natural and built environment, meteorology, instrument errors,
time-activity patterns, and differential infiltration of air pollutants into indoor
environments. The influence of these factors on effect estimates also depends on
xlv
-------
epidemiologic study design. For example, when longitudinal studies depend on spatial
contrasts in exposure estimates, it is important that the exposure estimates correspond in
space to the population of interest. Likewise for time-series studies, the temporal
variability of the exposure estimate must correspond temporally to the true exposures of
the study population.
b. Epidemiology
In addition to the general study quality considerations discussed above, the U.S. EPA
evaluates quality of individual epidemiologic studies for inference about health effects by
considering whether a given study: (1) presents information on associations with short- or
long-term pollutant exposures at or near conditions relevant to ambient exposures;
(2) addresses potential confounding, particularly by other pollutants; (3) assesses
potential effect modifiers; (4) evaluates health endpoints and populations, groups, or
lifestages not previously extensively researched; and (5) evaluates important
methodological issues related to interpretation of the health evidence (e.g., lag or time
period between exposure and effects, model specifications, thresholds).
In evaluating epidemiologic evidence, one important consideration is potential
confounding. Confounding is "... a confusion of effects. Specifically, the apparent effect
of the exposure of interest is distorted because the effect of an extraneous factor is
mistaken for or mixed with the actual exposure effect (which may be null)" (Rothman
and Greenland. 1998). A confounder is associated with both the exposure and the effect;
for example, confounding can occur between correlated pollutants that are associated
with the same effect. One approach to remove spurious associations due to possible
confounders is to control for characteristics that may differ between exposed and
unexposed persons; this is frequently termed "adjustment." Scientific judgment is needed
to evaluate likely sources and extent of confounding, together with consideration of how
well the existing constellation of study designs, results, and analyses address the potential
for erroneous inferences.
Several statistical methods are available to detect and control for potential confounders;
however, none of these methods is completely satisfactory. Multivariable regression
models constitute one tool for estimating the association between exposure and outcome
after adjusting for characteristics of participants that might confound the results. Because
much of the uncertainty in inferring causality may be due to potential confounding by
copollutants, evaluation of copollutant confounding in individual studies is of particular
importance. The use of copollutant regression models has been the prevailing approach
for controlling for potential confounding by copollutants in air pollution health effects
xlvi
-------
studies. Trying to determine whether an individual pollutant is independently associated
with the health outcome of interest from copollutant regression models is made difficult
by the possibility that one or more air pollutants is acting as a surrogate for an
unmeasured or poorly measured pollutant or for a particular mixture of pollutants. In
addition, pollutants may independently exert effects on the same system; for example,
several pollutants may be associated with a respiratory effect through either the same or
different modes of action. Despite these limitations, the use of copollutant models is still
the prevailing approach employed in most air pollution epidemiologic studies and can
provide some insight into the potential for confounding or interaction among pollutants.
Confidence that unmeasured confounders are not producing the findings is increased
when multiple studies are conducted in various settings using different subjects or
exposures, each of which might eliminate another source of confounding from
consideration. For example, multicity studies can provide insight on potential
confounding through the use of a consistent method to analyze data from across locations
with different concentrations of copollutants and other covariates. Intervention studies,
because of their quasi-experimental nature, can be particularly useful in characterizing
causation.
Another important consideration in the evaluation of epidemiologic studies is
effect-measure modification, which occurs when the effect differs between subgroups or
strata; for example, effect estimates that vary by age group or a potential risk factor. As
stated by Rothman and Greenland (1998):
"Effect-measure modification differs from confounding in several ways.
The main difference is that, whereas confounding is a bias that the
investigator hopes to prevent or remove from the effect estimate,
effect-measure modification is a property of the effect under study .... In
epidemiologic analysis one tries to eliminate confounding but one tries to
detect and estimate effect-measure modification."
When a risk factor is a confounder, it is the true cause of the association observed
between the exposure and the outcome; when a risk factor is an effect modifier, it
changes the magnitude of the association between the exposure and the outcome in
stratified analyses. For example, the presence of a pre-existing disease or indicator of low
socioeconomic status (SES) may act as an effect modifier if it is associated with
increased risk of effects related to air pollution exposure. It is often possible to stratify the
relationship between health outcome and exposure by one or more of these potential
effect modifiers. For variables that modify the association, effect estimates in each
stratum will be different from one another and different from the overall estimate,
xlvii
-------
indicating a different exposure-response relationship may exist in populations represented
by these variables.
c. Controlled Human Exposure and Animal
Toxicology
Controlled human exposure and animal toxicological studies experimentally evaluate the
health effects of administered exposures in human volunteers and animal models under
highly controlled laboratory conditions. Controlled human exposure studies are also
referred to as human clinical studies. In controlled human exposure and animal
toxicological experiments, investigators expose subjects or animals to known
concentrations of air pollutants under carefully regulated environmental conditions and
activity levels. In addition to the general quality considerations discussed previously,
evaluation of controlled human exposure and animal toxicological studies includes
assessing the design and methodology of each study with focus on (1) characterization of
the intake dose, dosing regimen, and exposure route; (2) characterization of the
pollutant(s); (3) sample size and statistical power to detect differences; and (4) control of
other variables that could influence the occurrence of effects. The evaluation of study
design generally includes consideration of factors that minimize bias in results, such as
randomization, blinding, and allocation concealment of study subjects, investigators, and
research staff, and unexplained loss of animals or withdrawal/exclusion of subjects.
Additionally, studies must include appropriate control groups to allow for accurate
interpretation of results relative to exposure. Emphasis is placed on studies that address
concentration-dependent responses or time-course of responses and studies that
investigate potentially at-risk lifestages or populations (e.g., older adults, groups with
pre-existing disease).
Controlled human exposure or animal toxicological studies that approximate expected
human exposures in terms of concentration, duration, and route of exposure are of
particular interest. Relevant pollutant exposures are considered to be those generally
within two orders of magnitude of recent ambient concentrations. This range in relevant
exposures is intended to account for differences in dosimetry, toxicokinetics, and
biological sensitivity of various species, strains, or potentially at-risk populations. Studies
using higher concentration exposures or doses will be considered to the extent that they
provide information relevant to understanding mode of action or mechanisms,
inter-species variation, or at-risk human populations. In vitro studies may provide
mechanistic insight for effects examined in vivo or in epidemiologic studies.
xlviii
-------
d. Ecological and Other Welfare Effects
Ecological effects evaluated in the ISAs typically include several of the topics given as
examples by the CAA definition in Section 302(h) related to effects on welfare, including
soils, water, vegetation, animals, and wildlife. Additional topic areas that may be
evaluated in an ISA include visibility, weather, and climate, as well as materials damage,
economic values, and impacts to personal comfort and well-being. In evaluating studies
that consider welfare effects, in addition to assessing the general quality considerations
discussed previously, emphasis is placed on studies that evaluate effects at or near
ambient concentrations of the air pollutant(s). Studies conducted in any country that
contribute meaningfully to the general understanding of air pollutant effects may be
evaluated for relevancy to U.S. air quality considerations and inclusion in the ISA.
Studies at higher pollutant concentrations are used to evaluate ecological effects only
when they are part of a range of concentrations that also include more typical values, or
when they inform understanding of modes of action and illustrate the wide range of
sensitivity to air pollutants across taxa or across biomes and ecoregions. In evaluating
quantitative exposure-response relationships, emphasis is placed on findings from studies
conducted in the U.S. and Canada as having ecological and climatic conditions most
relevant for review of the NAAQS. The type of experimental approach used in the study
(e.g., controlled laboratory exposure, growth chamber, open-top chamber, mesocosm,
gradient, field study) is also evaluated when considering the applicability of the results to
the review of criteria air pollutant effects.
In evaluating studies on climate and visibility, emphasis is placed on studies that use
well-established measurement and modeling techniques, especially those that report
uncertainty or compare results from an ensemble of techniques. Novel methods may also
be informative in addressing knowledge gaps not well characterized by existing
techniques. Relevant climate studies include those evaluating direct and indirect climate
impacts of criteria air pollutants at a global scale, while for visibility, studies conducted
in the U.S. and Canada provide information more applicable for review of the NAAQS.
In both cases, studies that evaluate effects by source sector or region, such as regional
climate modeling studies, are particularly informative. Studies that report impacts of
multiple PM components for visibility and multiple criteria pollutants for climate are
useful in evaluating interactions and the relative contributions of atmospheric
constituents. For example, in evaluating the climate forcing effects of ozone (Os), it is
useful to understand the atmospheric chemistry involving CO and NOx (the sum of nitric
oxide and nitrogen dioxide) that affects atmospheric concentrations of Os. Visibility
preference and valuation studies that explicitly separate preferences for visibility from
xlix
-------
concerns about health risks of air pollution are particularly relevant in considering a
welfare-based secondary NAAQS for pollutants that affect visibility.
5. Evaluation, Synthesis, and Integration of Evidence across
Disciplines and Development of Scientific Conclusions and
Causal Determinations
The U.S. EPA has developed an approach for integrating the scientific evidence gained
from the array of study types discussed above in order to draw conclusions regarding the
causal nature of ambient air pollutant-related health or welfare effects. Evidence from all
disciplines is integrated to evaluate consistency and inconsistency in the pattern of effects
as well as strengths and limitations of the evidence across disciplines. Part of this
approach includes a framework for making determinations regarding the extent to which
a causal relationship exists between the pollutant in ambient air, and health or welfare
effects (described in Section 5.b). This framework establishes a uniform approach and
language to characterizing causality and brings specificity to the conclusions.
a. Evaluation, Synthesis, and Integration of Evidence
across Disciplines
The ISA focuses on evaluation of the findings from the body of evidence across
disciplines, drawing upon the results of all studies judged of adequate quality and
relevance per the considerations described previously. Evidence across scientific
disciplines for related and similar health or welfare effects is evaluated, synthesized, and
integrated to develop conclusions and causal determinations. This process includes
evaluating strengths and weaknesses in the overall collection of studies across disciplines.
Confidence in the collective body of evidence is based on evaluation of study design and
quality. The roles of different types of evidence in drawing the conclusions varies by
pollutant or assessment, as does the availability of different types of evidence for causal
determination. Conclusions on health effects are informed largely by controlled human
exposure, epidemiologic, and toxicological studies. Evidence on ecological and other
welfare effects may be drawn from a variety of experimental approaches
(e.g., greenhouse, laboratory, field) and numerous disciplines (e.g., community ecology,
biogeochemistry, paleontological/historical reconstructions). Other evidence, including
mechanistic, toxicokinetics, and exposure assessment, may be highlighted if it is relevant
to the evaluation of health and welfare effects and is of sufficient importance to affect the
overall evaluation. Causal inference can be strengthened by integrating evidence across
disciplines. A weak inference from one line of evidence can be addressed by other lines
-------
of evidence, and coherence of these lines of evidence can add support to a cause-effect
interpretation of the association. Interpretation of the body of epidemiologic associations
as evidence of causal relationships involves assessing the full evidence base with regard
to elimination of alternative explanations for the association.
Evaluation and integration of evidence must also include consideration of uncertainty,
which is inherent in scientific findings. "Uncertainty" can be defined as a deficit of
knowledge to describe the existing state or future outcome with accuracy and precision
(e.g., the lack of knowledge about the correct value for a specific measure or estimate).
Uncertainty analysis may be qualitative or quantitative in nature. In many cases, the
analysis is qualitative and can include professional judgment or inferences based on
analogy with similar situations. Quantitative uncertainty analysis may include use of
simple measures (e.g., ranges) and analytical techniques. Quantitative uncertainty
analysis might progress to more complex measures and techniques, if needed for decision
support. Various approaches to evaluating uncertainty include classical statistical
methods, sensitivity analysis, or probabilistic uncertainty analysis, in order of increasing
complexity and data requirements. However, data may not be available for all aspects of
an assessment, and those data that are available may be of questionable or unknown
quality. Ultimately, the assessment is based on a number of assumptions with varying
degrees of uncertainty. While the ISA may include quantitative analysis approaches such
as meta-regression in some situations, generally qualitative evaluation of uncertainties is
used to assess the evidence across studies.
Publication bias is another source of uncertainty that can impact the magnitude of
estimated health or welfare effects. It is well understood that studies reporting non-null
findings are more likely to be published than reports of null findings. Publication bias can
result in overestimation of effect estimate sizes (loannidis. 2008). For example, effect
estimates from single-city epidemiologic studies have been found to be generally larger
than those from multicity studies. This is an indication of publication bias because null or
negative single-city results may be reported in multicity analyses but might not be
published independently (Bell et al.. 2005).
Health-specific Considerations
Potential strengths and limitations of the body of studies can vary across disciplines and
are evaluated during data synthesis and integration. Direct evidence of a relationship
between pollutant exposures and health effects may come from controlled human
exposure studies. These studies can also provide important information on the biological
plausibility of associations observed in epidemiologic studies and inform determinations
of factors that may increase or decrease the risk of health effects in certain populations. In
-------
some instances, controlled human exposure studies can be used to characterize
concentration-response relationships at pollutant concentrations relevant to ambient
conditions. Controlled human exposures are typically conducted using a randomized
crossover design, with subjects exposed both to the pollutant and a clean air control. In
this way, subjects serve as their own experimental controls, effectively limiting the
variance associated with potential interindividual confounders. Limitations that must be
considered in evaluating controlled human study findings include the generally small
sample size and short exposure time used, and that severe health outcomes are not
assessed. By experimental design, controlled human exposure studies are structured to
evaluate physiological or biomolecular outcomes in response to exposure to a specific air
pollutant and/or combination of pollutants. In addition, the study design generally
precludes inclusion of subjects with serious health conditions or heightened risks of
exposure, and therefore, the results often cannot be generalized to an entire population,
which includes populations or lifestages at potentially increased risk of air
pollutant-induced effects. Although some controlled human exposure studies have
included health-compromised individuals, such as those with mild or moderate
respiratory or cardiovascular disease, these individuals may also be relatively healthy and
may not represent the most sensitive individuals in the population. Thus, observed effects
in these studies may underestimate the response in certain populations. In addition, the
study design is limited to exposures and endpoints that are not expected to result in
severe health outcomes.
Epidemiologic studies provide important information on the associations between health
effects and exposure of human populations to ambient air pollution. In epidemiologic or
observational studies of humans, the investigator tends not to control exposures or
intervene with the study population. Broadly, observational studies can describe
associations between exposures and effects. These studies fall into several categories and
include, for example, cross-sectional, prospective cohort, time-series, and panel studies.
Each type of study has various strengths and limitations. Cross-sectional ecologic studies
use health outcome, exposure, and covariate data available at the community level
(e.g., annual mortality rates and pollutant concentrations), but do not have
individual-level data. Cross-sectional studies may have limited power to evaluate an
extensive set of confounding factors because these studies examine between-subject or
between-location comparisons. Prospective cohort studies include some data collected at
the individual level, typically health outcome data, and in some cases, individual-level
data on exposure and covariates are collected. Time-series and case-crossover studies are
often used to evaluate the relationship between day-to-day changes in air pollution
exposures and a specific health outcome at the population-level (i.e., mortality, hospital
admissions, or emergency department visits). Panel studies may include repeated
measurements of health outcomes (e.g., respiratory symptoms, heart rate variability) at
lii
-------
the individual level and include exposure data at the individual- or group-level. "Natural
experiments" offer the opportunity to investigate changes in health related to a change in
exposure, such as closure of a pollution source.
When evaluating the collective body of epidemiologic studies, many study design factors
and limitations must be considered to properly inform their interpretation. One key
consideration is the evaluation of the potential independent contribution of the criteria
pollutant to a health outcome when the criteria pollutant is a component of a complex air
pollutant mixture. Reported effect estimates in epidemiologic studies may reflect
(1) independent effects on health outcomes, (2) effects of the pollutant acting as an
indicator of a copollutant or a complex ambient air pollution mixture, and (3) effects
resulting from interactions between that pollutant and copollutants.
The third main type of health effects evidence, animal toxicological studies, provides
information on the biological action of a pollutant under controlled and monitored
exposure circumstances. Although biological differences among species must be taken
into account, animal toxicological studies contribute to our understanding of potential
health effects, exposure-response relationships, and modes of action. Further, animal
models can inform determinations of factors that may increase or decrease the risk of
health effects in certain populations. These studies evaluate the effects of exposures to a
variety of pollutants in a highly controlled laboratory setting and allow exploration of
toxicological pathways or mechanisms by which a pollutant may cause effects.
Understanding the biological mechanisms underlying various health outcomes can be
crucial in establishing or negating causality. In the absence of human studies data,
extensive, well-conducted animal toxicological studies can support determinations of
causality, if the evidence base indicates that similar responses are expected in humans
under ambient exposure conditions.
Interpretations of animal toxicological studies are affected by limitations associated with
extrapolation between animal and human responses. The differences between humans
and other species have to be considered, including metabolism, hormonal regulation,
breathing pattern, and differences in lung structure and anatomy. Also, in spite of a high
degree of homology and the existence of a high percentage of orthologous genes across
humans and rodents (particularly mice), extrapolation of molecular alterations at the gene
or protein level is complicated by species-specific differences in transcriptional
regulation and/or signaling. Given these differences, uncertainties are associated with
quantitative extrapolations of observed pollutant-induced pathophysiological alterations
between laboratory animals and humans, as those alterations are under the control of
widely varying biochemical, endocrine, and neuronal factors.
liii
-------
Ecological- and Welfare-specific Considerations
For ecological effects assessment, both laboratory and field studies (including field
experiments and observational studies) can provide useful data for causal determination.
Because conditions can be controlled in laboratory studies, responses may be less
variable and smaller effects may be easier to detect. However, the control conditions may
limit the range of responses (e.g., animals may not be able to seek alternative food
sources) or incompletely reflect pollutant bioavailability, so the responses under
controlled conditions may not reflect responses that would occur in the natural
environment. In addition, larger scale processes are difficult to reproduce in the
laboratory.
Field observational studies measure biological changes in uncontrolled situations with
high natural variability (in organismal genetics or in abiotic seasonal, climatic, or
soil-related factors) and describe an association between a disturbance and an ecological
effect. Field data can provide important information to assess multiple stressors or
circumstances where site-specific factors significantly influence exposure. Field data are
also often useful for analyzing pollutant effects at larger geographic scales and higher
levels of biological organization. However, because conditions are not controlled,
variability of the response is expected to be higher and may mask effects. Field surveys
are most useful for linking stressors with effects when stressor and effect levels are
measured concurrently. The presence of confounding factors can make it difficult to
attribute observed effects to specific stressors.
Ecological impacts of pollutants are also evaluated in studies "intermediate" between the
lower variability typically associated with laboratory exposures and high natural
variability usually found in field studies. Some studies use environmental media collected
from the field to examine the biological responses under controlled laboratory conditions.
Other studies are experiments performed in the natural environment that control for
some, but not all, of the environmental or genetic variability (e.g., mesocosm studies).
This type of study in manipulated natural environments can be considered a hybrid
between a field experiment and laboratory study because some sources of response
variation are removed through use of control conditions, while others are included to
mimic natural variation. Such studies make it possible to observe community and/or
ecosystem dynamics and provide strong evidence for causality when combined with
findings of studies that have been made under more controlled conditions.
liv
-------
b. Considerations in Developing Scientific
Conclusions and Causal Determinations
In its evaluation and integration of the scientific evidence on health or welfare effects of
criteria pollutants, the U.S. EPA determines the weight of evidence in support of
causation and characterizes the strength of any resulting causal classification. The
U.S. EPA also evaluates the quantitative evidence and draws scientific conclusions, to the
extent possible, regarding the concentration-response relationships and the loads to
ecosystems, exposures, doses or concentrations, exposure duration, and pattern of
exposures at which effects are observed.
Approaches to assessing the separate and combined lines of human health evidence
(e.g., epidemiologic, controlled human exposure, animal toxicological studies) have been
formulated by a number of regulatory and science agencies, including the National
Academy of Sciences (NAS) Institute of Medicine (TOM. 2008). the International
Agency for Research on Cancer (IARC. 2006). the U.S. EPA (2005). and the Centers for
Disease Control and Prevention; (CDC. 2004). Causal inference criteria have also been
described for ecological effects evidence (U.S. EPA. 1998a; Fox. 1991). These
formalized approaches offer guidance for assessing causality. The frameworks of each
are similar in nature, although adapted to different purposes, and have proven effective in
providing a uniform structure and language for causal determinations.
The 1964 Surgeon General's report on tobacco smoking defined "cause" as a
"significant, effectual relationship between an agent and an associated disorder or disease
in the host" (HEW. 1964). More generally, a cause is defined as an agent that brings
about an effect or a result. An association is the statistical relationship among variables,
but alone, it is insufficient proof of a causal relationship between an exposure and a
health outcome. Unlike an association, a causal claim supports the creation of
counterfactual claims; that is, a claim about what the world would have been like under
different or changed circumstances (IOM. 2008).
Many of the health and environmental outcomes reported in studies have complex
etiologies. Diseases such as asthma, coronary heart disease, or cancer are typically
initiated by multiple agents. Outcomes depend on a variety of factors, such as age,
genetic background, nutritional status, immune competence, and social factors (IOM.
2008; Gee and Payne-Sturges. 2004). Effects on ecosystems are also often multifactorial
with a complex web of causation. Further, exposure to a combination of agents could
cause synergistic or antagonistic effects. Thus, the observed risk may represent the net
effect of many actions and counteractions.
Iv
-------
To aid judgment, various "aspects"1 of causality have been discussed by many
philosophers and scientists. The 1964 Surgeon General's report on tobacco smoking
discussed criteria for the evaluation of epidemiologic studies, focusing on consistency,
strength, specificity, temporal relationship, and coherence (HEW. 1964). Sir Austin
Bradford Hill (Hill, 1965) articulated aspects of causality in epidemiology and public
health that have been widely used (lOM. 2008: IARC. 2006: U.S. EPA. 2005: CDC.
2004). These aspects (Hill. 1965) have been modified (Table I) for use in causal
determinations specific to health and welfare effects for pollutant exposures (U.S. EPA.
2009a).2 Although these aspects provide a framework for assessing the evidence, they do
not lend themselves to being considered in terms of simple formulas or fixed rules of
evidence leading to conclusions about causality (Hill. 1965). For example, one cannot
simply count the number of studies reporting statistically significant results or
statistically nonsignificant results and reach credible conclusions about the relative
weight of evidence and the likelihood of causality. Rather, these aspects provide a
framework for systematic appraisal of the body of evidence, informed by peer and public
comment and advice, which includes weighing alternative views on controversial issues.
In addition, it is important to note that the aspects in Table I cannot be used as a strict
checklist, but rather to determine the weight of evidence for inferring causality. In
particular, not meeting one or more of the principles does not automatically preclude a
determination of causality [see discussion in (CDC. 2004)1.
1 The "aspects" described by Sir Austin Bradford Hill (Hill. 1965) have become, in the subsequent literature, more
commonly described as "criteria." The original term "aspects" is used here to avoid confusion with "criteria" as it is
used, with different meaning, in the Clean Air Act.
2 The Hill aspects were developed for interpretation of epidemiologic results. They have been modified here for use
with a broader array of data (i.e., epidemiologic, controlled human exposure, ecological, animal lexicological
studies, in vitro data) and to be more consistent with the EPA Guidelines for Carcinogen Risk Assessment.
Ivi
-------
Table I
Aspects to aid in judging causality.
Aspect
Description
Consistency
An inference of causality is strengthened when a pattern of elevated risks is observed
across several independent studies. The reproducibility of findings constitutes one of the
strongest arguments for causality. Statistical significance is not the sole criterion by which
the presence or absence of an effect is determined. If there are discordant results among
investigations, possible reasons such as differences in exposure, confounding factors, and
the power of the study are considered.
Coherence
An inference of causality from one line of evidence (e.g., epidemiologic, controlled human
exposure, animal, welfare studies) may be strengthened by other lines of evidence that
support a cause-and-effect interpretation of the association. There may be coherence in
demonstrating effects from evidence across various fields and/or across multiple study
designs or related health endpoints within one scientific line of evidence. For example,
evidence on welfare effects may be drawn from a variety of experimental approaches
(e.g., greenhouse, laboratory, field) and subdisciplines of ecology (e.g., community
ecology, biogeochemistry, paleontological/historical reconstructions).
Biological plausibility
An inference of causality is strengthened by results from experimental studies or other
sources demonstrating biologically plausible mechanisms. A proposed mechanism, which
is based on experimental evidence and which links exposure to an agent to a given effect,
is an important source of support for causality.
Biological gradient
(exposure-response
relationship)
A well-characterized exposure-response relationship (e.g., increasing effects associated
with greater exposure) strongly suggests cause and effect, especially when such
relationships are also observed for duration of exposure (e.g., increasing effects observed
following longer exposure times).
Strength of the
observed association
The finding of large, precise risks increases confidence that the association is not likely
due to chance, bias, or other factors. However, it is noted that a small magnitude in an
effect estimate may or may not represent a substantial effect in a population.
Experimental evidence
Strong evidence for causality can be provided through "natural experiments" when a
change in exposure is found to result in a change in occurrence or frequency of health or
welfare effects.
Temporality of the
observed association
Evidence of a temporal sequence between the introduction of an agent and appearance of
the effect constitutes another argument in favor of causality.
Specificity of the
observed association
Evidence linking a specific outcome to an exposure can provide a strong argument for
causation. However, it must be recognized that rarely, if ever, does exposure to a pollutant
invariably predict the occurrence of an outcome, and that a given outcome may have
multiple causes.
Analogy
Structure activity relationships and information on the agent's structural analogs can
provide insight into whether an association is causal. Similarly, information on mode of
action for a chemical, as one of many structural analogs, can inform decisions regarding
likely causality.
Mi
-------
Consistency of findings across studies is informed by the repeated observation of effects
or associations across multiple independent studies. Further strength is provided by
reproducibility of findings in different populations under different circumstances.
However, discordant results among independent investigations may be explained by
differences in study methods, random errors, exposure, confounding factors, or study
power, and thus may not be used to rule out a causal connection.
In evaluating the consistency of findings across studies, the U.S. EPA emphasizes
examination of the pattern of results across various studies and does not focus solely on
statistical significance or the magnitude of the direction of the association as criteria of
study reliability. Statistical significance is influenced by a variety of factors including,
but not limited to, the size of the study, exposure and outcome measurement error, and
statistical model specifications. Statistical significance may be informative; however, it is
just one of the means of evaluating confidence in the observed relationship and assessing
the probability of chance as an explanation. Other indicators of reliability such as the
consistency and coherence of a body of studies as well as other confirming data may be
used to justify reliance on the results of a body of epidemiologic studies, even if results in
individual studies lack statistical significance. Traditionally, statistical significance is
used to a larger extent to evaluate the findings of controlled human exposure and animal
toxicology studies. Understanding that statistical inferences may result in both false
positives and false negatives, the U.S. EPA considers both trends in data and
reproducibility of results. Thus, in drawing judgments regarding causality, the U.S. EPA
emphasizes statistically significant findings from experimental studies but does not limit
its focus or consideration to statistically significant results in epidemiologic studies.
In evaluating the strength of the observed association, the U.S. EPA considers both the
magnitude and statistical precision (i.e., width of confidence interval) of the association
in epidemiologic studies. In a large study that accounts for several potential confounding
factors, a strong association can serve to increase confidence that a finding is not due to a
weak unmeasured confounder, chance, or other biases. However, in a study that accounts
for several potential confounding factors and other sources of bias, a weak association
does not rule out a causal connection. The health effects evaluated in the ISAs tend to
have multiple risk factors that likely vary in strength of effect, and the magnitude of
effect of air pollution exposure will depend on the prevalence of other risk factors in the
study population. Further, a small effect size can be important from a public health
impact perspective. The air pollution-related change in a health effect observed in a study
can represent a shift in the distribution of responses in the study population and
potentially an increase in the proportion of individuals with clinically important effects.
Iviii
-------
In making judgments regarding causality, the U.S. EPA considers biological plausibility
of effects resulting from air pollutant exposure. Experimental results from in vivo studies
involving animal models and humans, as well as from in vitro studies when appropriate,
may be used to establish biological plausibility and to interpret other lines of evidence
(e.g., health effects from epidemiologic studies). Biological plausibility is often provided
from understanding the mode of action by which exposure to a pollutant leads to health
effects. This understanding may encompass several different levels of biological
organization including, but not limited to, molecular and cellular events in the pathways
leading to disease. While a complete understanding of the mode of action is not
considered necessary for making causal determinations within the ISA, biological
plausibility plays a key role.
c. Framework for Causal Determinations
In the ISA, the U.S. EPA assesses the body of relevant literature, building upon evidence
available during previous NAAQS reviews, to draw conclusions on the causal
relationships between relevant pollutant exposures and health or environmental effects.
ISAs use a five-level hierarchy that classifies the weight of evidence for causation.1 This
weight-of-evidence evaluation is based on the integration of findings from various lines
of evidence across health and environmental effect disciplines that are integrated into a
qualitative statement about the overall weight of the evidence and causality. The five
descriptors for causal determination are described in Table II.
1 The CDC and IOM frameworks use a four-category hierarchy for the strength of the evidence. A five-level
hierarchy is used here to be consistent with the five-level hierarchy used in the U.S. EPA Guidelines for Carcinogen
Risk Assessment and to provide a more nuanced set of categories.
lix
-------
Table II Weight of evidence for causal determination.
Health Effects
Ecological and Welfare Effects
Causal Evidence is sufficient to conclude that there is a causal
relationship relationship with relevant pollutant exposures
(e.g., doses or exposures generally within one to two
orders of magnitude of recent concentrations). That is,
the pollutant has been shown to result in health effects
in studies in which chance, confounding, and other
biases could be ruled out with reasonable confidence.
For example: (1) controlled human exposure studies
that demonstrate consistent effects, or
(2) observational studies that cannot be explained by
plausible alternatives or that are supported by other
lines of evidence (e.g., animal studies or mode of
action information). Generally, the determination is
based on multiple high-quality studies conducted by
multiple research groups.
Evidence is sufficient to conclude that there is a
causal relationship with relevant pollutant exposures.
That is, the pollutant has been shown to result in
effects in studies in which chance, confounding, and
other biases could be ruled out with reasonable
confidence. Controlled exposure studies (laboratory or
small- to medium-scale field studies) provide the
strongest evidence for causality, but the scope of
inference may be limited. Generally, the determination
is based on multiple studies conducted by multiple
research groups, and evidence that is considered
sufficient to infer a causal relationship is usually
obtained from the joint consideration of many lines of
evidence that reinforce each other.
Likely to be a
causal
relationship
Evidence is sufficient to conclude that a causal
relationship is likely to exist with relevant pollutant
exposures. That is, the pollutant has been shown to
result in health effects in studies where results are not
explained by chance, confounding, and other biases,
but uncertainties remain in the evidence overall. For
example: (1) observational studies show an
association, but copollutant exposures are difficult to
address and/or other lines of evidence (controlled
human exposure, animal, or mode of action
information) are limited or inconsistent, or (2) animal
toxicological evidence from multiple studies from
different laboratories demonstrate effects, but limited or
no human data are available. Generally, the
determination is based on multiple high-quality studies.
Evidence is sufficient to conclude that there is a likely
causal association with relevant pollutant exposures.
That is, an association has been observed between
the pollutant and the outcome in studies in which
chance, confounding, and other biases are minimized
but uncertainties remain. For example, field studies
show a relationship, but suspected interacting factors
cannot be controlled, and other lines of evidence are
limited or inconsistent. Generally, the determination is
based on multiple studies by multiple research groups.
Suggestive of,
but not
sufficient to
infer, a causal
relationship
Evidence is suggestive of a causal relationship with
relevant pollutant exposures but is limited, and chance,
confounding, and other biases cannot be ruled out. For
example: (1) when the body of evidence is relatively
small, at least one high-quality epidemiologic study
shows an association with a given health outcome
and/or at least one high-quality toxicological study
shows effects relevant to humans in animal species, or
(2) when the body of evidence is relatively large,
evidence from studies of varying quality is generally
supportive but not entirely consistent, and there may be
coherence across lines of evidence (e.g., animal
studies or mode of action information) to support the
determination.
Evidence is suggestive of a causal relationship with
relevant pollutant exposures, but chance,
confounding, and other biases cannot be ruled out.
For example, at least one high-quality study shows an
effect, but the results of other studies are inconsistent.
Inadequate to
infer a causal
relationship
Evidence is inadequate to determine that a causal
relationship exists with relevant pollutant exposures.
The available studies are of insufficient quantity,
quality, consistency, or statistical power to permit a
conclusion regarding the presence or absence of an
effect.
Evidence is inadequate to determine that a causal
relationship exists with relevant pollutant exposures.
The available studies are of insufficient quality,
consistency, or statistical power to permit a conclusion
regarding the presence or absence of an effect.
Not likely to be
a causal
relationship
Evidence indicates there is no causal relationship with
relevant pollutant exposures. Several adequate studies,
covering the full range of levels of exposure that human
beings are known to encounter and considering at-risk
populations and lifestages, are mutually consistent in
not showing an effect at any level of exposure.
Evidence indicates there is no causal relationship with
relevant pollutant exposures. Several adequate
studies examining relationships with relevant
exposures are consistent in not showing an effect at
any level of exposure.
Ix
-------
This standardized language was drawn from sources across the federal government and
wider scientific community, especially the U.S. EPA Guidelines for Carcinogen Risk
Assessment (U.S. EPA. 2005). U.S. Surgeon General's report, The Health Consequences
of Smoking (CDC. 2004). and NAS IOM document, Improving the Presumptive
Disability Decision-Making Process for Veterans (IOM. 2008). a comprehensive report
on evaluating causality.
This framework:
• describes the kinds of scientific evidence used in making determinations on causal
relationships between exposure and health or welfare effects,
• summarizes the key aspects of the evaluation of evidence necessary to reach a
conclusion about the existence of a causal relationship,
• identifies issues and approaches related to uncertainty, and
• classifies and characterizes the weight of evidence in support of a general causal
determination.
Determination of causality involves evaluating and integrating evidence for different
types of health, ecological, or welfare effects associated with short- and long-term
exposure periods. In drawing conclusions regarding causality, evidence is evaluated for
major outcome categories or groups of related endpoints (e.g., respiratory effects,
vegetation growth), integrating evidence from across disciplines, and evaluating the
coherence of evidence across a spectrum of related endpoints. In discussing the causal
determination, the U.S. EPA characterizes the evidence on which the judgment is based,
including strength of evidence for individual endpoints within the outcome category or
group of related endpoints.
In drawing judgments regarding causality for the criteria air pollutants, the ISA focuses
on evidence of effects in the range of relevant pollutant exposures or doses and not on
determination of causality at any particular dose. Emphasis is placed on evidence of
effects at doses (e.g., blood Pb concentration) or exposures (e.g., air concentrations) that
are relevant to, or somewhat above, those currently experienced by the population or that
exist in the environment. The extent to which studies of higher concentrations are
considered varies by pollutant and major outcome category, but generally includes those
with doses or exposures in the range of one to two orders of magnitude above current or
ambient conditions to account for intraspecies variability and toxicokinetic or
toxicodynamics differences between experimental animals and humans. Studies that use
higher doses or exposures may also be considered to the extent that they provide useful
information to inform understanding of mode of action, interspecies differences, or
factors that may increase risk of effects for a population and if biological mechanisms
have not been demonstrated to differ based on exposure concentration. Thus, a causal
Ixi
-------
determination is based on weight-of-evidence evaluation for health or welfare effects,
focusing on the evidence from exposures or doses generally ranging from recent ambient
concentrations to one or two orders of magnitude above recent ambient concentrations.
In addition, the U.S. EPA evaluates evidence relevant to understanding the quantitative
relationships between pollutant exposures and health or welfare effects. This includes
evaluating the form of concentration-response or dose-response relationships and, to the
extent possible, drawing conclusions on the concentrations at which effects are observed.
The ISA also draws scientific conclusions regarding important exposure conditions for
effects and populations and lifestages that may be at greater risk for effects, as described
in the following two sections on public health and public welfare impacts.
6. Public Health Impact
Once a determination is made regarding the causality of relationship between the
pollutant and outcome category, the public health impact of exposure to the pollutant is
evaluated. Important questions regarding the public health impact include:
• What populations and lifestages appear to be differentially affected (i.e., at greater
or less risk of experiencing effects)?
• What exposure conditions (dose or exposure, duration, and pattern) are important?
• What is the severity of the effect (e.g., clinical relevance)?
• What is the concentration-response, exposure-response, or dose-response
relationship in the human population?
• What is the interrelationship between incidence and severity of effect?
To address these questions, the entirety of quantitative evidence is evaluated to
characterize pollutant concentrations and exposure durations at which effects were
observed for exposed populations, including populations and lifestages potentially at
increased risk. To accomplish this, evidence is considered from multiple and diverse
types of studies, and a study or set of studies that best approximates the
concentration-response relationships between health outcomes and the pollutant may be
identified. Controlled human exposure studies provide the most direct and quantifiable
exposure-response data on the human health effects of pollutant exposures, although they
tend to examine potential at-risk populations and lifestages to a limited extent and tend to
have small sample sizes for between-group comparisons. To the extent available, the ISA
evaluates results from epidemiologic studies that characterize the shape of the
relationship between a pollutant and a health outcome. Animal data may also inform
evaluation of concentration-response relationships, particularly relative to modes of
action and characteristics of at-risk populations.
Mi
-------
a. Approach to Identifying, Evaluating, and
Characterizing At-Risk Factors
A critical part of assessing the public health impact of an air pollutant is the
identification, evaluation, and characterization of populations potentially at greater risk of
an air pollutant-related health effect. Under the CAA, the primary NAAQS are intended
to protect public health with an adequate margin of safety. In doing so, protection is
provided for both the population as a whole and those groups potentially at increased risk
for health effects from exposure to a criteria air pollutant. To inform decisions on the
NAAQS, the ISA evaluates the currently available information regarding those factors
(e.g., lifestage, pre-existing disease) that could contribute to portions of the population
being at greater risk for an air pollutant-related health effect.
Studies often use a variety of terms to classify factors and subsequently populations that
may be at increased risk of an air pollutant-related health effect, including "susceptible,"
"vulnerable," "sensitive," and "at-risk," with recent literature introducing the term
"response-modifying factor" (Vinikoor-Imler et al.. 2014; Sacks et al.. 2011; U.S. EPA.
201 Ob. 2009a). The inconsistency in the definitions for each of these terms across the
scientific literature has shifted the focus away from answering the key questions: Which
populations are at increased risk and what evidence forms the basis of this conclusion
(Vinikoor-Imler et al.. 2014)? Due to the lack of a consensus on terminology in the
scientific community, the term "susceptible populations" was used in reviews and
previous ISAs (Sacks et al.. 2011; U.S. EPA. 2010b. 2009a) to encompass these various
factors. However, it was recognized that even using the term "susceptible populations"
was problematic because it often refers to populations at increased risk specifically due to
biological or intrinsic factors such as pre-existing disease or lifestage. As such, starting
with the ISA for Ozone and Related Photochemical Oxidants (U.S. EPA. 2013d). the
terminology "at-risk" was introduced to define populations and lifestages potentially at
increased risk of an air pollutant-related health effect. In assessing the overall public
health impact of an air pollutant, the ISA focuses on identifying, evaluating, and
characterizing "at-risk" factors to address the main question of what populations and
lifestages are at increased risk of an air pollutant-related health effect. Each "at-risk"
factor is evaluated with a focus on identifying whether the factor contributes to a
population at increased risk of an air pollutant-related health effect. Some factors may
lead to a reduction in risk, and these are acknowledged during the evaluation process.
However, for the purposes of identifying those populations or lifestages at increased risk
to inform decisions on the NAAQS, the ISA focuses on characterizing those factors that
may increase risk.
Ixiii
-------
A population or lifestage may be at increased risk for various reasons, which generally
are grouped into four broad categories. The first category of factors often is referred to as
intrinsic. Intrinsic factors can increase risk for an effect through a biological mechanism
and include genetic or developmental factors, race, sex, lifestage, or the presence of
pre-existing diseases. For example, people in this category would have a steeper
concentration-risk relationship and a greater or more severe effect at a given pollutant
concentration compared to those not in the category. The second category often is
referred to as extrinsic or nonbiological. These factors include SES (e.g., educational
attainment, income, access to healthcare), activity pattern, and exercise level. The third
category includes factors that can increase risk by increasing internal dose at a given
exposure concentration. Individuals in this category could have a greater dose of
delivered pollutant because of breathing patterns and could include children who are
typically more active outdoors. In addition, some groups could have greater exposure
(concentration x time) regardless of the delivered dose, such as outdoor workers. The
final category encompasses factors that may increase risk for experiencing a greater
exposure based on exposure to a higher concentration. For example, populations that live
near roadways could be exposed to higher pollutant concentrations. Some factors
described above are multifaceted and may influence the risk of an air pollutant-related
health effect through a combination of ways (e.g., SES). Additionally, it is recognized
that some portions of the population or lifestages may be at increased risk of an air
pollutant-related health effect because they experience insults from a combination of
factors. The emphasis is to identify and understand the factors that potentially increase
the risk of air pollutant-related health effects, regardless of whether the increased risk is
due to intrinsic factors, extrinsic factors, increased dose/exposure, or a combination due
to the often interconnectedness of factors.
To identify at-risk factors that potentially lead to some portions of the population being at
increased risk of air pollution-related health effects, the evidence is systematically
evaluated across relevant scientific disciplines (i.e., exposure sciences, dosimetry,
toxicology, epidemiology). The evaluation process consists of evaluating studies that
conduct stratified analyses (i.e., epidemiologic, controlled human exposure) to compare
populations or lifestages exposed to similar air pollutant concentrations within the same
study design. Experimental studies also provide an important line of evidence in
evaluating factors that can lead to increased risk of an air pollutant-related health effect.
Specifically, toxicological studies conducted using animal models of disease and
controlled human exposure studies that examine individuals with underlying disease or
genetic polymorphisms can provide coherence with the health effects observed in
epidemiologic studies as well as an understanding of biological plausibility. The potential
increased risk of an air pollutant-related health effect may also be determined from
studies that examine factors that result in differential air pollutant exposures. The
Ixiv
-------
characterization of each at-risk factor consists of evaluating the evidence across scientific
disciplines and assessing the overall confidence that a specific factor may result in a
population or lifestage being at increased risk of an air pollutant-related health effect. The
categories considered for describing the potential increased risk of an air pollutant-related
health effect are "adequate evidence," "suggestive evidence," "inadequate evidence," and
"evidence of no effect." They are described in more detail in Table III.
Table
Characterization of evidence for potential at-risk factors.
Classification
Health Effects
Adequate evidence
There is substantial, consistent evidence within a discipline to conclude that a factor results in
a population or lifestage being at increased risk of air pollutant-related health effect(s) relative
to some reference population or lifestage. Where applicable, this evidence includes coherence
across disciplines. Evidence includes multiple high-quality studies.
Suggestive The collective evidence suggests that a factor results in a population or lifestage being at
evidence increased risk of air pollutant-related health effect(s) relative to some reference population or
lifestage, but the evidence is limited due to some inconsistency within a discipline or, where
applicable, a lack of coherence across disciplines.
Inadequate The collective evidence is inadequate to determine whether a factor results in a population or
evidence lifestage being at increased risk of air pollutant-related health effect(s) relative to some
reference population or lifestage. The available studies are of insufficient quantity, quality,
consistency, and/or statistical power to permit a conclusion to be drawn.
Evidence of no There is substantial, consistent evidence within a discipline to conclude that a factor does not
effect result in a population or lifestage being at increased risk of air pollutant-related health effect(s)
relative to some reference population or lifestage. Where applicable, the evidence includes
coherence across disciplines. Evidence includes multiple high-quality studies.
b.
Evaluating Adversity of Human Health Effects
In evaluating health evidence, a number of factors can be considered in delineating
between adverse and nonadverse health effects resulting from exposure to air pollution.
Some health outcomes, such as hospitalization for respiratory or cardiovascular diseases,
are clearly adverse. It is more difficult to determine the extent of change that constitutes
adversity in more subtle health measures. These more subtle health effects include a wide
variety of responses, such as alterations in markers of inflammation or oxidative stress,
changes in pulmonary function or heart rate variability, or alterations in neurocognitive
function measures. The challenge is to determine the magnitude of change in these
measures when there is no clear point at which a change becomes adverse. The extent to
which a change in health measure constitutes an adverse health effect may vary between
Ixv
-------
populations and lifestages. Some changes that may not be considered adverse in healthy
individuals would be potentially adverse in more at-risk individuals.
Professional scientific societies may evaluate the magnitude of change in an outcome or
event that is considered adverse. For example, in an official statement titled What
Constitutes an Adverse Health Effect of Air Pollution? (ATS. 2000b). the American
Thoracic Society described transient decrements in lung function as adverse when
accompanied by clinical symptoms. Additionally, an air pollution-induced shift in the
population distribution of a given risk factor for a health outcome was viewed as adverse,
even though it may not increase the risk of any one individual to an unacceptable level.
For example, a population with asthma could have a distribution of lung function such
that no identifiable individual has a level associated with significant impairment.
Exposure to air pollution could shift the distribution such that no identifiable individual
experiences clinically relevant effects. This shift toward decreased lung function,
however, could be considered adverse because individuals within the population would
have diminished reserve function and therefore would be at increased risk to further
environmental insult. The committee also observed that elevations of biomarkers, such as
cell number and types, cytokines, and reactive oxygen species, may signal risk for
ongoing injury and clinical effects or may simply indicate transient responses that can
provide insights into mechanisms of injury, thus illustrating the lack of clear boundaries
that separate adverse from nonadverse effects.
The more subtle health outcomes may be connected mechanistically to health events that
are clearly adverse. For example, air pollution may affect markers of transient myocardial
ischemia such as ST-segment (segment of the electrocardiograph between the end of the
S wave and beginning of the T wave) abnormalities or onset of exertional angina. These
effects may not be apparent to the individual, yet may still increase the risk of a number
of cardiac events, including myocardial infarction and sudden death. Thus, small changes
in physiological measures may not appear to be clearly adverse when considered alone,
but may be a part of a coherent and biologically plausible chain of related health
outcomes that range up to responses that are very clearly adverse, such as hospitalization
or mortality.
c. Concentration-Response Relationships
An important consideration in characterizing the public health impacts associated with
exposure to a pollutant is whether the concentration-response relationship is linear across
the range of concentrations or if nonlinear relationships exist along any part of this range.
The shape of the concentration-response curve at and below the level of the current
Ixvi
-------
NAAQS is of particular interest. Various sources of variability and uncertainty, such as
low data density in the lower concentration range, possible influence of exposure
measurement error, and variability among individuals with respect to air pollution health
effects, tend to smooth and "linearize" the concentration-response function and thus can
obscure the existence of a threshold or nonlinear relationship. Because individual
thresholds vary from person-to-person due to individual differences such as genetic
differences or pre-existing disease conditions (and even can vary from one time to
another for a given person), it can be difficult to demonstrate that a threshold exists in a
population study. These sources of variability and uncertainty may explain why the
available human data at ambient concentrations for some environmental pollutants
(e.g., PM, Os, Pb, environmental tobacco smoke, radiation) do not exhibit
population-level thresholds for cancer or noncancer health effects, even though likely
mechanisms include nonlinear processes for some key events.
7. Public Welfare Impact
Once a determination is made regarding the causality of relationships between the
pollutant and outcome category, important questions regarding the public welfare impact
include:
• What endpoints or services appear to be differentially affected (i.e., at greater or
less risk of experiencing effects)? What elements of the ecosystem (e.g., types,
regions, taxonomic groups, populations, functions) appear to be affected, or are
more sensitive to effects? Are there differences between locations or materials in
welfare effects responses, such as impaired visibility or materials damage?
• What is concluded from the evidence with regard to other types of welfare
effects?
• Under what exposure conditions (amount deposited or concentration, duration,
and pattern) are effects seen?
• What is the shape of the concentration-response, exposure-response, or
dose-response relationship?
To address these questions, the entirety of quantitative evidence is evaluated to
characterize pollutant concentrations and exposure durations at which effects were
observed. To accomplish this, evidence is considered from multiple and diverse types of
studies, and a study or set of studies that best approximates the concentration-response
relationships between welfare outcomes and the pollutant may be identified. Controlled
experimental studies provide the most direct and quantifiable exposure-response data on
the effects of pollutant exposures. To the extent available, the ISA also evaluates results
from less controlled field studies that characterize the shape of the relationship between a
pollutant and an outcome. Other types of data may also inform evaluation of
Ixvii
-------
concentration-response relationships, particularly relative to modes of action and
characteristics of at-risk ecosystems.
a. Evaluating Adversity of Ecological and Other
Welfare Effects
The final step in assessing the public welfare impact of an air pollutant is the evaluation
of the level considered to be adverse. A secondary standard, as defined in
Section 109(b)(2) of the CAA must "specify a level of air quality the attainment and
maintenance of which, in the judgment of the Administrator, based on such criteria, is
requisite to protect the public welfare from any known or anticipated adverse effects
associated with the presence of such air pollutant in the ambient air." In setting standards
that are "requisite" to protect public health and welfare, as provided in Section 109(b),
the U.S. EPA's task is to establish standards that are neither more nor less stringent than
necessary for these purposes.
Adversity of ecological effects can be understood in terms ranging in biological level of
organization from the cellular level to the individual organism and to the population,
community, and ecosystem levels. In the context of ecology, a population is a group of
individuals of the same species, and a community is an assemblage of populations of
different species that inhabit an area and interact with one another. An ecosystem is the
interactive system formed from all living organisms and their abiotic (physical and
chemical) environment within a given area (TPCC. 2007). The boundaries of what could
be called an ecosystem are somewhat arbitrary, depending on the focus of interest or
study. Thus, the extent of an ecosystem may range from very small spatial scales to,
ultimately, the entire Earth (IPCC. 2007V
Effects on an individual organism are generally not considered to be adverse to public
welfare. However if effects occur to enough individuals within a population, then
communities and ecosystems may be disrupted. Changes to populations, communities,
and ecosystems can in turn result in an alteration of ecosystem processes. Ecosystem
processes are defined as the metabolic functions of ecosystems, including energy flow,
elemental cycling, and the production, consumption, and decomposition of organic matter
(U.S. EPA. 2002). Growth, reproduction, and mortality are species-level endpoints that
may be clearly linked to community and ecosystem effects and are considered to be
adverse when negatively affected. Other endpoints, such as changes in behavior and
physiological stress, can decrease ecological fitness of an organism but are harder to link
unequivocally to effects at the population, community, and ecosystem level. Support for
consideration of adversity beyond the species level by making explicit the linkages
Ixviii
-------
between stress-related effects at the species and effects at the ecosystem level is found in
A Framework for Assessing and Reporting on Ecological Condition: an SAB report (U.S.
EPA. 2002). Additionally, the National Acid Precipitation Assessment Program
(NAPAP. 1991) uses the following working definition of "adverse ecological effects" in
the preparation of reports to Congress mandated by the CAA: "any injury (i.e., loss of
chemical or physical quality or viability) to any ecological or ecosystem component, up
to and including the regional level, over both long and short terms."
Beyond species-level impacts, consideration of ecosystem services allows for evaluation
of how pollutant exposure may adversely impact species or processes of particular
economic or cultural importance to humans. On a broader scale, ecosystem services may
provide indicators for ecological impacts. Ecosystem services are the benefits that people
obtain from ecosystems (UNEP. 2003). According to the Millennium Ecosystem
Assessment, ecosystem services include "provisioning services such as food and water;
regulating services such as regulation of floods, drought, land degradation, and disease;
supporting services such as soil formation and nutrient cycling; and cultural services such
as recreational, spiritual, religious, and other nonmaterial benefits" (UNEP. 2003). For
example, a more subtle ecological effect of pollution exposure may result in a clearly
adverse impact on ecosystem services if it results in a population decline in a species that
is recreationally or culturally important.
A consideration in evaluating adversity of climate-related effects is that criteria air
pollutants have both direct and indirect effects on radiative forcing. For example, CO has
a relatively small direct forcing effect, but it influences the concentrations of other
atmospheric species, such as Os and methane (CH4), which are important contributors to
climate forcing. PM has both direct and indirect effects. For example, black carbon and
sulfate contribute directly to warming and cooling, respectively, while aerosols are
involved in cloud formation, which affect climate indirectly. Thus, it is crucial to
consider the role of multiple pollutants together in evaluating the climate impact of
criteria pollutants. Although climate effects of criteria air pollutants impact terrestrial and
aquatic environments in diverse ways over multiple time scales, their effect on
temperature is the main metric of adversity, with some consideration of proximate effects
such as precipitation and relatively rapid feedbacks impacting the composition of the
troposphere. Downstream effects such as land use changes are more difficult to link back
to changes in concentrations of individual pollutants regulated under the NAAQS. The
relative adversity of U.S. versus global emissions and concentrations is informed by
regional climate modeling studies, including consideration of uncertainty and spatial and
temporal variability.
Ixix
-------
The adversity of visibility impacts may be expressed in terms of psychological stress,
such as impairment of aesthetic quality or enjoyment of the environment, or in monetary
terms, such as willingness to pay to improve air quality. Understanding the relationship
between pollutant concentration and perception of visibility, including distinguishing
between concerns about health risks due to air pollution and perceived visibility
impairment, can be crucial in evaluating the level of protection provided by a
welfare-based secondary NAAQS when impacts on visibility are among the welfare
effects that are potentially relevant for a pollutant.
Adversity of materials damage is evaluated considering the impact to human and
economic well-being. Physical damage and soiling impair aesthetic qualities and function
of materials. Additionally, damage to property and cultural heritage sites due to pollutant
deposition may be considered adverse.
b. Quantitative Relationships: Effects on Welfare
Evaluations of causality generally consider the probability of quantitative changes in
welfare effects in response to exposure. A challenge to the quantification of
exposure-response relationships for ecological effects is the great regional and local
spatial variability, as well as temporal variability, in ecosystems. Thus,
exposure-response relationships are often determined for a specific ecological system and
scale, rather than at the national or even regional scale. Quantitative relationships,
therefore, are estimated site by site and may differ greatly between ecosystems.
Ixx
-------
PREFACE
Legislative Requirements for the Review of the National Ambient
Air Quality Standards
Two sections of the Clean Air Act (CAA) govern the establishment, review, and revision
of the National Ambient Air Quality Standards (NAAQS). Section 108 [42 U.S. Code
(U.S.C.) 7408] directs the Administrator of the United States Environmental Protection
Agency (U.S. EPA) to identify and list certain air pollutants and then to issue air quality
criteria for those pollutants. The Administrator is to list those air pollutants that in her
"judgment, cause or contribute to air pollution which may reasonably be anticipated to
endanger public health or welfare;" "the presence of which in the ambient air results from
numerous or diverse mobile or stationary sources;" and "for which ... [the Administrator]
plans to issue air quality criteria ..." [42 U.S.C. 7408(a)(l); (CAA. 1990a)1. Air quality
criteria are intended to "accurately reflect the latest scientific knowledge useful in
indicating the kind and extent of all identifiable effects on public health or welfare, which
may be expected from the presence of [a] pollutant in the ambient air ..." [42 U.S.C.
7408(b)]. Section 109 [42 U.S.C. 7409; (CAA. 1990b)1 directs the Administrator to
propose and promulgate "primary" and "secondary" NAAQS for pollutants for which air
quality criteria are issued. Section 109(b)(l) defines a primary standard as one "the
attainment and maintenance of which in the judgment of the Administrator, based on
such criteria and allowing an adequate margin of safety, are requisite to protect the public
health."1 A secondary standard, as defined in Section 109(b)(2), must "specify a level of
air quality the attainment and maintenance of which, in the judgment of the
Administrator, based on such criteria, is requisite to protect the public welfare from any
known or anticipated adverse effects associated with the presence of [the] air pollutant in
the ambient air."2
The requirement that primary standards provide an adequate margin of safety was
intended to address uncertainties associated with inconclusive scientific and technical
information available at the time of standard setting. It was also intended to provide a
1 The legislative history of Section 109 indicates that a primary standard is to be set at".. .the maximum permissible
ambient air level... which will protect the health of any [sensitive] group of the population," and that for this purpose
"reference should be made to a representative sample of persons comprising the sensitive group rather than to a
single person in such a group" S. Rep. No. 91:1196, 91st Cong., 2d Sess. 10 (1970).
2 Under CAA Section 302(h) [42 U.S.C. 7602(h)], language referring to "effects on welfare" includes, but is not
limited to, "effects on soils, water, crops, vegetation, man-made materials, animals, wildlife, weather, visibility and
climate, damage to and deterioration of property, and hazards to transportation, as well as effects on economic
values and on personal comfort and well-being" (CAA. 2005).
Ixxi
-------
reasonable degree of protection against hazards that research has not yet identified.1 Both
kinds of uncertainty are components of the risk associated with pollution at levels below
those at which human health effects can be said to occur with reasonable scientific
certainty. Thus, in selecting primary standards that provide an adequate margin of safety,
the Administrator is seeking not only to prevent pollution levels that have been
demonstrated to be harmful but also to prevent lower pollutant levels that may pose an
unacceptable risk of harm, even if the risk is not precisely identified as to nature or
degree. The CAA does not require the Administrator to establish a primary NAAQS at a
zero-risk level or at background concentration levels, but rather at a level that reduces
risk sufficiently so as to protect public health with an adequate margin of safety.2 In so
doing, protection is provided for both the population as a whole and those groups
potentially at increased risk for health effects from exposure to the air pollutant for which
each NAAQS is set.
In addressing the requirement for an adequate margin of safety, the U.S. EPA considers
such factors as the nature and severity of the health effects involved, the size of the
sensitive group(s), and the kind and degree of the uncertainties. The selection of any
particular approach to providing an adequate margin of safety is a policy choice left
specifically to the Administrator's judgment.3
In setting standards that are "requisite" to protect public health and welfare as provided in
Section 109(b), the U.S. EPA's task is to establish standards that are neither more nor less
stringent than necessary for these purposes. In so doing, the U.S. EPA may not consider
the costs of implementing the standards.4 Likewise, "[Attainability and technological
feasibility are not relevant considerations in the promulgation of national ambient air
quality standards."5
Section 109(d)(l) requires that "not later than December 31, 1980, and at 5-year intervals
thereafter, the Administrator shall complete a thorough review of the criteria published
under Section 108 and the national ambient air quality standards... and shall make such
revisions in such criteria and standards and promulgate such new standards as may be
1 See Lead Industries Association v. EPA, 647 F.2d 1130, 1154 [(District of Columbia Circuit (D.C. Cir) 1980];
American Petroleum Institute v. Costle, 665 F.2d 1176, 1186 (D.C. Cir. 19&1)', American Farm Bureau Federation
v. EPA, 559 F. 3d 512, 533 (D.C. Cir. 2009); Association of Battery Recyclersv. EPA, 604 F. 3d 613, 617-18 (D.C.
Cir. 2010).
2 See Lead Industries v. EPA, 647 F.2d at 1156 n.51; Mississippi v. EPA, 723 F. 3d 246, 255, 262-63 (D.C. Cir.
2013).
3 See Lead Industries Association v. EPA, 647 F.2d at 1161-62; Mississippi v. EPA, 723 F. 3d at 265.
4 See generally, Whitman v. American Trucking Associations, 531 U.S. 457, 465-472, 475-476 (2001).
5 See American Petroleum Institute v. Costle, 665 F. 2d at 1185.
Ixxii
-------
appropriate..." Section 109(d)(2) requires that an independent scientific review
committee "shall complete a review of the criteria... and the national primary and
secondary ambient air quality standards... and shall recommend to the Administrator any
new... standards and revisions of existing criteria and standards as may be
appropriate...." Since the early 1980s, this independent review function has been
performed by the Clean Air Scientific Advisory Committee (CASAC).1
Overview and History of the Review of the Primary National
Ambient Air Quality Standards for Nitrogen Dioxide
Nitrogen dioxide (NCh) is the indicator for gaseous oxides of nitrogen [e.g., NO2, nitric
oxide (NO)]. Consistent with Section 108(c) of the CAA (42 U.S.C.21 7408), the U.S.
EPA considers the term oxides of nitrogen to refer to all forms of oxidized nitrogen,
including multiple gaseous species (e.g., NO2, NO) and particulate species (e.g., nitrates).
The review of the primary NO2 NAAQS focuses on evaluating the health effects
associated with exposure to the gaseous oxides of nitrogen. The atmospheric chemistry,
exposure, and health effects associated with nitrogen compounds present in particulate
matter (PM) were most recently considered in the U.S. EPA's review of the NAAQS for
PM. The welfare effects associated with oxides of nitrogen are being considered in a
separate assessment as part of the review of the secondary NAAQS for NO2 and sulfur
dioxide [SO2; (U.S. EPA. 2013f)1.
NAAQS are defined by four basic elements: indicator, averaging time, level, and form.
The indicator defines the pollutant to be measured in the ambient air for the purpose of
determining compliance with the standard. The averaging time defines the time period
over which air quality measurements are to be obtained and averaged or cumulated,
considering evidence of effects associated with various time periods of exposure. The
level of a standard defines the air quality concentration (i.e., an ambient concentration of
the indicator pollutant) used in determining whether the standard is achieved. The form of
the standard defines the air quality statistic that is compared to the level of the standard in
determining whether an area attains the standard. For example, the form of the current
primary 1-hour NO2 standard is the 3-year average of the 98th percentile of the annual
distribution of 1-hour daily maximum NO2 concentrations. The Administrator considers
these four elements collectively in evaluating the protection to public health provided by
the primary NAAQS.
In 1971, the U.S. EPA added nitrogen oxides to the list of criteria pollutants under
Section 108(a)(l) of the CAA and issued the initial air quality criteria [36 Federal
1 Lists of CASAC members and of members of the CASAC Oxides of Nitrogen Primary NAAQS Review Panel are
available at: httrj://vosemite.epa.gov/sab/sabproduct.nsf/WebCASAC/CommitteesandMembership?OpenDocument.
Ixxiii
-------
Register (FR) 1515, January 30, 1971]. Based on these air quality criteria, the U.S. EPA
promulgated NAAQS for nitrogen oxides using NC>2 as the indicator (36 FR 8186,
April 30, 1971). Both primary and secondary standards were set at 100 ug/m3 [equal to
0.053 parts per million (ppm)], annual average. The standards were based on scientific
information contained in the 1971 Air Quality Criteria Document for Nitrogen Oxides
(U.S. EPA. 1971). Since then, the Agency has completed multiple reviews of the air
quality criteria upon which the primary NCh NAAQS are set and the primary standards
themselves. Table I provides a brief summary of these reviews.
Table I History of the primary National Ambient Air Quality Standards for
nitrogen dioxide since 1971.
Final Rule/
Decisions
Indicator Averaging Time Level Form
1971
36 FR 8186
April 30, 1971
NO2
1 year
53 ppba Annual arithmetic average
1985
50 FR 25532
June 19,1985
Primary NO2 standard retained, without revision.
1996
61 FR 52852
Octobers, 1996
Primary NO2 standard retained, without revision.
2010
75 FR 6474
February 9, 2010
NO2
1 hour
100 ppb 3-year average of the 98th percentile of the
annual distribution of daily maximum 1-hour
concentrations
Primary annual NO2 standard retained, without revision.
FR = Federal Register, NO2 = nitrogen dioxide, ppb = parts per billion.
aThe initial standard level of the annual NO2 standard was 100 |jg/m3 which is equal to 0.053 parts per million or 53 ppb. The units
for the standard level were officially changed to ppb in the final rule issued in 2010 (75 FR 6531, February 9, 2010).
The U.S. EPA retained the primary and secondary NC>2 standards, without revision, in
reviews completed in 1985 and 1996 (50 FR 25532, June 19, 1985; 61 FR 52852,
October 8, 1996). These decisions were informed, respectively, by scientific information
contained in the 1982 Air Quality Criteria Document for Oxides of Nitrogen [(U.S. EPA.
1982) which updated the scientific criteria upon which the initial NO2 standards were
based] and the 1993 Air Quality Criteria Document for Oxides of Nitrogen (U.S. EPA,
Ixxiv
-------
1993a). In the latter of the two decisions, the U.S. EPA concluded that "the existing
annual primary standard appears to be both adequate and necessary to protect human
health against both long- and short-term NCh exposures" and that retaining the existing
annual standard is consistent with the scientific data assessed in the 1993 Air Quality
Criteria Document (U.S. EPA. 1993a) and the Staff Paper (U.S. EPA. 1995a) and with
the advice and recommendations of CASAC" (61 FR 52854, October 8, 1996).!
The last review of the air quality criteria for oxides of nitrogen (health criteria) and the
primary NCh standard was initiated in December 2005 (70 FR 73236,
December 9, 2005).2>3 The Agency's plans for conducting the review were presented in
the Integrated Review Plan (IRP) for the Primary National Ambient Air Quality Standard
for NO2 (U.S. EPA. 2007a). which included consideration of comments received during a
CASAC consultation as well as public comment on a draft IRP. The science assessment
for the review was described in the 2008 Integrated Science Assessment (ISA) for Oxides
of Nitrogen—Health Criteria (U.S. EPA. 2008c). multiple drafts of which received
review by CASAC and the public. The U.S. EPA also conducted quantitative human risk
and exposure assessments, after consultation with CASAC and receiving public comment
on a draft analysis plan (U.S. EPA. 2007b). These technical analyses were presented in
the Risk and Exposure Assessment (REA) to Support the Review of the NO2 Primary
National Ambient Air Quality Standard (U.S. EPA. 2008e). multiple drafts of which
received CASAC and public review.
Over the course of the last review, the U.S. EPA made several changes to the NAAQS
review process. An important change was the discontinuation of the Staff Paper, which
traditionally contained staff evaluations to bridge the gap between the Agency's science
assessments and the judgments required of the U.S. EPA Administrator in determining
whether it was appropriate to retain or revise the NAAQS.4 In the course of reviewing the
second draft REA, however, CASAC expressed the view that the document would be
1 In presenting rationale for the final decision, the U.S. EPA noted that "a 0.053 ppm annual standard would keep
annual NCh concentrations considerably below the long-term levels for which serious chronic effects have been
observed in animals" and that "[retaining the existing standard would also provide protection against short-term
peak NO2 concentrations at the levels associated with mild changes in pulmonary function and airway
responsiveness observed in controlled human [exposure] studies" (61 FR 52854, October 8, 1996; 60 FR 52874,
52880, October 11, 1995).
2 Documents related to reviews completed in 2010 and 1996 are available at:
http ://www. epa. gov/ttn/naaqs/standards/nox/s noxindex, html.
3 The U.S. EPA conducted a separate review of the secondary NO2 NAAQS jointly with a review of the secondary
SO2 NAAQS. The Agency retained those secondary standards, without revision, to address the direct effects on
vegetation of exposure to oxides of nitrogen and sulfur (77 FR 20218, April 3, 2012).
4 Initial changes to the NAAQS review process included a policy assessment document reflecting Agency (rather
than staff) views published as an advanced notice of public rulemaking (ANPR). Under this process, the ANPR
would have been reviewed by CASAC (Peacock. 2006).
Ixxv
-------
incomplete without the addition of a policy assessment chapter presenting an integration
of evidence-based considerations and risk and exposure assessment results. CASAC
stated that such a chapter would be "critical for considering options for the NAAQS for
NCh" s. In addition, within the period of CASAC's review of the second draft REA, the
U.S. EPA's Deputy Administrator indicated in a letter to the CASAC chair, addressing
earlier CASAC comments on the NAAQS review process, that the risk and exposure
assessment will include "a broader discussion of the science and how uncertainties may
effect decisions on the standard" and "all analyses and approaches for considering the
level of the standard under review, including risk assessment and weight of evidence
methodologies" (Peacock. 2008). Accordingly, the final 2008 REA included a policy
assessment chapter that considered the scientific evidence in the 2008 ISA and the
exposure and risk results presented in other chapters of the 2008 REA as they related to
the adequacy of the then-current primary NC>2 standard and potential alternative standards
for consideration (U.S. EPA. 20086).' CASAC discussed the final version of the 2008
REA, with an emphasis on the policy assessment chapter during a public teleconference
on December 5, 2008 (73 FR 66895, November 12, 2008). Following that teleconference,
CASAC offered comments and advice on the primary NO2 standard in a letter to the
Administrator (Samet. 2008).
After considering an integrative synthesis of the body of evidence on human health
effects associated with the presence of NO2 in the air and the exposure and risk
information, the Administrator determined that the then-existing primary NC>2 NAAQS,
based on an annual arithmetic average, was not sufficient to protect public health from
the array of effects that could occur following short-term exposures to ambient NC>2. In
so doing, the Administrator particularly noted the potential for adverse health effects to
occur following exposures to elevated NO2 concentrations that can occur around major
roads (75 FR 6482). In a notice published in the Federal Register on July 15, 2009, the
U.S. EPA proposed to supplement the existing primary annual NC>2 standard by
establishing a new short-term standard (74 FR 34404). In a notice published in the
Federal Register on February 9, 2010, the U.S. EPA finalized a new short-term standard
with a level of 100 ppb, based on the 3-year average of the 98th percentile of the annual
distribution of daily maximum 1-hour concentrations. The U.S. EPA also retained the
existing primary annual NC>2 standard with a level of 53 ppb, annual average (75 FR
6474). The U.S. EPA's final decision included consideration of CASAC (2009) and
1 Subsequent to the completion of the 2008 REA, the U.S. EPA Administrator Jackson called for additional key
changes to the NAAQS review process including reinstating a policy assessment document that contains staff
analysis of the scientific bases for alternative policy options for consideration by senior Agency management prior
to rulemaking (Jackson. 2009). A Policy Assessment will be developed for the current review as discussed in
Chapter 7 of the 2014 Integrated Review Plan for the Primary National Ambient Air Quality Standards for Nitrogen
Dioxide (U.S. EPA. 2014b).
Ixxvi
-------
public comments on the proposed rule. The U.S. EPA's final rule was upheld against
challenges in a decision issued by the U.S. Court of Appeals for the District of Columbia
Circuit on July 17, 2012.1
Revisions to the NAAQS were accompanied by revisions to the data handling
procedures, the ambient air monitoring and reporting requirements, and the Air Quality
Index (AQI).2 One aspect of the new monitoring network requirements included
requirements for states to locate monitors near high-traffic roadways in large urban areas
and in other locations where maximum NC>2 concentrations can occur. Subsequent to the
2010 rulemaking, the U.S. EPA revised the deadlines by which the near-road monitors
are to be operational in order to implement a phased deployment approach (78 FR 16184,
March 14, 2013). The near-road NC>2 monitors will become operational between
January 1, 2014 and January 1, 2017.
1 See American Petroleum Institute v. EPA, 684 F. 3d 1342 (D.C. Cir. 2012).
2 The current federal regulatory measurement methods for NCh are specified in 40 Code of Federal Regulations
(CFR) part 50, Appendix F and 40 CFR part 53. Consideration of ambient air measurements with regard to judging
attainment of the standards is specified in 40 CFR part 50, Appendix S. The NC>2 monitoring network requirements
are specified in 40 CFR part 58, Appendix D, Section 4.3. The U.S. EPA revised the Air Quality Index for NCh to
be consistent with the revised primary NCh NAAQS as specified in 40 CFR part 58, Appendix G. Guidance on the
approach for implementation of the new standards was described in the Federal Register notices for the proposed
and final rules (74 FR 34404; 75 FR 6474).
Ixxvii
-------
EXECUTIVE SUMMARY
Purpose and Scope of the Integrated Science Assessment
This Integrated Science Assessment (ISA) is a thorough evaluation and synthesis of the
policy-relevant science aimed at characterizing exposures to ambient oxides of nitrogen
and relationships with health effects. As such, this ISA serves as the scientific foundation
for the review of the primary (health-based) National Ambient Air Quality Standards
(NAAQS) for nitrogen dioxide (NCh).1 NC>2 is the indicator for gaseous oxides of
nitrogen (i.e., oxidized nitrogen compounds), which also include nitric oxide and gases
produced from reactions involving NCh and nitric oxide (Section 2.2, Figure 2-1).2-3 In
2010, the U.S. Environmental Protection Agency (EPA) retained the NAAQS of 53 parts
per billion (ppb) annual average concentration to protect against health effects potentially
related to long-term NCh exposures. In addition, the U.S. EPA set a new 1-hour NAAQS
at a level of 100 ppb, based on the 3-year average of each year's 98th percentile of the
highest daily 1-hour concentration. The 1-hour NAAQS was set to protect against
respiratory effects related to short-term NC>2 exposures in populations potentially at
increased risk, such as people with asthma or people who spend time on or near
high-traffic roads. The U.S. EPA also set requirements for a network of monitors to
measure NO2 near high-traffic roads, one of the places where the highest concentrations
are expected to occur.
This ISA updates the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) with studies
and reports published from January 2008 through August 2014. The U.S. EPA conducted
searches to identify peer-reviewed literature on relevant topics such as health effects,
ambient concentrations, and exposure. The Clean Air Scientific Advisory Committee (a
formal independent panel of scientific experts) and the public also recommended studies
and reports. To fully describe the state of the science, the U.S. EPA also identified
relevant studies from previous assessments to include in this ISA.
As in the 2008 ISA, this ISA determines the causality of relationships with health effects
only for NCh (Chapter 5 and Chapter 6). Key to interpreting the health effects evidence is
understanding the sources, chemistry, and distribution of NC>2 in the ambient air
(Chapter 2) that influence exposure (Chapter 3). the uptake of inhaled NO2 in the
1 The ecological effects of oxides of nitrogen are being considered in a separate assessment as part of the review of
the secondary (welfare-based) NAAQS for NO2 and sulfur dioxide (U.S. EPA. 2013f).
2 Total oxides of nitrogen also include several paniculate species such as nitrates. Section 108(c) of the Clean Air
Act, 42 U.S.C. § 7408(c) specifies that criteria for oxides of nitrogen include consideration of nitric and nitrous
acids, nitrites, nitrates, nitrosamines, and other derivatives of oxides of nitrogen. Health effects associated with the
paniculate species are addressed in the review of the NAAQS for paniculate matter (U.S. EPA. 2014C).
3 The blue electronic links can be used to navigate to other parts of this ISA and to information on cited references.
Ixxviii
-------
respiratory tract, and subsequent biological mechanisms that may be affected (Chapter 4).
Further, the ISA aims to characterize the independent effect of NO2 exposure on health
rather than its role as just a marker for other air pollutants. The ISA also provides
understanding of policy-relevant issues (Section 1.6). such as (1) exposure durations and
patterns associated with health effects; (2) concentration-response relationship(s),
including evidence of potential thresholds for effects; and (3) populations or lifestages at
increased risk for health effects related to NO2 exposure (Chapter 7).
Sources and Human Exposure to Nitrogen Dioxide
A main objective of the ISA is to characterize health effects related to ambient NO2
exposure. This requires understanding what factors affect exposure to ambient NO2 and
the ability to estimate that exposure well. It also requires accounting for the influence of
factors that are related to NO2 exposure, such as other pollutants and demographic
characteristics. For the U.S. as a whole and for major cities, motor vehicle emissions are
the largest single contributor to NC>2 in the ambient air (Section 2.3.1, Figure 2-3).
Electric power plants, industrial facilities, other forms of transportation, soil, and
wildfires also can contribute considerably to ambient NO2 concentrations on a national
scale and to differences in concentrations and population exposures among locations.
Because many sources of NC>2 are ubiquitous, the potential for exposure to NO2 is
widespread. However, given that motor vehicles are a major source, air concentrations of
NO2 can be highly variable within neighborhoods (Section 2.5.2). depending on distance
to roads. NO2 concentrations tend to decrease over a distance of 200-500 m from the
road (Section 2.5.3). The first year of data from the U.S. near-road monitoring network
show that annual average NO2 concentrations range from 9 to 27 ppb at near-road sites
and 1 to 25 ppb at other sites, but concentrations are higher near roads than at most other
sites within a given urban area (Section 2.5.3.2. Table 2-10). The range in the day's
highest 1-hour NO2 concentration is 35-90 ppb at near-road sites and 12-73 ppb at other
sites, and concentrations are not always higher at the near-road sites. This is because in
addition to distance from road, local sources besides traffic, chemical reactions with
ozone in the air (Figure 2-1). season, wind direction, and physical features of the
environment (Sections 2.2 and 2.5.3) affect the distribution of NC>2 concentrations.
Because ambient NO2 concentrations show variability among geographic regions, within
communities, and overtime, ambient NC>2 exposure can vary considerably among people.
Differences in the outdoor and indoor locations where people spend time and the amount
of time spent in those locations also contribute to variation in ambient NC>2 exposure
(Sections 3.4.1 and 3.4.3. Figure 3-3). NC>2 concentrations vary by the type of location,
including inside vehicles and buildings (Figure 3-1). and the ventilation of buildings can
Ixxix
-------
affect the amount of NCh that penetrates indoors (Section 3.4.3.3). And so, understanding
the extent to which the methods used to estimate exposure adequately account for
variation in ambient concentrations across locations and people's activity patterns is
essential to characterize relationships between ambient NO2 exposure and health effects.
In this ISA, many health effects are examined in relation to ambient NCh concentrations
measured at community monitoring sites. These monitors do not cover all locations
where people live or spend their time and are not sited to capture the variability in NO2
concentrations observed within cities, including near roads. Thus, NCh measurements at
these sites have some error in representing people's actual exposures. This error may be
reflected in the wide range of relationships observed between total personal NO2
exposure and ambient concentrations averaged over periods up to 1 week (Section 3.4.2).
Such relationships are not well characterized for exposure periods of months to years.
Although these uncertainties exist, one cannot necessarily conclude that ambient NO2
concentrations are poor measures of the ambient portion of personal exposure because
variation among people in indoor or in-vehicle exposures and activity patterns may
obscure relationships between ambient concentrations and ambient exposure.
Error in estimating exposure can impact associations observed between ambient NO2
concentrations and health effects. In studies of short-term exposure that examine changes
in NCh overtime (e.g., day to day), NO2 from community monitors has shown lower
magnitude and/or more uncertain associations with health effects (Section 3.4.5)
compared with NCh measured at people's locations. In studies of long-term exposure that
compare people in locations that vary in ambient NCh concentrations, NCh from
community monitors has shown both smaller and larger associations with health effects
compared with NCh concentrations estimated for people's locations. The impact on
health effect associations of using NO2 concentrations at community sites to represent
near-road exposures is not clear. Given the impact of exposure error, this ISA draws
conclusions about health effects related to NC>2 exposure by considering the availability
of results for NCh measured at community monitoring sites versus other locations where
people live or spend time and by considering how well the method of a particular study
represented differences in exposure overtime or across locations. For example, there is
more confidence in the evidence for respiratory effects because many studies examined
exposure metrics that accounted for local variability in NCh concentrations and people's
activity patterns. These metrics included short-term personal, home, and school NCh
measurements and long-term average concentrations estimated at people's homes with
models that well captured the spatial pattern in ambient concentrations in the study areas.
The important contribution of motor vehicles to ambient NCh concentrations not only has
implications for estimating NCh exposure but also indicates the need to consider other
pollutants emitted from vehicles. For example, NCh concentrations often are moderately
Ixxx
-------
to highly correlated with pollutants such as elemental or black carbon, carbon monoxide,
PM2 5, and ultrafine particles : (Section 3.4.4.1. Figure 3-6). These pollutants show effects
on many of the same biological processes and health outcomes as NO2 (Appendix to the
ISA). Thus, in characterizing relationships of NO2 with health effects, this ISA evaluates
the extent to which an effect of NO2 can be separated from that of other traffic-related
pollutants and PM2 5. Experimental studies are key because they can indicate whether
NO2 exposure has a direct effect on health outcomes and biological processes. Though
epidemiologic studies that statistically adjust the NO2 association for another pollutant
cannot conclusively show an independent effect, some provided supporting evidence.
Health Effects of Nitrogen Dioxide Exposure
In this ISA, information on NO2 exposure, the potential influence of other traffic-related
pollutants, and health effects from epidemiologic, controlled human exposure, and
toxicological studies is integrated to form conclusions about the causal nature of
relationships between NO2 exposure and health effects. Health effects examined in
relation to the full range of NO2 concentrations relevant to ambient conditions are
considered. Based on peak concentrations (Section 2.5) and the ISA definition that
ambient-relevant exposures be within one to two orders of magnitude of current
conditions (Preamble. Section 5.c). NO2 concentrations up to 5,000 ppb2 are defined to be
ambient relevant. A consistent and transparent framework (Preamble. Table II) is applied
to classify the health effects evidence according to a five-level hierarchy:
1) Causal relationship
2) Likely to be a causal relationship
3) Suggestive of, but not sufficient to infer, a causal relationship
4) Inadequate to infer a causal relationship
5) Not likely to be a causal relationship
The conclusions presented in Table ES-1 are informed by recent findings and whether
recent findings integrated with information from the 2008 ISA for Oxides of Nitrogen
(U.S. EPA. 2008c) support a change in conclusion. Important considerations include
judgments of error and uncertainty in the collective body of available studies; the
consistency of findings integrated across epidemiologic, controlled human exposure, and
toxicological studies to inform understanding about an independent effect of NO2
1 PM25: In general terms, paniculate matter with a nominal mean aerodynamic diameter less than or equal to
2.5 nm, a measure of fine particles. UFP: Definitions vary but often refer to particles with a nominal mean
aerodynamic diameter less or equal to 0. 1 ^m.
2 The 5,000-ppb upper limit applies mostly to animal toxicological studies and also a few controlled human
exposure studies. Experimental studies examining NO2 exposures greater than 5,000 ppb were included if they
provided information on the uptake of NO2 in the respiratory tract or on potential biological mechanisms.
Ixxxi
-------
exposure and potential biological pathways; the extent to which epidemiologic studies
adequately represented NO2 exposure; and examination in epidemiologic studies of the
potential influence of other traffic-related pollutants and other factors that could bias
associations observed with NO2 exposure (described in the Appendix to the ISA).
Table ES-1 Causal determinations for relationships between nitrogen dioxide
exposure and health effects from the 2008 and 2016 Integrated
Science Assessment for Oxides of Nitrogen.
Causal Determination13
Exposure Duration and
Health Effects Category3 2008 Integrated Science Assessment 2016 Integrated Science Assessment
Short-Term Nitrogen Dioxide Exposure (minutes up to 1 month)
Respiratory effects
Section 5.2. Table 5-39
Sufficient to infers likely causal
relationship
Causal relationship
Cardiovascular effects
Section 5.3, Table 5-52
Inadequate to infer the presence or
absence of a causal relationship
Suggestive of, but not sufficient to infer, a
causal relationship
Total mortality
Section 5.4. Table 5-57
Suggestive of, but not sufficient to infer,
a causal relationship
Suggestive of, but not sufficient to infer, a
causal relationship
Long-Term Nitrogen Dioxide Exposure (more than 1 month to years)
Respiratory effects
Section 6.2. Table 6-5
Suggestive of, but not sufficient to infer,
a causal relationship
Likely to be a causal relationship
Cardiovascular effects and
diabetes0
Section 6.3, Table 6-11
Inadequate to infer the presence or
absence of a causal relationship
Suggestive of, but not sufficient to infer, a
causal relationship
Reproductive and
developmental effects0
Sections 6.4.2, 6.4.3, and
6.4.4. Table 6-14
Inadequate to infer the presence or
absence of a causal relationship
Fertility, reproduction, and pregnancy:
Inadequate to infer a causal relationship
Birth outcomes:
Suggestive of, but not sufficient to infer, a
causal relationship
Postnatal development:
Inadequate to infer a causal relationship
Total mortality
Section 6.5, Table 6-18
Inadequate to infer the presence or
absence of a causal relationship
Suggestive of, but not sufficient to infer, a
causal relationship
Cancer
Section 6.6. Table 6-20
Inadequate to infer the presence or
absence of a causal relationship
Suggestive of, but not sufficient to infer, a
causal relationship
aAn array of outcomes is evaluated as part of a broad health effects category: physiological measures (e.g., airway
responsiveness), clinical outcomes (e.g., hospital admissions), cause-specific mortality. Total mortality includes all nonaccidental
causes of mortality, and conclusions are informed by findings for the spectrum of morbidity effects (e.g., respiratory,
cardiovascular) that can lead to mortality. The sections and tables referenced include a detailed discussion of the evidence that
supports the causal determinations and the NO2 concentrations with which health effects have been associated.
bSince the 2008 ISA for Oxides of Nitrogen, the phrasing of causal determinations has changed slightly, and the weight of
evidence that describes each level in the hierarchy of the causal framework has been more explicitly characterized.
°ln this ISA, the conclusion is based on cardiovascular effects and diabetes, which are related and share risk factors. Reproductive
and developmental effects are separated into smaller subcategories of outcomes based on varied underlying biological processes
and exposure patterns over different lifestages.
Ixxxii
-------
Short-Term Nitrogen Dioxide Exposure and Respiratory Effects
A causal relationship is determined for short-term NC>2 exposure and respiratory effects.
The conclusion is strengthened from the 2008 ISA for Oxides of Nitrogen from likely to
be a causal relationship (Table ES-1) based on the evidence indicating that NO2 exposure
can trigger asthma attacks. There is some evidence relating short-term NCh exposure to
chronic obstructive pulmonary disease, respiratory infection, respiratory effects in
healthy populations, and respiratory mortality but uncertainty as to whether the effects of
NC>2 exposure are independent of other traffic-related pollutants (Table 5-39).
The key evidence that short-term NO2 exposure independently can trigger an asthma
attack is the increased airway responsiveness and allergic inflammation induced by NO2
exposure in controlled human exposures studies. Although reactions with antioxidants
typically are beneficial, such reactions for inhaled NO2 can form reactive species in the
fluid lining the lung (Section 4.2.2). These reactive species can enhance allergic
inflammation and airway responsiveness (Figure ES-1). so this evidence further links
NO2 exposure to asthma attacks. Allergic inflammation and airway responsiveness are
hallmarks of asthma attacks; thus, this evidence supports epidemiologic results, which
consistently link short-term increases in ambient NC>2 concentration with increases in
hospital admissions and emergency department visits for asthma, increases in respiratory
symptoms and airway inflammation in people with asthma, and decreases in lung
function in children with asthma (Section 5.2.9). These associations exist not only with
community-average ambient NC>2 concentrations but also with personal NC>2 and NO2
measured outside children's schools and inside their homes (Sections 5.2.9.3 and 5.2.9.6).
Because outdoor and indoor sources (e.g., vehicles, gas stoves) emit a different mix of
pollutants, NC>2 is more weakly related to other traffic-related pollutants for total personal
exposures than for ambient concentrations. The same may be true for indoor exposures.
So, associations with personal and indoor NO2 may be less influenced by pollutants that
are related to outdoor NO2. Further, studies that measured pollutants at people's locations
tend to show that NO2 remains associated with asthma-related effects after accounting for
PM25 or, as examined in fewer studies, a traffic-related pollutant such as elemental or
black carbon, metals, or ultrafine particles (Figures 5-16 and 5-17).
The 2008 ISA described much of the same evidence and determined a likely to be causal
relationship, citing uncertainty as to whether the epidemiologic results for NC>2 reflected
the effects of other traffic-related pollutants. The 2008 ISA did not explicitly evaluate the
extent to which various lines of evidence supported effects on asthma attacks. In this ISA,
the determination of a causal relationship is not based on new evidence as much as it is
on the integrated findings for asthma attacks with due weight given to experimental
studies. The epidemiologic evidence for asthma attacks and controlled human exposure
study findings for increased airway responsiveness and allergic inflammation together
Ixxxiii
-------
demonstrate that short-term NC>2 exposure has an independent relationship with
respiratory effects and is not just an indicator for other traffic-related pollutants.
Activation of neural
pathways
t Release of
.» immune mediators
frommastcells
t Permeability of
airways
Exposure to
an asthma
Short-term
Exposure
Reactions in
respiratory tract
lining fluid and tissue
Formation of
oxidation/nitration
products
t Allergic
responses
t Airway
oxidative stress
; t Airway trlgg% Asthma
responsiveness attack
Long-term
Exposure
Repeated/
Persistent
Persistent reactions in
respiratory tract
lining fluid and tissue
Persistent formation of
oxidation/nitration
products
Development of
* allergic responses
t Airway
responsiveness
Note: NO2 = nitrogen dioxide. Adapted from Figures 4-1 and 4-2 (Section 4.3.5). White boxes and solid arrows describe pathways
well supported by available evidence. Gray boxes and dotted arrows describe potential pathways for which evidence is limited or
inconsistent.
Figure ES-1 Evidence for relationships of short-term and long-term nitrogen
dioxide exposure with asthma presented as biological pathways.
Long-Term Nitrogen Dioxide Exposure and Respiratory Effects
There is likely to be a causal relationship between long-term NO2 exposure and
respiratory effects based on the evidence for development of asthma (Section 6.2.9. Table
6-5). The conclusion is strengthened from the 2008 ISA (Table ES-1) because where
previous epidemiologic findings were inconsistent, recent studies consistently observe
NO2-related increases in asthma development in children who are followed over time and
are supported by previous experimental studies. A key strength is that asthma
development is linked to ambient NCh concentrations measured near children's homes or
schools or estimated at homes with models that well predicted the concentration pattern
Ixxxiv
-------
within the community. Associations between NC>2 and asthma development are
independent of factors such as socioeconomic status and exposure to smoking, but the
influence of other traffic-related pollutants is not well studied. There is some support for
an independent effect of long-term NO2 exposure on asthma development provided by
findings of increased airway responsiveness in rodents (Figure ES-1). Also, evidence
relating short-term NCh exposure to airway inflammation in epidemiologic studies of
healthy people and allergic responses in experimental studies of rodents and healthy
people indicates that repeated short-term NO2 exposure could lead to the development of
asthma. Together, the epidemiologic and experimental evidence for asthma development
supports a relationship between long-term NCh exposure and respiratory effects, but
because experimental evidence is limited, there remains some uncertainty about the
potential influence of other traffic-related pollutants in the epidemiologic evidence.
Nitrogen Dioxide Exposure and Other Health Effects
There is more uncertainty about relationships of NC>2 exposure with health effects outside
of the respiratory system. NC>2 itself is unlikely to enter the bloodstream, and reactions
caused by ambient-relevant concentrations of NO2 in the airways do not clearly affect
concentrations of reaction products, such as nitrite, in the blood. Some but not all results
suggest that substances that can cause inflammation or oxidative stress may enter the
OO J
blood from the respiratory tract in response to NC>2 exposure (Section 4.3.2.9). This
uncertainty about the effects of NO2 exposure on underlying biological mechanisms is
common to nonrespiratory health effects.
For short-term and/or long-term NCh exposure, evidence is suggestive of, but not
sufficient to infer, a causal relationship with cardiovascular effects and diabetes, total
mortality, birth outcomes, and cancer (Table ES-1). For short-term NCh exposure, recent
epidemiologic studies continue to show associations with total mortality and add support
for cardiovascular effects by indicating a possible effect on triggering heart attacks.
Where there was little previous support, increases in recent epidemiologic evidence led to
strengthening conclusions for total mortality and cancer related to long-term NC>2
exposure. New epidemiologic findings for heart disease and diabetes and reduced fetal
growth point to possible relationships of long-term NCh exposure with health effects
categories new to this ISA. For fertility, reproduction, and pregnancy, as well as postnatal
development, evidence is inadequate to infer a causal relationship with long-term NO2
exposure (Table ES-1) because neither epidemiologic nor toxicological studies clearly
show effects. For all nonrespiratory effects, epidemiologic studies do not adequately
account for the potential influence of other traffic-related pollutants, which combined
with the few or inconclusive results from controlled human exposure or toxicological
Ixxxv
-------
studies, produces large uncertainty as to whether short-term or long-term NC>2 exposure
has independent relationships with health effects outside of the respiratory system.
Policy-Relevant Considerations for Health Effects Associated
with Nitrogen Dioxide Exposure
Multiple durations of short-term and long-term NC>2 exposure are observed to be
associated with health effects (Section 1.6.1). For short-term exposure, asthma-related
effects are associated with total personal NO2 exposure and NO2 measured at children's
schools or community monitors averaged over 1 to 5 days. These associations are
observed with both daily average and the daily highest 1-hour NCh concentration. No
particular duration of exposure shows a stronger effect. Controlled human studies
demonstrate increased airway inflammation and airway responsiveness in adults with
asthma following NC>2 exposures of 15 to 60 minutes. These results support the
epidemiologic evidence showing that NO2 exposures of 2 or 5 hours near high-traffic
roads are associated with similar respiratory effects in adults with and without asthma.
For long-term exposure, asthma development in children is associated with NC>2 exposure
estimates for homes or schools that are averaged over 1 to 10 years, representing various
time periods, such as infancy, childhood, and lifetime. It is not clear what pattern of NO2
exposure in time may underlie these associations, but some evidence from experimental
studies in humans and rodents suggests that repeated exposure over many days or weeks
can induce allergic responses that are involved in asthma development.
Information on the shape of the NC>2 concentration-health effects relationship is provided
mostly by epidemiologic studies. Based on the few results that are available, asthma
emergency department visits increase with increasing short-term average ambient NC>2
concentrations (Section 1.6.3). In Atlanta, GA, an association is present with daily
highest 1-hour NC>2 concentrations from 37 to 11 ppb but is uncertain at lower
concentrations. The lower bound of NC>2 concentrations where an association is present
also is uncertain because concentrations were averaged across sites in Atlanta, GA, which
may not reflect the range of concentrations in the city or range of exposures among
individuals.
Health effects related to NO2 exposure potentially have a large public health impact.
Many people in the U.S. live, work, or spend time near roads and may have higher
exposures to NC>2. Higher NCh exposure also is suggested for urban, low socioeconomic
status, and nonwhite populations. Further, people with asthma, children (especially ages
0-14 years), and older adults (especially ages 65 years and older) are identified as being
at increased risk of NCh-related health effects (Chapter 7). Evidence does not clearly
identify other at-risk populations in terms of other diseases or behavioral, genetic, or
Ixxxvi
-------
sociodemographic factors. Short-term and long-term NC>2 exposure is linked to clinically
relevant increases in airway responsiveness, emergency department visits and hospital
admissions for asthma, and development of asthma, which can have a large impact on
public health. Given that asthma is the leading chronic illness and the leading cause of
missed school days and hospital admissions among U.S. children, NCh-related asthma
attacks and asthma development have the potential to affect children's overall well-being.
Summary of Major Findings
Expanding on findings from the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c).
recent epidemiologic studies show associations of short-term and long-term NC>2
exposure with an array of health effects. However, except for respiratory effects, there
remains large uncertainty about whether NC>2 exposure has an effect that is independent
of other traffic-related pollutants. As in the 2008 ISA, recent information shows that
motor vehicle emissions are the largest single source of NO2 in the air and that NO2
concentrations tend to be variable within communities, decreasing with increasing
distance from roads. Information to assess whether NO2 exposure estimates adequately
represent the variability in ambient NC>2 concentrations and people's activity patterns
varies among the health effects evaluated in this ISA. The major findings from this ISA
about NC>2 exposure and health effects and related uncertainties are summarized below.
• Evidence for asthma attacks supports a causal relationship between short-term
NC>2 exposure and respiratory effects. Evidence for development of asthma
supports a likely to be causal relationship between long-term NO2 exposure and
respiratory effects. These are stronger conclusions than those determined in the
2008 ISA for Oxides of Nitrogen.
• There is more uncertainty as to whether short-term or long-term NO2 exposure is
related to cardiovascular effects, diabetes, reproductive and developmental
effects, total mortality, and cancer.
• People with asthma, children, and older adults are at increased risk for
NO2-related health effects.
• People living or spending time near or on roads, low socioeconomic status
populations, and nonwhite populations may have increased NO2 exposure.
• The first year of data from the U.S. near-road monitoring network indicate that
near-road sites tend to have higher NO2 concentrations on average but do not
always have the highest 1-hour NO2 concentration within an urban area.
• Epidemiologic studies link asthma attacks and asthma development to NO2
measures that appeared to well represent exposure, including personal measures
and concentrations where participants live or spend a lot of time.
• No specific NO2 averaging time, duration, or age of exposure is more strongly
associated with asthma attacks or asthma development. It is not clear whether
there is an exposure concentration below which effects do not occur.
Ixxxvii
-------
CHAPTER 1 INTEGRATED SUMMARY
1.1 Purpose and Overview of the Integrated Science Assessment
The Integrated Science Assessment (ISA) is a comprehensive evaluation and synthesis of
the policy-relevant science "useful in indicating the kind and extent of identifiable effects
on public health or welfare which may be expected from the presence of [a] pollutant in
ambient air" (CAA. 1990a). This ISA communicates critical science judgments of the
health criteria for a broad category of gaseous oxides of nitrogen (i.e., oxidized nitrogen
compounds) for which nitrogen dioxide (NO2) is the indicator. As such, this ISA serves
as the scientific foundation for the review of the current primary (health-based) National
Ambient Air Quality Standards (NAAQS) forNO2. Gaseous oxides of nitrogen include
NO2, nitric oxide (NO), and their various reaction products (Section 2.2. Figure 1-1).'
There also are particulate oxides of nitrogen (e.g., nitrates, nitro-polycyclic aromatic
hydrocarbons),2 which are being considered in the review of the NAAQS for particulate
matter (PM) (U.S. EPA. 2014c). The welfare effects of oxides of nitrogen are being
evaluated separately as part of the review of the secondary (welfare-based) NAAQS for
NO2 and sulfur dioxide [SO2; (U.S. EPA. 2013f)1.
This ISA evaluates relevant scientific literature published since the 2008 ISA for Oxides
of Nitrogen (U.S. EPA. 2008c). integrating key information and judgments contained in
the 2008 ISA and the 1993 Air Quality Criteria Document for Oxides of Nitrogen (U.S.
EPA. 1993a). Thus, this ISA updates the state of the science that was available for the
2008 ISA, which informed decisions on the primary NO2 NAAQS in the review
completed in 2010. In 2010, the U.S. Environmental Protection Agency (EPA) retained
the existing annual average (avg) NO2 NAAQS with a level of 53 parts per billion (ppb)
to protect against health effects potentially associated with long-term exposure. The
U.S. EPA established a new 1-hour (h) NAAQS at a level of 100 ppb NO2 based on the
3-year (yr) avg of each year's 98th percentile of 1-h daily maximum (max)
concentrations.3 The 1-h standard was established to protect against a broad range of
respiratory effects associated with short-term exposures in potential at-risk populations
such as people with asthma and people who spend time on or near high-traffic roads. In
2010, the U.S. EPA also set requirements for a monitoring network in urban areas that
1 The blue electronic links can be used to navigate to other parts of this ISA and to information on cited references.
2 Section 108(c) of the Clean Air Act, 42 U.S.C. § 7408(c) specifies that criteria for oxides of nitrogen include
consideration of nitric and nitrous acids, nitrites, nitrates, nitrosamines, and other derivatives of oxides of nitrogen,
including multiple gaseous and particulate species.
3 The legislative requirements and history of the NCh NAAQS are described in detail in the Preface to this ISA.
1-1
-------
include placing monitors near [within 50 meters (m)] high-traffic roads, one of the
locations where the highest NO2 concentrations are expected to occur (U.S. EPA. 2010c).
This review of the primary NC>2 NAAQS is guided by several policy-relevant questions
that have been identified in The Integrated Review Plan for the Primary National
Ambient Air Quality Standard for Nitrogen Dioxide (U.S. EPA. 2014b). To address these
questions and update the scientific judgments in the 2008 ISA, this ISA aims to:
• Characterize the evidence for health effects associated with short-term (minutes
up to 1 month) and long-term (more than 1 month to years) exposure to oxides of
nitrogen by integrating findings across scientific disciplines and across related
health outcomes and by considering important uncertainties identified in the
interpretation of the scientific evidence, including the role of NC>2 within the
broader ambient mixture of pollutants.
• Inform understanding of policy-relevant issues related to quantifying health risks,
such as exposure concentrations, durations, and patterns associated with health
effects; concentration-response relationships and evidence of thresholds below
which effects do not occur; and populations and lifestages with increased risk of
health effects related to NO2 exposure.
Although the scope of the ISA includes all gaseous oxides of nitrogen, much of the
information on the distribution of oxides of nitrogen in the air, human exposure and dose,
impact of errors associated with exposure assessment methods, and health effects is for
NC>2. There is limited information for NO and the sum of NO and NO2 (NOx) as well as
large uncertainty in relating health effects to NO or NOx exposure. In the body, NO is
produced from nitrates and nitrites derived from diet and through enzymatic pathways
that are enhanced during inflammation. Ambient NO concentrations generally are in the
range of endogenous NO concentrations exhaled from the respiratory tract. It is not clear
whether ambient-relevant NO exposures substantially alter endogenous NO production in
the respiratory tract or pathways affected by endogenous NO (Section 4.2.3). Thus, the
potential for detrimental health effects occurring from ambient-relevant NO exposure is
unclear. This lack of evidence leaves NO2 as the component of NOx to consider in
evaluating health effects in relation to NOx exposure. Because the ratio of NO2 to NOx
varies across locations, time of day, and season (Section 2.5). NOx may not represent
NO2 exposure consistently. The lack of evidence that ambient-relevant NO exposure can
lead to detrimental health effects and the measurement error related to using NOx to
represent NO2 exposure are the rationale for determining the causal nature of health
effects only for NO2 exposure.
In addressing policy-relevant questions, this ISA aims to characterize the independent
health effects of NO2 exposure, not the role of NO2 as just a marker for a broader mixture
of pollutants in the ambient air. The potential influence of other traffic-related pollutants
was the main uncertainty in the conclusions drawn in the 2008 ISA Oxides of Nitrogen
1-2
-------
(U.S. EPA. 2008c). As discussed in this ISA, evidence combined from controlled human
exposure and epidemiologic studies sufficiently describes a coherent, biologically
plausible relationship between short-term NO2 exposure and respiratory effects indicative
of asthma exacerbation. These effects include increased airway responsiveness as well as
increased symptoms, emergency department (ED) visits, and hospital admissions. New
epidemiologic evidence supports a relationship of long-term NO2 exposure with
respiratory effects, specifically, the development of asthma in children, and a small body
of previous experimental studies provide some indication thatNCh exposure may have an
independent effect. Recent epidemiologic studies continue to suggest that short-term NCh
exposure may be associated with cardiovascular effects and mortality, and new findings
potentially link long-term NCh exposure to cardiovascular effects, diabetes, poorer birth
outcomes, mortality, and cancer. However, for nonrespiratory effects, epidemiologic
studies have not adequately accounted for effects of other traffic-related pollutants, and
findings from experimental studies continue to be limited. The information in the ISA
forming the basis for these judgments will serve as the scientific foundation for the
review of the current primary 1-hour and annual NCh NAAQS.
1.2 Process for Developing Integrated Science Assessments
The U.S. EPA uses a structured and transparent process for evaluating scientific
information and determining the causality of relationships between air pollution
exposures and health effects (Preamble). This process includes approaches for literature
searches, guidelines for selecting and evaluating relevant studies, and a framework for
evaluating the weight of evidence and determining causality. As part of this process, the
ISA is reviewed by the Clean Air Scientific Advisory Committee (CASAC), a formal
independent panel of scientific experts, and the public. As this ISA informs the review of
the primary NC>2 NAAQS, it assesses information relevant to characterizing exposure to
gaseous oxides of nitrogen and potential effects on health. Studies on atmospheric
chemistry, spatial and temporal trends, and exposure assessment are relevant, as are
analyses by the U.S. EPA of air quality and emissions data. Also relevant are
epidemiologic, controlled human exposure, and toxicological studies on health effects, as
well as studies on dosimetry and modes of action.
The U.S. EPA initiated the current review of the primary NAAQS for NC>2 in February
2012 with a call for information from the public (U.S. EPA. 2012c). Thereafter, the
U.S. EPA routinely conducted literature searches to identify relevant peer-reviewed
studies published since the previous ISA (i.e., from January 2008 through August 2014).
Multiple search methods were used (Preamble. Section 2) including searches in databases
such as PubMed and Web of Science. Also, CASAC and the public recommended
1-3
-------
studies. The U.S. EPA identified additional studies considered to be the definitive work
on particular topics from previous assessments to include in this ISA. Some studies were
judged to be irrelevant (i.e., did not address a topic described in the preceding paragraph)
based on title and were excluded. Studies judged to be potentially relevant based on
review of the abstract or full text and considered for inclusion in the ISA are documented
and can be found at the Health and Environmental Research Online (HERO) website. The
HERO project page for this ISA (http://hero.epa.gov/oxides-of-nitrogen) contains the
references that are cited in the ISA, the references that were considered for inclusion but
not cited, and electronic links to bibliographic information and abstracts.
Health effects were considered for evaluation in this ISA if they were examined in
previous assessments by the U.S. EPA for oxides of nitrogen or multiple recent studies
(e.g., neurodevelopment). Literature searches identified one or two recent epidemiologic
studies each on outcomes such as gastrointestinal effects, bone density, headache, and
depression [Supplemental Table Sl-1; (U.S. EPA, 2015f)1. A review of these studies
indicated they are similar in design and conducted in areas and populations for which
associations between ambient NO2 concentrations and other health effects have been
documented. These few studies were excluded from this ISA because they do not provide
new information on particular geographic locations, potential at-risk populations or
lifestages, or range of ambient NO2 concentrations. These studies also are more likely to
be subject to publication bias.
The Preamble describes the general framework for evaluating scientific information,
including criteria for assessing the strength of inference of a study and developing
scientific conclusions. Aspects specific to evaluating studies of NO2 are described in the
Appendix. For epidemiologic studies, emphasis is placed on studies that characterize
quantitative relationships between NO2 and health effects, examine exposure metrics that
well represent the variability in concentrations in the study area, consider the potential
influence of other air pollutants and factors correlated with NO2, examine potential at-risk
populations and lifestages, or combine information across multiple cities. With respect to
the evaluation of controlled human exposure and toxicological studies, emphasis is
placed on studies that examine effects relevant to humans and NO2 concentrations that
are defined in this ISA to be relevant to ambient exposures. Based on peak ambient
concentrations (Section 2.5) and the ISA definition that ambient-relevant exposures be
within one to two orders of magnitude of current levels, NO2 concentrations of
5,000 ppb1 or less are defined to be ambient relevant. Experimental studies with higher
exposure concentrations were considered if they examined dosimetry or potential modes
of action. For the evaluation of human exposure to ambient NO2, emphasis is placed on
1 The 5,000-ppb upper limit applies largely to animal toxicological studies but also a few controlled human exposure
studies.
1-4
-------
studies that examine the adequacy of methods used to assess exposures, such as central
site monitors, land use regression (LUR) models, and personal exposure monitors. The
ISA also emphasizes studies that examine factors that influence exposure, such as
time-activity patterns and building ventilation characteristics.
Integrating information across scientific disciplines and related health outcomes and
synthesizing evidence from previous and recent studies, the ISA draws conclusions about
relationships between NO2 exposure and health effects. Determinations are made about
causation not just association and are based on judgments of aspects such as the
consistency, coherence, and biological plausibility of observed effects (i.e., evidence for
effects that can be linked in a mode of action) as well as related uncertainties. As such,
determinations of causation are made not on evidence for individual disciplines or
individual outcomes but on the integrated body of evidence. The ISA uses a formal
causal framework (Table II of the Preamble) to classify the weight of evidence according
to the five-level hierarchy summarized below.
• Causal relationship: the consistency and coherence of evidence integrated
across scientific disciplines and related health outcomes are sufficient to rule out
chance, confounding, and other biases with reasonable confidence.
• Likely to be a causal relationship: there are studies where results are not
explained by chance, confounding, or other biases, but uncertainties remain in the
evidence overall. For example, the influence of other pollutants is difficult to
address, or evidence among scientific disciplines may be limited or inconsistent.
• Suggestive of, but not sufficient to infer, a causal relationship: evidence is
generally supportive but not entirely consistent or overall is limited. Chance,
confounding, and other biases cannot be ruled out.
• Inadequate to infer a causal relationship: there is insufficient quantity, quality,
consistency, or statistical power of results from studies.
• Not likely to be a causal relationship: several adequate studies, examining the
full range of human exposure concentrations and potential at-risk populations and
lifestages, consistently show no effect.
1.3 Content of the Integrated Science Assessment
The ISA consists of the Preamble. Preface (legislative requirements and history of the
primary NCh NAAQS), Executive Summary, and seven chapters. Chapter 1 synthesizes
the scientific evidence that best informs policy-relevant questions that frame this review
of the primary NCh NAAQS. Chapter 2 characterizes the sources, atmospheric processes
of oxides of nitrogen, and trends in ambient concentrations. Chapter 3 describes methods
to estimate human exposure to oxides of nitrogen and the impact of error in exposure
estimates on associations with health effects. Chapter 4 describes the dosimetry and
1-5
-------
potential modes of action for NCh and NO. Chapter 5 and Chapter 6 evaluate and
integrate epidemiologic, controlled human exposure, and toxicological evidence for
health effects related to short-term and long-term exposure to oxides of nitrogen,
respectively. Chapter 7 evaluates information on potential at-risk populations and
life stages.
The purpose of this chapter is not to summarize each of the aforementioned chapters but
to synthesize the key findings on each topic that are considered in characterizing NC>2
exposure and relationships with health effects. This chapter also integrates information
across the ISA to address policy-relevant issues such as NO2 exposure durations and
patterns associated with health effects, concentration-response relationships, and the
public health impact of NO2-related health effects (Section 1.6). A key consideration in
the health effects assessment is the extent to which evidence indicates that NO2 exposure
independently causes health effects versus indicating that NO2 may be serving just as a
marker for a broader mixture of air pollutants, especially those related to traffic. To that
end, this chapter draws upon information about the sources, distribution, and exposure to
ambient NCh and identifies pollutants and other factors correlated with the distribution of
or exposure to ambient NO2 that can potentially influence epidemiologic associations
observed between health effects and NO2 exposure (Section 1.4.3). The discussions of the
health effects evidence and causal determinations (Section 1.5) describe the extent to
which epidemiologic studies accounted for these factors and the extent to which findings
from controlled human exposure and animal toxicological studies support independent
relationships between NO2 exposure and health effects.
1.4 From Emissions Sources to Exposure to Nitrogen Dioxide
Characterizing human exposure is key to understanding the relationships between
ambient NO2 exposure and health effects. The sources of oxides of nitrogen and the
transformations that occur in ambient air influence the spatial and temporal pattern of
NC>2 concentrations in the air. These patterns have implications for variation in exposure
in the population, the adequacy of methods used to estimate exposure, and in turn, the
strength of inferences that can be drawn from associations observed in epidemiologic
studies between NO2 exposure and health effects.
1.4.1 Emission Sources and Distribution of Ambient Concentrations
The strength and distribution of emissions sources are important determinants of the
distribution of NC>2 in the ambient air, and in turn, human exposure. Information on
1-6
-------
emissions is available for NOx, which is emitted primarily as NO. NO rapidly reacts with
radicals and ozone (Os) to form NO2 in the air. Based on the 2011 National Emissions
Inventory, the largest single source of NOx emissions in the U.S. overall and in major
population centers (city and surrounding communities) is highway vehicles (40-67%;
Section 2.3, Table 2-1). Sources such as electric utilities, commercial and residential
boilers, and industrial facilities are more variable across locations but can be important
contributors to ambient NO2 concentrations for the U.S. as a whole and in certain
populated areas. Some of these smaller sources can affect local air quality with large,
transient emissions of NOx. Natural sources such as microbial processes in soil and
wildfires contribute 2% of emissions in U.S. population centers, and emissions from
natural and anthropogenic sources from continents other than North America (i.e., North
American Background) account for less than 1% (typically 0.3 ppb) of ambient
concentrations (Section 2.5.6). Although highway vehicles are a large, ubiquitous source
of NOx, the varying presence and mix of specific emissions sources across locations can
contribute to heterogeneity in ambient NO2 concentrations regionally and locally, which
has implications for variation in exposure to ambient NO2 within the population.
In addition to emissions sources, factors that influence NO2 ambient concentrations
include chemical transformations, transport to other locations, meteorology, and
deposition to surfaces (Figure 1-1 and in more detail, Figure 2-1). NO and NO2 react with
gas phase radicals and Os to form other oxides of nitrogen such as peroxyacetyl nitrate
(PAN) and nitric acid (HNOs; Section 2.2). NO and NO2 also are involved in reaction
cycles with radicals produced from volatile organic compounds (VOCs) to form Os. The
reactions of NO and NO2 into other oxides of nitrogen typically occur more slowly than
the interconversion between NO2 and NO does, and NO and NO2 are the most prevalent
oxides of nitrogen in populated areas. HNOs and PAN can make up a large fraction of
ambient oxides of nitrogen downwind of major emission sources.
Sources, atmospheric transformations, and meteorology contribute to the temporal trends
observed in ambient NO2 concentrations. As a result of pollution control technologies on
vehicles and electric utilities (Section 2.3.2). NOx emissions from highway vehicles and
fuel combustion decreased by 49% in the U.S. from 1990 to 2013 (Figure 2-2). During
that time (1990-2012), U.S.-wide annual average NO2 concentrations decreased by 48%
(Figure 2-22). In addition to long-term trends, ambient NO2 concentrations show seasonal
trends, with higher concentrations measured in the winter than summer. Reflecting trends
in traffic, ambient concentrations at most urban sites are higher on weekdays than
weekends, and within a day, concentrations peak in early mornings, decrease until late
afternoon, then increase again in early evening corresponding with morning and evening
commutes. Diurnal trends in ambient NO2 also are affected by meteorology, with
1-7
-------
concentrations rising during the night when atmospheric mixing is reduced because of
low wind speeds and low mixing layer heights.
vl
Long range transport to remote
regions at low temperatures
r~ *"
INORGANIC
PRODUCTS
(e.g., nitric
acid)
L
ORGANIC
PRODUCTS
(e.g., peroxyacetyl
nitrate)
deposition
emissions
Note: The inner shaded box depicts NOX [sum of nitric oxide (NO) and nitrogen dioxide (NO2)]. The outer box contains oxides of
nitrogen formed from reactions of NOX (NOZ). Oxides of nitrogen in the outer and inner boxes (NOX + NOZ) are collectively referred
to as NOY by the atmospheric sciences community.
Source: National Center for Environmental Assessment. For more details on the various reactions, see Figure 2-1.
Figure 1-1
Reactions of oxides of nitrogen species in the ambient air.
The spatial variation in emissions sources and chemical transformation of oxides of
nitrogen likely contribute to the variability in ambient NC>2 concentrations observed at
regional, urban, neighborhood, and near-road scales (Section 2.5). Measurements from
U.S. air monitoring networks1 of several hundred sites (Section 2.5.1) show wide
variation in ambient NC>2 concentrations across the U.S. Across central site monitors, the
mean 1-h daily maximum ambient NC>2 concentration for 2011-2013 was 19 ppb, and the
1 The air monitoring networks serve many objectives: determining compliance with the NAAQS, providing the
public with air pollution data in a timely manner, and providing estimates of ambient exposure for research studies.
1-8
-------
5th to 99th percentile range was 2-55 ppb (Table 2-3). The mean annual average NO2
concentration was 8.6 ppb, and the 5th to 99th percentile range was 1.4-22.5 ppb
(Table 2-4). Ambient NO2 concentrations are higher in large cities than in less populated
areas (Figures 2-11 and 2-12). Ambient NO2 concentrations also can vary widely across
sites within cities where vehicle emissions are the major source (Figure 2-14. Table 2-5).
Some sites agree well with each other in terms of temporal correlations or magnitude of
concentration. However, the siting of most monitors away from sources likely means that
the monitors do not capture the extent of variability in ambient NCh in a city. Preliminary
data from the first year of the near-road network for 41 U.S. cities show that near-road
(within 50 m) sites have higher mean NC>2 concentrations than many other sites within an
urban area but not always the highest 1-hour concentrations (Table 2-10). Across
near-road sites, means for 1-h daily maximum NC>2 concentrations were 9-27 ppb, and
98th percentiles were 35-74 ppb. For durations of 1 hour or less, studies measured NC>2
concentrations of 5.8 to 120 ppb within 20 m of a road, which are up to 100% higher than
concentrations 80 to 400 m from the same road (Section 2.5.3. Table 2-8). The wide
variation in ambient NC>2 concentrations across spatial and temporal scales, largely
influenced by vehicle emissions, can contribute to variation in NCh exposure within the
population and has important implications for adequately characterizing exposure.
1.4.2 Assessment of Nitrogen Dioxide Exposure in Health Effect Studies
Characterizing the adequacy of various exposure assessment methods to represent the
variability in ambient concentrations in a location is key in drawing inferences from
epidemiologic associations with health effects. Exposure is determined by concentrations
in specific ambient, indoor, and in-vehicle locations and time spent in those locations
(Section 3.4.1). People vary in the locations where they spend time and time spent in
those locations (Section 3.4.3.1). and NCh concentrations can vary widely across outdoor,
indoor, and in-vehicle locations (Figure 3-1). Measures of NC>2 exposure that do not fully
account for the variability in ambient concentrations and people's activity patterns have
some amount of error, and this error can impact the characterization of relationships
between NCh exposure and health effects. The extent and impact of error can differ by
exposure assessment method and by study design. Errors in representing the temporal and
spatial variability in short-term and long-term averages, respectively, of ambient NO2
concentrations in a given area and exposures of the population can attenuate relationships
between NCh exposure and health effects. For some long-term NO2 exposure estimates,
the mismatch in where NC>2 is measured and where people are located can inflate health
effect estimates. Exposure error also can impact the precision [i.e., 95% confidence
interval (CI)] of health effect estimates due to variable relationships between personal
1-9
-------
and ambient NO2 across people and time and differences in nonambient exposures. Thus,
for short-term exposure, this ISA emphasizes studies indicating that exposure metrics
well captured temporal (e.g., day to day) changes in NC>2. For long-term exposure, this
ISA emphasizes studies that well captured variability among people living in locations
that differ in ambient NC>2 concentrations (Section 3.4.5).
Ambient NC>2 concentrations at central site monitors represent both short-term and
long-term exposure with some amount of error. Central site monitors do not cover all
locations where people live or spend their time and also are not likely to capture the
temporal or spatial variability in ambient NC>2 concentrations in a given area. Long-term
personal NO2 exposures and their relationships with ambient NC>2 concentrations are not
well characterized. A wide range of correlations (0.12 to 0.43; Table 3-6) is observed
between short-term total (ambient plus nonambient components) personal and ambient
NC>2 concentrations and in ambient NC>2 concentrations across sites within some cities
(Section 2.5.2). On one hand, poor correlations do not necessarily mean that
concentrations at central sites are inadequate exposure metrics because the data may not
reflect relationships between ambient NO2 concentrations and the ambient component of
personal exposure (Section 3.4.2). On the other hand, the correlations could mean that
there is variation among individuals in how well short-term temporal changes in NO2
concentrations at central site monitors represent temporal changes in ambient exposure.
Proximity to roads may contribute substantially to short-term and long-term ambient NO2
exposure among people living or working near roads or commuting on roads, and the
2008 ISA for Oxides of Nitrogen cited the potential for in-vehicle exposures to dominate
short-term personal exposure (U.S. EPA. 2008c). Data from the U.S. near-road
monitoring network are too preliminary to allow for meaningful comparisons of the
temporal or spatial patterns in NO2 near and away from roads. However, annual avg NO2
concentrations often are higher at near-road sites than other sites within an urban area,
which is consistent with NO2 being formed from NO emitted by vehicles on the road.
These data indicate that central site monitors may not represent the magnitude of
long-term average NO2 concentrations near roads. Whether NO2 concentrations at central
sites and near-road sites differ with respect to correlations with personal exposures is
unknown. Thus, it is unclear how error produced from using ambient NO2 concentrations
at central site monitors to represent near-road exposures impacts health effect
associations. Another issue in estimating exposure from central site monitors is that the
chemiluminescence measurement method tends to overestimate ambient NO2
concentrations because of interference from other oxides of nitrogen. However,
interference generally is less than 10% in urban areas (Section 2.4.1) and may not vary
widely day to day (Section 3.4.3.4) to produce substantial error in characterizing daily
changes in NO2 concentration. It is not clear how interference compares among locations
1-10
-------
and what impact interference may have on comparisons of long-term average NC>2
concentrations among locations.
Given the aforementioned sources of uncertainty, there is confidence in some results
relating asthma exacerbation (Sections 5.2.2.2 and 5.2.2.5) to NO2 measured at central
site monitors based on demonstrations of good correlation in short-term NC>2 averages
among sites within a city or with total personal exposure. Results for short-term averages
of NO2 measured at people's locations (Sections 5.2.2.2 and 5.2.2.5) also are a source of
confidence. Such metrics include total NO2 exposure as well as NO2 measured outdoors
at schools and indoors at homes and schools. Further, personal ambient NO2 was
examined in natural experiments in which people spent well-defined periods of time in
outdoor locations (Section 5.2.2.2). A time-weighted average of NO2 concentrations in
people's locations correlated well with total personal short-term NCh exposures
(Section 3.4.3.1). Thus, although NC>2 concentrations in a specific location may not
represent potentially important exposures across the range of locations where people
spend time, they can represent a component of personal exposure and aid in inference
about NO2-related health effects. Spatially resolved exposure metrics also have shown
larger magnitude associations with health effects compared to NO2 measured at a single
central site monitor or averaged over multiple monitors in a city (Section 3.4.3.1).
Inference for results relating asthma exacerbation to short-term total personal or indoor
NC>2 concentrations also is strong because these metrics can help distinguish NCh-related
effects from the potential influence of other traffic-related pollutants. Because the mix of
sources differs indoors and outdoors, correlations between NO2 and some copollutants
are lower for total personal or indoor metrics than ambient metrics (Section 3.4.4.3. Table
3-13. Results for total personal and indoor NC>2 concentrations also can aid understanding
of health effects related to ambient exposure for populations whose indoor exposures are
affected by the penetration of ambient NO2 from open windows or other factors that
increase building air exchange rate (Section 3.4.3.3). In the case of asthma exacerbation,
one study indicated a good personal-ambient NO2 correlations.
As with short-term exposure, many studies indicate that their long term NC>2 exposure
metrics adequately capture the variation in ambient NO2 exposures among people. For
example, asthma development is associated with long-term average ambient NO2
concentrations measured at central site monitors 1 km from children's homes or schools
(Section 6.2.2.1). For asthma development and other health effects, there is an increase in
recent studies that use LUR models to estimate long-term NO2 exposures at the
neighborhood scale or at an individual's residence. Compared with NO2 estimated by
LUR, long-term average NO2 concentrations at central site monitors often show smaller
associations with a health effect but larger associations in some studies (Section 3.4.5.2).
Many epidemiologic studies in this ISA demonstrated their models to predict well the
1-11
-------
patterns in long-term average ambient NC>2 concentrations in the study areas. In these
studies, LUR models appear to account for differences among people in distance between
home and sources of NO2 (Section 3.5).
For short-term and long-term exposure, evaluating how well NC>2 metrics capture the
variability in ambient concentrations or exposure and the potential impact of exposure
error is a key consideration in drawing inferences about NC>2-related health effects from
epidemiologic studies. Particularly for asthma exacerbation and asthma development,
associations are observed with personal, central site, location-specific, or LURNCh
metrics that are indicated to well represent temporal and spatial variability in short-term
and long-term NC>2 exposure, respectively.
1.4.3 Factors Potentially Correlated with Nitrogen Dioxide Exposure to
Consider in Evaluating Relationships with Health Effects
The large influence of motor vehicle emissions on the distribution of ambient NO2
concentrations not only affects the assessment of NC>2 exposure but also has implications
for co-exposure to other traffic-related pollutants. NO2 concentrations are higher near
roads as are concentrations of elemental or black carbon (EC/BC), ultrafine particles
(UFP), carbon monoxide (CO), and VOCs (Section 3.3.1). The exact nature of gradients
varies among pollutants, but concentrations of traffic-related pollutants, including NO2,
decrease with increasing distance from the road. PM2 s1 and organic carbon (OC) do not
show clear gradients; however, a portion of PIVb 5 and OC comes from vehicle emissions.
These correlations and evidence that the copollutants show relationships with many of the
same health effects as NO2 and have similar modes of action (Appendix to this ISA)
point to the importance of evaluating the potential for NO2-related health effects to be
confounded (i.e., biased) by PM2s or traffic-related pollutants or for NO2 to represent a
mixture of such pollutants. Common sources, atmospheric reactions, or similar trends due
to meteorologic conditions extend the potential for co-exposures to pollutants beyond
those emitted from vehicles. Factors such as socioeconomic status (SES), season, and
temperature also show correlations with NO2 concentrations and relationships with
similar health effects. The potential for a particular factor to confound NO2-health effect
associations varies depending on the extent of correlation with NO2 concentrations, the
nature of the relationship with the health effect, and study design (i.e., whether temporal
variation in short-term exposure or spatial variation in long-term exposure is examined).
1 In general terms, paniculate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 um, a
measure of fine particles.
1-12
-------
Short-term average NO2 concentrations show a range of correlations with PM2 5 and
traffic-related copollutants (Figure 3-6. Table 3-10). but high correlations are observed
often. For example, for averaging times of 1 to 24 hours, the 25th to 75th percentile
ranges of correlation coefficients are 0.41-0.61 for PM2s, 0.58-0.67 for EC, and
0.59-0.96 for CO. Limited data indicate similar correlations with short-term averages of
VOCs, and lower correlations with OC. Long-term average ambient NO2 concentrations
show correlations with PM25 and CO similar to short-term averages, but the distribution
of correlations is shifted to higher values. Correlations of long-term averages of NO2 with
EC/BC, VOCs, OC, and UFP are not well characterized (Figure 3-6). Information on
seasonal correlations between ambient concentrations of NO2 and key copollutants is
sparse, but there is some indication of lower NO2-PM2 5 correlations for short-term
averages in the warm season (Section 3.4.4.1). These data point to potentially lower
confounding by PM2 5 in the warm season. Although traffic-related copollutants and
PM2.5 have been associated with many of the same health effects as NO2 (Appendix to
this ISA), the wide range of correlations with short-term and long-term average NO2
concentrations indicates variation among locations in confounding potential.
Much of the data characterizing correlations of NO2 with PM2 5 and traffic-related
copollutants are based on measurements at central site monitors. The varying spatial
patterns among pollutants may obscure true correlations across study areas or correlations
in personal exposure. Except for UFP, the few available data do not show systematically
higher correlations near roads (Figure 3-6). However, compared with ambient
concentrations, correlations can be weaker for short-term average personal exposures of
NO2 with PM25 (r = 0.06 to 0.38), EC (r = 0.22 to 0.49), and VOCs (r = -0.42 to 0.14)
(Table 3-13. Section 3.4.4.3). Correlations of short-term averages of NO2 with PM2s and
BC sometimes can be lower indoors than outdoors (Table 3-12 and Table 3-14). These
limited data indicate that associations of short-term personal or indoor NO2 exposures
with health effects may be less subject to confounding by PM2 5 or certain traffic-related
copollutants. In some locations, short-term average ambient NO2 concentrations are
related more strongly to personal PM than personal NO2 exposure. However, recent data
show negative to moderate correlations between ambient NO2 concentrations and
personal PM2 5 or EC (r = -0.19 to 0.44; Table 3-11). suggesting that ambient NO2
concentrations are not necessarily just a surrogate for personal PM exposure. The varying
correlations for short-term average concentrations of NO2 with other traffic-related
pollutants and PM2 5 across various microenvironments indicate that the potential for
confounding by the copollutants of primary concern varies by the exposure assessment
method. Similar information to compare copollutant correlations among
microenvironments is not available for long-term average NO2 concentrations.
1-13
-------
Other potential confounding factors to consider for long-term NC>2 exposure are measures
of traffic proximity or intensity, which could represent exposure to other pollutants that
display gradients with distance to road. Although NO2 is not unique to vehicle emissions
and can indicate sources such as off-highway vehicles and electric utilities (Section 2.3).
distance to roads, the length of nearby roads, and vehicle counts are predictors of ambient
NC>2 concentrations in LUR models (Section 3.2.2.1). Given recent findings linking
residential proximity to roads with respiratory effects and possibly with cardiovascular
effects and mortality (HEI. 2010). roadway proximity could confound NCh-health effect
associations by indicating exposure to traffic pollution. Studies considering the influence
of exposure to traffic, including residential proximity to roads, are another line of
evidence used to assess whether long-term NCh exposure independently affects health.
Short-term and long-term averages of NO2 also show a range of correlations with the
copollutants PMio,1 SC>2, and Os. Short-term and long-term average NO2 concentrations
tend to be moderately correlated with PMio (r for 25th-75th percentiles = 0.40-0.66 for
short-term averages, 0.44-0.75 for long-term averages) and 862 (Figure 3-6. Table 3-10).
Short-term averages of Os often are inversely correlated with NO2, and peak correlations
are moderate (r for 25th-75th percentile = -0.51 to 0.32) even in the summer, when Os
concentrations are higher (Table 3-10). Higher correlations are observed between
long-term averages of NO2 and Os (r for 25th-75th percentiles = 0.26-0.63). The wide
range of correlations observed for short-term and long-term average concentrations of
NC>2 with PMio, SCh, and Os indicates the variable potential for these pollutants to
confound health effect associations for NC>2. For short-term average NO2 concentrations,
the distributions of correlations with PMio and 862 are shifted to lower values compared
to correlations with most traffic-related pollutants, indicating the lower potential for
confounding. Specific to long-term exposure, relationships of long-term SC>2 and Os
exposure with many of the health effects evaluated in this ISA are uncertain (Appendix to
this ISA) as is their potential to confound NCh-health effect associations.
Residence near traffic has been linked to higher noise or stress levels, but information on
whether noise or stress confounds health effect associations with short-term or long-term
NC>2 exposure is limited. Weak to moderate correlations tend to be reported between
noise and short-term (r = 0.14-0.62) and long-term (r = 0.22-0.46) average ambient NC>2
concentrations, but high correlations have been observed for short-term NC>2 averages
(r = 0.83; Section 3.4.4.4). The impact of short-term changes in noise or stress on health
effects is not well characterized, but some data link long-term noise exposure and stress
to cardiovascular effects (Section 6.3.2) and decreases in cognitive function
1 In general terms, paniculate matter with a nominal mean aerodynamic diameter less than or equal to 10 ^im, a
measure of thoracic particles.
1-14
-------
(Section 6.4.4). Thus, the potential for stress or noise to confound NCh-health effect
associations is uncertain for short-term exposure but may exist for long-term exposure.
Other potential confounding factors to consider include temperature and humidity for
associations of health effects with short-term NO2 exposure because of similar
time-varying patterns as ambient NC>2 concentrations and health effects. Also, similar to
many health effects, short-term averages of ambient NO2 concentration vary by day of
the week and season and exhibit long-term time trends. For studies of long-term NC>2
exposure that compare individuals living in different locations, it is important to evaluate
confounding by factors such as SES, race (Sections 7.5.2 and 7.5.3). and age, all of which
can covary with long-term NCh exposures among individuals and spatially with
long-term ambient NC>2 concentrations among communities.
For studies reviewed in this ISA, the main method to account for potential confounding is
multivariable models that include NO2 concentrations and the putative confounder. The
NC>2 effect estimate represents the effect of NO2, keeping the level of the covariate
constant. Confounding is assessed by examining the change in the magnitude of the effect
estimate and width of the 95% CI, not a change in statistical significance. There are
limitations to multivariable models, and correlations between variables and the exposure
assessment method are important considerations in drawing inferences about
confounding (Section 5.1.2.1). High correlations between NC>2 concentrations and the
potential confounder can misleadingly decrease or increase the magnitude or precision of
the effect estimate for NO2 or the covariate and are a particular concern for models that
include a traffic-related copollutant or include three or more pollutants in the same
model. Potential differences in exposure measurement error between NO2 and the
copollutant also limit inferences from copollutant models about an independent NO2
association. Inference from copollutant models may be stronger for pollutants measured
at people's locations and for personal exposure than for pollutants measured at central
site monitors. As in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). a key issue
in this ISA is the adequacy in which epidemiologic studies examined potential
confounding by traffic-related copollutants and the extent to which other lines of
evidence support independent relationships between NO2 exposure and health effects.
1.5 Health Effects of Nitrogen Dioxide Exposure
This ISA evaluates relationships between an array of health effects and short-term
(Chapter 5) and long-term (Chapter 6) exposures to NO2 as examined in epidemiologic,
controlled human exposure, and animal toxicological studies. Short-term exposures are
defined as those with durations of minutes up to 1 month, with most studies examining
1-15
-------
effects related to exposures in the range of 1 hour to 1 week. Long-term exposures are
defined as those with durations of more than 1 month to years. Drawing from the health
effects evidence described in detail in Chapter 5 and Chapter 6 information on dosimetry
and modes of action presented in Chapter 4. as well as issues regarding exposure
assessment and potential confounding described in Chapter 3 and Section 1.4, the
subsequent sections and Table 1-1 present the key evidence that informs the causal
determinations for relationships between NC>2 exposure and health effects.
1.5.1 Respiratory Effects
The strongest evidence for relationships of short-term and long-term NC>2 exposure with
respiratory effects is that for asthma exacerbation and asthma development, respectively.
Such relationships also are supported by information on the dosimetry for inhaled NC>2
and by evidence for effects that can be linked together in a mode of action. Although it is
unclear how ambient-relevant NO2 exposures compare with NO2 produced endogenously
in the lung during inflammation and other immune responses (Section 4.2.2.4).
ambient-relevant concentrations of inhaled NO2 are absorbed throughout the respiratory
tract. The conducting airways have the primary role in asthma, and dosimetry models
predict that total NO2 dose is relatively constant across the tracheobronchial region
(Section 4.2.2.3). NC>2 is a reactive gas that rapidly reacts with antioxidants and other
constituents of the epithelial lining fluid of the respiratory tract. While antioxidant
reactions often are thought to reduce oxidant species, reactions with NO2 lead to the
formation of secondary oxidation products (Section 4.2.2.1). Antioxidant levels vary
across regions of the respiratory tract, and the variable physical and chemical nature of
the respiratory tract may influence the site in the respiratory tract of NC>2 uptake and
NO2-induced effects. The formation of secondary oxidation products likely is the
initiating event in the mode of action proposed for NCh (Section 4.3.2.1). These products
can induce oxidative stress, inflammation, allergic responses, and altered immune
function, all of which are events in the mode of action proposed for NC>2-related asthma
exacerbation and asthma development (Figures 1-2 and 4-1) as described in the sections
that follow.
1-16
-------
t Release of immune
mediatorsfrommast [•
cells
Inhalation
of NO2
Reactions in
respiratory tract
lining fluid and tissue
Formation of
oxidation/nitration
products
t Permeability of airways
f Allergic
responses
f Airway
inflammation
t Oxidative
\ Airway
responsiveness
,
\
Allergic
sensitization
persistent
Short-term
Exposure
Asthma
exacerbation
Development
of asthma
Long-term
Exposure
Airway
remodeling
i
Migration of inflammatory
and othermediatorsto
blood
Systemic inflammation
Oxidative stress
i
I Activation of cells lining
bloodvessels
Changes in heart rate,
heart rate variability
Other organ
effects
Short-term
or Long-term
Exposure
Cardiovascular
effects
Note: NO2 = nitrogen dioxide. Modified from Figures 4-1. 4-2. and 4-3 in Section 4.3.5 to depict the strength of evidence for effects
occurring at ambient-relevant concentrations. Solid arrows and dark boxes represent pathways for which there is consistent
evidence. Dotted lines and white boxes represent uncertain pathways because evidence is limited or inconsistent.
Figure 1-2 Characterization of the evidence for health effects related to
nitrogen dioxide exposure in a mode of action framework.
Respiratory Effects and Short-Term Exposure to Nitrogen Dioxide
A causal relationship exists between short-term NO2 exposure and respiratory effects
based on evidence for asthma exacerbation. The conclusion is strengthened from the
likely to be a causal relationship determined in the 2008 ISA for Oxides of Nitrogen
because the combined controlled human exposure and epidemiologic evidence can be
linked in a coherent and biologically plausible pathway to explain how NO2 exposure can
trigger an asthma exacerbation (Table 1-1). There is some evidence indicating that
short-term NC>2 exposure may be related to other respiratory effects, such as exacerbation
of allergy or chronic obstructive pulmonary disease (COPD), respiratory infection,
respiratory mortality, and respiratory effects in healthy people. However, because of
inconsistency across disciplines and/or limited information to support biological
plausibility, there is uncertainty whether short-term NCh exposure has independent
relationships with nonasthma respiratory effects (Section 5.2.9. Table 5-39).
1-17
-------
Although indicating NCh-associated asthma exacerbation, epidemiologic evidence on its
own does not rule out the influence of other traffic-related pollutants (Section 1.4.3). The
key evidence that NO2 exposure can independently exacerbate asthma are the findings
from previous controlled human exposure studies for increases in airway responsiveness
in adults with asthma following NCh exposures of 200 to 300 ppb for 30 minutes and
100 ppb for 1 hour. Airway hyperresponsiveness can lead to poorer control of symptoms
and is a hallmark of asthma. A recent meta-analysis shows that NCh exposure reduced by
one-half the dose of a challenge agent required to increase airway responsiveness, which
is a measure of a clinically relevant change. This evidence for clinically relevant
increases in airway responsiveness induced by NCh exposures that are not much higher
than peak ambient concentrations (Section 2.5) provides plausibility that asthma can be
exacerbated by ambient NO2 exposures. Biological plausibility also is supported by
experimental studies of adults with asthma showing NO2 exposures of 260 ppb for 15 or
30 minutes to enhance allergic inflammation, which by increasing airway responsiveness,
is a key event in the mode of action proposed for asthma exacerbation (Figure 1-2).
The NO2-induced increases observed in airway responsiveness and allergic inflammation
indicate that the epidemiologic evidence for increases in hospital admissions, ED visits,
and symptoms for asthma, as well as decreases in lung function in children with asthma
in association with short-term increases in NCh concentration, can plausibly be attributed
to NCh exposure. As uncontrolled symptoms are the major reason for seeking medical
treatment, coherence also is demonstrated among the various asthma-related outcomes
examined in epidemiologic studies. Associations are observed in studies with maximum
concentrations of 48-106 ppb for 24-h avg NCh and 59-306 ppb for daily 1-h max NCh.
Epidemiologic evidence is consistent across the methods used to estimate NO2 exposure
and include personal ambient and total NO2 measurements, NO2 measured outside
children's schools, NCh measured inside children's schools and homes, and ambient NC>2
concentrations averaged across central site monitors in a city. NCh measured at people's
locations, whether outdoors, indoors, or all locations combined, likely represent exposure
better than NCh measured at central site monitors and lend confidence in epidemiologic
evidence base relating short-term NCh exposure to asthma exacerbation. Further, the
results for airway responsiveness and allergic inflammation increasing after NCh
exposures of 100-300 ppb for up to 1 hour support the few epidemiologic results of
increased respiratory effects in adults with asthma and healthy adults associated with NCh
exposure (range 5.7-154 ppb) occurring over 2 or 5 hours at locations near roads.
Not all evidence supports NCh-related respiratory effects. NCh exposure has variable
effects on oxidative stress in experimental studies. NCh-related decreases in lung function
are observed in epidemiologic but not controlled human exposure studies. In this ISA,
lung function is distinguished from airway responsiveness assessments by co-exposure to
1-18
-------
a bronchoconstrictor in the latter but not the former. Neural reflexes do not appear to be
involved (Figure 1-2. Section 4.3.2.2). but NC>2-induced (500 ppb) mast cell
degranulation in rats suggests airway obstruction, which could lead to decreases in lung
function. Thus, additional coherence can be drawn among these results, evidence for
allergic inflammation, and the epidemiologic findings for NC>2-related respiratory effects
in populations with asthma that also had high prevalence of allergy.
NO2 associations with asthma-related effects persist with adjustment for temperature,
humidity, season, long-term time trends, as well as PMio, 862, or Os. Recent studies add
findings for NC>2 associations that persist with adjustment for a key copollutant such as
PM25 or those from traffic such as EC/BC, UFP, or CO (examined in few studies). Only
in a few studies are NO2 associations eliminated with adjustment for EC/BC, UFP, or a
VOC. Confounding by OC, PM metal species, or VOCs is poorly studied, but NO2
associations with asthma exacerbation tend to persist in the few available copollutant
models. In some cases, single-pollutant models indicate asthma-related effects in
association with NO2 but not PM2 5 or EC/BC, which were moderately correlated with
NO2 (r = 0.22-0.57). Recent epidemiologic results also suggest asthma exacerbation in
relation to indices that combine NO2 with EC, PIVb 5, Os, and/or SO2 concentrations, but
neither epidemiologic nor experimental studies strongly indicate synergistic effects
between NO2 and copollutants. Although causality cannot be confirmed from copollutant
models, results based on personal exposure or pollutants measured at people's locations
provide support for NO2 associations that are independent of PlVfcs, EC/BC, OC, or UFP
because of comparable measurement error among pollutants. Associations with personal
total and indoor NO2 measurements also support an independent effect of NO2 exposure
because the lower (e.g., r = -0.37 to 0.31) correlations observed with many traffic-related
copollutants compared to ambient NO2 concentrations indicate that the findings for
personal and indoor NO2 may be less prone to confounding by the same traffic-related
copollutants than findings for ambient NO2 concentrations (Section 1.4.3). In the indoor
studies, the relative contribution of indoor and outdoor sources to indoor NO2
concentrations are unknown. And, while associations of outdoor school NO2 with
asthma-related effects persist with adjustment for indoor NO2 in one group of children, it
is unclear whether indoor exposure alters responses of people to outdoor NO2 exposure.
The nature of the evidence from epidemiologic and experimental studies largely was
similar in the 2008 ISA. However, the 2008 ISA did not explicitly evaluate the coherence
and biological plausibility for specific respiratory outcome groups. Rather than new
evidence, the integrated experimental and epidemiologic evidence for asthma
exacerbation, with due weight to controlled human exposure studies, supports a causal
relationship between short-term NO2 exposure and respiratory effects. This includes the
uptake of NO2 in the respiratory tract and formation of reactive oxidation products.
1-19
-------
Further, the allergic inflammation and airway responsiveness shown in controlled human
exposure studies, asthma symptoms, hospital admissions, and ED visits, associations with
NO2 measured in people's locations (which may better represent exposure), and results
from copollutant models with a traffic-related copollutant describe a coherent,
biologically plausible pathway linking short-term NCh exposure to asthma exacerbation.
Respiratory Effects and Long-Term Exposure to Nitrogen Dioxide
There is likely to be a causal relationship between long-term NCh exposure and
respiratory effects based on evidence for the development of asthma. The conclusion is
strengthened from that determined in the 2008 ISA for Oxides of Nitrogen because
whereas previous epidemiologic findings were limited and inconsistent, recent evidence
consistently indicates associations between ambient NO2 concentrations and asthma
incidence in children and is supported by experimental studies that characterize a
potential mode of action for NO2 (Table 1-1). As with short-term NO2 exposure, the
evidence base varies across respiratory outcomes, and there is more uncertainty as to
whether long-term NC>2 exposure decreases lung function or lung development or
increases risk of COPD, respiratory infection, or respiratory mortality.
Providing a strong basis for relating long-term NCh exposure to asthma development,
many studies estimated NC>2 exposures at or near children's homes or schools. Asthma
incidence is associated with NO2 measured at sites 1 km from schools or homes and with
NO2 exposures estimated from LUR models that were shown to well predict measured
concentrations in the communities studied (R2 = 0.68 or 0.69; Section 6.2.9. Table 6-5).
Results also are consistent for less spatially resolved ambient NO2 concentrations at
central site monitors. Another strength of the recent epidemiologic studies is their aim to
isolate the development of asthma from the exacerbation of pre-existing asthma by
following children overtime, in several cases from birth, and examining NC>2 exposure
for periods preceding asthma diagnosis. Asthma incidence is associated with the average
NO2 concentration for the first year of life and NC>2 averaged over multiple years (study
means: 14 to 28 ppb), and no single critical exposure period is identified.
Associations with asthma are found with adjustment for SES, smoking exposure, gas
stove use, community of residence, and in one study, psychosocial stress. However,
potential confounding by traffic-related pollutants or proximity to roads is not examined.
The uncertainty in the epidemiologic evidence as to whether NCh exposure has an
independent effect on asthma development is reduced partly by the biological plausibility
provided by a small body of previous experimental studies that characterize a potential
mode of action linking NC>2 exposure with asthma development. NC>2 exposure (1,000 to
4,000 ppb) for 6-12 weeks increased airway responsiveness and allergic responses in
1-20
-------
rodents. Also lending support is the coherent mode of action information between studies
of short-term and long-term NC>2 exposure (Figure 1-2). Long-term NC>2 exposure also is
shown to increase oxidative stress and inflammation but not consistently across studies.
The temporal pattern of NO2 exposure underlying the epidemiologic associations with
asthma is not well delineated. However, a few experimental studies show that repeated
short-term NC>2 exposures over 4 to 14 days led to the development of allergic responses
in healthy adults and healthy rodent models (2,000-4,000 ppb) and to increased airway
responsiveness in rodents (4,000 ppb). This evidence for short-term NC>2 exposure
supports a relationship between long-term NC>2 exposure and asthma development
because it demonstrates the development of asthma-related effects in healthy humans and
animal models and indicates that repeated increases in exposure may be important. NO2
exposures that induce effects related to asthma development are higher than those that
induce effects related to asthma exacerbation as described in the preceding section but are
within the range of exposures considered to be ambient relevant (Section 1.2).
Epidemiologic studies continue to show associations of long-term NC>2 exposure with
decreases in lung function and development and increased respiratory disease severity in
children. These outcomes are associated with similar NO2 concentrations, durations, and
exposure assessment methods as asthma development (Table 6-5). However, there is
more uncertainty whether long-term NCh exposure independently can decrease lung
function or development or increase respiratory disease severity. Associations of
long-term NC>2 exposure with bronchitic symptoms or lung function persisted when
adjusted for PM2 5, EC, OC, or distance to freeway, but such findings are few in number
and inconsistent. Further, NC>2 exposure does not alter lung function in animal models,
and the hyperproliferation of lung epithelial cells and fibrosis in adult animals are not
related to the lung function changes described in children. While associations of lung
function with long-term NC>2 persist after adjustment for short-term NC>2 exposure, most
studies of symptoms do not assess the potential influence of short-term NO2 exposure.
Together, evidence from recent epidemiologic studies and previous experimental studies
supporting effects on the development of asthma indicates there is likely to be a causal
relationship between long-term NC>2 exposure and respiratory effects. Epidemiologic
studies observe associations with NO2 exposure estimated at or near children's homes or
schools, which may better represent differences in ambient NC>2 exposure among subjects
compared with less spatially resolved NO2 measurements from central site monitors.
Potential confounding by traffic-related copollutants largely is unexamined for asthma
development. However, findings from experimental studies for increased airway
responsiveness and allergic responses, which are part of the mode of action proposed for
asthma development, are considered to provide some support for an independent effect of
long-term NC>2 exposure. Because such evidence is limited, some uncertainty remains in
1-21
-------
attributing epidemiologic associations between long-term NO2 exposure and asthma
development specifically to NO2 among the array of traffic-related pollutants.
1.5.2 Health Effects beyond the Respiratory System
Epidemiologic studies show associations between NO2 exposure and health effects in
various organ systems, and associations are observed with a similar range of short-term
and long-term NCh concentrations as respiratory effects (Table 1-1). However, compared
to respiratory effects, there is more uncertainty in relationships with NC>2 exposure,
largely in identifying an independent effect from other traffic-related pollutants. For some
health effects, epidemiologic findings also are inconsistent. A common source of
uncertainty across nonrespiratory health effects is the limited availability of controlled
human exposure and/or toxicological studies to inform understanding of how
ambient-relevant exposures to NCh may affect biological processes that underlie the
health effects observed beyond the respiratory system. NO2 itself is not likely to enter the
blood (Section 4.2.2). Among the various products of NO2 reactions that occur in the
epithelial lining fluid of the respiratory tract, nitrite has been identified in the blood.
However, nitrite produced from inhaled NC>2 may not appreciably alter levels derived
from diet or induce potentially detrimental health effects (Section 4.2.3). Nitrite can react
with red blood cell hemoglobin to form methemoglobin. Methemoglobin has been linked
with health effects but has not been found with ambient-relevant NO2 exposure
concentrations (Section 4.3.4.1). A recent controlled human exposure study suggests that
mediators from the respiratory tract may migrate into the blood. This migration could
lead to systemic inflammation and oxidative stress (Figure 1-2. Section 4.3.5). providing
a potential mechanism by which NC>2 exposure could lead to health effects beyond the
respiratory system.
Cardiovascular Effects and Diabetes
Although it is not clear how inhaled NC>2 affects underlying biological pathways,
epidemiologic evidence indicates associations of short-term NCh exposure with
cardiovascular effects and long-term exposure with cardiovascular effects and diabetes.
For both short-term and long-term NC>2 exposure, the 2008 ISA for Oxides of Nitrogen
concluded that evidence was inadequate to infer a causal relationship with cardiovascular
effects (U.S. EPA. 2008c). There was supporting evidence for short-term NCh exposure
but uncertainty about potential confounding by traffic-related copollutants. Additional
findings relating short-term NCh exposure to the triggering of myocardial infarction
support a suggestive of, but not sufficient to infer, a causal relationship with
1-22
-------
cardiovascular effects (Table 1-1). A similar determination is made for long-term NO2
exposure, but the health effect category is expanded to include diabetes. Supporting
evidence previously was lacking, but new findings relate long-term NCh exposure to the
development of diabetes and heart disease. Evidence is inconsistent for the effects of
short-term and long-term NC>2 exposure on cardiovascular effects, such as arrhythmia,
cerebrovascular diseases, and hypertension. There still is uncertainty whether NC>2
exposure has effects that are independent of other traffic-related pollutants.
Recent epidemiologic studies continue to indicate that short-term NCh exposure may
trigger a myocardial infarction. There are consistent findings for associations between
short-term increases in ambient NC>2 concentration and hospital admissions or ED visits
for myocardial infarction, angina, and their underlying cause, ischemic heart disease
(Section 5.3.11.1, Table 5-52). Coherence is found with epidemiologic evidence for
NO2-related ST segment changes, a nonspecific marker of myocardial ischemia, and
increases in cardiovascular mortality, of which ischemic heart disease is the leading cause
(Tinegold et al.. 2013). The robustness of epidemiologic findings is demonstrated by the
fact that associations are consistently observed in studies conducted over several years, in
diverse geographic locations, and with data pooled from multiple cities. Also, as with
findings for asthma exacerbation (Section 1.5.1). associations of short-term NCh
exposure with effects related to myocardial infarction persist with adjustment for
meteorology, long-term time trends, and a copollutant such as PMio, SC>2, or Os.
(Section 5.3.11.1). Most of the epidemiologic evidence is based on NO2 exposures
assigned as the average ambient concentration across multiple monitors within a city;
however, ST segment changes are associated with outdoor residential NO2, which may
belter represent temporal changes in subjects' personal exposures.
New epidemiologic evidence for increases in diabetes and heart disease in relation to
long-term NC>2 exposure is suggestive of, but not sufficient to infer, a causal relationship
(Section 6.3.9. Table 6-11). The study reviewed in the 2008 ISA observed a weak
association with cardiovascular events. The most consistent recent findings are for
diabetes. Similar to asthma development, diabetes is associated with ambient NO2
estimated at subjects' homes using LUR models that were demonstrated to well predict
ambient NCh concentrations in the study areas. Most studies examine concurrent 1-yr avg
NO2 concentrations, but some aim to represent longer exposures more relevant to disease
development by examining people who did not change residence. There is also some
support for heart disease and mortality from ischemic heart disease related to long-term
NO2 exposure. Heart disease is associated with 1- or 2-yr avg NO2 concentrations
estimated at a neighborhood scale from central site monitors or dispersion models or at
subjects' homes with LUR. Most studies assess heart disease by acute cardiovascular
events such as myocardial infarction or hospital admissions without considering the
1-23
-------
potential influence of short-term NC>2 exposure. Some studies assess exposures for
periods after the cardiovascular event, and it is uncertain the extent to which these
periods represent exposures during disease development. In addition to assessing
residential NO2 exposures, many studies of heart disease and diabetes are noteworthy for
their large sample sizes, prospective follow-up of subjects (up to 20 years), and
adjustment for potential confounding by age, sex, SES, and comorbid conditions.
Despite the epidemiologic evidence relating cardiovascular effects and diabetes to
short-term and/or long-term NC>2 exposure, studies do not adequately account for
potential confounding by PNfe.s or traffic-related copollutants, as was the case in the 2008
ISA. In limited examination of copollutant models with PlVfcs, UFP, or CO, associations
of short-term NC>2 exposure with effects related to myocardial infarction are not
consistently observed. Confounding by other traffic-related pollutants has not been
examined. Also in contrast with findings for asthma exacerbation (Section 1.5.1).
copollutant model results are based on NO2 and copollutant concentrations measured at
central site monitors. Differential exposure measurement error may limit the reliability of
copollutant model results. Studies of long-term NO2 exposure and heart disease and
diabetes do not examine potential confounding by stress, PIVb 5, or traffic-related
copollutants. Evidence for NO2 associations that are independent of noise also is limited.
New findings from experimental studies point to the potential for NC>2 exposure to induce
cardiovascular effects and diabetes but are not sufficient to address the uncertainties in
the epidemiologic evidence. Consistent with findings that reactive products of inhaled
NC>2 or mediators of inflammation may migrate from the respiratory tract to the blood
(Figure 1-2). some recent experimental studies find increases in mediators of
inflammation and oxidative stress in the blood or heart tissue of healthy humans and
rodent models in response to short-term NC>2 exposure (Section 5.3.11.1). Evidence does
not strongly support the involvement of neural reflexes as examined by decreases in heart
rate variability or indirectly by changes in respiratory rate (Figure 1-2. Sections 4.3.2.2
and 5.3.11.2). Findings for increases in inflammation and oxidative stress describe early,
nonspecific changes induced by NCh exposure that have the potential to lead to
myocardial infarction. Although the findings are mostly for single-day exposures, they
also may describe a possible way for recurrent NO2 exposures to lead to the development
of heart disease or diabetes. Limited findings of dyslipidemia in rats and epidemiologic
findings of vascular damage in adults in relation to long-term NC>2 exposure also describe
potential pathways for NO2 exposure to lead to heart disease. The limited extent and
consistency of findings from experimental studies and nonspecific nature of most of the
evidence is not sufficient to demonstrate an independent effect of NO2 exposure.
1-24
-------
In conclusion, evidence is suggestive of, but not sufficient to infer, causal relationships
for cardiovascular effects and diabetes with short-term and/or long-term NO2 exposure.
Conclusions were changed from the 2008 ISA based on more epidemiologic evidence
linking myocardial infarction to short-term exposure and new evidence linking heart
disease and diabetes to long-term exposure. However, an independent effect of NO2
exposure is not clearly demonstrated. Examination of confounding by PM2 5 and
traffic-related copollutants is absent for long-term NO2 exposure and gives inconsistent
results for short-term NO2 exposure. Some but not all recent experimental studies show
that short-term NO2 exposure increases inflammation and oxidative stress in the blood or
heart tissue. Increases in inflammation and oxidative stress describe a potential way for
short-term or long-term NO2 exposure to lead to cardiovascular effects and diabetes, but
because the findings are not linked to any specific health effect, unlike the mode of action
information for asthma exacerbation or development (Section 1.5.1), they do not rule out
chance, confounding, and other biases in the epidemiologic evidence.
Total Mortality
Similar to the evidence described above for cardiovascular effects and diabetes,
epidemiologic evidence supports associations of both short-term and long-term NO2
exposure with total mortality from all nonaccidental causes. However, potential
confounding by PM2 5 and traffic-related copollutants remains largely unresolved, and it
is not clear what biological processes NO2 exposure may affect to lead to mortality. This
uncertainty weighed with the supporting epidemiologic evidence is the basis for
concluding that evidence is suggestive of, but not sufficient to infer, a causal relationship
for both short-term and long-term NO2 exposure with total mortality (Table 1-1). For
short-term exposure, the nature of the evidence has not changed substantively, resulting
in the same conclusion as the 2008 ISA. For long-term NO2 exposure, whereas evidence
in the 2008 ISA was limited, inconsistent, and inadequate to infer a causal relationship,
several recent epidemiologic studies report associations with total mortality, supporting a
stronger causal determination.
Evidence is suggestive of, but not sufficient to infer, a causal relationship between
short-term NO2 exposure and total mortality based on consistent epidemiologic findings
across geographic locations, including several studies pooling data across cities
(Section 5.4.8. Table 5-57). Ambient NO2 exposures were assessed as the average
concentration across central site monitors within a city, which has uncertainty in
adequately representing the temporal pattern in personal NO2 exposures. Similar to
findings for asthma exacerbation (Section 1.5.1), associations with mortality persist with
adjustment for meteorological factors, long-term time trends, and a copollutant among
PMio, SO2, or Os. A multicontinent study suggests interaction between NO2 and PMio,
1-25
-------
with higher PMio-mortality associations observed for periods of higher ambient NC>2
concentrations. However, in contrast with asthma exacerbation, potential confounding of
associations between short-term NCh exposure and total mortality by PM2 5 or
traffic-related copollutants remains unexamined.
The generally supportive evidence from the large number of recent epidemiologic studies
is suggestive of, but not sufficient to infer, a causal relationship between long-term NO2
exposure and total mortality (Section 6.5.3. Table 6-18). Epidemiologic associations are
observed in large cohorts in diverse locations followed for long durations up to 26 years.
Increases in total mortality are found in association with NCh concentrations averaged
over 1 to 16 years and assessed for the year of death and for periods up to 20 years before
death. Not all studies observe associations, but the inconsistency does not appear to be
due to differences among studies in long-term average ambient NO2 concentrations or the
exposure period examined. Total mortality is associated with long-term NO2 exposure
assigned from central site monitors and exposures estimated at people's homes by LUR
models that well represented the spatial variability in ambient NO2 concentrations in the
study areas (R2 = 0.61 and 0.71). NCh associations persist with adjustment for potential
confounding by age, sex, smoking, education, and comorbid conditions. In a few studies,
associations between long-term NO2 exposure and mortality persist with adjustment for
traffic density or proximity, but confounding remains a concern because NC>2 associations
are inconsistently observed with adjustment for PIVb 5 or BC exposures estimated from
central site monitors or LUR models.
Evidence relating NO2 exposure to cardiovascular and respiratory effects can provide
understanding of whether NO2 exposure has an independent effect on mortality by
indicating whether NCh exposure affects the underlying causes of mortality. In the U.S.,
cardiovascular disease, namely ischemic heart disease, is the leading cause of death [35%
as cited in (Hoyert and Xu. 2012)1. Respiratory causes comprise a smaller fraction of
mortality (9% in the U.S.), but COPD and respiratory infections are among the leading
causes of all mortality in the world. As described in the preceding sections, independent
effects of short-term and long-term ambient NC>2 exposure on myocardial infarction,
heart disease, diabetes, COPD, and respiratory infection are uncertain. Strong evidence
demonstrates NCh-related asthma exacerbation, but asthma is not a leading cause of
mortality. Thus, it is not clear what spectrum of cardiovascular and respiratory effects
NC>2 exposure may induce to lead to mortality and by what biological processes
short-term or long-term NCh exposure may lead to mortality.
In conclusion, evidence is suggestive of, but not sufficient to infer, a causal relationship
for total mortality with both short-term and long-term NCh exposure based on supporting
epidemiologic evidence. The evidence bases for total mortality related to short-term and
1-26
-------
long-term NC>2 exposure share many characteristics. Although there is supporting
epidemiologic evidence, studies do not adequately account for potential confounding by
PM2 5 or traffic-related copollutants. Thus, it is uncertain the extent to which
epidemiologic findings for total mortality can be attributed specifically to short-term or
long-term NC>2 exposure. Also uncertain are the independent effects of NO2 exposure on
the cardiovascular and respiratory morbidity conditions that make up the leading causes
of mortality. Because potential confounding by traffic-related copollutants is largely
unaddressed and the biological processes underlying the effects of NO2 exposure on
mortality are unclear, chance, confounding, and other biases cannot be ruled out in the
epidemiologic evidence for short-term and long-term NC>2 exposure with total mortality.
Reproductive and Developmental Effects
The 2008 ISA for Oxides of Nitrogen concluded that evidence was inadequate to infer a
causal relationship between NO2 exposure and a heterogeneous group of reproductive and
developmental effects based on limited and inconsistent epidemiologic and animal
toxicological evidence for effects on birth outcomes. This ISA presents separate
conclusions for more defined categories of outcomes that are likely to occur by different
biological processes and exposure patterns over different stages of development:
(1) fertility, reproduction, and pregnancy (Section 6.4.2); (2) birth outcomes
(Section 6.4.3); and (3) postnatal development (Section 6.4.4). For all three categories,
there is a recent increase in epidemiologic studies. However, there is reasonable
consistency only in the finding for birth outcomes to support strengthening the causal
determination to suggestive of, but not sufficient to infer, a causal relationship with
long-term NC>2 exposure (Table 1-1). For all three categories of reproductive and
developmental effects, there is large uncertainty in identifying an independent effect of
NC>2 exposure. In particular, animal toxicological evidence to support biological
plausibility remains limited and inconclusive.
Fertility, Reproduction, and Pregnancy
Evidence is inadequate to infer a causal relationship between long-term NO2 exposure
and effects on fertility, reproduction, and pregnancy (Section 6.4.5. Table 6-14). This
conclusion is based heavily on findings from the epidemiologic studies of pre-eclampsia,
a pregnancy complication related to hypertension and protein in the urine (Table 1-1).
Associations are inconsistently observed with ambient NC>2 exposures estimated at homes
by LUR models that well predicted ambient NO2 concentrations in the study areas
(R2 = 0.59 to 0.86). Studies that observe associations considered confounding by maternal
age, smoking, SES, diabetes, and parity, but few examine other traffic-related pollutants
to assess the potential for confounding. Other lines of evidence to inform biological
1-27
-------
plausibility are not available. Toxicological studies have not examined effects related to
pre-eclampsia, and there is a lack of coherence with epidemiologic findings for
conditions that contribute to pre-eclampsia, such as gestational hypertension and
placental function. Inconsistent and limited findings from animal toxicological and/or
epidemiologic studies for detrimental effects on sperm quantity and quality, fertility,
maternal weight gain in pregnancy, and litter size add to the uncertainty regarding a
relationship of NC>2 exposure with fertility, reproduction, and pregnancy.
Birth Outcomes
Evidence is suggestive of, but not sufficient to infer, a causal relationship between NO2
exposure and effects on birth outcomes based primarily on recent epidemiologic
associations with fetal growth restriction (Section 6.4.5. Table 6-14). The combined
epidemiologic and toxicological findings for effects on birth weight and infant mortality
are inconsistent as are epidemiologic findings for preterm birth and birth defects.
Evidence for NCh-related decreases in fetal growth is not entirely consistent, but many
studies observe associations with ambient NO2 concentrations at homes estimated by
LUR models that well predict NO2 concentrations in the study areas (R2 = 0.68 to 0.91;
Table 1-1). A few studies observe stronger associations for children whose mothers spent
more time at home and less time outdoors in locations other than home, which may be
due to stronger correlations between residential ambient NO2 and personal exposures.
Other strengths of recent studies include fetal or neonatal physical measurements and
analysis of confounding by season of conception, maternal age, smoking, SES, and in one
study, noise. However, epidemiologic studies do not examine potential confounding by
traffic-related copollutants. Further, toxicological studies have not examined fetal growth,
and a potential mode of action for NO2 cannot be proposed (Figure 1-2). Prenatal ambient
NC>2 exposure is associated with a marker of inflammation in fetal cord blood but not
maternal blood. The role of inflammation in affecting birth outcomes is not clearly
established, and epidemiologic findings do not rule out effects of other pollutants. Thus,
despite the supporting evidence for fetal growth restriction, there is considerable
uncertainty in attributing epidemiologic findings specifically to NC>2 exposure.
Postnatal Development
Evidence is inadequate to infer a causal relationship between NO2 exposure and effects
on postnatal development based largely on the inconclusive findings across several recent
epidemiologic studies of cognitive function in children (Section 6.4.5. Table
6-14).Associations are inconsistently found for concurrent, infancy, or prenatal NC>2
exposure estimated at children's homes or schools with LUR models that well represent
the variability in ambient NCh concentrations in the study areas (R2 = 0.64 to 0.85;
Table 1-1). Further, confounding by traffic-related copollutants or stress is unexamined,
1-28
-------
although one study shows an association with decreases in memory, adjusting for noise.
The recent study indicating that short-term NC>2 exposure of adult rats induced oxidative
stress and neuronal degeneration, which potentially could lead to impaired cognitive
function, is not sufficient to address the uncertainties in epidemiologic findings. Findings
for other effects on postnatal development are both limited and inconsistent. Specifically,
evidence integrated from epidemiologic and toxicological studies is inconclusive for
motor function and psychological or emotional distress. Evidence is inconsistent for
decrements in attention and limited for autism as examined in epidemiologic studies and
for physical development as examined in toxicological studies.
Cancer
The best evidence base pointing to a possible relationship between NCh exposure and
cancer is that for lung cancer (Table 1-1). A few recent epidemiologic studies indicate
associations between NO2 exposure and leukemia, bladder cancer, and prostate cancer,
but findings for NC>2 exposure inducing carcinogenicity or mutagenicity in bone marrow,
spermatocytes, and lymphocytes is inconsistent and based on higher than
ambient-relevant NCh exposures. The findings for associations of NC>2 exposure with
lung cancer incidence and mortality from some recent epidemiologic studies combined
with some previous findings in rodents that NC>2 exposure may be involved in lung tumor
promotion is the basis for strengthening the causal determination from inadequate to infer
a causal relationship in the 2008 ISA for Oxides of Nitrogen to suggestive of, but not
sufficient to infer, a causal relationship (Section 6.6.9. Table 6-20).
Among the many recent epidemiologic studies, some report associations for NC>2 with
lung cancer incidence or mortality, but others do not. Findings are inconsistent for NC>2
exposure assessed from central site monitors and estimated at subjects' homes with
well-validated LUR models. In studies observing associations, NCh concentrations were
averaged over 1 year at the beginning of the study up to 30 years before the outcome.
Thus, there is evidence for associations with exposure durations considered to be relevant
for cancer. However, it is not clear whether LUR or dispersion models predicting
concentrations for periods a few years before cancer or mortality adequately account for
decreases in ambient NO2 concentration over years or represent longer duration
exposures because most studies do not report on changes in residence. Studies not finding
associations do not differ in mean NC>2 concentrations or exposure duration examined.
Many studies examined large numbers of cancer cases, followed adults for 7-30 years,
and adjusted for potential confounding by SES, smoking, diet, and occupational
exposures. One study observes an association of residential NO2 exposure with lung
cancer mortality that persists with adjustment for PIVb 5. But, examination of confounding
by diesel exhaust and other traffic-related pollutants is absent.
1-29
-------
NC>2 exposure does not independently induce lung tumor formation in various animal
models or transform other chemicals in the body into carcinogens at ambient-relevant
concentrations. However, some findings indicate a potential role forNCh in tumor
promotion. In some but not all studies, ambient-relevant NC>2 exposures increased lung
tumors incidence in rodents with spontaneously high tumor rates, with co-exposure to a
carcinogen, or injection with metastatic cancer cells. Increases in secondary oxidation
products in the respiratory tract (Section 1.5.1) and limited evidence for NC>2-induced
increases in hyperplasia of the lung epithelium of rodents are early events that have the
potential to mediate NCh-related lung cancer. While NO2 exposure impairs host defense
in animal models (Section 5.2.9). parameters more directly linked to antitumor immunity,
such as cytotoxic or regulatory T cells and interferon-gamma, have not been studied.
In conclusion, evidence is suggestive of, but not sufficient to infer, a causal relationship
between long-term NO2 exposure and cancer based on findings for lung cancer.
Associations between ambient NO2 concentrations and lung cancer incidence and
mortality are found in some but not all epidemiologic studies. NCh exposures, some at
higher than ambient-relevant concentrations, show an effect on lung tumor promotion in
rodents but do not directly induce carcinogenesis. Potential confounding by diesel
exhaust particles and other traffic-related copollutants is unaddressed and information to
support biological plausibility is limited. Therefore, chance, confounding, and other
biases cannot be ruled out based on the associations of long-term NO2 exposure with lung
cancer incidence and mortality observed in some epidemiologic studies.
1-30
-------
Table 1-1 Key evidence contributing to causal determinations for nitrogen dioxide exposure and health effects
evaluated in the Integrated Science Assessment for Oxides of Nitrogen.
Health Effect Category3 and Causal Determination13
NO2 Concentrations
Associated with Effects
Respiratory Effects and Short-Term Exposure (Section 5.2)
2016 ISA—Causal relationship. 2008 ISA—Sufficient to infer a likely causal relationship.
Key evidence
(Table 5-39)
Reason for change in
causal determination
Uncertainty remaining
Strongest evidence is for effects on asthma exacerbation. Controlled human exposure studies
demonstrate independent effect of NO2. In adults with asthma, NO2 exposures not much higher than
peak ambient concentrations induce clinically relevant increases in airway responsiveness and
increases in allergic responses, which are part of the proposed mode of action linking NO2 and asthma
exacerbation. Inconsistent experimental results for effects on lung function and respiratory symptoms in
absence of challenge agent.
Evidence from controlled human exposures provides plausibility for consistent epidemiologic evidence
for decreases in lung function and increases in respiratory symptoms in children with asthma and
increases in asthma hospital admissions and ED visits. Associations observed with NO2 measured at
central site monitors and at subjects' locations (i.e., personal ambient, outdoor school). Copollutant
models, based on pollutants measured at subjects' locations, show NO2 associations that are
independent of PlVh.s or, as examined in fewer studies, EC/BC, OC, UFP, VOCs, PM metals. NO2
associations persist with adjustment for meteorology, medication use, PM-io, SO2, orOs. Coherent
findings available for total personal and indoor NO2 with lower potential for copollutant confounding.
Uncertainty in the independent effect of NO2 on other respiratory effects (i.e., allergy exacerbation,
COPD exacerbation, respiratory infection, respiratory effects in healthy populations) due to limited
coherence among findings from epidemiologic and experimental studies.
Evidence from controlled human exposure studies plus epidemiologic evidence for NO2 exposures
assessed for subjects' locations and in copollutant models with PlVh.s or a traffic-related copollutant
demonstrate consistency, coherence, and biological plausibility for effect of NO2 exposure on asthma
exacerbation to rule out chance, confounding, and other biases with reasonable confidence.
Strength of inference from copollutant models about independent associations of NO2, especially with
pollutants measured at central site monitors. Potential exists for NCb-copollutant mixture effects.
Airway responsiveness: 200
to 300 ppbforSO min,
100 ppbfor 1 h
Allergic inflammation:
260 for 15 min and
581 ppb for 30 min
Overall study ambient
maximums
Central site monitors:
24-h avg: 55 to 80 ppb
1-h max: 59 to 306 ppb
Outdoor school:
24-h avg: 7.5 and 16.2 ppb
Personal ambient:
2-h avg: 77.7 and 154 ppb
Total personal:
24-h avg: 48 and 106 ppb
1-31
-------
Table 1-1 (Continued): Key evidence contributing to causal determinations for nitrogen dioxide exposure and
health effects evaluated in the Integrated Science Assessment for Oxides of Nitrogen.
Health Effect Category3 and Causal Determination13
NO2 Concentrations
Associated with Effects
Respiratory Effects and Long-Term Exposure (Section 6.2)
2016 ISA—Likely to be a causal relationship. 2008 ISA—Suggestive of, but not sufficient to infer, a causal relationship.
Key evidence
(Table 6-5)
Reason for change in
causal determination
Strongest evidence is for effects on asthma development. Consistent epidemiologic evidence from
recent cohort studies for associations of ambient NO2 averaged over 1-10 years with asthma incidence
in children. Associations found with NO2 estimated at homes and measured at central site monitors 1
km from homes or schools. NO2 associations persist with adjustment for SES and smoking exposure.
Potential confounding by traffic-related copollutants or proximity to roads not examined.
Small body of experimental studies show NO2 effects on hallmarks of asthma. Long-term exposure
increases allergic responses and airway responsiveness in rodents. Short-term exposure induces
development of allergic responses in humans and rodents. Inconsistent epidemiologic associations
between long-term NO2 exposure and development of allergic responses in children.
More uncertainty in relationships with other respiratory effects because of limited coherence among
disciplines. Epidemiologic evidence for increased severity of respiratory disease and decreased lung
function and lung development in children. Animal toxicological evidence for respiratory infection.
New epidemiologic evidence for associations of ambient NO2 exposure estimated at/near homes or
schools with asthma development and biological plausibility from a small body of experimental studies.
Uncertainty remaining
Some uncertainty remains in identifying an independent effect of NO2 exposure from traffic-related
copollutants because evidence from experimental studies for effects related to asthma development is
limited, and epidemiologic analysis of confounding is lacking.
Overall study ambient
means: 14 to 28 ppb for
residential annual avg
estimates
Individual city ambient
means: 9.6 to 51.3 ppb for
annual avg; 7.3 to 31.4 ppb
for 10-yr avg
Allergic responses: 2,000
ppb for 4 days in humans;
3,000 ppb for 2 weeks and
_ 4,000 ppb for 12 weeks in
rodents
Airway responsiveness:
• 1,000 to 4,000 ppb in
rodents for 6 or 12 weeks
Cardiovascular Effects and Short-Term Exposure (Section 5.3)
2016 ISA—Suggestive of, but not sufficient to infer, a causal relationship. 2008 ISA—Inadequate to infer a causal relationship.
Key evidence
(Table 5-52)
Reason for change in
causal determination
Strongest evidence is for effects related to triggering myocardial infarction. Consistent epidemiologic
evidence for ST segment changes, increases in hospital admissions and ED visits for myocardial
infarction and ischemic heart disease, and cardiovascular mortality. Most evidence is based on NO2
averaged across central site monitors in a city. Associations persist with adjustment for meteorology,
PM-io, SO2, or Os. NO2 associations inconsistent in copollutant models with PM2.5 or CO.
Some, but not entirely consistent, findings from experimental studies for early, nonspecific effects with
the potential to lead to myocardial infarction: increases in markers of inflammation and oxidative stress
in plasma of humans and heart tissue of rats. Inconsistent epidemiologic findings for inflammation.
Inconsistent evidence for cerebrovascular effects, arrhythmia, and hypertension.
Additional epidemiologic evidence for array of effects related to the triggering of myocardial infarction.
Uncertainty remaining
Effect of NO2 independent from traffic-related copollutants is uncertain because experimental evidence
is limited and not specific to myocardial infarction, and epidemiologic analysis of confounding is limited.
Potential exposure error associated with NO2 measured at central site monitors not well characterized.
Individual city ambient
24-h avg: 90th: 22 to 53
ppb; maximums: 58 to 135
ppb
Overall study ambient
1-h max: 90th: 68 ppb
Oxidative stress in rats:
5,320 ppb for 6 h/day, 7
days; inflammation in rats:
'2,660 and 5,320 ppb for
6 h/day, 7 days
. Inflammation in human cells
exposed to human plasma;
oxidative stress in human
plasma: 500 ppb for 2 h
1-32
-------
Table 1-1 (Continued): Key evidence contributing to causal determinations for nitrogen dioxide exposure and
health effects evaluated in the Integrated Science Assessment for Oxides of Nitrogen.
Health Effect Category3 and Causal Determination13
NO2 Concentrations
Associated with Effects
Cardiovascular Effects and Diabetes and Long-Term Exposure (Section 6.3)
2016 ISA— Suggestive of, but not sufficient to infer, a causal relationship. 2008 ISA—Inadequate to infer a causal relationship.
Key evidence: Strongest evidence is for development of diabetes and heart disease. Generally supportive, but not
(Table 6-11) entirely consistent, epidemiologic evidence from recent cohort studies for associations of diabetes,
myocardial infarction, and heart failure with ambient NO2 averaged over 1-2 year periods around time of
outcome assessment. Coherence with evidence for cardiovascular mortality. Associations found with
NO2 estimated at homes and measured at central site monitors. NO2 associations persist with
adjustment for age, sex, SES, comorbid conditions, and in a few cases, noise. Potential confounding by
traffic-related copollutants, proximity to roads, or stress not examined.
Some, but not entirely consistent, findings from experimental studies for early, nonspecific effects with
the potential to lead to heart disease or diabetes: dyslipidemia in rats with long-term NO2 exposure,
increases in markers of inflammation and oxidative stress in plasma of humans and heart tissue of rats
with short-term NO2 exposure. Inconsistent epidemiologic associations between long-term NO2
exposure and inflammation.
Reason for change in
causal determination
Large increase in recent epidemiologic studies of heart disease and diabetes, with generally supportive,
but not entirely consistent evidence. New evidence for estimates of residential NO2 exposure.
Uncertainty remaining
Effect of NO2 independent from traffic-related copollutants is uncertain because experimental evidence
is limited and not specific to heart disease or diabetes, and epidemiologic analysis of confounding is
lacking.
Overall study ambient
means: 4.2 to 31.9 ppb for
residential annual avg
estimates; 34 ppb for 9.5-yr
avg at central site monitors
Dyslipidemia in rats:
160 ppb for 32 weeks
Oxidative stress in rats:
5,320 ppb for 6 h/day, 7
days; inflammation in rats:
2,660 and 5,320 ppb for
6 h/day, 7 days
Inflammation in human cells
exposed to human plasma;
oxidative stress in human
plasma: 500 ppb for 2 h
Total Mortality and Short-Term Exposure (Section 5.4)
2016 ISA and 2008 ISA—Suggestive of, but not sufficient to infer, a causal relationship.
Key evidence:
(Table 5-57)
Consistent epidemiologic evidence for increases in total mortality in association with NO2 averaged
across central site monitors in a city. Associations persist with adjustment for meteorology, long-term
time trends, PM-io, SO2, orOs. Potential confounding by traffic-related copollutants not examined.
Evidence does not clearly describe independent NO2 effects on biological processes leading to
mortality. Large percentage of mortality is due to cardiovascular causes, for which independent effect of
NO2 is uncertain. The strongest evidence for respiratory morbidity is for asthma and is more limited or
inconsistent for COPD and respiratory infection, which are larger causes of mortality in adults.
Reason for no change in Effect of NO2 independent from traffic-related copollutants is uncertain because epidemiologic analysis
causal determination of confounding is lacking, and the independent effect of NO2 on biological processes (i.e., effects on
morbidity) that lead to mortality not clearly demonstrated. Potential exposure error associated with NO2
Uncertainty remaining measured at central site monitors not well characterized.
Individual city ambient
24-h avg maximums: 55 to
135 ppb
Individual city ambient
1-h max:
90th: 33 to 133 ppb
Maximums: 96 to 147 ppb
1-33
-------
Table 1-1 (Continued): Key evidence contributing to causal determinations for nitrogen dioxide exposure and
health effects evaluated in the Integrated Science Assessment for Oxides of Nitrogen.
Health Effect Category3 and Causal Determination13
NO2 Concentrations
Associated with Effects
Total Mortality and Long-Term Exposure (Section 6.5)
2016 ISA— Suggestive of, but not sufficient to infer, a causal relationship. 2008 ISA—Inadequate to infer a causal relationship.
Key evidence:
(Table 6-18)
Reason for change in
causal determination
Generally supportive, but not entirely consistent, epidemiologic evidence from recent cohort studies,
including those with extended follow-up (up to 26 years) of existing cohorts. Associations found with
NO2 averaged over 1 to 16 years for periods 0 to 20 years before death. Most evidence is based on
NO2 measured at central site monitors, but associations also observed with NO2 estimated at homes.
Associations found with adjustment for age, sex, smoking, education, comorbid conditions, and in some
cases, neighborhood-level SES. In limited analysis, NO2 associations persist with adjustment for traffic
proximity or density but mostly are attenuated in copollutant models with PlVhs or BC.
Evidence does not clearly describe independent NO2 effects on biological processes leading to
mortality. Large percentage of mortality is due to cardiovascular causes, for which independent effect of
NO2 is uncertain. The strongest evidence for respiratory morbidity is for asthma and is more limited or
inconsistent for COPD and respiratory infection, which are larger causes of mortality in adults.
Large increase in recent epidemiologic studies, with generally supportive, but not entirely consistent,
evidence. New evidence for estimates of residential NO2 exposure in some but not all recent studies.
Overall study ambient
means:
12.1 to21.7ppbfor
residential annual avg
estimates
13.9 to 33.6 ppbfor 1-yrto
15-yr avg at central site
monitors
Uncertainty remaining
Effect of NO2 independent from traffic-related copollutants is uncertain because epidemiologic analysis
of confounding is limited and inconclusive, and the independent effect of NO2 on biological processes
(i.e., effects on morbidity) that lead to mortality not clearly demonstrated. Potential exposure error
associated with NO2 measured at central site monitors not well characterized.
Reproductive and Developmental Effects Long-Term Exposure0
2008 ISA—Inadequate to infer a causal relationship for broad category.
Fertility, Reproduction, and Pregnancy (Section 6.4.2)
2016 ISA—Inadequate to infer a causal relationship.
Key evidence
(Table 6-14)
Reason for no change in
causal determination
Uncertainty remaining
Heterogeneous group of indicators of a successful pregnancy with little support for relationship with NO2
exposure. Inconsistent epidemiologic evidence among several recent studies for associations of
pre-eclampsia, increases in blood pressure, and systemic inflammation in pregnancy with NO2
estimated at homes with LUR or measured at central site monitors. Studies adjust for maternal age,
smoking, SES, diabetes, and parity. Lack of toxicological studies to inform a potential effect of NO2.
More limited and inconsistent epidemiologic evidence for effects on fertility. No effect on fertility in
rodents, but change in reproductive cycle found. No epidemiologic or toxicological evidence for effects
on sperm count or quality. Limited, inconclusive evidence in rodents for changes in pregnancy weight.
Increase in recent epidemiologic studies, but results lack sufficient consistency, including those for
residential estimates of NO2 exposure. Limited and inconclusive toxicological evidence does not provide
insight on a potential effect of NO2.
Overall study ambient mean
for pre-eclampsia:
31 ppb for residential 3rd
trimester avg estimate
1-34
-------
Table 1-1 (Continued): Key evidence contributing to causal determinations for nitrogen dioxide exposure and
health effects evaluated in the Integrated Science Assessment for Oxides of Nitrogen.
Health Effect Category3 and Causal Determination13
NO2 Concentrations
Associated with Effects
Birth Outcomes (Section 6.4.3)
2016 ISA— Suggestive of, but not sufficient to infer, a causal relationship.
Key evidence
(Table 6-14)
Reason for change in
causal determination
Strongest evidence is for fetal growth restriction. Generally supportive but not entirely consistent recent
epidemiologic evidence for decreased head circumference and fetal or birth length, particularly as
assessed with fetal or neonatal physical measurements. Associations found with NO2 estimated at
homes and measured at central site monitors. NO2 associations persist with adjustment for maternal
age, SES, smoking, alcohol use, and season of conception. Potential confounding by traffic-related
copollutants not examined, and no available toxicological studies to inform a potential effect of NO2.
Evidence for decreased birth weight in a study of rats, but large epidemiologic evidence base is
inconsistent. Inconsistent epidemiologic evidence for associations with preterm birth, birth defects, early
life mortality, and no or inconclusive toxicological evidence to inform a potential effect of NO2.
Large increase in epidemiologic studies, with generally supportive, but not entirely consistent, evidence
for associations between residential ambient NO2 exposure and fetal growth restriction.
Overall study ambient
means:
Entire pregnancy: 15.5 to
20 ppb
Specific trimesters: 7.8 to
36 ppb
Decreased birth weight in
rats: 1,300 ppb for 3 mo
Uncertainty remaining
Effect of NO2 independent from traffic-related copollutants is uncertain because evidence from
experimental studies and epidemiologic analysis of confounding are lacking.
Postnatal Development (Section 6.4.4)
2016 ISA—Inadequate to infer a causal relationship.
Key evidence
(Table 6-14)
Reason for no change in
causal determination
Uncertainty remaining
Inconsistent recent epidemiologic evidence for associations with neurodevelopmental effects, such as
cognitive function, attention, motor function, and emotional responses. Association found with indoor
NO2, but not consistently with ambient NO2 exposure estimated at home or school by LUR. Associations
found with adjustment for SES and, in one study, noise. Potential confounding inconsistently examined
for smoking and not examined for stress or traffic-related copollutants.
Limited and inconclusive toxicological evidence for effects on motor function and emotional responses.
In a study of adult rats, short-term NO2 exposure induced neurodegeneration and oxidative stress,
which have the potential to lead to neurodevelopmental effects.
Limited and inconclusive toxicological evidence for impaired physical development in rats and no
analogous epidemiologic investigation.
Large increase in epidemiologic studies of cognitive function, but results lack sufficient consistency,
including those for residential or school estimates of NO2 exposure. Limited and inconclusive
' toxicological evidence does not provide insight on a potential effect of NO2.
Overall study ambient
means for cognitive function:
16.5 ppb for concurrent
school annual avg estimate
15.7 ppb for prenatal home
annual avg estimate
Neurodegeneration in rat
brains: 2,500 ppb for 7 days
Oxidative stress in rat
brains: 5,320 ppb for 7 days
1-35
-------
Table 1-1 (Continued): Key evidence contributing to causal determinations for nitrogen dioxide exposure and
health effects evaluated in the Integrated Science Assessment for Oxides of Nitrogen.
Health Effect Category3 and Causal Determination13
NO2 Concentrations
Associated with Effects
Cancer and Long-Term Exposure (Section 6.6)
2016 ISA— Suggestive of, but not sufficient to infer, a causal relationship. 2008 ISA—Inadequate to infer a causal relationship.
Key evidence
(Table 6-20)
Reason for change in
causal determination
Uncertainty remaining
Best evidence is for lung cancer. Some, but not consistent, recent epidemiologic evidence from cohorts
followed for 7-30 years for associations of lung cancer incidence and mortality with NO2 exposures
averaged over 1 to 30 years. Inconsistency observed for NO2 estimated at homes and measured at
central site monitors. Associations persist with adjustment for smoking, diet, SES, and occupational
exposures, but confounding by diesel exhaust or other traffic-related copollutants largely not examined.
Lack of toxicological evidence for direct effect of NO2 in lung tumor induction, but findings in some
studies suggest a possible role for NO2 in lung tumor promotion with carcinogen co-exposure or with
metastatic cancer. Evidence for formation of secondary oxidation products in the respiratory tract and
limited evidence for hyperplasia of lung epithelium, which have the potential to lead to carcinogenicity.
Limited epidemiologic evidence for associations with cancers of other sites, but inconsistent findings for
mutagenic and genotoxic effects in experimental animals to support an independent effect of NO2.
Evidence in some, but not all, epidemiologic studies for lung cancer incidence and mortality, including
associations with residential estimates of NO2 exposure. Some, not entirely consistent, toxicological
evidence for role of NO2 in lung tumor promotion.
Effect of NO2 independent from traffic-related copollutants is uncertain because epidemiologic analysis
of confounding and results from experimental studies that NO2 acts as a direct carcinogen are lacking.
Overall study ambient
means:
12.1 to 23.2 ppb for
residential annual avg
estimates
Individual city ambient
means:
6.4 to 32.4 ppb for 10-yr avg
at central site monitors
6.4 to 32.4 ppb for 3-yr avg
. at central site monitors
Lung tumor promotion in
rodents: inconsistent 250 to
5,000 ppb for 6 to
' 17 months
Avg = average; BC = black carbon; CO = carbon monoxide; COPD = chronic obstructive pulmonary disease; EC = elemental carbon; ED = emergency department; h = hour;
ISA = Integrated Science Assessment; km = kilometer; min = minutes; max = maximum; mo = months; NO2 = nitrogen dioxide; O3 = ozone; OC = organic carbon; PM25 = particulate
matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PM10 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm;
ppb = parts per billion; SES = socioeconomic status; SO2 = sulfur dioxide; UFP = ultrafine particles; VOC = volatile organic compound; yr = year.
aA large spectrum of outcomes is evaluated as part of a broad health effect category including physiological measures (e.g., airway responsiveness, lung function), clinical outcomes
(e.g., respiratory symptoms, hospital admissions), and cause-specific mortality. Total mortality includes all nonaccidental causes of mortality and conclusions are informed by the
nature of the evidence for the spectrum of morbidity effects (e.g., respiratory, cardiovascular) that can lead to mortality. The sections and tables referenced include a detailed
discussion of the available evidence that informed the causal determinations.
bSince the completion of the 2008 ISA for Oxides of Nitrogen, the phrasing of causal determinations has changed slightly, and the weight of evidence that describes each level in the
hierarchy of the causal framework has been more explicitly characterized.
°ln the 2008 ISA, a single causal determination was made for the broad category of reproductive and developmental effects. In this ISA, separate causal determinations are made for
smaller subcategories of reproductive and developmental effects based on varying underlying biological processes and exposure patterns over different lifestages.
1-36
-------
1.6 Policy-Relevant Considerations
As described in the Preamble and Section 1.1. this ISA addresses policy-relevant issues
that are aimed at characterizing quantitative aspects of relationships between ambient
NO2 exposure and health effects and the impact of these relationships on public health.
To that end, this section integrates information from the ISA to describe NCh exposure
durations and patterns related to health effects, the shape of the concentration-response
relationship, regional heterogeneity in relationships, the adverse nature of health effects,
and at-risk populations and lifestages. In addressing these policy-relevant issues, this
section focuses on respiratory effects, for which the evidence indicates there is a causal
and likely to be a causal relationship, respectively, with short-term and long-term NC>2
exposure. Because of uncertainty in the independent effects of NO2 exposure, other
health effects are discussed if they potentially provide new insight on a particular issue.
1.6.1 Durations of Nitrogen Dioxide Exposure Associated with Health Effects
The primary NC>2 NAAQS are based on 1-h daily max concentrations (3-yr avg of each
year's 98th percentile) and annual average concentrations. These NAAQS were set to
protect against an array of respiratory effects associated with short-term NCh exposures
and various health effects potentially associated with long-term exposure (Section 1.1).
Thus, an important consideration in the review of the primary NCh NAAQS is whether
the nature of the health effects evidence varies by NO2 exposure duration.
For short-term exposure, the majority of previous and recent evidence associates health
effects with 24-h avg ambient NO2, but the small body of evidence is equally consistent
for subdaily averages, such as 1 or 8-h max NO2 and NO2 averaged over periods of 2 or
5 hours. The 24-h avg and 1-h max ambient NC>2 metrics, assessed primarily from
concentrations averaged across multiple monitors within a city, are associated with a
spectrum of effects related to asthma exacerbation. In the few within-study comparisons
and based on typical increases in 24-h avg and 1-h max ambient NO2 concentrations (20
and 30 ppb, respectively; Section 5.1.2.2). effect estimates for the two highly correlated
NC>2 metrics did not clearly differ (Sections 5.2.2 and 5.2.7). A study of asthma-related
ED visits in Atlanta, GA observed similar associations for 1-h max and 24-h avg NC>2
with a 1-day lag, and a slightly larger association for 6-h nighttime avg NC>2 [12:00 ante
meridiem (a.m.)-6:00 a.m.; Section 5.2.2.41. Based on measurements from central site
monitors, the distribution of concentrations and spatial heterogeneity varied among the
array of NCh averaging times, which may account for differences in associations with
asthma ED visits. For example, nighttime avg NC>2 had a wider range of concentrations
1-37
-------
than 24-h avg NCh. Nighttime avg NC>2 was similar to 1-h max NC>2 in spatial
heterogeneity but lower in concentration. The spatial heterogeneity in ambient NC>2
concentrations within urban areas and with distance to roads (Sections 2.5.2 and 2.5.3)
and diurnal trends with higher concentrations measured during morning commute hours
(Section 2.5.4) are not unique to Atlanta, GA. This heterogeneity in ambient NC>2
concentrations, along with diurnal variation in people's time-activity patterns, suggest
that the array of NO2 averaging times vary in the extent to which they represent people's
exposures, which could obscure true differences in association with health effects.
NC>2 measurements aligned with subjects' locations, including total and ambient personal,
outdoor and indoor school, and indoor home NC>2, are associated with asthma-related
effects (Sections 5.2.2.2 and 5.2.2.5) and mostly are integrated over 1 or multiple days.
These results do not necessarily mean that continuous exposure is required, as any diurnal
pattern of NO2 exposure that may underlie associations with asthma-related effects
cannot be discerned. The relative importance of daily average exposures or acute peaks in
exposure occurring as a result of diurnal variation in ambient concentrations is not clear.
Any contribution of acute peaks in indoor NC>2 exposures (Table 3-4) to associations
observed between 3-day or 4-week avg indoor NO2 and asthma-related effects also is not
known. However, NC>2 exposures of 2 or 5 hours during time spent outdoors are related
to pulmonary inflammation and lung function decrements in adults (Section 5.2.9.3).
Inference from these results is strong because they are based on personal ambient NO2
measurements or NC>2 measured at the locations of outdoor exposures. Controlled human
exposure studies showing clinically relevant increases in airway responsiveness
(Section 5.2.2.1) and allergic inflammation (Section 5.2.2.5) in adults with asthma in
response to 100-400 ppb NO2 exposures in the range of 30 minutes to 6 hours provides
biological plausibility for subdaily ambient NC>2 exposures inducing asthma exacerbation.
With respect to long-term ambient NC>2 exposure, asthma development in children is
associated with 1-yr avg concentrations estimated at homes by LUR models with good
predictive accuracy and 10-yr avg concentrations measured at central site monitors 1 km
from homes or schools (Section 6.2.2.1). The NC>2 concentrations averaged over 1 year
during prenatal or infancy periods could represent critical time windows of exposure for
asthma development or represent longer durations of NCh exposure for subjects who
remain in the same home or neighborhood. Experimental studies do not provide direct
insight into what the epidemiologic findings may be indicating are important periods of
long-term NC>2 exposure for asthma development because experimental studies examined
NO2 exposures of less than one year in adulthood. However, findings for increased
allergic responses and airway responsiveness in humans or rodents indicate that repeated
increases in NC>2 exposure over multiple days or exposures over 1 to 3 months may play a
role in asthma development (Section 6.2.2.3).
1-38
-------
Overall, asthma exacerbation and asthma development are linked to a range of short-term
and long-term durations of NC>2 exposure, respectively. There is no indication of a
stronger association for any particular short-term or long-term duration of NC>2 exposure.
1.6.2 Lag Structure of Relationships between Nitrogen Dioxide Exposure and
Health Effects
Characterizing the NC>2 exposure lags (i.e., time between exposure and effect) associated
with health effects can aid in understanding the nature of relationships between NO2
exposure and health effects. The lag structure for associations with NO2 exposure may
vary among health effects depending on differences in the time course by which
underlying biological processes occur. Identifying important lag structures can depend on
whether the lag structure varies within the population according to differences among
individuals in time-activity patterns, pre-existing disease, or other factors that influence
exposure and responses to exposure. Another consideration in drawing inferences about
important lag structures is that differences in associations among exposure lags,
particularly single-day and multiday average NC>2 concentrations, may not only have a
biological basis but may be influenced by differences in the extent to which single-day
and multiday average ambient NO2 concentrations represent people's actual exposures.
Epidemiologic panel studies of children with asthma observed increases in pulmonary
inflammation and respiratory symptoms and decreases in lung function in association
with increases in NC>2 concentration lagged 0 day (same day as outcome) or 1 day and
multiday averages of 2 to 7 days (Section 5.2.2). Consistent with these findings, increases
in asthma-related hospital admissions and ED visits were observed in association with
NO2 concentrations lagged 0 or 1 day or averaged over 2 to 5 days. Whereas no particular
lag of NO2 exposure was more strongly associated with decreases in lung function,
several studies indicate larger increases in pulmonary inflammation, respiratory
symptoms, and asthma-related hospital admissions and ED visits for increases in
multiday averages of NO2 than single-day lags. Asthma-related effects also were
associated with multiday average NO2 concentrations (i.e., 2 to 4 days) for measures of
personal ambient and total NCh, outdoor school NO2, and indoor NC>2, which may better
represent exposure compared with measurements from central site monitors.
Studies in which adults with asthma and healthy adults were exposed for 2 or 5 hours in
outdoor traffic and nontraffic locations indicate decreases in lung function and increases
in pulmonary inflammation immediately or 2 hours after exposures (Sections 5.2.2 and
5.2.7). In both populations, decreases in lung function also were found the day after
exposures. In healthy adults, increases in pulmonary inflammation did not persist the day
1-39
-------
after outdoor exposure (Section 5.2.7.4). These data based on personal ambient exposure
assessment or NCh measured at the locations of people's outdoor exposures support other
epidemiologic findings showing increases in respiratory effects at lag 0 or 1 day of NCh
exposure and also indicate a similar lag structure for respiratory effects in people with
and without asthma. Experimental studies show that NCh exposure affects the biological
processes underlying the asthma-related effects observed in epidemiologic studies on a
similar time frame. Controlled human exposure studies found airway responsiveness in
adults with asthma to increase immediately after or 20 minutes to 4 hours after a single
NC>2 exposure and over 4 days of repeated exposure (Section 5.2.2.1). In experimental
studies, NO2 exposure enhanced allergic inflammation 30 minutes up to 19 hours after a
single- or 2-day exposure in humans and 7 days after exposure in rats (Section 5.2.2.5).
Thus, the findings from experimental studies provide biological plausibility for the
asthma-related effects observed in epidemiologic studies in association with 2- or 5-hour
exposures, same-day NO2 exposures, as well as exposures averaged over multiple days.
1.6.3 Concentration-Response Relationships and Thresholds
Characterizing the shape of the concentration-response relationship aids in quantifying
the public health impact of NO2 exposure. A key issue is whether the relationship is
linear across the full range of ambient concentrations or whether there are deviations
from linearity at and below the levels of the current 1-h NAAQS of 100 ppb and annual
NAAQS of 53 ppb. Also important for the review of the primary NC>2 NAAQS is
identifying ambient NC>2 concentrations below which there is uncertainty in the
relationship with health effects. Characterization of the concentration-response
relationship in epidemiologic studies is complicated by fewer observations in the low
range of ambient concentrations, the influence of other pollutants or risk factors for the
health effects, and variability among individuals in the population in their response to air
pollution exposures. The shape of the concentration-response relationship for health
effects related to short-term NCh exposure is examined in a limited number of
epidemiologic studies and for respiratory hospital admissions and ED visits and total
mortality rather than for other health effects.
Recent U.S. studies suggest a linear relationship between short-term NO2 exposure and
asthma ED visits in children (Section 5.2.2.4). In Atlanta, GA during 1993-2004, a linear
association was observed for 1-h maxNCh concentrations (lag 0-2 day avg) combined
across urban monitors by placing more weight on concentrations in more populated areas.
Risk estimates increased across quintiles of NO2 between 28 and 181 ppb (with NO2 less
than 28 ppb as the reference). Also, in nonparametric models, asthma ED visits in the
warm season (May-October) increased with increasing 1-h maxNCh concentrations
1-40
-------
between 11 and 37 ppb (5th to 95th percentiles). There is similar confidence in the
relationship throughout this range of concentrations; the 95% CI is relatively narrow even
at 11 ppb NO2. A relationship is uncertain at 1-h max NC>2 concentrations less than
11 ppb because effect estimates were reported to be unstable. In Atlanta, GA, the
distribution of 1-h max NCh varied across monitors, with higher concentrations at the
downtown site (mean 42 ppb). Thus, while a population-weighted average of NO2 may
better represent concentrations where people live and spend time, they may not clearly
indicate concentrations at which an association is not present. Analysis of 24-h avg NC>2
in Detroit, MI during 2004-2006 does not indicate deviation from a linear relationship.
Risk estimated assuming linearity across the range of concentrations did not differ from
risk estimated for 24-h avg NCh concentrations above 23 ppb (point of deviation from
linearity, between the 82nd and 85th percentiles) in the nonlinear model. NC>2
concentrations were averaged between two Detroit, MI sites, and comparisons of ambient
NO2 concentrations between sites were not reported. These limited findings from U.S.
cities suggest that the association between short-term NCh exposure and asthma ED visits
in children is present at NC>2 concentrations typical of U.S. urban areas (Section 2.5.1).
The concentration-response relationship for short-term NCh exposure and asthma-related
effects is not well examined in controlled human exposure or animal toxicological
studies. Combining data across multiple studies, a recent meta-analysis observed that
NC>2 exposure cut in half the dose of the challenge agent required to induce an increase in
airway responsiveness (i.e., provocative dose) in adults with asthma, but the provocative
dose did not change with increasing NC>2 concentration in the range of 100-500 ppb
(Figure 5-1). Experimental studies do not provide insight on whether asthma responses
increase with increasing NC>2 concentration because few studies examined multiple NO2
exposure concentrations, and the range of these NO2 concentrations (greater than
100 ppb) exceed those examined in epidemiologic studies of concentration-response.
Linear concentration-response relationships also are observed for mortality associated
with short-term NC>2 averages in the U.S., Canada, and Asia based on comparisons of
linear and various nonlinear models with natural and cubic splines or quadratic and cubic
terms for NC>2 (Section 5.4.7). A few previous results point to nonlinear associations but
for health effects for which the concentration-response relationship has not been widely
examined, including cough in children in the general population or cardiovascular
hospital admissions in adults. These studies tend to find NCh-related increases in effects
that are larger in magnitude per increment in NC>2 concentration in the lower range of
NO2 concentrations than in the upper range of concentrations. The implications of results
for these nonasthma health effects is less clear given the uncertainty as to whether NO2
exposure has independent relationships with nonasthma health effects.
1-41
-------
For long-term NC>2 exposure, information on the shape of the concentration-response
relationship with asthma development is too limited to draw inferences. In analyses of
tertiles or quartiles of estimates of residential NO2 exposure (Section 6.2.2.2). a linear
concentration-response is indicated in one study but not another. In the study observing a
linear relationship, annual average NO2 concentrations ranged from 1.8 to 24 ppb, but
because tertiles of NO2 concentration were not reported, the range of NO2 concentrations
where there may be more or less uncertainty in the relationship with asthma development
cannot be assessed. Also based on categories of NO2 concentration or splines, linear
associations are observed for long-term averages of NO2 with asthma symptoms in
children, chronic bronchitis in adults, and asthma hospital admissions in adults
(Section 6.2.3). These findings may not be attributable specifically to long-term NO2
exposure but rather, reflect associations with short-term NC>2 exposure. Analysis of the
concentration-response with categories of long-term average NO2 concentrations does not
provide a strong basis for assessing whether there is a threshold for respiratory effects.
In summary, the shape of the concentration-response relationship is better characterized
in epidemiologic studies and for short-term NC>2 exposure than long-term exposure. Few
controlled human exposure or toxicological studies of asthma-related effects examined
multiple NO2 exposure concentrations; therefore, that evidence lacks strong insight into
the concentration-response relationship. Based on an array of methods, including analysis
of splines, higher order terms for NC>2 (e.g., quadratic, cubic), and categories of NO2
concentration, previous and recent evidence indicates a linear relationship between
short-term NC>2 exposure and hospital admissions or ED visits for asthma and multiple
respiratory conditions combined. In Atlanta, GA, a linear relationship with asthma ED
visits is indicated for 1-h maxNCh concentrations averaged over 3 days, with similar
confidence in the relationship across the range of 11 to 37 ppb. There is uncertainty in the
relationship at concentrations less than 11 ppb. Another source of uncertainty is that
24-h avg or 1-h max NC>2 concentrations were averaged across multiple central site
monitors within a city, which may not reflect varying distributions of concentrations
within the city or population exposures.
1.6.4 Regional Heterogeneity in Effect Estimates
In addition to examining the shape of the concentration-response relationship for
NO2-related health effects across the distribution of concentrations, studies have
examined whether associations vary across geographical regions. In one study,
heterogeneity was noted among Asian cities in the shape of the NCh-mortality
relationship. Information on regional heterogeneity is limited, particularly for the U.S.
and for relationships of NC>2 exposure with asthma exacerbation or development. There is
1-42
-------
no strong indication of heterogeneity in associations of short-term NCh exposure with
respiratory symptoms in children in the general population among Korean cities
(Section 5.2.7.3). A few studies observe regional heterogeneity in associations between
short-term NC>2 exposure and total mortality among European and Asian cities
(Section 5.4.7). A nonlinear concentration-response relationship observed in one of four
Asian cities was hypothesized to be due to differences among cities in mortality from
infection, air conditioning use, time spent by the population outdoors, or temperature. On
a smaller geographic scale, NCh-related respiratory effects do not clearly differ between
two cities in Ohio with similar ambient NO2 concentrations (Section 5.2.2.4) or
neighboring urban and suburban communities in Europe that differed in ambient NO2
concentrations (Section 7.5.5). Limited results point to potential within-city differences in
asthma exacerbation in relation to short-term NO2 exposure. NO2-related asthma ED
visits were larger in Bronx than Manhattan, NY (Section 5.2.2.4), and NO2-related lung
function and pulmonary inflammation among children with asthma differed between two
El Paso, TX schools (Sections 5.2.2.2 and 5.2.2.5). The reasons for the heterogeneity
were not explicitly analyzed. In the El Paso study, the schools differed in proximity to
road, ambient NO2 concentrations, racial composition, and asthma medication use.
For long-term NO2 exposure, differences are observed between Chicago, IL; Houston,
TX; San Francisco, CA; New York, NY; and Puerto Rico in the association with asthma
prevalence among Latino and African American individuals ages 8-21 years
(Section 6.2.2.1). A test for heterogeneity was not statistically significant, but
associations are observed only in the San Francisco, CA and New York, NY cohorts.
Odds ratios for the average ambient NO2 concentration for the first year or first 3 years of
life are largest in the San Francisco, CA cohort, which comprised only African American
individuals. The reasons for heterogeneity among the locations were not explicitly
analyzed, but the locations differed in the distribution of ambient NO2, SO2, and PM25
concentrations, which may indicate varying air pollution mixtures among locations. San
Francisco, CA had lower ambient NO2 and SO2 concentrations than New York, NY.
PM2 5 and SO2 were associated with asthma prevalence in Houston, TX but not in New
York, NY or San Francisco, CA.
In summary, with limited available information, including one U.S. study of asthma
prevalence, it is not clear whether there is regional heterogeneity in the relationship
between short-term or long-term NO2 exposure and respiratory effects. There is some
evidence of heterogeneity in associations of short-term NO2 exposure with mortality
among cities in Europe and Asia. Given the uncertainty as to whether NO2 exposure has
an independent relationship with mortality, the extent to which the regional heterogeneity
in risk is applicable specifically to NO2 exposure is uncertain.
1-43
-------
1.6.5 Public Health Impact
The public health impact of air pollution-related health effects is determined by the
adverse nature of the health effects that are observed, the size of the population exposed
to the air pollutant or affected by the health outcome, and the presence of populations or
lifestages with higher exposure or increased risk of air pollution-related health effects.
Characterizing Adversity of Health Effects
Both the World Health Organization (WHO) and the American Thoracic Society (ATS)
have provided guidance in describing what health effects may be considered adverse.
WHO defines health as "the state of complete physical, mental, and social well-being and
not merely the absence of disease or infirmity" (WHO. 1948). By this definition, changes
in health outcomes that are not severe enough to result in a diagnosis of a clinical effect
or condition can be considered adverse if they affect the well-being of an individual. ATS
also has considered a wide range of health outcomes in defining adverse effects.
Distinguishing between individual and population risk, ATS described its view that small
air pollution-related changes in an outcome observed in individuals might be considered
adverse on a population level. This is because a shift in the distribution of population
responses resulting from an increase in air pollution exposure might increase the
proportion of the population with clinically important effects or at those at increased risk
of a clinically important effect that could be caused by another risk factor (ATS. 2000b).
Increases in ambient NO2 concentrations are associated with a broad spectrum of health
effects related to asthma, including those characterized as adverse by ATS such as ED
visits and hospital admissions (ATS. 2000b). ATS also describes lung function changes
occurring with symptoms as adverse, but experimental studies do not show symptoms
increasing after NO2 exposures of a few hours. NO2 exposure also is associated with
more subtle effects such as increases in airway responsiveness and pulmonary
inflammation and decreases in lung function (Section 1.5.1). Increases in airway
responsiveness and pulmonary inflammation are proposed as part of mode of action
linking NO2 exposure to asthma exacerbation and asthma development (Figure 1-2) and
show a distribution within populations. Based on ATS guidance, NO2-associated changes
in airway responsiveness or pulmonary inflammation may be considered adverse on a
population level because they can increase the proportion of the population with
clinically important changes that can lead to exacerbation or development of asthma. A
meta-analysis of controlled human exposure studies demonstrates thatNO2 exposures of
140-200 ppb for 1-2 hours reduces by one-half the dose of a challenge agent required to
increase airway responsiveness in adults with asthma (Section 5.2.2.1). Such observations
that NO2 concentrations not much higher than peak ambient concentrations can induce
1-44
-------
clinically relevant effects related to asthma exacerbation further support a role for
ambient NO2 exposures in inducing adverse health effects.
At-Risk Populations and Lifestages for Health Effects Related to Nitrogen
Dioxide Exposure
The primary NAAQS are intended to protect public health with an adequate margin of
safety. In so doing, protection is provided for both the population as a whole and those
groups potentially at increased risk for health effects from exposure to the air pollutant
for which each NAAQS is set (Preface to the ISA). Hence, the public health impact of
NC>2 exposure also is determined by whether specific lifestages or population groups are
identified as being at increased risk of NCh-related health effects. The large proportion of
the U.S. population living near roads, where ambient NO2 concentrations are higher
compared to many other locations (Section 2.5.3). indicates the widespread potential for
elevated ambient NO2 exposures. In 2009, 17% of U.S. homes were estimated to be
within 91m of sources of ambient NC>2 such as a four-lane highway, railroad, or airport
(Section 7.5.6). The percentage of the population with elevated NCh exposures may be
greater in cities. For example, 40% of the Los Angeles, CA population was estimated to
live within 100 m of a major road (Section 7.5.6). People spending time near roads and
commuting or working on roads also have the potential for elevated NO2 exposure, and in
turn, potential for increased risk of NCh-related health effects.
At-risk populations or lifestages also can be characterized by specific biological,
sociodemographic, or behavioral factors, among others. Since the 2008 ISA for Oxides of
Nitrogen and as used in the recent ISAs for Ozone (U.S. EPA. 2013e) and Lead (U.S.
EPA. 2013c). the U.S. EPA has developed a framework for drawing conclusions about
the role of such factors in modifying risk of air pollutant-related health effects (Table III
of the Preamble). Conclusions describe the confidence in the evidence based on
judgments of consistency and coherence within and across disciplines (Chapter 7).
Briefly, the evaluation is based primarily on studies that compare exposure or health
effect relationships among groups that differ according to a particular factor (e.g., people
with and without asthma). Where available, information on exposure, dosimetry, and
modes of action is evaluated to assess coherence with health effects evidence and provide
understanding of how a particular factor may increase risk of NO2-related health effects
(e.g., by increasing exposure, increasing biological effect for a given dose).
There is adequate evidence that people with asthma, children, and older adults are at
increased risk for NO2-related health effects, specifically effects indicative of asthma
exacerbation (Table 7-27). These conclusions are substantiated by the clear evidence of
an independent relationship of asthma exacerbation with short-term NO2 exposure
1-45
-------
(Section 1.5.1). Limited, supporting evidence suggests that females, people of low SES,
and people with low antioxidant diets have increased risk for NCh-related health effects.
The inconsistent evidence is inadequate to determine whether genetic variants, COPD,
cardiovascular disease, diabetes, obesity, race/ethnicity, smoking, urban residence, or
proximity to roads increase NC>2-related health effects. For many of these factors, the
common uncertainty is that the evidence is for cardiovascular effects, diabetes, or
mortality, which are not clearly related to NC>2 exposure.
A causal relationship between short-term NO2 exposure and respiratory effects is based
on the evidence for asthma exacerbation (Section 1.5.1). The increased risk for people
with asthma is supported further by controlled human exposure studies demonstrating
increased airway responsiveness at lower NO2 concentrations in adults with asthma than
in healthy adults (Section 7.3.1). Differences in NC>2 dosimetry (Section 4.2.2) or
exposure among people with asthma are not well described. Epidemiologic evidence does
not consistently indicate differences in NCh-related respiratory effects between children
with asthma and without asthma. However, because asthma is a heterogeneous disease
and the populations examined varied in prevalence of asthma medication use and atopy,
the inconsistent epidemiologic results are not considered to be in conflict with controlled
human exposure studies, which examined primarily adults with mild, atopic asthma.
The increased risk of NO2-related asthma hospital admissions and ED visits for children
(Section 7.5.1.1) and older adults (Section 7.5.1.2) suggests that among people with
asthma the effects of NO2 exposure may vary by lifestage. Although not clearly
delineated for NO2, several physiological and behavioral traits may contribute to the
increased risk for children. Compared with adults, children have developing respiratory
systems and increased oronasal breathing and ventilation rates (Section 4.2.2.3). Limited
data do not clearly indicate higher personal NO2 exposures in children (Table 3-5) but do
indicate more time and vigorous activity outdoors (Section 7.5.1.1). Thus, children may
have greater NO2 uptake in the respiratory tract and/or less exposure measurement error.
Many studies reported a higher proportion of asthma ED visits or hospital admissions
among children than other lifestages. Thus, higher incidence of asthma exacerbation in
children may be a reason for their increased risk.
Because the respiratory system continues to develop throughout childhood, it is possible
that critical time windows of exposure exist for NC^-related asthma development.
However, the evidence shows that asthma development in children is associated with
several different time windows of long-term NC>2 exposure: the prenatal period, infancy,
year of diagnosis, or lifetime exposure (Section 7.5.1.1). Studies do not consistently
identify a specific time window of long-term NCh exposure more strongly associated
with the development of asthma as ascertained in children ages 4-18 years.
1-46
-------
Children not only comprise a large proportion of the U.S. population (24% in the 2010
U.S. census) but also have a higher rate of asthma health care encounters than adults
(e.g., 10.7 vs. 7.0 per 100 persons with asthma).1 Further, asthma is the leading chronic
illness (9.5% prevalence) and reason for missed school days in children in the U.S. Many
U.S. schools are located near high-traffic roads (7% within 250 m; Section 7.5.6). NCh
concentrations outside schools are associated with asthma-related effects in children
(Sections 5.2.2.2 and 5.2.2.5). and school could be an important source of NC>2 exposure.
Based on the large number of children in the U.S., the high prevalence of asthma
morbidity among children, and potential for high NC>2 exposures, higher risks of asthma
exacerbation for children compared with adults can translate into large numbers of people
affected, magnifying the potential public health impact of NC>2 exposure.
The public health impact of NO2-related health effects also is magnified by the growing
proportion of older adults in the U.S. As with children, it is not well understood why
older adults have increased risk for NCh-related hospital admissions for asthma. Older
adults did not consistently have a higher proportion of asthma hospital admissions
compared with younger adults, so higher incidence of asthma exacerbation does not seem
to explain their higher NCh-related risk estimates. Differences in NO2 dosimetry also are
not described for older adults (Section 4.2.2.3). Time-activity patterns have been shown
to differ between older and younger adults, but there is not a clear difference in time
spent in a particular location that could explain differential exposure to NO2 in older
adults (Section 7.5.1.2). Older adults have higher prevalence of many chronic diseases
compared to younger adults (Table 7-2). COPD, cardiovascular diseases, and diabetes did
not consistently modify NO2-related health effects, but studies have not examined
whether co-occurring morbidity contributes to the increased risk of NO2-related asthma
exacerbation among older adults or whether age alone influences risk.
Although evidence does not clearly identify increased NO2-related health effects in
populations of low SES or nonwhite race or populations living near roads or in urban
areas, there is an indication of higher NO2 exposure among these groups. In particular,
some communities are characterized as having both higher ambient NO2 concentrations
and higher proportions of nonwhite and low SES populations (Section 7.5.2). Further, a
few studies characterize schools located near high-traffic roads as having high nonwhite
and low SES populations compared to schools located farther away from roads
(Section 7.5.6). Nonwhite and low SES populations also are recognized to have higher
risks of certain illnesses or diseases, including asthma, although it is not clear whether
higher NO2 exposure and higher risk of negative health effects interact to influence
NO2-related health effects in these groups. A recent study observed higher risk of
National Center for Health Care Statistics Data Brief. Available:
http ://www. cdc. gov/nchs/data/databriefs/db94. htm.
1-47
-------
NO2-related asthma hospital admissions among Hispanic children compared with white
children only in the low SES group (Section 7.5.2). While these findings suggest that
co-occurring risk factors in a population could influence the risk of NO2-related health
effects, information at present is too limited to draw firm conclusions.
In summary, the public health impact of NC>2 exposure is supported by many lines of
evidence. A large proportion of the U.S. population lives near roads or spends time near
or on roads, resulting in a large number of people potentially with elevated ambient NO2
exposure. NO2 exposure is linked to health effects that are clearly adverse, such as ED
visits and hospital admissions for asthma and development of asthma. NCh-related
increases in airway responsiveness can be considered adverse at a population level
because an increase in NO2 exposure can lead to an increase in the number of people with
clinically important effects. The public health impact of NC>2 exposure also is supported
by the increased risk for people with asthma, children, and older adults. The roles of
co-occurring risk factors or combined higher NC>2 exposure and health risk within a
population in influencing risk of NC>2-related health effects is not well understood. The
large proportions of children and older adults in the U.S. population and the high
prevalence of asthma in children can translate into a large number of people affected by
NO2 and thus magnify the public health impact of ambient NC>2 exposure.
1.7 Conclusions
There is a causal relationship between short-term NO2 exposure and respiratory effects.
This conclusion is stronger than that determined in the 2008 ISA for Oxides of Nitrogen
and is supported by the evidence integrated from controlled human exposure and
epidemiologic studies for asthma exacerbation. Asthma-related effects continue to be
associated with NCh concentrations at central site monitors, but recent epidemiologic
studies add evidence for associations with personal ambient and total NO2 measurements
as well as NO2 concentrations outside schools and inside homes. Epidemiologic evidence
continues to show independent associations of NO2 exposure with asthma-related effects
in copollutant models with PIVb 5 or a traffic-related pollutant such as EC/BC, OC, UFP,
CO, or a VOC. The potential influence of the full array of traffic-related pollutants or
mixtures has not been examined. Thus, the key evidence for an independent effect of NO2
are the findings from previous controlled human exposure studies that NO2 exposure not
much higher than peak ambient concentrations enhances allergic inflammation and
induces clinically relevant increases in airway responsiveness. These effects are
hallmarks of asthma exacerbation and suggest a mode of action linking NO2 exposure to
asthma exacerbation.
1-48
-------
There is likely to be a causal relationship between long-term NO2 exposure and
respiratory effects. The conclusion is strengthened from the 2008 ISA based on new
epidemiologic evidence for associations of asthma development in children combined
with biological plausibility from experimental studies. Epidemiologic studies did not
examine confounding by traffic-related copollutants. However, a small body of previous
experimental studies, which show that long-term and short-term NC>2 exposure increases
airway responsiveness and allergic responses in healthy humans and rodent models,
provide some indication that long-term NCh exposure may have an independent effect on
asthma development. For both short-term and long-term exposure, results for NC>2
measured or estimated in subjects' locations that were shown to well represent exposure,
provide a stronger basis for inferring relationships with respiratory effects.
Evidence is suggestive of, but not sufficient to infer, a causal relationship for short-term
NC>2 exposure with cardiovascular effects and total mortality and for long-term NCh
exposure with cardiovascular effects and diabetes, poorer birth outcomes, and cancer.
While there is continued or new supporting epidemiologic evidence, a large uncertainty
remains whether NC>2 exposure has an effect independent of traffic-related copollutants.
Epidemiologic studies have not adequately accounted for confounding, and there is a
paucity of support from experimental studies. Some recent experimental studies show
NO2-induced increases in systemic inflammation or oxidative stress. Such changes are
not consistently observed or necessarily linked to any health effect, unlike the mode of
action information available for asthma. The insufficient consistency of epidemiologic
and toxicological evidence is inadequate to infer a causal relationship for long-term NCh
exposure with fertility, reproduction, and pregnancy, as well as postnatal development.
As described above, key considerations in drawing conclusions about relationships
between ambient NO2 exposure and health effects include evaluating the adequacy of
NC>2 exposure estimates to represent the temporal or spatial patterns in ambient NO2
concentrations in a given study and separating the effect of NO2 from that of other
traffic-related pollutants. Although motor vehicle emissions in the U.S. have decreased
greatly over the last few decades, vehicles still are the largest single source of ambient
NC>2 in U.S. population centers and can contribute to spatial and temporal heterogeneity
in ambient NO2 concentrations. Recent information combined with that in the 2008 ISA
for Oxides of Nitrogen (U.S. EPA. 2008c) shows that ambient NC>2 concentrations can be
higher at locations within 200-500 m of a road compared with locations farther away.
Additionally, the first year of data from the U.S. near-road monitoring network show that
near-road sites on average have higher NO2 concentrations at than most other sites within
a given urban area but not always the day's highest 1-hour NCh concentration.
1-49
-------
As in the 2008 ISA, many studies assess exposure with ambient NC>2 concentrations
measured at monitors whose siting away from sources likely does not capture the
variability in ambient NC>2 concentrations within an area. The resulting error in
representing temporal variation in short-term exposure and spatial variation in long-term
exposure can produce smaller magnitude or less precise associations with health effects.
Such findings are similar to those reported in the 2008 ISA (U.S. EPA. 2008c). This ISA
additionally indicates that error produced from using NC>2 concentrations at central site
monitors to represent long-term exposure in some cases can increase health effect
estimates compared with residential NO2 exposure metrics from LUR models. Thus,
spatial misalignment of study subjects and ambient NCh concentrations potentially can
overestimate health effect associations with long-term NO2 exposure if the difference in
exposure between groups that differ in the health effect systematically is underestimated.
Given the potential impact of exposure measurement error, the additional epidemiologic
findings for exposures assessed for people's locations (e.g., ambient or total personal,
outdoor or indoor home or school) increases confidence in inferences about relationships
between ambient NO2 exposure with asthma exacerbation or asthma development. There
is confidence in this evidence also because relationships between personal and ambient
NO2 concentrations are variable for short-term averages and largely uncharacterized for
long-term averages. Data from the near-road monitoring network may help address gaps
in the understanding of the variability in ambient NC>2 concentrations and people's
exposures within urban areas and the potential importance of the near-road environment
as a source of NO2 exposure contributing to health effects.
In addition to determining causality, characterizing quantitative aspects of NO2-related
health effects is key to the review of the primary NC>2 NAAQS. Limited investigation
suggests a linear association for short-term ambient NO2 exposure with asthma ED visits.
The association is present at 1-h max NO2 concentrations frequently observed in U.S.
urban areas but uncertain at the lowest end of the concentration distribution. Recent
evidence continues to indicate that people with asthma, children, and older adults are at
increased risk for NC>2-related health effects. While recent evidence points to higher NO2
exposure among people of low SES or nonwhite race or people living in urban areas or
close to roads, it is not clear whether this higher NC>2 exposure leads to increased health
effects. Large numbers of people in the U.S. live near (e.g., within 100 m) or travel on
major roads and potentially have elevated exposures to ambient NO2 compared with
people away from roads. The large numbers of children and older adults in the U.S.
population and the high prevalence of asthma in children can translate into a large
number of people potentially affected by NO2 exposure and thus magnify the public
health impact of ambient NCh exposure.
1-50
-------
CHAPTER 2 ATMOSPHERIC CHEMISTRY AND
AMBIENT CONCENTRATIONS OF
OXIDES OF NITROGEN
2.1 Introduction
This chapter presents concepts and findings relating to emissions sources, atmospheric
science, and spatial and temporal concentration patterns for oxides of nitrogen. It is
intended as a prologue for detailed discussions on the evidence for human exposure to
and health effects of oxides of nitrogen that follow in the subsequent chapters, and as a
source of information to help interpret those effects in the context of data about
atmospheric concentrations.
In this Integrated Science Assessment (ISA), the term "oxides of nitrogen" (NOy) refers
to all forms of oxidized nitrogen (N) compounds, including nitric oxide (NO), nitrogen
dioxide (NCh), and all other oxidized N-containing compounds formed from NO and
NO2. NO and NO2, along with volatile organic compounds (VOCs), are precursors in the
formation of ozone (Os) and photochemical smog. NO2 is an oxidant and can react to
form other photochemical oxidants such as peroxyacyl nitrates (PANs) and toxic
compounds such as nitro-substituted poly cyclic aromatic hydrocarbons (nitro-PAHs).
NO2 can also react with a variety of atmospheric species to produce organic and
inorganic nitrates, which make substantial contributions to the mass of atmospheric
particulate matter (PM) and the acidity of clouds, fog, and rainwater. The abbreviation
NOx refers specifically to the sum of NO and NO2. This chapter describes the origins,
distribution, and fate of gaseous oxides of nitrogen. Aspects of particulate nitrogen
species [such as particulate nitrate (pNOs)] are addressed in the review of the National
Ambient Air Quality Standards (NAAQS) for PM [see 2009 ISA for Particulate Matter
(U.S. EPA. 2009a)1 and 2014 Call for Information (U.S. EPA. 2014c).
2.2 Atmospheric Chemistry and Fate
The chemistry of oxidized nitrogen compounds in the atmosphere was reviewed in the
2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). The role of NOx in O3 formation
was reviewed in Chapter 3 of the 2013 ISA for Ozone (U.S. EPA. 2013e) and has been
discussed in numerous texts [e.g., (Jacobson. 2002; Jacob, 1999; Seinfeld and Pandis.
1998)1. The main points from the 2008 ISA for Oxides of Nitrogen will be presented here
along with updates based on recent material.
2-1
-------
The overall chemistry of reactive, oxidized nitrogen compounds in the atmosphere is
summarized in Figure 2-1. Sources include naturally occurring processes associated with
wildfires, lightning, and microbial activity in soils. Anthropogenic sources are dominated
by emissions from motor vehicles and electricity generating units. Oxidized nitrogen
compounds are emitted into the atmosphere mainly as NO, with only 10% or less emitted
as NO2. Further details about the composition of sources is given in Section 2.3. Freshly
emitted NO is primarily converted to NO2 by reacting with Os, and NO is recycled during
the day by photolysis of NO2. Thus, NO and NO2 are often grouped together into their
own group or family, which the atmospheric sciences community refers to as NOx
(shown in the inner box in Figure 2-1). A large number of oxidized nitrogen species in
the atmosphere are formed from the oxidation of NO and NO2. These include nitrate
radicals (NOs), nitrous acid (HONO), nitric acid (FINOs), dinitrogen pentoxide (ISbOs),
nitryl chloride (C1NO2), peroxynitric acid (FiNO4), PAN and its homologues (PANs),
other organic nitrates like alkyl nitrates [including isoprene nitrates(IN)], and pNOs.
These reactive oxidation products are referred to collectively as NOz. All of the species
shown within the dashed lines of Figure 2-1 constitute NOy (NOy = NOx + NOz). The
boxes labeled "inorganic" and "organic" in Figure 1-1 (Chapter 1) contain the species
shown in the left and right halves of Figure 2-1.
2-2
-------
Long-range transport to remote
regions at low temperatures
- -NO, .
CINO, *L *,.
u_ R-C=C-R
N03 > RONO2
liopcene nitrates
alky! nltralet
NOZ
N0; - NO, - NO,
^->RONO,
•I
NOX
deposition
I°0Q
IA
MIXA/^
1-J
deposition
emissions
Note: The inner shaded box contains NOX (= NO + NO2). The outer box contains other species (NOZ) formed from reactions of NOX.
All species shown in the outer and inner boxes are considered "oxides of nitrogen" and collectively referred to as NOY by the
atmospheric sciences community.
hv = solar photon, M = species transferring/removing enough energy to cause a molecule to decompose/stabilize,
MPP = multiphase processes, R = organic radical.
Source: National Center for Environmental Assessment.
Figure 2-1 Schematic diagram of the cycle of reactive, oxidized nitrogen
species in the atmosphere.
High NO concentrations found near heavy traffic and in power plant plumes are typically
associated with Os concentrations much lower than in surrounding areas because Os can
be titrated away, or consumed, by reacting with NO. In addition, the reaction of NO with
Os can produce appreciable amounts of NO2 rather quickly. For example, 10 ppb NO2 can
be formed in about 20 seconds [for an initial NO concentration of 30 ppb and initial
Os = 40 ppb at 298 K (25°C)].1 Higher temperatures and concentrations of reactants
1 Sample calculation based on solution to an equation for a second-order reaction dx/dt = &([NO]0 - x)([O3]0 - x),
where x = concentration of each species reacted; k = rate coefficient for the reaction, 3 x 1CT12 e(~1500/7) cm3/sec-
molecule (Sander etal. 2011): T = temperature inkelvin; [NO]0 and [Oslo = initial concentrations of NO and Os.
2-3
-------
result in shorter times, while dispersion and depletion of reactants increase this time. A
rough estimate of the time for transport away from a broad boulevard is about a minute
(During etal. 2011); this time is shorter for more open conditions and ranges up to about
an hour in midtown Manhattan street canyons (Richmond-Bryant and Reff. 2012). The
time for reaction must be compared to the time for mixing away from the road and for
replenishment of Os because the interplay between these factors determines how far NO
will travel downwind before it is oxidized. These dependencies imply seasonal variability
and geographic variability in the time scale for the reaction. In general, cooler months
present the most favorable conditions for NO to travel further before it is oxidized (lower
temperature, decreased vertical mixing of Os to the surface, generally lower Os). At any
time of the year, if loss of Os has been extensive near the surface as happens in many
locations at night, then NO could travel a kilometer or more before being oxidized,
resulting in a more uniform downwind distribution of NO2 than if NO were being
oxidized right at its source. The NO2 that is formed depletes hydroxyl radicals (OH) so
that they cannot oxidize hydrocarbons to continue the cycle of new Os formation. During
the day, NO2 photolyzes back to NO within a few minutes, setting up the cycle shown in
Figure 2-1. Although the assumption of a photostationary state to describe the relations in
the NO/NO2/Os triad might not be strictly valid, several studies [e.g., During et al.
(2011); Clapp and Jenkin (2001)] have shown the assumption of a photostationary state
can provide a useful approximation of the relationship among these species. Once the sun
sets, NO2 no longer photolyzes to reform NO. If very little or no Os is present due to
titration in a statically stable, near-surface boundary layer, then NO2 accumulates through
the night solely from direct emissions.
Because of the interplay between dispersion and chemical reaction, the distribution of
NO2 downwind of roads would likely differ from that of a traffic pollutant that is present
in ambient air mainly as the result of direct emissions, such as ultrafine particles (UFP) or
carbon monoxide (CO) (Section 2.5.3). In addition, day-night differences in both
transport and chemistry will also result in day-night differences in the patterns of spatial
and temporal variability of NO2. Examples of the behavior of NO2 and NOx downwind of
streets and highways are examined in Section 2.5.3. In summary, the major influences on
NO2 concentrations within and downwind of urban centers are the fraction of emissions
of NOx as NO2, dispersion, and the NO/NO2/Os equilibrium, which is established on a
time scale of a few minutes during daylight.
All the other species mentioned above in the definition of NOy (i.e., NOz) are products of
reactions of NO or NO2. Inorganic NOz species are shown on the left side of the outer
box, and organic species are shown on the right side of the outer box in Figure 2-1.
Ammonium nitrate and other inorganic particulate species [e.g., sodium (Na+), calcium
(Ca2+) nitrates] are formed from species shown on the left side of the figure; organic
2-4
-------
nitrates are formed from species shown on the right side of Figure 2-1. The conversion of
NOx into the inorganic and organic species in the outer box (collectively referred to as
NOz) typically takes place on much longer time scales than do interconversions between
NO and NCh, which occur on timescales of seconds to minutes. For example, conversion
of NOx to NOz takes about an hour for conditions in Houston, TX, in April-May of 2009
(Renet al.. 2013) but it likely takes longer in many other areas, especially those at higher
latitudes and generally during the cold season. As a result, NOx emitted during morning
rush hour by vehicles can be converted almost completely to products by late afternoon
during warm, sunny conditions. However, note the conversion of NO2 to HNOs and
hence the atmospheric lifetime of NOx depends on the concentration of OH radicals,
which in turn depends on the concentration of NO2 [e.g., Valinetal. (2013); Hameed et
al. (1979)1.
Inorganic NOz species shown on the left side of the outer box of Figure 2-1 include
HONO, HNO3, C1NO2, HNO4, and pNO3. Pernitric acid (HNO4) is unlikely to represent
an important reservoir for NOx except perhaps under extremely cold conditions. Mollner
et al. (2010) identified pernitrous acid (HOONO), an unstable isomer of nitric acid, as a
product of the major gas-phase reaction forming HNOs. However, because HOONO is
unstable, it is also not a substantial reservoir for NOx. With consideration of the
troposphere as a whole, most of the mass of products shown in the outer box of
Figure 2-1 is in the form of PAN and HNOs. The stability of PAN at low temperatures
allows its transport to remote regions where it has been shown to exert strong influence
on the local production of Os [see Fischer etal. (2014) and references therein]. Other
organic nitrates (e.g., alkyl nitrates, isoprene nitrates) increase in importance in the
planetary boundary layer (PEL), particularly at locations closer to sources (Perring et al..
2013; Horowitz et al.. 2007; Singh et al.. 2007).
In addition to the above compounds, there is a broad range of gas-phase organic nitrogen
compounds that are not shown in Figure 2-1. They are emitted by combustion sources
and formed in the atmosphere from reactions of NO, NO2, and NOs. These compounds
include nitro-aromatics (e.g., nitrotoluene), nitro-PAHs [e.g., nitro-naphthalene; (Nishino
et al., 2008)1, nitrophenols [e.g., (Harrison et al., 2005)1. nitriles [e.g., ethane-nitrile; (de
Gouw etal.. 2003)]. and isocyanic acid (Roberts et al.. 2014).
Sources of NOx are distributed with height, with some occurring at or near ground level
and others aloft as indicated in Figure 2-1. NOx emitted by elevated sources can be
oxidized to NOz products and/or be transported to the surface, depending on time of day,
abundance of oxidants, and strength of vertical mixing. During times of rapid convection,
typically in the afternoon on hot sunny days, vertical mixing through the PEL can take
place in about 1 hour [e.g., Stull (2000)1. and fresh emissions can be brought rapidly to
2-5
-------
the surface. After sunset, turbulence subsides, and emissions entrained into the nocturnal
residual boundary layer are not mixed downward to the surface. Also, because the
prevailing winds aloft are generally stronger than those at the surface, emissions from
elevated sources (e.g., the stacks of electrical utilities) can be distributed over a wider
area than those emitted at the surface (e.g., motor vehicles). Emissions from elevated
sources entrained into the nocturnal residual boundary layer can be transported over long
distances, up to a few hundred kilometers overnight depending on location [e.g., Husar et
al. (1978)]. Oxidation of NOx can occur during the night and in the morning in the
residual layer before it breaks up. Turbulence then mixes NOx and its oxidation products
downward. Emissions directly into the free troposphere are unlikely except in areas such
as the Intermountain West where PEL heights can be <200 m during winter, or even
<100 m in some locations. Because people live closer to surface sources, such as motor
vehicles, they are more likely to be exposed to NO and NO2 from these sources. Thus,
atmospheric chemical reactions determine the partitioning of a person's exposure to NO2
and its reaction products from different sources, and the sources of a person's exposure
cannot be judged solely by the source strengths given in the National Emissions
Inventory (NEI). Issues related to the transport and dispersion of NOx emitted by traffic
are discussed in depth in Section 2.5.3.
Oxidized nitrogen compounds are ultimately lost from the atmosphere by wet and dry
deposition to the Earth's surface. Soluble species are taken up by aqueous aerosols and
cloud droplets and are removed by wet deposition by rainout (i.e., incorporation into
cloud droplets that eventually coagulate into falling raindrops). Both soluble and
insoluble species are removed by washout (i.e., impaction with falling raindrops, another
component of wet deposition), and by dry deposition (i.e., impaction with the surface and
gas exchange with plants). NO and NO2 are not very soluble, and therefore wet
deposition is not a major removal process for them. However, a major NOx reservoir
species, HNOs, is extremely soluble, and its deposition (both wet and dry) represents a
major sink for NOy.
Many of the species shown in Figure 2-1. including pNOs and gas-phase HONO, are
formed by multiphase processes. Data collected in Houston, TX as part of TexAQS-II
summarized by Olaguer et al. (2009) indicate that concentrations of HONO are much
higher than can be explained by gas-phase chemistry and by tailpipe emissions.
N2Os is the acid anhydride of HNOs, and its uptake on aqueous aerosol represents a major
sink for NOx. The uptake of N2Os by atmospheric aerosols or cloud droplets leads to the
loss of Os and NOx and the production of aqueous-phase nitric acid, aerosol nitrate, and
gaseous halogen nitrites. Maclntyre and Evans (2010) showed that the sensitivity of key
tropospheric species, such as Os, varies from very small to high over the range of uptake
2-6
-------
coefficients (y) for N2Os obtained in laboratory studies. For example, global Os loss
ranges from 0 to over 10%, with large regional variability over the range of reported
N2Os uptake coefficients. However, uptake coefficients for^Os [yCNbOs)] on
atmospheric particles are not well defined, largely due to uncertainty and variability in
aerosol composition. As noted by Brown and Stutz (2012). yflSbOs) is largest («0.02) for
aqueous inorganic aerosols and water droplets, except for nitrate in aerosol, which can
reduce y^Os) by up to an order of magnitude. The uptake of ^Os by mineral particles
could also represent an important removal process. For example, values of y(N2Os) for
calcite and Saharan dust are about 0.03. However, as noted by Tang et al. (2014) not
enough is known to permit a global assessment of the importance of ^Os uptake on
mineral surfaces. Organic aerosol and soot can reduce y(N2Os) by two orders of
magnitude or more, further complicating the task of assessing the importance of uptake of
N2Os on aerosol surfaces.
The uptake of N2Os by aqueous aerosols containing chloride (Cl~) and bromide (Br ) has
been associated with the release of gaseous CINCh from marine aerosol [sea-spray;
(Osthoff et al., 2008)1. CINCh has been found not only in coastal and marine
environments, but also well inland. For example, Thornton et al. (2010) found production
rates of gaseous C1NO2 near Boulder, CO from reaction of ^Os with particulate Cl~ at
levels similar to those found in coastal and marine environments. They also found that
substantial quantities of N2Os are recycled through C1NO2 back into NOx instead of
forming HNOs. C1NO2 readily photolyzes to yield Cl and NO2 and can represent a
significant source of reactive Cl, capable of initiating the oxidation of hydrocarbons
(generally with much higher rate coefficients than OH radicals). Riedeletal. (2014)
found increases in the production of radicals by 27% and of Os by 15% during the 2010
CalNex [California Research at the Nexus of Air Quality and Climate Change in May to
June 2010 in Southern California; (Ryerson et al.. 2013)] field study. However, C1NO2
was found to cause only modest Os increases (e.g., ~1 to 1.5 ppb for nominal Os
concentrations between 60 and 85 ppb) in a model study of the Houston, TX airshed
(Simon et al., 2009). Differences are likely related to differences in the NOx sensitivity of
the two airsheds. Therefore, caution is advised in extrapolating results obtained in one
airshed to another.
As mentioned earlier, NO and NO2 are important precursors of Os formation. However,
because Os changes in a nonlinear way with changes in the concentrations of its
precursors (NOx and VOCs), Os is unlike many other atmospheric species with rates of
formation that vary directly with emissions of their precursors. At the low NOx
concentrations found in environments ranging from remote continental areas to rural and
suburban areas downwind of urban centers, the net production of Os typically increases
with increasing NOx. In this low-NOx regime, the overall effect of the oxidation of
2-7
-------
VOCs is to generate (or at least not consume) radicals, and Os production varies directly
with NOx. In the high-NOx regime, NC>2 reacts with OH radicals to form HNOs [e.g.,
Hameed et al. (1979)1. Otherwise, these OH radicals would oxidize VOCs to produce
peroxy radicals, which in turn would oxidize NO to NO2. In this regime, Os production is
limited by the availability of radicals (Tonnesen and Jeffries. 1994). and Os shows only a
weak dependence on NOx concentrations. Reaction of Os with NO in fresh motor vehicle
exhaust depletes Os in urban cores, but Os can be regenerated during transport downwind
of urban source areas, and additional chemical production of Os can occur, resulting in
higher Os concentrations than found upwind of the urban center. Similar depletion of Os
can occur in power plant plumes with subsequent Os regeneration downwind.
Brown etal. (2012) conducted a field study comparing nighttime chemistry in the plumes
of two power plants in Texas, one with selective catalytic reduction (SCR) NOx
emissions controls and the other without these controls. They noted that the plume from
the power plant with SCR controls did not have enough NOx to deplete all of the Os
present in background air. As a result, almost all of the NOx in the plume was oxidized to
NOz species. This situation contrasts with that in the plume from the power plant without
controls. In that plume, there was minimal formation of NOz species. Instead, NOx was
more nearly conserved.
2.3 Sources
2.3.1 Overview
Estimated total NOx emissions in the United States (U.S.) from all sources decreased by
49% over the period from 1990 to 2013, as shown in Figure 2-2. The NEI is a national
compilation of emissions sources collected from state, local, and tribal air agencies as
well as emission estimates developed by the U.S. Environmental Protection Agency
(EPA) from collected or estimated data by source sector. Emissions after 2011 for mobile
sources and electric utilities are regularly added to the 2011 NEI, but emissions for the
other sectors are based on 2011 estimates. Through this process, some of the major
sectors in the 2011 NEI have emission estimates more recent than 2011, while emissions
from other source sectors are based on 2011 data. When emissions from these sources are
added for later years, the inventory is still referred to as a version of the 2011 NEI.
2-8
-------
25 -
-£• 20 -
c
o
15 -
Ul
g
LJJ
10 -
5 -
Total NOV Emissions
ID
U3
O
M
O
o
o
M
o
o
M
o
o
NJ
M
O
o
LO
K>
O
O
M M
O O
O O
t_n CTi
NJ M
O O
O O
-«J 00
M
O
o
o
I—'
o
M
O
M
O
o
h-»
OJ
Year
Source: National Center for Environmental Assessment 2014 analysis of 2011 National Emissions Inventory data (U.S. EPA.
201 3a).
Figure 2-2 U.S. national average NOx (sum of nitrogen dioxide and nitric
oxide) emissions from 1990 to 2013.
The NEI program develops data sets, blends data from multiple sources, and performs
quality assurance steps that further enhance and augment the compiled data. The
inventory database does not include sector emissions uncertainty estimates. The accuracy
of individual emission estimates may vary from facility to facility or county to county,
and for some sources, data may be incomplete or lacking. For example, there is no
lightning data in the NEI, and the 2008 NEI for oil and gas was incomplete, although an
oil and gas production estimation tool was developed for subsequent inventories. While
uncertainties are difficult to predict, the NEI undergoes continuous improvement by the
U.S. EPA with the assistance of state, local, and tribal agencies by their reporting
emissions information for facilities, other stationary sources, and mobile sources. Each
3-year cycle of NEI development incorporates improvements based on lessons learned
from the previous cycles, and estimation procedures for emissions sectors typically
2-9
-------
evolve over time in response to identified deficiencies as the data are used. As a result, in
spite of inexact and potentially unknown uncertainties, the NEI largely meets the needs
for general emissions assessments and national trends reporting. For example, NOx data
from the NEI has done a reasonable job of predicting ozone concentrations, resulting in
decision making that has significantly improved air quality over the years.
The major sources of NOX in the U.S. identified from the 2008 and 2011 NEI (U.S. EPA.
2013a. 201 la) are described in Figure 2-3. The values shown are U.S. nationwide
averages and may not reflect the mix of sources relevant to individual exposure in
populated areas. For most sources, data are generally available for all 50 states and the
District of Columbia (in some cases, such as agricultural burning, data available in the
NEI exclude Alaska and Hawaii). Biogenic emissions were estimated using 2011
meteorology and land use information using the Biogenic Emission Inventory System,
version 3.14 [BEIS (Biogenic Emission Inventory System)3.14] model. Although the
BEIS domain includes Canada and Mexico, the NEI uses BEIS estimates from counties
that make up the contiguous 48 states.
2-10
-------
Highway Vehicles
Off-Highway Vehicles & Engines
Fuel Combustion-Utilities
Fuel Combustion-Other
Other Anthropogenic
Biogenics and Wildfires
0
2 4 6 8 10
Millions of Tons
Source: National Center for Environmental Assessment 2014 analysis of 2011 National Emissions Inventory data (U.S. EPA. 2013a.
2011 a).
Figure 2-3 Major sources of NOx (sum of nitrogen dioxide and nitric oxide)
emissions averaged over the U.S. from the 2008 and 2011
National Emissions Inventories.
The source categories displayed in Figure 2-3 represent groups of similar NEI source
sectors. Highway Vehicles include all on-road vehicles, including light-duty as well as
heavy-duty vehicles, both gasoline- and diesel-powered. Off-Highway Vehicles and
Engines include aircraft, commercial marine vessels, locomotives, and nonroad
equipment. Fuel Combustion-Utilities includes electric power generating units (EGUs). It
includes all types of fuels, but is dominated by coal combustion, which accounts for 85%
2-11
-------
of all NOx emissions from utilities in the 2011 NEI. Fuel Combustion-Other includes
commercial/institutional, industrial, and residential combustion of biomass, coal, natural
gas, oil, and other fuels. Other Anthropogenic sources include field burning, prescribed
fires, and various industrial processes (e.g., cement manufacturing, oil and gas
production). On a national scale, field burning and prescribed fires are the greatest
contributors to the Other Anthropogenic sources category. Biogenics and Wildfires
include NEI emission estimates for biogenic (plant and soil) emissions and wildfires. For
NOx, biogenic emissions are dominated by soil emissions, which are one to two orders of
magnitude greater than vegetation emissions.
Highway Vehicles are the largest source in the 2011 NEI, contributing 37% of the total
NOx emissions. Off-Highway Vehicles and Engines account for 20% of emissions, Fuel
Combustion-Utilities (by EGUs) for 14%, Fuel Combustion-Other for 11%, Other
Anthropogenic sources for 10%, and Biogenics and Wildfires for 8% of 2011 NEI
national emissions of NOx. Nationwide estimates of total NOx emissions in the 2011 NEI
are 13% lower than 2008 NEI estimates, decreasing from 18.0 megatons to
15.6 megatons. This decrease reflects lower emission estimates in the 2011 NEI than in
the 2008 NEI for the four largest categories in Figure 2-3: 17% lower for Highway
Vehicles, 10% lower for Off-Highway Vehicles and Engines, 33% lower for Fuel
Combustion-Utilities, and 6% lower for Fuel Combustion-Other. However, estimated
emissions were 17% higher for Other Anthropogenic sources, with the greatest increases
observed for oil and gas production, agricultural field burning, prescribed fires, and
mining. Although Biogenics and Wildfire emissions have increased as a proportion of
total national emissions, Anthropogenic sources (i.e., the other categories) still account
for more than 90% of emissions in the 2011 NEI.
A somewhat different source mixture than the U.S. national average occurs in the most
populated areas. Figure 2-4 compares contributions from different groups of sources in
the 21 core-based statistical areas (CBSAs) of the U.S. with populations greater than
2.5 million, where 39% of the U.S. population lives. Relative to the national average, the
urban areas have greater contributions to total NOx emissions from both Highway
Vehicle emissions and Off-Highway Vehicle and Engine emissions, and smaller
contributions from Fuel Combustion-Utilities (EGUs), Other Anthropogenic emissions,
and Biogenics and Wildfires. Table 2-1 provides details on source distributions for
individual CBSAs.
2-12
-------
LJ- U U' 1
1 r* U IU.-IM.-
Other Anthropogenic ^HI '
0% 10% 2C
• N<
)% 3C
ational
)% 4C
• Urban
)% 5C
)% 6C
%
Source: National Center for Environmental Assessment 2014 analysis of 2011 National Emissions Inventory data (U.S. EPA.
2013a).
Figure 2-4 Percentage contributions from major sources of the annual NOx
(sum of nitrogen dioxide and nitric oxide) emissions averaged
over the 21 largest U.S. core-based statistical areas with
populations greater than 2.5 million compared to the national
average.
2-13
-------
Table 2-1 Source distribution of the annual NOx (sum of nitrogen dioxide and
nitric oxide) emissions in the 21 largest U.S. core-based statistical
areas with populations greater than 2.5 million— 2011 National
Emissions Inventory.
Highway Off-Highway
Vehicles Vehicles and
(%) Engines (%)
New York, NY
Los Angeles, CA
Chicago, IL
Dallas, TX
Houston, TX
Philadelphia, PA
Washington, DC
Miami, FL
Atlanta, GA
Boston, MA
San Francisco, CA
Riverside, CA
Phoenix, AZ
Detroit, Ml
Seattle, WA
Minneapolis, MN
San Diego, CA
Tampa, FL
St. Louis, MO
Baltimore, MD
Denver, CO
Urban Average
44.3
59.6
40.1
53.9
46.0
51.2
55.2
50.5
57.7
46.4
52.8
53.0
64.0
52.5
66.5
50.2
64.3
52.7
56.9
48.5
55.6
51.9
27.6
26.2
27.1
21.5
25.6
22.5
23.4
32.2
19.6
27.3
30.6
20.8
25.9
19.4
25.7
19.4
25.5
23.2
13.6
21.4
22.5
23.8
Biogenics
Fuel and
Utilities Combustion- Other Anthro- Wildfires
(%) Other (%) pogenic(%) (%)
4.4
0.6
11.1
1.2
3.0
4.0
7.8
7.6
15.1
2.7
0.8
1.6
1.7
8.4
0.1
11.3
0.6
13.5
15.3
11.5
3.2
6.4
21.6
10.4
14.8
9.2
9.5
14.7
10.9
4.8
5.7
18.7
9.7
8.5
5.2
14.4
5.1
13.9
4.3
5.6
6.7
11.0
12.2
11.6
1.8
2.5
5.4
9.5
11.4
6.7
1.5
2.7
1.1
4.6
4.9
12.4
0.6
4.1
2.3
3.3
0.8
2.1
5.2
6.6
2.9
4.6
0.4
0.7
1.4
4.7
4.5
0.9
1.3
2.2
0.9
0.4
1.2
3.7
2.7
1.2
0.4
2.0
4.6
2.9
2.4
1.0
3.6
1.8
NY = New York; CA = California; IL = Illinois; TX = Texas; PA = Pennsylvania; FL = Florida; GA = Georgia; MA = Massachussets;
AZ = Arizona; Ml = Michigan; WA = Washington; MN = Minnesota; MO = Missouri; MD = Maryland; CO = Colorado.
Source: National Center for Environmental Assessment 2014 analysis of 201 1 National Emissions Inventory data (U.S. EPA,
2013a).
2-14
-------
2.3.2 Highway Vehicles
Nationally, Highway Vehicles account for about 37% of NOx emissions, according to the
2011 NEI. In the 21 largest CBSAs in the U.S. represented in Figure 2-4. more than half
of the urban NOx emissions are from Highway Vehicles, ranging from 40% in Chicago,
IL to 67% in Seattle, WA. Together, Highway Vehicles and Off-Highway Vehicles and
Engines account for more than three-quarters of total emissions. Other estimates of high
contributions from Highway Vehicles have also been reported. For example, on-road
vehicles were estimated to account for about 80% of anthropogenic NOx concentrations
in the Los Angeles, CA area (McDonald et al.. 2012) and 72% in the Atlanta, GA area
(Pachon et al.. 2012). Highway Vehicle NOx emissions nationwide are roughly equally
split between light-duty gasoline engines (48%) and heavy-duty diesel engines (46%),
according to the 2011 NEI. This is in spite of a national vehicle fleet distribution of more
than 230 million mostly gasoline-powered light-duty vehicles compared to only
10 million mostly diesel-powered heavy-duty vehicles.1 McDonald et al. (2012)
estimated that diesel engines were the dominant on-road NOx sources in the San Joaquin
Valley, CA, accounting for up to 70% of on-road NOx emissions. In contrast in Fulton
County, GA it was estimated that 60% of on-road NOx emissions were from gasoline
vehicles and 40% from diesel (Pachon etal.. 2012). McDonald et al. (2012) estimated
that in California, gasoline engine-related NOx emissions steadily decreased by 65% over
the period from 1990 to 2010. The study authors also found that the ratio of NOx
emission factors for heavy-duty diesel versus light-duty gasoline engines grew from ~3 to
~8 between 1990 and 2010 due to improved effectiveness of catalytic converters on
gasoline engines.
However, NOx emissions from on-road diesel engines in the U.S. have also decreased
substantially due to stricter emission standards, and emissions continue to decline
(McDonald et al.. 2012). Emission standards for heavy-duty diesel trucks were first
established at 10.7 g/bhp-h in 1988, and the current standard of 0.20 g/bhp-h was
gradually phased in for model years 2007 through 2010 (U.S. EPA. 2001). so that
emission standards from heavy-duty diesel trucks were reduced by more than a factor of
50 between 1988 and 2010. The current standard is achieved using a urea-based SCR
catalyst in engine exhaust placed downstream of a diesel oxidation catalyst (DOC) and a
catalyzed diesel particulate filter (DPF) used for PM emissions control. In extensive
testing of diesel engines, substantial reductions in NOx were observed, averaging 61%
relative to the 2010 standard requirements and 97% relative to the 2004 standard
requirements (Southwest Research Institute. 2013). However, while total diesel NOx
emissions have substantially decreased because of urea-based SCR control, the NO2/NOx
https://www.fhwa.dot.gov/policvinformation/statistics/2010/vml .cfm.
2-15
-------
ratio has increased. But these reductions for diesel emissions together with the recent
final Tier 3 rule for gasoline engine emissions and lower S 41 gasoline (U.S. EPA.
2014a) are likely to result in a substantial decline in NOx emissions as newer vehicles
penetrate into the on-road fleet over the next several years.
2.3.3 Off-Highway Vehicles and Engines
Off-Highway Vehicles and Engines constitute the next largest group of NOx emission
sources after Highway Vehicles, both on a nationwide basis and in large U.S. urban
CBSAs as shown in Figure 2-4 and Table 2-1. Emissions from the nonroad source sector
can also significantly contribute to local and national air quality. The 2011 NEI estimated
that approximately 20% of nationwide NOx was from Off-Highway Vehicles and
Engines. Zhu et al. (2011) estimated that nonroad diesel engines contribute 12% of total
nationwide NOx emissions from mobile sources. Off-Highway Vehicle and Engine
sources include aviation, marine, and railroad engines, as well as nonroad agricultural
and industrial equipment, all of which emit NOx through combustion processes.
Examples of nonroad equipment include farm tractors, excavators, bulldozers, and wheel
loaders. Nationally, agricultural and industrial equipment accounts for more than half of
Off-Highway Vehicle and Engine NOx emissions, mostly from diesel-powered
equipment (U.S. EPA. 2013a). The U.S. EPA has set a series of standards to reduce NOx
emissions from nonroad diesel, referred to as Tier 1-4 standards. The most recent
standard, Tier 4, was introduced in May 2004, and the phase-in is currently underway,
covering a time period between 2008 and 2015. In most cases, advanced diesel engine
design, exhaust gas recirculation, and/or SCR have been used to comply with these
standards, with DOC/DPFs used in several engine categories.
Although Fuel Combustion-Utilities is generally a smaller contributor to total NOx in
urban areas than it is nationally, emergency generators are an emerging concern. In urban
areas, emissions of NOx have been observed to increase substantially on days of near
peak electricity demand because of small natural gas- and petroleum-powered steam
turbines used to generate additional electrical power to meet demand. These generators
are classified in the NEI as nonroad equipment that fall into the category of Off-Highway
Vehicles and Engines. They are typically operated in densely populated areas. They are
usually older units with higher emissions and lower stack heights than larger generators
and are often located close to residential neighborhoods. Because of these factors,
emergency generators can have substantial impacts on local air quality. For example,
Gilbraith and Powers (2013) estimated that reducing emissions from emergency
generators could decrease NOx emissions in New York, NY alone by 70 tons per year.
2-16
-------
Aircraft, commercial marine transport, and locomotive emissions account for the
remaining 40% of Off-Highway Vehicle and Engine emissions, nationally. Aircraft
includes all aircraft types used for public, private, and military purposes, classified into
four types: commercial, air taxis, general aviation, and military. Airport-related NOx
emissions can significantly impact local and regional air quality. In the U.K., within a
2-3-km radius of London Heathrow Airport, Carslaw et al. (2006) reported that airport
emissions can comprise up to 15% of total ambient NOx. In Atlanta, GA, Unal et al.
(2005) showed that roughly 2.6% of regional NOx concentrations can be attributed to
emissions from activities at Hartfield-Jackson International Airport. Compared to
airport-related emissions of other gaseous pollutants [e.g., ammonia (NH3), CO, sulfur
dioxide (802), VOCs], airport NOx emissions had the largest contribution to decreased
regional air quality in Atlanta, GA.
Commercial marine vessels include boats and ships used either directly or indirectly in
the conduct of commerce or military activity. Globally, marine transport is a significant
source of NOx emissions, accounting for more than 14% of all global nitrogen emissions
from fossil fuel combustion [mostly NOx; (Corbett et al.. 1999)1. On a regional scale, the
contribution of shipping emissions to total NOx emissions is variable and can be a
substantial fraction near port cities (KimetaL 2011; Williams et al.. 2009; Vutukuru and
Dabdub. 2008). In Los Angeles, CA, Vutukuru and Dabdub (2008) estimated that
commercial shipping contributed 4.2% to total NOx emissions in 2002. Using the
NEI-05, Kim et al. (2011) estimated that roughly 50% of NOx concentration near the
Houston Ship Channel is associated with commercial shipping emissions. However, this
estimation is much higher than observed in satellite and aircraft measurements.
Locomotives powered by diesel engines are a source of NOx emissions. Using a
fuel-based approach to quantify emissions, Dallmann and Harley (2010) estimated that
diesel locomotives emitted on average 50% of total NOx from all nonroad mobile sources
and roughly 10% of total NOx from all mobile sources in the U.S. during 1996-2006
(Dallmann and Harley. 2010). Locomotives can comprise a much larger fraction of NOx
emissions for areas in or near large rail yard facilities (>90% of emissions), including
NO2 nonattainment areas (U.S. EPA. 2010a). In a year-long study at the Rougemere Rail
Yard facility near Dearborn, MI, 98% of NOx emissions was attributed to locomotive
operation, with only minimal impacts from other sources such as on-road mobile sources
and stationary sources (U.S. EPA. 2009a). Cahill et al. (2011) measured gaseous and PM
pollutants during a 5-week period near the Roseville Rail Yard in Placer County, CA.
They observed several transient NOx emission events, where NO levels between 200 ppb
and 500 ppb, or roughly seven times larger than the observed urban background NO,
were observed downwind of the Roseville Rail Yard.
2-17
-------
2.3.4 Fuel Combustion—Utilities and Other
Fuel combustion for electric power generation and for industrial, residential, commercial,
and institutional purposes (excluding motor vehicles and nonroad equipment) accounts
for about 25% of NOx emissions nationwide. As indicated in Figure 2-3. Fuel
Combustion-Utilities accounts for about 14% of total NOx emissions nationally.
Nationally, about 85% of the NOx emissions from power generation is from coal
combustion. In urban areas, as shown in Figure 2-4. and fuel combustion for purposes
other than electric power generation (Fuel Combustion-Other) appears to be a greater
source of emissions than Fuel Combustion-Utilities.
In contrast to Fuel Combustion-Utilities, coal accounts for only about 1% of Fuel
Combustion-Other emissions. However, Fuel Combustion-Other is still dominated by
fossil fuels, with natural gas contributing about 68% and oil combustion contributing
about 14% of other fuel combustion emissions. Although biofuels are an important NOx
source globally (Jaegle et al.. 2005). only about 10% of Fuel Combustion-Other
emissions in the U.S. are due to biomass burning. For Fuel Combustion-Utilities and Fuel
Combustion-Other combined, fossil fuels account for more than 90% of U.S. stationary
source fuel combustion, and biomass only 4%. Combustion of biofuels accounts for only
about 1% of total NOx emissions nationwide.
Fuel Combustion-Other accounts for an additional 12% of urban NOx emissions, but
ranges as high as 22% in New York, NY and 19% in Boston, MA as shown in Table 2-1.
Figure 2-5 shows that the contribution of Fuel Combustion-Other to overall urban NOx
emissions varies with average January temperatures. This trend suggests that winter
heating is the driving factor for Fuel Combustion-Other emissions, and that in winter the
Fuel Combustion-Other contribution is likely to be considerably greater than the
contribution presented on an annual basis in Table 2-1. possibly rivaling Highway
Vehicle emissions in winter.
2-18
-------
Q.
O
Ol
6000
5000
4000
£ 3000
(U
Q.
cf 2000
l/l
c
o
1000
^- Min
Chicago
B
neapolis
„ Del
•
;
oston -1
Denver
Wash
troit
^-StLo
r- Phila
I Balti
»
ington
X
A1
uis
delphia
more
Dallas
,
lanta Ph
i- Housl
.-
oenix
ton
San Fran
^ Los Ange
^^Tam
•-
San Die
cisco
•les
pa
— Miami
5°
10 20 30 40 50 60
Average January Temperature (°F)
80
Source: National Center for Environmental Assessment 2014 analysis of 2011 National Emissions Inventory data (U.S. EPA.
2013a).
Figure 2-5 Fuel Combustion-Other emissions of NOx (sum of nitrogen
dioxide and nitric oxide) versus average ambient January
temperature for the 21 largest U.S. core-based statistical areas
with populations greater than 2.5 million.
2.3.5 Other Anthropogenic Sources
Other Anthropogenic sources include prescribed and agricultural fires as well as
industrial operations such as oil and gas production and mining. As emissions estimates
from other major source categories have decreased in the U.S. between 2008 and 2011,
emissions from these sources have increased by 17%, from about 1.4 megatons in 2008 to
more than 1.6 megatons in 2011. On a national scale, agricultural burning and prescribed
fires are responsible for a large fraction of the Other Anthropogenic sources category and
the increase in national emissions for Other Anthropogenic sources between 2008 and
2011. However, in urban areas, fires are less of a contributor, and Other Anthropogenic
sources are mainly industrial. Other Anthropogenic sources vary considerably among the
2-19
-------
21 largest U.S. CBSAs with populations greater than 2.5 million. In three CBSAs as
described below, NOx emissions from Other Anthropogenic sources exceed 10,000 tons
per year. Emissions in these three CBSAs are separated by industrial sector in Table 2-2.
In Chicago, IL, emissions from several different sources contribute to Other
Anthropogenic emissions. In contrast, Other Anthropogenic NOx emissions in Dallas, TX
are dominated by oil and gas production, which is not an important source in Chicago, IL.
The oil and gas production sector is an increasing source of NOx, with a 2011 emission
estimate of more than 600,000 tons, compared to slightly more than 400,000 tons in
2008. Pacsi et al. (2013) estimated that routine operating activities from the Barnett Shale
production facility near Dallas, TX can emit roughly 30 to 46 tons NOx/day, depending
on the demand for natural gas electricity generation. Nonroutine gas flares can also result
in episodic peaks of large NOx emissions, affecting local air quality (Olaguer. 2012).
Houston, TX presents yet another variation, with anthropogenic emissions mainly coming
from petroleum refining and chemical manufacturing. These data demonstrate that
sources with relatively small nationwide or annual emissions may contribute substantially
to emissions on a local scale. For example, cement manufacturing, which is listed in
Table 2-2 as an important source in the local Dallas, TX, airshed, accounts for less than
1% of annual national emissions, but has been characterized by variable emissions with
high peaks (Walters et al.. 1999).
2-20
-------
Table 2-2 Relative contributions to Other Anthropogenic NOx (sum of nitrogen
dioxide and nitric oxide) sources in selected U.S. cities.3
Chicago, IL Dallas, TX Houston, TX
Bulk gasoline terminals <1 <1 <1
Fires — agricultural field burning
Fires — prescribed fires
Fires — wildfires
Gas stations
Industrial processes — cement manufacturing
Industrial processes — chemical manufacturing
Industrial processes — ferrous metals
Industrial processes — not elsewhere classified
Industrial processes — nonferrous metals
Industrial processes — oil & gas production
Industrial processes — petroleum refineries
Industrial processes — pulp & paper
<1 1 3
1 NRb 1
<1 4 <1
NRb <1 <1
NRb 19 NRb
8 <1 43
9 2 <1
35 6 3
21 <1 <1
0 66 9
13 <1 37
<1 <1 NRb
Industrial processes — storage and transfer <1 <1 <1
Miscellaneous nonindustrial not elsewhere classified <1 <1 <1
Solvent — degreasing
NRb <1 <1
Solvent — graphic arts <1 <1 <1
Industrial surface coating & solvent use
Waste disposal
Total
1 <1 1
11 1 4
100 100 100
aNOx (sum of NO and NO2) emissions as percent of "Other Anthropogenic sources" emissions in the Core-Based Statistical Area.
bNR indicates that no emissions were reported for this sector (i.e., there were no sources with emissions above the reporting
threshold).
Source: National Center for Environmental Assessment 2014 analysis of 2011 National Emissions Inventory data (U.S. EPA.
2013a).
2-21
-------
2.3.6 Biogenics and Wildfires
The NEFs Biogenics sector includes emissions from plants and soil. In the case of NOx,
biogenic emissions are dominated by emissions from soil. Biogenic emissions account for
about 6% of total NOx emissions in the 2011 NEI. However, spatial and temporal
variability in NOx emissions from soil leads to considerable variability in biogenic
emission estimates. For example, estimates obtained from satellite observations indicated
that 15-40% of the total NO2 column in various locations over the Great Plains region
can be attributed to soil emissions in spring and summer months (Hudman et al.. 2010).
This is consistent with geographic differences in soil contributions described in the 2011
NEI, in which soil contributions accounted for 13-34% of NOx emissions in Iowa,
Kansas, Nebraska, North Dakota, and South Dakota. About 60% of the total NOx emitted
from soils is estimated to occur in the central corn belt of the U.S. Because of low
population density and the wide area over which emissions are distributed, soil emissions
are a less important concern for exposure than more concentrated sources in more highly
populated areas.
Biogenic emissions for the 2011 NEI were computed based on the BEIS model. The
BEIS modeling domain includes the contiguous 48 states in the U.S., parts of Mexico,
and Canada. The NEI uses the biogenic emissions from counties from the contiguous
48 states and DC. Both nitrifying and denitrifying organisms in the soil can produce
NOx, mainly in the form of NO. Emission rates depend mainly on the amount of applied
fertilizer, soil temperature, and soil moisture. As a result, a high degree of uncertainty is
associated with soil emissions, and estimates obtained from satellite observations can be
greater than source-based estimates (Jaegle et al.. 2005).
Emissions from wildfires can produce enough NOx to cause local and regional
degradation of air quality in some regions (Pfister et al.. 2008). Roughly 15% of global
NOx emissions are from biomass burning (Penman et al., 2007). Burling et al. (2010)
reported that NOx emissions from southwestern U.S. vegetation ranged from 2.3 to
5.1 g/kg, with the majority of the NOx present as NO. Emissions vary considerably
among different species of biota, making it difficult to estimate emissions for key
ecosystems, such as extratropical forests (McMeeking et al.. 2009). Emissions from
forest wildfires can be more than double per amount of energy released than for shrub
wildfires (Mebust et al.. 2011).
2-22
-------
2.3.7 Emissions Summary
Major categories of NOx emissions in the U.S. are Highway Vehicles, Off-Highway
Vehicles and Engines, Fuel Combustion-Utilities, Fuel Combustion-Other, Other
Anthropogenic emissions, and Biogenics and Wildfire emissions. Of these,
Fuel-Combustion-Utilities and Biogenics and Wildfire emissions are less important in
populated U.S. urban areas with the highest NO2 concentrations, and thus, potentially
have less impact on human exposure to NO2. Instead, in urban areas, emissions are
generally dominated by Highway Vehicles and Off-Highway Vehicles and Engines,
which make up more than three-quarters of emissions in the 21 largest CBSAs with
populations greater than 2.5 million. Other sources can make important contributions. For
example, in cities with average January temperatures below freezing, NOx emissions
from Fuel Combustion-Other can also be important, and episodic emissions from Other
Anthropogenic sources can be important locally. Advances in emission control standards
and technology have led to substantial reduction in NOx emissions from Highway
Vehicles, and hold promise for further reductions. However, Highway Vehicles is
generally the greatest source of NOx emissions in urban areas.
2.4 Measurement Methods
2.4.1 Federal Reference and Equivalent Methods
This discussion focuses on current methods and on promising new technologies for
measuring oxides of nitrogen. No attempt is made here to cover in detail the development
of these methods, or of methods such as wet chemical techniques, which are no longer in
use. More detailed discussions of the histories of these methods can be found elsewhere
(U.S. EPA. 1996a. 1993a).
NO is routinely measured using the chemiluminescence induced by its reaction with Os at
low pressure. The Federal Reference Method (FRM) for NO2 makes use of this technique
of NO detection with a prerequisite step that is meant to reduce NO2 to NO on the surface
of a molybdenum oxide (MoOx) substrate heated to between 300 and 400°C. On June 1,
2012, an automated Federal Equivalent Method (FEM) for measuring NO2 using a
photolytic converter to reduce NO2 to NO met the equivalency specifications outlined in
40 Code of Federal Regulations (CFR) Part 53 and was approved by the U.S. EPA
(2012a). Although photolytic converters have lower conversion efficiencies than
FRM-based analyzers, they have been found to be stable over a period of at least two
2-23
-------
months (Pollack et al.. 2011). In addition, two monitors using cavity attenuated phase
shift (CAPS) spectroscopy have been approved more recently as FEMs (U.S. EPA.
2014d. 2013b). These techniques are described below.
Because the chemiluminescence based FRM cannot detect NO2 specifically, the
concentration of NC>2 is determined as the difference between the NO in the air stream
passed over the heated MoOx substrate and the NO in the air stream that has not passed
over the substrate.
However, the reduction of NO2 to NO on the MoOx catalyst substrate also reduces other
oxidized nitrogen compounds that are present in the sample (i.e., NOz compounds shown
in the outer box of Figure 2-1) to NO. This interference by NOz compounds has long
been recognized following Winer et al. (1974) who found over 90% conversion of PAN,
ethyl nitrate, ethyl nitrite, and n-propyl nitrate and 6-7% conversion of nitroethane to NO
with a MoOx converter. FINOs produced a response, but its form could not be
determined. As a result of their experiments, Winer et al. (1974) concluded that "the NOx
mode of commercial chemiluminescent analyzers must be viewed to a good
approximation as measuring total gas phase 'oxides of nitrogen,' not simply the sum of
NO and NO2." Numerous later studies have confirmed these results (Dunlea etal.. 2007;
Steinbacher et al.. 2007; U.S. EPA. 2006; McClenny et al.. 2002; Parrish and Fehsenfeld.
2000: Nunnermacker et al.. 1998: Croslev. 1996: U.S. EPA. 1993a: Rodgers and Davis.
1989: Fehsenfeld et al.. 1987). The sensitivity of the FRM to potential interference by
individual NOz compounds was found to be variable, depending on characteristics of
individual monitors, such as the design of the instrument inlet, the temperature and
composition of the reducing substrate, and on the interactions of atmospheric species
with the reducing substrate.
Only recently have attempts been made to systematically quantify the magnitude and
variability of the interference by NOz species in ambient measurements of NO2. Dunlea
et al. (2007) found an average of about 22% of ambient NO2 (~9 to 50 ppb), measured in
Mexico City over a 5-week period during the spring of 2004, was due to interference
from NOz compounds. However, similar comparisons have not been carried out under
conditions typical for State and Local Air Monitoring Stations (SLAMS) monitoring sites
in the U.S. Dunlea et al. (2007) compared NO2 measured using the conventional
chemiluminescent instrument with other (optical) techniques. The main sources of
interference were FINOs and various organic nitrates. Efficiency of conversion was
estimated to be -38% for HNOs and -95% for PAN and other organic nitrates. Peak
interference of up to 50% was found during afternoon hours and was associated with Os
and NOz compounds, such as FINOs and the alkyl and multifunctional alkyl nitrates.
2-24
-------
Lamsal et al. (2008) used data for the efficiency of reduction of NOz species on the
MoOx catalytic converters to estimate seasonal correction factors for NO2 measurements
across the entire U.S. These factors range from <10% in winter to >80%, with the highest
values found during summer in relatively unpopulated areas. In general, interference by
NOz species in the measurement of NO2 is expected to be larger downwind of urban
source areas and in relatively remote areas due to the conversion of NO2 to NOz during
transport downwind of source areas.
In a study in rural Switzerland, Steinbacher et al. (2007) compared continuous
measurements of NO2 from a chemiluminescence analyzer with a MoOx catalytic
converter (CL/MC) with measurements from a chemiluminescence analyzer with a
photolytic converter (CL/PC) that reduces NO2 to NO. They found the conventional
technique using catalytic reduction (as in the FRM) overestimated the measured NO2
compared to the photolytic technique on average by 10% during winter and 50% during
summer.
Villena et al. (2012) and Kleffmann et al. (2013) suggested that negative interference in
the chemiluminescent method using the photolytic converter could occur from production
of HO2 and RO2 radicals by the photolysis of VOCs (e.g., glyoxal) in the photolytic
converter. Subsequent to photolysis and prior to detection, these radicals react with NO
that is produced either by the photolytic converter or already in the sampling stream.
Because the chemiluminescent techniques rely on detection of NO, a negative artifact
results. The most direct evidence for this artifact was found at high concentrations in a
smog chamber containing 1 ppm glyoxal, a concentration more than a thousand times
higher than typically found in ambient air. Similar indications were also found by
Kleffmann et al. (2013) in a street canyon (at the University of Wuppertal, Germany) and
in an urban background environment (University of Santiago, Chile). However,
Kleffmann et al. (2013) also found that the magnitude of the negative artifact is smaller
when a light source with a smaller spectral range is used and that this artifact is expected
to be most apparent under high VOC conditions, such as in street canyons.
Within the urban core of metropolitan areas, where many of the ambient monitors are
sited in areas influenced by strong NOx sources such as motor vehicles on busy streets
and highways (i.e., where NO2 concentrations are highest), the positive artifacts due to
the NO2 oxidation products are much smaller on a relative basis. Conversely, the positive
artifacts are larger on a relative basis away from NOx sources. Data for PAN and HNOs
were collected in Houston, TX in April and May of 2009 during the Study of Houston
Atmospheric Radical Precursors (SHARP) campaign (Olaguer et al.. 2014). Median
concentrations of PAN and HNOs during the afternoon were 181 [interquartile range
(IQR) 94] parts per trillion (ppt) and 164 (IQR 158) ppt, respectively, forNO2 <1 ppb
2-25
-------
measured by CL/PC during SHARP and 157 (IQR 54) ppt and 146 (IQR 402) ppt,
respectively, for NC>2 >10 ppb. These results suggest that potential interference in CL/MC
caused by HNOs and PAN is estimated to be <1 ppb using the conversion efficiencies
obtained by Dunlea et al. (2007) and concentrations of HNOs and PAN obtained during
SHARP. However, the extent of interference could be expected to be most problematic
forNO2<~lppb.
In summary, the current FRM for determining ambient NOx concentrations and then
reporting NCh concentrations by subtraction of NO is subject to a consistently positive
interference by NOx oxidation products, including HNOs, PAN and its analogues, and
total organic nitrates (RONO2). The magnitude of this positive bias is largely unknown as
measurements of these oxidation products in urban areas are sparse. However, it is likely
to be less important in urban areas influenced by fresh NOx combustion emissions than in
remote areas where NOx oxidation has had more time to proceed.
2.4.2 Other Methods for Measuring Nitrogen Dioxide
Optical methods such as those using differential optical absorption spectroscopy (DOAS)
or laser-induced fluorescence (LIF) are available for use in ambient monitoring.
However, these particular methods, even those that have been commercialized (e.g.,
DOAS), can be more expensive than either the FRM monitors or photolytic reduction
technique and require specialized expertise to operate; moreover, the DOAS obtains a
path-integrated rather than a point measurement. Cavity attenuated phase shift (CAPS)
monitors are an alternative optical approach requiring much less user intervention and
expense than either DOAS or LIF (Kebabian et al.. 2008). At first glance, it might appear
that this technique is not highly specific to NO2, as it is subject to interference by species
that absorb at 440 nm such as 1,2-dicarbonyl compounds. However, this source of
interference is expected to be small (-1%), and if necessary, the extent of this
interference can be limited by shifting the detection to longer wavelengths and adjusting
the lower edge of the detection band to 455 nm. In principle, NO2 detection limits could
be <30 ppt for a 60-second time scale.
Lee etal. (2011 a) describe the development of a dual continuous-wave mode quantum
cascade-tunable infrared laser differential absorption spectrometer (QC-TILDAS) to
measure NO2 and HONO simultaneously. The one-second detection limit [signal-to-noise
ratio (S/N) = 3] for NO2 is 30 ppt. A field comparison of measurements of NO2 between
CAPS and CL/MC is shown in Figure 2-6. The CAPS—CL/MC (Thermo Electron 421)
data were obtained over 4 days in a parking lot located -200 m from a major arterial
highway (Route 3 in Billerica, MA) in October 2007. Figure 2-7 shows the results of a
2-26
-------
comparison of NCh measured by QC-TILDAS to NO2 measured by CL/PC. The
QC-TILDAS-CL/PC data were collected in Houston, TX in April and May of 2009
during the SHARP campaign (Olaguer et al. 2014). Both comparisons show very high R2
(>0.99) and close agreement over concentrations ranging from <1 ppb to >30 ppb, and
both comparisons are characterized by small nonzero intercepts. For the CAPS
instrument (Figure 2-6). slightly higher values than those reported by the CL/MC monitor
are seen at concentrations <~2 ppb. Figure 2-7 shows that the QC-TILDAS obtains
slightly lower concentrations than reported by CL/PC for NC>2 concentrations <~1 ppb.
Although CAPS presents a practical alternative to chemiluminescence for NO2
measurements, an important consideration in routine network deployment of CAPS or
any other method that only measures NC>2 (e.g., does not measure NO) is the potential
loss of NO and NOx data, which has been used as an indicator for traffic- or other
combustion-related pollution.
2-27
-------
en
CL-
= 1.007x +0.0357
R2 = 0.9939
i 10
CL/MC NO2 (ppb)
Note: CAPS = cavity attenuated phase shift, CL/MC = chemiluminescence/MoOx catalytic converter, NO2 = nitrogen dioxide.
Source: National Center for Environmental Assessment 2013 analysis of data from Kebabian et al. (2008).
Figure 2-6 Comparison of nitrogen dioxide measured by cavity attenuated
phase shift spectroscopy to nitrogen dioxide measured by
chemiluminescence/molybdenum oxide catalytic converter for
4 days in October 2007 in Billerica, MA.
2-28
-------
-Q
Q.
•a 10
U 1
a
y = l.OOllx - 0.1182
R2 = 0.9924
1 10
CL/PC NO2 (ppb)
Note: NO2 = nitrogen dioxide, PC = photolytic converter, QC-TILDAS = quantum cascade-tunable infrared differential absorption
spectroscopy.
Source: National Center for Environmental Assessment 2013 analysis of data from Lee et al. (2013).
Figure 2-7 Comparison of nitrogen dioxide measured by quantum
cascade-tunable infrared differential absorption spectroscopy to
nitrogen dioxide measured by chemiluminescence with photolytic
converter during April and May 2009 in Houston, TX.
Villena et al. (2011) describe the development of a long path absorption photometer
(LOPAP) to measure NC>2. In this technique, NO2 is sampled in a stripping coil using a
modified Griess-Saltzman reagent with the production of an azo dye whose visible
absorption is measured by long-path photometry. This reaction was the basis for a much
earlier manual method for measuring NC>2 (Saltzman. 1954). Interference, which can be
minimized by additional stripping coils, could be caused by HONO, Os, and PAN. In an
intercomparison with a CL/PC carried out over four days in March 2007 on the fifth floor
balcony of a building at the University of Wuppertal in Germany, very good agreement
(mean deviation of 2%) was obtained. Interestingly, in the entire range of NC>2
measurements (~0.5 ppb to ~40 ppb), the relation between LOPAP and CL/PC can be
characterized by LOPAP (ppb) = 0.984 x CL/PC - 0.42 (ppb). However, if the range
2-29
-------
<6 ppb only is considered, the relation becomes LOPAP (ppb) = 0.998 x CL/PC + 0.19
(ppb).
Diode laser-based cavity ring-down spectroscopy (CRDS) has also been used to detect
NO2. Fuchs et al. (2009) developed a portable instrument that relies on NO2 absorption at
404 nm, with 22 ppt detection limit at 1 second (S/N = 2). As opposed to
chemiluminescence monitors that measure NC>2 indirectly based on direct measurement
of NO, NO2 (formed by reaction of NO with excess Os) is directly measured in CRDS.
NO is then determined by subtracting NO2 measured in the first cavity from the sum of
NO2 and NO (i.e., NOx) measured in the second cavity. The Os is generated by
photolysis of O2 in the Schumann-Runge bands at 185 nm. This conversion should be
much more complete than relying on the reduction of NO2 and NOz species with variable
efficiency on a MoOx converter. Note that the optical methods relying on NO2 absorption
at -400 nm described above (i.e., CAPS, CRDS) might be subject to positive interference
from absorption by trace components (e.g., glyoxal and methyl glyoxal). However,
absorption cross sections for these dicarbonyls are much lower than for NO2 at this
wavelength, and concentrations for these potentially interfering species are generally
lower than those for NO2. Furthermore, it is possible that thermal decomposition of NOz
species, such as PAN, in inlets or their reduction on inlet surfaces or in optical cavities
can be a source of NO2 in these or other instruments requiring an inlet.
2.4.3 Satellite Measurements of Nitrogen Dioxide
Remote sensing by satellites is an approach that could be especially useful in areas where
surface monitors are sparse. Retrieving NO2 column abundances from satellite data
involves three steps: (1) determining the total NO2 integrated line-of-sight (slant)
abundance by spectral fitting of solar backscatter measurements; (2) removing the
stratospheric contribution by using data from remote regions where the tropospheric
column abundance1 is small; and (3) applying an air mass factor to convert tropospheric
slant columns into vertical columns. The retrieval uncertainty is largely determined by
Steps 1 and 2 over remote regions where there is little tropospheric NO2, and by Step 3,
over regions of elevated tropospheric NO2 (Boersma et al., 2004; Martin etal. 2002).
Satellite retrievals are largely limited to cloud fractions <20%. The algorithm used here to
derive the tropospheric column of NO2 is given in Bucselaetal. (2013). This algorithm
was used to generate the maps in Figure 2-8 for 2005 to 2007 and in Figure 2-9 for 2010
to 2012 showing seasonal average NO2 columns obtained by the Ozone Monitoring
Instrument (OMI) on the AURA satellite. Other algorithms, for example the Berkeley
1 Column refers to the integrated line-of-sight abundance in a unit cross section, such that its units are
molecules/cm2.
2-30
-------
High-Resolution product (Russell et al., 2011). which is based on higher resolution input
fields (topography, albedo, and NCh vertical profile shape) in the retrievals, can reduce
the uncertainty in the measurements.
OMI Tropospheric NO2 (1O1S molec. crn E)
1O
8
Note: Images shown were constructed by Dr. Lok Lamsal of Universities Space Research Association from data obtained by the
Ozone Monitoring Instrument (OMI) on the AURA satellite (http://aura.asfc.nasa.gov/scinst/omi.htmn using the algorithm described
in Bucsela et al. (2013). Top panel (winter; DJF: December, January, February). Lower panel (summer; JJA: June, July, August).
Figure 2-8 Seasonal average tropospheric column abundances for nitrogen
dioxide (1015 molecules/cm2) derived by ozone monitoring
instrument for winter (upper panel) and summer (lower panel) for
2005 to 2007.
2-31
-------
OMI Tropospheric NO2 (1O15 molec. cm2)
I
10
8
7
6
5
O.8
O.1
Note: Images shown were constructed by Dr. Lok Lamsal of Universities Space Research Association from data obtained by the
Ozone Monitoring Instrument (OMI) on the AURA satellite (http://aura.asfc.nasa.gov/scinst/omi.htmn using the algorithm described
in Bucsela et al. (2013). Top panel (winter; DJF: December, January, February). Lower panel (summer; JJA: June, July, August).
Figure 2-9 Seasonal average tropospheric column abundances for nitrogen
dioxide (1015 molecules/cm2) derived by ozone monitoring
instrument for winter (upper panel) and summer (lower panel) for
2010 to 2012.
Areas of high column NC>2 abundance are found over major source areas during both
2-year periods shown in Figures 2-8 and 2-9. High column abundances are found over
many major urban areas, such as Los Angeles, CA; Houston, TX; Chicago, IL; and New
York, NY; and over major power plant complexes such as the Four Corners (Colorado,
New Mexico, Arizona, and Utah) and the Ohio River Valley. A diffuse area with column
abundances above background is found over the Bakken Shale fields in northwestern
North Dakota in winter. However, in general, the area of high column abundance of NCh
2-32
-------
(shown in orange and red) is smaller in the 2010 to 2012 composite than from 2005 to
2007. The photochemical lifetime of NC>2 is longer in winter than in summer resulting in
lower column abundances of NO2 in summer than in winter during the two 3-year periods
shown in Figures 2-8 and 2-9.
Because satellite instruments do not return surface concentrations directly, information
on NC>2 surface concentrations must be inferred from the column measurements. Lamsal
et al. (2008) and Lamsal etal. (2010) combined satellite data for column NO2 from OMI
with results from the Goddard Earth Observing System (GEOS)-Chem global scale
chemistry-transport model to derive surface concentrations of NC>2 (see Figure 2-13 for
an example of seasonally averaged surface NO2 concentrations derived by this method).
This method accounts for the feedback from the abundance of NCh on the lifetime of
NO2. Note, however, that data are collected only during the daily satellite overpass in
early afternoon and this method has only been applied for the time of satellite overpass.
Some other means must be used to extend the time period of applicability, for example by
scaling the afternoon value by the diel variation in a model, provided the model bias in
simulating NC>2 has been characterized over the times of interest in a 24-hour cycle
(Stavrakou etal.. 2008: Kim et al.. 2006b).
2.4.4 Measurements of Total Oxides of Nitrogen in the Atmosphere
Commercially available NOx monitors have been converted to NOy monitors by moving
the MoOx converter to interface directly with the sample inlet. Because of losses on inlet
surfaces and differences in the efficiency of reduction of NOz compounds on the heated
MoOx substrate, NOx concentrations cannot be considered as a universal surrogate for
NOy. However, most of the NOy is present as NOx close to sources of fresh combustion
emissions, such as highways during rush hour. To the extent that all the major oxidized
nitrogen species can be reduced quantitatively to NO, measurements of NOy
concentrations should be more reliable than those for NOx concentrations, particularly at
typical ambient levels of NO2. Exceptions might apply in locations near NOx sources,
where NOx measurements are likely to be less biased and confidence in measurement
accuracy increases.
Alternatively, multiple methods for observing components of NOy have been developed
and evaluated in some detail. As a result of these methods, as applied in the field and the
laboratory, knowledge of the chemistry of odd-N species has evolved rapidly. Recent
evaluations of methods can be found in Arnold et al. (2007) for HNOs, Wooldridge et al.
(2010) for speciated PANs, and Pinto etal. (2014) for HONO. However, it is worth
reiterating that the direct measurements of NO are still the most reliable method. Reliable
2-33
-------
measurements of NOy and NCh concentrations, especially at the low concentrations
observed in many areas remote from sources, are also crucial for evaluating the
performance of three-dimensional, chemical transport models of oxidant and acid
production in the atmosphere.
2.4.5 Ambient Sampling Network Design
Figure 2-10 shows that approximately 500 monitoring sites operate routinely across the
U.S. for oxidized nitrogen in ambient air. Four networks are highlighted:
(1) regulatory-based SLAMS designed to determine NAAQS compliance; (2) Clean Air
Status and Trends Network (CASTNET), which provides weekly averaged values of total
nitrate (HNOs and pNOs) in rural locations; (3) the National Core (NCore) Network, a
subset of SLAMS comprised of approximately 70 stations designed to capture
area-representative multiple-pollutant concentrations that provides routinely measured
NOy; and (4) the Southeast Aerosol Research Characterization (SEARCH), a privately
funded network of 6-10 sites including direct measurements of true NC>2 as well as NOy
and other nitrogen species (oxidized and reduced forms). Relative to the Os and PM
monitoring networks, the ambient NC>2 monitor density is significantly lower.
2-34
-------
SLAMS
CASTNET
NCORE_
SEARCH
Note: SLAMS = State and Local Air Monitoring Stations, CASTNET = Clean Air Status and Trends Network, NCORE = National
Core Network, SEARCH = Southeast Aerosol Research Characterization.
Source: U.S. Environmental Protection Agency 2013 analysis of data from monitoring networks.
Figure 2-10 Map of monitoring sites for oxides of nitrogen in the U.S. from
four networks.
Currently, with the exception of 4-6 sites in the SEARCH network (Hansen et al.. 2003).
direct or true NCh is sparingly measured on a routine basis. The regulatory networks rely
mainly on chemiluminescence difference techniques that provide NO concentration
directly and report a calculated NO2 concentration as the difference between NOx
concentration and NO concentration as discussed in Section 2.4.1. Criteria for siting
ambient NO2 and NOy monitors are laid out in 40 CFR Part 58, Appendix D. NO2
monitors are meant to be representative of several scales: microscale (in close proximity,
up to 100 m from the source), middle (several city blocks, 100 to 500 m), neighborhood
(0.5 to 4 km), and urban (4 to 50 km). Microscale to neighborhood-scale monitors are
used to determine the highest concentrations and source impacts, while neighborhood-
and urban-scale monitors are used for relatively wider area concentrations.
The U.S. EPA promulgated new minimum monitoring requirements in February 2010,
mandating that state and local air monitoring agencies install near-road NO2 monitoring
2-35
-------
stations within the near-road environment in larger urban areas, and other monitoring
stations including area-wide measurements, and measurements in areas having vulnerable
and susceptible populations. With regard to the near-road monitors, under these new
requirements, state and local air agencies will operate one near-road NO2 monitor in any
CBSA with a population of 500,000 or more, and two near-road NCh monitors in CBSAs
with 2,500,000 or more persons or roadway segments carrying traffic volumes of 250,000
or more vehicles. These monitoring data are intended to represent the highest population
exposures that may be occurring in the near-road environment throughout an urban area
over the averaging times of interest. The near-road NO2 network is intended to focus
monitoring resources on near-road locations where peak ambient NCh concentrations are
expected to occur because of on-road mobile source emissions and to provide a clear
means to determine whether the NAAQS is being met within the near-road environment
throughout a particular urban area. The network is now being phased in, and the first
phase became operational in January of 2014.
2.5 Ambient Concentrations of Oxides of Nitrogen
This section provides a brief overview of ambient concentrations of NO2 and associated
oxidized N compounds in the U.S.; it also provides estimates of background
concentrations used to inform risk and policy assessments for the review of the NAAQS.
In the 2008 ISA for Oxides of Nitrogen, NC>2 concentrations were summarized with an
explanation that the annual average NCh concentrations of ~15 ppb reported by the
regulatory monitoring networks were well below the level of the NAAQS (53 ppb), but
that the 1-hour daily maximum concentrations can be much greater in some locations,
especially in areas with heavy traffic (U.S. EPA. 2008c).
2.5.1 National-Scale Spatial Variability
In the 2008 ISA for Oxides of Nitrogen, data were analyzed for NO2 measured at
monitoring sites located within urbanized areas in the U.S. (U.S. EPA. 2008c). NO2
concentrations were -15 ppb for averaging periods ranging from a day to a year, and the
1-hour daily maximum NO2 concentration was -30 ppb, about twice as high as the
24-h avg. Data on NOz concentrations were very limited but indicated that HNOs and
HONO concentrations were considerably lower than NO2 concentrations. HNOs
concentrations ranged from <1 to >10 ppb, and HONO concentrations were reported as
<1 ppb even under heavily polluted conditions. HNOs concentrations were highest
downwind of an urban center. HONO was present in areas with traffic, at concentrations
2-36
-------
several percent of NO2 concentrations (U.S. EPA. 2008c). Field study results indicating
much higher NOz concentrations than NOx concentrations in relatively unpolluted rural
air were also described (U.S. EPA. 2008c).
Figure 2-11 presents a national map of the U.S. 98th percentile of 1-hour daily maximum
concentrations based on 2011-2013 data, and Figure 2-12 presents annual average NO2
concentrations based on 2013 calendar year data. In both figures, data are included only
for monitors with 75% of days reported for each calendar quarter over the 3-year period
and only for days with 75% of all hours reported. Because of the completeness
requirements, there are cases where sites have valid annual average data but not valid
1-hour daily maximum concentrations. The highest concentrations are in the Northeast
Corridor, California, and other urbanized regions, and the lowest concentrations are in
sparsely populated regions, most notably in the West. These observations are consistent
with those described in the 2008 ISA (U.S. EPA. 2008c).
Table 2-3 presents summary data on 1-hour daily maximum NC>2 concentrations and
Table 2-4 presents annual average NO, NO2 and NOx concentrations for the period
2011-2013. Table 2-3 also includes summary data by individual years and by quarters
averaged over the 3 years, as well as summary data for selected urban areas that are
examined in recent U.S. epidemiologic studies on the health effects of NO2 (Chapters 5
and 6). Nationally, 1-hour daily maximum concentrations rarely exceeded 60 ppb for the
3-year period, but 99th percentile concentrations were greater than 60 ppb in New York,
NY; Los Angeles, CA; and Denver, CO. The 50th percentile 1-hour daily maximum
concentration nationwide was 16 ppb, but varied among cities in Table 2-3. ranging from
8 ppb in Atlanta to near 40 ppb in Denver, CO. Annual average NO2 concentrations from
Table 2-4 were mostly less than 10 ppb, and the highest concentrations never exceeded
26 ppb. The 50th percentile annual average NO concentrations were less than those for
NO2 but the ratio of NO concentration to NO2 concentration increased with increasing
NOx concentration. For 99th percentile and maximum, annual NO concentrations were
higher than those for NO2, indicating that on the most polluted days, the ratio of NO to
NO2 is higher, consistent with fresh combustion emissions. Annual average NOx
concentrations for a given location were usually under 20 ppb and never exceeded
70 ppb.
2-37
-------
Legend
2011-2013 98th Percentile Daily 1-Hour
Maximum NC»2 Concentration (ppb)
I I 25-37
| 38-46
I 1*7-56
^H 57 - 73
Note: NO2 = nitrogen dioxide. Concentrations indicated are the highest concentration in the county and do not represent countywide
concentrations.
Source: U.S. Environmental Protection Agency 2014 analysis of data from state and local air monitoring stations.
Figure 2-11 U.S. 98th percentiles of 1-hour daily maximum nitrogen dioxide
concentrations for 2011-2013.
2-38
-------
Legend
Nth Concentration (ppb)
^B 0-5
| 6- 10
I I 11 -14
| | 15-19
^B 20 - 25
Note: NO2 = nitrogen dioxide. Concentrations indicated are the highest concentration in the county and do not represent countywide
concentrations.
Source: U.S. Environmental Protection Agency 2014 analysis of data from state and local air monitoring stations.
Figure 2-12 U.S. annual average nitrogen dioxide concentrations for 2013.
2-39
-------
Table 2-3 Summary statistics for 1-hour daily maximum nitrogen dioxide
concentrations (ppb) based on state and local air monitoring
stations.
Percent! les
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
Atlanta, GAa
Atlanta, GA— allb
Boston, MAa
Boston, MA— allb
Denver, COa
Denver, CO— allb
Houston, TXa
Houston, TX— allb
Los Angeles, CAa
Los Angeles, CA — allb
New York, NYa
New York, NY— allb
Seattle, WAa
Seattle, WA— allb
Year
2011-2013
2011
2012
2013
1st Quarter
2nd Quarter
3rd Quarter
4th Quarter
2011-2013
2011-2013
2011-2013
2011-2013
2011-2013
2011-2013
2011-2013
2011-2013
2011-2013
2011-2013
2011-2013
2011-2013
2011-2013
2011-2013
n
390,713
127,610
130,170
132,933
94,612
96,962
99,125
100,101
3,215
3,215
6,246
10,986
966
2,184
9,525
16,610
8,328
30,612
9,469
11,803
none
1,649
Mean
19
19
18
18
22
16
16
21
13
13
25
19
38
41
21
18
27
28
27
27
13
1
1
1
1
1
1
1
1
1
2
2
5
1
6
9
1
1
4
4
1
1
3
5
2
2
2
2
2
2
2
2
2
2
8
3
14
21
3
3
7
7
3
3
4
10
3
4
3
3
4
3
3
4
3
3
11
4
22
26
5
4
10
10
5
5
5
25
8
8
8
7
10
6
7
10
4
4
16
9
30
33
10
8
16
17
13
15
7
50
16
16
16
15
20
12
13
20
8
8
24
17
39
41
18
15
26
28
27
27
11
75
27
28
27
26
32
22
22
31
18
18
32
28
46
48
29
26
36
38
38
38
17
90
38
39
37
37
41
33
32
40
34
34
39
36
53
55
45
36
44
47
47
47
24
95
44
45
43
43
47
40
38
46
41
41
44
41
58
61
45
43
49
52
52
52
31
99
55
57
55
54
58
52
50
58
52
52
52
49
68
73
56
54
60
63
64
63
46
NO2 = nitrogen dioxide; GA = Georgia; MA = Massachussets; CO = Colorado; TX = Texas; CA = California; NY = New York;
WA = Washington.
aCity name only rows contain hourly data that meet 75% completeness criteria.
bCity—all rows report data regardless of whether completeness criteria are met.
Source: Office of Air Quality Planning and Standards and National Center for Environmental Assessment 2014 analysis of Air
Quality System network data 2011-2013.
2-40
-------
Table 2-4 Summary statistics for nitrogen dioxide, nitric oxide, and sum of nitrogen dioxide and nitric oxide
annual average concentrations (ppb) based on state and local air monitoring stations.
Percentiles
Pollutant Year
n
Mean
5
10
25
50 75
90
95
99
Max
NO2
2011-2013
2011
2012
2013
1,041
338
347
356
8.6
9.0
8.5
8.3
1.4
1.5
1.4
1.3
2.2
2.5
2.2
2.1
4.3
4.7
4.2
4.2
8.1 11.
8.4 12.
8.1 11.
7.7 11.
8
,3
,6
,6
16.2
16.8
15.9
15.8
18.6
19.6
18.6
18.1
22.5
23.9
22.1
21.8
26.0
25.3
26.0
24.6
NO
2011-2013
2011
2012
2013
1,127
363
377
387
4.8
5.0
4.8
4.6
0.1
0.1
0.1
0.1
0.3
0.3
0.2
0.2
1.0
1.1
1.0
1.0
2.9 6.
3.1 7.
8
,4
2.9 6.6
2.6 6.
,5
11.3
12.7
10.9
11.0
15.3
15.1
15.0
15.7
25.3
23.9
27.7
21.5
48.8
46.9
48.8
36.2
NOx
2011-2013
2011
2012
2013
1,011
320
342
349
13.4
13.7
13.2
13.3
1.5
1.5
1.3
1.7
2.6
2.6
2.6
2.6
5.4
5.8
5.2
5.4
11.3 18.
11.8 19.
11.2 18.
10.9 18.
,6
3
,5
,3
28.1
28.9
26.8
28.0
31.8
31.7
31.3
32.7
45.4
44.8
48.9
44.1
68.4
68.4
61.0
61.7
Max = maximum, Min = minimum; NO = nitric oxide; NO2 = nitrogen dioxide; NOX = sum of NO2 and NO.
Source: Office of Air Quality Planning and Standards and National Center for Environmental Assessment 2014 analysis of Air Quality System network data 2011-2013.
2-41
-------
As described in Section 2.2. the lifetime of NO2 with respect to conversion to NOz
species can be as short as an hour. This relatively short NO2 lifetime results in gradients
and low concentrations away from major sources that are not adequately captured by the
existing monitoring networks (see Figure 2-10 for location of monitoring sites). Satellite
data coupled with model simulations might be more useful for showing large-scale
features in the distribution of NC>2. Winter and summer seasonal average NO2
concentrations for 2009-2011 derived from the OMI on the AURA satellite and the
GEOS-Chem global, three-dimensional chemistry-transport model are shown in
Figure 2-13. In this method, integrated vertical column abundances of NC>2 derived from
the OMI instrument are scaled to surface mixing ratios using scaling factors derived from
GEOS-Chem [see (Lamsaletal.. 2010: Lamsal et al.. 2008): also see Section 2A for
more complete descriptions of the method]. A nested version of GEOS-Chem at
50 km x 50 km horizontal resolution is used in this method. A description of the
capabilities of GEOS-Chem and other three-dimensional chemistry transport models is
given in the 2013 ISA for Ozone (U.S. EPA. 2013e).
Large variability in NO2 concentrations is apparent in Figure 2-13. As expected, the
highest NO2 concentrations are seen in large urban regions, such as in the Northeast
Corridor, and lowest values are found in sparsely populated regions located mainly in the
West. Minimum hourly values can be less than -10 ppt, leading to a large range between
maximum and minimum concentrations. Although overall patterns of spatial variability
are consistent with the current understanding of the behavior of NO2, not much
confidence should be placed on values <~100 ppt due to limitations in the satellite
retrievals. NO2 concentrations tend to be higher in January than in July, largely reflecting
lower planetary boundary layer heights in winter. Such seasonal variability is also evident
on a local scale, as measured by surface monitors. For example, in Atlanta, GA, NOx
measurements also exhibited higher concentrations in winter and lower concentrations in
summer, when NOx is more rapidly removed by photochemical reactions (Tachon et al..
2012).
2-42
-------
O Ml-derived surface NO2 (ppb)
Note: NO2 = nitrogen dioxide, OMI = Ozone Monitoring Instrument. Images shown were constructed by Dr. Lok Lamsal of
Universities Space Research Association from data obtained by the OMI instrument on the AURA satellite
(http://aura.asfc.nasa.gov/scinst/omi.htmn using the algorithm described in Bucsela et al. (2013). Output from the GEOS-Chem,
global-scale, three-dimensional, chemistry-transport model is used to derive surface concentration fields from the satellite data as
described in Lamsal et al. (2008) and Lamsal etal. (2010).
Top panel (winter; DJF: December, January, February). Lower panel (summer; JJA: June, July, August).
Figure 2-13 Seasonal average surface nitrogen dioxide concentrations in ppb
for winter (upper panel) and summer (lower panel) derived by
ozone monitoring instrument/Goddard Earth Observing
System-Chem for 2009-2011.
2-43
-------
2.5.2 Urban-Scale Spatial Variability
Figure 2-14 describes 1-hour daily maximum concentration agreement between pairs of
SLAM monitors from 2011-2013 for selected U.S. CBSAs with more than one monitor.
Agreement is expressed as coefficient of divergence (COD), which has been widely used
to assess spatial variability of air pollutant concentrations (U.S. EPA, 2008c; Wilson et
al.. 2005; Pinto et al.. 2004). In practical terms, a COD = 0 indicates perfect agreement,
and COD values increase as spatial variability increases. COD values in Figure 2-14
generally range from about 0.1 to 0.4, with a few higher values. This indicates a range of
variability across CBSAs from fairly uniform to a moderate degree of variability (Wilson
et al.. 2005). At first glance, distance between sites does not appear to be an important
factor for explaining variability between site pairs on an urban scale. However, for
extremely short distances, a trend with distance is observed, especially for data within the
same city. For example, for Boston, MA, the six observations with the shortest distances
between them exhibit a trend of increasing COD with distance, from about 3 km to about
10 km. These data are for the four sites closest to the city center. As indicated by the
COD values, there is a substantial degree of variability for all but the closest sites, with
CODs ranging above 0.4 even for comparison between site pairs near the city center.
2-44
-------
0.6-
0.5-
8
5
E> n
a
•&
e
•
•r>
E 0.3-
8
o
0.2-
0.1-
A o Los Angeles. CA
+ Phoenix, AZ
A x Philadelphia. PA
A A Boston. MA
A
O - ^ O
A A °
o o o
A O
£
A A O + Q O o
o o o
o o
A Q On 6 o O ^ °
A X o X ^ o o °
A _y O O O
A 0 0 0 x ^ ^-n00 ° o 0 0
A 9, @ o ° ? °° °
+o ° ^?^fto a oP° ° < ° o
O O c^> O O *" O
oo S3 oo
Ao++>^9 0 0®
I I 1 t I I
0 20 40 60 80 100
Distance between Monitors {km)
Source: National Center for Environmental Assessment 2014 analysis of Air Quality System network data.
Figure 2-14 Coefficient of divergence for ambient nitrogen dioxide
concentrations between monitor pairs in four U.S. cities.
A similar trend is observed for Los Angeles, CA but over a broader scale. In Los
Angeles, CA, the highest observed COD at a given distance increases regularly with
distance up to about 40 km, but at distances greater than 40 km it appears to level off
with distance at about 0.4. In addition, for all sites within 15 km of each other, a high
degree of agreement is observed. The difference between Boston, MA and Los Angeles,
CA in how much COD changes with distance is consistent with their different extents.
Boston, MA has a much smaller land surface area than Los Angeles, CA. Another major
difference between Boston, MA and Los Angeles, CA is that good agreement (COD
~0.1) is often observed between sites up to 50 km or more apart in Los Angeles, CA,
suggesting that other factors besides distance are important. Five of the Los Angeles, CA
monitors (Main Street Los Angeles, Burbank, Pasadena, Pomona, and Long Beach
North) form a subset of monitors with distinctly lower variability than the area as a
2-45
-------
whole, with low CODs for each possible combination of monitors within this group, as
shown in Figure 2-15. Other monitors near the ocean or mountains exhibit poorer
agreement with these five monitors, even if the distances between monitors are shorter.
Coefficient of Divergence
o o o
b P h-> P k)
O Un h-» un M un
•
•
• •
•
•
•
• •
•
•
•
•
•
) 10 20 30 40 50 60
Distance between Monitors (km)
Source: National Center for Environmental Assessment 2014 analysis of Air Quality System network data.
Figure 2-15 Coefficient of divergence for ambient nitrogen dioxide
concentrations among a subset of five Los Angeles, California
monitors.
Yet another pattern is observed for Phoenix, Arizona (AZ) and Philadelphia, PA. For
these cities, low COD values are observed for all sites except rural locations outside of
the urbanized area. The Phoenix, AZ data in Figure 2-14 fall into two clusters: one for
urban site pairs ranging up to 10 km distance from each other and one for urban-rural site
pairs 40 to 60 km from each other. All of the comparisons between urban sites exhibit a
COD <0.2, but poorer agreement is observed between urban and rural site pairs.
Similarly, good agreement (COD ~0.1) is observed between two monitoring sites
operating within the city of Philadelphia, PA about 10 km apart, but poorer agreement is
observed for more distant suburban sites. This result is consistent with observations of
Sarnat et al. (2010). who observed that using monitors in rural areas of counties
considered part of the Atlanta, GA metropolitan area affected relative risk estimates for
2-46
-------
associations with health effects, but that using different urban monitors within
approximately 15 to 30 km of a study subject did not.
To summarize urban-scale spatial variability for NC>2, good agreement among nearby
sites in city cores is common and was observed for 2011 to 2013 data for all sites in
Philadelphia, PA and Phoenix, AZ. In Los Angeles, CA good agreement was also usually
observed over similar distances to those compared in Philadelphia, PA and Phoenix, AZ
(i.e., among sites separated by less than 15 km). In contrast, agreement among monitors
in Boston, MA became poorer over a shorter distance. NCh concentrations followed a
trend of increasing variability with distance among sites over 3 to 10 km, a smaller spatial
scale than the other cities.
Similar results are observed for annual averages. Tables 2-5A and 2-5B present the
difference in annual average NO2 concentrations between pairs of sites divided by the
average between the two sites for that year as a measure of the percent difference in
concentration for Boston, MA and Los Angeles, CA. The CODs of annual average NO2
concentrations show wide ranges in agreement similar to those reported for 1-hour daily
maximum NC>2 concentrations measured in both Boston, MA and Los Angeles, CA. The
nearest site pairings in Boston, MA agree within 3 to 20%, while the other two site
pairings exhibit poorer agreement ranging from 38 to 65% and 31 to 90%.
The 14 sites in Los Angeles County that reported data for 2011 are shown in Table 2-5B.
A number of site pairings agree within 10 to 15%. For example, concentrations at the
Pico Rivera (1602 in Table 2-5B). Pomona (1701 in Table 2-5B). and Long Beach
Hudson (4006 in Table 2-5B) sites all agree within 10% of the concentrations reported at
the Los Angeles Main Street site.
Table 2-5A Percent difference in annual average nitrogen dioxide concentration
between monitors in Boston, Massachussets.
2011
2012
2013
0002 vs. 0040
41
65
38
ID's of Monitors Compared
0002 vs. 0042
10
20
3
0040 vs. 0042
31
47
90
Source: National Center for Environmental Assessment 2014 analysis of Air Quality System network data 2011-2013.
2-47
-------
Table 2-5B Percent difference in annual average nitrogen dioxide concentration
between monitors in Los Angeles, California for 2011.
Monitor
ID
0016
0113
1103
1201
1302
1602
1701
2005
4002
4006
5005
6012
9033
0002
7
17
21
16
3
22
26
7
7
12
32
35
47
0016
20
57
22
35
58
62
44
31
49
5
3
10
0113
38
2
15
39
43
24
10
29
15
18
30
Monitor ID
1103 1201 1302 1602 1701 2005 4002 4006 5005 6012
36
23 13
1 37 25
5 41 28 4
14 23 10 15 19
28 9 4 29 33 14
9 28 15 10 14 5 19
52 17 30 53 57 39 25 44
55 20 32 56 60 42 28 47 3
66 32 45 67 71 54 40 58 4 13
Source: National Center for Environmental Assessment 2014 analysis of Air Quality System network data 2011.
While these results indicate that relatively good agreement in 1-hour daily maximum and
annual average NO2 concentrations between pairs of nearby urban monitors in the same
metropolitan area occurs in some cases, they do not rule out the possibility of greater
variability on a smaller spatial scale. Vardoulakis et al. (2011) described a distinction
between "intra-urban" and "street-scale" variability, explaining that long-term monitoring
sites tend to be situated away from sources and hot spots that can strongly influence
variability. They compared results from long-term monitoring sites to short-term
networks of passive samplers. The passive samplers were placed in areas among the
long-term monitors at varying distances from key roads and intersections. Results
indicated that "street-level" variability determined from the passive sampler
measurements placed between long-term monitors exhibited greater variability than
2-48
-------
"intra-urban" variability based on long-term monitors themselves. Spatial variability near
roads is described in detail in Section 2.5.3.
2.5.3 Microscale to Neighborhood-Scale Spatial Variability, Including near
Roads
2.5.3.1 Near-Road Gradient Observations
General Observations
The spatial trends described in this section provide a background for understanding the
traffic-related NO2 exposure on and near roads described in Section 3.3.1.1. Numerous
observations have been summarized in several recent reviews, each of which concluded
that a zone of elevated NO2 concentration typically extends from 200 to 500 m from
roads with heavy traffic (HEI.2010; KarneretaL 2010; Zhou and Levy. 2007).
Tables 2-6 and 2-7 describe observations from studies that were included in these reviews
and/or in the 2008 Risk and Exposure Assessment for the review of the primary NC>2
NAAQS (U.S. EPA. 2008e) to estimate on-road concentrations, as well as more recent
observations. Concentrations measured relatively near the road (Cnear), concentrations
measured relatively farther from the road (Gar), and the difference between them in
concentration units (Cnear - Car) and as a fraction of the concentration measured farthest
from the road ([Cnear - Gar]/Gar) are summarized in Tables 2-6 and 2-7. Cnear was
measured from 0 to 60 m from the road, and Gar was measured at distances from 80 m to
more than 350m, depending on the study. Table 2-6 describes observations based on
averaging times of a half day or longer and includes important early studies based on
passive sampling methods that typically require sampling periods of 1 to 2 weeks to
collect a sufficient amount of sample. Table 2-7 describes observations based on
averaging times of 1 hour or shorter, and includes several recent studies with shorter time
resolution. A direct comparison of the observations included in these tables is not
appropriate because experimental designs, measurement methods, averaging times,
distances from the road, time of year, and other important factors vary among studies.
However, Tables 2-6 and 2-7 provide a broad overview of the magnitudes of NO2
concentration differences observed with distance from road and the spatial extent over
which differences in concentration have been observed.
2-49
-------
Table 2-6 Summary of near-road nitrogen dioxide concentration gradients from studies with passive samplers
and averaging times of 12 hours to 1 month.
Author
Gilbert etal. (2003)
Monnetal. (1997)
Pleiiel et al. (2004)
Roorda-Knape et al.
(1998)
Singer etal. (2004)
Smarqiassi et al.
(2005)
fBeckerman et al.
(2008)
Zou et al. (2006)
Gonzales et al.
(2005)
Cape etal. (2004)
Bell and Ashenden
(1997)
Location
Montreal,
Canada
Zurich,
Switzerland
Rural
Sweden
the
Netherlands
Oakland, CA
Montreal,
Canada
Toronto,
Canada
Shanghai,
China
El Paso, TX
Scotland,
U.K.
Rural
Wales, U.K.
Traffic
Count
(vehicles/
day)
185,000
NA
18,900-
32,500d
80,000-
152,000e
90,000-
210,000f
>150,000
349,1 00 &
395,400h
NA
NA
240-
85,000k
2,000-
6,000'
Method/
Averaging
Time
Passive/
1 week
Passive/
1 week
Passive/
1 month
Passive/
2 weeks
Passive/
1 week
Passive/
1 day
Passive/
1 week
Passive/
2 weeks
Passive/
1 week
Passive/
1 week
Passive/
1 week
Time of
Year
September
Summer
Winter
Not reported
Spring/
Summer
Fall/spring
May-June
August
All year
Winter
April-May
All year
Cnear Average
or Range
(distance from
road)
29 ppbb (0 m)
20ppbc(15m)
25 ppbc(15 m)
8-18ppbd
(10m)
24-25 ppbe
(15-32m)
30 ppbf (60 m)
33 ppb9(<10m)
19&28ppbh
(4 & 47 m)
50-65 ppb1
(Om)
25 ppW (0.25 m)
3-50ppbk(1 m)
8-28 ppb1
(<1 m)
Cfar Average or
Range
(distance from
road)
18 ppbb(200m)
12ppbc(80m)
25 ppbc (80 m)
4-10ppbd
(100m)
16-17 ppbe
(260-305 m)
20 ppbf
(>350 m)
20 ppb9
(>1,000 m)
14&15ppbh
(>380 m)
39-48 ppb'
(350 m)
15ppW
(10 m)
2-14 ppb1
(200-350 m)
Spatial
Extent
(m)
200
>80
None
500
>300
350
NA
300
>350
>3,750
>10
100
% Difference3
in
Cnear and Cfar
Average or
Range
60b
>30C
~oc
80-1 00d
AC cne
60f
709
30&100h
30-40'
-70'
<0-70k
20-660'
ppb
Difference3 in
Cnear and Cfar
Average or
Range
11b
~8C
~oc
2-8d
8e
11f
139
4&15h
12-18'
10J
0-11k
3-20'
2-50
-------
Table 2-6 (Continued): Summary of near-road nitrogen dioxide concentration gradients from studies with passive
samplers and averaging times of 12 hours to 1 month.
Author
Signal et al. (2007)
Kodama et al. (2002)
Maruo et al. (2003)
Nittaetal. (1993)
Location
Rural
England,
U.K.
Tokyo,
Japan
Sapporo,
Japan
Tokyo,
Japan
Traffic
Count
(vehicles/
day)
74,000-
94,000m
50,000-
60,000"
NA
>30,000
Method/
Averaging
Time
Passive/
11 -17 days
Passive/
48 h
Sensor/
1/2 day
Colorimetry/
1 week
Time of
Year
NA
All year
July
NA
Cnear Average
or Range
(distance from
road)
25 ppbm
25-50 ppb"
(<50 m)
32 ppb°
34-57 ppbP
(Om)
Cfar Average or
Range
(distance from
road)
10-20 ppbm
(250 m)
20-45 ppb"
(>200 m)
22ppb°(150m)
(150m)
Spatial
Extent
(m)
100
>200
150
150
% Difference3
in
Cnear and Cfar
Average or
Range
25-1 50m
NA
45°
10-50P
ppb
Difference3 in
Cnear and Cfar
Average or
Range
5-15m
<10"
10°
8-17P
Cfar = concentrations measured at the farthest distance; Cnear = concentrations measured at the nearest distance; NA = not available.
a% Difference refers to (Cnear - Cfar)/Cfar; ppb Difference refers to Cnear - Cfar.
bBased on a single set of samples.
°Average for nine winter and eight summer sets of samples.
dRange for five sets of samples at two different roads (three at one road, two at the other).
eAverages for two different roads based on eight sets of samples at one road and nine at the other.
'Average for 14 spring and 8 fall sets of samples with distance measured to different roads for different samples.
9Average for 15 sets of samples.
hLow and high values from single experiments at two different roads.
'Range for 12 sets of samples with 3 from each of four seasons.
'Based on a single set of samples with distance measured to different roads for different samples.
kRange of annual average concentrations at 14 locations; annual average concentrations based on six bimonthly sets of samples.
'Range of concentrations for 26 sets of samples from two different road segments of the same road.
mRange of averages for three different road segments (two on the same road); averages are for eight sets of samples for each segment.
"Range of average concentrations from each of 10 sets of samples simultaneously measured outside homes at varying distances from the road. Number of samples in each set of
samples ranged 34-103.
"Average for 28 sets of samples.
""Average of single measurements at seven locations.
fStudy published since the 2008 ISA for Oxides of Nitrogen.
2-51
-------
Table 2-7
Author
Rodes and
Holland (1981)
fMassoli et al.
(2012)
fPolidori and
Fine (201 2b)
tKimbrouah et al.
(2013)
fMcAdam et al.
(2011)
fDurant et al.
(2010)
Summary of near-road nitrogen
of 1 hour or less.
Location
Los Angeles,
CA— high O3
Los Angeles,
CA — medium O3
Los Angeles,
CA— low O3
New York, NY
Los Angeles, CA
Las Vegas, NV
Downwind only
Ontario, Canada
Somerville, MA
Traffic
Count
(vehicles/
day)
200,000
210,000
NRe
161,500
34,000
>1 50,000
Method/
Averaging
Time
Chemilum
/1 h
Chemilum/
real-time0
Chemilum/1 h
Chemilum/1 h
Chemilum
/1 h
TILDAS/
real-time0
dioxide concentration
Time of
Year
July-
August
July
Summer
Winter
All year
Summer
January
Cnear
Average or Range
(distance from road)
~120ppbb(8m)
-80 ppbb (8 m)
-70 ppbb (8 m)
25-40 ppbd (10m)
28ppbf(15m)
37 ppb9 (15m)
25 ppbh (20 m)
28 ppb1 (20 m)
5.8 ppb' (10m)
~15-35ppbk(<50m)
gradients from studies
Cfar Average or
Range
(distance from road)
-40 ppbb (388 m)
-40 ppbb (388 m)
-40 ppbb (388 m)
25-40 ppbd (500 m)
18ppbf(80m)
32 ppb9 (80 m)
20 ppbh (300 m)
23 ppb1 (300 m)
4.5 ppb' (60 m)
-1 0-30 ppbk (400m)
Spatial
Extent (m)
>400
>400
>400
None
>80
>80
>100
>100
None
100-250
with averaging times
% Difference3
in Cnear & Cfar
Average or
Range
~200b
~100b
~80b
~0d
56f
159
30h
201
~25i
>0k
ppb Difference9 in
Cnear & Cfar
Average or
Range
80b
40b
30b
~0d
10f
59
5h
5
1'
<10k
Cfar = concentrations measured at the farthest distance; Cnear = concentrations measured at the nearest distance; Chemilum = chemiluminescence; NR = not reported; O3 = ozone;
TILDAS = tunable infrared differential absorption spectroscopy.
a% Difference refers to (Cnear ~ Cfar)/Cfar; ppb Difference refers to Cnear ~ Cfar.
bAverage of at least 27 hourly measurements for each category.
°Described by authors as real-time monitoring with a mobile platform.
dRange for five time periods from early to late morning with each time period averaged over two samples from different days.
eTotal daily traffic count not reported, truck traffic at maximum count (10:00 a.m.-2:00 p.m.) >600 vehicles/hour winter and >400 vehicles/hour summer. 20% of vehicles were
heavy-duty diesel trucks.
'Seasonal average of more than 900 hourly measurements for each season.
9Average of 7,390 hourly measurements.
hAverage of 2,913 hourly measurements.
'Average of 231 hourly measurements at both 10m and 30 m, 281 measurements at both 30 m and 60 m.
'Range for five sets of samples at different times of day on the same day.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
2-52
-------
If sufficient detail is given on individual experiments, ranges for Cnear, Cfar, and
differences between them are provided in Tables 2-6 and 2-7. Otherwise, averages over
the entire study or over various categories such as season, wind direction, location, or Os
concentration are given. Earlier studies, including some of the studies in Table 2-6. were
mainly limited to passive samplers that required collection for 1 or 2 weeks, making it
difficult to explore effects of time of day or wind direction, which typically shifts on
shorter time scales. Several recent studies, summarized in Table 2-6. have used
chemiluminescence, QC-TILDAS, and other methods that not only provide greater time
resolution, but also result in the collection of larger numbers of samples. Both are useful
for better understanding the factors influencing near-road concentration patterns. There
are essentially three types of experimental designs used in the studies listed in Tables 2-6
and 2-7. (1) samples are collected simultaneously at varying distance from the same road;
(2) samples are collected by a mobile laboratory with high time resolution (samples are
collected at different distances from a road, not simultaneously, but with minimal elapsed
time between sampling at different distances from the road); or (3) samples are collected
over a wider spatial scale at varying distances from a number of heavily trafficked roads
and distance parameters are not linked to the same road for all samples.
Most of the studies listed in Tables 2-6 and 2-7 conclude that the spatial extent of
elevated NO2 concentrations is within the range of 200 to 500 m from the road as
described by HEI (2010) or Zhou and Lew (2007). Some recent studies concluded that
the influence of the road on NO2 concentrations can extend even farther, up to several
kilometers, but with smaller differences in concentration (Gilbert et al.. 2007; Gonzales et
al.. 2005). Bell and Ashenden (1997) and Cape et al. (2004) also observed remarkably
greater differences in NO2 concentration with distance within the first 10 to 20 m from
road than at further distances from the road, suggesting the possibility of an exponential
relationship of decreasing concentration with distance from the road with a steeper
decrease right next to the road.
2-53
-------
Several investigators have attempted to fit NO2 concentration data as a function of
distance from the road. NO2 concentrations followed a logarithmic function with distance
from a road over a range of 100 m (Pleijel et al.. 2004). more than 300 m (Roorda-Knape
etal.. 1998). and more than 1,000 m (Gilbert et al.. 2003):
Cx = Cb + Cv — klogx
Equation 2-1
where
x = distance from the road
k = decay constant derived from empirical data
Cx = NO2 concentration at a distance x from a road
Cb = NO2 concentration contribution away from the influence of the road
Cv = NO2 concentration contribution from vehicles on a roadway
Cape et al. (2004) used an exponential decay function to fit NO2 concentrations measured
from 1 to 10m from the road:
Equation 2-2
A shifted power law model has also been used (Zou et al.. 2006):
Equation 2-3
Compared to NO, UFP, and other traffic-related pollutants, NO2 concentrations decrease
less rapidly with distance from the road over a range of about 200 to 500 m, and exhibit a
somewhat greater spatial extent of elevated concentration (Section 3.3.1.1). This has been
attributed to chemical production occurring downwind of roads (Section 2.2) and to other
nontraffic-related sources of NO2 (Chaney et al.. 2011; Zhou and Levy. 2007; Rodes and
Holland. 1981). Because of the interaction between dispersion and chemical reaction
described in Section 2.2. the distribution of NO2 downwind of roads would likely differ
from that of a strictly primary traffic pollutant. For example, Massoli etal. (2012) found
that the concentrations of carbon dioxide [CCh] and NOx decreased by approximately
2-54
-------
50% within 150m downwind of the LIE, but that the concentration of NC>2 was nearly
constant over this distance from the road.
A slight effect of wind conditions has also been observed. NO2 concentration varies with
distance from the road under all wind conditions, but is more pronounced downwind
from the road (Kimbrough et al.. 2013; McAdametal.. 2011; Beckerman et al.. 2008;
Roorda-Knape etal.. 1998). When air is sampled both upwind and downwind of the road,
more gradual gradients are observed on the downwind side of the roadway (Durant et al..
2010; Clements et al.. 2009; Hu et al.. 2009; Beckerman etal.. 2008). Also, higher
concentrations are observed at low wind speeds, especially for winds blowing from the
road (Kimbrough et al., 2013).
Because of the long sampling duration required for passive monitors, earlier studies in
Table 2-6 were limited to a few samples and it was not possible to focus on spatial
differences over short time periods. In many early studies using passive monitors with
usual sampling periods of 1 to 2 weeks, Cnear ranged from 30 to 100% higher than Car.
These results are consistent with more recent observations of an approximately 40%
decrease in NO2 concentration due to road closures in Boston, also measured by passive
sampling for 1 week (Levy et al., 2006).
Thousands of hourly chemiluminescence measurements from two recent studies
(Kimbrough et al.. 2013; Polidori and Fine. 2012b) support observations from the earlier
passive sampling studies in Table 2-6. As described in Table 2-7. Kimbrough et al.
(2013) reported average concentrations of more than 7,000 hourly measurements, and
showed that NO2 concentrations 20 m from the road were an average of 27% higher than
at 300 m from the road in Las Vegas, NV. In Los Angeles, CA, Polidori and Fine (2012b)
reported that the average of hourly NCh concentrations at 15 m was 56% higher in
summer and 15% higher in winter than at 80 m. Averaging over the two seasons gives an
NC>2 concentration 34% higher at 15 m than at 80 m, which is remarkably similar to the
observation of Kimbrough et al. (2013). Table 2-8 summarizes hourly NC>2 data from two
field studies for which samples were collected 10-15 m from a major road and
simultaneously 80-100 m from the road. In Los Angeles, CA, samples were collected as
part of a near-road monitoring study by the South Coast Air Quality Management District
on the 1-710 freeway, on which heavy-duty diesel trucks account for about 20% of the
total number of vehicles (Polidori and Fine. 2012b). Samples in Detroit, MI were
collected at the Eliza Howell Park monitoring sites (140,500 vehicles per day) at 10 m
and 100 m from the road, and 26,000 hours of data were available 2011-2014. On
average over the entire data set including both seasons, hourly concentrations were 34%
higher at 15 m than at 80 m from the road in Los Angeles, CA and 70% higher at 10 m
than at 100 m from the road in Detroit, MI. The largest differences were 40 ppb in Los
2-55
-------
Angeles, CA and 52 ppb in Detroit, MI. However, the 99th percentile differences were
less than 30 ppb, 98th percentile differences were less than 25 ppb, and 95th percentile
differences were less than 20 ppb in both locations.
Table 2-8 Distribution of differences in higher 1-hour nitrogen dioxide
concentrations 10-15 m and lower 1-hour nitrogen dioxide
concentrations at 80-100 m from the road at two locations with
heavy traffic.
Location
Los Angeles, CA
1-710 freeway with
heavy diesel
traffic13
Detroit, Ml
Eliza Howell Park
near 1-96, 140,500
vehicles/day13
Parameter
Difference (ppb)c
Percentage
difference11
Difference (ppb)c
Percentage
difference11
Percentiles
Mean3 50th 90th 95th 98th 99th
7.0 5.9 16 19 24 27
34% 30% 72% 86% 99% 116%
5.3 4.0 12 16 20 23
70% 45% 167% 233% 350% 450%
Maximum3
40
170%
52
e
aMean and maximum concentrations all of concentrations for the entire data set, including all seasons and years.
"Los Angeles, CA data were collected 1/29/2009 to 3/11/2009 and 6/30/2009 to 8/19/09. Detroit, Ml data were collected 10/1/2011 to
12/31/2014.
°Difference in concentration between monitors at 15 and 80 m from the road in Los Angeles, CAand 10 and 100 m in Detroit, Ml.
Percentage difference in concentration relative to the concentration farthest from the road (C15 - C8o)/C8o in Los Angeles, CA and
(C10 - C10o)/Cioo in Detroit, Ml.
eMaximum is infinite because some concentrations at 100 m (Cioo) are below detection limit.
Source: National Center for Environmental Assessment 2015 analysis of Los Angeles, CA data obtained from Polidori and Fine
(2012b) and Detroit, Ml data obtained from Air Quality System database.
In Los Angeles, CA, near-road hourly NC>2 concentrations were rarely more than 100%
higher than hourly concentrations farther from the road. Such large differences in
concentration occurred more frequently in Detroit, MI. On average, the difference in
concentration at 10 and 100 m of the road was 70%, and the 90th percentile was 167%.
This is probably due in part to the lower concentrations observed in Detroit, MI
(Section 2.5.3.1).
The magnitude of the difference in 1-hour NC>2 concentrations near and farther away
from the road are not directly comparable to earlier studies based on passive sampling
(Beckerman et al.. 2008; Singer et al.. 2004; Gilbert et al.. 2003; Roorda-Knape et al..
1998) because the passive sampling results are for longer averaging times of 1 week or
2-56
-------
longer. However, the hourly data presented in Table 2-8 averaged over several months or
years and the passive sampling data presented in Table 2-6 both indicate that NO2
concentrations are consistently higher near the road than at a greater distance from the
road.
The absolute differences in measured NO2 concentrations between the nearest and
farthest locations (Cnear - Cfar) in Tables 2-6 and 2-7 are also consistent across most
studies, with concentration differences rarely exceeding 20 ppb, regardless of averaging
time. The exception is the Rodes and Holland (1981) study from Los Angeles, CA in the
early 1980s. Because this is an older study than the others, the vehicle fleet was not
strictly regulated for NOx emissions. As a result, the concentrations observed may not be
relevant to current conditions. With this study excluded, the range in Cnear - Cfar is
somewhat smaller than the range for Cfar across all of the studies, which implies that a
ratio of concentrations at different distances from the road could be more strongly
influenced by the concentration away from the road than by the concentration nearest the
road.
Seasonal and Diurnal Patterns
It is worth noting that in most of the earlier passive sampling studies in Table 2-6.
samples were collected mostly in warmer months, between May and September,
depending on the study (Beckerman et al.. 2008; Singer et al.. 2004; Gilbert et al.. 2003;
Roorda-Knape etal.. 1998). In a comparison between seasons by Monnet al. (1997)
presented in Table 2-6. similar results were observed in summer, but very little difference
in NCh concentration was observed in winter. More recently, a number of field studies
based on hourly measurements using the chemiluminescence Federal Reference Method
(Section 2.4) have been conducted. Based on these measurements it is possible to
evaluate concentration trends over shorter time periods, to examine seasonal and diurnal
patterns, and to determine concentration averages and distributions of hourly data over
thousands of hours. From Table 2-6. it is evident that in studies with finer time resolution,
more observations of a lack of any difference between concentrations nearest the road
and farther from the road (Cnear - Cfar = ~0) are reported. The lack of a near-road NO2
concentration gradient appears to be especially common in early morning measurements
(Massoli et al.. 2012; McAdametal.. 2011).
Table 2-9 shows how NO2 concentration differences near and far away from the road are
influenced by season and time of day. Data are divided into a warm season and a cold
season at each location. Averages of hourly concentrations are higher 10-15 m from the
road than 80-100 m regardless of location, season, or time of day. The near-road
influence is greater during the day in the warmer months and smallest at night in the
2-57
-------
winter. On summer days, near-road concentrations are on average 8.5 ppb and 133%
higher in Detroit, MI and 10.1 ppb and 54% higher in Los Angeles, CA than
concentrations farther from the road. In contrast, average concentration differences on
winter nights are less than 3 ppb in both locations. In Los Angeles, CA this corresponds
to a less than 10% higher concentration near the road than at the 80-m distance. For
individual nighttime hours in winter, there was frequently little or no difference in 1-hour
NO2 concentration between the near-road measurement and measurements farther from
the road, similar to results reported in other studies focused on early morning
measurements (Massoli et al., 2012; McAdam et al., 2011).
Table 2-9 Seasonal and diurnal variation of differences in 1-hour nitrogen
dioxide concentrations 10-15 m and 80-100 m from the road at two
locations with heavy traffic.
Mean Concentration Mean
80-100 m of Mean Percentage
Location Season Time of Day roada(ppb) Differenceb(ppb) Difference0 (%)
Los Angeles,
CAd
1-710 freeway
with heavy
diesel traffic13
Detroit, Mld
Eliza Howell
Park near
1-96, 140,500
vehicles/day13
Warm
Cold
Warm
Cold
Ort Mar
7:00 a.m. to 6:00 p.m.
7:00 p.m. to 6:00 a.m.
7:00 a.m. to 6:00 p.m.
7:00 p.m. to 6:00 a.m.
7:00 a.m. to 6:00 p.m.
7:00 p.m. to 6:00 a.m.
7:00 a.m. to 6:00 p.m.
7:00 p.m. to 6:00 a.m.
19.2
20.4
28.8
35.9
9.1
12.3
13.4
15.0
10.1
8.0
6.6
2.8
8.5
3.8
6.7
2.7
54
41
27
9
133
48
73
32
Aug = August; a.m. = ante meridiem; Feb = February.
aMean concentration 80 m from the road in Los Angeles, CA and 100 m from the road in Detroit, Ml. Mean of all hourly
concentrations in time of day and season specified.
bMean difference in concentration between monitors at 15 m and 80 m from the road in Los Angeles, CA and 10m and 100 m in
Detroit, Ml. Mean of all hourly concentrations in time of day and season specified.
°Mean percentage difference in concentration relative to the concentration farthest from the road (C15 - C8o)/C8o in Los Angeles,
CA and (Cio - Cioo)/Cioo in Detroit, Ml. Mean of all hourly concentrations in time of day and season specified.
dLos Angeles, CA data were collected 1/29/2009 to 3/11/2009 and 6/30/2009 to 8/19/09. Detroit, Ml data were collected 10/1/2011
to 12/31/2014.
Source: National Center for Environmental Assessment 2015 analysis of Los Angeles, CA data obtained from Polidori and Fine
(2012b) and Detroit, Ml data obtained from Air Quality System database.
2-58
-------
Figure 2-16 describes the evolving nature of NO2 concentrations and roadway gradients
during different seasons and hours of the day. NO2 roadway concentrations typically
increase during morning rush hour (6:00-10:00 a.m.) then gradually decrease from late
morning to mid-afternoon as the atmospheric mixing layer expands. Roadway NCh
concentrations begin to increase again during afternoon rush hour and nighttime, and are
generally similar to or slightly lower than NC>2 concentrations during morning rush hour.
This diurnal profile is more evident in the winter compared to the summer.
Notably, while maximum concentrations tend to occur during morning rush hour and
nighttime, especially during the winter, the NC>2 roadway gradient is largest during
afternoon hours (10:00 a.m.-5:00 p.m.). This trend is further demonstrated in
Figure 2-17. which shows the absolute difference in NC>2 concentrations between
near-road and downwind sites during winter and summer for Los Angeles, CA and
Detroit, MI. In both cities, the absolute difference between sites is below 15 ppb during
morning rush hour and nighttime, whereas a somewhat larger difference is observed
during mid-afternoon hours (12:00 p.m.-5:00 p.m.).
2-59
-------
Detroit Winter
Detroit Summer
10m 100m
0246
10 12 14 16 18 20 22
Hour
8 -
_
CL
Q.
10m
100m
0 2 4 6 8 10 12 14 16 18 20 22
Hour
8 -I
Los Angeles Winter
0 2 4 6 8 10 12 14 16 18 20 22
-S <=• H
Q. to
Los Angeles Summer
• 15m
80m
0 2 4 6 8 10 12 14 16 18 20 22
Hour
Note: NO2 = nitrogen dioxide. The box represents the interquartile range of concentrations observed during a given hour, with the
10th and 90th percentiles of concentrations shown by bottom and top whiskers, respectively. Red: within 15 m of a major interstate.
Gray: within 100 m of a major interstate.
Source: National Center for Environmental Assessment 2014 analysis of data obtained from Polidori and Fine (2012b) and Vette et
al. (2013).
Figure 2-16 Diurnal variation of differences in 1-hour nitrogen dioxide
concentrations 10-15 m and 80-100 m from the road in Los
Angeles, CA and Detroit, Ml.
2-60
-------
Detroit
Los Angeles
™ o
O co
o
o
O
E
o
Winter
Summer
_D
Q.
Q.
o
O CN
00
IT)
Winter
Summer
02468 11 14 17 20 23
Hour
02468 11 14 17 20 23
Hour
Note: NO2 = nitrogen dioxide. The box represents the interquartile range of concentrations observed during a given hour, with the
10th and 90th percentiles of concentrations shown by bottom and top whiskers, respectively. Red: winter. Gray: summer.
Source: National Center for Environmental Assessment 2014 analysis of data obtained from Polidori and Fine (2012b) and Vette et
al. (2013).
Figure 2-17 Absolute difference in 1-hour nitrogen dioxide concentrations
10-15 m and 80-100 m from the road in Los Angeles, CA and
Detroit, Ml.
The results for seasonal differences in 1-hour NCh concentration near and farther away
from the road are an important new contribution provided by the hourly data. A similar
seasonal pattern has also been observed in a few other studies using passive samplers
(Bell and Ashenden. 1997; Monn et al.. 1997). but the hourly data summarized in
Table 2-9 and in Figures 2-16 and 2-17 provides a more complete description of seasonal
and diurnal behavior. There have been a few recent observations of little or no variation
of NC>2 concentration with distance to the road for short time intervals before sunrise
(Gordon et al.. 2012; Massoli et al.. 2012). The data presented here based on more than
27,000 hours of NC>2 measurements in two cities build on these early studies to indicate a
clear trend of greater concentration differences between samples collected 10-15 m from
the road and those collected 80-100 m from the road during daytime than during
nighttime hours.
Concentration Dependence
The absolute concentration of NO2 also influences the magnitude of the road impact. In
studies with both one week or longer averaging times (Table 2-6). and one hour or shorter
2-61
-------
averaging times (Table 2-7); a few observations of NO2 concentrations more than 100%
higher at the location nearest the road than at the location farthest from the road were
reported, mostly when Cfar was much lower than usual. This is illustrated in Figure 2-18.
which shows that on a major road in a rural area of Great Britain (Bell and Ashenden.
1997). percentage differences in NO2 concentrations ranged up to 600%, but the greatest
differences were observed when Cfar was lower than usual. Differences were consistently
greater than 200% when Cfar was less than 4 ppb, but less than 100% when Cfar exceeded
10 ppb. Because Cfar was so low, even for the greatest differences in concentrations
observed by Bell and Ashenden (1997). the absolute difference in concentration between
distances of <1 m and 200 m never exceeded 20 ppb. Differences of similar magnitude
were observed by Bignal et al. (2007) for a British rural area where Cfar ranged from 5 to
10 ppb. Because data were collected in a rural area, the differences observed by Bignal et
al. (2007) would not necessarily be applicable for absolute differences that might be
observed in urban areas where NO2 concentrations are typically higher. Thus, Figure 2-18
clearly demonstrates that Cnear/Cfar at a lower concentration could be greater than Cnear/Cfar
observed at higher concentrations.
2-62
-------
700%
J 600%
•Is
X. 500%
ra
2
*g 400%
c
Ol
£ 300%
Q
01
CUO
2 200%
Ol
£ 100%
0%
(
«
4
•.
»
*
•
•
•
•
•/
9 •
) 2 4 6 8 10 12 14 16
Cfar(ppb)
Source: National Center for Environmental Assessment 2014 analysis of data from Bell and Ashenden (1997).
Figure 2-18 Influence of nitrogen dioxide concentration magnitude on the
ratio of nitrogen dioxide concentrations at <1 m from the road
(nearest concentration) to concentrations at 200-350 m (farthest
concentration) for 1-week averaging times in rural Wales.
This concentration effect is also evident for short averaging times presented in Table 2-7.
In Table 2-7. the greatest percentage differences in concentration between 10-15 m and
80-100 m distance from the road tend to occur at the times during which all sites
experience the generally lowest concentrations. For example, the greatest percentage
difference occurs April-September and 7:00 a.m. to 6:00 p.m. in Detroit, MI, the period
and location with the lowest concentration of all locations and time periods in Table 2-9.
In fact, Detroit, MI consistently had both the lowest concentrations and a greater
percentage difference than Los Angeles, CA at all time periods. Similarly, summer days
have the greatest percentage difference but lowest concentrations in both Los Angeles,
CA and Detroit MI, while winter nights have the highest concentrations and smallest
percentage differences. All of these observations concerning differences with location,
season, and time of day are also consistent with an inverse relationship between
2-63
-------
concentration and percentage difference in concentration with distance to road. This
relationship is clearly illustrated in Figure 2-19.
In Figure 2-19 the percentage difference in concentrations between measurements
10-15 m from the road and measurements 80-100 m from the road is plotted as a
function of concentration at 80-100 m at both Los Angeles, CA and Detroit, MI for the
same data summarized in Tables 2-8 and 2-9. Hourly near-road NCh concentrations are
sometimes several times higher than concentrations 80-100 m from the road, but only
when NC>2 concentrations at 80-100 m are below about 30 ppb. At 80-100 m NO2
concentrations greater than 33 ppb, near road concentrations are always less than 100%
higher than 80-100 m NC>2 concentrations, and when 80-100 m NC>2 concentrations are
greater than 50 ppb, near road NCh concentrations are always less than 50% higher than
80-100 m NC>2 concentrations. A smooth decrease of the upper limit of the percentage
difference in NC>2 concentration is evident in Figure 2-19. This pattern is consistent with
the concentration differences described in Table 2-7 and with earlier studies based on
passive sampling with longer averaging times described in Table 2-6 and Figure 2-18.
Low Cfar measurements do not explain all of the high ratios of (Cnear - Cfar)/Cfar in
Tables 2-6 and 2-7. In Table 2-6. Rodes and Holland (1981) observed percent differences
for (Cnear - Cfar)/Cfar ranging from 100 to 200% for averaging times of 1 h for average Cfar
concentrations of about 40 ppb, and attributed this to rapid formation of NO2 between the
road and monitor because of high Os concentrations. Most of the NOx emitted from
vehicles is emitted as NO, which can be rapidly converted into NO2 in the presence of Os
as described in Section 2.2. However, differences this large are not likely to be
representative of today's near road environment because at the time of the study, the
vehicle fleet was not strictly regulated for NOx emissions. In general, the observations in
Tables 2-6 and 2-7 indicate that NO2 concentrations nearest the road rarely appear to be
more than 100% higher than concentrations 80 to 400 m from the road for either 1-hour
or 1-week averaging times, except at very low concentrations.
To summarize, a zone of elevated NO2 concentration typically extends up to a distance of
200 to 500 m from roadways. NO2 concentrations for averaging times from 1 hour to
1 month measured 0-20 m from the road range up to 30 ppb higher or up to 100% higher
than concentrations measured at 80-400 m from a road, with greater differences during
daylight hours, in the summer, and at low concentrations. The difference in concentration
could be more strongly influenced by concentrations farther from the road than by
concentration nearest the road.
2-64
-------
500%
-100%
20 40 60
Concentration at 80 or 100 m (ppb)b
80
Note: NO2 = nitrogen dioxide. aPercentage difference in concentration relative to the concentration farthest from the road
(Cis - C8o)/C8o in Los Angeles, CA and (Cio - Cioo)/Cioo in Detroit, Ml, where Cn = concentration at n meters from the road.
""Concentration at 80 m from the road in Los Angeles, CA 100 m from the road in Detroit, Ml. Los Angeles, CA data collected from
1/29/2009 to 3/11/2009 and 6/30/2009 to 8/19/09 next to 1-710 freeway with heavy diesel traffic. Detroit, Ml data collected from
10/1/2011 to 12/31/2014 at Eliza Howell Park near-road monitoring site, 140,500 vehicles/day.
Source: National Center for Environmental Assessment 2015 analysis of Los Angeles, CA data obtained from Polidori and Fine
(2012b) and Detroit, Ml data obtained from Air Quality System database.
Figure 2-19 Percentage difference in 1-hour nitrogen dioxide concentration
between 10-15 m distance and 80-100 m distance from a road
with heavy traffic in two U.S. cities.
2.5.3.2 Near-Road Monitoring
The near-road monitoring network described in Section 2.4.5 was scheduled to be
implemented in three phases, with monitors in the first phase to become operational
January 1, 2014. As of July 2015, 56 monitoring sites were operational. Of these,
certified data for 2014 were available for 41 monitors in the Air Quality System database.
NO2 concentrations from this first year of near-road monitoring at these 41 sites are
2-65
-------
summarized and compared to concentrations at nonnear-road monitors in the same city in
Table 2-10. All near-road monitoring sites are within 50 m of a road with fleet equivalent
annual average daily traffic (FE-AADT) greater than 100,000 vehicles per day, and 57%
of them are within 20 m of the road. Many sites became operational after January 1, 2014
and did not accumulate a complete year of certified data. The number of days of available
data is also noted in Table 2-10. Because data presented here are for only a single year or
less, concentration trends and patterns should be considered very preliminary.
During 2014, 98th percentile daily 1-hour maximum NC>2 concentrations at all near-road
monitors in Table 2-10 were below the 1-hour daily maximum NAAQS of 100 ppb. No
near-road monitoring site had a 98th percentile 1-hour daily maximum NC>2 concentration
greater than 90 ppb or an estimated annual average concentration based on available data
of greater than 27 ppb. The highest 98th percentile 1-hour daily maximum concentrations
were observed for New York, NY; Denver, CO; Seattle, WA; and Los Angeles, CA, each
of which had concentrations greater than 65 ppb. At all other near-road monitors, 98th
percentile 1-hour daily maximum concentrations were less than 60 ppb.
High NC>2 concentrations were observed for near-road monitors with the highest traffic
counts. The three near-road monitors with target roads having FE-AADT of greater than
600,000 vehicles per day (New York, NY; Los Angeles, CA; and Phoenix, AZ) also had
among the six highest 98th percentile 1-hour daily maximum concentrations. The Seattle,
WA near-road monitor is targeting one of the highest FE-AADT counts in the network
and measured one of the highest 98th percentile 1-hour daily maximum concentrations.
Denver, CO and Houston, TX are important exceptions to this trend. In Denver, CO, the
second-highest 98th percentile daily 1-hour maximum concentration of all near-road
monitors was observed, but the target road FE-AADT was lower than that for most other
CBS As in Table 2-10. In contrast, the Houston, TX near road site targets one of the
highest FE-AADT counts among all near-road sites, but measured a 98th percentile
1-hour daily maximum concentration that was lower than that for most other CBSAs in
Table 2-10. Overall, the very highest 98th percentile 1-hour maximum concentrations
were generally observed at the monitors adjacent to roads with the highest traffic counts.
2-66
-------
Table 2-10 Comparison of nitrogen dioxide concentrations at U
non-near-road monitors for 2014.
Annual Average
(PPb)
Number
CBSAa of Days
New York, NY
Denver, CO
Seattle, WA
Los Angeles, CA
Cincinnati, OH
Phoenix, AZ
Indianapolis, IN
Boston, MA
Milwaukee, Wl
San Jose, CA
San Francisco,
CA
Providence, Rl
Baltimore, MD
Philadelphia, PA
Detroit, Ml
Nashville, TN
Birmingham, AL
St. Louis, MO
Atlanta, GA
Hartford, CT
Minneapolis, MN
Austin, TX
Houston, TX
New Orleans, LA
Columbus, OH
Kansas City, MO
San Antonio, TX
268
354
236
355
356
321
318
351
359
122
330
271
275
353
357
166
358
355
199
354
362
247
331
283
365
357
345
Near-
Road
19
25
24
27
23
21
17
17
16
20
17
20
18
16
16
15
14
14
20
14
16
14
13
12
12
12
11
Nonnear-
Roadb
4-22
18-23
12
8-22
4-11
9-25
9-14
4-17
10
13
3-14
1-10
11-16
6-18
12
10
9
5-12
3-11
9
5-9
5
2-13
7
10
11-13
5-6
1-h Maximum 98th
Percentile (ppb)
Near-
Road
90
70
69
66
59
59
58
53
53
52
52
51
51
51
51
51
51
50
50
49
48
48
48
48
47
46
46
Nonnear-
Roadb
41-70
64-73
47
40-85
31-45
37-64
46-49
25-62
43
55
17-58
12-44
47-52
34-59
49-52
40
41
34-45
17-53
45
28-50
31
18-52
42
51
51-53
31-37
.S. near-road and
Highest 1-h
Maximum (ppb)
Near-
Road
258
97
91
79
68
62
64
64
62
65
65
56
56
65
62
63
67
72
58
80
53
57
55
64
53
52
51
Nonnear-
Roadb
51-90
71-136
60
52-136
40-50
57-102
54-58
31-68
62
58
21-84
22-50
54-62
43-73
65-66
43
83
41-54
23-58
60
43-70
37
23-98
56
63
63-78
38-48
Near-
Road
AADT
311,234
249,000
237,000
272,000
163,000
320,138
189,760
198,239
133,000
191,000
216,000
186,300
186,750
124,610
140,500
144,204
141,190
159,326
284,920
159,900
277,000
188,150
324,119
68,015
142,361
114,495
201,840
Near-
Road
FE-AADT0
612,212
263,118
471,630
695,776
386,380
624,315
362,110
251,761
133,000
294,140
424,008
416,790
452,309
257,460
188,200
338,879
215,527
360,077
406,256
231,855
387,250
350,712
496,226
129,229
286,050
347,582
405,295
2-67
-------
Table 2-10 (Continued): Comparison of nitrogen dioxide concentrations at U.S.
near-road and nonnear-road monitors for 2014.
Annual Average 1-h Maximum 98th Highest 1-h
(ppb) Percentile (ppb) Maximum (ppb)
CBSAa
Richmond, VA
Louisville, KY
Tampa, FL
Boise, ID
Jacksonville, FL
Memphis, TN
Pittsburgh, PA
Dallas, TX
Buffalo, NY
Portland, OR
Charlotte, NC
Cleveland, OH
Raleigh, NC
Des Moines, IA
Number
of Days
262
226
258
239
267
183
121
273
268
225
121
152
313
349
Near-
Road
14
13
12
12
12
12
13
10
10
12
11
10
10
9
Nonnear-
Roadb
5-8
11
5
NAd
8
8
3-11
3-10
9
8
9
12
12
6
Near-
Road
45
45
45
43
44
44
40
40
40
38
38
36
36
35
Nonnear-
Roadb
37-44
49
30
NAd
40
42
21-45
24-28
55
35
41
48
48
35
Near-
Road
54
70
59
48
70
48
42
58
50
49
44
45
45
41
Nonnear-
Roadb
47-56
75
36-79
NAd
47
53
24-56
29-63
71
40
51
66
66
47
Near-
Road
AADT
151,000
163,000
190,500
103,000
139,000
140,850
87,534
235,790
131,019
156,000
153,000
153,660
141,000
110,000
Near-
Road
FE-AADT
259,720
247,600
327,660
162,000
304,062
292,968
148,248
431,027
NAd
289,052
260,830
287,580
203,280
150,140
AADT = annual average daily traffic; CBSA = core-based statistical area; FE-AADT = fleet-equivalent annual average daily traffic;
NA = not available; NY = New York; CO = Colorado; WA = Washington; CA = California; OH = Ohio; AZ = Arizona; IN = Indiana;
MA = Massachussetts; Wl = Wisconsin; Rl = Rhode Island; MD = Maryland; PA Pennsylvania; Ml = Michigan; TN = Tennessee;
AL = Alabama; MO = Missouri; GA = Georgia; CT = Connecticut; MN = Minnesota; TX = Texas; LA = Louisiana; VA = Virginia;
KY = Kentucky; FL = Flroida; ID = Idaho; OR = Oregon; NC = North Carolina; IA = Iowa.
aA core-based statistical area is a U.S. geographic area that centers on an urban center and adjacent areas that are
socioeconomically tied to the urban center by commuting. For CBSAs that are identified by more than one urban center, only the
first city used to identify the CBSA is used, without regard to monitor location. For example, the San Francisco-Oakland-Hayward
CBSA is identified in the table as San Francisco, CA even though the near-road monitor is in Oakland, CA. CBSAs are listed in
decreasing order of 98th percentile daily 1-hour maximum concentration.
bNonnear-road monitors are all monitors that report data to the Air Quality System database that do not meet criteria for
near-road monitors. These can be intended to be representative of area wide (AW), near source, or background concentrations.
Data are reported for the range of concentrations across nonnear-road monitors in a city or the concentration at the single
nonnear-road monitor.
°FE-AADT = (AADT- HDC) + (HDm x HDC) where AADT is annual average daily traffic, HDC is total number of heavy-duty vehicles
on a road segment, HDm is a multiplier [estimated as 10; (U.S. EPA, 2012b)1 that represents heavy-duty to light-duty emission
ratios on the road segment.
dBoise, ID does not have a nonnear-road monitor. Buffalo, NY does not have fleet equalivalent annual average daily traffic count.
Source: National Center for Environmental Assessment and Office of Air Quality Planning and Standards 2014 analysis of Air
Quality System network data.
The highest near-road annual average NO2 concentrations were observed at Los Angeles,
CA (27 ppb); Denver, CO (25 ppb); and Seattle, WA (24 ppb). In New York, NY, the
annual average concentration was considerably lower (19 ppb), but winter concentrations
were not included because the site was not operational until April 1. Annual average
concentrations of 20 ppb or greater were also observed at Cincinnati, OH; Phoenix, AZ;
2-68
-------
San Jose, CA; Providence, RI; and Atlanta, GA. At all other near-road monitors, annual
average concentrations were less than 20 ppb. For context, the only other monitors in the
national network with annual average NC>2 concentrations greater than 20 ppb in 2014
were two other sites within 200 m of highways with more than 200,000 vehicles per day
(Greenwood in Phoenix, AZ and Elizabeth Lab in Elizabeth, NJ), and several
nonnear-road monitors in Los Angeles, CA and Denver, CO. Because annual average
concentrations greater than 20 ppb were only observed at either near-road network
monitors (or other monitors strongly influenced by heavy traffic) or in the Denver, CO
and Los Angeles, CA CBSAs, it is interesting that the very highest annual average NO2
concentrations observed nationwide in 2014 were at the Los Angeles, CA and Denver,
CO near-road monitors.
For those CBSAs that have a near-road monitor and at least one nonnear-road monitor,
annual average concentrations were usually higher at near-road sites than at nonnear-road
counterparts within the same CBSA. This was the case even though approximately half of
the near road sites (i.e., those sites in Table 2-10 that operated for less than approximately
270 days) were not yet operational during the winter months, when concentrations are
likely to be highest (see Table 2-9).
In almost half of the CBSAs in Table 2-10. both the highest 98th percentile 1-hour daily
maximum concentration in the CBSA and the highest annual average concentration in the
CBSA were observed at the near-road monitoring site. In most of the remaining CBSAs,
the highest annual average concentration was observed at the near-road site, but not for
the highest 98th percentile 1-hour daily maximum concentration value of the available
data. The highest estimated annual average concentration was observed at the near-road
monitor in more than 80% of the CBSAs.
The differences between near-road and nonnear-road concentrations in Table 2-10 are not
directly comparable to the differences observed in the near-road gradient studies
discussed in Section 2.5.3.1. The range of nonnear-road concentrations in Table 2-10
includes observations not only from monitors sited to measure typical concentrations in
areas of high population, but also monitors sited to determine the highest concentration
expected to occur in the area, or to determine the impact of other significant sources. An
analysis of monitor siting prior to implementation of near-road monitoring requirements
indicated that across the entire network 177 monitors were sited for general population
exposure, 58 to measure the highest concentration in the area, 69 to measure general or
upwind background concentrations, and 19 for source-oriented measurements (U.S. EPA.
2010a). It should be noted that any monitoring site can have multiple of these monitoring
objectives, as they are not mutually exclusive. In that context, Table 2-10 indicates how
2-69
-------
near-road concentrations fit into a wider range of urban concentrations rather than how
they compare to an urban background with less traffic influence.
The Los Angeles, CA CBSA provides an example. It contains one of the busiest ports as
well as one of the busiest airports in the U.S. (Section 2.5.3.3). Out of 18 monitors in the
Los Angeles, CA CBSA, three of the five highest 98th percentile 1-hour maximum
concentrations were observed at the near-road site, the site nearest the port, and the site
adjacent to the airport. At the LAX Hastings monitoring site adjacent to Los Angeles
International Airport, the 98th percentile 1-hour daily maximum NC>2 concentration for
2014 was 66 ppb, identical to the concentration at the Los Angeles, CA near-road site,
although the annual average concentration of 12 ppb was much lower. The highest 1-hour
NO2 concentration in the Los Angeles, CA CBSA in 2014 (136 ppb) was observed at
Long Beach North monitor, the site closest to the port of Long Beach, CA. The 98th
percentile 1-hour daily maximum concentration at the site was also the highest in the Los
Angeles, CA CBSA (85 ppb), and far exceeded the 98th percentile 1-hour daily
maximum concentration at the near-road site. However, in Los Angeles, CA as in most of
the CBSAs with near-road monitors, the annual average concentration was highest at the
near-road monitor.
Many of the nonnear-road concentration ranges in Table 2-10 also include concentrations
that are unusually low for urban areas. The lowest values of the range are more indicative
of whether the CBSA contains monitors sited for background measurements than how
concentrations compare among CBSAs. For example, in New York, NY, annual average
nonnear-road NC>2 concentrations range from 4 to 22 ppb, while in Denver, CO they
range from 18 to 23 ppb. This is not an indication that concentrations are much lower in
New York, NY than those in Denver, CO. Rather, the Chester monitoring site is in a rural
area of New Jersey upwind of the New York, NY CBSA and is identified as a
background site, and its annual average NO2 concentration in 2014 was 4 ppb. This
concentration is much lower than the near-road concentration of 19 ppb for the New
York, NY CBSA, but also much lower than the concentration ranges for many other
CBSAs in the U.S. in Table 2-10. Without the two designated background sites for the
New York, NY CBSA, the range of nonnear-road concentrations would be 16 to 22 ppb,
more similar to Denver, CO, which does not have a background monitor.
Nonnear-road monitors can also be influenced by traffic. One of the highest 1-hour daily
maximum NO2 concentrations was 136 ppb, which was observed at a Denver, CO
nonnear-road site. As indicated in Table 2-10. this is much higher than the maximum
1-hour concentration of 97 ppb observed at the Denver, CO near-road monitor. The
136-ppb concentration was observed at the Childhood Asthma Management Program
(CAMP) monitor located approximately 3 km downwind of the near-road monitor, but
2-70
-------
one block from high-rise buildings that form the edge of the high-density central business
district (which lies between the two monitors). At the CAMP monitor, local traffic is also
a likely source, in addition to commercial heating and other activities. Recent traffic
counts on the nearest streets to the CAMP monitor, Broadway and 22nd Street, were
44,850 (in 2014) and 23,389 (in 2013) vehicles per day, respectively. Traffic counts on
other streets within one block were 22,000 (20th St.), 13,000 (Park Ave.), 5,000 (Champa
St.), and 2,490 (Curtis St.) vehicles per day according to the Denver Regional Council of
Governments data.1 This adds up to more than 100,000 vehicles per day on streets within
one block of this nonnear-road monitor.
While the near-road network has not been operating long enough to evaluate long-term
trends in near-road concentrations, there are older monitors in the U.S. that are
informative, even though they do not strictly meet new requirements for near-road
monitoring. The Elizabeth Lab site in Elizabeth, NJ does not meet near-road monitoring
requirements because it is more than 50 m from the road. Some of the highest NC>2
concentrations in the U.S. have been observed at this site, and long-term NCh
concentration trends are described in Section 2.5.5.
Outside of the U.S. (e.g., London, U.K.), routine near-road monitoring has been
conducted for a longer time. The preliminary results from the U.S. near-road network are
similar to data from the London, U.K. network, despite potential differences from the
U.S. in fleet mix (including fraction of vehicles with diesel engines), distance from road,
traffic mitigation policies, and small geographic scope that may limit generalizability.
London. U.K. data were analyzed because the city has a well-established system of
roadside and urban background monitors. Air quality data were obtained from the
Airbase database (EIONET. 2014) for 2004 to 2006 and 2010 to 2012 in the form of
hourly NC>2 measurements, and monitors of interest were those whose city was listed as
London and were within 10m of the roadway to capture NC>2 primarily derived from
mobile sources. The site with the highest concentration, Marylebone Road, had a traffic
count of 70,000 vehicles per day (Dall'Osto et al., 2011). and was within 2 m of a road, or
close enough to approximate on-road conditions. Overall, there were large differences in
NO2 concentrations between roadside and urban background monitors, which ranged
from 2.4 to 9.8 km apart as shown in Tables 2-11A and 2-11B. The differences in
24-h avg NO2 concentrations ranged from approximately 24% lower to 170% higher at
the roadside than urban background site. The largest relative differences in 24-h avg NO2
concentrations were observed when the ambient urban background concentrations were
less than 20 ppb. NO2 concentrations at all roadside monitors were positively correlated
with concentrations at the overall urban background monitors. Interquartile ranges were
http://gis.drcog.org/trafficcounts/.
2-71
-------
generally similar between roadside monitor-urban background monitor pairs, indicating
that while in the majority of cases roadside monitors had higher NCh concentrations than
urban background monitors, temporal variability was similar between the two monitors.
As with the preliminary results from the U.S. near-road network, the results for London.
U.K. suggest that while NO2 concentrations measured at roadside monitors were
generally higher than those measured at urban background monitors, there was a wide
range in mean differences between roadside and background.
2-72
-------
Table 2-11 A Roadside and urban background nitrogen dioxide concentrations in
Monitor Pairs
Roadside
Urban bkg
Roadside
Urban bkg
Roadside
Urban bkg
London Marylebone Rd
London Bloomsbury
Camden Kerbside
London N. Kensington
Haringey Roadside
London Bloomsbury
Distance
between
Monitors
(km)
2.4
2.8
9.8
Mean
Concentration3 A Mean
(PPb) (%)
52.3 68
31.2
43.8 124
19.6
23.8 -24
31.2
98th Percentile
of 1 -Hour Daily A 98th
Maximum Percentile
(PPb) (%)
140.5 102
69.7
132.0 108
63.6
64.2 -7.9
69.7
London
24-Hour
Avg IQR
(PPb)
25.7
12.0
16.9
12.6
12.0
12.0
, U.K. 2010-2012.
1-Hour
Max IQR
(PPb)
59.6
14.4
32.4
16.0
18.6
14.4
24-Hour Avg
Correlation with
Urban Background
Monitors (95% Cl)
0.30
(0.25, 0.36)
0.74
(0.71,0.77)
0.84
(0.83, 0.86)
A = difference between roadside and urban background monitors; avg = average; bkg = background; Cl = confidence interval; IQR = interquartile range.
a3-year average.
Source: National Center for Environmental Assessment 2014 analysis of European Air Quality Database data from 2010-2012.
2-73
-------
Table 2-11 B Roadside and urban background nitrogen dioxide concentrations in London, U.K.
Distance 98th Percentile
between Mean 1 -Hour of 1-Hour Daily
Monitors Concentration A Mean3 Maximum13
Monitor Pairs (km) (ppb) (%) (ppb)
Roadside
Urban bkg
Roadside
Urban bkg
Roadside
Urban bkg
Roadside
Urban bkg
Roadside
Urban bkg
Roadside
Urban bkg
Roadside
Urban bkg
Roadside
Urban bkg
London Marylebone Rd 2.4
London Bloomsbury
Southwark Roadside 3.3
London Eltham
London Cromwell Rd 2 3.4
London Bexley
Camden Kerbside 3.8
London N. Kensington
Tower Hamlets Roadside 4.1
London Hackney
Haringey Roadside 4.5
London Hackney
London Bromley 5.2
London Lewisham
London A3 Roadside 6.2
London Teddington
58.3 88
31.1
33.0 96
16.84
42.7 131
18.5
37.0 74
21.2
32.3 26
25.7
24.4 -5
25.7
24.7 -5
26.0
35.3 170
13.1
163.5
70.6
80.7
53.7
97.7
64.5
116.3
71.8
80.1
91.2
66.1
91.2
80.4
87.4
93.0
53.5
A 98th 24-Hour
Percentile Avg IQR
(%) (PPb)
132 28.6
13.1
50 11.1
10.7
51 12.1
11.9
62 15.7
12.7
-12 15.2
14.3
-28 12.0
14.3
-8 12.4
12.6
74 14.7
11.2
2004-2006.
1-Hour
Max
IQR
(PPb)
45.5
16.0
13.3
16.5
21.0
16.0
26.9
17.0
19.2
19.7
16.5
19.7
17.6
17.0
19.2
19.2
24-Hour Avg
Correlation
with Urban
Background
Monitors
(95% Cl)
0.24
(0.18, 0.29)
0.86
(0.84, 0.88)
0.63
(0.59, 0.66)
0.78
(0.75, 0.80)
0.81
(0.79, 0.83)
0.80
(0.78, 0.82)
0.70
(0.67, 0.73)
0.64
(0.77, 0.81)
A = difference between roadside and urban background monitors; avg = average; bkg = background; Cl = confidence interval; IQR = interquartile range.
aRoadside vs. urban background comparison.
b3-year average.
Source: National Center for Environmental Assessment 2014 analysis of European Air Quality Database data from 2004-2006.
2-74
-------
While NC>2 measurements are more widely used than NOx for exposure estimates in
epidemiologic studies, NO2 accounts for only a fraction of NOx near roads with heavy
traffic. For example, Clements et al. (2009) measured concentrations of NO, NO2, and
NOx, 5 m downwind from a state road in Austin, TX, and observed NOx concentrations
of approximately 40-50 ppb, NO concentrations of approximately 15-40 ppb, and NO2
concentrations of approximately 5-15 ppb under downwind conditions. NO2 accounted
for 10-38% of the NOx.
It follows that NO is often a greater contributor to NOx near roads. Baldauf etal. (2008a)
presented a time series of pollutants that were measured 5 m from 1-40 in Raleigh, NC,
and reported that NO concentrations reached near 250 ppb between 8:00 a.m. and
9:00 a.m., with minimum NO concentrations around 50 ppb during that time period. The
predominance of NO (rather than NO2) in the near-road environment contrasts with
nationwide annual average concentrations in Table 2-4. for which NO2 (rather than NO)
accounts for more than 60% of the annual average ambient concentration of NOx.
Wind speed and atmospheric stability also impact roadway NOx concentrations. Peak
roadway concentrations are often observed during presunrise hours when winds are weak
and atmospheric inversions are present (Gordon etal.. 2012; Durantet al.. 2010; Hu et
al.. 2009). During these presunrise hours, the NOx concentrations exhibit a more gradual
decay from the roadway than after sunrise. Hu et al. (2009) observed this effect for NO
during a near-road field campaign in Santa Monica, CA. They observed elevated NO
concentrations (90-160 ppb) as far as 1,200 m downwind of the roadway during
presunrise hours, which is much larger than the expected spatial extent of NO
(100-300 m; Section 3.3.1.1). NOx concentration gradients continue to change
throughout the day as atmospheric stability evolves. After sunrise, near-road NOx
concentrations drop as vertical mixing increases (Gordon etal.. 2012; Durantetal.. 2010)
until concentrations reach a minimum during the late afternoon (Gordon et al.. 2012). In
some studies, no clear gradient is observed in NOx concentrations (or other traffic-related
species) during mid-morning or early evening hours (Gordon et al.. 2012; Durant et al..
2010). However, the exact response of the horizontal concentration gradient to changes in
boundary layer height is unresolved to some extent.
Dispersion of NOx in the near-road environment is influenced by several factors:
atmospheric turbulence, vehicle-induced turbulence, and roadway-induced turbulence
(Baldauf et al.. 2009; Wang and Zhang. 2009). Atmospheric turbulence occurs because of
meteorological factors within the urban boundary layer. Vehicle-induced turbulence
results from the air disturbances caused by the direction and speed of vehicle motion.
Roadway-induced turbulence happens when wind-driven air masses undergo separation
following impact with a roadway structure in the built environment. These sources of
2-75
-------
turbulence interact with each other to create complex, unique dispersion profiles at a
given road segment to influence NOx concentrations. This discussion addresses the
physical factors influencing dispersion of NOx.
Several atmospheric conditions affect regional or urban airflow profiles and potentially
can impact the dispersion profile of NOx even in the absence of adjacent buildings,
roadway structures, or traffic-related turbulence. In urban areas, effects of the built
environment can be seen at regional-, urban-, neighborhood-, and street-level scales
(Fernando. 2010; Britter and Hanna. 2003). Roughness created by upstream buildings
contributes to local turbulence levels, even in the absence of adjacent buildings. Land
forms such as slopes and valleys can affect the atmospheric turbulence level because they
interact with atmospheric stability conditions to restrict air movement. Finnetal. (2010)
observed that tracer gas concentration increased with increasing atmospheric stability.
This finding is consistent with results with other studies (Gordon et al.. 2012; Durant et
al., 2010; Hu et al.. 2009) that observed the highest concentrations of NO, NO2, and NOx
before sunrise when traffic levels and atmospheric stability are high. Hu et al. (2009) also
argued that atmospheric stability potentially extends the decay profile of near-roadway
pollutants. Additionally, the presence of slopes and valleys can cause spots where airflow
converges or diverges (Fernando. 2010). Heat flux can be sizeable in urban areas where
the "heat island" effect from roadways and buildings can raise local temperatures by
several degrees (Britter and Hanna. 2003); heat flux potentially contributes to convection
near roadways and other structures in the built environment. Underscoring the dominant
role of local turbulence on dispersion patterns, Venkatram et al. (2007) measured
meteorological factors potentially affecting NO concentrations near a road segment in
Raleigh, NC and found that among meteorological variables, vertical velocity
fluctuations had the largest effect on NO concentration.
Vehicle motion creating high levels of turbulence on and near roads can contribute to the
dispersion of traffic-related air pollution in the vicinity of a roadway (Baldauf et al..
2008a). An early description of this was provided by Sedefian et al. (1981) for the
General Motors experiments, in which groups of vehicles were driven along a test track
while towers with mounted anemometers measured mean and fluctuating velocities. It
was observed that vehicle-induced turbulence dissipates slowly under low mean wind
conditions and vice versa. Vehicle-induced turbulence was found in that study to
contribute to vertical dispersion of emitted pollutants. Computational fluid dynamics
(CFD) simulations by Wang and Zhang (2009) also found that vehicle-induced
turbulence contributed to vertical dispersion. Rao et al. (2002) observed large
measurements of turbulence kinetic energy in the wake of a vehicle outfitted with a trailer
carrying sonic anemometers driving along a runway. Sedefian et al. (1981) found that
advection of vehicle-induced turbulence away from the roadway was related to the speed
2-76
-------
and direction of mean winds, di Sabatino et al. (2003) showed that vehicle-induced
turbulence is related to traffic levels. In light traffic, the wake behind a vehicle is isolated,
but for increasing traffic, the wakes interact and turbulence is a function of the number of
vehicles and vehicle length scale. At congested traffic levels, the vehicle-induced
turbulence becomes independent of the number of vehicles. For street canyon simulations
and measurements, Kastner-Klein et al. (2003) observed that predictions of tracer
concentrations were overestimated when vehicle-induced turbulence was not considered;
this implies additional dispersion related to vehicle-induced turbulence. Traffic
directionality was investigated by He and Dhaniyala (2011) and Kastner-Klein et al.
(2001). He and Dhaniyala (2011) observed that turbulent kinetic energy from two-way
traffic was roughly 20% higher than that for one-way traffic, and increased with
decreasing distance among the traffic lanes. Kastner-Klein et al. (2001) observed that
two-way traffic suppresses the mean flow of vehicle-induced air motion along a street
canyon, whereas one-way traffic produces a piston-like effect [note that the Kastner-
Klein etal. (2001) study was for the geometrical case of a street canyon]. Substantially
higher turbulence levels were produced with two-way traffic compared with one-way
traffic for the Kastner-Klein et al. (2001) study as well.
The presence of near-road structures results in recirculating airflow regions that may trap
air pollutants on one side and disperse them on another side, depending on wind
conditions (Baldauf et al.. 2008b). Finn etal. (2010) simulated transport from a roadway
using a point source tracer gas with barrier and open terrain conditions. With airflow
from the simulated roadway and high atmospheric stability, high concentrations were
trapped in the roadway region with a negligible tracer gas in the wake downstream of the
barrier with considerable lateral and vertical plume dispersion. For open terrain, transport
of the tracer was characterized by a narrow plume. Hagler et al. (2011) used CFD to
model airflow and concentrations around barriers of different heights and similarly found
reductions in inert tracer concentration downwind of the barrier compared with the open
terrain case with trapping of air pollutants upstream of the barrier. With the barrier in
place, downwind tracer concentrations were observed at elevations of twice the barrier
height. Mean airflow vectors also illustrate a wind disturbance at elevations of twice the
barrier height. Even for the open terrain case, vertical dispersion occurs. In additional
simulations involving a service readjust downstream of the barrier, Hagler etal. (2011)
observed entrainment of tracer in the wake downstream of the barrier. Tokairin and
Kitada (2005) used CFD to investigate the effect of porous fences on contaminant
transport near roads and observed tracer gas retention and airflow recirculation when the
fences were designed with less than 40-50% porosity. Heist et al. (2009b) investigated
the effect of geometry of road cuts and noise barriers in wind tunnel tracer gas
experiments. They observed that elevated roadways, depressed roadways, and noise
barriers all resulted in lower downwind concentrations compared with the open terrain
2-77
-------
case with elevated roadways producing the least reduction in concentration. As in Hagler
et al. (2011). Heist et al. (20091)) observed measurable concentrations at elevations that
resulted from Gaussian dispersion for all geometries of the road cut or barrier, but
vertical dispersion was enhanced or dampened depending on the specific geometry.
Similarly, for wind tunnel simulations of a single tower above a matrix of street canyons,
the tower was shown to induce both airflow and tracer concentration along the leeward
edge of the building to a height exceeding the tower height (Brixey et al., 2009; Heist et
al.. 2009a).
For the special case of street canyons, retention time for traffic-based pollution increases
on the roadway with increasing building height-to-road-width ratio because recirculating
airflow forms closed streamlines within the canyon (Li et al.. 2005; Liu et al.. 2005). For
wind tunnel simulations of tracer emission at street level with and without traffic,
Kastner-Klein et al. (2001) observed measurable tracer concentrations near the top of the
street canyon but with some dispersion from maximum tracer levels at the canyon floor.
Dilution of NOx concentrations through these recirculating air structures leads to a steep
decrease in concentration with increasing distance from the ground (Lee et al.. 2012a).
For low-aspect-ratio street canyons, secondary recirculating structures can arise; while
contaminant retention still occurs in this case, ventilation occurs more readily than for the
high-aspect-ratio case (Simoens and Wallace. 2008; Simoens et al., 2007). Cheng et al.
(2008) used CFD to evaluate factors leading to contaminant retention in street canyons
and observed that the exchange rate for air and a tracer gas was driven by the turbulent
component of airflow at the roof-level interface of the street canyon. Subsequent
simulations showed that exchange rate was also aided by unstable atmospheric conditions
(Cheng et al.. 2009b). CFD simulations by Gu et al. (2010) of transport within a street
canyon with and without vegetation suggested that the recirculating flow is dampened by
the presence of vegetation.
2.5.3.3 Monitoring Near Nonroad Sources
Compared to near-road monitors, fewer monitors are located near other major sources. In
rare cases, monitors are adjacent to or within hundreds of meters of a major source, and
additional monitors are located within a few kilometers. Table 2-12 summarizes NO2
concentrations at selected monitoring sites that are likely to be influenced by nearby
ports, airports, border crossings, petroleum refining, or oil and gas drilling.
2-78
-------
Table 2-12 Selected nitrogen dioxide measurements with potential nonhighway
source influences for 2014.
Annual Average
(PPb)
Monitoring Site
Hudson
Bayonne, NJ
NOAA Storage
Facility
LAX Hastings
Schiller Park, IL
Otay Mesa
Calexico, CA
Chamizal
Fairbanks, AK
Capitol
Roosevelt
National Park,
ND
Vernal, UT
Potential
Influence(s)
Port
Port (Newark)
Port (Norfolk)
rail yard
Airport
Airport (O'Hare)
rail yard
Border crossing
Border crossing
Border crossing
Wood burning
Petroleum
refinery
Oil and gas
drilling
Oil and gas
drilling
CBSA Site
Los Angeles, CA 20.7
New York, NY 17.1
Norfolk, VA 8.3
Los Angeles, CA 11.9
Chicago, IL 19.0
San Diego, CA 17.8
None 12.3
El Paso, TX 14.0
None 10.7
Baton Rouge, .n _
LA 1U'&
None 1.6
None 7.3
Range
across
CBSA
8-27
4-22
NA
8-27
10-21
5-20
NA
12-14
NA
2-11
NA
NA
98th Percentile
of 1-Hour Daily
Maximum (ppb)
Site
85
61
42
66
58
70
62
60
75
46
14
88a
Range
across
CBSA
40-85
41-70
NA
40-85
50-67
25-70
NA
54-60
NA
13-49
NA
NA
Highest 1-Hour
Daily Maximum
(PPb)
Site
136
75
56
87
105
87
94
71
108
59
89
88a
Range
across
CBSA
52-136
51-90
NA
52-136
66-105
48-87
NA
69-79
NA
20-93
NA
NA
AK = Alaska; CBSA = core-based statistical area; IEPA = Illinois Environmental Protection Agency; LAX = Los Angeles Airport;
Max = maximum; NA = None available; NOAA = National Oceanic and Atmospheric Administration.
aFor Vernal, UT, maximum and 98th percentile concentrations are the same because there are so few measurements.
Source: National Center for Environmental Assessment 2015 analysis of Air Quality System network data.
Three of the sites in Table 2-12 are near major ports for commercial shipping. The ports
of Long Beach, CA (in the Los Angeles, CA CBSA); New York-Newark, NY-NJ (in the
New York CBSA); and Norfolk, VA are among the busiest ports in the U.S. The Long
Beach-Hudson monitor is approximately 3 km from the port of Long Beach, CA. This
monitor had the highest nationwide 98th percentile 1-hour daily maximum NCh
2-79
-------
concentration in 2014 (85 ppb) among sites that met completeness criteria and the highest
hourly NC>2 concentration in the Los Angeles, CA CBSA (136 ppb). However, other
sources, including the 1-710 freeway with heavy diesel traffic are also nearby. The
Bayonne, NJ monitor is approximately 1 km directly across Newark Bay from the Port of
Newark. At Bayonne, NJ, NO2 concentrations were higher than those at most near-road
monitors listed in Table 2-10 but similar to other sites in the New York, NY CBSA. The
Norfolk NOAA Storage Facility monitor, VA is located approximately 1 km across the
Elizabeth River from the Portsmouth Marine Terminals and approximately 1 km from the
Norfolk Southern rail yard. Here, NC>2 concentrations were substantially lower than those
at most near-road monitors in Table 2-10. Altogether, a wide range of NO2
concentrations was observed for three sites near major ports, and it is difficult to
generalize the impact of port emissions on concentrations at nearby monitors.
There are two NO2 monitoring sites in Table 2-12 located within 1 km of two of the
busiest airports in the U.S., Los Angeles International Airport (LAX) and O'Hare
International Airport in Chicago, IL. The 98th percentile 1-hour daily maximum
concentration at LAX was identical to that at the near-road monitor, although annual
average concentration was much lower. The highest hourly concentration in the Chicago,
IL CBSA in 2014 (105 ppb) was observed at the Schiller Park, IL monitoring site located
adjacent to the airport. However, this is also very close to a major rail yard, the Bedford
Park Rail Yard. As with ports, it is difficult to isolate the impact of airports on NC>2
concentrations at nearby monitors.
From Table 2-12. three of the highest 98th percentile 1-hour daily maximum NC>2
concentrations in the U.S. were observed at monitors near the U.S.-Mexico border at
Otay Mesa in the San Diego, CA CBSA (70 ppb); Chamizal, in the El Paso, TX CBSA
(60 ppb); and Calexico, CA (62 ppb). Each of these sites are within 4 km of one of the
five busiest ports of entry to the U.S. for international truck traffic in 2014.: Data from
the Otay Mesa site were instrumental in demonstrating that concentrations of traffic
pollutants in the San Ysidro community surrounding the Otay Mesa border crossing are
related to wind direction and border crossing wait times (Quintana et al.. 2014). The
Chamizal site is only 0.2 km from the principle border crossing in El Paso, TX, a city
where 81% of the variance in NO2 concentration has been attributed to elevation, distance
from highways and ports of entry (Gonzales et al.. 2005). Idling vehicles at the Calexico,
1 Otay Mesa 810,193 trucks in 2014; El Paso 759,125 trucks in 2014; Calexico East 325,243 trucks in 2014; from
U.S. Department of Transportation, Bureau of Transportation Statistics, accessed October 13, 2015.
http://transborder.bts.gov/programs/international/transborder/TBDR BC/TBDR BC Index.html
2-80
-------
CA border crossing have been proposed as a potentially important source of NO2 in that
community1 and efforts are underway to quantify their contribution to local pollution.2
The only other monitoring site in the U.S. with 98th percentile 1-hour daily maximum
NO2 concentrations greater than 60 ppb that is not in a large urban area is the Fairbanks,
Alaska (AK) Ncore site. PIVb 5 in Fairbanks, AK is primarily due to wood smoke,3 which
can also be an important local NOx source (U.S. EPA. 2013a). However, Fairbanks is
also impacted by coal-fired power generation, motor vehicles, and oil-fired heating
systems,4 and the high concentrations are enhanced by reduced reactivity of NOx in the
darkness and extremely cold temperatures characteristic of Fairbanks in winter (Joyce et
al.. 2014).
Monitoring data from near a major petroleum refinery and in areas with oil and gas
drilling activities are also included in Table 2-12. These NO2 concentrations are generally
lower than those influenced by other sources in Table 2-12. There are occasional high
hourly concentrations but 98th percentile and annual average concentrations are generally
not as high as those observed near roads in Table 2-10 or at sites influenced by other
sources in Table 2-12. For example, at Theodore Roosevelt National Park, ND, a
maximum NC>2 concentration of 89 ppb was observed, but the 98th percentile 1-hour
daily maximum concentration was only 14 ppb. The Vernal, UT monitor in Table 2-12
was only operated for 27 days in 2014, and the 98th percentile/highest daily 1-hour
maximum concentration was 88 ppb. However, the next highest 1-hour daily maximum
concentration at this site was only 47 ppb.
In general, it is more difficult to assess the impact of nontraffic sources using the national
monitoring network because there are few monitors near major sources and they are
located at a greater distance from sources than near-road monitors. However, 98th
percentile 1-hour daily maximum concentrations at monitoring network sites near ports,
airports, and border crossings have been observed to be among the higher concentrations
measured in the U.S. nationwide network.
1 Imperial County Air Pollution Control District. (2012) Annual Network Plan for Ambient Air Monitoring Imperial
County Air Pollution Control District. Available online at
http://www3 .epa. gov/ttn/amtic/files/networkplans/CAImperialPlan2012.pdf.
2 Imperial County Air Pollution Control District. (2014) Imperial County 2013 State Implementation Plan for the
2006 24-Hour PIVb 5 Moderate Nonattainment Area. Available online at
http://www.arb.ca.gov/planning/sip/planarea/imperial/Final PM2.5 SIP (Dec 2. 2014) Approved.pdf.
3 Alaska Department of Environmental Conservation. (2014) Alaska Department of Environmental Conservation
Annual Air Quality Monitoring Network Plan 2014-2015. Available online at
http://dec.alaska.gov/air/am/2014-15%20Monitoring%20Plan%20Final%208-29-14.pdf.
4 Alaska Department of Environmental Conservation. (2014) Alaska Department of Environmental Conservation
Annual Air Quality Monitoring Network Plan 2014-2015. Available online at
http://dec.alaska.gov/air/am/2014-15%20Monitoring%20Plan%20Final%208-29-14.pdf.
2-81
-------
2.5.4 Seasonal, Weekday/Weekend, and Diurnal Trends
Month-to-month variability in 24-h avg NC>2 concentrations was described in the 2008
ISA for Oxides of Nitrogen (U.S. EPA. 2008c). Strong seasonal variability in NC>2 was
reported, with higher concentrations in winter and lower concentrations in summer.
Monthly maxima varied regionally. Day-to-day variability in NC>2 concentration was
generally larger during the winter.
Recent data presented in Table 2-3 continue to show similar seasonal trends for average
seasonal NO2 concentrations across the 2011 to 2013 3-year period. Mean and 99th
percentile concentrations are highest in the first and fourth quarters. Concentration
patterns of NO and NO2 are affected strongly by emissions and meteorology, as
concentrations peak during early morning hours and in winter when PEL heights are
lowest (Figure 2-20). NO2 exhibits flatter profiles relative to NO as secondary formation
processes influence concentration patterns.
—January NO
—January NO2
-July NO
—July N02
123456789 101112131415161718192021222324
Hour of the Day
Note: NO = nitric oxide, NO2 = nitrogen dioxide.
Source: National Center for Environmental Assessment 2013 analysis of Air Quality System network data.
Figure 2-20 January and July hourly profiles of nitric oxide and nitrogen
dioxide (ppb) for the site in Atlanta, GA with the highest 1-hour
nitrogen dioxide concentrations.
Figure 2-20 shows a typical diurnal cycle for a nonnear-road site for NO and NO2. As
described in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). the NO2
2-82
-------
concentration typically exhibits a daily maximum during morning rush hour, although the
concentration maximum can also occur at other times of day. This pattern in Figure 2-20
is shown for Atlanta, GA, but it is also typical for other urban sites. Although the
concentration trends shown in Figure 2-20 are for a nonnear-road monitoring site, they
are similar to trends observed for the Los Angeles, CA and Detroit, MI near-road
concentration patterns in Figure 2-16. NO levels well above zero at night imply that Os
has been completely titrated.
Typically, weekday concentrations of NOx and particularly NO exceed weekend
concentrations, and diurnal cycles are more compressed on weekends. This pattern is
demonstrated for NO2 and NO concentrations at the same monitor in Figure 2-21. In
Atlanta, GA, NOx concentrations were 24% higher on weekdays than on weekends
(Pachon et al.. 2012). The weekend effect for NO was first observed by Cleveland et al.
(1974) and is a general characteristic of urban NO and NOx concentrations observed in
many locations (Tonse et al., 2008; Pun et al., 2003; Marr and Harley. 2002). Differences
between weekdays and weekends were also noted in the 2008 ISA for Oxides of Nitrogen
(U.S. EPA. 2008c). with more pronounced differences at sites more influenced by traffic.
Modeling simulations of weekly cycles of NOx based on summer satellite column data
also indicate higher concentrations on weekdays than on weekends (Choi etal.. 2012).
The satellite column data is converted to concentrations using a chemistry transport
model of the vertical NO2 distribution (see Section 2.4.5). Predicted concentrations agree
with empirical observations and higher concentrations on weekdays than on weekends
are observed regardless of land coverage, for urban, forest, and other regions (Choi et al..
2012). In southern California, NOx concentrations were an average of 46% lower on
weekends than on weekdays in ground-based measurements, and 34% lower in airborne
measurements (Pollack et al.. 2012).
2-83
-------
NO Weekday
NO Weekend
NO2 Weekday
NO2 Weekend
12345678 9101112131415161718192021222324
Hour of the Day
Note: NO = nitric oxide, NO2 = nitrogen dioxide.
Source: National Center for Environmental Assessment 2013 analysis of Air Quality System network data.
Figure 2-21 Weekend/weekday hourly profiles of nitric oxide and nitrogen
dioxide (ppb) for the site in Atlanta, GA with the highest nitrogen
dioxide concentrations.
2.5.5 Multiyear Trends in Ambient Measurements of Oxides of Nitrogen
From 1990 to 2012, the annual average NC>2 concentration across the U.S. based on
concentrations from 135 monitoring sites in the national air quality monitoring network
decreased by 48%, and from 1990 to 2014, the U.S. average annual 98th percentile of
1-hour daily maximum concentrations from 91 monitoring sites decreased by 45%
(http://www3.epa.gov/airtrends/nitrogen.html). The steady decline in NC>2 concentrations
over the years can be attributed mainly to reductions in emissions from mobile and
stationary sources (see Figure 2-2).Considerably fewer monitoring sites were operational
before 1990. However, if the 98th percentile of 1-hour daily maximum NC>2
concentrations were extended as far back as 1980, average U.S. concentrations would
have exceeded the current NAAQS for part of the period. Figure 2-22 shows the decrease
in average annual 98th percentile 1-hour daily maximum NC>2 concentrations for 24 sites
for which NC>2 concentration data are available from 1980 to 2014. Over this period the
concentration decreased by 57%, from 111 ppb in 1980 to 47 ppb in 2014. However, it
was greater than 100 ppb in 1980, 1981, 1982 and 1988, and greater than 90 ppb every
year from 1980 to 1991. Since 1990, concentrations decreased steadily, and by 2014,
2-84
-------
90% of 98th percentile 1-hour daily maximum NCh concentrations at these 24 sites were
less than about 60 ppb.
1980
1985 1990 1995 2000
Year
• Mean 10th and 90th Percentiles
2005
2010
2015
•National Standard
Source: National Center for Environmental Assessment 2014 analysis of Air Trends data
(http://www.epa.aov/airtrends/nitrogen.htmn.
Figure 2-22 U.S. national average of annual 98th percentile of 1-hour daily
maximum nitrogen dioxide concentration at 24 sites, 1980-2012.
Information on trends on a regional basis and at individual, local air monitoring sites can
be found at http://www.epa.gov/air/airtrends/nitrogen.html (National Trends in Nitrogen
Dioxide Levels). One example of particular relevance to the near-road environment is
The Elizabeth Lab site in Elizabeth, NJ. It is situated at the Interchange 13 tollbooth of
the New Jersey Turnpike, within 200 m of a segment of the Turnpike with more than
250,000 vehicles per day. The Elizabeth Lab site is also within 200 m of Interstate 278,
with 126,000 vehicles per day. In 2014, both the highest 98th percentile 1-hour daily
maximum NC>2 concentration (90 ppb) and the highest nonnear-road annual average NC>2
2-85
-------
concentration (22 ppb) in the New York, NY CBSA were observed at the Elizabeth Lab
monitor. Figure 2-23 shows annual average, maximum, and 98th percentile 1-hour daily
maximum concentrations from 1980 to 2014 at the Elizabeth Lab monitor.
°nn
jUU
^
250 j*
£ I i
I2oo : i 1
c
.2 <
4->
to
i 150
0)
• ;
r ' './
• i
i ' J
1 if \
o • •: — •..
c . | ': J
o A .1:
UN100
0
z
••v
i
JS^-BfS- j>
isr
n
it
• I
~n~
' i
\ ' \
i *
A
A"" "•
L ^^ *,
*• •" J
A-A —
>&JB-1B-iB~«
L
(
Y
v* A
•
/.
' i
* ' t
' *
/ m
f
A i\
\ • ^ 9 i * A
A v/ A a V i r>*
k .• •'. A - • \ i * / ^
'^ i' i
^.* . V *
'HA^if
s^^®^^^^^^
••,-
A'A<-
"•SJS— iSl-liS—inL DK:
" ~ ^^^"™'^*<»-S-S-S
1980 1985 1990 1995 2000 2005 2010 20
Year
• Average —A— 98% -»-Max
Note: Average = Annual average; 98% = 98th percentile of daily 1-h maximum concentration; max = highest daily 1-h maximum
concentration. Years 1981-1984 and 1996 are less than 75% complete.
Source: National Center for Environmental Assessment 2015 analysis of Air Quality System network data from 1980-2014.
Figure 2-23 Trend in nitrogen dioxide concentrations at Elizabeth Lab
monitoring site near New Jersey Turnpike 1980-2014.
2.5.6 Background Concentrations
In the context of the review of a NAAQS, the U.S. EPA generally defines "background
concentrations" in a way that distinguishes among concentrations that result from
precursor emissions that are relatively less controllable from those that are relatively
2-86
-------
more controllable through U.S. policies or through international agreements. The most
commonly used form in the past and in this document is North American Background
(NAB), which refers to simulated NO2 concentrations that would exist in the absence of
anthropogenic emissions from the U.S., Canada, and Mexico. This definition of
background includes contributions resulting from emissions from natural sources
(e.g., soils, wildfires, lightning) around the world. Other definitions can also be used. For
example, in the 2013 ISA for Ozone (U.S. EPA. 2013e). a U.S. background, which
includes emissions from Canada and Mexico in addition to those in the definition of a
North American background, and a natural background, which includes only emissions
from natural sources globally, were used. Background is used to inform policy
considerations regarding the current or potential alternative standards.
As can be seen from Figure 2-13. maximum seasonally averaged concentrations of NO2
occur along the Northeast Corridor, the Ohio River Valley, and in the Los Angeles, CA
basin. While NO2 concentrations are often above 5 ppb, NAB is less than 300 ppt over
most of the continental U.S., and less than 100 ppt in the eastern U.S. [see Figure 2.4-18
of the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c)1. The distribution of
background concentrations in the 2008 ISA was shown to reflect the distribution of soil
NO emissions and lightning, with some local increases due to biomass burning, mainly in
the western U.S. In the northeastern U.S., where present-day NO2 concentrations are
highest, NAB contributes <1% to the total.
The only updates to the results given in the 2008 ISA (U.S. EPA. 2008c) are the
global-scale model calculations of Lin et al. (2012). In addition to U.S. and other North
American sources, various NOy species from sources outside North America have long
enough residence times in the atmosphere enabling them to be transported to the U.S.
(Lin etal.. 2012). As noted in the 2013 ISA for Ozone (U.S. EPA. 20136). spring is the
dominant season for effects of intercontinental transport of pollution to be detected in the
U.S. Lin etal. (2012) calculated that transport of NOx from other continents contributes
less than 10 ppt to the regional background in the western U.S., but concentrations of
PAN could range from 50 to 80 ppt.
The annual median NO2 concentration of ~8 ppb reported by the SLAMS monitoring
network is well below the level of the current annual NAAQS (53 ppb) and the hourly
NAAQS (100 ppb). Background concentrations of NO2 are much lower than average
ambient concentrations and are typically less than 0.1 ppb over most of the U.S., with the
highest values found in agricultural areas. All of these values indicate that background
concentrations of NO2 are well beneath the level of the current NO2 NAAQS.
2-87
-------
2.6 Conclusions
A large number of oxidized nitrogen species occur in the atmosphere. They are emitted to
the atmosphere mainly as NO, which interconverts with NC>2. Thus, NO and NO2 are
often combined into their own group and referred to as NOx. NOx plays an important role
in the formation of atmospheric Os and PM. The conversion of NOx into other oxides of
nitrogen, such as PAN, HNOs, or particulate nitrate typically takes place on much longer
time scales than does interconversion between NO and NO2. As a result, near sources,
such as in heavily populated areas or busy roads with heavy traffic, oxides of nitrogen are
mainly present as NOx. However, in remote areas downwind of major sources, more
oxidized species account for a greater fraction of oxides of nitrogen than in populated
areas.
NOx emissions in the U.S. have been roughly cut in half since 1990. In most of the
largest urban areas in the U.S., motor vehicle traffic accounts for 40-67% of emissions
and Off-Highway diesel and gasoline engines contribute an additional 20-30%. Mobile
sources, electric power generation, other stationary fuel combustion, industrial and
agricultural process, and fires are all important NOx sources on a national scale, with
Highway Vehicles, Off-Highway Vehicles and Engines, and stationary fuel combustion
especially important in urban areas. Urban stationary fuel combustion emissions account
for a greater fraction of NOx emissions in colder climates. In some cities, specific
industrial sources like oil and gas production, petroleum refining, or cement
manufacturing account for a greater fraction of NOx emissions locally than they do
nationally. However, traffic emissions are generally responsible for the greatest share of
NOx in the U.S., especially in populated areas.
NO and NO2 are most commonly measured by a Federal Reference Method based on
chemiluminescence of NO induced by its reaction with Os. NO2 is measured by first
reducing it to NO, and then measuring the chemiluminescence of NO. Recent
advancements in NO2 measurements include improved methods of conversion of NO2 to
NO, development of optical methods to measure NO2 directly, and development of
satellite measurement methods. NO2 is measured at hundreds of monitors in several
national monitoring networks. The new near-road monitoring network was initiated in
recognition that millions of people live within a few hundred meters of a major roadway,
and that concentrations of NO2 typically decrease with increasing distance from a major
road.
If annual average NO2 concentrations for individual monitoring sites are averaged over
all monitoring sites in the U.S., the overall average is about 15 ppb. Similarly, the
average daily 1-hour maximum NO2 concentration over all U.S. monitoring sites is about
2-88
-------
30 ppb. Average NO2 concentrations are usually somewhat higher in winter than in
summer. Concentrations are highest in populated urban areas where sources are
dominated by vehicle emissions. Near roads with heavy traffic annual average
concentrations exceeded 20 ppb and 98th percentile 1-hour daily maximum
concentrations exceeded 60 ppb at several sites in 2014. Within urban areas, COD's
typically range from near 0.1 to as high as 0.6, but with a large fraction below 0.2,
indicating that there can be a high degree of spatial variability in some locations, but that
concentrations can be fairly uniform in others. Concentrations within urban areas are
usually highest near major roadways and major stationary sources. Near roadways, there
is often a NO2 concentration gradient, which is strongest in the summer and during
daylight hours. NC>2 concentrations are typically up to 20 ppb higher within 20 m of a
major road than at a distance a few hundred meters from the road, and the spatial extent
of elevated concentration typically ranges from 200 to 500 m. Preliminary results from
the U.S. EPA's new near-road monitoring network indicate that 98th percentile 1-hour
daily maximum NC>2 concentrations at all near-road monitors were usually below 60 ppb,
and always below the 1-hour daily maximum NAAQS of 100 ppb at all sites. However,
annual average NO2 concentrations for 2014 were usually higher at near road monitoring
sites than in other locations in the same city.
Much of the most recent research on atmospheric NO2 and NOx has focused on their role
as a traffic pollutants and their spatial variability, especially in proximity to major roads.
Because traffic is the largest source of NOx in the U.S., especially in populated areas, this
research is highly relevant to human exposure, and the results described in this chapter
provide a useful context for characterization of NO2 exposures and associated health
effects.
2-89
-------
CHAPTER 3 EXPOSURE TO OXIDES OF
NITROGEN
3.1 Introduction
Assessment of exposure to ambient oxides of nitrogen builds from the characterization of
concentrations and atmospheric chemistry presented in Chapter 2. The primary
conclusions from Chapter 2 were that NCh concentrations have declined over the past
20 years, but concentrations are still elevated near roads and in urban areas, with
vehicular traffic and off-highway vehicles contributing the majority of NC>2 emissions.
For this reason, NC>2 exposure assessment focuses predominantly on urban and near-road
settings.
Total personal exposure to ambient oxides of nitrogen is given by the concentration of
oxides of nitrogen emitted from ambient sources and encountered by an individual over a
given time. Personal exposure to ambient oxides of nitrogen is influenced by a number of
factors, including:
• time-activity in different microenvironments (e.g., vehicle, residence, workplace,
outdoor);
• climate (e.g., weather, season);
• characteristics of indoor microenvironments (e.g., window openings, draftiness,
air conditioning); and
• microenvironmental emission sources (e.g., roadways, construction equipment,
indoor gas stoves) and concentrations.
Surrogates for personal exposure to ambient oxides of nitrogen include ambient NO2
concentrations measured at a central site monitor or modeled using spatial techniques
such as land use regression (LUR), Gaussian dispersion models, or chemical transport
models (CTM). All exposure surrogates are subject to measurement errors related to
spatial and temporal variability of the ambient concentration field, quality of additional
input data, representativeness of predictor variables, and accuracy of the monitoring or
modeling methodology. The following sections describe methodological considerations
for use of exposure data, characterization of NO2 exposures, and exposure assessment
and epidemiologic inference. This chapter focuses on the ambient component of personal
exposure to NO2, because the NAAQS regulates ambient oxides of nitrogen, for which
NO2 is the indicator. However, studies using total personal NO2 measurements and
indoor NC>2 concentrations to represent exposure can also inform the understanding of
exposure and related health effects and so are included as supporting evidence where
appropriate. This chapter focuses on studies of exposure among the general population.
3-1
-------
Exposure of at-risk groups, based for example on socioeconomic status, race, and
proximity to roadways, is addressed in Chapter 7; occupational exposures to ambient
NO2 are discussed in Chapter 7 within the subsections for socioeconomic status and
proximity to roadways. The information provided in this chapter will be used to help
interpret the health effects studies of NO2 exposure presented in Chapter 5. Chapter 6,
and Chapter 7.
3.2 Methodological Considerations for Use of Exposure Data
The following sections outline various facets of NO2 measurement and estimation,
including FRMs (i.e., central site monitors) and personal NC>2 exposure sampling
techniques and NCh exposure modeling. The section ends with a discussion of the
application of measurement and modeling techniques in epidemiologic studies of
different designs.
3.2.1 Measurement
3.2.1.1 Central Site and Near-Road Monitoring
Monitoring of NCh concentrations by chemiluminescent sampling is described in detail in
Section 2.4.1 along with limitations of the monitoring methodology. In summary, NC>2
concentrations are calculated by FRM as the difference between NO concentration
measured in the air stream that has passed over a heated MoOx substrate (measuring total
oxides of nitrogen) and NO concentration in the air stream that was diverted away from
the substrate. FRMs are subject to positive bias because oxidized nitrogen compounds
other than NO2 are often detected by the MoOx substrate. A FEM is also available to
measure NO2 concentration directly using a photolytic converter to reduce NO2 to NO.
Evaluation of the chemiluminescent method is provided in Section 2.4.1 along with a
description of the measuring technique. Monitors set up by state agencies as part of the
SLAMS network that report to the air quality system (AQS) are typically centrally sited,
although the same monitors are used in select cases for near-road monitoring. See
Section 2.4.5 for more details.
In addition to judging compliance with the NAAQS, NO2 concentrations measured by
centrally sited or near-road FRMs and FEMs are frequently used by epidemiologic
researchers as surrogates for exposure in studies of the health effects of exposure to
oxides of nitrogen, as described further in Section 3.4. Central site monitoring data can be
3-2
-------
used in epidemiologic studies of short-term exposure to NCh when focused on the time
series of exposure or in epidemiologic studies of long-term exposure when comparing
average NO2 concentrations from different geographic areas. Section 3.4.3 explores the
factors causing errors associated with siting central site or near-road monitors at a single
location, and Section 3.4.5 considers the influence of those errors on health effect
estimates. Briefly, with respect to time-series exposure estimation for epidemiologic
studies of short-term exposure, correlation decreases as distance increases between two
monitors. For epidemiologic studies of long-term exposure to NC>2, difference between
the measured concentration and the true exposure would result in exposure error. The
limited number of samplers in the network could potentially increase exposure error
further.
3.2.1.2 Personal and Area Sampling
Personal sampling for NC>2 exposure is most commonly used in epidemiologic panel
studies. Personal sampling for NC>2 was described in detail in Annex 3.3 to the 2008 ISA
for Oxides of Nitrogen (U.S. EPA. 2008a) and is briefly summarized here. Active
sampling systems typically involve air pumped past a chemiluminescent device; they
enable measurements of NO2 over short time periods to produce near real-time data.
Given the weight of most active sampling systems, they are not used extensively for
personal sampling. Passive samplers based on Pick's first law of diffusion are more
commonly deployed for personal or area NCh sampling in a badge, tube, or radial
manifold. These are typically deployed over periods ranging from a few days to several
weeks. Passive sampling results are integrated over the time period during which the
sorbent material is exposed, which is selected by the user and usually spans days to
weeks. The 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) reported that, depending
on the sorbent material, personal NO2 samplers may be subject to biases related to
interferences from HONO, PAN, HNOs (Gair et al., 1991). and high relative humidity
(RH) (Centre di Ricerche Ambientali. 2006). These biases also depend on ambient
temperature and atmospheric levels of the copollutants.
Recent work has been performed to evaluate passive sampling device performance.
Sather et al. (2007) compared Ogawa passive samplers with a collocated NO2 FRM
monitor over a 4-week field study in El Paso, TX and observed good agreement between
the techniques, with an average absolute difference of 1.2 ppb and R2 = 0.95. For
measurements in Umea, Sweden, Hagenbjork-Gustafsson et al. (2009) observed that,
when using the manufacturer's recommended uptake rates to calculate concentration,
passive NO2 measurements were negatively biased by 9.1%, and NOx concentration
measurements were positively biased by 15% compared with an FRM. When uptake rates
3-3
-------
were derived in the field based on the chemiluminescent FRM, NO2 measurements were
positively biased by 2%, and NOx concentration measurements were unbiased compared
with the FRM. These results suggest that deviation from temperature conditions under
which the samplers were laboratory tested may lead to biased results. Sanchez Jimenez et
al. (2011) used Palmes-type passive diffusion tubes to measure both NO2 and NOx
concentrations and investigated specific sources of biases in their measurements. They
found that, within the passive diffusion tubes, NO and Os were reacting to form NO2,
causing NO measurements to be negatively biased while NO2 measurements were
positively biased. Wind was also a source of positive bias in the NO2 and NOx
concentration measurements because increased airflow effectively reduced the diffusion
lengths of the gas collection tubes. In laboratory and field evaluation of NO2 passive
diffusion tubes, Buzica et al. (2008) observed negligible difference between the diffusion
tubes and FRM measurements; however, uncertainty increased with decreasing
concentration. When comparing biases among samplers, note that the FRM is subject to
positive biases related to sensitivity to PAN, RONO2, and FINOs (Sections 2.4.1 and
3.2.1.1).
Triethanolamine (TEA) is often employed as a sorbent material in denuders used for
capturing NO2 during active sampling and in passive sampling because it can be applied
in an even coating. However, sampling efficiency is sensitive to sampler flow rate (Vichi
and De Santis. 2012). relative humidity (Poddubny and Yushketova. 2013; Sereviciene
and Paliulis. 2012; Vardoulakis et al.. 2009), averaging time (Vardoulakis et al.. 2009).
and ambient temperature (Poddubny and Yushketova. 2013). Heal (2008) found that NO2
bias was sensitive to the method of application of the TEA to the substrate. Sekine et al.
(2008) and Nishikawa et al. (2009) experimented with size and number of filters,
respectively, in a passive sampler and found minimal effect on NO2 or NOx
concentration. Ozden and Dogeroglu (2008) observed that TEA-complexed NO2 was
sensitive to photodegradation if not stored in a dark glass tube, resulting in
underprediction of NO2 exposure.
Recent attention has been given to using passive or miniature active monitors for
saturation sampling, i.e., siting monitors over a dense grid. This is typically done in urban
areas. For example, Ross etal. (2013) sited roughly 25 passive NO2 monitors during six
two-week periods at a total of 150 locations across the five boroughs of New York City,
NY to create a dense concentration map for exposure estimates and to provide training
and validation data for LUR. Saturation sampling was also conducted in nine MESA-Air
communities with up to 105 Ogawa passive badges deployed to measure NO2 during
three two-week sampling periods; noise was also measured at two monitoring stations in
Chicago, IL and one in Riverside, CA (Section 3.4.4.4). Similarly, Shmool et al. (2014)
deployed Ogawa passive badges for NO2 sampling, along with measuring PM2 5, BC,
3-4
-------
relative humidity, and barometric pressure across metropolitan Pittsburgh, PA. The
monitoring boxes were sited to capture air pollution gradients along the
urban-to-suburban land use gradient and included areas influenced by industrial sources
and highways. Skouloudis and Kassomenos (2014) deployed sensors for NCh, NOx, CO,
Os, and benzene (CeHe) to correspond to the population distribution on the island of
Malta. Active samplers were used in this scheme, with a global positioning system (GPS)
and data transmission capabilities for near real-time analysis. Skouloudis and
Kassomenos (2014) proposed that data from these dense area samplers could also be
assimilated with satellite measurements to improve the accuracy of the exposure
estimates.
3.2.2 Modeling
Computational models can be used in epidemiologic studies to estimate exposure when
measurements are not available at locations and/or times needed to estimate spatial and
temporal variability in NC>2 concentration. These methods can sometimes account for
complex urban morphometry and meteorology, which can interact to cause turbulence
that may affect pollutant residence times (Fernando. 2010) or incorporate localized
sources that might not otherwise be detected by central site monitoring (Goldman et al..
2012). Such estimates can then be used as inputs to exposure models described in
Section 3.4. These modeling approaches produce data at times and/or locations where
exposures are uncharacterized, but each method carries its own uncertainty (Fuentes.
2009). Detailed descriptions of computational models used for predicting spatially
resolved concentration profiles for exposure assessment have been provided in
Section AX 3.6 of the 2008 ISA for Oxides of Nitrogen Annex (U.S. EPA. 2008a) and
Section 3.8 of the 2009 ISA for Particulate Matter (U.S. EPA. 2009a). Methods include
LUR models, spatial interpolation through statistical techniques, CTM, and dispersion
models.
3.2.2.1 Statistical Modeling
Land Use Regression Models
LUR modeling has been applied extensively to estimate the spatial distribution of
ambient NO2 or NO concentration for neighborhood or urban-scale exposure assessment
in epidemiologic studies of long-term exposure (Clougherty et al.. 2013; Hatzopoulou et
al.. 2013; Cesaroni et al., 2012; Gonzales et al., 2012; Mukerjee et al., 2012a; Mukerjee
etal.. 2012b; Oiamo etal.. 2012; Esplugues et al.. 2011; Fernandez-Somoano et al.. 2011;
3-5
-------
Hvstadetal..2011: Oiatno etal.. 2011; Rose etal.. 2011; Smith etal.. 2011; Szpiro et al..
201 la; Adamkiewicz et al.. 2010; Aguilera et al.. 2009; Cohen et al.. 2009; Hart et al..
2009; Iniguez et al.. 2009; Karr et al.. 2009; Mukerjee et al.. 2009; Su et al.. 2009b;
Aguilera etal.. 2008; Atari et al.. 2008; Cesaroni et al.. 2008; Rosenlund et al.. 2008a;
Jerrett etal.. 2007). LUR fits a multiple linear regression model of concentration data as a
function of land use data and then applies that model to locations without monitors to
increase the spatial resolution of the concentration field (Marshall et al.. 2008). LUR
models for NCh are typically calibrated using data from passive sampler measurements.
Given that most passive measurement methods are not designed for short-term sampling,
LUR models are typically based on several days, weeks, or years of data and hence do
not account for short-term temporal variability well. Hence, LUR is commonly used to
estimate air pollution exposure in epidemiologic studies of long-term NC>2 exposure
(Chapter 6).
Finer spatial resolution of calibration points can improve goodness of fit and
representativeness of the model. Using 155 monitoring sites throughout New York City,
NY, Clougherty et al. (2013) ran an LUR with resolutions down to 50 m with in-sample
sequential R2 = 0.671. Parenteau and Sawada (2012) examined LUR model performance
when basing the model on successively finer spatial resolution from 2 km down to 50 m,
with the geographic borders of the finely resolved regions tied to population groupings
based on population density mapping. The two finer resolution approaches yielded better
agreement with measured NO2 data (in-sample R2 = 0.80-0.81) than the less spatially
resolved approach (in-sample R2 = 0.70). Root mean squared error (RMSE) was
computed for a cross-validation data set, and RMSE = 1.05 ppb. Basagana et al. (2012)
evaluated LUR models for 24 to 120 NO2 measurement sites in Girona, Spain. Basagana
et al. (2012) observed that leave one out cross-validation (LOOCV) resulted in a higher
R2 compared with out-of-sample R2 computed using distinct validation sites. At the same
time, LOOCV R2 declined with increasing number of training sites, while out-of-sample
validation R2 increased within increasing number of training sites. Johnson et al. (201 Ob)
evaluated LUR performance in New Haven, CT when the LUR model was fit with NO2
data from 25 to 285 measurement sites and found that the LOOCV sites produced R2
much smaller than in-sample R2. The LOOCV R2 increased with increasing number of
training sites. Wang etal. (2012) also evaluated LUR performance when fit with 24 to
120 NO2 monitors distributed across the Netherlands. They compared LOOCV R2 with
external validation R2. Qualitatively, their results were the same as those of Basagana et
al. (2012).
^'Out-of-sample" refers to validation of the model with a data set not used to fit the model; in this case, the
neighborhood-level simulations were cross validated against the whole-city measurements and vice versa. "In-
sample" refers to a comparison between the model and measurements used to fit the model.
3-6
-------
Recent studies have applied an LUR model among multiple cities. Recently, LUR has
been implemented to examine local-scale concentration estimates across the contiguous
U.S. (Beckermanetal.. 2013b; Novotnyetal.. 2011; Hart etal.. 2009) and Canada
(Hystad et al.. 2011). Allen et al. (2011) developed separate LUR models for two
Canadian cities (Winnipeg, Manitoba and Edmonton, Alberta) with 50 calibration points
each and then applied the models to the other city to compare performance. As
anticipated, locally generated model performance (NCh: in-sample R2 = 0.81-0.84;
out-of-sample R2 = 0.75-0.77) was superior to performance of the model applied to the
other city (NCh: R2 = 0.37-0.52) and to bivariate local models using only road proximity
(in-sample R2 < 0.16). NC>2 models consistently performed better than NO models. Wang
etal. (2014) developed a LUR model for NO2 based on data from 23 European study
areas (containing 20-40 sites within each study area) with NC>2, PlVfc 5, land use, and
traffic data. Given the continental design of the study, a regional background
concentration variable was also imposed on the model. The in-sample LUR model fit was
R2 = 0.59 for all of the urban areas combined. After fitting the LUR model, Wang et al.
(2014) tested the LUR model's ability to predict concentrations for different
configurations of cities by LOOCV. They found comparable results (LOOCV R2 = 0.50).
Generally, both in-sample and out-of-sample R2 for multiple city studies were either
comparable or lower than the respective R2 for single city studies. This would be
expected given the smoothing effect of fitting a model over a large geographic area.
Selection of predictor variables, such as meteorology, traffic, land use, and population
density, influences the ability of the LUR model to predict concentrations of oxides of
nitrogen and depends on the specific city for which the model is fit. Su et al. (2008a) and
Ainslie et al. (2008) developed the Source Area-LUR (SA-LUR) to incorporate the
effects of meteorology (and hence to incorporate the effects of temporal variability) on
the model results. The SA-LUR integrates data for wind speed, wind direction, and cloud
cover variables in estimates for NO and NO2. It was found to perform better when
seasonal variability in concentrations was high. Su et al. (2008b) included a street canyon
aspect ratio as a LUR predictor variable to account for retention of pollutants in street
canyons. They observed that, upon adding the aspect ratio to the LUR model, in-sample
R2 increased from 0.56 to 0.67 for NO2. Similarly, when Clougherty et al. (2013) added
"built space within 1 km" to their LUR model of NO2, in-sample R2 increased by 0.41.
Franklin et al. (2012) explored bivariate correlations between NO2 concentrations and
several predictors reflecting traffic, population, elevation, and land use in twelve southern
California communities. Pearson correlations of NO2 concentration with distance to road
were r = -0.42 and -0.35 for freeway and nonfreeway roads, respectively. Their model
produced a -8.2% change in concentration per IQR increase in distance in the LUR
model. Correlations with traffic volume within a 300-m buffer were r = 0.41, and traffic
volume within a 300-m buffer produced a 2.4% change in the LUR prediction per IQR.
3-7
-------
Correlation with neighborhood elevation was r = -0.50, and neighborhood elevation
produced a -6.7% change in LUR-modeled concentration per IQR increase in elevation.
Su et al. (2009a) developed a method to optimize the SA-LUR variable selection process
in which correlations between several land use variables and NO2 concentrations were
computed across a 3-km buffer of the NO2 measurement (1.5-km buffer for traffic-related
variables), and the data for correlation versus distance were fit to a curve describing that
relationship. The variable with highest correlation at the optimum buffer distance was
added to the model if its addition produced a statistically significant change (p < 0.1) in
the model. Su et al. (2009a) found the important variables to be distance from monitor,
24-hour traffic levels, expressway casement, open land use, railway, major road, land
grade, population density, and distance to coast. Beckerman et al. (2013a) adopted an
addition\substitution\deletion machine learning approach to variable selection. This
method employs a v-fold cross-correlation and computes the least-square error for each
model having different numbers and combinations of predictor variables. The algorithm
selects the model that optimally minimizes both the least-square error and the size of the
model. The minimum out-of-sample R2 = 0.83, and the minimum least-square error was
0.118. This analysis of cross-validation statistics showed the point where inclusion of
more variables produced relatively small gains in cross-validation so that model
parsimony could be maintained.
Several studies of LUR have considered seasonality in the model. Grouse et al. (2009a)
and Dons et al. (2014) evaluated LUR across seasons and found that spatial variability in
the NO2 concentration profile did not change substantially with season. Therefore, the
authors concluded that an annual average would be acceptable for LUR simulations.
However, Arain etal. (2009) observed seasonal changes in the spatial distribution of NO2
concentration in a study of NC>2 concentrations over the greater Toronto-Hamilton
airshed. Seasonal deviations in observed spatial NO2 concentration patterns would imply
that an LUR model fit would also need to account for seasonality either through a
variable or through stratification.
LUR models applied several years after model development have demonstrated
predictive ability in a few studies. Eeftens et al. (2011) compared LUR obtained from
NO2 concentration measurements at 35 locations in the Netherlands over the years
1999-2000 with LUR developed from NC>2 concentration measurements at 144 locations
in the Netherlands during 2007. Both the NCh concentration measurements and the LUR
models agreed well for the two time periods studied; the comparison between models
from the different time periods produced p = 0.9998 and R2 = 0.89. Wang etal. (2013b)
tested stability of an LUR model for Vancouver, Canada between 2003 (based on
116 sites) and 2010 (based on 116 sites, with 73 from the 2003 study). Wang et al.
(2013b) evaluated the model by testing how much variability in the measurements was
3-8
-------
predicted by models from the other year. Linear regression for comparison of the 2003
model with 2010 measurements produced R2 = 0.52-0.61 for NO2, while comparison of
the 2010 model with 2003 measurements produced R2 = 0.44-0.49 for NO2. Wang et al.
(2013b) attributed the diminished performance for the 2003 model using 2010 data
(compared with using the 2010 model for 2003 data) to reductions in NO and NC>2
concentrations over the 7-year time period. Visual inspection of the NO and NO2
concentration maps from the Wangetal. (2013b) study suggests that changes in spatial
correlation over time may have contributed to reduced model performance in comparison
with the Eeftensetal. (2011) study.
LUR evaluation depends on the validation algorithm, model conditions, and basis for
validation (i.e., to what the modeling results are compared when computing
out-of-sample R2). In a recent study of LUR application in 20 European study areas,
Wang et al. (2013a) found that LOOCV produced higher R2 for NO2 concentration
compared with hold-out evaluation (HEV) (LOOCV: R2 = 0.83; HEV: R2 = 0.52).
LOOCV involves repeatedly withholding a fraction of the monitoring sites from the
fitting process for performance evaluation and then computing an ensemble R2, whereas
HEV entails prediction with the LUR at locations not fit by the model. Therefore, HEV
provides a more independent data set for validation. Mercer et al. (2011) compared
10-fold cross-validated LUR with universal kriging (UK), in which a surface of
concentrations was built based on measured values for three seasons in Los Angeles, CA
with roughly 150 measurement sites. UK performance was slightly better than LUR for
all seasons, and model performance did not vary much among the seasons (UK: 10-fold
cross-validation out-of-sample R2 = 0.75, 0.72, and 0.74; LUR: R2 = 0.74, 0.60, 0.67). LJ
et al. (2012b) developed a new formulation for LUR using generalized additive models
(GAM) and cokriging to boost the performance of LUR. They evaluated this approach for
Los Angeles, CA. GAM allowed localized nonlinear effects to be incorporated among the
prediction covariates, while cokriging was intended to improve spatial smoothing. The
LUR using GAM and cokriging had the highest LOOCV (R2 = 0.88-92), compared with
universal kriging (R2 = 0.68-0.75) and multiple linear LUR (R2 = 0.42-0.64).
LUR comparison with other models has produced variable results, in part because the
comparison data do not always have the same spatial resolution or account for the same
physical phenomena. Beelen et al. (2010) compared LUR with a dispersion model
incorporating a near-road module for modeling NO2 concentrations in a Rotterdam,
Netherlands neighborhood. The dispersion model agreed better with NO2 measurements
(out-of-sample Pearson r = 0.77) compared with the agreement between LUR and
measurements (out-of-sample r = 0.47) from 18 evaluation sites. Dijkema et al. (201 la)
also compared LUR for the city of Amsterdam and a larger geographic portion of
northwest Netherlands with a dispersion model and found belter agreement of the
3-9
-------
dispersion models with observations for both the citywide model (dispersion: R2 = 0.581;
LUR: out-of-sample R2 = 0.48) and the large-area model (dispersion: R2 = 0.74;
LUR: out-of-sample R2 = 0.57). Marshall et al. (2008) compared LUR with inverse
distance-weighted (IDW) spatial interpolation of NO and NO2 measurements, nearest NO
and NO2 measurements, and a Community Multiscale Air Quality (CMAQ) model run
for Vancouver, Canada. The LUR location was matched to each CMAQ grid cell centroid
and compared with the grid cell concentration. LUR and CMAQ produced similar
average absolute difference in the concentration compared with measured central site
concentrations for NO (LUR: 42%, CMAQ: 47%) andNO2 (LUR: 17%, CMAQ: 17%),
while nearest monitor and spatial interpolation methods produced less than 5% difference
for both pollutants and methods. However, it is important to recognize that these methods
were compared to a central site monitor, which cannot capture the spatial variability of
the NO2 concentration distribution. Specifically, IDW, central site monitoring of NO2
concentration, and nearest monitor NO2 concentration estimation approaches do not
account well for localized sources unless the sources are close to the monitors. Therefore,
to inform inference for epidemiological studies, the comparison of the modeled estimates
to measured values should be at locations that are relevant to the intended epidemiologic
study.
Recent studies have explored combination of LUR and other models. For example,
Wilton et al. (2010) added a covariate for concentrations computed with the CALINE3
dispersion model in their LUR to estimate NOx and NO2 concentrations in Los Angeles,
CA and Seattle, WA. They observed modest improvements in model R2 (Los Angeles,
NO2: LOOCV R2 = 0.77 vs. R2 = 0.71-0.73; Seattle, NO2: LOOCV R2 = 0.67 vs.
R2 = 0.53-0.63) when CALINE3-computed concentration was included as one variable
along with land use, roadway length, and traffic density variables. Molter et al. (2010a)
also used dispersion modeling data in lieu of measurement data when fitting an LUR for
Greater Manchester, U.K. and found reasonable agreement of LUR-predicted NO2
concentrations with a separate monitoring data set where 25% of the data were set aside
for cross-validation (out-of-sample R2 = 0.62). Note that the nature of the monitoring data
(i.e., central site or other) was not explicitly stated in the Molter et al. (2010a) study.
Janssen et al. (2012) proposed using LUR to improve performance of a CTM by
downscaling the CTM to the LUR. Downscaling entails a redistribution of the
CTM-modeled concentrations through a statistical model that conforms to measured
concentrations at points in space where measurements are available using the
LUR-derived regression parameters. Janssen etal. (2012) found that the spatial
representativeness of the CTM for NO2 improved by roughly 20% when incorporating
^'In-sample" and "out-of-sample" terminology is not used for dispersion, chemical transport, or scale models,
because those models do not require input concentration data. All comparisons with monitoring data are therefore
out-of-sample.
3-10
-------
the LUR downscaler, based on a comparison of the CTM and downscaled CTM with
central site monitor measurements. It is worth noting that any errors and uncertainties
associated with a particular LUR run would transfer to the downscaled result if LUR
were used as a basis for downscaling CTM results. Beckerman et al. (2013b) had
improved results when combining LUR with a Bayesian maximum entropy model to
capture PM2 5 concentration across the contiguous U.S., with fivefold cross-validation
producing R2 = 0.79. It is not clear whether NC>2 would produce as good of a validation,
because NO2 is more spatially variable.
Spatiotemporal Modeling
Spatiotemporal modeling uses advanced statistical techniques to model concentration
variation over space and time. These models decompose the concentration at each study
location into a trend and a residual component (Sampson et al.. 2011; Szpiro et al.. 2010).
The mean trend was modeled at each location in the study domain as a linear
combination of time-series basis functions of spatial covariates and kriging, and the basis
functions were selected using singular value decomposition. Examples of spatial
covariates included distance to major road, population density, and land use categories.
Sampson et al. (2011) showed good validation, with no outliers and
RMSE = 0.88-2.42 ppb (depending on site) using LOOCV across the six MESA Air
cities. In Szpiro et al. (2010). cross-validation produced an out-of-sample R2 = 0.67 with
RMSE = 4.21 ppb. (Li et al.. 2013) adopted this approach to model NO2 and NOx for
both time-series and long-term average exposures in southern California and found that
out-of-sample R2 was high for both temporal treatments (NCh: time series: R2 = 0.84,
long-term average R2 = 0.89; NOx: time series R2 = 0.81, long-term average R2 = 0.77).
Across the six MESA Air cities, Keller etal. (2015) observed a 10-fold LOOCV
R2 = 0.85-0.96 for NO2 and R2 = 0.00-0.98 for NOX. Lindstrom et al. (2013) adapted this
Spatiotemporal model by replacing the land use variables with dispersion model output.
Lindstrom et al. (2013) compared their results to the Spatiotemporal model employing
land use covariates and found no appreciable improvement (R2 within ± 0.04) in model
performance for a variety of averaging times (daily "snap shot," 2-week, 10-year).
3-11
-------
3.2.2.2 Mechanistic Models
Chemical Transport Models
CTMs can be used to develop estimates of NO, NO2, or NOx concentrations. CTMs, such
as CMAQ, are deterministic models of chemical transport that account for physical
processes including advection, dispersion, diffusion, gas-phase reaction, and mixing
while following the constraint of mass conservation (Byun and Schere. 2006). Temporal
resolution of CTMs can be as fine as 1 hour, although larger temporal aggregation is
often used to maintain reasonable data file size. These models provide regional
concentration estimates and are typically run with surface grid resolutions of 4, 12, or
36 km. No studies of CMAQ grid size convergence have been found, and U.S. EPA
(1999) points out that testing for convergence properties is prohibitive due to the large
computational demands of the CMAQ program. Shao et al. (2007) compared simulation
results among 12-km, 4-km, and 1.33-km resolutions and found that discontinuities at the
grid cell boundaries increased with increasing grid size.
CTMs can be applied in epidemiologic studies of either short- or long-term exposure to
NC>2 or NOx but are more commonly used in long-term exposure studies. These models
are used to compute interactions among atmospheric pollutants and their transformation
products, the production of secondary aerosols, the evolution of particle size distribution,
and transport and deposition of pollutants. CTMs are driven by emissions inventories for
primary species such as NO2, SO2, NHs, VOCs, and primary PM, and by meteorological
fields produced by other numerical prediction models. Given observed biases in the
CTMs [e.g., Shi and Zhang (2008) for NO2, and larger biases in organic carbon, PM2 5,
nitrate, and other compounds (Foley etal.. 2010; Eder and Yu. 2005)1. much attention
has been given recently to bias correction of these models for application in exposure
assessment, as detailed below in the Subgrid Scale and Data Fusion Models section.
Dispersion Models
Dispersion models, or Gaussian plume models, estimate the transport and dispersion of
ambient air pollutants emanating from a point or line source by solving for an equation
that estimates the spread of the pollutant to follow a Gaussian curve that is a function of
distance from the source. Given that dispersion models typically capture average
concentrations, they are most commonly used in epidemiologic studies of long-term
exposure. Several studies of health effects related to NOx exposure employ dispersion
models to estimate NOx concentrations [e.g., Gruzieva et al. (2013). McConnell et al.
(2010a). and Oftedal et al. (2009a)] because NO2 has high local spatial variability
(Section 2.5.3). The grid spacing in regional CTMs, usually between 1 and 12 km2, is too
3-12
-------
coarse to resolve spatial variations on the neighborhood scale. More finely resolved
spatial scales that better represent human exposure scales are provided by local-scale
dispersion models. Several dispersion models are available to simulate concentration
fields near roads, and each has its own set of strengths and weaknesses.
Several line-source Gaussian dispersion models are available to simulate the dispersion
of emissions from a roadway. The CALINE family of models does not include NOx
transformation chemistry. Benson (1992) validated the CALINE3 and CALINE4 model
versions using data from field studies at U.S. Highway 99 in Sacramento, CA and a
General Motors test track in Michigan. Benson (1992) found that more than 85% of
model predictions fell within a factor of two of measured observations for sulfur
hexafluoride (SF6) (an inert tracer gas). Among those that fell outside the factor of two
envelope, 85% were positively biased and mostly occurred when wind speeds were
below 1 m/s. Additionally, Benson (1992) tested the NO2 module of CALINE4 under a
limited set of conditions and recommended that CALINE4 not be used to predict NO2
dispersion under parallel wind conditions without ample data to calibrate the model
predictions.
The University of California, Davis (UCD) 2001 model was designed to improve upon
the design of CALINE by using an array of point sources to represent a three-dimensional
highway source of emissions and by using power law functions for wind speed and
vertical eddy diffusivity (Held et al.. 2003). UCD 2001 exhibited improved performance
for parallel, low-speed winds (<0.5 m/s), with 87% and 83% reduction in error compared
with CALINE3 and CALINE4, respectively, for the General Motors SFe evaluation data
set. Snyderetal. (2013) recently released a Research Line-source (RLINE) dispersion
model that incorporates improved formulations of horizontal and vertical dispersion and
found that the predictions were within a factor of two of the observations for neutral,
convective, and weakly stable atmospheric conditions, but negative bias was observed for
stable conditions based on a line source SF6 experiment in Idaho Falls, ID. During
comparison with the U.S. 99 data set, 81% of data were within a factor of two for
downwind measurements, but only 19% for upwind measurements when winds were
within 30°of perpendicular to the road. Seventy-five percent of downwind predictions
were within a factor of two of observations when winds were less than 1.5 m/s, and 88%
were within a factor of two for wind speeds greater than 1.5 m/s. Only 51% were within a
factor of two when winds were within 30°of parallel to the road. Additionally, an
optimization model fitting CALINE3 line-dispersion calculations for concentration to
observations of NCh was developed and applied in the greater Tel Aviv, Israel area
(Yuval et al., 2013). Cross-validation was reported to have negligible bias in the model
predictions with 36% error; the authors did not clearly distinguish bias and error in this
manuscript.
3-13
-------
The American Meteorological Society/Environmental Protection Agency Regulatory
Model (AERMOD; http://www.epa.gov/scramOOl/dispersion_prefrec.htm) is a
steady-state point source plume model formulated as a replacement to the Industrial
Source Complex (ISC3) dispersion model (Cimorelli et al.. 2005). In the stable boundary
layer, the model assumes the concentration distribution to be Gaussian in both the vertical
and horizontal dimensions. In the convective boundary layer, the horizontal distribution
is also assumed to be Gaussian, but the vertical distribution is described with a
bi-Gaussian probability density function. AERMOD has provisions that can be applied to
flat and complex terrain and multiple source types (including point, area, and volume
sources) in both urban and rural areas. It incorporates air dispersion based on the
structure of turbulence in the planetary boundary layer and scaling concepts and is meant
to treat surface and elevated sources, in both simple and complex terrain in rural and
urban areas. The dispersion of emissions from line sources like highways in AERMOD is
handled as a source with dimensions set using an area or volume source algorithm in the
model; however, actual emissions usually are not in a steady state.
Most simple dispersion models, including AERMOD, are designed without explicit
chemical mechanisms but have nondefault options to estimate conversion of NO to NO2
based on a NOx/Os titration model. Hendrick et al. (2013) evaluated two modules used
with AERMOD to compute NO2 concentrations: the plume volume molar ratio method
(PVMRM) and the ozone limiting method (OLM). Both methods assume ratios of
NO2-to-NOx that are based on the concentration of co-occurring Os. Hendrick et al.
(2013) validated the models against more than 12 months of hourly observations taken
near a small power plant in Wainwright, AK, and they observed that the PVMRM
overpredicted NO2 at low concentrations and underpredicted at high concentrations,
although the average bias was small; the OLM also overpredicted NO2 concentrations at
high observed NO2.
AERMOD results have been compared with measurements and other models to evaluate
relative performance. Gibson etal. (2013) found poor agreement with respect to
magnitude of NOx concentrations and correlations (R2 = 0.001-0.003) at hourly,
monthly, and annual timescales when comparing AERMOD results with observations in
Halifax, Canada where several industrial facilities emit NOx. Cohan etal. (2011)
compared AERMOD output with 24-hour central site monitoring observations averaged
over August 2005 from San Jose, CA, where there are combined emissions from a port,
rail yard, and roadways. They observed that the AERMOD model consistently
underpredicted the observations; negative bias was more pronounced for simulations
from January compared with August. Misraetal. (2013) compared AERMOD with the
Quick Urban and Industrial Complex (QUIC) model. QUIC approximates average
airflow around buildings in urban environments then models pollution parcels based on
3-14
-------
Lagrangian particle dispersion. In this case, AERMOD underpredicted NOx
concentrations in an urban street canyon, while most QUIC predictions were within a
factor of two of the observed NOx concentrations.
There are also nonsteady state models for different types of sources. For example,
CALPUFF (http://www.src.com/calpuff/calpuffl.htm). which is the U.S. Environmental
Protection Agency's (EPA's) recommended dispersion model for transport in ranges
>50 km, is a nonsteady state puff dispersion model that simulates the effects of time- and
space-varying meteorological conditions on pollution transport, transformation, and
removal and has provisions for calculating dispersion from surface sources (U.S. EPA.
1995b). However, CALPUFF was not designed to treat the dispersion of emissions from
roads, and like AERMOD, it has some limited chemistry options to estimate production
of secondary pollutants. The distinction between a steady state and time-varying model
may not be important for studying health effects for which long-exposure timescales are
relevant; however, when short-exposure timescales are of interest (e.g., 1-hour), it would
be more important to approximate the short-term variability in concentrations. CALPUFF
was validated against SF6 data at two military test sites in Nevada (Chang et al.. 2003).
where it was shown that 52% of CALPUFF predictions were within a factor of two of
SF6 observations for one site and 29% of predictions were within a factor two of the
observations at a second site. The second test site had surrounding mountains which
increased vertical dispersion; CALPUFF did not account well for vertical dispersion. Cui
et al. (2011) evaluated CALPUFF by releasing SFe from a weather tower at the bank of
the Gan Jiang River in China, an area that has a combination of open field, agricultural
land, and forest. CALPUFF was found to be negatively biased with only 25-27% of data
within a factor of two of the observations. The authors concluded that CALPUFF did not
predict hourly dispersion well. Similarly, Ghannam and El-Fadel (2013) compared NO2
concentrations calculated using CALPUFF with NO2 measurements and observed that the
model severely underpredicted the measurements, sometimes by up to three orders of
magnitude. It was stated to have captured the temporal variability, although correlations
were not reported. Ghannam and El-Fadel (2013) attributed this underprediction to
underestimation of the emissions input to the model. The results of Cui etal. (2011) and
Ghannam and El-Fadel (2013). which indicated negative bias, are consistent with those of
Chang et al. (2003) for the sites where vertical dispersion may have played a larger role
in the airflow characteristics.
An example of where AERMOD has been used to better understand the relationship
between ambient concentrations and health risks is found in Maantay et al. (2009). These
researchers coupled AERMOD with geographic information system proximity buffers
around a stationary point source in Bronx, NY. They observed that buffers based on the
predicted plume shape for concentrations of NOx, PMio, PNfe.s, CO, and SO2
3-15
-------
corresponded better with asthma hospitalization rates compared with circular buffers
centered around the emissions source.
Subgrid Scale and Data Fusion Models
Substantial uncertainties at the subgrid scale remain when using CTM to model
concentrations at resolutions of 4-36 km (U.S. EPA. 2008c). In densely populated
regions of the country, monitor density may be finer than CTM surface grid resolution.
Moreover, CMAQ and other CTMs suffer from pollutant-specific concentration biases,
such as underestimation of total nitrate (Tuentes and Raftery. 2005). that require
correction prior to interpretation for exposure assessment. Bayesian Maximum Entropy
models for merging CMAQ and concentration data (Tuentes and Raftery. 2005) and
downscaling (Berrocal et al.. 2010a. b) have recently been developed to improve spatial
resolution and provide bias correction for the modeled concentration used as an exposure
surrogate, but such methods must be used with caution. For instance, Chen et al. (2014b)
ran a 36-km resolution CMAQ simulation for NO2, NOx, and other copollutants, fused
the CMAQ results with monitor observations, and compared both the raw and fused
model results with monitor observation data. The raw CMAQ simulations overpredicted
NO2 and NOx concentrations, particularly in the winter. These overpredictions were
substantially reduced (and in some cases the model slightly underpredicted
concentrations) for the fused model. Isakov et al. (2009) modeled subgrid spatial
variability within CMAQ using the AERMOD dispersion model prior to linking the
modeled results with microenvironmental exposure models to predict annual and seasonal
variation in urban population exposure within urban microenvironments. In each case,
these papers have referred to other air pollutants, but the methodology is still applicable
for NO2 exposure prediction.
Berrocal et al. (201 Ob) proposed a downscaling approach combining monitoring and
CMAQ modeling data to improve the accuracy of spatially resolved Os model data.
Specifically, a Bayesian model was developed to regress CMAQ model estimates of Os
concentration on monitoring data, and then the regression model was used to predict
concentrations using the CMAQ model results as an input field. Berrocal et al. (2010a)
extended the approach to include two pollutants (ozone and PM2 5) in a single modeling
framework, and Berrocal et al. (2012) added smoothing processes that incorporate spatial
autocorrelation and correction for spatial misalignment between monitoring and modeled
data. Although these papers did not specifically use NO2 concentration data, the methods
can be applied for NO2 as they have been for Os and PM2 5. Bentayeb et al. (2014)
applied a similar data assimilation method in which local measurements and elevation
data were combined with CTM output in a geostatistical forecasting model. This
algorithm was applied for NO2, PMio, PM2s, SO2, CeHe, and Os. Correlations between
3-16
-------
assimilated values and measurements ranged from Pearson r = 0.75-0.90. Debry and
Mallet (2014) also employed data assimilation for forecasting but combines three CTMs
in an ensemble average to minimize the influence of their errors in conjunction with
assimilation of observation data. The method of Debry and Mallet (2014) reduced error in
hourly, daily, and peak NO2 concentrations by 19, 26, and 20%, respectively.
In a slightly different approach, Crooks and Isakov (2013) blended CMAQ, AERMOD,
and monitoring data for NOx, PM2 5, and CO using a Bayesian model based on a wavelet
basis series. In this method, the true exposure is represented by the B-spline wavelet
series, and then the CMAQ grid cell concentrations, AERMOD receptor concentrations,
and measurement points are represented by the wavelet field modified by some assumed
error. These components each comprise linear contributions to a Gaussian likelihood
model. For NOx, the model was found to favor CMAQ data when modeling background
and monitor data in dense urban areas where spatial variability is higher. The blended
model results had lower prediction error and bias compared with kriging when smaller
numbers of points were used for the kriging surface, although the blended model did not
perform as well as kriging when densely gridded data were available for that purpose.
Similarly, Robinson et al. (2013) used geographically weighted regression, which used a
combination of dispersion model results and monitoring data as input for a regression
model, to compute concentrations in local population centers. The authors then used
kriging to fill in gaps between those population centers. When compared with other
kriging methods, the geographically weighted regression approach produced the smallest
residual mean squared errors when modeling average NO2 concentrations across the U.K.
for the year 2004. Beevers et al. (2012) also blended CMAQ with a near-road dispersion
model and applied the blended model to estimate NOx concentrations in London, U.K.
(Beevers et al.. 2013). Predicted peak rush hour (6:00 a.m.-9:00 a.m.) NOx
concentrations exceeded observed NOx concentrations by roughly 25% at a heavily
trafficked road.
Scale Models
Scale models characterize relationships between parameters to create a simplified
relationship that can be applied despite the physical scale of the setting. For example in
atmospheric science, scale models may be used to compare parameters of mean and
turbulent air velocity or inertial and viscous forces to understand the relative influence of
each. Although scale models are not currently used for exposure assessment in
epidemiologic studies, they are described briefly here as emerging methods for potential
use in exposure assessment. Existing wind tunnel and observational data have been used
in scale models of wind movement that support NOx fate and transport modeling in the
presence of built structures. For example, the Operational Street Pollution Model
3-17
-------
(OSPM) uses scale modeling but is developed specifically to capture street canyon
recirculation. Berkowicz et al. (2008) developed a model that includes a turbulent mixing
velocity in the street canyon and free convection. Monthly and 6-month avg NC>2
concentrations were calculated using a turbulent plume model. Modeled concentrations
were compared with NO2 concentration measurements from a 1995 panel study and
found to agree reasonably well (6-12% negative bias; R2 = 0.75-0.81). However, Jensen
et al. (2009) made a more recent comparison of OSPM to NO2 and NOx concentration
data from a two-week passive sampling campaign in New York City as part of the MESA
Air study (obtained in 2005-6). Regression of the OSPM results against the passive
sampler data showed that the model underestimated NO2 and NOx concentrations by 56%
and 47%, respectively, with R2 = 0.28 and 0.51.
3.2.2.3 Microenvironmental Exposure Models
Microenvironmental models are not used for exposure assessment in epidemiologic
modeling because the stochastic component of the model can add measurement error to
the health effect estimate. However, microenvironmental exposure models inform the risk
assessment performed as part of the NAAQS review process. The state of the science for
microenvironmental exposure models has not changed substantially since the 2008 ISA
for Oxides of Nitrogen, as described in detail in Annex 3.6 (U.S. EPA. 2008a). Examples
of microenvironmental exposure models include Air Pollution Exposure (APEX),
Stochastic Human Exposure and Dose Simulation (SHEDS), and Exposure in Polis (or
cities) (EXPOLIS), which involve stochastic treatment of the model input factors (Kruize
et al.. 2003; Burke etal.. 2001). Dionisio etal. (2013) compared estimates of NOx
concentration from a central site monitor and AERMOD with estimates from the APEX
and SHEDS models. They observed that the microenvironmental models captured the
spatial variability of the concentration distribution well, but temporal variability produced
by the models differed from other concentration estimation methods.
Hybrid Microenvironmental Models
Hybrid microenvironmental exposure models use ambient air quality input from either
dispersion models or CTMs rather than from central site monitoring data. For example, in
a hybrid microenvironmental model developed by Isakov et al. (2009). the CMAQ model
is used to simulate concentrations for a coarse discrete grid. Next, local-scale
concentrations from point and mobile sources are estimated using Gaussian dispersion
modeling through AERMOD. In combination, these models produce an ambient air
quality estimate at the location of the receptor that is then input into APEX or SHEDS to
3-18
-------
estimate total human exposure. Isakov et al. (2009) observed that the omission of specific
point and traffic sources led to an underestimate in median concentration by up to a factor
of two, depending on location; these simulations were for benzene and PM2 5; NOx tends
to be comparable in spatial variability with benzene and more spatially variable compared
with PM2.5 (Beckerman et al., 2008).
Recent studies have considered the variability and uncertainty associated with hybrid
microenvironmental exposure modeling. Ozkaynak et al. (2009) considered uncertainty
and variability in simulations that involved estimating concentration, exposure, and dose
in separate compartments of a model. They found that uncertainty and variability
propagated from one compartment to the next. Zidek et al. (2007) addressed uncertainty
and variability in hybrid microenvironmental exposure modeling by using distributions of
input parameters in the exposure model framework rather than point estimates. These
models estimate time-weighted exposure for modeled individuals by summing exposure
in each microenvironment visited during the exposure period. Zidek et al. (2007) found
that use of distributions of input data allowed them to examine cases for potential
subpopulations with common characteristics. Note that both of these studies model PM
concentrations, but the findings are applicable to NOx concentrations.
Sarnat et al. (2013b) compared risks of cardiovascular and respiratory morbidity with
24-hour NOx concentrations and those of other primary and secondary air pollutants in
Atlanta, GA using various exposure metrics and models. Epidemiologic results based on
the mean, median, and 95th percentile of the estimated exposure distributions from
APEX were compared with measures from a central site monitor, regional background,
AERMOD, and a hybrid model that merged AERMOD output with regional background
data. NOx concentrations modeled with APEX were generally higher than those obtained
with the hybrid model, likely because the APEX model incorporates road activity levels
in their exposure estimates. Epidemiologic analyses for asthma/wheeze produced
statistically significant risk ratios for the APEX mean, median, and 95th percentile, but
these risk ratios were not statistically significantly higher compared with those computed
using alternate exposure assignment approaches.
3.2.3 Choice of Exposure Metrics in Epidemiologic Studies
Appropriateness of the exposure metric for a given study depends in part on
epidemiologic study design and spatial variability of the pollutant concentration.
Table 3-1 summarizes the methods described in Sections 3.2.1 and 3.2.2. Based on
epidemiologic studies using various methods for exposure assessment, Figure 3-1
illustrates the range of NO2 concentrations to which people may be exposed in different
3-19
-------
locations (HEI, 2010). Because this figure is the result of the HEI (2010) review, the data
points included were sampled over different temporal scales and using different sampler
types. The figure illustrates variability in concentrations across locations and also the
variability measured within a type of location. Given the natural variability of
concentrations over space and time, and given nuances of the specific exposure
assessment techniques, it is important to recognize the specific applicability and
limitations of each approach, as summarized in Table 3-1. Differences in sampling
methodologies may cause some of the variability observed in the figure.
Table 3-1 Summary of exposure estimation methods, their typical use in
nitrogen dioxide epidemiologic studies, and related errors and
uncertainties.
Method
Epidemiologic Application
Errors and Uncertainties in Exposure Estimates
Central site Short-term community time-series
monitors exposure of a population within a
city
Correlation between true outdoor concentrations and
outdoor measurements typically decreases with increasing
distance from the monitor (Sections 3.4.3 and 3.4.5),
potentially leading to reduced precision and exposure bias
towards the null
Long-term exposure to compare
populations among different cities
Potential for exposure bias and reduced precision if the
monitor site does not correspond to the exposed population
(Section 3.4.5)
Positive instrument bias in the central site monitor from
other ambient oxidized nitrogen products may lead to
exposure bias and reduced precision (Section 3.2.1.1)
Passive monitors
Short-term panel (personal
monitoring or fixed local site,
e.g., residence, school, work)
Positive instrument bias in the passive monitor from other
ambient oxidized nitrogen products or internal reaction of
NO and Os to produce additional NO2 within the passive
sampler body may lead to exposure bias and reduced
precision (Section 3.2.1.2)
Long-term exposure
characterization or LUR model fit
(monitors distributed across a city
or at fixed local sites)
Positive instrument bias as described for the short-term
panel studies above may add exposure bias and reduced
precision when LUR models are fit to passive sampler data
(Section 3.2.1.2)
Potential for exposure bias and reduced precision for
monitors sited at fixed locations (Section 3.4.5)
3-20
-------
Table 3-1 (Continued): Summary of exposure estimation methods, their typical
use in nitrogen dioxide epidemiologic studies, and related
errors and uncertainties.
Method
Epidemiologic Application
Errors and Uncertainties in Exposure Estimates
LUR
Long-term exposure, usually across Potential for exposure bias and reduced precision if the
a city but sometimes fit among solution grid is not finely resolved (Section 3.2.2.1)
multiple cities Potential for bias and reduced precision if the model is
misspecified or applied to a location different from where
the model was fit (Section 3.4.5)
IDW and kriging Long-term exposure across a city Potential for negative bias and reduced precision if sources
are not captured or overly smoothed (Section 3.2.2.1)
Spatiotemporal
modeling
Not reported
Not yet well understood (Section 3.2.2.1)
CTM
Long-term exposure, sometimes
within a city but more typically
across a larger region
Potential for exposure error and reduced precision when
grid cells are too large to capture spatial variability of
exposures (Section 3.2.2.2)
Gaussian plume Long-term exposure within a city
rlisnprsinn
dispersion
modeling
Potential for bias and reduced precision where the
dispersion model does not capture boundary conditions
and resulting fluid dynamics well (e.g., in large cities with
urban topography affecting dispersion) (Section 3.2.2.2)
Scale modeling Not reported
Not yet well understood (Section 3.2.2.2)
CTM = chemical transport model; IDW = inverse distance weighting; LUR = land use regression; NO = nitric oxide;NO2 = nitrogen
dioxide; O3 = ozone
Source: National Center for Environmental Assessment.
3-21
-------
1 uw
S""
€
c
O 1 nn
co
^— »
i
0
O-m
IU
1
H
1
„ I
— * X ;
: i I
*
.i..
i
Sampling time
A < 24 hr
j ' : :
i : !
• ! . i :
jt ft.
*
•
Nitrogen Dioxide
Note: hr = hour; ng/m3 = micrograms per cubic meters. Data presented in this figure were obtained using different types of samplers
and over different averaging periods.
Source: Reprinted with the permission of the Health Effects Institute, HEI (2010).
Figure 3-1 Average nitrogen dioxide concentrations measured in studies
using different monitor siting.
Concentrations measured by central site or near-road monitors are commonly used as
surrogates for human exposure in studies of both short- and long-term exposure to NC>2
(Section 3.2.1.1). Central site measurements are subject to positive bias from instrument
error. Correlation between measured central site concentration and concentration at some
distant point decreases with distance. Therefore, reductions in correlation between
measurements at a central site monitor and the true exposure at the location of individuals
in an epidemiologic study, caused by human activity and variation in sources over space
and time, can lead to bias towards the null and loss of precision in the exposure estimate
for studies of short-term exposure to ambient NCh. For epidemiologic studies of
3-22
-------
long-term exposure to ambient NC>2, the difference between the measured concentration
and the true exposure would result in exposure error.
Passive sampling can be used for panel studies, or when samples are integrated or
averaged over several weeks or months, as input for long-term studies (Section 3.2.1.2).
The integrated nature of the passive samples limits their application in time-resolved
studies. Passive sampling techniques such as Palmes tube measurements are subject to
positive instrumentation biases. Additionally, passive monitors left in place for sampling
durations of weeks may produce data having errors and uncertainties that are similar to
those associated with using a fixed-site monitor to capture exposures for a population that
is dispersed over space and moving in time. The influence of exposure error in passive
sampling methods is discussed in more detail in Sections 3.4.5.2 and 3.4.5.3. Passively
sampled concentrations are also used commonly as input for LUR model fitting
(Section 3.2.2.1).
LUR is often employed to estimate NO2 concentrations for use in long-term exposure
studies. The quality of the exposure metric provided by the model depends on several
factors, including spatial resolution of the model, the number of model training sites, the
selection of predictor variables, consideration of seasonality, and model validation. IDW
is also used for exposure estimation between spatially distributed NO2 concentration
measurements (Section 3.2.2.1). However, if too few monitors are used, then the IDW
might not capture the spatial variability of the true exposures.
CTMs and dispersion models are based on physics of air flow and contaminant transport
(Section 3.2.2.2). Like central site monitors, CTMs can be used to compare NCh
concentrations among different cities for long-term exposure studies. However, coarse
spatial resolution of CTMs limits their applicability within cities. Dispersion models are
frequently used for within-city NC>2 concentration estimation in long-term exposure
studies, but the simplifying assumption of Gaussian dispersion can add error to the
concentration estimate if meteorology or topography of the built environment are
complex. Given this complexity, the direction of exposure error is not predictable. Biases
in dispersion model output can occur in either direction, and they depend strongly on the
specific environment (i.e., topography, meteorology, source representation) being
modeled. Correction methods may sometimes be applied to minimize such error for a
given location, but the effectiveness of error minimization must be determined on a
case-by-case basis. Subsequent sections will describe characterization of NO2
concentrations, a conceptual model of exposure, relationships among exposure metrics,
sources of exposure error, confounding, and implications of exposure error for
epidemiologic studies of different designs.
3-23
-------
3.3 Characterization of Nitrogen Dioxide Exposures
Chapter 2 presents detailed data on NCh sources and concentrations. The purpose of this
section is to present NO2 concentration data used as surrogates for human exposure to
NO2. It is broken into two parts: NC>2 concentration as an indicator of source-based
mixtures and indoor dynamics. The section on NC>2 concentration as an indicator of
source-based mixtures presents ambient NO2 concentration data reported in studies of
human exposure to mobile source emissions and other outdoor sources. The section on
indoor dynamics describes sources, sinks, and penetration of ambient NC>2 into indoor
environments and the chemistry influencing indoor concentrations of NC>2. This section
provides context for the discussion of exposure assessment and factors influencing
epidemiologic inference in Section 3.4.
3.3.1 Nitrogen Dioxide Concentration as an Indicator of Source-Based
Mixtures
3.3.1.1 Mobile Source Emissions
Seventeen percent of U.S. homes, or 22,064,000 homes, are located within 91 m of a
highway with four or more lanes, a railroad, or an airport (U.S. Census Bureau. 2009).
Moreover, 7% of U.S. schools serving 3,152,000 school children are located within
100 m of a major roadway, and 15% of U.S. schools serving 6,357,000 school children
are located within 250 m of a major roadway based on data from the National Center for
Education Statistics (NCES) (Kingsley etal.. 2014). The NCES did not specifically
define traffic in terms of annual average daily traffic [AADT], predominant fuel type
used on the roadway, or other criteria besides number of lanes. Average one-way
commuting times for the U.S. labor force working outside the home are 19.3 minutes for
bicyclists, 11.5 minutes for walkers, and 25.9 minutes for all other modes of
transportation. Among the populace working outside the home, 15.6% spend 45 minutes
or more commuting to work each day (U.S. Census Bureau. 2007). Based on Figure 2-4.
the proportion of NOx emissions from mobile sources in the 21 CBSAs with at least
2.5 million residents is 16% higher than it is among the general population. Hence, a
large share of the U.S. population is exposed to the on- and near-road environment on a
regular basis, and those exposures are likely to be higher for the 38% of the population
living in urban areas (U.S. Census Bureau, 2013b). This has implications for potential
NO2 exposure. Section 2.5.3 describes spatial patterns of NCh concentrations near roads
as a background for understanding traffic-related NO2 exposure. This section builds on
3-24
-------
the observations of NO2 concentration gradients described in Chapter 2 to consider how
near-road concentrations influence traffic-related NCh and NOx exposure.
Time spent in traffic can be an important determinant of personal NCh exposure. Molter
et al. (2012) calculated associations with time spent in several home, transit, and school
microenvironments for a cohort of 12-13-year-old children from Greater Manchester,
U.K. based on 2-day sampling periods per season; traffic data were not provided. They
observed that time spent in transit was positively associated with both NO2 exposure and
mean prediction error of a microenvironmental model of personal NC>2 exposure, where
mean prediction error compares the microenvironmental model with NCh measurements.
Together, these findings suggest that exposures are higher on roads and consequently that
time spent in transit may comprise a larger share of daily NC>2 exposure compared with
the proportion of time in a day that is spent in transit. Ragettli etal. (2014) estimated the
exposures of commuters who reported the times and routes of their commutes and modes
of transport to the 2010 Swiss Mobility and Transport Microcensus in Basel, Switzerland.
The authors used concentration estimates from a combination of dispersion modeling and
LUR. Traffic data were not provided. Ragettli etal. (2014) found that travel in motor
vehicles on highways and class 1 roads (AADT limits not defined for these road types)
produced the highest exposure (reported as the product of concentration and time),
followed closely by bicyclists and those taking public transit. Pedestrians had measurably
lower exposures.
Health studies often focus on the independent effects of NO2 exposure or use NO2
concentration as a surrogate for exposure to traffic pollution mixtures when
measurements of other pollutants are unavailable. NO2 concentration is routinely
measured at sampling sites nationwide, and NO2 is a prevalent reaction product of NO,
which is a component of vehicle exhaust (Section 2.2). Correlation data from several
studies presented in Section 3.4.4 illustrate that NO2 concentration generally correlates
with concentrations of other traffic-related pollutants in urban areas. NO2 concentration
has also been observed in at least one study to correlate with nonconcentration measures
of traffic. With respect to exposure, these observations make it hard to distinguish NO2
from other pollutants when considering the health impacts potentially attributable to each.
As a surrogate for traffic-related exposure, NO2 concentration may do an adequate job of
capturing spatial and temporal trends of traffic pollution. Microscale spatial variability of
NO2 concentrations near roads has been studied extensively, and NO2 concentration
gradients from a number of studies are summarized and compared in Section 2.5.3. Based
on 1-2 weeks of passive sampling measurements for NO2, Wheeler et al. (2008) and
Beckerman et al. (2008) reported correlations among NCh and several traffic-related air
pollutants, including benzene (Pearson r = 0.85) and toluene (r = 0.63). Beckerman et al.
3-25
-------
(2008) reported that 349,100 to 395,400 vehicles traveled daily on the highways studied;
Wheeler et al. (2008) did not report traffic data. The near-road air pollutant gradients
displayed in the review by Karner etal. (2010) suggested that NO2 concentration is
correlated with concentrations of traffic-related air pollutants across various distances
from a roadway. McCreanor et al. (2007) similarly found much higher personal exposure
measurements of NO2 when subjects in a scripted exposure study walked alongside a
heavily trafficked (traffic data not provided) road in London that is limited to diesel truck
and bus traffic compared with personal exposure measurements of NO2 when walking in
a park (road: median NO2 = 75.5 ppb; park: median NO2 = 18.4 ppb). These studies
concluded that gradients in NCh concentrations were spatially correlated with gradients in
traffic-related pollution.
The size and shape of the near-road gradient for NO2 concentration determines the spatial
zone where near-road exposures are most likely. Observations of the structure of the NO2
near-road concentration gradient are summarized in Tables 2-6 and 2-7 in Section 2.5.3.
Although NO2 concentration tends to correlate with most roadway pollutants in a
near-road environment, the NC>2 concentration gradient tends to be shallower than
gradients for other primary traffic-related pollutants (e.g., CO, UFP). These gradients
influence how exposure and copollutant correlations change spatially across the near-road
environment. Karner etal. (2010) performed an analysis of 125 near-road NO2
concentration studies and observed a concentration reduction of 42% from the edge of the
roadway, in line with values presented in Tables 2-6 and 2-7. The review of Karner et al.
(2010) also showed that the NO2 concentration gradient was much less steep compared
with the concentration gradients for NO and NOx, with decay to background levels
within 550 m. In contrast, Karner et al. (2010) reported a 79-86% reduction in UFP
concentration and a 90% reduction in CO concentration from the roadway edge. These
results suggest that, although NO2 concentration may capture many aspects of pollutant
gradients from the roadway, NO2 concentration used as a marker for traffic may
underestimate the magnitude of the concentration gradient for other near-road pollutants,
such as UFP and CO. Figure 3-2 presents the spatial variability of NO2 and copollutants
at various gradients from the roadway reported in the Karner etal. (2010) paper to
illustrate the comparison of the spatial near-road gradient of NO2, NO, and NOx
concentrations with those of other traffic-related pollutants. In a later study of near-road
concentrations in Medford, MA, Padro-Martinez et al. (2012) used continuous
instrumentation mounted on a mobile sampling unit operated over the course of a year, to
illustrate a similar gradient; traffic levels were not provided, but heavy-duty vehicles
comprised nearly 20% of total traffic.
3-26
-------
V
I
o
1,
-2 L2
o n.s
e
E
O D 6
V
-S ,
O
u
I
I
1 1 1 1 1
Rapid: >50% drop by 150 m
f,f* fl*4l
EC<4fl)
NO (67)
UF1 Particte no (76)
lf\ UF2 Particle no. (93)
£*. VOC1 (80)
kt--l1 ^- ' ~ • — • •*„ ^ •- ~
,\~* * - - - - '"^' — - ^
V\ ~'")s'ir.-i .-,--'
\ X " --» ,
'*-,."^"~r
0 100 200 300 400
t i i i i
Less rapid or gradual decay
- . PW, . Hfi1 \
^Xrr_ _
i i i i i
No trend
^•s.- .r""i-"~~.T."—
*••*»
VQC2 (3i)
100 200 300 400
100 200 300 400
nee from eriqe (m)
Note: Concentrations are normalized by measurements at the edge of the road. NO2, NO, and NOX concentration gradients are
presented in the center panel. NO2 = nitrogen dioxide, NO = nitric oxide, NOX = sum of NO2 and NO; CO = carbon monoxide;
m = meter; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PM10 = particulate
matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; EC = elemental carbon; VOC1 = volatile organic
compounds whose concentrations varied with distance from the road; VOC2 = volatile organic compounds whose concentrations
did not vary with distance from the road; UF1 = ultrafine particles larger than 3 nm; UF2 = ultrafine particles larger than 15 nm.
Data presented from Karneret al. (2010) were synthesized from 41 peer-reviewed references, 1 1 of which reported data for NO2,
5 of which reported data for NO, and 6 of which reported data for NOX. The number in parentheses refers to regression sample size.
UF1 and UF2 are measures of ultrafine particle number.
Source: Reprinted with permission of the American Chemical Society, Karneret al. (2010);
Figure 3-2 Spatial variability in concentrations of near-road nitrogen dioxide,
nitric oxide, the sum of nitric oxide and nitrogen dioxide, carbon
monoxide, and other near-road pollutants.
As pointed out in Section 2.5.2. near road NO concentrations are typically much higher
than near road NO2 concentrations; Table 2-6 describes the near road concentration
gradient for NC>2 only. Table 3-2 expands on these observations to consider on-road
concentrations of NC>2 and NO while in transit. Recent on-road and near-road
measurements of both NO and NO2 concentrations indicate that on-road NO exposures
can be much higher than on-road NO2 exposures immediately upon their emission. In
particular, the Los Angeles, CA data for NOx and NO concentrations suggest that rush
hour NO2 concentrations are roughly 50-60 ppb, but NO concentrations reach roughly
200-360 ppb in the morning and 95-260 ppb in the afternoon, based on 2-hour avg of
1-minute data (Tujita et al.. 2011). These studies were conducted on four Los Angeles,
CA highways with traffic levels ranging from approximately 190,571 to 289,167 vehicles
per day with roughly 4-15% of traffic comprised of heavy-duty diesel traffic. Beckerman
et al. (2008) measured 1-week integrated NO and NO2 concentration samples next to two
highways in Toronto, Canada and also observed that mean NO concentrations were
3-4 times higher than mean NO2 concentrations.
3-27
-------
The relationship between NO2 concentration and traffic metrics informs exposure
assessment because it establishes potential for exposure among those commuting or
living in the near-road environment. In Minneapolis, MN, Pratt et al. (2014) compared
direct traffic metrics, such as traffic volume, with LUR-computed NCh concentrations
(which were not estimated from traffic volume although road length was included in the
model). They observed a correlation (type unstated) of 0.58 between NCh concentration
and traffic density (AADT per km2), with a slope of 0.103 on a log-log model of NO2
versus traffic density. Gauderman et al. (2005) measured the correlation between NO2
concentrations and various traffic metrics in 12 southern California communities. Traffic
on major roadways within these communities was stated to range from 50,000 to
270,000 vehicles per day. On average across the communities, the Spearman correlation
between NCh concentration and increasing distance to freeway was r = -0.54, but the
correlation between NO2 concentration and traffic volume within 150m of a freeway was
r = 0.24.
The contribution of mobile source emissions to NC>2 concentration varies with strength of
additional sources. For example, Ducret-Stich et al. (2013) modeled NO2 concentration as
a function of background NC>2 concentration; light-duty (highway: 19,010 vehicles per
day, main road: 7,047 vehicles per day) and heavy duty (highway: 3,030 vehicles per day,
main road: 143 vehicles per day) traffic counts; and meteorological, topographic, and
temporal variability in the Swiss Alps with a model out-of-sample R2 = 0.91. They
observed that background NO2 concentration contributed 83% of the variability in the
model, while heavy-duty and light-duty traffic counts contributed 8 and 7%, respectively.
Similarly, NOx has been found to have mixed correlation with traffic density in a
nationwide long-term exposure epidemiologic study of the U.S. Veterans Cohort
[1976-2001 (Lipfert et al.. 2009)1. In this study, annual vehicle-miles traveled per square
mile averaged 12.17 (standard deviation: 15.3). In areas deemed high traffic density
(higher traffic than the average 1985 traffic density), Pearson r = 0.27, while for areas of
low traffic density (lower traffic than the average 1985 traffic density), r = 0.56.
3-28
-------
Table 3-2 Near- and on-road measurements of nitrogen dioxide, nitric oxide, and the sum of nitric oxide and
nitrogen dioxide.
Reference
fBeckerman et
al. (2008)
fZhu et al.
(2008)
tFuiita et al.
(2011)
Location and Date
Toronto, Canada Site
1, Aug2004
Toronto, Canada Site
2, Aug 2004
Los Angeles, CA
1-710 (mostly diesel
trucks), NR
Los Angeles, CA
I-405 (mostly autos),
NR
Los Angeles, CA
1-110 (mostly autos),
Sep-Dec2004
Los Angeles, CA
I-405 (mostly autos),
Sep-Dec2004
Los Angeles, CA
SR-60 (mostly autos),
Sep-Dec2004
Distance to Road
(m)a Traffic Counts
28,47, 57, 107,
126, 194, 209, 382, 395,400 veh/day
507, 742, 986
4,28, 38,56, 105,
1 1 4, 1 75, 246, 335, 349, 1 00 veh/day
346, 438, 742, 875
n NR; up to
25% diesel
0 NR; about 5% diesel
259,500-261,000
_ veh/day;
12,631-13,375
heavy-duty veh/day
276,857-289,167
veh/day; 12,755-
13,100 heavy-duty
veh/day
249,333 veh/day;
0 16,775 heavy-duty
\/ph/rla\/
Averaging Time NO (ppb) NO2 (ppb)
1-week integrated 44.2 (19.9)b; 77. 6C 14.6 (2.8)b; 18. 6C
1-week integrated 70.5 (62.7)b; 239.3C 17.5 (4.6)b; 28.2C
2-havgof1-min .|R .|R
data unfiltered
2-havgof1-min
data unfiltered NR NR
2-havgof1-min b
data (morning) wr (ttt) INK
2-havgof1-min 198(g4)b NR
data (morning) lae^ NK
2-havgof1-min QOQM14V MR
data (morning) 329(114) NR
NOx (ppb)
NR
NR
432 (66.3)b
267(114)b
411 (250)b
245(100)b
388(120)b
3-29
-------
Table 3-2 (Continued): Near- and on-road measurements of nitrogen dioxide, nitric oxide, and the sum of nitric
oxide and nitrogen dioxide.
Reference
tFuiita et al.
(2011)
(Continued)
fFruin et al.
(2008)
Location and Date
Los Angeles, CA
1-710 truck route,
Sep-Dec2004
Los Angeles, CA
1-110 (mostly autos),
Sep-Dec2004
Los Angeles, CA
I-405 (mostly autos),
Sep-Dec2004
Los Angeles, CA
SR-60 (mostly autos),
Sep-Dec2004
Los Angeles, CA
truck route, Sep-Dec
2004
Los Angeles, CA 1-10
(mostly autos),
Feb-April 2003
Los Angeles, CA
1-710 (mostly diesel
trucks), Feb-April
2003
Distance to Road
(m)a
0
0
0
0
0
0
0
Traffic Counts
190,571 veh/day;
28,502 heavy-duty
veh/day
259,500-261,000
veh/day; 12,631-
13,375 heavy-duty
veh/day
276,857-289,167
veh/day; 12,755-
13,100 heavy-duty
veh/day
249,333 veh/day;
16,775 heavy-duty
veh/day
190,571 veh/day;
28,502 heavy-duty
veh/day
Approx. 10,000
heavy-duty veh/day
Approx. 25,000
heavy-duty veh/day
Averaging Time NO (ppb)
2-havgof1-min b
data (morning) M\ (\^)
2-havgof1-min 95 (4gy>
data (afternoon) ( >
2-havgof1-min qft ,„,„
data (afternoon) ao l°Dj
2-havgof1-min H2(55)b
data (afternoon) { >
2-haygof1-min 258(114)b
data (afternoon) v '
2-to-4-h avg of d
20-s data Zti(J
2-to-4-h avg of d
20-s data 39°
N02 (ppb) NOx (ppb)
NR 426(154)b
NR 148(62)b
NR 140(64)b
NR 170(65)b
NR 321 (125)b
NR NR
NR NR
3-30
-------
Table 3-2 (Continued): Near- and on-road measurements of nitrogen dioxide, nitric oxide, and the sum of nitric
oxide and nitrogen dioxide.
Reference
Location and Date
Distance to Road
(m)a
Traffic Counts Averaging Time
NO (ppb)
N02 (ppb)
NOx (ppb)
tMacNauqhton Boston, MA bike path
et al. (2014) separate from vehicle
traffic, NR
12,900 veh-km
Average over 40
3-h sampling
periods with 1-min
data
NR
14.7(0.582)b
NR
Boston, MA bike lane
adjacent to vehicle
traffic, NR
6,250 veh-km
Average over 40
3-h sampling
periods with 1-min
data
NR
19.5(0.343)b
NR
Boston, MA
designated bike lane
shared between bikes
and buses, NR
5,220 veh-km
Average over 40
3-h sampling
periods with 1-min
data
NR
24.2(1.72)b
NR
Aug = August; avg = average; CA = California; Dec = December; Feb = February; h = hour; I = interstate; m = meters; MA = Massachusetts; min = minute; NO = nitric oxide;
NO2 = nitrogen dioxide; NOX = the sum of NO and NO2; NR = not reported; ppb = parts per billion; s = second; Sep = September; SR = state route; veh/day = vehicles per day;
veh-km = vehicle-kilometers.
aDistance of 0 m indicates on-road measurements.
bAverage (standard deviation).
°Maximum.
dAverage of medians.
fStudies published since 2008 ISA for Oxides of Nitrogen.
3-31
-------
Natural experiments provide an opportunity to test the sensitivity of ambient NC>2
concentration to changes in traffic conditions. For example, Levy et al. (2006) measured
NO2 concentrations across the Boston, MA metropolitan area using one-week integrated
samples that were deployed the week prior to the 2004 Democratic National Convention
(DNC), during the week of the 2004 DNC, and during the week after the 2004 DNC.
Traffic data were not reported for this study. Levy et al. (2006) hypothesized that there
were four types of sites: (1) sites with concentration decreases around closed-down roads
that were not near alternate routes, (2) sites with concentration increases around alternate
routes but not near closed roads, (3) sites with no change because they were not near
closures or alternate routes, and (4) sites with unclear impacts. Sites hypothesized to have
decreasing NO2 concentration did in fact have a median 42% reduction in concentration.
Likewise, sites hypothesized to have increasing NCh concentration did have a median
15% increase in concentration. Sites hypothesized to have no change or unclear impacts
had 12% and 30% reductions in NCh concentration, respectively.
Several recent studies have evaluated the use of central site NO2 or NOx concentration as
a surrogate for personal exposure to traffic pollution mixtures. In a near-road
environment, NOx concentration can be correlated with concentrations of pollutants that
are also associated with health effects, including UFP and water soluble metals (Sanchez
Jimenez et al.. 2012): PAHs (Brook et al.. 2007): sum of the VOCs benzene, toluene,
ethylbenzene, and xylene (BTEX) (Beckerman et al.. 2008): and EC (Minguillon et al..
2012) or BC (Clougherty et al.. 2013). Correlations generally in the range of 0.6-0.8 of
NC>2 concentration with CO, NOx, and EC (or BC) concentration forms the basis for a
proposed multipollutant mobile source indicator that combines these three species into an
Integrated Mobile Source Indicator (IMSI) for traffic-related air pollution. The IMSI is a
weighted average of mobile source pollutant concentrations weighted by the ratio of
mobile source to total emissions for each pollutant, which Pachon et al. (2012) developed
using CO, NOx, and EC. Although the IMSI is not currently used in any epidemiologic
studies of the health effects of NO2 or NOx, it is an informative tool that may shed light
on the relationship between traffic-related sources and human exposures, as shown in
Equation 3-1.
EmissionEc,mobiie c/ EmissionNOxmobile EmissionCo,mobiie c/
EmissionEC
-------
Note that C" = average concentration normalized by the standard deviation of
concentration. Urban street-side (mostly street canyon) NO and NC>2 concentrations have
been measured and compared with downwind sites, including those located in parks and
reference sites (i.e., sites that are located away from or upwind from traffic-related
emissions). Criteria pollutant concentrations were sampled using high-density siting
throughout the five boroughs of New York City, NY with 2-week integrated samples per
season (25 samplers deployed for 2 weeks during 6 sampling sessions per season to make
150 sites within New York City, NY) (Ross etal.. 2013). Consistent with Karneretal.
(2010). the street-side sites generally showed higher NO concentrations compared with
NO2 concentrations (NO: mean 31.82 ppb, max 151.76 ppb; NO2: mean 27.60 ppb, max
87.18 ppb) in Ross etal. (2013) (see Table 3-3) for concentration, traffic, and land use
summary data). The NO concentration on average was lower than the NO2 concentration
away from the road, for example at park sites (NO: mean 18.88 ppb, max 45.15 ppb;
NO2: mean 22.13 ppb max 36.94 ppb). The ranges for overall and truck traffic density,
census population, and building areas were all higher for the street-side sites compared
with the park sites. In a mobile van study of street canyons in Helsinki, Finland operating
continuous monitors during rush hour on a roadway with approximately 40,000 vehicles
per day (roughly 10% diesel, with sampling interval: 1-minute), Pirjolaetal. (2012)
found that the topographical characteristics of the roadway influenced the concentration
gradient. They studied concentration profiles on the upwind and downwind sides within a
street canyon and observed downwind-to-upwind concentration ratios of 0.28 and
0.70 for NO and NO2, respectively, when the street canyon aspect ratio (building
height-to-street width) was 0.55. When the aspect ratio increased to 0.70,
downwind-to-upwind concentration ratios decreased to 0.18 and 0.65 for NO and NO2,
respectively.
3-33
-------
Table 3-3 Summary (mean, range) within 300 m of monitoring sites, by site
type, in a spatially dense monitoring campaign in New York City, NY,
based on 2-week integrated samples per season.
NO2 concentration (ppb)
NO concentration (ppb)
Roadway length (km)
Traffic density (vehicle-km/h)
Truck density (vehicle-km/h)
2000 census population (number)
Building area (m2)
Residential space area (m2)
Commercial space area (m2)
Industrial space area (m2)
Street-Side Sites3
(n = 138)
27.6 (8.32-87.2)d
31.8(2.69-152)d
4.3-6.0
561-2,800
13.4-83.2
1,316-5,819
90.7-382
53.79-242
15.6-105
0-7.19
Park Sites3
(n = 12)
22.1 (8.10-36.9)d
18.9(4.93-45.2)d
2.1-3.7
302-2,560
0.910-24.4
117-3,455
0-163
0-124
0-29.0
0-4.65
Reference Sitesb
(n = 5)
20.2 (9.43-38.2)d
15.9(5.42-54.8)d
1.9-2.5
119-783
5.80-13.5
0-522.7
0-38.5
0-30.7
0-18.7
0-0
Regulatory Sites0
(n = 5)
22.7(17.1-34.2)d
12.1 (3.30-40.0)d
km = kiilometer; km/h = kilometer per hour; m2 = meters squared; n = sample size; NAAQS = National Ambient Air Quality
Standards; NO2 = nitrogen dioxide; NO = nitric oxide; NY = New York; ppb = parts per biillion.
a150 passive monitoring sites sampled at 25 sites over two weeks and then moved over six sampling periods, repeated over three
seasons for a total of 48 weeks of sampling.
bFive continuous monitoring sites distributed in each borough of New York City, NY.
°Five continuous monitoring sites maintained by New York State to assess NAAQS compliance.
davg (range).
Source: Modified from Table 1 and Table 5 of Matte et al. (2013).
NC>2 and NO emissions, concentrations, and therefore exposures, are also subject to
interventions in the built environment. After a tunnel was built in Sydney, Australia to
reduce urban pollution levels, Cowie et al. (2012) observed statistically significant
reductions in NC>2 and NOx concentrations by 1.4 and 4.6 ppb, adjusted for meteorology,
based on 2-week passive sampler measurements taken at three periods during fall
2006-2008. Traffic was reported to be 90,000 vehicles per day on the surface road and
43,446-57,814 vehicles per day in the tunnel. Beevers and Carslaw (2005) studied the
impact on annual NOx emissions of the London congestion pricing zone implemented in
2003 to reduce traffic in central London, U.K. Overall, they reported a 12% decrease in
NOx emissions within the congestion pricing zone and a 1.5% increase in NOx emissions
at the surrounding ring road, related to some individuals rerouting their drives to the
surrounding ring road where no payment was required. Traffic count data were not
3-34
-------
provided. Similarly, Panteliadis et al. (2014) studied the impact of congestion pricing in
Amsterdam, the Netherlands and observed a 6.6% reduction in NO2 concentrations at a
roadside measurement, with an 11% reduction in the traffic contribution to ambient NCh
concentrations. Approximately 15,000 vehicles per day travelled on the road. However,
Masiol et al. (2014) analyzed the effects of traffic-free Sundays over 13 years on air
quality in the Po Valley of Italy and saw no appreciable change in NO2 levels for a
roadway with traffic of more than 75,000 vehicles per day. Rao etal. (2014) studied the
influence of tree canopies onNCh concentrations in Portland, OR using LUR modeling
and observed a 38% reduction in NO2 concentrations related to increasing the tree canopy
at higher elevations in the city; traffic count data were not reported. MacNaughton et al.
(2014) measured NO2 exposures of bicyclists in Boston, MA using real-time monitoring
(3-hour avg of 1-minute data) equipment and GPS and observed that riding in a shared
bicycling/bus lane, traffic density, background NC>2 concentration, and vegetation density
were associated with measured NO2 exposures. The city of Beijing, China restricted
traffic during the 2008 Olympics, thus creating a natural experiment in pollution
reduction. (Zhang et al.. 2013) reported that the average of 1-hour NO2 concentration
measurements dropped from 25.6 ± 3.66 ppb to 14.6 ± 3.76 ppb when comparing periods
before (June 2-July 20) and during (July 21-September 19) the Olympic games. After
the Olympics (September 20-October 30), concentrations increased back up to
41.4 ± 3.81 ppb. Huang et al. (2012a) reported concentration reductions of 21.6% and
12.9% for the periods before and during the Olympics compared with the previous year.
The reduced NO2 concentrations that followed these interventions suggest that controls
can lead to reduced NO2 exposures. Traffic count data were not reported for the Beijing,
China Olympics studies.
3.3.1.2 Other Outdoor Sources
As described in Section 2.3. other sources contributing to ambient NOx emissions include
nonroad mobile sources, electric generating units, industrial sources, and wildfires.
Nonroad mobile sources, such as airports, shipping ports, and rail yards, can contribute
substantially to local and regional ambient NOx concentrations (Kim etal.. 2011;
Williams et al.. 2009; Vutukuru and Dabdub. 2008; Carslaw et al.. 2006; Unal et al..
2005). Carslaw et al. (2012a) took advantage of the natural experiment of the Icelandic
volcano eruption of 2010, when airports across Europe were shut down for 6 days, to
evaluate the local effect on airport NOx concentrations. Downwind of the airport, a 38%
reduction in average NOx concentrations (from 42 ppb down to 26 ppb) was observed. At
shipping ports and airports, traffic from ground-level support activities can also
contribute a large portion to NOx emissions from these sources (Klapmever and Marr.
3-35
-------
2012; Kim et al., 2011). Outside of urban centers where traffic is not a dominant source,
other sources of NOx may include wildfires and residential wood burning. As such, NOx
concentration may not always be a reliable proxy for traffic pollution. Section 2.3
discusses different sources of NOx in more detail.
3.3.2 Indoor Dynamics
3.3.2.1 Sources, Sinks, and Penetration
The general understanding of oxide of nitrogen production indoors has not changed since
the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). Indoor sources of oxides of
nitrogen are combustion-based, including gas stoves, gas heating, oil furnaces, coal
stoves, wood burning stoves, kerosene heaters, smoking, candle burning, and to a lesser
extent, electric cooking. The magnitude of indoor oxides of nitrogen depends on
ventilation of the indoor space and appliances, source strength, and rate of pollutant
reaction. Recent studies show associations between indoor combustion and indoor NO2
levels (Vrijheidetal., 2012; Kornartit etal. 2010; Park et al., 2008) or indoor NOx levels
(Cattaneo et al.. 2014). depending on what was measured during the study. HONO can
also be emitted directly during combustion or through surface reactions. Park et al.
(2008) measured HONO and NO2 concentrations during combustion and compared their
results with older studies in the peer-reviewed literature, as shown in Table 3-4. High
peak-to-mean ratios suggest high temporal variability of exposures that might be
differentiated from exposures of outdoor origin through time-series analysis. This review
also generally found higher HONO concentrations in the presence of indoor combustion
sources. Oxides of nitrogen can be lost through indoor deposition and ventilation (U.S.
EPA. 2008c). Sarwar et al. (2002) reported deposition velocities of 6-7 x 10~5 m/seconds
for NO2, HONO, HNO3, HO2NO2, NO3 ~, and N2O5. Much lower deposition velocities
(not detected-2 x 10~6 m/s) were reported for NO, PAN, and organic NOs species.
3-36
-------
Table 3-4 Indoor nitrogen dioxide and nitrous acid concentrations in the
presence and absence of combustion.
Study
Braueretal. (1990)3
Braueretal. (1990)b
Braueretal. (1991)c
Spenqler et al.
(1993)d
Simon and Dasqupta
(1995)e
Leaderer et al.
(1999)'
Khoder (2002)9
Lee et al. (2002)h
Combustion
Source
No source
(background)
Gas range3
Convective space
heater3
No source
Gas rangeb
Unknown
Gas range, stove,
furnace
Kerosene heater
No source'
Gas stoves'
Kerosene heaters'
No source'
Gas stoves'
Gas appliances
(summer)
Gas appliances
(winter)
Gas range, etc.
Measurement ~
Frequency
15 min
15 min
15 min
15 min
15 min
15 min
24 h
8 min
24 h
24 h
24 h
24 h
24 h
24 h
24 h
6 day
NO2
Peak
29
157
955
5.0
37
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
(PPb)
24-h avg
17
36
209
1.8
8
NR
fin c~)A 1 -I £>\
NR
NR
NR
NR
NR
NR
on /on 7Q\
65(27-120)
28 (4.3-52.0)
MONO
Peak
8
35
106
3.5
31
NR
NR
ID
NR 0
NR 4.
NR 6.
NR 2.
NR 5.
NR 3
NR 6.
NR 4.
(PPb)
24-h avg
5
13
42
3.4
9.6
1-12
47 C~) R\
./ (Z ti)
NR
.8 (0.0-2.9)
0(0.0-11.3)
8(0.2-35.9)
4(0.1-20.1)
5(0.4-20.1)
7 M Q 7 Q\
8(1.6-12.5)
6/n -i o-i -i \
3-37
-------
Table 3-4 (Continued): Indoor nitrogen dioxide and nitrous acid concentrations in
the presence and absence of combustion.
Study
Jarvis et al. (2005)'
tHonq et al. (2007V
tParketal. (2008)k
Combustion
Source
Gas hob
Gas oven
Gas range
Gas range
Measurement ~
Frequency
4 min
4 min
NO2
Peak
NR
NR
81.1
189.3
(PPb)
24-h avg
12.8
12.8
NR
19.4
MONO
Peak
NR
NR
9.3
15.2
(PPb)
24-h avg
4.1
5.0
NR
2.1
avg = average; h = hour; MONO = nitrous acid; IL = Illinois; kcal/h = kilocalories per hour; MA = Massachusetts; min = minute;
NM = New Mexico, NO2 = nitrogen dioxide; NR = not reported; ppb = parts per billion.
al_ocation: Chicago, IL research home, unvented combustion condition; gas range operation hours: 1 h (with one burner and
2,320 kcal/h); convective space heater operation hours: 4 h (with one burner and 2,785 kcal/h).
"•Location: Maryland research home, unvented combustion condition; gas range operation hours: 1 h (with one burner and
2,320 kcal/h).
°Location: 11 Boston, MA homes (winter).
dLocation: 10 homes in Albuquerque, NM (winter).
eLocation: Four different home environments with a small kerosene heater (2,270 kcal/h).
'Location: 58 homes (summer) and 223 homes (winter) in southwest Virginia and Connecticut; 39 inside homes without gas stoves
(summer); 19 inside homes with gas stoves (summer); 74 inside kerosene-heater homes (winter); 96 inside homes without
kerosene heaters and gas stoves (winter); 52 inside homes without kerosene heaters and with gas stoves (winter).
9Location: Four homes in suburban residential areas in Greater Cairo, Egypt.
hLocation: 119 homes in southern California (spring).
'Location: Homes in European community.
'Location: Living room of an apartment in Gwangju, Korea (May 2006).
location: Korean apartment (city and year unspecified, October).
Source: Reprinted with permission of Elsevier, Parket al. (2008).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
3.3.2.2 Indoor Chemistry
The 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) described well-established
reactions involving oxides of nitrogen and other indoor air pollutants for gas-phase and
surface chemistry that serves as both a source and sink for oxides of nitrogen. Knowledge
of indoor chemistry helps identify potential sources of uncertainty in estimates of indoor
exposure to ambient oxides of nitrogen. Moreover, epidemiologic studies of indoor
exposure may providing supporting evidence to the assessment of health effects from
ambient NCh exposure. Identification of the uncertainty in those exposure estimates can
aid interpretation of those studies.
For gas-phase reactions, indoor NO can be oxidized to NO2 via reaction with Os or HO2
radicals generated by indoor Os chemistry or VOCs found in household products. NO2
can react with Os to form NOs radicals that may subsequently oxidize organic
compounds. NO2 also reacts with free radicals to produce PAN. NO2 removed through
3-38
-------
surface reactions was known to contribute to NO levels indoors either by surface
reduction of NCh or by reaction of NC>2 with aqueous HONO on indoor surfaces (Spicer
et al., 1989). Conversion of NC>2 to HONO occurs through a number of indoor surface
reactions, and the reaction increases with increased relative humidity (U.S. EPA. 2008c).
A recent study has demonstrated the role of irradiance in humidity-driven surface
reaction of NO2 to HONO on paints (Bartolomei et al.. 2014). Surface reactions of NO
and OH radicals may also produce HONO, but the reaction rate is slower than forNO2.
Indoor combustion can lead to direct emission of NO and HONO, and conversion of NO
to NO2 can lead to secondary HONO production from heterogeneous reactions involving
NO2 on indoor surfaces. Park et al. (2008) observed HONO to be correlated with both
NO (Spearman r = 0.64) and NO2 (r = 0.68) during combustion. They noted that HONO
concentrations were 4-8% of NO2 concentrations during gas range operations but rose to
-25% of NO2 concentrations after combustion ceased, which underscores the role of
surface reaction as the major source of HONO production. In a model of combustion
products for oxides of nitrogen during candle and incense burning, Loupa and
Rapsomanikis (2008) observed simultaneous NO and HONO production, the latter of
which were in agreement with older test chamber results of HONO production during
combustion (De Santis etal. 1996). These studies on surface reactions of NO2 provide
insight into indoor NO2 sinks that may reduce NO2 exposures as well as exposures to
HONO, of which health effects are less well understood.
Recent gas-phase indoor chemistry work has shed light on processes involving organic
compounds and/or secondary organic aerosols (SOA). Carslaw et al. (2012b) modeled
indoor reactions forming SOA and observed that for their base case simulation, organic
nitrates constituted 64% of the overall SOA, while PANs constituted an additional 21%.
In sensitivity tests varying ambient concentrations and meteorological conditions, organic
nitrates varied from 23-76% of the SOA, and PAN varied from 6-42%. N0jgaard et al.
(2006) investigated the interference of NO2 in ozonolysis of monoterpenes in a
simulation of indoor air chemistry and observed that NO2 reacted with Os and hence
reduced SOA formation from ozonolysis of alkenes a-pinene and /?-pinene while
increasing the mode of the SOA size distribution. Intermediate NOs products may play a
role in this process, as described above. However, the presence of NO2 had less effect on
ozonolysis of J-limonene, and this is thought to occur because the ozonolysis reaction
rate is faster. In chamber experiments and computational chemistry models, Cao and Jang
(2008) and Cao and Jang (2010) tested toluene SOA formation in the presence of low
(<3 ppb), medium (90-135 ppb), and high (280-315 ppb) NOx concentrations. They
found that the organic matter component of the toluene SOA yield generally decreased
with increasing NOx concentrations, especially when high NO concentrations
(-222-242 ppb) were present. Ji etal. (2012) explored rate constants of NO2 reactions
3-39
-------
with various low molecular weight aldehydes found indoors and observed that the
reaction rates, k, increased in the following order:
^formaldehyde < fecetaldehyde < ^propanal < &butanal. JJ 6t al. (2012) Concluded from this observation
that NC>2 reacts more with longer chain, low-molecular-weight aldehydes compared with
shorter chain, low-molecular-weight aldehydes. RC(=O) radicals and HONO were both
observed to be products of these reactions. These sinks may result in lower NO2
exposures, but little information is available regarding organic nitrate reaction product
exposures.
Reactions involving N2Os (formed by reaction of NC>2 and NOs in the presence of another
molecule) in an indoor context have been studied in recent years. In an examination of
NOs and ^Os (measured as the sum of those two species) in an office building, N0jgaard
(2010) observed that alkenes remove more indoor NOs and N2Os than either ventilation
or surface deposition. Griffiths et al. (2009) studied N2Os uptake by organic aerosols in a
reaction cell and large chamber (260 m3) and observed little N2Os uptake by solid organic
aerosols, more efficient uptake by liquid aerosols, and uptake that increased with
increasing RH. ^Os uptake by dicarboxylic acids (oxalic acid, malonic acid, succinic
acid, and glutaric acid) was 30-90% of that by (NH^SCU and (NH^SCU-mixed
dicarboxylic acid aerosols at similar RH. ^Os uptake by malonic or azelaic acid in the
presence of higher RH is consistent with findings of Thornton et al. (2003) for
experiments conducted in a reaction cell. Raff et al. (2009) suggested that N2Os
autoionizes to NO2 + NOs and then reacts quickly with water to form HNOs; it is
possible that HNOs might then participate in the liquid aerosol reactions described by
Griffiths et al. (2009) and Thornton et al. (2003). Raff etal. (2009) also proposed
autoionization of N2Os as a likely mechanism for reaction with HC1, which would result
in nitrosyl chloride (C1NO) and HNOs formation while NC>2 and water vapor experienced
an intermediate surface reaction to form HONO, which would react with HC1.
Complexity of reactions involving ^Os in creating NO2 as an intermediary reaction
product also lends uncertainty to NO2 exposure assessment. This uncertainty may lead to
variability in personal or indoor NO2 exposure measurements.
3.4 Exposure Assessment and Epidemiologic Inference
The 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) examined several factors
influencing assessment of exposure to ambient oxides of nitrogen and measurements used
to represent exposures. These factors include high spatial and temporal variability of NO2
concentrations in urban areas and near roads, location of NO2 samplers, and ventilation of
indoor microenvironments. The 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c)
concluded that errors associated with the use of NO2 concentrations measured at central
3-40
-------
site monitors as exposure metrics for epidemiologic studies tended to bias the health
effect estimate towards the null for both short-term exposure and long-term exposure
epidemiologic studies. The following sections explore new evidence regarding a
conceptual exposure model, exposure metrics employed in epidemiologic studies,
personal-ambient relationships, factors that introduce exposure error, potential
confounding, and how the exposure errors may or may not introduce bias and uncertainty
into epidemiologic health effect estimates, depending on the epidemiologic study design.
3.4.1 Conceptual Model of Total Personal Exposure
Total personal exposure (£>) integrates the product of microenvironmental concentration
and fraction of time spent in a microenvironment across an individual's
microenvironmental exposures:
n
ET = / Citj
j
Equation 3-2
where Q = average NO2 concentration in the/th microenvironment, tj = fraction of total
time spent in the/th microenvironment, and n = total number of microenvironments
which the individual has encountered (U.S. EPA. 2008c; Klepeis etal.. 2001). Hence,
both the microenvironmental NC>2 concentration and time-activity aspects of total
exposure must be considered.
Alternatively, based on the principle of mass balance, an individual's total NO2 exposure
can be expressed as the sum of its ambient NO2 exposure (Ea) and nonambient NO2
exposure (Ena) components (U.S. EPA. 2008c; Wilson and Brauer. 2006):
Equation 3-3
Ea represents the amount of NC>2 exposure derived from outdoor sources, and Ena
represents the amount of NO2 exposure from indoor sources. The microenvironmental
formulation presented in Equation 3-2 and the component formulation presented in
Equation 3-3 can be rectified by recognizing that Ea and Em can both be expressed in
terms of microenvironmental concentrations and time spent in each outdoor and indoor
microenvironment. During the fraction of a day spent in each outdoor microenvironment
(yoj), ambient exposure to NO2 having an outdoor concentration of C0j is:
3-41
-------
Equation 3-4
Indoor NC>2 exposures in the/th microenvironment (Ey) are more complicated because
some part of indoor exposure may emanate from nonambient sources, and some part of
indoor exposure infiltrates from outdoors. Indoor exposures from nonambient sources are
given as Ena,j. Exposures in each indoor microenvironment from ambient sources are also
influenced by infiltration of outdoor NC>2 (INF}), time spent indoors (yy), and COJ:
Ei,j = ytjINFj • CoJ +
Equation 3-5
Infiltration is a function of the/th microenvironment' s air exchange rate (a/), air pollutant
penetration (/*,), and decay rate (&,):
Equation 3-6
Hence, indoor NCh exposure for microenvironment/ is the sum of the ambient and
nonambient components:
EiJ = yi.APJ aj/(aj + kj)]co.j + Ena,j
Equation 3-7
Finally, Ea can be described as the sum of the outdoor NO2 exposure and the ambient
component of the indoor NC>2 exposure, summed over/ indoor microenvironments (U.S.
EPA. 2008c: Wilson and Brauer. 2006: Wilson et al.. 2000):
7 = 1 7 = 1
Equation 3-8
Ambient concentration of NCh is often used as a surrogate for human exposure. In
concert, a second simplifying assumption is often made that the exposed individual
resides in one indoor microenvironment, such that time-activity data are reduced to "time
indoors" and "time outdoors." Errors associated with this approach, which may vary
depending on the epidemiologic study design in which the exposure surrogate is used, are
described in detail in Section 3.4.3. In this case, outdoor microenvironmental NC>2
exposures (E0) are expressed simply as the product of the fraction of all time spent
3-42
-------
outdoors (y0) and ambient NCh concentration (Ca): E0 = y0Ca. Furthermore, based on the
assumption that the individual occupies only one indoor and one outdoor
microenvironment, then the infiltration term can be simplified to yi[(P x a)l(a + k)}, and
because j0 +J; = 1:
Ea=
Equation 3-9
Then, an exposure factor (a) can be defined to express the influence of time -weighting
and infiltration on NC>2 exposure:
a = y0 + (1 -y0)[(P x o)/(o + fc)]
Equation 3-10
Last, an approximate expression for total personal exposure is obtained:
ET = aCa + Ena
Equation 3-11
Comparison of Equations 3-3. 3-9, and 3-11 reveals that a can also be approximated as
the ratio EJCa. Subsequent sections examine how Ea, a, and Ca are modeled or measured,
and how errors and uncertainties in the simplifying assumptions behind Equations 3-9.
3-10. and 3-11 may influence health effect estimates computed from epidemiologic
studies of varying design.
3.4.2 Personal-Ambient Relationships and Nonambient Exposures
Personal exposure measurements typically capture both ambient and nonambient
exposure contributions; for the purpose of this document, these are referred to as "total
personal exposure" measurements. The 2008 ISA for Oxides of Nitrogen (U.S. EPA.
2008c) concluded that literature relating ambient NCh concentrations measured by a
central site monitor to personal NCh exposures was mixed for studies of both short-term
and long-term NCh exposure, with some studies finding associations between the
personal and central site monitors and other studies finding no association. These
inconsistencies reflected various factors that influence exposure in respective studies,
including proximity and strength of sources of ambient and nonambient NOx,
spatiotemporal variability of NO2 concentrations, and time-activity behavior of the
exposed sample population. Recent studies have found that personal NCh concentration
measurements taken for adults and children tend to be more highly correlated with indoor
concentrations compared with outdoor or ambient concentrations, although wide
variability in the correlations was observed (Tables 3-5 and 3-6). Personal-outdoor
3-43
-------
(i.e., measurements taken outdoors but not at a central site monitor) correlations also
tended to be higher for summer compared with winter. This is not surprising because
open windows and greater time spent outdoors during summer likely increase exposure to
outdoor air (Brown et al.. 2009). The study results indicate that, for epidemiologic studies
of short-term exposure, indoor sources of NCh can add noise to the ambient NC>2
exposure signal. As described further in Section 3.4.5.1. uncertainty in the NC>2 exposure
term can lead to negative bias and added uncertainty in the epidemiologic health effect
estimate for short-term exposure studies.
3-44
-------
Table 3-5 Ambient, outdoor, transport, indoor, and personal nitrogen dioxide measurements (ppb) across
studies.
Study
fSarnat et al.
(2012)
fWilliams et al.
(2012b):
tMenq et al.
(2012a)
tSuh and
Zanobetti(2010b)
Location Time Period
El Paso, TX Jan-May,
(large city) 2008
Ciudad
Juarez,
Mexico (large
city)
Detroit, Ml Summer,
(large city) 2004-2007
Winter,
2004-2007
Metropolitan Fall, 1999-
Atlanta, GA Spring, 2000
(large city)
Ambient Outdoor
Sampling (Central (Outside
Interval Site) Residence) Transport Indoor
96-h 14.0-20.6C 4.5-14.2c NRd 4.0-8.1C
NR 18. 7-27. 2C NR 23.1-120.8
24-h Williams: NR NR NR
22.0e;
Meng:
22.0e; 22.7C
24.0e;23.9c NR NR NR
24-h 17.968; NR NR NR
17.13C
Personal
NR
NR
Total:
Williams: 25. 5C;
Meng: 25. 4C
Ambient: 16. Oe;
21. Oc
Total:
24.0e; 35.6C
Ambient:
18.0e;20.4c
8.08e; 11.60C
Personal-Ambient
Slopea'b
NR
NR
Meng: 0.24; 0.13f
Meng: 0.08; 0.07f
NR
3-45
-------
Table 3-5 (Continued): Ambient, outdoor, transport, indoor, and personal nitrogen dioxide measurements (ppb)
across studies.
Study
tBrown et al.
(2009)
fDelfino et al.
(2008a)
tPelqado-Saborit
(2012)
Location
Metropolitan
Boston, MA
(large city)
Riverside and
Whittier, CA
(SoCAB)
(large city)
Birmingham,
U.K. (large
city)
Ambient Outdoor
Sampling (Central (Outside
Time Period Interval Site) Residence) Transport
Nov 24-h 25.89; 26.8C NR NR
1999-Jan
2000
Jun-Jul2000 22.09; 22.8C NR NR
Jul-Dec2003 24-h 25.3e; 25.0C NR NR
(Riverside);
Jul-Dec2004
(Whittier)
Jul-Oct2011 5-min 47C 64C Car: 40C
Bus: 71C
Bike: 125C
Train: 58C
Personal-Ambient
Indoor Personal Slopea>b
NR 10.49; 12.9= All: 0.19
Windows closed:
0.09
Windows open:
0.31
LowAER: 0.21
HighAER: 0.15
NR 13.99; 17.4C All: 0.23
Windows closed:
0.64
Windows open:
0.10
LowAER: 0.34
HighAER: 0.19
NR 26.7e; 28.6C NR
Office: 14C All: 23C 1-h avg: 0.044
Home: 17C Gas oven: 31C Sampling event:
Electric oven: °-14
19C
3-46
-------
Table 3-5 (Continued): Ambient, outdoor, transport, indoor, and personal nitrogen dioxide measurements (ppb)
across studies.
Study
tKornartit et al.
(2010)
Ambient Outdoor
Sampling (Central (Outside Personal-Ambient
Location Time Period Interval Site) Residence) Transport Indoor Personal Slopea>b
Hertfordshire, Winter 2000 7-day NR NR NR Electric oven: Electric oven: NR
U.K. (Greater Bedroom: 7. 8° 8.1C
London Area) Living room:
(large city) 7.9C _
Kitchen: 7.1= Ga*™"'
Gas oven:
Bedroom:
10.8C
Living room:
13.7C
Kitchen: 20. 6C
Summer 2001 NR NR NR Electric oven: Electric oven: NR
Bedroom: 13.3C
12.7C
Living room: _
*% 1C Gas oven:
Kitchen: 11. Oc 146C
Gas oven:
Bedroom:
14.3C
Living room:
14.7C
Kitchen: 14.2C
3-47
-------
Table 3-5 (Continued): Ambient, outdoor, transport, indoor, and personal nitrogen dioxide measurements (ppb)
across studies.
Ambient Outdoor
Sampling (Central (Outside
Study Location Time Period Interval Site) Residence)
tLeeetal. (2013) Seoul, South Jul 2008 NR 29. 5f; 30. 7C NR
Korea (large
city)
Seoul, South
Korea (large
Clty) Jan 2009 NR 29.59; 31. 1C NR
Daegu, South Jul 2008 NR 19.99; 21. 1C NR
Korea (mid-
sized city)
Jan 2009 NR 23.09; 24.3C NR
Asan, South Jul 2008 NR 26.09; 27.9C NR
Korea (small
city)
Jan 2009 NR 21.69;23.9C NR
Personal-Ambient
Transport Indoor Personal Slopea>b
NR Home: 25.39; 27C
24.49; 25.7C
Work: 19.29;
21. 5C
NR Home: 22.59; 24.2C
20.99; 24.9C
Work: 27.99;
29.9C
NR Home: 21.49;22.6C
19. 39; 20. 3C
Work: 21. 39;
22.8C
NR Home: 20.39; 21. 7C
23.39; 25.1C
Work: 20.39;
22.9C
NR Home: 22.69; 24.3C
23.89; 24.9C
Work: 21. 19;
25.6C
NR Home: 19.99; 22.3C
20.39; 22.9C
Work: 13.09;
18.6C
NR
NR
NR
NR
NR
NR
3-48
-------
Table 3-5 (Continued): Ambient, outdoor, transport, indoor, and personal nitrogen dioxide measurements (ppb)
across studies.
Ambient Outdoor
Sampling (Central (Outside Personal-Ambient
Study Location Time Period Interval Site) Residence) Transport Indoor Personal Slopea>b
tLeeetal. (2013) Suncheon, Jul 2008 NR 15.09; 15. 9C
(Continued) South Korea
(rural)
Jan 2009 NR 12.59; 15.2C
Total Jul 2008 NR 21.79;23.7C
Jan 2009 NR 20.69; 23.6C
tDu etal. (2011) Beiiinq, China Oct2006 Varied with NR
transit times
NR NR Home: 14.09; 15.3=
13.09; 14. 30
Work: 12.09;
14.5C
NR NR Home: 12.99; 15.7=
15.99; 20.4C
Work: 9.39;
12.9C
NR NR Home: 20.59; 22.6C
19.59; 21.2C
Work: 18.49;
21. 4C
NR NR Home: 18.69; 21. Oc
19. 99; 23. 3C
Work: 16.49;
21. 1C
NR Subway: 20C; NR NR
Nonsubway:
22C;
Taxi drivers:
25C
NR
NR
NR
NR
NR
3-49
-------
Table 3-5 (Continued): Ambient, outdoor, transport, indoor, and personal nitrogen dioxide measurements (ppb)
across studies.
Study
tPhvsick et al.
(2011)
fSchembari et al.
(2013)
tMolloy et al.
(2012)
tPeqas et al.
(2012)
Location
Melbourne,
Australia
(large city)
Barcelona,
Spain (large
city)
Melbourne,
Australia
(large city)
Aveiro,
Portugal
(small city
center,
suburb)
Time Period
May 2006;
Jun 2006; Apr
2007; May
2007
Nov2008 and
Oct 2009
Aug
2008-Dec
2008; Jan
2009-Apr
2009
Apr-Jun 2010
Ambient Outdoor
Sampling (Central (Outside Personal-Ambient
Interval Site) Residence) Transport Indoor Personal Slopea>b
Ambient: 6:00 p.m.to NR NR Home: 17. 2e; Total: 12. 2h
1 h; Personal: 8:00 a.m.: 16. 8C Home' 8 2h
Participants 19.8e; 18.7' Work: 21.6- work- 14 7"
wore two sets 8: oo a.m. to 21. 7C T
of passive 6.0o p m • Transit: 23.4"
samplers. 20.3e;21.2c Other: 17.4h
One was
worn for 48 h.
One was
worn only
during the
hours spent at
home, at
work, in
transit, or
while
performing
other
activities.
7-day NR 18.79*1; NR 19.29*1; 20.6c<
7-day NR 9.5e; 10.0C NR 7.9e; 8.4C NR
7-day NR City center: NR City center: NR
10.5C<
NR
1.01k
0.9k
NR
3-50
-------
Table 3-5 (Continued): Ambient, outdoor, transport, indoor, and personal nitrogen dioxide measurements (ppb)
across studies.
Study
fChatzidiakou et
al. (2014)
fRivas et al.
(2014)
Sampling
Location Time Period Interval
Suburban Nov2011 5-day
London, U.K.
London, U.K.
Barcelona Jan-Jun 2012 4-day
and Sant
Cugat del
Valles, Spain
Ambient
(Central
Site)
NR
NR
NR
NR
NR
NR
22C; 20d
Outdoor
(Outside
Residence)
7.4
5.1
5.1
19
20
22
25C; 24d
Transport
NR
NR
NR
NR
NR
NR
NR
Indoor
3.7'
2.9'
2.7'
13'
16'
18'
16C; 16d
Personal
NR
NR
NR
NR
NR
NR
NR
Personal-Ambient
Slopea'b
NR
NR
NR
NR
NR
NR
NR
AER = air exchange rate; a.m. = ante meridiem; Aug = August; avg = average; CA = California; Dec = December; GA = Georgia; h = hour; Jan = January; MA = Massachusetts;
Ml = Michigan; min = minute; Nov = November; NR = not reported; Oct = October; ppb = parts per billion; SoCAB = South Coast Air Basin; TX =Texas; U.K. = United Kingdom.
aUnadjusted models only.
Total personal NO2 exposure vs. ambient concentration unless noted otherwise.
°Average.
dMedian.
ePersonal exposure to ambient NO2 vs. ambient concentration.
'Geometric mean.
9Data provided by the authors for Figure 1 of Phvsick et al. (2011).
hReported in |jg/m3 and converted to ppb assuming 25°C and 760 mm Hg.
'Averaged over 4 classrooms and 2 weeks.
'Indoor-outdoor ratio, rather than slope, is reported for Schembari et al. (2013).
Integrated measurement over 2 weeks.
'Estimated from reported indoor-outdoor ratio and outdoor NO2 concentration.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
3-51
-------
Table 3-6 Correlations between measured nitrogen dioxide concentrations
from personal, outdoor, indoor, and ambient monitors.
Study
tSarnatetal. (201 2)c
tWilliams et al. (2012a)c
tSuh and Zanobetti
(2010b'c
tBrown et al. (2009)
tDelfinoetal. (2008a)
Kousaetal. (2001)
tDelqado-Saborit(2012)
tLeeetal. (2013)
Location
Ciudad Juarez, Mexico; El
Paso, TX
Wayne County, Ml
Atlanta, GA
Boston, MA
2 southern California cities
Helsinki, Finland; Basel,
Switzerland; Prague,
Czech Republic
Birmingham, U.K.
Seoul, South Korea
Daegu, South Korea
Asan, South Korea
Suncheon, South Korea
All 4 cities
Personal-
Ambient3
NR
All Subjects:
0.11;
Vest-compliant
(>60%)e: 0.14
0.12
Winter: 0.00;
Summer: 0.03
0.43
NR
1-h NO2: 0.024;
Sampling event
NO2: 0.15
NR
NR
NR
NR
NR
Personal-
Outdoor13
NR
NR
NR
NR
NR
0.61
NR
Summer:
0.39;
Winter: 0.47
Summer:
0.43;
Winter: 0.47
Summer:
0.62;
Winter: 0.11
Summer:
0.46;
Winter: 0.56
Summer:
0.58;
Winter: 0.53
Personal-
Indoor
NR
NR
NR
NR
NR
0.73
NR
Summer:
0.50;
Winter: 0.55
Summer:
0.32;
Winter: 0.59
Summer:
0.63;
Winter: 0.37
Summer:
0.46;
Winter: 0.60
Summer:
0.60;
Winter: 0.55
Outdoorb-
Indoor
CJ-A: 0.36;
CJ-B: 0.92;
EP-A: 0.66;
EP-B: 0.01
NR
NR
NR
NR
0.66
NR
Summer:
0.71;
Winter: 0.22
Summer:
0.65;
Winter: 0.57
Summer:
0.67;
Winter: 0.37
Summer:
0.77;
Winter: 0.80
Summer:
0.78;
Winter: 0.55
3-52
-------
Table 3-6 (Continued): Correlations between measured nitrogen dioxide
concentrations from personal, outdoor, indoor, and
ambient monitors.
Study
tSahsuvaroqlu et al.
(2009)d
tSchembarietal. (2013)d
tVieiraetal. (201 2)c
tVan Roosbroeck et al.
(2008)°
Location
Lake Ontario, Canada
(winter)
Lake Ontario, Canada
(spring)
Lake Ontario, Canada
(summer)
Lake Ontario, Canada
(all seasons)
Barcelona, Spain
Sao Paolo, Brazil
the Netherlands
(3 schools)
Personal- Personal-
Ambient3 Outdoor13
NR All
Subjects:
0.002; Non-
ETS: 0.020
NR All
Subjects:
0.233; Non-
ETS: 0.187
NR All
Subjects:
0.067;
Non-ETS:
0.011
NR All
Subjects:
0.517;
Non-ETS:
0.540
NR 0.58
NR <0.35
NR 0.35
Personal-
Indoor
All Subjects:
0.430;
Non-ETS:
0.283
All Subjects:
0.589;
Non-ETS:
0.599
All Subjects:
0.822;
Non-ETS:
0.783
All Subjects:
0.729;
Non-ETS:
0.693
0.78
NR
NR
Outdoorb-
Indoor
NR
NR
NR
NR
0.53
All subjects:
0.13;
Non-ETS:
0.42
NR
CJ-A = Ciudad Juarez Site A; CJ-B = Ciudad Juarez Site B; EP-A = El Paso Site A; EP-B = El Paso Site B; ETS = Environmental
Tobacco Smoke; GA = Georgia; h = hour; MA = Massachusetts; Ml = Michigan; NO2 = nitrogen dioxide; NR = not reported;
TX = Texas; U.K. = United Kingdom.
aAmbient = central site monitor.
bOutdoor = outside residence.
°Spearman coefficient.
dPearson coefficient.
eSubjects wore the sampling vests at least 60% of the sampling period.
fStudies published since the 2008 ISA for Oxides of Nitrogen
Several studies have investigated factors that influence the relationship between
short-term personal exposure measurements and ambient concentrations. It was observed
that, even when the median or average total personal NO2 exposures and ambient
concentrations were comparable, the total personal exposure measurements and central
site monitor concentrations might not have always been correlated. For example,
Williams et al. (2012a) measured total personal NO2 exposures for the Detroit Exposure
3-53
-------
and Aerosol Research Study (DEARS) population of nonsmoking adults in 24-hour
intervals and found a low association (Spearman r = 0.14 for participants complying with
study protocols; r = 0.11 for all participants) between total personal NO2 exposure with
NCh concentrations measured at central site monitors. This result indicated the influence
of nonambient sources on the DEARS participants' total personal NCh exposures,
suggesting that total personal NO2 exposures and ambient NO2 concentrations are not
always well correlated. Likewise, (Suh and Zanobetti 201 Ob) measured correlation of
Spearman r = 0.12 between 24-hour total personal NCh exposure and central site NCh
concentration measurements among an Atlanta panel of 30 adults. Vieiraet al. (2012)
calculated Spearman correlations between 12-hour outdoor NCh concentration, indoor
NO2 concentration, and personal NO2 exposure measurements. All correlations between
personal and outdoor NO2 concentration measurements were below r = 0.35. Indoor and
outdoor NCh concentrations were more correlated (r = 0.42), although when smokers
were included, correlation between indoor and outdoor NCh concentration dropped
(r = 0.13). Van Roosbroeck et al. (2008) compared personal NC>2 exposure measurements
for children obtained over 1 to 4 weeks in a panel study with NCh concentration
measurements taken outside the children's schools, and they observed Pearson
correlation of r = 0.35. Outdoor school NCh concentrations underestimated personal NCh
exposures when used as a surrogate, but when additional variables representing indoor
exposures (such as exposure to gas cooking and unvented water heaters) were added to
the model, r increased to 0.77, suggesting that indoor sources played a large role in NCh
exposure among the study participants. Bellander et al. (2012) measured personal NCh
exposure using 7-day integrated diffusion samplers and modeled it as a function of NCh
concentrations measured at an urban area, rural area, roadside, and outside of the
participants' homes and places of work in Stockholm County, Sweden. They observed
slopes ranging from 0.25-0.37 (R2 = 0.01-0.20). Kousaet al. (2001) developed a
time-weighted microenvironmental model of NCh exposure based on time-activity data
and 48-hour microenvironmental NCh concentration measurements. The
microenvironmental model agreed well with personal exposure measurements (/? = 0.90;
R2 = 0.74).
Meng etal. (2012b) performed a random effects meta-analysis of 15 studies that
calculated slopes and correlations between personal NCh measurements of ET and central
site ambient NCh concentrations for 32 sample populations, of which 7 were from daily
average analyses, 8 were from longitudinal panel analyses, and 17 were from analyses
whose correlations were pooled over short time periods up to 1 week in length.
Metaregression results are shown in Table 3-7. Meng et al. (2012b) found that the
magnitude and correlation of associations between personal NCh exposure and ambient
NCh concentration depended on several factors, including study design (pooled data
across days, longitudinal panel, or daily average), season, meteorological conditions,
3-54
-------
ambient PIVb 5 concentration, and pre-existing cardiopulmonary disease of the exposure
subjects. Together, the low associations reported in these studies indicate that most of the
total personal NO2 exposure measurements for these studies were influenced either by
nonambient sources or by spatially variable NO2 concentrations not well detected by the
central site monitor. However, Meng et al. (2012b) also stated that the longitudinal panel
studies included in their meta-analysis had several measurements below detection limit
that could have erroneously reduced the correlations, which otherwise would be expected
to be higher.
Table 3-7 Metaregression results from 15 studies examining the relationship
between personal nitrogen dioxide exposure measurements and
ambient concentrations.
Study Design
Based on Original Studies
Slope
Correlation
Corrected for Publication Bias
Slope
Correlation
Pooled3
0.40
0.42
0.30
0.37
Longitudinal panelb
0.14
0.16
0.14
0.16
Temporal average0
0.29
0.72
0.20
0.45
h = hour.
aPooled analyses: Piechocki-Minguv et al. (2006). Linnet al. (1996). Liard et al. (1999). Gauvinet al. (2001). Aim et al. (1998).
Brown et al. (2009). Sarnat et al. (2006). Delfino et al. (2008a). Averaging period varies among the studies between 24 h and
13 weeks.
"Longitudinal analyses: Sarnat etal. (2005), Sarnat et al. (2001), Sarnat et al. (2000), Linakeret al. (2000), Kim et al. (2006a),
Koutrakis et al. (2005). 24-h measurements were made between 5-12 days during one or more seasons.
Temporal average analyses:. Averaging period varies among the studies. Mukala et al. (2000): 13 1-week periods; Liard et al.
(1999): 140 24-h periods; (Aim etal.. 1998): 6 1-week periods.
Source: Reprinted with permission from Elsevier; Meng et al. (2012b).
3.4.3 Factors Contributing to Error in Estimating Exposure to Ambient
Nitrogen Dioxide
Recent studies of factors influencing exposure error build from the existing literature
presented in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008cX which have
focused on time-activity patterns, spatial variability of ambient NO2 concentrations,
infiltration, nonambient exposures, and instrument accuracy and precision, as described
in the subsequent subsections. These factors can influence epidemiologic results for
studies of short-term and long-term NC>2 exposure, as detailed further in Section 3.4.5.
3-55
-------
3.4.3.1 Activity Patterns
Temporal Patterns
The complex human activity patterns across the population (all ages) are illustrated in
Figure 3-3 (Klepeis et al., 2001) for data from the National Human Activity Pattern
Survey (NHAPS). This figure is presented to illustrate the diversity of daily activities
among the entire population as well as the proportion of time spent in each
microenvironment. Time-activity data become an important source of uncertainty when
considering that ambient exposures vary in different microenvironments (e.g., transit,
residential), and that exposure assignment is typically based on the assumption that study
participants are in one location (residential) for the study duration.
ooooooooooooooooo
oooooooooooooooop
c-j ^H <>i rn -it-
-------
Different time-activity patterns have been found when analyzing data for different
populations or lifestages. For example, Wu etal. (2010) observed activity patterns for a
panel of adults and children from Camden, NJ communities with larger percentages of
nonwhites (85%) and those below the poverty line (33%) compared with NHAPS. The
study participants spent more time outdoors compared with the nationwide cohort
(3.8 hours vs. 1.8 hours nationally); note that Wu etal. (2010) undersampled participants
ages 65+ years, and the median age of the population studied in Wuetal. (2010) was
27 years compared with 35 years nationwide. Other recent time-activity panel studies
have included working adults (Isaacs etal.. 2013; Bellanderetal.. 2012; Kornartitet al..
2010). pregnant women (Iniguez et al.. 2009). adolescents (DeCastro et al.. 2007). and
children (Molter et al.. 2012; Xue etal.. 2004). In many cases, the time-activity data were
limited to residential, occupational, school, and outdoor location categories to simplify
assignment of concentrations to which the subjects were estimated to be exposed in each
microenvironment. The implication of these findings is, given that time-activity data vary
among different populations, the one-location assumption used in many studies varies in
accuracy among those different populations. However, because few studies are as large as
NHAPS, it would be premature to make conclusions about time-activity data for smaller
cohorts.
Time spent in different locations has also been found to vary by age. Table 3-8
summarizes NHAPS data reported for four age groups, termed Very Young (0-4 years),
School Age (5-17 years), Working (18-64 years), and Retired (65+ years) (Klepeis et al..
1996). The working population spent the least amount of time outdoors, while the school
age population spent the most time outdoors. NHAPS respondents aged 65 and over spent
somewhat more time outdoors than adults aged 18-64, with a greater fraction of time
spent outdoors at a residence. Children aged 0-4 also spent most of their outdoor time in
a residential outdoor location. On average, the fraction of time spent outdoors by school
age respondents was 2.62 percentage points higher than working respondents,
corresponding to approximately 38 minutes more time outdoors per day. Moreover, in a
comparison of children (mostly less than age 8 years), adults mostly under age 55 years,
and adults older than age 55 years, a larger proportion of children reported spending over
30 minutes performing vigorous outdoor physical activity (Wu etal.. 201 Ib). Increased
time spent outdoors or more time outdoors performing vigorous physical activity not only
could have implications for differential exposure error in these age groups but also could
influence NO2 exposure of children and older adults and their risk of NO2-related health
effects (Section 7.5.1).
3-57
-------
Table 3-8 Mean fraction of time spent in outdoor locations by various age
groups in the National Human Activity Pattern Survey study.
Age Group
0-4 yr
5-1 7 yr
18-64yr
65+ yr
Residential-Outdoor
5.38%
5.05%
2.93%
4.48%
Other Outdoor
0.96%
2.83%
2.33%
1 .27%
Total Outdoors
6.34%
7.88%
5.26%
5.75%
yr = year.
Source: Data from Klepeiset al. (1996).
Recently, Kornartit et al. (2010) tested the associations between time-weighted exposure
estimates from area samples with personal sampling measurements for a London, U.K.
panel study. Kornartit et al. (2010) measured NO2 concentration for 1 week with passive
Palmes tube samplers in several outdoor and indoor microenvironments for 55 subjects
aged 21-60 years and correlated a time-weighted average of those microenvironmental
NO2 concentration measurements with personal NC>2 exposure measurements, also
measured with Palmes tubes. They observed a slope of 0.94 for the relationship between
time-weighted average microenvironmental NC>2 concentrations and personal NC>2
exposures (R2 = 0.85) in winter and a slope of 0.59 (R2 = 0.65) in summer. Higher levels
of NO2 were observed for both time-weighted average concentrations and personal
exposures in summer compared with winter. However, correlations between personal
NC>2 exposure and time-weighted microenvironmental NC>2 concentrations were higher in
winter, implying panel studies using personal NCh exposure measurements may be more
dominated by indoor sources during cold-weather months. The authors concluded that the
time-weighting approach provided a reasonable approximation of personal exposure but
sometimes underestimated it.
Exertion Levels
Together with location, exertion level is an important determinant of exposure. Table 3-9
summarizes ventilation rates for different age groups at several levels of activity as
presented in Table 6-2 of the U.S. EPA's Exposure Factors Handbook (U.S. EPA.
201 Ib). Most of the age-related variability is seen for moderate and high intensity
activities, except for individuals under 1 year. For moderate intensity, ventilation rate
increases with age through childhood and adulthood until age 61, after which a moderate
3-58
-------
decrease is observed. Ventilation rate is most variable for high intensity activities.
Children aged 1 to <11 years have ventilation rates of approximately 40 L/minute, while
children aged 11+ years and adults have ventilation rates of approximately 50 L/minute.
The peak is observed for the 51 to <61 year age group, at 53 L/minute, with lower
ventilation rates for older adults. The role of physical activity as a modifier of health
effect estimates is discussed in Section 7.6.3.
Table 3-9
Age Group
Birth to <1
1 to<2
2to<3
3to<6
6 to <1 1
11 to<16
16to<21
21 to <31
31 to <41
41 to <51
51 to <61
61 to <71
71 to <81
81 +
Mean ventilation rates (L/min)
age groups.
(yr) Sleep or Nap
3.0
4.5
4.6
4.3
4.5
5.0
4.9
4.3
4.6
5.0
5.2
5.2
5.3
5.2
Sedentary/Passive
3.1
4.7
4.8
4.5
4.8
5.4
5.3
4.2
4.3
4.8
5.0
4.9
5.0
4.9
at different
activity levels
Moderate
Light Intensity Intensity
7.6
12
12
11
11
13
12
12
12
13
13
12
12
12
14
21
21
21
22
25
26
26
27
28
29
26
25
25
for different
High Intensity
26
38
39
37
42
49
49
50
49
52
53
47
47
48
yr = year.
Source: Data from Exposure Factors Handbook (U.S. EPA, 201 1b)
A dramatic increase in ventilation rate occurs as exercise intensity increases. For children
and adults <31 years, high intensity activities result in nearly double the ventilation rate
for moderate activity, which itself is nearly double the rate for light activity. Children
have other important differences in ventilation compared to adults. As discussed in
Chapter 5. children tend to have a greater oral breathing contribution than adults, and
3-59
-------
they breathe at higher minute ventilations relative to their lung volumes. Both of these
factors tend to increase dose normalized to lung surface area.
Longitudinal activity pattern information is also an important determinant of exposure, as
different people may exhibit different patterns of time spent outdoors over time due to
age, gender, employment, and lifestyle-dependent factors. These differences may
manifest as higher mean exposures or more frequent high-exposure episodes for some
individuals. The extent to which longitudinal variability in individuals contributes to the
population variability in activity and location can be quantified by the ratio of
between-person variance to total variance in time spent in different locations and
activities (the intraclass correlation coefficient, ICC). Xue et al. (2004) quantified ICC
values in time-activity data collected by Harvard University for 160 children aged
7-12 years in southern California (Gevhetal.. 2000). For time spent outdoors, the ICC
was approximately 0.15, indicating that 15% of the variance in outdoor time was due to
between-person differences. The ICC value might be different for other population
groups.
The U.S. EPA's National Exposure Research Laboratory (NERL) has consolidated many
of the most important human activity databases into one comprehensive database called
the Consolidated Human Activity Database (CHAD). The current version of CHAD
contains data from 19 human activity pattern studies (including NHAPS), which were
conducted between 1982 and 1998 and evaluated to obtain over 33,000 person-days of
24-hour human activities in CHAD (McCurdy et al.. 2000). The surveys include
probability-based recall studies conducted by the U.S. EPA and the California Air
Resources Board, as well as real-time diary studies conducted in individual U.S.
metropolitan areas using both probability-based and volunteer subject panels. All ages of
both genders are represented in CHAD. The data for each subject consist of one or more
days of sequential activities, in which each activity is defined by start time, duration,
activity type, and microenvironmental classification (i.e., location). Activities vary from
one minute to one hour in duration, with longer activities being subdivided into
clock-hour durations to facilitate exposure modeling. CHAD also provides information
on the level of exertion associated with each activity, which can be used by exposure
models, including the APEX model, to estimate ventilation rate and pollutant dose.
3.4.3.2 Spatial Variability in Nitrogen Dioxide Concentrations
Data for spatial variability in ambient NO, NO2, and NOx concentrations are provided in
Section 2.5 for national, urban, neighborhood, and micro scales. The data illustrate that
3-60
-------
national variation in wintertime concentrations largely follows the degree of urbanization,
while variation at urban and smaller scales is influenced by source location, source
strength, meteorology, and natural and urban topography. Gradients in near-road
concentrations of NC>2 and NO indicate spatial variability at finer scales within 500 m of
the road (see Figures 2-16 and 2-17 in Section 2.5.3 and Figure 3-2 in Section 3.3.1.1).
Figure 3-4 illustrates regional-scale variability in background levels of daily 1-hour max
NO2 concentration based on Pearson correlation between monitor pairs for urban and
rural monitors across the U.K. (Butland et al.. 2013). Likewise, Figure 3-5 depicts
urban-scale variability for NC>2 and NOx, based on a semivariogram function (Goldman
etal..201Q).
The correlation between the true exposure and the measured NO2 concentration will
decrease with increasing distance from the monitor. Moreover, the magnitude of the error
in exposure estimation may increase with distance between the monitor and the subject.
This is an issue for both central site monitors and fixed site passive monitors (Table 3-1
and Section 3.2.1). Hence, there is a potential for exposure error if the ambient NO2
concentration measured at a given site differs from the concentration at the location of an
epidemiologic study participant, and this issue is present regardless of the spatial scale of
the epidemiology study. Similarly, when a spatial model (including LUR, IDW, and
CTM) is not sufficiently finely resolved, then the estimated concentration assigned as a
participant's exposure may have additional error (Table 3-1 and Section 3.2.2).
C CO
O '
o
o
o
<2
(0
o>
Urban background loge nitrogen dioxide
Fitted line
P = 0 71027 - 000073 x D
R-sq=0 71
200 400 600 800 1000
Distance in km (D)
Rural loge nitrogen dioxide
C 00
O -
O
O Tj-
o
£2
-------
0
NO2
NOx
0 50
Distance (km)
50 100
Distance (km)
Note: km = kilometer; NO2 = nitrogen dioxide; NOX = the sum of nitric oxide and nitrogen dioxide.; y' = semivariogram.
On the y-axis, y' denotes the semivariogram (i.e., a unitless function that describes the ratio between spatial and temporal variance
of the differences between two observations).
Source: Reprinted with the permission of American Chemical Society, Goldman et al. (2010).
Figure 3-5 Urban-scale variability in nitrogen dioxide and the sum of nitric
oxide and nitrogen dioxide in Atlanta, GA.
3.4.3.3 Infiltration and Building Ventilation
Given that people spend the majority of their time indoors, building air exchange rates
influence exposure to ambient NC>2. In an analysis of daily average NO2 concentration
and exposure data from the DEARS, Meng et al. (2012a) observed seasonal differences,
with slopes of 0.24 ± 0.04 for ET versus the concentration measured at a central site
monitor, Ca,csm, and of 0.13 ± 0.06 for Ea versus Ca,csm for summer measurements. For
winter measurements, the associations were lower (ET vs. Ca,csm: slope = 0.08 ± 0.05;
Ea vs. Ca,csm: slope = 0.07 ± 0.07). Meng etal. (2012a) found that high air exchange rate
(>1.3 air changes per hour), no central air conditioning, use and nonuse of window fans,
and presence of old carpeting were determinants of a, the exposure factor defined in
Equation 3-10 and approximated by the ratio of Ea to Ca, for NC>2 in summer; none of
these factors were determinants of a for NO2 in winter. In Molter et al. (2012). outdoor
exposures were calculated with LUR, while indoor exposures were calculated using the
probabilistic model for indoor pollution exposures (INDAIR) model that accounts both
for infiltration due to home ventilation characteristics and indoor sources. Sensitivity to
air exchange rate of INDAIR predictions of indoor NC>2 in the absence of indoor sources
3-62
-------
underscores potential for bias and uncertainty in a, which depends on air exchange rate,
penetration, and indoor deposition (Dimitroulopoulou et al.. 2006).
3.4.3.4 Instrument Accuracy and Precision
The influence of instrument error (Section 2.4.1. Section 3.2.1.2) on health effect
estimates from epidemiologic studies varies with study design. Intermonitor comparison
is often used to estimate instrument precision.
For epidemiologic studies of short-term exposure, Goldman et al. (2010) investigated
instrument precision error at locations where ambient monitors were collocated.
Instrument precision error increased with increasing concentration. If instrument error
and concentration are positively correlated, then error in the exposure estimates will be
larger in locations where there are more prevalent or stronger sources or at times when
NC>2 emissions are higher for a given location. Moreover, if error is positively correlated
with concentration, then it would be anticipated that the magnitude of the instrument
error is largest at times of day when emissions are highest, such as rush hour. Depending
on specific conditions such as sampler type (e.g., passive vs. continuous), meteorological
conditions, or presence of interferants, instrument errors may vary in total magnitude or
direction (Section 3.2.1) so that error is not always positively correlated with
concentration. Instrument error was also observed to exhibit some autocorrelation at
1- and 2-day lags in the Goldman et al. (2010) study. Hence, the diurnal variability in
relative NO2 instrument error does not change substantially from day to day. For
epidemiologic studies of short-term NCh exposure, the influence of instrument error
would not be expected to change if the health data were obtained on a daily basis.
Instrumentation bias could be anticipated to influence exposure estimates used in
long-term NC>2 exposure studies in some situations. For example, LUR exposure may be
overestimated when the LUR is fit using passive monitoring data, if the passive monitors
are positively biased (Section 3.2.1.2). Ambient temperature and relative humidity would
not be expected to vary greatly within a city. Because climate and ambient sources are
more likely to differ among cities, instrumentation error could have a larger influence on
the comparison of exposures among cities.
3.4.4 Confounding
To assess the independent effects of NO2 in an epidemiologic study of health effects, it is
necessary to identify (Bateson et al.. 2007): (1) which copollutants (e.g., PlVfc 5, UFP, BC)
and additional exposures (e.g., noise, traffic levels) are potential confounders of the
3-63
-------
health effect-NCh relationship so that their correlation with NO2 can be tested and, if
needed, they are accounted for in the epidemiologic model; (2) the time period over
which correlations might exist so that potential confounders are considered appropriately
for the time period relevant for the epidemiologic study design (e.g., pollutants or other
factors that are correlated over the long term might not be important for a short-term
exposure epidemiologic study); and (3) the spatial correlation structure across multiple
pollutants, if the epidemiologic study design is for long-term exposure. Given that a
covariate must be correlated with both the exposure and the health effect to be a
confounder, the potential for confounding of NC>2-related health effects can vary by the
health endpoint of interest.
For monitors that do show high correlations, copollutant epidemiologic models may be
appropriate to adjust the effect estimate for each pollutant for confounding by the other
pollutant (Tolbert et al.. 2007). As discussed in the 2010 ISA for Carbon Monoxide (U.S.
EPA. 201 Ob), copollutant models can help identify which is the better predictor of the
effect, particularly if the etiologically linked pollutant is measured with more error than
the other pollutant. Because NO2 exhibits a relatively high degree of exposure error
compared with other criteria pollutants (Section 3.4.3). copollutant models in which the
NO2 effect estimate remains robust provide additional evidence for an independent health
effect of NO2.
This section considers temporal copollutant correlations and how relationships among
copollutants may change in space. Temporal copollutant correlations are computed from
the time series of concentrations for two different collocated pollutants. Temporal
correlations are informative for epidemiologic studies of short-term NCh exposure when
the sampling interval is a month or less for each of the copollutants. Temporal
correlations are informative for epidemiologic studies of long-term NC>2 exposures when
sampling intervals are months to years. Spatial relationships are evaluated by comparing
within-pollutant variation across space for different pollutants. The following sections
review co-exposures that can potentially confound the relationship between a health
effect and NO2 exposure over different temporal and spatial resolutions.
3.4.4.1 Temporal Relationships among Ambient Nitrogen Dioxide and Copollutant
Exposures
Studies and analyses reported in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c)
demonstrated that ambient NCh concentration was correlated with several traffic-related
pollutants in urban and suburban areas generally in the range of Pearson r = 0.5 to 0.8 for
PM2.5 and CO and r = 0.8 to 0.9 for EC. These results suggest that in some cases NO2
concentration can be a surrogate for traffic pollution (U.S. EPA. 2008c). In contrast,
3-64
-------
correlations between NO2 and Os concentrations ranged from r = -0.7 to 0.1. Numerous
air quality, exposure, and epidemiologic studies have more recently evaluated
associations between concentrations of ambient NO2 and those of other pollutants. Many
of these studies report Pearson or Spearman correlations of ambient NO2 concentrations
with those of other criteria pollutants, mainly focusing on concentrations related to traffic
sources (PM2 5, CO, PMio). A few studies have explored associations between NO2
concentrations and those of other traffic-related pollutants, such as EC, UFP, and VOCs.
Data for correlations between NO2 concentrations and concentrations of other criteria
pollutants are summarized in Table 3-10. broken into short- and long-term exposure
studies. Figure 3-6 plots data for correlations between NO2 concentrations and
concentrations of all copollutants for which data were available, including PM2 5, PMio
PMio-2.5, Os, CO, SO2, EC, OC, UFP, particle number concentration (PNC), toluene, and
benzene. Figure 3-6 separates the data by averaging period. "Within-hourly" denotes
averaging time ranging from 20 seconds to 1-hour daily max. "Within daily" is noted for
averaging time ranging from 3 to 24 hours. Three-hour averaging times are typically
applied during rush hour measurement periods. "Within monthly" refers to averaging
times ranging from 84 hours to 1 month. "Annual or longer-term correlations" are for
studies that averaged the data over a period of 1 to 5 years. The studies presented in
Table 3-10 only include monitored data and not correlations computed from LUR studies.
Some of these studies used personal or area sampling in lieu of central site monitoring.
Note that, while Table 3-10 and Figure 3-6 are informative for considering the influence
of averaging time on correlations, small sample sizes for any given pollutant and
averaging period preclude making definitive conclusions about the observations. In
particular, the number of near-road studies reporting correlations between NO2
concentrations and concentrations of copollutants was too small to make any conclusions
about differences in NO2-copollutant correlations between near-road and central site or
personal measures.
The higher the copollutant correlation, the more difficult it is to disentangle the health
effects of NO2 exposure from those of the copollutants. This is particularly true of
traffic-related copollutants, and recent evidence indicates that copollutant confounding
adds such uncertainty. Figure 3-6 shows the range of temporal NO2-copollutant
correlation coefficients among the studies in Table 3-10 plus one additional measurement
study that did not include other criteria air pollutants (Williams et al., 2012a). Existing
studies indicate that NO2 concentration has, in general, correlations over Pearson
r = 0.5 with concentrations of other NAAQS and traffic-related pollutants. Similar to
findings in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). the strongest
temporal correlations are typically observed for NO2 concentrations with concentrations
of primary traffic-related pollutants, such as benzene, CO, EC, and PNC. A wide range of
temporal correlations is observed for NO2 concentrations with PM2 5, PMio, and SO2
3-65
-------
concentrations. Correlations of NC>2 concentrations with PIVb 5 and PMio concentrations
tend to be positive for the within-hourly, within-daily, and long-term metrics. For the
within-monthly measures, median correlations are closer to zero. The reason for this
difference is unknown, but fewer data are available for the within-monthly correlations.
The lowest temporal correlations are typically observed for NO2 concentrations with Os
and PMio-2.5 concentrations, with correlations having a wide range in magnitude
(r = -0.71 to 0.66; median r = 0.15). These observations are not surprising given the
nonlinear relationship between NC>2 concentration and instantaneous Os production rate
observed close to the location of emission (Pusede and Cohen. 2012; LaFranchi et al.,
2011; Murphy et al.. 2007. 2006). Temporal correlations for near-road studies are
highlighted in red for Figure 3-6. It is notable that the near-road correlations did not
appear to be systematically different from the urban scale correlations. Statistical testing
for near-road versus urban scale interpollutant correlations was not performed given the
small number of near-road studies.
Short-Term Temporal Correlations
For the shorter time periods (within hourly and within daily), UFP, BC, CO, and EC
concentrations tended to have higher correlations with NC>2 concentration, while Os
concentration had several negative correlations with NO2 concentration. The within-daily
category had the most data for PM2 5 and PMio concentration, and a wide range of
correlations was observed with NO2 concentrations for each of those copollutants. Fewer
data were available for within-monthly correlations. Black carbon, benzene, and toluene
concentrations were observed to have the highest correlations with NO2 concentration in
this temporal category. Across time-averaging periods, there is not a discernible pattern
with respect to correlations of near-road measurements.
3-66
-------
Table 3-10 Synthesis of nitrogen dioxide ambient-ambient copollutant correlations from measurements
reported in the literature.
Study3 Averaging Time
Location Scale
Correlation
Measure
CO O3 SO2 PM2.5 PMio
Short-term exposure studies
tPolidori and Fine 1 min
(2012a)
tLevy et al. (2014) <2 min
Los Angeles, CA (15 m Near road
downwind of 1-710) summer
Los Angeles, CA (80 m Near road
downwind of 1-710) summer
Los Angeles, CA (background) Urban
summer
Los Angeles, CA (15 m Near road
downwind of 1-710) winter
Los Angeles, CA (80 m Near road
downwind of 1-710) winter
Los Angeles, CA (background) Urban
winter
Montreal, Canada (all year) Urban
Montreal, Canada (summer) Urban
Montreal, Canada (fall) Urban
Montreal, Canada (winter) Urban
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
0.65 NR NR NR NR
0.65 NR NR NR NR
0.66 NR NR NR NR
0.60 NR NR NR NR
0.62 NR NR NR NR
0.79 NR NR NR NR
OAQ n A^I n -I -I n OQ n ^Q
0.77 -0.74 0.17 0.34 0.35
0.40 -0.33 0.25 0.26 0.30
0.16 -0.36 0.04 0.34 0.35
3-67
-------
Table 3-10 (Continued): Synthesis of nitrogen dioxide ambient-ambient correlations from measurements reported
in the literature.
Study3
fPadro-Martinez et
al. (2012)
tChuanq et al.
(2008)
fStrickland et al.
(2010)
fVilleneuve et al.
(2007)
fJalaludin et al.
(2007)
Mortimer et al.
(2002)
Burnett et al. (2000)
Mar et al. (2000)
Tolbert et al. (2007)
Averaging Time
2 min
Hourly
1-h daily max
1-h daily max
1-h daily max
1-h daily max
1-h daily max
1-h daily max
1-h daily max
Location
Boston, MA
Boston, MA
Atlanta, GA (cold season)
Atlanta, GA (warm season)
Edmonton, Canada
Sydney, Australia
8 U.S. cities
8 Canadian cities
Phoenix, AZ
Atlanta, GA
Scale
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Correlation
Measure
Spearman
Pearson
Spearman
Spearman
Pearson
NR
NR
NR
NR
Spearman
CO
0.51
NR
0.59
0.54
0.74
0.6
NR
0.65
0.87
0.7
03
NR
NR
0.11
0.42
NR
0.25
0.27
0.12
NR
0.44
SO2
NR
NR
0.36
0.37
NR
0.46
NR
0.49
0.57
0.36
PM2.5
0.21
0.38
0.37
0.36
NR
0.65
NR
0.53
0.77
0.47
PMio
NR
0.33
0.46
0.44
NR
0.48
NR
0.53
0.53
0.53
3-68
-------
Table 3-10 (Continued): Synthesis of nitrogen dioxide ambient-ambient correlations from measurements reported
in the literature.
Study3
fDarrow et al.
(2011 a)
Moshammer et al.
(2006)
fDarrow et al.
(2011 a)
Averaging Time Location
1-h daily max Atlanta, GA
Morning Atlanta, GA
commute (7:00
a.m.-10:00 a.m.)
Daytime Atlanta, GA
(8:00a.m.-7:00
p.m.)
Nighttime Atlanta, GA
(12:00a.m.-6:00
a.m.)
8-h avg Linz, Austria
24-h avg Atlanta, GA
Correlation
Scale Measure
Urban Partial
Spearman
Urban Partial
Spearman
Urban Partial
Spearman
Urban Partial
Spearman
Urban Pearson
Urban Partial
Spearman
CO O3 SO2 PM2.5 PMio
0.61 0.40 NR 0.50 NR
0.57 -0.16 NR 0.46 NR
0.53 -0.07 NR 0.41 NR
0.66 -0.66 NR 0.52 NR
NR NR NR 0.54 0.62
0.66 -0.15 NR 0.20 NR
3-69
-------
Table 3-10 (Continued): Synthesis of nitrogen dioxide ambient-ambient correlations from measurements reported
in the literature.
Study3 Averaging Time
fFaustini et al. 24-h avg
(2011)
tSamolietal. (2011) 24-h avg
Ko et al. (2007a) 24-h avg
tMehta et al. (2013) 24-h avq
Location
Milan, Italy
Mestre, Italy
Turin, Italy
Bologna, Italy
Florence, Italy
Pisa, Italy
Rome, Italy
Cagliari, Italy
Taranto, Italy
Palermo, Italy
Athens, Greece
Hong Kong, China
Ho Chi Minh City, Vietnam (dry
season)
Ho Chi Minh City, Vietnam (wet
season)
Scale
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Correlation
Measure
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
NR
Pearson
NR
NR
CO
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
03
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
0.34
0.44
0.17
SO2
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
0.55
0.66
0.29
0.01
PM2.5
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
0.44
NR
NR
PMio
0.79
0.66
0.72
0.66
0.65
0.57
0.5
0.23
0.19
0.22
NR
0.4
0.78
0.18
3-70
-------
Table 3-10 (Continued): Synthesis of nitrogen dioxide ambient-ambient correlations from measurements reported
in the literature.
Study3 Averaging Time
fAndersen et al. 24-h avg
(2008a)
Mannes et al. (2005) 24-h avg
Schildcrout et al. 24-h avq
(200S) 24-h avg
tLiu et al. (2009b) 24-h avq
tStraketal. (201 3a) 24-h avg
fO'Connor et al. 24-h avg
(2008)
Timonen et al. 24-h avg
(2006)
Location
Copenhagen, Denmark
Sydney, Australia
Albuquerque, NM
Baltimore, MD
Boston, MA
Denver, CO
San Diego, CA
St. Louis, MO
Toronto, Canada
Ontario, Canada
Locations across the
Netherlands
Inner-cities across the U.S.
Amsterdam, the Netherlands
Erfurt, Germany
Helsinki, Finland
Scale
Near road
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Correlation
Measure
Spearman
Pearson
NR
NR
NR
NR
NR
NR
NR
Spearman
Spearman
NR
Spearman
Spearman
Spearman
CO
NR
0.57
0.76
0.69
0.8
0.85
0.92
0.71
0.63
NR
NR
0.54
0.76
0.86
0.32
03
-0.58
0.29
0.04
0.44
0.47
0.24
0.39
0.42
0.4
-0.51
ORO
-0.31
NR
NR
NR
SO2
NR
NR
NR
0.49
0.68
0.56
0.23
0.58
0.63
0.18
NR
0.59
NR
NR
NR
PM2.5
0.41
0.66
NR
NR
NR
NR
NR
NR
NR
0.71
0.45
0.59
0.49
0.82
0.35
PMio
0.43
0.47
0.26
0.62
0.48
0.64
0.55
0.45
0.64
NR
0.49
NR
NR
NR
NR
3-71
-------
Table 3-10 (Continued): Synthesis of nitrogen dioxide ambient-ambient correlations from measurements reported
in the literature.
Study3
tGuo et al. (2009)
Roias-Martinez et al.
(2007a)
Sarnatetal. (2001)
Sarnat et al. (2005)
Kim et al. (2006a)
Roberts and Martin
(2006)
Andersen et al.
(2007)
tChen et al. (2008)
tArhami et al. (2009)
Averaging Time
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
Location
Beijing, China
Mexico City, Mexico
Baltimore, MD (summer)
Baltimore, MD (winter)
Boston, MA (summer)
Boston, MA (winter)
Toronto, Canada
Cleveland, OH
Nashville, TN
Copenhagen, Denmark
Shanghai, China
San Gabriel Valley, CA
(summer/fall)
San Gabriel Valley, CA
(fall/winter)
Riverside, CA (summer/fall)
Riverside, CA (fall/winter)
Scale
Urban
Urban
Urban
Urban
Near road
Near road
Near road
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Correlation
Measure
Pearson
Pearson
Spearman
Spearman
Spearman
Spearman
Spearman
NR-pairwise
NR-pairwise
Spearman
NR
Spearman
Spearman
Spearman
Spearman
CO
NR
NR
0.75
0.76
NR
NR
0.72
0.67
0.36
0.74
NR
NR
NR
NR
NR
03
NR
0.17
0.02
-0.71
NR
NR
NR
0.36
0.26
NR
NR
NR
NR
NR
NR
SO2
0.53
NR
NR
-0.17
NR
NR
NR
0.56
0.08
NR
0.73
NR
NR
NR
NR
PM2.5
0.67
NR
0.37
0.75
0.44
0.64
0.44
NR
NR
NR
NR
0.1
0.44
0.07
0.56
PMio
NR
0.25
NR
NR
NR
NR
NR
0.63
0.44
0.42
0.71
0.31
0.34
0.21
0.64
3-72
-------
Table 3-10 (Continued): Synthesis of nitrogen dioxide ambient-ambient correlations from measurements reported
in the literature.
Study3 Averaging Time
tPelfinoetal. (2009) 24-h avg
tBaxteretal. (2013) 24-h avg
fWilliams et al. 24-h avg
(2012c)
fWilliams et al. 24-h avg
(2012a)
fDelfino et al. 24-h avg
(2008a)
tSuh and Zanobetti 24-h avq
(2010b)
tSchembari et al. 24-h avq
(2013)
Location
San Gabriel Valley and
Riverside, CA (aggregated)
Boston, MA
Pittsburgh, PA
Memphis, TN
Detroit, Ml
Milwaukee, Wl
San Diego, CA
Riverside, CA
Research Triangle Park, NC
Detroit, Ml
Los Angeles, CA
Atlanta, GA
Barcelona, Spain
Scale
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Near road
Urban
Urban
Urban
Correlation
Measure
NR
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
CO
0.79
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
03
0/1O
NR
NR
NR
NR
NR
NR
NR
-0.12
NR
NR
NR
NR
SO2
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
PM2.5
0.19
0.41
0.46
0.27
0.59
0.55
0.57
0.37
0.03
NR
0.36
0.47
0.41
PMio
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
3-73
-------
Table 3-10 (Continued): Synthesis of nitrogen dioxide ambient-ambient correlations from measurements reported
in the literature.
Study3 Averaging Time
fLaurent et al. 24-h avg
(2013)
tPeters et al. (2009) 24-h avg
tSanchez Jimenez 24-h avq
etal. (2012)
fSteinvil et al. 24-h avg
(2009)
fSteinvil et al. 24-h avg
(2008)
tTaoetal. (2012) 24-h avg
tWichmann et al. 24-h avq
(2012)
Location
Los Angeles and Orange
counties, CA
Erfurt, Germany
Glasgow, U.K.
Glasgow, U.K.
Glasgow, U.K.
London, U.K.
London, U.K.
Tel Aviv, Israel
Tel Aviv, Israel
Guangzhou, Foshan,
Zhongshan, and Zhuhai, China
Copenhagen, Denmark (warm
period)
Copenhagen, Denmark (cold
period)
Scale
Urban
Urban
Near road
Background
Background
Near road
Background
Urban
Urban
Urban-
regional
Urban
Urban
Correlation
Measure
Pearson
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Partial Pearson
Partial Pearson
Pearson
Spearman
Spearman
CO
0.83
0.68
0.6
0.4
0.74
0.3
0.61
0.75
0.86
0.72
0.62
0.72
03
-0.81
Occ
NR
NR
NR
NR
NR
-0.34
-0.78
0.17
NR
NR
SO2
NR
0.54
NR
NR
NR
NR
NR
0.70
0.72
0.82
NR
NR
PM2.5
0.77
0.63
NR
NR
NR
0.49
0.42
NR
NR
NR
NR
NR
PMio
0.70
0.64
0.83
0.69
NR
0.67
0.37
0.076
0.082
0.82
0.47
0.46
3-74
-------
Table 3-10 (Continued): Synthesis of nitrogen dioxide ambient-ambient correlations from measurements reported
in the literature.
Study3 Averaging Time Location
tDimitriou and 24-h avg London, U.K. (cold period)
Kassomenos (2014)
London, U.K. (warm period)
Paris, France (cold period)
Paris, France (warm period)
Copenhagen, Denmark (cold
period)
Copenhagen, Denmark (warm
period)
Scale
Urban
Near road
Urban
Near road
Urban
Near road
Urban
Near road
Urban
Near road
Urban
Near road
Correlation
Measure
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
CO
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
03
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
SO2
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
PM2.5
0.52
0.49
0.63
0.60
0.65
0.60
0.54
0.75
0.31
0.36
0.42
0.53
PMio
0.49
0.70
0.56
0.67
0.71
0.68
0.50
0.83
0.35
0.37
0.42
0.55
3-75
-------
Table 3-10 (Continued): Synthesis of nitrogen dioxide ambient-ambient correlations from measurements reported
in the literature.
Study3 Averaging Time
tDimitriou and 24-h avg
Kassomenos (2014)
(Continued)
(Continued)
tClouqhertv et al. 84-h avg
(2013)
tSarnatet al. (2012) 96-h avg
tGreenwald et al. 96-h avg
Location
Hamburg, Germany (cold
period)
Hamburg, Germany (warm
period)
Stockholm, Sweden (cold
period)
Stockholm, Sweden (warm
period)
New York, NY
El Paso, TX (Site A)
El Paso, TX (Site B)
Ciudad Juarez, Mexico (Site A)
Ciudad Juarez, Mexico (Site B)
2 sites in El Paso, TX
Scale
Urban
Near road
Urban
Near road
Urban
Near road
Urban
Near road
Urban
Urban
Near road
Urban
Near road
Urban
Correlation
Measure
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Pearson
Spearman
Spearman
Spearman
Spearman
Pearson
CO
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
03
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
SO2
NR
NR
NR
NR
NR
NR
NR
NR
0.51
NR
NR
NR
NR
-0.22
PM2.5
0.21
0.40
0.50
0.69
0.20
0.49
0.38
0.58
0.74
-0.39
-0.28
-0.28
0
0.2
PMio
0.23
0.52
0.51
0.70
0.24
0.45
0.45
0.52
NR
-0.3
-0.1
-0.1
0.11
0.31
3-76
-------
Table 3-10 (Continued): Synthesis of nitrogen dioxide ambient-ambient correlations from measurements reported
in the literature.
Study3
fWheeler et al.
(2008)
fTrasande et al.
(2013)
Long-term exposure
tDadvand et al.
(2014c)
fKatanoda et al.
(2011)
tDonqetal. (2011)
tHwanq and Lee
(2010)
tHeinrich et al.
(2013)
fDucret-Stich et al.
(2013)
Averaging Time
2 week
1 -month avg
studies
9 month
1 -yr avg
1 -yr avg
1 -yr avg
1 -yr avg
1 -yr avg
Location
Windsor, Canada (all year)
Windsor, Canada (winter)
Windsor, Canada (spring)
Windsor, Canada (summer)
Windsor, Canada (fall)
United States
Barcelona, Spain
Japanese cities
7 cities across China
14 Taiwanese communities
North Rhine-Westphalia,
Germany
Swiss Alps
Scale
Urban
Urban
Urban
Urban
Urban
Varies
Urban
Urban
Urban
Urban
Urban
Near road
On highway
Correlation
Measure
Spearman
Spearman
Spearman
Spearman
Spearman
Pearson
Spearman
Pearson
NR
NR
Spearman
Spearman
Spearman
CO
NR
NR
NR
NR
NR
0.12
NR
NR
0.23
0.86
NR
NR
NR
03
NR
NR
NR
NR
NR
-0.023
NR
NR
0.66
-0.07
NR
NR
NR
SO2
0.85
0.84
0.61
0.51
0.66
-0.10
NR
0.76
0.52
0.55
NR
NR
NR
PM2.5
NR
NR
NR
NR
NR
-0.090
0.48
NR
NR
0.37
0.50
NR
NR
PMio
NR
NR
NR
NR
NR
-0.011
0.33
NR
0.7
NR
NR
0.51
0.04-
0.63
3-77
-------
Table 3-10 (Continued): Synthesis of nitrogen dioxide ambient-ambient correlations from measurements reported
in the literature.
Study3 Averaging Time Location
fEeftens et al. 1-yravg Oslo, Norway
(2012)
Stockholm County, Sweden
Helsinki/Turku, Finland
Copenhagen, Denmark
Kaunas, Lithuania
Manchester, U.K.
London/Oxford, U.K.
the Netherlands/
Belgium
Ruhr Area, Germany
Munich/Augsberg, Germany
Vorarlberg, Austria
Paris, France
Gyor, Hungary
Lugano, Switzerland
Turin, Italy
Rome, Italy
Barcelona, Spain
Scale
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Correlation
Measure
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
Spearman
CO
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
03
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
SO2
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
PM2.5
0.24
0.75
0.71
0.40
0.04
0.40
0.84
0.57
0.69
0.29
0.04
0.86
0.02
0.66
0.65
0.73
0.90
PMio
0.34
0.80
0.80
0.60
0.17
0.59
0.82
0.74
0.65
0.67
0.35
0.91
0.12
0.83
0.67
0.75
0.69
3-78
-------
Table 3-10 (Continued): Synthesis of nitrogen dioxide ambient-ambient correlations from measurements reported
in the literature.
Study3
fEeftens et al.
(2012)
(Continued)
McConnell et al.
(2003)
Averaging Time
1 -yr avg
(Continued)
4-yr avg
Location
Cataluna, Spain
Athens, Greece
Heraklion, Greece
12 communities in southern
California
Scale
Urban
Urban
Urban
Urban
Correlation
Measure
Spearman
Spearman
Spearman
Pearson
CO
NR
NR
NR
NR
03
NR
NR
NR
0.59
SO2
NR
NR
NR
NR
PM2.5
0.72
0.49
0.18
0.54
PMio
0.63
0.70
0.37
0.2
tGan et al. (2012a) 5-yr avg
Vancouver, Canada
Urban
Spearman
NR
NR
NR
0.47
NR
a.m. = ante meridiem; avg = average; AZ = Arizona; CA = California; CO = Colorado; CO = carbon monoxide; GA = Georgia; h = hour; I = interstate; m = meter; MA = Massachusetts;
max = maximum; MD = Maryland; Ml = Michigan; min = minute; MO = Missouri; NC = North Carolina; NM = New Mexico; NR = not reported; NY = New York; O3 = ozone;
OH = Ohio;PA = Pennsylvania; PM25 = in general terms, particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm, a measure of fine particles;
PMio = in general terms, particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm, a measure of thoracic particles (i.e., that subset of inhalable particles
thought small enough to penetrate beyond the larynx into the thoracic region of the respiratory tract); SO2 = sulfur dioxide; TN = Tennessee; TX = Texas; U.K. = United Kingdom;
U.S. = United States of America; Wl = Wisconsin; yr = year.
Correlation data computed from land use regression studies are not included here.
fStudies published since the 2008 ISA for Oxides of Nitrogen
3-79
-------
Within-Hourly Correlations
Within-Daily Correlations
PNC
PH2.5]
SOBI
PNC
Within-Monthly Correlations
wra -
PNC -
-.- -
-•.'i
•J
..... *
03
SOC1
FNC
Annual or Longer-term Correlations
Correlation
Correlation
BC = black carbon; CO = carbon monoxide; EC = elemental carbon; LUR = land use regression; O[3] = ozone; OC = organic
carbon; PM[2.5] =particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PM[10] = particulate
matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; PM[10-2.5] = particulate matter with a nominal
mean aerodynamic diameter less than or equal to 10 |jm and greater than 2.5 |jm; PNC = particle number concentration;
SO[2] = sulfur dioxide; UFP = ultrafine particles.
Notes: Boxes represent the interquartile range of the data with the median line plotted, and 90th and 10th percentile of the data are
plotted as the whiskers. Correlation data computed from LUR studies are not included here. Correlations shown by closed red
circles come from near-road studies, and correlations shown by open black circles either come from urban-regional scale studies or
do not specify the study's spatial scale.
Source: National Center for Environmental Assessment 2014 analysis of data from studies referenced in Table 3-10.
Figure 3-6 Summary of temporal nitrogen dioxide-copollutant correlation
coefficients from measurements reported in studies listed in
Table 3-10, sorted by temporal averaging period.
3-80
-------
Fewer studies have explored seasonal correlations between NO2 concentration and
concentrations of copollutants. Among these, a majority of studies report correlations of
NO2 concentration with PM2 5 and PMio concentrations. In general, studies show stronger
correlations of NO2 concentration with PM2s and PMio concentrations during cooler
seasons compared with warmer seasons. Connell et al. (2005) investigated associations
between PM2 5 concentration and gaseous copollutant concentration in Steubenville, OH
using linear regression. NO2 concentration was more strongly correlated with PM2 5
concentration during the fall (R2 = 0.53) and winter (R2 = 0.53) seasons compared with
the spring (R2 = 0.27) and summer (R2 = 0.086) seasons. Similarly, Sarnat et al. (2005)
found positive associations between PM2 5 concentration and NO2 concentration during
both seasons (summer: ft = 0.44; winter: ft = 0.64), with stronger associations in the
winter in Baltimore, MD. Arhami et al. (2009) evaluated relationships between ambient
copollutants at two sites in southern California (San Gabriel Valley, CA and Riverside,
CA) for warmer and cooler seasons. During the warm season, the Spearman correlation
coefficient (average among sites) was r = 0.09 between NO2 concentration and PM2 5
concentration, whereas during the winter the correlation was r = 0.50. However, they did
not observe a consistent seasonal trend between NO2 concentration and PMio
concentration. While associations between NO2 concentration and PMio concentration
were substantially lower during the summer (r = 0.21) at the Riverside, CA site,
correlations were relatively similar during both seasons at the San Gabriel Valley, CA
site (summer PMio: r = 0.31; winter PMio: r = 0.34). In contrast, for a study of
copollutant variation in Montreal, Canada, Levyetal. (2014) reported higher magnitude
Pearson correlations for concentrations of several copollutants in summer (CO: r = 0.77;
O3: r = -0.74; SO2: r = 0.17; PM25: r = 0.34; UFP: r = 0.77; BC: r = 0.80;
PMio: r = 0.35) compared with winter (CO: r = 0.16; O3: r = -0.36; SO2: r = 0.04;
PM2.5: r = 0.34; UFP: r = 0.71; BC: r = 0.55; PMio: r = 0.35). The Levy etal. (2014)
study measured the pollutants' concentrations using near-real-time instrumentation with
recording intervals ranging from 1 second to 2 minutes.
The relationship between NO2 concentration and Os concentration may also have
seasonal patterns, although limited seasonal data exist between these two pollutants. In
the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). ambient concentrations of NO2
and Os from several sites across Los Angeles, CA were compared during a multiyear
period. Slightly positive correlations between these two pollutants were observed during
the summer (Spearman r = 0.0 to 0.4), while negative correlations were observed during
the winter (r = -0.5 to -0.8). The slightly positive correlations during the summer can be
attributed in part to increased photochemical activity, resulting in enhanced Os formation.
Higher Os concentrations increase the ratio of NO2 concentration to NO concentration
3-81
-------
due to enhanced oxidation, thereby resulting in a stronger correspondence between NO2
concentration and Os concentration during the summer. The magnitude of the relationship
between NCh concentration and Os concentration may be dampened by the nonlinear
relationship between the two species (Pusede and Cohen. 2012). Only one study in
Table 3-10 reported seasonal differences in the correlation between NO2 concentration
and Os concentration. Sarnatet al. (2001) measured daily concentrations of gaseous and
PM pollutants during different seasons in Baltimore, MD. Similar to the trends reported
in the 2008 ISA for Oxides of Nitrogen, they observed a negative correlation between
NO2 concentration and Os concentration during the winter (r = -0.71) and a near-zero
correlation during the summer (r = 0.02). However, because there is a lack of studies
reporting such correlations, it is uncertain whether or not this seasonal trend exists
between the two pollutants in different locations.
Recent studies have also compared NO2-copollutant temporal correlations across
different regions in the U.S., based on central site monitoring data. Baxter et al. (2013)
studied differences in air pollution for the Northeast (Boston, MA; Pittsburgh, PA), South
(Memphis, TN; Birmingham, AL), Midwest (Milwaukee, WI; Detroit, MI), and West
(San Diego, CA; Riverside, CA). Average Spearman correlation coefficients between
PM2.5 concentration and NO2 concentration for each region were different (Northeast:
r = 0.44; South [data available for Memphis only]: r = 0.27; Midwest: r = 0.57; West:
r = 0.47). Schildcrout et al. (2006) compared a number of gaseous and particulate
pollutants in different cities across the U.S., including Albuquerque, NM; Baltimore,
MD; Boston, MA; and Denver, CO. While correlations between ambient NO2
concentration and CO concentration were relatively similar in all four locations, larger
differences were observed between correlations of NO2 concentration and PMio
concentration, ranging from a Spearman correlation of r = 0.64 in Denver, CO to
r = 0.26 in Albuquerque, NM. Other multicity studies conducted outside of the U.S. show
that NO2-copollutant correlations are widely variable across cities (Faustini et al.. 2011;
Dales etal.. 2010. 2009b: Stieb et al.. 2008: Timonen et al.. 2006).
A small subset of studies investigated temporal correlations between NO2 concentration
and concentrations of traffic-related VOCs, such as BTEX. In these studies, correlations
between NO2 concentration and VOC concentrations are variable. Brook et al. (2007)
demonstrated that concentrations of benzo(e)pyrene and hopanes, specific mobile source
tracers, were more strongly correlated with NO2 concentration (Spearman r = 0.27-0.80)
compared to PM2 5 concentration (r = 0.26-0.62) at several urban sites in Canada.
Beckerman et al. (2008) observed correlations between NO2 concentration and BTEX
concentration of Pearson r = 0.46-0.85 in a near-road field campaign. In a panel study,
Greenwald et al. (2013) compared ambient concentrations of traffic pollutants monitored
outside two schools in El Paso, TX, including one school within close proximity to a
3-82
-------
major roadway with heavy diesel truck traffic. A Pearson correlation of r = 0.77 was
observed between NO2 concentration and BTEX concentration, suggesting that both
pollutants are related to traffic sources.
Long-Term Temporal Correlations
Epidemiology studies of long-term NO2 exposure for which interpollutant correlations
were computed were substantially less numerous than for epidemiology studies of
short-term exposure (Atkinson et al.. 2013; Heinrich et al.. 2013; Gan et al.. 2012a;
Darrowetal.. 2011 a; Dong etal.. 2011: Katanoda et al.. 2011; Hwang and Lee. 2010;
Delfino et al.. 2009; Delfino et al.. 2008a; McConnell etal.. 2003). For long-term
averages, most of the studies collected data for correlations with PM2 5 and PMio
concentrations. In each case, the median correlations were near 0.5, and the correlations
were positive and ranging from near 0 to near 0.9. The sample size for other copollutants
was low in the long-term averages. Median correlations were comparable between
long-term exposure and short-term exposure epidemiology studies for concentrations of
CO, 862, PM2s, BC, and PMio. The largest difference was for the correlation between
NC>2 concentration and Os concentration, which was 0.59 over the long-term exposure
epidemiology studies and 0.17 for all studies pooled. However, given that only three
long-term studies were available to compute correlation between NO2 concentration and
Os concentration and one of those three studies reported a negative correlation, there is
insufficient information to make a conclusion regarding independence of the effects of
NC>2 concentration and Os concentration. Long-term correlations were not computed for
concentrations of UFP, EC, OC, PNC, PMio-2.5, benzene, and toluene, and the small
relative number of long-term exposure epidemiology studies compared with short-term
exposure epidemiology studies reporting temporal correlations add uncertainty to these
numbers.
3.4.4.2 Spatial Variability among Ambient Nitrogen Dioxide and Copollutants
When an epidemiologic study design relies on spatial contrasts to draw conclusions, such
as for an epidemiologic study of long-term exposure, unmeasured spatial correlation
between copollutants may lead to positive bias in the health effect estimate for each of
the pollutants included in the model. Paciorek (2010) performed simulations and
analyzed case study data (of the relationship between birth weight data and BC
concentrations in eastern Massachusetts) to test the effect of spatial errors on health effect
estimates in long-term exposure epidemiologic studies. He identified unmeasured spatial
confounding as a key driver in biasing health effect estimates in a spatial regression.
3-83
-------
Paciorek (2010) maintained that bias can be reduced when variation in the exposure
metric occurs at a smaller spatial scale than that of the unmeasured confounder.
Dionisio et al. (2013) compared the coefficient of variation (CV = o/u) of six air
pollutants' concentrations across space using a hybrid AERMOD-background model of
concentrations in the Atlanta, GA metropolitan area. They observed the following ordinal
relationship of the covariates' CVs: NOX (0.88) > CO (0.58) > EC (0.50) > PM25
(0.13) > O3 (0.07) > SO4 (0.05). Dionisio etal. (2013) did not report the CV for NO2
concentration, which would be expected to have a lower CV than NOx concentration.
Likewise, Goldman et al. (2012) and Ivy et al. (2008) both used monitoring data from the
Atlanta, GA metropolitan area to estimate spatial correlation functions, and they observed
that NO2 concentration and NOx concentration, along with CO, SO2, and EC
concentrations, had substantially steeper spatial correlograms than Os, PMio, PM2 5, SO4,
NOs, NFL, and OC concentrations. Sajani etal. (2011) also observed that spatial
correlation decreased more substantially with distance between monitoring sites for NO2
concentration compared with PMio and Os concentrations when looking at six Italian
cities.
Changes in correlations across space have been observed in a small number of studies.
For their long-term near-road study, Ducret-Stich et al. (2013) point out that the temporal
correlations of NO2 concentration with EC and PNC concentrations were high close to
the highway where they obtained measurements and decreased with increasing distance
from the road. This suggests that the influence of NO2 exposure on health effects might
be better detected in an epidemiologic study of long-term exposure when the participants
are further from the road so that an independent effect can be detected. Atari et al. (2009)
tested the relationship between NO2 concentration and SO2 concentration across
individual-level and census tract-level spatial resolutions, which were estimated by a
LUR model developed for testing odor threshold in Sarnia, Canada. They observed
higher spatial correlation when averaging over a census tract (r = 0.65) compared with
individual-level resolution (r = 0.49). These findings illustrate greater spatial variability
for NO2, NOx, CO, SO2, and EC concentrations compared with concentrations of the
other pollutants. Based on the conclusions of Paciorek (2010), the observations noted in
Dionisio etal. (2013). Goldman et al. (2012). Ivy et al. (2008). Saiani etal. (2011). Atari
et al. (2009). and Sanchez Jimenez et al. (2012) suggest that differences in the spatial
variability of NO2 concentration compared with copollutants having different spatial
variation make it unlikely that copollutant confounding will occur everywhere in space.
This is consistent with the findings of Ducret-Stich et al. (2013) regarding differences in
copollutant correlations over space.
3-84
-------
3.4.4.3 Relationships among Personal, Indoor, and Ambient Nitrogen Dioxide and
Copollutant Exposures
Many studies have investigated the relationship between personal exposure and ambient
concentrations of NC>2 and other pollutants to evaluate the use of central site
measurements as a proxy for personal exposure to ambient air pollution. Other studies
have explored relationships between indoor NCh concentration and copollutant
concentrations to understand sources and personal exposure in an indoor environment.
Tables 3-11, 3-12. 3-13. and 3-14 present correlations of ambient NC>2 concentration,
personal NO2 exposure, or indoor NC>2 concentration with similar measurements of
copollutants. A limited number of studies reported in the 2008 ISA for Oxides of
Nitrogen (U.S. EPA. 2008c) investigated the relationship between personal NO2 exposure
and personal exposures or ambient concentrations of other pollutants (e.g., PM2s, EC,
CO, volatile organic compounds, and HONO). Short-term correlation of personal NO2
exposure with these pollutants ranged from Spearman r = 0.26 to r = 0.71. Similar to the
results in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). correlations of
r = -0.33 to r = 0.44 were observed between personal NO2 exposure and personal
exposures or ambient concentrations of other regional (PlVfc 5) and traffic-related
pollutants (e.g., EC, OC). Additionally, personal exposures or ambient concentrations of
Os consistently showed a negative or no correlation with personal exposures or ambient
concentrations of NO2. More recent studies report indoor NO2-copollutant correlations
and observe a broader range of correlations between NO2 concentration and EC
concentration of r = -0.37 to r = 0.66.
3-85
-------
Table 3-11 Pearson correlation coefficients between ambient nitrogen dioxide and personal copollutants.
Study Location n Averaging Times PIVh.s EC OC Os
tPelfinoetal. (2008a) Los Angeles, CA <170 All: 24 h 0.32 0.2 0.16 NR
tSuhandZanobetti(2010b) Atlanta, GA <277 All: 24 h 0.25 0.33 NR -0.09
tWilliams et al. (2012a) Chapel Hill, NC
tSchembari et al. (2013) Barcelona, Spain
<357 All: 24 h
<65 NO2: 7 day;
PM2.5/EC: 2 day
O-i Q n -1 7 MD n f"H
0.21 0.44 NR NR
CA = California; EC = elemental carbon; GA = Georgia; h = hour; n = sample size; NC = North Carolina: NO2 = nitrogen dioxide; NR = not reported; O3 = ozone; OC = organic
carbon; PM25 = participate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
Table 3-12 Pearson correlation coefficients between personal nitrogen dioxide and ambient copollutants.
Study Location n Averaging Times PIVh.s EC OC Os
tDelfinoetal. (2008a) Los Angeles, CA <170 All: 24 h 0.21 0.2 0.18 NR
tSuhandZanobetti(2010b) Atlanta, GA <277 All: 24 h 0.2 0.22 NR NR
tWilliams et al. (2012a)
tSchembari et al. (2013)
Chapel Hill, NC
Barcelona, Spain
<326
<65
All: 24 h
NO2: 7 day;
PM2.5/EC: 2 day
0.33
0.28
OO
0.22
NR
NR
OOfi
NR
CA =California; EC = elemental carbon; GA = Georgia; h = hour; n = sample size; NC = North Carolina; NO2 = nitrogen dioxide; NR = not reported; O3 = ozone; OC = organic carbon;
PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
3-86
-------
Table 3-13 Pearson correlation coefficients between personal nitrogen dioxide and personal copollutants.
Study
tPelfinoetal. (2008a)
tSuhandZanobetti(2010b)
tWilliams et al. (2012a)
tSchembari et al. (2013)
Location
Los Angeles, CA
Atlanta, GA
Chapel Hill, NC
Barcelona, Spain
n
<486
<277
<326
<65
Averaging Times
All: 24 h
All: 24 h
All: 24 h
NO2: 7 day;
PM2.5/EC: 2 day
PM2.5
0.38
0.29
0.06
0.11
EC
0.22
0.49
0.33
0.3
oc
0.2
NR
NR
NR
03
NR
OHQ
0-1 -1
NR
CA = California; EC = elemental carbon; GA = Georgia; h =hour; n = samples size; NC = North Carolina; NO2 = nitrogen dioxide; NR = not reported; O3 = ozone; OC = organic
carbon; PM25 = participate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
3-87
-------
Table 3-14 Correlation coefficients between indoor nitrogen dioxide and indoor copollutants.
Study Location
fSarnat et al. El Paso, TX (Site A)
(2012)a
El Paso, TX (Site B)
Ciudad Juarez, Mexico (Site A)
Ciudad Juarez, Mexico (Site B)
fGreenwald et al. 2 sites in El Paso, TX
(201 3)b
n Averaging Times
15 NO2:4day;
PM2.5/EC:2 day
15 NO2:4day;
PM2.5/EC:2 day
15 NO2:4day;
PM2.s/EC:2 day
15 NO2:4day;
PM2.5/EC:2 day
18-26 All: 4 day
PM EC OC
-0.35(PM2.5) 0.58 NR
-0.26 (PMio-2.s)
-0.19 (PMio)
0.06 (PM2.s) -0.37 NR
0.28 (PMio-2.s)
0.12 (PMio)
-0.29 (PM2.s) 0.66 NR
-0.58 (PMio-2.s)
-O.S(PMio)
-0.04 (PM2.s) 0.45 NR
-0.5 (PMio-2.s)
-0.34 (PMio)
0.76 (PM2.s) 0.45 NR
0.83 (PMio)
03
NR
NR
NR
NR
NR
EC = elemental carbon; NO2 = nitrogen dioxide; NR = not reported; O3 = ozone; OC = organic carbon; PM = particulate matter; PM25 = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |jm; PM10 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; PM10-25 = particulate matter
with a nominal mean aerodynamic diameter less than or equal to 10 |jm and greater than 2.5 |jm; TX = Texas.
aSpearman correlation.
bPearson correlation.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
3-8
-------
In addition to these findings, higher correlations were typically observed between
ambient concentrations of NO2 and other traffic-related pollutants (Section 3.4.3.1)
compared to personal measurements [e.g., correlations among personal exposure
measurements in Table 3-13; (Schembari et al., 2013; Williams et al., 2012a; Suh and
Zanobetti. 201 Ob; Delfino et al.. 2008aVI. For example, (Suh and Zanobetti. 201 Ob)
observed a stronger relationship between ambient NO2:EC (r = 0.61) and ambient
NO2:PM2 5 (r = 0.47) compared to personal NO2:EC (r = 0.49) and personal NO2:PM2 5
(r = 0.29). Delfino et al. (2008a) observed similar results in the NO2:EC relationship in a
health study investigating the relationship between traffic-related pollution and lung
function decrements in Los Angeles, CA. While the ambient NO2:EC correlation was
r = 0.61, lower correlations were observed for personal NO2:EC (r = 0.22). Additionally,
a small number of time-series studies have used NO2 concentration in receptor models to
relate health effects to sources/factors (Baxter et al.. 2013; Cakmak et al.. 2009; Halonen
et al., 2009; Mar et al., 2000). Each of these studies used factor analysis, the U.S. EPA
positive matrix factorization method,1 or PCA analysis and found high loadings of NO2
and traffic-related copollutants (e.g., EC, OC, CO) on the same factor, which was
attributed to traffic-related pollution.
Correlations between NO2 and VOC concentrations also suggest different sources for
personal exposure. For example, Martins etal. (2012) estimated personal NO2 and BTEX
exposures during four 1-week periods using a microenvironmental approach that
combined outdoor and indoor concentrations with time-activity patterns. It consistently
observed correlations of r = -0.42 to r = 0.14 between NO2 concentration and BTEX
concentration during different seasons. The lack of correlation between these pollutants
can be attributed in part to differences in sources between indoor and outdoor
microenvironments. While exposure to VOCs, namely benzene, was attributed mainly to
indoor sources, NO2 concentration was largely associated with traffic sources. These
studies emphasize that proximity to roadways and time spent in various indoor and
outdoor microenvironments can impact the relationship between NO2 and traffic-related
VOCs.
Weaker correlations observed between personal measurements of NO2 exposure and
other traffic-related pollutant exposures (compared to ambient concentration correlations)
suggest that personal exposure to NO2 may include a number of outdoor and indoor
sources comprising traffic and nontraffic emissions (e.g., gas stoves, residential wood
burning, biomass burning). These observations provide further evidence that nonambient
sources of NO2 provide interference to the ambient NO2 measurement signal. At the same
time, the weaker correlations between total personal NO2 exposure and copollutant
1 http ://intranet.epa. gov/heasd/products/pmf/pmf. htm.
3-89
-------
exposures indicate that for panel studies of total personal NO2 exposure, ambient
copollutants would be less likely to confound health effect estimates for personal NO2
exposure. Titration conditions for NO, NC>2, and Os also likely differ from indoors to
outdoors, given variation in solar radiation and other atmospheric factors that influence
atmospheric chemistry. Additionally, personal exposures are influenced by building air
exchange rate and time-activity patterns that differ among study participants. This is in
contrast to ambient NCh concentrations, which appear to be largely driven by variability
in traffic pollution in many areas. This type of exposure error associated with ambient
concentrations is discussed in more detail in Section 3.4.3.3.
Few studies have reported indoor NCh-copollutant correlations for short-term averaging
times, focusing on correlations between NCh concentration and PM concentration in
different size fractions as well as NO2 concentration and BC concentration. In these
studies, correlations of Spearman r = -0.37 to 0.66 were observed between indoor NCh
concentration and EC concentration; however, lower correlations are observed for indoor
NO2 concentration and PM concentration compared with NCh concentration and EC
concentration. Sarnatet al. (2012) measured indoor concentrations of NC>2, EC, PM2 5,
PMio-25, and PMio at four elementary schools in two cities near the U.S.-Mexico border:
El Paso, TX and Ciudad Juarez, Mexico. NO2 and PM concentrations showed weaker
and/or inverse correlations at all four elementary schools (r = -0.58 to 0.12). Greenwald
et al. (2013) later conducted a follow-up study to Sarnat et al. (2012) and measured
similar pollutants at the same schools in El Paso, TX. Although Greenwald et al. (2013)
reported similar NO2-EC correlations to those reported in Sarnat etal. (2012). stronger
correlations were observed between NO2 and PM2 5 concentrations (r = 0.76) and
between NCh and PMio concentrations (r = 0.83). Differences in the NO2-PM correlations
between these two studies reflect that NO2 and PM can have many different sources in
indoor environments, which impact their temporal and spatial patterns. Moreover, the
results of Greenwald et al. (2013) suggest the potential for confounding of NCh health
effect estimates by PM based on indoor concentrations. Taken together, the existence and
extent of such confounding is uncertain.
In general, ambient NO2 concentration would not necessarily be expected to correlate
well with personal exposures of copollutants. For example, in the case where the exposed
population spends time at residences or workplaces sufficiently far from the near-road
environment, personal NC>2 exposure would not be expected to correlate with ambient
copollutants of traffic-related origin. Low correlations between ambient NCh
concentration and personal exposures to copollutants could support inferences regarding
the independent effects of NC>2.
3-90
-------
3.4.4.4 Traffic and Noise as Confounders
For the purpose of inferring causality from the body of epidemiologic studies of
short-term and long-term exposure to traffic-related pollutants, the Health Effects
Institute Report on Traffic-Related Air Pollution (HEI. 2010) raised the concern that
distance-to-road models are especially subject to confounding the associations between
health effects and exposures because traffic indicators may encompass additional
information, such as noise, unmeasured air pollutants, stress, and socioeconomic status,
that may also be associated with the health effects of interest. However, recent evidence
is mixed regarding the correlations of NO and NC>2 concentrations with traffic and noise
levels. Most of these studies are for short-term exposure. Hence, the role of traffic and
noise as confounders or independent variables in the relationship between health effects
and NO or NO2 exposure is unclear.
Several studies have examined the relationship of traffic-related noise with NO and NO2
concentrations. Kheirbek et al. (2014) added noise level meters to the dense New York,
NY monitoring project described in Ross et al. (2013) and observed that 1-week avg
noise level, obtained at 60 locations during Fall 2012, correlated with Pearson
r = 0.59 for NO2 concentration and r = 0.61 for NO concentration. Davies et al. (2009)
measured 2-week avg of NO2 and NOx concentrations concurrently with 5-minute noise
samples at 103 sites and observed correlations of r = 0.53 for NO2 concentration and
r = 0.64 for NOx concentration. Ganetal. (2012b) calculated the correlations among air
pollutants and noise from road traffic and aircraft using 5-minute data from 103 sites in
Vancouver, Canada during 2003 (dates not stated). They observed lower correlations for
NO2 concentration with road traffic noise (Spearman r = 0.33) and aircraft noise
(r = 0.14) compared with the correlation of NO concentration with these two noise
sources (road traffic: r = 0.41; aircraft: r = 0.26). For both NO2 and NO concentration,
correlations were higher for road traffic noise than aircraft noise. Over a 5-year avg, Gan
et al. (2012a) reported the correlation between NO2 concentration and noise from road
traffic of Spearman r = 0.33 from Ganetal. (2012b). as well as a correlation between NO
concentration and noise from road traffic of Spearman r = 0.39.
Ross etal. (2011) also examined relationships of different frequency noises with NO and
NO2 concentrations using continuous monitors collecting 48,000 samples per second for
six 24-hour periods in August 2009. Ross etal. (2011) measured the relationships
between traffic level, noise, and concentrations of NO2 and NO in New York, NY as part
of the Ross etal. (2013) study. Unweighted noise of all frequencies was uncorrelated
with NO2 concentration (Spearman r = -0.01) but correlation increased for NO
concentration (Spearman r = 0.43) for all times. Correlations were higher for medium
frequency noise (NO2: r = 0.22; NO: r = 0.57). Correlations between noise and traffic
3-91
-------
counts segregated by fleet mix were generally higher for cars (unweighted noise:
r = 0.37; medium frequency: r = 0.33), trucks (unweighted noise: r = 0.64; medium
frequency: r = 0.71), and buses (unweighted noise: r = 0.61; medium frequency: r = 0.60)
compared with the correlations with nonsegregated traffic data. Likewise, at night, high
frequency noise was correlated with NC>2 concentration (r = 0.83) and NO concentration
(r = 0.73).
Distance to road has also been observed to influence the relationship between noise and
NO2 concentration for both long-term and short-term noise and NC>2 exposure studies.
For the years 1987-1996, Beelen et al. (2009) estimated correlations among 1-year avg
NC>2 concentration, traffic level, and noise, and they observed correlations between traffic
and noise depending on spatial designation (r = 0.30-0.38) and for the correlation of NC>2
concentration and noise (r = 0.46). When segregating loud noise >65 dBA, correlation
dropped (r = 0.22). Note that Beelen et al. (2009) did not specify whether Pearson or
Spearman correlations were computed. Ross et al. (2011) noted within-day variability in
these relationships, where truck and car traffic are correlated (r = 0.81) during the
morning rush hour but inversely correlated at night (r = -0.67). Dadvand et al. (2014c)
measured 24-hour avg noise, NOx concentration, and NCh concentration at 50-m, 200-m,
500-m, and beyond 500-m buffers from the road in Barcelona, Spain from 2001-2005
and observed that all three decreased with increasing distance from the road. Measured
temporal Spearman correlation of noise was r = 0.45 for NC>2 concentration and r = 0.56
for NOx concentration. Allen et al. (2009) also studied the relationship between NO2
concentration, UFP concentration, and 5-minute avg A-weighted equivalent noise for
105 locations in Chicago, IL and Riverside, CA using measurements taken in December
2006 and April 2007. After adjustment for regional unspecified air pollutant
concentration gradients, Pearson correlations with noise were r = 0.16-0.62 for NO2
concentration (winter Chicago, IL: r = 0.16; spring Chicago, IL: r = 0.41; spring
Riverside, CA: r = 0.62) and 0.49-0.62 for NO concentration. In Chicago, IL,
correlations of noise with NO and NO2 concentrations were higher within a 100-m buffer
of the road, while correlations of noise with NO and NO2 concentrations were lower
within a 100-m buffer in Riverside.
For short-term exposure studies, more evidence is available to consider the relationship
between traffic-related noise and NO2 concentration compared with long-term exposure
studies. Collectively, these studies suggest that potential for confounding of NO2 effects
by noise may be influenced by temporal and spatial resolution of the data, noise
frequency, and fleet mix. Specifically, confounding is less probable as distance from the
road increases. However, total noise may be unlikely to act as a confounder. It should be
noted that noise would also have to be etiologically related to the health outcome under
3-92
-------
consideration to confound the relationship between the health effect and NO2 exposure.
When noise is decomposed by frequency, confounding is more likely.
3.4.5 Implications for Epidemiologic Studies of Different Designs
Estimates of NO2 exposures are subject to errors that can vary in nature, as described in
Section 3.4.3. Classical error is defined as error scattered around the true personal
exposure and independent of the measured exposure. Classical error results in bias of the
epidemiologic health effect estimate that is often towards the null. Classical error can also
cause inflation or reduction of the standard error of the health effect estimate. Berkson
error is defined as error scattered around the exposure surrogate (in most cases, the
central site monitor measurement) and independent of the true value (Goldman etal..
2011; Reeves etal.. 1998).
Definitions for Berkson-like and classical-like errors were developed for modeled
exposures. These errors depend on how exposure metrics are averaged across space.
Szpiro etal. (2011 a) defined Berkson-like and classical-like errors as errors sharing some
characteristics with Berkson and classical errors, respectively, but with some differences.
Specifically, Berkson-like errors occur when the modeled exposure does not capture all
of the variability in the true exposure. Berkson-like errors increase the variability around
the health effect estimate in a manner similar to pure Berkson error, but Berkson-like
errors are spatially correlated and not independent of predicted exposures, unlike pure
Berkson errors. Szpiro and Paciorek (2013a) simulated Berkson-like errors' influence on
health effect estimates (see also, Szpiro and Paciorek (2013b)). For the case simulated
where spatial variability in the exposure estimates from measured concentrations
exceeded the spatial variability in the true exposures (which were modeled to be
uniform), the health effect estimates were biased away from the null. For the case
simulated where covariates were included in the health model but not the exposure
model, the health effect estimates were biased towards the null. Hence, Berkson-like
error can lead to bias of the health effect estimate in either direction. Classical-like errors
can add variability to predicted exposures and can bias health effect estimates in a
manner similar to pure classical errors, but they differ from pure classical errors in that
the variability in estimated exposures is also not independent across space.
The results of Meng et al. (2012b). described in Section 3.4.2. illustrated that
epidemiologic study design can influence the relationship between personal exposure to
NO2 and ambient concentrations (Table 3-7). This meta-analysis found that correlations
were highest for short-term exposure community time-series epidemiology studies
(designated as "daily average" in Table 3-7). and correlations were lowest for
3-93
-------
longitudinal panel cohort studies. The following sections consider how exposure
assessment errors may influence interpretation of health effect estimates for
epidemiologic studies of different designs.
3.4.5.1 Community Time-Series Studies
In most short-term exposure epidemiologic studies of the health effects of NO2, the health
effect endpoint is modeled as a function of ambient exposure, Ea, which is defined as the
product of ambient concentration, Ca, and a, a term encompassing time-weighted
averaging and infiltration of NC>2 (Section 3.4.1). Community time-series epidemiologic
studies capturing the exposures and health outcomes of a large cohort frequently use the
concentration measured at a central site monitor (Ca,csm) as a surrogate for Ea in an
epidemiologic model (Wilson et al.. 2000). At times, an average (unweighted or
weighted) of central site monitored concentrations is used for the Ea surrogate. For
studies involving thousands of participants, it is not feasible to measure personal
exposures. Moreover, for community time-series epidemiology studies of short-term
exposure, the temporal variability in concentration is of primary importance to relate to
variability in the health effect estimate (Zeger et al., 2000). The magnitude of bias in the
health effect estimate will decrease and the precision of the health effect estimate will
increase as the temporal correlation ofCa,csm with the true air pollutant exposure
increases. Spatial variability in NO2 concentrations across the study area could attenuate
an epidemiologic health effect estimate if the exposures are not correlated in time with
Cucsm when central site monitoring is used to represent exposure. If exposure assessment
methods that more accurately capture spatial variability in the concentration distribution
over a study area are employed, then the confidence intervals around the health effect
estimate may decrease. The following several paragraphs describe studies that tested the
influence of different types of exposure error on the health effect estimate. Because the
majority of these studies were conducted for one metropolitan area (Atlanta, GA), caution
must be taken when interpreting the study results described.
Goldman et al. (2011) simulated the effect of classical and Berkson errors due to
spatiotemporal variability among ambient or outdoor air pollutant concentrations over a
large urban area on health effect estimates of emergency department (ED) visits for a
time-series study of cardiovascular disease. The relative risk (RR) per ppm was
negatively biased in the case of classical error (1-hour daily max NCh: -1.3%;
1-hour daily max NOx: -1.1%) and negligibly positively biased in the case of Berkson
error (1-hour daily max NO2: 0.0042%; 1-hour daily max NOX: 0.0030%). The 95%
confidence interval range for RR per ppm was wider for Berkson error (1-hour daily max
3-94
-------
: 0.028; 1-hour daily max NOx: 0.023) compared with classical error (1-hour daily
maxNO2: 0.0025; 1-hour daily max NOX: 0.0043).
Recent studies have explored the effect of spatial exposure measurement error on health
effect estimates to test the appropriateness of using central site monitoring for time-series
studies. Goldman et al. (2010) simulated spatial exposure measurement error based on a
semivariogram function across monitor sites with and without temporal autocorrelation at
1- and 2-day lags. Their goal was to analyze the influence of spatiotemporal variability
among ambient or outdoor concentrations over a large urban area on a time-series study
of ED visits for cardiovascular disease. A random term was calculated through Monte
Carlo simulations based on the data distribution from the semivariogram, which
estimated the change in spatial variability in exposure with distance from the monitoring
site. The average of the calculated random term was added to a central site monitor
concentration time series (considered in this study to be the base case) to estimate
population exposure to NO2 subject to spatial error. For the analysis with temporal
autocorrelation considered, RR per ppm for 1-hour daily max NC>2 dropped slightly to
1.0046 (95% CI: 1.0026, 1.0065), and RR per ppm for 1-hour daily max NOX dropped to
1.0079 (95% CI: 1.0057, 1.0100) when both were compared with the central site monitor
RR per ppm = 1.0139 (for all air pollutants).1 When temporal autocorrelation was not
considered, RR per ppm dropped to 1.0044 for 1-hour daily max NC>2 and 1.0074 for
1-hour daily max NOx. The results of Goldman et al. (2010) suggest that spatial exposure
measurement error from use of central site monitoring concentration data results in
biasing the health effect estimate towards the null, but the magnitude of the change in
effect was small.
Goldman et al. (2012) also studied the effect of different types of spatial averaging of the
exposure surrogate on bias in the health effect risk ratio and the effect of correlation
between measured and "true" ambient exposures of NO2 and NOx to analyze the
influence of spatiotemporal variability among ambient or outdoor concentrations over a
large urban area on health effect estimates. Concentrations were simulated at alternate
monitoring locations using the geostatistical approach described above for Goldman et al.
(2010) for the 20-county Atlanta, GA metropolitan area for comparison with
concentration measurements obtained directly from monitors at those sites.
Geostatistical-simulated concentrations were considered to be "true" in this study, and
other exposure assignment methods were assumed to have some error. Five different
exposure assignment approaches were tested: using concentrations from a single central
site monitor, averaging the simulated concentrations across all monitoring sites,
performing a population-weighted average of concentrations across all monitoring sites,
that 95% CIs were not reported for the central site monitor RR or for the cases where temporal autocorrelation
was not considered.
3-95
-------
performing an area-weighted average of concentrations across all monitoring sites, and
population-weighted averaging of the geostatistically simulated concentrations
(Table 3-15). Goldman et al. (2012) observed that the exposure measurement error was
somewhat correlated with both the measured and true values, reflecting both Berkson and
classical error components. For the central site monitor, the exposure measurement errors
were somewhat inversely correlated with the true value but had relatively higher positive
correlation with the measured value. For the other exposure assignment methods, the
exposure measurement errors were inversely correlated with the true exposures, while
they had positive but lower magnitude correlation with the measured concentrations. At
the same time, the exposure measurement bias, given by the ratio of the exposure
measurement error to the measured concentration, was much higher in magnitude at the
central site monitor than for the other exposure assignment methods for NCh
concentrations. For NOx concentrations, exposure measurement bias for the central site
monitor was much higher than for the other exposure assignment methods with the
exception of the area-weighted average, which produced a large negative exposure
measurement bias. Hence, compared with other exposure assignment methods, the health
effect estimate would likely have greater bias towards the null with reduced precision
when a central site monitor is used to measure NO2 concentration as a surrogate for
exposure. However, exposure error is also likely to cause some bias and imprecision in
the effect estimate for other exposure surrogate methods. These findings suggest more
Berkson error in the more spatially resolved exposure assignment methods compared
with the central site monitor and more classical error for the central site monitor estimate
compared with the other exposure assignment techniques. Hence, more bias and less
precision would be anticipated for the health effect estimate calculated from the central
site monitor compared with the more spatially resolved methods. It was observed that the
more spatially variable air pollutants studied in Goldman et al. (2012) also had more bias
in the health effect estimates. This was noted across exposure assignment methods but
was more pronounced for the central site measurement data.
Butland et al. (2013) conducted a simulation study to test how spatial resolution of the
NO2 concentration measures used for exposure assignment influences health effect
estimates in a time-series epidemiologic model of mortality in urban and rural areas. The
test domain was subdivided into squares ranging in area from 1 km2 to 25 km2. Health
effect estimates simulated using the 1-km2 resolution area were considered to be "true,"
and mortality estimates were sampled from a Poisson distribution of mortality data.
Monitor data were simulated based on a lognormal distribution using the correlogram
among pairs of NC>2 concentration monitors to establish the variability of the distribution
as a function of distance. The error structure in the model was constructed to include both
Berkson and classical components. Health effect estimates for mortality based on NO2
exposures were attenuated by 29 and 38% for urban and rural areas, respectively, when
3-96
-------
reducing the spatial resolution from 1 km2 to 25 km2 over a 3-year time-series analysis.
Because the Butlandetal. (2013) study was conducted in the U.K., where the NCh
monitoring network has a larger number of near-road sites, the nature and magnitude of
the exposure error can be expected to differ for the U.S.
Table 3-15 The influence
Exposure Estimation Approach
of exposure
Bias[(Z-Z*)/Z]a
metrics
R2(Z, Z*)b
on error in health
/?[(Z-Z*), Z*]c
effect estimates.
/?[(Z-Z*),Zf
NO2
Central site monitor
Unweighted average
Population-weighted average
Area-weighted average
Geostatistical model —
population-weighted average
0.62
0.25
0.18
-0.07
N/A
0.24
0.38
0.38
0.38
0.45
-0.46
-0.73
-0.78
-0.87
-0.82
0.61
0.20
0.14
-0.04
0.0017
NOx
Central site monitor
Unweighted average
Population-weighted average
Area-weighted average
Geostatistical model —
population-weighted average
0.71
0.31
0.03
-0.88
N/A
0.33
0.45
0.46
0.47
0.52
-0.11
-0.63
-0.81
-0.96
-0.80
0.81
0.29
0.02
-0.31
-0.00042
N/A = not applicable; NO2 = nitrogen dioxide; NOX = the sum of nitric oxide and NO2; R = Pearson correlation; R2 = coefficient of
determination; Z = the measured concentration; Z* = the true concentration.
Note: Model errors were based on comparisons between measured data and simulated data at several monitoring sites. Errors
were estimated for a single central site monitor, various monitor averages, and values computed from a geostatistical model. Z
denotes the measured concentration, and Z* denotes the true concentration, considered here to be from the chemical transport
model. Bias in the exposure metric is given as the proportion of error between the measurement and true value to the
measurement.
aData are from Figure 5 and provided by the authors (Goldman. 2013).
bData are from Figure 4 and provided by the authors (Mulholland. 2013).
°Pearson correlation.
Source:Data compiled from Table 1, Figure 4, and Figure 5; used with permission of Elsevier, Goldman et al. (2012).
3-97
-------
Sarnat et al. (2010) studied the spatial variability of concentrations of NO2, CO, Os, and
PM2 5 in the Atlanta, GA metropolitan area and how it affects interpretation of
epidemiologic results, using time-series data for circulatory disease ED visits. Sensitivity
to spatial variability was examined at slightly greater than neighborhood scale (8 km) in
this study. Interestingly, Sarnat et al. (2010) found that relative risk varied with distance
between the monitor and study population when comparing urban to rural locations, but
distance of the study population to the monitor was not an important factor when
comparing urban population groups. This suggests that, even for spatially heterogeneous
NO2, urban-scale concentration measures may produce results comparable to
neighborhood-scale concentration measures if the sites were comparable throughout the
city, for example, as a result of similar traffic patterns. However, Sarnat etal. (2010)
cautioned that, because their study was limited to 8-km radii, it is not possible to interpret
this work with respect to near-road and on-road microscale concentrations.
In a study of the effect of concentration metric choice (central site, arithmetic average
across space, or population-weighted average) used to assign exposure in a time-series
epidemiologic model, Strickland et al. (2011) found that choice of the concentration
metric resulted in large differences in the observed associations between ED visits for
pediatric asthma and exposure for spatially heterogeneous NO2 but not for spatially
homogeneous PM2 5 when using a unit standardization for computing the relative risk.
However, when Strickland et al. (2011) used IQR for standardization, there were little
differences among the relative risk estimates across the concentration metrics. The
differences observed between unit and IQR standardization are due to the fact that the
IQR reflects the spatial variability in the exposure metrics for the spatial and
population-weighted averages.
Error type also influences the health effect estimate from time-series studies. Dionisio et
al. (2014) decomposed the exposure measurement error into spatial and population-based
components. Spatial error was defined as the difference between concentration simulated
by an AERMOD dispersion model and concentration measured at a central site monitor,
and population error was defined as the difference between the SHEDS exposure model
(using only ambient sources) and the dispersion model. Errors were computed for each
ZIP code centroid. Three pollutants with high spatial variability (NOx, CO, EC) termed
"local" and three pollutants with low spatial variability (PM2 5, Os, SO/O termed
"regional" by the authors were included in the study. Although NO2 concentration was
not included explicitly, the local pollutant results are relevant. Dionisio et al. (2014)
observed more variability in both the spatial and population components of the exposure
measurement error across the ZIP codes for the local pollutants compared with the
regional pollutants. Attenuation of the health effect estimate by the spatial error
component was much larger for the local pollutants compared with the regional
3-98
-------
pollutants, and the amount of bias by the spatial error component was roughly the same
for NOx, CO, and EC. However, the population error component caused much more
attenuation of the health effect estimate for NOx compared with CO and EC. In fact, CO
had negligible bias of the health effect estimate due to the population error component.
This discrepancy is possibly related to the deposition rate of NOx and differences in
sunlight affecting the NO2/NOx ratio indoors compared with CO, which has a zero
deposition rate and is modeled as not reactive in SHEDS. Given that NO2 has a higher
deposition rate than NO, the results of Dionisio et al. (2014) suggest that health effect
estimates modeled in time-series studies of NOx exposure are likely extendable to NO2
(see Section 3.3.2.1 for information related to deposition of indoor NO2). Hence, it is
likely that spatial variability and indoor deposition both cause bias in the health effect
estimate for studies of NO2 exposure.
Nonambient sources of NO2 tend to diminish the correlation between NO2 concentration
measured at a central site monitor and total personal NO2 exposure measurements
(Section 3.4.2). Analyses of time-series epidemiologic studies have suggested that
nonambient contributions introduce Berkson error into the exposure term, where the error
does not bias health effect estimates for ambient NO2 assuming that nonambient NO2
sources are independent of ambient sources, but it does cause the confidence intervals
around the health effect estimates to widen (Sheppard. 2005; Wilson etal.. 2000). No
data from cohort studies are available to test if this theory can be applied more broadly to
all epidemiologic studies. Sheppard et al. (2005) simulated the effect of nonambient
sources for a time-series study of the health effects of PM exposure and found that, as
long as the ambient and nonambient sources were uncorrelated, the nonambient
exposures would widen the confidence interval around the health effect estimates but
would not bias the health effect estimate. This result is generalizable to NO2 because it
did not depend on the particle size distribution. Moreover, the data in Table 3-6 and
Section 2.3 illustrate seasonal variability in ambient NO2 concentrations and in the
relationship between ambient concentrations of NO2 and personal NO2 exposure.
Therefore, it can be anticipated that the influence of nonambient NO2 exposures on the
confidence interval around the health effect estimate would vary with season.
Exposure measurement error related to instrument precision has a smaller effect on health
effect estimates in time-series studies compared with error related to spatial gradients in
the concentration because instrument precision would not be expected to modify the
ability of the instruments to respond to changes in concentration overtime. Goldman et
al. (2010) investigated the influence of instrument error on health effect estimates in a
time-series epidemiology study by studying differences in exposure assignments and
health effect estimates obtained using copollutant monitors. In this study, a random error
term based on observations from copollutant monitors was added to a central site
3-99
-------
monitor's concentration time series to simulate population estimates for ambient air
concentrations subject to instrument precision error in 1,000 Monte Carlo simulations.
Very little changes in the risk ratios were observed for 1-hour daily max NCh and
1-hour daily max NOx concentrations. For 1-hour daily max NCh concentration, the RR
per ppm of NCh concentration with simulated instrument precision error was
1.0133 compared with RR per ppm = 1.0139 for the central site monitor. For 1 -hour daily
max NOx concentration with simulated instrument precision error, RR per
ppm = 1.0132 compared with the central site monitor's RR of 1.0139. The amount of bias
in the health effect estimate related to instrument precision was very small.
3.4.5.2 Long-Term Average Cohort Studies
For cohort epidemiologic studies of long-term human exposure to NCh, where the
difference in the magnitude of the concentration is of most interest, ifCa,csm is used as a
surrogate for Ea, then a can be considered to encompass the exposure measurement error
related to uncertainties in the time-activity data and air exchange rate. Spatial variability
in NCh concentrations across the study area could lead to bias in the health effect estimate
ifCa,CSm is not representative ofEa. This could occur, for example, if the study participants
are clustered in a location where their NO2 exposure is higher or lower than the exposure
estimated at a modeled or measurement site. There is limited information regarding
whether Ca,Csm is a biased exposure surrogate in the near-road environment for
epidemiologic studies of long-term exposure.
Sensitivity of the epidemiologic model to the temporal and spatial characteristics of
exposure data depends on the temporal characteristics of the disease process. Birth
outcome studies serve as an example where the exposure window becomes an important
consideration that helps to delineate short-term exposure from long-term exposure
epidemiologic study design. For example, Ross et al. (2013) studied the role of spatial
and temporal resolution of NCh estimates in the application of LURto study the
relationship between birth outcome data in New York City (NY) and NC>2 exposure.
Seasonal variability was more evident when averaging NCh estimates across the final
6 weeks of gestation compared with the entire gestation period, but temporal variation
had less influence on NCh predictions compared with PIVb 5 predictions. This finding
reflects the fact that variability in NCh concentrations is more prominent in space than in
time compared with PIVb 5 concentrations. Additionally, Brauer et al. (2008) studied the
influence of NCh exposure models (IDW of central site monitoring data and LUR) on
health effect estimates for birth outcomes data in Vancouver, Canada between
1999-2002. IDW produced monthly average NCh concentrations matched to the month
of pregnancy, while LUR was built using a dense passive sampling network deployed in
3-100
-------
2003. Brauer et al. (2008) observed higher adjusted odds ratios for IDW compared with
LUR (which produced health effect estimates closer to null, see Section 6.4.3). This
finding may have been related to temporal coincidence of the monitoring and health data
for the IDW and nearest monitor approaches; temporal coincidence is not possible for
LUR. Clark et al. (2010) compared IDW with LUR for the analysis of asthma risk, based
on hospitalizations in British Columbia (BC), Canada, from in utero and first-year-of-life
exposure to NCh, NO, and other pollutants. They observed comparable adjusted odds
ratios for the first year of NC>2 exposure and higher adjusted odds ratio for IDW
compared with LUR for in utero NC>2 exposures (Section 6.4.3). The biologically relevant
time period in eliciting a birth outcome likely determines whether spatial or temporal
variation in concentration is more important to the epidemiologic model. It is possible
that, if the biologically relevant time period is short, then temporal variability may play a
larger role. In that case, the seasonal differences in NO2 concentration become more
important for measuring an effect. If the biologically relevant time period is longer, then
the spatial contrasts evident in concentration maps become more important so that
exposure error can lead to over- or under-estimation of the effect.
Spatial resolution of the exposure estimates has been evaluated to examine the influence
of spatial exposure error in cohort studies. This has been considered with spatially
resolved alternatives to central site monitoring data, such as data from a LUR, to describe
exposure of individuals within a cohort that is spatially dispersed within a study area
(Section 3.2.2). Sellier et al. (2014) and Lepeule et al. (2010) evaluated various
approaches to estimate exposure (nearest central site monitor, geostatistical model, LUR
model, dispersion model) in a study of birth weight among a French mother-child cohort
in the French cities of Nancy and Poitiers. Correlations among the methods varied with
respect to methodology, distance, and land use type. For example, the correlation
between LUR and dispersion modeling had a minimum Pearson r = 0.58 (for urban
locations), while the correlation between central site monitoring and LUR had a
minimum r = 0.20 (also for urban locations). No effect of the method was observed on
change in birth weight, but confidence intervals around the health effect estimate
generally increased for dispersion models, which tended to be the most spatially
heterogeneous among the four methods studied.
The influence of spatial exposure error on health effect estimates varies with the
particular study parameters, such as model selection and location. Madsen et al. (2010)
compared odds ratios for birth weight from the National Birth Registry of Norway per
quartiles of NCh concentrations estimated from a near-road monitoring station and a
dispersion model. Higher exposure variability was captured by the dispersion model, but
the adjusted odds ratio showed an effect only for the near-road monitoring station
exposure data, where time-averaged or residential exposures were likely to be
3-101
-------
overestimated. Wu et al. (201 la) compared health effect estimates for birth outcomes
from four hospitals in Los Angeles, CA and Orange counties, CA given
NO2 concentrations as estimated using nearest monitors and LUR. Odds ratios for NO2
concentrations were comparable for nearest monitor and LUR for Los Angeles County,
CA, where the LUR was fit, but the odds ratio decreased for Orange County in
comparison with nearest monitor. This is consistent with studies reporting higher
exposure error when LUR models are fit in one city and applied elsewhere, as described
in Section 3.2.2.1. Ghosh et al. (2012a) compared health effect estimates for low birth
weight based on birth certificate data and NO2 concentration estimates from LUR (scaled
to account for seasonal fluctuations in concentration) to nearest monitoring station in Los
Angeles County, CA and found negligible difference between the health effect estimates
obtained with each exposure assignment method.
Minimization of error in the exposure estimate does not always minimize error in the
health effect estimate. Szpiro et al. (201 la) performed a simulation study to evaluate bias
and uncertainty of the health effect estimate obtained when using correctly specified and
misspecified exposure simulation conditions, where correct specification was considered
for comparison purposes to be the use of three spatial prediction variables and
misspecification implied unmeasured error in the model. LUR was used to simulate
exposure; the misspecified model omitted a geographic covariate in the LUR. Szpiro et
al. (201 la) also reduced the amount of variability in the third covariate when simulating
the monitoring network data in an additional set of simulations. Prediction accuracy of
the exposure estimate was higher for the correctly specified model compared with the
misspecified model. However, the health effect estimate was more variable for the
correctly specified model compared with the misspecified model when the variability in
the exposure covariate in the monitoring data decreased. The results of Szpiro et al.
(2011 a) suggested that use of more accurately defined exposure metrics in a cohort study
does not necessarily improve health effect estimates, and their influence depends on the
relative variability of the exposure covariates. The Szpiro et al. (201 la) simulations were
for a generic air pollutant but are relevant for NO2.
Basagana et al. (2013) also investigated the effect of differences in LUR model fitting on
error in the epidemiologic health effect estimates in a simulation study based on
cardiovascular disease data from the Girona (Italy) Heart Registry. For the exposure
estimate, Basagana et al. (2013) fit three LUR models with 20, 40, or 80 measurement
locations. For this simulation study, the model considered correctly specified contained
five covariates. As a comparison case, Basagana et al. (2013) fit misspecified models
containing 20 or 100 covariates (including the five original covariates). The
misspecification effectively added error to the model. The simulated exposure error
produced a combination of Berkson-like and classical-like errors on the health effect
3-102
-------
estimate. Compared with the true health effect estimate, bias towards the null was
observed to increase with decreasing number of measurement locations used to fit the
LUR model. At the same time, the mean squared error of the health effect estimate
increased with decreasing number of measurement locations. Moreover, bias towards the
null and mean squared error also grew with increasing the number of covariates from 5 to
20 to 100. Notably, in-sample R2 did not trend with the number of variables while
out-of-sample R2 increased with increasing number of sites (based on sites not collocated
with the samplers used for model fitting), suggesting that in-sample R2 is not a sufficient
measure of LUR model quality.
Error correction is a relatively new approach to estimate the correct standard error and
potentially correct for bias in longitudinal cohort studies (Szpiro etal.. 201 Ib). Szpiro
and Paciorek (2013a) established that two conditions must hold for the health effect
estimate to be predicted correctly: the exposure estimates from monitors must come from
the same underlying distribution as the true exposures, and the health effect model
includes all covariates relevant to the population. Szpiro and Paciorek (2013a) performed
several simulations to investigate what happens when these conditions are violated. In
one set of simulations, the distribution of the exposure was varied. When the assigned
exposure measurements were set to be uniform across space, the health effect estimate
was biased away from the null with different standard error compared with the case when
the exposure subjects were collocated with the study participants. When an additional
spatial covariate was omitted, the health effect estimate was biased towards the null with
different standard errors compared with the correctly specified model. Bias correction
and bootstrap calculation of the standard errors improved the model prediction, even
when the true model contained several degrees of freedom. (Spiegelman. 2013) noted that
the new measurement error correction methods developed by Szpiro and Paciorek
(2013a) are a version of regression calibration. This study illustrated the influence of
classical-like and Berkson-like errors on long-term exposure cohort study health effect
estimates through these simulations.
Not accounting for time-activity patterns of study participants adds uncertainty to
exposure estimates obtained via spatial modeling such as LUR. Setton et al. (2011)
investigated how both spatial variability and unaccounted study participant mobility bias
health effect estimates in long-term exposure epidemiologic models of health effects
from NC>2 exposure in a simulation study based on data from cohorts in southern
California and Vancouver, Canada. In this case, concentration at each participant's home
was modeled (using the Comprehensive Air Quality Model with Extensions [CAMx] for
southern California and using LUR and IDW interpolation of monitoring data for
Vancouver). Populations were simulated using human activity data for Vancouver,
Canada and transportation survey data for southern California. Bias in the health effect
3-103
-------
estimate increased in magnitude towards the null with distance from home and time spent
away from home. Moreover, when spatial variability increased (through comparison of
spatially variable LUR-derived NCh concentrations with a smoother monitor-based
approach for mapping NCh concentrations for the Vancouver, Canada data), the health
effect estimate obtained from the IDW-based approach was closer to the null compared
with the LUR-based health effect estimate. Setton etal. (2011) interpreted this finding as
evidence of the influence of smoothing spatially heterogeneous concentration profiles on
the health effect estimate.
Instrumentation bias could be anticipated to influence health effect estimates from
epidemiologic studies of long-term NCh exposures in some situations. Section 3.2.1.2
describes how passive monitors are likely to overestimate exposure given the influences
of ambient temperature, relative humidity, and presence of copollutants. Therefore, LUR
exposure may be overestimated when the LUR is fit using passive monitoring data.
Sections 2.4.1 and 3.2.1.1 describe how the presence of copollutants can also cause NO2
concentrations measured using central site monitors to be overestimated. Overestimating
exposure can bias health effect estimates. Ambient temperature and relative humidity
would not be expected to vary greatly within a city. However, local copollutant
concentrations may be spatially variable such that an LUR model fit, and resulting health
effect estimates, could have some differential bias in the health effect estimates across a
city related to instrument error. Because climate and ambient sources are more likely to
differ among cities, instrumentation error could have a larger influence on the
comparison of health effect estimates among cities when LUR or central site monitors are
used to estimate exposures.
In the case of long-term exposure cohort studies, nonambient contributions to the total
personal exposure estimates would be expected to widen the confidence interval around
the health effect estimates by adding noise to the exposure signal. Also, addition of any
non-negative nonambient component to the personal exposure measurement would result
in an overestimate of exposure to ambient NO2, because the average total personal NO2
exposure would have to be either equal to or greater than the average personal exposure
to ambient NC>2. This exposure error could bias the health effect estimate towards the
null.
3.4.5.3 Panel Studies
Consideration of errors in use of Ca,csm as a surrogate for Ea provides information on the
impact of this proxy measure on health effect estimates in panel studies. Van Roosbroeck
et al. (2008) evaluated health effect estimates among a panel of children for associations
3-104
-------
of four respiratory outcomes with 48-hour NCh data from a single monitor located at the
children's school. These health effect estimates were compared with those obtained from
personal NO2 monitoring to capture spatial variability in NC>2 concentrations and
time-activity data. Van Roosbroeck et al. (2008) observed that health effect estimates
were biased towards the null by roughly one-half to one-third when using a single
monitor outside the school in lieu of personal exposure monitors. In this case, bias in the
single-monitor health effect estimate was likely influenced by the spatial variability of the
NC>2 concentration profile, time-activity of the study participants, and infiltration of
ambient NO2 indoors. The authors also adjusted the health effect estimate for nonambient
sources, including parental smoking, gas cooking, and presence of an unvented water
heater.
Sarnat etal. (2012) considered the influence of exposure surrogate on health effect
estimates obtained for a panel of school children. This study was conducted along the
U.S.-Mexico border in El Paso, TX and Ciudad Juarez, Mexico, and 96-hour avg
concentrations measured from central site chemiluminescent monitors, passive monitors
outside the children's schools, and passive monitors inside the children's schools were all
used as surrogates for exposure to NCh. The largest health effect estimate was observed
for measurements outside the school. In comparison, the health effect estimates for NO2
measured inside the schools and at central site monitors were several times smaller
(Table 5-16). Based on the comparison between outdoor and central site monitoring
results, Sarnat etal. (2012) concluded that exposure error from using central site
measurements, in lieu of measurements at the site of exposure, could lead to biasing the
health effect estimate towards the null. They proposed that this bias was related to the
failure of central site monitors to capture intra-urban spatial variability. The 2008 ISA for
Oxides of Nitrogen (U.S. EPA. 2008c) also did not find conclusive evidence of the
influence of exposure measurement error on health effect estimates from panel
epidemiologic studies of NO2 exposure. In general, there is uncertainty regarding the
influence of NC>2 monitor placement on the magnitude and directionality of bias of the
health effect estimate as related to use of central site monitors in lieu of localized
monitors in panel studies. As for epidemiologic studies of long-term NC>2 exposure
(Section 3.4.5.2). panel studies with multiple sites could be affected by instrumentation
error, which could lead to overestimates of exposure at some but not all locations. This
could have a differential influence on health effect estimates, especially for intercity
comparisons.
3-105
-------
3.5 Conclusions
This chapter presents the current state of the science for assessment of human exposure to
NO2. It builds upon the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). which
concluded that errors associated with the use of NO2 concentrations measured at central
site monitors as exposure metrics for epidemiologic studies tended to bias the health
effect estimate towards the null for both short- and long-term exposure epidemiologic
studies. As detailed within this chapter, recent studies provide support for the conclusions
presented in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) for short-term
exposure studies but differ in some cases for long-term exposure studies.
Commonly used exposure assessment methods include central site monitors, passive
monitors, LUR, CTM, and dispersion models (Section 3.2). The influence of
measurement errors from each of these techniques varies with study design. These
methods are listed in Table 3-1, along with their application (i.e., the design of the study
in which they are used) and associated errors. Community time-series studies of
short-term NO2 exposure typically use central site monitoring. Panel studies tend to
employ central site monitors or, in some cases, passive monitors. Studies of long-term
NO2 exposure often use a variety of methods, including central site monitors, LUR,
dispersion models, spatial smoothing techniques, and spatiotemporal models. Errors
associated with these methods vary in importance based on their application. LUR
estimates of NO2 exposure have been validated by independent methods when the model
was trained and applied in the same general location so that the exposure estimates and
true exposures are assumed to come from the same data distribution. Dispersion
modeling can be subject to errors related to simplifying assumptions about the
meteorology, urban or natural topography, or photoreactivity of NO to form NO2.
Additionally, NO2 exposure estimates from inverse distance weighting or other spatial
smoothing techniques can be subject to error if the spatial scale of monitoring does not
capture all sources. Studies employing exposure estimates obtained using these methods
often report R2 (whether in-sample or out-of-sample varies with study), bias, and/or mean
squared error to describe the quality of the exposure estimates. Given that these metrics
do not always correlate, caution must be taken to interpret the quality of exposure data
from an individual study on the basis of one metric.
Factors contributing to error in NO2 exposure assessment include temporal activity of
epidemiologic study participants, spatial variability of NO2 concentrations across the
study area, infiltration of NO2 indoors, and instrument accuracy and precision
(Section 3.4.3). With respect to time-activity data, variability within and among different
populations causes the limitation of having only one monitoring location in many studies
to have varying influence on exposure estimates within and among those different
3-106
-------
populations. In general, spatial misalignment can occur when the time-activity patterns of
study participants are not factored into the study design or when the location where NC>2
exposure is estimated does not coincide with the residential, school, or work location of
interest. Spatial variability of human exposure can add uncertainty to the exposure
estimate if it is not characterized by the monitors. As a result, there is a potential for
exposure error if the ambient NC>2 concentration measured at a given site differs from that
at the location of an epidemiologic study participant, and this issue is present regardless
of the spatial scale of the epidemiology study. At the same time, the influence of spatial
variability depends strongly on the temporal design of the epidemiologic study, as
described in the paragraphs below. Infiltration and air exchange rate influence indoor
levels of NO2 in the absence of indoor sources and hence presents the potential for bias
and uncertainty in a, which depends on air exchange rate, penetration, and indoor
deposition. NCh monitors are often subject to positive biases resulting from interference
byNOy species.
Community time-series epidemiologic studies most commonly use central site monitors
to estimate human exposure to ambient NO2 (Section 3.4.5.1). Temporal variability in
exposure is the relevant feature of the exposure data in a community time-series study.
Additionally, personal exposure measurements cannot feasibly be obtained for health
studies with large numbers of participants. There is some uncertainty associated with
using central site measurements of NC>2 concentrations to represent personal exposure
because the temporal variability of the central site exposure estimate may differ from the
temporal variability of the true exposure. Exposure estimates using NO2 concentration
measurements from central site monitors do not capture the spatial variability of the
concentration field, which becomes a more important source of error for time-series
epidemiology studies if the NO2 concentrations at the locations of the study participants
are not well correlated with measurements at the central site monitor. Nonambient
contributions and differential infiltration of NC>2 can also add error or uncertainty to a
health effect estimate. Instrument precision and accuracy are not thought to have a
substantial influence on health effect estimates in time-series studies. Simulation studies
testing the influence of exposure error in time-series studies suggest that exposure error
may widen the confidence intervals of the health effect estimate and bias the estimate
towards the null. This implies that reported health effect estimates for time-series studies
of NC>2 exposure are potentially lower than true health effect estimates or that the
reported confidence intervals around those health effect estimates are wider than the true
confidence intervals.
Long-term exposure epidemiology studies compare subjects or populations at different
locations (Section 3.4.5.2). Therefore, spatial, rather than temporal, contrasts are more
important in epidemiologic studies of long-term exposure. NC>2 concentrations measured
3-107
-------
at central site monitors are often used to represent exposures when human health cohorts
are compared among cities. There is some uncertainty associated with using central site
measurements of NO2 concentrations to represent personal exposure, because the
distribution of concentrations measured at a central site monitor may vary from the
distribution of true exposures. LUR models, CTMs, dispersion models, spatial smoothing
models, and spatiotemporal models may be used to estimate exposure at the residential
locations of study participants in epidemiologic studies of long-term exposure, because
those models are designed to capture spatial variability of NO2 concentration within a
geographic area, such as a city. Differences between the exposure estimates and the true
exposures can add bias or reduce the precision of the health effect estimate. Moreover,
positive biases from measurement of NOy artifacts have the potential to enhance spatial
contrasts in exposure models. The magnitude and direction of bias and the size of
confidence intervals depend on differences between the distribution of true exposures and
the distribution of concentrations estimated by the exposure assignment method.
Panel epidemiologic studies of NO2 exposure using central site monitors are subject to
exposure error due to spatial misalignment between the monitored ambient NCh
concentration and the true personal exposure to ambient NO2 (Section 3.4.5.3). Available
panel studies that compare health effect estimates among exposure assessment techniques
have suggested that such spatial misalignment leads to attenuating the health effect
estimate. However, only a limited number of panel studies have examined the influence
of exposure measurement error on health effect estimates. For this reason, it is difficult to
reach a conclusion about the magnitude and direction of error in the health effect
estimates related to exposure error.
Confounding can occur when common sources emit multiple pollutants and other
stressors (e.g., noise) and therefore have the potential to increase uncertainty in
identifying whether the copollutants are independently associated with a health effect
(Section 3.4.4). Studies of noise suggest that total noise may be unlikely to act as a
confounder. However, when noise is decomposed by frequency, confounding of the
independent effect of NO2 is more likely for high frequency noises that are associated
with truck traffic. For traffic-related pollutants, NO (reacting to NCh), CO, EC, UFP, and
benzene are commonly co-emitted and can be highly correlated with NO2 in time and
space. During winter, NO2 emitted from heating fuel sources can also be highly
correlated with PIVbs and PMio. For both short-term exposure and long-term exposure
epidemiologic studies, it is difficult to distinguish the health effect associated with NO2
exposure among health effects attributed to other highly correlated pollutants. The
temporal correlations among copollutants may vary over space. For epidemiologic
studies of long-term NO2 exposure, bias related to copollutant confounding can be
reduced when the spatial scale of the NO2 exposure metric is smaller than the spatial
3-108
-------
scale of the correlated copollutants. Bias related to copollutant confounding may be less
likely for unstable copollutants (e.g., UFP) or air pollutants that disperse more quickly
than NO2 (e.g., CO), compared with more spatially homogeneous pollutants (e.g., PM2 5)
However, panel studies based on personal exposure measurements or outdoor residential
measurements do not appear to have high copollutant confounding, especially when
receptors live far from busy roads. Therefore, panel studies may be the best design for
demonstrating if NC>2 has independent health effects.
3-109
-------
CHAPTER 4 DOSIMETRY AND MODES OF
ACTION FOR INHALED OXIDES
OF NITROGEN
4.1 Introduction
This chapter has two main purposes. The first is to describe the principles that underlie
the dosimetry of nitrogen dioxide (NCh) and nitric oxide (NO) and to discuss factors that
influence it. The second is to describe the modes of action that may lead to the health
effects that will be presented in Chapter 5 and Chapter 6. This chapter is not intended to
be a comprehensive overview, but rather to update the basic concepts derived from the
NC>2 and NO literature presented in the 1993 Air Quality Criteria for Oxides of Nitrogen
(AQCD) and the 2008 Integrated Science Assessment (ISA) for Oxides of Nitrogen (U.S.
EPA. 2008a. c, 1993a. b) and to introduce the recent relevant literature.
In Section 4.3. particular attention is given to chemical properties of inhaled NO2 and NO
that affect absorption, distribution, metabolism, and elimination. Inhaled NO2 and NO,
and subsequent reaction products, are discussed in relation to endogenous production of
these chemical species. Because few NO2 dosimetry studies have been published since
the 1993 AQCD (U.S. EPA. 1993a). much of the information from that report has been
pulled forward into the current document and is discussed in the context of more recent
research. The topics of dosimetry and modes of action are bridged by reactions of NO2
with components of the epithelial lining fluid (ELF) and by reactions of NO with heme
proteins, processes that play roles in both uptake and biological responses.
Section 4.3 highlights findings of studies published since the 2008 ISA (U.S. EPA.
2008c) that provide insight into the biological pathways affected by exposure to NO2 and
NO. Earlier studies that represent the current state of the science are also discussed.
Studies conducted at more environmentally relevant concentrations of NO2 and NO
(i.e., <5,000 parts per billion [ppb], Section 1.1) are of greater interest because biological
pathways responsible for effects at higher concentrations may not be identical to those
occurring at lower concentrations. Some studies at higher concentrations are included if
they were early demonstrations of key biological pathways or if they are recent
demonstrations of potentially important new pathways. This information is used to
develop a mode of action framework for inhaled NO2 and NO that serves as a guide to
interpreting health effect evidence presented in subsequent chapters; in Chapter 5 and
Chapter 6.
4-1
-------
4.2 Dosimetry of Inhaled Oxides of Nitrogen
4.2.1 Introduction
This section provides a brief overview of NCh and NO dosimetry and updates
information provided in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008a. c).
Dosimetry refers to the measurement or estimation of the amount of a compound, or its
reaction products, absorbed and/or generated at specific sites in the respiratory tract
during an exposure. New to this ISA is the inclusion of basic information regarding the
endogenous production of NO2 and NO. It is important to consider inhaled NO2 and NO
and their subsequent reaction products in relation to endogenous production of these
chemical species. To establish an environmentally relevant context, ambient NO2 and NO
concentrations are briefly discussed below; more detail is provided in Chapter 2.
Ambient concentrations of NO2 and NO are variable. For example, ambient NO2
concentrations are highest in the winter months, near major roadways, during weekday
morning hours, and decrease moderately during the afternoon (see Atlanta, GA data in
Figures 2-20 and 2-21). One-hour average, near-road (15m) NO2 concentrations in Los
Angeles, CA ranged from 3 to 80 ppb with median values of about 40 ppb in the winter
and 30 ppb in the summer months of 2009 (Polidori and Fine. 2012b). Away from major
roadways, 1-hour average NO2 concentrations may still reach 50 to 70 ppb with median
NO2 concentrations between roughly 10 to 30 ppb depending on the season and distance
from roadways (Polidori and Fine. 2012b). As will be discussed, the uptake of inhaled
NO2 may potentially increase levels of NO2-derived reaction products beyond levels
endogenously occurring in the respiratory tract.
Similar to NO2, ambient NO concentrations are highest in the winter months near major
roadways during weekday morning hours, but decrease to very low levels during the
afternoon (see Atlanta, GA data in Figures 2-20 and 2-21). One-hour average, near-road
(15m) NO concentrations in Los Angeles, CA ranged from 0 ppb to over 400 ppb with
median values of about 50 ppb in the winter and 20 ppb in the summer months of 2009
(Polidori and Fine. 2012b). Away from major roadways, 1-hour average NO
concentrations may still reach 250 ppb, but median NO concentrations are 5 ppb or less
(Polidori and Fine. 2012b). For the same roadway (Interstate 710), Zhu et al. (2008)
reported on-road NOx (i.e., the sum of NO and NO2) concentrations of around 400 ppb
(average of eight 2-hour samples collected between 10:00 a.m. and noon during the
period from June 2006 to May 2007). As will be discussed, these ambient NO
concentrations are generally in the range of those occurring endogenously in the
respiratory tract.
4-2
-------
4.2.2 Dosimetry of Nitrogen Dioxide
NO2 is a highly reactive gas that occurs as a radical (technically a resonance structure)
wherein the unpaired electron is more localized to the nitrogen atom than either of the
oxygen atoms. Once inhaled, NC>2 first encounters the aqueous phase of the ELF, which
is a contiguous but biologically complex aqueous fluid layer that covers all of the
respiratory tract surfaces (Bastacky et al.. 1995). The ELF composition shows
appreciable heterogeneity with respect to anatomic site and species. The ELF of alveolar
surfaces and conducting airway surfaces has a monomolecular layer of surface active
lipids (Bernhard et al.. 2004; Hohlfeld. 2002; Mercer et al.. 1994). largely fully saturated,
which reduces surface tension and may provide a resistive barrier to the interfacial
transfer of NC>2 (Sections 4.2.2.1.2 and 4.2.2.1.3). Upon dissolution into the ELF, NC>2 is
converted from a gas to a nonelectrolyte solute, and thus becomes subject to partitioning
and reaction/diffusion. Thus, the ELF represents the initial barrier between NCh
contained within the intra-respiratory tract gas phase and the underlying epithelia
(Postlethwait and Bidani. 1990). NCh chemically interacts with antioxidants, unsaturated
lipids, and other compounds in the ELF. It preferentially reacts with one electron donors
(e.g., small molecular weight antioxidants, protein thiols, etc.), undergoes radical-radical
addition reactions, may abstract allylic hydrogen atoms from polyunsaturated fatty acids,
and through a complex series of reactions, can add to unsaturated fatty acids to generate
nitrolipids (Bonacci et al.. 2012: Rudolph etal.. 2010: O'Donnell et al.. 1999). The
compounds thought responsible, in large part, for the health effects of inhaled NO2 are
the reaction products themselves or the metabolites of these products in the ELF.
Quantifications of absolute NO2 absorption reported in the 1993 AQCD and the 2008
ISA (U.S. EPA. 2008c. 1993a) are briefly discussed below for thoroughness.
4.2.2.1 Mechanisms of Absorption of Nitrogen Dioxide
At the time of the 1993 AQCD (U.S. EPA. 1993a). it was thought that inhaled NO2
probably reacted with the water molecules in the ELF to form nitrous acid (HNCh) and
nitric acid (HNOs). However, some limited data suggested that the absorption of NO2 was
linked to reactive substrates in the ELF and subsequent nitrite (N(V) production. By the
time of the 2008 ISA (U.S. EPA. 2008c). chemical reactions between NO2 and ELF
substrates were more readily recognized as governing NO2 absorption in the respiratory
tract.
4-3
-------
4.2.2.1.1 Reactions of Nitrogen Dioxide with Water and Solutes
Previous studies have demonstrated that it is not NO2 but instead the NO2 dimer,
dinitrogen tetroxide (^Ch), that reacts with water to yield N(V and nitrate [N(V;
(Tinlayson-Pitts et al.. 2003; Schwartz and White. 1983; England and Corcoran. 1974)].
However, in aqueous solutions, NCh rapidly reacts with many solutes (e.g., ascorbate and
urate), particularly those that are easily oxidized. At environmentally relevant
concentrations of NCh (e.g., around 100 ppb), the direct reactions of NO2 with dissolved
substrates become important because, at equilibrium, there is very little N2O4 compared
to NCh. For example, using the difference in Gibbs energies for the formation of gaseous
NO2 and N2O4 (Chase. 1998). one can calculate that at equilibrium, when the
concentration of NO2 is 1,000 and 100 ppb, there are 1.48 x 105 and 1.48 x 106,
respectively, molecules of NO2 for each molecule of N2O.4. Thus, at environmental
exposure levels there are approximately 1.5 million NO2 molecules for each N2O4
molecule. At these concentrations, it is far more likely for NCh (compared to ^Ch) to
penetrate into the aqueous milieu of the ELF. Ensuing reactions of NCh with dissolved
reactive substrates become more likely than reaction with a second NO2 molecule (to
form N2O4). During uptake by pure water, all reactions occur via N2O4 regardless of the
concentration of NCh. However, in the presence of dissolved reactive substrates and at
low, environmentally relevant concentrations of NCh, this process (i.e., reactions
occurring via N2O4) becomes unlikely, and instead uptake occurs via direct reactions of
NO2 with reactive substrates. This resembles reactive uptake of NC>2 by the ELF that
would entail direct reactions of NCh with, for example, dissolved small molecular weight
antioxidants like glutathione (GSH), ascorbate, or urate.
Enami et al. (2009) revisited the discussions regarding NCh reaction with water versus
ELF solutes. Because the authors postulated that NCh effects are largely due to nitrate
formation and acidification via proton production, this issue warrants some discussion.
The claim by Enami etal. (2009) that "antioxidants catalyze the hydrolytic
decomposition of NCh.. .but are not consumed in the process" is problematic in view of
the vast existing environmental health literature that regards NCh as an oxidant gas (Pryor
et al.. 2006: Augusto et al.. 2002: Ford et al.. 2002: Kirsch et al.. 2002: Wardman. 1998:
Postlethwait et al.. 1995: Huie. 1994: Netaetal.. 1988: Finlavson-Pitts et al.. 1987:
Kikugawa and Kogi. 1987: Priitz etal.. 1985: Pryor and Lightsey. 1981). However,
Enami et al. (2009) measured nitrate without measuring nitrite, and therefore, their data
do not strongly support their contention, except to suggest that some hydrolysis of NCh
may be occurring because nitrate was detected. Nitrite data are important because any
excess nitrite found (reaction with water generally yields a 1:1 ratio of nitrite and nitrate;
thus, a yield of nitrite above 1 would be considered in excess) would indicate that it is the
4-4
-------
main product formed as a result of one-electron oxidations by NCh. Thus, by not
measuring nitrite, an important index to assess oxidation by NCh was missed.
Note that Enami et al. (2009) conducted their experiments in the absence of oxygen,
which makes their model inapplicable to the lung. At environmentally relevant
concentrations and physiologic temperatures, intra-pulmonary gas phase NC>2 will exist in
its monomeric form. Furthermore, in the presence of aqueous-phase reactive substrates,
nitrite, but little or no nitrate, is formed during controlled in vitro exposures. Thus, broad
reactivity of NC>2 with a diversity of reactive substrates (solutes) within the ELF
facilitates chemical interactions with antioxidants, lipids, and proteins/peptides/amino
acids.
4.2.2.1.2 Governing Determinants of Nitrogen Dioxide Absorption within the Respiratory
Tract
The absorption of inhaled NCh into the ELF is governed by a process termed "reactive
absorption" that involves dissolution followed by chemical reaction with reactive
substrates in the ELF (Postlethwait and Bidani. 1990). as well as reactions within the
interfacial region. Due to the limited aqueous solubility of NC>2 and thus the rapid
saturation of the aqueous phase interfacial thin film (Bidani and Postlethwait 1998). the
net flux of NO2 into reactant-free water is constrained by the relatively slow direct
reaction of NO2 with water (see above) compared with its radical reactions with
biological substrates (further discussion below). Thus, rapid reactions with ELF
substrates provide the net driving force for NCh mass transfer from the intra-pulmonary
gas phase into the ELF (Bidani and Postlethwait. 1998; Postlethwait and Bidani. 1994;
Postlethwait et al.. 1991a; Postlethwait and Bidani. 1990). Concentrations of "free" solute
NC>2 are likely negligible due to reaction-mediated removal. Empirical evidence suggests
that acute NO2 uptake in the lower respiratory tract is rate governed by chemical
reactions of NO2 with ELF constituents rather than solely by gas solubility in the ELF,
wherein the reaction between NO2 and water does not significantly contribute to the
absorption of inhaled NCh (Postlethwait and Bidani. 1994. 1990). Absorption was also
observed to increase with increasing temperature, an indication of chemical reaction
rather than aqueous solubility, where solubility increases with temperature decrements
(Postlethwait and Bidani. 1990). Postlethwait et al. (1991b) proposed that inhaled NO2
(<10,000 ppb) did not penetrate the ELF to reach underlying sites and suggested that
cytotoxicity likely was initiated by products formed during NC>2 reactions with ELF
constituents. Subsequently, the reactive absorption of NO2 was examined in a number of
studies that sought to identify the substrates that predominantly drive NCh reactive
absorption and to quantify the mass transfer kinetics of NC>2 in the respiratory tract.
4-5
-------
Uptake was observed to be first-order with respect to NO2 at concentrations less than
10,000 ppb, to be aqueous substrate dependent, and to be saturable, meaning that the
absolute amount of NO2 uptake would reach a maximum value even if reactive substrate
concentrations were in significant excess (Postlethwait et al.. 199la. b).
The absorption of inhaled NCh is thought to be coupled with either radical-mediated
hydrogen abstraction to form HNCh (Postlethwait and Bidani. 1994. 1989) or electron
transfer from ELF anionic species that directly reduces NC>2 to nitrite (Adgent et al..
2012). Both mechanisms produce an organic radical from the initial ELF substrate. At
physiologic pH, any formed HNC>2 subsequently dissociates to hydrogen ion (H+) and
nitrite. The concentration of the resulting nitrite is likely insufficient to alter
physiological function because basal nitrite levels may not change appreciably in either
the respiratory tract or the circulation due to ambient NC>2 exposure. This is, in part,
because nitrite will diffuse into the underlying epithelial cells and vascular space where,
in the presence of red blood cells, it is oxidized to nitrate [(Postlethwait and Bidani. 1989;
Postlethwait and Mustafa. 1981): Section 4.2.2.5]. Consequently, by default, effects are
probably attributable to the organic radical secondary oxidants formed (Adgent et al..
2012; Velsoretal.. 2003: Velsor and Postlethwait. 1997) and/or to the proton load,
although the ELF buffering capacity is anticipated to compensate for environmentally
relevant exposure-related proton generation.
Postlethwait et al. (1995) sought to determine the preferential absorption substrates for
NC>2 in the ELF lavaged from male Sprague-Dawley rats. Because bronchoalveolar
lavage (BAL) fluid collected from rats may be diluted up to 100 times relative to the
native ELF (the dilution will be procedure specific), the effect of concentrating the BAL
fluid on NO2 absorption was also investigated. A linear association was found between
the first-order rate constant for NCh absorption and the relative concentration of the BAL
fluid constituents. This suggested that concentration of the reactive substrates in the ELF
determines, in part, the rate of NO2 absorption. The absorption due to specific ELF
constituents was also examined in chemically pure solutions. Albumin, reduced cysteine,
glutathione, ascorbate, and urate were the hydrophilic moieties found to be the most
active substrates for NO2 absorption. Unsaturated fatty acids (such as oleic, linoleic, and
linolenic) were also identified as active absorption substrates and thought to account for
up to 20% of NO2 absorption. Vitamins A and E exhibited the greatest reactivity of the
substrates that were examined. However, the low concentrations of urate (the ELF of
rodents and some primates contains significantly less urate than the ELF of humans due
to differences in nitrogenous waste metabolism) and vitamins A and E were thought to
preclude them from being appreciable substrates in vivo. The authors concluded that
ascorbate and glutathione were the primary NC>2 absorption substrates in rat ELF.
Postlethwait et al. (1995) also found that the pulmonary surfactant component,
4-6
-------
dipalmitoyl phosphatidylcholine (DPPC), was relatively unreactive towards NO2, and
subsequent studies documented that compressed monomolecular interfacial films of
DPPC inhibit NO2 absorption in vitro (Connor et al.. 2001). Documenting whether
surface active phospholipids (which comprise surfactant) inhibit NC>2 mass transfer
in vivo is extremely challenging because any in situ manipulations that disrupt the
surface tension-lowering actions of a surfactant lead to a plethora of pathophysiologic
sequelae. However, even though such potentially important influences on NC>2 mass
transfer have not been verified in vivo, modeling studies could estimate how such effects
would influence the intra-pulmonary distribution of inhaled NC>2, local mass transfer
rates, and thus dosimetry.
4.2.2.1.3 Reaction/Diffusion of Nitrogen Dioxide in the Epithelial Lining Fluid, Potential for
Penetration to Underlying Cells
Because the uptake of NO2 from inhaled air into the ELF is governed by reactive
absorption, it may be postulated that rapid ELF reactions prevent NCh from reaching
underlying respiratory tract tissues. To evaluate this supposition, consideration must be
given to the time required for NC>2 to diffuse through some thickness of the ELF versus
the rate of NO2 reactions with substrates in that ELF.
The ELF varies in composition and thickness with distal progression into the lung. The
ELF of most of the tracheobronchial region may generally be described as consisting of
two layers: an upper mucus layer and a periciliary layer, which surrounds the cilia
(Button etal.. 2012: Widdicombe. 2002: Widdicombe and Widdicombe. 1995: Van As.
1977). The length of motile human cilia is about 7 urn in the distal nasal airways, trachea,
and bronchi and around 5 um in the bronchioles (Yaghi etal.. 2012: Song et al.. 2009:
Clary-Meinesz et al.. 1997: Widdicombe and Widdicombe. 1995). In the healthy lung,
the thickness of the periciliary layer is roughly the length of the cilia (Song et al.. 2009:
Widdicombe and Widdicombe. 1995). This periciliary layer forms a continuous liquid
lining over the tracheobronchial airways; whereas the upper mucus layer is discontinuous
and diminishes or is absent in smaller bronchioles (Widdicombe. 2002: Van As. 1977).
The periciliary layer may be the only ELF layer (i.e., there is little to no overlaying
mucus) in the ciliated airways of infants and healthy adults who are unaffected by
pathology related to disease, infection, or other stimuli (Bhaskar et al.. 1985).
The ELF covering the alveolar surface is considerably thinner than the periciliary layer
found in the tracheobronchial region. The alveolar ELF consists of two layers: an upper
surfactant layer and a subphase fluid (Ng et al.. 2004). Bastacky et al. (1995) conducted a
low-temperature scanning electron microscopy analysis of rapidly frozen samples
(9 animals; 9,339 measurements) of rat lungs inflated to approximately 80% total lung
4-7
-------
capacity. The alveolar ELF was found to be continuous, but of varied depth. Three
distinct ELF areas were described: (1) a thin layer [0.1 urn median depth, geometric
standard deviation (GSD) -2.16]l over relatively flat areas and comprising 80% of the
alveolar surface, (2) a slightly thinner layer (0.08 um, GSD ~1.79) over protruding
features and accounting for 10% of the surface, and (3) a thick layer (0.66 um, GSD
-2.18) occurring at alveolar junctions and accounting for 10% of the surface. Based on
these distributions of thicknesses, 10% of the alveolar region is covered by an ELF layer
of 0.04 um or less. Presuming that these depths would also occur in humans at 80% total
lung capacity and assuming isotropic expansion and contraction, depths should be
expected to be 20-40% greater during normal tidal breathing (rest and light exercise)
when the lung is inflated to between 50-60% total lung capacity averaged across the
respiratory cycle. During tidal breathing, a median ELF depth of 0.12-0.14 um would be
expected over 80% of the alveolar surface with 10% of the alveolar surface having a
median depth of around 0.05 um or less. Considering the entire distribution of depths
during tidal breathing, about 30, 60, and 90% of the alveolar surface would be estimated
to have a lining layer thickness of less than or equal to 0.1, 0.2, and 0.5 um, respectively.
The root mean square distance (d) that NO2 can diffuse in some time (t) is given by the
Einstein-Smoluchowski equation:
d = ^/2Dt
Equation 4-1
where D is the molecular diffusion coefficient of NC>2. A D value for NO2 in water at
25°C of 1.4 x 10~9 m2/sec has been reported and will be used in the calculations (Ford et
al.. 2002). In the lung, the D for NC>2 would be increased by temperature and decreased
by the higher viscosity of the ELF compared to water. The time available for diffusion
can be estimated based on the half-time for reactions between NO2 and reactive
substrates, assuming pseudo first-order kinetics apply. This half-time (T) has the form:
T =
L,i «-ii-i
Equation 4-2
yn^
where ^i l l is the summation of the products of the second-order rate constants (k,)
and substrate concentrations (c,) for the primary reactive substances in the ELF.
1 Although the authors stated that the distributions appeared to be log-normal, they did not report the geometric
standard deviation (GSD) for the three distinct areas they described. The GSD values were calculated from 25, 50,
and 75th percentiles of the distributions.
4-8
-------
Substituting T for t in Equation 4-1 yields:
_ 2Dln(2)
Equation 4-3
and approximates the distance NO2 may diffuse before it chemically reacts with ELF
constituent molecules (e.g., antioxidants, proteins, lipids). A similar approach of
comparing the half-time in Equation 4-2 to the time for diffusion through the ELF or
other phase boundaries such as a membrane bilayer (see Equation 4-1 and solve for t)
was originally applied by Prvor (1992) and later by Ford et al. (2002).
In considering the classes of ELF biomolecules that react with NO2, one may focus on the
water-soluble, small molecular weight antioxidants (e.g., ascorbate, urate, glutathione),
which exist in the ELF in high concentrations and are very reactive toward NC>2 and
consequently have large kid terms. Lipids, on the other hand, would not be expected to
considerably decrease the transit time of NO2 because only those lipids containing fatty
acids with two or more double bonds have significant reactivity towards NO2, and the
lipids in the ELF are highly saturated.
The reaction rate constants of 3.5 x 107 M^sec"1, 2 x 107 M^sec"1, and 2 x 107 M^sec"1
were assumed for the small molecular weight antioxidants ascorbate, urate, and
glutathione, respectively (Ford et al.. 2002). These rates were determined in solution
using the pulse radiolysis fast kinetics technique. The kinetics of ascorbate and urate were
directly monitored, while in the case of glutathione, ABTS2 [2,2'-azino-bis
(3-ethylbenzothiazoline-6-sulfonic acid)] was used to produce the intense chromophore
ABTS* (note, here and elsewhere the superscript 'designates a radical species) from its
reaction with the glutathiyl radical.
Species and anatomical loci must be considered when selecting appropriate
concentrations of reactive ELF biomolecules. Table 4-1 illustrates the small molecular
weight antioxidant composition differences between human and rat bronchoalveolar ELF
and the differences between human nasal and bronchoalveolar ELF (Squadrito etal..
2010; van derVliet etal.. 1999). Predicted by Equation 4-3 and shown in Table 4-1. NO2
is predicted to penetrate 0.2 to 0.6 um into the ELF and would not likely reach airway
tissues in the bronchi or bronchioles. Even extending the time for diffusion to ST, NCh
would only be predicted to penetrate 0.5 to 1.3 um into the ELF, which does not approach
the 5 um depth expected in the ciliated airways. However, minimal NO2 diffusion
through the ELF in the bronchi and bronchioles does not preclude the potential for NC>2
4-9
-------
reactions with cilia or potential importance of reaction products reaching the underlying
tissues in these regions.
Table 4-1 Small molecular weight antioxidant concentrations in epithelial lining
fluid and predicted penetration distances for nitrogen dioxide.
Species — Site
n
/ kici
Ascorbate Urate Glutathione
ELF ELF
Penetration Thickness
Distance (urn) Wm>
Substrate Concentration, ci (uM)a
Human — nasal
Human —
bronchoalveolar
Rat — bronchoalveolar
Rate Constant, ki (M~1
28 ±19 225 ±105 <0.5 5.5 x 103
40 ±18 207 ±167 109 ±64 7.7 x 103
1,004 ±325 81 ±27 43 ±15 3.8 x 104
sec-1)b
0.6 7C
0.5 5-7c'd
<0.1-0.5e
0.2 5-7c'd
<0.1-0.5e
3^xin7 9x1 n? 9 x 1 n?
,\J " I U ^ ^ I U ^^lU
ELF = epithelial lining fluid.
aSubstrate concentrations from van der Vliet et al. (1999) for human and from Squadrito et al. (2010) for rat.
""Reaction rate constants from Ford et al. (2002).
°Based on length of cilia.
dTracheobronchial airways.
eAlveolar airways.
In the alveolar region, the thickness of the ELF is sufficiently thin (<0.2 urn over 60% of
the alveolar surface) for NO2 to diffuse through. There are some important differences
between the ELF of the alveolar region and the ELF of the tracheobronchial airways. In
studies modeling NO2 and ozone (Os) uptake, a first-order rate constant has been
assumed for the alveolar ELF, which is 60-times slower than that of the tracheobronchial
ELF (Miller et al.. 1985; 1982). The slower reaction rate in the alveolar ELF would
increase the estimated potential diffusion distance to nearly 4 urn, well beyond the depth
of the alveolar ELF. Additionally, the presence of DPPC, a principle component of
pulmonary surfactant, has been shown in vitro to reduce the uptake of NCh and Os by
inhibiting their ability to reach and react with the underlying subphase fluid containing
ascorbate, glutathione, and uric acid (Connor etal.. 2004; Connor et al.. 2001). The
physical properties of the interfacial saturated phospholipids may act to reduce the
diffusivity of NC>2. Both the DPPC and the overall slower reaction rate in the alveolar
4-10
-------
ELF would increase diffusive resistance and increase the back diffusion of NO2 from the
surfactant into the gas phase. Nonetheless, the time for NO2 diffusion through a 0.2-um
alveolar ELF is over two orders of magnitude faster than the NCh reaction rate half-time
in the alveolar ELF. Thus, of the inhaled NCh reaching the alveolar region and diffusing
into the ELF, an appreciable amount of NO2 may reasonably be expected to diffuse
through the ELF to reach underlying tissues over much of the alveolar surface. Reaction
rates in these underlying tissues are expected to exceed those in the alveolar and
tracheobronchial ELF and would more rapidly consume NC>2 (Pryor. 1992; Miller et al..
1985).
4.2.2.2 Epithelial Lining Fluid Interactions with Nitrogen Dioxide
Small molecular weight antioxidants vary appreciably across species. For example, due
to the lack of urate oxidase, humans, primates, and select other species have increased
levels of urate. Conversely, rodent concentrations of urate are small compared to humans.
Such differences need to be recognized when considering preferential reactive absorption
substrates and the profile of products formed via reaction with NC>2. Glutathione and
ascorbate are the primary NCh-absorption substrates in rat ELF with near
1:1 stoichiometric yields of NC>2 uptake to nitrite formation, suggesting that one-electron
reduction of NCh is a predominant reaction pathway that also yields the corresponding
organic radical (Postlethwait et al.. 1995).
Beyond cell-specific differential susceptibility and the airway lumen concentration of
NC>2, site-specific injury was proposed to depend on the rate of bioactive reaction product
formation relative to the extent of quenching (detoxification) of these products within the
ELF. Velsor and Postlethwait (1997) investigated the mechanisms of acute cellular injury
from NC>2 exposure. In an in vitro test system using red blood cells, the maximal levels of
membrane oxidation were observed at low antioxidant levels versus null (absent
antioxidants) or high antioxidant levels. Glutathione- and ascorbate-related membrane
oxidation was superoxide- and hydrogen peroxide-dependent, respectively. The authors
proposed that increased absorption of NCh occurred at the higher antioxidant
concentrations, but little secondary oxidation of the membrane occurred because the
reactive species (e.g., superoxide and hydrogen peroxide) generated during absorption
were quenched. A lower rate of NO2 absorption occurred at the low antioxidant
concentrations, but oxidants were not quenched and so were available to interact with the
cell membrane. Further in vitro analyses also suggested that exposure-related responses
may not be strictly linear with respect to the inhaled NC>2 dose (concentration and/or
time) because the dependence of NO2 absorption and biologic target oxidation
demonstrated a bell-shaped function with respect to the initial antioxidant concentration
4-11
-------
(Adgent et al., 2012; Velsor et al., 2003). Because the ELF varies throughout the
respiratory tract, the heterogeneous distribution of epithelial injury observed following
NO2 exposures may be explained, in part, by the ELF-dependent effects on local NCh
uptake and product formation. However, it should be noted that while these
dose-response relationships have been documented in vitro, in vivo validation has not yet
been accomplished due to the complexities in reproducibly modulating in situ ELF
compositions. Importantly, such in vitro results are difficult to directly extrapolate to the
in vivo situation, as precise rates of NO2 uptake, and thus product formation, are a
function of many factors including gas-phase NC>2 concentration, aqueous substrate
concentrations, surface area, gas flow, and pH of the ELF (Adgent et al.. 2012; Bidani
and Postlethwait 1998). However, an in vivo study of healthy male albino mice (5 weeks
old) suggested that a low dose of ascorbate (25 mg/kg) may exacerbate inflammatory
responses in terminal bronchial tissues following NC>2 exposure (20,000 ppb; 4 hour/day,
10 days); whereas at a higher dose of ascorbate (100 mg/kg), NCh-exposed mice tissues
were similar to tissues from filter air-exposed controls (Zhang et al.. 201 Ob). These
in vivo responses seem parallel to those observed in vitro.
Antioxidant levels also vary spatially between lung regions and temporally with NC>2
exposure. While in vitro studies have clearly illustrated the role of antioxidants in
mediating NC>2 uptake and membrane oxidation, the temporal dynamics of biological
responses to NCh that occur in vivo are far more complex. Given the rapid reactions of
inhaled NO2 with various biological substrates, the short half-life of some primary and
secondary reaction products as well as the continuous turnover of the ELF, specific
chemical species do not likely persist at any given anatomic locale for any appreciable
time. Kelly et al. (1996a) examined the effect of a 4-hour NCh (2,000 ppb) exposure on
antioxidant levels in bronchial lavage (BL) fluid and BAL fluid of 44 healthy
nonsmoking adults (19-45 years, median 24 years). The baseline concentrations of urate
and ascorbate were strongly correlated between the BL fluid and BAL fluid within
individuals (r = 0.88,/> < 0.001; r = 0.78,/> = 0.001; respectively); whereas the
concentrations of glutathione in the BL fluid and BAL fluid were not correlated. At
1.5 hours after the NCh exposure, urate and ascorbate were significantly reduced in both
lavage fractions, while glutathione levels were significantly increased but only in BL
fluid. By 6 hours post-exposure, ascorbate levels had returned to baseline in both lavage
fractions, but urate had become significantly increased in both lavage fractions and
glutathione levels remained elevated in BL fluid. By 24 hours post-exposure, all
antioxidant levels had returned to baseline. The levels of glutathione in BAL fluid did not
change from baseline at any time point in response to NC>2 exposure.
The depletion of urate and ascorbate, but not glutathione, has also been observed with
ex vivo exposure of human BAL fluid to NC>2. Kelly et al. (1996b) collected BAL fluid
4-12
-------
from male lung cancer patients (n = 16) and exposed the BAL fluid ex vivo at 37°C to
NO2 (50 to 2,000 ppb; 4 hours) or O3 (50 to 1,000 ppb; 4 hours). Kelly and Tetlev (1997)
also collected BAL fluid from lung cancer patients (n = 12; 54 ± 16 years) and exposed
the BAL fluid ex vivo to NO2 (50 to 1,000 ppb; 4 hours). Both studies found that NO2
depletes urate and ascorbate, but not glutathione, from BAL fluid. Kelly et al. (1996b)
noted a differential consumption of the antioxidants, with urate loss being greater than
that of ascorbate, which was lost at a much greater rate than glutathione. Kelly and Tetley
(1997) found that the rates of urate and ascorbate consumption were correlated with their
initial concentrations in the BAL fluid, such that higher initial antioxidant concentrations
were associated with a greater rate of antioxidant depletion. Illustrating the complex
interaction of antioxidants, these studies also suggest that glutathione oxidized by NO2
may be again reduced by urate and/or ascorbate.
Human and animal results stemming from samples obtained after exposure should be
viewed with appropriate caution. As detailed below, secondary reactions within the ELF,
sample handling, and importantly, the temporal sequence of exposure relative to sample
acquisition may all confound data interpretation. Because the ELF is a dynamic
compartment, samples obtained after exposure (>30 minutes) may not reflect biochemical
conditions present during exposure. This is a critical point, because while there is some
value in quantifying the net short-term effects on ELF composition due to exposure, the
biological consequences of exposure are largely a function of the ELF conditions during
exposure, which initiate a cascade of events leading to alterations in cell signaling, cell
injury, inflammation, and so forth. Thus, measurements of ELF components should be
interpreted in the context of ELF turnover time, clearance of "stable" reaction products,
and species generated/regenerated as a consequence of secondary redox reactions.
Reported measurements may reflect net effects on individual antioxidants but lend
limited insights into the initial reactions of NO2 within the ELF, and by extension, into
what bioactive products may be formed and how differences in ELF constituent profiles
govern biological outcomes. A clear example is evident in the work of Ford et al. (2002).
who characterized the reaction of the GSH radical (GS*) with urate (UH2 ) at a pH (6.0)
slightly below the recognized ELF pH (~6.8 to 7.0). NO2 more readily reacts with
glutathione than urate, producing GS* and NO2~. However, the subsequent reaction
GS* + UH2 -»GSH + UH* has a rate constant of ~3 x 1071VT1 sec"1, which could
translate to an initial NO2 reaction with glutathione followed by reduction of the thiyl
radical by urate. This could result in an apparent, but potentially inaccurate, conclusion of
direct loss of urate during subsequent analyses. In addition, some reports have suggested
observations that include significant levels of the ascorbate oxidation product
dehydroascorbate (DHA). As with the example of secondary urate oxidation, such
observations need to be evaluated with caution as the half-life of DHA under biological
conditions is very short (minutes; the ascorbyl radical dismutation produces reduced
4-13
-------
ascorbate and DHA; and DHA spontaneously decomposes to its keto acid). Furthermore,
because high redox couples are maintained in the ELF, and the ELF is constantly turning
over due to secretion and mucociliary clearance, it is unlikely that any appreciable
accumulation of DF£A would occur. Therefore, care must be taken to avoid introducing
methodological artifacts (e.g., ascorbate oxidation during sample acquisition, handling,
and/or storage) that could significantly confound data interpretation. Consequently, an
understanding of the precise and preferential substrates is needed to discern the genesis of
species differences and the products formed that account for NO2 exposure-related
cellular perturbations.
Given the above considerations, variability in antioxidant concentrations and reactions
among species may affect NO2 dose and health outcomes. Guinea pigs and mice have a
lower basal activity of glutathione transferase and glutathione peroxidase and lower
a-tocopherol levels in the lung compared to rats dchinose et al.. 1988; Sagai et al.. 1987).
Human nasal lavage fluid has a high proportion of urate and low levels of ascorbate;
whereas mice, rats, and guinea pigs have high levels of ascorbate and undetectable levels
of urate. Glutathione is not detected in the nasal lavage fluid of most of these species,
except monkeys. Guinea pigs and rats have a higher antioxidant-to-protein ratio in nasal
lavage fluid and BAL fluid than humans (Hatch. 1992). The BAL fluid profile differs
from that of the nasal lavage fluid. Humans have a higher proportion of glutathione and
less ascorbate in their BAL fluid compared to guinea pigs and rats (Slade et al.. 1993;
Hatch. 1992). Rats have the highest antioxidant-to-protein mass ratio in their BAL fluid
(Slade et al.. 1993). Antioxidant defenses also vary with age (Servais et al.. 2005) and
exposure history (Duanetal.. 1996). In the case of another reactive gas, Os, some studies
have found that differences in antioxidant levels among species and lung regions did not
appear to be the primary factor affecting Os-induced tissue injury (Duan et al.. 1996;
1993). However, close correlations have been observed between site-specific Os dose and
the degree of epithelial injury, as well as the depletion of reduced glutathione in monkeys
(Plopper et al.. 1998). For both NCh and Os, differences in reactive substrates among
species and regions of the respiratory tract are recognized, but the importance of these
differences in relation to tissue injury is not fully understood.
4.2.2.3 Regional and Total Respiratory Absorption of Nitrogen Dioxide
Very limited work related to the quantification of NC>2 uptake has been published since
the 1993 AQCD (U.S. EPA. 1993a) or the subsequent 2008 ISA for Oxides of Nitrogen
(U.S. EPA. 2008c). Consequently, only an abbreviated discussion of this is included.
4-14
-------
4.2.2.3.1 Experimental Studies of Nitrogen Dioxide Uptake
Upper Respiratory Tract Absorption
The nasal uptake of NO2 has been experimentally measured in dogs, rabbits, and rats
under conditions of unidirectional flow. Yokoyama (1968) reported 42.1 ± 14.9%
[mean ± standard deviation (SD)] uptake of NO2 in the isolated nasal passages of two
dogs (3.5 L/min) and three rabbits (0.75 L/min) exposed to 4,000 and 41,000 ppb NO2.
Uptake did not appear to depend on the exposure concentration and was relatively
constant over a 10- to 15-minute period. Cavanagh and Morris (1987) measured 28 and
25% uptake of NO2 (40,400 ppb) in the noses of four naive and four previously exposed
rats (0.10 L/min; 4-hours; 40,400 ppb), respectively, and uptake was constant over the
24-minute period it was monitored.
Kleinman and Mautz (1991) measured the penetration of NCh through the upper airways
during inhalation in six tracheostomized dogs exposed to 1,000 or 5,000 ppb NC>2.
Uptake in the nasal passages was significantly greater at 1,000 ppb than at 5,000 ppb,
although the magnitude of this difference was not reported. The mean uptake of NO2
(1,000 ppb) in the nasal passages decreased from 80 to 70% as the ventilation rate
increased from about 3 to 7 L/min. During oral breathing, uptake was not dependent on
concentration. The mean oral uptake of NCh (1,000 and 5,000 ppb) decreased from 60 to
30% as the ventilation rate increased from 3 to 7 L/min. Although nasal uptake tended to
be greater than oral uptake, the difference was not statistically significant. The tendency
for greater nasal than oral uptake on NO2 is consistent with that observed for Os as
described in Chapter 5 of the 2013 ISA for Ozone (U.S. EPA. 2013e).
Overall, NO2 fractional absorption (uptake efficiency) in the upper respiratory tract is
greater in the nasal passage than in the oral passage and decreases with increasing
ventilation rates. As a result, a greater proportion of inhaled NO2 is delivered to the lower
respiratory tract at higher ventilation rates associated with exercise. In humans, exercise
causes a shift in the breathing pattern from nasal to oronasal relative to rest. Because the
nasal passages scrub gas-phase NO2 more efficiently than the mouth and because uptake
efficiency decreases with increasing flow, exercise delivers a disproportionately greater
quantity of the inhaled mass to the lower respiratory tract, where the NO2 is readily
absorbed.
Additionally, children tend to have a greater oral breathing contribution than adults at rest
and during exercise (Bennett et al., 2008; Becquemin et al., 1999). Chadha et al. (1987)
found that the majority (11 of 12) of patients with asthma or allergic rhinitis also breathe
oronasally at rest. Thus, compared to healthy adults, children and individuals with asthma
might be expected to have greater NO2 penetration into the lower respiratory tract.
4-15
-------
Furthermore, normalized to body mass, median daily ventilation rates (m3/kg per day)
decrease over the course of life (Brochu etal.. 2011). This decrease in ventilation relative
to body mass is rapid and nearly linear from infancy through early adulthood. Relative to
normal-weight adults (25-45 years of age), ventilation rates normalized to body mass are
increased 1.5-times in normal-weight children (7-10 years of age) and doubled in
normal-weight infants (0.22-0.5 years of age). Relative to their body mass, children, and
especially infants, respire greater amounts of air and associated pollutants than adults and
have a greater portion of respired pollutants reaching the lower respiratory tract than
adults.
Lower Respiratory Tract Absorption
Postlethwait and Mustafa (1989) investigated the effect of exposure concentration and
breathing frequency on the uptake of NC>2 in isolated perfused rat lungs. To evaluate the
effect of exposure concentration, the lungs were exposed to NO2 (4,000 to 20,000 ppb)
while ventilated at 50 breaths/min with a tidal volume (Vr) of 2.0 mL. To examine the
effect of breathing frequency, the lungs were exposed to NO2 (5,000 ppb) while
ventilated at 30-90 breaths/min with a VT of 1.5 mL. All exposures were for 90 minutes.
The uptake of NO2 ranged from 59 to 72% with an average of 65% and was not affected
by exposure concentration or breathing frequency. A combined regression analysis
showed a linear relationship between NO2 dose to the lungs and total inhaled dose.
Illustrating variability in NC>2 uptake measurements, Postlethwait and Mustafa (1989)
observed 59% NO2 uptake in lungs ventilated at 30 breaths/min with a VT of 1.5 mL;
whereas Postlethwait and Mustafa (1981) measured 35% NO2 uptake for the same
breathing condition. In another study, 73% uptake of NC>2 was reported for rat lungs
ventilated at 50 breaths/min with a VT of 2.3 mL (Postlethwait et al.. 1992). It should be
noted that typical breathing frequencies are around 80, 100, and 160 breaths/min for rats
during sleep, rest, and light exercise, respectively (de Winter-Sorkina and Cassee. 2002).
Hence, the breathing frequencies at which NO2 uptake has been measured are lower than
for rats breathing normally. Furthermore, one must consider the potential impacts of the
methods used to measure NO2 uptake (mass balance; wet chemical versus automated
analyzer which may or may not include a dilution component due to the sampling rate)
and the lack of perfusion of the bronchial circulation in isolated rat lungs (Postlethwait et
al.. 1990). In addition to measuring uptake in the upper respiratory tract, Kleinman and
Mautz (1991) also measured NO2 uptake in the lower respiratory tract of tracheostomized
dogs. In general, about 90% NO2 uptake in the lung was independent of ventilation rates
from 3 to 16 L/min.
4-16
-------
Total Respiratory Tract Absorption
Bauer et al. (1986) measured the uptake of NO2 (300 ppb) in 15 adults with asthma
exposed for 30 minutes (20 minutes at rest, then 10 minutes exercising on a bicycle
ergometer) via a mouthpiece during rest and exercise. There was a statistically significant
increase in uptake from 72% during rest to 87% during exercise. The minute ventilation
also increased from 8.1 L/min during rest to 30.4 L/min during exercise. Hence, exercise
increased the NO2 dose rate by 4.5-times in these subjects. In an earlier study by Wagner
(1970). seven healthy adults inhaled a NO2/NO mixture containing 290 to 7,200 ppb NO2
for brief (but unspecified) periods. The average NO2 uptake during 4,100-ppb and
7,200-ppb exposures was 82% during normal respiration (Vr, 0.4 L) and 92% during
maximal respiration (Vr, 2 to 4 L). Kleinman and Mautz (1991) also measured the total
respiratory tract uptake of NO2 (5,000 ppb) in nontracheostomized female beagle dogs
standing at rest or exercising on a treadmill. The dogs breathed through a small face
mask. Total respiratory tract uptake of NO2 was 78% during rest and increased to 94%
during exercise. This increase in uptake may largely be due to the increase in VT from
0.18 L during rest to 0.27 L during exercise. Coupled with an increase in minute
ventilation from 3.8 L/min during rest to 10.5 L/min during exercise, the dose rate of NO2
was 3.3-times greater for the dogs during exercise than rest.
4.2.2.3.2 Dosimetry Models of Nitrogen Dioxide Uptake
Few theoretical studies have investigated NO2 dosimetry. The original seminal dosimetry
model of Miller etal. (1982) was developed before much of the above information
regarding NO2 reaction/diffusion within the ELF had been obtained. In this model, there
was a strong distinction between uptake and dose. Uptake referred to the amount of NO2
being removed from gas phase per lung surface area (ug/cm2); whereas dose referred to
the amount of NO2 per lung surface area (ug/cm2) that diffused through the ELF and
reached the underlying tissues.
Miller etal. (1982). and subsequently Overton (1984). did not attempt to predict the
amount of reactants in the ELF or the transport of reaction products to the tissues. They
assumed that reactions of NO2 with constituents in the ELF were protective in that these
reactions reduced the flux of NO2 to the tissues. Others have postulated that NO2 reaction
products formed in the ELF, rather than NO2 itself, could mediate responses (Velsor and
Postlethwait. 1997; Postlethwait and Bidani. 1994; Overton. 1984). Overall, these
modeling studies predict that the net NO2 uptake (NO2 flux to air-liquid interface) is
relatively constant from the trachea to the terminal bronchioles and then rapidly decreases
in the pulmonary region. The pattern of net NO2 uptake rate is expected to be similar
among species and unaffected by age in humans. However, the NO2 uptake per unit
4-17
-------
surface area may be several times higher in infants compared to adults, because children
under age 5 have a much smaller surface area in the extrathoracic (nasal) and alveolar
regions (Sarangapani et al.. 2003).
The predicted tissue dose and dose rate of NO2 (NC>2 flux to liquid-tissue interface) are
low in the trachea, increase to a maximum in the terminal bronchioles and the first
generation of the pulmonary region, and then decrease rapidly with distal progression.
The site of maximal NC>2 tissue dose is predicted to be fairly similar among species,
ranging from the first generation of respiratory bronchioles in humans to the alveolar
ducts in rats. However, estimates of NO2 penetration in Table 4-1 showed that NO2 is not
expected to go deeper than 0.2 to 0.6 urn into the ELF of the ciliated airways before
reacting with substrates. The production of toxic NC>2 reaction products in the ELF and
the movement of the reaction products to the tissues have not been modeled.
Contrary to what in vitro studies have shown (Velsor and Postlethwait 1997). modeling
studies have generally considered NO2 reactions in the ELF to be protective. The
complex interactions among antioxidants, spatial differences in antioxidants across
respiratory tract regions, temporal changes in ELF constituent levels in response to NC>2
exposure, and species differences in antioxidant defenses need to be considered in the
next generation of dosimetric models. Current NC>2 dosimetry models are inadequate to
put response data collected from animals and humans on a comparative footing with each
other and with exposure conditions in epidemiologic studies. Total dose or liquid dose of
NC>2 could be used as a first approximation for inter-species dosimetric comparisons
using currently available NC>2 models.
As stated above, the total dose or uptake (fig per cm2 surface area) of NO2 is predicted to
be relatively constant across the tracheobronchial airways with a rapid decrease in dose
with progression into the gas exchange region (Miller et al.. 1982). The model used by
Miller et al. (1982) for NC>2 was generally the same as that subsequently used by Miller et
al. (1988) for O3. Miller etal. (1988) predicted that the total dose of O3 is relatively
similar among several mammalian species (namely, the rabbit, guinea pig, rat, and
human). The total dose of NO2 would also be expected to be relatively similar among
these mammalian species. Although it may not be strictly appropriate to apply identical
reaction rates for each of these species, varying the reaction rate from zero to that of Os
increased the predicted total dose of NO2 by less than 5 times in the trachea and bronchi.
This is small relative to the 400-times decrease in total dose from the first generation of
respiratory bronchioles to the alveolar sacs (Miller etal.. 1982).
Asgharian et al. (2011) recently developed a model for soluble and reactive gas uptake
that applied many of the basic concepts described by Miller et al. (1985). Unlike Miller et
al. (1985). who separately considered liquid and tissue layers, Asgharian et al. (2011)
4-18
-------
lumped the liquid layer lining the airways and the tissue layer together with the same
diffusion and reaction rates. The model predicted that formaldehyde could penetrate to a
maximum of 200 urn in tracheal tissue during inhalation before being removed by
reactions. Because predictions were for a single breath, it is possible that deeper tissues
may be reached during continuous breathing. Applying the model to experimental Os
data, Asgharian et al. (2011) estimated a first-order reaction rate of
105 sec"1 (i.e., half-time of only 7 usec). By comparison, the rate of 0.018 sec"1
(i.e., half-time of 39 sec) was used for formaldehyde. Lumping the liquid and tissue
layers may be appropriate for the relatively slow-reacting formaldehyde, but it is perhaps
less so for Os and NO2, which are expected to be removed by reactions within the liquid
layer of ciliated airways (Table 4-1). For the rapidly reacting gases Os and NO2, a
distinction between liquid and tissue compartments may be mechanistically important to
discern whether the gas itself or its reaction products are associated with health outcomes.
Existing dosimetric models can predict the total dose per surface area of distinct areas of
the lungs (e.g., individual generations of the tracheobronchial airways and alveolar
region). This total dose appears to be very similar among several mammalian species.
Similarly, the site of maximal NC>2 tissue dose, near the beginning of the gas exchange
region, is also predicted to be fairly similar among species. However, differences in
potential NO2 reactive substrates and reaction products among species have not been
considered in modeling efforts. Thus, despite the predicted similarities in total NO2 dose
and site of maximal tissue dose, there is uncertainty related to inter-species differences in
concentrations of reactive substrates and reaction products formed within the ELF and
tissues. The importance of specific reaction products in mediating health effects in
different species is similarly unclear. With regard to humans, individuals with asthma are
more likely to experience health effects from ambient NO2 exposures than healthy
individuals (Section 7.3.1). Specific aspects of asthma pathology that may affect NO2
uptake and disposition and that may be included in dosimetric models have not been
identified. Furthermore, most models have focused on the lungs and have not considered
the inter-species differences in the dose to nasal passages nor the potential importance of
neural or other pathways in affecting health outcomes. Although total dose in the
tracheobronchial airways and tissue dose in the alveolar region can be predicted,
modeling efforts do not sufficiently link these endpoints to subsequent downstream
events.
4-19
-------
4.2.2.4 Endogenous Generation, Metabolism, Distribution, and Elimination of
Nitrogen Dioxide
Along with carbon monoxide (CO), NO2 is a criteria pollutant believed to be produced
endogenously in the lung and other tissues. Evidence in support of a claim for
endogenously produced Os [e.g., Babior et al. (2003)] has received much criticism (Pryor
et al.. 2006; Kettle et al.. 2004; Sies. 2004; Smith. 2004) and is here considered
controversial. A useful discussion of the issues can be found in Drahl (2009).
This endogenous production and function of NO2 may have important implications in
interpreting health effects studies. NO2 can be produced endogenously by various
processes, including the acidification of nitrite
(2 H+ + 2 NO2" -» 2 HNO2 -»H2O + N2O3 -»NO + NO2 + H2O) (as can transpire in
phagolysosomes), the decomposition of peroxynitrite and/or the nitrosoperoxylcarbonate
anion (ONOCT + CO2 -» ONOOCO2" -» CO3*~ + NO2), and the action of peroxidases
when using nitrite and hydrogen peroxide (H2O2) as substrates. Nitrated proteins form
when tyrosine residues are first oxidized to atyrosyl radical intermediately followed by
radical-radical addition of NO2 to produce 3-nitrotyrosine. NO2 is the terminal nitrating
agent, and the presence of nitrated proteins provides solid evidence for the endogenous
production of NO2 per se. Endogenous NO2 is expected to increase with dietary
consumption of nitrite and nitrate (the latter of which occurs in substantial concentrations
in some leafy vegetables like spinach) as well as during immune responses and
inflammation. There is no known antioxidant enzymatic process for the decomposition of
NO2, probably because NO2 undergoes spontaneous reactions with small molecular
weight antioxidants, such as glutathione and ascorbate, which result in formation of
nitrite and antioxidant radicals. These reactions are so fast that they only allow NO2 to
diffuse small distances in the submicrometer range before reacting (Table 4-1) NO2 is
slightly hydrophobic (Squadrito and Postlethwait. 2009) and faces no significant physical
barriers to prevent it from readily traversing biological membranes. But due to its high
reactivity, NO2 is unlikely to become systemically distributed; therefore, its endogenous,
steady-state levels in distant tissues are unlikely to be affected by inhaled NO2.
Regarding the lung, understanding the balance between endogenous products and those
derived from inhaled ambient NO2 is a complex and challenging issue. Because inhaled
NO2 predominantly undergoes univalent reduction to nitrite during reactive absorption,
changes in nitrite concentrations can be used as a surrogate for initial considerations of
how inhaled NO2 compares with that produced endogenously. For example, rat lung ELF
contains low (uM to nM) levels of nitrite, with nitrate being substantially more prevalent.
Due to salivary and gut microflora nitrate reductase activity and to reactions of nitrite,
especially with heme proteins to yield nitrate, there is a constant cyclic flux between
4-20
-------
nitrite and nitrate, with nitrate being the primary excretion product in urine. In a rat, there
would be a net accumulation of approximately 0.3 umol of nitrite, assuming for
simplicity a gas phase concentration of 200 ppb NCh, a minute ventilation of
150 mL/min, an exposure time of 4 hours, quantitative conversion of NCh to nitrite,
100% uptake efficiency, an ELF volume of 150 uL, and no ELF clearance [even though
nitrite has been shown to diffuse out of the ELF quickly (Postlethwait and Bidani. 1989)].
If the NCh-derived nitrite were evenly distributed throughout the ELF pool, this would
equate to an additional 2 mM concentration of nitrite. However, in vitro studies using
isolated lungs have not reported increases of this magnitude consequent to
10,000-20,000-ppb NCh exposures, well above ambient concentrations, demonstrating
that the ELF is a dynamic compartment and that small molecular weight reaction
products (though charged) move readily from the respiratory tract surface to the vascular
space.
Both nitrite and nitrate levels are very diet dependent, and diet represents the primary
source for both ions. Although environmental exposures at current ambient NCh
concentrations would likely have a minimal effect on the overall balance of nitrite and
nitrate outside the respiratory tract, how inhaled NCh compares with endogenous
production rates or amounts within the respiratory tract remains essentially unknown.
However, the uptake of inhaled NCh may potentially increase levels of nitrite and/or
other reaction products beyond levels endogenously occurring in the respiratory tract.
4.2.2.5 Metabolism, Distribution, and Elimination of Products Derived from Inhaled
Nitrogen Dioxide
As stated earlier, NCh absorption may generate some HNCh, which subsequently
dissociates to H+ and nitrite. Nitrite enters the underlying epithelial cells and
subsequently the blood. In the presence of red blood cells and/or heme proteins, nitrite is
oxidized to nitrate (Postlethwait and Mustafa. 1981). Nitrate is the primary stable oxide
of nitrogen product, and it is subsequently excreted in the urine. Nitrate can also be
converted to nitrite by bacterial reduction in saliva, the gastrointestinal tract, and the
urinary bladder.
There has been concern that inhaled NCh may lead to the production of N-nitrosamines,
many of which are carcinogenic because inhaled NCh can contribute to blood levels of
nitrite and nitrate. Nitrite has been found to react with secondary amines to form
N-nitrosamines. However, nitrosamines are not detected in tissues of animals exposed by
inhalation to NCh unless precursors to nitrosamines and/or inhibitors of nitrosamine
metabolism are co-administered. Rubenchik et al. (1995) could not detect
N-nitrosodimethylamine (NDMA) in tissues of mice exposed to 4,000 to 4,500 ppb NCh
4-21
-------
for 1 hour. However, NDMA was found in tissues when mice were simultaneously given
oral doses of amidopyrine and 4-methylpyrazole, an inhibitor of NDMA metabolism.
Van Stee etal. (1983) reported N-nitrosomorpholine (NMOR) production in mice
gavaged with 1 g of morpholine/kg body weight per day and then exposed (5-6 hours
daily for 5 days) to 16,500-20,500 ppb NCh. NMOR is a nitrosamine and potent animal
carcinogen. The single site containing the greatest amount of NMOR was the
gastrointestinal tract, as would be expected due to the pH-dependent facilitation of
N-nitrosation chemistry. Later, Van Stee et al. (1995) exposed mice to approximately
20,000 ppb 15NO2 and 1 g/kg morpholine simultaneously. N-nitrosomorpholine was
found in the body of the exposed mice. Of the NMOR in the body, 98.4% was labeled
with 15N which was derived from the inhaled 15NO2, and 1.6% was derived presumably
from endogenous sources.
Inhaled NO2 may also be involved in the production of mutagenic (and carcinogenic)
nitro derivatives of other co-exposed compounds, such as polycyclic aromatic
hydrocarbons (PAHs), via nitration reactions. Miyanishi et al. (1996) co-exposed rats,
mice, guinea pigs, and hamsters to 20,000 ppb NO2 and various PAHs (pyrene,
fluoranthene, fluorene, anthracene, or chrysene). Nitro derivatives of these PAHs, which
were found to be highly mutagenic in the AmesAS'. typhimurium assay, were excreted in
the urine of these animals. Specifically, the nitrated metabolites of pyrene
(l-nitro-6/8-hydroxypyrene and l-nitro-3-hydroxypyrene) was detected in the urine.
Further studies indicated that these metabolites are nitrated by an ionic reaction in vivo
after the hydroxylation of pyrene in the liver.
Endogenous NO2 production and the cyclic inter-conversion of nitrite and nitrate may
provide the precursors that drive formation of nitrosamine and other nitro derivatives.
However, because ambient NO2 contributes only modest amounts of nitrite/nitrate
relative to dietary intake, any substantial contribution to systemic formation of
nitrosamines and other nitro derivatives is not likely. The relative importance of inhaled
NO2 in formation of N-nitrosamines or other nitro derivatives has yet to be demonstrated.
4.2.3 Dosimetry of Nitric Oxide
NO occurs within the respiratory tract gas phase due to the following: (1) inhalation of
ambient NO and (2) off-gassing from its endogenous production within pulmonary
tissues, airspace surface inflammatory cells, and blood. The net uptake of NO within the
gas exchange regions depends on the balance between the intra-pulmonary gas phase
concentration (discussed below) and the inhaled ambient concentration.
4-22
-------
While NO exists as a radical species, it is much less reactive than many other radical
species. However, it selectively participates in radical-radical reactions such as with
superoxide radical anions [62"~, which produces peroxynitrite (ONOO )], thiyl radicals
[e.g., cysteine (Cys*), glutathione (GS*), which produce S-nitrosothiols (RSNO)], and
organic peroxyl radicals (Madej et al., 2008; Goldstein et al., 2004). In addition, NO
reacts with heme-containing proteins such as hemoglobin (Pacher et al.. 2007). Although
the radical-based reactions generally occur at near diffusion-controlled rates, the
prevalence of non-NO radical species at any given time is low. Thus, in terms of the
overall uptake and tissue diffusion of NO within the lung, interception due to reactions is
not expected to consume appreciable amounts of the total NO involved in mass transfer
from the alveolar to the vascular space. Inhaled NO uptake occurs against the background
of endogenous NO production, which is derived primarily from the catalytic activities of
the several isoforms of nitric oxide synthase [NOS; (Forstermann and Sessa, 2012)1.
Estimates of nitrite and/or nitrate stemming from NO production via NOS suggest that
endogenous NO production, even during inflammatory states, is at best modest compared
to dietary intake; although, under specific conditions, plasma levels have been shown to
transiently increase due to nondietary, endogenous biological activities. Additional
endogenously generated NO may also occur from the acidification of nitrite in the
presence of electron donors, such as within phagolysosomes, by dissociation of RSNO,
and by complex interactions within red blood cells that likely lead to the release of NO
(Weitzberg et al.. 2010). In combination, these processes result in the appearance of NO
within the intra-pulmonary gas phase, the concentration of which can be measured in
expired breath (eNO).
Reported eNO concentrations from the lower respiratory tract span a broad range (~5 to
>300 ppb), with nasal/sinus concentrations generally accepted as being greater than that
measured from the lower respiratory tract [e.g., See and Christiani (2013); Alexanderson
etal. (2012); Gelb et al. (2012); Nodaetal. (2012); Taylor (2012); Bautistaetal. (2011);
Linhares et al. (2011); Olin et al. (1998)1. Levels of eNO are affected by a variety of
factors including disease state, diet, sex (or height), species, smoking history, and
environmental exposures. Although eNO from the lower respiratory tract is increased by
asthma, this is not the case for nasal NO (ATS/ERS. 2005).
For the general U.S. population, results of the 2007-2011 National Health and Nutrition
Examination Survey show a geometric mean eNO of 9.7 ppb in children (n = 1,855;
6-11 years of age; 10% with current asthma) and 13.3 ppb in teenagers and adults
[n = 11,420; 12-80 years of age; 8% with current asthma; (See and Christiani. 2013)]. In
healthy, never-smokers [558 males (M), 573 females (F); 25-75 years of age], Olin et al.
(2007) reported a geometric mean eNO of 16.6 ppb (95% reference interval, 6 to 47 ppb).
The eNO levels increased with age and height of the individuals but did not depend on
4-23
-------
sex. In healthy children (23 M, 28 F; 1-5 years of age), a geometric mean eNO of 7 ppb
(95% CI: 3,12) has been reported (van der Heijden et al.. 2014). The eNO levels in these
children were unrelated to age, height, weight, or sex. These eNO levels correspond to
NO output rates of about 40-50 nL/min from the lower respiratory tract of healthy adults
and about 20-30 nL/min for healthy children.
Kharitonov et al. (2005) reported nasal NO concentrations of 750 ppb (95% CI: 700, 810)
in children [n = 20; 10 ± 3 (SD) years] and 900 ppb (95% CI: 870, 930) in adults (n = 29;
38 ± 11 years of age). Another study of healthy adults (n = 10; 18-35 years of age) found
a nasal NO concentration of 670 ppb. Higher NO concentrations (9,100 ± 3,800 ppb;
n = 5) have been reported for the paranasal sinuses of healthy adults (Lundberg et al.,
1995). Asthma and current rhinitis do not appear to affect nasal NO concentrations
(Alexanderson et al.. 2012; Kharitonov et al.. 2005). Nasal NO is reduced by exercise
(ATS/ERS. 2005). The nasal NO concentrations described above correspond to NO
output rates of about 300 nL/min for the nasal airways of adults with or without asthma
and 230 nL/min for children with or without asthma. Nasal NO output rates of healthy
primates are in the range of 200 to 450 nL/min (ATS/ERS. 2005). With a NO output of
730 nL/min, a large contribution to nasal NO appears to derive from the paranasal
sinuses. Based on these NO output rates, the nasal passages may contribute, on average,
roughly 15-20 ppb NO to the lower respiratory tract during rest.
The other primary approach to noninvasive assessment of the respiratory tract surface is
exhaled breath condensate (EEC), which captures aerosolized materials contained in
exhaled air, including those directly related to reactive nitrogen chemistry (e.g., nitrite,
nitrate, 3-nitrotyrosine). Unfortunately, measurements of eNO and EEC (which rely on
relatively new analytical methods) do not necessarily produce comparable results (Rava
etal.. 2012; Dressel etal.. 2010; Malinovschi et al.. 2009; Cardinale et al.. 2007; Vints et
al.. 2005; Chambers and Ayres. 2001; Olin etal.. 2001; Zetterquist et al.. 1999; Olin et
al.. 1998; Jilmaetal.. 1996). Given the endogenous production of NO and the lack of a
correlation between the two measurements, neither eNO nor EEC can be employed as a
metric of exposure history with any significant degree of specificity for inhaled ambient
NO.
The absorption of inhaled NO proceeds similarly to oxygen and CO. In a study of seven
healthy adults, Wagner (1970) observed an average NO (5,000 ppb) uptake of 88%
during normal respiration (Vr, 0.4 L) and 92% during maximal respiration (Vr, 2 to 4 L).
Because blood acts as a near "infinite" sink for NO, it has been proposed as an alternative
to CO for measuring pulmonary diffusing capacity [e.g., (Chakraborty et al. (2004);
Heller etal. (2004))]. NO absorption follows Henry's law for dissolution into the aqueous
phase, followed by diffusion into the vascular space where it interacts with red blood cell
4-24
-------
hemoglobin to ultimately form nitrate. Thus, due to its chemical conversion, NO net flux
from alveolar gas phase to the blood occurs when the alveolar concentration exceeds that
found in tissue and/or blood. Mass transfer resistances may be encountered (Borland et
al.. 2010; Chakraborty et al.. 2004). but their combined effects are likely small due to the
low (ppb) concentrations of NO.
The formation of RSNO within the ELF may contribute to the overall epithelial cell
uptake via an L-type amino acid transporter [LAT (Toroket al.. 2012; Brahmajothi et al..
2010)1. An in vitro study by Brahmajothi et al. (2010) showed that pre-incubation of
cultured alveolar epithelial cells with L-cysteine increased intracellular RSNO
concentrations by 3-times compared with diffusive transport. This increase involved
transport via the LAT. LAT transport was further augmented by addition of glutathione
and was independent of sodium transport. The authors concluded that NO gas uptake by
alveolar epithelium occurred predominantly by forming extracellular
S-nitroso-L-cysteine, which was then transported by LAT rather than by diffusion.
Subsequently, Toroket al. (2012) also showed that LAT transport exceeded diffusive
transport in isolated mice lungs. However, the precise extent of contribution of LAT
transport remains unclear because formation of RSNO requires several steps due to the
slow direct reactivity of NO with reduced thiols. In vivo, the time for these reactions may
exceed the time for diffusion into and through alveolar epithelial cells. Furthermore,
because blood acts as a sink for NO (i.e., a near zero boundary condition), lower
intracellular concentrations of NO would occur in vivo compared to the nonzero
boundary conditions in cell cultures and isolated lungs (Asgharian et al.. 2011). While
diffusive transport of NO is known and relatively well characterized, the importance of
LAT transport in vivo has not been determined.
Ambient NO levels are likely similar to those endogenously occurring within the lung
airspaces, except during morning commutes or near major roadways where they may
possibly exceed endogenous levels. It is not known whether periods of high ambient NO
exposure could alter endogenous NO production within the respiratory tract or alter
pathways affected by endogenous NO. It is important to note that in the clinical setting,
therapeutic administration is a very different situation wherein > 10,000 ppb NO may be
administered continuously for prolonged periods.
4.2.4 Summary of Dosimetry
The uptake of inhaled NO2 in the respiratory tract is governed by "reactive absorption,"
which involves chemical reactions with antioxidants, unsaturated lipids, and other
compounds in the ELF. In vitro studies have clearly illustrated the role of antioxidants in
4-25
-------
mediating NC>2 uptake. The rapid reactions of NO2 with tracheobronchial ELF substrates
provides a net driving force for NCh mass transfer from the gas phase into the ELF.
Concentrations of "free" solute NO2 are likely negligible due to its reaction-mediated
removal. Thus, it is not NC>2 itself, but rather its reaction products that are believed to
interact with the apical surfaces of the tracheobronchial epithelium. At high substrate
concentrations, oxidants/cytotoxic products are at least partially quenched due to
secondary antioxidant reactions. At low substrate concentration, ELF-derived
oxidants/cytotoxic products have a lower probability of being intercepted by unreacted
antioxidants and instead may reach underlying targets.
Within the alveolar region, much of the inhaled NC>2 entering the ELF will diffuse
through rapidly enough to avoid reactions and will reach underlying tissue surfaces. A
principle component of pulmonary surfactant, DPPC, may partially reduce the uptake of
NC>2 by slowing its diffusion and decreasing reaction with substrates in the subphase
fluid. Reducing the reactive absorption increases diffusive resistance and back diffusion
into the air phase, thereby reducing uptake from the gas phase. Nonetheless, rapid
reactions of NO2 with tissues will maintain a concentration gradient for NC>2 through the
alveolar ELF to the underlying tissues.
Exercise, relative to rest, increases the dose rate of NO2 to the respiratory tract because of
greater NO2 penetration through the extrathoracic airways and a greater intake rate of
NO2. The uptake of NO2 by the upper respiratory tract decreases with increasing
ventilation rates occurring with activity. This causes a greater proportion of inhaled NC>2
to be delivered to the lower respiratory tract. In humans, exercise results in a shift in the
breathing pattern from nasal to oronasal relative to rest. Because the nasal passages scrub
gas-phase NO2 more efficiently than the mouth and because uptake efficiency decreases
with increasing flow, exercise delivers a disproportionately greater quantity of the inhaled
mass to the lower respiratory tract, where the NC>2 is readily absorbed. Experimental
studies have shown that exercise increases the dose rate of NO2 to the respiratory tract by
3- to 5-times compared to resting exposures.
Compared to healthy adults, children and individuals with asthma might be expected to
have greater NO2 penetration into the lower respiratory tract. Children tend to have a
greater oral breathing contribution than adults at rest and during exercise. Limited data
also suggest that patients with asthma or allergic rhinitis breathe oronasally at rest.
Because the nasal passages scrub gas-phase NO2 more efficiently, a greater quantity of
the inhaled NO2 may reach the lower respiratory tract of oronasally breathing individuals.
The dose rate to the lower airways of children compared to adults is increased further
because children breathe at higher minute ventilations relative to their lung volumes.
4-26
-------
Current dosimetry models for NC>2 do not adequately consider reactive absorption and
secondary reactions that affect the probability of oxidants and/or cytotoxic products
reaching target sites. Differences in potential NO2 reactive substrates and reaction
products among species have not been considered in modeling efforts. Although the
models predict similar total NO2 dose in the tracheobronchial airways and sites of
maximal NC>2 tissue dose (i.e., near the beginning of the gas exchange region) among
several mammalian species, the models do not sufficiently link these NO2 doses to
specific reaction products and downstream events. It is unclear to what extent
environmental exposures at current ambient NC>2 concentrations might affect the overall
balance of nitrite and nitrate in the respiratory tract or how ambient NC>2 uptake compares
with endogenous production rates/amounts in the respiratory tract. Systemic nitrite and
nitrate levels are highly dependent on diet and not likely affected by ambient NO2
exposures. However, the uptake of inhaled NO2 could increase levels of nitrite and/or
other reaction products beyond levels that are endogenously occurring in the respiratory
tract.
The uptake of inhaled NO occurs against the background of endogenous NO production
in the respiratory tract. In terms of the overall uptake and tissue diffusion of NO within
the lung, interception due to reactions is not expected to consume appreciable amounts of
the total NO involved in mass transfer from the alveolar to the vascular space. The
absorption of inhaled NO proceeds similarly to oxygen and CO. Blood acts as a near
"infinite" sink for NO. Absorption of NO follows Henry's law for dissolution into the
aqueous phase, and is followed by diffusion into the vascular space, where it interacts
with red blood cell hemoglobin to ultimately form nitrate. Ambient NO concentrations
are likely similar to those endogenously occurring within the lung airspaces, except
during morning commutes or near major roadways, where they may possibly exceed
endogenous levels. It is not known whether periods of high ambient NO exposure could
alter endogenous NO production within the respiratory tract or alter pathways affected by
endogenous NO.
4.3 Modes of Action for Inhaled Oxides of Nitrogen
4.3.1 Introduction
The purpose of this section is to describe the biological pathways that underlie health
effects resulting from short- and long-term exposures to NO2 and NO. Extensive research
carried out over several decades in humans and in laboratory animals has yielded much
information about these pathways. This section will discuss some of the representative
4-27
-------
studies, placing particular emphasis on studies published since the 2008 ISA for Oxides
of Nitrogen (U.S. EPA. 2008a. c) and on studies in humans. This information will be
used to develop a mode of action framework for inhaled NO2 and NO.
Mode of action refers to a sequence of key events, endpoints, and outcomes that result in
a given toxic effect (U.S. EPA. 2005). In contrast, elucidation of mechanism of action
requires a more detailed understanding of key events, usually at the molecular level (U.S.
EPA. 2005). The framework developed in this chapter will include some mechanistic
information on initiating events at the molecular level but mainly will focus on the effects
of NO2 and NO at the cellular, tissue, organ, and organism level.
NO2 is a radical species and a highly reactive oxidant gas [(Tukuto et al.. 2012):
Table 4-21. It is well known that oxidation and nitration products, which are formed as a
result of NO2 exposure, initiate numerous responses at the cellular, tissue, and whole
organ level of the respiratory system. Exposure to NO2 may also have effects outside the
respiratory tract. NO is a radical species and a gas that is more selective in its reactivity
than NO2 [(Fukuto et al., 2012); Table 4-21. Once inhaled, NO rapidly crosses the
alveolar capillary barrier into the vascular compartment and avidly binds to hemoglobin.
Subsequent reactions with hemoglobin lead to the generation of circulating nitrate, nitrite,
and methemoglobin.
Table 4-2 Chemical properties of nitrogen dioxide and nitric oxide that
contribute to proposed modes of action.
NO2
NO
Radical species
Radical species
Somewhat hydrophobic
Very hydrophobic
Very reactive
Selectively reactive
Less diffusible
More diffusible
Reactions with:
(1) unsaturated fatty acids
(2)thiols
(3) low molecular weight antioxidants
Radical-radical reactions with:
(1) superoxide to form peroxynitrite
(2) thiyl radicals to form RSNO
(3) organic peroxyl radicals
Reacts with amino acids, proteins, and lipids to form
nitrated species
Reacts with heme-containing proteins, transition metals, and
oxygen
Initiates radical reactions and lipid peroxidation
Quenches radical reactions
Metabolites include nitrite and nitrate
Metabolites include nitrite and nitrate
RSNO = S-nitrosothiols.
4-28
-------
Both NO2 and NO are formed endogenously in cells and tissues (Sections 4.2.2.4 and
4.2.3). Formation of endogenous NO is catalyzed by NOS. Three pathways contribute to
the formation of endogenous NO2: (1) acidification of nitrite, usually occurring in the
phagolysosomes; (2) reaction of peroxynitrite with carbonate to form
nitrosoperoxylcarbonate anion, which decomposes to carbonate anion and NO2; and
(3) reaction of peroxidases using nitrite and hydrogen peroxide as substrates. These
enzymatic and nonenzymatic pathways are enhanced during immune responses and
inflammation, leading to higher endogenous levels of NO and NO2. Furthermore, dietary
consumption of nitrate leads to increased levels of NO in the stomach and to increased
circulating levels of nitrite due to activity of the enterosalivary cycle (Weitzberg and
Lundberg. 2013; Lundberg et al.. 2011). The contribution of environmentally relevant
concentrations of inhaled NO2 and NO to levels of circulating nitrite and nitrate is
thought to be minimal (Sections 4.2.2.4 and 4.2.3). However, inhaled NO2 may act on the
same targets as endogenous NO2 produced during inflammation in the respiratory tract
(Ckless et al.. 2011). Endogenous NO2 is thought to contribute to the development of
lung disease; inhaled NO2 may further this process.
The following subsections describe the current understanding of biological pathways that
may be responsible for the pulmonary and extrapulmonary effects of inhaled NO2 and
NO. For NO2, this includes the formation of oxidation and nitration products
(Section 4.3.2.1), activation of neural reflexes (Section 4.3.2.2). initiation of
inflammation (Section 4.3.2.3). alteration of epithelial barrier function (Section 4.3.2.4).
enhancement of bronchial smooth muscle reactivity (Section 4.3.2.5). modification of
innate/adaptive immunity (Section 4.3.2.6). and remodeling of airways and alveoli
(Section 4.3.2.7). The potential induction of carcinogenesis is also briefly described
(Section 4.3.2.8). While NO2 exposure may result in effects occurring outside of the
respiratory tract, biological pathways underlying extrapulmonary effects of NO2 are not
well understood (Section 4.3.2.9). Activation of neural reflexes and release of
inflammatory or vasoactive mediators from the lung to the bloodstream are possibilities.
Inhaled NO impacts the pulmonary and systemic vasculature mainly through interaction
with heme proteins (Section 4.3.3). Other effects of NO may be due to circulating
metabolites (such as nitrite, nitrate, and methemoglobin), interactions with redox-active
transition metals, or reactions with thiyl and superoxide radicals. Because endogenous
NO is an important mediator of cell signaling, inhaled NO has the potential to disrupt cell
signaling.
4-29
-------
4.3.2 Nitrogen Dioxide
4.3.2.1 Formation of Oxidation and Nitration Products
The 2008 ISA and the 1993 AQCD for Oxides of Nitrogen (U.S. EPA. 2008a. c, 1993a)
summarized biochemical effects observed in the respiratory tract after NO2 exposure.
These effects have been attributed to the strong oxidizing potential of NO2, resulting in
the formation of reactive oxygen species (ROS). Key responses include oxidation of
membrane polyunsaturated fatty acids, thiol groups, and antioxidants. Chemical
alterations of lipids, amino acids, proteins, and enzymes can lead to functional changes in
membranes, enzymes, and oxidant/antioxidant status. For example, lipid peroxidation of
unsaturated fatty acids in membranes may alter membrane fluidity and permeability. As a
result, epithelial barrier functions may be impaired, and phospholipases may be activated
leading to the release of arachidonic acid. In addition, oxidation of protein thiols may
result in enzyme dysfunction. Further, consumption of low molecular weight antioxidants
by NC>2 may result in decreased antioxidant defenses. Effects may occur directly through
the action of NO2 or secondarily due to its reaction products, such as organic radicals,
ROS, or reactive nitrogen species (RNS). Later effects may occur due to the release of
ROS and/or RNS by leukocytes responding to cell damage.
As summarized in the 2008 ISA and the 1993 AQCD (U.S. EPA. 2008a. c, 1993a).
considerable attention has been paid to the effects of NO2 on the antioxidant defense
system in the ELF and in respiratory tract tissue. Studies in humans and animals exposed
to NO2 have demonstrated changes in low molecular weight antioxidants, such as
glutathione, ascorbate, and a-tocopherol, and in the activities of enzymes responsible for
glutathione synthesis or maintenance of redox status. For example, a controlled human
exposure study found depletion of urate and ascorbate, but not glutathione, in BAL fluid
1.5 hours following a 4-hour exposure to 2,000 ppb NO2 (Kelly etal.. 1996a). While
these results may be interpreted as evidence that NO2 prefers to react with urate or
ascorbate over glutathione, an alternative interpretation is that glutathione reacts with
NO2 and that the product of the reaction is reduced by urate or ascorbate
(Section 4.2.2.2). Other studies have found that antioxidant status modulates the effects
of NO2 inhalation. For example, in a controlled human exposure study, supplementation
with ascorbate and a-tocopherol decreased the levels of lipid peroxidation products found
in BAL fluid following a 3-hour exposure to 4,000 ppb NO2 (TVIohsenin. 1991).
Additionally, changes in lung antioxidant enzyme activity have been reported in animals
exposed to NO2 (U.S. EPA. 2008c). For example, long-term exposure to NO2 resulted in
decreased glutathione peroxidase activity in weanling mice that were a-tocopherol
4-30
-------
deficient, while supplementation with a-tocopherol resulted in an increase in glutathione
peroxidase activity (Ayaz and Csallany. 1978). Thus, NCh inhalation is capable of
perturbing glutathione-dependent reactions. These changes may reflect altered cell
populations because injury induced by NC>2 exposure may result in the influx of
inflammatory cells or the proliferation of resident epithelial or mesenchymal cells.
Changes in cell populations due to proliferative repair may also account for the
upregulation of Phase II, Phase I, and glycolytic enzymes that have been observed
following NC>2 exposure.
As discussed in Section 4.2.2.1.2. reactive absorption of NO2 gas occurs by reactions
with antioxidants and other components of the ELF. Studies employing in vitro and
in vivo systems point to the ability of antioxidants to both react with NC^to form reactive
intermediates and to quench those reactive intermediate species. NO2 exposure in the
presence of ELF antioxidants resulted in the formation of superoxide and hydrogen
peroxide in an in vitro cell system (Velsor and Postlethwait. 1997). In this study,
quenching of NCh-derived secondary oxidants was dependent on antioxidant
concentration, with lower concentrations promoting and higher concentrations reducing
oxidative injury. A recent in vivo study provided additional support for this mechanism.
Supplementation of mice with ascorbate had a biphasic effect, with a lower dose of
ascorbate promoting and a higher dose of ascorbate reducing lung injury and
inflammation induced by exposure to 20,000 ppb NC>2 (Zhang et al.. 201 Ob). Thus,
toxicity resulting from NC>2 exposure may be due to a product derived from the initial
ELF substrate and/or to secondary reaction products formed. These reaction products
may not be long lived due to short half-lives and/or continuous turnover of the ELF.
Further, quenching of reaction products by ELF antioxidants may limit damage to the
respiratory epithelium. The heterogeneous distribution of epithelial injury due to reactive
intermediates formed from inhaled NC>2 may reflect ELF-dependent local effects because
the ELF is nonuniform in composition and quantity along the respiratory tract.
Furthermore, localized endogenous formation of NC>2 in the respiratory tract could
overwhelm the antioxidant capacity and contribute to epithelial injury.
Nitrogen-based metabolites and RNS are also formed in the ELF as a result of NO2
exposure. Nitrite is the primary product of the chemical reactions of NC>2 in the
respiratory tract. As discussed in Section 4.2.2.1.2. nitrite formed in the ELF can diffuse
into respiratory tract epithelial cells and subsequently into the vascular space. While
levels of nitrite may not change appreciably in the respiratory tract, a localized effect of
nitrite on epithelial cells cannot be ruled out. However, based on numerous studies
investigating the effects of increased systemic nitrite on various tissues and organs, it
seems unlikely that nitrite is responsible for the toxicity of NC>2. Nitrite has been found to
protect against ischemia-reperfusion injury in the heart and other organs (Weitzberg and
4-31
-------
Lundberg. 2013). In addition, systemic nitrite administration prevented airway and
epithelial injury due to exposure to chlorine gas in rats (Yadavetal.. 2011). Further,
nitrite is known to have a direct relaxing effect on smooth muscle [(Folinsbee. 1992);
Section 4.3.4.1]. suggesting that it may play a role in bronchodilation.
RNS, such as RSNO and nitrated proteins, fatty acids, and lipids, may be formed in the
respiratory tract following NC>2 exposure. Evidence for these reaction products is mainly
provided by in vitro cell systems and ex vivo systems (Section 4.3.4). However, Matalon
et al. (2009) recently demonstrated the nitration of surfactant protein D (SP-D) in mice
exposed to 20,000 ppb NO2 for 4 hours. SP-D nitration was accompanied by protein
cross-linking and a decrease in SP-D aggregating activity, which could potentially impact
microbial clearance, immune regulation, and surfactant metabolism. In addition to
inhibiting protein function, nitration of proteins may induce antigenicity or trigger
immune reactions (Daiber and Muenzel. 2012). Further, the presence of nitrated amino
acids such as 3-nitrotyrosine in cells or tissues is an indicator of endogenous NC>2 and
peroxynitrite formation. Other potential RNS formed may have less deleterious effects.
For example, nitrated (or nitro) fatty acids have a direct relaxing effect on smooth
muscle, perhaps even on airway smooth muscle (Que et al.. 2009; Lima et al.. 2005). In
addition, RSNOs are known to be bronchodilators (Que etal.. 2009). Additional
discussion of the biological effects of these products of NO2 metabolism is found in
Section 4.3.4.
Collectively, these studies provide evidence for the formation of oxidation and nitration
products as a result of NO2 exposure. Responses included lipid peroxidation and nitration
of SP-D in the respiratory tract. In addition, NC>2 exposure altered antioxidant systems.
While studies in humans involved ambient-relevant exposures to NCh,, studies conducted
in experimental animals or in vitro systems mainly involved higher concentrations of
NO2 (i.e., >5,000 ppb).
4.3.2.2 Activation of Neural Reflexes
NO2 is classified as a pulmonary irritant (Alarie. 1973). Pulmonary irritants stimulate
afferent nerve endings in the lung, resulting in increased respiratory rate, decreased VT,
and subsequent rapid shallow breathing. Sometimes pulmonary irritants also stimulate
mild bronchoconstriction, bradycardia, and hypotension (Alarie. 1973). All of these
pathways involve the vagus nerve.
Numerous studies investigated pulmonary irritant effects of NO2 exposure using
respiratory rate as an indicator of neural reflex activation. In guinea pigs, NC>2 exposure
(5,200-13,000 ppb; 2-4 hours) by nose-cone resulted in statistically significant increases
4-32
-------
in respiratory rate and decreases in VT (Murphy et al., 1964). These responses were
concentration and time dependent and were reversible when animals were returned to
clean air. In contrast, no changes in these respiratory parameters were observed with
4-hour exposures to 16,000 and 50,000 ppb NO. In another study, guinea pigs exposed to
7,000-146,000 ppb NO2 for 1 hour demonstrated a concentration-dependent increase in
respiratory rate 10 minutes following exposure and a concentration-dependent decrease in
VT 10 minutes, 2 hours, and 19 hours following exposure (Silbaugh et al., 1981). NO2
exposure-induced increases in respiratory rate have also been reported in rats (Freeman et
al., 1966) and mice (McGrath and Smith. 1984). In mice, statistically significant
increases in respiratory rate and decreases in VT were found in response to an 8-minute
exposure to 100,000 ppb NO2, but not to 15,000 or 50,000 ppb NO2 (McGrath and Smith.
1984). In this latter study, continuous pre-exposure to 5,000 ppb NO2 for 3 days lessened
the response to 100,000 ppb NO2, suggesting the development of tolerance or an
attenuated response to NO2 (U.S. EPA. 1993a). In rats, continuous exposure to 800 ppb
and higher concentrations of NO2 resulted in elevated respiratory rates throughout life
(Freeman et al.. 1966). However, no NO2 exposure-induced increases in respiratory rate
in human subjects have been reported. In fact, respiratory rates tended to decrease in
humans exposed to 0-480 ppb for 20 minutes (Bvlinet al.. 1985). The authors proposed
that NO2 in this range of concentrations did not act as a pulmonary irritant in humans.
NO2 has been shown to elicit a small increase in airway resistance, which is consistent
with mild bronchoconstriction, in humans but not in rabbits or guinea pigs [Alarie (1973)
and studies cited below]. One study in human subjects at rest found a nonmonotonic
response to NO2 in terms of airway resistance (Bvlinet al.. 1985). In this study, specific
airway resistance was increased after 20 minutes of exposure to 250 ppb NO2 and was
decreased after 20 minutes of exposure to 480 ppb NO2. The authors suggested that reflex
bronchoconstriction occurred at the lower concentration and that other mechanisms
counteracted this effect at the higher concentration. Other controlled human exposure
studies found no change in airway resistance with acute exposures of 530-1,100 ppb NO2
and increases in airway resistance with acute exposures above 1,600-2,500 ppb in
healthy human subjects (U.S. EPA. 1993a). Human subjects with chronic lung disease
exposed for 5 minutes to 2,100 ppb NO2 also exhibited increased airway resistance (von
Nieding and Wagner. 1979). In addition, both forced expiratory volume in 1 second
(FEVi) and forced vital capacity were decreased in healthy human subjects exposed to
2,000 ppb NO2 for 4 hours (Blomberg et al.. 1999). These changes in pulmonary function
are consistent with reflex bronchoconstriction. Because the response was lessened with
each successive exposure on 4 consecutive days, Blomberg et al. (1999) suggested the
development of tolerance or an attenuated response.
4-33
-------
Some evidence points to NO2 exposure-induced histamine release from mast cells, rather
than reflex bronchoconstriction, as the mechanism underlying changes in airway
resistance. A study in rats showed that mast cell degranulation occurred after acute
exposure to NCh [500 ppb for 4 hours; 1,000 ppb for 1 hour; (Thomas et al.. 1967)]. In
addition, a histamine-suppressive agent, but not atropine or /^-agonists, blocked
NO2-mediated increases in airway resistance in healthy humans and in humans with
chronic lung disease exposed to 5,000-8,000 ppb NCh for 5 minutes (von Nieding and
Wagner. 1979). Because atropine inhibits vagal responses, these findings indicate that
neural reflexes were not involved in NC>2-induced changes in pulmonary function in
human subjects. More recent studies in animals have provided experimental evidence for
a relationship between lipid peroxidation/oxidative stress and the release of histamine by
allergen-activated mast cells (Beaven. 2009; Gushchin et al.. 1990). Taken together, these
studies suggest that NCh exposure may lead to lipid peroxidation, which may promote
mast cell-mediated changes in pulmonary function, albeit at high concentrations.
There is some experimental support for NC>2 exposure-induced cardiovascular reflexes.
An acute exposure to NO2 in an occupational setting resulted in tachycardia in one case
report (U.S. EPA. 1993a: Bates etal.. 1971). while rats exposed acutely to 20,000 ppb or
higher concentrations of NC>2 exhibited bradycardia (U.S. EPA. 1993a: Tsubone et al..
1982). This latter response was abolished by injection of atropine, which inhibits vagal
responses. Further, a decreased heart rate, which was not accompanied by an increase in
respiratory rate, was observed in mice exposed to 1,200 and 4,000 ppb NO2 for 1 month
(Suzuki etal.. 1981). The lack of respiratory rate response suggests that the decreased
heart rate was due to a different mechanism than rapid stimulation of irritant receptors by
NC>2. Controlled human exposure studies have also examined the effects of NO2 on heart
rate and heart rate variability (Section 5.3.10.1). Older studies and one recent study failed
to find statistically significant changes in heart rate at ambient-relevant concentrations of
NO2. A recent controlled human exposure study involving a 1-hour exposure to 400 ppb
NC>2 failed to find an effect on heart rate variability in subjects with coronary heart
disease (Scaife et al.. 2012). However, a second recent controlled human exposure study
reported an effect on heart rate variability resulting from a 2-hour exposure to 500 ppb
NO2 (Huang et al.. 2012b). Altered heart rate variability found in epidemiologic studies
(Section 5.3.10.1) is consistent with a possible effect of NO2 exposure on autonomic
tone.
Collectively, these studies show that NO2 is a pulmonary irritant that may affect airway
function and cardiac function through activation of neural reflexes involving the vagus
nerve. Experimental NCh exposures in animals resulted in increased respiratory rate,
decreased VT, reflex bronchoconstriction, and bradycardia. Responses were rapid,
concentration dependent, and variable among species. Evidence that reflex responses
4-34
-------
occur in humans is weak because no increases in respiratory rate have been reported as a
result of NO2 exposure. Further, modest increases in airway resistance in human subjects
exposed to NO2 were not blocked by atropine, which inhibits vagal responses. Findings
attributed to reflex bronchoconstriction in humans may be due to alternative pathways
such as mast cell degranulation. The recent demonstration that NO2 exposure (500 ppb;
2 hours) resulted in altered heart rate variability suggests the possible activation of a
neural reflex in humans. However, the clearest evidence for reflex responses mediated by
the vagus nerve involved exposures of experimental animals to NO2 at concentrations
higher than those considered ambient-relevant (i.e., >5,000 ppb).
4.3.2.3 Initiation of Inflammation
As summarized in the 2008 ISA and the 1993 AQCD for Oxides of Nitrogen (U.S. EPA.
2008a. c, I993a), NC>2 exposure-induced membrane perturbations resulted in the release
of arachidonic acid and the formation of eicosanoid products (Sections 5.2.2.5 and
5.2.7.4). Animal toxicological and controlled human exposure studies have found
increases in concentrations of eicosanoids in BAL fluid immediately or 1 hour following
a 2-3-hour exposure to 1,000 ppb NO2 (Torres et al.. 1995; Schlesingeretal.. 1990).
Eicosanoids play an important role in the recruitment of neutrophils. Interestingly, higher
concentrations (3,000 and 10,000 ppb; 2 hours) or longer durations (500 ppb; 12 hours)
of exposure to NO2 resulted in inhibited eicosanoid production (Robison and Forman.
1993; Schlesingeretal.. 1990).
Recently, acute exposure of mice to 10,000 ppb and higher concentrations of NO2 was
shown to activate the transcription factor, nuclear factor kappa-light-chain enhancer of
activated B cells (NFKB), in the airway epithelium (Ather et al.. 2010; Bevelander et al..
2007). NFKB activation resulted in the production of pro-inflammatory cytokines.
Inflammation and acute lung injury in this model were found to be dependent on an
active NFKB pathway. Controlled human exposure studies demonstrated increased levels
of cytokines IL-6 and IL-8 in BL fluid following NC>2 exposure. IL-8 levels were
increased at 1.5 and 16 hours following a 4-hour exposure to 2,000 ppb NCh, while levels
of IL-6 were increased at 16 hours following the exposure [(U.S. EPA. 2008c; Devlin et
al.. 1999; Blomberg et al.. 1997); Section 5.2.7.4]. Because IL-6 and IL-8 are under
NFKB transcriptional regulation, these studies suggest that exposure to 2,000 ppb NO2
may have led to an increase in IL-6 and IL-8 by stimulating the NFKB pathway in human
subjects.
Airway inflammation often occurs following NC>2 exposure. Studies in rodents exposed
acutely (1 hour to 3 days) to NCh (500-30,000 ppb) have demonstrated airway
4-35
-------
inflammation, mainly consisting of neutrophils and macrophages and sometimes of mast
cells and lymphocytes, by histological technique or sampling of BAL fluid [as
summarized in (Poynter et al. (2006); Paganietal. (1994); Sandstrom et al. (1990))1.
Numerous studies in healthy human subjects exposed to NC>2 have documented airway
inflammation in endobronchial biopsy tissue and in sputum, BL fluid, and BAL fluid.
Many of these studies were conducted while subjects were exercising intermittently and
exposed to 1,500-4,000 ppb NCh for a few hours. Neutrophilia was a prominent feature
(U.S. EPA. 2008c; Frampton et al.. 2002; Devlin etal.. 1999; Azadniv et al.. 1998;
Blomberg et al.. 1997). In addition, other types of inflammatory cells, including
macrophages, lymphocytes, and mast cells, have been demonstrated (Frampton et al..
2002; Sandstrom et al.. 1991; Sandstrom et al.. 1990).
Controlled human exposure studies have also evaluated the effects of repeated NO2
exposure on airway inflammation in healthy adults. Persistent neutrophilic inflammation,
demonstrated by increased numbers of neutrophils and increased levels of
myeloperoxidase in the BL fluid, was observed following 4 consecutive days of 4-hour
exposure to 2,000 ppb NCh (Blomberg et al.. 1999). Repeated exposure also led to the
upregulation of cytokines characteristic of the T helper cell 2 (Th2) inflammatory
response and also of inter-cellular adhesion molecule 1 (ICAM-1) in respiratory
epithelium (U.S. EPA. 2008c; Pathmanathan et al.. 2003). Upregulation of ICAM-1
suggests a potential mechanism for the persistent neutrophil influx that was observed
(Blomberg et al., 1999). A study of repeated exposure to 4,000 ppb (exposure every other
day for a total of six exposures) found inflammatory responses that differed from those
observed after a single exposure (Sandstrom et al.. 1992a). In particular, numbers of mast
cells and lymphocytes in the lavage fluid, which were increased following a single
exposure, were not increased following repeated exposure. Furthermore, repeated
exposure to 1,500 ppb NCh (by the same protocol) resulted in smaller numbers of some
lymphocyte subpopulations in BAL obtained following exposure compared with numbers
in BAL obtained prior to exposure (Sandstrom et al.. 1992b). In contrast, no changes in
lymphocyte subpopulations were reported following repeated exposure to 600 ppb NCh
(4 exposures over 6 days), with the exception of a slight increase in natural killer cells
(Rubinstein et al.. 1991).
Recently, a controlled human exposure study investigated the effects of repeated NO2
exposure on eosinophilic airway inflammation in subjects with atopic asthma (Ezratty et
al.. 2014). Subjects were exposed to 200 or 600 ppb NO2 for 30 minutes on the first day
and twice for 30 minutes on the second day. Compared with baseline, the number and
percentage of eosinophils and the amount of eosinophil cationic protein (ECP) in sputum
were significantly increased after the three exposures to 600, but not 200 ppb NO2.
Furthermore, ECP was highly correlated with eosinophil counts in sputum. No increases
4-36
-------
in either of these parameters were observed 6 hours after the first exposure to 600 ppb
NO2. While the design of this study did not include an allergen challenge, several other
studies examined eosinophilic inflammation and other allergic responses to NC>2 and an
allergen (Sections 4.3.2.6.2. 4.3.2.6.3. and 5.2.2.5). Studies investigating eosinophilic
inflammation suggest that exposure to NC>2 may prime eosinophils for subsequent
activation by allergens in previously sensitized individuals (Davies etal.. 1997; Wang et
al.. 1995b).
Collectively, these studies in animals and humans demonstrate that NO2 exposure
initiated inflammation in the respiratory tract. Responses included increases in
eicosanoids, cytokines, neutrophils, and eosinophils in the airways. Many, but not all, of
these studies involved ambient-relevant exposures to NCh.
4.3.2.4 Alteration of Epithelial Barrier Function
Lipid peroxidation and altered phospholipid composition in the respiratory tract
following NO2 exposure may affect membrane fluidity and airway epithelial barrier
function. NC>2 exposure-induced inflammation may further impair epithelial barrier
function. Increases in vascular permeability may occur, leading to the influx of plasma
proteins such as albumin into the airway lumen.
As summarized in the 2008 ISA and the 1993 AQCD for Oxides of Nitrogen (U.S. EPA.
2008a. c, I993a), numerous studies have demonstrated increases in biomarkers of
increased permeability, such as protein and albumin, as well as biomarkers of cellular
injury, such as lactate dehydrogenase (LDH) and shed epithelial cells, in BAL fluid
following exposure to NC>2 (Section 5.2.7.4). Because LDH can be oxidatively
inactivated, use of this indicator may underestimate the extent of injury during oxidative
stress. Many, but not all, of these effects were observed at NO2 concentrations that are
higher than ambient-relevant levels. Notably, one controlled human exposure study found
increased albumin levels in BL fluid following 4 consecutive days of 4-hour exposure to
2,000 ppb NO2 (Blomberg et al.. 1999).
Several studies in experimental animals found that antioxidant deficiency worsened the
cellular injury and/or impaired epithelial barrier function following NC>2 exposure.
Ascorbate deficiency enhanced protein levels in the BAL fluid of NCh-exposed guinea
pigs, suggesting a role for BAL fluid ascorbate in preventing the deleterious effects of
NO2 (Hatchet al.. 1986). Similarly, a-tocopherol deficiency enhanced lipid peroxidation
in NCh-exposed rats (Sevanian et al.. 1982a). Recently, selenium deficiency was found to
enhance the injury response in rats exposed to 1,000-50,000 ppb (acute, subacute, and
chronic exposures) NCh (de Burbure et al.. 2007). Levels of both BAL fluid total protein
4-37
-------
and serum club cell secretory protein were increased in selenium-deficient rats exposed to
NO2. Selenium supplementation diminished this response, which suggests that the
selenium-containing enzyme, glutathione peroxidase, may have played an important
mitigating role; however, a role for other selenium-dependent processes cannot be ruled
out.
Increases in lung permeability due to high concentrations of NC>2 (100,000 ppb and
above) are known to cause death from pulmonary edema (Lehnert et al.. 1994; Gray et
al.. 1954). At lower concentrations, more subtle effects have been reported. Exposure of
rats to 5,000 and 10,000 ppb NC>2 for 3 or 25 days resulted in epithelial degeneration and
necrosis and in proteinaceous edema (Earth et al.. 1995). while exposure to
800-10,000 ppb NO2 for 1 and 3 days resulted in concentration-dependent increases in
BAL fluid protein (Miiller et al.. 1994). BAL fluid protein was also elevated in guinea
pigs exposed for 1 week to 400 ppb NO2 (Sherwin and Carlson. 1973).
High concentrations of NC>2 (70,000 ppb; 30 minutes) were found to enhance
translocation of instilled antigen from the lung to the bloodstream of guinea pigs
(Matsumura. 1970). More subtle increases in lung permeability due to NC>2 exposure
could enhance the translocation of an antigen to local lymph nodes and subsequently to
the circulation (U.S. EPA. 2008c; Gilmour et al.. 1996) and/or to the immunocompetent
and inflammatory cells underlying the epithelium that are involved in allergic reactions
(Jenkins etal.. 1999). However, increased lung permeability following exposure to NC>2
does not always lead to allergic sensitization (Alberg et al.. 2011). Increased epithelial
permeability may alternatively contribute to the activation of neural reflexes and the
stimulation of smooth muscle receptors by allowing greater access of an agonist (Dimeo
etal.. 1981).
Susceptibility to NO2 exposure-induced lung injury was investigated in several mice
strains with differing genetic backgrounds (Kleeberger et al.. 1997). Lavageable total
protein, a biomarker for increased lung permeability, was variable among mouse strains
following a 3-hour exposure to 15,000 ppb NCh. In addition, repeated exposure to NCh
(10,000 ppb, 6 hour/day, 5 consecutive days) resulted in adaptation of the permeability
response in one of the tested strains but not in the other. Although specific genes were not
identified, this study provided evidence that genetic components conferred susceptibility
to NO2, at least in terms of lung permeability.
Collectively, these studies in animals and humans demonstrate that NO2 exposure
increased airway permeability. Responses included increased protein and albumin in
BAL or BL fluid and were enhanced by antioxidant deficiency. While many of these
studies involved exposures that were at higher concentrations than those considered
ambient-relevant, a study in humans and several studies in animals provided evidence of
4-38
-------
increased airway permeability following exposure to ambient-relevant concentrations of
NO2.
4.3.2.5 Enhancement of Bronchial Smooth Muscle Reactivity
Exposure to NCh enhanced the inherent reactivity of airway smooth muscle in human
subjects with and without asthma [(Folinsbee. 1992); Section 5.2.2.11 and in animal
models (see below). This "airway responsiveness" is defined as the sensitivity of airways
to a variety of natural or pharmacological stimuli (O'Byrne et al.. 2009). Airway
hyperresponsiveness (AHR) is a key feature of asthma, which is a chronic inflammatory
disease of the airways. As summarized in the 2008 ISA for Oxides of Nitrogen (U.S.
EPA. 2008c) and in Section 5.2.2.1. numerous studies found that human subjects exposed
to NCh were more sensitive to the nonspecific stimuli methacholine than human subjects
exposed to air. Subjects with asthma exhibited greater sensitivity than subjects without
asthma when similarly exposed. In addition, several studies found that NCh exposure
enhanced airway responsiveness to specific stimuli, such as allergens, in subjects with
mild allergic asthma.
Exercise during exposure to NCh appeared to modify airway responsiveness in subjects
with asthma [(Folinsbee. 1992); Section 5.2.2.11. Mechanisms by which this occurs are
not understood, but two hypotheses have been postulated. First, exercise-induced
refractoriness, which has been demonstrated in some subjects with asthma, may alter
responsiveness to NCh (Magnussen et al.. 1986). A second hypothesis is that nitrite
formed by reactions of NCh in the ELF mediates compensatory relaxation of airway
smooth muscle (Folinsbee. 1992). Exercise would increase the total dose of NCh to the
respiratory tract, thus increasing nitrite formation. Recent studies have shown that RNS
have bronchodilatory effects. For example, endogenous RSNOs are an important
modulator of airway responsiveness in subjects with asthma and in eosinophilic
inflammation (Lee et al.. 201 Ib; Que etal.. 2009).
Animal toxicological studies have also demonstrated NCh-induced AHR to nonspecific
and specific challenges, as summarized in the 2008 ISA and the 1993 AQCD (U.S. EPA.
2008a. c, 1993a) and in Section 6.2.2.3. Exposures ranged from acute to subchronic in
these studies, and results suggest that more than one mechanism may have contributed to
the observed AHR. Acute exposure of guinea pigs to NCh (10 minutes; 7,000 ppb and
higher) resulted in concentration-dependent AHR to histamine, which was administered
immediately after exposure (Silbaugh et al.. 1981). This response was short-lived because
no enhanced responsiveness was seen at 2 and 19 hours post-exposure to NCh. The
rapidity of the response and the concomitant change in respiratory rate suggest enhanced
4-39
-------
vagally mediated reflex responses (Section 4.3.2.2) as a possible underlying mechanism.
A 7-day exposure to 4,000 ppb NCh also induced AHR to histamine in guinea pigs
(Kobayashi and Shinozaki. 1990). Eicosanoids were proposed to play a role in this
response. In addition, a study in mice sensitized and challenged with ovalbumin found
that short-term exposure to NCh (25,000 ppb, but not 5,000 ppb; 3 days) resulted in AHR
to methacholine (Poynter et al.. 2006). This enhanced sensitivity correlated with an
increase in numbers of eosinophils, suggesting eosinophilic inflammation as a possible
underlying mechanism in this model of allergic airway disease. A subchronic study
demonstrated dose-dependent AHR to histamine in NO2-exposed guinea pigs
[1,000-4,000 ppb, 24 hour/day, 6-12 weeks; (U.S. EPA. 2008c: Kobavashi and Miura.
1995)1. Specific airway resistance in the absence of a challenge agent also was increased,
which indicates the development of airway obstruction. This finding suggests airway
remodeling as a possible underlying mechanism for AHR. Another subchronic exposure
study (5,000 ppb NC>2, 4 hour/day, 5 days/week, 6 weeks) found a delayed bronchial
response, which was measured as increased respiratory rate and was suggestive of AHR,
in guinea pigs sensitized and challenged with Candida albicans and exposed to NC>2
(Kitabatake et al.. 1995).
Mechanisms underlying the effects of NC>2 on airway responsiveness are not well
understood. Effects of NO2 exposure on redox status in the respiratory tract should be
considered because asthma pathogenesis, including airway inflammation, responsiveness,
and remodeling, may be under redox control (Comhair and Erzurum. 2010; Kloek et al.,
2010). In support of this mechanism, supplementation with the antioxidant ascorbate was
found to prevent nonspecific AHR in subjects with asthma who were exposed to 2,000
ppb NO2 (Mohsenin. 1987b).
Several different inflammatory pathways may underlie the increased airway
responsiveness following NC>2 exposure (Krishna and Holgate. 1999). First, mast cell
activation may contribute to NC>2 exposure-induced AHR. As discussed in
Section 4.3.2.2. acute exposure to NO2 led to mast cell activation in rats and possibly in
human subjects. Histamine released by mast cells can directly bind to receptors on
smooth muscle cells and cause contraction. This response would have the appearance of
reflex bronchoconstriction but would not involve neural pathways. Secondly, neutrophilic
and eosinophilic inflammation, which have been demonstrated following single and
repeated exposures to NO2 (Section 4.3.2.3). may play a role. Neutrophils and other
inflammatory cell types release mediators, such as IL-13, IL-17A, and tumor necrosis
factor-a (TNF-a), which can alter the calcium sensitivity of the smooth muscle and
enhance a contractile response to a stimulus (Kudoetal.. 2013b). Eosinophils can release
ECP and other mediators involved in allergen-induced asthmatic responses. This pathway
may contribute to the enhanced immune response to allergens demonstrated following
4-40
-------
NC>2 exposure (Section 4.3.2.6.2). Eosinophil release of ECP may also cause damage to
the airway epithelium in allergic airway disease (Ohashi et al.. 1994). This damage may
result in epithelial shedding and mucociliary dysfunction, which may allow greater access
of allergens to the airway epithelium and submucosa. In addition, epithelial shedding may
lead to greater exposure of sensory nerve endings on nerve fibers and to enhanced
activation of neural reflexes and airway smooth muscle contraction (Hesterberg et al..
2009; Cockcroft and Davis. 2006c). These processes may explain the close relationship
that has been observed between epithelial shedding and AHR (Ohashi etal.. 1994). Thus,
neutrophilic and/or eosinophilic airway inflammation following NC>2 exposure may
contribute to AHR through the release of mediators or by impairing epithelial barrier
function (Section 4.3.2.4). Thirdly, chronic airway inflammation may cause structural
changes in the airway walls that enhance the contractile response of the smooth muscle to
a given stimuli (Cockcroft and Davis. 2006c).
Evidence also supports a role for endogenous NO2 in mediating AHR. Increased
peroxynitrite formation occurs during inflammatory states, resulting from the reaction of
NO and superoxide. Peroxynitrite subsequently reacts with €62 to form
nitrosoperoxylcarbonate anion, which decomposes to carbonate radical and NO2
(Section 4.2.2.4). Recent studies provided evidence that endogenous peroxynitrite
contributes to AHR in animal models of allergic airway disease (Section 4.3.2.6.2). These
studies demonstrate that NO metabolism is dysfunctional in inflamed lungs and results in
enhanced peroxynitrite formation. Amelioration of the dysfunction resulted in less
nitrative stress, airway remodeling and airway responsiveness (Ahmad etal.. 2011;
Mabalirajan et al.. 201 Ob; Maarsingh et al.. 2009; Maarsingh et al., 2008).These studies
highlight the possibility that inhaled NO2 can add to the lung burden of endogenous NO2,
which contributes to AHR and allergic airway disease in animal models
(Section 4.3.2.6.2).
Collectively, these studies in animals and humans demonstrate that NO2 exposure
enhanced bronchial smooth muscle reactivity. The majority of these studies involved
ambient-relevant exposures to NO2. Antioxidant supplementation reduced this response
in humans. Experimental studies in animals suggest the involvement of mast cells,
neutrophilic and eosinophilic inflammation, airway remodeling, and endogenous NO2 in
enhancing bronchial smooth muscle reactivity.
4.3.2.6 Modification of Innate/Adaptive Immunity
Host defense depends on effective barrier function and on innate and adaptive immunity
(Al-Hegelan et al.. 2011). The effects of NO2 on barrier function in the airways were
4-41
-------
discussed above (Section 4.3.2.4). This section focuses on the mechanisms by which
exposure to NCh could impact innate and adaptive immunity. Both tissue damage and
foreign pathogens are triggers for the activation of the innate immune system. Innate
immune system activation results in the influx of inflammatory cells, such as neutrophils,
mast cells, basophils, eosinophils, monocytes, and dendritic cells, and the generation of
cytokines, such as TNF-a, IL-1, IL-6, keratinocyte chemoattractant, and IL-17. Further,
innate immunity encompasses complement, collectins, and the phagocytic functions of
macrophages, neutrophils, and dendritic cells. In addition, the airway epithelium
contributes to innate immune responses. Innate immunity is highly dependent on cell
signaling networks involving toll-like receptor (TLR) 4 in airway epithelium and other
cell types. Adaptive immunity provides immunologic memory through the actions of B
and T lymphocytes. Important links between the two systems are provided by dendritic
cells and antigen presentation.
4.3.2.6.1 Impairment of Host Defenses
As summarized in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008a. c), potential
mechanisms by which NC>2 exposure may impair host defenses include ciliary dyskinesis,
damage to ciliated epithelial cells, and altered alveolar macrophage function, all of which
may contribute to altered mucociliary transport and/or clearing infectious and
noninfectious particles from the lung. Altered alveolar macrophage function and other
potential mechanisms, such as increases in pro-inflammatory mediators and cytokines,
increased immunoglobulin E (IgE) concentrations, interactions with allergens, and altered
lymphocyte subsets, reflect modification of innate and/or adaptive immunity. These
changes may underlie susceptibility to infection, which has been observed in animals
exposed to NO2 (Section 5.2.5.1). Furthermore, dosimetric considerations suggest that
inhaled NO2 may diffuse partly through the ELF in the tracheobronchial region before
reacting with ELF solutes (Section 4.2.2.1.3). Diffusion may allow NC>2 to reach cilia
before transforming into other products. Thus, NC>2 may impact the structure or function
of cilia.
Controlled human exposure studies have demonstrated reduced mucociliary activity due
to depressed ciliary function, depressed phagocytic activity, and superoxide production in
alveolar macrophages, and altered humoral- and cell-mediated immunity following
exposure to 1,500-4,000 ppb NC>2 for a few hours [(Frampton et al.. 2002; Devlin et al..
1999; Helledav et al.. 1995; Sandstrom et al.. 1992a: Sandstrom et al.. 1992b: Sandstrom
et al.. 1991); Section 5.2.5.4]. Studies involving repeated daily exposure to 1,500 ppb
NC>2 (but not 600 ppb NCh) found reductions in lymphocyte subpopulations (Sandstrom
et al.. 1992b: Rubinstein et al.. 1991; Sandstrom et al.. 1990). Furthermore, repeated daily
4-42
-------
exposure to 2,000 ppb NCh resulted in upregulation of ICAM-1 in bronchial biopsy
specimens (Pathmanathan et al. 2003). These findings suggest a potential mechanism
underlying susceptibility to viral infection because ICAM-1 is a major receptor for
rhinoviruses and respiratory syncytial viruses. Finally, enhanced susceptibility of airway
epithelium to influenza viral infection was suggested in a study involving exposure to
1,000-3,000 ppb NO2 over 3 days, although the study results did not achieve statistical
significance (Goings etal.. 1989). Humans exposed to 600 and 1,500 ppb NO2 for
3 hours exhibited an increased injury response, as measured in bronchial epithelial cells,
resulting from influenza and respiratory syncytial virus (Frampton et al., 2002).
Epidemiologic evidence for associations between exposure to NO2 and increased
respiratory infections is somewhat inconsistent (Sections 5.2.5.2 and 5.2.5.3).
As summarized in the 2008 ISA and 1993 AQCD for Oxides of Nitrogen (U.S. EPA.
2008a. c, 1993a). studies in NC>2-exposed animals (500-10,000 ppb) have demonstrated
altered mucociliary clearance and several changes in alveolar macrophages. These
changes include morphological evidence of damage to alveolar macrophages (membrane
bleb formation and mitochondrial damage), decreased viability, and decreased function
(decreased superoxide production, decreased phagocytic capacity, and decreased
migration towards a stimulus) (Robison et al.. 1993; Davis etal.. 1992; Rose et al..
1989b: Schlesinger et al.. 1987; Schlesinger and Gearhart. 1987; Suzuki et al.. 1986;
Greene and Schneider. 1978; Powell etal.. 1971). A recent study involving exposure to
20,000 ppb NC>2 demonstrated nitration of SP-D, a surfactant protein that functions as a
collectin (Matalon et al.. 2009). Nitration was accompanied by cross-linking and a
decrease in SP-D aggregating activity, which could impact the role of SP-D in microbial
clearance and surfactant metabolism. Infectivity models have shown increased mortality
and decreased bactericidal activity (U.S. EPA. 2008c: Jakab. 1987; Miller et al.. 1987;
Ehrlich. 1980; Ehrlich et al.. 1977) as a result of NC>2 exposure (Sections 5.2.5.1 and
6.2.7).
Collectively, these studies in animals and humans demonstrate that NO2 exposure
impaired host defense. Responses in humans included reduced mucociliary activity and
altered humoral and cell-mediated immunity, while responses in experimental animals
included changes in alveolar function and SP-D nitration. Many of these studies involved
ambient-relevant exposures to NCh.
4.3.2.6.2 Exacerbation of Allergic Airway Disease
Inhaled allergens activate an acute immune response in allergen-sensitive individuals.
This response is characterized by early and late phases. Key players in the early asthmatic
response are mast cells and basophils, which release mediators following allergen binding
4-43
-------
to IgE receptors on their cell surfaces. These mediators include histamine and cysteinyl
leukotrienes, which bind airway smooth muscle receptors and induce contraction.
Mediators also activate T lymphocyte subsets (i.e., CD4+ T cells), resulting in the release
of Th2 cytokines that can recruit mast cells and cause airway smooth muscle contraction.
Th2 cytokines also promote the influx and activation of eosinophils and neutrophils.
Airway mucosal eosinophilia is characteristic of asthma and rhinitis. Eosinophils exert
their effects via degranulation and cytolysis, resulting in release of ECP and other
mediators (Erjefalt et al.. 1999). Th2 cytokines also activate B lymphocytes, resulting in
the production of allergen-specific IgE. These responses initiated by Th2 cytokines
contribute to the late asthmatic response, which is characterized by airway obstruction
generally occurring 3-8 hours following an antigen challenge (Cockcroft and Davis.
2006c). and to other responses occurring greater than >8 hours following an antigen
challenge.
Exogenous Nitrogen Dioxide
As summarized in the 2008 ISA (U.S. EPA. 2008c) and in Section 5.2.2.1. exposure to
NC>2 affects the acute immune response to inhaled allergens. Several controlled human
exposure studies found that NO2 exposure enhanced airway responsiveness to specific
stimuli, such as house dust mite (HDM) allergen (Jenkins et al.. 1999; Tunnicliffe et al..
1994) in subjects with mild allergic asthma. Further, repeated exposure to NO2 resulted in
an enhanced response to a dose of allergen that was asymptomatic when given alone
(Strand et al.. 1998). Airway responses were measured during the first 2 hours after
allergen challenge which falls within the timeline of the early phase asthmatic response.
These results provide evidence that NO2 exposure exacerbates the early phase asthmatic
response to allergen challenge, as measured by enhanced contraction of airway smooth
muscle cell.
Controlled human exposure studies also demonstrated that NCh exposure exacerbated the
late phase asthmatic response to allergen challenge in subjects with mild allergic asthma.
Airway obstruction, measured as a spontaneous fall in FEVi occurring after resolution of
the early asthmatic response (generally 3-8 hours after an antigen challenge), was
observed in subjects with asthma exposed to 400 ppb NO2 for 1 hour (Tunnicliffe et al..
1994) and to 250 ppb NC>2 for 30 minutes for 4 consecutive days (Strand et al.. 1998).
Other studies measured cell counts and mediators characteristic of the late phase
asthmatic response. Increased numbers of neutrophils and increased levels of ECP in
BAL and BL fluid, both indicators of inflammatory response to allergen challenge, were
reported following exposure to 260 ppb NC>2 for 15-30 minutes (Barck et al.. 2005a;
Barck et al.. 2002). Furthermore, increased ECP levels were observed in sputum and
blood, and an increase in myeloperoxidase (indicator of neutrophil activation) was seen
4-44
-------
in blood. In subjects with allergic rhinitis, NC>2 exposure (400 ppb for 6 hours) increased
eosinophil activation, measured by ECP in nasal lavage, following nasal allergen
provocation (Wang et al., 1995a). These studies suggest that exposure to NO2 may prime
eosinophils for subsequent activation by an allergen in previously sensitized individuals
(Davies et al., 1997; Wang et al., 1995b). However, another study found decreased
sputum eosinophils 6 hours after HDM challenge in subjects with HDM-sensitive allergic
asthma exposed to 400 ppb NO2 for 3 hours (Witten et al., 2005).
Late phase allergic responses were also investigated in animal models of allergic airway
disease (Section 5.2.2.5). Increased specific immune response to HDM allergen,
including enhanced antigen-specific serum IgE, and increased lung inflammation were
demonstrated in Brown Norway rats sensitized to and challenged with HDM allergen
followed by 3-hour exposure to 5,000 ppb NC>2 (Gilmour et al., 1996). Similarly, a recent
study showed that NO2 exposure (25,000 ppb, 6 hour/day for 3 days) increased the degree
and duration of the allergic inflammatory response in mice sensitized and challenged with
ovalbumin (Poynter et al.. 2006). Both neutrophilic and eosinophilic airway inflammation
were found in these studies; exposure of mice to a lower concentration of NO2
(5,000 ppb) failed to induce this response. Two other studies in ovalbumin-sensitized and
ovalbumin-challenged mice found decreased eosinophilic inflammation in response to
5,000 ppb NC>2; however, one of these studies found an increase in eosinophils following
exposure to 20,000 ppb NO2 (Hubbard et al.. 2002; Proust et al.. 2002). This increase in
eosinophils was accompanied by increased levels of eosinophil peroxidase, a marker of
activation. Both responses were observed 3 days, but not 1 day, after the 3-hour NO2
exposure. These results in animal models provide some evidence of NCh-mediated
enhancement of late phase allergic responses, albeit in many cases at higher
concentrations than those considered ambient relevant. It is important to note that
eosinophil activation and eosinophil influx reflect different processes and that only the
study by Hubbard et al. (2002) measured markers of activation. The ovalbumin-sensitized
and ovalbumin-challenged mouse model may not mimic the eosinophil degranulation and
cytolysis that are characteristic of asthma and allergic rhinitis in humans (Malm-Erjefalt
et al.. 2001). Hence, interspecies differences may account for the differing results of
animal and controlled human exposure studies.
Collectively, these studies demonstrate that inhaled NO2 enhanced both early and late
phase responses to inhaled allergens in humans with asthma and allergy. Furthermore,
exposure to NO2 augmented allergic inflammation in some rodent models of allergic
airway disease. Many, but not all, of these studies involved ambient-relevant exposures to
NO2. These results provide evidence for NCh-induced exacerbation of allergic airway
disease in the presence of an allergen challenge. Evidence for NCh-induced airway
4-45
-------
eosinophilia in the absence of an allergen challenge was described in Section 4.3.2.3.
Hence, NO2 exposure may lead to asthma exacerbations by multiple pathways.
Endogenous Nitrogen Dioxide
Several recent animal toxicological studies have explored the role of endogenous NO and
peroxynitrite, the latter of which decomposes to form NO2, on allergic airway disease in
animal models. The decomposition of peroxynitrite to NC>2 is one of the three pathways
contributing to the formation of endogenous NC>2 (Section 4.3.1). In one study,
upregulating the enzyme endothelial nitric oxide synthase (eNOS; and presumably NO
production) decreased airway inflammation, airway remodeling, and airway
responsiveness in a mouse model of asthma (Ahmad etal.. 2011). Asthma
phenotype-related features, such as cell infiltrates, mucus hypersecretion, peribronchial
collagen, and Th2 cytokines, were also diminished. Further, decreased inducible nitric
oxide synthase (iNOS) expression and 3-nitrotyrosine immunostaining in the airway
epithelium were reported, as were diminished epithelial injury and apoptosis. Because
3-nitrotyrosine is a marker of NO2/peroxynitrite formation, these findings suggest that an
increase in NO may have resulted in reduced peroxynitrite. While it is known that NO
rapidly reacts with superoxide to form peroxynitrite and that superoxide levels are
increased in inflammation, it is also known that excess NO will react with peroxynitrite
and quench peroxynitrite's reactivity. In fact, Stenger et al. (2010) found that high
concentrations of inhaled NO (10,000 ppb) prevented the formation of 3-nitrotyrosine in
the lungs of neonatal mice exposed to hyperoxia.
In a second set of studies, increased levels of the NOS substrate L-arginine were found to
decrease airway inflammation and airway responsiveness in a guinea pig model of
asthma (Maarsingh et al.. 2009). Similarly, increased L-arginine levels reduced
peroxynitrite formation and airway responsiveness in a mouse model of asthma
(Mabalirajan et al., 2010b). Markers of allergic inflammation (e.g., eosinophilia and Th2
cytokines), markers of oxidative and nitrative stress, and markers of airway remodeling,
such as goblet cell metaplasia and subepithelial fibrosis, were also decreased. Further,
increased L-arginine levels reduced mitochondrial dysfunction and airway injury
(Mabalirajan et al.. 2010a). Limited availability of L-arginine is known to uncouple NOS
enzyme activity, resulting in the production of superoxide in addition to NO. This
situation is commonly found in disease models and leads to peroxynitrite formation.
Increasing L-arginine availability is a common strategy used to prevent enzyme
uncoupling and peroxynitrite formation. Another approach was employed in a study by
North et al. (2009) where inhibition of the enzyme arginase 1 (arginase 1 decreases
arginine availability) was found to decrease airway responsiveness in a mouse model of
asthma. Similar findings were reported using arginase inhibition in a guinea pig model of
4-46
-------
allergic asthma where arginase was upregulated (Maarsingh et al., 2008). Inhibition of
arginase resulted in amelioration of the asthma phenotype. These effects were attributed
to decreased enzyme uncoupling, thus promoting the formation of NO, diminishing the
generation of superoxide, and reducing the formation of peroxynitrite. In contrast, a
different study found that arginase inhibition resulted in increased S-nitrosylated and
nitrated proteins, increased inflammation, mucous cell metaplasia, NFxB activation, and
increased airway responsiveness in a mouse model of asthma (Ckless et al., 2008).
However, the same study also found reduced levels of antigen-specific IgE and IL-4.
Thus, only some features of the asthma phenotype were ameliorated by arginase
inhibition. The authors suggested that peroxynitrite, whose presence was indicated by the
increase in nitrated proteins in mice treated with arginase, may have contributed to
increased airway responsiveness in this model.
Evidence for similar pathways in humans is provided by a study in which endogenous
markers of reactive nitrogen and oxygen chemistry were measured in individuals with
and without asthma (Anderson et al. 2011). Levels of total nitrite and nitrate were higher
in the BAL fluid of subjects with asthma compared to healthy subjects. In subjects with
asthma, upregulation of iNOS was observed, and it was greater in distal airways
compared with more proximal airways. In addition, levels of dihydroethidium-positive
cells, which are capable of producing ROS (such as superoxide), were higher in both the
BL and BAL fluid of subjects with asthma compared with healthy subjects. Levels of
arginase were also higher in BAL fluid of subjects with asthma compared with healthy
subjects. These results suggest that uncoupling of NOS and/or NOS dysfunction,
resulting in enhanced peroxynitrite/NO2 formation, may contribute to the asthma
phenotype in human subjects. These findings also provide biological plausibility for
results of another study demonstrating a correlation between increased airway
responsiveness and the induction of iNOS, the induction of arginase, and the production
of superoxide in subjects with asthma.
Collectively, these studies provide evidence that the balance between endogenous NO
and peroxynitrite influenced features of the asthma phenotype in animal models of
allergic airway disease and possibly in adults with asthma. Enhanced levels of
superoxide, which are characteristic of asthma and other inflammatory states, favored the
formation of peroxynitrite at the expense of NO. Evidence from experimental studies
indicated that peroxynitrite and other RNS are found in and contribute to allergic airway
disease in animal models. Thus, inhaled NO2 may exacerbate allergic airway disease by
adding to the lung burden of RNS in inflammatory states.
4-47
-------
4.3.2.6.3 T-Helper Cell 2 Skewing and Allergic Sensitization
A controlled human exposure study demonstrated that repeated daily exposures of
healthy adults to NC>2 resulted in increased expression of the interleukins IL-5, IL-10,
IL-13, and the protein ICAM-1 in the respiratory epithelium following the last exposure
[(Pathmanathan et al.. 2003); Section 5.2.7.41. These interleukins are characteristic of a
Th2 inflammatory response. IL-5 is known to promote eosinophilia, while IL-13 is
known to promote mucus production and AHR (Bevelander et al., 2007). These findings
suggest a potential mechanism whereby repeated exposure to NCh may exert a
pro-allergic influence. Further, upregulation of ICAM-1 suggests a potential mechanism
for leukocyte influx. A separate study by these same investigators found persistent
neutrophilic inflammation following the 4 days of repeated exposure (Blomberg et al.,
1999).
In addition, two studies in animals examined the effects of longer term exposures to NO2
on the development of allergic responses (Sections 5.2.7.4 and 6.2.2.3). In one study,
exposure of guinea pigs to 3,000 or 9,000 ppb NCh increased the numbers of eosinophils
in nasal epithelium and mucosa after 2 weeks (Ohashi et al.. 1994). In the other, exposure
to 4,000 ppb NCh for 12 weeks led to enhanced IgE-mediated release of histamine from
mast cells isolated from guinea pigs (Fujimaki and Nohara. 1994). This response was not
found in mast cells from rats similarly exposed. Both studies provide further evidence for
NC>2 having a pro-allergic influence.
Furthermore, a recent study in mice provides evidence that NC>2 may act as an adjuvant
promoting the development of allergic airway disease in response to a subsequent
inhalation exposure to ovalbumin (Bevelander et al.. 2007). Findings included AHR,
mucous cell metaplasia, and eosinophilic inflammation. In addition, ovalbumin-specific
IgE and IgGl, CD4+ T cells biased toward Th2, and a T helper cell 17 (Thl7) phenotype
in the blood were demonstrated. These results are consistent with an allergic asthma
phenotype in humans. The eosinophilic inflammation, mucus gene upregulation, and
ovalbumin-specific IgE production were found to be dependent on TLR2 and myeloid
differentiation primary response gene (88) pathways. TLR2 is known to promote
maturation of dendritic cells, inflammation, and Th2 skewing. A subsequent study in the
same model found that NO2 exposure had several effects on pulmonary CD1 lc+ dendritic
cells, including increased cytokine production, upregulation of maturation markers,
increased antigen uptake, migration to the lung-draining lymph node, and improved
ability to stimulate naive CD4+ T cells (Hodgkins et al., 2010). Dendritic cells are key
players in adaptive immune responses by regulating CD4+-mediated T cell responses
through the presentation of antigens in the draining lymph node. Further, dendritic cells
can express a distinct pattern of co-stimulatory molecules and produce cytokines that
4-48
-------
create an environment for T cell polarization, thus skewing the T helper cell response.
Changes reported in these two studies are consistent with the promotion of allergic
sensitization and suggest a role for TLR2 in mediating this effect. A third study by these
same investigators found that NO2 exposure resulted in antigen-specific IL-17A
generation from ThlV cells, which is characteristic of the severe asthma phenotype that is
unresponsive to glucocorticoid treatment in humans (Martin etal.. 2013). Although all
studies involved 1-hour exposures to high concentrations of NC>2 (10,000-15,000 ppb),
they are included here because they describe potentially new mechanisms by which
inhaled NO2 may exert its effects. It should additionally be noted that airway
inflammation was seen in mice exposed to 15,000 ppb, but not to 10,000 ppb, NC>2 for
1 hour and that pulmonary damage was minimal in this model (Martin et al., 2013).
In contrast, a similar study failed to find that NC>2 acted as an adjuvant in a mouse model
of allergic airway disease (Alberg et al.. 2011). The exposure consisted of 5,000 or
25,000 ppb NC>2 for 4 hours and followed exposure to ovalbumin which was administered
intra-nasally. Adjuvant activity was measured as the production of allergen-specific IgE
antibodies. Methodological differences in study design with respect to the timing of
ovalbumin and NCh exposures and the route of ovalbumin exposure may account for
differences in findings between this study and others. In fact, Bevelander et al. (2007)
found that NO2 promoted allergic sensitization when NC>2 exposure occurred before, but
not after, ovalbumin exposure.
It has been hypothesized that both endogenous and exogenous ROS and/or RNS can alter
the balance between tolerance and allergic sensitization from an inhaled agent (Ckless et
al.. 2011). Some activities of dendritic cells and T cells, such as maturation of the antigen
presenting capacity of dendritic cells, dendritic cell stimulation of CD4+ T cells, and
polarization of T cells, are redox sensitive. Endogenous ROS and RNS are produced by a
variety of respiratory tract cells, including epithelial cells, dendritic cells, T lymphocytes,
macrophages, neutrophils, and eosinophils, especially during inflammation. Peroxynitrite
formation, myeloperoxidase activity and/or nitrite acidification may also be enhanced
during inflammation and contribute to endogenous NCh levels. ROS and RNS are
thought to promote the allergic phenotype. Air pollution-derived exogenous ROS and
RNS can potentially contribute to oxidative and/or nitrative stress in the respiratory tract
and influence the adaptive immune response that occurs once dendritic cells are
activated. Thus, recent studies suggest the possibility of an interaction between inhaled
NO2 and the NO2 endogenously formed in the respiratory tract.
Collectively, these studies in humans and animals provide evidence that NO2 exposure
can lead to the development of allergic responses in nonallergic individuals or animals
via Th2 skewing and allergic sensitization. Many, but not all, of these studies involved
4-49
-------
ambient-relevant exposures to NCh. It should also be noted that increased prevalence of
allergic sensitization was found in a few epidemiologic studies in relation to NCh
exposure (Section 6.2.4.1).
4.3.2.7 Remodeling of Airways and Alveoli
As summarized in the 2008 ISA and the 1993 AQCD for Oxides of Nitrogen (U.S. EPA.
2008a. c, 1993a). numerous studies have examined morphological changes in the
respiratory tract resulting from chronic NC>2 exposure. The sites and types of
morphological lesions produced by exposure to NCh were similar in all species when
effective concentrations were used (U.S. EPA, 1993a). The centriacinar region, including
the terminal conducting airways, the alveolar ducts, and the alveoli, exhibited the greatest
sensitivity to NC>2 exposure, while the nasal cavity was minimally affected. The cells
most injured in the centriacinar region were the ciliated cells of the bronchiolar
epithelium and the Type I cells of the alveolar epithelium. These were replaced with
nonciliated bronchiolar cells and Type II cells, respectively, which were relatively
resistant to continued NO2 exposure. Some lesions rapidly resolved post-exposure. One
study found that collagen synthesis rates were increased in NO2-exposed rats. Because
collagen is an important structural protein in the lung and because increased total lung
collagen is characteristic of pulmonary fibrosis, it was proposed that NO2 exposure may
cause fibrotic-like diseases. Exposure to NO2 also enhanced pre-existing emphysema-like
conditions in animal models (U.S. EPA. 2008c). Other studies demonstrated that NO2
exposure induced air space enlargements in the alveolar region and suggested that
chronic exposures could result in permanent alterations resembling emphysema-like
diseases (U.S. EPA. 1993a). A recent study confirmed and extended these findings. NO2
exposure in rats (10,000 ppb for 21 days) caused increased apoptosis of alveolar
epithelial cells and enlargement of air spaces (Fehrenbach et al.. 2007). Further, alveolar
septal cell turnover was increased, and changes in extracellular matrix were reported.
However, there was no loss of alveolar walls (i.e., total alveolar wall volume or total
alveolar surface area), indicating that the lesions induced did not meet the 1985 National
Heart, Lung, and Blood Institute definition of human emphysema (U.S. EPA. 1993a).
A chronic-duration study in rats exposed to 9,500 ppb NO2 for 7 hour/day, 5 days/week
for 24 months found an additional effect on morphology (Mauderly et al., 1990).
Bronchiolar epithelium was observed in centriacinar alveoli, and this response progressed
with increasing length of exposure. This has been termed "alveolar bronchiolization"
(Nettesheim et al.. 1970). reflecting the replacement of one type of epithelium by another.
Long-term consequences of alveolar bronchiolization are not known.
4-50
-------
The relationship between NO2 exposure-induced morphologic changes in animal models
and impaired lung development seen in epidemiological studies is not clear. Effects of
NO2 exposure on lung morphology in rats have been shown to be age dependent (U.S.
EPA. 2008a. c, 1993a). Six-week-old rats exposed to NCh for 6 weeks were more
sensitive to the effects of NO2 exposure than 1-day-old rats exposed for 6 weeks (Chang
et al.. 1986). In humans, the respiratory and immune systems are immature in newborns,
and the respiratory system continues to develop until about 20 years of age. This suggests
the potential for NO2 exposure-induced permanent morphological changes in humans if
exposure should occur during critical windows of development. However, experimental
evidence to substantiate this claim is currently lacking.
Evidence from animal models of allergic airway disease suggests a role for endogenous
NO2 in airway remodeling. These studies, described above in Section 4.3.2.6.2. found
that decreased NO bioavailability during inflammation favored the formation of
peroxynitrite, which decomposes to NC>2. Interventions that reduced peroxynitrite
formation, as evidenced by decreased 3-nitrotyrosine immunostaining, resulted in an
amelioration of airway remodeling, as measured by mucus hypersecretion, peribronchial
collagen, goblet cell metaplasia, subepithelial fibrosis, and epithelial apoptosis (Ahmad et
al.. 2011; Mabalirajan et al.. 2010b). Exposure to inhaled NCh was found to enhance
allergic airway inflammation and airway responsiveness in experimental animals
previously sensitized and challenged with an allergen (Poynter et al.. 2006). Airway
remodeling was not evaluated in this study which involved acute exposures to NC>2.
Whether repeated or chronic exposures to NO2 lead to airway remodeling in the context
of allergic airway disease is not known. However, in nonallergic guinea pigs, subchronic
exposure to NO2 (60-4,000 ppb, 24 hour/day, 6-12 weeks) enhanced both airway
responsiveness and specific airway resistance, suggesting that airway remodeling may
have contributed to the development of AHR (Kobayashi and Miura. 1995).
Collectively, these studies in animals demonstrate that NO2 exposure altered respiratory
tract morphology. This included lesions in the centriacinar region and the alveolar region.
Many, but not all, of these studies involved higher than ambient-relevant exposures to
NC>2. A role for endogenous NCh in airway remodeling has also been demonstrated in
allergic animals. Increased specific airway resistance, a physiologic change, suggests that
subchronic exposure to ambient-relevant concentrations of NC>2 may also lead to airway
remodeling.
4-51
-------
4.3.2.8 Potential Induction of Carcinogenesis
Some studies have explored the potential carcinogenicity of NC>2. There is no clear
evidence that NO2 acts as a carcinogen [(U.S. EPA, 2008a. c, 1993a); Section 6.6.91.
However, NO2 may act as a tumor promoter at the site of contact, possibly due to its
ability to produce cellular damage and promote regenerative cell proliferation. In
addition, it has been shown to be genotoxic and mutagenic in some systems, including
human nasal epithelial mucosa cells ex vivo exposed to urban-level concentrations
[100 ppb; (Koehleretal.. 2011. 2010)]. Some studies demonstrated that inhaled NO2 at
high concentrations (e.g., 20,000 ppb) can contribute to the formation of mutagens and
carcinogens if other precursor chemicals are found in the body (e.g., N-nitrosomorpholine
from morpholine and nitro-pyrene from pyrene) [(U.S. EPA. 2008c): Section 4.2.2.5].
4.3.2.9 Transduction of Extrapulmonary Responses
While the respiratory tract has been viewed as the primary target of inhaled NC>2, effects
outside the respiratory tract have been demonstrated in numerous controlled human
exposure and toxicological studies (U.S. EPA. 2008a. c, 1993a). These include
hematological effects and effects on the heart, central nervous system, liver, and kidneys
and on reproduction and development. Epidemiologic evidence of associations between
NC>2 exposure and some extrapulmonary effects has also been described (Sections 5.3.
63 and
Some NO2-induced effects that have been demonstrated are briefly described here. Two
controlled human exposure studies involving NCh inhalation over several hours found
effects on circulating red blood cells, including reduced hemoglobin and hematocrit
levels; one of these also found reduced acetylcholinesterase activity [(Frampton etal..
2002; Posin etal.. 1978); Section 5.3.10]. Changes in lymphocyte numbers and subsets in
the peripheral blood have been demonstrated in human subjects following exposure to
NC>2 (Frampton et al.. 2002; Sandstrom et al.. 1992a). A recent controlled human
exposure study found altered blood lipids (Huang et al.. 2012b). Studies in experimental
animals have demonstrated decreases in red blood cell number as well as increases in
diphosphoglycerate, sialic acid, and methemoglobin following several days of NO2
exposure (Section 5.3.10). However, changes in hematocrit and hemoglobin did not occur
following longer term exposure to NCh. Increases in blood glutathione levels and altered
blood lipids resulting from NC>2 exposure have also been reported (U.S. EPA. 2008c).
More recent studies in rats exposed for 7 days to NCh (2,660 or 5,320 ppb NCh) have
shown mild pathology of brain and heart tissue, which was accompanied by markers of
inflammation and/or oxidative stress [(Li etal.. 2012a; Li etal.. 201 la); Section 5.3.10].
4-52
-------
In addition, animal studies demonstrated reproductive and developmental effects
resulting from exposure to NCh during gestation. These included decreased litter size and
neonatal weight and effects on postnatal development (Section 6.4).Many, but not all, of
these extrapulmonary effects in animal models have been observed at concentrations of
NC>2 that are higher than ambient-relevant concentrations.
Given the reactivity of NO2, extrapulmonary effects are likely due to NCh-derived
reaction products rather than to NCh itself. One pathway by which a reaction product
could mediate extrapulmonary effects of NO2 would be the activation of pulmonary
irritant receptors that results in cardiovascular reflex responses (Section 4.3.2.2).
Evidence suggests that the reduction in heart rate observed after acute exposure of
experimental animals to high concentrations of NC>2 may be due to stimulation of
pulmonary irritant receptors. However, much weaker evidence exists for activation of
pulmonary irritant receptors in humans because studies have observed no increases in
respiratory rate or decreases in heart rate, which are indicators of pulmonary irritant
receptor involvement. A recent controlled human exposure study found altered heart rate
variability following exposure to an ambient-relevant concentration of NC>2 [500 ppb;
(Huang et al.. 2012b)]: whether this effect was due to pulmonary irritant receptor
stimulation is unclear.
Alternatively, NC>2-derived reaction products in the lung may diffuse or migrate into the
circulation. One reaction product of inhaled NO2, nitrite, is known to gain access to the
circulation. In the presence of red blood cell hemoglobin, nitrite is oxidized to nitrate
(Postlethwait and Mustafa. 1981) and nitrosylhemoglobin and methemoglobin are
formed. Nitrite has known effects on blood cells, vascular cells, and other tissues. Much
recent attention has been paid to nitrite's systemic vasodilatory effects that occur under
hypoxic conditions. As discussed in the 2008 ISA and the 1993 AQCD for Oxides of
Nitrogen (U.S. EPA. 2008a, c, 1993a). one controlled human exposure study
demonstrated that NO2 exposure (4,000 ppb, 75 minutes, intermittent exercise) resulted in
a reduction in blood pressure (Linn et al.. 1985b). which is consistent with the systemic
vasodilatory properties of nitrite under conditions of low oxygen. However, studies from
other laboratories did not see this effect (Section 5.3.6.2). Furthermore, dosimetric
considerations suggest that contributions of nitrite derived from ambient NO2 to plasma
levels of nitrite are small compared to nitrite derived from dietary sources
(Section 4.2.2.4).
Besides nitrite and nitrate, other NO2-derived reaction products may translocate to the
circulation. The formation of fatty acid epoxides followed by transport to the circulation
and then to the liver was postulated to explain the effect of NO2 exposure (250 ppb,
3 hours) on pentobarbital-induced sleeping time in mice (Miller et al.. 1980). Findings of
4-53
-------
lipid peroxidation and markers of oxidative stress in some animal studies (Li et al..
2012a: Li et al.. 201 la), which utilized higher than ambient-relevant concentrations of
NC^ also suggest the presence of circulating ROS or RNS. However, there is no
experimental evidence to date for the translocation of NCh-derived ROS or RNS to the
circulation following NCh exposure.
A third pathway by which a NC>2-derived reaction product may transduce extrapulmonary
responses is the diffusion or migration of inflammatory or vasoactive mediators from the
lung into the circulation. This possibility is consistent with changes in peripheral blood
inflammatory cells and in tissue markers of inflammation that have been observed
following exposure to NC>2. Confirmation that this mechanism occurs in human subjects
exposed to ambient-relevant concentrations of NO2 was provided by a recent study
[(Channell et al.. 2012); Section 5.3.10.41. Exposure of healthy human subjects to NC>2
(500 ppb for 2 hours) resulted in circulating pro-inflammatory factors in the plasma.
Application of plasma to cultured endothelial cells resulted in upregulation of ICAM-1
and vascular cell adhesion molecule 1, as well as the release of IL-8 into the supernatant
of the cultured cells. Furthermore, the amount of soluble lectin-like receptor for oxidized
low-density lipoprotein (sLOX) was increased in plasma obtained 24 hours
post-exposure. Changes in plasma high density lipoprotein levels were observed in a
separate study employing the same exposure parameters (Huang et al.. 2012b). These
findings point to a pathway by which inhaled NC>2 leads to circulating soluble factors that
promote inflammatory signaling in the vasculature.
Collectively, these studies demonstrate that NO2 exposure resulted in extrapulmonary
effects in humans and animals. Responses included altered blood lipids and cells, altered
pentobarbital-induced sleeping time, altered heart rate variability, increased plasma
sLOX, and changes in the brain and heart. Mechanisms underlying these responses are
unclear although there is evidence for circulating pro-inflammatory factors. Many of
these studies were conducted using ambient-relevant concentrations of NC>2.
4.3.3 Nitric Oxide
As summarized in the 2008 ISA and 1993 AQCD for Oxides of Nitrogen (U.S. EPA.
2008a. c, 1993a) and a recent review (Hill et al.. 2010). the synthesis of endogenous NO
in cells is catalyzed by three different isoforms of NOS (eNOS, iNOS, neuronal NOS).
NO is involved in intra-cellular signaling in virtually every cell and tissue. In general,
low levels of endogenous NO play important roles in cellular homeostasis, while higher
levels are important in cellular adaptation and still higher levels are cytotoxic. Further,
4-54
-------
signaling functions of NO may be altered in the presence of acute inflammation (Hill et
al.. 2010V
Like NO2, NO is a radical species [(Fukuto et al.. 2012): Table 4-2]. However, it is more
selectively reactive than NO2 (Hill etal.. 2010). In addition, it is more hydrophobic and
can more easily cross cell membranes and diffuse much greater distances compared with
NO2. As a result, there may be overlap between endogenous and exogenous NO in terms
of biological targets and pathways. The following discussion focuses on common
mechanisms underlying the effects of both endogenous and exogenous NO.
Because NO has a high affinity for heme-bound iron, many of its actions are related to its
interactions with heme proteins (Hill et al.. 2010). For example, activation of the heme
protein, guanylate cyclase, is responsible for smooth muscle relaxation and vasodilation
of pulmonary and systemic vessels and possibly for bronchodilator effects. Inhaled NO
rapidly reacts with soluble guanylate cyclase in the pulmonary arterial smooth muscle. At
the same time, inhaled NO rapidly diffuses into the circulation and reacts with red blood
cell hemoglobin to form nitrosylhemoglobin, which is subsequently oxidized to
methemoglobin and nitrate. Increased blood concentrations of nitrosylhemoglobin and
methemoglobin have been reported in mice exposed for 1 hour to 20,000-40,000 ppb
NO, as well as in mice exposed chronically to 2,400 and 10,000 ppb NO (U.S. EPA.
1993a). Some S-nitrosohemoglobin may be formed in partially deoxygenated blood
(Wennmalm et al., 1993). NO can also disrupt iron-sulfur centers in proteins (Hill et al..
2010). Furthermore, redox reactions of NO and transition metals, such as iron and
copper, facilitate S-nitrosylation of protein and nonprotein thiols. Binding of NO to
iron- and copper-containing proteins in the mitochondria may play an important role in
mitochondrial respiration. NO also rapidly reacts with superoxide, an oxygen-derived
radical species, to produce the potent oxidant peroxynitrite (Hill et al.. 2010).
Peroxynitrite subsequently reacts with CO2 to form the nitrosoperoxylcarbonate anion,
followed by decomposition to carbonate radical and NO2 (Section 4.2.2.4).
Endogenous NO is formed in the respiratory tract at high levels (Section 4.2.3). and it has
physiologic functions. The paranasal sinuses are a major source of NO in air derived
from the nasal airways, with average levels of 9,100 ppb NO (n = 5) measured in the
sinuses (Lundberg et al.. 1995). Expression of iNOS was found to be higher in epithelial
cells of the paranasal sinuses than in epithelial cells of the nasal cavity. This NO
produced by nasal airways is thought to play a role in sinus host defense through
bacteriostatic activity. In addition, NO produced by nasal airways was found to modulate
pulmonary function in humans through effects on pulmonary vascular tone and blood
flow (Lundberg et al.. 1996). In healthy subjects, a comparison of nasal and oral
breathing demonstrated that nasal airway NO enhanced transcutaneous oxygen tension.
4-55
-------
In intubated patients, nasal airway NO increased arterial oxygenation and decreased
pulmonary vascular resistance. Additionally, endogenous NO has been shown to act as a
bronchodilator (Belvisi et al., 1992). Endogenous NO produced at high concentrations by
phagocytic cells is also known to participate in the killing of bacteria and parasites; this
contributes to host defense (U.S. EPA. 2008c). Another effect of endogenous NO on host
defense is modulation of ciliary beat frequency (Jain etal.. 1993). Specifically, NO
derived from more distal airways was found to increase ciliary beat frequency.
Furthermore, endogenous NO production can be upregulated during inflammation
(Anderson et al.. 2011). In fact, induction of iNOS in proximal or distal airways of
subjects with asthma results in levels of NO in exhaled breath as high as 20-50 ppb
lYAlving et al.. 1993; Hamidetal.. 1993): Section 4.2.31.
Endogenous NO has known pro- and anti-inflammatory effects; thus its role in
inflammatory lung disease is not clear. While both eNOS and iNOS contribute to NO
production in the lung, the relatively low levels of NO produced by eNOS are thought to
be more important in metabolic homeostasis (Ahmad etal.. 2011). Some evidence points
to a role of iNOS-derived NO in the pathogenesis of asthma because it has been
correlated with inflammation, epithelial injury, and clinical exacerbations of asthma
[(Anderson et al.. 2011): Section 4.3.2.6.2]. Furthermore, iNOS was upregulated to a
greater extent in the distal airways than in more proximal airways in subjects with
asthma. This is of interest because asthma is a disease of the small airways. As mentioned
above, signaling functions of NO may be altered in the presence of acute inflammation
(Hill etal.. 2010). which is characterized by enhanced levels of superoxide. Superoxide
reacts with NO to form peroxynitrite, which has been shown in animal models to play a
role in the pathogenesis of allergic airway disease (Section 4.3.2.6.2).
NO exposure has been shown to alter pulmonary function, pulmonary morphology, and
vascular function (U.S. EPA. 2008a. c, 1993a). Studies in animals have demonstrated that
inhaled NO reverses acute methacholine-induced bronchoconstriction (Hogman et al..
1993; Dupuyetal.. 1992). This reversal was observed with exposures of 5,000 ppb NO in
guinea pigs and 80,000 ppb in rabbits. Chronic inhalation exposures have been found to
alter the morphology of the alveolar septal units in rats (Mercer et al.. 1995). This effect
was not seen with chronic inhalation exposures to NO2 at similar concentrations (500 ppb
with twice daily spikes of 1,500 ppb). In addition, inhaled NO has been shown to alter
transferrin and red blood cells in mice. Further, acute inhalation exposure of NO
decreased pulmonary vascular resistance in pigs and reduced pulmonary arterial pressure
in a rodent model of chronic pulmonary hypertension. A recent study also found that
inhaled NO (1,000, 5,000, 20,000, and 80,000 ppb) selectively dilated pulmonary blood
vessels, improved ventilation-perfusion mismatch, and reduced hypoxemia-induced
pulmonary vascular resistance in a pig model (Lovich et al., 2011).
4-56
-------
Inhaled NO is used clinically at concentrations higher than those that are environmentally
relevant. Although it can cause both pulmonary and systemic vasodilation, effects on
pulmonary vasculature occur at lower concentrations than those required for vasodilation
of systemic vessels. This selectivity for pulmonary vasculature is likely due to the rapid
scavenging of NO by hemoglobin in the blood. Hence, inhaled NO has been used to
mitigate pulmonary hypertension in newborns and adults. High concentrations of inhaled
NO are also known to alter ciliary beating and mucus secretion in the airways, to increase
renal output, to alter distribution of systemic blood flow, to alter coagulation, fibrinolysis,
and platelet functions, and to modulate the inflammatory response (U.S. EPA, 2008c).
Endogenous NO is an important mediator of cardiovascular homeostasis. It has
anti-inflammatory and antithrombotic effects, is cytoprotective, and induces antioxidant
defenses (Wang and Widlanskv. 2009). Two recent studies in animal models demonstrate
that high concentrations of inhaled NO may result in vascular toxicity. One of these
studies found rapid formation of plasma nitrites/nitrates in rats exposed for 1 hour to
3,000-10,000 ppb NO (Knuckles etal.. 2011). Plasma nitrites/nitrates doubled after an
hour of exposure to 3,000 ppb NO and tripled after an hour of exposure to 10,000 ppb
NO. These changes were accompanied by an enhanced constriction response to
endothelin-1 in coronary arterioles, which reflected altered vasomotor tone. Although this
latter effect appears to run counter to the vasodilator role of NO, it should be noted that
the high concentrations of NO used in this study are known to inhibit eNOS activity in
other models (Griscavage et al., 1995). The increase in aortic eNOS content reported is
consistent with enzyme inactivation and turnover. Another recent animal toxicological
study conducted in ApoE~'~ mice, a model of atherosclerosis, found that exposure to very
high concentrations of inhaled NO over the course of a week (17,000 ppb NO for
6 hour/day for 7 days) led to increases in messenger ribonucleic acid for aortic
endothelin-1 and matrix metalloproteinase (MMP)-9, as well as to enhanced vascular
gelatinase activity (Campen et al., 2010). These effects, which are biomarkers of vascular
remodeling and plaque vulnerability, were not seen with 2,000 ppb NO2. The authors
suggested that the activity of eNOS was uncoupled, resulting in oxidative stress due to
the production of superoxide instead of, or in addition to, NO. Both of these studies
suggest that inhaled NO has the potential to disrupt normal signaling processes mediated
by endogenous NO.
As mentioned above, endogenous NO plays key signaling roles in virtually every cell and
tissue (Hill etal., 2010) and, as such, is an important mediator of homeostasis. Inhaled
NO at high concentrations has the potential to have beneficial or deleterious effects on
multiple organ systems. An important consideration is whether effects are mediated by an
NO metabolite, by the release of NO from a metabolite that serves as a storage pool of
4-57
-------
NO, or through methemoglobin formation in the blood. Further discussion of the
biological functions of NO metabolites is found below.
4.3.4 Metabolites of Nitric Oxide and Nitrogen Dioxide
4.3.4.1 Nitrites/Nitrates
Rapid appearance of nitrite and nitrate in the blood was demonstrated in rats exposed for
1-2 hours to 5,000-40,000 ppb NO2 (Odaetal.. 1981). Elevated levels of blood nitrite
and nitrate were maintained as long as the exposure to NO2 continued. A small increase
in levels of nitrosylhemoglobin, but not methemoglobin, was detected in blood. The lack
of accumulation of methemoglobin was likely due to reduction of methemoglobin to
hemoglobin catalyzed by methemoglobin reductase. Two other studies measured
methemoglobin in the blood of mice exposed to NO2, with conflicting results (U.S. EPA.
1993a). Rapid formation of plasma nitrites/nitrates has also been demonstrated in rats
exposed for 1 hour to 3,000-10,000 ppb NO (Knuckles et al.. 2011).
Recently, it has been proposed that nitrite is a storage form of NO because it can be
reduced back to NO under conditions of low oxygen tension in a reaction catalyzed by
deoxyhemoglobin (Gladwin et al., 2005). In addition, nitrite is a signaling molecule in its
own right and does not require conversion to NO for this activity (Bryan. 2006). Nitrite
can increase cyclic guanosine monophosphate (cGMP) levels and heat shock protein 20
expression, decrease cytochrome P450 activity, and alter heme oxygenase-1 expression
(Bryan et al., 2005). Nitrite is also bactericidal (Major et al., 2010). Furthermore, under
acidic conditions, nitrite can react with thiols to form RSNOs. Nitrite also reacts with
hemoglobin to form iron-nitrosyl-hemoglobin and with oxyhemoglobin to form nitrate.
Nitrite acts as a vasodilator under hypoxic conditions, through a reaction catalyzed by
deoxyhemoglobin (Cosby etal., 2003). The venous circulation may be more sensitive to
nitrite than the arterial circulation (Maher et al.. 2008).
A recent study found that inhaled nitrite decreased pulmonary blood pressure in newborn
lambs with hemolysis-induced pulmonary vasoconstriction (Blood et al., 2011). Nitrite
was converted to NO in lung tissue by a mechanism that did not require reaction with
deoxyhemoglobin in the circulation. This mechanism resulted in increased exhaled NO
gas as well as the relaxation of vascular smooth muscle, which led to pulmonary
vasodilation. Although concentrations of inhaled nitrite employed were high (0.87 mol/L
sodium nitrite), this study is discussed here because it illustrates a novel biological
4-58
-------
activity of lung nitrite that is normally formed by reactions of NC>2 and NO in the ELF
and/or the blood.
4.3.4.2 S-Nitrosothiols
RSNOs are found endogenously in tissues and extracellular fluids. High concentrations of
RSNOs are found in the lung, and their levels may vary depending on disease status (Oue
et al.. 2009). For example, levels of RSNOs in BAL fluid were higher in individuals with
asthma compared with healthy subjects (Oue etal.. 2009). Transport of RSNOs from
extracellular compartments into isolated perfused lungs and cultured alveolar epithelial
cells occurs via a specific amino acid transport pathway (Torok et al., 2012; Brahmajothi
etal..201Q).
Some S-nitrosohemoglobin may be formed in partially deoxygenated blood following
inhalation of NO (Wennmahn et al.. 1993). However, inhaled NO mainly reacts with red
blood cell hemoglobin to form nitrosylhemoglobin, which is subsequently oxidized to
methemoglobin and nitrate (Hill et al.. 2010). The exact mechanisms by which RSNO
formation occurs are not completely clear (Fukuto etal.. 2012). NO does not react
directly with thiol groups, but it can form RSNOs via reactions with thiyl groups and
through intermediate formation of N2Os or metal nitrosyls, such as nitrosylhemoglobin
(Fukuto et al.. 2012; Hill etal.. 2010). Recent evidence suggests that NO may diffuse into
extracellular fluid and be transformed to RSNOs (Torok et al.. 2012; Brahmajothi et al..
2010). These experiments were conducted ex vivo in isolated perfused lungs and in vitro
in cultured lung epithelial cells, neither of which is a blood-perfused system. Hence it is
not clear whether this mechanism contributes to RSNO formation in vivo where the
majority of inhaled NO diffuses rapidly across the alveolar capillary barrier and binds to
hemoglobin.
RSNOs are thought to serve as a storage or delivery form of NO and to play a role in cell
signaling (Fukuto et al., 2012; Hill et al.. 2010). They may mediate protein
S-glutathionylation and thiol oxidation reactions that can act as redox switches to initiate
cell signaling events or alter enzyme activity (Hill etal. 2010).
In the lung, RSNOs act as endogenous bronchodilators (Oue et al.. 2009) and suppress
inflammation by decreasing activation of the transcription factor NFxB (Marshall and
Stamler. 2001). Furthermore, augmentation of airway RNSOs by ethyl nitrite inhalation
protected against lipopolysaccharide-induced lung injury in an animal model (Marshall et
al., 2009). Several findings suggest an inverse relationship between endogenous airway
RSNO levels and airway responsiveness. First, levels of airway S-nitrosoglutathione
levels were decreased in children with asthmatic respiratory failure and in adults with
4-59
-------
asthma (Que et al.. 2009; Gastonetal.. 1998). Second, the enzyme nitrosoglutathione
reductase (GSNOR), which regulates airway S-nitrosoglutathione content, was expressed
at higher levels in BAL cell lysates in human subjects with asthma than in healthy
subjects (Que et al.. 2009). GSNOR expression was inversely correlated with
S-nitrosoglutathione content. In addition, GSNOR activity in BAL fluid was increased
and was inversely correlated with airway responsiveness in human asthma (Que et al..
2009). Third, levels of airway RSNOs were inversely correlated with airway
responsiveness in human subjects with eosinophilic inflammation (Lee et al.. 201 Ib).
4.3.4.3 Nitrated Fatty Acids and Lipids
Nitration of unsaturated fatty acids and lipids can occur during inflammation and
ischemia/reperfusion by reactions with NO and nitrite-derived species. (Higdon et al..
2012; Khoo et al.. 2010). However, there is no firm evidence that these reactions occur
following exposure to inhaled NO2. Nitrated fatty acids (also known as nitro-fatty acids)
can release NO, which stimulates vascular smooth muscle relaxation through
cGMP-dependent pathways in vitro (Lima etal.. 2005). However, most of the cell
signaling effects of nitrated fatty acids in vivo are likely due to post-translational
modification of proteins (Khoo etal.. 2010). These electrophilic species react with
susceptible thiol groups in transcription factors (Higdon etal.. 2012; Bonacci et al..
2011).
Nitro-fatty acids, such as nitro-oleic acid and nitro-linoleic acid, are anti-inflammatory
(Bonacci et al.. 2011) and vasculoprotective (Khoo etal.. 2010). These effects are
mediated via activation of the peroxisome proliferator-activated receptor gamma
(PPARy) and antioxidant response element (ARE) pathways and suppression of NFxB
and signal transducer and activator of transcription 1 pathways (Bonacci et al.. 2011). In
a mouse model, nitro-oleic acid upregulated vascular eNOS and heme oxygenase-1 and
inhibited angiotensin Il-inducedhypertension (Khoo etal.. 2010; Zhang etal.. 2010a).
Nitro-oleic acid protected against ischemia/reperfusion injury in a mouse model (Rudolph
et al.. 2010). Nitro-oleic acid also activated MMPs (a pro-inflammatory effect) through
thiol alkylation in vitro and inhibited MMP expression in macrophages through activation
of PPARy (Bonacci etal.. 2011). Expression of MMP was also suppressed in a mouse
model of atherosclerosis.
4-60
-------
4.3.4.4 Nitrated Amino Acids and Proteins
Peroxynitrite and NC>2 can react with amino acids to produce nitrated amino acids and
proteins (Hill et al., 2010). These products can also be formed from nitrite and peroxide
in a reaction catalyzed by myeloperoxidase. Nitration of proteins may cause inhibition of
protein function and/or induce antigenicity. Specific antibodies formed against nitrated
proteins may trigger immune reactions (Daiber and Muenzel. 2012). The presence of
nitrated amino acids, such as 3-nitrotyrosine, in cells or tissues is an indicator of NC>2
and/or peroxynitrite formation. A recent animal toxicological study reported formation of
nitrated SP-D resulting from in vivo exposure to 20,000 ppb NC>2 (Matalon et al., 2009).
This modification was accompanied by cross-linking and loss of aggregating activity.
4.3.5 Mode of Action Framework
This section describes the key events, endpoints, and outcomes that comprise the modes
of action proposed for inhaled NO2 and NO. Here, key events are subclinical effects,
endpoints are effects that are generally measured in the clinic, and outcomes are health
effects at the organism level. Biological pathways discussed above that may contribute to
health effects resulting from short- and long-term exposures to NC>2 and NO (Chapter 5
and Chapter 6) are summarized as a part of this analysis. These proposed modes of action
are based on the available evidence and may not reflect all of the pathophysiology
underlying health effects.
Figure 4-1 depicts the proposed mode of action linking respiratory effects to short-term
exposure to NO2.
4-61
-------
Mast cell
degranulation
Bronchoconstriction
Redox reactions in
respiratory tract
ELF and tissue:
formation of
oxidation/nitration
products
Legend
Pollutant
• Key Events
• Endpoints
• Outcomes
Airway
hyperresponsiveness
Asthma
exacerbation
trigger
Impaired epithelial
barrier function
*•
*
Impaired
host defense
Respiratory
infections
^Alveolar macrophage
function
x|,Mucociliary clearance
Note: Pathways indicated by a dotted line are those for which evidence is limited to findings from experimental animal studies, while
evidence from controlled human exposure studies is available for pathways indicated by a solid line. Dashed lines indicate proposed
links to the outcomes of asthma exacerbation and respiratory tract infections. Key events are subclinical effects, endpoints are
effects that are generally measured in the clinic, and outcomes are health effects at the organism level. NO2 = nitrogen dioxide;
ELF = epithelial lining fluid.
Source: National Center for Environmental Assessment.
Figure 4-1 Summary of evidence for the mode of action linking short-term
exposure to nitrogen dioxide and respiratory effects.
Because inhalation of NO2 results in redox reactions in the respiratory tract, the initiating
event in the development of respiratory effects is the formation of oxidation and/or
nitration products in the ELF and possibly in airway or alveolar epithelium. Reactive
intermediates formed are responsible for a variety of downstream key events that may
include respiratory tract inflammation and/or oxidative stress, impaired epithelial barrier
function, altered mucociliary clearance, activation and/or sensitization of neural reflexes,
mast cell degranulation, and increased allergic responses. These key events may
collectively lead to several endpoints including bronchoconstriction, AHR, and impaired
host defenses. Bronchoconstriction is characteristic of an asthma attack, and AHR often
leads to bronchoconstriction in response to a trigger. Thus, the endpoints of
bronchoconstriction and AHR may be linked to the outcome of asthma exacerbation
(Section 5.2.2). while the endpoint of impaired host defenses may be linked to the
outcome of respiratory tract infection (Section 5.2.5).
The strongest evidence for this mode of action comes from controlled human exposure
studies. NCh exposure resulted in enhanced inflammatory mediators (e.g., eicosanoids,
4-62
-------
interleukins) and neutrophils in BL and/or BAL fluid of healthy subjects. Repeated
exposure of healthy subjects led to increased albumin levels in BL fluid, suggesting
impaired epithelial barrier function. In addition, repeated exposure of subjects with
asthma to NO2 enhanced allergic responses such as increased numbers of eosinophils and
levels of abiomarker of eosinophil activation in sputum. Increased airway resistance was
demonstrated following NO2 exposure in healthy human subjects; this response did not
involve vagally mediated neural reflexes. There may be a role for mast cell degranulation
in mediating the increase in airway resistance in humans; however, evidence for this is
indirect as it was based on use of a histamine-suppressive agent. NO2 exposure also
enhanced airway responsiveness to nonspecific challenges, especially in subjects with
asthma. Antioxidant supplementation dampened NO2 exposure-induced lipid
peroxidation and airway responsiveness, providing support for in vitro findings
implicating redox reactions in the ELF. In addition, both early and late asthmatic
responses to an allergen challenge (e.g., AHR, neutrophil, and eosinophil activation) were
enhanced by NO2 exposure. NCh-induced impairment of ciliary function and alveolar
macrophage phagocytic activity suggested impairment of host defenses.
Experimental studies in animals suggest that vagally mediated neural reflexes, mast cell
degranulation, and production of eicosanoids, which may sensitize receptors on nerve
fibers and signal the influx of neutrophils, may contribute to NC>2 exposure-induced
AHR. Exposure to NC>2 also enhanced allergic responses (e.g., IgE, eosinophilic, and
neutrophilic inflammation). Nitration of the collectin protein SP-D and inhibition of its
aggregating activity were also observed following exposure to very high concentrations
of NO2. This may potentially impact microbial clearance and surfactant metabolism. NO2
exposure-induced alteration of mucociliary clearance and alveolar macrophages has also
been demonstrated, suggesting an impairment of host defenses.
Furthermore, there is some evidence for enhanced endogenous formation of peroxynitrite,
which decomposes to NO2, in both human subjects with asthma and animal models of
allergic airway disease. In experimental animals, endogenous peroxynitrite/NCh
formation was associated with AHR and allergic inflammatory responses. Reduction of
peroxynitrite formation lessened airway responsiveness, allergic inflammation, and
airway remodeling. These findings raise the possibility that inhaled NC>2 can add to the
lung burden of endogenous NCh which is found in and contributes to AHR and allergic
airway disease.
Figure 4-2 depicts the proposed mode of action for respiratory effects due to long-term
exposure to NCh.
4-63
-------
Legend
H Pollutant
• Key Events
• Endpoints
• Outcomes
/
, , MS
Recurrent/chronic
respiratory tract
inflammation/
\
Allergic
sensitization
^Allergic
responses
Airway
inflammation
Airway
remodeling
Airway
hyperresponsiveness
New onset asthma/
Asthma exacerbation
trigger
Note: Pathways indicated by a dotted line are those for which evidence is limited to findings from experimental animal studies, while
evidence from controlled human exposure studies is available for pathways indicated by a solid line. The dashed line indicates a
proposed link to the outcome of new onset asthma/asthma exacerbation. Key events are subclinical effects, endpoints are effects
that are generally measured in the clinic, and outcomes are health effects at the organism level. NO2 = nitrogen dioxide.
Source: National Center for Environmental Assessment.
Figure 4-2 Summary of evidence for the mode of action linking long-term
exposure to nitrogen dioxide and respiratory effects.
The initiating events in the development of respiratory effects due to long-term NC>2
exposure are recurrent and/or chronic respiratory tract inflammation and oxidative stress.
These are the driving factors for potential downstream key events, allergic sensitization,
airway inflammation, and airway remodeling, that may lead to the endpoint AHR. The
resulting outcome may be new asthma onset, which presents as an asthma exacerbation
that leads to physician-diagnosed asthma (Section 6.2.2).
The strongest evidence for this mode of action in humans comes from controlled human
exposure studies involving repeated exposures of healthy subjects to NO2 over several
days. Findings included upregulation of Th2 cytokines in respiratory epithelium, which is
characteristic of allergic skewing and part of the allergic sensitization pathway. In
addition, persistent airway neutrophilia and upregulation of 1C AM-1 in airway epithelium
were observed. Reductions in lymphocyte subpopulations, suggesting impaired host
defense, also occurred. Although these were short-term exposure studies, findings
suggest that cumulative effects may occur over time and the possibility that recurrent or
chronic exposure to NCh may lead to the development of asthma.
Studies in experimental animals exposed to NO2 for several weeks found nasal
eosinophilia and enhanced mast cell responses. Other evidence suggests that endogenous
NC>2 acts as an adjuvant promoting the development of allergic airway disease in
4-64
-------
response to an inhaled allergen. This is consistent with mechanistic studies that suggest
that allergic sensitization involves several redox-sensitive steps and that ROS and RNS
promote the development of an allergic phenotype.
Findings that reduction of endogenous peroxynitrite production decreased airway
remodeling in animal models of allergic airway disease suggest that endogenous NO2
may contribute to airway remodeling. Subchronic exposure to NCh enhanced both airway
responsiveness and specific airway resistance, suggesting that airway remodeling may
have contributed to the development of AHR in this nonallergic animal model. Thus,
evidence points to the possibility that inhaled NO2 can add to the lung burden of
endogenous NO that contributes to airway remodeling. Mechanistic studies indicate that
inflammatory mediators and structural changes occurring due to airway remodeling can
alter the contractility of airway smooth muscle. Thus, persistent inflammation, allergic
sensitization, and airway remodeling due to enhanced endogenous NO2 production or to
long-term NC>2 exposure may contribute to the development of AHR. The development
of AHR may be linked to the outcome of new onset asthma.
Figure 4-3 depicts the proposed mode of action for extrapulmonary effects due to
short- or long-term exposure to NC>2.
4-65
-------
Legend
Pollutant
• Key Events
• Endpoints
• Outcomes
Diffusion/
migration of
inflammatory/
vasoactive
Redox reactions in
respiratory tract
ELF and tissue:
formation of
oxidation/nitration
products
mediators
t
Respiratory tract
inflammation/
oxidative stress
Endothelial
dysfunction
Altered heart
rate/heart rate
variability
Cardiovascular
effects
Note: Pathways indicated by a dotted line are those for which evidence is limited to findings from experimental animal studies, while
evidence from controlled human exposure studies is available for pathways indicated by a solid line. Dashed lines indicate potential
links to outcomes related to cardiovascular or other organ effects. Key events are subclinical effects, endpoints are effects that are
generally measured in the clinic, and outcomes are health effects at the organism level. NO2 = nitrogen dioxide; ELF = epithelial
lining fluid.
Source: National Center for Environmental Assessment.
Figure 4-3 Summary of evidence for the mode of action linking exposure to
nitrogen dioxide with extrapulmonary effects.
There is more uncertainty regarding the mode of action for extrapulmonary effects of
inhaled NC>2. However, evidence suggests the following. The initiating events occur in
the respiratory tract, where redox reactions lead to the formation of oxidation and/or
nitration products in the ELF. Reactive intermediates formed are responsible for
downstream key events that may include activation/sensitization of neural reflexes and
respiratory tract inflammation/oxidative stress. This latter key event may lead to diffusion
or migration of inflammatory or vasoactive mediators into the circulation. Circulating
mediators may result in systemic inflammation and/or oxidative stress, which may affect
other organs (Sections 5.3. 6.3 and 6.4). Alternatively, circulating mediators may result in
inflammatory activation of endothelial cells, which may lead to the endpoint endothelial
dysfunction. Activation of neural reflexes may lead to the endpoint of altered heart rate
and/or heart rate variability. Endothelial dysfunction and altered heart rate/heart rate
variability may result in cardiovascular effects (Sections 5.3 and 6.3). It should be noted
that activation of neural pathways was also depicted in Figure 4-1 (short-term exposure
and respiratory effects) because activation of neural pathways may impact airway
function as well as cardiac function.
4-66
-------
The strongest evidence for this mode of action in humans comes from recent controlled
human exposure studies. Both altered heart rate variability and altered blood lipids have
been demonstrated. Whether altered heart rate variability was due to stimulation of
pulmonary irritant receptors is unclear because no studies in humans exposed to NC>2
have observed increases in respiratory rate or decreases in heart rate. In addition, plasma
from human subjects exposed to NCh was found to contain increased levels of sLOX
compared with plasma from control subjects. This plasma also stimulated endothelial cell
activation in an in vitro assay. These results indicate that diffusion or migration of an
inflammatory or vasoactive mediator into the circulation occurred that may transduce a
downstream effect in the vasculature or in other organs. This possibility is consistent with
changes in peripheral blood lymphocyte number and subsets, as well as with altered
blood lipids, which have been observed in humans following exposure to NC>2. These
findings point to a pathway by which inhaled NCh leads to circulating soluble factors that
promote inflammatory signaling in the vasculature and/or other organs.
In experimental animal studies, findings of altered blood glutathione levels and lipids,
decreased pentobarbital-induced sleeping time, and mild pathology in brain and heart
accompanied by tissue markers of oxidative stress and inflammation, are consistent with
the possibility that exposure to NO2 results in circulating soluble factors that promote
inflammatory signaling and/or oxidative stress. There is also some support for activation
of neural reflexes because bradycardia was demonstrated in experimental animals that
were exposed to very high concentrations of NC>2. This response was inhibited by
atropine indicating the involvement of pulmonary irritant receptors and the vagus nerve.
These findings are consistent with NC>2-induced changes in respiratory rate demonstrated
in other studies in experimental animals.
Figure 4-4 depicts the mode of action linking extrapulmonary effects with exposure to
NO.
4-67
-------
major ,.
.***
_ . Diffusion across
NO *••„ ,
cell membranes
*
minor **«.
Legend
H Pollutant
• Kev Events
Binding to
* hemoglobin in
red blood cells
Binding to soluble
guanylate cyclase — *•
in airway smooth muscle
Reactions
hemoglobin
Enzyme
activation and —
mediator
production
Circulating
products
Relaxation of airway
smooth muscle
Note: Pathways indicated by a dotted line are those for which evidence is limited to findings from experimental animal studies. No
links to endpoints or outcomes are proposed. Key events are subclinical effects. NO = nitric oxide.
Source: National Center for Environmental Assessment.
Figure 4-4 Summary of evidence for the mode of action linking exposure to
nitric oxide with extrapulmonary effects.
Because NO has a high affinity for heme proteins and because there is no barrier to its
diffusion across membranes, it rapidly crosses cell membranes and binds to heme
proteins. Inhaled NO diffuses across the alveolar capillary barrier and binds to
hemoglobin in red blood cells. To a lesser extent, inhaled NO also diffuses across airway
epithelium to react with soluble guanylate cyclase in airway smooth muscle. Diffusion
across cell membranes and binding to heme proteins comprise the initiating events in the
mode of action for inhaled NO. The resulting key events include reactions with
hemoglobin to form nitrosylhemoglobin, methemoglobin, nitrate, and possibly
S-nitrosohemoglobin, and activation of soluble guanylate cyclase, which produces a
mediator that relaxes airway smooth muscle. Because health effects of inhaled NO have
not been identified in Chapter 5 and Chapter 6. no endpoints or outcomes have been
included in this analysis.
4.4 Summary
This chapter provides a foundation for understanding how exposure to the gaseous air
pollutants NO2 and NO may lead to health effects. This discussion encompasses the many
steps between uptake into the respiratory tract or the circulation and the biological
responses that ensue. While NO2 reacts with components of the ELF and with the
respiratory epithelium, NO reacts with heme proteins in the circulation. These chemical
interactions are responsible for targeting these oxides of nitrogen species to different
tissues (i.e., NO2 to the respiratory tract and NO to the circulation). Biologic responses to
4-68
-------
inhaled NO2 and NO were organized in a mode of action framework that may be used to
guide interpretation of health effects evidence presented in subsequent chapters.
4-69
-------
CHAPTER 5 INTEGRATED HEALTH EFFECTS
OF SHORT-TERM EXPOSURE TO
OXIDES OF NITROGEN
5.1 Introduction
5.1.1 Scope of Chapter
The preceding chapters describe the widespread potential for human exposure to ambient
oxides of nitrogen (Chapters 2 and 3) and the capability for ambient-relevant
concentrations of inhaled NC>2 to initiate a cascade of molecular, cellular, and organ
responses, particularly in the airways (Chapter 4). These lines of evidence point to the
potential for ambient exposure to oxides of nitrogen to induce health effects. However,
the preceding chapters also identify the importance of assessing exposure measurement
error due to heterogeneity in ambient concentrations of oxides of nitrogen, effects of
other correlated pollutants, and the extent to which information on a proposed mode of
action is available to support biological plausibility. With consideration of these issues,
this chapter summarizes, integrates, and evaluates the evidence for relationships between
various health effects and short-term (i.e., minutes up to 1 month, Section 1.5) exposure
to oxides of nitrogen. The chapter sections comprise evaluations of the epidemiologic,
controlled human exposure, and animal toxicological evidence for the effects of
short-term exposure to oxides of nitrogen on health outcomes related to respiratory
effects (Section 5.2). cardiovascular effects (Section 5.3). and total mortality
(Section 5.4). Reproductive and developmental effects also have been examined in
relation to short-term exposure to oxides of nitrogen. This evidence is evaluated with
studies of long-term exposure in Chapter 6 because associations are often compared
among various short- and long-term exposure periods that are difficult to distinguish.
Individual sections for broad health categories (i.e., respiratory, cardiovascular, mortality)
begin with a summary of conclusions from the 2008 ISA for Oxides of Nitrogen followed
by an evaluation of recent (i.e., published since the completion of the 2008 ISA for
Oxides of Nitrogen) studies that builds upon evidence from previous reviews. Within
each of these sections, results are organized into smaller outcome groups [e.g., asthma
exacerbation, myocardial infarction (MI)] that comprise a continuum of clinical to
subclinical outcomes and events. The discussion of individual outcomes and events is
then organized by specific scientific discipline (i.e., epidemiology, controlled human
5-1
-------
exposure, toxicology). This organization permits clear description of the extent of
coherence and biological plausibility for the effects of oxides of nitrogen on a group of
related outcomes, and in turn, transparent characterization of the weight of evidence in
drawing conclusions.
Sections for each of the broad health categories conclude with an integrated assessment
of the evidence and a conclusion regarding causality. A determination of causality was
made for each broad health category by evaluating the evidence for each category
independently with the causal framework (described in the Preamble to the ISA).
Findings for mortality informed multiple causal determinations. Findings for
cause-specific mortality (i.e., respiratory, cardiovascular) were used to assess the
continuum of effects and inform the causal determinations for respiratory and
cardiovascular effects. A separate causal determination was made for total mortality
based on the evidence for nonaccidental causes of mortality combined and also based on
the extent of biological plausibility provided by evidence for the spectrum of
cardiovascular and respiratory effects that are underlying causes of mortality. Judgments
of causality were made by evaluating the evidence over the full range of concentrations in
animal toxicological, controlled human exposure, and epidemiologic studies defined in
this ISA to be relevant to ambient exposure (i.e., up to 5,000 ppb NCh or NO;
Section 1.2). Experimental studies that examined higher NO2 or NO concentrations were
evaluated particularly to inform judgments about plausible modes of action.
5.1.2 Evidence Evaluation and Integration to Form Causal Determinations
5.1.2.1 Evaluation of Individual Studies
As described in the Preamble to the ISA (Section 5 .a), causal determinations were
informed by the evidence integrated across scientific disciplines (e.g., exposure, animal
toxicology, epidemiology) and related outcomes and judgments of the strength of
inference from individual studies. These judgments were formed by evaluating strengths
as well as various sources of bias and uncertaintyjelated to aspects such as study design,
study population characterization, exposure assessment, outcome assessment,
consideration of confounding, and statistical methodology. This evaluation was applied to
controlled human exposure, animal toxicological, and epidemiologic studies included in
this ISA from previous assessments and those published since the 2008 ISA for Oxides of
Nitrogen. The aspects are described in the Appendix to the ISA and are consistent with
current best practices employed in other approaches for reporting or evaluating health
5-2
-------
science data.1 Additionally, the aspects are compatible with published the U.S. EPA
guidelines related to cancer, neurotoxicity, reproductive toxicity, and developmental
toxicity (U.S. EPA. 2005. 1998b. 1996b. 1991).
These aspects were not used as a checklist or criteria to define the quality of a study, and
judgments were made without considering the results of a study. The presence or absence
of particular features did not necessarily define a less informative study or exclude a
study from consideration in the ISA. These aspects were not criteria for a particular
determination of causality in the five-level hierarchy. As described in the Preamble.
causal determinations were based on judgments of the overall strengths and limitations of
the collective body of available studies and the coherence of evidence across scientific
disciplines and related outcomes. Further, the study aspects described in the Appendix are
not intended to be a complete list that affect the strength of inference from a study, but
they comprise the major aspects considered in this ISA to evaluate studies. Where
possible, study considerations, for example, exposure assessment and confounding
(i.e., bias due to a relationship with the outcome and correlation with exposures to oxides
of nitrogen), are framed to be specific to oxides of nitrogen. Thus, judgments of the
strength of inference from a study can vary depending on the specific pollutant being
assessed.
Confounding in epidemiologic studies was a key source of bias evaluated in this ISA.
Epidemiologic studies of short-term exposure to oxides of nitrogen relied primarily on
temporal variation in exposure (e.g., day-to-day changes in ambient NCh concentrations)
and health effects. Other risk factors for health effects also exhibit similar temporal trends
as oxides of nitrogen and include meteorological variables, season, long-term time trends,
medication use, and copollutant exposures. These factors and others specified in the
Appendix are important to evaluate as potential confounders of associations for oxides of
nitrogen, particularly given the small effect sizes typically observed. Epidemiologic
studies reviewed in this ISA varied in the extent to which they considered potential
confounding. Because no single study considered all potential confounders, and not all
factors were examined in the collective body of studies, residual confounding by
unmeasured factors is possible. Residual confounding also is possible by poorly
measured factors. In this ISA, potential confounding was assessed as the extent to which
the collection of studies examined factors that are well documented in the literature to be
associated with exposure to oxides of nitrogen and health outcomes.
In epidemiologic studies evaluated in this ISA, confounding was assessed primarily using
multivariable models that include NO2 concentrations and the putative confounder in the
1 See, for example, NTP OHAT approach (Rooney etal. 2014). IRIS Preamble (U.S. EPA. 2013p). ToxRTool
(Klimisch et al. 1997). STROBE guidelines (von Elm etal.. 2007). and ARRIVE guidelines (Kilkenny etal.. 2010).
5-3
-------
same model. The NO2 effect estimate represents the effect of NO2 keeping the level of
the covariate constant. In the ISA, confounding is assessed by examining the change in
the magnitude of the effect estimate and width of the 95% CI for NO2 in multivariable
models, not just a change in statistical significance. The limitations of multivariable
models are well recognized. If NO2 and the potential confounder are highly correlated,
the collinearity (i.e., covariates predict each other) introduced by including them in the
same model can misleadingly decrease or increase the magnitude or precision of the
effect estimates for NO2 or the potential confounder. Collinearity can occur, for example,
if pollutants are from the same sources or are derived from NO2 [e.g., ozone (Os)], or if
meteorology affects formation of both pollutants. Adding correlated but noncausal
variables can produce models that fit the data poorly, and residual confounding is
possible if confounders are excluded or poorly measured.
Studies reviewed in this ISA predominantly evaluated copollutant confounding by
copollutant models (NCh plus one copollutant). Inference about the independent effects
of NO2 from copollutant models can be limited because differences in the spatial
distributions of NC>2 and the copollutant may not satisfy the assumptions of equal
measurement error or constant correlations for NC>2 and the copollutant (Gryparis et al..
2007). Further, copollutant models for NC>2 assumed linear relationships with the
copollutant, and nonlinear relationships are possible because of varying near-road
gradients (Figure 3-2). Other methods for evaluating copollutant confounding do not
require the aforementioned assumptions, including a hierarchical Bayesian approach that
estimates single-pollutant effects in a particular location then combines these
single-pollutant effects across locations in a model as the predictor and outcome,
respectively (Gryparis et al.. 2007; Schwartz and Coull. 2003). Such Bayesian models are
unavailable for NC>2. Models examining joint effect or interaction terms for NC>2 and a
copollutant potentially can provide information on confounding and synergistic effects.
These are available only to a limited extent, particularly for traffic-related copollutants.
Because examination of copollutant confounding is based largely on copollutant models,
their limitations are considered in drawing inferences about independent associations for
NO2. Emphasis is placed on results based on exposure assessment methods that likely
produce comparable measurement error for NO2 and copollutants, such as ambient or
total personal and microenvironmental exposure assessment.
5.1.2.2 Integration of Scientific Evidence
In addition to strength of inference from individual studies, causal determinations were
based on integrating multiple lines of evidence. As detailed in the Preamble, evidence
integration involved evaluating the consistency and coherence of findings within and
5-4
-------
across disciplines as well as within and across related outcomes. Based on this
evaluation, judgments were formed on the extent to which chance, confounding, and
other biases could be ruled out with reasonable confidence in the evidence base as a
whole. Examples of evidence integration are summarized below.
To increase comparability of results across epidemiologic studies, the ISA presents effect
estimates for associations with health outcomes scaled to the same increment of oxide of
nitrogen concentration.1 The increments for standardization vary by averaging time
(e.g., 24-h avg, 1-h max) and oxide of nitrogen. For 24-h average, effect estimates were
scaled to a 20-ppb increase for NO2 or NO and a 40-ppb increase for NOx. For a
1-h maximum, effect estimates were scaled to a 30-ppb increase for NO2, an 80-ppb
increase for NO, and a 100-ppb increase for NOx. For an 8-h maximum, the increments
for standardization are 25 ppb for NO2, 45 ppb for NO, and 65 ppb for NOx. These
increments were derived by calculating the U.S.-wide percentile distributions for a given
averaging time and then calculating the approximate difference between the median (a
typical pollution day) and the 95th percentile (a more polluted day) for a given averaging
time [for 1-h maximum, see Table 2-3 and for 24-h average and 8-h maximum, see
Supplemental Table S5-1; (U.S. EPA.2015g)1.
There were common exceptions to this standardization method. Averaging times other
than 24-h average or 1-h maximum were examined, for example, 2- to 15-h averages.
Effect estimates based on these averaging times were not standardized but are presented
in the ISA as reported in their respective studies. Some studies reported effect estimates
in terms of (ig/m3 increases in oxides of nitrogen, which could be converted to ppb and
standardized for NO2 and NO but not NOx. This conversion could not be made for NOx
because the proportions of NO2 and NO are unknown for the various NOx metrics. Also,
data are not available to calculate the percentiles of NOx concentrations in (ig/m3 at a
national scale for the U.S. or other countries. Therefore, the ISA presents effect estimates
based on (ig/m3 of NOx as they are reported in their respective studies.
Integrating evidence across scientific disciplines can help address uncertainties within a
particular discipline. For example, controlled human exposure and animal toxicological
studies can provide direct evidence for health effects related to NO2 or NO exposures.
Experimental evidence for effects from a controlled exposure and coherence with
epidemiologic findings may help understand whether epidemiologic associations with
health outcomes plausibly reflect an independent effect of ambient NO2 exposure or
could be confounded by other factors. Experimental studies additionally can provide
biological plausibility for observed effects by identifying key events in the modes of
1 This is in contrast with reported effect estimates that are scaled to various changes in concentration, such as
interquartile range for the study period or an arbitrary unit such as 10 ppb.
5-5
-------
action. Thus, the integration of evidence across a spectrum of related outcomes and
across disciplines was used to inform uncertainties for any particular outcome or
discipline due to factors such as chance, publication bias, selection bias, and confounding
by copollutant exposures or other factors. The evaluation of health effects also drew upon
information on potential error associated with various exposure assessment methods and
the uptake and distribution of oxides of nitrogen in the body. The subsequent sections
assess strength of inference from studies and integrate multiple lines of evidence to
characterize relationships between oxides of nitrogen and various health effects.
5.2 Respiratory Effects
5.2.1 Introduction
The 2008 ISA for Oxides of Nitrogen concluded that evidence was sufficient to infer a
likely to be causal relationship between short-term exposure to NO2 and respiratory
effects (U.S. EPA. 2008c). emphasizing a large body of epidemiologic evidence. In
studies that were not available until after the completion of the 1993 AQCD for Oxides of
Nitrogen (U.S. EPA. 1993a). short-term increases in ambient NO2 concentrations were
consistently associated with increases in respiratory-related hospital admissions and ED
visits. The coherence of these findings with NO2-related increases in respiratory
symptoms in children with asthma supported an effect of NO2 exposure on asthma
exacerbation. NO2 was not consistently related to lung function decrements across
epidemiologic and controlled human exposure studies or across populations with varying
respiratory conditions such as asthma or COPD. However, epidemiologic studies of
children and adults with asthma observed associations with lung function measured by
supervised spirometry (U.S. EPA. 2008c).
The 2008 ISA identified multiple lines of evidence as supporting an independent
relationship between short-term NO2 exposure and respiratory effects. Controlled human
exposure studies demonstrated NO2-induced increases in airway responsiveness in adults
with asthma. These findings for increased airway responsiveness, a characteristic feature
of asthma, provided biological plausibility for epidemiologic evidence for asthma
exacerbation. Further, airway responsiveness was increased following <1 to 6-hour
exposures to NO2 at concentrations in the range of 100 to 300 ppb, which are not much
higher than peak ambient concentrations (Section 2.5).
5-6
-------
Previous epidemiologic studies also indicated independent associations for NC>2. Personal
and indoor NC>2 were associated with respiratory effects, and associations with both
personal and ambient NO2 were observed in copollutant models that adjusted for fine
particulate matter (PIVbs) or a traffic-related copollutants such as carbon monoxide (CO).
In the few available results, NCh-related respiratory effects were observed with
adjustment for elemental carbon (EC), organic carbon (OC), or ultrafine particles (UFP);
other traffic-related copollutants were not examined for potential confounding.
Controlled human exposure and animal toxicological studies also demonstrated
NC>2-induced impairments in host defense, changes in the oxidant/antioxidant balance,
and increases in pulmonary inflammation at concentrations of 1,500 to 5,000 ppb NCh,
higher than those demonstrated to increase airway responsiveness (U.S. EPA. 2008c).
The 2008 ISA did not explicitly link these NC>2-induced biochemical and immunological
changes to lines of evidence for asthma exacerbation. Although there was coherence of
evidence across related outcomes and disciplines supporting a relationship between
short-term ambient NC>2 exposure and respiratory effects, due to the high correlations of
NC>2 with other traffic-related pollutants and limited analysis of potential confounding,
sufficient uncertainty was noted in the epidemiologic evidence about the role of NO2 as
an indicator for another traffic-related pollutant or a mixture of such pollutants.
As will be described in the following sections, consistent with the body of evidence
presented in the 2008 ISA for Oxides of Nitrogen, recent studies continue to demonstrate
respiratory effects related to short-term NO2 exposure. The majority of the recent
evidence is from epidemiologic studies, which expand on findings for associations
between ambient NO2 and a broad array of respiratory effects from subclinical increases
in pulmonary inflammation to respiratory mortality, but particularly for effects related to
asthma exacerbation. Because there are few recent controlled human exposure and animal
toxicological studies, previous findings are a large basis of the characterization and
integration of evidence. Where available, results from recent studies are evaluated in the
context of results from previous studies. To clearly characterize differences in the weight
of evidence and the extent of coherence among disciplines and related outcomes, the
discussion of scientific information is organized by respiratory outcome group
(e.g., asthma exacerbation, allergy exacerbation, respiratory infection).
5.2.2 Asthma Exacerbation
As detailed in the preceding section, previous studies provided several lines of evidence
in support of a relationship between short-term NO2 exposure and asthma exacerbation,
represented as respiratory effects in populations with asthma. This evidence is
corroborated by recent studies. In characterizing the current state of the evidence, this
5-7
-------
section begins with effects on increasing airway responsiveness and decreasing lung
function. These are indications of bronchoconstriction and airway obstruction, which can
lead to poorer control of asthma symptoms and potentially hospital admissions or ED
visits for asthma. The evaluation of clinical indicators of asthma exacerbation follows
with discussion of pulmonary inflammation and oxidative stress, which are part of the
proposed mode of action for asthma exacerbation and mediate decreases in lung function
and increases in airway responsiveness (Figure 4-1).
5.2.2.1 Airway Responsiveness in Individuals with Asthma
Overview
Controlled human exposure studies evaluating the effect of inhaled NO2 on the inherent
responsiveness of the airways to challenge by bronchoconstricting agents have had mixed
results. However, meta-analyses show statistically significant effects of NO2 on the
airway responsiveness of individuals with asthma. This section describes analyses
showing that a statistically significant fraction (i.e., 70% of individuals with asthma
exposed to NO2 at rest) experience increases in airway responsiveness following
30-minute exposures to NC>2 in the range of 200 to 300 ppb and following 60-minute
exposures to 100 ppb. The distribution of changes in airway responsiveness is
log-normally distributed. About a quarter of the individuals exposed at rest experience a
clinically relevant reduction in their provocative dose due to NCh relative to air exposure.
A variety of factors that may affect the assessment of airway responsiveness and how
those factors may directionally bias the results of individual studies are briefly
considered.
Background
Bronchial challenge agents can be classified as nonspecific [e.g., histamine, sulfur
dioxide (802), cold air] or specific (i.e., an allergen). Nonspecific agents can be
differentiated between "direct" stimuli (e.g., histamine, carbachol, and methacholine)
which act on airway smooth muscle receptors and "indirect" stimuli (e.g., exercise, cold
air) which act on smooth muscle through intermediate pathways, especially via
inflammatory mediators (Cockcroft and Davis. 2006c). Specific allergen challenges
(e.g., house dust mite, cat allergen) also act "indirectly" via inflammatory mediators to
initiate smooth muscle contraction and bronchoconstriction. This section focuses on
changes in airway responsiveness to bronchial challenge attributable to NO2 in
individuals with asthma. Discussed in Section 4.3.2.5. toxicological studies have
5-8
-------
demonstrated increased airway responsiveness to nonspecific challenges following
short-term exposure. Described in Sections 5.2.2.5 and 4.3.2.6. altered responses to
specific allergens following NO2 exposure have also been demonstrated in human and
animal studies.
Responses to bronchial challenge are typically quantified in terms of the provocative dose
(PD) or provocative concentration (PC) of an agent required to produce a 20% reduction
in FEVi (PD2o or PC20, respectively) or a 100% increase in specific airway resistance
(sRaw) (PDioo or PCioo, respectively). In the general population, airway responsiveness is
log-normally distributed with individuals having airway hyperresponsiveness (AHR)
tending to be those with asthma (Postma and Boezen. 2004; Cockcroft et al.. 1983).
Along with symptoms, variable airway obstruction, and airway inflammation, AHR is a
primary feature in the clinical definition and characterization of asthma severity (Reddel
et al.. 2009). However, not all individuals with asthma experience airway
hyperresponsiveness. The range in airway responsiveness among individuals with asthma
extends into the range of healthy individuals without asthma (Cockcroft. 2010). In
asthma, there is a strong relationship between the degree of nonspecific airway
responsiveness and the intensity of the early airway response to specific allergens to
which individuals have become sensitized (Cockcroft and Davis. 2006a).
In studies investigating the effect of NO2 exposure on airway responsiveness, individuals
with asthma generally have a lower PD of a bronchial challenge agent than healthy
individuals to produce a given reduction in lung function. In the study by Morrow and
Utell (1989a). the average PD of carbachol producing a given change in lung function in
individuals with mild-to-moderate asthma was 16 times lower than in age-matched
healthy controls. Similarly, Hazucha et al. (1983) reported a 10-12 times lower average
baseline PDioo to methacholine in individuals with mild asthma than healthy age-matched
controls. The PDs for asthma in Morrow and Utell (1989a) did not overlap with those of
the healthy controls, whereas Hazucha et al. (1983) observed an overlap with 2 of
15 subjects with asthma being relatively unresponsive to bronchial challenge. Thus,
individuals with asthma are generally at risk at baseline relative to healthy individuals
without NC>2 or other agents, further increasing their airway responsiveness. The
bronchoconstrictive response to indirect acting agents (especially specific allergens) can
be more difficult to predict and control than the bronchoconstrictive response to
nonspecific agents that act directly on airway smooth muscle receptors (O'Byrne et al..
2009). Consequently, most of the available literature relevant to the evaluation of the
effects of NO2 on airway responsiveness has focused primarily on the responses of
individuals with asthma to bronchial challenge with "nonspecific" bronchoconstricting
agents (e.g., methacholine, SCh, cold air).
5-9
-------
In healthy adults without asthma or AHR, there is likely little or no clinical significance
of transient, small increases in airway responsiveness following low-level NC>2 inhalation
exposures. In individuals with asthma, however, transient changes in airway
responsiveness in response to NC>2 may have clinical consequences due to their tendency
to already have greater baseline airway responsiveness than healthy individuals.
Increased airway responsiveness is linked with airway inflammation and airway
remodeling (Chetta et al.. 1996). increased risk for exacerbation (Van Schayck et al..
1991). reduced lung function (Xuan et al.. 2000). and increased symptoms (Murray et al..
1981). A variety of environmental challenges can transiently increase AHR and worsen
asthma control, including allergen exposures (Strand et al.. 1997; Brusasco et al.. 1990).
viral infections (Cheung etal.. 1995; Fraenkel et al.. 1995). cigarette smoke (Tashkin et
al.. 1993). O3 (Kehrletal.. 1999). and other respiratory irritants (Kinsella et al.. 1991).
Transient increases in airway responsiveness following NC>2 or other pollutant exposures
have the potential to increase symptoms and worsen asthma control, even if the pollutant
exposure does not cause acute decrements in lung function.
Four meta-analyses in the peer-reviewed literature have assessed the effects of NCh
exposure on airway responsiveness in individuals with asthma (Brown. 2015; Goodman
et al.. 2009; Kjaergaard and Rasmussen. 1996; Folinsbee. 1992). Kjaergaard and
Rasmussen (1996) reported statistically significant effects of NO2 exposure on the airway
responsiveness of subjects with asthma exposed to less than or equal to 300 ppb NCh but
not for exposures in excess of 300 ppb NCh. With consideration given to activity level
during exposure, Folinsbee (1992) found statistically significant increases in airway
responsiveness of subjects with asthma exposed to NO2 at rest across all concentration
ranges (namely, <200 ppb, 200 to 300 ppb, and >300 ppb). However, there was no
statistically significant effect of NO2 exposures on responsiveness during exercise. For
instance, following exposures between 200 and 300 ppb NO2, 76% of subjects exposed at
rest had increased responsiveness, which was statistically significant, whereas only 52%
of subjects exposed while exercising had increased responsiveness, which was not a
statistically significant change. The analyses of Folinsbee (1992) and Kjaergaard and
Rasmussen (1996) in effect assessed nonspecific responsiveness because few studies of
allergen responsiveness were available.
The analyses conducted by Folinsbee (1992) were detailed in Chapter 15 of the 1993
AQCD for Oxides of Nitrogen (U.S. EPA. 1993a). Results of these analyses appeared in
Table 15-10 of the 1993 AQCD and supported the conclusion that NO2 exposure
increases airway responsiveness in individuals with asthma. The results of a slightly
modified analysis focusing exclusively on nonspecific responsiveness appeared in
Tables 3.1-3 of the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008b. c). The overall
conclusion of that modified analysis was that NO2 exposures as low as 100 ppb (the
5-10
-------
lowest concentration experimentally evaluated) conducted during rest, but not exercise,
increased nonspecific responsiveness of individuals with asthma. Due to differences in
study protocols (e.g., rest versus exercise) in the NCh-airway responsiveness literature,
the original (Folinsbee. 1992) and updated meta-analyses in the 2008 ISA for Oxides of
Nitrogen (U.S. EPA, 2008b) assessed only the fraction of individuals experiencing
increased or decreased airway responsiveness following NO2 exposure.
Goodman et al. (2009) provided meta-analyses and meta-regressions evaluating the
effects of NO2 exposure on airway responsiveness in subjects with asthma. By
considering studies of specific allergen and nonspecific responsiveness following NO2
exposure, Goodman et al. (2009) evaluated a larger number of studies than the analysis in
the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008^). which was limited to
nonspecific responsiveness in subjects with asthma in an attempt to reduce the
heterogeneity among studies. Goodman et al. (2009) evaluated changes in three endpoints
following NO2 exposure relative to a control air exposure: (1) the fraction of subjects
with asthma experiencing increases in responsiveness, (2) the PD of the bronchial
challenge agent, and (3) the FEVi response to the challenge agent. Overall, statistically
significant effects of NO2 exposure on each of these three endpoints were observed.
Consistent with the meta-analysis provided in the 2008 ISA for Oxides of Nitrogen (U.S.
EPA. 2008c). Goodman et al. (2009) found 64% (95% CI: 58, 71) of subjects with
asthma exposed at rest to NO2 experienced an increase in airway responsiveness, whereas
there was no effect of NO2 exposure during exercise with 52% (95% CI: 43, 60) having
an increase in responsiveness. Additionally, NO2 exposure resulted in statistically
significant reductions in PD as well as increases in the FEVi decrement following
bronchial challenge.
Goodman et al. (2009) concluded that, "NO2 is not associated with clinically relevant
effects on AHR at exposures up to 600 ppb based primarily on the small magnitude of
effects and the overall lack of exposure-response associations." Relative to therapeutic
agents used to treat airway responsiveness, which may be considered effective if they
more than double the PD for methacholine, the authors concluded that a -50% change in
the PD due to NO2 exposure would be considered adverse. Using the summary statistics
provided in individual studies, the effect of NO2 exposure was a -27% (95% CI: -37,
-18) change in the PD. Stratifying by rest and exercise exposure, the NO2-induced
changes in PD were -30% (95% CI: -38, -22) and -24% (95% CI: -40, -7),
respectively. Thus, the authors concluded that the effects of NO2 exposure on airway
responsiveness were sufficiently small so as not to be considered adverse. Based on the
lack of a monotonic increase in responsiveness with exposure, the authors also suggested
that NO2 is not a causal factor. However, as airway responsiveness data is log-normally
distributed (Postma and Boezen. 2004; Cockcroft et al.. 1983), use of arithmetic mean PD
5-11
-------
data may affect the validity of some analyses in the Goodman et al. (2009) study. For
example, in the study by Bylin etal. (1988) following exposure to 140 ppb NO2, there
was an arithmetic mean increase of 17% in the PD relative to filtered air, which was
driven by a few individuals; whereas, the median and geometric mean show a 24% and
16% decrease, respectively, in the PD following NO2 relative to filtered air exposure.
Based on this example, incorrectly assuming data to be normally distributed can lead to
the conclusion that airway responsiveness is decreased following NCh exposure, whereas
it was actually increased in most individuals.
None of the above described meta-analyses provided a comprehensive assessment of the
clinical relevance of changes in airway responsiveness, the potential for methodological
biases in the original papers, or the distribution of responses. The remainder of this
section provides such analyses of airway responsiveness data and a discussion of factors
that may have affected the experimental determination of airway responsiveness as
presented by Brown (2015). In review of this ISA, Roux and Frey (2015) expressed
support of this meta-analysis and suggested that it facilitated the inferences that could be
drawn from the original controlled human exposure studies (see page 2 of cover letter and
page 7 of consensus comments). Detailed descriptions of individual studies are provided
in the 1993 AQCD for Oxides of Nitrogen (U.S. EPA. 1993a) and 2008 ISA for Oxides
of Nitrogen (U.S. EPA. 2008c). As done in the 2008 ISA for Oxides of Nitrogen (U.S.
EPA. 2008c). the fraction of individuals having an increase in airway responsiveness
following NO2 exposure was assessed. Due to considerable variability in exposure
protocols and the potential for this variability in protocols to affect estimates of PD (see
Factors Affecting Airway Responsiveness and Dose-response), the magnitude of
NO2-induced changes in PD was not evaluated in the original work by Folinsbee (1992)
or in related documents of the U.S. EPA (U.S. EPA. 2008b. c, 1993a). Herein, the
magnitude of the PD change for nonspecific agents is evaluated in studies that presented
individual subject data for persons with asthma exposed to NO2 at rest. The focus on
resting exposures and nonspecific challenges when assessing the magnitude of change in
PD (dPD) was due to the statistically significant effects of NO2 exposure on airway
responsiveness for these conditions as reported in the 2008 ISA for Oxides of Nitrogen
[see Section 3.1.3.2 of U.S. EPA (2008c)1. In assessing the magnitude of PD change,
additional consideration was given to individuals experiencing a doubling-dose change in
PD following exposure to NO2 relative to filtered air. In a joint statement of the American
Thoracic Society (ATS) and European Respiratory Society, one doubling dose change in
PD is recognized as a potential indicator, although not a validated estimate, of clinically
relevant changes in airway responsiveness (Reddel et al.. 2009). Additional analyses also
evaluate the distribution of PD responses to NO2 and the concentration/dose-response
relationship.
5-12
-------
New Analyses
As an update to Table 3.1-2 in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c).
Tables 5-1 and 5-2 present studies selected for inclusion into the meta-analyses by Brown
(2015). Relative to Table 3.1-2 in the 2008 ISA, Tables 5-1 and 5^2 add data that either
are new or were previously excluded for 155 subject exposures from nine studies (Riedl
etal.. 2012: Witten etal.. 2005: Barck et al.. 2002: Jenkins et al.. 1999: Strand et al..
1998: Strand etal.. 1997: Tunnicliffe et al.. 1994: Morrow and Utell. 1989a: Oreheket
al.. 1976). The meta-analysis in the 2008 ISA focused on nonspecific airway
responsiveness, although studies examining allergen challenges were discussed in
Section 3.1.3.1 and Annex Table AX5.3-2 of the 2008 ISA. With respect to the studies in
Tables 5-1 and 5-2. subjects recruited for these studies ranged in age from 18 to 50 years
with the exception of Avol et al. (1989) who studied children ages 8-16 years. The
disease status of subjects was mild asthma in most studies, but ranged from inactive
asthma up to severe asthma in a few studies.
For studies that assessed airway responsiveness at multiple time points post-exposure or
over repeated days of exposure, the data from the first time point and first day of
exposure were selected for inclusion in Tables 5-1 and 5-2 to reduce the heterogeneity
among studies. Selection of the earliest time point assessing airway responsiveness was,
in part, due to late phase responses (3-8 hours post-allergen challenge) being
mechanistically different from early phase responses (<30 minutes post-allergen
challenge) (O'Bvrne et al.. 2009: Cockcroft and Davis. 2006c). Tables 5-1 and M are
sorted by NO2 exposure concentration, so studies that evaluated multiple NO2 exposure
concentrations appear in multiple rows.
5-13
-------
Table 5-1 Resting exposures to nitrogen dioxide and airway responsiveness in
individuals with asthma.
Reference N
Ahmed et al. 20
(1983a)
Ahmed et al. 20
(1983b)
Hazucha et al. 15
(1983)
Oreheketal. 20
(1976)
Tunnicliffe et al. 8
(1994)
Bylinetal. (1988) 20
Oreheketal. 4
(1976)
Jorres and 14
Maqnussen (1990)
Barck et al. (2002) 13
Strand et al. 18
(1997)
Strand et al. 16
(1998)
Bvlin etal. (1988) 20
Tunnicliffe et al. 8
(1994)
Time
Chall- Post-
NO2 Exp. enge End exp
ppb (min) Type Point min
100 60 CARB sGaw NA
100 60 RAG sGaw IM
100 60 METH sRaw 20
100 60 CARB sRaw IM
100 60 HDM FEVi IM
140 30 HIST sRaw 25
200 60 CARB sRaw IM
250 30 SO2 sRaw 27
260 30 SIR, FEVi 240
TIM
260 30 SIR, sRaw 240
TIM
260 30 SIR FEVi 240
270 30 HIST sRaw 25
400 60 HDM FEVi IM
Change in
ARa Average
+ - Air
13 7 6.0 ±2.4
10 8 9.0 ±5.7
6 7 1.9±0.4
14 3 0.56 ±0.08
3 5 -14.62
AFEVi
14 6 0.39 ±0.07
3 0 0.60 ±0.10
11 2 46.5 ±5.1
5 7 -5 ± 2
AFEVi
9 9 860 ± 450
11 4 -0.1 ±0.8
AFEVi
14 6 0.39 ±0.07
8 0 -14.62
AFEVi
PD±SEb
NO2 p-valuec
2.7 ±0.8 NA
11. 7 ±7.6 n.s.
2.0±1.0 n.s.
0.36 ±0.05 <0.01d
-14.41 n.s.
AFEVi
0.28 ±0.05 n.s.
0.32 ±0.02 n.s.
37.7 ±3.5 <0.01
-4 ± 2 n.s.
AFEVi
970 ±450 n.s.
-2.5 ±1.0 0.03
AFEVi
0.24 ±0.04 <0.01
-18.64 0.009
AFEVi
5-14
-------
Table 5-1 (Continued): Resting exposures to nitrogen dioxide and airway
responsiveness in individuals with asthma.
Reference
Bvlin etal. (1985)
Mohsenin (1987a)
Bvlin etal. (1988)
N
8
10
20
NO2
ppb
480
500
530
Exp.
(min)
20
60
30
Chall-
enge
Type
HIST
METH
HIST
End
Point
sRaw
pEF
sRaw
Time
Post-
ex p
min
20
IM
25
Change in
ARa Average
+ - Air
5 0 >30
7 2 9.2 ±4.7
12 7 0.39 ±0.07
PD±SEb
NO2
>20
4.6 ±2.6
0.34 ±0.08
p-valuec
0.04
0.042
n.s.
BIR = birch; CARB = carbachol; Exp. = exposure, FENA, = forced expiratory volume in 1 second; HDM = house dust mite allergen;
HIST = histamine; IM = immediately after exposure; METH = methacholine; NA = not available; NO2 = nitrogen dioxide; n.s. = less
than marginal statistical significance, p > 0.10; pEF = partial expiratory flow at 40% vital capacity; RAG = ragweed; SO2 = sulfur
dioxide; sGaw = specific airway conductance; sRaw = specific airway resistance; TIM = timothy pollen.
aChange in airway responsiveness (AR): number of individuals showing increased (+) or decreased (-) airway responsiveness
after NO2 exposure compared to air.
bPD ± SE: arithmetic or geometric mean provocative dose (PD) ± standard error (SE). See individual papers for PD calculation and
dosage units. AFEVi indicates the change in FEVi response at a constant challenge dose.
Statistical significance of increase in AR to bronchial challenge following NO2 exposure compared to filtered air as reported in the
original study unless otherwise specified. Statistical tests varied between studies (e.g., sign test, West, analysis of variance).
dStatistical significance for all individuals with asthma from analysis by Dawson and Schenker (1979). Orehek et al. (1976) only
tested for differences in subsets of individuals classified as "responders" and "nonresponders."
Table 5-2 Exercising exposures to nitrogen dioxide and airway
in individuals with asthma.
NO2 Exp.
Reference n ppb Min
Roqer etal. (1990) 19 150 80
Kleinmanetal. 31 200 120
(1983)
Jenkins etal. (1999) 11 200 360
Jorres and 11 250 30
Maqnussen (1991)
Strand etal. (1996) 19 260 30
Avol etal. (1988) 37 300 120
Avol etal. (1989) 34 300 180
Tj Change
Post- in ARa Average
Challenge End exp
Type Point min + - Air
METH sRaw 120 10d 7d 3.3 ± 0.7
METH FEVi IM 20 7 8.6 ± 2.9
HDM FEVi IM 6 5 2.94
METH sRaw 60 6 5 0.41 ± 1.6
HIST sRaw 30 13 5 296 ± 76
COLD FEVi 60 11d 16d -8.4 ±1.8
AFEVi
COLD FEVi 60 12d 21 d -5 ± 2
AFEVi
responsiveness
PD±SEb
NO2
3.1 ±0.7
3.0± 1.1
2.77
0.41 ± 1.6
229 ± 56
-10.7 ±2.0
AFEVi
-4 ±2
AFEVi
P-
value0
n.s.
<0.05
n.s.
n.s.
0.08
n.s.
n.s.
Bauer etal. (1986) 15 300 30
COLD FEVi 60
0.83 ±0.12 0.54 ±0.10 <0.05
5-15
-------
Table 5-2 (Continued): Exercising exposures to nitrogen dioxide and airway
responsiveness in individuals with asthma.
Reference
Morrow and
(1989a)
Roger et al.
Utell
(1990)
Rubinstein et al.
(1990)
tRiedletal.
tRiedletal.
(2012)
(2012)
Jenkins et al. (1999)
Witten et al.
(2005)
Avoletal. (1988)
Roger et al.
(1990)
n
20
19
9
15
15
10
15
37
19
NO2
ppb
300
300
300
350
350
400
400
600
600
Exp.
Min
240
80
30
120
120
180
180
120
80
Time
Post-
Challenge End exp
Type Point min
CARB
METH
SO2
METH
CA
HDM
HDM
COLD
METH
FEVi 30
sRaw 120
sRaw 60
FEVi 90
FEVi 90
FEVi IM
FEVi IM
FEVi 60
sRaw 120
Change
in ARa Average
+ - Air
7e 2e 3.31 ± 8.64e
AFEVi
8d 9d 3.3 ±0.7
4 5 1.25 ±0.23
6 7 7.5 ±2.6
4 11 -6.9 ±1.7
AFEVi
7 3 3.0
8 7 550 ± 240
13e 16e -8.4 ±1.8
AFEVi
11d 8d 3.3 ±0.7
PD±SEb
NO2
-6.98 ± 3.35e
AFEVi
3.3 ±0.8
1.31 ±0.25
7.0 ±3.8
-0.5 ± 1.7
AFEVi
2.78
160 ±60
-10.4 ±2.2
AFEVi
3.7± 1.1
P-
value0
n.s.
n.s.
n.s.
n.s.
<0.05f
0.018
n.s.
n.s.
n.s.
CARB = carbachol; CA = cat allergen; COLD = cold-dry air; Exp. = exposure; FEVi = forced expiratory volume in 1 second;
HDM = house dust mite allergen; HIST = histamine; IM = immediately after exposure; METH = methacholine; NO2 = nitrogen
dioxide; n.s. = less than marginal statistical significance, p > 0.10; SO2 = sulfur dioxide; sRaw = specific airway resistance.
aChange in airway responsiveness (AR): number of individuals showing increased (+) or decreased (-) airway responsiveness
after NO2 exposure compared to air.
bPD ± SE: arithmetic or geometric mean provocative dose (PD) ± standard error (SE). See individual papers for PD calculation and
dosage units. AFEVi indicates the change in FEVi response at a constant challenge dose.
Statistical significance of increase in AR to bronchial challenge following NO2 exposure compared to filtered air as reported in the
original study. Statistical tests varied between studies (e.g., sign test, t-test, analysis of variance).
dNumber of individuals having an increase or decrease in airway responsiveness is from Folinsbee (1992).
eData for 0.25% carbachol challenge from Appendix H of Morrow and Utell (1989b).
'Significantly greater AFEVi in response to a constant challenge dose following exposure to filtered air than NO2 (i.e., a protective
effect of NO2 exposure).
fStudy published since the 2008 ISA for Oxides of Nitrogen.
Fraction of Individuals with Nitrogen Dioxide-Induced Increase in
Airway Responsiveness
Tables 5-1 and 5-2 present all of the studies with data on the fraction of individuals
experiencing a change (increase or decrease) in airway responsiveness following both
NC>2 and filtered air exposures. The statistical significance reported in the original
publications for changes in airway responsiveness following NCh exposure compared to
filtered air is also provided in these tables. Based on all listed studies, the general
5-16
-------
tendency of most studies is toward increased airway responsiveness following NO2
exposure with some studies reaching statistical significance. Fewer studies showed no
effect or a tendency for decreased airway responsiveness following NC>2. Published since
the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). one study reported a statistically
significant decrease in airway responsiveness following NO2, but the authors attributed
the protective effect of NO2 to chance (Riedl etal.. 2012).
Based on the summary data in Tables 5-1 and 5-2. the fraction of individuals
experiencing an NC>2-induced increase in airway responsiveness was assessed in a
manner consistent with the analysis conducted by Folinsbee (1992). Specifically, a
two-tailed sign test was used to assess the statistical significance of directional changes in
airway responsiveness between the NO2 and filter air exposure days. The nonparametric
sign test assumes only that the responses of each subject are independent and makes no
assumptions about the distribution of the response data. This test allows estimation of
whether a statistically significant fraction of individuals experience an increase or
decrease in airway responsiveness, but does not provide information on the magnitude of
the change in that endpoint.
Table 5-3 provides the fraction of individuals experiencing an NCh-induced increase in
airway responsiveness to nonspecific agents. Footnotes for this table indicate the group
from Tables 5-1 and 5-2 that were included in the analyses. For example, in Table 5-3
footnote c, the results for resting exposures (see Table 5-1) to 100 ppb NO2 are for the
33 individuals having an increase in nonspecific responsiveness and the 17 individuals
having a decrease in nonspecific responsiveness in the studies by Ahmed et al. (1983a).
Hazucha etal. (1983). and Orehek etal. (1976). Table 5-3 updates Table 3.1-3 of the
2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) and is consistent with the prior
conclusion that statistically significant increases in nonspecific airway responsiveness
(following resting NO2 exposures) occur in the range of 200 and 300 ppb NO2 for
30-minute exposures and at 100 ppb NO2 for 60-minute exposures in individuals with
asthma. Similar to the sign test, the fraction of individuals experiencing an increase in
airway responsiveness in Table 5-3 is relative to those individuals having a change in
responsiveness and does not consider those individuals having no change in
responsiveness. A fraction of 0.5 indicates an equal number of individuals having
increases and decreases in responsiveness and no effect of NO2 exposure on
responsiveness. Increases in airway responsiveness were not observed following
exercising exposures to NO2. In general, statistically significant effects of NO2 exposure
on airway responsiveness to allergen challenge were not found by Brown (2015). except
at NO2 concentrations over 300 ppb. This may be, in part, due to the small number of
individuals in the analysis. Considering both specific and nonspecific challenges,
statistically significant effects of NO2 on airway responsiveness were again found for
5-17
-------
resting but not exercising exposures (Brown, 2015). Given differing mechanisms of
effect (see discussion of bronchial challenge agent later in this section), preference was
given to the analysis of nonspecific responsiveness (Table 5-3).
Table 5-3 Fraction of individuals with asthma having nitrogen dioxide-induced
increase in airway responsiveness to a nonspecific challenge.
NO2 Concentration (ppb)
[N02]
100 <
200 <
[N02]
= 100
[NO2] < 200
[NO2] < 300
>300
All [NO2]
All Exposures3'13
0.66 (50;
0.66 (87;
0.59(199
0.57 (94
0.60 (380
P
P
:p
:p
:p
= 0.033)
= 0.005)
= 0.011)
= 0.18)
< 0.001)
Exposure with
0.59(17;
0.55(163
0.49(61;
0.54 (241
-
Exercisea'b
p = 0.63)d
:p
p
:p
= 0.21)f
= 1.0)h
= 0.25)
Exposure at
0.66 (50;
0.67 (70;
0.78 (36;
0.73(33;
0.71 (139
P =
P =
P =
P =
Resta'b
0.033)c
0.006)e
0.001)9
0.014)1
;p<0.001)
aData are the fraction of subjects with asthma having an increase in airway responsiveness following NO2 versus air exposure.
Values in parentheses are number of individuals with asthma having a change (±) in responsiveness and the p-value for a
two-tailed sign test.
"•Analysis is for the 380 subjects with asthma in Tables 5-1 and 5-2 having a change (±) in nonspecific airway responsiveness.
°33 increases, 17 decreases; 100 ppb data from Ahmed et al. (1983a). Hazuchaet al. (1983). and Oreheket al. (1976).
d10 increases, 7 decreases; 150 ppb data from Roger et al. (1990).
e47 increases, 23 decreases; 100 ppb data from Ahmed et al. (1983a). Hazuchaet al. (1983). and Oreheket al. (1976): 140 ppb
data from.Bvlin etal. (1988).
'90 increases, 73 decreases; 200 ppb data from Kleinman et al. (1983): 250 ppb data from Jb'rres and Magnussen (1991): 260 ppb
data from Strand etal. (1996): 300 ppb data from Avol et al. (1988). Avol etal. (1989). Bauer et al. (1986). Morrow and Utell
(1989a). Roger etal. (1990). and Rubinstein et al. (1990).
928 increases, 8 decreases; 200 ppb data from Oreheketal. (1976): 250 ppb data from Jorres and Magnussen (1990): 270 ppb
data from Bvlinetal. (1988).
h30 increases, 31 decreases; 350 ppb data from Riedletal. (2012): 600 ppb data from Avol etal. (1988) and Roger etal. (1990).
'24 increases, 9 decreases; 480 ppb data from Bvlin et al. (1985): 500 ppb data from Mohsenin (1987a): 530 ppb data from Bvlin
etal. (1988).
Magnitude and Distribution of Nitrogen Dioxide-Induced Increase in
Airway Responsiveness
Individual subject airway responsiveness data for nonspecific challenges following
resting exposures to filtered air and NO2 were available from five studies (Jorres and
Magnussen. 1990: Bvlinetal.. 1988: Mohsenin. 1987a: Bvlinetal.. 1985: Orehek et al..
1976). Data were obtained for 72 individuals and 116 NCh exposures. Twenty individuals
in the Bvlin etal. (1988) study were exposed to three NO2 concentrations and four
5-18
-------
individuals in the Orehek et al. (1976) study were exposed to two NO2 concentrations.
The dPD due to NO2 for each individual was assessed as:
Equation 5-1
where: PDNo2 and PDan are the PD following NC>2 and air exposures, respectively. Given
that airway responsiveness is recognized as being log-normally distributed (Postma and
Boezen. 2004; Cockcroft et al.. 1983). this method of assessing dPD provides
nonnegative values for log transformation and plotting. The distribution of dPD data
[median and geometric standard deviation (GSD)] was determined for each study and
overall for all subjects as described by Brown (2015).
To assess the potential "adversity" or clinical relevance of changes in dPD, a sign test
was utilized to determine whether there were a statistically greater number of individuals
experiencing a doubling dose reduction in dPD (<0.5) versus those having a doubling
dose increase in dPD (>2). A sensitivity analysis was performed to ensure that no single
study or group of exposures affected the distribution of dPD and assessment of a
doubling dose change. Finally, dose-response was assessed by regressing the logarithms
of dPD against NO2 exposure concentration and against the product of NO2 exposure
concentration and duration.
As described above, the dPD for each individual was calculated as the PD following NO2
divided by the PD following air exposure. Hence, a dPD greater than one suggests
reduced responsiveness, whereas a dPD less than one suggests increased responsiveness
following NC>2 exposure. The dPD from the five studies providing individual PD data
following resting exposures to NC>2 and filtered air are illustrated in Figure 5-1. All of the
median responses illustrated in Figure 5-1 show increased responsiveness following NC>2
exposure (i.e., an NCh-induced reduction in the PD). Note that the dPD values are on a
log scale. The untransformed dPD data from Bylin et al. (1988) and Mohsenin (1987a)
were positively skewed with a few individuals having large values of dPD. This results in
a large difference between the median dPD and arithmetic mean dPD. For example, at the
140 ppb concentration in the Bvlin etal. (1988) study, the median dPD of 0.73 suggests
NC>2 increased responsiveness, which is consistent with 14 individuals having an increase
in responsiveness versus 6 having a decrease, whereas the arithmetic mean dPD of
1.15 erroneously suggests a reduction in responsiveness. The untransformed dPD data
from Bvlin et al. (1985). Torres and Magnussen (1990). and Orehek etal. (1976) were
more symmetrical than that from Bvlin etal. (1988) and Mohsenin (1987a).
5-19
-------
10.0 -T
1.0
0.1
0
ID
00
oo
CQ
ID
O 00
en oo
:O >•
--co
LD Is* 00
00 00 00
>- O >•
CO ^ CO
200
400
600
N02 concentration (ppb)
Note: dPD = provocative dose; NO2 = nitrogen dioxide; ppb = parts per billion. Points illustrate the responses of 72 individual
subjects, and bars are median responses. Doubling dose changes are illustrated by horizontal dotted lines. Data are from Or76
(Oreheketal.. 1976). By88 (Bvlinetal.. 1988). J690 (Jorres and Magnussen. 1990). By85 (Bvlinetal.. 1985). and Mo87 (Mohsenin.
1987a).
Figure 5-1 Change in provocative dose due to exposure to nitrogen dioxide
in resting individuals with asthma.
A clinically relevant NCh-induced increase in responsiveness (dPD <0.5) was observed in
24% of the data, while 8% had a double dose decrease in responsiveness (dPD >2). Of
the 28 responses where dPD was <0.5, 17 were from Bvlin et al. (1988). Of the nine
responses where dPD was >2, eight were again from the Bvlinetal. (1988) study.
Subject 1 in the Bvlinetal. (1988) study had the three highest dPD in Figure 5-1, which
generally reflects the reproducibility of response. For all subjects in the Bylin et al.
(1988) study, the Spearman's rank correlation between the 140 and 530 ppb exposures
was 0.56 (p = 0.01) and was 0.48 (p = 0.03) between the 270 ppb exposure and both the
140 and 530 ppb exposures. Clearly this study had the potential to affect both the
assessment of a doubling dose change in dPD as well as the distribution of responses and
illustrates the necessity of the sensitivity analysis. Both Bylin etal. (1988) and Orehek et
al. (1976) exposed subjects to more than one NO2 concentration. As stated above, a
doubling dose, NO2-induced increase in responsiveness (dPD <0.5), was observed in 24%
of the data. This fraction may be affected by the multiple exposures in these two studies.
5-20
-------
If only a single exposure concentration from the Bylin et al. (1988) study is considered
and only the lower concentration from the Oreheket al. (1976) study is considered, a
doubling dose, NO2-induced increase in responsiveness (dPD <0.5), is observed in
21-22% of the 72 individual subjects in Figure 5-1. The variability in this fraction of
individuals (i.e., 21% or 22%) depends on what parts of the Bvlin et al. (1988) study are
excluded to avoid double counting.
Figure 5-2 illustrates a log-probability plot of the dPD data. The data are log-normally
distributed with an estimated (from fitted line on plot) median of 0.75 and a GSD of 1.88.
The lowest and highest dPD were assigned the cumulative probabilities of 0.1% and
99.9%. Removing these two values did not affect the median and only slightly reduced
the geometric standard deviation from 1.88 to 1.87. Most of the data (namely 69%)
suggests an NCh-induced increase in responsiveness (dPD <1) due to NO2 exposure,
while 24% of the data suggests decrease responsiveness (dPD >1). Even though this 69%
was relative to all responses (i.e., increases, decreases, and no change in responsiveness),
it is remarkably similar to the 71% for all NC>2 concentrations in Table 5-3 having an
increase in responsiveness relative to only those having a change in responsiveness and
ignoring those having no change in responsiveness. Also consistent with the results in
Table 5-3. a two-tailed sign test shows a statistically significant (p < 0.001) reduction in
the dPD in 74% of the 108 dPD responses not equal to one. Of the 37 dPD having more
than a doubling dose change, 76% show a clinically relevant NCh-induced reduction in
dPD (p = 0.003; two-tailed sign test).
5-21
-------
10.0
1 5 10 20 30 50 70 80 90 95 99 99.9
Cumulative Percent Less Than Indicated dPD
Note: dPD = provocative dose. Data are for 72 individuals and 116 NO2 exposures illustrated in Figure 5-1. Line is log-normal fit
(0.75, median dPD; 1.88, geometric standard deviation). Table within figure is the number of observations within intervals of dPD.
Doubling dose changes are illustrated by horizontal dotted lines. The discontinuity between the 70th and 77th percentiles is due to 8
of the 116 dPD being equal to one.
Figure 5-2 Log-normal distribution of change in provocative dose due to
exposure to nitrogen dioxide in resting individuals with asthma.
The sensitivity analysis performed by Brown (2015) showed that the NCh-induced
increase in airway responsiveness overall and the clinically relevant, doubling dose
increase in responsiveness were robust to exclusion of individual studies and subparts of
studies with multiple exposures. Also evaluated in this sensitivity analysis, the
concentration range of the data set was split into roughly halves and thirds to determine
whether effects were more marked for a specific range of concentrations. That analysis
suggested more of an NO2 effect on airway responsiveness following lower concentration
(<140 ppb) exposures.
5-22
-------
Using the full dPD data set of 116 exposures, linear regression did not show an
association between log-transformed dPD and either NO2 concentration (p = 0.44) or
concentration x exposure duration (p = 0.89).
Factors Affecting Airway Responsiveness and Dose-Response
Exercise
In considering why increases in airway responsiveness occurred only after resting
exposure to NCh, Folinsbee (1992) and Bylin (1993) suggested that exercise itself may
affect the mechanisms responsible for increased responsiveness. Recent literature
continues to support the possibility that exercise may lead to a period of reduced airway
responsiveness. The review by O'Byrne et al. (2009) noted with repeated bouts of
exercise, the bronchoconstrictive response to exercise can be abolished in many
individuals with asthma. Refractory periods (i.e., periods during which airway
responsiveness to challenge is diminished) following exercise of 40 minutes to 3 hours
have been reported (Dryden et al.. 2010).
A comparison of two studies that utilized the same challenge agent following the same
duration of NCh exposure and nearly the same exposure concentration supports the
conclusion that exercise may diminish the subsequent responsiveness to bronchial
challenge. Torres and Magnussen (1990) found a statistically significant increase in
airway responsiveness to a 862 challenge in subjects with asthma following exposure to
250 ppb NO2 for 30 minutes at rest; whereas, Rubinstein et al. (1990) found no change in
responsiveness to a SCh challenge following exposure of subjects with asthma to 300 ppb
NC>2 for 30 minutes with 20 minutes of exercise. An effect of exercise refractoriness is
consistent with greater increases in airway responsiveness following resting than
exercising exposures to NC>2 as shown in Table 5-3.
Bronchial Challenge Delivery and Assessment
Variations in methods for administering the bronchoconstricting agents may substantially
affect the results (Cockcroft and Davis. 2006b; Cockcroft et al.. 2005). A repeated
measures study of 55 subjects with asthma evaluating two ATS-recommended methods
of methacholine delivery found a highly statistically significant (p < 0.00001), twofold
difference in PC20 which was attributable to the delivery method (Cockcroft and Davis.
2006b). Even in the same subjects exposed by the same investigators in the same facility
to the same bronchial challenge agent, there can be a doubling dose difference due to the
delivery method. The difference observed by Cockcroft and Davis (2006b) may, in part,
be due to the use of full vital capacity inspirations with breath-hold as part of the delivery
5-23
-------
technique that yielded the higher PC20. The maximal lung inflations are recognized to
induce bronchodilation.
The full vital capacity inspiration required for FEVi measurements when assessing
airway response to challenge may cause a partial reversal of bronchospasm versus the use
of other measures such as specific airway resistance or conductance (Jackson et al.. 2004;
Beaupre and Orehek. 1982; Orehek et al., 1981). Brown (2015) suggested that the use of
forced vital capacity (FVC) maneuvers likely contributed to the lack of statistically
significant effects in NC>2 studies employing exercising exposures. For nonspecific
challenges (Table 5-3). responsiveness was assessed using FVC maneuvers in only 6% of
139 individuals exposed at rest versus 62% of 241 individuals exposed during exercise.
However, inconsistent with the supposition by Brown (2015) for studies with exercise
exposure protocols, Table 5-2 shows that the studies showing statistically significant
effects utilized full FVC maneuvers.
Bronchial Challenge Agent
Bronchial challenge agents differ in the mechanisms by which they cause
bronchoconstriction, acting either "directly" or "indirectly" on bronchial smooth muscle
receptors. Even similarly delivered nonspecific, direct acting agents may affect the lung
differently. In a comparison of responses to methacholine and histamine in healthy
volunteers not having AHR, Verbanck et al. (2001) reported that histamine caused an
overall narrowing of the airways (i.e., similar between parallel lung regions), whereas
methacholine caused a differential narrowing of parallel airways, which altered
ventilation distribution. The differential effects of these two direct acting agents may, in
part, be due to their differing target receptors and the distribution of these receptors in the
airways (O'Byrne et al.. 2009). Comparison of the airway responsiveness among
bronchial challenge agents is complicated by the differing mechanisms by which they
initiate bronchoconstriction.
The lack of statistical significance for changes in responsiveness to allergen challenges
reported by Brown (2015) does not necessarily diminish the potential importance of
allergen exposures. First, use of FVC maneuvers in NO2 studies may have biased results
toward not finding an effect on airway responsiveness, although this did not appear to be
true for studies with exercise protocols as was discussed above. Second, 80% of children
with asthma are thought to be sensitized to common household allergens (O'Byrne et al..
2009). Third, individuals with asthma may experience an early phase response to allergen
challenge with declines in lung function within 30 minutes, which primarily reflects
release of histamine and other mediators by airway mast cells. Approximately half of
those individuals having an early phase response also have a late phase response with a
decline in lung function 3-8 hours after the challenge, which reflects enhanced airway
5-24
-------
inflammation and mucous production (O'Byrne et al., 2009; Cockcroft and Davis, 2006c).
The early response may be reversed with bronchodilators, whereas the late response
requires steroidal treatment. Studies have reported NCVinduced effects on allergen
responsiveness for both the early phase (Jenkins et al.. 1999; Strand et al.. 1998;
Tunnicliffe et al. 1994) and late phase (Strand et al.. 1998; Tunnicliffe et al.. 1994).
These effects were observed following 30-minute resting exposures to concentrations as
low as 260 ppb NO2. Finally, the response to an allergen is not only a function of the
concentration of inhaled allergen, but also the degree of sensitization as measured by the
level of allergen-specific IgE and responsiveness to nonspecific agents (Cockcroft and
Davis. 2006a). These factors make it difficult to predict the level of responsiveness to an
allergen, and although rare, severe bronchoconstriction can occur with inhalation of very
low allergen concentrations (O'Byrne et al.. 2009). It is a concern, given the ubiquity of
allergens and potential severity of responses to allergen inhalation, that NO2 exposure
might augment these responses. The responsiveness to allergens in animals and humans
is also addressed in Sections 4.3.2.6 and 5.2.2.5.
Subject Inclusion/Exclusion
Exercise is a major trigger of asthma symptoms in between 60 and 90% of people with
asthma (Dryden et al.. 2010). In their study of NC>2 effects on airway responsiveness,
Roger etal. (1990) reported that all of their volunteers with asthma experienced either
cold air or exercise-induced bronchoconstriction. Morrow and Utell (1989a) reported
that, "Many of the asthmatic subjects were unable to undertake the carbachol challenge
after either NO2 or air exposures, presumably because of pre-existing exercise-induced
bronchoconstriction." Consequently, in their study, data on changes in airway
responsiveness were only available for 9 of 20 subjects. Thus, the existence of
exercise-induced bronchospasm and symptoms may have caused an underlying
difference in the health status of subjects for which airway responsiveness was evaluated
between studies involving resting versus exercising exposures. Assessing those
individuals with less responsive airways could bias results toward not finding an effect of
NC>2 on airway responsiveness in studies utilizing exercising exposures.
Medication Usage
There was a wide range in restrictions on asthma medication usage among NO2 studies. It
is recommended that short-acting bronchodilators be stopped 8 hours before and
long-acting bronchodilators 36 hours before the bronchial challenge (Reddel et al.. 2009).
Even after withholding salmeterol (a long-acting bronchodilator) for 24 hours, there is
still a greater than twofold reduction in airway responsiveness relative to an unmedicated
baseline (Reddel et al.. 2009). In their NO2 study, Hazuchaetal. (1983) required that
subjects not receive steroid therapy or daily bronchodilator therapy for a month prior to
5-25
-------
bronchial challenge testing. Other NCh study investigators recorded asthma medication
usage and asked subjects to refrain from using for defined periods of time depending on
the medication, such as 8 hours for short-acting bronchodilators [e.g., (Witten etal..
2005; Avol et al.. 1988)]. Restrictions were far less in some studies, for example,
Kleinman et al. (1983) asked subjects to withhold bronchodilators for at least 4 hours
prior to exposure, but subjects were not excluded for noncompliance because medication
usage was generally balanced between filtered air and NO2 exposure days. Still other
studies provided no indication of asthma medications or prohibitions for study inclusion
[e.g., (Bylin et al.. 1988)1. Pretreatment (500 mg, 4 times per day for 3 days) with
ascorbic acid was shown to prevent NCh-induced increases in airway responsiveness of
healthy individuals (Mohsenin. 1987b). Thus, the use of asthma medications or dietary
antioxidants may have reduced the ability of studies to identify an effect of NO2 on
airway responsiveness and may have affected observed provocative doses.
Effect of Challenge Time Following Nitrogen Dioxide Exposure
With respect to the data in Tables 5-1 and 5-2. bronchial challenges were delivered an
average of 60 minutes post-exposure. For nonspecific agents, on average, challenges
were delivered 16 minutes following resting exposures and 67 minutes following exercise
exposures (p < 0.01). Although challenges may take upwards of 40 minutes to complete
(Mohsenin. 1987a). the difference in the time when challenge agents were delivered
could plausibly affect differences in airway responsiveness among studies. The existing
data on airway responsiveness following NO2 exposure are insufficient to assess the
influence of challenge delivery timing on airway responsiveness in those studies.
Effect of Repeated Nitrogen Dioxide Exposures
Two studies evaluated repeated sequential daily exposures to NO2 on airway
responsiveness (Ezratty et al.. 2014; Strand etal.. 1998). From these studies, it is unclear
whether repeated ambient NO2 exposures would have little effect or augment responses
observed.
Extraneous Factors
Although some early studies progressively increased NC>2 exposure concentrations for
safety purposes, the majority of controlled human exposure studies investigating the
effects of NC>2 are of a randomized, controlled, crossover design in which subjects were
exposed, without knowledge of the exposure condition and in random order to clean
filtered air (the control), and depending on the study, to one or more NC>2 concentrations.
The filtered air control exposure provides an unbiased estimate of the effects of the
experimental procedures on the outcome(s) of interest. Comparison of responses
following this filtered air exposure to those following NC>2 exposure allows for estimation
5-26
-------
of the effects of NO2 itself on an outcome measurement while controlling for independent
effects of the experimental procedures.
Dose-/?esponse
Folinsbee (1992) noted that greater NO2 doses occur with exercise due to both the
increased ventilation rates and a tendency for increased exposure duration. However, in
his and other meta-analyses, the effects of NO2 exposure on airway responsiveness were
found following resting exposures, but not exercising exposures to NC>2.
The dose-response of NO2 on airway responsiveness may be modulated by a number of
factors that have been described in this section. The finding of greater airway
responsiveness following exposures at rest compared to exercise, despite a lower intake
dose of NCh during the resting exposures, is consistent with an effect of exercise
refractoriness. Issues related to subject selection and medication may have reduced
observed effects of NO2 on airway responsiveness and contributed to variability within
and among studies. Both the choice of bronchial challenge agent and method of delivery
would also have likely contributed to variability among studies. Methodological
differences, if randomly occurring, among studies such as the choice of challenge agents,
challenge delivery method, and asthma medication usage would likely add variability to
assessment of airway responsiveness and thereby bias data toward the null of no
discernible dose-response.
A few studies investigated the effects of NO2 exposure on airway responsiveness at more
than one concentration during resting exposures (Tunnicliffe et al.. 1994; Bylin et al..
1988; OreheketaL 1976). However, these studies provide, at best, limited support for
increasing airway responsiveness with increasing NC>2 concentration in individuals with
asthma. Additionally, linear regression performed by Brown (2015) did not show an
association between log-transformed dPD in Figure 5-1 and either NO2 concentration
(p = 0.44) or concentration x exposure duration (p = 0.89). The study by Roger et al.
(1990) that used an exercise protocol also showed no dose-response relationship. In
reviewing a draft of this ISA, a CASAC Panel member commented that a dose-response
relationship would not necessarily be expected to be identified from a relatively small
group of only 72 individuals [see page A-8 of (Roux and Frey. 2015)].
Summary
There is a wide range of airway responsiveness influenced by many factors, including
exercise, medications, cigarette smoke, air pollutants, respiratory infections, disease
status, and respiratory irritants. In the general population, airway responsiveness is
log-normally distributed with individuals having asthma generally being more responsive
5-27
-------
than healthy age-matched controls. Nonspecific bronchial challenge agents causing
bronchoconstriction may act directly (i.e., histamine, carbachol, and methacholine) on
airway smooth muscle receptors or act indirectly (i.e., exercise, cold air) though
intermediate pathways, especially via inflammatory mediators. Specific challenge agents
(i.e., allergens) also act indirectly on smooth muscle to initiate bronchoconstriction.
Likely affecting the observed changes in airway responsiveness due to NO2 exposure,
there are methodological differences among NO2 studies including subject activity level
(rest versus exercise) during NO2 exposure, asthma medication usage, choice of airway
challenge agent (e.g., direct and indirect nonspecific stimuli), method of administering
the bronchoconstricting agents, and the physiological endpoint used to assess airway
responsiveness. Most of these intra-study differences likely contributed to variability and
uncertainty in comparison among studies of provocative doses and lung function
responses to bronchial challenge agents.
The analyses excerpted from Brown (2015) show that the airway responsiveness of
individuals with asthma is increased by brief exposures to NC>2. There was a statistically
significant fraction of individuals with asthma exposed to NCh at rest who experienced an
increase in responsiveness. About 70% had an increase in nonspecific airway
responsiveness following 30-minute exposures to NCh in the range of 200 to 300 ppb and
following 60-minute exposures to 100 ppb. The median response of these individuals is
an NO2-induced reduction in dPD to 0.75 (1.88, geometric standard deviation). About a
quarter of the exposed individuals experienced a clinically relevant, doubling dose
reduction in their dPD due to NC>2 exposure. The fraction experiencing a doubling dose
increase in responsiveness was also statistically significant and robust to exclusion of
individual studies. Results showed minimal change in airway responsiveness for
individuals exposed to NCh during exercise.
5.2.2.2 Lung Function Changes in Populations with Asthma
The lung function endpoint described in this section differs from airway responsiveness
in Section 5.2.2.1 in that subjects are not given a bronchoconstricting agent prior to
measurement of airflow or volume. Compared with evidence for airway responsiveness,
the 2008 ISA for Oxides of Nitrogen reported weak evidence in controlled human
exposure studies for the effects of NO2 exposure on lung function changes in adults with
asthma (U.S. EPA. 2008c). Epidemiologic evidence in people with asthma also was
weak. Most recent studies were epidemiologic and support associations between ambient
NC>2 concentrations and lung function decrements in children with asthma.
5-28
-------
Epidemiologic Studies
Collectively, previous and recent studies found associations between increases in ambient
NO2 concentrations and decrements in supervised spirometry measures (primarily FEVi)
in children with asthma. Across the various populations examined, results are less
consistent for lung function measured under unsupervised conditions, primarily peak
expiratory flow (PEF) at home. Results also are inconsistent for NO and NOx. Ambient
concentrations of NC>2, locations, and time periods for epidemiologic studies of lung
function are presented in Table 5-4.
Table 5-4 Mean and upper percentile concentrations of nitrogen dioxide in
epidemiologic studies of lung function in populations with asthma.
Study3
Location
Study Period
Upper
Mean/Median Percentile
NO2 Metric Concentration Concentration
Analyzed (ppb) (ppb)
fDelfino et al.
(2008a)
tSmarqiassi et al.
(2014)
fO'Connor et al.
(2008)
Riverside, CA
Whittier, CA
Montreal, QC, Canada
Boston, MA;
Bronx, NY;
Chicago, IL;
Dallas, TX;
New York, NY;
Seattle, WA;
Tucson, AZ
Jul-Dec 2003
Jul-Dec 2004
Oct 2009-
Apr2010
Aug 1998-
Jul2001
24-h avg total 28.6
personal
24-h avg 25.0
central site
24-h avg total 6.3
personal
24-h avg NR
Max: 105.7
Max: 29.2
75th: 7.4
Max: 70.6
NR
tGillespie-Bennett
etal. (2011)
Bluff, Christchurch, Sep 2006
Dunedin, Porirua, Hutt
Valley, New Zealand
4-week avg
3.9
NR
tWiwatanadate and Chiang Mai, Thailand
Trakultivakorn
(2010)
Aug 2005- 24-h avg
Jun 2006
17.2 90th: 26.5
Max: 37.4
5-29
-------
Table 5-4 (Continued): Mean and upper percentile concentrations of nitrogen
dioxide in epidemiologic studies of lung function in
populations with asthma.
Study3
Mortimer et al.
(2002)
Just et al. (2002)
tOdajima et al.
(2008)
Delfinoetal. (2003)
tHolquin et al.
(2007)
fBarraza-Villarreal
et al. (2008)
tLiuetal. (2009b)
Dales et al. (2009a)
tHernandez-
Cadena et al.
(2009)
fMartins et al.
(2012)
fGreenwald et al.
(2013)
tSpira-Cohen et al.
(2011)
fYamazaki et al.
(2011)
Location
Bronx and East
Harlem, NY;
Chicago, IL;
Cleveland, OH;
Detroit, Ml;
St. Louis, MO;
Washington, DC
Paris, France
Fukuoka, Japan
Los Angeles, CA
(Huntington Park area)
Ciudad Juarez, Mexico
Mexico City, Mexico
Windsor, ON, Canada
Mexico City, Mexico
Viseu, Portugal
El Paso, TX
Bronx, NY
Yotsukaido, Japan
Study Period
Jun-Aug 1993
Apr-Jun 1996
Apr-Sep 2002
Oct 2002-
Mar2003
Nov 1999-
Jan 2000
2001-2002
Jun 2003-
Jun 2005
Oct-Dec 2005
May-Sep
2005
Jan and Jun
2006 and 2007
Mar-Jun 2010
Spring 2002,
Spring/Fall
2004, Spring
2005
Oct-Dec 2000
NO2 Metric
Analyzed
4-h avg
(6 a.m.-
10 a.m.)
24-h avg
3-h avg
I' P-m-
10 p.m.)
1-h max
8-h max
1-week avg
8-h max
24-h avg
1-h max
1-weekavgb
96-h avg
6-h avg
(9 a.m. -3 p.m.)
1-h avg
(6 p.m.-7 p.m.)
Mean/Median
Concentration
(PPb)
NR
28.6c
20.0
11.0
7.2
5.9
18.2
37.4
19.8
57
Across 4 periods:
4.5, 3.5, 9.8, 8.2C
School A: 6.5
SchoolB: 17.5
NR
32.6
Upper
Percentile
Concentration
(PPb)
NR
Max: 59. Oc
Max: 51.3
Max: 49.0
90th: 9 Max: 14
90th: 8 Max: 11
NR
Max: 77.6
95th: 29.5
75th: 69
Max: 116
Max across
4 periods: 4.6,
4.0, 10.9, 9.4C
NR
NR
NR
5-30
-------
Table 5-4 (Continued): Mean and upper percentile concentrations of nitrogen
dioxide in epidemiologic studies of lung function in
populations with asthma.
Study3
tQian et al. (2009a)
Laqorio et al. (2006)
McCreanor et al.
(2007)
tMaestrelli et al.
(2011)
fCanova et al.
(2010)
Hiltermann et al.
(1998)
tWiwatanadate and
Liwsrisakun (2011)
Parketal. (2005)
Location
Boston, MA;
New York, NY;
Madison, Wl;
Denver, CO;
Philadelphia, PA;
San Francisco, CA
Rome, Italy
London, U.K.
Padua, Italy
Padua, Italy
Bilthoven, the
Netherlands
Chiang Mai, Thailand
Incheon, South Korea
Study Period
Feb1997-
Jan 1999
May-Jun,
Nov-Dec 1999
Nov-Mar
2003-2005
1999-2003
Summer/Fall
2004, Winter/
Summer/Fall
2005
July-Oct1995
Aug 2005-
Jun 2006
Mar-Jun 2002
NO2 Metric
Analyzed
24-h avg
24-h avg
2-h avg
(10:30 a.m.-
12:30 p.m.)
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
Mean/Median
Concentration
(PPb)
20.8
37.6C
Oxford St: 75.5C
Hyde Park: 11. 5C
Across seasons
and years:
20.9-37.0C
27.2C
11. 2C
17.2
Control days:
31.6
Dust days: 20.7
Upper
Percentile
Concentration
(PPb)
75th: 28.8
Max: 60.7
Max: 54. 3C
Max: 154C
Max: 77.7C
Range of 75th:
23.0^2.5C
48.1c
22.5C
90th: 26.5
Max: 37.4
NR
a.m. = ante meridiem; Aug = August; avg = average; AZ = Arizona; CA = California; CO = Colorado; DC = Disrict of Columbia;
Dec = December; Feb = February; IL = Illinois; MA = Massachusetts; Ml = Michigan; MO = Missouri; NO2 = nitrogen dioxide;
NR = not reported; NY = New York; OH = Ohio; ON = Ontario; PA = Pennsylvania; QC = Quebec; TX = Texas; WA = Washington;
Wl = Wisconsin; UK= United Kingdom.
aStudies presented in order of first appearance in the text of this section.
""Subject-level exposure estimates calculated from outdoor NO2 at schools and other locations plus time-activity patterns.
Concentrations converted from |jg/m3 to ppb using the conversion factor of 0.532 assuming standard temperature (25°C) and
pressure (1 atm).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
Children with Asthma
In contrast with studies reviewed in the 2008 ISA for Oxides of Nitrogen (U.S. EPA.
2008c), several recent studies of children with asthma conducted spirometry under
5-31
-------
supervised conditions, and most indicate a relationship with short-term NC>2 exposure
(Figure 5-3 and Table 5-5). Studies of supervised spirometry measured lung function
daily, weekly, biweekly, or seasonally, whereas lung function at home was measured
daily. Among the latter studies, some reported an association with NO2 (Gillespie-
Bennettetal., 2011; Delfino et al.. 2008a; O'Connor et al.. 2008). whereas others did not
(Wiwatanadate and Trakultivakorn. 2010; Odaiima et al.. 2008; Just et al.. 2002;
Mortimer et al.. 2002). Several studies that reported no association with home lung
function measurements did not provide quantitative results, including NCICAS (Odajima
et al.. 2008; Delfino etal.. 2003; Just et al.. 2002; Mortimer etal.. 2002). Thus, the
relative magnitude and precision of their results cannot be assessed. Results also are
inconsistent among U.S. multicity studies [National Cooperative Inner-city Asthma Study
(NCICAS), Inner City Asthma Study (ICAS)] (O'Connor et al.. 2008; Mortimer et al..
2002). However, a relationship between ambient NO2 and PEF is indicated in children
with asthma in a recent meta-analysis ("Weinmayr et al.. 2010) that included mostly
European studies as well as some studies reviewed in the 2008 ISA for Oxides of
Nitrogen.
5-32
-------
Study
NO2 Metrics Analyzed
McCreanor et al. (2007) 2-h avg, lag 0 h
Smargiassi et al. (2014) 24-h avg, lag 0 day
Greenwald etal. (2013) 24-h avg,
lag 0-3 day avg
Holguin et al. (2007)
Martins et al. (2012)
Spira-Cohen et al.
Liu et al. (2009)
Barraza-Villarreal et al.
(2008)
24-h avg,
lag 0-6 day avg
24-h avg,
lag 0-4 day avg
6-h avg, lag 0 day
24-h avg, lag 0 day
1-h max,
lag 1 -4 day avg
Personal outdoor
Personal total
School
School
Central sites
Central site
^
^-
I . ,
I
|
I
— d —
I
-•r
4
Delfino et al. (2008)b 24-h avg, lag 0 day Total personal
Dales et al. (2009)b 12-h avg, lag 0 day Central sites
O'Connor et al. (2008)b 24-h avg, Central sites
lag 1 -5 day avg
Maestrelli et al. (2011) 24-h avg, lag 0 day Central sites
Lagorio et al. (2009) 24-h avg, lag 0 day Central sites
I
-25 -20 -15 -10 -50 5 10 15 20 25 30
Percent change in FEV1 per increase in NO2 (95% Cl)a
All subjects 4
No bronchodilator 4
bronchodilator -"
-25 -20 -15 -10 -50 5 10 15 20 25 30
Change in % predicted FEV1 per increase in NO2 (95% Cl)a
Note: avg = average; Cl = confidence interval; FENA, = forced expiratory volume in 1 second; h = hour; max = maximum;
NO2 = nitrogen dioxide. Black = studies from the 2008 Integrated Science Assessment for Oxides of Nitrogen, red = recent studies.
Results are separated into two plots for the two most common indices of FEVi examined in studies. Results from more informative
studies in terms of the exposure assessment method and potential confounding considered are presented first in each plot. Study
details and quantitative results are reported in Table 5-5. Table 5-5 presents results for an array of lung function indices; some of
these did not have sufficient numbers to present in a figure.
aEffect estimates are standardized to a 20-ppb increase for 24-h avg NO2 and a 30-ppb increase in 1-h maximum NO2. Effect
estimates for 1-h average to 12-h average NO2 are not standardized but presented as they are reported in their respective studies
(Section 5.1.2.2).
"•Studies with home-based FENA, measurements. All other studies conducted supervised spirometry.
Figure 5-3 Associations of nitrogen dioxide ambient concentrations or
personal exposure with percentage change in forced expiratory
volume in 1 second (top plot) and change in percent predicted
forced expiratory volume in 1 second (bottom plot) in children
and adults with asthma.
5-33
-------
Table 5-5 Epidemiologic studies of lung function in children and adults with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics
Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
Children with asthma: studies with small spatial scale exposure assessment and/or examination of copollutant confounding
tDelfinoetal. (2008a)
Riverside, Whittier, CA
n = 53, ages 9-18 yr, persistent asthma and exacerbation in
previous 12 mo
Repeated measures. Home spirometry; measurements
checked daily by research staff. Examined daily for 10 days.
519 observations. Recruitment by referral from school nurses.
Parent report of physician-diagnosed asthma. Nonsmokers
from nonsmoking homes. No information on participation rate.
Mixed effects model with random effect for subject with
pollutant concentrations centered on subject mean and
adjusted for personal relative humidity, personal temperature,
and follow-up period. Adjustment for city, beta agonist use,
weekend, gas stove use did not alter results.
NO2-total personal
24-h avg
Monitoring checked
daily; all samples
above detection limit
of 2.1 ppb (Staimer et
al.. 2005)
Central site and
personal NO2
moderately correlated.
r=0.43.
0-1 avg
% predicted FEVi
All subjects
-1.7 (-3.2,-0.19)
All subjects
-1.5 (-2.3, -0.57)
No bronchodilator, n = 37
-1.7 (-2.7, -0.75)
NO2-central site
24-h avg
Site within 8 or 16 km
of homes.
Bronchodilator use, n
-0.70 (-2.9, 1.5)
= 16
All subjects
-1.3 (-2.4,-0.15)
With 1-h maxPM2.s:
-1.3 (-2.8, 0.22)
Moderate correlation with NO2.
• Spearman r= 0.38 for
personal PM2.5, 0.36 for central
site PM2.5.
. PM2.5 not altered by
adjustment for NO2.
EC, OC not associated with
_FEVi.
Central site NO2 with personal
PM2.5: -0.86 (-2.6, 0.89).
5-34
-------
Table 5-5 (Continued): Epidemiologic studies of lung function in children and adults with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics
Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
tSmarqiassi et al. (2014)
Montreal, QC, Canada
n = 72, ages 7-12 yr, 29% with ED visit in previous 12 mo, 43%
using steroid medication during study
Repeated measures. Supervised spirometry. Examined daily
for 10 days. 700 observations. Residence near oil refineries &
high traffic areas. Recruitment from asthma clinic or schools.
Asthma ascertained by respirologist or parental report of
physician diagnosis. No information on participation rate. Linear
mixed effects models with random effect for subject, random
and fixed effect for study day and adjusted for age, sex, height,
month, day of week, asthma medication use, parental
education, ethnicity, personal temperature, personal humidity.
NO2-total personal
24-h avg
99% samples above
limit of detection
65% time spent
indoors
FEVi: 0.56% (-1.8, 2.9)
FVC: 0.36% (-1.4, 2.2)
FEF25-75%:
0.35% (-4.7, 5.4)
No copollutant model.
No consistent associations
with personal PlVh.s, benzene,
total polycyclic aromatic
hydrocarbons.
Correlations among
pollutants = -0.11 to 0.11.
tMartins et al. (2012). Martins (2013)
Viseu, Portugal
n = 51, mean age 7.3 (SD: 1.1)yr, 53% with atopy.
Repeated measures. Supervised spirometry. Four
measurements over two different seasons. Recruitment from
urban and suburban schools. -66% participation rate. Parental
report of wheeze in previous 12 mo. GEE adjusted forage, sex,
parental smoking, parental education, atopy, time of visit,
average temperature, relative humidity. Also included height,
weight, older siblings, mold or dampness in home, fireplace in
home, pets in home because they changed at least one
pollutant effect estimate >10%.
NO2-modeled personal 0-4 avg
outdoor
24-h avg
Estimated from school
outdoor NO2, 20 city
locations,
MM5/CHIMERE
modeling, and daily
activity patterns.
20% time spent at
school, 65% at home.
FEVi:
-22% (-38,-1.5)
FEV-i/FVC:
-10% (-20, 0.83)
FEF25-75%:
-33% (-54, -2.6)
FEVi after bronchodilator:
19% (3.5, 37)
ForFEV-i:
with PMio: -27% (-67, 60)
with benzene:
-3.6% (-29, 31)
with ethylbenzene:
-17% (-41, 17)
Benzene unaltered by
adjustment for NO2.
Ethylbenzene & PMio
attenuated.
Correlations with NO2 negative
or weakly positive. Spearman
r= -0.82 to -0.55 for PMio,
-0.42 to 0.14 for VOCs.
5-35
-------
Table 5-5 (Continued): Epidemiologic studies of lung function in children and adults with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics
Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
tGreenwald et al. (2013)
El Paso, TX
n = 38, mean age 10 yr, 76% Mexican-American
Repeated measures. Supervised spirometry. Examined weekly
for 13 weeks. 413-441 observations. Recruitment from schools
in low- and high-traffic area. No information on participation
rate. School record of physician-diagnosed asthma. GLM with
subject as random effect and adjusted for school, temperature,
relative humidity, indoor NO.
NO2-school outdoor
School A: residential 0-3 avg
area
School B: 91 m from
major road.
NO2-school indoor
Most samples above
limit detection of 2.88
ppb (Ravsoni et al..
2011)
All 24-h avg
FEVi:
School A: 25% (-15, 84)
School B:
-17% (-32, 0.12)
School A: 38% (-12, 116)
SchoolB: -14% (-32, 7.2)
No copollutant model.
BC, SO2 (central site)
associated with FEVi.
Moderate correlation with NO2.
Pearson r= 0.62, -0.22.
• School BTEX associated with
FEVi, highly correlated with
NO2. r=0.77.
tHolquin et al. (2007)
Ciudad Juarez, Mexico
n = 95, ages 6-12 yr, 78% mild, intermittent asthma, 58% atopy
Repeated measures. Supervised spirometry. Examined
biweekly for 4 mo. 87% participation. Self-report of physician-
diagnosed asthma. Linear and nonlinear mixed effects model
with random effect for subject and school adjusted for sex,
body mass index, day of week, season, maternal and paternal
education, passive smoking exposure.
NO2-school outdoor
24-h avg
Schools located
239-692 m from
homes.
0-6 avg FEVi: -2.4% (-5.1, 0.24)
No copollutant model.
No association with PM2.5, EC.
Weak to moderate correlations
with NO2. Spearman r= 0.30
for PM2.5, 0.49 for EC.
Road density at home not
school associated with lung
function.
tSpira-Cohen (2013), Spira-Cohen et al. (2011)
Bronx, NY
n = 40, ages 10-12 yr, 100% nonwhite, 44% with asthma ED
visit or hospital admission in previous 12 mo
Repeated measures. Supervised spirometry. Examined daily
for 1 mo. 454 observations. No information on participation
rate. Recruitment from schools by referrals from school nurses.
Parental report of physician-diagnosed asthma. Mixed effects
model with subject as random effect adjusted for height, sex,
temperature. Adjustment for school (indicator of season) did not
alter results.
NO2-school outdoor 0
6-h avg
(9 a.m.-3 p.m.)
Schools 53-737 m
from highways with
varying traffic counts.
Most children walk to
school.
89% time spent
indoors.
FEVi: 0.56% (-3.9, 5.1)
PEF: 2.2% (-2.4, 6.8)
Per60-ppb increase NO2
(5th-95th percentile
change)
NO2 effect estimate adjusted
for personal EC not reported.
Personal EC associated with
lung function and not altered
by NO2 adjustment.
Personal EC-School NO2
correlation NR. School EC-
School NO2 moderately
correlated. r= 0.36.
5-36
-------
Table 5-5 (Continued): Epidemiologic studies of lung function in children and adults with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics
Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
tGillespie-Bennett et al. (2011)
Bluff, Dunedin, Christchurch, Porirua, Hutt Valley, New Zealand
n = 358, ages 6-1 Syr
Cross-sectional. Home spirometry. Multiple measures of lung
function, one NO2 measurement. Recruitment from a home
heating intervention. 77% participation. Mixed effects model
with log-transformed NO2 and random effect for subject.
Adjustment for age, sex, region, ethnicity, intervention, income,
temperature did not alter results.
NO2-home outdoor
4-week Per log increase NCb:
av9 Evening FEVi
-88 (-191, 15) mL
No copollutant model.
No other pollutants examined.
NO2-home indoor
Mean 6.1 ppb; no
information on limit of
detection
Evening FEVi
-13 (-26, -0.38) mL
tLiu(2013), Liuetal. (2009b)
Windsor, ON, Canada
n = 182, ages 9-14 yr
Repeated measures. Supervised spirometry. Examined weekly
for 4 weeks, same day of week. 672 observations. Recruitment
from schools. No information on participation rate. Parental
report of physician-diagnosed asthma. Mixed effect model with
random effect for subject and adjusted for testing period,
temperature, relative humidity, daily medication use.
NO2-central site 0
24-h avg
Average of 2 sites
99% subjects live
within 10 km of sites. 0-2 avg
FEVi:-1.2% (-3.2, 0.84)
FEF25-75%:
-4.8% (-8.6,-0.94)
FEVi: -2.3% (-5.5, 0.92)
FEF25-75%:-8.0(-14, -1.6)
For lag 0-2 avg NO2 and FEVi
withPM2.5: 1.2% (-3.8, 6.4)
withSO2: -1.5% (-4.9, 2.2)
PM2.5 association not altered
by NO2 adjustment, SO2
attenuated. NO2 highly
correlated with PM2.5.
Spearman r= 0.71 for PM2.5,
0.18forSO2.
tDalesetal. (2009a)
Windsor, ON, Canada
n = 182, ages 9-14 yr
Repeated measures. Same cohort as above. Unsupervised
peak flow. Examined daily for 4 weeks, same day of week.
672 observations. Recruitment from schools. No information on
participation rate. Parental report of physician-diagnosed
asthma. Mixed effect model with random effect for subject and
adjusted for sex, testing period, day of week, daily mean
temperature, relative humidity, time spent outdoors.
NO2-central site
12-h avg
(8 a.m.-8 p.m.)
Average of 2 sites;
99% subjects live
within 10 km of sites.
Mean 1.6 and
2.2 h/day spent
outdoors.
Evening % predicted FEVi: Copollutant model results only
-0.10 (-0.31, 0.10)
Diurnal change FEVi:
-0.34% (-0.64, -0.04)
Per 9.8 ppb increase in
NO2 (interquartile range)
in figure.
Evening FEVi: NO2 becomes
positive with PM2.5 adjustment.
Diurnal change FEVi: NO2 not
altered by adjustment for PM2.5
or SO2. SO2 and PM2.5 not
altered by adjustment for NO2.
NO2 highly correlated with
PM2.5. Pearson r= 0.68.
5-37
-------
Table 5-5 (Continued): Epidemiologic studies of lung function in children and adults with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics
Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3 Copollutant Examination
Children with asthma: studies with central site exposure assessment and no examination of copollutant confounding
Delfinoetal. (2003)
Los Angeles, CA (Huntington Park area)
n = 22, ages 10-16 yr, 100% Hispanic, 27% on anti-
inflammatory medication
Repeated measures. Home peak flow. Recruitment from
schools. 92% follow-up participation. Nonsmoking children from
nonsmoking homes. Self or parental report of physician-
diagnosed asthma. General linear mixed effects model with
autoregressive parameter and subject specific intercept and
adjusted for respiratory infections. Adjustment for weekend or
max temperature did not alter results.
NO2-central site 0
8-h max 1
1 site within 4.8 km of
home and school.
No quantitative data. Only
reported no statistically
significant association with
PEF.
No copollutant model.
Associations found with EC,
OC, PM-iobutnotVOCs.
Moderate to high correlations
with 8-h max NO2.
Spearman r = 0.38 (PM-io) to
0.62 (OC). For VOCs, r= 0.57
(benzene) to 0.72 (xylene).
tBarraza-Villarreal et al. (2008)
Mexico City, Mexico
n = 163-179, ages 6-14 yr, 54% persistent asthma, 89% atopy
Repeated measures. Supervised spirometry. Examined every
15 days for mean 22 weeks. 1,503 observations. No
information of participation rate. Recruitment from pediatric
clinic. Asthma severity assessed by pediatric allergist. Linear
mixed effects model with random effect for subject and
adjusted for minimum temperature, time, sex, body mass index,
ICS use. Adjustment for outdoor activity, smoking exposure,
allergy medication, season did not alter results.
NO2-central site
1-h max
Site within 5 km of
school or home.
Low correlation for
school vs. central site:
Spearman r= 0.21
1-4avg FEVi:0% (1.1, 1.04)
FVC: -0.11% (-1.2, 0.95)
No copollutant model
1 PM2.5 associated with FEV-i
and FVC.
Moderate correlation with NO2.
Pearson r= 0.61
tHernandez-Cadena etal. (2009) NO2-central site
Mexico City, Mexico 1-h max
n = 85, ages 7-12 yr, 62% mild, intermittent asthma, 90% atopy site within 5 km of
FEVi response to
bronchodilator:
-39% (-64, 5.4)
Cross-sectional. Supervised spirometry. Recruitment from
asthma and allergy clinic. Atopy and asthma severity assessed
at clinic. Linear regression adjusted for sex, pet ownership in
previous 12 mo, visible mold in home, lag 1 max temperature.
Adjustment for age and passive smoking exposure did not alter
results. Did not examine potential confounding by SES.
home or school.
24-h avg and 8-h max
similar results but less
precise.
No copollutant model.
03, not PM2.5 associated with
FEVi response.
03 moderately correlated with
NO2. r=0.35.
5-38
-------
Table 5-5 (Continued): Epidemiologic studies of lung function in children and adults with asthma.
Study
Population Examined and Methodological Details
Mortimer et al. (2002)
Bronx and East Harlem, NY; Chicago, IL; Cleveland, OH;
Detroit, Ml; St. Louis, MO; Washington, DC (NCICAS)
n = 846, ages 4-9 yr
Repeated measures. Home peak flow. Examined daily for four
2-week periods. Recruitment from ED visits and clinics. Parent
report of physician-diagnosed asthma and symptoms in
previous 12 mo or asthma symptoms for >6 weeks, or family
history of asthma. Participation from 55% full cohort. Sample
representative of full cohort except for greater asthma
medication use. Mixed effects model adjusted for city, follow-up
period, day of study, 24-h rainfall, 12-h avg temperature.
tO'Connoret al. (2008)
Boston, MA; Bronx, NY; Chicago, IL; Dallas, TX; New York, NY;
Seattle, WA; Tucson, AZ (ICAS)
n = 861, ages 5-12 yr, persistent asthma and atopy, 82% black
or Hispanic
Repeated measures. Home spirometry. Examined for 2 weeks
every 6 mo for 2 yr. 70% of maximum data obtained. Recruited
from intervention study. Mixed effects model adjusted for site,
month, sitexmonth interaction, temperature, intervention group.
NO2 Metrics
Analyzed
NO2-central site
4-h avg
(6 a.m.-10 a.m.)
Average of all city
monitors.
NO2-central site
24-h avg
All monitors close to
home and not near
industry. Median
distance to
site = 2.3 km.
Lag Day
Single-
day lags
1 to 6
1 -5 avg
1 -4 avg
0-4 avg
0—3 avg
1 -5 avg
Effect Estimate (95% Cl)
Single-Pollutant Model3
No quantitative data. Only
reported no association
with PEF.
% predicted FEVi:
-1.3 (-1.9, -0.78)
% predicted PEF:
-1.6 (-2. 2, -1.1)
Copollutant Examination
No copollutant model.
Os associated with PEF. Weak
correlation with NO2. r= 0.27.
Only three-pollutant model
analyzed with PM2.5 and Os.
Associations also found with
PM2.5, CO, SO2, and Os.
Moderate correlations with
NO2.r =0.59 for PM2.5, 0.54
for CO, 0.59 for SO2. Weak
correlation with Os. r= -0.31.
tYamazaki et al. (2011)
Yotsukaido, Japan
n = 17, ages 8-15 yr, 100% atopy
Repeated measures. Supervised peak flow before medication
use. Examined daily during long-term hospital stay. No air
conditioning in hospital. Permitted to go outside if asthma
stable. Poor generalizability. 1,198 observations. GEE adjusted
for sex, age, height, temperature, day of week, temporal trends.
NO2-central site
1 -h avg
(various times of day)
Monitor adjacent to
hospital. No major
roads nearby.
No quantitative data. PEF
decreases with increasing
NO2 in preceding 0 to
23 hours.
Strongest associations at
Oh and 12 h.
Only three-pollutant model
analyzed with PM2.5 and Os.
PM2.5, not Os, also associated
with evening PEF.
PM2.5 moderately to highly
correlated with NO2. r= 0.54 to
0.78 depending on time of day.
5-39
-------
Table 5-5 (Continued): Epidemiologic studies of lung function in children and adults with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics
Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
Just et al. (2002)
Paris, France
n = 82, ages 7-15 yr, asthma attack in previous 12 mo and
daily asthma medication use, 90% atopy
Repeated measures. Home peak flow. Examined daily for
3 mo. 82% participation. Recruitment from hospital outpatients.
GEE adjusted for time trend, day of week, pollen, temperature,
humidity.
NO2-central site
24-h avg
Average of 14 sites
NR No quantitative data. Only
reported no relationship
with PEF.
No copollutant model.
Os associated with PEF. No
correlation with NO2. Pearson
r=0.09.
tOdaiima et al. (2008) NCb-central site
Fukuoka, Japan 3-h avg
n = 70, ages 4-11 yr, 66% with asthma exacerbation 24-h avg
Repeated measures. Home peak flow. Examined daily for 1 yr. 1 site
>15,000 observations. Participation rate not reported.
Recruitment from hospital where received treatment. GEE
adjusted forage, sex, height, growth of child, temperature.
No quantitative data. Only
reported no association
with PEF.
Only three-pollutant model
analyzed with suspended PM
and Os.
Suspended PM associated
with PEF in warm season.
Weak correlation with NO2.
r= 0.30 for 24-h avg.
tWiwatanadate and Trakultivakorn (2010)b
Chang Mai, Thailand
n = 31, ages 4-11 yr, asthma plus symptoms in previous
12 mo, 52% mild intermittent
Repeated measures. Home peak flow. Examined daily for 1 yr.
97% participation. Recruitment from allergy clinic. GLM with
random effect for subject and adjusted for time trend, day of
week, height, weight, atmospheric pressure, temperature,
sunshine duration.
NO2-central site
24-h avg 0
1 site within 25 km of 1
homes.
PEF
-1.8 (-5.4, 1.8)L/min
2.6 (-1.2, 6.4)L/min
No copollutant model.
No consistent (across various
lags of exposure) associations
found for PM2.5, CO, PM-io,
SO2, or 03.
5-40
-------
Table 5-5 (Continued): Epidemiologic studies of lung function in children and adults with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics
Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3 Copollutant Examination
Adults with Asthma: studies with small spatial scale exposure assessment and/or examination of copollutant confounding
McCreanor et al. (2007)
London, U.K.
n = 31 mild asthma, 32 moderate asthma, ages 19-55 yr, all
with airway hyperresponsiveness, 84% with atopy
Randomized cross-over natural experiment. Supervised
spirometry. Exposure on busy road and park. 55 observations.
Participation rate not reported. Recruitment from
advertisements and volunteer databases. Mixed effects model
with random effect for subject and adjusted for temperature,
relative humidity.
NO2-personal outdoor 0-h
2-h avg
Measured next to
subjects during
outdoor exposures.
FEVi:
-0.22% (-0.40, -0.05)
FEF25-75%:
-0.78% (-1.4, -0.13)
22-h FEVi:
Post. -0.13% (-0.35, 0.10)
exposure FEF25-75%:
-0.75% (-1.6, 0.10)
per 5.3 ppb NO2
For FEF25-75%
with UFP: -0.47% (-1.3, 0.39)
with EC: -0.43% (-1.1, 0.26)
with PM2.s: -0.48% (-1.3, 0.3)
Moderate correlations with
NO2. Spearman r= 0.58 for
UFP and EC, 0.60 for PM2.5.
tQian et al. (2009b)
Boston, MA; New York, NY; Denver, CO; Philadelphia, PA; San
Francisco, CA; Madison, Wl
n = 119, ages 12-65 yr, persistent asthma, nonsmokers
Repeated measures. Home PEF. No information on
participation rate. Study population representative of full cohort.
Examined daily for 16 weeks. >14,000 observations. Trial of
asthma medication, a priori comparison of medication regimen.
Linear mixed effects model adjusted forage, sex, center,
season, race/ethnicity, week, daily average temperature and
humidity. Adjustment for viral infections did not alter results.
NO2-central site
24-h avg
Average of all
monitors within 32 km
of subject ZIP code
centroid.
PEF
All subjects
-0.68% (-1.3, -0.06)
Placebo
-0.29% (-1.4, 0.80)
Beta-agonist
-1.1% (-2.1, -0.05)
ICS
-0.61% (-1.7, 0.39)
with SO2: -0.11% (-0.87, 0.64)
with PMio: -0.80% (-1.7, 0.10)
withOs: -0.68% (-1.3, -0.05)
SO2 slightly attenuated with
NO2 adjustment.
PMio, 03 not associated with
PEF.
Correlations NR.
Adults with Asthma: studies with central site exposure assessment and no examination of copollutant confounding
tMaestrelli et al. (2011)
Padua, Italy
n = 32, mean age 39.6 (SD: 7.5) yr, 81% persistent asthma
Repeated measures. Supervised spirometry. 6 measures over
2 yr. 166 observations. Selected from database of beta-agonist
users (>6/yr for 3 yr), diagnosis clinically confirmed. 76%
follow-up participation. Dropouts did not differ from participants.
GEE adjusted for daily average temperature, atmospheric
pressure, humidity, asthma medication use, current smoking.
NO2-central site
24-h avg
Average of 2 city sites
% predicted FEVi:
1.1 (-6.6, 8.7)
No copollutant model.
CO associated with FEVi. No
association with personal or
central site PM2.5.
No association for central site
PMio, SO2, 03.
Correlations with NO2 NR.
5-41
-------
Table 5-5 (Continued): Epidemiologic studies of lung function in children and adults with asthma.
Study
Population Examined and Methodological Details
Laqorio et al. (2006)
Rome, Italy
n = 11, ages 18-64 yr, 100% mild, intermittent asthma
NO2 Metrics
Analyzed
NO2-central site
24-h avg
Average of 5 city sites
Lag Day
0
0-1 avg
Effect Estimate (95% Cl)
Single-Pollutant Model3
% predicted FEVi:
-4.1 (-6.7, -1.6)
-4.8 (-7.5, -2.1)
Copollutant Examination
No copollutant model.
CO at lag 0-2 avg associated
with FEVi. No association for
Repeated measures. Supervised spirometry. Examined 2/week
for two 1-mo periods. Mean 9, 15 observations/subject.
Participation rate not reported. Recruitment of nonsmokers
from outpatient clinic. GEE adjusted for season, temperature,
humidity, beta-agonist use.
PlVh.s, PM-io, PM-io-2.5, Os.
Low to moderate correlations
with NO2. Spearman r= 0.05
for CO, 0.17 for O3,0.43-0.51
for PM.
tCanova et al. (2010)
Padua, Italy
n = 19, ages 15-44 yr, 81% moderate/severe asthma
Repeated measures. Home PEF/FEV-i. Examined for five
30-day periods for 2 yr. Recruitment from prescription database
of subjects with mean >6 prescription/yr for 3 yr. 50% subjects
provided fewer than 33% maximum observations. GEE
adjusted for temperature, humidity, atmospheric pressure, ICS
use, smoking status.
NO2-central site
24-h avg
2 city sites
0, 1,2, 3,
0-1 avg,
0-3 avg
Results reported only in a
figure. NO2 shows null
associations with PEF and
FEVi.
No copollutant model.
CO associated with PEF.
Moderate correlation with NO2.
Spearman r = 0.48.
Parketal. (2005) NO2-central site
Incheon, South Korea 24-h avg
n = 64 with asthma, ages 16-75 yr, 31% with severe asthma 10 city sites
Repeated measures. Home PEF. Examined daily for 3-4 mo.
Recruited from medical center. GEE model, covariates NR.
PEF
0.45 (-1.1, 1.9)L/min
No copollutant model.
CO, PM-io, Os associated with
PEF.
No association for SO2.
tWiwatanadate and Liwsrisakun (2011)
Chiang Mai, Thailand
n = 121 with asthma and symptoms in previous 12 mo, ages
13-78 yr, 48% moderate/severe persistent asthma
Repeated measures. Home PEF. Examined daily for 10 mo.
Recruited from allergy clinic patients. GLM with random effect
for subject and adjusted for sex, age, asthma severity, day of
week, weight, pressure, temperature, sunshine duration, rain.
NO2-central site
24-h avg
1 city site
Evening PEF:
1.0(0.0,2.0)
Average PEF:
1.6(0.60,2.6)
Units of PEF not reported.
Only multipollutant models
analyzed. No associations with
PM-io, SO2, Os.
Interactions between NO2 and
copollutant or meteorological
variables reported not to be
statistically significant.
5-42
-------
Table 5-5 (Continued): Epidemiologic studies of lung function in children and adults with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics
Analyzed
Lag Day
Effect Estimate (95% Cl)
Single-Pollutant Model3
Copollutant Examination
Hiltermann et al. (1998)
Bilthoven, the Netherlands
n = 60 with asthma, ages 18-55 yr, all with airway
hyperresponsiveness, 87% with atopy
Repeated measures. Home PEF. Examined daily for 4 mo.
Recruitment from outpatient clinic. Model adjusted for allergen
concentrations, smoking exposure, day of week, temperature,
linear and quadratic term for study day
NO2-central site
24-h avg
1 city site
Site within 20 km of
subjects' homes. 3 city
sites highly correlated.
r=0.88.
Diurnal change PEF
0 -0.75 (-8.1, 6.6) L/min
0-6 avg -3.0 (-16, 10) L/min
No copollutant model.
BS at lag 0 associated with
PEF. No association with PM-io
orOs.
Note: More informative studies in terms of the exposure assessment method and potential confounding considered are presented first.
a.m. = ante meridiem; avg = average; BC = black carbon; BS = black smoke; AZ = Arizona; BTEX = benzene, toluene, ethylbenzene, xylene; CA = California; Cl = confidence
interval; CO = carbon monoxide, Colorado; DC = District of Columbia; EC = elemental carbon; ED = emergency department; FEF2s-75% = forced expiratory flow from 25% to 75% of
vital capacity; FEVi = forced expiratory volume in 1 second; FVC = forced vital capacity; GEE = generalized estimating equations; GLM = generalized linear model; ICAS = Inner City
Asthma Study; ICS = inhaled corticosteroid; IL = Illinois; MA = Massachusetts; Ml = Michigan; MO = Missouri; NCICAS = National Cooperative Inner-city Asthma Study;
NO2 = nitrogen dioxide; NR = not reported; NY = New York; O3 = ozone; OC = organic carbon; OH = Ohio; ON = Ontario; PA = Pennsylvania; PEF = peak expiratory flow;
PM2 5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PM10 = particulate matter with a nominal mean aerodynamic diameter less than or
equal to 10 |jm; QC = Quebec; SD = standard deviation; SES = socioeconomic status; SO2 = sulfur dioxide; TX = Texas; UFP = ultrafine particles, UK = United Kingdom;
VOC = volatile organic compound; WA = Washington; Wl = Wisconsin.
aEffect estimates were standardized to a 20-ppb increase in 24-h avg NO2, a 25-ppb increase in 8-h max NO2, and a 30-ppb increase 1-h max NO2. Effect estimates for other
averaging times (1-h avg to 12-h avg) are not standardized but presented as they are reported in their respective studies (Section 5.1.2.2).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-43
-------
With respect to the populations with supervised spirometry, most studies identified
children with asthma by parental report of physician-diagnosed asthma. Children were
recruited mostly from schools, supporting the likelihood that study populations were
representative of the general population of children with asthma. Based on a priori
hypotheses, results did not demonstrate larger NCh-associated decrements in lung
function in children with asthma than children without asthma (Barraza-Villarreal et al..
2008; Holguin et al.. 2007). Study populations represented a range of asthma severity, as
ascertained by Global Initiative for Asthma guidelines, ED visit or hospital admission for
asthma in the previous year, or medication use. Post hoc analyses pointed to stronger
associations among children with asthma not taking inhaled corticosteroid (ICS)
(Hernandez-Cadena et al., 2009; Liu et al., 2009b) or not taking controller
bronchodilators (Delfino et al.. 2008a). It is not clear what these results indicate about
risk varying by asthma severity, but bronchodilator use has been shown to reduce airway
responsiveness in response to a challenge agent (Section 5.2.2.1). Further, the larger
associations in ICS nonusers together with observations of NC^-associated lung function
decrements in populations with high prevalence of atopy (53%, 58%) (Martins et al..
2012; Holguin et al., 2007) are supported by findings for NCh-induced increases in
allergic inflammation (Section 5.2.2.5) and findings for mast cell degranulation (which
leads to histamine release) mediating NCh-induced lung function decrements
(Section 4.3.2.2).
For children with asthma, key evidence for NCh-associated decrements in lung function
measured under supervised conditions is provided by studies that assessed exposure in
subjects' locations [total personal NO2 (Smargiassi et al., 2014; Delfino et al., 2008a).
personal outdoor NCh estimated from measurements at school and other locations and
time-activity data (Martins et al., 2012) or outdoor school NO2 (Greenwald et al., 2013;
Spira-Cohen et al.. 2011; Holguin et al.. 2007)]. Among these studies, few reported
participation rates (66%, 87%; Table 5-5); however, none reported issues with selective
participation according to any subject characteristic or NC>2 exposure. These studies
examined limited lags of NO2 exposure but were similar in finding associations with
multiday averages (i.e., lag 0-1 avg, 0-4 avg) of 24-h avg NO2. Studies that measured or
estimated personal exposures provide evidence of an effect of outdoor NO2 exposure on
decreasing lung function. There is good indication that the time-weighted average of
microenvironmental NC>2 concentrations for children with wheeze in Portugal (Martins et
al.. 2012) well represented their personal outdoor exposure. Other studies show
agreement between microenvironmental and personal NC>2 concentrations
(Section 3.4.4.1). and Martins et al. (2012) found school and home indoor NCh
concentrations to be below the limit of detection.
5-44
-------
Studies of total personal NO2 produced contrasting results. Among children with asthma
in the Los Angeles, CA area, slightly larger decrements in percent predicted FEVi were
found with total personal NO2 (-1.5 [95% CI: -2.3, -0.57] per 20-ppb increase in
lag 0 day NO2) than central site NO2 (-1.3 [95% CI: -2.4, -0.15]) (Delfino et al.. 2008a).
A Spearman correlation of 0.43 between personal and central site NO2 indicated that
ambient NO2 had some influence on personal exposures. In contrast, total personal NO2
exposure was not associated with lung function among children with asthma living near
NO2 emissions sources (i.e., oil refineries, high traffic roads) in Montreal, Canada
(Smargiassi et al., 2014). Both studies of total personal exposure examined children for
10 consecutive days, and the total number of observations was higher in Smargiassi et al.
(2014) than in Delfino (2006) (-700 versus -500). In both studies, all or nearly all
personal NO2 measures were above the limit of detection (Table 5-5). However, it is
uncertain whether the Montreal study was sufficiently powered to detect associations
with NO2 exposures, which were far lower than in the Los Angeles, CA study (mean:
6.3 ppb versus 28.6 ppb) and showed low variability among children and days (IQR:
2.9 ppb versus 16.8 ppb). Also, the Montreal study did not provide evidence that lung
function was associated with total personal exposures to PM2 5, VOCs, or polycyclic
aromatic hydrocarbons (Smargiassi et al., 2014).
Among studies of outdoor school NO2, associations with FEVi were found in populations
in El Paso, TX, and Ciudad Juarez, Mexico, which are located along the U.S./Mexico
border [(Greenwald et al., 2013; Holguin et al., 2007); Figure 5-3 and Table 5-5].
Between two El Paso schools, associations were limited to the school located near a
major road and characterized by higher outdoor pollutant concentrations and a larger
percentage of Mexican-American children (Greenwald et al.. 2013). No association with
FEVi was found in children with asthma in Bronx, NY for school NO2 averaged over the
6-h school day (Spira-Cohen et al.. 2011). An effect of NO2 is supported by similar FEVi
decrements for outdoor and indoor NO2 (means for both - 17 ppb) in an El Paso school
(Greenwald et al.. 2013). Larger lung function decrements were observed in association
with home outdoor than indoor NO2 among children in five New Zealand towns
(Gillespie-Bennett et al.. 2011). but the results have weak implications as multiple daily
FEVi measures were related to a single 4-week average of NO2.
Compared with NO2 exposures estimated for subjects' locations, NO2 measured at central
sites is less clearly associated with lung function decrements. Among studies that
measured ambient NO2 at central sites, some found associations with lung function
decrements (Yamazaki et al.. 2011; Dales et al.. 2009a; Hernandez-Cadena et al.. 2009;
Liu et al.. 2009b; O'Connor et al., 2008). Many studies reported lack of association
(Wiwatanadate and Trakultivakorn. 2010; Barraza-Villarreal et al.. 2008; Odajima et al..
2008; Just et al.. 2002; Mortimer et al.. 2002) (Table 5-5). In this group were studies that
5-45
-------
did not report quantitative results, and it is not possible to assess the extent to which their
findings do or do not suggest associations. Across the studies examining central site NO2,
exposures were assigned as ambient measurements from a site located within 5 or 10 km
of subjects' homes or schools, measurements averaged among city monitors, or
measurements from one site. The central site NO2 assessment method did not appear to
influence results; however, in Mexico City, a low correlation (r = 0.21) between central
site and school NCh suggests that the central site may not have adequately represented the
variability within the study area (Barraza-Villarreal et al.. 2008).
Adjustment for potential confounding varied among studies but in most cases included
temperature. Several studies adjusted for (or considered in preliminary analyses) relative
humidity; a few adjusted for day of the week, smoking exposure, or asthma medication
use (Table 5-5). Lung function was associated with PMio, 862, and Os, which showed
negative or weakly positive correlations with NC>2 (e.g., -0.72 for PMio to 0.18 for 802).
Copollutant models were analyzed for SO2, and NCh associations with FEVi diurnal
change were unaltered, although both pollutants were measured at central sites within
10 km of homes (Dales et al.. 2009a).
Most studies found associations with PM2 5 and the traffic-related pollutants
black/elemental carbon (BC/EC) and VOCs. Copollutant models with PM2 5 produced
positive associations for 12-h avg and 24-h avg NCh with FEVi for children in Windsor,
Canada but unaltered associations with FEVi diurnal change [(Dales et al.. 2009a; Liuet
al.. 2009b): Table 5-5]. PM2 5 effect estimates adjusted for NO2 did not change. For this
study, high NO2-PM2 5 correlations (r = 0.71) and differential exposure error with central
site measurements limit inference from Copollutant models. Among studies with stronger
inference due to exposure assessment in subjects' locations, there is evidence of
confounding by the VOC benzene but evidence for NO2 associations that are independent
of PM2 5 or EC. For children with wheeze in Portugal, the association of modeled
personal outdoor NO2 with FEVi was attenuated (-3.7% [95% CI: -33, 25] per 20-ppb
increase in 1-week avg NCh) with adjustment for benzene (Spearman r = -0.42 to 0.14).
Outdoor school NO2 but not PM2 5 or EC was associated with FEVi in children with
asthma in Ciudad Juarez, Mexico (Holguin et al., 2007). In a detailed analysis of total
personal and central site measures, Delfino et al. (2008a) found the association of
personal NO2 with FEVi to be robust (-1.3-point [95% CI: -2.8, 0.22] change in percent
predicted FEVi per 20-ppb increase in NCh) to adjustment for personal PM2 5, which was
weakly correlated with personal NC>2 (Spearman r = 0.38). Adjustment for personal PM2 5
(r = 0.32) reduced the association of central site NO2 with FEVi (-0.86-point [95% CI:
-2.6, 0.89] change per 20-ppb increase in NCh). The attenuation could indicate that
ambient NCh is an indicator of personal PM25 but also could indicate less exposure
measurement error for personal PM2 5 than central site NCh. Nonetheless, the moderate
5-46
-------
personal-ambient NO2 correlation (r = 0.43) and the copollutant model results for
personal NO2 provide evidence for effects of ambient NO2 on FEVi that are independent
of PM25 exposure.
Adults with Asthma
Most previous and recent studies of lung function in adults with asthma were based on
PEF measured at home. Except for the recent multicity U.S. study (Qian et al.. 2009b).
studies indicated no association or inconsistent associations among the various lung
function parameters or NC>2 exposure lags examined (Wiwatanadate and Liwsrisakun.
2011: Canovaetal.. 2010: Park etal.. 2005: Hiltermann et al.. 1998V Subjects generally
were recruited from outpatient clinics or doctors' offices, and the nonrandom selection of
the general population may produce study populations less representative of the asthma
population. Most studies examined 24-h NO2 that was assessed primarily from central
site measurements, and results are equally inconsistent for NC>2 exposures assigned from
one site or averaged from multiple city sites.
Although the few results for supervised spirometry are inconsistent overall [(Maestrelli et
al.. 2011: McCreanor et al.. 2007: Lagorio et al.. 2006): Table 5-5]. the strongest study
with personal outdoor pollutant measurements shows associations for NC>2 that are
independent of PM25 or the traffic-related copollutants EC or UFP. Across studies, lung
function was associated with EC, UFP, PM2 5, and CO, which were moderately correlated
with NO2 (Spearman r = 0.43-0.60) (McCreanor et al.. 2007: Lagorio etal.. 2006).
Potential confounding by CO was not examined. Among adults in London, U.K. with
mild to moderate asthma, NO2 measured next to subjects while walking outdoors (next to
a high-traffic road allowing only diesel buses and taxis and in a park) was associated with
decrements in FEVi and forced expiratory flow from 25 to 75% of vital capacity
(FEF25-75%) (McCreanor et al.. 2007). NO2-related decrements occurred 2 to 22 hours
after exposure. A 5.3-ppb increase in 2-h avg NO2 was associated with a -0.22% (95%
CI: -0.40, -0.05) change in FEVi and -0.78% (95% CI: -1.4, -0.13) change in
FEF25-75%. NO2-associated decrements in FEVi were attenuated to near null with
adjustment for UFP (McCreanor et al.. 2007). Associations with FEF25-75% decreased in
magnitude and precision with adjustment for UFP, EC, or PM2 5 but remained negative
(e.g., -0.47% [95% CI: -1.3, 0.39] per 5.3-ppb increase in 2-h avg NO2 with adjustment
for UFP, Spearman r = 0.58). Effect estimates for UFP, EC, and PM2 5 were unaltered or
less attenuated with adjustment for NO2. Thus, results indicate NO2 associations that are
weaker but independent of those for UFP, EC, and PM2 5. PMio and SO2 also were
associated with lung function in adults with asthma and were more weakly correlated
with NO2. In the U.S. multicity study, NO2-PEF associations were attenuated with
adjustment for SO2 (Qian et al.. 2009b). However, with exposures assessed from
5-47
-------
measurements averaged across central site monitors within 32 km of subjects' ZIP codes,
it is not clear how well subjects' actual exposures to NO2 or 862 were represented.
Controlled Human Exposure Studies
Most controlled human exposure studies of lung function examined adults and were
reviewed in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). Consistent with
epidemiologic findings, most controlled human exposure studies did not report effects on
lung function in adults with asthma. Exposures ranged from 200 to 4,000 ppb NC>2 for
30 minutes to 6 hours, and most studies incorporated exercise in the exposure period to
assess lung function during various physiological conditions (Table 5-6).
Several studies examined NO2 exposures in the range of 120 to 400 ppb, and most did not
observe concurrent or subsequent changes in lung function or airway resistance in
adolescents with asthma (Koenig etal.. 1987) or adults with asthma (Riedl etal.. 2012;
Jenkins et al.. 1999; Vagaggini etal.. 1996; Torres and Magnussen. 1991; Kleinman et al..
1983). Only Bauer et al. (1986). who exposed adults with asthma to 300 ppb NO2 for
30 minutes, reported statistically significant decrements in forced expiratory flow rates. A
few studies examined both airway responsiveness and lung function (Jenkins etal.. 1999;
Bauer et al.. 1986; Kleinman et al.. 1983). and only Bauer etal. (1986) observed
NO2-induced changes in both lung function and airway responsiveness (Section 5.2.2.1).
Evidence for NO2-induced changes in lung function in adults with asthma are equally
weak for higher NO2 exposures, with 1,000 ppb NO2 exposure resulting in small
reductions in FEVi (Torres et al.. 1995) but 4,000 ppb NO2 producing no changes in
airway resistance (Linn etal.. 1985b).
There is no strong evidence for interactions between NO2 and Os in controlled human
exposure studies. Jenkins etal. (1999) found no change in lung function in adults with
asthma following exposure to 200 ppb NO2 for 6 hours (with or without 200 ppb Os) or
400 ppb NO2 for 3 hours (without 400 ppb Os). Statistically significant decreases in FEVi
were found following the 3-hour exposure to Os and Os + NO2. Thus, results indicated an
effect of Os exposure alone.
5-48
-------
Table 5-6 Characteristics of controlled human exposure studies of lung
function in individuals with asthma.
Study
Disease Status;
Sample Size; Sex; Age
(mean ± SD)
Exposure Details
Endpoints Examined
Koeniq et al.
(1987)
Asthma;
(1)n=4M,6F
(2)n = 7M, 3F;
Healthy;
(1)n = 3M, 7F
(2) n = 4 M, 6 F;
14.4 yr (range: 12-19)
(1)120ppbNO2,
(2)180ppbNO2;
(1-2) Exposures were 30 min at rest with
10 min of exercise at VE = 32.5 L/min
Pulmonary function tests
before, during, and after
exposure.
Symptoms immediately
after and 1 day after.
Jenkins et al.
(1999)
Asthma;
n = 9 M, 2 F;
31.2±6.6yr
(1)200ppbNO2for6h
(2) 200 ppb NO2 + 100 ppb O3 for 6 h
(3)400ppbNO2for3h
(4) 400 ppb NO2 + 200 ppb O3 for 3 h
(1-4) Exercise 10 min on/40 min off at
VE = 32 L/min)
Pulmonary function tests
before and after
exposure.
Kleinman et al. Asthma;
(1983) n = 12M, 19 F;
31 ± 11 yr
200 ppb for 2 h;
Exercise 15 min on/15 min off atVE = ~2
times resting
Pulmonary function
testing before and after
exposure.
Symptoms before,
immediately after, and
day after exposure.
Jorres and
Maqnussen
(1991)
Bauer et al.
(1986)
Vaqaqqini et al.
(1996)
tRiedletal.
(2012)
Asthma; n = 9 M, 2 F;
29 yr (range: 17-55)
Asthma;
n = 15; 33 ± 7.8 yr
Asthma; n = 4 M, 4 F;
29 ± 14 yr
Healthy; n = 7 M;
34 ± 5 yr
Asthma
Phase 1: methacholine
challenge;
n = 10M, 5F;
37.3 ± 10.9 yr
Phase 2: cat allergen
challenge;
n=6M, 9F;
36.1 ± 12.5 yr
250 ppb for 1 h;
Rest for 20 min followed by 10 min of
exercise (VE = 30 L/min)
300 ppb for 30 min (20 min at rest, 10 min of
exercise at VE > 3 times resting)
300 ppb for 1 h;
Exercise at VE = 25 L/min
350 ppb for 2 h;
Exercise 15 min on/15 min off at
VE = 15-20 L/min
Airway resistance
measured before, during,
and after exposure.
Pulmonary function
before, during, and after
exposure.
Symptoms before and
2 h after exposure.
Cell counts in sputum
2-h post-exposure.
Symptoms before,
during, 1-22 h after
exposure.
5-49
-------
Table 5-6 (Continued): Characteristics of controlled human exposure studies of
lung function in individuals with asthma.
Study
Disease Status;
Sample Size; Sex; Age
(mean ± SD)
Exposure Details
Endpoints Examined
Jorres et al.
(1995)
Asthma; n = 8 M, 4 F;
27 ± 5 yr
Healthy; n = 5 M, 3 F;
27 yr (range: 21-33)
1,000 ppbforS h;
Exercise 10 min on/10 min off at individual's
maximum workload
Pulmonary function tests
before, during, and after
exposure.
Symptoms immediately,
6 h, and 24 h after
exposure.
BAL fluid analysis 1 h
after exposure (cell
counts, histamine,
prostaglandins).
Linn et al. Asthma; n = 12 M, 11 F; 4,000 ppb for 75 min;
£1985b] range: 18-34 yr Two 15_min periods of exercise at yE = 25
Healthy; n = 16 M, 9 F; L/min and 50 L/min
range: 20-36 yr
Airway resistance before,
during, and after
exposure.
Symptoms before,
during, immediately after,
1 day after and 1 week
after exposure.
BAL = bronchoalveolar lavage; F = female, M = male; NO2 = nitrogen dioxide; O3 = ozone; SD = standard deviation; VE = minute
ventilation.
fStudy published since the 2008 ISA for Oxides of Nitrogen
5.2.2.3 Respiratory Symptoms and Asthma Medication Use in Populations with
Asthma
The preceding epidemiologic evidence describing associations between short-term
increases in ambient NC>2 concentrations and decreases in lung function in children with
asthma, particularly those with atopy, supports evidence for NCh-related increases in
respiratory symptoms in children with asthma. Decreased lung function can indicate
airway obstruction (Section 4.3.2). which can cause symptoms. Also providing a
biologically plausible link between NCh exposure and respiratory symptoms is evidence
for NC>2-induced increases in airway responsiveness (Section 5.2.2.1) and pulmonary
inflammation (Section 5.2.2.5 and Figure 4-1). Epidemiologic studies reviewed in the
2008 ISA for Oxides of Nitrogen consistently found increased respiratory symptoms in
children with asthma in association with increases in indoor, personal, and ambient NCh
concentrations (U.S. EPA. 2008c). These epidemiologic findings largely are unsupported
by controlled human exposure studies of symptoms in adolescents or adults with asthma.
Recent studies, most of which were epidemiologic, continue to indicate associations
5-50
-------
between short-term increases in ambient NC>2 concentration and increases in respiratory
symptoms in children with asthma.
Epidemiologic Studies
Epidemiologic studies examined respiratory symptoms in relation to ambient NCh
concentrations rather than NO or NOx, and evidence is stronger for children with asthma
than adults with asthma. Across the various populations examined, symptom data were
collected by having subjects or their parents complete daily diaries for periods of 2 weeks
to several months. Heterogeneity in the number of consecutive days of follow-up and the
frequency of diary collection from study subjects do not appear to influence results.
Ambient NCh concentrations, locations, and time periods for epidemiologic studies of
respiratory symptoms are presented in Table 5-7.
Table 5-7 Mean and upper percentile concentrations of nitrogen dioxide in
epidemiologic studies of respiratory symptoms in populations with
asthma.
Study3
Schildcrout et al.
(2006)
Romieu et al.
(2006)
Seqalaetal. (1998)
tPateletal. (2010)
fBarraza-Villarreal
et al. (2008),
fEscamilla-Nunez
et al. (2008)
Location
Albuquerque, NM
Baltimore, MD
Boston, MA
Denver, CO
San Diego, CA
St. Louis, MO
Toronto, ON,
Canada
Mexico City,
Mexico
Paris, France
New York City and
nearby suburb, NY
Mexico City,
Mexico
NO2 Metric
Study Period Analyzed
Nov1993- 24-h avg NO2
Sep 1995
Oct 1998-Apr 1-h max NO2
2000
Nov 1992- 24-h avg NO2
May 1993
2003-2005, 24-h avg NO2
months NR
Jun 2003- 8-h max NO2
Jun 2005
Mean/Median
Concentration
ppb
Across cities:
17.8-26.0
66
30.3b
NR
37.4
Upper Percentile
Concentrations
ppb
90th: across cities
26.7-36.9 ppb
Max: 298
Max: 64.9b
NR
Max: 77.6
5-51
-------
Table 5-7 (Continued): Mean and upper percentile concentrations of nitrogen
dioxide in epidemiologic studies of respiratory symptoms
in populations with asthma.
Study3
tMann et al. (2010)
tZoraetal. (2013)
Jalaludin et al.
(2004)
tSpira-Cohen et al.
(2011)
fSarnat et al.
(2012)
tHolquin et al.
(2007)
tGillespie-Bennett
etal. (2011)
tGent et al. (2009)
Delfinoetal. (2003)
Delfino et al. (2002)
fO'Connor et al.
(2008)
Location
Fresno/Clovis, CA
El Paso, TX
Western and
Southwestern
Sydney, Australia
Bronx, NY
El Paso, TX and
Ciudad Suarez,
Mexico
Ciudad Juarez,
Mexico
Bluff, Dunedin,
Christchurch,
Porirua, Hutt
Valley, New
Zealand
New Haven
County, CT
Los Angeles, CA
(Huntington Park
area)
Alpine, CA
Boston, MA
Bronx, NY
Chicago, IL
Dallas, TX
New York, NY
Seattle, WA
Tucson, AZ
Study Period
Winter-
Summer,
2000-2005
Mar-Jun 2010
Feb-Dec 1994
Spring 2002,
spring/fall
2004, spring
2005
Jan-Mar 2008
2001-2002
Sep 2006
Aug 2000-
Feb 2004
Nov 1999-Jan
2000
Mar-Apr 1996
Aug 1998-
Jul2001
NO2 Metric
Analyzed
24-h avg NO2
96-h avg NO2
15-h avg NO2
(6 a.m.-
9 p.m.)
6-h avg NO2
(9 a.m.-
3 p.m.)
96-h avg NO2
1-week avg
NO2
4-week avg
NO2
NO2 — avg
time NR
1-h max NO2
8-h max NO2
1-h max NO2
8-h max NO2
24-h avg NO2
Mean/Median
Concentration
ppb
Median: 18.6
School 1: 9.3
School 2: 3.4
15.0
NR
El Paso schools:
4.5, 14.2; central
sites: 14.0, 18.5,
20.5
Ciudad Juarez
schools: 18.7, 27.2;
central site: none
18.2
3.9
NR
7.2
5.9
24
15
NR
Upper Percentile
Concentrations
ppb
75th: 24.7
Max: 52.4
Max: 16.2
Max: 7.5
Max: 47.0
NR
NR
NR
NR
NR
90th: 9.0; max: 14
90th: 7.9; max: 11
Max: 53
Max: 34
NR
5-52
-------
Table 5-7 (Continued): Mean and upper percentile concentrations of nitrogen
dioxide in epidemiologic studies of respiratory symptoms
in populations with asthma.
Study3
Mortimer et al.
(2002)
Just et al. (2002)
tOstroetal. (2001)
Boezen et al.
(1998)
Forsberq et al.
(1998)
von Klot et al.
(2002)
fMaestrelli et al.
(2011)
tWiwatanadate and
Liwsrisakun (2011)
Hiltermann et al.
(1998)
fLaurent et al.
(2009)
fCarlsen et al.
(2012)
Location
Bronx & East
Harlem, NY
Chicago, IL
Cleveland, OH
Detroit, Ml
St. Louis, MO
Washington, DC
Paris, France
Los Angeles and
Pasadena, CA
Amsterdam,
Meppel,
the Netherlands
Landskrona,
Sweden
Erfurt, Germany
Padua, Italy
Chiang Mai,
Thailand
Bilthoven, the
Netherlands
Strasbourg, France
Reykjavik, Iceland
Study Period
Jun-Aug 1993
Apr-Jun 1996
Aug-Oct 1993
Winter 1993-
1994
Jan-Mar, yr
NR
Sep 1996-
Nov1997
1999-2003
Aug 2005-
Jun 2006
Jul-Oct1995
2004, all yr
Mar 2006-
Dec 2009
NO2 Metric
Analyzed
4-h avg NO2
(6 a.m.-
10 a.m.)
24-h avg NO2
1-h max NO2
24-h avg NO2
24-h avg NO2
24-h avg NO2
24-h avg NO2
24-h avg NO2
24-h avg NO2
24-h avg NO2
Dispersion
model
24-h avg NO2
1-h max
Mean/Median
Concentration
ppb
NR
28.6b
Los Angeles: 79.5
Pasadena: 68.1
24.5b
14.2b
16.2b
24.5b
Across seasons and
yr: 20. 9-37. Ob
17.2
11. 2b
18.6b
11. 7b
27.4b
Upper Percentile
Concentrations
ppb
NR
Max: 59. Ob
Max: 220
Max: 170
Max: 40.4b
Max: 28.9b
Max: 38. 1b
Max: 63.3b
75th: 23.0-42.5b
90th: 26.5
Max: 37.4
Max: 22.5b
NR
97.5th: 30.4b
97.5th: 61. Ob
5-53
-------
Table 5-7 (Continued): Mean and upper percentile concentrations of nitrogen
dioxide in epidemiologic studies of respiratory symptoms
in populations with asthma.
Study3
tKimetal. (2012)
tKarakatsani et al.
(2012)
fFeo Brito et al.
(2007)
NO2 Metric
Location Study Period Analyzed
Seoul and Kyunggi 2005-2009 24-h avg NO2
Province, South
Korea
Amsterdam, Oct 2002- 24-h avg NO2
the Netherlands Mar 2004
Athens, Greece
Birmingham, U.K.
Helsinki, Finland
Ciudad Real May-Jun 24-h avg NO2
Puertollano, Spain 2000-2001
Mean/Median
Concentration
ppb
Asthma
exacerbation
34.3 spring,
26.6 summer,
30.6 fall,
38.8 winter
No asthma
exacerbation:
32.7 spring,
26.0 summer, 30.6
fall, 37.7 winter
20.4b
21. 2b
18.3b
12.1b
17.4b
29.5b
Upper Percentile
Concentrations
ppb
75th: asthma
exacerbation:
41.3 spring,
35.3 summer,
42.0 fall,
46.8 winter
No asthma
exacerbation:
41.4 spring,
35.2 summer,
41. 6 fall,
48.9 winter
Max: 51. 8b
Max: 59.0b
Max: 44.2b
Max: 41. 4b
Max: 35.6b
Max: 100.5b
a.m. = ante meridiem; Aug = August; avg = average; AZ = Arizona; CA = California; CO = Colorado; CT = Connecticut;
DC = District of Columbia; Dec = December; Feb = February; IL = Illinois; MD = Maryland; MA = Massachusetts; Ml = Michigan;
MO = Missouri; NM = New Mexico; NO2 = nitrogen dioxide; NR = not reported; NY = New York; OH = Ohio; ON = Ontario;
ppb = parts per billion; TX = Texas; UK = United Kingdom; WA = Washington.
aStudies presented in order of first appearance in the text of this section.
bConcentrations converted from |jg/m3 to ppb using the conversion factor of 0.532 assuming standard temperature (25°C) and
pressure (1 atm).
fStudiies published since the 2008 ISA for Oxides of Nitrogen.
Children with Asthma
Several recent studies add to the evidence for increases in respiratory symptoms in
children with asthma associated with short-term increases in ambient NC>2. Across
previous and recent studies, there is heterogeneity in the magnitude and precision (width
of 95% CIs) of the association. However, the results collectively indicate a pattern of
elevated risk of respiratory symptoms across the various symptoms and lags of NC>2
exposure examined (Figure 5-4 and Table 5-9). The consistency of findings also is
supported by a meta-analysis of 24 mostly European studies and some U.S. studies,
including several reviewed in the 2008 ISA for Oxides of Nitrogen. In the meta-analysis,
5-54
-------
there was some evidence of publication bias with exclusion of the multicounty European
PEACE studies, but with adjustment for publication bias, an increase in 24-h avg NO2
was associated with increased risk of asthma symptoms (Weinmavr et al.. 2010). Across
individual studies reviewed in this ISA, the most consistent results were for total
respiratory or asthma symptoms, wheeze, and cough. Increases in ambient NC>2
concentrations were not consistently associated with increases in rescue inhaler or
beta-agonist use in children with asthma (Patel et al.. 2010; Gent et al.. 2009; Romieu et
al.. 2006; Schildcrout et al.. 2006; Segalaetal.. 1998).
Study populations were recruited from schools, asthma or allergy clinics, or doctors'
offices. Asthma was identified by parental report of physician-diagnosed asthma or
clinical examination. Neither of these methodological issues appeared to affect whether
an association was found. In studies that reported data on follow-up participation, rates
were 7-92%, and no study reported selective dropout among a particular group within the
study population. A priori-determined comparisons of children with and without asthma
were inconsistent. Patel et al. (2010) found stronger associations in children with asthma.
Another study found stronger associations in children without asthma, but 72% of that
group had atopy (Barraza-Villarreal et al.. 2008; Escamilla-Nunez et al.. 2008).
Atopy could influence NCh-related respiratory symptoms. Many asthma study
populations had high prevalence of atopy (47-89%), and larger NCh-associated increases
in symptoms were found in children with asthma who also had allergies (Zoraetal..
2013; Mann et al.. 2010). These results were based on 16 to 47% of the study populations
but are coherent with experimental evidence for NCh-induced allergic responses in adults
with asthma and animal models of allergic disease (Section 5.2.2.5). Study populations
also varied in asthma severity; some comprised mostly children with mild, intermittent
asthma and others comprised children with persistent asthma. Comparisons by asthma
severity indicated larger NCh-related increases in respiratory symptoms among children
with mild, intermittent asthma than severe or moderate asthma (Mann etal.. 2010; Segala
et al.. 1998). but these results also were based on small numbers of subjects. Jalaludin et
al. (2004) found that elevated risk was limited to children with more severe asthma
(asthma plus airway hyperresponsiveness). But, results were based on a three-pollutant
model, which can produce unreliable results because of potential multicollinearity.
5-55
-------
Study
Symptom Composite
Spira-Cohen et al. (2011)
Delfino et al. (2003)
Delfino et al. (2002)
Schildcroutetal. (2006)
Mortimer etal. (2002)
Just et al. (2002)
Segalaetal. (1998)
O'Connor et al. (2008)
Wheeze
Spira-Cohen et al. (2011)
Gent et al. (2009)
Mann etal. (2010)
Jalaludin et al. (2004)
Escamilla-Nunezetal. (2008)
Barraza-Villarreal et al. (2008)
Ostroetal. (2001)
Pateletal. (2010)
Asthma Medication Use
Schildcroutetal. (2006)
Jalaludin et al. (2004)
Romieu et al. (2006)
Segalaetal. (1998)
NO2 Metrics Analyzed
6-h avg, lag 0 day
8-h max, lag 0 day
1-h max, lag 0 day
8-h max, lag 0 day
24-h avg, lag 0 day
24-h avg, 0-2 day sum
4-h avg, lag 1 -6 day avg
24-h avg, lag 0 day
24-h avg, lag 0 day
24-h avg, lag 3 day
Exposure
Assessment
School
Central site
Central site
Central sites
Central sites
Central sites
Central sites
24-h avg, lag 1-19 day avg Central sites
6-h avg, lag 0 day School
Averaging time NR, lag 0 day Central site
24-h avg, lag 2 day Central site
15-h avg, lag 0 day
1-h max, lag 1 day
1-h max, lag 0 day
1-h max, lag 3 day
24-h avg, lag 0 day
24-h avg, lag 0 day
15-h avg, lag 0 day
1 -h max, lag 1 -6 day avg
24-h avg, lag 0 day
Central site
Central sites
Central sites
Central sites
Central sites
Central sites
Central site
Central site
Central sites
Subgroup 1
1.
1 .
1 .*
1 .
U-
1
1 ft
1 •
All subjects | — •
k
c
1
T
GSTM1 null -»f
GSTM1 positive .-•-
I
0.0 0.5 1.0 1.5 2.0 2.5
Odds ratio per increase in NO2 (95% Cl)a
3.0
Note: avg = average; Cl = confidence interval; GSTM1 = glutathione S-transferase mu 1; h = hour; max = maximum; NO2 = nitrogen
dioxide; NR = not reported. Black = studies from the 2008 Integrated Science Assessment for Oxides of Nitrogen, red = recent
studies. Results from more informative studies in terms of the exposure assessment method and potential confounding considered
are presented first within an outcome group. Study details and quantitative results are reported in Table 5-8. The figure presents a
subset of results included in Table 5-8 for which quantitative results were available for NO2 examined as a linear variable and for
specific outcomes examined by multiple studies.
aEffect estimates are standardized to a 20-ppb, 25-ppb, and 30-ppb increase for 24-h avg, 8-h max, and 1-h max NO2, respectively.
Effect estimates for other averaging times (4-h avg to 15-h avg) are not standardized and presented as reported in their respective
studies (Section 5.1.2.2).
Figure 5-4 Associations of ambient nitrogen dioxide concentrations with
respiratory symptoms and asthma medication use in children
with asthma.
5-56
-------
Table 5-8 Epidemiologic studies of respiratory symptoms and asthma medication use in children with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics Odds Ratio (95% Cl)
Analyzed Lag Day Single-Pollutant Modela
Copollutant Examination
Studies with small spatial scale exposure assessment and/or examination of copollutant confounding
tZoraetal. (2013)
El Paso, TX
n = 36, mean age 9.3 (SD: 1.5) yr, 47% with atopy
Repeated measures. Asthma control questionnaire given weekly at
school for 13 weeks. Questionnaire ascertains symptoms, activity
limitations, asthma medication use. Recruitment from schools via school
nurses. Parent report of physician-diagnosed asthma. No information on
participation rate. Linear mixed effects model adjusted for random subject
effect and humidity, temperature, school.
NO2-school
outdoor
24-h avg
1 school 91 m
from major road,
1 school in
residential area.
0-4 avg Change in asthma control
score (higher score
indicates poorer control):
Allergy, n = 17
0.56 (-0.10, 1.2)
No allergy, n = 19
-0.29 (-1.1, 0.49)
No copollutant models
analyzed for subgroups.
BC, benzene, toluene, also
associated with poorer
asthma control.
• Correlations with NO2 weak
to high. Spearman r= 0.29
to 0.56 for BC, 0.37 to 0.71
for benzene, 0.16 to 0.71
for toluene.
tSarnatetal. (2012)
El Paso, TX and Ciudad Juarez, Mexico
n = 29 per city, ages, 6-12 yr, asthma and current symptoms
Repeated measures. Daily symptom diaries. Recruitment from schools
representing a gradient of traffic, subjects from nonsmoking homes. No
information on participation rate. Self-report of physician-diagnosed
asthma. GLM with subject as random effect and adjustment for school,
temperature, relative humidity. Adjustment for medication use, cold
symptoms did not alter results.
NO2-school
outdoor
24-h avg
In each city,
1 school 91 m
from major road,
1 in residential
area.
0-4 avg No quantitative results
reported; associations
reported to be consistent
with the null.
No copollutant model.
5-57
-------
Table 5-8 (Continued): Epidemiologic studies of respiratory symptoms and asthma medication use in children
with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics Odds Ratio (95% Cl)
Analyzed Lag Day Single-Pollutant Modela
Copollutant Examination
tHolquin et al. (2007)
Ciudad Juarez, Mexico
n = 95, ages 6-12 yr, 78% mild, intermittent asthma, 58% with atopy
Repeated measures. Daily symptom diaries given by parents for 4 mo,
checked biweekly. 87% participation. Parent report of physician-
diagnosed asthma. Linear and nonlinear mixed effects model with
random effect for subject and school adjusted for sex, body mass index,
day of week, season, maternal and paternal education, passive smoking
exposure.
NO2-school
outdoor
24-h avg
Schools located
239-692 m from
homes.
0-6 avg No quantitative results
reported. No air pollutants
reported to be associated
with respiratory symptoms
No copollutant model.
Road density at home and
school reported not to be
associated with respiratory
symptoms.
tSpira-Cohen et al. (2011): Spira-Cohen (2013)
Bronx, NY
n = 40, ages 10-12 yr, 100% nonwhite, 44% with asthma ED visit or
hospital admission in previous 12 mo
Repeated measures. Daily symptom diaries for 1 mo, checked daily.
454 observations. Recruitment from schools by referrals from school
nurses. Parental report of physician-diagnosed asthma. 89% time
indoors. No information on participation rate. Mixed effects model with
subject as random effect adjusted for temperature, height, sex.
Adjustment for school (indicator of season) did not alter results.
NO2-school
outdoor
6-h avg
(9 a.m.-3 p.m.)
Schools 53-737
m from
highways with
varying traffic
counts. Most
children walk to
school.
Total symptoms:
1.1 (0.84, 1.5)
Wheeze:
1.2(0.75, 1.9)
OR per60-ppb increase
NO2 (5th to 95th percentile)
Personal EC associated
with symptoms with NO2
adjustment.
No quantitative data
reported.
tGillespie-Bennett et al. (2011)
Bluff, Dunedin, Christchurch, Porirua, Hutt Valley, New Zealand
n = 358, ages 6-1 Syr
Cross-sectional. Daily symptom diaries for 112 days. Recruitment from a
home heating intervention. 77% participation. Mixed effects model with
log-transformed NO2 and random effect for subject. Adjustment for age,
sex, region, ethnicity, intervention, income, temperature did not alter
results.
NCb-home
outdoor
24-h avg
1 measure per
subject.
NO2-home
indoor
24-h avg
Up to 4
measures per
subject.
4-week Lower respiratory
avg symptom:
1.1 (0.78, 1.5)
Reliever inhaler:
1.5(0.96,2.3)
OR per log increase NO2
Lower respiratory
symptom:
1.14(1.12, 1.16)
Reliever inhaler:
1.14(1.11, 1.17)
OR per log increase NO2
No copollutant model.
No other pollutants
examined.
5-58
-------
Table 5-8 (Continued): Epidemiologic studies of respiratory symptoms and asthma medication use in children
with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics Odds Ratio (95% Cl)
Analyzed Lag Day Single-Pollutant Modela
Copollutant Examination
tGent et al. (2009)
New Haven county, CT
n = 149, ages 4-12 yr
Repeated measures. Daily symptom diaries reported monthly.
Recruitment from larger cohort, pediatric asthma clinic, and school.
Parent report of physician-diagnosed asthma. No information on
participation rate. GEE adjusted for season, day of week, date, motor
vehicle factor obtained by source apportionment.
NO2-central site 0
Avg time not
reported
1 site 0.9-30 km
of homes (mean
10.2km).
NR
With source apportionment
factor of EC, zinc, lead,
copper, selenium
Wheeze: 1.1 (0.99, 1.2)
Inhaler: 1.0(0.97, 1.1)
Factor results not altered
by NO2 adjustment.
Moderate correlation with
NO2. Pearson r= 0.49.
Delfinoetal. (2003)
Los Angeles, CA (Huntington Park)
n = 16, ages 10-16 yr, 100% Hispanic, 27% on anti-inflammatory
medication
Repeated measures. Daily symptom diaries for 3 months, collected
weekly. Recruitment from schools of nonsmoking children from
nonsmoking homes. Self or parental report of physician diagnosed
asthma. 92% follow-up participation. GEE with autoregressive parameter
and adjusted for respiratory infections. Excluded potential confounding by
weekend, maximum temperature.
NO2-central site
4.8 km of home
& school
8-h max
1-h max
8-h max
1-h max
Asthma symptoms not
interfering with daily activity
1.3(1.1, 1.5)
1.2(0.96, 1.4)
Asthma symptoms
interfering with daily activity
1.4(1.02, 1.9)
1.3(0.81,2.0)
ORs per 1.4 ppb increase
in 8-h max and 2.0-ppb
increase in 1-h max NO2
(interquartile ranges).
Copollutant model results
only in figure. ORs for NO2
not altered by xylene or
toluene adjustment.
Smaller but positive ORs
for NO2, wider 95% Cl with
adjustment for benzene,
ethylbenzene,
acetylaldehyde,
formaldehyde. Moderate to
high correlations with 8-h
max NO2. Spearman
r= 0.57 (benzene) to
0.72 (xylene). No
interactions between NO2
and VOCs. ORs for VOCs
attenuated with NO2
adjustment.
No copollutant model with
ECorOC. r= 0.54 & 0.62,
respectively. No
association with CO.
5-59
-------
Table 5-8 (Continued): Epidemiologic studies of respiratory symptoms and asthma medication use in children
with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics Odds Ratio (95% Cl)
Analyzed Lag Day Single-Pollutant Modela
Copollutant Examination
Delfino et al. (2002)
Alpine, CA and adjacent areas
n = 22, ages 9-19 yr, 36% with mild persistent or more severe asthma,
77% with atopy
Repeated measures. 92% follow-up. Daily symptom diaries for 61 days,
collected weekly or biweekly. 1,248 observations (94% of expected).
Recruitment from schools. Asthma diagnosis based on referrals from
health maintenance organization and newspaper advertisements.
Subjects were nonsmokers from nonsmoking homes. GEE with
autoregressive lag 1 correlation matrix with no covariates. Adjustment for
day of week, linear trend, temperature, humidity did not alter results.
Adjustment for respiratory infection increased pollutant ORs.
NO2-central site 0
1-h max
8-h max
1 site 1-4.7 km
from subjects'
homes
Symptoms interfering with
daily activity
All subjects:
1.4(0.82,2.2)
No anti-inflammatory
medication, n = 12
1.8(0.89, 3.6)
On anti-inflammatory
medication, n = 10
0.91 (0.21,4.0)
All subjects:
1.7(0.94,2.9)
No anti-inflammatory
medication, n = 12
2.3(1.1,4.6)
On anti-inflammatory
medication, n = 10
1.1 (0.22, 5.5)
Positive interaction for 8-h
max NO2 with 1-h max
PMio(p< 0.01) and 1-h
maxOs(p = 0.12).
Fungi and pollen allergen
associated with symptoms.
No NO2-allergen
interactions. No
quantitative results for NO2-
allergen copollutant
. models, but ORs reported
to decrease.
Moderate correlations with
NO2. Pearson r= 0.29 for
fungi, 0.27 for pollen, 0.55
for PMm
Schildcrout et al. (2006)
Albuquerque, NM; Baltimore, MD; Boston, MA; Denver, CO; San Diego,
CA; St. Louis, MO; Toronto, ON, Canada (CAMP cohort)
n = 990, ages, 5-12 yr, mild to moderate asthma
Repeated measures. Daily symptom diaries for 21-201 days. No
information on participation rate. GEE for individual cities combined for
study-wide estimates. City-specific models adjusted for day of week,
ethnicity, annual family income, response to methacholine, maximum
temperature, humidity, temperature x humidity, calendar date. Pollutant
analyzed as daily deviation from subject mean.
NO2-central site
24-h avg
Average of
multiple sites
within 80 km of
ZIP code.
0 Asthma symptoms:
0-2 sum 1.06(1.00,1.12)
Rescue Inhaler use:
1.04(1.00, 1.08)
Asthma symptoms:
1.05(1.01, 1.09)
Joint effect models
NO2+CO: 1.07(1.0, 1.14)
NO2+SO2: 1.06(0.98, 1.15)
NO2+PM-IO:
1.06(0.99, 1.13)
Moderate to high
correlations with NO2.
r= 0.23 to 0.58 for SO2,
0.26 to 0.64 for PMio, 0.63
to 0.92 for CO.
5-60
-------
Table 5-8 (Continued): Epidemiologic studies of respiratory symptoms and asthma medication use in children
with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics Odds Ratio (95% Cl)
Analyzed Lag Day Single-Pollutant Modela
Copollutant Examination
tMann et al. (2010)
Fresno, Clovis, CA
n = 280, ages 6-11 yr, 47% mild persistent asthma, 25% moderate to
severe asthma, 63% with atopy
Repeated measures. Daily symptom diaries for 14 days every 3 mo.
Recruitment from schools, advertisements, physician's offices, local
media. Imputed wheeze values for 7.6% days. Participation from 89% of
original cohort. Group examined representative of original cohort. GEE
adjusted for fitted daily mean wheeze, home ownership, race, sex,
asthma severity, panel group, 6-mo cohort, 1-h minimum temperature.
Adjustment for medication use did not alter results.
NO2-central site 2
24-h avg
1 site within
20 km of
homes.
Wheeze:
All subjects:
1.2(1.1, 1.5)
Fungi allergic, n = 85
1.6(1.2,2.1)
Cat allergic, n = 49
1.7(1.1,2.6)
Boys, intermittent asthma,
n = 47
2.6(1.6,4.1)
With PMio-2.5, all subjects:
1.1 (0.95, 1.4).
PM-io-2.5 association not
' altered by NO2 adjustment.
Weak correlation with NO2.
r=0.12.
Mortimer et al. (2002)
Bronx and East Harlem, NY; Chicago, IL; Cleveland, OH; Detroit,
St.
Louis, MO; Washington, DC (NCICAS cohort)
n = 846, ages 4-9 yr
Repeated measures. Daily symptom data collected for 2-week periods
every 3 mo. Recruitment from ED visits and clinics. Parent report of
physician-diagnosed asthma and symptoms in previous 12 mo or asthma
symptoms for >6 weeks, or family history of asthma. Participation from
55% full cohort. Sample representative of full cohort except for greater
asthma medication use. Mixed effects model adjusted for city, follow-up
period, day of study, 24-h rainfall, 12-h avg temperature.
NO2-central site
4-h avg
(6 a.m.-10 a.m.)
Average of all
city monitors.
Lag 1-6 Morning symptoms:
avg 1.5(1.0,2.2)
OR per20-ppb increase in
NO2 (interquartile range).
With Os (summer):
1.4(0.93,2.1)
Weak correlation with NO2.
r=0.27.
Os effect estimate also
slightly attenuated.
SO2 and PM-io also
associated with symptoms.
Correlations with NO2 not
reported.
5-61
-------
Table 5-8 (Continued): Epidemiologic studies of respiratory symptoms and asthma medication use in children
with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics Odds Ratio (95% Cl)
Analyzed Lag Day Single-Pollutant Modela
Copollutant Examination
Studies with central site exposure assessment and no examination of copollutant confounding
Jalaludin et al. (2004) NCb-central site
Sydney, Australia 15-h avg
n = 125, mean age 9.6 yr, 45 with wheeze, asthma, and airway (6 a.m.-9 p.m.)
hyperresponsiveness, 60 with wheeze and asthma, 20 with wheeze site within 2 km
Repeated measures. Daily symptom diary mailed in monthly for 11 mo. of schools.
Recruitment from schools. Parent report of physician-diagnosed asthma.
84% follow-up participation. GEE adjusted for time trend, temperature,
humidity, number of hours spent outdoors, total pollen and Alternaria,
season.
Wheeze:
1.0(0.94, 1.1)
Wet cough:
1.05(1.00, 1.10)
Beta agonist use:
1.01 (0.97, 1.05)
OR per 8.2 ppb increase in
NO2 (interquartile range)
NO2 associations found in
children with asthma/airway
hyperresponsiveness but
examined only in
multipollutant model with
Os and PMm
Negative or weak
correlations with NO2.
r=-0.31 forOs, 0.26 for
PMio.
tEscamilla-Nufiez et al. (2008)
Mexico City, Mexico
n = 147, ages 6-14 yr, 43% with persistent asthma, 89% atopy
Repeated measures. Symptom data collected every 15 days for mean
22 weeks. Children with asthma recruited from pediatric clinic. Asthma
severity assessed by pediatric allergist. No information on participation
rate. Linear mixed effects model with random effect for subject and
adjusted for asthma severity, atopy, lag 1 minimum temperature, time,
sex. Adjustment for outdoor activities, smoking exposure, season did not
alter results.
NO2-central site 1
1-h max
Site within 5 km
of school or
home.
Cough: 1.07(1.02, 1.12)
Wheeze: 1.08(1.02, 1.14)
No copollutant model.
PM2.5 and Os also
associated with symptoms.
No statistically significant
interaction between NO2
and PM2.5 or Os.
Quantitative results not
reported.
tBarraza-Villarreal et al. (2008)
Mexico City, Mexico
n = 126, ages 6-14 yr, 44% persistent asthma, 89% with atopy
Part of same cohort as above. No information on participation rate. Linear
mixed effects model with random effect for subject and adjusted for sex,
body mass index, lag 1 minimum temperature, ICS use, time. Adjustment
for outdoor activities, smoking exposure, anti-allergy medication use,
season did not alter results.
NO2-central site 0
1-h max
Site within 5 km
of school or
home.
Low correlation
for central site
vs. school:
Spearman
r=0.21
Wheeze: 1.09(1.03, 1.15)
Cough: 1.09(1.04, 1.14)
No copollutant model.
PM2.5 and Os also
associated with symptoms.
Moderate correlations with
NO2. Pearson r = 0.61 for
PM2.5, 0.28 for Os.
5-62
-------
Table 5-8 (Continued): Epidemiologic studies of respiratory symptoms and asthma medication use in children
with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics Odds Ratio (95% Cl)
Analyzed Lag Day Single-Pollutant Modela
Copollutant Examination
Romieu et al. (2006)
NO2-central site 1-6avg Cough by genotype:
Mexico City, Mexico 1-h max
n = 151, mean age 9 yr, mild or moderate asthma Site within 5 km
Repeated measures. Daily symptom diaries 61 -92 days per subject, of home-
collected weekly. Recruitment from allergy clinic as part of a Vitamin C/E
supplementation trial. Diagnosis by clinical examination. 99% follow-up
participation. GEE adjusted for supplementation group, minimum
temperature, smoking exposure, asthma severity, time.
GSTM1 null
1.09(1.00, 1.19)
GSTM1 positive
1.19(1.11, 1.27)
GSTP1 lie/lie or Ile/Val
1.19(1.11, 1.27)
GSTP1 Val/Val
1.08(0.99, 1.18)
BD use by genotype:
GSTM1 null
0.94(0.87, 1.02)
GSTM1 positive
1.09(1.02, 1.17)
GSTP1 lie/lie or Ile/Val
1.08(1.02, 1.14)
GSTP1 Val/Val
0.94(0.85, 1.04)
No copollutant model.
Associations with Os found
with different variants than
NO2.
Moderate correlation with
NO2. Pearson r= 0.57 for
Osand PMm
Ostroetal. (2001)
Central Los Angeles and Pasadena, CA
n = 138 (83% LA), ages 8-13 yr, 85% mild or moderate asthma, 100%
African-American
Repeated measures. 90% follow-up. Daily symptom diaries for 13 weeks,
mailed in weekly. Excluded subjects returning diaries after 2 weeks.
9,126 observations. Recruitment from hospitals, urgent care clinics,
medical practices, asthma camps in Los Angeles and school nurses in
Pasadena. GEE adjusted for day of study, age, income, town, lag 1
temperature, lag 1 humidity.
NO2-central site
1-h max
Los Angeles site
within 16 km of
90% of subjects'
homes.
Pasadena site
within 8 km of
subjects'
homes.
Shortness of breath:
1.08(0.99, 1.18)
Wheeze: 1.08(1.02, 1.13)
Cough: 1.07(1.00, 1.15)
No quantitative results for
extra medication use but
reported not to be
associated with NO2.
No copollutant model.
Symptoms associated with
PM2.5, PM-io, fungi.
Weak to moderate
correlations with NO2.
r= 0.18 for pollen,
0.26-0.48 for fungi, 0.34
for PM2.5, 0.63 for PMm
5-63
-------
Table 5-8 (Continued): Epidemiologic studies of respiratory symptoms and asthma medication use in children
with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics Odds Ratio (95% Cl)
Analyzed Lag Day Single-Pollutant Modela
Copollutant Examination
tPateletal. (2010)
New York City and nearby suburb, NY
n = 57, ages 14-20 yr
Repeated measures. Daily symptom diaries for 4-6 weeks, collected
weekly. Recruitment from schools. Self-report of physician-diagnosed
asthma. 75-90% participation across schools. GLMM with random effect
for subject and school and adjusted for weekend, daily 8-h max Os, urban
location. Adjustment for season, pollen counts did not alter results.
NO2-central site 0
24-h avg
1 site 2.2-9.0
km from
schools, 1 site
40 km from
schools.
Wheeze: 1.16(0.93, 1.45)
Chest tightness:
1.26(1.00, 1.58)
No copollutant model with
BC.
BC also associated with
symptoms.
Across locations, moderate
to high correlations with
NO2. Spearman
r= 0.56-0.90.
Just et al. (2002) NCb-central site
Paris, France 24-h avg
n = 82, ages 7-15 yr, asthma attack in previous 12 mo and daily asthma Average of
medication use, 90% atopy 11 sites
Repeated measures. Daily symptom diaries for 3 mo, collected weekly.
Recruitment from hospital outpatients. 82% participation. GEE adjusted
for time trend, day of week, pollen, temperature, humidity.
Asthma attack:
1.75(0.82,3.70)
Night cough:
2.11 (1.20,3.71)
No copollutant model
BS associated with cough.
High correlation with NO2.
Pearson r= 0.92.
Seqala etal. (1998) NCb-central site
Greater Paris area, France 24-h avg
n = 43 mild asthma, 41 moderate asthma, 89% atopy, 69% ICS users, Average of
ages 7-1 Syr 8 sites
Repeated measures. Daily symptom diary for 25 weeks, collected
weekly. Recruitment from outpatients of children's hospital. 84% follow-
up participation. GEE adjusted for day of week, time trend, temperature,
humidity, age, sex.
Incident asthma:
Mild asthma, n = 43
1.89(1.13, 3.15)
Moderate asthma, n = 41
1.31 (0.85,2.03)
No copollutant model.
Associations also found
withBS, PM13, &SO2.
• Moderate correlations with
NO2. Pearson r= 0.61 for
BS, 0.55forPM13, 0.54 for
SO2.
Beta-agonist use:
Mild asthma, n = 43
1.27(0.82, 1.98)
Moderate asthma, n = 41
1.56(0.51,4.73)
5-64
-------
Table 5-8 (Continued): Epidemiologic studies of respiratory symptoms and asthma medication use in children
with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics Odds Ratio (95% Cl)
Analyzed Lag Day Single-Pollutant Modela
Copollutant Examination
tO'Connoret al. (2008)
Boston, MA; Bronx, NY; Chicago, IL; Dallas, TX; New York, NY; Seattle,
WA; Tucson, AZ (ICAS cohort)
n = 861, ages 5-12 yr, persistent asthma and atopy, 82% black or
Hispanic
Repeated measures. Symptom data collected for 2 week period every
2 mo for 2 yr. Recruitment from intervention of physician feedback. 89%
of maximum possible diaries obtained. Mixed effects model adjusted for
site, mo, sitexmo interaction, temperature, intervention group.
NO2-central site 1-19avg Wheeze-cough:
24-h avg
All monitors
near home, not
near industry.
Median distance
to site = 2.3 km.
1.17(0.99, 1.37)
Slow Play:
1.25(1.04, 1.51)
Missed school in 2 week
period:
1.65(1.18,2.32)
Only 3-pollutant model
analyzed.
Associations also found
with PM2.5 and CO.
Moderate correlations with
NO2.r =0.59 for PM2.5,
0.54 for CO.
Note: More informative studies in terms of the exposure assessment method and potential confounding considered are presented first.
a.m. = ante meridiem; avg = average; AZ = Arizona; BC = black carbon; BD = bronchodilator; BS = black smoke; CA = California; CAMP = Childhood Asthma Management Program;
Cl = confidence interval; CO = carbon monoxide, Colorado; CT = Connecticut; DC = District of Columbia; EC = elemental carbon; ED = emergency department; GEE = generalized
estimating equations; GLM = generalized linear model; GLMM = generalized linear mixed model; GSTM1 = glutathione S-transferase mu 1; GSTP1 = glutathione s-transferase Pi 1;
ICAS = Inner City Asthma Study; ICS = inhaled corticosteroid; IL = Illinois; MA = Massachusetts; MD = Maryland; Ml = Michigan; MO = Missouri; NCICAS = National Cooperative
Inner-city Asthma Study; NM = New Mexico; NO2 = nitrogen dioxide; NY = New York; O3 = ozone; OC = organic carbon; OH = Ohio; ON = Ontario; OR = odds ratio;
PM2 5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PMio = particulate matter with a nominal mean aerodynamic diameter less than or
equal to 10 |jm; SD = standard deviation; SO2 = sulfur dioxide; TX = Texas; Val = valine; WA = Washington; VOC = volatile organic compound.
aEffect estimates are standardized to a 20 ppb for 24-h avg NO2, 25 ppb for 8-h max, and a 30-ppb increase for 1-h max NO2. Effect estimates for other averaging times are not
standardized but presented as they are reported in their respective studies (Section 5.1.2.2).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-65
-------
Several studies are noteworthy for NO2 exposure assessment in subjects' locations or
analysis of the influence of other traffic-related pollutants onNO2 associations. NO2
concentrations at home or school, where subjects spend large portions of time, may better
represent ambient exposures for that location. In a group of 17 children with asthma and
allergy in El Paso, TX, a 20-ppb increase in outdoor school 4-day avg NO2 was
associated with a 0.56 (95% CI: -0.10, 1.2)-point poorer asthma control score (composite
of symptoms, activity limitation and asthma medication use) (Zoraet al.. 2013). Among
children in Bronx, NY, the wide 95% CIs did not provide strong evidence for
associations of 6-h avg school-day NO2 (9 a.m.-3 p.m.) with total symptoms (OR: 1.1
[95% CI: 0.84, 1.5] per 60-ppbNO2) or wheeze (OR: 1.2 [95% CI: 0.75, 1.91)(Spira-
Cohen etal.. 2011). Other studies did not indicate associations of school or home NO2
with respiratory symptoms in children with asthma, but it is unclear whether the results
represent inconsistency in the evidence base. Studies conducted in El Paso, TX and
Ciudad Juarez, Mexico only reported that NO2 was not associated with respiratory
symptoms in children with asthma but did not report quantitative results (Sarnat et al..
2012; Holguin et al.. 2007). Outdoor home NO2 was associated with reliever inhaler use
but not respiratory symptoms among children with asthma in multiple New Zealand
towns (Gillespie-Bennett et al.. 2011). However, daily outcomes were analyzed with a
single 4-week sample of NO2, which cannot represent temporal variability in exposure.
Home indoor NO2, which was represented as up to four measurements per subject,
showed stronger associations with both outcomes.
Most studies observed NO2-related increases in respiratory symptoms with adjustment for
temperature, humidity, season, and day of week. A few studies additionally adjusted for
asthma medication use, colds, smoking exposure, and allergens (Table 5-9). In addition to
NO2, studies with school and central site exposure assessment found symptoms
associated with the traffic-related pollutants EC/BC, black smoke (BS), OC, CO, and
VOCs, which showed a wide range of correlations with NO2 (r = 0.16-0.92) (Table 5-9).
Analysis of confounding by traffic-related copollutants is limited in both the number of
studies and array of copollutants. The only study that analyzed copollutant models for
school measurements found that neither NO2 nor BC was associated with asthma control
in a copollutant model (Zoraetal.. 2013). However, these results were based on the
whole study population. NO2 and BC were associated with asthma control only in
children with asthma and allergies; thus, the copollutant model results do not clearly
inform potential confounding. There are NO2 associations with asthma symptoms that are
independent of EC or a VOC, but potential differential exposure error for pollutants
measured at central sites limits inference. Among children in Los Angeles, CA, CO was
not associated with symptoms, and NO2-asthma symptom associations were relatively
unchanged with adjustment for various VOCs. NO2 concentrations were assigned from a
5-66
-------
central site within 4.8 km of children's homes, and although it is not certain whether
variability within neighborhoods is adequately represented, the high correlations
observed among monitors in that Los Angeles area (Section 2.5.3) provide some
confidence in the NO2 metrics to represent the broad variability the study area. In New
Haven County, CT, NC>2 was associated with wheeze with adjustment for a source
apportionment factor comprising EC, zinc, lead, copper, and selenium (OR: 1.08 [95%
CI: 0.99, 1.18] per unspecified increase in lag 0 day NCh) (Gentetal.. 2009). However,
the central site monitor was located 0.9 to 30 km (mean 10 km) from children's homes.
In limited analysis, interactions between NC>2 and traffic-related copollutants in affecting
asthma symptoms are not demonstrated. These analyses of interactions or joint effects
neither inform the potential for confounding. For children in Los Angeles, CA, no
NO2-VOC interaction was found (Delfino et al.. 2003). In the multicity Childhood
Asthma Management Program (CAMP) study, the joint effect of NO2 and CO (OR: 1.07
[95% CI: 1.00, 1.14] for a 20-ppb increase in lag 0-2 day sum of 24-h avg NO2) was
similar to the NO2 (OR: 1.05 [95% CI: 1.01, 1.09]) and CO single-pollutant ORs
(Schildcrout et al.. 2006). While these results indicate a lack of multiplicative interaction
between NO2 and CO, they may not be reliable given that pollutants were averaged from
sites up to 80 km from subjects' ZIP code centroids. Interactions with NO2 were not
found consistently for PMio and not found at all for SO2, Os, or allergens (Schildcrout et
al.. 2006; Delfino et al.. 2002). These copollutants were not examined as potential
confounding factors. An NO2-wheeze association decreased in magnitude and precision
(i.e., wider 95% CI) with adjustment for PMio-2.5 (r = 0.12) (Mann etal.. 2010). based on
exposures assessed from a site up to 20 km of children's homes.
Other studies largely corroborate the aforementioned evidence but do not provide a
strong basis for assessing an independent effect of NO2 exposure on respiratory
symptoms in children with asthma because of both central site exposure assessment and
no examination of potential confounding by other traffic-related pollutants (Table 5-8).
Confounding by PMio, SO2, or Os was not examined either. These pollutants were
moderately correlated with NO2 (r = 0.28-0.31) in most studies, although some reported
higher correlations (r = 0.54-0.68) (Ostro etal.. 2001; SegalaetaL 1998). Multipollutant
models were analyzed for traffic-related copollutants but can produce unreliable results
because of potential collinearity (Kim etal.. 2012; Escamilla-Nunez et al.. 2008;
O'Connor et al.. 2008). These multi- and single-city studies used central sites located
2-16 km from children's homes or schools (Patel etal.. 2010; Barraza-Villarreal et al..
2008; Escamilla-Nunez et al.. 2008; Romieu et al.. 2006; Jalaludin et al.. 2004; Ostro et
al.. 2001). averaged across city sites (O'Connor et al.. 2008; Just et al.. 2002; Mortimer et
al.. 2002). or an unspecified method (SegalaetaL. 1998). Low within-city correlations in
NO2 were reported for Mexico City, Mexico (Barraza-Villarreal et al.. 2008; Escamilla-
5-67
-------
Nunez et al.. 2008). and high correlations for NC>2 are reported for Los Angeles/Pasadena
(Ostro et al.. 2001) (Section 2.5.3). For other locations, information was not reported to
assess whether the temporal variation in NO2 metrics represented the variation across the
study areas. The recent multicity ICAS found increases in symptoms, slow play, and
missed school in association with a 19-day avg of 24-h avg NO2 (O'Connor et al., 2008),
but there is potential for residual temporal confounding for associations with NO2
exposure on the order of weeks. ICAS could not examine shorter lags because symptom
data were collected with a time resolution of 2 weeks.
In addition to the limited findings from copollutant models with traffic-related pollutants,
studies that show increases in respiratory symptoms in association with increases in
indoor NC>2 averaged over 3 to 7 days provide some support for an independent effect of
NO2 exposure (Luetal.. 2013; Hansel etal.. 2008). Previous findings indicated
reductions in respiratory symptoms after an intervention to switch to flued gas heaters led
to a reduction in indoor classroom NC>2 concentrations (Pilotto et al.. 2004). Although
potential differences in pollutant mixtures between the indoor and outdoor environments
are not well characterized, a recent study found that correlations of NO2 with BC, PM,
and SC>2 differed between the indoor and outdoor school environments (Sarnat et al..
2012). At most of the schools, indoor concentrations of NC>2 (medians 3.5-83.9 ppb), BC
(medians 0-1.1 (ig/m3), PM2.s (medians 7-22.8 (ig/m3), and PMio-2.5 (medians 6.5-38.9
(ig/m3) were well above the limits of detection [2.88 ppb NO2, 0.42 (ig/m3 BC, 4.62
(ig/m3 PM2.5, 2.69 (ig/m3 PMio-2.5 (Ravsoni etal.. 2011)1. In Hansel et al. (2008). most
indoor NC>2 samples (25th percentile 13.7 ppb) were above the 6.8-ppb limit of detection,
and indoor-outdoor NO2 correlations were low (R2 = 0.06). These results suggest that
NC>2 may exist as part of a different pollutant mixture in the indoor and outdoor
environments. The differing sources for indoor and outdoor NC>2 indicate that health
effects related to indoor NO2 exposure may not be confounded by the same traffic-related
copollutants as outdoor NO2.
Adults with Asthma
Previous and recent evidence indicates associations of ambient NO2 concentrations with
respiratory symptoms (Maestrelli et al.. 2011; Wiwatanadate and Liwsrisakun. 2011; von
Klot et al.. 2002; Boezenetal.. 1998; Forsberg et al.. 1998) and asthma medication use or
sales (Carlsen et al.. 2012; Laurent et al.. 2009; von Klot et al.. 2002; Forsberg et al..
1998; Hiltermann et al.. 1998) among adults with asthma or bronchial
hyperresponsiveness. Most studies were conducted in Europe and recruited subjects
primarily from clinics, doctors' offices, and administrative databases. Subjects
represented a mix of asthma severity and prevalence of ICS use and atopy. A few studies
did not find associations with symptoms, including Kim et al. (2012). who analyzed only
5-68
-------
a multipollutant model with SC>2, PMio, Os, and CO, the results of which can be unstable.
Null results also were reported in studies with more reliable statistical analysis: one
conducted in four European countries (Karakatsani et al.. 2012) and one with adults with
asthma and allergy (Feo Brito et al.. 2007). Results from the latter study contrast those
from experimental studies showing NC>2-induced allergic inflammation in humans with
asthma and animal models of allergic disease (Section 5.2.2.5). Across studies,
respiratory symptoms were associated with Lag Day 0 NC>2. Medication use or sales were
associated more strongly with multiday averages of NC>2 (i.e., lag 3-5 avg, 0-5 avg,
0-6 avg) than with single-day lags (Carlsen et al., 2012; von Klot et al., 2002;
Hiltermann et al.. 1998). and Carlsen etal. (2012) found a stronger association for
beta-agonist sales with 1-h max than 24-h avg NC>2.
Despite some supporting evidence, there is limited basis for inferring NC>2 effects on
respiratory symptoms in adults with asthma. Most studies assigned NC>2 exposure from a
single central site monitor located in the community and did not report information to
assess whether the NCh metrics were representative of the temporal variability across the
study areas and of subjects' ambient exposures. With exception of Boezen et al. (1998).
the studies found associations with PM2 5 and the traffic-related pollutants CO, BS, UFP,
and few of the studies analyzed potential confounding. Confounding is an uncertainty
also for the association between beta-agonist sales and block-level NO2 estimated with a
dispersion model (Laurent et al.. 2009). Dispersion model estimates were highly
correlated with measured concentrations (r = 0.87), but other traffic-related pollutants
were not examined. Only von Klot et al. (2002) conducted copollutant modeling and
found an association between lag 0-4 day avg NO2 and beta-agonist use with adjustment
for PM25 or UFP (OR: 1.22 [95% CI: 1.05, 1.43] per 20-pbb increase in NO2, with
adjustment for UFP, Pearson r = 0.66). The NO2-wheeze association was attenuated with
adjustment for UFP (OR: 1.02 [95% CI: 0.86, 1.21]). Copollutant effect estimates were
attenuated with NO2 adjustment. Thus, an independent NO2 association was found for
medication use, but an independent association with wheeze was not discerned for either
NO2 or UFP.
Controlled Human Exposure Studies
Controlled human exposure studies, including the single recent study, do not provide
strong evidence for NO2-induced increases in respiratory symptoms in adults or
adolescents with asthma. NO2 exposures of 120-350 ppb for 30 minutes to 2 hours with
exercise (Table 5-9) tended to result in no change in symptoms during or after
(immediately to 1 week) exposure to NO2 (Vagaggini et al.. 1996; Koenig etal.. 1987;
Kleinman et al.. 1983). NO2 exposures of 1,000 and 4,000 ppb also did not induce
respiratory symptoms in adults with asthma (Torres etal.. 1995; Linn etal.. 1985b).
5-69
-------
Vagaggini et al. (1996) reported a small, but statistically significant increase in symptom
score during 300 ppb NC>2 exposures in healthy adults but not those with asthma. Riedl et
al. (2012) reported an increase in symptom score in adults with asthma during, but not
after, exposure to 350 ppb NCh for 2 hours with alternating periods of exercise. The
increase in symptom score corresponded to a mild increase in any two symptoms or
moderate elevation of any one symptom. Unlike studies of airway responsiveness
(Section 5.2.2.1). most symptom studies did not include a challenge agent with NC>2
exposure. Riedl et al. (2012) was unique in assessing symptoms in subjects exposed to
allergen after NO2 exposure but found no change in symptoms.
Table 5-9 Characteristics of controlled human exposure studies of respiratory
symptoms in individuals with asthma.
Study
Koeniq et al.
(1987)
Kleinman et al.
(1983)
Vaqaqqini et al.
(1996)
tRiedletal.
(2012)
Disease Status; Sample Size; Sex;
Age (mean ± SD)
Asthma; n = 4 or 7 M, 3 or 6 F
Healthy; n = 3 or 4 M, 6 or 7 F
14.4 yr (range: 12-19)
Asthma; n = 12 M, 19 F;
31 ± 1 yr
Asthma; n = 4 M, 4 F; 29 ± 14 yr
Healthy; n = 7 M; 34 ± 5 yr
Asthma
Phacp 1 f mptharhnlinp rhallpnnpV
Exposure Details
120 ppb or 180 ppb NCb;
Exposures were 30 min at rest,
10 min of moderate exercise.
200 ppb for 2 h;
Exercise 15 min on/15 min off at
VE = ~2 times resting
300 ppb for 1 h;
Exercise at VE = 25 L/min
350 ppb for 2 h;
FvprriQp 1 R min nn/1R min nff at
Time of Symptom
Assessment
Immediately after, a
day after, and a week
after exposure.
Before, immediately
after, and day after
exposure.
Before and 2 h after
exposure.
Before, during, 1-22 h
after exposure.
n = 10 M, 5F; 37.3 ± 10.9 yr
Phase 2 (cat allergen challenge);
n = 6M, 9F; 36.1 ±12.5yr
VE = 15-20 L/min
Jorres et al. Asthma; n = 8 M, 4 F; 27 ± 5 yr
(1995) Healthy; n = 5 M, 3 F;
27 yr (range: 21-33)
1,000 ppb for 3 h;
Exercise 10 min on/10 min off at
individual's maximum workload
Immediately and 6 and
24 h after exposure.
Linn et al.
(1985b)
Asthma; n = 12 M, 11 F;
(range: 18-34yr)
Healthy; n = 16 M, 9 F;
(range: 20-36 yr)
4,000 ppb for 75 min;
Two 15 min periods of exercise
at VE = 25 L/min and 50 L/min
Before, during,
immediately after,
1 day after, and
1 week after exposure.
F = female; L/min = liters per minute; M = male; NO2 = nitrogen dioxide; ppb = parts per billion; SD = standard deviation;
VE = minute ventilation; yr = year.
fStudy published since the 2008 ISA for Oxides of Nitrogen.
5-70
-------
5.2.2.4 Hospital Admissions and Emergency Department Visits for Asthma
The evidence for NCh-related effects on increasing airway responsiveness, decreasing
lung function, and increasing respiratory symptoms detailed in the preceding sections are
all indicators of asthma exacerbation that may lead people with asthma to seek medical
interventions. Thus, the preceding evidence is coherent with associations observed
between short-term increases in ambient NO2 concentrations and hospital admissions and
ED visits for asthma. Since the completion of the 2008 ISA for Oxides of Nitrogen,
epidemiologic studies have continued to examine the association between short-term
exposure to ambient NOx or NO2 and respiratory-related hospital admissions and ED
visits, but are primarily limited to single-city studies. As summarized in Section 5.2.6.2.
the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) included the first thorough
evaluation of respiratory morbidity in the form of respiratory-related hospital admissions
and ED visits, including those for asthma. Previous studies of asthma hospital admissions
and ED visits consistently reported positive associations with short-term increases in
ambient NO2 concentration (Figure 5-7 and Table 5-13). The few studies that analyzed
copollutant models with CO or PM25 generally observed robust NO2 associations (U.S.
EPA. 2008c). Confounding by other traffic-related pollutants was not examined. The
strongest evidence for associations between short-term NO2 exposures and
asthma-related hospital admissions and ED visits was from studies of all ages and
children.
For asthma as well as other respiratory outcome groups, studies of hospital admissions
and ED visits are evaluated separately because often only a small percentage of
respiratory-related ED visits will be admitted to the hospital. Therefore, ED visits may
represent potentially less serious, but more common, outcomes. The air quality
characteristics of the city, or across all cities, and the exposure assignment approach used
in each asthma hospital admission and ED visit study evaluated in this section are
presented in Table 5-10. Other recent studies of asthma hospital admissions and ED visits
are not the focus of this evaluation because they were conducted in small individual
cities, encompass a short study duration, and/or had insufficient sample size. In addition
to these limitations, none examined potential copollutant confounding. The full list of
these studies, as well as study specific details, can be found in Supplemental Table S5-3
(U.S. EPA. 2015h).
5-71
-------
Table 5-10 Mean and upper percentile concentrations of oxides of nitrogen in studies of asthma hospital
admissions and emergency department visits.
Study
Location
Years
Mean/Median
Concentration Upper Percentile of
Exposure Assignment Metric ppb Concentrations (ppb) Copollutant Examination
Hospital Admissions
Linnetal. (2000)
Los Angeles,
CA
(1992-1995)
Average of NO2 24-h avg 3.4 NR Correlations (r):
concentrations over all Range across seasons
monitors- rn- n aa-n u
Burnett et al. (1999)
Toronto, ON,
Canada
(1980-1994)
Average of NO2
concentrations from 4
monitors.
24-h avg
25.2
NR
PMio: 0.67-0.88
O3: -0.23 to 0.35
Copollutant models: none
Correlations (r):
PlVh.s: 0.55
PMio-2.s: 0.38
PMio: 0.57
CO: 0.64
SO2: 0.54
O3: -0.08
Copollutant models: none
tSamolietal. (2011)
Athens,
Greece
(2001-2004)
Average of NO2
concentrations across 14
monitors.
1-h max
44.4
75th: 53.1 Correlations (r):
SOi. 0.55
Copollutant models: PMio, SO2, Os
5-72
-------
Table 5-10 (Continued): Mean and upper percentile concentrations of oxides of nitrogen in studies of asthma
hospital admissions and emergency department visits.
Study
Location
Years
Exposure Assignment Metric
Mean/Median
Concentration
ppb
Upper Percentile of
Concentrations (ppb) Copollutant Examination
tlskandaretal. (2012)
Copenhagen,
Denmark
(2001-2008)
All hospitals located
within 15 km of a central
site monitor.
24-h avg
NO2: 11.3
NOx: 14.5
75th:
NO2: 14.2
NOx: 17.7
Correlations (r):
NOx: 0.93
PMio: 0.43
PlVh.s: 0.33
UFP: 0.51
Copollutant models: NOx, PIVhs,
PMio, UFP
tKo et al. (2007b)
Hong Kong,
China
(2000-2005)
Average of NO2
concentrations across
14 monitors.
24-h avg
28.3
75th: 33.8
Max: 79.5
Correlations (r):
SO2: 0.57
PMio: 0.76
O3: 0.41
PIvh.s: 0.77
Copollutant models: Os
tSon etal. (2013)
8 South Average of hourly
Korean cities ambient NO2
(2003-2008) concentrations from
monitors in each city.
24-h avg
11.5-36.9
NR
Correlations (r):
PMio: 0.5
O3: -0.1
SO2: 0.6
CO: 0.7
Copollutant models: none
5-73
-------
Table 5-10 (Continued): Mean and upper percentile concentrations of oxides of nitrogen in studies of asthma
hospital admissions and emergency department visits.
Study
Location
Years
Exposure Assignment Metric
Mean/Median
Concentration
ppb
Upper Percentile of
Concentrations (ppb) Copollutant Examination
ED Visits
Peel et al. (2005)
Atlanta, GA Average of NO2
(1993-2000) concentrations from
monitors for several
monitoring networks.
1-h max
45.9
NR
Correlations (r):
PlVh.s: 0.46
PMio: 0.49
PMio-2.s: 0.46
UFP: 0.26
PIvh.s water soluble Metals: 0.32
PM2.5sulfate:0.17
PIvh.s acidity: 0.10
PIvh.s OC: 0.63
PIvh.s EC: 0.61
Oxygenated HCs: 0.30
O3: 0.42
CO: 0.68
SO2: 0.34
Copollutant models: none
Tolbert et al. (2000)
Atlanta, GA Concentrations from one
(1993-1995) central site NOx monitor.
1-h max
81.7
Max: 306 Correlations (r):
PMio: 0.44
O3: 0.51
Copollutant models: none
5-74
-------
Table 5-10 (Continued): Mean and upper percentile concentrations of oxides of nitrogen in studies of asthma
hospital admissions and emergency department visits.
Study
Jaffe et al. (2003)
Location
Years
2 Ohio cities
(Cincinnati
and
Cleveland)
(1991-1996)
Exposure Assignment Metric
When more than 1 NO2 24-h avg
monitor operating in a
day, monitor with highest
24-h avg concentration
used.
Mean/Median
Concentration
ppb
Cincinnati: 50
Cleveland: 48
Upper Percentile of
Concentrations (ppb) Copollutant Examination
NR Correlations (r):
Cincinnati
PMio: 0.36
O3: 0.60
SO2: 0.07
Cleveland
PMio: 0.34
SO2: 0.28
O3: 0.42
Copollutant models:
none
Ito et al. (2007) New York, Average of NO2
NY concentrations from
(1999-2002) 15 monitors.
24-h avg
31.1
NR
Correlations (r): NR
Copollutant models: PlVh.s, Os, SO2,
CO
ATSDR (2006)
Bronx and NO2 concentrations from 24-h avg
Manhattan, 1 monitor in Bronx and 1
NY in Manhattan.
(1999-2000)
Bronx: 36
Manhattan: 31
NR Correlations (r):
Bronx
Os: 0.03
crw n /\7
FRM PM2.5: 0.61
Max PM2.s: 0.55
Manhattan: NR
Copollutant models: Os, FRM and
Max PM2.5, SO2
tStrickland et al.
(2010)
Atlanta, GA
(1993-2004)
Combined daily NO2
concentrations across
monitors using
population weighting.
1-h max
23.3
NR
Correlations (r): NR
Copollutant models:
5-75
-------
Table 5-10 (Continued): Mean and upper percentile concentrations of oxides of nitrogen in studies of asthma
hospital admissions and emergency department visits.
Study
tSarnatetal. (201 3a)
fVilleneuve et al.
(2007)
tJalaludin et al. (2008)
tStieb et al. (2009)
tOrazzo et al. (2009)
Location
Years
Atlanta, GA
(1999-2002)
Edmonton,
AB, Canada
(1992-2002)
Sydney,
Australia
(1997-2001)
7 Canadian
cities
(1992-2003)
6 Italian
cities
(1996-2002)
Exposure Assignment
NOx concentrations
predicted using fused
spatially interpolated
background pollutant
concentrations and
local-scale AERMOD
output for 186 ZIP codes.
Average of NO2
concentrations across 3
monitoring stations.
Average of NO2
concentrations across
14 monitoring stations.
Average NO2
concentrations from all
monitors in each city.
Number of NO2 monitors
in each city ranged from
1-14.
Average of NO2
concentrations from all
monitors in each city.
Mean/Median
Concentration Upper Percentile of
Metric ppb Concentrations (ppb) Copollutant Examination
24-h avg NOx: 30.1 75th: 40.1 Correlation (r) with NOx:
95th: 94.4 CO: 0.93
Max: 51 7.8 O3: -0.03
PlVh.s: 0.40
Copollutant models: none
24-h avg 50th: 17.5 75th: 22.0 summer Correlations (r):
Summer 75th: 35.5 winter CO: 0.74
50th: 28.5 Winter Copollutant models: CO
1-h max 23.2 Max: 59.4 Correlations (r):
PMio: 0.67
PM2.s: 0.68
O3: 0.21
CO: 0.71
SO2: 0.52
Copollutant models: PMio, PM2.5, Os,
CO, SO2
24-h avg 9.3-22.7 75th: 12.3-27.6 Correlations (r) only reported by city
and season.
Copollutant models: none
24-h avg 21.4-41.2 NR Correlations (r): NR
Copollutant models: none
5-76
-------
Table 5-10 (Continued): Mean and upper percentile concentrations of oxides of nitrogen in studies of asthma
hospital admissions and emergency department visits.
Study
tStrickland et al.
(2011)
Location
Years
Atlanta, GA
(1993-2004)
Exposure Assignment Metric
NO2 concentrations 1-h max
obtained from 3 networks
Mean/Median
Concentration
ppb
Central monitor:
42.0
Upper Percentile of
Concentrations (ppb)
NR
Copollutant Examination
Correlations (r): NR
Copollutant models: none
Each air pollutant
measured by at least 3
monitoring stations.
3 exposure metrics used:
(1) 1 downtown monitor
was selected to be the
central site monitor, (2) all
monitors used to calculate
unweighted average of
pollutant concentrations
for all monitors, and
(3) population-weighted
average concentration.
Unweighted
average: 27.7
Population-
weighted average:
22.0
tLietal. (2011 b)
Detroit, Ml
(2004-2006)
Average of NO2
concentrations from 2
monitors in Detroit
metropolitan area that
measure NO2.
24-h avg
15.7
75th: 21.2
Max: 55.2
Correlations (r), range across
monitors:
CO: 0.37-0.40
PM2.5: 0.56-0.66
SO2: 0.42-0.55
Copollutant models: none
tGassetal. (2014)
Atlanta, GA
(1999-2009)
Population-weighted
average NO2
concentrations based on
same methods as
Strickland et al. (2010).
24-h avg
NR
NR
Correlations (r): NR
Copollutant models: none
5-77
-------
Table 5-10 (Continued): Mean and upper percentile concentrations of oxides of nitrogen in studies of asthma
hospital admissions and emergency department visits.
Study
Location
Years
Exposure Assignment Metric
Mean/Median
Concentration
ppb
Upper Percentile of
Concentrations (ppb) Copollutant Examination
tWinquist et al. (2014) Atlanta, GA
(1998-2004)
Population-weighted
average of NO2
concentrations.
1-hr max
Warm (May-Oct):
22.3
Cold (Nov-April):
25.6
75th:
Warm: 28.7
Cold: 31.7
Correlations (r):
Warm:
O3: 0.54
CO: 0.75
SO2: 0.44
PlVh.s: 0.52
EC: 0.68
Sulfate: 0.27
Secondary PlVh.s: 0.31
Cold:
O3: 0.30
CO: 0.74
SO2: 0.41
PIvh.s: 0.49
EC: 0.57
Sulfate: 0.08
Secondary PlVh.s: 0.12
5-78
-------
Table 5-10 (Continued): Mean and upper percentile concentrations of oxides of nitrogen in studies of asthma
hospital admissions and emergency department visits.
Study
Location
Years
Mean/Median
Concentration Upper Percentile of
Exposure Assignment Metric ppb Concentrations (ppb) Copollutant Examination
Physician Visits
tBurra et al. (2009)
tSinclairetal. (2010)
Toronto, ON,
Canada
(1992-2001)
Atlanta, GA
(1998-2002)
Average of NO2 1-h max 39.2 95th: 60 Correlations (r): NR
concentrations from 6 Max: 10s Copollutant models: none
monitors.
NO2 concentrations 1-h max 1998-2000:49.8 NR Correlations (r): NR
collected as part of 2000-2002: 35.5 Copollutant models: none
AIRES at SEARCH „„„„ „„„„ „„ _,
Jefferson street site. 1998-2002:41.7
Avg = average; AERMOD = American Meteorological Society/Environmental Protection Agency Regulatory Model; ATSDR = Agency for Toxic Substances and Disease Registry;
AIRES = Aerosol Research Inhalation Epidemiology Study; AB = Alberta; CA = California; CO = carbon monoxide; EC = elemental carbon; ED = emergency department;
FRM = Federal Reference Method; GA = Georgia; Ml = Michigan; NO2 = nitrogen dioxide; NOX = sum of NO and NO2; NR = not reported; NY = New York; O3 = ozone; OC = organic
carbon; ON = Ontario; PM25 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PMio = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 10 |jm; SEARCH = Southeastern Aerosol Research and Characterization; SO2 = sulfur dioxide; UFP = ultrafine particles.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-79
-------
Hospital Admissions
Generally, studies evaluated in the 2008 ISA for Oxides of Nitrogen that examined the
effect of short-term NO2 exposures on asthma hospital admissions were limited to single
cities. The results of these studies should be viewed with caution because they tended to
include ages <5 years in the study population, which is problematic considering the
difficulty in reliably diagnosing asthma within this age range [National Asthma
Education and Prevention Program Expert (NAEPP. 2007)1. However, it is unlikely the
inclusion of these individuals in a study would introduce a systematic positive bias. In
contrast, the majority of studies on asthma ED visits (discussed in the next section) have
excluded ages <2 years in analyses to account for this difficulty.
In a time-series study conducted in Athens, Greece, Samoli etal. (2011) evaluated the
association of multiple ambient air pollutants and pediatric asthma hospital admissions
for ages 0-14 years. In an all-year analysis, the authors reported a positive association
with NO2 (6.4% [95% CI: -3.8, 17.6]; lag 0 increase for a 30-ppb increase in 1-h max
NO2 concentrations). An examination of additional lags (quantitative results not
presented) revealed a similar pattern of associations at lag 2 and a 0-2 days distributed
lag. In copollutant analyses, NO2 risk estimates were robust when Os (7.6% [95% CI:
-2.7, 19.0]) was included in the model, and were attenuated but remained positive with
wide confidence intervals when including PMio in the model (3.1% [95% CI: -7.3,
14.6]). There was evidence of confounding of the NO2 association when SO2 was
included in the model as demonstrated by an effect estimate and confidence interval for
NO2 reflective of a null association (-4.3% [95% CI: -16.9, 10.2]). Of the two
copollutants examined, SO2 was most highly correlated with NO2 (r = 0.55).
The association between short-term air pollution exposures and asthma hospital
admissions in children (0-18 years of age) was also examined in a study conducted by
Iskandar et al. (2012) in Copenhagen, Denmark. In a time-stratified case-crossover
analysis using an a priori lag of 0-4 days, the authors reported positive associations for
both NO2 (OR: 1.3 [95% CI: 1.1, 1.6] for a 20-ppb increase in 24-h avg NO2
concentrations) and NOx (OR: 1.6 [95% CI: 1.3,2.1] for a 40-ppb increase in 24-h avg
NOx concentrations), which are larger in magnitude than those observed in Samoli et al.
(2011). Within this study NOx and NO2 were highly correlated (r = 0.93). Correlations
for NOx and NO2 with PM2 5 and UFPs ranged from, r = 0.28-0.33 and r = 0.45-0.51,
respectively. The high correlation between NOx and NO2, and the fact NO2 is part of
NOx, suggests that these pollutants should not be included in the same model due to the
inability to clearly examine whether one pollutant has an independent effect compared to
5-80
-------
the other. In additional copollutant models, NCh and NOx associations remained
relatively unchanged in models with PIVb 5 or UFP (Table 5-11).
Table 5-11 Copollutant model results from Iskandar et al. (2012) for a 20-ppb
increase in 24-h average nitrogen dioxide (NO2) concentrations and a
40-ppb increase in 24-h average NOx (sum of NO and NO2)
concentrations.
Pollutant
NOx
NO2
Copollutant
—
NO2
PMio
PM2.5
UFP
—
NO2
PMio
PM2.5
UFP
Odds Ratio (95% Cl)
1.6(1.3,2.1)
1.7(0.8, 3.5)
1.4(1.1, 1.8)
1.6(1.2,2.1)
1.6(1.2,2.2)
1.3(1.1, 1.6)
1.0(0.6, 1.6)
1.3(1.0, 1.5)
1.4(1.2, 1.7)
1.5(1.2, 1.8)
Cl = confidence interval; NO2 = nitrogen dioxide; NOX = sum of NO and NO2; PM25 = participate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |jm; PMio = particulate matter with a nominal mean aerodynamic diameter less
than or equal to 10 |jm; ppb = parts per billion; UFP = ultrafine particles.
Source: Iskandar et al. (2012)
Ko et al. (2007b) examined the association between short-term increases in NO2 and
asthma hospital admissions for all ages at both single- and multi-day lags in Hong Kong,
China. In a time-series analysis the authors reported positive associations at single-day
lags that were smaller in magnitude than those observed in Samoli etal. (2011)
[e.g., 3.4% (95% Cl: 1.9, 5.4); lag 0 for a 20-ppb increase in 24-h avg NO2
concentrations]. However, the results of Ko et al. (2007b) are consistent with those of
Son etal. (2013) in eight South Korean cities, who found the strongest association at lag
0 between short-term NCh exposures and asthma as well as allergic disease hospital
admissions, which encompasses asthma (3.6% [95% Cl: 0.5, 6.8] and 3.8% [95% Cl: 1.0,
6.6], respectively for a 20-ppb increase in 24-h avg NCh concentrations). Unlike Samoli
5-81
-------
et al. (2011) and Son et al. (2013). Ko et al. (2007b) found the strongest evidence of an
association between short-term NCh exposures and asthma hospital admissions at
multiday lags of 0-3 (10.9% [95% CI: 8.1, 13.8]) and 0-4 (10.9% [95% CI: 8.1, 13.4])
days. In a copollutant model with Os, although risk estimates remained positive (2.3%
[95% CI: -0.8, 5.8]; lag 0-4 days), evidence indicated a reduction in NO2 risk estimates.
The attenuation occurred even though NCh and Os were not strongly correlated (r = 0.41)
in Hong Kong, China. In contrast, Samoli et al. (2011) observed an increase in NC>2 risk
estimates with adjustment for Os but attenuation with adjustment for PMio or SCh.
Emergency Department Visits
Similar to the asthma hospital admission studies evaluated in the 2008 ISA for Oxides of
Nitrogen, the majority of ED visit studies were limited to single-city studies. However,
these studies provided additional information regarding potential seasonal differences in
risk estimates, indicating some evidence of larger associations during warmer months.
Strickland et al. (2010) examined the association between NCh exposure and pediatric
asthma ED visits (ages 5-17 years) in Atlanta, GA, using air quality data over the same
years as two studies [Darrow et al. (201 la) and Tolbert et al. (2007)1 that focused on total
respiratory ED visits (Section 5.2.6). However, unlike Darrow et al. (201 la) and Tolbert
et al. (2007). which used a single-site, centrally located monitor and the average of
multiple monitors to assign exposure, Strickland et al. (2010) used population weighting
to combine daily pollutant concentrations across monitors. In this study, the authors
developed a statistical model using hospital-specific, time-series data that is essentially
equivalent to a time-stratified case-crossover analysis (i.e., using interaction terms
between year, month, and day of week to mimic the approach of selecting referent days
within the same month and year as the case day). Strickland et al. (2010) reported an
8.6% (95% CI: 4.2, 13.3) increase in ED visits for a 30-ppb increase in 1-h max NC>2
concentrations at lag 0-2 days in an all-year analysis. The potential confounding effects
of other pollutants on the NC^-asthma ED visit relationship was only examined in a
copollutant model with Os, and correlations between pollutants were not presented. In the
copollutant model, NCh risk estimates were found to be relatively unchanged upon the
inclusion of Os (quantitative results not presented).
The magnitude of the association between short-term NCh concentrations and asthma ED
visits observed in Strickland et al. (2010) is larger than that observed in Sarnat et al.
(2013a) in a study also conducted in Atlanta, GA, which focused on the influence of air
exchange rates on air pollution-asthma ED visit associations detailed in Chapter 3.
Instead of using monitored NC>2 concentrations, Sarnat etal. (2013a) estimated NOx
exposures by "fus(ing) spatially interpolated background concentrations and the
5-82
-------
local-scale air quality model AERMOD output for the 186 ZIP code centroids" in the
Atlanta metro area. Also, focusing on a lag of 0-2 days, the authors reported a 1.3%
increase in asthma ED visits (95% CI: 0.0, 2.4) for a 40-ppb increase in 24-h avg NOx
concentrations. They did not examine copollutant models but found NOx to be highly
correlated with CO (r = 0.93). The magnitude of the association differs between
Strickland et al. (2010) and Sarnatet al. (2013a). which could be a reflection of:
(1) exposure measurement error and differences in exposure assessment methods for NOx
compared to NO2 and (2) the different age ranges included in both studies. The latter
explanation is supported by earlier studies conducted in Atlanta, GA by Tolbert et al.
(2000) and Peel et al. (2005) that focused on all ages and that reported associations
similar in magnitude to that observed in Sarnatet al. (2013a) (Figure 5-7).
Additional evidence for an association between short-term increases in NO2
concentrations and asthma ED visits comes from studies conducted in Edmonton, Canada
(Villeneuve et al.. 2007) and Sydney, Australia (Jalaludin et al.. 2008). Villeneuve et al.
(2007) reported evidence of positive associations between short-term NO2 concentrations
and asthma ED visits for multiple lag structures (lag 1, lag 0-2, and lag 0-4 days) in the
population aged 2 years and older. The authors observed the strongest association for
lag 0-4 days (4.5% [95% CI: 0, 7.5] for a 20-ppb increase in 24-h avg NO2
concentrations). There was no evidence of an association at lag 0. In this study, NO2 and
CO were strongly correlated (r = 0.74), and as a result, associations were examined in
copollutant models for each age group examined in the study, focusing on the warm
season (April-September). In copollutant models with CO, NO2 associations with asthma
ED visits were relatively similar to single-pollutant results except for one age group,
15-44 years, but in all instances NO2 associations were larger in magnitude than those
for CO (quantitative results not provided).
In a study focusing on children 1-14 years old, Jalaludin et al. (2008) examined air
pollution associations with asthma ED visits for single day lags up to 3 days as well as
the average of 0-1 day lags. Jalaludin et al. (2008) observed a similar magnitude of an
association for both lag 0 (7.5% [95% CI: 4.5, 10.5]) and lag 0-1 days (7.8% [95% CI:
4.5, 11.1] for a 30-ppb increase in 1-h max NO2 concentrations). An examination of the
potential confounding effects of other pollutants was assessed in copollutant models with
PMio, PM25, Os, CO, or SO2. NO2 was moderately to weakly correlated with each of these
pollutants (r ranging from 0.21-0.71). In copollutant models, the NO2-asthma ED visit
association remained positive, but was slightly attenuated with the magnitude of the
association ranging from a 4.2-6.1% increase in asthma ED visits. In addition to
analyzing ages 1-14 years, the authors examined whether risks varied among age ranges
within this study population (see Chapter 7).
5-83
-------
In contrast with the majority of the evidence, short-term increases in NC>2 concentrations
were not associated with asthma ED visits in a multicity study conducted in seven
Canadian cities (Stieb etal.. 2009). Compared to the other asthma ED visit studies
evaluated, mean NO2 concentrations across the cities included in this study were the
lowest with all cities having mean 24-h avg concentrations <23 ppb (Table 5-10). Stieb et
al. (2009) examined the association between short-term NCh exposure and a number of
respiratory-related ED visits for all ages. There was no evidence that NO2 was associated
with asthma ED visits at single-day lags of 0 to 2 days (0.0% [95% CI: -2.6, 2.7]; lag 2
for a 20-ppb increase in 24-h avg NC>2 concentrations). Additionally, there was no
evidence of associations between respiratory-related ED visits, including asthma, and air
pollution averaged over subdaily time scales (i.e., 3-h avg of ED visits versus 3-h avg
pollutant concentrations).
Emergency Department Visits for Wheeze
As stated previously [National Asthma Education and Prevention Program Expert
(NAEPP. 2007)1. asthma is difficult to diagnose in children less than 5 years of age;
however, asthma-like symptoms in children within this age range are often presented in
the form of transient wheeze. Although studies that examine ED visits for wheeze do not
directly inform the relationship between short-term NCh exposures and asthma, they can
add supporting evidence. Notably, some studies that examine asthma ED visits, as well as
hospital admissions, often include International Classification of Diseases (ICD) codes
for wheeze in the definition of asthma [e.g., (Sarnat et al.. 2013a)1. Orazzo et al. (2009)
examined the association between NO2 and wheeze ED visits in children, (ages
0-2 years) in six Italian cities. Daily counts of wheeze were examined in relation to air
pollution using a time-stratified case-crossover approach in which control days were
matched on day of week in the same month and year as the case day. PMio, 862, CO, and
Os were also evaluated, but correlations with NO2 were not reported nor were copollutant
analyses conducted. The authors reported positive associations between short-term
24-h avg NC>2 exposures and wheeze ED visits when examining various multiday lags
(0-1 through 0-6 days) with risk estimates ranging from 1.1% (95% CI: -1.2, 3.4) for
lag 0-1 days to 2.5% (95% CI: -0.9, 6.0) for lag 0-6 days.
Outpatient and Physician Visit Studies
Several recent studies examined the association between ambient NO2 concentrations and
less severe asthma exacerbation, which are often encountered through physician or
outpatient (nonhospital, nonED) visits. Burra et al. (2009) examined asthma physician
visits among patients aged 1-17 and 18-64 years focusing on differences by sex and
5-84
-------
income within age categories in Toronto, Canada. The authors reported evidence of
consistently positive associations between short-term increases in NCh concentrations
and asthma physician visits across the single- and multi-day lags examined (i.e., 0, 0-1,
0-2, 0-3, and 0-4 days). The magnitude of the effect estimates were found to be similar
between sexes, income quintiles, and both within and between ages. In a study conducted
in Atlanta, GA, Sinclair et al. (2010) examined the association between air pollution and
a number of respiratory-related (including asthma) outpatient visits to a managed care
organization. The authors separated the analysis into two time periods to compare the air
pollutant concentrations and relationships for acute respiratory visits for the 25-month
time period examined in Sinclair and Tolsma (2004) (i.e., August 1998-August 2000)
and an additional 28-month time period of available data from the Atlanta Aerosol
Research Inhalation Epidemiology Study (AIRES) (i.e., September 2000-December
2002). Across the two time periods, mean 1-h max NCh concentrations were lower in the
28-month versus the 25-month time period, 35.5 versus 49.8 ppb, respectively (Table
5-10). A comparison of the two time periods indicated that risk estimates across
outcomes tended to be larger in the earlier 25-month period compared to the later
28-month period, with evidence of consistently positive associations at lags of 0-2 and
3-5 days for asthma, but confidence intervals were relatively large.
Examination of Seasonal Differences
In addition to examining the association between short-term NC>2 concentrations and
asthma hospital admissions and ED visits in all-year analyses, some studies also
conducted seasonal analyses. Overall, these studies generally provide evidence of larger
associations in the warm or summer season compared to cooler months (Figure 5-7).
Notably, these studies did not examine potential copollutant confounding by season, and
the correlation between personal and ambient measures differ by season (Section 3.4.2).
both of which could influence the results of the studies presented below.
In the study of eight South Korean cities, Sonetal. (2013) examined potential seasonal
differences across respiratory hospital admission outcomes, including asthma and allergic
disease. For both outcomes, the association with NO2 was largest in magnitude during the
summer (asthma: 16.2% [95% CI: 5.1, 28.6], lag 0; allergic disease: 15.9 [95% CI: 4.6,
28.5], lag 0 for a 20-ppb increase in 24-h avg NC>2 concentrations) despite the lowest NO2
concentrations during the summer season (<20 ppb compared to >24 ppb in the other
seasons) across the eight cities. However, when using the warm season as the referent in
Hong Kong, China, Ko et al. (2007b) reported evidence of larger effects in the winter
(i.e., December to March), suggesting that differences in seasonal associations may vary
by geographic location. The difference in seasonal associations by geographic location is
further highlighted in a study by Samoli et al. (2011) conducted in Athens, Greece that
5-85
-------
reported results consistent with Son et al. (2013). Although risk estimates for asthma
hospital admissions were relatively consistent across winter, spring, and autumn, ranging
from a 13.1 to a 13.8% increase per 20-ppb increase in 24-h avg NC>2, the largest
percentage increase was observed forthe summer (28.7% [95% CI: -3.4, 71.3]).
The asthma ED visit studies that conducted seasonal analyses also reported seasonal
patterns similar to those observed in the hospital-admission studies. Villeneuve et al.
(2007) reported associations to be generally stronger in the warm season (e.g., 21.4%
[95% CI: 13.6, 31.0] at lag 0-4 days for a 20-ppb increase in 24-h avg NO2
concentrations) than in the cold season (-2.9% [95% CI: -7.3, 1.5]) in Edmonton,
Canada. Additionally, Jalaludin et al. (2008) found evidence of larger effects during the
warm months (November-April) compared to the cold months (May-October) in
Sydney, Australia (Figure 5-7). These results are consistent with Strickland et al. (2010).
which reported stronger associations during the warm season (i.e., May-October) (16.0%
[95% CI: 9.1, 23.5]; lag 0-2 days) than the cold season (3.8% [95% CI: -1.9, 9.6]; lag
0-2 days) in a study of pediatric asthma ED visits in Atlanta, GA. Additional support for
these seasonal differences in associations was presented by Orazzo et al. (2009). who
focused on wheeze ED visits in six Italian cities, where associations were slightly larger
in the summer compared to the winter, but the confidence intervals were wide and
overlapping (quantitative results not provided). In the study of seven Canadian cities,
Stieb etal. (2009) also conducted seasonal analyses but did not present detailed results.
However, the authors did state that there was no evidence of consistent associations
during the winter months (October-March) between any pollutant and respiratory
outcomes, including asthma.
Additional evidence for potential seasonal differences in NO2-associations with asthma
hospital admissions and ED visits comes from the analysis of asthma physician visits by
Sinclair et al. (2010). When focusing on asthma in children, the authors reported larger
risk estimates in the warm season at all lags for the 25-month period (e.g., warm: 9.6%
[95% CI: -7.4, 30.0]; cold: 1.2% [95% CI: -12.4, 16.8] at lag 0-2 days for a 30-ppb
increase in 1-h maxNO2 concentrations), with less consistent evidence for seasonal
differences in the 28-month period.
Concentration-Response Relationship
To date, few studies have examined the concentration-response (C-R) relationship
between NO2 exposures and respiratory morbidity. In recent studies, Strickland et al.
(2010) and Li et al. (20 lib) examined the shape of the NO2-pediatric asthma ED visit
relationship using different analytical approaches. Strickland et al. (2010) examined the
C-R relationship by conducting quintile and locally weighted scatterplot smoothing
5-86
-------
(LOESS) C-R analyses. In the quintile analysis, NO2 associations were positive and
stronger at quintiles representing higher concentrations, ranging from 28 ppb to
>181 ppb, relative to the first quintile (i.e., NCh concentrations <28 ppb). Additionally,
the LOESS C-R relationship analysis provides evidence indicating elevated NO2
associations along the distribution of concentrations from the 5th to 95th percentile
(Figure 5-5). Collectively, these analyses do not provide evidence of a threshold.
in
(N
in
o
'•*-
m
ro in
o: p H
m
O) -|
d
Nitrogen Dioxide Warm Season
15 20 25 30
Concentration (ppb)
35
Note: ppb = parts per billion. Solid line = locally weighted scatterplot smoothing concentration-response estimates. Dashed
lines = twice-standard error estimates. Results are from generalized additive models. Results are presented for the 5th to 95th
percentiles of nitrogen dioxide concentrations.
Source: Reprinted with permission of the American Thoracic Society, (Strickland et al.. 2010).
Figure 5-5 Concentration-response function for the association between
3-day average (lag 0-2) nitrogen dioxide concentrations and
emergency department visits for pediatric asthma in the Atlanta,
GA area.
In a study conducted in Detroit, MI, Li et al. (20lib) focused on the C-R relationship by
examining whether there is evidence of a deviation from linearity. Associations were
5-87
-------
examined in both a time-series and time-stratified case-crossover study design assuming:
(1) no deviation from linearity and (2) a change in linearity at 23 ppb [i.e., the maximum
likelihood estimate within the 10th to 95th percentile concentration where a change in
linearity may occur (~80th percentile)]. The analysis assumed a deviation in linearity but
did not assume zero risk below the inflection point. The focus of the analysis was on
identifying whether risk increased above the risk observed in the linear models at NO2
concentrations above 23 ppb. In the analyses assuming linearity, effect estimates varied
across models fora 0-4-day lag (time series: 2.9% [95% CI: -7.9, 15.1]; case-crossover:
9.1% [95% CI: -0.83, 20.2] for a 20-ppb increase in 24-h avg NO2 concentrations). In the
models that assumed a deviation from linearity, the authors did not observe evidence of
higher risk in either the time-series or case-crossover analyses at NCh concentrations
greater than 23 ppb.
Exposure Assignment
Questions often arise in air pollution epidemiologic studies with regard to the method
used to assign exposure. Strickland et al. (2011) assessed this question in a study
conducted in Atlanta, GA focusing on pediatric asthma ED visits. Using data from the
warm season from a previous analysis (Strickland et al.. 2010). Strickland et al. (2011)
examined the relative influence of different exposure assignment approaches (i.e., central
site monitor, unweighted average across available monitors, and population-weighted
average) on the magnitude and direction of associations between NO2 and pediatric
asthma hospital admission. Strickland et al. (2011) reported that effect estimates per IQR
increase in NC>2 were similar across the metrics; however, based on a standardized
increment, the magnitude of the association between NO2 and pediatric asthma ED visits
varied (central site monitor: 7.9% [95% CI: 4.2, 11.8] < unweighted average: 12.1%
[95% CI: 6.7, 17.9] 2 concentrations at lag 0-2 days). Although Strickland et
al. (2011) represents one study in one location, the results suggest that the different
approaches used to assign exposure across the studies evaluated may alter the magnitude,
but not direction, of the associations observed.
Nitrogen Dioxide within the Multipollutant Mixture
Another important question often encountered during the review of any criteria air
pollutant is whether the pollutant has an independent effect on human health. In the case
of NC>2, this is questioned because NO2 is often found to be highly correlated with other
traffic-related pollutants. However, ambient exposures to criteria air pollutants are in the
form of mixtures, which make answering this question difficult and primarily limited to
5-88
-------
examining copollutant models. Recent studies conducted by Gass et al. (2014) and
Winquist et al. (2014). both of which use pediatric asthma ED visit data from Atlanta,
GA, use novel approaches to assess whether specific mixtures are more strongly
associated with health effects compared to others. Although the primary objective of
these types of studies is not to directly assess the independent effects of a pollutant, the
studies can inform the role of NC>2 in the air pollution mixture.
Gass etal. (2014) used a classification and regression tree (C&RT) approach to examine
the association between short-term exposures to unique daily multipollutant mixtures of
NC>2, CO, PM25, and Os, and pediatric (i.e., ages 2-18 years) asthma ED visits in Atlanta.
C&RT is a supervised learning approach that creates various groupings of pollutants
based on an outcome variable, which differs from similar techniques, such as principal
component analysis, that do not consider the outcome (Gass etal.. 2014). For this
approach, daily pollutant concentrations were divided into quartiles with the referent
group comprised of all days in which each pollutant was in the lowest quartile. The
C&RT analysis identified 13 different unique daily pollutant combinations or terminal
nodes. Similar to Strickland et al. (2010). Gass etal. (2014) examined the relationship
between each combination and pediatric asthma ED visits using a Poisson model in the
context of a time-referent case-crossover analysis. Of the 13 unique combinations, 5 of
the largest relative risks (RRs) (i.e., RR ranging from 1.05 to 1.08) were observed for
combinations where NO2 concentrations were in the 3rd or 4th quartile. Of note for three
of the five combinations with the largest RRs, PIVb 5 concentrations were also high, with
concentrations in the 4th quartile. However, the RR largest in magnitude was observed
for a combination where NO2 concentrations were low (1st and 2nd quartiles) and PM2 5
concentrations were high (4th quartile). Overall, these results suggest that high daily
concentrations of NO2 in combination with low and high daily concentrations of PIVbs
can impact respiratory morbidity.
Winquist et al. (2014) took a different approach to examining multipollutant mixtures by
estimating the joint effect (i.e., the combined effect) of pollutants often associated with
specific air pollution sources. Associations between short-term NO2 exposures and
pediatric asthma ED visits (i.e., ages 5-17) were examined in single-pollutant models and
also in a multipollutant context in joint models for pollutant combinations representative
of oxidant gases (i.e., Os, NO2, 802), traffic (i.e., CO, NO2, EC), and criteria pollutants
(i.e., O3, CO, NO2, SO2, PM2 5). Using the model detailed in Strickland etal. (2010). the
authors reported results for an IQR increase for lag 0-2 days in single-pollutant analyses
as well as three types of joint effect models [i.e., no interaction terms (primary),
first-order multiplicative interactions between pollutants (interactions), and nonlinear
pollutant terms (nonlinear)] (Figure 5-6).
5-89
-------
1 3n
i 7Q
1 9O
O-I 1 C
1
OC i in
ft- 1 n^
1.00 -
n Q^
0.85
.."
«S
t
i
M
1
1
1
1
IE
Oxidanl
Gases
i
SPI
>;
• '
1.
! 1
«
Secondan
Comb
i. ,T..
J
1
m
1
SK
i
i
!
4
I
N
!*
X
Traffic
Combin
•II
Jn
TJ
Ǥ
SPf
i
1 »
i
•
Power Plan!
Coml
ina
ior
«
-1
t
T
3 8 | 8
1 1 1
in
i
1
'11
IE
Criteria Pollutants
lombi
Cation
Combination
.
i
1
't
1
i
•
i
1
1
J
••
" u r
till "i TT i T
I .. t
I |
M
l\'
1
Oxkhnt
Gases
ui
»
SK
!!i
•
Secondary
Combina
Combination
•i"|H §s n 8S*si fi
1 1 lil* 1 HI* s *
IK « iM * SK «
Traffic Power Plant Criteria Pollutants
ion Combination Combination Combination
Warm Season
Cold Season
Note: CO = carbon monoxide; EC = elemental carbon; h = hour; JE = joint model estimate; max = maximum; NO2 = nitrogen
dioxide; O3 = ozone; PM25 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; ppb = parts
per billion; SO2 = sulfur dioxide; SO4 = sulfate; SPE = single-pollutant model estimate. Results are presented for an interquartile
range increase in each pollutant. Interquartile range for nitrogen dioxide = 12.87 ppb.
Source: Reprinted with permission from Wolters Kluwer Health, Winquist et al. (2014).
Figure 5-6 Rate ratio and 95% confidence intervals for asthma-related
emergency department visits in single-pollutant and joint effect
models for each pollutant at lag 0-2 days.
Across pollutant combinations that contained NC>2, in the warm season, joint effect
models reported consistent positive associations with pediatric asthma ED visits. For each
pollutant combination the association observed was larger in magnitude than any
single-pollutant association, including NC>2, but not equivalent to the sum of each
individual pollutant association for a specific combination. Furthermore, in the warm
season analysis, associations across the different joint effects models were found to be
relatively similar. The results during the cold season were inconsistent; however, for the
combination of traffic pollutants, results from the joint effects models were relatively
similar to the single-pollutant results. The results of Winquist et al. (2014) suggest that
NO2 in combination with other pollutants is associated with asthma ED visits, but also
highlight the difficulty in separating out the independent effect of a pollutant that is part
of a mixture where multiple pollutants are often highly correlated.
5-90
-------
Summary of Asthma Hospital Admissions and Emergency Department
Visits
Recent studies that examined the association between short-term NO2 exposure and
asthma hospital admissions and ED visits consistently report positive associations and
support the results of U.S. and Canadian studies evaluated in the 2008 ISA for Oxides of
Nitrogen (Figure 5-7 and Table 5-12). Across asthma hospital admission and ED visit
studies, there was some evidence of a different pattern of associations for each outcome,
with more immediate effects (i.e., lag 0) for asthma hospital admissions and evidence of
prolonged effects for asthma ED visits, with a number of studies showing effects at
multiday lags ranging from 0-2 to 0-4 days. The studies that examined potential
confounding by PM2 5 or the traffic-related pollutants UFP or CO showed evidence that
associations between short-term NO2 exposures and asthma hospital admissions and ED
visits remained relatively unchanged in copollutant models with (i.e., similar in
magnitude or attenuated slightly, but remaining positive). NO2 is often found to be highly
correlated with these copollutants; therefore, the ability to determine whether short-term
NO2 exposures are independently associated with asthma hospital admissions and ED
visits is limited. Recent studies of multipollutant exposures further inform the effect of
short-term NO2 exposures on respiratory morbidity, specifically asthma. These studies
demonstrate that high daily concentrations of NO2 in combination with high daily
concentrations of other pollutants, such as PM2 5, can impact respiratory morbidity and
that associations are observed between asthma ED visits and NO2 in combination with
other traffic-related pollutants, oxidants, and criteria pollutants.
A number of recent studies also examined whether there was evidence that the
association between short-term NO2 exposures and asthma hospital admissions and ED
visits was modified by season or some other individual- or population-level factor
(Chapter 7). An examination of seasonal differences in NO2-asthma hospital admission
and ED visit associations provide some evidence of NO2 effects being larger in
magnitude in the summer or warm season, and that seasonal associations may vary by
geographic location. Studies of individual- and population-level factors, provide evidence
of differences in associations by lifestage, with larger NO2 effects for children and older
adults, and more limited evidence for differences by sex, race/ethnicity, and
socioeconomic status (SES), specifically insurance status (Chapter 7). Additionally, there
is evidence that exposure differences, specifically whether a population lives in housing
with low or high air exchange rates, may influence the association between short-term
NOx exposures and asthma ED visits.
Some recent studies also examined various study design issues, including model
specification and exposure assignment. An examination of model specification, as
detailed in Section 5.2.6. indicates that the relationship between short-term NO2
5-91
-------
exposures and respiratory-related hospital admissions, including those for asthma and
allergic disease, are sensitive to using less than 6 degrees of freedom (df) per year to
account for temporal trends, but robust to alternative lags and df, ranging from 3 to 6, for
weather covariates (Son et al.. 2013). An examination of various exposure assignment
approaches including single central site, average of multiple monitors, and
population-weighted average, suggests that each approach can influence the magnitude,
but not direction, of the NCVasthma ED-visit risk estimate (Strickland et al., 2011).
Finally, a few recent studies examined whether the shape of the NCh-asthma ED visit
relationship is linear or provides evidence of a threshold. These studies provide evidence
of a linear, no-threshold relationship between short-term NC>2 exposures and asthma ED
visits (Lietal..2011b: Strickland etal. 2010).
5-92
-------
Study
Burnett etal. (1999)
Sonetal. (2013)a
Ko etal. (2007)
Samoli etal. (2011)a
Location
Toronto, CAN
8 South Korean cities
Hong Kong
Athens, Greece
Age Lag
All
All
All
0-14
Iskandaretal. (2012) Copenhagen, Denmark 0-18
Linn etal. (2000)
Son etal. (2013)a,b
Tolbertetal. (2000)
Peel etal. (2005)
Sarnateta). (2013)d
Itoetal. (2007)
ATSDR (2006)
ATSDR 2006
Stieb etal. (2009)
Jalaludinetal. (2007)
Peel etal. (2005)
Li etal. (2011)
Villeneuve etal. (2007)
Strickland etal. (2010)
Jaffe etal. (2003)
Los Angeles, CA <30
8 South Korean cities All
Atlanta, GA All
Atlanta, GA All
Atlanta, GA All
New York, NY All
Bronx, NY All
Manhattan, NY All
7 Canadian cities All
Sydney, Australia 1-14
Atlanta, GA 2-18
Detroit, Ml 2-18
Edmonton, Canada 2+
Atlanta, GA 5-17
2 Ohio cities 5-34
0
0
0
0
0-4
0
0
0
0-4c
0-4d
0
0
0
0
1
0-2
0-2
0-1
0-4
0-4
2
0-1
0
0
0-2
0-4e
0-4f
0-4
0-2
Hospital Admissions
62 (95% Cl: 25,107)-
o
ED Visits
-e-
-5.0
5.0
15.0 25.0
% Increase (95% Cl)
35.0
Note: CA = California; Cl = confidence interval; ED = emergency department; GA = Georgia; Ml = Michigan; NO2 = nitrogen dioxide; NOX = sum of NO2 and nitric oxide; NY = New
York. Black = studies from the 2008 ISA for Oxides of Nitrogen, red = recent studies. Circles = NO2, triangles = NOX, solid symbols = all year, horizontal stripes = warm/summer
months, vertical stripes = cool/winter months. Results are standardized to a 20-ppb increase in 24-h avg NO2, a 30-ppb increase in 1-h max NO2, a 40-ppb increase in 24-h avg NOX,
and a 100-ppb increase in 1-h max NOX. a = results were presented for four seasons; however the summer and winter estimates represented the largest and smallest estimates
across seasons; b = this estimate is for allergic disease, which includes asthma; c = risk estimate for NO2; d = risk estimate for NOX; e = time-series results; f = case-crossover results.
Figure 5-7 Percentage increase in asthma hospital admissions and emergency department visits in relation
to short-term increases in ambient nitrogen dioxide concentrations.
5-93
-------
Table 5-12 Corresponding risk estimates for studies presented in Figure 5-7.
Study Location Age
Avg Time Season
% Increase
Lag (95% Cl)
Hospital Admissions
Burnett et al. (1999) Toronto, ON, Canada All
tSon etal. (2013) 8 South Korean cities All
Ko et al. (2007b) Hong Kong, China All
tSamoli etal. (2011) Athens, Greece 0-14
flskandar et al. Copenhagen, Denmark 0-18
(2012)
Linn etal. (2000) Los Angeles, CA <30
tSon etal. (2013)c'd 8 South Korean cities All
24-h avg All
24-h avg All
Summer
Winter
24-h avg All
1-h max All
Summer
Winter
24-h avg All
24-h avg All
24-h avg All
Summer
Winter
0 2.6(0.5,4.9)
0 3.6(0.5,6.8)
16.2(5.1,28.6)
-1.1 (-6.5,4.5)
0-4 10.9(8.1,13.8)
0 6.4 (-3.8, 17.6)
28.7 (-3.4, 71.3)
12.9 (-6.6, 36.5)
0-4 34.0(13.0, 58.0)a
62.0(25.0, 107)b
0 2.8(0.8,4.9)
0 3.8(1.0,6.6)
15.9(4.6,28.4)
-0.3 (-5.4, 5.1)
ED Visits
Tolbert et al. (2000) Atlanta, GA All
Peel et al. (2005) Atlanta, GA All
tSarnatetal. (201 3a) Atlanta, GA All
Ito et al. (2007) New York, NY All
ATSDR (2006) Bronx, NY All
ATSDR (2006) Manhattan, NY All
tStieb et al. (2009) 7 Canadian cities All
1-h max All
1-h max All
24-h avg All
24-h avg All
24-h avg All
24-h avg All
24-h avg All
1 2.4 (-2.6, 8.0)b
0-2 2.1 (-0.4, 4.5)
0-2 1.3(0.0, 2.4)b
0-1 12.0(7.0,15.0)
0-4 6.0(1.0,10.0)
OA ^ n ( 1 ft n *\A n\
2 0.0 (-2.6, 2.7)
5-94
-------
Table 5-12 (Continued): Corresponding risk estimates for studies presented in
Figure 5-7.
Study
fJalaludin et al.
(2008)
Peel et al. (2005)
tLietal. (201 1b)
fVilleneuve et al.
(2007)
fStrickland et al.
(2010)
Jaffe et al. (2003)
Location Age Avg Time Season
Sydney, Australia 1-14 1-h max All
Warm
Cold
Atlanta, GA 2-18 1-h max All
Detroit, Ml 2-18 24-h avg All
All
Edmonton, AB, Canada 2+ 24-h avg All
Warm
Cold
Atlanta, GA 5-17 1-h max All
Warm
Cold
2 Ohio cities 5-34 24-h avg Summer
Lag
0-1
0
0
0-2
0-4e
0-4f
0-4
0-2
1
% Increase
(95% Cl)
7.8(4.5, 11.1)
8.4(4.2, 12.5)
4.4 (-1.7, 10.4)
4.1 (0.8, 7.6)
2.9 (-7.9, 15.1)
9.1 (-0.8, 20.2)
4.5(0.0, 7.5)
21.4(13.6, 31.0)
-2.9 (-7.3, 1.5)
8.6(4.2, 13.3)
16.0(9.1,23.5)
3.8 (-1.9, 9.6)
6.1 (-2.0, 14.0)
AB = Alberta; avg = average; ATSDR = Agency for Toxic Substances and Disease Registry; CA = California; Cl = confidence
interval; ED = emergency department; GA = Georgia; max = maximum; Ml = Michigan; NY = New York; ON = Ontario.
aRisk estimate for NO2.
"Risk estimate for NOX.
°Results were presented for four seasons; the summer and winter estimates represented the largest and smallest estimates for
each season.
dEstimate for allergic disease, which includes asthma.
eTime-series analysis results.
'Case-crossover analysis results.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5.2.2.5 Subclinical Effects Underlying Asthma Exacerbation: Pulmonary Inflammation
and Oxidative Stress
The evidence described in the preceding sections for NCh-related increases in airway
responsiveness (Section 5.2.2.1). decreases in lung function and increases in respiratory
symptoms in children with asthma (Sections 5.2.2.2 and 5.2.2.3). and asthma hospital
admissions and ED visits (Section 5.2.2.4) is coherent and consistent with a sequence of
key events by which NCh can plausibly lead to asthma exacerbation. Adding to the
5-95
-------
proposed mode of action is evidence indicating NO2 exposure-related pulmonary
inflammation, a key early event in asthma exacerbation that can mediate increases in
airway responsiveness (Section 4.3.2.5). The initiation of inflammation by NC>2 exposure
is supported by observations of NCh-induced increases in eicosanoids, which mediate
recruitment of neutrophils (Section 4.3.2.3). The 2008 ISA for Oxides of Nitrogen
described evidence for NCh-induced increases in pulmonary inflammation in some
controlled human exposure studies and animal toxicological studies (U.S. EPA. 2008c).
with generally consistent findings for increased allergic inflammation in humans. Such
findings were coherent with results from the few available epidemiologic studies in
children with asthma indicating associations between short-term increases in ambient
NO2 concentrations and increases in exhaled nitric oxide (eNO). Other potential early
events linking NO2 exposure to asthma exacerbation are increases in reactive oxygen
species (ROS) and reactive nitrogen species. Many transcription factors regulating
expression of pro-inflammatory cytokines are redox sensitive, and reactive species also
can induce pulmonary injury (Section 4.3.2.1). Most information on the effects of NO2 on
pulmonary oxidative stress and injury is in healthy people and animal models, and
findings are inconsistent at ambient-relevant concentrations (Section 5.2.7.4). The key
evidence providing insight into a potential mode of action for NO2-induced asthma
exacerbation continues to be the findings from previous experimental studies for allergic
inflammation and the recent epidemiologic studies that continue to find NO2-associated
increases in pulmonary inflammation and oxidative stress. Biological indicators of
pulmonary inflammation and oxidative stress included those measured in exhaled breath;
bronchoalveolar, bronchial, and nasal lavage fluid; and sputum. Indicators of systemic
inflammation in blood are evaluated in the context of cardiovascular effects in
Section 5.3.
Experimental Studies
As described in Section 5.2.2.1. controlled human exposure studies in adults with asthma
and allergy demonstrated increases in airway responsiveness in response to NO2 exposure
with or without allergen challenge. These observations are supported by findings in
experimental studies involving adults with asthma and allergy and in a rat model of
allergic airway disease that NO2 exposure with or without an allergen challenge resulted
in increased indicators of allergic inflammation. The indicators include increases in IgE
and the influx and/or activation of eosinophils and neutrophils. Results provide evidence
that NO2 exposure can lead to exacerbation of allergic airways disease (discussed below
and in Section 4.3.2.6) and also provide support for epidemiologic evidence of
NO2~associated increases in inflammation in children with asthma and allergy.
5-96
-------
The 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) described several studies that
examined inflammatory responses in adults with mild allergic asthma who were exposed
to NO2 followed by a specific allergen challenge (Table 5-13). In a series of studies from
the Karolinska Institute in Sweden, adults at rest were exposed to air or 260 ppb NO2 for
15-30 minutes followed by an antigen (birch or timothy pollen) challenge 4 hours later.
BAL and bronchial wash fluids were collected 19 hours after the allergen challenge. NO2
exposure for 30 minutes increased polymorphonuclear cells (PMN) in the BAL and
bronchial wash fluids and increased ECP in the bronchial wash fluid compared with air
exposure (Barck etal.. 2002). Reduced cell viability of BAL cells and reduced volume of
BAL fluid were also reported. ECP is released by activated eosinophils; it is toxic to
respiratory epithelial cells and thought to play a role in the pathogenesis of airway injury
in asthma. In a subsequent study, Barck et al. (2005a) exposed adults with mild allergic
asthma to air or NCh for 15 minutes on Day 1 and twice on Day 2, and for 15 minutes
with allergen challenges following all of the exposures. NO2 exposure induced an
increased level of ECP in both sputum and blood and increased myeloperoxidase levels
in blood. These results suggest that NO2 may prime circulating eosinophils and enhance
activation of airway eosinophils and neutrophils in response to an inhaled allergen. Nasal
responses to nasal allergen challenge were also examined following a 30-minute
exposure to NCh (Barck et al., 2005b). No enhancement of nasal allergen responses was
observed in adult subjects. As noted in the 2008 ISA for Oxides of Nitrogen (U.S. EPA.
2008c), these studies indicate that brief exposures to 260 ppb NO2 can enhance allergen
responsiveness in individuals with asthma.
5-97
-------
Table 5-13 Characteristics of controlled human exposure studies of pulmonary
inflammation in populations with asthma.
Study
Disease Status;
Sample Size;
Sex; Age
(mean ± SD)
Exposure Details
Endpoints Examined
Barck et al. (2002)
Adults with mild
asthma and
allergy to birch or
timothy pollen;
n = 6 M, 7 F; 29 yr
Histamine inhalation test to confirm
airway hyperresponsiveness.
266 ppb NO2for30 min
Inhaled allergen challenge 4 h after
pollutant exposure.
Albumin in serum samples.
BW and BAL cell
parameters-volume recovered, cell
viability, total cell counts,
macrophage concentrations, % of
neutrophils, # eosinophils, # mast
cells (performed 19 h after allergen
challenge).
ECP, MPO, IL-5, IL-8, eotaxin,
ICAM-1.
Barck et al.
(2005a)
Adults with mild
asthma and
allergy to birch or
timothy pollen;
n = 10 M, 8F;
32 yr
260 ppb NO2
Day 1: one 15 min exposure with
bronchial challenge 4 h after
exposure.
Day 2: two 15 min exposures with
bronchial challenge 3 h after 2nd
exposure.
Total and differential cells counts of
induced sputum and venous blood
(samples taken on morning of
Days 1-3).
ECP, MPO in sputum.
Barck et al. Adults with rhinitis Seasonal allergy confirmed by
(2005b) and mild asthma; positive nasal challenge of allergen.
AHR confirmed by histamine test.
260 ppb NO2
Nasal allergen challenge 4 h after
exposure.
n = 9 M, 7 F;
31 Vn
Total and differential cell counts and
cell viability in NAL (performed
before exposure, before allergen
challenge, and 1 h, 4 h, and 18 h
after challenge).
ECP and MPO in NAL fluid and
blood.
Wang et al.
(1995a); Wanqet
al. (1995b)
Adults with
seasonal rhinitis;
n=6M, 10 F;
26 yr;
Nasal provocation with grass pollen Nasal lavage for inflammatory
allergen to confirm increase in nasal mediators fluid-ECP, MCT, MPO,
airway resistance. IL-8 (30 min after allergen
(1)400ppbN02for6h challenge).
(2) 400 ppb NO2 for 6 h + allergen
challenge
Wang et al. (1999) Adults with grass Nasal airway resistance tests at rest, NAL—total and differential cell
allergy;
n = 8 M, 8 F; 32 yr
after saline, and after allergen
challenge to confirm reactivity for
inclusion in study.
(1) 200 ug Fluticasone propionate
(FP) + 400ppbNO2for6h
(2) Matched placebo + 400 ppb NO2
for6h
counts (30 min after allergen
challenge).
Immunoassay of NAL fluid-ECP,
RANTES.
5-98
-------
Table 5-13 (Continued): Characteristics of controlled human exposure studies of
pulmonary inflammation in populations with asthma
Study
Disease Status;
Sample Size;
Sex; Age
(mean ± SD)
Exposure Details
Endpoints Examined
Witten et al.
(2005)
Adults with
asthma and house
dust mite allergy;
n = 6 M, 9 F; 32 yr
Inhaled allergen challenge to
determine predicted allergen PC20.
400 ppb NO2 for 3 h w/intermittent
exercise
2nd inhaled allergen challenge,
starting at 4 doubling doses less
than APC20 and doubling until 20%
decrease in FEV-i.
Total and differential cell counts in
induced sputum—macrophages,
lymphocytes, neutrophils, and
eosinophils (samples taken at 6 and
26 h after allergen challenge).
fRiedl et al.
(2012)
Vaqaqqini et al.
(1996)
Jorres et al.
(1995)
Phase 1: adults
with mild asthma;
n = 10 M, 5F;
0-7 .,r
61 yr
Phase 2: adults
with mild asthma
and cat allergy;
n = 6 M, 9 F; 36 yr
Asthma; n = 4 M,
4F;29± 14 yr
Healthy; n = 7 M;
34 ± 5 yr
Asthma; n = 8 M,
4 F; 27 ± 5 yr
Healthy; n = 5 M,
3 F; 27 yr
(range: 21-33)
Inhalation challenge to detect
bronchoconstrictive response
Phase 1: methacholine; Phase 2: cat
allergen).
(1 ) 1 00 ug/m3 DEP for 2 h with
intermittent exercise
(2) 350 ppb NO2 control for 2 h with
intermittent exercise
300 ppb for 1 h;
Exercise at VE = 25 L/min
1,000 ppb for 3 h;
Exercise 10 min on/10 min off at
individual's maximum workload.
Total counts and differential cell
counts (alveolar macrophages,
lymphocytes, PMNs, eosinophils) in
induced sputum (taken 22 h after
exposure).
Induced sputum fluid
assay-RANTES, eotaxin, ECP, IgG,
lgG4, IgA, IgM, IgE.
Cat-specific IL-4, IL-5, IL-8, IL-12,
GM-CSF, IFN-Y, TNF-a, tryptase.
Cell counts in sputum 2-h
post-exposure.
BAL fluid analysis 1 h after
exposure (cell counts, histamine,
prostaglandins).
AHR = airway hyperresponsiveness; BAL = bronchoalveolar lavage; BW = bronchial wash; DEP = diesel exhaust particles;
ECP = eosinophil cationic protein; F = female; FE\A = forced expiratory volume in 1 second; GM-CSF = granulocyte
macrophage-colony stimulating factor; h = hour; HDM = house dust mite; ICAM-1 = intercellular adhesion molecule 1;
IL = interleukin; L/min = liters per minute; M = male; min = minutes; MPO = myeloperoxidase; NAL = nasal lavage; NO2 = nitrogen
dioxide; PC = provocative concentration; PEF = peak expiratory flow; PMN = polymorphonuclear cells; yr = year.
fStudy published since the 2008 ISA for Oxides of Nitrogen.
Additional studies have been performed using longer NO2 exposures (Table 5-13). (Wang
etal. (1999); Wangetal. (1995a): Wangetal. (1995b)) found that exposure of adults to
400 ppb NO2 for 6 hours enhanced allergen responsiveness in the nasal mucosa in
subjects with allergic rhinitis. Mixed grass pollen was used as the challenge agent and
was administered immediately after the NO2 exposure. Responses included increased
numbers of eosinophils and increased levels of myeloperoxidase and ECP in nasal lavage
fluid collected 30 minutes after the allergen challenge. Witten et al. (2005) did not
observe enhanced airway inflammation with allergen challenge in adults with asthma and
5-99
-------
allergy to HDM allergen who were exposed to 400 ppb NCh for 3 hours with intermittent
exercise. HDM allergen was administered immediately after the NCh exposure and a
decrease in sputum eosinophils was found 6 hours later (Witten et al., 2005). Sputum
ECP levels were increased although this change did not reach statistical significance. The
authors suggested that their findings may be explained by a decreased transit of
eosinophils across the bronchial mucosa occurring concomitantly with NC^-induced
eosinophilic activation. Other investigators have noted that numbers of eosinophils do not
always correlate with allergic disease activity (Erjefalt et al.. 1999). Airway mucosal
eosinophilia is a characteristic feature of asthma and rhinitis; eosinophils exert their
effects via degranulation or cytolysis resulting in release of ECP and other mediators.
However, under conditions favoring eosinophil cytolysis, ECP concentrations may be
high and numbers of eosinophils may be low.
As noted in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). differing findings
between the studies in allergic individuals could be due to differences in timing of the
allergen challenge, the use of multiple- or single-allergen challenges, the use of BAL
fluid versus sputum versus nasal lavage fluid, exercise versus rest during exposure,
differences in subjects, or chance. Furthermore, study protocols varied in the timing of
biological sample collection post-exposure to NO2 or allergen. A recent study of adults
with mild allergic asthma also did not find enhanced airway inflammatory responses
following exposure to NO2 (350 ppb NO2,2 hours, intermittent exercise) [(Riedl et al.,
2012); Table 5-14]. Subjects exposed to NO2 followed by methacholine challenge
1.5 hours later had increased levels of blood IgM and decreased levels of sputum IgG4,
interleukin (IL)~4, eotaxin, RANTES, and fibrinogen measured 22 hours after exposure.
Subjects exposed to NO2 followed by cat allergen 1.5 hours later did not exhibit changes
in sputum cell counts measured 22 hours after exposure. While these results are not
consistent with NO2 enhancing airway inflammatory responses, importantly, markers of
eosinophil activation were not measured.
Several other studies investigated allergic inflammation following NO2 exposure in the
absence of a challenge. Vagaggini et al. (1996) observed a decrease in eosinophils in
sputum collected from adults with asthma following a 1-hour exposure to 300 ppb NO2,
though this decrease was not statistically significant. In contrast, a recent controlled
human exposure study reported an increase in eosinophils and ECP following repeated
NO2 exposure in adults with atopic asthma Ezrattyetal. (2014). Subjects were exposed to
203 or 581 ppb NO2 for 30 minutes on one day and twice for 30 minutes on the second
day. Compared with baseline, statistically significant increases in the amount of ECP and
the number and percentage of eosinophils in sputum were observed after the three
exposures to 600, but not 200 ppb NO2. Furthermore, ECP was highly correlated with
eosinophil count in sputum. No increases in either of these parameters were observed
5-100
-------
6 hours after the first exposure to 600 ppb NC>2. Torres et al. (1995) exposed healthy
adults and those with asthma and allergy to 1,000 ppb NO2 for 3 hours and performed
bronchoscopy 1 hour later. The macroscopic appearance of the bronchial epithelium was
altered after exposure in adults with asthma compared to healthy controls; however, no
accompanying changes in cell counts in the BAL fluid were observed. Eicosanoid levels
were also measured; thromboxane B2 was increased in healthy adults and those with
asthma following NC>2 exposure while prostaglandin D2 was increased and 6-keto
prostaglandin Fla was decreased after exposure only in adults with asthma. Because
eicosanoids are known mediators of inflammation, these results suggest that exposure to
NO2 resulted in activation of cell signaling pathways associated with inflammation.
Allergic inflammatory responses were also investigated in animal models of allergic
airways disease (Table 5-14). These studies involved sensitization and challenge with an
antigen followed by exposure to NCh. In several studies in mice, which were sensitized
and challenged with ovalbumin, NCh exposure over several hours or days failed to
increase allergic inflammatory responses. Exposures to 700 or 5,000 ppb NCh for 3 hours
on a single day, for 2 hours on 3 consecutive days or for 6 hours on 3 consecutive days
either reduced or had no effect on indicators of eosinophil inflammation such as
eosinophil counts, eosinophil peroxidase activity, and total cellularity (Poynteretal..
2006; Hubbard et al., 2002; Proust etal.. 2002). Other findings included decreases in IL-5
levels in the BAL fluid at both 24 and 72 hours after exposure to 5,000 ppb NO2 and
reductions in perivascular and peribronchial cellular infiltrates after exposure to 700 ppb
NO2. Others have noted that the ovalbumin-induced airway inflammation in mice does
not involve substantial eosinophil degranulation or cytolysis, which is characteristic of
asthma and allergic rhinitis in humans (Malm-Erjefalt et al.. 2001). This suggests that
species-related differences may account for NO2-induced decreases in eosinophilic
inflammation seen in mouse models. Mechanisms underlying the NO2~induced decrease
in airways eosinophilia are unknown.
In contrast with evidence in mice, NO2 exposure of rats, which were sensitized and
challenged with HDM allergen, enhanced specific immune responses and increased the
numbers of lymphocytes, neutrophils, and eosinophils in the airways, albeit with
exposure (3 hours) to 5,000 ppb NO2 (Gilmour et al., 1996). In this study, the most
pronounced responses occurred when rats were exposed to NO2 immediately after
sensitization and immediately after challenge with HDM antigen. Rats exposed to NO2
twice had increased levels of antigen-specific IgG and IgA and increased levels of IgE in
BAL fluid 7 days post-exposure to NO2. In addition, an increase in the ratio of
inflammatory cells (i.e., lymphocytes, neutrophils, eosinophils) to alveolar macrophages
was observed 7 days post-exposure to NO2, although the total number of lavagable cells
did not change.
5-101
-------
In summary, several controlled human exposure studies of adults with asthma and allergy
found that exposures to 260 ppb NO2 for 15-30 minutes or 400 ppb NC>2 for 6 hours
increased inflammatory responses to an allergen challenge. These responses included
increases in number and activation of eosinophils and neutrophils. In the absence of an
allergen challenge, repeated exposure to 600 ppb NCh for 30 minutes also enhanced
allergic inflammation in subjects with asthma and allergy. Other studies involving a
single exposure to NCh (300-350 ppb, 1-2 hours; 1,000 ppb, 3 hours) did not show these
responses. Allergic inflammation was also enhanced by a 3-hour exposure to 5,000 ppb
NO2 in a rat model of allergic airways disease, as demonstrated by increases in IgE levels
and numbers of eosinophils and neutrophils. These results provide evidence for
NC>2-induced exacerbation of allergic airways disease both in the presence and absence of
an allergen challenge (Section 4.3.2.6).
5-102
-------
Table 5-14 Characteristics of animal toxicological studies of pulmonary
inflammation.
Study
Species (Strain);
Sample Size; Sex;
Age
Exposure Details
Endpoints Examined
Proust et al.
(2002)
Mice (BALB/c);
n = 5/group; M; 6-7
weeks;
Immunization with injection of 10 ug Endpoints examined 24 h after
OVA (Day 0 and Day 7)
Challenge with either 10 ug OVA or
saline control (Day 14)
Exposure following OVA/saline
challenge: 5,000 ppb NO2
Challenge to 0.1 M aerosol of
methacholine for 20 sec
exposure:
BAL fluid total and differential cell
counts
Eosinophil peroxidase activity
Immunoassay of IL-4, IL-5
Anti-OVA IgE and lgG1 in serum
Lung histology
Hubbard et al.
Mice (CB57BI/6);
n = NR; M/F; adult
Sensitization by weekly injections of
25 ug OVA for 3 weeks
Challenge with 20 mg/m3 OVA
aerosol for 1 h for 3 days or 10 days
Exposure following OVA aerosol
challenge:
(1)700ppbNO2for2h
(2) 5,000 ppb NO2 for 2 h
Total and differential cell counts
from lung lavage (24 h after
exposure)
Histology analysis (24 h after
exposure)
Povnter et al.
(2006)
Mice (C57BL/6); n,
sex, and age NR
Sensitization by 20 ug of OVA via i.p.
injections on Days 0 and 7
Challenge with OVA aerosol (1% in
phosphate buffered saline) for
30 min on Days 14-16
Exposures subsequent to OVA
challenge:
(1) 5,000 ppb NO2 for 6 h/day for 1,
3, 5 days
Select groups given 20-day recovery
period
Methacholine challenge (0, 3.125,
12.5, 50 mg/mL in aerosol)
Endpoints examined after last day
of exposure or after 20 day
recovery:
BAL fluid—total and differential cell
counts; LDH
Histopathology analysis
mRNA levels of Gob5, MucSAC,
Th2, dendritic cell chemokine
CCL20 and eotaxin-1
Gilmouret al. Rats (brown Immunization with 100 ug antigen
(1996) Norway); (D. farina and D.
n = 5/group; F; pteronyssinuss) + killed Bordetella
6 weeks pertussis in 0.3 mL saline
Challenge with 50 ug allergen
(2 weeks after immunization),
followed by:
5,000 ppb NO2 for 3 h
Endpoints examined 7 days after
exposure:
Total and differential cell counts
from lung lavage
Antigen-specific IgG, IgA, IgE
antibodies in serum and lavage
fluid
Lymphocyte proliferation
responsiveness
BAL = bronchoalveolar lavage; F = female; h = hour; IL = interleukin; LDH = lactate dehydrogenase; M = male;
mRNA = messenger RNA; NO2 = nitrogen dioxide; NR = not reported; OVA = ovalbumin; Th2 = T-derived lymphocyte helper 2.
5-103
-------
Epidemiologic Studies of Populations with Asthma
The observations described in the preceding sections for NCh-induced increases in
allergic inflammation provide support for the epidemiologic associations observed for
ambient or personal NC>2 with increases in inflammation in children with asthma and
allergy. The limited evidence in adults with asthma is inconclusive. The number of these
epidemiologic studies of pulmonary inflammation has increased dramatically since the
2008 ISA for Oxides of Nitrogen, and recent studies expand on previous studies with
exposure assessment conducted in subjects' locations (e.g., homes, schools) and
additional examination of potential confounding by traffic-related copollutants. Ambient
NO2 concentrations, locations, and time periods for epidemiologic studies of pulmonary
inflammation and oxidative stress are presented in Table 5-15.
As in previous studies, the majority of evidence is for eNO. Across studies, eNO was
collected with a similar protocol, following the guidelines established by ATS (2000a).
eNO assessment methods also accounted for NO in the collection room, although eNO
has not been shown to be a reliable indicator of NO exposure (Section 4.2.3). eNO has
not been examined in controlled human exposure or animal toxicological studies of NO2
exposure, but several observations support the epidemiologic findings. NO2 exposure has
been shown to increase some pro-inflammatory cytokines and increase neutrophils and
eosinophils (Section 4.3.2.6). which can activate inducible nitric oxide synthase or
produce NO in the lung during an inflammatory response (Barnes and Liew. 1995).
Higher eNO has been associated with higher eosinophil counts (Brody et al.. 2013).
Further, eNO commonly is higher in children and adults with asthma and increases
during acute exacerbation (Soto-Ramos et al.. 2013; Carraro et al.. 2007; Jones et al..
2001; Kharitonov and Barnes. 2000).
5-104
-------
Table 5-15 Mean and upper percentile concentrations of nitrogen dioxide in
epidemiologic studies of pulmonary inflammation and oxidative
stress in populations with asthma.
Study3
tLiu et al.
(2009b)
tBarraza-
Villarreal et al.
(2008)
Delfino et al.
(2006)
fDelfino et al.
(2013)
fMartins et al.
(2012)
fSarnat et al.
(2012)
fGreenwald et al.
(2013)
tHolquin et al.
(2007)
tFlamant-Hulin
etal. (2010)
tLinetal. (2011)
tLiu et al.
(2014a)
tBerhane et al.
(2011)
Location
Windsor, ON,
Canada
Mexico City,
Mexico
Riverside, CA
Whittier, CA
Riverside, CA
Whittier, CA
2 sites combined
Viseu, Portugal
El Paso, TX and
Ciudad Juarez,
Mexico
El Paso, TX
Ciudad Juarez,
Mexico
Clermont-Ferrand,
France
Beijing, China
Munich and
Wesel, Germany
13 southern
California
communities
NO2 Metric
Study Period Analyzed
Oct-Dec 2005 24-h avg NO2
Jun 2003-Jun 8-h max NO2
2005
Aug-Dec 2003 24-h avg NO2
Jul-Nov2004
8-h max NO2
24-h avg NO2
Jan and Jun, 1-week avg NO2b
2006 and 2007
Jan-Mar 2008 96-h avg NO2
Mar-Jun2010 96-h avg NO2
2001-2002 1-week avg NO2
NR 5-day avg NO2
Jun 2007 24-h avg NO2
Sep 2007
Dec 2007
Jun 2008
Sep 2008
NR 24-h avg NO2
Sep-Jun 24-h avg NO2
2004-2005
Mean
Concentration
ppb
19.8
37.4
Personal: 24.3
Personal: 30.9
Central site: 39.3
Central site: 35.1
Central site: 27.4
Across 4 periods:
4.5, 3.5, 9.8, 8.2C
El Paso school:
4.5, 14.2, central
sites: 14.0, 18.5,
20.5
Ciudad Juarez
school: 18.7,27.2,
central site: none
School A: 6.5
School B: 17.5
18.2
Schools <14 ppb:
10.1
Schools >14 ppb:
17.4
24.3
30.4
45.3
26.6
25.9
15.9C
NR
Upper Percentile
Concentrations
ppb
95th: 29.5
Max: 77.6
Max: 47.6
Max: 106
Max: 72.4
Max: 96
Max: 73.8
Max across 4
periods: 4.6, 4.0,
10.9, 9.4C
NR
NR
NR
Across schools:
75th: 14.0C
Max: 19.7C
NR
NR
NR
NR
NR
95th: 29.7C
NR
5-105
-------
Table 5-15 (Continued): Mean and upper percentile concentrations of nitrogen
dioxide in epidemiologic studies of pulmonary
inflammation and oxidative stress in populations with
asthma.
Study3
fRomieu et al.
(2008)
fQian et al.
(2009a)
fMaestrelli et al.
(2011)
Location
Mexico City,
Mexico
Boston, MA; New
York City, NY;
Philadelphia, PA;
Madison, Wl;
Denver, CO; San
Francisco, CA
Padua, Italy
Study Period
Jan-Oct 2004
Feb 1997-Jan
1999
1999-2003
NO2 Metric
Analyzed
8-h max NO2
24-h avg NO2
24-h avg NO2
Mean
Concentration
ppb
35.3
23.6
Range across
seasons and
years: 20. 9-37. Oc
Upper Percentile
Concentrations
ppb
Max: 73.5
75th: 28.8
Max: 48.1
Range of 75th:
oo n /o ^c
Avg = average; Aug = August; CA = California; CO = Colorado; Dec = December; Feb = February; MA = Massachusetts;
NO2 = nitrogen dioxide; NR = not reported; NY = New York; ON = Ontario; PA = Pennsylvania; TX = Texas; Wl Wisconsin.
aStudies presented in order of first appearance in the text of this section.
"•Subject-level exposure estimates calculated from outdoor NO2 at schools and other locations plus time-activity patterns.
°Concentrations converted from |jg/m3 to ppb using the conversion factor of 0.532 for NO2 assuming standard temperature (25°C)
and pressure (1 atm).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
Children with Asthma
Several recent and previous studies found associations between short-term increases in
ambient NO2 concentration and increases in pulmonary inflammation in children with
asthma. Children were recruited mostly from schools, supporting the likelihood that study
populations were representative of the general population of children with asthma.
Asthma was identified as self- or parental report of physician-diagnosed asthma, but the
studies varied in whether they assessed asthma severity or required the presence of
current symptoms in subjects. Larger associations were found in children not using ICS
in some (Sarnat et al., 2012; Liu et al., 2009b), but not all, (Delfino et al., 2006) studies.
Because of the heterogeneity in the definition of ICS use and lack of assessment of ICS
compliance, it is not clear whether ICS use represents well-controlled or more severe
asthma across populations. Across studies, associations varied in magnitude and
statistical significance; however, the consistent pattern of increasing eNO with increasing
short-term NC>2 exposure provides evidence of an association (Figure 5-8 and
Table 5-17). Most studies analyzed multiple endpoints, pollutants, lags of exposure, or
subgroups; however, with a few exceptions (Liu et al.. 2009b; Barraza-Villarreal et al..
2008), a pattern of association was found across the multiple comparisons, thus reducing
the likelihood of associations found by chance alone or from publication bias.
5-106
-------
Study
NO2 Metrics
Analyzed
Children with Asthma
Delfino et al. (2006) 24-h avg
lag 0-1 day avg
Exposure Assessment Subgroup
Total personal
Delfino etal. (2013)
Martins et al. (2012)
Sarnat etal. (2012)
8-h max
lag 0-1 day avg
24-h avg
lag 0-1 day avg
24-h avg
lag 0-4 day avg
24-h avg
lag 0-4 day avg
Central sites
Central sites
Subject modeled
School
All subjects
No anti-inflamm medtra*
Anti-inflamm med use
ICS use
Anti-LT and ICS use
All subjects
Greenwald et al. (2013)
Linetal. (2011)
Liu et al. (2009)
Liu et al. (In press)
Barraza-Villarreal et al.
(2008)
Berhane et al. (2011)
Adults with Asthma
Qian et al. (2009)
24-h avg
lag 0-3 day avg
24-h avg
lag 0 day
24-h avg
lag 0 day
24-h avg
lag 0 day
8-h max
lag 0 day
24-h avg
lag 1-6 day avg
Central site
School
Central site
Central sites
Central site
Central site
Central site
All subjects
No ICS use
ICS use
All subjects
School A
School B
24-h avg
lag 0 day
Central sites
All subjects
Placebo
Beta-agonist use
ICS use
u
e
-10 -5 0 5 10 15 20 25 30 35 40 45 50 55
Percent change in eNO per increase in NO2 (95% Cl)a
Note: Anti-inflamm med = anti-inflammatory medication; Anti-LT = anti-leukotriene medication; avg = average; Cl = confidence
interval; eNO = exhaled nitric oxide; h = hour; ICS = inhaled corticosteroid; max = maximum; NO2 = nitrogen dioxide. Black = studies
from the 2008 Integrated Science Assessment for Oxides of Nitrogen, red = recent studies. Results from more informative studies in
terms of exposure assessment method and potential confounding considered are presented first. Study details and quantitative
results reported in Table 5-16. Table 5-16 presents results for an array of indications of inflammation and oxidative stress for which
there was not sufficient numbers to present in a figure. For some studies, eNO results could not be presented in the figure because
results were not reported in terms of percentage change eNO.
aEffect estimates are standardized to a 20-ppb increase for 24-h avg NO2 and 30-ppb increase for 1-h max NO2.
Figure 5-8 Associations of personal or ambient nitrogen dioxide with
exhaled nitric oxide in populations with asthma.
5-107
-------
Table 5-16 Epidemiologic studies of pulmonary inflammation and oxidative stress in children and adults with
asthma.
Study
Population Examined and Methodological Details
NO2 Metrics Effect Estimate (95% Cl)
Analyzed Lag Day Single-Pollutant Model3
Copollutant Examination
Children with asthma: studies with small spatial scale exposure assessment and/or examination of copollutant confounding
Delfino et al. (2006)
Riverside, Whittier, CA
n = 45, ages 9-18 yr, persistent asthma and exacerbation in
previous 12 mo
Repeated measures. Examined daily for 10 days,
372 observations. Recruitment in schools of nonsmokers from
nonsmoking homes. No information on participation rate. Self-
report of physician-diagnosed asthma. Mixed effects model
with random effect for subject with pollutant concentrations
centered on subject mean and adjusted for personal measures
of relative humidity, measures of personal temperature, follow-
up period. Adjustment for city, daily beta-agonist use, weekend
did not alter results.
NO2-total personal
24-h avg
Compliance
assessed with
motion detectors.
Monitoring checked
daily; all samples
above detection limit
of 2.1 ppb (Staimer
etal.,2005)
eNO:
All subjects: 1.2% (-2.0, 4.3)
0-1 avg All subjects: 7.5% (2.0, 13)
No anti-inflammatory medication,
n = 14: 2.6% (-9.9, 15)
Anti-inflammatory medication,
n = 31: 9.3% (3.1, 16)
ICSuse, n = 19: 7.0% (0.23, 14)
Anti-leukotrienes + ICS use, n = 12
9.1% (-3.7, 22)
NO2-central site 0 All subjects: 0.81% (-4.5, 6.1)
8-h max 0-1 avg All subjects: 11% (3.2, 19)
Copollutant model results
in figure only.
- With PM2.5, EC, or OC:
NO2 results robust but
. increase in 95% Cl.
Copollutant results robust
to NO2 adjustment.
. Weak correlations for
personal exposures.
Spearman r = 0.20-0.33.
Stronger correlations for
central site pollutants.
Pearson r= 0.25-0.70.
' Central site CO not
associated with eNO.
5-108
-------
Table 5-16 (Continued): Epidemiologic studies of pulmonary inflammation and oxidative stress in children and
adults with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics Effect Estimate (95% Cl)
Analyzed Lag Day Single-Pollutant Model3
Copollutant Examination
tPelfinoetal. (2013)
Riverside, Whittier, CA
Same population and methodology as Delfino et al. (2006)
above.
Analysis also indicated lack of confounding by respiratory
infections.
NO2-central site
24-h avg
1 site Riverside
within 12 km of
subjects' homes
2 sites Whittier
averaged, distance
NR
eNO:
0 -0.12% (-3.8, 3.7)
1 5.0% (1.2, 9.1)
0-1 avg 9.0% (2.9, 15)
For lag 0-1 avg:
With oxidative potential of
PlVh.s: 3.8% (-5.1, 14)
With in vitro ROS from
PM2.5: 5.8% (-1.9, 14)
Copollutant associations
attenuated with NO2
adjustment.
Moderate correlations with
NO2. Spearman r=0.43
for ROS, 0.49 for oxidative
potential.
tMartins (2013), Martins et al. (2012)
Viseu, Portugal
n = 51, mean age 7.3 (SD: 1.1)yr, 53% with atopy
Repeated measures. 4 measurements over 2 different
seasons. Recruitment from urban and suburban schools.
-66% participation rate. Parental report of wheeze in previous
12 mo. GEE adjusted forage, sex, parental smoking, parental
education, atopy, time of visit, average temperature, relative
humidity. Also included height, weight, older siblings,
mold/dampness in home, fireplace in home, pets in home
because their inclusion changed the effect estimate for at least
1 pollutant by >10%.
NO2-subject
modeled outdoor
24-h avg
Estimated from
school outdoor NO2,
20 city locations,
MM5/CHIMERE
modeling, and daily
activity patterns.
20% time spent at
school, 65% at
home.
0-4 avg eNO: 14% (-12, 40)
Exhaled breath condensate pH:
-2.6% (-3.9, -1.3)
For EEC pH only:
With PMio:
0.30 (-3.0, 3.6)
With benzene:
-1.7 (-3.6, 0.26)
With ethylbenzene:
-1.6 (-3.7, 0.49)
PMio robust to adjustment
for NO2. VOCs attenuated
to null. Negative or weakly
positive correlations with
NO2. Spearman r= -0.82
to -0.55 for PMio, -0.42 to
0.14 for various VOCs.
5-109
-------
Table 5-16 (Continued): Epidemiologic studies of pulmonary inflammation and oxidative stress in children and
adults with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics Effect Estimate (95% Cl)
Analyzed Lag Day Single-Pollutant Model3
Copollutant Examination
tSarnatetal. (2012)
El Paso, TX and Ciudad Suarez, Mexico
n = 29 per city, ages 6-12 yr, asthma and current symptoms
Repeated measures. Examined weekly for 16 weeks,
697 observations. Recruitment from schools representing a
gradient of traffic, subjects from nonsmoking homes. No
information on participation rate. Self-report of physician-
diagnosed asthma. GLM with subject as random effect and
adjustment for school, temperature, relative humidity, indoor
NO. Adjustment for medication use, cold symptoms did not
alter results. Most indoor samples above limit detection of 2.88
ppb (Ravsoni et al., 2011).
NO2-school outdoor
Each city: one
school 91 m from
major road, one in
residential area.
NO2-school indoor
NO2-central site
1 site in El Paso, TX
near major road.
All 24-h avg
eNO:
0-4 avg All subjects: 6.3% (2.5, 10)
With Os: 8.8% (4.6, 13)
No copollutant model with
PM2.5 or PM-io-2.5, which
No ICS use, n = 10: 6.6% (2.6, 11) were associated with
eNO. No association with
BC among all subjects.
Weak to moderate
correlations with NO2.
All subjects: 0.53% (0.11, 1.0) Spearman r=-0.39 to
0.32 for PM25; -0.24 to
All subjects: 1.7% (-1.0, 4.5) 0.04 for PMio-2.s.
ICSuse, n = 19: 1.1% (-8.9, 12)
tGreenwald et al. (2013)
El Paso, TX
n = 38, mean age 10 yr, 76% Mexican-American
Repeated measures. Examined weekly for 13 weeks,
436 observations. Recruitment from schools in low- and high-
traffic area. No information on participation rate. School record
of physician-diagnosed asthma. GLM with subject as random
effect and adjusted for school, temperature, relative humidity,
indoor NO. Most indoor samples above limit detection of 2.88
ppb (Ravsoni et al.. 2011).
NO2-school outdoor
School A: residential 0-3 avg
area, School B:
91 m from major
road.
NO2-school indoor
All 24-h avg
eNO:
School A: -0.86% (-38, 58)
School B: 30% (-3.1, 73)
School A: -16% (-53, 47)
SchoolB: 5.6% (-19, 37%)
No copollutant model.
BC, VOCs (central site)
associated with eNO.
Moderate correlations with
NO2. Pearson
• r= 0.47-0.62.
BTEX associated with
eNO. Highly correlated
with NO2. r=0.77.
tHolquin et al. (2007)
Ciudad Juarez, Mexico
n = 95, ages 6-12 yr, 78% mild asthma, 58% with atopy
Repeated measures. Examined biweekly for 4 mo. 87%
participation. Self-report of physician-diagnosed asthma.
Linear and nonlinear mixed effects model with random effect
for subject and school adjusted for sex, body mass index, day
of week, season, maternal and paternal education, passive
smoking exposure.
NO2-school outdoor
24-h avg
Schools located
239-692 m from
homes.
0-6 avg No quantitative results reported for
eNO. No association was reported.
No copollutant model.
Road density but not
PM2.5 or EC associated
with eNO.
5-110
-------
Table 5-16 (Continued): Epidemiologic studies of pulmonary inflammation and oxidative stress in children and
adults with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics
Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
tZhu(2013): Linetal. (2011)
Beijing, China
n = 36, ages 9-12 yr, 22% with asthma
Repeated measures before and after Olympics. Examined
daily for five 2-week periods. 1,581 observations. Recruitment
from school. Selection from 437 (60%) students who
responded to initial survey, 95% follow-up participation. GEE
adjusted for temperature, relative humidity, body mass index.
NO2-central site
24-h avg
Site 650 m from
school.
eNO:
0 All subjects: 22% (18, 26)
Asthma: 23% (16, 31)
1 Asthma: 12% (4.0, 20)
Among all subjects:
With BC: 5.6% (0.38, 11)
With PIvh.s: 14% (9.5, 19)
No change in BC with NO2
adjustment. PIvh.s reduced
but positive. NO2 highly
correlated with BC
(r= 0.68), moderately
correlated with PIvh.s
(r=0.30).
tFlamant-Hulin et al. (2010)
Clermont-Ferrand, France
n = 34, mean age: 10.7 (SD: 0.7) yr, 44% with atopy
Cross-sectional. Recruitment from schools. 69% participation
rate. Self- or parental-report of lifetime asthma. For some
subjects, eNO measured up to 1 week before pollutants. GEE
adjusted for atopy, mother's birth region, parental education,
family history of allergy, prenatal and childhood smoking
exposure. Did not consider potential confounding by weather.
NO2-school outdoor
24-h avg
NO2-school indoor
24-h avg
No information on
limit of detection
0-4 avg log eNO comparing >14.3 vs.
<14.3ppb NO2:
0(-0.14, 0.14)
0(-0.13, 0.14)
No copollutant model.
PIvh.s, acetylaldehyde
associated with eNO.
tLiu(2013). Liuetal. (2009b)
Windsor, ON, Canada
n = 182, ages 9-14 yr
Repeated measures. Examined weekly for 4 weeks, same day
of week. 672 observations. Recruitment from schools. No
information on participation rate. Parental report of physician-
diagnosed asthma. Mixed effect model with random effect for
subject and adjusted for testing period, temperature, relative
humidity, daily medication use.
NO2-central site
24-h avg
Average of 2 sites.
99% subjects live
within 10 km of
sites.
0
1
0-2 avg
0
1
0-2 avg
eNO:
17% (-5.8, 47)
7.7% (-12, 32)
1.5% (-32, 50)
TEARS:
48% (3.9, 111)
22% (-11, 67)
131% (23, 334)
For TEARS only:
with PIvh.s:
31% (-30, 145)
withSO2:43%(-10, 126)
Small decrease in PIvh.s
estimate with adjustment
for NO2. NO2 highly
correlated with PIvh.s
(Spearman r= 0.71),
weakly with SO2 (r = 0.18).
5-111
-------
Table 5-16 (Continued): Epidemiologic studies of pulmonary inflammation and oxidative stress in children and
adults with asthma.
Study
Population Examined and Methodological Details
tLiuetal. (2014a)
NO2 Metrics
Analyzed
NO2-central site
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
0 eNO:
Copollutant Examination
With PMio:
Munich and Wesel, Germany
n = 192, age 10 yr
Cross-sectional. Recruitment from GINIplus, LISAplus birth
cohort studies. No information reported on participation rate or
ascertainment. Parental report of physician-diagnosed asthma.
GAM adjusted for cohort, city, sex, parental education,
parental history of atopy, indoor gas pollution, current pets,
maternal prenatal smoking, smoking exposure at age 10 yr,
temperature. Results not altered by adjustment for asthma
medication use or annual avg NO2 estimated from land use
regression models.
24-h avg
1 site per city in
suburban locations.
Both cities: 51% (-11, 154)
Results in figure show association
only in Munich, null in Wesel.
23% (-37, 137) among
children with asthma.
PMio results not altered
with NO2 adjustment.
34% (15, 56) among all
1,985 children. PMio
association attenuated
with NO2 adjustment.
Moderate correlated with
NO2. Spearman r=0.59.
Children with asthma: studies with central site exposure assessment and no examination of copollutant confounding
tBarraza-Villarreal et al. (2008)
Mexico City, Mexico
n = 119-129, ages 6-14 yr, 54% persistent asthma, 89%
atopy
Repeated measures. Examined every 15 days for mean
22 weeks. 1,004 observations. Recruited from pediatric clinic.
Asthma severity assessed by pediatric allergist. No information
on participation rate. Linear mixed effects model with random
effect for subject and adjusted for sex, body mass index, lag
one minimum temperature, ICS use, time. Adjustment for
outdoor activities, smoking exposure, anti-allergy medication
use, season did not alter results.
NO2-central site
8-h max
Monitors within 5 km
of school or home.
Low correlation for
school vs. central
site: Spearman
r=0.21
0 eNO: 8.4% (7.9, 9.0)
lnterleukin-8: 1.2% (1.1, 1.3)
Exhaled breath condensate pH:
-0.5% (-1.5, 0.50)
No copollutant model.
' PM2.5 and Os associated
with eNO and IL-8.
' Moderate or weak
correlation with NO2.
Pearson r = 0.61 for PM2.5,
0.28 for O3.
5-112
-------
Table 5-16 (Continued): Epidemiologic studies of pulmonary inflammation and oxidative stress in children and
adults with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics Effect Estimate (95% Cl)
Analyzed Lag Day Single-Pollutant Model3
Copollutant Examination
tRomieu et al. (2008)
Mexico City, Mexico
n = 107, mean age 9.5 yr. 48% persistent asthma, 90% atopy
Repeated measures. EEC collected every 2 weeks for
2-16 weeks. 480 observations. Recruitment from allergy clinic.
No information on participation rate. 25% EEC samples below
detection limit, assigned random value 0-4.1 nmol. Malondi-
aldehyde associated with wheeze and asthma medication use.
GEE model adjusted for sex, school shift, temperature,
chronological time. Adjustment for outdoor activities, parental
smoking did not alter results.
NO2-central site
8-h max
Similar results for
1-h max and 24-
h avg.
Monitors within 5 km
of school or home.
Log malondialdehyde:
0.13 (-0.10, 0.35)
No copollutant model.
PM2.5, distance to closest
Avenue; 4.5-h traffic
count, and Os also
associated with
malondialdehyde.
Moderate correlation with
NO2. Pearson r = 0.44 for
Os and 0.54 for PM2.5.
tBerhane et al. (2011)
13 Southern CA towns
n = 169, ages 6-9 yr
Cross-sectional. Recruitment from schools. Parental report of
physician-diagnosed asthma and history of respiratory allergy.
Linear regression adjusted for community, race/ethnicity, age,
sex, asthma, asthma medication use, history of respiratory
allergy, eNO collection time, body mass index, smoking
exposure, parental education, questionnaire language,
season, multiple temperature metrics, eNO collected outdoors.
NO2-central site
24-h avg
Sites in each
community. # sites
in each community
NR.
1-6 avg eNO: -6.7% (-31 26%)
No copollutant model.
PM2.5, PMlO, Os
associated with eNO.
Moderate or weak
correlations with NO2.
Pearson r = 0.47 for PM2.5,
0.49 for PMio, 0.15forOs.
Adults with asthma: central site exposure assessment, no examination of potential confounding by traffic-related copollutants
tQian et al. (2009a)
Boston, MA; New York, NY; Denver, CO; Philadelphia, PA;
San Francisco, CA; Madison, Wl.
n = 119, ages 12-65 yr, persistent asthma, nonsmokers
Repeated measures. Examined every 2-4 weeks for
16 weeks. 480 person-days. No information on participation
rate. Study population representative of full cohort. Asthma
medication trial and a priori comparison of medication
regimens. Linear mixed effects model adjusted forage, sex,
race/ethnicity, center, season, week, daily average
temperature, daily average humidity. Adjustment for viral
infections did not alter results.
NO2-central site
24-h avg
Average of all
monitors within
51 km of subject
ZIP code centroid.
eNO: With PMio:
All subjects: 1.1% (0.52, 1.7) 0.69% (-0.09, 1.5)
Placebo: 0.79% (-0.08, 1.7) With °3: °-94% (°-43> 1 -5)
Beta-agonist use: 0.86% (0.08, 1.6) With s°2: 12% (°-52> 1-9)
ICS use: 1.8% (0.62, 2.9) Copollutant effect
estimates attenuated with
adjustment for NO2.
Correlations NR.
0-3 avg All subjects: 0.94% (0.09, 1.8)
5-113
-------
Table 5-16 (Continued): Epidemiologic studies of pulmonary inflammation and oxidative stress in children and
adults with asthma.
Study
Population Examined and Methodological Details
NO2 Metrics Effect Estimate (95% Cl)
Analyzed Lag Day Single-Pollutant Model3
Copollutant Examination
tMaestrelli et al. (2011)
Padua, Italy
n = 32, mean age 39.6 (SD: 7.5) yr, 81% persistent asthma
Repeated measures. Examined 6 times over 2 yr. Selected
from database of beta-agonist users (>6/yr for 3 yr), diagnosis
clinically confirmed. 76% follow-up participation. Dropouts did
not differ from participants. GEE adjusted for daily average
temperature, humidity, atmospheric pressure, asthma
medication use, current smoking status.
NO2-central site
24-h avg
2 sites in city
eNO (ppb):
All subjects: 3.1 (-14, 21)
Nonsmokers, n = 22: 2.9 (-20, 26)
Exhaled breath condensate pH:
All subjects: 0(-0.19, 0.21)
Nonsmokers: -0.09 (-0.24, 0.05)
No copollutant model.
Personal and central site
PM2.5 and PM-io not
. associated with eNO. No
associations with central
site CO. Association found
with 03 and SO2.
Correlations NR.
Note: More informative studies in terms of the exposure assessment method and potential confounding considered are presented first.
avg = average; BC = black carbon; BTEX = sum of the VOCs benzene, toluene, ethybenzene, xylene; Cl = confidence interval; CA = California; CO = carbon monoxide, Colorado;
EEC = exhaled breath condensate; EC = elemental carbon; eNO = exhaled nitric oxide; GAM = generalized additive models; GEE = generalized estimating equation;
GINIPIus = German Infant Nutritional Intervention plus environmental and genetic influences; GLM = generalized linear model; ICS = inhaled corticosteroid; IL = interleukin;
MA = Massachusetts; NO2 = nitrogen dioxide; NR = not reported; NY = New York; O3 = ozone; OC = organic carbon; ON = Ontario; PA = Pennsylvania; PM25 = particulate matter
with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PMio = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm;
ROS = reactive oxygen species; SD = standard deviation; TEARS = thiobarbituric acid reactive substances; TX = Texas; VOC = volatile organic compound; Wl = Wisconsin.
aEffect estimates are standardized to a 20 ppb increase for 24-h avg NO2 and 25 ppb increase for 8-h max NO2.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-114
-------
Key evidence is provided by studies with NO2 exposures assessed for subjects' locations,
comparison of various exposure metrics, and/or examination of confounding by
traffic-related copollutants. These studies examined a limited number of exposure lags
but specified them a priori. Across studies, associations were found with multiday
averages of NO2 (i.e., 0-1 avg to 0-6 avg) (Figure 5-8 and Table 5-16). with Delfino et
al. (2006) finding a stronger association of eNO with lag 0-1 avg than lag 0 or 1 day
NO2. As reported by a few studies, participation rates were high (87, 95%, Table 5-16).
Selective participation by certain groups was not indicated. Strong exposure assessment
was characterized as personal monitoring (Delfino et al.. 2006): estimation of individual
outdoor exposures based on monitoring, modeling, and daily activity patterns (Martins et
al.. 2012); monitoring at or near schools (Greenwald et al.. 2013: Sarnatetal.. 2012: Lin
et al.. 2011: Holguin et al.. 2007): or examination of central site ambient concentrations
that were temporally correlated with total personal NO2 measures (Delfino et al.. 2013).
In comparisons with central site NC>2, associations with eNO were similar to personal
NC>2 among children with asthma in Riverside and Whittier, CA [(Delfino et al.. 2013:
Delfino et al.. 2006): Figure 5-8 and Table 5-16]. An increase in 8-h max NO2 assigned
from each child's community central site was associated with a similar increase in eNO
as 24-h avg personal NO2 based on the interquartile ranges of NO2 (1.4% [95% CI: 0.39,
2.3] per 12-ppb increase in 8-h max central site NO2 and 1.6% [95% CI: 0.43, 2.8] per
17-ppb increase in 24-h avg personal NO2). Personal and central site NO2 were
moderately correlated (Spearman r = 0.43). Thus, despite the potential for greater
exposure measurement error due to within-community variability in ambient NO2
concentrations and variation in time-activity patterns (Section 3.4.4). daily variation in
ambient NO2 is to some extent represented in daily variation in personal NO2 exposures
of these children that is associated with eNO. Such results provide a rationale for drawing
inferences about ambient NO2 exposure from associations observed with total personal
NO2 exposures. All total personal NO2 exposures were above the LOD, supporting the
reliability of the associations with eNO (Delfino et al.. 2006: Staimer et al.. 2005).
Among children with wheeze in Portugal, a 20-ppb increase in 1-week avg individual
estimates of ambient NO2 exposure was associated with a 14% (95% CI: -12, 40)
increase in eNO and a -2.6% (95% CI: -3.9, -1.3) change in exhaled breath condensate
(EEC) pH (Martins et al.. 2012). School and home indoor NO2 concentrations were
nondetectable, providing support for an association with ambient NO2. No issues with
LOD were reported for ambient measurements. Further, time-weighted averages of
microenvironmental NO2 have shown good agreement with personal NO2
(Section 3.4.3.1). Children were reported to spend 85% of time at home or school,
underscoring the importance of the individual-level exposure estimation in this study.
5-115
-------
Evidence also points to associations of eNO in children with asthma with NO2
concentrations measured outside schools. Of the studies conducted in communities along
the Texas/Mexico border, most found NCh-associated increases in eNO. eNO was more
strongly associated with outdoor school NO2 than central site NO2 (Sarnat et al.. 2012) or
school indoor NO2 [(Greenwald et al.. 2013; Sarnat etal.. 2012); Figure 5-8 and
Table 5-16]. In the Texas/Mexico study, a 20-ppb increase in 96-h avg NO2 concentration
was associated with increases in eNO of 6.3% (95% CI: 2.5, 10) for outdoor school, 0.5%
(95% CI: 0.1,1.0) for indoor school, and 1.7% (95% CI: -1.0, 4.5) for central site. Most
indoor NO2 measurements were above the LOD. NO2 from the single central site in El
Paso was moderately to strongly correlated (Spearman r = 0.63-0.91) with school NO2
(Sarnat et al.. 2012). suggesting that for some schools, the central site measures captured
temporal variation in school-based measures. However, the variability in NO2 found
across schools (coefficient of variation = 59%) indicates that the stronger associations
with outdoor school NO2 may be attributable to school measurements better representing
variability in NO2 within the area and exposures of children. Misrepresenting temporal
variability in short-term exposure has been shown to influence health effect estimates
(Section 3.4.5.1). Holguin et al. (2007) did not find an association with eNO in children
with asthma in Ciudad Juarez schools. No association was found in a study of children in
France (Flamant-Hulin et al.. 2010). However, this study had weaker methodology
because of its cross-sectional design, comparison of eNO between low and high NO2
(means 10.1 and 17.4 ppb), and for some subjects, measurement of eNO 1 week before
NO2. NO2 measured within 650 m of subjects' schools (lag 0 day of 24-h avg) was
associated with eNO among children in Beijing, China examined before and after the
2008 Olympics (Linet al.. 2011).
With regard to confounding, most studies that assessed exposures in subjects' locations
adjusted for temperature and humidity, with a few additionally adjusting for asthma
medication use (Sarnat et al.. 2012; Delfino et al.. 2006). An array of copollutants was
examined, and most studies found associations with PM2 5 and with the traffic-related
copollutants EC/BC, OC, and VOCs. These copollutants showed a wide range of
correlations with NO2 (Pearson or Spearman r = -0.43 to 0.77). There is some evidence
for NO2 effects on pulmonary inflammation that are independent from PM2 5, EC/BC, or
OC. NO2-eNO associations were found with adjustment for personal PM2 5, EC, or OC.
Personal exposure measures, all of which were above the limit of detection (LOD), were
more weakly correlated with NO2 (Spearman r = 0.20-0.33) than central site measures
(r = 0.20-0.70) (Delfino etal.. 2006). For central site NO2, associations with eNO
decreased but remained positive with adjustment for BC or the oxidative potential of
ambient PM2 5 extracts measured in vitro [(Delfino et al.. 2013; Lin et al.. 2011);
Table 5-16. Figures 5-16 and 5-17]. The latter results support an independent association
with NO2 because oxidative stress is a key early event in the mode of action proposed for
5-116
-------
NC>2, traffic-related copollutants, and PlVfcs (Appendix to the ISA). The studies conducted
in El Paso, TX and Ciudad Juarez, Mexico did not analyze copollutant models with
EC/BC or PM25 However, NC>2 associations were less variable across schools than were
PM2 5 associations, and in Ciudad Juarez, NO2, but not EC/BC or PM2 5, was associated
with eNO (Sarnatetal.. 2012; Holguin etal.. 2007).
Pulmonary inflammation also was associated with VOCs (Greenwald et al.. 2013;
Martins etal.. 2012). In the El Paso schools, because of the high correlation (Pearson
r = 0.77) between NO2 and the sum of benzene, toluene, ethylbenzene, xylene (BTEX),
an independent association is not discernible for either pollutant. Reporting
copollutant-adjusted results only for EEC pH, Martins et al. (2012) found that
associations for individual estimates of outdoor NO2 exposure were similar after
adjustment for a VOC, which showed no or negative correlations with NC>2 (range of
Spearman r across four visits = -0.42 to 0.14; Table 5-16). VOC estimates were
attenuated to the null with adjustment for NCh; thus, NO2 may have confounded
associations for VOCs. There was no report of outdoor or indoor VOC or outdoor NO2
measurements being below the detection limit. Other pollutants, Os, SO2, PMio, and
PMio-2.5, were associated with pulmonary inflammation and oxidative stress but did not
show strong positive correlations with NO2 (r = -0.72 to 0.18). NO2 effect estimates
increased with adjustment for Os measured at school (Sarnatetal.. 2012) and became
null with individual estimates of PMio (Martins et al.. 2012). But PMio and NO2 were
strongly to moderately negatively correlated (r = -0.82 to -0.55).
Other studies have a weaker basis for inferring an independent effect of NO2 on
pulmonary inflammation and oxidative stress in children with asthma. They all assigned
NO2 exposure as ambient concentrations from one city central site or sites 5 km or 10 km
from subjects' homes. While the studies adjusted for potential confounding by
meteorological factors and asthma medication use, most did not examine confounding by
traffic-related copollutants. Findings were variable for indicators of inflammation among
eNO, IL-8, and EEC pH, as well as indicators of oxidative stress related to lipid
peroxidation. Some studies found associations with ambient NO2 (Liu etal.. 2014a;
Barraza-Villarreal et al., 2008) or inconsistent associations across the lags of exposure or
specific endpoints examined (Liu et al.. 2009b). In the others, effect estimates with wide
95% CIs did not support associations [(Berhane et al., 2011; Romieu et al., 2008);
Figure 5-8 and Table 5-16]. Studies also found associations with traffic proximity and
volume and with PlVfc 5, which was moderately to highly correlated with NO2
(r = 0.49-0.71) (Liu etal.. 2014a; Berhane etal.. 2011; Liu et al.. 2009b; Romieu et al..
2008);(Barraza-Villarreal et al.. 2008). Copollutant modeling was conducted in a study of
children in Windsor, Canada, and effect estimates for NO2 were largely attenuated with
adjustment for PM25 [(Liu et al.. 2009b); Table 5-171. PM2s estimates were less altered
5-117
-------
with adjustment for NCh; however, the reliability of the copollutant model is questionable
because of the high NO2-PM2 5 correlation (r = 0.71). In the limited analysis of
copollutant models with PMio (r = 0.59) or 862 (r = 0.18), NCh remained associated with
pulmonary inflammation or oxidative stress (Liu et al.. 2014a: Liu et al.. 2009b).
Adults with Asthma
Epidemiologic studies of pulmonary inflammation in adults with asthma, which were not
available for the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). show contrasting
associations with ambient NC>2. Both studies examined adults predominately with
persistent asthma, assessed NO2 exposure from central site monitors, and adjusted for
temperature and humidity. In a U.S. multicity (Boston, MA; New York, NY;
Philadelphia, PA; San Francisco, CA; Madison, WI) study nested within an asthma
medication trial, a 20-ppb increase in lag 0 day of 24-h avg NCh was associated with a
0.26-ppb (95% CI: 0.12, 0.40) increase in eNO (Oian et al.. 2009a'). A similar increase in
eNO was found for lag 0-3 day avg NO2 but not lags 1, 2 or 3. A larger effect was
estimated in the daily ICS group than the placebo or beta-agonist groups only for lag
0 day NO2. Among children and adults with asthma in Padua, Italy, a large percentage of
whom reported ICS use, lag 0 day of 24-h avg ambient NCh was not associated with eNO
or exhaled breath condensate pH (Maestrelli et al., 2011). The U.S. multicity study did
not indicate whether the NO2 concentration averaged from monitors within 32 km of
subjects' homes adequately represented the temporal variability in exposure and did not
examine whether the association for ambient NO2 was independent of other traffic-related
pollutants. Copollutant models were examined only for PMio, SO2, and Os, in which NO2
remained associated with eNO (Qian et al.. 2009a). Adjustment for NO2 attenuated the
effect estimates for PMio, SO2, and Os, indicating that the copollutant associations were
confounded by NO2.
5.2.2.6 Summary of Asthma Exacerbation
Evidence integrated across the array of health outcomes and disciplines strongly supports
a relationship between short-term NO2 exposure and asthma exacerbation. The evidence
for allergic inflammation, increased airway responsiveness, and clinical events, such as
respiratory symptoms in populations with asthma as well as ED visits and hospital
admissions for asthma, is consistent with the sequence of events in the proposed mode of
action linking short-term NO2 exposure and asthma exacerbation (Figure 4-1) and
supports the biological plausibility for a relationship. Much of this evidence, especially
from experimental studies, was described in the 2008 ISA for Oxides of Nitrogen (U.S.
EPA. 2008c). Recent findings, primarily from epidemiologic studies, continue to indicate
5-118
-------
NCh-associated increases in asthma exacerbation. Many recent epidemiologic studies
contribute additional exposure assessment in subjects' locations and examination of
potential confounding by or interactions with other traffic-related pollutants.
The primary evidence demonstrating that NC>2 can independently exacerbate asthma is
that NC>2-induced increases in airway responsiveness in adults with asthma, particularly
in response to nonspecific bronchoconstrictors and exposures in which subjects did not
exercise (Section 5.2.2.1). Of all the health outcomes examined in controlled human
exposure studies of NO2, increased airway responsiveness was induced by the lowest
NC>2 concentrations, 100 ppb for 1 hour and 200-300 ppb for 30 minutes. Further, a
meta-analysis indicates a clinically relevant doubling reduction in provocative dose in
adults with asthma in response to NO2 exposure relative to air exposure. Increased airway
responsiveness can lead to poorer asthma control. Thus, the evidence for relatively low
NC>2 exposures inducing clinically relevant increases in airway responsiveness in adults
with asthma provides key support that ambient concentrations of NC>2 can exacerbate
asthma. Most experimental studies did not report lung function changes, respiratory
symptoms, or an array of inflammatory responses following exposure to NC>2 (120-4,000
ppb for 30 minutes-6 hours) in adults with asthma or animal models of allergic disease in
the absence of challenge with a bronchoconstrictor (Sections 5.2.2.2. 5.2.2.3. and
5.2.2.5). However, there are several observations of MVinduced allergic inflammation,
most consistently indicated as increases in eosinophil number and activation of
eosinophils and/or neutrophils following exposures with and without an allergen
challenge (Section 5.2.2.5). Similar to airway responsiveness, allergic inflammation was
enhanced by lower NO2 exposures than many other health effects examined in
experimental studies: 260 ppb NO2 for 15-30 minutes or 400 ppb NC>2 for 6 hours. The
evidence for NCh-induced allergic inflammation also demonstrates that NO2 exposure
plausibly can lead to an asthma exacerbation.
Epidemiologic studies generally did not find NCh-associated changes in inflammatory
cell counts in populations with asthma; however, they did consistently indicate ambient
or personal NCh-associated increases in eNO (Figure 5-8 and Table 5-16). These findings
are coherent with experimental evidence for allergic inflammation because increases in
eosinophils and neutrophils are linked with NO production during an inflammatory
response. The limited studies in adults with asthma produced conflicting results, but the
large body of findings in children with asthma shows a consistent pattern of association
across the various lags of exposure and outcomes examined. T-derived lymphocyte
helper 2 (Th2)-mediated airway obstruction can lead to a decrease in lung function. Thus,
the evidence for NO2-induced allergic inflammation supports the epidemiologic
associations observed for ambient or personal NO2 concentrations with lung function
decrements in children with asthma as measured by supervised spirometry
5-119
-------
(Section 5.2.2.2). In fact, the evidence for NCh-related increases in allergic inflammation
provides biological plausibility for the associations observed for NO2 with the array of
asthma-related effects in populations with asthma that have a high prevalence of atopy
(53-84%) and groups of children with asthma not using anti-inflammatory ICS
(Hernandez-Cadena et al.. 2009; Liu et al.. 2009b: Barraza-Villarreal et al.. 2008;
Escamilla-Nufiez et al.. 2008). Although information is limited, a role for allergic
inflammation in NC>2-induced lung function decrements is supported by evidence for
changes in lung function mediated by mast cell degranulation. Mast cell degranulation
leads to histamine release (Section 4.3.2.2). Neural reflexes do not appear to mediate lung
function changes in response to ambient-relevant NC>2 exposures.
Consistent with the evidence for airway responsiveness and allergic inflammation,
epidemiologic studies consistently demonstrate associations of increases in ambient NO2
concentration with increases in asthma symptoms in children (Section 5.2.2.3) and
asthma hospital admissions and ED visits among subjects of all ages and children
(Section 5.2.2.4). The robustness of evidence is demonstrated by associations found in
studies conducted in diverse locations in the U.S., Canada, and Asia, including several
multicity studies. NC>2 was associated with the use or sale of asthma medication in adults
with asthma but not children with asthma. Individual epidemiologic studies examined
multiple outcomes and lags of exposure; however, a pattern of association was
consistently observed with NCh, which does not point to a higher probability of findings
due to chance alone.
The evidence for asthma is substantiated by several studies with strong exposure
assessment characterized by measuring NO2 concentrations in subjects' location(s).
Respiratory symptoms, lung function decrements, and pulmonary inflammation were
associated with personal total and ambient NO2 exposures (Martins et al.. 2012; Delfino
et al.. 2008a; McCreanor et al.. 2007) and NO2 measured outside schools (Greenwald et
al.. 2013: Zoraetal.. 2013: Sarnatet al.. 2012: Holguin et al.. 2007). Imparting
confidence in results for personal exposure metrics, no issues were reported regarding a
large number of NC>2 measurements being near the LOD. Given the high variability in
NC>2 concentrations, measurements in subjects' locations may better represent temporal
variation in ambient NC>2 exposures than area-wide central site concentrations
(Sections 2.5.3 and 3.4.4). Observations that daily variation in central site ambient NC>2
was related to variation in total personal NCh (r = 0.43) (Delfino et al.. 2008a) and that
indoor home or school NC>2 concentrations were negligible (Martins et al.. 2012) provide
additional support for a relationship of asthma exacerbation with ambient NO2 exposure.
A key uncertainty noted in the 2008 ISA for Oxides of Nitrogen was whether NO2 had an
effect independent of other traffic-related pollutants and PM2 5 (U.S. EPA. 2008c).
5-120
-------
Epidemiologic studies of asthma-related respiratory effects found associations with NO2
as well as with PM2 5, CO, BC/EC, UFP, other PM constituents, and VOCs (Figures 5-16
and 5-17). Among the studies that examined confounding by one of these copollutants,
most indicate an independent association with NC>2. The predominant method for
evaluation was copollutant models, which have well-recognized limitations for
distinguishing independent pollutant associations (Section 5.1.2.2). However, several
studies with strong exposure assessment provide a sound basis for inferring an
independent NO2 association, when integrated with experimental evidence.
In populations with asthma, personal total and ambient NO2 and NO2 measured outside or
0.65 km from children's schools were associated with respiratory symptoms, decreased
lung function, and pulmonary inflammation with adjustment for BC/EC, UFP, OC, or
PM2.5 (Martins etal..2012; Lin etal.. 2011; Delfino et al.. 2008a; McCreanor et al..
2007; Delfino et al.. 2006) (Figures 5-16 and 5-17). Also supporting an independent
association for NCh, some studies found associations with school or personal NC>2 but not
EC, OC, or PM2 5 (Samat et al.. 2012; Delfino et al.. 2008a; Holguin etal.. 2007). In a
few cases, adjustment for UFP or a VOC attenuated the NO2 association with one
outcome in a study but not another (Martins et al.. 2012; McCreanor et al.. 2007).
indicating the potential for confounding to differ by outcome. Among children with
asthma in El Paso, TX, school NO2 was not associated with asthma control score after
adjusting for BC (Zoraet al.. 2013). However, a copollutant model was not examined in
the group with atopy, to whom the association with NO2 was limited. Copollutant
associations adjusted for NO2 were robust in some cases (Lin etal.. 2011; Delfino et al..
2006) and attenuated in other cases (Martins etal.. 2012; Delfino et al.. 2008a). Thus, in
some studies, NO2 appeared to confound associations for traffic-related copollutants. The
spatial alignment of NO2 with subjects' location(s) may have reduced differences in
exposure measurement error between NO2 and copollutant, thereby improving the
reliability of copollutant model results. Correlations between NO2 and traffic-related
copollutants varied widely (r = -0.42 to 0.75), and inference from copollutant model
results also is improved by the low NO2-copollutant correlations found for personal
measurements (r = 0.20-0.33 for EC, OC, PM2 5 and -0.42 to 0.08 for benzene and
ethylbenzene) (Martins etal.. 2012; Delfino et al.. 2006). Neither study reported issues
regarding measurements being below LOD for any of the pollutants examined.
Consistent with findings for NO2 exposures in subjects' locations, copollutant models
based on central site concentrations indicate ambient NO2 remains associated with
asthma-related effects with adjustment for CO, UFP, a source apportionment factor
comprising EC and various metals, PM2s, or oxidative potential of PM2 5 extracts
(Delfino etal.. 2013; Iskandar etal.. 2012; Dales et al.. 2009a; Gent et al.. 2009; Jalaludin
et al.. 2008; Villeneuve et al.. 2007; Delfino etal.. 2003; von Klot et al.. 2002). For
5-121
-------
pollutants measured at central sites, the impact of copollutant adjustment on NCh
associations also differed by outcomes within studies (Liu et al.. 2009b: Andersen et al..
2008a; von Klot et al., 2002). Central site NO2 tended to show moderate correlations with
traffic-related copollutants (r = 0.28-0.56, but 0.66 for UFP and 0.71 for PM2 5), but
differences in spatial distributions may result in differential exposure measurement error
and render copollutant model results unreliable. Differential error may influence findings
from Delfino et al. (2008a). where the association between ambient NO2 and lung
function remained positive but was reduced with adjustment for personal PIVb 5.
Epidemiologic evidence also indicates that NO2 associations with asthma-related effects
are independent of nontraffic pollutants and other temporally varying factors (Tables 5-5.
5-8. and 5-16). In most cases, NC>2 associations were found with adjustment for 862,
PMio, PMio-2.5, or O3 (Liuetal.. 2014a; Sarnatet al.. 2012: Mann etal.. 2010: Patel et al..
2010: Strickland etal.. 2010: Dales et al.. 2009a: Oian et al.. 2009a: Jalaludin etal.. 2008:
Mortimer et al.. 2002). In some copollutant models, associations for NO2 as well as SO2
or PMio were attenuated (Martins etal.. 2012: Liu et al.. 2009b: Qian et al.. 2009a). and
an independent or confounding effect was not distinguished for either NC>2 or copollutant.
In exception, Samoli etal. (2011) indicated that the NO2 association with asthma ED
visits was confounded by SCh or PMio but not vice versa. Most epidemiologic studies
found associations between NO2 and asthma-related effects with adjustment for potential
confounding by temperature, humidity, and season. As examined in fewer studies, NC>2
associations persisted with adjustment for day of week, smoking, and asthma medication
use.
Gass etal. (2014) and Winquist et al. (2014) show increases in asthma ED visits when
ambient concentrations of NC>2 are jointly high with PM2s, CO, EC, Os, and/or 862. Such
joint effect analyses do not provide insight into independent effects of NC>2. These and
other epidemiologic studies (Schildcrout et al.. 2006: Delfino et al.. 2003) do not provide
evidence of synergistic interactions between NO2 and PM2 5, CO, EC, or VOCs. Such
interactions are not clearly demonstrated in controlled human exposure studies either
(Jenkins et al.. 1999: Devalia et al.. 1994: Hazucha et al.. 1994: Adams etal.. 1987).
Another line of evidence indicating that short-term NO2 exposure may have an
independent effect on asthma exacerbation is the coherence of evidence for indoor and
outdoor NO2. Except for Greenwald et al. (2013). indoor home or school NO2
concentrations were associated with respiratory symptoms and pulmonary inflammation
in children with asthma (Luet al.. 2013: Sarnat etal.. 2012: Hansel et al.. 2008). In both
cohorts, most indoor measurements were above the LOD (Raysoni et al.. 2011: Diette et
al.. 2007). Sarnatet al. (2012) found that correlations between NO2 and copollutants
differed between the indoor and outdoor environments for BC, PM, and 862, suggesting
5-122
-------
that NC>2 may exist as part of different pollutant mixtures in the indoor and outdoor
environments.
In summary, evidence for increased airway responsiveness and allergic inflammation in
experimental studies clearly demonstrates that short-term NC>2 exposure can induce
effects related to asthma exacerbation. That these hallmarks of asthma are enhanced by
exposures to 100-400 ppb NO2 substantiates the supposition that ambient NO2 exposures
can exacerbate asthma. Further, these effects are coherent with epidemiologic
associations consistently observed for short-term increases in NCh concentrations with
asthma-related hospital admissions, ED visits, symptoms, and pulmonary inflammation.
Key epidemiologic evidence comprises the associations with outdoor and indoor NO2
exposures assessed in subjects' locations, many of which persist in copollutant models
with another traffic-related pollutant or PM2 5 Not all asthma-related effects are
associated with NCh or show coherence between epidemiologic and experimental studies,
particularly lung function assessed in the absence of a bronchoconstricting agent.
Potential confounding has not been assessed for all correlated traffic-related pollutants,
and reliable methods are not available for simultaneous control for multiple pollutants.
However, the integrated evidence for airway responsiveness and allergic inflammation in
experimental studies and asthma-related effects in epidemiologic studies sufficiently
provides a biologically plausible link between short-term NCh exposure and asthma
exacerbation.
5.2.3 Allergy Exacerbation
The evidence from experimental studies for the effects of short-term NO2 and allergen
co-exposure on increasing allergic inflammation in adults with asthma and animal models
of allergic disease (Section 5.2.2.5) not only supports NCh-related asthma exacerbation
but also indicates that NCh-induced allergy exacerbation may be biologically plausible.
Support also is provided by in vitro findings that NC>2 can increase the allergenicity of
pollen (Cuinica et al., 2014; Sousaet al., 2012). Studies examining clinical indications of
allergy exacerbation have become available since the 2008 ISA for Oxides of Nitrogen
(U.S. EPA. 2008c). In contrast with asthma exacerbation, short-term NCh exposure is not
clearly related with clinical indications of allergy exacerbation.
Equivocal epidemiologic evidence in adults with allergies is indicated by associations of
ambient NO2 with physician visits, allergic rhinitis, or nonspecific respiratory symptoms
that are either inconsistent across the lags of exposure examined (Yilleneuve et al..
2006b), negative, or positive but with wide 95% CIs (Annesi-Maesano et al., 2012b; Feo
Brito etal.. 2007). Results from Annesi-Maesano et al. (2012b) are based on a
5-123
-------
multipollutant model, which can produce unreliable results (Table 5-17). Studies only
reported that ambient NO2 exposure was assigned as concentrations from one central site
or the average of multiple sites. And it is uncertain whether the temporal variability in
these NC>2 metrics adequately represented the variability in ambient concentrations across
the study area or in subjects' ambient exposures. Similarly inconclusive, a recent
controlled human exposure study of adults with allergic asthma did not find increases in
allergic inflammation after a 3-hour exposure to 350 ppb NC>2 and found increases in
respiratory symptoms only during exposure (Riedl et al.. 2012).
In children with allergies, increases in ambient NC>2 were associated with decreases in
lung function (Correia-Deur et al., 2012; Barraza-Villarreal et al., 2008). with an
association with cough found in a cohort in Mexico City (Barraza-Villarreal et al.. 2008;
Escamilla-Nufiez et al.. 2008). Strengths of these studies include the clinical assessment
of allergy and the supervised measurement of lung function. Although not specific to
allergy exacerbation, lung function can decrease during an allergy exacerbation due to
airway obstruction caused by Th2 cytokine-mediated inflammation. The studies in
children aimed to account for heterogeneity in ambient NCh concentrations. In one
cohort, exposures were assigned from sites within 5 km of children's home or school, but
a Pearson correlation of r = 0.21 between school and central site NO2 indicates the
variability at the central site may not represent the variability in the subjects' locations
(Barraza-Villarreal et al.. 2008). Another study examined NO2 from a central site in the
backyard of the subjects' school (Correia-Deur et al., 2012). providing a stronger basis
for inference of NC>2 effects. Increases in NO2 lagged 2 hours, averaged over the same
day, and averaged over 3 days were associated with decreases in PEF. However, counter
to expectation, associations were observed for the 36 children identified as having atopy
with a less stringent definition (one positive test among skin prick test, serum IgE, or
blood eosinophils: -0.87% [95% CI: -1.7, -0.04] per 20-ppb increase in lag 0 day NO2),
not the 28 children with atopy defined more stringently (all three tests positive: -0.30%
[95%CI:-1.7, 1.1]).
5-124
-------
Table 5-17 Epidemiologic studies of allergy exacerbation.
Study
Population Examined and
Methodological Details
NO2 Metrics
Analyzed
Effect Estimate
(95% Cl)
Single-Pollutant
Model3
Copollutant
Examination
tCorreia-Deur et al. (2012)
Sao Paolo, Brazil, Apr-Jul 2004
n = 36, one test positive (see below); n =
28, three tests positive, ages 9-11 yr
Repeated measures. Daily supervised
spirometry for 15 school days. Number of
observations not reported. Recruitment from
school. 86% participation. Allergic
sensitization ascertained by skin prick test,
blood eosinophils, and serum IgE. GEE with
autoregressive correlation matrix adjusted for
date, school absence, temperature, humidity.
NO2-outdoor school
24-h avg, lag 0 day
Mean: 69.9 ppbb
75th: 84.5 ppbb
90th: 102ppbb
% change PEF:
Group with 1 positive
test
-0.87% (-1.7, -0.04)
Group with 3 positive
tests
-0.30% (-1.7, 1.1)
All subjects, lag 0
With CO (r= 0.51)
-1.5% (-3.0, 0)
With SO2(r= 0.60)
-1.9% (-3.3, -0.37)
With PMio(r=0.59)
-0.75% (-4.4, 3.1)
With Os(r= 0.40)
-1.5% (-3.3, 0.38)
Associations for CO &
Os not altered by NO2
adjustment. SO2 &
PM-io attenuated.
tBarraza-Villarreal et al. (2008)
Mexico City, Mexico, Jun 2003-Jun 2005
n = 50, ages 6-14 yr, 72% with atopy
Repeated measures. Examined every
15 days for mean 22 weeks. Participation rate
not reported. 1,503 observations. Recruitment
from friends or schoolmates of subjects with
asthma. Clinical assessment of allergy.
Supervised spirometry. Linear mixed effects
model with random effect for subject and
adjusted for sex, BMI, temperature, ICS use,
time. Adjustment for outdoor activities,
smoking exposure, anti-allergy medication
use, season did not alter results.
NO2-central site
8-h max NO2
Closest site, within
5 km of homes or
schools.
R = 0.21 for central
site and school
NO2.
Mean: 37.4 ppb
Max: 77.6 ppb
OR for cough:
lag 0-1 day avg
1.28(1.04, 1.57)
% change FEV-i: lag
1-4 day avg
-0.64% (-2.1, 0.82)
No copollutant model.
PM2.5 associated with
lung function and
cough.
Moderate correlation
with NO2.
Pearson r= 61.
tEscamilla-Nufiez et al. (2008)
Mexico City, Mexico, Jun 2003-Jun 2005
n = 50, ages 6-14 yr, 79% with atopy
Part of same cohort as above. Participation
rate not reported. Linear mixed effects model
with random effect for subject and adjusted
for atopy, temperature, time, sex. Adjustment
for outdoor activities, smoking exposure,
season did not alter results.
NO2-central site OR for cough:
1-h maxNO2, lag 1.23(1.03,1.47)
0-1 day avg
Closest site, within
5 km of homes or
schools
Mean: 68.6 ppb
Upper percentile:
NR
Only multipollutant
model with Osand
PM2.5 analyzed.
Moderate correlation
between PM2.5 and
NO2. Pearson
r=0.62.
5-125
-------
Table 5-17 (Continued): Epidemiologic studies of allergy exacerbation.
Study
Population Examined and
Methodological Details
Villeneuve et al. (2006b)
Toronto, ON, Canada, 1995-2000
n = 52,971 physician visits for allergic rhinitis,
ages 65 yr or older
Time-series analysis. GLM adjusted for
temperature, relative humidity, daily number
visits for influenza, allergen levels, natural
spline for time trend.
NO2 Metrics
Analyzed
NO2-central site
24-h avg, lag 0 day
Average of 9 city
sites.
Mean: 25.4 ppb
Max: 71.7 ppb
Effect Estimate
(95% Cl)
Single-Pollutant
Model3
Quantitative results NR
NO2 associated with
physician visits for
allergic rhinitis at lag
0 day. Negative or null
associations at lag
1 day to 6 days.
Copollutant
Examination
No copollutant model.
Association also
observed for SO2 but
not PM2.5 or CO.
Correlations with NO2
NR.
tFeo Brito et al. (2007)
Ciudad Real and Puertollano, Spain
May-Jun 2000 or 2001
n = 137, ages NR, mild/moderate asthma and
pollen allergy
Repeated measures, 90% follow-up
participation. Daily symptom diaries. Number
of observations not reported. Recruitment
from allergy clinics. Clinical assessment of
allergy. Poisson regression adjusted only for
linear and quadratic terms for season.
NO2-central site
24-h avg
4 sites in
Puertollano, 1
mobile site in
Ciudad Real
Mean and Max
Ciudad Real: 17.4b,
35.6b
Puertollano: 29.5b,
100b
% change in
symptoms:
Ciudad Real, Lag
Day 4
4.75% (-5.75, 16.4)
Puertollano, Lag Day 3
-3.00% (-9.55, 4.03)
No copollutant model.
PM-io, SO2, Os
associated with
symptoms only in
Puertollano. Moderate
correlation with NO2.
R=0.67, 0.36, 0.36.
Pollen associated
with symptoms only in
Ciudad Real. Rwith
NO2 =-0.10 & 0.16
for two pollen types.
tAnnesi-Maesano et al. (2012b)
Multiple metropolitan locations, France,
May-Aug 2004
n = 3,708 with severe allergic rhinitis, ages
6 yr and older, 82% adults
Cross-sectional. Recruitment from physicians'
offices. No information on participation rate.
Clinical assessment of allergy and symptom
severity. Multilevel model adjusted for age,
date of physician visit, asthma status, postal
code. Did not consider confounding by
meteorology or SES.
NO2-central site
24-h avg, Lag
Day 1
Site in postal code
of home.
Mean: 9.9b
Max: 38.9b
NR
Only multipollutant
model analyzed with
SO2, Os, PM-io, pollen.
Correlation only
reported for pollen,
r=-0.12.
Note: More informative studies in terms of the exposure assessment method and potential confounding considered are presented
first.
Avg = average; Aug = August; BMI = body mass index; Cl = confidence interval; CO = carbon monoxide; FEVi = forced expiratory
volume in 1 second; GEE = generalized estimating equations; GLM = generalized linear model; ICS = inhaled corticosteroid;
NO2 = nitrogen dioxide; NR = not reported; O3 = ozone; ON = Ontario; OR = odds ratio; PEF = peak expiratory flow;
PM2.s = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PM10 = particulate matter with
a nominal mean aerodynamic diameter less than or equal to 10 |jm; ppb = parts per billion; SES = socioeconomic status;
SO2 = sulfur dioxide.
aEffect estimates are standardized to a 20 ppb for 24-h avg NO2.
""Concentrations converted from |jg/m3 to ppb using the conversion factor of 0.532 assuming standard temperature (25°C) and
pressure (1 atm).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
Respiratory effects in populations with allergy were associated with the traffic-related
pollutant CO and with PIVb.5, so it is unclear whether the supporting epidemiologic
5-126
-------
evidence represents an independent effect of NC>2. Only Correia-Deur et al. (2012)
examined copollutant models, which indicated that the NCh-PEF association in children
with and without allergies combined is independent of CO (Table 5-17). School-based
NO2 and CO may have similar exposure error, providing basis for good inference from
the copollutant model results. NO2 also remained associated with PEF with adjustment
for school SO2 concentrations. However, both NO2 and PMio (r = 0.59) associations were
attenuated when adjusted for each other, and a confounding or independent effect cannot
be distinguished for either pollutant. NO2 effects could be confounded by Os in the warm
season, and interactions between Os and allergens have been reported (U.S. EPA. 2013e).
However, NO2 was associated with PEF after adjustment for Os (r = 0.40) [(Correia-Deur
etal.. 2012); Table 5-171.
In summary, the evidence does not clearly indicate whether NO2 exposure independently
induces allergy exacerbation. While there is evidence for effects on key events in the
proposed mode of action, the limited evidence for effects on clinical events related to
allergy exacerbation is inconclusive. Further, in the limited analysis of key copollutants,
there is evidence for an effect of NO2 on lung function decrements independent of CO
measured at children's schools but uncertainty regarding confounding by PlVfc 5 or the
array of traffic-related copollutants that were not examined (Appendix to the ISA).
5.2.4 Exacerbation of Chronic Obstructive Pulmonary Disease
COPD is characterized by deterioration of lung tissue and airflow limitation. In
exacerbation of COPD, episodes of reduced airflow, which can be indicated by decreases
in lung function, can lead to symptoms such as cough, sputum production, and shortness
of breath. Severe exacerbation can lead to ED visits or hospital admissions. This
spectrum of outcomes comprises the majority of investigations of the effects of
short-term NO2 exposure on COPD exacerbation, and as described in the sections that
follow, the consistency of findings from previous and recent studies varies among
outcomes. A key early event in COPD exacerbation is pulmonary inflammation, which
mediates narrowing of the airways and reduces airflow. As described in Section 5.2.4.3.
limited recent information does not show NO2-related increases in pulmonary
inflammation in adults with COPD.
5-127
-------
5.2.4.1 Lung Function Changes and Respiratory Symptoms in Adults with Chronic
Obstructive Pulmonary Disease
Evidence does not clearly indicate a relationship for NC>2 exposure with changes in lung
function or respiratory symptoms in adults with COPD. Evidence is inconsistent in both
controlled human exposure and epidemiologic panel studies, many of which examine
both respiratory symptoms and lung function. Most of these studies were reviewed in the
2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). and the only recent study, which is
epidemiologic, does not support associations for ambient NO2 concentrations with either
respiratory symptoms or lung function decrements in adults with COPD.
Lung Function Changes
Epidemiologic studies recruited adults with COPD from clinics, and the nonrandom
selection of the general population may produce study populations less representative of
the COPD population. NO2 exposures were assessed primarily as 24-h averages of central
site NO2 measurements, and results are equally inconsistent for NO2 exposures assigned
from one site or averaged from multiple city sites (Table 5-18). Most previous and recent
epidemiologic studies of adults with COPD assessed lung function with unsupervised
home measurements, and associations with ambient NO2 concentrations were
inconsistent among the various lung function parameters (e.g., FEVi, PEF) or NO2
exposure lags (0-, 1-, 2-, or 2- to 7-day avg) examined (Peacock et al.. 2011; Silkoff et
al.. 2005; Higgins etal.. 1995). Lagorio et al. (2006) found an association between
ambient NO2 concentrations and FEVi (Table 5-18). with similar effects estimated for
adults with COPD and asthma.
In addition to the inconsistent evidence for changes in lung function in adults with
COPD, there is uncertainty regarding an independent association of NO2 from that of
copollutants. Studies did not examine abroad array of traffic-related copollutants, and
inference about confounding is limited by potential differential exposure error for
pollutants measured at central sites. Lagorio et al. (2006) found FEVi decrements in
association with NO2 but not PM2 5, which was moderately correlated with NO2. Only
Peacock etal. (2011) conducted copollutant modeling, and the NO2-PEF effect estimate
was attenuated with adjustment for BS. In contrast, the effect estimate for BS was
relatively unchanged with adjustment for NO2. With respect to PMio, no association was
found with FEVi (Lagorio et al.. 2006). or the NO2 association with PEF was attenuated
with adjustment for PMio (Peacock et al.. 2011).
5-128
-------
Table 5-18 Epidemiologic panel studies of adults with chronic obstructive
pulmonary disease.
Study
Population Examined and
Methodological Details
tPeacock et al. (2011)
London, U.K., Oct 1995-Oct 1997
n = 28-94, ages 40-83 yr
Repeated measures. Home PEF.
Examined daily for 21 -709 days.
Recruitment from outpatient clinic. 75%
follow-up. GEE adjusted for
temperature, season. Lung function
measures adjusted for indoor
temperature and time spent outdoors.
Silkoffetal. (2005)
Denver, CO, Winters 1999-2000 and
2000-2001
n = 34 with COPD, mean age 66 yr,
75% severe COPD
Repeated measures. Home PEF.
Recruitment from outpatient clinics,
research registries, advertisements.
93-96% diaries completed. Mixed
effects model with random effect for
subjects and adjusted for temperature,
relative humidity, barometric pressure.
Desqueyroux et al. (2002)
Oxide of Nitrogen
Metrics Analyzed
NO2-central site
1-h max NO2, lag 1 day
1 city site
Mean: 51.4 ppb
75th: 56 ppb
NO2-central site
24-h avg, lag 0, 1, 2
days
1 city site
Means:
1999-2000: 16 ppb
2000-2001: 29 ppb
y^th anrl May
/ \J\.\ I dl IU IVIdA.
1999-2000: 30, 54 ppb
2000-2001: 36, 54 ppb
NO2-central site
Effect Estimate
(95% Cl)
Single-Pollutant
Model3
PEF:
0.1 7% (0.03, 0.32)
PEF >20% below
predicted:
OR: 1.0(0.86, 1.2)
Symptomatic fall in
PEF:
OR: 1.1 (0.97, 1.3)
No quantitative data.
Negative, positive, and
null associations with
symptoms across NO2
lags.
Physician visits for
Copollutant
Examination
Symptomatic fall in
PEF
WithBS: 1.1 (0.84, 1.3)
With PMio:
0.97(0.81, 1.2)
Correlations NR. 95%
Cl forBS also
increased. No change
in OR for PMio with
NO2 adjustment.
No copollutant model.
Mixed positive,
negative, null
associations for PM2.5,
PMio, 03.
with Os.
Paris, France, Oct 1995-Mar 1996,
Apr-Sept 1996
n = 39, severe COPD, mean age 67 yr
Repeated measures. Recruitment from
physicians' offices. No information on
participation. GEE adjusted for FEV-i,
smoking, CO2 pressure, oxygen
treatment, dyspnea, temperature,
humidity, season, holiday.
24-h avg, lag 1-5 day
avg
Average of 15 city sites
Means for Periods 1 &
2
31.4,26.1 ppbb
Max for Periods 1 &2
68.1, 56.4 ppbb
COPD exacerbation
OR: 0.76(0.28,2.10)
0.47(0.02,9.45)
C>3 association robust
to NO2 adjustment.
Correlations not
reported.
SO2 and PMio not
associated with COPD.
Laqorio et al. (2006)
Rome, Italy, May-Jun, Nov-Dec 1999
n = 11, ages 40-64 yr, nonsmokers
Repeated measures. Supervised
spirometry. Examined 2/weekfortwo
1-mo periods. Mean observations per
subject = 15. Recruitment from
outpatient clinic. Participation rate NR.
GEE adjusted for season, temperature,
humidity, beta-agonist use.
NO2-central site
24-h avg, lag 0 day
Average of 5 city sites
within 2 km of subjects'
census tracts.
Mean: 37.6 ppbb
Max: 54.3 ppbb
% predicted FEVi:
-2.3 (-3.6,-1.0)
No copollutant model.
Lung function
associated with PM2.5,
PMio. Moderate
correlation with NO2.
Spearman r = 0.43 for
PM2.5, 0.45 for PMio.
5-129
-------
Table 5-18 (Continued): Epidemiologic panel studies of adults with chronic
obstructive pulmonary disease.
Study
Population Examined and
Methodological Details
Oxide of Nitrogen
Metrics Analyzed
Effect Estimate
(95% Cl)
Single-Pollutant
Model3
Copollutant
Examination
Harreetal. (1997)
Christchurch, New Zealand, Jun-Aug
1994
n = 40, ages 55-83 yr, nonsmokers
Repeated measures. Home PEF.
Recruitment from doctors' offices,
COPD support group, advertisements.
66% participation. Log-linear model
adjusted for day of study, temperature,
wind speed, CO, PM-io, SO2.
NO2-central site
24-h avg, lag 1 day
Number of sites NR.
Concentrations NR
PEF:
-0.72% (-1.5, 0.07)
Only multipollutant
model analyzed.
PM-io, CO, SO2 not
associated with PEF.
tBruske (2014): Bruske et al. (2010)
Erfurt, Germany, Oct 2001 -May 2002
n = 38, ages 35-78 yr, all male, 53%
also with asthma
Repeated measures. Examined 2/mo
for 6 mo. 381 observations after
excluding concurrent fever or infection.
Method of recruitment and COPD
assessment and participation rate NR.
Additive mixed models with random
intercept for subject and adjusted for
infection/antibiotic use in previous
2 weeks, long-term time trend,
temperature, humidity as linear terms or
penalized splines. Also evaluated
confounding by barometric pressure and
corticosteroid use.
NO2-central site
24-h avg, lag 0 0-23 h
before blood collection
1 site 3.5 km from
subjects' homes.
Mean: 13.5 ppbb
75th: 16.6 ppbb
NO-central site
Mean: 10.8 ppbb
75th: 14.0 ppbb
PMN:
-8.0% (-18, 3.1)
Lymphocytes:
8.4% (-5.0, 24)
PMN:
-0.80% (-10, 9.9)
Lymphocytes:
13% (-1.9, 23)
NO2 and NO reported
not to be associated
with eosinophils. No
quantitative results.
PMN with UFP:
7.3% (-14, 34).
Lymphocytes with
UFP:
8.4% (-7.2, 27)
CO associated with
lymphocytes.
NO2 highly correlated
with UFP and CO.
Spearman r= 0.66,
0.78.
No copollutant model
for NO.
Note: More informative studies in terms of the outcome examined, exposure assessment method, and potential confounding
considered are presented first.
Avg = average; Aug = August; BS = black smoke; Cl = confidence interval; CO = carbon monoxide, Colorado; CO2 = carbon
dioxide; COPD = chronic obstructive pulmonary disease; Dec = December; FEVi = forced expiratory volume in 1 second;
GEE = generalized estimating equations; NO = nitric oxide; NO2 = nitrogen dioxide; NR = not reported; O3 = ozone; OR = odds
ratio; PEF = peak expiratory flow; PM2 5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5
|jm; PM10 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; PMN = polymorphonuclear
cells; SO2 = sulfur dioxide; UFP = ultrafine particles; UK = United Kingdom.
aEffect estimates were standardized to a 20-ppb increase in 24-h avg NO2 or a 30-ppb increase 1-h max NO2.
""Concentrations converted from |jg/m3 to ppb using the conversion factor of 0.532 assuming standard temperature (25°C) and
pressure (1 atm).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
Similar to the epidemiologic studies, the controlled human exposure studies, which were
reviewed in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). have mixed results
regarding decrements in lung function with NC>2 exposure. Studies examined older adults
diagnosed with COPD, and most incorporated exercise in the exposure period to assess
lung function during varying physiological conditions (Table 5-19). Effects are
5-130
-------
inconsistent with 1- to 4-hour exposures in the range of 200 to 500 ppb NO2, with
decrements in FVC and FEVi observed following exposure to 300 ppb NO2 (Vagaggini
etal.. 1996; Morrow et al.. 1992) but not 400 or 500 ppb NO2 (Gong et al.. 2005; Linnet
al.. 1985a). No effect was observed after exposure to 1,000 or 2,000 ppb NO2 (Linn et al..
1985a). Furthermore, Gong et al. (2005) did not find NC>2 to enhance lung function
decrements induced by PM exposures.
Respiratory Symptoms
The limitations and uncertainties described above for the evidence base relating NO2
exposure to lung function changes in adults with COPD largely apply to the evidence
base for respiratory symptoms. Most epidemiologic panel studies recruited adults with
COPD from outpatient clinics or doctors' offices. Association between ambient NO2 and
respiratory symptoms are either null (Peacock et al.. 2011; Desqueyroux et al.. 2002) or
inconsistent across the lags of exposure or range of outcomes examined [(Silkoff et al..
2005; Harre etal.. 1997); Table 5-18]. Results are equally inconsistent for individual
symptoms of cough, wheeze, dyspnea, total symptoms, and medication use (Table 5-18).
No pattern of association was found for either 24-h avg or 1-h max NC>2 or for a
particular lag day of exposure examined (0, 1, or longer). Most of these studies assigned
exposures from a single central site, but associations with symptoms and medication were
inconsistent for NC>2 assigned from the closest site (Desqueyroux et al.. 2002) or site
within 5 km (Harre etal.. 1997).
In the studies that found associations with specific symptoms or lags of NO2, associations
also were found with PM2 5 and the traffic-related pollutants BS and CO (Peacock et al..
2011; Silkoff et al.. 2005; Harre etal.. 1997). Among adults in New Zealand, an increase
in 24-h avg NO2 was associated with an increase in inhaler use in a multipollutant model
with CO, PMio, and SO2 (Harre etal.. 1997). which has weak implications because of
multicollinearity. A recent study of adults in London, U.K. found that associations
between Lag Day 1 of 1-h max NO2 and dyspnea were null with adjustment for BS or
PMio (Peacock et al.. 2011). but the potential differential exposure error for pollutants
measured at central sites limits inference from the results. Thus, in the few associations
found between increases in ambient NO2 concentration and increases in symptoms or
medication among adults with COPD, there is uncertainty as to whether ambient NO2 has
effects independent of other traffic-related pollutants. The equally inconsistent findings
from controlled human exposure studies (Table 5-19) do not address uncertainties in the
epidemiologic evidence base. Some studies reported no change in symptom score in
adults with COPD (Gong etal.. 2005; Morrow etal.. 1992). though some studies reported
small, but statistically significant increases in symptom scores during NO2 exposures of
300-2,000 ppb for 1 hour with exercise (Vagaggini et al.. 1996; Linn etal.. 1985a).
5-131
-------
Table 5-19 Characteristics of controlled human exposure studies of adults with
chronic obstructive pulmonary disease.
Study
Disease Status; Sample
Size; Sex; Age
(mean ± SD)
Exposure Details
Endpoints Examined
Morrow et al.
(1992)
COPD; n = 13 M, 7F
(14 current smokers,
6 former smokers);
59.9 ± 7.0 yr
Healthy; n = 10M, 10 F
(13 never smokers,
4 former smokers,
3 current smokers)
300 ppbfor4 h;
Three 7-min periods of
exercise at VE = ~4 times
resting
Pulmonary function tests before, during,
and after exposure and 24-h
post-exposure.
Symptoms before, during, and after
exposure and 24-h post-exposure.
Vaqaqqini et COPD; n = 7 M;
al. (1996) 58±12yr
Healthy; n = 7 M;
34 ± 5 yr
300 ppb for 1 h;
Exercise at VE = 25 L/min
Pulmonary function tests before and 2 h
after exposure.
Symptoms before and 2 h after
exposure.
Gong et al.
(2005)
COPD; n = 9 M, 9 F;
72 ± 7 yr
Healthy; n = 2 M, 4 F;
68 ± 11 yr
(1)400ppbNO2for2h
(2) 200 ug/m3 CAPs for 2 h
(3) 400 ppb NO2 + 200 ug/m3
CAPs for 2 h
(1-3) Exercise 15 min on/
15 min off at VE = ~2 times
resting
Pulmonary function tests before and
immediately after exposure and 4 h and
22-h post-exposure.
Symptoms before, during, and after
exposure.
Linnetal. COPD; n = 13 M, 9 F
(1985a) (1 never smoker,
13 former smokers,
8 current smokers);
60.8 ± 6.9 yr
500, 1,000, or 2,000 ppb for
1 h;
Exercise 15 min on/15 min off
VE = 16 L/min
Pulmonary function tests before, during,
and after exposure.
Symptoms before, during, immediately
after, 1 day after, and 1 week after
exposure.
CAPs = concentrated ambient particles; COPD = chronic obstructive pulmonary disease; F = female; h = hours; L/min = liters per
minute; M = male; min = minutes; NO2 = nitrogen dioxide; SD = standard deviation; yr = years.
5.2.4.2 Hospital Admissions and Emergency Department Visits for Chronic
Obstructive Pulmonary Disease
In contrast with the inconsistent evidence for the effects of short-term NC>2 exposure on
lung function changes and respiratory symptoms in adults with COPD (Section 5.2.4.1).
epidemiologic evidence is consistent for NCh-related increases in hospital admissions and
ED visits for COPD. The few studies of COPD hospital admissions or ED visits
evaluated in the 2008 ISA for Oxides of Nitrogen provided initial evidence of a positive
association between short-term NO2 exposures and COPD hospital admissions and ED
5-132
-------
visits, with more studies focusing on hospital admissions (Figure 5-9 and Table 5-21).
However, these studies were more limited in their evaluation of potential confounders
and other factors that may modify the relationships of NCh exposure with COPD hospital
admissions and ED visits. Consistent with the 2008 ISA for Oxides of Nitrogen, a few
recent studies have examined COPD hospital admissions and ED visits and generally add
to the initial evidence of a positive association observed in the 2008 ISA for Oxides of
Nitrogen. The air quality characteristics of the study cities and the exposure assignment
approach used in each study evaluated in this section are presented in Table 5-20. Other
recent studies of COPD hospital admissions and ED visits are not the focus of this
evaluation, as detailed in Section 5.2.2.4. but the full list of these studies and study
details, can be found in Supplemental Table S5-3 (U.S. EPA. 2015h).
5-133
-------
Table 5-20 Mean and upper percentile concentrations of nitrogen dioxide in studies of hospital admission and
emergency department visits for chronic obstructive pulmonary disease.
Study Location Years
Mean
Concentration Upper Percentile of Copollutant
Exposure Assignment Metric (ppb) Concentrations (ppb) Examination
Hospital Admissions
tFaustini et al. 6 Italian cities
(2013) (2001-2005)
Ko et al. (2007a) Honq Konq, China
(2000-2004)
Average of NO2 concentrations over all 24-h avg 24.1-34.6
monitors within each city. Number of
NO2 monitors in each city ranged from
1-5a
Average of NO2 concentrations across 24-h avg 27.2
14 monitors.
NR Correlations (r),
across cities:
PMio: 0.22-0.79
Copollutant models:
PMio
75th: 34.0 Correlations (r):
Max: 83.8 PM2.5: 0.44
PMio: 0.40
SO2: 0.66
O3: 0.34
Copollutant models:
none
fQiu et al.
(2013b)
tWonq et al.
(2009)
Hong Kong, China
(1998-2007)
Hong Kong, China
(1996-2002)
Of 14 monitors, average NO2 based on 24-h avg 30.9
data from 10 monitors. 3 monitors sited
near roads and 1 monitor on a remote
island were excluded.
Average of NO2 concentrations across 24-h avg 31 .2
8 monitors.
NR Correlations (r): NR
Copollutant models:
PMio
75th: 37.0 Correlations (r): NR
Max: 89.4 Copollutant models:
none
5-134
-------
Table 5-20 (Continued): Mean and upper percentile concentrations of nitrogen dioxide in studies of hospital
admission and emergency department visits for chronic obstructive pulmonary disease.
Study
Location Years Exposure Assignment
Mean
Concentration
Metric (ppb)
Upper Percentile of
Concentrations (ppb)
Copollutant
Examination
Emergency Department Visits
fStieb et al.
(2009)
tArbex et al.
(2009)
7 Canadian cities
(1992-2003)
Sao Paulo, Brazil
(2001-2003)
Average NO2 concentrations from all 24-h avg 9.3-22.7
monitors in each city. Number of NO2
monitors in each city ranged from 1-14.
Average of NO2 concentrations across 1-h max 63.0
4 monitors.
75th: 12.3-27.6
75th: 78.6
Max: 204.6
Correlations (r) only
reported by city and
season.
Copollutant models:
none
Correlations (r):
PMio: 0.60
SO2: 0.63
CO: 0.56
Copollutant models:
none
avg = average; CO = carbon monoxide; max = maximum; NO2 = nitrogen dioxide; NR = not reported; O3 = ozone; PM2.5 = particulate matter with a nominal mean aerodynamic
diameter less than or equal to 2.5 |jm; PMio = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; SO2 = sulfur dioxide.
aMonitoring information obtained from Colais et al. (2012).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-135
-------
Hospital Admissions
Consistent with the 2008 ISA for Oxides of Nitrogen, relatively few recent studies have
focused on the outcome of COPD hospital admissions, but these studies build upon the
initial evidence of a positive association (Figure 5-9). Faustini et al. (2013) examined the
relationship between short-term air pollution exposures and respiratory-related hospital
admissions, including COPD, specifically on the adult population (i.e., individuals
35 years of age and older) in six Italian cities. In a time-series analysis, the authors
examined the lag structure of associations through single-day lags as well as cumulative
lags using cubic polynomial distributed lags to identify whether the NO2 effect on
respiratory-related hospital admissions was immediate (lag 0, lag 0-1 days), delayed (lag
2-5 days), or prolonged (lag 0-3, 0-5 days). For COPD hospital admissions, the authors
observed stronger evidence for immediate (lag 0: 4.6% [95% CI: 0.64, 8.6] for a 20-ppb
increase in 24-h avg NO2 concentrations) NO2 effects on COPD hospital admissions.
Smaller associations were observed when examining prolonged effects, (3.3% for lag
0-3 days and 3.1% for lag 0-5 days). There was no evidence for delayed effects (lag
2-5 days). In a copollutant model with PMio at lag 0, the association between NO2 and
COPD hospital admissions remained relatively unchanged compared to the
single-pollutant model results (3.9% [95% CI: -1.7, 9.8]).
In a study conducted in Hong Kong, China from 2000-2004, Ko et al. (2007a) also
examined the lag structure of associations between short-term air pollution exposures and
COPD hospital admissions. In analyses of both single-day lags and multiday averages,
Ko et al. (2007a) observed the largest magnitude of an association at lags ranging from
0-3 to 0-5 days (10.1% [95% CI: 8.5, 12.2] for a 20-ppb increase in 24-h avg NO2
concentrations at both 0-3 and 0-5 day lags). These associations are larger in magnitude
than those reported by Oiu et al. (2013b) at lag 0-3 (4.7% [95% CI: 3.3, 6.2] for a 20 ppb
increase in 24-h avg NO2 concentrations) for a study also conducted in Hong Kong,
China, but for a longer duration (1998-2007). Although Ko et al. (2007a) reported
associations larger in magnitude for multiday averages, the authors also observed a
positive association across single-day lags, with lag 0 having one of the stronger
associations (3.4% [95% CI: 1.9, 5.0]), which is of similar magnitude to the lag 0 effect
observed in Faustini etal. (2013). Ko et al. (2007a) only examined the potential
confounding effects of copollutants through the use of three- and four-pollutant models,
which are difficult to interpret. In comparisons of the single-pollutant results for NO2and
the other pollutants examined (Os, PM2s, and PMio), similar patterns of associations were
observed across pollutants. Additionally, Ko et al. (2007a) examined whether there was
evidence of seasonal differences in NO2-COPD hospital admission associations. When
using the warm season as the referent, the authors reported evidence of larger
5-136
-------
associations in the cold season (i.e., December to March). These results are consistent
with the results of Ko et al. (2007b) for asthma (Section 5.2.2.4) and support potential
differences in seasonal associations by geographic location.
In addition to examining the association between short-term air pollution exposures and
COPD hospital admissions, Qiu et al. (2013b) also examined whether air pollution
associations with COPD hospital admissions were modified by the interaction between
season and humidity. In models stratifying by both season (warm: May-October; cold:
November-April) and humidity (high or humid: >80%; low or dry: <80%) the authors
found larger NO2 associations in the cool season and high humidity days (5.6 and 6.3%,
respectively) compared to the warm season and low humidity days (3.8 and 4.6%,
respectively) for a 20-ppb increase in 24-h avg NO2 concentrations at lag 0-3 days.
When examining the joint effect of season and humidity, Qiuetal. (2013b) found that the
magnitude of the association was larger when season and humidity were considered
together. Specifically, the largest associations were observed for the combination of
warm season and humid days (7.3% [95% CI: 3.7, 11.1]; lag 0-3) and cool season and
dry days (9.3% [95% CI: 6.2, 12.5]; lag 0-3). In a series of copollutant models with PMio
and all combinations of season and humidity, NO2-COPD hospital admission associations
were attenuated. Null associations were observed in models for warm season plus dry
days and cool season plus humid days. These results further highlight the different
seasonal patterns in NO2 associations that have been reported across different geographic
areas as well as the potential influence of different weather conditions on NO2-related
health effects.
Emergency Department Visits
As in the 2008 ISA for Oxides of Nitrogen, relatively few studies have examined the
relationship between short-term NO2 exposures and ED visits, compared to hospital
admissions. In the seven Canadian cities discussed previously, consistent with the asthma
ED visits results, Stieb et al. (2009) did not find evidence of associations between
24-h avg NO2 and COPD ED visits at individual lags ranging from 0 days (0.1% [95%
CI: -6.1, 6.8] for a 20-ppb increase in 24-h avg NO2) to 2 days (-5.2% [95% CI: -12.4,
2.7]). Additionally, there was no evidence of consistent associations between any
pollutant and COPD ED visits at subdaily time scales (i.e., 3-h avg of ED visits versus
3-h avg pollutant concentrations).
Arbex et al. (2009) also examined the association between COPD and several ambient air
pollutants, including NO2, in a single-city study conducted in Sao Paulo, Brazil, for
individuals over age 40 years. Associations between NO2 exposure and COPD ED visits
were examined in both single-day lags (0 to 6 days) and a polynomial distributed lag
5-137
-------
model (0-6 days). However, for NO2, only those results that were statistically significant
were presented, that is, for individuals 65 years of age and older for lag 5 days (4.3%
[95% CI: 0.5, 8.3] for a 20-ppb increase in 24-h avg NO2 concentrations) and a
distributed lag of 0-5 days (9.6% [95% CI: 0.2, 19.9]). The authors did not analyze
copollutant models but reported moderate correlations between NO2 and PMio (r = 0.60),
SO2 (r = 0.63), and CO (r = 0.56).
Summary of Chronic Obstructive Pulmonary Disease Hospital Admissions
and Emergency Department Visits
In combination with those studies evaluated in the 2008 ISA for Oxides of Nitrogen,
recent studies add to the growing body of literature that has examined the association
between short-term NO2 exposures and COPD hospital admissions and ED visits.
Overall, these studies have reported consistent positive associations with evidence of
NO2-COPD hospital admissions and ED visits occurring immediately (lag 0) as well as a
few days after exposure (average of lags up to 5 days) (Figure 5-9). However, caution
should be used in inferring the independent effects of NO2 exposure due to the relative
sparseness of copollutant model analyses as well as the high correlation often observed
between NO2 and other traffic-related pollutants (e.g., CO, PM2s). Additionally, studies
that have focused on COPD hospital admissions and ED visits have not thoroughly
examined potential seasonal differences in associations; however, initial evidence
suggests that the combination of season and weather conditions, such as humidity, may
have a larger effect on NO2-COPD hospital admission associations than either
individually. Additionally, these studies have provided limited information on individual-
or population-level factors that could modify the NO2-hospital-admission or ED visit
relationship, or the shape of the C-R relationship.
5-138
-------
Study
Koetal. (2007)
Qiuetal. (2013)
Wong etal. (2009)
v-ar»« /at ^1 ff)r\r\f^\
Ysng et 31. (zuuoj
IV/|ftf-J/-io\ /l^o r ^OnfT^
iviooiyavKar ^zuuoj
_ . /rt«-«,JO\
hsustini et si. (zuloj
D««I Q+ oi fonn'^
r eei en ai. ^zuuo^
tie eta. (
Arbex et al. (2009)
Location
Hong Kong
Hong Kong
Hong Kong
Vancouver, CAN
P \f P +\ PA
UOOK uounty, UA
LA County, CA
_ .
b Itslisn cities
A+l + f^A
'
i Csnsdisn cities
S§o Psulo, Brazil
Age
All
All
All
cc>
CC _|_
DO+
„_
DO
65+
OC i
00+
All
MM
All
MM
CCi
bo+
Lag
0-3
0-3
0-1
n 1
u- i
1
0
Oc r\\
-O UL
n o
Oc ni
-O UL
. , !
Hospital Admissions
— • —
-•—
— • —
•k
^
-•-
ED Visits
\
w
i , , , , , , , , , ,
-4-20 2 4 6 8 10 12 14 16 18 20
% Increase (95% Cl)
Note: CA = California; CAN = Canada; Cl = confidence interval; DL =distributed lag; ED = emergency department; GA = Georgia;
LA = Los Angeles. Black = studies from the 2008 Integrated Science Assessment for Oxides of Nitrogen, red = recent studies. Effect
estimates are standardized to a 20-ppb increase in 24-h avg nitrogen dioxide and 30-ppb increase in 1-h max nitrogen dioxide.
Figure 5-9 Percentage increase in chronic obstructive pulmonary disease
hospital admissions and emergency department visits in relation
to nitrogen dioxide concentrations.
5-139
-------
Table 5-21 Corresponding risk estimate for studies presented in Figure 5-9.
Study
Location Age
Averaging Time
% Increase
Lag (95% Cl)
Hospital Admissions
Ko et al. (2007a)
fQiu et al.
(2013b)a
tWonq et al.
(2009)a
tYanq et al.
(2005)
tMoolqavkar
(2003)
fFaustini et al.
(2013)
Hong Kong, All
China
Hong Kong, All
China
Hong Kong, All
65+
Vancouver, BC, 65+
Canada
Cook County, IL 65+
LA County, CA 65+
6 Italian cities 35+
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
0-3 10.1 (8.5, 12.2)
0-3 2.2(0.8,4.4)
0-1 7.1(5.1,9.1)
0-1 4.6(2.4,6.8)
1 19.0(4.0,37.0)
0 4.9(1.6,8.2)
0 3.6(2.8,4.5)
0-5 DL 3.1 (-2.6, 9.2)
Emergency Department Visits
Peel et al. (2005)
tStieb et al.
(2009)
tArbex et al.
(2009)
Atlanta, GA All
7 Canadian cities All
Sao Paulo, Brazil 65+
1-h max
24-h avg
24-h avg
0-2 5.3(0.9,9.9)
0 0.1 (-6.1,6.8)
0-5 DL 9.6(0.2,19.9)
avg = average; BC = British Columbia; CA = California; Cl = confidence interval; DL = distributed lag; GA = Georgia; IL = Illinios.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5.2.4.3 Subclinical Effects Underlying Chronic Obstructive Pulmonary
Disease—Pulmonary Inflammation
Exacerbation of COPD can be precipitated by increases in airway responsiveness and
pulmonary inflammation. While there is some supporting evidence for an effect of NC>2
exposure in initiation of inflammation (Sections 4.3.2.3 and 4.3.2.1). the effects of NC>2
on airway responsiveness and inflammation are not well characterized in adults with
COPD. Thus, little information is available to propose a mode of action that would
5-140
-------
support the associations observed between ambient NO2 concentrations and hospital
admissions and ED visits for COPD (Section 5.2.4.1). In a recent epidemiologic study,
neither ambient NCh nor NO was associated with indicators of inflammation such as
increases in the numbers of blood eosinophils or lymphocytes in adults with COPD
consistently among Lag Days 0, 1, and 2 [(Bruske et al., 2010); Table 5-181. NO2 at Lag
Day 0 was associated with an increase in neutrophils only with adjustment for UFP.
However, the 95% CI was wide, indicating an imprecise association. Further limiting
inference from the copollutant model, NO2 and UFP were measured at central sites and
were highly correlated (Spearman r = 0.68). UFP and OC were associated with decreases
in neutrophils, of which the relation to COPD exacerbation is not clear. Vagaggini etal.
(1996) found no changes in either adults with COPD or healthy adults in inflammatory
cell counts in sputum following a 1-hour exposure to 300 ppb NO2.
5.2.4.4 Summary of Exacerbation of Chronic Obstructive Pulmonary Disease
Evidence for the effects of short-term NO2 exposure on COPD exacerbation is
inconsistent among the various outcomes examined and across scientific disciplines. In
epidemiologic studies, short-term increases in ambient NO2 concentration are
consistently associated with increases in hospital admissions and ED visits for COPD
(Section 5.2.4.2). However, NO2-related increases in respiratory symptoms or decreases
in lung function in adults with COPD are not consistently observed in epidemiologic or
controlled human exposure studies (Section 5.2.4.1). Further, a proposed mode of action
for NO2 effects on COPD exacerbation is not clear. In limited examination, an
epidemiologic and controlled human exposure study do not indicate NO2-related
increases in inflammation in adults with COPD (Section 5.2.4.3). Epidemiologic studies
assigned NO2 exposure as central site ambient concentrations (average of multiple
monitors, nearest site), and many found associations with PM2 5 and with the
traffic-related pollutants CO, BS, and UFP. Epidemiologic studies have not adequately
informed the potential for confounding by PIVbs or traffic-related copollutants,
particularly for the associations observed between NO2 and COPD hospital admissions
and ED visits. Because of the inconsistent evidence across disciplines for effects on
clinical indications of COPD exacerbation and the lack of evidence for effects on
underlying mechanisms, there is uncertainty regarding a relationship between short-term
NO2 exposure and COPD exacerbation.
5-141
-------
5.2.5 Respiratory Infection
The respiratory tract is protected from exogenous pathogens and particles through various
lung host defense mechanisms that include mucociliary clearance, particle transport and
detoxification by alveolar macrophages, and innate and adaptive immunity. The 2008
ISA for Oxides of Nitrogen reported clear evidence from animal toxicological studies for
NO2-induced susceptibility to bacterial or viral infection with some coherence with
results from controlled human exposure and epidemiologic studies (U.S. EPA. 2008c).
There is some mechanistic support for these observations, with NC>2-induced impairments
in alveolar macrophage (AM) function found in some but not all animal toxicological
studies. Effects on mucociliary clearance and activity were not in a consistent direction,
but the exact mechanism by which mucociliary clearance could impair host defense is not
well characterized. Recent contributions to the evidence base are limited to epidemiologic
studies. These studies show associations between increases in ambient NO2
concentrations and increases in hospital admissions and ED visits for respiratory
infections but do not consistently show associations with respiratory infections reported
or diagnosed in children.
5.2.5.1 Susceptibility to Bacterial or Viral Infection in Experimental Studies
A large body of evidence, provided by studies reviewed in the 2008 ISA for Oxides of
Nitrogen (U.S. EPA. 2008c). demonstrates increased susceptibility of rodents to viral or
bacterial infection following short-term NO2 exposure. These studies used a variety of
experimental approaches but in most cases included an infectivity model of exposing
animals to NO2 or filtered air and then combining treatment groups for a brief exposure
to an aerosol of a viable agent, such as Streptococcus zooepidemicus, Streptococcus
pyogenesi, Staphylococcus aureus, and Klebsiella pneumoniae. Toxicological studies
measured mortality over a specified number of days following the challenge, and both
toxicological and controlled human exposure studies examined endpoints such as
bacterial clearance and viral inactivation by cells isolated from BAL or nasal fluid
(Table 5-22). While there are differences in sensitivity across species to various
infectious organisms, host defense mechanisms are shared, and the infectivity model is
well accepted as an indicator of impaired or weakened pulmonary defense.
Mortality from Bacterial Infection
Compared to clear air exposure, short-term exposures of mice to 500 ppb NO2 did not
increase mortality following bacterial infection (Ehrlich. 1980; Ehrlich et al.. 1979;
Gardner et al., 1979). Each of these studies demonstrated that exposures of 500 ppb NO2
5-142
-------
increased mortality or decreased survival time with long-term exposures (>1 month), but
higher concentrations of 1,500 to 5,000 ppb were required to increase mortality due to
bacterial infection for short-term exposures. In such examinations of the relationship
between NCh concentration and time, concentration was found to be more important than
time in determining mortality due to bacterial infection (Ehrlich et al., 1979; Gardner et
al.. 1979). Other studies also reported bacterial infection-related mortality or shorter
survival time in mice in response to NC>2 exposures (3 hours) of 2,000 or 3,000 ppb
(Sherwood etal.. 1981: Illingetal.. 1980: Ehrlich et al.. 1977). but 2-hour exposures to
NC>2 at concentrations less than 5,000 ppb did not increase mortality in Purvis and
Ehrlich (1963).
Table 5-22 Characteristics of experimental studies of susceptibility to infection.
Study
Species (Strain);
Sample Size;
Sex; Age
Exposure Details
Endpoints Examined
Mortality from Infection
Ehrlich
(1,2) Mice;
(1) 500 ppb NO2 continuously for 1 week-1 yr (1-3) Mortality
(3) Mice, hamsters, (2) 1,500 ppb NO2 continuously for 2 h-3 mo
and squirrel
monkeys;
(3) 1,500-50,000 ppb NO2for2 h
n > 88/group; (1 ~3) Klebsiella pneumoniae challenge
6-8 weeks immediately after exposure.
Ehrlich et al.
(1979)
Mice(CD2F1,
Cd-1);
n = 99-127/group;
F; 6-8 weeks
(1) 500 ppb NO2 3 h/day, 5 days/week for 1
mo
(2) 100 ppb NO2 continuously + 500 ppb 3/h
day, 5 days/week for 1 mo
(1-2) Streptococcus pyogenes challenge
immediately after exposure.
(1-2) Mortality
Gardner et Mice (Swiss albino); (1) 500 ppb NO2 continuously for 7 days-1 yr Mortality
al. (1979) n = 20/group; F; (2) 1,500 ppb NO2 continuously for
age NR 2 h-21 days
(3) 1,500 ppb NO2 7 h/day for 7 h-11 days
(3) 3,500 ppb NO2 continuously for
30 min-16 days
(4) 3,500 ppb NO2 7 h/day for 7 h-13 days
(1-4) Streptococcus pyogenes challenge
immediately after exposure.
5-143
-------
Table 5-22 (Continued): Characteristics of experimental studies of susceptibility
to infection.
Study
Species (Strain);
Sample Size;
Sex; Age
Exposure Details
Endpoints Examined
Ehrlich et al.
(1977)
Mice (CF-1);
n = 5-88/group; F;
5-8 weeks
0, 1,500, 2,000, 3,500, or 5,000 ppb NO2 for
3h
Streptococcus pyogenes challenge
immediately after exposure.
Mortality
Illinq et al.
(1980)
Mice(CD-l);
n = 16/group; F;
5-6 weeks
1,000 ppb, 3,000 ppb NO2, or air for 3 h;
With or without continuous exercise;
Streptococcus pyogenes challenge
immediately after exposure.
Mortality after
Streptococcus pyogenes
challenge.
Purvis and Mice (Swiss
Ehrlich Webster and
(1963) albino);
1,500, 2,500, 3,500, or 5,000 ppb NO2 for 2 h; Mortality
Klebsiella pneumoniae challenge 0-27 h
n > 25/group; sex post-exposure.
and age NR
Graham et al. Mice (CD-1): (1) 4,500 ppb NO2 for 1, 3.5, and 7 h Mortality
(1987) n = 5-12/group; sex (2) 1,500 ppb NO2 continuously with a daily
NR; 4-6 weeks spike of 4]50o ppb for 1, 3.5, and 7 h;
(1-2) Streptococcus zooepidemicus challenge
immediately and 18 h after exposure.
Bacterial Clearance and Virus Inactivation
Davis et al.
Mice (C57BL/6N);
8-10 weeks;
n = 6/group
0, 500, 1,000, 2,000, or 5,000 ppb NO2 for Bacterial clearance,
4 h; bactericidal activity.
Mycoplasma pulmonis challenge immediately
after exposure.
Parker et al. Mice (C57BL/6N
(1989) and C3H/HeN);
6-10 weeks
0 and 5,000 ppb NO2 for 4 h;
Mycoplasma pulmonis challenge immediately
after exposure.
Histopathological
evaluation, bacterial
infection and clearance
4 h up to 7 days
post-challenge, BAL fluid
cell counts.
Jakab (1988) Mice (Swiss);
n = 6-10/group; F;
age NR
0, 2,500, 4,000, or 5,000 ppb NO2 for 4 h
Challenge with Staphylococcus aureus,
Proteus mirabilis, or Pasteurella
pneumotropica immediately before exposure,
after exposure, or in between two NO2
exposures
Bactericidal activity
immediately after bacteria
and NO2 exposure or 4
hours after NO2 and
bacteria exposure.
Sherwood et
al. (1981)
Mice (Swiss albino);
n = 8-24/group; M;
age NR
1,000 ppb NO2 for 24 and 48 h;
Streptococcus (Group C) challenge
immediately after exposure.
Bacterial counts 0-48 h
post-challenge, bacterial
clearance,
histopathological
evaluation, mortality.
5-144
-------
Table 5-22 (Continued): Characteristics of experimental studies of susceptibility
to infection.
Study
Species (Strain);
Sample Size;
Sex; Age
Exposure Details
Endpoints Examined
Goldstein et
al. (1974)
Goldstein et
al. (1973)
Frampton et
al. (2002)
Mice (Swiss albino);
n = 30/group; M;
age NR
Mice (Swiss albino);
n = 30/group; M;
age NR
Human;
(1,2)n = 12 M, 9F;
F = 27.1 ±4.1 yr
M = 26.9 ± 4.5 yr
(1)1,740ppbNO2+ 110 ppb Os
(2) 1,490 ppb NO2 + 200 ppb O3
(3) 2,300 ppb NO2 + 200 ppb O3
(4) 1,780 ppb NO2 + 270 ppb O3
(5) 4,180 ppb NO2 + 210 ppb O3
(1-5) 17 h; Staphylococcus aureus challenge
immediately after exposure.
(1) Staphylococcus aureus challenge
immediately before exposure;
0, 1 ,900, and 3,800 ppb NO2 for 4 h
(2) Staphylococcus aureus challenge
immediately after exposure;
0, 1,000, or2,300ppbNO2for17h;
(1 ) 600 ppb for 3 h,
(2) 1,500 ppb for 3 h;
(1,2) Exercise 10 min on/20 min off at
VE = 40 L/min
Bacterial counts,
bactericidal activity, and
bacterial clearance 0 h
and 4 h after challenge.
(1) Bacterial counts and
bactericidal activity 5 h
after challenge (i.e., 1 h
after exposure).
(2) Bacterial counts and
bactericidal activity 0 h
and 4 h after challenge.
Viral titers in AM and
bronchial epithelial cells
after influenza and RSV
challenge.
Frampton et
al. (1989)
Human;
(1)n = 7M, 2F;
30 yr (range:
24-37)
(2)n = 11 M, 4F;
25 yr (range:
19-37)
(1) 600 ppb for 3 h,
(2) Three 15-min peak exposures to
2,000 ppb with continuous 500 ppb for 3 h;
(1,2) Exercise 10 min on/20 min off at
VE = ~4 times resting
Inactivation of influenza
virus by BAL cells.
Goings et al.
(1989)
Human;
(1)n = 44
(2) n = 43
(3)n = 65; sex NR;
range: 18-35 yr
(1) 2,000 ppb for 2 h
(2) 3,000 ppb for 2 h
(3) 1,000 or 2,000 ppb for 2 h
Nasal wash virus isolation
and count 4 days after
virus administration.
Serum and nasal wash
antibody response
4 weeks after virus
administration.
AM = alveolar macrophage; BAL = bronchoalveolar fluid; F = female; h = hour; IL = interleukin; L/min = liters per minute;
M = male; min = minutes; NO2 = nitrogen dioxide; NR = not reported; O3 = ozone; ppb = parts per billion; RSV = respiratory
syncytial virus; SD = standard deviation; yr = years.
5-145
-------
Most studies in mice examined continuous NC>2 exposures, but there is no indication that
peak exposures superimposed on a lower continuous background level of NCh alter
susceptibility to infection. Compared to clean air exposure, 4,500 ppb NO2 spikes alone
and the spikes superimposed on a 1,500 ppb NO2 background exposure similarly
increased mortality when infection occurred immediately after NC>2 exposure (Graham et
al.. 1987). Mortality increased in proportional to duration of the 4,500 ppb spike from 1
to 3.5 to 7 hours. In mice challenged 18 hours after NCh exposure, increases in mortality
were statistically significant only with 3.5- and 7-hour exposures to 4,500 ppb NC>2
(combining groups with spike alone and with 1,500 background NC>2 exposure).
Limited analysis indicates that effects of NO2 exposure on mortality from infection may
differ among species. Increases in K. pneumoniae infection mortality were observed in
mice after a 2-hour exposure to 3,500 ppb NC>2 but in hamsters and squirrel monkeys
only at higher than ambient-relevant NC>2 concentrations (greater than 35,000 ppb and
50,000 ppb, respectively) (Ehrlich. 1980). The effects of NCh also were observed to vary
by infectious organism. NC>2 exposure (1,000 ppb, 24 or 48 hours) shortened survival
time in mice after infection with the virulent group C Streptococci but not
Staphylococcus aureus (Sherwood et al.. 1981).
Bacterial Clearance, Virus Inactivation
Consistent with findings for mortality from bacterial infection, NC>2 exposures of mice in
the range of 1,000 to 5,000 ppb reduced bacterial clearance and/or bactericidal activity
from the lungs (Davis etal.. 1991; Parker etal.. 1989; Jakab. 1988; Sherwood et al..
1981; Goldstein et al.. 1974; Goldstein et al.. 1973). In these studies, mice were
challenged with radiolabeled bacteria either immediately before or after NO2 exposures
of 4 to 48 hours. For four-hour exposures, bacterial clearance was not affected with 500
ppb NO2 (Davis etal.. 1991) but decreased with NO2 concentrations of 5,000 ppb and
higher (Parker etal.. 1989; Jakab. 1988; Goldstein et al.. 1973). Concentration-dependent
decreases in bactericidal activity were observed but only with higher than
ambient-relevant NO2 exposures (Jakab. 1988; Goldstein et al.. 1974; Goldstein etal..
1973). For NCh concentrations in the range of 1,000 to 2,500 ppb, bacterial clearance was
reduced only with 17- to 48-hour exposures (Sherwood et al.. 1981; Goldstein et al..
1973) or in immunocompromised mice (Jakab. 1988). NC>2-induced reductions in
bactericidal activity and bacterial clearance also were observed in the lungs of mice
strains known to be free of all mice pathogens (Davis etal.. 1991; Parker etal.. 1989). As
with mortality, the effects of NC>2 on bacterial clearance and killing varied by infectious
organism. Increased bacterial proliferation was observed in NO2-exposed (1,000 ppb, 24
or 48 hours) mice after infection with the virulent group C Streptococci but not
5-146
-------
Staphylococcus aureus (Sherwood et al., 1981). A 4-hour exposure to 5,000 ppb NCh
resulted in a decrease in bactericidal activity after challenge with Staphylococcus aureus;
however, NO2 exposure at concentrations less than 20,000 ppb did not affect killing of
Proteus mirabilis or Pasteurellapneumotropica (Jakab. 1988).
Compared with animal toxicological studies, controlled human exposure studies are few
in number and do not provide evidence for NC>2-induced infectivity. Controlled human
exposure studies also differed from toxicological studies in examining infectivity in
response to viral not bacterial exposure. NO2 exposures of 600 ppb did not lead to
inactivation of live influenza virus in AMs or bronchial epithelial cells collected from
adults or alteration in titers of influenza or respiratory syncytial virus (Frampton etal..
2002; Frampton et al.. 1989). Frampton et al. (1989) reported a trend of decreased
inactivation of influenza virus, although results were not statistically significant.
Repeated exposures to higher NO2 concentrations of 1,000 to 3,000 ppb also did not
affect inactivation of administered influenza, whether NC>2 exposure was repeated over a
single day with a lower background NCh exposure (Frampton et al.. 1989) or over 3
consecutive days (Goings et al.. 1989). The 3-day exposure study lacked a sham control;
thus, results have weaker implications.
5.2.5.2 Respiratory Infections Reported or Diagnosed in Children
In contrast with findings in mouse models, epidemiologic evidence does not clearly
indicate a relationship between short-term NO2 exposure and respiratory infection in
children (ages 0-15 years) as reported by self or parents or more objectively ascertained
as laboratory-confirmed or physician-diagnosed cases. Some studies found associations
(Espositoetal.. 2014; Luetal.. 2014; Stern etal.. 2013; Ghosh etal.. 2012b; Just et al..
2002); others did not find associations or found inconsistent associations among the
outcomes examined [(Altug etal.. 2014; Stern etal., 2013; Xuet al., 2013); Table 5-231.
Xuet al. (2013) did not provide strong evidence for an association of ambient NO2 with
laboratory-confirmed cases of influenza (RR: 1.01 [95% CI: 0.97, 1.04] for an unreported
increment in NCh); however, study limitations preclude strong inferences from the
results. There were a mean of only two influenza cases per day, and potential collinearity
in a multipollutant model with PMio and Os (Spearman r for correlation with NC>2 = 0.62
and -0.42, respectively) limits inference about NCh effects.
Results indicating associations between ambient NO2 or NOx and respiratory infections
also have weak implications (Esposito et al.. 2014; Lu etal.. 2014; Stern etal.. 2013;
Ghosh et al., 2012b; Just etal., 2002). All of these studies assigned exposure from central
site concentrations (one city site or average of multiple sites) (Table 5-23). None reported
5-147
-------
information on the spatial distribution of subjects around the monitoring site(s) or the
within-city variability in NC>2 or NOx concentrations to ascertain potential exposure
measurement error and its impact on effect estimates. A few studies examined more
spatially resolved exposure metrics but also have uncertain implications. Ghosh et al.
(2012b) reported similar results in analyses restricted to homes for which central site
NOx better represented exposure but did not report how these homes were selected.
Further, the adequacy of NOx to serve as an indicator of NO2 could vary among subjects
because of varying NO2 to NOx ratios across locations (Section 2.5.3). NO2 at the central
site nearest to schools was associated with pneumonia prevalence (Lu et al., 2014).
However, pneumonia was ascertained as "ever having a diagnosis" and may not be
temporally matched to exposure assessed for a 3-year period. The only study that
measured NO2 in subjects' location (i.e., school) did not observe associations with colds
(Altug et al., 2014). The importance of the microenvironmental measures in this study is
underscored by the variability in traffic volume and road length reported within the study
area.
Uncertainty regarding confounding by PM2 5 and traffic-related copollutants also limits
any strong inferences from the results of these studies. In addition to NO2, respiratory
infections were associated with BS and PM25 Other copollutants, such as PMio, SO2, and
Os, also were associated with respiratory infection (Table 5-23). Studies did not examine
other traffic-related PM components, CO, copollutant models, or other methods to assess
the independent or mixture effects of NO2. Where reported, NO2 was moderately to
highly correlated with copollutants (r = 0.92 for BS; 0.6-0.8 for unspecified copollutants)
(Ghosh et al.. 2012b: Just et al.. 2002).
5-148
-------
Table 5-23 Epidemiologic studies of respiratory infections reported or
diagnosed in children.
Study
Population Examined and
Methodological Details
Oxide of Nitrogen
Metrics Analyzed
Effect Estimate (95% Cl)
Single-Pollutant Model3
Copollutant
Examination
tAltuqetal. (2014)
Eskisehir, Turkey, Feb-Mar, 2007
n =605, ages 9-1 Syr
Cross-sectional. Self-reported
respiratory infections. Recruitment
from schools of participants of a larger
study. Participation rate not reported.
Logistic regression adjusted for sex,
age, asthma, parental smoking, coal or
wood stove use, parental education,
height, weight, daily average
temperature.
NO2-outdoor school
24-h avg, lag 0-6 day avg
1 site at each of
16 schools
Means & max(ppb)
Suburban: 9.4, 13
Urban: 13, 18
Urban-traffic: 21, 28
Common cold last 7 days:
OR: 1.86(0.41, 8.42)
Common cold at the
moment:
OR: 4.59 (0.79, 26)
No copollutant
model.
03 associated with
colds. Strong
inverse correlation
with NO2. Pearson
r=-0.80.
NO2 and PlVh.s
reported to be
highly correlated.
tEsposito et al. (2014)
Milan, Italy, Jan-Dec 2012
n = 718, ages 2-18 yr, 329 with
wheeze or asthma, 389 healthy
children
Repeated measures. Daily symptom
diaries for 12 mo. Diaries checked
weekly, clinic visits conducted every
2 mo. Recruited from respiratory
disease section (wheeze/asthma) and
outpatient surgery (healthy) sections of
pediatric clinic. 89% follow-up
participation. Followed cohort similar to
cohort at baseline. GEE adjusted for
age, sex, siblings, parental education,
smokers in home, season, day of
week, temperature, humidity.
NO2-central site
1-h max, Lag 0-2 day avg
8 city sites, 7 surrounding
area
Weighted avg at
municipality level
Tertiles (T) in ppb
1: <47.3b
2: 47.3-60.1b
3: >60.1b
RR for pneumonia with T1
as reference
Children with asthma:
T2: 1.20(0.75, 1.90)
T3: 1.56(1.01,2.42)
Healthy children:
T2: 1.45(0.80,0.63)
T3: 1.02(0.93, 1.12)
No copollutant
model.
PM-io associated
with pneumonia.
Correlations NR.
Just et al. (2002)
Paris, France, Apr-Jun 1996
n = 82, ages 7-15 yr, children with
asthma, 90% atopy
Repeated measures. Daily symptom
diaries for 3 mo, collected weekly.
Recruitment from hospital outpatients.
82% follow-up participation. GEE
adjusted for time trend, day of week,
pollen, temperature, humidity.
NO2-central site
24-h avg, lag 0 day
Average of 11 sites
Mean: 28.6 ppbb
Max: 59.0 ppbb
Respiratory infection:
OR: 7.19(2.53,20.4)
No copollutant
model.
BS associated
with cough and
infection. High
correlation with
NO2. Pearson
r=0.92.
5-149
-------
Table 5-23 (Continued): Epidemiologic studies of respiratory infections reported
or diagnosed in children.
Study
Population Examined and
Methodological Details
Oxide of Nitrogen
Metrics Analyzed
Effect Estimate (95% Cl)
Single-Pollutant Model3
Copollutant
Examination
tSternetal. (2013)
Bern, Basel, Switzerland, Apr
1999-Feb2011
n = 366, ages 0-1 yr
Repeated measures. Symptoms
reported weekly by telephone for 1 yr.
Recruitment from birth cohort. 95%
follow-up participation. GAM adjusted
for study week, sex, siblings, nursery
care, prenatal maternal smoking,
postnatal maternal smoking, birth
weight, maternal atopy, parental
education.
NO2-central site
2-h avg, lag 5-11 day avg
2 sites
Rural mean: 8.1 ppbb
Urban mean: 25.6 ppbb
Upper percentiles NR
Incidence respiratory tract
infection:
RR: 1.20(0.82, 1.75)
Days with respiratory tract
infection:
NO2 < 26 ppb:
reference category
NO2 > 26 ppb:
18%(0, 39)
No copollutant
model.
PM-io lag 7 days
associated with
respiratory
infection.
Correlation NR.
tXuetal. (2013)
Brisbane, Australia, 2001-2008,
winters only
n = 2,922 influenza cases, ages
0-14 yr
Time-series. Laboratory-confirmed
cases of influenza referred by public or
private health sector. Only mean
2/day. No information available on
subjects' residential location. Poisson
regression adjusted for lag 0-9 day
avg temperature, lag 0-9 day avg
PM-io, lag 0-9 day avg Os, PM-io-
temperature interaction.
NO2-central site
24-h avg, lag 0-9 day avg
# sites NR
Mean: 5.9 ppbb
75th: 7.3 ppbb
Max: 13.3 ppbb
Daily influenza counts:
RR: 1.01 (0.97, 1.04)
increment of NO2 NR.
Results are presented
only for a multipollutant
model that also includes
PM-ioand Os.
Only multipollutant
model analyzed.
tGhoshetal. (2012b)
Teplice and Prachatice, Czech
Republic, May 1994-Jun 2003
n = 1,113 children, ages 0-4.5 yr
Repeated measures.
Physician-diagnosed infections
between ages 0-4.5 yr ascertained
from medical records. Recruitment
from birth cohort. Participation rate not
reported. GEE with exchangeable
correlation and adjusted for city, year
of birth, day of week, fuel used for
heating and cooking, season, 7-day
avg temperature. Restricted analyses
to children for whom central site may
better represent exposure (method not
reported).
NOx-central site
24-h avg, lag 0-2 day avg
2 sites
Mean, 75th (ug/m3)
Teplice: 59.2, 73.3
Prachatice: 20.3, 24.4
Acute bronchitis:
Birth to age 2 yr:
RR: 1.09(1.01, 1.16) per
35 ug/m3 NOx
Age 2 yr-4.5 yr:
RR: 1.05(0.94, 1.14) per
35 ug/m3 NOx
No copollutant
model.
Association with
PM2.5 reported in
separate paper.
Moderate to high
correlations
reported with
unspecified
copollutants.
r= 0.6-0.8
5-150
-------
Table 5-23 (Continued): Epidemiologic studies of respiratory infections reported
or diagnosed in children.
Study
Population Examined and
Methodological Details
Oxide of Nitrogen
Metrics Analyzed
Effect Estimate (95% Cl)
Single-Pollutant Model3
Copollutant
Examination
tLuetal. (2014)b
Changsha, China, Sep 2011-Jan 2012
n = 2706, ages 3~6 yr
Cross-sectional. Recruitment from
schools. 59% participation. Potential
temporal mismatch of exposure
(2008-2011) and ever having
pneumonia diagnosis. Two-level
model. Pneumonia first adjusted for
parental atopy, antibiotic use, new
furniture in home, coal, wood or gas
used in home, painted walls/air
conditioning in home, pets in home,
visible mold/dampness in home.
Adjusted pneumonia prevalence
regressed with NO2. Confounding by
meteorological factors not examined.
NO2-central site
24-h avg
Nearest to school,
distance NR
Concentrations NR
2008-2011:
7 days >63.8 ppbb
standard before 2012
89 days >42.6 ppbb
standard 2012
OR for NO2 > 63.8 ppb
1 day/yr
1.04(1.02, 1.05)
No copollutant
model.
PM-ioand SO2
associated with
pneumonia.
Correlations NR.
Note: More informative studies in terms of the exposure assessment method and potential confounding considered are presented
first.
avg = average; BS = black smoke; Cl = confidence interval; Dec = December; Feb = February; GAM = generalized additive model;
GEE = generalized estimating equations; NO2 = nitrogen dioxide; NOX = sum of NO and NO2; NR = not reported; O3 = ozone;
OR = odds ratio; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm;
PM10 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; ppb = parts per billion;
RR = relative risk; SO2 = sulfur dioxide; yr = year.
aEffect estimates are standardized to a 20 ppb for 24-h avg NO2. NOX effect estimates are presented as reported in the study
(Section 5.1.2.2).
"•Concentrations converted from |jg/m3 to ppb using the conversion factor of 0.532 assuming standard temperature (25°C) and
pressure (1 atm).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-151
-------
5.2.5.3 Hospital Admissions and Emergency Department Visits for Respiratory
Infections
To date, relatively few studies have examined the association between short-term NO2
exposures and hospital admissions and ED visits due to respiratory infections. The 2008
ISA for Oxides of Nitrogen identified studies that evaluated a number of respiratory
infection outcomes, such as upper respiratory infections (URIs), pneumonia, bronchitis,
allergic rhinitis, and lower respiratory disease. Across these outcomes, studies have
generally not provided consistent evidence of an association between NO2 and hospital
admissions and ED visits due to respiratory infections (U.S. EPA. 2008c). Recent studies
add to the body of literature evaluated in the 2008 ISA for Oxides of Nitrogen, but
compared to other respiratory-related hospital admission and ED visit outcomes the total
body of literature remains limited. The air quality characteristics of the city, or across all
cities, and the exposure assignment approach used in each respiratory infection-related
hospital admission and ED visit study evaluated in this section are presented in
Table 5-24. As detailed in Section 5.2.2.4. other recent studies of respiratory
infection-related hospital admissions and ED visits are not the focus of this evaluation,
and the full list of these studies, as well as study details, can be found in Supplemental
Table S5-3 (U.S. EPA. 2015h).
5-152
-------
Table 5-24 Mean and upper percentile concentrations of nitrogen dioxide in studies of hospital admissions and
emergency department visits for respiratory infection.
Study
Location
Years
Type of Visit
(ICD9/10) Exposure Assignment
Mean Upper Percentile
Concentration Concentrations Copollutant
Metric ppb ppb Examination
Hospital Admissions
Burnett et al.
(1999)
Lin et al.
(2005)
Karr et al.
(2006)
Toronto, ON,
Canada
(1980-1994)
Toronto, ON,
Canada
(1998-2001)
Southern Los
Angeles County,
CA
(1995-2000)
Respiratory Average of NO2 concentrations
infection (464, across 4 monitors.
466, 480-7, 494)
Respiratory Average of NO2 concentrations
infection (464, across 7 monitors.
466, 480-487)
Acute bronchiolitis 34 NO2 monitors, exposure
(466.1) assigned based on nearest
monitor to residential ZIP code.
24-h avg 25.2 NR Correlations (r):
PlVh.s: 0.55
PMio-2.s: 0.38
PMio: 0.57
CO: 0.64
SO2: 0.54
O3: -0.08
Copollutant
models: none
24-h avg 25.5 75th: 29.3 Correlations (r):
PM2.5: 0.48
PMio-2.s: 0.40
PMio: 0.54
CO: 0.20
SO2: 0.61
O3: 0.0
Copollutant
models: none
1-h max 59 75th: 69 Correlations (r): NR
90th: 90 Copollutant
models: none
5-153
-------
Table 5-24 (Continued): Mean and upper percentile concentrations of nitrogen dioxide in studies of hospital
admissions and emergency department visits for respiratory infection.
Study
Zanobetti and
Schwartz
(2006)
tHEl (2012)
fMehta et al.
(2013)
tSeqala et al.
(2008)
Location
Years
Boston, MA
(1995-1999)
Ho Chi Minh
City, Vietnam
(2003-
2005)
Paris, France
(1997-2001)
Type of Visit
(ICD9/10) Exposure Assignment
Pneumonia Average of NO2 concentrations
(480-487) across 5 monitors.
Acute lower Average of NO2 concentrations
respiratory across 9 monitors.
infection (J13-16,
18, 21)
Bronchiolitis Average of NO2 concentrations
from 21 monitors, representative
of urban background.
Mean Upper Percentile
Concentration Concentrations Copollutant
Metric ppb ppb Examination
24-h avg NR 50th: 23.2 Correlations (r):
PlVh.s: 0.55
BC: 0.70
CO: 0.67
O3: -0.14
PM nontraffic: 0.14
Copollutant
models: none
24-h avg 11.7 Max: 29.2 Correlations (r):
Dry season:
PMio: 0.78
O3: 0.44
SO2: 0.29
Rainy season:
PMio: 0.18
O3: 0.17
SO2: 0.01
Copollutant
models: SO2, PMio,
03
24-h avg 27.0 Max: 90.4 Correlations (r):
BS: 0.83
PMio: 0.74
SO2: 0.78
Copollutant
models: none
5-154
-------
Table 5-24 (Continued): Mean and upper percentile concentrations of nitrogen dioxide in studies of hospital
admissions and emergency department visits for respiratory infection.
Location Type of Visit
Study Years (ICD9/10)
fFaustini et 6 Italian cities LRTI (466,
al. (2013) (2001-2005) 480-487)
Mean Upper Percentile
Concentration Concentrations Copollutant
Exposure Assignment Metric ppb ppb Examination
Average of NO2 concentrations 24-h avg 24.1-34.6
over all monitors within each city.
Number of NO2 monitors in each
city ranged from 1-5.a
NR Correlations (r),
across cities:
PMio: 0.22-0.79
Copollutant
models: PMio
ED Visits
Peel et al.
(2005)
Atlanta, GA
(1993-2000)
Upper respiratory
infection (460-
6,477)
Pneumonia
(480-486)
Average of NO2 concentrations 1-h max 45.9
from monitors for several
monitoring networks.
NR Correlations (r):
PIvh.s: 0.46
PMio: 0.49
PMio-2.s: 0.46
UFP: 0.26
PIvh.s water-soluble
metals: 0.32
PIvh.s sulfate: 0.17
PIvh.s acidity: 0.10
PIvh.s OC: 0.63
PIvh.s EC: 0.61
Oxygenated HCs:
0.30
O3: 0.42
CO: 0.68
SO2: 0.34
Copollutant
models: none
5-155
-------
Table 5-24 (Continued): Mean and upper percentile concentrations of nitrogen dioxide in studies of hospital
admissions and emergency department visits for respiratory infection.
Study
tStieb et al.
(2009)
tSeqala et al.
(2008)
fZemek et al.
(2010)
Location
Years
7 Canadian
cities
(1992-2003)
Paris, France
(1997-2001)
Edmonton, AB,
Canada
(1992-2002)
Type of Visit
(ICD9/10)
Respiratory
infection (464,
466, 480-487)
Bronchiolitis
Otitis media
(382.9)
Exposure Assignment Metric
Average NO2 concentrations from 24-h avg
all monitors in each city. Number
of NO2 monitors in each city
ranged from 1-14.
Average of NO2 concentrations 24-h avg
across 21 monitors,
representative of urban
background.
Average of NO2 concentrations 24-h avg
across 3 monitors.
Mean
Concentration
ppb
9.3-22.7
27.0
21.9
Upper Percentile
Concentrations Copollutant
ppb Examination
75th: 12.3-27.6 Correlations (r)
reported by city
season.
Copollutant
models: none
Max: 90.4 Correlations (r):
BS: 0.83
PMio: 0.74
SO2: 0.78
Copollutant
models: none
75th: 27.6 Correlations (r):
Copollutant
models: none
only
and
NR
Physician Visits
fSinclair et
al. (2010)
Atlanta, GA
Upper respiratory
infection
Lower respiratory
infection
NO2 concentrations collected as 1-h max
part of AIRES at SEARCH
Jefferson Street site.
1998-2000:
49.8
2000-2002:
35.5
1998-2002:
41.7
NR Correlations (r):
Copollutant
models: none
NR
AB = Alberta; AIRES = Aerosol Research Inhalation Epidemiology Study; avg = average; BC = black carbon; BS = black smoke; CA = California; CO = carbon monoxide;
EC = elemental carbon; ED = emergency department; GA = Georgia; HC = hydrocarbons; LRTI = lower respiratory tract infection; MA = Massachusetts; Ml = Michigan;
NO2 = nitrogen dioxide; NR = not reported; NY = New York; O3 = ozone; OC = organic carbon; ON = Ontario; PM2 5 = particulate matter with a nominal mean aerodynamic diameter
less than or equal to 2.5 |jm; PMio = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; SO2 = sulfur dioxide; UFP = ultrafine particles.
aMonitoring information obtained from Colais et al. (2012).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-156
-------
Hospital Admissions
Few recent studies have examined the association between short-term NCh exposures and
respiratory infection hospital admissions. A time-series study conducted in Ho Chi Minh
City, Vietnam (Mehtaet al.. 2013; HEI. 2012) examined the association between
short-term air pollution exposures and pediatric (ages 28 days-5 years) hospital
admissions for acute lower respiratory infections (ALRI, including bronchiolitis and
pneumonia). In a time-stratified case-crossover analysis focused only on the average of a
1-6 day lag, there was no evidence of an association between NO2 and ALRI hospital
admissions in the all-year analysis (-4.0% [95% CI: -18.0, 12.5] for a 20-ppb increase in
24-h avg NO2 concentrations).
In an additional study that also examined respiratory infections (i.e., bronchiolitis) in
children, Segalaetal. (2008) focused on associations with winter (October-January) air
pollution because it is the time of year when respiratory syncytial virus (RSV) activity
peaks. It has been hypothesized that air pollution exposures may increase the risk of
respiratory infections, including bronchiolitis due to RSV (Segalaetal.. 2008). Focusing
on children <3 years of age in Paris, France, the authors conducted a bidirectional
case-crossover analysis along with a time-series analysis to examine air pollution
(i.e., PMio, BS, NC>2, SO2) associations with bronchiolitis ED visits and hospital
admissions. Although the authors specify the bidirectional case-crossover approach was
used to "avoid time-trend bias," other have shown the bidirectional approach to bias
results (Segala et al.. 2008; Levy et al.. 2001). In the case-crossover analysis, NO2 was
associated with bronchiolitis hospital admissions (15.9% [95% CI: 7.7, 29.0], lag
0-4 days for a 20-ppb increase in 24-h avg NC>2 concentrations); NCh was not examined
in the time-series analysis. Although a positive association was observed, the authors did
not analyze copollutant models. The lack of copollutant analyses complicates the
interpretation of these results because the pollutants were highly correlated, ranging from
an r = 0.74 to 0.83.
Faustini et al. (2013). in the analysis of air pollution in six Italian cities, also examined
associations with lower respiratory tract infection (LRTI) hospital admissions. However,
the authors only focused on LRTIs in individuals with COPD over the age of 35 years. In
this population, the largest associations were observed at lag 2-5 days (10.0% [95% CI:
-2.7, 24.3]), and there was no evidence of an immediate effect. This is in contrast to the
results for COPD hospital admissions, where the strongest associations were observed at
lag 0 and 0-1 days. The authors examined the NO2 association with LRTI hospital
admissions in copollutant models with PMio at lag 0-5 days, and consistent with the
5-157
-------
other endpoints examined, reported that the association remained positive but was
attenuated (7.8% [95% CI: -6.5, 24.2]).
Emergency Department Visits
Studies that examined the effect of air pollution on ED visits attributed to respiratory
infections have focused on similar outcomes to those examined in the studies of hospital
admissions. Stieb et al. (2009). in their study of seven Canadian cities, also examined the
association between short-term NCh concentrations and respiratory infection ED visits.
The authors reported positive associations at lags of 1 and 2 days, but the confidence
intervals were wide, providing little evidence of an association. However, Segala et al.
(2008) in the study of winter (October-January) air pollution in Paris, France (discussed
above) reported evidence of an association between short-term NCh concentrations and
bronchiolitis ED visits (11.8% [95% CI: 7.7, 20.1]; lag 0-4 day avg for a 20-ppb increase
in 24-h avg NC>2 concentrations) in a bidirectional case-crossover analysis. As mentioned
previously, the interpretation of these results is complicated by the lack of copollutant
analyses and the high correlation between pollutants examined (r = 0.74 to 0.83).
In an additional study conducted in Edmonton, Alberta, Canada, Zemek et al. (2010)
examined otitis media (i.e., ear infections) ED visits, for ages 1-3 years. Associations
were examined for single-day lags of 0 to 4 days in all-year as well as seasonal analyses.
The authors observed that NO2 was associated with increases in ED visits for otitis media
in the all-year analysis at lag 2 days (7.9% [95% CI: 1.6, 12.8] for a 20-ppb increase in
24-h avg NC>2 concentrations). When examining whether there was evidence of seasonal
patterns in associations, the authors found that the magnitude of the association was
larger in the warm months of April-September (16.1% [95% CI: 3.1, 31.2]) compared to
the cold months of October-March (4.7% [95% CI: 0, 11.2]) at lag 2 days for a 20-ppb
increase in 24-h avg NO2 concentrations. Importantly, the pattern of associations for CO
were similar to that observed for NO2, and the authors did not report correlations between
pollutants or analyze copollutant models.
Outpatient and Physician Visits
In addition to examining severe occurrences of a respiratory infection that would require
a trip to a hospital, studies have begun to explore whether air pollution may lead to less
severe cases, which would be indicated by trips to an outpatient facility. In a study
conducted in Atlanta, GA, Sinclair et al. (2010) examined respiratory infection
(e.g., upper respiratory infections, lower respiratory infections) outpatient visits to a
managed care organization. As detailed in Section 5.2.2.4, the authors separated the
analysis into two time periods to compare air pollutant concentrations and relationships
5-158
-------
for acute respiratory visits. The two time periods were the 25-month time period
examined in Sinclair and Tolsma (2004) and an additional 28-month time period of
available data from AIRES. Across the two time periods, mean 1-h max NO2
concentrations were lower in the 28-month versus the 25-month time period, 49.8 ppb
versus 35.5 ppb, respectively (Table 5-24). For both outcomes, the daily number of
outpatient visit counts varied, with LRI being rather small (i.e., 12 per day) compared to
that for URI (i.e., 263 per day). A comparison of the two time periods indicated that risk
estimates for LRI and URI tended to be larger in the earlier 25-month period compared to
the later 28-month period with relatively wide confidence intervals for both outcomes.
Additionally, the lag structure of associations varied between each time period. For LRI,
the largest magnitude of an association was for both lag 0-2 and 3-5 day avg in the
earlier time period, but only lag 3-5 day avg in the latter time period. For URI, the largest
associations were for lag 0-2 and 3-5 days for the earlier time period, but a positive
association was only observed for lag 6-8 days in the latter time period. The authors also
examined potential seasonal differences in associations, but the inconsistent results
between the two time periods with respect to the lag structure of associations complicates
the interpretation of seasonal results.
Summary of Respiratory Infection Hospital Admissions and Emergency
Department Visits
Recent studies that examined the association between short-term NO2 exposure and
hospital admissions and ED visits due to respiratory infections add to the body of
evidence detailed in the 2008 ISA for Oxides of Nitrogen, but studies have not
consistently examined similar respiratory infection outcomes (Figure 5-9 and
Table 5-25). Of the studies evaluated, the strongest associations are for studies that
focused on children, specifically less than 5 years of age. These studies demonstrate
associations with respiratory infection, bronchiolitis, and otitis media, specifically during
certain times of the year depending on geographic location. The relatively small number
of studies that have examined hospital admissions and ED visits due to respiratory
infections has resulted in an inadequate assessment of the lag structure of associations
and potential copollutant confounding.
5-159
-------
Study
Mehtaetal. (2013)
Lin et al. (2005)
Faustini et al. (2013)b
Burnett et al. (1999)
Karr et al. (2006)
Segala et al. (2008)
Zanobetti and Schwartz (2006)
Stieb et al. (2009)
Peel et al. (2005)
Segala et al. (2008)
Peel et al. (2005)
Zemeketal. (2010)
Location
Ho Chi Minh, Vietnam
Toronto, CAN
6 Italian citiss
Toronto, CAN
LA County, CA
Paris, France
Boston, MA
7 Canadian cities
Atlanta, GA
Paris, France
Atlanta, GA
Edmonton, Canada
Age
28 days - 5
n 1 A
U- I 'f
ric _i_
oO-t-
All
0-1
<3
65+
All
All
<3
All
1-3
Lag
"IP ^ ^^
1 fia
I -Od
n ^
u-o
Or- r-\|
2
1 -(D-
0-4
0-1 —
2 H
0-2
0-4
0-2 -<
2
Hospital Admissions
• t
^
.»
Bronchiolitis
(D
-• Pneumonia
ED Visits
J- Respiratory Infection
-9-
— (D Bronchiolitis
>— Pneumonia
• — Otitis Media
-10.0 0.0 10.0 20.0 30.0 40.0
% Increase (95% Cl)
50.0
Note: CA =California; CAN = Canada; Cl = confidence interval; DL = distributed lag; ED = emergency department; GA = Georgia;
MA = Massachusetts. Black = studies from the 2008 Integrated Science Assessment for Oxides of Nitrogen, red = recent studies.
Solid symbols = all year, horizontal stripes = warm/summer months, vertical stripes = cool/winter months, a = results are for the dry
season (November-April); b = Lower Respiratory Infection in people with chronic obstructive pulmonary disease. Effect estimates
are standardized to a 20-ppb increase in 24-h avg nitrogen dioxide and 30-ppb increase in 1-h max nitrogen dioxide.
Figure 5-10 Percentage increase in respiratory infection-related hospital
admissions and emergency department visits in relation to
nitrogen dioxide concentrations.
5-160
-------
Table 5-25 Corresponding risk estimate for studies presented in Figure 5-10.
Study
Location
Avg % Increase
Age Time Lag Season (95% Cl)
Hospital Admissions
Respiratory Infection
tMehtaetal. (2013)
Lin et al. (2005)
tFaustini et al. (201 3)b
Burnett et al. (1999)
Ho Chi Minh, Vietnam 28 days- 24-h avg 1-6 All
y
Dry3
Toronto, ON, Canada 0-14 24-h avg 0-5 All
6 Italian cities 35+ 24-h avg 0-5 DL All
Toronto, ON, Canada All 24-h avg 2 All
-4.0 (-18.0, 12.5)
35.9 (3.0, 79.3)
41.1 (15.6,73.7)
6.9 (-4.3, 19.4)
5.4 (3.5, 7.4)
Bronchiolitis
Karretal. (2006)
tSeqalaetal. (2008)
LA County, CA 0-1 1-h max 1 Winter
Paris, France <3 24-h avg 0-4 Winter
0 C / C Q -1 O\
15.9(7.7,29.0)
Pneumonia
Zanobetti and Schwartz
(2006)
Boston, MA
65+
24-h avg 0-1 All
2.7 (-3.0, 8.4)
ED Visits
Respiratory Infection
tStieb et al. (2009)
Peel et al. (2005)
7 Canadian cities
Atlanta, GA
All 24-h avg 2 All 0.7 (-1.5, 2.8)
All 1-h max 0-2 All 2.9(0.9,4.7)
Bronchiolitis
tSeqalaetal. (2008)
Paris, France
<3 24-h avg 0-4 Winter 11.8(7.7,20.1)
Pneumonia
Peel et al. (2005)
Atlanta, GA
All 1-h max 0-2 All 0.0 (-2.5, 2.9)
5-161
-------
Table5-25 (Continued): Corresponding risk estimate for studies presented in
Figure 5-10.
Study
Location
Avg
Age Time
% Increase
Lag Season (95% Cl)
Otitis Media
tZemeketal. (2010)
Edmonton, AB,
Canada
1 -3 24-h avg
2 All 7.9(1.6,12.8)
Summer 16.1 (3.1, 31.2)
Winter 4.7(0.0,11.2)
AB = Alberta; avg = average; CA = California; Cl = confidence interval; DL = distributed lag, ED = emergency department;
GA = Georgia; MA = Massachusetts; ON = Ontario.
aDry season was defined as November-April.
"LRTI in people with COPD.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5.2.5.4 Subclinical Effects Underlying Respiratory Infections
Overall, NO2 exposure has not been related with consistent changes in subclinical effects
that have been identified as key events in the proposed mode of action for respiratory
infections (Figure 4-1). with the direction of change varying across studies. Some support
for the effects of NO2 on respiratory infection morbidity and mortality observed in
toxicological studies and some epidemiologic studies is provided by toxicological
findings for NC>2-induced impairments in alveolar macrophage function. There is
uncertainty about the effects of NO2 on alveolar macrophages and immunoglobulin
antibody responses as examined in controlled human exposure and epidemiologic studies,
respectively.
Mucociliary and Alveolar Clearance
Airborne substances small enough to be respired may be trapped in the epithelial lining
fluid in the conducting airways and physically removed or cleared from the airway by
ciliated epithelial cells. This pulmonary clearance consists of mucociliary and alveolar
clearance. Recent animal toxicological studies and studies reviewed in the 2008 ISA for
Oxides of Nitrogen (U.S. EPA. 2008c) demonstrated that NO2 exposures higher than
5,000 ppb, which are above those considered ambient-relevant, functionally impair
pulmonary clearance and damage the ciliated epithelium of the airway. However, animal
toxicological and controlled human exposure studies with NO2 exposures of 5,000 ppb or
less, which were reviewed in the 2008 ISA, often show increased pulmonary clearance.
5-162
-------
Exposures of rabbits to 300 ppb NO2 for 2 hours, given as a single exposure or repeated
over 14 days [(Schlesingeretal.. 1987; Schlesinger and Gearhart. 1987; Vollmuth et al..
1986); Table 5-261 had no effect on mucociliary or alveolar clearance or accelerated
clearance as measured by retention of radioactive tracer microspheres. Results were
similar in rabbits and rats exposed to 1,000 ppb NC>2. Increased particle clearance was
observed after a single-day 2-hour exposure (Vollmuth et al.. 1986) and after repeated
2- or 7-hour exposures over 11 to 22 days (Schlesinger and Gearhart. 1987; Ferin and
Leach. 1975). Another study observed no change in clearance in rabbits exposed to
1,000 ppb NC>2 for 2 hours per day for 14 days (Schlesinger et al.. 1987).
An effect of NO2 exposure on pulmonary clearance also is unclear based on ciliary
activity measurements. The uncertainty is due largely to limitations in study design.
Guinea pigs exposed to 3,000 ppb NO2 for 6 hours/day and 6 days/week for 2 weeks had
concentration-dependent reductions in ciliary activity (Ohashi et al.. 1994). However,
ciliary beat (measured by light refraction) was examined in nasal tissues excised after
animals were exposed. This method could have affected the outcome and be less
representative of changes that occur from human ambient exposure. In a controlled
human exposure study of healthy adults, there was no ciliary activity 45 minutes after a
20-minute exposure to 1,500 or 3,500 ppb NCh (Helleday et al.. 1995). In contrast,
increases in ciliary activity were reported 24 hours after a 4-hour exposure to 3,500 ppb
NC>2. Importantly, baseline measurements for each subject were used as control values,
and therefore, the study lacked air controls and subject blinding.
5-163
-------
Table 5-26 Characteristics of experimental studies of subclinical lung host
defense effects.
Study
Species (Strain);
Sample Size;
Sex; Age
Exposure Details
Endpoints Examined
Mucociliary and alveolar clearance
Vollmuth et al.
(1986)
Rabbits (New Zealand
White); n = 5/group; M;
age NR
300, 1,000, or 3,000 ppb for 2 h Retained tracer particles for 14 days
following exposure.
Schlesinqer and
Gearhart(1987)
Schlesinqer et
al. (1987)
Rabbits (New Zealand
White); n = 5/group; M;
age NR
300 or 1,000 ppb for 2h/day up
to 14 days
Retained radioactively labeled tracer
particles after each of the 14 days of
NO2 exposure. Retention in alveolated
and tracheobronchial regions of lung.
Ferin and Leach
Rats (Long-Evans),
n = 5-10/group; sex
and age NR
1,000-24,000 ppb NO2 for 11
or 22 days (7 h/day,
5 days/week)
Retained titanium dioxide particles at 8,
25, and 130 days post-exposure.
Ohashi et al.
(1994)
Helleday et al.
(1995)
Guinea pigs (Hartley);
n = 10/group; sex and
age NR
Humans; n = 14 M,
10 F; 27 yr (range:
23-30 yr)
3,000 or 9,000 ppb for 6 h/day,
6 days/week for 2 weeks
(1) 1,500 ppb for 45 min
(2) 3,500 ppb for 45 min
(3) 3,500 ppb for 4 h
Baseline/control obtained
2 weeks prior.
Ciliary beat in excised nasal tissue 24 h
after exposure.
Fiberoptic bronchoscopy to record
mucociliary activity frequency.
(1) and (2) 45 min following exposure.
(3) 24 h following exposure.
Function and Morphology of Alveolar Macrophages
Goldstein et al. Rats 500, 1,000, or 2,400 ppb NO2
(1977) (Sprague-Dawley); n = for 1 and 2 h
NR; F; age NR
Agglutination of AMs.
Rombout et al. Rats (Wistar);
(1986) n = 3-6/group; F;
6 weeks
500, 1,390, or2,800ppbNO2
forl, 2, 4, 8, 16, and 28 days
Histopathological evaluation.
Mochitate et al.
Rats (Wistar);
n = 6/group; M;
19-23 weeks
4,000 ppb NO2 continuously up
to 10 days
BAL fluid cell counts and AM function
and morphology.
Suzuki et al.
Rats (Fischer 344);
n = 8/group; M;
7 weeks
4,000 NO2 ppb for 1, 3, 5, 7,
and 10 days
AM activity (phagocytosis and
superoxide production), SOD and
glucose-6-phosphate dehydrogenase
activity.
Hooftman et al.
Rats (Wistar);
n = 10/group; M; age
NR
4,000, 10,000, or 25,000 ppb
NO2 for 6 h/day, 5 days/week
for 7-21 days
Histopathological evaluation, analysis
of BAL fluid, and AM function and
morphology.
5-164
-------
Table 5-26 (Continued): Characteristics of experimental studies of subclinical
lung host defense effects.
Study
Azadnivet al.
(1998)
Dowell et al.
(1971)
Robison et al.
(1990)
Robison et al.
(1993)
Amoruso et al.
(1981)
Devlin et al.
(1999)
Schlesinqer
(1987b)
Rose et al.
(1988)
Rose et al.
(1989a)
Species (Strain);
Sample Size;
Sex; Age
Humans; n = 11 M,
4F;
Early phase:
28.1 ±3.5yr
Late phase:
27.4 ± 4.2 yr
Dogs (beagle);
n = 11; sex and age
NR
Rats
(Sprague-Dawley); n,
sex, and age NR
Rats (Sprague
Dawley); n > 4/group;
sex and age NR
Rats (Sprague-
Dawley); n = 4/group
F; age NR
Humans; n = 11 M;
range: 18-35 yr
Rabbits (New Zealand
White); n = 5/group
M; age NR
Mice (CD-1 );
n > 4/group; sex NR;
4-6 weeks
Exposure Details
2,000 ppb for 6 h;
Exercise for approximately 10
of every 30 min at
VE = 40 L/min
3,000ppb NO2for1 h
1 00, 500, or 1 ,000 ppb NO2 for
1 h; AMs exposed ex vivo
500 ppb NO2 for 8 h/day for
0.5, 1, 5, or 10 days
1,300, 1,900, or6,100 ppb NO2
forSh
2,000 ppb for 4 h;
Exercise for 15 min on/15 min
off at VE = 50 L/min
310 or 1,030 ppb NO2 for
2 h/day for 2, 6, and 1 3 days
(1)1,000,2,500, or 5,000 ppb
NO2 for 6 h/day for 2 days;
intratracheal inoculation with
murine Cytomegalovirus',
4 additional days (6 h/day) of
exposure.
(2) re-inoculation 30 days (air)
post-primary inoculation.
Endpoints Examined
Alveolar macrophage function 1 h (early
phase) and 18 h (late phase) after
exposure.
Histopathological evaluation and lung
surfactant properties.
Viability, LTB4 production, neutrophil
chemotaxis, superoxide production.
BAL fluid cell counts and arachidonate
metabolite levels, AM arachidonate
metabolism, respiratory burst activity,
and glutathione content.
Analysis of BAL fluid and superoxide
production by AMs (PMA stimulation).
BAL fluid macrophage superoxide
production and phagocytosis.
Viability and AM activity (mobility,
attachment, and phagocytosis).
Infection 5 and 10 days
post-inoculation, histopathological
evaluation, and analysis of BAL fluid
(LDH, albumin, macrophages).
Pinkston et al.
Human AMs isolated
from 14 M and 1 F;
29±3.9yr
5,000 ppb for 3 h (ex vivo)
Cell viability and release of neutrophil
chemotactic factor and IL-1.
5-165
-------
Table 5-26 (Continued): Characteristics of experimental studies of subclinical
lung host defense effects.
Study
Species (Strain);
Sample Size;
Sex; Age
Exposure Details
Endpoints Examined
Other Subclinical Effects
Frampton et al.
(2002)
Pathmanathan
et al. (2003)
Humans;
(1,2) n = 12 M, 9F;
F = 27.1 ±4.1 yr
M = 26.9 ± 4.5 yr
Humans; n = 8 M, 4 F
26 yr (range: 21-32)
(1)600ppbfor3h,
(2)1,500ppbfor3h;
(1,2) Exercise 10 min
on/20 min off at VE = 40 L/min
2,000 ppbfor4 h/day for
4 days;
Bronchial and alveolar lavage fluid cell
viability and differential counts 3.5 h
post-exposure, influenza and RSV
challenge in BAL cells, peripheral blood
characterization.
Biomarkers in bronchial epithelium-
exotoxin, GM-CSF, Gro-a, I-CAM 1,
Exercise 15 min on/15 min off
at 75 watts
IL-5, IL-6, IL-8, IL-10, IL-13, total and
active NF-K(3, and TNF-a (fiberoptic
bronchoscopy after end of last
exposure).
AM = alveolar macrophages; BAL = bronchoalveolar lavage; F = female; GM-CSF = granulocyte macrophage-colony stimulating
factor; h = hour; I-CAM = intercellular adhesion molecule; IL = interleukin; L/min = liters per minute; LDH = lactate dehydrogenase;
M = male; min = minute; NF-K(3 = nuclear factor kappa-light chain-enhancer of activated B cells; NO2 = nitrogen dioxide; NR = not
reported; PMA = phorbol myristate acetate; ppb = parts per billion; RSV = respiratory syncitial virus; SD = standard deviation;
SOD = superoxide dismutase; TNF-a = tumor necrosis factor; yr = years.
Function and Morphology of Alveolar Macrophages
NO2 exposure shows variable effects on AM numbers, morphology, and superoxide
production, but there is indication of effects on AM phagocytosis. The inconsistencies
present across studies could be the result of strain or sex differences in response to NO2
or reflect true uncertainty. In limited examination, 500 ppb NC>2 exposures of rats (see
Table 5-26 for study details) were demonstrated to decrease AM agglutination (Goldstein
et al.. 1977) but not to increase AM numbers in BAL fluid (Rombout et al.. 1986). NC>2
exposures of 4,000 ppb NCh (repeated over 7-21 days) tended to increase numbers of
AM in BAL from rats (Mochitate etal.. 1986; Suzuki etal.. 1986) but not in all studies
(Hooftman et al.. 1988). Changes in AM morphology were not observed in human
exposed to 2,000 ppb NO2 (Azadniv et al.. 1998) but were observed in dogs and rats
exposed to 3,000 or 4,000 ppb NO2 (Hooftman etal.. 1988; Powell etal.. 1971).
NO2 exposure was shown to decrease the ability of AMs to produce superoxide anion
(indicating reduced respiratory burst) but not in all studies. NC>2 exposures of 100 or
500 ppb (details in Table 5-26) did not consistently decrease superoxide production in
AM isolated from rats (Robison et al.. 1993; Robison etal.. 1990). although a decrease
was observed after 500 ppb NO2 exposure that was repeated up to 10 days. NO2
exposures in the range of 1,000 to 4,000 ppb tended to decrease superoxide anion
5-166
-------
production in AMs from male and female rats (Robison et al.. 1990; Suzuki et al.. 1986).
with Robison etal. (1990) demonstrating a concentration-dependent decrease following a
1-hour exposure to NC>2 at concentrations of 1,000 ppb and higher. Conversely, AMs
isolated from Sprague Dawley female rats showed no change in superoxide production in
response to 1,300 or 1,900 ppb NCh exposure (Amoruso et al.. 1981). Similar to animal
toxicological studies, controlled human exposure studies did not demonstrate a consistent
effect of ambient-relevant NO2 exposures on superoxide production from AM, with 2,000
ppb NCh exposures (Table 5-26) of healthy subjects with intermittent exercise leading to
decreased production (Devlin et al.. 1999) or increased release (Azadniv et al.. 1998)
from AMs.
There is more consistent evidence for ambient-relevant NO2 exposures to decrease
phagocytic capacity of AMs. Such effects were observed in AMs from rabbits exposed to
300 ppb NCh for 2 or 6 days (Schlesinger. 1987b). However, all animals were co-exposed
to 0.5 mg/m3 sulfuric acid, and an independent effect of NO2 exposure could not be
assessed. Decreased phagocytosis was observed in AMs of rabbits, rodents (Rose et al..
1989a: Schlesinger. 1987b: Suzuki etal.. 1986). and humans (Devlin et al.. 1999)
following exposures of 1,000 to 5,000 ppb NCh for 1 to 7 days. However, NCh exposure
increased uptake of murine Cytomegalovirus in AMs from rats (Rose et al.. 1989a). In
contrast, Hooftman et al. (1988) found no changes in phagocytosis of latex microspheres
by AMs from rats below 10,000 ppb NCh at 1, 2, or 3 weeks. In vitro exposure of human
AMs to 5,000 ppb NCh for 3 hours did not result in statistically significant changes in
release of neutrophil chemotactic factor (IL-8) or IL-1 or changes in markers of
macrophage activity (Pinkston et al.. 1988).
Other Subclinical Effects
As examined in a few studies of humans, bronchial epithelial cells showed increased
virus-induced cytotoxicity as measured by lactate dehydrogenase (LDH) release
following exposure to 600 and 1,500 ppb NCh (Frampton et al.. 2002) as well as
increased expression of intercellular adhesion molecule 1 (ICAM-1), an extracellular
receptor for viruses following exposure to 2,000 ppb NCh (Pathmanathan et al.. 2003).
Immunoglobulin Antibody Response
Immunoglobulin M antibodies increase in response to infections, and a recent
epidemiologic study of adults infected with human immunodeficiency virus and
hospitalized for pneumocystis pneumonia found a 34% (95% CI: 6.5, -60) diminished
antibody response to pneumocystis proteins per 20-ppb increase in 24-h avg ambient NCh
concentrations (lag 0-2 day avg) (Blount et al.. 2013). Potential confounding by
5-167
-------
traffic-related copollutants or factors such as meteorology, sex, and SES was not
examined. Further, because subjects were distributed at varying distances of the single
central site monitor, which was located within 1 km of major roads, the impact of
exposure measurement error on the results is uncertain. Thus, the results do not strongly
inform the understanding of the effects of NCh on respiratory infections.
5.2.5.5 Summary of Respiratory Infection
The strongest evidence for an effect of short-term NO2 exposure on respiratory infections
is from toxicological studies, although results vary by exposure concentration. NC>2
exposures of 500 or 600 ppb, whether for a few hours, 1 week, or 1 month, did not affect
infection-induced mortality or bacterial clearance or killing in experimental animals or
affect virus titers or inactivation in humans (Table 5-22). Susceptibility to bacterial
infection did increase in experimental animals following NC>2 exposures of 2 to 48 hours
in the range of 1,000 to 5,000 ppb. Thus, there is some evidence from toxicological
studies to support the associations observed in some epidemiologic studies between
increases in ambient NC>2 concentrations (5- to 7-day avg) and increases in respiratory
infections. Although many epidemiologic studies observed null or imprecise associations
with wide 95% CIs for hospital admissions, ED visits, and parental reports of infection
(Table 5-23 and Figure 5-10). some observed associations with hospital admissions or
ED visits for bronchiolitis, ear infection, or any respiratory infection. Epidemiologic
associations were observed in study populations with respiratory disease (i.e., children
with asthma, adults with COPD). Most epidemiologic studies did not examine
copollutant confounding, and respiratory infections also were associated with PM2 5, the
traffic-related pollutants BS, OC, and CO, as well as other highly correlated pollutants
such as PMio and SCh (r = 0.74 to 0.92). The toxicological evidence is not conclusive,
but there is some biological plausibility for NCh-induced impaired host defense. Some
studies demonstrated effects on potential mechanistic events underlying susceptibility to
infection. Results vary across studies, with NC>2 exposures in the range of 1,000 to 5,000
ppb showing variable effects on pulmonary clearance and superoxide production by AMs
but decreasing phagocytic activity in AMs isolated from exposed experimental animals
and humans. In both experimental animals and humans, NC>2 exposures of 100 to 500 ppb
(Table 5-26) did not consistently affect pulmonary clearance or AM numbers or
superoxide production. There was heterogeneity across studies in animal species, strain,
and sex that may or may not have contributed to inconsistencies observed in response to
NO2. Although NO2 exposure shows inconsistent effects on various endpoints related to
respiratory infection, there is some supporting epidemiologic evidence for incidence of
respiratory infections, evidence in mice for susceptibility to infection, and evidence from
5-168
-------
experimental studies for decreased AM function that provides some indication that
short-term NC>2 exposure may increase risk of respiratory infection.
5.2.6 Aggregated Respiratory Conditions
In addition to individual respiratory conditions, epidemiologic studies examined
respiratory effects as an aggregate of multiple respiratory conditions (e.g., asthma,
COPD, respiratory infections). The studies from the 2008 ISA for Oxides of Nitrogen
(U.S. EPA. 2008c) and recent studies consistently show associations between short-term
increases in ambient NC>2 concentration and increases in aggregated respiratory
conditions. This evidence is based primarily on hospital admissions and ED visits for all
respiratory diseases, which are the focus of the discussion in this section. Other outcomes
include lung function in adults with asthma or COPD and sales of medication for asthma
and COPD combined or cough and mucus combined. As described in preceding sections,
evidence for the effects of short-term NO2 exposure varies among specific respiratory
outcome groups. Thus, it is not clear whether the evidence for aggregated respiratory
conditions reflects associations with each respiratory condition equally or a particular
condition(s).
5.2.6.1 Respiratory Symptoms, Lung Function, and Medication Use
Outcomes such as lung function decrements in adults with asthma and/or COPD (Rice et
al.. 2013; Higgins et al.. 2000) and increases in the sale of medication for asthma and
COPD combined or for cough and mucus combined (Pitard et al.. 2004; Zeghnoun et al..
1999) were associated with ambient NO2 (24-h avg, lagged 0 to 7 days or 0-1 day avg) in
a small group of epidemiologic studies, with exception of the Higgins et al. (1995) study.
However, uncertainties in these studies result in weak inference of the independent
effects of NO2. Associations with medication sales were modeled with GAM in S-plus
(Pitard et al.. 2004). which can produce biased results (U.S. EPA. 2006). In the
Framingham cohort study, lung function was associated with PM2 5 (r = 0.63) (Rice et al..
2013). and a copollutant model was not analyzed. The other lung function studies did not
report what potential confounding factors were examined (Higgins et al.. 2000; 1995).
Another uncertainty is potential exposure measurement error produced by the use of
central site ambient concentrations to represent ambient exposure. The Framingham
study averaged NO2 concentrations from sites in the Boston, MA area (Rice etal.. 2013).
With one or two observations per subject collected over 3-9 years, the analysis relied on
both temporal and spatial contrasts in exposure. With individuals distributed across a
40 km area and variability in ambient NO2 observed across a range of 3 to 10 km in
5-169
-------
Boston (Section 2.5.2). it is not clear how well the average area concentration represents
ambient exposure of study subjects.
5.2.6.2 Hospital Admissions and Emergency Department Visits for All Respiratory
Diseases
Epidemiologic studies examining the association between short-term NO2 exposures and
respiratory-related hospital admissions or ED visits were not available until after the
completion of the 1993 AQCD for Oxides of Nitrogen. As a result, the 2008 ISA for
Oxides of Nitrogen (U.S. EPA. 2008c) contained the first thorough evaluation of
respiratory morbidity in the form of respiratory-related hospital admissions and ED visits.
The majority of the studies evaluated consisted of single-city, time-series studies that
examined all respiratory hospital admissions or ED visits with additional cause-specific
studies, as discussed in previous sections. Studies of all respiratory hospital admissions
and ED visits consistently reported positive associations with short-term NO2 exposures
(Figure 5-13 and Table 5-28). In the few analyses of copollutant models with PM2 5 or a
traffic-related copollutant among CO, UFP, benzene, or BS, NO2 associations were
generally found to be robust (U.S. EPA. 2008c). The evidence supporting NO2-associated
increases in all respiratory disease hospital admission and ED visits contributed heavily
to the 2008 ISA for Oxides of Nitrogen conclusion that "there is a likely causal
relationship between short-term exposure to NO2 and effects on the respiratory system"
(U.S. EPA. 2008c). The air quality characteristics of the cities and the exposure
assignment approach used in each study evaluated in this section are presented in
Table 5-27. As detailed in Section 5.2.2.4. other recent studies of all respiratory disease
hospital admissions and ED visits are not the focus of this evaluation, and the full list of
these studies, as well as study details, can be found in Supplemental Table S5-3 (U.S.
EPA. 2015h).
5-170
-------
Table 5-27 Mean and upper percentile concentrations of nitrogen dioxide in
studies of hospital admissions and emergency department visits for
aggregated respiratory conditions.
Study
Yana et al.
(2003)
Funq et al.
(2006)
Burnett et
al. (2001)
fCakmak
etal.
(2006)
tWonq et
al. (2009)
fDales et
al. (2006)
Location
Years
Vancouver,
BC, Canada
(1986- 1998)
Vancouver,
BC, Canada
(1995-1999)
Toronto, ON,
Canada
(1980- 1994)
10 Canadian
cities
(1993-2000)
Hong Kong,
China
(1996-2002)
11 Canadian
cities
(1986-2000)
Exposure
Assignment
Average of NO2
concentrations from
30 monitors.
Average of NO2
concentrations over
all monitors.
Average of NO2
concentrations from
4 monitors.
Average of NO2
concentrations over
all monitors within
each city.
Average of NO2
concentrations
across 8 monitors.
Average of NO2
concentrations over
all monitors within
each city.
Upper
Mean Percentile
Concentration Concentrations Copollutant
Metric ppb ppb Examination
24-h avg 18.7 NR Correlations (r):
O3: -0.32
Copollutant
models: Os
24-h avg 16.8 Max: 33.9 Correlations (r):
CO: 0.74
CoH; 0.72
O3: -0.32
SO2: 0.57
PMio: 0.54
PlVh.s: 0.36
PMio-2.s: 0.52
Copollutant
models: none
1-h max 44.1 146 Correlations (r):
O3: 0.52
Copollutant
models: Os
24-h avg 21.4 Max: 44-134 Correlations (r):
NR
Copollutant
models: none
24-h avg 31.2 75th: 37.0 Correlations (r):
Max: 89.4 NR
Copollutant
models: none
24-h avg 21.8 95th: 21-43 Correlations (r),
across cities:
PMio: -0.26 to
0.69
O3: -0.55 to
0.05
SO2: 0.20-0.67
CO: 0.13-0.76
Copollutant
models: none
5-171
-------
Table 5-27 (Continued): Mean and upper percentile concentrations of nitrogen
dioxide in studies of hospital admissions and emergency
department visits for aggregated respiratory conditions.
Study
fSon etal.
(2013)
tAtkinson
etal.
(2012)
tFaustini
etal.
(2013)
Location
Years
8 South
Korean cities
(2003- 2008)
Meta-analysis
(Asia)
(years NR)
6 Italian cities
(9001 90051
Exposure
Assignment
Hourly ambient NO2
concentrations from
monitors in each city.
NR
Average of NO2
concentrations over
all monitors within
each city. Number of
NO2 monitors in each
city ranged from
1-5.a
Upper
Mean Percentile
Concentration Concentrations Copollutant
Metric ppb ppb Examination
24-h avg 11.5-36.9 NR Correlations (r):
PMio: 0.5
O3: -0.1 SO2:
0.6
CO: 0.7
Copollutant
models: none
24-h avg NR NR Correlations (r):
NR
Copollutant
models: none
24-h avg 24.1-34.6 NR Correlations (r),
across cities:
PMio:
0.22-0.79
Copollutant
models: PMio
5-172
-------
Table 5-27 (Continued): Mean and upper percentile concentrations of nitrogen
dioxide in studies of hospital admissions and emergency
department visits for aggregated respiratory conditions.
Study
Location
Years
Exposure
Assignment
Metric
Upper
Mean Percentile
Concentration Concentrations Copollutant
ppb ppb Examination
Peel et al.
(2005)
Atlanta, GA
(1993-2000)
Average of NO2
concentrations from
monitors for several
monitoring networks.
1-h max
45.9
NR
Correlations (r):
PlVh.s: 0.46
PMio: 0.49
PMio-2.s: 0.46
UFP: 0.26
PIvh.s Water
Soluble Metals:
0.32
PIvh.s Sulfate:
0.17
PIvh.s Acidity:
0.10
PIvh.s OC: 0.63
PIvh.s EC: 0.61
Oxygenated
HCs: 0.30
O3: 0.42
CO: 0.68
SO2: 0.34
Copollutant
models: none
Tolbert et
al. (2007)
Atlanta, GA
(1993-2004)
Average of NO2
concentrations from
monitors for several
monitoring networks.
1-h max
81.7
306
Correlations (r):
PIvh.s: 0.47
PMio: 0.53
PMio-2.s: 0.4
PIvh.s Sulfate:
0.14
PIvh.s OC: 0.62
PIvh.s EC: 0.64
PIvh.s TC: 0.65
PIvh.s Water
Soluble Metals:
0.32
Oxygenated
HCs: 0.24
Os: 0.44
CO: 0.70
SO2: 0.36
Copollutant
models: CO,
PMio, Os
5-173
-------
Table 5-27 (Continued): Mean and upper percentile concentrations of nitrogen
dioxide in studies of hospital admissions and emergency
department visits for aggregated respiratory conditions.
Study
fDarrow
etal.
(2011 a)
Location Exposure
Years Assignment
Atlanta, GA Epidemiologic
(1993 2004) analysis used N02
concentrations from
1 centrally located
monitor. Assessment
of spatial
heterogeneity relied
upon all monitors
from ARIES and the
U.S. EPA.
Mean
Concentration
Metric ppb
1-h max 1-h max: 43
24-h avg 24-h avg: 22
Commute13 Commute: 21
Daytimeb Daytime: 17
Nighttime13 Nighttime: 25
Upper
Percentile
Concentrations
ppb
75th:
1-h max: 53
24-h avg: 28
Commute: 27
Daytime: 22
Nighttime: 35
Max:
1-h max: 181
24-h avg: 74
Commute: 97
Daytime: 82
Nighttime: 97
Copollutant
Examination
Correlations (r),
for averaging
times specified
in current
NAAQS:
CO, 1-h: 0.61
O3, 8-h: 0.34
PM2.5, 24-h:
0.42
Copollutant
models: none
Avg = average; BC = Brittish Columbia; CO = carbon monoxide; CoH = coefficient of haze; EC = elemental carbon;
ED = emergency department; GA = Georgia; HC = hydrocarbon; NAAQS = National Ambient Air Quality Standards;
NO2 = nitrogen dioxide; NR = not reported; O3 = ozone; OC = organic carbon; ON = Ontario; PM2.s = particulate matter with a
nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PMio = particulate matter with a nominal mean aerodynamic
diameter less than or equal to 10 |jm; SO2 = sulfur dioxide; TC = total carbon; UFP = ultrafine particles; U.S. EPA = United States
Environmental Protection Agency.
aMonitoring information obtained from Colais et al. (2012).
bCommute = 7 ante meridiem (a.m.)- 10a.m., 4 p.m.-7 p.m., Daytime = 8 a.m.-7 p.m., Nighttime = 12 a.m.-6 a.m.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
Hospital Admissions
Multicity studies conducted in Canada (Cakmak et al.. 2006; Dales et al.. 2006). Italy
(Faustini et al.. 2013) and Korea (Sonet al.. 2013). as well as a single-city study
conducted in Hong Kong, China (Wong et al.. 2009) observed associations between
short-term NC>2 concentrations and hospital admissions for all respiratory diseases, each
focusing on a different age range (Figure 5-13 and Table 5-28). Additional support for an
association between short-term NCh exposures and respiratory hospital admissions comes
from a meta-analysis of studies conducted in Asian cities (Atkinson et al.. 2012).
Cakmak et al. (2006) focused on all ages in 10 Canadian cities with the primary objective
of the study being to examine the potential modification of the effect of ambient air
pollution on daily respiratory hospital admissions by education and income using a
time-series analysis conducted at the city level (the effect measure modification analysis
is discussed in Chapter 7). The authors calculated a pooled estimate across cities for each
pollutant using a random effects model by first selecting the lag day with the strongest
5-174
-------
association from the city-specific models. For NC>2, the mean lag day across cities that
provided the strongest association and for which the pooled effect estimate was
calculated was 1.4 days. At this lag, Cakmak et al. (2006) reported a 2.3% increase
(95% CI: 0.2, 4.5) in respiratory hospital admissions for a 20-ppb increase in 24-h avg
NC>2 concentrations. This result is consistent with a study conducted in Hong Kong,
China that examined whether influenza modifies the relationship between air pollution
exposure and hospital admissions (Wong etal.. 2009). Wong et al. (2009) observed a
3.2% (95% CI: 1.9, 4.5) increase in all respiratory disease hospital admissions for all ages
at lag 0-1 days for a 20-ppb increase in 24-h avg NO2 concentrations, with an association
slightly smaller in magnitude for acute respiratory disease (2.1% [95% CI: -0.1, 4.3]),
which comprises approximately 39% of all respiratory disease hospital admissions in
Hong Kong, China. Cakmak et al. (2006) also examined the potential confounding by
other pollutants but only through the use of a multipollutant model (i.e., two or more
additional pollutants included in the model). These models are difficult to interpret due to
the multicollinearity between pollutants and are not evaluated in this ISA.
In an additional multicity study conducted in 11 Canadian cities, Dales et al. (2006)
focused on NCh-associated respiratory hospital admissions in neonatal infants (ages
0-27 days). The investigators used a statistical analysis approach similar to Cakmak et al.
(2006) (i.e., time-series analysis to examine city-specific associations, and then a random
effects model to pool estimates across cities). Dales et al. (2006) observed that the mean
lag day across cities that provided the strongest association for NO2 was 1 day, which
corresponded to a 6.5% (95% CI: 3.5, 9.6) increase in neonatal respiratory hospital
admissions for a 20-ppb increase in 24-h avg NC>2 concentrations. Similar to Cakmak et
al. (2006). Dales et al. (2006) only examined the potential confounding effects of other
pollutants on the NCh-respiratory hospital admission association through the use of
multipollutant models, which are not informative due to multicollinearity between
pollutants.
The results of Cakmak et al. (2006) and Wong et al. (2009). which focus on all ages, are
further supported by Son et al. (2013). a study that examined the association between
short-term exposures to air pollution and respiratory-related hospital admissions in eight
South Korean cities. South Korea has unique demographic characteristics with some
indicators more in line with other more developed countries (e.g., life expectancy,
percentage of population living in urban areas), but because it represents a rapidly
developing Asian country, it is likely to have different air pollution, social, and health
patterns than less industrialized Asian nations or Western nations that developed earlier
(Sonet al.. 2013). In a time-series analysis using a two-stage Bayesian hierarchical
model, Son et al. (2013) examined both single-day lags and cumulative lags up to 3 days
(i.e., lag 0-3). The authors only presented NC>2 results for the strongest lag and observed
5-175
-------
a 3.6% increase (95% CI: 1.0, 6.1) in respiratory disease hospital admissions at lag 0 for a
20-ppb increase in 24-h avg NC>2 concentrations. These results are consistent with those
of a meta-analysis of studies conducted in Asian cities by Atkinson et al. (2012). which in
a random effects model based on five estimates, reported a 3.5% increase (95% CI: 0.6,
6.5) in respiratory hospital admissions for a 20-ppb increase in 24-h avg NC>2
concentrations.
Son etal. (2013) did not conduct copollutant analyses; however, similar patterns of
associations were observed across pollutants that were moderately [PMio (r = 0.5); SCh
(r = 0.6)] to highly correlated [CO (r = 0.7)] with NO2. Son etal. (2013) also examined
potential seasonal differences in all respiratory disease hospital-admission associations.
The authors reported that the association with NO2 was largest in magnitude during the
summer (8.3% [95% CI: 2.8, 14.3], lag 0). However, across the eight cities, NO2
concentrations were lowest during the summer season (<20 ppb compared to >24 ppb in
the other seasons), which complicates the interpretation of these results.
Faustini et al. (2013) focused on examining the relationship between short-term air
pollution exposures and respiratory hospital admissions, specifically on the adult
population (i.e., individuals 35 years of age and older) in six Italian cities. In a time-series
analysis the authors examined the lag structure of associations through single-day lags as
well as cumulative lags, using cubic polynomial distributed lags, in an attempt to identify
whether the NO2 effect on respiratory-related hospital admissions was immediate (lag 0,
lag 0-1 days), delayed (lag 2-5 days), or prolonged (lag 0-3, 0-5 days). The authors
reported that NO2 was most strongly associated with all respiratory hospital admissions at
lag 0-5 days (4.6% [95% CI: 0.87, 8.3] for a 20-ppb increase in 24-h avg NO2
concentrations), which differs from Cakmak et al. (2006) and Dales et al. (2006) where
the strongest effects were observed at lags less than 2 days. However, Faustini et al.
(2013) did observe positive associations, although smaller in magnitude (ranging from
2.5-2.9%) at the shorter lags (i.e., lag 0 and 0-1 days). Faustini etal. (2013) only
examined potential copollutant confounding of NC>2 associations in models with PMio,
and reported that the NO2 association with respiratory hospital admissions at lag
0-5 days was attenuated slightly, but remained positive (3.3% [95% CI: -1.1, 7.8]).
Emergency Department Visits
Studies of ED visits for aggregated respiratory conditions that were evaluated in the 2008
ISA for Oxides of Nitrogen were few in number and focused almost exclusively on study
populations consisting of all ages, and U.S. studies were limited to Atlanta, GA. Building
on the previous studies conducted in Atlanta, GA (Tolbert et al.. 2007; Peel et al.. 2005).
Darrow et al. (201 la) examined associations between short-term air pollution exposures
5-176
-------
and all respiratory ED visits. To examine the association between the various NO2
exposure metrics and respiratory ED visits, the authors conceptually used a time-stratified
case-crossover framework in which control days were selected as those days within the
same calendar month and maximum temperature as the case day. However, instead of
conducting a traditional case-crossover analysis, the authors used a Poisson model with
indicator variables for each of the strata (i.e., parameters of the control days). Darrow et
al. (201 la) only reported results for a 1 day lag in NC>2 concentrations. For a 30-ppb
increase in 1-h maxNCh concentrations, the authors reported a 1.4% increase (95% CI:
0.8, 2.1) in all respiratory ED visits. These results are slightly smaller than those reported
by Peel et al. (2005) and Tolbert et al. (2007). but this could be attributed to the fact that
the latter two studies used a multiday average of NO2 concentrations (i.e., lag 0-2 days)
instead of the single-day lag used in Darrow et al. (201 la).
Model Specification—Sensitivity Analyses
A question that often arises in the examination of associations between air pollution and a
health effect is whether the statistical model employed adequately controls for the
potential confounding effects of temporal trends and meteorological conditions. Sonetal.
(2013). in the study of eight South Korean cities, conducted a sensitivity analysis to
identify whether risk estimates changed depending on the abused to control for temporal
trends and meteorology covariates (i.e., temperature, humidity, and barometric pressure).
Similar to the other respiratory-related hospital admission outcomes examined, the
authors reported that the association between short-term NO2 exposures and all
respiratory disease hospital admissions was sensitive to using less than 6 dfper year to
control for temporal trends, but was stable when using 6-10 dfper year. Additionally,
when varying the number of df used for the meteorology covariates from 3 to 6 dfas well
as the lag structure (i.e., lag 0 and lag 0-3 days), the NO2 association remained robust
(i.e., relatively unchanged).
Exposure Assignment
In addition to model specification, the method used to assign exposure in epidemiologic
studies has been suggested to influence the magnitude and direction of air
pollution-health effects associations. As discussed in Section 5.2.2.4. Strickland et al.
(2011) examined exposure assignment in the case of asthma ED visits in Atlanta, GA and
found that different exposure assignment approaches could influence the magnitude, but
not direction of associations. Darrow et al. (201 la) also used data from Atlanta, GA to
examine the influence of alternative exposure metrics on the association between
short-term NC>2 concentrations and all respiratory ED visits along with the spatial
variability of each exposure metric.
5-177
-------
To examine whether all respiratory ED visits associations differed depending on the
exposure metric used, Darrow et al. (201 la) used five different exposure metrics:
(1) 1-h max; (2) 24-h avg; (3) commuting period (7:00 a.m. to 10:00 a.m. and 4:00 p.m.
to 7:00 p.m.); (4) daytime avg (8:00 a.m. to 7:00 p.m.); and (5) nighttime avg (12:00 a.m.
to 6:00 a.m.). The authors reported relatively consistent results (using an a priori lag of
1 day) across exposure metrics with the largest estimate found for the nighttime avg and
the smallest for the daytime metrics (Figure 5-11). The larger risk estimate for the
nighttime metric could be a reflection of NC>2 during this exposure duration being a better
surrogate for NCh concentrations on the previous day (Darrow et al., 201 la). The
correlations of 1-h max NCh with most of the other NC>2 metrics were lower than those
for other pollutants examined in the study (i.e., r < 0.60). However, 1-h max NC>2 was
highly correlated with 24-h avg NC>2 (r = 0.79), which is the other NC>2 metric often
examined in epidemiologic studies.
5-178
-------
fg
o
1.03
o 1 02 -
Q? | 1-01 -
J£ *
f^/f 4^
K 2 1 00
•
— o.99 -
Partial
Spearman r
J
i
1 0.79 0.59 0.55 0.44
3 $
C Cw
I I I
CM
NO.
Note: ave = average; hr = hour; IQR = interquartile range; max = maximum; NO2 = nitrogen dioxide. Partial Spearman correlation
coefficient between a priori metrics (shaded in gray) and other pollutant metrics shown above the x-axis.
Source: Reprinted with permission of Nature Publishing Group (Darrow et al.. 2011 a).
Figure 5-11 Risk ratio and 95% confidence intervals for associations between
various lag 1 day nitrogen dioxide metrics and respiratory
emergency department visits.
In the analysis of the spatial correlation of exposure metrics for NO2, Darrow et al.
(2011 a) found that unlike Os and PM2 5, which were spatially homogenous, there was
evidence that correlations for NO2 metrics decreased dramatically as distance from the
central site monitor increased (Figure 5-12). This was especially true for the 1-h max and
nighttime metrics (r < 0.20) at 60 km. The 24-h avg metric was also reduced (r = -0.40),
but not as dramatically as the 1-h max. Although reduced at greater distances, moderate
correlations (r = -0.50) were reported with the central site monitor for the daytime and
5-179
-------
commute time metrics. Overall, these results suggest evidence of potential exposure
misclassification for NC>2 with increasing distance from the central site monitor across
exposure metrics.
1.0
c
g 0.8
o
8 0.6
8 0.4
c
ra
0.2
00
N02
o
--V--
_ _A _ .
— -• —
24-hr ave
commute
day
night
10 20 30 40 50 60
Distance from Central Monitor (km)
70
Note: ave = average; hr = hour; km = kilometer; max = maximum; NO2 = nitrogen dioxide.
Source: Reprinted with permission of Nature Publishing Group, Darrow et al. (2011 a).
Figure 5-12 Spatial correlations for nitrogen dioxide metrics in the Atlanta, GA
area.
As detailed within this section, hospital admission and ED visit studies of all respiratory
diseases consistently report positive associations with short-term increases in ambient
NC>2 concentrations. As presented in Figure 5-13 and Table 5-28. associations are
consistently observed in studies evaluated in the 2008 ISA for Oxides of Nitrogen as well
as recent studies.
5-180
-------
Study
Cakmaketal. (2006)
Atkinson et al. (2012)
Son etal. (2013)
Dales etal. (2006)
Ri irnatt at a I fOC\C\^ ^
Durnen ei ai. ^zuu ij
Vann ot al fOrin^
T any siai. ^zuuo^
Faustini et al. (2013)
Pi inn at al fOfinK\
rung ei ai. ^zuuoj
Wong etal. (2009)
Vann ot a I fl)r\r\'V\
T any siai. ^zuuo^
Wong etal. (2009)a
Peel et al. (2005)
Tolbertetal. (2007)
Darrow etal. (2011)
Location
10 Canadian cities
Meta-analysis (Asia)
8 South Korean cities
1 1 Canadian cities
I oronto, CAN
Vancouver, CAN
6 Italian cities
Vancouver, CAN
Hong Kong
Vancouvsr, CAN
Hong Kong
Atlanta, GA
Atlanta, GA
Atlanta, GA
Age
All
All
All
0-27 days
e- 9
*> Z
-------
Table 5-28 Corresponding risk estimate for studies presented in
Study
Location
Age
Avg Time
Lag
Figure 5-13.
% Increase
(95% Cl)
Hospital Admissions
tCakmak et al. (2006)
t Atkinson et al. (2012)
tSon etal. (2013)
tDalesetal. (2006)
Burnett etal. (2001)
Yang et al. (2003)
tFaustini et al. (2013)
Funq et al. (2006)
tWonq et al. (2009)a
Yanq et al. (2003)
tWonq et al. (2009)a
Emergency Department
Peel et al. (2005)
Tolbert et al. (2007)
tDarrowet al. (2011 a)
10 Canadian cities
Meta-analysis (Asia)
8 South Korean cities
11 Canadian cities
Toronto, ON, Canada
Vancouver, BC, Canada
6 Italian cities
Vancouver, BC, Canada
Hong Kong, China
Vancouver, BC, Canada
Hong Kong, China
Visits
Atlanta, GA
Atlanta, GA
Atlanta, GA
All
All
All
0-27 days
<2
<3
35+
65+
All
65+
65+
All
65+
All
All
All
24-h avg
24-h avg
24-h avg
24-h avg
1-h max
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
1-h max
1-h max
1-h max
1.4
0
1
0-1
1
0-5 DL
0-2
0-1
0-1
1
0-1
0-1
0-2
0-2
1
2.3(0.2,4.5)
3.5(0.6,6.5)
3.6(1.0, 6.1)
6.5(3.5, 9.6)
13.3(5.3,22.0)
19.1 (7.4, 36.3)
4.6(0.9, 8.3)
9.1 (1.5, 17.2)
3.2(1.9,4.5)
4.0 (2.4, 5.7)
19.1 (11.2,27.5)
2.1 (-0.1,4.3)
1.7 (-0.6, 4.0)
2.4(0.9,4.1)
2.0(0.5, 3.3)
1.4(0.8,2.1)
avg = average; BC = British Columbia; Cl = confidence interval; DL = distributed lag; GA = Georgia; max = maximum;
ON = Ontario.
aThis estimate is for acute respiratory diseases, which comprise approximately 39% of all respiratory disease hospital admissions
in Hong Kong, China.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-182
-------
5.2.6.3 Summary of Aggregated Respiratory Conditions
Previous and recent epidemiologic studies consistently indicate associations between
short-term increases in ambient NC>2 concentrations and increases in respiratory effects
aggregated across specific conditions such as asthma, COPD, and respiratory infections.
A majority of the available evidence is for hospital admissions for all respiratory diseases
combined (Figure 5-13 and Table 5-29). with a few additional studies of ED visits for all
respiratory diseases, lung function in adults with asthma or COPD, or medication sales
for unspecified respiratory effects. Associations of NC>2 with respiratory disease hospital
admissions and ED visits are observed to be larger for children and older adults; limited
evidence points to differences in risk by sex and SES.
With respect to the lag structure of associations, evidence indicates that the largest
increase in all respiratory hospital admissions and ED visits occurs within the first few
days after NO2 exposure, specifically lags of 0 to 2 days. An examination of model
specification indicated the NCh-respiratory hospital admission relationship was robust to
alternative lags and df for weather covariates (Son et al.. 2013). Thus, varying approaches
to modeling weather did not appear to be a source of confounding. NO2 effect estimates
were sensitive to using less than 6 dfper year to account for temporal trends, but most
studies did not model temporal trends with fewer df. The limited analysis of potential
seasonal differences suggests that NO2 associations with all respiratory disease hospital
admissions are stronger during the summer (Son et al.. 2013). In a comparison of
averaging times, similar associations with all respiratory disease ED visits were observed
for 1-h max and 24-h avg NCh (Darrow et al.. 201 la).
The epidemiologic evidence for associations of NO2 with aggregated respiratory effects is
based on exposure assessment from central site monitors. In two study locations, Boston,
MA (Section 2.5.2) and Atlanta, GA (Darrow et al.. 201 la), between-monitor correlation
in ambient NO2 concentration decreased with increasing distance. Thus, it is unclear the
extent to which temporal variation in central site NC>2 concentrations represent variation
in exposure among subjects. Also, studies of aggregated respiratory effects did not
thoroughly examine potential confounding by traffic-related copollutants, which in many
studies, showed moderate to high (r = 0.61-0.76, Table 5-28) correlations with NC>2. In
limited analysis of copollutant models, NO2 associations with all respiratory hospital
admissions and ED visits persisted with adjustment for CO or PIVb 5 (Tolbert et al.. 2007)
(Figures 5-16 and 5-17). However, potential differential measurement error resulting
from central site exposure assessment limits inference from the copollutant model results.
Further, given the variable nature of evidence for the effects of short-term NO2 exposure
among specific respiratory conditions (Sections 5.2.2. 5.2.4. and 5.2.5). it is not clear
5-183
-------
whether the evidence for aggregated respiratory conditions reflects associations with each
respiratory condition equally or a particular condition(s).
5.2.7 Respiratory Effects in Healthy Populations
Similar to populations with asthma and COPD, an array of respiratory outcomes has been
examined in relation to short-term exposure to NO2 in healthy populations. The 2008 ISA
for Oxides of Nitrogen did not draw inferences specifically about respiratory effects of
NO2 exposure in healthy populations (U.S. EPA. 2008c) but described epidemiologic
associations of short-term increases in ambient NO2 concentration with increases in
respiratory symptoms and decreases in lung function in children. Evidence from
experimental studies varied across outcomes, indicating no effects on respiratory
symptoms or lung function in healthy adults. However, NO2 exposure did affect
underlying key events, inducing increases in airway responsiveness and PMNs in healthy
adults generally at 1,000-ppb NO2 exposure or higher. Recent evidence, which is from
epidemiologic studies, continues to indicate NO2-related respiratory effects in healthy
populations, most consistently seen as increases in pulmonary inflammation.
5.2.7.1 Airway Responsiveness in Healthy Individuals
Recent studies are not available, but the 2008 ISA for Oxides of Nitrogen reported that
increases in nonspecific airway responsiveness were observed in healthy adults following
1- to 3-hour NO2 exposures in the range of 1,500 to 3,000 ppb (U.S. EPA. 2008c).
Studies of airway responsiveness in healthy individuals were generally conducted using
volunteers ages 18 to 35+ years. Mohsenin (1988) found that a 1-hour resting exposure to
2,000 ppb NO2 increased responsiveness to methacholine. A mild increase in
responsiveness to carbachol was observed following a 3-hour exposure to 1,500 ppb NO2
with moderate intermittent exercise (VE = 40 L/min; 10 of 30 minutes) (Frampton et al..
1991). Kulle and Clements (1988) also showed a tendency for greater FEVi decrements
from methacholine challenge following 2-hour resting exposures to 2,000 and 3,000 ppb
NO2. Resting exposures to 100 ppb NO2 for 1 hour did not affect carbachol or
methacholine responsiveness in healthy subjects (Ahmed etal.. 1983a; Hazucha et al..
1983). Two meta-analyses of the available literature confirm statistically significant
effects of NO2 exposures above 1,000 ppb, but not below, on airway responsiveness in
healthy individuals (Kjaergaard and Rasmussen. 1996; Folinsbee. 1992).
5-184
-------
5.2.7.2 Lung Function Changes in Healthy Populations
Compared with evidence for airway responsiveness to an inhaled bronchoconstrictor, the
2008 ISA for Oxides of Nitrogen reported weak evidence for the effects of NCh exposure
on changes in lung function in the absence of a bronchoconstrictor in controlled human
exposure and epidemiologic studies of healthy adults (U.S. EPA. 2008c). A small body of
epidemiologic studies of children in the general population indicated associations
between increases in ambient NO2 concentration and decrements in lung function
measured by supervised spirometry. Several recent studies, which are epidemiologic,
contribute inconsistent evidence for ambient NCh-associated lung function decrements in
children in the general population.
Epidemiologic Studies of Children in the General Population
As in other populations, ambient NC>2 concentrations are more consistently associated
with lung function decrements in children in the general population as measured by
supervised spirometry than by home PEF. However, many studies of supervised
spirometry did not find associations with ambient NC>2 concentrations. Locations, time
periods, and ambient concentrations of oxides of nitrogen for these studies are presented
in Table 5-29. The studies recruited children from schools, supporting the likelihood that
study populations were representative of the population of children in the study areas.
The most informative studies of lung function in children are those examining NC>2
concentrations outdoor schools or at a central site adjacent to schools, which may
represent a component of the subjects' ambient exposures. These metrics were
inconsistently associated with lung function in children (Altug et al.. 2014; Castro et al.,
2009; Moshammer et al., 2006; Scarlett et al., 1996). The inconsistent evidence does not
appear to be related to the health status of the study population. NO2 was not associated
with lung function in children without respiratory symptoms (Altug etal. 2014). but
results were inconsistent in groups of children with prevalence of asthma or wheeze of 5
or 9% (Castro et al., 2009; Scarlett et al., 1996). Associations were found with same-day
NO2 and NO2 averaged over 3 to 8 days but were inconsistent for Lag Day 1. Linn et al.
(1996) found that a 20-ppb increase in lag 0 of central site NO2 was associated with a
-5.2 mL (95% CI: -13, 2.3) change in evening FEVi among children in three southern
California communities. These results have relatively strong inference about an
association with ambient NC>2 exposure because daily average personal and ambient NO2
were reported to be well correlated (r = 0.63).
5-185
-------
Table 5-29
Study3
tAltuq et al.
(2014)
fCastro et al.
(2009)
Moshammer et
al. (2006)
Scarlett et al.
(1996)
Linn et al.
(1996)
Oftedal et al.
(2008)
tPadhi and
Padhy (2008)
fEenhuizen et
al. (2013)
tChanq et al.
(2012)
tBaqheri
Lankarani et
al. (2010)
Steerenberq et
al. (2001)
Mean and upper percentile oxides of nitrogen concentrations in
epidemiologic studies of lung function in the general population.
Location
Eskisehir, Turkey
Rio de Janeiro, Brazil
Linz, Austria
Surrey, U.K.
Upland, Rubidoux,
Torrance, CA
Oslo, Norway
West Bengal, India
3 study areas in the
Netherlands
Taipei, Taiwan
Tehran, Iran
Utrecht, the
Netherlands
Bilthoven, the
Netherlands
Study Period
Feb-Mar 2007
May, Jun, Sep,
Oct 2004
School yr
2000-2001
Jun-Jul 1994
School yr
1992-1994
Nov 2001 -Dec
2002
Jun 2006-Jul
2007
Oct 2000-Nov
2001
Dec 1996-May
1997
NR
Feb-Mar 1998
Exposure
Metric
Analyzed
24-h avg NO2
24-h avg NO2
8-h avg NO2
(12-8 a.m.)
24-h avg NO2
1-h max NO2
24-h avg NO2
24-h avg NO2
24-h avg NO2
indoor
24-h avg NO
indoor
24-h avg NO2
6-day avg NO2
24-h avg NO2
24-h NO
24-avg NOx
24-h avg NO2
24-h avg NO
24-h avg NO2
24-h avg NO
Mean/Median
Concentration
ppb
Suburban: 9.4,
Urban: 13.0
Traffic: 21. 2
49.2b
NR
9.3b
34.9
33
14.4b
LPG: 37.7,
Biomass fuel:
71.7
LPG: 27.5,
Biomass fuel:
46.7
16.0b
31.8
75.5, 17.6b
51.6, 40.4b
72.9, 38.8b
28.2b
30.2b
25.5b
7.4b
Upper Percentile
Concentrations
ppb
Max: 13.1
Max: 17.7
Max: 28.2
Max: 115b
NR
75th: 11. 4b
Max: 82
Max: 96
Max: 59.2b
75th: LPG: 44,
Biomass fuel: 90
75th: LPG: 30,
Biomass fuel: 55
75th: 23.2b
Max: 47.9b
75th: 41. 7
Max: 119, 25.5b
Max: 85.1, 11 Ob
Max: 122, 94.7b
Max: 44.7b
Max: 168b
Max: 49. 5b
Max: 85.6b
5-186
-------
Table 5-29 (Continued): Mean and upper percentile oxides of nitrogen
concentrations in epidemiologic studies of lung function
in the general population.
Study3
Location
Exposure
Metric
Study Period Analyzed
Mean/Median Upper Percentile
Concentration Concentrations
ppb ppb
Peacock et al. Rochester-upon-
(2003) Medway, U.K.
Nov 1996-Feb 24-h avg NO2
1997 1-hmaxNO2
17.4,17.1,19.2 Max: 39, 39, 43
28.5,28.1,31.8 Max: 67, 71, 98
tCorreia-Deur
etal. (2012)
van der Zee et
al. (2000)
van der Zee et
al. (1999)
Ranzi et al.
(2004)
Ward et al.
(2000)
Roemeret al.
(1998)
Sao Paolo, Brazil
Rotterdam,
Nunspeet,
Bodegraven/Reeuwij,
Amsterdam, Meppel,
the Netherlands
Emiglia-Romagna,
Italy
West Midlands, U.K.
Sweden, Germany,
Finland, Hungary,
Apr-Jul 2004 24-h avg
Three winters 24-h avg NO2
1992-1993
1993-1994
1994-1995
Feb-May1999 24-h avg NO2
Jan-Mar 1997 24-h avg NO2
May-Jul 1997
Winter 24-h avg NO2
1993-1994
Mean: 69. 9b
27.1, 17.6b
25.5, 13.3b
25.0, 11. 7b
Urban: 37.0b
Rural: 18.51b
NR
Across locations:
6.7-39.8b
75th:
90th:
Max:
Max:
Max:
NR
NR
NR
NR
84.5b
102b
50, 44.2b
40.4, 28.7b
43.6, 30.3b
Norway, Italy,
Greece,
Czech Republic, the
Netherlands
Timonen and Kuopio, Finland
Pekkanen
(1997)
Feb-Apr 1994 24-h avg NO2
Urban: 14.9b Max:41.5b
Suburban: 7.4b Max: 27.1b
Schindleret al. Aarau, Basel, Davos, NR
(2001) Geneva, Lugano,
Montana, Payerne,
Wald, Switzerland
24-h avg NO2 19.5b
tCakmak et al. 14 Canadian cities Mar2006-Mar 24-h avg N02
(2011 a) 2007
12.6
Max: 69.3b
fSteinvil et al. Tel Aviv, Israel
(2009)
Sep2002-Nov 24-h avg NO2 19.3
2007
75th: 25.3
Max: 59.9
95th: 29.4
tLepeule etal. Boston, MA area 1999-2009 24-h avg N 02
(2014)
20.2b
95th: 23.9b
tSon etal. Ulsan, South Korea 2003-2007 24-h avg NO2
(2010)
21.4
75th: 26.1
Max: 44.8
5-187
-------
Table 5-29 (Continued): Mean and upper percentile oxides of nitrogen
concentrations in epidemiologic studies of lung function
in the general population.
Study3
tAqarwal et al.
(2012)
fWeichenthal
etal. (2011)
fThaller et al.
(2008)
fStrak et al.
(2012)
t Dales et al.
(2013)
Location
Patiala, Punjab area,
India
Ottawa, ON, Canada
Galveston, TX
Bilthoven, the
Netherlands
Sault Ste. Marie, ON,
Canada
Study Period
Aug-Jan
2007-2009
NR
Summers 2002,
2003, 2004
Mar-Oct 2009
May-Aug 2010
Exposure
Metric
Analyzed
1-mo avg NO2
1-h avg NO2
24-h avg NO2
1-h max NO2
5-h avg NOx
5-h avg NO2
10-h avg NO2
(8 a.m.-6 p.m.)
Mean/Median
Concentration
ppb
For 2008
Aug-Sep: 8.4b
Oct-Nov:21.9b
Dec-Jan: 17. 4b
High traffic: 4.8
Low traffic: 4.6
1.2
3.2
36
20
Near steel plant:
7.1
Distant site: 4.5
Upper Percentile
Concentrations
ppb
NR
Max: 1 1
Max: 10
Max: 7.1
Max: 27.7
Max: 96
Max: 34
NR
a.m. = ante meridiem; Aug = August; avg = average; CA = California; Dec = December; Feb = February; LPG = liquefied
petroleum gas; MA = Massachusetts; max = maximum; NO = nitric oxide; NO2 = nitrogen dioxide; NOx = sum of NO and NO2;
NR = not reported; ON = Ontario; ppb = parts per billion; TX = Texas; U.K. = United Kingdom.
aStudies presented in order of first appearance in the text of this section.
bNO2 concentrations converted from |jg/m3 to ppb by multiplying by 0.532, NO concentrations converted by multiplying by 0.815.
Both conversions assume standard temperature (25°C) and pressure (1 atm).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
In the studies examining ambient NO2 metrics representing subjects' school or personal
(total or indoor) exposure, there is uncertainty regarding confounding by PM2 5 or
traffic-related copollutants. Associations were found with CO, PMi, and PM2 5 (Castro et
al.. 2009; Padhi and Padhy. 2008; Moshammer et al.. 2006; Linn etal.. 1996); other
traffic-related pollutants were not examined. There is some information on potential
confounding of NO2-related lung function decrements in children by PM2 5, and results
are inconclusive. Oftedal et al. (2008) observed high correlations (r = 0.83-0.95) among
NO2, PM2 5, and PMio estimated by a dispersion model. Linnet al. (1996) did not provide
quantitative results and indicated only that NO2 effect estimates lost statistical
significance with adjustment for PM2s measured at schools, which could have differential
exposure error than NO2 measured at central sites. Among children in Austria, with
pollutants measured at a site adjacent to the school, NO2 effect estimates were unchanged
with adjustment for moderately correlated PM2 5 (r = 0.54) (Moshammer et al.. 2006). A
25-ppb increase in lag 1 of 8-h avg NO2 (12-8 a.m.) was associated with a -4.1% change
5-188
-------
(95% CI: -6.4, -1.7) in FEVi in the single-pollutant model and a -4.7% change (95% CI:
-7.3, -2.0) with adjustment for PM2s. PM2s effect estimates were attenuated or became
positive with adjustment for NC>2. While these results indicate an independent association
with NCh, other model covariates were not specified, and potential confounding by other
factors such as weather cannot be assessed.
Among studies of supervised spirometry, evidence was inconsistent for associations with
NO2 and NO ascertained from central sites (Eenhuizen et al.. 2013; Chang etal.. 2012;
Bagheri Lankarani et al.. 2010; Oftedal et al.. 2008; Steerenberg etal.. 2001). Results
were inconsistent for PEF as well as FEVi, and no association was found with a measure
of airway resistance. Controlled human exposure studies, conducted in healthy adults, do
not consistently indicate effects on ambient-relevant NO2 exposures on FEVi (see below)
or airway resistance (Section 4.3.2.2). In addition to the inconsistent findings, there is
uncertainty as to whether the NC>2 concentrations from an average of area central sites or
one central site represent the variability in NC>2 concentrations across the study area or
subjects' ambient exposure, particularly in the many cross-sectional studies that make up
the evidence base. Inconsistencies also were found between studies that measured NO2 at
sites located 2 km from children's schools (Chang etal.. 2012; Steerenberg et al.. 2001).
Repeated measures and cross-sectional studies found associations with adjustment for
time-varying factors such as weather as well as between-subject factors such as height,
weight, smoking exposure, and SES. However, copollutant confounding was not
examined, and lung function also was associated with the traffic-related pollutants CO
and BS (Chang etal.. 2012: Steerenberg etal.. 2001) as well as PMio, SO2, and O3.
A fairly large body of studies, conducted in various European countries, does not strongly
support NO2-associated decrements in home measurements of PEF in children. These
studies were similar to studies of supervised lung function in that they examined
populations that included children with respiratory symptoms, asthma, or atopy and
measured NO2 concentrations at central sites and schools. Outdoor school NO2
concentrations were associated with an increase in PEF in children with 25% wheeze
prevalence (Peacock et al.. 2003). Associations with central site NO2 tended to be
positive (Roemer et al.. 1998; Timonen and Pekkanen. 1997) or null (Ranzi et al.. 2004;
Ward et al.. 2000; van der Zee et al.. 1999). A recent study found an NO2-associated
decrease in PEF among children that was independent of CO (Correia-Deur et al.. 2012)
(Table 5-30) but did not report information to assess whether NO2 and CO averaged
across multiple sites in the city adequately represented exposure or had comparable
exposure measurement error.
5-189
-------
Table 5-30 Epidemiologic studies of lung function in children and adults in the general population.
Study
Population Examined and
Methodological Details
Oxide of Nitrogen
Metrics Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
Children in the General Population
Linnetal. (1996)
Upland, Rubidoux, Torrance, CA
n = 269, 4th-5th grades
Repeated measures. Supervised spirometry.
Examined 1 week/season for 2 yr. Recruitment
from schools. 75-90% follow-up participation
across communities. Repeated measures ANOVA
adjusted for year, day, temperature, rain. Time
spent outdoors = 101-136 min across seasons
and communities.
NO2-central site
24-h avg
# sites NR, no site in
Torrance
r = 0.63 correlation
with personal NO2
p.m. FEVi:
-5.2mL(-13, 2.3)
p.m. FVC:
-3.6 ml_ (-12, 4.6)
Diurnal change FEVi:
-7.8 (-14, -1.5)
Diurnal change FVC:
-2.2 (-9.6, 4.9)
No quantitative results. NO2
association reported to lose
statistical significance with
adjustment for PlVh.s measured at
school. Lung function weakly
associated with Os.
Weak correlation with PlVhs.
r=0.25.
tCastro et al. (2009)
Rio de Janeiro, Brazil
n = 118, ages 6-15 yr, 18.4% with asthma
Repeated measures. Supervised PEF.
Recruitment from school. Examined daily for
6 weeks. 9-122 observations per subject. No
information on participation rate. Mixed effects
model with random effect for subject and adjusted
for weight, height, sex, age, asthma, smoking
exposure, time trend, temperature, relative
humidity.
NO2-school outdoor
24-h avg
School was within
2 km of homes
PEF (L/min):
1 0.04 (-0.58, 0.65)
1-2 avg -0.60 (-1.3, 0.14)
1-3 avg -0.83 (-1.7, 0.02)
No copollutant model.
Associations also found with PM-io.
Associations with CO and SO2
had wide 95% CIs.
5-190
-------
Table 5-30 (Continued): Epidemiologic studies of lung function in children and adults in the general population.
Study
Population Examined and
Methodological Details
Oxide of Nitrogen
Metrics Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
Scarlett et al. (1996)
Surrey, U.K.
n = 154, ages 7-11 yr, 9% with wheeze
Repeated measures. Supervised spirometry.
Examined daily for 6 weeks. Recruitment from
school. No information on participation rate. Lung
function adjusted for machine, operator, day of
week. Individual subject regressions adjusted for
temperature, humidity, pollen. Pooled estimates
obtained using weighting method.
NO2-school outdoor
1-h max
FEVo./s: 0.30% (-0.29, 0.89)
FVC: 5.5% (-5.1, 17)
No copollutant model.
Association found with PM-io.
No to moderate correlations with
NO2.
r = 0.07 for PM-io, 0.50 for 8-h max
03.
Moshammer et al. (2006) NCb-central site
Linz, Austria 8-h avg
n = 163, ages 7-10 yr (12 a.m.-8 a.m.)
Repeated measures. Supervised spirometry. Site adjacent to
Examined every 2 weeks for school yr. school
Recruitment from schools. No information on
participation rate. GEE model, covariates not
specified.
FEVi: -4.1% (-6.4,-1.7)
FVC: -2.7% (-5.1, -0.33)
With PM2.5:-4.7% (-7.3, -2.0)
PM2.5 results attenuated or
become positive. Associations
also found for PM-i, PM-io.
Moderate correlations with NO2.
r= 0.53 for PMi, 0.54 for PlVh.s,
0.62 for PMio.
tAltuqetal. (2014)
Eskisehir, Turkey; Feb-Mar 2007
n = NR, ages 9-13 yr, no upper respiratory
symptoms
Cross-sectional. Supervised spirometry.
Recruitment from schools of participants of a larger
study. No information on participation rate. Logistic
regression adjusted for sex, age, asthma, parental
smoking, coal or wood stove use, parental
education, height, weight, daily average
temperature.
NO2-outdoor school
24-h avg
1 site at each of
16 schools
0-6 avg FEVi: 0% (-14, 17)
FVC: 3.8% (-7.3, 16)
No copollutant model.
Os associated with PEF only.
Strong inverse correlation with
NO2. Pearson r= -0.80.
NO2 and PlVh.s reported to be
highly correlated.
5-191
-------
Table 5-30 (Continued): Epidemiologic studies of lung function in children and adults in the general population.
Study
Population Examined and
Methodological Details
Oxide of Nitrogen
Metrics Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
Oftedal et al. (2008)
Oslo, Norway
n =2,170, ages 9-10 yr, 5.5% with asthma
Cross-sectional. Supervised spirometry.
Recruitment from a birth cohort. 67% participation,
60% follow-up. Examined subjects had more
"Westernized" parents. Linear regression adjusted
for age, sex, height, BMI, current asthma, early life
maternal smoking, parental ethnicity, education,
smoking, and atopy, lag 1-3 temperature,
neighborhood variables (% married, % with
income < median, etc.), long-term NO2.
NO2-dispersion model 1-3 avg
NO2-central site 1-7 avg
24-h avg 1 -30 avg
1 city site
Quantitative results not
reported.
Association observed with lag
1-3-day avg and 1-7-day avg.
Larger effect estimated for lag
1-30-day avg.
Central site no association.
No copollutant model.
No association reported for PlVh.s.
Correlations among
pollutants = 0.83-0.95.
Short-term association attenuated
with adjustment for early or
lifetime NO2. r= 0.46-0.77 among
NO2 metrics.
Steerenberq et al. (2001)
Utrecht and Bilthoven, the Netherlands
n = 126, ages 8-13 yr, 28% respiratory disease,
20% allergy
Repeated measures. Supervised PEF. Examined
1/week for 7-8 weeks. Recruitment from urban and
suburban schools. 65% participation. Mixed effects
NO2-central site
15-h avg
(8 a.m.-11 p.m.)
24-h avg
Site within 2 km of
schools
PEF (mL/min):
0 Urban:-17 (-35,0)
Suburban: 7, p > 0.05
0-2 avg Urban: 0, p > 0.05
Suburban: 6, p > 0.05
No copollutant model.
BS associated with PEF.
Correlation with NCband NO NR.
Association also found with PM-io.
model adjusted for sex, age, # cigarettes smoked
in home, presence of a cold, history of respiratory
symptoms and allergy. No consideration for
potential confounding by meteorological factors.
tChanqetal. (2012)
Taipei, Taiwan
n =2,919, ages 12-16 yr
Cross-sectional. Supervised spirometry.
Recruitment from schools. No information on
participation rate. Regression model adjusted for
residence in district, age, sex, height, weight,
temperature, rainfall.
NO-central site
1 5-h avg
24-h avg
NO2-central site
4-h avg
(8 a.m.-12 p.m.)
1 0-h avg
(8 a.m. -6 p.m.)
Average of 5 city sites
within 2 km of schools
15-h avg Urban: 1, p > 0.05
Suburban: 0, p > 0.05
0-2 avg Urban: -6 (-12, 0)
Suburban: 6, p > 0.05
FEVi (mL): No copollutant model.
0 -25 (-57, 7.5) Associations also found with SO2,
1 -41 (-70, -11) CO, Os, PMio.
2 -2.5 (-50, 45)
5-192
-------
Table 5-30 (Continued): Epidemiologic studies of lung function in children and adults in the general population.
Study
Population Examined and
Methodological Details
Oxide of Nitrogen
Metrics Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
tEenhuizen etal. (2013)
3 study areas in the Netherlands
n = 880, age 8 yr
Cross-sectional. Recruitment from intervention
study of mattress allergy covers. Valid data on
49% subjects, who had higher parental education,
less likely to have pets. Linear regression adjusted
for sex, age, height, weight, prenatal smoke
exposure, smoking in home, gas stove, parental
allergy, dampness in home, parental education,
season, temperature, humidity.
NO2-central site
1 site
Interrupter resistance
0 (kPAxs/L):
1 0 (-0.04, 0.04)
-0.02 (-0.06, 0.03)
Positive effect estimate
indicates increase in resistance
No associations with PM-io or BS.
Moderate correlations with NO2.
Pearson r= 0.47 for PM-io, 0.60 for
BS.
tBaqheri Lankarani et al. (2010)
Tehran, Iran
n = 562, elementary school age
Repeated measures. Examined daily for 6 weeks.
No information on participation rate.
158 case-days. Case crossover with control dates
as 2 weeks before and after case date. Conditional
logistic regression adjusted for daily temperature,
lag 0-6-day avg PM-io.
NO-central site
24-h avg
2 city sites
0-6 avg
PEF <50% predicted:
OR: 18(1, 326)
No copollutant model.
PM-io associated with decreased
odds of large PEF decrement.
tPadhiand Padhy(2008)
West Bengal, India
n = 755 from biomass fuel homes, 372 from
liquefied petroleum gas homes, ages 5-10 yr
Cross-sectional. Supervised spirometry.
Recruitment method and participation not reported.
Multiple regression adjusted for unspecified
covariates.
NO2-indoor home
24-h avg
NR Biomass fuel homes
Lung function units NR
FEVi: -1.05 (-1.75,-0.35)
FVC: -1.09 (-1.58, -0.61)
Liquefied gas petroleum homes
FEVi: -5.41 (-8.33, -2.50)
FVC: -5.17 (-9.17, -1.17)
No copollutant model.
CO also associated with lung
function. Correlation with NO2NR.
SPM, SO2, Os also associated
with lung function.
5-193
-------
Table 5-30 (Continued): Epidemiologic studies of lung function in children and adults in the general population.
Study
Population Examined and
Methodological Details
Oxide of Nitrogen
Metrics Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
tCorreia-Deur et al. (2012)
Sao Paolo, Brazil
n = 31, ages 9-11 yr, no allergic sensitization
Repeated measures. Daily supervised spirometry
for 15 school days. Number of observations not
reported. Recruitment from school. 86%
participation. Allergic sensitization ascertained by
skin prick test, blood eosinophils, and serum IgE.
GEE with autoregressive correlation matrix
adjusted for date, school absence, temperature,
humidity.
NO2-outdoor school
24-h avg
PEF: -1.0% (-1.7, -0.35)
Lag 0, all subjects
With CO: -1.5% (-3.0, 0)
Moderate correlation with NO2.
r= 0.51. CO association persists
with NO2 adjustment.
withSO2: -1.9% (-3.3, -0.4)
with PMio: -0.8% (-4.4, 3.1)
withOs: -1.5% (-3.3, 0.38)
Moderate correlations with NO2. r
= 0.59, 0.60, 0.40.
03 association persists with NO2
adjustment. SO2 & PMio
attenuated.
Peacock et al. (2003)
Rochester-upon-Medway, U.K.
n = 177, ages 7-13 yr, 25% with wheeze
Repeated measures. Home PEF. Examined daily
for 13 weeks. 14-63 observations/subject.
Recruitment from rural and urban schools. No
information on participation rate. Individual subject
regressions adjusted for day of week, date,
temperature. Estimates pooled using weighting
method.
NO2-outdoor school
24-h avg
0-4 avg PEF: -0.20 (-3.0, 2.6)
OR for PEF > 20%:
2.3(1.0, 5.4)
1-h max
PEF: 1.2 (-1.5, 3.9)
OR for PEF > 20%:
1.3(0.5, 3.4)
No copollutant model.
PM2.5 also associated with PEF
decrement >20%.
Correlation NR.
van derZee et al. (1999)
Rotterdam, Bodegraven/Reeuwijk, Amsterdam,
Meppel, Nunspeet, the Netherlands
n = 633, ages 7-11 yr, 50% with symptoms, 33%
with asthma
Repeated measures. Home PEF. Examined daily
for 3 mo. Recruitment from school and mail. 47%
responded to initial survey, 80% follow-up
participation. Logistic regression adjusted for
minimum temperature, day of week, time trend,
influenza.
NO2-central site
24-h avg
1 site per community
ORs
0 Urban: 0.96(0.79, 1.2)
Suburban: 0.77(0.54, 1.1)
0-4 avg Urban: 1.1 (0.93, 1.3)
Suburban: 0.99(0.72, 1.4)
Associations found for PMio, BS,
SO4, and SO2 in urban area.
Correlations NR.
5-194
-------
Table 5-30 (Continued): Epidemiologic studies of lung function in children and adults in the general population.
Study
Population Examined and
Methodological Details
Oxide of Nitrogen
Metrics Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
Roemeret al. (1998)
Germany, Finland, the Netherlands, Czech
Republic, Norway, Italy, Greece, Hungary,
Sweden—26 locations
n = 2,010, ages 6-12 yr, atopy prevalence: 7-81%
Repeated measures. Home PEF. Examined daily
for 2 mo. 85% of enrolled included in analysis.
Regression model adjusted for minimum
temperature, school day, time trend. Individual
panel results combined in a meta-analysis.
NO2-central site
24-h avg
0
0-6 avg
PEF (L/min):
0.15 (-0.19, 0.49)
0.23 (-1.2, 1.6)
Association found with PM-io and
BS, but not consistently across
lags.
Ranzi et al. (2004)
Emiglia-Romagna, Italy
n = 118, ages 6-11 yr, 77% with asthma, 67% with
atopy
Repeated measures. Home PEF. Examined daily
for 12 weeks. 98.4% follow-up participation.
Recruited from schools. GLM adjusted for sex,
medication use, symptoms, temperature, humidity
NO2-central site
24-h avg
# sites NR
No quantitative data. Figure
shows no association in group
with and without atopy.
PM2.5 associated with PEF in
urban group.
Ward et al. (2000)
West Midlands, U.K.
n = 147, age 9 yr, 24% with symptoms, 31% with
atopy
Repeated measures. Home PEF. Examined daily
for two 8-week periods. Recruitment from schools.
Individual subject regressions adjusted for time
trend, day of week, meteorological variables,
pollen count. Individual regressions pooled with
weighting method.
NO2-central site
24-h avg
2 sites
0, 1,2, 3, 0-4
avg
No quantitative data. Figure
shows no association across
lags, except at Lag Day 0 in
symptomatic group.
No copollutant model.
Associations with PlVh.s equally
inconsistent.
5-195
-------
Table 5-30 (Continued): Epidemiologic studies of lung function in children and adults in the general population.
Study
Population Examined and
Methodological Details
Oxide of Nitrogen
Metrics Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
Timonen and Pekkanen (1997)
Kuopio, Finland
n = 169, ages 7-12 yr, children with cough
Repeated measures. Home PEF. Examined daily
for 3 mo. Recruitment from schools. 86%
participation. Linear mixed model adjusted for time
trend, weekend, minimum temperature, relative
humidity.
NO2-central site
24-h avg
# sites NR
26% missing data
were modeled,
r=0.58
FEVi
Urban: 11 (-14, 35)
Suburban: -6.5 (-40, 27)
Associations found for SO2 in
urban group. Weak correlations
with NO2. r= 0.22.
1 -4 avg
PEF
Urban: 13 (-24, 50)
Suburban: -22 (-87, 43)
Adults in the General Population
tStraketal. (2012)
Utrecht area, the Netherlands
n = 31, adults ages 19-26 yr, all healthy,
nonsmoking
Repeated measures. Supervised spirometry.
Examined 3-7 times. 107 observations.
Recruitment from university. No information on
participation rate. Well-defined outdoor exposures
at various traffic/nontraffic sites. Heart rate
maintained during intermittent exercise. Higher
probability of associations found by chance alone.
Mixed effects model adjusted for temperature,
relative humidity, season, high/low pollen,
respiratory infection.
NO2-personal outdoor
5-h avg 0-h
Measured next to 2-h
subjects during -|8-h
outdoor exposures
FVC post-exposure:
-4.3% (-7.4, -1.0)
-3.5% (-6.5, -0.43)
-4.5% (-7.4, -1.4)
NOx-personal outdoor
5-h avg
0-h
2-h
18-h
-1.6% (-2.6, -0.51)
-2.0% (-4.9, -0.16)
-2.5% (-5.4, -0.69)
FVC with PNC:
NO2: -3.0% (-7.2, 1.4)
NOx: -0.11% (-2.6, 2.5)
Moderate to high correlation with
NO2. Spearman r= 0.56, 0.75.
PNC association attenuated with
adjustment for NO2 or NOx
5-196
-------
Table 5-30 (Continued): Epidemiologic studies of lung function in children and adults in the general population.
Study
Population Examined and
Methodological Details
Oxide of Nitrogen
Metrics Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
tDalesetal. (2013)
Sault Ste. Marie, ON, Canada
n = 59, adults mean (SD) age 24.2 (5.8) yr, all
healthy
Repeated measures. Supervised spirometry.
Examined 10 times. Total observations NR.
Recruitment from university. No information on
participation rate. Well-defined outdoor exposures
near steel plant and university campus 4.5 km
away. Exposures occurred at rest except for
30-min exercise to increase heart rate to 60%
predicted maximum. Mixed effect model with
autoregressive correlation matrix and adjusted for
site, day of week, mean temperature, humidity.
Well-defined outdoor exposures near steel plant
and university campus 4.5 km away. Exposures
occurred at rest except for 30-min exercise to
increase heart rate to 60% predicted maximum.
Mixed effect model with autoregressive correlation
matrix and adjusted for site, day of week, mean
temperature, humidity.
NO2-on site of
outdoor exposure
10-h avg
(8 a.m.-6 p.m.)
0-h % predicted FEVi:
Post-exposure -10.9 (-13.3,-8.6)
% predicted FVC:
-9.2 (-14.5, -3.9)
No copollutant model.
Associations found with UFP and
PM2.5. Correlations NR. All
pollutants higher at steel plant
than at university campus.
Associations also found with SO2
and 03.
tWeichenthal et al. (2011)
Ottawa, ON, Canada
n = 42, adults ages 19-58 yr, from nonsmoking
homes, 95% white, 62% with allergies, 33% with
asthma
Repeated measures. Supervised spirometry. Most
examined 3 times. 118 observations. 1-h outdoor
exposures during cycling in low and high traffic
areas. Recruitment from public advertisements. No
information on participation rate. Differential
exposure measurement error for personal PM and
VOCs and central site NO2. Mixed effects models
with random subject effect adjusted for
temperature during cycling, average heart rate.
Adjustment for relative humidity, day of week did
not affect results.
NO2-central site FEVi (L):
1-h avg 1-h 0.54 (-0.15, 1.2)
1 site 4-h 0.40 (-0.12, 0.92)
Post-exposure
No copollutant model.
Lung function not associated with
O3 or VOCs, UFP, BC, PM2.5.
5-197
-------
Table 5-30 (Continued): Epidemiologic studies of lung function in children and adults in the general population.
Study
Population Examined and
Methodological Details
Oxide of Nitrogen
Metrics Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
tThaller et al. (2008)
Galveston, TX
n = 142, lifeguards at work, ages 16-27 yr, 13%
with asthma, 22% with allergies
Repeated measures. Supervised spirometry.
Recruitment from worksite. 1,140 observations.
Self-report of physician-diagnosed asthma. 81%
follow-up participation. GLM, covariates not
specified.
NO2 & N Ox-centra I
site
24-h avg, 1-h max
1 site 4-12 km from
beaches
No quantitative data. NO2 and
NOx reported not to be
significantly associated with
lung function.
No copollutant model.
Schindler et al. (2001)
Aarau, Basel, Davos, Geneva, Lugano, Montana,
Payerne, Wald, Switzerland
n = 3,912, ages 18-60 yr, nonsmokers
Cross-sectional. Supervised spirometry.
Recruitment from registry and SALPADIA cohort.
Sample representative of full cohort. Regression
model adjusted for sex, age, height, weight, day of
week, temperature, relative humidity. Adjustment
for asthma medication or wheeze did not alter
results.
NO2-central site
24-h avg
1 site per city
FEVi:
0 -2.5% (-4.5, -0.48)
0-3 avg -2.9% (-5.9, 0.21)
WithTSP: -1.2% (-3.8, 1.6)
van derZee et al. (2000)
Rotterdam, Bodegraven/Reeuwijk, Amsterdam,
Meppel, Nunspeet, the Netherlands
n = 274, ages 50-70 yr, no symptoms in previous
12 mo
Repeated measures. Home PEF. Examined daily
for 3 mo. Recruitment from mailings. 81% enrolled
included in final analysis. Logistic regression
adjusted for minimum temperature, day of week,
time trend, influenza.
NO2-central site
24-h avg
1 site per community
OR for PEF decrease >10%
Urban: 0.85(0.59, 1.2)
Suburban: 0.72(0.50, 1.05)
0-4 avg Urban: 0.46 (0.20, 1.08)
Suburban: 0.56(0.27, 1.16)
No copollutant model.
PEF associated with PM-io and
SO4 in urban group.
- Wide range of correlations with
NO2. Spearman r= 0.16-0.72 for
PMio, 0.25-0.65 for BS.
5-198
-------
Table 5-30 (Continued): Epidemiologic studies of lung function in children and adults in the general population.
Study
Population Examined and
Methodological Details
Oxide of Nitrogen
Metrics Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
tCakmaketal. (2011 a)
15 cities, Canada
n = 5,01 1 , ages 6-79 yr, mean age 39 yr
NO2-central site
24-h avg
# sites NR
% predicted FEV-i:
0 -1.6 (-2.9, -0.35)
No copollutant model.
03 and PIVh s also associated with
lung function. Correlations NR.
Cross-sectional. Supervised spirometry.
Recruitment by random sampling of households.
No information on participation rate. GLMM
adjusted forage, sex, income, education, smoking,
random effect for site. Adjustment for temperature
and relative humidity did not alter results.
tLepeule et al. (2014), Lepeule (2014)
Boston, MA area
n = 776, all male, mean (SD) age at baseline 72.3
(6.8) yr, Normative Aging Study
Longitudinal. Supervised spirometry. Examined
1-4 times over 10 yr. No information on
recruitment over follow-up participation. Linear
mixed effects model adjusted for age, log-height,
race, education, standardized weight, smoking
status, pack-years smoking, chronic lung condition,
methacholine responsiveness, medication use
season, day of week, visit number, temperature,
humidity. Adjusting for cardiovascular disease did
not alter results.
NO2-central site
24-h avg
Average of 5 sites in
Boston area. Median
21.4 km from
subjects' homes.
FEVi:
0 -0.08% (-1.92, 1.80)
0-2 avg -1.00% (-3.45, 1.51)
0-27 avg -13.0% (-17.9, 7.75)
Low IL-6 gene methylation:
-11.6% (-17.5, -5.32)
High IL-6 gene methylation:
-13.0% (-18.7, -6.95)
No copollutant model.
Associations found with BC, CO,
PM2.5. Moderate correlation with
NO2. Spearman r= 0.59, 0.52,
0.62, respectively.
BC and PM2.5 measured at one
Boston site.
Association also found with Os.
r=-0.31.
tSteinvil et al. (2009)
Tel Aviv, Israel
n = 2,380, mean age 43 (SD: 11) yr, healthy
nonsmokers
Cross-sectional. Supervised spirometry.
Recruitment from ongoing survey of individuals
attending health center. No information on
participation rate. Linear regression adjusted for
sex, age, height, BMI, exercise intensity,
education, temperature, relative humidity, season,
year.
NO2-central site
24-h avg
3 sites within 11 km
of homes
0
5
0-6 avg
FEVi (mL):
-16 (-64, 33)
-55 (-103, -6.3)
-97 (-181, -13)
w/CO(lag5):-19(-88, 50)
w/SO2 (lag 5): -7.8 (-72, 56)
SO2 and CO results persist with
adjustment for NO2.
High correlations with NO2.
Pearson r = 0.75 for CO, 0.70 for
SO2.
5-199
-------
Table 5-30 (Continued): Epidemiologic studies of lung function in children and adults in the general population.
Study
Population Examined and
Methodological Details
Oxide of Nitrogen
Metrics Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
tSon etal. (2010)
Ulsan, South Korea
n ~~ 9 1 09 mpan anp 4^ fm~V 17^ \/r mpan %
predicted FEVi: 83%
Cross-sectional. Supervised spirometry.
Recruitment during meeting of residents. No
information on participation. Regression model
adjusted forage, sex, BMI. Did not consider
potential confounding by weather, season, or time
trend. High correlation among exposure
assessment methods. r= 0.84-0.96.
NO2-central site
13 site average
Nearest site
Inverse distance
weighting
Kriging
All 24-h avg
% predicted FVC:
0-2 avg -7.9 (-10, -5.6)
-6.9 (-8.8
-6.9 (-9.1
-7.4 (-9.8
, -5.0)
, -4.7)
,-5.1)
Associations found with PM-io, Os,
SO2, CO. NO2 effect estimate
slightly reduced with adjustment
for Os. No copollutant model with
PMioorSO2.
tAqarwal etal. (2012) NCfe-central site
5 locations with agricultural burning around Patiala 24-h avg
City, Punjab, India. 1 site per |ocation
n = 50, ages 13-53 yr, 80% adults, no respiratory
conditions
Repeated measures. Supervised spirometry.
Examined 2 times/mo for 6 mo in each of 3 years.
Total observations NR. No information on
recruitment method. 40% follow-up participation.
Linear regression. Did not report whether
covariates were included.
1-moavg FEVi: -8.9%, p = 0.054
FVC: -7.5%, p = 0.064
No copollutant model.
Association found with PlVh.s.
Correlation with NO2 NR.
Association also found with PM-io
and SO2.
Note: Studies are organized by population examined, and more informative studies in terms of exposure assessment method and potential confounding considered are presented
first.
a.m. = ante meridiem; ANOVA = analysis of variance; avg = average; BC = black carbon; BMI = body mass index; BS = black smoke; CA = California; Cl = confidence interval;
CO = carbon monoxide; Feb = February; FEV075 = forced expiratory volume in 0.75 seconds; FENA, = forced expiratory volume in 1 second; FVC = forced vital capacity;
GEE = generalized estimating equations; GLM = Generalized linear model; GLMM = generalized linear mixed model; (kPAxs/L): = kilopascals times seconds per liter;
IgE = immunoglobulin E; IL = interleukin; MA = Massachusetts; max = maximum; NO = nitric oxide; NO2 = nitrogen dioxide; NOX = sum of NO and NO2; NR = not reported;
O3 = ozone; ON = Ontario; OR = odds ratio; PEF = peak expiratory flow; PM2 5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm;
PMio = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; PNC = particle number concentration; SALPADIA = Study on Air Pollution and
Lung Disease in Adults; SD= standard deviation; SO2 = sulfur dioxide; SO4 = sulfate; SPM = suspended particulate matter; TSP = total suspended particles; TX = Texas;
UFP = ultrafine particles; U.K. = United Kingdom; VOC = volatile organic compound.
aEffect estimates were standardized to a 20-ppb increase in 24-h avg NO2 and a 30-ppb increase 1-h max NO2. Effect estimates for other averaging times (1-h avg to 15-h avg) are
not standardized but presented as they are reported in their respective studies (Section 5.1.2.2).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-200
-------
Adults in the General Population
In studies of adults in the general population reviewed in the 2008 ISA for Oxides of
Nitrogen (U.S. EPA. 2008c). increases in ambient NC>2 concentration were associated
with decrements in lung function as measured by supervised spirometry (Schindler et al..
2001) but not home peak flow (van der Zee et al.. 2000). Recent studies conducted
supervised spirometry, and while the results are inconsistent overall, the studies with
stronger exposure assessment and/or examination of copollutant confounding indicate
ambient NCh-associated decreases in lung function in healthy adults. Overall, studies
examined a wide range of ages (i.e., 18-79 years) and a mix of healthy populations and
those including adults with asthma or allergies, but these factors did not appear to
influence the results, van der Zee et al. (2000) found no association in adults with or
without respiratory symptoms. Locations, time periods, and ambient concentrations of
oxides of nitrogen for these studies are presented in Table 5-30.
Many studies that found lung function decrements in adults in the general population in
association with higher ambient NO2 concentrations do not strongly inform the
independent effects of NO2 exposure (Lepeule etal.. 2014; Cakmak et al.. 201 la: Son et
al., 2010; Steinvil et al.. 2009). A major uncertainty is potential confounding. Son et al.
(2010) did not examine confounding by meteorological or other time-varying factors.
Studies found associations with PIVb 5 and the traffic-related pollutants CO and BC,
which were moderately to highly correlated with NO2 (r = 0.56-0.75) (Lepeule et al..
2014; Agarwal etal.. 2012; Cakmak etal.. 201 la; Sonet al.. 2010; Steinvil etal.. 2009).
Copollutant models were not analyzed, except in Steinvil et al. (2009). where the NO2
effect estimate was attenuated with CO adjustment, and NO2 results were mixed among
the various lags examined. Lung function also was associated with PMio, total suspended
particles (TSP), SO2, and Os, and in copollutant models, NO2 associations remained with
adjustment for TSP or Os (Son etal.. 2010; Schindler et al.. 2001) but not for highly
correlated SO2 (r = 0.70) (Steinvil et al.. 2009). Another uncertainty is whether NO2
concentrations measured at the nearest central site, one site, averaged across multiple
sites, or spatially interpolated by inverse distance weighting or kriging were equally
representative of ambient exposure among subjects distributed within a city or across
multiple communities. Differences in exposure measurement error between subjects may
influence results, particularly in cross-sectional studies (Cakmak et al.. 201 la; Son et al..
2010; Steinvil et al.. 2009; Schindler et al.. 2001) and a longitudinal study collecting one
to four measures of lung function over 10 years (Lepeule et al.. 2014).
Ambient concentrations may better represent exposures in situations when people are
outdoors. In adults cycling in various traffic and nontraffic locations or lifeguards
working outdoors, lung function before and after repeated outdoor exposures were not
5-201
-------
associated with NO2 assessed from a central site (Weichenthal et al.. 2011; Thaller et al..
2008). However, in healthy adults, lung function decrements were associated with NO2
measured on site of outdoor activity in locations that varied in traffic (Strak etal.. 2012)
or distance from a steel plant (Dales et al.. 2013). Lung function decreased immediately
after and 2 to 18 hours after the outdoor exposure period (Dales et al., 2013; Strak et al..
2012). These studies have stronger inference than the aforementioned central site studies
because NO2 measurements are aligned with subjects in both time and space. Both
outdoor exposure studies found associations with PM2 5 and/or the highly correlated
(r = 0.67-0.87) traffic-related PM2s absorbance, EC, metal components of PM2s, and
UFP/particle number concentration (PNC). Only Strak etal. (2012) examined copollutant
models and found that NO2 associations persisted with adjustment for PNC, EC, PM2 5, or
another PM2 5 component. A 25-ppb increase in 5-h avg NO2 was associated with a
-4.3% (95% CI: -7.4, -1.0) change in FVC and a -3.0% (95% CI: -7.2, 1.4) change
with adjustment for PNC. The NOx association was attenuated with adjustment for PNC.
Effect estimates for EC, absorbance, and PNC were attenuated with adjustment for NO2,
indicating that NO2 may have confounded associations for copollutants.
Controlled Human Exposure Studies
Similar to the epidemiologic studies, controlled human exposure studies generally did not
report effects of NO2 on lung function in healthy adults. Overall, exposures ranged from
200 to 4,000 ppb NO2 for 40 minutes to 5 hours, and most studies incorporated exercise
in the exposure period to assess lung function during various physiological conditions
(Table 5-31). As examined in many studies, NO2 exposures of 120-600 ppb (40 minutes
to 4 hours) did not affect lung function of healthy adolescents (Koenig etal.. 1987).
young adults (Huang et al.. 2012b: Frampton et al.. 2002; Vagaggini etal.. 1996;
Hazuchaetal.. 1994; Frampton et al.. 1991; Frampton et al.. 1989; Adams et al.. 1987).
or older adults (Gong etal.. 2005; Morrow et al.. 1992). NO2 exposures in this range
(400, 500 ppb) did not show additive or synergistic effects with exposure to concentrated
ambient particles (CAPs) in the size range of PM2 5 (Huang et al.. 2012b: Gong et al..
2005). which in the ambient air often is moderately to highly correlated with NO2
(Figure 3-6). Lack of additivity or synergy also was observed for co-exposures of NO2
and Os, which often are weakly correlated in the ambient air (Figure 3-6). Decreases in
lung function and increases in airway resistance in response to Os exposure were not
affected by co-exposures to 600 ppb NO2 (Hazuchaetal.. 1994; Adams etal.. 1987).
Lung function generally was not altered by higher NO2 exposures of 1,000-4,000 ppb
(1.5 to 5 hours) (Frampton et al.. 2002; Devlin etal.. 1999; Torres etal.. 1995; Frampton
etal.. 1989; Linn et al.. 1985b). including intermittent spikes of 2,000 ppb NO2
superimposed on a 3-hour background exposure to 50 ppb NO2 (Frampton et al.. 1991).
5-202
-------
A few controlled human exposure studies did report NCh-induced changes in lung
function in healthy adults. Rasmussen et al. (1992) observed statistically nonsignificant
increases in FVC and FEVi during and after a 5-hour exposure to 2,300 ppb NCh. NCh
exposures repeated over multiple days had contrasting effects, with a 4-hour exposure to
2,000 ppb NC>2 inducing a decrease in FEVi and FVC only on the first day of a 4-day
exposure (Blomberg et al.. 1999). and a 2-hour exposure to 1,000 ppb NO2 inducing a
1.5% drop in FVC only on the second day of a 2-day exposure (Hackney et al.. 1978).
Table 5-31 Characteristics of controlled human exposure studies of lung
function and respiratory symptoms in healthy adults.
Disease Status3; Sample
Study Size; Sex; Age (mean ± SD)
Exposure Details
Endpoints Examined
Koeniq et al.
(1987)
Healthy;
(1)n = 3M, 7F
(2) n = 4 M, 6 F
Asthma;
(1)n=4M,6F
(2)n = 7M, 3F;
14.4 yr (range: 12-19)
(1) 120ppbNO2,
(2)180ppbNO2;
(1-2) Exposures were 30 min at
rest with 10 min of moderate
exercise
Pulmonary function tests
before, during, and after
exposure.
tHuanq et al.
(2012b)
(1)n = 11 M, 3F
(2) n = 6 M, 7 F
(3) n = 7 M, 6 F;
24.6 ± 4.3 yr
(1)500ppbNO2for2h,
(2) 500 ppb
NO2 + 73.4 ± 9.9 ug/m3 CAPs for
2h,
(3) 89.5 ± 10.7 ug/m3 CAPs for 2 h;
(1-3) Exercise 15 min on/15 min
off at VE = 25 L/min
Pulmonary function tests
before, immediately after, and
18 h after exposure.
Frampton et (1,2) n = 12 M, 9 F;
al. (2002) F = 27.1 ± 4.1 yr
M = 26.9 ±4. Syr
(1) 600 ppb for 3 h,
(2) 1,500 ppb for 3 h;
(1,2) Exercise 10 min on/20 min off
at VE = 40 L/min
Pulmonary function tests
before and after exposure.
Vaqaqqini et
al. (1996)
Healthy; n = 7; M;
34 ± 5 yr
Asthma; n = 4 M, 4 F;
29 ± 14 yr
COPD; n = 7; M;
58 ± 12 yr
300 ppb for 1 h;
Exercise at VE = 25 L/min
Pulmonary function tests
before and 2 h after exposure.
5-203
-------
Table 5-31 (Continued): Characteristics of controlled human exposure studies of
lung function and respiratory symptoms in healthy
adults.
Disease Status3; Sample
Study Size; Sex; Age (mean ± SD)
Exposure Details
Endpoints Examined
Hazucha et al. n = 21; F; 22.9 ± 3.6 yr
(1994)
(1) 600 ppb NO2 for 2 h, air for 3 h,
300 ppb 03 for 2 h,
(2) Air for 5 h, 300 ppb O3 for 2 h;
(1,2) Exercise for 15 min on/15 min
off at VE = 35 L/min
Pulmonary function tests
before, during, and after
exposure; airway reactivity
after exposure.
Times for symptoms
assessment not reported.
Frampton et
al. (1991)
(1) n = 7 M, 2 F; (1) 600 ppb for 3 h,
29.9 ± 4.2 yr (2) 1,500 ppb for 3 h,
(2) n = 12 M, 3 F; 25.3 ± 4.6 yr (3) 50 ppb for 3 h + 2]000 ppb
(3) n = 1 1 M, 4 F; 23.5 ± 2.7 yr peak for 15 min/h;
(1-3) Exercise 10 min on/20 min
off at VE = ~4 times resting
Pulmonary function tests
before, during, and after
exposure; airway reactivity
30 min after exposure.
Symptoms after exposure.
Frampton et (1) n = 7 M, 2 F; 30 yr (range:
al. (1989) 24-37)
(2)n = 11 M, 4F;25yr
(range: 19-37)
(1) 600 ppb for 3 h, Pulmonary function tests
(2) 1,500 ppb for 3 h; before, during, and after
sxDosurs
(1,2) Exercise 10 min on/20 min off
at VE = ~4 times resting
Adams et al.
(1-3)n = 20M, 20 F;
F = 21.4± 1.5yr
M = 22.7 ± 3.3 yr
(1)600ppbNO2for1 h,
(2) 300 ppb O3 for 1 h,
(3) 600 ppb NO2 and 300 ppb Os
for 1 h;
(1-3) Exercise during entire
exposure at VE = 75 L/min (M) and
VE = 50 L/min (F)
Pulmonary function before
and after exposure.
Symptoms following
exposure.
Gong et al.
(2005)
Healthy; n = 2 M, 4 F;
68 ± 11 yr
COPD; n = 9M, 9 F; 72 ± 7 yr
(1)400ppbNO2for2h,
(2)200ug/m3CAPsfor2h,
(3) 400 ppb NO2 + 200 ug/m3
CAPs for 2 h;
(1-3) Exercise 15 min on/15 min
off at VE = ~2 times resting
Pulmonary function tests
before and immediately after
exposure and 4 h and 22 h
after exposure.
Symptoms before, during, and
after exposure.
Morrow et al. Healthy; n = 10 M, 10 F
(1992) (13 never smokers, 4 former
smokers, 3 current smokers)
COPD; n = 13 M, 7F
(14 current smokers, 6 former
smokers);
59.9 ± 7.0 yr
300 ppb for 4 h;
Three 7-min periods of exercise at
VE = ~4 times resting
Pulmonary function tests
before, during, immediately
after, and 24 h after exposure.
Symptoms 24 h after
exposure.
Devlin et al. n = 11; M; range: 18-35 yr
2,000 ppb for 4 h;
Exercise for 15 min on/15 min off
at VE = 50 L/min
Aerosol bolus dispersion
(deposition, FEV-i and sRaw).
5-204
-------
Table 5-31 (Continued): Characteristics of controlled human exposure studies of
lung function and respiratory symptoms in healthy
adults.
Disease Status3; Sample
Study Size; Sex; Age (mean ± SD) Exposure Details
Endpoints Examined
Jorres et al.
(1995)
Healthy; n = 5 M, 3 F;
27 yr (range: 21-33)
Asthma; n = 8 M, 4 F;
27 ± 5 yr
1,000 ppbforS h;
Exercise 10 min on/10 min off at
individual's maximum workload
Symptoms immediately and 6
and 24 h after exposure.
Linn et al.
Healthy; n = 16 M, 9 F; range: 4,000 ppb for 75 min;
20-36 yr
Airway resistance before,
during, and after exposure.
Asthma; n = 12 M, 11 F;
range: 18-34 yr
Two 15-min periods of exercise at
VE = 25 L/min and 50 L/min
Symptoms before, during,
immediately after, and 24 h
after exposure.
Rasmussen et n = 10 M, 4 F; 34.4 yr (range: 2,300 ppb for 5 h
al. (1992). 22-66)
Pulmonary function tests
before, 2 times during, and
3 times after exposure.
Symptoms before, during, and
after exposure.
Blomberq et
al. (1999)
n = 8M, 4F;
26 yr (range: 21-32)
2,000 ppb, 4 h/day for 4 days;
Exercise 15 min on/15 min off at
workload of 75 watts
Pulmonary function before
and after exposure.
Hackney et al. n = 16; M; 26.9 ± 5.0 yr
(1978)
1,000 ppb, 2 h/day for 2 days;
Exercise 15 min on/15 min off at
VE = 2 times resting
Pulmonary function tests
before and after each
exposure.
Symptoms after each
exposure.
CAPs = concentrated ambient particles; COPD = chronic obstructive pulmonary disease; F = female; FEVi = forced expiratory
volume in 1 second; M = male; NO2 = nitrogen dioxide; O3 = ozone; ppb = parts per billion; SD = standard deviation;
sRaw = specific airway resistance; VE = minute ventilation.
aSubjects were healthy individuals unless described otherwise.
fStudy published since the 2008 ISA for Oxides of Nitrogen
5.2.7.3 Respiratory Symptoms in Healthy Populations
Epidemiologic Studies of Children in the General Population
Respiratory symptoms in relation to short-term NC>2 exposure have not been examined in
epidemiologic studies of healthy adults; however, associations are indicated in
school-aged children in the general population (Table 5-32 and 5-33). NCh-associated
increases in respiratory symptoms also were found in infants (Stern et al.. 2013;
Andersen et al.. 2008a: Peel et al.. 2007) (Table 5-33). These results have weaker
5-205
-------
implications because symptoms such as wheeze are common in infancy and may not
clearly distinguish children who do and do not develop respiratory conditions like asthma
later in life (Cano Garcinuno and Mora Gandarillas. 2013). Further, Peel et al. (2007)
examined apnea in infants on home cardiorespiratory monitors, a group unrepresentative
of the general population. Another uncertainty is whether the temporal variation in
ambient NO2 concentrations from one central site in the area adequately represents
variation in ambient NC>2 exposure of infants, particularly those on cardiorespiratory
monitors, who may not spend much time outdoors away from home.
In school-aged children, not all results were statistically significant, but a pattern of
elevated odds ratios indicates consistency in association between short-term NO2
exposure and respiratory symptoms (Table 5-33). Evidence is stronger for cough than
wheeze, which is identified more with asthma. Children were recruited primarily from
schools but also from a birth cohort, suggesting study populations were representative of
the general populations. A wide range of participation rates was reported (Table 5-33).
but no study reported issues with differential participation by a particular group. The
health status of study populations was not always specified, and it is not clear whether the
NO2-associated increases in respiratory symptoms reflect associations among all children
or those with a respiratory disease. For example, associations were reported in
populations with parental history of asthma (Rodriguez et al.. 2007) or with 27% asthma
prevalence (Ward et al.. 2002). Findings for symptoms are uncertain in healthy children.
NC>2 was not associated with respiratory symptoms in children without asthma (Patel et
al.. 2010) but was associated with new diagnosis of asthma in children (Wendt et al..
2014) (Table 5-33). Findings that an increase in a 5-day avg of ambient NCh
concentrations may induce respiratory symptoms that precipitate an asthma diagnosis
have uncertain implications. Asthma diagnosis was ascertained from Medicaid claims as
a record of an outpatient or inpatient visit for asthma or dispensing events of asthma
medication during a three-year period (Wendt etal.. 2014). Among children older than
age 3 years, what is defined as a diagnosis instead could represent an exacerbation of
previously diagnosed asthma. Among children younger than age 3 years, the reliability of
an asthma diagnosis is uncertain. The uncertainty of basing a new asthma diagnosis on a
three-year review of medical records is underscored by observations that NO2
associations were stronger in children older than age 4 years than in children ages
1-4 years (Table 5-33).
5-206
-------
Table 5-32 Mean and upper percentile concentrations of nitrogen dioxide in
epidemiologic studies of respiratory symptoms in children in the
general population.
Study3
Location
Study Period
Upper
Mean/Median Percentile
NO2 Metric Concentration Concentrations
Analyzed ppb ppb
tStern et al. (2013)
fAndersen et al.
(2008a)
tPeeletal. (2011)
Rodriguez et al.
(2007)
Ward et al. (2002)
tPateletal. (2010)
fWendt et al.
(2014)
Schwartz et al.
(1994)
Bern, Basel,
Switzerland
Copenhagen,
Denmark
Atlanta, GA
Perth, Australia
Birmingham,
Sandwell, U.K.
New York City and
nearby suburb, NY
Harris County, TX
(Houston area)
Watertown, MA;
Steubenville, OH;
Topeka, KS; St.
Louis, MO; Portage,
Wl; Kingston-
Harriman, TN
Apr 1999-Feb 2011
Dec 1998-Dec2004
Aug 1998-Dec2002
Jun 1996-Jul 1998
Jan-Mar 1997
May-Jul 1997
2003-2005, mo NR
2005-2007
Apr-Aug 1984-1988
24-h avg NO2
24-h avg NO2
1-h max NO2
1-h max NO2
24-h avg NO2
24-h avg NO2
24-h avg NO2
1-h max NO2
24-h avg NO2
Rural: 8.1b
Urban: 25.6b
11.8
41.7
18
7
18
13.3
NR
39.26
13.3
NR
NR
75th: 14.6
90th: 65.6
Max: 109.2
Max: 48
Max: 24
Max: 35
Max: 29
NR
75th: 48.00
Max: 108
75th: 24.1
Max: 44.2
tMoon et al. (2009) Seoul, Incheon,
Busan, Jeju, South
Korea
Apr-May 2003
24-h avg NO2
NR
NR
Aug = August; avg = average; Dec = December; Feb = February; GA = Georgia; KS = Kansas; MA = Massachusetts;
max = maximum; MO = Missouri; NO2 = nitrogen dioxide; NR = not reported; NY = New York; OH = Ohio; ppb = parts per billion;
TN = Tennessee; TX = Texas; U.K. = United Kingdom; Wl = Wisconsin.
aStudies presented in order of first appearance in the text of this section.
""Concentrations converted from |jg/m3 to ppb using the conversion factor of 0.532 assuming standard temperature (25°C) and
pressure (1 atm).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
Despite the associations found with respiratory symptoms in children, there is uncertainty
regarding the extent to which the results reflect an independent relationship with NC>2.
Ambient NC>2 exposures were assigned from one central site per city or the average
across multiple sites per city. In the study of asthma diagnosis, effect estimates were
5-207
-------
similar for 1-h max NC>2 assigned to subjects as the average NO2 across 17 sites covering
the 4,400-km2 area of Harris County, TX and the nearest site within 9.7 km of the ZIP
code centroid (Wendt et al.. 2014). The two NO2 exposure metrics differed in mean
concentration, 39.3 ppb versus 27.7 ppb, but the temporal variability of each metric was
not reported to assess whether large differences in temporal variability in NO2 occurred
within the study area.
Studies of respiratory symptoms in children also did not adequately examine whether
NO2 associations were independent of PM2 5 and traffic-related copollutants. Symptoms
also were related to PIVb 5, CO, BS, and UFP, which tended to be highly correlated with
NO2 (r = 0.61-0.75) (Wendtet al.. 2014; Moon et al.. 2009; Andersen et al.. 2008a;
Rodriguez et al. 2007; Ward et al.. 2002). In Harris County, TX, 1-h max NCh remained
associated with diagnosis of asthma in a copollutant model with 24-h avg PM2s (Table
5-33) (Wendt et al.. 2014). NO2 was weakly correlated with PM25 (r = 0.21), and while
the variability in ambient PIVb 5 concentrations was not reported, NC>2 concentrations
were reported to vary across the county. In infants, ORs for both NC>2 and UFP decreased
with mutual adjustment (Table 5-33); thus, an independent or confounding effect was not
discerned for either pollutant (Andersen et al.. 2008a). Copollutant models were
examined in the U.S. Six Cities study for PMioand 862. ORs for cough decreased with
PMio or SO2 adjustment to 1.37 (95% CI: 0.98, 2.12) and 1.42 (95% CI: 0.90, 2.22) for a
20-ppb increase in NO2, respectively (Schwartz et al.. 1994). The width of 95% CIs is
inflated when presented for a 20-ppb increase in NO2, which is double the 10-ppb
interquartile range for the study areas. The OR for PMio was robust to NO2 adjustment.
Thus, PMio may partly confound NO2 associations. While the positive ORs for NO2 in
the U.S. Six Cities study suggest an independent association for NO2 as well, the
potentially differential exposure measurement error for central site NO2, SO2, and PMio
limits inference from the copollutant model results. This limited inference also applies to
the aforementioned results from copollutant models with PM2 5 or UFP.
5-208
-------
Table 5-33 Epidemiologic studies of respiratory symptoms in children in the general population.
Study
Population Examined and Methodological Details
NO2 Metrics Analyzed Lag Day
Odds Ratio (95% Cl)
Single-Pollutant Model3
Copollutant Examination
tWendtetal. (2014)
Harris County, TX (Houston area)
n = 18,264 cases in incident asthma, ages 1-17 yr
Case-crossover. Incident asthma cases ascertained for
2004-2007 from Medicaid database. Medicaid enrollment
required only for 13 mo. Date of diagnosis defined as
earliest date of asthma diagnosis on inpatient or outpatient
record or earliest of four asthma medication dispensing
events in a year. Uncertainty as to whether outcome
represents incident asthma. Conditional logistic regression
adjusted for temperature, humidity, mold spores, tree pollen,
grass pollen, weed pollen.
NO2-central site
1-h max
Average of 17 sites in
4,400 km2 area
Mean: 39.3 ppb
Site within 9.7 km of ZIP
code centroid
Mean: 27.6 ppb
0-5 avg Asthma diagnosis
May-Oct
All ages:
1.22(1.09, 1.36)
1 -4 yr:
1.05(1.00, 1.09)
15-17 yr:
1.57(1.06,2.32)
Nov-Apr
All ages:
1.03(0.93, 1.14)
No quantitative results.
Slightly higher OR than
that for 17 site avg.
All ages, NO2 average over
17 county sites, May-Oct.
With 24-h avg PM2.5:
1.20(1.06, 1.36)
With 8-h max O3:
1.11 (0.90, 1.37)
Low or moderate correlations
with NO2. r= 0.21 for PM2.5,
0.49 for 03. 03 means similar
for county average and site
within 9 km.
. PM2.5 and Os attenuated with
adjustment for NO2.
Schwartz et al. (1994)
Watertown, MA; Kingston-Harriman, TN; St. Louis, MO;
Steubenville, OH; Portage, Wl; Topeka, KS
n = 1,844, grades 2-5
Repeated measures. Daily symptom diaries for 5 mo,
collected every 2 weeks. Recruitment from schools. No
information on participation rate. Logistic regression
adjusted for Lag Day 1 temperature, day of week, city.
NO2-central site
24-h avg
1 site per community
0
0-3 avg
1
Cough:
1.21 (0.92, 1.59)
1.61 (1.08,2.40)
Lower respiratory
symptoms:
1.44(0.96,2.16)
For cough
With PMio: 1.37(0.98,2.12)
WithOs: 1.61 (1.08,2.41)
WithSO2: 1.42(0.90,2.22)
PMio and Os robust to
adjustment for NO2. SO2
reduced.
Moderate correlations with
NO2.r =0.36 for PMio, 0.35
for PM2.5, 0.51 for SO2, -0.28
for Os.
5-209
-------
Table 5-33 (Continued): Epidemiologic studies of respiratory symptoms in children in the general population.
Study
Population Examined and Methodological Details
NO2 Metrics Analyzed Lag Day
Odds Ratio (95% Cl)
Single-Pollutant Model3
Copollutant Examination
tMoon et al. (2009) NCb-central site
Seoul, Incheon, Busan, Jeju, South Korea 24-h avg
n =696, ages NR # sites NR
Repeated measures. Daily symptom diaries for 2 mo.
Recruitment from schools. 69% participation rate. GEE
adjusted for temperature, relative humidity.
0 Lower respiratory
symptoms
All subjects:
1.02(1.00, 1.05)
Seoul:
1.08(0.99, 1.18)
Incheon:
1.08(0.99, 1.18)
Busan:
1.04(0.96,1.12)
Jeju:
0.97(0.89, 1.06)
No copollutant model.
Association also found with
CO. Correlation NR.
tPateletal. (2010)
New York City and nearby suburb, NY
n = 192 children without asthma, ages 14-20 yr
Repeated measures. Daily symptom diaries for 4-6 weeks,
collected weekly. Recruitment from schools. Self-report of
physician-diagnosed asthma. 75-90% participation across
schools. GLMM with random effect for subject and school
and adjusted for weekend, 8-h max Os, urban location.
Adjustment for season, pollen counts did not alter results.
NO2-central site
24-h avg
1 site 2.2-9.0 km from
schools, 1 site 40 km from
schools
Wheeze:
0.88(0.75, 1.03)
Chest tightness
0.96(0.75, 1.23)
No copollutant model with BC.
BC also associated with
symptoms. Across locations,
moderately to highly
correlated with NO2.
Spearman r= 0.56-0.90 for
BC.
Ward et al. (2002)
Birmingham, Sandwell, U.K.
n = 162, age 9 yr, 27% with asthma, 31% with atopy
Repeated measures. Daily symptom diaries for two 8-week
periods, collected weekly. Recruitment from schools. 61%
participation rate. Logistic regression adjusted for time trend,
temperature, school day.
NO2-central site
24-h avg
Multiple sites
Cough
Winter:
0.78(0.57, 1.09)
Summer:
1.14(1.01, 1.27)
No copollutant model.
PlVh.s associated with cough.
5-210
-------
Table 5-33 (Continued): Epidemiologic studies of respiratory symptoms in children in the general population.
Study
Population Examined and Methodological Details
NO2 Metrics Analyzed Lag Day
Odds Ratio (95% Cl)
Single-Pollutant Model3
Copollutant Examination
Rodriguez et al. (2007)
Perth, Australia
n = 263, ages 0-5 yr, 1 parent with asthma or other atopic
disease
Repeated measures. Daily symptom diary from birth to age
5 yr. Recruitment from birth cohort. >80% follow-up
participation until yr4 and 5. GEE adjusted for temperature,
humidity.
NO2-central site
24-h max
10-site average
24-h avg
0 Wheeze (unit NR):
1.00(0.99, 1.01)
Cough: 1.01 (1.00, 1.02)
Wheeze: 1.01 (0.98, 1.04)
Cough: 1.03(1.00, 1.06)
No copollutant model.
Associations also found for
PM2.5, BS at lag 0.
tAndersen et al. (2008a)
Copenhagen, Denmark
n = 205, ages 0-3 yr, all with maternal asthma
Repeated measures. Daily symptom diaries from birth to
3 yr, collected every 6 mo. Recruitment from birth cohort.
95% follow-up participation. Mean 805 observations/subject.
GEE adjusted for age, sex, smoking exposure, paternal
asthma, temperature, calendar season.
NO2-central site
24-h avg
1 site within 15 km of
homes
N Ox-centra I site
24-h avg
Wheeze
Age 0-1 yr:
3.13(1.27, 7.77)
Age 2-3 yr:
1.71 (0.94, 3.10)
Age 0-1 yr:
3.26(1.14,9.26)
Age 2-3 yr:
1.80(0.87,3.72)
Forage 0-1 yr
WithUFP: 1.19(0.14, 75)
With PMio: 2.46 (0.72, 8.4)
UFP & PMio associations
attenuated with adjustment for
-NO2.
UFP & CO highly correlated
with NO2. Spearman r= 0.67,
0.75. Moderate correlation for
PMio. r=0.43.
tStern et al. (2013)
Bern, Base, Switzerland
n = 366, ages 0-1 yr
Repeated measures. Symptoms reported weekly by
telephone for 1 yr. Recruitment from birth cohort. 95%
follow-up participation. GAM adjusted for study week, sex,
siblings, nursery care, maternal atopy, birth weight, prenatal
& post-natal maternal smoking, parental education.
NO2-central site
1-week avg
2 site, urban and rural
Daytime respiratory
symptom composite:
1.20(1.04, 1.39)
No copollutant model.
PMio lag 7 associated with
respiratory symptoms.
Correlation NR.
5-211
-------
Table 5-33 (Continued): Epidemiologic studies of respiratory symptoms in children in the general population.
Study
Population Examined and Methodological Details NO2 Metrics Analyzed
tPeel et al. (201 1 ) NO2-central site
Atlanta, GA area 1-h max
n = 4,277, mean age 46 days, 84% premature births 1 sjte
Repeated measures. Followed for mean of 42 days.
1 1 1 ,000 person-days. Recruitment from referral center for
home cardiorespiratory monitoring of infants. Limited
generalizability. Apnea events collected electronically. No
information on participation rate. GEE adjusted for long-term
trends, age.
Odds Ratio (95% Cl)
Lag Day Single-Pollutant Model3
0-1 avg Apnea:
1.02(0.96, 1.08)
Copollutant Examination
WithOs: 1.00(0.96, 1.05)
C>3 association robust to NO2
adjustment. Moderate
correlation with NO2.
Spearman r= 0.45. No
association with PMio, coarse
PM.
Note: Studies are organized by population examined, and more informative studies in terms of the exposure assessment method and potential confounding considered are presented
first.
avg = average; BC = British Columbia; BS = black smoke; Cl = confidence interval; CO = carbon monoxide; GA = Georgia; GAM = generalized additive model; GEE = generalized
estimating equations; GLMM = Generalized linear mixed model; KS = Kansas; MA = Massachusetts; max = maximum; MO = Missouri; NO2 = nitrogen dioxide; NOX = sum of NO and
NO2; NR = not reported; NY = New York; O3 = ozone; OH = Ohio; OR = odds ratio; PM = particulate matter; PM25 = particulate matter with a nominal mean aerodynamic diameter
less than or equal to 2.5 |jm; PMio = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; ppb = parts per billion; SO2 = sulfur dioxide;
TN = Tennessee; TX = Texas; UFP = ultrafine particles; U.K. = United Kindgom; Wl = Wisconsin.
aEffect estimates are standardized to 20 ppb for 24-h avg NO2, 25 ppb for 8-h max, a 30-ppb increase for 1-h max NO2, and 40-ppb increase in 24-h avg NOX.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-212
-------
Controlled Human Exposure Studies
Controlled human exposure studies of respiratory symptoms in healthy populations do
not strongly inform whether the epidemiologic findings for NO2-related increases in
respiratory symptoms in children in the general population plausibly could reflect an
independent effect of NO2 exposure. The controlled human exposure studies were
reviewed in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). and most with NO2
exposures of 300-600 ppb for 1-3 hours (Table 5-31) did not observe changes in
symptom score 24 hours later (Gong et al.. 2005; Hazucha et al.. 1994; Morrow etal..
1992; Adams etal.. 1987). NO2 exposures of 400 or 600 ppb also did not affect
respiratory symptoms with simultaneous or sequential Os or PM2 5 CAPs co-exposures
[(Gong et al.. 2005; Hazucha et al.. 1994; Adams et al.. 1987): Table 5-311. In exception,
Frampton etal. (1991) reported statistically nonsignificant increases in symptom score
following 600 ppb NO2 exposure. NO2 exposures of 1,000 to 4,000 largely had no effect
on respiratory symptoms in healthy adults either (Torres etal.. 1995: Rasmussen et al..
1992: Linn et al.. 1985b: Hackney et al.. 1978).
5.2.7.4 Subclinical Respiratory Effects in Healthy Populations: Pulmonary
Inflammation, Injury, and Oxidative Stress
Pulmonary inflammation, injury, and oxidative stress are mediators of respiratory
symptoms and decreases in lung function (Section 4.3.5). Consistent with the evidence
described in the preceding sections, epidemiologic studies show ambient NO2-related
increases in pulmonary inflammation and oxidative stress in children and adults in the
general population. The few analyses of copollutant models indicate associations forNO2
persist with adjustment for another traffic-related pollutant. Supporting an independent
effect of NO2 in healthy populations, experimental studies report evidence for pulmonary
inflammation as PMN increases. Also, limited evidence from experimental studies
indicates development of an allergic phenotype with repeated NO2 exposures. Effects on
other indicators of inflammation and oxidative stress were observed more consistently at
higher than ambient-relevant NO2 concentrations.
Epidemiologic Studies
Together, the few epidemiologic studies from the 2008 ISA for Oxides of Nitrogen (U.S.
EPA. 2008c) and most recent studies found associations between increases in ambient
oxides of nitrogen and increases in pulmonary inflammation or oxidative stress among
children and adults in the general population and healthy populations. Locations, time
5-213
-------
periods, and ambient concentrations of oxides of nitrogen for these studies are presented
in Table 5-34. In this group of studies are several with exposure assessment methods that
aim to account for the high variability in ambient oxides of nitrogen.
Table 5-34 Mean and upper percentile concentrations of oxides of nitrogen in
epidemiologic studies of pulmonary inflammation and oxidative
stress in the general population.
Study3
Exposure
Metric
Location Study Period Analyzed
Mean Concentration
ppb
Upper Percentile
Concentrations (ppb)
5-day avg N62 Schools <14.0: 10.1 Across schools:
etal. (2010)
tLinetal.
(2011)
tLiu et al.
(2014a)
tBerhane et al.
(2011)
Ferrand, France
Beijing, China
Munich, Wesel,
Germany
13 southern
California
communities
Jun 2007
Sep 2007
Dec 2007
Jun 2008
Sep 2008
NR
Sep-Jun
2004-2005
Schools >14.0: 17.4
24-h avg NO2 24.3
30.4
45.3
26.6
25.9
24-h avg NO2 15.9C
24-h avg NO2 NR
75th: 14.0b, Max: 19.7b
NR
NR
NR
NR
NR
95th: 29.7b
NR
tPatel et al. New York, NY May-Jun 2005
(2013)
24-h avg NO2 Median: 23.3
NR
tAltuq et al. Eskisehir,
(2014) Turkey
Feb-Mar2007 24-h avg NO2 Suburban: 9.4b Max: 13.1b
Urban: 13.0b Max: 17.7b
Urban-traffic: 21.2b Max: 28.2b
Holquin etal. Ciudad Juarez, 2001-2002
(2007) Mexico
1-week avg NO2 18.2
NR
Steerenberq et Utrecht Feb-Mar1998
al-(2001) Bilthoven, the
Netherlands
24-h avg NO2 28.2b
24-h avg NO 30.2b
24-h avg NO2 25.5b
24-h avg NO 7.4b
Max: 44.7b
Max: 168b
Max: 49.5b
Max: 85.6b
tChen et al. New Taipei City, Oct-Jun 2007;
(2012a) Taiwan Jun-Nov2009
24-h avg NO2 21.7
NR
5-214
-------
Table 5-34 (Continued): Mean and upper percentile concentrations of oxides of
nitrogen in epidemiologic studies of pulmonary
inflammation and oxidative stress in the general
population.
Study3
tSalam et al.
(2012)
Steerenberq et
al. (2003)
fSteenhof et
al. (2013)
tStrak et al.
(2012)
Adamkiewicz
et al. (2004)
fWeichenthal
etal. (2011)
fChimenti et
al. (2009)
fMadsen et al.
(2008)
Location
13 southern
California
communities
the
Netherlands,
city NR
the
Netherlands,
city NR
Steubenville,
OH
Ottawa, ON,
Canada
Palermo, Sicily,
Italy
Oslo, Norway
Study Period
2004-2007,
school year
May-Jun; year
not reported
Mar-Oct 2009
Sep-Dec2000
NR
Nov
Feb
Jul; year NR
Jan-Jun 2000
Exposure
Metric
Analyzed
24-h avg NO2
24-h avg NO2
24-h avg NO
5-h avg NO2
5-h avg NOx
1-h avg NO2
24-h avg NO2
1-h avg NO
24-h avg NO
1-h avg NO2
7-day avg NO2
24-h avg NO2
7-day avg NO2
Mean Concentration
ppb
19.0
17.3b
6.3b
36
20
9.2
10.9
15
11.2
High traffic: 4.8
Low traffic: 4.6
31. 7b
27. 1b
33.9b
NR
NR
Upper Percentile
Concentrations (ppb)
Max: 39.4
Max: 28.3b
Max: 34.5b
Max: 96
Max: 34
75th: 12.8, Max:
75th: 14.6, Max:
75th: 16.1, Max:
75th: 14.2, Max:
Max: 1 1
Max: 10
NR
NR
NR
NR
NR
32.9
23.8
215
70.7
avg = average; Dec = December; Feb = February; h = hour; Max = maximum; NO = nitric oxide; NO2 = nitrogen dioxide; NOX = sum of
NO and NO2; NR = not reported.
aStudies presented in order of first appearance in the text of this section.
bConcentrations converted from |jg/m3 to ppb using the conversion factor of 0.532 for NO2 and 0.815 for NO assuming standard
temperature (25°C) and pressure (1 atm).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
Children in the General Population
Ambient NCh was associated with pulmonary inflammation and oxidative stress in
populations of children, in which the prevalence of asthma ranged from 7.5 to 59% and
prevalence of allergy ranged from 20 to 56% (Patel etal.. 2013; Berhane etal.. 2011; Lin
etal.. 2011; Steerenberg et al.. 2001). Except for Altugetal. (2014). studies
demonstrated associations in groups without asthma or allergy, with no consistent
difference in magnitude of association between children with and without respiratory
5-215
-------
disease [(Liu et al.. 2014a; Berhane et al.. 2011; Lin et al.. 2011); Figure 5-14 and
Table 5-35]. These findings suggest associations between NO2 exposure and pulmonary
inflammation in healthy children.
Among children, associations for NC>2 varied among the various indicators of oxidative
stress and inflammation. As examined in one study each, associations were not observed
with PMNs, eosinophils, exhaled breath condensate pH, or methylation of inducible nitric
oxide synthase (iNOS) (Pateletal.. 2013: Chenetal.. 2012a: Salam et al.. 2012:
Steerenberg et al.. 2001). But several study results pointed to associations with eNO
(Berhane etal.. 2011: Linet al.. 2011: Steerenberg etal.. 2001) (Figure 5-14 and
Table 5-35). Most of these studies assigned exposure from one central site per
community located between 1 and 14 km of subjects' homes. Further, associations were
found with CO, BC, BS, and PM2 5, and copollutant models were not analyzed. Moderate
to strong correlations were reported for NC>2 with PM2 5 and BC (r = 0.47-0.80) (Patel et
al., 2013: Berhane et al., 2011). Thus, the extent to which the results for central site NO2
reflect an independent association with NO2 is uncertain.
Other studies examined confounding by traffic-related copollutants and/or ambient NO2
measurements spatially aligned with a location of subjects, which may better represent
ambient microenvironmental exposure. Outdoor school NO2, averaged over 5 or 7 days,
was not associated with pulmonary inflammation in children without respiratory disease
(Altug etal.. 2014: Flamant-Hulin et al.. 2010: Holguin et al.. 2007). However, Holguin
et al. (2007) did not report quantitative results to assess whether there was suggestion of
association. And, the cross-sectional comparison of high versus low NO2 in Flamant-
Hulin etal. (2010) lacks the sensitivity to discern incremental changes in eNO that may
occur with incremental changes in NO2 exposure. Also, for some subjects, eNO was
measured days before NO2 was measured.
The study with the strongest inference about NO2-related increases in pulmonary
inflammation in healthy children was conducted in Beijing, China before and after the
2008 Olympics (Lin etal.. 2011). Although results were based on 28 children without
asthma, a large number of measurements was collected per child. NO2 and copollutants
were measured at a site 0.65 km from schools, improving the spatial alignment of
pollutants with subjects over the aforementioned central site studies. A 20-ppb increase in
lag 0 day of 24-h avg NO2 was associated with a 22% (95% CI: 18, 26) increase in eNO.
This effect estimate was attenuated two to fourfold with adjustment for BC or PM2 5 but
remained positive (e.g., 5.6% [95% CI: 0.38, 11] with adjustment for BC). Adjustment
for NO2 attenuated the association of eNO with PM2 5 but not BC. Thus, the NO2
association was partly confounded, by BC especially. However, the results also indicate
associations for NO2 that are independent of PM2 5 or BC in this population of children
5-216
-------
without asthma. Supporting inferences from the copollutant models, NO2, PM2 5, and BC
measured near children's school may have comparable exposure measurement error.
Although Linetal. (2011) found that eNO increased in relation to ambient NO2 measured
near subjects' school and independently of PM25 or the traffic-related BC, other studies
had weaker inference. Outdoor school NO2 was not associated with eNO in children
without respiratory disease, but these studies either did not report quantitative results or
had other methodological limitations. eNO was consistently associated with NO2
measured at central sites and also with PM2 5 and the traffic-related copollutants CO, BC,
and BS. Thus, there is uncertainty in the epidemiologic evidence as a whole regarding an
independent association of NO2 exposure with pulmonary inflammation and/or oxidative
stress in healthy children.
Study
Children
Linetal. (2011)
NO2 Metrics
Analyzed
24-h avg
lag 0 day
Exposure Assessment Subgroup
Central site
0.65 km from school
Berhane et al. (2011) 24-h avg Central sites
lag 1 -6 day avg
Steerenbergetal.
(2001)
Adults
Straketal. (2012)
Weichenthal et al.
(2011)
Adamkiewicz et al.
(2004)
24-h avg Central site
lag 0-6 day avg 1.9 km from schools
All subjects
No asthma
No asthma
No respiratory allergy
Urban
Suburban
5-h avg
lag 0 h
1 -h avg
lag 1 h
24-h avg
lag 0 day
On location of outdoor
exposure
Central site
Central site
-10 -5 0 5 10 15 20 25 30 35 40 45
Percent change in eNO per increase in NO2 (95% Cl)a
Note: avg = average; Cl = confidence interval; eNO = exhaled nitric oxide; h = hour; km = kilometer; NO2 = nitrogen dioxide.
Black = studies from the 2008 Integrated Science Assessment for Oxides of Nitrogen, red = recent studies. Results are presented
first for children then adults. Within each of these groups, results from more informative studies in terms of the exposure
assessment method and potential confounding considered are presented first. Study details and quantitative results reported in
Table 5-35.
aEffect estimates are standardized to a 20-ppb increase for 24-h avg NO2. Effect estimates for 5-h or 1-h avg NO2 are not
standardized but are presented as reported in their respective studies (Section 5.1.2.2).
Figure 5-14 Associations between ambient nitrogen dioxide concentrations
and exhaled nitric oxide among children and adults in the general
population.
5-217
-------
Table 5-35 Epidemiologic studies of pulmonary inflammation, injury, and oxidative stress in children and adults
in the general population.
Study NO2 Metrics
Population Examined and Methodological Details Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
Children in the general population: studies with small spatial scale exposure assessment and/or examination of copollutant confounding
tZhu(2013): Linetal. (2011)
Beijing, China
n = 36, ages 9-12 yr, 8 with asthma, 28 without asthma
Repeated measures before and after Olympics. Examined
daily for five 2-week periods. 1,581 observations.
Recruitment from school. 60% responded to initial survey,
95% follow-up participation. GEE adjusted for temperature,
relative humidity, body mass index.
NO2-central site
24-h avg
Site 0.65 km from
schools
eNO:
All subjects: 22% (18, 26)
No asthma: 22% (18, 26)
No asthma: 9.5% (5.8, 13)
With BC: 5.6% (0.38, 11)
With PlVhs: 14% (9.5, 19)
BC robust to adjustment for
NO2, PM2.5 reduced but
positive.
Moderate correlations with
NO2. Spearman r = 0.30 for
PM2.5, 0.68 for BC.
tFlamant-Hulin et al. (2010)
Clermont-Ferrand, France
n = 70 without asthma, mean age: 10.7 (SD: 0.7) yr, 75% no
atopy
Cross-sectional. Recruitment from schools. 69%
participation. Self or parental report of no asthma. For some
subjects, eNO measured up to 1 week before pollutants.
GEE adjusted for atopy, mother's birth region, parental
education, family history of allergy, smoking exposure. Did
not consider confounding by meteorology.
NO2-school outdoor
24-h avg
0-4 avg log eNO comparing >14.3 vs.
<14.3ppbNO2
-0.09 (-0.22, -0.04)
NO2-school indoor
24-h avg
0-4 avg log eNO comparing >16.3 vs.
<16.3ppb NO2
-0.16 (-0.11, -0.20)
No copollutant model.
Acetylaldehyde and PlVh.s
associated with eNO.
1 Correlations with NO2 not
reported.
5-218
-------
Table 5-35 (Continued): Epidemiologic studies of pulmonary inflammation, injury, and oxidative stress in children
and adults in the general population.
Study NO2 Metrics
Population Examined and Methodological Details Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
Children in the general population: studies with central site exposure assessment and no examination of copollutant confounding
tChenetal. (2012a)
New Taipei City, Taiwan
n = 100, mean age 10.6 (SD: 2.5) yr, 33% asthma, 33%
atopy
Repeated measures. Examined 3-4 times/mo for 10 mo.
824 observations. Recruited from schools. A priori
recruitment of children with and without asthma or atopy.
Participants similar to nonparticipants. Mixed effects model
adjusted for school, age, sex, body mass index, upper
respiratory infection, asthma/allergic rhinitis attack, asthma
medication use, temperature, humidity, day of week,
sampling time, parental education, smoking exposure at
home.
NO2-central site
24-h avg
1 site 2.5 km from
schools, most
homes 1 km of
schools
0 No quantitative data. NO2
1 reported not to affect
2 eosinophils, PMNs,
monocytes, IL-8.
O
No copollutant model.
Associations found for PlVhs,
Os but not CO.
Moderate to no correlation with
NO2. Pearson r= 0.61 for
PM2.5, -0.01 forOs.
Steerenberq et al. (2001)
Utrecht (Urban, near busy roadway) and Bilthoven
(Suburban), the Netherlands
n = 126, ages 8-13 yr, 28% respiratory disease,
20% allergy
Repeated measures. Examined 1/weekfor 7-8 weeks.
Recruitment from urban and suburban schools. 65%
participation. Nonstandardized eNO collection. Mixed
effects model adjusted for sex, age, # cigarettes smoked in
home, presence of a cold, history of respiratory symptoms,
allergy. No consideration for potential confounding by
meteorological factors.
NO2-central site 0-6 avg eNO:
Urban: 35% (0, 70)b
Suburban: 3.0%, p > 0.05
IL-8 (units NR)
Urban OR: 1.08, p> 0.05
Suburban OR: 1.03, p > 0.05
No copollutant model.
PM-io and BS also associated
with eNO, IL-8, uric acid, urea.
NO-central site
All 24-h avg
Site within 1.9 km of
schools
0-6 avg eNO:
Urban: 6.6% (0, 13)b
Suburban: 7.3% (0, 15)b
IL-8 (units NR)
Urban OR: 1.05, p> 0.05
Suburban OR: 0.95, p > 0.05
5-219
-------
Table 5-35 (Continued): Epidemiologic studies of pulmonary inflammation, injury, and oxidative stress in children
and adults in the general population.
Study NO2 Metrics
Population Examined and Methodological Details Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
tPateletal. (2013)
New York City, NY
n = 36, ages 14-19 yr, 94% nonwhite, 50% with asthma
Repeated measures. EEC collected 2/week for 4 weeks.
217 observations. Recruitment from schools. 89-90%
participation rate. A priori recruitment of children with and
without asthma or atopy. Self-report of physician-diagnosed
asthma and symptoms in previous 12 mo. Mixed effects
model with random effects for subject and adjusted for
school, daily average temperature, 8-h max Os. Adjustment
for day of week and humidity did not alter results.
NO2-central site
24-h avg
Site 14 km from
schools
EEC 8-isoprostane:
0 1.7(0.63, 2.7) log units
0-3 avg 3.1 (1.3, 4.9) log units
EEC pH:
0 -0.05 (-0.79, 0.68)
0-3 avg -0.11 (-1.2, 1.0)
No copollutant model.
EC also associated with EEC
pH and 8-isoprostane.
• School EC moderately to highly
correlated with NO2. Pearson
r=0.62, 0.80.
tBerhaneetal. (2011)
Anaheim, Glendora, Long Beach, Mira Loma, Riverside,
San Dimas, Santa Barbara, Upland, CA, Children's Health
Study
n = 169, ages 6-9 yr
Cross-sectional. Recruitment from schools. Parental report
of physician-diagnosed asthma and history of respiratory
allergy. Two different methods used foreNO measurement.
No information on participation rate. Linear regression
adjusted for community, age, sex, race/ethnicity, asthma,
asthma medication use, history of respiratory allergy, eNO
collection time, body mass index percentile, smoking
exposure, parental education, questionnaire language,
season, multiple temperature metrics, eNO collected
outdoors.
NO2-central site
24-h avg
Sites in each
community. # sites
in each community
NR
1-6 avg eNO
No asthma:
11% (-3.2, 28)
No respiratory allergy:
9.1% (-3.6, 23)
No copollutant model.
PM2.5, PM-io, Os associated
with eNO.
Moderate or weak correlations
with NO2.
Pearson r for warm and cold
season = 0.47, 0.65 for PM2.s;
0.49, 0.55 for PMio; 0.15, -0.4
for Os.
tSalametal. (2012)
Same cohort as above
n = 940, ages 6-11 yr, 14% asthma, 56% respiratory allergy
Cross-sectional. Recruitment from schools. Subjects
representative of full cohort. Linear regression model
adjusted for age, sex, ethnicity, asthma, respiratory allergy,
parental education, smoking exposure, community, month
of eNO collection. No consideration for confounding by
meteorology.
NO2-central site
24-h avg
Sites in each
community. # sites
in each community
NR
1-7 avg iNOS promoter methylation:
0.40% (-1.0, 1.8)
iNOS methylation not strong
predictor of eNO.
No copollutant model.
PM2.5 associated with higher
iNOS promoter methylation.
Moderate correlation with NO2.
Spearman r = 0.36.
5-220
-------
Table 5-35 (Continued): Epidemiologic studies of pulmonary inflammation, injury, and oxidative stress in children
and adults in the general population.
Study NO2 Metrics
Population Examined and Methodological Details Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
Adults in the general population: studies with small spatial scale exposure assessment and/or examination of copollutant confounding
tStrak (2013): Strak et al. (2012)
tSteenhofetal. (2013)
Utrecht area, the Netherlands
n = 31, adults ages 19-26 yr, all healthy, nonsmoking
Repeated measures. Examined 3-7 times.
107 observations. Recruitment from university. Well-defined
outdoor exposures at various sites: underground train
station, two traffic sites, farm, and urban background site.
Heart rate maintained during intermittent exercise. Multiple
comparisons could results in higher probability of
associations found by chance alone. No information on
participation rate. Mixed effects model adjusted for
temperature, relative humidity, season, high/low pollen,
respiratory infection.
NO2 and NOx-on
site of outdoor
activity
5-h avg
0-h eNO:
post- NO2:6.9%(-1.9, 16)
exposure NOX: 4.7% (-1.8, 11)
Per 10.54 ppb increase in NO2
and 28.05 ppb increase in NOx
2-h NO2
IL-6:66%(-10, 142)
NAL protein:
60% (0, 121)b
With PNC:
-7.4% (-19, 3.9)forNO2
-5.8% (-14, 2.4) for NOx
With EC:
4.1% (-6.0, 14)forNO2
2.0% (-7.3, 11) for NOx
PNC association persists
NO2/NOx adjustment. EC &
Abs attenuated.
ForlL-6:
With PNC: 95% (0, 190)
With OC: 67% (-10, 144)
Copollutant results robust.
Moderate to high correlations
with NO2 & NOx. Spearman
r=0.56, 0.75 for PNC, 0.74,
0.87 for Abs, 0.67, 0.87 for EC.
5-221
-------
Table 5-35 (Continued): Epidemiologic studies of pulmonary inflammation, injury, and oxidative stress in children
and adults in the general population.
Study NO2 Metrics
Population Examined and Methodological Details Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
Adults in the general population: studies with central site exposure assessment and no examination of copollutant confounding
tChimenti et al. (2009)
Palermo, Sicily, Italy
n = 9, male adults mean age 40 (SD: 3.8) yr, all healthy,
nonsmoking
Repeated measures. Examined during 3 outdoor races. No
information on participation rate. Statistical analyses limited
to correlation analyses. No consideration for potential
confounding factors or repeated measures.
NO2-central site
Averaging time NR
10 sites
NR No correlations with plasma
PMN or eosinophils.
No results reported for CC16.
No copollutant model.
Associations found with Os and
PM2.5.
tWeichenthal et al. (2011)
Ottawa, Canada
n = 42, adults ages 19-58 yr, from nonsmoking homes,
95% white, 62% with allergies, 33% with asthma
Repeated measures. Most examined 3 times.
118 observations. 1-h outdoor exposures during cycling in
low and high traffic areas. Recruitment from public
advertisements. No information on participation rate. Mixed
effects models with random subject effect adjusted for
temperature during cycling, average heart rate. Adjustment
for relative humidity, day of week did not affect results.
NO2-central site eNO:
1-h avg 1-h -0.01% (-0.08,0.06)
1 site 4-h -0.04% (-0.09, 0.01)
post- Per 4-ppb increase in NO2
exposure
No copollutant model.
PM2.5 associated with eNO.
Moderate correlation with NO2.
Spearman r = 0.31 for low
traffic site, 0.45 for high traffic
site.
Potential differential exposure
error for personal PM species
and VOCs vs. central site NO2.
tMadsen et al. (2008)
Oslo, Norway
n = 1,004, male adults ages 67-77 yr, 10% with respiratory
disease
Cross-sectional. Recruitment from a larger cohort to
represent a range of home outdoor NO2. No information on
participation rate. GLM adjusted for age, respiratory
disease, alcohol consumption, smoking status, #
cigarettes/day, smoking exposure, education, hour of exam,
body mass index, temperature.
NO2-central site
NO2-dispersion
model
No information on
model validation.
0-7 avg
CC16:
30% (7.8, 57)
3.8% (-7.3, 16)
No copollutant model.
PM2.5 (central site and home)
associated with CC16.
Moderate correlation with NO2.
Spearman rfor home = 0.59.
5-222
-------
Table 5-35 (Continued): Epidemiologic studies of pulmonary inflammation, injury, and oxidative stress in children
and adults in the general population.
Study NO2 Metrics
Population Examined and Methodological Details Analyzed
Effect Estimate (95% Cl)
Lag Day Single-Pollutant Model3
Copollutant Examination
Adamkiewicz et al. (2004) NCb-central site
Steubenville, OH 24-h avg
n = 29, adults ages 53-90 yr, nonsmoking, 28% with
asthma, 24% with COPD NO-central site
Repeated measures. Examined weekly for 12 weeks. 24-h avg
138-244 total observations. No information on participation -| sjte
rate. GLM with subject-specific intercept and adjusted for
time of day, day of week, study week, temperature,
pressure, relative humidity. Several NO2 measurements
missing.
eNO:
0 3.8% (-7.3, 16%)
0 30% (7.8, 57%)
No copollutant model for NO2.
NO with PlVh.s: 9.2% (-1.7, 20)
' PM2.5 result robust.
Correlations NR.
Ambient NO robust to
adjustment for indoor NO.
Note: Studies are organized by population examined, and more informative studies in terms of the exposure assessment method and potential confounding considered are presented
first.
Abs = absorbance coefficient; avg = average; BC = British Columbia; BS = black smoke; CC16 = club cell protein; Cl = confidence interval; CO = carbon monoxide; COPD = chronic
obstructive pulmonary disease; EEC = exhaled breath condensate; EC = elemental carbon; eNO = exhaled nitric oxide; GEE = generalized estimating equation; GLM = generalized
linear mixed effects model; IL = interleukin; iNOS = inducible nitric oxide synthase; m = meters; NAL = nasal lavage; NO = nitric oxide; NO2 = nitrogen dioxide, NOX = sum of NO and
NO2; NR = not reported; O3 = ozone; OC = organic carbon; OR = odds ratio; PM = particulate matter; PM25 = particulate matter with a nominal mean aerodynamic diameter less than
or equal to 2.5 |jm; PMio = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; PMN = polymorphonuclear leukocyte; PNC = particle number
concentration; SD = standard deviation; VOC = volatile organic compound; yr = years.
aEffect estimates are standardized to 20 ppb for 24-h avg N02 or NO and 25 ppb for 8-h max NO2. Effect estimates for other averaging times (1-h avg to 15-h avg) are not
standardized but presented as they are reported in their respective studies (Section 5.1.2.2).
b95% Cl estimated for p = 0.05 based on reported p-value < 0.05.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-223
-------
Adults in the General Population
Among a few studies of older adults reviewed in the 2008 ISA for Oxides of Nitrogen
(U.S. EPA. 2008c) and recent studies of older adults and adults performing outdoor
exercise, several results point to increases in pulmonary inflammation in association with
increases in ambient NO2 concentrations. Pulmonary inflammation was indicated as
increases in eNO, nasal lavage IL-6, and indicators of pulmonary injury and lung
permeability such as club cell protein (CC16) and nasal lavage protein levels
(Table 5-35). Associations adjusted for PM25 or a traffic-related copollutant were found
for 24-h avg NO (Adamkiewicz et al.. 2004) and with 5-h avg NO2 for some outcomes
(Steenhof et al.. 2013; Strak et al., 2012). The epidemiologic findings have some support
from controlled human exposure and toxicological studies (described in sections that
follow), although the evidence for pulmonary injury is inconsistent.
In populations of mostly healthy adults performing outdoor exercise for <1 to 5 hours,
increases in pulmonary inflammation were associated with NO2 measured at the locations
of outdoor exposures but not at central sites. Compared with studies that do not account
for time-activity patterns, examination of subjects during time spent outdoors may better
reflect effects related to ambient exposures, particularly when pollutants are measured in
subjects' outdoor locations. In these studies, subjects had 3-5 separate outdoor exposure
periods. In some studies, exposures occurred in locations that represented a gradient of
traffic volume. Among adults running or cycling outdoors for 35-90 minutes, eNO and
inflammatory cell counts (as measured by PMNs and eosinophils) were not associated
with NO2 measured at central sites (Weichenthal et al.. 2011; Chimenti et al.. 2009)
(Figure 5-14 and Table 5-35). However, increases in eNO and nasal lavage IL-6 and
protein were found in healthy adults in association with 5-h avg NOx and NO2 measured
on the site of outdoor exposures (Steenhof et al.. 2013; Strak etal.. 2012). which account
for variability in exposure better than central site measurements. Increases in eNO and
nasal lavage IL-6 and protein were found immediately after and 2 hours after exposures
ended but not the morning after, indicating a transient increase in pulmonary
inflammation. Multiple analyses were conducted across pollutants, including several
PM2 5 components, but the consistency in results does not support the likelihood that the
NO2 associations were found by chance alone (Strak etal.. 2012).
Among healthy adults, eNO also was associated with EC, absorbance coefficient (Abs),
and PNC (Strak etal.. 2012); IL-6 also was associated with PM2 5 and OC (Steenhof et
al.. 2013). In copollutant models, associations of eNO with NOx and NO2 were
attenuated with adjustment for EC or Abs and became negative with adjustment for PNC
(Strak etal.. 2012). The PNC effect estimate was robust to adjustment for NOx or NO2.
NOx and NO2 were highly correlated with PNC and EC (e.g., r = 0.75 for NOx and PNC
5-224
-------
and 0.71 forNCh and EC). However, NC>2 remained associated with nasal lavage IL-6
and protein after adjustment for PNC, PM2 5, EC, or OC (e.g., 67% [95% CI: -10, 144]
increase in IL-6 per 30-ppb increase in 5-h avg NC>2 and 95% [95% CI: 0, 190] with
adjustment for PNC). Thus, in this study of well-defined outdoor exposures, there is
evidence of confounding of NC^-eNO associations by PNC but associations of NO2 with
IL-6, nasal lavage protein, and lung function that are independent of many key
traffic-related copollutants (Figures 5-16 and 5-17).
Increases in pulmonary inflammation were associated with 24-h avg NO or NCh
measured at central monitoring sites among older adults (ages: 53-90 years) (Madsen et
al.. 2008; Adamkiewicz et al., 2004). Multiday averages of NO2 (e.g., lag 0-4 day avg,
0-7 day avg) were associated with CC16 (Madsen et al.. 2008). However, there is
uncertainty in older adults regarding independent associations for NC>2 as Madsen et al.
(2008) found an association with central site not home NC>2, and each study found
associations with PIVb 5 and traffic-related copollutants. Among older adults in
Steubenville, OH, the association between NO and eNO decreased with adjustment for
PM2 5 but remained positive (Adamkiewicz et al., 2004). The PIVb 5 effect estimate
increased. The copollutant model result has uncertain inference because of potential
differential exposure measurement error and unreported correlations for NO2 and PM2 5.
Controlled Human Exposure and lexicological Studies of Pulmonary
Inflammation
As reviewed in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). several
controlled human exposure and toxicological studies examined the effects of NO2
exposure on markers of pulmonary inflammation such as differential inflammatory cell
counts, eicosanoids, and cytokines. In these studies, the typical protocol consisted of a
single- or multi-day exposure to NO2 (100-5,000 ppb) followed 1 to 24 hours later by
collection of bronchial wash or BAL fluid (Tables 5-36 and 5-37). The consistency and
biological significance of effects across studies is difficult to evaluate given the variety of
exposure circumstances and timing when effects were measured, but there is evidence for
NO2-induced pulmonary inflammation in healthy adults that is most consistently
demonstrated by increases in PMNs.
In limited analysis, NO2 exposures (1-3 hours) of 300-600 ppb tended not to affect PMN
levels (Huang etal.. 2012b: Vagaggini et al.. 1996). (Frampton et al. (2002): Frampton et
al. (1989)) observed a statistically nonsignificant increase after a 3-hour exposure to
600 ppb. In most groups of healthy adults, PMNs did increase after exposure to NO2 in
the range of 1,000 to 3,500 ppb given for 20 minutes to 6 hours as a single exposure
(Frampton et al.. 2002; Devlin et al.. 1999; Azadniv et al.. 1998; Torres etal.. 1995;
5-225
-------
Helleday et al.. 1994) or given for 4 hours over 3 or 4 days (Solomon et al.. 2000;
Blomberg et al.. 1999). In exception, 2,000-ppb NO2 spikes superimposed on a
background exposure of 50 ppb NO2 did not affect PMNs (Frampton et al.. 1989).
Helleday et al. (1994) found that a 20-minute exposure to 3,500 ppb NCh increased
bronchial PMNs in nonsmoking adults and increased alveolar PMNs in smoking adults.
In contrast with findings in healthy humans, NC>2 exposures in the range of 300 to 5,000
ppb did not affect PMN levels in mice, rats, or rabbits whether exposure occurred on a
single day or was repeated over 2 to 13 days (Poynter et al.. 2006; Pagani et al.. 1994;
Schlesinger et al., 1990). PMNs did increase in one study of rabbits, but an independent
effect of 1,000 ppb NO2 could not be discerned because exposure co-occurred with
H2SO4 (Schlesinger. 1987a).
Experimental studies of healthy humans and animal models provide inconsistent evidence
for the effects of short-term NO2 exposure on eicosanoid levels. In humans, exposure to
1,000 or 2,000 ppb NCh for 3 or 4 hours increased thromboxane 62 levels but not an array
of other prostaglandins (Devlin et al.. 1999; Torres et al.. 1995). Increases in thromboxane
also were observed in rabbits after a 2-hour exposure to 1,000 ppb but not 3,000 ppb NO2
(Schlesinger et al.. 1990). Lower NO2 exposures of 100 ppb (4 hours) increased enzymes
that form eicosanoids with ex vivo exposure of rat AMs, but eicosanoid levels increased
in a statistically nonsignificant manner with ex vivo exposure and decreased with in vivo
exposure (Robison and Forman. 1993).
Both controlled human exposure and toxicological studies tended to show lack of effect
of short-term NC>2 exposure (300-5,000 ppb for 1 hour-7 days) on indicators of
pulmonary inflammation such as BAL fluid levels of IL-6 and IL-8 (Huang et al.. 2012b:
Pathmanathan et al.. 2003) as well as lymphocytes (Huang et al.. 2012b; Frampton et al..
2002: Vagaggini et al.. 1996: Torres etal.. 1995: Muller et al.. 1994: Pagani etal.. 1994).
There were a few controlled human exposure studies of healthy adults that showed
increased pulmonary inflammation as increased lymphocytes with 600 ppb NC>2 exposure
(Frampton et al., 2002) or increased IL-6 and IL-8 (Devlin et al.. 1999) or ICAM-1
(Pathmanathan et al.. 2003) with 2,000 ppb NC>2 exposure. In examination of
NC>2-copollutant co-exposures, there was no additive or synergistic effect with PM2 5
CAPs for lymphocytes or PMNs in humans (Huang et al.. 2012b). However, there is
evidence of synergistic effects with Os for some eicosanoids in rabbits (Schlesinger et al..
1990).
Controlled Human Exposure and Toxicological Studies of Pulmonary Injury
In contrast with pulmonary inflammation, controlled human exposure studies of healthy
adults generally did not show NCh-induced pulmonary injury as measured by BAL or
5-226
-------
bronchial wash fluid levels of protein, LDH, or albumin. Null effects were observed in
response to a wide range of NO2 exposures (600-3,500 ppb; Table 5-37) given for
20 minutes to 6 hours on a single day or repeated over 3 or 4 days (Frampton et al.. 2002;
Solomon et al.. 2000: Blomberg et al.. 1999: Devlin et al.. 1999: Azadniv et al.. 1998:
Helleday et al.. 1994). Pulmonary injury in healthy adults was indicated in a few studies
as an increase in LDH following a 2-hour 500-ppb NCh exposure (Huang et al.. 2012b)
and an increase in albumin following a 4-hour 2,000-ppb NC>2 exposure repeated over 4
days (Blomberg et al.. 1999). A wide range of NCh exposures (100-5,000 ppb for 1
hour-20 days; Table 5-38) showed no effect BAL protein, LDH, albumin, lipid, or
surfactant in rodents (Muller et al.. 1994: Pagani et al.. 1994: Robison and Forman. 1993:
RoseetaL 1989a: Last and Warren. 1987: Selgrade et al.. 1981). When BAL fluid
protein, LDH, or albumin increased in rodents, NC>2 exposures often were 5,000 ppb for a
few hours on a single day or on 2-5 days (Poynter et al., 2006: Miiller et al., 1994: Rose
etal.. 1989a: Last and Warren. 1987: Hatch etal.. 1986: Gregory et al.. 1983). Lower
NO2 exposures of 400-3,000 ppb increased indicators of pulmonary injury in rodents
deficient in dietary antioxidant vitamin intake (Selgrade etal.. 1981: Sherwin and
Carlson. 1973) or with longer duration exposures of 1-3 weeks (Gregory et al., 1983:
Elsaved and Mustafa. 1982: Sherwin et al.. 1972).
In rats, 5,000 ppb NC>2 exposure (not lower) over 1-7 days induced minor morphologic
changes in the respiratory tract indicative of mild pulmonary injury. Such changes
included thickened interstitium and inflammatory cell accumulation (Miiller et al.. 1994).
increased collagen synthesis, a feature of fibrosis (Last and Warren. 1987). slight
interstitial edema (Earth etal.. 1995). a few necroses of the bronchiolar epithelium
(bronchi were normal), and increased proliferative index in bronchioles and bronchi
(Earth and Miiller. 1999). As examined by (Earth and Miiller (1999): Earth etal. (1995)).
neither edema nor club cell proliferative index were increased after a 25-day exposure to
5,000 ppb NO2.
Controlled Human Exposure and lexicological Studies of Oxidative Stress
and Antioxidant Status
Although toxicological studies demonstrate NCh-induced pulmonary oxidative stress and
antioxidant capacity at higher than ambient-relevant concentrations of NC>2, effects in
toxicological and controlled human exposure studies of healthy rodent models and
humans are more variable with ambient-relevant exposures. There is heterogeneity
observed across the array of indicators of pulmonary oxidative stress and antioxidant
capacity examined: lipid peroxidation, antioxidant enzymes, uric acid, and glutathione or
glutathione-related enzymes (Tables 5-37 and 5-38). Indicators of lipid peroxidation
increased in BAL fluid of healthy adults exposed to 4,000 ppb NO2 for 3 hours
5-227
-------
(Mohsenin. 1991) but were not altered with 2,000 ppb NO2 (4 hours) (Kelly et al.,
1996a). Inconsistency also was observed in rats, with 400 ppb NCh (2 weeks) inducing no
effect in rats or guinea pigs (Ichinose and Sagai. 1989) but 3,000 ppb NO2 (1 week)
increasing lipid peroxidation in rats deficient in antioxidant Vitamin E (neutralizes
reactive oxygen species) (Sevanian et al., 1982b).
Effects on antioxidants were similarly variable. Blomberg et al. (1999) found no changes
in glutathione, ascorbic acid, or uric acid levels in humans 1.5 hours after 4 consecutive
days of exposure to 2,000 ppb NO2 for 4 hours. A study of the kinetics of antioxidant
response in the respiratory tract indicates that NC>2 exposure may have a transient effect.
A single exposure to 2,000 ppb NO2 reduced levels of uric acid and ascorbic acid in
bronchial wash and BAL fluid 1.5 hours post-exposure (Kelly etal.. 1996a). Six and
24 hours after exposure, the levels of these antioxidants returned to baseline or increased.
Glutathione increased in the bronchial wash 1.5 and 6 hours after exposure, but no
changes in glutathione were found in the BAL fluid or for reduced glutathione at any
time after exposure. In rodents, 400 ppb NC>2 exposure for 2 weeks did not affect
Vitamin C or Vitamin E levels (neutralize reactive oxygen species) in lung homogenates
(Ichinose and Sagai. 1989). but a 5,000 ppb NO2 exposure for 24 hours increased
oxidized glutathione in BAL fluid (Pagani etal.. 1994). Total glutathione was slightly
diminished in BAL fluid but was increased in the peripheral blood. Because of the
heterogeneity across studies in NO2 exposure and antioxidants examined as well as the
time of antioxidant measurement, it is not clear whether the variable results indicate
inconsistent effects on oxidative stress, an increase in antioxidants in response to the
increased presence of reactive oxidant species induced by NO2 exposure, and/or oxidative
stress due to depleted antioxidant capacity.
The lack of clarity as to whether NO2 exposure stimulates antioxidants in response to
increased oxidant species and/or results in oxidative stress also applies to results for
changes in antioxidant enzymes. Enzyme levels or activities were examined in rodents
with exposures of 1 to 4 weeks. Exposures of rodents to 400 or 500 ppb NC>2 did not alter
activity of glutathione peroxidase (GPx), glutathione S-transferase (GST), superoxide
dismutase (SOD), glutathione reductase, glucose-6-phosphate dehydrogenase, or
6-phosphogluconate dehydrogenase (Ichinose etal.. 1988; Ayaz and Csallany. 1978).
Higher exposures of 1,000 to 5,000 ppb NC>2 had effects in varying directions.
Continuous 4-week exposure of rats to 1,000 ppb NO2 led to decreased GPx (Ayaz and
Csallany. 1978). Four-week exposure to 1,000 and 5,000 ppb NCh for 6 hours a day
increased GPx, GST, and SOD in BAL fluid of rats (de Burbure etal.. 2007). With 1,000
ppb NO2, SOD levels returned to control levels by 48 hours post-exposure. Exposure of
mice to 4,800 ppb NO2 for 8 hours a day for 8 days induced no change in GST or
glutathione reductase, but a statistically nonsignificant decrease in GPx was observed
5-228
-------
(Mustafa et al.. 1984). For both 400 ppb NO2 (Ichinose and Sagai. 1989) and 4,800 ppb
NO2 (Mustafa et al.. 1984). combined exposure with Os did increase GPx synergistically,
indicating responses may vary depending on the level of oxidants produced.
Another explanation for the variable effects of NO2 exposure on indicators of pulmonary
oxidative stress, antioxidant capacity, and pulmonary injury may be variation among
subjects in antioxidant status. Discussed in detail in Section 7.6.1. Vitamin C and/or E
supplementation attenuated the effects of NO2 exposure (1,000 or 4,000 ppb) on reduced
antioxidant enzyme activity in rodents and increased lipid peroxidation (as well as airway
responsiveness, Section 5.2.2.1) in healthy humans. As additional support, NC>2 exposure
(1,000-4,800 ppb) increased indicators of pulmonary injury and oxidative stress in
rodents deficient in dietary Vitamin C or E.
Controlled Human Exposure and lexicological Studies of Development of a
Pro-Allergic Phenotype
A few experimental studies indicate that repeated exposures to NO2 may promote Th2
skewing, which may have implications for allergic sensitization and development of
Th2-related conditions such as asthma (Section 4.3.2.6). In guinea pigs, 2-week exposure
to 3,000 ppb NCh led to an increase in eosinophils in the nasal epithelium and submucosa
lYOhashi et al.. 1994): Table 5-371. In healthy adults, 2,000-ppb NO2 exposure for
6 hours on 4 consecutive days increased expression of the Th2 cytokines IL-5, IL-10, and
IL-13 in the bronchial epithelium (Pathmanathan et al.. 2003). IL-5 promotes
eosinophilia, and IL-13 promotes airway hyperresponsiveness and mucus production.
5-229
-------
Table 5-36 Characteristics of controlled human exposure studies of pulmonary
inflammation, injury, and oxidative stress in healthy adults.
Study
Disease Status3;
Sample Size;
Sex; Age (Mean ± SD)
Exposure Details
Endpoints Examined
Azadnivet al.
(1998)
n = 11 M, 4F;
Early phase: 28.1 ± 3.Syr
Late phase: 27.4 ± 4.2 yr
2,000 ppb for 6 h;
Exercise for approximately 10 of
every 30 min at VE = 40 L/min
BAL fluid analysis 1 h and 18 h
after exposure. Protein
concentration, differential cell
counts.
Blomberq et al.
(1999)
n = 8 M, 4 F; 26 yr
(range: 21-32)
2,000 ppb, 4 h/day for 4 days;
Exercise 15 min on/15 min off at
workload of 75 watts
Cell counts from bronchial biopsies,
BW, and BAL fluid 1.Shatter
exposure; protein concentration,
IL-8, MPO, hyaluronic acid,
glutathione, ascorbic acid, and uric
acid in BAL fluid and BW1.5h
after exposure, blood parameters.
Devlin et al.
(1999)
n = 10 M;
range: 18-35 yr
2,000 ppb for 4 h;
Exercise for 15 min on/15 min off
at VE = 50 L/min
Bronchial and alveolar lavage fluid
contents 16 h after exposure. LDH
activity, tissue plasminogen factor
activity, IL-6 activity, IL-8 activity,
PGE2 levels, total protein,
ascorbate, urate, and glutathione.
Frampton et al. (1) n = 7 M, 2 F;
(1989) 30 yr (range: 24-37)
(2)n = 11 M, 4F;
25 yr (range: 19-37)
(1) 600 ppb for 3 h,
(2) 50 ppb for 3 h + 2,000 ppb
peak for 15 min/h;
(1,2) Exercise 10 min on/20 min
off at VE = ~4 times resting
BAL fluid cell viability and
differential counts 3.5 h after
exposure, IL-1 activity in BAL fluid
cells.
Frampton et al. (1,2) n = 12 M, 9 F;
(2002) F = 27.1 ± 4.1 yr
M = 26.9 ± 4.5 yr
(1) 600 ppb for 3 h,
(2) 1,500 ppb for 3 h;
(1,2) Exercise 10 min on/20 min
off at VE = 40 L/min
Bronchial and alveolar lavage fluid
cell viability and differential counts
3.5 h after exposure, peripheral
blood characterization.
Helleday et al.
(1994)
n = 8 nonsmokers; sex
NR; median: 26 yr
(range: 24-35)
n = 8 smokers; sex NR;
median: 29 yr
(range: 28-32)
3,500 ppb for 20 min;
Exercise last 15 min at 75 watts
BW and BAL fluid analysis. Protein
concentration, differential cell
counts.
tHuanq et al.
(2012b)
(1)n = 11 M, 3F
(2)n=6M, 7F
(3) n = 7 M, 6 F;
24.6 ± 4.3 yr
(1)500ppbNO2for2h,
(2) 500 ppb NO2
+ 73.4 ± 9.9 ug/m3 CAPs for 2 h,
(3) 89.5 ± 10.7 ug/m3 CAPs for
2h;
(1-3) Exercise 15 min on/15 min
off at VE = 25 L/min
Cell counts and concentrations of
IL-6, IL-8, a1-antitrypsin, and LDH
in BAL fluid 18 h after exposure.
5-230
-------
Table 5-36 (Continued): Characteristics of controlled human exposure studies of
pulmonary inflammation, injury, and oxidative stress in
healthy adults.
Study
Disease Status3;
Sample Size;
Sex; Age (Mean ± SD)
Exposure Details
Endpoints Examined
Jorres et al.
(1995)
Healthy; n = 5 M, 3 F;
27 yr (range: 21-33)
Asthma; n = 8 M, 4 F;
27 ± 5 yr
1,000 ppbforS h;
Exercise 10 min on/10 min off at
individual's maximum workload
BAL fluid analysis 1 h after
exposure (cell counts, histamine,
eicosanoids).
Kelly et al.
(1996a)
n = 44; sex NR;
median: 24 yr
(range: 19-45)
2,000 ppbfor4 h;
Exercise 15 min on/15 min off at
75 watts
Antioxidant concentrations and
malondialdehyde in BAL fluid and
BWat 1.5, 6, or 24 h after
exposure.
Mohsenin
(1991)
n = 10 M, 9F;
range: 21-33yr
4,000 ppbforS h;
Prior to exposure, 4-week course
of daily placebo or Vitamin C and
Vitamin E.
BAL fluid immediately after
exposure (a1-protease inhibitor,
elastase inhibitory capacity,
TEARS, conjugated dienes, and
phospholipid phosphorus in lipid
extraction, albumin).
Pathmanathan
et al. (2003)
n = 8M, 4F;
26 yr (range: 21-32)
2,000 ppb for 4 h/day for 4 days;
Exercise 15 min on/15 min off at
75 watts
Biomarkers in bronchial epithelium-
exotoxin, GM-CSF, Gro-a, I-CAM
1, IL-5, IL-6, IL-8, IL-10, IL-13, total
and active NFKB, and TNF-a
(fiberoptic bronchoscopy after end
of last exposure).
Solomon et al.
n = 11 M, 4F;
29.3 ± 4.8 yr
2,000 ppb for 4 h/day for 3 days;
Exercise 30 min on/30 min off at
VE = 25 L/min
BW and BAL fluid analysis
immediately after exposure.
Differential cell counts, LDH,
peripheral blood parameters.
Vaqaqqini et Healthy; n = 7 M; 34 ± 5 yr 300 ppb for 1 h;
al- (1996) Asthma; n = 4 M, 4 F; Exercise at VE = 25 L/min
29 ± 14 yr
COPD; n = 7M; 58 ± 12 yr
Cell counts in sputum 2 h after
exposure.
BAL = bronchoalveolar lavage; BW = bronchial wash; CAPs = concentrated ambient particles; COPD = chronic obstructive
pulmonary disease; F = female; GM-CSF = granulocyte macrophage-colony stimulating factor; I-CAM = intercellular adhesion
molecule; IL = interleukin; LDH = lactate dehydrogenase; M = male; MPO = myeloperoxidase; NFKB = nuclear factor kappa-light-
chain-enhancer of activated B cells; NO2 = nitrogen dioxide; NR = not reported; PGE2 = prostaglandin E2; ppb = parts per billion;
SD = standard deviation; TEARS = thiobarbituric acid reactive substances; TNF-a = tumor necrosis factor alpha; VE = minute
ventilation; yr = year.
aSubjects were healthy individuals unless described otherwise.
fStudy published since the 2008 ISA for Oxides of Nitrogen.
5-231
-------
Table 5-37 Characteristics of animal toxicological studies of pulmonary
inflammation, injury, and oxidative stress.
Study
Ayaz and
Csallany
(1978)
Earth etal.
(1995)
Barth and
Muller
(1999)
Species (Strain);
Sample Size;
Sex; Age
Mice (C57BL/6J); n =
24-40/group; F;
18 months
Rats (Sprague
Dawley); n = 7/group;
M; age NR
Rats (Sprague
Dawley); n = 5/group;
M; age NR
Exposure Details
500 or 1 ,000 ppb NO2 for 1 7 mo,
continuous;
Animals had Vitamin E-deficient or
Vitamin E-supplemented diets.
5,000, 1 0,000, or 20,000 ppb NO2 for 3
or 25 days
5,000, 1 0,000, or 20,000 ppb NO2 for 3
or 25 days
Endpoints Examined
Glutathione peroxidase activity.
Histological evaluation,
morphometry, parenchymal and
vascular damage, pulmonary
arterial thickness, average medial
thickness.
Club cell morphology, cellular
proliferation, epithelial
proliferation.
de Burbure Rats (Wistar);
et al. (2007) n = 8/group; M
8 weeks
(1)1,OOOppbNO2for6h/day,
5 days/week for 4 weeks;
(2) 10,000 ppb NO2for6 h/day,
5 days/week for 4 weeks;
(3) 5,000 ppb NO2 for 6 h/day for
5 days;
(1-3) Animals had selenium-deficient or
selenium-supplemented diets.
BAL fluid lipid peroxidation,
antioxidant enzyme levels, protein
concentration, cell counts, oxidant
production, selenium levels, and
peripheral blood parameters.
Elsaved and Rats (Sprague-
Mustafa Dawley);
(1982) n = 12/group; sex NR;
8 weeks
3,000 ppb for 7 days;
Animals had Vitamin E-deficient or
Vitamin E-supplemented diets.
Lung tissue protein content, lipid
peroxidation, content, induction of
antioxidant enzymes.
Gregory et
al. (1983)
Rats (Fischer 344);
n = 4-6/group; sex NR;
14-16 weeks
(1) 1,000 or 5,000 ppb NO2 for 7 h/day
for 5 days/week for up to 15 weeks;
(2) 1,000 ppb NO2 for 0.5 h, 5,000 ppb
NO2for1.5h;
(3) 1,000 ppb NO2 for 3 h, 5,000 ppb
NO2for1.5h;
(4) 1,000 ppb NO2 for 0.5 h for
5 days/week for up to 15 weeks
Histopathological evaluation, BAL
fluid and lung homogenate
biochemical analysis (protein
concentration, LDH, glucose-6-
phosphate dehydrogenase,
alkaline phosphatase, glutathione
reductase, and glutathione
peroxidase).
Hatch et al.
Guinea pigs (Hartley);
n = >3/group; sex NR;
young adult
4,800 ppb NO2 for 3 h in deficient and
normal animals; 4,500 ppb NO2 for 16 h;
Animals had Vitamin C deficient or
normal diets.
BAL fluid protein and antioxidant
concentrations, 16 h after the 3 h
exposure and within 2 h after the
16 h exposure.
5-232
-------
Table 5-37 (Continued): Characteristics of animal toxicological studies of
pulmonary inflammation, injury, and oxidative stress.
Study
Species (Strain);
Sample Size;
Sex; Age
Exposure Details
Endpoints Examined
Ichinose and
Saqai (1989)
Rats (Wistar), guinea
pig (Hartley);
n = 6/group; M;
10 weeks
400 ppb NO2, 400 ppb Os, or 400 ppb
NO2 + 400 ppb Os for 24 h/day for
2 weeks
Lipid peroxidation, antioxidant
protective enzymes, total
proteins, TEA reactants,
nonprotein sulfhydryls in lung
homogenates, immediately after
exposure.
Last and Rats (Sprague
Warren Dawley); n >4/group;
(1987) M; age NR
5,000 ppb NO2, 1.0 mg/m3 NaCI or
hhSCM, 5,000 ppb NO2 + 1.0 mg/m3
NaCI, 5,000 ppb NO2 + 1.0 mg/m3
H2SO4for23.5 h/day for 1, 3, or 7 days
Collagen synthesis, BAL fluid
protein content and lavagable
enzyme activities, immediately
after exposure.
Mulleretal.
(1994)
Rats (Sprague
Dawley); n = 4; M; age
NR
800, 5,000, or 10,000 ppb NO2 for 1 and
3 days
BAL fluid cell counts and protein
concentration, phospholipid
component, SP-A, morphological
changes.
Mustafa et
al. (1984)
Mice (Swiss Webster);
n = 6/group; M;
8 weeks
(1)4,800 ppb NO2;
(2)4,500 ppbOs;
(2) 4,800 ppb NO2 + 4,500 ppb O3;
(1-3) for 8 h/day for 7 days
Physical and biochemical lung
parameters (lung weight, DNA,
protein contents, oxygen
consumption, sulfhydryl
metabolism, NADPH-generating
enzyme activities), immediately
after exposure.
Ohashi et al.
(1994)
Guinea pigs (Hartley);
n = 10/group; F; age
NR
3,000 or 9,000 ppb NO2 for 6 h/day,
6 times/week for 2 weeks
Pathology of mucosal samples:
accumulation of eosinophils,
epithelial injury, mucociliary
dysfunction (taken 24 h after end
of exposure period).
Pagan! et al.
(1994)
Rats (CD Cobs);
n = 5/group; M; age
NR
5,000 or 10,000 ppb NO2 for 24 h and
7 days
Analysis of BAL fluid and
superoxide anion production by
AMs.
Poynter et
al. (2006)
Mice (C57BL/6);
n = 5/group; sex and
age NR
5,000 or 25,000 ppb NO2 for 6 h/day for
1, 3, or 5 days
Analysis of BAL fluid and
histopathological evaluation
immediately or 20 days after
exposure.
Robison and
Forman
Rat (Sprague Dawley);
n = 3/group; M; age
NR
100, 1,000, 5,000, or 20,000 ppb NO2
for 1, 2, and 4 h ex vivo
Enzymatic production of
arachidonate metabolites in AMs,
cyclooxygenase products.
Robison et
al. (1993)
Rats (Sprague
Dawley);
n > 4/group; sex and
age NR
500 ppb NO2 for 8 h/day for 0.5, 1, 5, or
10 days
BAL fluid cell counts and
arachidonate metabolite levels,
AM arachidonate metabolism,
respiratory burst activity, and
glutathione content.
5-233
-------
Table 5-37 (Continued): Characteristics of animal toxicological studies of
pulmonary inflammation, injury, and oxidative stress.
Study
Species (Strain);
Sample Size;
Sex; Age
Exposure Details
Endpoints Examined
Rose et al.
(1989a)
Mice (CD-1);
n > 4/group; sex NR;
4-6 weeks
(1) 1,000, 2,500, or 5,000 ppb NO2 for Infection 5 and 10 days
6 h/day for 2 days; intra-tracheal
inoculation with murine
Cytomegalovirus; 4 additional days
(6 h/day) of exposure;
(2) Re-inoculation 30 days (air)
post-primary inoculation
post-inoculation, histopathological
evaluation, and analysis of BAL
fluid (LDH, albumin,
macrophages).
Belgrade et Guinea pigs (Hartley);
al. (1981) n = 7-31/group; M;
age NR
(1) 400, 1,000, 3,000, or 5,000 ppb NO2
for 72 h
(2) 400 ppb NO2 for 1 week
(3) 5,000 ppb NO2 for 3 h;
(1-3) Animals had Vitamin C-
supplemented diet or normal diet.
Protein concentration in BAL fluid,
lipid profile of BAL fluid,
histological evaluation of lung.
Sevanian et Rats (Sprague
al. (1982b) Dawley); n = 8/group;
sex NR; 8 weeks
3,000 ppb NO2 for 7 days
Animals had Vitamin E-deficient or
Vitamin E-supplemented diets.
Protein content, fatty acid
composition, lipid peroxidation in
lung microsomes.
Sherwin et
al. (1972)
Guinea pigs;
n = 4/group; M; age
NR
2,000 ppb NO2 continuously for 7, 14, or
21 days
Histopathological evaluation,
cellular damage by LDH staining.
Sherwin and
Carlson
(1973)
Guinea pigs;
n = 9/group; M; age
NR
400 ppb NO2 continuously for 1 week Protein concentration in BAL fluid.
Schlesinqer
(1987a)
Rabbit (New Zealand
White); M;
n = 5/group
0.5 mg/m3 H2SCM + 300 ppb NO2,
0.5 mg/m3 H2SCM + 1,000 ppb NO2 for
2 h/day for 2, 6, or 13 days
Cell counts in BAL fluid, AM
function 24 h after exposure.
Schlesinqer
etal. (1990)
Rabbits (New Zealand
White); n = 3/group; M;
age NR
(1) 1,000, 3,000, or 10,000 ppb NO2 for
2h;
(2) 3,000 ppb NO2 + 300 ppb O3 for 2 h;
(3) 100, 300, or 1,000 ppbO3for2h
Eicosanoids in BAL fluid,
immediately and 24 h after
exposure.
AM = alveolar macrophage; BAL = bronchoalveolar lavage; DNA = deoxyribonucleic acid; H2SO4 = sulfuric acid; F = female;
LDH = lactate dehydrogenase; M = male; NaCI = sodium chloride; NADPH = reduced nicotinamide adenine dinucleotide
phosphate; NO2 = nitrogen dioxide; NR = not reproted; O3 = ozone; ppb = parts per billion; SP-A = surfactant protein A;
TEA = thiobarbituric acid.
5-234
-------
5.2.7.5 Summary of Respiratory Effects in Healthy Individuals
The 2008 ISA for Oxides of Nitrogen did not make a specific assessment about the
respiratory effects of short-term exposure to oxides of nitrogen in healthy populations
(U.S. EPA. 2008c). However, previous and recent epidemiologic evidence indicates
ambient NO2-associated increases in respiratory symptoms and pulmonary inflammation
in children in the general population. Results from experimental studies of healthy
populations are variable; some show effects on events that may mediate the occurrence of
respiratory symptoms with NO2 exposures of 1,000 ppb and above but not lower.
Epidemiologic studies consistently show associations of short-term increases in ambient
NO2 with cough and pulmonary inflammation in school-aged children (Tables 5-34 and
5-36). While evidence overall is inconsistent, NO2-associated lung function decrements
were observed in healthy adults in studies characterized as having strong exposure
assessment with NO2 and copollutants measured on site of outdoor exposures near busy
roads or a steel plant (Dales etal.. 2013; Straketal.. 2012). In children, cough and
pulmonary inflammation were associated with same-day or 2- to 5-day averages (no clear
difference among lags) of 24-h avg NO2. However, there is some indication of lung
function decrements and increased pulmonary inflammation in healthy adults related to
shorter-duration NO2 exposures (5- or 10-hour) that persist 0 to 18 hours after exposure.
Although there is some epidemiologic evidence for NO2-related respiratory effects in
healthy adults, it is not clear to what extent the evidence in children reflects effects in
healthy children as most studies did not report the health status of study populations.
Despite the epidemiologic evidence supporting NO2-related respiratory effects in healthy
populations, there is uncertainty as to whether the results can be attributed to NO2
exposure specifically. Associations also were observed with PM2 5 and an array of
traffic-related pollutants, and few studies examined confounding by these copollutants.
As examined for lung function and pulmonary inflammation, NO2 associations persisted
in copollutant models with PM2 5 or a traffic-related copollutant among BC, OC, PNC, or
PM2.5 metal component (Steenhof etal., 2013; Straketal.. 2012; Lin etal., 2011). That
some of the copollutant effect estimates were attenuated with adjustment for NO2
indicates that NO2 may have confounded copollutant associations. Providing good
inference from copollutant model results, NO2 and copollutants were measured near
school or on location of outdoor exposures. Such informative epidemiologic studies in
healthy populations are few in number. Studies of symptoms in healthy populations did
not assess confounding by traffic-related copollutants, and while NO2 associations in the
U.S. multicity study persisted with adjustment for PMio or SO2 (Schwartz et al.. 1994).
5-235
-------
symptom studies assigned NO2 exposure from central sites. Limited evidence associating
indoor NC>2 at ice arenas with respiratory symptoms in hockey players (Salonen etal..
2008) provides some coherence for the findings for outdoor NC>2.
There is limited evidence from experimental studies demonstrating that NC>2 exposure
independently induces respiratory effects in healthy populations and reducing uncertainty
in the epidemiologic evidence base. Controlled human exposure studies do not indicate
an effect of NO2 exposures of 200-4,000 ppb (40 minutes to 5 hours) on respiratory
symptoms or lung function in healthy adults (Sections 5.2.7.2 and 5.2.7.3). Both
controlled human exposure and toxicological studies show variable effects on pulmonary
inflammation, injury, and oxidative stress (Section 5.2.7.4). Controlled human exposure
studies show increases in airway responsiveness and PMNs as well as development of a
pro-allergic phenotype in healthy adults, but effects are observed with NO2 exposures of
1,000 ppb and above, not lower concentrations (Sections 5.2.7.1 and 5.2.7.4). These are
key events underlying respiratory symptoms or development of allergic disease or asthma
(Section 4.3.5): however, the variable findings among NCh exposure concentrations
and/or specific endpoints limit the extent of support to the epidemiologic evidence for
respiratory effects in healthy populations.
5.2.8 Respiratory Mortality
Studies evaluated in the 2008 ISA for Oxides of Nitrogen that examined the association
between short-term NCh exposure and cause-specific mortality consistently found
positive associations with respiratory mortality, with some evidence indicating that the
magnitude of the effect was larger compared to total and cardiovascular mortality. Recent
multicity studies conducted in Asia (Wong et al.. 2008). China (Meng et al.. 2013; Chen
et al.. 2012b). and Italy (Faustini et al.. 2013; Chiusolo et al.. 2011). as well as a
meta-analysis of studies conducted in Asian cities (Atkinson et al.. 2012) add to the initial
body of evidence indicating larger respiratory mortality effects (Section 5.4.3 and
Figure 5-23). However, an additional multicity study conducted in Italy (Bellini et al..
2007). an extension of Biggeri et al. (2005). observed relatively consistent risk estimates
across mortality outcomes, which differs from the results of the original analysis and
complicates interpretation of whether there is differential risk among mortality outcomes.
The initial observation of consistent positive NC>2 associations with respiratory mortality
was examined in a few studies that conducted copollutant analyses. As with the
interpretation of NC>2 associations with total mortality (Section 5.4.4). it is difficult to
examine whether NC>2 is independently associated with respiratory mortality because
NC>2 is often highly correlated with other traffic-related pollutants. In the 17 Chinese
5-236
-------
cities study [China Air Pollution and Health Effects Study (CAPES)], Chen etal. (2012b)
found that NO2 risk estimates for respiratory mortality were slightly attenuated, but
remained positive in copollutant models with PMio and SO2 (9.8% [95% CI: 5.5, 14.2];
with PMio: 6.7% [95% CI: 2.9, 10.7]; with SO2: 7.0% [95% CI: 3.2, 11.0]; for a 20-ppb
increase in 24-h avg NO2 concentrations at lag 0-1 days). These results are consistent
with those of Meng et al. (2013) for COPD mortality in a study of four Chinese cities
(i.e., 7.1% [95% CI: 5.4, 8.9]; lag 0-1 for a 20-ppb increase in 24-h avg NO2
concentrations; with PMio: 6.0% [95% CI: 3.2, 8.8]; and with SO2: 6.9% [95% CI: 4.2,
9.5]). Chiusolo et al. (2011) also found evidence that associations between short-term
NO2 exposure and respiratory mortality remained robust in copollutant models in a study
of 10 Italian cities. In both an all-year analysis of NO2 with PMio (NO2: 13.7% [95% CI:
2.9, 25.8]; NO2 with PMio: 13.4% [95% CI: 2.9, 24.9]; for a 20-ppb increase in NO2
concentrations at lag 1-5 days), and a warm season (April-September) analysis of NO2
with O3 (NO2: 41.3% [95% CI: 16.2, 71.7]; NO2 with O3: 43.4% [95% CI: 14.6, 79.5]; for
a 20-ppb increase in NO2 concentrations at lag 1-5 days) NO2 associations with
respiratory mortality were relatively unchanged. However, when focusing on a subset of
respiratory mortality, specifically those deaths occurring out-of-hospital, in six Italian
cities, Faustini etal. (2013) reported evidence of an attenuation of the NO2-repiratory
mortality association in copollutant models with PMio (NO2: 24.5% [95% CI: 7.4, 44.2];
lag 0-5 for a 20-ppb increase in 24-h avg NO2 concentrations; NO2 with PMio: 11.8%
[95% CI: -7.5, 35.0]). Overall, the limited number of studies that have examined the
potential confounding effects of copollutants on the NO2-respiratory mortality
relationship generally indicate that associations remain relatively unchanged, but it is
difficult to disentangle the independent effects of NO2.
Of the studies evaluated, only the studies conducted in Italy examined potential seasonal
differences in the NO2-respiratory mortality relationship (Chiusolo et al.. 2011; Bellini et
al.. 2007). In a study of 15 Italian cities, Bellini et al. (2007) found that risk estimates for
respiratory mortality were dramatically increased in the summer from 1.4 to 9.1% for a
20-ppb increase in 24-h avg NO2 concentrations at lag 0-1 days, respectively, with no
evidence of an association in the winter. These results were further confirmed in a study
of 10 Italian cities (Chiusolo et al.. 2011). which also observed an increase in risk
estimates for respiratory mortality in the warm season (i.e., April-September) compared
to all-year analyses. Chiusolo et al. (2011) did not conduct analyses only for the winter
season. Although the respiratory mortality results are consistent with those observed in
the total mortality analyses conducted by Bellini et al. (2007) and Chiusolo et al. (2011).
as discussed in Section 5.4. studies conducted in Asian cities observed different seasonal
patterns, and it remains unclear whether the seasonal patterns observed for total mortality
would be similar to those observed for respiratory mortality in these cities.
5-237
-------
An uncertainty that often arises when examining the relationship between short-term air
pollution exposures and cause-specific mortality is whether the lag structure of
associations and the C-R relationship provide results that are consistent with what is
observed for total mortality. Chiusolo et al. (2011) in a study of 10 Italian cities found the
strongest evidence for an effect of NO2 on respiratory mortality at longer lags, with the
largest association at lag 2-5 days, which is indicative of a delayed effect (Figure 5-24).
These results are supported by the study of (Faustini et al., 2013) in six Italian cities,
which found the strongest evidence of an NCh-association with out-of-hospital
respiratory mortality at lags 2-5 and 0-5 days. Evidence of an immediate effect at lag
0-1 day avg was also observed, but the magnitude of the association was smaller
compared to lags 2-5 and 0-5 days. However, Chenetal. (2012b) in CAPES reported
the largest effect at single-day lags of 0 and 1 and the average of lag 0-1 days providing
support for an immediate effect of NO2 on respiratory mortality (Figure 5-25). When
examining longer lags, Chen et al. (2012b) reported that the magnitude of the association
was similar, albeit slightly smaller, for a 0-4 day lag, suggesting a potential prolonged
effect. In a study of COPD mortality in four Chinese cities (all four are examined in
CAPES), Mengetal. (2013) reported slightly different results than CAPES' respiratory
mortality results. When examining single-day lags from 0 to 7 days, the authors reported
the largest association for Lag Day 0. However, larger associations were observed in
multiday lag analyses with a similar magnitude of an association observed for lags 0-1
and 0-7 days, and the largest magnitude of an association overall for lag 0-4 days.
To date, analyses detailing the C-R relationship between air pollution and cause-specific
mortality have been limited. In the analysis of four Chinese cities, Meng etal. (2013) also
examined the NO2 and COPD mortality C-R relationship in each individual city. To
examine the assumption of linearity, the authors fit both a linear and spline model to the
city-specific NC^-COPD mortality relationship. Meng etal. (2013) then computed the
deviance between the two models to determine if there was evidence of nonlinearity. An
examination of the deviance did not indicate that the spline model improved the overall
fit of the NO2-COPD mortality relationship across the cities examined (Figure 5-15).
5-238
-------
**?
o
£
ra
o
E
Q
Q.
O
O
0>
-
0} O
V o H
O)
o
Beijing
Shanghai
• — •••• Guangzhou
Hongkong
50
100
NO2 concentration at lag 01 day
150
COPD = chronic obstructive pulmonary disease; NO2 = nitrogen dioxide.
Source: Reprinted with permission of Elsevier, Meng et al. (2013).
Figure 5-15 City-specific concentration-response curves of nitrogen dioxide
and daily chronic obstructive pulmonary disease mortality in four
Chinese cities.
5.2.9 Summary and Causal Determination
Evidence indicates that there is a causal relationship between short-term NO2 exposure
and respiratory effects based on the coherence among multiple lines of evidence and
biological plausibility for effects on asthma exacerbation. There is some support for
NO2-related exacerbation of respiratory allergy and COPD, respiratory infection,
respiratory mortality, and respiratory effects in healthy populations. However, because of
inconsistency among lines of evidence and consequent uncertainty about the effects of
NO2 exposure, evidence for these other nonasthma respiratory effects does not strongly
contribute to the determination of a causal relationship.
5-239
-------
The determination of a causal relationship represents a change from the "sufficient to
infer a likely causal relationship" concluded in the 2008 ISA for Oxides of Nitrogen (U.S.
EPA. 2008c). Consistent with previous findings, recent epidemiologic results indicate
associations between ambient NO2 concentrations and asthma-related respiratory effects.
Biological plausibility continues to be provided by the NCh-induced increases in airway
responsiveness and allergic inflammation demonstrated in experimental studies. The
2008 ISA cited uncertainty as to whether NC>2 has an effect independent from other
traffic-related pollutants, and additional copollutant model results show ambient
NO2-associated increases in asthma-related effects with adjustment for PlVfc 5, BC/EC,
UFP, OC, metals, VOCs, or CO. Thus, much of the evidence for NO2-related respiratory
effects was available in the 2008 ISA. However, the 2008 ISA emphasized epidemiologic
findings and did not assess the coherence and biological plausibility for various
respiratory conditions separately, which is important given that the weight of evidence
varies among respiratory conditions. More than new findings, the evidence integrated
across outcomes related to asthma exacerbation, with due weight given to the
experimental evidence, is sufficient to rule out chance, confounding, and other biases
with reasonable confidence and support a change in conclusion from likely to be causal to
causal relationship. The evidence for a causal relationship is detailed below using the
framework described in the Preamble (Table ID. The key evidence as it relates to the
causal framework is presented in Table 5-39.
5.2.9.1 Evidence on Asthma Exacerbation
A causal relationship between short-term NO2 exposure and respiratory effects is strongly
supported by evidence for effects on clinical events and pulmonary responses that
indicate and mediate asthma exacerbation. The evidence from controlled human exposure
studies for NO2-induced increases in airway responsiveness in adults with asthma is
sufficient to support the biological plausibility for the effects of NO2 exposure on asthma
exacerbation. Increased airway responsiveness can lead to asthma symptoms such as
wheeze. Increases in airway responsiveness and doubling reduction in provocative dose
are demonstrated in adults with asthma following 200 to 300 ppb NO2 exposures at rest
for 30 minutes and 100 ppb for 1 hour [(Brown. 2015; Folinsbee. 1992); Section 5.2.2.11.
The findings for clinically relevant airway responsiveness with NO2 exposures not much
higher than peak ambient concentrations (Section 2.5.3) particularly support an effect on
asthma exacerbation of ambient NO2 exposures. Further linking short-term NO2 exposure
to asthma exacerbation is evidence for NO2 exposures of 260-400 ppb enhancing allergic
inflammation (e.g., eosinophil activation, Th2 cytokines, PMNs) in humans with allergic
asthma and a rat model of allergic disease [(Ezratty et al., 2014; Barck et al., 2005a;
5-240
-------
Barck et al.. 2002; Wang etal.. 1999); Sections 4.3.2.6 and 5.2.2.51. NO2-associated
increases in pulmonary inflammation also were found in epidemiologic studies of
populations with asthma (Section 5.2.2.5). In experimental studies, ambient-relevant NO2
exposures increased eicosanoids (involved in PMN recruitment) but did not consistently
affect pulmonary injury or oxidative stress (Section 5.2.7.4). These inconsistent findings
for other events in the proposed mode of action linking NC>2 exposure and asthma
exacerbation are not considered to weaken the evidence for a relationship with NO2
because the effects were studied mostly in healthy humans and animal models.
Also supporting a causal relationship is the coherence of NCh-induced increases in
airway responsiveness and allergic inflammation with the epidemiologic associations
observed between short-term increases in ambient NO2 concentration and increases in
asthma symptoms in children [(Zoraet al.. 2013; Gent et al.. 2009; Delfino et al.. 2003;
Delfino et al.. 2002); Section 5.2.2.3] and increases in asthma hospital admissions and
ED visits among subjects of all ages and children Rlskandar et al., 2012; Strickland et al.,
2010; Jalaludin et al.. 2008; Villeneuve et al.. 2007); Section 5.2.2.4]. Epidemiologic
evidence for NCh-related decreases in lung function in populations with asthma is
inconsistent as a whole, but associations were found with lung function measured under
supervised conditions [(Greenwald et al.. 2013; Martins etal.. 2012; Delfino et al..
2008a; Holguin et al.. 2007; McCreanor etal.. 2007; Delfino et al.. 2003);
Section 5.2.2.2]. Most controlled human exposure studies found no effect of NO2
exposure (120-4,000 ppb) on respiratory symptoms or lung function in adults with
asthma. In contrast with airway responsiveness, symptom and lung function assessments
did not include challenge with a bronchoconstrictor.
Individual epidemiologic studies examined multiple outcomes and lags of exposure, and
not all studies had statistically significant results. However, the pattern of association
observed for NCh supports the consistency of evidence and does not indicate a high
probability of associations found by chance alone. Consistency also is demonstrated as
evidence for NCh-related asthma exacerbation across diverse locations in North America,
Europe, and Asia, including recent multicity studies. Most evidence was for multiday
lags of NCh exposure of 2 to 5 days, but associations also were found with lags of 0 or
1 day. A larger magnitude of association is not clearly indicated for a particular lag of
exposure. Asthma hospital admissions and ED visits were associated with 24-h avg and
1-h max NCh, and risk estimates ranged from a 4.5 to 34% increase per 20-ppb increase
in 24-h avg NCh or 30-ppb increase in 1-h max NCh. Symptoms and lung function in
children with asthma were associated primarily with 24-h avg NCh. The recruitment of
children from schools supports the likelihood that study populations were representative
of the general population of children with asthma. Issues with selective participation by
certain groups were not reported. The concentration-response relationship was analyzed
5-241
-------
for pediatric asthma ED visits in Atlanta, GA and Detroit, MI, and neither a threshold nor
deviation from linearity was found in the range of 24-h avg or 1-h max ambient NC>2
concentrations examined (Li etal.. 20lib; Strickland et al.. 2010).
Allergic inflammation promotes bronchoconstriction and airway obstruction. Thus,
strengthening the link between NCh-related increases in allergic inflammation and airway
responsiveness and epidemiologic evidence for asthma exacerbation are the
NO2-associated increases in symptoms and decreases in lung function in populations of
children with asthma with high prevalence of atopy (e.g., 47-84%). Airway obstruction
in response to allergens can lead to lung function decrements and respiratory symptoms.
The evidence for NCh-induced increases in airway responsiveness and allergic
inflammation provides sufficient biological plausibility for the asthma-related effects
observed in epidemiologic studies.
5.2.9.2 Evidence on Nonasthma Respiratory Effects
Epidemiologic studies demonstrate associations of ambient NO2 concentrations with
hospital admissions and ED visits for all respiratory causes combined (Table 5-39).
suggesting that the respiratory effects of short-term NC>2 exposure may extend beyond
exacerbation of asthma. However, when other respiratory conditions are evaluated
individually, there is uncertainty about relationships with NC>2 because of inconsistency
among disciplines and/or inconsistency of findings across the array of clinical and
subclinical effects. Where epidemiologic associations were found, limited examination of
potential confounding by traffic-related copollutants results in weak inference about NCh
effects. Experimental evidence for NCh-induced increases in airway responsiveness and
allergic inflammation supports effects on allergy exacerbation, but epidemiologic
evidence is inconsistent (Section 5.2.3). For COPD exacerbation and respiratory infection
(Sections 5.2.4 and 5.2.5). evidence from epidemiologic, controlled human exposure, and
toxicological studies is inconsistent across outcomes such as hospital admissions, ED
visits, symptoms, lung function, and immune cell function; thus, a direct effect of NC>2
exposure is not clearly demonstrated (Table 5-39). Epidemiologic studies consistently
found NC>2-associated increases in respiratory mortality (Section 5.2.8). but the spectrum
of respiratory effects that can lead to mortality is not entirely clear. Among the leading
causes of mortality, COPD, and respiratory infections are the ones related to respiratory
causes (Hoyert and Xu. 2012). but these conditions are not clearly related to NO2
exposure. Epidemiologic evidence also indicates ambient NCh-associated respiratory
effects in healthy populations (Section 5.2.7). as cough and pulmonary inflammation in
children in the general population and healthy adults. However, an independent effect of
NC>2 is uncertain because of limited support from experimental studies.
5-242
-------
5.2.9.3 Evaluation of Nitrogen Dioxide Exposure Assessment
Most epidemiologic evidence indicating ambient NCh-related asthma exacerbation is
based on exposure assessment from central site monitors. Substantiating the
epidemiologic evidence are several findings for associations with NO2 measurements in
subjects' location(s), including personal total and outdoor NO2 as well as NCh measured
outside schools (Greenwald et al.. 2013; Zoraet al.. 2013; Martins et al.. 2012; Sarnat et
al.. 2012; Delfino et al.. 2008a: Holguin et al.. 2007; McCreanor et al.. 2007). Ambient
NO2 concentrations are highly variable across locations (Sections 2.5.2 and 2.5.3). Thus,
compared to measurements at central sites, NO2 measurements in subjects' locations may
belter represent temporal variation in subjects' ambient exposures in those locations. NO2
concentrations summed across individuals' microenvironments have shown good
agreement with total personal NO2 (Section 3.4.3.1). demonstrating that
microenvironmental ambient concentrations are important determinants of exposure.
Imparting confidence in results for personal NC>2, no issues were reported regarding
measurements being near the LOD. Further supporting asthma exacerbation in relation to
ambient NCh exposure, for some study areas, central site concentrations were reported to
be correlated with total personal NO2 (Delfino et al.. 2008a). outdoor school NCh (Sarnat
et al., 2012). or NO2 measured at other central sites in the area (Section 2.5.2). In support
of exposure assessment from central sites, larger ambient NCh-associated increases in
respiratory hospital admissions and ED visits were found in the warm season.
Personal-ambient NCh correlations are higher in the warm than cold season
(Section 3.4.4.3). pointing to lower potential NC>2 exposure error.
The studies with microenvironmental exposure assessment provide some, albeit far from
conclusive, indication that short-term NC>2 exposures near sources may be related to
respiratory effects. Respiratory effects were associated with ambient NO2 measured
across locations with varying traffic intensities or distance to highways (Steenhof et al..
2013; Strak et al.. 2012). Other studies compared NO2 associations among locations,
observing respiratory effects in association with NC>2 at a school in a high but not low
traffic area (Greenwald et al.. 2013; Sarnat etal.. 2012) or larger respiratory effects near a
steel plant than in a residential area (Dales etal.. 2013). However, none of the latter
studies examined whether the findings were attributable to NO2 independently of
correlated copollutants or differences between schools in population characteristics such
as race/ethnicity, body mass index, or asthma medication use. Informing this uncertainty,
McCreanor et al. (2007) found that adults with asthma had larger decreases in lung
function and increases in eosinophil activation after walking along a high traffic road in
London, U.K. than after walking in a park. Results from copollutant models provide
evidence that respiratory effects associated with personal ambient NO2 exposures near
high traffic roads were independent of personal ambient EC, UFP, or PIVb 5 exposures.
5-243
-------
Whether estimated from central sites or subjects' locations, NO2 exposure metrics largely
were integrated over 24 hours or 2-15 hours. The diurnal temporal pattern of exposure
(e.g., acute peaks) underlying the associations of respiratory effects with daily average or
multiday averages of NO2 is unclear. However, studies conducted in outdoor locations
with varying traffic intensities indicate increases in pulmonary inflammation and
decreases in lung function in association with 2- or 5-h avg NC>2 exposures that ranged
between 5.7 and 153.7 ppb (Straketal.. 2012; McCreanor et al.. 2007).
5.2.9.4 Evaluation of Confounding
Also indicating an independent effect of NO2 exposure on asthma exacerbation are
epidemiologic associations found for NCh with statistical adjustment for potential
confounding factors such as temperature, humidity, season, medication use, and, in
particular, copollutants. Based on a common source and moderate to high correlations
with NO2, confounding by other traffic-related pollutants is a major concern
(Sections 1.4.3 and 5.1.2.1). Copollutant models were the predominant method used for
evaluating copollutant confounding, and most of these studies found that NO2
associations persisted with adjustment for PlVfc 5, BC/EC, OC, UFP, PNC (Figure 5-16
and Table 5-38). PM2 5 metal components, VOCs, or CO (Figure 5-17 and Table 5-38).
Copollutant models also indicated that NO2 associations with asthma and other
respiratory effects were independent of PMio-2.5, PMio, SC>2, or Os [Supplemental
Figure S5-1 (U.S. EPA. 2015a)1. O3 (r = -0.61 to 0.45) and PMio (r = -0.71 to 0.59)
showed a wide range of correlations with NO2, from strongly negative to moderately
positive; 862 was moderately correlated with NO2 (r = 0.31-0.56). Inference regarding
confounding by PIVb 5 and traffic-related copollutants is strongest for exposures assessed
in subjects' locations. Exposure measurement error may be similar for NC>2 and
copollutants, thereby improving the reliability of copollutant models. These studies
reported a wide range of correlations for NCh with PM2 5 and traffic-related copollutants
(r = -0.42 to 0.68). Also strengthening inference from copollutant models, in some
studies, personal NCh was not strongly positively correlated with personal copollutants
(r = 0.20-0.33 for EC, OC, PM2 5; -0.42 to 0.08 for ethylbenzene and benzene) (Martins
et al.. 2012; Delfino et al.. 2006). No issues were reported regarding personal exposure
metrics for any of these pollutants being near the LOD. As examined in multiple
populations with asthma, associations of lung function and pulmonary inflammation with
personal total or outdoor NO2 and NO2 measured within 650 m of a children's school
persisted with adjustment for PIVb 5 or BC/EC (Martins et al.. 2012; Lin et al.. 2011;
Delfino et al.. 2008a; McCreanor et al.. 2007; Delfino et al.. 2006). Results were similar
in a study of healthy adults (Steenhof et al.. 2013; Strak et al.. 2012). In some cases, the
5-244
-------
95% CIs for NO2 associations are exaggerated because the increment used to standardize
effect estimates is far larger than the variability in NO2 concentrations during the study
period (Martins etal.. 2012; Strak etal.. 2012). Most studies examining exposures at
subjects' locations did not examine CO in single- or co-pollutant models. However,
among children in the general population, the association between outdoor school NO2
and lung function persisted with adjustment for CO (Correia-Deur et al.. 2012).
As examined in only one or two studies with exposures assessed in subjects' locations,
NO2 associations persisted with adjustment for OC or metal PM2 5 components such as
iron and copper (Steenhof et al.. 2013; Strak et al.. 2012; Delfino et al.. 2006).
Information on potential confounding by UFP/PNC or VOCs also is limited, and results
from copollutant models are more variable. However, rather than clearly demonstrating
confounding of NO2 associations, results show that adjustment for UFP/PNC or benzene
attenuated one outcome in a study but not another (Steenhof et al.. 2013; Martins et al..
2012; Strak et al.. 2012; McCreanor et al., 2007). There also was some evidence that NO2
exposure confounded associations for traffic-related copollutants or PM25. Some
associations of personal ambient PM25, EC/BC, OC, copper, UFP/PNC, or benzene with
respiratory effects were attenuated with adjustment for personal ambient NO2 (Martins et
al.. 2012; Strak et al.. 2012; McCreanor et al.. 2007). Also indicating an independent
association for NO2, some studies found associations with school or personal NO2 but not
EC, OC, or PM2 5 (Sarnat et al.. 2012; Delfino et al.. 2008a; Holguin etal.. 2007).
Copollutant models based on central site concentrations indicate that ambient NO2
remains associated with asthma- and nonasthma-related respiratory effects with
adjustment for PM2 5 (Iskandar etal.. 2012: Dales et al.. 2009a; Jalaludin et al.. 2008;
Villeneuve et al., 2007; von Klot et al., 2002) or as examined in fewer studies, UFP, CO,
VOCs, or a source apportionment factor comprising EC and various metals (Delfino et
al.. 2013; Gent et al.. 2009; Tolbert et al.. 2007; Delfino et al.. 2003). Several
traffic-related PM constituents are shown to induce oxidative stress (Appendix), and
Delfino etal. (2013) found an NO2 association with adjustment for the oxidative potential
of PM2 5 extracts. Observations that ambient NO2-associated increases in respiratory
hospital admissions and ED visits are larger in the warm than cold season also point to an
NO2 association that may be independent of PM2 5. NO2-PM2 5 correlations are lower in
the warm season (Section 3.4.4.1). pointing to lower potential confounding by PM2 5.
NO2 and Os are not strongly positively correlated in the warm season. As with NO2
measured in subjects' locations, some central site NO2 associations were attenuated with
adjustment for PM2 5 or UFP for some but not all outcomes within the same studies
(Dales et al.. 2009a; Liu et al.. 2009b; von Klot et al.. 2002). and a clear confounding
effect was not demonstrated. Similar to NO2 measured in subjects' locations, central site
NO2 showed a range of correlations with traffic-related pollutants (r = 0.43-0.74).
5-245
-------
However, because spatial distributions may differ (Section 3.3.1.1). exposure
measurement error may differ between central site concentrations of NC>2 and
traffic-related copollutants, resulting in weaker inference from copollutant models.
Confounding by any particular copollutant was examined to a limited extent, and not all
potentially correlated pollutants were examined. Further, inference from copollutant
models can be limited (Section 5.1.2.2). and methods to adjust for multiple copollutants
simultaneously are not reliable. Thus, residual confounding is likely. However, evidence
integrated from copollutant models based on pollutants measured in subjects' locations
and from controlled human exposure studies support an effect of ambient NO2 exposure
on asthma exacerbation independent of other traffic-related pollutants.
5.2.9.5 Evaluation of Nitrogen Dioxide-Copollutant Mixture Effects
As a component of an air pollution mixture, NC>2 potentially can induce health effects in
combination with other pollutants. Controlled human exposure studies, with well-defined
NO2-copollutant co-exposures, do not provide strong evidence that NO2 exposure affects
lung function and airway responsiveness differentially when occurring alone or as part of
a mixture with PM2 5 (Gong et al.. 2005). SO2 (Devalia et al.. 1994). or O3 (simultaneous
or sequential exposure) (Jenkins et al.. 1999; Hazuchaet al.. 1994; Adams et al.. 1987).
Interactions with CO, EC/BC, or UFP have not been examined in controlled human
exposure studies. Limited epidemiologic findings point to increases in asthma-related
symptoms and ED visits when short-term averages of ambient NC>2 concentration are
jointly high with PIVb 5, the traffic-related pollutants CO and EC, and/or Os and SO2
(Gass etal.. 2014; Winquist et al.. 2014; Schildcrout et al.. 2006). However, there is no
clear indication that the combined effects of NO2 with PM2 5, CO, EC, or VOCs are larger
than effects estimated for NO2 alone (Gass etal.. 2014; Winquist et al.. 2014; Schildcrout
et al.. 2006; Delfino et al.. 2003). These epidemiologic studies of joint effects or
interactions have limited inference because exposure measurement error in the central
site pollutant metrics may obscure interactions between personal exposures. They also do
not provide information on whether NO2 has an independent effect from other pollutants.
5.2.9.6 Indoor Nitrogen Dioxide-Related Asthma Exacerbation
A causal relationship between NO2 and respiratory effects also is supported by the
coherence of asthma-related effects associated with ambient and indoor NO2 (Table
5-39). No issues were reported regarding a large number of indoor measurements being
near the LOD. In schools in Ciudad Juarez, Mexico, correlations of NO2 with BC,
5-246
-------
PMio, PMio-2.5, and SC>2 differed indoors and outdoors, suggesting that NC>2 was part of a
different pollutant mixture indoors and outdoors. NC>2 also may be part of different
mixtures inside homes and classrooms because gas heaters and not stoves are a major
source of classroom NC>2. Cooking has been shown to be a more important determinant
of indoor UFP than heating systems (Weichenthal et al., 2007). Thus, associations with
indoor classroom NC>2 may be less likely to be confounded by UFP than are associations
with indoor home NC>2 or ambient NC>2. Mean concentrations of indoor NC>2, averaged
over 3 to 7 days, were in the range of ambient concentrations (Table 5-39). except for a
mean of 121 ppb at one school. Indoor NO2 concentrations, particularly at home, can
exhibit acute peaks that deviate from the mean (Table 3-4). As with ambient NO2, the
temporal pattern of NO2 concentrations underlying associations of asthma-related effects
with multiday averages of indoor NCh is not understood.
5.2.9.7 Conclusion
Multiple lines of evidence support a relationship between short-term NO2 exposure and
asthma exacerbation. Some findings point to effects on allergy, COPD, respiratory
infection, respiratory effects in healthy populations, and respiratory mortality, but there is
inconsistency among disciplines and outcomes. The NC>2-induced increases in allergic
inflammation and airway responsiveness in controlled human exposure studies of adults
with asthma comprise the key evidence that NO2 exposure can independently exacerbate
asthma and support the epidemiologic evidence for asthma hospital admissions and ED
visits, as well as symptoms, lung function decrements, and pulmonary inflammation in
populations with asthma. These studies found associations with 24-h avg and 1-h max
NC>2 concentrations, at lags of 0 or 1 day and multiday averages of 2 to 5 days. The range
of mean ambient concentrations was 11.3-30.9 ppb for 24-h avg NC>2, 75.5 ppb for
2-h avg NC>2, and 23.0-44.4 ppb for 1-h max NC>2. The epidemiologic evidence is
substantiated by findings for NC>2 measured in subjects' location(s), including personal,
ambient school, ambient near-road, or indoor concentrations. Further, associations of
personal total or ambient NO2 or school NO2 with asthma-related effects persist with
adjustment for PM2 5 or a traffic-related pollutant such as BC/EC, UFP, OC, a PM2 5
metal, or VOC. Inference from copollutant models is limited as is the breadth of analysis
of traffic-related copollutants and copollutant interactions. Thus, epidemiologic evidence
for NO2-associated asthma exacerbation and biological plausibility from NCh-induced
increases in airway responsiveness and allergic inflammation in adults with asthma
together are sufficient to conclude that there is a causal relationship between short-term
NC>2 exposure and respiratory effects.
5-247
-------
Study
PM2.5
McCreanor et al. (2007)a FEF25-75%
Delfino et al. (2008a)
Correlation with NO2
Exposure Assessment
0.60
Personal ambient
0.38
Total personal
0.36
Central site within 8,16 km of homes
Lin etal. (2011)
Moshammer et al. (2006)£
ATSDR (2006)
Delfino et al. (201 3)b
Burnett etal. (1997)
Jalaludin etal. (2008)
Liu et al. (2009)
Iskandaret al. (2012)
von Mot et al. (2UU2)
von o e a . ( )
eNO
FEV1
Asthma ED visits
eNO
Respiratory Hospital Ad.
Asthma ED visits
FEV1
Asthma Hospital Ad.
Wheeze
e a-agoni se
0.30 I —9 —
Central site 0.65 km from school | — O
0.54 I •
Central site next to school | O
0.61 I •——
Central site - 1 site -fO
Central site - 1 or 2 sites per city — 1 O
0.45 1 •
Central sites - city average — KIH—
0.45 warm , 0.68 cold 1 •
Central sites - city average 1 — O—
0.71 -te-
Central site within 1 0 km of homes — C|—
r t 'ti ' 1 "~
Central site 0.5 km from traffic O
EC/BC/BS
McCreanor et al.
(2007)a FEF25-75%
0.58
Personal ambient
Straket al. (2012)
Straket al. (2012)
LJn etal. (2011)
Gent et al. (2009)c
Peacock etal. (2011)
FVC
eNO
eNO
Wheeze
Symptomatic Fall PEF
0.67
Location of outdoor exposures
0.67
Location of outdoor exposures
0.68
Central site 0.65 km from scho(
0.49
Central site
NR
PNC/UFP
McCreanor et al. (2007)a
Central site -1 site
0.58
Personal Ambient
Straket al. (2012)
Iskandaret al. (2012)
von Mot et al. (2UU2)
von Mot et al. (2\JU2)
Bruske et al. (2010)
FVC
Asthma Hospital Ad.
beta agonist Use
Lymphocytes
0.56
Location of outdoor exposures
Location of outdoor exposures -*O
r t
n fi'fi J '" liari'L
r t 't ff
—
u
*
Percent change in respiratory outcome per increase
in NO2 (95% Cl)d in single- and co-pollutant models
Note: BS = black smoke; Cl = confidence interval; EC/BC = elemental/black carbon, ED = emergency department; eNO = exhaled
nitric oxide; FEF2s-75% = forced expiratory flow between 25 and 75% of forced vital capacity; FENA = forced expiratory volume in 1
second; FVC = forced vital capacity; km = kilometer; m = meter; NO2 = nitrogen dioxide, PEF = peak expiratory flow;
PM2.5 = particles with a nominal mean aerodynamic diameter less than equal to 2.5 |jm, PNC = particle number concentration;
UFP = ultrafine particles. Black = studies reviewed in the 2008 Integrated Science Assessment for Oxides of Nitrogen, Red = recent
studies, Closed circles = NO2 effect estimate in a single-pollutant model; open circles = NO2 effect estimate adjusted for a
copollutant. Magnitude and precision of effect estimates should not be compared among different outcomes. Results are organized
by copollutant analyzed then by exposure assessment method. Percentage change in FEF25-75o/0, FENA,, or FVC refers to percentage
decrease. Quantitative results presented in Table 5-38.
aTo fit results in the figure, effect estimates are multiplied by 10. bCopollutant is ROS generated from PM25 extract. °Copollutant is a
source apportionment factor comprising EC and various metals. dEffect estimates standardized to a 20-ppb increase for 24-h avg
NO2 and a 30-ppb increase for 1-h max NO2. Effect estimates for 2-h, 5-h, or 15-h avg NO2 are not standardized but presented as
reported in their respective studies (see Section 5.1.2.2).
Figure 5-16 Associations of ambient or personal nitrogen dioxide with
respiratory effects adjusted for fine particulate matter,
elemental/black carbon, or particle number
concentration/ultrafine particles.
5-248
-------
Study
Benzene
Martins et al. (2012)
Martins et al. (2012)
Outcome
EBCpH
FEV1
Correlation with NO2
Exposure Assessment
-0.43 to 0.14
Individual TWA
-0.43-0.14
Individual TWA •*
Ethvlbenzene
Martins etal. (2012) EEC pH
Martins etal. (2012) FEV1
CO
Correia-Deur et al. (2012) PEF
Tolbert et al. (2007) Respiratory ED visits
Jalaludin et al. (2008) Asthma ED visits
-0.43 to 0.14
Individual TWA
-0.43 to 0.14
Individual TWA
0.51
Outdoor school
0.70
Central sites - city average
0.71 warm, 0.55 cold
Central sites - city average
O
O
-10
0
10
20
30
40
Percent change in respiratory outcome per increase
in NO2 (95% Cl)ain single-and co-pollutant models
Note: Cl = confidence interval; CO = carbon monoxide; EEC = exhaled breath condensate; ED = emergency department;
FENA, = forced expiratory volume in 1 second; NO2 = nitrogen dioxide; PEF = peak expiratory flow; TWA = time-weighted average;
VOC = volatile organic compound. Black = studies reviewed in the 2008 Integrated Science Assessment for Oxides of Nitrogen,
Red = recent studies, Closed circles = NO2 effect estimate in a single-pollutant model; open circles = NO2 effect estimate adjusted
for a copollutant. Magnitude and precision of effect estimates should not be compared among different outcomes. Results are
organized by copollutant analyzed then by exposure assessment method. Percentage change in EEC pH, FENA,, and PEF refers to
percentage decrease. Quantitative results presented in Table 5-38.
aEffect estimates standardized to a 20-ppb increase for 24-avg NO2 and a 30-ppb increase for 1-h max NO2.
Figure 5-17 Associations of ambient nitrogen dioxide with respiratory effects
adjusted for a volatile organic compound or carbon monoxide.
5-249
-------
Table 5-38 Corresponding effect estimates for nitrogen dioxide-associated
respiratory effects in single- and copollutant models presented in
Figures 5-16 and 5-17.
Respiratory
Study Outcome
McCreanor et FEF2s-7s%
al. (2007)
tDelfino et al. FEVi %
(2008a) predicted
tZhu(2013): eNO
Lin et al.
(2011)
Moshammer FEVi
et al. (2006)
ATSDR (2006) Asthma ED
visits
fDelfino etal. eNO
(2013)
Jalaludin et al. Asthma ED
(2008) visits
NO2
Averaging
Time and
Lag
2-h avg
Lag 0 h
24-h avg
Lag 0-1 -day
avg
24-h avg
Lag 0 day
8-h avg
(12-8 a.m.)
Lag 0 day
24-h avg
Lag 0-4-day
avg
24-h avg
Lag 0-1 -day
avg
1-h max
Lag 0-1 -day
avg
% Change in Outcome (95% Cl)
per Increase in NO2a
Exposure
Assessment Correlation
Method with NO2
Personal 0.60
ambient
0.58
0.58
Total personal 0.38
Central site 0.36
within 8 or
16 km of
homes
Central site 0.30
0.65 km of
school
0.68
Central site 0.54
next to school
Central site: 0.61
1 site
Central site: 1 0.43
or 2 sites per
city
Central site: 0.45 warm
city average 0.68 cold
Single-
Pollutant
Model
7.8(1.3, 14)b
per 5.3 ppb
NO2
1.7(0.19, 3.2)
1.3(0.15,2.4)
22(18,26)
8.9(3.7, 14)b
per 5.32 ppb
NO2
5.8(0.59, 11)
9.0(2.9, 15)
7.4(4.5, 10)
Copollutant
Model
With PIVb.s:
4.8 (-2.5, 13)b
With EC:
4.3 (-2.6, 11)b
With UFP:
4.7 (-3.9, 13)b
With PIvh.s:
1.3 (-0.22, 2.8)
With PIvh.s:
0.86 (-0.89, 2.6)
With PlVh.s:
14(9.5, 19)
With BC:
5.6(0.38, 11)
With PlVh.s:
10(4.2, 16)b
With PlVh.s:
3.5 (-1.8, 9.0)
With PlVh.s:
5.8 (-1.9, 14)c
With PlVh.s:
4.5(1.3, 7.8)
5-250
-------
Table 5-38 (Continued): Corresponding effect estimates for nitrogen dioxide-
associated respiratory effects in single- and copollutant
models presented in Figures 5-16 and 5-17.
Respiratory
Study Outcome
tLiu(2013); FEVi
Liu et al.
(2009b)
flskandar et Asthma
al. (2012) hospital
admissions
von Klot et al. Wheeze
(2002)
Beta agonist
use
tStrak(2013): FVC
Strak et al.
(2012)
eNO
tGent et al. Wheeze
(2009)
fPeacock et Symptomatic
al. (2011) fall in PEF
NO2
Averaging
Time and
Lag
24-h avg
Lag 0— 2-day
avg
24-h avg
avg
24-h avg
Lag 0-4-day
avg
5-h avg
Lag 0 h
NR
Lag 0 day
1-h max
Lag 1 day
% Change in Outcome (95% Cl)
per Increase in NO2a
Exposure
Assessment Correlation
Method with NO2
Central site 0.71
within 10 km
of homes
Central site 0.33
within 15 km
of hospital
0.51
Central site 50 0.68
m from traffic
0.66
0.68
0.66
Location of 0.67
outdoor
exposures
0.56
0.67
0.56
Central site 0.49
Central site: NR
1 site
Single-
Pollutant
Model
1.2 (-0.84, 3.2)
30(10,60)
15(2.0,28)
20(5.0, 37)
1.8(0.44, 3.2)
per 10.54 ppb
NO2
6.9 (-1.9, 16)
per 10.54 ppb
NO2
NR
13 (-3.0, 31)
Copollutant
Model
With PlVh.s:
-1.2 (-6.4, 3.8)
With PIvh.s:
40 (20, 70)
With UFP:
50 (20, 80)
With PlVh.s:
12 (-7.0, 35)
With UFP:
2.0 (-14, 21)
With PlVh.s:
25 (5.0, 49)
With UFP:
22 (5.0, 43)
With EC:
2.3(0,4.6)
With PNC:
1.3 (-0.58, 3.1)
With EC:
4.1 (-6.0, 14)
With PNC:
-7.4 (-19, 3.9)
With source
apportionment
factor of EC, zinc,
copper, lead:
8.0 (-1.0, 18)
With BS:
6.2 (-17, 34)
5-251
-------
Table 5-38 (Continued): Corresponding effect estimates for nitrogen dioxide-
associated respiratory effects in single- and copollutant
models presented in Figures 5-16 and 5-17.
% Change in Outcome (95% Cl)
per Increase in NO2a
Study
tBruske
(2014): Bruske
etal. (2010)
fMartins
(2013):
Martins et al.
(2012)
fCorreia-Deur
etal. (2012)
Tolbert (2009):
Tolbert et al.
(2007)
Jalaludin et al.
(2008)
Respiratory
Outcome
Lymphocytes
in BAL fluid
EBCpH
FEVi
PEF
Respiratory
ED visits
Asthma ED
visits
IMO2
Averaging
Time and
Lag
24-h avg
I in n OQ h
24-h avg
Lag 0-4-day
avg
24-h avg
Lag 0 day
1-h max
avg
1-h max
Lag 0-1 -day
avg
Exposure Single-
Assessment Correlation Pollutant
Method with NO2 Model
Central site 0.66 8.4 (-5.0, 24)
within 3.5 km
of homes
Individual -0.42 to 2.6(1.3,3.9)
TWA based 0.14 across
on outdoor time periods
monitoring,
modeling,
time-activity
data
22(1.5, 38)
Outdoor 0.51 1.9 (-0.38, 4.1)
school
Central sites: 0.70 2.6(1.3,3.9)
city average
Central sites: 0.71 warm 7.4 (4.5, 10)
city average 0.55 cold
Copollutant
Model
With UFP:
8.4 (-7.2, 27)
With benzene:
1.7 (-0.26, 3.6)
With
ethylbenzene:
1.6 (-0.49, 3.7)
With benzene:
3.6 (-31, 29)
With
ethylbenzene:
17 (-17, 41)
With CO:
1.5(0, 3.0)
With CO:
2.2(0.78,3.7)
With CO:
4.2(0.78,3.7)
avg = average; a.m. = ante meridiem; ATSDR = Agency for Toxic Substances and Disease Registry; BC = black carbon;
BS = black smoke; CO = carbon monoxide; EC = elemental carbon; eNO = exhaled breath condensate; EEC = exhaled breath
condensate; ED = emergency department; FEF2s-75% = forced expiratory flow between 25 and 75% of forced vital capacity;
FENA, = forced expiratory flow in 1 second; FVC = forced vital capacity; NR = not reported; PEF = peak expiratory flow;
PM2.5 = particles with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm, PNC = particle number concentration;
TWA = time-weighted average; UFP = ultrafine particles.
aSingle- and copollutant model results are standardized to a 20-ppb increase in 24-h avg NO2 and 30-ppb increase in 1-h max
NO2. Results based on other averaging times are not standardized but presented as reported in their respective studies
(Section 5.1.2.2). Percentage change in FEF25-75o/0 FEV-i, FVC, PEF, and EEC pH refers to percentage decrease.
To fit results in Figure 5-16. results are multiplied by 10.
°Copollutant specifically is reactive oxygen species generated from ambient PM25 extracts.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-252
-------
Table 5-39 Summary of evidence for a causal relationship between short-term
nitrogen dioxide exposure and respiratory effects.
Rationale for Causal
Determination3
Key Evidence13
Key References'3
NO2 Concentrations
Associated with
Effects0
Asthma Exacerbation
Consistent evidence
from multiple,
high-quality controlled
human exposure studies
Rules out chance,
confounding, and other
biases with reasonable
confidence
NO2 increases airway
responsiveness in adults with asthma
exposed at rest following nonspecific
or allergen challenge in several
individual studies and meta-analyses.
Clinical relevance supported by
findings of a doubling reduction in
provocative dose in response to NO2.
Folinsbee (1992).
tBrown(2015),
tGoodman et al. (2009)
Section 5.2.2.1,
Table 5-1
tBrown(2015)
100 ppb for 1 h
200-300 ppb for 30 min
100 ppb for 1 h,
140 ppb for 30 min
Epidemiologic evidence NO2 associations with lung function Delfino et al. (2006), Personal outdoor mean
from multiple,
high-quality studies
provides some support
for independent NO2
association
and pulmonary inflammation persist
in copollutant models with PlVh.s or a
traffic-related copollutant—EC/BC,
OC, UFP, or VOC in studies with
exposure assessment in subjects'
locations.
Some studies show weak-moderate
correlations for personal ambient and
total NO2 with traffic-related
copollutants (r = -0.42 to 0.49).
tDelfino et al. (2008a),
tMartins et al. (2012),
McCreanoret al. (2007)
Figures 5-16 and 5-17,
Table 5-38
2-h avg at near-road
site: 75.5 ppb
Total personal mean
24-h avg: 28.6 ppb
Most central site NO2 associations
persist with adjustment for PlVh.s,
EC/metals factor, UFP, or CO.
Potential for differential exposure
measurement error limits inference
from copollutant models with central
site measurements.
Asthma hospital
admissions, ED visits:
tVilleneuve et al.
(2007). tJalaludin etal.
(2008), Ito et al. (2007),
tlskandaretal. (2012).
ATSDR (2006)
Symptoms, medication
use: tGent et al. (2009),
von Klot et al. (2002)
Overall study mean
24-h avg:
11.3-31.3 ppb
Overall study mean
1-h max: 23-44 ppb
Some associations were attenuated
with adjustment for PlVh.s or UFP.
tLiu et al. (2009b), von
Klot et al. (2002)
Indoor NO2 associated with
increases in asthma-related effects in
children.
tSarnatetal. (2012).
tLuetal. (2013).
tHansel et al. (2008)
No association in
tGreenwald et al.
(2013)
Means of 3-to
7-day avg:
18.7-121 ppb
75th: 31 ppb
Max: 394 ppb
5-253
-------
Table 5-39 (Continued): Summary of evidence for a causal relationship between
short-term nitrogen dioxide exposure and respiratory
effects.
Rationale for Causal
Determination3
Key Evidence13
Key References13
NO2 Concentrations
Associated with
Effects0
(Continued)
(Epidemiologic evidence
from multiple,
high-quality studies
provides some support
for independent NO2
association)
Most associations for
microenvironmental and central site
NO2 persist in copollutant models
with PMio, SO2, or O3.
NO2 associations persist with
adjustment for meteorology, time
trends, season, medication use.
Supplemental
Figure S5-1
(U.S. EPA. 2015a)
(Continued)
Consistent
epidemiologic evidence
for NO2 metrics with
lower potential for
exposure measurement
error
Asthma-related effects associated
with NO2 measured in subjects'
locations (personal total and ambient,
school outdoor) or combined across
central sites based on population
density.
Better alignment with subjects'
locations compared to central site
NO2.
tGreenwald et al.
(2013), tHolquinetal.
(2007), tDelfino et al.
(2008a). McCreanor et
al. (2007). tSarnat et al.
(2012),
tZoraetal. (2013),
Delfino et al. (2006),
tStrickland et al. (2010)
No association:
tSmarqiassi et al.
(2014), tSpira-Cohen et
al. (2011)
Personal outdoor mean
2-h avg: 75.5 ppb
Total personal mean
24-h avg: 28.6 ppb
Outdoor school mean
1-week avg:
3.4-18.2 ppb
Central site
population-weighted
mean 1-h max: 23.3
ppb
Consistent
epidemiologic evidence
from other studies with
more uncertainty
regarding confounding
and exposure
measurement error
Increases in asthma hospital
admissions, ED visits in diverse
populations in association with
24-h avg and 1-h max NO2, lags 0
and 3 to 5-day avg among all ages
and children.
No association in recent Canadian
multicity study.
Ito et al. (2007),
tlskandaretal. (2012).
ATSDR (2006)
Section 5.2.2.4.
Figure 5-7
tStieb et al. (2009)
Mean 24-h avg:
21.4-41.2 ppb
Coherence with increases in
respiratory symptoms and
decrements in lung function in
populations with asthma in
association with 24-h avg, 2-4 h avg
NO2, 1-h max, lags 0, 3 to 6-day avg.
Panel studies of children examined
representative populations recruited
from schools. No reports of selective
participation by particular groups.
Sections 5.2.2.2 and
5.2.2.3, Figures 5-3 and
5-4
City mean 24-h avg:
17.8-26 ppb
Evidence for key events
in proposed mode of
action
Allergic responses
Increases in eosinophil activation,
IgE, Th2 cytokines in adults with
asthma.
Barck et al. (2005a),
Barck et al. (2002),
Wanqetal. (1995a),
tEzrattv et al. (2014)
Sections 4.3.2.6 and
5.2.2.5
Figure 4-1
Humans: 260 ppb
15-30 min, 400 ppb
6 h, 581 ppb for 30 min,
2 days
5-254
-------
Table 5-39 (Continued): Summary of evidence for a causal relationship between
short-term nitrogen dioxide exposure and respiratory
effects.
Rationale for Causal
Determination3
Key Evidence13
Key References13
NO2 Concentrations
Associated with
Effects0
Evidence for key events
in proposed mode of
action
Inflammation
Increases in PMNs and
prostaglandins in healthy adults.
Section 5.2.7.4
1,500-3,500 ppb 20
min or 3-4 h
Increases in eNO in children with
asthma in association with 24-h avg
NO2.
Delfino et al. (2006),
tSarnatetal. (2012),
tMartins et al. (2012)
Section 5.2.7.4
Total personal mean
24-h avg: 24.3,
30.9 ppb
Ambient mean
1-week avg:
4.5-20 ppb
Inconsistent effects on
oxidative stress,
pulmonary injury
See Respiratory Effects in Healthy
Individuals below.
COPD Exacerbation
Inconsistent
epidemiologic evidence
and uncertainty
regarding NO2
independent effects
Increases in COPD hospital
admissions and ED visits.
Inconsistent associations with lung
function decrements and symptoms
in adults with COPD.
tFaustini et al. (2013),
tKo et al. (2007b),
tArbex et al. (2009)
Section 5.2.4.2
Section 5.2.4.1
Mean 24-h avg:
24.1-63.0 ppb
Mean 1-h max:
63.0 ppb
Inconsistent evidence
from controlled human
exposure studies
Lack of evidence to
propose a mode of
action
Lung function decrements not
consistently found in adults with
COPD.
Increased inflammation in healthy
adults but not in adults with COPD.
Morrow et al. (1992),
Vaqaqqini et al. (1996)
Section 5.2.4.1
tBruskeetal. (2010)
Sections 5.2.4.3 and
5.2.7.4
300 ppb for 1 h, 4 h
1,500-3,500 ppb
20 min or 3-4 h
Respiratory Infection
Consistent animal
toxicological evidence
with relevant NO2
exposures
Mortality from bacterial or viral
infection in animals with NO2
exposures of 1,500 ppb and higher,
not lower concentrations.
Ehrlich etal. (1977),
Ehrlich et al. (1979).
Ehrlich (1980),
Graham etal. (1987)
Section 5.2.5.1
1,500-5,000 ppb for
3h
1,500 ppb with
4,500 ppb spike of
1-7.5 h
Inconsistent
epidemiologic evidence
and uncertainty
regarding NO2
independent effects
Associations with hospital
admissions and ED visits for
respiratory infections. All results
based on central site NO2, and some
have wide 95% CIs.
Inconsistent evidence for parental
reports of infection or laboratory-
confirmed infections.
tZemeketal. (2010),
tMehta etal. (2013).
tStieb et al. (2009),
tFaustini et al. (2013),
Just etal. (2002),
tSternetal. (2013)
Sections 5.2.5.3 and
5.2.5.2.
Overall study mean
24-h avg:
11.7-28.6 ppb
City mean 24-h avg:
9.3-34.6 ppb
5-255
-------
Table 5-39 (Continued): Summary of evidence for a causal relationship between
short-term nitrogen dioxide exposure and respiratory
effects.
Rationale for Causal
Determination3
Key Evidence13
Key References13
NO2 Concentrations
Associated with
Effects0
Limited evidence for key
events in proposed
mode of action
Decreased AM function indicated by
diminished superoxide production.
No consistent effect on pulmonary
clearance.
Section 5.2.5.4
Respiratory Effects in Healthy Individuals
Limited epidemiologic
evidence and
uncertainty regarding
NO2 independent effects
Consistent evidence for respiratory
symptoms in children. All based on
central site NO2 and no examination
of confounding by traffic-related
copollutants.
Schwartz et al. (1994)
Section 5.2.7.3,
Table 5-33
Mean 24-h avg: 13 ppb
Lung function not consistently
associated with NO2 measured at
subjects' locations or central site NO2
correlated (r = 0.63) with total
personal. But, personal ambient NO2
associations persist with adjustment
forPM2.s, EC, OC, copper, iron, or
UFP.
tStraketal. (2012),
Moshammer et al.
(2006), Linn etal.
(1996)
Section 5.2.7.2,
Table 5-30
Max for 5-h avg: 96 ppb
Max for 24-h avg:
96 ppb
75th for 24-h avg:
11.4 ppb
Limited and inconsistent
evidence from controlled
human exposure studies
Increases in airway responsiveness
found in healthy adults above
1,000 ppb NO2, not lower
concentrations.
Folinsbee (1992),
Kiaerqaard and
Rasmussen (1996)
Section 5.2.7.1
1,000-2,000 ppb for 3
h
Respiratory symptoms or lung
function examined in adults; changes
generally not found.
Sections 5.2.7.2 and
5.2.7.3
200-4,000 ppb for
2-5 h
Limited evidence for key Increases in PMNs and
events in proposed prostaglandins in healthy adults.
mode of action
Inflammation
Frampton et al. (2002).
Frampton etal. (1989)
Section 5.2.7.4
1,500-3,500 ppb for 3
h
Limited epidemiologic evidence for
associations of NO2 measured in
subjects' locations with increases in
pulmonary inflammation in children
and adults. Associations persist with
adjustment for BC/EC, OC, UFP, or
PM2.5.
tStraketal. (2012).
tSteenhofetal. (2013),
tLinetal. (2011)
Section 5.2.7.4.
Table 5-35
Mean 24-h avg: 9.3,
33 ppb
Mean 5-h avg across
sites with varying
traffic: 20 ppb
5-256
-------
Table 5-39 (Continued): Summary of evidence for a causal relationship between
short-term nitrogen dioxide exposure and respiratory
effects.
Rationale for Causal
Determination3
Key Evidence13
Key References13
NO2 Concentrations
Associated with
Effects0
Inconsistent effects on
oxidative stress,
pulmonary injury
Inconsistent changes in antioxidants Sections 4.3.2.3 and
in experimental studies but found in 5.2.7.4
humans and rodents with lower
dietary antioxidant vitamins.
Humans: 2,000 ppb for
4 h, 1 day or 4 days
Rodents:
1,000-5,000 ppb for
3-7 days
Increases in LDH, CC16, BAL fluid
protein inconsistently found in
humans, rodents. Limited evidence
for impaired epithelial barrier
function.
Sections 4.3.2.4 and
5.2.7.4
Humans:
600-2,000 ppb for 3-4
h, 1-4 days
Animal models:
400-2,000 ppb for
1-3 weeks
Respiratory Mortality
Consistent
epidemiologic evidence
but uncertainty
regarding NO2
independent effect
Multicity studies consistently observe
associations of respiratory mortality
with 24-h avg NO2 at lag 0-1 days.
Results based on NO2 averaged
across central sites.
Potential confounding by traffic-
related copollutants not assessed.
NO2 results robust to adjustment for
PM-io, SO2, orOs.
tWonq et al. (2008),
tChenetal. (2012b),
tChiusolo et al. (2011),
tBellinietal. (2007).
Bigger! et al. (2005)
Section 5.2.8
Means across cities for
24-h avg:
13.5-55.5 ppb
Uncertainty due to
limited coherence with
respiratory morbidity
evidence
Evidence for asthma exacerbation in
adults but limited coherence among
lines of evidence for effects on
COPD and respiratory infection.
Uncertainty regarding spectrum of
effects that can lead to respiratory
mortality.
AM = alveolar macrophage; ATSDR = Agency for Toxic Substances and Disease Registry; avg = average; BAL =
bronchoalveolar lavage; BC = black carbon; CC16 = club cell protein; Cl = confidence interval; CO = carbon monoxide;
COPD = chronic obstructive pulmonary disease; EC = elemental carbon; ED = emergency department; eNO = exhaled nitric
oxide; IgE = immunoglobulin E; LDH = lactate dehydrogenase; max = maximum; NO2 = nitrogen dioxide; O3 = ozone;
OC = organic carbon; PM25 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm;
PMio = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; PMN = polymorphonuclear
cells; ppb = parts per billion; SO2 = sulfur dioxide; Th2 = T-derived lymphocyte helper 2; UFP = ultrafine particles; VOC = volatile
organic compound.
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Tables I and N. of the
Preamble.
Describes the key evidence and references, supporting or contradicting, contributing most heavily to causal determination and,
where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where full body of evidence is
characterized.
°Describes the NO2 concentrations with which the evidence is substantiated (for experimental studies, £ 5,000 ppb).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-257
-------
5.3 Cardiovascular Effects
5.3.1 Introduction
The 2008 ISA for Oxides of Nitrogen concluded that the "available evidence on the
effects of short-term exposure to NO2 on cardiovascular health effects was inadequate to
infer the presence or absence of a causal relationship" (U.S. EPA. 2008c). Multiple
studies found associations between short-term exposure to NO2 and rates of hospital
admission or ED visits for cardiovascular diseases (CVDs), yet it was unclear at that time
whether these results supported a direct effect of short-term NO2 exposure on
cardiovascular morbidity or were confounded by other correlated pollutants.
Additionally, epidemiologic studies available at the time of the last review provided
inconsistent evidence for associations between short-term NO2 exposure and other
cardiovascular events such as arrhythmia among patients with implanted cardioverter
defibrillators and subclinical measures associated with cardiovascular events, such as
heart rate variability (HRV) and electrocardiographic (ECG) markers of cardiac
repolarization. Experimental studies available at the time of the 2008 ISA for Oxides of
Nitrogen did not provide biological plausibility for the cardiovascular effects observed in
epidemiologic studies. There was limited evidence from controlled human exposure
studies demonstrating a reduction in hemoglobin and some evidence from toxicological
studies for effects of NO2 on various hematological parameters in animals, but these
studies were limited and inconsistent. Overall, the experimental studies could not address
the uncertainty related to copollutant confounding in epidemiologic studies of hospital
admission or ED visits for CVDs in the 2008 ISA for Oxides of Nitrogen.
The following sections review the published studies pertaining to the cardiovascular
effects of short-term exposure to oxides of nitrogen in humans, animals, and cells. When
compared to the 2008 ISA for Oxides of Nitrogen, the recent epidemiologic and
toxicological studies provide evidence for effects of NO2 exposure on a broader array of
cardiovascular effects and mortality. Still, substantial uncertainties remain concerning
potential confounding by other traffic-related pollutants, exposure measurement error,
and the limited mechanistic evidence to describe a role for NO2 in the manifestation of
cardiovascular diseases, including key events in the proposed mode of action. The
majority of the recent evidence is from epidemiologic studies, which suggest that
exposure to NO2 may result in the triggering of MI. To clearly characterize the evidence
underlying causality, the discussion of the evidence is organized into groups of related
5-258
-------
outcomes [e.g., MI including ischemic heart disease (IHD), arrhythmia, and cardiac
arrest]. Evidence for subclinical effects (e.g., HRV, blood biomarkers of cardiovascular
effects) that potentially underlie the development, progression, or indication of various
clinical events is discussed in Section 5.3.10. and may provide biological plausibility for
multiple outcomes.
5.3.2 Myocardial Infarction
Several lines of evidence are evaluated to assess the relationship between short-term NO2
exposure and triggering of an MI: hospital admissions and ED visits for MI, IHD, or
angina as well as ST-segment amplitude changes. An MI or heart attack occurs as a
consequence of IHD, resulting in insufficient blood flow to the heart that overwhelms
myocardial repair mechanisms and leads to muscle tissue death. In addition, IHD
includes the diagnosis of angina. Symptoms of angina are similar to those of MI;
however, where MI results in damage to the heart muscle, angina does not result in
myocardial necrosis. As angina may indicate an increased risk for future MI, studies of
angina are part of the evaluation of a relationship between short-term NCh exposure and
MI. Finally, acute MI may be characterized by deviations in ST segment amplitude,
which may serve as a nonspecific marker of myocardial ischemia.
5.3.2.1 Hospital Admissions and Emergency Department Visits for Myocardial
Infarction and Ischemic Heart Disease
The 2008 ISA for Oxides of Nitrogen concluded that the epidemiologic evidence
consistently supported the associations between short-term increases in ambient NO2
concentrations and hospital admissions or ED visits for cardiac diseases (U.S. EPA.
2008c). This conclusion continues to be supported by studies published since the 2008
ISA, as reviewed below (Figure 5-18 and Table 5-40). However, potential copollutant
confounding, especially from other traffic-related pollutants (e.g., EC, CO), and limited
mechanistic evidence are still key uncertainties, and make it difficult to interpret the
results of these studies. Additionally, all of the studies in this section use central site
monitors to estimate ambient NO2 exposure, which may result in misclassification of the
exposure due to the high variability in NO2 (Section 3.4.5.1).
A number of studies rely on clinical registries, which are generally less susceptible to
misclassification of the outcome and exposure. The strongest evidence of an association
between ambient NO2 and the risk of MI comes from a study using clinical registry data
from the U.K.'s Myocardial Ischaemia National Audit Project (Bhaskaran et al.. 2011).
5-259
-------
which found a 5.8% (95% CI: 1.7, 10.6) increase in risk of MI per 30-ppb increase in
1-h max NO2 concentrations in the 6 hours preceding the event. This study is unique
because it included detailed data on the timing of MI onset in more than 79,000 patients
from 15 conurbations in England and Wales, which allowed examination of association
with ambient NC>2 in the hours preceding MI. NC>2 results were robust to a number of
sensitivity analyses that evaluated key aspects of study design and model specification
(e.g., stricter diagnosis criteria, different time strata). Additionally, Bhaskaran et al.
(2011) restricted analyses to urban areas to reduce heterogeneity that may have resulted
in measurement bias from the use of fixed site monitors to assess NO2 exposure. The
findings for NC>2 were more pronounced in those aged between 60 and 80 years, among
those with prior coronary heart disease, and for events occurring in the autumn and
spring. Conversely, in a smaller study of only 429 MI events, Turin et al. (2012) did not
observe a consistent positive association using data from the Takashima County Stroke
and AMI Registry in Central Japan. Cases were cross-checked by research physicians,
epidemiologists, and cardiologists, thereby minimizing potential misclassification of the
outcome.
5-260
-------
Study
Thach et al. (2010)
Qiuet al. (2013a)
Hsiehetal. (2010)
Tsai et al. (2012)
Belief al. (2008)
Goggins et al. (2013)
Szyszkowicz (2009)
Szyszkowicz (2007)
Stieb et al (2009)
vencloviene et al. (2011)
.
Nuvolone et al. (2011)
Wichmannetal.(2012)
Larrieuetal. (2007)
Mann et al. (2002)
Polonieckiet al. (1997)
Linn et al. (2000)
Wongetal. (1999)
Ponka and V.rtanen (1996)
Barnett et al. (2006)
Bhaskaran et al. (2011)
Outcome
IHD
IHD
Ml
Ml
IHD
Ml
Angina
IHD
Ml
Ml
Ml
IHD
IHD
Ml
IHD
Ml
Angina
Ml
IHD
IHD
Angina
Ml
Mean
Concentration
31.2
30.85
29.88
27.59
26.4
23.5-29.0
20.1
19.4
18.4
18.4
15.2-21.1
11.9-24.9
37.2
35
28-41
27.3
72.8
15.7-23.2
16.5
Lag Day
0-1 avg
0-3 avg
0-3 avg
0-3 avg
0-3 avg
0-3 avg
0-2 avg
0-2 avg
0-2 avg
0-2 avg
0
1
0-3 avg
0
1
0-3 avg
0
1
0-3 avg
0
0
0
0
1
0
1
0-1 avg
0-1 avg
0
1
1
0-1 avg
0
1
0-1 avg
0-1 avg
0
1
0-1 avg
0
1
0-1 avg
0
0-1 avg
1
1
1
0
0-1 avg
1
0
1
0-1 avg
0-1 avg
0-1 avg
0
1-6 h
7-12 h
13-18 h
19-24 h
Notes
whole year
warm season
cool season
23+ C
<23 C
23+ C; hypertension
23+ C; no hypertension
<23 C; no hypertension
all 13 monitors
5 city monitors
8 correlated monitors
Hong Kong
Taipei, Taiwain
Kaohsiung, Taiwan
<65 yr
yr
Cold
sARR
sCHF
No secondary disease
NO
£ 65 yr
15-64 yr
15-64 yr
' 24-h avg
1*
1 ~°~
9
1 •
i j
1 *
1 ,•
9
1 9
~l — •
|0
-o-
1 •
1°
1 — • —
|O
1 .
— •"•
•l
( 1 *
1 •
1 ,
k
-J 9
1 •
1 -•-
1.
1 ^
•r
!•
I*
\
p
i»
r*~
A
* — A —
1 ,
-|~"
I \
1
* J l-h max
-4-
0-2 days avg — •=-
Atkinson et al. (1999)
Peel et al. (2007)
Jalaludinet al. (2006)
Simpson et al. (2005a)
IHD
IHD
IHD
IHD
50.3
45.9
23.2
16.3-23.7
0
0
0-2 avg
0-2 avg
0
1
0-1 avg
0-64 yr
65+ yr
case-crossover
time series
£65yr
&
§-•-
1*
^
~l* —
•^
O.75
1 1.25 1.5 1.75 2
Relative Risk and 95% Cl
Note: Cl = confidence interval; h = hours; IHD = ischemic heart disease; Ml = myocardial infarction; NO = nitric oxide;
NO2 = nitrogen dioxide; sARR = secondary arrhythmia; sCHF = secondary congestive heart failure; yr = years. Black = studies from
the 2008 Integrated Science Assessment for Oxides of Nitrogen, red = recent studies. Circles = NO2, triangles = NO. Relative risks
are standardized to a 20-ppb or 30-ppb increase in NO2 or NO concentration for 24-h average and 1-h maximum metrics,
respectively. Studies are organized first by averaging time, then by recent study or previous study, then in descending order of
mean NO2 concentration (in parts per billion).
Figure 5-18 Associations between short-term exposure to oxides of nitrogen
and hospital admissions for ischemic heart disease.
5-261
-------
Table 5-40 Corresponding risk estimates for hospital admissions for ischemic
heart disease for studies presented in Figure 5-18.
Study
Location
Health
Effect
Relative Risk3 (95% Cl) Copollutant Examination13
tThach et al.
(2010)
Hong Kong, China IHD
Lag 0-1 avg:
1.04(1.02, 1.05)
No copollutant models.
tQiu et al.
(2013a)
Hong Kong, China IHD
Lag 0-3 avg
All year:
1.09(1.08, 1.11)
Warm season:
1.05(1.03, 1.08)
Cool season:
1.15(1.12, 1.18)
No copollutant models.
NO2 and PM-io correlation:
Pearson r= 0.76.
tHsieh etal.
(2010)
Taipei, Taiwan
Ml
Lag 0-3 avg
>23°C:
1.24(1.16, 1.35)
<23°C:
1.26(1.18, 1.35)
NO2: robust to PM-io, SO2, CO,
or Os inclusion in copollutant
models.
Copollutants: all but Os
attenuated by NO2 adjustment.
NO2 correlations (Pearson r):
PMio: 0.55; SO2: 0.51; CO:
0.71; Os: 0.02.
tTsai et al.
(2012)
Kaohsiung, Taiwan Ml
Lag 0-2 avg
Hypertension
>23°C:
1.24(1.12, 1.38)
<23°C:
1.29(1.16, 1.44)
No hypertension
>23°C:
1.29(1.18, 1.40)
<23°C:
1.24(1.14, 1.35)
No copollutant models.
NO2 correlations (Pearson r):
PMio: 0.48; SO2: 0.45; CO:
0.77; Os: -0.01.
5-262
-------
Table 5-40 (Continued): Corresponding risk estimates for hospital admissions for
ischemic heart disease for studies presented in
Figure 5-18.
Study
Location
Health
Effect
Relative Risk3 (95% Cl) Copollutant Examination13
tChenq et al.
(2009a)
Kaohsiung, Taiwan
Ml
Lag 0-2 avg
>25°C:
1.23(1.06, 1.44)
<25°C:
1.76(1.55,2.02)
NO2: attenuated by CO or Os
adjustment on warm days and
PM-io on cool days. Robust to
SO2 adjustment.
Copollutants: all but CO and O
on warm days attenuated by
NO2 adjustment.
NO2 correlations (Pearson r):
PMio: 0.73; SO2: 0.53; CO:
0.66; O3: 0.09.
tBelletal. Taipei, Taiwan IHD
(2008)
tGoqqins et al. Hong Kong, China; Ml
(2013) Taipei, Taiwan;
Kaohsiung, Taiwan
tSzyszkowicz 6 Canadian cities Angina
(2009)
tSzyszkowicz Montreal, QC, Canada IHD
(2007)
All 13 monitors
Lag 0:
1.10(1.02, 1.18)
Lag1:
1.05(0.98, 1.13)
Lag 0-3 avg:
1.03(0.98, 1.21)
5 city monitors
Lag 0:
1.09(1.02, 1.16)
Lag1:
1.05(0.98, 1.12)
Lag 0-3 avg:
1.08(0.99, 1.20)
8 correlated monitors
Lag 0:
1.09(1.02, 1.17)
Lag1:
1.05(0.98, 1.12)
Lag 0-3 avg:
1.08(0.97, 1.20)
LagO
Hong Kong:
1.04(1.02, 1.07)
Taipei:
1.09(1.05, 1.13)
Kaohsiung:
1.05(0.98, 1.13)
LagO:
1.04(1.03, 1.05)
Lag1:
1.13(1.04, 1.22)
No copollutant models.
No copollutant models.
No copollutant models.
No copollutant models.
5-263
-------
Table 5-40 (Continued): Corresponding risk estimates for hospital admissions for
ischemic heart disease for studies presented in
Figure 5-18.
Health
Study Location Effect
fStieb et al. 7 Canadian cities Ml
(2009)
fVencloviene et Kaunas, Lithuania Ml
al. (2011)
tTurin et al. Takashima County, Ml
(2012) Japan
tNuvolone et al. Tuscany, Italy Ml
(2011)
Relative Risk3 (95% Cl) Copollutant Examination13
Lag 0:
1.03(1.00, 1
Lag1:
1.03(1.00, 1
Lag 0-1 avg
<65 yr:
1.19(0.96, 1
>65 yr:
0.97(0.81, 1
Lag 0:
1.14(0.92, 1
Lag1:
0.87(0.70, 1
Lag 0:
1.04(0.97, 1
Lag1:
1.08(1.00, 1
.05)
.06)
.48)
.17)
.40)
.08)
.12)
.15)
NO2 association attenuated by
CO.
No results provided for other
pollutants.
NO2: robust to TSP or SO2
adjustment. Attenuated by Os
adjustment.
Copollutants: TSP and SO2
attenuated by NO2 adjustment.
No associations between Os
and Ml regardless of NO2
adjustment.
NO2: robust to PM-io
adjustment. Attenuated by CO
adjustment.
Copollutants: PM-io not
associated with Ml after NO2
adjustment. No association
between CO and Ml regardless
of NO2 adjustment.
No correlations reported.
tWichmann et al. Copenhagen,
(2012) Denmark
Ml
Warm season
Lag 0:
0.99(0.86, 1.14)
Lag1:
1.10(0.96, 1.26)
No copollutant models.
Lag 0-1 avg:
1.08(0.932, 1.27)
Cool season
LagO:
1.01 (0.91, 1.11)
Lag 1:
1.08(0.98, 1.19)
Lag 0-1 avg:
1.07(0.95, 1.20)
tLarrieu et al.
(2007)
8 French cities
IHD Lag 0-1 avg
1.07(1.03, 1.10)
No copollutant models.
5-264
-------
Table 5-40 (Continued): Corresponding risk estimates for hospital admissions for
ischemic heart disease for studies presented in
Figure 5-18.
Study
Location
Health
Effect
Relative Risk3 (95% Cl) Copollutant Examination13
tRich et al.
(2010)
New Jersey
Lag 0
Transmural:
1.14(0.96, 1.32)
NO2: slightly attenuated by
PM2.5 adjustment.
Copollutants: PIVhs association
attenuated by adjustment for
NO2.
NO2 and PIVhs correlation:
r=0.44.
Mannetal. Los Anqeles, CA IHD
(2002)
With sARR
LagO:
1.04(1.02, 1.06)
Lag 1:
1.02(1.00, 1.04)
Lag 0-1 avg:
1.03(1.01, 1.05)
With sCHF
LagO:
1.05(1.01, 1.08)
Lag 1:
1.04(1.01, 1.08)
Lag 0-1 avg:
1.05(1.02, 1.09)
No secondary disease
LagO:
1.03(1.01, 1.04)
Lag 1:
1.03(1.01, 1.04)
Lag 0-1 avg:
1.03(1.02, 1.05)
No copollutant models.
Ml Lag 0
1.04(1.02, 1.06)
Poloniecki et al. London, U.K. IHD,
(199Z) Ml,
angina
Linn et al. (2000) Los Anqeles, CA Ml
Wonq et al. Honq Konq, China IHD
(1999)
Lag1:
1.00(0.98, 1.02)
Lag 1:
1.02(1.01, 1.03)
Lag1:
1.01 (1.00, 1.03)
Lag 0:
1.02(1.00, 1.04)
Lag 0-1 :
1.04(1.00, 1.08)
NO2: robust to O; Attenuated
by CO, SO2, and BS.
No copollutant models.
No copollutant models.
5-265
-------
Table 5-40 (Continued): Corresponding risk estimates for hospital admissions for
ischemic heart disease for studies presented in
Figure 5-18.
Study
Location
Health
Effect
Relative Risk3 (95% Cl) Copollutant Examination13
Ponka and Helsinki, Finland
Virtanen(1996)
IHD NO2
LagO:
1.17(0.96, 1.42)
Lag 1:
0.95(0.77, 1.16)
NO
LagO:
1.01 (0.96, 1.05)
Lag 1:
1.10(1.05, 1.15)
No copollutant models.
Barnett et al.
(2006)
7 Australian and New
Zealand cities
IHD Lag 0-1 avg
>65 yr:
1.10(1.04, 1.17)
15-64yr:
1.03(0.96, 1.10)
No copollutant models.
Ml
Lag 0-1 avg
>65 yr:
1.18(1.04, 1.35)
15-64yr:
1.07(0.96, 1.20)
von Klot et al.
(2005)
5 European cities
Ml Lag 0 NCb: robust to PM-io or
Ages >35 yr, Ml survivors: adjustment.
1.14(0.99, 1.32)
Angina Lag 0:
1.16(1.03, 1.31)
tBhaskaran et
al. (2011)
England and Wales,
U.K.
Lag 1-6h:
1.06(1.02, 1.11)
Lag 7-12 h:
0.95(0.90,0.99)
Lag 13-18 h:
0.99(0.94, 1.05)
Lag 19-24h:
1.00(0.96, 1.05)
Lag 0-2 day avg:
0.98(0.93, 1.02)
No copollutant models.
NO2 correlations: PMm 0.48;
O3: -0.58; CO: 0.61; SO2: 0.31.
Atkinson et al.
(1999b)
London, U.K.
IHD Lag 0
0-64 yr:
1.01 (0.99, 1.04)
65+ yr:
1.03(1.01, 1.04)
No copollutant models
analyzed for IHD.
5-266
-------
Table 5-40 (Continued): Corresponding risk estimates for hospital admissions for
ischemic heart disease for studies presented in
Figure 5-18.
Study Location
Peel et al. (2007) Atlanta, GA
Jalaludin et al. Sydney, Australia
(2006)
Simpson et al. 4 Australian cities
(2005a)
Health
Effect Relative Risk3 (95% Cl)
IHD Lag 0-2 avg
Case crossover:
1.04(1.00, 1.07)
Time series:
1.04(1.01, 1.08)
IHD Ages >65 yr
LagO:
1.07(1.01, 1.13)
Lag 1:
1.02(0.97, 1.08)
Lag 0-1 avg:
1.01 (0.99, 1.03)
IHD Lag 0-1 avg:
1.06(1.03, 1.09)
Copollutant Examination13
No copollutant models.
No copollutant models
analyzed for IHD.
No copollutant models
analyzed for IHD.
avg = average; BS = black smoke; CA = California; Cl = confidence interval; CO = carbon monoxide; GA = Georgia;
IHD = ischemic heart disease; Ml = myocardial infarction; NO = nitric oxide; NO2 = nitrogen dioxide; O3 = ozone;
PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PMio = particulate matter with
a nominal mean aerodynamic diameter less than or equal to 10 |jm; QC = Quebec; sARR = secondary arrhythmia;
sCHF = secondary congestive heart failure; SO2 = sulfur dioxide, TSP = total suspended particles; U.K. = United Kingdom.
aEffect estimates are standardized to a 20-ppb or 30-ppb increase in NO2 or NO for 24-h avg and 1-h max metrics, respectively.
""Relevant relative risks for copollutant models can be found in Supplemental Figures S5-2, S5-3, S5-4, and S5-5 (U.S. EPA.
2015b. c, d, e).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
A number of studies based on administrative data have also been published since the
2008 ISA for Oxides of Nitrogen. In six areas in central Italy, Nuvolone et al. (2011)
found an 8% (95% Cl: 0, 15) increase in risk of hospital admission for MI per 20-ppb
increase in 24-h avg NCh on the previous day. Similar associations were seen in relation
to lags 2 to 4 days prior to hospital admission. The association at lag 2 was robust to
adjustment for PMio in a copollutant model, and remained positive, though somewhat
attenuated, by adjustment for CO [Supplemental Figures S5-2, (U.S. EPA. 2015b) and
S5-3, (U.S. EPA. 2015c)1. The association with NO2 was somewhat more pronounced
among females and in the cold season. Using data from 14 hospitals in seven Canadian
cities, Stieb et al. (2009) found a 2.8% (95% Cl: 0.2, 5.4) increase in risk of ED visits for
the composite endpoint of acute MI or angina per 20-ppb increase in 24-h avg NO2 on the
same day. However, the overall association was heavily influenced by the association
observed in Edmonton, and exclusion of the data from Edmonton from the analysis
attenuated the results. Furthermore, the association observed from the data including
Edmonton was weakened in magnitude and precision (wider 95% Cl) in a copollutant
5-267
-------
model adjusting for CO (1.3% [95% CI: -2.9, 5.6] increase per 20 ppb increase in 24 h
avg NO2 on the same day). Larrieu et al. (2007) observed a positive association between
hospital admissions for IHD and NC>2 concentrations in eight French cities. The
magnitude of the association was higher for older adults (i.e., >65 years) than for the
general population. In large single-city studies, Szyszkowicz (2007). Thach et al. (2010).
and Franck etal. (2014) found that NCh was associated with increased risk of hospital
admission for IHD in Montreal, Canada; Hong Kong, China; and Santiago, Chile,
respectively. Qiu et al. (2013a) also reported an overall association between NO2
concentrations risk of ED visits for IHD in Hong Kong, China that was stronger in the
cool season and on low humidity days.
In New Jersey, Rich etal. (2010) found a relative risk of 1.14 (95% CI: 0.96, 1.32) per
20-ppb increase in 24-h avg NC>2 for hospitalization for transmural Mis, but that
association was attenuated by adjustment for PM25 in a copollutant model [1.05 (95% CI:
0.85, 1.28)]. No results were reported for all Mis or for nontransmural infarcts. NC>2 was
positively associated with hospital admissions for MI in Taipei, Taiwan (Goggins et al..
2013; Tsai etal.. 2012; Hsieh etal.. 2010); Kaohsiung, Taiwan (Tsai etal.. 2012; Cheng
et al.. 2009a): and Hong Kong, China (Goggins et al.. 2013). The associations reported by
Hsieh etal. (2010) remained relatively unchanged after adjustment for PMio, 862, CO, or
Os in copollutant models, as did the results from Cheng et al. (2009a). with the exception
of CO and Os on warm days. NO2 was also positively associated with hospital admissions
for IHD in Taipei, Taiwan (Bell et al.. 2008). In an effort to reduce uncertainty related to
the use of central site monitors, Bell et al. (2008) estimated NO2 exposure over the entire
Taipei area (average of 13 monitors), within Taipei City only (average of 5 monitors),
and using a subset of monitors where all pairs of monitors had NO2 correlations greater
than 0.75 (8 monitors). The authors reported consistent results across the multiple
exposure metrics, with the exception of stronger associations observed using the city or
correlated monitors at lag 0-3 avg (Table 5-40). Wichmann et al. (2012) found that NO2
was positively associated with risk of acute MI hospital admissions in Copenhagen,
Denmark, but only in the warm months of the year. NO2 was not associated with risk of
hospital admission for acute coronary syndrome in Lithuania (Vencloviene et al.. 2011).
5.3.2.2 Hospital Admissions and Emergency Department Visits for Angina Pectoris
The preceding epidemiologic evidence describing associations between short-term
increases in ambient NO2 concentrations and increased hospital admissions and ED visits
for MI and IHD is supported by evidence for increases in hospital admissions and ED
visits for angina. Angina pectoris results from an imbalance between the demand for
oxygen in the heart and the delivery by the coronary artery. Reduction in coronary blood
5-268
-------
flow due to atherosclerosis is a common cause of this imbalance. Unstable angina, where
the coronary artery is not completely occluded, can lead to MI.
The 2008 ISA for Oxides of Nitrogen did not include specific discussion of angina but
did report results from two studies that examined associations between ambient NO2
concentrations and angina hospital admissions (U.S. EPA. 2008c). In a study of five
European cities, von Klot et al. (2005) examined the relationship between short-term air
pollution and hospital readmissions of myocardial infarction survivors. The authors
reported a 16% (95% CI: 3, 31) increase in risk of hospital readmissions for angina
pectoris per 20-ppb increase in 24-h avg NCh on the same day. Poloniecki et al. (1997)
observed a smaller, but statistically significant association between NO2 concentrations
on the previous day and angina hospital admissions in London, U.K. (Table 5-40).
Neither study evaluated copollutant models.
More recent studies add to the limited, but consistent evidence of an association between
ambient NO2 exposure and angina hospital admissions and ED visits. Szyszkowicz
(2009) found that NC>2 concentrations were associated with risk of ED visits for chest
pain in six Canadian cities. The magnitude of association was stronger in the warm
season, with a 5.9% increase in risk (95% CI: 3.3, 8.6) than in the cold season, with a
3.2% increase in risk (95% CI: 1.5, 5.0) at lag 1 per 20-ppb increase in 24-h avg NO2. As
discussed in Section 5.3.2.1. Stieb et al. (2009) examined the composite endpoint of acute
MI or angina ED visits in a study of seven Canadian cities that included overlapping data
with Szyszkowicz (2009). Stieb et al. (2009) observed a positive association between
ambient NO2 and Ml/angina that was still positive, but attenuated, imprecise, and no
longer statistically significant after adjustment for CO in a copollutant model. In addition
to limited interpretability from using a composite endpoint, the results were also largely
influenced by data from one city, as detailed in Section 5.3.2.1.
5.3.2.3 ST-Segment Amplitude
ST-segment changes (either ST-segment elevation or depression) on the
electrocardiogram are considered a nonspecific marker of myocardial ischemia. While
the 2008 ISA for Oxides of Nitrogen did not review any epidemiologic studies of ambient
oxides of nitrogen concentrations and markers of myocardial ischemia (U.S. EPA.
2008c). a few recent studies report associations (Table 5-41). Chuang et al. (2008)
conducted a repeated-measures study of Boston-area adults with a history of coronary
heart disease and examined the association between ambient pollutants and ST-segment
changes. The authors reported a RR of 3.29 (95% CI: 1.82, 5.92) for ST-segment
depression of >0.1 mm per 20-ppb increase in 24-h avg NO2 concentrations over the
5-269
-------
previous 24 hours. This finding was robust to additional adjustment for PIVb 5 in a
copollutant model (RR: 3.29 [95% CI: 1.65, 6.59]).
Delfino etal. (2011) used a similar design to study 38 older, nonsmoking adult residents
of four retirement homes in the Los Angeles area with a documented history of coronary
artery disease. A particular strength of this study is that the authors measured pollutant
concentrations outside of the residence, which improved spatial matching of NC>2
concentrations to subjects' locations. The authors observed an OR of 3.83 (95% CI: 1.20,
12.16) for ST-segment depression > 1.0 mm per 17.4-ppb increase in mean 1-hour NC>2
concentrations preceding measurement over the previous 3 days. Other averaging periods
from 8 hours to 4 days gave similar or slightly weaker results. NO2 was more strongly
associated with ST depression than was NOx. No copollutant models were evaluated.
5-270
-------
Table 5-41 Epidemiologic studies of ST-segment amplitude.
Study
Location Sample Mean NO2
Size ppb
Exposure Assessment
Selected Effect Estimates3
(95% Cl)
tChuanq et al.
(2008)
Boston, MA 24-h avg NO2 Citywide avg
n=48 21.4
75th: 24.9
Max: 44.5
ST segment change (mm)
12-h: -0.02 (-0.05, 0.00)
24-h: -0.08 (-0.12, -0.05)
RR for ST-segment
depression >0.1 mm
12-h: 1.15(0.72, 1.82)
24-h: 3.29(1.82, 5.92)
tDelfino etal.
(2011)
Los Angeles, CA
n = 38
1-h NO2:
27.5
1-h NOx:
46.6
Outdoor monitor at retirement
community
OR for ST-segment
depression >1.0 mm
NO2 per 17.4-ppb increase in
1-h mean:
1-h: 1.18(0.90, 1.54)
8-h: 1.65(1.08,2.52)
24-h: 2.47 (1.27, 4.78)
2-day: 3.22(1.26,8.23)
3-day: 3.83(1.20, 12.16)
4-day: 2.68 (0.78, 9.20)
NOx per 42.3-ppb increase in
1-h mean:
1-h: 1.17(0.97, 1.42)
8-h: 1.33(0.96, 1.86)
24-h: 1.56(0.88,2.76)
2-day: 1.69(0.76,3.72)
3-day: 1.81 (0.57,5.72)
4-day: 1.53(0.33,7.02)
avg = average; CA = California; Cl = confidence interval; h = hours; MA = Massachussets; max = maximum; NO2 = nitrogen
dioxide; NOX = sum of NO and NO2; OR = odds ratio; ppb = parts per billion; RR = relative risk; ST-segment = segment of the
electrocardiograph between the end of the S wave and beginning of the T wave.
aEffect estimates are standardized to a 20-ppb or 30-ppb increase in NO2 or 40-ppb or 100-ppb increase in NOX concentration for
24-h avg and 1-h max metric, respectively.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-271
-------
5.3.2.4 Summary of Myocardial Infarction
In summary, the epidemiologic data available continue to support associations between
short-term increases in ambient NO2 concentrations and increased risk of triggering an
** oo o
MI. However, potential copollutant confounding by traffic-related pollutants was not
examined extensively in these studies. In the studies that did analyze copollutant models
to adjust for another traffic pollutant, the findings were generally inconsistent.
Associations between ambient NO2 and risk of hospital admissions or ED visits for MI
and IHD were attenuated by adjustment for CO in two studies (Nuvolone et al.. 2011:
Stieb etal.. 2009) but remained robust in a few studies conducted in Taiwan (Hsieh et al..
2010; Cheng et al.. 2009a: Yang. 2008). Additionally, Rich etal. (2010) reported that an
association between short-term NO2 exposure and hospital admission for MI was
attenuated by the inclusion of PM2 5 in a copollutant model. There is limited, but
consistent evidence of an association between NO2 and angina pectoris (Stieb et al..
2009; Szyszkowicz. 2009: von Klot et al.. 2005: Poloniecki et al.. 1997). However, only
one study examined a copollutant model, in which the NO2 association was attenuated in
magnitude and precision after adjustment for CO (Stieb etal.. 2009). None of the
reviewed studies of MI, IHD, or angina utilized copollutant models to adjust for potential
confounding by EC or VOCs. Additionally, all of the studies of Mi-related health effects
used central site monitors to assess ambient NO2 exposures. Central site monitors have
noted limitations in capturing the variation in NO2 (Section 3.4.4.2). and none of the
studies of MI reported information on the extent to which concentrations at central site
monitors captured the temporal variation in NO2 across the study area
In addition to hospital admission and ED visit studies, a few available epidemiologic
studies report an association between short-term increases in NO2 and ST-segment
changes on the electrocardiogram of older adults with a history of coronary artery
disease. These results potentially indicate an association between NO2 and increased risk
of myocardial ischemia in a population with pre-existing cardiovascular disease. No
studies from the 2008 ISA of Oxides of Nitrogen are available on ST-segment changes
for comparison. Once again, there was limited assessment of potential confounding by
traffic pollutants in copollutant models, although Chuang et al. (2008) reported that the
association between NO2 and ST-segment changes was robust to PM2 5 adjustment.
5-272
-------
5.3.3 Arrhythmia and Cardiac Arrest
5.3.3.1 Panel Epidemiologic Studies
The 2008 ISA for Oxides of Nitrogen found little epidemiologic evidence of an
association between short-term changes in ambient NO2 concentrations and cardiac
arrhythmias (U.S. EPA. 2008c). There continues to be limited epidemiologic evidence for
such an association, either from panel studies of patients with ICDs or panel studies of
arrhythmias detected on ambulatory ECG recordings (Table 5-42).
In a study of patients with ICDs, Ljungman et al. (2008) found that NO2 was associated
with increased risk of confirmed ventricular tachyarrhythmias (VT). The association with
PMio and PM2 5 was stronger than the association for NO2. There was no evidence of
effect measure modification by city, distance from the nearest ambient monitor at the
time of the event, number of events, type of event (ventricular fibrillation versus
ventricular tachycardia), age, history of IHD, left ventricular ejection fraction, diabetes,
body mass index, or use of beta blockers. However, a stronger association between NO2
and risk of VT was observed for the 22 subjects who were outdoors at the time of ICD
activation. Because the authors accounted for personal activity/behavior, exposure
measurement error may have been reduced for subjects who were outdoors given that
more time spent outdoors is likely to correspond to a greater personal-ambient correlation
(Section 3.4.4.1). In a similar study, Anderson et al. (2010) observed generally null
associations between ICD activation and ambient NO, NO2, or NOx concentrations.
Anderson etal. (2010) only had the study cardiologist review the electrocardiograms
from about 60% of ICD activations (confirming 87% of those cases as VT), potentially
leading to greater misclassification of the outcome than in the study by Ljungman et al.
(2008). Recently, Link etal. (2013) examined a panel of patients with dual chamber
ICDs. They observed positive associations between ICD-detected arrhythmias and atrial
fibrillations >30 seconds and NO2 concentrations that were generally stronger when the
authors used a 2-hour lag compared to a 2-day lag. Finally, Metzger et al. (2007)
observed generally null associations between NO2 concentrations and VT events over a
10-year period in Atlanta, GA.
Using a different approach, Bartell etal. (2013) used ECG monitors to evaluate VT
events in 50 older adult, nonsmoking residents of four retirement communities in the
greater Los Angeles area. The study reported a 35% (95% CI: -1, 82) increase in the
daily rate of VT events per 40-ppb increase in 24-h avg NOx. The estimated effect of
3- and 5-day avg NOx on the daily rate of VT was somewhat stronger, though markedly
less precise (i.e., wider confidence limits around the effect estimates). Bartell et al. (2013)
measured pollutant concentrations outside of each of the retirement communities, which
5-273
-------
improved spatial matching of NC>2 concentrations to subjects' locations. Conversely,
Barclay et al. (2009) generally observed weak and inconsistent associations between NO2
or NO and incident arrhythmias detected on ambulatory ECG recordings in a repeated
measures study of nonsmoking adults with stable heart failure.
Table 5-42 Epidemiologic studies of arrhythmia and cardiac arrest.
Location Mean NOz
Study Sample Size (ppb)
Exposure Assessment
Selected Effect Estimates
(95% Cl)a
tLjunqman Gothenburg
and
Stockholm,
Sweden
n=211
(266 events)
24-h avg
NO2
Central monitor
1 in Gothenburg average of 2 in
Gothenburg: Stockholm
11.8
Stockholm:
8.3
Ventricular tachyarrhythmia (OR)
2-havg: 1.37(0.53, 3.64)
24-h avg: 1.26(0.49, 3.32)
tAnderson
etal.
(2010)
London, U.K. 24-h avg
n = 705 NO* 12.1
(5,462 device 24-h avg
activations) NOx: 24.1
24-h avg
NO: 19.4
Central monitor
City wide avg
ICD activations (OR)
NO2, lag 0-1 avg: 0.93 (0.70, 1.24)
NOx, lag 0-1 avg: 0.92 (0.86, 1.08)
NO, lag 0-1 avg: 0.96 (0.93, 1.04)
tLink etal.
(2013)
Boston, MA
n = 176 (328
atrial
fibrillation
episodes >30
sec)
24-h avg
NO2: 16.1
Central monitor
City wide avg
ICD-detected arrhythmias (OR)
24-h lag: 1.23(0.75,2.10)
2-h lag: 1.57(0.97,2.47)
Metzqer et Atlanta, GA
al. (2007)
1-h max
NO2: 44.9
90th: 68
Max: 181
Central monitor
All arrhythmia events (OR)
Allyr: 1.00(0.95, 1.05)
Warm season: 1.00(0.93, 1.08)
Cold season: 1.01 (0.94, 1.08)
Events resulting in cardiac pacing or
defibrillation:
Allyr: 1.01 (0.94, 1.10)
Events resulting in defibrillation:
Allyr: 1.07(0.93, 1.23)
tBartell et Los Angeles, 24-h avg
al. (2013) CA NOx: 42.3
n = 50(302 Max: 183.7
subject h of
VT observed)
Monitors on trailers at each of 4
retirement communities
Ventricular tachyarrhythmia (RR)
24-h avg: 1.35(0.99, 1.82)
3-day avg: 1.74(0.47,6.40)
5-day avg: 1.65(0.56,4.93)
5-274
-------
Table 5-42 (Continued): Epidemiologic studies of arrhythmia and cardiac arrest.
Location Mean NOz
Study Sample Size (ppb)
Exposure Assessment
Selected Effect Estimates
(95% Cl)a
Aberdeen,
Scotland,
U.K.
n = 132
24-h avg
NO2: 30.1
NO: 14.7
Central monitor
All arrhythmias (change in magnitude)
NO2: 3.193 (-3.600, 9.985)
NO: 3.524 (-3.059, 10.107)
Ventricular ectopic beats
NO2: 3.642 (-4.837, 12.121)
NO: 4.588 (-3.628, 12.803)
Ventricular couplets
NO2: 0.356 (-7.395, 8.106)
NO: -0.085 (-7.601, 7.431)
Ventricular runs
NO2: 2.443 (-2.537, 7.422)
NO: 2.219 (-2.618, 7.055)
Supraventricular ectopic beats
NO2: 2.888 (-4.833, 10.608)
NO: -2.688 (-10.170, 4.794)
Supraventricular couplets
NO2: 5.209 (-1.896, 12.313)
NO: 1.366 (-5.542, 8.274)
Supraventricular runs
NO2: 3.441 (-1.760,8.641)
NO: 2.298 (-2.753, 7.348)
avg = average; CA = California; Cl = confidence interval; GA = Georgia; h = hours; ICD = implantable cardioverter defibrillators;
MA = Massachussets; NO = nitric oxide; NO2 = nitrogen dioxide; NOX = sum of NO and NO2; OR = odds ratio; RR = relative risk;
U.K. = United Kingdom; VT = ventricular tachyarrhythmias.
aEffect estimates are standardized to a 20-ppb or 30-ppb increase in NO2 or NO or 40-ppb or 100-ppb increase in NOX
concentration for 24-h avg and 1-h max metrics, respectively.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5.3.3.2 Out-of-Hospital Cardiac Arrest
The majority of out-of-hospital cardiac arrests are due to cardiac arrhythmias.
Dennekamp et al. (2010) observed generally positive, though weak, associations between
NO2 concentrations and risk of out-of-hospital cardiac arrest (Table 5-43). A similar
approach was used by Silverman et al. (2010) using data from out-of-hospital cardiac
arrests in New York City and observed generally null associations with NC>2
concentrations in all year and cold season analyses, and a positive association in the
warm season analysis. Straney et al. (2014) also reported null associations between
out-of-hospital cardiac arrest and ambient NO2 concentrations from a case-crossover
study in Perth, Australia. In other studies of out-of-hospital cardiac arrest, Ensor et al.
5-275
-------
(2013) found inconsistent and weak associations with ambient NC>2 concentrations in
Houston, while Wichmann et al. (2013) reported similarly inconsistent associations with
NOx in Copenhagen. However, Wichmann et al. (2013) observed a positive association
between ambient NOx concentration and out-of-hospital cardiac arrest in females (46%
[95% CI: 8, 99] increase per 40-ppb increase in 24-h avg NOx at lag 3), although there
were slightly under two thirds the amount of cases observed in females compared to
males. None of the out-of-hospital cardiac arrest studies examined potential copollutant
confounding of NO2 or NOx associations. No studies from the 2008 ISA for Oxides of
Nitrogen are available for comparison.
Table 5-43 Epidemiologic studies of out-of-hospital cardiac arrest.
Study
tDennekamp et al.
(2010)
fSilverman et al.
(2010)
Location
Sample Size
Melbourne,
Australia
n = 8,434
New York City, NY
n = 8,216
Mean NO2
(ppb) Exposure Assessment
24-h avg Central site monitor
NO2
12.0
75th: 15.16
24-h avg Central site monitor
N°2 Citywide avg
50th: 27
75th: 32
95th: 43
Selected Effect Estimates3
(95% CI)
% change in out-of-hospital
cardiac arrest
Lag 0:3.23 (-10.19, 18.51)
Lag 1:7.69 (-7.29, 25.11)
Lag 2: -4.51 (-17.53, 10.56)
Lag 3: 7.37 (-7.11, 24.13)
Lag 0-1 avg:
9.28 (-7.54, 29.14)
No quantitative results
presented for NO2.
5-276
-------
Table 5-43 (Continued): Epidemiologic studies of out-of-hospital cardiac arrest.
Study
Location
Sample Size
Mean NO2
(PPb)
Exposure Assessment
Selected Effect Estimates3
(95% Cl)
tStraney et al.
(2014)
Perth, Australia
n = 8,551
1-h max NO2 Nearest central site monitor OR
50th: 3.0
75th: 8.1
95th: 19.8
(avg and/or max distances
not specified)
Lag 0-h:
1.008(0.992, 1.030)
Lag 1-h:
1.000(0.979, 1.021)
Lag 2-h:
0.987(0.967, 1.004)
Lag 3-h:
0.992(0.971, 1.013)
LagO-1-h:
1.004(0.987, 1.026)
Lag 0-3-h:
0.996(0.975, 1.017)
LagO-12-h:
0.996(0.971, 1.026)
tEnsoretal. (2013) Houston, TX
n = 11,677
24-h avg
NO2
9.11
75th: 11.66
95th: 16.87
Central site monitor
Citywide avg
% change in out-of-hospital
cardiac arrests
Lag 0:3.2 (-11.3, 18.9)
Lag 1:-2.5 (-14.7, 11.0)
Lag 2:-1.4 (-13.8, 12.6)
Lag 3: 3.2 (-9.6, 17.7)
Lag 4: 1.1 (-11.5, 15.3)
Lag 0-1: -0.4 (-14.4, 16.1)
Lag 1-2: -2.8 (-16.3, 12.9)
tWichmann et al.
(2013)
Copenhagen,
Denmark
n= 4,657
24-h avg
NOx
14.75
75th: 18.35
Central site monitor
% change in out-of-hospital
cardiac arrests
LagO: -13.5 (-28.6, 5.0)
Lag 1: 5.4 (-12.7, 27.4)
Lag 2: -9.2 (-25.0, 11.6)
Lag 3: 16.0 (-4.0, 40.2)
Lag 4: 8.7 (-10.2, 31.2)
Lag 5: 5.4 (-13.1, 27.9)
Males, lag 3:
2.2 (-19.7, 30.1)
Females, lag 3:
46.1 (7.8, 98.7)
avg = average; Cl = confidence interval; NO2 = nitrogen dioxide; NOX = sum of NO and NO2; NY = New York; OR = odds ratio;
TX = Texas.
aEffect estimates are standardized to a 20-ppb or 30-ppb increase in NO2 or NO or 40-ppb or 100-ppb increase in NOX
concentration for 24-h avg and 1-h max metrics, respectively.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-277
-------
5.3.3.3 Hospital Admissions and Emergency Department Visits for Arrhythmias
There are a limited number of studies examining associations between short-term NCh
exposure and hospital admissions with a primary discharge diagnosis related to
arrhythmias. Using data from 14 hospitals in seven Canadian cities, Stieb et al. (2009)
found no association between NCh and risk of hospital admission for arrhythmias.
However, Tsai et al. (2009) reported a positive association in Taipei, Taiwan that was
stronger on cool days (OR: 1.34 [95% CI: 1.25, 1.44] per 20-ppb increase in 24-h avg
NO2) than warm days (OR: 1.19 [95%CI: 1.10, 1.28] per 20-ppb increase in 24-h avg
NO2). Both cool and warm day associations remained robust in copollutant models with
PMio, SO2, CO, or Os; however, potential confounding by most of the traffic-related
pollutants of concern was not evaluated.
5.3.3.4 Summary of Arrhythmia and Cardiac Arrest
In summary, there is currently inconsistent epidemiologic evidence for an association
between 24-h avg NO2, NO, or NOx and risk of cardiac arrhythmias as examined in
patients with ICDs, continuous ECG recordings, out-of-hospital cardiac arrest, and
hospital admissions. The reviewed studies rarely adjusted for copollutant confounding by
traffic pollutants, focused almost exclusively on ventricular arrhythmias, and are
potentially limited by misclassification of the outcome. Additionally, the majority of
studies used central site monitors to estimate ambient NO2 exposure. Central site
monitors have noted limitations in capturing the variation in NO2 (Section 3.4.4.2). and
none of the studies of arrhythmia and cardiac arrest reported information on the extent to
which concentrations at central site monitors captured the temporal variation in NO2
across the study area
5.3.4 Cerebrovascular Disease and Stroke
5.3.4.1 Hospital Admissions and Emergency Department Visits
The 2008 ISA for Oxides of Nitrogen found that the epidemiologic evidence for
associations between short-term changes in NO2 concentrations and hospital admissions
or ED visits for cerebrovascular diseases was generally inconsistent and provided little
support for an independent NO2 effect (U.S. EPA. 2008c). Recent studies also provide
inconsistent evidence (Figure 5-19 and Table 5-44).
5-278
-------
Generally, studies based on clinical registries are less susceptible to misclassification of
the outcome and exposure, which may explain the stronger evidence provided by these
studies than that based on administrative data. Wellenius et al. (2012) reviewed the
medical records of 1,705 Boston-area patients hospitalized with neurologist-confirmed
acute ischemic stroke and found an OR for ischemic stroke onset of 1.32 (95% CI: 1.08,
1.63) per 20-ppb increase in NC>2 concentration averaged over the 24 hours preceding
hospitalization for stroke. A unique strength of this study was the availability of
information on the date and time of stroke symptom onset in most patients, thereby
potentially reducing misclassification of the exposure. Copollutant models were not
evaluated.
Andersen et al. (2010) obtained data on strokes in Copenhagen, Denmark from the
Danish National Indicator Project and found a positive association between ambient NOx
concentrations and risk of ischemic stroke but not hemorrhagic stroke. The strongest
association was observed in relation to NOx levels 4 days earlier and for those suffering a
mild stroke, but the association was attenuated after adjustment for UFP. Using data from
a stroke registry in Como, Italy, Vidale etal. (2010) found that NO2 was associated with
risk of ischemic stroke hospital admission. On the other hand, Turin etal. (2012) did not
observe any association using data from the Takashima County Stroke and AMI Registry
in central Japan. Similarly, Oudin etal. (2010) found no association between modeled
residential NOx concentration and risk of ischemic or hemorrhagic stroke within the
context of a Swedish quality register for stroke.
5-279
-------
Study
Thachetal. (2010)
Zheng etal. (2013)
Bell etal. (2008)
Xiang etal. (2013)
Villeneuveetal. (2012)
Turin etal. (2012)
Welleniusetal. (2012)
Andersen et al. (2010)
Larrieu et al. (2007)
Polonieckietal. (1997)
Linn etal. (2000)
Tsaietal. (2003)
Wong etal. (1999)
Outcome Mean Concentration
Stroke 31.2
28.2
CBV
CBV
CBV
26.4
Lag
0
0-3
0
1
0-3
0
1
0-3
0
1
0-3
0
0-2
0
0-2
Stroke 16.7
Ischemic Stroke
Hemorrhagic Stroke
Transient ischemic Stroke
Stroke
Ischemic Stroke
Hemorrhagic Stroke
Transient (schemic Stroke
Stroke 16
Cerebral Infarction
Intracerebral Hem
Subarachnoid hem
Ischemic Stroke
Ischemic Stroke 15.3
Hemorrhagic Stroke
Mild Ischemic
Stroke 11.9-24.9
CBV
Occ Stroke
Cerebral Stroke
Ischemic Stroke
Cerebral Stroke
Ischemic Stroke
CBV
35
28-41
27.3
24
Villeneuve et al. (2006a) Ischemic Stroke
Hemorrhagic Stroke
Cerebral Stroke
Welleniusetal. (2005) Ischemic Stroke 23.5
Hemorrhagic Stroke
Ponka and Virtanen (1996) CBV 20.7
Ballesteretal. (2001)
Peel etal. (2007)
Chen etal. (2014c)
Jalaludinetal. (2006)
CBV
CBV
Ischemic Stroke
61.8
45.9
0
1
0-4
0
1
0-4
0-4
1
0
0-2
0-2
2
0-2
0
1
0-1
Notes
All 13 monitors
5 city monitors
8 correlated monitors
Warm season
Cool season
Warm season
Cool season
24-h
+-
-O-
^p-
-a-
All ages 4
Ages > 65 yr ^
20+C
<20C
ges > 65 yr
£
f
1-h
Case-crossover
Time series
All year
Warm season
Cold season
Ages > 65 yr
0.0 0.5 1.0 1.5 2.0
Relative Risk (95% Cl)
2.5
Note: CBV = cerebrovascular; Cl = confidence interval; hr = hour; yr = years. Black = studies from the 2008 Integrated Science
Assessment for Oxides of Nitrogen, red = recent studies. Circles = NO2, diamonds = NOX. Relative risks are standardized to a
20-ppb or 30-ppb increase in NO2 concentration and 40 ppb or 100 ppb for NOX concentrations for 24-h avg and 1-h max metrics,
respectively. Studies are organized first by averaging time, then recent versus previous study, then descending order of mean
concentration (in parts per billion).
Figure 5-19 Associations between short-term exposure to oxides of nitrogen
and hospital admissions for cerebrovascular disease and stroke.
5-280
-------
Table 5-44 Corresponding risk estimates for hospital admissions for
cerebrovascular disease and stroke for studies presented in
Figure 5-19.
Study
Location
Health Effect
Selected Relative Risks3
95% Cl
Copollutant Examination13
fThach et al.
(2010)
tZhenq et al.
(2013)
Hong Kong,
China
Lanzhou,
China
Stroke
Cerebrovascular
disease
Lag 0-1 avg:
1.01 (1.00, 1.03)
Lag 0: 1.05(1.02, 1.08)
I an 0—3 avrr
No copollutant models.
NO2: associations were
robust to adjustment for
1.06(1.02, 1.10)
SO2, associations increased
with adjustment for PM-io.
Copollutants: SO2 (positive)
associations and PM-io
(negative) associations
robust to adjustment for
NO2.
NO2 correlations (Spearman
r): PMio: 0.64; SO2: 0.64.
tBell et al. (2008) Taipei,
Taiwan
Cerebrovascular
disease
All 13 monitors
Lag 0: 1.01 (0.95, 1.07)
Lag 1:0.97(0.92, 1.03)
Lag 0-3 avg:
1.04(0.96, 1.12)
5 city monitors
Lag 0: 1.01 (0.96, 1.06)
Lag 1:0.97(0.92, 1.02)
Lag 0-3 avg:
1.04(0.96, 1.12)
8 correlated monitors
LagO: 1.01 (0.96, 1.06)
Lag 1:0.97(0.92, 1.02)
Lag 0-3 avg:
1.04(0.96, 1.12)
No copollutant models.
tXianqetal. (2013) Wuhan,
China
Stroke
Warm season
Lag 0: 0.99 (0.93, 1.06)
Lag 0-2 avg:
0.96(0.88, 1.05)
Cool season
LagO: 1.11 (1.05, 1.18)
Lag 0-2 avg:
1.12(1.04, 1.21)
: cold season
association robust to PMio
adjustment.
Copollutants: PMio no
longer associated with
stroke hospital admissions
in the cold season after NO2
adjustment.
No correlations reported.
5-281
-------
Table 5-44 (Continued): Corresponding risk estimates for hospital admissions for
cerebrovascular disease and stroke for studies presented
in Figure 5-19.
Study
tVilleneuve et al.
(2012)
Location
Edmonton,
Canada
Health Effect
Stroke
Selected Relative Risks3
95% Cl
Lag 0-2 avg
Warm- 1 49 m Q4 9 "m
Copollutant Examination13
Ischemic stroke during
warm season
Cool: 0.98(0.84, 1.15)
Ischemic stroke
Lag 0-2 avg
Warm: 2.37 (1.27, 4.41)
Cool: 0.90(0.69, 1.13)
Hemorrhagic stroke Lag 0-2 avg
Warm: 1.50(0.59,4.32)
Cool: 0.98(0.64, 1.50)
Transient ischemic
stroke
Lag 0-2 avg
Warm: 0.67(0.33, 1.42)
Cool: 1.11 (0.80, 1.45)
NO2: associations robust to
. adjustment for 862; slightly
attenuated but positive after
adjustment for CO, Os, or
PM2.5.
Copollutants: CO, Os, and
• PlVh.s associations
attenuated by adjustment
for NO2. No association
between SO2 and ischemic
stroke.
Hemorrhagic stroke during
warm season
NO2: associations
attenuated after adjustment
forSO2and Os, but
increased after adjustment
for CO or PlVh.5.
Copollutants: SO2 and Os
associations robust to NO2
adjustment; CO no longer
associated with
hemorrhagic stroke after
NO2 adjustment.
tTurin et al. (2012) Takashima
County,
Japan
fWellenius et al. Boston, MA
(2012)
Stroke
Cerebral infarction
Intracerebral
hemorrhage
Subarachnoid
hemorrhage
Ischemic stroke
Lag 0: 0.98 (0.89, 1.08)
Lag 1:0.98(0.89, 1.08)
Lag 0: 0.98 (0.87, 1.10)
Lag 1: 1.00(0.89, 1.12)
LagO: 1.06(0.85, 1.33)
Lag 1: 0.94(0.75, 1.16)
LagO: 1.12(0.80, 1.56)
Lag 1: 1.02(0.73, 1.42)
Lag 24 h preceding event:
1.32(1.08, 1.63)
No evidence of an
association between NO2
and stroke. Copollutant
models did not change the
results.
No copollutant models.
5-282
-------
Table 5-44 (Continued): Corresponding risk estimates for hospital admissions for
cerebrovascular disease and stroke for studies presented
in Figure 5-19.
Study
tAndersen et al.
(2010)
Location Health Effect
Copenhagen, Ischemic stroke
Denmark
Selected Relative Risks3
95% Cl
NOx
LagO: 1.20(0.96, 1.48)
Copollutant Examination13
NOx: no longer associated
with ischemic stroke after
Lag 1:0.96(0.79, 1.20)
Lag 0-4 avg:
1.36(0.96, 1.89)
adjustment for UFP.
Copollutants: UFP
association robust after
. adjustment for NOx.
Hemorrhagic stroke
LagO: 0.96(0.48, 1.81)
Lag 1: 1.14(0.59,2.20)
tLarrieu et al.
(2007)
Poloniecki et al.
(1997)
Linnetal. (2000)
Tsai et al. (2003)
Mild ischemic
stroke
8 French Stroke
cities
London, U.K. Cerebrovascular
disease
Los Angeles, Cerebrovascular
CA disease
Occlusive stroke
Kaohsiung, Cerebral stroke
Taiwan
Ischemic stroke
Lag 0-4 avg:
0.43(0.13, 1.36)
Lag 0-4 avg:
1.61 (0.79, 3.30)
Alleges: 0.99(0.96, 1.03)
>65yr: 1.01 (0.97, 1.05)
Lag 1: 0.99(0.98, 1.00)
Lag 0: 1.01 (0.99, 1.02)
LagO: 1.04(1.02, 1.06)
Lag 0-2 avg
20+°C: 1.68(1.38,2.04)
<20°C: 0.78(0.44, 1.37)
Lag 0-2 avg
20+°C: 1.67(1.48, 1.87)
<20°C: 1.19(0.78, 1.84)
No copollutant models.
No copollutant models
examined.
No copollutant models.
NO2 correlations: PMm
0.67 to 0.88; Os: -0.23 to
0.35; CO: 0.84 to 0.94
NO2: Ischemic stroke and
hemorrhagic stroke
associations robust to SO2,
CO, or Os adjustment.
Attenuated, but positive
after PM-io adjustment.
Copollutants: PM-io, SO2,
CO, and Os ischemic stroke
and hemorrhagic stroke
associations attenuated by
adjustment for NO2.
5-283
-------
Table 5-44 (Continued): Corresponding risk estimates for hospital admissions for
cerebrovascular disease and stroke for studies presented
in Figure 5-19.
Selected Relative Risks3
Study
Wonqetal. (1999)
Location
Hong Kong,
China
Health Effect
Cerebrovascular
disease
95% Cl
Lag 0-1 avg:
1.03(0.99, 1.07)
Copollutant Examination13
No copollutant models.
Villeneuve et al. Edmonton, Ischemic stroke
(2006a) AB, Canada
Ages 65 yr and older
LagO: 1.04(0.96, 1.15)
Lag 1: 1.07(0.97, 1.17)
Ischemic stroke during
warm season
NO2: warm season
associations robust to
adjustment for SO2 or CO;
increase with adjustment for
PM-io or PlVh.s; attenuated
with adjustment for Os.
Hemorrhagic stroke Lag 0: 1.07 (0.96, 1.21)
Lag 1: 1.06(0.94, 1.20)
Hemorrhagic stroke during
warm season
NO2: warm season
associations robust to SO2,
03, PM2.5, orPM-io
adjustment (large increases
in CIs in models with PM);
but attenuated with
adjustment for CO.
NO2 warm season
correlations (Pearson r):
SO2: 0.22; O3: -0.09; CO:
0.59; PIvh.s: 0.52; PMm
0.57.
Wellenius et al.
(2005)
Ponka and
Virtanen(1996)
Ballester et al.
(2001)
9 U.S. cities
Helsinki,
Finland
Valencia,
Spain
Cerebral stroke
Ischemic stroke,
hemorrhagic stroke
Cerebrovascular
disease
Cerebrovascular
disease
LagO: 0.99(0.90, 1.07)
Lag 1: 0.91 (0.84, 1.00)
LagO: 1.05(1.03, 1.07)
LagO: 1.01 (0.96, 1.06)
Lag 0: 0.96 (0.87, 1.07)
Lag 1:0.98(0.87, 1.09)
Lag 2: 1.22(1.04, 1.44)
No copollutant models.
No copollutant models.
NO2: associations were
robust to adjustment for
SO2 orBS.
Peel et al. (2007) Atlanta, GA Cerebrovascular
disease
Lag 0-2 avg
Case crossover:
1.05(1.01, 1.09)
Time series:
1.06(1.02, 1.11)
No copollutant models.
5-284
-------
Table 5-44 (Continued): Corresponding risk estimates for hospital admissions for
cerebrovascular disease and stroke for studies presented
in Figure 5-19.
Study
tChen et al.
(2014c)
Jalaludin et al.
(2006)
Location Health Effect
Edmonton, Ischemic stroke
AB, Canada
Sydney, Stroke
Australia
Selected Relative Risks3
95% Cl
Lag 1 -8 h avg
All year: 1.06(0.98, 1.14)
Warm season:
1.39(1.13, 1.71)
Cold season:
0.98(0.89, 1.08)
Ages 65 yr and older
i an rv n QF; m aa 1 rr?^
Copollutant Examination13
No copollutant models.
No copollutant models
analyzed for
Lag 1:0.96 (0.90, 1.03) cerebrovascular disease.
Lag 0-1 avg:
0.95(0.88, 1.02)
AB = Alberta; avg = average; BS = black smoke; CA = California; Cl = confidence interval; CO = carbon monoxide; GA = Georgia;
MA = Massachusetts; NO2 = nitrogen dioxide; NOX = sum of NO and NO2; O3 = ozone; PM2 5 = particulate matter with a nominal
mean aerodynamic diameter less than or equal to 2.5 |jm; PMio = particulate matter with a nominal mean aerodynamic diameter
less than or equal to 10 |jm; SO2 = sulfur dioxide; UFP = ultrafine particles; U.K. = United Kingdom; U.S. = United States.
aEffect estimates are standardized to a 20-ppb or 30-ppb increase in NO2 or 40-ppb or 60-ppb increase in NOX concentration for
24-h avg and 1-h max metrics, respectively.
bRelevant relative risks for copollutant models can be found in Supplemental Figures S5-2, S5-3, S5-4,. and S5-5 (U.S. EPA.
2015b. c, d, e).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
Additional studies based on administrative data are also available. A number of the
administrative data studies were conducted in Edmonton, Canada and used similar or
identical data sources. The most thorough Edmonton study observed an association
between NO2 and ED visits for ischemic stroke in the warm season (OR: 2.37 [95% Cl:
1.27, 4.41] per 20-ppb increase in lag 0-2 day avg NO2) and an imprecise (i.e., wide 95%
Cl) association between hemorrhagic stroke in the warm season (OR: 1.50 [95% Cl: 0.59,
4.32] per 20-ppb increase in lag 0-2 day avg NO2) (Villeneuve et al., 2012). Villeneuve
et al. (2012) also examined copollutant models in which ischemic stroke associations
were robust to adjustment for SO2 (OR: 2.34 [95% Cl: 1.25, 4.37]), and remained
positive, but slightly attenuated after adjustment for CO (OR: 2.05 [95% Cl: 0.92, 4.61]),
O3 (OR: 1.92 [95% Cl: 0.98, 3.78]), and PM25 (OR: 1.98 [95% Cl: 0.94, 4.20]). The
hemorrhagic stroke associations were attenuated after adjustment for SO2 or Os but were
robust to adjustment for the traffic-related pollutant CO or PM2 5. There is the potential
for misclassification of the exposure due to differences in the timing of stroke symptoms
and the corresponding ED visit; however, after surveying a subset of the study
population, Villeneuve et al. (2012) observed that roughly 75% of patients visited the
5-285
-------
emergency room on the same day that their symptoms presented. Further, when the
authors adjusted the assigned pollution levels from the day of the ED visit to the day of
symptom presentation, they observed no systematic differences in assigned pollution
levels. Szyszkowicz (2008) observed a positive association between 24-h avg NO2 and
ED visits for ischemic stroke in Edmonton, Canada but only within specific subgroups
according to sex, season, and age. In a recent related study, Chenetal. (2014c) used the
same data, but applied hourly NO2 values to their models. The authors reported an
association between NO2 and ED visits for acute ischemic stroke that remained relatively
consistent across lag days. However, that association was almost entirely influenced by
the association observed in the warm season, as the association in the cold season was
null.
Zheng etal. (2013) conducted a time-series study in Lanzhou, China and found a positive
association between NO2 and all cerebrovascular hospital admissions. The strongest
relationships were observed on same-day and 3-day cumulative lags. Zheng etal. (2013)
also reported stronger associations in women and the elderly. Xiang etal. (2013)
observed a positive association between NO2 and hospital admissions for all strokes in
the cold season in Wuhan, China that was robust in a copollutant model including PMio.
Conversely, in Taipei, Taiwan, Bell et al. (2008) did not observe an association between
NO2 and cerebrovascular disease. As mentioned in Section 5.3.2.1. Bell et al. (2008)
attempted to reduce uncertainty related to the use of central site monitors by estimating
NC>2 exposure over the entire Taipei area (average of 13 monitors), within Taipei City
only (average of 5 monitors), and using a subset of monitors where all pairs of monitors
had NC>2 correlations greater than 0.75 (8 monitors). The null findings were consistent
across the three exposure assignment techniques. In a 7-year study of Hong Kong, China
residents, Thach etal. (2010) also reported no association between NO2 and all
cerebrovascular hospital admissions.
5.3.4.2 Summary of Cerebrovascular Disease and Stroke
In summary, the epidemiologic data provide generally inconsistent evidence for a
potential association between ambient NO2 concentrations and risk of hospital admission
for cerebrovascular disease and stroke. Clinical registry studies reported both positive
(Wellenius etal..2012; Andersen etal.. 2010; Vidale etal.. 2010) and null (Turin et al..
2012; Oudin etal.. 2010) associations. Evidence for an association based on
administrative databases came from studies using similar or identical data sets (Chen et
al.. 2014c; Villeneuve et al.. 2012; Szyszkowicz. 2008). A limited number of studies
evaluated potential confounding by PM25 or traffic pollutants (e.g., UFP, CO, BS), and
associations with NCh were not consistently observed in copollutant models.
5-286
-------
Additionally, the majority of the studies of cerebrovascular disease and stroke used
central site monitors to estimate ambient NC>2 exposure. Central site monitors have noted
limitations in capturing the variation in NC>2 (Section 3.4.4.2). and none of the studies of
cerebrovascular disease and stroke reported information on the extent to which
concentrations at central site monitors captured the temporal variation in NC>2 across the
study area.
5.3.5 Decompensation of Heart Failure
Two recent studies found associations between short-term increases in ambient NO2
concentration and hospital admissions or ED visits for heart failure. In the study of seven
Canadian cities described in Table 5-40. Stieb et al. (2009) observed a 5.1% (95% CI:
1.3, 9.2) increase in risk of ED visits for heart failure per 20-ppb increase in 24-h avg
NO2. Unlike the results for the composite endpoint of MI or acute angina, the increased
risk of ED visits for heart failure was not dominated by results from a single city. In
Taipei, Taiwan, Yang (2008) found that risk of hospital admission for heart failure were
associated with NO2 concentrations but only on days where the mean ambient
temperature was >20°C. The association on warm days remained relatively unchanged
after copollutant adjustment for PMio, 862, CO, or Os.
5.3.6 Increased Blood Pressure and Hypertension
5.3.6.1 Epidemiologic Studies
Epidemiologic studies of NO2 and blood pressure (BP) were not available for the 2008
ISA for Oxides of Nitrogen (U.S. EPA. 2008c). but several studies are now available for
review (Table 5-45). There is little evidence from longitudinal studies of the association
between NO2 and BP. A number of longitudinal studies measured BP in subjects in
Beijing before, during, and after the 2008 Beijing Olympics when citywide air pollution
control measures substantially reduced ambient levels of most criteria pollutants. One
study reported that NO2 concentrations during the Olympics were reduced by close to
22% versus the previous month and 13% versus the same period the previous summer
(Huang et al.. 2012a). Other ambient pollutants (except Os) were reduced by similar or
larger amounts. Huang et al. (2012a) measured BP repeatedly in participants with
pre-existing cardiovascular disease in Beijing and found no association between NO2 and
either systolic or diastolic BP. Focusing on healthy young adults, Zhang et al. (2013) and
Rich etal. (2012) each observed no clear association between NO2 and either systolic or
5-287
-------
diastolic BP among participants assessed before, during, and after the 2008 Beijing
Olympics.
Table 5-45 Epidemiologic studies of blood pressure.
Study
fWilliams et al.
(2012a)
tHuanq et al.
(2012a)
tRich et al.
(2012) and
tZhanq et al.
(2013)
tLiuetal.
(2014c)
fCakmak et al.
(2011 a)
Location
Sample Size
Detroit, Ml
n = 65
Beijing, China
n = 40
Beijing, China
n = 125
Sault Ste.
Marie, ON,
Canada
n = 61
Canada
n = 5,604
Mean NO2 (ppb)
24-h avg NO2: 24.0
75th: 28.0
Max: 100.0
2007, Visit 1: 33.8
2007, Visit 2: 26.3
2008, Visit 3: 29.2
2008, Visit 4: 22.9
24-h avg NCb:
Entire study: 27.0
Before: 26.0
During: 13.9
After: 41. 4
1-h max NO2:
Site 1: 3.9
95th: 9.5
Site 2: 5.8
95th: 13.8
24-h avg NCb:
12.6
Exposure
Assessment
Personal
monitoring &
central site
monitor
Central site
monitor
Central site
monitor
Central site
monitor
Central site
monitor
Citywide avg
Selected Effect Estimates3 (95% Cl)
No quantitative results presented.
Change in SBP (mmHg) per IQR (NR)
increase in NO2
30-min: 1.8 (-0.5, 4.0)
2-h: 0.0 (-2.6, 2.6)
12-h: 0.7 (-3.2, 4.6)
24-h: -0.8 (-6.6, 5.0)
Change in DBP (mmHg) per IQR (NR)
increase in NO2
30-min: 1.1 (-0.9,3.0)
2-h: -0.1 (-2.3,2.1)
12-h: 0.7 (-2.1, 4.5)
24-h: 1.5 (-3.4, 6.4)
No quantitative results presented; results
presented graphically. Generally
inconsistent results with SBP: positive
and negative associations across lags.
Generally null and inconsistent
associations with DBP across lags 0-6.
Change in SBP (mmHg)
Lag 0: -1.04 (-4.20, 2.12)
Lag 1:2.08 (-1.36, 5.52)
Change in DBP (mmHg)
Lag 0: -1.32 (-3.88, 1.28)
Lag 1: 1.64 (-1.36, 4.60)
Change in resting SBP (mmHg)
LagO: 1.76(0.35, 3.17)
Change in resting DBP (mmHg)
Lag 0:2.11 (1.12, 3.10)
5-288
-------
Table 5-45 (Continued): Epidemiologic studies of blood pressure.
Location
Study Sample Size
tChuanq et al. Taiwan
(2010) n = 7]578
fChen et al. Taiwan
(2012c) n = 9]238
tChoi et al. Incheon,
(2007) South Korea
n = 10,459
Mean NO2 (ppb)
24-h avg NO2:
22.4
Max: 65.5
24-h avg NCb:
13. 9 to 26.1 across
locations
Max: 34.3 to 49.1
24-h avg NCb:
Warm season: 22.5
75th: 26.9
Max: 49.3
Cool season: 29.2
75th: 34.7
Max: 74.0
Exposure
Assessment Selected Effect Estimates3 (95% Cl)
Nearest central No quantitative results presented for
site monitor NO2.
(within 10 km)
Central site Change in SBP (mmHg)
monitor Lag 0: -Q.81 (-2.16, 0.55)
Citywide avg Lag 0-1 avg: -1.17 (-2.34, -0.01)
Lag 0-2 avg: -4.20 (-5.22, -3.17)
Change in DBP (mmHg)
LagO: 1.03(0.11, 1.95)
Lag 0-1 avg: 1.54(0.75,2.32)
Lag 0-2 avg: -0.01 (-0.71, 0.68)
Pulse Pressure Change
LagO: -2.55 (-3.62, -1.48)
Lag 0-1 avg: -2.09 (-3.02, -1.18)
Lag 0-2 avg: -3.22 (-4.04, -2.40)
Central site Change in SBP (mmHg)
monitor Lag 0: 2.24 (p = 0.002)
Citywide avg Lag 1: 2.40 (p < 0.001)
Lag 2: -0.04 (p = 0.534)
Change in DBP (mmHg)
Lag 0: 2.02 (p = 0.645)
Lag 1: 2.12 (p = 0.016)
Lag 2: -0.04 (p = 0.331)
Change in SBP (mmHg)
Lag 0: 2.06 (p = 0.181)
Lag 1: 2.06 (p = 0.195)
Lag 2: -0.06 (p = 0.223)
Change in DBP (mmHg)
LagO: -0.02 (p = 0.573)
Lag 1 : 2.00 (p = 0.445)
Lag 2: 2.02 (p = 0.445)
avg = average; Cl = confidence interval; DBP = diastolic blood pressure; h = hours; IQR = interquartile range; Max = maximum;
Ml = Michigan; NO2 = nitrogen dioxide; NR = not reported; ON = Ontario; ppb = parts per billion; SBP = systolic blood pressure.
aEffect estimates are standardized to a 20-ppb or 30-ppb increase in NO2 for 24-h avg and 1-h max metrics, respectively.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
In the Detroit, MI area, Williams et al. (2012a) measured BP up to 10 times in each of
65 adult participants and found no association between BP and either total personal or
ambient NO2 concentrations. A strength of this study was the authors' use of personal
5-289
-------
exposure measurements, which are less susceptible to exposure error due to the
variability in NC>2 concentrations and variation in time-activity patterns than central site
monitoring (Section 3.4.4). Similarly, in a randomized cross-over study designed to
examine the cardiovascular effects of exposure to steel plant emissions in Ontario,
Canada, Liu et al. (2014c) measured NO2 exposure near subjects' randomized exposure
location and reported no association between NO2 and either systolic or diastolic BP.
Results of cross-sectional studies of the association between NO2 and BP measured on
the same day or with the NO2 measurement lagged 1-3 days before the BP measurement
have also been mixed. Cakmak et al. (201 la) used cross-sectional data from a national
population-based survey of children and adults in Canada and found a 1.76-mmHg (95%
CI: 0.35, 3.17 mmHg) increase in systolic BP and a 2.11-mmHg (95% CI: 1.12, 3.10)
increase in diastolic BP per 20-ppb increase in 24-h avg NCh on the same day. Chuang et
al. (2010) used cross-sectional data from a national population-based health screening of
adults in Taiwan and reported finding no association between BP and NO2 levels,
although quantitative results were not presented. On the other hand, Chenetal. (2012c)
used cross-sectional data from a different population-based health screening in adults
across six townships in Taiwan and found a 4.20-mmHg decrease (95% CI: -5.22, -3.17)
in systolic BP per 20-ppb increase in 24-h avg NO2 at lag 0-2 day avg and a 1.54 mmHg
increase (95% CI: 0.75, 2.32) in diastolic BP per 20-ppb increase in 24-h avg NO2 at
lag 0-2 day avg. Choi et al. (2007) observed positive associations between NO2
concentrations and systolic BP during the warm and cold seasons at lags 0 and 1, though
the associations with diastolic BP were generally null.
5.3.6.2 Controlled Human Exposure Studies
The 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) reviewed controlled human
studies of cardiac output and BP (Table 5-50); several of these studies also examined
heart rate (HR) as described in Section 5.3.10.1. NO2 exposure generally did not increase
cardiac output or BP in healthy adults or those with COPD. These endpoints have not
been evaluated in recent controlled human exposure studies of NO2.
Cardiac output is the volume of blood pumped out by each of the two ventricles per
minute. It is directly related to HR, as the output of each ventricle is the product of the
HR (beats/minute) and the stroke volume (mL of blood/beat). BP is the product of
cardiac output and vascular resistance. Cardiac output, vascular resistance, and BP
interact moment-to-moment to ensure systemic circulatory demands are met.
Folinsbee et al. (1978) exposed three groups of five young healthy adult males to 600 ppb
NO2 for 2 hours with intermittent exercise. The authors reported no changes in cardiac
5-290
-------
output or BP. Drechsler-Parks (1995) exposed eight older healthy adults to filtered air,
600 ppb NC>2, 450 ppb Os, and NO2 + Os for 2 hours with intermittent exercise. There
was no change in stroke volume or cardiac output following exposure to NC>2 or Os alone
compared to filtered air; however, a decrease in cardiac output was observed following
NC>2 + Os exposure compared to Os and filtered air exposures (p < 0.05). Gong etal.
(2005) reported no change in BP after exposure to 400 ppb NC>2 for 2_hours with
intermittent exercise in volunteers with COPD and healthy volunteers. One controlled
human exposure study examined exposure to higher concentrations of NC>2. Linn et al.
(1985b) reported a small, but statistically significant decrease in BP after exposure to
approximately 4,000 ppb NC>2 for 75 minutes with exercise. In both healthy volunteers
and those with asthma, the mean BP decrease was about 5 mmHg relative to controls.
5.3.6.3 Hospital Admissions and Emergency Department Visits
In contrast with findings for changes in BP, the limited number of available studies report
associations between NO2 and ED visits for hypertension. In Beijing, China, Guo et al.
(2010) found that NC>2 was associated with ED visits for hypertension, and the
association remained relatively unchanged in copollutant models adjusting for PMio or
862. Similarly, in Edmonton, Canada, Szyszkowicz et al. (2012) found that ED visits for
hypertension were positively associated with NC>2 in single-pollutant models. The
association was attenuated in a multipollutant model adjusting for SO2 and PMio, but
results from multipollutant models are difficult to interpret given the potential for
multicollinearity among pollutants. Importantly, neither study evaluated confounding by
traffic-related pollutants.
5.3.6.4 Summary of Blood Pressure and Hypertension
In summary, there is little evidence from available epidemiologic studies to suggest that
short-term exposure to ambient NC>2 is associated with increased BP in the population
overall. There is no evidence of an association from longitudinal studies and mixed
evidence from cross-sectional studies. However, cross-sectional studies have inherent
limitations in the establishment of temporal relationships and are more prone to
confounding by factors that differ between individual participants. Controlled human
exposure studies show no evidence to suggest that short-term exposure to
ambient-relevant concentrations of NC>2 alone alter BP or cardiac output.
5-291
-------
5.3.7 Venous Thromboembolism
Venous thromboembolism is a term that includes both deep vein thrombosis (DVT) and
pulmonary embolism (PE). DVT occurs when a blood clot develops in the deep veins,
most commonly in the lower extremities. A part of the clot can break off and travel to the
lungs, causing a PE, which can be life threatening.
Two recent studies found associations between NO2 and venous thrombosis and/or PE;
however, both studies were small, and neither evaluated potential copollutant
confounding. A study covering the metropolitan region of Santiago, Chile, found a 9.7%
(95%CI:4.1, 15.4) and 8.4% (95% CI: 5.0, 11.8) increase in hospital admissions for
venous thrombosis and PE, respectively, per 20-ppb increase in 24-h avg NC>2
concentrations (Dales etal.. 2010). Spieziaetal. (2014) also examined the association
between ambient air pollution and PE hospital admissions in a small case-control study of
105 adults in Padua, Italy. The authors observed an increase in the risk of unprovoked PE
for subjects who were in the upper tertile of NOx exposure (average exposure for the
month leading up to hospitalization >124 (ig/m3) compared to those in the bottom two
exposure tertiles (OR: 2.35 [95% CI: 0.76, 7.25]).
5.3.8 Aggregated Cardiovascular Effects
Many epidemiologic studies consider the composite endpoint of all cardiovascular
diseases, which typically includes all diseases of the circulatory system (e.g., heart
diseases and cerebrovascular diseases). Most studies reviewed in the 2008 ISA for
Oxides of Nitrogen found positive associations between ambient NO2 concentrations and
risk of hospital admissions or ED visits for all cardiovascular diseases (U.S. EPA. 2008c)
(Figure 5-20 and Table 5-46). However, it was unclear at that time whether these results
truly indicated effects of NO2 or were confounded by other correlated pollutants. Several
additional studies are now available with broadly consistent results, though uncertainty
still remains with regard to potential confounding by PM2 5 and traffic-related pollutants.
Ito etal. (2011) observed that risk of CVD hospital admission was associated with NO2
concentrations at lag 0 in New York City. Results from copollutant models were not
reported. Zheng et al. (2013) and Son etal. (2013) observed seasonal variation in the
strength of association between NO2 and CVD hospital admission in Lanzhou, China and
8 cities in South Korea, respectively. In contrast, Ito etal. (2011) did not find any
seasonal differences in New York, NY. A study in Santiago, Chile reported an increase in
risk of CVD hospital admissions per increase in 24-h avg NO2; the effect estimate
remained relatively unchanged after adjustment for highly correlated PM2 5 (r = 0.87) and
5-292
-------
the traffic-related copollutant CO [r = 0.94; no quantitative results, results presented
graphically; (Franck et al.. 2014)]. In Beijing, China, Guo et al. (2009) reported an
association between ambient NO2 concentrations and risk of CVD hospital admissions at
lag 0 (OR: 1.05 [95% CI: 1.00, 1.11] per 20-ppb increase in 24-h avg NO2), but this
association was attenuated and had a wide 95% CI in copollutant models adjusting for
either PM2 5 (OR: 1.02 [95% CI: 0.96, 1.09]) or SO2 (OR: 1.01 [95% CI: 0.94, 1.08]).
Sarnat etal. (2013b) reported a positive association between NOx concentrations and
CVD ED visits in Atlanta. This study compared the strength of the association across
exposure assessment techniques and estimated larger effects using spatially refined
ambient concentration metrics (AERMOD, Air Pollution Exposure model, and a hybrid
model of background concentrations and AERMOD) in contrast to central site
monitoring data. However, there is uncertainty regarding the extent to which an
association with NOx reflects an association with NO2 (Sections 1.1 and 2.5).
In Shanghai, China, Chen et al. (201 Ob) found a 1.4% (95% CI: -2, 5) increased risk of
hospital admission for CVD per 20-ppb increase in 24-h avg NO2 concentrations (lag
0-l_day avg). This association was robust to additional adjustment for PMio but was
attenuated after adjustment for SO2 [Supplemental Figure S5-5; (U.S. EPA. 2015e)]. A
study in Sao Paulo, Brazil also found a positive association with some evidence that the
association was stronger among patients with a secondary diagnosis of diabetes mellitus
(Filho et al.. 2008). Jevtic et al. (2014) reported a positive association that was robust to
the inclusion of SO2 in a copollutant model in Novi Sad, Serbia. Studies from
Copenhagen, Denmark (Andersen et al.. 2008b): Madrid, Spain (Linares and Diaz. 2010):
Reykjavik, Iceland (Carlsen etal.. 2013); and Taipei, Taiwan (Chan et al.. 2008) reported
null or negative associations between NO2 concentrations and risk of hospital admission
for CVD. A study in Guangzhou, China also found no clear association between NO2 and
CVD hospital admissions, with observed associations alternating between positive and
negative depending on the lags examined (Zhang etal.. 2014).
In summary, evidence reported in the 2008 ISA for Oxides of Nitrogen combined with
recent epidemiologic data continues to consistently show associations between ambient
NO2 concentrations and risk of hospital admission for cardiovascular diseases
(Figure 5-20 and Table 5-46). However, despite generally consistent evidence, a limited
number of studies evaluated potential confounding by correlated copollutants,
particularly PM2 5 and the traffic-related copollutants EC, UFP, and VOCs, resulting in
uncertainty about the independent effect of NO2 on cardiovascular disease hospital
admissions and ED visits (Table 5-46). Further, exposures were represented as ambient
concentrations at central site monitors, which have noted limitations in capturing the
variation in NO2 (Section 3.4.4.2). None of the studies reported information on the extent
5-293
-------
to which concentrations at central site monitors captured the temporal variation in NC>2
across the study area.
5-294
-------
Mean
study Concentration
Quo et al. (2009)
Chen etal. (2010)
Zhang etal. (2014)
Itoetal. (2011)
Son etal. (2013)
Larrieu et al. (2007)
Sarnat etal. (2013)
Poloniecki etal. (1997)
Chang etal. (2005)
Yang et al. (2004)
Linn et al. (2000)
Wong etal. (1999)
von Klot et al. (2005)
Andersen etal. (2008b)
Hi nwood etal. (2006)
Llorca et al. (2005)
Filho et al. (2008)
Atkinson etal. (1999)
Peel et al. (2007)
Tolbertetal. (2007)
Fung et al. (2005)
Jalaludin et al. (2006)
36.3
30.3
29.8
28.7
17.9-26.8
11.9-24.9
6.3-30.0
35
31.5
28.17
28-41
27.3
12.1 -37.2
11
10.3
11.3
9.76
61.1
50.3
45.9
43.2
38.9
23.2
Lag
0
1
0
1
0-1 avg
0
1
2
0
1
0
1
0-1 avg
0-2 avg
1
0-2 avg
0-2 avg
0
0-1 avg
0
0-3 avg
1
0
0
0
1
0-1 avg
0
1
0-1 avg
0
0-2 avg
0-2 avg
0
0-1 avg
0
0-1 avg
0
0-1 avg
Notes
Central site
Background
AERMOD
Hybrid
>20C
25+ C
<25C
1
j^
-6-
K5-
-U-
— J» —
-»£•
1
1 O
P
!o
$
k>
j>
i— + —
?
f
£ 35 yr; Ml survivors
All Ages
Diabetes
No diabetes
Case-crossover
Time-series
< 65 yr
>65yr
>65yr
-
•_
•
-•-
— • —
—
-•-
_^ —
~* —
i
§
k
l-e-
I*
I*
J
•I ,
dt
1
| +
!-•-
24-h
— • —
2.54 (2.27, 2.84) *•
1-h
Simpson et al. (2005a) 16.3-23.7 0
0-1 avg
Ballesteretal.(2006) 12.4-40.5 0-1 avg 24-h avg
61.8 0 1-h max
24-h avg and
1-h max
Morgan etal. (1998)
15
29
24-h avg
1-h max
0.5 0.75 1 1.25
Relative Risk (95% Cl)
1.5
1.75
Note: AERMOD = American Meteorological Society/Environmental Protection Agency Regulatory Model;
AERMOD/BG = AERMOD/background concentration hybrid model; Cl = confidence interval; CS = central site; h = hour;
Ml = myocardial infarction; NO = nitric oxide; NO2 = nitrogen dioxide; yr = years. Black = studies from the 2008 Integrated Science
Assessment for Oxides of Nitrogen, red = recent studies. Circles = NO2, triangles = NO, diamonds = NOX. Relative risks are
standardized to a 20-ppb or 30-ppb increase in NO2 or NO concentrations and 40 ppb or 100 ppb for NOX concentrations for
24-h avg and 1-h max metrics, respectively. Studies are organized first by averaging time, then by recent versus previous studies,
then in descending order of mean concentration (in parts per billion, ppb). Franck et al. (2014) not presented due to lack of
quantitative results.
Figure 5-20 Associations between short-term exposure to oxides of nitrogen
and hospital admissions for all cardiovascular disease.
5-295
-------
Table 5-46 Corresponding effect estimates for hospital admissions for all
cardiovascular disease studies presented in Figure 5-20.
Study
fGuo et al.
(2009)
fChen et al.
(2010b)
tZhanq et al.
(2014)
tltoetal.
(2011)
fSon et al.
(2013)
fLarrieu et al.
(2007)
fSarnat et al.
(2013b)
Location
Beijing, China
Shanghai, China
Guangzhou,
China
New York City,
NY
8 cities, South
Korea
8 French cities
Atlanta, GA
Relative Risk3
(95% Cl)
LagO: 1.05(1.00, 1.11)
Lag 1: 1.03(0.985, 1.09)
Lag 0: 0.997 (0.970, 1.025)
Lag 1: 1.02(0.99, 1.05)
Lag 0-1 avg: 1.014(0.98, 1.05)
LagO: 1.03(0.98, 1.08)
Lag 1: 1.02(0.95, 1.08)
Lag 2: 0.97(0.91, 1.03)
Lag 0: 1.04(1.03, 1.05)
Lag 1: 1.01 (1.00, 1.02)
LagO: 1.04(1.02, 1.06)
Lag 1: 1.00(0.98, 1.01)
Lag 0-1 avg: 1.02(1.00, 1.04)
Lag 0-2 avg NOx
r.pntral <;itp- 1 D1 M DD 1 D"n
Copollutant Examination13
NO2: associations attenuated by PlVh.s or
SO2 adjustment.
Copollutants: PlVh.s and SO2 associations
robust to adjustment for NO2.
NO2 correlations (Pearson r): PIvh.s: 0.67;
SO2: 0.53.
NO2: associations robust to adjustment for
PM-io; attenuated by SO2 adjustment.
Copollutants: PM-io and SO2 associations
attenuated by adjustment for NO2.
NO2 correlations (Pearson r): PM-io: 0.70;
SO2: 0.76.
No copollutant models.
NO2 correlations (Spearman r):
PMio: 0.82; SO2: 0.60.
No copollutant models.
No copollutant models.
NO2 correlations (Pearson r):
PMio: 0.5; SO2: 0.6; CO: 0.7; O3: -0.1.
No copollutant models.
No copollutant models.
Background (km2 weighting): 1.05
(1.01, 1.09)
AERMOD: 1.01 (1.00, 1.02)
Hybrid background-AERMOD: 1.01
(1.00, 1.02)
Poloniecki et
al. (1997)
London, U.K.
Lag 1: 1.02(1.00, 1.04)
No copollutant models analyzed for CVD.
Chang et al.
(2005)
Taipei, Taiwan
Lag 0-2 avg
>20°C: 1.39(1.32, 1.45)
<20°C: 1.23(1.12, 1.37)
NO2: associations robust to adjustment for
PMio, SO2, CO, or 03, with the exception of
PMio on cold days.
5-296
-------
Table 5-46 (Continued): Corresponding effect estimates for hospital admissions
for all cardiovascular disease studies presented in
Figure 5-20.
Study
Yana et al.
(2004)
Linn et al.
(2000)
Wonq et al.
(1999)
von Klot et al.
(2005)
Andersen et
al. (2008b)
Hinwood et al.
(2006)
Llorca et al.
(2005)
fFilhoetal.
(2008)
Atkinson et al.
(1999b)
Peel et al.
(2007)
Location
Kaohsiung,
Taiwan
Los Angeles, CA
Hong Kong,
China
5 European
cities
Copenhagen,
Denmark
Perth, Australia
Torrelavega,
Spain
Sao Paulo,
Brazil
London, U.K.
Atlanta, GA
Relative Risk3
(95% Cl)
Lag 0-2 avg
>25°C: 1.46(1.31, 1.62)
<25°C: 2.54 (2.27, 2.84)
LagO: 1.03(1.02, 1.04)
Lag 0-1 avg: 1.05(1.03, 1.08)
Lag 0: 1.16(1.07, 1.27)
Lag 0-3 avg: 1.00(0.93, 1.10)
Lag 1: 1.08(1.04, 1.13)
LagO
NO2: 1.11 (1.05, 1.17)
NO: 1.13(1.07, 1.19)
Diabetes
LagO: 1.00(1.00, 1.00)
Lag 1: 1.00(0.99, 1.00)
Lag 0-1 avg: 1.00(1.00, 1.00)
No diabetes
Lag 0: 1.00(1.00, 1.00)
Lag 1: 1.00(0.99, 1.00)
Lag 0-1 avg: 1.00(1.00, 1.00)
Lag 0: 1.01 (1.00, 1.02)
Lag 0-2 avg
Case crossover: 1.04 (1.02, 1.06)
Time series: 1.04 (1.02, 1.06)
Copollutant Examination13
NO2: associations robust to adjustment for
PM-io, SO2, CO, or 03 on cold days. Positive
but somewhat attenuated after adjustment
on warm days. Robust with SO2 adjustment.
No copollutant models.
NO2 correlations: PM-io: 0.67 to 0.88;
O3: -0.23 to 0.35; CO: 0.84 to 0.94.
No copollutant models.
NO2: associations robust to adjustment for
PM-io orOs.
No evidence of an association between NO2
and CVD. Copollutant models did not
change the results.
No copollutant models.
No copollutant models.
No copollutant models.
NO2 correlations (Pearson r): PMm 0.68;
SO2: 0.62; CO: 0.58; O3: 0.41.
NO2: association attenuated by adjustment
for BS.
Copollutants: BS robust to adjustment for
NO2.
No copollutant models.
5-297
-------
Table 5-46 (Continued): Corresponding effect estimates for hospital admissions
for all cardiovascular disease studies presented in
Figure 5-20.
Study
Location
Relative Risk3
(95% Cl)
Copollutant Examination13
Tolbert et al. Atlanta, GA
(2007)
LagO-2avg: 1.02(1.01, 1.03)
NO2: association attenuated after
adjustment for CO or PlVh.s TC.
Copollutants: CO and PlVh.s TC
associations robust to adjustment for NO2.
NO2 correlations (Spearman r). CO: 0.70;
PM2.5 TC: 0.65.
Fung et al.
(2005)
Windsor, ON,
Canada
<65 yr
Lag 0: 0.99(0.90, 1.13)
Lag 0-1 avg: 1.04(0.93, 1.16)
>65yr
LagO: 1.02(0.96, 1.07)
Lag 0-1 avg: 1.02(0.98, 1.05)
Copollutant results not reported for NO2.
Jalaludin et al. Sydney,
(2006) Australia
LagO: 1.06(1.02, 1.09)
Lag 0-1 avg: 1.01 (0.98, 1.05)
Lag 1: 1.04(1.01, 1.08)
NO2: associations robust to adjustment for
PM-io, PM2.5, SO2, Os, or BS in adults aged
65 yr and older. Attenuated after CO
adjustment.
Copollutants: CO, PIvh.s, SO2, Os, and BS
associations robust to NO2 adjustment;
PM-io association attenuated.
NO2 correlations: BS: 0.35 to 0.59; PM-io:
0.44 to 0.67; PIvh.s: 0.45 to 0.68; O3: 0.21 to
0.45; CO: 0.55 to 0.71; SO2: 0.52 to 0.56.
Simpson et al.
(2005a)
4 Australian
cities
LagO:
1.07(1.05, 1.09)
Lag 0-1 avg:
1.07(1.05, 1.09)
NO2: associations robust to adjustment for
BS; attenuated, but positive after Os
adjustment.
Copollutants: Os negative association
robust to adjustment for NO2; BS
association attenuated, but positive after
adjustment for NO2.
Ballester et al.
(2006)
Spain
24-h NO2
Lag 0-1 avg:
1.01 (1.00, 1.03)
1-h NO2
LagO:
1.04(0.99, 1.09)
NO2: associations robust to adjustment for
Os; attenuated but positive with CO or SO2
adjustment.
Copollutants: CO, BS, PM-io, SO2, and Os
associations robust to NO2 adjustment.
5-298
-------
Table 5-46 (Continued): Corresponding effect estimates for hospital admissions
for all cardiovascular disease studies presented in
Figure 5-20.
Study
Location
Relative Risk3
(95% Cl)
Copollutant Examination13
Morgan et al.
(1998)
Sydney,
Australia
LagO
24-h NO2:
1.09(1.06, 1.12)
1-h NO2:
1.06(1.04, 1.09)
No copollutant models.
NO2 correlations: Os. -0.09; PM: 0.53.
avg = average; AERMOD = American Meteorological Society/Environmental Protection Agency Regulatory Model; BS = black
smoke; CA = California; Cl = confidence interval; CO = carbon monoxide; CVD = cardiovascular disese; GA = Georgia; h = hours;
NO = nitric oxide; NO2 = nitrogen dioxide; O3 = ozone; ON = Ontario; NY = New York; PM25 = particulate matter with a nominal
mean aerodynamic diameter less than or equal to 2.5 |jm; PMio = particulate matter with a nominal mean aerodynamic diameter
less than or equal to 10 |jm; SO2 = sulfur dioxide; TC = total carbon; U.K. = United Kingdom.
aRelative risks are standardized to a 20-ppb or 30-ppb increase in NO2 or NO or 40-ppb or 100-ppb increase in NOX concentration
for 24-h avg and 1-h max metrics, respectively.
""Relevant relative risks for copollutant models can be found in Supplemental Figures S5-2, S5-3, S5-4, and S5-5 (U.S. EPA,
2015b. c, d, e).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5.3.9 Cardiovascular Mortality
Studies that examined the association between short-term increases in ambient NO2
concentration and cause-specific mortality and that were evaluated in the 2008 ISA for
Oxides of Nitrogen consistently reported positive associations. Across studies, there was
evidence that the magnitude of the NCh-cardiovascular mortality relationship was similar
or slightly larger than that for total mortality. Recent multicity studies as well as a
meta-analysis of studies conducted in Asian cities (Atkinson et al.. 2012) provide
evidence that is consistent with those studies evaluated in the 2008 ISA for Oxides of
Nitrogen (Section 5.4 and Figure 5-23).
The NO2-cardiovascular mortality relationship was further examined in a few studies that
analyzed copollutant models. Importantly, it is difficult to examine whether NO2 is
independently associated with cardiovascular mortality because NO2 often is highly
correlated with other traffic-related pollutants. No study examined potential confounding
by traffic-related pollutants or PM25 In the 17 Chinese cities study (CAPES), Chen et al.
(2012b) found that NO2 risk estimates for cardiovascular mortality were slightly
attenuated but remained positive in copollutant models with PMio or SO2 (6.9% [95% Cl:
3.8, 10.1] for a 20-ppb increase in 24-h avg NO2 concentrations at lag 0-1 avg; 4.6%
[95% Cl: 1.1, 8.1] with PMio; 5.7% [95% Cl: 2.5, 9.0] with SO2). Chen etal. (2013b)
reported similar results when examining stroke mortality in a subset of eight CAPES
5-299
-------
cities [i.e., 5.6% increase in stroke mortality (95% CI: 3.4, 8.0) at lag 0-1 day avg for a
20-ppb increase in 24-h avg NC>2 concentrations. A slight attenuation of the association
was observed in copollutant models with PMio (4.5% [95% CI: 1.8, 7.3]) or SO2 (5.2%
[95% CI: 2.1, 8.3])]. Also, Chiusolo et al. (2011) found evidence that associations
between short-term NCh exposure and cardiovascular mortality remained robust in
copollutant models in a study of 10 Italian cities. In an all-year analysis, a 20-ppb
increase in NC>2 at lag 0-5 avg was associated with a 10.5% (95% CI: 5.9, 14.8) increase
in cardiovascular mortality and a 10.1% (95% CI: 4.0, 16.4) increase adjusted for PMio.
In a warm season analysis (April-September), the NCh effect estimate was 19.2% (95%
CI: 11.4, 27.4) and 18.8% (95% CI: 10.7, 27.5) with adjustment for O3. Overall, the
limited number of studies that have examined the potential confounding effects on the
NC>2-cardiovascular mortality relationship indicate that associations remain relatively
unchanged with adjustment for PMio or SC>2, but it remains difficult to disentangle the
independent effects of NO2 as confounding by more highly traffic-related copollutants
has not been examined.
Of the multicity studies evaluated, only the studies conducted in Italy examined potential
seasonal differences in the NO2-cause-specific mortality relationship (Chiusolo et al..
2011; Bellini et al.. 2007). Additional information with regard to whether there is
evidence of seasonal differences in NO2-cardiovascular mortality associations is provided
by single-city studies conducted in the U.S. (Sacks et al.. 2012; Ito et al.. 2011). In a
study of 15 Italian cities, Bellini et al. (2007) found that risk estimates for cardiovascular
mortality were dramatically increased in the summer from 1.5 to 7.3% for a 20-ppb
increase in 24-h avg NO2 concentrations at lag_0-l avg, respectively, with no evidence of
an association in the winter. These results were corroborated in a study of 10 Italian cities
(Chiusolo et al., 2011). which observed an increase in risk estimates for cardiovascular
mortality in the warm season (i.e., April-September) compared to all-year analyses.
Chiusolo etal. (2011) did not conduct analyses with only the winter season. U.S. studies
conducted in New York, NY (Ito etal.. 2011) and Philadelphia, PA (Sacks etal.. 2012)
do not provide consistent evidence indicating seasonal differences (Section 5.4.6).
Overall, the cardiovascular mortality results from the multicity studies conducted in Italy
are consistent with those observed in the total mortality analyses conducted by Bellini et
al. (2007) and Chiusolo et al. (2011). However, as discussed in Section 5.4.3. studies
conducted in Asian cities observed very different seasonal patterns, and it remains
unclear if the seasonal patterns observed for total mortality would be similar to those
observed for cardiovascular mortality in these cities.
An uncertainty that often arises when examining the relationship between short-term air
pollution exposures and cause-specific mortality is whether analyses that examine
statistical modeling parameters, the lag structure of associations, and the C-R relationship
5-300
-------
provide results that are consistent with those observed for total mortality. In a study
conducted in Philadelphia, PA, Sacks et al. (2012) examined whether the various
modeling approaches to control for both temporal trends/seasonality and weather used in
a number of multicity studies [e.g., National Morbidity, Mortality, and Air Pollution
Study (NMMAPS), Air Pollution and Health: A European Approach (APHEA)]
influence air pollution-cardiovascular mortality associations when using the same data
set. Across models, the authors reported that associations of NO2 with cardiovascular
mortality were relatively consistent, with the percentage increase in cardiovascular
mortality for a 20-ppb increase in 24-h avg NC>2 concentrations ranging from 1.4 to 2.0%.
The results of Sacks etal. (2012) support those of Chen et al. (2013b). which found that
NO2-stroke mortality associations were robust to using 4 to 10 dfper year to control for
temporal trends.
Studies that examined the lag structure of associations for cardiovascular mortality
reported results consistent with those observed for total mortality (Section 5.4.7). In a
study of 10 Italian cities, Chiusolo etal. (2011) reported evidence of an immediate effect
of NO2 at lag 0-1 avg on cardiovascular mortality but also provided evidence for a
prolonged effect due to the magnitude of the association being larger at lag 0-5 avg
(Figure 5-24). These results are consistent with those of Chen et al. (2012b) in the
CAPES study. The authors found the largest effect at single-day lags of 0 and 1 and the
average of lag 0-l_days providing support for an immediate effect of NO2 on
cardiovascular mortality (Figure 5-25). However, when examining longer lags Chen et al.
(2012b) reported that the magnitude of the association was slightly larger for a 0-4_day
lag suggesting a potential prolonged effect. In an analysis of stroke mortality, Chen et al.
(2013b) reported similar results in a subset of eight Chinese cities from CAPES.
To date, analyses detailing the C-R relationship between air pollution and cause-specific
mortality have been limited. In the analysis of eight Chinese cities, Chen etal. (2013b)
also examined the air pollution and stroke mortality C-R relationship. To examine the
assumption of linearity, the authors fit both a linear and spline model to the NCh-stroke
mortality relationship. Chen etal. (2013b) then computed the deviance between the two
models to determine if there was any evidence of nonlinearity. An examination of the
deviance did not indicate that the spline model improved the overall fit of the NCh-stroke
mortality relationship (Figure 5-21).
5-301
-------
(0
•c
o
0)
"S
-------
5.3.10.1 Heart Rate and Heart Rate Variability
HRV provides a noninvasive marker of cardiac autonomic nervous system function. The
rhythmic variation in the intervals between heart beats can be quantified in either the time
domain or the frequency domain (TFESC and NASPE. 1996). Common time-domain
measures of HRV include the standard deviation of all normal-to-normal intervals
(SDNN, an index of total HRV) and the root-mean-square of successive differences
(rMSSD, an index influenced mainly by the parasympathetic nervous system). In the
frequency domain, HRV is usually divided into the high frequency (HF) and low
frequency (LF) components, as well as the ratio of the LF to HF components (LF/HF)
(TFESC and NASPE. 1996). Decreases in indices of HRV have been associated with
increased risk of cardiovascular events in prospective cohort studies (La Rovere et al..
2003; Kikuva et al.. 2000; Tsuiietal.. 1996; TsuiietaL 1994).
Epidemiologic Studies
The 2008 ISA for Oxides of Nitrogen reported that there was insufficient evidence to
determine whether exposure to oxides of nitrogen was associated with changes in cardiac
autonomic control as assessed by indices of HRV (U.S. EPA. 2008c). Additional studies
are now available for review (Table 5-47) that provide evidence for an association
between exposure to NCh and HRV among those with pre-existing disease but not in
healthy individuals.
5-303
-------
Table 5-47 Epidemiologic studies of heart rate/heart rate variability.
Study
Timonen et
al. (2006)
fZanobetti et
al. (2010)
tBartell et al.
(2013)
Location and
Sample Size
Amsterdam, the
Netherlands; Erfurt,
Germany; Helsinki,
Finland
n = 131
Boston, MA
n - 4fi
(aged 43-75 yr)
Los Angeles, CA
n = 50
Pre-Existing Mean NO2 Exposure
Condition ppb Assessment
Coronary artery 24-h avg NO2 Central monitor
disease Amsterdam: 22.7
Erfurt: 15.4
Helsinki: 16.5
Coronary artery 2-h avg NO2 Central monitor
disease 50th: 21 Citywide avg
75th: 27
95th: 36
72-h avg NO2
50th: 21
75th: 25
95th: 31
Coronary artery 24-h avg NOx: 42.3 Monitors on
disease Max' 183 7 trailers parked
at each of
4 retirement
communities
Selected Effect Estimates3 (95% Cl)
SDNN (msec) LF/HF (% change)
Lag 0: -1.05 (-3.50, 1.39) Lag 0: -3.01 (-15.4,
Lag 1: -1.28 (-3.98, 1.43) Lag 1: -16.5 (-30.1,
Lag 2: -3.01 (-5.94, -0.11) Lag 2: -17.7 (-32.0,
Lag 3: -0.68 (-3.42, 2.07) Lag 3: -1.88 (-15.4,
Lag 0-4: -4.59 (-9.32, 0.15) Lag 0-4: -26.0 (-50.
HF (% change)
Lag 0:4.51 (-33.5, 18.4)
Lag 1:4.51 (-10.2, 19.6)
Lag 2: 1.88 (-13.9, 18.1)
Lag 3: -1.88 (-16.9, 13.2)
Lag 0-4: 4.51 (-22.2, 30.8)
HF (% change)
2-h: -6.7 (-10.8, -2.5) per 11-ppb increase
120-h: -9.4 (-14.1, -4.4) per6-ppb increase
All other results presented graphically, no quantitative
No quantitative results presented; results presented
graphically.
Generally null associations between NOx and SDNN
medication use in participants taking and not taking
acetylcholine esterase inhibitors.
9.77)
-3.01)
-3.01)
11.7)
1, -1.88)
results.
5-304
-------
Table 5-47 (Continued): Epidemiologic studies of heart rate/heart rate variability.
Study
Location and
Sample Size
Pre-Existing
Condition
Mean NO2
ppb
Exposure
Assessment
Selected Effect Estimates3 (95% Cl)
tBarclay et
al. (2009)
Aberdeen,
Scotland, U.K.
n = 132
Stable heart
failure
24-havgNO2: 30.1
24-havgNO: 14.7
Central monitor
HR(bpm)
NO2: 0.40(0.00,0.80)
NO: 0.35 (-0.04, 0.74)
SDNN (msec)
NO2: 0.62 (-0.59, 1.83)
NO: 0.61 (-0.56, 1.78)
SDANN (msec)
NO2: 0.51 (-0.87, 1.89)
NO: 0.57 (-0.77, 1.91)
rMSSD (msec)
NO2: 0.40(0.00,0.80)
NO: 0.35 (-0.04, 0.74)
pNNSO(%)
NO2: 1.57 (-3.85, 6.99)
NO: 0.91 (-4.35,6.17)
LF power (units NR)
NO2: 2.35 (-1.05, 5.76)
NO: 1.94 (-1.33, 5.21)
LF normalized (units NR)
NO2: -0.86 (-2.82, 1.10)
NO: -0.18 (-2.07, 1.70)
HF power (units NR)
NO2: 3.37 (-1.17, 7.90)
NO: 2.90 (-1.46, 7.25)
HF normalized (units NR)
NO2: 0.72 (-1.55, 3.00)
NO: 1.41 (-0.78, 3.56)
LF/HF ratio
NO2: -1.09 (-3.93, 1.75)
NO: -1.05 (-3.78, 1.67)
tGoldberq et
al. (2008)
tSuh and
Zanobetti
(201 Oa)
Montreal, QC,
Canada
n = 31
Atlanta, GA
n-30
Stable heart 24-h avg NO2
failure 17 g
Max: 54.1
Ml or COPD 24-h avg NO2
Ambient: 17.1
Personal: 11.6
Central monitor
citywide avg
Central monitor
Citywide avg
Personal
Pulse rate (mean difference)
Lag 0: -0.07 (-0.09, 0.80)
Lag 1:0.78 (-0.14, 1.71)
Lag 0-2: 0.99 (-0.34, 2.32)
SDNN (% change)
Ambient: -0.64 (-11.1, 10.4)
Personal: -3.48 (-10.7, 3.9)
HF (% change)
Ambient:
-1.49 (-37.1, 41
.3)
rMSSD (% change)
Ambient: -6.60 (-30.6, 20.9)
Personal: -14.5 (-29.9, 1.70)
pNNSO (% change)
Ambient: 0.30 (-38.3, 47.4)
Personal: -32.3 (-56.5, -5.65)
Personal: -21.4 (-44.9, 4.48)
LF/HF (% change)
Ambient: 13.7 (-4.11, 33.1)
Personal: 9.7 (-2.34, 22.2)
5-305
-------
Table 5-47 (Continued): Epidemiologic studies of heart rate/heart rate variability.
Study
Location and
Sample Size
Pre-Existing
Condition
Mean NO2
ppb
Exposure
Assessment
Selected Effect Estimates3 (95% Cl)
tHuanq et al. Beijing, China
n=40
CVD
1-h max NO2
2007, Visit 1:33.8
2007, Visit 2: 26.3
2008, Visit 3: 29.2
2008, Visit 4: 22.9
Central monitor
SDNN (% change)
1-h: -1.9 (-3.4, -0.3)
4-h: -3.9 (-5.7, -2.2)
12-h: -3.6 (-5.5, -1.6)
rMSSD (% change)
1-h: 1.4 (-1.1, 3.9)
4-h: -2.2 (-5.7, 1.5)
12-h: -2.2 (-6.1, 2.0)
LF (% change)
1-h: -5.4 (-9.3, -1.4)
4-h: -8.9 (-13.2, -4.3)
12-h: -7.9 (-12.8, -2.8)
HF (% change)
1-h: -3.5 (-8.2, 1.4)
4-h: -5.1 (-11.0, 1.3)
12-h: -3.7 (-10.4, 3.5)
fWilliams et
al. (2012b)
fLaumbach
etal. (2010)
tPeel et al.
(2011)
tChuanq et
al. (2007a)b
tJia et al.
(2011)
Detroit, Ml
n - 65
New Brunswick, NJ
n - 91
Atlanta, GA
n= 4,277
Taipei, Taiwan
n - 76
Beijing, China
n=20
CVD risk factors
(hypertension,
hyperlipidemia,
diabetes)
Diabetes
Healthy infants
Healthy
Healthy
24-h avg NO2
24.0
75th: 28.0
Max: 100.0
NO2
50th: 25.9
75th: 32.8
Max: 61.1
1-h max NO2
41.7
90th: 65.6
Max: 109.2
24-h avg NO2
17.3
Max: 53.1
24-h avg NOx
35.0
Personal HR (bpm)
monitor -2.95 (-4.82, -0.80)
In-vehicle HF (% change)
monitor -11.9 (-105, 80.8)
LF/HF ratio (% change)
-107 (-298, 83.4)
Central monitor Bradycardia (OR)
1.04(1.00, 1.08)
Central monitor "We found no associations between HRV indices and NO2."
No quantitative results presented.
Central monitor "No significant effects are found between daily average... NOx
on HRV indices."
No quantitative results presented.
5-306
-------
Table 5-47 (Continued): Epidemiologic studies of heart rate/heart rate variability.
Study
Location and
Sample Size
Pre-Existing
Condition
Mean NO2
ppb
Exposure
Assessment
Selected Effect Estimates3 (95% Cl)
tParket al.
(2010)
Baltimore, MD;
Chicago, IL;
Forsyth County, NC;
Los Angeles, CA;
New York, NY;
St. Paul, MN
n = 5,465
Healthy
24-h avg NO2
Lag 0-1: 23.5
Central monitor
Citywide avg
"There were no significant associations of HRV with gaseous
pollutants (data not shown)."
No quantitative results presented.
tChuanq et
al. (2007b)
Taipei, Taiwan
n=46
Healthy
1-h max NO2 Central monitor "...NO2...exposures were not associated with any HRV indices
384 Avg of monitors in our study participants (data not shown)."
within 1 km of No quantitative results presented.
residence
tMin et al.
(2008)
Taein Island,
South Korea
n = 1,349
Healthy
24-h avg NO2 Central monitor SDNN (% change)
24 6-h:-2.45 (-6.28, 1.53)
75th: 30 9-h:-3.89 (-8.31, 0.71)
Max: 119 12-h: -3.81 (-8.75, 1.34)
24-h: -1.72 (-6.71, 3.51)
48-h: 2.93 (-2.33, 8.42)
72-h: 1.20 (-3.81, 6.42)
LF (% change)
6-h: -8.61 (-16.9, 0.31)
9-h: -12.2 (-21.5, -2.11)
12-h: -12.3 (-22.6, -0.88)
24-h: -5.71 (-16.6, 6.33)
48-h: 3.69 (-8.22, 16.9)
72-h: 5.84 (-6.19, 18.5)
HF (% change)
6-h: -1.08 (-10.8, 9.47)
9-h: -3.31 (-14.3, 8.88)
12-h: -2.38 (-14.7, 11.5)
24-h: -4.53 (-16.6, 8.94)
48-h: 4.42 (-8.72, 19.1)
72-h: 4.18 (-8.52, 18.4)
5-307
-------
Table 5-47 (Continued): Epidemiologic studies of heart rate/heart rate variability.
Study
fWeichenthal
etal. (201 1)b
Location and Pre-Existing
Sample Size Condition
Ottawa, ON, Canada Healthy
n -49
Mean NO2
ppb
1-h max NO2
4 8
Exposure
Assessment Selected Effect Estimates3 (95% Cl)
Central monitor ALF (msec2)
1-h- -539 f-9879 1807^
ASDNN (msec)
1-h- -18 8 f-113 79 m
2-h: 12.0 (-2468, 2490)
3-h: 578 (-2055, 3218)
4-h: -398 (-3533, 2025)
AHF (ms2)
1-h: -420 (-1785, 953)
2-h: -488 (-1612, 638)
3-h: -24.0 (-1020, 975)
4-h: -248 (-1418, 923)
ALF:HF
1-h: 5.70 (-2.10, 13.5)
2-h: 10.5(2.63, 18.8)
3-h: 12.8(4.20,21.8)
4-h: 7.50 (-1.80, 17.3)
2-h: -75.0 (-150, -2.55)
3-h: -39.8 (-120, 40.5)
4-h: -12.0 (-82.5, 61.5)
ArMSSD (msec)
1-h: -12.0 (-48.8, 24.8)
2-h: -12.0 (-41.3, 17.3)
3-h: 2.33 (-30.0, 34.5)
4-h: -2.10 (-33.0, 29.3)
ApNNSO (%)
1-h: -3.30 (-31.5, 24.8)
2-h: -8.25 (-33.0, 15.8)
3-h: -3.23 (-29.3, 22.5)
4-h: 1.28 (-26.3, 29.3)
fShields et
al. (2013)
Mexico City, Mexico Healthy
n ~ 16
1-h max NO2
130
In-vehicle
Monitor
LF (% change)
-0.69 (-1.91, 0.57)
HF (% change)
-0.24 (-1.80, 1.47)
LF/HF (% change)
-0.45 (-1.53, 0.64)
SDNN (% change)
-1.03 (-1.55, -0.49)
5-308
-------
Table 5-47 (Continued): Epidemiologic studies of heart rate/heart rate variability.
Study
fRich et al.
(2012)b
tZhanq et al.
(201 3)b
Location and
Sample Size
Beijing, China
n = 12
Beijing, China
n = 125
Pre-Existing Mean NO2
Condition ppb
Healthy 24-h avg NO2
Entire study: 27.0
Before: 26.0
During: 13.9
After: 41. 4
Healthy 24-h avg NO2
Before: 25.6
During: 14.6
After: 41. 4
Exposure
Assessment Selected Effect Estimates3 (95% Cl)
Central monitor No quantitative results presented; results presented
graphically. Positive, but statistically nonsignificant increase in
heart rate, generally consistent across lags from 0 to 6.
Central monitor No quantitative results presented; results presented
graphically. Statistically significant decreases in SDNN and
rMSSD in the early lags (0 to 1); measurable but statistically
nonsignificant decreases across lags 2 and 3; and generally
null associations in lags 4, 5, and 6.
avg = average; bpm = beats per minute; CA = California; Cl = confidence interval; COPD = chronic obstructive pulmonary disease; CVD = cardiovascular disease; GA = Georgia;
HF = high frequency; HR = heart rate; HRV = heart rate variability; LF = low frequency; LF/HF = LF to HF components; IL = Illinois; max = maximum; MD = Maryland;
Ml = myocardial infarction or Michigan; MN = Minnesota; NJ = New Jersey; NC = North Carolina; NO = nitric oxide; NO2 = nitrogen dioxide; NOX = sum of NO and NO2; NY = New
York; OR = odds ratio; pNNSO = percentage of pairs of successive normal sinus intervals that exceeds 50 milliseconds divided by the total number of successive pairs of normal
sinus intervals; ppb = parts per billion; rMSSD = root-mean-square of successive differences; SDANN = standard deviation of average normal-to-normal intervals; SDNN = standard
deviation of all normal-to-normal intervals; U.K. = United Kingdom; U.S. = United States.
"•Relative risks are standardized to a 20-ppb or 30-ppb increase in NO2 or NO or 40-ppb or 100-ppb increase in NOX concentration for 24-h avg and 1-h max metrics, respectively.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-309
-------
The multicountry ULTRA study assessed the longitudinal association between ambient
pollution and HRV among older adults with stable coronary artery disease in Amsterdam,
the Netherlands; Erfurt, Germany; and Helsinki, Finland (Timonen et al., 2006). In each
participant, HRV was assessed multiple times over a 6-month period, resulting in a total
of 1,266 repeated measures. Pooling results across the three cities, the authors found a
3.01 msec (95% CI: -5.94, -0.11) decrease in SDNN and a 17.67% (95% CI: -31.95,
-3.01) decrease in LF/HF associated with a 20-ppb increase in 24-hour average NO2
concentrations at Lag Day_2. The magnitudes of these associations were somewhat larger
in relation to the 5-day moving average of NO2. The authors report that these effects were
robust to adjustment for other pollutants in copollutant models, including UFP, PM2 5, or
CO, but detailed results were not provided. These results were reportedly similar in men
and women and after exclusion of those exposed to environmental tobacco smoke at
home. Most associations with HF were positive.
Huang et al. (2012a) measured HRV repeatedly in participants with pre-existing
cardiovascular disease in Beijing in the summer of 2007 and again in the summer of
2008, including one measurement period during the 2008 Beijing Olympics when
citywide air pollution control measures substantially reduced ambient concentrations of
most criteria pollutants as described in more detail in Section 5.3.6.1. In single-pollutant
models, an unspecified IQR increase in 1-h max NO2 was associated with a 3.6%
decrease (95% CI: -5.5, -1.6) in SDNN, and a 7.9% decrease (95% CI: -12.8, -2.8) in
LF. The association with SDNN was stronger among those with a higher C-reactive
protein (CRP), women, and those without a history of diabetes, but body mass index
(BMI) did not appear to modify the association. Richetal. (2012) also examined the
association between heart rate and NO2 concentrations before, during and after the 2008
Beijing Olympics. The authors observed increases in heart rate that were generally
consistent in magnitude across lags from 0 to 6 days. In expanded results from the same
protocol, Zhang etal. (2013) reported that NO2 was inversely associated with SDNN and
rMSSD, with stronger associations in the earlier lags (0 to 3). The HR association with
NO2 was somewhat attenuated, but still positive after adjustment for PM2 5, CO, SO2, or
OC, and no longer positive after adjustment for EC. The decrements in rMSSD and
SDNN associated with increased ambient NO2 remained relatively unchanged after
adjustment for PM25, CO, SO2, OC, or EC.
Some studies (Laumbach et al.. 2010; Suh and Zanobetti. 201 Ob) assessed personal
exposures, which tend to reduce uncertainty in the NO2 exposure estimate when
compared to the use of citywide averages (Section 3.4.4.2). Suh and Zanobetti (2010a)
examined people that had either recently experienced an MI or had COPD. Same-day
total personal exposures to NO2 were associated with decreased HRV. Decreases in
pNN50 (proportion of pairs of successive normal sinus intervals exceeds 50 milliseconds
5-310
-------
divided by the total number of successive pairs of normal sinus intervals) were the largest
among the individuals with COPD, while NCh-associated decrements in HF were the
largest among individuals with a recent MI, but were less precise when all individuals or
individuals with COPD were included in the analysis. Associations were more
pronounced when examining personal as opposed to ambient measures of NCh.
Copollutant confounding was not assessed. Laumbach et al. (2010) studied the effects of
in-vehicle exposure to traffic-related pollutants among a group of individuals with
diabetes. The authors did not observe any strong evidence of an association between HF
HRV and in-vehicle exposure to NC>2. Weichenthal et al. (2011) measured NO2 at central
site monitors and carried out a cross-over trial with 42 healthy adults who cycled for
IJiour on high- and low-traffic routes as well as indoors. Results were inconsistent
among the many measures of HRV examined, but mean concentrations of NC>2 were
associated with decreases in SDNN and increases in LF/HF.
In a repeated-measures study of Boston-area patients with clinically significant coronary
artery disease, Zanobetti et al. (2010) found that HF was inversely associated with
ambient NO2 concentrations. This association remained robust after adjustment for PM2 5
in a copollutant model. Among a population reporting a substantial prevalence of
cardiovascular risk factors (i.e., hypertension, diabetes, hyperlipidemia), Williams et al.
(2012a) observed a strong association between NO2 concentrations and reduced heart
rate. On the other hand, Barclay et al. (2009) reported no association between NO2 or NO
and indices of HRV in a repeated-measures study of nonsmoking patients with stable
heart failure. (Bartell etal.. 2013) observed generally null associations between NO2 and
SDNN medication use in retirement residents with coronary artery disease in the greater
Los Angeles area. Also, Goldberg et al. (2008) followed 31 Montreal-area participants
with stable heart failure for 2_months and found no association between pulse rate and
NO2 concentrations.
Infants are potentially at greater risk of pollution-related health effects (AAP. 2004). Peel
etal. (2011) examined data from 4,277 Atlanta-area infants prescribed home
cardiorespiratory monitors and observed a slightly elevated risk of bradycardia (OR: 1.04
[95% CI: 1.00, 1.08]) per 30-ppb increase in 1-h maxNO2 concentrations averaged over
the previous 2 days and measured at a central site monitor. The clinical or public health
significance of this finding is unclear.
The majority of the above studies focused on infants or participants with a documented
history of heart disease, with the exception of the Beijing Olympics studies (Zhang etal..
2013; Rich etal.. 2012). In contrast to the pre-existing disease studies, there is little
evidence that HRV is associated with NO2 concentrations in healthy participants. For
example, a repeated-measures study of young healthy participants in Taipei, Taiwan
5-311
-------
found no association between NO2 and HRV indices (Chuang et al., 2007a). Another
repeated-measures study in Mexico City observed small decrements in SDNN associated
with increases in NO2, but no association between NO2 and LF/HF (Shields et al., 2013).
In Beijing, Jiaet al. (2011) assessed HRV two times in each of 20 healthy participants
and reported no association between oxides of nitrogen and FIRV. However, this study
was quite small, and detailed results were not shown.
Cross-sectional analyses of populations with or without a history of heart disease have
also tended to yield null results. In a cross-sectional analysis of 5,465 participants, ages
45-84 years, from the multicity Multiethnic Study of Atherosclerosis, Parket al. (2010)
found no association between NO2 concentrations and indices of HRV. A cross-sectional
study from Taipei also observed no association between NO2 and HRV among 46 older
adults with cardiovascular disease risk factors (Chuang et al.. 2007b). A cross-sectional
study of 1,349 healthy participants in Taein Island, South Korea by Min et al. (2008).
found that NO2 was associated with decreases in the LF component of HRV, but not with
changes in SDNN or the HF component.
In summary, current evidence suggests that among participants with pre-existing or
elevated risk for cardiovascular disease, ambient NO2 concentrations are associated with
alterations in cardiac autonomic control as assessed by indices of HRV; however,
evidence for differential effects between populations with and without pre-existing
diseases and conditions is limited. In this specific subgroup of the population, NCh seems
to be associated with changes in HRV, which is consistent with relative increases in
sympathetic nervous system activity and/or decreases in parasympathetic nervous system
activity. In contrast, this association has not been commonly apparent among healthier
participants. In the two studies that examined copollutant models with PM2 5 or a
traffic-related pollutant among UFP, CO, or OC, NO2 associations in the pre-existing
disease populations generally persisted. However, inference about the independent effect
of NO2 is limited given the high correlations among pollutants and potential differential
exposure measurement error resulting from use of central site ambient pollutant
measurements.
Controlled Human Exposure Studies
Controlled human exposure studies evaluating HRV were not available for review in the
2008 ISA for Oxides of Nitrogen; since then, two studies are available (Table 5-50).
Huang et al. (2012b) evaluated changes in various HRV parameters following NO2
exposure in healthy adult volunteers performing intermittent exercise. Exposure to
500 ppb NO2 did not alter HRV time domain intervals. LF/HF slightly increased,
although this change was not statistically significant. The authors reported an 11.6 and
5-312
-------
13% decrease in the HF domain normalized for heart rate (HFn) 1 and 18 hours after
exposure, respectively. Combined exposure to NCh and PM2 5 CAPs increased LF/HF
(1 hour;/? = 0.062), as well as the low frequency domain normalized for heart rate
(1 hour;/? = 0.021) and cardiac t-wave amplitude (18 hour;/? = 0.057). CAPs exposure
alone did not induce such changes. Epidemiologic studies found NCh-associated
decreases in HRV primarily in adults with or at risk for cardiovascular disease. However,
a recent study in resting adults with stable coronary heart disease and impaired left
ventricular systolic function showed no statistically significant changes in HRV after
exposure to 400 ppb NO2 for 1 hour while seated (Scaife et al., 2012); however, the study
had only 75% power to detect significant differences in the HF domain of 50% or less.
The few studies reviewed in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c.
1993a) reported mixed effects of NO2 exposure on HR; a recent study shows no effect.
Folinsbee et al. (1978) and Drechsler-Parks (1995) exposed healthy adult males and
healthy older adults, respectively, to approximately 600 ppb NC>2 for 2 hours and reported
no changes in HR. Changes in HR were also examined in potentially at-risk populations
exposed to NCh. Exposure to 400 ppb NC>2 did not alter HR in adults with coronary heart
disease (Scaife et al.. 2012) and resulted in a statistically nonsignificant increase in adults
with COPD and healthy volunteers (Gong etal.. 2005). Among healthy volunteers and
those with asthma, NO2 exposure resulted in no change in HR (Linn etal.. 1985a).
In summary, there is limited, inconsistent evidence from controlled human exposure
studies to suggest NO2 alone or in combination with CAPs exposure during exercise
alters HRV. Additionally, there appears to be no evidence from controlled human
exposure studies that NO2 exposure alters HR.
lexicological Studies
Toxicology studies examining HRV changes were not available for review in the 2008
ISA for Oxides of Nitrogen. Similar to controlled human exposure studies, a recent study
in rats found mixed evidence for changes in HR and HRV (Table 5-51). Ramos-Bonilla
etal. (2010) examined body weight, HR, and HRV, following exposure of aged inbred
mice to an ambient mixture consisting of PM, CO, and NO2. Animals were exposed to
either filtered or unfiltered outdoor Baltimore air for 6 hours daily for 40 weekdays. In
multipollutant models, statistically significant declines in HR were associated with NO2
at lag 3 and the 7-day cumulative concentration with adjustment for PM and CO.
However, HRV changes were not associated with NO2 exposure. The independent effects
of each pollutant are difficult to distinguish in multipollutant models because of potential
multicollinearity among pollutants.
5-313
-------
5.3.10.2 QT-lnterval Duration
The QT interval provides an ECG marker of ventricular repolarization. Prolongation and
increased variability of the QT interval is associated with increased risk of
life-threatening ventricular arrhythmias. Consistent with the weak epidemiologic
evidence for associations of NO2 exposure with arrhythmias (Section 5.3.3). the limited
evidence from epidemiologic and controlled human exposure studies does not clearly
indicate an effect of short-term NO2 exposure on markers of ventricular repolarization.
In the Normative Aging Study, Bajaet al. (2010) found imprecise associations between
heart-rate-corrected QT interval (QTc) and 10-hour moving average of NCh
concentrations among older, generally white men but observed associations with NC>2
concentrations at lags 3 and 4 hours (longer lags or moving averages were not
considered) (Table 5-48). The only study from the 2008 ISA for Oxides of Nitrogen (U.S.
EPA. 2008c) available for comparison found that 24=h avg NO2 concentrations were
positively associated with increased QTc duration, but this association was imprecise
(i.e., had wide confidence intervals), and the 6-hour moving average of NCh was not
associated with an increase in QTc duration (Henneberger et al.. 2005).
Table 5-48 Epidemiologic studies of QT-interval duration.
Study
Location
Sample Size
Mean
Concentration
ppb
Exposure
Assessment
Selected Effect Estimates3
(95% Cl)
tBaiaetal. (2010) Boston, MA
n = 580
1-h max NO2
19 ppb during ECG
monitoring
21 ppb 10 h before
monitoring
Central site Change in QTc (msec)
10-hlag: 5.91 (-2.03, 13.85)
4-h lag: 6.28 (-0.02, 12.55)
Henneberqer et al. Erfurt, Germany
(2QQ5J n = 56
24-h avg NCb: 18.2 Citywide avg
75th: 22.6
Max: 36.4
24-h avg NO: 19.4
75th: 24.2
Max: 110.1
QTc (msec)
NO2, lag 6-11 h: 9.77(2.23, 17.3)
T-wave complexity (%)
NO, lag 0-23: 0.15(0.02, 0.28)
T-wave amplitude (uV)
NO, lag 0-5 h: -2.10 (-4.16, -0.03)
avg = average; Cl = confidence interval; ECG = electrocardiographic; MA = Massachusetts; max = maximum; NO = nitric oxide;
NO2 = nitrogen dioxide; QTc = corrected QT interval.
aEffect estimates are standardized to a 20-ppb or 30-ppb increase in NO2 or NO for 24-h avg and 1-h max metrics, respectively.
fStudy published since the 2008 ISA for Oxides of Nitrogen.
5-314
-------
There were no controlled human exposure studies evaluating changes in the QT interval
available for the 2008 ISA for Oxides of Nitrogen, and the single recent study found
NO2-induced changes in QT interval that are in the opposite direction as that associated
with arrhythmias. Huang et al. (2012b) found a small (quantitative results not reported)
decrease in QTc at 1 and 18 hours after a 2-hour exposure to 500 ppb NO2 in healthy
exercising adults (Table 5-50). NCh exposure also induced a 29.9% decrease (p = 0.001)
in the QT variability index (QTVI). However, when volunteers were exposed to both
PM2 5 and NCh, the QTVI synergistically increased.
5.3.10.3 Vascular Reactivity
The vascular endothelium plays a fundamental role in the maintenance of vascular tone
that is involved in the regulation of blood pressure and blood flow. In a controlled human
exposure study, Langrish et al. (2010) examined the effects of NO2 on vascular
endothelial tone and fibrinolytic function. In a random crossover double-blind study,
healthy male volunteers were exposed to 4,000 ppb of NCh for 1 hour with intermittent
exercise. This study employed infusion of endothelial-dependent vasodilators, bradykinin
and acetylcholine, and endothelial-independent vasodilators, sodium nitroprusside and
verapamil, to examine vascular endothelial tone. The results demonstrated that NO2 did
not attenuate the vasodilator response to these vasoactive agents.
Epidemiologic studies provide inconsistent evidence regarding a potential association
between NCh and vascular function. In the EPA's Detroit Exposure and Aerosol
Research Study, Williams et al. (2012a) found that total personal NO2 concentrations
were associated with inconsistent changes in brachial artery diameter (positive
association at lag 1 and negative association at lag 2) an d increases (i.e., presumably
beneficial) in flow-mediated dilatation. No associations were observed with ambient
measures of NC>2. Ljungman et al. (2014) reported no consistent associations between 1-,
2-, 3-, 5-, and 7-day moving averages of NOx and peripheral arterial tonometry ratio in
the Offspring and Third Generation Cohorts of the Framingham Heart Study.
In summary, the available controlled human exposure and epidemiologic evidence is
limited and inconsistent, and therefore does not support the presence of an association
between ambient NO2 concentrations and vascular reactivity.
5.3.10.4 Blood Biomarkers of Cardiovascular Effects
Several epidemiologic and toxicological studies have explored the potential relationship
between NO2 and biomarkers of cardiovascular risk. In particular, markers of
5-315
-------
inflammation, cell adhesion, coagulation, and thrombosis have been evaluated in a
number of epidemiologic studies published since the 2008 ISA for Oxides of Nitrogen
(U.S. EPA. 2008c). These biomarkers also have been examined in controlled human
exposure and animal toxicological studies.
Epidemiologic Studies
Levels of some circulating systemic inflammatory markers appear to be related to NC>2
concentrations among participants with a history of heart disease (Table 5-49). Delfino et
al. (2008b) followed nonsmoking elderly subjects with a history of coronary artery
disease living in retirement communities in Los Angeles, CA and measured plasma
biomarkers weekly over a 12-week period. The authors observed that indoor and/or
outdoor NO2 concentrations measured at the retirement homes were associated with
increases IL-6 and the soluble tumor necrosis factor a receptor II (sTNFa-RII), markers
of systemic inflammation, but not associated with a number of other biomarkers of
inflammation and vascular injury including CRP, P-selectin, D-dimer, TNF-a, soluble
intercellular adhesion molecule-1 (sICAM-1), or soluble vascular adhesion molecule-1
(sVCAM-1). In subsequent analysis of overlapping populations, Delfino et al. (2009) and
Delfino et al. (2010) found that NCh and NOx were both associated with circulating
levels of IL-6. Delfino et al. (2009) also observed positive associations with P-selectin,
TNF-RII, and CRP. Working with the same study population, Wittkopp et al. (2013) also
found an association between NOx concentrations and increases in IL-6 and TNF-a, but
only for participants with mitochondrial haplogroup H, which has been linked to
oxidative damage and increased risk of age-related diseases. Similarly, Ljungman et al.
(2009) repeatedly measured plasma IL-6 in 955 MI survivors from six European cities,
and found that NO2 was associated with increased levels of IL-6, and that the strength of
the association varied in individuals with specific variants of inflammatory genes.
However, in studies conducted among patients with stable chronic heart failure, no
associations were observed between any biomarkers (including hematological markers
and markers of inflammation) and NO2 concentrations (Barclay et al.. 2009; Wellenius et
al.. 2007). None of these studies examined potential confounding by traffic-related
copollutants.
In Augsburg, Germany, Briiske etal. (2011) measured lipoprotein-associated
phospholipase A2 (Lp-PLA2), a marker of vascular inflammation and an independent
predictor of coronary heart disease events and stroke, up to six times in 200 participants
with a history of myocardial infarction. They found that Lp-PLA2 was associated with
both NO and NO2. However, the association was negative at short lags and positive at
longer lags, making interpretation of these results difficult.
5-316
-------
Table 5-49 Epidemiologic studies of biomarkers of cardiovascular effects.
Study
Location
Sample Size
Pre-
Existing
Condition
Mean NO2 ppb
Exposure
Assessment
Selected Effect Estimates3 (95% Cl)
tPelfinoetal. (2008b)
Los Angeles,
CA
n = 29
Coronary
artery
disease
24-h avg NO2
Outdoor:
33.1
Max: 59.8
Indoor:
32.3
Max: 53.5
Indoor and
outdoor home
measurements
Outdoor:
CRP (ng/mL)
LagO: 1,125 (-314, 2,565)
Lag 0-2: 1,027 (-465, 2,520)
Fibrinogen (ug/mL)
LagO: -110 (-504, 283)
Lag 0-2: -110 (-502, 281)
IL-6 (pg/mL)
LagO: 1.32(0.48,2.18)
Lag 0-2: 1.17(0.28,2.08)
IL-6R (pg/mL)
Lag 0: -493 (-9,387, -249)
Lag 0-2: -3,212 (-7,789,
1,365)
TNF-a (pg/mL)
LagO: 0.13 (-0.26, 0.52)
Lag 0-2: 0.15 (-0.22, 0.54)
TNF-RII (pg/mL)
Lag 0:290 (-41, 623)
Lag 0-2: 240 (-82, 562)
P-selectin (ng/mL)
Lag 0:5.13 (-1.02, 11.3)
Lag 0-2: 1.49 (-5.04, 8.02)
VCAM-1 (ng/mL)
Lag 0:53.7 (-11.4, 119)
Lag 0-2: 18.3 (-45.5, 82.0)
ICAM-1 (pg/mL)
Lag 0:5.4 (-9.0, 19.7)
Lag 0-2: 0.58 (-13.5, 14.6)
SOD (U/g Hb)
LagO:-541 (-1,021, -63)
Lag 0-2: -571 (-1,036, -106)
GPx (U/g Hb)
LagO:-1.99 (-3.68, -0.26)
Lag 0-2: 1.15 (-2.81, 0.58)
MPO (ng/mL)
LagO:-5.34 (-14.92, 4.33)
Lag 0-2: -1.15 (-10.81, 8.44)
5-317
-------
Table 5-49 (Continued): Epidemiologic studies of biomarkers of cardiovascular effects.
Pre-
Location Existing
Study Sample Size Condition
tDelfino et al. (2009) Los Angeles, Coronary
CA artery
n = 60 disease
tDelfino et al. (2010) Los Angeles, Coronary
CA artery
n = 60 disease
tWittkopp et al. (2013) LosAnqeles, Coronary
CA artery
n = 36 disease
Exposure
Mean NO2 ppb Assessment
24-h avg NO2 Hourly outdoor
Phase 1:26.4 home air
Phase 2: 28.3 measurements
24-h avg NOx
Phase 1: 37.2
Phase 2: 53.9
Warm season Hourly outdoor
24-h avg NO2: home air
26.4 measurements
24-h avg NOx:
37.2
Cool season
24-h avg NCb:
28.3
24-h avg NOx:
53.9
24-h avg NOx: Hourly outdoor
45.35 home air
Max: 188.00 measurements
Selected Effect Estimates3 (95% Cl)
NOx: NOx:
IL 6 (pg/mL) CRP (ng/mL)
Lag 0: 0.31 (0.16, 0.46) Lag 0: 626 (284, 969)
Lag 0-2: 0.31 (0.13, 0.47) Lag 0-2: 544 (148, 940)
P-selectin (ng/mL) SOD (U/g Hb)
Lag 0: 2.03 (0.13, 3.92) Lag 0: -134 (-269, 2.88)
Lag 0-2: 3.05 (0.90, 5.20) Lag 0-2: -128 (-286, 29.8)
TNF-RII (pg/mL) GPx (U/g Hb)
Lag 0: 88.5 (10.6, 166) Lag 0: -0.23 (-0.81, 0.35)
Lag 0-2: 1 1 5 (25.0, 207) Lag 0-2: -0. 1 8 (-0.85, 0.48)
TNF-a (pg/mL)
Lag 0:0.01 (-0.08, 0.10)
Lag 0-2: 0.06 (-0.04, 0.15)
IL-6 (pg/mL)
NO2: 0.48 (-0.06, 1.05)
NOx: 0.61 (0.26, 0.95)
No quantitative results presented; results presented graphically.
Statistically significant positive associations between 1-, 2-, 3-,
and 5-day avg NOx and IL-6 (pg/mL) and TNF-a (pg/mL) in
haplogroup H participants. Statistically nonsignificant, but
negative associations between 1-, 2-, 3-, and 5-day avg NOx and
IL-6 (pg/mL) and TNF-a (pg/mL) in haplogroup U participants.
5-318
-------
Table 5-49 (Continued): Epidemiologic studies of biomarkers of cardiovascular effects.
Pre-
Location Existing
Study Sample Size Condition
tLiunqman etal. (2009) 6 European Ml
cities
n = 955
(total n = 5,539
measurements)
tBruske etal. (2011) Auqsburq, Ml
Germany
n = 200
Exposure
Mean NO2 ppb Assessment Selected Effect
24-h avg NO2 Central monitor IL-6 (% change)
22.6 Citywide avg Overall: 4.02 (0.47, 8.04)
IL-6 genetic variants
IL-6rs2069832(1,1):
7.33(2.13, 12.8)
IL-6 rs2069832 (1,2):
2.84 (-1.18, 7.09)
IL-6 rs2069832 (2,2):
-1.18 (-8.27, 5.91)
IL-6rs2069840(1,1):
4.26 (-0.95, 9.46)
IL-6 rs2069840 (1,2):
4.02 (0.00, 8.04)
IL-6 rs2069840 (2,2):
4.02 (-3.55, 11.6)
Estimates3 (95% Cl)
IL-6rs2069845(1,1):
6.62(1.18, 12.1)
IL-6 rs2069845 (1,2):
3.07 (-0.95, 7.33)
IL-6 rs2069845 (2,2):
0.47 (-6.15, 7.57)
IL-6rs2070011 (1,1):
4.96 (-0.24, 10.2)
IL-6rs2070011 (1,2):
3.78 (-0.24, 7.80)
IL-6rs2070011 (2,2):
2.60 (-4.26, 9.69)
I L-6rs 1800790 (1,1):
2.36 (-2.13, 6.86)
IL-6 rs1 800790 (1,2):
6.62(1.42, 11.8)
IL-6 rs1 800790 (2,2):
10.4(0.24,21.0)
24-h avg NO2 Central monitor Lp-PLA2 (% Change)
20.8 NO2, lag 4: 7.28(3.00, 11.56)
75th: 24.7 NO, lag 4: 2.74 (-0.21 , 5.70)
Max: 38.2 "Inverse associations were observed for ... NO2 with Lp-PLA2 at
Lag Days 1-2 and positive associations were estimated ...with
24-h avg NO Lp-PLA2 lagged 4 and 5 days."
24.0
75th: 25.8
Max: 141.1
5-319
-------
Table 5-49 (Continued): Epidemiologic studies of biomarkers of cardiovascular effects.
Study
Location
Sample Size
Pre-
Existing
Condition
Mean NO2 ppb
Exposure
Assessment
Selected Effect Estimates3 (95% Cl)
tBarclay et al. (2009)
Aberdeen,
Scotland, U.K.
n = 132
Stable
chronic
heart failure
24-h avg NO2:
30.1
24-h avg NO:
14.7
Central monitor
Hemoglobin
NO2: 0.04 (-0.29, 0.36)
NO: -0.01 (-0.33, 0.31)
Mean corpuscular hemoglobin
NO2: 0.05 (-0.16, 0.26)
NO: -0.04 (-0.24, 0.17)
Platelets
NO2: -0.05 (-0.87, 0.77)
NO: 0.25 (-0.56, 1.05)
Hematocrit
NO2: -0.02 (-0.35, 0.32)
NO: 0.10 (-0.23, 0.43)
WBC
NO2: -0.72 (-1.67, 0.23)
NO: -0.71 (-1.64, 0.22)
CRP
NO2: 0.42 (-5.26, 6.11)
NO: 0.89 (-4.69, 6.47)
IL-6
NO2: 6.28 (0.59, 11.9)
NO: 2.77 (-2.81, 8.34)
vWF
NO2: 2.16 (-0.33, 4.66)
NO: 3.52(1.09, 5.95)
E-selectin
NO2: 1.16 (-0.37, 2.70)
NO: 0.48 (-1.02, 1.99)
Fibrinogen
NO2: -0.22 (-1.76, 1.32)
NO: 0.20 (-1.32, 1.71)
Factor VII
NO2: 0.27 (-1.44, 1.99)
NO: 0.34 (-1.35, 2.02)
D-dimer
NO2: -0.24 (-2.78, 2.29)
NO: -0.32 (-2.81, 2.18)
Welleniusetal. (2007) Boston, MA
n = 28
Stable
chronic
heart failure
24-h avg NO2 Central monitor "No significant associations were observed between any other
20.7 City wide avg pollutant and BNP levels at any of the lags examined."
No quantitative results presented.
tHildebrandt et al.
(2009)
Erfurt,
Germany
n = 38
COPD
24-h avg NO2
13.5
24-h NO
10.7
Central monitor Increases in fibrinogen and prothrombin fragment 1 + 2
associated with NO concentrations. A decrease in vWF was
associated with NO2 concentrations. No quantitative results
presented for NO or NO2.
5-320
-------
Table 5-49 (Continued): Epidemiologic studies of biomarkers of cardiovascular effects.
Study
Location
Sample Size
Pre-
Existing
Condition
Mean NO2 ppb
Exposure
Assessment
Selected Effect Estimates3 (95% Cl)
tDadvand et al. (2014b) Barcelona,
Spain
n = 242
COPD 24-h avg NCfe: Residential land CRP (% change)
30.7 use regression Lag 1: 2.99 (-21.6, 34.2)
Lag 2: 26.1 (-3.47,64.1)
Lag 5: 54.9 (23.2, 94.5)
TNF-a (% change)
Lag 1:3.90 (-24.6, 41.5)
Lag 2: 10.6 (-19.4, 51.4)
Lag 5: 26.5 (-4.51, 66.9)
IL-6 (% change)
Lag 1: 10.5 (-13.5, 40.1)
Lag 2: 4.31 (-18.3, 32.9)
Lag 5: 21.3 (-2.47, 50.4)
IL-8 (% change)
Lag 1: 7.94 (-2.07, 19.0)
Lag 2: 8.09 (-2.00, 19.1)
Lag 5: 3.44 (-5.37, 13.1)
Fibrinogen (% change)
Lag 1: 3.57 (-1.84, 9.09)
Lag 2: 3.26 (-2.00, 8.72)
Lag 5: 10.4(5.59, 15.6)
HGF (% change)
Lag 1: 3.11 (-3.91, 10.6)
Lag 2: 5.57 (-1.58, 13.2)
Lag 5: 9.99(3.44, 17.0)
tKhafaie et al. (2013)
tBindetal. (2012)
tRen etal. (2011)
Pune City, Type II 24-h avg NOx:
India diabetes 21.1
n = 1,392
Boston, MA Healthy 24-h avg NO2
n = 704 18
95th: 35
Boston, MA Healthy 24-h avg NO2
n = 320 17.8
Central site
monitor
citywide avg
Central site
monitor
citywide avg
Central site
monitor
No quantitative results presented; results presented graphically.
NOx was statistically significantly associated with increases in
CRP at lags 0, 1, 2, 4, and 0-7. There were no measurable
differences between winter and summer associations.
Fibrinogen (% change)
Lag 0-2: 8.18(4.73, 11.6)
8-OhdG (% change)
Lag 0:28.5 (-19.4, 76.4)
Lag 0-6: 90.0 (-12.2, 192)
Lag 0-13: 167(28.8, 306)
Lag 0-20: 195(44.9, 345)
tThompson et al. (2010) Toronto,
Canada
n = 45
Healthy 24-h avg NO2 Central site
23.8 monitor
Quantitative results not presented.
5-321
-------
Table 5-49 (Continued): Epidemiologic studies of biomarkers of cardiovascular effects.
Study
Location
Sample Size
Pre-
Existing
Condition
Mean NO2 ppb
Exposure
Assessment
Selected Effect Estimates3 (95% Cl)
tRudez et al. (2009)
Rotterdam, the
Netherlands
n = 40
Healthy
24-h avg NO2
50th: 19.7
75th: 25.5
Max: 43.1
24-h NO:
50th: 5.6
75th: 12
Max: 130.4
Central site Maximal platelet aggregation
monitor (% change)
Lag 0-6 h
NO2: -4.11 (-13.0,4.82)
NO: 5.42 (-18.3, 29.6)
Lag 0-12 h
NO2: -4.64 (-15.0, 5.89)
NO: 2.92 (-22.5, 28.3)
Lag 0-24 h
NO2: -5.36 (-18, 7.68)
NO: 7.92 (-12.5, 28.8)
Lag 24-48 h
NO2: -1.07 (-11.8, 9.46)
NO: 5.00 (-17.1, 27.1)
Lag 48-72 h
NO2: 10.0(2.68, 17.3)
NO: 25.42 (10.00, 40.42);
Late aggregation (% change)
Lag 0-6 h
NO2: 5.89 (-9.46, 21.1)
NO: 33.8 (-5.00, 72.1)
Lag 0-12 h
NO2: 13.4 (-4.11, 30.7)
NO: 35.4 (-2.92, 73.3)
Lag 0-24 h
NO2: 17.7 (-4.46, 39.8)
NO: 37.1 (4.67,69.2)
Lag 24-48 h
NO2: 3.39 (-16.1, 22.7)
NO: 22.9 (-6.25, 51.7)
Thrombin generation—peak
(% change)
Lag 0-6 h
NO2: -2.68 (-9.82, 4.46)
NO: -1.67 (-15.0, 11.7)
Lag 0-12 h
NO2: -1.25 (-9.11, 6.61)
NO: -1.67 (-12.9, 9.58)
Lag 0-24 h
NO2: -1.07 (-9.46, 7.32)
NO: -2.50 (-16.3, 10.8)
Lag 24-48 h
NO2: 14.3(4.29,24.3)
NO: 17.1 (4.58,30.0)
Lag 48-72 h
NO2:6.61 (-2.68, 16.1)
NO: 5.00 (-6.67, 16.7)
Lag 72-96 h
NO2: -0.36 (-8.57, 7.86)
NO: 14.6(1.67,27.9)
Lag 0-96 h
NO2: 1.79 (-7.3, 10.7)
NO: 12.9 (-7.1, 32.5)
Thrombin generation—lag time
(% change)
Lag 0-6 h
NO2: 0.00 (-2.86, 2.86)
NO: -0.42 (-5.83, 4.58)
Lag 0-12 h
NO2: 0.00 (-3.21, 3.04)
NO: 0.00 (-4.58, 4.17)
5-322
-------
Table 5-49 (Continued): Epidemiologic studies of biomarkers of cardiovascular effects.
Study
tRudez et al. (2009)
(Continued)
tSteenhofetal. (2014)
tStraketal. (201 3b)
Pre-
Location Existing
Sample Size Condition
Rotterdam, the Healthy
Netherlands
n = 40 (Continued)
(Continued)
Utrecht, the Healthy
Netherlands
n = 31
Utrecht, the Healthy
Netherlands
n = 31
Mean NO2 ppb
24-h avg NO2
50th: 19.7
75th: 25.5
Max: 43.1
24-h NO:
50th: 5.6
75th: 12
Max: 130.4
(Continued)
5-h avg NO2: 20
5-h avg NO2: 20
Exposure
Assessment
Central site
monitor
(Continued)
Central site
monitor at each
of 5 sites
Central site
monitor at each
of 5 sites
Selected Effect Estimates3 (95% Cl)
Late aggregation (% change)
(Continued):
Lag 48-72 h
NO2: 15.9(4.64,27.1)
NO: 32.9 (9.58, 55.8)
Lag 72-96 h
NO2: 8.57 (-7.68, 24.8)
NO: 14.2 (-23.8, 52.5)
Lag 0-96 h
NO2: 28.8 (8.93, 48.6)
NO: 54.2 (20.4, 87.9)
Thrombin generation — lag time
(% change)
(Continued):
Lag 0-24 h
NO2: 0.36 (-2.86, 3.57)
NO: 2.50 (-2.50, 7.50)
Lag 24-48 h
NO2: -5.54 (-9.11, -1.79)
NO: -7.50 (-12.1, -2.92)
Lag 48-72 h
NO2: -4.46 (-7.68, 1.07)
NO: -3.33 (-7.50, 1.25)
Lag 72-96 h
NO2: 0.00 (-3.21, 3.04)
NO: -5.83 (-10.8, -0.83)
Lag 0-96 h
NO2: -1.25 (-4.46, 1.96)
NO: -4.58 (-11. 7, 2.08)
No quantitative results presented; results presented graphically.
NO2 was statistically significantly associated with decreases in
eosinophils and lymphocytes 2-h after exposure. Null
associations were observed between NO2 and WBC count,
neutrophils, or monocytes.
Endogenous thrombin [in Factor Xll-mediated thrombin
generation pathway (% Change)].
All sites: 65.5 (7.63, 145)
Outdoor: 76.1 (-2.23, 155)
5-323
-------
Table 5-49 (Continued): Epidemiologic studies of biomarkers of cardiovascular effects.
Study
Location
Sample Size
Pre-
Existing
Condition
Mean NO2 ppb
Exposure
Assessment
Selected Effect Estimates3 (95% Cl)
Steinvil et al. (2008)
Tel Aviv, Israel
n = 3,659
Healthy
24-h avg NO2
19.5
75th: 25.3
Central site
monitor
citywide avg
CRP (% change)
LagO
Men: 0.31 (-7.87, 12.6)
Women: -4.72 (-17.3, 9.45)
Lag1
Men: -7.87 (-17.3, 9.45);
Women: -3.15 (-15.8, 11.0)
Lag 2
Men: -1.57 (-11.02, 11.02);
Women: 0.00 (-12.60, 15.75)
Fibrinogen (mg/dL)
Lag 0
Men: -9.92 (-15.6, -4.25)
Women: -12.4 (-19.8, -5.20)
Lag 1
Men: -7.87 (-13.9, -2.05)
Women: -5.51 (-12.9, 1.89)
Lag 2
Men: -7.09 (-13.1, -1.10)
Women: -1.42 (-9.45, 6.46)
WBC (cells/|jL)
Lag 0
Men: 22.1 (-156,200)
Women: -83.5 (-306, 139)
Lag1
Men: 39.4 (-146, 224)
Women: -20.5 (-244, 203)
Lag 2
Men: -36.2 (-227, 154)
Women: 18.9 (-219, 255)
tHildebrandt et al.
(2009)
Erfurt,
Germany
n = 38
Healthy
24-h avg NO2
13.5
24-h NO
10.7
Central monitor Increases in fibrinogen and prothrombin fragment 1 + 2
associated with NO concentrations. A decrease in vWF was
associated with NO2 concentrations. No quantitative results
presented for NO or NO2.
tKhafaie et al. (2013) Pune City,
India
n = 1,392
Healthy 24-h avg NOx: Citywide avg No quantitative results presented; results presented graphically.
21.1 NOx was statistically significantly associated with increases in
CRP at lags 0, 1, 2, 4, and 0-7. There were no measurable
differences between winter and summer associations.
tKelishadi et al. (2009) Isfahan, Iran Healthy 24-h avg: 35.8
(2004-2005) 75th: 47.2
n = 374 Max: 271
Citywide avg NO2 positively associated with CRP and markers of oxidative
stress (oxidized-LDL, malondialdehyde, and conjugated diene).
5-324
-------
Table 5-49 (Continued): Epidemiologic studies of biomarkers of cardiovascular effects.
Study
tLeeetal. (201 1c)
Chuanq et al. (2007a)
Location
Sample Size
Allegheny
County, PA
(1997-2001)
n = 2,211
Taipei, Taiwan
n = 76
Pre-
Existing
Condition
Healthy
Healthy
Mean NO2 ppb
7-day avg: 8.4
75th: 10.1
Max: 25.4
24-h avg NO2
17.3
Max: 53.1
Exposure
Assessment
City wide avg
Central monitor
Selected Effect Estimates3 (95% Cl)
No quantitative results presented. "... NO2 ... associations [with
CRP] were negligible for both the entire population and
nonsmokers only."
"There was no association between NO2 and any of the
markers." No quantitative results presented.
blood
Baccarelli etal. (2007) Lombardia, Healthy
Italy
n = 1,213
24-h avg NO2 Central site
Median: 22.7 monitor
75th: 33.7 citywide avg
Max: 194.2
Homocysteine difference
(% change)
Lag 24 h: 0.24 (-2.86, 3.57)
Lag 0-6 day: -2.21 (-6.0, 1.72)
Homocysteine,
post-methionine-load
(% change)
Lag 24 h: 0.00 (-2.86, 2.86)
Lag 0-6 day: 0.49 (-2.9, 4.17)
tRich etal. (2012)
tZhanqetal. (2013)
Beijing, China Healthy
n = 125
Beijing, China Healthy
n = 125
24-h avg NO2
Entire study:
27.0
Before: 26.0
During: 13.9
After: 41.4
24-h avg NO2
Before: 25.6
During: 14.6
After: 41.4
Central site No quantitative results presented; results presented graphically.
monitor Positive and statistically significant increase in P-selectin,
generally consistent across lags from 0 to 6. Generally null
associations with soluble CD40 ligand across lags from 0-6.
Positive, statistically significant increases in vWF and fibrinogen
at early lags (lag 0, lag 1 ) but null, or negative at later lags.
Generally null or negative associations with WBC across lags
0-6.
Central monitor No quantitative results presented; results presented graphically.
Statistically significant increase in fibrinogen at lag 0. Positive, but
statistically nonsignificant at lags 1, 2, 3, and 6.
8-OhdG = urinary 8-hydroxy-deoxyguanosine; avg =average; CA = California; Cl = confidence interval; COPD = chronic obstructive pulmonary disease; CRP = C-reactive protein;
GPx = glutathione peroxidase; h = hours; HGF = hepatocyte growth factor; ICAM-1 = intercellular adhesion molecule 1; IL = interleukin; IL-6R = interleukin 6 receptor; LDL = low
density lipoprotein; Lp-PLA2 = lipoprotein-associated phospholipase A2; MA = Massachussets; max = maximum; Ml = myocardial infarction; MPO = myeloperoxidase; NO = nitric
oxide; NO2 = nitrogen dioxide; NOX = sum of NO and NO2; PA = Pennsylvania; SOD = superoxide dismutase; TNF-a = tumor necrosis factor alpha; TNF-RII = TNF-receptor II;
VCAM-1 = vascular adhesion molecule-1; vWF = von Willebrand factor; WBC = white blood cells.
aEffect estimates are standardized to a 20-ppb or 30-ppb increase in NO2 or NO or 40-ppb or 100-ppb increase in NOX concentration for 24-h avg and 1-h max metrics,
respectively.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-325
-------
The results have been more heterogeneous in participants without a history of heart
disease. One semiexperimental design assessed changes in blood biomarker levels in
healthy participants exposed to ambient air pollution at five locations in the Netherlands
with contrasting air pollution characteristics (Steenhof et al., 2014; Straket al.. 2013b). A
particular strength of these studies is the measurement of pollutants at the location of
participants' outdoor exposure, which minimizes measurement error from time-activity
patterns and variability in NO2 concentration (Sections 3.4.4.1 and 3.4.4.2). Steenhof et
al. (2014) reported that NO2 was negatively associated with both eosinophil and
lymphocyte counts. Importantly, this could either be due to eosinophils and lymphocytes
leaving the blood and infiltrating stressed tissue, or a decrease in formation of eosinophils
and lymphocytes. The associations were relatively unchanged after adjustment for PM2 5,
EC, OC, or PMio in copollutant models. Straketal. (2013b) observed an increase in
thrombin generation in the endogenous pathway (Factor XH-mediated) associated with
ambient outdoor NO2, that was robust to adjustment for EC or OC, and slightly
attenuated after adjustment for PM2 5.
Among older men participating in the Normative Aging Study, Bindetal. (2012) found
that NO2 was associated with fibrinogen, sVCAM-1, and sICAM-1, but not CRP. In this
same cohort, Ren et al. (2011) found that NO2 was positively linked with urinary
8-hydroxy-29-deoxyguanosine (8-OhdG) concentrations, a marker of oxidative stress
resulting in deoxyribonucleic acid (DNA) damage. Thompson et al. (2010) analyzed the
baseline data on IL-6 and fibrinogen from 45 nonsmoking subjects who participated in a
controlled human exposure study in Toronto, Canada. Using baseline blood samples
allowed the authors to measure the association between systemic inflammation and
ambient NO2, prior to controlled exposure. The authors found that NO2 concentrations
were not associated with either IL-6 or fibrinogen overall, but IL-6 was associated with
NO2 in the winter months. In Rotterdam, the Netherlands, Rudez et al. (2009) measured
CRP, fibrinogen, and markers of platelet aggregation and thrombin generation up to
13 times in 40 healthy participants. Both NO2 and NO were associated with markers of
platelet aggregation and thrombin generation, but neither NO2 nor NO was associated
with CRP or fibrinogen.
During the Beijing Olympics, NO2 was positively associated with increases in biomarkers
indicative of the thrombosis-endothelial dysfunction mechanism (i.e., P-selectin) and
increases in fibrinogen among healthy young adults (Zhang etal.. 2013; Rich et al..
2012). The association with P-selectin was attenuated, but remained positive after
adjustment for PM2 5, CO, Os, SO2, EC, or OC; whereas the association between NO2 and
fibrinogen was generally robust to the above pollutants, with the exception of EC and
OC. Among 3,659 individuals in Tel-Aviv, Israel, Steinvil et al. (2008) found a null
5-326
-------
association between NO2 concentrations and CRP and a negative association with
fibrinogen and white blood cell counts. Baccarelli et al. (2007) observed generally null
associations between NO2 concentrations and total homocysteine among subjects in
Lombardia, Italy. Similarly, Chuang et al. (2007a) observed no association between NCh
and any blood markers, including markers of systemic inflammation and oxidative stress,
as well as fibrinolytic and coagulation factors.
Other subgroups that might be at increased risk of pollution-related health effects have
also been studied. In a cross-sectional study of COPD patients in Barcelona, Spain, there
was evidence of a positive association between NO2 and multiple biomarkers of
inflammation and tissue repair, including CRP, TNFa, IL-6, IL-8, fibrinogen, and
hepatocyte growth factor (HGF) (Dadvand et al.. 2014b). These associations were
generally strongest at lags of 4 or 5 days. A particular strength of this study is that the
authors used validated land use regression models to estimate ambient NO2 exposure at
residential locations. In a repeated-measures study of male patients with chronic
pulmonary disease in Germany, Hildebrandt et al. (2009) reported that NO was positively
associated with fibrinogen and prothrombin levels but not other markers of coagulation;
however, detailed results were not presented in the paper. Khafaie et al. (2013) observed
a positive association between NO2 and CRP in a cross-sectional study of Type II
diabetes patients in Pune City, India. In another cross-sectional analysis of pregnant
women in Allegheny County, PA, there was no association between NO2 and CRP (Lee
etal., 20lie). Among 374 Iranian children aged 10-18 years, Kelishadi et al. (2009)
found that NO2 was associated with CRP and markers of oxidative stress.
Controlled Human Exposure Studies
Markers of inflammation, oxidative stress, cell adhesion, coagulation, and thrombosis
have been evaluated in a few controlled human exposure studies published since the 2008
ISA for Oxides of Nitrogen [(U.S. EPA. 2008c): and Table 5-501. Similar to
epidemiologic studies, controlled human exposure studies also report evidence for
increases in some inflammatory markers, but not consistently across all studies. There is
also evidence for hematological changes following NO2 exposure, and a recent study
reported endothelial cell activation.
5-327
-------
Table 5-50 Characteristics of controlled human exposure studies of
cardiovascular effects.
Study
fChannell et al.
(2012)
Drechsler-Parks
(1995)
Folinsbee et al.
(1978)
Frampton et al.
(2002)
Gonq et al.
(2005)
tHuanq et al.
(2012b)
tLanqrish et al.
(2010)
Disease Status3;
Sample Size;
Sex; Age
(mean±SD)
Primary hCAECs
from n = 7 adults;
M/F; 25.3 ± 5.5 yr
n = 8; M/F;
65. 9 ± 9yr
n = 5/group; M;
20-25 yr
n = 12; M;
26. 9 ± 4.5 yr
n = 9; F;
27.1 ±4.1 yr
Healthy
nonsmokers; n = 6;
68 ± 11 yr
Ex-smokers with
COPD; n = 18;
72 ± 7 yr
n = 23; M/F;
24.56 ± 4.28 yr
n = 10; M; median
age 24 yr
Exposure Details
500 ppb NO2for2 h
Intermediate intermittent exercise
(15 min on/off; VE = 25 L/min per m2
of BSA). Plasma samples collected
before exposures, immediately after,
and 24-h post-exposure. hCAECs
treated with 10 or 30% diluted plasma
samples for 24 h.
600 ppb NO2for2 h
Intermittent exercise (20 min on/off;
VE = 26-29 L/min)
600 ppb NO2 for 2 h
Exercise for 15, 30, or 60 min;
VE = 33 L/min.
600 and 1 ,500 ppb NO2 for 3 h
Intermittent exercise (10 min
on/20 min off; VE = 40 L/min)
400 ppb NO2 for 2 h
Intermittent exercise (15 min on/off;
VE = 22-26 L/min)
1)500ppbNO2for2h
2) 500 ppb NO2 + 73.4 ± 9.9 ug/m3
CAPs for 2 h
Intermittent exercise (15 min on/off;
VE = 25 L/min per m2 of BSA)
4,000 ppb NO2 for 1 h
Intermittent exercise (VE = 25 L/min)
Vasodilator administered 4 h after
exposure:
5, 10, 20 ug/min acetylcholine
inn ^nn 1 nnn nmnl/min hrarl\/Hnin
Endpoints Examined
LOX-1 protein measured from
plasma pre-, immediately post-, and
24-h post-exposure. ICAM-1 and
VCAM-1 mRNA from hCAECs and
IL-8 and MCP-1 protein from cell
supernatant measured immediately
post-exposure to plasma.
HR calculated throughout exposure.
Cardiac output measured during the
last 2 min of each exercise period.
HR, BP, and cardiac output
measured during exposure.
Venous blood collected for
hematocrit, hemoglobin, and red
blood cell count 3.5 h post-
exposure.
HR and BP measured immediately
post, 4-h post, and Day 2.
IL-6, coagulation factors, and lipid
panel in peripheral blood; HRV; and
HR measured 1 and 18 h post-
exposure.
Hemoglobin concentration
measured 4 and 6 h after exposure.
Forearm blood flow and tissue-
plasminogen activator and
plasminogen-activator inhibitor
Type I measured 4 h post-exposure.
2, 4, 8 ug/min sodium nitroprusside
10, 30, 100 ug/min verapamil.
Infusion in brachial artery for
6 min/dose during forearm venous
occlusion plethysmography. Each
vasodilator administration separated
by a 20-min washout period.
5-328
-------
Table 5-50 (Continued): Characteristics of controlled human exposure studies of
cardiovascular effects.
Study
Linn et al.
(1985a)
Disease Status3;
Sample Size;
Sex; Age
(mean±SD)
With asthma;
n = 23; M/F;
18-34yr
Without asthma;
n = 25; 20-36 yr
Exposure Details
3,850-4,210 ppb NO2 for 75 min
Light and heavy intermittent exercise
(1 5 min of each; light VE = 25 L/min;
heavy VE = 50 L/min)
Endpoints Examined
HR and BP measured throughout
exposure.
Posin et al.
n = 8-10; sex and
age NR
1,000 and 2,000 ppb NO2 for 2.5 h
Light intermittent exercise (15 min
on/off)
Acetylcholinesterase, glutathione,
glucose-6-phosphate
dehydrogenase, lactate
dehydrogenase, erythrocyte
glutathione reductase, erythrocyte
glutathione peroxidase, alpha-
tocopherol, TEARS, serum
glutathione reductase, 2,3
diphosphoglycerate, hemoglobin,
hematocrit.
tRiedl et al.
(2012)
(1)n=10M, 5F;
37.3 ± 10.9 yr;
(2)n = 6M, 9F;
36.1 ±2.5yr
(1-2)350ppbNO2for2h
Intermittent exercise (15 min on/off;
VE= 15-20 L/min x m2 BSA)
(1) Methacholine challenge post-
exposure.
(2) Cat allergen challenge post-
exposure.
Serum levels of IL-6, ICAM-1,
fibrinogen, factor VII, and vWF.
Serum collected 22.5 h post-
exposure.
tScaife et al.
(2012)
Stable coronary
heart disease or
impaired left
ventricular systolic
function; n = 18;
M/F; median age
68 yr
400 ppb NO2 for 1 h
HR and HRV monitored
continuously for 24 h post-exposure.
BP = blood pressure; BSA = body surface area; CAP = concentrated air particle; COPD = chronic obstructive pulmonary disease;
F = female; hCAEC = human coronary artery endothelial cells; HR = heart rate; HRV = heart rate variability; ICAM-1 = intercellular
adhesion molecule 1; IL = interleukin; LOX-1 = receptor for oxidized low-density lipoprotein; M = male; MCP-1 = monocyte
chemoattractant protein-1; mRNA = messenger RNA; NO2 = nitrogen dioxide; NR = not reported; ppb = parts per billion; SD =
standard deviation; TEARS - thiobarbituric acid reactive substances; VCAM-1 = vascular adhesion molecule-1; VE = minute
ventilation; vWF = von Willebrand factor.
"•Subjects are heathly unless otherwise specified.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
In healthy adults, exposure to 500 ppb NO2 for 2 hours with intermittent exercise did not
alter circulating IL-8, a pro-inflammatory cytokine, or coagulation factors and induced a
statistically nonsignificant increase in the pro-inflammatory cytokine, IL-6 (Huang et al.,
2012b). Lipid profile changes were also reported. There was a 4.1% increase in blood
total cholesterol (p = 0.059) and a 5.9% increase in high density lipoprotein cholesterol
5-329
-------
(p = 0.036) 18 hours after exposure, but no changes in low density lipoprotein or very
low density lipoprotein cholesterol or triglycerides.
The controlled human exposure study by Langrish et al. (2010) examined the effects of
NO2 on fibrinolytic function. The endogenous fibrinolytic pathway was assessed by
sampling venous concentrations of tissue-plasminogen activator and
plasminogen-activator inhibitor Type I at baseline and 4 hours after exposure.
Concentrations of these proteins were not affected by exposure to NCh.
Atherosclerosis is a chronic inflammatory disease; early stages of the disease include
inflammatory activation of endothelial cells and adhesion of leukocytes to the vascular
endothelium. Channell et al. (2012) reported endothelial cell activation in an in vitro
model following NC>2 exposure. Plasma samples were collected from healthy volunteers
exposed to filtered air or 500 ppb NO2 for 2 hours with intermittent exercise. Primary
human coronary artery endothelial cells (hCAECs) were then treated with a dilution of
these plasma samples (10 or 30% in media) for 24 hours. Increases in messenger RNA
(mRNA) expression levels of endothelial cell adhesion molecules, vascular adhesion
molecule-1 (VCAM-1) and ICAM-1, from hCAECs were observed for both
post-exposure time points compared to control. Cells treated with plasma (30%) collected
immediately post NC>2 exposure had a statistically significant greater release of IL-8 but
not monocyte chemoattractant protein-1 (MCP-1). In addition, plasma collected 24 hours
post NO2 exposure had a significant increase (30%) in soluble lectin-like oxLDL receptor
(LOX-1) levels, a protein recently found to play a role in the pathogenesis of
atherosclerosis (Sections 4.3.2.9 and 4.3.5).
Riedl et al. (2012) reported on the cardiovascular effects of healthy volunteers and
individuals with asthma exposed to 350 ppb NO2 for 2 hours with intermittent exercise.
No statistically significant differences were found in IL-6, ICAM-1, and blood
coagulation factors [i.e., factor VII, fibrinogen, and von Willebrand factor (vWF)] the
morning after NCh exposure.
The 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) reported NO2-induced
hematological changes. Frampton et al. (2002) reported decreases in hematocrit,
hemoglobin, and red blood cell (RBC) count in healthy volunteers 3.5 hours after
exposure to 600 or 1,500 ppb NC>2 for 3 hours with intermittent exercise. Results from
this study support those of Posin et al. (1978). in which hematocrit and hemoglobin levels
were decreased in young males exposed to 1,000 or 2,000 ppb NCh for approximately 2.5
hours with intermittent exercise. However, a recent study reported no change in
hemoglobin levels 4 and 6 hours post-exposure to 4,000 ppb NO2 for 1 hour (Langrish et
al.. 2010).
5-330
-------
Toxicological Studies
Similar to epidemiologic and controlled human exposure studies, several recent
toxicological studies examined the relationship between short-term NO2 exposure and
biomarkers of cardiovascular effects, including markers of oxidative stress, inflammation,
and cell adhesion (Table 5-51). These observations add to those from the 2008 ISA for
Oxides of Nitrogen (U.S. EPA. 2008c) on various hematological parameters in animals
including RBC turnover and methemoglobin levels.
Recently, the effects of NO2 on markers of oxidative stress were examined by Li et al.
(201 la). Rats exposed to 2,660 or 5,320 ppb NO2 for 7 days had a small, but statistically
significant decrease in the activity of the antioxidant enzyme Cu/Zn-SOD and, at the
higher dose, an increase in malondialdehyde (an indicator of lipid peroxidation) in heart
tissue. These changes were accompanied by mild pathological changes in the heart.
However, there were no changes in Mn-SOD or GPx activity or protein carbonyl levels at
either exposure concentration. Campen et al. (2010) reported apolipoprotein E knockout
mice (ApoE~'~) exposed to 200 or 2,000 ppb NO2 had a concentration-dependent decrease
(statistically significant linear trend) in the expression of the antioxidant enzyme HO-1 in
the aorta. Together, these results demonstrate that NO2 inhalation may perturb the
oxidative balance in the heart and aorta.
The effects of NO2 on antioxidant capacity were also examined in the context of diet (de
Burbure et al.. 2007). Rats were placed on low (Se-L) or supplemented (Se-S) selenium
(Se) diets and were exposed to 5,000 ppb NO2 for 5 days. Se is an integral component of
the antioxidant enzyme GPx. GPx levels in RBCs increased in both groups immediately
and 48 hours after exposure; however, plasma levels were decreased in Se-L diet rats at
both time points. SOD activity in RBCs also decreased in Se-L diet rats at both time
points but increased in Se-S diet rats 48 hours after exposure. Overall, NO2 exposure
stimulated antioxidant mechanisms with high Se but were mixed with low Se.
5-331
-------
Table 5-51 Characteristics of animal toxicological studies of cardiovascular
effects.
Study
Species (Strain);
Sample Size;
Sex; Age
Exposure Details
Endpoints Examined
tCampen et al.
(2010)
de Burbure et al.
(2007)
Kunimoto et al.
(1984)
tLietal. (2011 a)
Mersch et al.
(1973)
Mochitate and
Miura(1984)
Mice (ApoE-/-);
n = 5-10/group; M;
8 weeks
Rats (Wistar);
n = 8/group; M;
8 weeks
Rats (Wistar);
n = 6/group; M;
16-20 weeks
Rats (Wistar);
n = 6/group; M;
adults
Guinea pigs; n = 8;
age and sex NR
Rats (Wistar);
n=6; M;
16-20 weeks
200 ppb or 2, 000 ppb NCbfor
6 h/day for 7 days
High-fat diet
(1 ) 1,000 ppb NO2 for 6 h/day,
5 days/week for 28 days
(2) 5,000 ppb NO2 for 6 h/day
for 5 days
(1-2) 30 min Selenium:
6 ug/day or 1.3 ug/day
4,000 ppb NO2 continuously for
1, 4, 7, and 10 days
2,660 or 5,320 ppb NO2,
6 h/day for 7 days
360 ppb NO2 continuously for
7 days
4,000 ppb NO2 continuously for
1, 3, 5, 7, and 10 days
ET-1, MMP-9, HO-1, and TIMP-1 mRNA
expression in aorta. TEARS and
MMP-2/9 activity in aorta. Endpoints
measured 18 h post-exposure.
GPx in plasma and RBC lysate; SOD
activity in RBC lysate; GST activity in
RBC lysate; TEARS in plasma.
Endpoints examined immediately and
48 h after exposure.
ATPase activity, sialic acid, and hexose
in RBC membranes measured after 1, 4,
7, and 10 days of exposure.
Haematoxylin and eosin staining of heart
tissue; SOD activity, GPx activity, MDA
level, and PCO level in heart tissue;
ET-1, eNOS, TNF-a, IL-1, and ICAM-1
mRNA and protein levels in heart tissue;
cardiac myocyte apoptosis. Endpoints
examined 18 h post-exposure.
D-2,3-diphosphoglycerate content in
RBCs; collection time NR.
PK and PFK activity and hemoglobin
content in RBC measured after 1, 3, 5, 7,
and 10 days of exposure.
Nakajima and Mice (ICR); n NR;
Kusumoto (1968) M; 4 weeks
800 ppb NO2 continuously for
5 days
Methemoglobin in blood from the heart
taken immediately after exposure.
tRamos-Bonilla
etal. (2010)
Mice (AKR/J);
n = 3/group; M;
180 days
Low-pollution: 21.2 ppb NO2,
465 ppb CO, 11.5ug/m3PM
High-pollution: 36.1 ppb NO2,
744 ppb CO, 36.7 ug/m3 PM
6 h/day, 5 days/week for
40 weekdays
HR, various heart rate variability
parameters by electrocardiograph, body
weight.
Endpoints measured throughout
exposure.
CO = carbon monoxide; eNOS = endothelial nitric oxide synthase; ET-1 = endothelin-1; GPx = glutathione peroxidase;
GST = glutathione-S-transferase; HO-1 = heme-oxygenase-1; HR = heart rate; ICAM-1 = intercellular adhesion molecule 1;
IL = interleukin; M = male; MDA = malondialdehyde; MMP = matrix metalloproteinase; mRNA = messenger RNA; NO2 = nitrogen
dioxide; NR = not reported; PCO = protein carbonyl; PFK = phosphofructokinase; PK = pyruvate kinase; ppb = parts per billion;
PM = particulate matter; RBC = red blood cell; Se = seleium; SOD = superoxide dismutase; TEARS = thiobarbituric acid reactive
substances; TFN-a = tumor necrosis factor-alpha; TIMP-1 = tissue inhibitor or metalloproteinases-1.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-332
-------
The effects of NO2 on vascular tone modifiers, endothelin-1 (ET-1), and endothelial nitric
oxide synthase (eNOS) were recently examined in two studies (Li etal.. 201 la: Campen
etal.. 2010). ET-1 is a potent vasoconstrictor while the enzyme eNOS catalyzes the
production of NO, which induces vasodilation. Campen et al. (2010) did not see a
statistically significant increase in ET-1 expression level in the aorta after exposure of
mice to 200 or 2,000 ppb NO2. However, higher than ambient-relevant NO2 exposures
induced a statistically significant increase in ET-1 in the heart at the mRNA (10,640 ppb)
and protein level (5,320 and 10,640 ppb) (Li etal.. 201 la). eNOS mRNA and protein
levels were increased at both 2,660 and 5,320 ppb NO2 and decreased to control levels at
10,640 ppb NO2.
Studies also reported changes in some inflammatory markers and adhesion molecules
after NO2 exposure in animals, and some were observed with ambient-relevant
exposures. ICAM-1 transcription and protein levels were increased in the heart after both
2,660 and 5,320 ppb NO2 exposures. These results are consistent with the increase in
ICAM-1 mRNA (Channell et al.. 2012) found in vitro as described above. Li et al.
(2011 a) observed a statistically significant increase in TNF mRNA levels in the heart at
5,320 ppb NO2. In addition, IL-1 expression and protein levels were increased; however,
this effect was in response to a higher NO2 concentration.
The 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) reported on several animal
studies examining hematological parameters. Three studies indicate elevated levels of a
younger population of RBCs following NO2 exposure. In RBCs, levels of
D-2,3-diphosphoglycerate, important in hemoglobin-oxygen dissociation, were increased
in guinea pigs following a 7-day continuous exposure to 360 ppb NO2 (Mersch et al..
1973). Kunimoto et al. (1984) reported an increase in RBC sialic acid after 24 hours of
exposure to 4,000 ppb NO2. Similarly, Mochitate and Miura (1984) reported an elevation
of the glycolytic enzymes, pyruvate kinase and phosphofructokinase, after a 7-day
continuous exposure to 4,000 ppb NO2. These results suggest an increase in RBC
turnover after NO2 exposure. Nakajima and Kusumoto (1968) reported that mice exposed
to 800 ppb NO2 continuously for 5 days had no change in the oxygen-carrying
metalloproteins hemoglobin and methemoglobin.
Summary of Blood Biomarkers of Cardiovascular Effects
In summary, the evidence across disciplines for changes in blood biomarkers of
cardiovascular effects is inconsistent, and supportive evidence for NO2-induced systemic
inflammation is limited. Some epidemiologic evidence suggests the presence of an
association between NO2 concentrations and some markers of systemic inflammation
among participants with a history of heart disease. This association is not consistently
5-333
-------
observed in healthy individuals. Other potentially at-risk populations are not clearly
identified due to contrasting or limited evidence. An important limitation of the
epidemiologic evidence is that potential confounding by traffic-related pollutants was not
evaluated in these studies. Thus, the possibility remains that the observed associations are
the artifact of correlated pollutants. Controlled human exposure studies evaluating
systemic inflammation demonstrated inconsistent results with ambient-relevant NC>2
exposures. Toxicological studies reported an increase in some inflammatory mediators, as
well as oxidative stress effects in RBC, the heart, and aorta of rodents. Short-term NO2
exposure causes a slight reduction in hematocrit and hemoglobin levels associated with a
decrease in RBC levels in controlled human exposure studies and an increase in RBC
turnover in toxicological studies. The clinical significance of these hematological
changes is unknown (Section 4.3.2.9). Evidence has not shown NC>2 to alter circulating
blood coagulation factors or modify responses to vasodilators in controlled human
exposure studies. However, in rats, higher than ambient-relevant NC>2 exposure was
found to induce the expression and production of the vasoconstrictor ET-1.
Overall, there is preliminary evidence, albeit not entirely consistent, from controlled
human exposure and toxicological studies that suggests systemic inflammation and
oxidative stress can occur after exposure to NO2. However, changes in other blood
biomarkers, such as coagulation or vasomotor response, are not observed in relation to
NO2 exposure.
5.3.11 Summary and Causal Determination
Available evidence is suggestive of, but not sufficient to infer, a causal relationship
between short-term exposure to NO2 and cardiovascular effects. The strongest evidence
comes from epidemiologic studies of adults and consistently demonstrates associations
between short-term increases in ambient NO2 concentration and triggering of an MI.
Such associations are indicated by NO2-associated hospital admissions and ED visits for
MI, IHD, and angina; ST-segment alterations; and mortality from cardiovascular disease.
There is a lack of experimental studies that evaluate similar clinical outcomes in order to
assess the coherence across disciplines. However, some controlled human exposure and
toxicological studies provide limited evidence for potential biologically plausible
mechanisms, including inflammation and oxidative stress. Evidence for other
cardiovascular effects is inconsistent.
This conclusion represents a change from the 2008 ISA for Oxides of Nitrogen, which
concluded the "available evidence on the effects of short-term exposure to NO2 on
cardiovascular health effects was inadequate to infer the presence or absence of a causal
5-334
-------
relationship at this time" (U.S. EPA. 2008c). Specifically, the epidemiologic panel
studies and toxicological studies available at the time of the last review were inconsistent.
Most epidemiologic studies reviewed in the 2008 ISA for Oxides of Nitrogen found
positive associations between ambient NO2 concentrations and risk of hospital
admissions or ED visits for all cardiovascular diseases (U.S. EPA. 2008c). However, it
was unclear at that time whether these results supported a direct effect of short-term NO2
exposure on cardiovascular morbidity or were confounded by other correlated pollutants.
Recent epidemiologic studies have further evaluated this uncertainty using copollutant
models and comparing associations of NO2 with those of other criteria pollutants. While
some recent studies show independent associations of NO2 with cardiovascular effects
after adjusting for some pollutants, uncertainties still remain regarding the potential for
NC>2 to serve as an indicator for other traffic-related pollutants or mixtures. Specifically,
there are limited epidemiologic studies evaluating PIVbs or traffic-related pollutants (i.e.,
BC/EC, UFP, OC, VOCs) in copollutant models with NC>2. In some studies, associations
of NO2 and cardiovascular effects were attenuated in with adjustment for PM2 5 or UFP
[Supplemental Figure S5-2, (U.S. EPA. 2015bVI.
There continues to be a lack of experimental evidence that is coherent with the
epidemiologic studies to strengthen the inference of causality for NCh-related
cardiovascular effects, including MI. Further, the limited mechanistic evidence to
describe a role for NO2 in the triggering of cardiovascular diseases, including key events
in the proposed mode of action, remains from the 2008 ISA for Oxides of Nitrogen. The
evidence for cardiovascular effects, with respect to the causal determination for
short-term exposure to NO2, is detailed below using the framework described in the
Preamble (Tables I and II). The key evidence, supporting or contradicting, as it relates to
the causal framework, is summarized in Table 5-52.
5.3.11.1 Evidence on Triggering a Myocardial Infarction
The causal determination for the relationship between short-term NO2 exposure and
cardiovascular effects is based on the evidence for effects related to triggering an MI,
including findings for hospital admissions and ED visits for IHD, MI, or angina and
ST-segment amplitude changes. Time-series studies of adults in the general population
consistently report positive associations between 24-h avg and 1-h max NO2
concentrations and hospital admissions and ED visits for IHD and MI among adults
(Section 5.3.2.1 and Figure 5-18). Risk estimates ranged from 0.87 to 1.76 per a 20- or
30-ppb increase in NO2; most of the risk estimates were greater than 1.00. Symptoms of
MI are similar to those of angina; however, where MI results in damage to the heart
muscle, angina does not result in myocardial necrosis. However, angina may indicate an
5-335
-------
increased risk for future MI. IHD is an over-arching category of related ischemic events
that includes both acute MI and angina, as well as events related to older MI and other
IHD-related events. Observations of increased hospital admissions and ED visits for MI
and IHD are coherent with epidemiologic studies reporting increased hospital admissions
and ED visits for angina (Section 5.3.2.2). Among those hospitalized, ST-segment
decreases are considered a nonspecific marker of myocardial ischemia. A small number
of epidemiologic panel studies reported associations between short-term exposure to NO2
and ST-segment changes on the electrocardiogram of older adults with a history of
coronary artery disease (Section 5.3.2.3).
Coherent with the increase in hospital admissions and ED visits for IHD, MI, and angina,
single-city studies from the U.S. (Ito et al.. 2011: Peel et al. 2007: Tolbert et al.. 2007)
and multicity studies conducted in Europe and Australia and New Zealand (Larrieu et al..
2007: Ballester et al.. 2006: Barnett et al.. 2006: von Klot et al.. 2005) report positive
associations with all CVD hospital admissions in adults with adjustment for numerous
potential confounding factors, including weather and time trends (Section 5.3.8). In turn,
this evidence is coherent with positive associations reported in epidemiologic studies of
short-term increases in ambient NC>2 concentration and cardiovascular mortality in adults
(Section 5.3.9). These include studies reviewed in the 2008 ISA for Oxides of Nitrogen
and recent multicity studies that generally report a similar or slightly larger magnitude for
the NO2 cardiovascular mortality relationship compared to total mortality.
Recent controlled human exposure and animal toxicological studies provide preliminary
evidence for a potentially biologically plausible mechanism for short-term exposure to
NO2 leading to cardiovascular disease, including IHD. Reactive intermediates or
inflammatory mediators that have migrated from the respiratory tract into the circulation
could result in systemic inflammation and/or oxidative stress, which could mediate
effects in the heart and vasculature (Sections 4.3.2.9 and 4.3.5). These nonspecific effects
could promote the triggering of an MI. There is limited and not entirely consistent
evidence in humans and animals for increased systemic and tissue specific oxidative
stress (Channell et al.. 2012: Li et al.. 201 la). In addition, evidence in animal and cell
models and in some controlled human exposure studies report NO2-mediated increases in
inflammatory markers (Channell et al.. 2012: Huang et al.. 2012b: Li et al.. 201 la).
A key uncertainty that remains since the 2008 ISA for Oxides of Nitrogen is the potential
for confounding by other correlated traffic-related pollutants given a common source and
moderate to high correlations with NO2. Recent studies evaluated this uncertainty using
copollutant models and comparing associations of NO2 with those of other pollutants.
NO2 associations with cardiovascular disease outcomes persisted in some but not all
copollutant models with CO [Supplemental Figure S5-3, (U.S. EPA. 2015c)1. Further, not
5-336
-------
all analyses reported NO2 as the strongest predictor of cardiovascular effects. Two studies
reported that associations with cardiovascular hospital admissions were not robust in
models adjusting for ambient CO concentrations (Tolbert et al.. 2007; Barnett et al..
2006). Tolbert et al. (2007) also reported associations with CO, EC, and OC that were
stronger or similar in magnitude to those for NO2. Confounding of NO2-cardiovascular
effect associations by PM2 5 or UFP was examined to a limited extent, and EC/BC and
VOCs were generally not examined as confounders, resulting in the potential for
unmeasured confounding. A larger number of studies examined copollutant models with
PMio or TSP [Supplemental Figure S5-2, (U.S. EPA. 2015b)1. O3 [Supplemental
Figure S5-4; (U.S. EPA. 2015d)1. or SO2 [Supplemental Figure S5-5; (U.S. EPA. 2015e)1
and reported robust NO2 associations with various cardiovascular disease outcomes and
cardiovascular mortality (Chen et al.. 2012b: Chiusolo et al.. 2011). While copollutant
models are a common statistical tool used to evaluate the potential for confounding,
inferences from their results can be limited (Section 5.1.2.2) due to high correlations
among pollutants. Further, copollutant models with cardiovascular effects were based on
pollutants measured at central site monitors, which could result in differential exposure
measurement error. Residual confounding due to unmeasured copollutants
(Section 5.1.2.2) also is possible because reliable methods to adjust for multiple
copollutants simultaneously are not available. Without consistent and reproducible
experimental evidence that is coherent with the effects observed in epidemiologic studies,
uncertainty still exists concerning the role of correlated pollutants in the associations
observed with NO2. The limited or inconsistent results from copollutant models
evaluating confounding by yCO, PM2 5, BC/EC, UFP, or VOCs raises the concern that
NO2 associations could be a result of NO2 serving as a marker for effects of other
traffic-related pollutants or mixtures of pollutants.
5.3.11.2 Evidence on Other Cardiovascular Effects
There is inconclusive evidence from epidemiologic, controlled human exposure, and
animal toxicological studies for other cardiovascular effects from short-term exposure to
NO2. Epidemiologic studies provide inconsistent evidence for an association between
24-h avg NO2 and risk of cardiac arrhythmias as examined in patients with ICDs,
continuous ECG recordings, out-of-hospital cardiac arrest, and hospital admissions
(Section 5.3.3). Similarly, epidemiologic studies provide inconsistent evidence for a
potential association between ambient NO2 concentrations and risk of hospital admission
for cerebrovascular disease and stroke (Section 5.3.4). Both epidemiologic and controlled
human exposure studies provide little to no evidence to indicate that short-term exposure
to ambient NO2 is associated with increased BP or hypertension (Section 5.3.6). Other
5-337
-------
outcomes have an insufficient quantity of studies to evaluate the effects. A small number
of epidemiologic studies have found associations between NO2 concentrations and
hospital admissions or ED visits for heart failure (Section 5.3.5) and hospital admission
for venous thrombosis and pulmonary embolism (Section 5.3.7).
Various subclinical effects have been investigated that are not clearly associated with a
particular clinical cardiovascular event but may be key events in a proposed mode of
action for cardiovascular effects other than MI. There is limited evidence from
epidemiologic and controlled human exposure studies to suggest that NO2 exposure
results in alterations of cardiac autonomic control. Recent epidemiologic studies
generally reported associations between ambient NO2 concentrations and decreases in
indices of HRV (Section 5.3.10.1) and changes in ventricular repolarization
(Section 5.3.10.2) among populations with pre-existing or at elevated risk for
cardiovascular disease. Experimental studies also evaluated changes in HRV and
ventricular repolarization parameters. Although changes were not observed across all
endpoints, a recent controlled human exposure study reported decreased HFn in healthy
exercising adults exposed to NO2, indicating a potential disruption in the normal cardiac
autonomic control (Huang et al.. 2012b). However, similar measures of autonomic
control in another controlled human exposure study showed statistically nonsignificant
increases after exposure to NO2 (Scaife et al.. 2012).
5.3.11.3 Conclusion
In conclusion, consistent epidemiologic evidence from multiple studies at relevant NC>2
concentrations is suggestive of, but not sufficient to infer, a causal relationship between
short-term NC>2 exposure and cardiovascular health effects. The strongest evidence
supporting this determination comprises outcomes related to triggering of an MI.
However, uncertainty remains regarding exposure measurement error and potential
confounding by traffic-related copollutants. Experimental studies provide some evidence
that could propose a potential mode of action but do not provide evidence that is coherent
with the epidemiologic studies to help rule out chance, confounding, and other biases.
Evidence for other cardiovascular effects is inconclusive, including effects on cardiac
arrest and arrhythmia, cerebrovascular disease and stroke, increased blood pressure and
hypertension, and decompensation of heart failure. Studies of adults consistently
demonstrate NCh-associated hospital admissions and ED visits for IHD, MI, and angina,
as well as all cardiovascular diseases. This is coherent with evidence for NCh-related ST
segment decrements and mortality from cardiovascular disease. These results have been
replicated by different researchers in different locations and have adjusted for numerous
potential confounding factors including meteorological factors and time trends. However,
5-338
-------
due to limited analysis of potentially correlated pollutants and recognized limitations of
copollutant models, uncertainty remains regarding the extent to which NC>2 is
independently associated with cardiovascular effects or if NO2 serves primarily as a
marker for the effects of another traffic-related pollutant or mix of pollutants. Thus, the
combined evidence from epidemiologic and experimental studies is suggestive of, but not
sufficient to infer, a causal relationship between short-term NO2 exposure and
cardiovascular effects.
Table 5-52 Summary of evidence, which is suggestive of, but not sufficient to
infer, a causal relationship between short-term nitrogen dioxide
exposure and cardiovascular effects.
Rationale for Causal
Determination3
Key Evidence13
Key References'3
NO2 Concentrations
Associated with
Effects0
Triggering a myocardial infarction
Consistent epidemiologic
evidence from multiple,
high-quality studies at
relevant NO2
concentrations
Increases in hospital admissions and
ED visits for IHD and Ml in adults in
multiple studies, including multicity
studies, in diverse locations.
tLarrieu et al. (2007):
tStiebetal. (2009):Peel
et al. (2007):von Klot et al.
(2005): Mann et al. (2002)
Figure 5-13:
Section 5.3.2.1
Mean 24-h avg:
11.9-37.2 ppb
Mean 1-h max:
45.9 ppb
Coherence with limited evidence for
increased hospital admissions and
ED visits for angina in adults in
multiple studies, including multicity
studies.
tSzvszkowicz
(2009):Poloniecki et al.
(1997): von Klot etal.
(2005)
Section 5.3.2.2
Mean 24-h avg:
12.1-37.2 ppb
Increases in hospital admissions and tLarrieu et al. (2007): flto Mean 24-h avg:
ED visits for all CVD in adults in
multiple studies, including multicity
studies, in diverse locations.
etal. (2011): Peel etal.
(2007):
Tolbert et al. (2007): von
Klot et al. (2005):
Ballesteretal. (2006):
Barnett et al. (2006)
Section 5.3.8
11.9-40.5 ppb
Mean 1-h max:
43.2-45.9 ppb
Coherence with ST-segment
depression in adults with pre-existing
coronary heart disease in association
with 24-h avg and 1-h avg NO2.
tChuanq etal. (2008):
tDelfino et al. (2011)
Section 5.3.2.3
Mean 24-h avg: 21.4
ppb; Mean
1-h max: 27.5 ppb
Consistent evidence for increased
risk of cardiovascular mortality in
adults applying differing model
specifications in diverse locations.
tBellinietal. (2007):
tWonq et al. (2008):
tChenetal. (2012b):
tChiusolo et al. (2011)
Section 5.3.9
Mean 24-h avg:
13.5-35.4 ppb
5-339
-------
Table 5-52 (Continued): Summary of evidence, which is suggestive of, but not
sufficient to infer, a causal relationship between short-
term nitrogen dioxide exposure and cardiovascular
effects.
Rationale for Causal
Determination3
Key Evidence13
Key References'3
NO2 Concentrations
Associated with
Effects0
Uncertainty regarding
exposure measurement
error
Majority of evidence from time-series
studies that rely on central site
exposure estimates.
Uncertainty regarding
potential confounding by
traffic-related
copollutants
Inability to disentangle the effects of
traffic-related pollutants because of
lack of examination of BC/EC and
VOCs, and inconsistent NO2
associations with adjustment for
PM2.5, UFP, or CO.
NO2 associations with ED visits,
hospital admissions, and mortality
found with adjustment for other
potential confounding factors
including meteorological factors and
time trends as well as PM-io, Os, or
SO2.
Supplemental
Figures S5-2, S5-3, S5-4,
and S5-5 (U.S. EPA,
2015b. c, d, e)
Limited evidence for key
events in the proposed
mode of action
Oxidative stress
Limited and supportive evidence of
increased oxidative stress in heart
tissue in rats with relevant NO2
exposures (i.e., MDA) and plasma
from NO2-exposed humans
(i.e., LOX-1).
tLietal. (2011 a)
Section 4.3.2.9.
Figure 4-3
Rats: 5,320 ppb but
not 2,660 ppb NO2
Inflammation
Limited and supportive toxicological
evidence of increased transcription of
some inflammatory mediators in vitro
(i.e., IL-8, ICAM-1, VCAM-1) and in
rats (i.e., ICAM-1, TNF-a).
tChannelletal. (2012)
Human cells exposed
to plasma from
healthy adults:
500 ppb NO2
tLietal. (2011 a)
Rats: 2,660 and
5,320 ppb NO2
Limited and inconsistent evidence in
controlled human exposure studies
(i.e., IL-6, IL-8, ICAM-1).
tHuanqetal. (2012b):
tRiedletal. (2012)
Adults: 350, 500 ppb
NO2
Inconsistent epidemiologic evidence
for changes in CRP, IL-6, and
TNF-RII.
Section 5.3.10.4
5-340
-------
Table 5-52 (Continued): Summary of evidence, which is suggestive of, but not
sufficient to infer, a causal relationship between short-
term nitrogen dioxide exposure and cardiovascular
effects.
Rationale for Causal
Determination3
Key Evidence13
Key References'3
NO2 Concentrations
Associated with
Effects0
Other cardiovascular effects
Inconclusive evidence
from epidemiologic,
controlled human
exposure and
toxicological studies
Inconsistent epidemiologic evidence
for an association between NO2 and
risk of cardiac arrest and
arrhythmias, cerebrovascular disease
and stroke, and increased blood
pressure and hypertension.
Sections 5.3.3, 5.3.4, and
5.3.6
Insufficient quantity of studies
evaluating decompensation of heart
failure and venous thrombosis and
pulmonary embolism.
tStieb et al. (2009):
tYanq(2008)
Section 5.3.5
tDalesetal. (2010)
Section 5.3.7
Inconsistent changes in HRV in
controlled human exposure studies.
tHuanqetal. (2012b)
Healthy adults:
500 ppb NO2
tScaifeetal. (2012)
Section 5.3.10.1
Adults with
pre-existing CVD:
400 ppb NO2
Limited epidemiologic evidence for
changes in HRV and ventricular
repolarization.
Stronger associations observed in
groups of individuals with pre-existing
cardiovascular disease.
HRV:
Timonen et al. (2006):
tSuh and Zanobetti
(201 Oa): tZanobetti et al.
(2010)
Section 5.3.10.1
QT interval:
Henneberqeret al. (2005)
Section 5.3.10.2
avg = average; BC= black carbon; CO = carbon monoxide; CRP = C-reactive protein; CVD = cardiovascular disease;
EC = elemental carbon; ED = emergency department; h = hour; HRV = heart rate variability; ICAM-1 = intercellular adhesion
molecule 1; IHD = ischemic heart disease; IL = interleukin; LOX-1 = receptor for oxidized low-density lipoprotein; max = maximum;
MDA = malondialdehyde; Ml = myocardial infarction; NO = nitric oxide; NO2 = nitrogen dioxide; O3 = ozone; PM25 = particulate
matter with a nomail aerodynamic diameter less than or equal to 2.5 pm; PM10 = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 10 |jm; ppb = parts per billion; SO2 = sulfur dioxide; TNF-RII = tumor necrosis factor
receptor-ll; UFP = ultrafine particles; VCAM-1 = vascular adhesion molecule-1; VOC = volatile organic compound.
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Tables I and \[ of the
Preamble.
"•Describes the key evidence and references, supporting or contradicting, that contribute most heavily to causal determination and,
where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where full body of evidence is
described.
°Describes the NO2 concentrations with which the evidence is substantiated.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-341
-------
5.4 Total Mortality
5.4.1 Introduction and Summary of the 2008 Integrated Science Assessment
for Oxides of Nitrogen
Prior to the 2008 ISA for Oxides of Nitrogen, epidemiologic studies had not been
identified that examined whether an association exists between short-term NO2 exposure
and mortality. The 2008 ISA for Oxides of Nitrogen evaluated a collection of studies,
including multicity studies, conducted in the U.S., Canada, and Europe, and a
meta-analysis (U.S. EPA. 2008c). All of these studies reported evidence of an association
between short-term NO2 exposure and mortality with estimates ranging from 0.5 to 3.6%
for a 20-ppb increase in 24-h avg or 30-ppb increase in 1-h max NO2 concentrations. A
limitation of this collection of studies was that the majority focused specifically on PM
and did not conduct extensive analyses to examine the relationship between short-term
NO2 exposure and mortality.
Multicity studies conducted in the U.S. (HEI. 2003). Canada (Brook et al. 2007; Burnett
et al.. 2004). and Europe (Samoli et al.. 2006). as well as a large study conducted in the
Netherlands (Hoek. 2003). consistently reported positive associations between short-term
NO2 exposure and mortality, specifically at lag 1. Associations were robust in copollutant
models with PMio, SO2, or Os. These results were confirmed in a meta-analysis that did
not include any of the aforementioned multicity studies (Stieb et al.. 2002).
Of the studies evaluated in the 2008 ISA for Oxides of Nitrogen, a limited number
provided additional information (i.e., seasonal analyses, examination of cause-specific
mortality, examination of effect modifiers) on the NO2-mortality relationship. Initial
evidence indicated a larger NO2-mortality association during the warmer months (Brook
et al.. 2007; Burnett et al.. 2004; HEI. 2003). Additionally, an examination of total and
cause-specific mortality found associations similar in magnitude across mortality
outcomes (total, respiratory, and cardiovascular); however, some studies reported
stronger NO2 associations for respiratory mortality (Biggeri et al.. 2005; Simpson et al..
2005b). Potential effect modifiers of the NO2-mortality relationship were examined only
within the APHEA study, which found that within the European cities, geographic area
and smoking prevalence modified the NO2-mortality relationship. It is worth noting that
additional multicity European studies that focused on PM (Agaetal.. 2003; Katsouyanni
et al.. 2003) reported that cities with higher NO2 concentrations also had higher PM risk
estimates indicating that NO2 and PM may be potential effect modifiers of each other.
5-342
-------
In summary, the multicity studies evaluated in the 2008 ISA for Oxides of Nitrogen
consistently observed positive associations between short-term NO2 exposure and
mortality. These studies indicated that associations were found to occur within the first
few days after exposure and are potentially influenced by season. However, uncertainties
remained in the NCh-mortality relationship, which led to the 2008 ISA for Oxides of
Nitrogen (U.S. EPA. 2008c) concluding that the evidence "was suggestive but not
sufficient to infer a causal relationship." These uncertainties and data gaps included
whether: NO2 is acting primarily as an indicator for another pollutant or a mix of
pollutants; there is evidence for potential copollutant confounding; specific factors
modify the NO2-mortality relationship; there is seasonal heterogeneity in mortality
associations; NO2 associations are stronger with specific mortality outcomes; and the
shape of the NO2-mortality concentration-response relationship is linear.
5.4.2 Associations between Short-Term Nitrogen Dioxide Exposure and
Mortality
Since the completion of the 2008 ISA for Oxides of Nitrogen, the body of epidemiologic
literature that has examined the association between short-term NO2 exposure and
mortality has grown. However, similar to the collection of studies evaluated in the 2008
ISA for Oxides of Nitrogen, most of the recent studies did not focus specifically on the
NO2-mortality relationship but on other pollutants. Of the studies identified, a limited
number have been conducted in the U.S., Canada, and Europe, with the majority being
conducted in Asia due to the increased focus on examining the effect of air pollution on
health in developing countries. Although these studies are informative in evaluation of
the relationship between oxides of nitrogen and mortality, the broad implications of these
studies in the context of results from studies conducted in the U.S., Canada, and Western
Europe are limited. This is because studies conducted in Asia encompass cities with
meteorological (Tsai et al.. 2010; Wong et al.. 2008). outdoor air pollution
(e.g., concentrations, mixtures, and transport of pollutants), and sociodemographic
(e.g., disease patterns, age structure, and socioeconomic variables) (Kan et al., 2010)
characteristics that differ from cities in North America and Western Europe, potentially
limiting the generalizability of results from these studies to other cities.
Overall, this section evaluates studies that examined the association between short-term
NO2 exposure and mortality and addresses the key uncertainties and data gaps in the
NO2-mortality relationship identified in the 2008 ISA for Oxides of Nitrogen: potential
confounding of NO2 associations, effect measure modification (i.e., sources of
heterogeneity in risk estimates across cities), seasonal heterogeneity in NO2 associations,
and the NO2-mortality C-R relationship. Other recent studies of mortality are not the
5-343
-------
focus of this evaluation because they were conducted in small single-cities, encompass a
short study duration, had insufficient sample size, and/or did not examine potential
copollutant confounding. The full list of the studies can be found in Supplemental
Table S5-4 (U.S. EPA. 20150.
5.4.3 Associations between Short-term Nitrogen Dioxide Exposure and
Mortality in All-Year Analyses
Multicity studies evaluated in the 2008 ISA for Oxides of Nitrogen reported consistent,
positive associations between short-term NO2 exposure and mortality in all-year analyses
(U.S. EPA. 2008c). However, when focusing on specific causes of mortality, some
studies reported similar risk estimates across total (nonaccidental), cardiovascular, and
respiratory mortality (Samoli et al., 2006; Burnett et al., 2004). while others indicated
larger respiratory mortality risk estimates compared to both total and cardiovascular
mortality (Atkinson etal., 2012; Bigger! etaL 2005; Simpson et al., 2005b). Additional
multicity studies focusing on COPD (Meng etal.. 2013) and stroke (Chenetal.. 2013b)
mortality further support potential differences in the NC^-mortality association by
mortality outcome. Although only a small number of multicity studies have been
conducted since the completion of the 2008 ISA for Oxides of Nitrogen, these studies
build upon and provide additional evidence for an association between short-term NO2
exposure and total mortality along with potential differences by mortality outcome. Air
quality characteristics and study specific details for the studies evaluated in this section
are provided in Table 5-53.
5-344
-------
Table 5-53 Air quality characteristics of studies evaluated in the 2008 Integrated Science Assessment for
Oxides of Nitrogen and recently published multicity and select single-city studies.
Study
Biqqeri et al.
(2005)
Brook etal. (2007)
Burnett et al.
(2004)
HEI (2003)
Hoek (2003)
Samoli etal.
(2006)
Simpson et al.
(2005b)
Stieb et al. (2003)
Location
(Years)
8 Italian cities
(1990-1999)
10 Canadian
cities
(1984-2000)
12 Canadian
cities
(1981-1999)
58 U.S. cities3
(1987-1994)
the
Netherlands
(1986-1994)
30 European
cities
(1990-1997)
4 Australian
cities
(1996-1999)
Meta-analysis,
worldwide
(years NR)
Mortality Outcome(s)
Total, cardiovascular,
respiratory
Total
Total, cardiovascular,
respiratory
Total
Total
Total, cardiovascular,
respiratory
Total, cardiovascular,
respiratory
Total
Mean
Averaging Concentration
Exposure Assignment Time ppb
Average of NO2 concentrations across all 24-h avg 30.1-55.0
monitors in each city (1-6 monitors). Monitors
influenced by local traffic excluded.
Average of NO2 concentrations across all 24-h avg NR
monitors in each city.
Average of NO2 concentrations across all 24-h avg 10.0-26.4
monitors in each city.
Average of NO2 concentrations across all 24-h avg 9.2-39.4
monitors in each city.
15 NO2 monitors across the study area, mean 24-h avg NR
concentration calculated in each region then
weighted by population density in each region.
Average of NO2 concentrations across all 1-h maxb 24.0-80.5
monitors in each city.
Average of NO2 concentrations across all 1-h max 16.3-23.7
monitors in each city.
NA NR NR
Upper Percentile
Concentrations
ppb
95th: 45.8-94.0
Max: 62.6-160.7
NR
NR
NR
NR
90th: 33.1-132.5
Max: 96.0-111.5
NR
5-345
-------
Table 5-53 (Continued): Air quality characteristics of studies evaluated in the 2008 Integrated Science
Assessment for Oxides of Nitrogen and recently published multicity and select single-
city studies.
Study
fAtkinson et al.
(2012)
fBellini et al.
(2007)
tBerqlind et al.
(2009)
fChen et al.
(2012b)
fChen et al.
(2013b)
tChiusolo et al.
(2011)
tKanetal. (2010):
Kan et al. (2008)
fFaustini et al.
(2013)
tltoetal. (2011)
tSacks et al.
(2012)
Location
(Years)
Meta-analysis,
Asia
(years NR)
15 Italian cities
(1996-2002)
5 European
cities
(1992-2002)
17 Chinese
cities
(1996-2010C)
8 Chinese
cities
(1996-2008d)
10 Italian
cities8
(2001-2005)
Shanghai,
China
(2001-2004)
6 Italian cities
(2001-2005)
New York, NY
(2000-2006)
Philadelphia,
PA
(1992-1995)
Mortality Outcome(s)
Total, cardiovascular,
respiratory
Total, cardiovascular,
respiratory
Total
Total, cardiovascular,
respiratory
Stroke
Total, cardiovascular,
cerebrovascular,
respiratory
Total, cardiovascular,
respiratory
Respiratory
(out-of-hospital)
Cardiovascular
Cardiovascular
Exposure Assignment
NA
NR
Average of NO2 concentrations across all
monitors in each city.
Average of NO2 concentrations across all
monitors in each city (2-13 monitors).8
Average of NO2 concentrations across all
monitors in each city (2-12 monitors).8
If more than 1 monitor, average of NO2
concentrations across all monitors in each city
(1-5 monitors).
Average of NO2 concentrations across
6 monitors.
Average of NO2 concentrations over all
monitors within each city (1-5 monitors).'
Average of NO2 concentrations across all
monitors.
Central site monitor.
Mean
Averaging Concentration
Time ppb
NR NR
24-h avg NR
24-havg 11.0-35.4
24-h avg 13.5-34.8
24-havg 19.7-35.6
24-havg 13.8-35.0
24-h avg 35.4
24-havg 24.5-35.1
24-h avg 28.7
1-h max 47.4
Upper Percentile
Concentrations
ppb
NR
NR
NR
Max: 55.1-132.1
NR
90th: 21. 7-48.8
75th: 42.1
Max: 134.9
NR
NR
Max: 146.7
5-346
-------
Table 5-53 (Continued): Air quality characteristics of studies evaluated in the 2008 Integrated Science
Assessment for Oxides of Nitrogen and recently published multicity and select single-
city studies.
Study
tMenq et al.
(2013)
Location
(Years)
4 Chinese
cities
(1996-2008C)
Mortality Outcome(s) Exposure Assignment
COPD Average of NO2 concentrations across all
monitors in each city (7-8 monitors).d
Averaging
Time
24-h avg
Mean
Concentration
ppb
30.6-35.4
Upper Percentile
Concentrations
ppb
NR
tMoolqavkar et al. 72 U.S. cities^ Total
(2013) (1987-2000)
Average of NO2 concentrations across all
monitors in each city.
24-h avg
NR
NR
tShin etal. (2012) 24 Canadian
cities
(1984-2004)
Cardiopulmonary
If more than 1 monitor, average of NO2 24-h avg
concentrations across all monitors in each city
(1-8 monitors).
8.7-25.0
NR
fStieb et al.
(2008)
tWonq et al.
(2010): Wong et
al. (2008)
12 Canadian
cities
(1981-2000)
4 Asian cities
(1996-2004h)
Total
Total cardiovascular
respiratory
If more than 1 monitor, average of NO2 3-h max
concentrations across all monitors in each city.
Average of NO2 concentrations across all 24-h avg
monitors in each city (6-10 monitors).
1981-1990:
24.7-42.6
1991-2000:
16.3-39.2
23.2-34.6
NR
75th: 28.5-41. 2
Max: 72.6-131.9
avg = average; COPD = chronic obstructive pulmonary disease; h = hour; max = maximum; NA = not available; NO2 = nitrogen dioxide; NR = not reported; NY = New York;
PA = Pennsylvania; ppb = parts per billion; U.S. = United States.
aOf the 90 cities included in the NMMAPS analysis only 58 had NO2 data.
bSamoli et al. (2006) estimated 1 -h max concentrations for each city by multiplying 24-h avg concentrations by 1.64.
°Study period varied for each city and encompassed 2 to 7 yr. Hong Kong, China was the only city that had air quality data prior to 2000.
dThese monitors were "mandated to not be in the direct vicinity of traffic or of industrial sources, and not be influenced by local pollution sources, and to avoid buildings, or those
housing large emitters, such as coal-, waste-, or oil-burning boilers, furnaces, and incinerators" (Chen et al.. 2013b: Meng et al.. 2013: Chen et al.. 2012b).
eOnly 9 cities (Cagliari was excluded) were included in the formal analysis of examining potential factors that could increase the risk of mortality due to short-term NO2 exposure.
'Information on the monitors used in this study were obtained from Colais et al. (2012).
9Of the 108 cities included in the analyses using NMMAPS data only 72 had NO2 data.
The study period varied for each city, Bangkok: 1999-2003, Hong Kong: 1996-2002, and Shanghai and Wuhan: 2001-2004.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-347
-------
As demonstrated in Figure 5-22 and Table 5-54. multicity studies evaluated in the 2008
ISA for Oxides of Nitrogen and those recently published, consistently provide evidence
of positive associations between short-term NO2 exposure and total (nonaccidental)
mortality. In these multicity studies, the associations observed were in analyses that
primarily examined all ages, the exceptions being Chiusolo et al. (2011) and Berglind et
al. (2009). who both focused on the risk of mortality attributed to air pollution in the
population >35 years of age. Across these studies, associations between short-term NCh
exposure and mortality were examined primarily in the total population; however,
Berglind et al. (2009) focused on a subset of the population (i.e., MI survivors). The large
effect estimate for Berglind et al. (2009) could be attributed to the larger mortality rate
for MI survivors: 30-day mortality rate of 14-15% and 1-year mortality rate of 22-24%,
compared to populations examined in the other multicity studies (Berglind et al.. 2009).
Study
Dominici et al. (2003)
Stieb et al. (2003)
Samoli et al. (2006)
Burnett et al. (2004)
Hoek (2003)
<-,. . - /i,-incu\
oimpson ei ai. ^ZUUDDJ
_ . , . ff\r\r\-r\
brooK et al. (^uu/)
Biggeri etal. (2005)
Stieb etal. (2008)
Moolgavkaretal. (2013)
Bellini etal. (2007)
Atkinson et al. (2012)
Wong et al. (2008)
Chen etal. (201 2b)
— . . . . . fjf\4 4\
• ( )
oerQlind et al. (2009)
Location
58 U.S. cities
Meta-analysis (Worldwide)
30 European cities
12 Canadian cities
Netherlands
A A * r •*•
4 /\usiranan cmes
-i n /~* -!• •*•
iu oanadian cities
8 Italian cities
12 Canadian cities
72 U.S. cities
15 Italian cities
Meta-Analysis (Asia)
4 Asian cities
17 Chinese cities
4 n it-
iu Italian cmes
5 European cities
Lag
1
...
0-1
0-2
0-6
0-1
- 1
0-1
1
1
0-1
04
- I
0-1
Oc
-O
-2.0 0
-0-
-•-
-•-
— •—
•
A
•
•
•
A
c
0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0
% Increase (95% Cl)
Note: Cl = confidence interval; Black = studies from the 2008 Integrated Science Assessment for Oxides of Nitrogen; Red = recent
studies. Results are presented per a 20-ppb increase in 24-h avg nitrogen dioxide concentrations or a 30-ppb increase in 1-h max
nitrogen dioxide concentrations.
Figure 5-22 Summary of multicity studies that examined the association
between short-term nitrogen dioxide exposure and total mortality.
5-348
-------
Table 5-54 Corresponding percentage
Study
Dominici et al. (2003)
Stieb et al. (2003)
Samoli etal. (2006)
Burnett et al. (2004)
Hoek (2003)
Simpson et al. (2005b)
Brook etal. (2007)
Biqqeri etal. (2005)
tStieb et al. (2008)
tMoolqavkar et al. (2013)
tBellinietal. (2007)
tAtkinson etal. (2012)
tWonq et al. (2008)
tChen etal. (2012b)
tChiusoloetal. (2011)
tBerqlind et al. (2009)
Location
58 U.S. cities
Meta-analysis (worldwide)
30 European cities
12 Canadian cities
the Netherlands
4 Australian cities
10 Canadian cities
8 Italian cities
12 Canadian cities
72 U.S. cities
15 Italian cities
Meta-analysis (Asia)
4 Asian cities
17 Chinese cities
10 Italian cities
5 European cities
increase in total mortality for Figure 5-22.
Age
All
All
All
All
All
All
All
All
All
All
All
All
All
All
>35
>35
Lag Averaging Time
1
—
0-1
0-2
0-6
0-1
1
0-1
1
1
0-1
—
0-1
0-1
0-5
0-1
24-h avg
24-h avg
1-h max
24-h avg
24-h avg
1-h max
24-h avg
1-h max
3-h max
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
% Increase (95% Cl)a
0.5(0.09,0.90)
0.8(0.20, 1.5)
1.8(1.3,2.2)
2.0(1.1,2.9)
2.6(1.2,4.0)
3.4(1.1, 5.7)
3.5(1.4, 5.5)
3.6(2.3, 5.0)
1.9(0.80,2.9)
2.1 (1.8,2.3)
2.2(1.0, 3.6)
3.7(2.1, 5.4)
4.7(3.2,6.2)
6.3(4.2, 8.4)
8.1 (3.7, 12.7
11. 6 (-5.9, 32.4)
avg = average; Cl = confidence interval; U.S. = United States.
aResults are presented for a 20-ppb increase in 24-h avg nitrogen dioxide concentrations or a 30-ppb increase in 1-h max nitrogen
dioxide concentrations.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
When focusing on cause-specific mortality, recent multicity studies reported similar
patterns of associations to those evaluated in the 2008 ISA for Oxides of Nitrogen with
some evidence of larger respiratory mortality risk estimates (Figure 5-23 and Table 5-55).
However, in a study of 15 Italian cities, Bellini et al. (2007) observed smaller
cardiovascular and respiratory mortality risk estimates compared to total mortality, which
contradicts the results of Biggeri et al. (2005) of which Bellini et al. (2007) is an
extension. Additionally, the total mortality results of Bellini et al. (2007) are smaller in
magnitude than those observed in Biggeri et al. (2005).
5-349
-------
Study
Samoli et al. (2006)
Location
30 European cities (APHEA2)
Burnett et al. (2004) 12 Canadian cities
Simpson et al. (2005b) 4 Australian cities
Biggeri et al. (2005) 8 Italian cities (MISA-1)
Bellini et al. (2007) 15 Italian cities (MISA-2)
Atkinson et al. (2012) Meta-Analysis (Asia)
Wong et al. (2008)
Chen et al.
Chen et al.
Chen et al.
Chen et al.
Meng etal.
2012b)
2012b
2013)
2012b)
2013)a
4 Asian cities (PAPA)
17 Chinese cities (CAPES)
17 Chinese cities (CAPES)
8 Chinese cities (CAPES)
17 Chinese cities (CAPES)
4 Chinese cities
Chiusolo et al. (2011)b 10 Italian cities
Mortality
All
Cardiovascular
Respiratory
All
Cardiopulmonary
Respiratory
All
Cardiovascular
Respiratory
All
Cardiovascular
Respiratory
All
Cardiovascular
Respiratory
All
Cardiovascular
Respiratory
All
Cardiovascular
Respiratory
All
Cardiovascular
Stroke
Respiratory
COPD
All
Cardiac
Respiratory
Lag
0-1
0-1
0-1
0-2
0-2
0-2
0-1
0-1
0-1
0-1
0-1
0-1
0-1
0-1
0-1
NR
NR
NR
0-1
0-1
0-1
0-1
0-1
0-1
0-1
0-1
0-5
0-5
1-5
-O-
-ffl-
-2.0 0.0 2.0
4.0 6.0 8.0 10.0
% Increase (95% Cl)
12.0 14.0 16.0
Note: APHEA2 = Air Pollution and Health: A European Approach 2; CAPES =China Air Pollution and Health Effects Study;
C = confidence interval; MISA = meta-analysis of the Italian studies on short-term effects of air pollution; NR = not reported;
PAPA = Public Health and Air Pollution in Asia; ppb = parts per billion. Black = studies from the 2008 Integrated Science
Assessment for Oxides of Nitrogen, red symbols = recent studies. Filled circles = total mortality, Crosshatch = cardiovascular
mortality, dots = respiratory mortality.
a Although the study was not part of the CAPES study, it included four of the cities also included in CAPES;
bStudy examined individuals 535 years of age while the other studies examined all ages. Results are presented per a 20-ppb
increase in 24-h avg nitrogen dioxide concentrations or a 30-ppb increase in 1-h max nitrogen dioxide concentrations.
Figure 5-23 Percentage increase in total, cardiovascular, and respiratory
mortality from multicity studies in relation to ambient nitrogen
dioxide concentrations.
5-350
-------
Table 5-55 Corresponding percentage increase for Figure 5-23.
Study
Samoli etal. (2006)
Burnett et al. (2004)
Simpson et al.
(2005b)
Bigger! etal. (2005)
tBellinietal. (2007)
tAtkinson et al.
(2012)
tWonq et al. (2008)
tChenetal. (2012b)
tChen etal. (201 3b)
tChen etal. (2012b)
tMenq etal. (2013)
tChiusolo et al.
(2011)
Location Age Lag
30 European cities All 0-1
12 Canadian cities All 0-2
4 Australian cities All 0-1
8 Italian cities All 0-1
15 Italian cities All 0-1
Meta-analysis All
(Asia)
4 Asian cities All 0-1
17 Chinese cities All 0-1
8 Chinese cities All 0-1
17 Chinese cities All 0-1
4 Chinese cities All 0-1
10 Italian cities >35 0-5
yr
1-5
Averaging
Time
1-h max
24-h avg
1-h max
1-h max
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
Mortality
Total
Cardiovascular
Respiratory
Total
Cardiopulmonary
Respiratory
Total
Cardiovascular
Respiratory
Total
Cardiovascular
Respiratory
Total
Cardiovascular
Respiratory
Total
Cardiovascular
Respiratory
Total
Cardiovascular
Respiratory
Total
Cardiovascular
Stroke
Respiratory
COPD
Total
Cardiac
Respiratory
% Increase
(95% Clf
1.8(1.3,2.2)
2.3(1.7, 3.0)
2.2(1.0, 3.4)
2.0(1.1,2.9)
2.0(0.5, 3.9)
2.1 (-0.3, 3.9)
3.4(1.1, 5.7)
4.3(0.9, 7.8)
11.4(3.4, 19.7)
3.5(2.2,4.9)
5.0(2.9, 7.1)
5.4(0.2, 11.0)
2.2(1.0, 3.6)
1.5 (-1.7, 4.0)
1.4 (-2.4, 6.7)
3.7(2.1, 5.4)
4.1 (2.2, 6.0)
6.7(3.2, 10.3)
4.7 (3.2, 6.2)
5.2 (3.4, 7.0)
5.7 (2.6, 8.8)
6.3 (4.2, 8.4)
6.9(3.8, 10.1)
5.6 (3.4, 8.0)
9.8(5.5, 14.2)
7.1 (5.4, 8.9)
8.1 (3.7, 12.7)
10.3(5.9, 14.8)
13.7(2.9,25.8)
avg = average; Cl = confidence interval; COPD = chronic obstructive pulmonary disease; h = hour; max = maximum; yr = years.
aResults are presented for a 20-ppb increase in 24-h avg nitrogen dioxide concentrations or a 30-ppb increase in 1-h max nitrogen
dioxide concentrations.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-351
-------
5.4.4 Potential Confounding of the Nitrogen Dioxide-Mortality Relationship
A key uncertainty of the NCh-mortality relationship identified in the 2008 ISA for Oxides
of Nitrogen (U.S. EPA. 2008c) was whether NC>2 acts primarily as a surrogate of another
unmeasured pollutant. As such, although the multicity studies evaluated in the 2008 ISA
for Oxides of Nitrogen reported consistent evidence of an association between short-term
NO2 exposure and mortality that persisted in copollutant models with PMio, SO2, or Os,
these studies often concluded that the observed mortality effects could not be attributed
solely to NO2. Copollutant analyses conducted in recent studies further attempted to
identify whether NO2 has an independent effect on mortality. Additionally, recent studies
have examined whether the extent of temporal adjustment employed adequately controls
for the potential confounding effects of season on the NO2-mortality relationship.
Copollutant Confounding
In the examination of the potential confounding effects of copollutants on the
NO2-mortality relationship, it is informative to evaluate whether NO2 risk estimates
remain robust in copollutant models, specifically with PM2 5 and traffic-related pollutants
(e.g., EC, CO), and whether NO2 modifies the effect of other pollutants. Recent multicity
studies examine the NO2-mortality relationship by taking into consideration both of these
aspects in different study designs and in different study locations (i.e., U.S., Canada,
Europe, and Asia). However, copollutant analyses in these studies did not include
traffic-related pollutants, complicating the overall interpretation of results regarding
whether there is an independent effect of short-term NO2 exposures on mortality.
In a study of 108 U.S. cities using data from the NMMAPS for 1987-2000 (of which 72
had NO2 data), Moolgavkar et al. (2013) used a subsampling approach where a random
sample of 4 cities was removed from the 108 cities over 5,000 bootstrap cycles to
examine associations between short-term air pollution concentrations and mortality. This
approach was used instead of the two-stage Bayesian hierarchical approach employed in
the original NMMAPS analysis, which assumes that city-specific risk estimates are
normally distributed around a national mean (Dominici et al., 2003). In a single-pollutant
model using 100 degrees of freedom (~7 dflyr, which is consistent with NMMAPS) to
control for temporal trends, Moolgavkar et al. (2013) reported a 2.1% (95% CI: 1.8, 2.3)
increase in total (nonaccidental) mortality at lag 1 day for a 20-ppb increase in 24-h avg
NO2 concentrations. The single-pollutant result is larger in magnitude than that observed
in (Dominici et al.. 2003). which only included 58 cities in the NO2 analysis
(Figure 5-22). In a copollutant analysis with PMio, the NO2-mortality risk estimate was
relatively unchanged (1.9% [95% CI: 1.3, 2.4]), and similar to the copollutant results in
(Dominici etal.. 2003).
5-352
-------
Stieb etal. (2008) reported results consistent with Moolgavkar et al. (2013) in a study
that focused on the development of a new air quality health index in Canada. Focusing on
Lag Day 1 and models using 10 dfper year, Stieb et al. (2008) examined whether
copollutants confounded the single-pollutant results in both copollutant and
multipollutant models with CO, Os, PMio, PNfe.s, and SCh. However, the study did not
clearly identify which results pertained to which model. As stated previously in this ISA,
multipollutant models are difficult to interpret due to the multicollinearity often observed
between pollutants and as a result are not used to assess whether there is evidence of
copollutant confounding. In models using all available data and limited to days with PMio
data the results of the copollutant and multipollutant analyses conducted by Stieb et al.
(2008) indicate that the NC^-mortality relationship remain relatively unchanged when
adjusted for other pollutants, including some traffic-related pollutants (quantitative results
not presented).
Additional studies conducted in Europe and Asia also provide evidence indicating that
NO2-mortality associations remain robust in copollutant models; however, these studies
also did not focus on traffic-related pollutants. Chiusolo etal. (2011) conducted a
multicity study of 10 Italian cities using a time-stratified, case-crossover approach as part
of the Italian Epi Air multicenter study "Air Pollution and Health: Epidemiological
Surveillance and Primary Prevention." The authors reported consistent, positive
associations for total and cause-specific mortality (i.e., cardiac, cerebrovascular, and
respiratory), ranging from an 8.1 to 13.7% increase for a 20-ppb increase in 24-hour NC>2
concentrations using an unconstrained distributed lag of 0-5 days (lag 1-5 days was used
for respiratory mortality). In copollutant models, NCh risk estimates remained robust in
models with PMio in all-year analyses and with Os in analyses restricted to the summer
season (i.e., April-September) (Table 5-56).
5-353
-------
Table 5-56 Percentage increase in total and cause-specific mortality for a
20-ppb increase in 24-hour average nitrogen dioxide concentrations
in single- and copollutant models with particulate matter in all-year
analyses or ozone in summer season analyses.
Mortality
All natural
Cardiac
Cerebrovascular
Respiratory
Season
All-year
April-September
All-year
April-September
All-year
April-September
All-year
April-September
Model
NO2 (lag 0-5)
With PMio (lag 0-5)
NO2 (lag 0-5)
With O3 (lag 0-5)
NO2 (lag 0-5)
With PMio (lag 0-5)
NO2 (lag 0-5)
With O3 (lag 0-5)
NO2 (lag 0-5)
With PMio (lag 0-5)
NO2 (lag 0-5)
With O3 (lag 0-5)
NO2(lag 1-5)
With PMio (lag 0-5)
NO2(lag 1-5)
With O3 (lag 0-5)
% Increase (95% Cl)
8.1 (3.7, 12.7)
7.5(1.9, 13.5)
17.8(12.3,23.6)
18.2(13.1,23.6)
10.3(5.9, 14.8)
10.1 (4.0, 16.4)
19.2(11.4,27.4)
18.8(10.7,27.5)
9.1 (-0.5, 19.7)
9.9 (-2.6, 24.1)
33.0(19.2,48.3)
30.2(13.9,48.8)
13.7(2.9,25.8)
13.4(2.9,24.8)
41.3(16.2, 71.7)
43.4(14.6, 79.5)
Cl = confidence interval; NO2 = nitrogen dioxide; O3 = ozone; PM10 = particulate matter with a nominal mean aerodynamic
diameter less than or equal to 10 |jm.
Concentrations converted from |jg/m3to ppb using the conversion factor of 0.532, assuming standard temperature (25°C) and
pressure (1 atm).
Source: Reproduced from Environmental Health Perspectives, (Chiusolo et al.. 2011).
The Public Health and Air Pollution in Asia (PAPA) study as well as CAPES collectively
found that the NCh-mortality association remains robust in copollutant models with a
nontraffic-related pollutant in analyses conducted in Asian cities. The PAPA study
examined the effect of air pollution on mortality in four cities, one in Thailand
(i.e., Bangkok) and three in China (i.e., Hong Kong, Shanghai, and Wuhan) (Wong et al..
2010; Wong et al.. 2008). In these study locations, PMio and 862 concentrations are
much higher than those reported in the U.S.; however, NC>2 and Os concentrations are
5-354
-------
fairly similar (Wong et al., 2010; Wong et al., 2008). Copollutant models were only
analyzed in the individual cities; a combined four-city analysis was not conducted. In
models using lag 0-1 days NCh concentrations in the Chinese cities, NC>2 mortality risk
estimates were relatively unchanged in copollutant models (quantitative results not
presented). However, in Bangkok, the NC^-mortality risk estimate was attenuated in
models with PMio.
The results from the Chinese cities in the PAPA study are consistent with those found in
CAPES (Chenet al., 2012b). In a two-stage Bayesian hierarchical model, where the first
stage followed the PAPA protocol, Chen et al. (2012b) reported a 6.3% increase
(95% CI: 4.2, 8.4) in total mortality, 6.9% increase (95% CI: 3.8, 10.1) for cardiovascular
mortality, and 9.8% increase (95% CI: 5.5, 14.2) for respiratory mortality for a 20-ppb
increase in 24-h avg NCh concentrations at lag 0-1 days. Although NC>2 was moderately
correlated with both PMio and SO2, 0.66 and 0.65, respectively, NCh-mortality
associations, although attenuated, remained positive across total, cardiovascular, and
respiratory mortality with the percentage increase in mortality ranging from 4.6-6.7% in
copollutant models with PMio and 5.2-7.0% in models with 862 for a 20-ppb increase in
24-h avg NC>2 concentrations.
In addition to examining whether copollutants confound the NCh-mortality relationship,
studies also conducted analyses to examine if there was any indication that NO2 modifies
the PMio-mortality relationship. The Air Pollution and Health: A European and North
American Approach study, although it focused specifically on examining the
PMio-mortality relationship, also conducted an analysis to identify whether NC>2 modifies
the PMio-mortality relationship. In both the European and U.S. data sets, as mean NCh
concentrations and the NCh/PMio ratio increased, there was evidence that the risk of PMio
mortality increased. These results are consistent with Katsouyanni et al. (2003) and
Katsouyanni et al. (2001). who reported higher PMio risk estimates in cities with higher
NO2 concentrations, suggesting that NC>2 and PMio may be effect modifiers of each other.
Temporal Confounding
Recent studies have also examined whether the NO2-mortality relationship is subject to
temporal confounding. These studies have focused on examining the effect of increasing
the number of df employed per year to control for temporal trends on NC^-mortality risk
estimates. Using the entire data set, which encompassed the years 1981-2000, Stieb et al.
(2008) examined the effect of using an alternative number of dfto adjust for seasonal
cycles on NCh-mortality risk estimates. In analyses of single-day lags from 0 to 2 days in
single-pollutant models, the authors reported comparable risk estimates for each
individual lag day when using 6, 8, 10, 12, and 14 dfper year. Similar to Stieb et al.
5-355
-------
(2008). the PAPA study also examined the impact of alternative approaches to
controlling for temporal trends on mortality risk estimates. In models using 4, 6, 8, 10, or
12 dfper year, Wong et al. (2010) also reported relatively similar results across the dfper
year specified, with some evidence for a slight attenuation of the NCh-mortality
association in Wuhan, China as the df per year increased.
Unlike Stieb et al. (2008) and Wong et al. (2010). who conducted a systematic analysis of
the influence of increasing the dfper year to control for temporal trends on the
NO2-mortality relationship, Moolgavkar et al. (2013) only compared models that used
50 df(~3.5 dfper year) or 100 df(~7 dfper year) in the statistical model. However,
similar to both Stieb et al. (2008) and Wong etal. (2010). Moolgavkar etal. (2013)
reported similar results regardless of the number of df used, 2.0% (95% CI: 1.8, 2.3) for a
20-ppb increase in 24-h avg NC>2 concentrations at lag 1 day in the 50 df model and 2.1%
(95% CI: 1.8, 2.3) in the WO df model
5.4.5 Modification of the Nitrogen Dioxide-Mortality Relationship
To date, a limited number of studies have examined potential effect measure modifiers of
the NO2-mortality relationship. In the 2008 ISA for Oxides of Nitrogen (U.S. EPA.
2008c). Samoli et al. (2006) provided evidence of regional heterogeneity in
NO2-mortality associations and higher NC^-mortality risk estimates in cities with a lower
prevalence of smoking as part of the APHEA-2 study. Recent multicity studies conducted
in Italy (Chiusolo etal.. 2011). Chile (Cakmaket al.. 20 lib), and Asia (Chen et al..
2012b) conducted extensive analyses of potential effect measure modifiers of the
NO2-mortality relationship and identified specific factors that may characterize
populations potentially at increased risk of NCh-related mortality (Chapter 7). These
studies presented evidence indicating that older adults (>65 years of age), females,
individuals with pre-existing cardiovascular or respiratory diseases, and individuals of
lower SES, specifically lower income and educational attainment, are at greater risk.
Despite these findings, demographic as well as socioeconomic differences between
countries may complicate the interpretation of results across these studies, and
subsequently the ability to make generalizations across locations regarding the factors
that may modify the NC^-mortality association.
5-356
-------
5.4.6 Potential Seasonal Differences in the Nitrogen Dioxide-Mortality
Relationship
Studies evaluated in the 2008 ISA for Oxides of Nitrogen indicated seasonal differences
in the NC^-mortality relationship with evidence of larger associations in the warm or
summer season. Recent multicity studies conducted in Canada (Shin etal.. 2012; Stieb et
al.. 2008) and Italy (Chiusolo etal.. 2011; Bellini et al.. 2007) further support these
previous findings but also raise additional questions in light of the seasonal patterns in
NO2 concentrations observed in the U.S. and Canada (i.e., higher concentrations in the
winter months compared with the summer months) and the higher personal-ambient
relationship in the summer compared with the winter (Section 2.5.4). Additionally,
limited information indicates higher correlation between personal and ambient measures
in the warm season (Section 3.4.2). which could be a factor in larger NC>2 risk estimates
in the warm season.
In the 12 Canadian city study, Stieb et al. (2008) reported that NCh-mortality risk
estimates were larger in the warm season (April-September) compared with the cool
season (October-March) (quantitative results not presented). These results are consistent
with those reported by Shin etal. (2012) in a study that examined year-to-year changes in
the association between short-term NO2 exposure and mortality (i.e., cardiopulmonary
and noncardiopulmonary) across 24 Canadian cities during 1984-2004. In seasonal
analyses, NO2 associations with cardiopulmonary mortality at lag 0-2 days were
observed to be stronger in the warm season (April-September) compared with the cold
season (October-March). Shin etal. (2012) suggest that the larger NO2 mortality effects
in the warm season could be due to the role of NO2 in the atmospheric reactions that form
Os, and subsequently suggests that the relationship between NO2 and Os does not allow
for a clear assessment of the independent effects of NO2. However, in Canada, as well as
the U.S., NO2 concentrations are higher in the cold season compared to the warm season.
Additionally, NO2 and Os are not well correlated during the summer (r ranging from 0.0
to 0.40), which makes it less likely Os is a confounder of the NO2-mortality relationship
(Section 3.4.4.1).
To date, U.S.-based multicity studies have not examined whether the seasonal patterns of
NO2-mortality associations observed in Canadian multicity studies are similar in the U.S.
However, a few single-city U.S.-based studies that focused on cardiovascular mortality
inform upon whether there is evidence of seasonal differences in NO2-total mortality
associations (Sacks etal.. 2012; Ito etal.. 2011). In a study conducted in New York City
that examined the association between short-term exposure to air pollution and
cardiovascular mortality, Ito etal. (2011) reported similar effect estimates in all-year
[1.8% (95% CI: 0.17, 3.3) for a 20-ppb increase in 24-h avg NO2 concentrations at lag
5-357
-------
1 day] and seasonal [warm: 1.8% (95% CI: -0.4, 3.9); cold: 2.3% (95% CI: 0.0, 4.7)]
analyses. The study did not conduct copollutant analyses and the NCh-mortality pattern
of associations was similar to that observed for PM2 5 and EC.
Sacks et al. (2012) also examined potential seasonal differences in the
NC>2-cardiovascular mortality association in a study conducted in Philadelphia, PA that
examined the influence of various approaches to control for seasonality and the potential
confounding effects of weather on the air pollution-cardiovascular mortality relationship.
Across models, the authors found that either: NCh-mortality associations were similar
between warm and cold seasons; or that associations were slightly larger in magnitude
during the warm season. These results suggest that the modeling approach employed may
influence the NCh-mortality associations observed, specifically with regard to whether
there is evidence of seasonal differences in associations, but the various approaches did
not influence the direction of the observed association.
Multicity studies conducted in Italy provide evidence consistent with that observed in the
Canadian multicity studies. In MISA-2, Bellini et al. (2007) reported larger
NO2-mortality risk estimates in the summer (April-September) compared with the winter
(October-March) for total (6.4 versus 0.9% for a 20-ppb increase in 24-h avg NC>2
concentrations at lag 0-1 days), respiratory (9.1 versus -0.04%), and cardiovascular (7.3
versus -0.2%) mortality. In an analysis of 10 Italian cities, Chiusolo et al. (2011)
supported the results of Bellini et al. (2007) by indicating larger NCh-mortality risk
estimates in the warm season compared with all-year (Table 5-56) for total
(nonaccidental) mortality and cause-specific mortality (i.e., cardiac, cerebrovascular,
respiratory).
The evidence for increased NCh-mortality associations in the warm season, as presented
in the Canadian and Italian multicity studies (Shin etal.. 2012; Stieb et al.. 2008; Brook
et al.. 2007; Burnett et al.. 2004). differs from the seasonal patterns observed in a study
conducted in Shanghai as part of the PAPA study (Kan etal., 2010; Kan et al., 2008). The
authors reported evidence of increased NO2-mortality risk estimates in the cold season
compared with the warm for total (nonaccidental) mortality (cold: 4.7 versus Warm:
1.7% for a 20-ppb increase in 24-h avg NC>2 at lag 0-1 days), cardiovascular (cold: 4.8
versus Warm: 1.1%), and respiratory mortality (cold: 10.4 versus Warm: -5.1%). Across
all of the gaseous pollutants examined, mortality risk estimates were double the size or
larger in the cool season, whereas PMio mortality risk estimates were similar across
seasons except for respiratory mortality (larger in the cool season). The authors speculate
these seasonal differences could be due to seasonal exposure differences specific to
Shanghai (i.e., limited time spent outdoors and increased air conditioning use in the warm
season because of high temperature and humidity and heavy rain, versus more time spent
5-358
-------
outdoors and open windows in the cool season) (Kan et al., 2010; Kan etal.. 2008). The
results of (Kan et al.. 2010; Kan et al.. 2008) highlight the complexity of clearly
identifying seasonal patterns in NC^-mortality associations across locations with
drastically different seasonal weather patterns.
5.4.7 Nitrogen Dioxide-Mortality Concentration-Response Relationship and
Related Issues
Lag Structure of Associations
The 2008 ISA for Oxides of Nitrogen found consistent evidence across studies indicating
that NO2-mortality effects occur within the first few days after exposure, with multiple
studies demonstrating the largest effect occurring the day after exposure (i.e., lag 1 day)
(U.S. EPA. 2008c). Recent multicity studies have conducted additional analyses
examining multiday lags, which further inform the lag structure of associations between
short-term NC>2 exposure and mortality.
In the analysis of 10 Italian cities, Chiusolo et al. (2011) examined the lag structure of
associations between mortality and short-term NCh exposure through both single-day and
multiday lag analyses. Multiday analyses consisted of a priori defined lags (i.e., 0-1, 2-5,
and 0-5 days) examined using an unconstrained distributed lag model. In addition to
examining single-day lags of 0 to 5 days, the authors also explored the pattern of
associations observed over each individual day using a constrained polynomial
distributed lag model. The individual lag days of a constrained distributed lag model are
not directly interpretable; however, this analysis allowed Chiusolo et al. (2011) to
visually display the potential latency of the NO2 effect on mortality. Collectively, the
single- and multi-day lag analyses support an immediate effect of NCh on mortality but
also provide evidence for a prolonged effect extending out to 5 days for all mortality
outcomes (Figure 5-24).
5-359
-------
ta
(Q
3.00
2.50
2,00
1.50
1.00
0.50
-0.50
1 nn
All natural mortality
,,tl
QSingle-lag models
• Distributed-lag models
0123
c
U '
1
[
1
1
Cardiac mortality
T T T T T
1 i „ -
1 5 1 \ 1 3 *
,i,
n
]
4 5 0-1 2-50-5 012345 0-1 2-5 0-5
Lag (days) Lag (days)
5.00
4.00
01
re 3.00
0)
| 2.00
'i i.oo
i °
°~ -1.00
7.00
[
]
c
[
]
]"
1
1 L
lereb
I [
rovascular mortality
i 1 }*
| i
i
i
i
"T""1""""
I
I1
Re
If
I
spiratory mortality 1
:Hf-l
c
i
i
[
]
234 50-1 2-5 0-5
Lag (days)
0123450-12-5 0-5
Lag (days)
Source: Reproduced from Environmental Health Perspectives, (Chiusolo et al.. 2011).
Figure 5-24 Percentage increase in total and cause-specific mortality
associated with short-term increases in ambient nitrogen dioxide
concentration at single day lags, individual lag days of a
constrained polynomial distributed lag model, and multiday lags
of an unconstrained distributed lag model.
Chen etal. (2012b) also conducted an extensive analysis of the lag structure of
associations for the NCh-mortality relationship as part of CAPES. Multiday lags were
examined by averaging multiple single lag days and using a constrained polynomial
distributed lag model of 0-4 days. Chen etal. (2012b) reported the largest effect at single
day lags of 0 and 1 and the average of lags 0-1 days indicating an immediate effect of
NO2 on mortality (Figure 5-25). However, the similar or larger magnitude results for lag
0-4 day avg and the distributed lag model provide some evidence for a delayed NO2
effect on total, cardiovascular, and respiratory mortality, which is consistent with the
results of Chiusolo et al. (2011) (Figure 5-24). These results were further supported by
studies of cause-specific mortality. Chen et al. (2013b) as part of CAPES, in a subset of
eight Chinese cities, reported the largest magnitude of an NO2 effect on stroke mortality
5-360
-------
at lag 0-1 days, but the association remained positive and statistically significant in an
analysis of lag 0-4 days (Section 5.3.9). In an analysis of COPD mortality in four
Chinese cities, Meng etal. (2013) also provided evidence of associations larger in
magnitude for multiday averages, suggesting a prolonged effect, with the largest
association at lag 0-4 and slightly smaller associations for a lag of 0-1 days
(Section 5.2.8). These results are consistent with Faustini etal. (2013) in a study of
out-of-hospital respiratory mortality in six Italian cities that found upon examining both
single- and multi-day lags the strongest associations with NCh were for lags of 2-5 and
0-5 days.
4.0-i
Co"
2- 3.5-
o5 3.0-
0
E 2.5-
c
'CD 2.0-
cn
1 I-*'
c
c
(1)
CJ
Q) °-5~
Q.
-0.5-
"•.,'
0 1 2 3 4 01
i
i
04 pLivj
Total mortality
<
i
0
»
I
f
2
<
i
3
I
i
1
1
«
1
01
]
|
^
i
04 pLIv]
Cardiovascular mortality
i
i
0
i
i
1
i
2
4
:
i
3
i
i
4
<
I
01
i
1 1
••
04 PLIV
Respiratory mortality
DIM = polynomial distributed lag model, representing the cumulative effects of NO2. Percentage increase (mean and 95% Cl) of
daily mortality associated with a 10-ug/m3 (5.3-ppb) increase of nitrogen dioxide concentrations, using different lag structures.
Multiday average lag 01 corresponds to 2-day moving average, and lag 04 corresponds to 5-day moving average of nitrogen dioxide
concentration of the current and previous 4 days.
Source: Reprinted with permission of Elsevier, (Chen et al.. 2012b).
Figure 5-25 Percentage increase in total and cause-specific mortality
associated with short-term increases in ambient nitrogen dioxide
concentration in single- and multi-day lag models in a multicity
study in China.
Additional studies that examined associations between NO2 and mortality at single-day
lags or multiday averages provide evidence that is consistent with those studies evaluated
in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). which demonstrated strong
associations between NO2 and mortality at lag 1. In the analysis of 12 Canadian cities,
5-361
-------
Stieb etal. (2008) found the strongest association between short-term NCh exposure and
mortality at lag 1 when examining single-day lags of 0-2 days. Wong et al. (2008) and
Wong etal. (2010) examined single and multiday lags in each individual city in the
PAPA study. In the three Chinese cities, similar to Stieb et al. (2008). the authors
reported evidence of immediate effects of NO2 on mortality; with the strongest
association occurring for a 0-1 day lag. However, in Bangkok, the lag structure of
associations was different and more in line with those observed in Chiusolo etal. (2011)
and Chen et al. (2012b). with the strongest association occurring at a lag of 0-4 days.
Concentration-Response Relationship
The studies evaluated in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) that
examined the association between short-term NCh exposure and mortality did not
conduct formal analyses of the C-R relationship. Recent studies published since the
completion of the 2008 ISA for Oxides of Nitrogen have examined the NO2-mortality
C-R relationship in both multi- and single-city analyses, focusing on the shape of the C-R
curve and whether a threshold exists.
Using a subsampling approach, Moolgavkar et al. (2013) examined the shape of the C-R
relationship between short-term air pollution exposures and mortality in the NMMAPS
data set by applying a nonlinear function (i.e., natural splines with 6df)to each pollutant.
This analysis provides support for a linear relationship between short-term NO2
exposures and mortality (Figure 5-26). Although Moolgavkar et al. (2013) stated that the
C-R relationship for NO2 "suggest(s) nonlinearity and threshold-like behavior," the
widening of the confidence intervals at the tails of the distribution prevents a clear
interpretation of the shape of the curve where the data density is low. Notably, the
confidence intervals approach zero at the low end of the NO2 distribution due to the way
the model is structured.
5-362
-------
0.06
0.04
0.02
-0.02
V _.t
20 40
Lag-1 N02
60
Note: NO2 = nitrogen dioxide in parts per billion; RR = relative risk. Pointwise means and 95% confidence intervals adjusted for size
of the bootstrap sample.
Source: Reproduced from Environmental Health Perspectives, (Moolgavkar et al.. 2013).
Figure 5-26 Flexible ambient concentration-response relationship between
short-term nitrogen dioxide concentrations and mortality at Lag
Day 1.
The evidence for a linear C-R relationship between short-term NO2 exposure and
mortality was further supported by Stieb et al. (2008) in a pooled analysis of 12 Canadian
cities. The authors examined three functional forms (i.e., linear, quadratic, and cubic
polynomial) and assessed the model fit using the sum of the Akaike Information
Criterion. Stieb et al. (2008) indicated that the linear function was the best fit of the
NO2-mortality relationship (quantitative results not presented).
Multicity studies conducted in Asia examined the NCh-mortality C-R relationship
through either a combined analysis using data from all cities or by examining the C-R
relationship in individual cities. Chenetal. (2012b) examined the shape of the
NO2-mortality C-R curve across all cities as part of CAPES for total, cardiovascular, and
respiratory mortality using 24-h avg NCh concentrations at lag 0-1 days. To limit the
influence of extreme NO2 concentrations on the shape of the C-R curve, concentrations
greater than 120 (ig/m3 (62.4 ppb), which represented only 3% of the data, were
excluded. The authors used a cubic spline with two knots at different concentrations for
each of the mortality outcomes [40 (ig/m3 (20.8 ppb) and 70 (ig/m3 (36.4 ppb) for total
5-363
-------
mortality, 50 (ig/m3 (26.0 ppb) and 70 (ig/m3 (36.4 ppb) for cardiovascular mortality, and
40 (ig/m3 (20.8 ppb) and 60 (ig/m3 (31.2 ppb) for respiratory mortality]. Chen et al.
(2012b) found evidence of a linear relationship between short-term NO2 exposure and
total and cause-specific mortality (Figure 5-27). which was confirmed by the lack of a
statistically significant difference in the deviance between the spline and linear fit
models. These results are further supported by examinations of the C-R relationship for
the cause-specific mortality outcomes of stroke [(Chenetal.. 2013b); Section 5.3.91 and
COPD [Meng et al. (2013): Section 5.2.8]. which also provided evidence of a linear
relationship.
o
E o
ro
0)
w
ra
2! 0
Q)
I
v> -
^^— Total mortality
~~ ~ Cardiovascular mortality
1 - " Respiratory mortality
120
20 40 60 80 100
Two-day average N02 concentrations
Note: NO2 = nitrogen dioxide. NO2 concentrations on the x-axis are in the unit of |jg/m3.
Source: Reprinted with permission of Elsevier, (Chen et al.. 2012b).
Figure 5-27 China Air Pollution and Health Effects Study
concentration-response curve for the association between total
and cause-specific mortality and 24-hour average nitrogen
dioxide concentrations at lag 0-1 days.
5-364
-------
The four-city PAPA study (Wong etal., 2010; Wong et al., 2008) also examined the
NO2-mortality C-R relationship but only focused on the shape of the C-R curve in each
individual city. The C-R curve for the NO2-mortality relationship was assessed by
applying a natural spline smoother with 3 dfto NCh concentrations. To examine whether
the NO2-mortality relationship deviates from linearity, the deviance between the
smoothed (nonlinear) pollutant model and the unsmoothed (linear) pollutant model was
examined. The C-R curves in the three Chinese cities further support the results from
Stieb et al. (2008) and Chen et al. (2012b) by indicating a linear relationship between
short-term NC>2 concentrations and mortality (Figure 5-28). Specifically, the evidence for
linearity was strongest between the 25th and 75th percentiles of the NO2 concentrations
in each city with some uncertainty in the shape of the C-R curve at lower concentrations
where the data density is low, generally below the 25th percentile. The results of the
analysis for Bangkok, which provides evidence for nonlinearity, are consistent with what
has been observed in examinations of city-specific C-R curves for other air pollutants
(e.g., PMio, Os). That is, the heterogeneity in city-specific risk estimates can translate into
heterogeneity in the shape of the C-R curve, which has often been hypothesized to be due
to city-specific exposure characteristics and demographics. The results from the Bangkok
analysis highlight the difficulty in interpreting a combined C-R curve across cities, when
there is evidence for city-to-city differences in the association between short-term NO2
exposure and mortality.
5-365
-------
0.3
0.2
_*
V)
•= 0.1
O)
o
0.0
-0.1
Bangkok
0.3
0.2
0.1
0.0
-0.1
Hong Kong
_^-
20 40 60 80 100 120
N02 concentration (ug/m3)
20 40 60 80 100 120 140
N02 concentration (fig/m3)
0.3
0.2
.M
00
•= 0.1
3
0.0
-0.1
Shanghai
0.3
0.2
0.1
0.0
-0.1
Wuhan
50 100 150 200
N02 concentration (ug/m3)
20 40 60 80 100 120
N02 concentration (ug/m3)
Note: ng/m3 = micrograms per cubic meter; NO2 = nitrogen dioxide. Thin vertical lines represent interquartile range of NO2
concentrations in each city. The thick line was included by Wong et al. (2008) to depict where the World Health Organization 1 -year
averaging time standard for NO2 of 40 |jg/m3 (20.8 ppb) could be found along the distribution of NO2 concentrations in each city.
Source: Reproduced from Environmental Health Perspectives, (Wong et al.. 2008).
Figure 5-28 Concentration-response curve for association between total
mortality and 24-hour average nitrogen dioxide concentrations at
lag 0-1 days in the four cities of the Public Health and Air
Pollution in Asia study.
5-366
-------
5.4.8 Summary and Causal Determination
Recent multicity studies evaluated since the completion of the 2008 ISA for Oxides of
Nitrogen continue to provide consistent evidence of positive associations between
short-term NC>2 exposures and total mortality. Although the body of evidence is still
consistent, key uncertainties and data gaps still remain; thus, the evidence for short-term
NO2 exposures and total mortality is suggestive of, but not sufficient to infer, a causal
relationship. This conclusion is the same as that reached in the 2008 ISA for Oxides of
Nitrogen (U.S. EPA. 2008c). Recent multicity studies evaluated have further informed
key uncertainties and data gaps in the NO2-mortality relationship identified in the 2008
ISA for Oxides of Nitrogen including confounding, modification of the NO2-mortality
relationship, potential seasonal differences in NO2-mortality associations, and the shape
of the NO2-mortality C-R relationship. However, questions remain regarding whether
NO2 is independently associated with mortality, specifically due to the lack of copollutant
model analyses with traffic-related pollutants. This section describes the evidence for
total mortality with respect to the causal determination for short-term NO2 exposure,
using the framework described in Table II of the Preamble. The key evidence, as it relates
to the causal framework, is summarized in Table 5-57.
Collectively, the evidence from recent multicity studies of short-term NO2 exposures and
mortality consistently demonstrate the NO2-mortality association is robust in copollutant
models with PMio, Os, or SO2. However, NO2 is often highly correlated with other
traffic-related pollutants complicating the ability to disentangle the independent effects of
NO2 from those of other measured or unmeasured pollutants associated with traffic
(Section 3.4.4 and Figure 3-6). adding uncertainty to the interpretation of the association
between NO2 and total mortality. Studies that focused on PMio and examined whether
NO2 modified the PMio-mortality relationship reported that PMio risk estimates increased
as NO2 concentrations increased or the ratio of NO2/PMio increased. These results
suggest that NO2 and PMio may be effect modifiers of each other. This is consistent with
the conclusions of the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). In addition to
copollutant models, recent studies examined the influence of the extent of temporal
adjustment on NO2-mortality risk estimates and reported similar results across a range of
degrees of freedom per year.
An examination of factors that may contribute to increased risk of NO2-related mortality,
as discussed in Chapter 7. indicate that older adults (>65 years of age), females,
individuals with pre-existing cardiovascular or respiratory diseases, and individuals of
lower SES, specifically lower income and educational attainment, are at greater risk.
Studies that examined whether there are seasonal differences in the NO2-mortality
relationship found greater effects in the warm or summer months in multicity studies
5-367
-------
conducted in Canada and Europe. However, these results are contradicted by a study
conducted in Asia where larger effects were observed in the cold season. These
between-study differences in seasonal associations are more than likely a reflection of the
different seasonal weather patterns observed between countries (Kan et al.. 2010; Kan et
al.. 2008).
Those studies that examined the lag structure of associations for the NCh-mortality
relationship observed that there continues to be evidence of an immediate effect (i.e., lag
0 to 1 day), which is consistent with studies evaluated in the 2008 ISA for Oxides of
Nitrogen. Recent studies also provided evidence for a prolonged effect on mortality in
distributed lag models with lags ranging from 0-4 to 0-5 days (Chen et al.. 2012b;
Chiusolo etal.. 2011). Multicity studies examined the shape of the C-R relationship and
whether a threshold exists in both a multi- and single-city setting. These studies used
different statistical approaches and consistently demonstrated a linear relationship with
no evidence of a threshold within the range of NCh concentrations currently found in the
U.S. However, consistent with observations from C-R analyses conducted for other
criteria pollutants [e.g., PMio (U.S. EPA. 2009a) and O3 (U.S. EPA. 2013e)1. an
examination of the C-R relationship in individual cities, specifically in China,
demonstrated heterogeneity in the shape of the curve across cities (Wong etal.. 2010;
Wong et al.. 2008).
Overall, recent epidemiologic studies build upon and support the conclusions of the 2008
ISA for Oxides of Nitrogen for total mortality. However, the biological mechanisms that
could lead to mortality as a result of short-term NO2 exposures have not been clearly
characterized. This is evident when evaluating the underlying health effects
(i.e., cardiovascular effects in Section 5.3 and respiratory effects in Section 5.2) that
could lead to cardiovascular (-35% of total mortality) and respiratory (~9% of total
mortality) mortality, the causes of total mortality most thoroughly evaluated (Hovert and
Xu. 2012).
Epidemiologic studies that examined the relationship between short-term NO2 exposure
and cardiovascular effects found consistent evidence for myocardial infarction, but
epidemiologic and experimental evidence for other cardiovascular endpoints is
inconclusive. However, important uncertainties remain especially in disentangling
whether there is an independent effect of NO2 on cardiovascular effects, which is the
same uncertainty in total mortality studies. Overall this evidence provides limited
coherence and biological plausibility for NO2-related cardiovascular mortality. For
respiratory effects, there is causal evidence for NO2-related asthma exacerbation
supported by controlled human exposure studies demonstrating increased airway
responsiveness in response to short-term NO2 exposures (Section 5.2.2.1) as well as
5-368
-------
epidemiologic studies reporting associations with asthma-related hospital admissions, ED
visits, and symptoms (Section 5.2.2.4). However, the biological mechanism that explains
the continuum of effects that could lead to respiratory-related mortality also remains
unclear. Additionally, studies that examine the association between short-term NCh
exposures and mortality rely on central site monitors, which may contribute to exposure
measurement error and underestimate NC>2 effects if the temporal variation in ambient
NO2 concentrations reflects variation in people's NO2 exposure (Section 3.4.4.1). In
conclusion, inference from the consistently positive associations observed across various
multicity studies is limited by the uncertainty as to whether NCh is independently
associated with total mortality as well as the uncertainty in the biological mechanism that
could lead to NCh-induced mortality. Thus, the evidence is suggestive of, but not
sufficient to infer, a causal relationship between short-term NO2 exposure and total
mortality.
Table 5-57 Summary of evidence, which is suggestive of, but not sufficient to
infer, a causal relationship between short-term nitrogen dioxide
exposure and total mortality.
Rationale for Causal
Determination3
Key Evidence13
Key References'3
NO2
Concentrations
Associated with
Effects0
Consistent epidemiologic
evidence from multiple,
high-quality studies at
relevant NO2
concentrations
Increases in mortality in multicity Section 5.4.3
studies conducted in the U.S., Canada, jab|e 5.54
Europe, and Asia.
Mean 24-h avg:
9.2-55.0 ppb
Mean 1-h max:
16.3-80.5 ppb
Mean 3-h max:
16.3-42.6 ppb.
Table 5-53
Uncertainty regarding NO2 associations were relatively
potential confounding by unchanged in copollutant models with
traffic-related copollutants PM-io, SO2, orOs, but confounding by
highly correlated traffic-related
pollutants (e.g., EC, CO) or PM2.5 not
examined. Unclear whether NO2 is
independently associated with total
mortality.
tMoolgavkar et al. (2013):
tChenetal. (2012b),
tChiusolo et al. (2011):
tWonqetal. (2010):
tStieb et al. (2008):
tWonq et al. (2008)
Section 3.4.4, Figure 3-6:
Section 5.4.4
NO2 and PM-io may be effect modifiers
of each other.
tKatsouvanni et al.
(2009): Katsouvanni et al.
(2003): Katsouvanni et al.
(2001)
5-369
-------
Table 5-57 (Continued): Summary of evidence, which is suggestive of, but not
sufficient to infer, a causal relationship between short-
term nitrogen dioxide exposure and total mortality.
Rationale for Causal
Determination3
Key Evidence13
Key References'3
NO2
Concentrations
Associated with
Effects0
Uncertainty regarding
exposure measurement
error
Studies that examine the association
between short-term NO2 exposures and
mortality rely on central site monitors.
Sections 3.4.5.1 and 3.5
Uncertainty due to limited
coherence and biological
plausibility with
cardiovascular morbidity
evidence
Consistent epidemiologic evidence for
myocardial infarction. Inconclusive
epidemiologic and experimental
evidence for other cardiovascular
endpoints.
Section 5.3.11
Table 5-52
Uncertainty due to limited
coherence and biological
plausibility with
respiratory morbidity
evidence
Consistent evidence for asthma
exacerbation from experimental studies
demonstrating increased airway
responsiveness and epidemiologic
studies. Uncertainty as to the biological
mechanism that explains the continuum
of effects leading to NCb-related
cardiovascular mortality and respiratory
mortality, which comprise 35% and
-8% of total mortality, respectively.11
Section 5.2.9
Table 5-39
avg = average; CO = carbon monoxide; EC = elemental carbon; max = maximum; NO2 = nitrogen dioxide; O3 = ozone;
PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PMio = particulate matter with
a nominal mean aerodynamic diameter less than or equal to 10 |jm; ppb = parts per billion; SO2 = sulfur dioxide.
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Tables I and N. of the
Preamble.
""Describes the key evidence and references, supporting or contradicting, contributing most heavily to causal determination and,
where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where full body of evidence is
described.
°Describes the NO2 concentrations with which the evidence is substantiated (for experimental studies, below 5,000 ppb).
dStatistics taken from American Heart Association (2011).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
5-370
-------
CHAPTER 6 INTEGRATED HEALTH EFFECTS
OF LONG-TERM EXPOSURE TO
OXIDES OF NITROGEN
6.1 Scope and Issues Considered in Health Effects Assessment
6.1.1 Scope of Chapter
This chapter summarizes, integrates, and evaluates the evidence for a broad spectrum of
health effects associated with long-term exposure (i.e., more than 1 month to years) to
oxides of nitrogen. As in the preceding chapter on short-term exposure, key
considerations in the evaluation include exposure measurement error, effects of other
correlated pollutants, and mode of action information to support biological plausibility.
This chapter comprises evaluations of the epidemiologic and toxicological evidence for
the effects of long-term exposure to oxides of nitrogen on health outcomes related to
respiratory effects (Section 6.2). cardiovascular effects and diabetes (Section 6.3).
reproductive and developmental effects (Section 6.4). and mortality (Section 6.5).
Chapter 6 concludes with a discussion of the evidence for cancer effects (Section 6.6). To
characterize the weight of evidence for reproductive and developmental effects in a
cohesive manner, results from both short-term (i.e., up to 1 month) and long-term
exposure studies are included in this chapter. These results are identified according to
exposure duration in the text and tables throughout Section 6.4.
Individual sections for broad health categories (e.g., respiratory effects) begin with a
summary of conclusions from the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c)
followed by an evaluation of recent (i.e., published since the completion of the 2008 ISA
for Oxides of Nitrogen) studies that builds upon evidence from previous reviews. Within
each of these sections, results are organized into smaller outcome groups (e.g., asthma
development) that are made up of a continuum of clinical to subclinical events and
outcomes. The discussion of individual events and outcomes is then organized by specific
scientific discipline (i.e., epidemiology, toxicology). This organization permits a clear
description of the extent of coherence and biological plausibility for the effects of oxides
of nitrogen on a group of related outcomes, and in turn, provides a transparent
characterization of the weight of evidence in drawing conclusions.
Sections for each of the broad health categories (e.g., respiratory effects, cardiovascular
effects and diabetes) conclude with an integrated assessment of evidence and conclusions
6-1
-------
regarding causality. A determination of causality was made for a broad health category
(e.g., respiratory effects) or smaller group of related outcomes (e.g., birth outcomes) by
evaluating the evidence for each category or group independently with the causal
framework (described in the Preamble). Findings for mortality informed multiple causal
determinations. Findings for cause-specific mortality (i.e., respiratory, cardiovascular)
were used to assess the continuum of effects and inform the causal determinations for
respiratory and cardiovascular effects. A separate causal determination was made for total
mortality (Section 6.5). based primarily on the evidence for nonaccidental causes of
mortality combined but also based on the extent of biological plausibility provided by
evidence for the spectrum of cardiovascular and respiratory effects that are underlying
causes of mortality. Judgments of causality were made by evaluating the evidence for the
full range of concentrations in animal toxicological and epidemiologic studies defined in
this ISA to be relevant to ambient exposure (i.e., up to 5,000 ppb NCh; Section 1.2).
Experimental studies that examined higher concentrations were evaluated mainly to
inform judgments about plausible modes of action.
6.1.2 Evidence Evaluation and Integration to Form Causal Determinations
As was done for relationships of health effects with short-term exposure, judgments
regarding causality were made by evaluating evidence for the consistency of findings
across multiple studies, the coherence of findings across related endpoints and across
disciplines, and the extent to which chance, confounding (i.e., bias due to a correlation
with NO2 exposures or ambient concentrations and relationship with the outcome), and
other biases could be ruled out with reasonable confidence. This evaluation involved
integrating various lines of evidence and a consideration of the strength of inference from
individual studies (detailed in the Appendix to the ISA).
Epidemiologic studies of long-term NC>2 exposure generally rely on differences in
exposure between subjects. For example, studies may base exposure contrasts on
differences in residential location (spatial differences) or time periods that vary in
long-term ambient NC>2 concentrations. For the assessment of potential confounding,
long-term exposure epidemiologic studies were evaluated for the extent to which they
considered other factors associated with health outcomes and correlated with exposures
to oxides of nitrogen. These potential confounding factors can include socioeconomic
status (SES), diet, smoking or exposure to environmental tobacco smoke, medication use,
and copollutant exposures (Appendix). Epidemiologic studies varied in the extent to
which they considered potential confounding. Because no single study examined all
potential confounding factors and not all potential confounding factors were examined in
the collective body of evidence, residual confounding by unmeasured factors is possible.
6-2
-------
Residual confounding is also possible by poorly measured factors. The evidence was
evaluated based on factors well documented in the literature to be associated with NO2
exposure and health outcomes. Other considerations in drawing inferences about the
independent effects of NO2 were the limitations of multivariable models, including
copollutant models, to examine potential confounding (Section 5.1.2.1). Specific to
copollutant confounding, the magnitude of correlations between NC>2 and copollutants
was assessed, and emphasis was placed on particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 um (PIVb 5) and traffic-related pollutants
(e.g., BC/EC, carbon monoxide [CO]). The potential for differential measurement error
for NO2 and copollutants also was considered.
This ISA presents epidemiologic effect estimates for associations with health outcomes
scaled to the same increment of oxides of nitrogen to increase comparability among
studies.1 For long-term exposure metrics, effect estimates are scaled to a 10-ppb increase
in NO2 or NO and a 20-ppb increase in NOx. These increments were derived by
calculating the U.S. nationwide percentile distributions for annual average concentrations
(Table 2-2) and then calculating the approximate difference between the median (a
typical pollution year) and the 95th percentile (a more polluted year) concentrations
among monitors in the State and Local Air Monitoring Stations network. Long-term
averages of ambient oxides of nitrogen are lower in concentration than short-term
averages, less variable across time, and do not differ widely among multimonth averages,
annual averages, or multiyear averages [Supplemental Table S6-1; (U.S. EPA. 2015j)1.
Thus, all long-term exposure metrics were scaled to the same increment. Effect estimates
that were reported in terms of ug/m3 were converted to ppb and standardized for NO2 and
NO but not NOx. Because the proportions of NO2 and NO are unknown for the various
NOx metrics, concentrations could not be converted from ug/m3 to ppb. And data are not
available to calculate the percentiles of NOx concentrations in ug/m3 at a national scale
for the U.S. or other countries. Therefore, the ISA presents effect estimates based on
ug/m3 of NOx as they are reported in individual studies.
To form causal determinations, evidence was integrated across a spectrum of related
endpoints, including cause-specific mortality, and across disciplines to assess the extent
to which chance, confounding, and other biases could be ruled out with reasonable
confidence. Animal toxicological studies can provide direct evidence for health effects
related to NO2 exposures. Coherence between toxicological and epidemiologic findings
can address uncertainties such as whether epidemiologic associations with health
outcomes plausibly reflect an independent effect of ambient NO2 exposure or could be
confounded by other factors. Experimental studies also can provide biological plausibility
1 This is in contrast with reported effect estimates that are scaled to various changes in concentration such as
interquartile range for the study period or an arbitrary unit such as 5 ppb.
6-3
-------
by identifying key events in the modes of action for health effects. Thus, integration of
evidence was used to inform uncertainties for any particular outcome or discipline due to
factors such as publication bias, selection bias, exposure measurement error, or
confounding by copollutant exposures. The subsequent sections assess strength of
inference from studies and integrate evidence across multiple lines of evidence to
characterize relationships between oxides of nitrogen and various health effects.
6.2 Respiratory Effects
6.2.1 Introduction
The 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) examined the epidemiologic
and toxicological evidence on long-term exposure to NO2 and respiratory effects and
concluded that the evidence was suggestive of, but not sufficient to infer, a causal
relationship. The key supporting evidence comprised epidemiologic associations linking
higher NC>2 exposure with decrements in lung function and partially irreversible
decrements in lung development in children. However, several sources of uncertainty
were acknowledged, including confounding by traffic-related copollutants. For example,
the southern California Children's Health Study (CHS) decrements in lung development
in children not only in association with higher ambient NC>2 concentrations but also with
EC, proximity to traffic (<500 m), and PM2 5 (Gauderman et al.. 2004). Because of the
high correlation of long-term averages of NC>2 with such copollutants, an independent
effect of NC>2 could not be discerned in the evidence base as a whole. Animal
toxicological studies demonstrated that long-term exposure to NO2 resulted in permanent
morphologic changes to the lung, particularly in the centriacinar region and bronchiolar
epithelium. However, such effects were indicative of emphysema-like disease and were
not related to the epidemiologic associations observed between NO2 and decreases in
lung function or development in children. Another source of uncertainty was the
inconsistent cross-sectional evidence for associations between long-term exposure to NCh
and increases in asthma prevalence. Epidemiologic studies conducted in both the U.S.
and Europe also reported inconsistent results regarding an association between long-term
exposure to NCh and respiratory symptoms.
This section evaluates the current body of evidence examining the relationship between
long-term exposure to NC>2 and respiratory effects. The strongest evidence is that for
asthma development in children, particularly from recent longitudinal epidemiologic
studies, and is presented first. Evidence for respiratory disease severity, development of
6-4
-------
allergic disease, altered lung function and development, altered lung morphology,
respiratory infections, and COPD is discussed thereafter. No animal toxicological studies
evaluating respiratory effects of long-term NCh exposure have been published since the
release of the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). but previous studies
are evaluated to inform the biological plausibility for the array of respiratory effects
examined.
6.2.2 Development of Asthma or Chronic Bronchitis
Asthma is a chronic disease characterized by inflammation, variable airflow obstruction,
airway hyperresponsiveness (AHR), and in some cases, airway remodeling. To
characterize the evidence for a relationship between long-term NCh exposure and asthma
development, this section evaluates asthma incidence in children in longitudinal cohort
studies. This section also evaluates evidence for airway responsiveness, allergic
responses, and pulmonary inflammation, which are key events in the proposed mode of
action linking long-term NC>2 exposure and asthma development (Figure 4-2). A few
studies examined chronic bronchitis alone or as part of a composite index with asthma.
6.2.2.1 Asthma or Chronic Bronchitis in Children
Recent prospective and retrospective longitudinal cohort studies comprise a strong
evidence base that generally demonstrates a positive relationship between long-term NO2
exposure and asthma incidence in children (Figure 6-1). Unless stated otherwise,
associations were observed with annual average NO2 concentrations. The consistency is
supported by a pooled analysis (Macintyre et al.. 2014a) and many meta-analyses
(Anderson et al., 2013; Gasana et al., 2012; Powers et al., 2012; Takenoue et al.. 2012;
Braback and Forsberg. 2009). However, some meta-analyses included children and adults
as well as both cross-sectional and prospective studies. Longitudinal studies in children
were conducted in North America, Europe, and Asia (detailed in Table 6-1). Most studies
used physician-diagnosed asthma as the indicator for asthma incidence. Cross-sectional
studies of asthma prevalence were reviewed and are discussed to inform understanding of
potential copollutant confounding and other policy-relevant issues [Tables AX6 3-15,
AX6 3-16, and AX6 3-17 of the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c)
describe previous studies]. However, longitudinal studies were emphasized because they
can better characterize the temporality between exposure and incidence of a health effect.
Namely, the prospective designs distinguished between onset of asthma and the
exacerbation of asthma by defining asthma incidence as diagnosis of asthma by a
physician in the time since the previous follow-up period. As described in the sections
6-5
-------
that follow, other strengths of the longitudinal studies are follow-up of children from
birth to ages 7 to 12 years or from ages 8 to 18 years. The reliability of asthma diagnosis
improves in children ages 5 years and older. Additionally, several exposure assessment
methods were used across studies, and many studies aimed to estimate exposure for
individual study participants.
Study
Carlstenetal.(2011) Birth yr
Period of exposure Enrollment Years
age Followed
Gruzievaetal.(2013) Birth yr
Clougherty et al. (2007) Yr of diagnosis
Clark etal. (2010) Birth yr
Gehring et al. (2010) Birth yr
Ranzi et al. (2014)
Oftedal et al. (2009)
Shima et al. (2002)
Birth to age 7 yr
Birth
Birth
Birth
Birth
Birth
Birth
Exposure assessment
LUR
12 Dispersion
At least 4 to 10 LUR
3 to 4
Birth yr Birth
Birth to age 4-10 yr Birth
9 to 10
9 to 10
10-yr avg
around follow-up
Syr
8
Jerrett et al. (2008) Age 14 or 17 yr 10 yr
McConnell et al. (2010) During follow-up 4.8to9yr 3
Central site-IDW
LUR
LUR adjusted for region
LUR unadjusted for
region
LUR
Dispersion
Dispersion
Central site-1 km from
home/school
Residential
Central site-1 per
community
Dispersion
Nishimura etal. (2013) Birth yr
8 to 21 yr None
Central site-IDW
0.5 1 1.5 2 2.5 3 3.5
Risk or Odds Ratio and 95 % Cl
Note: avg = average; IDW = inverse distance weighted; LUR = land use regression; yr = year(s). Black = study from the 2008
Integrated Science Assessment for Oxides of Nitrogen; red = recent studies. Circles = nitrogen dioxide; triangles = nitric oxide;
diamond = sum of nitrogen dioxide and nitric oxide. All effect estimates are standardized to 10 ppb, with the exception of Gruzieva
et al. (2013) who examined NOX in jig/m3 and Oftedal et al. (2009a) who did not report increments for the effect estimates for the
birth to age 4 years or birth to age 10 years exposure periods. See Table 6-1 for study details and quantitative results.
Figure 6-1 Associations of long-term exposure to oxides of nitrogen with
asthma incidence in longitudinal cohort studies of children.
6-6
-------
Table 6-1 Longitudinal studies of long-term exposure to oxides of nitrogen and asthma incidence in children.
Study3
Exposure Assessment Copollutant Correlation Statistical Methods
Comments
Effect Estimates
(95% Cl)b
Vancouver, Canada
tCarlsten et al. (2011 c)
n = 184 children followed
from birth to age 7 yr
Cohort with high risk of
asthma: 1 parent/sibling
with asthma or
2 parents/siblings with
other allergic disease.
Related publications:
Carlsten etal. (2011 a),
Carlsten etal. (2011 b),
Henderson et al. (2007).
Marshall etal. (2008)
LUR model
Annual avg at birth
residence.
Estimates for 1995
generated by temporally
adjusting high resolution
(10 m)2003 annual
averages.
LOOCV (in sample):
Mean error = 0;
SD = 2.75(15%).
R2 for comparisons with
measurements at central
sites and by mobile
monitoring = 0.69, 0.44
(Henderson etal., 2007).
Mean (SD):
17.3 (13.1) ppb
Pearson r:
NO2-PM2.5 = 0.7
NO2-BC = 0.5
NO2-NO = 0.8
NO-PM25 = 0.5
NO-BC = 0.3
LUR models for .
and BC showed poorer
predictive accuracy.
Multiple logistic
regression adjusted for
maternal education,
history of asthma in
mother, father or siblings,
atopic status at age 1 yr.
63% follow-up
participation at age 7 yr.
Key characteristics of
children followed did not
differ from those in the
original cohort.
OR for NO2 among all
children:
2.9(0.8, 10.9)
Association observed
with PM2.5 with wide Cl;
no association observed
with BC.
OR for NO among
13 children with both
allergist diagnosis of
asthma and bronchial
hyperactivity:
1.2(0.9,1.7)
6-7
-------
Table 6-1 (Continued): Longitudinal studies of long-term exposure to oxides of nitrogen and asthma incidence in
children.
Study3
Exposure Assessment Copollutant Correlation Statistical Methods
Comments
Effect Estimates
(95% Cl)b
Children, Allergy, Milieu, Stockholm, Epidemiology Survey (BAMSE), Stockholm, Sweden
tGruzieva et al. (2013)
n = 3,633 children
followed from birth
(1994-1996) to age
12yr.
Related publications:
Gruzieva et al. (2012),
Nordlinq etal. (2008),
Wickman et al. (2002)
Dispersion model
Annual avg NOx at all
residences from birth to
age 12 yr (1994-2008).
Time and activity
patterns used to estimate
exposure.
Model validated against
measurements at central
site monitors. R = 0.74
for correlation of NO2 at
monitors and
traffic-related NO2
estimated from NOx.
r= 0.96 for NOx-PMio
Multinomial
regression/generalized
estimating equation
adjusted for municipality,
SES,year house was
built, mother or father
with doctor diagnosis of
asthma and asthma
medication.
Associations were
stronger for the oldest
children and for
nonallergic asthma.
Follow-up participation:
96%yr1,94%yr2, 91%
yr 4, 84% yr 8, and 82%
yr12.
Study population and the
original cohort had
similar characteristics.
OR for NOx during the
first yr of life and
development of incident
asthma at 12 yr of age.
1.87(1.0, 3.44) per 46.8
ug/m3 NOx
Maternal-Infant Smoking Study of East Boston, MA
tClouqherty et al. (2007)
n = 413 children followed
from birth (1987-1993).
End of follow-up NR.
N = 255 lifetime
residents.
Median age of asthma
diagnosis 5 yr.
LUR model
Annual avg for various
time windows and avg
from birth to diagnosis for
all residences during a
given period.
Model developed based
on monthly
measurements
1987-2004. No
information on validation.
Mean (SD) for year of
diagnosis: 27.5 (4.3) ppb
NR
Regression model
adjusted for maternal
asthma, education, and
smoking before and after
pregnancy, child's sex
and age.
Association limited to
group with high exposure
to violence.
OR for NO2 in the yr of
diagnosis in group with
high exposure to
violence:
3.12(1.36,7.15)
6-8
-------
Table 6-1 (Continued): Longitudinal studies of long-term exposure to oxides of nitrogen and asthma incidence in
children.
Study3
Exposure Assessment Copollutant Correlation Statistical Methods
Comments
Effect Estimates
(95% Cl)b
Southwest British Columbia, Canada
tClarketal. (2010)
n =2,801 children
followed from birth
(1999-2000) to age 3-4
yr.
All births in southwest
British Columbia eligible.
Related publications:
Henderson et al. (2007).
Marshall etal. (2008)
LUR model and IDWof
central site monitors
Annual avg for first yr of
life. LUR estimates with
postal code resolution.
IDWof 3 closest
monitors within 50 km.
LUR model LOOCV (in
sample): Mean
error = 0.0; SD = 2.75
(15%). R2with
measurements at central
sites and sites that vary
in traffic = 0.69, 0.44
(Henderson etal., 2007).
Poorer predictability of
variability due to local
traffic.
Mean (SD):
16.3 (12.3) ppb
Correlations among
pollutants were generally
high. Quantitative results
reported only for Os.
R = -0.9 to -0.7.
Conditional logistic
regression adjusted for
native status,
breast-feeding, maternal
smoking, income quartile,
maternal age, birth
weight, and gestational
length.
To address the lower
reliability of asthma
diagnosis in young
children, asthma was
defined as hospital
admission or at least two
outpatient diagnoses.
OR for IDW:
1.24(1.14, 1.34)
OR for LUR:
1.26(1.08, 1.48)
Associations observed
with CO, BC, and
proximity to point
sources. Traffic-related
pollutants associated
with the highest risks.
ORs were smaller for
PM2.5 than for NO2 for
both LUR and IDW.
Associations also
observed with PM-io, SO2.
6-9
-------
Table 6-1 (Continued): Longitudinal studies of long-term exposure to oxides of nitrogen and asthma incidence in
children.
Study3
Exposure Assessment Copollutant Correlation Statistical Methods
Comments
Effect Estimates
(95% Cl)b
Prevention and Incidence of Asthma and Mite Allergy (PIAMA), the Netherlands
tGehrinq et al. (2010)
n = 3,863 children
followed from birth to age
8 yr
Related publications:
Wijgaetal. (2014),
Eeftensetal. (2011),
Hoek et al. (2008),
Braueretal. (2007).
Braueret al. (2003),
Braueretal. (2002)
LUR model
Annual avg at birth
residence.
Developed from 40 sites:
16 urban/suburban,
12 regional, 12 traffic.
5-10% of population
lived near major roads.
LOOCV (in sample):
R2 = 0.68.
Good agreement among
measured and modeled
NO2 concentrations for
2007 and 1999-2000 at
the same locations.
Mean(10th-90th
percentile):
13.5 (7.8-18.5) ppb
NO2-PM2.5:r=0.93
NO2-soot: r= 0.96
PlVh.s-soot: r= 0.97
Generalized estimating
equations adjusted for
sex, study arm, use of
mite-impermeable
mattress covers,
maternal and paternal
allergies, maternal and
paternal education,
maternal prenatal
smoking, breastfeeding,
presence of a gas stove
in the child's home,
having older siblings, and
any smoking at home.
Follow-up participation:
94.4% yr 1, 82% yr 8.
Characteristics of the
original cohort and
studied groups are
similar.
OR: 1.32(0.93, 1.85)
with adjustment for study
region
OR: 1.36(1.09, 1.67)
without adjustment for
study region
Similar ORs for Plvh.s
and soot.
Gene and Environmental Prospective Study in Italy (GASPII), Rome, Italy
tRanzi etal. (2014)
n = 486 children followed
from birth (2003-2004) to
age 7 yr
Related publication:
Cesaroni etal. (2012)
LUR model
Annual avg at birth
residence, current
residence, lifetime avg.
NO2 measured
simultaneously at
78 locations in winter,
spring, and fall 2007.
LOOCV (in sample):
R2 = 0.66.
Mean(10th-90th
percentile) at age 7 yr:
20.0 (15.5-25.4) ppb.
NO2-O3:
Spearman r= -0.34.
Logistic regression
adjusted for sex, age,
breastfeeding at 3 mo,
day care attendance,
presence of any pets in
the home, siblings,
maternal and paternal
smoking, maternal
prenatal smoking,
maternal and paternal
education, presence of
molds or dampness at
home, familial asthma or
allergies.
Follow-up participation:
99-71% from age 6 mo
to 7 yr.
OR for lifetime avg:
1.17(0.63,2.20)
OR adjusted for Os:
1.11 (0.54,2.25)
6-10
-------
Table 6-1 (Continued): Longitudinal studies of long-term exposure to oxides of nitrogen and asthma incidence in
children.
Study3
Exposure Assessment Copollutant Correlation Statistical Methods
Comments
Effect Estimates
(95% Cl)b
Oslo, Norway
tOftedal et al. (2009a)
n =2,329 children
followed from birth
(1992-1993) to age
9-1 Oyr.
Related publications:
(Oftedaletal..2009b),
Laupsa and Slordal
(2003). Walker et al.
(1999), Gronskei et al.
(1993)
Dispersion model
Annual avg at residence
in first yr of life, birth to
asthma diagnosis, and yr
before diagnosis.
Modeled estimates well
correlated with
measurements from
10 central site monitors.
R = 0.76.
Mean (25th-75th
percentile) for first yr of
life:
20.9 (13.2-28.1) ppb
NO2-PM2.5 & NO2-PM10:
r= 0.79-0.91.
Cox proportional hazard
regression and logistic
regression adjusted for
sex, parental atopy,
maternal smoking in
pregnancy, paternal
education, and maternal
marital status at the
child's birth.
RR for NO2 in first yr of
life and asthma onset at
any age:
0.87(0.76, 1.00)
Average NO2 before
diagnosis not associated
with asthma diagnosed at
any age or after age 4 yr.
Chiba prefecture, Japan
Shima et al. (2002)
n = 1,910 children in
eight communities
followed from 1st grade
(age 6 yr)to 6th grade.
Enrolled 1989-1992.
Central site monitors
10-yr avg (1988-1997).
Almost all children's
homes and schools were
about 1 km from sites.
Range across
communities:
7.3-31.4 ppb
NR
Logistic regression
adjusted for sex, history
of allergic diseases,
respiratory diseases prior
to age 2 yr, parental
history of allergic
diseases, maternal
smoking habits, type of
heater used in winter in
the home, and
construction elements of
the house.
67% follow-up
participation. Lower
follow-up among urban
children than rural
because of more
frequent changes in
residence. Exposure data
not calculated for 944
children due to missing
residential data for 3-yr
period before enrollment
(missing questionnaire).
OR: 1.71 (1.04,2.79)
6-11
-------
Table 6-1 (Continued): Longitudinal studies of long-term exposure to oxides of nitrogen and asthma incidence in
children.
Study3
Exposure Assessment Copollutant Correlation Statistical Methods
Comments
Effect Estimates
(95% Cl)b
Children's Health Study (CHS), southern California
tJerrett et al. (2008)
n = 217 children followed
from ages 10 to 18 yr.
Enrolled in 1993 or 1996
from 11 southern
California communities.
Asthma assessed over
8 yr of follow-up.
Residential outdoor
monitors.
Season and annual avg
estimated from Palmes
tubes outside home for
2 weeks summer and
winter.
Mean (SD) annual avg
NO2 range across
11 communities:
9.6(2.5)-51.3(4.4)ppb
No quantitative data.
Correlations of residential
NO2 with measures of
traffic proximity or
modeled pollutant
concentrations reported
to be moderate to high.
Random-effects Cox
proportional hazards
model adjusted for
median household
income, proportion of
respondents with low
education, percentage of
males unemployed,
percentage living in
poverty, temperature,
and humidity.
Within-community effects HR: 1.51 (1.12, 2.05)
(deviation of individual
from mean) similar to
between-community
effects (community
mean). Suggests
influence of both regional
and local pollution.
Smaller range in NO2
within communities than
between communities.
tMcConnell et al.
(201 Oa)
n = 2,497 children ages
4.8-9.0 yr followed for 3
yr.
Enrolled 2002-2003 in
13 southern California
communities.
Related publications:
Wu et al. (2005). Peters
et al. (1999), Benson
(1984)
NO2: central site, 1 per NR
community
NOx: dispersion model
for home and school.
Annual avg, time period
Within-community
variability of personal
NO2 estimates:
± 20-40%. Lower for EC,
PMio, PM2.5, and CO (Wu
etal., 2005).
For different CHS cohort,
residential NO2 and
freeway-related NO2 from
dispersion model
correlated with r= 0.56
(Gauderman et al..
2005). Residential and
modeled NO2 may have
some level of
independence.
Mean (range):
20.4 (8.7-23.6) ppb
Multilevel Cox
proportional hazards
model adjusted for
sociodemographic
characteristics, exposure
to cigarette and wildfire
smoke, health insurance,
housing characteristics,
history of allergy, and
parental asthma.
74% follow-up
participation. Lower
follow-up among
Hispanic children and
children with lower SES.
NO2 association was
similar after adjusting for
these factors.
Risk for NOx was higher
in children with high
parental stress compared
to low parental stress
(Shankardass et al.,
2009).
HR for central site NO2:
1.39(1.07, 1.80)
HR for modeled NOx
from freeways:
1.67(1.32, 2.12) near
homes
1.88(1.10, 3.19) near
schools.
OR for PM2.5 central site
(per 3.5 ug/m3):
1.66(0.91, 3.05)
6-12
-------
Table 6-1 (Continued): Longitudinal studies of long-term exposure to oxides of nitrogen and asthma incidence in
children.
Study3
Exposure Assessment Copollutant Correlation Statistical Methods
Comments
Effect Estimates
(95% Cl)b
Genes-environment & admixture in Latino Americans and the study of African Americans, asthma, genes, & environments
tNishimura etal. (2013)
n = 4,320, ages 8-21 yr
Multicity study: Chicago,
IL; Bronx, NY; Houston,
TX; San Francisco Bay
Area, CA; Puerto Rico.
IDW of central site
monitors
Average for first yr and
first 3 yr of life.
IDW estimates based on
4 closest monitors within
50 km of home.
Mean (SD) across cities:
9.9 (2.9) to 32.1 (5.7)
NR
Logistic regression
adjusted for age, sex,
ethnicity, and composite
SES. Sensitivity analysis
conducted with: maternal
prenatal smoking,
smoking in the household
ages 0-2 yr, and
maternal language of
preference.
Associations varied
among cities. Cities
varied in racial/ethnic
make-up of study
population, air pollution
concentrations.
OR for first yr of life, all
cities combined:
1.37(1.08,1.73)
OR for PM2.5
(per 1 ug/m3):
1.03(0.90, 1.18)
BAMSE = Children, Allergy, Milieu, Stockholm, Epidemiology Survey; BC = black carbon; CHS = Children's Health Study; Cl = confidence interval; CO = carbon monoxide;
EC = elemental carbon; GASPII = Gene and Environmental Prospective Study in Italy; HR = hazard ratio; IDW = inverse distance weighting; LOOCV = leave-one-out cross
validation, LUR = land use regression; NO = nitric oxide; NO2 = nitrogen dioxide; NOX = sum of NO and NO2; NR = not reported; OR = odds ratio; PIAMA = Prevention and Incidence
of Asthma and Mite Allergy; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PMio = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 10 |jm; RR = risk ratio, relative risk; SD = standard deviation; SES = socioeconomic status.
aStudies are presented in the order of appearance in the text.
""Results are presented for a 10 ppb change in NO2 unless otherwise specified.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
6-13
-------
The majority of studies administered an annual questionnaire that asked parents whether
a doctor has ever diagnosed the child as having asthma. Asthma incidence was defined as
a diagnosis at follow-up, with no diagnosis at any previous time of evaluation. The use of
questionnaires to determine asthma incidence is a best practice (Burr. 1992; Ferris. 1978)
and adds to the strength of inference from the available studies. Carlsten et al. (201 la) is
particularly noteworthy for having a pediatric allergist apply uniform criteria to assess
asthma in children when they were 7 years old. NC>2 also was associated with a composite
measure of doctor-diagnosed asthma or asthmatic/spastic/obstructive bronchitis (Kramer
et al.. 2009). which could represent conditions other than asthma. Among children age 12
years, Gruzieva et al. (2013) found an association of NOx estimated from a dispersion
model with asthma incidence defined as at least 4 episodes of wheeze in the last
12 months, or at least one episode in combination with prescription of inhaled
corticosteroids. Although wheeze may not necessarily indicate an asthma diagnosis, a
prescription for corticosteroids likely would have resulted from a physician making a
diagnosis of asthma. An uncertainty in this study is whether the NOx exposure estimate
represents NO2 exposure equally among subjects. Further, the very high correlations that
have been observed for dispersion model estimates of NOx with PIVb 5 as well as
traffic-related copollutants such as CO and EC (Table 6-1) produces large uncertainty in
attributing associations to NOx specifically.
Transient wheezing is common in infants and often resolves as the child ages (Martinez
et al.. 1995). Thus, asthma diagnosis in infants and young children may have lower
reliability. As a child progresses in age, the reliability of diagnosis of asthma would be
expected to improve as would the strength of inference regarding associations with NO2
exposure. Although a few studies indicated asthma assessment in young children (Clark
et al.. 2010; Clougherty et al.. 2007). a strength of the evidence base is the many studies
that followed children to age 7 to 18 years to ascertain asthma (Figure 6-1 and Table 6-1).
Consistent with reliability of asthma diagnosis increasing with age, Gehring etal. (2010)
observed associations of NO2 that were greater in magnitude at later age evaluation and
with longer follow-up time. Results from models with air pollution-age interaction terms
indicated small age-related differences, but larger odds ratios (ORs) were observed at
ages 6-8 years (Figure 6-2). Gruzieva et al. (2013) found an association with asthma
incidence at age 12 years but not at earlier ages. However, the association was with NOx.
In contrast, a few studies did not observe associations of NO2 with asthma diagnosed at
any age examined (Ranzi etal.. 2014; Oftedal et al.. 2009a).
6-14
-------
Incident asthma past 12 mo.
Tol TT J,i
-
)
4
, 1
' fo
iti " 11J
> C^
)C
>«
c
1
>^
)«
1
(
|
I
4
1 *
><
'<
i
$1$
I!
-3 3.6
.2 ^
2 .S 2.8
^ 2 exposure in the first 3 years of life (Nishimura et al.. 2013) or year
of diagnosis (Clougherty et al.. 2007). Gruzieva et al. (2013) observed that asthma was
associated with NOx for the first year of life but not with average NOx concentrations
since the date of the previous follow-up or during the preceding 12 months. Both of the
latter exposure periods were periods of lower exposure. Often, the various early life
exposure periods that are evaluated are highly correlated with one another, making it
difficult to interpret the results or identify a single exposure window of concern.
Exposure measurement error also may vary across time periods due to changes in
time-activity patterns, for example.
6-15
-------
Confidence in the evidence base also is based on the high follow-up participation
reported in many studies. The range across studies was 63 to 82%. The high follow-up
participation increases the likelihood that the results are representative of the original
study population. No study indicated that results could have been unduly influenced by
selective drop-out due to exposure or asthma status. An uncertainty in the evidence base
is whether there is a monotonic concentration-dependent increase in NCh-related asthma
risk. Analysis of the concentration-response relationship is limited and does not clearly
indicate a concentration-dependent increase. Shima et al. (2002) observed a linear
relationship for NCh-related asthma across communities in Japan, with higher asthma
incidence observed in communities with higher 10-year average NC>2 concentrations.
Analyzing individual subjects, Carlsten et al. (201 Ic) observed higher risk estimates in
the second and third tertiles of NO2, but the very wide and overlapping CIs (Table 6-1)
do not strongly demonstrate a linear relationship. These studies did not conduct analysis
to evaluate whether there is a threshold effect.
As described in Section 3.4.5.2. misrepresenting the differences between subjects in
long-term NC>2 exposure due to the high variability in ambient concentrations observed in
many locations can bias health effects associations. Thus, a key issue in evaluating the
strength of inference about NO2-related asthma development from epidemiologic studies
is the extent to which the NC>2 exposure assessment method used in a study captured the
variability in exposure among study subjects. The set of studies examining asthma among
children used a variety of exposure assessment techniques including land use regression
(LUR) models, monitors outside each subject's home, a single or nearest community
monitoring site, monitoring site measurements combined by inverse distance weighting
(IDW), and dispersion models (Table 6-1). These studies of asthma incidence
demonstrate various levels of effort to design exposure assessments that might
characterize exposure for individual subjects. However, the exposure assessment of many
studies of asthma development provide a strong basis for drawing inferences about the
relationship between long-term NC>2 exposure and asthma development in children. Such
judgments are based on how well studies considered the strengths and limitations of the
exposure assessment method as described in Section 3.4.5.2.
Because children spend a large portion of time near their home, NO2 concentrations at
home have the potential to represent exposure well. Many recent longitudinal studies
found increased risk of asthma incidence in association with ambient NC>2 concentrations
estimated outside each subject's home using LUR models (Table 6-1), including a pooled
analysis of six birth cohort studies (5,115 children) (Macintyre et al.. 2014a). Annual
average NO2 assigned to each child's birth address by LUR was associated with asthma
(up to ages 7-8 years) with an odds ratio (OR) of 1.48 (95% CI: 1.06, 2.06) per 10-ppb
increase in NC>2. Many LUR models were developed based on a large number of sites that
6-16
-------
varied in urbanicity and traffic to represent the range locations in the study area,
measurements of multiple seasons, and various traffic indicators (Cesaroni et al.. 2012;
Cloughertv et al.. 2007; Henderson et al.. 2007; Brauer et al.. 2003).
Some individual studies demonstrated LUR models to have good accuracy with respect to
predicting the spatial pattern of ambient NO2 concentrations in the study area. As shown
in Table 6-1, leave-one-out cross validation (LOOCV) indicated good to excellent
predictive accuracy (Ranzi et al.. 2014; Carlsten et al.. 201 Ic: Gehring etal.. 2010). For
example, in the Netherlands cohort, the LOOCV for the LUR model had an R2 of 0.68
(Eeftens et al.. 2011; Gehring et al.. 2010). In many studies, the LUR model was
developed with NO2 measurements taken during a different time period as the exposure
period examined in relation to asthma. The Netherlands study demonstrated the temporal
validity of the LUR model. There was good agreement among LUR-predicted NO2
concentrations and NO2 measurement at the same sites in 1999-2000 and 2007 (Eeftens
et al., 2011; Gehring et al., 2010). NO2 exposures for the Italian cohort also were
estimated with an LUR model shown to have good predictive accuracy [LOOCV
R2 = 0.67; (Ranzi etal.. 2014; Cesaroni et al., 2012)1. and no association with asthma was
observed. Most of the LOOCV procedures were based on removing one measurement
from the sample of sites used to develop the LUR model. This in-sample validation may
not represent how well the model can predict NO2 concentrations at other sites not used
in the model. For the Vancouver, Canada cohort, LOOCV was performed with NO2
measured at nonLUR sites. The model had better predictive accuracy for central site
monitors than sites with low or high traffic [R2 = 0.69 versus 0.44; (Carlsten et al.. 201 Ic;
Henderson etal.. 2007)1.
Unlike the aforementioned LUR studies, Clark etal. (2010) assigned NO2 exposure at the
postal code or block levels, which have the potential for greater exposure measurement
error than estimates outside the home (Section 3.4.5.2). Clark etal. (2010) did not
provide information on whether the block-level estimates adequately represented the
variation in NO2 concentrations in the study area or between-subject differences in
ambient NO2 exposure. Information on model validation was not reported for the East
Boston cohort either (Cloughertv et al.. 2007). Thus, these studies have weaker inference
compared to the other LUR studies. However, the well-validated NO2 exposure estimates
in Gehring etal. (2010) and Carlsten et al. (201 Ic) provide a good basis for inferring a
relationship between NO2 exposure and asthma development in children.
Adding to the LUR studies are studies that related asthma incidence to NO2
measurements spatially aligned with subjects' homes and/or schools. The spatial
alignment of monitors to children's residences and/or schools increases confidence in the
NO2 metrics to capture between-subject variation in NO2 exposures. Although NO2
6-17
-------
measurements from central site monitors have well-known limitations in capturing the
spatial heterogeneity in concentrations within an area (Section 3.4.5.2). Shima et al.
(2002) reduced the potential for exposure measurement error by estimating NC>2 exposure
from sites located 1 km (for a majority of subjects) to 2 km from a subject's home and/or
school. Such estimates may better capture the spatial heterogeneity in ambient NO2
concentrations compared to concentrations from a single site in the community or
averaged across monitoring sites.
In a CHS cohort, asthma incidence was associated with annual average NO2 exposure
estimated from two 2-week passive sampling periods outside children's homes (Jerrett et
al., 2008). However, the NO2 monitoring for some children occurred after asthma
diagnosis, and information was not reported on the extent of temporal mismatch. While
long-term trends in NC>2 concentrations tend to be similar within a community, it is not
known what percentage of the cohort moved to different communities during follow-up.
In a separate CHS cohort, asthma incidence was associated with NO2 measured at one
central site per community (McConnell et al.. 2010a). In the study area, NC>2
concentrations show high within-community variability (Table 6-1). Thus, it is unknown
how well the central site measurements represent between-subject variation in exposure.
Jerrett et al. (2008) modeled the effects of the within- and between-community variation
in NO2 This approach allowed examination of the potentially different contributions of
local NC>2 and regional NCh to the associations with asthma. Both within-community
variation and between-community variation in NC>2 were associated with the
development of asthma. The study did not examine whether associations for NO2 are
independent of traffic-related copollutants or PM2 5; thus, the results provide evidence
that both regional and local pollution contributed to the observed associations.
In contrast with Shima et al. (2002). other studies assigned NO2 exposure based on
central site measurements with coarser spatial resolution by combining concentrations
across monitors by inverse distance weighting (IDW) (Nishimura et al.. 2013; Clark et
al.. 2010). IDW is used to account for spatial variability in that greater weight is placed
on measurements in closer proximity to subjects. However, the representativeness of
IDW estimates to the fine-scale spatial pattern of NO2 concentrations may vary across
locations, depending on the presence of localized sources between measurement sites
(Section 3.2.1.1). Studies in the U.S. and British Columbia, Canada aimed to characterize
NC>2 exposure for individual subjects or at the postal code level by combining NO2
concentrations across the three or four closest ambient monitors within 50 km by IDW
(Nishimura et al.. 2013; Clark et al.. 2010). These studies did not provide information to
assess the extent to which the IDW estimates based on a 50-km buffer captured local
sources in the study areas or represented between-subject variation in NO2 exposure, and
thus the studies have uncertainty regarding associations with asthma incidence. For the
6-18
-------
British Columbia, Canada cohort, Clark et al. (2010) found similar ORs for asthma
incidence in relation to NC>2 exposures estimated by LUR and IDW; however, both
methods estimated exposures at the postal code level and may have similar uncertainty in
representing subjects' exposures. In the British Columbia, Canada study area, NO2
concentrations estimated by LUR and IDW for the postal code spatial scale agreed well
with concentrations at central site monitors (Marshall et al.. 2008). but it is unclear
whether the LUR or IDW estimates represent NO2 variability at smaller spatial scales.
Dispersion models have the potential to produce uncertain exposure estimates due to
inaccuracies in estimating within-community conditions that result from simplifying
assumptions of the NOx reaction model and meteorological conditions (Section 3.4.5.2).
However, these limitations do not necessarily apply across locations. Oftedal et al.
(2009a) did not observe an increased risk of asthma development in children in relation to
long-term NC>2 exposure. However, there is indication that the NC>2 exposure metrics
captured some spatial variability in NO2 within the study area. There was high correlation
(r = 0.76; Table 6-1) in NC>2 concentrations estimated by dispersion model and measured
at 10 central site monitors in the city.
The ability to distinguish among the effects of correlated factors is important in making
inferences about the effects of NO2 on asthma development. The longitudinal studies of
children observed associations with NO2 after adjusting for various confounding factors
such as sociodemographic and housing characteristics, cigarette smoking exposure,
history of asthma and parental asthma, and education. A key concern for NC>2 is potential
confounding by PIVb 5 and traffic-related pollutants such as CO, EC, and ultrafine
particles (UFP). No studies of NCh and asthma development examined copollutant
models with traffic-related copollutants or with PM2 5 In Japan, Hasunuma et al. (2014)
linked a reduction in ambient NC>2 concentrations from 1997 to 2009 to a reduction in the
prevalence of asthma. While these results support an association of asthma with NO2, key
copollutants were not examined. The limited information available on the potential for
confounding by key copollutants is the correlations reported with NC>2 and associations
for copollutants. Hwang et al. (2005) did not analyze copollutant models with
traffic-related copollutants but observed that the association for NOx with asthma
prevalence remained relatively unchanged with adjustment for either sulfur dioxide
(802), PMio, or Os. The cross-sectional design, lack of analysis of NO2, and use of central
site measurements limits inferences from the copollutant model results from this study.
For key copollutants, the level of detail on correlations varies across studies. In some
cases, the correlations are not reported or a statement of moderate to high correlation is
reported without quantitative results (Table 6-1). No data were reported for the
correlations of NO2 with CO or UFP. For PIVb 5, correlations range from about 0.7 to
6-19
-------
0.93. The strong correlations often observed make it difficult to distinguish whether there
is an independent effect for NC>2. For the Vancouver cohort, residential estimates of BC
and NO2 were moderately correlated (r = 0.5) (Carlsten et al.. 201 Ic). However, the
lower predictive accuracy for the LUR model for BC may result in differential exposure
error and misrepresent the correlation with NC>2. Thus, the moderate correlation may not
reflect the true potential for confounding.
PM2 5 was evaluated in several of the studies relating NO2 exposure to asthma incidence
in children (Table 6-1). McConnell et al. (2010a). assigning exposure from a single
community central site, observed a smaller risk estimate for PIVb 5 than for NC>2 and a
wider CI. They did not report a correlation between NO2 and PIVb 5. The association for
NO2 was attenuated with adjustment for dispersion model estimates of residential NOx,
which was nearly perfectly correlated with traffic-related copollutants such as EC and
CO. NOx results were not altered by NO2 or PM2 5 adjustment. These results may reflect
differences in exposure measurement error for central site NO2 and residential estimates
of NOx or may indicate that NOx represents NO2 as well as other traffic-related
pollutants. Carlsten et al. (201 Ic) reported a correlation between NO2 and PM2 5 of
r = 0.7; and NO2 and BC of r = 0.5. A larger OR but wider 95% CI was observed for
PM2.5 than NO2, and no risk for BC. However, the LUR models for PIVb 5 and BC were
demonstrated to have poor predictive accuracy. Nishimura et al. (2013) did not report
correlations between pollutants and observed a smaller OR for PM2 5 than NO2, both of
which were estimated by IDW. Gehring et al. (2010) reported a correlation between NO2
and PM2s of 0.93 and observed similar ORs for NO2 and PM2s. Both pollutants were
estimated by LUR models that were shown to have similar predictive accuracy. Clark et
al. (2010) reported high correlations between NO2 and the other pollutants but did not
provide quantitative data. Based on exposure assessment by LUR and IDW at the postal
code level, Clark etal. (2010) observed PM2 5 effect estimates that were smaller than
those for NO2 with wider 95% CIs. Thus, stronger effect estimates with smaller CIs were
generally observed for NO2 than for PM2 5. However, for BC estimated by LUR, the odds
ratio was larger than that for NO2. Risk estimates were not always larger for NO2
compared with BC. Potential differential exposure measurement error in many studies
between NO2 and PM2 5 or BC may limit inferences that can be drawn about confounding
by comparing the magnitude of associations with asthma.
Long-term exposure to ambient NO2 is consistently associated with the development of
asthma in children as examined in several longitudinal studies. Associations are observed
with various periods of exposure, including the first year of life, the year prior to asthma
diagnosis, and cumulative exposure. Strengths of the evidence base include the general
timing of asthma diagnosis, which lends confidence that the NO2 exposure preceded
asthma development. Further, physician-diagnosed asthma, whether by parental report or
6-20
-------
clinical assessment, along with the assessment of children older than age 5 years provides
reliability in the outcome to represent asthma development. Also providing a good basis
for inferring associations of NC>2 with asthma incidence, some studies estimated
residential NO2 from LUR models that were demonstrated to predict well the variability
in NCh in study locations or examined NC>2 measured at locations 1-2 km of subjects'
school or home (Table 6-1). A key uncertainty that remains when examining the
epidemiologic evidence alone is the inability to determine whether NO2 exposure has an
independent effect from that of other pollutants in the ambient mixture. The strong
correlations reported between NO2 and PIVbs, the paucity of data on correlations with
other traffic-related pollutants such as CO and EC/BC, and the lack of examination of
potential confounding by traffic-related copollutants introduces uncertainty of the extent
to which NC>2 has an independent effect or serves primarily as a surrogate for other
highly correlated pollutants based on just epidemiologic results.
6.2.2.2 Asthma or Chronic Bronchitis in Adults
Since the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). recent longitudinal
studies have examined asthma and chronic bronchitis in adults in relation to long-term
exposure to NO2 and observed positive associations. In the European Community
Respiratory Health Survey (ECRHS) cohort, this relationship in adults was examined
using various definitions of asthma or chronic bronchitis and exposure assessment
approaches.
Home outdoor NO2 was associated with chronic bronchitis defined as productive
(phlegm-producing) chronic cough more than 3 months each year (Sunyer et al.. 2006).
The follow-up time period was 8.9 years. Indoor kitchen and outdoor (at the kitchen
window) residential NO2 was measured using Palmes tubes during a 14-day period in
1,634 households of ECRHS subjects who did not change residences during the
follow-up. This was repeated in 659 households (45%) 6 months later. A linear
concentration-response relationship with NO2 was observed only in females. NO2 was
associated with chronic bronchitis after adjustment for traffic intensity at the residence.
New onset asthma was related to NO2 exposure in the ECRHS adult cohort as ascertained
by a positive response to the question "Have you ever had asthma?" (Jacquemin et al..
2009b) and a continuous asthma score (Jacquemin et al.. 2009a). Asthma incidence was
defined as reporting asthma in the follow-up (1999 to 2001) but not at baseline (1991 to
1993). Outdoor NO2 estimates developed from NOx emissions were linked to subjects'
home addresses. For "ever having asthma," the adjusted OR was 1.96 (95% CI: 1.04,
6-21
-------
3.70) per 10-ppb increase in NC>2. The OR for asthma incidence based on the ratio of the
mean asthma score was 1.48 (95% CI: 1.18, 1.85) per 10-ppb increase in NC>2.
In a preliminary examination of a smaller group of the ECRHS cohort, the prospective
Respiratory Health in Northern Europe cohort study, Modig et al. (2009) used dispersion
models and two definitions of asthma for 3,824 adults (ages 20-44 years at baseline):
(1) the cumulative number of onset cases of asthma and (2) incident cases of asthma.
Asthma was defined as no asthma attacks during the last 12 months and no current use of
asthma medication at the start of the study but having asthma or ever being diagnosed
with asthma at follow-up. NCh concentrations estimated at the home were associated with
risk of developing asthma. The OR was 2.04 (95% CI: 1.14, 3.65) for the cumulative
number of onset cases of asthma and 2.25 (95% CI: 1.00, 5.07) for the incident definition
of cases per 10-ppb increase in NO2 concentration. The OR for asthma increased across
NO2 tertiles, indicating a concentration-dependent relationship. With the first tertile as
the reference, the OR was higher for the third tertile (ORonset 1.58 [95% CI: 0.96, 2.6];
OR^ent 2.06 [95% CI: 0.98, 4.32]) than for the second tertile (ORonset 1.17 [95% CI:
0.70, 1.94]; ORncMent 1.77 [95% CI: 0.86, 3.64]).
Castro-Giner et al. (2009) prospectively examined asthma incidence and prevalence in
the large (2,577 subjects at follow-up) ECRHS cohort from 13 European cities. In the
longitudinal analysis, for the 120 subjects who developed asthma during the follow-up
period, NO2 was associated with new-onset asthma with an OR of 2.20 (95% CI: 1.17,
4.10) per 10-ppb increase. For asthma prevalence, an association was indicated among
subjects who changed homes rather than subjects who lived in the same home during
follow-up (movers OR: 2.53 [95% CI: 1.16, 5.56]; nonmovers OR: 1.04 [95% CI: 0.66,
1.64]). However, for new-onset asthma, evidence for association was stronger among
nonmovers than movers [nonmovers OR: 2.39 (95% CI: 1.10, 5.22); movers OR: 2.09
(95% CI: 0.70, 6.12)].
In a meta-analysis, Cai et al. (2014) cross-sectionally assessed the associations of outdoor
air pollution on the prevalence of chronic bronchitis symptoms in adults in five cohort
studies participating in the European Study of Cohorts for Air Pollution Effects
(ESCAPE) project. Annual average NO2, NOx, as well as PMio, PIVb 5, PMabSOrbance, and
PMCoarse from 2008-2011 were assigned to home addresses by LUR. Symptoms examined
were chronic bronchitis (cough and phlegm for >3 months of the year for >2 years),
chronic cough (with/without phlegm), and chronic phlegm (with/without cough). Overall,
there were no associations with any air pollutant or traffic exposure.
In summary, among adults, long-term NO2 exposure generally is associated with asthma
incidence and chronic bronchitis. The longitudinal design of the studies lends strength to
the interpretation of results. Studies aimed to produce individual estimates of exposure at
6-22
-------
the residence, and in most cases, NO2 exposures estimated from dispersion models were
demonstrated to be well correlated with measured concentrations in the study locations
(R2 = 0.60-0.79). However, the strength of inference is limited because findings are
based on one cohort, and none of the studies in adults considered confounding by PIVb 5
or traffic-related pollutants.
6.2.2.3 Subclinical Effects Underlying Development of Asthma or Chronic Bronchitis
Animal toxicological studies demonstrate that long-term NO2 exposure enhances both
responsiveness of airways and the development of allergic responses. Some animal
toxicological studies and epidemiological studies of long-term exposure show increases
in pulmonary inflammation and oxidative stress. Thus, there is some evidence that
suggests a mechanistic basis for the development of asthma in relation to NO2 exposure.
Airway Responsiveness
Animal toxicological studies have demonstrated that NO2 exposure enhances
responsiveness of airways to nonspecific and specific challenges. A subchronic-duration
study demonstrated concentration-dependent increases in airway responsiveness to
histamine in NCh-exposed guinea pigs (Kobayashi and Miura. 1995). In this study, one
experiment demonstrated AHR after 6 weeks of exposure to 4,000 ppb, but not 60 or
500 ppb NO2. In another experiment, AHR was observed in guinea pigs exposed to
4,000 ppb NC>2 for 6 weeks; 2,000 ppb for 6 and 12 weeks; and 1,000 ppb for 12 weeks.
Specific airways resistance in the absence of a challenge agent was increased in guinea
pigs exposed to 2,000 and 4,000 ppb NO2 for 12 weeks. AHR occurring with increased
airway resistance suggests the involvement of airway remodeling. Another
subchronic-duration exposure study found delayed bronchial responses, measured as
increased respiration rate, in guinea pigs sensitized and challenged with C. albicans and
exposed to NCh [4,760 ppb, 4 h/day, 5 days/week, 6 weeks (Kitabatake etal.. 1995)].
However, NO2 exposure (4,000 ppb, 2 h/day, 3 months) failed to alter airway
responsiveness to a nonspecific challenge in rabbits sensitized at birth with house dust
mite antigen (Douglas et al., 1995). Overall, results are consistent with those in rodents
with short-term exposure to NC>2 (Section 4.3.2.5). In addition, they are supported by
findings for effects underlying development of AHR, including inflammation, and
allergic sensitization (Section 4.3.2).
6-23
-------
Development of Allergic Responses
lexicological studies provide some evidence that is coherent with the development of
allergic responses seen in some of the epidemiologic studies (Section 6.2.4). One
subchronic-duration study showed that exposure to 4,000 ppb NC>2 for 12 weeks led to
enhanced immunoglobulin E (IgE)-mediated release of histamine from mast cells isolated
from guinea pigs (Fujimaki and Nohara. 1994). This response was not found in mast cells
from rats similarly exposed in the same study. Furthermore, two short-term studies
provide evidence that exposure to NO2 leads to T-derived lymphocyte helper 2 (Th2)
skewing and/or allergic sensitization in healthy human adults exposed repeatedly to 2,000
ppb NO2 and in naive animals exposed repeatedly to 3,000 ppb NO2, as discussed in
Sections 4.3.2.6 and 5.2.7.4 (Pathmanathan et al., 2003; Ohashietal.. 1994). Findings of
increased histamine release from mast cells, increased nasal eosinophils, and increased
Th2 cytokines seen in humans and animal models exposed to NC>2 provide support for the
epidemiologic evidence relating NO2 exposure to asthma development and the findings in
some of the epidemiologic studies for the association of NO2 exposure with the
development of allergic responses.
Pulmonary Inflammation and Oxidative Stress
Inflammation and oxidative stress are identified as key events in the proposed mode of
action for development of asthma (Figure 4-2). Long-term NC>2 exposure has been shown
to induce pulmonary inflammation or oxidative stress in toxicological and epidemiologic
studies, but results are not entirely consistent. Similarly, there is some evidence for a
relationship of short-term exposure to NC>2 with pulmonary inflammation and oxidative
stress (Section 5.2.7.4) to describe a potential pathophysiologic basis for development of
asthma in response to repeated NO2 exposures.
Epidemiologic Evidence in Children
In the CHS cohort of 1,211 schoolchildren from eight southern California communities,
annual average NO2 was associated with a longitudinal increase in exhaled nitric oxide
(eNO; using a flow rate of 50 mL/sec) in 2006-2007 and 2007-2008 (Berhane et al..
2014). This association was observed with adjustment for short-term NC>2 assessed from
central monitoring sites and was independent of asthma status. Based on prior findings in
CHS (Bastain et al.. 2011) that elevated eNO is associated with increased risk of new
onset asthma, an effect of long-term exposure to NO2 on increases in eNO over time is
consistent with a role for NO2 in asthma pathogenesis. However, NO2 exposure was
estimated from a single monitor in each of the study communities, and confounding by
PM25 or traffic-related pollutants was not examined.
6-24
-------
Using LUR models to estimate annual average NO2 exposure, Liu et al. (2014a) observed
null associations with eNO among all children (n = 1,985, age 10 years, ESCAPE) and
those without asthma (n = 1,793) in both the single and copollutant (PMio) models. NC>2
was positively associated with eNO in the 192 children with asthma.
Using a cross-sectional prevalence design, Dales et al. (2008) examined the relationship
of eNO and NO2 in a cohort of 2,402 healthy school children. NO2 was estimated for
each child's residential postal code. Quantitative results were not reported; the study
authors only indicate that NO2 showed positive but statistically nonsignificant
associations with eNO. An eNO-roadway density association persisted after adjustment
for air pollutant concentrations (NO2, SO2, and PlVfc 5) within the previous 24 and
48 hours of the eNO measure, indicating that the association with roadway density was
unlikely to be confounded by an unmeasured short-term exposure effect.
The short-term evidence base provides support for the development of asthma in relation
to long-term NO2 exposure. Epidemiologic studies indicate associations between
short-term NO2 exposure and increases in oxidative stress and pulmonary inflammation
in the general population of children and in healthy adults (Section 5.2.7.4).
Toxicological Evidence
Similar to studies of short-term NO2 exposure (Section 5.2.7.4). some animal
toxicological studies of long-term exposure show increases in pulmonary inflammation,
oxidative stress, and injury. Details from these studies, all of which were reviewed in the
2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). are presented in Table 6-2.
Many studies investigating NO2-induced injury and oxidative stress in the airway
measured changes in lipids, which are necessary for both lung function and defense.
Sagaietal. (1982) reported that rats exposed to 40, 400, or 4,000 ppb NO2 for 9 or
18 months had increased ethane exhalation and that exposure to 40 ppb for 9 months
resulted in increased lipid peroxidation. Arner and Rhoades (1973) showed that rats
exposed to 2,900 ppb NO2 for 9 months had decreased lipid content leading to increased
surface tension of the lung surfactant and altered lung mechanics.
Histopathological assessment of lung tissue showed that long-term exposure to NO2
resulted in alveolar macrophage (AM) accumulation and areas of hyperinflation (Gregory
etal.. 1983). Kumae and Arakawa (2006) exposed rats to 200, 500, or 2,000 ppb NO2
prenatally (embryonic group) or postnatally during the weanling period (5 weeks old,
weanling group) and assayed bronchoalveolar lavage (BAL) fluid at 8 and 12 weeks of
age. In the embryonic group, exposure to 500 ppb NO2 resulted in increases in
lymphocytes at 8 weeks and increases in macrophages and neutrophils at 12 weeks. No
changes in differential cell counts were observed in the weanling group at 8 weeks of age,
6-25
-------
but at 12 weeks of age, macrophages and lymphocytes were increased with exposures at
and above 500 ppb and neutrophils were increased at 2,000 ppb. The embryonic group
also had increased tumor necrosis factor alpha (TNF-a) and interferon gamma (IFN^y) at
8 weeks but not at 12 weeks. In the weanling group, TNF-a was increased at both 8 and
12 weeks, and IFN-y was increased only at 12 weeks.
Long-term NCh exposure can modify oxidant balance in the airway. However, similar to
short-term exposures (Section 5.2.2.5). long-term ambient-relevant NC>2 exposures do not
consistently induce effects on antioxidant levels or enzyme activity across species.
Long-term NCh exposure has increased, decreased, and unaltered activity of enzymes
involved in the glutathione cycle (Sagai et al., 1984; Gregory et al., 1983; Ayaz and
Csallany. 1978). Sagai etal. (1984) reported increased nonprotein sulfhydryl levels and
decreased glutathione S-transferase (GST) activity in adult male rats after 9 and
18 months of exposure to 400 ppb NC>2 and decreased glutathione peroxidase (GPx)
activity and increased glucose-6-phosphate dehydrogenase activity after exposure to
4,000 ppb NO2. There were no changes in the activity of 6-phosphogluconate
dehydrogenase, superoxide dismutase (SOD), or disulfide reductase after exposure to
400 ppb NC>2. Gregory et al. (1983) reported increased GPx activity in BAL fluid after
6 weeks of exposure to 5,000 ppb NCh; however, at 15 weeks, enzyme activity returned
to control levels although slight changes in pathology were reported. Ayaz and Csallany
(1978) showed that continuous exposure to 1,000 ppb NC>2 for 17 months decreased GPx
activity in Vitamin E-deficient mice but increased GPx activity in
Vitamin E-supplemented mice.
6-26
-------
Table 6-2 Characteristics of animal toxicological studies of long-term nitrogen
dioxide exposure and respiratory effects.
Species (Strain); Sample
Study Size; Sex; Age
Arnerand Rats (Long-Evans);
Rhoades(1973) n = NR; M; age NR
Exposure Details
2,900 ppb for 5 days/week for 9 mo
Endpoints Examined
Histopathologic
evaluation and
morphometry
Ayaz and Mice (C57BL/6J); n = 120;
Csallany(1978) F; age NR
500 ppb or 1,000 ppb continuously for Morphometry
17 mo
Blair et al. Mice (strain NR);
(1969) n = 4/group; sex & age NR
500 ppb for 6, 18, or 24 h/day,
7 days/week for 3-12 mo
Histopathologic
evaluation
Chang et al.
(1986)
Rat (Fisher 344);
n = 8/group; M; 1-day or
6 weeks
(1) 500 ppb continuously with two, daily
1-h spikes of 1,500 ppb, 5 days/week for
6 weeks
(2) 2,000 ppb continuously for
7 days/week for 6 weeks; Two 1-h spikes
daily to 6,000 ppb (6-week rats only)
Histopathologic
evaluation and lung
morphometry
Crapo et al. Rat (CD, Fisher 344); n =
(1984) NR;M;6week
2,000 ppb for 23 h/day; two daily 30-min
spikes of 6,000 ppb
Morphometric analysis of
proximal alveolar and
distal alveolar regions
Ehrlich and
Henry (1968)
Mice (Swiss albino);
n = >30/group,
n = 4-8/group; F; age NR
(1) 500 ppb continuously
(2) 500 ppb for 6 h/day
(3) 500 ppb for 18 h/day
(1-3) for 1 to 12 mo; challenged with
Klebsiella pneumoniae after exposure
Mortality, hematology,
serum LDH, body weight,
bacterial clearance
Fujimaki and Rats (Wistar); n = 10/group; 1,000, 2,000, or 4,000 ppb continuously
Nohara (1994) M; 8 weeks for 12 weeks
Guinea pigs (Hartley);
n = 10/group; sex NR;
8 weeks
Furiosi et al. Monkey (Macaca
(1973) speciosa), n = 4-5/group;
M/F; maturing
Rat (Sprague-Dawley);
n = 15-25/group; M;
weanling
(1) 2,000 ppb NO2 continuously
(2) 330 ug/m3 NaCI continuously
(3) 2,000 ppb NO2 + 330 ug/m3 NaCI
continuously
(1-3) for 14 mo
Mast cell counts and
histamine release
Histopathologic
evaluation, hematology
Greene and Baboons; n = 6; M/F; 3 to 2,000 ppb 8 h/day, 5 days/week for 6 mo Immunologic and
Schneider 4 yr histopathologic
(1978) evaluation
6-27
-------
Table 6-2 (Continued): Characteristics of animal toxicological studies of
long-term nitrogen dioxide exposure and respiratory
effects.
Study
Species (Strain); Sample
Size; Sex; Age
Exposure Details
Endpoints Examined
Gregory et al.
(1983)
Rat (Fischer 344);
n = 4-6/group; sex NR;
14-16 weeks
(1)1,000ppb
(2)5,000ppb
Histopathological
evaluation, BAL fluid
(3) 1,000 ppb with two daily, 1.5-h spikes ana|ysis (LDH> ALKP>
of 5,000 ppb
(1-3) 7 h/day for 5 days/week for up to
15 weeks
glutathione peroxidase),
antioxidant enzymes in
lung homogenates
Hayashi et al.
(1987)
Henry et al.
(1970)
Kumae and
Arakawa (2006)
Kubota et al.
(1987)
Lafuma et al.
(1987)
Mercer et al.
(1995)
Miller et al.
(1987)
Saqai et al.
(1982)
Saaai et al.
(1984)
Sherwin and
Richters(1982)
Rat (Wistar);
n = 18-160/group; M; age
NR
Squirrel monkeys; n = 37;
M; age NR
Rats (Brown -Norway);
n = 5-47/group; F; age NR
Rat (JCL Wistar);
n = 3-4/group; M; 2 mo
Hamster (Golden Syrian);
n = 7-9/group; M; age NR
Rats (Fischer 344);
n = 5/group; M; 7 weeks
Mice(CD-l);
n = 18-217 group; F;
4-6 weeks
Rats (JCL, Wistar);
n = 6-12/group; M;
8 weeks
Rats (JCL Wistar);
n = 4-6/group; M; 8 weeks
Mice (Swiss Webster);
n = 30/group; M; young
500 ppb or 5,000 ppb continuously for up
to 1 9 mo
5,000 ppb continuously for 2 mo;
challenge with Klebsiella pneumoniae or
influenza after exposure
200, 500, or 2,000 ppb pre- and
post-natal for up to 12 postnatal weeks
40, 400, or 4,000 ppb continuously for 9,
18, and 27 mo
2,000 ppb NO2 for 8 h/day for
5 days/week for 2 mo
500 ppb continuously with 2 daily, 1-h
peaks of 1 ,500 ppb for 9 weeks
(1)200 ppb
(2) 200 ppb daily continuously for
7 days/week with 2 daily, 1-h peaks of
780 ppb 5 days/week
(1-2)16, 32, or 52 weeks
1 0,000 ppb continuously for 2 weeks
40, 400, or 4,000 ppb continuously for 9,
18, or 27 mo
340 ppb for 6 h/day for 5 days/week for
6 weeks
Morphological changes,
histology
Infection resistance,
mortality, peripheral
blood markers, and
respiratory function
Immunologic evaluation
(alveolar macrophage
activity) and BAL fluid
cell counts and cytokines
Serological examination
and lung morphometry
Lung histopathology and
morphometry, lung
mechanics, serum
elastase activity, and
protease inhibitor
capacity
Histopathologic
evaluation and
morphometry
Histopathologic
evaluation, pulmonary
function, and
antibacterial host
defenses
Antioxidant levels,
enzyme activity, lipid
peroxidation
Histopathologic
evaluation and
morphometry
Type II pneumocytes in
the lungs and alveolar
adults
wall area
6-28
-------
Table 6-2 (Continued): Characteristics of animal toxicological studies of
long-term nitrogen dioxide exposure and respiratory
effects.
Study
Stevens et al.
(1988)
Tepperet al.
(1993)
Waqner et al.
(1965)
Species (Strain); Sample
Size; Sex; Age
Rat (Fischer 344);
n = 1 ore/group; M;
neonate, young adult
Rats (Fischer 344);
n = 11-16/group; M;
60 days
Dog (mongrels);
n = 6-10/group; M; age NR
Exposure Details
500, 1,000, or 2, 000 ppb continuously
with two daily, 1-h spikes at 1,500, 3,000,
or 6,000 ppb for 5 days/week for 6 weeks
500 ppb continuously 7 days/week with
two daily, 2-h spikes of 1,500 ppb,
5 days/week for up to 78 weeks
1 ,000 or 5,000 ppb continuously for
10-18 months
Endpoints Examined
Pulmonary function
Pulmonary function and
lung disease
Pulmonary function and
histopathology
Rabbit; n = 4-8/group; M;
age NR
Guinea pig (English);
15-31/group; M; age NR
Rat (Sherman);
n = 20-40/group; M; age
NR
Mice (HLA, C57BI/6J,
CAF/Jax);
n = 60-110/group; M; age
NR
ALKP = alkaline phosphatase; BAL = bronchoalveolar lavage; F = female; LDH = lactate dehydrogenase; M = Male;
NaCI = sodium chloride; NO2 = nitrogen dioxide; NR = not reported.
6.2.2.4 Summary of Development of Asthma or Chronic Bronchitis
The recent evidence base adds several longitudinal studies that consistently find a
positive association of various NO2 exposure measures with asthma incidence in children
at several ages. Many studies observe associations with individual, residential ambient
NO2 exposure estimated by LUR that were demonstrated to well predict ambient NO2
concentrations in the study locations. Another study observed an association with NC>2
measured at sites 1-2 km from the subjects' school or home. Asthma incidence was also
associated with neighborhood-level ambient NO2 concentrations estimated by IDW or
measured at a single community central site monitor. For the latter, studies did not report
information on the representativeness of the neighborhood-level exposure estimates. In
adults, positive associations are also observed; however, this evidence base is limited
primarily to one adult cohort in Europe. None of the studies of children or adults
examined whether there was evidence for an association of NO2 with health effects
independent from PM2s or traffic-related pollutants such as EC/BC, CO, or UFP.
6-29
-------
Toxicological and controlled human exposure studies reduce some of the uncertainty in
the epidemiologic evidence by providing biological plausibility for a relationship
between long-term NO2 exposure and asthma development. In the pathophysiology of
asthma, recurrent pulmonary inflammation, allergic sensitization, and subsequent
development of AHR play important roles (Section 4.3.5 and Figure 4-2).
Long-term-exposure toxicological studies demonstrate NO2-induced AHR, and
experimental studies of repeated short-term exposures provide evidence for NCh-induced
development of allergic responses in healthy adults and animal models as well as
increases in neutrophils in healthy adults. In one study of guinea pigs, NCh-induced
(1,000-4,000 ppb) increases in AHR was accompanied by an increase in specific airways
resistance, suggesting that airway remodeling may contribute to the development of AHR
[(Kobayashi and Miura. 1995); Section 4.3.2.5]. Mechanistic studies indicate that
inflammatory mediators and structural changes occurring due to airway remodeling can
alter the contractility of airway smooth muscle. There also is some evidence for
pulmonary oxidative stress induced by short-term NCh exposure in healthy adults
(Section 5.2.7.4) and long-term exposure in rodents (Section 6.2.2.3). although results
overall are not consistent. Epidemiologic evidence points to associations between
short-term increases in ambient NC>2 concentrations and increases in pulmonary
inflammation in healthy children and adults (Section 5.2.7.4). but such evidence is
limited and inconsistent for long-term NC>2 exposure. The positive relationship between
NO2 exposures and asthma in longitudinal epidemiologic studies and the small body of
evidence indicating NO2 effects on inflammation, allergic sensitization, and AHR, which
are key events in the proposed mode of action for the development of asthma, indicate
that long-term NC>2 exposure could have an independent role in asthma development.
6.2.3 Severity of Asthma, Chronic Bronchitis, and Chronic Obstructive
Pulmonary Disease: Respiratory Symptoms and Hospital Admissions
In the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). there was inconsistent
evidence for an association between long-term exposure to NO2 and respiratory disease
severity assessed as occurrence of respiratory symptoms. There was a single longitudinal
study, and for the many cross-sectional studies, there was uncertainty related to the
temporality of exposure and occurrence of symptoms. There are several recent
longitudinal studies of respiratory disease severity, which are the focus of this evaluation.
Studies that evaluated indoor NO2 concentrations are discussed first, followed by studies
of outdoor NO2 concentrations. Cross-sectional studies were reviewed, and results
generally do not differ from those in longitudinal studies (Annesi-Maesano et al.. 2012a:
Ghosh etal.. 2012b; Dong etal.. 2011; Mi et al.. 2006; Pattenden et al.. 2006; Nicolai et
al.. 2003; Brauer et al.. 2002; Gehring etal.. 2002; Zempetal.. 1999). Previous
6-30
-------
cross-sectional studies are summarized in Annex Table AX6.3-17 of the 2008 ISA for
Oxides of Nitrogen (U.S. EPA. 2008a). A key consideration in evaluating respiratory
disease severity is whether the study design or statistical methods accounted for the
potential effects of short-term exposure.
6.2.3.1 Indoor Nitrogen Dioxide and Respiratory Symptoms in Children and Adults
Effects of indoor NC>2 may not be confounded by all of the same copollutants as outdoor
NO2, although there could be confounding by other indoor pollutants such as those
emitted from heating sources. Coherence between associations of respiratory disease
severity for indoor and ambient NC>2 exposure metrics can aid in drawing inferences
about the effects of ambient NO2 exposure. For long-term NO2 exposures, the recent
indoor prospective study of school-aged children (Belanger et al.. 2013) and the adult
indoor prospective study (Hansel et al., 2013) provide evidence that supports a
relationship between long-term NC>2 exposure and respiratory disease severity.
Belanger et al. (2013) observed positive associations of asthma severity score, wheeze,
nighttime symptoms, and rescue medication use with indoor residential NC>2 where the
mean monitoring length was 33 [standard deviation (SD): 7] days. Figure 6-3 illustrates
the concentration-response relationships between indoor NC>2 and asthma-related effects
with threshold functions for each outcome. In adjusted models with quintiles of NC>2
exposure, concentrations >14.3 ppb compared with the reference level (<6 ppb,
designated as the threshold value) were associated with elevated risk of a one-level
increase in asthma severity score (OR: 1.43 [95% CI: 1.08, 1.88]). These same exposures
were also associated with increased risks of wheeze (OR: 1.53 [95% CI: 1.16, 2.02]),
night symptoms (OR: 1.59 [95% CI: 1.24, 2.01]), and rescue medication use (OR: 1.74
[95% CI: 1.34, 2.26]). Every fivefold increase in NO2 exposure >6 ppb was associated
with increases in asthma-related outcomes (Table 6-3).
Recent studies of infants are consistent with previous results (Samet et al.. 1993) that
showed no association between 2-week avg exposure to NO2 and the incidence and
duration of respiratory illness. Raaschou-Nielsen et al. (201 Ob) and Sonnenschein-Van
der Voort et al. (2012) found no associations between indoor NO2 exposure and wheezing
in infants.
6-31
-------
A Asthma Severity Score
B wheeze
2.0-,
ro -.1
QC -1
•D
O
0.5.-'
0.3 J
6 ppb X
32
128
128
NO2 (ppb, log scale)
0.3
C Night Symptoms
NO2 (ppb, log scale)
D Rescue Medication Use
2.0
; -1.0-
-o
TJ
O
0.5
0.3
32
128
32
128
600 All subjects
400
200
o|
Gas stove users
NO2 (ppb, log scale)
8 32
NO, (ppb, log scale)
128
Note: ppb = parts per billion; NO2 = nitrogen dioxide. Solid lines = constrained, natural spline functions; small dashed lines = 95%
confidence intervals; bold dashed line = threshold function. Also shown is a histogram of NO2 concentrations measured in subjects'
homes (lower portion of panel D) for all observations (thin border) and observations taken in homes of gas stove users (bold
border). Indoor NO2 was modeled as a continuous variable of log-transformed concentrations.
Source: Reprinted with permission of Wolters Kluwer Health, Belanger et al. (2013).
Figure 6-3 Concentration-response relationships between asthma-related
effects and indoor nitrogen dioxide illustrated with constrained,
natural spline and threshold functions in hierarchical ordered
logistic regression models.
6-32
-------
Table 6-3 Longitudinal studies of long-term nitrogen dioxide exposure and respiratory symptoms in children.
Study3
Exposure
Assessment
Pollutant
Correlation
Statistical Methods
Comments
Results
(95% Cl)b
Study of Traffic, Air Quality, and Respiratory Health, Connecticut and Western Massachusetts
tBelanqeretal. (2013)
n = 1,642 children with asthma,
ages 5-10 yr.
Followed for 1 yr during
2006-2009.
Indoor residential.
Repeated
measurements with
Palmes tubes in
bedrooms and
dayroom for 4 weeks
for 4 seasons.
No other pollutants
examined.
Hierarchical ordered logistic
regression adjusted for age,
sex, atopy, season of
monitoring, race/ethnicity,
mother's education, smoking
in the home, and all five
variables for combined
specific sensitization and
exposure to indoor allergens.
Also adjusted for
maintenance medication use
because it was also
associated with SES.
Asthma severity score
consisted of
symptoms and
medication use based
on the Global
Initiative for Asthma
(NHLBI, 2002).
OR per fivefold increase in
NO2 exposure above
6 ppb
Asthma severity score:
1.37(1.01, 1.89)
Wheeze:
1.49(1.09,2.03)
Night symptoms:
1.52(1.16,2.00)
Rescue medication use:
1.78(1.33,2.38)
6-33
-------
Table 6-3 (Continued): Longitudinal studies of long-term nitrogen dioxide exposure and respiratory symptoms in
children.
Study3
Children's
Exposure
Assessment
Health Study (CHS), southern California
Pollutant
Correlation Statistical Methods Comments
Results
(95% Cl)b
McConnell et al. (2003)
n = 475 children with asthma or
bronchitic symptoms, ages
9-10yr(4th grade) and 12-13
yr (7th grade).
Followed 1996-1999.
Completed two or more yearly
follow-up questionnaires.
Central site—1 in
each of 12
communities.
Within community:
yearly deviation in
each community from
overall 4-yr mean.
Between
communities: 4-yr
average across
communities.
Mean (SD) 4-yr avg
across communities:
19.4(11.3)ppb.
Within-community
correlations differed
in that NO2 could be
distinguished from
most other major
pollutants except
OC and I ACID.
EC: 0.54
OC: 0.67
PlVh.s: 0.54
O3: 0.59
PMio: 0.20
PMio-2.s: -0.22
I ACID: 0.65
OACID: 0.48
Three-stage regression
adjusted for age, maternal
smoking history, child's sex,
maternal and child's race.
Within-community estimates
were adjusted for
between-community effects
of the pollutant and vice
versa.
Overall participation
rate was high (82%).
OR within community
1.97(1.22,3.18)
OR between communities
1.22(1.00, 1.49)
Copollutant model results:
Within community
NO2withEC: 1.05C
NO2withOC: 1.04NS
NO2with PIvh.s: 1.05NS
NO2withO3: 1.06NS
NO2withPMio: 1.07C
NO2with PMio-2.s: 1.08d
NO2with I ACID: 1.09d
NO2 with OACID: 1.07C
Between communities
NO2withEC: 1.01 NS
NO2withOC: 1.01 NS
NO2 with PlVh.s: 1.01 NS
NO2withO3: 1.02C
NO2with PMio: 1.01 NS
NO2with PMio-2s.: 1.02C
NO2with I ACID: 1.02NS
NO2 with OACID: 1.02NS
6-34
-------
Table 6-3 (Continued): Longitudinal studies of long-term nitrogen dioxide exposure and respiratory symptoms in
children.
Study3
Exposure
Assessment
Pollutant
Correlation
Statistical Methods
Comments
Results
(95% Cl)b
Prevention and Incidence of Asthma and Mite Allergy (PIAMA), the Netherlands
tGehrinq et al. (2010)
n = 3,863 children followed from
birth to age 8 yr
LUR model
Annual avg at birth
residence.
NO2-PM2.5: 0.93
NO2-soot: 0.96
GEE adjusted for sex, study
arm (intervention or natural
history), use of
mite-impermeable mattress
covers, allergies of mother
and father, maternal and
paternal education, maternal
prenatal smoking,
breastfeeding, presence of a
gas stove in the child's home,
presence of older siblings,
smoking, signs of dampness
and pets in the child's home,
day care attendance, and
Dutch nationality.
NO2 not associated
with atopic eczema,
allergic sensitization,
or bronchial
hyperresponsiveness.
No copollutant
models.
OR for asthma symptoms:
1.17(0.98, 1.39) without
adjustment for study
region.
OR for wheeze:
1.27(1.07, 1.50) without
adjustment for study
region.
Children, Allergy, Milieu, Stockholm, Epidemiology Survey (BAMSE), Stockholm, Sweden
tGruzieva et al. (2013)
n = 3,633 children followed from
birth (1994-1996) to age 12 yr.
Related publications:
Melen et al. (2008)
Nordlinq etal. (2008)
Dispersion model
Annual avg NOx at all
residences from birth
to age 12 yr
(1994-2008).
NOx-PMio for first
yr of life: 0.96
Multinomial regression/GEE
adjusted for municipality,
SES, yrthe house was built,
and heredity.
OR for wheeze at 12 yr of
age, 3 or more episodes:
1.35(0.79, 2.29) per
20 ppb NOx.
No association of NOx
after the first yr of life with
asthma symptoms.
BAMSE = Children, Allergy, Milieu, Stockholm, Epidemiology Survey; CHS = Children's Health Study; Cl = confidence interval; EC = elemental carbon; GEE = generalized estimating
equations; I ACID = inorganic acid; LUR = land use regression; NO2 = nitrogen dioxide; NOX = sum of NO and NO2; NR = not reported; NS = not statistically significant; O
ACID = organic acid; O3 = Ozone; OC = organic carbon; OR = odds ratio; PIAMA = Prevention and Incidence of Asthma and Mite Allergy; PM2.s = particulate matter with a nominal
mean aerodynamic diameter less than or equal to 2.5 |jm; PM10 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; PM10-2.5 = particulate
matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm and greater than a nominal mean of 2.5 |jm; SD = standard deviation; SES = socioeconomic status.
aStudies are presented in the order of appearance in the text.
bResults are presented per 10-ppb increase in NO2 unless otherwise specified.
°p<0.01.
dp < 0.05.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
6-35
-------
Hansel etal. (2013) investigated indoor NCh and PIVb 5 concentrations in relation to
respiratory effects among former smokers with COPD in Baltimore, MD. Pollutants and
symptoms were measured at baseline, 3 months, and 6 months. Pollutants were measured
in the bedroom and the main living area, rooms where participants reported spending the
most time. No interaction was indicated between PIVb 5 and NO2, and increasing NC>2
concentrations in the main living area were independently associated with increased
dyspnea and increased rescue medication use with adjustment for PM25 Higher bedroom
NO2 concentrations were associated with increased risk of nocturnal awakenings (OR:
1.59 [95% CI: 1.05, 2.42] per 10-ppb increase) and severe exacerbations (OR: 1.65 [95%
CI: 1.02, 2.64]). NO2 concentrations were not associated with lung function. There was
indication of outdoor NO2 concentrations contributing to indoor NO2. Among the
26 subjects who lived within 4.8 km of a central monitoring site, outdoor NO2
concentrations explained 25% of the variance in indoor NO2 concentrations.
6.2.3.2 Outdoor Nitrogen Dioxide and Respiratory Symptoms in Children
A number of studies (Table 6-3) observed an association between various respiratory
symptoms in children and long-term exposure to outdoor NO2. McConnell et al. (2003)
examined children with asthma for bronchitic symptoms, including daily cough for
3 months in a row, congestion or phlegm 3 months in a row, or bronchitis. Thus, while
these symptoms may have started with acute exacerbation of asthma, they were likely to
represent chronic indolent symptoms. In copollutant models, the effects of yearly
variation in NO2, ascertained from a central monitoring site in each community, were
only modestly reduced by adjusting for either PM2 5 or a traffic-related copollutant such
as EC or organic carbon (OC; Figure 6-4 and Table 6-3).
Gehring etal. (2010) examined a composite of asthma symptoms (one or more attacks of
wheeze, shortness of breath, prescription of inhalation steroids) and wheeze (transient,
late onset, persistent) and observed positive associations with LUR modeled NO2
exposures. Gruzieva et al. (2013) examined wheeze, categorized as either one or more
episodes or three or more episodes in the past year and observed an association with NOx
concentrations from a dispersion model.
Hwang and Lee (2010) provide information on the potential for copollutant confounding
of NO2 associations with respiratory symptoms but based on a cross-sectional analysis.
The associations between NO2 and respiratory symptoms in children with asthma did not
appreciably change in copollutant models with PM2 5 (ORs for NO2 adjusted for PM2 5:
2.25 [95% CI: 1.17, 4.33] for bronchitis; 1.60[95%CI: 0.76, 3.34] for chronic phlegm;
6-36
-------
1.30 [95% CI: 0.53, 3.12] for chronic cough; 2.21 [95% CI: 1.23, 3.97] for bronchitic
symptoms per 10-ppb increase in
Risk of Bronchitic Symptoms as a Function
of Yearly Deviation in NO2
1.0 -
Q.
^- 1.3 -
*& 1?
| •-
'•'
T3
o 1.0 -
n Q -
LJL_LiJ_J
I
I I
_L
-•£>
0°
Adjustment Air Pollutants
Notes: EC = elemental carbon; I ACID = inorganic acid; NO2 = nitrogen dioxide; O3 = ozone; O ACID = organic acid; OC = organic
carbon; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PMio = particulate
matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; PMio-2.s = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 10 |jm and greater than a nominal mean of 2.5 |jm; ppb = parts per billion.
Source: Reprinted with permission of the American Thoracic Society, McConnell et al. (2003).
Figure 6-4 Within-community odds ratios for bronchitis symptoms
associated with nitrogen dioxide adjusted for a copollutant in the
12 communities of the Children's Health Study.
Outdoor Nitrogen Dioxide and Respiratory Symptoms in Adults with
Asthma
The relationship between long-term NO2 exposure and respiratory symptoms in adults
was examined in many prospective studies of asthma incidence in adults discussed in
Section 6.2.2.2. Jacquemin et al. (2009b) reported that associations between NO2 and all
of the examined asthma symptoms at ECHRS II were positive. The strongest was for
waking with a feeling of chest tightness in the last 12 months. Symptoms in the last
12 months at ECRHS II among people without asthma at baseline were also associated
with NO2. NO2 exposures estimated from dispersion models were demonstrated to be
moderately correlated with measured concentrations in the study locations (R2 = 0.60).
6-37
-------
In the cross-sectional Swiss Study on Air Pollution and Lung Disease in Adults
(SAPALDIA), Zemp etal. (1999) report an association between NO2 and prevalence of
respiratory symptoms in adults. Bentayeb et al. (2010) reported cross-sectional
associations to be weakly positive for cough and phlegm in adults (>65 years old, in
Bordeaux, France) in relation to NO2 exposure.
Outdoor Nitrogen Dioxide and Asthma Hospital Admissions in Adults
Recent studies represent the first evaluation of the association between long-term NO2
exposure and hospital admissions for asthma. The Danish Diet, Cancer and Health cohort
examined asthma hospital admissions [International Classification of Diseases (ICD)-IO:
J45-46] in adults ages 50-65 years at baseline (Andersen et al.. 2012a). Associations
between NO2 concentration estimated by the Danish Air geographic information system
(GIS) dispersion modelling system and hospital admission were found in the full cohort
[hazard ratio (HR) per-10 ppb NO2: 1.44 [95% CI: 1.14, 1.84]). NO2 was estimated to
have a similar effect on the first asthma hospital admission (HR: 1.36 [95% CI: 1.03,
1.80]), but people with a previous asthma hospital admission were at greater risk for
re-admission (HR: 3.05 [95% CI: 1.57, 5.90]). NO2 was associated with a much larger
risk of asthma hospital admission among people with previous admission for COPD (HR:
2.34 [95% CI: 1.25, 4.40]). Some of the observed effects could possibly be ascribed to
the short-term effects of increases in air pollution on the days prior to asthma admission.
The 3 5-year avg, 15-year avg, and 1-year avg NO2 preceding admission were highly
correlated (r = 0.88, 0.92) and were more strongly associated with asthma hospital
admission than was 1-year avg NO2 at baseline. The authors indicated that they could not
discern whether the results reflected the importance of more recent exposures or the
better performance of the dispersion model in more recent years. The NO2 exposure
estimates may be less certain for earlier time periods because of uncertainty in emission
factors and traffic counts that are used as inputs to the dispersion model.
An ecological time-series study (Delamater et al.. 2012) and a cross-sectional study
(Meng etal.. 2010) provide inconsistent results with regard to asthma-related emergency
department (ED) visits or hospital admissions. A time-series study observed that monthly
NO2 concentrations were associated with monthly asthma hospital admission rates in
Los Angeles, CA, but NO2 concentrations were averaged over the county (Delamater et
al.. 2012). A cross-sectional study in the San Joaquin Valley, CA examined more
spatially resolved estimates of NO2 exposure for subjects, ages 1 to 65+ years, who
reported physician-diagnosed asthma. Annual average NO2 concentrations were assigned
to subjects from the closest air monitoring station within 8 km of the residential ZIP code,
but data on duration of residence were not available. No quantitative results were shown
6-38
-------
for NC>2, but NC>2 was reported not to be associated with asthma-related ED visits or
hospital admissions (Meng et al.. 2010).
6.2.3.3 Summary of Severity of Asthma, Chronic Bronchitis, and Chronic Obstructive
Pulmonary Disease
Longitudinal studies observe associations between long-term ambient NO2 exposure
metrics and an array of respiratory symptoms in school-age children. Results in infants
are inconsistent. Transient symptoms are common in infants; thus, symptoms in infants
may not have strong implications for development of respiratory disease. For children,
ambient NCh exposure was assessed using outdoor residential measurements, an LUR
that estimated exposure at subjects' homes, and central site measurements. The
McConnell et al. (2003) study is unique in that it is the only prospective study examining
bronchitic symptoms in children with asthma. The study authors report stronger
associations for NCh variation within communities (within-community associations are
less prone to confounding by time-fixed personal covariates) than for NO2 variation
between communities.
Further supporting a relationship with NC>2, indoor NC>2 was associated with asthma
symptoms and medication use in children with asthma (Belanger et al.. 2013) and
respiratory symptoms in former smokers with COPD (Hansel et al.. 2013). The effect
estimates for indoor NC>2 were generally larger than those reported in the studies of
outdoor NC>2, and Belanger et al. (2013) provided evidence for a concentration-dependent
increase in NC^-related symptoms. These indoor NCh exposures may be part of a
different mix of air pollutants than is NC>2 in the ambient air and support an independent
effect of NO2.
An uncertainty in the evidence base is the potential influence of short-term NC>2
exposure. While many studies aimed to characterize chronic symptoms, they did not
examine whether associations were independent of short-term NO2 exposure. Another
uncertainty is the potential confounding by other traffic-related pollutants. Long-term
averages of NO2 showed a correlation of 0.54 for ambient EC and 0.96 for soot; no data
are available for CO (Table 6-3). For PIVb 5, the correlations are 0.54 for central site
measurements and 0.93 (Gehring et al.. 2010; McConnell et al.. 2003). In limited
analysis, NC>2 associations with symptoms persisted with adjustment for EC or PM2 5 as
measured at central sites and was somewhat attenuated with adjustment for OC. The
weaker NCh-copollutant correlations and copollutant model results are based on central
site measurements and could reflect differential exposure measurement error. The
collective evidence from this group of prospective studies is supportive of a relationship
of long-term exposure to NC>2 and increased respiratory symptoms using various
6-39
-------
indicators in children with asthma, but evidence identifying an independent association of
long-term NC>2 exposure is limited.
6.2.4 Development of Allergic Disease
6.2.4.1 Epidemiologic Studies of Children or Adults
Recent cross-sectional studies report results for various aspects of allergic responses in
relation to long-term exposure to NC>2. Allergic sensitization indicators included
measures of IgE, allergic rhinitis, skin prick test, and reporting of respiratory allergy/hay
fever. Various age groups were examined, including children less than 6 years old,
children aged about 10 years, and adults. As described in Section 6.2.2.3. the few
available experimental studies support an effect of long-term or repeated short-term NO2
exposure on development of allergic responses.
In a nationally representative sample of the U.S. population, Weiretal. (2013) linked
annual average concentrations of NO2 to allergen-specific IgE data for participants aged
6 years and older in the 2005-2006 National Health and Nutrition Examination Survey
using both monitor-based (within 32.2 km) air pollution estimates and the Community
Multiscale Air Quality model (36 km) and observed that increased concentrations of NO2
were associated with positive IgE to any allergen.
In the German Infant Nutritional Intervention (GINI) and Life style-Related Factors on
the Immune System and the Development of Allergies in Childhood (LISA) cohorts,
analysis of individual-based exposure to NCh derived from LUR and allergic disease
outcomes during the first 6 years of life (Morgenstern et al.. 2008) indicated associations
with eczema. Some associations with allergen-specific IgE and hay fever were positive
but imprecise with wide 95% CIs. Previous analyses of these cohorts did not indicate
associations with runny nose and sneezing at age 2 years (Morgenstern et al.. 2007;
Gehring et al.. 2002). A longitudinal study of the LISA and GINI cohorts (Fuertes et al..
2013) found no evidence that NO2 exposure in the birth year increases the prevalence of
allergic rhinitis or increases risk of aeroallergen sensitization as determined by
allergen-specific IgE in children examined at age 10 years. Air pollution concentrations
decreased in the study areas during this time.
Annesi-Maesano et al. (2007) related lifetime prevalence of allergic conditions in
5,338 schoolchildren (ages 10.4 ± 0.7 years) attending 108 randomly chosen schools in
six French cities to NC>2 concentrations in school yards and at central site monitors. The
6-40
-------
authors used a 5-day avg NO2 concentration to represent long-term exposure to NC>2. NCh
was positively associated with flexural dermatitis and skin prick test to indoor allergens
but not with allergic rhinitis or atopic dermatitis. In the same study population, Annesi-
Maesano et al. (2012a) also evaluated a 5-day mean concentration for indoor classroom
NO2, categorized into tertiles independent of the city (low <9.7 ppb, medium >9.7 to
<12.9 ppb, high >12.9 ppb NC>2). Between-school and within-school variability of the
measured indoor pollutants were estimated using linear mixed models for longitudinal
data. Among children with atopy (n = 1,719), high NCh was related to previous-year
allergic asthma but not allergic rhinitis.
Similar to NC>2, NOx is inconsistently associated with development of allergic responses.
Nordling et al. (2008) reported that exposure to dispersion-modeled NOx from traffic
during the first year of life was associated with sensitization (measured as specific IgE) to
inhalant allergens, especially pollen (OR: 1.24 [95% CI: 1.04, 1.49] per 10-ppb increase
in NO2). In the Children, Allergy, Milieu, Stockholm, Epidemiology Survey (BAMSE)
cohort, NOx assessed from a dispersion model was not associated with risk of allergic
sensitization in children at age 4 years (Gruzieva et al.. 2012).
Cross-sectional studies also do not consistently link long-term NO2 exposure to allergic
conditions in children or adults. Among 30,139 Chinese children aged 3 to 12 years,
Dong et al. (2011) observed positive associations between 3-year avg of NO2 and allergic
rhinitis in the 26,004 children without allergic predisposition (n = 26,004), mainly among
males. Among children with an allergic predisposition, associations were detected in
males and females. In school children in Taiwan, Hwang et al. (2006) observed that a
10-ppb increase in NO2 was associated with a higher prevalence of allergic rhinitis, with
an OR of 1.11 (95% CI: 1.08, 1.15). Parker etal. (2009) evaluated respiratory allergy/hay
fever in the 1999-2005 U.S. National Health Interview Survey of approximately
70,000 children and observed no associations with NO2. In 2,644 adults aged 18-70 years
living in Nottingham, U.K., Pujades-Rodriguez et al. (2009) found generally null
associations between NO2 concentration and skin test positivity, total IgE, and
questionnaire-reported eczema or hay fever. Total IgE levels were not related to NO2
concentrations in 369 adults with asthma in five French centers as part of the
Epidemiological Study on the Genetics and Environment of Asthma (Rage et al.. 2009)
but were related to Os concentrations.
6.2.4.2 Summary of Development of Allergic Responses
A few available experimental studies demonstrate the effects of repeated short- or
long-term NO2 exposure on development of an allergic phenotype in healthy adults and
6-41
-------
animal models (Section 6.2.2.3). These findings not only suggest the possibility that
recurrent or chronic exposure to NC>2 may lead to the development of asthma but also
support a role for NC>2 exposure in the development of allergic conditions. Long-term
NC>2 exposure has been linked to indicators of allergic sensitization such as
allergen-specific IgE or skin prick test in a few cross-sectional studies of children around
age 10 years but not the longitudinal study. Results also are inconsistent for outcomes
such as allergic rhinitis or hay fever. In children 6 years and younger and in adults,
indicators of allergic responses are not related to NO2 exposure. NO2 metrics aimed at
characterizing individual exposures, such as 5-day measurements in school yards and
residential estimates from LUR, produced inconsistent results. Thus, the evidence base
for a relationship between long-term NCh exposure and allergic responses is limited.
6.2.5 Lung Function and Lung Development
The 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) characterized longitudinal
studies as showing a relationship between NO2 concentrations and decrements in lung
function and lung development in children. A key uncertainty in these studies was the
high correlation of NC>2 concentrations with other traffic-related pollutants and the
potential for confounding. Recent longitudinal epidemiologic studies add to the evidence
base linking long-term NC>2 exposure to decrements in lung function and lung
development as assessed by supervised spirometry. As with severity of respiratory
disease, a key consideration in evaluating the evidence for relationships of lung function
and lung development with long-term exposure is whether studies accounted for the
effects of short-term exposure.
6.2.5.1 Lung Function and Development in Children
The key longitudinal epidemiologic studies continue to show associations between
long-term NC>2 exposure and decrements in lung function, especially as children reach
adolescence (Figure 6-5 and Table 6-4). Lung function continues to increase through
early adulthood with growth and development, then declines with aging (Stanojevic et al.
2008; Zeman and Bennett. 2006; Thurlbeck. 1982). Thus, the relationship between
long-term NC>2 exposure and decreased lung function over time in school-age children
into early adulthood is an indicator of decreased lung development.
6-42
-------
Study Period of Exposure
Gauderman et al. (2004) 8-yr avg
Years
Enrollment Age Followed Lung Function Metric Exposure Assessment
Mean 10 yr
Breton etal. (2011)
Urmanetal. (2014)
Schultz etal. (2012)
Annual avg before testing Mean 10 yr
annual avg
Birth yr-all subjects
Birth yr-with atopy
Molter et al. (2013) 4- or 6-yr avg before testing
Rojas-Martinez et al. (2007a) 6-mo avg before testing
5 to 7
Syr
Syr
Syr
FEV1 (mL)
FVC (mL)
MMEF (mL/sec)
FEV1 (mL)
FVC (mL)
MMEF (mL/sec)
FEV1 (% change)
FVC (% change)
FEV1 (mL)
Central site-1 per community
Central site-1 per community
LUR
Dispersion model
Oftedal et al. (2008)
Birth yr
9-1 Oyr
FEV1 (% pred)
FEV1 (mL)-girls
FEV1 (mL)-boys
FVC (mL)-girls
FVC (mL)-boys
FEF25-75 (mL/sec)-girls
FEF25-75 (mL/sec)-boys
FEV1 (mL)
FVC (mL)
PEF (mL/sec)
FEF50% (mL/sec)
Microenvironmental
Central site-nearest school
Dispersion model
-120 -100 -80 -60 -40 -20 0 20 40
Change in Lung Function Metric (95% Cl)
Note: avg = average; Cl = confidence interval; FEF25-75 = forced expiratory flow between 25 and 75% of forced vital capacity; FEVi = forced expiratory volumne in 1 second;
FVC = forced vital capacity; LUR = land use regression; MMEF = maximum (or maximal) midexpiratory flow; mL = milliliters; mL/sec = milliliters per second; mo = month; PEF = peak
expiratory flow; yr = year(s). Black = study from the 2008 Integrated Science Assessment for Oxides of Nitrogen; red = recent studies. Circles = nitrogen dioxide (NO2);
Diamonds = sum of NO2 and nitric oxide (NOX). Effect estimates are standardized to a 10-ppb increase in NO2 and a 20-ppb increase in NOX concentration. Effect estimates from
Schultz et al. (2012) are not standardized because NOX is examined in |jg/m3.
Figure 6-5 Associations of oxides of nitrogen with lung function or lung development from longitudinal
studies of children.
6-43
-------
Table 6-4 Longitudinal studies of long-term nitogen dioxide exposure and lung function and lung development
in children.
Study3
Exposure Assessment Pollutant Correlation Statistical Methods
Comments
Results (95% Cl)b
Children's Health Study (CHS), southern California
Gauderman et al.
(2004)
n = 1,759 children
followed ages
10-1 Syr).
Follow-up began in
1993.
Central site monitors, 1 in
each of 12 communities.
8-yr avg.
Monitoring conducted
1994-2000.
Range in mean NO2
across communities:
34.6 ppb.
NO2-EC: 0.94
NO2-acid vapor: 0.87
NCb-PIVh.s: 0.79
NO2-PMio: 0.67
NO2-OC: 0.64
Two-stage linear
regression adjusted for log
height, BMI, BMI squared,
race, Hispanic ethnicity,
doctor-diagnosed asthma,
any tobacco smoking by
the child in the preceding
yr, exposure to
environmental tobacco
smoke, and exercise; or
respiratory tract illness on
the day of the test and
indicator variables for the
field technician and the
spirometer.
Follow-up participation:
n = 1,414 in 1995, 1,252 in
1997, 1,031 in 1999, and 747
in 2001 (10% loss peryr).
Children who moved away
from recruitment community
were classified as lost to
follow-up.
Model fit no better in
copollutantthan in
single-pollutant models. No
quantitative data shown.
Adjustment for 3-day avg NO2
before each lung function test
did not alter association for
long-term NO2.
Change over 8-yr
period:
FVC: -27.5
(-54.7, -0.2) ml_
FEVi: -29.3
(-47.5, -11.1) ml_
MMEF: -61.0
(-109.1, -12.8)
mL/sec
Gauderman et al.
(2007)
n = 3,677 children,
mean age 10 yr (SD
0.44)
12 CHS communities
Central site monitors, 1 in
each of 12 communities.
8-yr avg.
Monitoring conducted
1994-2000.
Range in mean NO2
across communities:
34.6 ppb.
NR
Regression models
adjusted for local traffic
exposure, height, height
squared, BMI, BMI
squared, present asthma
status, exercise or
respiratory illness on the
day of the test, any
tobacco smoking by the
child in the previous yr,
and indicator variables for
field technician.
NO2 and distance to freeways
were independently
associated with decrements
in lung development.
Compared with living >1,500
away from a freeway, living
within 500 m of a freeway
was associated with a mean
percent-predicted FEVi of
97.0% (94.6, 99.4) and
MMEF of 93.4% (89.1, 97.7).
Change in FEVi over
8-yr period:
-32 mL/sec
95% Cl NR.
6-44
-------
Table 6-4 (Continued): Longitudinal studies of long-term nitrogen dioxide exposure and lung function and lung
development in children.
Study3
Exposure Assessment Pollutant Correlation Statistical Methods
Comments
Results (95% Cl)b
tBretonetal. (2011)
1993 cohort 1:
n = 1,759
1996 cohort 2:
n= 2,004
Both cohorts followed
from age 10 to 18 yr.
Central site monitors, 1 in
each of 12 communities.
Annual avg.
Monitoring conducted
1994-2000.
Range in mean NO2
across communities:
34.6 ppb.
NO2-PM2.5 = 0.79
NO2-O3 = -0.11
Hierarchical mixed effects
models adjusted for
height, height squared,
BMI, BMI squared, current
asthma status, exercise or
respiratory illness on the
day of the test, any
tobacco smoking by the
child in the last yr,
glutathione S-transferase
mu 1 genotype, and
indicator variables for the
field technician.
Main purpose was to
determine whether sequence
variation in genes in the
glutathione synthesis
pathway alters susceptibility
to air pollution effects on lung
function.
Haplotype "0100000" was
associated with a 39.6-mL,
29.1-mL, and 51.0-mL/sec
reduction in FEV-i, FVC, and
MMEF, respectively, over 8-yr
follow-up.
FEVi: -29.8
(-50.0, -9.7) ml_
FVC: -29.8
(-54.7, -5.0) ml_
MMEF: -54.4
(-90.8,-18.0) mL/sec
tUrmanetal. (2014)
n = 1,811 children,
mean age 11 yr.
82% of the active 2002
cohort from 8 CHS
communities.
LUR model for near-
roadway NO2and NOx.
Central site monitor for
NO2.
LUR: annual avg at
residence.
Central site: 6-yr avg
LUR model developed
from 900 monitoring sites
in CHS communities
(Franklin etal.. 2012).
10-fold cross-validation R2
for NO2 = 69% in basin
(higher pollution) and 72%
out of basin (lower
pollution)
Quantitative data on mean
NO2 or NOx NR.
Central site
concentrations:
NO2-PM2.5: 0.60
NO2-PMio: 0.06
Near-roadway NO,
NO2, and NOx (within
communities): >0.90
Linear regression models
for near-roadway NO2 and
NOx with fixed effects for
study community.
Mixed model for central
site pollutants with
random intercept for
community and adjusted
for near-roadway NO2,
NO, or NOx.
Adjusted for log height,
height squared, BMI, BMI
squared, sex, age,
sex x age interaction,
race, Hispanic ethnicity,
respiratory illness at time
of test, and field
technician and study
community.
NO2 was associated with
FEVi but not FVC.
Lung function deficits of
2-3% were associated with
PM2.s, PM-io, and Os.
Associations with central site
NO2 and near-roadway NOx
were independent in
copollutant models.
Associations for
near-roadway NOx were not
modified by central site
pollutant concentrations.
Near-roadway NOx:
FVC: -1.7%
(-2.9, -0.55)
FEVi: -1.2%
(-2.4, -0.01)
Associations observed
in all communities and
also for NO2 and NO.
Residential proximity
to a freeway was
associated with a
reduction in FVC.
6-45
-------
Table 6-4 (Continued): Longitudinal studies of long-term nitrogen dioxide exposure and lung function and lung
development in children.
Study3
Exposure Assessment Pollutant Correlation Statistical Methods
Comments
Results (95% Cl)b
Children, Allergy, Milieu, Stockholm, Epidemiology Survey (BAMSE), Stockholm, Sweden
tSchultz et al. (2012)
n = 1,924 children
followed from birth to
age 8 yr.
Related Publications:
Nordlinqetal. (2008)
Dispersion model NOx
Time-weighted average
exposures for various time
windows estimated based
on lifetime residential, day
care, and school
addresses.
Short-term average
concentrations assessed
from central sites.
NR
Linear regression adjusted
for sex, age, height,
municipality, and heredity
for asthma and/or allergy.
Additional adjustments for
temperature, relative
humidity, Os, and PM-io
levels during 3-7 days
before each child's
pulmonary function test
showed little effect on the
estimates for long-term
avg NOx.
Specific adjustment for
short-term NOx was not
discussed.
OR for NOx in first yr of life
FEVi 80% of predicted:
2.1 (0.6, 8.1)
FEVi 85% of predicted:
3.4(1.6, 7.4)
FEVi per 47 ug/m3
increase in NOx in first
yr of life:
-34.9
(-80.1, 10.4)mL
Group sensitized
against any common
inhalant or food
allergens, and those
with asthma at 8 yr:
-98.9 (-169, -28.4)
ml_
No clear association
with NOx exposure
after infancy.
Prevention and Incidence of Asthma and Mite Allergy (PIAMA), the Netherlands
tEenhuizen et al.
(2013)
n = 880 children, age
4 yr
LUR model
Annual avg NO2 at birth
residence.
Mean: 10.4 ppb.
Daily average NO2
concentrations on the day
Rint measurement
obtained from central
sites.
For birth address:
NO2-PM2.5: 0.93
NO2-soot: 0.96
NO2 on test day and
long-term NO2: 0.55
NO2 on the day before
the test and long-term
NO2: 0.57
Multiple linear regression
adjusted for sex, age at
examination (days),
height, weight, maternal
prenatal smoking, any
smoking in the child's
home, use of gas for
cooking, parental allergy,
dampness in the home,
education of the parents,
season, temperature, and
humidity on the day of the
Rint measurement.
First report of an association
in 4-yr old children. Rint at
age 4 yr predicted asthma
and wheeze at age 8 yr.
Long-term average Plvh.s and
soot associated with Rint. A
monotonic increase of Rint
with increasing NO2
concentration was seen, with
no threshold identified.
Change in Rint:
0.05(0.001, 0.11)
kPaxs/L
6-46
-------
Table 6-4 (Continued): Longitudinal studies of long-term nitrogen dioxide exposure and lung function and lung
development in children.
Study3
Exposure Assessment Pollutant Correlation
Statistical Methods Comments
Results (95% Cl)b
Manchester Asthma and Allergy Study (MAAS), Manchester, U.K.
tMolter etal. (2013)
tMolter etal. (201 Oa)
tMolter et al. (201 Ob)
tMolter et al. (2012)
n = 1,185.
Followed from birth
(1995-1 997) to age
11 yr.
Personal NCb-PM-io
4-or6-yravg Pearson r= 0.59 to
Estimated from 0.89.
Microenvironmental
Exposure Model that
incorporated children's
time-activity patterns and
LUR modeled
concentrations.
The modeled estimates
agreed well with
measured NO2
concentrations.
Mean (SD) NO2:
14.76 (6.6) ppb
GEE adjusted for age,
atopy, asthma or wheeze,
presence of a gas cooker
in the home, hospital
admission for infection in
first two yr or life, pubertal
stage of development,
3-day avg PM-io.
Also considered sex, BMI,
ethnicity, parental atopy,
parental smoking, day
care attendance during
first 2 yr of life, family
history of asthma, dog or
cat in the home, visible
signs of dampness or
mold in the home,
duration of breast feeding,
and paternal income.
Change in percent
predicted FEV-i during
age 5-11 yr:
-15.6 (-26.1, -5.3)
Based on the average
predicted FEV-i in
cohort of 1.65 L,
change equals total
decrease in FEV-i of
263 ml_.
Change in percent
predicted
post-bronchodilator
FEVi during age
5-11 yr'
-19.8 (-32.5, -7.1).
Equivalent to total
413 ml_ decrease.
Mexico City, Mexico
Roias-Martinez et al.
(2007a) Roias-Martinez
et al. (2007b)
Central site NO2 24-h avg NO2 and
6-mo avg 8-h av9 °3'- 0.166
Closest site to school. 24-h avg NO2 and
General linear mixed NO2, Os, and PM-io were
models adjusted for age, associated with decrements
BMI, height, height by in lung development after
Girls
FVC- -40
(-46, -34) ml_
n = 3,170 children
followed ages 8-11 yr
(1996-1999).
Recruited from 31
schools.
Within 2-1 Okm.
Mean (SD) NO2 across
communities:
27.2 (10.9) to 42.6 (13.2)
24-h avg PMm 0.250
age, weekday time spent
in outdoor activities, and
environmental tobacco
smoke.
adjusting for short-term
averages (day before lung
function measurement) for
the pollutants.
FEVi: -27
(-33, -22) ml_
FEF25-75%:
7 (-8, 18)mL/sec
Boys
FVC: -38
(-44, -31)mL
FEVi: -22
(-28, -16) mL
FEF25-75%:
3 (-10, 16)mL/sec
6-47
-------
Table 6-4 (Continued): Longitudinal studies of long-term nitrogen dioxide exposure and lung function and lung
development in children.
Study3
Exposure Assessment Pollutant Correlation Statistical Methods
Comments
Results (95% Cl)b
Oslo, Norway
Oftedal et al. (2008)
n =2,307 children
followed from birth
(1992-1993) to age
9-10 yr.
Related Publications:
SI0rdal et al. (2003)
Dispersion model NO2
Annual avg at residence in
first yr of life and lifetime
avg.
Modeled estimates well
correlated with
measurements from
10 central site monitors.
r=0.76.
Mean NCb:
16 ppb in the first yr of life
11.86 ppb for lifetime avg
NO2 and PM2.5 or PM-io
r= 0.83-0.95
Multiple linear regression
adjusted for sex, height,
age, BMI, birth weight,
temperature lagged
1-3 days before the lung
function test, current
asthma, indicator for
participation in the Oslo
Birth Cohort study,
maternal smoking in early
lifetime, parental ethnicity,
education, and smoking.
NO2 associations stronger in
girls.
In models that included both
short- and long-term NO2
exposures, only the
association with long-term
NO2 remained.
NO2 in first yr of life
among all children:
FEVi: -6.0
(-18.0, 6.2) ml_
FVC: -1.4
(-14.6, 11.8)mL
PEF: -57.9
(-92.5, -22.3) mL/sec
FEFso%: -37.3
(-71.2, -3.5) mL/sec
Avg = average, BAMSE = Children, Allergy, Milieu, Stockholm, Epidemiology Survey; BMI = body mass index; CHS = Children's Health Study; Cl = confidence interval;
EC = elemental carbon; FEF = forced expiratory flow; FEF50% = forced expiratory flow at 50% of forced vital capacity. FEF25-75% = forced expiratory flow between 25 and 75% of
forced vital capacity; FEVi = forced expiratory volume in 1 second; FVC = forced vital capacity; GEE = generalized estimating equations; kPaxs/L = kilopascal times seconds per
liter; LUR = land use regression; MAAS = Manchester Asthma and Allergy Study; MMEF = maximum (or maximal) midexpiratory flow; NO = nitric oxide; NO2 = nitrogen dioxide;
NOX = sum of NO and NO2; NR = not reported; O3 = ozone; OC = organic carbon; PEF = peak expiratory flow; PIAMA = Prevention and Incidence of Asthma and Mite Allergy;
PM2 5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PM10 = particulate matter with a nominal mean aerodynamic diameter less than or
equal to 10 |jm; SD = standard deviation.
aStudies are presented in the order of appearance in the text.
bResults are presented for a 10-ppb change in NO2 and 20-ppb change in NOX unless otherwise specified.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
6-48
-------
The CHS has examined three separate cohorts for pollutant effects on lung function and
development [1993 cohort in Gauderman et al. (2004). 1993 and 1996 cohorts in
Gauderman et al. (2007) and Breton etal. (2011). and 2002 cohort in Urman et al.
(2014)1. The results of Breton et al. (2011) are consistent with earlier results from
Gauderman et al. (2004). Both Gauderman et al. (2007) and Urman et al. (2014) observed
that associations with NO2 assessed from central sites persisted with adjustment for
distance to freeway or near-roadway NOx, and Urman etal. (2014) observed that the
association with near-roadway NOx was not modified by PM2 5 measured at a central site.
Inference from these copollutant model results are limited because of the potential for
differential exposure measurement error.
Urman et al. (2014) examined lung function in 1,811 children (82% of the active cohort)
from eight communities in the CHS cohort established in 2002-2003. NO2 was assessed
from a single central site monitor in each of the study communities. LUR models were
developed from 900 monitoring sites in the CHS communities (Franklin et al.. 2012) to
estimate near-roadway NOx, NO2, and NO at children's homes. LUR models for NOx
used estimates from a dispersion model. For FEVi, there was little change in the
association of near-roadway NOx after adjusting for PM2 5 measured at the central site.
NOx associations also persisted with adjustment for PMio or Os. Central site NO2
remained associated with FEVi after adjustment for near-roadway NOx. Near-roadway
NO2 also was associated with decrements in lung function. Near-roadway NOx was
associated with lung function decrements in children with and without asthma, suggesting
that traffic-related pollution may affect all children. The association for
within-community near-roadway NOx was somewhat less than that for
between-community regional NOx, although the two effect estimates were based on
increments of pollution that may not be directly comparable.
Gauderman et al. (2007) reported results of an 8-year follow-up on 3,677 children who
participated in the CHS. Throughout the 8-year follow-up, around an 11% loss of study
participants per year was observed. The FEVi reduction was -31.5 mL (95% CI not
reported) for an increase in NO2 of 10 ppb. Children living <500 m from a freeway
(n = 440) had deficits in lung development over the 8-year follow-up compared to
children who lived at least 1,500 m from a freeway. When examined in the same model,
both distance to freeway and NO2 measured at community central sites were associated
with decrements in lung development. There was no evidence that the association for
NO2 differed according to distance to freeway or vice versa. Acid vapor, EC, PMio, and
PM2 5, but not Os, were associated with reduced lung development.
Gauderman et al. (2004) examined the 1993 CHS cohort of 1,759 children aged 10 to
18 years and report that although the average increase in FEVi over time was larger in
6-49
-------
boys than in girls, the associations of lung development with NC>2 measured at central site
monitor did not differ between the sexes. As depicted by the regression line in Figure 6-6.
for both sexes combined, the average difference in FEVi growth over the 8-year period
between the communities with the lowest and highest 8-year avg NC>2 concentration
(34.6 ppb difference) was -101.4 mL (95% CI: -165, -38.4). Gauderman et al. (2004)
further indicated that NO2 exposure over the 8-year follow-up was associated with
clinically relevant decrements in attained lung function at the age of 18 years (Figure
6-7). Clinical relevance was defined as FEVi less than 80% of the predicted value for
height, BMI, sex, race/ethnicity, and asthma status. Across the 12 communities, higher
NO2 was associated with an increase in the percentage of children with FEVi less than
80% predicted.
1450-
=• 1420-
-| 1390-
b
SP 1360H
;; 1330H
£ 1300H
c
^ 1270-
2 1240-1
1210-
0
0
O Girls
• Boys
Q>
O
10 15 20 25
N02 (ppb)
30 35
-2485
•2455 ^
-2425 5.
QQ
•2395 go
o
•2365 |
•2335 JJJ
c
•2305 ic
1
•2275 2
-2245
•0
40
Note: FENA, = forced expiratory volume in 1 second; NO2 = nitrogen dioxide; ppb = parts per billion.
Source: Reprinted with permission of the Massachusetts Medical Society, Gauderman et al. (2004).
Figure 6-6 Community-specific average growth in forced expiratory volume
in 1 second (mL) among girls and boys from 1993 to 2001, plotted
against average nitrogen dioxide concentrations from 1994
through 2000.
6-50
-------
Several recent studies examined lung function in cohorts other than the CHS. In
1,924 school-age children in the Swedish birth cohort BAMSE Schultzetal. (2012)
observed that NOx exposure during the first year of life was associated with a deficit in
FEVi. The ORs of having a deficit of 80% and 85% of predicted FEVi were 2.1 (95% CI:
0.6, 8.1), and 3.4 (95% CI: 1.6, 7.4), respectively, for a 47 ug/m3 increase in NOX.
Authors did not account for the influence of short-term NOx exposure but did observe the
association for long-term PMio to persist after accounting for short-term PMio.
o
I
1
00
V
lO-i
8-
6-
4-
2-
0
R=0.75
P=0.005
LA
0
10
» LN|
20
N02(ppb)
30
40
Note: AL = Alpine; AT = Atascadero; FE\A = forced expiratory volume in 1 second; LE = Lake Elsinore; LA = Lake Arrowhead;
LN = Lancaster; LM = Lompoc; LB = Long Beach; ML = Mira Loma; NO2 = nitrogen dioxide; P = p value; ppb = parts per billion; R =
correlation coefficent; RV = Riverside; SD = San Dimas; SM = Santa Maria; and UP = Upland.
Source: Reprinted with permission of the Massachusetts Medical Society, Gauderman et al. (2004).
Figure 6-7 Community-specific proportion of 18-year-olds with a forced
expiratory volume in 1 second below 80% of the predicted value,
plotted against the average concentrations of nitrogen dioxide
from 1994 through 2000.
6-51
-------
Limited data are available on NCh-related lung function changes in young children, such
as those age 4 years or younger because of the difficulties of lung function examinations
in this age group. Eenhuizen et al. (2013) measured interrupter resistance (Rint), an
indicator of airway resistance, in 4-year-old children participating in the Prevention and
Incidence of Asthma and Mite Allergy (PIAMA) Dutch birth cohort study. Of the
original invited 1,808 children, a total of 880 children were in the final analysis. The
children with valid Rint data did not have different characteristics than the population
recruited for the study. Long-term average concentrations of NC>2, PNfe.s, and soot at the
residential address at birth were assessed using LUR models as discussed in
Section 6.2.2.1. and daily average air pollution concentrations on the day of clinical
examination were obtained. Positive associations were observed between long-term
average NO2 concentrations and Rint. Such findings are supported by the study showing
NC>2-induced increased airway resistance in guinea pigs (Section 6.2.2.3). The findings
also support a relationship between NCh and asthma, given that Rint at age 4 years was a
predictor of asthma and wheeze at age 8 years. A monotonic increase in Rint with
increasing NC>2 concentration, with no suggestion of a threshold, was observed.
Short-term exposure was not associated with interrupter resistance. NO2 concentrations
on the test day and the day before the test were moderately correlated with long-term
concentrations. This is the first report of an association in 4-year-old children. NCh was
highly correlated with PIVb 5 and soot (quantitative data not reported), and an independent
association could not be discerned for any of the examined pollutants.
The long-term effects of PMio and NO2 exposure on specific airway resistance (sRaw)
and FEVi before and after bronchodilator treatment was examined within the Manchester
Asthma and Allergy Study (MAAS) birth cohort (Molter et al.. 2013). At age 11 years,
the cohort size was 813. A Microenvironmental Exposure Model incorporated an LUR
model and children's time-activity patterns to produce total person exposure estimates.
The model was validated, and there was good agreement between modeled and measured
total personal NO2 concentrations for short-term averaging times (Molter et al.. 2012).
Higher lifetime exposure to NC>2 was associated with a smaller increase in percent
predicted FEVi overtime, both before (16.0% [95% CI: -26.0, -0.5] fora 10-ppb
increase in NO2) and after bronchodilator treatment (23% [95% CI: -37.0, -9.0]).
As part of ESCAPE, Gehring et al. (2013) analyzed data from birth cohort studies
conducted in Germany, Sweden, the Netherlands, and the U.K. that measured lung
function at 6 to 8 years of age. The five birth cohorts [BAMSE, MAAS, German Infant
Nutritional Intervention South (GINI SOUTH), GINI/LISA, and PIAMA] are discussed
in Section 6.2.2.1. Annual average exposures to NO2, NOx, PM2 5, PMio, PM coarse, and
PM absorbance at the birth address and current address were estimated by LUR models,
except for the BAMSE cohort, for which a dispersion model was used. Associations of
6-52
-------
lung function with estimated air pollution concentrations and traffic indicators were
examined for each cohort using linear regression analysis, and then combined by random
effects meta-analysis. Across the five cohorts, annual mean (SD) for NO2 ranged from
7.44 (2.87) to 12.6 (1.91) ppb. Long-term associations were adjusted for short-term
changes in pollutants measured at central sites. NO2 and NOx estimated for the current
address were associated with decrements in both FEVi and forced vital capacity (FVC) as
were PM25 and PM25 absorbance. NO2 and PM25 at the current address also were
associated with peak expiratory flow (PEF). For the five cohorts combined, NO2 at the
birth address was associated with a smaller decrease in lung function than was NO2 for
the current address. Short-term (7-day avg) exposure to NO2 also was associated with
lung function decrements. Traffic intensity on the nearest street and traffic load on major
roads within a 100-m buffer were associated with deficits in lung function, although the
effect estimates had wide confidence intervals, indicating imprecise associations. Annual
average concentrations of NO2, NOx, PM2 5, and PMio at the current address were
associated with clinically relevant lung function decrements (FEVi < 85% predicted). In
copollutant models with NO2 and PM2 5, effect estimates for both pollutants were
reduced, but the relative impact on NO2 and PM2 5 differed among lung function indices.
Associations for FEVi and PEF were reduced more for NO2 than for PM25. In contrast,
the association for FVC was reduced more for PM2 5 than for NO2. Thus, the results do
not clearly discern an independent effect for either PM2 5 or NO2.
In Mexico City, Mexico, Rojas-Martinez et al. (2007b) and Rojas-Martinez et al. (2007a)
evaluated lung development in a prospective cohort of children aged 8 years at baseline.
Long-term NO2 exposures were assigned from the closest central monitoring site located
within 2 km of schools. An unspecified number of children were lost to follow-up during
the study, mainly because they moved to another area of the city or to another city
altogether. Information was obtained from atotal of 3,170 children. A 10-ppb increase in
NO2 was associated with an annual deficit in FEVi of 27 (95% CI: 22, 33) mL in girls
and 22 (95% CI: 16, 28) mL in boys. The negative association for NO2 persisted in
copollutant models with Os or PMio. A deficit in lung development was observed for NO2
after adjusting for the short-term associations with NO2 (previous-day concentrations).
A cohort study in Oslo, Norway, examined associations of short- and long-term NO2 and
other pollutant exposures on PEF and forced expiratory flow at 25% of forced vital
capacity and 50% of forced vital capacity in 2,307 children ages 9-10 years (Oftedal et
al., 2008). In models that included both short- and long-term NO2 exposures estimated
from dispersion models, only the association with long-term NO2 remained. Adjusting for
a contextual socioeconomic factor diminished the association with NO2.
6-53
-------
In summary, epidemiologic findings consistently indicate associations of long-term NCh
exposure with decrements in lung function and lung development in school children.
However examination of potential for confounding by PM2 5 and traffic-related
copollutants is limited, and it is unclear whether NC>2 exposure has an independent effect.
6.2.5.2 Lung Function in Adults
Both longitudinal and cross-sectional (Forbes et al. 2009b: Sekine et al. 2004) studies
are inconsistent in showing associations between long-term NCh exposure and lung
function. In the ECRHS cohort Gotschi et al. (2008). FEVi and FVC were assessed at
baseline and after 9 years of follow-up from 21 European centers (followed-up sample
N = 5,610). Quantitative results were not reported; NO2 was reported only to show no
statistically significant association with average lung function. This is in contrast to the
results from Ackermann-Liebrich et al. (1997) (SAPALDIA) and Schikowski et al.
(2005) [Study on the Influence of Air Pollution on Lung, Inflammation, and Aging
(SALIA)], which examined far more homogenous populations than the ECRHS cohort.
A recent study (Boogaard et al.. 2013) evaluated the impact on lung function of a
reduction in outdoor pollution concentrations resulting from restricting old heavy-duty
vehicles in all inner cities and other related policies. At 12 locations in the Netherlands,
NO2 was measured on the street where participants lived within 500 m of subjects'
homes. Respiratory health was measured in 2008 and 2010, during which air pollution
concentrations decreased. The study population included both children and adults, but
84% was above age 30 years at baseline. The participation rate was around 10%. Over
the two time periods, 585 subjects were re-evaluated for spirometry. Reductions in
concentrations of NC>2 and NOx as well as soot, copper (Cu), and iron (Fe) were
associated with increases in FVC. Airway resistance decreased with a decline in PMio
and PM2 5, although these associations were somewhat less consistent. No associations
were found with eNO. Results were driven largely by the small group of residents living
at the one urban street where traffic flow as well as air pollution were drastically reduced.
In a Nottingham, U.K. cohort of adults aged 18-70 years, lung function changes were
evaluated in a cross-sectional analysis of 2,599 subjects at baseline and a longitudinal
analysis of 1,329 subjects followed up 9 years later (Pujades-Rodriguez et al.. 2009).
There were no substantial cross-sectional associations between home proximity to the
roadside and NO2 concentration with lung function or any other outcome. Also, neither
exposure was associated with a decline in FEVi overtime. Insufficient contrast in NO2
exposure (interquartile range: 18.1-19.1 ppb) may be a factor in the inability to detect
any associations for NC>2 in this study population.
6-54
-------
6.2.5.3 Toxicological Studies of Lung Function
The 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) reported inconsistent evidence
of changes in lung function in animals after long-term ambient-relevant NC>2 exposures.
No recent studies are available. Arner and Rhoades (1973) exposed rats to 2,900 ppb NO2
continuously for 5 days/week for 9 months and reported changes in lipid composition in
the airway that could be related to observed functional consequences, including decreased
lung volume and compliance and increased surface tension, although these changes have
not been consistently observed in animal studies.
Tepper etal. (1993) exposed rats to a background concentration of 500 ppb NO2 for
16 h/day followed by a 6 hour peak of 1,500 ppb and 2 hours of downtime for up to
78 weeks. Frequency of breath was substantially slower in these animals and was
paralleled by a trend toward increased tidal volume, expiratory resistance, and inspiratory
and expiratory time, although changes were not statistically significant. Mercer etal.
(1995) and Miller et al. (1987) examined similar exposures in rats and mice, respectively,
and also reported that NO2 exposure did not alter lung function, although mice tended to
have slightly decreased vital capacity from 16 to 52 weeks of exposure.
Inconsistent effects were also described in studies of long-term NO2 exposure in the
range of 6-8 weeks. Stevens etal. (1988) exposed 1-day- and 7-week-old rats to 500,
1,000, and 2,000 ppb NC>2 continuously with two daily peaks at three times the baseline
concentration (1,500, 3,000, and 6,000 ppb) for 1-7 weeks and observed different results
among age groups. Rats exposed from 1 day of age had increased lung compliance after
3 weeks of exposure that returned to control levels by 6 weeks (1,000 ppb with 3,000 ppb
peaks). In rats exposed from 7 weeks of age, compliance was decreased after 6 weeks of
exposure at 1,000 and 2,000 ppb NC>2. In an 8-week study, Lafuma et al. (1987). reported
increased lung volumes in animals exposed to 2,000 ppb (8 h/day, 5 days/week), but vital
capacity and compliance were not affected.
6.2.5.4 Summary of Lung Function
Recent longitudinal studies continue to indicate associations between early-life NO2
exposure and decrements in lung function and lung development in school-aged children.
An association also was observed in young children aged 4 years. A linear
concentration-response relationship was observed in one study. Such associations also
are observed in cross-sectional studies (Gao etal.. 2013; Svendsen et al., 2012; Lee et al.,
201 Id; Rosenlund et al.. 2009b; Tager etal.. 2005; Sekine et al.. 2004; Moseler et al..
1994). A meta-analysis across five birth cohorts in Europe using LUR exposure estimates
reported results consistent with the rest of the evidence base. Evidence in adults is more
6-55
-------
limited and inconsistent. In children, much of the evidence is for FEVi, which reflects the
mechanical properties of the airways. There is less evidence for FVC, which represents
lung volume. As examined in two longitudinal studies, short-term NC>2 exposures did not
explain the association between long-term NO2 exposure and lung function. Associations
were observed in various locations using varied exposure assessment methods, lung
function measurements, and time of follow-up with children. Many studies demonstrated
that LUR models used to estimate individual-level residential NO2 exposure were able to
well predict measured NO2 concentrations in the study area.
An important uncertainty in the evidence base is whether NC>2 exposure has an
independent effect on lung function or lung development. Results for NC>2 were
inconsistent with PIVb 5 adjustment. In copollutant models with PIVb 5, NC>2 remained
associated with FVC but not FEVi or PEF. NOx association persisted with adjustment for
PM2 5, but there likely is differential measurement error for NOx exposure estimated for
individual children and PlVfc 5 measured at central sites. Confounding by traffic-related
pollutants is unexamined, and high copollutant correlations often are observed. Animal
studies do not address this uncertainty in the epidemiologic evidence as they demonstrate
inconsistent effects of long-term NC>2 exposure on lung function. However, age may be
an influential factor on the effect of NO2 on lung function that has not been adequately
addressed by the existing body of toxicological evidence.
6.2.6 Changes in Lung Morphology
While no recent studies are available, the 2008 ISA for Oxides of Nitrogen (U.S. EPA.
2008c) reported that animal toxicological studies demonstrate morphological changes to
the respiratory tract resulting from exposure to NO2 but variable responses to
concentrations below 5,000 ppb. Study characteristics are presented in Table 6-2. Wagner
et al. (1965) exposed dogs, rabbits, guinea pigs, rats, hamsters, and mice to 1,000 or
5,000 ppb NO2 for up to 18 months and found enlarged air space and edema and areas of
mild to moderately thickened septae with chronic inflammatory cells. However, some of
these observations were also made in control animals and were not considered to be
statistically significant in any species. Importantly, this study demonstrated differences in
sensitivity to NO2 across species. Furiosi et al. (1973) exposed monkeys and rats to
2,000 ppb NO2 continuously for 14 months and also found species-specific responses;
monkeys experienced hypertrophy of the bronchiolar epithelium that was most notable in
the respiratory bronchioles in addition to development of a cuboidal phenotype in the
squamous proximal bronchiolar epithelium. In rats, these effects were more occasional
under identical exposure conditions.
6-56
-------
The majority of other morphologic studies employed rodent models to evaluate effects of
NO2 exposure. Chang etal. (1986) compared responses in mature and juvenile rats to an
urban exposure pattern of NO2 for 6 weeks (500 ppb continuously with two daily peaks at
1,500 ppb). Mature rats were more sensitive to NO2 exposure and exhibited increased
surface density of the alveolar basement membrane and decreased air space in the
proximal alveolar regions. This was accompanied by an increase in lung volume
attributable to Type II cell hyperplasia and increases in fibroblasts, alveolar macrophages,
and extracellular matrix. In the juvenile rats, effects of exposure were limited to thinning
of Type II cells that were spread over more surface area compared to controls. Mercer et
al. (1995) found more subtle effects in rats with NC>2 exposure; frequency of fenestrae
was increased in the alveolar epithelium. However, there were no changes found in the
extracellular matrix or interstitial cells. Also, lungs did not appear to have differences in
alveolar septal thickness, parenchymal cell populations, or cellular size and surface area
after 9 weeks of exposure. Crapo etal. (1984) conducted a 6-week study in rats with a
similar exposure pattern at higher concentrations (2,000 ppb NC>2 for 23 h/day with two
30-minute peaks of 6,000 ppb) and reported hypertrophy and hyperproliferation of the
alveolar epithelium. In another study, rats were exposed to a similar urban exposure
pattern in addition to a single high concentration for up to 15 weeks; these animals had
subpleural alveolar macrophage accumulation and areas of focal hyperinflation, although
the mean linear intercept (MLI), a measure of free distance in the air space, was not
changed (Gregory et al.. 1983). Conversely, Lafumaet al. (1987) reported that hamsters
exposed to 2,000 ppb NO2 for 8 h/day, 5 days/week for 8 weeks had increased MLI and
decreased internal surface area, but no lesions were found in the bronchiole or
bronchiolar epithelium, alveolar ducts, or alveolar epithelium.
Kubota et al. (1987) assessed pathology of the airway in rats exposed continuously for 27
months to 40, 400, or 4,000 ppb NC>2. At the highest exposure, rats had increased
bronchial epithelial proliferation after 9 and 18 months, and by 27 months, proliferation
and edema resulted in fibrosis. Exposure to 400 ppb produced similar morphological
changes in the bronchial epithelium that was not apparent until 27 months. Exposure to
40 ppb NO2 did not result in morphological changes that could be identified by
microscopic techniques. Studies conducted at similar concentrations and durations,
500 ppb NC>2 for up to 19 months reported analogous effects. Blair et al. (1969) described
an increase in alveolar size after 3 months of exposure with loss of cilia in respiratory
bronchioles, which persisted at 12 months. After 4 months of exposure, Hayashi et al.
(1987) reported Type II cell hypertrophy and interstitial edema leading to thickened
alveolar septa at 6 months and fibrous pleural thickening at 9 months. Similarly, exposure
to 500 ppb for 7 months resulted in interstitial edema and Type II cell hyperplasia in rats,
and additional injury at 1,000 ppb included loss of cilia in the terminal bronchioles
(Yamamoto and Takahashi. 1984). Type II cell hyperplasia was also documented by
6-57
-------
Sherwin and Richters (1982) as was an increase in the MLI. These studies demonstrate
that long-term exposure to NC>2 can result in changes in lung morphology including
Type II cell hyperplasia, loss of cilia in the bronchiolar region, change in phenotype of
bronchiolar epithelium, fibrosis, and enlarged airspace, some of which are permanent,
while others may be transient.
6.2.7 Respiratory Infection
Toxicological studies, as reviewed in the 2008 ISA for Oxides of Nitrogen (U.S. EPA.
2008c), demonstrated NCh-induced mortality from infection in experiment animals as
well as changes in host defense mechanisms. Epidemiologic investigation of the
relationship between long-term NC>2 exposure and respiratory infection is limited,
particularly in longitudinal studies.
6.2.7.1 Epidemiologic Studies of Respiratory Infection
NO2 exposure is inconsistently associated with respiratory infection in infants and young
children up to 3 years of age (Aguilera et al.. 2013; Sunyer et al.. 2004). In a multicity
longitudinal study, Sunyer et al. (2004) observed no associations between 2-week indoor
NO2 exposure and lower respiratory tract infections during the first year of life. In
analyzing infections during the first year of life, authors used the 2-week NC>2
measurement to represent long-term exposure. Aguilera et al. (2013) observed an
association between increased NO2 exposure estimated by LUR and increased risk of
upper and lower respiratory tract infections in infants.
In a population-based case-control study in Hamilton, Ontario, Canada, Neupane et al.
(2010) examined hospital admission for community-acquired pneumonia in 345 adults
aged 65 years or more. Control participants (n = 494) aged 65 years or more were
randomly selected by telephone calls from the same community as cases from July 2003
to April 2005. Annual average NO2 exposures were estimated by three methods: central
site monitors, LUR models, and IDW. Participants had to present to the emergency room
with at least two signs and symptoms for pneumonia and have a new opacity on a chest
radiograph interpreted by a radiologist as being compatible with pneumonia. NO2 and
PM2 5 were associated with hospital admission for community-acquired pneumonia, but
SO2 was not. NO2 exposure estimated by all three methods were associated with
pneumonia (ORs per 10 ppb increase: 3.20 [95% CI: 1.37, 7.45] for IDW; 1.97 [95% CI:
1.21,3.19] for bicubic spline; 1.93 [95% CI: 1.00, 3.74] for LUR). There was no mention
of adjustment for short-term exposure effects, and it is not clear what the relative impacts
6-58
-------
on respiratory infections are for short-term versus long-term exposure. Associations are
observed between short-term NO2 exposure and hospital admissions for pneumonia
(Table 5-25). but quantitative comparisons with long-term NC>2 exposure effect estimates
may not be informative given the differences in exposure assessment methods and
distribution of NC>2 concentrations.
Parent report of physician-diagnosed pneumonia, croup, and otitis media during early
childhood was examined in 10 European birth cohorts: BAMSE (Sweden), Gene and
Environmental Prospective Study in Italy [GASPII (Italy)], Gene and Environmental
Prospective Study in Italy plus environmental and genetic influences (GINIplus),
Lifestyle-Related factors on the Immune System and the Development of Allergies in
Childhood plus the influence of traffic emissions and genetics [LISAplus (Germany)],
MAAS (U.K.), PIAMA (the Netherlands), and four Infancia y Medio Ambiente cohorts
[Spain; (Maclntyre et al.. 2014b)]. Annual average NO2 exposure was estimated using
LUR models and assigned to children based on their residential address at birth. Identical
protocols were used to develop LUR models for each study area. There was a complete
outcome (at least one), exposure (a minimum of NO2 and NOx), and potential confounder
information for 16,059 children across all 10 cohorts. For pneumonia, the meta-analysis
produced a combined adjusted OR of 1.64 (95% CI: 1.02, 1.65) per 10-ppb increase in
NO2. NO2 was associated with otitis media but not croup. NO2 measurements to build the
LUR models were made 2008-2011, but children in the study cohorts were born as early
as 1994. To address this temporal mismatch, a sensitivity analyses was conducted using
data from central site monitors to back-extrapolate LUR estimates and produced results
generally consistent with the main findings. NC>2 associations with pneumonia were
attenuated and had wide 95% CIs with adjustment PM2s or PM25 absorbance. ORs (95%
CI) forNC>2 in copollutant models, respectively, were: 1.32 (0.72, 2.42) and 1.36 (0.57,
3.28). Correlations between NO2and PM25 ranged between 0.42 and 0.80, and
correlations between NO2 and PM2s absorbance ranged between 0.40 and 0.93.
6.2.7.2 lexicological Studies of Respiratory Infection
Long-term NCh exposure has been shown to increase susceptibility of experimental
animals to infection. In Henry etal. (1970). squirrel monkeys exposed to 5,000 ppb NC>2
for 2 months and then exposed to Klebsiella pneumoniae or influenza had increased
markers of infection, white blood cell counts and erythrocyte sedimentation rate (ESR),
3 days post-infection. Furthermore, two of the seven monkeys exposed to NCh died at 3
and 10 days post-infection. When influenza virus was given 24 hours prior to NCh
exposure and after NO2 exposure, tidal volume and respiratory rate increased and the
ESR increased. One of the three exposed monkeys died 5 days post-infection. Ehrlich and
6-59
-------
Henry (1968) and Ehrlich (1980) also studied the effects of NO2 on Klebsiella
pneumonias infection in mice. Exposures were either continuous or intermittent (6 or
18 h/day) at a concentration of 500 ppb NCh, and bacterial challenge was administered at
1, 3, 6, 9, and 12 months. Continuous exposure to NC>2 for 3 months or longer increased
mortality rates after infection, whereas intermittent exposures increased mortality at 6, 9,
and 12 months. Likewise, Miller et al. (1987) showed increased mortality in mice
exposed to a base of 200 ppb NCh with two daily 1-hour peaks of 800 ppb and
subsequent challenge with Streptococcus zooepidemicus at 16, 32, and 52 weeks.
6.2.7.3 Subclinical Effects Underlying Respiratory Infection
Impaired host defense mechanisms can increase susceptibility to bacterial and viral
infection, and toxicological studies in experimental animals demonstrate that
ambient-relevant NO2 exposures for periods greater than 6 weeks can modulate lung host
defense including alter characteristics of AMs. Characteristics of these studies, which
were also reviewed in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c). are
presented in Table 6-2.
Alveolar macrophages play a critical role in removing pathogens from the airways, and
impaired function can increase susceptibility to infection and injury. Chang et al. (1986)
showed that exposure to 500 ppb NO2 continuously with 1,500 ppb 1-hour peaks twice
daily for 6 weeks increased the number of AMs in the alveoli and their cellular volume.
Gregory et al. (1983) reported similar findings and observed AM accumulation in lung
sections by light microscopy after exposure to 5,000 ppb NC>2 or a base of 1,000 ppb NO2
with 5,000 ppb spikes twice each day for 15 weeks.
Greene and Schneider (1978) investigated the function of AMs isolated from
antigen-sensitized baboons exposed to 2,000 ppb NCh for 8 h/day, 5 days/week for
6 months and found that AMs had diminished response to migration inhibitory factor
obtained from antigen-stimulated lymphocytes. However, sample size in this study was
small: 3 exposed to NO2 and antigen, 1 exposed to NO2 alone, 1 exposed to antigen
alone, and 1 air control. Other studies have not reported on this endpoint.
In addition to AMs, mast cells play an important role in host defense and inflammatory
processes. Fujimaki and Nohara (1994) exposed both rats and guinea pigs to 1,000,
2,000, or 4,000 ppb NO2 continuously for 12 weeks. Although the number of mast cells
in the airway increased after exposure to 2,000 and 4,000 ppb, these changes were not
statistically significant. Histamine, released by mast cells, was reduced in rats at
2,000 ppb NO2 and increased in guinea pigs at 4,000 ppb. This observation suggests
species differences in response to NO2 exposure.
6-60
-------
6.2.7.4 Summary of Respiratory Infection
In the small body of epidemiologic studies, long-term NO2 exposure estimated for
subjects' homes by LUR was associated with respiratory infections in school children and
pneumonia hospital admissions in adults, ages 65 years or older. Results are inconsistent
in infants. Particularly for hospital admissions, it is not clear whether the association
observed for long-term NO2 exposure is independent of an association with short-term
exposure. As examined in school children, associations for long-term NO2 exposure were
positive with adjustment for PM2 5 or PM2 5 absorbance, but the 95% CIs were very wide.
Thus, an independent association for NO2 is not clearly indicated. A small body of
toxicological studies provide support for an independent effect of NO2 exposure on
respiratory infections, showing that mice and monkeys exposed to 500 or 5,000 ppb NO2
for periods greater than 6 weeks have increased infection-induced mortality and AM
numbers in the airways.
6.2.8 Chronic Obstructive Pulmonary Disease
Recent epidemiologic studies examined associations between long-term NO2 exposure
and effects related to COPD, including the study of indoor NO2 and respiratory
symptoms in adults with COPD (Hansel et al.. 2013), described in Section 6.2.3.1. Few
studies examined COPD development, and results are inconsistent. In a longitudinal
cohort study, Andersen et al. (2011) estimated outdoor annual average NO2 and NOx
since 1971 by a validated LUR model for residential locations and calculated
time-weighted averages for 15-, 25- and 35-year periods (Raaschou-Nielsen et al.. 2000).
No other pollutants were considered. COPD hospital admissions were ascertained from
1976, and incidence of COPD was defined as first hospital admission between
1993-1997 and June 2006. COPD incidence was associated with the 35- and 25-year
mean concentration of NO2 (HR: 1.28; [95% CI: 1.07, 1.54] and 1.22 [95% CI: 1.03,
1.45] per 10-ppb increase) and 35-year mean concentration of NOx (1.16 [95% CI: 1.04,
1.31] per 20-ppb increase). Weaker positive associations were observed with 25-year
mean NOx, 15-year mean NO2 and NOx, and baseline residence traffic proxies (major
road within 50m, traffic load within 200 m). The associations with NO2 were stronger
than those with NOx. The association was stronger in people with diabetes and asthma
compared to the rest of the cohort, but no difference in association was observed by
smoking or occupational exposure. COPD incidence was most strongly associated with
35-year avg NO2, suggesting that long-duration, possibly lifetime exposure may be
associated with development of COPD.
6-61
-------
Gan etal. (2013) evaluated a population-based cohort in Canada that included a 5-year
exposure period and a 4-year follow-up period. All residents aged 45-85 years of
Metropolitan Vancouver, Canada, during the exposure period and did not have known
COPD at baseline were included in this study (n = 467,994). Five-year average
residential exposures to NCh and NO were estimated using LUR models, incorporating
changes in exposure overtime due to changes in residences. COPD incidence was
ascertained from a hospital admissions database and defined as admission during the
follow-up period. Mortality also was studied and is discussed in Section 6.5.2. The
association of 5-year NO2 with COPD hospital admission was null. The exposure period
examined in this study was shorter than that in Andersen et al. (2011) (i.e., 25-35 years).
In the ECRHS cohort, the association of NOx with prevalence of COPD and related
symptoms was investigated by two methods for assessing exposure to power
plant-specific emissions of NOx (Amster et al.. 2014). NOx exposures (8-year avg)
related to power plant emissions were estimated for subjects' residences (n = 2,244)
based on kriging ambient concentrations from 20 central site monitors downwind of the
power plant (source approach), and peak emission events (event approach) were defined
as 30-minute concentrations that exceeded 125 ppb NO2. Neither source-based nor
event-based power plant NOx emissions were associated with COPD prevalence.
Respiratory symptoms were associated with source-based NOx but not event-based NOx.
In a cross-sectional study, Wood et al. (2009) examined respiratory phenotype (PiZZ
type) in alpha 1-antitrypsin deficiency (a-ATD) from the U.K. a-ATD registry. This
deficiency leads to exacerbated responses to inflammatory stimuli. In total, 304 PiZZ
subjects underwent full lung-function testing and quantitative high-resolution computed
tomography to identify the presence and severity of COPD emphysema. Annual average
NO2 was estimated for subjects' homes with dispersion models. NO2 was associated with
improved gas transfer and less severe emphysema. Similar associations were observed
with SO2 and particles. In contrast, Os was associated with worse gas transfer and more
severe emphysema, albeit accounting for only a small proportion of the lung function
variability. NO2 was negatively correlated with Os, which might explain NO2 associations
with gas transfer and emphysema severity. No information was provided on how well the
dispersion model captured the spatial pattern of long-term NO2 concentrations.
6.2.9 Summary and Causal Determination
There is likely to be a causal relationship between long-term NO2 exposure and
respiratory effects, based strongly on evidence integrated across disciplines for a
relationship with asthma development. There is more uncertainty in relationships with
6-62
-------
lung function and partially irreversible decrements in lung development in children,
respiratory disease severity, chronic bronchitis/asthma incidence in adults, COPD
hospital admissions, and respiratory infection.
The conclusion of a "likely to be causal relationship" represents a change from the
"suggestive of, but not sufficient to infer, a causal relationship" determined in the 2008
ISA for Oxides of Nitrogen (U.S. EPA. 2008c). The main difference in the evidence base
is several recent longitudinal studies that indicate associations between asthma incidence
in children and long-term NCh exposures estimated for at or near children's homes or
schools and biological plausibility from previous experimental studies. In contrast, the
2008 ISA for Oxides of Nitrogen reported inconsistent findings from a limited number of
cross-sectional studies that examined asthma prevalence. An additional uncertainty
identified in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) was the potential
for NO2 to serve primarily as a surrogate for another traffic-related pollutant or mixture.
Because of the high correlations among traffic-related pollutants and limited examination
of copollutant confounding, the independent effects of long-term NO2 exposure could not
be clearly discerned in the last review. While this uncertainty continues to apply to the
epidemiologic evidence across the respiratory effects examined, coherence of
epidemiologic evidence for asthma incidence with the limited previous toxicological
evidence for both AHR and development of allergic responses, which are key events in
the proposed mode of action for asthma development, provides support for an
independent effect of long-term exposure to NO2 on development of asthma. The key
evidence supporting the "likely to a causal relationship" is detailed in Table 6-5 using the
framework described in Table II of the Preamble to this ISA.
6.2.9.1 Evidence on Development of Asthma
Multiple longitudinal studies demonstrate associations between higher ambient NO2
concentrations measured in the first year of life, in the year of diagnosis, or over a
lifetime and asthma incidence in children. Results are consistent across locations based
on various study designs and cohorts (Table 6-1). Consistency across studies in the use of
questionnaires to ascertain parent report of physician-diagnosed asthma, a best practice
(Burr. 1992; Ferris. 1978). adds to the strength of inference about associations with NO2.
A pooled analysis across six birth cohorts relating NO2 with ever asthma (OR: 1.48 [95%
CI: 1.06, 2.04] per 10 ppb increase) (Macintyre et al.. 2014a) is consistent with results
from individual studies. A strength of several studies is NO2 exposures estimated at or
near children's homes or schools from central sites 1 km away or from LUR models that
were demonstrated to represent well the spatial variability in the study areas. LUR
6-63
-------
models were developed with traffic metrics assessed at various buffers [e.g., 100 m
(Gehring et al.. 2010). 500 m (Clougherty et al.. 2007)]. Associations were observed
across a range of ambient NO2 concentrations [(Carlsten et al., 201 Ic; Gehring et al..
2010: Jerrett et al.. 2008: Cloughertv et al.. 2007): Table 6-11. In limited analysis of the
concentration-response relationship, results did not consistently indicate a linear
relationship in the range of ambient NC>2 concentrations examined (Carlsten et al.. 201 Ic:
Shima et al.. 2002). These studies did not conduct analyses to evaluate whether there is a
threshold for effects. Limited supporting evidence for incidence of asthma or chronic
bronchitis incidence in adults is provided in the ECRHS cohort in relation to NC>2
measurements outside subjects' homes (Sunyer et al.. 2007) and estimated by dispersion
models that were validated against ambient measurements.
Epidemiologic studies of asthma development in children have not clearly characterized
potential confounding by PlVfcs or traffic-related pollutants [e.g., CO, BC/EC, volatile
organic compounds (VOCs)]. In the longitudinal studies, correlations with PIVb 5 and BC
were often high (e.g., r = 0.7-0.96), and no studies of asthma incidence evaluated
copollutant models to address copollutant confounding, making it difficult to evaluate the
independent effect of NO2. Across studies that examined both NO2 and PIVb 5, PIVb 5
concentrations were associated with asthma development (Nishimura et al.. 2013: Clark
et al.. 2010: McConnell et al.. 2010a). Effect estimates were smaller in magnitude for
PM2 5 compared to NO2, but because of the potential for differential exposure
measurement error, comparisons of effect estimates may not be meaningful.
The uncertainty in the epidemiologic evidence base is partly reduced by the biological
plausibility provided by findings from experimental studies that demonstrate
NO2-induced effects on key events in the mode of action proposed for the development of
asthma (Figure 4-2). Though not consistently demonstrated, AHR was induced in guinea
pigs after 6-12 weeks of exposure to NO2 [1,000-4,000 ppb; (Kobayashi and Miura.
1995)]. and there is some evidence that airway remodeling was involved. Experimental
studies also indicate that short-term exposure repeated over several days and long-term
NC>2 exposure can induce Th2 skewing/allergic sensitization by showing increased Th2
cytokines, airway eosinophils, and IgE-mediated responses (Sections 4.3.5 and 6.2.2.3).
Epidemiologic evidence for NCh-related pulmonary inflammation is inconsistent
(Section 6.2.2.3). but reported in a recent longitudinal study (Berhane et al.. 2014). The
association was independent of short-term changes in NO2 concentrations, and elevated
eNO was associated with increased risk of new onset asthma in the cohort. Recurrent
pulmonary inflammation and oxidative stress are identified as key early events in the
proposed mode of action for asthma development (Figure 4-2). While the effects of
long-term NC>2 exposure on oxidative stress in toxicological studies are variable and
transient (Section 6.2.2.3). some evidence supports a relationship between short-term
6-64
-------
NC>2 exposure and increased pulmonary inflammation. Evidence from controlled human
exposure studies indicates that repeated NO2 exposure increases neutrophils in healthy
adults, and epidemiologic evidence also points to associations between ambient NO2
concentrations and increases in pulmonary inflammation in healthy children and adults
(Section 5.2.2.5). Findings for short-term NCh exposure support an effect on asthma
development by describing a potential role for repeated exposures to lead to recurrent
inflammation and allergic responses.
The limited evidence base for NCh-related development of AHR and allergic responses
and increases in pulmonary inflammation combined with the consistent epidemiologic
evidence for NCh-related development of asthma in children are coherent and indicate a
biologically plausible sequence of events by which long-term NO2 exposure could lead to
asthma development.
6.2.9.2 Evidence on Lung Function
Another line of evidence indicating a relationship between long-term NC>2 exposure and
respiratory effects includes multiple, longitudinal epidemiologic studies observing
associations between long-term NO2 exposure and decrements in lung function and
partially irreversible decrements in lung development in children. Expanding on evidence
reviewed in the 2008 ISA, recent studies consistently demonstrate associations with
individual-level NO2 exposure estimates based on time-activity patterns and/or LUR
(Urman et al.. 2014; Eenhuizen et al.. 2013; Molter et al.. 2013). Associations are also
observed with NCh assessed from central sites. Gauderman et al. (2004) found an NCh
concentration-dependent decrement in lung development but based on comparisons
among communities not individual subjects.
Potential confounding of long-term NC^-related decrements in lung function and lung
development by traffic-related copollutants has not been evaluated, although an
association was observed with adjustment for Os or PMio. Toxicological studies do not
clearly support epidemiologic findings. NC>2-induced changes in lung function were
inconsistently demonstrated in animal models [(Tepper et al., 1993; Stevens etal. 1988;
Lafumaetal.. 1987); Section 6.2.5.3]. Long-term NC>2 exposure was observed to alter
lung morphology in adult experimental animals but not juvenile animals (Section 6.2.6).
but the changes observed do not appear to contribute to altered lung function or explain
the effects observed in epidemiologic studies. Thus, it remains unclear whether NO2
exposure has an independent effect on lung function or lung development.
6-65
-------
6.2.9.3 Evidence on Respiratory Disease Severity
Several longitudinal studies consistently demonstrate increases in respiratory symptoms
in children with asthma in relation to increased ambient NC>2 concentrations
(Section 6.2.3 and Table 6-3). Associations were observed with NO2 estimated from
central sites and NO2 estimated for children's homes using LUR. Studies did not examine
whether associations of long-term NC>2 were independent of short-term exposure;
however, McConnell et al. (2003) assessed chronic symptoms as a daily cough for
3 months or congestion/phlegm for 3 months. Limited information from longitudinal
studies of indoor NC>2 support an association with respiratory symptoms in children with
asthma and adults with COPD (Belanger et al.. 2013; Hansel et al.. 2013). Findings for
indoor NC>2 exposure provide support for an independent relationship between NO2 and
respiratory effects because NO2 may exist as part of a different air pollutant mixture
indoors than in the ambient air (Section 5.2.9.6). In limited analysis of copollutant
models, associations of NO2 with respiratory symptoms in children persisted with
adjustment for PM25 or the traffic-related pollutants EC or OC. Potentially limiting
inference from these results, pollutants were measured from central sites, and correlations
with NO2 were high for PM2 5 and EC (0.75 and 0.92, respectively).
6.2.9.4 Evidence on Respiratory Infection
In the limited body of epidemiologic studies, findings do not consistently indicate
associations between long-term NO2 exposure and respiratory infection. Findings in
infants are inconsistent, and associations with pneumonia hospital admissions in adults
could be due to short-term exposure. An evaluation of 10 European birth cohorts
demonstrated associations of residential estimates of NO2 exposure with parental report
of physician-diagnosed pneumonia and otitis media (Maclntyre et al.. 2014b).
Adjustment for PIVb 5 or PIVb 5 absorbance produced associations for NC>2 with wide 95%
CIs, limiting inferences about independent NO2 associations (Section 6.2.7.2). The
strongest toxicological evidence for long-term NC>2 exposure leading to respiratory
effects is that for respiratory infection (Section 6.2.7.2). Exposure to 200-5,000 ppb NC>2
for 1 month up to 1 year increased mortality in rodents and squirrel monkeys following
bacterial challenge (Miller et al.. 1987; Henry et al.. 1970) and altered AM numbers and
morphology (Gregory et al.. 1983; Aranyi et al.. 1976). The latter are identified as key
events in the proposed mode of action linking NC>2 exposure to respiratory infection.
6-66
-------
6.2.9.5 Analysis of Potential Confounding by Traffic-Related Copollutants
Potential confounding of long-term NC>2 associations with respiratory effects by
traffic-related copollutants has been examined to a limited extent, particularly in
longitudinal analyses of asthma incidence. In longitudinal studies of children, copollutant
models were analyzed for chronic bronchitic symptoms (McConnell et al., 2003) and
lung function and respiratory infection (Maclntyre et al.. 2014b: Gehring et al.. 2013).
NC>2 associations varied with adjustment for EC, OC, PIVb 5 absorbance, or PM2 5, from
not changing or being modestly reduced to being attenuated. In some cases, copollutants
were moderately correlated with NO2 (r = 0.37-0.46 for PIVb 5, 0.52 for PIVb 5
absorbance, 0.58 for OC). However, high correlations often were reported (r = 0.72-0.80
for PM2 5 or 0.75-0.92 for PIVb 5 absorbance or EC). Although some studies indicate an
independent association for long-term NCh exposure with some respiratory effects,
inconsistency in the evidence and limited analysis of the array of potential confounding
traffic-related copollutants make it difficult to disentangle the independent effect of NO2
from other traffic-related pollutants or mixtures in the epidemiologic studies.
6.2.9.6 Conclusion
Taken together, recent epidemiologic studies and previous experimental studies provide
evidence that there is likely to be a causal relationship between long-term NO2 exposure
and respiratory effects (Table 6-5). This conclusion is based on the evidence for asthma
development. Evidence for other respiratory effects such as respiratory disease severity,
lung function changes, and respiratory infection is more uncertain because the combined
epidemiologic and/or experimental evidence does not clearly demonstrate an independent
effect of long-term NO2 exposure. Recent epidemiologic studies consistently indicate
increases in asthma incidence in children particularly in association with NC>2 exposures
estimated at or near children's homes or schools. Potential confounding by copollutants
of greatest concern, PlVfcs and traffic-related copollutants, largely is unexamined.
Experimental evidence indicating AHR induced by long-term NC>2 exposure and
development of an allergic phenotype with repeated short-term or long-term NC>2
exposure provides biological plausibility by characterizing a potential mode of action by
which long-term NC>2 exposure may lead to asthma development. However, because this
experimental evidence is limited, there remains some uncertainty regarding an
independent effect of long-term NO2 exposure on asthma development. Thus, the
combined consistent epidemiologic evidence and consistent but limited experimental
evidence for development of asthma is sufficient to conclude that there is likely to be a
causal relationship between long-term NO2 exposure and respiratory effects.
6-67
-------
Table 6-5 Summary of key evidence for a likely to be a causal relationship
between long-term nitrogen dioxide exposure and respiratory
effects.
Rationale for Causal
Determination3
Key Evidence13
Key References13
NO2 Concentrations
Associated with Effects0
Asthma Development
Consistent epidemiologic
evidence from multiple,
high-quality studies with
relevant NO2
concentrations
Consistent evidence for
increases in asthma
incidence in diverse cohorts
of children in U.S., Europe,
Canada, and Asia.
Asthma ascertainment by
parental report of doctor
diagnosis.
t Carlsten et al. (2011 c),
tClouqherty et al. (2007),
tGehrinq et al. (2010),
tJerrett et al. (2008),
Shima et al. (2002)
Weak evidence:
tRanzietal. (2014)
Section 6.2.2.1, Table 6-1,
Figure 6-1
Supporting evidence for
asthma incidence or chronic
bronchitis in the ECHRS
cohort of adults.
tJacquemin etal. (2009b).
tModiq et al. (2009),
Sunveret al. (2006)
Section 6.2.2.2
Consistent evidence for
NO2 metrics with lower
potential for exposure
measurement error
NO2 estimated for children's
homes with well-validated
LUR models or by monitoring
at or near children's
homes/schools.
tCarlsten et al. (2011c).
tGehrinq et al. (2010),
tJerrett et al. (2008),
Shima et al. (2002)
Section 6.2.2.1
Means across studies of
LUR model: 13.5, 17.3,
27.5 ppb
75th percentile: 15.1,
15.4 ppb, Max: 69.4 ppb
Range in mean residential
NO2 across communities:
9.6 to 51.3 ppb
Range in mean central site
NO2 across communities:
7.3-31.4 ppb
Uncertainty regarding
potential confounding by
PM2.5 or traffic-related
copollutants
Correlations with PM2.5 and
EC often were high
(r= 0.7-0.96). Copollutant
models not analyzed.
Associations found with
adjustment for SES, family
history of asthma, smoking
exposure, gas stove in home
tMcConnelletal. (2010a)
Table 6-1
Some evidence for key
events in the proposed
mode of action
Allergic responses
Increased IgE-mediated
histamine release in mast
cells from rodents.
Fujimaki and Nohara
(1994)
Section 6.2.4.1
4,000 ppb for 12 weeks
Experimental findings for
development of Th2
phenotype with short-term
NO2.
Pathmanathan et al.
(2003). Ohashietal. (1994)
Section 5.2.7.4
2,000 ppb over
4 consecutive days;
3,000 ppb for 2 weeks
Inconsistent epidemiologic
evidence
Section 6.2.4.1
6-68
-------
Table 6-5 (Continued): Summary of key evidence for a likely to be a causal
relationship between long-term nitrogen dioxide exposure
and respiratory effects.
Rationale for Causal
Determination3
Key Evidence13
Key References13
NO2 Concentrations
Associated with Effects0
Airway
hyperresponsiveness
Increased airway
responsiveness with short-
term exposure in
experimental studies of
healthy adults and guinea
pigs
Humans: Section 5.2.7.1
Guinea pigs: Kobayashi
and Shinozaki (1990)
Humans: 1,000-2,000 ppb
for 3 h but not below
Guinea pigs: 4,000 ppb for
7 days
Limited evidence in guinea
pigs with long-term
exposure; increased airway
resistance suggests airway
remodeling.
Kobayashi and Miura
(1995)
1,000-4,000 ppb for
6-12 weeks
Inflammation
Increases in lymphocytes,
PMNs, in rats with long-term
exposure.
Increases in PMNs in healthy
adults with repeated
short-term exposure.
Kumae and Arakawa
(2006),
Blomberq et al. (1999)
Section 6.2.2.3,
Section 5.2.7.4
500 or 2,000 ppb; prenatal
or postnatal exposure up to
12 weeks of age for 4 days
Inconsistent epidemiologic
evidence with exposure
assessment by LUR and
central site.
tBerhane et al. (2014),
tLiuetal. (2014a)
Section 6.2.2.3
Limited epidemiologic
evidence in healthy children
and adults with short-term
exposures assessed in
subjects' locations and
associations adjusted for
BC/EC, OC, PNC, or PM2.5.
tStraketal. (2012),
tSteenhofetal. (2013),
tLinetal. (2011)
Section 5.2.7.4
Max for 5-h avg: 96 ppb
Means for 24-h avg across
seasons: 24.3-45.3 ppb
Oxidative stress
Varying and transient effects
on antioxidant levels and
enzyme activity.
Avaz and Csallanv (1978), 400, 1,000, 5,000 ppb for
Gregory et al. (1983), Saqai 6 weeks to 18 mo
etal. (1984)
Section 6.2.2.3
6-69
-------
Table 6-5 (Continued): Summary of key evidence for a likely to be a causal
relationship between long-term nitrogen dioxide exposure
and respiratory effects.
Rationale for Causal
Determination3
Key Evidence13
Key References13
NO2 Concentrations
Associated with Effects0
Severity of Asthma
Consistent epidemiologic
evidence but uncertainty
regarding NO2
independent effects
Consistent evidence for
increases in respiratory
symptoms in children with
asthma. Exposure
assessment by central site
measurements and LUR
model.
McConnell et al. (2003),
tGehrinq et al. (2010)
Table 6-3
Section 6.2.3.2
Associations with respiratory
symptoms remain robust with
adjustment for a
traffic-related copollutant:
PM2.5, EC, orOC. But,
analysis is limited and based
on central site exposure
assessment.
In limited analysis,
associations with respiratory
symptoms remain robust with
adjustment for Os, SO2,
PMio-2.5, or PM-io.
McConnell et al. (2003),
tHwanq and Lee (2010)
Table 6-3
Residential NO2 by LUR
model:
Mean: 13.5 ppb
10th-90th percentile:
7.8-18.5 ppb
Central site:
. Mean, Max 4-yr avg for
12 communities:
19.4, 38.0 ppb
McConnell etal. (2003),
tHwanq and Lee (2010)
Evidence for associations
between indoor NO2 and
respiratory symptoms in
children with asthma ages
5-10 yr; inconsistent
evidence in younger children
and infants.
tBelanqeretal. (2013)
Section 6.2.3.1
Mean daily indoor NCb:
10.6 ppb
75th: 12.5 ppb
Lung Function and Development
Consistent epidemiologic
evidence from multiple,
high-quality studies but
uncertainty regarding NO2
independent effects
Epidemiologic evidence for
decrements in lung function
and partially irreversible
decrements in lung
development in children.
Gauderman et al. (2004).
Rojas-Martinez et al.
(2007a),
tMolter et al. (2013),
tGehrinq et al. (2013).
tUrmanetal. (2014).
tEenhuizen etal. (2013)
Section 6.2.5.1
NO2 by LUR model:
Means across
communities: 7.4-12.6 ppb
Overall study mean:
13.5 ppb, 75th: 15.4 ppb
Central site NO2 mean
across communities:
27.2-42.6 ppb
In limited analysis,
associations are inconsistent
with adjustment for PM2.5 but
robust with adjustment for
PM-io or Os.
Residential NO2-PM2.5
correlations vary across
cohorts. Pearson
r= 0.31-0.76.
tGehrinq et al. (2013),
Rojas-Martinez etal.
(2007b)
6-70
-------
Table 6-5 (Continued): Summary of key evidence for a likely to be a causal
relationship between long-term nitrogen dioxide exposure
and respiratory effects.
Rationale for Causal
Determination3
Key Evidence13
Key References13
NO2 Concentrations
Associated with Effects0
Uncertain relevance of
toxicological evidence
Changes in lung morphology
including increases in
edema, hypertrophy of lung
epithelium, fibrotic changes
in adult not juvenile animals.
Uncertain relevance to
epidemiologic findings.
Kubotaetal. (1987),
Hayashietal. (1987)
Section 6.2.6
500 ppbfor 19 mo,
4,000 ppbfor9-27 mo
Respiratory Infection
Consistent toxicological
evidence
Increased mortality of mice
and monkeys with NO2
exposure and challenge with
bacterial or viral infection.
Henry et al. (1970),
Ehrlich and Henry (1968).
Ehrlich(1980).
Miller etal. (1987)
Section 6.2.7.2
500 ppb for 3 mo,
5,000 ppb for 2 mo,
200 ppb base plus daily
spike of 800 ppb for
16-52 weeks
Limited and inconsistent
epidemiologic evidence
Associations with
physician-diagnosed
pneumonia and otitis media
in multicounty European
cohort study but not
consistently in other studies.
tMacintyre etal. (2014a)
Section 6.2.7.1
Range in mean across
10 birth cohorts:
7.5-23.7 ppb
Limited evidence for key
events in proposed mode
of action
Increased AM infiltration to
lung tissue or increased
lymphocytes in BAL fluid of
experimental animals.
Gregory etal. (1983)
Section 6.2.7.3
5,000 ppb for 15 weeks
COPD
Limited and inconsistent
epidemiologic evidence
Inconsistent evidence for
hospital admissions for
COPD in adults. Unclear
whether independent of
short-term exposure effects.
tAndersen et al. (2011).
tGanetal. (2013)
Section 6.2.8
AM = alveolar macrophage; BAL = bronchoalveolar lavage; BC = black carbon; COPD = chronic obstructive pulmonary disease;
EC = elemental carbon; ECHRS = European Community Respiratory Health Survey; IgE = immunoglobulin E; LUR = land use
regression; NO2 = nitrogen dioxide; O3 = ozone; OC = organic carbon; PM2 5 = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |jm; PM10 = particulate matter with a nominal mean aerodynamic diameter less
than or equal to 10 |jm; PM10-25 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm and
greater than a nominal mean of 2.5 |jm; PMN = polymorphonuclear cell(s), polymorphonuclear leukocyte; PNC = particle number
concentration; SES = socioeconomic status; SO2 = sulfur dioxide; Th2 = T-derived lymphocyte helper 2.
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Tables I and N. of the
Preamble.
""Describes the key evidence and references, supporting or contradicting, contributing most heavily to causal determination and,
where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where full body of evidence is
characterized.
°Describes the NO2 concentrations with which the evidence is substantiated (for experimental studies, < 5,000 ppb).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
6-71
-------
6.3 Cardiovascular Effects and Diabetes
6.3.1 Introduction
The 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) concluded that "the available
epidemiologic and toxicological evidence was inadequate to infer the presence or absence
of a causal relationship" between cardiovascular effects and long-term NO2 exposure.
This section updates the previous review with the inclusion of recent studies on the
cardiovascular effects of NO2 and NOx exposure in humans, animals, and cells.
Additionally, recent evidence on the relationship between diabetes and exposure to NO2
and NOx is discussed; there were no studies available on this relationship at the time of
the last review. Data from individual studies related to cardiovascular health effects and
diabetes can be found in summary tables in each section, and an integrated summary of
the evidence is presented in Section 6.3.9.
At the completion of the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) one
epidemiologic study of the association of cardiovascular disease (CVD) with long-term
exposure to NO2 was available for review. Miller et al. (2007) studied
65,893 post-menopausal women (50-79 years old) without previous CVD from 36 U.S.
metropolitan areas. Exposures to air pollution were estimated by assigning the annual
(2000) mean air pollutant concentration measured at the monitor nearest to the subject's
five-digit residential ZIP code centroid. In single-pollutant models, PM2 5 showed the
strongest associations with the CVD events [myocardial infarction (MI),
revascularization, angina, congestive heart failure, coronary heart disease (CHD) death],
followed by SO2. The association of NO2 with overall CVD events was 1.04 (95% CI:
0.96, 1.12) per 10-ppb increase, and NO2 was not associated with CVD events when the
data set was restricted to those study participants with nonmissing exposure data.
Previous animal toxicological studies were limited to examination of changes in heart
rate, vagal response, and alterations in specific hematological parameters
[e.g., hematocrit, hemoglobin, erythrocytes; (U.S. EPA. 2008c. 1993a)1.
Large, prospective studies with consideration of potential confounding and other sources
of bias have become available since the completion of the 2008 ISA for Oxides of
Nitrogen (U.S. EPA. 2008c). and are emphasized in this section (see Appendix for study
evaluation guidelines). The exposure assessment method was also an important
consideration in the evaluation of long-term exposure and cardiovascular effects and
diabetes, given the spatial variability typically observed in ambient NO2 concentrations
6-72
-------
(Section 2.5.3). Exposure assessment was evaluated drawing upon discussions in
Section 3.2 and Section 3.4.5. A select number of recent studies employed exposure
assessment methods to account for the spatial variability of NC>2. For example, LUR
model predictions generally have been found to correlate well with outdoor NO2
measurements (Section 3.2.2.1). For long-term NC>2 exposure, exposure assessment was
evaluated by the extent to which the method represented the spatial variability in NO2
concentrations in a given study. For modeled estimates, such information includes
statistics indicating the correlation between predicted and measured NO2 concentrations.
Several recent epidemiologic studies report positive associations of NO2 and NOx
exposure with heart disease, stroke, hypertension, and diabetes. The body of evidence is
generally consistent for heart disease and includes several large, longitudinal studies with
consideration of multiple potential confounding factors including age, sex, BMI,
smoking, and pre-existing conditions (Table 6-6). There is also a consistent body of
evidence for diabetes comprising large, longitudinal studies that consider multiple
potential confounding factors including age, sex, BMI, smoking, and pre-existing
conditions (Table 6-10). Some of these studies employed validated exposure assessment
methods such as LUR, which were demonstrated to capture the spatial variability of NO2
concentration. The extent to which the studies inform the independent effect of NC>2
exposure through their consideration of copollutants of greatest concern (i.e., CO, BC,
PM2.5) and noise is discussed in this section. A small number of experimental animal
studies examining the effect of NO2 on oxidative stress and the progression of vascular
disease provide limited support for the biological plausibility of the cardiovascular effects
observed in the epidemiologic studies.
6.3.2 Heart Disease
Several studies published since the 2008 ISA for Oxides of Nitrogen examine the
association of long-term NO2 exposure and heart disease. Although the evidence from the
epidemiologic studies is not entirely consistent, several prospective studies and/or studies
with exposure assessment strategies designed to capture the spatial variability of NO2
report positive associations. Most studies adjust for a wide array of potential confounders
such as age, sex, BMI, smoking, and pre-existing conditions (Table 6-6). but uncertainty
remains on the extent to which the findings can be explained by correlated copollutant
exposures and noise.
Cesaroni et al. (2014) reported an increased risk of incident coronary events of 1.06
(95% CI: 0.96, 1.16) per 10-ppb increase in NO2. This large study of 11 cohorts from
5 countries used LUR to assign exposure at each participant's residence. Authors
6-73
-------
reported good performance of exposure models based on their comparison of predicted
estimates and concentrations measured at 40 sites (R2 > 0.61). A study of pulmonary
patients in Toronto, Canada (Beckerman et al., 2012) reported an increased risk of 1.17
(95% CI 1.01, 1.36) per 10-ppb increase in NCh with ischemic heart disease (IHD)
prevalence after adjustment for individual covariates as well as simultaneous adjustment
for Os and PM2 5. In this study, LUR was also used to estimate NO2 concentrations, which
were assigned at the post code centroid level (typically one block area or specific
building in this study area). In another prospectively designed study, Gan et al. (2011)
examined the association of long-term exposure to BC, PIVb 5, NCh, and NO with CHD
hospital admission and mortality among participants (45-85 year-olds) residing in
Vancouver, Canada enrolled in the universal health insurance system. In this study, LUR
was used to predict NO2 concentrations at a resolution of 10 m. These predicted
concentrations were adjusted using factors derived from regulatory monitoring data and
then linked to each participant's postal code of residence (typically one block or specific
building in this study area). After adjustment for potential confounders, NC>2 and NO
were inversely associated with CHD hospitalization (HR: 0.93 [95% CI: 0.89, 0.98] and
HR: 0.96 [95% CI: 0.92, 0.98] per 10 ppb, respectively); however, positive associations
of NO2 and NO with CHD mortality were observed (Section 6.5.2).
6-74
-------
Table 6-6 Epidemiologic studies of long-term exposure to oxides of nitrogen
and heart disease.
Study
Cohort, Location,
Study Period
Exposure Assessment and
Concentration (ppb)
Effect Estimates (95% Cl)a
tCesaroni et al.
(2014)
ESCAPE Project,
11 cohorts
5 countries in
Europe
2008-2012
LUR model
Annual avg NO2, NOx
Model developed from
40 monitoring sites and linked to
geocoded addresses.
NO2 means and ranges across
cohorts: 4.2 (3.2-5.8) to 31.9
(22.3-40.9)
Coronary events
NO2HR: 1.06(0.96, 1.16)
NOxHR: 1.01 (0.98, 1.05)
per 20 ug/m3 NOx
Covariate adjustment: marital
status, education, occupation,
smoking status duration and
intensity, SES.
Copollutant adjustment: none
tGanetal. (2011)
Population-based
cohort
Vancouver, Canada
1999-2002
n = 452,735
LUR model
5-yr avg (NO2 and NO, 1995-1998)
and 4 yr avg (1999-2002; 10-m
spatial resolution).
Concentrations assigned to postal
code centroids (typically ~1 city
block in urban areas and larger in
less populated areas).
NO2: Mean 16.3, IQR4.5
NO: Mean 26.1, IQR 10.8
CHD hospital admission
(ICD-9 410-414)
NO2RR: 0.93(0.89,0.98)
NORR: 0.96(0.92, 0.98)
per 10 ppb NO2and NO
Covariates: age, sex, pre-existing
diabetes, COPD, hypertension,
SES.
Copollutant adjustment: none
tBeckerman et al.
(2012)
Cohort of pulmonary
patients
Toronto, Canada
1992-1999
n = 2,414
LUR model
Average of fall 2002 and spring
2004
NO2 exposures assigned at the
postal code centroid (typically 1
block, or single building and larger
in less populated areas). Cross-
validation showed mean 4%
difference between modeled and
measured NO2.
NO2: Median 22.9, IQR 4.0
IHD prevalence (ICD-9
412-414)—old Ml, angina or other
IHD
RR: 1.24(1.01, 1.40)
Covariate adjustment: sex, age,
pack-yr smoking, BMI, deprivation
index, diabetes.
RR 1.17(1.01, 1.36)after
adjustment for covariates above
plus 03 and PM2.s (in same
model).
tRosenlund et al.
(2009a)
SHEEP Study
Stockholm, Sweden
1985-1996
n = 24,347 cases,
276,926 controls
Dispersion model
5-yr avg NO2
Modeling of traffic-related
emissions. 25-m resolution inner
city, 100-m urban, 500-m
regional/countryside.
Concentrations assigned to
residential address.
5th-95th: 15.9 cases
Median: 6.9 cases, 6.3 controls
First nonfatal Ml
OR: 0.93(0.78, 1.12)
Covariates: age, sex, calendar yr,
Copollutant adjustment: none
6-75
-------
Table 6-6 (Continued): Epidemiologic studies of long-term exposure to oxides of
nitrogen and heart disease.
Study
Cohort, Location,
Study Period
Exposure Assessment and
Concentration (ppb)
Effect Estimates (95% Cl)a
tHartetal. (2013)
NHS cohort
11 States in the U.S.
1990-2008
n = 121,700
Dispersion model
Annual avg (2000) NO2
Assigned to residential address.
Main results were for traffic
proximity.
Concentrations NR
Incident Ml
HR: 1.22(0.99, 1.50) per 1-ppb
increase in NO2 between
addresses
Covariate adjustment: BMI,
physical activity, healthy diet
score, alcohol,
hypercholesterolemia, high blood
pressure, diabetes, family history
of Ml, smoking status, mental
health status, father's occupation,
marital status, husband's
education, education level,
employment, median
income/home value.
Copollutant adjustment: none
tLipsett et al.
(2011)
CIS cohort
California, U.S.
Jun 1996-Dec2005
n = 124,614
Central site monitor concentrations
combined by IDW
Gridded pollutant surface (250-m
spatial resolution) created and
concentrations linked to geocoded
residential address. Defined
representative range of 3-5 km
(neighborhood and regional
monitors, respectively) for NOx and
NO2 to account for spatial variability
of pollutant.
NO2: Mean: 33.6, IQR: 10.3
NOx: Mean: 95.6, IQR: 58.3
Ml incidence
NO2HR: 1.06(0.88, 1.23)
NOxHR: 1.00(0.95, 1.05)
Covariate adjustment: age, race,
smoking second-hand smoke,
BMI, lifetime physical activity,
nutritional factors, alcohol, marital
status, menopausal status,
hormone therapy, hypertension
medication and aspirin, family
history of Ml/stroke.
Copollutant adjustment: none
tAtkinson et al.
(2013)
General practice
patient national
cohort
U.K.
2003
n = 836,557
Dispersion model
Annual average NO2 (2002)
Model incorporated all known
emissions sources (1 * 1 km
resolution). Concentrations linked to
residential post code centroids that
typically include 13 residential
addresses.
Mean: 12.0, IQR: 5.7
Ml incidence
HR: 0.97(0.90, 1.04)
Arrhythmia incidence
HR: 0.98(0.91, 1.04)
Heart failure incidence
HR: 1.11 (1.02, 1.21)
Covariates: age, sex, smoking,
BMI, diabetes, hypertension,
index of multiple deprivation.
Copollutant adjustment: none
tde Kluizenaar et
al. (2013)
Eindhoven, the
Netherlands
1991-2003
n = 18,213
Dispersion model
Annual avg NO2
1 x 1 km resolution, linked to
residential address.
5th-95th percentile range: 7.5
CBVDand IHD:
1.16(0.95,1.45)
Covariates: age, sex, BMI,
smoking, education, exercise,
marital status, alcohol use, work
situation, financial difficulties
1.18(0.93, 1.48) after adjustment
for the covariates above plus
noise
6-76
-------
Table 6-6 (Continued): Epidemiologic studies of long-term exposure to oxides of
nitrogen and heart disease.
Study
Cohort, Location,
Study Period
Exposure Assessment and
Concentration (ppb)
Effect Estimates (95% Cl)a
tDonq et al.
(2013a)
33 communities in
11 districts of 3 cities
in Liaoning Province,
China
2006-2008
n = 24,845
Central site monitor
Communities were 1 km of a
monitor (selected to maximize intra-
and inter-city gradients).
District-specific 3-yr avg NO2
Mean: 18.7, Median: 17.5, IQR: 4.8
Self-reported CVD
OR: 1.04(0.60, 1.87)
Copollutant adjustment: none
Miller etal. (2007)
WHI cohort Central site monitor
36 U.S. cities Annual avg (2000) NO2
1994-1998 Nearest monitor to residence
ZIP code centroid (overall effect
based on intra- and inter-city
gradients).
Concentrations NR
Incident CVD events
HR: 1.04(0.96, 1.12)
Covariates: age, ethnicity,
education, household income,
smoking, diabetes, hypertension,
systolic blood pressure, BMI,
hypercholesterolemia.
Copollutant adjustment: none
BMI = body mass index; CBVD = cerebrovascular disease; CHD = coronary heart disease; Cl = confidence interval;
COPD = chronic obstructive pulmonary diesease; CIS = California Teachers Study; CVD = cardiovascular disease;
Dec = December; ESCAPE = European Study of Cohorts for Air Pollution Effects; HR = hazard ratio; ICD = international
Classification of Diseases; IDW = inverse distance weighting; IHD = ischemic heart disease; IQR = interquartile range; LUR = land
use regression; Ml = myocardial infarction; NHS = Nurses Health Study; NO = nitric oxide; NO2 = nitrogen dioxide; NOX = sum of
NO and NO2; NR = not reported; O3 = ozone; OR = odds ratio; PM25 = particulate matter with a nominal mean aerodynamic
diameter less than or equal to 2.5 |jm; RR = risk ratio(s), relative risk; SES = socioeconomic status; SHEEP = Stockholm Heart
Epidemiology Program; WHI = Women's Health Initiative.
aEffect estimates are reported per 10-ppb increase in NO2 or NO and 20-ppb increase in NOX unless otherwise specified. NOX
results that are originally reported in ng/m3 are not standardized if the molecular weight needed to convert to ppb is not reported.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
Several other studies characterized NO2 exposure using IDW estimates of concentration
from central site monitors or dispersion models that captured a range of spatial
resolutions. Uncertainties associated with these models are described in detail in
Sections 3.2.2 and 3.2.3. Briefly, estimates derived from IDW monitor concentrations
may not capture the true variability in NC>2 concentration from local sources if the
monitor coverage is not adequately dense. Biases in dispersion model output can occur in
either direction and depend on the complexity of the topography, meteorology, and
sources that are modelled. Two studies addressed these potential uncertainties by
demonstrating that dispersion models well represented the spatial pattern of NO2
concentrations in the study communities. Dispersion model estimates were shown to
agree well with NO2 concentrations measured at sites in the communities (Atkinson et al..
2013; Rosenlund et al.. 2009a).
Lipsettetal. (2011) determined the association of incident MI with long-term exposure to
NC>2, NOx, other gases (CO, Os, 802), and PM in a prospective study. These authors
followed a cohort of California public school teachers aged 20-80 years old
(n = 124,614). Each participant's geocoded residential address was linked to a pollutant
6-77
-------
surface with a spatial resolution of 250 m, which was determined by IDW interpolation
of NC>2 concentrations measured at central site monitors within 3 km of a home. Those
living outside the radial range for which the monitor was intended to provide
representative data were excluded from the analysis. The authors observed a positive
association between NO2 and incident MI (HR: 1.06 [95% CI: 0.88, 1.23] per 10 ppb). In
a study of women enrolled in the Nurses' Health Study (NHS), Hartet al. (2013) reported
an increased risk of incident MI associated with living consistently near sources of traffic.
Although the main analyses in this study was for distance to roadway, the authors used a
dispersion model to predict the change in NC>2 concentration among those who moved
from one address to another. They observed an increased risk of incident MI in
association with current NC>2 compared with NCh concentration at the previous address
(Table 6-6).
Rosenlund et al. (2009a) conducted a case-control study of first MI using the registry of
hospital discharges and deaths for Stockholm County, Sweden and randomly selected
population-based controls. Predicted 5-year avg NO2 concentrations were determined and
linked to each participant's geocoded address using dispersion models. The resolution of
the predicted concentrations corresponded to 500 m in the countryside, 100 m in urban
areas, and 25 m in the inner city. Five-year average NC>2 concentration was associated
with fatal MI (OR: 1.14 [95% CI: 1.09, 1.19] per 10 ppb) but not with nonfatal MI (OR:
0.93 [95% CI: 0.78, 1.12] per 10 ppb). CO and PMio were also associated with fatal cases
of MI in this population. Atkinson et al. (2013) examined the association of incident
cardiovascular disease with NO2. These authors studied patients (aged 40-89 years)
registered with 205 general practices across the U.K. Predicted annual average NO2
concentrations within 1x1 km grids, estimated using dispersion models, were assigned
to participants based on their residential postal code. Cardiovascular disease outcomes
included in the analysis were MI, arrhythmias, and heart failures. The authors reported a
positive association between NO2 and heart failure in fully adjusted models (HR: 1.11
[95% CI: 1.02, 1.21] per 10 ppb). Incident MI and arrhythmia were not associated with
NO2 concentration in this analysis. A similar pattern of findings was observed for the
associations between PM and these outcomes (associations with CHD and MI were null
while the association of PMio with heart failure was positive).
de Kluizenaar et al. (2013) assigned NO2 exposure to participants' residential addresses,
based on dispersion modelling with a 1 * 1 km resolution, reported an association of NO2
with CHD and cerebrovascular disease hospital admissions that was robust to adjustment
for individual-level covariates and noise (HR 1.18 [95% CI: 0.93, 1.48] per 10 ppb
increase in NO2). As discussed above, the only study available for inclusion in the
previous assessment was (Miller et al.. 2007). which compared annual average
concentration assigned at the ZIP code centroid level to study participants across 36 U.S.
6-78
-------
cities. This study reported a null association between NO2 and incident CVD events.
Dong etal. (2013a) reported a small, imprecise increase in the prevalence of self-reported
CVD comparing 3-year avg concentrations for communities within 1 km of an air
monitoring station across 3 cities in Liaoning, China (OR: 1.04 [95% CI: 0.60, 1.87] per
10 ppb increase NCh).
Overall, several epidemiologic studies, including some large studies with prospective
designs; adjustment for known risk factors for cardiovascular disease such as age, sex,
BMI, and smoking; and NO2 exposures estimated at or near homes with well-validated
LUR models (Cesaroni et al.. 2014; Beckerman et al.. 2012). provide evidence for
associations between long-term exposure to NO2 and risk of heart disease. Although
positive associations between MI and CHD were not observed consistently across studies,
with some reporting null or inverse associations with CHD or MI morbidity, others
reported increased risk of mortality due to CHD or MI (Gan etal.. 2011; Rosenlund et al..
2009a), and a positive association with heart failure was reported by Atkinson etal.
(2013). The few studies that examined confounding by PM2 5 (Beckerman et al.. 2012) or
noise (de Kluizenaar et al.. 2013) provide showed that NO2 estimates are robust to
adjustment for these factors. However, PM25 confounding was analyzed with Os in a
multipollutant model, which is prone to produce unreliable results. In general, the
epidemiologic studies were not designed to distinguish the independent effect of NO2
from the effects of other traffic-related pollutants (e.g., BC, EC, CO), noise, or stress.
6.3.3 Cerebrovascular Disease and Stroke
Several studies published since the 2008 ISA for Oxides of Nitrogen examine the
association of long-term NO2 exposure and stroke (Table 6-7). Evidence from
epidemiologic studies is not consistent, and there is uncertainty in the extent to which the
findings can be explained by noise or exposures to traffic-related copollutants.
A hospital-based, case-control study in Edmonton, Canada reported a positive association
of NO2 exposure with ischemic stroke (OR: 1.06 [95% CI: 0.88, 1.28] per 10 ppb
increase) and a stronger positive association with hemorrhagic stroke (OR: 1.14 [95% CI
0.85, 1.54]) but not with transient ischemic attack [OR: 0.90 (95% CI: 0.74, 1.10);
(Johnson et al.. 2013)1. This was the only study of stroke to use LUR to estimate NO2
concentration at the participants' residences, but no information was reported on how
well the model predicted NO2 concentrations in the study area. Findings were similar in
an ecological analysis of annual incidence of stroke also conducted in Edmonton,
Canada. Positive, imprecise associations with hemorrhagic and nonhemorrhagic stroke
incidence were observed with IDW average NO2 concentration assigned based on
6-79
-------
residential postal code (Johnson et al., 2010a). Associations of stroke with CO and traffic
density were also observed in this study.
Table 6-7 Epidemiologic studies of long-term exposure to oxides of nitrogen
and cerebrovascular disease or stroke.
Cohort, Location,
Study Study Period
Exposure Assessment and
Concentration (ppb)
Effect Estimates (95% Cl)a
tJohnson et
al. (2013)
Edmonton, Canada
Jan 2007-Dec 2009
n = 4,696 cases,
37,723 controls
LUR model
NO2 concentrations matched to
residential postal code (spatial
resolution <50 m). No information
reported on model validation.
Stroke hospital admissions:
All stroke: 1.02(0.88, 1.17)
IS: 1.06(0.88, 1.28)
TIA: 0.90(0.74, 1.10)
HS: 1.14(0.85, 1.54)
Means: 15.4 for cases, 15.2 for controls Covariate adjustment: age, sex,
Copollutant adjustment: none
tJohnson et
al. (2010a)
Edmonton, Canada
Jan 2003-Dec 2007
Central site monitor concentrations
combined by IDW
5-yr avg NO2
Concentrations assigned at postal code
centroid level.
Mean: 15.7, IQR: 2.2
Ecological analysis of stroke
incidence rates:
HS ED visits
Q1 RR 1.0 (reference)
Q2RR: 0.88(0.68, 1.14)
Q3RR: 1.03(0.79, 1.20)
Q4RR: 1.13(0.90, 1.43)
Q5RR: 1.14(0.92, 1.52)
nonHS ED visits
Q1 RR 1.0 (reference)
Q2RR: 1.0(0.85, 1.18)
Q3RR: 1.05(0.92, 1.20)
Q4RR: 1.02(0.87, 1.18)
Q5RR: 1.08(0.91, 1.27)
Covariate adjustment: age, sex,
household income.
Copollutant adjustment: none
tS0rensen et
al. (2014)
Diet, Cancer, and
Health cohort
Copenhagen or
Aarhus, Denmark
1993/1997-Jun 2006
n = 57,053
LUR and dispersion model combined
Annual avg NO2
Concentrations linked to geocoded
residential address, r = 0.67 for
modeled and measured NO2 in
high-traffic street canyon.
Median: 8.1, 10th: 6.3, 90th: 12.4,
IQR: 3.0
Noise-adjusted IS incidence
NO2 (at the time of diagnosis) IRR:
1.04(0.85, 1.24)
NOxIRR: 0.97(0.92, 1.03)
per 20 |ig/m3 NOx
Covariate adjustment: age, sex,
education, municipality, SES,
smoking status and intensity,
intake of fruits, vegetables,
alcohol, and coffee, physical
activity, BMI, calendar yr.
Copollutant adjustment: none
6-80
-------
Table 6-7 (Continued): Epidemiologic studies of long-term exposure to oxides of
nitrogen and cerebrovascular disease or stroke.
Cohort, Location,
Study Study Period
Exposure Assessment and
Concentration (ppb)
Effect Estimates (95% Cl)a
tAndersen et
al. (2012b)
Diet, Cancer, and
Health cohort
Copenhagen or
Aarhus, Denmark
1993/1997-Jun 2006
n = 57,053
LUR and dispersion model combined
Annual avg NO2
Concentrations linked to geocoded
residential address, r = 0.67 for
modeled and measured NO2 in
high-traffic street canyon.
Median: 8.1, 10th: 6.3, 90th: 12.4,
IQR: 3.0
IS incidence
NO2HR: 1.19(0.88, 1.61)
HS incidence
NO2HR: 0.80(0.53, 1.23)
Covariate adjustment: smoking
status, duration, intensity; ETS;
sex; BMI; education; sports
activity; alcohol, fruit, fat
consumption; hypertension;
hypercholesterolemia.
Copollutant adjustment: none
Eindhoven, the
uizenaar et Netherlands
al. (2013) 1991-2003
n = 18,213
Dispersion model
Annual avg NO2
1 x 1 km resolution, linked to residential
address.
5th-95th percentile range: 7.5
IHD or CBVD hospital admissions
HR: 1.16(0.95, 1.45)
Covariate adjustment: age, sex,
BMI, smoking, education, marital
status, exercise, alcohol use, work
situation, financial difficulty.
Copollutant adjustment: none
HR: 1.18(0.93, 1.48) after
adjustment for covariates above
plus noise
tLipsett et al.
(2011)
CIS cohort
California, U.S.
Jun 1996-Dec2005
n = 133,479
Central site monitor concentrations
combined by IDW
Gridded pollutant surface (250-m spatial
resolution) created and concentrations
linked to geocoded residential address.
Defined representative range of 3-5 km
(neighborhood and regional monitors,
respectively) for NOx and NO2 to
account for spatial variability of
pollutant.
NO2: Mean: 33.6, IQR: 10.3
NOx: Mean: 95.6, IQR: 58.3
Stroke incidence
NOxHR: 1.01 (0.98, 1.05)
NO2HR: 1.02(0.90, 1.16)
Covariate adjustment: age, race,
smoking, second-hand smoke,
BMI, lifetime physical activity,
nutritional factors, alcohol, marital
status, menopausal status,
hormone therapy, hypertension,
medication and aspirin, family
history of Ml/stroke.
Copollutant adjustment: none
tAtkinson et General practice Dispersion model
al (2013) patient national cohort Annual average NO2 (2002)
U.K. Model incorporated all known emissions
2003 sources (1 x 1 km resolution).
n = 836,557 Concentrations linked to residential post
code centroids that typically include
13 residential addresses.
Mean: 12.0, IQR: 5.7
Stroke incidence
HR: 0.98(0.91, 1.06)
Covariates: age, sex, smoking
BMI, diabetes, hypertension, index
of multiple deprivation.
Copollutant adjustment: none
6-81
-------
Table 6-7 (Continued): Epidemiologic studies of long-term exposure to oxides of
nitrogen and cerebrovascular disease or stroke.
Cohort, Location,
Study Study Period
Exposure Assessment and
Concentration (ppb)
Effect Estimates (95% Cl)a
tDonq et al.
(2013a)
33 communities in
11 districts of 3 cities
in Liaoning Province,
China
2006-2008
n = 24,845
Central site monitor
Communities were 1 km of a monitor
(selected to maximize intra- and
inter-city gradients).
District-specific 3-yr avg NO2
Mean: 18.7, Median: 17.5, IQR: 4.8
Self-reported stroke:
OR: 1.27(0.92, 1.76)
Sex-specific results also
presented.
Covariate adjustment: age,
gender, education, occupation,
family income, BMI, hypertension,
family history of stroke, family
history of CVD, smoking status,
drinking, diet, exercise.
Copollutant adjustment: none
BMI = body mass index; Cl = confidence interval; CIS = California Teachers Study; CBVD = cerebrovascular disease;
CVD = cardiovascular disease; Dec = December; ED = emergency department; ETS = environmental tobacco smoke;
HR = hazard ratio; HS = hemorrhagic stroke; IDW = inverse distance weighting; IHD = ischemic heart disease; IQR = interquartile
range; IS = ischemic stroke; LUR = land use regression; Ml = myocardial infarction; NO2 = nitrogen dioxide; NOX = sum of NO and
NO2; nonHS = nonhemhorragic stroke; OR = odds ratio; Q1 = 1st quintile; Q2 = 2nd quintile: Q3 = 3rd quintile; Q4 = 4th quintile;
Q5 = 5th quintile; RR = risk ratio(s), relative risk; SD = standard deviation; SES = socioeconomic status; TIA = transient ischemic
attack.
aEffect estimates are reported per 10-ppb increase in NO2 or NO and 20-ppb increase in NOX unless otherwise specified. NOX
results that are originally reported in (xg/m3 are not standardized if the molecular weight needed to convert to ppb is not reported.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
Andersen et al. (2012b) conducted a study of long-term traffic-related NO2 exposure and
incident stroke using data from a large cohort study of residents of Copenhagen,
Denmark enrolled in the Danish Diet, Cancer, and Health Study. The Danish GIS-based
air pollution and human exposure dispersion modelling system was used to predict NC>2
concentrations for geocoded residential address histories up to approximately 35 years in
duration. The authors reported an increase in ischemic stroke incidence (HR 1.19 [95%
CI: 0.88, 1.61]) but not hemorrhagic stroke incidence (HR: 0.80 [95% CI: 0.53, 1.23). In
an analysis of the same data set that adjusted for traffic-related road noise as a potential
confounder, S0rensen et al. (2014) reported a substantially attenuated risk estimate
between NCh concentration at the time of diagnosis and ischemic stroke [incidence rate
ratios (IRR: 1.04 [95% CI: 0.85, 1.24]). The association for the combined effect of the
highest tertile of noise and the highest tertile of NO2 was increased (IRR: 1.28 [95% CI:
1.09, 1.52]), and the association with fatal strokes persisted after adjustment for noise. A
study of IHD and cerebrovascular diseases combined reported that associations with NO2
were robust to adjustment for noise and other individual-level covariates (HR: 1.18
[95% CI 0.9, 1.48] per 10 ppb increase) (de Kluizenaaretal.. 2013).
Lipsettetal. (2011) analyzed the association of incident stroke with long-term exposure
to NO2, NOx, other gases (CO, Os, 862), and PM. These authors analyzed data from a
cohort of California public school teachers and assigned exposure by linking IDW NC>2
concentrations from monitors within 3 km of participants' geocoded addresses. An
association with incident stroke that was close to the null value (HR: 1.02 [95% CI: 0.90,
6-82
-------
1.16] per 10-ppb increase inNCh) was observed. Estimates for the association of other
pollutants (PMio, PM2 5, 862, and Os) with incident stroke were increased.
Atkinson et al. (2013) examined the association of incident cardiovascular disease with
NO2. These authors studied patients (aged 40-89 years) registered with 205 general
practices across the U.K. Predicted annual average NO2 concentrations within 1x1 km
grids, estimated using dispersion models, were assigned to participants based on their
residential postal code. Incident stroke was not associated with NO2 concentration in this
analysis. An increase in NC>2 concentration for communities within 1 km of an air
monitoring station was associated with self-reported stroke prevalence in a multicity
study in China (OR 1.27 [95% CI: 0.92, 1.76] per 10 ppb increase) (Dong et al.. 2013a).
Oudin etal. (2011) reported no association between long-term NOx exposure and
ischemic stroke in a population- and registry-based, case-control study conducted in
Scania, Sweden. Exposure was characterized using dispersion models to estimate outdoor
NOx concentrations within 500 x 500 m grids and linking those predicted concentrations
to geocoded residential addresses. Although no association of NOx exposure with stroke
was observed, modification of the association of diabetes and stroke by long-term NOx
exposure was reported in this study.
Although several studies report an increased risk between NO2 exposure and stroke
and/or cerebrovascular disease, estimates are generally imprecise (de Kluizenaar et al..
2013; Dong etal.. 2013a; Johnson et al.. 2013; Andersen et al.. 2012b; Johnson et al..
2010a). Some studies reported weak or null associations (Atkinson et al.. 2013; Lipsett et
al.. 2011). The positive associations observed for stroke were not consistent across stroke
subtype. Johnson etal. (2013) observed a larger increased risk for hemorrhagic compared
to ischemic stroke in a LUR study, while Andersen et al. (2012b) observed an increase
for ischemic not hemorrhagic stroke in the Danish Diet, Cancer, and Health Study. The
association with ischemic stroke observed by Andersen et al. (2012b) was diminished
after further adjustment for noise, although an interaction between the highest tertile of
NO2 and highest tertile of noise was observed (S0rensen et al.. 2014). Evidence from
epidemiologic studies is not consistent; one study used LUR to estimate NO2 exposures
at homes but did not report information on how well the model captured the fine-scale
variability of NO2 concentrations. Further, there is uncertainty in the extent to which
findings can be explained by noise or exposures to traffic-related copollutants.
6.3.4 Hypertension
There were no studies of the effect of long-term NO2 or NOx exposure on hypertension in
the 2008 ISA for Oxides of Nitrogen. Several recent studies of both children and adults
6-83
-------
add to the evidence base (Table 6-8). Overall, findings from studies of both adults and
children are inconsistent. Further, the independent effect of NC>2 is not distinguished from
the effect of noise and other traffic pollutants in the epidemiologic studies reporting
positive associations.
Coogan et al. (2012) examined the association of long-term NOx exposure with incident
hypertension among black women residing in Los Angeles, CA. An LUR model was
used to estimate exposure at each participant's residential address, and a cross-validation
R2 of 0.92 indicated that the model predicted well the pattern of NOx concentrations in
the study area. These authors reported an increased risk of 1.24 (95% CI: 1.05, 1.45) per
20-ppb increase in NOx after adjustment for a wide array of potential confounders
including traffic-related noise exposure. Slight attenuation in the effect estimate for NOx
was reported after adjustment for PM2 5 Although the correlation between NOx and
PM25 was low (r = 0.27), PIVbs concentration was estimated using kriging of central site
monitor measurements and may be subject to differential exposure measurement error.
Further, correlations between NOx and other traffic-related copollutants were not
reported.
In a cross-sectional study, Foraster et al. (2014) reported that NO2 was associated with an
increase in systolic blood pressure that was attenuated to varying degrees depending on
the method of adjustment for medication use but not with hypertension. In this study,
LUR was used to estimate NO2 exposure among participants (35-83 years) of a large
population-based cohort study. A cross-validation R2 of 0.63 indicated good agreement
between modeled and measured concentrations. Both short-term exposure to NO2 and
noise were adjusted for in the analysis, in addition to an array of other potential
confounders. Associations with systolic blood pressure were stronger among those with
cardiovascular disease, those living alone, and those living in areas with high traffic load
and traffic noise. In a large study involving Danish adults (50-64 year old), S0rensen et
al. (2012) reported an inverse association between NO2 and both systolic and diastolic
blood pressure at baseline while largely null findings were reported for self-reported
incident hypertension. A validated dispersion model was used in this study to predict
annual and 5-year avg NO2 concentration and to assign exposure based on residential
address history [r = 0.90 for measured and predicted 1/2-year avg NO2 concentration in
greater Copenhagen, Denmark area; (S0rensen et al.. 2012)1. Dong et al. (2013b) reported
imprecise associations of prevalent hypertension and increased blood pressure with
average NO2 concentration from monitoring stations located within 1 km of the
community where study participants resided. Stronger associations of hypertension with
PMio, SO2, and Os were reported in this study.
6-84
-------
Table 6-8 Epidemiologic studies of long-term exposure to oxides of nitrogen
and hypertension and blood pressure.
Study
Cohort, Location,
Study Period
Exposure Assessment and
Concentrations (ppb)
Effect Estimates (95% Cl)a
Studies of Adults
tCooqan et al.
(2012)
BWHS cohort
Los Angeles, CA
2006
n = 3,236
LUR model
Annual avg NOx
Model incorporated traffic, land use,
population, and physical geography
to estimate exposure at the
participants' residences. Summer
and winter field measurements taken
at 183 sites in Los Angeles, CA.
Cross-validation with 16 sites,
R2 = 0.92
Mean: 43.3, Median: 41.6, IQR: 12.4
Hypertension
IRR: 1.24(1.05, 1.45)
Covariate adjustment: age, BMI, yr
of education, income, number of
people in the household, smoking,
drinks per week, h/week of physical
activity, neighborhood SES score,
noise.
Copollutant adjustment: none
IRR: 1.19(1.00, 1.40) after
adjustment for covariates above and
PM2.5
tForaster et al.
(2014)
REGICOR cohort
Girona, Spain
2007-2009
n = 3,836
LUR model
Annual avg NO2
Primary model inputs were air
sampler height and traffic-related
variables. NO2 concentrations
estimated at participant's geocoded
address.
Cross-validation R2 = 0.61
Median: 14.1, IQR: 6.22
Systolic BP change (mm Hg)
Participants taking BP-lowering
medications: 2.24 (-2.58, 7.06)
Participants not taking BP-lowering
medications: 2.52 (0.26, 4.80)
Covariates: age, age squared, sex,
living alone, education, diabetes,
BMI, smoking, alcohol consumption,
deprivation, daily NO2, temperature,
nighttime railway and traffic noise.
Copollutant adjustment: none
tS0rensen et
al. (2012)
Diet, Cancer, and
Health cohort
Copenhagen and
Aarhus, Denmark
2000-2002
n= 45,271
LUR and dispersion model combined
Annual and 5-yr avg NOx
Model sums local air pollution from
street traffic based on traffic location
and density. NOx, NO2, and NO
concentrations predicted at each
participant's residence.
1-yr avg (ug/m3)
Baseline: Median 20.2, IQR 72.5
Follow-up: Median 20.0, IQR 71.1
5-yr avg
Baseline: Median 19.6, IQR 73.2
Follow-up: Median 19.3, IQR: 71.4
Systolic BP change (mm Hg)
per doubling NOx
1-yr period: -0.53 (-0.88, -0.19)
5-yr period: -0.84 (-0.84, -0.16)
OR for hypertension
5-yr period: 0.96(0.91,1.00)
Covariates: traffic noise, short term
NOx, temperature, relative humidity,
season, age, sex, calendar yr,
center of enrollment, length of
school attendance, BMI, smoking
status, alcohol intake, intake of fruit
and vegetables, sport during leisure
time.
Copollutant adjustment: none
6-85
-------
Table 6-8 (Continued): Epidemiologic studies of long-term exposure to oxides of
nitrogen and hypertension and blood pressure.
Study
Cohort, Location,
Study Period
Exposure Assessment and
Concentrations (ppb)
Effect Estimates (95% Cl)a
tDonq et al.
(2013b)
33 communities in
11 districts of
3 cities in Liaoning
Province, China
2006-2008
n = 24,845
Central site monitor
Communities were 1 km of a monitor
(selected to maximize intra- and
inter-city gradients).
District-specific 3-yr avg NO2
Mean: 18.7, Median: 17.5, IQR: 4.8
BP change (mm Hg)
Diastolic: 0.46 (-0.10, 1.03)
Systolic: 0.48 (-0.44,1.42)
Estimated change in the prevalence
of hypertension
OR: 1.20(1.00, 1.46)
Covariate adjustment: smoking
status, duration, intensity; ETS; sex;
BMI; education; sports activity;
alcohol, fruit, fat consumption;
hypertension; hypercholesterolemia.
Copollutant adjustment: none
Studies of Children
tLiuetal.
(2014b)
GINIplus and
LISAplus cohorts
Germany
Oct 2008-Nov 2009
n= 2,368
LUR model
Annual avg NO2
NO2 measurements taken during
three 2-week periods (warm, cold,
and intermediate seasons) at
40 sites. Exposure estimated for
each participant's residence.
Cross-validation R2: 0.67 for Munich
Mean: 12.4, Median: 18.8, IQR: 3.4
BP change (mm Hg)
Systolic: 0.32 (-1.32, 1.96)
Diastolic: -0.18 (-1.41, 1.05)
Covariate adjustment: cohort study,
area, sex, age, BMI, physical
activity, maternal smoking during
pregnancy, parental education,
parental history of hypertension,
7-day avg NO2, 7-day temperature.
Additionally adjusted for road-traffic
noise (n = 605)
Systolic: -0.70 (-3.43, 2.02)
Diastolic: -2.58 (-4.89, -0.23)
Copollutant adjustment: none
tBilenko et al.
(2015)
PIAMA birth cohort
the Netherlands
Feb2009-Feb2010
n = 1,432
LUR model
Annual avg NO2
NO2 measurements taken during
three 2-week periods (warm, cold,
and intermediate seasons) at
80 sites. Exposure estimated at each
participant's residence.
Median: 11.6, IQR: 4.1
BP change (mm Hg)
Diastolic: 0.80 (-0.43, 2.03)
Systolic: -0.07 (-1.69, 1.55)
Covariate adjustment: age, sex,
height, BMI, cuff size, gestational
age at birth, birthweight, weight gain
during first yr of life, breast feeding,
maternal smoking during pregnancy,
parental smoking in child's home,
physical activity, puberty
development scale, maternal
education, maternal hypertension
during pregnancy, pneumonia,
and/or otitis media during first 2 yr of
life, ambient temperature, room
temperature.
Copollutant adjustment: none
6-86
-------
Table 6-8 (Continued): Epidemiologic studies of long-term exposure to oxides of
nitrogen and hypertension and blood pressure.
Study
Cohort, Location,
Study Period
Exposure Assessment and
Concentrations (ppb)
Effect Estimates (95% Cl)a
tClarket al.
(2012)
RANCH study
U.K.
2001-2003
n = 719 children,
ages (9-10 yr) in 11
schools
LUR and dispersion model combined
Annual avg NO2
Concentrations estimated at
20 x 20 m resolution at each school.
No associations reported with
systolic or diastolic blood pressure.
Covariate adjustment: age, gender,
employment status, crowding, home
ownership, mother's educational
level, long-standing illness, language
spoken at home, parental support for
school work, classroom window
glazing type, and noise.
Copollutant adjustment: none
BMI = body mass index; BP = blood pressure; BWHS = Black Women's Health Study; Cl = confidence interval;
ETS = environmental tobacco smoke; Feb = February; GINIplus = German Infant Nutritional Intervention plus environmental and
genetic influences; IQR = interquartile range; IRR = incidence rate ratios; LISAplus = Lifestyle-Related factors on the Immune
System and the Development of Allergies in Childhood plus the influence of traffic emissions and genetics; NO = nitric oxide;
NO2 = nitrogen dioxide; NOX = sum of NO and NO2; RANCH = Road Traffic and Aircraft Noise Exposure and Children's Cognition
and Health; LUR = land use regression model; OR = odds ratio; PIAMA = Prevention and Incidence of Asthma and Mite Allergy;
PM2 5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; REGICOR = Registre Gironi
del Cor; SES = socioeconomic status.
aEffect estimates are reported per 10-ppb increase in NO2 or NO and 20-ppb increase in NOX unless otherwise specified. NOX
results that are originally reported in (xg/m3 are not standardized if the molecular weight needed to convert to ppb is not reported.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
Additional studies examined the association of NO2 with blood pressure in children. Liu
et al. (2014b) reported an association with increased diastolic blood pressure that was
diminished after adjustment for traffic-related noise exposure. Bilenko et al. (2015).
however, reported an association between NCh and diastolic blood pressure that was
robust to adjustment for noise. A study designed to evaluate the effect of aircraft noise on
cognition among school children (9-10 years old) reported no association between NC>2
and blood pressure after adjustment for noise (Clark et al.. 2012). All studies used LUR
models to assign NO2 exposure at participants' residential addresses, but only Liu et al.
(2014b) reported on model performance (cross-validation R2 = 0.67).
Overall, findings from studies of adults and studies of children report weak, inconsistent
results for the association between NO2 and hypertension and increased blood pressure,
although one prospective study using LUR to estimate exposure and adjusting for
cardiovascular disease risk factors reported an association of NOx with hypertension
(Coogan etal.. 2012). Uncertainties remain regarding the independent effect of NO2 on
hypertension and blood pressure, specifically whether confounding by correlated
traffic-related pollutants or noise can explain the positive associations observed.
6-87
-------
6.3.5 Cardiovascular Mortality
Results of studies of long-term exposure to NCh and cardiovascular diseases are coherent
with findings reporting associations of long-term NC>2 exposure with total and
cardiovascular mortality. Positive associations with total mortality, as well as deaths due
to cardiovascular disease have been observed in cohort studies conducted in the U.S. and
Europe (Figure 6-9. Table 6-16. and Section 6.5.2). Specifically, the strongest evidence
comes from a number of recent studies that have observed positive associations between
exposure to NCh and NOx and IHD mortality (Cesaroni et al.. 2013; Chen et al.. 2013a:
Lipsett et al.. 2011; Yorifuji et al.. 2010). mortality due to coronary heart disease (Gan et
al.. 2011; Rosenlund et al.. 2008b). and circulatory mortality (Yorifuji et al.. 2010; Jerrett
et al.. 2009). Coherence for the effect of long-term exposure and cardiovascular effects is
also provided by the evidence from studies of short-term cardiovascular mortality and
morbidity (Section 5.3).
6.3.6 Markers of Cardiovascular Disease or Mortality
Some recent epidemiologic and toxicological studies (Table 6-9) have investigated the
effects of long-term NO2 exposure on risk factors and markers of cardiovascular disease
risk or mortality, such as arterial stiffness, subclinical atherosclerosis, circulating lipids,
and heart rate variability (HRV). Previous information was limited. The 1993 AQCD for
Oxides of Nitrogen (U.S. EPA. 1993a) reported a significant reduction in heart rate in
rats exposed to 1,200 and 4,000 ppb NO2 for 1 month but not after lower concentrations
or longer durations of exposure (Suzuki etal.. 1981). There were no changes in vagal
responses in rats exposed to 400 ppb NO2 for 4 weeks (Tsubone and Suzuki. 1984).
Several recent cross-sectional analyses of long-term exposure to NCh evaluated vascular
markers of cardiovascular disease. Rivera etal. (2013) estimated NO2 concentrations with
LUR models whose performance in predicting measured concentrations varied among
study areas (Cross-validation R2 = 0.32-0.61). Increases in carotid intima-media
thickness (cIMT) observed in crude models were attenuated in fully adjusted models
while a positive association between NO2 and high ankle brachial index (ABI >1.3)
remained. A study of vascular damage in healthy young adults also used LUR to estimate
NO2 exposure but did not report information on model validation (Lenters etal.. 2010).
Increases in pulse wave velocity and augmentation index, but not cIMT, were observed in
association with NC>2 exposure. In a study that combined central site concentrations by
IDW to estimate residential exposure to NO2 during childhood lifestages, no association
was observed with cIMT among young adults (Breton et al.. 2012).
6-88
-------
The effects of NO2 in relation to autonomic function in a random selection of Swiss
cohort study participants were examined by Felber Dietrich et al. (2008). Measures of
HRV were linked to annual NC>2 concentration at each participant's residential address
estimated by dispersion model predictions supplemented with land use and
meteorological data. The model demonstrated good predictability of outdoor home NO2
concentrations, which were measured for some study participants (R2 = 0.77-0.86 across
areas). Annual average NO2 concentration was associated with decreased standard
deviation of beat-to-beat (NN) intervals, an index of total HRV, nighttime low-frequency
component of HRV (LF), and LF/high frequency (HF) component of HRV ratio in
women. No associations with other parameters of HRV were observed in these data.
A recent toxicological study by Seilkop et al. (2012) reported changes in markers that are
characteristic of vascular disease development and progression. Mice were exposed for
50 days to various multipollutant atmospheres (diesel or gasoline exhaust, wood smoke,
or simulated "downwind" coal emissions) comprising varying concentrations of NC>2
(0-3,670 ppb) and other pollutants. A data mining technique known as multiple additive
regression trees analysis was employed to identify associations between the 45 different
exposure component categories, including NC>2, and various indicators of cardiovascular
disease stability and progression [e.g., endothelin-1 (ET-1), matrix metalloproteinase
(MMP)-3, MMP-7, MMP-9, tissue inhibitor of metalloproteinase-2 (TIMP-2)]. The
results demonstrated that NO2 was one of the strongest predictors of responses. More
specifically, NO2 ranked among the top three predictors for ET-1 and TIMP-2; however,
the study design did not allow for the independent effects of NCh to be evaluated.
In another study, Takano et al. (2004) reported that obese rats (Otsuka Long-Evans
Tokushima Fatty) had elevated levels of triglycerides and decreased high-density
lipoprotein (HDL) and HDL/total cholesterol levels after long-term exposure to 160 ppb
NC>2 compared to control rats breathing clean air. HDL levels were also decreased after
800 ppb NO2 exposure in the obese strain and in the nonobese rats (Long-Evans
Tokushima). The authors suggested that obese animals were at greater risk of
dyslipidemia following NC>2 exposure.
Overall, a limited number of epidemiologic and toxicological studies have evaluated
long-term NC>2 exposure on markers of cardiovascular disease or mortality. There is some
evidence for increased arterial stiffness, increased markers for cardiovascular disease
stability and progression, dyslipidemia, decreased HRV, and reduced HR; however, these
effects have only been reported in one study each. Findings from several studies of cIMT
are inconsistent. Further, an independent effect of NCh is not clearly distinguished in the
available body of epidemiologic and toxicological evidence.
6-89
-------
Table 6-9 Characteristics of toxicological studies of long-term nitrogen dioxide
exposure and cardiovascular effects.
Study
Species (Strain);
Sample size;
Sex; Age
Exposure Details
Endpoints Examined
de Burbure et Rats (Wistar);
al. (2007) n = 8/group; M
8 weeks
High (6 pg/day) or low (1.3 ug/day)
selenium;
(1)1,000ppb, 28 days, 6 h/day,
5 days/week (Se+/Se~);
(2) 5,000 ppb, 5 days, 6 h/day
GPx in plasma and RBC lysate; SOD
activity in RBC lysate; GST activity in RBC
lysate; TEARS in plasma. Endpoints
examined immediately and 48 h after
exposure.
Renters et al.
(1973)
Furiosi et al.
(1973)
tSeilkop et
al. (2012)
Suzuki et al.
(1981)
Takano et al.
(2004)
Tsubone and
Suzuki (1984)
Wagner et al.
(1965)
Squirrel monkeys;
adult; n = 4; M;
age NR
Monkeys (Macaca
spec/osa); n = 4;
M/F; adult
Rats
(Sprague-Dawley);
n = 8; M; 4 weeks
Mice (ApoE-/-);
n = 8-10; M;
10 weeks
Rats (strain NR);
n = 6; sex and age
NR
Rats (OLETF and
LETO); n = 10-14;
M; 4 weeks
Rats (Wistar);
n = 6; M;
9-13 weeks
Dogs;
n = 6-10/group;
M; adult
1 ,000 ppb NO2 continuously for
16 mo; challenged with influenza
virus
(1 ) 2,000 ppb NO2 continuously for
14 mo
260, 745, and 3,670 ppb (along
with dilutions of 1/3 and 1/10) NO2
6 h/day, 7 days/week for 50 days;
co-exposure with 700 other
components; fed a high-fat diet
400, 1,200, and 4,000 ppb NO2;
1, 2, and 3 mo
160, 800, or 4,000 ppb NO2
continuously for 32 weeks
400 and 4,000 ppb NO2
continuously for 1 and 4 weeks,
respectively. Immediately after
exposure, animals injected with
5 ug/kg BW phenyl diguanide
1,000 or 5, 000 ppb NO2
continuously for 18 mo
Hemoglobin and hematocrit levels
measured throughout the study.
Erythrocyte, hematocrit, and hemoglobin
levels measured throughout the study.
ET-1, VEGF, MMP-3, MMP-7, MMP-9,
TIMP-2, HO-1, TEARS in proximal aorta
18-h after exposure.
Heart rate and hemoglobin levels
measured after 1, 2, and 3 mo exposures.
BW, triglyceride, HDL, total cholesterol,
HDL/total cholesterol, and sugar measured
8 weeks after exposure.
Heart rate measured 10 sec after injection.
Hemoglobin and hematocrit levels
measured quarterly throughout exposure.
BW = body weight; ET-1 = endothelin-1; GPx = glutathione peroxidase; GST = glutathione s-transferase; HDL = high density
lipoprotein; HO-1 = heme oxygenase-1; LETO = Long-Evans Tokushima; MMP = matrix metalloproteinase; NO2 = nitrogen
dioxide; NR = not reported; OLETF = Otsuka Long-Evans Tokushima Fatty; RBC = red blood cell; SOD = superoxide dismutase;
TEARS = thiobarbituric acid reactive substances; TIMP-2 = tissue inhibitor of metalloproteinase-2; VEGF = vascular endothelial
growth factor.
fStudy published since the 2008 ISA for Oxides of Nitrogen.
6-90
-------
6.3.7 Diabetes
There were no epidemiologic studies examining the association of NO2 exposure with
diabetes or insulin deficiency in the 2008 ISA for Oxides of Nitrogen. Recent large
prospective studies using exposure assessment methods designed to achieve high spatial
resolution, provide some evidence of an association (Table 6-10). However, studies
overall have not distinguished an independent effect of NO2 on diabetes.
Cooganetal. (2012) examined the association of long-term NOx exposure with incident
diabetes among black women residing in Los Angeles, CA. An LUR model was applied
to estimate exposure at each participant's residential address, and a cross-validation R2 of
0.92 indicated that the model predicted well the pattern of NOx concentrations in the
study area. An increased risk of 1.43 (95% CI: 1.12, 1.84) per 20-ppb increase in NOx
was observed after adjustment for a wide array of potential confounders including
traffic-related noise exposure. PM2 5 and NOx concentrations were poorly correlated
(r = 0.27), and negligible attenuation in the NOx effect estimate was reported after
adjustment for PM25. However, PM2 5 concentrations were estimated using kriging of
central site monitor measurements and may be subject to differential exposure
measurement error. Further, correlations between NOx and NO2 or other traffic-related
copollutants were not reported. An increased risk of Type II diabetes in association with
LUR estimates of NO2 was reported among older adult women living in the Ruhr district
of West Germany (HR: 1.55 [95% CI: 1.20, 1.99] per 10-ppb increase inNO2; (Kramer et
al.. 2010)). In this study, nondiabetic women (age 54-55) were followed over 16 years
(1990-2006), and alternate NO2 exposure assessment methods (mean monitor
concentration and emission inventory-based methods) were compared. Relative risks
determined by these alternative methods were smaller and less precise compared to those
obtained using LUR models, which showed good predictability of NO2 concentrations
measured at central site monitors (r = 0.66). Although diabetes status was self-reported in
this study, a validation study comparing self-reported diabetes from the questionnaire to
answers obtained during a clinical exam interview indicated 99% concordance. In an
analysis of a subgroup (n = 363) of these women, Teichert et al. (2013) observed positive
associations of NO2 and NOx exposure (estimated for the period 10-20 years prior to the
baseline exam) with impaired glucose metabolism (IGM). Risk estimates were robust to
adjustment for an array of biomarkers of subclinical inflammation.
6-91
-------
Table 6-10 Epidemiologic studies of long-term exposure to oxides of nitrogen
and diabetes or diabetes-related effects.
Study
Cohort, Location,
Study Period
Exposure Assessment and
Concentrations (ppb)
Effect Estimates (95% Cl)a
tCooqan et
al. (2012)
BWHS cohort
Los Angeles, CA
1995-2005
n = 3,236
LUR model
Annual avg NOx
Model incorporated traffic, land use,
population, and physical geography to
estimate exposure at the participants'
residences. Summer and winter field
measurements taken at 183 sites in Los
Angeles, CA.
Cross-validation with 16 sites, R2 = 0.92
Mean: 43.3, Median: 41.6, IQR: 12.4
Diabetes
IRR: 1.20(1.06, 1.36)
Covariate adjustment: age, BMI,
yr of education, income, number
of people in the household,
smoking, drinks per week,
h/week of physical activity,
neighborhood SES, family history
of diabetes.
IRR: 1.19(1.04, 1.35) adjusted
for above covariates plus PIVh.s
tKramer et
al. (2010)
SALIA cohort
Ruhr district, West
Germany
1990-2006
n = 1,755
Older adult women,
ages 54-55 yr at
enrollment
LUR model, nearest central site monitor,
emissions inventory
Annual avg for LUR and emissions inventory,
5-yr avg (1986-1990) for central site
LUR model estimated NO2 at participant's
residence (r = 0.66 for correlation with
concentrations at central site monitors).
Emission inventory estimated traffic-related
NO2 at 1 x 1 km grid. Central site monitors
covered 8 x 8 km grid to capture broad scale
variability in NO2.
LUR: Median, 18.3, IQR: 8.0
Central site monitor: Median: 22.2, IQR: 13.2
Emissions inventory: Median: 6.4, IQR: 10.1
Diabetes
LURHR: 1.55(1.20, 1.99)
Central site monitor HR:
1.25(1.02, 1.53)
Emission inventory HR:
1.15(1.04,1.27)
Covariate adjustment: age, BMI,
heating with fossil fuels,
workplace exposure to
dust/fumes, extreme
temperature, smoking, and
education.
Copollutant adjustment: none
tTeichert et
al. (2013)
SALIA cohort
Ruhr district, West
Germany
2003-2009
Subgroup, n = 363
Older adult women
ages 54-55 yr at
enrollment
LUR for NO2 and NOx, nearest central site
monitor for NO2
Annual avg NO2 and 10 to 20-yr avg NO2
and NOx for LUR, 5-yr avg (2003-2007) for
central site NO2
LUR model estimated NO2 at participant's
residences (40 sites, cross-validation R2 =
0.84). 10 to 20-yr avg NO2 and NOx prior to
disease estimated by back-extrapolation.
LUR estimates multiplied by ratio of
concentration at baseline to follow-up period.
Central site monitors covered 8 * 8 km grid to
capture broad scale variability in NO2.
NO2
Group with IGM: Mean: 21.1, SD: 5.8
Group without IGM: Mean: 20.1, SD: 4.1
NOx (pg/m3)
Group with IGM: Mean: 74.1, SD: 31.2
Group without IGM: Mean: 69.3, SD: 30.0
IGM for LUR exposure estimates
NO2OR: 1.63(1.06,2.51)
NOx OR: 1.41 (1.01,1.97)
per 43.2 ug/m3 NOx
Covariate adjustment: age, BMI,
smoking status, passive
smoking, education, exposure to
indoor mold, and season of blood
sampling.
Copollutant adjustment: none
6-92
-------
Table 6-10 (Continued): Epidemiologic studies of long-term exposure to oxides of
nitrogen and diabetes or diabetes-related effects.
Study
Cohort, Location,
Study Period
Exposure Assessment and
Concentrations (ppb)
Effect Estimates (95% Cl)a
tEze et al.
(2014)
SAPALDIA cohort
Switzerland
1991-2002
n= 6,372
LUR model and dispersion model combined
10-yr avg NO2
Dispersion model (200 * 200 m resolution)
incorporated LUR to reduce underestimation
at background sites. Annual NO2 combined
with residential history to estimate 10-yr avg
exposure. Ratio of model predictions and
measurements at central sites < 0.04.
Mean: 15.0, IQR: 6.1
Diabetes prevalence
OR: 1.43(1.08,1.88)
Covariate adjustments: age, sex,
educational level, neighborhood
SEI, lifestyle, BMI, noise,
hypertension, hs-CRP, and
dyslipidemia.
OR: 1.08(0.72, 1.62) adjusted
for above covariates plus PM-io
tBrooket
al. (2008)
Patients who
attended 2
respiratory clinics
Hamilton and
Toronto, Canada
2002-2004
n = 7,634
LUR model
NO2 measured at -250 sites selected using a
location-allocation model, and concentrations
estimated at participants' residences.
Difference between predicted and measured
NO2: 1-7% for Hamilton, < 4% for Toronto
Hamilton
Female: Median: 15.3, IQR: 3
Male: Median: 15.2, IQR: 3.2
Toronto
Female: Median: 22.9, IQR: 3.9
Male: Median: 23, IQR: 20.8
Diabetes mellitus
Female OR: 1.48(1.00,2.16)
Male OR: 0.90(0.60,1.34)
Both sexes combined OR:
1.16(0.82, 1.61)
Covariate adjustment: age, BMI,
and neighborhood income.
Copollutant adjustment: none
tDiikema et
al. (2011b)
Westfriesland, the
Netherlands
(semirural)
2007
n = 8,018
LUR model
NO2 concentration estimate at residential
address at the time of recruitment.
Cross-validation R2 = 0.82
Ranges for quartiles
Q1 4.7-7.5, Q2: 7.5-8.1, Q3: 8.1-8.8,
Q4: 8.8-19.1
Type II diabetes prevalence
Q1 OR: 1.00 (reference)
Q2OR: 1.03(0.82, 1.31)
Q3OR: 1.25(0.99, 1.56)
Q4OR: 0.8(0.63, 1.02)
Covariate adjustment: average
monthly income, age, sex.
tAndersen
etal.
Diet, Cancer, and
Health cohort
Copenhagen and
Aarhus, Denmark
Jan 1995-Jun 2006
n = 57,053
Dispersion model
Model sums local air pollution from street
traffic based on traffic location and density.
AirGIS used to predict the NO2
concentrations at each participant's
residence, r = 0.77 between predicted and
measured 6-mo avg NO2.
1971-end of follow-up: Median: 7.7, IQR: 2.6
1991-end of follow-up: Median: 8.1, IQR: 3.0
Diabetes from 1971-end of
follow-up
All diabetes HR: 1.00(0.89,1.12)
Confirmed diabetes HR:
1.16(1.00, 1.34)
Diabetes from 1991-end of
follow-up
All diabetes HR: 1.04(0.93, 1.16)
Confirmed diabetes HR:
1.16(1.04, 1.30)
Covariate adjustment: sex, BMI,
waist-to-hip ratio, smoking status,
smoking duration, smoking
intensity, environmental tobacco
smoke, educational level,
physical/sports activity in leisure
time, alcohol, fruit, fat
consumption, calendar yr.
Copollutant adjustment: none
Confirmed diabetes defined as
exclusion of cases identified
solely by glucose blood test.
6-93
-------
Table 6-10 (Continued): Epidemiologic studies of long-term exposure to oxides of
nitrogen and diabetes or diabetes-related effects.
Cohort, Location, Exposure Assessment and
Study Study Period Concentrations (ppb) Effect Estimates (95% Cl)a
tThierinq et GINIplus and LUR model Insulin resistance percentage
al. (2013) LISAplus cohorts Annual average NO2 difference: 29.77 (6.85, 57.43)
Germany NO2 measurements taken during three Covariate adjustment: sex, age,
Oct 2008-Nov 2-week periods (warm, cold, and intermediate BMI, birth weight, study center,
2009 seasons) at 40 sites, and concentrations parental education, study, study
n = 397 estimated at each participant's residence. design, puberty status, exposure
Cross-validation R2: 0.67 for Munich to smoke.
Mean: 11.5, SD: 2.8 Copollutant adjustment: none
BMI = body mass index; BWHS = Black Women's Health Study; Cl = confidence interval; GINIplus = German Infant Nutritional
Intervention plus environmental and genetic influences; HOMA = homeostatic model assessment; HR = hazard ratio;
hs-CRP = high sensitivity C-reactive protein; IGM = impaired glucose metabolism; IQR = interquartile range; IRR = incidence rate
ratios; LISAplus = Lifestyle-Related Factors on the Immune System and the Development of Allergies in Childhood plus the
influence of traffic emissions and genetics; LUR = land use regression model; NO2 = nitrogen dioxide; NOX = sum of NO and NO2;
OR = odds ratio; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm;
PMio = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; Q1 = 1st quartile; Q2 = 2nd
quartile; Q3 = 3rd quartile; Q4 = 4th quartile; SALIA = Study on the Influence of Air Pollution on Lung, Inflammation, and Aging;
SAPALDIA = Swiss study on Air Pollution and Lung Disease in adults; SD = standard deviation; SEI = socioeconomic index; SES
= socioeconomic status.
aEffect estimates are reported per 10-ppb increase in NO2 or NO and 20-ppb increase in NOX unless otherwise specified. NOX
results that are originally reported in ng/m3 are not standardized if the molecular weight needed to convert to ppb is not reported.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
In a large study of prevalent diabetes ascertained by self-report among randomly selected
adults (ages 18-65 years), which also used LUR, Eze et al. (2014) reported a positive
association (OR: 1.43 [95% CI: 1.08, 1.88] per 10 ppb increase in NO2). Noise was
included among the array of potential confounders for which the final model was
adjusted. The association of NO2 with diabetes was attenuated after adjustment for PMio
in a copollutant model (OR: 1.08 [95% CI: 0.72, 1.62] per 10-ppb increase in NO2). In
another study of diabetes prevalence using LUR, Brook et al. (2008) reported an
increased risk in the prevalence of diabetes mellitus (Types I and II) of 1.48 (95% CI:
1.00, 2.16) per 10-ppb increase in NO2 among female respiratory patients in two
Canadian cities. NO2 exposure was not associated with diabetes in male patients,
however. Prevalent diabetes was not associated with NO2 exposure estimated by LUR in
a semirural population in the Netherlands (Dijkema et al.. 201 Ib). All of these studies
demonstrated that LUR models well predicted measured NO2 concentrations in the study
areas.
A large prospective study examined the association of NO2 exposure with diabetes
incidence among participants of the Danish Diet, Cancer, and Health Cohort (Andersen et
al.. 2012c). A validated dispersion model [r > 0.75 between measured and predicted 6-mo
avg NO2 concentration; (Hertel etal.. 2014)] was used to assign mean NO2 concentration
since 1971 based on residential address history. No association between NO2 exposure
and diabetes was observed in fully adjusted models (HR: 1.00 [95% CI: 0.89, 1.12] per
6-94
-------
10-ppb increase in NCh); however, after restricting the analyses to confirmed cases of
diabetes, a positive association was observed (HR: 1.16 [95% CI: 1.04, 1.30] per 10-ppb
increase in NCh). Long-term exposure to traffic noise was not associated with a higher
risk of diabetes in this population (S0rensen et al.. 2013). Another study designed to
evaluate the association of long-term exposure to aircraft noise with diabetes found that,
although associations with metabolic outcomes such as waist circumference were
observed, no association of Type II diabetes or BMI with noise was present. (Eriksson et
al.. 2014).
A study in children examined the association of NC>2 concentration with insulin
resistance, which plays a key role in the development of Type II diabetes mellitus.
Among 10-year-old children (n = 397), Thiering et al. (2013) reported that a 10-ppb
increase in NC>2 was associated with a 30% increase (95% CI: 6.9, 57.4) in the
homeostatic model assessment (HOMA) of insulin resistance, a metric derived from
blood glucose and serum insulin measurements. Additionally, two recent studies reported
contrasting findings regarding the associations between short-term NC>2 exposure and
measures of insulin resistance. In contrast with results from Thiering et al. (2013).
Kelishadi et al. (2009) reported a lack of an association between 24-h avg NO2 and
insulin resistance in a study of 374 Iranian children aged 10-18 years. Coherent with the
results of studies examining long-term exposure and diabetes in adults, a panel study of
older adults in Korea observed a 1.33 uU/mL (95% CI: 0.54, 2.11) increase in insulin
resistance and a 0.52 mean (95% CI: 0.24, 0.77) increase in the HOMA of insulin
resistance [fasting insulin x (fasting glucose + 22.5)] per 20-ppb increase in 24-h avg
NO2 at lag 7 (Kim and Hong. 2012). The association was stronger in participants with a
history of diabetes mellitus but still present for those without. Both of the short-term
exposure studies relied on central site monitoring for exposure estimation, and neither
evaluated potential confounding by other traffic-related pollutants.
Generally consistent associations of NO2 (Teichert et al.. 2013; Andersen et al.. 2012c:
Kramer etal.. 2010) as well as NOx (Coogan et al.. 2012) with diabetes or impaired
insulin metabolism are reported in prospective studies using LURto assign exposure.
Associations of NO2 with prevalent diabetes among females and respiratory patients are
reported in some (Eze etal.. 2014; Brook et al.. 2008) but not all studies (Dijkema et al..
201 Ib). All of studies estimating exposure with LUR models indicated that the modeled
estimates agreed well with measured concentrations in the study areas (Table 6-10).
Findings regarding the potential for noise exposure to confound observed associations of
NO2 and NOx with diabetes are limited, as is coherence with short-term exposure studies
of insulin resistance. Overall, studies have not distinguished an independent effect of NO2
from other traffic-related exposures on diabetes.
6-95
-------
6.3.8 Subclinical Effects Underlying Cardiovascular Disease and Diabetes
Inflammation and oxidative stress have been shown to play a role in the progression of
chronic heart disease and diabetes. Although studies of inflammation and oxidative stress
were not generally available for review in the 2008 ISA for Oxides of Nitrogen (U.S.
EPA. 2008c). a number of null findings related to changes in hematological parameters
were reported. Hematocrit and hemoglobin levels were unchanged in squirrel monkeys
(Fenters et al.. 1973). rats (Suzuki et al.. 1981). or dogs exposed to <5,000 ppb NO2
(Wagner etal.. 1965). However, Furiosi et al. (1973) reported polycythemia due to
reduced mean corpuscular volume and an increased trend in the ratio of neutrophil to
lymphocytes in the blood of NC>2-exposed monkeys and similar increases in erythrocyte
counts in NCh-exposed rats.
A limited number of studies published since the 2008 ISA for Oxides of Nitrogen have
evaluated markers of inflammation and oxidative stress. Forbes et al. (2009a) examined
the association of annual average NO2 concentrations with C-reactive protein (CRP) and
fibrinogen among a population in England. Multilevel linear regression models were used
to determine pooled estimates across three cross-sectional surveys conducted during
different years. Each participant's postal code of residence was linked to a predicted
annual average NO2 concentration derived from dispersion models. NO2 was not
associated with increased CRP or fibrinogen in these data nor were PMio, SO2, or Os. A
study conducted among men and women (ages 45-70 years) in Stockholm, Sweden
reported an association of 30-year avg traffic-related NO2 concentrations estimated using
dispersion models with increases in interleukin-6 (IL-6) and CRP but not with TNF-a,
fibrinogen, or PAI-1 (Panasevich et al.. 2009). Associations between several metrics of
SO2 exposure and increased IL-6 and CRP were observed in this study. In another
analysis from this study, long-term exposure to NO2 interacted with IL-6 and TNF
polymorphisms on an additive scale with regard to increased MI risk (Panasevich et al..
2013). In a study of COPD patients, annual average NO2 concentration was associated
with increases in interleukin-8 (IL-8) but not with the other markers studied including
CRP, TNF-a, IL-6, fibrinogen, and hepatocyte growth factor (Dadvand et al.. 2014b).
de Burbure et al. (2007) examined oxidative stress markers in rats on a low selenium
(Se-L) or supplemented selenium (Se-S) diet exposed to 1,000 ppb NO2 for 28 days.
Blood Se levels decreased significantly in both groups immediately after the 28-day
exposure and continued to decrease in the Se-S diet rats following a 48-hour recovery
period. GPx, of which Se is an integral component, also decreased immediately and
48 hours after exposure only in the plasma of Se-S diet rats. However, GPx levels
increased in red blood cells (RBC) of Se-L diet rats immediately after the 28-day
exposure and increased in both groups 48 hours later. RBC SOD activity increased in
6-96
-------
both groups immediately after the exposure and decreased in Se-L diet rats 48 hours later.
GST was increased for both groups immediately after the 28-day exposure and continued
to increase after the 48-hour recovery period, potentially compensating for the transient
increase in thiobarbituric acid reactive substances (TEARS) immediately after exposure.
As discussed in Section 6.3.6. Seilkop et al. (2012) examined the effects of NO2
exposure, in a multipollutant context, on markers of oxidative stress [heme oxygenase-1
(HO-1) expression and TBARs, indicator of lipid peroxidation] in ApoE-/- mice fed a
high-fat diet. Mice were exposed to various atmospheres (diesel or gasoline exhaust,
wood smoke, or simulated "downwind" coal emissions) with varying concentrations of
NC>2 (0-745 ppb) for 50 days. Associations between the oxidative stress indicators and
the 45 different exposure component categories were determined using a data mining
technique known as multiple additive regression trees analysis. The results demonstrated
NC>2 was among one of the strongest predictors of response for TBARS but not HO-1.
Overall, a limited number of epidemiologic and toxicological studies have evaluated
long-term NO2 exposure on inflammation and oxidative stress with some, but not all,
studies reporting positive associations. In general, findings of the epidemiologic studies
do not separate the effect of NO2 from copollutants, and the few animal toxicological
studies do not clearly demonstrate an effect of NO2 exposure.
6.3.9 Summary and Causal Determination
Overall, the evidence is suggestive of, but not sufficient to infer, a causal relationship
between long-term exposure to NO2 and cardiovascular health effects and diabetes. This
conclusion is based heavily on recent epidemiologic studies reporting associations of
NO2 with heart disease and diabetes. While well-conducted studies of NOx are also
available, these studies are less informative regarding the independent effect of NO2
exposure on cardiovascular effects and diabetes. The current conclusion represents a
change from the conclusion drawn in the 2008 ISA for Oxides of Nitrogen, which stated
that the evidence was inadequate to infer the presence or absence of a causal relationship
(U.S. EPA. 2008c). Previously, Miller et al. (2007) reported a positive association
between long-term NO2 exposure and cardiovascular events among post-menopausal
women but did not distinguish an independent effect of NO2. Toxicological evidence
reported no changes in hematocrit or hemoglobin and increased erythrocyte count and did
not demonstrate that long-term NO2 exposure induces cardiovascular effects. In contrast,
although not with entire consistency, several recent studies report positive associations of
NO2 exposure with heart disease and diabetes. As in the 2008 ISA, uncertainty remains as
to whether NO2 exposure has an independent effect of other traffic-related pollutants and
6-97
-------
noise for both cardiovascular effects and diabetes. Evidence from experimental studies is
limited and inconsistent and does not clearly support biological plausibility. The key
evidence as it relates to the causal determination is detailed in Table 6-11 using the
framework described in Table II of the Preamble to the ISA.
Evidence from recent, large and well-conducted prospective epidemiologic studies
generally supports an association of long-term exposure to NCh with heart disease.
Studies were prospective in design and did not report evidence that findings were likely
to be biased by selective participation or missing data. Additionally, these studies
adjusted for a wide array of cardiovascular risk factors and used exposure assessment
methods that captured the fine scale variability in NO2 concentration. The evidence for
the association of long-term NCh exposure with stroke and hypertension is less
consistent. The strongest evidence for heart disease comes from a large multicohort
prospective study (Cesaroni et al.. 2014) with supporting evidence from cross-sectional
study of CHD hospital admissions (Beckerman et al., 2012). Both of these studies
demonstrated LUR modeled estimates to agree well with measured NO2 concentrations in
the study areas. Consistent findings from multiple epidemiologic studies of
cardiovascular mortality support these morbidity findings (Section 6.5.2). Studies using
dispersion models or IDW for exposure assessments were less consistent, with some
reporting positive associations (Lipsett et al., 2011) and others reporting null associations
with heart disease (Atkinson et al.. 2013; Rosenlund et al.. 2009a). IDW and dispersion
modeling may not adequately capture the spatial variability in NO2 concentrations
(Section 3.2.3) and produce biased exposure estimates and health effect estimates.
However, in the null studies, dispersion model estimates were shown to represent well
the spatial pattern of measured NO2 concentrations in the study areas. Epidemiologic
studies have not adequately accounted for confounding by PIVb 5, noise, or traffic-related
copollutants, and there is limited coherence and biological plausibility for NO2-related
development of heart disease. Epidemiologic evidence does not consistently link
long-term NC>2 exposures to markers of cardiovascular risk [e.g., increased high ABI
(Rivera et al., 2013). arterial stiffness (Lenters et al.. 2010). cIMT, and markers of
inflammation [CRP and IL-6; (Panasevich et al.. 2009)]. Experimental studies also do not
clearly demonstrate an independent effect of NO2 exposure on heart disease development.
Long-term NCh exposure induced dyslipidemia in rats (Takano et al.. 2004). but
short-term or long-term NCh exposure did not consistently increase inflammation or
oxidative stress in controlled human exposure or animal toxicological studies (Channell
etal.. 2012; Huang etal.. 2012b; Riedl etal. 2012; Li etal.. 201 la: de Burbure et al..
2007). Further, these are early events in the proposed mode of action and not specific to
development of heart disease.
6-98
-------
Evidence from recent, large and well-conducted prospective epidemiologic studies
generally supports the association of long-term exposure to NCh with diabetes. Several
large prospective studies also using LUR report increased risk of diabetes, impaired
glucose metabolism, and increased insulin resistance with NO2 exposure (Teichertet al..
2013; Thieringetal.. 2013; Kramer etal.. 2010). These studies indicated that LUR
modeled estimates agreed well with measured NO2 concentrations in the study areas.
Several studies reporting positive associations were prospective in design and did not
report evidence that findings were likely to be biased by selective participation or missing
data. As with heart disease, epidemiologic studies have not adequately accounted for
confounding by PM2 5, noise, or traffic-related copollutants, and there is limited
coherence and biological plausibility for NCh-related development of diabetes.
Long-term NCh exposure induced dyslipidemia in rats (Takano et al.. 2004). but
short-term or long-term NCh exposure did not consistently increase inflammation or
oxidative stress in controlled human exposure or animal toxicological studies (Channell
etal.. 2012; Huang etal.. 2012b; Riedl etal.. 2012; Li etal.. 201 la; de Burbure et al..
2007). These are early events and not specific to development of diabetes.
Several epidemiologic studies report positive associations of NO2 exposure with both
heart disease and diabetes, and the annual average or multi-year avg residential exposures
were typically considered surrogates for long-term exposure, and residential stability was
assumed (or sometimes required for eligibility). Confounding by correlated traffic-related
pollutants and noise remains an uncertainty, and experimental studies do not clearly
support the biological plausibility of associations observed in the epidemiologic studies.
Further, most studies did not disentangle the effects of long-term from short-term
exposure. These general limitations introduce some uncertainty with regard to the
specific patterns of exposure associated with the observed effects. Overall, the evidence
from some epidemiologic studies of heart disease and diabetes but uncertainty about an
independent effect of NCh is suggestive of, but not sufficient to infer, a causal
relationship between long-term NC>2 exposure and cardiovascular effects and diabetes.
6-99
-------
Table 6-11 Summary of evidence, which is suggestive of, but not sufficient to
infer, a causal relationship between long-term nitrogen dioxide
exposure and cardiovascular effects and diabetes.
Rationale for Causal
Determination3
Key Evidence13
Key References13
NO2 Concentrations
Associated with Effects0
Heart Disease
Evidence from
epidemiologic studies
generally supportive
but not entirely
consistent
Findings from a large multicohort
prospective study using
validated LUR provides evidence
that NO2 is associated with
coronary events.
tCesaroni et al. (2014)
Section 6.3.2
Range of mean annual avg
across cohorts: 4.2-31.9
ppb
Supporting evidence from
cross-sectional study of CHD
hospital admissions and study of
Ml incidence using LUR or IDW
to estimate NO2 exposure.
tBeckerman et al. (2012),
tLipsett et al. (2011)
Median: 22.9 ppb
Mean: 33.6 ppb
Inverse or null associations of
NO2 with CHD or Ml reported in
some studies.
tAtkinson et al. (2013).
tRosenlund et al.
(2009a), tGan et al.
(2011)
Annual avg mean: 12.0 ppb
5-yr avg: median: 6.9
(cases), 6.3 (controls) ppb,
mean: 16.3 ppb
Limited coherence with
evidence for effects on
cardiovascular disease
risk factors
Associations with some markers
of vascular damage observed in
some but not all epidemiologic
studies.
tLenters et al. (2010)
tRiveraetal. (2013)
Section 6.3.6
Means: 19.7 ppb for annual
avg, 11.0 ppb for 10-yr avg
Coherence with
evidence for
cardiovascular mortality
Strongest evidence of mortality
from IHD, CHD, and circulatory
diseases, including supporting
evidence from studies that report
weak or null associations with
cardiovascular morbidity.
tRosenlund et al.
(2009a): tGan et al.
(2011)
Section 6.5.2
5-yr avg: median: 6.9
(cases), 6.3 (controls) ppb,
mean: 16.3 ppb
Uncertainty regarding
potential confounding
by traffic-related
copollutants and noise
Overall, studies did not adjust for Section 6.3.2
PM2.5, BC, EC, or CO.
Confounding by noise examined
in one study.
Cerebrovascular Disease and Stroke
Inconsistent
epidemiologic evidence
Some studies of variable quality
report increased, typically
imprecise, risks of stroke and/or
cerebrovascular disease with
NO2. Inconsistency across
stroke subtype.
tJohnson etal. (2013),
tJohnson etal. (2010a).
tAndersen etal. (2012b),
tde Kluizenaar et al.
(2013), tDonq etal.
(2013a)
Section 6.3.3
Means for annual avg and
3- or 5-yr avg:
15.2-18.7 ppb
Median for annual avg:
8.1 ppb
Other studies reported weak or t Atkinson et al. (2013). Means for annual avg: 12.0,
null associations. tLipsett et al. (2011) 33.6
6-100
-------
Table 6-11 (Continued): Summary of evidence, which is suggestive of, but not
sufficient to infer, a causal relationship between long-
term nitrogen dioxide exposure and cardiovascular
effects and diabetes.
Rationale for Causal
Determination3
Key Evidence13
Key References13
NO2 Concentrations
Associated with Effects0
Uncertainty regarding
potential confounding
by traffic-related
copollutants and noise
Results not consistently robust
to adjustment for noise. No
studies examined confounding
by traffic pollutants.
tS0rensen et al. (2014),
tde Kluizenaar et al.
(2013)
Section 6.3.3
Hypertension
Inconsistent
epidemiologic evidence
Cross-sectional association with tForaster et al. (2014)
increased blood pressure but not Section 6.3.4
hypertension.
Median for annual avg: 14.1
Uncertainty regarding
potential confounding
by traffic-related
copollutants and noise
Studies did not adjust for PlVhs,
BC, EC, or CO. Few studies
adjusted for noise.
Section 6.3.4
Association observed with NOx
may not represent effect of NO2.
tCooqan etal. (2012)
Diabetes
Evidence from
epidemiologic studies
generally consistent
and supportive
Large prospective studies using
LUR report increased risk of
diabetes incidence, impaired
glucose metabolism, and
increased insulin resistance with
NO2 exposure.
tKramer et al. (2010),
tTeichert et al. (2013).
tThierinq et al. (2013)
Section 6.3.7
Median for annual avg: 18.3
ppb. Means for annual avg:
11.5-21 ppb
Supporting evidence that NO2
exposure is associated with
prevalent diabetes.
tEzeetal. (2014),
tBrook et al.
(2008)—females only
Means: 15.0 ppb for 10-yr
avg, 15, 23 ppb for annual
avg
Association observed among
confirmed cases of diabetes but
not overall.
tAndersen etal. (2012c)
Medians for multi-yr avg:
7.7, 8.1 ppb
Uncertainty regarding
potential confounding
by traffic-related
copollutants and noise
Consistent but limited evidence
that associations are robust to
adjustment for noise. Overall,
studies did not adjust for PM2.5,
BC, EC, CO.
tEzeetal. (2014):
tS0rensen etal. (2013):
tErikssonetal. (2014)
Section 6.3.7
Associations observed with NOx
may not represent effect of NO2.
tCooqan etal. (2012)
Limited toxicological
evidence for NO2
independent effect
Dyslipidemia—increased
triglycerides and decreased
HDL—in rats.
Takano et al. (2004)
Section 6.3.8
160 ppb for 32 weeks
6-101
-------
Table 6-11 (Continued): Summary of evidence, which is suggestive of, but not
sufficient to infer, a causal relationship between long-
term nitrogen dioxide exposure and cardiovascular
effects and diabetes.
Rationale for Causal
Determination3
Key Evidence13
Key References13
NO2 Concentrations
Associated with Effects0
Limited Evidence for Key Events in the Proposed Mode of Action for Cardiovascular Effects and Diabetes
Oxidative stress
Some, but not entirely consistent de Burbure etal. (2007), Rats: 1,000 ppb for 28 days,
evidence of increased oxidative
stress in rats (i.e., MDA, TEARS)
with short-term and long-term
NO2 exposures and in plasma
from NO2-exposed humans
(i.e., LOX-1).
tLietal. (2011 a).
tChannelletal. (2012)
Section 4.3.2.9,
Figure 4-3
5,320 ppb but not 2,660 ppb
for 7 days
Human cells exposed to
plasma from healthy adults:
500 ppb for 2 h
Inflammation
Limited and supportive
toxicological evidence of
increased transcription of some
inflammatory mediators in vitro
(i.e., IL-8, ICAM-1, VCAM-1) and
in rats (i.e., ICAM-1, TNF-a) after
short-term NO2 exposure.
tChannelletal. (2012)
tLietal. (2011 a)
Section 4.3.2.9,
Figure 4-3
Human cells exposed to
plasma from healthy adults:
500 ppb for 2 h
Rats: 2,660 and 5,320 ppb
for 7 days
Limited and inconclusive
evidence in controlled human
exposure studies (i.e., IL-6, IL-8,
ICAM-1).
tHuanqetal. (2012b),
tRiedletal. (2012)
350 and 500 ppb for 2 h
Inconsistent epidemiologic
evidence for increases in CRP
and IL-6 in adults.
Section 6.3.8
Dyslipidemia (for heart Increased triglycerides and
disease) decreased HDL—in rats.
Takano et al. (2004)
Section 6.3.8
160 ppb for 32 weeks
BC = black carbon; CHD = coronary heart disease; CO = carbon monoxide; CRP = C-reactive protein; EC = elemental carbon;
HDL = high-density lipoprotein; ICAM = intercellular adhesion molecule 1; IDW = inverse distance weighting; IGM = impaired
glucose metabolism; IHD = ischemic heart disease; IL-6 = interleuken-6; IL-8 = interleukin-8; LOX-1 = lectin-like oxidized low
density lipoprotein receptor 1; LUR = land use regression; MDA = malondialdehyde; Ml = myocardial infarction; NO2 = nitrogen
dioxide; NOX = sum of NO and NO2; NR = not reported; PM25 = particulate matter with a nominal mean aerodynamic diameter less
than or equal to 2.5 |jm; TEARS = thiobarbituric acid reactive substances; TNF-a = tumor necrosis factor alpha;
VCAM-1 = vascular cell adhesion molecule 1.
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Tables I and N. of the
Preamble.
Describes the key evidence and references, supporting or contradicting, contributing most heavily to causal determination and,
where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where full body of evidence is
characterized.
°Describes the NO2 concentrations with which the evidence is substantiated (for experimental studies, £ 5,000 ppb).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
6-102
-------
6.4 Reproductive and Developmental Effects
6.4.1 Introduction
The body of literature characterizing reproductive and developmental effects associated
with exposure to NC>2 is large and has grown considerably since the 2008 ISA for Oxides
of Nitrogen (U.S. EPA. 2008c). with much of the research focused on birth outcomes.
Well-designed studies with consideration of potential confounding and other sources of
bias are emphasized in this section (see Appendix for study evaluation guidelines). Due
to the growth in the quantity of literature, as well as in the breadth of the health endpoints
evaluated, the reproductive and developmental effects are divided into three separate
categories: (1) Fertility, Reproduction, and Pregnancy (i.e., the ability to achieve and
maintain a healthy pregnancy, with emphasis on the health of potential parents); (2) Birth
Outcomes [i.e., measures of birth weight and fetal growth, preterm birth (PTB), birth
defects, and infant mortality, with emphasis on the perinatal health of the child]; and (3)
Developmental Effects (i.e., effect on development through puberty/adolescence).
Separate causal determinations are made for each of these categories. Among the
epidemiologic studies of birth outcomes, various measures of birth weight and fetal
growth, such as low birth weight (LEW), small for gestational age (SGA), intrauterine
growth restriction (IUGR), and preterm birth (<37-week gestation) have received more
attention in air pollution research, while congenital malformations are less studied. There
is some examination of effects on fertility and pregnancy conditions; however, studies on
any particular endpoint remain limited. In toxicological studies, outcomes analogous to
fetal growth and birth weight in humans include litter size and birth weight. Nervous
system outcomes after early life exposures to NO2 are examined in the toxicological and
epidemiologic literature.
A major issue in studying environmental exposures and reproductive and developmental
effects (including infant mortality) is selecting the relevant exposure period because the
biological mechanisms leading to these outcomes and the critical periods of exposure are
poorly understood. Exposures proximate to death may be most relevant if exposure
causes an acute effect. However, exposure occurring in gestation or early life might affect
critical growth and development. For some developmental effects, results may be
observable later in the first year of life, or cumulative exposure during the first year of
life may be the most important determinant. To account for this, many epidemiologic
studies evaluate multiple exposure periods including long-term exposure periods, such as
6-103
-------
the entirety of pregnancy; and individual trimesters or months of pregnancy or short-term
(days to weeks) exposure periods, such as the days and weeks immediately preceding
birth. Due to the shorter length of gestation in rodents (about 18-24 days, on average),
animal toxicological studies investigating the effects of NO2 on pregnancy generally
utilize short-term exposure periods, which cover an entire lifestage. Thus, a study in
humans that uses the entire pregnancy as the exposure period is considered to have a
long-term exposure period (about 40 weeks, on average), while a toxicological study
conducted with rats that also uses the entire pregnancy as the exposure period (about
18-24 days, on average) is defined as a short-term exposure. In order to characterize the
weight of evidence for the effects of NO2 on reproductive and developmental effects in a
consistent, cohesive, and integrated manner, results from both short-term and long-term
exposure periods are included in this section and are identified accordingly in the text and
tables throughout this section.
Although the biological mechanisms are not fully understood, several hypotheses have
been proposed for the effects of NCh on reproductive and developmental outcomes; these
include oxidative stress, systemic inflammation, vascular dysfunction, and impaired
immune function. The study of these outcomes can be difficult given the need for
detailed exposure data and potential residential movement of mothers during pregnancy.
Air pollution epidemiologic studies reviewed in the 2008 ISA for Oxides of Nitrogen
(U.S. EPA. 2008c) examined impacts on birth-related endpoints, including intrauterine,
perinatal, post-neonatal, and infant deaths; premature births; intrauterine growth
restriction; very low birth weight (weight < 1,5 00 g) and low birth weight (weight
<2,500 g); and birth defects. However, in the limited number of studies included in the
2008 ISA for Oxides of Nitrogen, no associations were found between NO2 and birth
outcomes, with the possible exception of birth defects. Overall, the evidence evaluated in
the 2008 ISA for Oxides of Nitrogen was inconsistent and lacked coherence and
plausibility, and was determined to be inadequate to infer the presence or absence of a
causal relationship.
Several recent articles reviewed methodological issues relating to the study of outdoor air
pollution and adverse birth outcomes (Chen et al., 2010a; Woodruff et al., 2009; Ritz and
Wilhelm. 2008; Slama et al.. 2008). Some of the key challenges to interpretation of these
study results include the difficulty in assessing exposure as most studies use existing
monitoring networks to estimate individual exposure to ambient air pollution; the
inability to control for potential confounders such as other risk factors that affect birth
outcomes (e.g., smoking); evaluating the exposure window (e.g., trimester) of
importance; and limited evidence on the physiological mechanism of these effects (Ritz
and Wilhelm. 2008; Slama et al.. 2008). Recently, an international collaboration was
formed to better understand the relationships between air pollution and adverse birth
6-104
-------
outcomes and to examine some of these methodological issues through standardized
parallel analyses in data sets from different countries (Woodruff et al.. 2010). Initial
results from this collaboration have examined PM and birth weight (Parker et al.. 2011);
work on NC>2 has not yet been performed. The exposure assessment method was an
important consideration in the evaluation of long-term exposure and reproductive and
developmental outcomes, given the spatial variability typically observed in ambient NO2
concentrations (Section 2.5.3). Exposure assessment was evaluated drawing upon
discussions in Section 3.2 and Section 3.4.5. Many recent studies employed exposure
assessment methods to account for the spatial variability of NC>2. For example, LUR
model predictions generally have been found to correlate well with outdoor NO2
measurements (Section 3.2.2.1). For long-term NC>2 exposure, exposure assessment was
evaluated by the extent to which the method represented the spatial variability in NO2
concentrations in a given study. For modeled estimates, such information includes
statistics indicating the correlation between predicted and measured NO2 concentrations.
Overall, the number of studies examining the relationship of NC>2 exposure with
reproductive and developmental effects has increased tremendously, yet evidence for an
independent effect of NO2 exposure remains relatively uncertain. Although a previous
animal study (Shalamberidze and Tsereteli. 1971a. b) found that exposure to NCh during
pregnancy led to some abnormal birth outcomes in rats, epidemiologic studies to date
have reported inconsistent results for the association of ambient NO2 concentrations and a
range of reproductive and developmental effects, though the evidence has been generally
supportive for some outcomes such as fetal growth.
6.4.2 Fertility, Reproduction, and Pregnancy
6.4.2.1 Effects on Sperm
A limited amount of research has been conducted to examine the association between
NC>2 and male reproductive outcomes, specifically semen quality. To date, the
epidemiologic studies have considered various exposure durations before semen
collection that encompass either the entire period of spermatogenesis (i.e., 90 days) or
key periods of sperm development that correspond to epididymal storage, development of
sperm motility, and spermatogenesis.
An occupational study of male motorway company employees reported that men with the
highest NO2 exposures in the workplace (near-road environment, -160 ppb) had lower
sperm motility, but no difference in sperm count, compared to men with lower exposures
6-105
-------
[~80 ppb; (Boggia et al.. 2009)1. Two epidemiologic studies evaluated the relationship
between sperm quality and ambient concentrations of NCh based on ambient and personal
monitoring (Rubes et al.. 2010) or statistical models (Sokol et al.. 2006) and observed no
associations; while a cross-sectional study observed associations between NO2 measured
at central site monitors and some semen quality parameters (Zhou et al.. 2014). No recent
toxicological studies have examined the effect of NO2 exposure on semen quality. Kripke
and Sherwin (1984) found no statistically significant effects on spermatogenesis, or on
germinal and interstitial cells of the testes of a small group of LEW/f mai rats (n = 6)
after 21 days of exposure to a single concentration of NCh, 1,000 ppb 7 h/day,
5 days/week (Table 6-13). Overall, there is little and inconsistent epidemiologic evidence
and no toxicological evidence of effects of NO2 exposure on sperm or semen quality.
6.4.2.2 Effects on Reproduction
Several recent studies have examined the association between exposure to NO2 during
pregnancy and the ability to reproduce. Gametes (i.e., ova and sperm) may be even more
at risk, especially outside of the human body, as occurs with assisted reproduction.
Smokers require twice the number of in vitro fertilization (IVF) attempts to conceive as
nonsmokers (Feichtinger et al.. 1997). suggesting that a preconception exposure can be
harmful to pregnancy. A recent study estimated daily concentrations of criteria pollutants
at addresses of women undergoing their first IVF cycle and at their IVF labs from 2000 to
2007 in the northeastern U.S. (Legro et al.. 2010). Increasing NC>2 concentration
estimated at the patient's address during ovulation induction (short-term exposure,
-12 days) was associated with a decreased chance of live birth (OR: 0.80 [95% CI: 0.71,
0.91] per 10-ppb increase). Similar risks were observed when the exposure period was
the daily concentration averaged over the days from oocyte retrieval through embryo
transfer, and the days from embryo transfer through the pregnancy test (14 days). The
authors also observed a decreased odds of live birth when exposed from embryo transfer
to live birth [long-term exposure, -200 days; OR: 0.76 [95% CI: 0.56, 1.02] per 10-ppb
increase). After adjusting for Os in a copollutant model, NO2 continued to be associated
with IVF failure. The results of this study suggest that both short- and long-term exposure
to NO2 during ovulation and gestation was detrimental and reduced the likelihood of a
live birth. In a more general population, increased NO2 exposure in the 30 days before
initiation of unprotected intercourse also was associated with reduced fecundability
[fecundability ratio per 10 ppb: 0.50 [95% CI: 0.32, 0.76]) (Slamaetal.. 2013)1.
Similarly, in a cross-sectional study of fertility rates, Nieuwenhuijsen et al. (2014)
observed lower fertility in census tracts with higher NO2 and NOx concentrations
estimated at the level of the census tract. Importantly, census tract associations may not
6-106
-------
reflect associations for individuals. Further, none of these studies examined confounding
by traffic-related copollutants.
In contrast, NC>2 exposure has not been shown to induce effects on fertility in animals. A
breeding study found that exposures of rats to 67 or 1,300 ppb NC>2 12 h/day for 3 months
(Table 6-13) produced no change in the number of dams that became pregnant after
mating with an unexposed male (Shalamberidze and Tsereteli, 1971a. b). At the higher
dose, NO2 exposure impaired estrous cyclicity (cycle prolongation, increased duration of
diestrus, decreased number of normal and total estrus cycles), and the exposed females
had a decreased number of ovarian primordial follicles.
6.4.2.3 Effects on Pregnancy
Epidemiologic Evidence
Evidence suggests that exposure to air pollutants may affect maternal and fetal health
during pregnancy. Systemic inflammation has been proposed as a potential biological
mechanism through which air pollution could result in adverse pregnancy outcomes
(Slama et al., 2008; Kannan et al., 2006). Recent studies have investigated the
relationship between CRP, a marker for systemic inflammation, measured in maternal
blood during early pregnancy and in umbilical cord blood (as a measure of fetal health)
and NC>2 concentrations, van den Hooven et al. (2012a) observed generally null
associations between exposure to NO2 and maternal CRP levels but did observe a
positive, linear relationship between quartiles of NO2 exposure and fetal CRP levels. This
association was evident when exposure was measured 1, 2, and 4 weeks prior to delivery
but was strongest when exposure to NO2 was measured over the entire pregnancy.
Similarly, Lee et al. (20lie) observed generally null associations between short-term
exposure (i.e., 1 to 29 days) to NO2 and elevated maternal CRP levels.
Pregnancy-associated hypertension is a leading cause of perinatal and maternal mortality
and morbidity. A large body of research has linked changes in blood pressure to ambient
air pollution; however, evidence is inconsistent for NC>2 (Sections 5.3.6 and 6.3.4). A few
recent studies have examined whether increases in NO2 concentrations are associated
with blood pressure changes in women who are pregnant. The results of these studies are
not consistent. Hampel etal. (2011) observed that increases in NO2 concentrations
measured at central site monitors were associated with decreases in systolic blood
pressure but found no clear associations between NO2 concentrations and diastolic blood
pressure. Lee etal. (2012b) observed associations between exposure to NO2 estimated for
the maternal ZIP code using kriging interpolation and changes in blood pressure that
6-107
-------
were null for the entire population and when the population was restricted to nonsmokers.
van den Hooven et al. (2011) observed small increases in systolic blood pressure
associated with increases in NCh concentrations estimated from a GIS spatiotemporal
model across all three trimesters of pregnancy but did not observe a similar association
with diastolic blood pressure. Mobasher et al. (2013) observed a positive association
between exposure to NCh (estimated with IDW and spatial mapping) during the first
trimester and hypertensive disorders of pregnancy, though the association was imprecise
and was reduced when exposure was averaged over the second and third trimesters. The
same pattern was observed when analyses were restricted to nonobese women, but among
obese women, the effect estimate was below 1.00 for each trimester. Xuet al. (2014)
observed positive associations between NO2 concentrations measured at central site
monitors during the entire pregnancy and first trimester and hypertensive disorders of
pregnancy. Associations remained positive after adjustment for CO or PIVb 5; however,
differential exposure measurement error may limit inference regarding a potential
independent association for NC>2.
New-onset gestational hypertension can contribute to pre-eclampsia, a common
pregnancy complication diagnosed after 20 weeks of pregnancy. Wu et al. (2009)
observed a 45% increase (95% CI: 23, 64) in the risk of pre-eclampsia associated with a
20-ppb increase in NOx averaged over the entire pregnancy. When NOx was examined as
categories, the association was consistent with a linear concentration-response
relationship. Similarly, NC>2 concentrations during pregnancy were associated with an
increased risk of pre-eclampsia among a cohort of Australian women (Pereira et al..
2013). with the strongest association observed when exposure was limited to the third
trimester. Malmqvist et al. (2013) also observed a positive association between NOx
concentrations in the third trimester of pregnancy and pre-eclampsia consistent with a
linear concentration-response relationship in a Swedish cohort. Dadvand et al. (2013)
observed increases in odds of pre-eclampsia, particularly late-onset pre-eclampsia, with
increased NO2 exposure during the third and first trimesters, and with entire pregnancy
exposures. A number of other studies, of similar quality and using similar study designs,
did not observe positive associations for NO2 exposure and risks of pregnancy-induced
hypertension or pre-eclampsia across different exposure periods including exposure over
entire pregnancy (Nahidi et al.. 2014; van den Hooven et al.. 2011) and the first trimester
exposure (Olsson et al.. 2013)1. NO2 exposure estimated at each subject's residential
address using LUR or dispersion models that were demonstrated to predict well the
ambient NO2 measurements in the study areas showed both positive [cross-validation R2
= 0.68; (Pereira et al.. 2013)] and null [r = 0.77 for correlations between modeled and
measured NO2 (van den Hooven et al.. 2011)1 associations with pre-eclampsia. A
meta-analysis of pre-eclampsia studies reported a combined OR for NO2 of 1.23 (95%
CI: 1.04, 1.42), though there was a large amount of heterogeneity between studies
6-108
-------
particularly in outcome definition; removal of an influential study produced an OR of
1.11 (95% CI: 1.04, 1.17) with no observed heterogeneity (Tedersen et al.. 2014). None
of the studies of pre-eclampsia examined confounding by PM25 or traffic-related
copollutants.
Other pregnancy complications that have recently been evaluated and found to be
associated with NO2 include gestational diabetes (Malmqvist et al., 2013) and markers of
placental growth and function (van den Hooven et al.. 2012b). Overall, the evidence for
the effects of NO2 on pregnancy outcomes is inconsistent. Key studies examining the
association between exposure to NO2 and pregnancy-related effects can be found in Table
6-12. Supplemental Table S6-2 (U.S. EPA, 2013g) provides an overview of all of the
epidemiologic studies of pregnancy-related health effects.
lexicological Evidence
Evidence from animal toxicological studies (Table 6-13) does not clearly indicate that
NO2 exposure affects pregnancy. NO2 exposure of rats to 1,500 or 3,000 ppb over the
duration of pregnancy did not alter dam weight gain over pregnancy, assessed as a
percentage of body weight at conception or Gestational Day (GD) 0 (Pi Giovanni et al..
1994). Fetal lethality in toxicological studies is measured by counting pup loss or
resorption sites. This directly affects litter size, i.e., number of live pups born. Rat dams
that received 1,300 ppb NO2 12 h/day for 3 months before pregnancy had decreased litter
size (Shalamberidze and Tsereteli. 1971a, b). However, litter size was not affected in rat
dams exposed to 1,500 or 3,000 ppb NO2 exposure over the duration of pregnancy (Di
Giovanni et al., 1994)
6-109
-------
Table 6-12 Key epidemiologic studies of oxides of nitrogen and reproductive
and developmental effects.
Location
Study Sample Size
Exposure Assessment and
Concentrations (ppb)
Selected Effect Estimates (95% Cl)a
Fertility, Reproduction, and Pregnancy
tLeqro et al. Northeastern
(2010) U.S.
n = 7,403
tSlama et al. Teplice, Czech
(2013) Republic
n = 1,916
tWu et al. southern
(2009) California, U.S.
n = 81,186
tPereira et al. Perth, Australia
(2013) n = 23,452
Central site monitors combined by
kriging
3. 6- to 199-day avg NO2
Means: -19 for each exposure
window evaluated
Central site monitor
Within 12 km of residence
1- and 2-mo avg NO2
Median for 1-mo avg: 19
Dispersion model
Pregnancy- and trimester-avg NOx
r= 0.91 for NOx and PM2.5
Means: EP 7.2, T1 7.5,127.3,
T3: 7.1
LUR model
Pregnancy- and trimester-avg NO2
Odds of live birth following IVF
Medication start to oocyte retrieval:
0.80(0.71,0.91)
Oocyte retrieval to embryo transfer:
0.87(0.79,0.96)
Embryo transfer to pregnancy test
(14 days): 0.76(0.66, 0.86)
Embryo transfer to live birth:
0.76(0.56, 1.02)
Fecundity ratio
30 days before unprotected intercourse
(Lag 1): 0.50(0.32, 0.76)
30 days before Lag 1 (Lag 2):
1.10(0.69, 1.80)
Lag 1 + Lag 2:
0.52 (0.28, 0.94)
Month post-outcome:
1.17(0.76, 1.85)
Pre-eclampsia
EP: 1.44(1.23, 1.68)
Pre-eclampsia
T1: 1.04(0.94, 1.16)
NO2 estimated at home address. No
information on model validation.
Means: EP: 23.0, T1 23.3, T2 23.3,
T3 22.5
T2: 1.02(0.91, 1.15)
T3: 1.17(1.04, 1.32)
EP: 1.22(1.02, 1.49)
tMalmqvist et Sweden Dispersion model
al. (2013) N = 81,110 Trimester-avg NOx
Median forT1: 7.5
NOx estimated at home address.
Dispersion model had a spatial
resolution of 500 * 500 m.
Pre-eclampsia, T3 exposure
Q1: reference
Q2: 1.28(1.13, 1.46)
Q3: 1.33(1.17, 1.52)
Q4: 1.51 (1.32, 1.73)
Gestational diabetes, T3 exposure
Q1: reference
Q2: 1.19(0.99, 1.44)
Q3: 1.52(1.28, 1.82)
Q4: 1.69(1.41,2.03)
6-110
-------
Table 6-12 (Continued): Key epidemiologic studies of oxides of nitrogen and
reproductive and developmental effects.
Study
tDadvand et al.
(2013)
Location Exposure Assessment and
Sample Size Concentrations (ppb)
Barcelona, Spain LUR model
n = 8,398 Pregnancy- and trimester-avg NO2
Selected Effect Estimates (95% Clf
Pre-eclampsia
T1: 1.07(0.94, 1.22)
NO2 estimated at home address.
Cross-validation R2 = 0.68.
Means: EP 30, T1 30, T2 31, T3 31
T2: 1.03(0.90, 1.19)
T3: 1.11 (0.99, 1.23)
EP: 1.09(0.94, 1.27)
Birth Outcomes
tAquilera et al.
(2010)
Catalonia, Spain
n = 562
LUR model
Trimester avg NO2
NO2 estimated at home address.
Root mean squared error for cross-
validation = 0.85 ppb
Means: T1 17.3, T2 16.9, T3 17.1
Fetal length (% change)
T1: -2.04 (-7.01, 2.95)
T2: -1.69 (-7.05, 3.69)
T3: 0.33 (-4.06, 4.72)
Head circumference (% change)
T1: 0.25 (-5.42, 5.91)
T2: 1.70 (-3.69, 7.07)
T3: 0.23 (-4.32, 4.77)
Abdominal circumference (% change)
T1: -2.82 (-8.24, 2.59)
T2: -0.13 (-5.64, 5.38)
T3: 0.74 (-3.92, 5.40)
Biparietal diameter (% change)
T1: 3.87 (-2.04, 9.75)
T2: 4.90 (-0.34, 10.11)
T3: 1.48 (-3.41, 6.35)
Estimated fetal weight (% change)
T1: -2.22 (-7.39, 2.98)
T2: 0.46 (-5.82, 6.72)
T3: 0.91 (-3.65, 5.45)
tlfiiquez et al.
(2012)
Valencia, Spain LUR model
n = 818 Trimester-avg NO2
NO2 estimated at home address. No
information on model validation
Median: 20.2 for EP
Fetal length (ratio)
T1: 0.97(0.92, 1.02)
T2: 0.96(0.92, 1.00)
T3: 0.97(0.92, 1.02)
Abdominal circumference (ratio)
T1: 0.96(0.92,0.99)
T2: 0.98(0.94, 1.02)
T3: 0.98(0.94, 1.03)
Biparietal diameter (ratio)
T1: 0.96(0.92, 1.00)
T2: 0.97(0.92, 1.01)
T3: 0.98(94, 1.02)
Estimated fetal weight (ratio)
T1: 0.96(0.92, 1.00)
T2: 0.98(0.94, 1.02)
T3: 0.97(0.93, 1.02)
6-111
-------
Table 6-12 (Continued): Key epidemiologic studies of oxides of nitrogen and
reproductive and developmental effects.
Study
Location Exposure Assessment and
Sample Size Concentrations (ppb)
Selected Effect Estimates (95% Clf
tvan den
Hooven et al.
(2012c)
Rotterdam, the
Netherlands
n = 7,772
Dispersion model, GIS-based
techniques, and central site
monitors combined
Pregnancy- and trimester-avg NO2
Pearson r = 0.77 for modeled and
measured annual avg NO2 at 18
sites
Mean: 21.2, Median: 21.1,
75th: 22.4, Max: 30.3
Head circumference (mm), T3
Q1: reference
Q2: -0.40 (-1.00, 0.20)
Q3: -0.81 (-1.42, -0.20)
Q4: -1.28 (-1.96, -0.61)
Length (mm), T3
Q1: reference
Q2: -0.02 (-0.17, 0.13)
Q3: -0.09 (-0.24, 0.06)
Q4: -0.33 (-0.50, -0.16)
SGA, EP
Q1: reference
Q2: 0.93(0.66, 1.31)
Q3: 1.25(0.90, 1.73)
Q4: 1.35(0.94, 1.94)
tHansen et al.
(2008)
Brisbane,
Australia
n = 15,623
Nearest central site monitor
1-mo avg NO2
Homes within 2-14 km of 1 of
17 monitors
Mean: 9.8 for city 1992-2003
Head circumference (mm)
M1: 0.54 (-1.88, 2.94)
M2: -0.16 (-2.54, 2.20)
M3: -0.60 (-3.18, 2.00)
M4: -0.30 (-2.30, 1.68)
Biparietal diameter (mm)
M1: 0.14 (-0.62, 0.88)
M2: -0.20 (-0.88, 0.50)
M3: -0.12 (-0.82, 0.58)
M4: -0.16 (-0.74, 0.42)
Abdominal circumference (mm)
M1: 0.48 (-1.98, 2.94)
M2: 0.98 (-1.40, 3.34)
M3: 0.20 (-2.12, 2.52)
M4: 0.30 (-1.80, 2.40)
Femur length (mm)
M1: 0.06 (-0.50, 0.62)
M2: -0.18 (-0.78, 0.44)
M3: 0.02 (-0.52, 0.56)
M4: -0.26 (-0.80, 0.26)
tRitz et al.
(2014)
Los Angeles, CA
LUR: n = 501
Central site
monitors: n = 98
LUR model
Trimester-avg NO2
NO2 estimated at home address.
Cross-validation R2 = 0.87
Mean: 22.7
Biparietal diameter (mm)
GW 0-19:-0.41 (-1.07, 0.23)
GW 19-29: 0.39 (-0.25, 1.02)
GW 29-37: -0.50 (-1.23, 0.23)
Central site monitors
Measurements combined by IDW
Means: T1 37.3 T2 37.6 T3 39.3
GW 0-19:-4.45 (-10.55, 1.55)
GW 19-29: 4.92 (0.03, 9.83)
GW 29-37: -8.33 (-13.83, -2.83)
6-112
-------
Table 6-12 (Continued): Key epidemiologic studies of oxides of nitrogen and
reproductive and developmental effects.
Study
Location Exposure Assessment and
Sample Size Concentrations (ppb)
Selected Effect Estimates (95% Clf
tEstarlich et al.
(2011)
Asturias,
Gipuzkoa,
Sabadell,
Valencia, Spain
n= 2,337
LUR model
Pregnancy-avg NO2
NO2 estimated at home address. No
information on model validation
Means: Overall 15.5, Urban 15.9,
Rural: 8.7
Birth length (cm)
EP: -1.69 (-0.34, -0.02)
Head circumference (cm)
EP: -0.01 (0.13, 0.11)
tBallesteretal.
(2010)
Valencia, Spain
n = 785
LUR model and central site monitors Head circumference (cm)
combined
Pregnancy-avg NO2
NO2 estimated at home address. No
information on model validation.
Mean: 19.6
-0.11 (-0.25, 0.03)
Birth length (cm)
-0.09 (-0.27, 0.10)
SGA—weight
1.59(0.89,2.84)
SGA—length
1.48(0.628, 3.49)
tHansen et al. Brisbane,
(2007) Australia
n= 26,617
Central site monitors
Trimester-avg NO2
Measurements averaged across city
City-wide concentrations
1999-2003: Median 7.8, 75th: 11.4,
Max: 24.2
Head circumference (cm)
T1: 0.05 (-0.05, 0.17)
T2: 0.08 (-0.02, 0.19)
T3: 0.00 (-0.10, 0.10)
Crown-heel length (cm)
T1: 0.24(0.05,0.42)
T2: 0.07 (-0.10,0.24)
T3: -0.15 (-0.25, -0.05)
tDarrow et al. Atlanta, GA Central site monitors
(2011b) n = 406,627 Pregnancy-, trimester-, and first
28-day avg of 1-h max NO2
Population-weighted average of all
monitors in city
Mean: 23.6 for first 28-day avg
Birth weight (g)
Entire pregnancy: -18.4 (-28.0, -9.0)
First 28 days: 0.8 (-3.6, 5.2)
T3
All subjects: -9.0 (-17.0, -1.2)
Non-Hispanic white: -9.2 (-18.6, 0.2)
Non-Hispanic black: -7.8 (-17.4, 1.6)
Hispanic: -11.6 (-24.8, 1.4)
tBelletal.
(2007)
Connecticut and
Massachusetts,
U.S.
n = 358,504
Central site monitors
Pregnancy-avg
Measurements averaged across
county
Mean: 17.4
Birth weight (g), EP exposure
All subjects: -18.5 (-22.5, -14.6)
Black mothers: -26.5 (-37.5, -15.6)
White mothers: -17.3 (-21.7 -13.1)
LEW: 1.06(1.00, 1.11)
6-113
-------
Table 6-12 (Continued): Key epidemiologic studies of oxides of nitrogen and
reproductive and developmental effects.
Study
Location Exposure Assessment and
Sample Size Concentrations (ppb)
Selected Effect Estimates (95% Clf
Postnatal Development
tvan Kempen
etal. (2012)
the Netherlands
n=485
LUR model—school
Annual avg NO2
Concurrent exposure. Cross-
validation R2 = 0.85.
Mean: 16.5
Change in score, adjusted for traffic
noise
Memory: -0.30 (-0.55, 0.04)
Measures of attention
SRTT, reaction time (msec):
-2.23 (-22.1, 17.7)
SAT block, # errors: -0.02 (-0.42, 0.38)
SAT block, reaction time (msec):
13.9 (-16.7, 43.9)
SAT switch, # errors: -1.19 (-3.62, 1.26)
SAT switch, reaction time (msec):
21.5 (-45.2, 88.2)
Locomotion: 0.08 (-0.08, 0.25)
LUR model—home
Annual avg NO2
Concurrent exposure. Cross-
validation R2 = 0.85.
Mean: 16.4
Change in score, adjusted for traffic
noise
Memory: 0.17 (-0.08, 0.42)
Measures of attention
SRTT, reaction time (msec):
-2.11 (-21.0, 16.7)
SAT block, # errors: -0.04 (-0.40, 0.32)
SAT block, reaction time (msec):
15.9 (-11.3, 43.0)
SAT switch, # errors: -1.23 (-3.32, 0.87)
SAT switch, reaction time (msec):
-20.2 (-74.9, 34.5)
Locomotion: 0.06 (-0.08, 0.21)
tClarketal.
(2012)
U.K.
n = 719
LUR model and dispersion model
combined—school
Annual avg NO2
Concurrent exposure. No
information on model validation.
Mean: 22.7
Change in score, adjusted for traffic
noise
Reading comprehension:
0.08 (-0.17, 0.34)
Information recall:
0.28 (-0.62, 1.17)
Working memory:
0.06 (-5.55, 5.66)
Physiological distress:
0.47 (-0.62, 1.57)
tFreire et al.
(2010)
Spain
n=210
LUR model—home
Annual avg NO2
Concurrent exposure. Cross-
validation R2 = 0.64.
Mean: 11.1
Change in score in group with
NO2 >13.2 ppb compared with group
with NO2 <8.2 ppb
General cognitive index: -4.2 (-14, 5.6)
Verbal: -3.09 (-13.31, 7.13)
Quantitative: -6.71 (-17.91, 4.49)
Memory: -5.52 (-16.18, 5.13)
Executive function: -4.93 (-14.90, 5.05)
Gross motor function: -8.6 (-19, 1.7)
Fine motor skills: 0.91 (-10.22, 12.05)
6-114
-------
Table 6-12 (Continued): Key epidemiologic studies of oxides of nitrogen and
reproductive and developmental effects.
Study
Location Exposure Assessment and
Sample Size Concentrations (ppb)
Selected Effect Estimates (95% Clf
tGuxens et al. n = 9,482
(2014)
LUR model—home
Annual avg NO2
Prenatal exposure. Exposure
estimated by back-extrapolation
using central site measurements.
Cross-validation R2 = 0.49-0.87
across cities.
Change in cognitive or motor function
score
Ruhr, Germany
Heraklion,
Greece
Asturias, Spain
Gipuzkoa, Spain
Valencia, Spain
Sabadell, Spain
Granada, Spain
Rotterdam, the
Netherlands
Poitiers, France
Nancy, France
Rome, Italy
Median: NR
Median: 6.1
Median: NR
Median: NR
Median: NR
Median: 23.4
Median: NR
Median: NR
Median: NR
Median: NR
Median: NR
Mental development: -3.61 (-8.53, 1.32)
Motor function: -5.04 (-11, 0.49)
Mental development: 1.90 (-2.33, 6.13)
Motor function: -0.83 (-5.39, 3.74)
Mental development: -1.39 (-3.12, 0.34)
Motor function: -2.03 (-3.82, -0.24)
Mental development: -1.11 (-6.65, 4.44)
Motor function: 0.17 (-1.73, 5.34)
General cognition: -1.35 (-3.74, 1.03)
Motor function: -3.72 (-6.37, -1.07)
General cognition: -0.15 (-2.42, 2.12)
Motor function: 0.71 (-1.71, -3.14)
General cognition: 3.18 (-0.26, 6.62)
Motor function: 1.80 (-1.73, 5.34)
Motor function: -0.17 (-1.60, 1.26)
Motor function: -0.64 (-6.75, 5.47)
Motor function: -2.84 (-5.64, -0.04)
Motor function: -1.97 (-4. 44, 0.49)
tGuxens et al. Valencia,
(2012) Sabadell,
Asturias,
Gipuzkoa, Spain
n = 1,889
LUR model—home
Pregnancy-avg NO2
Exposure estimated by back-
extrapolation using central site
measurements. Cross-validation R2
= 0.75, 0.77.
Means: Overall 15.7, Valencia 19.6,
Sabadell 17.1, Asturias 12.3,
Gipuzkoa: 10.7
Change in mental development indexb
Location
All regions: -0.95 (-3.90, 1.89)
Gipuzkoa: -5.15 (-8.04, -2.27)
Asturias: 0.17 (-2.71, 3.04)
Sabadell: 1.98 (-1.69, 5.66)
Valencia:-0.43 (-2.86, 2.01)
Maternal fruit and vegetable intake
<405 g/day: -4.13 (-7.06, -1.21)
>405g/day: 0.25 (-3.63, 4.12)
Maternal Vitamin D circulation
Low: -2.49 (-6.87, 1.89)
Medium: -0.55 (-3.48, 2.39)
High: -0.11 (-2.72,2.49)
6-115
-------
Table 6-12 (Continued): Key epidemiologic studies of oxides of nitrogen and
reproductive and developmental effects.
Study
Location Exposure Assessment and
Sample Size Concentrations (ppb)
Selected Effect Estimates (95% Clf
tBecerra et al.
(2013)
Los Angeles
County, CA
n = 83,385
LUR model—home
Pregnancy- and trimester-avg NO2
Monthly exposure estimated using
central site measurements.
Cross-validation R2 = 0.87.
Mean: 28.0
Odds ratio for autism
EP: 1.05(0.98, 1.12)
T1: 1.03(0.98, 1.08)
T2: 1.03(0.98, 1.08)
T3: 1.04(0.98, 1.09)
Central site monitor
Nearest central site monitor to birth
residence.
Mean: 30.8
EP: 1.04(0.98, 1.10)
T1: 1.04(0.99, 1.08)
T2: 1.01 (0.97, 1.06)
T3: 1.02(0.97, 1.07)
tVolk et al.
(2013)
California, U.S.
n = 524
Dispersion model
Pregnancy- and first yr-avg NOx.
Exposure estimated within 5 km of
child's home, r ~ 0.99 for correlation
with EC and CO.
Q1: <9.7
Q2: 9.7-16.9
Q3: 16.9-31.8
Q4: >31.8
Odds ratio for autism relative to Q1
First yr of life
Q2: 0.91 (0.56, 1.47)
Q3: 1.00(0.62, 1.62)
Q4: 3.10(1.76, 5.57)
EP
Q2: 1.26(0.77,2.06)
Q3: 1.09(0.67, 1.79)
Q4: 1.98(1.20, 3.31)
Central site monitors
Measurements from sites within
50 km of homes, combined by IDW
Odds ratio for autism
First yr of life: 1.67(1.25,2.23)
EP: 1.52(1.16,2.00)
avg = average; CA = California; Cl = confidence interval; cm = centimeter; EC = elemental carbon; EP = entire pregnancy;
GIS = geographic information system; GW = gestational week; IDW = inverse distance weighting; IVF = in vitro fertilization;
LBW = low birth weight; LUR = land use regression; M1 = Month 1; M2 = Month 2; M3 = Month 3; M4 = Month 4; msec =
milliseconds; NO2 = nitrogen dioxide; NOX = sum of NO and NO2; NR = not reported; Q1 = 1st quartile; Q2 = 2nd quartile; Q3 = 3rd
quartile; Q4 = 4th quartile; SAT = switching attention test; SGA = small for gestational age; SRTT = simple reaction time test;
T1 = 1st trimester; T2 = 2nd trimester; T3 = 3rd trimester.
aRelative risk per 10-ppb change in NO2 or 20-ppb change in NOX, unless otherwise noted.
bPer doubling in NO2 concentration.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
6.4.3 Birth Outcomes
6.4.3.1
Fetal Growth
Fetal growth is influenced by maternal, placental, and fetal factors. The biological
mechanisms by which air pollutants may influence the developing fetus remain largely
unknown. LBW has often been used as an outcome measure because it is easily available
6-116
-------
and accurately recorded on birth certificates. However, LEW may result from either short
gestation or inadequate growth in utero. Most of the studies investigating air pollution
exposure and LEW limited their analyses to term infants to focus on inadequate growth.
A number of studies were identified that specifically addressed growth restriction in utero
by identifying infants who failed to meet specific growth standards. Usually, these infants
had birth weight less than the 10th percentile for gestational age, using an external
standard.
A limitation of environmental studies that use birth weight as a proxy measure of fetal
growth is that patterns of fetal growth during pregnancy cannot be assessed. This is
particularly important when investigating pollutant exposures during early pregnancy as
birth weight is recorded many months after the exposure period. The insult of air
pollution may have a transient effect on fetal growth, where growth is hindered at one
point in time but catches up at a later point. For example, maternal smoking during
pregnancy can alter the growth rate of individual body segments of the fetus at variable
developmental stages, as the fetus experiences selective growth restriction and
augmentation (Lampl and Jeanty. 2003).
The terms SGA, which is defined as a birth weight <10th percentile for gestational age
(and often sex and/or race), and IUGR are often used interchangeably. However, this
definition of SGA does have limitations. For example, using it for IUGR may
overestimate the percentage of "growth-restricted" neonates as it is unlikely that 10% of
neonates have growth restriction ("Wolhnann. 1998). On the other hand, when the 10th
percentile is based on the distribution of live births at a population level, the percentage
of SGA among PTB is most likely underestimated (Hutcheon and Platt. 2008).
Nevertheless, SGA represents a statistical description of a small neonate, whereas the
term IUGR is reserved for those with clinical evidence of abnormal growth. Thus, all
IUGR neonates will be SGA, but not all SGA neonates will be IUGR (Wollmann. 1998).
In the following section, the terms SGA and IUGR are referred to as each cited study
used the terms.
The 2008 ISA for Oxides of Nitrogen reviewed three studies that evaluated the
relationship between exposure to NO2 and fetal growth (Mannes et al.. 2005; Salam et al..
2005; Liu et al., 2003) and concluded that they "did not consistently report associations
between NO2 exposure and intrauterine growth retardation" (U.S. EPA. 2008c). In recent
years, a number of studies have examined various metrics of fetal growth restriction.
Several of these recent studies have used anthropometric measurements (e.g., head
circumference, abdominal circumference) measured via ultrasound at different periods of
pregnancy in order to evaluate patterns of fetal growth during pregnancy and to detect
growth restriction that may occur early in pregnancy, but which may no longer be
6-117
-------
detectable at birth. In a mother and child cohort study conducted in Spain, ultrasound
measurements were recorded at 12, 20, and 32 weeks of gestation, and these
anthropometric measurements were recorded again at birth (Iniguez et al.. 2012; Aguilera
et al.. 2010). Aguilera et al. (2010) observed that exposure to NO2 early in pregnancy was
associated with impaired growth in head circumference from Weeks 12 to 20 of gestation
and abdominal circumference and estimated fetal weight from Weeks 20 to 32. Similarly,
Iniguez etal. (2012) reported decreased fetal length and decreased biparietal diameter
measured by ultrasound in association with exposure to NC>2 during Weeks 12-20 of
gestation. Decreased birth length and head circumference measured at birth were also
associated with exposure to NO2 during this same period. Examining fetal growth
characteristics assessed by ultrasound during each trimester of pregnancy, van den
Hooven et al. (2012c) observed decreases in head circumference and fetal length in the
second and third trimesters associated with exposure to NC>2. Hansen et al. (2008) used
ultrasound measurements during Weeks 13-26 of pregnancy and did not observe
associations between exposure to relatively low concentrations of NO2 (mean: 9.8 ppb)
and head circumference, biparietal diameter, abdominal circumference, or fetal length.
Ritz etal. (2014) used multiple ultrasound measures to examine fetal growth parameters
across gestation and observed that higher exposure to NO2 during Gestational Weeks
29-37 was associated with decrements in biparietal diameter at 37 weeks; no consistent
associations were found for head circumference, femur length, or abdominal
circumference.
Several studies made use of anthropometric measurements made immediately afterbirth
to evaluate fetal growth. Estarlich et al. (2011). Ballester et al. (2010). and Hansen et al.
(2007) observed decreases in body length associated with exposure to NC>2. This
association persisted when NO2 exposure was estimated for each trimester of pregnancy
in the study by Estarlich et al. (2011). Ballester et al. (2010) observed the strongest
association with NC>2 exposure during the first trimester, while Hansen et al. (2007)
reported that the association was strongest for NO2 exposure measured at the end of the
pregnancy.
When using SGA as an indicator of fetal growth restriction, several studies observed
associations with exposure to NO2, NOx, or NO (Sathyanarayana et al.. 2013; Le et al..
2012; Pereira etal.. 2012; Malmqvist etal.. 2011; Ballester etal.. 2010; Rich et al.. 2009;
Brauer et al.. 2008; Mannes et al.. 2005). These associations were most often observed
for exposure to NO2 during the second trimester (Pereira et al.. 2012; Ballester etal..
2010; Rich et al.. 2009; Mannes et al.. 2005). Gehring et al. (2011 a). Hansen et al.
(2007). Olsson etal. (2013). Kashima etal. (2011). and Hannam etal. (2014) did not
observe an increased risk of SGA associated with exposure to NO2. All of the studies that
used IUGR as an indicator of fetal growth restriction observed an association with
6-118
-------
exposure to NCh, and this association was strongest for exposures at the beginning of
pregnancy [i.e., first month or first trimester; (Liu et al.. 2007; Salam et al.. 2005; Liu et
al.. 2003)1.
When evaluating the association between fetal growth and exposure to NO2, many
studies estimated NO2 exposure at the maternal residence using LUR models (Iniguez et
al.. 2012; Pereiraet al.. 2012; Estarlich etal.. 2011; Gehring etal.. 2011 a: Aguilera et al..
2010; Ballester et al.. 2010; Brauer et al.. 2008) and dispersion models (van den Hooven
etal.. 2012c). Generally, the results of studies that relied on estimates of NO2 from LUR
models were not substantially different from those that estimated exposure to NO2 using
concentrations measured at central site monitors. Given the differences among the study
designs, it cannot be concluded that the inconsistencies are related to exposure
assessment method or length of follow-up periods. However, in a study that assigned
exposure to NCh using both a LUR model and IDW of measured NO2 concentration from
monitors, Brauer et al. (2008) found higher risks for SGA using the monitoring data (OR:
1.28 [95% CI: 1.18, 1.36]) compared to the risks observed with the NO2 estimates from
the LUR model (OR 0.94 [95% CI: 0.80, 1.10]). In general, epidemiologic studies that
estimate long-term NO2 exposure from central site monitors can carry uncertainty
because the exposure error resulting from spatial misalignment between subjects and
monitor locations can overestimate or underestimate associations with health effects
(Section 3.4.5.2). While many studies did not report on the extent to which exposures
estimated from central site monitors or models represented the spatial pattern in ambient
NO2 concentrations, some studies demonstrated that their LUR or dispersion models had
good accuracy in predicting ambient NO2 concentrations in the study areas (van den
Hooven et al.. 2012c: Aguilera et al.. 2010).
Several studies incorporated data on activity patterns in order to decrease potential error
in residential exposure estimates. Aguilera et al. (2010) and Estarlich et al. (2011) only
analyzed subjects who spent 15 or more hours per day at home or subjects who spent less
than 2 hours a day in an outdoor environment other than at their primary residence and
found stronger associations between measures of decreased fetal growth and exposure to
NO2. In contrast, when Gehring et al. (201 la) limited their analyses to participants who
did not move during pregnancy or did not have paid employment outside of the home,
there were no consistent associations between SGA and exposure to NO2.
In summary, there is generally consistent evidence for an association between exposure
to NO2 and fetal growth restriction, including recent evidence from studies that used fetal
anthropometric measurements made via ultrasound and anthropometric measurements
made immediately after birth. Some studies demonstrated the validity of models used to
estimate residential exposure for individual subjects. These are consistent with the studies
6-119
-------
of the clinical measurement of IUGR and the statistical definition of SGA. Key studies
can be found in Table 6-12. Supplemental Table S6-3 (U.S. EPA. 2013h) provides an
overview of all of the epidemiologic studies of fetal growth effects. The evidence is less
certain when it comes to assessing the time period of pregnancy when exposure to NO2 is
associated with the highest risks. Some studies find the highest risks associated with NO2
when NO2 is measured in early pregnancy, while in other studies, the time period
associated with the greatest risk is toward the end of pregnancy. Others find the greatest
risk when exposure is assigned for the entire pregnancy period. A major uncertainty is
whether NCh exposure has an independent effect on fetal growth restriction because
epidemiologic studies did not examine the potential for confounding by traffic-related
copollutants, toxicological investigation of these outcomes is lacking.
6.4.3.2 Preterm Birth
PTB is a syndrome (Romero et al.. 2006) that is characterized by multiple etiologies. It is,
therefore, unusual to be able to identify an exact cause for each PTB. In addition, PTB is
not an adverse outcome in itself but an important determinant of health status
(i.e., neonatal morbidity and mortality). Although some overlap exists for common risk
factors, different etiologic entities related to distinct risk factor profiles and leading to
different neonatal and post-neonatal complications are attributed to PTB and measures of
fetal growth. Although both restricted fetal growth and PTB can result in LEW,
prematurity does not have to result in LEW or growth-restricted babies.
A major issue in studying environmental exposures and PTB is selecting the relevant
exposure period because the biological mechanisms leading to PTB and the critical
periods of vulnerability are poorly understood (Bobak. 2000). Short-term exposures
proximate to birth may be most relevant if exposure causes an acute effect. However,
exposure occurring in early gestation might affect placentation, with results observable
later in pregnancy, or cumulative exposure during pregnancy may be the most important
determinant. The studies reviewed have dealt with this issue in different ways. Many
have considered several exposure metrics based on different periods of exposure. Often
the time periods used are the first month (or first trimester) of pregnancy and the
last month (or 6 weeks) prior to delivery. Using a time interval prior to delivery
introduces an additional problem because cases and controls are not in the same stage of
development when they are compared. For example, a preterm infant delivered at
36 weeks is a 32-week fetus 4 weeks prior to birth, while an infant born at term
(40 weeks) is a 36-week fetus 4 weeks prior to birth.
6-120
-------
Recently, investigators have examined the association of PTB with both short-term
(i.e., hours, days, or weeks) and long-term (i.e., months or years) exposure periods.
Time-series studies have been used to examine the association between air pollution
concentrations during the days immediately preceding birth. An advantage of these
time-series studies is that this approach can remove the influence of covariates that do not
vary across individuals over a short period of time. Retrospective cohort and case-control
studies have been used to examine long-term exposure periods, often averaging air
pollution concentrations over months or trimesters of pregnancy.
Studies of PTB fail to show consistency in the periods during pregnancy when pollutants
are associated with an effect. For example, while some studies find the strongest effects
associated with exposures early in pregnancy, others report effects when the exposure is
limited to the second or third trimester. Many studies of PTB compare exposure in
quartiles, using the lowest quartile as the reference (or control) group. No studies use a
truly unexposed control group. If exposure in the lowest quartile confers risk, then it may
be difficult to demonstrate additional risk associated with a higher quartile. Thus,
negative studies must be interpreted with caution.
Preterm birth occurs both naturally (idiopathic PTB) and as a result of medical
intervention (iatrogenic PTB). Ritz et al. (2000) excluded all births by Cesarean section
to limit their studies to idiopathic PTB. No other studies attempted to distinguish the type
of PTB, although air pollution exposure may be associated with only one type. This is a
source of potential effect misclassification. One study examined preterm premature
rupture of membranes, observing positive ORs with NO2 exposure (Dadvand etal..
2014a).
A number of recent studies evaluated the association between exposure to NO2 and PTB,
and the results generally are inconsistent. The body of literature that observed an
association between NO2 or NOx and PTB (Gehring et al.. 2014; Trasande et al.. 2013;
Leetal.. 2012; Olssonet al.. 2012; Wuetal.. 201 la: Llop etal.. 2010: Darrow et al..
2009: Wu et al.. 2009: Jiang et al.. 2007: Leem et al.. 2006: Maroziene and
Grazuleviciene. 2002: Bobak. 2000) is generally the same (in both the quantity and
quality of studies) to the one that finds no consistent pattern in the association between
NO2 and PTB (Hannam etal.. 2014: Olssonet al.. 2013: Gehring etal.. 2011 a: Gehring et
al.. 201 Ib: Kashimaet al.. 2011: Basuet al.. 2010: Brauer et al.. 2008: Jalaludin et al..
2007: Ritz et al.. 2007: Hansen et al.. 2006: Liuet al.. 2003: Ritz etal.. 2000). Among the
studies that observe an association between exposure to NO2 and PTB, the association
seems to be strongest for exposure to NO2 late in pregnancy, including the third trimester
(Llop etal.. 2010: Leem et al.. 2006: Bobak. 2000). the last 8 weeks of pregnancy (Jiang
6-121
-------
et al.. 2007), the last 6 weeks of pregnancy (Darrow et al.. 2009). month of birth
(Trasande et al.. 2013). or the last week of pregnancy (Olsson et al.. 2012).
Several studies examined very preterm birth (VPTB, <30 weeks gestation). Some
observed positive associations with NO2 for VPTB when none were observed for PTB
(Brauer et al.. 2008). or observed stronger associations for NO2 or NOx with VPTB
compared to those for PTB fWuet al.. 201 la: Wuetal.. 2009).
Several studies of PTB estimated NO2 concentrations for subjects' homes or postal codes
with LURmodels (Gehring et al.. 2014; Gehring et al.. 201 la: Gehring etal.. 20lib:
Kashima et al.. 2011: Wuetal.. 201 la: Llop etal.. 2010: Brauer etal.. 2008) or
dispersion models (Wu et al.. 201 la). Results for an association between PTB and
exposure to NCh were inconsistent among studies that demonstrated that LUR models
predicted well ambient NO2 concentrations in the study areas (cross-validation R2 =
0.63-0.87). In general, epidemiologic studies that estimate long-term NO2 exposure from
central site monitors can carry uncertainty because the exposure error resulting from
spatial misalignment between subjects and monitor locations can overestimate or
underestimate associations with health effects (Section 3.4.5.2). Studies did not report on
the extent to which exposures estimated from central site monitors represented the spatial
pattern in ambient NC>2 concentrations. The results of studies that used NO2
concentrations from central site monitors were similarly inconsistent as those that
estimated residential NO2 exposure. Brauer et al. (2008) assigned exposure to NO2 using
both a LUR model and IDW of NO2 concentration from monitors and found comparable
risk estimates per 10 ppb for VPTB using the monitoring data (OR: 1.24, [95% CI: 0.80,
1.88]) andNO2 estimates from the LUR model (OR: 1.16 [95% CI: 0.93, 1.61]).
However, both LUR and IDW estimated exposures for subjects' postal code and may
have similar exposure error. In summary, the evidence for an association between NO2
exposure and PTB is inconsistent [Supplemental Table S6-4 (U.S. EPA. 2013i)1. Given
the differences among the study designs, it cannot be concluded that the inconsistencies
are related to exposure assessment method or length of follow-up periods.
6.4.3.3 Birth Weight
With birth weight routinely collected in vital statistics and being a powerful predictor of
infant mortality, it is the most studied outcome within air pollution-birth outcome
research. Air pollution researchers have analyzed birth weight as a continuous variable
and/or as a dichotomized variable in the form of LEW [<2,500 g (5 Ibs., 8 oz.)].
Birth weight is primarily determined by gestational age and intrauterine growth but also
depends on maternal, placental, and fetal factors as well as on environmental influences.
6-122
-------
In both developed and developing countries, LEW is the most important predictor for
neonatal mortality and is an important determinant of post-neonatal mortality and
morbidity. Studies report that infants who are smallest at birth have a higher incidence of
diseases and disabilities that continue into adulthood (Hack and Fanaroff. 1999).
A number of recent studies evaluated the association between exposure to NO2 and birth
weight, and the results generally are inconsistent. When examining birth weight as a
continuous variable, several studies observed decreases in birth weight associated with
increases in NC>2 exposure (Gehring et al.. 2014; Laurent et al.. 2014; Savitz et al.. 2014;
Darrowetal.. 20lib; Estarlichet al.. 2011; Ballester etal.. 2010; Morello-Frosch et al..
2010; Bell et al., 2007). Generally, these studies observed the largest decreases in birth
weight when exposure to NC>2 was averaged over the entire pregnancy. There are also a
number of studies that examined birth weight as a continuous variable that found no
consistent decreases in birth weight associated with increases in NO2 exposure averaged
over the entire pregnancy or specific trimesters of pregnancy (Hannam et al., 2014;
Sellier etal.. 2014; Pedersenetal.. 2013; Geer etal.. 2012; Rahmaliaetal.. 2012;
Gehring etal.. 201 la; Gehring etal.. 20 lib; Kashima etal.. 2011; Lepeule etal.. 2010;
Aguilera etal.. 2009: Hansen et al.. 2007; Salam etal.. 2005; Gouveia et al.. 2004). With
LEW examined as the risk of having a baby weighing less than 2,500 g, the study results
remain inconsistent, with some study authors observing an association between LEW and
exposure to NO2 (Dadvand et al.. 2014c; Ebisu and Bell. 2012; Ghosh etal.. 2012a:
Wilhelm etal.. 2012; Morello-Frosch et al.. 2010; Brauer et al.. 2008; Bell et al.. 2007;
Lee et al.. 2003). while others reporting no consistent association (Pedersen et al.. 2013;
Kashima etal.. 2011; Slama et al.. 2007; Salam etal.. 2005; Wilhelm and Ritz. 2005;
Gouveia et al.. 2004; Liu etal.. 2003; Maroziene and Grazuleviciene. 2002; Bobak.
2000). One study observed decreases in effect estimates for both LEW and change in
birth weight with increases in NC>2 exposure (Laurent et al.. 2013). Generally, the studies
that observed the largest risks for LEW averaged exposure to NO2 over the entire
pregnancy.
Some studies incorporated data on activity patterns in order to reduce potential error in
residential exposure estimates. Estarlich et al. (2011) only analyzed subjects who spent
15 or more hours per day at home or subjects who spent less than 2 hours a day in an
outdoor environment other than at their primary residence and found stronger
associations between birth weight and NO2 exposure. These sensitivity analyses did not
consistently change the associations observed by Aguilera et al. (2009). When Gehring et
al. (201 la) limited their analyses to participants who did not move during pregnancy, or
did not have paid employment outside of the home, they continued to observe no
consistent associations between birth weight and exposure to NC>2.
6-123
-------
Several birth weight studies estimated NO2 concentrations for subjects' homes or postal
code with LUR models, kriging of monitored concentrations (Ghosh etal.. 2012a:
Wilhelmetal.. 2012; Estarlich etal.. 2011; Gehring etal.. 201 la: Gehring etal.. 20lib:
Kashimaetal.. 2011: Ballester etal.. 2010: Lepeule etal.. 2010: Aguilera et al. 2009:
Brauer et al.. 2008: Slama et al.. 2007). or dispersion models (Rahmalia et al.. 2012: van
den Hooven et al.. 2012b: Madsen et al.. 2010). Results for an association with birth
weight are inconsistent among studies that demonstrated that LUR models predicted well
ambient NO2 concentrations in the study areas (cross-validation R2 = 0.63-0.87) or that
dispersion model estimates correlated well with monitored concentrations. In general,
epidemiologic studies that estimate long-term NO2 exposure from central site monitors
can carry uncertainty because the exposure error resulting from spatial misalignment
between subjects and monitor locations can overestimate or underestimate associations
with health effects (Section 3.4.5.2). Studies did not report on the extent to which
exposures estimated from central site monitors represented the spatial pattern in ambient
NO2 concentrations, and their results were similarly inconsistent as studies that estimated
residential NO2 exposure. Some studies compared the use of central site monitor data to
assign exposure to LUR or kriging, and concluded that while the monitoring data may
include larger errors in estimated exposure, these errors had little impact on the
association between exposure to NO2 and birth weight calculated using two exposure
assessment methods that different in spatial resolution (Lepeule et al.. 2010: Madsen et
al.. 2010). Given the differences among the study designs, it cannot be concluded that the
inconsistencies are related to exposure assessment method or length of follow-up periods.
In summary, epidemiologic evidence for an association between NC>2 exposure and birth
weight is generally inconsistent, with some studies observing an association, while other
studies observe no consistent pattern of association [Table 6-12 for key studies and
Supplemental Table S6-5 (U.S. EPA. 2013J) for all studies]. Further, supporting evidence
from toxicological studies is limited. Albino rats exposed to 1,300 ppb NO2 12 h/day for
3 months prior to breeding (Table 6-13) produced pups with statistically significant
decreased birth weights (Shalamberidze and Tsereteli. 197la. b). Statistically significant
body-weight decrements continued to be observed at postnatal day (PND) 4 and PND12.
6.4.3.4 Birth Defects
Despite the growing body of literature evaluating the association between ambient air
pollution and various adverse birth outcomes, relatively few studies have investigated the
effect of temporal variations in ambient air pollution on birth defects. Heart defects and
oral clefts have been the focus of the majority of these recent studies, given their higher
prevalence than other birth defects and associated mortality.
6-124
-------
A recent study investigated the association between NO or NO2 and cardiac birth defects
(Padula et al.. 2013a) and other noncardiac birth defects (Padula et al.. 2013b) in the San
Joaquin Valley in California. The authors observed no associations between heart defects
and NO or NO2 but did observe an association between neural tube defects and both NO
and NO2. In a further analysis of noncardiac/nonneural tube defects, Padula et al. (2013c)
observed no associations between NO or NO2 and any of the defects studied. A nine-state
cardiac birth defect case-control study observed associations between NO2 and
coarctation of the aorta, pulmonary valve stenosis, and left ventricular outflow tract
obstructions (Stingone et al., 2014). A Barcelona, Spain-based case-control study of
18 congenital anomaly groups found coarctation of the aorta and digestive system defects
associated with increases in NO2 (Schembari et al., 2014). Two studies examining
trisomy risk observed no correlations/associations with NO2, and a correlation between
NO and Trisomy 21 (Chung et al., 2014; Jurewicz et al., 2014). In general, studies of
birth defects have focused on cardiac defects, and the results from these studies are not
entirely consistent. This inconsistency could be due to the absence of true associations
between NO2 and risks of cardiovascular malformations; it could also be due to
differences in populations, pollution concentrations, outcome definitions, or analytical
approaches. Also, the lack of consistency of associations between NO2 and
cardiovascular malformations might be due to issues relating to statistical power or
measurement error. A recent meta-analysis of air pollution and congenital anomalies
observed elevated summary effects for NO2 and coarctation of the aorta (OR: 1.17 [95%
CI: 1.00, 1.36] per 10-ppb NO2), tetralogy of Fallot (OR: 1.20 [95% CI: 1.02, 1.42]), and
atrial septal defects (1.10 [95% CI: 0.91, 1.33]) (Vriiheidet al.. 2011). Ventral septal
defects exhibited an elevated summary estimate, but there also was high heterogeneity
between studies. Another meta-analysis found association only between coarctation of
the aorta and NO2 (Chenetal.. 2014a). These authors note that heterogeneity in the
results of these studies may be due to inherent differences in study location, study design,
and/or analytic methods, and comment that these studies have not employed some recent
advances in exposure assessment used in other areas of air pollution research that may
help refine or reduce this heterogeneity. Further, none of the birth defect studies of NO2,
including meta-analyses, examined confounding by PNfe.s or traffic-related pollutants.
These studies are characterized in Supplemental Table S6-6 (U.S. EPA. 2013k).
6.4.3.5 Early Life Mortality
An important question regarding the association between NO2 and infant mortality is the
critical window of exposure during development for which infants are at risk. Several age
intervals have been explored: neonatal (<1 month); post-neonatal (1 month to 1 year);
6-125
-------
and an overall interval for infants that includes both the neonatal and post-neonatal
periods (<1 year). During the neonatal and post-neonatal periods, the developing lung is
highly sensitive to environmental toxicants. The lung is not well developed at birth, with
80% of alveoli being formed postnatally. The studies below reflect a variety of study
designs, exposure periods, regions, and included many relevant potential confounders
except for traffic-related copollutants. As discussed below, a handful of studies have
examined the effect of ambient air pollution on neonatal and post-neonatal mortality,
with the former the least studied. These studies varied somewhat with regard to the
outcomes and exposure periods examined and study designs employed. Because it is
unknown whether early life mortality follows similar biological pathways or modes of
action to adult mortality, studies on adult mortality are discussed in Section 6.5.
Overall, the evidence for an association between exposure to NO2 and infant mortality is
inconsistent. Recent epidemiologic studies examined the association between long-term
exposure to NO2 measured at central site monitors and stillbirths, with one study
observing an association (Taizetal.. 2012) and another observing associations near the
null value (Hwang et al.. 2011). Faiz et al. (2013) observed positive ORs for stillbirth
with NCh exposures 2 days before birth. A case-control study of spontaneous abortion
before 14 weeks of gestation found a positive OR for NO2 exposure (Moridi etal. 2014).
Hou et al. (2014) also found positive ORs for fetal loss before 14 weeks of gestation with
NO2 exposure; however, these estimates have large confidence intervals and may not be
reliable. Enkhmaa et al. (2014) found high correlations between air pollutants and fetal
loss before 20 weeks in Mongolia including NO2, but these results were unadjusted for
other factors or copollutants. One study investigated the association between short-term
exposure to NO2 and mortality during the neonatal period (Lin et al.. 2004a) and did not
observe a positive association. More studies examined the association between exposure
to NO2 and mortality during the post-neonatal period. Son et al. (2008). Tsai et al. (2006).
and Yang et al. (2006) examined the association between short-term exposure to NO2 and
post-neonatal mortality, while Ritz et al. (2006) investigated the association between
long-term exposure to NO2 and post-neonatal mortality; none observed a consistent,
positive association. Finally, two studies examined the association between NO2 and
sudden infant death syndrome. Dales et al. (2004) and Ritz et al. (2006) observed positive
associations with short-term and long-term exposure to NO2, respectively. Supplemental
Table S6-8 (U.S. EPA. 2013m) provides a brief overview of the epidemiologic studies of
infant mortality.
Toxicological evidence for NO2-related infant mortality is similarly inconsistent.
Tabacova et al. (1985) exposed rodent dams to 50, 500, or 5,300 ppb NO2 (6 h/day,
7 days/week, GD 0-21; Table 6-13). Statistically significant decreased pup viability was
seen at PND21 with 5,300 ppb NO2. Another study in which male pups received prenatal
6-126
-------
exposure to NCh via a different daily exposure duration (dam continuous exposure to
1,500 or 3,000 ppb, GD 0-20) showed no statistically significant effects on pup postnatal
mortality to PND21 (Pi Giovanni et al., 1994).
6.4.4 Postnatal Development
The role of prenatal air pollution exposure has assumed increasing importance overtime
for effects on postnatal development. Ambient air pollution exposures of pregnant
women have been associated with negative birth outcomes. Additionally, the prenatal and
early postnatal periods are critical periods for extensive growth and development, and air
pollution exposures during this period have been linked to health effects in the first years
of life. Thus, air pollution-related effects in both the developing fetus and infant have
implications for effects on postnatal development. This evaluation of the relationship of
postnatal developmental with NO2 exposure consists primarily of neurodevelopmental
outcomes and limited toxicological information on physical development. Studies
examining the effects of NO2 exposure on development of the respiratory system
(Sections 6.2.5 and 6.2.6) inform the evaluation of the relationship between long-term
NO2 exposure and respiratory effects (Section 6.2.9).
6.4.4.1 Neurodevelopmental Effects
Epidemiologic studies of neurodevelopment in children were not available for the 2008
ISA for Oxides of Nitrogen (U.S. EPA. 2008c). but several have been published since
then. As described in the sections that follow, associations with NO2 are inconsistent for
cognitive function, which was most extensively examined, and for attention-related
behaviors, motor function, psychological distress, and autism, which were examined in a
few studies each. Table 6-12 details the key studies, and Supplemental Table S6-7 (U.S.
EPA. 20131) provides an overview of all of the epidemiologic studies of
neurodevelopmental effects. Strengths of the studies overall are assessment of
neurodevelopment with widely used, structured neuropsychological tests, spatial
alignment of ambient NO2 concentrations to subjects' school or home locations, and
examination of potential confounding by multiple SES indicators. While some studies
considered birth outcomes and traffic noise exposure as potential confounding factors,
smoking and stress were inconsistent or not considered. Further, in most studies, NC>2
was the only air pollutant examined, and uncertainty remains regarding potential
confounding by copollutants that are well-characterized risk factors for decrements in
neurodevelopmental function such as lead, PlVfc 5, or PM components such as polycyclic
aromatic hydrocarbons.
6-127
-------
Cognitive Function
NC>2 is not consistently associated with cognition in children. Among children age
4 years, indoor home NCh at age 3 months was associated with multiple measures of
cognitive function, from a general cognition index to memory, verbal, and quantitative
skills (Morales et al., 2009). These associations were limited to children with a GST Pi 1
valine (Val) -105 allele [isoleucine (Ile)/Val or Val/Val versus lie/lie genotype], which is
associated with lower oxidative metabolism. In contrast with indoor NC>2, ambient NCh
assessed concurrently with cognitive function or for the prenatal period was not clearly
associated with cognitive function in school children or infants (Guxens et al., 2014;
Clark etal.. 2012; Guxens etal.. 2012; van Kempen et al.. 2012; Freire etal.. 2010;
Wang etal.. 2009a). Within studies, results were inconsistent among the multiple indices
of cognitive function examined. Results also were inconsistent across studies, including
those for indices of memory, which was examined in most studies. As was done in the
2013 ISA for Lead (U.S. EPA. 2013c). evidence is evaluated separately for cognition in
school children and infants.
A common strength of the studies conducted in school children is the assessment of NC>2
exposures for home or school locations using well-validated LUR models
(Section 3.2.2.1). van Kempen et al. (2012) is particularly noteworthy for assessing
exposures outside both home and school and examining potential confounding by traffic
noise. The LUR model well predicted ambient NCh concentrations in the study area (R2
for cross-validation = 0.85). School, not home, NC>2 was associated with memory with
adjustment for noise (Table 6-12). Neither school nor home NC>2 was associated with the
ability to process information. A similar study found school-based aircraft noise, but not
NO2, to be associated with cognitive function (Clark etal.. 2012). NO2 estimated for
home locations also was inconsistently associated with cognitive function. Another study
of concurrent NCh exposure observed that higher home outdoor NC>2 was associated with
poorer cognitive function, but the wide 95% CIs call into question the reliability of
findings (Freire etal.. 2010). Associations for prenatal residential NO2 exposure were
similarly inconsistent among cohorts in three cities in Spain, with negative, null, and
positive associations observed with cognition [(Guxens et al.. 2014); Table 6-12]. Across
studies, NO2 exposures were estimated from LUR models that varied in performance
across locations; however, the inconsistency in findings does not appear to be related to
model performance. In Freire etal. (2010). the cross-validation R2 was 0.64 for urban and
nonurban areas combined. There was indication that the model was better for nonurban
areas, where 84% of subjects lived. Additionally, NO2 was not associated with
decrements in cognitive function in cities in Spain where LUR models predicted ambient
NC>2 concentrations well [cross-validation R2 = 0.75, 0.77; (Guxens et al., 2012; Estarlich
etal.. 2011)]. Wang et al. (2009a) also produced inconsistent findings but has weaker
6-128
-------
implications because of its ecological comparison of school locations that differed in
ambient NCh concentrations not direct analysis of NC>2.
The Bayley Scales of Infant Development is a widely used and reliable test for infant
development. The mental development index is a measure of sensory acuity, memory,
and early language skills. However, the Bayley Scales of Infant Development scores are
not necessarily correlated with development of children at older ages, and the tests at age
1 year or younger do not assess many outcomes that are analogous to those assessed at 2
and 3 years old. Across studies, mental development in infants ages 6 to 24 months was
not associated with trimester-specific prenatal NO2 exposure or NO2 exposure from birth
to age 6 months (Linet al.. 2014) and was inconsistently associated with NC>2 averaged
over pregnancy (Guxens et al.. 2014; Kim et al.. 2014; Guxens et al.. 2012). In an
analysis of cohorts across multiple European countries, associations were found in
locations where the LUR model better predicted ambient NO2 [cross-validation
R2 = 0.69-0.84 versus 0.45 or 0.51; (Beelenet al.. 2013; Estarlich et al.. 2011)1.
Associations for four cohorts in Spain differed between publications [(Guxens etal..
2014; Guxens etal.. 2012); Table 6-121. Other than sample sizes and possibly different
pregnancy addresses used to estimate NO2 exposure, an explanation for divergent results
is not clear (Guxens et al.. 2014; Estarlich et al.. 2011). Mental development of infants
also was inconsistently associated with NO2 exposure assessed by averaging
concentrations across central site monitors within a city (Lin etal.. 2014) or combining
concentrations by IDW (Kim et al.. 2014). Kim etal. (2014) illustrated the uncertainty of
the IDW method in capturing the spatial heterogeneity in ambient NC>2 concentrations in
their study area by indicating a moderate or weak correlation (r = 0.42, 0.21) between
ambient NO2 concentrations estimated by IDW and measured outside homes in a subset
of subjects. Another uncertainty in these studies is the lack of examination of potential
confounding by benzene (Guxens et al.. 2012) or PMio (Kim etal.. 2014). which showed
a similar pattern of association as did NO2 and were highly or moderately correlated with
NO2 (r = 0.70 and 0.40, respectively).
Attention-Related Behaviors
The few studies of attention-related behaviors produced contrasting results for
associations with NCh. Morales et al. (2009) observed that exposure to gas appliances
and higher indoor NC>2 at age 3 months were associated with elevated odds ratios for
symptoms of Attention Deficit Hyperactivity Disorder at age 4 years. The association
was attributable mainly to inattention, as hyperactivity was not associated with NC>2. As
was observed for cognitive function, associations were limited to children with a
GST Pi 1 Val-105 allele. In contrast, outdoor school and home NCh concentrations (with
or without adjustment for road traffic and aircraft noise) were not associated with poorer
6-129
-------
performance on multiple tests of sustained and switching attention (van Kempen et al..
2012). There was some evidence of an MVroad traffic noise interaction, as home
outdoor NO2 was associated with poorer attention switching among children in the
highest noise category. Home NC>2 and road traffic noise were moderately correlated
(r = 0.30). The ecological study did not find attention performance test results to differ
consistently with respect to school locations (Wang et al.. 2009a).
Motor Function
Evidence does not strongly indicate that NC>2 exposure affects motor function of children.
Whereas higher indoor home NC>2 exposure at age 3 months was associated with poorer
motor function in 4-year olds (Morales et al.. 2009). findings are inconsistent for ambient
NO2 exposure. A combined analysis of multiple European cohorts found thatNCh
exposure ascertained for the birth address by LUR was associated with poorer motor
function overall, but associations were limited to half of the individual cohorts (Guxens et
al.. 2014). PM25, PM2s absorbance (an indicator of EC), PMio, and coarse PM were not
associated with motor function in all of the same cohorts as NC>2. Thus, confounding by
these copollutants does not seem to fully explain the NO2 associations. There was no
clear pattern of association by gross or fine motor function or by age at which motor
function was assessed. NCh was associated with poorer motor function among children
ages 1-6 years in some locations but not others. The inconsistency in findings does not
appear to depend on the adequacy of the LUR models to represent ambient NO2
concentrations in the study area. LUR models showed a similar range in performance
[cross-validation^2 = 0.31 to 0.87; (Beelen etal.. 2013; Estarlich et al.. 2011)] in
locations where associations were observed and not observed. A limitation of the LUR
models is that they were constructed based on ambient concentrations measured after the
birth of some subjects. When the analysis was restricted to subjects whose birth dates
coincided with the period of ambient monitoring, the effect estimate decreased; however,
there was evidence of association between NO2 and motor function for the combined
cohorts (Table 6-12).
Results also are inconsistent for concurrent exposure. Among children ages 9-11 years,
neither concurrent NC>2 nor noise exposure, alone or combined, at school or home, was
associated with fine motor function (van Kempen et al.. 2012). Among children age
4 years, higher concurrent outdoor home NC>2 exposure was associated with poorer gross
motor function but not fine motor skills (Freire etal.. 2010). Children attending schools
with higher ambient NC>2 had poorer motor function compared to children attending
schools with lower NC>2 (Wang et al.. 2009a); however, attributing the findings to NO2
versus another factor that differed between schools is not possible. Like home- or
school-based exposure estimates, NO2 exposure assessed from central sites was not
6-130
-------
clearly associated with motor function in infants. Prenatal and lifetime exposure was
inconsistently associated with motor function at 6 months of age, and no associations
were observed in infants ages 12 to 24 months (Kim et al.. 2014; Lin et al.. 2014).
Limited evidence from toxicological studies also shows mixed effects of NO2 exposure
on motor function. Tabacova et al. (1985) found deficits in motor function and postural
gait in rat pups exposed gestationally to 50, 500, or 5,300 ppb NC>2 (Table 6-13; 6 h/day,
7 days/week, GD 0-20). In the open field test, female and male animals exhibited
retarded locomotor development, with stronger effects earlier in life (testing done until
3 months of age). On PND9, reductions were noted in horizontal motility and head
raising with prolonged periods of immobility, hypotonia, tremor, and equilibrium deficits.
Gait deficits including hindlimb dragging, crawling in lieu of walking, pivoting, and
impaired body raising ability were observed out to PND14 even in animals in the lowest
dose group. Tabacova et al. (1985) also found deficits in righting reflex and the auditory
startle reflex. In a separate study, prenatal exposure to 1,500 or 3,000 ppb NO2 [(Pi
Giovanni etal.. 1994); dam exposure GD 0-20] did not have a statistically significant
effect on motor function in 10- to 15-day-old male pups as measured by infrared sensors.
Psychological Distress
The two available studies produced equivocal evidence for the effects of NO2 exposure
on psychological distress. Among children ages 9-10 years, an index of emotional,
social, and conduct problems was not associated with concurrent NO2 or with aircraft or
road traffic noise at school, either alone or after mutual adjustment (Clark et al.. 2012). In
rats, Di Giovanni et al. (1994) reported that 3,000 ppb continuous NCh exposure of dams
during GD 0-21 resulted in decreased pup vocalization, an indicator of emotionality, in
males removed from the nest at PND5, PND10, or PND15.
Autism
Autism is a neurodevelopmental disorder characterized by impaired social interaction,
verbal and nonverbal communication deficits, and repetitive or stereotypic behavior.
Although the causes of autism are not fully understood, genetic conditions, family
history, and older parental age have been implicated as risk factors. Case-control studies
in California, U.S. observed that higher NO2 concentrations during the prenatal period
and during the first year of life were associated with higher odds ratios for autism in
children ages 24-71 months (Becerra et al.. 2013; Volk et al.. 2013). In both studies,
cases were identified from regional referral centers contracted by the Department of
Developmental Services. Controls were selected as birth certificate records not having a
matching record of autism with the referral centers. Controls were matched to cases by
6-131
-------
age, sex, and wide geographic area. However, matching by area of residence was based
only on birth addresses. These studies also observed stronger associations for autism
among children with mothers with less than a high school education compared with
higher education (Becerra et al.. 2013) and children with the CC MET genotype
compared with CC/GG genotype (Volk et al.. 2014). The CC MET genotype is associated
with decreased MET protein in the brain and has been associated with autism risk.
Between studies, inference about NO2 is stronger in (Becerra et al.. 2013). Residential
NO2 exposure was assessed using a well-validated LUR model [cross-validation
R2 = 0.87; (Suetal.. 2009a)]. In contrast, Volketal. (2013) examined central site
ambient NO2 concentrations and NOx estimated with a dispersion model. There are large
uncertainties with these exposure measures. Central site NO2 concentrations were
assigned from a site 5 km from homes, if available, or by IDW over a 50-km area. The
authors did not report what proportion of subjects were assigned NO2 exposures at the
5-km or 50-km scale, but neither scale may adequately capture the spatial heterogeneity
in NCh concentrations (Section 3.2.3). Inference is poor for NOx as it was nearly
perfectly correlated (r ~ 0.99) with EC and CO. In each study, PM2 5 also was associated
with autism, and (Becerra etal.. 2013) found that NO2 associations were robust to
adjustment for PM2 5 as well as the traffic-related copollutant CO. NO2 associations also
were robust to adjustment for Os or PMio. However, the reliability of the copollutant
model results is uncertain as copollutant concentrations were assessed from central sites,
and exposure measurement error likely varies between central site copollutant
concentrations and residential estimates of NO2.
Neuronal Degeneration and Nervous System Oxidative Stress
A recent study found that short-term NO2 exposure induced neuronal degeneration and
oxidative stress in the brains of adult male Wistar rats. Seven-day (6 h/day) exposure to
2,500-5,320 ppb NO2 (Li et al., 2012a) had no effect on body weight; however,
concentration-dependent reductions in brain-to-body-weight ratios were observed, with
statistically significant differences at 5,320 ppb NO2. Histopathological analysis of
cerebral cortex demonstrated a concentration-dependent increase in swollen or shrunken
nuclei and a concentration-dependent, statistically significant increase in apoptotic cell
number in all NO2-exposed rats. Statistically significant changes in antioxidant enzyme
activities [Cu/zinc (Zn) SOD, MnSOD, and GPx], protein carbonyls, and
malondialdehyde were observed in response to 5,320 ppb NO2. While rats exposed to
2,500 ppb NO2 demonstrated a statistically significant increase in the protein level of
p-53, rats exposed to the higher concentrations exhibited statistically significant increases
in mRNA and protein levels of c-fos, c-jun, p-53, and bax. These results, especially at
higher concentrations of NO2, are consistent with oxidative stress.
6-132
-------
6.4.4.2 Physical Development
Limited information from toxicological studies does not clearly indicate that NC>2
exposure affects physical postnatal development. Distinct exposure periods and test
endpoints produced disparate results in two studies on postnatal body-weight gain in pups
whose dams were exposed to NC>2. Albino rats with prenatal exposures to 1,300 ppb NO2
for 12 hours a day during the 3 months prior to breeding showed decrements in postnatal
body weight at PND4 and 12 in (Shalamberidze and Tsereteli. 197la. b). Continuous
NO2 exposure during gestation [continuous exposure of dam to 1,500 or 3,000 ppb NCh,
GD 0-20; (Pi Giovanni etal.. 1994)1 produced no statistically significant differences in
weight gain at PND1, 11, or 21 in male Wistar rat pups. Tabacova et al. (1985) saw
concentration-dependent delays in eye opening and incisor eruption in rodents after
maternal exposure to NO2 during pregnancy (dam exposure: GD 0-21, 5 h/day, 25, 500
or 5,300 ppb NO2) (Table 6-13).
6.4.4.3 Summary of Postnatal Development
The collective evidence does not consistently indicate a relationship between NO2
exposure and effects on postnatal development. Very few outcomes were similar between
epidemiologic and toxicological studies. Physical development, examined as postnatal
weight gain, eye opening, and incisor eruption in only a few studies of rats, was not
clearly affected by prenatal NO2 exposures in the range of 500 to 5,320 ppb. As
examined primarily in epidemiologic studies, prenatal, early life, or concurrent
school-age NO2 exposure was not consistently associated with cognitive function,
attention-related behaviors, motor function, or psychological distress in infants or school
children. While epidemiologic associations were observed for indoor home NO2 (Morales
et al., 2009). evidence is equivocal for ambient NCh, including exposure metrics spatially
aligned with subjects' home and school locations using LUR models that well
represented the spatial heterogeneity in the study areas (Guxens et al.. 2014; van Kempen
et al.. 2012; Freire et al.. 2010). In limited examination of children in California, U.S.,
autism was associated with residential prenatal NO2 exposure (Becerraetal.. 2013).
In addition to the inconsistent or limited evidence for NCh-related neurodevelopmental
effects, there is uncertainty regarding confounding by factors spatially correlated with
NC>2 at the level of individuals or communities. Studies observed associations with NO2
with adjustment for SES indicators and birth outcomes. However, analysis of
confounding was absent for stress and was very limited for smoking, noise, and
traffic-related copollutants. Neurodevelopmental effects were associated with noise,
PM2 5, and the traffic-related copollutants benzene and CO. There were observations of
6-133
-------
NCh-associated memory decrements with adjustment for traffic noise and NO2-associated
autism with adjustment for PM2 5 or CO. The reliability of the copollutant model results is
uncertain because of potential differential exposure measurement error between
residential NO2 and central site copollutant measurements. Other pollutants characterized
to be associated with neurodevelopment such as lead or polycyclic aromatic
hydrocarbons (U.S. EPA. 2013c. 2009b) were not examined as potential confounding
copollutants. Toxicological evidence is far more limited than epidemiologic evidence and
is similarly uncertain. Prenatal NO2 exposure induced psychological distress (Di
Giovanni et al., 1994) in rat offspring but showed mixed effects on motor function (Di
Giovanni et al.. 1994; Tabacova et al.. 1985). Further, although some studies showed
effects on postnatal development, there is very little information to propose possible
modes of action for ambient-relevant NO2 exposures. In a recent study of adult rats,
short-term exposure to 5,320 ppb NO2 induced increases in the neuronal apoptotic and
oxidative stress markers (Li etal.. 2012a). which have been linked to cognitive function.
Table 6-13 Characteristics of lexicological studies of nitrogen dioxide exposure
and reproductive and developmental effects.
Reference
Kripke and Sherwin
(1984)
Shalamberidze and
Tsereteli(1971a),
Shalamberidze and
Tsereteli(1971b)
Tabacova et al.
(1985)
Di Giovanni et al.
(1994)
Species
(Strain);
Sample Size;
Sex; Age
Rats (LEW/f
mai); n = 6; M;
young adult
Rats (albino);
n = 7; F; pup and
adult
Rats (Wistar);
n = 20; F; pup to
adult
Rats (Wistar);
n = 7; M; pup and
adult
Exposure Details
1 ,000 ppb NO2 for 7 h/day,
5 days/week for 21 days.
67 or 1,300 ppb NO2
Exposure of dams for 12 h/day
for 3 mo before pregnancy.
25, 50, 500, or 5,300 ppb NO2
Exposure of dams for 5 h/day
during GD 0-21
1,500 or 3, 000 ppb NO2
Exposure of dams continuously
GD 0-20
Endpoints Examined
Spermatogenesis, germinal cells
histology, testicular interstitial cell
histology.
Litter size, birth weight, postnatal weight
gain (body weight).
Pup viability, physical development (eye
opening, incisor eruption); neuromotor
(righting reflex, postural gait, geotaxis);
hepatic lipid peroxidation; hepatic
drug-metabolizing enzyme activity.
Progeny followed up to PND60.
Neurobehavior (ultrasonic vocalization),
maternal body weight during pregnancy,
litter size, postnatal body weight, early
life mortality, motor function. Progeny
tested for vocalizations on PND5,
PND10, andPND15.
F = female; GD = gestational day; h = hour; M = male; mo = month; NO2 = nitrogen dioxide; PND = postnatal day.
6-134
-------
6.4.5 Summary and Causal Determination
Overall, the evidence is suggestive of, but not sufficient to infer, a causal relationship
between exposure to NCh and birth outcomes and is inadequate to infer the presence or
absence of a causal relationship between exposure to NO2 and fertility, reproduction and
pregnancy as well as postnatal development. Separate conclusions are made for these
groups of reproductive and developmental effects because they are likely to have
different etiologies and critical exposure patterns over different lifestages. At the time of
the 2008 ISA for Oxides of Nitrogen, a limited number of epidemiologic and
toxicological studies had assessed the relationship between NO2 exposure and
reproductive and developmental effects. The 2008 ISA concluded that there was not
consistent evidence for an association between NO2 and birth outcomes and that evidence
was inadequate to infer the presence or absence of a causal relationship with reproductive
and developmental effects overall (U.S. EPA. 2008c). The change in the causal
determination for birth outcomes reflects the larger number of studies that observed
associations with fetal growth restriction and the improved outcome assessment
(e.g., measurements throughout pregnancy via ultrasound) and exposure assessment
(e.g., well-validated LUR models) employed by these studies. The key evidence as it
relates to the causal determinations is summarized in Table 6-14 using the framework
described in Table II of the Preamble to the ISA.
Fertility, Reproduction, and Pregnancy
Relationships of outcomes related to fertility, reproduction, and pregnancy with NO2
exposure have only recently been evaluated, and thus, the number of studies for any one
endpoint is limited. One study (Legro et al., 2010) observed a decreased odds of live birth
associated with higher NO2 concentrations during ovulation induction and the period
after embryo transfer; while another (Slamaet al., 2013) observed decreased
fecundability with higher NO2 exposure near conception. Both studies assessed NO2
exposure from central site monitors, and neither evaluated confounding by traffic-related
copollutants. NO2-exposed rats showed impaired estrus cyclicity and a decrease in
number of primordial follicles, which could indicate an effect on reproduction. There is
inconsistent evidence for an association of NO2 exposure estimated from well-validated
LUR models with pre-eclampsia or reduced placental growth and function. Toxicological
studies examined different effects on pregnancy and gave divergent results for
NO2-related maternal weight gain during pregnancy and reduced litter size. There is
generally no evidence for an effect of NO2 exposure on sperm quality in either
epidemiologic or toxicological studies. Collectively, the limited evidence is of
6-135
-------
insufficient consistency and is inadequate to infer a causal relationship between NO2
exposure and effects on fertility, reproduction, and pregnancy.
Birth Outcomes
While the collective evidence for many of the birth outcomes examined is not entirely
consistent, there are several well-designed, well-conducted studies that indicate an
association between NO2 and poorer birth outcomes, particularly fetal growth restriction.
For example, the Spanish cohort that used anthropometric fetal measurements throughout
pregnancy (liiiguez et al.. 2012; Estarlich et al.. 2011; Aguilera et al.. 2010; Ballester et
al.. 2010) observed small, yet consistent associations with impaired fetal growth. Some
studies of fetal growth restriction demonstrated that LUR models used to estimate NO2
exposure predicted well ambient NO2 concentrations in the study areas. Studies that
examined PTB, birth weight, birth defects, and infant mortality had inconsistent results
(i.e., positive and null associations). For PTB, many associations very close to the null
value. Several different methods for exposure assessment were used in studies of birth
outcomes, and results were inconsistent across methods. Generally, studies of birth
outcomes did not evaluate confounding by PIVb 5 or other traffic-related pollutants,
resulting in uncertainty in an independent effect of NCh exposure. NCh-related
decrements in rat pup birth weight were reported in a toxicology study (Shalamberidze
and Tsereteli. 1971a). which provides limited support for the associations with fetal
growth restriction observed in epidemiologic studies. Collectively, the epidemiologic
evidence for fetal growth restriction but uncertainty regarding an independent effect of
NO2 exposure is suggestive of, but not sufficient to infer, a causal relationship between
NO2 exposure and effects on birth outcomes.
Postnatal Development
There is inconsistent evidence from both epidemiologic and toxicological studies for a
relationship between prenatal and childhood NO2 exposure and effects on postnatal
development. Findings across the several recent epidemiologic studies of
neurodevelopment do not consistently support associations of NC>2 with cognitive
function, attention-related behaviors, motor function, or psychological distress in
children. Many of these studies estimate ambient NO2 exposures for children's homes or
schools using LUR models that predicted well ambient NO2 concentrations in study
areas. NO2 exposures were related to autism in children in recent epidemiologic studies,
but such findings are limited to a few studies. In the small group of epidemiologic studies
observing associations with neurodevelopmental effects, examination of confounding by
noise, stress, or traffic-related copollutants was absent or unreliable. Toxicological
6-136
-------
evidence for effects on neurodevelopment also is limited and mixed. NO2 exposure
impaired vocalization of rat pups in one study but did not affect motor function (e.g.,
startle and righting reflex, postural gait, impaired walking, head raising) consistently.
Toxicological evidence for impaired physical development as well is mixed. Collectively,
the evidence is of insufficient consistency or quantity and is inadequate to infer a causal
relationship between exposure to NO2 and effects on postnatal development.
Table 6-14 Summary of evidence supporting the causal determinations for
relationships between long-term nitrogen dioxide exposure and
reproductive and developmental effects.
Rationale for Causal
Determination3
Key Evidence13
Key References'3
NO2 Concentrations
Associated with Effects0
Fertility, Reproduction, and Pregnancy—Inadequate to Infer a Causal Relationship
Inconsistent Inconsistent associations when
epidemiologic NO2 exposure is assessed
evidence for across entire pregnancy and
pre-eclampsia after adjustment for many
potential confounders.
Uncertainty regarding potential
confounding by traffic-related
copollutants.
tDadvand et al. (2013),
tvan den Hooven et al.
(2011), tPereira et al.
(2013)
Section 6.4.2.3
Association with mean
31 ppb for trimester avg
No association with means
21.2, 23.0, and 30 ppb for
pregnancy-avg
Inconsistent
epidemiologic and
toxicological evidence
for other
pregnancy-related
effects
Limited and inconsistent
epidemiologic evidence for
associations with
pregnancy-induced
hypertension and placental
growth and function.
tHampel et al. (2011),
tLeeetal. (2012b).
tMobasher et al. (2013),
tXuetal. (2014), tvan den
Hooven etal. (2012b)
Section 6.4.2.3
Inconsistent associations
Mean: 10.9, 11.5 ppb for
7-day avg
Means: 18.7-30 ppb for
trimester-avg
Mean: 21.2 for
pregnancy-avg
No effect in rats on maternal
weight gain during pregnancy.
Pi Giovanni etal. (1994)
Section 6.4.2.3
1,500 and 3,000 ppb for 21
days
Inconsistent
epidemiologic
evidence for in vitro
fertilization failure
Decreased odds of live birth
associated with higher NO2
concentrations during ovulation
induction and the period after
embryo transfer.
tLegroetal. (2010),
tSlamaetal. (2013)
Section 6.4.2.2
Mean: 19 ppb for 3.6-to
200-day avg
Median: 19 ppb for 1 mo-avg
Lack of supporting
epidemiologic and
toxicological evidence
for effects on sperm
Limited number of toxicological Rats: Kripke and Sherwin
and epidemiologic studies (1984)
provide no evidence for effects Humans: tRubes etal.
on sperm count or motility, (2010), Sokol et al. (2006)
spermatogenesis. Sectjon 6421
Rats: 1,000 ppb for 21 days
Humans: Mean 16.8 ppb for
90-day avg, 30.1 for24-h
avg
6-137
-------
Table 6-14 (Continued): Summary of evidence supporting the causal
determinations for relationships between long-term
nitrogen dioxide exposure and reproductive and
developmental effects.
Rationale for Causal
Determination3
Key Evidence13
Key References13
NO2 Concentrations
Associated with Effects0
Limited evidence for
key events in
proposed mode of
action
Impaired estrus cyclicity and
decreased number of
primordial follicles in rats.
Shalamberidze and
Tsereteli(1971a)
Section 6.4.2.2
1,300 ppb for 3 mo before
pregnancy
Birth Outcomes—Suggestive of, but Not Sufficient to Infer, a Causal Relationship
Evidence from multiple
epidemiologic studies
of fetal growth
restriction is generally
supportive but not
entirely consistent
Strongest evidence from
well-conducted Spanish cohort
studies that observe
associations with fetal growth
restriction. Some demonstrate
LUR model to predict well
ambient NO2 concentrations.
Supported by consistent
evidence for SGA and IUGR.
Outcomes assessed with
anthropometric fetal
measurements.
tAquilera et al. (2010).
tlniquez et al. (2012),
tEstarlich et al. (2011),
tBallesteretal. (2010)
Section 6.4.3.1
Mean trimester-avg:
7.8-36.1 ppb
Limited and
inconsistent
epidemiologic
evidence for other birth
outcomes
Some studies observe an
association with PTB, birth
weight, birth defects, and infant
mortality while other studies
observe no consistent pattern
of association.
Section 6.4.3.2,
Section 6.4.3.3,
Section 6.4.3.4,
Section 6.4.3.5
Mean trimester-avg (PTB):
8.8-37.6 ppb
Mean trimester-avg (birth
weight): 6.2-62.7 ppb
Mean early pregnancy avg
(birth defects): 8.2-28.0 ppb
Mean 24-h avg (infant
mortality): 20.3-50.3 ppb
Limited and
inconsistent
toxicological evidence
with relevant NO2
exposures
Mixed evidence for effects on
litter size and late embryonic
lethality in rats.
Shalamberidze and
Tsereteli (1971 a), Di
Giovanni et al. (1994),
Tabacova et al. (1985)
Section 6.4.3.3 and
Section 6.4.3.5
1,300 ppb for 3 mo,
500 and 5,300 ppb for 20
days, 1,500 and 3,000 ppb
for 21 days
Weak evidence for key
events in proposed
mode of action
Inflammation
Increase in C-reactive protein
concentration in human
umbilical cord blood but not
maternal blood.
tvan den Hooven et al.
(2012a)
Section 6.4.2.3
Mean 1-week avg before
delivery: 21.4 ppb
6-138
-------
Table 6-14 (Continued): Summary of evidence supporting the causal
determinations for relationships between long-term
nitrogen dioxide exposure and reproductive and
developmental effects.
Rationale for Causal
Determination3
Key Evidence13
Key References13
NO2 Concentrations
Associated with Effects0
Postnatal Development—Inadequate to Infer a Causal Relationship
Limited and
inconsistent
epidemiologic and
toxicological evidence
for effects on
neurodevelopment
Some but not all epidemiologic
studies showed associations
with cognitive function
decrements in infants and
schoolchildren.
Inconsistent evidence NO2
exposures estimated for
children's homes or schools
using LUR models that predict
well study area ambient NO2
concentrations.
Uncertainty regarding potential
confounding by traffic-related
copollutants.
tvan Kempen et al. (2012),
tMorales et al. (2009).
tGuxensetal. (2012)
No association:
tClarketal. (2012), fFreire
et al. (2010). tGuxens et al.
(2014)
Section 6.4.4.1
Mean concurrent:
16.5, 16.4 ppb
Mean prenatal:
15.7 ppb
More limited and inconsistent
epidemiologic evidence for
attention-related behaviors,
motor function, psychological
distress.
Section 6.4.4.1
Increased emotionality in rat
pups, but effects on motor
function inconsistent with
prenatal exposure.
Tabacova et al. (1985), Pi
Giovanni et al. (1994)
Section 6.4.4.1
50-5,300 ppb for 20 days,
1,500 and 3,000 ppb for 21
days
Prenatal NO2 exposure
associated with early childhood
autism in California, U.S.
tBecerra etal. (2013)
Section 6.4.4.1
Mean: 30.8 ppb
Limited evidence for key
events in proposed mode of
action. Increased apoptotic
factors and oxidative stress in
brain of adult rats.
tLietal. (2012a)
Section 6.4.4.1
2,500 and 5,320 ppb for
7 days
Limited and
inconsistent
toxicological evidence
for physical
development
Delayed postnatal eye opening
and incisor eruption but mixed
effects on postnatal growth.
Tabacova et al. (1985). Pi 500 and 5,300 ppb for 20
Giovanni etal. (1994) days, 1,500 and 3,000 ppb
Section 6.4.4.2 for 21 days
avg = average; NO2 = nitrogen dioxide; PTB = preterm birth; SGA = small for gestational age; IUGR = intrauterine growth
restriction.
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Tables I and N. of the
Preamble.
Describes the key evidence and references, supporting or contradicting, contributing most heavily to causal determination and,
where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where full body of evidence is
characterized.
°Describes the NO2 concentrations with which the evidence is substantiated (for experimental studies, £ 5,000 ppb).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
6-139
-------
6.5 Total Mortality
6.5.1 Introduction and Review of Evidence from 2008 Integrated Science
Assessment for Oxides of Nitrogen
At the time of the 2008 ISA for Oxides of Nitrogen, a limited number of epidemiologic
studies had assessed the relationship between long-term exposure to NO2 and mortality in
adults, including cause-specific and total mortality. The 2008 ISA concluded that the
evidence was "inadequate to infer the presence or absence of a causal relationship" (U.S.
EPA. 2008c). In this ISA, findings for cause-specific mortality (i.e., respiratory,
cardiovascular) are used to assess the continuum of effects and inform the causal
determinations for respiratory and cardiovascular effects. The causal determination for
total mortality contained herein (Section 6.5) is based primarily on the evidence for
nonaccidental mortality but also is informed by the extent to which evidence for the
spectrum of cardiovascular and respiratory effects provides biological plausibility for
NO2-related total mortality. The exposure assessment method was an important
consideration in the evaluation of long-term exposure and mortality, given the spatial
variability typically observed in ambient NC>2 concentrations (Section 2.5.3). Exposure
assessment was evaluated drawing upon discussions in Section 3.2 and Section 3.4.5.
Several recent studies of long-term exposure to NO2 and mortality employed exposure
assessment methods to account for the spatial variability of NC>2. For example, LUR
model predictions have been found to correlate well with outdoor NO2 concentration
measurements (Section 3.2.2.1). For long-term NC>2 exposure, exposure assessment was
evaluated by the extent to which the method represented the spatial variability in NO2
concentrations in a given study. Supplemental Table S6-9 (U.S. EPA. 2013n) provides an
overview of the epidemiologic studies of long-term exposure to NCh or NOx and
mortality, including details on exposure assessment and mean concentrations from the
study locations.
Two seminal studies of long-term exposure to air pollution and mortality among adults
have been conducted in the United States; the American Cancer Society (ACS) and the
Harvard Six Cities (HSC) cohorts have undergone extensive independent re-analyses and
have reported extended results including additional years of follow-up. The initial reports
from the ACS (PopeetaL 1995) and the HSC (Dockery et al. 1993) studies did not
include results for gaseous pollutants. However, as reported in the 2008 ISA for Oxides
of Nitrogen (U.S. EPA, 2008c). in re-analyses of these studies, Krewski et al. (2000)
examined the association between gaseous pollutants, including NO2, and mortality.
Krewski etal. (2000) observed a positive association between long-term exposure to NO2
6-140
-------
and mortality in the HSC cohort, with effect estimates1 similar in magnitude to those
observed with PM2 5 The effect estimates were positive for various causes of mortality
but were the strongest for cardiopulmonary and total mortality. In a re-analysis of the
ACS cohort data (Krewski et al.. 2000). long-term exposure to NO2 estimated from
central site monitors was not associated with mortality. An extended study of the ACS
cohort (Pope et al.. 2002) doubled the follow-up time and tripled the number of deaths
compared to the original study but still observed no association between long-term
exposure to NCh and mortality.
A series of studies (Lipfert et al.. 2006a: Lipfert et al.. 2006b: Lipfert et al.. 2003. 2000)
characterized a national cohort of over 70,000 male U.S. military veterans who were
diagnosed as having hypertension in the mid-1970s and were followed through 2001. In
the earlier studies, the authors reported increased risk of mortality associated with
exposure to NCh; these excess risks were in the range of 5-9% (Lipfert et al.. 2003.
2000). In the later studies, the authors focused on traffic density in this cohort. Lipfert et
al. (2006b) and Lipfert et al. (2006a) reported that traffic density was a better predictor of
mortality than NC>2, though they still observed a positive association between mortality
and NC>2 exposure. The results from the series of studies characterizing the Veterans
cohort are indicative of a traffic-related air pollution effect on mortality, but the study
population (lower SES, males with hypertension and a very high smoking rate) was not
representative of the general U.S. population.
In another cohort conducted in the U.S. [the California Seventh-Day Adventist cohort
(AHSMOG)], Abbey etal. (1999) enrolled young adult, nonsmoking Seventh-Day
Adventists throughout California. Generally, NC>2 was not associated with total,
cardiopulmonary, or respiratory mortality in either men or women. The authors observed
large risk estimates for lung cancer mortality for most of the air pollutants examined,
including NCh, but the number of lung cancer deaths in this cohort was very small (12 for
females and 18 for males out of a total of 5,652 subjects); therefore, it is difficult to
interpret these results.
Several studies conducted in European countries have examined the relationship between
long-term exposure to traffic-related pollutants (including NC>2 and NOx) and mortality
among adults. Hoek et al. (2002) observed an association between NO2 and mortality in
the Netherlands Cohort Study on Diet and Cancer (NLCS), though the association with
living near a major road was stronger in magnitude. On the other hand, Gehring et al.
(2006) observed that NO2 was generally more strongly associated with mortality than an
1 Quantitative effect estimates from studies reviewed in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 20080) can
be found alongside effect estimates from more recent studies in Figure 6-8. Figure 6-9. and Figure 6-10 (and in
corresponding Table 6-15. Table 6-16. and Table 6-17 respectively).
6-141
-------
indicator for living near a major road in a cohort of women from Germany. Results from
the Air Pollution and Chronic Respiratory Diseases survey conducted in France,
demonstrated increased risk between long-term exposure to NO2 and total,
cardiopulmonary, and lung cancer mortality (Tilleul et al.. 2005). Similarly, Nafstad et al.
(2004) observed an association between NOx and total mortality, as well as deaths due to
respiratory causes, lung cancer, and ischemic heart disease in a cohort of Norwegian men.
Nyberg et al. (2000) observed similar results for lung cancer mortality in a case-control
study of men in Stockholm, Sweden. Naess et al. (2007) investigated the
concentration-response relationships between NO2 and cause-specific mortality among a
cohort from Oslo, Norway, aged 51-90 years. Total mortality, as well as death due to
cardiovascular causes, lung cancer, and COPD were associated with NO2 for both men
and women in two different age groups, 51-70 and 71-90 years. Naess et al. (2007)
reported that the effects appeared to increase at NO2 levels higher than 21 ppb in the
younger age group (with little evidence of an association below 21 ppb), while a linear
effect was observed between 10 and 31 ppb in the older age group.
The results from these studies led to the conclusion that the evidence was inadequate to
infer the presence or absence of a causal relationship in the 2008 ISA for Oxides of
Nitrogen (U.S. EPA. 2008c). The 2008 ISA also noted that potential confounding by
copollutants was an important uncertainty when interpreting the evidence for the
association between long-term exposure to NO2 and mortality. Collinearity among
criteria pollutants is another uncertainty that needs to be considered; several studies
reported moderate-to-high correlations between NO2 and PM indices (i.e., >0.5). The
2008 ISA acknowledged that NO2 could be serving primarily a surrogate or marker for
traffic-related pollution. These uncertainties do not preclude the possibility of an
independent effect of NO2 or of NO2 playing a role in interactions among traffic-related
pollutants.
6.5.2 Recent Evidence for Mortality from Long-Term Exposure to Oxides of
Nitrogen
Several recent studies provide extended analyses of existing cohort studies of adult
populations. Because it is unknown whether early life mortality follows similar biological
pathways or modes of action to adult mortality, studies on early life mortality are
discussed in Section 6.4.3. In a re-analysis that extended the follow-up period for the
ACS cohort to 18 years (1982-2000), Krewski et al. (2009) reported generally null
associations between long-term exposure to NO2 estimated from central site monitors and
total and cause-specific mortality, similar to what was reported in the initial reanalysis of
this cohort (Krewski et al.. 2000). In an update to the ACS study including cohort
6-142
-------
members residing in California, Jerrettet al. (2013) estimated long-term (i.e., 15 years)
exposures to NCh at the home addresses of each of the cohort members using LUR
models that predicted well ambient NO2 concentrations in the study area (cross-validation
R2 = 0.71). The authors observed positive associations between predicted NO2 exposures
and total, CVD, IHD, stroke, and lung cancer mortality, but not for respiratory mortality.
The strongest associations were observed for deaths due to lung cancer and stroke. The
associations with CVD and IHD mortality were attenuated in copollutant models that
included PIVb 5 (also estimated from an LUR model with similar predictive capacity),
while the association with lung cancer was generally unchanged in copollutant models. In
an update to the Veterans cohort study, Lipfert etal. (2009) looked at markers for specific
emission sources, including NOx as a marker of traffic, and their relationship with
mortality, using a 26-year follow-up period now available for this cohort. The authors
observed an association between long-term exposures to NOx estimated from a plume-in-
grid model and mortality, and noted that this association was stronger among men living
in areas with high traffic density compared to men living in areas with lower traffic
density. The authors also demonstrate that traffic-related air pollutants (including NOx)
are better predictors of mortality than a measure of traffic density in this cohort. Updated
results also were reported for the NLCS cohort [the same effect estimates are reported by
both Beelen et al. (2008b) and Brunekreef etal. (2009)]. Consistent with previous results
from this cohort, the authors observe an association with total mortality. In the updated
results, the authors observe the strongest effect between long-term exposure to NO2
estimated from central site monitors and respiratory mortality; this association is stronger
than any observed with the traffic variables and total or cause-specific mortality.
In updates to a cohort of women in Germany (Gehring et al.. 2006). Schikowski et al.
(2007) observed a positive association between ambient NO2 concentrations measured at
central site monitors and cardiovascular mortality among older women, though this
association was not modified by lung function status (i.e., normal versus impaired lung
function). Heinrich et al. (2013) included five additional years of follow-up and twice as
many fatalities compared to the original analysis. In the updated analyses, the authors
observed positive associations between NO2 concentrations measured at central site
monitors and total and cardiopulmonary mortality. The effect estimates were highest for
women living within 50 m of a road with median daily traffic volume of 5,000 cars or
greater. The effect estimates for the associations between total and cardiopulmonary
mortality and NO2 were generally lower for the follow-up period compared to the
original analysis.
Several recent U.S. cohort studies examined the association between long-term exposures
to NO2 and mortality in occupational cohorts. Hart etal. (2011) examined the association
between residential exposure to NO2 estimated from a spatial smoothing model and
6-143
-------
mortality among men in the U.S. trucking industry in the Trucking Industry Particle
Study (TrIPS). The authors observed an increase in cardiovascular disease mortality and
a decrease in COPD mortality associated with NO2 exposure. The association between
NC>2 exposure and total mortality was robust to the inclusion of PMio or SCh in
copollutant models. This association was stronger when the cohort was restricted to truck
drivers that maintained local routes, and long haul drivers were excluded. COPD
mortality was positively associated with NO2 exposure in the sensitivity analysis
excluding long haul drivers. The associations for other causes of death (i.e., lung cancer,
IHD, respiratory disease) were generally positive. Another recent U.S. cohort study, The
California Teachers Study (Lipsett et al.. 2011) examined the association between
long-term exposure to NOx and NC>2 measured at central site monitors and mortality
among current and former female public school teachers. The authors observed the
strongest associations between IHD mortality and exposure to NOx and NO2; the
associations for other causes of death (i.e., CVD, cerebrovascular, respiratory, lung
cancer, and total) were less consistent and generally close to the null value. Hart et al.
(2013) examined the association between long-term exposure to NO2 and total mortality
among a cohort of female nurses in the Nurses' Health Study. The authors used spatial
modeling to estimate exposure to NO2 and observed a small increase in the risk of total
mortality. In a sensitivity analysis examining women that moved during study follow-up,
the authors observed even higher risks among women that moved to areas with higher
concentrations of NO2.
A number of recent studies examined the association between long-term exposure to NO2
and mortality in Canadian cities. All three studies estimated long-term NO2 exposure
using well-validated LUR models (predicted versus measured: R2 = 0.90 for Toronto,
Ontario, R2 = 0.69, 0.44 for Vancouver, British Columbia, 1-7% difference for Hamilton,
Ontario, 4% difference for Windsor, Ontario). Chen etal. (2013a) conducted a cohort
study in three cities in Ontario and observed that long-term exposure to NO2 was
associated with an increased risk of cardiovascular mortality. The association was
stronger when mortality from IHD was evaluated separately. In a single-city study
conducted in Toronto, Ontario, Jerrettetal. (2009) examined subjects from a respiratory
clinic and observed positive associations with total and circulatory mortality. The
associations with respiratory and lung cancer mortality were also positive, though less
precise. In a model that included both NO2 and proximity to traffic, the effect estimate for
NO2 remained robust, and the effect attributable to traffic was attenuated. Gan et al.
(2013) and Gan etal. (2011) conducted a single-city, population-based cohort study in
Vancouver, British Columbia to evaluate the association between traffic-related
pollutants and risk of mortality due to CHD and COPD, respectively. LUR models were
used to estimate exposure over a 5-year period, (1994-1998) and the cohort was followed
for 4 years (1999-2002). The authors observed the strongest associations (i.e., highest
6-144
-------
magnitude) for exposures to NO2 and CHD mortality; however, these associations were
greatly attenuated when PM2 5 or BC were included in the model. The correlations
between NCh and PM2 5 and BC were low to moderate (r <0.5), and LUR models for
PM2 5 and BC showed poorer predictive accuracy. The authors observed positive
associations between both NO and NO2 concentrations and COPD mortality, which were
slightly attenuated when PM2 5 or BC were included in the model.
A recent multicenter European study pooled data from 22 existing cohort studies and
used a strictly standardized protocol to investigate the associations between long-term
concentrations of NC>2 and NOx and total (Beelen et al.. 2014a). respiratory
(Dimakopoulou et al., 2014). and cardiovascular (Beelen et al.. 2014b) mortality. The
authors used LUR models to assign exposure and observed generally null associations
with total, respiratory, and cardiovascular mortality. For most cohorts, LUR models were
shown to predict well ambient NC>2 concentrations (cross-validation R2 = 0.46-0.87). In
copollutant models, the total mortality null associations did not change after adjustment
for PM2 5 or PMio-2.5.
Several studies examined the association between long-term exposure to NO2 and
mortality in England. Carey etal. (2013) conducted a cohort study using an
emissions-based model to assign exposure. Model validation was good (cross-validation
R2 = 0.57-0.80), and model estimates for NO2 were highly correlated with PMio and
PM2 5 (r = 0.9). The authors observed positive associations with total mortality; these
associations were stronger for respiratory and lung cancer deaths, and somewhat
attenuated when restricted to cardiovascular deaths. Tonne and Wilkinson (2013)
evaluated the association between long-term exposure to NCh and NOx estimated from a
Gaussian dispersion model among survivors of hospital admissions for acute coronary
system in England and Wales and observed evidence of a null association after
adjustment for PIVbs. In a single-city study, Maheswaran et al. (2010) compiled a cohort
of stroke survivors and estimated NO2 exposures from an emissions model across
London, U.K. Modeled NO2 concentrations were well correlated with measurements at
56 sites (r = 0.91). The authors observed a nearly 30% increase in total mortality per 10-
ppb increase in exposure to NO2.
Rome, Italy was the setting for a number of single-city cohort studies. Cesaroni etal.
(2013) observed positive associations between long-term exposure to NO2 estimated from
an LUR model and total, cardiovascular, IHD, respiratory, and lung cancer mortality
among the adult population in the Rome Longitudinal Study (RoLS). As examined for
total mortality, the association was robust to the inclusion of PM2 5 in the model. The
modeled estimates for both NO2 (cross-validation R2 = 0.61) and PM2 5 (r = 0.83) agreed
well with measured concentrations in the study area, but NO2 and PM2 5 were highly
6-145
-------
correlated (r = 0.79). Later, these authors used several different LUR models to predict
NO2 in Rome, Italy (Cesaroni et al.. 2012) and observed that the modest, positive
association between total mortality and NO2 concentrations was consistent across all
models evaluated. Rosenlund et al. (2008b) conducted a cohort study in Rome, Italy to
investigate the effects of long-term exposure to NCh and cardiovascular deaths, including
mortality among previous MI survivors. The authors observed a positive association
between long-term exposures to NO2 estimated from an LUR model and fatal coronary
events, though they did not observe an association with mortality among survivors of a
first coronary event. Modeled NCh estimates correlated well (r = 0.77) with
measurements, but exposures were estimated at the level of census blocks.
A Danish study evaluated the association between long-term exposure to NO2 (estimated
from a dispersion model) and diabetes-related mortality (Raaschou-Nielsen et al.. 2013b).
The authors reported a 30% increase in risk of diabetes-related mortality associated with
NC>2 concentrations. In Brisbane, Australia, Wang et al. (2009b) examined the association
between long-term exposure to NO2 estimated from central site monitors and
cardiorespiratory mortality. The relative risk for NC>2 and cardiorespiratory mortality was
near the null value.
A number of studies were conducted in Asian countries to evaluate the association
between long-term-exposure to NO2 and mortality. In a national study covering
16 provinces in eastern China, Cao etal. (2011) observed positive associations between
ambient NOx concentrations from central site monitors and total, cardiovascular,
respiratory, and lung cancer mortality. The association between total mortality and NOx
was relatively unchanged in a copollutant model with total suspended particles (TSP) but
was reduced by half in a copollutant model with 862. The associations between NOx and
cardiovascular, respiratory, and lung cancer mortality were all attenuated in copollutant
models including either TSP or SO2. In a single-city study in Shenyang, China, the
authors observed a strong, positive association between long-term exposure to NO2
estimated from central site monitors and respiratory (Dong etal.. 2012). total,
cardiovascular, and cerebrovascular (Zhang etal.. 2011) mortality. In Shizuoka, Japan,
Yorifuji et al. (2010) observed positive associations between NO2 estimated from an
LUR model and total, cardiopulmonary, IHD, and respiratory disease mortality, with the
strongest effects observed for IHD mortality. When the analysis was restricted to
nonsmokers, a positive association was observed with lung cancer mortality. Similar
observations were reported for lung cancer by Katanoda et al. (2011) among a cohort in
Tokyo, Japan, and Liu et al. (2008) for a study of women living in Taiwan. In a related
study, Liu et al. (2009a) also observed a positive association between long-term exposure
to NO2 and bladder cancer mortality.
6-146
-------
The quantitative results of mortality studies are characterized in Figure 6-8. Figure 6-9.
and Figure 6-10; and Table 6-15. Table 6-16. and Table 6-17.
STUDY
Abbey etal. (1999)
Lipsettetal. (2011)
Krewski et al. (2000)
Caoetal. (2011)b
Filleul et al. (2005)
Zhang etal. (2011)
Cesaronietal. (2012)
Cesaronietal. (2013)
Tonne and Wilkinson (2013)b
Lipfert et al. (2000)
Maheswaran et al. (2010)
Pope et al. (2002)
Lipfert et al. (2006b)
Brunekreefetal. (2009)
Gehring et al. (2006)
Heinrichetal. (2013)
Hoeketal. (2002)
Jerrett et al. (2009)
Lipfert et al. (2006a)
Hart etal. (2011)
Hart etal. (2013)
Yorifuji et al. (2010)
Jerrett etal. (2013)
Carey etal. (2013)
Nafstad et al. (2004)
Tonne and Wilkinson (2013)
Krewski et al. (2000)
Lipfert et al. (2009)
Beelenetal. (2014a)
Beelenetal. (2014a)b
COHORT
AHSMOG
AHSMOG
CA Teachers
MEAN
CONCENTRATION NOTES
36.7 Men
36.7 Women
33.6 Women
ACS 27.9
China National Hypertension 50
PAARC
Shenyang, China
RoLS
RoLS
U.K.
Veterans
South London Stroke
ACS -Extended
Veterans
NLCS-AIR
German Women's Health
German Women's Health
Toronto, Canada
Veterans
TrIPS
Nurses Health Study
Shizuoka Elderly
ACS-CA
England, U.K.
Oslo, Norway, men
U.K.
HSC-Re-analysis
Veterans
ESCAPE
ESCAPE
24.5
24.4
23.3-24.2
23.4
22.6
21.5-27.8
21 5
21.4-27.9
19.8-27.2
20.7
20.7
20.7
A Q E
ly.u
19.5
16.3
14.2
13.9
13.3
12.3
11.9
10.6
10; 6. 5
6.1-21.9
5.58
2.8-31.7
8.7-107.3
24 areas 4
1 8 areas
StrokB Survivors
1
Full Cohort
Case Cohort — 0-
1-yravg
5-yr avg
^
Full Cohort
Excluding Long Haul Drivs
Female Nurses
!
Pooled Analysis
Pooled Analysis
— 0—
-0—
0
O
\
*
o-
-
— • —
•
0-
*
rs-0-
»
•
-O-
-O
0-
-0-
— 0—
o
D
Ik
0.5
Hazard Ratios (95% Cl)a
Note: ACS = American Cancer Society; AHSMOG = California Seventh-Day Adventists Cohort; CA = California; Cl = confidence
interval; ESCAPE = European Study of Cohorts for Air Pollution Effects; HSC = Harvard Six Cities; NLCS = Netherlands Cohort
Study on Diet and Cancer; NLCS-AIR = Netherlands Cohort Study on Air Pollution and Mortality; PAARC = Air Pollution and Chronic
Respiratory Diseases; RoLS = Rome Longitudinal Study; TrIPS = Trucking Industry Particle Study. Black = studies from the 2008
Integrated Science Assessment for Oxides of Nitrogen; Red = recent studies and those not included in 2008 Integrated Science
Assessment. Circles = NO2; triangles = NOX. Studies are presented in descending order of mean concentration.
aHazard ratios are standardized to a 10-ppb increase in NO2 and a 20-ppb increase in NOX concentration.
bEffect estimates from studies measuring NOX in |jg/m3 are not standardized.
Figure 6-8 Association between long-term exposure to oxides of nitrogen
and total mortality.
6-147
-------
Table 6-15 Corresponding risk estimates for Figure 6-8.
Study
Abbey et al. (1999)
tLipsett et al. (2011)
Krewski et al. (2000)
tCaoetal. (2011)
Filleuletal. (2005)
tZhanqetal. (2011)
tCesaroni et al. (2012)
Location
U.S.
California, U.S.
U.S.
China
France
China
Italy
Notes
Men
Women
Women, NO2
Women, NOx
NOx
24 areas
18 areas
Hazard Ratio (95% Clf
1.02(0.95, 1.08)
0.99(0.94, 1.05)
0.97(0.94, 1.05)
1.02(0.98, 1.06)
0.99(0.99, 1.00)
1.02(1.00, 1.03)b
0.98(0.96, 1.00)
1.22(1.10, 1.34)
5.39(4.94,5.94)
1.07(1.05, 1.11)
tCesaroni et al. (2013)
tTonne and Wilkinson (2013)
Lipfert et al. (2000)
tMaheswaran et al. (2010)
Pope etal. (2002)
Lipfert et al. (2006b)
tBrunekreef et al. (2009)
Gehrinq et al. (2006)
tHeinrich et al. (2013)
Hoek et al. (2002)
tJerrett et al. (2009)
Lipfert et al. (2006a)
tHartetal. (2011)
tHartetal. (2013)
tYorifuii et al. (2010)
tJerrett et al. (2013)
Italy
England and Wales, U.K. NOx
U.S.
England, U.K. Stroke survivors
U.S.
U.S.
the Netherlands Full cohort
Case cohort
Germany 1-yravg
5-yr avg
Germany
the Netherlands
Canada
U.S.
U.S. Full cohort
Excluding long haul drivers
U.S. Female nurses
Japan
U.S.
1.06(1.04, 1.06)
1.03(1.01, 1.05)b
1.07(1.04, 1.10)
1.59(1.22,2.09)
1.00(0.98, 1.02)
1.03(0.98, 1.02)
1.05(1.00, 1.10)
0.92(0.79, 1.06)
1.20(1.02, 1.41)
1.23(1.02, 1.47)
1.21 (1.08, 1.36)
1.15(0.89, 1.49)
1.48(1.00,2.16)
1.04(0.97, 1.13)
1.10(1.06, 1.15)
1.19(1.13, 1.26)
1.03(1.00, 1.06)
1.04(0.93, 1.32)
1.08(1.02, 1.14)
6-148
-------
Table 6-15 (Continued): Corresponding risk estimates for Figure 6-8.
Study
Location
Notes
Hazard Ratio (95% Clf
tCarevetal. (2013)
Nafstad et al. (2004)
tTonne and Wilkinson (2013)
Krewski et al. (2000)
tLipfert et al. (2009)
tBeelen etal. (2014a)
England, U.K.
Norway
England and Wales, U.K.
U.S. NOx
U.S.
Europe Pooled analysis, NO2
Pooled analysis, NOx
1.11 (1.05, 1.15)
1.16(1.12, 1.22)
1.12(1.06, 1.20)
1.15(1.04, 1.27)
1.04(1.03, 1.05)
1.02(0.98, 1.06)
1.02(1.00, 1.04)b
avg = average; Cl = confidence interval; NO2 = nitrogen dioxide; NOX = sum of NO and NO2
aEffect estimates are standardized to a 10-ppb increase in NO2 and a 20-ppb increase in NOX concentration, unless otherwise
specified.
bEffect estimates for NOX based on |jg/m3are not standardized. Effect estimate in Cao etal. (2011) and Tonne and Wilkinson
(2013) is per 10 |jg/m3 increase and in Beelen et al. (2014a) is per 20 |jg/m3 increase.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
6-149
-------
MEAN
STUDY
Naess etal. (2007)
Lipsettetal. (2011)
Cao etal. (2011 )b
Zhang etal. (2011)
Cesaronietal. (201 3)
Brunekreefet al. (2009)
Hart etal. (20 10)
Jerrettetal. (2013)
Chen etal. (201 3)
Carey etal. (201 3)
Beelen etal. (201 4b)
Abbey etal. (1999)
Krewski et al. (2000)
Filleul et al. (2005)
c^ehring et al. (^OOb)
Heinrich etal. (2013)
Wangetal. (2009b)
Yorifuji etal. (2010)
Krewski et al. (2000)
Rosenlund et al. (20081
COHORT CONCENTRATION NOTES |
Oslo, Norway
CA Teachers
42.1
33.6
China National Hypertension 26. 5
Shenyang, China
RoLS
NLCS-AIR
TrIPS
ACS-CA
Ontario, Canada
National English
ESCAPE
AHSMOG
ACS
PAARC Survey
c^erman women s Health
German Women's Health
Brisbane, Australia
NLCS
Shizuoka Elderly
HSC Re-analysis
Rome. Italy
24.4
23.4
20.7
14.2
12.3
12.1-21.7
11.9
2.8-31.7
8.7-107
36.7
27.9
24.5
20.7
19.8
195
13.3
6.1-21.9
-23.9
Women l»-
Women — 4 —
j^
|
.•
Full Cohort -U-
Case Cohort •-! —
Full Cohort |-«-
Excluding Long Haul Drivers . •
' — • —
1 -•-
f
|
Men -f-
Women -»-
b
24 areas -f-
18 areas — = — •
1 yr avg
yr aVQ ( •
*j—
| •
Out of hosoital * — •—
Cardiovascular
Cardio pulmonary
Gan etal. (2011]
Carey etal. (2013]
Lipsettetal. (2011]
Vancouver, Canada
National English
In hospital -|-<
Following non-fatal event—• i
Krewski et al. (2000)
Cesaronietal. (201 3)
Hart etal. (2011)
Yorifuji etal. (2010)
Jerrettetal. (2013)
Nafstad et al. (2004)b
Chen etal. (201 3)
Carey etal. (201 3)
Beelen etal. (201 4b)
Beelen etal. (201 4b)b
Beelen etal. (201 4b)
Beelen etal. (201 4b)b
Jerrettetal. (2009)
Yorifuji etal. (2010)
ACS
RoLS
TrIPS
Shizuoka Elderly
ACS-CA
Oslo, Norway, men
Ontario, Canada
National English
ESCAPE
Toronto, Canada
Shizuoka Elderly
27.9 f
23.4 ! »
14.2 Full Cohort — •—
Excluding Long Haul Drivers -§-• —
1° ° •
11.5-21.7 '-±
12.1-21.7 | -•-
11.9 Ml — • —
2.8-31.7 All IHD 4
8.7-107 All IHD -k-
2.8-31.7 Ml *\
8.7-107 Ml -i-
195 1 O
Circulatory
Cerbrovascular
Lipsettetal. (2011) CA Teachers
Zhang et al. (2011)
Cesaronietal. (2013)
Yorifuji etal. (2010)
Nafstad et al. (2004)b
Chen etal. (2013)
Beelen etal. (2014b)
Beelen etal. (2014b)b
Shenyang, China
RoLS
Shizuoka Elderly
Oslo, Norway, men
Ontario, Canada
ESCAPE
Jerrett et al. (2013) ACS-CA
Carey et al. (2013) National English
33.6 V
24.4
23.4
13.3
11.5-21.7
12.1-21.7
2.8-31.7
8.7-107
12.3
11.9
Carey etal. (2013) National English
Raaschou-Nielsen et al.
(2013b)
11.9
Diet, Cancer and Health 9
Heart Failure
Diabetes
Hazard Ratios (95% Cl)1
Note: ACS = American Cancer Society; AHSMOG = California Seventh-Day Adventists Cohort; CA = California; CHD = coronary
heart disease; Cl = confidence interval; ESCAPE = European Study of Cohorts for Air Pollution Effects; HSC = Harvard Six Cities;
IHD = ischemic heart disease; NLCS = Netherlands Cohort Study on Diet and Cancer; NLCS-AIR = Netherlands Cohort Study on
Air Pollution and Mortality; PAARC = Air Pollution and Chronic Respiratory Diseases; RoLS = Rome Longitudinal Study;
SALIA = Study on the Influence of Air Pollution on Lung, Inflammation, and Aging; TrIPS = Trucking Industry Particle Study.
Black = studies from the 2008 Integrated Science Assessment for Oxides of Nitrogen, red = recent studies and those not included in
the 2008 Integrated Science Assessment. Circles = NO2; triangles = NOX; squares = NO. Studies are presented in descending order
of mean concentration (in parts per billion [ppb]).
aHazard ratios are standardized to a 10-ppb increase in NO2 and NO, and a 20-ppb increase in NOX concentration.
bEffect estimates from studies measuring NOX in |jg/m3 are not standardized.
Figure 6-9 Associations between long-term exposure to oxides of nitrogen
and cardiovascular mortality.
6-150
-------
Table 6-16 Corresponding risk estimates for Figure 6-9.
Study
Location Notes
Hazard Ratio (95% Cl)a
Cardiovascular Disease
Naess et al. (2007)
tLipsettetal. (2011)
tCaoetal. (201 1)b
tSchikowski et al. (2007)
tZhanqetal. (2011)
tCesaroni et al. (2013)
tBrunekreef et al. (2009)
tHartetal. (2011)
tJerrett et al. (2013)
tChenetal. (201 3a)
tCarey et al. (2013)
tBeelenetal. (2014b)
Norway Women
California, U.S. Women, NO2
China NOx
Germany
China
Italy
the Netherlands Full cohort
Case cohort
U.S. Full cohort
Excluding long haul drivers
U.S.
Canada
England
Europe Pooled analysis, NO2
1.06(1.00, 1.12)
0.98(0.88, 1.09)
1.05(0.99, 1.12)
1.02(1.01, 1.04)
1.86(1.26,2.74)
5.43(4.82,6.16)
1.06(1.04, 1.08)
1.04(0.96, 1.13)
0.92(0.77, 1.10)
1.09(1.01, 1.17)
1.14(1.03, 1.25)
1.12(1.02, 1.22)
1.17(1.10, 1.23)
1.05(1.00, 1.13)
1.02(0.94, 1.12)
1.02(0.99, 1.06)c
Cardiopulmonary Disease
Abbey et al. (1999)
Krewski et al. (2000)
Filleuletal. (2005)
Gehrinq et al. (2006)
tHeinrich et al. (2013)
tWanq et al. (2009b)
Hoek et al. (2002)
tYorifuji et al. (2010)
Krewski et al. (2000)
U.S. Men
Women
U.S.
France 24 areas
18 areas
Germany 1-yravg
5-yr avg
Germany
Australia
the Netherlands
Japan
U.S.
1.01 (0.93, 1.09)
1.02(0.95, 1.09)
1.01 (1.00, 1.02)
1.00(0.96, 1.04)
1.16(0.93, 1.45)
1.70(1.28,2.26)
1.92(1.35,2.71)
1.67(1.36,2.05)
0.95(0.74, 1.22)
1.45(0.99,2.13)
1.32(1.12, 1.54)
1.17(1.03, 1.34)
6-151
-------
Table 6-16 (Continued): Corresponding risk estimates for Figure 6-9.
Study
Location
Notes
Hazard Ratio (95% Cl)a
CHD
tRosenlund et al. (2008b)
tGanetal. (2011)
tCarevetal. (2013)
Italy
Canada
England
Out of hospital
In hospital
Following nonfatal coronary event
NO2
NO
1.16(1.04, 1.26)
1.10(0.94, 1.30)
0.91 (0.80, 1.04)
1.09(1.02, 1.19)
1.09(1.03, 1.15)
0.98(0.90, 1.07)
IHD
tLipsett et al. (2011)
Krewski et al. (2000)
tCesaroni et al. (2013)
tHartetal. (2011)
tYorifuji et al. (2010)
tJerrett et al. (2013)
Nafstad et al. (2004)
tChenetal. (201 3a)
tCarevetal. (2013)
tBeelenetal. (2014b)
California, U.S.
U.S.
Italy
U.S.
Japan
U.S.
Norway
Canada
England
Europe
Women, NO2
Women, NOx
Full cohort
Excluding long haul drivers
NOx
Ml only
Pooled analysis, all IHD; NO2
Pooled analysis all IHD; NOx
Pooled analysis, Ml only; NO2
Pooled analysis, Ml only; NOx
1.07(0.92, 1.24)
1.09(1.00, 1.19)
1.02(1.00, 1.03)
1.10(1.06, 1.14)
1.01 (0.92, 1.11)
1.07(0.95, 1.21)
1.57(1.04,2.36)
1.17(1.04, 1.31)
1.08(1.03, 1.12)b
1.19(1.08, 1.30)
1.00(0.88, 1.13)
1.00(0.84, 1.18)
1.02(0.95, 1.09)b
0.96(0.79, 1.18)
0.99(0.90, 1.07)b
Circulatory Disease
tJerrett et al. (2009)
tYorifuji et al. (2010)
Canada
Japan
2.53(1.27,5.11)
1.30(1.06, 1.59)
Cerebrovascular Disease
tLipsett et al. (2011)
tZhanqetal. (2011)
tCesaroni et al. (2013)
tYorifuji et al. (2010)
Nafstad et al. (2004)
California, U.S.
China
Italy
Japan
Norway
Women, NO2
Women, NOx
NOx
0.86(0.71, 1.06)
1.01 (0.90, 1.14)
5.35(4.67,6.11)
1.02(0.98, 1.06)
1.18(0.89, 1.57)
1.04(0.94, 1.15)b
6-152
-------
Table 6-16 (Continued): Corresponding risk estimates for Figure 6-9.
Study
tChenetal. (201 3a)
tBeelenetal. (2014b)
Location Notes
Canada
Europe Pooled analysis, NO2
Pooled analysis, NOx
Hazard Ratio (95% Cl)a
0.92(0.81, 1.10)
1.02(0.87, 1.20)
1.00(0.93, 1.08)b
Stroke
tJerrett et al. (2013)
tCareyetal. (2013)
U.S.
England
1.20(1.04, 1.39)
1.00(0.91, 1.09)
Heart Failure
tCarevetal. (2013)
England
1.09(0.91, 1.32)
Diabetes
tRaaschou-Nielsen et al. (2013b)
Denmark
1.66(0.96,2.89)
avg = average; CHD = coronary heart disease; Cl = confidence interval; IHD = ischemic heart disease; Ml = myocardial infarction;
NO = nitric oxide; NO2 = nitrogen dioxide; NOX = sum of NO and NO2
aEffect estimates are standardized to a 10-ppb increase in NO2 and NO, or a 20-ppb increase in NOX concentration.
bEffect estimates for NOX based on |jg/m3 are not standardized. Effect estimate in Cao et al. (2011) and Nafstad et al. (2004) is per
10 |jg/m3 increase. Effect estimate in Beelen et al. (2014b) is per 20 |jg/m3 increase.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
6-153
-------
STUDY
Abbey etal. (1999)
Lipsettetal. (2011)
Caoetal. (201 1)b
Dong etal. (2012)
Cesaroni et al. (2013)
Brunekreef et al. (2009)
Heinrich etal. (2013)
Jerrett et al. (2009)
Hart etal. (2011)
Yorifuji et al. (2010)
Jerrett etal. (2013)
Carey etal. (2013)
Nafstadetal. (2004) b
Dimakopoulou etal. (201 4)b
Katanoda et al. (2011)
Naess et al. (2007)
Ganetal. (2013)
Hart etal. (2011)
Yorifuji et al. (2010)
Carey etal. (2013)
Abbey etal. (1999)
Lipsettetal. (2011)
Naess et al. (2007)
Krewski et al. (2000)
Caoetal. (201 1)b
Filleul et al. (2005)
Cesaroni et al. (2013)
Brunekreef et al. (2009)
Heinrich etal. (2013)
Hart etal. (2011)
Nyberg et al. (2000)
Yorifuji et al. (2010)
Jerrett etal. (2013)
Carey etal. (2013)
Nafstadetal. (2004) b
Krewski et al. (2000)
Katanoda et al. (2011)
COHORT MEAN NOTES • Respiratory
CONCENTRATION '
AHSMOG 36.7 Men »-|-
Women — • —
CA Teachers 33.6 Women «-J
Women — A 1
China National Hypertension 26.5 It
Shenyang, China 24.5 | •
RoLS ' 23.4 .-•-
NLCS-AIR 20.7 Full Cohort •
German Women's Health
Toronto, Canada
TrIPS
Shizuoka Elderly
ACS-CA
National English
Oslo, Norway, men
ESCAPE
Three-prefecture
Oslo, Norway
Vancouver, Canada
TrIPS
Shizuoka Elderly
National English
AHSMOG
CA Teachers
Oslo, Norway
ACS
China National Hypertensic
PA ARC Survey
RoLS
NLCS-AIR
German Women's Health
TrIPS
Stockhom, Sweden
Shizuoka Elderly
ACS-CA
National English
Oslo, Norway, men
HSC Re-analysis
Three-prefecture
14.2 Full Cohort — (-•
12.3 •
11.9 -o-
10.6 1 — A-
2.8-28.1 »-l —
1.2-33.7 All subjects | »-
Men 1 -•-
Women . -•-
COPD
27.5-44.9 Men ' — • —
Women — ^» —
26 t"-
14.2 Full Cohort «|
11.9 J — • —
1 Lung Cancer
00 f^ \At 1.
Women A
27.5-44.9 Men -j-0—
Women — • —
27.9 J
m 26.5 It
24.5 24 areas -*-t
23.4 . •*-
20.7 Full Cohort »-j—
14.2 Full Cohort -!-• —
Excluding Long Haul Drivers-|— •
12.3 ' •
11.9 1 — 0—
10.6 I-A-
1.2-33.7 All subjects | -•-
Men * -•-
Women — • —
Hazard Ratio (95% Cl)1
Note: ACS = American Cancer Society; AHSMOG = California Seventh-Day Adventists Cohort; CA = California; Cl = confidence
interval; COPD = chronic obstructive pulmonary disease; ESCAPE = European Study of Cohorts for Air Pollution Effects; HSC =
Harvard Six Cities; NLCS-AIR = Netherlands Cohort Study of Air Pollution and Mortality; PAARC = Air Pollution and Chronic
Respiratory Diseases; RoLS = Rome Longitudinal Study; TrIPS = Traffic Industry Particle Study; yr = year. Black = studies from the
2008 Integrated Science Assessment for Oxides of Nitrogen. Red = studies not included in 2008 Integrated Science Assessment;
Circles = NO2; triangles = NOX; squares = NO. Studies are presented in descending order of mean concentration (in parts per billion
[ppb]).
aHazard ratios are standardized to a 10-ppb increase in NO2 or NO and a 20-ppb increase in NOX concentration.
bEffect estimates from studies measuring NOX in |jg/m3 are not standardized.
Figure 6-10 Associations between long-term exposure to oxides of nitrogen
and respiratory mortality.
6-154
-------
Table 6-17 Corresponding risk estimates for Figure 6-10.
Study
Location
Notes
Hazard Ratio (95% Cl)a
Respiratory
Abbey et al. (1999)
tLipsett et al. (2011)
tCaoetal. (2011)
tDonqetal. (2012)
tCesaroni et al. (2013)
tBrunekreef et al. (2009)
tHeinrich et al. (2013)
tJerrett et al. (2009)
tHartetal. (2011)
tYorifuii et al. (2010)
tJerrett et al. (2013)
tCarevetal. (2013)
Nafstad et al. (2004)
tDimakopoulou etal.
(2014)
tKatanoda et al. (2011)
U.S.
California, U.S.
China
China
Italy
the Netherlands
Germany
Canada
U.S.
Japan
U.S.
England
Norway
Europe
Japan
Men
Women
Women, NO2
Women, NOx
NOx
Full cohort
Case cohort
Full cohort
Excluding long haul drivers
NOx
Pooled analysis, NO2
Pooled analysis, NOx
All subjects
Men
Women
0.93(0.82, 1
0.98(0.87, 1
0.93(0.76, 1.
0.94(0.83, 1.
1.03(1.00, 1.
6.1 (5.2, 7.2)
1.06(1.00, 1
1.22(1.00, 1
1.16(0.83, 1
1.15(0.67,2,
1.16(0.37,2,
1.07(0.91, 1
1.26(1.01, 1
1.39(1.04, 1
1.00(0.83, 1
1.28(1.18, 1
1.32(1.12, 1.
0.94(0.80, 1.
0.99(0.90, 1.
1.16(1.12, 1
1.11 (1.05, 1
1.25(1.18, 1
.07)
.11)
,15)
,07)
,06)b
.12)
.48)
.62)
.00)
.71)
.27)
.56)
.83)
.20)
.38)
,54)b
,10)
,09)b
.21)
.18)
.33)
COPD
Naess et al. (2007)
tGanetal. (2013)
tHartetal. (2011)
Norway
Canada
U.S.
Men
Women
NO2
NO
Full cohort
Excluding long haul drivers
1.18(1.04, 1
1.05(0.93, 1
1.09(0.91, 1.
1.06(0.97, 1.
0.97(0.79, 1
1.01 (0.82, 1
.33)
.18)
,32)
,15)
.19)
.44)
6-155
-------
Table 6-17 (Continued): Corresponding risk estimates for Figure 6-10.
Study
tYorifuji et al. (2010)
tCareyetal. (2013)
Location
Japan
England
Notes
Hazard Ratio (95% Cl)a
1.22(0.63,
1.13(0.98,
2.31)
1.28)
Lung Cancer
Abbey et al. (1999)
tLipsett et al. (2011)
Naess et al. (2007)
Krewski et al. (2000)
tCaoetal. (2011)
Filleuletal. (2005)
tCesaroni et al. (2013)
tBrunekreef et al. (2009)
tHeinrich et al. (2013)
tHartetal. (2011)
Nvberq et al. (2000)
tYorifuii et al. (2010)
tJerrett et al. (2013)
tCarevetal. (2013)
Nafstad et al. (2004)
Krewski et al. (2000)
tKatanoda et al. (2011)
U.S.
California, U.S.
Norway
U.S.
China
France
Italy
the Netherlands
Germany
U.S.
Sweden
Japan
U.S.
England
Norway
U.S.
Japan
Men
Women
Women, NO2
Women, NOx
Men
Women
NOx
24 areas
18 areas
Full cohort
Case cohort
Full cohort
Excluding long haul drivers
30-yr exposure
10-yr exposure
NOx
All subjects
Men
Women
1.35(0.96,
1.69(1.07,
1.00(0.76,
0.98(0.90,
1.06(0.97,
1.20(1.09,
0.99 (0.97,
1.03(0.99,
0.94 (0.89,
1.12(0.77,
1.08(1.04,
0.94(0.81,
0.87(0.66,
1.56(0.91,
1.07(0.96,
1.09(0.95,
1.10(0.87,
1.20(0.94,
0.91 (0.63,
1.29(1.05,
1.20(1.09,
1.11 (1.03,
1.09(0.76,
1.17(1.10,
1.18(1.11,
1.13(1.01,
1.90)
2.65)
1.32)
1.07)
1.15)
1.32)
1.01)
1.07)b
1.02)
1.61)
1.14)
1.09)
1.14)
2.69)
1.19)
1.25)
1.36)
1.48)
1.34)
1.59)
1.32)
1.19)b
1.57)
1.26)
1.26)
1.27)
Cl = confidence interval; COPD = chronic obstructive pulmonary disease; NO = nitric oxide; NO2 = nitrogen dioxide; NOX = sum of
NO and NO2.
aEffect estimates are standardized to a 10-ppb increase in NO2 and NO or a 20-ppb increase in NOX concentration, unless
otherwise specified.
bEffect estimates for NOX based on |jg/m3 are not standardized. Effect estimate in Cao et al. (2011) and Nafstad et al. (2004) is per
10 |jg/m3 increase. Effect estimate in Dimakopoulou et al. (2014) is per 20 |jg/m3 increase.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
6-156
-------
6.5.3 Summary and Causal Determination
Collectively, the evidence is suggestive of, but not sufficient to infer, a causal
relationship between long-term exposure to NO2 and mortality among adults.1 The
strongest evidence comes from cohort studies conducted in the U.S., Canada, and Europe,
which generally show positive associations with total mortality, as well as deaths due to
respiratory and cardiovascular disease (Chen et al.. 2013a; Gan et al.. 2013; Hart et al..
2013: Heinrichetal.. 2013: Jerrettet al.. 2013: Gan etal.. 2011: Hart etal.. 2011: Lipsett
etal.. 2011: Brunekreef etal.. 2009: Jerrett et al.. 2009: Beelen et al.. 2008b: Schikowski
et al.. 2007: Krewski et al.. 2000). The results from these studies are coherent with
studies that have observed associations between long-term exposure to NO2 and
respiratory hospital admissions (Andersen et al.. 2012a: Andersen et al.. 2011) and
cardiovascular effects (Hart etal.. 2011: Lipsett et al.. 2011). The evidence for short- and
long-term respiratory and cardiovascular morbidity provides limited biological
plausibility for mortality. In the 2008 ISA for Oxides of Nitrogen, a limited number of
epidemiologic studies assessed the relationship between long-term exposure to NCh and
mortality in adults. The 2008 ISA concluded that the scarce amount of evidence was
"inadequate to infer the presence or absence of a causal relationship" (U.S. EPA. 2008c).
The stronger conclusion in this ISA is based on evidence in recent studies for an
association between long-term exposure to NO2 and mortality from extended analyses of
existing cohorts as well as original results from new cohorts in the U.S., Europe, and
Asia. The key evidence as it relates to the causal determination is summarized in Table
6-18 using the framework described in Table II of the Preamble.
Many of the studies evaluating the associations between long-term exposure to NO2 and
mortality used concentrations measured at central site monitors to assign exposure. Many
recent studies employed exposure assessment methods such as LUR or dispersion
modeling to account for the spatial variability of NO2 and estimate exposure at subjects'
homes. There was no distinguishable pattern or trend in the results of this body of
evidence that could be attributed to the use of central site monitors, LUR, or dispersion
models to assign exposure. However, demonstrations in many studies that LUR or
dispersion modeled estimates of NCh exposures well predicted NO2 concentrations in the
study areas lend confidence in the results for associations with mortality. While the
results were generally consistent across studies, there were several well-designed,
well-conducted studies that did not observe an association between long-term exposure to
NC>2 and mortality (Beelen et al., 2014a: Beelen et al., 2014b: Dimakopoulou et al., 2014:
For early life mortality, see Section 6.4.3.
6-157
-------
Krewski et aL 2009; Pope et al., 2002; Abbey etaL 1999). Many of these studies not
observing associations also estimated NO2 exposure with we 11-validated models.
Recent studies examined the potential for copollutant confounding by PM2 5, the traffic-
related pollutant BC, or measures of traffic proximity or density in copollutant models
(Beelen et al.. 2014a; Jerrett et al.. 2013; Hart et al.. 2011). These recent studies address a
previously identified data gap. The NO2 results from these models were generally
attenuated with the adjustment for PM2 5 or BC, though analysis of confounding is limited
and not all key traffic-related copollutants (i.e., CO) were evaluated. It remains difficult
to disentangle the independent effect of NO2 from the potential effect of the
traffic-related pollution mixture or other components of that mixture. Further, the
evidence does not demonstrate that long-term NO2 exposure has an independent effect on
the cardiovascular and respiratory morbidity outcomes that are major underlying causes
of mortality. In conclusion the generally supporting epidemiologic evidence but
uncertainty regarding an independent NCh effect is suggestive of, but not sufficient to
infer, a causal relationship between long-term exposure to NO2 and total mortality.
6-158
-------
Table 6-18 Summary of evidence, which is suggestive of, but not sufficient to
infer, a causal relationship between long-term nitrogen dioxide
exposure and total mortality.
Rationale for
Causal
Determination3
Key Evidence13
Key References'3
NO2 Concentrations
Associated with
Effects0
High-quality
epidemiologic
studies are
generally
supportive but not
entirely consistent
Positive association with cardiovascular
mortality in the HSC cohort and total
mortality a subset of the ACS cohort, with
effect estimates similar in magnitude to
those observed with PlVhs, even after
adjustment for common potential
confounders.
Krewski et al. (2000)
Means across cities
(1980): 6.1-21.9 ppb
tJerrett et al. (2013)
Sections 6.5.1 and 6.5.2
Mean (1988-2002):
12.3 ppb
Updated results from the NLCS report a
positive association with total mortality,
effects for respiratory mortality stronger
than any observed with traffic variables
and total or other cause-specific mortality.
tBeelenetal. (2008b),
tBrunekreef et al.
(2009)
Mean: 2.8-31.7 ppb for
annual avg.
Median: 20.2 ppb for 10-
yravg
Updated results from German women's
cohort report positive associations with
total and cardiopulmonary mortality.
tHeinrich et al. (2013),
tSchikowski et al.
(2007)
Mean: 20.7 ppb for 15-yr
avg.
Median: 24.4 ppb for
5-yr avg
Recent cohort studies in the U.S. observe
increases in total mortality and mortality
due to cardiovascular disease in separate
cohorts of men and women.
tHartetal. (2011)
Mean (1985-2000):
14.2 ppb
tLipsett et al. (2011)
Mean (1996-2005): 33.6
ppb; max: 67.2 ppb
tHartetal. (2013)
Median (2000): 13.9 ppb
Positive associations with total,
cardiovascular, respiratory, and lung
cancer mortality in Canadian cities. NO2
exposure estimated with well-validated
LUR models.
tChenetal. (2013a),
tJerrett et al. (2009),
tGanetal. (2011),
tGan etal. (2013)
Mean across cities:
12.1-21.7 ppb for
annual avg. Median:
22.9 ppb for annual avg
Mean: 17.0, 17.1 ppb for
5-yr avg
Some studies show No association in several re-analyses of
no association the ACS cohort.
Krewski et al. (2000).
Pope etal. (2002).
tKrewski et al. (2009)
Sections 6.5.1 and 6.5.2
Mean (1982-1998):
21.4-27.9 ppb
Mean (1982-1998): 27.9
ppb; max 51.1 ppb
No association in a multi-country
European study of 22 existing cohorts for
total, cardiovascular, or respiratory
mortality. NO2 exposure estimated with
well-validated LUR models.
tBeelen etal. (2014a),
tPimakopoulou etal.
(2014), tBeelen et al.
(2014b)
Means across cohorts:
2.8-31.7 ppb for annual
avg
No association with total,
cardiopulmonary, or respiratory mortality
intheAHSMOG.
Abbey etal. (1999)
Mean (1973-1992):
36.8 ppb
6-159
-------
Table 6-18 (Continued): Summary of evidence, which is suggestive of, but not
sufficient to infer, a causal relationship between long-
term nitrogen dioxide exposure and total mortality.
Rationale for NO2 Concentrations
Causal Associated with
Determination3 Key Evidence13 Key References'3 Effects0
Uncertainty Associations with mortality generally tJerrett et al. (2013),
remains regarding attenuated with adjustment for PIVhs or tGan et al. (2011),
independent effects BC. Analysis is limited and does not tGan et al. (2013)
of NO2 include many traffic-related copollutants Section 652
(i.e., CO). When reported, correlations
with copollutants were highly variable (low
to high).
Limited coherence Limited evidence for COPD hospital tAndersen et al. (2011), 35-yr mean: 9.0 ppb
with evidence for admissions in adults but uncertainty in t Andersen et al. 25-yr mean' 9 5 ppb
respiratory and independent NO2 effect. (2012a)
cardiovascular Section 6.2.8
morbidity
Some inconsistencies reported for tLipsett et al. (2011). Mean: 33.6 ppb for
cardiovascular morbidity. Coherent t Atkinson etal. (2013) multi-yravg, 12.0 ppb
evidence for Ml and heart failure, but Section 632 for annual av9
uncertainty in independent NO2 effect.
ACS = American Cancer Society; AHSMOG = Adventist Health Study of Smog; BC = black carbon; CO = carbon monoxide;
COPD = chronic obstructive pulmonary disease; HSC = Harvard Six Cities; LUR = land use regression; Ml = myocardial infarction;
NLCS = Netherlands Cohort Study on Diet and Cancer; NO2 = nitrogen dioxide; ppb = parts per billion; PM2.s = particulate matter
with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm.
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Tables I and N. of the
Preamble.
Describes the key evidence and references, supporting or contradicting, contributing most heavily to causal determination and,
where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where full body of evidence is
characterized.
°Describes the NO2 concentrations with which the evidence is substantiated (for experimental studies, £ 5,000 ppb).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
6.6 Cancer
6.6.1 Introduction
The 1993 AQCD for Oxides of Nitrogen and the 2008 ISA for Oxides of Nitrogen
reported that there was no clear evidence that NO2 or other oxides of nitrogen act as a
direct carcinogen. The U.S. Department of Health and Human Services and the
International Agency for Research on Cancer have not classified oxides of nitrogen for
potential carcinogenicity. The American Conference of Industrial Hygienists has
classified NO2 as A4 (not classifiable for humans or animals). The 2008 ISA for Oxides
of Nitrogen (U.S. EPA. 2008C) included a few epidemiologic studies of oxides of
nitrogen and cancer, both examining lung cancer incidence and reporting positive
associations. Several additional epidemiologic studies examined lung cancer and a few
6-160
-------
studies each examined leukemia, bladder cancer, breast cancer, and prostate cancer.
These are all described in more detail in Supplemental Table S6-10 (U.S. EPA. 2013o).
which includes information on the exposure assessment and duration, as well as effect
estimates. For the evaluation of long-term exposure and cancer, important considerations
were the exposure assessment method and potential confounding by PlVfc 5, diesel
exhaust, polycyclic aromatic hydrocarbons, and other traffic-related copollutants.
Ambient NCh concentrations typically show spatial variability (Section 2.5.3). and PM2 5
and traffic-related copollutants also show associations with cancer. Exposure assessment
was evaluated drawing upon discussions in Sections 3.2 and 3.4.5. Several recent studies
of cancer employed exposure assessment methods to account for the spatial variability of
NO2. For example, LUR model predictions have been found to correlate well with
outdoor NC>2 concentration measurements (Section 3.2.2.1). For long-term NC>2 exposure,
exposure assessment was evaluated by the extent to which the method represented the
spatial variability in NO2 concentrations in a given study.
6.6.2 Lung Cancer
6.6.2.1 Epidemiologic Studies
Lung Cancer Incidence
The two previous studies in the 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c)
reported positive associations between NO2 or NOx and lung cancer incidence (Nafstad et
al., 2003; Nvberg etal.. 2000). Nyberg et al. (2000) reported an association between NO2
and lung cancer at the highest 10-year avg concentrations of NC>2 with a 20-year lag. NC>2
exposure was estimated from NOx emissions and a geostatistical model. The association
was relatively unchanged after adjustment for SO2, which was not observed to be
associated with lung cancer (Pearson r with NO2 exposure = 0.64). Nafstad et al. (2003)
performed a study with 24 years of follow-up and reported a positive association between
NOx concentrations and lung cancer incidence during the early years of the study, but the
authors report more recent years had weaker associations (results were not provided).
Neither of the previous studies examined PM25, diesel exhaust, polycyclic aromatic
hydrocarbons, or other traffic-related pollutants or reported information on whether the
exposure assessment methods captured the spatial variability in the study areas.
Evidence from recent studies is inconsistent regarding an association between long-term
NO2 exposure and lung cancer incidence, including results for NO2 exposures estimated
using well-validated LUR models. A meta-analytical study combined individual
6-161
-------
estimates from 17 cohort studies across nine countries in Europe (Raaschou-Nielsen et
al.. 2013a). Positive associations for NO2 and lung cancer became null after adjusting for
smoking-related covariates, fruit intake, and area-level SES. No associations were
observed for PIVb 5, PNfe.s absorbance, and PMio-2.5, but PMio was positively associated
with lung cancer. LUR models for all of the examined pollutants were demonstrated to
predict well ambient concentrations in the study areas (cross-validation R2 = 0.60-0.80).
A smaller study in Toronto, Canada reported a positive association between NO2
concentrations and lung cancer only when using a population-based control group
(Villeneuve etal.. 2014). The association was present when using NCh concentration at
the time of the interview, from 10 years prior, and from a time-weighted average. Cross-
validation of the LUR model indicated a small difference (4%) between model
predictions and measured concentrations. There was no association in analyses adjusted
for personal (e.g., age, smoking, BMI) and neighborhood SES (e.g., neighborhood
unemployment rate) covariates using both hospital-based and population-based controls.
The traffic-related pollutant benzene also was associated with lung cancer incidence, and
confounding by benzene was not evaluated (r = 0.67 for correlation with NCh).
NC>2 exposure was not consistently associated with lung cancer incidence in studies with
more uncertain exposure estimates. In a Netherlands cohort, 5-yr and 10-yr avg NC>2
exposure was estimated for subjects' homes using LUR models, but information was not
reported on model validation. Over 11 years of follow-up, no associations were observed
in analyses using case-cohort and full cohort approaches. (Brunekreef et al., 2009; Beelen
et al.. 2008a). NO2 was highly correlated (r > 0.75) with PM2 5, PM2.s absorbance, and
black smoke (BS), which also were not associated with lung cancer incidence. In Canada,
20-yr avg NO2 exposure estimated from a spatiotemporal model was associated with lung
cancer incidence (Hystad et al., 2013). However, national model was used to estimate
exposure at the level of the postal code, and the model showed considerable variation
with measured concentrations (cross-validation R2 = 0.38). An association was present
for adenocarcinomas but not squamous cell carcinoma. Confidence intervals for estimates
of small cell and large cell carcinomas were wide. PIVbs demonstrated some associations
with lung cancer incidence, especially in the third and fourth, but not fifth, quintiles of
exposures. Copollutant models were not examined for NO2 and PM2 5 because of the high
correlation between the pollutants. Odds of lung cancer increased in association with NO2
concentrations when the analysis was limited to the closest monitor within 50 km, but it
is unclear whether the 50 km distance adequately represents the spatial variability in NO2
exposure. In stratified analyses, associations appeared to be greater among men; results
were null for women. No differences were apparent by education or smoking status.
NO2 exposure estimated from central site monitors was not associated with lung cancer
incidence. A case-control study of molecular changes and genetic susceptibility used
6-162
-------
multiple statistical techniques to evaluate the associations between air pollutants and lung
cancer incidence (Papathomas et al.. 2011). A high risk group was identified as having
combined high PMio and NC>2 exposures and residence in main roads; however, NC>2
exposure examined individually was not associated with lung cancer incidence. The same
was true of PMio. In another statistical model, PMio but not NCh was chosen as a
predictor. An ecologic study in Taiwan observed no associations for NC>2 concentrations
with all lung cancer cases combined, adenocarcinomas, or squamous cell carcinomas
(Tseng et al.. 2012). No association was observed for CO, but other traffic-related
pollutants were not examined. A positive association was observed in the highest quartile
of NO for adenocarcinomas but not for squamous cell carcinomas.
Similar to a previous study, recent analyses reported an association between increased
NOx concentrations and lung cancer incidence. In Danish cohorts, the increased
incidence with NOx exposure persisted in some models of specific cancer types, such as
squamous cell carcinomas (Raaschou-Nielsen et al., 2010a). For associations stratified by
sex, length of education, and smoking status, the precision decreased (i.e., wider 95%
CI), and no differences were observed between the groups. One of the Danish cohorts
was used in another study where the follow-up period was extended 5 years to include
more cases (Raaschou-Nielsen et al.. 201 Ib). Increased incidence rate of lung cancer was
observed in the highest quartile of NOx concentrations, and risk was greater in the group
with 8 years or more of schooling, the group with lower fruit intake, nonsmokers, and
females. In these studies, residential NOx exposure was estimated from a model with
emissions and geographical data as inputs. The modeled NOx estimates correlated well
with measured concentrations on a busy street (r = 0.88), but they also were nearly
perfectly correlated with traffic-related copolluants (e.g., r = 0.93 for particle number
concentration). Therefore, an independent NOx effect is not discernable.
Lung Cancer Mortality
Compared to lung cancer incidence, long-term NO2 exposure was more consistently
associated with lung cancer mortality. Positive associations were observed in many but
not all studies with NO2 exposure estimated at or near subjects' homes with well-
validated models or nearby central site monitors. As with lung cancer incidence, an
independent association with lung cancer mortality is uncertain. In most studies, PM2 5
was highly correlated with NO2 (r = 0.79-0.95), was estimated with models with similar
predictive capacity, and also associated with lung cancer mortality. Further, important
copollutants such as diesel exhaust particles or VOCs were not examined. In the ACS
Cancer Prevention Study II cohort, lung cancer mortality was positively associated with
NO2 exposure estimated by an LUR model that predicted well ambient NO2
concentrations in the study area (cross-validation R2 = 0.71) (Jerrett et al., 2013). NO2
6-163
-------
was moderately correlated with PlVfc 5 and Os (Pearson r = 0.55). The positive association
with NO2 was robust to adjustment with PM2 5 or Os. In a U.S. study of men employed by
the trucking industry in 1985, no association was observed with lung cancer mortality
ascertained from the National Death Index through 2000 (Hart etal.. 2011). NO2
exposure was estimated for subjects' homes using a well-validated spatial smoothing
model (cross-validation R2 = 0.88). There was no association after long-haul drivers who
are away from the home at least one night per week were excluded from the analyses.
Similar results were observed for PMio and SCh.
Long-term NCh exposure also was associated with lung cancer mortality in studies
conducted across Europe. These studies also estimated NO2 exposure with well-validated
LUR, emissions-based, or dispersion models (cross-validation R2 = 0.57-0.80 across
locations or r = 0.30-0.87 for correlation with measured NC>2 across seasons). In
England, lung cancer mortality was ascertained from a large nationally representative
database (Carey et al.. 2013). In Rome, Italy, the association demonstrated a linear
relationship (Cesaroni et al.. 2013). A study in Oslo, Norway linked lung cancer mortality
ascertained for a 4-year period to NO2 averaged over this follow-up period (Naess et al..
2007). The association was positive only among women. In a nonparametric smooth
analysis that combined the sexes, the increase in log odds for lung cancer appears to
begin around 21.3 ppb for the 51-70 year age group while the increase appears to occur
at lower concentrations among those aged 71-90 years. A study in Japan estimated NO2
exposure with an LUR model that did not perform as well as the aforementioned studies
(cross-validation R2 = 0.54). Among individuals ages 65-84 years at enrollment and
followed for about 6 years, NCh was not associated with lung cancer mortality (Yorifuji
et al.. 2010). With additional years of follow-up, associations for lung cancer mortality
were observed for NO2 exposure estimated for the year of mortality, averaged 1 to 3
years before death, and averaged over follow-up (Yorifuji et al.. 2013). In this Japanese
cohort, no effect measure modification was indicated by smoking status, age, sex, BMI,
diabetes, hypertension, or financial ability.
A few studies estimated NO2 exposure from central site monitors located in close
proximity to subjects. A multi-region study in Japan followed individuals for 10 years
and reported region-specific results (Katanoda et al.. 2011). NO2 was inconsistently
associated with lung cancer mortality between two regions where distance to the monitor
was less than 1 km. The region where the association was observed had the highest NO2
concentration (data on association by region only presented in figures). A study in France
monitored NCh in study areas that were 0.5 to 2.3 km wide (Filleul etal.. 2005). NC>2 was
associated with lung cancer mortality only after exclusion of areas with a strong influence
of heavy traffic (i.e., monitoring sites reporting a high ratio of NO to NO2). The small
area around most monitors and the exclusion of high-traffic areas suggest that the
6-164
-------
monitor may fairly well represent NC>2 concentrations across the study area. BS, which
was highly correlated with NC>2 (r = 0.72), was not associated with lung cancer mortality.
Associations with lung cancer mortality generally were not observed in studies with more
uncertainty in NC>2 exposure estimates. No association was reported with NOx in a
national study of urban areas in China (Cao et al.. 2011). NCh was not associated with
lung cancer mortality in the NCLS with follow-up of at least 10 years (Brunekreef et al.,
2009; Beelen et al.. 2008b). NCh exposure was estimated for subjects' homes by LUR,
but no information was reported on model validation. No association was observed with
PM2 5 or BS. Lung cancer mortality was not associated with NCh exposure estimated
from central site monitors in the ACS cohort (Krewski et al., 2009) or in a cohort of
women in Germany (Heinrich et al.. 2013). Neither study indicated the extent to which
the concentrations averaged over the metropolitan statistical area (ACS) or assigned from
the nearest monitor which in most cases was within 5 km (German cohort) represented
the spatial variability in NC>2 concentrations in the study area. Such information is lacking
also in a study in Taiwan, which used a case-control design (Liu etal.. 2008). The highest
tertile of NO2 concentration was positively associated with lung cancer mortality. Lung
cancer mortality was not associated with SO2, PMio, or Os but was associated with CO as
well as an exposure index combining the highest tertile of CO and NO2.
6.6.2.2 Animal lexicological Studies
Lung Tumors with Co-exposure with Known Carcinogens or Copollutants
The 1993 AQCD for Oxides of Nitrogen and the 2008 ISA for Oxides of Nitrogen
reported a lack of evidence that NO2 acts as a direct carcinogen but noted a possible role
for NO2 to act as a tumor promoter in the lung with co-exposure with known carcinogens
or diesel exhaust particles. Toxicological studies of NO2 and carcinogenicity and
genotoxicity are described in Table 6-19. There also is some evidence of hyperplasia of
respiratory epithelium with NO2 exposure, which is consistent with a role for NO2 in
tumor promotion (Section 6.2.6). Continuous exposure to 4,000 ppb NO2 for 17 months
after injection with the carcinogen N-bis(2-hydroxy-propyl) nitrosamine (BHPN) led to a
statistically nonsignificant fivefold increase in the number of rats that developed
adenomas or adenocarcinomas of the lungs compared to clean air exposure (Ichinose et
al., 1991). The number of rats with tumors following BHPN exposure and BHPN plus
NO2 co-exposure did not differ statistically. Another study by the same lab (Ichinose and
Sagai. 1992) showed statistically significant increases in rats with lung tumors with
combined 13-month exposure to 400 ppb NO2 and 50 ppb Os after BHPN injection (5
rats). BHPN injection alone did not induce tumors, and the increase with Os exposure
6-165
-------
alone (3 rats) or NCh plus H2SO4 exposure (3 rats) was not statistically significant. Rats
were not exposed to NO2 alone.
Higher than ambient-relevant NC>2 exposure (6,000 ppb, 8 months) combined with diesel
exhaust particle extract-coated carbon black particles (DEPcCBP) increased the number
of F344 rats with lung tumors (alveolar adenomas) for the animals (Ohyama et al.. 1999).
DEPcCBP exposure occurred by intratrachael (IT) installation once per week for 4 weeks
and did not induce tumors when given alone.
Lung Tumors in Animals with Spontaneously High Tumor Rates
The 2008 ISA for Oxides of Nitrogen described increased tumors in rodents with
spontaneously high tumor rates with higher than ambient-relevant NC>2 exposure but not
clearly with relevant exposure concentrations. In AKR/cum mice, 250 ppb NC>2 exposure
for up to 26 weeks decreased progression of spontaneous T cell lymphoma and increased
survival rates (Richters andDamji. 1990). A similar duration of exposure (6 months,
6 h/day, 5 days/week) to 1,000 or 5,000 ppb NC>2 had no effect on pulmonary adenomas
in A/J mice, but 10,000 ppb NC>2 induced a small, but statistically significant increase in
pulmonary adenomas [increased tumor multiplicity; (Adkins etal.. 1986)1. In CAFl/Jax
mice, continuous exposure to 5,000 ppb NC>2 produced statistically significant increases
in the number of mice with pulmonary tumors when compared with controls after 12
months but not after 14 or 16 months (Wagner et al.. 1965).
Facilitation of Lung Cancer Metastases
The 2008 ISA for Oxides of Nitrogen summarized a group of experiments by one lab that
focused on the role ofNO2 in facilitation of metastases, or more accurately, colonization
of the lung by tumor cells. Richters and Kuraitis (1981). Richters et al. (1985). and
Richters and Kuraitis (1983) exposed mice to 300-800 ppb NO2 for 10 or 12 weeks, then
injected mice intravenously (i.v.) with the B16 melanoma cell line. Lung tumors were
then counted, with results of some of the experiments showing significantly increased
numbers of tumors.
Genotoxicity in Airway Cells
Ex vivo exposure of human nasal epithelial mucosa cells cultured at the air-liquid
interface starting at 10 ppb NO2 (Koehler et al.. 2013; Koehler et al.. 2010) or 100 ppb
NO2 (Koehler et al.. 2011) produced increased deoxyribonucleic acid (DNA)
fragmentation measured with the single cell gel electrophoresis (comet) assay as early as
6-166
-------
30 minutes after exposure. Percentage of DNA content in the tail as detected with the
comet assay decreased with increasing exposure duration [0.5, 1, 2, and 3-hour exposure;
(Koehler et al., 2013)1. NO2 exposure induced micronuclei formation in some but not all
studies but did not affect cell proliferation.
Table 6-19 Characteristics of lexicological studies of carcinogenicity and
genotoxicity with exposure to nitrogen dioxide.
Reference
Species (Strain);
Sample Size; Sex;
Age
Exposure Details
Endpoints Examined
Ichinose et al.
(1991)
Rats (Wistar); n = 30;
M; adult
40, 400, or 4,000 ppb NO2 for 17 mo.
Followed injection with carcinogen BHPN.
Incidence of BHPN-induced
lung tumors (adenoma or
adenocarcinomas)
Ichinose and
Saqai(1992)
Rats (Wistar); n = 36;
M; adult
(1) 400 ppb NO2 + 50 ppb Os
(2) 400 ppb NO2 + 1 mg/m3 H2SO4
(1-2) Exposure 11 h/day for 13 mo,
followed by 11 mo of clean air exposure.
Followed injection with carcinogen BHPN.
Incidence of BHPN-induced
lung tumors (adenoma or
adenocarcinomas)
Ohyama et al.
(1999)
Rats (F344); n = 26; M; 6,000 ppb NO2 for 16 h/day for 8 mo,
adult
followed by 8 mo of clean air exposure.
Co-exposure to DEPcCBP. IT installation 1
day/week for 4 weeks.
Lung tumor incidence
(alveolar adenomas)
Richters and
Damji (1990)
Adkins et al.
(1986)
Mice (AKR/cum);
n = 50; F; adult
Mice (A/J,
spontaneously high
250 ppb NO2 for 7 h/day, 5 days/week for
up to 26 weeks.
1,000,5,000, or 10,000 ppb NO2 for
6 h/day, 5 days/week for 6 mo.
Tumor progression and
survival rate
Lung tumor multiplicity
(pulmonary adenomas)
tumor rates); n = 30; F;
adult
Wagner et al.
(1965)
Mice (CAF1/Jax);
n = 20; M; adult
5,000 ppb NO2 continuously for 16 mo.
Lung tumor multiplicity
measured at 12, 14, and
16 mo.
Richters and
Kuraitis(1981)
Mice (Swiss Webster);
n = 24; M; adult
Mice (C57BL/6J);
n = 90; m; adult
400 or 800 ppb NO2 for 8 h/day,
5 days/week for 10 weeks (Swiss mice) or
12 weeks (C57BL/6J mice).
Both strains then infused i.v. with B16
melanoma cells that are known to
metastasize to the lung.
Facilitation of lung tumor
metastasis (incidence of lung
tumors) 3 weeks after
infusion of melanoma cells
Richters et al.
(1985)
Mice (C57BL/6J);
n = 48; M; adult
400 ppb NO2 continuously for 12 weeks.
All animals infused i.v. with B16 melanoma
cells that are known to metastasize to the
lung.
Facilitation of lung tumor
metastasis (incidence of lung
tumors) 3 weeks after
infusion of melanoma cells
6-167
-------
Table 6-19 (Continued): Characteristics of animal toxicological studies of
carcinogenicity and genotoxicity with exposure to
nitrogen dioxide.
Reference
Species (Strain);
Sample Size; Sex;
Age
Exposure Details
Endpoints Examined
Richters and
Kuraitis(1983)
fKoehleret al.
(2013)
fKoehleret al.
(2010)
fKoehleret al.
(2011)
Mice (C57BL/6J); 300, 400, or 500 ppb NO2 for 7 h/day,
n = 25, 51 , 23; M; adult 5 days/week for 1 0 weeks.
All animals then infused i.v. with B16
melanoma cells that are known to
metastasize to the lung.
Human cells 10 ppb NCbforO, 0.5, 1, 2 and 3 h.
Ex vivo cell culture at the air-liquid
interface, primary human nasal epithelia
cells from n = 10 donors.
Human cells 10, 100, or 1,000 ppb NCbforO or 0.5 h.
Ex vivo cell culture at the air-liquid
interface, primary human nasal epithelia
cells from n = 10 donors.
Human cells 100 ppb NCbforO, 0.5, 1, 2, and 3 h.
Ex vivo cell culture at the air-liquid
interface, primary human nasal epithelia
cells from n = 10 donors.
Facilitation of lung tumor
metastasis (incidence of lung
tumors) 3 weeks after
infusion of melanoma cells
Comet assay, micronucleus
formation, proliferation
assay, apoptosis, necrosis,
cytotoxicity
Comet assay, micronucleus
formation, proliferation
assay, cytotoxicity
Comet assay, micronucleus
formation, proliferation
assay, apoptosis, necrosis,
cytotoxicity
BHPN = N-bis(2-hydroxypropvl) nitrosamine; DEPcCBP = diesel exhaust particle extract-coated carbon black
particles; IT = intratrachael; i.v. = intravenously; NO2 = nitrogen dioxide; Os = ozone.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
6.6.2.3 Summary of Lung Cancer
Multiple epidemiologic studies examined the associations of long-term average NO2 or
NOx concentrations with lung cancer incidence or mortality. Positive associations were
reported in multiple studies, but many studies reported no associations. The evidence is
more consistent for association of NC>2 exposure with lung cancer mortality. The
inconsistency in the evidence does not appear to be related to the length of exposure
period examined, length of follow-up period, or the method of exposure assessment.
Studies using central site monitors for exposure estimates can carry uncertainty because
the exposure error resulting from spatial misalignment between subjects and monitor
locations can overestimate or underestimate associations with health effects
(Section 3.4.5.2). However, a few studies estimated NO2 exposure from central sites
located near (e.g., within 1 km) subjects and/or in areas not influenced by heavy traffic,
and associations with lung cancer mortality were observed in some locations (Katanoda
et al.. 2011; Filleul etal.. 2005). Several studies estimated long-term NO2 exposure at
6-168
-------
subjects' homes with well-validated LUR or dispersion models. Predicted NO2
concentrations were demonstrated to correlate well with measured NO2 concentrations in
the study areas. These studies did not observe associations with lung cancer incidence
rVilleneuve et al.. 2014; Raaschou-Nielsen et al.. 2013a). but many observed associations
with lung cancer mortality (Carey et al., 2013; Cesaroni et al.. 2013; Jerrettet al., 2013;
Naess etal.. 2007). Most of these studies controlled for confounders, such as smoking
and age, but evaluation of confounding by important copollutants such as PM25, diesel
exhaust particles, polycyclic aromatic hydrocarbons, volatile organic compounds, and
other traffic-related pollutants generally is absent. Evidence from toxicological evidence
studies does little to address the uncertainty in epidemiologic studies. NO2 exposure is
not shown to independently induce lung tumors, and there is inconsistent evidence that
ambient-relevant NO2 exposures promote lung tumors when given in conjunction with a
known carcinogen. There is some evidence that NCh exposure may facilitate metastases
of tumors to the lung and induce some genotoxic effects in nasal cells.
6.6.3 Leukemia Incidence and Mortality
Leukemia was not examined in studies available for the 2008 ISA for Oxides of Nitrogen
(U.S. EPA. 2008c). but was examined in recent studies. Studies in children examining
estimates of prenatal NO2 exposure at the residence with well-validated LUR models did
not consistently observe associations with leukemia incidence. Another uncertainty
relates to the lack of examination of PM25, diesel exhaust, polycyclic aromatic
hydrocarbons, or other traffic-related pollutants to assess potential confounding by key
copollutants. Among children in Los Angeles County, CA, NC>2, NO, and NOx
concentrations averaged over the prenatal period were associated with incidence of acute
lymphoblastic leukemia between ages 0 and 5 years (Ghosh etal.. 2013). The cross-
validation R2 of the LUR models exceeded 0.87. A positive association was observed for
second trimester-average NO2 but not for the first or third trimesters. None of the oxides
of nitrogen was associated with acute myeloid leukemia, which comprised 13% of
leukemia cases in the study population. A similar case-control study in Italy observed no
association between prenatal NO2 exposure and leukemia incidence (Badaloni et al..
2013). This was true in analyses limited to children aged 0-4 years and children who
never moved. Controls were randomly selected from the population and matched to cases
by age, sex, and region of residence. Although the analysis was based on all leukemia,
acute lymphoblastic leukemia comprised 87% of cases. The authors also reported that
LUR models were validated. No association was observed with PM2 5 either, which was
highly correlated with PM2s (r = 0.78).
6-169
-------
Other studies in children reported associations between NO2 and leukemia incidence or
mortality but have more uncertainty regarding the exposure estimates. Information was
not reported on whether a smoothed map of NO2 concentrations with a 4 km2 resolution
(Amigou et al.. 2011). or concentrations from one central site in a municipality (Weng et
al., 2008) adequately represented the spatial variability of concentrations in the study
areas. Another uncertainty is the examination of exposure around the time of diagnosis.
Information was not reported on the extent to which recent exposure represented
exposure during development of leukemia.
A study in Denmark examined adults and reported no association between NOx and
leukemia incidence (Raaschou-Nielsen et al.. 201 la). NOx exposure was estimated for
subjects' residences from a dispersion model. Modeled estimates agreed well with
measured concentrations (r = 0.90) and also were highly correlated with other traffic-
related pollutants (e.g., r = 0.93 with particle number concentration).
6.6.4 Bladder Cancer Incidence and Mortality
Similar to leukemia, recent studies provide information on potential associations of
oxides of nitrogen and bladder cancer not available for the 2008 ISA for Oxides of
Nitrogen (U.S. EPA. 2008c). Similar to leukemia, a study of adults in Denmark observed
no association of bladder cancer incidence and estimates of residential NOx exposure
(Raaschou-Nielsen et al.. 201 la).
A study performed in Taiwan examined mortality records, comparing individuals
(matched on sex, year of birth, and year of death) with and without mortality due to
bladder cancer (Liu et al.. 2009a). Increased odds of bladder cancer mortality was
associated with increased NO2 concentrations. This trend was also observed for SO2. The
highest tertile of PMio concentration was also associated with bladder cancer mortality,
but no association was observed for CO or Os concentrations. Elevated ORs also were
observed for an index combining higher tertiles of NO2 and SO2 concentrations (relative
to the first tertile of <4.32 ppb NO2 and <20.99 ppb SO2). Although the point estimates
for NO2 and SO2 combined are higher than those observed for NO2 or SO2 alone
[Supplemental Table S6-10; (U.S. EPA. 2013o)1. the 95% CIs overlap. Therefore, it is
unclear whether NO2 and SO2 combined contribute to higher odds of mortality than either
pollutant on its own. More importantly, an independent association for NO2 is not
discernable as confounding by PM2 5 or traffic-related copollutants was not examined. An
additional uncertainty is the extent to which concurrent NO2 concentrations at one central
site monitor in each municipality adequately represent exposure contrasts between
subjects.
6-170
-------
6.6.5 Breast Cancer Incidence
Recent studies also examined associations of NO2 or NOx with breast cancer but have
uncertainty in inference from their results. A Canadian study of post-menopausal breast
cancer incidence using a hospital-based case-control study design estimated NO2
exposure at residential addresses from LUR models and from concentrations at central
site monitors combined by IDW (Grouse et al.. 2010). Modeled estimates of NO2 agreed
better with concentrations at sites left out of the model (r = 0.90) than with concentrations
at 9 central site monitors (R2 = 0.56). For NC>2 estimated from central sites and LUR, ORs
were elevated for annual average NO2 around the time of diagnosis and 10-years prior.
However, the wide 95% CIs produce uncertainty in the association for annual average
NO2 10 years before diagnosis, a period more relevant for cancer development. Exposure
was estimated for the current residential address, and to improve correspondence to past
exposure, sensitivity analyses were limited to subjects who were residents of the same
address for at least 10 years prior to the study. ORs remained elevated, but 95% CIs were
wide. Another uncertainty is potential bias in results due to the selection of controls,
which were hospital based and limit the generalizability of the study.
As with leukemia and bladder cancer, a study in Denmark observed no association
between NOx and breast cancer incidence (Raaschou-Nielsen et al.. 201 la). This study,
unlike the one performed in Canada, included all breast cancer cases instead of limiting
to post-menopausal cases. In an ecologic study of the U.S. by Wei et al. (2012). states
with the higher NOx emissions had higher breast cancer incidence rates (r = 0.89).
Similar correlations were observed for emissions of the traffic-related pollutants CO and
VOCs as well as PMio and SO2. This study is limited by its ecologic nature and the lack
of individual-level data. There is no control for potential confounding factors.
6.6.6 Prostate Cancer Incidence
Studies of prostate cancer were not available for the 2008 ISA for Oxides of Nitrogen
(U.S. EPA. 2008c). However, a recent study indicates an association with NO2 exposure
among men in Montreal, Canada (Parent et al.. 2013). Cases were recruited through
pathology departments. Population-based controls were identified through electoral lists
and frequency matched by 5-year age groups. NO2 exposure was estimated for the period
around diagnosis with an LUR model that predicted well NO2 concentrations at sites left
out of the model (r = 0.90). Prostate cancer also was associated with NO2 exposures
estimated by back-extrapolation for 10 and 20 years before diagnosis. Multiple sensitivity
analyses were performed, including the addition of smoking and alcohol consumption as
confounders, exclusion of proxy subjects, exclusion of subjects without a prostate cancer
6-171
-------
screening in the past 5 years, exclusion of subjects at their residence for less than
10 years, and comparisons of subjects with geocoding to their exact address or to a
centroid of their postal code. ORs were positive, although the wide 95% CIs in some
cases indicated uncertainty in the associations. A key uncertainty in this study is
confounding by PIVb 5 and traffic-related copollutants. NCh was the only pollutant
examined. As with other cancers, prostate cancer was not associated with NOx among
men in Denmark (Raaschou-Nielsen et al., 201 la).
6.6.7 Other Cancer Incidences and Mortality
A few recent studies examined incidence or mortality of an array of cancers, including
cervical, brain, buccal cavity/pharynx, esophageal, stomach, colon, rectal, liver,
pancreatic, laryngeal, uterine, ovarian, kidney, melanoma, non-Hodgkin lymphoma, and
myeloma. An array of tumor types were examined, including central nervous system
tumors, intracranial and intraspinal embryonal tumors, primitive neuroectodermal tumor,
other gliomas, neuroblastoma, retinoblastoma, unilateral retinoblastoma, Wilms tumor,
hepatoblastoma, rhabdomyosarcomas, germ cell tumors, extracranial and extragonadal
germ cell tumors, and teratoma). A few positive association were observed between NOx
and cervical or brain cancer in adults (Raaschou-Nielsen et al.. 201 la) or NOx and
retinoblastoma in children (Ghosh et al.. 2013). However, most associations were null,
including those forNO2 exposure. Both of these studies estimated NOx or NO2 exposure
with well-validated models. A study in Canada observed a positive association between
daily NO2 concentrations and mortality from all types of cancer combined (Goldberg et
al.. 2013). No association was reported for CO or PIVb 5; however, the biological
relevance of daily exposure to mortality from all types of cancer combined is not clear.
6.6.8 Genotoxicity
6.6.8.1 lexicological Studies
Results from toxicological studies do not clearly indicate that NO2 exposure induces
mutations or genotoxic effects. Few studies examined ambient-relevant NO2 exposures,
and findings are mixed even for higher than ambient-relevant exposures. Of the in vivo
assays reported in the 2008 ISA for Oxides of Nitrogen [(U.S. EPA. 2008a). Annex
Tables AX4-11, AX4-12, and AX4-13], two studies observed mutations or chromosomal
abnormalities in rat lung cells (50,000-560,000 ppb NO2 >12 days; 27,000 ppb NO2,
3 hours). No genotoxicity was seen in tests employing Drosophila recessive lethals
6-172
-------
(500,000-7,000,000 ppb NO2, 1 hour), Drosophila wing spot test (50,000-280,000 ppb
NO2, 2 days), or mouse bone marrow micronuclei (20,000 ppb, 23 hours). In vitro
exposures to NO2 yielded positive findings in a majority of the tests in rodent
(2,000-3,000 ppb NO2, 10 minutes) and human cell lines, bacteria (5,000-90,000 ppb
NO2, 30 minutes), and plants (5,000 ppb NO2, 24 hours).
Evidence for genotoxicity is equally inconclusive in limited examination of ambient-
relevant NO2 exposures. NO2 exposures of 100-5,000 ppb (6 hours) did not induce
chromosomal aberrations in mouse spermatocytes or lymphocytes [(U.S. EPA. 2008a).
Annex Table AX4-12]. A recent study observed that male rats exposed to 2,660 or
5,320 ppb NO2 for 6 h/day for 7 days had increases in DNA damage in the liver, lung,
and kidney, increases in micronuclei formation in bone marrow, and increases in
DNA-protein cross-links in the brain and liver (Han et al., 2013). Exposure to 5,320 ppb
NO2 also induced DNA damage in the spleen and heart.
The 2008 ISA for Oxides of Nitrogen discussed the possibility that NO2 could produce
cancer via nitrosamine formation (U.S. EPA, 2008c). N-nitroso compounds can be
generated endogenously in the human body from NO2 via processes that generate nitrite
(NO2 ) or nitrate. Further, NO2 is known to react with amines to produce nitrosamines,
known animal carcinogens. Such reactions have not been demonstrated with ambient-
relevant exposures. Peroxyacetyl nitrate (PAN) is an oxide of nitrogen produced by
reactions involving ultraviolet, NO2, and hydrocarbons (Section 2.2). PAN is weakly
mutagenic in the lungs of the highly susceptible big Blue (R) mice and in Salmonella and
produces a unique signature mutation (DeMarini et al., 2000). But, such effects were
observed with PAN exposures more than five order of magnitude above typical ambient
concentrations (Section 2.4.1).
6.6.8.2 Epidemiologic Studies
Similar to toxicological evidence, a recent epidemiologic study observed mixed evidence
for associations between NO2 exposure and genotoxicity in children in Italy living near
chipboard industries (Marcon et al., 2014). NO2 was measured 3 km from the industries,
and residential exposure was estimated by kriging. NO2 exposure was not associated with
results of comet assays (i.e., tail length, moment, or intensity) but was associated with
some results of micronucleus assays. A 10-ppb increase in NO2 concentration was
associated with a 1.16% (95% CI: 0.6, 1.7) change in binucleated cells and an increased
risk ratio of nuclear buds of 3.72 (95% CI: 1.67, 7.73). Formaldehyde was associated
with nuclear buds and comet tail intensity and moment but not with binucleated cells.
Potential confounding of NO2 associations was not examined.
6-173
-------
6.6.9 Summary and Causal Determination
The evidence for long-term NC>2 exposure and cancer is suggestive of, but not sufficient
to infer, a causal relationship. This conclusion is based primarily on some, but not
entirely consistent, epidemiologic and toxicological evidence for lung cancer. There is
not toxicological evidence that NO2 at ambient-relevant exposure concentrations is a
direct carcinogen. A few recent epidemiologic studies each indicate associations of NCh
exposure with leukemia, prostate cancer, and bladder cancer mortality. However, none of
these epidemiologic studies examined potential confounding by key traffic-related
pollutants, and biological plausibility from toxicological studies is lacking.
The current conclusion represents a change from the "inadequate to infer the presence or
absence of a causal relationship" determination made in the 2008 ISA for Oxides of
Nitrogen (U.S. EPA. 2008c). At that time, there was some toxicological evidence that
NC>2 exposure may act as a tumor promotor, and a few epidemiologic studies indicating
associations between long-term NO2 exposure and lung cancer incidence. The change to
"suggestive, but not sufficient, to infer a causal relationship" in this ISA is based on the
positive associations observed in some epidemiologic studies between NO2 exposure and
lung cancer incidence and mortality studies and the limited coherence provided by some
toxicological studies. The key evidence as it relates to the causal determination is
summarized in Table 6-20 using the framework described in Table II of the Preamble.
Epidemiologic studies of lung cancer incidence have mixed results, with some studies
reporting no associations and others reporting positive associations. Evidence is
inconsistent particularly among studies that estimated subjects' residential NO2 exposure
with well-validated models. Although epidemiologic findings for NCh-related lung
cancer mortality are not entirely consistent, more studies reported positive associations.
Many studies of lung cancer mortality estimated NC>2 exposure with LUR or dispersion
models that were shown to predict well ambient NO2 concentrations in the study areas.
Most of these studies included large sample sizes and aimed to account for many
potential confounders, including age, SES, and smoking exposures. A major limitation of
the epidemiologic evidence is that investigation into potential confounding by PM2 5,
diesel exhaust, polycyclic aromatic hydrocarbons, and VOCs is largely absent. In fact, in
many epidemiologic studies, NC>2 was the only pollutant examined. The uncertainty as to
whether NCh exposure independently is related to lung cancer also exists in the
toxicological evidence. Toxicological data provide no clear evidence of NO2 acting as a
complete carcinogen, and agencies that classify carcinogens including the Department of
Health and Human Services, the International Agency for Research on Cancer, and the
EPA have not classified oxides of nitrogen for potential carcinogenicity. The American
Conference of Industrial Hygienists has classified NC>2 as A4 (not classifiable for humans
6-174
-------
or animals). Genotoxic and mutagenic studies with NC>2 have mixed results. However,
some studies showed that NCh may act as a tumor promoter at the site of contact. Co-
exposure to a known carcinogen (nitrosamine), exposure of rodents with spontaneously
high tumor rates, or exposure of rodents injected with metastatic cancer cells ambient-
relevant NC>2 exposures increased lung tumor burden or the number of rodents with lung
tumors. A role in tumor promotion may be possible due to an effect of NC>2 on producing
cellular damage, inducing respiratory epithelial hyperplasia (Section 6.2.6). or promoting
regenerative cell proliferation.
In conclusion, while some studies observed no associations, some epidemiologic findings
point to lung cancer incidence and mortality in association with NC>2 exposure estimated
using well-validated models. Epidemiologic studies largely did not examine confounding
by key traffic-related copollutants, and toxicological studies provide no clear evidence of
NC>2 acting as a complete carcinogen. However, some toxicological studies in tumor-
prone rodents or with co-exposure with a known carcinogen indicate that ambient-
relevant NCh exposures can increase lung tumor incidence. The evidence from some
epidemiologic studies for lung cancer incidence and mortality combined with some
toxicological evidence for lung tumor promotion is suggestive of, but not sufficient to
infer, a causal relationship between long-term exposure to NO2 and cancer.
6-175
-------
Table 6-20 Summary of evidence, which is suggestive of, but not sufficient to
infer, a causal relationship between long-term nitrogen dioxide
exposure and cancer.
Rationale for Causal
Determination3
Key Evidence13
Key References'3
NO2 Concentrations
Associated with
Effects0
Lung Cancer
Evidence from
epidemiologic studies
generally supportive but
not consistent
Positive associations with lung
cancer incidence or mortality in
multiple studies conducted in
the U.S., Canada, and Europe.
Lack of examination of PIVh.s or
traffic-related copollutants.
Nvberq et al. (2000),
tCesaroni et al. (2013). Filleul
et al. (2005), tCarey et al.
(2013), tJerrett et al. (2013),
tHvstad etal. (2013)
Section 6.6.2
Means: 12, 23.2 ppb for
annual avg, 12.3, 15.4
ppb for multi-yr avg,
6.4-32.4 ppb across
cities in multicity study
No associations with lung tRaaschou-Nielsen et al.
cancer incidence or mortality in (2013a). tBrunekreef et al.
multiple studies conducted in (2009). tPapathomas et al.
the U.S. and Europe. (2011), tBeelen et al.
(2008b), tHartetal. (2011)
Means: 14.2 ppb for 15-
yr avg, 19.6 ppb for 20-yr
avg, 2.8-31.8 ppb for
annual avg across cities
in multicity study
Lack of toxicological
evidence for direct
effect on carcinogenesis
but limited evidence for
tumor promotion with
ambient-relevant
exposures
NO2 exposure alone does not
induce lung tumors. Mixed
evidence for tumor promotion
with co-exposure to known
carcinogen. Limited evidence
for facilitation of metastasis.
Adkins etal. (1986). Richters
and Damji (1990), Wagner et
al. (1965). Richters and
Kuraitis (1981). Richters and
Kuraitis (1983), Ichinose et al.
(1991)
Section 6.6.2.2
Tumor promotion with
nitrosamine exposure:
inconsistent 250-5,000
ppb for 6-17 mo.
Facilitation of
metastases: 300-800
ppbfor10or12 weeks
Limited evidence for key
events in proposed
mode of action
Finding of mutagenicity and
micronucleus formation in
ex vivo culture of primary
human nasal epithelial cells
exposed to NO2.
tKoehler et al. (2013),
tKoehler et al. (2011).
tKoehler etal. (2010)
Section 6.6.2.2
10, 100, 1,000,
10,000 ppb
Cancers of Sites Outside the Lung
Limited epidemiologic
evidence and
uncertainty regarding
independent effect of
NO2
Weak evidence for key
events in proposed
mode of action
Positive associations with tLiu et al. (2009a). tParent et Means or medians:
leukemia, bladder cancer, and al. (2013), tGhosh etal. 17-24 ppb for annual
prostate cancer. Lack of (2013) avg
examination of PM2.5 or traffic- Sections 6.6.3, 6.6.4, 6.6.6
related copollutants.
No mutagenicity or genotoxicity
in animal models with ambient-
relevant exposure. Mixed
evidence with higher exposure.
2008 NOx ISA Annex Tables
AX4-11, AX 4-12, and AX4-13
(U.S. EPA. 2008a)
h = hours; min = minutes; mo = months; NO2 = nitrogen dioxide; ppb = parts per billion; yr = yr.
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Tables I and N. of the
Preamble.
""Describes the key evidence and references, supporting and contradicting, contributing most heavily to causal determination and,
where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where the full body of evidence is
described.
°Describes the NO2 concentrations with which the evidence is substantiated (for experimental studies, < 5,000 ppb).
fStudies published since the 2008 ISA for Oxides of Nitrogen.
6-176
-------
CHAPTER 7 POPULATIONS AND LIFESTAGES
POTENTIALLY AT RISK FOR
HEALTH EFFECTS RELATED TO
NITROGEN DIOXIDE EXPOSURE
7.1 Introduction
This chapter aims to characterize populations and lifestages that may be at increased risk
for detrimental health effects in response to ambient NO2 exposure, using the approach
developed for the 2013 ISAs for Ozone and Lead (U.S. EPA. 2013c. e). The NAAQS are
intended to protect public health with an adequate margin of safety. Protection is
provided for both the population as a whole and those potentially at increased risk for
health effects in response to exposure to a criteria air pollutant (Preface to the ISA). The
scientific literature has used a variety of terms to describe populations at increased risk of
an air pollutant-related health effect or factors that increase people's risk. These terms
include susceptible, vulnerable, sensitive, and at-risk, with recent literature introducing
the term response-modifying factor (Vinikoor-Imler et al.. 2014) (Preamble to the ISA).
Due to the inconsistency in definitions for these terms across the scientific literature and
the lack of a consensus on terminology in the scientific community, as detailed in the
Preamble, recent ISAs have used the broadly applicable term "at-risk" population or
lifestage. Thus, this chapter identifies, evaluates, and characterizes at-risk factors to
address the main question of what populations and lifestages are at increased risk of an
NO2-related health effect. Some factors may reduce risk, and these are acknowledged in
this evaluation. However, for the purposes of identifying those populations or lifestages
at greater risk to inform decisions on the NAAQS, the focus of this chapter is on
characterizing those factors that may increase risk.
Individuals, and ultimately populations, could be at increased risk of an air
pollutant-related health effect for various reasons. As discussed in the Preamble, risk may
be modified by intrinsic or extrinsic factors, differences in internal dose, or differences in
exposure to air pollutant concentrations. The objective of this chapter is to identify,
evaluate, and characterize the evidence for factors that potentially increase the risk of
health effects related to exposure to NO2, regardless of whether the change in risk is due
to intrinsic factors, extrinsic factors, increased internal dose, increased exposure, or a
combination. It is important to note that although risk of an NC>2-related health effect is
evaluated for each factor individually, it is likely that portions of the population are at
increased risk of an NCh-related health effect due to a combination of co-occurring
factors (e.g., residential location and SES). However, information on the interaction
7-1
-------
among factors remains limited. Thus, the following sections characterize the overall
confidence for a particular factor to result in increased risk for NC^-related health effects.
7.2 Approach to Evaluating and Characterizing the Evidence for
At-Risk Factors
The systematic approach used to evaluate factors that may increase the risk of a
population or specific lifestage to an air pollutant-related health effect is described in
more detail in the Preamble. The evaluation emphasizes relevant health studies discussed
in Chapter 5 and Chapter 6 of this ISA building on the evidence presented in the 2008
ISA for Oxides of Nitrogen (U.S. EPA. 2008c) and the 1993 Air Quality Criteria for
Oxides of Nitrogen (U.S. EPA. 1993a). Based on the approach developed for the ISAs for
Ozone and Lead (U.S. EPA. 2013c. e), evidence is integrated across scientific disciplines,
across health effects, and where available, with information on exposure and dosimetry
(Chapters 3 and 4). Conclusions are drawn based on the overall confidence that a specific
factor may characterize a population or lifestage at increased risk of an NO2-related
health effect.
As discussed in the Preamble, this evaluation focuses on health effect studies that conduct
stratified analyses to compare populations or lifestages exposed to similar air pollutant
concentrations within the same study design. For the evaluation of these studies,
important considerations include whether the stratified analyses were planned a priori or
were post hoc analyses, whether the study conducted multiple comparisons, and whether
there were small sample sizes in individual strata. These study design issues can increase
the probability of finding associations by chance or reduce power to detect associations in
subgroup analyses. Experimental studies that examine health effects exclusively in
human subjects or animal models with a particular factor, such as pre-existing disease,
are also important lines of evidence to evaluate because they inform judgments of the
coherence and biological plausibility of effects observed in epidemiologic studies, as well
as the independent effect of NO2 exposure. Additionally, studies examining whether
factors may result in differential exposure to NO2 are included. Excluded from this
evaluation are a small group of studies that examined potential at-risk populations or
lifestages based on NOx exposures because they do not directly inform understanding of
whether a population is at increased risk of an NO2-related health effect (Section 1.1).
The objective of this chapter is not to detail individual study results but to integrate
information in preceding chapters and describe the basis for characterizing the evidence
for each factor to increase risk of NO2-related health effects as adequate, suggestive,
inadequate, or no effect (Table 7-1). Classifications are based on the cumulative evidence
7-2
-------
available for each factor, regardless of whether the risk is due to increased exposure,
internal dose, biological effect, or a combination of reasons. The classification provides a
consistent basis for evaluating the evidence for an at-risk population or lifestage and for
describing the overall confidence in the evidence. The classifications are not direct
assessments of causation but rather build on causal determinations to inform decisions on
whether a NAAQS provides an adequate margin of safety across groups in the
population. Thus, for NO2, judgments about at-risk factors emphasize evidence for
asthma exacerbation and asthma development, which are the basis for concluding that
relationships of short-term (Section 5.2.9) and long-term (Section 6.2.9) NC>2 exposure
with respiratory effects, respectively, are causal and likely to be causal. These are the
health effects for which there is the strongest evidence for NO2 exposure having effects
independent of traffic-related copollutants. The factors that are evaluated in this chapter
include pre-existing diseases and conditions (Section 7.3). genetic factors (Section 7.4).
sociodemographic factors (Section 7.5). and behavioral and other factors (Section 7.6).
Section 7.7 summarizes the evidence for each factor to increase risk of NO2-related
health effects.
Table 7-1 Characterization of evidence for factors potentially increasing the
risk for nitrogen dioxide-related health effects.
Classification Health Effects
Adequate There is substantial, consistent evidence within a discipline to conclude that a factor results in a
evidence population or lifestage being at increased risk of air pollutant-related health effect(s) relative to
some reference population or lifestage. Where applicable, this evidence includes coherence
across disciplines. Evidence includes multiple high-quality studies.
Suggestive The collective evidence suggests that a factor results in a population or lifestage being at
evidence increased risk of air pollutant-related health effect(s) relative to some reference population or
lifestage, but the evidence is limited due to some inconsistency within a discipline or, where
applicable, a lack of coherence across disciplines.
Inadequate The collective evidence is inadequate to determine whether a factor results in a population or
evidence lifestage being at increased risk of air pollutant-related health effect(s) relative to some reference
population or lifestage. The available studies are of insufficient quantity, quality, consistency,
and/or statistical power to permit a conclusion to be drawn.
Evidence of no There is substantial, consistent evidence within a discipline to conclude that a factor does not
effect result in a population or lifestage being at increased risk of air pollutant-related health effect(s)
relative to some reference population or lifestage. Where applicable, the evidence includes
coherence across disciplines. Evidence includes multiple high-quality studies.
7-3
-------
7.3 Pre-Existing Disease/Conditions
Individuals with pre-existing disease may be considered at greater risk for some air
pollution-related health effects because they may be in a compromised biological state
depending on the disease and severity. The 2008 ISA for Oxides of Nitrogen (U.S. EPA.
2008c) concluded that those with pre-existing pulmonary conditions were likely to be at
greater risk for NC>2-related health effects, especially individuals with asthma. Among
recent studies evaluating effect measure modification by pre-existing disease, the largest
group examined asthma [Section 7.3.1). Several studies are available on other diseases:
COPD (Section 7.3.2). CVD (Section 7.3.3). diabetes (Section 7.3.4). and obesity
(Section 7.3.5). Table 7-2 presents the prevalence of these diseases according to the
Centers for Disease Control's National Center for Health Statistics (Blackwell et al..
2014). including the proportion of adults with a current diagnosis categorized by age and
geographic region. The large proportions of the U.S. population affected by many chronic
diseases, including various cardiovascular diseases, indicate the potential importance of
characterizing the risk of NC>2-related health effects for affected populations.
7-4
-------
Table 7-2 Prevalence of selected respiratory, cardiovascular, and metabolic
diseases and disorders among adults by age and region in the
United States in 2012.
Chronic
Disease/Condition
All (n, in thousands)
Selected respiratory
Asthmac
COPD
Adults (18+)
n
(in thousands)
234,921
diseases
18,719
6,790
18-44
111,034
8.1
0.5
Age(
45-64
82,038
8.4
3.8
%)»
65-74
23,760
7.8
6.9
75+
18,089
6.0
8.6
North-
east
42,760
9.2
2.2
Region
Midwest
53,378
8.1
3.2
(%)b
South
85,578
7.3
3.0
West
53,205
7.8
2.1
Selected cardiovascular diseases/conditions
All heart disease
Coronary heart
disease
Hypertension
Stroke
26,561
15,281
59,830
6,370
3.8
0.9
8.3
0.6
12.1
7.1
33.7
2.8
24.4
16.2
52.3
6.3
36.9
25.8
59.2
10.7
10.0
5.3
21.4
1.8
11.6
6.5
24.1
2.5
11.6
7.0
26.6
3.0
9.3
5.1
21.5
2.5
Selected metabolic disorders/conditions
Diabetes
Obesity
(BMI > 30 kg/m2)
Overweight
(BMI 25-30 kg/m2)
21,319
64,117
78,455
2.4
26.0
31.4
12.7
33.7
36.8
21.1
29.7
40.7
19.8
18.0
38.6
7.6
25.1
34.3
8.4
29.9
34.1
10.0
29.9
34.2
7.3
25.2
35.3
BMI = body mass index; COPD = chronic obstructive pulmonary disease.
"•Percentage of individual adults within each age group with disease, based on n (at the top of each age column).
Percentage of individual adults within each geographic region with disease, based on n (at the top of each region column).
°Asthma prevalence is reported for "still has asthma."
Source: Blackwell et al. (2014): National Center for Health Statistics: Data from Tables 1 and 2; Tables 3 and 4; Tables 7 and 8;
and Tables 28 and 29 of the Centers for Disease Control and Prevention report.
7.3.1 Asthma
Approximately 8.0% of adults and 9.3% of children (age <18 years) in the U.S. currently
have asthma (Blackwell et al.. 2014; Bloom et al.. 2013), and it is the leading chronic
illness affecting children. This ISA concludes that there is a causal relationship between
short-term NC>2 exposure and respiratory effects, based primarily on evidence for effects
on asthma exacerbation (Section 5.2.9). The evidence demonstrating NCh-induced
7-5
-------
asthma exacerbation was the basis for the 2008 ISA for Oxides of Nitrogen (U.S. EPA.
2008c) concluding that individuals with pre-existing pulmonary conditions, particularly
those with asthma, are likely at greater risk of NCh-related health effects. The current
evaluation includes controlled human exposure and epidemiologic studies that compared
NC>2-related health effects between groups with or without asthma (Tables 7-3 and 7-4).
As a whole, results of these comparisons are variable; however, controlled human
exposure studies provide compelling evidence that people with asthma are at greater risk
for NO2-related respiratory effects than people without asthma.
Information is sparse to determine whether NC>2 exposure or internal dose differs between
people with and without asthma. Epidemiologic studies comparing NCh-related health
effects between children with and without asthma did not have measures of NCh
exposure for individual children; such measures were not required because the study
designs relied on temporal variation in short-term NC>2 exposure. People with asthma
tend to have oronasal breathing; however, the implications on differential uptake of NO2
in the respiratory tract have not been examined (Section 4.2.2.3). Based on limited study,
some data suggest that children with asthma spend less time outdoors on high air
pollution days (Mansfield et al.. 2006): other data point to children with asthma spending
more time outdoors in general but with less vigorous activity in summer (Avol et al..
1998). These limited data on outdoor activity also do not clearly indicate whether people
with asthma could have higher NO2 exposure or internal dose than people without
asthma. Nor are the data conclusive in indicating whether epidemiologic studies are
subject to greater exposure measurement error for people with asthma and thus less likely
to detect differences in NO2-related health effects.
Controlled human exposure studies demonstrating NC>2-induced increases in airway
responsiveness in adults with asthma provide key evidence for an independent, causal
relationship between NO2 exposure and respiratory effects (Section 5.2.9). This evidence
also demonstrates greater sensitivity of adults with asthma to short-term NO2 exposure
compared to adults without asthma. A meta-analysis conducted by Folinsbee (1992)
demonstrates that NO2 exposures in the range of 100-300 ppb increased airway
responsiveness in adults with asthma (Table 7-3). The meta-analysis combined groups
that varied with respect to respiratory symptoms and medication use at the time of
assessment but similarly had high prevalence (50-100%) of atopic asthma. Additionally,
in many studies, these participants were characterized as having mild asthma. Few studies
of healthy adults examined airway responsiveness forNC>2 exposures below 300 ppb, and
results indicate increased airway responsiveness in healthy adults only for NC>2 exposures
>1,000 ppb. In comparison to airway responsiveness, there is inconsistent evidence for
the effects of short-term NO2 exposure on lung function in adults with asthma in the
absence of a bronchoconstricting agent, and the limited evidence is inconclusive in
7-6
-------
demonstrating differences in response between healthy adults and those with asthma
KVagaggini et al.. 1996: Torres etal.. 1995: Linnetal.. 1985b): Table 7-31.
Epidemiologic evidence does not clearly show differences in NO2-related health effects
between children with and without asthma (Table 7-4). Evidence is inconsistent in studies
that explicitly noted a priori comparisons (Patel et al.. 2010: Barraza-Villarreal et al..
2008: Holguin et al., 2007) as well as studies judged to have strong exposure assessment,
such as those monitoring NC>2 at or near children's schools (Lin et al.. 2011: Flamant-
HulinetaL 2010: Holguin et al., 2007). Some of these studies had large numbers of
children with asthma for comparison. Effect measure modification by asthma also is
inconsistent for cardiovascular and metabolic effects (Table 7-4). although inference is
weak because of uncertainty as to whether NO2 exposure has independent relationships
with these effects (Section 6.3.9). A notable difference between epidemiologic and
controlled human exposure studies comparing people with and without asthma is that the
epidemiologic studies examined a more diverse set of respiratory outcomes and asthma
phenotypes. Asthma is a heterogeneous disease as demonstrated in Section 5.2.2. One
limitation of epidemiologic studies that may obscure potential differences among people
with and without asthma is that these studies commonly analyze all people with asthma
together, even though there are varying phenotypes and triggers of asthma as indicated by
varying atopy and asthma medication use.
Several lines of evidence indicate that people with asthma are at increased risk for
NO2-related health effects. The causal relationship determined for short-term NCh
exposure and respiratory effects is based on the evidence for asthma exacerbation
(Section 5.2.9). Controlled human exposure studies demonstrate that NO2 has an
independent effect on increasing airway responsiveness in adults with asthma and show
increased sensitivity of adults with asthma compared to healthy adults. Epidemiologic
studies do not clearly demonstrate differences between populations with and without
asthma. The study populations represent an array of asthma phenotypes, and limited
epidemiologic evidence indicates larger NCh-related respiratory effects in children with
asthma not using asthma medication. Thus, the epidemiologic results are not necessarily
incoherent with experimental findings from populations of mostly mild, atopic asthma.
Information is insufficient to determine whether the increased risk for people with asthma
is attributable to higher NO2 exposure or internal dose. There is clear evidence for an
effect of NC>2 exposure on asthma exacerbation and for increased sensitivity of adults
with asthma to NC>2-induced increases in airway responsiveness in controlled human
exposure studies. Therefore, evidence is adequate to conclude that people with asthma are
at increased risk for NCh-related health effects.
7-7
-------
Table 7-3 Controlled human exposure studies evaluating pre-existing asthma.
Factor
Evaluated
Reference
Category
Asthma-related outcomes
Asthma
n = 33
<300 ppb
NO2
Asthma
n = 12
Asthma
n=4
Asthma
n=23
Atopic
asthma
Asthma
Asthma
Asthma
Healthy
n = 36
<1,000 ppb
NO2
Healthy
n = 8
Healthy
n = 7
Healthy
n = 25
None
None
None
None
Direction of
Effect
Modification
or Effect3 Outcome
and short-term exposure
* Airway
' respons-
iveness
_ Airway
inflamma-
tion
* Lung
' function
decrement
Lung
function
decrement
Airway
resistance
Lung
function
decrement,
allergen
response
_ Lung
function
decrement
| Lung
' function
decrement
_ Airway
resistance
Study
Population
n = 355
n=20
Ages
21-37yr
n = 11
Mean age
31. Syr
n=48
Ages
18-36yr
n = 11
Mean age
31.2yr
n=41
Mean age
31 yr
n = 15
Mean age
33 yr
n = 11
Ages 7-55 yr
Study Details
Range: 100 ppb NO2for
1 h to 7,500 ppb NO2 for
2 h of exposure;
Exposures at rest
1,000 ppb NO2 for 3 h;
Exercise 10 min
on/10 min off at
individual's maximum
workload
300 ppb NO2 for 1 h;
Exercise at VE = 25 L/min
4,000 ppb NO2 for 75 min;
Two 15-min periods of
exercise at VE = 25 L/min
and 50 L/min
(1)200ppbNO2for6h
(2) 200 ppb NO2 and
100 ppb OsforB h
(3) 400 ppb NO2 for 3 h
(4)400 ppb NO2 and
200 ppb O3 for 3 h
(1-4) Exercise 10 min
on/40 min off at
VE = 32 L/min
200ppbNO2for2h;
Exercise 15 min
on/15 min off at
VE = ~2 times resting
300 ppb NO2 for 30 min
Exercise 10 min
on/20 min off at
VE >3 times resting
250 ppb NO2for30 min;
Exercise 10 min
on/20 min off at VE 3 times
resting
Study
Folinsbee
(1992)
Jorres et al.
(1995)
Vaqaqqini et
al. (1996)
Linn et al.
(1985b)
Jenkins et al.
(1999)
Kleinman et al.
(1983)
Bauer et al.
(1986)
Jorres and
Maqnussen
(1991)
NO2 = nitrogen dioxide; O3 = ozone; VE = minute ventilation.
aUp facing arrow indicates that the effect of NO2 is greater (e.g., larger lung function decrement, larger increase in airway
inflammation) in the group with the factor evaluated than in the reference group. Down facing arrow indicates that the effect of NO2
is smaller in the group with the factor evaluated than in the reference group. A dash indicates no difference in NO2-related health
effect between groups. In some studies, only a population with pre-existing disease was examined; therefore, the arrow or dash
represents the direction of the effect in that population after exposure to NO2 relative to exposure to filtered air.
7-8
-------
Table 7-4 Epidemiologic studies evaluating pre-existing asthma.
Factor
Evaluated
Reference
Category
Asthma-related outcomes
Asthma
n = 57
Asthma
n = 50,
89% with
atopy
Asthma
n = 8
Asthma
n = 100
Asthma
n = 34
Asthma
n = 169
Nonasthma
Asthma
n = 50
Asthma
n = 1,273
No asthma
n = 192
No asthma
n - 158
72% with
atopy
No asthma
n = 30
No asthma
n = 100
No asthma
n - 70
No asthma
n - 2,071
outcomes and
No asthma
n = 1,934
No asthma
n = 50,545
Direction of
Effect
Modification3
and short-term
t
1
-
—
—
1
Outcome
exposure
Asthma
symptoms
Pulmonary
inflammation,
lung function
decrement
Pulmonary
inflammation
Pulmonary
inflammation
Pulmonary
inflammation
Pulmonary
inflammation
Study
Population
n=249
Ages 14-20 yr
n=208
Ages 7.9-11 yr
n = 38
Ages 9-12 yr
n=200
Ages 6-12 yr
n = 104
Mean age
10.3 yr
n = 2,240
Ages 5-7 yr
Study Details
New York, NY,
2003-2005
Mexico City,
2003-2005
Beijing, China,
2008
Ciudad Juarez,
Mexico,
2001-2002
Clermont-
Ferrand, France
southern
California,
2004-2005
Study
Patel et al.
(201 0)t
Barraza-
Villarreal et al.
(2008VT
Lin et al.
(201 1)t
Holquin et al.
(2007)t
Flamant-Hulin
etal. (2010)t
Berhane et al.
(201 1)t
long-term exposure
t
t
Incident stroke
Fatal stroke
Diabetes
n = 1,984
Ages 50-65 yr at
baseline
n = 51,818
Ages 50-65 yr at
baseline
Copenhagen,
Aarhus
counties,
Denmark,
1993-2006
Copenhagen,
Aarhus
counties,
Denmark,
1993-2006
Andersen et
al. (2012b)t
Andersen et
al. (2012c)t
aUp facing arrow indicates that the effect of nitrogen dioxide (NO2) is greater (e.g., larger lung function decrement, larger increase
in pulmonary inflammation) in the group with the factor evaluated than in the reference group. Down facing arrow indicates that the
effect of NO2 is smaller in the group with the factor evaluated than in the reference group. A dash indicates no difference in
NO2-related health effect between groups.
fStudies published since the 2008 ISA for Oxides of Nitrogen
7.3.2 Chronic Obstructive Pulmonary Disease
Chronic lower respiratory disease, including COPD, was ranked as the third leading
cause of death in the U.S. in 2011 (Hoyert and Xu. 2012). COPD comprises chronic
bronchitis and emphysema and affects approximately 6.8 million adults in the U.S.,
respectively (Table 7-2). Given that people with COPD have compromised respiratory
7-9
-------
function and systemic inflammation, they could be at increased risk for an array of
NO2-related health effects. Evidence for differential NCh-related respiratory or
cardiovascular effects between adults with COPD and those without COPD is weak
(Tables 7-5 and 7-6). Information on potential NC>2 exposure or dosimetry differences is
lacking. Similar total personal NC>2 exposure was measured in adults with COPD and
adults with prior myocardial infarction (MI), but no comparison was made to healthy
adults (Suh and Zanobetti 201 Ob).
Compared with asthma exacerbation, there is greater uncertainty regarding a relationship
between short-term NCh exposure and COPD exacerbation (Section 5.2.4).This is
illustrated by the lack of consistent evidence from controlled human exposure studies for
changes in lung function or pulmonary inflammation in adults with COPD following NO2
exposure (Gong et al.. 2005; Linn etal.. 1985a). Only a few studies compared adults with
COPD and healthy adults, and not all indicated larger NO2-induced decrements in lung
function in adults with COPD (Table 7-5). Among adults with COPD, NO2 exposures of
300 ppb for 1 or 4 hours induced decreases in lung function of 4.8, 8.2 (Morrow et al..
1992) or 10% (Vagaggini et al., 1996) relative to air control exposures. In contrast, in
healthy adults, NO2 did not have any effect on lung function or resulted in increased lung
function. However, adults with COPD were older than healthy adults and/or had a higher
prevalence of smoking, which could have influenced results (Table 7-5). For example, in
one study, smokers had larger NO2-induced decrements in lung function independently of
COPD (Morrow etal.. 1992).
COPD was not observed to modify associations between long-term NO2 exposure and
diabetes (Eze etal.. 2014; Andersen et al.. 2012c). but some epidemiologic studies show
larger associations of short-term or long-term NO2 exposure with various cardiovascular
effects among adults with COPD (Table 7-6). Inference is limited from studies of
cardiovascular effects due to lack of comparison to healthy groups in the short-term
exposure studies (Suh and Zanobetti. 201 Ob: Peel et al.. 2007) and uncertainty regarding
an independent relationship with cardiovascular effects for both short-term and long-term
NO2 exposure (Sections 5.3.11 and 6.3.9).
In conclusion, some but not all epidemiologic evidence points to larger NO2-related
cardiovascular effects in adults with COPD, and there is uncertainty as to whether the
findings reflect an independent effect of NO2. NO2 exposure and internal dose differences
for people with COPD are unknown. Controlled human exposure studies do not clearly
demonstrate that NO2 exposure induces respiratory effects in adults with COPD, and the
limited findings for larger lung function decrements in adults with COPD relative to
healthy adults may be influenced by differences between groups in age or smoking. The
limited and inconsistent evidence from controlled human exposure studies for
7-10
-------
NO2-related changes in lung function is inadequate to determine whether people with
COPD are at increased risk for NCh-related health effects.
Table 7-5 Controlled human exposure studies evaluating pre-existing chronic
obstructive pulmonary disease.
Factor
Evaluated
Reference
Category
Direction of
Effect
Modification or
Effect3 Outcome
Study
Population Study Details
Study
Nonasthma outcomes and short-term exposure
COPD
n=20
COPD
n = 7
COPD
n = 18
COPD
Healthy
n = 20
Healthy
n = 7
Healthy
n = 6
None
-j- Lung function
' decrement
_ Symptoms,
respiratory
conductance
-j- Lung function
' decrement
_ Sputum cell
counts,
symptoms
_ Lung function
decrement,
heart rate,
blood
pressure,
symptoms
_ Lung function
decrement,
heart rate,
symptoms
n=40
Mean age
'59.9yr
n = 14
Mean age
" COPD: 58 yr
Healthy:
34 yr
n=24
Mean age
COPD: 72 yr
Healthy:
68 yr
n=22
Mean age
60. Syr
300 ppb NO2for4 h;
Three 7 min periods of
exercise at VE — 4 times
resting
300 ppb NO2 for 1 h;
Exercise at VE = 25 L/min
(1)400ppbNO2for2h
(2) 200 ug/m3 CAPs for 2 h
(3) 400 ppb NO2 and
200 ug/m3 CAPs for 2 h
(1-3) Exercise 15 min
on/15 min off at
VE = ~2 times resting
500, 1,000, or 2,000 NO2
ppb for 1 h;
Exercise 15 min on/15 min
offVE= 16 L/min
Morrow et
al. (1992)
Vaqaqqini
etal. (1996)
Gonq etal.
(2005)
Linn et al.
(1985a)
CAPs = concentrated ambient particles; COPD = chronic obstructive pulmonary disease; NO2 = nitrogen dioxide; VE = minute
ventilation.
aUp facing arrow indicates that the effect of NO2 is greater (e.g., larger lung function decrement, larger increase in symptoms) in
the group with the factor evaluated than in the reference group. Down facing arrow indicates that the effect of NO2 is smaller in the
group with the factor evaluated than in the reference group. A dash indicates no difference in NO2-related health effect between
groups. In some studies, only a population with pre-existing disease was examined; therefore, the arrow or dash represents the
direction of the effect in that population after exposure to NO2 relative to exposure to filtered air.
7-11
-------
Table 7-6 Epidemiologic studies evaluating pre-existing chronic obstructive
pulmonary disease.
Factor
Evaluated
Nonasthma
COPD
n = 18
COPD
n = 8% ED
visits
Nonasthma
COPD
n = 121
COPD
n=6
COPD
n= 2,058
COPD
n = 1,268
Reference
Category
outcomes and
Recent Ml
n = 12
No COPD
n = 92% ED
visits
outcomes and
No COPD
n = 1,863
No COPD
n = 136
No COPD
n= 49,760
No COPD
n = 5,124
Direction of
Effect
Modification3
Outcome
Study
Population
Study Details
Study
short-term exposure
t
-
t
HRV pNNSO
decrement
HRV r-MSSD
decrement
Cardiovascular-
related ED
visits
n = 30
Ages NR
n = 103,551
ED visits for
CVD; 31
participating
hospitals
Atlanta, GA,
1999-2000
Atlanta, GA,
1993-2000
Suh and
Zanobetti
(2010b)t
Peel etal.
(2007)
long-term exposure
—
t
-
Incident stroke
Fatal stroke
Diabetes
Diabetes
n = 1,984
Ages 50-65 yr
at baseline
n = 142
Ages 50-65 yr
at baseline
n = 51,818
Ages 50-65 yr
at baseline
n = 6,392
Ages 29-73 yr
Copenhagen,
Aarhus counties,
Denmark,
"1993-2006
Copenhagen,
Aarhus counties,
Denmark,
1993-2006
Switzerland, 2002
Andersen et
al. (2012b)t
Andersen et
al. (2012c)t
Eze et al.
(2014)t
COPD = chronic obstructive pulmonary disease; CVD = cardiovascular disease; HRV = heart rate variability; Ml = myocardial
infarction; NR = not reported.
aUp facing arrow indicates that the effect of nitrogen dioxide (NO2) is greater (e.g., larger decrement in HRV, larger increase in ED
visits) in the group with the factor evaluated than in the reference group. Down facing arrow indicates that the effect of NO2 is
smaller in the group with the factor evaluated than in the reference group. A dash indicates no difference in NO2-related health
effect between groups.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
7.3.3
Cardiovascular Disease
Cardiovascular disease is the primary cause of death in the U.S., and approximately 11%
of adults report a diagnosis of heart disease. In addition, hypertension has been diagnosed
in roughly 24% of the adult U.S. population (Blackwell et al.. 2014). For both short-term
(Section 5.3.11) and long-term (Section 6.3.9) NCh exposure, the collective body of
evidence is suggestive, but not sufficient, to infer a causal relationship with
cardiovascular effects. These conclusions are primarily based on the uncertainty in
7-12
-------
distinguishing an independent relationship for NCh with cardiovascular effects. In
addition to this uncertainty, the evidence base does not consistently show that
pre-existing CVD increases the risk for health effects related to short-term or long-term
NC>2 exposure (Table 7-7). NCh exposure or dose differences for people with CVD are
not characterized. Suh and Zanobetti (201 Ob) found that total personal NO2 exposure was
similar for adults with prior MI and adults with COPD but made no comparison to
healthy adults.
For short-term exposure, NC>2-related mortality was higher among individuals with CVD
(Chiusolo et al.. 2011): however, the majority of evidence, which is for cardiovascular
hospital admissions and ED visits, is inconsistent (Table 7-7). The strongest evidence for
a relationship between short-term NO2 exposure and cardiovascular effects is for MI
(Section 5.4.8). and most studies show no difference in the association in groups with
hypertension (Tsai et al.. 2012; Peel et al.. 2007; D'Ippoliti et al.. 2003). arrhythmia (Tsai
et al.. 2012; Mann et al.. 2002). or congestive heart failure (Tsai et al.. 2012; D'Ippoliti et
al.. 2003). Many studies of long-term NC>2 exposure compared groups with and without
hypertension and found no difference in the association with diabetes (Eze et al.. 2014;
Andersen et al.. 2012c) and no consistent difference in the association with stroke
(Andersen et al.. 2012b). For blood pressure examined as an outcome, associations with
long-term NC>2 exposure were larger in groups with pre-existing CVD (Foraster et al..
2014).
Among studies examining subclinical cardiovascular effects such as changes in HRV,
interleukin (IL)-6, or arrhythmic events recorded on electrocardiograms, most did not
observe that associations with short- or long-term NC>2 exposure differed between groups
with or without pre-existing CVD, whether defined as any CVD, ischemic heart disease
(IHD), or hypertension (Panasevich et al.. 2009; Felber Dietrich et al.. 2008; Ljungman et
al.. 2008). Risk factors for CVD, including higher systemic inflammation and
hypercholesterolemia, do not consistently modify NC>2-related cardiovascular effects
(Andersen et al.. 2012b; Huang et al.. 2012a). Experimental studies (Table 7-8) also do
not clearly support NC>2-induced subclinical cardiovascular effects in adults with CVD
(Scaife et al.. 2012) or a mouse model of CVD (Campen et al.. 2010).
For associations with short-term and long-term NC>2 exposure, people with and without
pre-existing CVD have been compared with respect to an array of cardiovascular
diseases, events, and subclinical effects. Studies are also diverse in the conditions by
which they define pre-existing CVD. No consistent difference in NC>2-related
cardiovascular effects is demonstrated between groups with and without pre-existing
CVD, and it is unclear whether people with CVD differ in NC>2 exposure or internal dose.
Additionally, there is limited experimental evidence supporting the biological plausibility
7-13
-------
of NO2-related health effects in response to pre-existing cardiovascular conditions. In
conclusion, the large evidence base lacks sufficient consistency in demonstrating that
pre-existing CVD modifies NC>2-related cardiovascular effects, and an independent effect
of NC>2 is uncertain overall. Therefore, the evidence is inadequate to determine whether
people with CVD are at increased risk for NCh-related health effects.
Table 7-7 Epidemiologic studies evaluating pre-existing cardiovascular
disease.
Factor
Evaluated
Reference
Category
Direction of
Effect
Modification3
Outcome
Study
Population Study Details Study
Nonasthma outcomes and short-term exposure
Hypertension
n = 30% visits
No
hypertension
n - 70% visits
t
-
Arrhythmia ED
visits
ED visits for IHD
orCHF
n = 103,551 Atlanta, GA, Peel et al.
ED visits for 1993-2000 (2007)
CVD; 31
participating
hospitals
Hypertension No
n = 40% visits hypertension
n = 60% visits
Hospital n = 27,563
admission for Ml hospital
admissions
Taipei, Taiwan Tsai et al.
1999-2009 (2012)t
CHF
n = 15% visits
Cardiac
arrhythmia
n = 11% visits
Hypertension
n = 1,648
Heart
conduction
disorder
n=414
Cardiac
dysrhythmia
n = 1,296
Heart failure
n = 703
No CHF
n = 85% visits
No cardiac _
arrhythmia
n = 89% visits
No _ First hospital n = 6,531 Rome, Italy, D'lppoliti et al.
hypertension admission for hospital 1995-1997 (2003)
n = 4,883 acute Ml records
No heart +
conduction '
disorder
n = 6,117
No cardiac _
dysrhythmia
n = 5,235
No heart _
failure
n = 5,828
7-14
-------
Table 7-7 (Continued): Epidemiologic studies evaluating pre-existing
cardiovascular disease.
Factor
Evaluated
Secondary
diagnosis of
arrhythmia
n = 34.5%
admissions
Secondary
diagnosis of
CHF
n = 14.1%
admissions
Pre-existing
heart disease
n = 525 with
stroke
IHD
n = 56
Pre-existing
CVD
n = 1.2-1 4%
Direction of
Reference Effect
Category Modification3
No secondary
diagnosis of
arrhythmia
n = 65.5%
admissions
No secondary +
diagnosis of '
CHF
n = 85.9%
admissions
No pre- *
existing heart '
disease
n - 2,214 with
stroke ~
No IHD
n = 32
No pre- *
existing CVD '
n = 86-98.8%
Outcome
Hospital
admission for
IHD
Hospital
admission for
ischemic stroke
Hospital
admission for
hemorrhagic
stroke
Ventricular
tachy-
arrhythmia
Total mortality
Study
Population
n = 54,863
hospital
admissions
n = 5,927
hospital
admissions
n = 88 with
implantable
cardioverter
defibrillators
Ages 28-85 yr
n = 276,205
natural deaths
Study Details
southern
California,
1988-1995
Edmonton,
Canada,
2003-2009
Gothenburg,
Stockholm,
Sweden,
2001-2006
10 cities, Italy
12% of
population
2001-2005
Study
Mann et al.
(2002)
Villeneuve et
al. (2012)t
Liunqman et
al. (2008)t
Chiusolo et al.
(201 1)t
Nonasthma outcomes and long-term exposure
Hypertension
n = 575
Hypertension
n = 38
Hypercholes-
terolemia
n=230
Hypercholes-
terolemia
n = 19
Pre-existing
CVD
n=269
No
hypertension
n = 1,409
No
hypertension
n = 104
No
hypercholes-
terolemia
n = 1,754
No
hypercholes-
terolemia
n = 123
No +
pre-existing '
CVD
n = 3,431
Incident stroke
Fatal stroke
Incident stroke
(confirmed by
hospital
admission)
Fatal stroke
(confirmed by
hospital
admission)
Systolic/diastolic
blood pressure
n = 1,984
Ages 50-65 yr
at baseline
n = 142
Ages 50-65 yr
at baseline
n = 1,984
Ages 50-65 yr
at baseline
n = 142
Ages 50-65 yr
at baseline
n = 3,700
Ages 35-83 yr
Copenhagen,
Aarhus
counties,
Denmark,
1993-2006
Copenhagen,
Aarhus
counties,
Denmark,
'1993-2006
Girona, Spain
Andersen et
al. (2012b)t
Andersen et
al. (2012b)t
Foraster et al.
(2014)t
7-15
-------
Table 7-7 (Continued): Epidemiologic studies evaluating pre-existing
cardiovascular disease.
Factor
Evaluated
Hypertension
n = 867
Hypertension
n = 19.4%
Reference
Category
No
hypertension
n = 669
No hyper-
tension
n = 80.6%
Direction of
Effect
Modification3 Outcome
Blood IL-6
levels
_ Diabetes
Study
Population
n = 1,536
Ages 45-70 yr
n = 6,392
Ages 29-73 yr
Study Details
Stockholm
county,
Sweden,
1992-1994
Switzerland,
2002
Study
Panasevich et
al. (2009)t
Eze et al.
(2014)t
CHF = congestive heart failure; CVD = cardiovascular disease; ED = emergency department; IHD = ischemic heart disease;
Ml = myocardial infarction.
aUp facing arrow indicates that the effect of nitrogen dioxide (NO2) is greater (e.g., larger increase in hospital admission, larger risk
of mortality) in the group with the factor evaluated than in the reference group. Down facing arrow indicates that the effect of NO2
is smaller in the group with the factor evaluated than in the reference group. A dash indicates no difference in NO2-related health
effect between groups.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
Table 7-8 Controlled human exposure and toxicological studies informing risk
due to pre-existing cardiovascular disease.
Factor
Evaluated
Nonasthma
Stable CHD
or impaired
left
ventricular
systolic
function
Athero-
sclerosis
Reference
Category
outcomes and
None
None
Direction of
Effect3 Outcome
short-term exposure
_ Heart rate,
HRV
decrement
* Oxidative
' stress in
heart tissue
Study
Population/
Animal Model
n = 18 humans
Mean age 68 yr
n = 5-10
mice/group,
ApoE-'-
Study Details
400 ppb NO2 for 1 h
200 or 2, 000 ppb
NO2, 6 h/day,
7 days
High fat diet
Study
Scaife et al.
(2012)t
Campen et al.
(201 0)t
CHD = coronary heart disease; HRV = heart rate variability; NO2 = nitrogen dioxide.
aThese studies only examined subjects with cardiovascular disease and have no reference group. A dash indicates that NO2 was
not observed to induce an effect in the group with cardiovascular disease evaluated relative to clean air exposure. An up-facing
arrow indicates that NO2 induced an effect on the outcome (e.g., cause a decrement in HRV) in the group with cardiovascular
disease.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
7.3.4
Diabetes
Diabetes mellitus is a group of diseases characterized by high blood glucose levels that
result from defects in the body's ability to produce and/or use insulin. High blood glucose
levels can damage blood vessels, increasing the risk of people with diabetes for heart
7-16
-------
disease or stroke. Diabetes and cardiovascular disease are also linked by common risk
factors such as hypertension and obesity. These relationships support the potential for
diabetes to influence the risk of cardiovascular disease. Diabetes has been examined as a
modifier of NCh-related health effects in adults, and the large number of affected adults
[21 million or 8.6% with diagnosis in the U.S. in 2012 (Blackwell etal.. 2014)1
underscores the potential public health impact of elevated risk from diabetes. However,
diabetes has not consistently been observed to modify epidemiologic associations of NO2
exposure with cardiovascular effects, and no consistent pattern is observed for short-term
or long-term exposure or for any particular outcome (Table 7-9). No difference by
diabetes was observed in studies of short-term NO2 exposure with hospital admissions or
ED visits for IHD or MI (Tsai etal.. 2012; Filho et al.. 2008) or of long-term NO2
exposure with stroke (Andersen et al.. 2012b). Diabetes also did not clearly modify
associations of short-term or long-term NCh exposure with the subclinical effects of heart
rate variability, ventricular tachyarrhythmia, blood pressure, and blood IL-6 levels
(Foraster et al., 2014; Huang etal.. 2012a; Panasevich et al., 2009; Ljungman et al..
2008). Associations of short-term and long-term NC>2 exposure with total mortality also
did not consistently differ between people with and without diabetes (Faustini et al..
2013; Chiusolo et al.. 2011; Maheswaran et al.. 2010). It is unknown whether people with
diabetes differ with respect to NO2 exposure or dosimetry. In conclusion, there is
inconsistent evidence that diabetes modifies NCh-related cardiovascular effects. This
finding along with uncertainty in an independent effect of NCh on cardiovascular
outcomes (Sections 5.3.11 and 6.3.9) makes the evidence inadequate to determine
whether people with diabetes are at increased risk for NCh-related health effects.
7-17
-------
Table 7-9
Factor
Evaluated
Nonasthma
Diabetes
n = 11.3%
admission
Diabetes
29.6%
admission
Diabetes
n = 700 ED
visits
Diabetes
n = 9
Diabetes
n = 12
Diabetes
n = 30,260
Nonasthma
Diabetes
n = 97
Diabetes
n = 9
Diabetes
n = 580
Diabetes
n = 121
Epidemiologic studies evaluating pre-existing diabetes.
Reference
Category
outcomes and
No diabetes
n = 88.7%
admission
No diabetes
70.4%
admission
No diabetes
n= 44,300
ED visits
No diabetes
n = 31
No diabetes
n = 76
No diabetes
n = 245,945
outcomes and
No diabetes
n = 1,887
No diabetes
n = 133
No diabetes
n = 3,120
No diabetes
n = 1,415
Direction of
Effect
Modification3 Outcome
short-term exposure
_ Respiratory
hospital
admission
_ Hospital
admission for Ml
* ED visits for
' hypertension
and cardiac
ischemic disease
i HRV decrement
_ Ventricular
tachy-arrhythmia
* Total mortality
long-term exposure
_ Incident stroke
_ Fatal stoke
_ Systolic blood
pressure
i Diastolic blood
^ pressure
_ Relative IL-6
level
Study
Population
n = 100,690
hospital
admissions
n = 27,563
hospital
admissions
n = 45,000
ED visits
n = 40 with
CVD
Mean age
66 yr
n = 88 with
implantable
cardioverter
defibrillators
Ages 28-85 yr
n = 276,205
natural deaths
n = 1,984
Ages 50-65 yr
at baseline
n = 142
Ages 50-65 yr
at baseline
n = 3,700
Ages 35-83 yr
n = 1,536
Ages 45-70 yr
Study Details
6 cities, Italy,
2001-2005
Taipei, Taiwan
1999-2009
Sao Paulo
Hospital,
January
2001 -July 2003
Beijing, China,
summer 2007
and summer
2008
Gothenburg,
Stockholm,
Sweden,
2001-2006
6 cities, Italy,
2001-2005
Copenhagen,
Aarhus counties,
Denmark,
-1993-2006
Girona, Spain
Stockholm
county, Sweden
1992-1994
Study
Faustini et al.
(201 3)t
Tsai et al.
(2012)t
Filhoetal.
(2008)t
Huanq et al.
(2012a)t
Ljunqman et al.
(2008Jf
Chiusolo et al.
(201 1)t
Andersen et al.
(2012b)t
Foraster et al.
(2014)t
Panasevich et
al. (2009)t
7-18
-------
Table 7-9 (Continued): Epidemiologic studies evaluating pre-existing diabetes.
Factor
Evaluated
Diabetes
n = 315
Diabetes
n = 1,045
Reference
Category
No diabetes
n = 1,541
No diabetes
n = 12,399
Direction of
Effect
Modification3
1
-
Outcome
Total mortality
Lung cancer
mortality
Study
Population
n = 3,320
Mean age
70 yr
n = 13,444
Ages >65 yr
Study Details
London, England
Follow-up:
1995-2005
NO2 assessed
for 2002
Shizuoka, Japan
1999-2006
Study
Maheswaran et
al. (201 0)t
Yorifuii et al.
(201 0)t
CVD = cardiovascular disease; ED = emergency department; HRV = heart rate variability; IL = interleukin; Ml = myocardial
infarction.
aUp facing arrow indicates that the effect of nitrogen dioxide (NO2) is greater (e.g., larger risk of ED visit, larger decrement in HRV)
in the group with the factor evaluated than in the reference group. Down facing arrow indicates that the effect of NO2 is smaller in
the group with the factor evaluated than in the reference group. A dash indicates no difference in effects between groups.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
7.3.5 Obesity
Obesity can be defined as a BMI of 30 kg/m2 or greater. It is a public health issue of
increasing importance as obesity rates have continually increased over several decades in
the U.S. among adults and children (Blackwell et al., 2014). Obesity or high BMI has
been examined as a modifier of NO2-related health effects only in adults. Among U.S.
adults, prevalence of both obesity and being overweight (BMI of 25-30 kg/m2) are high
(28 and 34.5%, respectively). Being obese or overweight could increase the risk of
NO2-related health effects through multiple mechanisms including persistent, low-grade
inflammation. Poor diet and chronic diseases often occur with obesity and could be part
of the pathway by which obesity increases the risk of NO2-related health effects or could
act in combination with obesity to increase risk.
The 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) did not evaluate obesity as a
potential factor that could increase the risk of NO2-related health effects. Recent studies
have examined obesity or BMI as a potential effect measure modifier, but overall the
evidence is inconsistent as to whether obesity leads to increased risk of NO2-related
health effects. And while activity and breathing patterns differ in obese individuals, it is
not known whether these factors contribute to differences in NO2 exposure or dosimetry.
Most studies examined whether obesity increases the risk of NO2-related cardiovascular
and metabolic effects, which is expected given the evidence that obesity is a risk factor
for both cardiovascular disease and diabetes. A study in rats provides evidence that
long-term NO2 exposure has larger effects on dyslipidemia, a known risk factor for
7-19
-------
cardiovascular disease, in obese rats compared to nonobese rats [(Takano et al.. 2004);
Table 7-10]. However, differences between obese and nonobese strains were limited to
160 ppb NO2 and not observed at higher NO2 exposures.
The epidemiologic evidence is not coherent with the results in rats. There is some
indication of larger NCh-associated cardiovascular or diabetes-related mortality in obese
groups (Raaschou-Nielsen et al.. 2012). However, most studies did not provide evidence
that associations between long-term NCh exposure and cardiovascular disease or diabetes
differed between obese and nonobese groups [(Eze et al.. 2014; Atkinson et al.. 2013;
Hart etal.. 2013; Mobasheretal.. 2013; Andersen etal.. 2012c; Andersen etal.. 2012b);
Table 7-111. Most studies used a similar definition of obese: BMI > 30 kg/m2. High BMI
also did not tend to modify associations of short-term and long-term NCh exposure with
subclinical cardiovascular effects (Dadvand et al.. 2014b; Huang etal.. 2012a; Baja et al..
2010; Ljungman et al.. 2008). The limited number of studies that examined NCh
associations with mortality from respiratory causes or lung cancer also did not provide
any evidence of increased risk for obese individuals (Dimakopoulou et al.. 2014; Yorifuji
etal.. 2010).
In conclusion, obesity increased NC>2-related cardiovascular effects in a single study in
rats, but a larger body of epidemiologic evidence largely shows no difference between
obese and nonobese adults. Obesity also does not tend to modify associations of
long-term NC>2 exposure with mortality or diabetes. In addition to the limited evidence
indicating that obese people may be at increased risk of NCh-related health effects,
uncertainty remains in the overall body of evidence regarding the independent effects of
NO2 on cardiovascular effects, diabetes (Section 6.3.9). and mortality (Section 6.5.3).
Further, information on potential NCh exposure or dosimetry differences by obesity is
lacking. Therefore, the evidence is inadequate to determine whether obese individuals are
at increased risk for NCh-related health effects.
Table 7-10 lexicological study evaluating pre-existing obesity.
Direction of
Factor Reference Effect
Evaluated Category Modification3 Outcome Animal Model Study Details Study
Nonasthma outcomes and long-term exposure
Obesity No obesity * Triglycerides, Rats (OLETF and 160, 800, or Takano et
n = 9-l3 n = 10-14 ' HDL, total LETO diabetes 4,000 ppb NO2; al. (2004)
cholesterol, blood models) n = 10-14 continuously for
sugar males/group 32 weeks
HDL = high density lipoprotein; NO2 = nitrogen dioxide.
aUp facing arrow indicates that the effect of NO2 is greater (e.g., larger increase in triglycerides) in the group with the factor
evaluated than in the reference group.
7-20
-------
Table 7-11 Epidemiologic studies evaluating pre-existing obesity.
Factor
Evaluated
Nonasthma
HighBMI
(>27)
n =29 or
30
HighBMI
(>25)
n = 16
HighBMI
(>30)
n = 27.6%
Nonasthma
HighBMI
(>30)b
HighBMI
(>30)
n = 109,104
HighBMI
(>30)
n - 84
HighBMI
(>30)
n = 366
HighBMI
(>30)
n = 19
Reference
Category
outcomes and
LowBMI
(<27)
n = 69
LowBMI
(<25)
n = 24
LowBMI
(<30)
n = 72.4%
outcomes and
LowBMI
(<30)b
LowBMI
(25-30)
n = 243,556
LowBMI
(<30)
n - 158
LowBMI
(<30)
n = 1,618
LowBMI
(<30)
n = 123
Direction of
Effect
Modification3 Outcome
short-term exposure
_ Ventricular
arrhythmia
_ HRV decrement
* Change in
' ventricular
repolarization
long-term exposure
_ Incidence Ml
_ Heart failure
_ C-reactive
protein
TNF-a
IL-6
| IL-8
_ Fibrinogen
_ Hepatocyte
growth factor
Incident stroke
_ Fatal stroke
Study
Population
n = 98 with
implantable
cardioverter
defibrillators
Ages 28-85 yr
n = 40
nonsmoking
adults with CVD
Mean age 66 yr
n = 580 males
Mean age 75 yr
n = 84,562
Ages 30-55 yr at
enrollment
n = 836,557
Ages 40-89 yr at
baseline
n = 242 adults
with clinically
stable COPD
Mean age 68 yr
n = 1,984
Ages 50-65 yr at
baseline
n = 142
Ages 50-65 yr at
baseline
Study Details
Gothenburg,
Stockholm,
Sweden,
2001-2006
Beijing, China,
summer 2007
and summer
2008
Boston, MA
area, Follow-up:
2000-2008
U.S.,
1990-2008
England,
2003-2007
Barcelona,
Spain,
2004-2006
Copenhagen,
Aarhus counties,
Denmark,
'1993-2006
Study
Liunqman et
al. (2008)t
Huanq et al.
(2012ajf
Baia et al.
(201 0)t
Hartetal.
(201 3)t
Atkinson et al.
(201 3)t
Dadvand et al.
(2014b)t
Andersen et
al. (2012b)t
7-21
-------
Table 7-11 (Continued): Epidemiologic studies evaluating pre-existing obesity.
Factor
Evaluated
High/very
highBMI
(225)
n= 28,937
High waist-
to-hip ratio
(>0.90 M,
>0.85 F)
n= 26,183
HighBMIb
(>30)
HighBMI
(>30)
n=68
HighBMIb
(>30)
HighBMI
(>25)
n = 1,950
HighBMI
(>30)
n = 96,076
person-yr
HighBMI
(>21.8)b
Reference
Category
LowBMI
(<25)
n = 22,881
Low waist-
to-hip ratio
(<0.90 M,
<0.85 F)
n = 25,635
LowBMIb
(<30)
LowBMI
(<30)
n = 213
LowBMIb
(<30)
LowBMI
(<18.5)
n = 1,010
LowBMI
(<25)
n = 298,503
person-yr
LowBMI
(<21.8)b
Direction of
Effect
Modification3 Outcome
_ Diabetes
t
_ Diabetes
_ Hypertensive
disorders of
pregnancy
i Respiratory
^ mortality
_ CVD mortality
* Diabetes-related
' mortality
_ Lung cancer
mortality
Study
Population
n = 51,818
Ages 50-65 yr at
baseline
n = 6,392
Ages 29-73 yr
n = 298
predominantly
Hispanic women
n = 307,553
Mean age
across
16 cohorts
41. 9-73.0 yr at
baseline
n = 9,941, 256
deaths
Ages 35-103 yr
n = 52,061
Ages 50-64 yr
n = 13,444
Ages >65 yr
Study Details
Copenhagen,
Aarhus counties,
Denmark,
1993-2006
Switzerland,
2002
Los Angeles,
CA,
1996-2008
Europe
Follow-up:
1985-2007
NO2 exposure
assessed for
2008-2011
Shenyang,
China
Follow-up:
1998-2009
NO2 exposure
assessed for
1998-2009
Denmark
Follow-up:
1971-2009
NO2 exposure
assessed for
1971-2009
Shizuoka,
Japan,
1999-2006
Study
Andersen et
al. (2012c)t
Eze et al.
(2014)t
Mobasher et
al. (2013)t
Dimakopoulou
etal. (2014)t
Zhanq et al.
(201 1)t
Raaschou-
Nielsen et al.
(2012)t
Yorifuji et al.
(201 0)t
BMI = body mass index; CA = California; CVD = cardiovascular disease; COPD = chronic obstructive pulmonary disease; F =
female; HRV = heart rate variability; M = male; Ml = myocardial infarction; NO2 = nitrogen dioxide; TNF = tumor necrosis factor;
IL = interleukin.
aUp facing arrow indicates that the effect of NO2 is greater (e.g., larger change in ventricular repolarization) in the group with the
factor evaluated than in the reference group. Down facing arrow indicates that the effect of NO2 is smaller in the group with the
evaluated factor than in the reference group. A dash indicates no difference in NO2-related health effect between groups.
bSample size not reported.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
7-22
-------
7.4 Genetic Factors
Genetic variation in the human population is known to contribute to numerous diseases
and differential physiologic responses. The 2008 ISA for Oxides of Nitrogen (U.S. EPA.
2008c) concluded that "it remains plausible that there are genetic factors that can
influence health responses to NO2, though the few available studies did not provide
specific support." Since then, many studies have examined whether specific genetic
polymorphisms increase the risk of NCh-related health effects. A strength of these studies
is that they use a targeted approach, focusing on specific genes that potentially are
involved in signaling pathways mediating biological responses to NCh. For example, NC>2
exposure can lead to the formation of oxidation products (Section 4.3.2.1) and also
modulate immune function (Section 4.3.2.6). and studies examined variants for genes
encoding antioxidant enzymes [e.g., glutathione S-transferases (GSTs)] and mediators of
immune response [e.g., tumor necrosis factor-alpha (TNF-a)]. A potential limitation for
drawing conclusions is the large number of genetic variants examined within studies,
which increases the probability of finding associations by chance alone. Thus,
consistency in findings across genetic variants is considered. Further, the functional or
biological consequence of some of the gene variants is unknown, and some variants may
be surrogates for another linked gene or a group of related genes. Thus, where available,
the variant effect is described (Table 7-12) and considered in conclusions.
Several studies examined gene variants for modification of NCh-asthma relationships,
and oxidative stress is described as a key process underlying asthma exacerbation and
development attributed to NCh exposure (Section 4.3.2.1). However, studies that
examined functional variants of GST Mu 1 or GST Pi 1 estimated similar (Castro-Giner
et al.. 2009) or lower (Romieu et al.. 2006) effects of short-term or long-term NC>2
exposure on asthma symptoms or asthma prevalence in groups with variants encoding
enzymes with null or reduced oxidative metabolizing activity (Table 7-12). These
variants are common in the population, and NCh-related health effects were compared
between groups with fairly similar numbers of people (Table 7-12). However, it is not
clear that either study planned comparisons a priori. Another limitation of the study that
observed no difference is the uncertain temporal sequence of current exposure and
current asthma status (Castro-Giner et al.. 2009).
The beta-2-adrenergic receptor (ADRB2) is an encoded G protein-coupled receptor that
plays an important role in the regulation of airway smooth muscle tone and is the
pharmacological target of beta-agonist asthma medications (Hizawa. 2011). NCh
exposure has been shown to increase airway responsiveness in adults with asthma
7-23
-------
(Section 5.2.2.1). providing a plausible role for variants in ADRB2 in modifying the risk
of NC>2-associated asthma. Higher methylation of the ADRB2 promoter, which is
associated with reduced expression of the receptor, was observed to increase the risk of
asthma severity in children associated with indoor residential NO2 exposure (Tu et al..
2012). There is mixed evidence for beta-agonist medication use in modifying
NO2-associated respiratory effects (Section 5.2.2.2). and it is not known whether the
response to beta-agonists is influenced by genetic variants in ADRB2.
Similar to variants with known functional differences, no clear evidence exists that
genetic variants with unknown functional differences increase NCh-related risk of
asthma. As examined in a single study that estimated long-term NO2 exposure at
subjects' homes, a variant in the antioxidant enzyme NADPH-quinone oxidoreductase
(NQO1) increased associations with asthma prevalence in adults. However, NQO1 was
one among many variants examined. No modification was observed by variants in
ADRB2 or the immune response genes TNF-a and toll-like receptor (TLR)4 [(Castro-
Giner et al.. 2009): Table 7-12]. which are known to have a role in oxidant-induced
inflammation and asthma pathogenesis.
Inflammation and oxidative stress are also linked to cardiovascular and metabolic effects
(Section 4.3.2.9). However, gene variants associated with increased inflammatory
mediators in the blood did not modify associations of long-term NC>2 exposure with
myocardial infarction (Panasevich et al.. 2013). Genetic variants with the potential for
elevated oxidative stress did increase NCh-related subclinical cardiovascular and
metabolic effects [(Kim and Hong. 2012; Baia et al.. 2010): Table 7-121. A strength of
(Bajaetal. (2010)) is that, rather than performing multiple comparisons of individual
variants, they analyzed the sum of gene variants with increased oxidative stress potential
as a cumulative index of oxidative stress potential. Despite some positive findings,
independent relationships between short-term NC>2 exposure and cardiovascular and
metabolic effects are uncertain. Thus, it is not clear the extent to which the findings for
modification by gene variants can be attributed to NO2 specifically. Genetic variants in
oxidative metabolism enzymes with unknown functional differences did not clearly
modify associations of long-term NCh exposure with decrements in lung development in
children. Associations were larger for some variants in glutathione metabolism pathway
genes, such as glutathione synthetase [GSS; (Breton et al.. 2011)1. However, the results
were not consistent across the multiple gene variants of glutathione examined
[glutathione reductase (GSR), glutamate-cysteine ligase, modifier subunit (GCLM),
glutamate-cysteine ligase, catalytic subunit (GCLC)] or NQO1 (rs!0517).
There is evidence for independent relationships of short-term and long-term NC>2
exposure, respectively, with exacerbation (Section 5.2.9) and development
7-24
-------
(Section 6.2.9) of asthma, and antioxidant modulation, immune-mediated inflammation,
and airway responsiveness are described as key events in the underlying modes of action
(Section 4.3.5). Evidence in rodents and humans that dietary antioxidants modify
NC>2-induced pulmonary oxidative stress (Section 7.6.1) would suggest a role for variants
in oxidative metabolism genes in modifying the effects of NC>2 on asthma exacerbation or
development. However, gene variants with greater potential for oxidative stress were not
observed to modify associations of short-term NC>2 exposure with asthma-related effects.
Variants in antioxidant and immune-related genes did modify some associations of
long-term NC>2 exposure with asthma, but results are inconsistent for any particular gene
variant or outcome and are based on multiple comparisons and post hoc analyses. While
gene variants for antioxidant enzymes and inflammatory cytokines modified
cardiovascular and metabolic effects, the findings were limited to subclinical outcomes
and did not include myocardial infarction. Overall, the findings for effect measure
modification by genetic variants are inconsistent for asthma-related effects, and the
interpretation of the results for cardiovascular and metabolic effects is complicated by the
uncertainty as to whether the results can be attributed specifically to NO2 exposure. Also
unknown is whether the gene variants alter oxidant species or inflammatory mediators in
response to NCh exposure. Additionally, because it currently is uncertain whether a
particular gene variant examined in isolation can clearly represent an at-risk population,
evidence for effect measure modification may provide insight only on biological
pathways mediating a health effect. For all of these reasons, the evidence is inadequate to
determine whether genetic variants, particularly for antioxidant enzymes and immune
responses, increase the risk for NCh-related health effects.
7-25
-------
Table 7-12 Epidemiologic studies evaluating genetic factors.
Factor
Evaluated/Gene
Function
Reference
Category
Direction of
Effect
Modification3 Outcome
Study Population
and other Details
Study
Asthma-related outcomes and functional gene variants
GSTM1 null,
n = 58
Null oxidant
metabolizing
capacity
GSTP1 VaMOSVal,
n = 54
(rsID 947894)
Reduced oxidant
metabolizing
capacity
GSTM1 null
n = 49%
Null oxidant
metabolizing
capacity
GSTP1 VaMOSVal
(rsID 1695)
n = 32%
Reduced oxidant
metabolizing
capacity
TNF-a 308 GA/AA
(rs1 800629)
n = 16%
Increased
expression
ADRB2b
Intermediate or
high methylation
levels
Reduced
expression
GSTM1 positive,
n = 93
Ile105lle,
n = 97
GSTM1 positive,
n = 51%
GSTP1 Ile105lle
orlle105Val
n = 68%
TNF-a 308 GG
n = 84%
ADRB2b
Low methylation
levels
i Asthma Mexico City, Mexico Romieu et al.
^ symptoms and n = 151 children with (2006)
medication use asthma
Ages NR
Short-term exposure
1
_ Asthma Umea and Uppsala, Castro-Giner et
prevalence Sweden: Ipswich al. (2009)t
and Norwich, U.K.;
Albacete, Barcelona,
Huelva, Galdakao,
and Oviedo Spain'
_ Erfurt, Germany;
Paris, Grenoble,
France; Antwerp,
Belgium
n = 2,920
Mean age 43 yr
Long-term exposure
* Asthma severity CT and Springfield, Fu et al. (2012)t
' Worcester, MA
n = 182
Ages 5-12 yr,
followed for 1 yr
Long-term exposure
7-26
-------
Table 7-12 (Continued): Epidemiologic studies evaluating genetic factors.
Factor
Evaluated/Gene
Function
Direction of
Reference Effect
Category Modification3 Outcome
Study Population
and other Details Study
Asthma-related outcomes and gene variants with unknown functional difference
NQO1 CC
(rs2917666)
n = 32%
TLR4 GG
(rs1 1536889)
n = 14%
ADRB2
rs 104271 3 G/G
rs 104271 4 C/C
rs 104271 8 C/C
rs 10427 19 G/G
n = 18-40%
NQO1 GC or
GG,
n = 68%
TLR4 GC or CC
n = 86%
G/A or A/A
C/G or G/G
C/A or A/A
G/C or C/C
n = 60-82%
-j- Asthma Multiple European Castro-Giner et
' prevalence countries (see al. (2009)t
above)
n - 2,920
'f Mean age 43 yr
Long-term exposure
Nonasthma outcomes and functional gene variants
TNF-a 308 GA/AA
Increased
expression
n = 17%
IL-6174CC
n = 48%
Increased blood
IL-6 levels
IL-6598AA
n = 47%
Increased blood
IL-6 levels
>4 variants with
increased oxidative
stress potential13
(GSTT1, GSTP1,
GSTM1, HMOX,
NQO1, HFE)
TNF-a 308 GG
n - 83%
IL-6174GG
n = 52%
IL-6 598 GG
n = 53%
<4 variants with
increased
oxidative stress
potential13
Ml Stockholm County Panasevich et
Sweden al. (201 3)t
n = 2,698
Ages 45-70 yr
Long-term exposure
^™
-j- Heart rate- Boston, MA area Baia et al.
' corrected QT n = 530 males (201 OVr
interval Mean age 75 yr
(ventricular „, . .
repolarization) Short-term exposure
7-27
-------
Table 7-12 (Continued): Epidemiologic studies evaluating genetic factors.
Factor
Evaluated/Gene
Function
GSTM1 null,
n=299
Null oxidant
metabolizing
capacity
GSTT1 null,
n=270
Null oxidant
metabolizing
capacity
GSTP1 VaMOSVal
orlle105Val
(rs1695), n = 179
Reduced oxidant
metabolizing
capacity (Val/Val)
GSTP1 VaMOSVal
orlle105Val
n = 198
Reduced oxidant
metabolizing
capacity
Reference
Category
GSTM1 positive,
n = 225
GSTT1 positive,
n = 254
GSTP1 Ile105lle,
n = 359
GSTP1 lie/lie
n = 152
Direction of
Effect
Modification3
t
t
t
t
t
—
t
Outcome
Fasting glucose
Insulin level
Fasting glucose
Insulin level
Fasting glucose
Insulin level
Cognitive
function
decrement
Study Population
and other Details
Seoul, Korea
n = 560
Ages 60-87 yr
Short-term exposure
Seoul, Korea
n = 560
Ages 60-87 yr
Short-term exposure
Menorca, Spain
n = 350 children
followed from birth to
age 4 yr
Long-term exposure
Study
Kim and Hong
(2012)t
Kim and Honq
(2012)t
Morales et al.
(2009)t
Nonasthma outcomes and gene variants with unknown functional difference
GSS haplotype
0100000,
n = 1,010
(rs1801310)
Unknown function
GSR, various SNPs
n = 3-21%
Unknown function
GCLM, various
SNPs,
n = 6-35%
Unknown function
GCLC, various
SNPs,
n = 4-54%
Unknown function
Other
haplotypes,
n = 1,096
Other haplotypes
n = 3-21%
Other
haplotypes,
n = 6-35%
Other
haplotypes,
n = 4-54%
t
—
^™
^™
Lung
development
decrement
Alpine, Atascadero,
Lake Elsinore, Lake
Arrowhead,
Lancaster, Lompoc,
Long Beach, Mira
Loma, Riverside,
San Dimas, Santa
Maria, Upland, CA
n = 2,106 children
followed ages
10-1 Syr
Long-term exposure
Breton et al.
(201 1)t
7-28
-------
Table 7-12 (Continued): Epidemiologic studies evaluating genetic factors.
Factor
Evaluated/Gene
Function
MET Tyrosine
receptor kinase CC
(rs1 858830)
n = 102
Reference
Category
MET Tyrosine
receptor kinase
CG/GG
n = 305
Direction of
Effect
Modification3 Outcome
-j- Autism
Study Population
and other Details
Multiple unspecified
locations, CA
n = 252 with autism,
156 without
Ages 2-5 yr
Long-term exposure
Study
Volketal.
(2014)t
ADRB2 = beta-2 adrenergic receptor; CA = California; GCLC = glutamate-cysteine ligase catalytic subunit;
GCLM = glutamate-cysteine ligase; modifier subunit; GSR = glutathione reductase; GSS = glutathione synthetase;
GSTM1 = glutathione S-transferase mu1; GSTP1 = glutathione s-transferase Pi 1; GSTT1 = glutathione S-transferase theta 1;
HFE = hemochromatosis; HMOX = heme oxygenase; IL = interleukin; Ml = myocardial infarction; NQO1 = NAD(P)H:quinone
oxidoreductase 1; NR = not reported; TLR = toll-like receptor; TNF = tumor necrosis factor.
aUp facing arrow indicates that the effect of nitrogen dioxide (NO2) is greater (e.g., larger increase in symptoms) in the group with
the factor evaluated than in the reference group. Down facing arrow indicates that the effect of NO2 is smaller (e.g., smaller
increase in symptoms) in the group with the factor evaluated than in the reference group. A dash indicates no difference in
NO2-related health effect between groups.
bSample size not reported.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
7.5 Sociodemographic Factors
7.5.1
Lifestage
Lifestage refers to a distinguishable time frame in an individual's life characterized by
unique and relatively stable behavioral and/or physiological characteristics that are
associated with development and growth (U.S. EPA. 2014b). The 2008 ISA for Oxides of
Nitrogen (U.S. EPA. 2008c) cited supporting evidence for increased risk of health effects
related to NO2 exposure among children and older adults. Differential health effects of
NC>2 across lifestages theoretically could be due to several factors:
• The human respiratory system is not fully developed until 18-20 years of age, and
therefore, it is plausible to consider children to have an intrinsic risk for
respiratory effects due to potential perturbations in normal lung development.
• Older adults (typically considered those 65 years of age or greater) have weakened
immune function, impaired healing, decrements in pulmonary and cardiovascular
function, and greater prevalence of chronic disease (Table 7-2).
• Exposure/internal dose of NO2 may vary across lifestages due to varying
ventilation and time-activity patterns.
7-29
-------
Studies in this ISA add to the evidence presented in the 2008 ISA indicating increased
risk of NO2-related health effects for children and older adults. Further, this evaluation of
lifestage as a factor that may increase risk for NCh-related health effects draws upon
information about time-activity patterns and ventilation patterns among different
lifestages to assess the potential for differences in NO2 exposure or internal dose among
life stages.
7.5.1.1 Children
According to the 2010 census, 24% of the U.S. population is less than 18 years of age,
with 6.5% less than age 6 years (Howden and Meyer. 2011). The large proportion of
children in the U.S. underscores the public health impact of characterizing the risk of
NO2-related health effects for children.
NC>2 exposure or internal dose differences in children are not well characterized.
However, children and adults differ with respect to time-activity patterns, which are
determinants of inter-individual variability in NC>2 exposure [(Molter et al.. 2012; Kousa
et al.. 2001); Section 3.4.3.11 and which potentially could lead to differences in NO2
exposure and internal dose. The National Human Activity Pattern Survey showed that
children spend more time outdoors compared with adults (Klepeis et al., 1996). and a
longitudinal study in California showed a larger proportion of children reported spending
over 30 minutes performing moderate or vigorous outdoor physical activity (Wu et al..
201 Ib). However, other than school or work, there was little variation among groups in
time spent in various microenvironments (Wu et al.. 201 Ib; Klepeis et al.. 1996).
Although a recent meta-analysis suggested a weaker association between ambient NO2
concentrations and personal NC>2 exposure of children (Meng etal.. 2012a). some studies
found ambient NC>2-related respiratory effects in children for whom there were moderate
personal-ambient correlations [r values of 0.43 and 0.63; (Delfino et al.. 2006; Linn et al..
1996)]. Such personal-ambient NC>2 relationships are consistent with children spending
greater amounts of time outdoors and could be an explanation for larger risk of
NC>2-related health effects for children. A recent analysis found children more likely than
adults to take part in vigorous activity or aerobic exercise [indoors and outdoors; (Wu et
al.. 20 lib)]. Higher activity along with higher ventilation rates relative to lung volume
and higher propensity for oronasal breathing could potentially result in greater NC>2
penetration to the lower respiratory tract of children; however, this has not been
examined for NC>2 (Section 4.2.2.3).
Epidemiologic evidence across diverse locations (U.S., Canada, Europe, Asia, Australia)
consistently demonstrates that short-term increases in ambient NC>2 concentration are
7-30
-------
associated with larger increases in asthma-related hospital admissions, ED visits, or
outpatient visits among children than adults [(Son etal.. 2013; Sinclair et al.. 2010; Ko et
al.. 2007b; Villeneuve et al., 2007; Hinwood et al., 2006; Peel etal.. 2005; Atkinson et
al.. 1999a; Anderson et al.. 1998); Table 7-13]. A few of these studies specified the aim
to compare lifestages a priori (Ko et al., 2007b; Villeneuve et al., 2007; Anderson et al.,
1998). Most results are based on comparisons between children ages 0-14 years and
people ages 15-64 years, and these show NCh-associated increases in asthma hospital
admissions that are 1.8 to 3.4-fold greater in children (Son etal.. 2013; Ko et al.. 2007b;
Atkinson et al., 1999a; Anderson et al.. 1998). Not all results demonstrated increased risk
for children, with some studies of asthma hospital admissions, outpatient visits, and
medication sales showing no difference in association with NC>2 between children and
adults or no association in either group (Burra et al.. 2009; Laurent et al.. 2009;
Migliaretti et al., 2005; Petroeschevsky et al., 2001). Except for Petroeschevsky et al.
(2001). these studies had similar sample sizes as those observing increased risk for
children (Table 7-13). A few results point to larger NCh-related increases in asthma
hospital admissions or ED visits among younger children (e.g., age 0-4 years, 2-4 years)
than older children ages 5-14 years (Samoli et al.. 2011; Villeneuve et al.. 2007);
however, inference from these findings is limited because of the questionable reliability
of asthma diagnosis in children below the age of 5 years (Section 5.2.2.4).
The single available toxicological study does not indicate greater NCh-related lung injury,
inflammation, or lung host defense among juvenile than mature rodents. The most
pronounced effects, including mortality, occurred with 10,000-ppb NC>2 exposure, above
that considered in this ISA to be ambient relevant [(Azoulay-Dupuis et al.. 1983);
Table 7-14]. Because the endpoints examined in the rodents largely were related to
pneumonia and emphysema and are not specific to asthma-related effects, they are not
considered to be in conflict with epidemiologic evidence.
Risk may vary among children according to the time window of exposure because there
are differences in lung development over the course of childhood. Across epidemiologic
studies, asthma development was associated with long-term NO2 exposures assessed for
various time windows, including birth, the first year of life, year of asthma diagnosis, and
lifetime exposure (Section 6.2.2.1). In limited comparisons of time periods in both
epidemiologic and toxicological studies, no single critical time window of exposure was
identified for the effects of short-term or long-term NC>2 exposure on outcomes related to
asthma exacerbation or development (Tables 7-13 and 7-14). In epidemiologic studies,
critical time windows were assessed from longitudinal studies that permitted
within-subject comparisons as children were followed overtime. In cohorts of children
diagnosed with asthma at a median age of 2 or 5 years, NO2 in the first year of life was
associated with similar or lower risk of asthma compared with NC>2 assessed for later in
7-31
-------
childhood [average of ages 1-3 years or average in year of diagnosis; (Nishimura et al.,
2013; Clougherty et al.. 2007)]. The young age of diagnosis in most of these children
limits inference about critical time windows of NC>2 exposure. In the Children's Health
Study (CHS) cohorts, both exposures and respiratory outcomes were examined at various
ages during follow-up from ages 5 or 10 years to 18 years. The heterogeneity among
studies in exposure assessment methods, statistical methods, and examination of
incidence or prevalence of outcomes is not amenable to quantitative comparisons.
However, NO2 exposure was associated with asthma and respiratory symptoms in
childhood (ages 9-13 or 10 years) and into adolescence [ages 13-16 years or
10-18 years; (Jerrett et al.. 2008; McConnell et al.. 2006; Gauderman et al.. 2005;
McConnell et al., 2003; McConnell et al.. 1999)1. also pointing to risk of NCh-associated
respiratory effects throughout childhood.
In conclusion, epidemiologic evidence generally demonstrates that NCh-related asthma
exacerbation is greater in children compared to adults. In a few cases, no difference was
observed by age for NCh-associated asthma outpatient visits and medication use.
However, there is sufficient consistency for asthma hospital admissions and ED visits and
for similar age comparisons (ages 0-14 years vs. 15-64 years). Limited toxicological
results suggest greater NCh-induced pulmonary injury and impaired host defense in
mature compared to juvenile animals, but most of the endpoints examined are related to
pneumonia and emphysema rather than asthma, and results are not considered to
contradict epidemiologic evidence. Neither epidemiologic nor toxicological evidence
clearly identifies a single critical time window of exposure in childhood (e.g., infancy,
later childhood) for asthma-related effects attributable to NO2 exposure. Children have
different time-activity and ventilation patterns than adults, but it is not clear whether
these contribute to higher NO2 exposure or internal dose or increased risk for NCh-related
asthma exacerbation in children. Overall, the consistent epidemiologic evidence for larger
NO2-related asthma exacerbation is adequate to conclude that children are at increased
risk for NCh-related health effects.
7-32
-------
Table 7-13
Factor
Evaluated
Epidemiologic studies evaluating childhood lifestage.
Direction of
Reference Effect
Category Modification3 Outcome
Study
Population
Study
Details
Study
Asthma-related outcomes and short-term NO2 exposure: age of health effect
Childhood
Ages 0-14 yr
n= 23,596
Childhood
Ages 0-14 yr
n = mean
8.7/day
Childhood
Ages 0-14 yr
n = mean
19.5/day
Childhood
Ages
0-14 yrb
Childhood
Ages 0-14 yr
n = mean
5.6/day
Childhood
Ages 0-14 yr
n = mean
2.6/day
Childhood
Ages 0-4 yr
n = median
1/day
Childhood
Ages 5-14 yr
n = 13,145
Childhood
Ages
2-18yrb
Adulthood
Ages 15-65 yr
n = 21,204
Adulthood
Ages 15-64 yr
n = mean
4.3/day
Adulthood
Ages 15-64 yr
n = mean
13. 1/day
Younger
adulthood
Ages
15-64yrb
All ages
n = mean
8.8/day
Adulthood
Ages 15-64 yr
n = mean
1.7/day
Childhood
Ages 5-14 yr
n = median
0/day
Childhood
Ages 15-44 yr
n = 24,916
Adulthood
Ages >19 yrb
-j- Asthma
' hospital
admission
-j- Asthma
' hospital
admission
-j- Asthma
' hospital
admission
i Asthma
^ hospital
admission
-j- Asthma
' hospital
admission
_ Asthma
hospital
admission
-j- Asthma
' hospital
admission
-j- Asthma ED
' visits
* Asthma ED
' visits
n = 69,176
admissions;
15 hospitals
n = 5.7-36.3
mean daily
admissions
across cities
n = mean
35.1 admissions
per day
n = 202,472
admissions for
asthma, heart
diseases,
nonrespiratory
diseases
n = 8.8 mean
asthma
admissions per
day
n = 13,246
asthma
admissions
n = 3,601
asthma
admissions,
3 children's
hospitals
n = 57,192 ED
visits,
5 hospitals
n = mean
39.0 asthma
visits/day,
31 hospitals
Hong Kong,
2000-2005
8 cities,
South
Korea,
2003-2008
London,
U.K.,
1987-1992
Turin, Italy,
1997-1999
Perth,
Australia,
1992-1998
Brisbane,
Australia,
1987-1994
Athens,
Greece,
2001-2004
Edmonton,
Canada,
1992-2002
Atlanta, GA,
1993-2000
Ko et al.
(2007b)t
Son et al.
(201 3)t
Anderson et al.
(1998)
Miqliaretti et al.
(2005)
Hinwood et al.
(2006)
Petroeschevskv
etal. (2001)
Samoli etal.
(201 1)t
Villeneuve et al.
(2007)t
Peel etal. (2005)
7-33
-------
Table 7-13 (Continued): Epidemiologic studies evaluating childhood lifestage.
Factor
Evaluated
Childhood
Ages 0-14 yr
n = mean
12.0/day
Childhood
Ages NR
n= 28,487
Childhood
Ages 1-17 yr
n = 1,146,215
Childhood
Ages 0-19 yr
n = 7,774
Reference
Category
Adulthood
Ages 15-64 yr
n = mean
1 1 .4/day
Adulthood
Ages NR
n = 19,085
Adulthood
Ages 18-64 yr
n = 1,558,071
Adulthood
Ages 20-39 yr
n = 7,347
Direction of
Effect
Modification3 Outcome
-j- Asthma ED
' visits
-j- Asthma
' outpatient
visits
_ Asthma
outpatient
visits
_ Asthma
medication
sales
Study
Population
n = 28,435
asthma visits,
12 hospitals
n = 47,572
asthma visits
n = 2,704,286
Asthma visits,
Ontario Health
Insurance Plan
Ages 1-64 yr
n = 261, 063
Ages 0-39 yr
Study
Details
London,
U.K.,
1992-1994
Atlanta, GA,
1998-2002
Toronto,
Canada,
1992-2001
Strasbourg,
France,
2004
Study
Atkinson et al.
(1999a)
Sinclair et
(201 0)t
Burra et al
(2009VT
Laurent et
(2009)t
aL.
.aL.
Asthma-related outcomes and long-term exposure: time window of childhood exposure
Exposure in
yr of
diagnosis
Exposure in
first yr of life
Exposure in
first yr of life
Exposure in
first 3 yr of life
•j> Asthma
' incidence
Median age of
diagnosis: 5 yr
_ Asthma
prevalence
Median age of
diagnosis: 2 yr
n = 417 children
n = 4,320
children
enrolled
between ages
of 8 and 21 yr
Boston, MA,
Follow-up:
prenatally
(1987-1993)
to 1 997
5 U.S. cities,
1996-2001
Clouqhertv
(2007)t
etal.
Nishimura et al.
(201 3)t
ED = emergency department; NR = not reported.
aUp facing arrow indicates that the effect of nitrogen dioxide (NO2) is greater (e.g., larger increase in hospital admission) in the
group with the factor evaluated than in the reference group. Down facing arrow indicates that the effect of NO2 is smaller in the
group with the factor evaluated than in the reference group. A dash indicates no difference in NO2-related health effect between
groups.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
7-34
-------
Table 7-14 Toxicological studies evaluating childhood lifestage.
Factor Reference
Evaluated Category
Direction of
Effect
Modification3
Outcome
Asthma-related outcomes and short-term exposure:
Prenatal/ Weanling
weanling exposure
exposure
Nonasthma outcomes
Juvenile Adult age
age
Juvenile Adult age
age
Juvenile Adult age
age
Juvenile Adult age
age
1
t
and short-term
-
—
-
1
Eosinophils,
neutrophils,
lymphocytes
Alveolar
macrophage
activity
Animal Model
Study Details Study
time window of childhood exposure
Rats (Brown
Norway, model
of allergic
n = 5-7/group
Females
Prenatal/weaning Kumae and
exposure: Breedinq pairs Arakawa
mated in 200, 500, or (2006)
2,000 ppb NO2. Litters
continuously exposed to
age 8 and 12 weeks
Weanling exposure:
5 week old rats exposed
to 200, 500, or 2, 000 ppb
NO2 continuously to age
8 and 12 weeks
exposure: age of health effect
Mortality
Inflammatory
cells in
- bronchial
airway,
alveolar
edema
Rats (Wistar)
n = 5— 8/group
Guinea pigs
(Hartley)
n = 5-8/group
Rats (Wistar)
n = 5-8/group
Guinea pigs
(Hartley)
n = 5-8/group
2,000 ppb NO2 for 3 days Azoulav-
ataaes 5. 10. 21. 45. 55. Dupuisetal.
and 60 days (1983)
2,000 ppb NO2 for 3 days
at 5, 10, 21, 45, 55, and
. 60 days of age
NO2 = nitrogen dioxide.
aUp facing arrow indicates that the effect of NO2 is greater (e.g., greater increase in alveolar macrophage activity) in the group with
the factor evaluated than in the reference group. Down facing arrow indicates that the effect of NO2 is smaller in the group with the
factor evaluated than in the reference group. A dash indicates no difference in NO2-related health effect between groups.
7.5.1.2
Older Adults
According to the 2012 National Population Projections issued by the U.S. Census
Bureau, 13% of the U.S. population was age 65 years or older in 2010, and by 2030, this
fraction is estimated to grow to 20% (Ortman et al., 2014). Thus, this lifestage represents
a substantial proportion of the U.S. population that is potentially at increased risk for
health effects related to NO2 exposure.
7-35
-------
It is not clear whether NC>2 exposure or uptake in the respiratory tract differs between
older adults and younger adults. The National Human Activity Pattern Survey indicated
that older adults spend more time outdoors at home but less time outdoors in other
locations or in vehicles (Klepeis et al.. 1996). A recent study in California did not
consistently indicate differences in time spent in particular microenvironments or time
engaged in vigorous or outdoor activity (Wuetal.. 201 Ib).
In contrast with exposure or dose information, epidemiologic evidence points to greater
risk of NO2-related health effects in older adults (ages 65 years and older) compared with
younger adults. The 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c) indicated that
older adults may be at increased risk for NCh-related respiratory effects and mortality,
and recent epidemiologic findings add to this body of evidence (Table 7-15).
Comparisons of older and younger adults with respect to NO2-related asthma
exacerbation are limited and generally show larger (one to threefold) effects in adults
ages 65 years or older than among individuals ages 15-64 years or 15-65 years (Ko et
al.. 2007b: Villeneuve et al.. 2007; Migliaretti etal.. 2005; Anderson et al.. 1998). Some
these studies specified the aim to compare lifestages a priori (Ko et al., 2007b; Villeneuve
et al.. 2007; Anderson et al.. 1998). A few studies showed no increased risk in older
adults [(Son etal..2013; Hinwood et al.. 2006): Table 7-15]. and it is not clear that the
sample sizes differed from those in studies observing increased risk for older adults.
Results for all respiratory hospital admissions combined also tend to show larger
associations with NO2 among older adults ages 65 years or older (Arbex et al.. 2009:
Wong et al.. 2009: Hinwood et al.. 2006: Atkinson et al.. 1999a). Controlled human
exposure studies of older adults did not examine asthma-related effects or draw
comparisons with younger adults. Examination of older adults was limited to healthy
adults, and the sparse evidence shows only statistically nonsignificant decrements in lung
function following NC>2 exposure [(Gong etal.. 2005: Morrow et al.. 1992): Table 7-16].
For nonrespiratory effects, associations of short-term NC>2 with total mortality in most
studies were larger in adults ages 65 or older than in younger adults (Table 7-15). with
evidence pointing to elevated risk among the oldest adults ages greater than 75 or
85 years (Chen et al.. 2012b: Cakmak et al.. 20 lib: Chiusolo etal.. 2011). Studies of
long-term NC>2 exposure do not provide strong evidence of elevated risk of health effects
among older adults, with inconsistent effect measure modification observed for total or
cause-specific mortality (Dimakopoulou et al.. 2014: Carey etal.. 2013: Cesaroni et al..
2013: Zhang etal.. 2011: Maheswaran etal.. 2010: Yorifuii etal.. 2010) and generally no
difference by age group observed for associations with cardiovascular effects or diabetes
(Eze etal.. 2014: Atkinson etal.. 2013: Rivera et al.. 2013: Wichmannet al.. 2013:
Andersen etal.. 2012b: Rosenlund et al.. 2009a: Ljungman et al.. 2008: Min etal.. 2008).
The age to define older adults varied among mortality and cardiovascular effect studies
7-36
-------
from 50 to 75 years. Further, it is uncertain the extent to which the findings in older
adults for mortality or cardiovascular effects can be attributable to NO2 because of
uncertainty in whether relationships of NC>2 exposure with these health effects are
independent of highly correlated copollutants (Sections 5.3.11. 5.4.8. 6.3.9. and 6.5.3).In
particular, older adults consistently have increased risk for mortality associated with
short-term PM2 5 exposure (U.S. EPA. 2009a). Differences in confounding among
lifestages are not well characterized for NC>2. As examined in a few studies of asthma
hospital admissions, adjustment for total suspended particles eliminated the NCh
association in older adults but not younger adults or children (Migliaretti et al.. 2005).
whereas the NO2 association in older adults persisted with adjustment for black smoke
(Anderson et al.. 1998). Other issues limiting inference are the lack of examination of
copollutants and exposure assessment based on measurements from central site monitors.
Although NC>2 exposure or dosimetry differences in older adults are not characterized,
there is generally supportive evidence for larger NCh-related risk of hospital admissions
and ED visits for asthma in older adults compared to younger adults. Most of these
studies compared adults ages 64-65 years or older with individuals ages 15 to 64 or
65 years. Controlled human exposure studies do not indicate increased risk in older
adults, but the limited evidence is based on healthy adults not those with asthma. Thus,
the evidence from controlled human exposure studies is not considered to conflict with
epidemiologic evidence. As described in Section 5.2.9. a relationship between NO2
exposure and respiratory effects in healthy populations is not clearly demonstrated. Older
adults have larger NO2-related increases in total mortality but not cardiovascular effects;
however, inferences about the risk for older adults from this evidence is limited because
of uncertainties regarding the independent effect of NO2 on nonasthma outcomes.
Overall, the consistent epidemiologic evidence for larger NO2-related asthma hospital
admissions and ED visits is adequate to conclude that older adults are at increased risk
for NO2-related health effects.
7-37
-------
Table 7-15
Factor
Evaluated
Epidemiologic studies evaluating
Reference
Category
Asthma-related outcomes and
Older
adulthood
Ages >65 yr
n= 24,916
Older
adulthood
Ages >65 yr
n = mean
2.6/day
Older
adulthood
Ages >64 yrb
Older
adulthood
Ages >65 yrb
Older
adulthood
Ages >65 yr
n= 4,705
Older
adulthood
Ages >65 yr
n = mean
3/day
Younger
adulthood
Ages 15-65 yr
n = 21,204
Younger
adulthood
Ages 15-64 yr
n = mean
13.1 /day
Younger
adulthood
Ages
15-64yrb
All agesb
Younger
adulthood
Ages 15-64 yr
n = 32,815
Younger
adulthood
Ages 15-64 yr
n = mean
4.3/day
Direction of
Effect
Modification3 Outcome
short-term exposure
-j- Asthma hospital
' admission
+ Asthma hospital
' admission
+ Asthma hospital
' admission
_ Asthma hospital
admission
-j- Asthma ED
' visits
_ Asthma and
allergic disease
hospital
admission
older adult lifestage.
Study
Population
n = 69,176
admissions,
15 hospitals
n = mean
35.1 admissions
per day
n = 202,472
admissions for
asthma, heart
diseases,
nonrespiratory
diseases
n = 8.8 mean
asthma
admissions per
day
n = 57,912 visits,
5 hospitals
n = 5.7-36.3
mean daily
admissions
across cities
Study Details
Hong Kong,
2000-2005
London, U.K.,
1987-1992
Turin, Italy,
1997-1999
Perth,
Australia,
1992-1998
Edmonton,
Canada,
1992-2002
8 South
Korean cities,
2003-2008
Study
Ko et al.
(2007b)t
Anderson et
al. (1998)
Miqliaretti et
al. (2005)
Hinwood et
(2006)
aL
Villeneuve et
al. (2007)t
Son et al.
(201 3)t
7-38
-------
Table 7-15 (Continued): Epidemiologic studies evaluating older adult lifestage.
Factor
Evaluated
Direction of
Reference Effect
Category Modification3 Outcome
Study
Population
Study Details Study
Nonasthma outcomes and short-term exposure
Older
adulthood
Ages >65 yrb
Older
adulthood
Ages >64 yr,
n = 789
Older
adulthood
Ages >65 yrb
Older
adulthood
Ages >65 yr,
n = 187,608
Older
adulthood
Ages >85 yr
n = 90,070
Older
adulthood
Ages 65-74 yr
n = 52,689
All agesb _
t
^
t
Younger *
adulthood '
Ages
40-64 yr,
n = 980
Younger *
adulthood, '
childhood
Ages 5-64 yrb
Younger *
adulthood, '
childhood
Ages <65 yr,
n = 91,253
Younger +
adulthood '
Ages 35-64 yr
n = 181,031
Younger i
adulthood ^
Ages 35-64 yr
n = 35,803
COPD hospital
admission
COPD hospital
admission with
influenza
Acute
respiratory
disease hospital
admission
Cardiovascular
hospital
admission
COPD ED visits
Total mortality
Total mortality
Total mortality
n = 91. 5-270.3
mean daily
- admissions
across
conditions,
14 hospitals
n = 1,769 COPD
ED visits,
40 hospitals
n = 11-119
mean daily
deaths across
cities
n = 7.29-15.8
mean daily
deaths across
locations
n = 276,205
natural deaths
Honq Konq, Wonq et al.
1996-2002 (2009)t
Sao Paulo, Arbex et al.
Brazil, (2009)t
2001-2003
17 Chinese Chen et al.
cities (2012b)t
Santiago Cakmak et al.
Province, (2011bVr
Chile (7 urban
centers),
1997-2007
10 Italian Chiusolo et al.
cities, (201 1)t
2001-2005
7-39
-------
Table 7-15 (Continued): Epidemiologic studies evaluating older adult lifestage.
Factor
Evaluated
Nonasthma
Older
adulthood
Ages >75 yrb
Older
adulthood
Ages >75 yrb
Older
adulthood
Ages >60 yr
n = 365,368
Older
adulthood
Ages >60 yr
n =4,061
Older
adulthood
Ages >60 yrb
Older
adulthood
Ages >70 yr
n = 1,329
Older
adulthood
Ages >50 yr
n=635
Reference
Category
outcomes and
Younger
adulthood
Ages <60 yr
Younger
adulthood
Ages
65-75 yrb
Younger
adulthood
Ages 40-60
n = 470,239
Younger
adulthood
Ages <60 yr
n = 5,880
Younger
adulthood
Ages <60 yr
Younger
adulthood
Ages <70 yr
n = 527
Younger
adulthood
Ages 20-50
n = 242
Direction of
Effect
Modification3 Outcome
long-term exposure
i Total mortality,
^ cardiovascular
b mortality
_ Lung cancer
mortality
_ Lung cancer or
cardiopulmonary
mortality
* Total mortality
yr
_ Cardiovascular
mortality
* Respiratory
' mortality
b
i Total mortality
* HRV decrement
' in low frequency
yr domain
Study
Population
n = 1,265,058
Ages >30 yr
n = 13,444
Ages >65 yr
n = 835,607
deaths
Ages 40-89 yr
n = 9,941, 256
deaths
Ages 35-103 yr
n = 307,553
Mean age
across
16 cohorts
41. 9-73.0 yr at
baseline
n = 3,320
Mean age 70 yr
n = 1,349
healthy subjects
Mean age 44 yr
Study Details
Rome, Italy,
2001-2010
Shizuoka,
Japan,
1999-2006
England
Follow-up:
2003-2007
NO2 exposure
assessed for
2002
Shenyang,
China
Follow-up:
1998-2009
NO2 exposure
assessed for
1998-2009
Europe
Follow-up:
1985-2007
NO2 exposure
assessed for
2008-2011
London,
England
Follow-up:
1995-2005
NO2 exposure
assessed for
2002
Taein Island,
South Korea,
2003-2004
Study
Cesaroni et al.
(201 3)t
Yorifuii et al.
(201 0)1-
Carev et al.
(201 3)t
Zhanq et al.
(201 1)t
Dimakopoulou
etal. (2014)t
Maheswaran
etal. (201 0)t
Min et al.
(2008)t
7-40
-------
Table 7-15 (Continued): Epidemiologic studies evaluating older adult lifestage.
Factor
Evaluated
Older
adulthood
Ages >75 yr
n = 1,995
Older
adulthood
Ages >65 yr
n = 50
Older
adulthood
Ages >65 yr
n = 137,184
Older
adulthood
Ages
65-89 yrb
Older
adulthood
Ages >56 yr
n = 1,297
Older
adulthood
Ages >56 yr,
n = 106
Older
adulthood
Ages >60 yr
Femaleb
Maleb
Reference
Category
Younger
adulthood
Ages <60 yr,
n = 1,252
Younger
adulthood
Ages
60-75 yr,
n = 1,410
Younger
adulthood
Ages <65 yr
n = 60
Younger
adulthood
Ages <65 yr
n = 417,156
Younger
adulthood
Ages
40-64 yrb
Younger
adulthood
Ages <56 yr
n = 687
Younger
adulthood
Ages <56 yr,
n = 36
Younger
adulthood
Ages <60 yr
Femaleb
Maleb
Direction of
Effect
Modification3 Outcome
_ Out-of-hospital
cardiac arrest
t
Ventricular
arrhythmia
Ml
_ Heart failure
Incidence stroke
_ Fatal stroke
_ Intima media
thickness cca
i Intima media
"*" thickness 6seg
+ Intima media
' thickness cca
_ Intima media
thickness 6seg
Study
Population
n = 4,657 events
n = 211 with
implantable
cardioverter
defibrillators
Ages 28-85 yr
n = 43,275
cases,
51 1,065 controls
Ages 15-79 yr
n = 836,557
Ages 40-89 yr
at baseline
n = 1,984
Ages 50-65 yr
at baseline
n = 142
Ages 50-65 yr
at baseline
n = 2,780
Median age
'58yr
Study Details
Copenhagen,
Denmark,
2000-2010
Gothenburg,
Stockholm,
Sweden,
2001-2006
Stockholm
County,
Sweden,
1985-1996
England,
2003-2007
Copenhagen,
Aarhus
counties,
Denmark,
1993-2006
Girona
Province,
Spain,
2007-2010
Study
Wichmann et
al. (201 3)t
Liunqman et
al. (2008)t
Rosenlund et
al. (2009a)t
Atkinson et al.
(201 3)t
Andersen et
al. (2012b)t
Rivera et al.
(201 3)t
7-41
-------
Table 7-15 (Continued): Epidemiologic studies evaluating older adult lifestage.
Factor
Evaluated
Older
adulthood
Ages >65 yr
n = 2,234
Older
adulthood
Ages >50 yrb
Reference
Category
Younger
adulthood
Ages 55-65 yr
n = 3,913
Younger
adulthood
Ages <50 yrb
Direction of
Effect
Modification3 Outcome
_ Prevalent
hypertension
_ Diabetes
Study
Population
n = 24,845
Mean age
45.59 yr
n = 6,392
Ages 29-73 yr
Study Details
Shenyang,
Anshan and
Jinzhou,
China,
2006-2008
Switzerland,
2002
Study
Donq et al.
(2013b)t
Eze et al.
(2014)t
COPD = chronic obstructive pulmonary disease; ED = emergency department; HRV = heart rate variability; Ml = myocardial
infarction.
aUp facing arrow indicates that the effect of nitrogen dioxide (NO2) is greater (e.g., larger risk of hospital admission, larger
decrement in HRV) in the group with the factor evaluated than in the reference group. Down facing arrow indicates that the effect
of NO2 is smaller in the group with the factor evaluated than in the reference group. A dash indicates no difference in NO2-related
health effect between groups.
bSample size not reported.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
Table 7-16 Controlled human exposure studies informing risk for older adult
lifestage.
Factor Reference
Evaluated Category
Direction of
Effect3 Outcome Study Population Study Details Study
Nonasthma outcomes and short-term exposure
Older None
adulthood
Older None
adulthood
_ Lung function
decrement
_ Lung function
decrement,
pulmonary
inflammation
n=20
(10 males,
10 females)
Mean age 61 yr
n=6
(2 males, 4 females)
Mean age 68 yr
300 ppb NO2 for
4 h with exercise
400 ppb NO2 for
2 h with exercise
Morrow et al.
(1992)
Gonq et al.
(2005)
NO2 = nitrogen dioxide.
aThese studies only examined older adults and have no reference group. A dash indicates that NO2 was not observed to induce an
effect in the older adults relative to clean air exposure.
7-42
-------
7.5.2 Socioeconomic Status
SES is a composite measure that usually consists of economic status measured by
income, social status measured by education, and work status measured by occupation.
Persons with lower SES generally have been found to have a higher prevalence of
pre-existing diseases, potential inequities in access to resources such as healthcare, and
possibly increased nutritional deficiencies. Neighborhoods comprising low SES
populations are characterized as having fewer resources and more psychosocial stressors
and air pollution sources (Diez Roux and Mair. 2010; O'Neill et al.. 2003); thus,
community-level SES may influence health. Any or a combination of these factors could
link low SES to increased risk for NC>2-related health effects. According to the
U.S. Census Bureau's American Community Survey, 15.8% (approximately 48.8 million)
of Americans were of poverty status in 2013 as defined by household income, which is
one metric used to define SES (Bishaw and Fontenot 2014). Across the indicators of SES
examined (e.g., education level, employment status, insurance status, social deprivation,
access to health care), there is some evidence for higher NO2 exposure and larger risk of
NC>2-related health effects among low SES groups in the population, but these
relationships are not uniformly observed (Table 7-17). A challenge in synthesizing
findings across studies is the array of SES indicators examined. Further, many studies
were conducted outside of the U.S., and because demographics, bureaucracy, and the
local economy vary among countries, relationships of SES metrics to extent of
deprivation or inequities also may vary among countries. In addition, the studies
examined SES for individual subjects (e.g., household income) and for communities
(e.g., mean income of census tract). Because individual- and community-level SES
indicators have been shown to have independent influences on health, these indicators
were evaluated separately as modifiers of NO2 exposure and NCh-related health effects.
Several studies relate higher NO2 exposure with indicators of low SES, but the
relationship varies across communities, levels of SES, and indicators of SES. Results
from studies conducted in the U.S., Canada, and Europe point to a relationship between
higher ambient NC>2 exposure among populations of low SES as determined by
household income, job class (e.g., unskilled, professional, skilled manual labor), or
education. Higher ambient NC>2 concentrations have been measured in communities in
Montreal, Canada and Los Angeles, CA with high proportions of nonwhite residents and
low SES residents (Suetal.. 2012; Molitoretal.. 2011; Grouse et al.. 2009b; Su et al..
2009c). While many of these studies examined community-level correlations based on
census block or tract SES and NO2 modeled at the neighborhood scale (Clark etal.. 2014;
Deguen and Zmirou-Navier. 2010; Namdeo and Stringer. 2008; Kruize et al.. 2007;
Chaix et al., 2006; Mitchell. 2005). studies with data from individuals also found
relationships between higher residential or personal NO2 exposure and lower SES (Llop
7-43
-------
et al., 2011; Deguen and Zmirou-Navier. 2010). A U.S.-wide analysis of census blocks
suggests inequities in NC>2 exposure in the low SES communities by age, with higher
exposures indicated for children and older adults (Clark et al.. 2014).
While most results indicate higher NC>2 exposure in low SES groups, some indicate that
the relationship between NCh exposure and SES varies in strength and direction. In some
cases, a nonlinear relationship is observed with either no difference in NC>2
concentrations among communities in the higher end of the income distribution [e.g., top
50%; (Kruize et al.. 2007)1 or higher NC>2 concentrations in some affluent communities in
the downtown core of a city (Grouse et al.. 2009b). Other studies find that the relationship
varies across communities (Stroh et al., 2005) and among particular SES indicators
[e.g., education but not occupation, country of birth but not education; (Stroh et al.. 2005;
Rotko etal. 2001)1. The relationship between NCh exposure and SES also may weaken
overtime as was forecasted for Leeds, U.K. Over the period 1993-2005, increased
prevalence of vehicles with more efficient emissions controls and surcharges for using
congested roads were predicted to reduce the discrepancy in NO2 exposures between
groups with high and low deprivation index [combining unemployment, noncar
ownership, nonhome ownership, and household overcrowding; (Mitchell. 2005)]. O'Neill
et al. (2003) noted that several factors might alter the relationship between NO2 exposure
and SES, including changing development, migration, and transportation patterns all of
which could result in individuals of high SES having high NO2 exposures.
There is also the possibility that a multitude of factors may interact to influence the risk
of NO2-related health effects in populations of low SES. The hypothesis of "double
jeopardy" describes interactions between higher air pollution exposure and social
inequities in health, whereby risk of health effects may be increased for low SES and/or
nonwhite populations because of increased exposure as well as increased psychosocial
stress or less access to health services (O'Neill et al., 2003). An index combining risk
factors such as air pollution concentrations, including NO2, with nonwhite population and
low SES population has been constructed for communities in a few California cities (Su
et al.. 2012; 2009c). However, health effects have not been examined in relation to such
indices comprising NO2, and for the studies evaluated in this ISA, the risk for certain SES
or nonwhite populations resulting from multiple stressors has not been characterized.
While there is strong evidence for relationships between short-term and long-term NO2
exposure and asthma exacerbation (Section 5.2.9) and development (Section 6.2.9).
evidence does not clearly indicate differences among groups of varying SES
(Table 7-17). All of these studies described their analyses of SES as a priori objectives.
U.S. studies observed higher risk of NO2-related asthma exacerbation and incidence in
groups made up of low SES children as defined by high psychosocial stress due to
7-44
-------
exposure to community violence (Clougherty et al.. 2007) or having no health insurance
(Grineski et al.. 2010). A study in Asia did not observe differences in NCh-related asthma
outpatient visits by insurance status (Kim et al.. 2007). Census tract- or block-level
income or a composite SES index did not modify associations of short-term NC>2
exposure with asthma hospital admissions, physician visits, or medication sales among all
ages (Burraetal.. 2009: Laurent et al.. 2009: Lin et al.. 2004b).
There is limited analysis of interactions between SES and other factors in modifying risk
of NO2-related asthma hospital admissions, but the results are not conclusive regarding
the potential risk due to multiple co-occurring factors within a population. For children,
the association did not differ between Hispanic and white children, except in the group
without health insurance (Grineski et al.. 2010). In another study, associations did not
differ between low- and high-income census tracts among children, adults, males, or
females (Burraetal.. 2009).
Several studies in various countries found larger associations for short-term and
long-term NC>2 exposures with total mortality in low SES compared to high SES groups
as indicated by education, income, or employment [(Carey etal.. 2013: Cesaroni et al..
2013: Chenetal.. 2012b: Cakmaket al.. 20lib: Chiusolo etal.. 2011): Table 7-171. The
increased risk for low SES was observed with both individual- and community-level SES
indicators. Despite the consistency of results, the extent to which the findings can be
attributed specifically to NO2 is uncertain because potential confounding of relationships
between NCh exposure and total mortality by traffic-related copollutants has not been
adequately assessed (Sections 5.4.8 and 6.5.3).
Evidence that SES modifies associations of long-term NO2 exposure with cardiovascular
effects, diabetes, reproductive effects, developmental effects, or cancer is unclear
(Table 7-17). However, independent relationships of NO2 exposure with these health
effects are uncertain (Sections 6.3.9. 6.4.5. and 6.6.9). Most studies found no difference
among SES groups (Eze etal.. 2014: Andersen et al.. 2012c: Guxens et al.. 2012: Pereira
etal.. 2012: Zhang etal.. 2011: Lenters etal.. 2010: Yorifuji etal.. 2010: Rosenlund et
al.. 2009a) or inconsistent effect measure modification among the outcomes examined
(Foraster et al.. 2014: Andersen et al.. 2012b). A few studies found larger associations
among lower SES groups (Becerra et al.. 2013: Morello-Frosch et al.. 2010). but just as
many observed weaker NC^-related effects among lower SES groups (Atkinson et al..
2013: Rivera etal.. 2013). A diverse set of SES indicators was examined, and results are
inconsistent even among studies examining education or income. Results are inconsistent
for both individual- and community-level SES indicators.
Interpreting the evidence for exposure or effect measure modification by SES is
challenging given the diversity of individual- and community-level SES indicators
7-45
-------
examined across studies, array of countries where studies were conducted, and
uncertainty in the independent effect of NC>2 on many of the health outcomes examined.
Evidence indicates higher NO2 exposure among low SES communities, although elevated
concentrations are also reported for some high SES communities. Associations between
short-term NC>2 exposure and asthma exacerbation do not consistently vary by SES, and
results for cardiovascular effects, diabetes, reproductive effects, developmental effects,
and cancer also are inconsistent. No clear pattern of effect measure modification is
observed for individual- or community-level SES indicators. Evidence consistently
demonstrates larger associations between NO2 exposure and total mortality among low
SES groups, but uncertainty remains in attributing the findings specifically to NC>2. The
mortality evidence combined with the evidence for higher NCh exposure is suggestive
that low SES populations are at increased risk for NCh-related health effects.
Table 7-17 Epidemiologic studies evaluating socioeconomic status.
Factor
Evaluated
Reference
Category
Asthma-related outcomes
No
insurance
n=205
Lowest
quintile
income-
based
insurance
premiums
n = 24%
High
exposure to
violence13
Insurance
n= 2,508
Private
n= 2,015
Medicaid
Highest
quintile
income-
based
insurance
premiums
n = 17%
Low
exposure to
violence13
Asthma-related outcomes
Low income
census
tracts'3
High income
census
tracts'3
Direction of
Effect
Modification3 Outcome
Study
Population
Study Details
Study
and individual-level SES indicators
* Asthma hospital
' admission
_ Asthma
emergency
outpatient visit
* Asthma
' incidence
and area-level SES indicators
_ Asthma hospital
admission
n = 4,728
asthma
admissions
Ages <14 yr
n = 254
mean visits
per day
Prior asthma
diagnosis
required
n = 417
children
followed from
prenatal
period
n = 3,822
admissions
Ages 6-12 yr
Phoenix, AZ,
2001-2003
Short-term exposure
Seoul, Korea, 2002
Short-term exposure
Boston, MA,
1987-1993, Follow-
up to 1997
Long-term exposure
Vancouver, Canada
(13 subdivisions),
1987-1998
Short-term exposure
Grineski et
al. (201 0)t
Kim et al.
(2007)
Clouqherty
et al. (2007)t
Lin et al.
(2004b)
7-46
-------
Table 7-17 (Continued): Epidemiologic studies evaluating socioeconomic status.
Factor
Evaluated
Lowest
quintile for
census tract
income
n =610,121
Lowest
stratum SES
census
blocks:
composite of
income, job,
education,
housing
n = 43,674
Reference
Category
Highest
quintile for
census tract
income
n = 527,385
Highest
stratum SES
census
blocks
n=49,111
Nonasthma outcomes and
Low
education
(illiterate/
primary
school)b
Blue collar
work, Low-
level white
collarwork
n = 57.9%
Low income
(mean of
controls)
n = 55.6%
High
education
(>high
school)
n=41.5%
High
education
n = 35%
Higher
education/
technician
n = 526
Direction of
Effect
Modification3 Outcome
_ Asthma
physician visits
_ Asthma
medication
sales
individual-level SES indicators
t Total mortality
I
Ml
_ Atherosclerosis
(carotid intima-
media
thickness)
i Atherosclerosis
"*" (carotid intima
media
thickness)
Study
Population
n = 2,704,286
asthma visits,
Ontario Health
Insurance
Plan
Ages 1-64 yr
n= 261, 063
Ages 0-39 yr
n = 11-119
mean daily
deaths across
cities
n = 43,275
cases,
511,065
controls
n = 745
Ages 26-30 yr
n = 2,780
Median age
58 yr
Study Details
Toronto, Canada,
1992-2001
Short-term exposure
Strasbourg, France,
2004
Short-term exposure
17 Chinese cities
Short-term exposure
Stockholm county,
Sweden,
1985-1996
Long-term exposure
Utrecht, the
Netherlands,
1999-2000
Long-term exposure
Girona Province,
Spain,
2007-2010
Long-term exposure
Study
Burra et al.
(2009)t
Laurent et al.
(2009)t
Chen et al.
(2012b)t
Rosenlund et
al. (2009a)t
Lenters et al.
(201 0)1-
Rivera et al.
(201 3)t
7-47
-------
Table 7-17 (Continued): Epidemiologic studies evaluating socioeconomic status.
Factor
Evaluated
Lower
medium
education
(<10yr)
n = 1,628
Low or
medium
education
(<10yr)
n = 110
Low or
medium
education
(<10yr)
n= 40,956
Low or
medium
education
(primary or
secondary)
n= 4,586
Illiterate/
primary
education
n = 1,540
Low income
(<200 mo)
n = 1,817
Low
education
n = 5,970
Low
education
(<8 yr)
n = 33%
Financially
incapable
(self-
reported)
n= 4,054
Reference
Category
High
education
(>10yr)
n = 356
High
education
(>10yr)
n = 32
High
education
(>10yr)
n = 10,862
High
education
(college or
university)
n = 1,806
Secondary/
university
education
n= 2,160
High income
(>800 mo)
n= 2,618
High
education
n = 3,971
High
education
(<10yr)
n=21%
Financially
capable
(self-
reported)
n = 7,340
Direction of
Effect
Modification3 Outcome
_ Incident stroke
* Fatal stroke
_ Diabetes
Diabetes
_ Systolic blood
pressure
-j- Diastolic blood
' pressure
_ Cardiovascular
mortality
—
_ Diabetes-
related mortality
_ Lung cancer
mortality
Study
Population
n = 1,984
Ages 50-65 yr
at baseline
n = 142
Ages 50-65 yr
at baseline
n = 51,818
Ages 50-65 yr
at baseline
n = 6,392
Ages 29-73 yr
n = 3,700
Ages 35-83 yr
n = 9,941, 256
deaths
Ages
oc -i no wr
n = 52,061
Ages 50-64 yr
n = 13,444
Ages >65 yr
Study Details
Copenhagen,
Aarhus counties,
Denmark,
1993-2006
Long-term exposure
Copenhagen,
Aarhus counties,
Denmark,
1993-2006
Long-term exposure
Switzerland, 2002
Long-term exposure
Girona Province,
Spain
Long-term exposure
Shenyang, China
Follow-up:
1998-2009
NO2 assessed for
1998-2009
Long-term exposure
Denmark
Follow-up:
1971-2009
NO2 exposure
assessed for
1971-2009
Long-term exposure
Shizuoka, Japan,
1999-2006
Long-term exposure
Study
Andersen et
al. (2012b)t
Andersen et
al. (2012c)t
Eze et al.
(2014)t
Foraster et
al. (2014)t
Zhanq et al.
(201 1)t
Raaschou-
Nielsen et al.
(2012)t
Yorifuji et al.
(201 0)t
7-48
-------
Table 7-17 (Continued): Epidemiologic studies evaluating socioeconomic status.
Factor
Evaluated
Low
parental
social class
(semi-
skilled/un-
skilled
occupation
high
school)
n = 3,926
Direction of
Effect
Modification3 Outcome
_ Mental
development
decrement in
infants at age
14 mo
* Autistic disorder
' in children
Study
Population
n = 1,889
children
followed from
prenatal
period
n = 7,603
children with
autism,
10 controls
per case
Study Details Study
4 Spanish cities, Guxens et al.
2003-2008 (2012)t
Long-term exposure
Los Angeles, CA, Becerra et
1998-2009 al. (201 3)t
Long-term exposure
Nonasthma outcomes and both individual-level and area-level SES indicators
Low
education
(80th
percentile
n = 38,681
| Total mortality n = 276,205
' natural deaths
Ages >35 yr
SES available
for 44% of
study
population
10 Italian cities, Chiusolo et
2001-2005 al. (201 1)t
Short-term exposure
7-49
-------
Table 7-17 (Continued): Epidemiologic studies evaluating socioeconomic status.
Factor
Evaluated
Low socio-
economic
position
census
block"
Lowest
quintile for
area-level
income
n = 12.5%
Lowest
deprivation
index area
n = 20%
Lowest
tertile for
area-level
SES
n = 7,556
High
neighbor-
hood level
poverty13
Reference
Category
High or
medium
socio-
economic
position
census
block"
Lowest
quintile for
area-level
income
n = 24.7%
Highest
deprivation
index area
n = 20%
Highest
tertile for
area-level
SES
n = 7,941
Low
neighbor-
hood level
poverty13
Direction of
Effect
Modification3
t
t
1
t
Outcome
Mortality — total,
cardiovascular,
IHD, lung
cancer
Total mortality
Heart failure
Small for
gestational age
or intrauterine
growth
restriction
Low birth weight
Study
Population
n = 1,265,058
Ages >30 yr
n = 835,607
Ages 40-89 yr
n = 836,557,
Ages 40-89 yr
at baseline
n = 23,452
women/infants
n = 3,545,177
births
Study Details
Rome, Italy,
2001-2010
Long-term exposure
England
Follow-up:
2003-2007
NO2 exposure
assessed for 2002
Long-term exposure
England,
2003-2007
Long-term exposure
Perth, Western
Australia,
2000-2006
Long-term exposure
California,
1996-2006
Long-term exposure
Study
Cesaroni et
al. (201 3)t
Carey et al.
(201 3)t
Atkinson et
al. (201 3)t
Pereira et al.
(2012)t
Morello-
Frosch et al.
(201 0)t
AZ = Arizona; CA = California; IHD = ischemic heart disease; I
SES = socioeconomic status.
= myocardial infarction; NO2 = nitrogen dioxide;
aUp facing arrow indicates that the effect of NO2 is greater (e.g., larger risk of hospital admission) in the group with the factor
evaluated than in the reference group. Down facing arrow indicates that the effect of NO2 is smaller in the group with the factor
evaluated than in the reference group. A dash indicates no difference in NO2-related health effect between groups.
bSample size not reported.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
7.5.3 Race/Ethnicity
In the 2010 U.S. census, 63.7% of the U.S. population identified themselves as
non-Hispanic white; 12.6% reported their race as non-Hispanic black; and 16.3%
reported being Hispanic (Humes et al.. 2011). Race and ethnicity are complex factors that
are often closely correlated with other factors including particular genetics, diet, and SES.
Therefore, race and ethnicity may influence any potential differences in NC>2-related
health effects through both intrinsic and extrinsic mechanisms.
7-50
-------
Information characterizing racial/ethnic differences in NC>2 exposure is sparse but
suggests higher exposure among nonwhite people independent of SES. For U.S. urban
areas, the population-weighted mean annual average NO2 for nonwhites was estimated to
be 4.6 ppb (38%) higher than for whites (Clark et al.. 2014). This difference was
observed across the distribution of census block household income. However, NO2 was
estimated from a national scale land use regression model and may reflect census block
differences other than race or in combination with race.
In contrast with exposure, NC^-related health effects do not clearly differ between
nonwhite and white populations (Table 7-18). This was shown in a study of asthma ED
visits among children that specifically aimed to analyze racial/ethnic differences
(Grineski et al.. 2010). Interestingly, there was a difference between Hispanic and white
children who had no health insurance (Section 7.5.2). as well as larger risk of
NO2-related asthma ED visit for black children compared to Hispanic children. Racial
and ethnic differences in NCh-related health effects also are not consistently found for
birth outcomes, although the implications of these findings are weak because an
independent relationship between NO2 exposure and birth outcomes is not certain. Some
studies estimated larger effects on birth weight or gestational age among babies of black
or Hispanic mothers (Rich et al.. 2009; Bell et al.. 2007); whereas others estimated larger
effects for babies of white mothers (Morello-Frosch et al.. 2010). or no difference among
races (Darrow et al.. 20lib: Madsenet al.. 2010).
There is some indication that NC>2 exposure may be higher among nonwhite compared to
white populations, but information on NC>2 exposure at the individual level is lacking.
NO2-related health effects do not consistently differ among racial and ethnic groups,
particularly, for asthma exacerbation, which is concluded to have an independent
relationship with short-term NC>2 exposure (Section 5.2.9). Additionally, it is unclear
whether higher NC>2 exposure in combination with higher prevalence of potential at-risk
factors impact the health of nonwhite populations (Section 7.5.2). Overall, the evidence
for potential differences in the risk of NC^-related health effects by race and ethnicity is
inconsistent and largely based on birth outcomes, for which an independent relationship
with NCh exposure is uncertain. Therefore, the evidence is inadequate to determine
whether race or ethnicity increases the risk for NC>2-related health effects.
7-51
-------
Table 7-18 Epidemiologic studies evaluating race/ethnicity.
Factor Evaluated
Reference Direction of Effect
Category Modification3
Outcome
Study
Population
Study
Details Study
Asthma-related outcomes and short-term exposure
Black race
n=635
Black race
n=635
Hispanic race
n = 1,454
White race _
n = 2,227
Hispanic *
race '
n = 1,454
White race _
n = 2,227
Asthma
hospital
- admission
n= 4,316
asthma
admissions
Ages <14 yr
Phoenix, Grineski
AZ, et al.
2001-2003 (201 0)t
Nonasthma outcomes and long-term exposure
Black maternal
race
n = 10.7%
Hispanic maternal
race
n = 14.3%
Non-Hispanic
black maternal
race
n = 40.5%
Non-Western
ethnicity
n = 24.3%
Hispanic maternal
race
n = 51.5%
Non-Hispanic
black maternal
race
n = 5.8%
White *
maternal '
race
n = 83.4%
White
maternal
race
n — A ^ on/
Western _
ethnicity
n = 75.7%
Non-Hispanic i
white ^
maternal
race
n = 32.2% i
Birth weight
decrement
Birth weight
decrement
Birth weight
decrement
Birth weight
decrement
n = 358,504
births
n = 406,627
full-term,
singleton births
n = 25,229
full-term,
singleton births
n = 3,545,177
singleton births,
37-44 week
gestation
CT; MA Bell et al.
1999-2002 (2007)
Atlanta, GA, Darrow et
1994-2004 al.
(2011 b)t
Oslo, Madsen et
Norway al.
(201 0)t
California, Morello-
1996-2006 Froschet
al.
(201 OVr
7-52
-------
Table 7-18 (Continued): Epidemiologic studies evaluating race/ethnicity.
Factor Evaluated
Hispanic maternal
race
n = 31%
Reference Direction of Effect
Category Modification3 Outcome
White or
African-
American
maternal
race
n = 69%
* Very small for
' gestational
age
Study
Population
n = 178,198
singleton births,
37-42 week
gestation,
birth weight
>500 g
Study
Details
New
Jersey,
1999-2003
Study
Rich et al.
(2009)t
AZ = Arizona; CT = Connecticut; MA = Massachusetts.
aUp facing arrow indicates that the effect of nitrogen dioxide (NO2) is greater (e.g., larger risk of hospital admission, larger
decrement in birth weight) in the group with the factor evaluated than in the reference group. Down facing arrow indicates that the
effect of NO2 is smaller in the group with the factor evaluated than in the reference group. A dash indicates no difference in
NO2-related health effect between groups.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
7.5.4
Sex
A vast number of health conditions and diseases have been shown to differ by sex, with
some indication that there may be differences by sex in the relationship between air
pollution and health effects. The 2010 U.S. census indicates an approximately equal
distribution of males and females in the U.S.: 49.2% male and 50.8% female (Howden
and Meyer. 2011). However, the distribution varies by age with a greater prevalence of
females above 65 years of age compared to males. Thus, the public health implications of
potential sex-based differences in air pollution-related health effects may vary among age
groups within the population.
Comparisons of NO2 internal dose are lacking, and limited evidence from a large
(n = 1,634) multi-European country study indicates no difference in NC>2 exposure
between males and females as shown by similar 2-week average residential outdoor NO2
concentrations (Sunver et al., 2006). With respect to NCh-related health effects, studies of
asthma exacerbation and asthma development do not show that associations with
short-term and long-term NC>2 exposure are consistently higher for either male or female
children (Table 7-19). Evidence equally points to increased risk for females (Clark et al..
2010; Kim et al.. 2004; Lin et al.. 2004b). increased risk for males (Carlsten etal. 20 lie:
Mannetal.. 2010). and no difference between sexes (Sarnat et al.. 2012; Liu et al..
2009b). Few studies indicated that comparisons were planned a priori (Mannetal.. 2010;
Lin etal.. 2004b). However, because the proportion of sexes tended to be approximately
equal, many studies had fairly large numbers of male and females to compare.
Differences between males and females also were inconsistent in studies that estimated
NO2 exposure at subjects' schools (Sarnat et al.. 2012) or at homes using land use
7-53
-------
regression models that were shown to predict ambient NO2 concentration in the study
areas well ITCarlsten et al.. 20lie: Clark etal.. 2010): Section 6.2.2.11.
Inconsistent evidence for differences by sex is also observed for associations of
short-term and long-term NC>2 exposures with respiratory infections, bronchitis, or
respiratory symptoms in children (Zemek et al.. 2010: Lin et al.. 2005). In contrast,
associations of long-term NC>2 exposure with lung function decrements (Rosenlund et al..
2009b: Oftedal et al.. 2008: Roias-Martinez et al.. 2007a: Peters etal.. 1999) are
consistently larger in female than male children. This effect measure modification is not
explained by lower baseline lung function in females. While many studies measured NC>2
concentrations near (2 km) subjects' schools or estimated concentrations at or near homes
using well-validated models [(Rosenlund et al.. 2009b: Oftedal et al.. 2008: Rojas-
Martinez et al.. 2007a): Section 6.2.5.11. there is uncertainty about potential confounding
by other traffic-related pollutants (Section 6.2.9). Thus, the extent to which the larger
NO2-related decreases in lung function among females reflect an independent effect of
NO2 exposure is unclear.
Beyond respiratory effects, the majority of studies observed no difference between males
and females in associations of long-term NO2 exposure with an array of cardiovascular
effects, diabetes, total mortality, cause-specific mortality, or lung cancer incidence as
described in Table 7-19 (Beelen et al.. 2014b: Eze etal.. 2014: Atkinson etal.. 2013:
Cesaroni et al.. 2013: Dong et al.. 2013a: Johnson et al.. 2013: Andersen et al.. 2012c:
Raaschou-Nielsen et al.. 2012: Raaschou-Nielsen et al.. 201 Ib: Zhang et al.. 2011:
Raaschou-Nielsen et al.. 2010a: Yorifuji etal.. 2010: Rosenlund et al.. 2009a: Abbey et
al.. 1999). In most cases, no difference between males and females was observed for NO2
associations with subclinical effects such as changes in blood pressure, atherosclerosis,
HRV, systemic inflammation, or insulin resistance (Bilenko et al.. 2015: Foraster et al..
2014: Atkinson etal.. 2013: Dong etal.. 2013b: Rivera etal.. 2013: Thiering etal.. 2013:
Lenters etal.. 2010: Panasevich et al.. 2009: Felber Dietrich et al.. 2008). In the relatively
small group of studies that found differences between males and females, most observed
greater risk among females for associations of short-term NO2 exposure with cardiac
arrest or mortality (Wichmann etal.. 2013: Cakmak et al.. 201 Ib: Kan et al.. 2008).
Similar to the evidence for lung function, there is uncertainty as to whether NC>2 exposure
has an effect on cardiovascular outcomes or mortality independent of other traffic-related
pollutants (Sections 6.3.9 and 6.5.3). Thus, the extent to which the greater risk for
females or lack of modification of NCh-related cardiovascular effects or mortality by sex
can be attributable to NO2 versus correlated copollutants is not clear.
The collective body of evidence does not clearly indicate thatNCh exposure or
NO2-related health effects differ between males and females. Differences between males
7-54
-------
and females are not consistently observed for associations of short-term and long-term
NO2 exposure with asthma exacerbation and asthma development, the health effects for
which evidence most strongly indicates independent relationships with NC>2 exposure
(Sections 5.2.9 and 6.2.9). Nonrespiratory health effects related to long-term NCh
exposure mostly do not differ between males and females, but lung function decrements
related to long-term NCh exposure and mortality and cardiovascular effects related to
short-term NC>2 exposure are increased consistently among females. Because it is
uncertain whether NC>2 exposure has an independent effect on these health outcomes, the
evidence is suggestive that females are at increased risk for NCh-related health effects.
Table 7-19 Epidemiologic studies evaluating sex.
Factor
Evaluated
Reference
Category
Asthma-related outcomes and
Female
n = 1,454
Female
n=20
Female
n=68
Female
n = 43.5%
Male
n = 2,368
Male
n = 38
Male
n = 114
Male
n = 56.5%
Direction of
Effect
Modification3
Outcome
Study
Population
Study
Details
Study
short-term exposure
t
^
^™
1
Asthma hospital
admission
Pulmonary
inflammation
Lung function
decrement,
Pulmonary
inflammation
Wheeze
n = 3,822
admissions
Ages 6-12 yr
n = 58 children
with asthma
Ages 6-12 yr
n = 182
children with
asthma
Ages 9-14 yr
n = 315
children with
asthma
Ages 6-11 yr
Vancouver,
Canada,
1987-1998
Ciudad
Juarez,
Mexico and
El Paso, TX
Windsor,
Canada,
2005
Fresno, CA,
2000-2005
Lin et al.
(2004b)
Sarnat et al.
(2012)t
Liu et al.
(2009b)t
Mann et al.
(201 0)t
7-55
-------
Table 7-19 (Continued): Epidemiologic studies evaluating sex.
Factor
Evaluated
Reference
Category
Asthma-related outcomes
Female
n = 52.6%
Female
n = 89
Female
n = 7,560
Nonasthma
Female
n= 2,137
Female
n= 2,784
Female
n = 8,055
Female
n=24
Female
n = 1,846
Female
n = 51.9%
Femaleb
Male
n = 47.4%
Male
n = 97
Male
n = 13,332
outcomes and
Male
n = 2,077
Male
n = 3,998
Male
n = 6,472
Male
n = 16
Male
n = 2,811
Male
n = 48.1%
Maleb
Direction of
Effect
Modification3
Outcome
Study
Population
Study
Details
Study
and long-term exposure
t
1
t
Asthma
prevalence and
incidence
Asthma
incidence
Asthma
incidence
n = 1,109
children
Grades 3-5
n = 184
children
followed from
birth to age
7yr
n= 20,892
children
followed from
birth to age
5 yr
San
Francisco,
CA, 2001
Vancouver,
Canada
Southwestern
British
Columbia,
Canada
Kim et al.
(2004)
Carlsten et al.
(2011c)t
Clark et al.
(201 0)t
short-term exposure
t
^™
t
t
t
1
t
Respiratory
hospital
admission
Respiratory
infection hospital
admission
Otitis media ED
visits
HRV
Out-of-hospital
cardiac arrest
Total mortality
Total mortality
n= 4,214
respiratory
admissions
n= 6,782
admissions in
children
Ages 0-14 yr
n = 14,527 ED
visits
Ages 1-3 yr
n=40
nonsmoking
adults with
CVD
Mean age
65.6 yr
n = 4,657
events
n = 276,205
natural deaths
Ages >35 yr
n = 7.29-15.8
mean daily
deaths across
locations
Windsor,
Canada,
1995-2000
Toronto,
Canada,
1998-2001
Edmonton,
Canada,
1992-2002
Beijing,
China,
summer 2007
and summer
2008
Copenhagen,
Denmark,
2000-2010
10 cities,
Italy,
2001-2005
Santiago
Province,
Chile
(7 urban
centers),
1997-2007
Luqinaah et
al. (2005)
Lin et al.
(2005)
Zemek et al.
(201 0)t
Huanq et al.
(2012a)t
Wichmann et
al. (201 3)t
Chiusolo et al.
(201 1)t
Cakmak et al.
(2011b)t
7-56
-------
Table 7-19 (Continued): Epidemiologic studies evaluating sex.
Factor
Evaluated
Female
Mean daily
deaths =
56.5
Nonasthma
Female
n = 49%
Female
n =
942-1,161
Female
n=648
Femaleb
Female
n = 52.6%
Female
n = 832
Female
n = 49.4%
Female
n = 508
Female
n = 395
Female
n = 53.6%
Female
n = 1,980
Direction of
Reference Effect
Category Modification3
Male *
Mean daily
deaths = 62. 5%
outcomes and long-term exposure
Male *
n = 51%
Male *
n — oyu i , IDU
Male +•
n = 711
Maleb *
Male _
n = 47.4%
Male _
n = 924
Male _
n = 50.6%
Male
n = 1,028
Male
n = 350
Male
n = 46.4%
Male
n = 1,720
Outcome
Total mortality
Lung function
decrement
Lung
development
decrement
Lung function
decrement
Lung function
decrement
Bronchitis
Respiratory
symptoms
Ml
Blood IL-6 level
Mean carotid
artery intima
media thickness
Atherosclerosis
(carotid intima
media thickness)
Systolic/diastolic
blood pressure
Study
Population
n = 173,911
deaths
n= 2,307
Ages 9-10 yr
n = 3,170
healthy
children
Age 8 yr
n = 1,760
Ages 9-14 yr
n = 3,293
Grades 4, 7, 10
n = 1,109
Grades 3-5
n = 1,756
full-term infants
Assessed at
ages 1 and 2 yr
n= 43,275
cases,
511,065 control
n = 1,536
Ages 45-70 yr
n = 745
Ages 26-30 yr
n= 2,780
Median age
58 yr
n = 3,700
Ages 35-83 yr
Study
Details
Shanghai,
China,
2001-2004
Oslo,
Norway,
2001-2002
Mexico City,
Mexico,
1996-1999
Rome, Italy,
2000-2001
Southern
California,
1986-1990
San
Francisco,
CA, 2001
3 German
cities,
1995-1999
Stockholm
county,
Sweden,
1985-1996
Stockholm
county,
Sweden,
1992-1994
Utrecht, the
Netherlands,
1999-2000
Girona
Province,
Spain,
2007-2010
Girona, Spain
Study
Kan et al.
(2008)t
Oftedal et al.
(2008)
Rojas-
Martinez et al.
(2007a)
Rosenlund et
al. (2009b)t
Peters et al.
(1999)
Kim et al.
(2004)
Gehrinq et al.
(2002)
Rosenlund et
al. (2009a)t
Panasevich et
al. (2009)t
Lenters et al.
(201 0)t
Rivera et al.
(201 3)t
Foraster et al.
(2014)t
7-57
-------
Table 7-19 (Continued): Epidemiologic studies evaluating sex.
Factor
Evaluated
Female
n= 431, 388
Female
n = 573
Female
n = 12,184
Female
n = 12,184
Female
n = 829
Female
n=63
Female
n = 18,085
Female
n = 725
Female with
CVD
n = 115
Female
without CVD
n =610
Female
n= 27,273
Female
n = 51.3%
Reference
Category
Male
n = 405, 169
Male
n = 574
Male
n = 12,661
Male
n = 12,661
Male
n = 1,155
Male
n = 79
Male
n = 29,030
Male
n = 683
Male with CVD
n = 121
Male
without CVD
n = 562
Male
n = 24,545
Male
n = 48.7%
Direction of
Effect
Modification3 Outcome
_ Heart failure
_ Systolic/diastolic
blood pressure
_ Incident CVD
_ Incident stroke
_ Incidence
hypertension
_ Absolute
increase in
arterial blood
pressure
_ Incident stroke
_ Fatal stroke
_ Stroke
_ HRV decrement
(SDNN)
—
t
Diabetes
_ Diabetes
Study
Population
n = 836,557
Ages 40-89 at
baseline
n = 1,147
Age 12 yr
n = 24,845
Mean age
41.7yr
n = 24,845
Mean age
"45.59 yr
n = 1,984
Ages 50-65 yr
at baseline
n = 142
Ages 50-65 yr
at baseline
n = 4,696
cases and
37,723 controls
Ages >20 yr
n = 1,408
Ages >50 yr
n = 51,818
Ages 50-65 yr
at baseline
n= 6,392
Ages 29-73 yr
Study
Details
England,
2003-2007
the
Netherlands
Shenyang,
Anshan and
Jinzhou,
China, 2009
Shenyang,
Anshan and
Jinzhou,
China,
2009-2010
Copenhagen,
Aarhus
counties,
. Denmark,
1993-2006
Edmonton,
Alberta,
Canada,
2007-2009
Switzerland
Follow-up:
1991 to
2001-2003
Copenhagen,
Aarhus
counties,
Denmark,
1993-2006
Switzerland,
2002
Study
Atkinson et al.
(201 3)t
Bilenko et al.
(201 5)t
Donq et al.
(2013a)t
Donq et al.
(2013b)t
Andersen et
al. (2012b)t
Johnson et al.
(201 3)t
Felber
Dietrich et al.
(2008)t
Andersen et
al. (2012c)t
Eze et al.
(2014VT
7-58
-------
Table 7-19 (Continued): Epidemiologic studies evaluating sex.
Factor
Evaluated
Female
n=222
Female
n =
47.7-77.5%
across
cohorts
Female
n = 51.47%
Female
n = 63%
Female
n = 52.5%
Female
n= 430,891
Femaleb
Female
n = 111
deaths
Reference
Category
Male
n = 175
Male
n =
22.5-52.3%across
cohorts
Male
n = 48.53%
Male
n = 37%
Male
n = 47.5%
Male
n = 404,716
Maleb
Male
n = 407 deaths
Direction of
Effect
Modification3 Outcome
_ Insulin
resistance
Respiratory
mortality
CVD mortality
CVD mortality
- Diabetes-related
mortality
j Total mortality
— Mortality
(total, CVD, IHD,
lung cancer)
Luno cancer
mortality
t Respiratory
mortality
Study
Population
n = 397
Age 10 yr
n = 307,553
Mean age
across
16 cohorts
41. 9-73.0 yr at
baseline
n = 9,941, 256
deaths
Ages
35-1 03 yr
n = 367,383
Mean age
41. 1-70.3 yr
across
22 cohorts
n = 52,061
Ages 50-64 yr
n = 835,607
deaths
Ages 40-89 yr
n = 1,265,058
Ages >30 yr
n = 63,520
Ages >40 yr
Study
Details
Munich and
Wesel,
Germany
Europe
Follow-up:
1985-2007
NO2
exposure
assessed for
2008-2011
Shenyang,
China
Follow-up:
1998-2009
NO2
exposure
assessed for
1998-2009
Europe
Follow-up:
1985-2007
NO2
exposure
assessed for
2008-2011
Denmark
Follow-up:
1971-2009
NO2
exposure
assessed for
1971-2009
England
Follow-up:
2003-2007
NO2
exposure
assessed for
2002
Rome, Italy,
2001-2010
3 prefectures,
Japan,
-1 QQO -1 QQC
Study
Thierinq et al.
(201 3)t
Dimakopoulou
etal. (2014)t
Zhanq et al.
(201 1)t
Beelen et al.
(2014b)t
Raaschou-
Nielsen et al.
(2012)t
Carey et al.
(201 3)t
Cesaroni et
al. (2013)t
Katanoda et
al. (201 1)t
7-59
-------
Table 7-19 (Continued): Epidemiologic studies evaluating sex.
Factor
Evaluated
Female
n = 203-
7,840 deaths
Female
n= 4,060
Female
n = 49%
Female
n = 1,206
Female
n= 27,788
Reference
Category
Male
n = 233- 4,531
deaths
Male
n = 2,278
Male
n = 51%
Male
n = 2,275
Male
n = 25,182
Direction of
Effect
Modification3 Outcome
t Lung cancer
mortality
CVD mortality
I COPD mortality
t Lung cancer
mortality
- Total mortality
- Cardiopulmonary
mortality
— Respiratory
mortality
- Lung cancer
mortality
incidence
Luno cancer
incidence
Study
Population
n = 138,977,
518 deaths
Ages 51-90 yr
n = 6,338
nonsmoking,
non-Hispanic
adults
Ages 27-95 yr
n = 13,444
Ages >65 yr
n = 3,481
Ages 20-93 yr
at enrollment
n = 52,970
Ages 50-64 yr
Study
Details
Oslo,
Norway,
-1 QQO -1 QQQ
California,
Follow-up:
1977-1992
NO2
exposure
assessed for
1973-1992
Shizuoka,
Japan,
1999-2006
Copenhagen,
Aarhus
counties,
Denmark,
1970-1997
Copenhagen,
Aarhus
counties,
Denmark,
1993-1997
Study
Naess et al.
(2007)
Abbey et al.
(1999)
Yorifuji et al.
(201 0)t
Raaschou-
Nielsen et al.
(2010a)t
Raaschou-
Nielsen et al.
(2011b)t
CA = California; COPD = chronic obstructive pulmonary disease; CVD = cardiovascular disease; ED = emergency department;
HRV = heart rate variability; IHD = ischemic heart disease; IL = interleukin; Ml = myocardial infarction; NO2 = nitrogen dioxide.
aUp facing arrow indicates that the effect of NO2 is greater (e.g., larger risk of hospital admission, larger decrement in HRV) in the
group with the factor evaluated than in the reference group. Down facing arrow indicates that the effect of NO2 is smaller in the
group with the factor evaluated than in the reference group. A dash indicates no difference in NO2-related health effect between
groups.
bSample size not reported.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
7-60
-------
7.5.5 Residence in Urban Areas
A majority (81%) of the U.S. population lives in urban areas, and U.S. census data
indicate that the urban population grew 12% from 2000 to 2010. Higher ambient NC>2
concentrations in urban than suburban areas and the large numbers of people potentially
having higher exposures highlight the public health impact of potential differences in
NC>2-related health effects in urban residents. Higher ambient NO2 concentrations have
been described for downtown versus suburban areas [14.9 ppb vs. 11.7 ppb; (Rotko et al..
2001)1. Higher ambient NC>2 concentrations also were related to building characteristics
such as high-rise building versus single family home and older versus newer construction
(before or after 1970). Proximity to roads has been shown to be a determinant of personal
NO2 exposure (Section 7.5.6). and the higher road density in urban areas and proximity to
major roads also may result in higher exposure of urban residents. The topography of
urban communities also may contribute to higher NCh exposure among residents because
the presence of street canyons enhances mixing at elevations closer to the street
canyon-urban boundary layer interface, resulting in higher NCh concentrations at lower
elevations (Section 2.5.3). This may have implications for higher NC>2 exposures for
pedestrians, outdoor workers, and those living on lower floors of buildings. Multiple lines
of evidence indicate that residing in urban areas may lead to increased exposure to NC>2.
Although the potential for higher exposure of urban residents to NC>2 is well
characterized, epidemiologic comparisons of NO2-related health effects between urban
and nonurban residents are limited and use variable definitions of urban and nonurban
residence (Table 7-20). The single study examining asthma exacerbation, which specified
comparisons a priori, did not observe that associations of short-term increases in NC>2
with changes in lung function differed between urban and suburban children with asthma
(Ranzi et al.. 2004). Associations of NO2 exposure with nonasthma outcomes did not
differ by urban residence (Atkinson et al., 2013; Steerenberg et al., 2001). However, there
is uncertainty regarding the extent to which NC>2 exposure is independently related to
respiratory effects in healthy populations (Section 5.2.9) or to cardiovascular effects
(Section 6.3.9).
Evidence indicates the potential for higher ambient NO2 exposure among urban residents,
but the limited epidemiologic evidence does not provide a strong basis for inferring
whether urban residence leads to increased risk for NO2-related health effects. Overall,
the limited epidemiologic evidence on urban residence is inconsistent and is based on
variable definitions of urban and nonurban residence and health effects not conclusively
linked to NC>2 exposure. As a result, the evidence is inadequate to determine whether
residence in urban areas increases the risk for NCh-related health effects.
7-61
-------
Table 7-20 Epidemiologic studies evaluating urban residence.
Factor
Evaluated
Reference
Category
Direction of
Effect
Modification3
Outcome
Study
Population
Study Details
Study
Asthma-related outcomes and short-term exposure
Urban-
industrial
residence
n=67
Rural
residence
n = 51
_ Lung function
variability
n = 118 children
with asthma or
respiratory
symptoms
Ages 6-11 yr
Urban & rural areas
Emilia-Romagna,
Italy,
1999
Ranzi et al.
(2004)
Nonasthma outcomes and short-term and long-term exposure
Urban
residence
n = 38
Residence
in London
n = 91,992
Suburban
residence
n = 44
Residence
in
North/South
UK
(excluding
London)
n = 744,565
-j- Lung function n = 82
' decrement Ages 8-1 Syr
_ Heart failure n = 836,557
Ages 40-89 yr at
baseline
Utrecht, Bilthoven,
the Netherlands
Short-term
exposure
UK,
2003-2007
Long-term exposure
Steerenberq
etal. (2001)
Atkinson et al.
(201 3)t
UK= United Kingdom.
aUp facing arrow indicates that the effect of nitrogen dioxide (NO2) is greater (e.g., larger decrement in lung function) in the group
with the factor evaluated than in the reference group. Down facing arrow indicates that the effect of NO2 is smaller (e.g., smaller
decrement in lung function) in the group with the factor evaluated than in the reference group. A dash indicates no difference in
NO2-related health effect between groups.
fStudy published since the 2008 ISA for Oxides of Nitrogen
7.5.6 Proximity to Roadways
A zone of elevated NO2 concentrations (on average 30% and up to 100%) typically
extends to within 200-500 m of roads with heavy traffic (Section 2.5.3). Thus,
individuals spending a substantial amount of time on or near high-traffic roadways,
including those living or working near highways and commuting on roadways, are likely
to be exposed to elevated NO2 concentrations.
Large proportions of the U.S. population potentially have elevated NO2 exposures as a
result of living within 200 to 500 m of roadways. Seventeen percent of U.S. homes are
located within 91m of a highway with four or more lanes, a railroad, or an airport (U.S.
Census Bureau. 2009). Specific to road traffic, Rowangould (2013) found that over 19%
of the U.S. population lives within 100 m of roads with an annual average daily traffic
7-62
-------
(AADT) of 25,000 vehicles, and 1.3% lives near roads with AADT greater than 200,000.
The proportion is much larger in certain parts of the country, mostly coinciding with
urban areas. For example in Los Angeles, CA, primary and secondary roads run through
neighborhoods with population density as high as 15,000-18,000 people per km2 [(U.S.
Census Bureau. 2014. 2013a); Figure 7-11. Among California residents, 40% lives within
100 m of roads with AADT of 25,000 (Rowangould. 2013).
Though far from generalizable across populations, there are examples indicating that
residence near a busy roadway may be associated with higher NO2 exposure. In the
southern California CHS cohort, closer proximity to a freeway showed a range of
correlations with residential NC>2 measurements, with values of-0.73 to -0.90 in some
communities (Gauderman et al. 2005). In a cohort of pregnant women who spent on
average 60% of time home indoors, traffic intensity within 250 and 500 m of homes was
moderately correlated with personal NO2 exposures [r = 0.3, 0.4, respectively;
(Schembari et al., 2013)1. However, the strongest correlation was not observed for traffic
intensity in closest proximity to homes (r = 0.2 for traffic intensity within a 100-m
buffer). Such results may be explained by the atmospheric chemistry of NC>2. Depending
on atmospheric stability, NC>2 concentrations can dilute with distance from the road or
extend beyond 1 km of the roadway (Section 2.5.3) and may be higher after some of the
NO reacts photochemically to become NC>2.
7-63
-------
0 4,250 8,500
17.000
Population Density per Sq Kilometer
j^H 110-4.600
I 4,700-9,100
| | 9.200 - 14,000
| | 15,000-18,000
| | 19,000-23.000
HH 24,000 - 27,000
Primary and Secondary Roads
J Los Angeles city boundry
A
N
Source: National Center for Environmental Assessment analysis of U.S. census data (U.S. Census Bureau, 2014, 2013a).
Figure 7-1 Map of population density in Los Angeles, CA, in relation to
primary and secondary roads.
Exposure in transport is found to be an important determinant of total personal NC>2
exposure (Son et al., 2004; Lee et al. 2000). although time in transport makes up a
relatively small proportion of people's activities. Such findings have implications for
people who commute on roadways as well as for professional drivers. Among the
populace working outside the home, 15.6% spend 45 minutes or more commuting to
7-64
-------
work each day (U.S. Census Bureau. 2007). Average one-way commuting times for the
U.S. labor force working outside the home are 19.3 minutes for bicyclists, 11.5 minutes
for walkers, and 25.9 minutes for all other modes of transportation. The potential of
higher NO2 exposure of people who commute on roadways and professional drivers is
supported by many observations of in-vehicle NC>2 concentrations approaching roadside
concentrations (Figure 3-2) and some evidence of higher personal NC>2 exposure during
transport than in outdoor or indoor environments (Delgado-Saborit 2012). The
relationship of health effects with NO2 exposure during commute or while driving for
work is not well characterized. In the CHS cohort, longer commuting time to school was
associated with increased wheeze but not asthma onset; commute-time NO2 exposures
were not measured (McConnell et al.. 2010b).
Children are characterized to be at increased risk for NCh-related health effects
(Section 7.5.1.1). and time spent near major roads could potentially be a source of higher
NC>2 exposure contributing to health effects. Attendance at schools or day care near major
roads may be an important determinant of NC>2 exposure, and ambient NO2
concentrations at schools have been associated with respiratory effects in children with
asthma (Sections 5.2.2.2 and 5.2.2.5). Seven percent of U.S. schools serving
3,152,000 school children are located within 100 m of a major roadway, and 15% of U.S.
schools serving 6,357,000 school children are located within 250 m of a major roadway
[not specifically defined in terms of AADT, number of lanes, or other criteria; (Kingsley
et al.. 2014)1. In California, 2.3% of public schools serving 150,323 children were
estimated to be located within 150m of high-traffic roads [>5 0,000 vehicles per day;
(Green et al.. 2004)1. Also in California, 1,534 daycare facilities serving 57,173 (7% of
those in daycare) children were within 200 m of roadways with AADT of >50,000, and
4,479 facilities serving 171,818 (21%) children were within 200 m of roadways with
AADT of 25,000-49,999 (Houston et al.. 2006). Though neither of these analyses
assessed NCh exposures, they identify the large numbers of children potentially exposed
to higher NC>2 concentrations in locations where they spend several hours per day.
There is some indication that traffic exposures differ among sociodemographic groups. In
California, schools or daycare in close proximity to high-traffic roadways had a higher
percentage of nonwhite students (Green et al.. 2004) or tended to be located in areas with
higher percentages of nonwhite residents (Houston et al., 2006). Analyses of U.S. census
blocks or tracts indicate associations of higher traffic or road density or proximity to
roadways with higher proportion of nonwhite residents (Rowangould. 2013; Tian et al..
2013). In some (Rowangould. 2013: Green et al.. 2004) but not all (Tian etal.. 2013)
cases, closer proximity to roadways or higher traffic density was associated with lower
SES at the school or census block level. In analyses not considering proximity or density
of traffic, higher NC>2 exposures are suggested for nonwhite (Section 7.5.3) or low SES
7-65
-------
(Section 7.5.2) populations. However, it is not understood whether observations of higher
NO2 exposures in certain sociodemographic groups are related to disparities in traffic
exposure.
Large proportions of the U.S. population live or attend school near roads or travel on
roads, and some evidence indicates higher NO2 exposure with proximity to roads. Traffic
proximity may be more prevalent among nonwhite and low SES groups, but the influence
of traffic proximity on differential NC>2 exposure in these groups is unclear. While traffic
proximity (HEI. 2010) and NC>2 exposure near traffic (Section 5.2.9.3) are linked to
asthma exacerbation or prevalence, studies have not examined whether NCh-related risk
of asthma differs for populations living near traffic (Table 7-21). Closer residential
proximity to freeway was associated with larger NCh-related decrements in lung
development among children (Gauderman et al., 2007). but NO2 concentrations as
measured at central sites were weakly correlated with traffic counts near homes, and an
independent effect of NO2 exposure on lung development is uncertain. Additionally,
results are inconclusive for cardiovascular effects and leukemia (Foraster et al.. 2014;
Hartetal.. 2013; Amigou et al.. 2011). The insufficient quantity and consistency of
evidence, based on health effects for which independent relationships with NO2 exposure
are uncertain, is inadequate to determine whether populations in close proximity to
roadways are at increased risk for NCh-related health effects.
7-66
-------
Table 7-21 Epidemiologic studies
Factor
Evaluated
Nonasthma
Residence
<500 m
from
freeway
n=440
Near road
n = 539MI
cases
Moved near
road
n=48MI
cases
Moved
away from
road
n=603MI
cases
Traffic
intensity of
nearest
road
>median
Traffic load
at 500 m
>median
Residence
<500 m of
main roads
n=48
Direction of
Reference Effect
Category Modification3
evaluating proximity
Outcome
Study
Population
to roadways.
Study Details
Study
outcomes and long-term exposure
Residence *
1,000-1, 500m '
from freewayb
Far from road *
(>50 m from '
road with
>z. lanes or
>1 50m from \
highway)
n = 2,841 Ml
cases
Traffic intensity _
of nearest road
500 m from all
road types
n = 954
Lung
development
decrement
(change
overtime)
Incident Ml
Systolic
blood
pressure
Diastolic
blood
pressure
Systolic
blood
pressure
Diastolic
blood
pressure
Leukemia
n = 3,677
children
followed
ages 10-18 yr
n = 84,562
Ages 30-55 yr
at enrollment
n = 3,700
Ages 35-83 yr
n = 763 cases,
1,681 controls
Ages <15 yr
Alpine, Lake
Elsinore, Lake
Arrowhead,
Atascadero,
Lancaster, San
Dimas, Long Beach,
Mira Loma, Lompoc,
Riverside, Santa
Maria, Upland, CA
Follow-up:
1993/1 996 to
2001/2004
U.S.,
1990-2008
Girona, Spain
France,
2003-2004
Gauderman
et al. (2007)
Hartetal.
(201 3)t
Foraster et
al. (2014)t
Amiqou et al.
(201 1)t
CA = California; Ml = myocardial infarction.
aUp facing arrow indicates that the effect of nitrogen dioxide (NO2) is greater (e.g., larger decrement in development, larger risk of
Ml) in the group with the factor evaluated than in the reference group. Down facing arrow indicates that the effect of NO2 is smaller
in the group with the factor evaluated than in the reference group. A dash indicates no difference in NO2-related health effect
between groups.
bSample size not reported.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
7-67
-------
7.6 Behavioral and Other Factors
7.6.1 Diet
Diet is an important influence on health and thus, plausibly could influence air
pollutant-related health effects. The 2008 ISA for Oxides of Nitrogen (U.S. EPA. 2008c)
did not discuss whether diet influences the risk of NCh-related health effects; however,
evidence from previous experimental studies indicates reduced or greater respiratory
effects in humans and rodents with supplementation of or deficiencies in antioxidant
vitamins, respectively (Table 7-22). It is not clear that diet would affect NO2 exposure,
although the two may be linked through a common relationship with SES. However, NC>2
reactions in the epithelial lining fluid of the respiratory tract form oxidation products
(Section 4.3.2.1), and dietary antioxidants could affect the availability of such oxidation
products to initiate subsequent events in the mode of action for NCh-related health
effects. Increased risk of NCh-related health effects due to dietary deficiencies in
antioxidants could have a large public health impact. Among U.S. adults, 37.7% and
22.6% report consuming fruits and vegetables, respectively, less than once a day (CDC.
2013). which could result in lower levels of Vitamins C and E.
The strongest evidence for diet modifying NCh-related health effects comes from a
controlled human exposure study that demonstrated that healthy adults with diets
supplemented with Vitamin C had less airway responsiveness following 2,000 ppb NCh
for 1 hour compared to adults with a normal diet (Mohsenin. 1987b). Airway
responsiveness is a hallmark of asthma exacerbation (Figure 4-1). The evidence that
higher antioxidant vitamin intake reduces NC>2-induced airway responsiveness is
supported by experimental evidence in humans and rodents that higher dietary Vitamin E
and/or C reduces NCh-induced pulmonary inflammation and modulates the
oxidant/antioxidant balance [(Mohsenin. 1991; Hatch etal. 1986; Elsayed and Mustafa.
1982; Sevanian et al.. 1982b: Selgrade et al.. 1981; Ayaz and Csallany. 1978):
Table 7-22]. Pulmonary inflammation and formation of oxidation products are early
events in the mode of action for NO2 effects on asthma exacerbation (Figure 4-1 and
Section 4.3.5). Despite the consistency and coherence of evidence, findings are limited,
particularly for changes that are indicative of health effects. The changes in NO2-induced
lipid peroxidation, antioxidant levels, and antioxidant enzyme activity observed in
relation to vitamin deficiencies or supplementation may or may not lead to health effects.
7-68
-------
Epidemiologic studies have not examined whether diet modifies NCh-related respiratory
effects. Limited information indicates that associations of long-term NCh exposure with
mental development in infants are larger in groups with low fruit intake (maternal
prenatal or concurrent, respectively) than groups with high fruit intake [(Guxens et al..
2012); Table 7-231. Fruits are a source of antioxidants; thus, the results for modification
by fruit intake are consistent with those for dietary antioxidant vitamins. However,
because evidence for NO2-related neurodevelopmental effects is overall inconclusive
(Section 6.4.5). the available epidemiologic evidence cannot adequately inform whether
diet deficiencies increase the risk for NCh-related health effects.
Experimental studies in humans and animals provide evidence that dietary intake of
Vitamin C or E modifies airway responsiveness, pulmonary inflammation, and oxidant
balance following NO2 exposure, with high vitamin intake reducing these effects and low
intake increasing effects. Epidemiologic evidence is available only for health effects for
which relationships with NC>2 are uncertain. Oxidative stress, pulmonary inflammation,
and airway responsiveness are key events in the mode of action for asthma exacerbation
(Figure 4-1); thus, a biologically plausible mechanism exists for dietary antioxidants to
reduce the risk of NCh-related health effects. Despite this biological plausibility, most
findings are for changes in oxidant/antioxidant balance rather than changes clearly
indicative of health effects, such as airway responsiveness. Thus, the findings for dietary
deficiencies may further support a role for oxidative stress in the biological pathways that
mediate the effects of NO2 exposure on asthma exacerbation. Therefore, there is
suggestive evidence that low dietary antioxidant intake increases the risk for NCh-related
health effects.
7-69
-------
Table 7-22 Controlled human exposure and toxicological studies evaluating
diet.
Factor Evaluated
Direction of
Reference Effect
Category Modification3
Outcome
Study
Population/
Animal Model
Study
Details
Study
Asthma-related outcomes and short-term exposure
Vitamin C
supplemented
diet 3 days
n = 11
Normal diet
n = 11
1
Airway
responsive-
ness
Humans
n = 8 male,
3 female
Ages 18-37yr
2,000 ppb
NO2 for 1 h,
randomized,
double-blind
Mohsenin
(1987b)
Nonspecific outcomes and short-term exposure
Vitamin C and E
supplemented
diet 4 weeks
n = 10
Vitamin C
supplemented
diet
n=4-5
Vitamin C
deficient diet
n = 5-8
Vitamin E
deficient diet,
birth-adolescence
n=6-7
Vitamin E
deficient diet,
birth-adolescence
n=6
Normal diet
n = 9
Vitamin C
normal diet
n = 9
Vitamin C
supplemented
diet
n = 15
Vitamin E
supplemented
diet
n=6-8
Vitamin E
supplemented
diet
n=6
Nonspecific outcomes and long-term
Vitamin E
deficient diet
n=6-10
Vitamin E
supplemented
diet
n =6-10
1
t
t
t
t
exposure
t
Lipid
peroxidation
in lavage fluid
Pulmonary
inflammation
Pulmonary
inflammation,
Antioxidant
reduction
Lipid
peroxidation,
Pulmonary
inflammation
Lipid
peroxidation,
Induction of
antioxidant
enzymes
Glutathione
peroxidase
activity
reduction
Humans
n = 10 male,
9 female
Ages 21 -33 yr
Guinea pigs
(Hartley)
n = 2-6
males/group
Guinea pigs
(Hartley)
n = 3-15
males/group
Rats (Sprague-
Dawley)
n = 6-8/group
Rats (Sprague-
Dawley)
n = 6/group
Mice
(C57BL/6J)
n = 120 females
4,000 ppb
NO2 for 3 h
400, 1,000,
3,000, or
5,000 ppb
NO2 for
3 days
4,800 ppb
NO2 for 3 h
3,000 ppb
NO2 for
1 week
3,000 ppb
NO2 for
1 week
500 or
1,000 ppb
NO2 for
17 mo
Mohsenin
(1991)
Belgrade et
al. (1981)
Hatch et al.
(1986)
Sevanian et
al. (1982b)
Elsaved and
Mustafa
(1982)
Avaz and
Csallany
(1978)
NO2 = nitrogen dioxide.
aUp facing arrow indicates that the effect of NO2 is greater (e.g., larger increase in airway responsiveness, larger increase in lipid
peroxidation) in the group with the factor evaluated than in the reference group.
7-70
-------
Table 7-23 Epidemiologic studies evaluating diet.
Factor
Evaluated
Nonasthma
Low
healthy
eating
index
(<109)b
Low
maternal
fruit intake
in 1st
trimester
(<405 g per
day)
n = 33.5%
Reference
Category
outcomes and
High healthy
eating index
(>109)b
Medium/
high
maternal
fruit intake in
1st trimester
(405 g per
day)
n = 66.5%
Direction of
Effect
Modification3 Outcome
long-term exposure
_ Ml incidence
-j- Decrement in
' mental
development
score in infants
at age 14 mo
Study
Population
n = 84,562
Ages 30-55 yr at
enrollment
n = 1,889
children followed
from prenatal
period
Study Details Study
U.S., Hartetal.
1990-2008 (201 3)t
Valencia, Guxens et al.
Sabadell, (2012)t
Gipuzkoa, and
Asturias, Spain
2003-2008
Ml = myocardial infarction.
aUp facing arrow indicates that the effect of nitrogen dioxide (NO2) is greater (e.g., larger decrement in mental development score)
in the group with the factor evaluated than in the reference group. A dash indicates no difference in NO2-related health effect
between groups.
bSample size not reported.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
7.6.2 Smoking
In the 2012 National Health Interview Survey, 18.1% of U.S. adults reported being
current smokers, and 21.3% reported being a former smoker (Blackwell et al.. 2014).
Smoking is a well-documented risk factor for many diseases, but it is unclear whether
smoking exacerbates health effects associated with air pollutant exposures, including
NC>2. Information on NC>2 exposure or internal dose differences between smokers and
nonsmokers also is lacking.
Although many controlled human exposure studies report smoking status, comparisons
between smokers and nonsmokers are infrequent due to small sample size. In addition to
being limited, the available evidence is based on nonasthma outcomes. Among healthy
adults, NC>2 exposure induced a larger decrement in lung function in smokers than in
nonsmokers [(Morrow et al., 1992); Table 7-241. There is a lack of epidemiologic studies
to draw direct coherence with this experimental evidence for respiratory effects. As
examined primarily for cardiovascular or diabetes morbidity and mortality, most
associations with long-term NC>2 do not differ between smokers and nonsmokers
7-71
-------
(Dadvandetal.. 2014b; Atkinson etal.. 2013; Hart etal.. 2013; Rivera et al.. 2013;
Andersen etal.. 2012b; Zhang etal.. 2011; Lenters etal.. 2010; Panasevich et al.. 2009)
or are larger among nonsmokers [(Carey et al.. 2013; Andersen et al.. 2012c; Raaschou-
Nielsen et al.. 2012; Maheswaran et al.. 2010); Table 7-25]. A similar lack of difference
between smokers and nonsmokers was observed for NO2 associations with lung cancer
incidence (Raaschou-Nielsen et al.. 201 Ib; Raaschou-Nielsen et al.. 2010a). although the
association with lung cancer mortality was larger in smokers when limited to males
(Katanoda etal..2011).
A controlled human exposure study demonstrated larger NCh-induced decrements in lung
function among smokers compared to nonsmokers, but there is no information from
experimental or epidemiologic studies on asthma-related effects. Most epidemiologic
comparisons of smokers and nonsmokers are for NC^-related cardiovascular effects,
diabetes, or mortality and do not indicate differences between the groups. Many studies
similarly defined smoking as current or former smoking, providing a basis for
comparisons across studies. Although there is lack of evidence for differences in risk for
NO2-related health effects by smoking status, there is also uncertainty as to whether NO2
has an independent relationship with cardiovascular effects, diabetes (Section 6.3.9). and
mortality (Section 6.5.3). the health effects for which smoking status was examined.
Therefore, the evidence is inadequate to determine whether smoking increases the risk of
NC>2-related health effects.
Table 7-24 Controlled human exposure study evaluating smoking.
Direction of
Factor Reference Effect Study
Evaluated Category Modification3 Outcome Study Population Details Study
Nonasthma outcomes and short-term exposure
Current or Never _ Lung function n = 20 healthy adults 300 ppb Morrow et al.
former smoking decrement (10 males, 10 females) NO2for4h (1992)
smoking n = 13 Mean age 61 yr with
n = 7 exercise
NO2 = nitrogen dioxide.
aA dash indicates no difference in NO2-related health effects in the group with the factor evaluated and the reference group.
7-72
-------
Table 7-25 Epidemiologic studies evaluating smoking status.
Factor
Evaluated
Nonasthma
Current or
former
smoking
n = 74%
Nonasthma
Current or
former
smokingb
Current or
former
smoking
n = 313,487
Current or
former
smoking
n = 1,503
Current or
former
smoking
n = 118
Current or
former
smoking
n = 33,380
Current or
former
smoking
n = 45%
Current or
former
smoking
n = 45.5%
Reference
Category
outcomes and
Never
smoking
n = 28.4%
outcomes and
Never
smokingb
Never
smoking
n = 396,647
Never
smoking
n = 481
Never
smoking
n = 24
Never
smoking
n = 18,438
Never
smoking
n = 55%
Never
smoking
n = 54.5%
Direction of
Effect
Modification3 Outcome
short-term exposure
i Change in
"*" ventricular
repolarization
long-term exposure
_ Incident Ml
_ Heart failure
_ Incident stroke
_ Fatal stroke
i Diabetes
_ Atherosclerosis
(carotid intima
media thickness)
Atherosclerosis
(carotid intima
media thickness)
Study
Population
n = 580 males
Mean age 75 yr
n = 84,562
Ages 30-55 yr at
enrollment
n = 836,557
Ages 40-89 yr in
2003
n = 1,984
baseline
n = 142
Ages 50-65 yr at
baseline
n = 51,818
baseline
n = 745
Ages 26-30 yr
n = 2,780
Median age
58 yr
Study Details
Boston, MA,
Follow-up:
2000-2008
U.S.,
1990-2008
England,
2003-2007
Copenhagen,
Aarhus counties,
Denmark,
1993-2006
Copenhagen,
Aarhus counties,
Denmark,
1993-2006
Utrecht, the
Netherlands,
1999-2000
Girona Province,
Spain,
2007-2010
Study
Baia et al.
(2010VT
Hartetal.
(201 3)t
Atkinson et al.
(201 3)t
Andersen et
al. (2012b)t
Andersen et
al. (2012c)t
Lenters et al.
(2010VT
Rivera et al.
(201 3)t
7-73
-------
Table 7-25 (Continued): Epidemiologic studies evaluating smoking status.
Factor
Evaluated
Current
smoking
n - 90
Current or
former
smoking
n = 917
Current or
former
smoking
n=25.5-
65% across
cohorts
Current or
former
smoking
n= 2,850
Current or
former
smoking
n = 64%
Current
smoking
n=608
Reference
Category
Former
smoking
n - 152
Never
smoking
n = 619
Never
smoking
n =
35-74.5%
across
cohorts
Never
smoking
n = 4,359
Never
smoking
n = 36%
Never or
former
smoking
n = 1,248
Direction of
Effect
Modification3
1
-
-
-
1
-
t
1
1
Outcome
C-reactive
protein
TNF-a
IL-6
IL-8
Fibrinogen
Hepatocyte
growth factor
Blood
IL-6 level
Respiratory
mortality
CVD mortality
Diabetes-related
mortality
Total mortality
Study
Population
n = 242 adults
with clinically
ctohla C^PlDn
Mean age 68 yr
n = 1,536
Ages 45-70 yr
n = 307,553
Mean age
across
16 cohorts
41. 9-73.0 yr at
baseline
n = 9,941, 256
deaths
Ages 35-1 03 yr
n = 52,061
Ages 50-64 yr
n = 3,320
Mean age 70 yr
Study Details
Barcelona,
Spain,
2004-2006
Stockholm
county, Sweden,
1992-1994
Europe
Follow-up:
1985-2007
NO2 exposure
assessed for
2008-2011
Shenyang,
China
Follow-up:
1998-2009
NO2 exposure
assessed for
1998-2009
Denmark
Follow-up:
1971-2009. NO2
exposure
assessed for
1971-2009
London,
England
Follow-up:
1995-2005
NO2 assessed
for 2002
Study
Dadvand et al.
(2014b)t
Panasevich et
al. (2009)t
Dimakopoulou
etal. (2014) t
Zhana et al.
(201 1)t
Raaschou-
Nielsen et al.
(2012)t
Maheswaran
etal. (2010)t
7-74
-------
Table 7-25 (Continued): Epidemiologic studies evaluating smoking status.
Factor
Evaluated
Current or
former
smoking
n = 322,766
Current
smoking,
male
n=292
deaths
Current or
former
smoking
n = 3,713
Current or
former
smoking
n = 3,372
Current or
former
smoking
n = 19,253
Reference
Category
Never
smoking
n = 386,591
Former
smoking,
male
n = 90
deaths
No smoking
n = 9,135
No smoking
n = 109
No smoking
n = 33,717
Direction of
Effect
Modification3 Outcome
i Total mortality
* Lung cancer
' mortality
i Lung cancer or
"*" cardiopulmonary
mortality
_ Lung cancer
incidence
_ Lung cancer
incidence
Study
Population
n = 835,607
deaths
Ages 40-89 yr
n = 63,520
Ages >40 yr
n = 13,444
Ages >65 yr
n = 3,481
Ages 20-93 yr at
baseline
n = 52,970
Ages 50-64 yr
Study Details
England
Follow-up:
2003-2007
NO2 exposure
assessed for
2002
3 prefectures,
Japan
1983-1985
Shizuoka,
Japan,
1999-2006
Copenhagen,
Aarhus counties,
Denmark,
1970-1997
Copenhagen,
Aarhus counties,
Denmark,
1993-1997
Study
Carey et al.
(201 3)t
Katanoda et
al. (2011 Vr
Yorifuji et al.
(201 0)t
Raaschou-
Nielsen et al.
(2010a)t
Raaschou-
Nielsen et al.
(2011b)t
CVD = cardiovascular mortality; IL = interleukin; Ml = myocardial infarction; NO2 = nitrogen dioxide; TNF = tumor necrosis factor.
aUp facing arrow indicates that the effect of NO2 is greater (e.g., larger risk of hypertension) in the group with the factor evaluated
than in the reference group. Down facing arrow indicates that the effect of NO2 is smaller in the group with the factor evaluated
than in the reference group. A dash indicates no difference in NO2-related health effect between groups.
bSample size not reported.
fStudies published since the 2008 ISA for Oxides of Nitrogen
7-75
-------
7.6.3 Physical Activity
Physical activity outdoors could lead to higher NC>2 exposure (Section 3.4.3.1) and
uptake in the respiratory tract. There is some evidence indicating that during physical
activity, increased respiratory rate and oronasal breathing can increase the deposition of
NC>2 in the lower respiratory tract (Section 4.2.2.3). Thus, physical activity could have
implications for increasing the risk of NC>2-related health effects. However, the role of
concurrent physical activity in modifying NO2-related health effects has not been
characterized. Rather, physical activity has been examined as an indicator of active
versus sedentary lifestyle or fitness. Further, outdoor activity has not been assessed. The
influence of general activity or fitness on NC>2 exposure and internal dose are not known.
Epidemiologic studies did not examine physical activity or exercise as a modifier of
asthma-related effects but of cardiovascular effects, diabetes, and mortality, for which
independent relationships with NCh are uncertain (Sections 6.3.9 and 6.5.3). These
studies have inconsistent results with respect to whether physical activity increases the
risk for NCh-related health effects (Table 7-26). Associations between long-term NO2
exposure and mortality from diabetes was higher in the group not engaging in exercise
(Raaschou-Nielsen et al.. 2012). but risk of diabetes was similar or lower among those
with low levels of physical activity (Eze et al.. 2014; Andersen et al.. 2012c). Similarly,
NO2-related cardiovascular mortality was greater in the group with no exercise (Zhang et
al.. 2011). but associations with other cardiovascular effects were similar between groups
with low or high physical activity (Hartetal.. 2013; Panasevich et al.. 2009).
Contributing to the uncertainty in the evidence base is the heterogeneity across studies in
how physical activity was defined, for example, the frequency or intensity of activity.
Overall, the inconsistent evidence based on health effects for which independent
relationships with NC>2 exposure are uncertain is inadequate to determine whether low or
high physical activity increases the risk for NC^-related health effects.
7-76
-------
Table 7-26 Epidemiologic studies evaluating physical activity.
Factor
Evaluated
Reference
Category
Direction of
Effect
Modification3 Outcome
Study
Population
Study Details
Study
Nonasthma outcomes and long-term exposure
Low physical
activity
(<18METS/
week)b
Physical
inactivity
n= 23,536
Low physical
activity
(<0.5 h/week)
n = 38%
Physical
inactivity
(inactive
leisure time)
n = 543
No Exercise
n = 45.7%
No Exercise
n = 5,795
High physical
activity
(>18METS/
week)b
Physical
activity or
playing sports
in leisure time
n = 28,282
Physical
activity
(>2 h/week)
n = 28%
Physical
activity
n = 993
Exercise
n = 54.3%
Exercise
n = 4,146
_ Incident Ml n = 84,562
Ages 30-55 yr
at baseline
i Diabetes n = 51,818
Ages 50-65 yr
at baseline
_ Diabetes n = 6,392
Ages 29-73 yr
Blood IL-6 n = 1,536
level Ages 45-70 yr
| Diabetes- n = 52,061
related Ages 50-64 yr
mortality
f CVD mortality n = 9,941, 256
' deaths
Ages 35-103 yr
U.S.,
1990-2008
Copenhagen,
Aarhus counties,
Denmark,
1993-2006
Switzerland,
2002
Stockholm county,
Sweden,
1992-1994
Denmark
Follow-up and NO2
exposure assessed
for 1971 -2009
Shenyang, China
Follow-up and NO2
exposure assessed
for 1998-2009
Hart etal.
(201 3)t
Andersen et
al. (2012cVr
Eze et al.
(2014)t
Panasevich
etal.
(2009)t
Raaschou-
Nielsen etal.
(2012)t
Zhanq et al.
(201 1)t
CVD = cardiovascular disease; IL = interleukin; METS = metabolic equivalents; Ml = myocardial infarction; NO2 = nitrogen dioxide.
aUp facing arrow indicates that the effect of NO2 is greater (e.g., larger risk of mortality) in the group with the factor evaluated than
in the reference group. Down facing arrow indicates that the effect of NO2 is smaller in the group with the factor evaluated than in
the reference group. A dash indicates no difference in NO2-related health effect between groups.
bSample size not reported.
fStudies published since the 2008 ISA for Oxides of Nitrogen.
7.7
Conclusions
This chapter evaluates factors that may characterize populations and lifestages at
increased risk for health effects related to NO2 exposure (Table 7-27). The evidence for
each factor was classified based on judgments of the consistency, coherence, and
7-77
-------
biological plausibility of evidence integrated across epidemiologic, controlled human
exposure, and toxicological studies as detailed in Table 7-1. The evaluation also drew
upon information presented in preceding chapters on exposure, dosimetry, modes of
action, and independent relationships of NO2 exposure with health effects.
Consistent with observations made in the 2008 ISA for Oxides of Nitrogen (U.S. EPA.
2008c), there is adequate evidence to conclude that people with asthma, children, and
older adults are at increased risk for NCh-related health effects. Not only does evidence
consistently indicate increased risk for these groups, but the evidence is based on findings
for short-term NC>2 exposure and outcomes related to asthma exacerbation. Asthma
exacerbation is the basis for concluding a causal relationship exists between short-term
NO2 exposure and respiratory effects (Section 5.2.9). In addition to the strong evidence
for a relationship between short-term NC>2 exposure and asthma exacerbation, the
conclusion that people with asthma are at increased risk of NCh-related health effects is
supported by results from a meta-analysis of controlled human exposure studies
demonstrating that NC>2 exposure increases airway responsiveness, a key feature of
asthma exacerbation, at lower concentrations in people with asthma compared to healthy
individuals. Epidemiologic evidence does not consistently demonstrate differences in
NO2-related respiratory effects in people with asthma. It is important to note that there is
evidence of heterogeneity in asthma severity and triggers within study populations; thus,
the epidemiologic evidence is not considered to be in conflict with experimental
evidence. Children and older adults consistently have larger magnitude associations
between NCh exposure and asthma hospital admissions and ED visits, compared to adults
or the general population. There is not clear evidence from controlled human exposure
studies that NCh induces respiratory effects in older adults, but examination is limited and
based on healthy adults, not adults with asthma. Time-activity patterns and ventilation
rates differ among age groups, but it is not understood whether these factors contribute to
increased NO2 exposure, internal dose, or risk of NCh-related health effects for children
and older adults.
7-78
-------
Table 7-27 Summary of evidence for potential increased nitrogen dioxide
exposure and increased risk of nitrogen dioxide-related health
effects.
Evidence
Classification
Factor Evaluated
Rationale for Classification
Adequate
evidence
Asthma (Section 7.3.1)
Lifestage (Section 7.5.1.1): Children
(Section 7.5.1.2): Older adults
Each factor: consistent evidence for
increased risk for NCb-related asthma
exacerbation.
Asthma: evidence from controlled human
exposure studies.
Lifestage: different time-activity patterns
and ventilation patterns but unclear
implications for differences in NO2
exposure or internal dose.
Suggestive
evidence
SES (Section 7.5.2): Low SES
Sex (Section 7.5.4): Females
Diet (Section 7.6.1): Reduced antioxidant intake
Each factor: limited and generally
supporting evidence for differences in
NO2-related health effects.
SES and females: findings based primarily
on mortality for SES and exposure and
lung function for females. Uncertainty in
independent relationships with NO2
exposure provides limited basis for
inferences about differential risk.
Reduced dietary antioxidant vitamin intake:
consistent evidence from experimental
studies for modification of NO2-related
respiratory effects, but changes in oxidant
balance may not necessarily indicate
health effects.
Inadequate
evidence
COPD (Section 7.3.2)
Cardiovascular disease (Section 7.3.3)
Diabetes (Section 7.3.4)
Genetic factors (Section 7.4)
Obesity (Section 7.3.5)
Smoking (Section 7.6.2)
Physical activity (Section 7.6.3)
Race/ethnicity (Section 7.5.3)
Residence in urban areas (Section 7.5.5)
Proximity to roadways (Section 7.5.6)
Epidemiologic findings inconsistently show
differences in NO2-related health effects,
show no difference, or are limited in
quantity.
Findings based primarily on cardiovascular
effects, diabetes, birth outcomes, and
mortality. Uncertainty in independent
relationships with NO2 provides limited
basis for inferences about differential risk.
Indication of higher NO2 exposure among
nonwhite populations, urban residents, and
people spending time or living near
roadways, but insufficient information to
assess increased risk of NO2-related
health effects.
Evidence of no None
effect
COPD = chronic obstructive pulmonary disease; NO2 = nitrogen dioxide; SES = socioeconomic status.
7-79
-------
There is suggestive evidence that people with low antioxidant diets, people of low SES,
and females are at increased risk for NO2-related health effects because of some
uncertainties in the evidence bases. While experimental studies indicate that dietary
intake of Vitamin C or E modifies NCh-related effects on airway responsiveness, much of
the evidence is for effects on oxidant balance, which are not necessarily indicative of
health effects. Evidence indicates that low SES populations have higher NO2 exposure
and larger NO2-related risk of mortality. For females, limited epidemiologic evidence
points to larger NCh-related decrements in lung function. However, for low SES
populations and females, the evidence is based on health effects for which independent
relationships with NC>2 exposure are uncertain (Sections 5.4.8 and 6.2.9).
There is inadequate evidence to determine whether pre-existing cardiovascular disease,
diabetes, COPD, genetic variants, obesity, smoking, or physically active lifestyle
increases the risk for NCh-related health effects. Studies show either inconsistent or no
modification of NCh-related health effects by these factors, and information is based
primarily on cardiovascular effects (Section 5.3.11 and 6.3.9) and mortality
(Sections 5.4.8 and 6.5.3) for which independent relationships with NCh are uncertain.
Evidence also is inadequate to determine whether race/ethnicity, urban residence, or
proximity to roadways increase the risk for NC>2-related health effects. While nonwhite
populations, urban residents, and people with close proximity to roadways (i.e., living,
attending school, working, or commuting on or near roadways) may have increased
exposure to NCh, there is limited or inconsistent evidence for larger NCh-related health
effects in these populations. Further, inferences about the potential differential risk for
these populations are limited by evidence that is based on cardiovascular effects
(Section 6.3.9) and birth outcomes (Section 6.4.5). for which independent effects of NC>2
exposure are uncertain. Additionally, it is important to note that many factors may be
acting in combination to influence risk, which may lead to a different public health
impact than is reflected when evaluating any one factor in isolation. However, at this time
information remains limited as to the impact of multiple factors and how they affect the
risk for NC^-related health effects.
In conclusion, evidence is adequate to conclude that people with asthma, children, and
older adults are at increased risk for NCh-related health effects. The large proportions of
the U.S. that encompass each of these groups and lifestages (i.e., 8% adults and 9.3%
children with asthma, 24% childhood lifestage, 13% older adult lifestage) underscores the
potential public health impact of NCh-related health effects.
7-80
-------
APPENDIX: EVALUATION OF STUDIES ON
HEALTH EFFECTS OF OXIDES
OF NITROGEN
This appendix describes the approach used in the Integrated Science Assessment (ISA)
for Oxides of Nitrogen to evaluate strength of inference from health effect studies. As
described in the Preamble to the ISA (Section 5.a), causal determinations were informed
by the integration of evidence across scientific disciplines (e.g., exposure, animal
toxicology, epidemiology) and related outcomes and judgments of the strength of
inference from individual studies. Table A-l describes aspects considered in evaluating
the strength of inference from controlled human exposure, animal toxicological, and
epidemiologic studies. This evaluation was applied to studies included in this ISA from
previous assessments and those published since the 2008 ISA for Oxides of Nitrogen.
The aspects found in Table A-l are consistent with current best practices employed in
other approaches for reporting or evaluating health science data.1 Additionally, the
aspects are compatible with guidelines published by the United States Environmental
Protection Agency related to cancer, neurotoxicity, reproductive toxicity, and
developmental toxicity (U.S. EPA. 2005. 1998b. 1996b. 1991).
These aspects were not used as a checklist or strict criteria to define the quality of a
study, and judgments on strength of inference were made without considering the results
of a study. The presence or absence of particular features in a study did not necessarily
define a less informative study or exclude a study from consideration in the ISA. Further,
these aspects were not criteria for a particular determination of causality in the five-level
hierarchy. As described in the Preamble, causal determinations were based on judgments
of the overall strengths and limitations of the collective body of available studies and the
coherence of evidence across scientific disciplines and related outcomes. Table A-l is not
intended to be a complete list of aspects that affect the strength of inference from a study,
but they comprise the major aspects considered in this ISA to evaluate studies. Where
possible, study considerations, for example, exposure assessment and confounding
(i.e., bias due to a relationship with the outcome and correlation with exposures to oxides
of nitrogen), are framed to be specific to oxides of nitrogen. Thus, judgments of the
strength of inference from a study can vary depending on the specific pollutant being
assessed.
1 For example, NTP OHAT approach (Rooney etal. 2014). IRIS Preamble (U.S. EPA. 2013p). ToxRTool
(Klimisch et al. 1997). STROBE guidelines (von Bimetal.. 2007). and ARRIVE guidelines (Kilkenny etal. 2010).
A-l
-------
Table A-1 Scientific considerations for evaluating the strength of inference
from studies on the health effects of oxides of nitrogen.
Study Aspect:
Study Design
Controlled Human Exposure
Inference is stronger for studies that clearly describe the primary and any secondary aims of the study or specific
hypotheses being tested. Study subjects should be randomly exposed without knowledge of the exposure condition.
Preference is given to balanced crossover (repeated measures) or parallel design studies that include control
(e.g., clean filtered air) exposures. In crossover studies, there should be sufficient and specified time between
exposure days to avoid carryover effects from prior exposure days. In parallel design studies, all study groups
should be matched for individual characteristics such as age, sex, race, anthropometric properties, and health
status. In studies evaluating effects of disease, appropriately matched healthy controls are desired for interpretative
purposes.
Animal Toxicology
Inference is stronger for studies that clearly describe the primary and any secondary aims of the study or specific
hypotheses being tested. Studies should include appropriately matched control exposures (e.g., clean filtered air,
time matched). Studies should use methods to limit differences in baseline characteristics of control and exposure
groups. Studies should randomize assignment to exposure groups and, where possible, conceal allocation to
research personnel. Experimental procedures and conditions as well as animal care (e.g., housing, husbandry)
should be identical between groups. Blinding of research personnel to the study group may not be possible due to
animal welfare and experimental considerations; however, differences in the monitoring or handling of animals in all
groups by research personnel should be minimized.
Epidemiology
Inference is stronger for studies that clearly describe the primary and any secondary aims of the study or specific
hypotheses being tested.
For short-term exposure, time-series, case crossover, and panel studies are emphasized over cross-sectional
studies because they examine temporal correlations and are less prone to confounding by factors that differ
between individuals (e.g., SES, age). Panel studies with scripted exposures, in particular, can have strong inference
because they have consistent, well-defined exposure durations across subjects, measure pollutants at the location
of exposures, and measure outcomes at consistent, well-defined lags after exposures. Studies with large sample
sizes and conducted over multiple years are considered to produce more reliable results. If other quality parameters
are equal, multicity studies carry more weight than single-city studies because they tend to have larger sample sizes
and lower potential for publication bias.
For long-term exposure, inference is considered to be stronger for prospective cohort studies and case-control
studies nested within a cohort (e.g., for rare diseases) than cross-sectional, other case-control, or ecologic studies.
Cohort studies can better inform the temporality of exposure and effect. Other designs can have uncertainty related
to the appropriateness of the control group or validity of inference about individuals from group-level data.
Study design limitations can bias health effect associations in either direction.
A-2
-------
Study Aspect:
Study Population/Test Model
Controlled Human Exposure
In general, the subjects recruited into study groups should be similarly matched for age, sex, race, anthropometric
properties, and health status. In studies evaluating effects of specific subject characteristics (e.g., disease, genetic
polymorphism), appropriately matched healthy controls are preferred. Relevant characteristics and health status
should be reported for each experimental group. Criteria for including and excluding subjects should be clearly
indicated. For the examination of populations with an underlying health condition (e.g., asthma), independent,
clinical assessment of the health condition is ideal, but self-report of physician diagnosis generally is considered to
be reliable for respiratory diseases as well as cardiovascular diseases and events.3 The loss or withdrawal of
recruited subjects during the course of a study should be reported. Specific rationale for excluding subject(s) from
any portion of a protocol should be explained.
Animal Toxicology
Ideally, studies should report species, strain, substrain, genetic background, age, sex, and weight. However,
differences among studies in these parameters do not necessarily exclude comparisons from being made across
studies. Unless data indicate otherwise, all animal species and strains are considered appropriate for evaluating
effects of NO2 or NO exposure. It is preferred that investigators test for effects in both sexes and multiple lifestages
and report the result for each group separately. All animals used in a study should be accounted for, and rationale
for exclusion of animals or data should be specified.
Epidemiology
There is greater confidence in results from studies with a population that is recruited from and representative of the
target population. Studies with high participation and low drop-out over time that is not dependent on exposure or
health status are considered to have low potential for selection bias. Clearly specified criteria for including and
excluding subjects can aid in assessing selection bias. For populations with an underlying health condition,
independent, clinical assessment of the health condition is valuable, but self-report of physician diagnosis generally
is considered to be reliable for respiratory diseases as well as cardiovascular diseases and events.3 Comparisons of
groups with and without an underlying health condition are more informative if groups are from the same source
population. Selection bias can influence results in either direction or may not affect the validity of results but rather
reduce the generalizability of findings to the target population.
Pollutant
Controlled Human Exposure
The focus is on studies testing NO2 exposure.
Animal Toxicology
The focus is on studies testing NO2 exposure.
Epidemiology
Most of the available health effect studies examine NCb; fewer examine NO or NOx. Studies that compare health
effect associations among these species are informative. Typically, one species is examined, and studies of NO2
are emphasized. It is not clear that ambient-relevant NO exposures induce detrimental health effects
(Section 4.2.3). The relationship of NOx to NO2 varies with distance from roads, and thus, may vary among
subjects. Hence, the extent to which associations with NOx reflect those for NO2 versus other pollutants from traffic
is uncertain.
A-3
-------
Study Aspect:
Exposure Assessment or Assignment
Controlled Human Exposure
Studies should well characterize pollutant concentration, temperature, and relative humidity and/or have measures
in place to adequately control the exposure conditions for subject safety. For this assessment, the focus is on
studies that use NO2 or NO concentrations less than or equal to 5,000 ppb (Section 1.2). Studies that use higher
exposure concentrations may provide information relevant to mode of action, dosimetry, or at-risk human
populations. Preference is given to balanced crossover or parallel design studies that include control exposures
(e.g., clean filtered air). Study subjects should be randomly exposed without knowledge of the exposure condition.
Method of exposure (e.g., chamber, facemask) should be specified and activity level of subjects during exposures
should be well characterized.
Animal Toxicology
Studies should characterize pollutant concentration, temperature, and relative humidity and have measures in place
to adequately control the exposure conditions. The focus is on inhalation exposure. Noninhalation exposure
experiments may provide information relevant to mode of action. In vitro studies generally are not included;
however, such studies may be included, particularly if conducted in airway cells, if they provide mechanistic insight
or examine similar effects as in vivo. All studies should include exposure control groups (e.g., clean filtered air). For
this assessment, the focus is on studies that use NO2 or NO concentrations less than or equal to 5,000 ppb
(Section 1.2). Studies that use higher exposure concentrations may provide information relevant to mode of action,
dosimetry, interspecies variation, or at-risk human populations.
Epidemiology
Of primary relevance are relationships of health effects with the ambient component of exposure to oxides of
nitrogen. However, information about ambient exposure rarely is available for individual subjects; most often,
inference is based on ambient concentrations. Studies that compare exposure assessment methods are considered
to be particularly informative. Inference is strong when the duration or lag of the exposure metric corresponds with
the time course for physiological changes in the outcome (e.g., up to a few days for symptoms) or latency of
disease (e.g., several years for cancer).
Given the spatial heterogeneity in ambient oxides of nitrogen and variable relationships between personal
exposures and ambient concentrations (Section 3.4.2). validated methods that capture the extent of variability for
the particular study design (temporal for short-term exposure studies versus spatial contrasts for long-term
exposure studies) and location carry greater weight. Central site measurements, whether averaged across multiple
monitors or assigned from the nearest or single available monitor, have well-recognized limitations in capturing
variation in oxides of nitrogen. Inference from central site measurements can be adequate if correlated with
personal exposures, closely located to study subjects, highly correlated across monitors within a location, used in
locations with well-distributed sources, or combined with time-activity information.
In studies of short-term exposure, metrics that may capture temporal variation in ambient oxides of nitrogen and
provide a good basis for inference include concentrations in subjects' microenvironments (e.g., outdoor home,
school, in-vehicle) and individual-level outdoor concentrations combined with time-activity data. Results for total
personal and indoor NO2 exposure are other lines of evidence that inform judgments about causality of NO2
because inference is based on an individual's microenvironmental exposures and the potential for copollutant
confounding may be lower or different than that for ambient concentrations. Results for total personal exposure can
inform understanding of the effects of ambient exposure when well correlated with ambient concentrations. For
long-term exposures, LUR models validated to well represent spatial variation in ambient NO2 in the study area can
provide good estimates of individual exposure. Less weight is placed on NOx from dispersion models because NOx
estimates often show near perfect correlations (r = 0.94-0.99) with EC, PlVh.5, and CO, and the effects of NOx
cannot be distinguished from traffic-related copollutants. Dispersion models may not well account for all sources of
NO2 or NOx or the emissions and transformation chemistry in an urban location, and thus, poorly estimate
neighborhood-scale variability (Section 3.2.2.2).
Exposure measurement error often attenuates health effect estimates or decreases the precision of the association
(i.e., wider 95% CIs), particularly associations based on temporal variation in short-term exposure (Section 3.4.5).
However, exposure measurement error can bias estimates away from the null, particularly for long-term exposures.
A-4
-------
Study Aspect:
Outcome Assessment/Evaluation
Controlled Human Exposure
Endpoints should be assessed in the same manner for control and exposure groups (e.g., time after exposure,
methods, endpoint evaluator) using valid, reliable methods. Blinding of endpoint evaluators is ideal, especially for
qualitative endpoints (e.g., histopathology). For each experiment and each experimental group, including controls,
precise details of all procedures carried out should be provided including how, when, and where. Time of the
endpoint evaluations is a key consideration that will vary depending on the endpoint evaluated. Endpoints should be
assessed at time points that are appropriate for the research questions.
Animal Toxicology
Endpoints should be assessed in the same manner for control and exposure groups (e.g., time after exposure,
methods, endpoint evaluator) using valid, reliable methods. Blinding of endpoint evaluators is ideal, especially for
qualitative endpoints (e.g., histopathology). For each experiment and each experimental group, including controls,
precise details of all procedures carried out should be provided including how, when, and where. Time of the
endpoint evaluations is a key consideration that will vary depending on the endpoint evaluated. Endpoints should be
assessed at time points that are appropriate for the research questions.
Epidemiology
Inference is stronger when outcomes are assessed or reported without knowledge of exposure status. Knowledge
of exposure status could produce artifactual associations. There is greater confidence when outcomes assessed by
interview, self-report, clinical examination, or analysis of biological indicators are defined by consistent criteria and
collected by validated, reliable methods. Independent clinical assessment is valuable for outcomes such as lung
function or incidence of disease, but report of physician diagnosis has shown good reliability.3 Outcomes assessed
at time intervals that correspond with the time course for physiological changes (e.g., up to a few days for
symptoms) are emphasized. When health effects of long-term exposure are assessed by acute events such as
symptoms or hospital admissions, inference is strengthened when results are adjusted for short-term exposure.
Validated questionnaires for subjective outcomes such as symptoms are regarded to be reliable,13 particularly when
collected frequently and not subject to long recall. For biological samples, the chemical stability of the substance of
interest and the sensitivity and precision of the analytical method is considered.
If not based on knowledge of exposure status, errors in outcome assessment tend to bias results toward the null.
A-5
-------
Study Aspect:
Potential Copollutant Confounding
Controlled Human Exposure
Exposure should be well characterized to evaluate independent effects of NO2 or NO.
Animal Toxicology
Exposure should be well characterized to evaluate independent effects of NO2 or NO.
Epidemiology
Confounding can occur by copollutants that are highly correlated with oxides of nitrogen and have similar modes of
action and health effects. Not accounting for copollutant confounding can produce artifactual associations; thus,
studies that examine copollutant confounding carry greater weight. The predominant method is copollutant
modeling, which is especially informative when measurement error is comparable for copollutants and correlations
are not high. Interaction and joint effect models are examined to a lesser extent. Copollutant confounding also can
be assessed by evaluating correlations between oxides of nitrogen and copollutants and comparing health
associations between oxides of nitrogen and copollutants in single-pollutant models if exposure measurement error
is comparable among pollutants. Studies that examine only oxides of nitrogen are considered to poorly inform the
potential for copollutant confounding.
Among copollutants, those of primary concern are traffic-related pollutants, which include CO, PlVh.s, BC/EC, OC,
UFP, metal PM components such as copper, zinc, and iron, as well as VOCs such as benzene, acetaldehyde,
toluene, ethylbenzene, and xylene. Short-term and long-term metrics for these pollutants often show moderate to
high correlations with oxides of nitrogen (Figure 3-6). Many traffic-related pollutants also are hypothesized to have
common modes of action.0 Common key events include formation of secondary oxidation products, inflammation,
and, for respiratory effects, increases in airway responsiveness. Traffic-related pollutants also show relationships
with many of the health effects evaluated in this ISAd except as follows: for long-term exposure, there is uncertainty
regarding confounding by UFP because of the short atmospheric lifetime; and for long-term exposure, CO is not
considered to be an important confounding copollutant for mortality or lung cancer.d
There generally is less concern of confounding by PM-io, SO2, and Os because they show varying and often lower
correlations with NO2 (Figure 3-6). Os generally is negatively or weakly positively correlated with NO2 but may be a
confounder where moderate positive correlations are observed. Os and SO2 in particular show similarities with NO2
in mode of action. PM-io, SO2, and Os show relationships with the health effects evaluated in this ISAd except as
follows: for short-term exposure, SO2 is not considered to be a strong confounding copollutant for cardiovascular
effects; and for long-term exposure, neither Os nor SO2 is considered to be a strong confounder for any health
effect.
A-6
-------
Study Aspect:
Other Potential Confounding Factors6
Controlled Human Exposure
Preference is given to studies using experimental and control groups that are matched for individual level
characteristics (e.g., body weight, smoking history, age) and time-varying factors (e.g., seasonal and diurnal
patterns).
Animal Toxicology
Preference is given to studies using experimental and control groups that are matched for individual level
characteristics (e.g., body weight, litter size, food and water consumption) and time-varying factors (e.g., seasonal
and diurnal patterns).
Epidemiology
Factors are considered to be potential confounders if demonstrated in the scientific literature to be related to health
effects and correlated with oxides of nitrogen and/or traffic indicators. Not accounting for confounders can produce
artifactual associations; thus, studies that statistically adjust for multiple factors or control for them in the study
design are emphasized. Less weight is placed on studies that adjust for factors that mediate the relationship
between oxides of nitrogen and health effects, which can bias results toward the null. In the absence of information
linking health risk factors to oxides of nitrogen or traffic indicators, a factor may be evaluated as a potential effect
measure modifier, but uncertainty is noted as to its role as a confounder. Confounders vary according to study
design, exposure duration, and health effects and include the following.
For time-series and panel studies of short-term exposure:
• Respiratory Effects—meteorology, day of week, season, medication use, allergen exposure (potential
effect modifier)
• Cardiovascular Effects—meteorology, day of week, season, medication use
• Total Mortality—meteorology, day of week, season, long-term temporal trends
For studies of long-term exposure:
• Respiratory Effects—socioeconomic status, race, age, medication use, smoking, stress
• Cardiovascular Effects, Diabetes, Reproductive Effects, and Developmental Effects—socioeconomic
status, race, age, medication use, smoking, stress, noise
• Total Mortality—socioeconomic status, race, age, medication use, smoking, comorbid health conditions
• Cancer—socioeconomic status, race, age, occupational exposure
A-7
-------
Study Aspect:
Statistical Methodology
Controlled Human Exposure
Statistical methods should be clearly described and appropriate for the study design and research question
(e.g., correction for multiple comparisons). Generally, statistical significance is considered in the evaluation of
findings of controlled human exposure studies. Detection of statistical significance is influenced by a variety of
factors including, but not limited to, the size of the study, exposure and outcome measurement error, and statistical
model specifications. Sample size is not a criterion for excluding a study; ideally, the sample size should provide
adequate power to detect hypothesized effects (e.g., sample sizes less than three are considered less informative).
Because statistical tests have limitations, consideration is given to both trends in data and reproducibility of results.
Animal Toxicology
Statistical methods should be clearly described and appropriate for the study design and research question
(e.g., correction for multiple comparisons). Generally, statistical significance is considered in the evaluations of
findings of animal toxicology studies. Detection of statistical significance is influenced by a variety of factors
including, but not limited to, the size of the study, exposure and outcome measurement error, and statistical model
specifications. Sample size is not a criterion for excluding a study; ideally, the sample size should provide adequate
power to detect hypothesized effects (e.g., sample sizes less than three are considered less informative). Because
statistical tests have limitations, consideration is given to both trends in data and reproducibility of results.
Epidemiology
Multivariable regression models that include potential confounding factors are emphasized. However, multipollutant
models (more than two pollutants) are considered to produce too much uncertainty from copollutant collinearity to
be informative. Models with interaction terms can help evaluate potential confounding as well as effect modification.
Sensitivity analyses with alternate specifications for potential confounding inform the stability of findings and aid in
judging the strength of inference of results. In the case of multiple comparisons, consistency in the pattern of
association can increase confidence that the associations were not found by chance alone. Statistical methods that
are appropriate for the power of the study carry greater weight. For example, categorical analyses with small
sample sizes can be prone to bias results toward or away from the null. Statistical tests such as t-tests and
chi-squared tests are not considered sensitive enough for adequate inferences regarding pollutant-health effect
associations. For all methods, the effect estimate and precision of the estimate (i.e., width of 95% Cl) are important
considerations rather than statistical significance.
BC = black carbon; Cl = confidence interval; CO = carbon monoxide; EC = elemental carbon; ISA = Integrated Science
Assessment; LUR = land use regression; NO = nitric oxide; NO2 = nitrogen dioxide; NOX = sum of NO and NO2; O3 = ozone;
OC = organic carbon; PM25 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm;
PMio = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; ppb = parts per billion;
SES = socioeconomic status; SO2 = sulfur dioxide; UFP = ultrafine particles; VOC = volatile organic compound.
aTorenetal. (1993): Murgia et al. (2014): Weaklev et al. (2013): Yang et al. (2011): Heckbert et al. (2004): Barr et al. (2002):
Muhaiarine et al. (1997).
"Burnev et al. (1989>.
information on modes of action for NO2 is described in Section 4.3. The characterization of similar modes of action for many
traffic-related pollutants is based on information described in the most recently completed ISAs (U.S. EPA. 2013e. 201 Ob. 2009a.
2008d) and the Health Effects Institute's 2010 review of traffic-related air pollution (HEI. 2010).
dJudgments regarding potential confounding by other criteria pollutants are based on studies evaluated in this ISA, causal
determinations made in the most recently completed ISAs (U.S. EPA. 2013e. 201 Ob. 2009a. 2008d). as well as recent reviews
published by the Health Effects Institute (HEI Review Panel on Ultrafine Particles. 2013: HEI. 2010). Judgments regarding
potential confounding by the PM components EC/BC, OC, metals, and UFP as well as VOCs should not be inferred as
conclusions regarding causality of their relationships with health effects. Their consideration as potential confounders is based on
associations with oxides of nitrogen and health effects observed in the studies examined in this ISA and reviews conducted by the
Health Effects Institute. Judgments regarding potential confounding by PMio should not be inferred as conclusions regarding
causality specifically for that size fraction. The 2009 ISA for Particulate Matter evaluated PMio studies but did not form individual
causal determinations for that size fraction because PMio comprises both fine and thoracic coarse particles.
eMany factors evaluated as potential confounders can be effect measure modifiers (e.g., season, comorbid health condition) or
mediators of health effects related to oxides of nitrogen (comorbid health condition).
A-8
-------
REFERENCES
AAP (American Academy of Pediatrics). (2004). Ambient air pollution: Health hazards to children. Pediatrics
114: 1699-1707. http://dx.doi.org/10.1542/peds.2004-2166
Abbey. DE: Nishino. N: McDonnell WF: Burchette. PJ: Knutsen. SF: Beeson. WL: Yang. JX. (1999). Long-
term inhalable particles and other air pollutants related to mortality in nonsmokers. Am J Respir Crit Care
Med 159: 373-382. http://dx.doi.org/10.1164/airccm. 159.2.9806020
Ackermann-Liebrich. U: Leuenberger. P: Schwartz. J: Schindler. C: Monn. C: Bolognini. G: Bongard. JP:
Brandli O: Domenighetti. G: Elsasser. S: Grize. L: Karrer. W: Keller. R: Keller-Wossidlo. H: Kiinzli. N:
Martin. BW: Medici. TC: Perruchoud. AP: Schonj MH: Tschopp. JM: Villiger. B: Wuthrich. B: Zellweger.
JP: Zemp. E. (1997). Lung function and long term exposure to air pollutants in Switzerland. Am J Respir Crit
Care Med 155: 122-129. http://dx.doi.0rg/10.1164/airccm.155.l.9001300
Adamkiewicz. G: Ebelt S: Syring. M: Slater. J: Speizer. FE: Schwartz. J: Suh. H: Gold. DR. (2004). Association
between air pollution exposure and exhaled nitric oxide in an elderly population. Thorax 59: 204-209.
http://dx.doi.org/10.1136/thorax.2003.006445
Adamkiewicz. G: Hsu. HH: Vallarino. J: Melly. SJ: Spengler. JD: Lew. JI. (2010). Nitrogen dioxide
concentrations in neighborhoods adjacent to a commercial airport: A land use regression modeling study.
Environ Health 9: 73. http://dx.doi.org/10.1186/1476-069X-9-73
Adams. WC: Brookes. KA: Schelegle. ES. (1987). Effects of NO2 alone and in combination with O3 on young
men and women. J Appl Physiol 62: 1698-1704.
Adgent MA: Squadrito. GL: Ballinger. CA: Krzywanski. DM: Lancaster. JR. Jr: Postlethwait EM. (2012).
Desferrioxamine inhibits protein tyro sine nitration: Mechanisms and implications. Free Radic Biol Med 53:
951-961. http://dx.doi.0rg/10.1016/i.freeradbiomed.2012.06.003
Adkins. B. Jr: Van Stee. EW: Simmons. JE: Eustis. SL. (1986). Oncogenic response of strain A/J mice to inhaled
chemicals. J Toxicol Environ Health 17: 311-322. http://dx.doi.org/10.1080/15287398609530825
Aga. E: Samoli. E: Touloumi. G: Anderson. HR: Cadum. E: Forsberg. B: Goodman. P: Goren. A: Kotesovec. F:
Kriz. B: Macarol-Hiti. M: Medina. S: Paldv. A: Schindler. C: Sunyer. J: Tittanen. P: Woityniak. B: Zmirou.
D: Schwartz. J: Katsouvanni. K. (2003). Short-term effects of ambient particles on mortality in the elderly:
Results from28 cities in the APHEA2 project. Eur Respir J 21: 28s-33s.
http://dx.doi.org/10.1183/09031936.03.00402803
Agarwal. R: Awasthi. A: Singh. N: Gupta. PK: Mittal SK. (2012). Effects of exposure to rice-crop residue
burning smoke on pulmonary functions and Oxygen Saturation level of human beings in Patiala (India). Sci
Total Environ 429: 161-166. http://dx.doi.0rg/10.1016/i.scitotenv.2012.03.074
Aguilera. I: Garcia-Esteban. R: Iniguez. C: Nieuwenhuijsen. MJ: Rodriguez. A: Paez. M: Ballester. F: Sunyer. J.
(2010). Prenatal exposure to traffic-related air pollution and ultrasound measures of fetal growth in the
INMA Sabadell cohort. Environ Health Perspect 118: 705-711. http://dx.doi.org/10.1289/ehp.0901228
Aguilera. I: Guxens. M: Garcia-Esteban. R: Corbella. T: Nieuwenhuijsen. MJ: Foradada. CM: Sunyer. J. (2009).
Association between GIS-based exposure to urban air pollution during pregnancy and birth weight in the
INMA Sabadell cohort. Environ Health Perspect 117: 1322-1327. http://dx.doi.org/10.1289/ehp.0800256
Aguilera. I: Pedersen. M: Garcia-Esteban. R: Ballester. F: Basterrechea. M: Esplugues. A: Fernandez-Somoano.
A: Lertxundi A: Tardon. A: Sunyer. J. (2013). Early-life exposure to outdoor air pollution and respiratory
health, ear infections, and eczema in infants from the INMA study. Environ Health Perspect 121: 387-392.
http://dx.doi.org/10.1289/ehp.1205281
Aguilera. I: Sunyer. J: Fernandez-Patier. R: Hoek. G: Aguirre-Alfaro. A: Meliefste. K: Bomboi-Mingarro. MT:
Nieuwenhuijsen. MJ: Herce-Garraleta. D: Brunekreef. B. (2008). Estimation of outdoor NOx, NO2, and
BTEX exposure in a cohort of pregnant women using land use regression modeling. Environ Sci Technol 42:
815-821. http://dx.doi.org/10.1021/es0715492
R-l
-------
Ahmad. T: Mabaliraian. U: Sharma. A: Aich. J: Makhija. L: Ghosh. B: Agrawal A. (2011). Simvastatin
improves epithelial dysfunction and airway hyperresponsiveness: From asymmetric dimethyl-arginine to
asthma. Am J Respir Cell Mol Biol 44: 531-539. http://dx.doi.org/10.1165/rcmb.2010-0041OC
Ahmed. T: Dougherty. R: Sackner. MA. (1983a). Effect of 0.1 ppm NO2 on pulmonary functions and non-
specific bronchial reactivity of normals and asthmatics [final report]. (CR-83/11/BI). Warren, MI: General
Motors Research Laboratories.
Ahmed. T: Dougherty. R: Sackner. MA. (1983b). Effect of NO2 exposure on specific bronchial reactivity in
subjects with allergic bronchial asthma [final report]. (CR-83/07/BI). Warren, MI: General Motors Research
Laboratories.
Ainslie. B: Steyn. DG: Su. J: Buzzelli M: Brauer. M: Larson. T: Rucker. M. (2008). A source area model
incorporating simplified atmospheric dispersion and advection at fine scale for population air pollutant
exposure assessment. Atmos Environ 42: 2394-2404. http://dx.doi.0rg/10.1016/i.atmosenv.2007.12.021
Al-Hegelan. M: Tighe. RM: Castillo. C: Hollingsworth. JW. (2011). Ambient ozone and pulmonary innate
immunity [Review]. Immunol Res 49: 173-191. http://dx.doi.org/10.1007/sl2026-010-8180-z
Alarie. Y. (1973). Sensory irritation by airborne chemicals [Review]. CRC Crit Rev Toxicol 2: 299-363.
http://dx.doi.org/10.3109/10408447309082020
Alberg. T: Nilsen. A: Hansen. JS: Nygaard. UC: Lovik. M. (2011). Nitrogen dioxide: No influence on allergic
sensitization in an intranasal mouse model with ovalbumin and diesel exhaust particles. Inhal Toxicol 23:
268-276. http://dx.doi.org/10.3109/08958378.2011.566898
Alexanderson. C: Olin. AC: Dahlman-Hoglund. A: Finizia. C: Toren. K. (2012). Nasal nitric oxide in a random
sample of adults and its relationship to sensitization, cat allergen, rhinitis, and ambient nitric oxide. Am J
Rhinol Allergy 26: e99-e!03. http://dx.doi.org/10.2500/aira.2012.26.3777
Allen. RW: Amram O: Wheeler. AJ: Brauer. M. (2011). The transferability of NO and NO2 land use regression
models between cities and pollutants. Atmos Environ 45: 369-378.
http://dx.doi.0rg/10.1016/i.atmosenv.2010.10.002
Allen. RW: Davies. H: Cohen. MA: Mallach. G: Kaufman. JD: Adar. SD. (2009). The spatial relationship
between traffic-generated air pollution and noise in 2 US cities. Environ Res 109: 334-342.
http://dx.doi.0rg/10.1016/i.envres.2008.12.006
Aim. S: Mukala. K: Pasanen. P: Tiittanen. P: Ruuskanen. J: Tuomisto. J: Jantunen. MJ. (1998). Personal NO2
exposures of preschool children in Helsinki. J Expo Anal Environ Epidemiol 8: 79-100.
Altug. H: Gaga. EO: Dogeroglu. T: Brunekreef B: Hoek. G: Van Doom. W. (2014). Effects of ambient air
pollution on respiratory tract complaints and airway inflammation in primary school children. Sci Total
Environ479-480: 201-209. http://dx.doi.0rg/10.1016/i.scitotenv.2014.01.127
Alving. K: Weitzberg. E: Lundberg. JM. (1993). Increased amount of nitric oxide in exhaled air of asthmatics.
Eur Respir J 6: 1368-1370.
American Heart Association. (2011). Cardiovascular disease statistics. Available online at
http://web.archive.org/web/20110410093532/http://www.americanheart.org/presenter.jhtml?identifier=4478
(accessed April 12, 2011).
Amigou. A: Sermage-Faure. C: Orsi L: Leverger. G: Baruchel. A: Bertrand. Y: Nelken. B: Robert. A: Michel
G: Margueritte. G: Perel Y: Mechinaud. F: Bordigoni. P: Hemon. D: Clavel. J. (2011). Road traffic and
childhood leukemia: The ESCALE Study (SFCE). Environ Health Perspect 119: 566-572.
http://dx.doi.org/10.1289/ehp. 1002429
Amoruso. MA: Witz. G: Goldstein. BD. (1981). Decreased superoxide anion radical production by rat alveolar
macrophages following inhalation of ozone or nitrogen dioxide. Life Sci 28: 2215-2221.
http://dx.doi.org/10.1016/0024-3205(81)90572-5
Amster. ED: Haim M: Dubnov. J: Brodav. DM. (2014). Contribution of nitrogen oxide and sulfur dioxide
exposure from power plant emissions on respiratory symptom and disease prevalence. Environ Pollut 186:
20-28. http://dx.doi.0rg/10.1016/i.envpol.2013.10.032
R-2
-------
Andersen. ZJ: Bennelykke. K: Hvidberg. M: Jensen. SS: Ketzel M: Loft. S: Serensen. M: Tjenneland. A:
Overvad. K: Raaschou-Nielsen. O. (2012a). Long-term exposure to air pollution and asthma hospitalisations
in older adults: A cohort study. Thorax 67: 6-11. http://dx.doi.0rg/10.l 136/thoraxjnl-2011-200711
Andersen. ZJ: Hvidberg. M: Jensen. SS: Ketzel. M: Loft. S: Sorensen. M: Tjonneland. A: Overvad. K:
Raaschou-Nielsen. O. (2011). Chronic obstructive pulmonary disease and long-term exposure to traffic-
related air pollution: A cohort study. Am J Respir Crit CareMed 183: 455-461.
http://dx.doi.org/10.1164/rccm.201006-0937OC
Andersen. ZJ: Kristiansen. LC: Andersen. KK: Olsen. TS: Hvidberg. M: Jensen. SS: Ketzel M: Loft S:
Serensen. M: Tjenneland. A: Overvad. K: Raaschou-Nielsen. O. (2012b). Stroke and long-term exposure to
outdoor air pollution from nitrogen dioxide: A cohort study. Stroke 43: 320-325.
http://dx.doi.0rg/10.l 161/STROKEAHA. 111.629246
Andersen. ZJ: Loft. S: Ketzel M: Stage. M: Scheike. T: Hermansen. MN: Bisgaard. H. (2008a). Ambient air
pollution triggers wheezing symptoms in infants. Thorax 63: 710-716.
http://dx.doi.org/10.1136/thx.2007.085480
Andersen. ZJ: Olsen. TS: Andersen. KK: Loft. S: Ketzel M: Raaschou-Nielsen. O. (2010). Association between
short-term exposure to ultrafme particles and hospital admissions for stroke in Copenhagen, Denmark. Eur
Heart J 31: 2034-2040. http://dx.doi.org/10.1093/eurhearti/ehql88
Andersen. ZJ: Raaschou-Nielsen. O: Ketzel M: Jensen. SS: Hvidberg. M: Loft S: Tjenneland. A: Overvad. K:
Serensen. M. (2012c). Diabetes incidence and long-term exposure to air pollution: A cohort study. Diabetes
Care 35: 92-98. http://dx.doi.org/10.2337/dcll-1155
Andersen. ZJ: Wahlin. P; Raaschou-Nielsen. O: Ketzel M: Scheike. T: Loft. S. (2008b). Size distribution and
total number concentration of ultrafine and accumulation mode particles and hospital admissions in children
and the elderly in Copenhagen, Denmark. Occup Environ Med 65: 458-466.
http://dx.doi.org/10.1136/oem.2007.033290
Andersen. ZJ: Wahlin. P: Raaschou-Nielsen. O: Scheike. T: Loft. S. (2007). Ambient particle source
apportionment and daily hospital admissions among children and elderly in Copenhagen. J Expo Sci Environ
Epidemiol 17: 625-636. http://dx.doi.org/10.1038/sj.jes.7500546
Anderson. HR: Armstrong. B: Haiat S: Harrison. R: Monk. V: Poloniecki. J: Timmis. A: Wilkinson. P. (2010).
Air pollution and activation of implantable cardioverter defibrillators in London. Epidemiology 21: 405-413.
http://dx.doi.org/10.1097/EDE.Ob013e3181d61600
Anderson. HR: Favarato. G: Atkinson. RW. (2013). Long-term exposure to air pollution and the incidence of
asthma: Meta-analysis of cohort studies. Air Qual Atmos Health 6: 47-56. http://dx.doi.org/10.1007/sll869-
011-0144-5
Anderson. HR: Ponce de Leon. A: Bland. JM: Bower. JS: Emberlin. J: Strachen. DP. (1998). Air pollution,
pollens, and daily admissions for asthma in London 1987-92. Thorax 53: 842-848.
http://dx.doi.org/10.1136/thx.53.10.842
Anderson. JT: Zeng. M: Li. Q: Staplev. R: Moore. PR: Chenna. B: Fineberg. N: Zmijewski. J: Eltoum. IE:
Siegal GP: Gaggar. A: Barnes. S: Velu. SE: Thannickal VJ: Abraham. E: Patel RP: Lancaster. J. R.:
Chaplin. DP: Dransfield. MT: Deshane. JS. (2011). Elevated levels of NO are localized to distal airways in
asthma. Free Radic Biol Med 50: 1679-1688. http://dx.doi.org/10.1016/j.freeradbiomed.2011.03.015
Annesi-Maesano. I: Hulin. M: Lavaud. F: Raherison. C: Kopferschmitt C: de Blav. F: Charpin. DA: Denis. C.
(2012a). Poor air quality in classrooms related to asthma and rhinitis in primary schoolchildren of the French
6 Cities Study. Thorax 67: 682-688. http://dx.doi.org/10.1136/thoraxinl-2011-200391
Annesi-Maesano. I: Moreau. D: Caillaud. D: Lavaud. F: Le Moullec. Y: Tavtard. A: Pauli G: Charpin. D.
(2007). Residential proximity fine particles related to allergic sensitisation and asthma in primary school
children. Respir Med 101: 1721-1729. http://dx.doi.0rg/10.1016/i.rmed.2007.02.022
Annesi-Maesano. I: Rouve. S: Desqueyroux. H: Jankovski. R: Klossek. JM: Thibaudon. M: Demoly. P: Didier.
A^ (2012b). Grass pollen counts, air pollution levels and allergic rhinitis severity. Int Arch Allergy Immunol
158: 397-404. http://dx.doi.org/10.1159/000332964
R-3
-------
Arain. MA: Blair. R: Finkelstein. N: Brook. J: Jerrett M. (2009). Meteorological influences on the spatial and
temporal variability of NO 2 in Toronto and Hamilton. Can Geogr 53: 165-190.
http://dx.doi.org/10.1111/i. 1541-0064.2009.00252.X
Aranyj C: Fenters. J: Erhlich. R: Gardner. D. (1976). Scanning electron microscopy of alveolar macrophages
after exposure to oxygen, nitrogen dioxide, and ozone [Abstract]. Environ Health Perspect 16: 180.
Arbex. MA: de Souza Conceicao. GM: Cendon. SP: Arbex. FF: Lopes. AC: Moyses. EP: Santiago. SL: Saldiva.
PHN: Pereira. LAA: Braga. ALF. (2009). Urban air pollution and chronic obstructive pulmonary disease-
related emergency department visits. J Epidemiol Community Health 63: 777-783.
http://dx.doi.org/10.1136/iech.2008.078360
Arhami. M: Polidori. A: Delfino. RJ: Tjoa. T: Sioutas. C. (2009). Associations between personal, indoor, and
residential outdoor pollutant concentrations: Implications for exposure assessment to size-fractionated
particulate matter. J Air Waste Manag Assoc 59: 392-404. http://dx.doi.Org/10.3155/1047-3289.59.4.392
Arner. EC: Rhoades. RA. (1973). Long-term nitrogen dioxide exposure: Effects on lung lipids and mechanical
properties. Arch Environ Health 26: 156-160. http://dx.doi.org/10.1080/00039896.1973.10666244
Arnold. JR: Hartsell BE: Luke. WT: Ullah. SMR: Dasgupta. PK: Huev. LG: Tate. P. (2007). Field test of four
methods for gas-phase ambient nitric acid. Atmos Environ 41: 4210-4226.
http://dx.doi.0rg/10.1016/i.atmosenv.2006.07.058
Asgharian. B: Price. OT: Schroeter. JD: Kimbell JS: Jones. L: Singal M. (2011). Derivation of mass transfer
coefficients for transient uptake and tissue disposition of soluble and reactive vapors in lung airways. Ann
Biomed Eng 39: 1788-1804. http://dx.doi.org/10.1007/sl0439-011-0274-9
Atari. DO: Luginaah. I: Xu. X: Fung. K. (2008). Spatial variability of ambient nitrogen dioxide and sulfur
dioxide in Sarnia, "Chemical Valley," Ontario, Canada. JToxicol Environ Health A 71: 1572-1581.
http://dx.doi.org/10.1080/15287390802414158
Atari. DO: Luginaah. IN: Fung. K. (2009). The relationship between odour annoyance scores and modelled
ambient air pollution in Sarnia, "Chemical Valley", Ontario. Int J Environ Res Public Health 6: 2655-2675.
http://dx.doi.org/10.3390/ijerph6102655
Ather. JL: Alcorn. JF: Brown. AL: Guala. AS: Suratt BT: Janssen-Heininger. YM: Poynter. ME. (2010).
Distinct functions of airway epithelial nuclear factor-kappaB activity regulate nitrogen dioxide-induced acute
lung injury. Am J Respir Cell Mol Biol 43: 443-451. http://dx.doi.org/10.1165/rcmb.2008-0416OC
Atkinson. RW: Anderson. HR: Strachaa DP: Bland. JM: Bremner. SA: Ponce de Leon. A. (1999a). Short-term
associations between outdoor air pollution and visits to accident and emergency departments in London for
respiratory complaints. Eur Respir J 13: 257-265.
Atkinson. RW: Bremner. SA: Anderson. HR: Strachaa DP: Bland. JM: Ponce de Leon. A. (1999b). Short-term
associations between emergency hospital admissions for respiratory and cardiovascular disease and outdoor
air pollution in London. Arch Environ Health 54: 398-411. http://dx.doi.org/10.1080/00039899909603371
Atkinson. RW: Carey. IM: Kent AJ: van Staa. TP: Anderson. HR: Cook. DG. (2013). Long-term exposure to
outdoor air pollution and incidence of cardiovascular diseases. Epidemiology 24: 44-53.
http://dx.doi.org/10.1097/EDE.Ob013e318276ccb8
Atkinson. RW: Cohen. A: Mehta. S: Anderson. HR. (2012). Systematic review and meta-analysis of
epidemiological time-series studies on outdoor air pollution and health in Asia. Air Qual Atmos Health 5:
383-391. http://dx.doi.org/10.1007/sll869-010-0123-2
ATS (American Thoracic Society). (2000a). Guidelines for methacholine and exercise challenge testing-1999.
Am J Respir Crit Care Med 161: 309-329. http://dx.doi.org/10.1164/airccm.16Ll.atsll-99
ATS (American Thoracic Society). (2000b). What constitutes an adverse health effect of air pollution? Am J
Respir Crit Care Med 161: 665-673. http://dx.doi.0rg/10.1164/airccm.161.2.ats4-00
R-4
-------
ATS/ERS (American Thoracic Society and European Respiratory Society). (2005). ATS/ERS recommendations
for standardized procedures for the online and offline measurement of exhaled lower respiratory nitric oxide
and nasal nitric oxide, 2005. Am J Respir Crit Care Med 171: 912-930.
http://dx.doi.org/10.1164/rccm.200406-710ST
ATSDR (Agency for Toxic Substances and Disease Registry). (2006). A study of ambient air contaminants and
asthma in New York City: Part A and B. Atlanta, GA: U.S. Department of Health and Human Services.
http://permanent.access.gpo.gov/lps88357/ASTHMA BRONX FINAL REPORT.pdf
Augusto. O: Bonini. MG: Amanso. AM: Linares. E: Santos. CC: De Menezes. SL. (2002). Nitrogen dioxide and
carbonate radical anion: Two emerging radicals in biology [Review]. Free Radic Biol Med 32: 841-859.
http://dx.doi.org/10.1016/S0891-5849(02)00786-4
Avol EL: Linn. WS: Peng. RC: Valencia. G: Little. D: Hackney. JD. (1988). Laboratory study of asthmatic
volunteers exposed to nitrogen dioxide and to ambient air pollution. Am Ind Hyg Assoc J 49: 143-149.
http://dx.doi.org/10.1080/15298668891379530
Avol EL: Linn. WS: Peng. RC: Whynot. JD: Shamoo. DA: Little. DE: Smith. MN: Hackney. JD. (1989).
Experimental exposures of young asthmatic volunteers to 0.3 ppm nitrogen dioxide and to ambient air
pollution. Toxicol Ind Health 5: 1025-1034.
Avol EL: Navidi. WC: Rappaport. EB: Peters. JM. (1998). Acute effects of ambient ozone on asthmatic,
wheezy, and healthy children. (82). Topsfield, MA: Health Effects Institute; Flagship Press.
Ayaz. KL: Csallany. AS. (1978). Long-term NO2 exposure of mice in the presence and absence of vitamin E. II.
Effect of glutathione peroxidase. Arch Environ Occup Health 33: 292-296.
http://dx.doi.org/10.1080/00039896.1978.10667350
Azadniv. M: Utell MJ: Morrow. PE: Gibb. FR: Nichols. J: Roberts. NJ. Jr: Speers. DM: Torres. A: Tsai. Y:
Abraham. MK: Voter. KZ: Frampton. MW. (1998). Effects of nitrogen dioxide exposure on human host
defense. Inhal Toxicol 10: 585-601. http://dx.doi.org/10.1080/089583798197556
Azoulav-Dupuis. E: Torres. M: Soler. P: Moreau. J. (1983). Pulmonary NO2 toxicity in neonate and adult guinea
pigs and rats. Environ Res 30: 322-339. http://dx.doi.org/10.1016/0013-9351(83)90218-9
Babior. BM: Takeuchi. C: Ruedi. J: Gutierrez. A: Wentworth. P. (2003). Investigating antibody-catalyzed ozone
generation by human neutrophils. PNAS 100: 3031-3034. http://dx.doi.org/10.1073/pnas.053Q251100
Baccarelli. A: Zanobetti. A: Martinelli I: Grillo. P: Hou. L: Lanzani. G: Mannucci PM: Bertazzi. PA: Schwartz.
I (2007). Air pollution, smoking, and plasma homocysteine. Environ Health Perspect 115: 176-181.
http://dx.doi.org/10.1289/ehp.9517
Badaloni. C: Ranucci. A: Cesaroni. G: Zanini. G: Vienneau. D: Al-Aidrous. F: De Hoogh. K: Magnani. C:
Forastiere. F: Group. SS. (2013). Air pollution and childhood leukaemia: a nationwide case-control study in
Italy. Occup Environ Med 70: 876-883. http://dx.doi.org/10.1136/oemed-2013-101604
Bagheri Lankarani. N: Kreis. I: Griffiths. DA. (2010). Air pollution effects on peak expiratory flow rate in
children. Iran J Allergy Asthma Immunol 9: 117-126. http://dx.doi.org/09.02/ijaai. 117126
Baja. ES: Schwartz. JD: Wellenius. GA: Coull BA: Zanobetti. A: Vokonas. PS: Suh. HH. (2010). Traffic-related
air pollution and QT interval: Modification by diabetes, obesity, and oxidative stress gene polymorphisms in
the Normative Aging Study. Environ Health Perspect 118: 840-846. http://dx.doi.org/10.1289/ehp.0901396
Baldauf. R: Thoma. E: Hays. M: Shores. R: Kinsev. J: Gullett B: Kimbrough. S: Isakov. V: Long. T: Snow. R:
Khlystov. A: Weinstein. J: Chen. FL: Seila. R: Olson. D: Gilmour. I: Cho. SH: Watkins. N: Rowley. P: Bang.
I (2008a). Traffic and meteorological impacts on near-road air quality: Summary of methods and trends
from the Raleigh near-road study. J Air Waste Manag Assoc 58: 865-878. http://dx.doi.0rg/10.3155/1047-
3289.58.7.865
Baldauf. R: Thoma. E: Khlystov. A: Isakov. V: Bowker. G: Long. T: Snow. R. (2008b). Impacts of noise barriers
on near-road air quality. Atmos Environ 42: 7502-7507. http://dx.doi.0rg/10.1016/i.atmosenv.2008.05.051
R-5
-------
Baldauf. R: Watkins. N: Heist. D: Bailey. C: Rowley. P: Shores. R. (2009). Near-road air quality monitoring:
Factors affecting network design and interpretation of data. Air Qual Atmos Health 2: 1-9.
http://dx.doi.org/10.1007/sll869-009-0028-0
Ballester. F: Estarlich. M: Iniguez. C: Llop. S: Ramon. R: Esplugues. A: Lacasana. M: Rebagliato. M. (2010).
Air pollution exposure during pregnancy and reduced birth size: A prospective birth cohort study in
Valencia, Spain. Environ Health 9:6. http://dx.doi.org/10.1186/1476-069X-9-6
Ballester. F: Rodriguez. P: Iniguez. C: Saez. M: Daponte. A: Galan. I: Taracido. M: Arribas. F: Bellido. J:
Cirarda. FB: Canada. A: Guillen. JJ: Guillen-Grima. F: Lopez. E: Perez-Hoyos. S: Lertxundi. A: Toro. S.
(2006). Air pollution and cardiovascular admisisons association in Spain: Results within the EMECAS
project. J Epidemiol Community Health 60: 328-336. http://dx.doi.org/10.1136/iech.2005.037978
Ballester. F: Tenias. JM: Perez-Hoyos. S. (2001). Air pollution and emergency hospital admissions for
cardiovascular diseases in Valencia, Spain. J Epidemiol Community Health 55: 57-65.
http://dx.doi.org/10.1136/iech.55.L57
Barck. C: Lundahl. J: Hallden. G: Bylin. G. (2005a). Brief exposures to NO2 augment the allergic inflammation
in asthmatics. Environ Res 97: 58-66. http://dx.doi.0rg/10.1016/i.envres.2004.02.009
Barck. C: Lundahl. J: Holmstrom. M: Bylin. G. (2005b). Does nitrogen dioxide affect inflammatory markers
after nasal allergen challenge? Am J Rhinol 19: 560-566.
Barck. C: Sandstrom. T: Lundahl. J: Hallden. G: Svartengren. M: Strand. V: Rak. S: Bylin. G. (2002). Ambient
level of NO2 augments the inflammatory response to inhaled allergen in asthmatics. Respir Med 96: 907-
917. http://dx.doi.org/10.1053/rmed.2002.1374
Barclay. JL: Miller. BG: Dick. S: Dennekamp. M: Ford. I: Hillis. GS: Ayres. JG: Seaton. A. (2009). A panel
study of air pollution in subjects with heart failure: negative results in treated patients. Occup Environ Med
66: 325-334. http://dx.doi.org/10.1136/oem.2008.039032
Barnes. PJ: Liew. FY. (1995). Nitric oxide and asthmatic inflammation [Review]. Immunol Today 16: 128-130.
http://dx.doi.org/10.1016/0167-5699(95)80128-6
Barnett AG: Williams. GM: Schwartz. J: Best TL: Neller. AH: Petroeschevsky. AL: Simpson. RW. (2006). The
effects of air pollution on hospitalizations for cardiovascular disease in elderly people in Australian and New
Zealand cities. Environ Health Perspect 114: 1018-1023. http://dx.doi.org/10.1289/ehp.8674
Barr. RG: Herbstman. J: Speizer. FE: Camargo. CA. Jr. (2002). Validation of serf-reported chronic obstructive
pulmonary disease in a cohort study of nurses. Am J Epidemiol 155: 965-971.
http://dx.doi.org/10.1093/aie/155.10.965
Barraza-Villarreal A: Sunyer. J: Hernandez-Cadena. L: Escamilla-Nunez. MC: Sienra-Monge. JJ: Ramirez-
Aguilar. M: Cortez-Lugo. M: Holguin. F: Diaz-Sanchez. D: Olin. AC: Romieu. I. (2008). Air pollution,
airway inflammation, and lung function in a cohort study of Mexico City schoolchildren. Environ Health
Perspect 116: 832-838. http://dx.doi.org/10.1289/ehp. 10926
Bartell. SM: Longhurst. J: Tjoa. T: Sioutas. C: Delfino. RJ. (2013). Particulate air pollution, ambulatory heart
rate variability, and cardiac arrhythmia in retirement community residents with coronary artery disease.
Environ Health Perspect 121: 1135-1141. http://dx.doi.org/10.1289/ehp.1205914
Earth. PJ: Miiller. B. (1999). Effects of nitrogen dioxide exposure on Clara cell proliferation and morphology.
Pathol Res Pract 195: 487-493. http://dx.doi.org/10.1016/S0344-0338(99)80052-l
Earth. PJ: Muller. B: Wagner. U: Bittinger. A. (1995). Quantitative analysis of parenchymal and vascular
alterations in NO2-induced lung injury in rats. Eur Respir J 8: 1115-1121.
http://dx.doi.org/10.1183/09031936.95.08071115
Bartolomei. V: Sorgel. M: Gligorovski. S: Alvarez. EG: Gandolfo. A: Strekowski. R: Quivet E: Held. A:
Zetzsch. C: Wortham. H. (2014). Formation of indoor nitrous acid (HONO) by light-induced NO2
heterogeneous reactions with white wall paint. Environ Sci Pollut Res Int 21: 9259-9269.
http://dx.doi.org/10.1007/sll356-014-2836-5
R-6
-------
Basagana. X: Aguilera. I: Rivera. M: Agis. D: Foraster. M: Marrugat J: Elosua. R: Kunzli. N. (2013).
Measurement error in epidemiologic studies of air pollution based on land-use regression models. Am J
Epidemiol 178: 1342-1346. http://dx.doi.org/10.1093/aie/kwtl27
Basagafia. X: Rivera. M: Aguilera. I: Agis. D: Bouso. L: Elosua. R: Foraster. M: de Nazelle. A: Nieuwenhuijsen.
M: Vila. J: Kiinzli. N. (2012). Effect of the number of measurement sites on land use regression models in
estimating local air pollution. Atmos Environ 54: 634-642. http://dx.doi.0rg/10.1016/i.atmosenv.2012.01.064
Bastacky. J: Lee. CY: Goerke. J: Koushafar. H: Yager. D: Kenaga. L: Speed. TP: Chen. Y: Clements. JA.
(1995). Alveolar lining layer is thin and continuous: Low-temperature scanning electron microscopy of rat
lung. J Appl Physiol 79: 1615-1628.
Bastain. T: Islam T: Berhane. K: Mcconnell R: Rappaport E: Salam. M: Linn. W: Avol. E: Zhang. Y:
Gilliland. F. (2011). Exhaled nitric oxide, susceptibility and new-onset asthma in the Children's Health
Study. EurRespir J 37: 523-531. http://dx.doi.org/10.1183/09031936.00021210
Basu. R: Malig. B: Ostro. B. (2010). High ambient temperature and the risk of preterm delivery. Am J Epidemiol
172: 1108-1117. http://dx.doi.org/10.1093/aie/kwql70
Bates. DV: Macklem. PT: Christie. RV. (1971). Respiratory function in disease: An introduction to the
integrated study of the lung (2nd ed.). Philadelphia, PA: W.B. Saunders Company.
Bateson. TF: Coull. BA: Hubbell B: Ito. K: Jerrett M: Lumlev. T: Thomas. D: Vedal S: Ross. M. (2007). Panel
discussion review: Session three - issues involved in interpretation of epidemiologic analyses - statistical
modeling. J Expo Sci Environ Epidemiol 17: S90-S96. http://dx.doi.org/10.1038/sj.jes.750Q631
Bauer. MA: Utell MJ: Morrow. PE: Speers. DM: Gibb. FR. (1986). Inhalation of 0.30 ppm nitrogen dioxide
potentiates exercise-induced bronchospasm in asthmatics. Am Rev Respir Dis 134: 1203-1208.
Bautista. AP: Eisenlohr. CP: Lanz. MJ. (2011). Nasal nitric oxide and nasal eosinophils decrease with
levocetirizine in subjects with perennial allergic rhinitis. Am J Rhinol Allergy 25: 383-387.
http://dx.doi.org/10.2500/aira.2011.25.3668
Baxter. LK: Duvall. RM: Sacks. J. (2013). Examining the effects of air pollution composition on within region
differences inPM(2.5) mortality risk estimates. J Expo Sci Environ Epidemiol 23: 457-465.
http://dx.doi.org/10.1038/ies.2012.114
Beaupre. A: Orehek. J. (1982). Factors influencing the bronchodilator effect of a deep inspiration in asthmatic
patients with provoked bronchoconstriction. Thorax 37: 124-128. http://dx.doi.0rg/10.1136/thx.37.2.124
Beaven. MA. (2009). Our perception of the mast cell from Paul Ehrlich to now. Eur J Immunol 39: 11-25.
http://dx.doi.org/10.1002/eji.200838899
Becerra. TA: Wilhelm M: Olsen. J: Cockburn. M: Ritz. B. (2013). Ambient air pollution and autism in Los
Angeles County, California. Environ Health Perspect 121: 380-386. http://dx.doi.org/10.1289/ehp. 1205827
Beckerman. B: Jerrett. M: Brook. JR: Verma. DK: Arain. MA: Finkelstein. MM. (2008). Correlation of nitrogen
dioxide with other traffic pollutants near a major expressway. Atmos Environ 42: 275-290.
http://dx.doi.0rg/10.1016/j.atmosenv.2007.09.042
Beckerman. BS: Jerrett. M: Finkelstein. M: Kanaroglou. P: Brook. JR: Arain. MA: Sears. MR: Stieb. D: Balmes.
J: Chapman. K. (2012). The association between chronic exposure to traffic-related air pollution and
ischemic heart disease. J Toxicol Environ Health A 75: 402-411.
http://dx.doi.org/10.1080/15287394.2012.670899
Beckerman. BS: Jerrett. M: Martin. RV: van Donkelaar. A: Ross. Z. ev: Burnett. RT. (2013a). Application of the
deletion/substitution/addition algorithm to selecting land use regression models for interpolating air pollution
measurements in California. Atmos Environ 77: 172-177. http://dx.doi.0rg/10.1016/i.atmosenv.2013.04.024
Beckerman. BS: Jerrett. M: Serre. M: Martin. RV: Lee. S: van Donkelaar. A: Ross. Z. ev: Su. J: Burnett RT.
(2013b). A hybrid approach to estimating national scale spatiotemporal variability of PM2.5 in the
contiguous United States. Environ Sci Technol 47: 7233-7241. http://dx.doi.org/10.1021/es400039u
R-7
-------
Becquemin. MM: Bertholon. JF: Bouchikhi. A: Malarbet JL: Roy. M. (1999). Oronasal ventilation partitioning
in adults and children: Effect on aerosol deposition in airways. Radiat Prot Dosimetry 81: 221-228.
Beelen. R: Hoek. G: Houthuijs. D: van den Brandt. PA: Goldbohm. RA: Fischer. P: Schouten. LJ: Armstrong. B:
Brunekreef. B. (2009). The joint association of air pollution and noise from road traffic with cardiovascular
mortality in a cohort study. Occup Environ Med 66: 243-250. http://dx.doi.org/10.1136/oem.2008.042358
Beelen. R: Hoek. G: van den Brandt. PA: Goldbohm. RA: Fischer. P: Schouten. LJ: Armstrong. B: Brunekreef.
B_. (2008a). Long-term exposure to traffic-related air pollution and lung cancer risk. Epidemiology 19: 702-
710. http://dx.doi.org/10.1097/EDE.Ob013e318181b3ca
Beelen. R: Hoek. G: van den Brandt PA: Goldbohm RA: Fischer. P: Schouten. LJ: Jerrett M: Hughes. E:
Armstrong. B: Brunekreef. B. (2008b). Long-term effects of traffic-related air pollution on mortality in a
Dutch cohort (NLCS-AIR study). Environ Health Perspect 116: 196-202.
http://dx.doi.org/10.1289/ehp. 10767
Beelen. R: Hoek. G: Vienneau. D: Eeftens. M: Dimakopoulou. K: Pedeli. X: Tsai. MY. i: Kuenzli. N:
Schikowskj T: Marcon. A: Eriksen. KT: Raaschou-Nielsen. O. le: Stephanou. E: Patelarou. E: Lanki. T: Yli-
Toumi. T: Declercq. C: Falq. G: Stempfelet M: Birk. M: Cyrys. J: vonKlot. S: Nador. G: Varro. MJ:
Dedele. A: Grazuleviciene. R: Moelter. A: Lindlev. S: Madsen. C: Cesaroni. G: Ranzi. A: Badaloni. C:
Hoffmann. B: Nonnemacher. M: Kraemer. U: Kuhlbusch. T: Cirach. M: de Nazelle. A: Nieuwenhuijsen. M:
Bellander. T. om: Korek. M: Olsson. D: Stromgren. M: Dons. E. vi: Jerrett M: Fischer. P: Wang. M:
Brunekreef. B: de Hoogh. K. (2013). Development of NO2 and NOx land use regression models for
estimating air pollution exposure in 36 study areas in Europe - The ESCAPE project. Atmos Environ 72: 10-
23. http://dx.doi.0rg/10.1016/i.atmosenv.2013.02.037
Beelen. R: Raaschou-Nielsen. O: Stafoggia. M: Andersen. ZJ: Weinmayr. G: Hoffmann. B: Wolf. K: Samoli. E:
Fischer. P: Nieuwenhuijsen. M: Vineis. P: Xun. WW: Katsouvanni. K: Dimakopoulou. K: Oudin. A:
Forsberg. B: Modig. L: Havulinna. AS: Lanki. T: Turunen. A: Oftedal B: Nystad. W: Nafstad. P: De Faire.
U: Pedersen. NL: Ostenson. CG: Fratiglioni. L: Penell. J: Korek. M: Pershagen. G: Eriksen. KT: Overvad. K:
Ellermann. T: Eeftens. M: Peeters. PH: Meliefste. K: Wang. M: Bueno-De-Mesquita. B: Sugiri. D: Kramer.
U: Heinrich. J: de Hoogh. K: Key. T: Peters. A: Hampel. R: Concin. H: Nagel. G: Ineichen. A: Schaffner. E:
Probst-Hensch. N: Kiinzli. N: Schindler. C: Schikowski. T: Adam. M: Phuleria. H: Vilier. A: Clavel-
Chapelon. F: Declercq. C: Grioni. S: Krogh. V: Tsai. MY: Ricceri. F: Sacerdote. C: Galassi. C: Migliore. E:
Ranzi. A: Cesaroni. G: Badaloni. C: Forastiere. F: Tamavo. I: Amiano. P: Dorronsoro. M: Katsoulis. M:
Trichopoulou. A: Brunekreef. B: Hoek. G. (2014a). Effects of long-term exposure to air pollution on natural-
cause mortality: an analysis of 22 European cohorts within the multicentre ESCAPE project. Lancet 383:
785-795. http://dx.doi.org/10.1016/S0140-6736(13)62158-3
Beelen. R: Stafoggia. M: Raaschou-Nielsen. O: Andersen. ZJ: Xun. WW: Katsouvanni. K: Dimakopoulou. K:
Brunekreef. B: Weinmayr. G: Hoffmann. B: Wolf. K: Samoli. E: Houthuijs. D: Nieuwenhuijsen. M: Oudin.
A: Forsberg. B: Olsson. D: Salomaa. V: Lanki. T: Yli-Tuomi. T: Oftedal. B: Aamodt. G: Nafstad. P: De
Faire. U: Pedersen. NL: Ostenson. CG: Fratiglioni. L: Penell. J: Korek. M: Pyko. A: Eriksen. KT:
Tjonneland. A: Becker. T: Eeftens. M: Bots. M: Meliefste. K: Wang. M: Bueno-De-Mesquita. B: Sugiri. D:
Kramer. U: Heinrich. J: de Hoogh. K: Key. T: Peters. A: Cyrys. J: Concin. H: Nagel. G: Ineichen. A:
Schaffner. E: Probst-Hensch. N: Dratva. J: Ducret-Stich. R: Vilier. A: Clavel-Chapelon. F: Stempfelet. M:
Grioni. S: Krogh. V: Tsai. MY: Marcon. A: Ricceri. F: Sacerdote. C: Galassi. C: Migliore. E: Ranzi. A:
Cesaroni. G: Badaloni. C: Forastiere. F: Tamavo. I: Amiano. P: Dorronsoro. M: Katsoulis. M: Trichopoulou.
A: Vineis. P: Hoek. G. (2014b). Long-term exposure to air pollution and cardiovascular mortality: An
analysis of 22 European cohorts. Epidemiology 25: 368-378.
http://dx.doi.org/10.1097/EDE.0000000000000076
Beelen. R: Voogt. M: Duyzer. J: Zandveld. P: Hoek. G. (2010). Comparison of the performances of land use
regression modelling and dispersion modelling in estimating small-scale variations in long-term air pollution
concentrations in a Dutch urban area. Atmos Environ 44: 4614-4621.
http://dx.doi.0rg/10.1016/i.atmosenv.2010.08.005
Beevers. SD: Carslaw. DC. (2005). The impact of congestion charging on vehicle emissions in London. Atmos
Environ 39: 1-5. http://dx.doi.0rg/10.1016/i.atmosenv.2004.10.001
R-8
-------
Beevers. SD: Kitwiroon. N: Williams. ML: Carslaw. DC. (2012). One way coupling of CMAQ and a road source
dispersion model for fine scale air pollution predictions. Atmos Environ 59: 47-58.
http://dx.doi.0rg/10.1016/i.atmosenv.2012.05.034
Beevers. SD: Kitwiroon. N: Williams. ML: Kelly. FJ: Ross Anderson. H: Carslaw. DC. (2013). Air pollution
dispersion models for human exposure predictions in London. J Expo Sci Environ Epidemiol 23: 647-653.
http://dx.doi.0rg/10.1038/ies.2013.6
Belanger. K: Holford. TR: Gent JF: Hill ME: Kezik. JM: Leaderer. BP. (2013). Household levels of nitrogen
dioxide and pediatric asthma severity. Epidemiology 24: 320-330.
http://dx.doi.org/10.1097/EDE.Ob013e318280e2ac
Bell ML: Dominici F: Samet JM. (2005). A meta-analysis of time-series studies of ozone and mortality with
comparison to the national morbidity, mortality, and air pollution study. Epidemiology 16: 436-445.
http://dx.doi.org/10.1097/01.ede.0000165817.40152.85
Bell ML: Ebisu. K: Belanger. K. (2007). Ambient air pollution and low birth weight in Connecticut and
Massachusetts. Environ Health Perspect 115: 1118-1124. http://dx.doi.org/10.1289/ehp.9759
Bell ML: Lew. JK: Lin. Z. (2008). The effect of sandstorms and air pollution on cause-specific hospital
admissions in Taipei, Taiwan. Occup Environ Med 65: 104-111. http://dx.doi.org/10.1136/oem.2006.031500
Bell. S: Ashenden. TW. (1997). Spatial and temporal variation in nitrogen dioxide pollution adjacent to rural
roads. Water Air Soil Pollut 95: 87-98. http://dx.doi.Org/10.1023/A:1026478127687
Bellander. T: Wichmann. J: Lind. T. (2012). Individual exposure to NO2 in relation to spatial and temporal
exposure indices in Stockholm, Sweden: The INDEX study. PLoS ONE 7: e39536.
http://dx.doi.org/10.1371/iournal.pone.0039536
Bellini. P: Baccini. M: Biggeri. A: Terracini. B. (2007). The meta-analysis of the Italian studies on short-term
effects of air pollution (MIS A): Old and new issues on the interpretation of the statistical evidences.
Environmetrics 18: 219-229. http://dx.doi.org/10.1002/env.836
Belvisi. MG: Stretton. CD: Miura. M: Verleden. GM: Tadjkarimi. S: Yacoub. MH: Barnes. PJ. (1992).
Inhibitory NANC nerves in human trachea! smooth muscle: A quest for the neurotransmitter. J Appl Physiol
73:2505-2510.
Bennett WD: Zeman. KL: Jarabek. AM. (2008). Nasal contribution to breathing and fine particle deposition in
children versus adults. J Toxicol Environ Health A 71: 227-237.
http://dx.doi.org/10.1080/15287390701598200
Benson. PE. (1984). CALINE4 - a dispersion model for predicting air pollutant concentrations near roadways.
(FHWA/CA/TL-84/15). Sacramento, California: California Department of Transportation, Office of
Transportation Laboratory.
https://ntrl.ntis.gov/NTRL/dashboard/searchResults.xhtml?searchQuery=PB85211498
Benson. PE. (1992). A review of the development and application of the CALINE3 and CALINE4 models.
Atmos Environ 26: 379-390. http://dx.doi.org/10.1016/0957-1272(92)90013-I
Bentaveb. M: Helmer. C: Raherison. C: Dartigues. JF: Tessier. JF: Annesi-Maesano. I. (2010). Bronchitis-like
symptoms and proximity air pollution in French elderly. Respir Med 104: 880-888.
http://dx.doi.0rg/10.1016/i.rmed.2010.01.004
Bentaveb. M: Stempfelet M: Wagner. V: Zins. M: Bonenfant S: Songeur. C: Sanchez. O: Rosso. A: Brulfert. G:
Rios. I: Chaxel. E: Virga. J: Armengaud. A: Rossello. P: Riviere. E: Bernard. M: Vasbien. F: Deprost. R.
(2014). Retrospective modeling outdoor air pollution at a fine spatial scale in France, 1989-2008. Atmos
Environ 92: 267-279. http://dx.doi.0rg/10.1016/j.atmosenv.2014.04.019
Berglind. N: Bellander. T: Forastiere. F: vonKlot. S: Aalto. P: Elosua. R: Kulmala. M: Lanki. T: Lowel. H:
Peters. A: Picciotto. S: Salomaa. V: Stafoggia. M: Sunyer. J: Nyberg. F. (2009). Ambient air pollution and
daily mortality among survivors of myocardial infarction. Epidemiology 20: 110-118.
http://dx.doi.org/10.1097/EDE.Ob013e3181878b50
R-9
-------
Berhane. K: Zhang. Y: Linn. WS: Rappaport. EB: Bastain. TM: Salam. MT: Islam. T: Lurmann. F: Gilliland.
FD. (2011). The effect of ambient air pollution on exhaled nitric oxide in the Children's Health Study. Eur
RespirJ37: 1029-1036. http://dx.doi.org/10.1183/09031936.00081410
Berhane. K: Zhang. Y: Salam. MT: Eckel SP: Linn. WS: Rappaport. EB: Bastain. TM: Lurmann. F: Gilliland.
FD. (2014). Longitudinal effects of air pollution on exhaled nitric oxide: the Children's Health Study. Occup
Environ Med 71: 507-513. http://dx.doi.org/10.1136/oemed-2013-101874
Berkowicz. R: Ketzel. M: Jensen. SS: Hvidberg. M: Raaschou-Nielsen. O. (2008). Evaluation and application of
OSPM for traffic pollution assessment for a large number of street locations. Environ Modell Softw 23: 296-
303. http://dx.doi.0rg/10.1016/i.envsoft.2007.04.007
Bernhard. W: Haslam. PL: Floros. J. (2004). From birds to humans: New concepts on airways relative to
alveolar surfactant [Review]. Am J Respir Cell Mol Biol 30: 6-11. http://dx.doi.org/10.1165/rcmb.2003-
0158TR
Berrocal. VJ: Gelfand. AE: Holland. DM. (2010a). A bivariate space-time downscaler under space and time
misalignment. Ann Appl Stat 4: 1942-1975. http://dx.doi.org/10.1214/10-AOAS351
Berrocal VJ: Gelfand. AE: Holland. DM. (2010b). A spatio-temporal downscaler for output from numerical
models. J Agric Biol Environ Stat 15: 176-197. http://dx.doi.org/10.1007/sl3253-009-0004-z
Berrocal VJ: Gelfand. AE: Holland. DM. (2012). Space-time data fusion under error in computer model output:
An application to modeling air quality. Biometrics 68: 837-848. http://dx.doi.0rg/10.l lll/i.1541-
0420.2011.01725.x
Bevelander. M: Mavette. J: Whittaker. LA: Paveglio. SA: Jones. CC: Robbins. J: Hemenwav. D: Akira. S:
Uematsu. S: Poynter. ME. (2007). Nitrogen dioxide promotes allergic sensitization to inhaled antigen. J
Immunol 179: 3680-3688.
Bhaskar. KR: O'Sullivan. DP: Seltzer. J: Rossing. TH: Drazen. JM: Reid. LM. (1985). Density gradient study of
bronchial mucus aspirates from healthy volunteers (smokers and nonsmokers) and from patients with
tracheostomy. Exp Lung Res 9: 289-308. http://dx.doi.org/10.3109/01902148509057529
Bhaskaran. K: Hajat. S: Armstrong. B: Haines. A: Herrett. E: Wilkinson. P: Smeeth. L. (2011). The effects of
hourly differences in air pollution on the risk of myocardial infarction: case crossover analysis of the MINAP
database. B M J 343: d5531. http://dx.doi.org/10.1136/bmi.d5531
Bidani. A: Postlethwait. EM. (1998). Kinetic determinants of inhaled reactive gas absorption. In MP Hlastala;
HT Robertson (Eds.), Complexity in structure and function of the lung (pp. 243-296). New York, NY:
Marcell Dekker Inc.
Biggeri. A: Baccini. M: Bellini. P: Terracini. B. (2005). Meta-analysis of the Italian studies of short-term effects
of air pollution (MISA), 1990-1999. Int J Occup Environ Health 11: 107-122.
http://dx.doi.0rg/10.1179/oeh.2005.ll.l.107
Bignal. KL: Ashmore. MR: Headlev. AD: Stewart. K: Weigert. K. (2007). Ecological impacts of air pollution
from road transport on local vegetation. Appl Geochem 22: 1265-1271.
http://dx.doi.0rg/10.1016/i.apgeochem.2007.03.017
Bilenko. N: Rossem. LV: Brunekreef. B: Beelen. R: Eeftens. M: Hoek. G: Houthuijs. D: de Jongste. JC:
Kempen. EV: Koppelman. GH: Meliefste. K: Oldenwening. M: Smit. HA: Wijga. AH: Gehring. U. (2015).
Traffic-related air pollution and noise and children's blood pressure: results from the PIAMA birth cohort
study. Euro J Prev Cardio 22: 4-12. http://dx.doi.org/10.1177/2047487313505821
Bind. MA: Baccarelli. A: Zanobetti. A: Tarantini. L: Suh. H: Vokonas. P: Schwartz. J. (2012). Air pollution and
markers of coagulation, inflammation, and endothelial function: Associations and epigene-environment
interactions in an elderly cohort. Epidemiology 23: 332-340.
http://dx.doi.org/10.1097/EDE.Ob013e31824523fO
Bishaw. A: Fontenot K. (2014). Poverty: 2012 and 2013 - American community survey briefs. (ACSBR/13-01).
Suitland, MD: U.S. Census Bureau.
https://www.census.gov/content/dam/Census/library/publications/2014/acs/acsbrl3-01.pdf
R-10
-------
Blackwell PL: Lucas. JW: Clarke. TC. (2014). Summary health statistics for U.S. adults: National health
interview survey, 2012. In Vital and health statistics. Hyattsville, MD: National Center for Health Statistics,
U.S Department of Health and Human Services, http://www. cdc.gov/nchs/data/series/sr 10/sr 10 260.pdf
Blair. WH: Henry. MC: Ehrlich. R. (1969). Chronic toxicity of nitrogen dioxide: II. Effect on histopathology of
lung tissue. Arch Environ Health 18: 186-192. http://dx.doi.org/10.1080/00039896.1969.10665391
Blomberg. A: Krishna. MT: Bocchino. V: Biscione. GL: Shute. JK: Kelly. FJ: Frew. AJ: Holgate. ST:
Sandstrom. T. (1997). The inflammatory effects of 2 ppmNO2 on the airways of healthy subjects. Am J
Respir Crit Care Med 156: 418-424. http://dx.doi.0rg/10.1164/airccm.156.2.9612042
Blomberg. A: Krishna. MT: Helledav. R: Soderberg. M: Ledin. MC: Kelly. FJ: Frew. AJ: Holgate. ST:
Sandstrom. T. (1999). Persistent airway inflammation but accomodated antioxidant and lung function
responses after repeated daily exposure to nitrogen dioxide. Am J Respir Crit Care Med 159: 536-543.
http://dx.doi.org/10.1164/airccm. 159.2.9711068
Blood. AB: Schroeder. HJ: Terry. MH: Merrill-Henry. J: Bragg. SL: Vrancken. K: Liu. T: Herring. JL: Sowers.
LC: Wilson. SM: Power. GG. (2011). Inhaled nitrite reverses hemolysis-induced pulmonary vasoconstriction
in newborn lambs without blood participation. Circulation 123: 605-612.
http://dx.doi.0rg/10.l 161/CIRCULATIONAHA. 110.001073
Bloom. B: Jones. LI: Freeman. G. (2013). Summary health statistics for U.S. children: National health interview
survey, 2012. In Vital and health statistics. Hyattsville, MD: National Center for Health Statistics, U.S
Department of Health and Human Services, http://www.cdc.gov/nchs/data/series/sr 10/srlO 258.pdf
Blount RJ: Djawe. K: Daly. KR: Jarlsberg. LG: Fong. S: Balmes. J: Miller. RF: Walzer. PD: Huang. L. (2013).
Ambient air pollution associated with suppressed serologic responses to Pneumocystis jirovecii in a
prospective cohort of HIV-infected patients with Pneumocystis pneumonia. PLoS ONE 8: e80795.
http://dx.doi.org/10.1371/iournal.pone.0080795
Bobak. M. (2000). Outdoor air pollution, low birth weight, and prematurity. Environ Health Perspect 108: 173-
176. http://dx.doi.org/10.2307/3454517
Boersma. KF: Eskes. HJ: Brinksma. EJ. (2004). Error analysis for tropospheric NO2 retrieval from space. J
Geophys Res 109: D04311. http://dx.doi.org/10.1029/2003JD003962
Boezen. M: Schouten. J: Rijcken. B: Vonk. J: Gerritsen. J: Van Per Zee. S: Hoek. G: Brunekreef. B: Postma. D.
(1998). Peak expiratory flow variability, bronchial responsiveness, and susceptibility to ambient air pollution
in adults. Am J Respir Crit Care Med 158: 1848-1854. http://dx.doi.0rg/10.1164/airccm.158.6.9804072
Boggia. B: Carbone. U: Farinaro. E: Zarrilli. S: Lombardi G: Colao. A: De Rosa. N: De Rosa. M. (2009).
Effects of working posture and exposure to traffic pollutants on sperm quality. J Endocrinol Invest 32: 430-
434. http://dx.doi.org/10.1007/BF03346481
Bonacci G: Baker. PR: Salvatore. S. R.: Shores. D: Khoo. NK: Koenitzer. J. R.: Vitturi. DA: Woodcock. S. R.:
Golin-Bisello. F: Cole. MP: Watkins. S: St Croix. C: Batthvany. CI: Freeman. BA: Schopfer. FJ. (2012).
Conjugated linoleic acid is a preferential substrate for fatty acid nitration. J Biol Chem 287: 44071-44082.
http://dx.doi.org/10.1074/ibc.M112.401356
Bonacci. G: Schopfer. FJ: Batthyany. CI: Rudolph. TK: Rudolph. V: Khoo. NK: Kellev. EE: Freeman. BA.
(2011). Electrophilic fatty acids regulate matrix metalloproteinase activity and expression. J Biol Chem 286:
16074-16081. http://dx.doi.org/10.1074/ibc.Ml 11.225029
Boogaard. H: Fischer. PH: Janssen. NA: Kos. GP: Weijers. EP: Cassee. FR: vanderZee. SC: de Hartog. JJ:
Meliefste. K: Wang. M: Brunekreef. B: Hoek. G. (2013). Respiratory effects of a reduction in outdoor air
pollution concentrations. Epidemiology 24: 753-761. http://dx.doi.org/10.1097/EDE.Ob013e31829el639
Borland. CDR: Dunningham. H: Bottrill. F: Vuylsteke. A: Yilmaz. C: Dane. DM: Hsia. CCW. (2010).
Significant blood resistance to nitric oxide transfer in the lung. J Appl Physiol 108: 1052-1060.
http://dx.doi.org/10.1152/iapplphysiol.00904.2009
R-ll
-------
Braback. L: Forsberg. B. (2009). Does traffic exhaust contribute to the development of asthma and allergic
sensitization in children: Findings from recent cohort studies [Review]. Environ Health 8: 17.
http://dx.doi.org/10.1186/1476-069X-8-17
Brahmajothi MV: Mason. SN: Whorton. AR: Mcmahon. TJ: Auten. RL. (2010). Transport rather than diffusion-
dependent route for nitric oxide gas activity in alveolar epithelium. Free Radic Biol Med 49: 294-300.
http://dx.doi.0rg/10.1016/i.freeradbiomed.2010.04.020
Brauer. M: Hoek. G: Smit HA: de Jongste. JC: Gerritsen. J: Postma. PS: Kerkhof. M: Brunekreef. B. (2007).
Air pollution and development of asthma, allergy and infections in a birth cohort. Eur Respir J 29: 879-888.
http://dx.doi.org/10.1183/09031936.00083406
Brauer. M: Hoek. G: van Vliet P: Meliefste. K: Fischer. P: Gehring. U: Heinrich. J: Cyrys. J: Bellander. T:
Lewne. M: Brunekreef. B. (2003). Estimating long-term average paniculate air pollution concentrations:
Application of traffic indicators and geographic information systems. Epidemiology 14: 228-239.
http://dx.doi.org/10.1097/00001648-200303000-00019
Brauer. M: Hoek. G: Van Vliet P: Meliefste. K: Fischer. PH: Wijga. A: Koopman. LP: Neijens. HJ: Gerritsen. J:
Kerkhof. M: Heinrich. J: Bellander. T: Brunekreef. B. (2002). Air pollution from traffic and the development
of respiratory infections and asthmatic and allergic symptoms in children. Am J Respir Crit Care Med 166:
1092-1098. http://dx.doi.org/10.1164/rccm.200108-007OC
Brauer. M: Koutrakis. P: Keeler. GJ: Spengler. JD. (1991). Indoor and outdoor concentrations of inorganic acidic
aerosols and gases. J Air Waste Manag Assoc 41: 171-181.
http://dx.doi.org/10.1080/10473289.1991.10466834
Brauer. M: Lencar. C: Tamburic. L: Koehoorn. M: Demers. P: Karr. C. (2008). A cohort study of traffic-related
air pollution impacts on birth outcomes. Environ Health Perspect 116: 680-686.
http://dx.doi.org/10.1289/ehp. 10952
Brauer. M: Ryan. PB: Suh. HH: Koutrakis. P: Spengler. JD: Leslie. NP: Billick. IH. (1990). Measurements of
nitrous acid inside two research houses. Environ Sci Technol 24: 1521-1527.
http://dx.doi.org/10.1021/es00080a011
Breton. CV: Salam MT: Vora. H: Gauderman. WJ: Gilliland. FD. (2011). Genetic variation in the glutathione
synthesis pathway, air pollution, and children's lung function growth. Am J Respir Crit Care Med 183: 243-
248. http://dx.doi.org/10.1164/rccm.201006-0849OC
Breton. CV: Wang. X: Mack. WJ: Berhane. K: Lopez. M: Islam TS: Feng. M: Lurmann. F: McConnell. R:
Hodis. HN: Kiinzli. N: Avol. E. (2012). Childhood air pollutant exposure and carotid artery intima-media
thickness in young adults. Circulation 126: 1614-1620.
http://dx.doi.org/10.1161/CIRCULATIONAHA. 112.096164
Britter. RE: Hanna. SR. (2003). Flow and dispersion in urban areas. Annu Rev Fluid Mech 35: 469-496.
http://dx.doi.org/10.1146/annurev.fluid.35.101101.161147
Brixev. LA: Heist. DK: Richmond-Bryant J: Bowker. GE: Perry. SG: Wiener. RW. (2009). The effect of a tall
tower on flow and dispersion through a model urban neighborhood: Part 2: Pollutant dispersion. J Environ
Monit 11: 2171-2179. http://dx.doi.org/10.1039/B907137g
Brochu. P: Brodeur. J: Krishnan. K. (2011). Derivation of physiological inhalation rates in children, adults, and
elderly based on nighttime and daytime respiratory parameters. Inhal Toxicol 23: 74-94.
http://dx.doi.org/10.3109/08958378.2010.543439
Brodv. DJ: Zhang. X: Kit BK: Dillon. CF. (2013). Reference values and factors associated with exhaled nitric
oxide: U.S. youth and adults. Respir Med 107: 1682-1691. http://dx.doi.0rg/10.1016/i.rmed.2013.07.006
Brook. JR: Burnett. RT: Dann. TF: Cakmak. S: Goldberg. MS: Fan. X: Wheeler. AJ. (2007). Further
interpretation of the acute effect of nitrogen dioxide observed in Canadian time-series studies. J Expo Sci
EnvironEpidemiol 17: S36-S44. http://dx.doi.org/10.1038/sj.jes.7500626
R-12
-------
Brook. RD: Jerreft. M: Brook. JR: Bard. RL: Finkelstein. MM. (2008). The relationship between diabetes
mellitus and traffic-related air pollution. J Occup Environ Med 50: 32-38.
http://dx.doi.org/10.1097/JOM.Ob013e31815dba70
Brown. JS. (2015). Nitrogen dioxide exposure and airway responsiveness in individuals with asthma. Inhal
Toxicol 27: 1-14. http://dx.doi.org/10.3109/08958378.2014.979960
Brown. KW: Sarnat JA: Suh. HH: Coull BA: Koutrakis. P. (2009). Factors influencing relationships between
personal and ambient concentrations of gaseous and paniculate pollutants. Sci Total Environ 407: 3754-
3765. http://dx.doi.0rg/10.1016/i.scitotenv.2009.02.016
Brown. SS: Dube. WP: Karamchandani P: Yarwood. G: Peischl. J: Ryerson. TB: Neuman. JA: Nowak. JB:
Hollowav. JS: Washenfelder. RA: Brock. CA: Frost GJ: Trainer. M: Parrish. DP: Fehsenfeld. FC:
Ravishankara. AR. (2012). Effects of NO x control and plume mixing on nighttime chemical processing of
plumes from coal-fired power plants. J Geophys Res 117: D07304. http://dx.doi.org/10.1029/2011JD016954
Brown. SS: Stutz. J. (2012). Nighttime radical observations and chemistry. Chem Soc Rev 41: 6405-6447.
http://dx.doi.org/10.1039/c2cs35181a
Brunekreef B: Beelen. R: Hoek. G: Schouten. L: Bausch-Goldbohm S: Fischer. P: Armstrong. B: Hughes. E:
Jerrett. M: van den Brandt. P. (2009). Effects of long-term exposure to traffic-related air pollution on
respiratory and cardiovascular mortality in the Netherlands: The NLCS-AIR Study. (139). Boston, MA:
Health Effects Institute. http://www.n65.nl/NCLS-AIR-Studv-2009.pdf
Brusasco. V: Crimi. E: Gianiorio. P: Lantero. S: Rossi. GA. (1990). Allergen-induced increase in airway
responsiveness and inflammation in mild asthma. J Appl Physiol 69: 2209-2214.
Briiske. I. (2014). Email from Dr. Irene Briiske to Molini Patel, cc Dr. Regina Hampel, Dr. Kathrin Wolf. WG:
question re: 2010 Inhalational toxicology paper (With attachments). Available online
Briiske. I: Hampel R: Baumgartner. Z: Riickerl R: Graven. S: Koenig. W: Peters. A: Schneider. A. (2011).
Ambient air pollution and lipoprotein-associated phospholipase A2 in survivors of myocardial infarction.
Environ Health Perspect 119: 921-926. http://dx.doi.org/10.1289/ehp.1002681
Bruske. I: Hampel. R: Socher. MM: Ruckerl. R: Schneider. A: Heinrich. J: Oberdorster. G: Wichmann. HE:
Peters. A. (2010). Impact of ambient air pollution on the differential white blood cell count in patients with
chronic pulmonary disease. Inhal Toxicol 22: 245-252. http://dx.doi.org/10.3109/08958370903207274
Bryan. NS. (2006). Nitrite in nitric oxide biology: Cause or consequence? A systems-based review [Review].
Free Radic Biol Med 41: 691-701. http://dx.doi.0rg/10.1016/j.freeradbiomed.2006.05.019
Bryan. NS: Fernandez. BO: Bauer. SM: Garcia-Saura. MF: Milsom AB: Rassaf T: Malonev. RE: Bharti. A:
Rodriguez. J: Feelisch. M. (2005). Nitrite is a signaling molecule and regulator of gene expression in
mammalian tissues. Nat Chem Biol 1: 290-297. http://dx.doi.org/10.1038/nchembio734
Bucsela. EJ: Krotkov. NA: Celarier. EA: Lamsal LN: Swartz. WH: Bhartia. PK: Boersma. KF: Veefkind. JP:
Gleason. JF: Pickering. KE. (2013). A new stratospheric and tropospheric NO2 retrieval algorithm for nadir-
viewing satellite instruments: Applications to OMI. Atmos Meas Tech Disc 6: 1361-1407.
http://dx.doi.org/10.5194/amtd-6-1361-2013
Burke. JM: Zufall MJ: Ozkaynak. H. (2001). A population exposure model for paniculate matter: Case study
results forPM2.5 in Philadelphia, PA. J Expo Anal Environ Epidemiol 11: 470-489.
Burling. IR: Yokelson. RJ: Griffith. DWT: Johnson. TJ: Veres. P: Roberts. JM: Warneke. C: Urbanski. SP:
Reardon. J: Weise. PR: Hao. WM: De Gouw. J. (2010). Laboratory measurements of trace gas emissions
from biomass burning of fuel types from the southeastern and southwestern United States. Atmos Chem Phys
10: 11115-11130. http://dx.doi.0rg/10.5194/acp-10-l 1115-2010
Burnett. RT: Brook. J: Dann. T: Delocla. C: Philips. O: Cakmak. S: Vincent. R: Goldberg. MS: Krewski. D.
(2000). Association between paniculate- and gas-phase components of urban air pollution and daily mortality
in eight Canadian cities. Inhal Toxicol 12: 15-39. http://dx.doi.org/10.1080/08958370050164851
R-13
-------
Burnett. RT: Smith-Doiron. M: Stieb. D: Cakmak. S: Brook. JR. (1999). Effects of paniculate and gaseous air
pollution on cardiorespiratory hospitalizations. Arch Environ Health 54: 130-139.
http://dx.doi.org/10.1080/00039899909602248
Burnett. RT: Smith-Doiron. M: Stieb. D: Raizenne. ME: Brook. JR: Dales. RE: Leech. JA: Cakmak. S: Krewski.
IX (2001). Association between ozone and hospitalization for acute respiratory diseases in children less than
2 years of age. Am J Epidemiol 153: 444-452. http://dx.doi.0rg/10.1093/aie/153.5.444
Burnett. RT: Stieb. D: Brook. JR: Cakmak. S: Dales. R: Raizenne. M: Vincent. R: Dann. T. (2004). Associations
between short-term changes in nitrogen dioxide and mortality in Canadian cities. Arch Environ Occup Health
59: 228-236. http://dx.doi.Org/10.3200/AEOH.59.5.228-236
Burnev. PG: Laitinen. LA: Perdrizet S: Huckauf. H: Tattersfield. AE: China S: Poisson. N: Heeren. A: Britton.
JR: Jones. T. (1989). Validity and repeatability of the IUATLD (1984) Bronchial Symptoms Questionnaire:
an international comparison. Eur Respir J 2: 940-945.
Burr. ML. (1992). Diagnosing asthma by questionnaire in epidemiological surveys [Editorial]. Clin Exp Allergy
22: 509-510. http://dx.doi.org/10.1111/i. 1365-2222.1992.tb00158.x
Burra. TA: Moineddin. R: Agha. MM: Glazier. RH. (2009). Social disadvantage, air pollution, and asthma
physician visits in Toronto, Canada. Environ Res 109: 567-574.
http://dx.doi.0rg/10.1016/i.envres.2009.03.004
Butland. BK: Armstrong. B: Atkinson. RW: Wilkinson. P: Heal. MR: Dohertv. RM: Vieno. M. (2013).
Measurement error in time-series analysis: a simulation study comparing modelled and monitored data. BMC
MedRes Methodol 13: 136. http://dx.doi.org/10.1186/1471-2288-13-136
Button. B: Cai LH: Ehre. C: Kesimer. M: Hill DB: Sheehan. JK: Boucher. RC: Rubinstein. M. (2012). A
periciliary brush promotes the lung health by separating the mucus layer from airway epithelia. Science 337:
937-941. http://dx.doi.org/10.1126/science.1223012
Buzica. D: Gerboles. M: Plaisance. H. (2008). The equivalence of diffusive samplers to reference methods for
monitoring O3, benzene and NO2 in ambient air. J Environ Monit 10: 1052-1059.
http://dx.doi.org/10.1039/b802260g
Bylin. G. (1993). Health risk evaluation of nitrogen oxide: Controlled studies on humans. Scand J Work Environ
Health 19: 37-43.
Bylin. G: Hedenstierna. G: Lindvall T: Sundin. B. (1988). Ambient nitrogen dioxide concentrations increase
bronchial responsiveness in subjects with mild asthma. Eur Respir J 1: 606-612.
Bylin. G: Lindvall. T: Rehn. T: Sundin. B. (1985). Effects of short-term exposure to ambient nitrogen dioxide
concentrations on human bronchial reactivity and lung function. Eur Respir J 66: 205-217.
Byun. D: Schere. KL. (2006). Review of the governing equations, computational algorithms, and other
components of the models-3 community multiscale air quality (CMAQ) modeling system [Review]. Appl
MechRev 59: 51-77. http://dx.doi.0rg/10.1115/l.2128636
CAA. Clean Air Act as amended by Pub. L. No. 101-549. section 108: Air quality criteria and control
techniques. 42 USC 7408. 42 USC (1990a). http://www.law.cornell.edu/uscode/text/42/7408
CAA (Clean Air Act). (1990b). Clean Air Act, as amended by Pub. L. No. 101-549, section 109: National
primary and secondary ambient air quality standards, 42 USC 7409.
http://www.epa.gov/air/caa/titlel.htrnMa
CAA. Clean Air Act, section 302: Definitions. 42 USC 7602. 42 USC (2005).
http://www.gpo.gov/fdsvs/pkg/USCODE-2005-title42/pdf/USCODE-2005-title42-chap85-subchapIII-
sec7602.pdf
Cahill. TA: Cahill. TM: Barnes. DE: Spada. NJ: Miller. R. (2011). Inorganic and organic aerosols downwind of
California's Roseville railyard. Aerosol Sci Technol 45: 1049-1059.
http://dx.doi.org/10.1080/02786826.2011.580796
R-14
-------
Cai Y: Schikowski. T: Adam. M: Buschka. A: CarsirL AE: Jacquemin. B: Marcon. A: Sanchez. M: Vierkotter.
A: Al-Kanaanj Z: Beelen. R: Birk. M: Brunekreef B: Cirach. M: Clavel-Chapelon. F: Declercq. C: de
Hoogh. K: de Nazelle. A: Ducret-Stich. RE: Valeria Ferretti V: Forsberg. B: Gerbase. MW: Hardy. R:
Heinrich. J: Hoek. G: Jarvis. D: Keidel D: Kuh. D: Nieuwenhuijsen. MJ: Ragettri. MS: Ranzj A: Rochat T:
Schindler. C: Sugiri. D: Temam. S: Tsai MY: Varraso. R: Kaufimann. F: Kramer. U: Sunyer. J: Kiinzlj N:
Probst-Hensch. N: Hansell AL. (2014). Cross-sectional associations between air pollution and chronic
bronchitis: an ESCAPE meta-analysis across five cohorts [Review]. Thorax 69: 1005-1014.
http://dx.doi.org/10.1136/thoraxinl-2013-204352
Cakmak. S: Dales. R: Leech. J: Liu. L. (201 la). The influence of air pollution on cardiovascular and pulmonary
function and exercise capacity: Canadian Health Measures Survey (CHMS). Environ Res 111: 1309-1312.
http://dx.doi.0rg/10.1016/i.envres.2011.09.016
Cakmak. S: Dales. RE: Judek. S. (2006). Respiratory health effects of air pollution gases: Modification by
education and income. Arch Environ Occup Health 61:5-10. http://dx.doi.Org/10.3200/AEOH.61.l.5-10
Cakmak. S: Dales. RE: Rubio. MA: Vidal CB. (201 Ib). The risk of dying on days of higher air pollution among
the socially disadvantaged elderly. Environ Res 111: 388-393. http://dx.doi.0rg/10.1016/i.envres.2011.01.003
Cakmak. S: Dales. RE: Vida. CB. (2009). Components of paniculate air pollution and mortality in Chile. Int J
Occup Environ Health 15: 152-158. http://dx.doi.org/10.1179/107735209799195844
Campen. MJ: Lund. AK: Doyle-Eisele. ML: Mcdonald. JD: Knuckles. TL: Rohr. AC: Knipping. EM: Mauderly.
JL. (2010). A comparison of vascular effects from complex and individual air pollutants indicates a role for
monoxide gases and volatile hydrocarbons. Environ Health Perspect 118: 921-927.
http://dx.doi.org/10.1289/ehp.0901207
Cano Garcinuno. A: Mora Gandarillas. I. (2013). Early patterns of wheezing in asthmatic and nonasthmatic
children. Eur Respir J 42: 1020-1028. http://dx.doi.org/10.1183/09031936.00148712
Canova. C: Torresan. S: Simonato. L: Scapellato. ML: Tessari. R: Visentin. A: Lotti M: Maestrelli P. (2010).
Carbon monoxide pollution is associated with decreased lung function in asthmatic adults. Eur Respir J 35:
266-272. http://dx.doi.org/10.1183/09031936.00043709
Cao. G: Jang. M. (2008). Secondary organic aerosol formation from toluene photooxidation under various NOx
conditions and particle acidity. Atmos Chem Phys Discuss 8: 14467-14495. http://dx.doi.org/10.5194/acpd-8-
14467-2008
Cao. G: Jang. M. (2010). An SOA model for toluene oxidation in the presence of inorganic aerosols. Environ Sci
Technol 44: 727-733. http://dx.doi.org/10.1021/es901682r
Cao. J: Yang. C: Li. J: Chen. R: Chen. B: Gu. B: Kan. H. (2011). Association between long-term exposure to
outdoor air pollution and mortality in China: A cohort study. J Hazard Mater 186: 1594-1600.
http://dx.doi.0rg/10.1016/i.jhazmat.2010.12.036
Cape. JN: Tang. YS: vanDijk. N: Love. L: Sutton. MA: Palmer. SCF. (2004). Concentrations of ammonia and
nitrogen dioxide at roadside verges, and their contribution to nitrogen deposition. Environ Pollut 132: 469-
478. http://dx.doi.0rg/10.1016/j.envpol.2004.05.009
Cardinale. F: Tesse. R: Fucilli. C: Loffredo. MS: lacoviello. G: Chinellato. I: Armenio. L. (2007). Correlation
between exhaled nitric oxide and dietary consumption of fats and antioxidants in children with asthma
[Letter]. J Allergy Clinlmmunol 119: 1268-1270. http://dx.doi.0rg/10.1016/j.jaci.2007.01.028
Carey. IM: Atkinson. RW: Kent. AJ: van Staa. T: Cook. DG: Anderson. HR. (2013). Mortality associations with
long-term exposure to outdoor air pollution in a national English cohort. Am J Respir Crit Care Med 187:
1226-1233. http://dx.doi.org/10.1164/rccm.201210-1758OC
Carlsen. HK: Forsberg. B: Meister. K: Gislason. T: Oudin. A. (2013). Ozone is associated with cardiopulmonary
and stroke emergency hospital visits in Reykjavik, Iceland 2003-2009. Environ Health 12: 28.
http://dx.doi.org/10.1186/1476-069X-12-28
R-15
-------
Carlsen. HK: Zoega. H: Valdimarsdottir. U: Gislason. T: Hrafnkelsson. B. (2012). Hydrogen sulfide and particle
matter levels associated with increased dispensing of anti-asthma drugs in Iceland's capital. Environ Res 113:
33-39. http://dx.doi.0rg/10.1016/i.envres.2011.10.010
Carlsten. C: Brauer. M: Dimich-Ward. H: Dybuncio. A: Becker. AB: Chan-Yeung. M. (201 la). Combined
exposure to dog and indoor pollution: Incident asthma in a high-risk birth cohort. Eur Respir J 37: 324-330.
http://dx.doi.org/10.1183/09031936.00187609
Carlsten. C: Dybuncio. A: Becker. A: Chan-Yeung. M: Brauer. M. (201 Ib). GSTP1 polymorphism modifies risk
for incident asthma associated with nitrogen dioxide in a high-risk birth cohort [Letter]. Occup Environ Med
In Press: In Press, http://dx.doi.org/10.1136/oem.2010.063560
Carlsten. C: Dybuncio. A: Becker. A: Chan-Yeung. M: Brauer. M. (201 Ic). Traffic-related air pollution and
incident asthma in a high-risk birth cohort. Occup Environ Med 68: 291-295.
http://dx.doi.org/10.1136/oem.2010.055152
Carraro. S: Gottardi. G: Bonetto. G: Baraldi. E. (2007). Exhaled nitric oxide in children with asthma and sinusitis
[Review]. PAI 18: 28-30. http://dx.doi.org/10.1111/i. 1399-3038.2007.00629.X
Carslaw. DC: Beevers. SD: Ropkins. K: Bell MC. (2006). Detecting and quantifying aircraft and other on-
airport contributions to ambient nitrogen oxides in the vicinity of a large international airport. Atmos Environ
40: 5424-5434. http://dx.doi.0rg/10.1016/i.atmosenv.2006.04.062
Carslaw. DC: Williams. ML: Barratt B. en. (2012a). A short-term intervention study - Impact of airport closure
due to the eruption of Eyjafjallajokull on near-field air quality. Atmos Environ 54: 328-336.
http://dx.doi.0rg/10.1016/i.atmosenv.2012.02.020
Carslaw. N: Mota. T: Jenkin. ME: Barley. MH: Mcfiggans. G. (2012b). A significant role for nitrate and
peroxide groups on indoor secondary organic aerosol. Environ Sci Technol 46: 9290-9298.
http://dx.doi.org/10.1021/es301350x
CASAC (Clean Air Scientific Advisory Committee). (2009). Comments and recommendations concerning
EPA's proposed rule for the revision of the National Ambient Air Quality Standards (NAAQS) for nitrogen
dioxide (unsigned) [EPA Report]. (EPA-CASAC-09-014). Washington, D.C.: U.S. Environmental Protection
Agency.
Castro-Giner. F: Kiinzli. N: Jacquemin. B: Forsberg. B: de Cid. R: Sunyer. J: Jarvis. D: Briggs. D: Vienneau. D:
Norback. D: Gonzalez. JR: Guerra. S: Janson. C: Anto. JM: Wist M: Heinrich. J: Estivill X: Kogevinas. M.
(2009). Traffic-related air pollution, oxidative stress genes, and asthma (ECHRS). Environ Health Perspect
117: 1919-1924. http://dx.doi.org/10.1289/ehp.0900589
Castro. HA: Cunha. MF: Mendonca. GA: Junger. WL: Cunha-Cruz. J: Leon. AP. (2009). Effect of air pollution
on lung function in schoolchildren in Rio de Janeiro, Brazil. Rev Saude Publica 43: 26-34.
http://dx.doi.org/10.1590/S0034-89102009000100004
Cattaneo. A: Tecce. N: Derudi M: Gelosa. S: Nano. G: Cavallo. DM. (2014). Assessment of modeled indoor air
concentrations of paniculate matter, gaseous pollutants, and volatile organic compounds emitted from
candles. Hum Ecol Risk Assess 20: 962-979. http://dx.doi.org/10.1080/10807039.2013.821902
Cavanagh. DG: Morris. JB. (1987). Mucus protection and airway peroxidation following nitrogen dioxide
exposure in the rat. J Toxicol Environ Health 22: 313-328. http://dx.doi.org/10.1080/15287398709531074
CDC (Centers for Disease Control and Prevention). (2004). The health consequences of smoking: A report of the
Surgeon General. Washington, DC: U.S. Department of Health and Human Services.
http://www.cdc.gov/tobacco/data statistics/sgr/2004/index. htm
CDC (Centers for Disease Control and Prevention). (2013). State indicator report on fruits and vegetables.
Atlanta, GA. http://www.cdc.gov/nutrition/downloads/State-Indicator-Report-Fruits-Vegetables-2013.pdf
Centro di Ricerche Ambientali (Centro di Ricerche Ambientali di Padova). (2006). Radiello(R) manual. Padova,
Italy: Fondazione Salvatore Maugeri Clinica del Lavoro e Delia Riabilitazione I.R.C.C.S., Centro di Ricerche
Ambientali. http://www.radiello.com/english/Radiello%27s%20manual%2001-06.pdf
R-16
-------
Cesaroni. G: Badaloni C: Gariazzo. C: Stafoggia. M: Sozzi. R: Davoli. M: Forastiere. F. (2013). Long-term
exposure to urban air pollution and mortality in a cohort of more than a million adults in rome. Environ
Health Perspect 121: 324-331. http://dx.doi.org/10.1289/ehp. 1205862
Cesaroni. G: Badaloni. C: Porta. D: Forastiere. F: Perucci. CA. (2008). Comparison between various indices of
exposure to traffic-related air pollution and their impact on respiratory health in adults. Occup Environ Med
65: 683-690. http://dx.doi.org/10.1136/oem.2007.037846
Cesaroni. G: Forastiere. F: Stafoggia. M: Andersen. ZJ: Badaloni. C: Beelen. R: Caracciolo. B: de Faire. U:
Erbel R: Eriksen. KT: Fratiglioni. L: Galassi C: Hampel R: Heier. M: Hennig. F: Hilding. A: Hoffmann. B:
Houthuijs. D: Jockel. KH: Korek. M: Lanki. T: Leander. K: Magnusson. PK: Migliore. E: Ostenson. CG:
Overvad. K: Pedersen. NL: J. JP: Penell. J: Pershagen. G: Pyko. A: Raaschou-Nielsen. O: Ranzi. A: Ricceri.
F: Sacerdote. C: Salomaa. V: Swart. W: Turunen. AW: Vineis. P: Weinmayr. G: Wolf. K: de Hoogh. K:
Hoek. G: Brunekreef. B: Peters. A. (2014). Long term exposure to ambient air pollution and incidence of
acute coronary events: prospective cohort study and meta-analysis in 11 European cohorts from the ESCAPE
Project. B M J (Online) 348:17412.
Cesaroni. G: Porta. D: Badaloni. C: Stafoggia. M: Eeftens. M: Meliefste. K: Forastiere. F. (2012). Nitrogen
dioxide levels estimated from land use regression models several years apart and association with mortality
in a large cohort study. Environ Health 11:48. http://dx.doi.org/10.1186/1476-069X-11-48
Chadha. TS: Birch. S: Sackner. MA. (1987). Oronasal distribution of ventilation during exercise in normal
subjects and patients with asthma and rhinitis. Chest 92: 1037-1041.
http://dx.doi.0rg/10.1378/chest.92.6.1037
Chaix. B: Gustafsson. S: Jerrett M: Kristersson. H: Lithman. T: Boalt A: Merlo. J. (2006). Children's exposure
to nitrogen dioxide in Sweden: Investigating environmental injustice in an egalitarian country. J Epidemiol
Community Health 60: 234-241. http://dx.doi.org/10.1136/iech.2005.038190
Chakrabortv. S: Balakotaiah. V: Bidani. A. (2004). Diffusing capacity reexamined: Relative roles of diffusion
and chemical reaction in red cell uptake of O2, CO, CO2, and NO. J Appl Physiol 97: 2284-2302.
http://dx.doi.org/10.1152/iapplphysiol.00469.2004
Chambers. DC: Ayres. JG. (2001). Effects of nitrogen dioxide exposure and ascorbic acid supplementation on
exhaled nitric oxide in healthy human subjects. Thorax 56: 774-778.
http://dx.doi.org/10.1136/thorax.56.10.774
Chan. CC: Chuang. KJ: Chen. WJ: Chang. WT: Lee. CT: Peng. CM. (2008). Increasing cardiopulmonary
emergency visits by long-range transported Asian dust storms in Taiwan. Environ Res 106: 393-400.
http://dx.doi.0rg/10.1016/i.envres.2007.09.006
Chanev. AM: Cryer. DJ: Nicholl EJ: Seakins. PW. (2011). NO and NO(2) interconversion downwind of two
different line sources in suburban environments. Atmos Environ 45: 5863-5871.
http://dx.doi.0rg/10.1016/i.atmosenv.2011.06.070
Chang. CC: Tsai SS: Ho. SC: Yang. CY. (2005). Air pollution and hospital admissions for cardiovascular
disease in Taipei, Taiwan. Environ Res 98: 114-119. http://dx.doi.0rg/10.1016/i.envres.2004.07.005
Chang. JC: Franzese. P: Chavantrakom K: Hanna. SR. (2003). Evaluations of CALPUFF, HP AC, and
VLSTRACK with two mesoscale field datasets. J Appl Meteorol 42: 453-466.
http://dx.doi.org/10.1175/1520-0450(2003)0420453 :EOCHAV>2.0.CO:2
Chang. LY: Graham. JA: Miller. FJ: Ospital JJ: Crapo. JD. (1986). Effects of subchronic inhalation of low
concentrations of nitrogen dioxide: I. The proximal alveolar region of juvenile and adult rats. Toxicol Appl
Pharmacol 83: 46-61. http://dx.doi.org/10.1016/0041-008X(86)90321-2
Chang. YK: Wu. CC: Lee. LT: Lin. RS: Yu. YH: Chen. YC. (2012). The short-term effects of air pollution on
adolescent lung function in Taiwan. Chemosphere 87: 26-30.
http://dx.doi.0rg/10.1016/i.chemosphere.2011.11.048
R-17
-------
Channel!. MM: Paffett ML: Devlin. RB: Madden. MC: Campen. MJ. (2012). Circulating factors induce
coronary endothelial cell activation following exposure to inhaled diesel exhaust and nitrogen dioxide in
humans: Evidence from a novel translational in vitro model. Toxicol Sci 127: 179-186.
http://dx.doi.org/10.1093/toxsci/kfs084
Chase. MW. Jr. (1998). NIST-JANAF thermochemical tables, monograph 9 (Fourth Edition ed.). Washington
DC: American Chemical Society, http://www.nist.gov/srd/monogr.cfm
Chatzidiakou. L: Mumovic. D: Summerfield. AJ: Hong. SM: Altamirano-Medina. H. (2014). A Victorian school
and a low carbon designed school: Comparison of indoor air quality, energy performance, and student health.
Indoor Built Environ 23: 417-432. http://dx.doi.org/10.1177/1420326X14532388
Chea BY: Chaa CC: Lee. CT: Cheng. TJ: Huang. WC: Jhou. JC: Haa YY: Chea CC: Quo. YL. (2012a). The
association of ambient air pollution with airway inflammation in schoolchildren. Am J Epidemiol 175: 764-
774. http://dx.doi.org/10.1093/aie/kwr380
Chea EK: Zmirou-Navier. D: Padilla. C: Deguen. S. (2014a). Effects of air pollution on the risk of congenital
anomalies: a systematic review and meta-analysis [Review]. Int J Environ Res Public Health 11: 7642-7668.
http://dx.doi.org/10.3390/ijerphll0807642
Chen. G: Li. J: Ying. Q: Sherman. S: Perkins. N: Rajeshwarj S: Mendola. P. (2014b). Evaluation of observation-
fused regional air quality model results for population air pollution exposure estimation. Sci Total Environ
485-486: 563-574. http://dx.doi.0rg/10.1016/i.scitotenv.2014.03.107
Chea G: Song. G: Jiang. L: Zhang. Y: Zhao. N: Chea B: Kan. H. (2008). Short-term effects of ambient gaseous
pollutants and paniculate matter on daily mortality in Shanghai, China. J Occup Health 50: 41-47.
http://dx.doi.org/10.1539/ioh.50.41
Chen. H: Goldberg. MS: Burnett. RT: Jerrett M: Wheeler. AJ: Villeneuve. PJ. (2013a). Long-term exposure to
traffic-related air pollution and cardiovascular mortality. Epidemiology 24: 35-43.
http://dx.doi.org/10.1097/EDE.Ob013e318276c005
Chea L: Bell EM: Caton. AR: Druschel CM: Lin. S. (2010a). Residential mobility during pregnancy and the
potential for ambient air pollution exposure misclassification. Environ Res 110: 162-168.
http://dx.doi.0rg/10.1016/i.envres.2009.ll.001
Chen. L: Villeneuve. PJ: Rowe. BH: Liu. L: Stieb. DM. (2014c). The Air Quality Health Index as a predictor of
emergency department visits for ischemic stroke in Edmonton, Canada. J Expo Sci Environ Epidemiol 24:
358-364. http://dx.doi.org/10.1038/ies.2013.82
Chea R: Chu. C: Tan. J: Cao. J: Song. W: Xu. X: Jiang. C: Ma. W: Yang. C: Chen. B: Gui. Y: Kaa H. (2010b).
Ambient air pollution and hospital admission in Shanghai, China. J Hazard Mater 181: 234-240.
http://dx.doi.0rg/10.1016/i.jhazmat.2010.05.002
Chea R: Samoli. E: Wong. CM: Huang. W: Wang. Z: Chen. B: Kaa H. (2012b). Associations between short-
term exposure to nitrogen dioxide and mortality in 17 Chinese cities: The China Air Pollution and Health
Effects Study (CAPES). Environ Int 45: 32-38. http://dx.doi.0rg/10.1016/i.envint.2012.04.008
Chea R: Zhang. Y: Yang. C: Zhao. Z: Xu. X: Kan. H. (2013b). Acute effect of ambient air pollution on stroke
mortality in the china air pollution and health effects study. Stroke 44: 954-960.
http://dx.doi.0rg/10.l 161/STROKEAHA. 111.673442
Chea SY: Su. TC: Lin. YL: Chan. CC. (2012c). Short-term effects of air pollution on pulse pressure among
nonsmoking adults. Epidemiology 23: 341-348. http://dx.doi.org/10.1097/EDE.Ob013e3182452fld
Cheng. MF: Tsai SS: Yang. CY. (2009a). Air pollution and hospital admissions for myocardial infarction in a
tropical city: Kaohsiung, Taiwan. J Toxicol Environ Health A 72: 1135-1140.
http://dx.doi.org/10.1080/15287390903091756
Cheng. WC: Liu. CH: Leung. DYC. (2008). Computational formulation for the evaluation of street canyon
ventilation and pollutant removal performance. Atmos Environ 42: 9041-9051.
http://dx.doi.0rg/10.1016/i.atmosenv.2008.09.045
R-18
-------
Cheng. WC: Liu. CH: Leung. DYC. (2009b). On the correlation of air and pollutant exchange for street canyons
in combined wind-buoyancy-driven flow. Atmos Environ 43: 3682-3690.
http://dx.doi.0rg/10.1016/i.atmosenv.2009.04.054
Chetta. A: Foresi A: Del Donno. M: Consigli GF: Bertorelli. G: Pesci A: Barbee. RA: Olivieri. D. (1996).
Bronchial responsiveness to distilled water and methacholine and its relationship to inflammation and
remodeling of the airways in asthma. Am J Respir Crit Care Med 153: 910-917.
http://dx.doi.0rg/10.1164/airccm.153.3.8630572
Cheung. D: Dick. EC: Timmers. MC: de Klerk. EP: Spaan. WJ: Sterk. PJ. (1995). Rhinovirus inhalation causes
long-lasting excessive airway narrowing in response to methacholine in asthmatic subjects in vivo. Am J
Respir Crit Care Med 152: 1490-1496. http://dx.doi.0rg/10.1164/airccm.152.5.7582282
Chimenti. L: Morici G: Paterno. A: Bonanno. A: Vultaggio. M: Bellia. V: Bonsignore. MR. (2009).
Environmental conditions, air pollutants, and airway cells in runners: A longitudinal field study. J Sports Sci
27: 925-935. http://dx.doi.org/10.1080/02640410902946493
Chiusolo. M: Cadum. E: Stafoggia. M: Galassi C: Berti G: Faustini. A: Bisanti L: Vigotti. MA: Dessl MP:
Cernigliaro. A: Mallone. S: Pacelli. B: Minerba. S: Simonato. L: Forastiere. F. (2011). Short term effects of
nitrogen dioxide on mortality and susceptibility factors in 10 Italian cities: The EpiAir Study. Environ Health
Perspect 119: 1233-1238. http://dx.doi.org/10.1289/ehp. 1002904
Choi. JH: Xu. OS: Park. SY: Kim JH: Hwang. SS: Lee. KH: Lee. HJ: Hong. YC. (2007). Seasonal variation of
effect of air pollution on blood pressure. JEpidemiol Community Health 61: 314-318.
http://dx.doi.org/10.1136/iech.2006.049205
Choi Y: Kim. H: Tong. D: Lee. P. (2012). Summertime weekly cycles of observed and modeled NOx and O3
concentrations as a function of satellite-derived ozone production sensitivity and land use types over the
Continental United States. Atmos ChemPhys 12: 6291-6307. http://dx.doi.org/10.5194/acp-12-6291-2012
Chuang. KJ: Chan. CC: Su. TC: Lee. CT: Tang. CS. (2007a). The effect of urban air pollution on inflammation,
oxidative stress, coagulation, and autonomic dysfunction in young adults. Am J Respir Crit Care Med 176:
370-376. http://dx.doi.org/10.1164/rccm.200611-1627OC
Chuang. KJ: Chan. CC: Su. TC: Lin. LY: Lee. CT. (2007b). Associations between paniculate sulfate and organic
carbon exposures and heart rate variability in patients with or at risk for cardiovascular diseases. J Occup
Environ Med 49: 610-617. http://dx.doi.org/10.1097/JOM.Ob013e318058205b
Chuang. KJ: Coull BA: Zanobetti A: Suh. H: Schwartz. J: Stone. PH: Litonjua. A: Speizer. FE: Gold. DR.
(2008). Paniculate air pollution as a risk factor for ST-segment depression in patients with coronary artery
disease. Circulation 118: 1314-1320. http://dx.doi.org/10.1161/CIRCULATIONAHA.108.765669
Chuang. KJ: Yan. YH: Cheng. TJ. (2010). Effect of air pollution on blood pressure, blood lipids, and blood
sugar: A population-based approach. J Occup Environ Med 52: 258-262.
http://dx.doi.org/10.1097/JOM.ObQ13e3181ceff7a
Chung. MK: Lao. TT: Ting. YH: Wong. TW: Leung. TY. (2014). Seasonality of fetal trisomy 21 - Have ambient
air pollutants played a role? J Matern Fetal Neonatal Med 28: 552-557.
http://dx.doi.org/10.3109/14767058.2014.924104
Cimorelli. AJ: Perry. SG: Venkatram A: Weil JC: Paine. R: Wilson. RB: Lee. RF: Peters. WD: Erode. RW.
(2005). AERMOD: A dispersion model for industrial source applications. Part I: General model formulation
and boundary layer characterization. J Appl Meteorol 44: 682-693. http://dx.doi.Org/10.1175/JAM2227.l
Ckless. K: Hodgkins. SR: Ather. JL: Martin. R: Poynter. ME. (2011). Epithelial, dendritic, and CD4(+) T cell
regulation of and by reactive oxygen and nitrogen species in allergic sensitization [Review]. Biochim
Biophys Acta 1810: 10251034. http://dx.doi.0rg/10.1016/i.bbagen.2011.03.005
Ckless. K: Lampert. A: Reiss. J: Kasahara. D: Poynter. ME: Irvin. CG: Lundblad. LKA: Norton. R: van der
Vliet A: Janssen-Heininger. YMW. (2008). Inhibition of arginase activity enhances inflammation in mice
with allergic airway disease, in association with increases in protein S-nitrosylation and tyrosine nitration. J
Immunol 181: 4255-4264.
R-19
-------
Clapp. L: Jenkin. M. (2001). Analysis of the relationship between ambient levels of O3, NO2 and NO as a
function of NOx in the UK. Atmos Environ 35: 6391-6405. http://dx.doi.org/10.1016/81352-2310(01100378-
8
Clark. C: Crombie. R: Head. J: van Kamp. I: van Kempen. E: Stansfeld. SA. (2012). Does traffic-related air
pollution explain associations of aircraft and road traffic noise exposure on children's health and cognition? A
secondary analysis of the United Kingdom sample from the RANCH project. Am J Epidemiol 176: 327-337.
http://dx.doi.org/10.1093/aie/kws012
Clark. LP: Millet DB: Marshall JD. (2014). National patterns in environmental injustice and inequality:
Outdoor NO2 air pollution in the United States. PLoS ONE 9: e94431.
http://dx.doi.org/10.1371/iournal.pone.0094431
Clark. NA: Demers. PA: Karr. CJ: Koehoorn. M: Lencar. C: Tamburic. L: Brauer. M. (2010). Effect of early life
exposure to air pollution on development of childhood asthma. Environ Health Perspect 118: 284-290.
http://dx.doi.org/10.1289/ehp.0900916
Clary-Meinesz. C: Mouroux. J: Huitorel P: Cosson. J: Schoevaert. D: Blaive. B. (1997). Ciliary beat frequency
in human bronchi and bronchioles. Chest 111: 692-697. http://dx.doi.org/10.1378/chest. 111.3.692
Clements. AL: Jia. Y: Denbleyker. A: McDonald-Buller. E: Fraser. MP: Allen. DT: Collins. PR: Michel E:
Pudota. J: Sullivan. D: Zhu. Y. (2009). Air pollutant concentrations near three Texas roadways, part II:
Chemical characterization and transformation of pollutants. Atmos Environ 43: 4523-4534.
http://dx.doi.0rg/10.1016/i.atmosenv.2009.06.044
Cleveland. WS: Graedel. TE: Kleiner. B: Warner. JL. (1974). Sunday and workday variations in photochemical
air pollutants in New Jersey and New York. Science 186: 1037-1038.
http://dx.doi.org/10.1126/science. 186.4168.1037
Cloughertv. JE: Kheirbek. I: Eisl. HM: Ross. Z: Pezeshkj G: Gorczynski. JE: Johnson. S: Markowitz. S: Kass.
D: Matte. T. (2013). Intra-urban spatial variability in wintertime street-level concentrations of multiple
combustion-related air pollutants: The New York City Community Air Survey (NYCCAS). J Expo Sci
Environ Epidemiol 23: 232-240. http://dx.doi.org/10.1038/ies.2012.125
Cloughertv. JE: Lew. JI: Kubzansky. LD: Ryan. PB: Suglia. SF: Canner. MJ: Wright PJ. (2007). Synergistic
effects of traffic-related air pollution and exposure to violence on urban asthma etiology. Environ Health
Perspect 115: 1140-1146. http://dx.doi.org/10.1289/ehp.9863
Cockcroft DW. (2010). Direct challenge tests. Airway hyperresponsiveness in asthma: Its measurement and
clinical significance [Review]. Chest 138: 18S-24S. http://dx.doi.org/10.1378/chest.10-0088
Cockcroft DW: Berscheid. BA: Murdock. KY. (1983). Unimodal distribution of bronchial responsiveness to
inhaled histamine in a random human population. Chest 83: 751-754.
http://dx.doi.0rg/10.1378/chest.83.5.751
Cockcroft DW: Davis. BE. (2006a). Airway hyperresponsiveness as a determinant of the early asthmatic
response to inhaled allergen [Review]. J Asthma 43: 175-178. http://dx.doi.org/10.1080/02770900600566348
Cockcroft DW: Davis. BE. (2006b). The bronchoprotective effect of inhaling methacholine by using total lung
capacity inspirations has a marked influence on the interpretation of the test result [Review]. J Allergy Clin
Immunol 117: 1244-1248. http://dx.doi.0rg/10.1016/i.jaci.2006.02.038
Cockcroft DW: Davis. BE. (2006c). Mechanisms of airway hyperresponsiveness [Review]. J Allergy Clin
Immunol 118: 551-559. http://dx.doi.0rg/10.1016/i.iaci.2006.07.012
Cockcroft DW: Davis. BE: Todd. DC: Smycniuk. AJ. (2005). Methacholine challenge: Comparison of two
methods. Chest 127: 839-844. http://dx.doi.0rg/10.1378/chest.127.3.839
Cohan. A: Wu. J: Dabdub. D. (2011). High-resolution pollutant transport in the San Pedro Bay of California.
Atmos Pollut Res 2: 237-246. http://dx.doi.org/10.5094/APR.2011.030
R-20
-------
Cohen. MA: Adar. SD: Allen. RW: Avol. E: Curl. CL: Gould. T: Hardie. D: Ho. A: Kinnev. P: Larsoa TV:
Sampson. P: Sheppard. L: Stukovsky. KD: Swan. SS: Liu. LJ: Kaufman. JD. (2009). Approach to estimating
participant pollutant exposures in the multi-ethnic study of atherosclerosis and air pollution (MESA Air).
Environ Sci Technol 43: 4687-4693. http://dx.doi.org/10.1021/es8030837
Colais. P: Faustini. A: Stafoggia. M: Berti. G: Bisanti L: Cadum. E: Cernigriaro. A: Mallone. S: Pacelli B:
Serinelli M: Simonato. L: Vigotti. MA: Forastiere. F. (2012). Paniculate air pollution and hospital
admissions for cardiac diseases in potentially sensitive subgroups. Epidemiology 23: 473-481.
http://dx.doi.org/10.1097/EDE.Ob013e31824d5a85
Comhair. SAA: Erzurum. SC. (2010). Redox control of asthma: Molecular mechanisms and therapeutic
opportunities [Review]. Antioxid Redox Signal 12: 93-124. http://dx.doi.org/10.1089/ARS.2008.2425
Connell DP: Withum. JA: Winter. SE: Statnick. RM: Bilonick. RA. (2005). The Steubenville Comprehensive
Air Monitoring Program (SCAMP): Associations among fine paniculate matter, co-pollutants, and
meteorological conditions. J Air Waste Manag Assoc 55: 481-496.
http://dx.doi.org/10.1080/10473289.2005.10464631
Connor. LM: Ballinger. CA: Albrecht TB: Postlethwait EM. (2004). Interfacial phospholipids inhibit ozone
reactive absorption-mediated cytotoxicity in vitro. Am J Physiol 286: L1169-L1178.
http://dx.doi.org/10.1152/aiplung.00397.2003
Connor. LM: Bidani. A: Goerke. J: Clements. JA: Postlethwait EM. (2001). NO2 interfacial transfer is reduced
by phospholipid monolayers. J Appl Physiol 91: 2024-2034.
Coogan. PF: White. LF: Jerrett M: Brook. RD: Su. JG: Seto. E: Burnett. R: Palmer. JR: Rosenberg. L. (2012).
Air pollution and incidence of hypertension and diabetes mellitus in black women living in Los Angeles.
Circulation 125: 767-772. http://dx.doi.org/10.1161/CIRCULATIONAHA. 111.052753
Corbett JJ: Fischbeck. PS: Pandis. SN. (1999). Global nitrogen and sulfur inventories for oceangoing ships. J
Geophys Res Atmos 104: 3457-3470. http://dx.doi.org/10.1029/1998JD100040
Correia-Deur. J: Claudio. L: Imazawa. AT: Eluf-Neto. J. (2012). Variations in peak expiratory flow
measurements associated to air pollution and allergic sensitization in children in Sao Paulo, Brazil. Am J Ind
Med55: 1087-1098. http://dx.doi.org/10.1002/aiim.2206Q
Cosby. K: Partovi KS: Crawford. JH: Patel. RP: Reiter. CD: Martyr. S: Yang. BK: Waclawiw. MA: Zalos. G:
Xu. X: Huang. KT: Shields. H: Kim-Shapiro. DB: Schechter. AN: Cannon. RO: Gladwin. MT. (2003).
Nitrite reduction to nitric oxide by deoxyhemoglobin vasodilates the human circulation. Nat Med 9: 1498-
1505. http://dx.doi.org/10.1038/nm954
Cowie. CT: Rose. N: Gillett. R: Walter. S: Marks. GB. (2012). Redistribution of traffic related air pollution
associated with a new road tunnel. Environ Sci Technol 46: 2918-2927. http://dx.doi.org/10.1021/es202686r
Crapo. JD: Barry. BE: Chang. LY: Mercer. RR. (1984). Alterations in lung structure caused by inhalation of
oxidants. J Toxicol Environ Health 13: 301-321. http://dx.doi.org/10.1080/15287398409530500
Crooks. J: Isakov. V. (2013). A wavelet-based approach to blending observations with deterministic computer
models to resolve the intraurban air pollution field. J Air Waste Manag Assoc 63: 1369-1385.
http://dx.doi.org/10.1080/10962247.2012.758061
Croslev. DR. (1996). NOy blue ribbon panel. J Geophys Res 101: 2049-2052.
http://dx.doi.org/10.1029/95JD02276
Grouse. PL: Goldberg. MS: Ross. NA. (2009a). A prediction-based approach to modelling temporal and spatial
variability of traffic-related air pollution in Montreal, Canada. Atmos Environ 43: 5075-5084.
http://dx.doi.0rg/10.1016/i.atmosenv.2009.06.040
Grouse. PL: Goldberg. MS: Ross. NA: Chen. H: Labreche. F. (2010). Postmenopausal breast cancer is
associated with exposure to traffic-related air pollution in Montreal, Canada: A casecontrol study. Environ
HealthPerspect 118: 1578-1583. http://dx.doi.org/10.1289/ehp.1002221
R-21
-------
Grouse. PL: Ross. NA: Goldberg. MS. (2009b). Double burden of deprivation and high concentrations of
ambient air pollution at the neighbourhood scale in Montreal, Canada. Soc Sci Med 69: 971-981.
http://dx.doi.0rg/10.1016/i.socscimed.2009.07.010
Cui H: Yao. R: Xu. X: Xin. C: Yang. J. (2011). A tracer experiment study to evaluate the CALPUFF real time
application in a near-field complex terrain setting. Atmos Environ 45: 7525-7532.
http://dx.doi.0rg/10.1016/i.atmosenv.2011.08.041
Cuinica. LG: Abreu. I: Esteves da Silva. J. (2014). Effect of air pollutant NO2 on Betula pendula, Ostrya
carpinifolia and Carpinus betulus pollen fertility and human allergenicity. Environ Pollut 186: 50-55.
http://dx.doi.0rg/10.1016/i.envpol.2013.12.001
D'Ippoliti. D: Forastiere. F: Ancona. C: Agabiti N: Fusco. D: Michelozzi. P: Perucci. CA. (2003). Air pollution
and myocardial infarction in Rome: a case-crossover analysis. Epidemiology 14: 528-535.
http://dx.doi.org/10.1097/01.ede.0000082046.22919.72
Dadvand. P: Basagana. X: Figueras. F: Martinez. D: Beelen. R: Cirach. M: de Nazelle. A: Hoek. G: Ostro. B:
Nieuwenhuijsen. MJ. (2014a). Air pollution and preterm premature rupture of membranes: a spatiotemporal
analysis. Am J Epidemiol 179: 200-207. http://dx.doi.org/10.1093/aie/kwt240
Dadvand. P: Figueras. F: Basagana. X: Beelen. R: Martinez. D: Cirach. M: Schembari. A: Hoek. G: Brunekreef.
B: Nieuwenhuijsen. MJ. (2013). Ambient air pollution and preeclampsia: A spatiotemporal analysis. Environ
Health Perspect 121: 1365-1371. http://dx.doi.org/10.1289/ehp. 1206430
Dadvand. P: Nieuwenhuijsen. MJ: Agusti. A: de Batlle. J: Benet M: Beelen. R: Cirach. M: Martinez. D: Hoek.
G: Basagana. X: Ferrer. A: Ferrer. J: Rodriguez-Roisin. R: Sauleda. J: Guerra. S: Anto. JM: Garcia-
Aymerich. J. (2014b). Air pollution and biomarkers of systemic inflammation and tissue repair in COPD
patients. Eur Respir J 44: 603-613. http://dx.doi.org/10.1183/09031936.00168813
Dadvand. P: Ostro. B: Figueras. F: Foraster. M: Basagana. X: Valentin. A: Martinez. D: Beelen. R: Cirach. M:
Hoek. G: Jerrett M: Brunekreef. B: Nieuwenhuijsen. MJ. (2014c). Residential proximity to major roads and
term low birth weight: The roles of air pollution, heat, noise, and road-adjacent trees. Epidemiology 25: 518-
525. http://dx.doi.org/10.1097/EDE.0000000000000107
Daiber. A: Muenzel T. (2012). Increased circulating levels of 3-nitrotyrosine autoantibodies marker for or
maker of cardiovascular disease? [Editorial]. Circulation 126: 2371-2373.
http://dx.doi.org/10.1161/CIRCULATIONAHA.112.143214
Dales. R: Burnett. RT: Smith-Doiron. M: Stieb. DM: Brook. JR. (2004). Air pollution and sudden infant death
syndrome. Pediatrics 113: e628-e631.
Dales. R: Chen. L: Frescura. AM: Liu. L: Villeneuve. PJ. (2009a). Acute effects of outdoor air pollution on
forced expiratory volume in 1 s: A panel study of schoolchildren with asthma. Eur Respir J 34: 316-323.
http://dx.doi.org/10.1183/09031936.00138908
Dales. R: Kauri. LM: Cakmak. S: Mahmud. M: Weichenthal SA: VanRyswyk. K: Kumarathasan. P: Thomson.
E: Vincent. R: Broad. G: Liu. L. (2013). Acute changes in lung function associated with proximity to a steel
plant: A randomized study. Environ Int 55: 15-19. http://dx.doi.0rg/10.1016/i.envint.2013.01.014
Dales. R: Wheeler. A: Mahmud. M: Frescura. AM: Smith-Doiron. M: Nethery. E: Liu. L. (2008). The influence
of living near roadways on spirometry and exhaled nitric oxide in elementary schoolchildren. Environ Health
Perspect 116: 1423-1427. http://dx.doi.org/10.1289/ehp. 10943
Dales. RE: Cakmak. S: Doiron. MS. (2006). Gaseous air pollutants and hospitalization for respiratory disease in
the neonatal period. Environ Health Perspect 114: 1751-1754.
Dales. RE: Cakmak. S: Vidal. CB. (2009b). Air pollution and hospitalization for headache in Chile. Am J
Epidemiol 170: 1057-1066. http://dx.doi.org/10.1093/aie/kwp217
Dales. RE: Cakmak. S: Vidal. CB. (2010). Air pollution and hospitalization for venous thromboembolic disease
in Chile. J Thromb Haemost 8: 669-674. http://dx.doi.0rg/10.llll/i.1538-7836.2010.03760.x
R-22
-------
Dall'Osto. M: Thorpe. A: Beddows. DCS: Harrison. RM: Barlow. JF: Dunbar. T: Williams. PI: Coe. H. (2011).
Remarkable dynamics of nanoparticles in the urban atmosphere. Atmos ChemPhys 11: 6623-6637.
http://dx.doi.org/10.5194/acp-ll-6623-2011
Dallmann. TR: Harlev. RA. (2010). Evaluation of mobile source emission trends in the United States. J Geophys
Res 115: D14305. http://dx.doi.org/10.1029/2010JD013862
Darrow. LA: Klein. M: Flanders. WD: Waller. LA: Correa. A: Marcus. M: Mulholland. JA: Russell AG:
Tolbert PE. (2009). Ambient air pollution and preterm birth: A time-series analysis. Epidemiology 20: 689-
698. http://dx.doi.org/10.1097/EDE.Ob013e3181a7128f
Darrow. LA: Klein. M: Sarnat. JA: Mulholland. JA: Strickland. MJ: Sarnat SE: Russell AG: Tolbert. PE.
(201 la). The use of alternative pollutant metrics in time-series studies of ambient air pollution and
respiratory emergency department visits. J Expo Sci Environ Epidemiol 21: 10-19.
http://dx.doi.org/10.1038/ies.2009.49
Darrow. LA: Klein. M: Strickland. MJ: Mulholland. JA: Tolbert. PE. (201 Ib). Ambient air pollution and birth
weight in full-term infants in Atlanta, 1994-2004. Environ Health Perspect 119: 731-737.
http://dx.doi.org/10.1289/ehp. 1002785
Davies. HW: Vlaanderen. JJ: Henderson. SB: Brauer. M. (2009). Correlation between co-exposures to noise and
air pollution from traffic sources. Occup Environ Med 66: 347-350.
http://dx.doi.org/10.1136/oem.2008.041764
Davies. PJ: Rusznak. C: Calderon. MA: Wang. JH: Abdelaziz. MM: Devalia. JL. (1997). Allergen-irritant
interaction and the role of corticosteroids [Review]. Allergy 52: 59-65. http://dx.doi.org/10.Illl/i.1398-
9995.1997.tb04873.x
Davis. JK: Davidson. M: Schoeb. TR. (1991). Murine respiratory mycoplasmosis: A model to study effects of
oxidants (pp. 1-29; discussion 31-43). Health Effects Institute. http://pubs.healtheffects.org/view.php?id=65
Davis. JK: Davidson. MK: Schoeb. TR: Lindsev. JR. (1992). Decreased intrapulmonary killing of Mycoplasma
pulmonis after short-term exposure to NO2 is associated with damaged alveolar macrophages. Am Rev
RespirDis 145: 406-411. http://dx.doi.0rg/10.1164/airccrn/145.2 Pt 1.406
Dawson. SV: Schenker. MB. (1979). Health effects of inhalation of ambient concentrations of nitrogen dioxide
[Editorial]. Am Rev Respir Dis 120: 281-292.
de Burbure. CY: Heilier. JF: Neve. J: Becker. A: Albrecht C: Borm. PJA: Gromadzinska. J: Wasowicz. W:
Rydzynski. K: Bernard. AM. (2007). Lung permeability, antioxidant status, and NO2 inhalation: A selenium
supplementation study in rats. J Toxicol Environ Health A 70: 284-294.
http://dx.doi.org/10.1080/15287390600884875
de Gouw. JA: Warneke. C: Parrish. DP: Hollowav. JS: Trainer. M: Fehsenfield. FC. (2003). Emission sources
and ocean uptake of acetonitrile (CH 3 CN) in the atmosphere. J Geophys Res 108: 4329.
http://dx.doi.org/10.1029/2002JD002897
de Kluizenaar. Y: vanLenthe. FJ: Visschedijk. AJ: Zandveld. PY: Miedema. HM: Mackenbach. JP. (2013).
Road traffic noise, air pollution components and cardiovascular events. Noise Health 15: 388-397.
http://dx.doi.org/10.4103/1463-1741.121230
De Santis. F: Allegrini. I: Fazio. MC: Pasella. D. (1996). Characterization of indoor air quality in the Church of
San luigi Dei Francesi, Rome, Italy. Int J Environ Anal Chem 64: 71-81.
http://dx.doi.org/10.1080/03067319608028336
de Winter-Sorkina. R: Cassee. FR. (2002). From concentration to dose: Factors influencing airborne paniculate
matter deposition in humans and rats (pp. 1-36). (RIVM Report 650010031). Bilthoven, Netherlands:
National Institute for Public Health and the Environment.
http://www.rivm.nl/en/Documents andjublications/Scientific/Reports/2003/februari/From concentration t
o dose factors influencing airborne_particulate matter deposition in humans and rats?sp=cml2bXE9Zm
Fsc2U7c2VhcmNoYmFzZT01MDI5MDtvaXZtcTlmYWxzZTs=&pagenr=5030
R-23
-------
Debry. E: Mallet V. (2014). Ensemble forecasting with machine learning algorithms for ozone, nitrogen dioxide
and PM10 on the Prev'Air platform. Atmos Environ 91: 71-84.
http://dx.doi.0rg/10.1016/i.atmosenv.2014.03.049
DeCastro. BR: Sax. AN: Chillrud. SN: Kinnev. PL: Spengler. JD. (2007). Modeling time-location patterns of
inner-city high school students in New York and Los Angeles using a longitudinal approach with generalized
estimating equations. J Expo Sci Environ Epidemiol 17: 233-247. http://dx.doi.org/10.1038/sj.jes.7500504
Deguen. S: Zmirou-Navier. D. (2010). Social inequalities resulting from health risks related to ambient air
quality—A European review [Review]. Eur J Public Health 20: 27-35.
http://dx.doi.org/10.1093/eurpub/ckp220
Delamater. PL: Finlev. AO: Banerjee. S. (2012). An analysis of asthma hospitalizations, air pollution, and
weather conditions in Los Angeles County, California. Sci Total Environ 425: 110-118.
http://dx.doi.0rg/10.1016/i.scitotenv.2012.02.015
Delfino. PJ. (2006). Think globally, breathe locally [Editorial]. Thorax 61: 184-185.
http://dx.doi.org/10.1136/thx.2005.047894
Delfino. RJ: Gillen. PL: Tjoa. T: Staimer. N: Polidori. A: Arhami. M: Sioutas. C: Longhurst J. (2011).
Electrocardiographic ST-segment depression and exposure to trafficrelated aerosols in elderly subjects with
coronary artery disease. Environ Health Perspect 119: 196-202. http://dx.doi.org/10.1289/ehp. 1002372
Delfino. RJ: Gone. H: Linn. WS: Pellizzari. ED: Hu. Y. (2003). Asthma symptoms in Hispanic children and
daily ambient exposures to toxic and criteria air pollutants. Environ Health Perspect 111: 647-656.
Delfino. RJ: Staimer. N: Gillen. D: Tjoa. T: Sioutas. C: Fung. K: George. SC: Kleinman. MT. (2006). Personal
and ambient air pollution is associated with increased exhaled nitric oxide in children with asthma. Environ
Health Perspect 114: 1736-1743.
Delfino. RJ: Staimer. N: Tjoa. T: Arhami. M: Polidori. A: Gillen. PL: George. SC: Shafer. MM: Schauer. JJ:
Sioutas. C. (2010). Associations of primary and secondary organic aerosols with airway and systemic
inflammation in an elderly panel cohort. Epidemiology 21: 892-902.
http://dx.doi.org/10.1097/EDE.Ob013e3181f20e6c
Delfino. RJ: Staimer. N: Tjoa. T: Gillen. D: Kleinman. MT: Sioutas. C: Cooper. D. (2008a). Personal and
ambient air pollution exposures and lung function decrements in children with asthma. Environ Health
Perspect 116: 550-558. http://dx.doi.org/10.1289/ehp.10911
Delfino. RJ: Staimer. N: Tjoa. T: Gillen. PL: Polidori. A: Arhami. M: Kleinman. MT: Vaziri. ND: Longhurst. J:
Sioutas. C. (2009). Air pollution exposures and circulating biomarkers of effect in a susceptible population:
Clues to potential causal component mixtures and mechanisms. Environ Health Perspect 117: 1232-1238.
http://dx.doi.org/10.1289/ehp.0800194
Delfino. RJ: Staimer. N: Tjoa. T: Gillen. PL: Schauer. JJ: Shafer. MM. (2013). Airway inflammation and
oxidative potential of air pollutant particles in a pediatric asthma panel. J Expo Sci Environ Epidemiol 23:
466-473. http://dx.doi.org/10.1038/ies.2013.25
Delfino. RJ: Staimer. N: Tjoa. T: Polidori. A: Arhami. M: Gillen. PL: Kleinman. MT: Vaziri. ND: Longhurst. J:
Zaldivar. F: Sioutas. C. (2008b). Circulating biomarkers of inflammation, antioxidant activity, and platelet
activation are associated with primary combustion aerosols in subjects with coronary artery disease. Environ
Health Perspect 116: 898-906. http://dx.doi.org/10.1289/ehp. 11189
Delfino. RJ: Zeiger. RS: Seltzer. JM: Street DH: Mclaren. CE. (2002). Association of asthma symptoms with
peak paniculate air pollution and effect modification by anti-inflammatory medication use. Environ Health
Perspect 110: A607-A617.
Delgado-Saborit JM. (2012). Use of real-time sensors to characterise human exposures to combustion related
pollutants. J Environ Monit 14: 1824-1837. http://dx.doi.org/10.1039/c2eml0996d
R-24
-------
DeMarini DM: Shelton. ML: Kohan. MJ: Hudgens. EE: Kleindienst TE: Ball LM: Walsh. D: de Boer. JG:
Lewis-Bevan. L: Rabinowitz. J. R.: Claxton. LD: Lewtas. J. (2000). Mutagenicity in lung of Big Blue(R)
mice and induction of tandem-base substitutions in Salmonella by the air pollutant peroxyacetyl nitrate
(PAN): Predicted formation of intrastrand cross-links. Mutat Res-Fundam Mol Mech Mutagen 457: 41-55.
http://dx.doi.org/10.1016/S0027-5107(00)00121-4
Penman. KL: Brasseur. G: Chidthaisong. A: Ciais. P: Cox. PM: Dickinson. RE: Hauglustaine. D: Heinze. C:
Holland. E: Jacob. D: Lohmann. U: Ramachandran. S: da Silva Dias. PL: Wofsv. SC: Zhang. X. (2007).
Couplings between changes in the climate system and biogeochemistry. In Climate change 2007: The
physical science basis (pp. 499-587). New York, NY: Cambridge University Press.
http://www.ipcc.ch/publications and data/publications ipcc fourth assessment report wgl report the_phy
sical science basis.htm
Dennekamp. M: Akram. M: Abramson. MJ: Tonkin. A: Sim. MR: Fridman. M: Erbas. B. (2010). Outdoor air
pollution as a trigger for out-of-hospital cardiac arrests. Epidemiology 21: 494-500.
http://dx.doi.org/10.1097/EDE.Ob013e3181e093db
Desqueyroux. H: Pujet JC: Prosper. M: Le Moullec. Y: Momas. I. (2002). Effects of air pollution on adults with
chronic obstructive pulmonary disease. Arch Environ Occup Health 57: 554-560.
http://dx.doi.org/10.1080/00039890209602088
Devalia. JL: Rusznak. C: Herdman. MJ: Trigg. CJ: Tarraf. H: Davies. PJ. (1994). Effect of nitrogen-dioxide and
sulfur-dioxide on airway response of mild asthmatic-patients to allergen inhalation. Lancet 344: 1668-1671.
http://dx.doi.org/10.1016/80140-6736(94)90458-8
Devlin. RB: Horstman. DP: Gerritv. TR: Becker. S: Madden. MC. (1999). Inflammatory response in humans
exposed to 2.0 PPM nitrogen dioxide. Inhal Toxicol 11:89-109. http://dx.doi.org/10.1080/089583799197195
Pi Giovanni. V: Cagiano. R: Carratu. MR: De Salvia. MA: Giustino. A: Cuomo. V. (1994). Alterations in the
ontogeny of rat pup ultrasonic vocalization produced by prenatal exposure to nitrogen dioxide.
Psychopharmacology 116: 423-427.
di Sabatino. S: Kastner-Klein. P: Berkowicz. R: Britter. RE: Fedorovich. (2003). The modelling of turbulence
from traffic in urban dispersion models Part I: Theoretical considerations. Environ Fluid Mech 3: 129-143.
http://dx.doi.org/10.1023/A: 1022063608237
Diette. GB: Hansel NN: Buckley. TJ: Curtin-Brosnan. J: Eggleston. PA: Matsui EC: Mccormack. MC:
Williams. PL: Brevsse. PN. (2007). Home indoor pollutant exposures among inner-city children with and
without asthma. Environ Health Perspect 115: 1665-1669. http://dx.doi.org/10.1289/ehp. 10088
Diez Roux. AV: Mair. C. (2010). Neighborhoods and health. AnnN Y Acad Sci 1186: 125-145.
http://dx.doi.org/10.1111/i. 1749-6632.2009.05333.X
Dijkema. MB: Gehring. U: van Strien. RT: vanderZee. SC: Fischer. P: Hoek. G: Brunekreef. B. (201 la). A
comparison of different approaches to estimate small-scale spatial variation in outdoor NO concentrations.
Environ Health Perspect 119: 670-675. http://dx.doi.org/10.1289/ehp.0901818
Dijkema. MBA: Mallant SF: Gehring. U: van den Hurk. K: Alssema. M: van Strien. RT: Fischer. PH: Nijpels.
G: Stehouwer. CD A: Hoek. G: Dekker. JM: Brunekreef. B. (201 Ib). Long-term exposure to traffic-related air
pollution and type 2 diabetes prevalence in a cross-sectional screening-study in the Netherlands. Environ
Health 10: 76. http://dx.doi.org/10.1186/1476-069X-10-76
Dimakopoulou. K: Samoli. E: Beelen. R: Stafoggia. M: Andersen. ZJ: Hoffmann. B: Fischer. P:
Nieuwenhuijsen. M: Vineis. P: Xun. W: Hoek. G: Raaschou-Nielsen. O: Oudin. A: Forsberg. B: Modig. L:
Jousilahti. P: Lanki. T: Turunen. A: Oftedal. B: Nafstad. P: Schwarze. PE: Penell. J: Fratiglioni. L:
Andersson. N: Pedersea N: Korek. M: De Faire. U: Eriksen. KT: Tjenneland. A: Becker. T: Wang. M:
Bueno-De-Mesquita. B: Tsai. MY: Eeftens. M: Peeters. PH: Meliefste. K: Marcon. A: Kramer. U:
Kuhlbusch. TA: Vossoughi. M: Key. T: de Hoogh. K: Hampel R: Peters. A: Heinrich. J: Weinmayr. G:
Concin. H: Nagel G: Ineichen. A: Jacquemin. B: Stempfelet. M: Vilier. A: Ricceri. F: Sacerdote. C: Pedeli.
X: Katsoulis. M: Trichopoulou. A: Brunekreef. B: Katsouvanni. K. (2014). Air pollution and nonmalignant
respiratory mortality in 16 cohorts within the ESCAPE project. Am J Respir Crit Care Med 189: 684-696.
http://dx.doi.org/10.1164/rccm.201310-1777OC
R-25
-------
Dimeo. MJ: Glenn. MG: Holtzman. MJ: Sheller. JR: Nadel JA: Boushev. HA. (1981). Threshold concentration
of ozone causing an increase in bronchial reactivity in humans and adaptation with repeated exposures. Am
Rev Respir Dis 124: 245-248.
Dimitriou. K: Kassomenos. P. (2014). Local and regional sources of fine and coarse paniculate matter based on
traffic and background monitoring. Theor Appl Climatol 116: 413-433. http://dx.doi.org/10.1007/s00704-
013-0961-6
Dimitroulopoulou. C: Ashmore. MR: Hill. MTR: Byrne. MA: Kinnerslev. R. (2006). INDAIR: A probabilistic
model of indoor air pollution in UK homes. Atmos Environ 40: 6362-6379.
http://dx.doi.0rg/10.1016/i.atmosenv.2006.05.047
Dionisio. KL: Baxter. LK: Chang. HH. (2014). An empirical assessment of exposure measurement error and
effect attenuation in bipollutant epidemiologic models. Environ Health Perspect 122: 1216-1224.
http://dx.doi.org/10.1289/ehp.1307772
Dionisio. KL: Isakov. V: Baxter. LK: Sarnat JA: Sarnat SE: Burke. J: Rosenbaum. A: Graham. SE: Cook. R:
Mulholland. J: Oezkaynak. H. (2013). Development and evaluation of alternative approaches for exposure
assessment of multiple air pollutants in Atlanta, Georgia. JExpo Sci Environ Epidemiol 23: 581-592.
http://dx.doi.org/10.1038/ies.2013.59
Dockerv. DW: Pope. CA. Ill: Xu. X: Spengler. JD: Ware. JH: Fay. ME: Ferris. BG. Jr: Speizer. FE. (1993). An
association between air pollution and mortality in six US cities. N Engl J Med 329: 1753-1759.
http://dx.doi.org/10.1056/NEJM199312093292401
Dominici F: McDermott. A: Daniels. M: Zeger. SL: Samet JM. (2003). Mortality among residents of 90 cities
[HEI]. In Revised analyses of time-series studies of air pollution and health (pp. 9-24). Boston, MA: Health
Effects Institute. http://pubs.healtheffects.org/view.php?id=4
Dong. G: Qian. Z: Wang. J: Chen. W: Ma. W: Trevathan. E: Xaverius. PK: DeClue. R: Wiese. A: Langston. M:
Liu. MM: Wang. D: Ren. W. (2013a). Associations between ambient air pollution and prevalence of stroke
and cardiovascular diseases in 33 Chinese communities. Atmos Environ 77: 968-973.
http://dx.doi.0rg/10.1016/i.atmosenv.2013.06.034
Dong. GH: Chea T: Liu. MM: Wang. D: Ma. YN: Ren. WH: Lee. YU Zhao. YD: He. PC. (2011). Gender
differences and effect of air pollution on asthma in children with and without allergic predisposition:
Northeast Chinese children health study. PLoS ONE 6: e22470.
http://dx.doi.org/10.1371/iournal.pone.0022470
Dong. GH: Qian. ZM: Xaverius. PK: Trevathan. E: Maalouf. S: Parker. J: Yang. L: Liu. MM: Wang. D: Ren.
WH: Ma. W: Wang. J: Zelicoff. A: Fu. Q: Simckes. M. (2013b). Association between long-term air pollution
and increased blood pressure and hypertension in China. Hypertension 61: 578-584.
http://dx.doi.org/10.1161/HYPERTENSIONAHA.lll.00003
Dong. GH: Zhang. P: Sun. B: Zhang. L: Chea X: Ma. N: Yu. F: Guo. H: Huang. H: Lee. YL: Tang. N: Chea J.
(2012). Long-term exposure to ambient air pollution and respiratory disease mortality in Shenyang, China: A
12-year population-based retrospective cohort study. Respiration 84: 360-368.
http://dx.doi.org/10.1159/000332930
Dons. E: Van Poppel M: Int Panis. L: De Prins. S: Berghmans. P: Koppen. G: Matheeussen. C. (2014). Land use
regression models as a tool for short, medium and long term exposure to traffic related air pollution. Sci
Total Environ 476-477: 378-386. http://dx.doi.0rg/10.1016/i.scitotenv.2014.01.025
Douglas. GJ: Price. JF: Page. CP. (1995). The effect of prolonged exposure to NO2 from birth on airways
responsiveness in rabbits sensitized at birth. Eur Respir J 8: 246-252.
http://dx.doi.org/10.1183/09031936.95.08020246
Powell. AR: Kilburn. KH: Pratt PC. (1971). Short-term exposure to nitrogen dioxide: Effects on pulmonary
ultrastructure, compliance, and the surfactant system. Arch Intern Med 128: 74-80.
http://dx.doi.org/10.1001/archinte.1971.00310190078009
Drahl C. (2009). Probing for in-body ozone. Chem Eng News 87: 40-42. http://dx.doi.org/10.1021/cen-
v087n025.p040
R-26
-------
Drechsler-Parks. DM. (1995). Cardiac output effects of O3 and NO2 exposure in healthy older adults. Toxicol
Ind Health 11: 99-109. http://dx.doi.org/10.1177/074823379501100109
Dressel H: Muller. F: Fischer. R: Rommelt H: Hohlfeld. JM: Behr. J: Huber. RM: Nowak. D: Torres. RA.
(2010). Independent information of nonspecific biomarkers in exhaled breath condensate. Respiration 80:
401-409. http://dx.doi.org/10.1159/000319945
Drvden. DM: Spooner. CH: Stickland. MK: Vandermeer. B: Tjosvold. L: Bialy. L: Wong. K: Rowe. BH. (2010).
Exercise-induced bronchoconstriction and asthma. (AHRQ Publication No. 10-E001). Rockville, MD:
Agency for Healthcare Research and Quality, http://www.ahrq.gov/research/findings/evidence-based-
reports/eibeia-evidence-report.pdf
Du. X: Fu. L: Ge. W: Zhang. S: Wang. H. (2011). Exposure of taxi drivers and office workers to traffic-related
pollutants in Beijing: A note. Transport Res Transport Environ 16: 78-81.
http://dx.doi.0rg/10.1016/i.trd.2010.08.002
Duan. X: Buckpitt AR: Pinkerton. KE: Ji. C: Plopper. CG. (1996). Ozone-induced alterations in glutathione in
lung subcompartments of rats and monkeys. Am J Respir Cell Mol Biol 14: 70-75.
http://dx.doi.org/10.1165/aircmb. 14.1.8534488
Duan. X: Buckpitt. AR: Plopper. CG. (1993). Variation in antioxidant enzyme activities in anatomic
subcompartments within rat and rhesus monkey lung. Toxicol Appl Pharmacol 123: 73-82.
http://dx.doi.org/10.1006/taap. 1993.1223
Ducret-Stich. RE: Tsai MY: Ragettli. MS: Ineichen. A: Kuenzli. N: Phuleria. HC. (2013). Role of highway
traffic on spatial and temporal distributions of air pollutants in a Swiss Alpine valley. Sci Total Environ 456-
457: 50-60. http://dx.doi.0rg/10.1016/i.scitotenv.2013.03.065
Dunlea. EJ: Herndon. SC: Nelson. DP: Volkamer. RM: San Martini. F: Sheehy. PM: Zahniser. MS: Shorter. JH:
Wormhoudt JC: Lamb. BK: Allwine. EJ: Gaffnev. JS: Marlev. NA: Grutter. M: Marquez. C: Blanco. S:
Cardenas. B: Retama. A: Ramon Villegas. CR: Kolb. CE: Molina. LT: Molina. MJ. (2007). Evaluation of
nitrogen dioxide chemiluminescence monitors in a polluted urban environment. Atmos Chem Phys 7: 2691-
2704.
Dupuv. PM: Shore. SA: Drazen. JM: Frostell C: Hill. WA: Zapol WM. (1992). Bronchodilator action of
inhaled nitric oxide in guinea pigs. J Clin Invest 90: 421-428. http://dx.doi.org/10.1172/JCI115877
Durant JL: Ash. CA: Wood. EC: Herndon. SC: Jayne. JT: Knighton. WB: Canagaratna. MR: Trull JB: Brugge.
D: Zamore. W: Kolb. CE. (2010). Short-term variation in near-highway air pollutant gradients on a winter
morning. Atmos Chem Phys 10: 8341-8352. http://dx.doi.org/10.5194/acp-10-8341-2010
During. I: Bachlin. W: Ketzel. M: Baum. A: Friedrich. U: Wurzler. S. (2011). A new simplified NO/NO2
conversion model under consideration of direct NO2-emissions. Meteor Z 20: 67-73.
http://dx.doi.org/10.1127/0941-2948/2011/0491
Ebisu. K: Bell ML. (2012). Airborne PM2.5 chemical components and low birth weight in the Northeastern and
Mid-Atlantic regions of the United States. Environ Health Perspect 120: 1746-1752.
http://dx.doi.org/10.1289/ehp. 1104763
Eder. B: Yu. S. (2005). A performance evaluation of the 2004 release of Models-3 CMAQ. Atmos Environ 40:
4811-4824. http://dx.doi.0rg/10.1016/i.atmosenv.2005.08.045
Eeftens. M: Beelen. R: Fischer. P: Brunekreef B: Meliefste. K: Hoek. G. (2011). Stability of measured and
modelled spatial contrasts inNO2 overtime. Occup EnvironMed 68: 765-770.
http://dx.doi.org/10.1136/oem.2010.061135
R-27
-------
Eeftens. M: Tsai MY: Ampe. C: Anwander. B: Beelen. R: Bellander. T: Cesaroni. G: Cirach. M: Cyrys. J: de
Hoogh. K: De Nazelle. A: de Vocht F: Declercq. C: Dedele. A: Eriksen. K: Galassi C: Grazuleviciene. R:
Grivas. G: Heinrich. J: Hoffmann. B: Lakovides. M: Ineichen. A: Katsouvannj K: Korek. M: Kraemer. U:
Kuhlbusch. T: Lanki. T: Madsen. C: Meliefste. K: Moelter. A: Mosler. G: Nieuwenhuijsen. M: Oldenwening.
M: Pennanen. A: Probst-Hensch. N: Quass. U: Raaschou-Nielsen. O. le: Ranzi. A: Stephanou. E: Sugiri. D:
Udvardv. O: Vaskoevi E. va: Weinmayr. G: Brunekreef B: Hoek. G. (2012). Spatial variation of PM2.5,
PM10, PM2.5 absorbance and PMcoarse concentrations between and within 20 European study areas and the
relationship with NO2 - Results of the ESCAPE project. Atmos Environ 62: 303-317.
http://dx.doi.0rg/10.1016/i.atmosenv.2012.08.038
Eenhuizen. E: Gehring. U: Wijga. AH: Smit HA: Fischer. PH: Brauer. M: Koppelman. GH: Kerkhof M: de
Jongste. JC: Brunekreef. B: Hoek. G. (2013). Traffic related air pollution is related to interrupter resistance in
four-year old children. Eur Respir J 41: 1257-1263. http://dx.doi.org/10.1183/09031936.00020812
Ehrlich. R. (1980). Interaction between environmental pollutants and respiratory infections. Environ Health
Perspect 35: 89-99.
Ehrlich. R: Findlav. JC: Fenters. JD: Gardner. DE. (1977). Health effects of short-term inhalation of nitrogen
dioxide and ozone mixtures. Environ Res 14: 223-231. http://dx.doi.org/10.1016/0013-9351(77)90034-2
Ehrlich. R: Findlav. JC: Gardner. DE. (1979). Effects of repeated exposures to peak concentrations of nitrogen
dioxide and ozone on resistance to streptococcal pneumonia. J Toxicol Environ Health 5: 631-642.
http://dx.doi.org/10.1080/15287397909529775
Ehrlich. R: Henry. MC. (1968). Chronic toxicity of nitrogen dioxide: I: Effect on resistance to bacterial
pneumonia. Arch Environ Health 17: 860-865. http://dx.doi.org/10.1080/00039896.1968.10665342
EIONET (European Environment Information and Observation Netork). (2014). European topic centre on air
pollution and climate change mitigation: AirBase- the European air quality database. Available online at
http://acm.eionet.europa.eu/databases/airbase
Elsaved. NM: Mustafa. MG. (1982). Dietary antioxidants and the biochemical response to oxidant inhalation: I:
Influence of dietary vitamin E on the biochemical effects of nitrogen dioxide exposure in rat lung. Toxicol
Appl Pharmacol 66: 319-328. http://dx.doi.org/10.1016/0041-008X(82)90298-8
Enami S: Hoffmann. MR: Colussi AJ. (2009). Absorption of inhaled NO(2) [Letter]. JPhys ChemB 113: 7977-
7981. http://dx.doi.org/10.1021/ip902667x
England. C: Corcoran. WH. (1974). Kinetics and mechanisms of the gas-phase reaction of water vapor and
nitrogen dioxide. IndEng ChemFundam 13: 373-384. http://dx.doi.org/10.1021/il60052a014
Enkhmaa. D: Warburton. N: Javzandulam. B: Uyanga. J: Khishigsuren. Y: Lodovsamba. S: Enkhtur. S:
Warburton. D. (2014). Seasonal ambient air pollution correlates strongly with spontaneous abortion in
Mongolia. BMC Pregnancy Childbirth 14: 146. http://dx.doi.org/10.1186/1471-2393-14-146
Ensor. KB: Raun. LH: Persse. D. (2013). A case-crossover analysis of out-of-hospital cardiac arrest and air
pollution. Circulation 127: 1192-1199. http://dx.doi.org/10.116 I/CIRCULATION AH A. 113.000027
Eriksson. C: Hilding. A: Pyko. A: Bluhm. G: Pershagen. G: Ostenson. CG. (2014). Long-term aircraft noise
exposure and body mass index, waist circumference, and type 2 diabetes: A prospective study. Environ
Health Perspect 122: 687-694. http://dx.doi.org/10.1289/ehp.1307115
Erjefalt JS: Greiff. L: Andersson. M: Matsson. E: Petersen. H: Linden. M: Ansari. T: Jeffery. PK: Persson. CG.
(1999). Allergen-induced eosinophil cytolysis is a primary mechanism for granule protein release in human
upper airways. Am J Respir Crit Care Med 160: 304-312. http://dx.doi.org/10.1164/airccm. 160.1.9809048
Escamilla-Nunez. MC: Barraza-Villarreal A: Hernandez-Cadena. L: Moreno-Macias. H: Ramirez-Aguilar. M:
Sienra-Monge. JJ: Cortez-Lugo. M: Texcalac. JL: del Rio-Navarro. B: Romieu. I. (2008). Traffic-related air
pollution and respiratory symptoms among asthmatic children, resident in Mexico City: The EVA cohort
study. Respir Res 9: 74. http://dx.doi.org/10.1186/1465-9921-9-74
R-28
-------
Esplugues. A: Ballester. F: Estarlich. M: Llop. S: Fuentes-Leonarte. V: Mantilla. E: Vioque. J: Ifiiguez. C.
(2011). Outdoor, but not indoor, nitrogen dioxide exposure is associated with persistent cough during the first
year of life. Sci Total Environ 409: 4667-4673. http://dx.doi.0rg/10.1016/i.scitotenv.2011.08.007
Esposito. S: Galeone. C: Lelii. M: Longhi. B: Ascolese. B: Senatore. L: Prada. E: Montinaro. V: Malerba. S:
Patria. MF: Principi N. (2014). Impact of air pollution on respiratory diseases in children with recurrent
wheezing or asthma. BMC Pulm Med 14:130. http://dx.doi.org/10.1186/1471-2466-14-130
Estarlich. M: Ballester. F: Aguilera. I: Fernandez-Somoano. A: Lertxundi A: Llop. S: Freire. C: Tardon. A:
Basterrechea. M: Sunyer. J: Ifiiguez. C. (2011). Residential exposure to outdoor air pollution during
pregnancy and anthropometric measures at birth in a multicenter cohort in Spain. Environ Health Perspect
119: 1333-1338. http://dx.doi.org/10.1289/ehp.1002918
Eze. 1C: Schaffner. E: Fischer. E: Schikowski. T: Adam. M: Imboden. M: Tsai M: Carballo. D: von Eckardstein.
A: Kunzli. N: Schindler. C: Probst-Hensch. N. (2014). Long-term air pollution exposure and diabetes in a
population-based Swiss cohort. Environ Int 70: 95-105. http://dx.doi.0rg/10.1016/i.envint.2014.05.014
Ezrattv. V: Guillossou. G: Neukirch. C: Dehoux. M: Koscielny. S: Bonav. M: Cabanes. PA: Samet JM: Mure.
P: Ropert. L: Tokarek. S: Lambrozo. J: Aubier. M. (2014). Repeated nitrogen dioxide exposures and
eosinophilic airway inflammation in asthmatics: a randomized crossover study. Environ Health Perspect 122:
850-855. http://dx.doi.org/10.1289/ehp. 1307240
Faiz. AS: Rhoads. GG: Demissie. K: Kruse. L: Lin. Y: Rich. DQ. (2012). Ambient air pollution and the risk of
stillbirth. Am JEpidemiol 176: 308-316. http://dx.doi.org/10.1093/aie/kws029
Faiz. AS: Rhoads. GG: Demissie. K: Lin. Y: Kruse. L: Rich. DQ. (2013). Does ambient air pollution trigger
stillbirth? Epidemiology 24: 538-544. http://dx.doi.org/10.1097/EDE.Ob013e3182949ce5
Faustini. A: Stafoggia. M: Berti. G: Bisanti L: Chiusolo. M: Cernigliaro. A: Mallone. S: Primerano. R: Scarnato.
C: Simonato. L: Vigotti. MA: Forastiere. F. (2011). The relationship between ambient paniculate matter and
respiratory mortality: A multi-city study in Italy. Eur Respir J 38: 538-547.
http://dx.doi.org/10.1183/09031936.00093710
Faustini. A: Stafoggia. M: Colais. P: Berti. G: Bisanti. L: Cadum E: Cernigliaro. A: Mallone. S: Scarnato. C:
Forastiere. F. (2013). Air pollution and multiple acute respiratory outcomes. Eur Respir J 42: 304-313.
http://dx.doi.org/10.1183/09031936.00128712
Fehrenbach. H: Zimmermann. G: Starke. E: Bratu. VA: Conrad. D: Yildirim. AO: Fehrenbach. A. (2007).
Nitrogen dioxide induces apoptosis and proliferation but not emphysema in rat lungs. Thorax 62: 438-446.
http://dx.doi.org/10.1136/thx.2006.062364
Fehsenfeld. FC: Dickerson. RR: Hubler. G: Luke. WT: Nunnermacker. LJ: Williams. EJ: Roberts. JM: Calvert.
JG: Oman. CM: Delany. AC: Eubank. CS: Fahev. DW: Fried. A: Gandrud. BW: Langford. AO: Murphy.
PC: Norton. RB: Pickering. KE: Ridley. BA. (1987). A ground-based intercomparison of NO, NOx, and NOy
measurement techniques. J Geophys Res 92: 14,710-714,722. http://dx.doi.org/10.1029/JD092iD12pl4710
Feichtinger. W: Papalambrou. K: Poehl. M: Krischker. U: Neumann. K. (1997). Smoking and in vitro
fertilization: A meta-analysis. J Assist Reprod Genet 14: 596-599.
http://dx.doi.Org/10.1023/A:1022584802711
Felber Dietrich. D: Gemperli. A: Gaspoz. JM: Schindler. C: Liu. LJ: Gold. PR: Schwartz. J: Rochat. T:
Barthelemy. JC: Pons. M: Roche. F: Probst Hensch. NM: Bridevaux. PO: Gerbase. MW: Neu. U:
Ackermann-Liebrich. U. (2008). Differences in heart rate variability associated with long-term exposure to
NO2. Environ Health Perspect 116: 1357-1361. http://dx.doi.org/10.1289/ehp. 11377
Fenters. JD: Findlav. JC: Port. CD: Ehrlich. R: Coffin. PL. (1973). Chronic exposure to nitrogen dioxide:
Immunologic, physiologic, and pathologic effects in virus-challenged squirrel monkeys. Arch Environ Health
27: 85-89. http://dx.doi.org/10.1080/00039896.1973.10666324
Feo Brito. F: Mur Gimeno. P: Martinez. C: Tobias. A: Suarez. L: Guerra. F: Borja. JM: Alonso. AM. (2007). Air
pollution and seasonal asthma during the pollen season: A cohort study in Puertollano and Ciudad Real
(Spain). Allergy 62: 1152-1157. http://dx.doi.0rg/10.llll/i.1398-9995.2007.01438.x
R-29
-------
Ferin. J: Leach. LJ. (1975). The effects of selected air pollutants on clearance of titanic oxide particles from the
lungs of rats. Inhaled Particles 4 Pt 1: 333-341.
Fernandez-Somoano. A: Estarlich. M: Ballester. F: Fernandez-Patier. R: Aguirre-Alfaro. A: Herce-Garraleta.
MD: Tardon. A. (2011). Outdoor NO(2) and benzene exposure in the INMA (Environment and Childhood)
Asturias cohort (Spain). Atmos Environ 45: 5240-5246. http://dx.doi.0rg/10.1016/i.atmosenv.2011.02.010
Fernando. HJS. (2010). Fluid dynamics of urban atmospheres in complex terrain. Annu Rev Fluid Mech 42: 365-
389. http://dx.doi.org/10.1146/annurev-fluid-121108-145459
Ferris. BG. (1978). Epidemiology standardization project II: Recommended respiratory disease questionnaires
for use with adults and children in epidemiological research. Am J Respir Crit Care Med 118: 7-53.
Filho. MAP: Pereira. LAA: Arbex. FF: Arbex. M: Conceicao. GM: Santos. UP: Lopes. AC: Saldiva. PHN:
Braga. ALF: Cendon. S. (2008). Effect of air pollution on diabetes and cardiovascular diseases in Sao Paulo,
Brazil. Braz J Med Biol Res 41: 526-532. http://dx.doi.org/10.1590/S0100-879X2008005000020
Filleul L: Rondeau. V: Vandentorren. S: Le Moual N: Cantagrel A: Annesi-Maesano. I: Charpin. D: Declercq.
C: Neukirch. F: Paris. C: Vervloet D: Brochard. P: Tessier. JF: Kauffmann. F: Baldi. I. (2005). Twenty five
year mortality and air pollution: Results from the French PAARC survey. Occup Environ Med 62: 453-460.
http://dx.doi.org/10.1136/oem.2004.014746
Finegold. JA: Asaria. P: Francis. DP. (2013). Mortality from ischaemic heart disease by country, region, and age:
Statistics from World Health Organisation and United Nations. Int J Cardiol 168: 934-945.
http://dx.doi.0rg/10.1016/i.ijcard.2012.10.046
Finlavson-Pitts. BJ: Sweetman. LL: Weissbart. B. (1987). A Fourier transform infrared spectrometry study of the
reactions of phosphatidylcholines with gaseous N2O5 and NO2. Toxicol Appl Pharmacol 89: 438-448.
http://dx.doi.org/10 J016/0041-008X(87)90163-3
Finlavson-Pitts. BJ: Wingen. LM: Sumner. AL: Svomin. D: Ramazan. KA. (2003). The heterogeneous
hydrolysis of NO2 in laboratory systems and in outdoor and indoor atmospheres: An integrated mechanism.
Phys Chem ChemPhys 5: 223-242. http://dx.doi.org/10.1039/b208564j
Finn. D: Clawson. KL: Carter. RG: Rich. JD: Eckman. RM: Perry. SG: Isakov. V: Heist DK. (2010). Tracer
studies to characterize the effects of roadside noise barriers on near-road pollutant dispersion under varying
atmospheric stability conditions. Atmos Environ 44: 204-214.
http://dx.doi.0rg/10.1016/i.atmosenv.2009.10.012
Fischer. EV: Jacob. DJ: Yantosca. RM: Sulprizio. MP: Millet DB: Mao. J: Paulot F: Singh. HB: Roiger. A:
Ries. L: Talbot RW: Dzepina. K: Pandev Deolal S. (2014). Atmospheric peroxyacetyl nitrate (PAN): a
global budget and source attribution. Atmos Chem Phys 14: 2679-2698. http://dx.doi.org/10.5194/acp-14-
2679-2014
Flamant-Hulin. M: Caillaud. D: Sacco. P: Penard-Morand. C: Annesi-Maesano. I. (2010). Air pollution and
increased levels of fractional exhaled nitric oxide in children with no history of airway damage. J Toxicol
Environ Health A 73: 272-283. http://dx.doi.org/10.1080/15287390903249206
Folev. KM: Roselle. SJ: Appel KW: Bhave. PV: Pleim JE: Otte. TL: Mathur. R: Sarwar. G: Young. JO:
Gilliam RC: Nolte. CG: Kelly. JT: Gilliland. AB: Bash. JO. (2010). Incremental testing of the Community
Multiscale Air Quality (CMAQ) modeling system version 4.7. Geoscientific Model Development 3: 205-226.
Folinsbee. LJ. (1992). Does nitrogen dioxide exposure increase airways responsiveness? Toxicol Ind Health 8:
273-283.
Folinsbee. LJ: Horvath. SM: Bedi. JF: Delehunt JC. (1978). Effect of 0.62 ppm NO2 on cardiopulmonary
function in young male nonsmokers. Environ Res 15: 199-205. http://dx.doi.org/10.1016/0013-
9351(78)90096-8
Foraster. M: Basagana. X: Aguilera. I: Rivera. M: Agis. D: Bouso. L: Deltell A: Marrugat J: Ramos. R: Sunyer.
J: Vila. J: Elosua. R: Kiinzli. N. (2014). Association of long-term exposure to traffic-related air pollution with
blood pressure and hypertension in an adult population-based cohort in Spain (the REGICOR Study).
Environ Health Perspect 122: 404-411. http://dx.doi.org/10.1289/ehp. 1306497
R-30
-------
Forbes. LJ: Patel MD: Rudnicka. AR: Cook. DG: Bush. T: Stedman. JR: Whincup. PH: Strachan. DP:
Anderson. HR. (2009a). Chronic exposure to outdoor air pollution and diagnosed cardiovascular disease:
meta-analysis of three large cross-sectional surveys. Environ Health 8: 30. http://dx.doi.org/10.1186/1476-
069X-8-30
Forbes. LJL: Kapetanakis. V: Rudnicka. AR: Cook. DG: Bush. T: Stedman. JR: Whincup. PH: Strachan. DP:
Anderson. HR. (2009b). Chronic exposure to outdoor air pollution and lung function in adults. Thorax 64:
657-663. http://dx.doi.org/10.1136/thx.2008.109389
Ford. E: Hughes. MN: Wardman. P. (2002). Kinetics of the reactions of nitrogen dioxide with glutathione,
cysteine, and uric acid at physiological pH. Free Radic Biol Med 32: 1314-1323.
http://dx.doi.org/10.1016/S0891-5849(02)00850-X
Forsberg. B: Stjernberg. N: Linne. R: Segerstedt B: Wall S. (1998). Daily air pollution levels and acute asthma
in southern Sweden. EurRespir J 12: 900-905.
Forstermann. U: Sessa. WC. (2012). Nitric oxide synthases: Regulation and function [Review]. Eur Heart J 33:
829-837. http://dx.doi.org/10.1093/eurhearti/ehr304
Fox. GA. (1991). Practical causal inference for ecoepidemiologists. J Toxicol Environ Health A 33: 359-373.
http://dx.doi.org/10.1080/15287399109531535
Fraenkel DJ: Bardin. PG: Sanderson. G: Lampe. F: Johnston. SL: Holgate. ST. (1995). Lower airways
inflammation during rhinovirus colds in normal and in asthmatic subjects. Am J Respir Crit Care Med 151:
879-886. http://dx.doi.org/10.1164/airccrn/15L3 Pt 1.879
Frampton. MW: Boscia. J: Roberts. NJ. Jr: Azadniv. M: Torres. A: Cox. C: Morrow. PE: Nichols. J: Chalupa. D:
Frasier. LM: Gibb. FR: Speers. DM: Tsai Y: Utell MJ. (2002). Nitrogen dioxide exposure: Effects on
airway and blood cells. Am J Physiol 282: L155-L165.
Frampton. MW: Morrow. PE: Cox. C: Gibb. FR: Speers. DM: Utell MJ. (1991). Effects of nitrogen dioxide
exposure on pulmonary function and airway reactivity in normal humans. Am Rev Respir Dis 143: 522-527.
http://dx.doi.0rg/10.1164/airccm/143.3.522
Frampton. MW: Smeglin. AM: Roberts. NJ. Jr: Finkelstein. JN: Morrow. PE: Utell MJ. (1989). Nitrogen
dioxide exposure in vivo and human alveolar macrophage inactivation of influenza virus in vitro. Environ
Res 48: 179-192. http://dx.doi.org/10.1016/S0013-9351(89)80033-7
Franck. U: Leitte. AM: Suppan. P. (2014). Multiple exposures to airborne pollutants and hospital admissions due
to diseases of the circulatory system in Santiago de Chile. Sci Total Environ 468-469: 746-756.
http://dx.doi.0rg/10.1016/i.scitotenv.2013.08.088
Franklin. M: Vora. H: Avol. E: Mcconnell. R: Lurmann. F: Liu. F: Penfold. B: Berhane. K: Gilliland. F:
Gauderman. WJ. (2012). Predictors of intra-community variation in air quality. J Expo Sci Environ
Epidemiol 22: 135-147. http://dx.doi.org/10.1038/ies.2011.45
Freeman. G: Furiosi NJ: Havdon. GB. (1966). Effects of continuous exposure of 0.8 ppm NO2 on respiration of
rats. Arch Environ Health 13: 454-456. http://dx.doi.org/10.1080/00039896.1966.10664595
Freire. C: Ramos. R: Puertas. R: Lopez-Espinosa. MJ: Julvez. J: Aguilera. I: Cruz. F: Fernandez. MF: Sunyer. J:
Olea. N. (2010). Association of traffic-related air pollution with cognitive development in children. J
Epidemiol Community Health 64: 223-228. http://dx.doi.org/10.1136/iech.2008.084574
Fruin. S: Westerdahl D: Sax. T: Sioutas. C: Fine. PM. (2008). Measurements and predictors of on-road ultrafine
particle concentrations and associated pollutants in Los Angeles. Atmos Environ 42: 207-219.
http://dx.doi.0rg/10.1016/i.atmosenv.2007.09.057
Fu. A: Leaderer. BP: Gent JF: Leaderer. D: Zhu. Y. (2012). An environmental epigenetic study of ADRB2 5'-
UTR methylation and childhood asthma severity. Clin Exp Allergy 42: 1575-1581.
http://dx.doi.org/10.1111/i. 1365-2222.2012.04055.X
Fuchs. H: Dube. WP: Lerner. BM: Wagner. NL: Williams. EJ: Brown. SS. (2009). A sensitive and versatile
detector for atmospheric NO2 and NOx based on blue diode laser cavity ring-down spectroscopy. Environ
SciTechnol43: 7831-7836. http://dx.doi.org/10.1021/es902067h
R-31
-------
Fuentes. M. (2009). Statistical issues in health impact assessment at the state and local levels. Air Qua! Atmos
Health 2: 47-55. http://dx.doi.org/10.1007/sl 1869-009-0033-3
Fuentes. M: Raftery. AE. (2005). Model evaluation and spatial interpolation by Bayesian combination of
observations with outputs from numerical models. Biometrics 61: 36-45. http://dx.doi.org/10. Ill l/j.0006-
341X.2005.030821.X
Fuertes. E: Standl. M: Cyrys. J: Berdel D: von Berg. A: Bauer. CP: Kramer. U: Sugiri. D: Lehmann. I:
Koletzko. S: Carlsten. C: Brauer. M: Heinrich. J. (2013). A longitudinal analysis of associations between
traffic-related air pollution with asthma, allergies and sensitization in the GINIplus and LISAplus birth
cohorts. Peer J 1: el93. http://dx.doi.org/10.7717/peerj.193
Fujimaki. H: Nohara. O. (1994). Changes in the response of lung mast cells isolated from rats and guinea pigs
exposed to nitrogen dioxide. Inhal Toxicol 6: 515-520. http://dx.doi.org/10.3109/08958379409040508
Fujita. EM: Campbell. DE: Zielinska. B: Arnott WP: Chow. JC. (2011). Concentrations of air toxics in motor
vehicle-dominated environments (pp. 3-77). (ISSN 1041-5505
156). Boston, MA: Health Effects Institute.
Fukuto. J: Carrington. SJ: Tantillo. DJ: Harrison. JG: Ignarro. LJ: Freeman. BA: Chen. A: Wink. DA. (2012).
Small molecule signaling agents: The integrated chemistry and biochemistry of nitrogen oxides, oxides of
carbon, dioxygen, hydrogen sulfide, and their derived species [Review]. Chem Res Toxicol 25: 769-793.
http://dx.doi.org/10.1021/tx2005234
Fung. KY: Khan. S: Krewski. D: Chen. Y. (2006). Association between air pollution and multiple respiratory
hospitalizations among the elderly in Vancouver, Canada. Inhal Toxicol 18: 1005-1011.
http://dx.doi.org/10.1080/08958370600904538
Fung. KY: Luginaah. I: Gorev. KM: Webster. G. (2005). Air pollution and daily hospital admissions for
cardiovascular diseases in Windsor, Ontario. Can J Public Health 96: 29-33.
Furiosi NJ: Crane. SC: Freeman. G. (1973). Mixed sodium chloride aerosol and nitrogen dioxide in air:
Biological effects on monkeys and rats. Arch Environ Occup Health 27: 405-408.
Gair. AJ: Penkett SA: Oyola. P. (1991). Development of a simple passive technique for the determination of
nitrogen-dioxide in remote continental locations. Atmos Environ 25A: 1927-1939.
Gan. W: Koehoorn. M: Davies. H: Demers. P: Tamburic. L: Brauer. M. (2011). Long-term exposure to traffic-
related air pollution and the risk of coronary heart disease hospitalization and mortality. Environ Health
Perspect 119: 501-507. http://dx.doi.org/10.1289/ehp. 1002511
Gan. WQ: Davies. HW: Koehoorn. M: Brauer. M. (2012a). Association of long-term exposure to community
noise and traffic-related air pollution with coronary heart disease mortality. Am J Epidemiol 175: 898-906.
http://dx.doi.org/10.1093/aie/kwr424
Gan. WQ: Fitzgerald. JM: Carlsten. C: Sadatsafavi M: Brauer. M. (2013). Associations of ambient air pollution
with chronic obstructive pulmonary disease hospitalization and mortality. Am J Respir Crit Care Med 187:
721-727. http://dx.doi.org/10.1164/rccm.201211-2004OC
Gan. WQ: Mclean. K: Brauer. M: Chiarello. SA: Davies. HW. (2012b). Modeling population exposure to
community noise and air pollution in a large metropolitan area. Environ Res 116: 11-16.
http://dx.doi.0rg/10.1016/i.envres.2012.04.001
Gao. Y: Chan. EY: Li. LP: He. 00: Wong. TW. (2013). Chronic effects of ambient air pollution on lung
function among Chinese children. ArchDis Child 98: 128-135. http://dx.doi.org/10.1136/archdischild-2011-
301541
Gardner. DE: Miller. FJ: Blommer. EJ: Coffin. PL. (1979). Influence of exposure mode on the toxicity of NO2.
Environ Health Perspect 30: 23-29.
Gasana. J: Dillikar. D: Mendv. A: Forno. E: Ramos Vieira. E. (2012). Motor vehicle air pollution and asthma in
children: A meta-analysis. Environ Res 117: 36-45. http://dx.doi.0rg/10.1016/j.envres.2012.05.001
R-32
-------
Gass. K: Klein. M: Chang. HH: Flanders. WD: Strickland. MJ. (2014). Classification and regression trees for
epidemiologic research: an air pollution example. Environ Health 13: 17. http://dx.doi.org/10.1186/1476-
069X-13-17
Gastoa B: Sears. S: Woods. J: Hunt. J: Ponaman. M: McMahon. T: Stamler. J. (1998). Bronchodilator S-
nitrosothiol deficiency in asthmatic respiratory failure. Lancet 351: 1317-1319.
http://dx.doi.org/10.1016/80140-6736(97)07485-0
Gauderman. WJ: Avol E: Gilliland. F: Vora. H: Thomas. D: Berhane. K: Mcconnell R: Kuenzli. N: Lurmann.
F: Rappaport. E: Margolis. H: Bates. D: Peters. J. (2004). The effect of air pollution on lung development
from 10 to 18 years of age. N Engl J Med 351: 1057-1067. http://dx.doi.org/10.1056/NEJMoa040610
Gauderman. WJ: Avol E: Lurmann. F: Kuenzli. N: Gilliland. F: Peters. J: McConnell R. (2005). Childhood
asthma and exposure to traffic and nitrogen dioxide. Epidemiology 16: 737-743.
http://dx.doi.org/10.1097/01.ede.0000181308.51440.75
Gauderman. WJ: Vora. H: McConnell. R: Berhane. K: Gilliland. F: Thomas. D: Lurmann. F: Avol. E: Kunzli. N:
Jerrett M: Peters. J. (2007). Effect of exposure to traffic on lung development from 10 to 18 years of age: A
cohort study. Lancet 369: 571-577. http://dx.doi.org/10.1016/S0140-6736(07)60037-3
Gauvin. S: Le Moullec. Y: Bremont F: Momas. I: Balducci. F: Ciognard. F: Poilve. MP: Zmirou. D. (2001).
Relationships between nitrogen dioxide personal exposure and ambient air monitoring measurements among
children in three French metropolitan areas: VESTA study. Arch Environ Occup Health 56: 336-341.
http://dx.doi.org/10.1080/00039890109604465
Gee. GC: Pavne-Sturges. DC. (2004). Environmental health disparities: A framework integrating psychosocial
and environmental concepts [Review]. Environ Health Perspect 112: 1645-1653.
http://dx.doi.org/10.1289/ehp.7074
Geer. LA: Weedon. J: Bell. ML. (2012). Ambient air pollution and term birth weight in Texas from 1998 to
2004. J Air Waste Manag Assoc 62: 1285-1295. http://dx.doi.org/10.1080/10962247.2012.707632
Gehring. U: Cyrys. J: Sedlmeir. G: Brunekreef. B: Bellander. T: Fischer. P: Bauer. CP: Reinhardt D:
Wichmann. HE: Heinrich. J. (2002). Traffic-related air pollution and respiratory health during the first 2 yrs
of life. Eur Respir J 19: 690-698. http://dx.doi.org/10.1183/09031936.02.01182001
Gehring. U: Gruzieva. O: Agius. RM: Beelen. R: Custovic. A: Cyrys. J: Eeftens. M: Flexeder. C: Fuertes. E:
Heinrich. J: Hoffmann. B: de Jongste. JC: Kerkhof. M: Klumper. C: Korek. M: Molter. A: Schultz. ES:
Simpson. A: Sugiri. D: Svartengren. M: von Berg. A: Wijga. AH: Pershagen. G: Brunekreef. B. (2013). Air
pollution exposure and lung function in children: the ESCAPE project. Environ Health Perspect 121: 1357-
1364. http://dx.doi.org/10.1289/ehp. 1306770
Gehring. U: Heinrich. J: Kramer. U: Grote. V: Hochadel M: Sugiri. D: Kraft M: Rauchfuss. K: Eberwein. HG:
Wichmann. HE. (2006). Long-term exposure to ambient air pollution and cardiopulmonary mortality in
women. Epidemiology 17: 545-551. http://dx.doi.org/10.1097/01.ede.0000224541.38258.87
Gehring. U: Tamburic. L: Sbihj H: Davies. HW: Brauer. M. (2014). Impact of noise and air pollution on
pregnancy outcomes. Epidemiology 25: 351-358. http://dx.doi.org/10.1097/EDE.0000000000000073
Gehring. U: van Eijsden. M: Dijkema. MB: van der Wai. MF: Fischer. P: Brunekreef. B. (2011 a). Traffic-related
air pollution and pregnancy outcomes in the Dutch ABCD birth cohort study. Occup Environ Med 68: 36-43.
http://dx.doi.org/10.1136/oem.2009.053132
Gehring. U: Wijga. AH: Brauer. M: Fischer. P: de Jongste. JC: Kerkhof. M: Oldenwening. M: Smit HA:
Brunekreef. B. (2010). Traffic-related air pollution and the development of asthma and allergies during the
first 8 years of life. Am J Respir Crit Care Med 181: 596-603. http://dx.doi.org/10.1164/rccm.2009Q6-
0858OC
Gehring. U: Wijga. AH: Fischer. P: de Jongste. JC: Kerkhof. M: Koppelman. GH: Smit HA: Brunekreef. B.
(201 Ib). Traffic-related air pollution, preterm birth and term birth weight in the PIAMA birth cohort study.
Environ Res 111: 125-135. http://dx.doi.0rg/10.1016/i.envres.2010.10.004
R-33
-------
Gelb. AF: Barnes. PJ: George. SC: Ricciardolo. FL: Dimaria. G: Zamel N. (2012). Review of exhaled nitric
oxide in chronic obstructive pulmonary disease [Review]. J Breath Res 6: 047101.
http://dx.doi.Org/10.1088/1752-7155/6/4/047101
Gent. JF: Koutrakis. P: Belanger. K: Triche. E: Holford. TR: Bracken. MB: Leaderer. BP. (2009). Symptoms and
medication use in children with asthma and traffic-related sources of fine particle pollution. Environ Health
Perspect 117: 1168-1174. http://dx.doi.org/10.1289/ehp.0800335
Geyh. AS: Xue. J: Ozkaynak. H: Spengler. JD. (2000). The Harvard Southern California chronic ozone exposure
study: Assessing ozone exposure of grade-school-age children in two southern California communities.
Environ Health Perspect 108: 265-270.
Ghannam. K: El-Fadel M. (2013). A framework for emissions source apportionment in industrial areas:
MM5/CALPUFF in a near-field application. J Air Waste Manag Assoc 63: 190-204.
http://dx.doi.org/10.1080/10962247.2012.739982
Ghosh. JK: Heck. JE: Cockburn. M: Su. J: Jerrett M: Ritz. B. (2013). Prenatal exposure to traffic-related air
pollution and risk of early childhood cancers. Am J Epidemiol 178: 1233-1239.
http://dx.doi.org/10.1093/aie/kwtl29
Ghosh. JKC: Wilhelm. M: Su. J: Goldberg. D: Cockburn. M: Jerrett M: Ritz. B. (2012a). Assessing the
influence of traffic-related air pollution on risk of term low birth weight on the basis of land-use-based
regression models and measures of air toxics. Am J Epidemiol 175: 1262-1274.
http://dx.doi.org/10.1093/aie/kwr469
Ghosh. R: Joad. J: Benes. I: Postal. M: Sram RJ: Hertz-Picciotto. I. (2012b). Ambient nitrogen oxides exposure
and early childhood respiratory illnesses. Environ Int 39: 96-102.
http://dx.doi.0rg/10.1016/i.envint.2011.10.001
Gibson. MD: Kundu. S: Satish. M. (2013). Dispersion model evaluation of PM2.5, NOX and SO2 from point
and major line sources in Nova Scotia, Canada using AERMOD Gaussian plume air dispersion model.
Atmos Pollut Res 4: 157-167. http://dx.doi.org/10.5094/APR.2013.016
Gilbert. NL: Goldberg. MS: Brook. JR: Jerrett. M. (2007). The influence of highway traffic on ambient nitrogen
dioxide concentrations beyond the immediate vicinity of highways. Atmos Environ 41: 2670-2673.
http://dx.doi.0rg/10.1016/i.atmosenv.2006.12.007
Gilbert. NL: Woodhouse. S: Stieb. DM: Brook. JR. (2003). Ambient nitrogen dioxide and distance from a major
highway. Sci Total Environ 312: 43-46. http://dx.doi.org/10.1016/80048-9697(03)00228-6
Gilbraith. N: Powers. SE. (2013). Residential demand response reduces air pollutant emissions on peak
electricity demand days in New York City. Energy Policy 59: 459-469.
http://dx.doi.0rg/10.1016/i.enpol.2013.03.056
Gillespie-Bennett J: Pierse. N: Wickens. K: Crane. J: Howden-Chapman. P. (2011). The respiratory health
effects of nitrogen dioxide in children with asthma. Eur Respir J 38: 303-309.
http://dx.doi.org/10.1183/09031936.00115409
Gilmour. MI: Park. P: Selgrade. MK. (1996). Increased immune and inflammatory responses to dust mite
antigen in rats exposed to 5 ppm NO2. Fundam Appl Toxicol 31: 65-70.
http://dx.doi.org/10.1006/faat. 1996.0076
Gladwin. MT: Schechter. AN: Kim-Shapiro. DB: Patel RP: Hogg. N: Shiva. S: Cannon. RO. Ill: Kelm M:
Wink. DA: Espev. MG: Oldfield. EH: Pluta. RM: Freeman. BA: Lancaster. JR: Feelisch. M: Lundberg. JO.
(2005). The emerging biology of the nitrite anion. Nat ChemBiol 1: 308-314.
http://dx.doi.org/10.1038/nchembioll05-308
Goggins. WB: Chan. EY: Yang. CY. (2013). Weather, pollution, and acute myocardial infarction in Hong Kong
and Taiwan. Int J Cardiol 168: 243-249. http://dx.doi.0rg/10.1016/i.iicard.2012.09.087
Goings. SAJ: Kulle. TJ: Bascom. R: Sauder. LR: Green. DJ: Hebel JR: Clements. ML. (1989). Effect of
nitrogen dioxide exposure on susceptibility to influenza A virus infection in healthy adults. Am Rev Respir
Dis 139: 1075-1081. http://dx.doi.0rg/10.1164/airccm/139.5.1075
R-34
-------
Goldberg. MS: Burnett. RT: Stieb. DM: Brophy. JM: Daskalopoulou. SS: Valois. MF: Brook. JR. (2013).
Associations between ambient air pollution and daily mortality among elderly persons in Montreal, Quebec.
Sci Total Environ 463-464: 931942. http://dx.doi.0rg/10.1016/i.scitotenv.2013.06.095
Goldberg. MS: Giannetti. N: Burnett. RT: Mayo. NE: Valois. MF: Brophy. JM. (2008). A panel study in
congestive heart failure to estimate the short-term effects from personal factors and environmental conditions
on oxygen saturation and pulse rate. Occup Environ Med 65: 659-666.
http://dx.doi.org/10.1136/oem.2007.034934
Goldman. GT. (2013). Email from Gretchen Tanner Goldman to Dr Richmond-Bryant; Response to Data
Request. Available online at http://www.regulations.gov/#!documentDetail:D=EPA-HQ-ORD-2013-0232-
0019
Goldman. GT: Mulholland. JA: Russell AG: Gass. K: Strickland. MJ: Tolbert PE. (2012). Characterization of
ambient air pollution measurement error in a time-series health study using a geostatistical simulation
approach. Atmos Environ 57: 101-108. http://dx.doi.0rg/10.1016/i.atmosenv.2012.04.045
Goldman. GT: Mulholland. JA: Russell AG: Srivastava. A: Strickland. MJ: Klein. M: Waller. LA: Tolbert. PE:
Edgerton. ES. (2010). Ambient air pollutant measurement error: characterization and impacts in a time-series
epidemiologic study in Atlanta. Environ Sci Technol 44: 7692-7698. http://dx.doi.org/10.1021/esl01386r
Goldmaa GT: Mulholland. JA: Russell. AG: Strickland. MJ: Kleia M: Waller. LA: Tolbert. PE. (2011). Impact
of exposure measurement error in air pollution epidemiology: Effect of error type in time-series studies.
Environ Health 10: 61. http://dx.doi.org/10.1186/1476-069X-10-61
Goldstein. BD: Hamburger. SJ: Falk. GW: Amoruso. MA. (1977). Effect of ozone and nitrogen dioxide on the
agglutination of rat alveolar macrophages by concanavalin A. Life Sci 21: 1637-1644.
http://dx.doi.org/10.1016/0024-3205(77)90242-9
Goldstein. E: Eagle. MC: Hoeprich. PP. (1973). Effect of nitrogen dioxide on pulmonary bacterial defense
mechanisms. Arch Environ Health 26: 202-204. http://dx.doi.org/10.1080/00039896.1973.10666256
Goldstein. E: Warshauer. D: Lippert W: Tarkington. B. (1974). Ozone and nitrogen dioxide exposure: Murine
pulmonary defense mechanisms. Arch Environ Health 28: 85-90.
http://dx.doi.org/10.1080/00039896.1974.10666444
Goldstein. S: Lind. J: Merenyi. G. (2004). Reaction of organic peroxyl radicals with NO 2 and NO in aqueous
solution: Intermediacy of organic peroxynitrate and peroxynitrite species. J Phys Chem A 108: 1719-1725.
http://dx.doi.org/10.1021/ipQ37431z
Gong. H: Linn. WS: Clark. KW: Anderson. KR: Geller. MD: Sioutas. C. (2005). Respiratory responses to
exposures with fine particulates and nitrogen dioxide in the elderly with and without COPD. Inhal Toxicol
17: 123-132. http://dx.doi.org/10.1080/08958370590904481
Gonzales. M: Myers. O: Smith. L: Olvera. HA: Mukerjee. S: Li. WW: Pingitore. N: Amava. M: Burchiel S:
Berwick. M. (2012). Evaluation of land use regression models for NO2 in El Paso, Texas, USA. Sci Total
Environ432: 135-142. http://dx.doi.0rg/10.1016/i.scitotenv.2012.05.062
Gonzales. M: Quails. C: Hudgens. E: Neas. L. (2005). Characterization of a spatial gradient of nitrogen dioxide
across a United States-Mexico border city during winter. Sci Total Environ 337: 163-173.
http://dx.doi.0rg/10.1016/i.scitotenv.2004.07.010
Goodman. JE: Chandalia. JK: Thakali. S: Seelev. M. (2009). Meta-analysis of nitrogen dioxide exposure and
airway hyper-responsiveness in asthmatics. Crit Rev Toxicol 39: 719-742.
http://dx.doi.org/10.3109/10408440903283641
Gordon. M: Staebler. RM: Liggio. J: Li. SM: Wentzell. J: Lu. G: Lee. P: Brook. JR. (2012). Measured and
modeled variation in pollutant concentration near roadways. Atmos Environ 57: 138-145.
http://dx.doi.0rg/10.1016/i.atmosenv.2012.04.022
R-35
-------
Gotschi. T: Sunyer. J: Chinn. S: de Marco. R: Forsberg. B: Gauderman. JW: Garcia-Esteban. R: Heinrich. J:
Jacquemin. B: Jarvis. D: Ponzio. M: Villanj S: Kiinzlj N. (2008). Air pollution and lung function in the
European Community Respiratory Health Survey. Int J Epidemiol 37: 1349-1358.
http://dx.doi.org/10.1093/ije/dvnl36
Gouveia. N: Bremner. SA; Novaes. HMD. (2004). Association between ambient air pollution and birth weight in
Sao Paulo, Brazil. J Epidemiol Community Health 58: 11-17. http://dx.doi.0rg/10.1136/iech.58.l.ll
Powers. AM: Cullinan. P; Ayres. JG: Anderson. HR: Strachan. DP: Holgate. ST: Mills. 1C: Maynard. RL.
(2012). Does outdoor air pollution induce new cases of asthma? Biological plausibility and evidence; a
review [Review]. Respirology 17: 887-898. http://dx.doi.0rg/10.llll/i.1440-1843.2012.02195.x
Graham. JA: Gardner. DE: Blommer. EJ: House. DE: Menache. MG: Miller. FJ. (1987). Influence of exposure
patterns of nitrogen dioxide and modifications by ozone on susceptibility to bacterial infectious disease in
mice. JToxicol Environ Health 21: 113-125. http://dx.doi.org/10.1080/15287398709531006
Gray. EL: Patton. FM: Goldberg. SB: Kaplan. E. (1954). Toxicity of the oxides of nitrogen. II. Acute inhalation
toxicity of nitrogen dioxide, red fuming nitric acid, and white fuming nitric acid. AMA Arch Ind Health 10:
418-422.
Green. RS: Smorodinsky. S: Kim. JJ: McLaughlin. R: Ostro. B. (2004). Proximity of California public schools to
busy roads. Environ Health Perspect 112: 61-66. http://dx.doi.org/10.1289/ehp.6566
Greene. ND: Schneider. SL. (1978). Effects of NO2 on the response of baboon alveolar macrophages to
migration inhibitory factor. J Toxicol Environ Health 4: 869-880.
http://dx.doi.org/10.1080/15287397809529708
Greenwald. R: Sarnat SE: Ravsonj AU: Li. WW: Johnson. BA: Stock. TH: Holguin. F: Sosa. T: Sarnat JA.
(2013). Associations between source-indicative pollution metrics and increases in pulmonary inflammation
and reduced lung function in a panel of asthmatic children. Air Qual Atmos Health 6: 487-499.
http://dx.doi.org/10.1007/sll869-012-0186-3
Gregory. RE: Pickrell JA: Halm. FF: Hobbs. CH. (1983). Pulmonary effects of intermittent subacute exposure to
low-level nitrogen dioxide. J Toxicol Environ Health 11: 405-414.
http://dx.doi.org/10.1080/15287398309530354
Griffiths. PT: Badger. CL: Cox. RA: Folkers. M: Henk. HH: Mentel TF. (2009). Reactive uptake of N2O5 by
aerosols containing dicarboxylic acids. Effect of particle phase, composition, and nitrate content. J Phys
Chem A 113: 5082-5090. http://dx.doi.org/10.1021/ip8096814
Grineski. SE: Staniswalis. JG: Peng. Y: Atkinson-Palombo. C. (2010). Children's asthma hospitalizations and
relative risk due to nitrogen dioxide (NO2): Effect modification by race, ethnicity, and insurance status.
Environ Res 110: 178-188. http://dx.doi.0rg/10.1016/i.envres.2009.10.012
Griscavage. JM: Hobbs. AJ: Ignarro. LJ. (1995). Negative modulation of nitric oxide synthase by nitric oxide
and nitroso compounds [Review]. InLJ Ignarro; F Murad (Eds.), Nitric oxide: Biochemistry, molecular
biology, and therapeutic implications (pp. 215-234). San Diego, CA: Academic Press.
http://dx.doi.org/10.1016/S1054-3589(08)61088-l
Gronskei KE; Walker. SE; Gram. F. (1993). Evaluation of a model for hourly spatial concentration distributions.
Atmos Environ 27: 105-120. http://dx.doi.org/10.1016/0957-1272(93)90050-G
Gruzieva. O; Bellander. T; Eneroth. K; Kull. I; Melen. E; Nordling. E; van Hage. M; Wickman. M; Moskalenko.
V; Hulchiy. O; Pershagen. G. (2012). Traffic-related air pollution and development of allergic sensitization
in children during the first 8 years of life. J Allergy Clin Immunol 129: 240-246.
http://dx.doi.0rg/10.1016/i.iaci.2011.ll.001
Gruzieva. O; Bergstrom. A; Hulchiy. O; Kull I; Lind. T; Melen. E; Moskalenko. V; Pershagen. G; Bellander. T.
(2013). Exposure to air pollution from traffic and childhood asthma until 12 years of age. Epidemiology 24:
54-61. http://dx.doi.org/10.1097/EDE.Ob013e318276clea
R-36
-------
Gryparis. A: Coull BA: Schwartz. J. (2007). Controlling for confounding in the presence of measurement error
in hierarchical models: A Bayesian approach. JExpo Sci Environ Epidemiol 17: S20-S28.
http://dx.doi.org/10.1038/si.ies.7500624
Gu. Z: Zhang. Y: Lei. K. (2010). Large eddy simulation of flow in a street canyon with tree planting under
various atmospheric instability conditions. Sci China Tech Sci 53: 1928-1937.
http://dx.doi.org/10.1007/sll431-010-3243-x
Guo. Y: Jia. Y: Pan, X: Liu. L: Wichmann. HE. (2009). The association between fine paniculate air pollution
and hospital emergency room visits for cardiovascular diseases in Beijing, China. Sci Total Environ 407:
4826-4830. http://dx.doi.0rg/10.1016/i.scitotenv.2009.05.022
Guo. Y: Tong. S: Li. S: Barnett AG: Yu. W: Zhang. Y: Pan, X. (2010). Gaseous air pollution and emergency
hospital visits for hypertension in Beijing, China: a time-stratified case-crossover study. Environ Health 9:
57. http://dx.doi.org/10.1186/1476-069X-9-57
Gushchin. IS: Petvaev. IM: Tsinkalovsky. OR. (1990). Kinetics of oxygen metabolism indices in the course of
histamine secretion from rat mast cells. Agents and Actions 30: 85-88.
http://dx.doi.org/10.1007/BF01969005
Guxens. M: Aguilera. I: Ballester. F: Estarlich. M: Fernandez-Somoano. A: Lertxundi. A: Lertxundi N: Mendez.
MA: Tardon. A: Vrijheid. M: Sunyer. J. (2012). Prenatal exposure to residential air pollution and infant
mental development: Modulation by antioxidants and detoxification factors. Environ Health Perspect 120:
144-149. http://dx.doi.org/10.1289/ehp. 1103469
Guxens. M: Garcia-Esteban. R: Giorgis-Allemand. L: Forns. J: Badaloni. C: Ballester. F: Beelen. R: Cesaroni.
G: Chatzi. L: de Agostini. M: de Nazelle. A: Eeftens. M: Fernandez. MF: Fernandez-Somoano. A: Forastiere.
F: Gehring. U: Ghassabian. A: Heude. B: Jaddoe. VW: Kliimper. C: Kogevinas. M: Kramer. U: Larroque. B:
Lertxundi. A: Lertxuni. N: Murcia. M: Navel V: Nieuwenhuijsen. M: Porta. D: Ramos. R: Roumeliotaki. T:
Slama. R: Serensen. M: Stephanou. EG: Sugiri. D: Tardon. A: Tiemeier. H: Tiesler. CM: Verhulst FC:
Vrijkotte. T: Wilhelm. M: Brunekreef B: Pershagen. G: Sunyer. J. (2014). Air pollution during pregnancy
and childhood cognitive and psychomotor development: Six European birth cohorts. Epidemiology 25: 636-
647. http://dx.doi.org/10.1097/EDE.0000000000000133
Hack. M: Fanaroff AA. (1999). Outcomes of children of extremely low birth weight and gestational age in the
1990s [Review]. Early HumDev 53: 193-218. http://dx.doi.org/10.1016/80378-3782(98)00052-8
Hackney. JD: Thiede. FC: Linn. WS: Pedersen. EE: Spier. CE: Law. DC: Fischer. DA. (1978). Experimental
studies on human health effects of air pollutants. IV. Short-term physiological and clinical effects of nitrogen
dioxide exposure. Arch Environ Health 33: 176-181. http://dx.doi.org/10.1080/00039896.1978.10667330
Hagenbiork-Gustafsson. A: Eriksson. K: Forsberg. B. (2009). Field evaluation of the Ogawa Diffusive Sampler
for NO(2)/NO(x) in a cold climate [Abstract]. Epidemiology 20: S161.
http://dx.doi.org/10.1097/01.ede.0000362548.83420.d7
Hagler. GSW: Tang. W: Freeman. MJ: Heist DK: Perry. SG: Vette. AF. (2011). Model evaluation of roadside
barrier impact on near-road air pollution. Atmos Environ 45: 2522-2530.
http://dx.doi.0rg/10.1016/i.atmosenv.2011.02.030
Halonen. JI: Lanki. T: Yli-Tuomi. T: Tiittanen. P: Kulmala. M: Pekkanen. (2009). Paniculate air pollution and
acute cardiorespiratory hospital admissions and mortality among the elderly. Epidemiology 20: 143-153.
http://dx.doi.org/10.1097/EDE.Ob013e31818c7237
Hameed. S: Pinto. JP: Stewart. RW. (1979). Sensitivity of the predicted CO-OH-CH4 perturbation to
tropospheric NOx concentrations. J Geophys Res Oceans 84: 763-768.
http://dx.doi.org/10.1029/JC084iC02pQ0763
Hamid. 0: Springall. PR: Riveros-Moreno. V: Chanez. P: Howarth. P: Redington. A: Bousquet. J: Godard. P:
Holgate. S: Polak. JM. (1993). Induction of nitric oxide synthase in asthma. Lancet 342: 1510-1513.
http://dx.doi.org/10.1016/S0140-6736(05)80083-2
R-37
-------
Hampel R: Lepeule. J: Schneider. A: Bottagisi S: Charles. MA: Ducimetiere. P: Peters. A: Slama. R. (2011).
Short-term impact of ambient air pollution and air temperature on blood pressure among pregnant women.
Epidemiology 22: 671-679. http://dx.doi.org/10.1097/EDE.Ob013e318226e8d6
Han. M: Guo. Z: Li. G: Sang. N. (2013). Nitrogen dioxide inhalation induces genotoxicity in rats. Chemosphere
90: 2737-2742. http://dx.doi.0rg/10.1016/i.chemosphere.2012.ll.057
Hannam. K: McNamee. R: Baker. P: Siblev. C: Agius. R. (2014). Air pollution exposure and adverse pregnancy
outcomes in a large UK birth cohort: use of a novel spatio-temporal modelling technique. Scand J Work
Environ Health 40: 518-530. http://dx.doi.org/10.5271/sjweh.3423
Hansel. NN: Brevsse. PN: McCormack. MC: Matsui EC: Curtin-Brosnan. J: Williams. DAL: Moore. JL:
Cuhran. JL: Diette. GB. (2008). A longitudinal study of indoor nitrogen dioxide levels and respiratory
symptoms in inner-city children with asthma. Environ Health Perspect 116: 1428-1432.
http://dx.doi.org/10.1289/ehp.11349
Hansel NN: McCormack. MC: Belli. AJ: Matsui. EC: Peng. RD: Aloe. C: Paulin. L: Williams. PL: Diette. GB:
Brevsse. PN. (2013). In-home air pollution is linked to respiratory morbidity in former smokers with chronic
obstructive pulmonary disease. Am J Respir Crit Care Med 187: 1085-1090.
http://dx.doi.org/10.1164/rccm.201211-1987OC
Hansen. C: Neller. A: Williams. G: Simpson. R. (2006). Maternal exposure to low levels of ambient air pollution
and pretermbirth in Brisbane, Australia. BJOG 113: 935-941. http://dx.doi.org/10.1111/i. 1471-
0528.2006.01010.x
Hansen. C: Neller. A: Williams. G: Simpson. R. (2007). Low levels of ambient air pollution during pregnancy
and fetal growth among term neonates in Brisbane, Australia. Environ Res 103: 383-389.
http://dx.doi.0rg/10.1016/i.envres.2006.06.010
Hansen. CA: Barnett AG: Pritchard. G. (2008). The effect of ambient air pollution during early pregnancy on
fetal ultrasonic measurements during mid-pregnancy. Environ Health Perspect 116: 362-369.
http://dx.doi.org/10.1289/ehp. 10720
Hansen. DA: Edgerton. ES: Hartsell BE: Jansen. JJ: Kandasamy. N: Hidv. GM: Blanchard. CL. (2003). The
southeastern aerorsol research and characterization study: Part I - overview. J Air Waste Manag Assoc 53:
1460-1471. http://dx.doi.org/10.1080/10473289.2003.10466318
Harre. ESM: Price. PD: Ayrev. RB: Toop. LJ: Martin. IR: Town. GI. (1997). Respiratory effects of air pollution
in chronic obstructive pulmonary disease: A three month prospective study. Thorax 52: 1040-1044.
http://dx.doi.org/10.1136/thx.52.12.1040
Harrison. MAJ: Barra. S: Borghesi D: Vione. D: Arsene. C: Olariu. RI. (2005). Nitrated phenols in the
atmosphere: a review [Review]. Atmos Environ 39: 231-248.
http://dx.doi.0rg/10.1016/i.atmosenv.2004.09.044
Hart. J: Yanosky. J: Puett R: Ryan. L: Dockery. D: Smith. T: Garshick. E: Laden. F. (2009). Spatial modeling of
PM10 and NO2 in the continental United States, 1985-2000. Environ Health Perspect 117: 1690-1696.
http://dx.doi.org/10.1289/ehp.0900840
Hart. JE: Garshick. E: Dockery. DW: Smith. TJ: Ryan. L: Laden. F. (2011). Long-term ambient multi-pollutant
exposures and mortality. Am J Respir Crit Care Med 183: 73-78. http://dx.doi.org/10.1164/rccm.200912-
1903OC
Hart. JE: Rimm. EB: Rexrode. KM: Laden. F. (2013). Changes in traffic exposure and the risk of incident
myocardial infarction and all-cause mortality. Epidemiology 24: 734-742.
http://dx.doi.org/10.1097/EDE.Ob013e31829d5dae
Hasunuma. H: Ishimaru. Y: Yoda. Y: Shima. M. (2014). Decline of ambient air pollution levels due to measures
to control automobile emissions and effects on the prevalence of respiratory and allergic disorders among
children in Japan. Environ Res 131: 111-118. http://dx.doi.0rg/10.1016/i.envres.2014.03.007
Hatch. GE. (1992). Comparative biochemistry of airway lining fluid. In RA Parent (Ed.), Comparative biology
of the normal lung, Vol 1: Treatise on pulmonary toxicology (pp. 617-632). Boca Raton, FL: CRC Press, Inc.
R-38
-------
Hatch. GE: Slade. R: Selgrade. MK: Stead. AG. (1986). Nitrogen dioxide exposure and lung antioxidants in
ascorbic acid-deficient guinea pigs. Toxicol Appl Pharmacol 82: 351-359. http://dx.doi.org/10.1016/0041-
008X(86)90212-7
Hatzopoulou. M: Weichenthal S: Barreau. G: Goldberg. M: Farrell W: Grouse. D: Ross. N. (2013). A web-
based route planning tool to reduce cyclists' exposures to traffic pollution: A case study in Montreal, Canada.
Environ Res 123: 58-61. http://dx.doi.0rg/10.1016/i.envres.2013.03.004
Havashj Y: Kohno. T: Ohwada. H. (1987). Morphological effects of nitrogen dioxide on the rat lung. Environ
Health Perspect 73: 135-145.
Hazucha. MJ: Folinsbee. LJ: Seal. E: Bromberg. PA. (1994). Lung function response of healthy women after
sequential exposures to NO2 and O3. Am J Respir Crit Care Med 150: 642-647.
http://dx.doi.0rg/10.1164/airccm.150.3.8087332
Hazucha. MJ: Ginsberg. JF: McDonnell. WF: Haak. ED. Jr: Pimmel. RL: Salaam. SA: House. DE: Bromberg.
PA. (1983). Effects of 0.1 ppm nitrogen dioxide on airways of normal and asthmatic subjects. J Appl Physiol
Respir Environ Exerc Physiol 54: 730-739.
He. M: Dhaniyala. S. (2011). A dispersion model for traffic produced turbulence in a two-way traffic scenario.
Environ Fluid Mech 11: 627-640. http://dx.doi.org/10.1007/sl0652-011-9215-2
Heal. MR. (2008). The effect of absorbent grid preparation method on precision and accuracy of ambient
nitrogen dioxide measurements using Palmes passive diffusion tubes. J Environ Monit 10: 1363-1369.
http://dx.doi.org/10.1039/b811230d
Heckbert SR: Kooperberg. C: Safford. MM: Psatv. BM: Hsia. J: McTiernan. A: Gaziano. JM: Frishman. WH:
Curb. JD. (2004). Comparison of serf-report, hospital discharge codes, and adjudication of cardiovascular
events in the Women's Health Initiative. Am J Epidemiol 160: 1152-1158.
http://dx.doi.org/10.1093/aie/kwh314
HEI (Health Effects Institute). (2003). Revised analyses of time-series studies of air pollution and health [HEI].
(HEI Special Report). Boston, MA: Health Effects Institute (HEI).
http://pubs.healtheffects.org/view.php?id=4
HEI (Health Effects Institute). (2010). Traffic-related air pollution: A critical review of the literature on
emissions, exposure, and health effects. (Special Report 17). Boston, MA: Health Effects Institute (HEI).
http://pubs.healtheffects.org/view.php?id=334
HEI (Health Effects Institute). (2012). Effects of short-term exposure to air pollution on hospital admissions of
young children for acute lower respiratory infections in Ho Chi Minh City, Vietnam (pp. 5-72; discussion 73-
83). (Research Report 169). Boston, MA: Health Effects Institute, HEI Collaborative Working Group on Air
Pollutino, Poverty, and Health in Ho Chi Minh City. http://pubs.healtheffects.org/view.php?id=380
HEI Review Panel on Ultrafine Particles. (2013). Understanding the health effects of ambient ultrafme particles.
(HEI Perspectives 3). Boston, MA: Health Effects Institute. http://pubs.healtheffects.org/view.php?id=394
Heinrich. J: Thiering. E: Rzehak. P: Kramer. U: Hochadel. M: Rauchfuss. KM: Gehring. U: Wichmann. HE.
(2013). Long-term exposure to NO2 and PM10 and all-cause and cause-specific mortality in a prospective
cohort of women. Occup Environ Med 70: 179-186. http://dx.doi.org/10.1136/oemed-2012-100876
Heist DK: Brixev. LA: Richmond-Bryant J: Bowker. GE: Perry. SG: Wiener. RW. (2009a). The effect of a tall
tower on flow and dispersion through a model urban neighborhood: Part 1. Flow characteristics. J Environ
Monit 11: 2163-2170. http://dx.doi.org/10.1039/B907135k
Heist DK: Perry. SG: Brixev. LA. (2009b). A wind tunnel study of the effect of roadway configurations on the
dispersion of traffic-related pollution. Atmos Environ 43: 5101-5111.
http://dx.doi.0rg/10.1016/i.atmosenv.2009.06.034
Held. T: Chang. DPY: Niemeier. DA. (2003). UCD 2001: An improved model to simulate pollutant dispersion
from roadways. Atmos Environ 37: 5325-5336. http://dx.doi.0rg/10.1016/i.atmosenv.2003.09.015
R-39
-------
Helledav. R: Huberman. D: Blomberg. A: Stiernberg. N: Sandstrom. T. (1995). Nitrogen dioxide exposure
impairs the frequency of the mucociliary activity in healthy subjects. Eur Respir J 8: 1664-1668.
http://dx.doi.org/10.1183/09031936.95.08101664
Helledav. R: Sandstrom. T: Stiernberg. N. (1994). Differences in bronchoalveolar cell response to nitrogen
dioxide exposure between smokers and nonsmokers. Eur Respir J 7: 1213-1220.
http://dx.doi.org/10.1183/09031936.94.07071213
Heller. H: Korbmacher. N: Gabler. R: Brandt. S: Breitbach. T: Juergens. U: Grohe. C: Schuster. KD. (2004).
Pulmonary 15NO uptake in interstitial lung disease. Nitric Oxide 10: 229-232.
http://dx.doi.0rg/10.1016/i.niox.2004.06.001
Henderson. SB: Beckerman. B: Jerrett M: Brauer. M. (2007). Application of land use regression to estimate
long-term concentrations of traffic-related nitrogen oxides and fine paniculate matter. Environ Sci Technol
41: 2422-2428. http://dx.doi.org/10.1021/es0606780
Hendrick. EM: Tino. VR: Hanna. SR: Egan. BA. (2013). Evaluation of NO2 predictions by the plume volume
molar ratio method (PVMRM) and ozone limiting method (OLM) in AERMOD using new field
observations. J Air Waste Manag Assoc 63: 844-854. http://dx.doi.org/10.1080/10962247.2013.798599
Henneberger. A: Zareba. W: Ibald-Mulli. A: Riickerl R: Cyrys. J: Couderc. JP: Mykins. B: Woelke. G:
Wichmann. HE: Peters. A. (2005). Repolarization changes induced by air pollution in ischemic heart disease
patients. Environ Health Perspect 113: 440-446. http://dx.doi.org/10.1289/ehp.7579
Henry. MC: Findlav. J: Spangler. J: Ehrlich. R. (1970). Chronic toxicity of NO2 in squirrel monkeys: III. Effect
on resistance to bacterial and viral infection. Arch Environ Health 20: 566-570.
http://dx.doi.org/10.1080/00039896.1970.10665663
Hernandez-Cadena. L: Holguin. F: Barraza-Villarreal A: Del Rio-Navarro. BE: Sienra-Monge. JJ: Romieu. I.
(2009). Increased levels of outdoor air pollutants are associated with reduced bronchodilation in children
with asthma. Chest 136: 1529-1536. http://dx.doi.org/10.1378/chest.08-1463
Hertel O: Jensen. SS: Ketzel. M: Becker. T: Peel RG: Orbv. PV: Skjoth. CA: Ellermann. T: Raaschou-Nielsen.
O: Sorensen. M: Brauner. EV: Andersen. ZJ: Loft S: Schlunssen. V: Bonlokke. JH: Sigsgaard. T. (2014).
Utilizing monitoring data and spatial analysis tools for exposure assessment of atmospheric pollutants in
Denmark [Chapter 6]. In LL McConnell; J Dachs; CJ Hapeman (Eds.), Occurrence, fate and impact fo
atmospheric pollutants on environmental and human health (pp. 95-122). Washington, D.C.: American
Chemical Society, http://pubs.acs.org/doi/abs/10.1021/bk-2013-1149.ch006
Hesterberg. TW: Bunn. WB: McClellan. RO: Hamade. AK: Long. CM: Valberg. PA. (2009). Critical review of
the human data on short-term nitrogen dioxide (NO2) exposures: Evidence for NO2 no-effect levels
[Review]. Crit Rev Toxicol 39: 743-781. http://dx.doi.org/10.3109/10408440903294945
HEW (U.S. Department of Health, Education and Welfare). (1964). Smoking and health: Report of the advisory
committee to the surgeon general of the public health service. Washington, DC: U.S. Department of Health,
Education, and Welfare. http://profiles.nlm.nih.gov/ps/retrieve/ResourceMetadata/NNBBMQ
Higdon. A: Piers. AR: Oh. J: Landar. A: Parley-Usmar. VM. (2012). Cell signalling by reactive lipid species:
New concepts and molecular mechanisms [Review]. Biochem J 442: 453-464.
http://dx.doi.org/10.1042/BJ20111752
Higgins. BG: Francis. HC: Yates. C: Warburton. CJ: Fletcher. AM: Pickering. CAC: Woodcock. AA. (2000).
Environmental exposure to air pollution and allergens and peak flow changes. Eur Respir J 16: 66.
Higgins. BG: Francis. HC: Yates. CJ: Warburton. CJ: Fletcher. AM: Reid. JA: Pickering. CAC: Woodcock. AA.
(1995). Effects of air pollution on symptoms and peak expiratory flow measurements in subjects with
obstructive airways disease. Thorax 50: 149-155. http://dx.doi.0rg/10.1136/thx.50.2.149
Hildebrandt K: Riickerl R: Koenig. W: Schneider. A: Pitz. M: Heinrich. J: Marder. V: Frampton. M:
Oberdorster. G: Wichmann. HE: Peters. A. (2009). Short-term effects of air pollution: A panel study of blood
markers in patients with chronic pulmonary disease. Part Fibre Toxicol 6: 25. http://dx.doi.org/10.1186/1743-
8977-6-25
R-40
-------
Hill AB. (1965). The environment and disease: Association or causation? Proc R Soc Med 58: 295-300.
Hill. BG: Dranka. BP: Bailey. SM: Lancaster. JR: Darlev-Usmar. VM. (2010). What part of NO don't you
understand? Some answers to the cardinal questions in nitric oxide biology [Review]. J Biol Chem 285:
19699-19704. http://dx.doi.org/10.1074/ibc.R110.101618
Hiltermann. TJN: Stolk. J: van der Zee. SC: Brunekreef. B: de Bruijne. CR: Fischer. PH: Ameling. CB: Sterk.
PJ: Hiemstra. PS: vanBree. L. (1998). Asthma severity and susceptibility to air pollution. Eur Respir J 11:
686-693.
Hinwood. AL: De Klerk. N: Rodriguez. C: Jacobv. P: Runnion. T: Rye. P: Landau. L: Murray. F: Feldwick. M:
Spickett. J. (2006). The relationship between changes in daily air pollution and hospitalizations in Perth,
Australia 1992-1998: A case-crossover study. Int J Environ Health Res 16: 27-46.
http://dx.doi.org/10.1080/09603120500397680
Hizawa. N. (2011). Pharmacogenetics of (32-agonists [Review]. Allergol Int 60: 239-246.
http://dx.doi.org/10.2332/allergolint.ll-RAI-0317
Hodgkins. SR: Ather. JL: Paveglio. SA: Allard. JL: LeClair. LA: Suratt BT: Boyson. JE: Poynter. ME. (2010).
NO2 inhalation induces maturation of pulmonary GDI lc+ cells that promote antigenspecific CD4+ T cell
polarization. Respir Res 11: 102. http://dx.doi.org/10.1186/1465-9921-ll-102
Hoek. G. (2003). Daily mortality and air pollution in The Netherlands [HEI]. In Revised analyses of time-series
studies of air pollution and health (pp. 133-141). Boston, MA: Health Effects Institute.
http://pubs.healtheffects.org/getfile.php?u=21
Hoek. G: Beelen. R: de Hoogh. K: Vienneau. D: Gulliver. J: Fischer. P: Briggs. D. (2008). A review of land-use
regression models to assess spatial variation of outdoor air pollution [Review]. Atmos Environ 42: 7561-
7578. http://dx.doi.0rg/10.1016/i.atmosenv.2008.05.057
Hoek. G: Brunekreef. B: Goldbohm. S: Fischer. P: van den Brandt. PA. (2002). Association between mortality
and indicators of traffic-related air pollution in the Netherlands: A cohort study. Lancet 360: 1203-1209.
http://dx.doi.org/10.1016/sO 140-6736(02) 11280-3
Hogman. M: Frostell. C: Arnberg. H: Hedenstierna. G. (1993). Inhalation of nitric oxide modulates
methacholine-induced bronchoconstriction in the rabbit. Eur Respir J 6: 177-180.
Hohlfeld. JM. (2002). The role of surfactant in asthma. Respir Res 3: 4. http://dx.doi.org/10.1186/rrl76
Holguin. F: Flores. S: Ross. Z: Cortez. M: Molina. M: Molina. L: Rincon. C: Jerrett. M: Berhane. K: Granados.
A: Romieu. I. (2007). Traffic-related exposures, airway function, inflammation, and respiratory symptoms in
children. Am J Respir Crit Care Med 176: 1236-1242. http://dx.doi.org/10.1164/rccm.200611-1616OC
Hong. JH: Lee. JH: Park. SS. (2007). Application of an in-situ measurement system to determine HONO levels
in an indoor environment. Journal of Korean Society for Atmospheric Environment 23: 191-202.
http://dx.doi.Org/10.5572/KOSAE.2007.23.2.191
Hooftman. RN: Kuper. CF: Appelman. LM. (1988). Comparative sensitivity of histo-pathology and specific lung
parameters in the detection of lung injury. J Appl Toxicol 8: 59-65. http://dx.doi.org/10.1002/iat.2550Q80110
Horowitz. LW: Fiore. AM: Milly. GP: Cohen. RC: Perring. A: Wooldridge. PJ: Hess. PG: Emmons. LK:
Lamarque. JF. (2007). Observational constraints on the chemistry of isoprene nitrates over the eastern United
States. J Geophys Res Atmos 112: D12S08. http://dx.doi.org/10.1029/2006JD007747
Hou. HY: Wang. D: Zou. XP: Yang. ZH: Li. TC: Chen. YQ. (2014). Does ambient air pollutants increase the
risk of fetal loss? A case-control study. Arch Gynecol Obstet 289: 285-291.
http://dx.doi.org/10.1007/s00404-013-2962-l
Houston. D: Ong. P: Wu. J: Winer. A. (2006). Proximity of licensed child care facilities to near-roadway vehicle
pollution. Am J Public Health 96: 1611-1617. http://dx.doi.org/10.2105/AJPH.2005.077727
Howden. LM: Meyer. JA. (2011). Age and sex composition: 2010. (2010 Census Briefs, C2010BR-03).
Washington, DC: U.S. Department of Commerce, Economics and Statistics Administration, U.S. Census
Bureau, http://www.census.gov/prod/cen2010/briefs/c2010br-03.pdf
R-41
-------
Hovert PL: Xu. J. (2012). Deaths: Preliminary data for 2011. National Vital Statistics Reports 61: 1-51.
Hsieh. YL: Yang. YH: Wu. TN: Yang. CY. (2010). Air pollution and hospital admissions for myocardial
infarction in a subtropical city: Taipei, Taiwan. J Toxicol Environ Health A 73: 757-765.
http://dx.doi.org/10.1080/15287391003684789
Hu. S: Fruin. S: Kozawa. K: Mara. S: Paulson. SE: Winer. AM. (2009). A wide area of air pollutant impact
downwind of a freeway during pre-sunrise hours. Atmos Environ 43: 2541-2549.
http://dx.doi.0rg/10.1016/i.atmosenv.2009.02.033
Huang. W: Zhu. T: Pan. X: Hu. M: Lu. SE: Lin. Y: Wang. T: Zhang. Y: Tang. X. (2012a). Air pollution and
autonomic and vascular dysfunction in patients with cardiovascular disease: Interactions of systemic
inflammation, overweight, and gender. Am J Epidemiol 176: 117-126. http://dx.doi.org/10.1093/aie/kwr511
Huang. YC: Rappold. AG: Graff. DW: Ohio. AJ: Devlin. RB. (2012b). Synergistic effects of exposure to
concentrated ambient fine pollution particles and nitrogen dioxide in humans. Inhal Toxicol 24: 790-797.
http://dx.doi.org/10.3109/08958378.2012.718809
Hubbard. AK: Symanowicz. PT: Thibodeau. M: Thrall RS: Schramm. CM: Cloutier. MM: Morris. JB. (2002).
Effect of nitrogen dioxide on ovalbumin-induced allergic airway disease in a murine model. J Toxicol
Environ Health A 65: 1999-2005. http://dx.doi.org/10.1080/00984100290071397
Hudman. RC: Russell AR: Valin. LC: Cohen. RC. (2010). Interannual variability in soil nitric oxide emissions
over the United States as viewed from space. Atmos Chem Phys 10: 9943-9952.
http://dx.doi.org/10.5194/acp-10-9943-2010
Huie. RE. (1994). The reaction kinetics of NO2 [Review]. Toxicology 89: 193-216.
http://dx.doi.org/10.1016/0300-483X(94)90098-1
Humes. KR: Jones. NA: Ramirez. RR. (2011). Overview of race and Hispanic origin: 2010. (2010 Census Briefs,
C2010BR-02). Washington, DC: U.S. Department of Commerce, Economics and Statistics Administration,
U.S. Census Bureau, http://www.census.gov/prod/cen2010/briefs/c2010br-02.pdf
Husar. RB: Patterson. DE: Husar. JD: Gillanj NV: Wilson. WE. Jr. (1978). Sulfur budget of a power plant
plume. Atmos Environ 12: 549-568. http://dx.doi.org/10.1016/0004-6981(78)90236-6
Hutcheon. JA: Platt RW. (2008). The missing data problem in birth weight percentiles and thresholds for
"small-for-gestational-age". Am J Epidemiol 167: 786-792. http://dx.doi.org/10.1093/aie/kwm327
Hwang. BF: Jaakkola. JJK: Lee. YL: Lin. YC: Guo. YLL. (2006). Relation between air pollution and allergic
rhinitis in Taiwanese schoolchildren. Respir Res 7: 23. http://dx.doi.org/10.1186/1465-9921-7-23
Hwang. BF: Lee. YL. (2010). Air pollution and prevalence of bronchitic symptoms among children in Taiwan.
Chest 138: 956-964. http://dx.doi.org/10.1378/chest.09-2600
Hwang. BF: Lee. YL: Jaakkola. JJ. (2011). Air pollution and stillbirth: A population-based case-control study in
Taiwan. Environ Health Perspect 119: 1345-1349. http://dx.doi.org/10.1289/ehp. 1003056
Hwang. BF: Lee. YL: Lin. YC: Jaakkola. JJK: Guo. YL. (2005). Traffic related air pollution as a determinant of
asthma among Taiwanese school children. Thorax 60: 467-473. http://dx.doi.org/10.1136/thx.2004.033977
Hystad. P: Demers. PA: Johnson. KC: Carpiano. RM: Brauer. M. (2013). Long-term residential exposure to air
pollution and lung cancer risk. Epidemiology 24: 762-772.
http://dx.doi.org/10.1097/EDE.Ob013e3182949ae7
Hystad. P: Setton. E: Cervantes. A: Poplawskj K: Deschenes. S: Brauer. M: van Donkelaar. A: Lamsal L:
Martin. R: Jerrett M: Demers. P. (2011). Creating national air pollution models for population exposure
assessment in Canada. Environ Health Perspect 119: 1123-1129. http://dx.doi.org/10.1289/ehp. 1002976
IARC (International Agency for Research on Cancer). (2006). IARC monographs on the evaluation of
carcinogenic risks to humans: Preamble. Lyon, France: World Health Organization.
http://monographs.iarc.fr/ENG/Preamble/
R-42
-------
Ichinose. T: Arakawa. K: Shimojo. N: Sagai M. (1988). Biochemical effects of combined gases of nitrogen
dioxide and ozone: II. Species differences in lipid peroxides and antioxidative protective enzymes in the
lungs. ToxicolLett42: 167-176.
Ichinose. T: Fujii. K: Sagai. M. (1991). Experimental studies on tumor promotion by nitrogen dioxide.
Toxicology 67: 211-225. http://dx.doi.org/10.1016/0300-483X(91)90144-P
Ichinose. T: Sagai. M. (1989). Biochemical effects of combined gases of nitrogen dioxide and ozone III
Synergistic effects on lipid peroxidation and antioxidative protective systems in the lungs of rats and guinea
pigs. Toxicology 59: 259-270. http://dx.doi.org/10.1016/0300-483X(89)90196-0
Ichinose. T: Sagai. M. (1992). Combined exposure to NO2, O3 and H2SO4-aerosol and lung tumor formation in
rats. Toxicology 74: 173-184. http://dx.doi.org/10.1016/0300-483X(92)90137-4
Illing. JW: Miller. FJ: Gardner. DE. (1980). Decreased resistance to infection in exercised mice exposed to NO2
and O3. J Toxicol Environ Health 6: 843-851. http://dx.doi.org/10.1080/15287398009529902
Iniguez. C: Ballester. F: Amoros. R: Estarlich. M: Murcia. M: Esplugues. A: Rebagliato. M. (2009). The impact
of outdoor NO2 exposure on fetal growth assessed by ultrasounds during pregnancy [Abstract].
Epidemiology 20: S78. http://dx.doi.org/10.1097/01.ede.0000362944.07976.ba
Iniguez. C: Ballester. F: Estarlich. M: Esplugues. A: Murcia. M: Llop. S: Plana. A: Amoros. R: Rebagliato. M.
(2012). Prenatal exposure to traffic-related air pollution and fetal growth in a cohort of pregnant women.
Occup Environ Med 69: 736-744. http://dx.doi.org/10.1136/oemed-2011-100550
Iniguez. C: Ballester. F: Estarlich. M: Llop. S: Fernandez-Patier. R: Aguirre-Alfaro. A: Esplugues. A. (2009).
Estimation of personal NO2 exposure in a cohort of pregnant women. Sci Total Environ 407: 6093-6099.
http://dx.doi.0rg/10.1016/i.scitotenv.2009.08.006
loannidis. JPA. (2008). Why most discovered true associations are inflated [Review]. Epidemiology 19: 640-
648. http://dx.doi.org/10.1097/EDE.Ob013e31818131e7
IOM (Institute of Medicine). (2008). Improving the presumptive disability decision-making process for veterans.
In JM Samet; CC Bodurow (Eds.). Washington, DC: National Academies Press.
http://dx.doi.org/10.17226/11908
IPCC (Intergovernmental Panel on Climate Change). (2007). Climate change 2007: Impacts, adaptation and
vulnerability. Cambridge, UK: Cambridge University Press.
https://www.ipcc.ch/publications and data/publications ipcc fourth assessment report wg2 report impact
s adaptation and vulnerabilitv.htm
Isaacs. K: McCurdv. T: Glen. G: Nysewander. M: Errickson. A: Forbes. S: Graham. S: McCurdv. L: Smith. L:
Tulve. N: Vallero. D. (2013). Statistical properties of longitudinal time-activity data for use inhuman
exposure modeling. J Expo Sci Environ Epidemiol 23: 328-336. http://dx.doi.org/10.1038/ies.2012.94
Isakov. V: Touma. JS: Burke. J: Lobdell DT: Palma. T: Rosenbaum. A: Ozkaynak. H. (2009). Combining
regional- and local-scale air quality models with exposure models for use in environmental health studies. J
Air Waste Manag Assoc 59: 461-472.
Iskandar. A: Andersen. ZJ: Bennelykke. K: Ellermann. T: Andersen. KK: Bisgaard. H. (2012). Coarse and fine
particles but not ultrafme particles in urban air trigger hospital admission for asthma in children. Thorax 67:
252-257. http://dx.doi.org/10.1136/thoraxinl-2011-200324
Ito. K: Mathes. R: Ross. Z: Nadas. A: Thurston. G: Matte. T. (2011). Fine paniculate matter constituents
associated with cardiovascular hospitalizations and mortality in New York City. Environ Health Perspect
119: 467-473. http://dx.doi.org/10.1289/ehp. 1002667
Ito. K: Thurston. GD: Silverman. RA. (2007). Characterization of PM2.5, gaseous pollutants, and meteorological
interactions in the context of time-series health effects models. J Expo Sci Environ Epidemiol 17: S45-S60.
http://dx.doi.org/10.1038/si.ies.7500627
Ivy. D: Mulholland. JA: Russell AG. (2008). Development of ambient air quality population-weighted metrics
for use in time-series health studies. J Air Waste Manag Assoc 58: 711-720. http://dx.doi.org/10.3155/1047-
3289.58.5.711
R-43
-------
Jackson. AC: Murphy. MM: Rassulo. J: Celli BR: Ingram. RH. (2004). Deep breath reversal and exponential
return of methacholine-induced obstruction in asthmatic and nonasthmatic subjects. J Appl Physiol 96: 137-
142. http://dx.doi.org/10.1152/iapplphysiol.00504.2003
Jackson. LP. (2009). [Letter from EPA Administrator Lisa P. Jackson to Dr. Jonathan M. Samet. Subject:
Changes to the NAAQS review process]. Available online at
http://vosemite.epa.gov/sab/sabproduct.nsf/WebCASAC/Jackson%2005-21-
09/$File/NAAQS%20Letter%20to%20CASAC%20Chair-Mav%202009.pdf
Jacob. DJ. (1999). Introduction to atmospheric chemistry. Princeton, NJ: Princeton University Press.
Jacobson. MZ. (2002). Atmospheric pollution: History, science, and regulation. New York: Cambridge
University Press, http://www.cambridge.org/us/academic/subjects/earth-and-environmental-
science/atmospheric-science-and-meteorology/atmospheric-pollution-history-science-and-regulation
Jacquemin. B: Sunyer. J: Forsberg. B: Aguilera. I: Bouso. L: Briggs. D: de Marco. R: Garcia-Esteban. R:
Heinrich. J: Jarvis. D: Maldonado. JA: Pavo. F: Rage. E: Vienneau. D: Kiinzli. N. (2009a). Association
between modelled traffic-related air pollution and asthma score in the ECRHS. Eur Respir J 34: 834-842.
http://dx.doi.org/10.1183/09031936.00138208
Jacquemin. B: Sunyer. J: Forsberg. B: Aguilera. I: Briggs. D: Garcia-Esteban. R: Gotschi. T: Heinrich. J:
Jarvholm. B: Jarvis. D: Vienneau. D: Kiinzli. N. (2009b). Home outdoor NO2 and new onset of self-reported
asthma in adults. Epidemiology 20: 119-126. http://dx.doi.org/10.1097/EDE.Ob013e3181886e76
Jaegle. L: Steinberger. L: Martin. RV: Chance. K. (2005). Global partitioning of NOx sources using satellite
observations: Relative roles of fossil fuel combustion, biomass burning and soil emissions. Faraday Discuss
130: 407-423. http://dx.doi.org/10.1039/b502128f
Jaffe. DH: Singer. ME: Rimm. AA. (2003). Air pollution and emergency department visits for asthma among
Ohio Medicaid recipients, 1991-1996. EnvironRes 91: 21-28. http://dx.doi.org/10.1016/S0013-
9351(02)00004-X
Jain. B: Rubinstein. I: Robbins. RA: Leise. KL: Sisson. JH. (1993). Modulation of airway epithelial cell ciliary
beat frequency by nitric oxide. Biochem Biophys Res Commun 191: 83-88.
http://dx.doi.org/10.1006/bbrc. 1993.1187
Jakab. GJ. (1987). Modulation of pulmonary defense mechanisms by acute exposures to nitrogen dioxide.
Experientia, Suppl 51: 235-242.
Jakab. GJ. (1988). Modulation of pulmonary defense mechanisms against viral and bacterial infections by acute
exposures to nitrogen dioxide (pp. 1-38). (HEI Research Report no. 20). Cambridge, MA: Health Effects
Institute. http://pubs.healtheffects.org/view.php?id=238
Jalaludin. B: Khalaj. B: Sheppeard. V: Morgan. G. (2008). Air pollution and ED visits for asthma in Australian
children: A case-crossover analysis. Int Arch Occup Environ Health 81: 967-974.
http://dx.doi.org/10.1007/s00420-007-0290-0
Jalaludin. B: Mannes. T: Morgan. G: Lincoln. D: Sheppeard. V: Corbett S. (2007). Impact of ambient air
pollution on gestational age is modified by season in Sydney, Australia. Environ Health 6: 16.
http://dx.doi.org/10.1186/1476-069X-6-16
Jalaludin. B: Morgan. G: Lincoln. D: Sheppeard. V: Simpson. R: Corbett. S. (2006). Associations between
ambient air pollution and daily emergency department attendances for cardiovascular disease in the elderly
(65+years), Sydney, Australia. JExpo Sci Environ Epidemiol 16: 225-237.
http://dx.doi.org/10.1038/si.iea.7500451
Jalaludin. BB: O'Toole. BI: Leeder. SR. (2004). Acute effects of urban ambient air pollution on respiratory
symptoms, asthma medication use, and doctor visits for asthma in a cohort of Australian children. Environ
Res 95: 32-42. http://dx.doi.org/10.1016/S0013-9351(03)00038-0
Janssen. S: Dumont. G: Fierens. F: Deutsch. F: Maiheu. B: Celis. D: Trimpeneers. E: Mensink. C. (2012). Land
use to characterize spatial representativeness of air quality monitoring stations and its relevance for model
validation. Atmos Environ 59: 492-500. http://dx.doi.0rg/10.1016/i.atmosenv.2012.05.028
R-44
-------
Jarvis. PL: Leaderer. BP: Chinn. S: Burnev. PG. (2005). Indoor nitrous acid and respiratory symptoms and lung
function in adults. Thorax 60: 474-479. http://dx.doi.org/10.1136/thx.2004.032177
Jenkins. HS: Devalia. JL: Mister. RL: Bevan. AM: Rusznak. C: Davies. RJ. (1999). The effect of exposure to
ozone and nitrogen dioxide on the airway response of atopic asthmatics to inhaled allergen: Dose- and time-
dependent effects. Am J Respir Crit Care Med 160: 33-39. http://dx.doi.0rg/10.1164/airccm.160.l.9808119
Jensen. SS: Larson. T: Deepti. KG: Kaufman. JD. (2009). Modeling traffic air pollution in street canyons in New
York City for intra-urban exposure assessment in the US Multi-Ethnic Study of atherosclerosis and air
pollution. Atmos Environ43: 4544-4556. http://dx.doi.0rg/10.1016/i.atmosenv.2009.06.042
Jerrett. M: Arain. MA: Kanaroglou. P: Beckerman. B: Grouse. D: Gilbert. NL: Brook. JR: Finkelstein. N:
Finkelstein. MM. (2007). Modeling the intraurban variability of ambient traffic pollution in Toronto, Canada.
J Toxicol Environ Health A 70: 200-212. http://dx.doi.org/10.1080/15287390600883018
Jerrett. M: Burnett. RT: Beckerman. BS: Turner. MC: Krewski. D: Thurston. G: Martin. RV: van Donkelaar. A:
Hughes. E: Shi. Y: Gapstur. SM: Thun. MJ: Pope. CA. III. (2013). Spatial analysis of air pollution and
mortality in California. Am J Respir Crit Care Med 188: 593-599. http://dx.doi.org/10.1164/rccm.201303-
0609OC
Jerrett. M: Finkelstein. MM: Brook. JR: Arain. MA: Kanaroglou. P: Stieb. DM: Gilbert. NL: Verma. D:
Finkelstein. N: Chapman. KR: Sears. MR. (2009). A cohort study of traffic-related air pollution and mortality
in Toronto, Ontario, Canada. Environ Health Perspect 117: Ill-Ill, http://dx.doi.org/10.1289/ehp.11533
Jerrett. M: Shankardass. K: Berhane. K: Gauderman. WJ: Kiinzli. N: Avol E: Gilliland. F: Lurmann. F: Molitor.
JN: Molitor. JT: Thomas. DC: Peters. J: McConnell R. (2008). Traffic-related air pollution and asthma onset
in children: A prospective cohort study with individual exposure measurement. Environ Health Perspect 116:
1433-1438. http://dx.doi.org/10.1289/ehp. 10968
Jevtic. M: Dragic. N: Bijelovic. S: Popovic. M. (2014). Cardiovascular diseases and air pollution in Novi Sad,
Serbia. Int J Occup Med Environ Health 27: 153-164. http://dx.doi.org/10.2478/sl3382-014-0239-y
Ji Y: Gao. Y: Li. G: An. T. (2012). Theoretical study of the reaction mechanism and kinetics of low-molecular-
weight atmospheric aldehydes (C1C4) withNO2. Atmos Environ 54: 288-295.
http://dx.doi.0rg/10.1016/i.atmosenv.2012.02.040
Jia. X: Song. X: Shima. M: Tamura. K: Deng. F: Guo. X. (2011). Acute effect of ambient ozone on heart rate
variability in healthy elderly subjects. JExpo Sci Environ Epidemiol 21: 541-547.
http://dx.doi.org/10.1038/ies.2011.18
Jiang. LL: Zhang. YH: Song. GX: Chen. GH: Chen. BH: Zhao. NQ: Kan. HP. (2007). A time series analysis of
outdoor air pollution and preterm birth in Shanghai, China. Biomed Environ Sci 20: 426-431.
Jilma. B: Kastner. J: Mensik. C: Vondrovec. B: Hildebrandt J: Krejcv. K: Wagner. OF: Eichler. H. (1996). Sex
differences in concentrations of exhaled nitric oxide and plasma nitrate. Life Sci 58: 469-476.
http://dx.doi.org/10.1016/0024-3205(95)02311-9
Johnson. JY: Rowe. BH: Villeneuve. PJ. (2010a). Ecological analysis of long-term exposure to ambient air
pollution and the incidence of stroke in Edmonton, Alberta, Canada. Stroke 41: 1319-1325.
http://dx.doi.0rg/10.l 161/STROKEAHA. 110.580571
Johnson. JYM: Rowe. BH: Allen. RW: Peters. PA: Villeneuve. PJ. (2013). A case-control study of medium-term
exposure to ambient nitrogen dioxide pollution and hospitalization for stroke. BMC Public Health 13: 368.
http://dx.doi.org/10.1186/1471-2458-13-368
Johnson. M: Isakov. V: Touma. JS: Mukerjee. S: Ozkaynak. H. (2010b). Evaluation of land-use regression
models used to predict air quality concentrations in an urban area. Atmos Environ 44: 3660-3668.
http://dx.doi.0rg/10.1016/i.atmosenv.2010.06.041
Jones. SL: Kittelson. J: Cowan. JO: Flannery. EM: Hancox. PJ: McLachlan. CR: Taylor. DR. (2001). The
predictive value of exhaled nitric oxide measurements in assessing changes in asthma control. Am J Respir
Crit Care Med 164: 738-743. http://dx.doi.0rg/10.1164/airccm.164.5.2012125
R-45
-------
Torres. R: Magnussen. H. (1990). Airways response of asthmatics after a 30 min exposure, at resting ventilation,
to 0.25 ppmNO2 or 0.5 ppm SO2. EurRespir J 3: 132-137.
Torres. R: Magnussen. H. (1991). Effect of 0.25 ppm nitrogen dioxide on the airway response to methacholine in
asymptomatic asthmatic patients. Lung 169: 77-85. http://dx.doi.org/10.1007/BF02714145
Torres. R: Nowak. D: Grimminger. F: Seeger. W: Oldigs. M: Magnussen. H. (1995). The effect of 1 ppm
nitrogen dioxide on bronchoalveolar lavage cells and inflammatory mediators in normal and asthmatic
subjects. Eur Respir J 8: 416-424. http://dx.doi.org/10.1183/09031936.95.08030416
Joyce. PL: von Glasow. R: Simpson. WR. (2014). The fate of NOx emissions due to nocturnal oxidation at high
latitudes: 1-D simulations and sensitivity experiments. Atmos ChemPhys 14: 7601-7616.
http://dx.doi.org/10.5194/acp-14-7601-2014
Jurewicz. J: Radwan. M: Sobala. W: Polanska. K: Radwan. P: Jakubowski. L: Ulanska. A: Hanke. W. (2014).
The relationship between exposure to air pollution and sperm disomy. Environ Mol Mutagen 56: 50-59.
http://dx.doi.org/10.1002/em.21883
Just J: Segala. C: Sahraoui F: Priol G: Grimfeld. A: Neukirch. F. (2002). Short-term health effects of
paniculate and photochemical air pollution in asthmatic children. Eur Respir J 20: 899-906.
http://dx.doi.org/10.1183/09031936.02.00236902
Kan. H: Chen. B: Zhao. N: London. SJ: Song. G: Chen. G: Zhang. Y: Jiang. L. (2010). Part 1. A time-series
study of ambient air pollution and daily mortality in Shanghai, China. In Public Health and Air Pollution in
Asia (PAPA): Coordinated studies of short-term exposure to air pollution and daily mortality in four cities
(pp. 17-78). Boston, MA: Health Effects Institute. http://pubs.healtheffects.org/view.php?id=348
Kan. H: London. SJ: Chen. G: Zhang. Y: Song. G: Zhao. N: Jiang. L: Chen. B. (2008). Season, sex, age, and
education as modifiers of the effects of outdoor air pollution on daily mortality in Shanghai, China: The
Public Health and Air Pollution in Asia (PAPA) Study. Environ Health Perspect 116: 1183-1188.
http://dx.doi.org/10.1289/ehp.10851
Kannan. S: Misra. DP: Dvonch. JT: Krishnakumar. A. (2006). Exposures to airborne paniculate matter and
adverse perinatal outcomes: A biologically plausible mechanistic framework for exploring potential effect
modification by nutrition. Environ Health Perspect 114: 1636-1642. http://dx.doi.org/10.1289/ehp.9081
Karakatsani. A: Analitis. A: Perifanou. D: Ayres. JG: Harrison. RM: Kotronarou. A: Kavouras. IG: Pekkanen. J:
Hameri. K: Kos. GP: de Hartog. JJ: Hoek. G: Katsouvannj K. (2012). Paniculate matter air pollution and
respiratory symptoms in individuals having either asthma or chronic obstructive pulmonary disease: A
European multicentre panel study. Environ Health 11: 75. http://dx.doi.org/10.1186/1476-069X-ll-75
Karner. AA: Eisinger. PS: Niemeier. DA. (2010). Near-roadway air quality: Synthesizing the findings from real-
world data. Environ Sci Technol 44: 5334-5344. http://dx.doi.org/10.1021/esl00008x
Karr. C: Lumlev. T: Shepherd. K: Davis. R: Larson. T: Ritz. B: Kaufman. J. (2006). A case-crossover study of
wintertime ambient air pollution and infant bronchiolitis. Environ Health Perspect 114: 277-281.
http://dx.doi.org/10.1289/ehp.8313
Karr. CJ: Rudra. CB: Miller. KA: Gould. TR: Larson. T: Sathyanaravana. S: Koenig. JQ. (2009). Infant exposure
to fine paniculate matter and traffic and risk of hospitalization for RSV bronchiolitis in a region with lower
ambient air pollution. Environ Res 109: 321-327. http://dx.doi.0rg/10.1016/i.envres.2008.ll.006
Kashima. S: Naruse. H: Yorifuji T: Ohki. S: Murakoshj T: Takao. S: Tsuda. T: Pol H. (2011). Residential
proximity to heavy traffic and birth weight in Shizuoka, Japan. Environ Res 111: 377-387.
http://dx.doi.0rg/10.1016/i.envres.2011.02.005
Kastner-Klein. P: Fedorovich. E: Ketzel. M: R. B: Britter. R. (2003). The modelling of turbulence from traffic in
urban dispersion models part II: Evaluation against laboratory and full-scale concentration measurements in
street canyons. Environ Fluid Mech 3: 145-172. http://dx.doi.org/10.1023/A: 1022049224166
Kastner-Klein. P: Fedorovich. E: Rotach. MW. (2001). A wind tunnel study of organised and turbulent air
motions in urban street canyons. J Wind Eng Ind Aerod 89: 849-861. http://dx.doi.org/10.1016/S0167-
6105(01)00074-5
R-46
-------
Katanoda. K: Sobue. T: Satoh. H: Taiima. K: Suzuki. T: Nakatsuka. H: Takezaki. T: Nakavama. T: Nitta. H:
Tanabe. K: Tominaga. S. (2011). An association between long-term exposure to ambient air pollution and
mortality from lung cancer and respiratory diseases in Japan. J Epidemiol 21: 132-143.
http://dx.doi.org/10.2188/iea.JE20100098
Katsouvanni. K: Samet JM: Anderson. HR: Atkinson. R: Le Tertre. A: Medina. S: Samoli. E: Touloumi. G:
Burnett. RT: Krewski. D: Ramsay. T: Dominici. F: Peng. RD: Schwartz. J: Zanobetti. A. (2009). Air
pollution and health: A European and North American approach (APHENA) (pp. 5-90). (ISSN 1041-5505
Research Report 142). Boston, MA: Health Effects Institute. http://pubs.healtheffects.org/view.php?id=327
Katsouvanni. K: Touloumi. G: Samoli. E: Gryparis. A: Le Tertre. A: Monopolis. Y: Rossi. G: Zmirou. D:
Ballester. F: Boumghar. A: Anderson. HR: Wojtyniak. B: Paldv. A: Braunstein. R: Pekkanen. J: Schindler.
C: Schwartz. J. (2001). Confounding and effect modification in the short-term effects of ambient particles on
total mortality: Results from 29 European cities within the APHEA2 project. Epidemiology 12: 521-531.
Katsouvanni. K: Touloumi. G: Samoli. E: Petasakis. Y: Analitis. A: Le Tertre. A: Rossi. G: Zmirou. D:
Ballester. F: Boumghar. A: Anderson. HR: Wojtyniak. B: Paldv. A: Braunstein. R: Pekkanen. J: Schindler.
C: Schwartz. J. (2003). Sensitivity analysis of various models of short-term effects of ambient particles on
total mortality in 29 cities in APHEA2. In Revised Analyses of Time-series Studies of Air Pollution and
Health (pp. 157-164). Boston, MA: Health Effects Institute.
Kebabian. PL: Wood. EC: Herndon. SC: Freedman. A. (2008). A practical alternative to chemiluminescence-
based detection of nitrogen dioxide: Cavity attenuated phase shift spectroscopy. Environ Sci Technol 42:
6040-6045. http://dx.doi.org/10.1021/es703204j
Kehrl. HR: Peden. DB: Ball. B: Folinsbee. LJ: Horstman. D. (1999). Increased specific airway reactivity of
persons with mild allergic asthma after 7.6 hours of exposure to 0.16 ppm ozone. J Allergy Clin Immunol
104: 1198-1204. http://dx.doi.org/10.1016/S0091-6749(99)70013-8
Kelishadi. R: Mirghaffari. N: Poursafa. P: Gidding. SS. (2009). Lifestyle and environmental factors associated
with inflammation, oxidative stress and insulin resistance in children. Atherosclerosis 203: 311-319.
http://dx.doi.0rg/10.1016/i.atherosclerosis.2008.06.022
Keller. JP: Olives. C: Kim. SY: Sheppard. L: Sampson. PD: Szpiro. AA: Oron. AP: Lindstrom. J: Vedal S:
Kaufman. JD. (2015). A unified spatiotemporal modeling approach for predicting concentrations of multiple
air pollutants in the multi-ethnic study of atherosclerosis and air pollution. Environ Health Perspect 123: 301-
309. http://dx.doi.org/10.1289/ehp.1408145
Kelly. FJ: Blomberg. A: Frew. A: Holgate. ST: Sandstrom. T. (1996a). Antioxidant kinetics in lung lavage fluid
following exposure of humans to nitrogen dioxide. Am J Respir Crit Care Med 154: 1700-1705.
http://dx.doi.0rg/10.1164/airccm.154.6.8970358
Kelly. FJ: Cotgrove. M: Mudwav. IS. (1996b). Respiratory tract lining fluid antioxidants: the first line of defence
against gaseous pollutants. Cent Eur J Public Health 4: 11-14.
Kelly. FJ: Tetlev. TD. (1997). Nitrogen dioxide depletes uric acid and ascorbic acid but not glutathione from
lung lining fluid. Biochemistry 325: 95-99.
Kettle. AJ: Clark. BM: Winterbourn. CC. (2004). Superoxide converts indigo carmine to isatin sulfonic acid:
Implications for the hypothesis that neutrophils produce ozone. J Biol Chem 279: 18521-18525.
http://dx.doi.org/10.1074/ibc.M400334200
Khafaie. MA: Salvi. SS: Ojha. A: Khafaie. B: Gore. SS: Yajnik. CS. (2013). Systemic inflammation (C-reactive
protein) in type 2 diabetic patients is associated with ambient air pollution in Pune City, India. Diabetes Care
36: 625-630. http://dx.doi.org/10.2337/dcl2-0388
Kharitonov. SA: Barnes. PJ. (2000). Clinical aspects of exhaled nitric oxide [Review]. Eur Respir J 16: 781-792.
Kharitonov. SA: Walker. L: Barnes. PJ. (2005). Repeatability of standardised nasal nitric oxide measurements in
healthy and asthmatic adults and children. Respir Med 99: 1105-1114.
http://dx.doi.0rg/10.1016/i.rmed.2005.02.019
R-47
-------
Kheirbek. I: Ito. K: Neitzel R: Kim. J: Johnson. S: Ross. Z: Eisl H: Matte. T. (2014). Spatial Variation in
Environmental Noise and Air Pollution in New York City. J Urban Health 91: 415-431.
http://dx.doi.org/10.1007/sll524-013-9857-0
Khoder. MI. (2002). Nitrous acid concentrations in homes and offices in residential areas in Greater Cairo. J
Environ Monit 4: 573-578. http://dx.doi.org/10.1039/b201818g
Khoo. NK: Rudolph. V: Cole. MP: Golin-Bisello. F: Schopfer. FJ: Woodcock. SR: Batthyany. C: Freeman. BA.
(2010). Activation of vascular endothelial nitric oxide synthase and heme oxygenase-1 expression by
electrophilic nitro-fatty acids. Free Radic Biol Med 48: 230-239.
http://dx.doi.0rg/10.1016/i.freeradbiomed.2009.10.046
Kikugawa. K: Kogi K. (1987). Oxidation products of linoleate and linolenate exposed to nitrogen dioxide.
ChemPharmBull (Tokyo) 35: 344-349. http://dx.doi.org/10.1248/cpb.35.344
Kikuya. M: Hozawa. A: Ohokubo. T: Tsuji. I: Michimata. M: Matsubara. M: Ota. M: Nagai. K: Araki. T: Satoh.
H: Ito. S: Hisamichj S: Imai Y. (2000). Prognostic significance of blood pressure and heart rate variabilities:
the Ohasama study. Hypertension 36: 901-906. http://dx.doi.Org/10.1161/01.HYP.36.5.901
Kilkenny. C: Browne. WJ: Cuthill 1C: Emerson. M: Altaian. DG. (2010). Improving bioscience research
reporting: The ARRIVE guidelines for reporting animal research [Review]. PLoS Biol 8: e!000412.
http://dx.doi.org/10.1371/iournal.pbio.1000412
Kim. D: Sass-Kortsak. A: Purdham JT: Dales. RE: Brook. JR. (2006a). Associations between personal
exposures and fixed-site ambient measurements of fine paniculate matter, nitrogen dioxide, and carbon
monoxide in Toronto, Canada. JExpo Sci Environ Epidemiol 16: 172-183.
http://dx.doi.org/10.1038/si.iea.7500446
Kim. E: Park. H: Hong. YC: Ha. M: Kim Y: Kim. BN: Kim. Y: Roh. YM: Lee. BE: Ryu. JM: Kim. BM: Ha.
EH. (2014). Prenatal exposure to PM10 and NO2 and children's neurodevelopment from birth to 24 months
of age: Mothers and Children's Environmental Health (MOCEH) study. Sci Total Environ 481: 439-445.
http://dx.doi.0rg/10.1016/i.scitotenv.2014.01.107
Kim. JH: Hong. YC. (2012). GSTM1, GSTT1, and GSTP1 polymorphisms and associations between air
pollutants and markers of insulin resistance in elderly Koreans. Environ Health Perspect 120: 1378-1384.
http://dx.doi.org/10.1289/ehp. 1104406
Kim. JJ: Smorodinsky. S: Lipsett M: Singer. BC: Hodgson. AT: Ostro. B. (2004). Traffic-related air pollution
near busy roads: the East Bay children's Respiratory Health Study. Am J Respir Crit Care Med 170: 520-526.
http://dx.doi.org/10.1164/rccm.200403-281OC
Kim. S: Kim. Y: Lee. MR: Kim. J: Jung. A: Park. JS: Jang. AS: Park. SW: Uh. ST: Choi. JS: Kim YH: Buckley.
T: Park. CS. (2012). Winter season temperature drops and sulfur dioxide levels affect on exacerbation of
refractory asthma in South Korea: a time-trend controlled case-crossover study using soonchunhyang asthma
cohort data. J Asthma 49: 679-687. http://dx.doi.org/10.3109/02770903.2012.702839
Kim. SW: Heckel. A: McKeen. SA: Frost. GJ: Hsie. E. -Y: Trainer. MK: Richter. A: Burrows. JP: Peckham. SE:
Grell. GA. (2006b). Satellite-observed U.S. power plant NOx emission reductions and their impact on air
quality. Geophys Res Lett 33: L22812. http://dx.doi.org/10.1029/2006GL027749
Kim. SW: McKeen. SA: Frost. GJ: Lee. SH: Trainer. M: Richter. A: Angevine. WM: Atlas. E: Bianco. L:
Boersma. KF: Brioude. J: Burrows. JP: de Gouw. J: Fried. A: Gleason. J: Hilboll. A: Mellqvist. J: Peischl. J:
Richter. D: Rivera. C: Ryerson. T: Te Lintel Hekkert. S: Walega. J: Warneke. C: Weibring. P: Williams. E.
(2011). Evaluations of NOx and highly reactive VOC emission inventories in Texas and their implications
for ozone plume simulations during the Texas Air Quality Study 2006. Atmos ChemPhys 11: 11361-11386.
http://dx.doi.org/10.5194/acp-ll-11361-2011
Kim. SY: O'Neill. MS: Lee. JT: Cho. Y: Kim. J: Kim. H. (2007). Air pollution, socioeconomic position, and
emergency hospital visits for asthma in Seoul, Korea. Int Arch Occup Environ Health 80: 701-710.
http://dx.doi.org/10.1007/s00420-007-0182-3
Kim. YM: Kim. YS: Jeon. SG: Kim. YK. (2013). Immunopathogenesis of allergic asthma: more than the Th2
hypothesis. Allergy Asthma Immunol Res 5: 189-196. http://dx.doi.0rg/10.4168/aair.2013.5.4.189
R-48
-------
Kimbrough. ES: Baldauf. RW: Watkins. N. (2013). Seasonal and diurnal analysis of NO2 concentrations from a
long-duration study conducted in Las Vegas, Nevada. J Air Waste Manag Assoc 63: 934-942.
http://dx.doi.org/10.1080/10962247.2013.795919
Kingslev. SL: Eliot MN: Carlson. L: Finn. J: Macintosh. PL: Suh. HH: Wellenius. GA. (2014). Proximity of US
schools to major roadways: a nationwide assessment. J Expo Sci Environ Epidemiol 24: 253-259.
http://dx.doi.0rg/10.1038/ies.2014.5
Kinsella. J: Carter. R: Reid. WH: Campbell D: Clark. CJ. (1991). Increased airways reactivity after smoke
inhalation. Lancet 337: 595-597.
Kirsch. M: Korth. HG: Sustmann. R: De Groot H. (2002). The pathobiochemistry of nitrogen dioxide [Review].
Biol Chem 383: 389-399. http://dx.doi.org/10.1515/BC.2002.043
Kitabatake. M: Yamamoto. H: Yuan. PF: Manjurul H: Murase. S: Yamauchi. T. (1995). Effects of exposure to
NO2 or SO2 on bronchopulmonary reaction induced by Candida albicans in guinea pigs. J Toxicol Environ
Health45: 75-82. http://dx.doi.org/10.1080/15287399509531981
Kjaergaard. SK: Rasmussen. TR. (1996). Clinical studies of effects of nitrogen oxides in healthy and asthmatic
subjects. Cent Eur J Public Health 4: 23-26.
Klapmever. ME: Marr. LC. (2012). CO2, NOx, and particle emissions from aircraft and support activities at a
regional airport. Environ Sci Technol 46: 10974-10981. http://dx.doi.org/10.1021/es302346x
Kleeberger. SR: Zhang. LY: Jakab. GJ. (1997). Differential susceptibility to oxidant exposure in inbred strains of
mice: Nitrogen dioxide versus ozone. Inhal Toxicol 9: 601-621. http://dx.doi.org/10.1080/089583797198006
Kleffmann. J: Tapia. GV: Bejan. I: Kurtenbach. R: Wiesen. P. (2013). NO2 measurement techniques: Pitfalls
and new developments. In I Barnes; KJ Rudzin'ski (Eds.), Disposal of dangerous chemicals in urban areas
and mega cities. New York, NY: Springer, http://dx.doi.org/10.1007/978-94-007-5034-0 2
Kleinman. MT: Bailey. RM: Linn. WS: Anderson. KR: Whynot JD: Shamoo. DA: Hackney. JD. (1983). Effects
of 0.2 ppm nitrogen dioxide on pulmonary function and response to bronchoprovocation in asthmatics. J
Toxicol Environ Health 12: 815-826. http://dx.doi.org/10.1080/15287398309530472
Kleinman. MT: Mautz. WJ. (1991). The effects of exercise on dose and dose distribution of inhaled automotive
pollutants. (Research Report no. 45). Cambridge, MA: Health Effects Institute.
http://pubs.healtheffects.org/view.php?id=45
Klepeis. NE: Nelson. WC: Ott WR: Robinson. JP: Tsang. AM: Switzer. P: Behar. JV: Hern. SC: Engelmann.
WH. (2001). The National Human Activity Pattern Survey (NHAPS): A resource for assessing exposure to
environmental pollutants. J Expo Anal Environ Epidemiol 11: 231-252.
http://dx.doi.org/10.1038/si.iea.7500165
Klepeis. NE: Tsang. AM: Behar. JV. (1996). Analysis of the national human activity pattern survey (NHAPS)
respondents from a standpoint of exposure assessment [EPA Report]. (EPA/600/R-96/074). Washington, DC:
U.S. Environmental Protection Agency.
http://exposurescience.org/pub/reports/NHAPS Reportl.pdfW....Local SettingsTemporary Internet
FilesContent.Outlook3JQ221FPB_Approaches_Population_Tables.docx
Klimisch. HJ: Andreae. M: Tillmann. U. (1997). A systematic approach for evaluating the quality of
experimental lexicological and ecotoxicological data. Regul Toxicol Pharmacol 25: 1-5.
http://dx.doi.org/10.1006/rtph.1996.1076
Kloek. J: Mortaz. E: van Ark. I: Lilly. CM: Nijkamp. FP: Folkerts. G. (2010). Glutathione prevents the early
asthmatic reaction and airway hyperresponsiveness in guinea pigs. J Physiol Pharmacol 61: 67-72.
Knuckles. TL: Buntz. JG: Paffett M: Channell. M: Harmon. M: Cherng. T: Lucas. SN: McDonald. JD: Kanagy.
NL: Campen. MJ. (2011). Formation of vascular S-nitrosothiols and plasma nitrates/nitrites following
inhalation of diesel emissions. J Toxicol Environ Health A 74: 828-837.
http://dx.doi.org/10.1080/15287394.2011.570225
R-49
-------
Ko. FW: Tarn. W: Wong. TW: Chan. DP: Tung. AH: Lai. CKW: Hui. DSC. (2007a). Temporal relationship
between air pollutants and hospital admissions for chronic obstructive pulmonary disease in Hong Kong.
Thorax 62: 780-785. http://dx.doi.org/10.1136/thx.2006.076166
Ko. FWS: Tarn W: Wong. TW: Lai. CKW: Wong. GWK: Leung. TF: Ng. SSS: Hui. DSC. (2007b). Effects of
air pollution on asthma hospitalization rates in different age groups in Hong Kong. Clin Exp Allergy 37:
1312-1319. http://dx.doi.0rg/10.llll/i.1365-2222.2007.02791.x
Kobavashi. T: Miura. T. (1995). Concentration- and time-dependent increase in specific airway resistance after
induction of airway hyperresponsiveness by subchronic exposure of guinea pigs to nitrogen dioxide. Fundam
Appl Toxicol 25: 154-158. http://dx.doi.org/10.1006/faat.1995.1049
Kobavashi. T: Shinozaki. Y. (1990). Effect of subacute exposure to nitrogen dioxide on the airway
responsiveness of guinea pig. Agents Actions Suppl 31: 71-74.
Kodama. Y: Arashidani. K: Tokui. N: Kawamoto. T: Matsuno. K: Kunugita. N: Minakawa. N. (2002).
Environmental NO2 concentration and exposure in daily life along main roads in Tokyo. Environ Res 89:
236-244. http://dx.doi.org/10.1006/enrs.2002.4350
Koehler. C: Ginzkev. C: Friehs. G: Hackenberg. S: Froelich. K: Scherzed. A: Burghartz. M: Kessler. M:
Kleinsasser. N. (2010). Aspects of nitrogen dioxide toxicity in environmental urban concentrations in human
nasal epithelium. Toxicol Appl Pharmacol 245: 219-225. http://dx.doi.0rg/10.1016/i.taap.2010.03.003
Koehler. C: Ginzkev. C: Friehs. G: Hackenberg. S: Froelich. K: Scherzed. A: Burghartz. M: Kessler. M:
Kleinsasser. N. (2011). Ex vivo toxicity of nitrogen dioxide in human nasal epithelium at the WHO defined
1-h limit value. Toxicol Lett 207: 89-95. http://dx.doi.0rg/10.1016/i.toxlet.2011.08.004
Koehler. C: Thielen. S: Ginzkev. C: Hackenberg. S: Scherzed. A: Burghartz. M: Paulus. M: Hagen. R:
Kleinsasser. NH. (2013). Nitrogen dioxide is genotoxic in urban concentrations. Inhal Toxicol 25: 341-347.
http://dx.doi.org/10.3109/08958378.2013.788104
Koenig. JQ: Covert. PS: Marshall SG: Van Belle. G: Pierson. WE. (1987). The effects of ozone and nitrogen
dioxide on pulmonary function in healthy and in asthmatic adolescents. Am J Respir Crit Care Med 136:
1152-1157. http://dx.doi.0rg/10.1164/airccm/136.5.1152
Kornartit. C: Sokhi. RS: Burton. MA: Ravindra. K. (2010). Activity pattern and personal exposure to nitrogen
dioxide in indoor and outdoor microenvironments. Environ Int 36: 36-45.
http://dx.doi.0rg/10.1016/i.envint.2009.09.004
Kousa. A: Monn. C: Rotko. T: Aim. S: Oblesbv. L: Jantunen. MJ. (2001). Personal exposures to NO2 in the
EXPOLIS-study: Relation to residential indoor, outdoor and workplace concentrations in Basel, Helsinki and
Prague. Atmos Environ 35: 3405-3412.
Koutrakis. P: Suh. HH: Sarnat JA: Brown. KW: Coull BA: Schwartz. J. (2005). Characterization of paniculate
and gas exposures of sensitive subpopulations living in Baltimore and Boston (pp. 1-65; discussion 67-75).
(ISSN 1041-5505131). Boston, MA: Health Effects Institute. http://pubs.healtheffects.org/view.php?id=91
Kramer. U: Herder. C: Sugiri. D: Strassburger. K: Schikowski. T: Ranft. U: Rathmann. W. (2010). Traffic-
related air pollution and incident type 2 diabetes: Results from the SALIA cohort study. Environ Health
Perspect 118: 1273-1279. http://dx.doi.org/10.1289/ehp.0901689
Kramer. U: Sugiri. D: Ranft. U: Krutmann. J: von Berg. A: Berdel D: Behrendt H: Kuhlbusch. T: Hochadel M:
Wichmann. HE: Heinrich. J. (2009). Eczema, respiratory allergies, and traffic-related air pollution in birth
cohorts from small-town areas. J Dermatol Sci 56: 99-105. http://dx.doi.0rg/10.1016/i.jdermsci.2009.07.014
Krewski. D: Burnett RT: Goldberg. MS: Hoover. K: Siemiatvcki. J: Jerrett. M: Abrahamowicz. M: White. WH.
(2000). Reanalysis of the Harvard Six Cities study and the American Cancer Society study of paniculate air
pollution and mortality. Cambridge, MA: Health Effects Institute.
http://pubs.healtheffects.org/view.php?id=6
R-50
-------
Krewski. D: Jerrett M: Burnett. RT: Ma. R: Hughes. E: Shi. Y: Turner. MC: Pope AC. I: Thurston. G: Calle.
EE: Thun. MJ: Beckerman. B: Deluca. P: Finkelstein. N: Ito. K: Moore. DK: Newbold. KB: Ramsay. T:
Ross. Z: Shin. H: Tempalski. B. (2009). Extended follow-up and spatial analysis of the American Cancer
Society study linking paniculate air pollution and mortality (pp. 5-114; discussion 115-136). (ISSN 1041-
5505, HEI Research Report 140). Boston, MA: Health Effects Institute.
http://pubs.healtheffects.org/view.php?id=315
Kripke. BJ: Sherwin. RP. (1984). Nitrogen dioxide exposure—influence on rat testes. Anesth Analg 63: 526-528.
Krishna. MT: Holgate. ST. (1999). Inflammatory mechanisms underlying potentiation of effects of inhaled
aeroallergens in response to nitrogen dioxide in allergic airways disease [Review]. Clin Exp Allergy 29: 150-
154. http://dx.doi.0rg/10.1046/j.1365-2222.1999.00512.x
Kruize. H: Driessen. PP: Glasbergen. P: vanEgmond. KN. (2007). Environmental equity and the role of public
policy: Experiences in the Rijnmond region. Environ Manage 40: 578-595. http://dx.doi.org/10.1007/s00267-
005-0378-9
Kruize. H: Hanninen. O: Breugelmans. O: Lebret E: Jantunen. M. (2003). Description and demonstration of the
EXPOLIS simulation model: two examples of modeling population exposure to paniculate matter. J Expo
Anal Environ Epidemiol 13: 87-99. http://dx.doi.org/10.1038/sj.jea.7500258
Kubota. K: Murakami. M: Takenaka. S: Kawai. K: Kyono. H. (1987). Effects of long-term nitrogen dioxide
exposure on rat lung: Morphological observations. Environ Health Perspect 73: 157-169.
Kudo. M: Ishigatsubo. Y: Aoki. I. (2013a). Pathology of asthma [Review]. FMICB 4: 1-16.
http://dx.doi.org/10.3389/fmicb.2013.00263
Kudo. M: Khalifeh Soltani. SM: Sakuma. SA: McKlerov. W: Lee. TH: Woodruff. PG: Lee. JW: Huang. K:
Chen. C: Arjomandi. M: Huang. X: Atabai. K. (2013b). MfgeS suppresses airway hyperresponsiveness in
asthma by regulating smooth muscle contraction. PNAS 110: 660-665.
http://dx.doi.org/10.1073/pnas. 1216673110
Kulle. TJ: Clements. ML. (1988). Susceptibility to virus infection with exposure to nitrogen dioxide (pp. 5-21).
(ISSN 1041-5505). Cambridge, MA: Health Effects Institute. http://pubs.healtheffects.org/view.php?id=243
Kumae. T: Arakawa. H. (2006). Comparison of effects of in vivo nitrogen dioxide exposure starting from
different periods on alveolar macrophage activity, assessed by a chemiluminescence technique in Brown-
Norway rats. Luminescence 21: 226-232. http://dx.doi.org/10.1002Mo.911
Kunimoto. M: Mochitate. K: Kava. K: Miura. T: Kubota. K. (1984). Effects of nitrogen dioxide on red blood
cells of rats: Alterations of cell membrane components and populational changes of red blood cells during in
vivo exposure to NO2. EnvironRes 33: 361-369.
La Rovere. MT: Pinna. GD: Maestri. R: Mortara. A: Capomolla. S: Febo. O: Ferrari. R: Franchini. M: Gnemmi.
M: Opasich. C: Riccardi. PG: Traversi. E: Cobelli. F. (2003). Short-term heart rate variability strongly
predicts sudden cardiac death in chronic heart failure patients. Circulation 107: 565-570.
LaFranchi. BW: Goldstein. AH: Cohen. RC. (2011). Observations of the temperature dependent response of
ozone to NOx reductions in the Sacramento, CA urban plume. Atmos Chem Phys 11: 6945-6960.
http://dx.doi.org/10.5194/acp-ll-6945-2011
Lafuma. C: Harf. A: Lange. F: Bozzi. L: Poncv. JL: Bignon. J. (1987). Effect of low-level NO2 chronic exposure
onelastase-induced emphysema. EnvironRes 43: 75-84. http://dx.doi.org/10.1016/S0013-9351(87)80059-2
Lagorio. S: Forastiere. F: Pistelli. R: lavarone. I: Michelozzi. P: Fano. V: Marconi. A: Ziemacki. G: Ostro. BD.
(2006). Air pollution and lung function among susceptible adult subjects: A panel study. Environ Health 5:
11. http://dx.doi.org/10.1186/1476-069X-5-11
Lampl. M: Jeantv. P. (2003). Timing is everything: A reconsideration of fetal growth velocity patterns identifies
the importance of individual and sex differences. Am J Hum Biol 15: 667-680.
http://dx.doi.org/10.1002/aihb. 10204
R-51
-------
Lamsal LN: Martin. RV: Van Donkelaar. A: Celarier. EA: Bucsela. EJ: Boersma. KF: Dirksen. R: Luo. C:
Wang. Y. (2010). Indirect validation of tropospheric nitrogen dioxide retrieved from the OMI satellite
instrument: Insight into the seasonal variation of nitrogen oxides at northern midlatitudes. J Geophys Res
115: D05302. http://dx.doi.org/10.1029/2009JD013351
Lamsal. LN: Martin. RV: Van Donkelaar. A: Steinbacher. M: Celarier. EA: Bucsela. E: Dunlea. EJ: Pinto. JP.
(2008). Ground-level nitrogen dioxide concentrations inferred from the satellite-borne ozone monitoring
instrument. J Geophys Res 113: D16308. http://dx.doi.org/10.1029/2007JD009235
Langrish. JP: Lundba'ck. M: Barath. S: Soderberg. S: Mills. NL: Newbv. DE: Sandstrom. T: Blomberg. A.
(2010). Exposure to nitrogen dioxide is not associated with vascular dysfunction in man. Inhal Toxicol 22:
192-198. http://dx.doi.org/10.3109/08958370903144105
Larrieu. S: Jusot JF: Blanchard. M: Prouvost H: Declercq. C: Fabre. P: Pascal. L: Le Tertre. A: Wagner. V:
Riviere. S: Chardon. B: Borelli. D: Cassadou. S: Eilstein. D: Lefranc. A. (2007). Short term effects of air
pollution on hospitalizations for cardiovascular diseases in eight French cities: The PSAS program. Sci Total
Environ 387: 105-112. http://dx.doi.0rg/10.1016/i.scitotenv.2007.07.025
Last JA: Warren. PL. (1987). Synergistic interaction between nitrogen dioxide and respirable aerosols of
sulfuric acid or sodium chloride on rat lungs. Toxicol Appl Pharmacol 90: 34-42.
http://dx.doi.org/10.1016/0041-008X(87)90303-6
Laumbach. RJ: Rich. DQ: Gandhi. S: Amorosa. L: Schneider. S: Zhang. J: Ohman-Strickland. P: Gong. J:
Lelyanov. O: Kipen. HM. (2010). Acute changes in heart rate variability in subjects with diabetes following a
highway traffic exposure. J Occup EnvironMed 52: 324-331.
http://dx.doi.org/10.1097/JOM.Ob013e3181d241fa
Laupsa. H: Slordal LH. (2003). Applying model calculations to estimate urban air quality with respect to the
requirements of the EU directives on NO2, PM10 and C6H6. Int J Environ Pollut 20: 309-320.
Laurent O: Hu. J: Li. L: Cockburn. M: Escobedo. L: Kleeman. MJ: Wu. J. (2014). Sources and contents of air
pollution affecting term low birth weight in Los Angeles County, California, 20012008. Environ Res 134:
488-495. http://dx.doi.0rg/10.1016/i.envres.2014.05.003
Laurent O: Pedrono. G: Filleul L: Segala. C: Lefranc. A: Schillinger. C: Riviere. E: Bard. D. (2009). Influence
of socioeconomic deprivation on the relation between air pollution and beta-agonist sales for asthma. Chest
135: 717-723. http://dx.doi.org/10.1378/chest.08-1604
Laurent O: Wu. J: Li. L: Chung. J: Bartell S. (2013). Investigating the association between birth weight and
complementary air pollution metrics: a cohort study. Environ Health 12: 18. http://dx.doi.org/10.1186/1476-
069X-12-18
Le. HQ: Batterman. SA: Wirth. JJ: Wahl. RL: Hoggatt KJ: Sadeghnejad. A: Hultin. ML: Depa. M. (2012). Air
pollutant exposure and preterm and term small-for-gestational-age births in Detroit, Michigan: Long-term
trends and associations. Environ Int 44: 7-17. http://dx.doi.0rg/10.1016/j.envint.2012.01.003
Leaderer. BP: Naeher. L: Jankun. T: Balenger. K: Holford. TR: Toth. C: Sullivan. J: Wolfson. JM: Koutrakis. P.
(1999). Indoor, outdoor, and regional summer and winter concentrations of PM10, PM25, SO42-, H+, NH4+,
NO3-, NH3, and nitrous acid in homes with and without kerosene space heaters. Environ Health Perspect
107:223-231.
Lee. BE: Ha. EH: Park. HS: Kim. YJ: Hong. YC: Kim. H: Lee. JT. (2003). Exposure to air pollution during
different gestational phases contributes to risks of low birth weight. Hum Reprod 18: 638-643.
http://dx.doi.org/10.1093/humrep/degl02
Lee. BH: Wood. EC: Zahniser. MS: Mcmanus. JB: Nelson. DP: Herndon. SC: Santoni. GW: Wofsv. SC:
Munger. JW. (201 la). Simultaneous measurements of atmospheric HONO andNO2 viaabsorption
spectroscopy using tunable mid-infrared continuous-wave quantum cascade lasers. Appl Phys B 102: 417-
423. http://dx.doi.org/10.1007/s00340-010-4266-5
Lee. BJ: Jeung. YJ: Lee. JY: Choi. DC. (201 Ib). Increased S-nitrosothiol levels in nonasthmatic eosinophilic
bronchitis compared with cough variant asthma. Int Arch Allergy Immunol 156: 99-103.
http://dx.doi.org/10.1159/000321919
R-52
-------
Lee. H: Gu. M: Kim. YJ: Hwang. J: Jung. U. (2012a). First-time remote sensing of NO2 vertical distributions in
an urban street canyon using Topographic Target Light scattering Differential Optical Absorption
Spectroscopy (ToTaL-DOAS). Atmos Environ 54: 519-528.
http://dx.doi.0rg/10.1016/i.atmosenv.2012.02.065
Lee. K: Xue. J: Geyh. AS: Ozkaynak. H: Leaderer. BP: Weschler. CJ: Spengler. JD. (2002). Nitrous acid,
nitrogen dioxide, and ozone concentrations in residential environments. Environ Health Perspect 110: 145-
150.
Lee. K: Yang. W: Bofinger. ND. (2000). Impact of microenvironmental nitrogen dioxide concentrations on
personal exposures in Australia. J Air Waste Manag Assoc 50: 1739-1744.
Lee. K: Yeom J: Yoon. C: Yang. W: Son. BS: Jeon. JM: Yu. SD: Park. CH. (2013). Seasonal and geographic
effects on predicting personal exposure to nitrogen dioxide by time-weighted microenvironmental model.
Atmos Environ 67: 143-148. http://dx.doi.0rg/10.1016/i.atmosenv.2012.ll.010
Lee. PC: Talbott EO: Roberts. JM: Catov. JM: Bilonick. RA: Stone. RA: Sharma. RK: Ritz. B. (2012b).
Ambient air pollution exposure and blood pressure changes during pregnancy. Environ Res 117: 46-53.
http://dx.doi.0rg/10.1016/i.envres.2012.05.011
Lee. PC: Talbott. EO: Roberts. JM: Catov. JM: Sharma. RK: Ritz. B. (201 Ic). Paniculate air pollution exposure
and C-reactive protein during early pregnancy. Epidemiology 22: 524-531.
http://dx.doi.org/10.1097/EDE.Ob013e31821c6c58
Lee. YL: Wang. WH: Lu. CW: Lin. YH: Hwang. BF. (20 lid). Effects of ambient air pollution on pulmonary
function among schoolchildren. Int J Hyg Environ Health 214: 369-375.
http://dx.doi.org/10.1016/Ujteh2011.05.004
Leem. JH: Kaplan. BM: Shim. YK: Pohl. HR: Gotwav. CA: Bullard. SM: Rogers. JF: Smith MM: Tylenda. CA.
(2006). Exposures to air pollutants during pregnancy and preterm delivery. Environ Health Perspect 114:
905-910. http://dx.doi.org/10.1289/ehp.8733
Legro. RS: Sauer. MV: Mottla. GL: Richter. KS: Li. X: Dodson. WC: Liao. D. (2010). Effect of air quality on
assisted human reproduction. Hum Reprod 25: 1317-1324. http://dx.doi.org/10.1093/humrep/deq021
Lehnert. BE: Archuleta. DC: Ellis. T: Session. WS: Lehnert NM: Gurlev. LR: Stavert. DM. (1994). Lung injury
following exposure of rats to relatively high mass concentrations of nitrogen dioxide. Toxicology 89: 239-
277. http://dx.doi.org/10.1016/0300-483X(94)90100-7
Lenters. V: Uiterwaal CS: Beelen. R: Bots. ML: Fischer. P: Brunekreef. B: Hoek. G. (2010). Long-term
exposure to air pollution and vascular damage in young adults. Epidemiology 21:512-520.
http://dx.doi.org/10.1097/EDE.Ob013e3181dec3a7
Lepeule. J. (2014). Email from Johanna Lepeule to Molini Patel. Re: question re: in press EHP paper (With
attachments). Available online
Lepeule. J: Bind. MA: Baccarelli. AA: Koutrakis. P: Tarantinj L: Litonjua. A: Sparrow. D: Vokonas. P:
Schwartz. JD. (2014). Epigenetic influences on associations between air pollutants and lung function in
elderly men: the normative aging study. Environ Health Perspect 122: 566-572.
http://dx.doi.org/10.1289/ehp. 1206458
Lepeule. J: Camj F: Bottagisi S: Galineau. J: Hulin. A: Marquis. N: Bohet A: Siroux. V: Kaminski. M:
Charles. MA: Slama. R. (2010). Maternal exposure to nitrogen dioxide during pregnancy and offspring birth
weight: Comparison of two exposure models. Environ Health Perspect 118: 1483-1489.
http://dx.doi.org/10.1289/ehp.0901509
Lew. D: Sheppard. L: Checkowav. H: Kaufman. J: Lumlev. T: Koenig. J: Siscovick. D. (2001). A case-
crossover analysis of paniculate matter air pollution and out-of-hospital primary cardiac arrest. Epidemiology
12: 193-199.
Lew. I: Mihele. C: Lu. G: Naravan. J: Hilker. N: Brook. JR. (2014). Elucidating multipollutant exposure across
a complex metropolitan area by systematic deployment of a mobile laboratory. Atmos ChemPhys 14: 7173-
7193. http://dx.doi.org/10.5194/acp-14-7173-2014
R-53
-------
Lew. JL: Baxter. LK: Cloughertv. JE. (2006). The air quality impacts of road closures associated with the 2004
Democratic National Convention in Boston. Environ Health 5: 16. http://dx.doi.org/10.1186/1476-069X-5-16
Li. H: Chen. L: Guo. Z: Sang. N: Li. G. (2012a). In vivo screening to determine neurological hazards of nitrogen
dioxide (NO2) using Wistar rats. J Hazard Mater 225-226: 46-53.
http://dx.doi.0rg/10.1016/i.jhazmat.2012.04.063
Li. H: Han. M: Guo. L: Li. G: Sang. N. (201 la). Oxidative stress, endothelial dysfunction and inflammatory
response in rat heart to NO inhalation exposure. Chemosphere 82: 1589-1596.
http://dx.doi.0rg/10.1016/i.chemosphere.2010.ll.055
Li. H: Li. QD: Wang. MS: Li. FJ: Li. QH: Ma. XJ: Wang. D. (2013). Smoking and air pollution exposure and
lung cancer mortality in Zhaoyuan County. Int J Hyg Environ Health 216: 63-70.
http://dx.doi.0rg/10.1016/i.iiheh.2012.06.003
Li. L: Wu. J: Wilhelm M: Ritz. B. (2012b). Use of generalized additive models and cokriging of spatial
residuals to improve land-use regression estimates of nitrogen oxides in Southern California. Atmos Environ
55: 220-228. http://dx.doi.0rg/10.1016/i.atmosenv.2012.03.035
Li. S: Batterman. S: Wasilevich. E: Wahl R: Wirth. J: Su. FC: Mukherjee. B. (201 Ib). Association of daily
asthma emergency department visits and hospital admissions with ambient air pollutants among the pediatric
Medicaid population in Detroit: Time-series and time-stratified case-crossover analyses with threshold
effects. Environ Res 111: 1137-1147. http://dx.doi.0rg/10.1016/i.envres.2011.06.002
Li. X: Liu. C: Leung. D. (2005). Development of a model for the determination of air exchange rates for street
canyons. Atmos Environ 39: 7285-7296. http://dx.doi.0rg/10.1016/i.atmosenv.2005.09.007
Liard. R: Zureik. M: Le Moullec. Y: Soussan. D: Glorian. M: Grimfeld. A: Neukirch. F. (1999). Use of personal
passive samplers for measurement of NO2, NO, and O3 levels in panel studies. EnvironRes 81: 339-348.
http://dx.doi.org/10.1006/enrs.1999.3993
Lima. ES: Bonini. MG: Augusto. O: Barbeiro. HV: Souza. HP: Abdalla. PS. (2005). Nitrated lipids decompose
to nitric oxide and lipid radicals and cause vasorelaxation. Free Radic Biol Med 39: 532-539.
http://dx.doi.0rg/10.1016/i.freeradbiomed.2005.04.005
Lin. CA: Pereira. LAA: Nishioka. DC: Conceicao. QMS: Graga. ALF: Saldiva. PHN. (2004a). Air pollution and
neonatal deaths in Sao Paulo, Brazil. Braz J Med Biol Res 37: 765-770. http://dx.doi.org/10.1590/S0100-
879X2004000500019
Lin. CC: Yang. SK: Lin. KG: Ho. WC: Hsieh. WS: Shu. BC: Chen. PC. (2014). Multilevel analysis of air
pollution and early childhood neurobehavioral development. Int J Environ Res Public Health 11: 6827-6841.
http://dx.doi.org/10.3390/iierphll0706827
Lin. M: Chen. Y: Villeneuve. PJ: Burnett RT: Lemyre. L: Hertzman. C: McGrail KM: Krewskj D. (2004b).
Gaseous air pollutants and asthma hospitalization of children with low household income in Vancouver,
British Columbia, Canada. Am J Epidemiol 159: 294-303. http://dx.doi.org/10.1093/aie/kwh043
Lin. M: Fiore. AM: Horowitz. LW: Cooper. OR: Naik. V: Hollowav. J: Johnson. BJ: Middlebrook. AM:
Oltmans. SJ: Pollack. IB: Ryersoa TB: Warner. JX: Wiedinmyer. C: Wilson. J: Wyman. B. (2012).
Transport of Asian ozone pollution into surface air over the western United States in spring. J Geophys Res
117: DOOV07. http://dx.doi.org/10.1029/2011JD016961
Lin. M: Stieb. DM: Chen. Y. (2005). Coarse paniculate matter and hospitalization for respiratory infections in
children younger than 15 years in Toronto: A case-crossover analysis. Pediatrics 116: 235-240.
http://dx.doi.org/10.1542/peds.2004-2012
Lin. W: Huang. W: Zhu. T: Hu. M: Brunekreef. B: Zhang. Y: Liu. X: Cheng. H: Gehring. U: Li. C: Tang. X.
(2011). Acute respiratory inflammation in children and black carbon in ambient air before and during the
2008 Beijing Olympics. Environ Health Perspect 119: 1507-1512. http://dx.doi.org/10.1289/ehp. 1103461
Linaker. CH: Chauhan. AJ: Inskip. HM: Holgate. ST: Coggon. D. (2000). Personal exposures of children to
nitrogen dioxide relative to concentrations in outdoor air. Occup Environ Med 57: 472-476.
R-54
-------
Linares. C: Diaz. J. (2010). Short-term effect of concentrations of fine paniculate matter on hospital admissions
due to cardiovascular and respiratory causes among the over-75 age group in Madrid, Spain. Public Health
124: 28-36. http://dx.doi.0rg/10.1016/i.puhe.2009.ll.007
Lindstrom. J: Szpiro. AA: Sampson. PD: Oron. AP: Richards. M: Larson. TV: Sheppard. L. (2013). A flexible
spatio-temporal model for air pollution with spatial and spatio-temporal covariates. Environ Ecol Stat 21:
411-433. http://dx.doi.org/10.1007/sl0651-013-0261-4
Linhares. D: Jacinto. T: Pereira. AM: Fonseca. JA. (2011). Effects of atopy and rhinitis on exhaled nitric oxide
values - a systematic review. Clin Transl Allergy 1: 8. http://dx.doi.org/10.1186/2045-7022-l-8
Link. MS: Luttmann-Gibson. H: Schwartz. J: Mittleman. MA: Wessler. B: Gold. PR: Dockery. DW: Laden. F.
(2013). Acute exposure to air pollution triggers atrial fibrillation. J Am Coll Cardiol 62: 816-825.
http://dx.doi.0rg/10.1016/i.jacc.2013.05.043
Linn. WS: Shamoo. DA: Anderson. KR: Peng. RC: Avol EL: Hackney. JD: Gong. H. Jr. (1996). Short-term air
pollution exposures and responses in Los Angeles area schoolchildren. J Expo Anal Environ Epidemiol 6:
449-472.
Linn. WS: Shamoo. DA: Spier. CE: Valencia. LM: Anzar. UT: Venet TG: Avol EL: Hackney. JD. (1985a).
Controlled exposure of volunteers with chronic obstructive pulmonary disease to nitrogen dioxide. Arch
Environ Health 40: 313-317. http://dx.doi.org/10.1080/00039896.1985.10545939
Linn. WS: Solomon. JC: Trim. SC: Spier. CE: Shamoo. DA: Venet. TG: Avol. EL: Hackney. JD. (1985b).
Effects of exposure to 4 ppm nitrogen dioxide in healthy and asthmatic volunteers. Arch Environ Occup
Health 40: 234-239.
Linn. WS: Szlachcic. Y: Gong. H. Jr: Kinnev. PL: Berhane. KT. (2000). Air pollution and daily hospital
admissions in metropolitan Los Angeles. Environ Health Perspect 108: 427-434.
Lipfert FW: Batv. JD: Miller. JP: Wyzga. RE. (2006a). PM2.5 constituents and related air quality variables as
predictors of survival in a cohort of U.S. military veterans. Inhal Toxicol 18: 645-657.
http://dx.doi.org/10.1080/08958370600742946
Lipfert. FW: Perry. HM. Jr: Miller. JP: Batv. JD: Wyzga. RE: Carmodv. SE. (2000). The Washington
University-EPRI veterans' cohort mortality study: Preliminary results. Inhal Toxicol 12: 41-73.
http://dx.doi.org/10.1080/713856640
Lipfert. FW: Perry. HM. Jr: Miller. JP: Batv. JD: Wyzga. RE: Carmodv. SE. (2003). Air pollution, blood
pressure, and their long-term associations with mortality. Inhal Toxicol 15: 493-512.
http://dx.doi.org/10.1080/08958370304463
Lipfert. FW: Wyzga. RE: Batv. JD: Miller. JP. (2006b). Traffic density as a surrogate measure of environmental
exposures in studies of air pollution health effects: Long-term mortality in a cohort of US veterans. Atmos
Environ 40: 154-169. http://dx.doi.0rg/10.1016/j.atmosenv.2005.09.027
Lipfert. FW: Wyzga. RE: Batv. JD: Miller. JP. (2009). Air pollution and survival within the Washington
University-EPRI veterans cohort: Risks based on modeled estimates of ambient levels of hazardous and
criteria air pollutants. J Air Waste Manag Assoc 59: 473-489. http://dx.doi.Org/10.3155/1047-3289.59.4.473
Lipsett MJ: Ostro. BD: Reynolds. P: Goldberg. D: Hertz. A: Jerrett M: Smith. DF: Garcia. C: Chang. ET:
Bernstein. L. (2011). Long-term exposure to air pollution and cardiorespiratory disease in the California
teachers study cohort. Am J Respir Crit Care Med 184: 828-835. http://dx.doi.org/10.1164/rccm.201012-
2082OC
Liu. C: Flexeder. C: Fuertes. E: Cyrys. J: Bauer. CP: Koletzko. S: Hoffmann. B: von Berg. A: Heinrich. J:
Groups. ftGaLS. (2014a). Effects of air pollution on exhaled nitric oxide in children: Results from the
GINIplus and LISAphis studies. Int J Hyg Environ Health 217: 483-491.
http://dx.doi.0rg/10.1016/i.iiheh.2013.09.006
R-55
-------
Liu. C: Fuertes. E: Tiesler. CMT: Birk. M: Babisch. W: Bauer. CP: Koletzko. S: von Berg. A: Hoffmann. B:
Heinrich. J. (2014b). The associations between traffic-related air pollution and noise with blood pressure in
children: Results from the GINIplus and LISAplus studies. Int J Hyg Environ Health 217: 499-505.
http://dx.doi.0rg/10.1016/i.iiheh.2013.09.008
Liu. CC: Tsai SS: Chiu. HF: Wu. TN: Chen. CC: Yang. CY. (2009a). Ambient exposure to criteria air pollutants
and risk of death from bladder cancer in Taiwan. Inhal Toxicol 21: 48-54.
http://dx.doi.org/10.1080/08958370802207326
Liu. CC: Tsai. SS: Chiu. HF: Wu. TN: Yang. CY. (2008). Ambient exposure to criteria air pollutants and female
lung cancer in Taiwan. Inhal Toxicol 20: 311-317. http://dx.doi.org/10.1080/08958370701866107
Liu. CH: Leung. DYC: Earth. MC. (2005). On the prediction of air and pollutant exchange rates in street
canyons of different aspect ratios using large-eddy simulation. Atmos Environ 39: 1567-1574.
http://dx.doi.0rg/10.1016/i.atmosenv.2004.08.036
Liu. L. (2013). Email from Dr. Liu to Dr. Patel; Response to data request. Available online at
http://www.re gulations.gov/#!documentDetail:D=EPA-HQ-ORD-2013-0232-0013
Liu. L: Kauri. LM: Mahmud. M: Weichenthal S: Cakmak. S: Shutt R: You. H: Thomson. E: Vincent R:
Kumarathasan. P: Broad. G: Dales. R. (2014c). Exposure to air pollution near a steel plant and effects on
cardiovascular physiology: A randomized crossover study. Int J Hyg Environ Health 217: 279-286.
http://dx.doi.0rg/10.1016/i.iiheh.2013.06.007
Liu. L: Poon. R: Chen. L: Frescura. AM: Montuschi. P: Ciabattonj G: Wheeler. A: Dales. R. (2009b). Acute
effects of air pollution on pulmonary function, airway inflammation, and oxidative stress in asthmatic
children. Environ Health Perspect 117: 668-674. http://dx.doi.org/10.1289/ehpll813
Liu. S: Krewski. D: Shi. Y: Chen. Y: Burnett. R. (2007). Association between maternal exposure to ambient air
pollutants during pregnancy and fetal growth restriction. J Expo Sci Environ Epidemiol 17: 426-432.
http://dx.doi.org/10.1038/si.ies.7500503
Liu. S: Krewski. D: Shi. Y: Chen. Y: Burnett. RT. (2003). Association between gaseous ambient air pollutants
and adverse pregnancy outcomes in Vancouver, Canada. Environ Health Perspect 111: 1773-1778.
Ljungman. P: Bellander. T: Schneider. A: Breitner. S: Forastiere. F: Hampel. R: Illig. T: Jacquemin. B:
Katsouvanni. K: vonKlot S: Koenig. W: Lanki. T: Nyberg. F: Pekkanen. J: Pistelli. R: Pitsavos. C:
Rosenqvist M: Sunyer. J: Peters. A. (2009). Modification of the interleukin-6 response to air pollution by
interleukin-6 and fibrinogen polymorphisms. Environ Health Perspect 117: 1373-1379.
http://dx.doi.org/10.1289/ehp.0800370
Ljungman. PL: Wilker. EH: Rice. MB: Schwartz. J: Gold. PR: Koutrakis. P: Vita. JA: Mitchell GF: Vasan. RS:
Benjamin. EJ: Mittleman. MA: Hamburg. NM. (2014). Short-term exposure to air pollution and digital
vascular function. Am J Epidemiol 180: 482-489. http://dx.doi.org/10.1093/aie/kwul61
Ljungman. PLS: Berglind. N: Holmgren. C: Gadler. F: Edvardsson. N: Pershagen. G: Rosenqvist M: Sjogren. B:
Bellander. T. (2008). Rapid effects of air pollution on ventricular arrhythmias. Eur Heart J 29: 2894-2901.
http://dx.doi.org/10.1093/eurheartj/ehn463
Llop. S: Ballester. F: Estarlich. M: Esplugues. A: Rebagliato. M: Iniguez. C. (2010). Preterm birth and exposure
to air pollutants during pregnancy. Environ Res 110: 778-785. http://dx.doi.0rg/10.1016/i.envres.2010.09.009
Llop. S: Ballester. F: Estarlich. M: Iniguez. C: Ramon. R: Gonzalez. MC: Murcia. M: Esplugues. A: Rebagliato.
M. (2011). Social factors associated with nitrogen dioxide (NO2) exposure during pregnancy: The INMA-
Valencia project in Spain. Soc Sci Med 72: 890-898. http://dx.doi.0rg/10.1016/i.socscimed.2010.12.018
Llorca. J: Salas. A: Prieto-Salceda. D: Chinchon-Bengoechea. V: Delgado-Rodriguez. M. (2005). Nitrogen
dioxide increases cardiorespiratory admissions in Torrelavega (Spain). J Environ Health 68: 30-35.
Loupa. G: Rapsomanikis. S. (2008). Air pollutant emission rates and concentrations in medieval churches. J
Atmos Chem60: 169-187. http://dx.doi.org/10.1007/sl0874-008-9116-z
R-56
-------
Lovich. MA: Bruno. NK: Plant. CP: Wei. AE: Vasauez. GB: Johnson. BJ: Fine. DH: Gilbert. RJ. (2011). Use of
ultra pure nitric oxide generated by the reduction of nitrogen dioxide to reverse pulmonary hypertension in
hypoxemic swine. Nitric Oxide 24: 204-212. http://dx.doi.0rg/10.1016/i.niox.2011.04.006
Lu. C: Deng. Q: Yu. CWF: Sundell J: Ou. C. (2014). Effects of ambient air pollution on the prevalence of
pneumonia in children: Implication for National Ambient Air Quality Standards in China. Indoor Built
Environ 23: 259-269. http://dx.doi.org/10.1177/1420326X13504423
Lu. KD: Brevsse. PN: Diette. GB: Curtin-Brosnan. J: Aloe. C: Williams. PL: Peng. RD: Mccormack. MC:
Matsui EC. (2013). Being overweight increases susceptibility to indoor pollutants among urban children
with asthma. J Allergy Clinlmmunol 131: 1017-1023.el013. http://dx.doi.0rg/10.1016/i.iaci.2012.12.1570
Luginaah. IN: Fung. KY: Gorev. KM: Webster. G: Wills. C. (2005). Association of ambient air pollution with
respiratory hospitalization in a government-designated "area of concern": The case of Windsor, Ontario.
Environ Health Perspect 113: 290-296. http://dx.doi.org/10.1289/ehp.7300
Lundberg. JO: Carlstrom. M: Larsen. FJ: Weitzberg. E. (2011). Roles of dietary inorganic nitrate in
cardiovascular health and disease [Review]. Cardiovasc Res 89: 525-532.
http://dx.doi.org/10.1093/cvr/cvq325
Lundberg. JON: Farkas-Szallasi T: Weitzberg. E: Rinder. J: Lidholm. J: Anggaard. A: Hokfelt. T: Lundberg.
JM: Alving. K. (1995). High nitric oxide production in human paranasal sinuses. NatMed 1: 370-373.
http://dx.doi.org/10.1038/nm0495-370
Lundberg. JON: Settergren. G: Gelinder. S: Lundberg. JM: Alving. K: Weitzberg. E. (1996). Inhalation of
nasally derived nitric oxide modulates pulmonary function in humans. Acta Physiol Scand 158: 343-347.
http://dx.doi.Org/10.1046/i.1365-201X.1996.557321000.x
Maantav. JA: Tu. J: Maroko. AR. (2009). Loose-coupling an air dispersion model and a geographic information
system (GIS) for studying air pollution and asthma in the Bronx, New York City. Int J Environ Health Res
19: 59-79. http://dx.doi.org/10.1080/09603120802392868
Maarsingh. H: Bossenga. BE: Bos. IS: Volders. HH: Zaagsma. J: Meurs. H. (2009). L-arginine deficiency causes
airway hyperresponsiveness after the late asthmatic reaction. Eur Respir J 34: 191-199.
http://dx.doi.org/10.1183/09031936.00105408
Maarsingh. H: Zuidhof. AB: Bos. IS: vanDuin. M: Boucher. JL: Zaagsma. J: Meurs. H. (2008). Arginase
inhibition protects against allergen-induced airway obstruction, hyperresponsiveness, and inflammation. Am
J Respir Crit Care Med 178: 565-573. http://dx.doi.org/10.1164/rccm.200710-1588OC
Mabahraian. U: Ahmad. T: Leishangthem. GD: Dinda. AK: Agrawal. A: Ghosh. B. (2010a). L-arginine reduces
mitochondrial dysfunction and airway injury in murine allergic airway inflammation. Int Immunopharmacol
10: 1514-1519. http://dx.doi.0rg/10.1016/j.intimp.2010.08.025
Mabahraian. U: Ahmad. T: Leishangthem. GD: Joseph. DA: Dinda. AK: Agrawal A: Ghosh. B. (201 Ob).
Beneficial effects of high dose of L-arginine on airway hyperresponsiveness and airway inflammation in a
murine model of asthma. J Allergy Clinlmmunol 125: 626-635. http://dx.doi.0rg/10.1016/i.jaci.2009.10.065
Macintyre. EA: Brauer. M: Melen. E: Bauer. CP: Bauer. M: Berdel. D: Bergstrom. A: Brunekreef B: Chan-
Yeung. M: Klumper. C: Fuertes. E: Gehring. U: Gref A: Heinrich. J: Herbarth. O: Kerkhof M: Koppelman.
GH: Kozyrskyj. AL: Pershagen. G: Postma. PS: Thiering. E: Tiesler. CM: Carlsten. C. (2014a). GSTP1 and
TNF gene variants and associations between air pollution and incident childhood asthma: The traffic, asthma
and genetics (TAG) study. Environ Health Perspect 122: 418-424. http://dx.doi.org/10.1289/ehp. 1307459
Maclntyre. EA: Gehring. U: Molter. A: Fuertes. E: Klumper. C: Kramer. U: Quass. U: Hoffmann. B: Gascon. M:
Brunekreef B: Koppelman. GH: Beelen. R: Hoek. G: Birk. M: de Jongste. JC: Smit HA: Cyrys. J: Gruzieva.
O: Korek. M: Bergstrom A: Agius. RM: de Vocht F: Simpson. A: Porta. D: Forastiere. F: Badaloni. C:
Cesaroni. G: Esplugues. A: Fernandez-Somoano. A: Lerxundi. A: Sunyer. J: Cirach. M: Nieuwenhuijsen.
MJ: Pershagen. G: Heinrich. J. (2014b). Air pollution and respiratory infections during early childhood: an
analysis of 10 European birth cohorts within the ESCAPE Project. Environ Health Perspect 122: 107-113.
http://dx.doi.org/10.1289/ehp.1306755
R-57
-------
Maclntyre. HL: Evans. MJ. (2010). Sensitivity of a global model to the uptake of N2O5 by tropospheric aerosol.
Atmos ChemPhys 10: 7409-7414. http://dx.doi.org/10.5194/acp-10-7409-2010
MacNaughton. P: Melly. S: Vallarino. J: Adamkiewicz. G: Spengler. JD. (2014). Impact of bicycle route type on
exposure to traffic-related air pollution. Sci Total Environ 490: 37-43.
http://dx.doi.0rg/10.1016/i.scitotenv.2014.04.lll
Madei E: Folkes. LK: Wardman. P: Czapskj G: Goldstein. S. (2008). Thiyl radicals react with nitric oxide to
form S-nitrosothiols with rate constants near the diffusion-controlled limit. Free Radic Biol Med 44: 2013-
2018. http://dx.doi.0rg/10.1016/i.freeradbiomed.2008.02.015
Madsen. C: Durand. KL: Nafstad. P: Schwarze. PE: Renningen. KS: Haheim LL. (2008). Associations between
environmental exposures and serum concentrations of Clara cell protein among elderly men in Oslo, Norway.
Environ Res 108: 354-360. http://dx.doi.0rg/10.1016/i.envres.2008.07.017
Madsen. C: Gehring. U: Walker. SE: Brunekreef. B: Stigum. H: Naess. O: Nafstad. P. (2010). Ambient air
pollution exposure, residential mobility and term birth weight in Oslo, Norway. Environ Res 110: 363-371.
http://dx.doi.0rg/10.1016/i.envres.2010.02.005
Maestrelli. P: Canova. C: Scapellato. ML: Visentin. A: Tessari. R: Bartolucci. GB: Simonato. L: Lotti. M.
(2011). Personal exposure to paniculate matter is associated with worse health perception in adult asthma. J
Investig Allergol Clin Immunol 21: 120-128.
Magnussen. H: Reuss. G: Torres. R. (1986). Airway response to methacholine during exercise induced
refractoriness in asthma. Thorax 41: 667-670.
Maher. AR: Milsom. AB: Gunaruwan. P: Abozguia. K: Ahmed. I: Weaver. RA: Thomas. P: Ashrafian. H: Born.
GVR: James. PE: Frenneaux. MP. (2008). Hypoxic modulation of exogenous nitrite-induced vasodilation in
humans. Circulation 117: 670-677. http://dx.doi.org/10.1161/CIRCULATIONAHA.107.719591
Maheswaran. R: Pearson. T: Smeeton. NC: Beevers. SD: Campbell MJ: Wolfe. CD. (2010). Impact of outdoor
air pollution on survival after stroke: population-based cohort study. Stroke 41: 869-877.
http://dx.doi.org/10.1161/STROKEAHA.109.567743
Major. TA: Panmanee. W: Mortensen. JE: Gray. LD: Hoglen. N: Hassett. DJ. (2010). Sodium nitrite-mediated
killing of the major cystic fibrosis pathogens Pseudomonas aeruginosa, Staphylococcus aureus, and
Burkholderia cepacia under anaerobic planktonic and biofilm conditions. Antimicrob Agents Chemother 54:
4671-4677. http://dx.doi.org/10.1128/AAC.00379-10
Malinovschi. A: Janson. C: Holm. L: Nordvall L: Alving. K. (2009). Basal and induced NO formation in the
pharyngo-oral tract influences estimates of alveolar NO levels. J Appl Physiol 106: 513-519.
http://dx.doi.org/10.1152/iapplphysiol.91148.2008
Malm-Erjefalt M: Persson. CG: Erjefalt JS. (2001). Degranulation status of airway tissue eosinophils in mouse
models of allergic airway inflammation. Am J Respir Cell Mol Biol 24: 352-359.
http://dx.doi.0rg/10.1165/aircmb.24.3.4357
Malmqvist E: Jakobsson. K: Tinnerberg. H: Rignell-Hydbom. A: Rylander. L. (2013). Gestational diabetes and
preeclampsia in association with air pollution at levels below current air quality guidelines. Environ Health
Perspect 121: 488-493. http://dx.doi.org/10.1289/ehp. 1205736
Malmqvist. E: Rignell-Hydbom. A: Tinnerberg. H: Bjork. J: Stroh. E: Jakobsson. K: Rittner. R: Rylander. L.
(2011). Maternal exposure to air pollution and birth outcomes. Environ Health Perspect 119: 553558.
http://dx.doi.org/10.1289/ehp. 1002564
Mann. JK: Balmes. JR: Bruckner. TA: Mortimer. KM: Margolis. HG: Pratt B: Hammond. SK: Lurmann. F:
Tager. IB. (2010). Short-term effects of air pollution on wheeze in asthmatic children in Fresno, California.
Environ Health Perspect 118: 1497-1502. http://dx.doi.org/10.1289/ehp.0901292
Mann. JK: Tager. IB: Lurmann. F: Segal M: Quesenberry. CP. Jr: Lugg. MM: Shan. J: VandenEeden. SK.
(2002). Air pollution and hospital admissions for ischemic heart disease in persons with congestive heart
failure or arrhythmia. Environ Health Perspect 110: 1247-1252.
R-58
-------
Mannes. T: Jalaludin. B: Morgan. G: Lincoln. D: Sheppeard. V: Corbett S. (2005). Impact of ambient air
pollution on birth weight in Sydney, Australia. Occup Environ Med 62: 524-530.
http://dx.doi.org/10.1136/oem.2004.014282
Mansfield. CA: Johnson. FR: Van Houtven. GL. (2006). The missing piece: Valuing averting behavior for
children's ozone exposures. Resource Energ Econ 28: 215-228.
http://dx.doi.0rg/10.1016/i.reseneeco.2006.02.002
Mar. TF: Norris. GA: Koenig. JQ: Larson. TV. (2000). Associations between air pollution and mortality in
Phoenix, 1995-1997. Environ Health Perspect 108: 347-353.
Marcon. A: Fracasso. ME: Marchetti P: Doria. D: Girardi P: Guarda. L: Pesce. G: Pironi. V: Ricci P: de
Marco. R. (2014). Outdoor formaldehyde and NO2 exposures and markers of genotoxicity in children living
near chipboard industries. Environ Health Perspect 122: 639-645. http://dx.doi.org/10.1289/ehp.1307259
Maroziene. L: Grazuleviciene. R. (2002). Maternal exposure to low-level air pollution and pregnancy outcomes:
A population-based study. Environ Health 1: 6. http://dx.doi.org/10.1186/1476-069X-l-6
Marr. LC: Harlev. RA. (2002). Modeling the effect of weekday-weekend differences in motor vehicle emissions
on photochemical air pollution in central California. Environ Sci Technol 36: 4099-4106.
http://dx.doi.org/10.1021/es020629x
Marshall HE: Potts. EN: Kelleher. ZT: Stamler. JS: Foster. WM: Auten. RL. (2009). Protection from
lipopolysaccharide-induced lung injury by augmentation of airway S-nitrosothiols. Am J Respir Crit Care
Med 180: 11-18. http://dx.doi.org/10.1164/rccm.200807-1186OC
Marshall HE: Stamler. JS. (2001). Inhibition of NF-kappa B by S-nitrosylation. Biochemistry 40: 1688-1693.
http://dx.doi.org/10.1021/bi002239Y
Marshall JD: Nethery. E: Brauer. M. (2008). Within-urban variability in ambient air pollution: Comparison of
estimation methods. Atmos Environ 42: 1359-1369. http://dx.doi.0rg/10.1016/j.atmosenv.2007.08.012
Martin. RA: Ather. JL: Lundblad. LKA: Suratt BT: Boyson. JE: Budd. RC: Alcorn. JF: Flavell RA: Eisenbarth.
SC: Poynter. ME. (2013). Interleukin-1 receptor and caspase-1 are required for the Thl7 response in NO2-
promoted allergic airway disease. Am J Respir Cell Mol Biol 48: 655-664.
http://dx.doi.org/10.1165/rcmb.2012-0423OC
Martin. RV: Chance. K: Jacob. DJ: Kurosu. TP: Spurr. RJD: Bucsela. E: Gleasoa JF: Palmer. PI: Bey. I: Fiore.
AM: Li. Q: Yantosca. RM: Koelemeijer. RBA. (2002). An improved retrieval of tropospheric nitrogen
dioxide from GOME. J Geophys Res 107: 4437. http://dx.doi.org/10.1029/2001JD001027
Martinez. FD: Wright. AL: Taussig. LM: Holberg. CJ: Halonen. M: Morgan. WJ. (1995). Asthma and wheezing
in the first six years of life. N Engl J Med 332:133-138. http://dx.doi.org/10.1056/NEJM199501193320301
Martins. P. (2013). Email from Dr. Martins to Dr Patel; Response to data request. Martins, P.
Martins. PC: Valente. J: Papoila. AL: Caires. I: Araujo-Martins. J: Mata. P: Lopes. M: Torres. S: Rosado-Pinto.
J: Borrego. C: Annesi-Maesano. I: Neuparth. N. (2012). Airways changes related to air pollution exposure in
wheezing children. Eur Respir J 39: 246-253. http://dx.doi.org/10.1183/09031936.00025111
Maruo. YY: Ogawa. S: Ichino. T: Murao. N: Uchiyama. M. (2003). Measurement of local variations in
atmospheric nitrogen dioxide levels in Sapporo, Japan, using a new method with high spatial and high
temporal resolution. Atmos Environ 37: 1065-1074. http://dx.doi.org/10.1016/81352-2310(02)00974-3
Masiol M: Agostinelli C: Formenton. G: Tarabotti E: Pavoni. B. (2014). Thirteen years of air pollution hourly
monitoring in a large city: Potential sources, trends, cycles and effects of car-free days. Sci Total Environ
494-495: 84-96. http://dx.doi.0rg/10.1016/i.scitotenv.2014.06.122
Massoli. P: Former. EC: Canagaratna. MR: Williams. LR: Zhang. Q: Sun. Y: Schwab. JJ: Trimborn. A: Onasch.
TB: Demerjian. KL: Kolb. CE: Worsnop. PR: Jayne. JT. (2012). Pollution gradients and chemical
characterization ofparticulatematter from vehicular traffic near major roadways: Results from the 2009
Queens College air quality study in NYC. Aerosol Sci Technol 46: 1201-1218.
http://dx.doi.org/10.1080/02786826.2012.701784
R-59
-------
Matalon. S: Shrestha. K: Kirk. M: Waldheuser. S: McDonald. B: Smith. K: Gao. Z: Belaaouaj. A: Crouch. EC.
(2009). Modification of surfactant protein D by reactive oxygen-nitrogen intermediates is accompanied by
loss of aggregating activity, in vitro and in vivo. FASEB J 23: 1415-1430. http://dx.doi.org/10.1096/fj.Q8-
120568
Matsumura. Y. (1970). The effects of ozone, nitrogen dioxide, and sulfur dioxide on the experimentally induced
allergic respiratory disorder in guinea pigs: II. The effects of ozone on the absorption and the retention of
antigen in the lung. Am Rev Respir Dis 102: 438-443.
Matte. TD: Ross. Z: Kheirbek. I: Eisl H: Johnson. S: Gorczynski. JE: Kass. D: Markowitz. S: Pezeshki G:
Cloughertv. JE. (2013). Monitoring intraurban spatial patterns of multiple combustion air pollutants in New
York City: Design and implementation. JExpo Sci Environ Epidemiol 23: 223-231.
http://dx.doi.org/10.1038/ies.2012.126
Mauderlv. JL: Cheng. YS: Gillett. NA: Griffith. WC: Henderson. RF: Pickrell. JA: Ulblff. RK. (1990). Influence
of preexisting pulmonary emphysema on susceptibility of rats to chronic inhalation exposure to nitrogen
dioxide. Inhal Toxicol 2: 129-150. http://dx.doi.org/10.3109/08958379009145250
McAdam. K: Steer. P: Perrotta. K. (2011). Using continuous sampling to examine the distribution of traffic
related air pollution in proximity to a major road. Atmos Environ 45: 2080-2086.
http://dx.doi.0rg/10.1016/i.atmosenv.2011.01.050
McClenny. WA: Williams. EJ: Cohen. RC: Stutz. J. (2002). Preparing to measure the effects of the NOx SIP
call-methods for ambient air monitoring of NO, NO2, NOy, and individual NOz species. J Air Waste Manag
Assoc 52: 542-562. http://dx.doi.org/10.1080/10473289.2002.10470801
McConnell R: Berhane. K: Gilliland. F: London. SJ: Vora. H: Avol E: Gauderman. WJ: Margolis. HG:
Lurmann. F: Thomas. DC: Peters. JM. (1999). Air pollution and bronchitic symptoms in southern California
children with asthma. Environ Health Perspect 107: 757-760.
McConnell. R: Berhane. K: Gilliland. F: Molitor. J: Thomas. D: Lurmann. F: Avol. E: Gauderman. WJ: Peters.
JM. (2003). Prospective study of air pollution and bronchitic symptoms in children with asthma. Am J Respir
Crit Care Med 168: 790-797. http://dx.doi.org/10.1164/rccm.200304-466OC
McConnell R: Berhane. K: Molitor. J: Gilliland. F: Kunzli. N: Thorne. PS: Thomas. D: Gauderman. WJ: Avol
E: Lurmann. F: Rappaport. E: Jerrett M: Peters. J. (2006). Dog ownership enhances symptomatic responses
to air pollution in children with asthma. Environ Health Perspect 114: 1910-1915.
http://dx.doi.org/10.1289/ehp.8548
McConnell R: Islam. T: Shankardass. K: Jerrett M: Lurmann. F: Gilliland. F: Gauderman. J: Avol E: Kunzli.
N: Yao. L: Peters. J: Berhane. K. (2010a). Childhood incident asthma and traffic-related air pollution at home
and school. Environ Health Perspect 118: 1021-1026. http://dx.doi.org/10.1289/ehp.0901232
McConnell. R: Liu. FF: Wu. J: Lurmann. F: Peters. J: Berhane. K. (201 Ob). Asthma and school commuting time.
J Occup Environ Med 52: 827-829. http://dx.doi.org/10.1097/JOM.Ob013e3181ebfla9
McCreanor. J: Cullinan. P: Nieuwenhuijsen. MJ: Stewart-Evans. J: Malliarou. E: Jarup. L: Harrington. R:
Svartengren. M: Han. IK: Ohman-Strickland. P: Chung. KF: Zhang. J. (2007). Respiratory effects of
exposure to diesel traffic in persons with asthma. N Engl J Med 357: 2348-2358.
http://dx.doi.org/10.1056/NEJMoa071535
McCurdv. T: Glen. G: Smith. L: Lakkadi. Y. (2000). The National Exposure Research Laboratory's consolidated
human activity database. J Expo Anal Environ Epidemiol 10: 566-578.
http://dx.doi.org/10.1038/si.iea.7500114
McDonald. BC: Dallmann. TR: Martin. EW: Harlev. RA. (2012). Long-term trends in nitrogen oxide emissions
from motor vehicles at national, state, and air basin scales. J Geophys Res 117: DOOV18.
http://dx.doi.org/10.1029/2012JD018304
McGrath. JJ: Smith. PL. (1984). Respiratory responses to nitrogen dioxide inhalation: Subtitle: Nitrogen
dioxide. J Environ Sci Health, Part A: Environ Sci Eng 19: 417-431.
http://dx.doi.org/10.1080/10934528409375168
R-60
-------
McMeeking. GR: Kreidenweis. SM: Baker. S: Carrico. CM: Chow. JC: Collett. JL. Jr: Hao. WM: Holden. AS:
Kirchstetter. TW: Malm. WC: Moosmiiller. H: Sullivan. AP: Wold. CE. (2009). Emissions of trace gases and
aerosols during the open combustion of biomass in the laboratory. J Geophys Res Atmos 114: D19210.
http://dx.doi.org/10.1029/2009id011836
Mebust AK: Russell AR: Hudman. RC: Valin. LC: Cohen. RC. (2011). Characterization of wildfire NOx
emissions using MODIS fire radiative power and OMI tropospheric NO2 columns. Atmos Chem Phys 11:
5839-5851. http://dx.doi.org/10.5194/acp-ll-5839-2011
Mehta. S: Ngo. LH: Dzung. DV: Cohen. A: Thach. TQ: Dan. VX: Tuan. ND: Giang. LT. (2013). Air pollution
and admissions for acute lower respiratory infections in young children of Ho Chi Minh City. Air Qual
Atmos Health 6: 167-179. http://dx.doi.org/10.1007/sll869-011-0158-z
Melen. E: Nyberg. F: Lindgren. CM: Berglind. N: Zucchelli. M: Nordling. E: Hallberg. J: Svartengren. M:
Morgenstern. R: Kere. J: Bellander. T: Wickman. M: Pershagen. G. (2008). Interactions between glutathione
S-transferase PI, tumor necrosis factor, and traffic-related air pollution for development of childhood allergic
disease. Environ Health Perspect 116: 1077-1084. http://dx.doi.org/10.1289/ehp. 11117
Meng. Q: Williams. R: Pinto. JP. (2012a). Determinants of the associations between ambient concentrations and
personal exposures to ambient PM2.5, NO2, and O3 during DEARS. Atmos Environ 63: 109-116.
http://dx.doi.0rg/10.1016/i.atmosenv.2012.09.019
Meng. QY: Svendsgaard. D: Kotchmar. DJ: Pinto. JP. (2012b). Associations between personal exposures and
ambient concentrations of nitrogen dioxide: A quantitative research synthesis. Atmos Environ 57: 322-329.
http://dx.doi.0rg/10.1016/i.atmosenv.2012.04.035
Meng. X: Wang. C: Cao. D: Wong. CM: Kan. H. (2013). Short-term effect of ambient air pollution on COPD
mortality in four Chinese cities. Atmos Environ 77: 149-154.
http://dx.doi.0rg/10.1016/i.atmosenv.2013.05.001
Meng. YY: Rull. RP: Wilhelm. M: Lombardi C: Balmes. J: Ritz. B. (2010). Outdoor air pollution and
uncontrolled asthma in the San Joaquin Valley, California. J Epidemiol Community Health 64: 142-147.
http://dx.doi.org/10.1136/iech.2008.083576
Mercer. LD: Szpiro. AA: Sheppard. L: Lindstrom. J: Adar. SD: Allen. RW: Avol EL: Oron. AP: Larson. T: Liu.
LJ: Kaufman. JD. (2011). Comparing universal kriging and land-use regression for predicting concentrations
of gaseous oxides of nitrogen (NOx) for the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA
Air). Atmos Environ 45: 4412-4420. http://dx.doi.0rg/10.1016/j.atmosenv.2011.05.043
Mercer. RR: Costa. PL: Crapo. JD. (1995). Effects of prolonged exposure to low doses of nitric oxide or
nitrogen dioxide on the alveolar septa of the adult rat lung. Lab Invest 73: 20-28.
Mercer. RR: Russell. ML: Roggli. VL: Crapo. JD. (1994). Cell number and distribution in human and rat
airways. Am J Respir Cell Mol Biol 10: 613-624.
Mersch. J: Dyce. BJ: Haverback. BJ: Sherwin. RP. (1973). Diphosphoglycerate content of red blood cells:
Measurements in guinea pigs exposed to 04 ppm nitrogen dioxide. Arch Environ Health 27: 94-95.
http://dx.doi.org/10.1080/00039896.1973.10666326
Metzger. KB: Klein. M: Flanders. WD: Peel JL: Mulholland. JA: Langberg. JJ: Tolbert. PE. (2007). Ambient air
pollution and cardiac arrhythmias in patients with implantable defibrillators. Epidemiology 18: 585-592.
http://dx.doi.org/10.1097/EDE.Ob013e318124ffOe
Mi. YH: Norback. D: Tao. J: Mi. YL: Perm. M. (2006). Current asthma and respiratory symptoms among pupils
in Shanghai, China: Influence of building ventilation, nitrogen dioxide, ozone, and formaldehyde in
classrooms. Indoor Air 16: 454-464. http://dx.doi.org/10.1111/i. 1600-0668.2006.00439.X
Migriaretti. G: Cadum. E: Migliore. E: Cavallo. F. (2005). Traffic air pollution and hospital admission for
asthma: a case-control approach in a Turin (Italy) population. Int Arch Occup Environ Health 78: 164-169.
http://dx.doi.org/10.1007/s00420-004-0569-3
Miller. FJ: Graham. JA: Illing. JW: Gardner. DE. (1980). Extrapulmonary effects of NO2 as reflected by
pentobarbital-induced sleeping time in mice. Toxicol Lett 6: 267-274.
R-61
-------
Miller. FJ: Graham. JA: Raub. JA: Illing. JW: Menache. MG: House. DE: Gardner. DE. (1987). Evaluating the
toxicity of urban patterns of oxidant gases. II. Effects in mice from chronic exposure to nitrogen dioxide. J
Toxicol Environ Health 21: 99-112. http://dx.doi.org/10.1080/15287398709531005
Miller. FJ: Overton. JH: Gerritv. TR: Graham. RC. (1988). Interspecies dosimetry of reactive gases. InU Mohr;
D Dungworth; R McClellan; G Kimmerle; W Stober; J Lewkowski (Eds.), Inhalation toxicology: The design
and interpretation of inhalation studies and their use in risk assessment (pp. 139-155). New York, NY:
Springer-Verlag.
Miller. FJ: Overton. JH. Jr: Jaskot RH: Menzel DB. (1985). A model of the regional uptake of gaseous
pollutants in the lung: I. The sensitivity of the uptake of ozone in the human lung to lower respiratory tract
secretions and exercise. Toxicol Appl Pharmacol 79: 11-27. http://dx.doi.org/10.1016/0041-008X(85)90364-
o
J_
Miller. FJ: Overton. JH: Myers. ET: Graham. JA. (1982). Pulmonary dosimetry of nitrogen dioxide in animals
and man. In T Schneider; L Grant (Eds.), Air pollution by nitrogen oxides (pp. 377-386). Amsterdam:
Elsevier Scientific Publishing Company.
Miller. KA: Siscovick. PS: Sheppard. L: Shepherd. K: Sullivan. JH: Anderson. GL: Kaufman. JD. (2007). Long-
term exposure to air pollution and incidence of cardiovascular events in women. N Engl J Med 356: 447-458.
http://dx.doi.org/10.1056/NEJMoa054409
Min. KB: Min. JY: Cho. SI: Paek. D. (2008). The relationship between air pollutants and heart-rate variability
among community residents in Korea. Inhal Toxicol 20: 435-444.
http://dx.doi.org/10.1080/08958370801903834
Minguillon. MC: Rivas. I: Aguilera. I: Alastuev. A: Moreno. T: Amato. F: Sunyer. J: Querol. X. (2012). Within-
city contrasts in PM composition and sources and their relationship with nitrogen oxides. J Environ Monit
14: 2718-2728. http://dx.doi.org/10.1039/c2em30469d
Misra. A: Roorda. MJ: Maclean. HL. (2013). An integrated modelling approach to estimate urban traffic
emissions. Atmos Environ73: 81-91. http://dx.doi.0rg/10.1016/i.atmosenv.2013.03.013
Mitchell G. (2005). Forecasting environmental equity: Air quality responses to road user charging in Leeds, UK.
J Environ Manage 77: 212-226. http://dx.doi.0rg/10.1016/i.jenvman.2005.04.013
Miyanishj K: Kinouchj T: Kataoka. K: Kanoh. T: Ohnishi. Y. (1996). In vivo formation of mutagens by
intraperitoneal administration of polycyclic aromatic hydrocarbons in animals during exposure to nitrogen
dioxide. Carcinogenesis 17: 1483-1490.
Mobasher. Z: Salam. MT: Goodwin. TM: Lurmann. F: Ingles. SA: Wilson. ML. (2013). Associations between
ambient air pollution and hypertensive disorders of pregnancy. Environ Res 123:9-16.
http://dx.doi.0rg/10.1016/i.envres.2013.01.006
Mochitate. K: Miura. T. (1984). In vivo effect of nitrogen dioxide on the activities of glycolytic enzymes in red
blood cells of rats. Toxicol Lett 22: 315-321.
Mochitate. K: Takahashj Y: Ohsumj T: Miura. T. (1986). Activation and increment of alveolar macrophages
induced by nitrogen dioxide. J Toxicol Environ Health 17: 229-239.
http://dx.doi.org/10.1080/15287398609530819
Modig. L: Toren. K: Janson. C: Jarvholm. B: Forsberg. B. (2009). Vehicle exhaust outside the home and onset of
asthma among adults. Eur Respir J 33: 1261-1267. http://dx.doi.org/10.1183/09031936.00101108
Mohsenin. V. (1987a). Airway responses to nitrogen dioxide in asthmatic subjects. J Toxicol Environ Health 22:
371-380. http://dx.doi.org/10.1080/15287398709531080
Mohsenin. V. (1987b). Effect of vitamin C on NO2-induced airway hyperresponsiveness in normal subjects: A
randomized double-blind experiment. Am Rev Respir Dis 136: 1408-1411.
http://dx.doi.0rg/10.1164/airccrn/136.6.1408
Mohsenin. V. (1988). Airway responses to 2.0 ppm nitrogen dioxide in normal subjects. Arch Environ Health
43: 242-246. http://dx.doi.org/10.1080/00039896.1988.9934941
R-62
-------
Mohsenin. V. (1991). Lipid peroxidation and antielastase activity in the lung under oxidant stress: Role of
antioxidant defenses. J Appl Physiol 70: 1456-1462.
Molitor. J: Su. JG: Molitor. NT: Rubio. VG: Richardson. S: Hastie. D: Morello-Frosch. R: Jerrett. M. (2011).
Identifying vulnerable populations through an examination of the association between multipollutant profiles
and poverty. Environ Sci Technol 45: 7754-7760. http://dx.doi.org/10.1021/esl04017x
Mollner. AK: Valluvadasaa S: Feng. L: Sprague. MK: Okumura. M: Milligan. DB: Bloss. WJ: Sander. SP:
Martien. PT: Harlev. RA: McCoy. AB: Carter. WRL. (2010). Rate of gas phase association of hydroxyl
radical and nitrogen dioxide. Science 330: 646-649. http://dx.doi.org/10.1126/science. 1193030
Molloy. SB: Cheng. M: Galbally. IE: Kevwood. MD: Lawson. SJ: Powell JC: Gillett R: Dunne. E: Selleck.
PW. (2012). Indoor air quality in typical temperate zone Australian dwellings. Atmos Environ 54: 400-407.
http://dx.doi.0rg/10.1016/i.atmosenv.2012.02.031
Molter. A: Agius. RM: de Vocht F: Lindlev. S: Gerrard. W: Lowe. L: Belgrave. D: Custovic. A: Simpson. A.
(2013). Long-term exposure to PM10 and NO2 in association with lung volume and airway resistance in the
MAAS birth cohort. Environ Health Perspect 121: 1232-1238. http://dx.doi.org/10.1289/ehp.1205961
Molter. A: Lindlev. S: de Vocht. F: Agius. R: Kerry. G: Johnson. K: Ashmore. M: Terry. A: Dimitroulopoulou.
S: Simpson. A. (2012). Performance of a microenviromental model for estimating personal NO2 exposure in
children. Atmos Environ 51: 225-233. http://dx.doi.0rg/10.1016/i.atmosenv.2012.01.030
Molter. A: Lindlev. S: de Vocht. F: Simpson. A: Agius. R. (2010a). Modelling air pollution for epidemiologic
research-Part I: A novel approach combining land use regression and air dispersion. Sci Total Environ 408:
5862-5869. http://dx.doi.0rg/10.1016/i.scitotenv.2010.08.027
Molter. A: Lindlev. S: de Vocht. F: Simpson. A: Agius. R. (2010b). Modelling air pollution for epidemiologic
research-Part II: Predicting temporal variation through land use regression. Sci Total Environ 409: 211-217.
http://dx.doi.0rg/10.1016/i.scitotenv.2010.10.005
Monn. C: Carabias. V: Junker. M: Waeber. R: Karrer. M: Wanner. HU. (1997). Small-scale spatial variability of
paniculate matter <10 mu m (PM10) and nitrogen dioxide. Atmos Environ 31: 2243-2247.
Moolgavkar. SH. (2003). Air pollution and daily deaths and hospital admissions in Los Angeles and Cook
counties. In Revised Analyses of the National Morbidity, Mortality, and Air Pollution Study, Part II (pp. 183-
198). Boston, MA: Health Effects Institute.
Moolgavkar. SH: McClellan. RO: Dewanji. A: Turim. J: Luebeck. EG: Edwards. M. (2013). Time-series
analyses of air pollution and mortality in the United States: A subsampling approach. Environ Health
Perspect 121: 73-78. http://dx.doi.org/10.1289/ehp. 1104507
Moon. JS: Kim. YS: Kim. JH: Son. BS: Kim. PS: Yang. W. (2009). Respiratory health effects among
schoolchildren and their relationship to air pollutants in Korea. Int J Environ Health Res 19: 31-48.
http://dx.doi.org/10.1080/09603120802272201
Morales. E: Julvez. J: Torrent M: de Cid. R: Guxens. M: Bustamante. M: Kiinzli. N: Sunyer. J. (2009).
Association of early-life exposure to household gas appliances and indoor nitrogen dioxide with cognition
and attention behavior in preschoolers. Am JEpidemiol 169: 1327-1336.
http://dx.doi.org/10.1093/aie/kwp067
Morello-Frosch. R: Jesdale. BM: Sadd. JL: Pastor. M. (2010). Ambient air pollution exposure and full-term birth
weight in California. Environ Health 9: 44. http://dx.doi.org/10.1186/1476-069X-9-44
Morgan. G: Corbett S: Wlodarczyk. J. (1998). Air pollution and hospital admissions in Sydney, Australia, 1990
to 1994. Am J Public Health 88: 1761-1766. http://dx.doi.org/10.2105/AJPH.88.12.1761
Morgenstern. V: Zutavern. A: Cyrys. J: Brockow. I: Gehring. U: Koletzko. S: Bauer. CP: Reinhardt D:
Wichmann. HE: Heinrich. J. (2007). Respiratory health and individual estimated exposure to traffic-related
air pollutants in a cohort of young children. Occup Environ Med 64: 8-16.
http://dx.doi.org/10.1136/oem.2006.028241
R-63
-------
Morgenstern. V: Zutavern. A: Cyrys. J: Brockow. I: Koletzko. S: Kramer. U: Behrendt H: Herbarth. O: von
Berg. A: Bauer. CP: Wichmann. HE: Heinrich. J. (2008). Atopic diseases, allergic sensitization, and
exposure to traffic-related air pollution in children. Am J Respir Crit Care Med 177: 1331-1337.
http://dx.doi.org/10.1164/rccm.200701-036OC
Moridi M: Ziaei S: Kazemnejad. A. (2014). Exposure to ambient air pollutants and spontaneous abortion. J
Obstet Gynaecol Res 40: 743-748. http://dx.doi.org/10.1111/iog. 12231
Morrow. PE: Utell MJ. (1989a). Responses of susceptible subpopulations to nitrogen dioxide [HEI] (pp. 1-45).
(ISSN 1041-5505 Research Report 23). Cambridge, MA: Health Effects Institute.
http://pubs.healtheffects.org/view.php?id=235
Morrow. PE: Utell MJ. (1989b). Responses of susceptible subpopulations to nitrogen dioxide: Appendix H.
Selected raw data [HEI]. (Research Report 23). Cambridge, MA: Health Effects Institute.
Morrow. PE: Utell MJ: Bauer. MA: Smeglin. AM: Frampton. MW: Cox. C: Speers. DM: Gibb. FR. (1992).
Pulmonary performance of elderly normal subjects and subjects with chronic obstructive pulmonary disease
exposed to 0.3 ppm nitrogen dioxide. Am J Respir Crit Care Med 145: 291-300.
http://dx.doi.0rg/10.1164/airccrn/145.2 Pt 1.291
Mortimer. KM: Neas. LM: Dockery. DW: Redline. S: Tager. IB. (2002). The effect of air pollution on inner-city
children with asthma. Eur Respir J 19: 699-705. http://dx.doi.org/10.1183/09031936.02.00247102
Moseler. M: Hendel-Kramer. A: Karmaus. W: Forster. J: Weiss. K: Urbanek. R: Kuehr. J. (1994). Effect of
moderate NO2 air pollution on the lung function of children with asthmatic symptoms. Environ Res 67: 109-
124. http://dx.doi.org/10.1006/enrs.1994.1068
Moshammer. H: Hutter. HP: Hauck. H: Neuberger. M. (2006). Low levels of air pollution induce changes of
lung function in a panel of schoolchildren. Eur Respir J27: 1138-1143.
http://dx.doi.org/10.1183/09031936.06.00089605
Muhajarine. N: Mustard. C: Roos. LL: Young. TK: Gelskev. DE. (1997). Comparison of survey and physician
claims data for detecting hypertension. J Clin Epidemiol 50: 711-718. http://dx.doi.org/10.1016/S0895-
4356(97)00019-X
Mukala. K: Aim. S: Tiittanen. P: Salonen. RO: Jantunen. M: Pekkanen. J. (2000). Nitrogen dioxide exposure
assessment and cough among preschool children. Arch Environ Health 55: 431-438.
Mukerjee. S: Smith. L: Neas. L: Norris. G. (2012a). Evaluation of land use regression models for nitrogen
dioxide and benzene inFour US cities. ScientificWorldJournal 2012: 865150.
http://dx.doi.org/10.1100/2012/865150
Mukerjee. S: Smith. LA: Johnson. MM: Neas. LM: Stallings. CA. (2009). Spatial analysis and land use
regression of VOCs and NO(2) from school-based urban air monitoring in Detroit/Dearborn, USA. Sci Total
Environ407: 4642-4651. http://dx.doi.0rg/10.1016/i.scitotenv.2009.04.030
Mukerjee. S: Willis. RD: Walker. JT: Hammond. D: Norris. GA: Smith. LA: Welch. DP: Peters. TM. (2012b).
Seasonal effects in land use regression models for nitrogen dioxide, coarse paniculate matter, and gaseous
ammonia in Cleveland, Ohio. Atmos Pollut Res 3: 352-361. http://dx.doi.org/10.5094/APR.2012.039
Mulholland. JA. (2013). Email from Dr. Mulholland to Dr Richmond-Bryant; Response to Data Request.
Available online at http://www.regulations.gov/#!documentDetail:D=EPA-HQ-ORD-2013-0232-0020
Miiller. B: Schafer. H: Earth. P: von Wichert. P. (1994). Lung surfactant components inbronchoalveolar lavage
after inhalation of NO2 as markers of altered surfactant metabolism. Lung 172: 61-72.
http://dx.doi.org/10.1007/BF00185077
Murgia. N: Brisman. J: Claesson. A: Muzj G: Olin. AC: Toren. K. (2014). Validity of a questionnaire-based
diagnosis of chronic obstructive pulmonary disease in a general population-based study. BMC Pulm Med 14:
49. http://dx.doi.org/10.1186/1471-2466-14-49
R-64
-------
Murphy. JG: Day. DA: Clearv. PA: Wooldridge. PJ: Millet. DB: Goldsteia AH: Cohen. RC. (2006). The
weekend effect within and downwind of Sacramento: Part 2. Observational evidence for chemical and
dynamical contributions. Atmos Chem Phys Discuss 6: 11971-12019. http://dx.doi.0rg/10.5194/acpd-6-
11971-2006
Murphy. JG: Day. DA: Clearv. PA: Wooldridge. PJ: Millet DB: Goldsteia AH: Cohen. RC. (2007). The
weekend effect within and downwind of Sacramento - Part 1: Observations of ozone, nitrogen oxides, and
VOC reactivity. Atmos Chem Phys 7: 5327-5339.
Murphy. SD: Ulrich. CE: Frankowitz. SH: Xintaras. C. (1964). Altered function in animals inhaling low
concentrations of ozone and nitrogen dioxide. Am Ind Hyg Assoc J 25: 246-253.
http://dx.doi.org/10.1080/00028896409342583
Murray. AB: Ferguson. AC: Morrison. B. (1981). Airway responsiveness to histamine as a test for overall
severity of asthma in children. J Allergy Clin Immunol 68: 119-124. http://dx.doi.org/10.1016/0091-
6749(81)90169-X
Mustafa. MG: Elsaved. NM: von Dohlen. FM: Hassett CM: Postlethwait EM: Quinn. CL: Graham. JA:
Gardner. DE. (1984). A comparison of biochemical effects of nitrogen dioxide, ozone, and their combination
in mouse lung. Toxicol Appl Pharmacol 72: 82-90.
NAEPP (National Asthma Education and Prevention Program). (2007). Expert panel report 3 (EPR-3):
Guidelines for the diagnosis and management of asthmasummary report 2007. J Allergy Clin Immunol 120:
S94-S138. http://dx.doi.0rg/10.1016/i.iaci.2007.09.029
Naess. O: Nafstad. P: Aamodt. G: Claussen. B: Rosland. P. (2007). Relation between concentration of air
pollution and cause-specific mortality: four-year exposures to nitrogen dioxide and paniculate matter
pollutants in 470 neighborhoods in Oslo, Norway. Am J Epidemiol 165: 435-443.
http://dx.doi.org/10.1093/aie/kwk016
Nafstad. P: Haheim. LL: Oftedal. B: Gram F: Holme. I: Hjermann. I: Leren. P. (2003). Lung cancer and air
pollution: A 27 year follow up of 16 209 Norwegian men. Thorax 58: 1071-1076.
http://dx.doi.org/10.1136/thorax.58.12.1071
Nafstad. P: Haheim. LL: Wisloff. T: Gram. F; Oftedal B: Holme. I: Hjermann. I: Leren. P. (2004). Urban air
pollution and mortality in a cohort of Norwegian men. Environ Health Perspect 112: 610-615.
Nahidi. F; Gholami. R: Rashidi. Y: Maid. HA. (2014). Relationship between air pollution and pre-eclampsia in
pregnant women: a case-control study. East Mediterr Health J 19: S60-S66.
Nakajima. T: Kusumoto. S. (1968). Effect of nitrogen dioxide exposure on the contents of reduced glutathione in
mouse lung. Food Sanitation Research (English) 6: 17-21.
Namdeo. A: Stringer. C. (2008). Investigating the relationship between air pollution, health and social
deprivation in Leeds, UK. Environ Int 34: 585-591. http://dx.doi.0rg/10.1016/j.envint.2007.12.015
NAPAP (National Acid Precipitation Assessment Program). (1991). The experience and legacy of NAPAP:
Report of the Oversight Review Board of the National Acid Precipitation Assessment Program. Washington,
DC.
Neta. P: Huie. RE: Ross. AB. (1988). Rate constants for reactions of inorganic radicals in aqueous solution. J
Phys ChemRef Data 17: 1027-1284. http://dx.doi.0rg/10.1063/l.555808
Nettesheim. P: Hanna. MG. Jr: Dohertv. DG: Newell. RF: Hellman. A. (1970). Effect of chronic exposure to air
pollutants on the respiratory tracts of mice: Histopathological findings. In P Nettesheim; MG Hanna, Jr.; JW
Deatherage, Jr. (Eds.), Morphology of experimental respiratory carcinogenesis: proceedings of a biology
division, Oak Ridge National Laboratory conference; May; Gatlinburg, TN. Oak Ridge, TN: U.S. Atomic
Energy Commission.
Neupane. B; Jerrett M; Burnett. RT; Marrie. T; Arain. A; Loeb. M. (2010). Long-term exposure to ambient air
pollution and risk of hospitalization with community-acquired pneumonia in older adults. Am J Respir Crit
Care Med 181: 47-53. http://dx.doi.org/10.1164/rccm.200901-0160OC
R-65
-------
Ng. AW: Bidanj A: Heming. TA. (2004). Innate host defense of the lung: effects of lung-lining fluid pH
[Review]. Lung 182: 297-317. http://dx.doi.org/10.1007/s00408-004-2511-6
NHLBI (National Institutes of Health, National Heart Lung and Blood Institute). (2002). Global initiative for
asthma: Global strategy for asthma management and prevention. U.S. Department of Health and Human
Services. http://www.ginasthma.org/Guidelines/guidelines-2002-original%3a-workshop-report%2c-global-
strategy-for-asthmamanagement-and-prevention. html
Nicolai T: Carr. D: Weiland. SK: Duhme. H: Von Ehrenstein. O: Wagner. C: VonMutius. E. (2003). Urban
traffic and pollutant exposure related to respiratory outcomes and atopy in a large sample of children. Eur
Respir 121:956-963.
Nieuwenhuijsen. MJ: Basagana. X: Dadvand. P: Martinez. D: Cirach. M: Beelen. R: Jacquemin. B. (2014). Air
pollution and human fertility rates. Environ Int 70: 9-14. http://dx.doi.0rg/10.1016/i.envint.2014.05.005
Nishikawa. Y: Murano. K: Mukai. H. (2009). Comparison of sampling resistance for one to three sheets of
membrane type passive sampler. Water Air Soil Pollut 197: 241-247. http://dx.doi.org/10.1007/sll270-008-
9807-8
Nishimura. KK: Galanter. JM: Roth. LA: Oh. SS: Thakur. N: Nguyen. EA: Thyne. S: Farber. HJ: Serebrisky. D:
Kumar. R: Brigino-Buenaventura. E: Davis. A: LeNoir. MA: Meade. K: Rodriguez-Cintron. W: Avila. PC:
Borrell LN: Bibbins-Domingo. K: Rodriguez-Santana. JR: Sea S: Lurmann. F: Balmes. JR: Burchard. EG.
(2013). Early-life air pollution and asthma risk in minority children: The GALA II and SAGE II studies. Am
J Respir Crit Care Med 188: 309-318. http://dx.doi.org/10.1164/rccm.201302-0264OC
Nishino. N: Atkinson. R: Arev. J. (2008). Formation of nitro products from the gas-phase OH radical-initiated
reactions of toluene, naphthalene, and biphenyl: effect of NO2 concentration. Environ Sci Technol 42: 9203-
9209. http://dx.doi.org/10.1021/es802046m
Nitta. H: Sato. T: Nakai. S: Maeda. K: Aoki. S: Ono. M. (1993). Respiratory health associated with exposure to
automobile exhaust: I. Results of cross-sectional studies in 1979, 1982, and 1983. Arch Environ Health 48:
53-58. http://dx.doi.org/10.1080/00039896.1993.9938393
Noda. N: Takeno. S: Fukuiri. T: Hirakawa. K. (2012). Monitoring of oral and nasal exhaled nitric oxide in
eosinophilic chronic rhino sinusitis: A prospective study. Am J Rhinol Allergy 26: 255-259.
http://dx.doi.org/10.2500/aira.2012.26.3772
Nejgaard. JK. (2010). Indoor measurements of the sum of the nitrate radical, NO3, and nitrogen pentoxide,
N2O5 in Denmark. Chemosphere 79: 898-904. http://dx.doi.0rg/10.1016/i.chemosphere.2010.02.025
Nojgaard. JK: Bilde. M: Stenbv. C: Nielsen. OJ: Wolkoff. P. (2006). The effect of nitrogen dioxide on particle
formation during ozonolysis of two abundant monoterpenes indoors. Atmos Environ 40: 1030-1042.
http://dx.doi.0rg/10.1016/i.atmosenv.2005.ll.029
Nordling. E: Berglind. N: Melen. E: Emenius. G: Hallberg. J: Nyberg. F: Pershagen. G: Svartengren. M:
Wickman. M: Bellander. T. (2008). Traffic-related air pollution and childhood respiratory symptoms,
function and allergies. Epidemiology 19: 401-408. http://dx.doi.org/10.1097/EDE.Ob013e31816alce3
North. ML: Khanna. N: Marsden. PA: Grasemann. H: Scott JA. (2009). Functionally important role for arginase
1 in the airway hyperresponsiveness of asthma. Am J Physiol Lung Cell Mol Physiol 296: L911-L920.
http://dx.doi.org/10.1152/aiplung.00025.2009
Novotny. EV: Bechle. MJ: Millet DB: Marshall JD. (2011). National satellite-based land-use regression: NO2
in the United States. Environ Sci Technol 45: 4407-4414. http://dx.doi.org/10.1021/esl03578x
Nunnermacker. LJ: Imre. D: Daum. PH: Kleinman. L: Lee. YN: Lee. JH: Springston. SR: Newman. L:
Weinstein-Llovd. J: Luke. WT: Banta. R: Alvarez. R: Senff. C: Sillman. S: Holdren. M: Keiglev. GW: Zhou.
X (1998). Characterization of the Nashville urban plume on July 3 and July 18, 1995. J Geophys Res 103:
28129-28148.
R-66
-------
Nuvolone. D: Balzi. D: Chini. M: Scala. D: Giovannini. F: BarchieHi. A. (2011). Short-term association between
ambient air pollution and risk of hospitalization for acute myocardial infarction: Results of the cardiovascular
risk and air pollution in Tuscany (RISCAT) study. Am J Epidemiol 174: 63-71.
http://dx.doi.org/10.1093/aie/kwr046
Nyberg. F: Gustavsson. P: Jarup. L: Bellander. T: Berglind. N: Jakobsson. R: Pershagen. G. (2000). Urban air
pollution and lung cancer in Stockholm. Epidemiology 11: 487-495.
O'Byrne. PM: Gauvreau. GM: Brannan. JD. (2009). Provoked models of asthma: What have we learnt?
[Review]. ClinExp Allergy 39: 181-192. http://dx.doi.org/10.1111/i.1365-2222.2008.03172.x
O'Connor. GT: Neas. L: Vaughn. B: Rattan. M: Mitchell H: Grain. EF: Evans. R. Ill: Gruchalla. R: Morgan. W:
Stout J: Adams. GK: Lippmann. M. (2008). Acute respiratory health effects of air pollution on children with
asthma in US inner cities. J Allergy Clin Immunol 121: 1133-1139.el 131.
http://dx.doi.0rg/10.1016/i.iaci.2008.02.020
O'Donnell VB: Eiserich. JP: Chumlev. PH: Jablonsky. MJ: Krishna. NR: Kirk. M: Barnes. S: Darlev-Usmar.
VM: Freeman. BA. (1999). Nitration of unsaturated fatty acids by nitric oxide-derived reactive nitrogen
species peroxynitrite, nitrous acid, nitrogen dioxide, and nitronium ion. Chem Res Toxicol 12: 83-92.
http://dx.doi.org/10.1021/tx980207u
O'Neill MS: Jerrett M: Kawachi. I: Lew. JI: Cohen. AJ: Gouveia. N: Wilkinson. P: Fletcher. T: Cifuentes. L:
Schwartz. J. (2003). Health, wealth, and air pollution: Advancing theory and methods [Review]. Environ
HealthPerspect 111: 1861-1870.
Oda. H: Tsubone. H: Suzuki. A: Ichinose. T: Kubota. K. (1981). Alterations of nitrite and nitrate concentrations
in the blood of mice exposed to nitrogen dioxide. Environ Res 25: 294-301.
Odajima. H: Yamazaki. S: Nitta. H. (2008). Decline in peak expiratory flow according to hourly short-term
concentration of paniculate matter in asthmatic children. Inhal Toxicol 20: 1263-1272.
http://dx.doi.org/10.1080/08958370802311151
Oftedal B: Brunekreef. B: Nystad. W: Madsen. C: Walker. SE: Nafstad. P. (2008). Residential outdoor air
pollution and lung function in schoolchildren. Epidemiology 19: 129-137.
http://dx.doi.org/10.1097/EDE.Ob013e31815c0827
Oftedal. B: Nystad. W: Brunekreef. B: Nafstad. P. (2009a). Long-term traffic-related exposures and asthma
onset in schoolchildren in Oslo, Norway. Environ Health Perspect 117: 839-844.
http://dx.doi.org/10.1289/ehp.11491
Oftedal B: Walker. S: Gram. F: Mclnnes. H: Nafstad. P. (2009b). Modelling long-term averages of local
ambient air pollution in Oslo, Norway: evaluation of nitrogen dioxide, PM10 and PM2.5. Int J Environ Pollut
36: 110-126. http://dx.doi.org/10.1504/IJEP.2009.021820
Ohashi. Y: Nakai. Y: Sugiura. Y: Ohno. Y: Okamoto. H: Tanaka. A: Kakinoki. Y: Havashi. M. (1994). Nitrogen
dioxide-induced eosinophilia and mucosal injury in the nose of the guinea pig. Acta Otolaryngol 114: 547-
551.
Ohyama. K: Ito. T: Kanisawa. M. (1999). The roles of diesel exhaust particle extracts and the promotive effects
of NO2 and/or SO2 exposure on rat lung tumorigenesis. Cancer Lett 139: 189-197.
http://dx.doi.org/10.1016/S0304-3835(99)00040-3
Oiamo. TH: Luginaah. IN: Atari. DO: Gorev. KM. (2011). Air pollution and general practitioner access and
utilization: a population based study in Sarnia, 'Chemical Valley,' Ontario. Environ Health 10: 71.
http://dx.doi.org/10.1186/1476-069X-10-71
Oiamo. TH: Luginaah. IN: Buzzelli. M: Tang. K: Xu. X: Brook. JR: Johnson. M. (2012). Assessing the spatial
distribution of nitrogen dioxide in London, Ontario. J Air Waste Manag Assoc 62: 1335-1345.
http://dx.doi.org/10.1080/10962247.2012.715114
Olaguer. EP. (2012). The potential near-source ozone impacts of upstream oil and gas industry emissions. J Air
Waste Manag Assoc 62: 966-977. http://dx.doi.org/10.1080/10962247.2012.688923
R-67
-------
Olaguer. EP: Kolb. CE: Lefer. B: Rappengliick. B: Zhang. RY: Pinto. JP. (2014). Overview of the SHARP
campaign: Motivation, design, and major outcomes. J Geophys Res 119: 2597-2610.
http://dx.doi.org/10.1002/2013JD019730
Olaguer. EP: Rappengliick. B: Lefer. B: Stutz. J: Dibb. J: Griffin. R: Brune. WH: Shauck. M: Buhr. M: Jeffries.
H: Vizuete. W: Pinto. JP. (2009). Deciphering the role of radical precursors during the Second Texas Air
Quality Study. J Air Waste Manag Assoc 59: 1258-1277. htto://dx.doi.org/10.3155/1047-3289.59.11.1258
Olin. AC: Aldenbratt A: Ekman. A: Ljungkvist. G: Jungersten. L: Alving. K: Toren. K. (2001). Increased nitric
oxide in exhaled air after intake of a nitrate-rich meal. RespirMed95: 153-158.
http://dx.doi.org/10.1053/rmed.2000.1010
Olin. AC: Bake. B: Toren. K. (2007). Fraction of exhaled nitric oxide at 50 mL/s: Reference values for adult
lifelong never-smokers. Chest 131: 1852-1856. http://dx.doi.org/10.1378/chest.06-2928
Olin. AC: Hellgren. J: Karlsson. G: Ljungkvist. G: Nolkrantz. K: Toren. K. (1998). Nasal nitric oxide and its
relationship to nasal symptoms, smoking and nasal nitrate. Rhinology 36: 117-121.
Olsson. D: Ekstrom. M: Forsberg. B. (2012). Temporal variation in air pollution concentrations and preterm
birth-a population based epidemiological study. Int J Environ Res Public Health 9: 272-285.
http://dx.doi.org/10.3390/iierph9010272
Olsson. D: Mogren. I: Forsberg. B. (2013). Air pollution exposure in early pregnancy and adverse pregnancy
outcomes: a register-based cohort study. BMJ Open 3. http://dx.doi.org/10.1136/bmjopen-2012-001955
Orazzo. F: Nespoli. L: Ito. K: Tassinari. D: Giardina. D: Funis. M: Cecchi. A: Trapani. C: Forgeschi. G: Vignini.
M: Nosetti. L: Pigna. S: Zanobetti. A. (2009). Air pollution, aeroallergens, and emergency room visits for
acute respiratory diseases and gastroenteric disorders among young children in six Italian cities. Environ
Health Perspect 117: 1780-1785. http://dx.doi.org/10.1289/ehp.0900599
Orehek. J: Massari. JP: Gayrard. P: Grimaud. C: Charpin. J. (1976). Effect of short-term, low-level nitrogen
dioxide exposure on bronchial sensitivity of asthmatic patients. J Clin Invest 57: 301-307.
http://dx.doi.org/10.1172/JCI108281
Orehek. J: Nicoli. MM: Delpierre. S: Beaupre. A. (1981). Influence of the previous deep inspiration on the
spirometric measurement of provoked bronchoconstriction in asthma. Am Rev Respir Dis 123: 269-272.
Ortman. JM: Velkoff. VA: Hogan. H. (2014). An aging nation: The older population in the United States (pp. 1-
28). (P25-1140). United States Census Bureau. http://www.census.gov/library/pubrications/2014/demo/p25-
1140.html
Osthoff. HP: Roberts. JM: Ravishankara. AR: Williams. EJ: Lerner. BM: Sommariva. R: Bates. TS: Coffman.
D: Quinn. PK: Dibb. JE: Stark. H: Burkholder. JB: Talukdar. RK: Meagher. J: Fehsenfeld. FC: Brown. SS.
(2008). High levels of nitryl chloride in the polluted subtropical marine boundary layer. Nat Geosci 1: 324-
328. http://dx.doi.org/10.1038/ngeol77
Ostro. B: Lipsett M: Mann. J: Braxton-Owens. H: White. M. (2001). Air pollution and exacerbation of asthma
in African-American children in Los Angeles. Epidemiology 12: 200-208.
Oudin. A: Stromberg. U: Jakobsson. K: Stroh. E: Bjork. J. (2010). Estimation of short-term effects of air
pollution on stroke hospital admissions in southern Sweden. Neuroepidemiology 34: 131-142.
http://dx.doi.org/10.1159/000274807
Oudin. A: Stromberg. U: Jakobsson. K: Stroh. E: Lindgren. AG: Norrving. B: Pessah-Rasmussen. H: Engstrom.
G: Bjork. J. (2011). Hospital admissions for ischemic stroke: does long-term exposure to air pollution
interact with major risk factors. Cerebrovasc Dis 31: 284-293. http://dx.doi.org/10.1159/000322600
Overton. JH. Jr. (1984). Physicochemical processes and the formulation of dosimetry models. In FJ Miller; DB
Menzel (Eds.), Fundamentals of extrapolation modeling of inhaled toxicants: Ozone and nitrogen dioxide
(pp. 93-114). Washington, DC: Hemisphere Publishing Corporation.
Ozden. O: Dogeroglu. T. (2008). Field evaluation of a tailor-made new passive sampler for the determination of
NO2 levels in ambient air. Environ Monit Assess 142: 243-253. http://dx.doi.org/10.1007/sl0661-007-9921-
R-68
-------
Ozkaynak. H: Frev. HC: Burke. J: Finder. RW. (2009). Analysis of coupled model uncertainties in source-to-
dose modeling of human exposures to ambient air pollution: A PM2.5 case study. Atmos Environ 43: 1641-
1649. http://dx.doi.0rg/10.1016/i.atmosenv.2008.12.008
Pacher. P: Beckman. JS: Liaudet L. (2007). Nitric oxide and peroxynitrite in health and disease [Review].
Physiol Rev 87: 315-424. http://dx.doi.org/10.1152/physrev.00029.2006
Pachon. JE: Balachandran. S: Hu. Y: Mulholland. JA: Darrow. LA: Sarnat JA: Tolbert. PE: Russell AG.
(2012). Development of outcome-based, multipollutant mobile source indicators. J Air Waste Manag Assoc
62: 431-442. http://dx.doi.org/10.1080/10473289.2012.656218
Paciorek. CJ. (2010). The importance of scale for spatial-confounding bias and precision of spatial regression
estimators. Stat Sci 25: 107-125. http://dx.doi.org/10.1214/10-STS326
Pacsi. AP: Alhajeri. NS: Zavala-Araiza. D: Webster. MD: Allen. DT. (2013). Regional air quality impacts of
increased natural gas production and use in Texas. Environ Sci Technol 47: 3521-3527.
http://dx.doi.org/10.1021/es3044714
Padhj BK: Padhy. PK. (2008). Domestic fuels, indoor air pollution, and children's health. Ann N Y Acad Sci
1140: 209-217. http://dx.doi.org/10.1196/annals. 1454.015
Padro-Martinez. LT: Patton. AP: Trull JB: Zamore. W: Brugge. D: Durant. JL. (2012). Mobile monitoring of
particle number concentration and other traffic-related air pollutants in a near-highway neighborhood over
the course of ayear. Atmos Environ61: 253-264. http://dx.doi.0rg/10.1016/j.atmosenv.2012.06.088
Padula. AM: Tager. IB: Carmichael SL: Hammond. SK: Lurmann. F: Shaw. GM. (2013a). The association of
ambient air pollution and traffic exposures with selected congenital anomalies in the San Joaquin Valley of
California. Am JEpidemiol 177: 1074-1085. http://dx.doi.org/10.1093/aie/kws367
Padula. AM: Tager. IB: Carmichael SL: Hammond. SK: Yang. W: Lurmann. F: Shaw. GM. (2013b). Ambient
air pollution and traffic exposures and congenital heart defects in the San Joaquin Valley of California.
Paediatr Perinat Epidemiol 27: 329-339. http://dx.doi.org/10.1111/ppe. 12055
Padula. AM: Tager. IB: Carmichael SL: Hammond. SK: Yang. W: Lurmann. FW: Shaw. GM. (2013c). Traffic-
related air pollution and selected birth defects in the San Joaquin Valley of California. Birth Defects Res A
ClinMol Teratol 97: 730-735. http://dx.doi.org/10.1002/bdra.23175
Pagani. P: Romano. M: Erroi. A: Ferro. M: Salmona. M. (1994). Biochemical effects of acute and subacute
nitrogen dioxide exposure in rat lung and bronchoalveolar lavage fluid. Arch Environ Contam Toxicol 27:
426-430. http://dx.doi.org/10.1007/BF00213182
Panasevich. S: Leander. K: Ljungman. P: Bellander. T: de Faire. U: Pershagen. G: Nyberg. F. (2013). Interaction
between air pollution exposure and genes in relation to levels of inflammatory markers and risk of
myocardial infarction. BMJ Open 3: e003058. http://dx.doi.org/10.1136/bmiopen-2013-003058
Panasevich. S: Leander. K: Rosenlund. M: Ljungman. P: Bellander. T: de Faire. U: Pershagen. G: Nyberg. F.
(2009). Associations of long- and short-term air pollution exposure with markers of inflammation and
coagulation in a population sample. Occup Environ Med 66: 747-753.
http://dx.doi.org/10.1136/oem.2008.043471
Panteliadis. P: Strak. M: Hoek. G: Weijers. E: van Per Zee. S: Dijkema. M. (2014). Implementation of a low
emission zone and evaluation of effects on air quality by long-term monitoring. Atmos Environ 86: 113-119.
http://dx.doi.0rg/10.1016/i.atmosenv.2013.12.035
Papathomas. M: Molitor. J: Richardson. S: Riboli. E: Vineis. P. (2011). Examining the joint effect of multiple
risk factors using exposure risk profiles: Lung cancer in non-smokers. Environ Health Perspect 119: 84-91.
http://dx.doi.org/10.1289/ehp.1002118
Parent ME: Goldberg. MS: Grouse. PL: Ross. NA: Chen. H: Valois. MF: Liautaud. A. (2013). Traffic-related
air pollution and prostate cancer risk: A case-control study in Montreal, Canada. Occup Environ Med 70:
511-518. http://dx.doi.org/10.1136/oemed-2012-101211
Parenteau. MP: Sawada. MC. (2012). The role of spatial representation in the development of a LUR model for
Ottawa, Canada. Air Qual Atmos Health 5: 311-323. http://dx.doi.org/10.1007/sll869-010-0094-3
R-69
-------
Park. JW: Lim. YH: Kyung. SY: An, CH: Lee. SP: Jeong. SH: Ju. YS. (2005). Effects of ambient paniculate
matter on peak expiratory flow rates and respiratory symptoms of asthmatics during Asian dust periods in
Korea. Respirology 10: 470-476. http://dx.doi.0rg/10.llll/i.1440-1843.2005.00728.x
Park. SK: Auchincloss. AH: O'Neill. MS: Prineas. R: Correa. JC: Keeler. J: Barr. RG: Kaufman. JD: Diez Roux.
AV. (2010). Particulate air pollution, metabolic syndrome, and heart rate variability: The Multi-Ethnic Study
of Atherosclerosis (MESA). Environ Health Perspect 118: 1406-1411. http://dx.doi.org/10.1289/ehp.0901778
Park. SS: Hong. JH: Lee. JH: Kim. YJ: Cho. SY: Kim. SJ. (2008). Investigation of nitrous acid concentration in
an indoor environment using an in-situ monitoring system. Atmos Environ 42: 6586-6596.
http://dx.doi.0rg/10.1016/i.atmosenv.2008.05.023
Parker. JD: Akinbami. LJ: Woodruff. TJ. (2009). Air pollution and childhood respiratory allergies in the United
States. Environ Health Perspect 117: 140-147. http://dx.doi.org/10.1289/ehp. 11497
Parker. JD: Rich. DQ: Glinianaia. SV: Leem. JH: Wartenberg. D: Bell ML: Bonzini. M: Brauer. M: Darrow. L:
Gehring. U: Gouveia. N: Grillo. P: Ha. E: van den Hooven. EH: Jalaludin. B: Jesdale. BM: Lepeule. J:
Morello-Frosch. R: Morgan. GG: Slama. R: Pierik. FH: Pesatori. AC: Sathyanaravana. S: Seo. J: Strickland.
M: Tamburic. L: Woodruff. TJ. (2011). The international collaboration on air pollution and pregnancy
outcomes: Initial results. Environ Health Perspect 119: 1023-1028. http://dx.doi.org/10.1289/ehp. 1002725
Parker. RF: Davis. JK: Cassell. GH: White. H: Dziedzic. D: Blalock. DK: Thorp. RB: Simecka. JW. (1989).
Short-term exposure to nitrogen dioxide enhances susceptibility to murine respiratory mycoplasmosis and
decreases intrapulmonary killing of Mycoplasma pulmonis. Am J Respir Crit Care Med 140: 502-512.
http://dx.doi.0rg/10.1164/airccm/140.2.502
Parrish. DP: Fehsenfeld. FC. (2000). Methods for gas-phase measurements of ozone, ozone precursors and
aerosol precursors. Atmos Environ 34: 1921-1957. http://dx.doi.org/10.1016/S1352-2310(99)00454-9
Patel MM: Chillrud. SN: Correa. JC: Hazi. Y: Feinberg. M: Deepti. KG: Prakash. S: Ross. JM: Lew. D:
Kinnev. PL. (2010). Traffic-related paniculate matter and acute respiratory symptoms among New York City
area adolescents. Environ Health Perspect 118: 1338-1343. http://dx.doi.org/10.1289/ehp.0901499
Patel. MM: Chillrud. SN: Deepti. KG: Ross. JM: Kinnev. PL. (2013). Traffic-related air pollutants and exhaled
markers of airway inflammation and oxidative stress in New York City adolescents. Environ Res 121: 71-78.
http://dx.doi.0rg/10.1016/i.envres.2012.10.012
Pathmanathan. S: Krishna. MT: Blomberg. A: Helledav. R: Kelly. FJ: Sandstrom T: Holgate. ST: Wilson. SJ:
Frew. AJ. (2003). Repeated daily exposure to 2 ppm nitrogen dioxide upregulates the expression of IL-5, IL-
10, IL-13, and ICAM-1 in the bronchial epithelium of healthy human airways. Occup Environ Med 60: 892-
896. http://dx.doi.org/10.1136/oem.6Q. 11.892
Pattenden. S: Hoek. G: Braun-Fahrlander. C: Forastiere. F: Kosheleva. A: Neuberger. M: Fletcher. T. (2006).
NO2 and children's respiratory symptoms in the PATY study. Occup Environ Med 63: 828-835.
http://dx.doi.org/10.1136/oem.2006.025213
Peacock. JL: Anderson. HR: Bremner. SA: Marston. L: Seemungal TA: Strachan. DP: Wedzicha. JA. (2011).
Outdoor air pollution and respiratory health in patients with COPD. Thorax 66: 591-596.
http://dx.doi.org/10.1136/thx.2010.155358
Peacock. JL: Symonds. P: Jackson. P: Bremner. SA: Scarlett. JF: Strachan. DP: Anderson. HR. (2003). Acute
effects of winter air pollution on respiratory function in schoolchildren in southern England. Occup Environ
Med 60: 82-89. http://dx.doi.0rg/10.1136/oem.60.2.82
Peacock. M. (2006). Memorandum from Marcus Peacock, Deputy Administrator of the U.S. EPA, to Dr. George
Gray, Assistant Administrator of ORD, and Bill Wehrum, Acting Assistant Administrator of OAR: Process
for reviewing National Ambient Air Quality Standards. Washington, D.C.: U.S. Environmental Protection
Agency.
Peacock. MC. (2008). Letter from Marcus C. Peacock to Rogene Henderson (regarding CASAC comments on
application of new NAAQS process). Available online at
http://www.epa.gov/ttn/naaqs/pdfs/PeacocklettertoCASAC090808.pdf
R-70
-------
Pedersen. M: Giorgis-Allemand. L: Bernard. C: Aguilera. I: Andersen. AM: Ballester. F: Beelen. RM: Chatzi. L:
Cirach. M: Danileviciute. A: Dedele. A: vanEijsden. M: Estarlich. M: Fernandez-Somoano. A: Fernandez.
MF: Forastiere. F: Gehring. U: Grazuleviciene. R: Gruzieva. O: Heude. B: Hoek. G: de Hoogh. K: van Den
Hooven. EH: Haberg. SE: Jaddoe. VW: Kliimper. C: Korek. M: Kramer. U: Lerchundi. A: Lepeule. J:
Nafstad. P: Nystad. W: Patelarou. E: Porta. D: Postma. D: Raaschou-Nielsen. O: Rudnai. P: Sunyer. J:
Stephanou. E: Sorensen. M: Thiering. E: Tuffnell D: Varro. MJ: Vrijkotte. TG: Wijga. A: Wilhelm. M:
Wright J: Nieuwenhuijsen. MJ: Pershagen. G: Brunekreef B: Kogevinas. M: Slama. R. (2013). Ambient air
pollution and low birthweight: a European cohort study (ESCAPE). Lancet Respir Med 1: 695-704.
http://dx.doi.org/10.1016/S2213-2600(13)70192-9
Pedersen. M: Stayner. L: Slama. R: Serensen. M: Figueras. F: Nieuwenhuijsen. MJ: Raaschou-Nielsen. O:
Dadvand. P. (2014). Ambient air pollution and pregnancy-induced hypertensive disorders: a systematic
review and meta-analysis [Review]. Hypertension 64: 494-500.
http://dx.doi.org/10.1161/HYPERTENSIONAHA. 114.03545
Peel JL: Klein. M: Flanders. WD: Mulholland. JA: Freed. G: Tolbert. PE. (2011). Ambient air pollution and
apnea and bradycardia in high-risk infants on home monitors. Environ Health Perspect 119: 1321-1327.
http://dx.doi.org/10.1289/ehp.1002739
Peel JL: Metzger. KB: Klein. M: Flanders. WD: Mulholland. JA: Tolbert. PE. (2007). Ambient air pollution and
cardiovascular emergency department visits in potentially sensitive groups. Am J Epidemiol 165: 625-633.
http://dx.doi.org/10.1093/aie/kwk051
Peel. JL: Tolbert. PE: Klein. M: Metzger. KB: Flanders. WD: Todd. K: Mulholland. JA: Ryan. PB: Frumkin. H.
(2005). Ambient air pollution and respiratory emergency department visits. Epidemiology 16: 164-174.
http://dx.doi.org/10.1097/01.ede.0000152905.42113.db
Pegas. PN: Nunes. T: Alves. CA: Silva. JR: Vieira. SLA: Caseiro. A: Pio. CA. (2012). Indoor and outdoor
characterisation of organic and inorganic compounds in city centre and suburban elementary schools of
Aveiro, Portugal. Atmos Environ 55: 80-89. http://dx.doi.0rg/10.1016/i.atmosenv.2012.03.059
Pereira. G: Cook. AG: Haggar. F: Bower. C: Nassar. N. (2012). Locally derived traffic-related air pollution and
fetal growth restriction: A retrospective cohort study. Occup Environ Med 69: 815-822.
http://dx.doi.org/10.1136/oemed-2011-100509
Pereira. G: Haggar. F: Shand. AW: Bower. C: Cook. A: Nassar. N. (2013). Association between pre-eclampsia
and locally derived traffic-related air pollution: A retrospective cohort study. J Epidemiol Community Health
67: 147-152. http://dx.doi.org/10.1136/iech-2011-200805
Perring. AE: Pusede. SE: Cohen. RC. (2013). An observational perspective on the atmospheric impacts of alkyl
and multifunctional nitrates on ozone and secondary organic aerosol. ChemRev 113: 5848-5870.
http://dx.doi.org/10.1021/cr300520x
Peters. A: Breitner. S: Cyrys. J: Stolzel M: Pitz. M: Wolke. G: Heinrich. J: Kreyling. W: Kiichenhoff H:
Wichmann. H. (2009). The influence of improved air quality on mortality risks in Erfurt, Germany. (137).
Boston, MA: Health Effects Institute. http://pubs.healtheffects.org/view.php?id=299
Peters. JM: Avol E: Gauderman. WJ: Linn. WS: Navidi W: London. SJ: Margolis. H: Rappaport. E: Vora. H:
Gong. H. Jr: Thomas. DC. (1999). A study of twelve southern California communities with differing levels
and types of air pollution: II. Effects on pulmonary function. Am J Respir Crit Care Med 159: 768-775.
http://dx.doi.0rg/10.1164/airccm.159.3.9804144
Petroeschevsky. A: Simpson. RW: Thalib. L: Rutherford. S. (2001). Associations between outdoor air pollution
and hospital admissions in Brisbane, Australia. Arch Environ Occup Health 56: 37-52.
http://dx.doi.org/10.1080/00039890109604053
Pfister. GG: Wiedinmyer. C: Emmons. LK. (2008). Impacts of the fall 2007 California wildfires on surface
ozone: Integrating local observations with global model simulations. Geophys Res Lett 35: L19814.
http://dx.doi.org/10.1029/2008gl034747
R-71
-------
Physick. W: Powell J: Cope. M: Boast K: Lee. S. (2011). Measurements of personal exposure to NO2 and
modelling using ambient concentrations and activity data. Atmos Environ 45: 2095-2102.
http://dx.doi.0rg/10.1016/i.atmosenv.2011.01.063
Piechocki-Minguy. A: Plaisance. H: Schadkowskj C: Sagnier. I: Saison. JY: Galloo. JC: Guillermo. R. (2006).
A case study of personal exposure to nitrogen dioxide using a new high sensitive diffusive sampler. Sci Total
Environ 366: 55-64. http://dx.doi.0rg/10.1016/j.scitotenv.2005.08.009
Pilotto. LS: Nitschke. M: Smith. BJ: Pisaniello. D: Ruffin. RE: McElroy. HJ: Martin. J: Killer. JE. (2004).
Randomized controlled trial of unflued gas heater replacement on respiratory health of asthmatic
schoolchildren. Int J Epidemiol 33: 208-214. http://dx.doi.org/10.1093/ije/dyh018
Pinkston. P: Smeglin. A: Roberts. NJ. Jr: Gibb. FR: Morrow. PE: Utell MJ. (1988). Effects of in vitro exposure
to nitrogen dioxide on human alveolar macrophage release of neutrophil chemotactic factor and interleukin-
1. Environ Res 47: 48-58. http://dx.doi.org/10.1016/S0013-9351(88)80020-3
Pinto. JP: Dibb. J: Lee. BH: Rappengliick. B: Wood. EC: Lew. M: Zhang. RY: Lefer. B: Ren. XR: Stutz. J:
Tsai. C: Ackermann. L: Golovko. J: Herndon. SC: Oakes. M: Meng. QY: Munger. JW: Zahniser. M: Zheng.
I (2014). Intercomparison of field measurements of nitrous acid (HONO) during the SHARP campaign. J
Geophys Res Atmos 119: 5583-5601. http://dx.doi.org/10.1002/2013JD020287
Pinto. JP: Lefohn. AS: Shadwick. PS. (2004). Spatial variability of PM2.5 in urban areas in the United States. J
Air Waste Manag Assoc 54: 440-449.
Pirjola. L: Lahde. T: Niemi. JV: Kousa. A: Ronkko. T: Karjalainen. P: Keskinen. J: Frev. A: Hillamo. R. (2012).
Spatial and temporal characterization of traffic emissions in urban microenvironments with a mobile
laboratory. Atmos Environ 63: 156-167. http://dx.doi.0rg/10.1016/i.atmosenv.2012.09.022
Pitard. A: Zeghnoun. A: Courseaux. A: Lambertv. J: Delmas. V: Fossard. JL: Villet. H. (2004). Short-term
associations between air pollution and respiratory drug sales. Environ Res 95: 43-52.
http://dx.doi.0rg/10.1016/i.envres.2003.08.006
Pleijel. H: Karlsson. GP: Gerdin. EB. (2004). On the logarithmic relationship between NO2 concentration and
the distance from a highroad. Sci Total Environ 332: 261-264.
http://dx.doi.0rg/10.1016/i.scitotenv.2004.03.020
Plopper. CG: Hatch. GE: Wong. V: Duan. X: Weir. AJ: Tarkington. BK: Devlin. RB: Becker. S: Buckpitt AR.
(1998). Relationship of inhaled ozone concentration to acute tracheobronchial epithelial injury, site-specific
ozone dose and glutathione depletion in rhesus monkeys. Am J Respir Cell Mol Biol 19: 387-399.
http://dx.doi.org/10.1165/aircmb. 19.3.3183
Poddubny. VA: Yushketova. NA. (2013). A physicochemical model of sorption processes in NO(2) passive
sampling with air humidity effects. Environ Monit Assess 185: 3819-3829. http://dx.doi.org/10.1007/sl0661-
012-2830-7
Polidori. A: Fine. PM. (2012a). Ambient concentrations of criteria and air toxic pollutants in close proximity to a
freeway with heavy-duty diesel traffic. South Coast Air Quality Management District.
www.aqmd.gov/tao/AQ-Reports/I710Fwv Studv.pdf
Polidori. A: Fine. PM. (2012b). Ambient concentrations of criteria and air toxic pollutants in close proximity to a
freeway with heavy-duty diesel traffic. Final report. South Coast Air Quality Management District.
http://www.aqmd.gov/docs/default-source/air-qualitv/air-qualitv-monitoring-studies/near-roadwav-
studv.pdf? sfVrsn=2
Pollack. IB: Lerner. BM: Ryerson. TB. (2011). Evaluation of ultraviolet light-emitting diodes for detection of
atmospheric NO2 by photolysis - chemiluminescence. J Atmos Chem 65: 111-125.
http://dx.doi.org/10.1007/sl0874-011-9184-3
R-72
-------
Pollack. IB: Ryerson. TB: Trainer. M: Parrish. DP: Andrews. AE: Atlas. EL: Blake. PR: Brown. SS: Commane.
R: Daube. BC: De Gouw. JA: Dube. WP: Flynn. J: Frost. GJ: Oilman. JB: Grossberg. N: Hollowav. JS:
Kofler. J: Kort. EA: Kuster. WC: Lang. PM: Lefer. B: Lueb. RA: Neuman. JA: Nowak. JB: Novell!. PC:
Peischl. J: Perring. AE: Roberts. JM: Santoni. G: Schwarz. JP: Spackman. J. R.: Wagner. NL: Warneke. C:
Washenfelder. RA: Wofsv. SC: Xiang. B. (2012). Airborne and ground-based observations of a weekend
effect in ozone, precursors, and oxidation products in the California South Coast Air Basin. J Geophys Res
117: DOOV05. http://dx.doi.org/10.1029/2011JD016772
Poloniecki. JD: Atkinson. RW: Ponce de Leon. A: Anderson. HR. (1997). Daily time series for cardiovascular
hospital admissions and previous day's air pollution in London, UK. Occup Environ Med 54: 535-540.
http://dx.doi.0rg/10.1136/oem.54.8.535
Ponka. A: Virtanen. M. (1996). Low-level air pollution and hospital admissions for cardiac and cerebrovascular
diseases inHelsinki. Am J Public Health 86: 1273-1280. http://dx.doi.Org/10.2105/AJPH.86.9.1273
Pope. CA. Ill: Burnett. RT: Thun. MJ: Calle. EE: Krewskj D: Ito. K: Thurston. GD. (2002). Lung cancer,
cardiopulmonary mortality, and long-term exposure to fine paniculate air pollution. JAMA 287: 1132-1141.
http://dx.doi.0rg/10.1001/iama.287.9.1132
Pope. CA. Ill: Thun. MJ: Namboodiri. MM: Dockery. DW: Evans. JS: Speizer. FE: Heath. CW. Jr. (1995).
Paniculate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am J Respir Crit
Care Med 151: 669-674. http://dx.doi.0rg/10.1164/airccm/151.3 Pt 1.669
Posin. C: Clark. K: Jones. MP: Patterson. JV: Buckley. RD: Hackney. JD. (1978). Nitrogen dioxide inhalation
and human blood biochemistry. Arch Environ Occup Health 33: 318-324.
http://dx.doi.org/10.1080/00039896.1978.10667355
Postlethwait EM: Bidani. A. (1989). Pulmonary disposition of inhaled NO2-nitrogen in isolated rat lungs.
Toxicol Appl Pharmacol 98: 303-312.
Postlethwait EM: Bidani. A. (1990). Reactive uptake governs the pulmonary air space removal of inhaled
nitrogen dioxide. J Appl Physiol 68: 594-603.
Postlethwait EM: Bidani. A. (1994). Mechanisms of pulmonary NO2 absorption [Review]. Toxicology 89: 217-
237.
Postlethwait EM: Bioani. A: Evans. MJ. (1990). The effect of NO2 exposure on perfusate distribution in
isolated rat lungs: Pulmonary versus bronchial circulation. Toxicol Appl Pharmacol 106: 456-461.
http://dx.doi.org/10.1016/0041-008X(90)90340-Z
Postlethwait EM: Langford. SD: Bidani. A. (1991a). Interfacial transfer kinetics of NO2 into pulmonary
epithelial lining fluid. J Appl Physiol 71: 1502-1510.
Postlethwait EM: Langford. SD: Bidani. A. (1991b). Transfer of NO2 through pulmonary epithelial lining fluid.
Toxicol Appl Pharmacol 109: 464-471. http://dx.doi.org/10.1016/0041-008X(91)90009-4
Postlethwait EM: Langford. SD: Bidani. A. (1992). Kinetics of NO2 air space absorption in isolated rat lungs. J
Appl Physiol 73: 1939-1945.
Postlethwait EM: Langford. SD: Jacobson. LM: Bidani. A. (1995). NO2 reactive absorption substrates in rat
pulmonary surface lining fluids. Free Radic Biol Med 19: 553-563. http://dx.doi.org/10.1016/0891-
5849(95)00058-6
Postlethwait EM: Mustafa. MG. (1981). Fate of inhaled nitrogen dioxide in isolated perfused rat lung. J Toxicol
Environ Health 7: 861-872. http://dx.doi.org/10.1080/15287398109530029
Postlethwait EM: Mustafa. MG. (1989). Effect of altered dose rate on NO2 uptake and transformation in
isolated lungs. J Toxicol Environ Health 26: 497-507. http://dx.doi.org/10.1080/15287398909531272
Postma. PS: Boezen. HM. (2004). Rationale for the Dutch hypothesis. Allergy and airway hyperresponsiveness
as genetic factors and their interaction with environment in the development of asthma and COPD [Review].
Chest 126: 96S-104S. http://dx.doi.0rg/10.1378/chest.126.2 suppl 1.96S
R-73
-------
Povnter. ME: Persinger. RL: Irvin. CG: Butnor. KJ: vanHirtura H: Blav. W: Heintz. NH: Robbins. J:
Hemenwav. D: Taaties. DJ: Janssen-Heininger. Y. (2006). Nitrogen dioxide enhances allergic airway
inflammation and hyperresponsiveness in the mouse. Am J Physiol Lung Cell Mol Physiol 290: L144-L152.
http://dx.doi.org/10.1152/aiplung.00131.2005
Pratt. GC: Parson. K: Shinoda. N: Lindgrea P: Dunlap. S: Yawn. B: Wollaa P: Johnson. J. (2014). Quantifying
traffic exposure. J Expo Sci Environ Epidemiol 24: 290-296. http://dx.doi.org/10.1038/jes.2013.51
Proust B: Lacroix. G: Robidel F: Marliere. M: Lecomte. A: Vargaftig. BB. (2002). Interference of a short-term
exposure to nitrogen dioxide with allergic airways responses to allergenic challenges in B ALB/c mice.
Mediators Inflamm 11: 251-260. http://dx.doi.org/10.1080/096293502900000113
Priitz. WA: Monig. H: Butler. J: Land. EJ. (1985). Reactions of nitrogen dioxide in aqueous model systems:
Oxidation of tyrosine units inpeptides and proteins. ArchBiochemBiophys 243: 125-134.
http://dx.doi.org/10.1016/0003-9861(85)90780-5
Pryor. WA. (1992). How far does ozone penetrate into the pulmonary air/tissue boundary before it reacts? Free
Radic Biol Med 12: 83-88. http://dx.doi.org/10.1016/0891-5849(92)90060-T
Pryor. WA: Houk. KN: Foote. CS: Fukuto. JM: Ignarro. LJ: Squadrito. GL: Davies. KJA. (2006). Free radical
biology and medicine: It's a gas, man! [Review]. Am J Physiol Regul Integr Comp Physiol 291: R491-R511.
http://dx.doi.org/10.1152/aipregu.00614.2005
Pryor. WA: Lightsev. JW. (1981). Mechanisms of nitrogen dioxide reactions: Initiation of lipid peroxidation and
the production of nitrous acid. Science 214: 435-437. http://dx.doi.org/10.1126/science.214.4519.435
Puiades-Rodriguez. M: Mckeever. T: Lewis. S: Whyatt D: Britten. J: Venn. A. (2009). Effect of traffic pollution
on respiratory and allergic disease in adults: Cross-sectional and longitudinal analyses. BMC Pulm Med 9:
42. http://dx.doi.org/10.1186/1471-2466-9-42
Pun. BK: Seigneur. C: White. W. (2003). Day-of-week behavior of atmospheric ozone in three US cities. J Air
Waste Manag Assoc 53: 789-801. http://dx.doi.org/10.1080/10473289.2003.10466231
Purvis. MR: Ehrlich. R. (1963). Effect of atmospheric pollutants on susceptibility to respiratory infection: II
Effect of nitrogen dioxide. J Infect Dis 113: 72-76.
Pusede. SE: Cohen. RC. (2012). On the observed response of ozone to NOx and VOC reactivity reductions in
San Joaquin Valley California 1995-present. Atmos ChemPhys 12: 8323-8339.
http://dx.doi.org/10.5194/acp-12-8323-2012
Qian. Z: Lin. HM: Chinchillj VM: Lehman. EB: Duan. Y: Craig. TJ: Wilson. WE: Liao. D: Lazarus. SC:
Bascom. R. (2009a). Interaction of ambient air pollution with asthma medication on exhaled nitric oxide
among asthmatics. Arch Environ Occup Health 64: 168-176. http://dx.doi.org/10.1080/19338240903240616
Qian. Z: Lin. HM: Chinchillj VM: Lehman. EB: Stewart. WF: Shah. N: Duan. Y: Craig. TJ: Wilson. WE: Liao.
D: Lazarus. SC: Bascom. R. (2009b). Associations between air pollution and peak expiratory flow among
patients with persistent asthma. J Toxicol Environ Health A 72: 39-46.
http://dx.doi.org/10.1080/15287390802445517
Qiu. H: Yu. IT: Wang. X: Tian. L: Tse. LA: Wong. TW. (2013a). Cool and dry weather enhances the effects of
air pollution on emergency IHD hospital admissions. Int J Cardiol 168: 500-505.
http://dx.doi.0rg/10.1016/i.iicard.2012.09.199
Qiu. H: Yu. ITS: Wang. X: Tian. L: Tse. LA: Wong. TW. (2013b). Season and humidity dependence of the
effects of air pollution on COPD hospitalizations in Hong Kong. Atmos Environ 76: 74-80.
http://dx.doi.0rg/10.1016/i.atmosenv.2012.07.026
Que. LG: Yang. Z: Stamler. JS: Lugogo. NL: Kraft M. (2009). S-nitrosoglutathione reductase: An important
regulator in human asthma. Am J Respir Crit Care Med 180: 226-231.
http://dx.doi.org/10.1164/rccm.200901-0158OC
R-74
-------
Quintana. PJE: Dumbauld. JJ: Garnica. L: Chowdhury. MZ: Velascosoltero. J: Mota-Raigoza. A: Flores. D:
Rodriguez. E: Panagon. N: Gamble. J: Irbv. T: Tran. C: Elder. J: Galaviz. VE: Hoffman. L: Zavala. M:
Molina. LT. (2014). Traffic-related air pollution in the community of San Ysidro, CA, in relation to
northbound vehicle wait times at the US-Mexico border Port of Entry. Atmos Environ 88: 353-361.
http://dx.doi.0rg/10.1016/i.atmosenv.2014.01.009
Raaschou-Nielsen. O: Andersen. ZJ: Beelen. R: Samoli. E: Stafoggia. M: Weinmayr. G: Hoffmann. B: Fischer.
P: Nieuwenhuijsen. MJ: Brunekreef. B: Xun. WW: Katsouvanni. K: Dimakopoulou. K: Sommar. J: Forsberg.
B: Modig. L: Oudin. A: Oftedal. B: Schwarze. PE: Nafstad. P: De Faire. U: Pedersen. NL: Ostenson. CG:
Fratiglioni. L: Penell. J: Korek. M: Pershagen. G: Eriksen. KT: Serensen. M: Tjonneland. A: Ellermann. T:
Eeftens. M: Peeters. PH: Meliefste. K: Wang. M: Bueno-De-mesquita. B: Key. TJ: De Hoogh. K: Concin. H:
Nagel G: Vilier. A: Grioni. S: Krogh. V: Tsai MY: Ricceri. F: Sacerdote. C: Galassi. C: Migliore. E: Ranzi.
A: Cesaroni. G: Badaloni. C: Forastiere. F: Tamavo. I: Amiano. P: Dorronsoro. M: Trichopoulou. A: Bamia.
C: Vineis. P: Hoek. G. (2013a). Air pollution and lung cancer incidence in 17 European cohorts: Prospective
analyses from the European Study of Cohorts for Air Pollution Effects (ESCAPE). Lancet Oncol 14: 813-
822. http://dx.doi.org/10.1016/S1470-2045(13)70279-l
Raaschou-Nielsen. O: Andersen. ZJ: Hvidberg. M: Jensen. SS: Ketzel. M: Sorensen. M: Hansen. J: Loft. S:
Overvad. K: Tjonneland. A. (201 la). Air pollution from traffic and cancer incidence: A Danish cohort study.
Environ Health 10: 67. http://dx.doi.org/10.1186/1476-069X-10-67
Raaschou-Nielsen. O: Andersen. ZJ: Hvidberg. M: Jensen. SS: Ketzel M: Serensen. M: Loft S: Overvad. K:
Tjonneland. A. (201 Ib). Lung cancer incidence and long-term exposure to air pollution from traffic. Environ
Health Perspect 119: 860-865. http://dx.doi.org/10.1289/ehp.1002353
Raaschou-Nielsen. O: Andersen. ZJ: Jensen. SS: Ketzel M: Serensen. M: Hansen. J: Loft. S: Tjenneland. A:
Overvad. K. (2012). Traffic air pollution and mortality from cardiovascular disease and all causes: A Danish
cohort study. Environ Health 11: 60. http://dx.doi.org/10.1186/1476-069X-ll-60
Raaschou-Nielsen. O: Bak. H: Serensen. M: Jensen. SS: Ketzel M: Hvidberg. M: Schnohr. P: Tjonneland. A:
Overvad. K: Loft S. (2010a). Air pollution from traffic and risk for lung cancer in three Danish cohorts.
Cancer Epidemiol Biomarkers Prev 19: 1284-1291. http://dx.doi.org/10.1158/1055-9965.EPI-10-0036
Raaschou-Nielsen. O: Hermansen. MN: Loland. L: Buchvald. F: Pipper. CB: Sorensen. M: Loft. S: Bisgaard. H.
(2010b). Long-term exposure to indoor air pollution and wheezing symptoms in infants. Indoor Air 20: 159-
167. http://dx.doi.org/10.1111/i. 1600-0668.2009.00635.X
Raaschou-Nielsen. O: Hertel O: Vignati E: Berkowicz. R: Jensen. SS: Larsen. VB: Lohse. C: Olsen. JH.
(2000). An air pollution model for use in epidemiological studies: Evaluation with measured levels of
nitrogen dioxide and benzene. J Expo Anal Environ Epidemiol 10: 4-14.
http://dx.doi.org/10.1038/sj.jea.7500070
Raaschou-Nielsen. O: Sorensen. M: Ketzel M: Hertel O: Loft. S: Tjenneland. A: Overvad. K: Andersen. ZJ.
(2013b). Long-term exposure to traffic-related air pollution and diabetes-associated mortality: A cohort
study. Diabetologia 56: 36-46. http://dx.doi.org/10.1007/s00125-012-2698-7
Raff. JD: Niegic. B: Chang. WL: Gordon. MS: Dabdub. D: Gerber. RB: Finlavson-Pitts. BJ. (2009). Chlorine
activation indoors and outdoors via surface-mediated reactions of nitrogen oxides with hydrogen chloride.
PNAS 106: 13647-13654. http://dx.doi.org/10.1073/pnas.0904195106
Rage. E: Jacquemin. B: Nadif. R: Oryszczyn. MP: Siroux. V: Aguilera. I: Kauffmann. F: Kiinzlj N. (2009).
Total serum IgE levels are associated with ambient ozone concentration in asthmatic adults. Allergy 64: 40-
46. http://dx.doi.org/10.1111/i. 1398-9995.2008.01800.X
Ragettli MS: Tsai. MY: Braun-Fahrlander. C: de Nazelle. A: Schindler. C: Ineichen. A: Ducret-Stich. RE:
Perez. L: Probst-Hensch. N: Kiinzli. N: Phuleria. HC. (2014). Simulation of population-based commuter
exposure to NO using different air pollution models. Int J Environ Res Public Health 11: 5049-5068.
http://dx.doi.org/10.3390/ijerphll0505049
Rahmalia. A: Giorgis-Allemand. L: Lepeule. J: Philippat C: Galineau. J: Hulin. A: Charles. MA: Slama. R.
(2012). Pregnancy exposure to atmospheric pollutants and placenta! weight: An approach relying on a
dispersion model. Environ Int 48: 47-55. http://dx.doi.0rg/10.1016/i.envint.2012.06.013
R-75
-------
Ramos-Bonilla. JP: Brevsse. PN: Dominici F: Geyh. A: Tankerslev. CG. (2010). Ambient air pollution alters
heart rate regulation in aged mice. Inhal Toxicol 22: 330-339. http://dx.doi.org/10.3109/08958370903349365
Ranzj A: Gambinj M: Spattinj A: Galassi C: Sesti. D: Bedeschi. M: Messori. A: Baroni. A: Cavagni. G:
Lauriola. P. (2004). Air pollution and respiratory status in asthmatic children: Hints for a locally based
preventive strategy AIRE study. Eur J Epidemiol 19: 567-576.
Ranzj A: Porta. D: Badaloni. C: Cesaroni. G: Lauriola. P: Davoli M: Forastiere. F. (2014). Exposure to air
pollution and respiratory symptoms during the first 7 years of life in an Italian birth cohort. Occup Environ
Med 71: 430-436. http://dx.doi.org/10.1136/oemed-2013-101867
Rao. KS: Gunter. RL: White. JR: Hosker. RP. (2002). Turbulence and dispersion modeling near highways.
Atmos Environ 36: 4337-4346. http://dx.doi.org/10.1016/S1352-2310(02)00353-9
Rao. M: George. LA: Rosenstiel TN: Shandas. V: Dinno. A. (2014). Assessing the relationship among urban
trees, nitrogen dioxide, and respiratory health. Environ Pollut 194: 96-104.
http://dx.doi.0rg/10.1016/i.envpol.2014.07.011
Rasmussen. TR: Kjaergaard. SK: Tarp. U: Pedersen. OF. (1992). Delayed effects of NO2 exposure on alveolar
permeability and glutathione peroxidase in healthy humans. Am Rev Respir Dis 146: 654-659.
http://dx.doi.0rg/10.1164/airccm/146.3.654
Rava. M: Varraso. R: Decoster. B: Huyvaert. H: Le Moual N: Jacquemin. B: Kiinzli. N: Kauffmann. F:
Zerimech. F: Matran. R: Nadif. R. (2012). Plasma and exhaled breath condensate nitrite-nitrate level in
relation to environmental exposures in adults in the EGEA study. Nitric Oxide 27: 169-175.
http://dx.doi.0rg/10.1016/i.niox.2012.06.007
Ravsonj AU: Sarnat JA: Sarnat SE: Garcia. JH: Holguin. F: Luevano. SF: Li. WW. (2011). Binational school-
based monitoring of traffic-related air pollutants in El Paso, Texas (USA) and Ciudad Juarez, Chihuahua
(Mexico). Environ Pollut 159: 2476-2486. http://dx.doi.0rg/10.1016/i.envpol.2011.06.024
Reddel HK: Taylor. PR: Bateman. ED: Boulet LP: Boushev. HA: Busse. WW: Casale. TB: Chanez. P: Enright
PL: Gibson. PG: de Jongste. JC: Kerstjens. HA: Lazarus. SC: Lew. ML: O'Byrne. PM: Partridge. MR:
Pavord. ID: Sears. MR: Sterk. PJ: Stoloff. SW: Sullivan. SD: Szefler. SJ: Thomas. MD: Wenzel. SE. (2009).
An official American Thoracic Society/European Respiratory Society statement: Asthma control and
exacerbations: Standardizing endpoints for clinical asthma trials and clinical practice. Am J Respir Crit Care
Med 180: 59-99. http://dx.doi.org/10.1164/rccm.200801-060ST
Reeves. GK: Cox. PR: Darby. SC: Whitlev. E. (1998). Some aspects of measurement error in explanatory
variables for continuous and binary regression models. Stat Med 17: 2157-2177.
http://dx.doi.org/10.1002/(SICI)1097-0258(19981015)17:19<2157::AID-SIM916>3.0.CO:2-F
Ren. C: Fang. S: Wright RO: Suh. H: Schwartz. J. (2011). Urinary 8-hydroxy-2'-deoxyguanosine as abiomarker
of oxidative DNA damage induced by ambient pollution in the Normative Aging Study. Occup Environ Med
68: 562-569. http://dx.doi.org/10.1136/oem.2010.056358
Ren. X: vanDuin. D: Cazorla. M: Chen. S: Mao. J: Zhang. L. i: Brune. WH: Flynn. JH: Grossberg. N: Lefer.
BL: Rappenglueck. B: Wong. K: Tsai C: Stutz. J: Dibb. JE: Jobson. BT: Luke. WT: Kellev. P. (2013).
Atmospheric oxidation chemistry and ozone production: Results from SHARP 2009 in Houston, Texas. J
Geophys Res Atmos 118: 5770-5780. http://dx.doi.org/10.1002/igrd.50342
Rice. MB: Ljungman. PL: Wilker. EH: Gold. PR: Schwartz. JD: Koutrakis. P: Washko. GR: O'Connor. GT:
Mittleman. MA. (2013). Short-term exposure to air pollution and lung function in the Framingham heart
study. Am J Respir Crit Care Med 188: 1351-1357. http://dx.doi.org/10.1164/rccm.201308-1414OC
Rich. DO: Demissie. K: Lu. SE: Kamat L: Wartenberg. D: Rhoads. GG. (2009). Ambient air pollutant
concentrations during pregnancy and the risk of fetal growth restriction. J Epidemiol Community Health 63:
488-496. http://dx.doi.org/10.1136/iech.2008.082792
Rich. DO: Ripen. HM: Huang. W: Wang. G: Wang. Y: Zhu. P: Ohman-Strickland. P: Hu. M: Philipp. C: Diehl
SR: Lu. SE: Tong. J: Gong. J: Thomas. D: Zhu. T: Zhang. JJ. (2012). Association between changes in air
pollution levels during the Beijing Olympics and biomarkers of inflammation and thrombosis in healthy
young adults. JAMA 307: 2068-2078. http://dx.doi.org/10.1001/iama.2012.3488
R-76
-------
Rich. DQ: Kipen. HM: Zhang. J: Kamat L: Wilson. AC: Kostis. JB. (2010). Triggering of transmural
infarctions, but not nontransmural infarctions, by ambient fine particles. Environ Health Perspect 118: 1229-
1234. http://dx.doi.org/10.1289/ehp.0901624
Richmond-Bryant J: Reff. A. (2012). Air pollution retention within a complex of urban street canyons: A two-
city comparison. Atmos Environ49: 24-32. http://dx.doi.0rg/10.1016/i.atmosenv.2011.12.036
Richters. A: Damji KS. (1990). The relationship between inhalation of nitrogen dioxide, the immune system,
and progression of a spontaneously occurring lymphoma in AKR mice. J Environ Pathol Toxicol Oncol 10:
225-230.
Richters. A: Kuraitis. K. (1981). Inhalation of NO2 and blood borne cancer cell spread to the lungs. Arch
Environ Health 36: 36-39.
Richters. A: Kuraitis. K. (1983). Air pollutants and the facilitation of cancer metastasis. Environ Health Perspect
52: 165-168.
Richters. A: Richters. V: Alley. WP. (1985). The mortality rate from lung metastases in animals inhaling
nitrogen dioxide (NO2). J Surg Oncol 28: 63-66. http://dx.doi.org/10.1002/iso.2930280115
Riedel TP: Wolfe. GM: Danas. KT: Oilman. JB: Kuster. WC: Boa DM: Vlasenko. A: Li. SM: Williams. EJ:
Lerner. BM: Veres. PR: Roberts. JM: Hollowav. JS: Lefer. B: Brown. SS: Thornton. JA. (2014). AnMCM
modeling study of nitryl chloride (C1NO2) impacts on oxidation, ozone production and nitrogen oxide
partitioning in polluted continental outflow. Atmos Chem Phys 14: 3789-3800. http://dx.doi.org/10.5194/acp-
14-3789-2014
Riedl. MA: Diaz-Sanchez. D: Linn. WS: Gong. H. Jr: Clark. KW: Effros. RM: Miller. JW: Cocker. PR:
Berhane. KT. (2012). Allergic inflammation in the human lower respiratory tract affected by exposure to
diesel exhaust [HEI] (pp. 5-43; discussion 45-64). (ISSN 1041-5505 Research Report 165). Boston, MA:
Health Effects Institute. http://pubs.healtheffects.org/view.php?id=373
Ritz. B: Qiu. J: Lee. PC: Lurmann. F: Penfold. B: Erin Weiss. R: Mcconnell R: Arora. C: Hobel C: Wilhelm.
M. (2014). Prenatal air pollution exposure and ultrasound measures of fetal growth in Los Angeles,
California. Environ Res 130: 7-13. http://dx.doi.0rg/10.1016/i.envres.2014.01.006
Ritz. B: Wilhelm. M. (2008). Ambient air pollution and adverse birth outcomes: Methodologic issues in an
emerging field [Review]. Basic Clin Pharmacol Toxicol 102: 182-190. http://dx.doi.org/10.1111/i. 1742-
7843.2007.00161.x
Ritz. B: Wilhelm. M: Hoggatt KJ: Ghosh. JK. (2007). Ambient air pollution and preterm birth in the
environment and pregnancy outcomes study at the University of California, Los Angeles. Am J Epidemiol
166: 1045-1052. http://dx.doi.org/10.1093/aie/kwml81
Ritz. B: Wilhelm. M: Zhao. Y. (2006). Air pollution and infant death in southern California, 1989-2000.
Pediatrics 118: 493-502. http://dx.doi.org/10.1542/peds.2006-0027
Ritz. B: Yu. F: Chapa. G: Fruin. S. (2000). Effect of air pollution on preterm birth among children born in
Southern California between 1989 and 1993. Epidemiology 11: 502-511.
Rivas. I: Viana. M: Moreno. T: Pandolfi. M: Amato. F: Reche. C: Bouso. L: Alvarez-Pedrerol M: Alastuev. A:
Sunyer. J: Querol X. (2014). Child exposure to indoor and outdoor air pollutants in schools in Barcelona,
Spain. Environ Int 69: 200-212. http://dx.doi.0rg/10.1016/i.envint.2014.04.009
Rivera. M: Basagana. X: Aguilera. I: Foraster. M: Agis. D: de Groot E: Perez. L: Mendez. MA: Bouso. L:
Targa. J: Ramos. R: Sala. J: Marrugat. J: Elosua. R: Kiinzli. N. (2013). Association between long-term
exposure to traffic-related air pollution and subclinical atherosclerosis: The REGICOR study. Environ Health
Perspect 121: 223-230. http://dx.doi.org/10.1289/ehp.1205146
Roberts. JM: Veres. PR: Vandenboer. TC: Warneke. C: Graus. M: Williams. EJ: Lefer. B: Brock. CA: Bahreini.
R: Ozturk. F: Middlebrook. AM: Wagner. NL: Dube. WP: De Gouw. JA. (2014). New insights into
atmospheric sources and sinks of isocyanic acid, HNCO, from recent urban and regional observations. J
Geophys Res Atmos 119: 1060-1072. http://dx.doi.org/10.1002/2013JD019931
R-77
-------
Roberts. S: Martin. MA. (2006). Using supervised principal components analysis to assess multiple pollutant
effects. Environ Health Perspect 114: 1877-1882. http://dx.doi.org/10.1289/ehp.9226
Robinson. DP: Lloyd. CD: Mckinlev. JM. (2013). Increasing the accuracy of nitrogen dioxide (NO2) pollution
mapping using geographically weighted regression (GWR) and geostatistics. Int J Appl Earth Obs Geoinf 21:
374-383. http://dx.doi.0rg/10.1016/i.jag.2011.ll.001
Robison. TW: Duncan. DP: Forman. HJ. (1990). Chemoattractant and leukotriene B4 production from rat
alveolar macrophages exposed to nitrogen dioxide. Am J Respir Cell Mol Biol 3: 21-26.
http://dx.doi.0rg/10.1165/aircmb/3.l.21
Robison. TW: Forman. HJ. (1993). Dual effect of nitrogen dioxide on rat alveolar macrophage arachidonate
metabolism. Exp Lung Res 19: 21-36.
Robison. TW: Murphy. JK: Beyer. LL: Richters. A: Forman. HJ. (1993). Depression of stimulated arachidonate
metabolism and superoxide production in rat alveolar macrophages following in vivo exposure to 0.5 ppm
NO2. J Toxicol Environ Health 38: 273-292. http://dx.doi.org/10.1080/15287399309531718
Rodes. CE: Holland. DM. (1981). Variations of NO, NO2 and O3 concentrations downwind of a Los Angeles
freeway. Atmos Environ 15: 243-250. http://dx.doi.org/10.1016/0004-6981(81)90024-X
Rodgers. MO: Davis. DP. (1989). A UV-photofragmentation/laser-induced fluorescence sensor for the
atmospheric detection of HONO. Environ Sci Technol 23: 1106-1112.
Rodriguez. C: Tonkin. R: Hevworth. J: Kusel M: De Klerk. N: Sly. PD: Franklin. P: Runnion. T: Blocklev. A:
Landau. L: Hinwood. AL. (2007). The relationship between outdoor air quality and respiratory symptoms in
young children. Int J Environ Health Res 17: 351-360. http://dx.doi.org/10.1080/09603120701628669
Roemer. W: Hoek. G: Brunekreef. B: Haluszka. J: Kalandidi. A: Pekkanen. J. (1998). Daily variations in air
pollution and respiratory health in a multicentre study: The PEACE project. Eur Respir J 12: 1354-1361.
Roger. LJ: Horstman. DH: McDonnell W: Kehrl. H: Ives. PJ: Seal. E: Chapman. R: Massaro. E. (1990).
Pulmonary function, airway responsiveness, and respiratory symptoms in asthmatics following exercise in
NO2. Toxicol Ind Health 6: 155-171. http://dx.doi.org/10.1177/074823379000600110
Rojas-Martinez. R: Perez-Padilla. R: Olaiz-Fernandez. G: Mendoza-Alvarado. L: Moreno-Macias. H: Fortoul T:
McDonnell W: Loomis. D: Romieu. I. (2007a). Lung function growth in children with long-term exposure to
air pollutants in Mexico City. Am J Respir Crit Care Med 176: 377-384.
http://dx.doi.org/10.1164/rccm.200510-1678OC
Rojas-Martinez. R: Perez-Padilla. R: Olaiz-Fernandez. G: Mendoza-Alvarado. L: Moreno-Macias. H: Fortoul. T:
McDonnell. W: Loomis. D: Romieu. I. (2007b). Lung function growth in children with long-term exposure
to air pollutants in Mexico City. Online data supplement [Supplemental Data]. Am J Respir Crit Care Med
[176].
Rombout PJ: Dormans. JA: Marra. M: Van Esch. GJ. (1986). Influence of exposure regimen on nitrogen
dioxide-induced morphological changes in the rat lung. Environ Res 41: 466-480.
Romero. R: Espinoza. J: Kusanovic. JP: Gotsch. F: Hassan. S: Erez. O: Chaiworapongsa. T: Mazor. M. (2006).
The preterm parturition syndrome [Review]. BJOG 113: 17-42. http://dx.doi.0rg/10.l 111/i. 1471-
0528.2006.01120.x
Romieu. I: Barraza-Villarreal. A: Escamilla-Nunez. C: Almstrand. AC: Diaz-Sanchez. D: Sly. PD: Olin. AC.
(2008). Exhaled breath malondialdehyde as a marker of effect of exposure to air pollution in children with
asthma. J Allergy Clinlmmunol 121: 903-909. http://dx.doi.0rg/10.1016/i.jaci.2007.12.004
Romieu. I: Ramirez-Aguilar. M: Sienra-Monge. JJ: Moreno-Macias. H: Del Rio-Navarro. BE: David. G:
Marzec. J: Hernandez-Avila. M: London. S. (2006). GSTM1 and GSTP1 and respiratory health in asthmatic
children exposed to ozone. Eur Respir J 28: 953-959. http://dx.doi.org/10.1183/09031936.06.00114905
Roonev. AA: Boyles. AL: Wolfe. MS: Bucher. JR: Thaver. KA. (2014). Systematic review and evidence
integration for literature-based environmental health science assessments. Environ Health Perspect 122: 711-
718. http://dx.doi.org/10.1289/ehp.1307972
R-78
-------
Roorda-Knape. MC: Janssen. NAH: De Hartog. JJ: Van Vliet PHN: Harssema. H: Brunekreef. B. (1998). Air
pollution from traffic in city districts near major motorways. Atmos Environ 32: 1921-1930.
http://dx.doi.org/10.1016/S1352-2310(97)00496-2
Rose. N: Cowie. C: Gillett R: Marks. GB. (2011). Validation of a spatiotemporal land use regression model
incorporating fixed site monitors. Environ Sci Technol 45: 294-299. http://dx.doi.org/10.1021/esl00683t
Rose. RM: Fuglestad. JM: Skornik. WA: Hammer. SM: Wolfthal SF: Beck. BD: Brain. JD. (1988). The
pathophysiology of enhanced susceptibility to murine cytomegalovirus respiratory infection during short-
term exposure to 5 ppm nitrogen dioxide. Am Rev Respir Dis 137: 912-917.
http://dx.doi.0rg/10.1164/airccm/137.4.912
Rose. RM: Pinkston. P: Skornik. WA. (1989a). Altered susceptibility to viral respiratory infection during short-
term exposure to nitrogen dioxide. (HEI Research Report no. 24). Cambridge, MA: Health Effects Institute.
Rose. RM: Pinkston. P: Skornik. WA. (1989b). Altered susceptibility to viral respiratory infection during short-
term exposure to nitrogen dioxide. Res Rep Health Eff Instl-24.
Rosenlund. M: Bellander. T: Nordquist T: Alfredsson. L. (2009a). Traffic-generated air pollution and
myocardial infarction. Epidemiology 20: 265-271. http://dx.doi.org/10.1097/EDE.Ob013e318190ea68
Rosenlund. M: Forastiere. F: Porta. D: De Sario. M: Badaloni. C: Perucci. CA. (2009b). Traffic-related air
pollution in relation to respiratory symptoms, allergic sensitisation and lung function in schoolchildren.
Thorax 64: 573-580. http://dx.doi.org/10.1136/thx.2007.094953
Rosenlund. M: Forastiere. F: Stafoggia. M: Porta. D: Perucci. M: Ranzi. A: Nussio. F: Perucci. C. (2008a).
Comparison of regression models with land-use and emissions data to predict the spatial distribution of
traffic-related air pollution in Rome. JExpo Sci Environ Epidemiol 18: 192-199.
http://dx.doi.org/10.1038/si.ies.7500571
Rosenlund. M: Picciotto. S: Forastiere. F: Stafoggia. M: Perucci. CA. (2008b). Traffic-related air pollution in
relation to incidence and prognosis of coronary heart disease. Epidemiology 19: 121-128.
http://dx.doi.org/10.1097/EDE.Ob013e31815cl921
Ross. Z: Ito. K: Johnson. S: Yee. M: Pezeshki. G: Cloughertv. JE: Savitz. D: Matte. T. (2013). Spatial and
temporal estimation of air pollutants in New York City: Exposure assignment for use in a birth outcomes
study. Environ Health 12. http://dx.doi.org/10.1186/1476-069X-12-51
Ross. Z: Kheirbek. I: Cloughertv. JE: Ito. K: Matte. T: Markowitz. S: Eisl H. (2011). Noise, air pollutants and
traffic: Continuous measurement and correlation at a high-traffic location in New York City. Environ Res
111: 1054-1063. http://dx.doi.0rg/10.1016/i.envres.2011.09.004
Rothman. KJ: Greenland. S. (1998). Modern epidemiology. In Modern Epidemiology (2nd ed.). Philadelphia,
PA: Lippincott, Williams, & Wilkins.
Rotko. T: Kousa. A: Aim. S: Jantunen. M. (2001). Exposures to nitrogen dioxide in EXPOLIS-Helsinki:
Microenvironment, behavioral and sociodemographic factors. J Expo Anal Environ Epidemiol 11: 216-223.
http://dx.doi.org/10.1038/si.iea.7500162
Roux. D: Frev. C. (2015). CASAC review of the EPAs integrated science assessment for oxides of nitrogen
Health criteria (Second external review draft January 2015) [EPA Report]. (EPA-CASAC-15-001).
Washington, DC: U.S. Environmental Protection Agency.
http://YOsemite.epa.gov/sab/sabproduct.nsf/264cbl227d55e02c85257402007446a4/6612DAF24438687B852
57EBB0070369C/$File/EPA-CASAC-15-001+unsigned.pdf
Rowangould. GM. (2013). A census of the US near-roadway population: Public health and environmental justice
considerations. Transport Res Transport Environ 25: 59-67. http://dx.doi.0rg/10.1016/i.trd.2013.08.003
Rubenchik. BL: Glavin. AA: Galenko. PM: Kilkichko. AA: Olenick. IP: Artemov. KV. (1995). Gaseous
nitrogen dioxide increases the endogenous synthesis of carcinogenic N-nitrosodimethylamine in animals. J
Environ Pathol Toxicol Oncol 14: 111-115.
R-79
-------
Rubes. J: Rybar. R: Prinosilova. P: Veznik. Z: Chvatalova. I: Solansky. I: Sram. RJ. (2010). Genetic
polymorphisms influence the susceptibility of men to sperm DNA damage associated with exposure to air
pollution. MutatRes 683: 9-15. http://dx.doi.0rg/10.1016/i.mrfmmm.2009.09.010
Rubinstein. I: Bigbv. BG: Reiss. TF: Boushev. HA. Jr. (1990). Short-term exposure to 0.3 ppm nitrogen dioxide
does not potentiate airway responsiveness to sulfur dioxide in asthmatic subjects. Am Rev Respir Dis 141:
381-385. http://dx.doi.0rg/10.1164/airccm/141.2.381
Rubinstein. I: Reiss. TF: Bigbv. BG: Stites. DP: Boushev. HA. Jr. (1991). Effects of 0.60 PPM nitrogen dioxide
on circulating and bronchoalveolar lavage lymphocyte phenotypes in healthy subjects. Environ Res 55: 18-
30.
Rudez. G: Janssen. NA: Kilinc. E: Leebeek. FW: Gerlofs-Nijland. ME: Spronk. HM: ten Gate. H: Cassee. FR: de
Maat MP. (2009). Effects of ambient air pollution on hemostasis and inflammation. Environ Health Perspect
117: 995-1001. http://dx.doi.org/10.1289/ehp.0800437
Rudolph. V: Rudolph. TK: Schopfer. FJ: Bonacci G: Woodcock. SR: Cole. MP: Baker. PR: Ramani. R:
Freeman. BA. (2010). Endogenous generation and protective effects of nitro-fatty acids in a murine model of
focal cardiac ischaemia and reperfusion. Cardiovasc Res 85: 155-166. http://dx.doi.org/10.1093/cvr/cvp275
Russell AR: Perring. AE: Valin. LC: Bucsela. EJ: Browne. EC: Wooldridge. PJ: Cohen. RC. (2011). A high
spatial resolution retrieval of NO 2 column densities from OMI: method and evaluation. Atmos Chem Phys
11: 8543-8554. http://dx.doi.org/10.5194/acp-ll-8543-2011
Ryerson. TB: Andrews. AE: Angevine. WM: Bates. TS: Brock. CA: Cairns. B: Cohen. RC: Cooper. OR: de
Gouw. JA: Fehsenfeld. FC: Ferrare. RA: Fischer. ML: Flagan. RC: Goldstein. AH: Hair. JW: Hardestv. RM:
Hostetler. CA: Jimenez. JL: Langford. AO: McCaulev. E: McKeen. SA: Molina. LT: Nenes. A: Oltmans. SJ:
Parrish. DP: Pederson. J. R.: Pierce. RB: Prather. K: Quinn. PK: Seinfeld. JH: Senff. CJ: Sorooshian. A:
Stutz. J: Surratt JD: Trainer. M: Volkamer. R: Williams. EJ: Wofsv. SC. (2013). The 2010 California
Research at the Nexus of Air Quality and Climate Change (CalNex) field study. J Geophys Res Atmos 118:
5830-5866. http://dx.doi.org/10.1002/igrd.50331
Sacks. JD: Ito. K: Wilson. WE: Neas. LM. (2012). Impact of covariate models on the assessment of the air
pollution-mortality association in a single- and multipollutant context. Am J Epidemiol 176: 622-634.
http://dx.doi.org/10.1093/aie/kwsl35
Sacks. JD: Stanek. LW: Luben. TJ: Johns. DO: Buckley. BJ: Brown. JS: Ross. M. (2011). Particulate-matter
induced health effects: Who is susceptible? [Review]. Environ Health Perspect 119: 446-454.
http://dx.doi.org/10.1289/ehp. 1002255
Sagai. M: Arakawa. K: Ichinose. T: Shimojo. N. (1987). Biochemical effects on combined gases of nitrogen
dioxide and ozone: I. Species differences of lipid peroxides and phospholipids in lungs. Toxicology 46: 251-
265. http://dx.doi.org/10.1016/0300-483X(87)90207-1
Sagai. M: Ichinose. T: Kubota. K. (1984). Studies on the biochemical effects of nitrogen dioxide. IV. Relation
between the change of lipid peroxidation and the antioxidative protective system in rat lungs upon life span
exposure to low levels of NO2. Toxicol Appl Pharmacol 73: 444-456.
Sagai. M: Ichinose. T: Oda. H: Kubota. K. (1982). Studies on biochemical effects of nitrogen dioxide. II.
Changes of the protective systems in rat lungs and of lipid peroxidation by acute exposure. J Toxicol Environ
Health 9: 153-164. http://dx.doi.org/10.1080/15287398209530150
Sahsuvaroglu. T: Su. JG: Brook. J: Burnett R: Loeb. M: Jerrett M. (2009). Predicting personal nitrogen dioxide
exposure in an elderly population: Integrating residential indoor and outdoor measurements, fixed-site
ambient pollution concentrations, modeled pollutant levels, and time-activity patterns. J Toxicol Environ
HealthA72: 1520-1533. http://dx.doi.org/10.1080/15287390903129408
Saiani. SZ: Hanninen. O: Marchesi S: Lauriola. P. (2011). Comparison of different exposure settings in a case-
crossover study on air pollution and daily mortality: Counterintuitive results. J Expo Sci Environ Epidemiol
21: 385-394. http://dx.doi.org/10.1038/ies.2010.27
R-80
-------
Salam. MT: Byun. HM: Lurmann. F: Breton. CV: Wang. X: Eckel SP: Gilliland. FD. (2012). Genetic and
epigenetic variations in inducible nitric oxide synthase promoter, paniculate pollution, and exhaled nitric
oxide levels in children. J Allergy Clinlmmunol 129: 232-239.e237.
http://dx.doi.0rg/10.1016/i.iaci.2011.09.037
Salam. MT: Millstein. J: Li. YF: Lurmann. FW: Margolis. HG: Gilliland. FD. (2005). Birth outcomes and
prenatal exposure to ozone, carbon monoxide, and paniculate matter: Results from the Children's Health
Study. Environ Health Perspect 113: 1638-1644. http://dx.doi.org/10.1289/ehp.8111
Salonen. RO: Pennanen. AS: Vahteristo. M: Korkeila. P: Aim. S: Randell JT. (2008). Health risk assessment of
indoor air pollution in Finnish ice arenas. Environ Int 34: 51-57.
http://dx.doi.0rg/10.1016/i.envint.2007.06.012
Saltzman. BE. (1954). Colorimetric microdetermination of nitrogen dioxide in the atmosphere. Anal Chem 26:
1949-1955.
Samet. JM. (2008). Letter from Dr. Jonathan M. Sametto The Honorable Stephen L. Johnson. Subject: Clean
Air Scientific Advisory Committees (CASAC) peer review of draft chapter 8 of EPAs risk and exposure
assessment to support the review of the NO2 primary National Ambient Air Quality Standard. Available
online at http://vosemite.epa.gov/sab/sabproduct.nsf/WebCASAC/Samet-NAAQSLtr-12-16-08/$File/Samet-
NAAQS-12-16-08.pdf
Samet JM: Lambert. WE: Skipper. BJ: Gushing. AH: Hunt. WC: Young. SA: McLaren. LC: Schwab. M:
Spengler. JD. (1993). Nitrogen dioxide and respiratory illness in children. Part I: Health outcomes. (HEI
Research Report no. 58). Cambridge, MA: Health Effects Institute.
Samoli. E: Aga. E: Touloumi. G: Nisiotis. K: Forsberg. B: Lefranc. A: Pekkanen. J: Woityniak. B: Schindler. C:
Niciu. E: Brunstein. R: Dodic Fikfak. M: Schwartz. J: Katsouvanni. K. (2006). Short-term effects of nitrogen
dioxide on mortality: An analysis within the APHEA project. Eur Respir J 27: 1129-1137.
http://dx.doi.org/10.1183/09031936.06.00143905
Samoli. E: Nastos. PT: Paliatsos. AG: Katsouvanni. K: Priftis. KN. (2011). Acute effects of air pollution on
pediatric asthma exacerbation: Evidence of association and effect modification. Environ Res 111: 418-424.
http://dx.doi.0rg/10.1016/i.envres.2011.01.014
Sampson. PD: Szpiro. AA: Sheppard. L: Lindstrom. J: Kaufman. JD. (2011). Pragmatic estimation of a spatio-
temporal air quality model with irregular monitoring data. Atmos Environ 45: 6593-6606.
http://dx.doi.0rg/10.1016/i.atmosenv.2011.04.073
Sanchez Jimenez. A: Heal MR: Beverland. IJ. (2011). Intercomparison study of NOx passive diffusion tubes
with chemiluminescence analysers and evaluation of bias factors. Atmos Environ 45: 3062-3068.
http://dx.doi.0rg/10.1016/i.atmosenv.2011.03.011
Sanchez Jimenez. A: Heal MR: Beverland. IJ. (2012). Correlations of particle number concentrations and metals
with nitrogen oxides and other traffic-related air pollutants in Glasgow and London. Atmos Environ 54: 667-
678. http://dx.doi.0rg/10.1016/j.atmosenv.2012.01.047
Sander. SP: Abbatt JPD: Barker. JR: Burkholder. JB: Friedl RR: Golden. DM: Huie. RE: Kolb. CE: Kurylo.
MJ: Moortgat GK: Orkin. VL: Wine. PH. (2011). Section 1. Biomolecular reactions. In Chemical kinetics
and photochemical data for use in atmospheric studies: Evaluation number 17 (JPL Publication 10-6) (pp. 1-1
-1-34). Pasadena, CA: Jet Propulsion Laboratory.
Sandstrom. T: Andersson. MC: Kormodin-Hedman. B: Stiernberg. N: Angstrom. T. (1990). Bronchoalveolar
mastocytosis and lymphocytosis after nitrogen dioxide exposure in man: A time-kinetic study. Eur Respir J
3: 138-143.
Sandstrom. T: Helledav. R: Bjermer. L: Stiernberg. N. (1992a). Effects of repeated exposure to 4 ppm nitrogen
dioxide on bronchoalveolar lymphocyte subsets and macrophages in healthy men. Eur Respir J 5: 1092-1096.
Sandstrom. T: Ledin. MC: Thomasson. L: Helledav. R: Stiernberg. N. (1992b). Reductions in lymphocyte
subpopulations after repeated exposure to 1.5 ppm nitrogen dioxide. Br J Ind Med 49: 850-854.
R-81
-------
Sandstrom. T: Stjernberg. N: Eklund. A: Ledin. MC: Bjermer. L: Kolmodin-Hedman. B: Lindstrom. K:
Rosenhall L: Angstrom. T. (1991). Inflammatory cell response in bronchoalveolar lavage fluid after nitrogen
dioxide exposure of healthy subjects: A dose-response study. Eur Respir J 4: 332-339.
Sarangapani. R: Gentry. PR: Covingtoa TR: Teeguarden. JG: Clewell HJ. III. (2003). Evaluation of the
potential impact of age- and gender-specific lung morphology and ventilation rate on the dosimetry of
vapors. Inhal Toxicol 15: 987-1016. http://dx.doi.org/10.1080/08958370390226350
Sarnat JA: Brown. KW: Schwartz. J: Coull BA: Koutrakis. P. (2005). Ambient gas concentrations and personal
paniculate matter exposures: Implications for studying the health effects of particles. Epidemiology 16: 385-
395. http://dx.doi.org/10.1097/01.ede.0000155505.04775.33
Sarnat. JA: Koutrakis. P: Suh. HH. (2000). Assessing the relationship between personal paniculate and gaseous
exposures of senior citizens living in Baltimore, MD. J Air Waste Manag Assoc 50: 1184-1198.
http://dx.doi.org/10.1080/10473289.2000.10464165
Sarnat. JA: Sarnat SE: Flanders. WD: Chang. HH: Mulholland. J: Baxter. L: Isakov. V: Ozkaynak. H. (2013a).
Spatiotemporally resolved air exchange rate as a modifier of acute air pollution-related morbidity in Atlanta.
J Expo Sci Environ Epidemiol 23: 606-615. http://dx.doi.org/10.1038/ies.2013.32
Sarnat JA: Schwartz. J: Catalano. PJ: Suh. HH. (2001). Gaseous pollutants in paniculate matter epidemiology:
Confounders or surrogates? Environ Health Perspect 109: 1053-1061.
Sarnat. SE: Coull. BA: Ruiz. PA: Koutrakis. P: Suh. HH. (2006). The influences of ambient particle composition
and size on particle infiltration in Los Angeles, CA, residences. J Air Waste Manag Assoc 56: 186-196.
http://dx.doi.org/10.1080/10473289.2006.10464449
Sarnat. SE: Klein. M: Sarnat JA: Flanders. WD: Waller. LA: Mulholland. JA: Russell AG: Tolbert. PE. (2010).
An examination of exposure measurement error from air pollutant spatial variability in time-series studies. J
Expo Sci Environ Epidemiol 20: 135-146. http://dx.doi.org/10.1038/ies.20Q9.10
Sarnat. SE: Ravsoni. AU: Li. WW: Holguin. F: Johnson. BA: Flores Luevano. S: Garcia. JH: Sarnat JA. (2012).
Air pollution and acute respiratory response in a panel of asthmatic children along the U. S.-Mexico border.
Environ Health Perspect 120: 437444. http://dx.doi.org/10.1289/ehp. 1003169
Sarnat. SE: Sarnat. JA: Mulholland. J: Isakov. V: Ozkaynak. H: Chang. HH: Klein. M: Tolbert. PE. (2013b).
Application of alternative spatiotemporal metrics of ambient air pollution exposure in a time-series
epidemiological study in Atlanta. J Expo Sci Environ Epidemiol 23: 593-605.
http://dx.doi.org/10.1038/ies.2013.41
Sarwar. G: Corsi R: Kumura. Y: Allen. D: Weschler. CJ. (2002). Hydroxyl radicals in indoor environments.
Atmos Environ 36: 3973-3988. http://dx.doi.org/10.1016/S1352-2310(02)00278-9
Sather. ME: Slonecker. ET: Mathew. J: Daughtrev. H: Williams. DP. (2007). Evaluation of ogawa passive
sampling devices as an alternative measurement method for the nitrogen dioxide annual standard in El Paso,
Texas. Environ Monit Assess 124: 211-221. http://dx.doi.org/10.1007/sl0661-006-9219-4
Sathvanaravana. S: Zhou. C: Rudra. CB: Gould. T: Larson. T: Koenig. J: Karr. CJ. (2013). Prenatal ambient air
pollution exposure and small for gestational age birth in the Puget Sound Air Basin. Air Qua! Atmos Health
6: 455-463. http://dx.doi.org/10.1007/sll869-012-0182-7
Savitz. DA: Bobb. JF: Carr. JL: Cloughertv. JE: Dominici F: Elston. B: Ito. K: Ross. Z: Yee. M: Matte. TD.
(2014). Ambient fine paniculate matter, nitrogen dioxide, and term birth weight in New York, New York.
Am J Epidemiol 179: 457-466. http://dx.doi.org/10.1093/aje/kwt268
Scaife. A: Barclay. J: Hillis. GS: Srinivasan. J: MacDonald. DW: Ross. JA: Ayres. JG. (2012). Lack of effect of
nitrogen dioxide exposure on heart rate variability in patients with stable coronary heart disease and impaired
left ventricular systolic function. Occup Environ Med 69: 587-591. http://dx.doi.org/10.1136/oemed-2011-
100126
Scarlett. JF: Abbott. KJ: Peacock. JL: Strachan. DP: Anderson. HR. (1996). Acute effects of summer air
pollution on respiratory function in primary school children in southern England. Thorax 51:1109-1114.
http://dx.doi.org/10.1136/thx.51.ll.1109
R-82
-------
Schembari. A: Nieuwenhuijsen. MJ: Salvador. J: de Nazelle. A: Cirach. M: Dadvand. P: Beelen. R: Hoek. G:
Basagana. X: Vrijheid. M. (2014). Traffic-related air pollution and congenital anomalies in Barcelona.
Environ Health Perspect 122: 317-323. http://dx.doi.org/10.1289/ehp. 1306802
Schembari. A: Triguero-Mas. M: De Nazelle. A: Dadvand. P: Vrijheid. M: Cirach. M: Martinez. D: Figueras. F:
Querol X: Basagafia. X: Eeftens. M: Meliefste. K: Nieuwenhuijsen. MJ. (2013). Personal, indoor and
outdoor air pollution levels among pregnant women. Atmos Environ 64: 287-295.
http://dx.doi.0rg/10.1016/i.atmosenv.2012.09.053
Schikowski T: Sugiri. D: Ranft U: Gehring. U: Heinrich. J: Wichmann. HE: Kramer. U. (2007). Does
respiratory health contribute to the effects of long-term air pollution exposure on cardiovascular mortality?
RespirRes 8: 20. http://dx.doi.org/10.1186/1465-9921-8-20
Schikowski. T: Sugiri. D: Ranft U: Gehring. U: Heinrich. J: Wichmann. HE: Kramer. U. (2005). Long-term air
pollution exposure and living close to busy roads are associated with COPD in women. Respir Res 22: 152-
161. http://dx.doi.org/10.1186/1465-9921-6-152
Schildcrout JS: Sheppard. L: Lumlev. T: Slaughter. JC: Koenig. JQ: Shapiro. GG. (2006). Ambient air pollution
and asthma exacerbations in children: An eight-city analysis. Am J Epidemiol 164: 505-517.
http://dx.doi.org/10.1093/aie/kwj225
Schindler. C: Kiinzli. N: Bongard. JP: Leuenberger. P: Karrer. W: Rapp. R: Monn. C: Ackermann-Liebrich. U.
(2001). Short-term variation in air pollution and in average lung function among never-smokers. The Swiss
Study on Air Pollution and Lung Diseases in Adults (SAPALDIA). Am J Respir Crit Care Med 163: 356-
361. http://dx.doi.0rg/10.1164/airccm.163.2.9911116
Schlesinger. RB. (1987a). Effects of intermittent inhalation exposures to mixed atmospheres of NO2 and H2SO4
on rabbit alveolar macrophages. J Toxicol Environ Health 22: 301-312.
http://dx.doi.org/10.1080/15287398709531073
Schlesinger. RB. (1987b). Intermittent inhalation of nitrogen dioxide: Effects on rabbit alveolar macrophages. J
Toxicol Environ Health 21: 127-139. http://dx.doi.org/10.1080/15287398709531007
Schlesinger. RB: Driscoll KE: Gunnison. AF: Zelikoff. JT. (1990). Pulmonary arachidonic acid metabolism
following acute exposures to ozone and nitrogen dioxide. J Toxicol Environ Health 31: 275-290.
http://dx.doi.org/10.1080/15287399009531456
Schlesinger. RB: Driscoll KE: Vollmuth. TA. (1987). Effect of repeated exposures to nitrogen dioxide and
sulfuric acid mist alone or in combination on mucociliary clearance from the lungs of rabbits. Environ Res
44: 294-301. http://dx.doi.org/10.1016/S0013-9351(87)80238-4
Schlesinger. RB: Gearhart JM. (1987). Intermittent exposures to mixed atmospheres of nitrogen dioxide and
sulfuric acid: effect on particle clearance from the respiratory region of rabbit lungs. Toxicology 44: 309-319.
http://dx.doi.org/10.1016/0300-483X(87)90032-1
Schultz. ES: Gruzieva. O: Bellander. T: Bottai M: Hallberg. J: Kull I: Svartengren. M: Melen. E: Pershagen. G.
(2012). Traffic-related air pollution and lung function in children at 8 years of age: A birth cohort study. Am
J Respir Crit Care Med 186: 1286-1291. http://dx.doi.org/10.1164/rccm.201206-1045OC
Schwartz. J: Coull. BA. (2003). Control for confounding in the presence of measurement error in hierarchical
models. Biostatistics 4: 539-553. http://dx.doi.Org/10.1093/biostatistics/4.4.539
Schwartz. J: Dockery. DW: Neas. LM: Wvpij. D: Ware. JH: Spengler. JD: Koutrakis. P: Speizer. FE: Ferris. BG.
Jr (1994). Acute effects of summer air pollution on respiratory symptom reporting in children. Am J Respir
Crit Care Med 150: 1234-1242. http://dx.doi.0rg/10.1164/airccm.150.5.7952546
Schwartz. SE: White. WH. (1983). Kinetics of reactive dissolution of nitrogen oxides into aqueous solution. In
SE Schwartz (Ed.), Trace atmospheric constituents: Properties, transformations, and fates (pp. 1-116). New
York, NY: Wiley.
Sedefian. L: Rao. ST: Czapski. U. (1981). Effects of traffic-generated turbulence on near-field dispersion. Atmos
Environ (1967) 15: 527-536. http://dx.doi.org/10.1016/0004-6981(81)90182-7
R-83
-------
See. KG: Christian!. DC. (2013). Normal values and thresholds for the clinical interpretation of exhaled nitric
oxide levels in the US general population: Results from the National Health and Nutrition Examination
Survey 2007-2010. Chest 143: 107-116. http://dx.doi.org/10.1378/chest.12-0416
Segala. C: Fauroux. B: Just J: Pascual. L: Grimfeld. A: Neukirch. F. (1998). Short-term effect of winter air
pollution on respiratory health of asthmatic children in Paris. Eur Respir J 11: 677-685.
Segala. C: Poizeau. D: Mesbah. M: Willems. S: Maidenberg. M. (2008). Winter air pollution and infant
bronchiolitis in Paris. Environ Res 106: 96-100. http://dx.doi.0rg/10.1016/i.envres.2007.05.003
Seilkop. SK: Campea MJ: Lund. AK: McDonald. JD: Mauderly. JL. (2012). Identification of chemical
components of combustion emissions that affect pro-atherosclerotic vascular responses in mice. Inhal
Toxicol 24: 270-287. http://dx.doi.org/10.3109/08958378.2012.667455
Seinfeld. JH: Pandis. SN. (1998). Atmospheric chemistry and physics: From air pollution to climate change.
New York: John Wiley & Sons, http://dx.doi.0rg/10.1063/l.882420
Sekine. K: Shima. M: Nitta. Y: Adachj M. (2004). Long term effects of exposure to automobile exhaust on the
pulmonary function of female adults in Tokyo, Japan. Occup Environ Med 61: 350-357.
http://dx.doi.org/10.1136/oem.2002.005934
Sekine. Y: Watts. SF: Rendell. A: Butsugan. M. (2008). Development of highly sensitive passive sampler for
nitrogen dioxide using porous polyethylene membrane filter as turbulence limiting diffuser. Atmos Environ
42: 4079-4088. http://dx.doi.0rg/10.1016/j.atmosenv.2008.01.029
Selgrade. MK: Mole. ML: Miller. FJ: Hatch. GE: Gardner. DE: Hu. PC. (1981). Effect of NO2 inhalation and
vitamin C deficiency on protein and lipid accumulation in the lung. Environ Res 26: 422-437.
Sellier. Y: Galineau. J: Hulin. A: Cainj F: Marquis. N: Navel V: Bottagisi S: Giorgis-Allemand. L: Jacquier. C:
Slama. R: Lepeule. J. (2014). Health effects of ambient air pollution: Do different methods for estimating
exposure lead to different results? Environ Int 66: 165-173. http://dx.doi.0rg/10.1016/i.envint.2014.02.001
Sereviciene. V: Paliulis. D. (2012). Influence of temperature and relative humidity on the performance of
nitrogen dioxide diffusive sampler. Res J of Chem Sci 2: 89-92.
Servais. S: Boussouar. A: Molnar. A: Doukj T: Pequignot JM: Favier. R. (2005). Age-related sensitivity to lung
oxidative stress during ozone exposure. Free Radic Res 39: 305-316.
http://dx.doi.org/10.1080/10715760400011098
Setton. E: Marshall. JD: Brauer. M: Lundquist. KR: Hystad. P: Keller. P: Cloutier-Fisher. D. (2011). The impact
of daily mobility on exposure to traffic-related air pollution and health effect estimates. J Expo Sci Environ
Epidemiol 21: 42-48. http://dx.doi.org/10.1038/ies.2010.14
Sevanian. A: Elsaved. N: Hacker. AD. (1982a). Effects of vitamin E deficiency and nitrogen dioxide exposure
on lung lipid peroxidation: Use of lipid epoxides and malonaldehyde as measures of peroxidation. J Toxicol
Environ Health 10: 743-756. http://dx.doi.org/10.1080/15287398209530292
Sevanian. A: Hacker. AD: Elsaved. N. (1982b). Influence of vitamin E and nitrogen dioxide on lipid
peroxidation in rat lung and liver microsomes. Lipids 17: 269-277'.
Shalamberidze. OP: Tsereteli. NT. (1971a). Effect of low concentrations of sulfur and nitrogen dioxides on the
estrual cycle and reproductive functions of experimental animals. Gig Sanit 36: 178-182.
Shalamberidze. OP: Tsereteli. NT. (1971b). [Effect of small concentrations of sulfur dioxide and nitrogen
peroxide on the estrus cycle and reproductive function in experiments on animals]. Gig Sanit 36: 13-17.
Shankardass. K: Mcconnell R: Jerrett M: Milam. J: Richardson. J: Berhane. K. (2009). Parental stress increases
the effect of traffic-related air pollution on childhood asthma incidence. PNAS 106: 12406-12411.
http://dx.doi.org/10.1073/pnas.0812910106
Shao. X: Stein. M: Ching. J. (2007). Statistical comparisons of methods for interpolating the output of a
numerical air quality model. J Stat Plan Inference 137: 2277-2293.
http://dx.doi.0rg/10.1016/i.ispi.2006.07.014
R-84
-------
Sheppard. L. (2005). Acute air pollution effects: Consequences of exposure distribution and measurements. J
Toxicol Environ Health A 68: 1127-1135. http://dx.doi.org/10.1080/15287390590935987
Sheppard. L: Slaughter. JC: Schildcrout J: Liu. J. -S: Lumlev. T. (2005). Exposure and measurement
contributions to estimates of acute air pollution effects. JExpo Anal Environ Epidemiol 15: 366-376.
http://dx.doi.org/10.1038/si.iea.7500413
Sherwin. RP: Carlson. DA. (1973). Protein content of lung lavage fluid of guinea pigs exposed to 0.4 ppm
nitrogen dioxide. Arch Environ Health 27: 90-93.
Sherwin. RP: Dibble. J: Weiner. J. (1972). Alveolar wall cells of the guinea pig: Increase in response to 2 ppm
nitrogen dioxide. Arch Environ Health 24: 43-47.
Sherwin. RP: Richters. V. (1982). Hyperplasia of type 2 pneumocytes following 0.34 ppm nitrogen dioxide
exposure: Quantitation by image analysis. Arch Environ Health 37: 306-315.
http://dx.doi.org/10.1080/00039896.1982.10667584
Sherwood. RL: Lippert WE: Goldstein. E: Tarkington. B. (1981). Effect of ferrous sulfate aerosols and nitrogen
dioxide on murine pulmonary defense. Arch Environ Occup Health 36: 130-135.
Shi. C: Zhang. B. (2008). Tropospheric NO2 columns over Northeastern North America: Comparison of CMAQ
model simulations with GOME satellite measurements. Adv Atmos Sci 25: 59-71.
http://dx.doi.org/10.1007/s00376-008-0059-8
Shields. KN: Cavallari. JM: Hunt. MJ: Lazo. M: Molina. M: Molina. L: Holguin. F. (2013). Traffic-related air
pollution exposures and changes in heart rate variability in Mexico City: A panel study. Environ Health 12:
7. http://dx.doi.org/10.1186/1476-069X-12-7
Shima. M: Nitta. Y: Ando. M: Adachi. M. (2002). Effects of air pollution on the prevalence and incidence of
asthma in children. Arch Environ Health 57: 529-535. http://dx.doi.org/10.1080/00039890209602084
Shin. HH: Stieb. D: Burnett R: Takahara. G: Jessiman. B. (2012). Tracking national and regional spatial-
temporal mortality risk associated with NO(2) concentrations in Canada: A Bayesian hierarchical two-level
model. Risk Anal 32: 513-530. http://dx.doi.0rg/10.llll/i.1539-6924.2011.01684.x
Shmool JLC: Michanowicz. PR: Cambal L: Tunno. B: Howell J: Gillooly. S: Roper. C: Tripathy. S: Chubb.
LG: Eisl HM: Gorczynski. JE: Holguin. FE: Shields. KN: Cloughertv. JE. (2014). Saturation sampling for
spatial variation in multiple air pollutants across an inversion-prone metropolitan area of complex terraing.
Environ Health 13: 28. http://dx.doi.org/10.1186/1476-069X-13-28
Sies. H. (2004). Ozone in arteriosclerotic plaques: Searching for the "smoking gun". Angew Chem Int Ed Engl
43: 3514-3515. http://dx.doi.org/10.1002/anie.200460118
Silbaugh. SA: Mauderly. JL: Macken. CA. (1981). Effects of sulfuric acid and nitrogen dioxide on airway
responsiveness of the guinea pig. J Toxicol Environ Health 8: 31-45.
http://dx.doi.org/10.1080/15287398109530048
Silkoff. PE: Zhang. L: Dutton. S: Langmack. EL: Vedal S: Murphy. J: Make. B. (2005). Winter air pollution and
disease parameters in advanced chronic obstructive pulmonary disease panels residing in Denver, Colorado. J
Allergy Clinlmmunol 115: 337-344. http://dx.doi.0rg/10.1016/i.jaci.2004.ll.035
Silverman. RA: Ito. K: Freese. J: Kaufman. BJ: De Claro. D: Braun. J: Prezant. DJ. (2010). Association of
ambient fine particles with out-of-hospital cardiac arrests in New York City. Am J Epidemiol 172: 917-923.
http://dx.doi.org/10.1093/aie/kwq217
Simoens. S: Ayrault M: Wallace. JM. (2007). The flow across a street canyon of variable width- Part 1:
Kinematic description. Atmos Environ 41: 9002-9017. http://dx.doi.0rg/10.1016/i.atmosenv.2007.08.013
Simoens. S: Wallace. JM. (2008). The flow across a street canyon of variable widthPart 2: Scalar dispersion
from a street level line source. Atmos Environ 42: 2489-2503.
http://dx.doi.0rg/10.1016/i.atmosenv.2007.12.013
R-85
-------
Simon. H: Kimura. Y: McGaughev. G: Allen. DT: Brown. SS: Osthoff. HP: Roberts. JM: Byun. D: Lee. D.
(2009). Modeling the impact of C1NO2 on ozone formation in the Houston area. J Geophys Res Atmos 114:
DOOF03. http://dx.doi.org/10.1029/2008jd010732
Simon. PK: Dasgupta. PK. (1995). Continuous automated measurement of the soluble fraction of atmospheric
paniculate matter. Anal Chem 67: 71-78. http://dx.doi.org/10.1021/ac00097a012
Simpson. R: Williams. G: Petroeschevsky. A: Best. T: Morgan. G: Denison. L: Hinwood. A: Neville. G.
(2005a). The short-term effects of air pollution on hospital admissions in four Australian cities. Aust N Z J
Public Health 29: 213-221.
Simpson. R: Williams. G: Petroeschevsky. A: Best. T: Morgan. G: Denison. L: Hinwood. A: Neville. G: Neller.
A^ (2005b). The short-term effects of air pollution on daily mortality in four Australian cities. Aust N Z J
Public Health 29: 205-212. http://dx.doi.org/10.1111/i. 1467-842X.2005.tb00758.x
Sinclair. AH: Edgerton. ES: Wyzga. R: Tolsma. D. (2010). A two-time-period comparison of the effects of
ambient air pollution on outpatient visits for acute respiratory illnesses. J Air Waste Manag Assoc 60: 163-
175. http://dx.doi.0rg/10.3155/1047-3289.60.2.163
Sinclair. AH: Tolsma. D. (2004). Associations and lags between air pollution and acute respiratory visits in an
ambulatory care setting: 25-month results from the Aerosol Research and Inhalation Epidemiological Study.
J Air Waste Manag Assoc 54: 1212-1218. http://dx.doi.org/10.1080/10473289.2004.10470979
Singer. BC: Hodgson. AT: Hotchi. T: Kim JJ. (2004). Passive measurement of nitrogen oxides to assess traffic-
related pollutant exposure for the East Bay Children's Respiratory Health Study. Atmos Environ 38: 393-403.
http://dx.doi.0rg/10.1016/i.atmosenv.2003.10.005
Singh. HB: Salas. L: Herlth. D: Kolyer. R: Czech. E: Avery. M: Crawford. JH: Pierce. RB: Sachse. GW: Blake.
PR: Cohen. RC: Bertram. TH: Perring. A: Wooldridge. PJ: Dibb. J: Huev. G: Hudman. RC: Turquetv. S:
Emmons. LK: Flocke. F: Tang. Y: Carmichael. GR: Horowitz. LW. (2007). Reactive nitrogen distribution
and partitioning in the North American troposphere and lowermost stratosphere. J Geophys Res 112:
D12S04. http://dx.doi.org/10.1029/2006JD007664
Skouloudis. AN: Kassomenos. P. (2014). Combining environment and health information systems for the
assessment of atmospheric pollution on human health. Sci Total Environ 488: 364-370.
http://dx.doi.0rg/10.1016/i.scitotenv.2014.02.116
Slade. R: Crissman. K: Norwood. J: Hatch. G. (1993). Comparison of antioxidant substances in bronchoalveolar
lavage cells and fluid from humans, guinea pigs, and rats. Exp Lung Res 19: 469-484.
http://dx.doi.org/10.3109/01902149309064358
Slama. R: Bottagisi. S: Solansky. I: Lepeule. J: Giorgis-Allemand. L: Sram R. (2013). Short-term impact of
atmospheric pollution on fecundability. Epidemiology 24: 871-879.
http://dx.doi.org/10.1097/EDE.Ob013e3182a702c5
Slama. R: Darrow. L: Parker. J: Woodruff. TJ: Strickland. M: Nieuwenhuijsen. M: Glinianaia. S: Hoggatt KJ:
Kannan. S: Hurley. F: Kalinka. J: Sram R: Brauer. M: Wilhelm. M: Heinrich. J: Ritz. B. (2008). Meeting
report: Atmospheric pollution and human reproduction. Environ Health Perspect 116: 791-798.
http://dx.doi.org/10.1289/ehp. 11074
Slama. R: Morgenstern. V: Cyrys. J: Zutavern. A: Herbarth. O: Wichmann. HE: Heinrich. J: Group. LS. (2007).
Traffic-related atmospheric pollutants levels during pregnancy and offspring's term birth weight: A study
relying on a land-use regression exposure model. Environ Health Perspect 115: 1283-1292.
http://dx.doi.org/10.1289/ehp. 10047
Slerdal LH: Walker. SE: Solberg. S. (2003). The urban air dispersion model EPISODE applied in
AirQUIS2003: Technical description. (TR 12/2003). Kjeller, Norway: Norwegian Institute for Air Research.
Smargiassi A: Baldwin. M: Pilger. C: Dugandzic. R: Brauer. M. (2005). Small-scale spatial variability of
particle concentrations and traffic levels in Montreal: A pilot study. Sci Total Environ 3 3 8: 243 -251.
http://dx.doi.0rg/10.1016/i.scitotenv.2004.07.013
R-86
-------
Smargiassi A: Goldberg. MS: Wheeler. AJ: Plante. C: Valois. MF: Mallach. G: Kauri. LM: Shutt R: Bartlett S:
Raphoz. M: Liu. L. (2014). Associations between personal exposure to air pollutants and lung function tests
and cardiovascular indices among children with asthma living near an industrial complex and petroleum
refineries. Environ Res 132: 38-45. http://dx.doi.0rg/10.1016/i.envres.2014.03.030
Smith. LA: Mukerjee. S: Chung. KG: Afghani. J. (2011). Spatial analysis and land use regression of VOCs and
NO2 in Dallas, Texas during two seasons. J Environ Monit 13: 999-1007.
http://dx.doi.org/10.1039/cOem00724b
Smith. LL. (2004). Oxygen, oxysterols, ouabain, and ozone: A cautionary tale [Review]. Free Radic Biol Med
37: 318-324. http://dx.doi.0rg/10.1016/i.freeradbiomed.2004.04.024
Snyder. MG: Venkatram A: Heist DK: Perry. SG: Petersen. WE: Isakov. V. (2013). RLINE: A line source
dispersion model for near-surface releases. Atmos Environ 77: 748-756.
http://dx.doi.0rg/10.1016/i.atmosenv.2013.05.074
Sokol. RZ: Kraft. P: Fowler. IM: Mamet R: Kim. E: Berhane. KT. (2006). Exposure to environmental ozone
alters semen quality. Environ Health Perspect 114: 360-365. http://dx.doi.org/10.1289/ehp.8232
Solomon. C: Christian. PL: Chen. LL: Welch. BS: Kleinmaa MT: Dunham. E: Erie. DJ: Balmes. JR. (2000).
Effect of serial-day exposure to nitrogen dioxide on airway and blood leukocytes and lymphocyte subsets.
EurRespir 115:922-928.
Son. B: Yang. W: Brevsse. P: Chung. T: Lee. Y. (2004). Estimation of occupational and nonoccupational
nitrogen dioxide exposure for Korean taxi drivers using a microenvironmental model. Environ Res 94: 291-
296. http://dx.doi.0rg/10.1016/j.envres.2003.08.004
Son. JY: Bell ML: Lee. JT. (2010). Individual exposure to air pollution and lung function in Korea: Spatial
analysis using multiple exposure approaches. Environ Res 110: 739-749.
http://dx.doi.0rg/10.1016/i.envres.2010.08.003
Son. JY: Cho. YS: Lee. JT. (2008). Effects of air pollution on postneonatal infant mortality among firstborn
infants in Seoul, Korea: Case-crossover and time-series analyses. Arch Environ Occup Health 63: 108-113.
http://dx.doi.Org/10.3200/AEOH.63.3.108-113
Son. JY: Lee. JT: Park. YH: Bell ML. (2013). Short-term effects of air pollution on hospital admissions in
Korea. Epidemiology 24: 545-554. http://dx.doi.org/10.1097/EDE.Ob013e3182953244
Song. Y: Namkung. W: Nielson. DW: Lee. JW: Finkbeiner. WE: Verkman. AS. (2009). Airway surface liquid
depth measured in ex vivo fragments of pig and human trachea: Dependence on Na+ and Cl- channel
function. Am J Physiol Lung Cell Mol Physiol 297: LI 131-L1140.
http://dx.doi.org/10.1152/aiplung.00085.2009
Sonnenschein-Van der Voort. AM: de Kluizenaar. Y: Jaddoe. VW: Gabriele. C: Raat H: Moll HA: Hofman. A:
Pierik. FH: Miedema. HM: de Jongste. JC: Duijts. L. (2012). Air pollution, fetal and infant tobacco smoke
exposure, and wheezing in preschool children: A population-based prospective birth cohort. Environ Health
11: 91. http://dx.doi.org/10.1186/1476-069X-ll-91
Sorensen. M: Andersen. ZJ: Nordsborg. RB: Becker. T: Tjenneland. A: Overvad. K: Raaschou-Nielsen. O.
(2013). Long-term exposure to road traffic noise and incident diabetes: A cohort study. Environ Health
Perspect 121: 217-222. http://dx.doi.org/10.1289/ehp. 1205503
Serensen. M: Hoffmann. B: Hvidberg. M: Ketzel. M: Jensen. SS: Andersen. ZJ: Tjenneland. A: Overvad. K:
Raaschou-Nielsen. O. (2012). Long-term exposure to traffic-related air pollution associated with blood
pressure and self-reported hypertension in a Danish cohort. Environ Health Perspect 120: 418-424.
http://dx.doi.org/10.1289/ehp. 1103631
Serensen. M: Liihdorf. P: Ketzel M: Andersen. ZJ: Tjenneland. A: Overvad. K: Raaschou-Nielsen. O. (2014).
Combined effects of road traffic noise and ambient air pollution in relation to risk for stroke? Environ Res
133: 49-55. http://dx.doi.0rg/10.1016/i.envres.2014.05.011
R-87
-------
Soto-Ramos. M: Castro-Rodriguez. JA: Hinojos-Gallardo. LC: Hernandez-Saldana. R: Cisneros-Castolo. M:
Carrillo-Rodriguez. V. (2013). Fractional exhaled nitric oxide has a good correlation with asthma control and
lung function in Latino children with asthma. J Asthma 50: 590-594.
http://dx.doi.org/10.3109/02770903.2013.792349
Sousa. R: Duque. L: Duarte. AJ: Gomes. CR: Ribeiro. H: Cruz. A: Esteves da Silva. JC: Abreu. I. (2012). In
vitro exposure of Acer negundo pollen to atmospheric levels of SO and NO: Effects on allergenicity and
germination. Environ Sci Technol 46: 2406-2412. http://dx.doi.org/10.1021/es2034685
Southwest Research Institute. (2013). Phase 2 of the advanced collaborative emissions study. (SwRI Final
Report 03.17124). Alpharetta, GA: Coordinating Research Council, Inc.
http://crcao.org/reports/recentstudies2013/ACES%20Ph2/03-17124 CRC%20ACES%20Phase2-
%20FINAL%20Report Khalek-R6-SwRI.pdf
Spengler. JD: Brauer. M: Samet JM: Lambert. WE. (1993). Nitrous acid in Albuquerque, New Mexico, homes.
Environ Sci Technol 27: 841-845. http://dx.doi.org/10.1021/es00042a005
Spicer. CW: Coutant. RW: Ward. GF: Joseph. DW: Gaynor. AJ: Billick. IH. (1989). Rates and mechanisms of
NO2 removal from indoor air by residential materials. Environ Int 15: 643-654.
http://dx.doi.org/10.1016/0160-4120(89)90087-l
Spiegelman. D. (2013). Regression calibration in air pollution epidemiology with exposure estimated by spatio-
temporal modeling. Environmetrics 24: 521-524. http://dx.doi.org/10.1002/env.2249
Spiezia. L: Campello. E: Boa M: Maggiolo. S: Pelizzaro. E: Simioni. P. (2014). Short-term exposure to high
levels of air pollution as a risk factor for acute isolated pulmonary embolism. Thromb Res Suppl 134: 259-
263. http://dx.doi.0rg/10.1016/j.thromres.2014.05.011
Spira-Cohen. A. (2013). Email from Dr. Spira-Cohen to Dr. Patel; Response to data request. Available online
Spira-Cohen. A: Chen. LC: Kendall M: Lall. R: Thurston. GD. (2011). Personal exposures to traffic-related air
pollution and acute respiratory health among Bronx schoolchildren with asthma. Environ Health Perspect
119: 559-565. http://dx.doi.org/10.1289/ehp. 1002653
Squadrito. GL: Postlethwait EM. (2009). On the hydrophobicity of nitrogen dioxide: Could there be a "lens"
effect for NO(2) reaction kinetics. Nitric Oxide 21: 104-109. http://dx.doi.0rg/10.1016/i.niox.2009.05.008
Squadrito. GL: Postlethwait. EM: Matalon. S. (2010). Elucidating mechanisms of chlorine toxicity: Reaction
kinetics, thermodynamics, and physiological implications [Review]. Am J Physiol Lung Cell Mol Physiol
299: L289-L300. http://dx.doi.org/10.1152/aiplung.00077.2010
Staimer. N: Delfino. RJ: Bufalino. C: Fine. PM: Sioutas. C: Kleinman. MT. (2005). A miniaturized active
sampler for the assessment of personal exposure to nitrogen dioxide. Anal Bioanal Chem 383: 955-962.
http://dx.doi.org/10.1007/s00216-005-0086-6
Stanojevic. S: Wade. A: Stocks. J: Hankinson. J: Coates. AL: Pan. H: Rosenthal M: Corey. M: Lebecque. P:
Cole. TJ. (2008). Reference ranges for spirometry across all ages: A new approach. Am J Respir Crit Care
Med 177: 253-260. http://dx.doi.org/10.1164/rccm.200708-1248OC
Stavrakou. T: Miiller. JF: Boersma. KF: De Smedt I: Van Per A. RJ. (2008). Assessing the distribution and
growth rates of NO x emission sources by inverting a 10-year record of NO 2 satellite columns. Geophys Res
Lett 35: L10801. http://dx.doi.org/10.1029/2008GL033521
Steenhof. M: Janssea NA: Strak. M: Hoek. G: Gosens. I: Mudwav. IS: Kelly. FJ: Harrison. RM: Pieters. RH:
Cassee. FR: Brunekreef. B. (2014). Air pollution exposure affects circulating white blood cell counts in
healthy subjects: The role of particle composition, oxidative potential and gaseous pollutants - the RAPTES
project. Inhal Toxicol 26: 141-165. http://dx.doi.org/10.3109/08958378.2013.861884
Steenhof. M: Mudwav. IS: Gosens. I: Hoek. G: Godri. KJ: Kelly. FJ: Harrison. RM: Pieters. RH: Cassee. FR:
Lebret E: Brunekreef. BA: Strak. M: Janssen. NA. (2013). Acute nasal pro-inflammatory response to air
pollution depends on characteristics other than particle mass concentration or oxidative potential: The
RAPTES project. Occup Environ Med 70: 341-348. http://dx.doi.org/10.1136/oemed-2012-100993
R-8
-------
Steerenberg. PA: Bischoff. EWM. A: de Klerk. A: Verlaan. APJ: Jongbloets. LMN: Van Loveren. H:
Opperhuizen. A: Zomer. G: Heisterkamp. SH: Hadv. M: Spieksma. FTM: Fischer. PH: Dormans. JAM. A:
van Amsterdam. JGC. (2003). Acute effect of air pollution on respiratory complaints, exhaled NO and
bio markers in nasal lavages of allergic children during the pollen season. Int Arch Allergy Immunol 131:
127-137. http://dx.doi.org/10.1159/000070928
Steerenberg. PA: Nierkens. S: Fischer. PH: Van Loveren. H: Opperhuizen. A: Vos. JG: van Amsterdam. JG.
(2001). Traffic-related air pollution affects peak expiratory flow, exhaled nitric oxide, and inflammatory
nasal markers. Arch Environ Health 56: 167-174. http://dx.doi.org/10.1080/00039890109604069
Steinbacher. M: Zellweger. C: Schwarzenbach. B: Bugmann. S: Buchmann. B: Ordonez. C: Prevot ASH:
Hueglin. C. (2007). Nitrogen oxides measurements at rural sites in Switzerland: Bias of conventional
measurement techniques. J Geophys Res Atmos 112: D11307. http://dx.doi.org/10.1029/2006JD007971
Steinvil A: Fireman. E: Kordova-Biezuner. L: Cohen. M: Shapira. I: Berliner. S: Rogowski. O. (2009).
Environmental air pollution has decremental effects on pulmonary function test parameters up to one week
after exposure. Am J Med Sci 338: 273-279. http://dx.doi.org/10.1097/MAJ.Ob013e3181adb3ed
Steinvil A: Kordova-Biezuner. L: Shapira. I: Berliner. S: Rogowski. O. (2008). Short-term exposure to air
pollution and inflammation-sensitive biomarkers. Environ Res 106: 51-61.
http://dx.doi.0rg/10.1016/i.envres.2007.08.006
Stenger. MR: Rose. MJ: Joshi. MS: Rogers. LK: Chicoine. LG: Bauer. JA: Nehru LD. (2010). Inhaled nitric
oxide prevents 3-nitrotyrosine formation in the lungs of neonatal mice exposed to >95% oxygen. Lung 188:
217-227. http://dx.doi.org/10.1007/s00408-010-9235-6
Stern. G: Latzin. P: Roosli. M: Fuchs. O: Proietti. E: Kuehni. C: Frev. U. (2013). A prospective study of the
impact of air pollution on respiratory symptoms and infections in infants. Am J Respir Crit Care Med 187:
1341-1348. http://dx.doi.org/10.1164/rccm.201211-2008OC
Stevens. MA: Menache. MG: Crapo. JD: Miller. FJ: Graham. JA. (1988). Pulmonary function in juvenile and
young adult rats exposed to low-level NO2 with diurnal spikes. J Toxicol Environ Health 23: 229-240.
http://dx.doi.org/10.1080/15287398809531109
Stieb. DM: Burnett. RT: Smith-Doiron. M: Brion. O: Shin. HH: Economou. V. (2008). A new multipollutant,
no-threshold air quality health index based on short-term associations observed in daily time-series analyses.
J Air Waste Manag Assoc 58: 435-450. http://dx.doi.Org/10.3155/1047-3289.58.3.435
Stieb. DM: Judek. S: Burnett RT. (2002). Meta-analysis of time-series studies of air pollution and mortality:
Effects of gases and particles and the influence of cause of death, age, and season. J Air Waste Manag Assoc
52: 470-484. http://dx.doi.org/10.1080/10473289.2002.10470794
Stieb. DM: Judek. S: Burnett RT. (2003). Meta-analysis of time-series studies of air pollution and mortality:
Update in relation to the use of generalized additive models. J Air Waste Manag Assoc 53: 258-261.
http://dx.doi.org/10.1080/10473289.2003.10466149
Stieb. DM: Szyszkowicz. M: Rowe. BH: Leech. JA. (2009). Air pollution and emergency department visits for
cardiac and respiratory conditions: A multi-city time-series analysis. Environ Health 8.
http://dx.doi.org/10.1186/1476-069X-8-25
Stingone. JA: Luben. TJ: Daniels. JL: Fuentes. M: Richardson. DB: Aylsworth. AS: Herring. AH: Anderka. M:
Botto. L: Correa. A: Gilboa. SM: Langlois. PH: Moslev. B: Shaw. GM: Siffel C: Olshan. AF. (2014).
Maternal exposure to criteria air pollutants and congenital heart defects in offspring: Results from the
national birth defects prevention study. Environ Health Perspect 122: 863-872.
http://dx.doi.org/10.1289/ehp.1307289
Strak. M. (2013). [Email from Dr. Strak to Dr. Patel; Response to data request]. Available online
Strak. M: Hoek. G: Godri. KJ: Gosens. I: Mudwav. IS: van Oerle. R: Spronk. HM: Cassee. FR: Lebret. E: Kelly.
FJ: Harrison. RM: Brunekreef. B: Steenhof. M: Janssen. NA. (2013a). Composition of PM affects acute
vascular inflammatory and coagulative markers - The RAPTES project. PLoS ONE 8: e58944.
http://dx.doi.org/10.1371/iournal.pone.0058944
R-89
-------
Strak. M: Hoek. G: Steenhof. M: Kilinc. E: Godri. KJ: Gosens. I: Mudwav. IS: van Oerle. R: Spronk. HM:
Cassee. FR: Kelly. FJ: Harrison. RM: Brunekreef. B: Lebret. E: Janssen. NA. (2013b). Components of
ambient air pollution affect thrombin generation in healthy humans: The RAPTES project. Occup Environ
Med 70: 332-340. http://dx.doi.org/10.1136/oemed-2012-100992
Strak. M: Janssen. NA: Godri. KJ: Gosens. I: Mudwav. IS: Cassee. FR: Lebret. E: Kelly. FJ: Harrison. RM:
Brunekreef. B: Steenhof. M: Hoek. G. (2012). Respiratory health effects of airborne paniculate matter: The
role of particle size, composition and oxidative potential - The RAPTES project. Environ Health Perspect
120: 1183-1189. http://dx.doi.org/10.1289/ehp.1104389
Strand. V: Rak. S: Svartengren. M: Bylin. G. (1997). Nitrogen dioxide exposure enhances asthmatic reaction to
inhaled allergen in subjects with asthma. Am J Respir Crit Care Med 155: 881-887.
http://dx.doi.0rg/10.1164/airccm.155.3.9117021
Strand. V: Salomonsson. P: Lundahl J: Bylin. G. (1996). Immediate and delayed effects of nitrogen dioxide
exposure at an ambient level on bronchial responsiveness to histamine in subjects with asthma. Eur Respir J
9: 733-740. http://dx.doi.org/10.1183/09031936.96.09040733
Strand. V: Svartengren. M: Rak. S: Barck. C: Bylin. G. (1998). Repeated exposure to an ambient level of NO2
enhances asthmatic response to a nonsymptomatic allergen dose. Eur Respir J 12: 6-12.
http://dx.doi.org/10.1183/09031936.98.12010006
Stranev. L: Finn. J: Dennekamp. M: Bremner. A: Tonkin. A: Jacobs. I. (2014). Evaluating the impact of air
pollution on the incidence of out-of-hospital cardiac arrest in the Perth Metropolitan Region: 2000-2010. J
Epidemiol Community Health 68: 6-12. http://dx.doi.org/10.1136/iech-2013-202955
Strickland. MJ: Darrow. LA: Klein. M: Flanders. WD: Sarnat JA: Waller. LA: Sarnat SE: Mulholland. JA:
Tolbert PE. (2010). Short-term associations between ambient air pollutants and pediatric asthma emergency
department visits. Am J Respir Crit Care Med 182: 307-316. http://dx.doi.org/10.1164/rccm.200908-1201OC
Strickland. MJ: Darrow. LA: Mulholland. JA: Klein. M: Flanders. WD: Winquist A: Tolbert. PE. (2011).
Implications of different approaches for characterizing ambient air pollutant concentrations within the urban
airshed for time-series studies and health benefits analyses. Environ Health 10: 36.
http://dx.doi.org/10.1186/1476-069X-10-36
Stroh. E: Oudin. A: Gustafsson. S: Pilesjo. P: Harrie. L: Stromberg. U: Jakobsson. K. (2005). Are associations
between socio-economic characteristics and exposure to air pollution a question of study area size? An
example from Scania, Sweden. Int J Health Geogr 4: 30-42. http://dx.doi.org/10.1186/1476-072X-4-30
Stall. R. (2000). Meteorology for scientists and engineers: A technical companion book with Ahrens'
Meteorology Today (2nd ed.). Pacific Grove, CA: Brooks/Cole.
Su. JG: Brauer. M: Ainslie. B: Steyn. D: Larson. T: Buzzelli. M. (2008a). An innovative land use regression
model incorporating meteorology for exposure analysis. Sci Total Environ 390: 520-529.
http://dx.doi.0rg/10.1016/i.scitotenv.2007.10.032
Su. JG: Brauer. M: Buzzelli. M. (2008b). Estimating urban morphometry at the neighborhood scale for
improvement in modeling long-term average air pollution concentrations. Atmos Environ 42: 7884-7893.
http://dx.doi.0rg/10.1016/i.atmosenv.2008.07.023
Su. JG: Jerrett. M: Beckerman. B. (2009a). A distance-decay variable selection strategy for land use regression
modeling of ambient air pollution exposures. Sci Total Environ 407: 3890-3898.
http://dx.doi.0rg/10.1016/i.scitotenv.2009.01.061
Su. JG: Jerrett. M: Beckerman. B: Wilhelm M: Ghosh. JK: Ritz. B. (2009b). Predicting traffic-related air
pollution in Los Angeles using a distance decay regression selection strategy. Environ Res 109: 657-670.
http://dx.doi.0rg/10.1016/i.envres.2009.06.001
Su. JG: Jerrett M: Morello-Frosch. R: Jesdale. BM: Kyle. AD. (2012). Inequalities in cumulative environmental
burdens among three urbanized counties in California. Environ Int 40: 79-87.
http://dx.doi.0rg/10.1016/i.envint.2011.ll.003
R-90
-------
Su. JG: Morello-Frosch. R: Jesdale. BM: Kyle. AD: Shamasunder. B: Jerrett. M. (2009c). An index for assessing
demographic inequalities in cumulative environmental hazards with application to Los Angeles, California.
Environ Sci Technol 43: 7626-7634. http://dx.doi.org/10.1021/es901041p
Suh. HH: Zanobetti. A. (2010a). Erratum: Exposure error masks the relationship between traffic-related air
pollution and heart rate variability [Erratum]. J Occup Environ Med 52: 1138.
http://dx.doi.org/10.1097/JOM.Ob013e3181fd2632
Suh. HH: Zanobetti. A. (2010b). Exposure error masks the relationship between traffic-related air pollution and
heart rate variability. J Occup Environ Med 52: 685-692. http://dx.doi.org/10.1097/JOM.Ob013e3181e8071f
Sunyer. J: Jarvis. D: Gotschi. T: Garcia-Esteban. R: Jacquemin. B: Aguilera. I: Ackerman. U: de Marco. R:
Forsberg. B: Gislason. T: Heinrich. J: Norback. D: Villani. S: Kiinzli. N. (2006). Chronic bronchitis and
urban air pollution in an international study. Occup Environ Med 63: 836-843.
http://dx.doi.org/10.1136/oem.2006.027995
Sunyer. J: Pekkanen. J: Garcia-Esteban. R: Svanes. C: Kiinzli. N: Janson. C: de Marco. R: Anto. JM: Burnev. P.
(2007). Asthma score: Predictive ability and risk factors. Allergy 62: 142-148.
http://dx.doi.org/10.1111/i. 1398-9995.2006.01184.x
Sunyer. J: Puig. C: Torrent. M: Garcia-Algar. O: Calico. I: Munoz-Ortiz. L: Barnes. M: Cullinan. P. (2004).
Nitrogen dioxide is not associated with respiratory infection during the first year of life. Int J Epidemiol 33:
116-120. http://dx.doi.org/10.1093/iie/dvh037
Suzuki. AK: Tsubone. H: Ichinose. T: Oda. H: Kubota. K. (1981). [Effects of subchronic nitrogen dioxide
exposure on arterial blood pHa, PaCO2 and PaO2 in rats]. Nippon Eiseigaku Zasshi 36: 816-823.
http://dx.doi.org/10.1265/ijh.36.816
Suzuki. T: Ikeda. S: Kanoh. T: Mizoguchi. I. (1986). Decreased phagocytosis and superoxide anion production
in alveolar macrophages of rats exposed to nitrogen dioxide. Arch Environ Contain Toxicol 15: 733-739.
http://dx.doi.org/10.1007/BF01054920
Svendsen. ER: Gonzales. M: Mukerjee. S: Smith. L: Ross. M: Walsh. D: Rhonev. S: Andrews. G: Ozkaynak. H:
Neas. LM. (2012). GIS-modeled indicators of traffic-related air pollutants and adverse pulmonary health
among children in El Paso, Texas. Am J Epidemiol 176: S131-S141. http://dx.doi.org/10.1093/aie/kws274
Szpiro. AA: Paciorek. CJ. (2013a). Measurement error in two-stage analyses, with application to air pollution
epidemiology. Environmetrics 24: 501-517. http://dx.doi.org/10.1002/env.2233
Szpiro. AA: Paciorek. CJ. (2013b). Rejoinder [Editorial]. Environmetrics 24: 531-536.
http://dx.doi.org/10.1002/env.2254
Szpiro. AA: Paciorek. CJ: Sheppard. L. (201 la). Does more accurate exposure prediction necessarily improve
health effect estimates? Epidemiology 22: 680-685. http://dx.doi.org/10.1097/EDE.Ob013e3182254cc6
Szpiro. AA: Sampson. PD: Sheppard. L: Lumlev. T: Adar. SD: Kaufman. JD. (2010). Predicting intra-urban
variation in air pollution concentrations with complex spatio-temporal dependencies. Environmetrics 21:
606-631. http://dx.doi.org/10.1002/env.1014
Szpiro. AA: Sheppard. L: Lumlev. T. (201 Ib). Efficient measurement error correction with spatially misaligned
data. Biostatistics 12: 610-623. http://dx.doi.org/10.1093/biostatistics/kxq083
Szyszkowicz. M. (2007). Air pollution and emergency department visits for ischemic heart disease in Montreal,
Canada. Int J Occup Med Environ Health 20: 167-173. http://dx.doi.org/10.2478/vl0001-007-0015-3
Szyszkowicz. M. (2008). Ambient air pollution and daily emergency department visits for ischemic stroke in
Edmonton, Canada. Int J Occup Med Environ Health 21: 295-300. http://dx.doi.org/10.2478/vl0001-008-
0029-5
Szyszkowicz. M. (2009). Air pollution and ED visits for chest pain. Am J Emerg Med 27: 165-168.
http://dx.doi.0rg/10.1016/i.aiem.2008.01.010
R-91
-------
Szyszkowicz. M: Rowe. BH: Brook. RD. (2012). Even low levels of ambient air pollutants are associated with
increased emergency department visits for hypertension. Can J Cardiol 28: 360-366.
http://dx.doi.0rg/10.1016/i.cica.2011.06.011
Tabacova. S: Nikiforov. B: Balabaeva. L. (1985). Postnatal effects of maternal exposure to nitrogen dioxide.
Neurotoxicol Teratol 7: 785-789.
Tager. IB: Balmes. J: Lurmann. F: Ngo. L: Alcorn. S: Kunzli. N. (2005). Chronic exposure to ambient ozone and
lung function in young adults. Epidemiology 16: 751-759.
http://dx.doi.org/10.1097/01.ede.0000183166.68809.bO
Takano. H: Yanagisawa. R: Inoue. KI: Shimada. A: Ichinose. T: Sadakane. K: Yoshino. S: Yamaki. K: Morita.
M: Yoshikawa. T. (2004). Nitrogen dioxide air pollution near ambient levels is an atherogenic risk primarily
in obese subjects: A brief communication. ExpBiolMed229: 361-364.
Takenoue. Y: Kaneko. T: Miyamae. T: Mori. M: Yokota. S. (2012). The influence of outdoor NO2 exposure on
asthma in childhood: Meta-analysis [Review]. Pediatrics International 54: 762-769.
http://dx.doi.org/10.1111/i. 1442-200X.2012.03674.x
Tang. MJ: Schuster. G: Crowlev. JN. (2014). Heterogeneous reaction of N2O5 with illite and Arizona test dust
particles. Atmos ChemPhys 14: 245-254. http://dx.doi.org/10.5194/acp-14-245-2014
Tao. Y: Huang. W: Huang. X: Zhong. L: Lu. SE: Li. Y: Dai. L: Zhang. Y: Zhu. T. (2012). Estimated acute
effects of ambient ozone and nitrogen dioxide on mortality in the Pearl River Delta of Southern China.
Environ Health Perspect 120: 393-398. http://dx.doi.org/10.1289/ehp.1103715
Tashkin. DP: Simmons. MS: Chang. P: Liu. H: Coulson. AH. (1993). Effects of smoked substance abuse on
nonspecific airway hyperresponsiveness. Am Rev RespirDis 147: 97-103.
http://dx.doi.0rg/10.1164/airccrn/147.l.97
Taylor. DR. (2012). Advances in the clinical applications of exhaled nitric oxide measurements [Review]. J
Breath Res 6: 047102. http://dx.doi.Org/10.1088/1752-7155/6/4/047102
Teichert T: Vossoughj M: Vierkotter. A: Sugiri. D: Schikowski. T: Schulte. T: Roden. M: Luckhaus. C: Herder.
C: Kramer. U. (2013). Association between traffic-related air pollution, subclinical inflammation and
impaired glucose metabolism: Results from the SALIA study. PLoS ONE 8: e83042.
http://dx.doi.org/10.1371/iournal.pone.0083042
Tepper. JS: Costa. PL: Winsett DW: Stevens. MA: Doerfler. PL: Watkinson. WP. (1993). Near-lifetime
exposure of the rat to a simulated urban profile of nitrogen dioxide: Pulmonary function evaluation. Fundam
Appl Toxicol 20: 88-96. http://dx.doi.0rg/10.1093/toxsci/20.l.88
TFESC and NASPE (Task Force of the European Society of Cardiology and the North American Society of
Pacing and Electrophysiology). (1996). Heart rate variability: Standards of measurement, physiological
interpretation and clinical use. Circulation 93: 1043-1065. http://dx.doi.org/10.1161/01.CrR.93.5.1043
Thach. TQ: Wong. CM: Chan. KP: Chau. YK: Thomas. GN: Ou. CO: Yang. L: Peiris. JSM: Lam. TH: Hedlev.
AL (2010). Air pollutants and health outcomes: Assessment of confounding by influenza. Atmos Environ 44:
1437-1442. http://dx.doi.0rg/10.1016/i.atmosenv.2010.01.036
Thaller. El: Petronella. SA: Hochman. D: Howard. S: Chhikara. RS: Brooks. EG. (2008). Moderate increases in
ambient PM2.5 and ozone are associated with lung function decreases in beach lifeguards. J Occup Environ
Med50: 202-211. http://dx.doi.org/10.1097/JOM.Ob013e31816386b4
Thiering. E: Cyrys. J: Kratzsch. J: Meisinger. C: Hoffmann. B: Berdel D: von Berg. A: Koletzko. S: Bauer. CP:
Heinrich. J. (2013). Long-term exposure to traffic-related air pollution and insulin resistance in children:
Results from the GINIplus and LISAplus birth cohorts. Diabetologia 56: 1696-1704.
http://dx.doi.org/10.1007/s00125-013-2925-x
Thomas. HV: Mueller. PK: Wright R. (1967). Response of rat lung mast cells to nitrogen dioxide inhalation. J
AirPollut Control Assoc 17: 33-35. http://dx.doi.org/10.1080/00022470.1967.10468939
R-92
-------
Thompson. AM: Zanobetti A: Silverman. F: Schwartz. J: Coull B: Urch. B: Speck. M: Brook. JR: Manno. M:
Gold. DR. (2010). Baseline repeated measures from controlled human exposure studies: Associations
between ambient air pollution exposure and the systemic inflammatory biomarkers IL-6 and fibrinogen.
Environ Health Perspect 118: 120-124. http://dx.doi.org/10.1289/ehp.0900550
Thornton. JA: Braban. CF: Abbatt JPD. (2003). N2O5 hydrolysis on sub-micron organic aerosol: The effect of
relative humidity, particle phase, and particle size. Phys Chem Chem Phys 5: 4593-4603.
http://dx.doi.org/10.1039/b307498f
Thornton. JA: Kercher. JP: Riedel TP: Wagner. NL: Cozic. J: Hollowav. JS: Dube. WP: Wolfe. GM: Quinn.
PK: Middlebrook. AM: Alexander. B: Brown. SS. (2010). A large atomic chlorine source inferred from mid-
continental reactive nitrogen chemistry. Nature 464: 271-274. http://dx.doi.org/10.1038/nature08905
Thurlbeck. WM. (1982). Postnatal human lung growth. Thorax 37: 564-571.
http://dx.doi.0rg/10.1136/thx.37.8.564
Tian. N: Xue. J: Barzyk. TM. (2013). Evaluating socioeconomic and racial differences in traffic-related metrics
in the United States using a GIS approach. J Expo Sci Environ Epidemiol 23: 215-222.
http://dx.doi.org/10.1038/ies.2012.83
Timonen. KL: Pekkanen. J. (1997). Air pollution and respiratory health among children with asthmatic or cough
symptoms. Am J Respir Crit Care Med 156: 546-552. http://dx.doi.0rg/10.1164/airccm.156.2.9608044
Timonen. KL: Vanninen. E: de Hartog. J: Ibald-Mulli. A: Brunekreef. B: Gold. PR: Henrich. J: Hoek. G: Lanki.
T: Peters. A: Tarkiainen. T: Tiittanen. P: Kreyling. W: Pekkanen. J. (2006). Effects of ultrafine and fine
paniculate and gaseous air pollution on cardiac autonomic control in subjects with coronary artery disease:
The ULTRA study. J Expo Sci Environ Epidemiol 16: 332-341. http://dx.doi.org/10.1038/sj.jea.7500460
Tokairin. T: Kitada. T. (2005). Study on the effect of porous fence on air quality and traffic noise level around a
double-decked road structure. Environ Monit Assess 105: 121-143. http://dx.doi.org/10.1007/sl0661-005-
3160-9
Tolbert. P. (2009). Personal communication from Dr. Paige Tolbert, Emory University to Dr. Ellen Kirrane,
NCEA with attachment. Available online
Tolbert. PE: Mulholland. JA: Macintosh. PL: Xu. F: Daniels. D: Devine. OJ: Carlin. BP: Klein. M: Dorlev. J:
Butler. AJ: Nordenberg. DF: Frumkin. H: Ryan. PB: White. MC. (2000). Air quality and pediatric emergency
room visits for asthma in Atlanta, Georgia. Am J Epidemiol 151: 798-810.
Tolbert. PE: Klein. M: Peel JL: Sarnat SE: Sarnat JA. (2007). Multipollutant modeling issues in a study of
ambient air quality and emergency department visits in Atlanta. J Expo Sci Environ Epidemiol 17: S29-S35.
http://dx.doi.org/10.1038/si.ies.7500625
Tonne. C: Wilkinson. P. (2013). Long-term exposure to air pollution is associated with survival following acute
coronary syndrome. European Heart Journal (Online) 34: 1306-1311.
http://dx.doi.org/10.1093/eurheartj/ehs480
Tonnesen. S: Jeffries. HE. (1994). Inhibition of odd oxygen production in the carbon bond four and generic
reaction set mechanisms. Atmos Environ 28: 1339-1349. http://dx.doi.org/10.1016/1352-2310(94)90281-X
Tonse. SR: Brown. NJ: Harlev. RA: Jin. L. (2008). A process-analysis based study of the ozone weekend effect.
Atmos Environ 42: 7728-7736. http://dx.doi.0rg/10.1016/j.atmosenv.2008.05.061
Toren. K: Brisman. J: Jarvholm. B. (1993). Asthma and asthma-like symptoms in adults assessed by
questionnaires: A literature review [Review]. Chest 104: 600-608. http://dx.doi.org/10.1378/chest. 104.2.600
Torok. JA: Brahmaiothi. MV: Zhu. H: Tinch. BT: Auten. RL: McMahon. TJ. (2012). Transpulmonary flux of S-
nitrosothiols and pulmonary vasodilation during nitric oxide inhalation: Role of transport. Am J Respir Cell
Mol Biol 47: 37-43. http://dx.doi.org/10.1165/rcmb.2011-0439OC
Trasande. L: Wong. K: Roy. A: Savitz. DA: Thurston. G. (2013). Exploring prenatal outdoor air pollution, birth
outcomes and neonatal health care utilization in a nationally representative sample. J Expo Sci Environ
Epidemiol 23: 315-321. http://dx.doi.org/10.1038/ies.2012.124
R-93
-------
Tsai DH: Wang. JL: Chuang. KJ: Chan. CC. (2010). Traific-related air pollution and cardiovascular mortality in
central Taiwan. Sci Total Environ 408: 1818-1823. http://dx.doi.0rg/10.1016/i.scitotenv.2010.01.044
Tsai. SS: Chen. CC: Hsieh. HJ: Chang. CC: Yang. CY. (2006). Air pollution and postneonatal mortality in a
tropical city: Kaohsiung, Taiwan. Inhal Toxicol 18: 185-189. http://dx.doi.org/10.1080/08958370500434214
Tsai. SS: Chen. PS: Yang. YH: Liou. SH: Wu. TN: Sung. FC: Yang. CY. (2012). Air pollution and hospital
admissions for myocardial infarction: Are there potentially sensitive groups? J Toxicol Environ Health A 75:
242-251. http://dx.doi.org/10.1080/15287394.2012.641202
Tsai. SS: Chiu. HF: Wu. TN: Yang. CY. (2009). Air pollution and emergency room visits for cardiac arrhythmia
in a subtropical city: Taipei, Taiwan. Inhal Toxicol 21: 1113-1118.
http://dx.doi.org/10.3109/08958370902758939
Tsai. SS: Goggins. WE: Chiu. HF: Yang. CY. (2003). Evidence for an association between air pollution and
daily stroke admissions in Kaohsiung, Taiwan. Stroke 34: 2612-2616.
http://dx.doi.org/10.1161/01.STR.0000095564.33543.64
Tseng. CY: Huang. YC: Su. SY: Huang. JY: Lai. CH: Lung. CC: Ho. CC: Liaw. YP. (2012). Cell type
specificity of female lung cancer associated with sulfur dioxide from air pollutants in Taiwan: An ecological
study. BMC Public Health 12: 4. http://dx.doi.org/10.1186/1471-2458-12-4
Tsubone. H: Oda. H: Suzuki. AK: Kubota. K. (1982). Electrocardiographic abnormalities in rats by acute
exposure to nitrogen dioxide. Toxicol Lett 12: 125-129. http://dx.doi.org/10.1016/0378-4274(82)90174-6
Tsubone. H: Suzuki. AK. (1984). Reflex cardiopulmonary responses by stimulation to type J receptors in rats
exposed to NO2. J Toxicol Environ Health 13: 905-917. http://dx.doi.org/10.1080/15287398409530550
Tsuii H: Larson. MG: Venditti. FJ. Jr: Manders. ES: Evans. JC: Feldman. CL: Lew. D. (1996). Impact of
reduced heart rate variability on risk for cardiac events. The Framingham Heart Study. Circulation 94: 2850-
2855. http://dx.doi.org/10.1161/01.CIR.94.ll.2850
Tsuii. H: Venditti. FJ: Manders. ES: Evans. JC: Larson. MG: Feldman. CL: Lew. D. (1994). Reduced heart rate
variability and mortality risk in an elderly cohort: The Framingham Heart Study. Circulation 90: 878-883.
http://dx.doi.Org/10.1161/01.CIR.90.2.878
Tunnicliffe. WS: Burge. PS: Ayres. JG. (1994). Effect of domestic concentrations of nitrogen dioxide on airway
responses to inhaled allergen in asthmatic patients. Lancet 344: 1733-1736. http://dx.doi.org/10.1016/s0140-
6736(94)92886-x
Turin. TC: Kita. Y: Rumana. N: Nakamura. Y: Ueda. K: Takashima. N: Sugihara. H: Morita. Y: Ichikawa. M:
Hirose. K: Nitta. H: Okavama. A: Miura. K: Ueshima. H. (2012). Short-term exposure to air pollution and
incidence of stroke and acute myocardial infarction in a Japanese population. Neuroepidemiology 38: 84-92.
http://dx.doi.org/10.1159/000335654
U.S. Census Bureau. (2007). 2007 American community survey [Database]. Washington, DC: U.S. Department
of Commerce, U.S. Census Bureau. Retrieved from http://www.census.gov/programs-survevs/acs/news/data-
releases.2007.html
U.S. Census Bureau. (2009). American housing survey for the United States: 2009. Washington, DC: U.S.
Government Printing Office, http://www.census.gov/prod/2011pubs/hl50-09.pdf
U.S. Census Bureau. (2013a). 2013 TIGER/Line shapefiles. Suitland, MD: U.S. Census Bureau, Geography
Division. https://www.census.gov/cgi-bin/geo/shapefiles2013/main
U.S. Census Bureau. (2013b). American community survey 2013 1-year estimate. Available online at
http://factfinder2.census.gov/faces/tableservices/jsf/pages/productview.xhtml?pid=ACS 13 1YR B01003&
prodType=table
U.S. Census Bureau. (2014). TIGER/Line with selected demographic and economic data. Suitland, MD: U.S.
Census Bureau, Geography Division. https://www.census.gov/geo/maps-data/data/tiger-data.html
R-94
-------
U.S. EPA (U.S. Environmental Protection Agency). (1971). Air quality criteria for nitrogen oxides [EPA
Report]. (AP-84). Washington DC: U.S. Environmental Protection Agency, Air Pollution Control Office.
http://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=20013K3B.txt
U.S. EPA (U.S. Environmental Protection Agency). (1982). Air quality criteria for oxides of nitrogen [EPA
Report]. (EPA/600/8-82-026). Research Triangle Park, NC: U.S. Environmental Protection Agency,
Environmental Criteria and Assessment Office.
U.S. EPA (U.S. Environmental Protection Agency). (1991). Guidelines for developmental toxicity risk
assessment. In US Environmental Protection Agency. (EPA/600/FR-91/001). Washington, DC: U.S.
Environmental Protection Agency, Risk Assessment Forum.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=23162
U.S. EPA (U.S. Environmental Protection Agency). (1993a). Air quality criteria for oxides of nitrogen, vol. 1-3
[EPA Report]. (EPA/600/8-9!/049aF-cF). Research Triangle Park, NC: U.S. Environmental Protection
Agency, Environmental Criteria and Assessment Office.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=40179
U.S. EPA (U.S. Environmental Protection Agency). (1993b). Ambient air quality surveillance: Final rule. Fed
Reg 58: 8452-8475.
U.S. EPA (U.S. Environmental Protection Agency). (1995a). Review of the national ambient air quality
standards for nitrogen dioxide: Assessment of scientific and technical information [EPA Report].
(EPA/452/R-95/005). Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air
Quality, Planning and Standards, http://nepis.epa.gov/exe/ZvPURL.cgi?Dockev=00002UBE.txt
U.S. EPA (U.S. Environmental Protection Agency). (1995b). A user's guide for the CALPUFF dispersion model
[EPA Report]. (EPA/454/B-95/006). Research Triangle Park, NC: U.S. Environmental Protection Agency,
Office of Air Quality, Planning, and Standards.
http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockev=2001206A.txt
U.S. EPA (U.S. Environmental Protection Agency). (1996a). Air quality criteria for ozone and related
photochemical oxidants [EPA Report]. (EPA/600/P-93/004AF). Research Triangle Park, NC.
U.S. EPA (U.S. Environmental Protection Agency). (1996b). Guidelines for reproductive toxicity risk
assessment. (EPA/630/R-96/009). Washington, DC: U.S. Environmental Protection Agency, Risk
Assessment Forum, http://www.epa.gov/raf/publications/guidelines-reproductive-tox-risk-assessment.htm
U.S. EPA (U.S. Environmental Protection Agency). (1998a). Guidelines for ecological risk assessment [EPA
Report]. (EPA/630/R-95/002F). Washington, DC: U.S. Environmental Protection Agency, Risk Assessment
Forum, http://www.epa.gov/raf/publications/guidelines-ecological-risk-assessment.htm
U.S. EPA (U.S. Environmental Protection Agency). (1998b). Guidelines for neurotoxicity risk assessment [EPA
Report]. (EPA/630/R-95/001F). Washington, DC: U.S. Environmental Protection Agency, Risk Assessment
Forum, http://www.epa.gov/raf/publications/guidelines-neurotoxicitv-risk-assessment.htm
U.S. EPA (U.S. Environmental Protection Agency). (1999). Science algorithms of the EPA models-3
Community Multiscale Air Quality (CMAQ) modeling system [EPA Report]. (EPA/600/R-99/030).
Washington, DC. http://nepis.epa.gov/exe/ZvPURL.cgi?Dockev=30003R9Y.txt
U.S. EPA (U.S. Environmental Protection Agency). (2001). Control of air pollution from new motor vehicles:
Heavy-duty engine and vehicle standards and highway diesel fuel sulfur control requirements. Fed Reg 66:
5002-5193.
U.S. EPA (U.S. Environmental Protection Agency). (2002). A framework for assessing and reporting on
ecological condition: An SAB report [EPA Report]. (EPA-SAB-EPEC-02-009). Washington, DC.
http://vosemite.epa.gov/sab%5CSABPRODUCT.NSF/7700D7673673CE83852570CA0075458A/$File/epec
02009.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2005). Guidelines for carcinogen risk assessment [EPA
Report]. (EPA/630/P-03/00IF). Washington, DC: U.S. Environmental Protection Agency, Risk Assessment
Forum, http://www2.epa.gov/osa/guidelines-carcinogen-risk-assessment
R-95
-------
U.S. EPA (U.S. Environmental Protection Agency). (2006). Air quality criteria for ozone and related
photochemical oxidants [EPA Report]. (EPA/600/R-05/004AF). Research Triangle Park, NC: U.S.
Environmental Protection Agency, National Center for Environmental Assessment.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=149923
U.S. EPA (U.S. Environmental Protection Agency). (2007a). Integrated review plan for the primary national
ambient air quality standard for nitrogen dioxide [EPA Report]. Research Triangle Park, NC: U.S.
Environmental Protection Agency, National Center for Environmental Assessment.
http://www.epa.gov/ttiVnaaqs/standards/nox/data/20070823 nox reviewjlanfmal.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2007b). Nitrogen dioxide health assessment plan: Scope
and methods for exposure and risk assessment. Draft [EPA Report]. Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards.
http://www.epa.gov/ttn/naaqs/standards/nox/data/20070927 risk exposure scope.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2008a). Integrated science assessment for oxides of
nitrogen - Health criteria: Annexes [EPA Report]. (EPA/600/R-08/072). Research Triangle Park, NC: U.S.
Environmental Protection Agency, National Center for Environmental Assessment.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=194645
U.S. EPA (U.S. Environmental Protection Agency). (2008b). Integrated science assessment for oxides of
nitrogen and sulfur: Ecological criteria [EPA Report]. (EPA/600/R-08/082F). Research Triangle Park, NC.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=201485
U.S. EPA (U.S. Environmental Protection Agency). (2008c). Integrated science assessment for oxides of
nitrogen Health criteria (Final report, 2008) [EPA Report]. (EPA/600/R-08/071). Research Triangle Park,
NC: U.S. Environmental Protection Agency, National Center for Environmental Assessment.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=194645
U.S. EPA (U.S. Environmental Protection Agency). (2008d). Integrated science assessment for sulfur oxides:
Health criteria [EPA Report]. (EPA/600/R-08/047F). Research Triangle Park, NC: U.S. Environmental
Protection Agency, National Center for Environmental Assessment.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=198843
U.S. EPA (U.S. Environmental Protection Agency). (2008e). Risk and exposure assessment to support the
review of the NO2 primary national ambient air quality standard [EPA Report]. (EPA-452/R-08-008a).
Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, http://www.epa.gov/ttn/naaqs/standards/nox/data/20081121 NO2REA final.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2009a). Integrated science assessment for paniculate matter
[EPA Report]. (EPA/600/R-08/139F). Research Triangle Park, NC: U.S. Environmental Protection Agency,
National Center for Environmental Assessment.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=216546
U.S. EPA (U.S. Environmental Protection Agency). (2009b). Integrated science assessment for paniculate matter
- Annexes with errata [EPA Report]. (EPA/600/R-08/139). Washington, DC: U.S. Environmental Protection
Agency, National Center for Environmental Assessment.
http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid=216546
U.S. EPA (U.S. Environmental Protection Agency). (2010a). Final regulatory impact analysis (RIA) for the NO2
National Ambient Air Quality Standards (NAAQS) [EPA Report]. Research Triangle Park, NC.
http://www.epa.gov/ttnecasl/regdata/RIAs/FinalNO2RIAfulldocument.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2010b). Integrated science assessment for carbon monoxide
[EPA Report]. (EPA/600/R-09/019F). Research Triangle Park, NC: U.S. Environmental Protection Agency,
Office of Research and Development, National Center for Environmental Assessment.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=218686
U.S. EPA (U.S. Environmental Protection Agency). (2010c). Primary national ambient air quality standards for
nitrogen dioxide. Fed Reg 75: 6474-6537.
R-96
-------
U.S. EPA (U.S. Environmental Protection Agency). (201 la). 2008 National Emissions Inventory data and
documentation. Available online at http://www.epa.gov/ttn/chief/net/2008inventorv.html (accessed February
17,2011).
U.S. EPA (U.S. Environmental Protection Agency). (201 Ib). Exposure factors handbook: 2011 edition (final)
[EPA Report]. (EPA/600/R-090/052F). Washington, DC: U.S. Environmental Protection Agency, Office of
Research and Development, National Center for Environmental Assessment.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=236252
U.S. EPA (U.S. Environmental Protection Agency). (2012a). Ambient air monitoring reference and equivalent
methods: Designation of three new equivalent methods. Fed Reg 77: 32632-32633.
U.S. EPA (U.S. Environmental Protection Agency). (2012b). Near-road NO2 monitoring technical assistance
document [EPA Report]. (EPA-454/B-12-002). Research Triangle Park, NC: U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards.
http://nepis.epa.gov/exe/ZvPURL.cgi?Dockev=P100EPRZ.txt
U.S. EPA (U.S. Environmental Protection Agency). (2012c). Notice of workshop and call for information on
integrated science assessment for oxides of nitrogen. Fed Reg 77: 7149-7151.
U.S. EPA (U.S. Environmental Protection Agency). (2013a). 2011 National Emissions Inventory data and
documentation. Available online at http://www.epa.gov/ttnchiel/net/2011inventorv.html (accessed December
1,2014).
U.S. EPA (U.S. Environmental Protection Agency). (2013b). Ambient air monitoring reference and equivalent
methods: Designation of five new equivalent methods. Fed Reg 78: 67360-67361.
U.S. EPA (U.S. Environmental Protection Agency). (2013c). Integrated science assessment for lead [EPA
Report]. (EPA/600/R-10/075F). Research Triangle Park, NC: U.S. Environmental Protection Agency,
National Center for Environmental Assessment.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=255721
U.S. EPA (U.S. Environmental Protection Agency). (2013d). Integrated science assessment for ozone and related
photochemical oxidants. (EPA/600/R-10/076F). Research Triangle Park, NC: U.S. Environmental Protection
Agency, National Center for Environmental Assessment.
http://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=247492
U.S. EPA (U.S. Environmental Protection Agency). (2013e). Integrated science assessment for ozone and related
photochemical oxidants (final report) [EPA Report]. (EPA/600/R-10/076F). Washington, DC: U.S.
Environmental Protection Agency, National Center for Environmental Assessment.
http://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=247492
U.S. EPA (U.S. Environmental Protection Agency). (20131). Notice of workshop and call for information on
integrated science assessment for oxides of nitrogen and oxides of sulfur. Fed Reg 78: 53452-53454.
U.S. EPA (U.S. Environmental Protection Agency). (2013g). Supplemental table S6-2. Epidemiologic studies of
exposure to oxides of nitrogen and pregnancy effects [EPA Report].
U.S. EPA (U.S. Environmental Protection Agency). (2013h). Supplemental table S6-3. Epidemiologic studies of
exposure to oxides of nitrogen and fetal growth [EPA Report].
U.S. EPA (U.S. Environmental Protection Agency). (2013i). Supplemental table S6-4. Epidemiologic studies
exposure to oxides of nitrogen and preterm birth [EPA Report].
U.S. EPA (U.S. Environmental Protection Agency). (2013J). Supplemental table S6-5. Epidemiologic studies of
exposure to oxides of nitrogen and birth weight [EPA Report].
U.S. EPA (U.S. Environmental Protection Agency). (2013k). Supplemental table S6-6. Epidemiologic studies of
exposure to oxides of nitrogen and birth defects [EPA Report].
U.S. EPA (U.S. Environmental Protection Agency). (20131). Supplemental table S6-7. Epidemiologic studies of
exposure to nitrogen dioxide and neurodevelopmental effects [EPA Report].
R-97
-------
U.S. EPA (U.S. Environmental Protection Agency). (2013m). Supplemental table S6-8. Epidemiologic studies of
exposure to nitrogen dioxide and infant mortality [EPA Report].
U.S. EPA (U.S. Environmental Protection Agency). (2013n). Supplemental table S6-9. Epidemiologic studies of
long-term exposure to oxides of nitrogen and mortality [EPA Report].
U.S. EPA (U.S. Environmental Protection Agency). (2013o). Supplemental table S6-10. Epidemiologic studies
of exposure to oxides of nitrogen and cancer incidence and mortality [EPA Report].
U.S. EPA (U.S. Environmental Protection Agency). (2013p). Toxicological review of trimethylbenzenes
(CASRN 25551-13-7, 95-63-6, 526-73-8, and 108-67-8) in support of summary information on the
Integrated Risk Information System (IRIS): revised external review draft [EPA Report].
(EPA/635/R13/171a). Washington, D.C.: U.S. Environmental Protection Agency, National Center for
Environmental Assessment.
http://YOsemite.epa.gov/sab/SABPRODUCT.NSF^5d8alce9b07293485257375007012b7/eele280e77586de
985257b65005d37e7!OpenDocument
U.S. EPA (U.S. Environmental Protection Agency). (2014a). Control of air pollution from motor vehicles: Tier3
motor vehicle emission and fuel standards. Fed Reg 79: 23414-23886.
U.S. EPA (U.S. Environmental Protection Agency). (2014b). Integrated review plan for the primary national
ambient air quality standards for nitrogen dioxide [EPA Report]. (EPA-452/R-14/003). Research Triangle
Park, NC: U.S. Environmental Protection Agency, National Center for Environmental Assessment.
http://www.epa.gov/ttn/naaqs/standards/nox/data/201406fmalirpprimarvno2.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2014c). Notice of workshop and call for information on
integrated science assessment for paniculate matter. Fed Reg 79: 71764-71766.
U.S. EPA (U.S. Environmental Protection Agency). (2014d). Office of research and development; Ambient air
monitoring reference and equivalent methods: Designation of four new equivalent methods. Fed Reg 79:
34734-34735.
U.S. EPA (U.S. Environmental Protection Agency). (2015a). Supplemental figure S5-1 associations of NO2 with
respiratory effects in copollutant models with PM10 PM10_25 SO2 O3.
U.S. EPA (U.S. Environmental Protection Agency). (2015b). Supplemental Figure S5-2. Results of single-
pollutant and copollutants models of short-term exposure to NO2 with and without PM and CVD HA.
U.S. EPA (U.S. Environmental Protection Agency). (2015c). Supplemental Figure S5-3. Results of single-
pollutant and copollutants models of short-term exposure to NO2 with CO and without CO and CVD HA.
U.S. EPA (U.S. Environmental Protection Agency). (2015d). Supplemental Figure S5-4 Results of single-
pollutant and copollutants models of short-term exposure to NO2 or NOX with and without O3 and hospital
admissions CVD.
U.S. EPA (U.S. Environmental Protection Agency). (2015e). Supplemental Figure S5-5 Results of single-
pollutant and copollutants models of short-term exposure to NO2 or NOX with and without SO2 and hospital
admissions CVD.
U.S. EPA (U.S. Environmental Protection Agency). (2015f). Supplemental table Sl-1. Epidemiologic studies of
health effects not evaluated in the ISA for oxides for nitrogen [EPA Report].
U.S. EPA (U.S. Environmental Protection Agency). (2015g). Supplemental Table S5-1. Calculation of
increments of oxides of nitrogen for standardizing epidemiologic effect estimates - short-term averages.
U.S. EPA (U.S. Environmental Protection Agency). (2015h). Supplemental Table S5-3. Studies of short-term
exposure to oxides of nitrogen and respiratory hospital admissions and ED visits [EPA Report].
U.S. EPA (U.S. Environmental Protection Agency). (2015i). Supplemental Table S5-4. Studies of short-term
exposure to oxides of nitrogen and mortality.
U.S. EPA (U.S. Environmental Protection Agency). (2015J). Supplemental Table S6-1 Calculation of increments
of oxides of nitrogen for standardizing effect estimates.
R-98
-------
Unal A: Hu. Y: Chang. ME: Talat Odman. M: Russell AG. (2005). Airport related emissions and impacts on air
quality: Application to the Atlanta International Airport. Atmos Environ 39: 5787-5798.
http://dx.doi.0rg/10.1016/i.atmosenv.2005.05.051
UNEP (United Nations Environment Programme). (2003). Ecosystems and human well-being: A framework for
assessment. Washington, DC: Island Press. https://www.islandpress.org/book/ecosvstems-and-human-well-
being-0
Urman. R: McConnell R: Islam. T: Avol EL: Lurmann. FW: Vora. H: Linn. WS: Rappaport. EB: Gilliland. FD:
Gauderman. WJ. (2014). Associations of children's lung function with ambient air pollution: Joint effects of
regional and near-roadway pollutants. Thorax 69: 540-547. http://dx.doi.org/10.1136/thoraxinl-2012-203159
Vagagginj B: Paggiaro. PL: Giannini. D: Franco. AD: Cianchetti S: Carnevali S: Taccola. M: Bacci. E:
Bancalari. L: Dente. FL: Giuntini. C. (1996). Effect of short-term NO2 exposure on induced sputum in
normal, asthmatic and COPD subjects. Eur Respir J 9: 1852-1857.
Valin. LC: Russell. AR: Cohen. RC. (2013). Variations of OH radical in an urban plume inferred from NO 2
column measurements. Geophys Res Lett 40: 1856-1860. http://dx.doi.org/10.1002/grl.50267
Van As. A. (1977). Pulmonary airway clearance mechanisms: A reappraisal [Editorial]. Am Rev Respir Dis 115:
721-726.
van den Hooven. EH: de Kluizenaar. Y: Pierik. FH: Hofman. A: van Ratingen. SW: Zandveld. PY: Lindemans.
J: Russcher. H: Steegers. EA: Miedema. HM: Jaddoe. VW. (2012a). Chronic air pollution exposure during
pregnancy and maternal and fetal C-reactive protein levels: The generation R study. Environ Health Perspect
120: 746-751. http://dx.doi.org/10.1289/ehp. 1104345
van den Hooven. EH: de Kluizenaar. Y: Pierik. FH: Hofman. A: van Ratingen. SW: Zandveld. PY: Mackenbach.
JP: Steegers. EA: Miedema. HM: Jaddoe. VW. (2011). Air pollution, blood pressure, and the risk of
hypertensive complications during pregnancy: The generation R study. Hypertension 57: 406-412.
http://dx.doi.org/10.1161/HYPERTENSIONAHA. 110.164087
van den Hooven. EH: Pierik. FH: de Kluizenaar. Y: Hofman. A: van Ratingen. SW: Zandveld. PY: Russcher. H:
Lindemans. J: Miedema. HM: Steegers. EA: Jaddoe. VW. (2012b). Air pollution exposure and markers of
placental growth and function: The Generation R Study. Environ Health Perspect 120: 1753-1759.
http://dx.doi.org/10.1289/ehp.1204918
van den Hooven. EH: Pierik. FH: de Kluizenaar. Y: Willemsen. SP: Hofman. A: van Ratingen. SW: Zandvel.
PYJ: Mackenbach. JP: Steegers. EAP: Miedema. HME: Jaddoe. VWV. (2012c). Air pollution exposure
during pregnancy, ultrasound measures of fetal growth, and adverse birth outcomes: A prospective cohort
study. Environ Health Perspect 120: 150-156. http://dx.doi.org/10.1289/ehp.1003316
van der Heijden. HHA. CM: Brouwer. ML: Hoekstra. F: van der Pol. P: Merkus. PJF. M. (2014). Reference
values of exhaled nitric oxide in healthy children 1-5 years using off-line tidal breathing. Pediatr Pulmonol
49: 291-295. http://dx.doi.org/10.1002/ppul.22796
van der Vliet. A: O'Neill CA: Cross. CE: Koostra. JM: Volz. WG: Halliwell. B: Louie. S. (1999). Determination
of low-molecular-mass antioxidant concentrations in human respiratory tract lining fluids. Am J Physiol 276:
L289-L296.
van der Zee. SC: Hoek. G: Boezen. HM: Schouten. JP: van Wijnen. JH: Brunekreef. B. (1999). Acute effects of
urban air pollution on respiratory health of children with and without chronic respiratory symptoms. Occup
Environ Med 56: 802-813. http://dx.doi.org/10.1136/oem.56.12.802
van der Zee. SC: Hoek. G: Boezen. MH: Schouten. JP: van Wijnen. JH: Brunekreef. B. (2000). Acute effects of
air pollution on respiratory health of 50-70 yr old adults. Eur Respir J 15: 700-709.
vanKempen. E: Fischer. P: Janssen. N: Houthuijs. D: vanKamp. I: Stansfeld. S: Cassee. F. (2012).
Neurobehavioral effects of exposure to traffic-related air pollution and transportation noise in primary
schoolchildren. Environ Res 115: 18-25. http://dx.doi.0rg/10.1016/i.envres.2012.03.002
R-99
-------
Van Roosbroeck. S: Li. R: Hoek. G: Lebret E: Brunekreef. B: Spiegelman. D. (2008). Traffic-related outdoor air
pollution and respiratory symptoms in children: The impact of adjustment for exposure measurement error.
Epidemiology 19: 409-416. http://dx.doi.org/10.1097/EDE.Ob013e3181673bab
Van Schavck. CP: Dompeling. E: Van Herwaarden. CL: Wever. AM: Van Weel C. (1991). Interacting effects of
atopy and bronchial hyperresponsiveness on the annual decline in lung function and the exacerbation rate in
asthma. Am Rev Respir Dis 144: 1297-1301. http://dx.doi.org/10.1164/ajrccm/144.6.1297
Van Stee. EW: Sloane. RA: Simmons. JE: Brunnemann. KD. (1983). In vivo formation of N-nitrosomorpholine
in CD-I mice exposed by inhalation to nitrogen dioxide and by gavage to morpholine. J Natl Cancer Inst 70:
375-379. http://dx.doi.0rg/10.1093/inci/70.2.375
Van Stee. EW: Sloane. RA: Simmons. JE: Moorman. MP: Brunnemann. KD. (1995). Endogenous formation of
N-nitrosomorpholine in mice from 15NO2 by inhalation and morpholine by gavage. Carcinogenesis 16: 89-
92. http://dx.doi.0rg/10.1093/carcin/16.l.89
Vardoulakis. S: Lumbreras. J: Solazzo. E. (2009). Comparative evaluation of nitrogen oxides and ozone passive
diffusion tubes for exposure studies. Atmos Environ 43: 2509-2517.
http://dx.doi.0rg/10.1016/i.atmosenv.2009.02.048
Vardoulakis. S: Solazzo. E: Lumbreras. J. (2011). Intra-urban and street scale variability of BTEX, NO2 and O3
in Birmingham, UK: Implications for exposure assessment. Atmos Environ 45: 5069-5078.
http://dx.doi.0rg/10.1016/i.atmosenv.2011.06.038
Velsor. LW: Ballinger. CA: Patel J: Postlethwait. EM. (2003). Influence of epithelial lining fluid lipids on NO2-
induced membrane oxidation and nitration. Free Radic Biol Med 34: 720-733.
http://dx.doi.org/10.1016/S0891-5849(02)01370-9
Velsor. LW: Postlethwait. EM. (1997). NO2-induced generation of extracellular reactive oxygen is mediated by
epithelial lining layer antioxidants. Am J Physiol 17: L1265-L1275.
Vencloviene. J: Grazuleviciene. R: Babarskiene. R: Dedele. A: Grazulevicius. T. (2011). Short-term nitrogen
dioxide exposure and geomagnetic activity interaction: contribution to emergency hospitalization for acute
coronary syndrome. Int J Environ Health Res 21: 149-160. http://dx.doi.org/10.1080/09603123.2010.515671
Venkatram A: Isakov. V: Thoma. E: Baldauf. R. (2007). Analysis of air quality data near roadways using a
dispersion model. Atmos Environ 41: 9481-9497. http://dx.doi.0rg/10.1016/i.atmosenv.2007.08.045
Verbanck. S: Schuermans. D: Noppen. M: Vincken. W: Paiva. M. (2001). Methacholine versus histamine:
Paradoxical response of spirometry and ventilation distribution. J Appl Physiol 91: 2587-2594.
Vette. A: Burke. J: Norris. G: Landis. M: Batterman. S: Breen. M: Isakov. V: Lewis. T: Gilmour. MI: Kamal. A:
Hammond. D: Vedantham. R: Bereznicki. S: Tian. N: Croghan. C. (2013). The Near-Road Exposures and
Effects of Urban Air Pollutants Study (NEXUS): Study design and methods. Sci Total Environ 448: 38-47.
http://dx.doi.0rg/10.1016/i.scitotenv.2012.10.072
Vichj F: De Santis. F. (2012). The measurement of the sink properties of triethanolamine (TEA) as a coating for
collecting NO2 by using annular diffusion denuders. Environ Technol 33: 1065-1069.
http://dx.doi.org/10.1080/09593330.2011.606848
Vidale. S: Bonanomi. A: Guidotti M: Arnaboldi M: Sterzi. R. (2010). Air pollution positively correlates with
daily stroke admission and in hospital mortality: A study in the urban area of Como, Italy. Neurol Sci 31:
179-182. http://dx.doi.org/10.1007/sl0072-009-0206-8
Vieira. SE: Stein. RT: Ferraro. AA: Pastro. LD: Pedro. SS: Lemos. M: da Silva. ER: Sly. PD: Saldiva. PH.
(2012). Urban air pollutants are significant risk factors for asthma and pneumonia in children: The influence
of location on the measurement of pollutants. Arch Bronconeumol 48: 389-395.
http://dx.doi.0rg/10.1016/i.arbres.2012.05.005
Villena. G: Beian. I: Kurtenbach. R: Wiesen. P: Kleffmann. J. (2011). Development of a new Long Path
Absorption Photometer (LOPAP) instrument for the sensitive detection of NO2 in the atmosphere. Atmos
Meas Tech 4: 1663-1676. http://dx.doi.org/10.5194/amt-4-1663-2011
R-100
-------
Villena. G: Bejan. I: Kurtenbach. R: Wiesen. P: Kleffmann. J. (2012). Interferences of commercial NO2
instruments in the urban atmosphere and in a smog chamber. Atmos Meas Tech 5: 149-159.
http://dx.doi.org/10.5194/amt-5-149-2012
Villeneuve. PJ: Chen. L: Rowe. BH: Coates. F. (2007). Outdoor air pollution and emergency department visits
for asthma among children and adults: A case-crossover study in northern Alberta, Canada. Environ Health
6: 40. http://dx.doi.org/10.1186/1476-069X-6-40
Villeneuve. PJ: Chen. L: Stieb. D: Rowe. BH. (2006a). Associations between outdoor air pollution and
emergency department visits for stroke in Edmonton, Canada. Eur J Epidemiol 21: 689-700.
http://dx.doi.org/10.1007/sl0654-006-9050-9
Villeneuve. PJ: Doiron. MS: Stieb. D: Dales. R: Burnett. RT: Dugandzic. R. (2006b). Is outdoor air pollution
associated with physician visits for allergic rhinitis among the elderly in Toronto, Canada? Allergy 61: 750-
758. http://dx.doi.org/10.1111/i. 1398-9995.2006.01070.X
Villeneuve. PJ: Jerrett M: Brenner. D: Su. J: Chen. H: Mclaughlin. J. R. (2014). A case-control study of long-
term exposure to ambient volatile organic compounds and lung cancer in Toronto, Ontario, Canada. Am J
Epidemiol 179: 443-451. http://dx.doi.org/10.1093/aie/kwt289
Villeneuve. PJ: Johnson. JY: Pasichnyk. D: Lowes. J: Kirkland. S: Rowe. BH. (2012). Short-term effects of
ambient air pollution on stroke: Who is most vulnerable? Sci Total Environ 430: 193-201.
http://dx.doi.0rg/10.1016/i.scitotenv.2012.05.002
Vinikoor-Imler. LC: Owens. EO: Nichols. JL: Ross. M: Brown. JS: Sacks. JD. (2014). Evaluating potential
response-modifying factors for associations between ozone and health outcomes: A weight-of-evidence
approach [Review]. Environ Health Perspect 122: 1166-1176. http://dx.doi.org/10.1289/ehp.1307541
Vints. AM: Oostveen. E: Eeckhaut G: Smolders. M: De Backer. WA. (2005). Time-dependent effect of nitrate-
rich meals on exhaled nitric oxide in healthy subjects. Chest 128: 2465-2470.
http://dx.doi.0rg/10.1378/chest.128.4.2465
Volk. HE: Kerin. T: Lurmann. F: Hertz-Picciotto. I: Mcconnell R: Campbell DB. (2014). Autism spectrum
disorder: Interaction of air pollution with the MET receptor tyro sine kinase gene. Epidemiology 25: 44-47.
http://dx.doi.org/10.1097/EDE.0000000000000030
Volk. HE: Lurmann. F: Penfold. B: Hertz-Picciotto. I: McConnell R. (2013). Traffic-related air pollution,
paniculate matter, and autism. Arch Gen Psychiatry 70: 71-77.
http://dx.doi.org/10.1001/iamapsvchiatry.2013.266
Vollmuth. TA: Driscoll. KE: Schlesinger. RB. (1986). Changes in early alveolar particle clearance due to single
and repeated nitrogen dioxide exposures in the rabbit. J Toxicol Environ Health 19: 255-266.
http://dx.doi.org/10.1080/15287398609530925
von Elm. E: Altaian. DG: Egger. M: Pocock. SJ: Getzsche. PC: Vandenbroucke. JP. (2007). The Strengthening
the Reporting of Observational Studies in Epidemiology (STROBE) statement: Guidelines for reporting
observational studies [Review]. PLoS Med 4: e296. http://dx.doi.org/10.1371/journal.pmed.0040296
vonKlot. S: Peters. A: Aalto. P: Bellander. T: Berglind. N: D'Ippoliti. D: Elosua. R: Hormann. A: Kulmala. M:
Lanki. T: Lowel H: Pekkanen. J: Picciotto. S: Sunyer. J: Forastiere. F. (2005). Ambient air pollution is
associated with increased risk of hospital cardiac readmissions of myocardial infarction survivors in five
European cities. Circulation 112: 3073-3079. http://dx.doi.org/10.1161/CIRCULATIONAHA.105.548743
vonKlot S: Wolke. G: Tuch. T: Heinrich. J: Dockery. DW: Schwartz. J: Kreyling. WG: Wichmann. HE: Peters.
A^ (2002). Increased asthma medication use in association with ambient fine and ultrafme particles. Eur
Respir J 20: 691-702. http://dx.doi.org/10.1183/09031936.02.01402001
von Nieding. G: Wagner. HM. (1979). Effects of NO2 on chronic bronchitics. Environ Health Perspect 29: 137-
142.
R-101
-------
Vrijheid. M: Martinez. D: Aguilera. I: Bustamante. M: Ballester. F: Estarlich. M: Fernandez-Somoano. A:
Guxens. M: Lertxundi. N: Martinez. MD: Tardon. A: Sunyer. J. (2012). Indoor air pollution from gas
cooking and infant neurodevelopment. Epidemiology 23: 23-32.
http://dx.doi.org/10.1097/EDE.Ob013e31823a4023
Vrijheid. M: Martinez. D: Manzanares. S: Dadvand. P: Schembari. A: Rankin. J: Nieuwenhuijsen. M. (2011).
Ambient air pollution and risk of congenital anomalies: A systematic review and meta-analysis [Review].
Environ Health Perspect 119: 598-606. http://dx.doi.org/10.1289/ehp. 1002946
Vutukuru. S: Dabdub. D. (2008). Modeling the effects of ship emissions on coastal air quality: A case study of
southern California. Atmos Environ 42: 3751-3764. http://dx.doi.0rg/10.1016/j.atmosenv.2007.12.073
Wagner. HM. (1970). [Absorption of NO and NO2 in MIK- and MAK-concentrations during inhalation].
[Absorption von NO und NO2 in MIK- und MAK-konzentrationen bei der inhalation]. Staub, Reinhaltung
derLuft 30: 380-381.
Wagner. WD: Duncan. BR: Wright. PG: Stokinger. HE. (1965). Experimental study of threshold limit of NO2.
Arch Environ Health 10: 455-466. http://dx.doi.org/10.1080/00039896.1965.10664029
Walker. SE: Slordal LH: Guerreiro. C: Gram F: Gronskei KE. (1999). Air pollution exposure monitoring and
estimation. Part II. Model evaluation and population exposure. J Environ Monit 1: 321-326.
http://dx.doi.org/10.1039/A902776I
Walters. LJ: May. MS: Johnson. DE: MacMann. RS: Woodward. WA. (1999). Time-variability of NO x
emissions from Portland cement kilns. Environ Sci Technol 33: 700-704.
http://dx.doi.org/10.1021/es980203b
Wang. J: Widlansky. ME. (2009). Lifestyle choices and endothelial function: Risk and relevance [Review]. Curr
Vase Pharmacol 7: 209-224. http://dx.doi.org/10.2174/157016109787455581
Wang. JH: Devalia. JL: Puddle. JM: Hamilton. SA: Davies. RJ. (1995a). Effect of six-hour exposure to nitrogen
dioxide on early-phase nasal response to allergen challenge in patients with a history of seasonal allergic
rhinitis. J Allergy Clin Immunol 96: 669-676.
Wang. JH: Devalia. JL: Rusznak. C: Bagnall A: Sapsford. RJ: Davies. RJ. (1999). Effect of fluticasone
propionate aqueous nasal spray on allergen-induced inflammatory changes in the nasal airways of allergic
rhinitics following exposure to nitrogen dioxide. Clin Exp Allergy 29: 234-240.
http://dx.doi.0rg/10.1046/i.1365-2222.1999.00440.x
Wang. JH: Puddle. J: Devalia. JL: Davies. RJ. (1995b). Nitrogen dioxide increases eosinophil activation in the
early-phase response to nasal allergen provocation. Int Arch Allergy Immunol 107: 103-105.
http://dx.doi.org/10.1159/000236944
Wang. M: Beelen. R: Basagana. X: Becker. T: Cesaroni. G: de Hoogh. K: Dedele. A: Declercq. C:
Dimakopoulou. K: Eeftens. M: Forastiere. F: Galassi. C: Grazuleviciene. R: Hoffmann. B: Heinrich. J:
lakovides. M: Kunzli. N: Korek. M: Lindlev. S: Molter. A: Mosler. G: Madsen. C: Nieuwenhuijsen. MJ:
Phuleria. H: Pedeli. X: Raaschou-Nielsen. O: Ranzi. A: Sugiri. D: Stempfelet M: Tsai. MY: Lanki. T:
Udvardv. O: Varro. MJ: Wolf. K: Weinmayr. G: Yli-Tuomi. T: Hoek. G: Brunekreef. B. (2013a). Evaluation
of land use regression models for NO2 and paniculate matter in 20 European study areas: The ESCAPE
project. Environ Sci Technol 47: 4357-4364. http://dx.doi.org/10.1021/es305129t
Wang. M: Beelen. R: Bellander. T: Birk. M: Cesaroni. G: Cirach. M: Cyrys. J: de Hoogh. K: Declercq. C:
Dimakopoulou. K: Eeftens. M: Eriksen. KT: Forastiere. F: Galassi. C: Grivas. G: Heinrich. J: Hoffmann. B:
Ineichen. A: Korek. M: Lanki. T: Lindlev. S: Modig. L: Molter. A: Nafstad. P: Nieuwenhuijsen. MJ: Nystad.
W: Olsson. D: Raaschou-Nielsen. O: Ragettli M: Ranzi. A: Stempfelet. M: Sugiri. D: Tsai. MY: Udvardv.
O: Varro. MJ: Vienneau. D: Weinmayr. G: Wolf. K: Yli-Tuomi. T: Hoek. G: Brunekreef. B. (2014).
Performance of multi-city land use regression models for nitrogen dioxide and fine particles. Environ Health
Perspect 122: 843-849. http://dx.doi.org/10.1289/ehp.1307271
Wang. M: Beelen. R: Eeftens. M: Meliefste. K: Hoek. G: Brunekreef. B. (2012). Systematic evaluation of land
use regression models for NO. Environ Sci Technol 46: 4481-4489. http://dx.doi.org/10.1021/es204183v
R-102
-------
Wang. R: Henderson. SB: Sbihi. H: Allen. RW: Brauer. M. (2013b). Temporal stability of land use regression
models for traffic-related air pollution. Atmos Environ 64: 312-319.
http://dx.doi.0rg/10.1016/i.atmosenv.2012.09.056
Wang. SO: Zhang. JL: Zeng. XD: Zeng. YM: Wang. SC: Chen. SY. (2009a). Association of traffic-related air
pollution with children's neurobehavioral functions in Quanzhou, China. Environ Health Perspect 117: 1612-
1618. http://dx.doi.org/10.1289/ehp.0800023
Wang. XY: Hu. W: Tong. S. (2009b). Long-term exposure to gaseous air pollutants and cardio-respiratory
mortality in Brisbane, Australia. Geospat Health 3: 257-263.
Wang. YJ: Zhang. KM. (2009). Modeling near-road air quality using a computational fluid dynamics model,
CFD-VIT-RIT. Environ Sci Technol 43: 7778-7783. http://dx.doi.org/10.1021/es9014844
Ward. DJ: Miller. MR: Walters. S: Harrison. RM: Ayres. JG. (2000). Impact of correcting peak flow for
nonlinear errors on air pollutant effect estimates from a panel study. Eur Respir J 15: 137-140.
Ward. DJ: Roberts. KT: Jones. N: Harrison. RM: Ayres. JG: Hussain. S: Walters. S. (2002). Effects of daily
variation in outdoor particulates and ambient acid species in normal and asthmatic children. Thorax 57: 489-
502. http://dx.doi.0rg/10.1136/thorax.57.6.489
Wardman. P. (1998). Nitrogen dioxide in biology: Correlating chemical kinetics with biological effects. In ZB
Alfassi (Ed.), N-centered radicals (pp. 155-179). Chichester, UK: Wiley.
Weaklev. J: Webber. MP: Ye. F: Zeig-Owens. R: Cohea HW: Hall. CB: Kelly. K: Prezant. DJ. (2013).
Agreement between obstructive airways disease diagnoses from self-report questionnaires and medical
records. Prev Med 57: 38-42. http://dx.doi.0rg/10.1016/i.vpmed.2013.04.001
Wei Y: Davis. J: Bina. WF. (2012). Ambient air pollution is associated with the increased incidence of breast
cancer in US. Int J Environ Health Res 22: 12-21. http://dx.doi.org/10.1080/09603123.2011.588321
Weichenthal S: Dufresne. A: Infante-Rivard. C: Joseph. L. (2007). Indoor ultrafme particle exposures and home
heating systems: A cross-sectional survey of Canadian homes during the winter months. J Expo Sci Environ
Epidemiol 17: 288-297. http://dx.doi.org/10.1038/si.ies.7500534
Weichenthal S: Kulka. R: Dubeau. A: Martin. C: Wang. D: Dales. R. (2011). Traffic-related air pollution and
acute changes in heart rate variability and respiratory function in urban cyclists. Environ Health Perspect
119: 1373-1378. http://dx.doi.org/10.1289/ehp.1003321
Weinmavr. G: Romeo. E: De Sario. M: Weiland. SK: Forastiere. F. (2010). Short-term effects of PM10 and NO2
on respiratory health among children with asthma or asthma-like symptoms: A systematic review and meta-
analysis [Review]. Environ Health Perspect 118: 449-457. http://dx.doi.org/10.1289/ehp.0900844
Weir. CH: Yeatts. KB: Sarnat. JA: Vizuete. W: Salo. PM: Jaramillo. R: Coha RD: Chu. H: Zeldia DC: Londoa
SJ. (2013). Nitrogen dioxide and allergic sensitization in the 2005-2006 National Health and Nutrition
Examination Survey. Respir Med 107: 1763-1772. http://dx.doi.0rg/10.1016/i.rmed.2013.08.010
Weitzberg. E: Hezel M: Lundberg. JO. (2010). Nitrate-nitrite-nitric oxide pathway: Implications for
anesthesiology and intensive care [Review]. Anesthesiology 113: 1460-1475.
http://dx.doi.org/10.1097/ALN.Ob013e3181fcf3cc
Weitzberg. E: Lundberg. JO. (2013). Novel aspects of dietary nitrate and human health [Review]. Annu Rev
Nutr33: 129-159. http://dx.doi.org/10.1146/annurev-nutr-071812-161159
Wellenius. GA: Burger. MR: Coull BA: Schwartz. J: Suh. HH: Koutrakis. P: Schlaug. G: Gold. PR: Mittleman.
MA. (2012). Ambient air pollution and the risk of acute ischemic stroke. Arch Intern Med 172: 229-234.
http://dx.doi.org/10.1001/archinternmed.2011.732
Wellenius. GA: Schwartz. J: Mittleman. MA. (2005). Air pollution and hospital admissions for ischemic and
hemorrhagic stroke among medicare beneficiaries. Stroke 36: 2549-2553.
http://dx.doi.org/10.1161/01.STR.0000189687.78760.47
R-103
-------
Wellenius. GA: Yeh. GY: Coull BA: Suh. HH: Phillips. RS: Mittleman. MA. (2007). Effects of ambient air
pollution on functional status in patients with chronic congestive heart failure: A repeated-measures study.
Environ Health 6. http://dx.doi.org/10.1186/1476-069X-6-26
Wendt JK: Symanski. E: Stock. TH: Chan. W: Du. XL. (2014). Association of short-term increases in ambient
air pollution and timing of initial asthma diagnosis among Medicaid-enrolled children in a metropolitan area.
Environ Res 131: 50-58. http://dx.doi.0rg/10.1016/i.envres.2014.02.013
Weng. HH: Tsai. SS: Chen. CC: Chiu. HF: Wu. TN: Yang. CY. (2008). Childhood leukemia development and
correlation with traffic air pollution in Taiwan using nitrogen dioxide as an air pollutant marker. J Toxicol
Environ Health A 71: 434-438. http://dx.doi.org/10.1080/15287390701839042
Wennmalm. A: Benthin. G: Edlund. A: Jungersten. L: Kieler-Jensen. N: Lundin. S: Westfelt UN: Petersson. AS:
Waagstein. F. (1993). Metabolism and excretion of nitric oxide in humans. An experimental and clinical
study. Circ Res 73: 1121-1127. http://dx.doi.Org/10.1161/01.RES.73.6.1121
Wheeler. AJ: Smith-Doiron. M: Xu. X: Gilbert. NL: Brook. JR. (2008). Intra-urban variability of air pollution in
Windsor, Ontario - Measurement and modeling for human exposure assessment. Environ Res 106: 7-16.
http://dx.doi.0rg/10.1016/i.envres.2007.09.004
WHO (World Health Organization). (1948). Preamble to the Constitution of the World Health Organization as
adopted by the International Health Conference, New York, 19-22 June, 1946. In Constitution of the World
Health Organization (pp. 2). Geneva, Switzerland.
http://whqlibdoc.who.int/hist/official records/constitution.pdf
Wichmann. J: Folke. F: Torp-Pedersen. C: Lippert. F: Ketzel M: Ellermann. T: Loft. S. (2013). Out-of-hospital
cardiac arrests and outdoor air pollution exposure in Copenhagen, Denmark. PLoS ONE 8.
http://dx.doi.org/10.1371/iournal.pone.0053684
Wichmann. J: Ketzel. M: Ellermann. T: Loft S. (2012). Apparent temperature and acute myocardial infarction
hospital admissions in Copenhagen, Denmark: A case-crossover study. Environ Health 11.
http://dx.doi.org/10.1186/1476-069X-ll-19
Wickman. M: Kull. I: Pershagen. G: Nordvall SL. (2002). The BAMSE project: Presentation of a prospective
longitudinal birth cohort study. PAI13: 11-13. http://dx.doi.0rg/10.1034/i.1399-3038.13.s.15.10.x
Widdicombe. JH. (2002). Regulation of the depth and composition of airway surface liquid [Review]. J Anat
201: 313-318. http://dx.doi.org/10.1046/i. 1469-7580.2002.00098.X
Widdicombe. JH: Widdicombe. JG. (1995). Regulation of human airway surface liquid [Review]. Respir Physiol
99: 3-12. http://dx.doi.org/10.1016/0034-5687(94)00095-H
Wijga. AH: Kerkhof. M: Gehring. U: de Jongste. JC: Postma. PS: Aalberse. R: Wolse. A: Koppelman. GH: van
Rossem. L: Oldenwening. M: Brunekreef. B: Smit HA. (2014). Cohort profile: The Prevention and
Incidence of Asthma and Mite Allergy (PIAMA) birth cohort. Int J Epidemiol 43: 527-535.
http://dx.doi.org/10.1093/iie/dvs231
Wilhelm. M: Ghosh. JK: Su. J: Cockburn. M: Jerrett M: Ritz. B. (2012). Traffic-related air toxics and term low
birth weight in Los Angeles County, California. Environ Health Perspect 120: 132-138.
http://dx.doi.org/10.1289/ehp. 1103408
Wilhelm. M: Ritz. B. (2005). Local variations in CO and paniculate air pollution and adverse birth outcomes in
Los Angeles County, California, USA. Environ Health Perspect 113: 1212-1221.
http://dx.doi.org/10.1289/ehp.7751
Williams. EJ: Lerner. BM: Murphy. PC: Herndoa SC: Zahniser. MS. (2009). Emissions of NOx, SO2, CO, and
HCHO from commercial marine shipping during Texas Air Quality Study (TexAQS) 2006. J Geophys Res
114. http://dx.doi.org/10.1029/2009JD012094
Williams. R: Brook. R: Bard. R: Conner. T: Shin. H: Burnett. R. (2012a). Impact of personal and ambient-level
exposures to nitrogen dioxide and paniculate matter on cardiovascular function. Int J Environ Health Res 22:
71-91. http://dx.doi.org/10.1080/09603123.2011.588437
R-104
-------
Williams. R: Jones. P: Croghan. C: Thornburg. J: Rodes. C. (2012b). The influence of human and environmental
exposure factors on personal NO(2) exposures. J Expo Sci Environ Epidemiol 22: 109-115.
http://dx.doi.org/10.1038/ies.2011.20
Williams. R: Rappold. A: Case. M: Schmitt M: Stone. S: Jones. P: Thornburg. J: Devlin. RB. (2012c). Multi-
pollutant exposures in an asthmatic cohort. Atmos Environ 61: 244-252.
http://dx.doi.0rg/10.1016/i.atmosenv.2012.07.049
Wilson. JG: Kingham. S: Pearce. J: Sturman. AP. (2005). A review of intraurban variations in paniculate air
pollution: Implications for epidemiological research [Review]. Atmos Environ 39: 6444-6462.
http://dx.doi.0rg/10.1016/i.atmosenv.2005.07.030
Wilson. WE: Brauer. M. (2006). Estimation of ambient and non-ambient components of paniculate matter
exposure from a personal monitoring panel study. J Expo Sci Environ Epidemiol 16: 264-274.
http://dx.doi.org/10.1038/si.ies.7500483
Wilson. WE: Mage. DT: Grant. LD. (2000). Estimating separately personal exposure to ambient and nonambient
paniculate matter for epidemiology and risk assessment: Why and how. J Air Waste Manag Assoc 50: 1167-
1183. http://dx.doi.org/10.1080/10473289.2000.10464164
Wilton. D: Szpiro. A: Gould. T: Larson. T. (2010). Improving spatial concentration estimates for nitrogen oxides
using a hybrid meteorological dispersion/land use regression model in Los Angeles, CA and Seattle, WA. Sci
Total Environ 408: 1120-1130. http://dx.doi.0rg/10.1016/i.scitotenv.2009.ll.033
Winer. AM: Peters. JW: Smith. JP: Pitts. JN. (1974). Response of commercial chemiluminescent nitric oxide-
nitrogen dioxide analyzers to other nitrogen-containing compounds. Environ Sci Technol 8:1118-1121.
http://dx.doi.org/10.1021/es60098a004
Winquist A: Kirrane. E: Klein. M: Strickland. M: Darrow. LA: Sarnat SE: Gass. K: Mulholland. J: Russell. A:
Tolbert. P. (2014). Joint effects of ambient air pollutants on pediatric asthma emergency department visits in
Atlanta, 1998-2004. Epidemiology 25: 666-673. http://dx.doi.org/10.1097/EDE.0000000000000146
Witten. A: Solomon. C: Abbritti. E: Arjomandi M: Zhai W: Kleinman. M: Balmes. J. (2005). Effects of
nitrogen dioxide on allergic airway responses in subjects with asthma. J Occup Environ Med 47: 1250-1259.
http://dx.doi.org/10.1097/01.iom.0000177081.62204.8d
Wittkopp. S: Staimer. N: Tjoa. T: Gillen. D: Daher. N: Shafer. M: Schauer. JJ: Sioutas. C: Delfmo. RJ. (2013).
Mitochondrial genetic background modifies the relationship between traffic-related air pollution exposure
and systemic biomarkers of inflammation. PLoS ONE 8: e64444.
http://dx.doi.org/10.1371/iournal.pone.0064444
Wiwatanadate. P: Liwsrisakun. C. (2011). Acute effects of air pollution on peak expiratory flow rates and
symptoms among asthmatic patients in Chiang Mai, Thailand. Int J Hyg Environ Health 214: 251-257.
http://dx.doi.0rg/10.1016/i.ijheh.2011.03.003
Wiwatanadate. P: Trakultivakorn. M. (2010). Air pollution-related peak expiratory flow rates among asthmatic
children in Chiang Mai, Thailand. Inhal Toxicol 22: 301-308. http://dx.doi.org/10.3109/08958370903300327
Wollmann. HA. (1998). Intrauterine growth restriction: Definition and etiology. Horm Res 49: 1-6.
http://dx.doi.org/10.1159/000053079
Wong. CM: Vichit-Vadakan. N: Kan. H: Qiaa Z. (2008). Public Health and Air Pollution in Asia (PAPA): A
multicity study of short-term effects of air pollution on mortality. Environ Health Perspect 116: 1195-1202.
http://dx.doi.org/10.1289/ehp. 11257
Wong. CM: Vichit-Vadakan. N: Vajanapoom. N: Ostro. B: Thach. TQ: Chau. PY: Chan. EK: Chung. RY: Ou.
CO: Yang. L: Peiris. JS: Thomas. GN: Lam. TH: Wong. TW: Hedlev. AJ: Kan. H: Chen. B: Zhao. N:
London. SJ: Song. G: Chen. G: Zhang. Y: Jiang. L: Qiaa Z: He. Q: Lin. HM: Kong. L: Zhou. D: Liang. S:
Zhu. Z: Liao. D: Liu. W: Bentlev. CM: Dan. J: Wang. B: Yang. N: Xu. S: Gong. J: Wei. H: Sua H: Oin. Z.
(2010). Part 5. Public health and air pollution in Asia (PAPA): A combined analysis of four studies of air
pollution and mortality. In Public Health and Air Pollution in Asia (PAPA): Coordinated studies of short-
term exposure to air pollution and daily mortality in four cities (pp. 377-418). Boston, MA: Health Effects
Institute. http://pubs.healtheffects.org/getfile.php?u=595
R-105
-------
Wong. CM: Yang. L: Thach. TO: Chau. PY: Chan. KP: Thomas. GN: Lam. TH: Wong. TW: Hedlev. AJ: Peiris.
JS. (2009). Modification by influenza on health effects of air pollution in Hong Kong. Environ Health
Perspect 117: 248-253. http://dx.doi.org/10.1289/ehp. 11605
Wong. TW: Lau. TS: Yu. TS: Neller. A: Wong. SL: Tarn. W: Pang. SW. (1999). Air pollution and hospital
admissions for respiratory and cardiovascular diseases in Hong Kong. Occup Environ Med 56: 679-683.
http://dx.doi.org/10.1136/oem.56.10.679
Wood. AM: Harrison. RM: Semple. S: Ayres. JG: Stocklev. RA. (2009). Outdoor air pollution is associated with
disease severity in al-antitrypsin deficiency. Eur Respir J 34: 346-353.
http://dx.doi.org/10.1183/09031936.00087908
Woodruff. TJ: Parker. JD: Adams. K: Bell ML: Gehring. U: Glinianaia. S: Ha. EH: Jalaludin. B: Slama. R.
(2010). International collaboration on air pollution and pregnancy outcomes (ICAPPO). Int J Environ Res
Public Health 7: 2638-2652. http://dx.doi.org/10.3390/ijerph7062638
Woodruff. TJ: Parker. JD: Darrow. LA: Slama. R: Bell ML: Choi. H: Glinianaia. S: Hoggatt KJ: Karr. CJ:
Lobdell DT: Wilhelm. M. (2009). Methodological issues in studies of air pollution and reproductive health.
Environ Res 109: 311-320. http://dx.doi.0rg/10.1016/i.envres.2008.12.012
Wooldridge. PJ: Perring. AE: Bertram. TH: Flocke. FM: Roberts. JM: Singh. HB: Huev. LG: Thornton. JA:
Wolfe. GM: Murphy. JG: Fry. JL: Rollins. AW: Lafranchi. BW: Cohen. RC. (2010). Total Peroxy Nitrates
(Sigma PNs) in the atmosphere: The Thermal Dissociation-Laser Induced Fluorescence (TD-LIF) technique
and comparisons to speciated PAN measurements. Atmos Meas Tech 3: 593-607.
http://dx.doi.org/10.5194/amt-3-593-2010
Wu. J: Lurmann. F: Winer. A: Lu. R: Turco. R: Funk. T. (2005). Development of an individual exposure model
for application to the Southern California children's health study. Atmos Environ 39: 259-273.
http://dx.doi.0rg/10.1016/i.atmosenv.2004.09.061
Wu. J: Ren. C: Delfino. RJ: Chung. J: Wilhelm. M: Ritz. B. (2009). Association between local traffic-generated
air pollution and preeclampsia and preterm delivery in the south coast air basin of California. Environ Health
Perspect 117: 17731779. http://dx.doi.org/10.1289/ehp.0800334
Wu. J: Wilhelm. M: Chung. J: Ritz. B. (201 la). Comparing exposure assessment methods for traffic-related air
pollution in an adverse pregnancy outcome study. Environ Res 111: 685-692.
http://dx.doi.0rg/10.1016/i.envres.2011.03.008
Wu. X: Bennett. DH: Lee. K: Cassadv. PL: Ritz. B: Hertz-Picciotto. I. (201 Ib). Longitudinal variability of time-
location/activity patterns of population at different ages: A longitudinal study in California. Environ Health
10. http://dx.doi.org/10.1186/1476-069X-10-80
Wu. XM: Fan. ZT: Ohman-Strickland. P. (2010). Time-location patterns of a population living in an air pollution
hotspot. J Environ Public Health 2010: 625461. http://dx.doi.org/10.1155/2010/625461
Xiang. H: Mertz. KJ: Arena. VC: Brink. LL: Xu. X: Bi. Y: Talbott EO. (2013). Estimation of short-term effects
of air pollution on stroke hospital admissions in Wuhan, China. PLoS ONE 8: e61168.
http://dx.doi.org/10.1371/iournal.pone.0061168
Xu. X: Hu. H: Ha. S: Roth. J. (2014). Ambient air pollution and hypertensive disorder of pregnancy. J Epidemiol
Community Health 68: 13-20. http://dx.doi.org/10.1136/iech-2013-202902
Xu. Z: Hu. W: Williams. G: Clements. AC: Kan. H: Tong. S. (2013). Air pollution, temperature and pediatric
influenza in Brisbane, Australia. Environ Int 59: 384-388. http://dx.doi.0rg/10.1016/i.envint.2013.06.022
Xuan. W: Peat JK: Toelle. BG: Marks. GB: Berry. G: Woolcock. AJ. (2000). Lung function growth and its
relation to airway hyperresponsiveness and recent wheeze: Results from a longitudinal population study. Am
J Respir Crit Care Med 161: 1820-1824. http://dx.doi.0rg/10.1164/airccm.161.6.9809118
Xue. J: McCurdv. T: Spengler. J: Ozkaynak. H. (2004). Understanding variability in time spent in selected
locations for 7-12-year old children. J Expo Anal Environ Epidemiol 14: 222-233.
http://dx.doi.org/10.1038/si.iea.7500319
R-106
-------
Yadav. AK: Doran. SF: Samal AA: Sharma. R: Vedagiri. K: Postlethwait EM: Squadrito. GL: Fanucchi. MV:
Roberts. LJ: Patel RP: Matalon. S. (2011). Mitigation of chlorine gas lung injury in rats by postexposure
administration of sodium nitrite. Am J Physiol Lung Cell Mol Physiol 300: L362-L369.
http://dx.doi.org/10.1152/aiplung.00278.2010
Yaghi. A: Zaman. A: Cox. G: Dolovich. MB. (2012). Ciliary beating is depressed in nasal cilia from chronic
obstructive pulmonary disease subjects. RespirMed 106: 1139-1147.
http://dx.doi.0rg/10.1016/i.rmed.2012.04.001
Yamamoto. I: Takahashj M. (1984). Ultrastructural observations of rat lung exposed to nitrogen dioxide for 7
months. Kitasato ArchExperMed 57: 57-65.
Yamazakj S: Shima. M: Ando. M: Nitta. H: Watanabe. H: Nishimuta. T. (2011). Effect of hourly concentration
of paniculate matter on peak expiratory flow in hospitalized children: A panel study. Environ Health 10.
http://dx.doi.org/10.1186/1476-069X-10-15
Yang. CL: To. T: Fotv. RG: Stieb. DM: Dell SD. (2011). Verifying a questionnaire diagnosis of asthma in
children using health claims data. BMC Pulm Med 11. http://dx.doi.org/10.1186/1471-2466-ll-52
Yang. CY. (2008). Air pollution and hospital admissions for congestive heart failure in a subtropical city: Taipei,
Taiwan. J Toxicol Environ Health A 71: 1085-1090. http://dx.doi.org/10.1080/15287390802114428
Yang. CY: Chen. YS: Yang. CH: Ho. SC. (2004). Relationship between ambient air pollution and hospital
admissions for cardiovascular diseases in Kaohsiung, Taiwan. J Toxicol Environ Health A 67: 483-493.
http://dx.doi.org/10.1080/15287390490276502
Yang. CY: Hsieh. HJ: Tsai SS: Wu. TN: Chiu. HF. (2006). Correlation between air pollution and postneonatal
mortality in a subtropical city: Taipei, Taiwan. J Toxicol Environ Health A 69: 2033-2040.
http://dx.doi.org/10.1080/15287390600746181
Yang. Q: Chen. Y: Krewskj D: Burnett RT: Shi. Y: McGrail KM. (2005). Effect of short-term exposure to low
levels of gaseous pollutants on chronic obstructive pulmonary disease hospitalizations. Environ Res 99: 99-
105. http://dx.doi.0rg/10.1016/i.envres.2004.09.014
Yang. Q: Chen. Y: Shi. Y: Burnett. RT: McGrail. KM: Krewski. D. (2003). Association between ozone and
respiratory admissions among children and the elderly in Vancouver, Canada. Inhal Toxicol 15: 1297-1308.
http://dx.doi.org/10.1080/08958370390241768
Yokovama. E. (1968). Uptake of SO2 and NO2 by the isolated upper airways. Koshu Eisei Kenkyu 17: 302-306.
Yorifuji. T: Kashima. S: Tsuda. T: Ishikawa-Takata. K: Ohta. T: Tsuruta. K: Doi H. (2013). Long-term
exposure to traffic-related air pollution and the risk of death from hemorrhagic stroke and lung cancer in
Shizuoka, Japan. Sci Total Environ 443: 397-402. http://dx.doi.0rg/10.1016/j.scitotenv.2012.10.088
Yorifuji. T: Kashima. S: Tsuda. T: Takao. S: Suzuki. E: Doi. H: Sugiyama. M: Ishikawa-Takata. K: Ohta. T.
(2010). Long-term exposure to traffic-related air pollution and mortality in Shizuoka, Japan. Occup Environ
Med 67: 111-117. http://dx.doi.org/10.1136/oem.2008.045542
Yuval: Bekhor. S: Brodav. DM. (2013). Data-driven nonlinear optimisation of a simple air pollution dispersion
model generating high resolution spatiotemporal exposure. Atmos Environ 79: 261-270.
http://dx.doi.0rg/10.1016/i.atmosenv.2013.06.005
Zanobetti A: Gold. PR: Stone. PH: Suh. HH: Schwartz. J: Coull BA: Speizer. FE. (2010). Reduction in heart
rate variability with traffic and air pollution in patients with coronary artery disease. Environ Health Perspect
118: 324-330. http://dx.doi.org/10.1289/ehp.0901003
Zanobetti. A: Schwartz. J. (2006). Air pollution and emergency admissions in Boston, MA. J Epidemiol
Community Health 60: 890-895. http://dx.doi.org/10.1136/iech.2005.039834
Zeger. SL: Thomas. D: Dominici F: Samet JM: Schwartz. J: Dockery. D: Cohen. A. (2000). Exposure
measurement error in time-series studies of air pollution: Concepts and consequences. Environ Health
Perspect 108: 419-426.
R-107
-------
Zeghnoun. A: Beaudeau. P: Carrat F: Delmas. V: Boudhabhav. O: Gavon. F: Guincetre. D: Czernichow. P.
(1999). Air pollution and respiratory drug sales in the city of Le Havre, France, 1993-1996. EnvironRes 81:
224-230. http://dx.doi.org/10.1006/enrs.1999.3974
Zeman. KL: Bennett. WD. (2006). Growth of the small airways and alveoli from childhood to the adult lung
measured by aerosol-derived airway morphometry. J Appl Physiol 100: 965-971.
http://dx.doi.org/10.1152/iapplphysiol.00409.2005
Zemek. R: Szyszkowicz. M: Rowe. BH. (2010). Air pollution and emergency department visits for otitis media:
A case-crossover study in Edmonton, Canada. Environ Health Perspect 118: 1631-1636.
http://dx.doi.org/10.1289/ehp.0901675
Zemp. E: Elsasser. S: Schindler. C: Kiinzlj N: Perruchoud. AP: Domenighetti. G: Medici. T: Ackermann-
Liebrich. U: Leuenberger. P: Monn. C: Bolognini. G: Bongard. JP: Brandli. O: Karrer. W: Keller. R: Schoni.
MH: Tschopp. JM: Villiger. B: Zellweger. JP. (1999). Long-term ambient air pollution and respiratory
symptoms in adults (SAPALDIA study). Am J Respir Crit Care Med 159: 1257-1266.
http://dx.doi.0rg/10.1164/airccm.159.4.9807052
Zetterquist W: Pedroletti C: Lundberg. JO: Alving. K. (1999). Salivary contribution to exhaled nitric oxide. Eur
Respir J 13: 327-333. http://dx.doi.0rg/10.1034/j.1399-3003.1999.13bl8.x
Zhang. J: Villacorta. L: Chang. L: Faa Z: Hamblia M: Zhu. T: Chen. CS: Cole. MP: Schopfer. FJ: Deng. CX:
Garcia-Barrio. MT: Feng. YH: Freeman. BA: Chen. YE. (2010a). Nitro-oleic acid inhibits angiotensin II-
induced hypertension. Circ Res 107: 540-548. http://dx.doi.org/10.1161/CIRCRESAHA. 110.218404
Zhang. J: Zhu. T: Kipen. H: Wang. G: Huang. W: Rich. D: Zhu. P: Wang. Y: Lu. SE: Ohman-Strickland. P:
Diehl. S: Hu. M: Tong. J: Gong. J: Thomas. D. (2013). Cardiorespiratory biomarker responses in healthy
young adults to drastic air quality changes surrounding the 2008 Beijing Olympics. Res Rep Health Eff Inst
174: 5-174.
Zhang. PF: Dong. GH: Sun. BJ: Zhang. LW: Chen. X: Ma. NN: Yu. F: Quo. HM: Huang. H: Lee. YL: Tang. NJ:
Chen. J. (2011). Long-term exposure to ambient air pollution and mortality due to cardiovascular disease and
cerebrovascular disease in Shenyang, China. PLoS ONE 6. http://dx.doi.org/10.1371/iournal.pone.0020827
Zhang. Y: Ma. C: Xiao. Y: Liu. C: Xu. Q: Wang. Y: Xu. X: Hao. L. (2010b). Dual role of vitamin C utilization
in NO2-induced oxidative stress in lung tissues of mice. Bull Environ Contam Toxicol 84: 662-666.
http://dx.doi.org/10.1007/s00128-010-0021-l
Zhang. Z: Wang. J: Chen. L: Chen. X: Sun. G: Zhong. N: Kan. H: Lu. W. (2014). Impact of haze and air
pollution-related hazards on hospital admissions in Guangzhou, China. Environ Sci Pollut Res Int 21: 4236-
4244. http://dx.doi.org/10.1007/sll356-013-2374-6
Zheng. S: Wang. M: Wang. S: Tao. Y: Shang. K. (2013). Short-term effects of gaseous pollutants and paniculate
matter on daily hospital admissions for cardio-cerebrovascular disease in Lanzhou: Evidence from a heavily
polluted city in China. Int J Environ Res Public Health 10: 462-477.
http://dx.doi.org/10.3390/ijerphl0020462
Zhou. N: Cui. Z: Yang. S: Han. X: Chea G: Zhou. Z: Zhai. C: Ma. M: Li. L: Cai. M: Li. Y: Ao. L: Shu. W: Liu.
J: Cao. J. (2014). Air pollution and decreased semen quality: A comparative study of Chongqing urban and
rural areas. Environ Pollut 187: 145-152. http://dx.doi.0rg/10.1016/i.envpol.2013.12.030
Zhou. Y: Lew. JI. (2007). Factors influencing the spatial extent of mobile source air pollution impacts: A meta-
analysis [Review]. BMC Public Health 7: 89. http://dx.doi.org/10.1186/1471-2458-7-89
Zhu. DZ: Nussbaum. NJ: Kuhns. HP: Chang. MCO: Sodeman. D: Moosmuller. H: Watson. JG. (2011). Real-
world PM, NO(x), CO, and ultrafme particle emission factors for military non-road heavy duty diesel
vehicles. Atmos Environ45: 2603-2609. http://dx.doi.0rg/10.1016/i.atmosenv.2011.02.032
Zhu. T. (2013). Email from Dr. Zhu to Dr. Patel; Response to data request. Available online
Zhu. Y: Fung. DC: Kennedy. N: Hinds. WC: Eiguren-Fernandez. A. (2008). Measurements of ultrafine particles
and other vehicular pollutants inside a mobile exposure system on Los Angeles freeways. J Air Waste Manag
Assoc 58: 424-434.
R-108
-------
Zidek. JV: Shaddick. G: Meloche. J: Chatfield. C: White. R. (2007). A framework for predicting personal
exposures to environmental hazards. Environ Ecol Stat 14: 411-431. http://dx.doi.org/10.1007/sl0651-007-
0028-x
Zora. JE: Sarnat SE: Ravsonj AU: Johnson. BA: Li. WW: Greenwald. R: Holguin. F: Stock. TH: Sarnat JA.
(2013). Associations between urban air pollution and pediatric asthma control in El Paso, Texas. Sci Total
Environ448: 56-65. http://dx.doi.0rg/10.1016/j.scitotenv.2012.ll.067
Zou. X: Shen. Z: Yuan. T: Yia S: Cai J: Chen. L: Wang. W. (2006). Shifted power-law relationship between
NO2 concentration and the distance from a highway: A new dispersion model based on the wind profile
model. Atmos Environ40: 8068-8073. http://dx.doi.0rg/10.1016/i.atmosenv.2006.09.012
R-109
------- |