4% rriA United States
Environmental Protection
Agency
May 2008
EPA/600/R-08/047
Integrated Science
Assessment for
Sulfur Oxides -
Health Criteria
(Second External Review Draft)
-------
May 2008
EPA/600/R-08/047
Integrated Science Assessment
for Sulfur Oxides - Health Criteria
National Center for Environmental Assessment-RTP Division
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
Disclaimer
This document is a second external review draft being released for review purposes
only and does not constitute U.S. Environmental Protection Agency policy. Mention of
trade names or commercial products does not constitute endorsement or recommendation
for use.
-------
Table of Contents
List of Tables vii
List of Figures ix
Abbreviations and Acronyms xiii
Authors, Contributors, Reviewers xix
SOx Project Team xxiii
Clean Air Scientific Advisory Committee for NOx and SOx Primary NAAQS xxv
Preface xxvii
Chapter 1. Introduction 1-1
1.1. Document Development 1-2
1.2. Document Organization 1-3
1.3. EPA Framework for Causal Determinations 1-4
1.3.1. Scientific Evidence Used in Establishing Causality 1-5
1.3.2. Association and Causation 1-5
1.3.3. Evidence for Going beyond Association to Causation 1-6
1.3.4. Multifactorial Causation 1-9
1.3.5. Uncertainty 1-9
1.3.6. Application of Framework 1-10
1.3.7. First Step—Determination of Causality 1-13
1.3.8. Second Step—Evaluation of Population Response 1-14
1.4. Conclusions 1-15
Chapter 2. Source to Tissue Dose 2-1
2.1. Sources of Sulfur Oxides 2-1
2.2. Atmospheric Chemistry 2-3
2.3. Measurement Methods and Associated Issues 2-5
2.3.1. Sources of Positive Interference 2-6
2.3.2. Sources of Negative Interference 2-7
2.3.3. Other Techniques for Measuring S02 2-8
2.4. Environmental Concentrations of SOx 2-8
2.4.1. Design Criteria for the NAAQS S02 Monitoring Networks 2-8
2.4.2. Monitor Locations in Selected Areas of the U.S. 2-11
2.4.3. Ambient S02 Concentrations in Relation to S02 Sources 2-14
2.4.4. Spatial and Temporal Variability of Ambient S02 Concentrations 2-21
2.4.5. 5-Minute Sample Data in the Monitoring Network 2-26
2.4.6. Policy Relevant Background Contributions to S02 Concentrations 2-33
2.5. Issues Associated with Evaluating S02 Exposure 2-38
2.5.1. General Considerations for Personal Exposure 2-38
2.5.2. Methods Used for Monitoring Personal Exposure 2-42
2.5.3. Relationship between Personal Exposure and Ambient Concentration 2-43
2.5.3.1. Indoor Versus Outdoor S02 Concentrations 2-43
2.5.3.2. Relationship of Personal Exposure to Ambient Concentration 2-46
2.5.4. Exposure Measurement Errors in Epidemiological Studies 2-52
2.5.4.1. Community Time-Series Studies 2-53
2.5.4.1.1. Relationship of Measured S02 to the True Concentration 2-53
2.5.4.1.2. Relationship of Day-to-day Variations in the Ambient Concentration of
S02 to Variations in the Community Average 2-54
May 2008
DRAFT—DO NOT QUOTE OR CITE
-------
2.5.4.1.3. Relationship of Community Average Concentration of S02 to Average
Personal Exposure to Ambient S02 2-55
2.5.4.2. Short-Term Panel Studies 2-56
2.5.4.3. Long-Term Cohort Studies 2-56
2.5.4.4. Summary of Evaluation of Exposure Measurement Error in Epidemiological
Studies 2-57
2.6. Dosimetry of Inhaled Sulfur Oxides 2-57
2.6.1. Gas Deposition 2-58
2.6.2. Particles and Sulfur Oxide Mixtures 2-62
2.6.3. Distribution and Elimination of Sulfur Oxides 2-62
Chapter 3. Integrated Health Effects 3-1
3.1. Respiratory Morbidity Associated with Short-Term Exposure 3-4
3.1.1. Summary of Findings from the Previous Review 3-4
3.1.2. Potential Mode of Action for Respiratory Health Effects 3-6
3.1.3. Respiratory Effects Associated with Peak Exposure 3-8
3.1.3.1. Respiratory Symptoms 3-9
3.1.3.2. Lung Function 3-10
3.1.3.3. Airway Inflammation 3-13
3.1.3.4. Evidence of the Effect of Peak Exposure from Animal Studies 3-14
3.1.3.5. Summary of Evidence on the Effect of Peak Exposure on Respiratory Health 3-15
3.1.4. Respiratory Effects Associated with Short-Term (> 1 h) Exposure 3-17
3.1.4.1. Respiratory Symptoms 3-17
3.1.4.1.1. Children 3-18
3.1.4.1.2. Adults 3-24
3.1.4.2. Lung Function 3-25
3.1.4.2.1. Children 3-26
3.1.4.2.2. Adults 3-27
3.1.4.3. Airway Inflammation 3-29
3.1.4.4. Airway Hyperresponsiveness and Allergy 3-30
3.1.4.5. Respiratory Illness-Related Absences 3-33
3.1.4.6. Emergency Department Visits and Hospitalizations for Respiratory Diseases 3-34
3.1.4.6.1. All Respiratory Diseases 3-35
3.1.4.6.2. Asthma 3-39
3.1.4.6.3. Chronic Obstructive Pulmonary Disease 3-42
3.1.4.6.4. Respiratory Diseases Other than Asthma or COPD 3-42
3.1.4.6.5. Summary of Evidence on Emergency Department Visits and
Hospitalizations for Respiratory Diseases 3-43
3.1.4.7. Summary of Evidence on the Effect of Short-Term (> 1 h) Exposure on
Respiratory Health 3-45
3.1.5. Mixtures and Interactive Effects 3-47
3.1.5.1. Evidence from Human Clinical Studies 3-47
3.1.5.2. Evidence from Animal Toxicological Studies 3-48
3.1.5.2.1. Effects of S02 Layered on Metallic Particles 3-49
3.1.5.2.2. Effects of S02 Layered on Carbon Particles 3-51
3.1.5.2.3. Effects of Sulfite Aerosols 3-51
3.1.5.2.4. Other Mixtures 3-52
3.1.5.2.5. Summary of Evidence on S02 Interactions with PM and Other Mixtures 3-52
3.1.6. Evidence of the Effect of S02 on Respiratory Morbidity from Intervention Studies 3-52
3.1.7. Summary of Evidence of the Effect of Short-Term S02 Exposure on Respiratory
Health 3-55
3.2. Other Morbidity Associated with Short-Term S02 Exposure 3-56
3.2.1. Summary of Findings from the Previous Review 3-56
3.2.2. Cardiovascular Effects Associated with Short-Term Exposure 3-56
3.2.2.1. Heart Rate and Heart Rate Variability 3-57
3.2.2.2. Repolarization Changes 3-60
May 2008
iv
DRAFT—DO NOT QUOTE OR CITE
-------
3.2.2.3. Cardiac Arrhythmias 3-60
3.2.2.4. Blood Pressure 3-62
3.2.2.5. Blood Markers of Cardiovascular Risk 3-63
3.2.2.6. Acute Myocardial Infarction 3-64
3.2.2.7. Emergency Department Visits and Hospitalizations for Cardiovascular
Diseases 3-65
3.2.2.7.1. All Cardiovascular Diseases 3-65
3.2.2.7.2. Specific Cardiovascular Diseases 3-67
3.2.2.7.3. Summary of Evidence on Emergency Department Visits and
Hospitalizations from Cardiovascular Diseases 3-68
3.2.2.8. Summary of Evidence on the Effect of Short-Term S02 Exposure on
Cardiovascular Health 3-68
3.2.3. Other Effects Associated with Short-Term S02 Exposure 3-69
3.3. Mortality Associated with Short-Term S02 Exposure 3-70
3.3.1. Summary of Findings from the Previous Review 3-70
3.3.2. Associations of Mortality and Short-Term S02 Exposure in Multicity Studies and
Meta-Analyses 3-71
3.3.2.1.1. Multicity Studies 3-71
3.3.2.1.2. National Morbidity, Mortality, and Air Pollution Study 3-71
3.3.2.1.3. Canadian Multicity Studies 3-72
3.3.2.1.4. Air Pollution and Health: A European Approach 3-73
3.3.2.1.5. The Netherlands Study 3-75
3.3.2.1.6. Other European Multicity Studies 3-76
3.3.2.2. Meta-Analyses of Air Pollution-Related Mortality Studies 3-77
3.3.2.2.1. Meta-Analysis of All Criteria Pollutants 3-77
3.3.2.2.2. Health Effects Institute Review of Air Pollution Studies in Asia 3-77
3.3.3. Evidence of the Effect of S02 on Mortality from an Intervention Study 3-78
3.3.4. Summary of Evidence on the Effect of Short-Term S02 Exposure on Mortality 3-80
3.4. Morbidity Associated with Long-Term S02 Exposure 3-84
3.4.1. Summary of Findings from the Previous Review 3-84
3.4.2. Respiratory Effects Associated with Long-Term Exposure to S02 3-86
3.4.2.1. Asthma, Bronchitis, and Respiratory Symptoms 3-86
3.4.2.2. Lung Function 3-89
3.4.2.3. Morphological Effects 3-91
3.4.2.4. Lung Host Defense 3-91
3.4.2.5. S02 Interactions with PM and Other Mixtures 3-92
3.4.2.6. Summary of Evidence on the Effect of Long-Term Exposure on Respiratory
Health 3-93
3.4.3. Carcinogenic Effects Associated with Long-Term Exposure 3-94
3.4.4. Cardiovascular Effects Associated with Long-Term Exposure 3-98
3.4.5. Prenatal and Neonatal Outcomes Associated with Long-Term Exposure 3-99
3.4.6. Other Organ System Effects Associated with Long-Term Exposure 3-104
3.5. Mortality Associated with Long-Term S02 Exposure 3-104
3.5.1. Summary of Findings from the Previous Review 3-104
3.5.2. Associations of Mortality and Long-Term Exposure in Key Studies 3-105
3.5.2.1. U.S. Cohort Studies 3-105
3.5.2.1.1. Harvard Six Cities Studies 3-105
3.5.2.1.2. American Cancer Society Cohort Studies 3-106
3.5.2.1.3. The EPRI-Washington University Veterans' Cohort Mortality Studies 3-108
3.5.2.1.4. Seventh-day Adventist Study 3-109
3.5.2.2. European Cohort Studies 3-110
3.5.2.3. Cross-Sectional Analysis Using Small Geographic Scale 3-111
3.5.3. Summary of Evidence on the Effect of Long-Term Exposure on Mortality 3-112
Chapter 4. Public Health Impact 4-1
4.1. Assessment of Concentration-Response Function and Potential Thresholds 4-1
May 2008
v
DRAFT—DO NOT QUOTE OR CITE
-------
4.1.1. Evidence from Human Clinical Studies 4-2
4.1.2. Evidence from Epidemiological Studies 4-4
4.1.3. Summary of Evidence on Concentration-Response Functions and Thresholds 4-8
4.2. Susceptible and Vulnerable Populations 4-9
4.2.1. Preexisting Disease as a Potential Risk Factor 4-10
4.2.1.1. Individuals with Respiratory Diseases 4-10
4.2.1.2. Individuals with Cardiovascular Diseases 4-12
4.2.2. Genetic Factors for Oxidant and Inflammatory Damage from Air Pollutants 4-13
4.2.3. Age-Related Susceptibility 4-15
4.2.4. Other Potentially Susceptible Populations 4-16
4.2.5. Factors that Potentially Increase Vulnerability to S02 4-18
4.3. Potential Public Health Impacts 4-20
4.3.1. Concepts Related to Defining Adverse Health Effects 4-20
4.3.2. Estimation of Potential Numbers of Persons in At-Risk Susceptible Population
Groups in the United States 4-22
Chapter 5. Summary and Conclusions 5-1
5.1. Emissions and Ambient Concentrations of S02 5-1
5.2. Health Effects of S02 5-2
5.3. Interpretation of the Epidemiological Evidence 5-9
5.4. Susceptible and Vulnerable Populations 5-11
5.5. Conclusions 5-12
May 2008
vi
DRAFT—DO NOT QUOTE OR CITE
-------
List of Tables
Table 1-1 Aspects to aid in judging causality. 1-11
Table 1-2. Weight of evidence for causal determination. 1-13
Table 2-1. Monitor counts for California and San Diego County, 2005. 2-11
Table 2-2. Monitor counts for Ohio and Cuyahoga County, 2005. 2-11
Table 2-3. Regional distribution of S02 and S042- ambient concentrations, averaged for
2003-05. 2-22
Table 2-4. Distributions of temporal averaging inside and outside CMSAs. 2-22
Table 2-5. Range of mean annual S02 concentrations and Pearson correlation coefficients
in urban areas having at least four regulatory monitors, 2003-2005. 2-24
Table 2-6. Locations, counts, and sampling periods of monitors reporting 5-minute
maximum S02 values, 1997-2006. 2-32
Table 2-7. Locations, counts, and sampling periods of monitors reporting all 12 5-minute
S02 values in each hour, 1997-2006. 2-32
Table 2-8. Relationships of indoor to outdoor S02 concentrations. 2-45
Table 2-9. Association between personal exposure and ambient concentration (longitudinal
correlations coefficients). 2-47
Table 2-10. Association between personal exposure and ambient concentration (pooled
correlations coefficients). 2-48
Table 3-1. Percentage of asthmatic individuals in controlled human exposures experiencing
S02-induced decrements in lung function. 3-16
Table 4-1. Gradation of individual responses to short-term S02 exposure in individuals with
impaired respiratory systems. 4-21
Table 4-2. Prevalence of selected respiratory disorders by age group in the United States
(2004 [U.S. adults] and 2005 [U.S. children] National Health Interview Survey). 4-23
Table 5-1. Key health effects of short-term exposure to S02 observed in human clinical
studies. 5-3
Table 5-2. Key respiratory health effects of exposure to S02 in animal toxicological studies. 5-5
Table 5-3. Key findings on the health effects of S02 exposure 5-13
Table 5-4. Effects of short-term exposure to S02 on respiratory symptoms among children. 5-16
Table 5-5. Effects of short-term S02 exposure on emergency department visits and hospital
admissions for respiratory outcomes. 5-18
May 2008
vii
DRAFT—DO NOT QUOTE OR CITE
-------
List of Figures
Figure 1-1. Exposure-disease-stress model for environmental health disparities. 1-6
Figure 2-1. Criteria pollutant monitor locations (A) and S02 monitor locations (B), California,
2005. Shaded counties have at least one monitor. 2-13
Figure 2-2. Criteria pollutant monitor locations (A) and S02 monitor locations (B), Ohio, 2005.
Shaded counties have at least one monitor. 2-13
Figure 2-3. Criteria pollutant monitor locations (A) and S02 monitor locations (B), Arizona,
2005. Shaded counties have at least one monitor. 2-15
Figure 2-4. Criteria pollutant monitor locations (A) and S02 monitor locations (B),
Pennsylvania, 2005. Shaded counties have at least one monitor. 2-16
Figure 2-5. Criteria pollutant monitor locations (A) and S02 monitor locations (B), New York,
2005. Shaded counties have at least one monitor. 2-17
Figure 2-6. Criteria pollutant monitor locations (A) and S02 monitor locations (B),
Massachusetts, 2005. Shaded counties have at least one monitor. 2-18
Figure 2-7. State-level S02 emissions, 1990-2005. 2-19
Figure 2-8. Annual mean ambient S02 concentration, 1989 through 1991 (a), and 2003
through 2005 (b). 2-20
Figure 2-9. Annual mean ambient S042- concentration, 1989 through 1991 (a), and 2003
through 2005 (b). 2-20
Figure 2-10. Annual S02 emissions for Acid Rain Program cooperating facilities, 2006. 2-21
Figure 2-11. Boxplot of hourly S02 concentrations across all cities in focus. 2-23
Figure 2-12. Steubenville, OH, 2003-2005. (a) Monthly mean, minimum, and maximum S02
concentrations, (b) Monthly mean, minimum, and maximum S042-
concentrations, (c) Monthly mean S042" concentrations as a function of S02
concentrations. 2-27
Figure 2-13. Philadelphia, 2003-2005. (a) Monthly mean, minimum, and maximum S02
concentrations, (b) Monthly mean, minimum, and maximum S042"
concentrations, (c) Monthly mean S042" concentrations as a function of S02
concentrations. 2-28
Figure 2-14. Los Angeles, 2003-2005. (a) Monthly mean, minimum, and maximum S02
concentrations, (b) Monthly mean, minimum, and maximum S042"
concentrations, (c) Monthly mean S042" concentrations as a function of S02
concentrations. 2-29
Figure 2-15. Riverside, CA, 2003-2005. (a) Monthly mean, minimum, and maximum S02
concentrations, (b) Monthly mean, minimum, and maximum S042"
concentrations, (c) Monthly mean S042" concentrations as a function of S02
concentrations. 2-30
Figure 2-16. Phoenix, 2003-2005. (a) Monthly mean, minimum, and maximum S02
concentrations, (b) Monthly mean, minimum, and maximum S042"
concentrations, (c) Monthly mean S042" concentrations as a function of S02
concentrations. 2-31
Figure 2-17. S02 monitors reporting maximum or continuous 5-minute average values for any
period, 1997-2006. 2-33
Figure 2-18. Annual mean model-predicted concentrations of S02 (ppb). 2-35
-------
Figure 2-19. 15-minute average ambient S02 concentrations measured at Hawaii Volcanoes
National Park monitoring sites, March 12, 13, and 15, 2007. 2-36
Figure 2-20. 15-minute average ambient S02 concentrations measured at the two National
Park monitoring sites at Hawaii Volcanoes NP, Hawaii on September 29, 2007. 2-37
Figure 2-21. Percentage of time spent in various environments in the United States. 2-40
Figure 2-22. Average annual indoor and outdoor S02 concentrations for each of the six cities
included in the analysis. 2-44
Figure 3-1. Distribution of individual airway sensitivity to S02. 3-12
Figure 3-2. Odds ratios (95% CI) for incidence of morning asthma symptoms of 846
asthmatic children from the National Cooperative Inner-City Asthma Study. 3-18
Figure 3-3. Odds ratios (95% CI) for daily asthma symptoms of 990 asthmatic children from
the Childhood Asthma Management Program Study. 3-20
Figure 3-4. Odds ratios (95% CI) for incidence of cough among children, grouped by season. 3-22
Figure 3-5. Odds ratios (95% CI) for the incidence of lower respiratory tract or asthma
symptoms among children, grouped by season. 3-23
Figure 3-6. Relative risks (95% CI) of S02-associated emergency department visits and
hospitalizations for all respiratory causes among all ages and separated by age
group. 3-37
Figure 3-7. Relative risks (95% CI) of S02-associated emergency department visits and
hospitalizations for asthma among all ages and age-specific groups. 3-41
Figure 3-8. Relative risks (95% CI) of S02-associated emergency department visits and
hospitalizations for all respiratory causes and asthma, with and without
copollutant adjustment. 3-44
Figure 3-9. Relative risks (95% CI) of S02-associated emergency department visits (*) and
hospitalizations for all cardiovascular causes, arranged by age group. 3-66
Figure 3-10. All cause mortality excess risk estimates for S02 from the National Morbidity,
Mortality, and Air Pollution Study. 3-72
Figure 3-11. Relative risks (95% CI) of S02-associated all-cause (nonaccidental) mortality,
with and without copollutant adjustment, from multicity and meta-analysis
studies. 3-81
Figure 3-12. Relative risks (95% CI) of S02-associated mortality for all (nonaccidental),
respiratory, and cardiovascular causes from multicity studies. 3-82
Figure 3-13. Relative risks (95% CI) for low birth weight, grouped by trimester of S02
exposure. 3-101
Figure 3-14. Relative risks (95% CI) of S02-associated all-cause (nonaccidental) mortality,
with and without adjustment for sulfate, from longitudinal cohort studies. 3-113
Figure 4-1. Percent of mild and moderate asthmatics (vE = 40-50 L/min) experiencing an
S02-induced increase in sRaw of > 100% or a decrease in FEV1 of > 15%,
adjusted for effects of moderate to heavy exercise in clean air. 4-3
Figure 4-2. S02-induced increase in sRaw among S02-sensitive mild and moderate
asthmatics (n=38) following 10-min exposures during moderate to heavy exercise
(VE = 40-50 L/min). 4-4
Figure 4-3. S02-induced decrease in FEV1 among S02-sensitive mild and moderate
asthmatics (n=41) following 10 min exposures during moderate to heavy exercise
(VE = 40-50 L/min). 4-5
-------
Figure 4-4. Adjusted odds ratios of asthma hospitalizations by groupings of 1 -h max S02
concentrations in Bronx County, New York. 4-6
Figure 4-5. Relative odds ratio of incidence of lower respiratory tract symptoms smoothed
against 24-h avg S02 concentrations on the previous day, controlling for
temperature, city, and day of week. 4-7
Figure 4-6. Relative risks (95% CI) of age-specific associations between short-term exposure
to S02 and respiratory ED visits and hospitalizations. 4-17
Figure 5-1. Odds ratios (95% CI) for the association between short-term exposures to
ambient S02 and respiratory symptoms in children. 5-6
Figure 5-2. Relative risks (95% CI) for the association between short-term exposures to
ambient S02 and emergency department (ED) visits/hospitalizations for all
respiratory diseases and asthma in children. 5-7
-------
Abbreviations and Acronyms
A
ACS
ADS
AHR
AM
APHEA
APEX
APIMS
ARIC
ARP
AQCD
asl
atm
P
B[a]P
BHR
BS
CAMP
CARB
CASAC
CASTNet
CDC
CHAD
CHF
CHS
CH3-S-H
CH3-S-S-CH
CI
CMSA
CO
CoH
CONUS
COPD
CS2
CVD
DEN
DEP
DMS
ED
ECG
EIB
ELF
May 2008
Alpha
American Cancer Society
annular denuder system
airways hyperreactiveness
alveolar macrophages
Air Pollution on Health: a European Approach (study)
Air Pollution Exposure (model)
atmospheric pressure ionization mass spectrometer
Atherosclerosis Risk in Communities (study)
Acid Rain Program
Air Quality Criteria Document
above sea level
Atmosphere
beta; the calculated Health Effect Parameter
benzo[a]pyrene
bronchial hyperresponsiveness
black smoke
Childhood Asthma Management Program
California Air Resources Board
Clean Air Scientific Advisory Committee
Clean Air Status and Trends Network
Centers for Disease Control and Prevention
Consolidated Human Activities Database
congestive heart failure
Children's Health Study
methyl mercaptan
dimethyl disulfide
confidence interval
consolidated metropolitan statistical area
carbon monoxide
coefficient of haze
continental United States
chronic obstructive pulmonary disease
carbon disulfide
cardiovascular disease
diethylnitrosamine
diesel exhaust particle
dimethyl sulfide
emergency department
electrocardiography; electrocardiogram
exercise-induced bronchial reactivity
epithelial lining fluid
xiii DRAFT—DO NOT QUOTE OR CITE
-------
EMECAM
Spanish Multicentre Study on Air Pollution and Mortality
EPA
U.S. Environmental Protection Agency
eNO
exhaled nitric oxide
ET
extrathoracic
Fe
iron
FEMs
Federal Equivalent Methods
FEVo.75
forced expiratory volume in 0.75 second
FEV-i
forced expiratory volume in 1 second
FPD
flame photometric detection
FPD-TA
flame photometric detection-thermal analysis
FRM
Federal Reference Method
FVC
forced vital capacity
GAM
Generalized Additive Model(s)
GIS
Geographic Information System
GLM
Generalized Linear Model(s)
GSH
glutathione; reduced glutathione
GST
glutathione S-transferase (e.g., GSTM1, GSTP1, GSTT1)
H+
hydrogen ion
HEADS
Harvard-EPAAnnular Denuder System
HEI
Health Effects Institute
HF
high frequency
hno2
nitrous acid
HN03
nitric acid
ho2
hydroperoxyl; hydroperoxy radical
h2o
water
h2o2
hydrogen peroxide
HR
heart rate
HRV
heart rate variability
H2S
hydrogen sulfide
hso3_
hydrogen sulfite, bisulfite
HSO4"
bisulfate ion
h2so4
sulfuric acid
hv
solar ultraviolet photon
IARC
International Agency for Research on Cancer
ICD9
International Classification of Diseases, Ninth Revision
ICDs
implanted cardioverter defibrillators
ig
immunoglobulin (e.g., IgA, IgE, IgG)
IHD
ischemic heart disease
NASA
International Institute for Applied Systems Analysis
IL
interleukin (e.g., IL-4, IL-6, IL-8)
IOM
Institute of Medicine
IQR
interquartile range
ISA
Integrated Science Assessment
ISAAC
International Study of Asthma and Allergies in Children
May 2008 xiv DRAFT—DO NOT QUOTE OR CITE
-------
IUGR intrauterine growth retardation
K mass transfer coefficient
LF low frequency
LOD limit of detection
LRD lower respiratory disease
MCh methacholine
MENTOR Modeling Environment for Total Risk for One-Atmosphere studies
Ml myocardial infarction
MEF50% maximal midexpiratory flow at 50% of forced vital capacity
MMEF maximal midexpiratory flow
Mn manganese
MONICA Monitoring Trend and Determinants in Cardiovascular Disease
(registry)
MOZART-2 Model for Ozone and Related Chemical Tracers, version 2
MSA metropolitan statistical area
N, n number of observations
NAAQS National Ambient Air Quality Standards
NaCI sodium chloride
NaC03 sodium carbonate
NADP National Atmospheric Deposition Program
NAMS National Air Monitoring Stations
NAPAP National Acid Precipitation Assessment Program
NAS National Academy of Sciences
NCAR National Center for Atmospheric Research
NCEP National Center for Environmental Prediction
NCICAS National Cooperative Inner-City Asthma Study
NCore National Core Monitoring Network
NERL National Exposure Research Laboratory
NH4+ ammonium ion
NHAPS National Human Activity Pattern Survey
NHANES National Health and Nutrition Examination Survey
NMMAPS National Morbidity, Mortality, and Air Pollution Study
NO nitric oxide
N02 nitrogen dioxide
N03 nitrate radical
N03 nitrate ion
NOAA National Oceanic and Atmospheric Administration
NOx oxides of nitrogen
NR not reported
NRC National Research Council
NTN National Trends Network
NTP National Toxicology Program
02 molecular oxygen, diatomic oxygen
03 ozone
OCS carbonyl sulfide
May 2008
xv
DRAFT—DO NOT QUOTE OR CITE
-------
OH hydroxyl radical
OR odds ratio
P, p probability value
PAARC Air Pollution and Chronic Respiratory Diseases (study)
PAH polycyclic aromatic hydrocarbon
PC(S02) provocative concentration of S02 that produces a 100% increase
in specific airway resistance
PD20FEV1 20% decrease in forced expiratory volume in 1 second
PD20 provocative dose that produces a 20% decrease in FEV1
PD100 provocative dose that produces a 100% increase in sRAW
PEACE Pollution Effects on Asthmatic Children in Europe (study)
PEC pulmonary endocrine cell
PEF peak expiratory flow
PEMs personal exposure monitors
PF pulsed fluorescence
PM particulate matter
PM2 5 particulate matter with 50% upper cut point aerodynamic
diameter of 2.5 jjm for sample collection; surrogate for fine PM
PM10 particulate matter with 50% upper cut point aerodynamic
diameter of 10 jjm for sample collection
PM10-2 5 particulate matter with 10 jjm as upper cut point aerodynamic
diameter and 2.5 jjm as lower cut point for sample collection;
surrogate for thoracic coarse PM (does not include fine PM)
PM13 particulate matter with 50% upper cut point aerodynamic
diameter of 13 jjm for sample collection
PMT photomultipliertube
ppb parts per billion
ppbv parts per billion by volume
ppm parts per million
pptv parts per trillion by volume
PRB policy relevant background
PS passive sample
R, r correlation coefficient
RAR rapidly activating receptor
RAS roll-around system
Raw airway resistance
RH relative humidity
r-MSSD root mean square of successive differences in R-R intervals.
RR rate ratio; relative risk
S2" sulfur radical
SAB Science Advisory Board
SAPALDIA Study of Air Pollution and Lung Diseases in Adults
SAVIAH Small-Area Variation in Air Pollution and Health (study)
SD standard deviation
SDNN standard deviation of normal R-R intervals
SES socioeconomic status
May 2008
xvi
DRAFT—DO NOT QUOTE OR CITE
-------
SHEDS Simulation of Human Exposure and Dose System
SIDS sudden infant death syndrome
SNP single nucleotide polymorphism
S sulfur-35 radionuclide
SLAMS State and Local Air Monitoring Stations
SO sulfur monoxide
502 sulfur dioxide
503 sulfur trioxide
SO32" sulfite ion
SO42" sulfate ion
SOx sulfur oxides
S20 disulfur monoxide
SPM suspended particulate matter
sRaw specific airway resistance
STN Speciation Trends Network
t tau; atmospheric lifetime
TBARS thiobarbituric acid reactive substances
TEA triethanolamine
TNF tumor necrosis factor (e.g., TNF-a)
TSP total suspended particles
URI upper respiratory infections
UV ultraviolet
VE minute ventilation
May 2008
xvii
DRAFT—DO NOT QUOTE OR CITE
-------
Authors, Contributors, Reviewers
Authors
Dr. Jee Young Kim (S0X Team Leader)—National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Jeffrey Arnold—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. James S. Brown—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Barbara Buckley—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Ila Cote—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Douglas Johns—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Ellen Kirrane—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Thomas Long—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Thomas Luben—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Qingyu Meng—Oak Ridge Institute for Science and Education, Postdoctoral Research Fellow to
National Center for Environmental Assessment, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Anu Mudipalli—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Joseph Pinto—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Mary Ross—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. David Svendsgaard—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Lori White—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. William Wilson—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
May 2008
xix
DRAFT—DO NOT QUOTE OR CITE
-------
Dr. Brett Grover—National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Douglas Bryant—Intrinsik Science, Mississauga, Ontario, Canada
Dr. Arlene Fiore—Geophysical Fluid Dynamics Laboratory/National Oceanographic & Atmospheric
Administration, Princeton, NJ
Dr. Panos Georgopoulos—Computational Chemodynamics Laboratory, Environmental and Occupational
Health Sciences Institute, Piscataway, NJ
Dr. Vic Hasselblad—Duke University Medical Center, Durham, NC
Dr. Larry Horowitz—Geophysical Fluid Dynamics Laboratory/National Oceanographic and Atmospheric
Administration, Princeton University Forrestal Campus, Princeton, NJ
Ms. Annette Ianucci—Sciences International, Alexandria, VA
Dr. Kazuhiko Ito—Department of Environmental Medicine, New York University School of Medicine,
Tuxedo, NY
Dr. Jane Koenig— Department of Environmental and Occupational Health Sciences, University of
Washington, Seattle, WA
Dr. Therese Mar— Department of Environmental and Occupational Health Sciences, University of
Washington, Seattle, WA
Dr. James Riddle—Sciences International, Alexandria, VA
Contributors
Ms. Rebecca Daniels—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Jason Sacks—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Dale Allen—Department of Atmospheric and Oceanic Sciences, University of Maryland, College
Park, MD
Ms. Louise Camalier—Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Russell Dickerson—Department of Atmospheric and Oceanic Science, University of Maryland,
College Park, MD
Dr. Tina Fan—Environmental and Occupational Health Sciences Institute, Piscataway, NJ
Mr. William Keene—Department of Environmental Sciences, University of Virginia, Charlottesville, VA
Dr. Randall Martin—Department of Physics and Atmospheric Science, Dalhousie University, Halifax,
Nova Scotia, Canada
Dr. Maria Morandi—Department of Environmental Sciences, School of Public Health, University of
Texas - Houston Health Science Center, Houston, TX
May 2008
xx
DRAFT—DO NOT QUOTE OR CITE
-------
Dr. William Munger—Division of Engineering and Applied Sciences, Harvard University, Cambridge,
MA
Mr. Charles Piety—Department of Meteorology, University of Maryland, College Park, MD
Dr. Sandy Sillman—Department of Atmospheric, Ocean, and Space Sciences, University of Michigan,
Ann Arbor, MI
Dr. Helen Suh—Department of Environmental Health, Harvard School of Public Health, Boston, MA
Dr. Charles Wechsler—Environmental and Occupational Health Sciences Institute, Piscataway, NJ
Dr. Clifford Weisel—Environmental and Occupational Health Sciences Institute, Piscataway, NJ
Dr. Jim Zhang—Environmental and Occupational Health Sciences Institute, Piscataway, NJ
Reviewers
Dr. Tina Bahadori—American Chemistry Council, Arlington, VA
Dr. Tim Benner—Office of Science Policy, U.S. Environmental Protection Agency, Washington, DC
Dr. Daniel Costa—National Program Director for Air, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Robert Devlin—National Health and Environmental Effects Research Laboratory, U.S. Environmental
Protection Agency, Chapel Hill, NC
Dr. Judy Graham—American Chemistry Council, Arlington, VA
Dr. Stephen Graham—Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Beth Hassett-Sipple—Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Gary Hatch—National Health and Environmental Effects Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Scott Jenkins—Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. David Kryak—National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. John Langstaff—Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Morton Lippmann—Department of Environmental Medicine, New York University School of
Medicine, Tuxedo, NY
Dr. Karen Martin—Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. William McDonnell—William F. McDonnell Consulting, Chapel Hill, NC
May 2008
xxi
DRAFT—DO NOT QUOTE OR CITE
-------
Dr. Dave McKee—Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Lucas Neas—National Health and Environmental Effects Research Laboratory, U.S. Environmental
Protection Agency, Chapel Hill, NC
Dr. Russell Owen—National Health and Environmental Effects Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Haluk Ozkaynak—National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Jennifer Peel—Department of Environmental and Radiological Health Sciences, Colorado State
University, Fort Collins, CO
Mr. Harvey Richmond—Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Steven Silverman—Office of General Counsel, U.S. Environmental Protection Agency, Washington,
DC
Dr. Michael Stewart—Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Susan Stone—Office of Air Quality Planning and Standards, Office of Air and Radiation, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Ms. Chris Trent—Office of Air Quality Planning and Standards, Office of Air and Radiation, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. John Vandenberg—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Alan Vette—National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Ms. Debra Walsh—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Ron Williams—National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC
May 2008
xxii
DRAFT—DO NOT QUOTE OR CITE
-------
S0X Project Team
Executive Direction
Dr. Ila Cote (Acting Director)—National Center for Environmental Assessment-RTP Division, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Ms. Debra Walsh (Deputy Director)—National Center for Environmental Assessment-RTP Division, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Mary Ross—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Scientific Staff
Dr. Jee Young Kim (SOx Team Leader)—National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Jeff Arnold—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. James S. Brown—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Barbara Buckley—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Douglas Johns—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Ellen Kirrane—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Dennis Kotchmar—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Thomas Long—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Thomas Luben—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Qingyu Meng—Oak Ridge Institute for Science and Education, Postdoctoral Research Fellow to
National Center for Environmental Assessment, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Joseph Pinto—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Jason Sacks—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
May 2008
xxiii
DRAFT—DO NOT QUOTE OR CITE
-------
Dr. David Svendsgaard—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Lori White—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. William Wilson—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Technical Support Staff
Ms. Emily R. Lee—Management Analyst, National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Ms. Ellen Lorang—Information Manager, National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Ms. Christine Searles—Management Analyst, National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Mr. Richard Wilson—Clerk, National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Document Production Staff
Ms. Barbra H. Schwartz—Task Order Manager, Computer Sciences Corporation, Morrisville, NC
Mr. John A. Bennett—Technical Information Specialist, Library Associates of Maryland, Rockville, MD
Mr. David Casson—Publication/Graphics Specialist, TekSystems, Raleigh, NC
Ms. Melissa Cesar—Publication/Graphics Specialist, Computer Sciences Corporation, Morrisville, NC
Ms. Rebecca Early—Publication/Graphics Specialist, TekSystems, Raleigh, NC
Mr. Eric Ellis—Records Management Technician, InfoPro, Inc., McLean, VA
Ms. Kristin Hamilton—Publication/Graphics Specialist, TekSystems, Raleigh, NC
Ms. Stephanie Harper—Publication/Graphics Specialist, TekSystems, Raleigh, NC
Ms. Sandra L. Hughey—Technical Information Specialist, Library Associates of Maryland, Rockville,
MD
Dr. Barbara Liljequist—Technical Editor, Computer Sciences Corporation, Morrisville, NC
Ms. Molly Windsor—Graphic Artist, Computer Sciences Corporation, Morrisville, NC
May 2008
xxiv
DRAFT—DO NOT QUOTE OR CITE
-------
Clean Air Scientific Advisory Committee for
NOx and SOx Primary NAAQS
Chairperson
Dr. Rogene Henderson*, Scientist Emeritus, Lovelace Respiratory Research Institute, Albuquerque, NM
Members
Mr. Ed Avol, Professor, Preventive Medicine, Keck School of Medicine, University of Southern
California, Los Angeles, CA
Dr. John R. Balmes, Professor, Department of Medicine, Division of Occupational and Environmental
Medicine, University of California, San Francisco, CA
Dr. Ellis Cowling*, University Distinguished Professor At-Large, North Carolina State University,
Colleges of Natural Resources and Agriculture and Life Sciences, North Carolina State University,
Raleigh, NC
Dr. James D. Crapo*, Professor, Department of Medicine, National Jewish Medical and Research
Center, Denver, CO
Dr. Douglas Crawford-Brown*, Director, Carolina Environmental Program; Professor, Environmental
Sciences and Engineering; and Professor, Public Policy, Department of Environmental Sciences and
Engineering, University of North Carolina at Chapel Hill, Chapel Hill, NC
Dr. Terry Gordon, Professor, Environmental Medicine, NYU School of Medicine, Tuxedo, NY
Dr. Dale Hattis, Research Professor, Center for Technology, Environment, and Development, George
Perkins Marsh Institute, Clark University, Worcester, MA
Dr. Patrick Kinney, Associate Professor, Department of Environmental Health Sciences, Mailman
School of Public Health, Columbia University, New York, NY
Dr. Steven Kleeberger, Professor, Laboratory Chief, Laboratory of Respiratory Biology, NIH/NIEHS,
Research Triangle Park, NC
Dr. Timothy Larson, Professor, Department of Civil and Environmental Engineering, University of
Washington, Seattle, WA
* Members of the statutory Clean Air Scientific Advisory Committee (CASAC) appointed by the EPA Administrator
May 2008
xxv
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Preface
Legislative Requirements
Section 109 (42 U.S. Code, 2003b) directs the Administrator to propose and promulgate
"primary" and "secondary" National Ambient Air Quality Standards (NAAQS) for pollutants
listed under section 108. Section 109(b)(1) 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 required to
protect the public welfare from any known or anticipated adverse effects associated with the
presence of [the] pollutant in the ambient air."2 The requirement that primary standards include
an adequate margin of safety was intended to address uncertainties associated with inconclusive
scientific and technical information available at the time of standard setting. It was also intended
to provide a reasonable degree of protection against hazards that research has not yet identified.
S qq Lead Industries Association v. EPA, 647 F.2d 1130, 1154 (D.C. Cir 1980), cert, denied. 449
U.S. 1042 (1980); American Petroleum Institute v. Costle, 665 F.2d 1176, 1186 (D.C. Cir. 1981)
cert, denied. 455 U.S. 1034 (1982). Both kinds of uncertainties are components of the risk
associated with pollution at levels below those at which human health effects can be said to
occur with reasonable scientific certainty. Thus, in selecting primary standards that include an
adequate margin of safety, the Administrator is seeking not only to prevent pollution levels that
have been demonstrated to be harmful but also to prevent lower pollutant levels that may pose an
unacceptable risk of harm, even if the risk is not precisely identified as to nature or degree.
In selecting a margin of safety, the EPA considers such factors as the nature and severity of
the health effects involved, the size of sensitive population(s) at risk, and the kind and degree of
the uncertainties that must be addressed. The selection of any particular approach to providing an
adequate margin of safety is a policy choice left specifically to the Administrator's judgment. See
Lead Industries Association v. EPA, supra. 647 F.2d at 1161-62.
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)]. (Senate., 1970)
2 Welfare effects as defined in section 302(h) [42 U.S.C. 7602(h)] include, but are not limited to, "effects on soils, water, crops, vegetation, man-made
materials, animals, wildlife, weather, visibility and climate, damage to and deterioration of property, and hazards to transportation, as well as effects
on economic values and on personal comfort and well-being."
May 2008
xxvii
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
In setting standards that are "requisite" to protect public health and welfare, as provided in
section 109(b), EPA's task is to establish standards that are neither more nor less stringent than
necessary for these purposes. In so doing, EPA may not consider the costs of implementing the
standards. See generally Whitman v. American Trucking Associations, 531 U.S. 457, 465472,
475-76 (D C. Cir. 2001).
Section 109(d)(1) requires that "not later than December 31, 1980, and at 5-year intervals
thereafter, the Administrator shall complete a thorough review of the criteria published under
section 108 and the national ambient air quality standards.. .and shall make such revisions in
such criteria and standards and promulgate such new standards as may be appropriate..." 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 (CAS AC) of EPA's Science
Advisory Board.
History of Reviews of the Primary NAAQS for Sulfur Oxides
On April 30, 1971, the EPA promulgated primary NAAQS for sulfur oxides (SOx). These
primary standards, which were based on the findings outlined in the original 1969 Air Quality
Criteria for Sulfur Oxides, were set at 0.14 parts per million (ppm) averaged over a 24-hour
period, not to be exceeded more than once per year, and 0.030 ppm annual arithmetic mean with
S02 as the indicator. In 1982, EPA published the Air Quality Criteria for Particulate Matter and
Sulfur Oxides (EPA, 1982) along with an addendum of newly published controlled human
exposure studies, which updated the scientific criteria upon which the initial standards were
based. In 1986, a second addendum was published presenting newly available evidence from
epidemiologic and controlled human exposure studies (EPA, 1986b). In 1988, EPA published a
proposed decision not to revise the existing standards (53 FR 14926). However, EPA specifically
requested public comment on the alternative of revising the current standards and adding a new
1-hour primary standard of 0.4 ppm.
As a result of public comments on the 1988 proposal and other post-proposal
developments, EPA published a second proposal on November 15, 1994 (59 FR 58958). The
May 2008
xxviii
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
1994 re-proposal was based in part on a supplement to the second addendum of the criteria
document, which evaluated new findings on short-term S02 exposures in asthmatics (EPA,
1994b). As in the 1988 proposal, EPA proposed to retain the existing 24-hour and annual
standards. The EPA also solicited comment on three regulatory alternatives to further reduce the
health risk posed by exposure to high 5-minute peaks of S02 if additional protection were judged
to be necessary. The three alternatives were: 1) revising the existing primary S02NAAQS by
adding a new 5-minute standard of 0.60 ppm S02; 2) establishing a new regulatory program
under section 303 of the Act to supplement protection provided by the existing NAAQS, with a
trigger level of 0.60 ppm S02, one expected exceedance; and 3) augmenting implementation of
existing standards by focusing on those sources or source types likely to produce high 5-minute
peak concentrations of S02. On May 22, 1996, EPA's final decision, that revisions of the NAAQS
for SOx were not appropriate at that time, was announced in the Federal Register (61 FR 25566).
In that decision, EPA announced an intention to propose guidance, under section 303 of the Act,
to assist states in responding to short-term peak levels of S02. The basis for the decision, and
subsequent litigation, is discussed in Annex A.
May 2008
xxix
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Chapter 1. Introduction
This second external review draft Integrated Science Assessment (ISA) presents a concise
synthesis of the most policy-relevant science to form the scientific foundation for the review of
the primary (health-based) NAAQS for SOx. This document is intended to "accurately reflect the
latest scientific knowledge useful in indicating the kind and extent of identifiable effects on
public health which may be expected from the presence of [a] pollutant in ambient air" (Clean
Air Act, Section 108, 2003a)1. Contained herein are the key information and judgments formerly
contained in the Air Quality Criteria Document (AQCD) for SOx; additional details of the
pertinent scientific literature published since the last review, as well as selected older studies of
particular interest, are included in a series of annexes to the draft ISA. This second external
review draft ISA thus serves to update and revise the information available at the time of the
previous review of the NAAQS for SOx in 1996.
S02 is the most important of the monomelic sulfur oxides (SOx) for both atmospheric
chemistry and health effects. SOx is usually defined to include SO3 and H2SO4 as well, but
neither is present in the atmosphere in concentrations significant for human exposures.
Descriptions of the atmospheric chemistry of SOx include both gaseous and particulate species; a
meaningful analysis would not be possible otherwise. Most studies on the health effects of
gaseous SOx focus on S02; effects of other gaseous species are considered as information is
available. The health effects of particulate SOx are included in the review of the NAAQS for
particulate matter (PM). In evaluating the health evidence, this second external draft ISA
considers possible influences of other atmospheric pollutants, including interactions of S02 with
other co-occurring pollutants such as PM, nitrogen oxides (NOx), carbon monoxide (CO), and
ozone (O3).
As discussed in the Integrated Plan for Review of the Primary NAAQS for SOx (EPA,
2007), a series of policy-relevant questions frames this review to provide a scientific basis for a
decision about whether the current primary NAAQS for SOx should be retained or revised. The
primary NAAQS for SOx, with S02 serving as the indicator, is set at 0.14 parts per million (ppm),
averaged over a 24-h period, not to be exceeded more than once per year, and 0.030 ppm annual
arithmetic mean. This second external review draft ISA focuses on evaluation of the newly
1A review of the secondary SOx NAAQS, in conjunction with a review of the secondary NAAQS for NOx, is underway independently, as is a review of
the primary NAAQS for NOx and a review of the primary and secondary effects of PM.
May 2008
1-1
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
available scientific evidence to best inform consideration of these framing questions, including
the following:
¦ How has new information altered/substantiated the scientific support for the occurrence
of health effects following short- and/or long-term exposure to levels of SOx found in
the ambient air?
¦ How does new information influence conclusions from the previous review regarding the
effects of SOx on susceptible populations?
¦ At what levels of SOx exposure do health effects of concern occur?
¦ How has new information altered conclusions from previous reviews regarding the
plausibility of adverse health effects caused by SOx exposure?
¦ To what extent have important uncertainties identified in the last review been reduced?
Have new uncertainties emerged?
¦ What are the air quality relationships between short-term and long-term exposures
to SOx?
1.1. Document Development
EPA initiated the current formal review of the NAAQS for SOx on May 15, 2006 with a
call for information from the public (FR, 2006). In addition to the call for information,
publications are identified through an ongoing literature search process that includes extensive
computer database mining on specific topics. Additional publications were identified by EPA
scientists in a variety of disciplines by combing through relevant, peer-reviewed scientific
literature obtained through these ongoing literature searches, reviewing previous EPA reports,
and a review of reference lists from important publications. All relevant epidemiological, human
clinical, and animal toxicological studies, including those related to exposure-response
relationships, mechanism(s) of action, or susceptible subpopulations published since the last
review were considered. Added to the body of research were EPA's analyses of air quality and
emissions data, studies on atmospheric chemistry, transport, and fate of these emissions, as well
as issues related to exposure to SOx. Further information was acquired from consultation with
content and area experts and the public. Annex A has more discussion of search strategies and
criteria for study selection.
May 2008
1-2
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
1.2. Document Organization
This second external review draft ISA is composed of five chapters. This introductory
chapter presents background information, discusses the purpose of the document, and
characterizes the search, evaluation and retrieval process of policy-relevant scientific studies.
Chapter 2 highlights key concepts or issues relevant to understanding the atmospheric chemistry,
sources, exposure, and dosimetry of SOx, following a "source-to-dose" paradigm. Chapter 3
evaluates and integrates epidemiological, human clinical, and animal toxicological information
relevant to the review of the primary NAAQS for SOx. Chapter 4 has information related to the
public health impact of ambient SOx exposure, with emphasis on potentially susceptible and
vulnerable population groups. Finally, Chapter 5 summarizes key findings and conclusions from
the atmospheric sciences, ambient air data analyses, exposure assessment, dosimetry, and health
effects for consideration in the review of the NAAQS for SOx.
A series of annexes supplement this second external review draft ISA. The annexes provide
additional details of the pertinent literature published since the last review, as well as selected
older studies of particular interest. These annexes contain information on:
¦ atmospheric chemistry of SOx as well as the sampling and analytic methods for
measurement of SOx1;
¦ environmental concentrations and human exposure to SOx;
¦ toxicological studies of health effects in laboratory animals;
¦ human clinical studies of health effects related to peak (5-10 min) and short-term (1-h or
longer) exposure to SOx; and
¦ epidemiological studies of health effects from short- and long-term exposure to SOx.
Detailed information about methods and results of health studies is summarized in tabular
format, and generally includes information about: concentrations of SOx and averaging times;
study methods employed; results and comments; and quantitative results for relationships
between effects and exposure to SOx.
1 This section also includes information on N02, in order to support the reviews of the primary and secondary NAAQS for both S02 and N02. The
atmospheric chemistry of NOx and SOx are intricately linked; discussion of their combined chemistry is more effective and more efficient than a
separate discussion of each pollutant.
May 2008
1-3
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
1.3. EPA Framework for Causal Determinations
It is important to have a consistent and transparent basis to evaluate the causal nature of air
pollution-induced health effects. The framework described below establishes uniform language
concerning causality and brings more specificity to the findings. It draws standardized language
from across the Federal government and wider scientific community, especially from the recent
National Academy of Sciences (NAS) Institute of Medicine (IOM) document, Improving the
Presumptive Disability Decision-Making Process for Veterans (IOM, 2007), the most recent
comprehensive work on evaluating the causality of health effects. This section:
¦ describes the kinds of scientific evidence used in establishing a general causal
relationship between exposure and health effects;
¦ defines cause, in contrast to statistical association;
¦ discusses the sources of evidence necessary to reach a conclusion about the existence of a
causal relationship;
¦ highlights the issue of multifactorial causation;
¦ identifies issues and approaches related to uncertainty; and
¦ provides a framework for classifying and characterizing the weight of evidence in support
of a general causal relationship.
Approaches to assessing the separate and combined lines of evidence (e.g.,
epidemiological, human clinical, animal toxicological, and in vitro studies) have been formulated
by a number of regulatory and science agencies, including the IOM of the National Academies of
Science (IOM, 2008), International Agency for Research on Cancer (IARC, 2006), EPA
Guidelines for Carcinogen Risk Assessment (EPA, 2005), Centers for Disease Control and
Prevention (CDC, 2004), and National Acid Precipitation Assessment Program (NAPAP, 1991).
Highlights or excerpts from the various decision framework documents are included in Annex A.
These formalized approaches offer guidance for assessing causality. The frameworks are
similar in nature, although adapted to different purposes, and have proved effective in providing
a uniform structure and language for causal determinations. Moreover, these frameworks must
support decision-making under conditions of uncertainty.
May 2008
1-4
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
1.3.1. Scientific Evidence Used in Establishing Causality
The most compelling evidence of a causal relationship between pollutant exposures and
human health effects comes from human clinical studies. This type of study experimentally
evaluates the health effects of administered exposures in humans under highly-controlled
laboratory conditions.
In epidemiological or observational studies of humans, the investigator does not control
exposures or intervene with the study population. Broadly, observational studies can describe
associations between exposures and effects. These studies fall into several categories: cross-
sectional, prospective cohort, time-series, and panel studies. "Natural experiments" occur
occasionally in epidemiology; these include comparisons of health effects before and after a
change in population exposures, such as closure of a pollution source.
Experimental animal data complements the clinical and observational data; these studies
can help characterize effects of concern, exposure-response relationships, sensitive
subpopulations and modes of action. In the absence of clinical or epidemiological data, animal
data alone may be sufficient to support a likely causal determination, assuming that humans
respond similarly to the experimental species.
1.3.2. Association and Causation
Association and causation are not the same. "Cause" conveys the notion of a significant,
effectual relationship between an agent and an associated disorder or disease in the population.
"Association" is the statistical dependence among multiple (two or more) events, characteristics,
or other variables. An association is merely prima facie evidence for causation; alone, it is not
sufficient for proof of a causal relationship between exposure and disease. Unlike an association,
a causal claim supports the creation of counterfactual claims; that is, a claim about what the
world would have been like under different or changed circumstances (IOM, 2008). Currently,
much of the newly available health information evaluated in the draft ISA comes from
epidemiological studies that report a statistical association between exposure and health
outcome.
It is recognized that many of the health outcomes evaluated in ISAs have complex
etiologies. Most diseases, such as cancer or coronary heart disease, result from a complex web of
causation, whereby one or more agents can initiate a disease process. The outcome could depend
May 2008
1-5
DRAFT—DO NOT QUOTE OR CITE
-------
1 on many factors, including age, genetic susceptibility, nutritional status, immune competence,
2 social factors, and others (Gee and Payne-Sturges, 2004; IOM, 2008). Figure 1-1 shows a
3 diagram of a variety of etiologic factors that contribute to disease. Exposure to multiple agents
4 together could result in synergistic or antagonistic effects that are different from what might
5 result from exposure to each agent separately.1 The results are the net effect of many actions and
6 counteractions.
Cumulative Risks
(The combined risks from aggregate exposures to multiple agents or stressors)
Race/Ethnicity
Residential Location
i i
Neighborhood
Community
Resources
Stressors
,_fc=a
Factors
Community
Stress
Environmental
Hazards &
Pollutants
Exposure
Community
Level
Vulnerability
Stress/Coping,
Life Stage/Style
Individual Stressors
I
Modified from Gee & Payne-Sturges, 2004
Internal dose
Biologically
effective dose
Individual
Level
Vulnerability
Health effects/
disparities
Figure 1-1. Exposure-disease-stress model for environmental health disparities.
1.3.3. Evidence for Going beyond Association to Causation
7 Moving from association to causation involves elimination of alternative explanations for
8 the association. Human clinical studies are experiments in which subjects in a population are
1 For example, a multiplicative interaction relative risk (RR) could be defined as RRM(™» ~ RR/WRRt *¦ RRs- An additive interaction RR could
bedefined as RRSnlledd) = RRjou - RRe - RRs + 1
May 2008
1-6
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
randomly allocated into groups, usually called study and control groups, and exposed to a
pollutant or a sham. The results are assessed by rigorous comparison of rates of appropriate
outcomes between the study and control groups. Randomized human clinical studies are
generally regarded as the most scientifically rigorous method of hypothesis testing available. By
assigning exposure randomly, the study design attempts to remove the effect of any factor that
might influence exposure. Done properly, and setting aside randomness, only a causal
relationship between exposure and health outcome should produce observed associations in
randomized clinical trials. In another type of human clinical study, the same subject is exposed to
a pollutant and a sham at different time points, and the responses to the two types of exposures
are compared. This study design is also effective at controlling for any potential confounders,
since the subject is serving as his/her own control. Alack of observation of effects from human
clinical studies does not necessarily mean that a causal relationship does not occur. Human
clinical studies are often limited because the study population is generally small, which restricts
the ability to discern statistically significant findings. In addition, the most susceptible
individuals may be explicitly excluded (for ethical reasons), and other susceptible individuals or
groups, such as those with nutritional deficits, may not be included.
Inferring causation from epidemiological studies requires consideration of potential
confounders. When associations are found in epidemiological studies, one approach to remove
spurious association from possible confounders is statistical control on characteristics that may
differ between exposed and unexposed persons; this is frequently termed "adjustment."
Multivariable regression models constitute one tool for estimating the association between
exposure and outcome after adjusting for characteristics of participants that might confound the
results. Another way to adjust for potential confounding is through stratified analysis, i.e.,
examining the association within homogeneous groups with the confounding variable. The use of
stratified analyses has an additional benefit: it allows examination of effect modification through
comparison of the effect estimates across different groups. If investigators successfully measured
characteristics that distort the results, adjustment of these factors help separate a spurious from a
true causal association. Appropriate statistical adjustment for confounders requires identifying
and measuring all reasonably expected confounders. Deciding which variables to control for in a
statistical analysis of the association between exposure and disease depends on knowledge about
May 2008
1-7
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
possible mechanisms. Identifying these mechanisms makes it possible to control for potential
sources that may result in a spurious association.
Measurement error is another problem encountered when adjusting for spurious
associations. In multivariate analyses, the effects of a well-measured covariate may be
overestimated, in contrast to a more poorly measured covariate. There are several components
that contribute to exposure measurement error in these studies, including the difference between
true and measured ambient concentrations, the difference between average personal exposure to
ambient pollutants and ambient concentrations at central monitoring sites, and the use of average
population exposure rather than individual exposure estimates.
It is difficult to identify and measure all potential confounders in epidemiological studies.
Confidence that unmeasured confounders are not producing the findings is increased when
multiple studies are conducted in various settings using different subjects or exposures; each of
which might eliminate another source of confounding from consideration. Thus, multi-city
studies which use a consistent method to analyze data from across locations with different levels
of covariates can provide insight on potential confounding in associations. The number and
degree of diversity of covariates, as well as their relevance to the potential confounders, remain
matters of scientific judgment. Intervention studies, because of their experimental nature, can be
particularly useful in characterizing causation.
In addition to clinical and epidemiological studies, the tools of experimental biology have
been valuable for developing insights into human physiology and pathology. Laboratory tools
have been extended to explore the effects of putative toxicants on human health, especially
through the study of model systems in other species. Background knowledge of the biological
mechanisms by which an exposure might or might not cause disease can prove crucial in
establishing, or negating, a causal claim. At the same time, species can differ from each other in
fundamental aspects of physiology and anatomy (e.g., metabolism, airway branching, hormonal
regulation) that may limit extrapolation. Testable hypotheses about the causal nature of proposed
mechanisms or modes of action are central to utilizing experimental data in causal
determinations.
May 2008
1-8
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
1.3.4. Multifactorial Causation
Scientific judgment is needed regarding likely sources and magnitude of confounding,
together with judgment about how well the existing constellation of study designs, results, and
analyses address this potential threat to inferential validity. One key consideration in this review
is evaluation of the potential contribution of SOx to health effects, when it is a component of a
complex air pollutant mixture. There are multiple ways by which SOx might cause or be
associated with adverse health effects. First, the reported SOx effect estimates in epidemiological
studies may reflect independent SOx effects on respiratory health. Second, ambient SOx may be
serving as an indicator of complex ambient air pollution mixtures that share the same source as
SOx (i.e., combustion of sulfur-containing fuels or metal smelting). Finally, copollutants may
mediate the effects of SOx or SOx may influence the toxicity of copollutants. Epidemiologists use
the term "interaction" or "effect modification" to denote the departure of the observed joint risk
from what might be expected based on the separate effects of the factors. These possibilities are
not necessarily exclusive. In addition, confounding can result in the production of an association
between adverse health effects and S02 that is actually attributable to another factor that is
associated with S02 in a particular study. Multivariate models are the most widely used strategy
to address confounding in epidemiological studies, but such models are not readily interpreted
when assessing effects of covarying pollutants such as PM, S02, and nitrogen dioxide (NO2).
1.3.5. Uncertainty
The science of estimating the causal influence of an exposure on disease is an uncertain
one. There are two distinct levels of uncertainty to be considered here:
¦ Model uncertainty—uncertainty regarding gaps in scientific theory required to make
predictions on the basis of causal inferences.
¦ Parameter uncertainty—uncertainty as to the statistical estimates within each model.
Assessment of model uncertainty involves:
¦ whether exposure causes the health outcome;
¦ the set of confounders associated with exposure and health outcome;
May 2008
1-9
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
¦ which parametric forms best describe the relations of exposure and confounders with
outcome; and
¦ whether other forms of bias could be affecting the association.
Model uncertainty is not limited to the qualitative causal structure: it also involves factors
such as uncertainty about the parametric form of the model specified, the variables included and
whether or not measurement error is modeled. When mechanistic knowledge exists, this
important source of uncertainty can be reduced. In contrast, uncertainty about the parameter
estimates (regression coefficients) for a given model is a well-studied problem. The important
point is that these reports of uncertainty are conditional on the model providing a sufficiently
adequate approximation of reality so that inferences are valid. The overall scientific inference
involves evaluation of model uncertainty and uncertainty about parameter estimates given to
each model.
There are systematic, quantitative approaches for including uncertainty about the model in
an assessment of overall uncertainty about a causal inference, such as sensitivity analysis and
model averaging. Sensitivity analysis attempts to quantify the sensitivity of the parameter
estimate in relation to assumptions about the model. Uncertainty ranges can be estimated using
classical analysis (Robinson, 1989) or the Monte Carlo technique (Eggleston, 1993). By
averaging over many different competing models, Bayesian Model Averaging incorporates
model uncertainty into conclusions about parameters and prediction.
1.3.6. Application of Framework
EPA uses a two-step approach to evaluate the scientific evidence on health effects of
exposure to criteria pollutants. These two steps address two policy relevant questions noted in
the beginning of this chapter - what are (if any) the effects of SOx on susceptible populations,
given the total body of evidence, and at what levels of SOx exposure do health effects of concern
occur. The first step determines the weight of evidence in support of causation and characterizes
the strength of any resulting causal classification. The second step includes further evaluation of
the quantitative evidence regarding the concentration-response relationships and the levels,
duration and pattern of exposures at which effects are observed.
May 2008
1-10
DRAFT—DO NOT QUOTE OR CITE
-------
1 To aid judgment, various "aspects"1 of causality have been discussed by many
2 philosophers and scientists. The most widely cited aspects of causality in epidemiology, and
3 public health in general, were articulated by Sir Austin Bradford Hill in 1965 and have been
4 widely used (EPA, 2005; IARC, 2006; Surgeon General, 2004; IOM, 2008). These nine aspects
5 (Hill, 1965) have been modified (below) for use in causal determinations specific to health and
6 environmental effects and pollutant exposures.2
Table 1-1 Aspects to aid in judging causality.
1. Consistency of the observed association. An inference of causality is strengthened
when a pattern of elevated risks is observed across several independent studies. The
reproducibility of findings constitutes one of the strongest arguments for causality. If
there are discordant results among investigations, possible reasons such as differences
in exposure, confounding factors, and the power of the study are considered.
2. 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. A
modest risk, however, does not preclude a causal association and may reflect a lower
level of exposure, an agent of lower potency, or a common disease with a high
background level.
3. Specificity of the observed association. As originally intended, this refers to increased
inference of causality if one cause is associated with a single effect or disease (Hill,
1965). Based on our current understanding this is now considered one of the weaker
guidelines for causality; for example, many agents cause respiratory disease and
respiratory disease has multiple causes. The ability to demonstrate specificity under
certain conditions remains, however, a powerful attribute of experimental studies.
Thus, although the presence of specificity may support causality, its absence does not
exclude it.
4. Temporal relationship of the observed association. A causal interpretation is
strengthened when exposure is known to precede development of the disease.
1The "aspects" describe by Hill (1965) have become, in the subsequent literature, more commonly described as "criteria." The original term "aspects"
is used here to avoid confusion with 'criteria' as it is used, with different meaning, in the Clean Air Act.
2 The Hill apects were developed for use with epidemiology data. They have been modified here for use with a broader array of data, i.e.,
epidemiological, controlled human exposure, and animal toxicological studies, as well as in vitro data, and to be more consistent with EPA's
Guidelines for Carcinogen Risk Assessment.
May 2008
1-11
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
5. Biological gradient (exposure-response relationship). A clear exposure-response
relationship (e.g., increasing effects associated with greater exposure) strongly suggests
cause and effect, especially when such relationships are also observed for duration of
exposure (e.g., increasing effects observed following longer exposure times). There are,
however, many possible reasons that a study may fail to detect an exposure-response
relationship. Thus, although the presence of a biologic gradient may support causality,
the absence of an exposure-response relationship does not exclude a causal
relationship.
6. Biological plausibility. An inference of causality tends to be strengthened by
consistency with data from experimental studies or other sources demonstrating
plausible biological mechanisms. Alack of biologic understanding, however, is not a
reason to reject causality.
7. Coherence. An inference of causality may be strengthened by other lines of evidence
(e.g., clinical and animal studies) that support a cause-and-effect interpretation of the
association. The absence of other lines of evidence, however, is not a reason to reject
causality.
8. Experimental evidence (from human populations). Experimental evidence is
generally available from human populations for the criteria pollutants. The strongest
evidence for causality can be provided when a change in exposure brings about a
change in adverse health effect or disease frequency in either clinical or observational
studies.
9. Analogy. Structure activity relationships and information on the agent's structural
analogs can provide insight into whether an association is causal. Similarly,
information on mode of action for a chemical, as one of many structural analogs, can
inform decisions regarding likely causality.
While these aspects provide a framework for assessing the evidence, they do not lend
themselves to being considered in terms of simple formulas or fixed rules of evidence leading to
conclusions about causality (Hill, 1965). For example, one cannot simply count the number of
studies reporting statistically significant results or statistically nonsignificant results for health
effects and reach credible conclusions about the relative weight of the evidence and the
likelihood of causality. Rather, these important considerations are taken into account with the
goal of producing an objective appraisal of the evidence, informed by peer and public comment
and advice, which includes weighing alternative views on controversial issues. Additionally, it is
important to note that the principles in Table 1-1 cannot be used as a strict checklist, but rather to
determine the weight of the evidence for inferring causality. In particular, the absence of one or
more of the principles does not automatically exclude a study from consideration (e.g., see
discussion in U.S. Surgeon General's Report, 2004).
May 2008
1-12
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
1.3.7. First Step—Determination of Causality
In the ISA, EPA assesses results of recent publications, in light of evidence available
during the previous NAAQS review, to draw conclusions on the causal relationships between
relevant pollutant exposures and health outcomes. This second external review draft ISA uses a
five-level hierarchy that classifies the weight of evidence for causation, not just association.1;
that is, whether the weight of scientific evidence makes causation at least as likely as not, in the
judgment of the reviewing group. In developing this hierarchy, EPA has drawn on the work of
previous evaluations, most prominently the IOM's Improving the Presumptive Disability
Decision-Making Process for Veterans (IOM, 2008), EPA's Guidelines for Carcinogen Risk
Assessment (EPA, 1986a), and the U.S. Surgeon General's smoking reports (U.S. Surgeon
General's Report, 2004). These efforts are presented in more detail in Annex A. In the draft ISA,
EPA uses a series of five descriptors to characterize the weight of evidence on whether
associations are in fact causal. This weight of evidence evaluation is based on various lines of
evidence from epidemiological studies, animal studies, or other mechanistic, toxicological, or
biological sources. These separate judgments are integrated into a qualitative statement about the
overall weight of the evidence and causality. The five descriptors for causal determination are
described in Table 1-2.
Table 1-2. Weight of evidence for causal determination.
Sufficient to infer a causal
relationship
Evidence is sufficient to conclude that there is a causal relationship between relevant
pollutant exposure and the outcome. That is, a positive association has been observed
between the pollutant and the outcome in studies in which chance, bias, and
confounding could be ruled out with reasonable confidence. For example, controlled
human exposures, epidemiologic "natural experiments," or observational studies
supported by other lines of evidence. Generally, determination is based on multiple
studies by multiple investigators.
Sufficient to infer a likely
causal relationship (i.e.,
more likely than
than not).
Evidence is sufficient to conclude a likely causal association between relevant pollutant
exposures and the outcome. That is, a positive association has been observed between
the pollutant and the outcome in studies in which chance and bias can be ruled out with
reasonable confidence but potential confounding issues remain. For example, a)
observational studies show positive associations but copollutant exposures are difficult
to address and/or other lines of evidence (controlled human exposure, animal, or
mechanism of action information) are limited or inconsistent; or b) animal evidence
from multiple studies, sex, or species is positive but limited or no human data are
available. Generally, determination is based on multiple studies by multiple
investigators.
1 It should be noted that the CDC and IOM frameworks use a four-category hierarchy for the strength of the evidence. A five-level hierarchy is used
here to be consistent with the EPA Guidelines for Carcinogen Risk Assessment and to provide a more nuanced set of categories.
May 2008
1-13
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Suggestive, but not
sufficient to infer a causal
relationship
Evidence is suggestive of an association between relevant pollutant exposures and the
outcome, but is limited because chance, bias and confounding cannot be ruled out. For
example, at least one high-quality study shows a positive association but the results of
other studies are inconsistent.
Inadequate to infer the
presence or absence of a
causal relationship
The available studies are of insufficient quality, consistency or statistical power to
permit a conclusion regarding the presence or absence of an association between
relevant pollutant exposure and the outcome. For example, studies fail to control for
confounding, have inadequate exposure assessment, or fail to address latency.
Suggestive of no causal
relationship
Several adequate studies, covering the full range of levels of exposure that human
beings are known to encounter and considering sensitive subpopulations, are mutually
consistent in not showing a positive association between exposure and the outcome at
any level of exposures. In addition, the possibility of a very small elevation in risk at
the levels of exposure studied can never be excluded.
1.3.8. Second Step—Evaluation of Population Response
Beyond judgments regarding causality are questions relevant to characterizing exposure
and risk to populations (i.e., at what levels do health effects occur). Such questions include:
¦ Under what exposure conditions (dose or exposure, duration and pattern) are effects
seen?
¦ What is the shape of the concentration-response or dose-response relationship?
¦ What population groups appear to be affected or more susceptible to effects?
On the population level, causal and likely causal claims typically characterize how risk—
the probability of health effects—changes in response to exposure. Initially, the response is
evaluated within the range of observation. Approaches to analysis of the range of observation of
epidemiological and human clinical studies are determined by the type of study and methods of
exposure/dose and response measurement. Extensive human data for concentration-response
analyses exists for all criteria pollutants, unlike most other environmental pollutants. Animal data
also can inform concentration-response, particularly relative to dosimetry, mechanisms of action,
and characteristics of sensitive subpopulations.
An important consideration in characterizing the public health impacts associated with
exposure to a pollutant is whether the concentration-response relationship is linear across the full
concentration range encountered, or if nonlinear relationships exist along any part of this range.
Of particular interest is the shape of the concentration-response curve at and below the level of
the current standards. The complex molecular and cellular events that underlie cancer and
noncancer toxicity are likely to be both linear and nonlinear, and vary depending on dose.
May 2008
1-14
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Additionally, many chemicals and agents may act by perturbing naturally occurring background
processes that lead to disease. At the human population level, however, various sources of
variability and uncertainty tend to smooth and "linearize" the concentration-response function
(such as the low data density in the lower concentration range, possible influence of
measurement error, and individual differences in susceptibility to air pollution health effects).
These attributes of population dose-response may explain why the available human data at
ambient concentrations for some environmental pollutants (e.g.,ozone, lead [Pb], PM,
secondhand tobacco smoke, radiation) do not exhibit evident thresholds for cancer or noncancer
health effects, even though likely mechanisms of action include nonlinear processes for some
key events. These attributes of human population dose-response relationships have been
extensively discussed in the broader epidemiological literature (e.g., Rothman and Greenland,
1998).
1.4. Conclusions
This second external review draft ISA strives to present a concise review, synthesis, and
evaluation of the most policy-relevant science, and communicates critical science judgments
relevant to the NAAQS review. It reviews the most policy relevant evidence from
epidemiological, human clinical, and animal toxicological studies, including mechanistic
evidence from basic biological science. Annexes to the ISA provide additional details of the
literature published since the last review. A framework for making critical judgments concerning
causality is presented in this chapter. It relies on a widely accepted set of principles and
standardized language to express evaluation of the evidence. This approach can bring rigor and
clarity to the current and future assessments. This ISA should assist EPA and others, now and in
the future, to represent accurately what is presently known—and what remains unknown—
concerning the effects of sulfur oxides on human health.
May 2008 1 -15 DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Chapter 2. Source to Tissue Dose
This chapter contains basic information about concepts and findings in atmospheric
sciences, human exposure assessment, and human dosimetry. It is meant to serve as a prologue
for the detailed discussions of health effects data in Chapters 3 and 4. Section 2.1 gives an
overview of the sources of S02. Atmospheric chemistry processes involved in the oxidation of
S02 and those involved in the production of S02 from reduced sulfur gases in the atmosphere are
discussed in Section 2.2. A description of S02 measurement methods and related issues are
presented in Section 2.3. Data for ambient S02 concentrations are characterized in Section 2.4.
Policy relevant background concentrations of S02, i.e., those concentrations defined to result
from uncontrollable emissions, are also presented in Section 2.4. Factors related to personal
exposure to S02 are discussed in Section 2.5. Finally, Section 2.6 covers the dosimetry of S02 in
the respiratory tract. This organization generally follows that given in the National Research
Council (NRC) paradigm for integrating air pollutant research (NRC, 1998).
2.1. Sources of Sulfur Oxides
Industrial emissions of S02 in the United States are mainly due to combustion of fossil
fuels by electrical utilities (-66 %) and industry (-29%); transportation-related sources
contribute minimally (-5%) (2002 statistics) (EPA, 2006d). Thus, most S02 emissions originate
from point sources. Annex B has a detailed breakdown of emissions by source category. Almost
all of the sulfur in fuel is released as volatile components (S02 or SO3) during combustion.
Hence, based on sulfur content in fuel stocks, sulfur emissions can be calculated to a higher
degree of accuracy than other pollutants such as nitrogen oxides or primary PM. However, these
estimates given above are national averages and may not accurately reflect the contribution of
specific local sources for determining individual exposure to S02 at a particular location and
time. For example, shipping and in-port activities may be a significant source of S02 in some
coastal cities (Wang et al., 2007).1
1 Ships and commercial boats contribute approximately 25% of the SO2 emissions in the South Coast Air Basin, and 50% of statewide SO2 emissions
(Dabdub and Vutukuru, 2008). Because of the importance of S02 emissions, the ports of Long Beach and Los Angeles are part of a Sulfur
Emissions Control Area in which sulfur contents of fuels are not to exceed 1.5%. Modeling studies by Vutukuru (2008) also indicate that ships
contribute just over 1 part per billion (ppb) S02 (for a 24-h avg) to Long Beach, and a few tenths of a ppb to locations further inland .
May 2008 2-1 DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
The largest natural sources of S02 are volcanoes and wildfires. Although S02 constitutes a
relatively minor fraction (0.005% by volume) of total volcanic emissions (Holland, 1978),
concentrations in volcanic plumes can be in the range of several to tens of ppm. Volcanic sources
of S02 in the U.S. are limited to the Pacific Northwest, Alaska, and Hawaii. Emissions of S02
from burning vegetation are generally in the range of 1 to 2% of the biomass burned (Levine and
Pinto, 1998). Sulfur is a component of amino acids in vegetation and is released during
combustion. Gaseous sulfur emissions from this source are mainly in the form of S02.
In addition to its role as an emitted primary pollutant, S02 is also produced by the
photochemical oxidation of reduced sulfur compounds such as dimethyl sulfide (CH3-S-CH3, or
DMS), hydrogen sulfide (H2S), carbon disulfide (CS2), carbonyl sulfide (OCS), methyl
mercaptan (CH3-S-H), and dimethyl disulfide (CH3-S-S-CH3). The sources for these compounds
are mainly biogenic (see Annex Table B-6). Emissions of reduced sulfur species are associated
typically with marine organisms living either in pelagic or coastal zones, and with anaerobic
bacteria in marshes and estuaries. Emissions of DMS from marine plankton represent the largest
single atmospheric source of reduced sulfur species (Berresheim et al., 1995). Other than OCS,
which is lost mainly by photolysis (e-folding lifetime, [t] ~6 months), species are lost mainly by
reaction with hydroxyl radical (OH) and NO3 radicals, and are relatively short-lived; lifetimes
range from a few hours to a few days (see Annex Table B-2). Reaction with NO3 radicals at night
most likely represents the major loss process for DMS and methyl mercaptan. Although the
mechanisms for the oxidation of DMS are not completely understood, excess sulfate in marine
aerosol appears related mainly to the production of S02 from the oxidation of DMS. Emissions of
sulfur from natural sources are small compared to industrial emissions within the U.S. However,
important exceptions occur locally as the result of volcanic activity, wildfires and in certain
coastal zones as described above.
Because OCS is relatively long-lived, it can survive oxidation in the troposphere and be
transported upward into the stratosphere. Crutzen (1976) proposed that its oxidation to sulfate in
the stratosphere serves as the major source of the stratospheric aerosol layer. However, Myhre
et al. (2004) proposed that S02 transported upward from the troposphere by deep convection is
the most likely source, since the flux of OCS is too small. In addition, in situ measurements of
the isotopic composition of sulfur in stratospheric sulfate do not match those of OCS (Leung,
May 2008
2-2
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
2002). Thus, anthropogenic S02 emissions could be important precursors to the formation of the
stratospheric aerosol layer.
2.2. Atmospheric Chemistry
The only forms of monomeric sulfur oxides of interest in tropospheric chemistry are S02
and SO3. SO3 can be emitted from the stacks of power plants and factories; however, it reacts
extremely rapidly with H20 in the stacks or immediately after release into the atmosphere to
form H2S04, which mainly condenses onto existing particles when particle loadings are high; it
can nucleate to form new particles under lower concentration conditions. Thus, only S02 is
present in the tropospheric boundary layer at concentrations of concern for human exposures.
The gas phase oxidation of S02 is initiated by the reaction
SO y + OH + M —~ HSO j + M (Reaction 2 1)
where M is an atmospheric constituent such as N2 and O2 that helps stabilize the reaction
product. Reaction 2-1 is followed by
+ 02^>S03 + HO2
SOj + H20 ¦—y H2SO4
(Reaction 2-2)
(Reaction 2-3)
Because the saturation vapor pressure of H2S04 is extremely low, it will be removed rapidly by
transfer to the aqueous phase of aerosol particles and cloud drops. Depending on atmospheric
conditions and concentrations of ambient particles and gaseous species that can participate in
new particle formation, it can also nucleate to form new particles. Rate coefficients for the
reactions of S02 with either the hydroperoxyl radical (H02) or N03 are too low to be significant
(Jet Propulsion Laboratory, 2003).
The major sulfur species in clouds are hydrogen sulfite (HSO3 ) and the sulfite ion (SO32 ).
Both are derived from the dissolution of S02 in water, and are referred to as S(IV); bisulfate ion
(HSO4 ) and sulfate (sulfate) are referred to as S(VI). The chief species capable of oxidizing
S(IV) to S(VI) in cloud water are O3, peroxides (either hydrogen peroxide [H2O2] or organic
peroxides), hydroxyl (OH) radicals, and ions of transition metals such as iron (Fe), manganese
(Mn) and copper (Cu) that can catalyze the oxidation of S(IV) to S(VI) by O2. The basic
May 2008
2-3
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
mechanism of the aqueous phase oxidation of S02 has long been studied and can be found in
numerous texts on atmospheric chemistry, e.g., Seinfeld and Pandis (1998), Finlayson-Pitts and
Pitts (1999), Jacob (1999), and Jacobson (2002). Following Jacobson (2002), the steps involved
in the aqueous phase oxidation of S02 can be summarized as
Dissolution of S02
S02{g)S02(aq)
The formation and dissociation of H2SO3
S02(aq) + H20(aq) <=> H2SO? <=> H+ + HSO$~ O 2H + + SOs ¦
(Reaction 2-4)
(Reaction 2-5)
In the pH range commonly found in rainwater (pH 2 to 6), the most important reaction
converting S(IV) to S(VI) is
HS03 - + H2 02 + //+<=> S04 2~ + ff20 + 2 ft (Reaction 2-6)
as SO32 is much less abundant than HSO3 .
For pH up to about 5.3, H202 is the dominant oxidant, while at pH > 5.3, 03 becomes
dominant, followed by Fe(III), using characteristic values found in Seinfeld and Pandis (1998).
However, differences in concentrations of oxidants result in differences in the pH at which this
transition occurs. It should also be noted that the oxidation of S02 by O3 and O2 tends to be self-
limiting: as sulfate is formed, the pH decreases and the rates of these reactions decrease. Higher
pH levels are expected to be found mainly in marine aerosols. However, in marine aerosols, the
chloride-catalyzed oxidation of S(IV) may be more important (Hoppel and Caffrey, 2005; Zhang
and Millero, 1991). Because the ammonium ion (NH4+) is so effective in neutralizing acidity, it
affects the rate of oxidation of S(IV) to S(VI) and the rate of dissolution of S02 in particles and
cloud drops.
A comparison of the relative rates of oxidation by gas and aqueous phase reactions by
Warneck (1999) indicates that on average only about 20% of S02 is oxidized by gas phase
reactions; the remainder is oxidized by aqueous phase reactions. In areas away from strong
pollution sources, the S02 t is ~7 days, based on measurements of the rate constant for Reaction
2-1 (Jet Propulsion Laboratory, 2003) and a nominal concentration for the OH radical of 106/cm3.
May 2008
2-4
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
However, the mechanism of S02 oxidation at a particular location depends on local
environmental conditions. For example, near stacks, oxidants such as OH radicals are depleted
and almost no S02 is oxidized in the gas phase. Further downwind, as the plume is diluted with
background air, the gas phase oxidation of S02 increases in importance. Finally, even further
downwind when conditions in the plume can become more oxidizing than in background air, the
S02 oxidation rate could exceed that in background air. S02 in the planetary boundary layer is
also removed from the atmosphere by dry deposition to moist surfaces, resulting in an
atmospheric t with respect to dry deposition of approximately 1 day to 1 week. Wet deposition of
sulfur naturally depends on the variable nature of rainfall, but in general results in a t of S02 ~7
days, too. These two processes, oxidation and deposition, lead to an overall lifetime of S02 in the
atmosphere of 3 to 4 days.
2.3. Measurement Methods and Associated Issues
Currently, ambient S02 is measured using instruments based on pulsed ultraviolet (UV)
fluorescence. The UV fluorescence monitoring method for atmospheric S02 was developed to
improve on the flame photometric detection (FPD) method, which in turn had replaced the
pararosaniline wet chemical method. This latter method is still the EPA's Federal Reference
Method (FRM) for atmospheric S02, but is rarely used due to its complexity and slow response,
even in its automated forms. Both the UV fluorescence and FPD methods are designated as
Federal Equivalent Methods (FEMs) by EPA, but UV fluorescence has largely supplanted the
FPD approach because of the UV method's inherent linearity and because the FPD method needs
consumable hydrogen gas.
In the UV fluorescence method, S02 molecules absorb UV light at one wavelength and
emit UV light at longer wavelengths in the process known as fluorescence, through excitation of
the S02 molecule to a higher energy (singlet) electronic state. Once excited, the molecule decays
nonradiatively to a lower-energy electronic state from which it then decays to the original or
electronic state by emitting a photon of light at a longer wavelength (i.e., a lower-energy photon)
than the original, incident photon. The intensity of the emitted light is thus proportional to the
number of S02 molecules in the sample gas.
In commercial analyzers, light from a high-intensity UV lamp passes through a bandwidth
filter, allowing only photons with wavelengths around the S02 absorption peak (near 214
May 2008 2-5 DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
nanometers [nm]) to enter the optical chamber. The light passing through the source bandwidth
filter is collimated using a UV lens and passes through the optical chamber, where it is detected
on the opposite side of the chamber by the reference detector. A photomultiplier tube (PMT) is
offset from and placed perpendicular to the light path to detect the S02 fluorescence. Since the
S02 fluorescence at 330 nm is different from its excitation wavelength, an optical bandwidth
filter is placed in front of the PMT to filter out any stray light from the UV lamp. A lens is
located between the filter and the PMT to focus the fluorescence onto the active area of the
detector and optimize the fluorescence signal. The limit of detection (LOD) for a non-trace level
S02 analyzer is required to be 10 ppb (FR, 2006). However, most commercial analyzers have
detection limits of about 3 ppb; many monitors might have lower effective detection limits. The
EPA, through its National Core (NCore) initiative (EPA, 2005) is in the process of supporting
state, local, tribal, and federal networks in the implementation of newer trace-level S02
instrumentation. These new trace-level instruments have detection limits of 0.1 ppb or lower.
More information related to S02 sampling and measurement is in Annex B.5.
2.3.1. Sources of Positive Interference
The most common source of interference to the UV fluorescence method for S02 is from
other gases that fluoresce in a similar fashion when exposed to UV radiation. The most signifi-
cant of these are polycyclic aromatic hydrocarbons (PAHs), of which naphthalene is a prominent
example. Xylene is another common hydrocarbon that can cause fluorescent interference. Conse-
quently, any such aromatic hydrocarbons in the optical chamber can act as positive interference.
To remove this source of interference, high-sensitivity S02 analyzers, such as those to be used in
the NCore network (EPA, 2005), have hydrocarbon scrubbers to remove these compounds from
the sample stream before the sample air enters the optical chamber.
Luke (1997) reported positive artifacts of a modified pulsed fluorescence detector
generated by the coexistence of nitric oxide (NO), carbon disulfide (CS2), and a number of
highly fluorescent aromatic hydrocarbons such as benzene, toluene, o-xylene, w-xylene,
p-xylene, w-ethyltoluene, ethylbenzene, and 1,2,4-trimethylbenzene. The positive artifacts could
be reduced by using a hydrocarbon "kicker" membrane. At a flow rate of 300 standard cc min 1
and a pressure drop of 645 torr across the membrane, the interference from ppm levels of many
aromatic hydrocarbons was eliminated. NO fluoresces in a spectral region close to that of S02.
May 2008
2-6
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
However, in high-sensitivity S02 analyzers, the bandpass filter in front of the PMT is designed to
prevent NO fluorescence from being detected at the PMT. Care must be exercised when using
multicomponent calibration gases containing both NO and S02, so that the NO rejection ratio of
the S02 analyzer is sufficient to prevent NO interference.
The most common source of positive bias (as contrasted with positive spectral
interference) in high-sensitivity S02 monitoring is stray light in the optical chamber. Since S02
can be electronically excited by a broad range of UV wavelengths, any stray light with an
appropriate wavelength that enters the optical chamber can excite S02 in the sample and increase
the fluorescence signal. Furthermore, stray light at the wavelength of the S02 fluorescence that
enters the optical chamber may impinge on the PMT and increase the fluorescence signal.
Several design features minimize stray light, including the use of light filters, dark surfaces, and
opaque tubing.
Nicks and Benner (2001) reported a sensitive S02 chemiluminescence detector based on a
differential measurement: response from ambient S02 is determined by the difference between
air containing S02 and air scrubbed of S02 when both air samples contain other detectable sulfur
species. Assuming monotonic efficiency of the sulfur scrubber, all positive artifacts should also
be reduced with this technique.
2.3.2. Sources of Negative Interference
Nonradiative deactivation (quenching) of excited S02 molecules can occur from collisions
with common molecules in air, including nitrogen, oxygen, and water. During collisional
quenching, the excited S02 molecule transfers energy, kinetically allowing the S02 molecule to
return to the original lower energy state without emitting a photon. Collisional quenching results
in a decrease in the S02 fluorescence and, hence, an underestimation of S02 concentration in the
air sample. Of particular concern is the variable water vapor content of air. Luke (1997) reported
that the response of the detector could be reduced by an amount of ~7 to 15% at water vapor
mixing ratios of 1 to 1.5 mole percent (relative humidity [RH] = 35 to 50% at 20 to 25°C and 1
atmosphere [atm] for a modified pulsed fluorescence detector [Thermo Environmental
Instruments, Model 43 s]). Condensation of water vapor in sampling lines must be avoided, as
water on the inlet surfaces can absorb S02 from the sample air. The simplest approach to avoid
condensation is to heat sampling lines to a temperature above the expected dew point and to
May 2008
2-7
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
within a few degrees of the controlled optical bench temperature. At very high S02
concentrations, reactions between electronically excited S02 and ground state S02 might occur,
forming S03 and SO (Calvert et al., 1978). However, the possibility that this artifact might be
affecting measurements at very high S02 levels has not been examined.
2.3.3. Other Techniques for Measuring so2
More sensitive techniques for measuring S02 are available, but most of these systems are
too complex and expensive for routine monitoring applications. However, techniques such as
those described by Luke (1997) can be used to improve the sensitivity of ambient S02 monitors
by eliminating sources of common interference. See descriptions in Annex section B.5.
2.4. Environmental Concentrations of SOx
2.4.1. Design Criteria for the NAAQS so2 Monitoring Networks1
Trace level S02 monitoring is currently required at the approximately 75 proposed NCore
sites, as noted in CFR 40 Part 58 Appendices C and D. Continued operation of existing State and
Local Air Monitoring Sites (SLAMS) for S02 using Federal Reference Methods (FRM) or
Federal Equivalent Methods (FEM) is required until discontinuation is approved by the EPA
Regional Administrator. Where SLAMS S02 monitoring is required, at least one of the sites must
be a maximum concentration site for that specific area. In 2007, there were -500 S02 monitors
reporting values to the EPA Air Quality System database (AQS).
The appropriate spatial scales for S02 SLAMS monitoring are the microscale, middle, and
possibly neighborhood scales.
¦ Micro and middle scale—Some data uses associated with microscale and middle scale
measurements for S02 include assessing the effects of control strategies to reduce
concentrations (especially for the 3-hour and 24-hour averaging times), and monitoring
air pollution episodes.
¦Neighborhood scale—This scale applies where there is a need to collect air quality data
as part of an ongoing S02 stationary source impact investigation. Typical locations
might include suburban areas adjacent to S02 stationary sources, for example, or for
1 This section is adapted from Code of the Federal Register 40 CFR Parts 53 and 58 and Appendix E to Part 58, as revised: Vol. 71, No. 200 / 17
October 2006
May 2008
2-8
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
determining background concentrations as part of studies of population responses to
S02 exposure.
Horizontal and Vertical Placement
The probe, or at least 80 percent of the monitoring path, must be located between 2 and 15
meters above ground level for all S02 monitoring sites. The probe, or at least 90 percent of the
monitoring path, must be positioned at least 1 meter vertically or horizontally from any
supporting structure, walls, parapets, penthouses, etc., and away from dusty or dirty areas. If the
probe, or a significant portion of the monitoring path, is located near the side of a building, it
should be located on the windward side relative to the prevailing wind direction during the
season of highest concentration potential for the pollutant being measured.
Spacing from Minor Sources
Local minor sources of a primary pollutant such as S02 can affect concentrations of that
particular pollutant at a monitoring site. If the objective for that site is to investigate these local
primary pollutant emissions, then the site should be located where the spatial and temporal
variability in these emissions can be captured. This type of monitoring site would likely be the
microscale type. If a monitoring site is to be used to determine air quality over a much larger
area, such as a neighborhood or city, a monitoring agency should avoid placing a monitor probe,
path, or inlet near local, minor sources. The plume from the local minor sources should not be
allowed to inappropriately influence the air quality data collected.
To minimize these potential interferences, the probe, or at least 90 percent of the
monitoring path, must be placed away from furnace or incineration flues, or other minor sources
of S02. The separation distance should take into account the heights of the flues, type of waste or
fuel burned, and the sulfur content of the fuel.
Spacing from Obstructions
Buildings and other obstacles may possibly scavenge S02, and can act to restrict airflow
for any pollutant. To avoid this interference, the probe, inlet, or at least 90 percent of the
monitoring path must have unrestricted airflow and be located away from obstacles. The distance
from the obstacle to the probe, inlet, or monitoring path must be at least twice the height of the
obstruction's protrusion. An exception can be made for measurements taken in street canyons or
May 2008
2-9
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
at source-oriented sites where buildings and other structures are unavoidable. Generally, a probe
or monitoring path located near or along a vertical wall is undesirable, because air moving along
the wall may be subject to possible removal mechanisms. A probe, inlet, or monitoring path must
have unrestricted airflow in an arc of at least 180 degrees. This arc must include the predominant
wind direction for the season of greatest pollutant concentration potential.
Special consideration must be devoted to the use of open path analyzers, due to their
inherent potential sensitivity to certain types of interferences, or optical obstructions. A
monitoring path must be clear of all trees, brush, buildings, plumes, dust, or other optical
obstructions, including potential obstructions that may move due to wind, human activity, growth
of vegetation, etc. Temporary optical obstructions, such as rain, particles, fog, or snow, should be
considered when locating an open path analyzer. Any temporary obstructions that are of
sufficient density to obscure the light beam will affect the ability of the open path analyzer to
measure pollutant concentrations continuously. Transient, but significant obscuration of
especially longer measurement paths could occur because certain meteorological conditions
(e.g., heavy fog, rain, snow) and/or aerosol levels are of sufficient density to prevent the
analyzer's light transmission. If certain compensating measures are not otherwise implemented at
the onset of monitoring (e.g., shorter path lengths, higher light source intensity), data recovery
during periods of greatest primary pollutant potential could be compromised. For instance, if
heavy fog or high particulate levels are coincident with periods of projected NAAQS-threatening
pollutant potential, the resulting data may not be representative for reflecting maximum pollutant
concentrations, despite the fact that the site may otherwise exhibit an acceptable, even
exceedingly high overall valid data capture rate.
Spacing from Trees
Trees can provide surfaces for S02 adsorption or reactions, and surfaces for particle
deposition. Trees can also act as obstructions in cases where they are located between the air
pollutant sources or source areas and the monitoring site, and where the trees are of sufficient
height and leaf canopy density to interfere with normal airflow around the probe, inlet, or
monitoring path. To reduce possible interference, the probe, inlet, or at least 90 percent of the
monitoring path must be at least 10 meters or further from the drip line of trees.
May 2008
2-10
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
For microscale sites, no trees or shrubs should be located between the probe and the source
under investigation, such as a roadway or a stationary source.
2.4.2. Monitor Locations in Selected Areas of the U.S.
Figures 2-1 through 2-6 illustrate the 2005 geospatial locations of monitors for S02, N02,
CO, particulate matter <10 [j,m (PMio), particulate matter <2.5 [j,m (PM2.5), and O3. These
locations, sited in several cities in six states, were selected as relevant for S02 health effects
studies; see the summaries and assessments of health effects in Chapter 4, and the discussion of
intracity S02 correlations that follows. For each state, Figure A shows locations of each monitor
for all six pollutants; Figure B shows only the S02 monitor locations. Totals for each monitor
type are included. These figures demonstrate the important point that not all S02 monitors in any
Consolidated Metropolitan Statistical Area (CMSA) are co-located with monitors for other
pollutants. Two examples are given below.
Table 2-1. Monitor counts for California and San Diego County, 2005.
so2
N02
03
CO
PM10
PM2.5
California (all)
35
105
176
86
177
97
San Diego County
4
9
10
6
7
7
Table 2-2. Monitor counts for Ohio and Cuyahoga County, 2005.
O
(O
no2
O3
CO
PM10
PM2.5
Ohio (all)
31
4
49
15
49
49
Cuyahoga County
4
2
3
4
6
7
Table 2-1 lists the totals for all criteria air pollutant monitors (except Pb) in California, as
well as the subset of these monitors in San Diego County. At each of the four sites where S02
was measured, NO2, CO, PM10, PM2.5, and O3 were also measured, with the exception of PM2.5
at one site (AQS ID 060732007) in Otay Mesa, CA. Table 2-2 lists the totals for all criteria air
pollutant monitors (except Pb) in Ohio, as well as the subset of in Cuyahoga County.
May 2008
2-11
DRAFT—DO NOT QUOTE OR CITE
-------
1 In Cuyahoga County, PMio and PM2.5 were measured at all four sites where S02 was also
2 measured in 2005, but O3 and CO were not measured at any of those four sites; NO2 was only
3 measured at one site (AQS ID 39050060) near Cleveland's city center and -0.5 km from the
4 intersection of Interstate Highways 77 and 90.
May 2008
2-12
DRAFT—DO NOT QUOTE OR CITE
-------
Sacramen
San FranciscoJ
San Josl
San Francis)
Sacramen
Franciscq|
San Josl
Highway
Monitor Location
Interstate
+
CO (86)
Federal
A
N02 (105)
State
V
03 (176)
O
PM10 (177)
~
PM25(177)
~
S02 (35)
San Diego
Source: US EPA Office of Air and Radiation AQS Database
Figure 2-1. Criteria pollutant monitor locations (A) and S02 monitor locations (B), California,
2005. Shaded counties have at least one monitor.
May 2008
2-13
DRAFT—DO NOT QUOTE OR CITE
-------
Cleveland
Toledo
Clevelan
incinnati
Highway
Monitor Location
Interstate
+
CO (15)
Federal
A
N02 (4)
State
V
03 (49)
o
PM10 (49)
~
PM25 (49)
~
S02 (31)
Source: US ERA Office of Air and Radiation AQS Database
Figure 2-2. Criteria pollutant monitor locations (A) and S02 monitor locations (B), Ohio, 2005.
Shaded counties have at least one monitor.
2.4.3. Ambient so2 Concentrations in Relation to so2 Sources
1 S02 data collected from the SLAMS and NAMS networks, like those illustrated in
2 Figures 2-1 through 2-6, show that the decline in S02 emissions from electric generating utilities
3 has substantially improved air quality. Not one monitored exceedance of the S02 annual ambient
May 2008
2-14
DRAFT—DO NOT QUOTE OR CITE
-------
1 air quality standard in the lower 48 States of the United States has been recorded since 2000,
2 according to the EPA Acid Rain Program (ARP) 2005 Progress Report (EPA, 2006a). EPA's
3 trends data (www.epa.gov/airtrends) reveal that the national composite average S02 annual mean
4 ambient concentration decreased by 48% from 1990 to 2005; the largest single-year reduction
5 was 1994-95, the ARP's first operating year (EPA, 2006a). Figure 2-7 depicts data for S02
6 emissions in the contiguous United States (CONUS) during those years, with state-level totals.
Tucson
Tucson
Highway
Monitor Location
Interstate
+
CO (20)
Federal
A
N02 (13)
State
V
03 (45)
O
P
CO
o
1
CL
~
PM25(16)
~
S02 (7)
Source: US EPA Office of Air and Radiation AQS Database
Figure 2-3. Criteria pollutant monitor locations (A) and S02 monitor locations (B), Arizona, 2005.
Shaded counties have at least one monitor.
May 2008
2-15
DRAFT—DO NOT QUOTE OR CITE
-------
AI lento wn
X* '
risburg
Highway
Monitor Location
Interstate
+
CO (25)
Federal
A
N02 (29)
State
V
03 (47)
O
PM10 (46)
~
PM2 5 (49)
•£r
S02 (42)
Source: US EPA Office of Air and Radiation AQS Database
Figure 2-4. Criteria pollutant monitor locations (A) and S02 monitor locations (B), Pennsylvania,
2005. Shaded counties have at least one monitor.
May 2008
2-16
DRAFT—DO NOT QUOTE OR CITE
-------
Albany1
'onkers ^
Highway
Monitor Location
Interstate
+
CO (11)
Federal
A
N02 (9)
State
V
03 (34)
O
PM10 (10)
~
PM25 (28)
~
S02 (25)
Buffalo
Buffalo
Source: US EPA Office of Air and Radiation AQS Database
Figure 2-5. Criteria pollutant monitor locations (A) and S02 monitor locations (B), New York,
2005. Shaded counties have at least one monitor.
May 2008
2-17
DRAFT—DO NOT QUOTE OR CITE
-------
Boston
'Springfield
Highway
Monitor Location
Interstate
+
CO (5)
Federal
A
N02 (13)
State
V
03 (16)
O
PM-io (11)
~
PM2 5 (22)
~
S02 (10)
Source: US EPA Office of Air and Radiation AQS Database
Figure 2-6. Criteria pollutant monitor locations (A) and S02 monitor locations (B), Massachusetts,
2005. Shaded counties have at least one monitor.
Boston
I
1 These emissions data trends are consistent with the trends in the observed ambient
2 concentrations from the Clean Air Status and Trends Network (CASTNet). Following
3 implementation of the Phase I controls on ARP sources between 1995 and 2000, significant
4 reductions in S02 and ambient SOf ~ concentrations were observed at CASTNet sites throughout
5 the eastern United States. The mean annual concentrations of S02 and S042 from CASTNet's
6 long-term monitoring sites can be compared using two 3-year periods, 1989-1991 and
7 2003-2005, shown in Figure 2-8 for S02 and Figure 2-9 for SO f
May 2008
2-18
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
¦ S02 Emissions in 1990
HB S02 Emissions in 1995
~ SO? Emissions in 2000
¦ S02 Emissions in 2005
Scale: Largest bar equals
2.2 million tons of S02
emissions in Ohio, 1990
Figure 2-7. State-level S02 emissions, 1990-2005.
Source: Environmental Protection Agency Clean Air Markets Division (www.epa.gov/airmarkets/index.html).
From 1989 through 1991—that is, in the years prior to implementation of the ARP
Phase I— the highest ambient mean concentrations of S02 and SO/ ~ were observed in western
Pennsylvania and along the Ohio River Valley: > 20 jig/nr' (-8 ppb) S02 and > 15 ug/nr' SOf .
As with S02, in the years since the ARP controls were enacted, both the magnitude of SO42
concentrations and their areal extent have been significantly reduced, with the largest decreases
again along the Ohio River Valley.
Figure 2-10 depicts the magnitude and spatial distribution of S02 emissions in 2006 from
sources in the ARP for the CONUS. This depiction clearly shows the continuing
overrepresentation of S02 sources in the United States east of the Mississippi River, a trend even
stronger in the central Ohio Ri ver Valley, as evident in the smoothed concentration plots in
Figure 2-8. As shown in Table 2-3, regional distributions of S02 and SOr concentrations
averaged for 2003-2005 reflect this geospatial emissions source difference as well.
May 2008
2-19
DRAFT—DO NOT QUOTE OR CITE
-------
Figure 2-8. Annual mean ambient S02 concentration, 1989 through 1991 (a), and 2003 through
2005 (b).
Source: U.S. EPACASTNet
S02
Figure 2-9. Annual mean ambient SO/~ concentration, 1989 through 1991 (a), and 2003 through
2005(b).
Source: U.S. EPACASTNet
May 2008
2-20
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
• Under 5.000 tons
• 5.000 to 25.000 tons
• 25.000 to 50.000 tons
• 50.000 to 100.000 tons
• 100,000 to 207.000 torn
Figure 2-10. Annual S02 emissions for Acid Rain Program cooperating facilities, 2006.
Dots represent monitoring sites. Lack of shading for Southern Florida indicates lack of monitoring coverage.
Source: Environmental Protection Agency, Clean Air Markets Division (wvwv.epa.gov/airmarkets/index.html).
2.4.4. Spatial and Temporal Variability of Ambient so2 Concentrations
S02 concentrations have been falling throughout all regions of the ("ONUS, as
demonstrated by the CASTNet data reviewed above. In and around most individual CMSAs, the
trends are also toward lower S02 levels. Table 2-4 shows that many annual and even 1-h mean
concentrations for the years 2003 through 2005 were consistently at or below the operating LOD
of ~3 ppb for the standard sensitivity UV fluorescence S02 monitors deployed in the regulatory
networks, while the aggregate mean value over all 3 years and all sites in and around the CMSAs
was just above the LOD at ~4 ppb, and identical to the 1-h and 24-h means. Hence, it appears
reasonable to aggregate up in time from available 1-h samples to daily and even annual exposure
estimates.
Figure 2-11 shows the composite diel variation in hourly S02 concentrations in boxplot
form from all monitors reporting S02 data into AQS. The AQS contains measurements of air
pollutant concentrations in the 50 states, plus the District of Columbia, Puerto Rico, and the
Virgin Islands, for the six criteria air pollutants (S02, NO2, PM, CO, Pb, O3), as well as for
hazardous air pollutants.
May 2008
2-21
DRAFT—DO NOT QUOTE OR CITE
-------
Table 2-3. Regional distribution of S02 and S042- ambient concentrations, averaged for 2003-05.
REGION
CONCENTRATION
S02 (ppb)
so42" (jjg/m3)
Mid-Atlantic
3.3
4.5
Midwest
2.3
3.8
Northeast
1.2
2.5
Southeast
1.3
4.1
Table 2-4. Distributions of temporal averaging inside and outside CMSAs.
AVERAGING TIME
N
MEAN
PERCENTILES
MAX
MONITOR LOCATIONS
1
5
10
25
30
50
70
75
90
95
99
1-h Max Concentration
Inside CMSAs
332405
13
1
1
1
3
4
7
13
16
30
45
92
714
Outside CMSAs
53417
13
1
1
1
1
2
5
10
13
31
51
116
636
1-h Avg Concentration
Inside CMSAs
7408145
4
1
1
1
1
1
2
4
5
10
15
34
714
Outside CMSAs
1197179
4
1
1
1
1
1
2
3
3
7
13
36
636
24-h Avg Concentration
Inside CMSAs
327918
4
1
1
1
1
2
3
5
6
10
13
23
148
Outside CMSAs
52871
4
1
1
1
1
1
2
3
4
8
12
25
123
Annual Avg Concentration
Inside CMSAs
898
4
1
1
1
1
2
4
5
6
8
10
12
15
Outside CMSAs
143
4
1
1
1
1
2
3
4
5
8
9
13
14
Aggregate 3-yrAvg Concentration, 2003-2005
Inside CMSAs
283
4
1
1
1
2
3
3
5
5
8
10
12
14
Outside CMSAs
42
4
1
1
1
2
2
3
4
5
8
9
13
13
* Values are ppb
** CMSA = Consolidated Metropolitan Statistical Area
1 To be sure, the max 1-h concentration observed at some sites in and around some CMSAs
2 still exceeded the mean by a large margin, with max 1-h values of > 600 ppb. However, the 50th
3 percentile maximum value outside CMSAs, 5 ppb, was only slightly greater than the 1-h, 24-h,
4 and annual mean value, 4 ppb. The 50th percentile maximum value inside CMSAs, 7 ppb, was
5 75% greater than these longer-term averages, reflecting heterogeneity in source strength and
May 2008
2-22
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
location. In addition, even with 1-h max values of > 600 ppb, the maximum annualized mean
value for all CMSAs was still < 16 ppb, which is below the current annual primary S02 NAAQS.
800-1
700- x
600- x
X
S" 500 ~ x „
a i„n * X *
•= 400- X
°*XX * v X XX
w 300- x x x „xiS**x xx
liniiiiiilliliilniili
I I I I I I I I I I I I I I I I I 1 M [ 1 I M I I I I I I I I I I I r I I I 1 1 I I M I I
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Sample Hour
Figure 2-11. Boxplot of hourly S02 concentrations across all cities in focus.
The strong west-to-east increasing gradient in S02 emissions described above is well-
replicated in the observed concentrations in individual CMSAs. For example, Table 2-5 shows
the mean annual concentrations from 2003-2005 for the 12 CMSAs with four or more S02
regulatory monitors. Values ranged from a reported low of ~1 ppb in Riverside, CAand San
Francisco, CAto a high of-12 ppb in Pittsburgh, PA and Steubenville, OH, in the highest S02
source region.
The Pearson correlation coefficients (r) for multiple monitors in these CMSAs were
generally very low for all cities, especially at the lower end of the observed concentration ranges,
and even negative at the very lowest levels on the West Coast (see Table 2-5). This reflects
strong heterogeneity in S02 ambient concentrations even within any one CMSA and, therefore,
indicates possibly different exposures of spatially distinct subgroups of humans in these CMSAs
to these very low concentrations of S02. At higher concentrations, the r values were also higher.
In some CMSAs, this heterogeneity may result from meteorological effects, whereby a generally
well-mixed subsiding air mass containing one or more S02 plumes with relatively high
concentration would be more uniformly spread than faster-moving plumes with lower
concentrations. However, instrument error may also play a role, because the highest r values, i.e.,
those > 0.7, correspond to the highest S02 concentrations, i.e., > 6 and > 10 ppb. Since the lowest
S02 concentrations are at or below the operating LOD, and demonstrate the lowest correlation
May 2008
2-23
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
across monitors that share at least some air mass characteristics most of the year, the unbiased
instrument error in this range may be confounding interpretation of any possible correlation. This
could be because the same actual ambient value would be reported by different monitors (with
different error profiles) in the CMSA as different values in this lowest concentration range.
To better characterize the extent and spatiotemporal variance of S02 concentrations within
each of the CMSAs having four or more S02 monitors, the means, minima, and maxima were
computed from daily mean data across all available monitors for each month for the years 2003
through 2005. Because many of these CMSAs with S02 monitors also reported SO42 , it is
possible to compute the degree of correlation between S02, the emitted species, and SO42 , the
most prominent oxidized product from S02. SO42 values, however, while averaged over all
available data at each site are generally available at their monitoring sites on a schedule of only 1
in 3 days or 1 in 6 days. Furthermore, S02 and SO42 monitors are not all co-located throughout
the CMSAs. For each of the five example CMSAs in Figures 2-12 through 2-16, monthly
aggregated values are depicted from daily means of: (a) the monthly mean, minimum, and
maximum S02 concentrations; (b) the monthly mean, minimum and maximum SO42
concentrations; and (c) a scatterplot of S02 versus SO42 concentrations.
Table 2-5. Range of mean annual S02 concentrations and Pearson correlation coefficients in
urban areas having at least four regulatory monitors, 2003-2005.
CMSA (# MONITORS)
MEAN S02 CONCENTRATION (ppb)
PEARSON CORRELATION COEFFICIENT
Philadelphia, PA (10)
O)
10
1
CD
CO
0.37-0.84
Washington, DC (5)
3.2-6.5
0.30-0.68
Jacksonville, FL (5)
1.7-3.4
-0.03-0.51
Tampa, FL (8)
2.0-4.6
-0.02-0.18
Pittsburgh, PA (10)
6.8-12
0.07-0.77
Steubenville, OH (13)
8.6-14
0.11 -0.88
Chicago, IL (9)
2.4-6.7
0.04-0.45
Salt Lake City, UT (5)
2.2-4.1
LO
CM
O
1
O
O
Phoenix, AZ (4)
1.6-2.8
-0.01 -0.48
San Francisco, CA (7)
1.4-2.8
-0.03-0.60
Riverside, CA (4)
1.3-3.2
-0.06-0.15
Los Angeles, CA (5)
1.4-4.9
I
0
O)
1
0
CO
May 2008
2-24
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
In Steubenville, OH (Figure 2-12), the area of highest S02 concentrations of all 12 CMSAs
with more than four monitors, all monthly mean S02 concentrations (a) were substantially < 30
ppb, though max daily means in some months were often > 60 ppb, or even > 90 ppb. Sulfate
data (b) at Steubenville were insufficient to make meaningful comparisons, though the 12
months of available SO42 data suggest no correlation with S02 (c).
Next, consider Philadelphia, PA (Figure 2-13). S02 in Philadelphia, PA (a) is present at
roughly one-half the monthly mean concentrations in Steubenville, OH, and demonstrates a
strong seasonality with S02 concentrations peaking in winter. By contrast, SO42 concentrations
in Philadelphia peak in the three summer seasons, with pronounced wintertime minima. This
seasonal anticorrelation still contains considerable monthly scatter, however.
Los Angeles, CA (Figure 2-14) presents a special case, since its size and power
requirements place a larger number of S02 emitters near it than would otherwise be expected on
the West Coast. Concentrations of S02 demonstrate weak seasonality in these 3 years, with
summertime means of ~3 to 4 ppb, and maxima generally higher than wintertime ones, though
the highest means and maxima occur during the winter of 2004-2005. S042 at Los Angeles
shows stronger seasonality, most likely because the longer summer days of sunny weather allow
for additional oxidation of S02 to SO42 than would be available in winter. Weak seasonal effects
in S02 likely explain the complete lack of correlation between S02 and SO42 here.
The Riverside, CA CMSA (Figure 2-15) presents the strongest example among the 12
examined for this study of correlation between S02 and S042 , though even here the R2 value is
merely 0.3. Seasonal peaks are obvious in summertime for S02 and SO42 , both at roughly one-
half the ambient concentrations seen in Los Angeles. This is very likely due to Riverside's
geographic location just downwind of the regionally large electric generating utility sources near
Los Angeles and the prevailing westerly winds in summer. Again, as with Los Angeles, the
summertime peaks in SO42 are most likely due to the combination of peaking S02 and favorable
meteorological conditions allowing more complete oxidation.
Phoenix, AZ was the CMSA with the lowest monthly mean S02 and SO42 concentrations
examined here (Figure 2-16). In Phoenix, nearly all monthly mean S02 values were at or below
the regulatory monitors' operating LOD of ~3 ppb. SO42 concentrations were equivalently low,
roughly one-half the concentrations seen in Riverside, CA, for example. The monthly mean data
show strong summertime peaks for even these very low-level SO42 observations, which, at ~1 to
May 2008
2-25
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
3 |ig/m3, were generally one-half of those in Philadelphia. This suggests some seasonality in S02,
though anticorrelated with SO42 ; however, the trend is very weak, as the correlation scatterplot
shows.
2.4.5. 5-Minute Sample Data in the Monitoring Network
Although the number of monitors across the CONUS varies somewhat each year, in 2006
there were -500 S02 monitors in the NAAQS monitoring network (http://www.epa.gov/air/data).
The state and local agencies responsible for these monitors are required to report 1-h avg
concentrations to the EPA AQS. In addition, a very small number of sites—only 108 total from
1997 to 2006, and not the same sites in all years—voluntarily reported 5-min block avg to AQS.
Of these, 104 reported only the max 5-minute average, 15 reported all 12 5-minute avg in each
hour, and 11 of those 15 reported all 12 values each hour and maximum values for some fraction
of time between 1997 and 2006. See Table 2-6 and Table 2-7 for a breakdown of these monitor
locations and sampling periods, and Figure 2-17 for the geospatial distribution of these monitors
across the CONUS.
May 2008
2-26
DRAFT—DO NOT QUOTE OR CITE
-------
(a)120 ¦¦
100 ¦¦
ff 80 +
Q.
~ 60 +
o
(0
40 ¦
20
S$>
.S$> .S$>
,jS>
or „ol
cJ*
.$£ .5$
<£
# ^ of -f of -f ^ ^ of' ^
&
(b) 40 T
36
32 ;¦
_ 28 ¦¦
CO
E 24 ¦¦
S 20 - •
16 ••
12 ¦•
8
4 +
M
o
CO
I
if
I
f
& t&
fl0>
^ <£> of* sJ"
^ „<>' .«*>'
5^
gr
5? <£>
¦f ^ ^ of y^' -f ^ i> of ^ -f ^ i* of ^
& & &
^ J>' .^'
sP & & &
v _<>'
(c) 24 ••
20 ¦¦
#T 16 ¦¦
E
O)
3 12 +
CNJ
o
w 8 +
4 +
0
O
o
S02 (ppb)
10
15
Figure 2-12. Steubenville, OH, 2003-2005. (a) Monthly mean, minimum, and maximum S02
concentrations, (b) Monthly mean, minimum, and maximum S042" concentrations, (c) Monthly
mean S042- concentrations as a function of S02 concentrations.
May 2008
2-27
DRAFT—DO NOT QUOTE OR CITE
-------
(a) 50 "
45 --
40 --
35 -- t
o. 30 +
Q.
™ 25 +
8 20
15 4-
10
5
0
I I
s$> .s?5
j$>
&
J? J>
N,#' ^ ^ ^ ^ ^ ^ ^
(b)
E
O)
"¦t
O
to
70 "
60 •¦
50 --
40 --
30 --
20 --
10 - -t
I" f:: I.. If* I j f I "I I
ffkiHlii
nift*
f
# .s$>
<$>
S?
& & &
&
& <& &
.& .S§>
&
)f ^ ^ ^ of' ^ ^ ^ ^ ^ ^ ^ ^
&
(c) 25
20 ¦•
E 15 f
O)
10 --
o
CO
5 --
~
o
O O
Oo ° ^OO
o
o
oo
o
o^ .
O O<>
SOz (ppb)
10
12
Figure 2-13, Philadelphia, 2003-2005. (a) Monthly mean, minimum, and maximum S02
concentrations, (b) Monthly mean, minimum, and maximum S042- concentrations, (c) Monthly
mean S042- concentrations as a function of S02 concentrations.
May 2008
2-28
DRAFT—DO NOT QUOTE OR CITE
-------
(a)
16
14
12
3"
10
a
Q.
8
O
to
6
4
2
0
I
& .&
.c? _&
5^ 5^ & ,
.5?
.<£ .5?
S$»
.<£>
& &
&
d* & <&
.J$> j£ cj>
_J4
^ ^ ^ if ^ ^ ^ ^ ^ of ^
(c) 20 j
18 -¦
16 ¦•
14 --
J 12 -
3 10 ¦¦
8 -¦
6 --
4 --
2 --
0 --
o
w
o o
o <> o $
O Oo
o o
o %
o
O o
0
S02(ppb)
Figure 2-14. Los Angeles, 2003-2005. (a) Monthly mean, minimum, and maximum S02
concentrations, (b) Monthly mean, minimum, and maximum S042" concentrations, (c) Monthly
mean S042- concentrations as a function of S02 concentrations.
May 2008
2-29
DRAFT—DO NOT QUOTE OR CITE
-------
(a) 16-
14 --
12
~ 10 +
Q.
3 8 +
O
& Si1" & & c$" & & jv3 & $> & jy5 &
^ ^ ^ ^ ® ^ ^ ^ ^ e**5 ^
(b)
E
O)
O
to
18 T
16 --
14 --
12
10 --
8 --¦
6 --
4
2
0
U-I
^il
il
Ul
5$> .J?
.5? .5?
c? s? .s^
pfc ~jx
^ s£ .<£
.S$> <£
^ «^' ^ ^ 9^ ^ ^ ^ ^ ^ -f' ^ ^ # cJ$ ^
&
(c) 10;-
9 ¦•
8 ;-
7 ~
c*>
6 ••
5 '"
4 --
3 -¦
2 -¦
1 -¦
0 •¦
o>
3.
o
w
oo
o
o o £ ~
o V o <$>
o o
o°
S02(ppb>
Figure 2-15. Riverside, CA, 2003-2005. (a) Monthly mean, minimum, and maximum S02
concentrations, (b) Monthly mean, minimum, and maximum S042" concentrations, (c) Monthly
mean S042- concentrations as a function of S02 concentrations.
May 2008
2-30
DRAFT—DO NOT QUOTE OR CITE
-------
(a) 10 j
8 ¦¦
n
a.
o.
O
in
6 ¦¦
4 -•
2 ¦¦
ll
ill
iif
ii
f
& .S? J? J? ^ ,5^ £>*" .£?* p1* & & d3 J> & &
^ ^ ^ ^ .J? ,<$> jS> .J? c5* .cJ* # & ,& d3 & & ^$P cy3 ,cy>
^ #V ^ ^ ^ ^ ^ ^ ^ ^
o
o
~
V
O V o
~
o^
0,0 0.5 1.0 1.5 2.0 2.5 3.0
S02 (ppb)
3.5
4.0 4.5
5.0
Figure 2-16. Phoenix, 2003-2005. (a) Monthly mean, minimum, and maximum S02 concentrations,
(b) Monthly mean, minimum, and maximum S042- concentrations, (c) Monthly mean S042-
concentrations as a function of S02 concentrations.
May 2008
2-31
DRAFT—DO NOT QUOTE OR CITE
-------
Although these 5-minute data meet AQS minimum quality assurance requirements, the
voluntary nature of this reporting and the high variability across space and time make these data
very difficult to use precisely.
Table 2-6. Locations, counts, and sampling periods of monitors reporting 5-minute maximum
S02 values, 1997-2006.
STATE
NUMBER OF
COUNTIES
NUMBER OF
MONITORS
NUMBER OF YEARS
YEARS OPERATING
Arkansas
2
3
10
1997-2006
Colorado
1
1
10
1997-2006
Delaware
1
1
2
1997-1998
D.C.
1
1
5
2000-2004
Iowa
6
9
5
2001-2005
Louisiana
1
1
4
1997-2000
Missouri
7
14
10
1997-2006
Montana
1
7
10
1997-2006
North Carolina
1
1
8
1997-2004
North Dakota
11
19
10
1997-2006
Pennsylvania
8
23
7
1997-2003
Table 2-7. Locations, counts, and sampling periods of monitors reporting all 12 5-minute S02
values in each hour, 1997-2006.
STATE
NUMBER OF
COUNTIES
NUMBER OF
MONITORS
NUMBER OF YEARS
YEARS OPERATING
D.C.
1
1
1
2007
Florida
1
1
4
2002-2005
Missouri
1
2
4
2003-2006
Montana
1
4
1
2002
North Carolina
1
1
4
1999-2002
Pennsylvania
2
5
5
2002-2006
West Virginia
2
2
5
2001-2005
May 2008
2-32
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Ambient Monitors
Counties
5 Minute S02
POLL
El all 5 minute vals
¦ max 5 minute val
Figure 2-17. S02 monitors reporting maximum or continuous 5-minute average values for any
period, 1997-2006.
2.4.6. Policy Relevant Background Contributions to so2 Concentrations
Background concentrations used for purposes of informing decisions about the NAAQS
are referred to as Policy Relevant Background (PRB) concentrations. PRB concentrations are
those concentrations that would occur in the United States in the absence of anthropogenic
emissions in continental North America (defined here as the United States, Canada, and Mexico).
PRB concentrations include contributions from natural sources everywhere in the world, and
from anthropogenic sources outside these three countries. Background levels so defined facilitate
separation of cases where pollution levels can be controlled by U.S. regulations (or through
international agreements with neighboring countries), from cases where pollution is generally
uncontrollable by the United States. EPA assesses risks to human health and environmental
effects from S02 levels in excess of PRB concentrations.
Contributions to PRB concentrations include natural emissions of S02 and photochemical
reactions involving reduced sulfur compounds of natural origin, as well as their long-range
transport from outside of North America from any source. As an example, transport of S02 from
Eurasia across the Pacific Ocean or the Arctic Ocean would carry PRB S02 into the U.S.
May 2008
2-33
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Annex B contains a schematic diagram showing the major photochemical processes involved in
the sulfur cycle, including natural sources of reduced sulfur species from anaerobic microbial
activity in wetlands and volcanic activity. Volcanoes and wildfires are the major natural source of
S02. Biogenic emissions from agricultural activities are not considered in the formation of PRB
concentrations. Discussions of the sources and estimates of emissions are given in Annex
Section B.6.
The MOZART-2 global model of tropospheric chemistry (Horowitz et al., 2003) is used to
estimate the PRB contribution to S02 concentrations. The model setup for the present-day
simulation, i.e., including all sources in the U.S. Canada and Mexico, was published in a series
of papers from a recent model intercomparison (Dentener et al., 2006; van Noije et al., 2006).
MOZART-2 is driven by the National Oceanic and Atmospheric Administration's National
Center for Environmental Prediction (NOAA/NCEP) meteorological fields and the International
Institute for Applied Systems Analysis (IIASA) 2000 emissions at a resolution of 1.9° x 1.9°
with 28 g (sigma) levels in the vertical, and includes gas- and aerosol-phase chemistry. Results
shown in Figure 2-18 are for the meteorological year 2001. An additional PRB simulation was
conducted in which continental North American anthropogenic emissions were set to zero.
The role of PRB in contributing to S02 concentrations in surface air is examined first.
Figure 2-18 shows the annual mean predicted S02 concentrations in surface air in the simulation
including all sources, or the "base case" (top panel); the PRB simulation (middle panel); and the
percentage contribution of the background to the total base case S02 (bottom panel). Maximum
concentrations in the base case simulation, > 5 ppb, occur along the Ohio River Valley (upper
panel). Background S02 concentrations are orders of magnitude smaller, below 10 parts per
trillion (ppt) over much of the United States (middle panel). Maximum PRB concentrations of
S02 are 30 ppt. In the Northwest where there are geothermal sources of S02, the contribution of
PRB to total S02 is 70 to 80%; however absolute S02 concentrations are still of the order of a
couple of ppb or less. With the exception of the West Coast where volcanic S02 emissions cause
high PRB concentrations, PRB contributes < 1% to present-day S02 concentrations in surface air
(bottom panel).
May 2008
2-34
DRAFT—DO NOT QUOTE OR CITE
-------
Total
50°N
45°N
40°N
35°N
30°N
25° N
iao°w ioo°w ao°w
< am 1.21 *41 3.60 4.80 6*00 ppb
Background
50°N
45°N
40°N
35°N
30°N
25"N
120°W 100°W ao°w
< O.o'ST-TSSr aoTT 0.015 002^™^25 ppb
Percent Background Contribution
50"N
45°N
40°N
3S°N
30"N
25°N
12o°w ioo°w ao°w
< 1 5 10 15 20 25
Figure 2-18. Annual mean model-predicted concentrations of S02 (ppb).
1 When estimating background concentrations it is instructive to consider measurements of
2 S02 at relatively remote monitoring sites, i.e., sites located in sparsely populated areas not
3 subject to obvious local sources of pollution. Berresheim et al. (1993) used a type of atmospheric
4 pressure ionization mass spectrometer (APIMS) at Cheeka Peak, WA (48.30°N 124.62°W,
5 480 m asl), in April 1991 during a field study for DMS oxidation products. S02 concentrations
6 ranged between 20 and 40 ppt. Thornton et al. (2002) have also used an APIMS with an
jfe/
hm
May 2008
2-35
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
isotopically labeled internal standard to determine background S02 levels. S02 concentrations of
25 to 40 ppt were observed in northwestern Nebraska in October, 1999 at 150 m above ground
using the National Center for Atmospheric Research (NCAR)'s C-130 research aircraft. These
data are comparable to remote central South Pacific convective boundary layer S02 data
(Thornton, 1999).
Figure 2-19. 15-minute average ambient S02 concentrations measured at Hawaii Volcanoes
National Park monitoring sites, March 12,13, and 15, 2007.
Source: National Park Service
As noted earlier, volcanic sources of S02 in the United States are found in the Pacific
Northwest, Alaska, and Hawaii. The greatest potential domestic effects from volcanic S02 occurs
on the island of Hawaii. Nearly continuous venting of S02 from Mauna Loa and Kilauea
produces S02 in high concentrations (see Figure 2-19 and Figure 2-20) at two National Park sites
near the Kilauea caldera and the nearby east rift zone. The latter emits several times as much S02
as the Kilauea caldera. The two measurement sites within the National Park are < 3 km from the
summit emission source and -10 km from the east rift source and are affected by the two sources
during southerly and easterly winds. A number of communities and population centers are within
the same distance from the east rift gas source that affects these two monitoring sites. When the
May 2008 2-36 DRAFT—DO NOT QUOTE OR CITE
Ambient SO, - Jaggar Museum - 15 minute averages
March 12
March 13
March 15
00:00 04:00 08:00 1 2:00 1 6:00 20:00 24:00
Hour
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
normal trade wind flows are disrupted, emissions from the sources can be brought directly to
these various communities. Since these communities are located at a similar distance from the
large east rift emission source as the Park monitoring stations, it is probable that these
communities experience S02 concentrations as high as those measured within Hawaii Volcanoes
National Park.
¦Q
Q_
CL
5000
4000 -
3000 -
2000 -
Sept 29 2007 -15-minute averages
8 1000 -
o -
Kila uea Visitor Center
— ¦ — Ja ggar Museum
L
- ... jj\i
1 1 1 1 1 1 1 1 1 1 1
0 200 400 BOO BOO 1000 1 200 1400 1 BOO 1 BOO 2000 2200 2400
Time
Figure 2-20. 15-minute average ambient S02 concentrations measured at the two National Park
monitoring sites at Hawaii Volcanoes NP, Hawaii on September 29, 2007.
Source: National Park Service
Since 1980, the Mount St. Helens volcano (46.20°N, 122.18°W, summit 2549 m asl) in the
Washington Cascade range has been a variable source of S02. Its major effects came in the
explosive eruptions of 1980, which primarily affected the northwestern United States. The
Augustine volcano near the mouth of the Cook Inlet in southwestern Alaska (59.363 °N,
153.43 °W, summit 1252 m asl) has emitted variable quantities of S02 since its last major
eruptions in 1986. Volcanoes in the Kamchatka peninsula in far eastern Siberia do not
particularly affect the surface concentrations in northwestern North America.
Overall, the background contribution to S02 over the United States is relatively small, with
a max PRB of 0.030 ppb S02, except for areas with volcanic activity.
May 2008
2-37
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
2.5. Issues Associated with Evaluating so2 Exposure
2.5.1. General Considerations for Personal Exposure
Human exposure to an airborne pollutant consists of contact between the human and the
pollutant at a specific concentration for a specified period of time. People spend various amounts
of time in different microenvironments characterized by different pollutant concentrations. The
integrated exposure of a person to a given pollutant is the sum of the exposures over all time
intervals for all microenvironments. Figure 2-21 represents a composite average of activity
patterns across all age groups in the United States, based on data collected in the National
Human Activity Pattern Survey (NHAPS) (Klepeis et al., 2001). The demographic distribution of
the respondents was designed to be similar to that of overall U.S. Census data. Different cohorts,
e.g., the elderly, young and middle-aged working adults, and children exhibit different activity
patterns.
A person's exposure to a pollutant, such as S02, can be represented by:
n
ET = ICiti
i=l (2-7)
where ET is an individual's total personal exposure for a specific time period, n is the total
number of microenvironments encountered, C, is the average concentration, and l, is the time
spent in the ith microenvironment. The exposure a person experiences can be characterized as: an
instantaneous exposure; a peak exposure such as might occur during cooking; an average
exposure; or an integrated exposure over all environments encountered. These distinctions are
important because health effects caused by long-term low-level exposures may differ from those
caused by short-term peak exposures.
An individual's total exposure (ET) can also be represented by:
E'T ~ Ea Ena ~ \J;o ^ 3;i \f>i @i/(ai + ^i)j } ^na ~ f ^ J'i ^inf} ^- a ^na
' ' (2-8)
subject to the constraint
y0 + Zj'i = 1
i (2-9)
May 2008
2-38
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
where Ea is the ambient component of personal exposure, Ena is the nonambient component of
personal exposure, y0 is the fraction of time spent outdoors, and is the fraction of time spent in
microenvironment i. , P,, at, and k, are the infiltration factor, penetration coefficient, air
exchange rate, and decay rate, respectively for microenvironment /. In the case where an
exposure occurs mainly in one microenvironment, Equation 2-8 may be approximated by
Equation 2-10 where y is the fraction of time spent outdoors, and a is the ratio of personal
exposure from a pollutant of ambient origin to the pollutant's ambient concentration (or the
ambient exposure factor). Other symbols have the same definitions as in Equations 2-8 and 2-9.
Et= Ea + E„a = {y + {l-y)[Pa/(a + k)]}Ca + Ena = a Ca + E„a (2_1Q)
If concentrations in a single microenvironment are considered, then Equation 2-10 can be recast
as
Cme = Ca+ C„a = [Pa /(a + k)]Ca + S/[V{a + *)] ^ 1}
where Cme is the concentration in a microenvironment, Ca and Cna are the contributions to Cme
from ambient and nonambient sources, S is the microenvironmental source strength, and V is the
volume of the microenvironment. (Bracketed symbols are same as Equation 2-8.) In this
equation, it is assumed that microenvironments do not exchange air with each other, but only
with the ambient environment.
Microenvironments in which people are exposed to air pollutants such as S02 typically
include residential indoor environments, other indoor locations, near-traffic outdoor
environments, other outdoor locations, and in vehicles, as shown in Figure 2-21. Indoor
combustion sources such as gas stoves and space heaters need to be considered when evaluating
exposures to S02. Exposure misclassification may result when total human exposure is not
disaggregated between various microenvironments, and this may obscure the true relationship
between ambient air pollutant exposures and health outcomes.
May 2008
2-39
DRAFT—DO NOT QUOTE OR CITE
-------
NHAPS - Nation, Percentage Time Spent
Total n = 9, 196
IN A RESIDENCE (68.7%)
OFFICE-FACTORY (5.4%)
"I TOTAL TIME SPENT
I 1 INDOORS (86.9%)
' OUTDOORS (7.6%)
IN A VEHICLE (5.5%)
OTHER INDOOR LOCATION (11%)
BAR-RESTAURANT (1.8%)
Figure 2-21. Percentage of time spent in various environments in the United States.1
Source: Klepeis et al. (2001).
1 In a given microenvironment, the ambient component of a person's microenvironmental
2 exposure to a pollutant is determined by the following physical factors:
3 ¦ ambient concentration Ca
4 ¦ air exchange rate a,
5 ¦ pollutant specific penetration coefficient /J,
6 ¦ pollutant specific decay rate kt
7 ¦ fraction of time an individual spends in the microenvironment yt
8 These factors are in turn affected by the following exposure factors:
9 ¦ environmental conditions, such as weather and season
1 For example, the cohort of working adults between the ages of 18 and 65 represents ~50% of the population. Of this total, about 60% work outside
the home, spending ~24% (40 h/168 h) of their time in factory/office environments. Thus, this cohort is likely to spend considerably more time in
offices and factories than shown in the figure (5.4%), which reflects the entire population, and is also likely to spend less time in a residence than
small children or the elderly.
May 2008
2-40
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
¦ dwelling conditions, such as the location of the house which determines proximity to
sources and geographical features that can modify transport from sources, the amount of
natural ventilation (e.g., open windows and doors, and the "draftiness" of the dwelling)
and ventilation system (e.g., removal efficiency and operation cycle)
¦ personal activities (e.g., the time spent cooking or commuting)
¦ indoor sources and sinks of a pollutant
Microenvironmental exposures can also be influenced by the individual-specific factors such as
age, gender, health or socioeconomic status.
Time-activity diaries, completed by study participants, are used to compile activity patterns
for input to exposure models and assessments. The EPA's National Exposure Research
Laboratory (NERL) has consolidated the majority of the most significant human activity
databases into one comprehensive database called the Consolidated Human Activity Database
(CHAD). Eleven different human activity pattern studies were evaluated to obtain over 22,000
person-days of 24-h human activities in CHAD (McCurdy et al., 2000). These data can be useful
in assembling population cohorts to be used in exposure modeling and analysis.
In general, the relationship between personal exposures and ambient concentrations can be
modified by microenvironments. During infiltration, ambient pollutants can be lost through
chemical and physical loss processes, and therefore, the ambient component of a pollutant's
concentration in a microenvironment is not the same as its ambient concentration but the product
of the ambient concentration and the infiltration factor (¥ inf or a if people spend 100% of their
time indoors). In addition, exposure to nonambient, microenvironmental sources modifies the
relationship between personal exposures and ambient concentrations.
In practice, it is extremely difficult to characterize community exposures by measurements
of each individual's personal exposures. Instead, the distribution of personal exposures in a
community, or the population exposure, is simulated by extrapolating measurements of personal
exposure using various techniques or by stochastic, deterministic or hybrid exposure modeling
approaches such as APEX, SHEDS, and MENTOR (see Annex Section C.2 for a description of
modeling methods). Variations in community-level personal exposures are determined by cross-
community variations in ambient pollutant concentrations and the physical and exposure factors
mentioned above. These factors also determine the strength of the association between
population exposure to S02 of ambient origin and ambient S02 concentrations.
May 2008
2-41
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Of major concern is the ability of S02, measured by ambient monitors, to serve as a
reliable indicator of personal exposure to S02 of ambient origin. The key question is what errors
are associated with using S02 measured by ambient monitors as a surrogate for personal
exposure to ambient S02 and/or its oxidation products in epidemiological studies. There are three
aspects to this issue: (1) ambient and personal sampling issues; (2) the spatial variability of
ambient S02 concentrations; (3) the associations between ambient concentrations and personal
exposures as influenced by exposure factors, e.g., indoor sources and time spent indoors and
outdoors. Items (1) and (3) are treated individually in the following sections; item (2) was treated
previously in Section 2.4.2.
2.5.2. Methods Used for Monitoring Personal Exposure
Three basic methods of analysis have been used as personal exposure monitors (PEMs) to
measure personal exposure to S02. The Harvard-EPA annular denuder system (HEADS) was
initially developed to measure particles and acid gases simultaneously (Brauer et al., 1999;
Koutrakis et al., 1988). The aerosol is initially sampled at 10 L/min through an impactor that is
attached to an annular denuder to remove particles. Subsequently, the aerosol is sampled through
an annular denuder coated with sodium carbonate (Na2C03). This denuder is used to trap S02,
nitric acid (HNO3), and nitrous acid (HNO2). Following sampling, the denuder is extracted with
ultrapure water and analyzed by ion chromatography. Collection efficiencies of S02 in the
denuder are typically around 0.993, which compares well with predicted values.
For a study conducted in Baltimore, MD, Chang et al. (2000) developed and employed a
personal roll-around system (RAS, an active sampling system designed to measure short-term
exposure) to measure personal exposure concentrations of several atmospherically relevant
species, including S02. For the measurement of S02, the RAS employed an NCVSC^ sorbent
denuder worn on a vest by the study participant. The hollow glass denuder, encased in an
aluminum jacket, is coated with triethanolamine (TEA) for the collection of S02 and NO2, and
aerosol is sampled through the denuder at 100 cc/min. Following sampling, the denuder can be
extracted and analyzed for S02 concentrations by ion chromatography. The detection limit for
1-h sampling of S02 was reported to be 62 ppb, which resulted in many of the 1-h samples being
below the LOD.
May 2008
2-42
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
The most commonly employed S02 PEM method for personal exposure studies is the
passive badge sampler. A personal multipollutant sampler has been developed to measure
particulate and gaseous pollutants simultaneously (Demokritou et al., 2001). A single elutriator,
operating at 5.2 L/min, is employed to sample particulate pollutants. A passive S02 badge is
attached diametrically to the elutriator, which has been coated with Teflon to minimize reactive
gas losses. The passive badge sample is coated with TEA for the collection of S02 and NO2.
Because wind speed can affect the collection rate of the passive badge sampler, this system
employs a constant face velocity across the passive badge sampler. For 24-h sampling times, the
estimated limit of detection (LOD) for S02 is 5 ppb.
Currently, limits exist for using PEM systems to measure personal exposure to S02.
Because S02 concentrations have been declining annually in the United States, little focus has
been placed on improving the methods of analysis. LODs for S02 PEMs (-5-10 ppb for 24 hr
sampling) are often greater than the concentrations of S02 that are typically observed in urban
ambient environments. However, much lower detection limits can be achieved by extending the
sampling time (Kasper-Giebl et al., 1999). Personal exposure monitoring studies often suffer
from having many of the daily S02 samples (e.g., 30 to 70%) collected below the sampler's LOD
(see Tables 2-10 and 2-11). Because of these issues, current methods can not characterize hourly
or shorter exposures unless these values are in the range of several tens to hundreds of ppb.
2.5.3. Relationship between Personal Exposure and Ambient
Concentration
Because S02 concentrations have declined markedly over the past few decades, relatively
few studies have focused on S02. Another consideration is that current indoor and outdoor levels
in many areas are often beneath detection limits for passive personal S02 monitors.
2.5.3.1. Indoor Versus Outdoor so2 Concentrations
Several studies in the United States, Canada, Europe, and Asia have examined the
relationships of indoor, outdoor, and personal concentrations of S02 to ambient S02
concentrations. Perhaps the most comprehensive set of indoor-outdoor data was obtained by
Spengler et al. (1979) during the Harvard Six Cities Study. These data are shown in Figure 2-22.
May 2008
2-43
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Twenty-four-hour ambient and indoor S02 concentrations were measured every sixth day for
1 year in a minimum of 10 homes or public facilities for each of the cities studied.
W
E
40
20
30
¦ Outdoor
50 - Indoor
10
0
PORT TOPE KING WAT STL STEU
Figure 2-22. Average annual indoor and outdoor S02 concentrations for each of the six cities
included in the analysis.
Source: Adapted from Spengler et al. (1979).
As can be seen from Table 2-8, a wide range is found in the ratio of indoor to outdoor
concentrations among the different studies. These differences among studies could be due in part
to differences in building characteristics (e.g., residences versus schools or other public
buildings), in activities affecting air exchange rates, and in analytical capabilities. In several
studies, high values for R2 were found, suggesting that indoor levels were largely driven by
outdoor levels. A few studies found higher levels of S02 indoors than outdoors in some samples.
This situation could have arisen if there were indoor sources or because of analytical
measurement issues. One would expect to find lower concentrations indoors than outdoors,
because S02 is consumed by reactions on indoor surfaces, especially those that are moist. Chao
acknowledged this point but could not account for the findings of this study. It was noted that
May 2008 2-44 DRAFT—DO NOT QUOTE OR CITE
PORT = Portage, Wl
WAT = Watertown, MA
TOPE = Topeka, KS
STL = St. Louis, MO
KING = Kingston, TN
STEU = Steubenville, OH.
-------
1 two samples had unusually high indoor to outdoor ratios and that the mean ratios would have
2 been much lower otherwise. Winter-summer differences in the indoor:outdoor ratio are
3 consistent with seasonal differences in air exchange rates, as noted by Brauer et al. (1991).
Table 2-8. Relationships of indoor to outdoor S02 concentrations.
REFERENCE
LOCATION
INDOOR TO OUTDOOR RATIO
V(# SAMPLES)
NOTES
Spengler et al. (1979)
Portage, Wl
0.67 (349)
One year during Harvard Six Cities Study.
West-Gaeke method.
Topeka, KS
0.50 (389)
Kingston, TN
0.08 (425)
Watertown, MA
0.33 (486)
St. Louis. MO
0.31 (543)
Steubenville, OH
0.39 (499)
Stock et al. (1985)
Houston, TX
0.54 (2425)
May to October, continuous FRM for indoor
and outdoor.
Meranger and Brule (1987)
Antigonish, NS, Canada
0.84 (8)
Early spring, 1 wk avg in 1 house with oil
furnace, FPD-TA
Brauer et al. (1989)
Boston, MA
0.23 (24)
Summer, HEADS
Li and Harrison (1990)
Essex, UK
0.22
Summer
Brauer et al. (1991)
Boston, MA
0.39 (geom. mean)
(29), R2 = 0.89
Summer, HEADS
0.05 (geom. mean)
(23), R2 = 0.73
Winter, HEADS
Chanet al. (1994)
Taipei, Taiwan
0.24 (15)
Summer, PS
0.23 (37)
Winter, PS
Lee et al. (1999)
Hong Kong
0.92, R2 = 0.56
Winter, PF
Patterson and Eatough
(2000)
Lindon, UT
0.027 ± 0.0023, R2 = 0.73
Winter, ADS, all samples
Kindzierski and Sembaluk
(2001)
Boyle, Alberta, Canada
0.12 (12)
Late Fall, PS
Sherwood Park, Alberta,
Canada
0.14 (13)
Chao (2001)
Hong Kong
1.01 ±0.78 (10)
Summer. Windows mainly kept closed, PS
Kindzierski and Ranganathan
(2006)
Fort McKay, Alberta,
Canada
0.35 (30)
Fall. All indoor levels < LOD and set =1/2
LOD, PS
FPD-TA = Flame Photometric Detection-Thermal Analysis PF = pulsed fluorescence
HEADS = Harvard-EPA Annular Denuder System FRM = Federal Reference Method
PS = passive sampler ADS = Annular Denuder System
4 Indoor, or nonambient, sources of S02 could complicate the interpretation of associations
5 between personal exposure to ambient S02 in exposure studies. Possible sources of indoor S02
6 are associated with the use of sulfur-containing fuels, with higher levels expected when
7 emissions are poorly vented. Brauer et al. (2002) noted that only one study (Biersteker et al.,
8 1965) conducted inferential analyses of potential determinants of exposure to indoor S02 levels.
May 2008
2-45
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
In the Biersteker et al. study, conducted in the Netherlands, indoor levels increased with oil, coal,
and gas heating, as well as smoking in homes and increased outdoor levels.
Triche et al. (2005) measured S02 levels in homes in which secondary heating sources
(fireplaces, kerosene heaters, gas space heaters, and wood stoves) were used. They found
elevated indoor levels of S02 when kerosene heaters were in use. Median levels of S02 when
kerosene heaters were used (6.4 ppb) were much higher than when they were not in use
(0.22 ppb). The maximum S02 level associated with kerosene heater use was 90.5 ppb. They did
not find elevated S02 levels when the other secondary heating sources were in use.
2.5.3.2. Relationship of Personal Exposure to Ambient Concentration
A few studies evaluated the association of personal exposure to S02 to ambient
concentrations (Brauer et al., 1989; Chang et al., 2000; Sarnat et al., 2000; 2001; 2005; 2006).
Some of these studies fall under the umbrella of the Health Effects Institute's Characterization of
Particulate and Gas Exposures of Sensitive Subpopulations Living in Baltimore and Boston
research plan (Koutrakis et al., 2005). However, the focus of many of these studies has been
exposure to particles, with acid gases included to evaluate confounder or surrogate issues.
Table 2-9 summarizes the longitudinal correlation coefficients between personal S02
exposures and ambient concentrations of S02, and Table 2-10 the pooled correlation coefficients.
Most of the studies examined lack the ability to quantify 24-h averaged personal S02 exposures
due to the low ambient S02 concentrations and the limitations of passive sampling, except two
studies conducted by Brauer et al. (1989) and Sarnat et al (2006).
Brauer et al. (1989) determined the slope of the regression line between personal and
ambient concentrations to be 0.13 ± 0.02, R2 = 0.43, based on 44 measurements made in Boston,
MA during the summer of 1988. Most if not all of the data points obtained using the HEADS
appeared to be above the working detection limits as defined by the authors in their publications
(Brauer et al., 1989; Koutrakis et al., 1988). Note that calculating detection limits in this way
could result in lower detection limits than if field blanks were used. The authors reported
significance at the p < 0.001 level, but the intercept was not significant at the p < 0.001 level.
Since the stationary monitoring site was located at an elevation of 250 m above street level, the
use of data from this ambient monitoring site will overestimate personal exposure, as the
concentration of S02 increases with height because it is emitted mainly by elevated point
May 2008
2-46
DRAFT—DO NOT QUOTE OR CITE
-------
sources. Indeed, the ambient concentrations are about a factor of two higher than the outdoor
concentrations. Sarnat et al. (2006) reported that ambient S02 was observed to be significantly
associated with personal S02 exposures during the fall (slope = 0.08 for overall population) in a
study in Steubenville, OH. The authors also observed the effect of ventilation on the association
between personal exposures and ambient concentrations (slope = 0.07 for subjects in buildings
with low ventilation rates, and 0.13 for subjects in buildings with high ventilation rates).
Table 2-9. Association between personal exposure and ambient concentration (longitudinal
correlations coefficients).
REFERENCE
STUDY DESIGN
SEASON
MEAN
CONC.
(PPb)
SLOPE
INTERCEPT
r, R2
COMMENTS
Sarnat et al.
(2000)
Longitudinal, Baltimore, 20 senior,
healthy, nonsmoking people (average
age 75), summer of 1998 and winter
of 1999, 1 day averaged sample, for
12 consecutive days for each subject;
four to six subjects were measured
concurrently during each 12-day
monitoring period.
Winter
Ambient:
6.6-10.2
Personal:
-0.8-1.2
NR
NR
-0.75 to
0.65 (r)
with a
median of
0.02 (14
subjects)
The LOD for 24-h
sampling was 6.5
ppb. All personal
samples were below
LOD.
Sarnat et al.
(2001)
Longitudinal, Baltimore, 56 seniors,
schoolchildren, and people with
COPD, summer of 1998 and winter of
1999, 14 of 56 subjects participated in
both sampling seasons; all subjects
were monitored for 12 consecutive
days (24-h avg samples) in each of
the one or two seasons, with the
exception of children who were
measured for 8 consecutive days
during the summer.
Winter
Ambient:
4-17
Personal:
-2-3
-0.05*
(N = 487
with 45
subjects)
0.54*
(N = 487 with
45 subjects)
-0.75 to 0.6
(r) with a
median of -
0.1
(44 sub-
jects)
1) Concentrations
are estimated from
Figure 1 in the paper.
2) Correlation coeffi-
cients are estimated
from Figure 2 in the
paper.
3) LOD was referred
to Sarnat et al
(2000), which was
6.5 ppb. Therefore,
all personal samples
were below LOD.
Sarnat et al.
(2005)
Longitudinal, Boston, 43 seniors and
schoolchildren, summer of 1999 and
winter of 2000, Similar study design
as Sarnat et al. (2001).
Summer
Ambient:
2.8-4.5
Personal:
0.3-0.5
0.00
(N = 335)
NR
-0.60 to
0.70 (r)
with a
median of
0.00
(Sample
size NR)
1) Correlation coeffi-
cients are estimated
from Figure 1 in the
paper.
2) LOD was 2.3 ppb,
and 96.5% of per-
sonal samples were
below LOD.
Winter
Ambient:
4.9-10.7
Personal:
-0.3-1.9
-0.02
(N = 299)
NR
-0.55 to
0.60 (r)
with a
median of
0.10
(Sample
size NR)
1) Correlation coeffi-
cients are estimated
from Figure 1 in the
paper.
2) LOD was 3.2 ppb,
and 95.4% of per-
sonal samples were
below LOD.
* significant at a = 0.05 level
7 The associations between personal exposure and ambient concentration cannot be exam-
8 ined in the other studies because almost all the personal exposure concentrations were beneath
May 2008 2-47 DRAFT—DO NOT QUOTE OR CITE
1
2
3
4
5
6
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
detection limits. For example, Chang et al. (2000) tested a new personal active sampling device
(a RAS with a TEA-based denuder) on volunteer participants to measure hourly personal
exposure to S02. However, the method detection limit was too high for S02 (62 ppb for 1-h
sampling) to generate a robust S02 exposure dataset to perform further analysis, and so the
authors did not use the S02 data.
Table 2-10. Association between personal exposure and ambient concentration (pooled
correlations coefficients).
REFERENCE
STUDY DESIGN
SEASON
MEAN
CONC.
(PPb)
SLOPE
INTERCEPT
r, R2
COMMENTS
Brauer et al.
(1989)
Pooled, Boston, study population was NR,
the number of participants was estimated
to be 48, July and August of 1988 for 24
days, 1 day averaged sample, two
subjects were monitored each day.
Summer
Ambient:
2.5-9.5
Personal:
0.4-1.8
0.13*
(N = 44)
Not significant
0.43
(R2)
1) Concentrations estimated
from Figure 2 in the paper.
2) Central site monitor was
250 m above the ground level.
3) LOD for personal samples
was -0.19 ppb based on the
way to determine the LOD for
an active sampling system.
Sarnat et al.
(2006)
Steubenville, 15 senior subjects, summer
and fall of 2000, two consecutive 24-h
samples were collected for each subject
for each wk, 23 wks total. Correlation
coefficients were calculated in the pooled
data set.
Summer
Ambient:
2.7 ±3.9
Personal:
1.5 ±3.3
0.03
(N =
106)
NR
0.00
(R2)
LOD was 5.5 ppb; 53.5% of
personal samples were below
LOD.
Fall
Ambient:
5.4 ±9.6
Personal:
0.7 ± 1.9
0.08*
(N =
152)
NR
0.15
(R2)
LOD was 3.8 ppb, and 31.6%
of personal samples were
below LOD.
* significant at a = 0.05 level
In the context of determining the effects of ambient pollutants on human health, the
association between the ambient component of personal exposures and ambient concentrations is
more relevant than the association between personal total exposures (ambient component +
nonambient component) and ambient concentrations. As described in Equations 2-8 and 2-10,
personal total exposure can be decomposed into two parts; an ambient and a nonambient
component. Usually, the ambient component of personal exposure is not directly measureable,
but it can be estimated by exposure models, or the personal total exposure can be regarded as the
personal exposure of ambient origin if there are no indoor or nonambient sources. It is expected
that the association between ambient concentrations and the ambient component of personal
exposures would be stronger than the association between ambient concentrations and personal
total exposures as long as the ambient and nonambient component of personal total exposure are
independent. None of the studies examined indoor sources, however, indoor sources are not
May 2008
2-48
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
expected to be present. The correlation coefficients between personal ambient S02 exposures and
ambient S02 concentrations in different types of exposure studies are relevant to different types
of epidemiologic studies.
There are three types of correlations generated from different study designs and ways to
analyze the data from exposure studies: longitudinal, "pooled," and daily-average correlations
(EPA, 2004). Longitudinal correlations1 are calculated when data from a study includes
measurements over multiple days for each subject (longitudinal study design). Longitudinal
correlations describe the temporal relationship between daily personal S02 exposure or
microenvironment concentration and daily ambient S02 concentration for the same subject. The
longitudinal correlation coefficient can differ between subjects (i.e., each person may have a
different correlation coefficient). The distribution of correlations for each subject across a
population could be obtained with this type of data (e.g., Sarnat et al., 2000; 2001; 2005). A
longitudinal correlation coefficient between the ambient component of personal exposures and
ambient concentrations is relevant to the panel epidemiological study design. In Table 2-9, most
longitudinal studies reported the association between personal total exposures and ambient
concentrations for each subject; for some subjects the associations were strong and for some
subjects the associations were weak. The weak personal and ambient associations do not
necessarily mean that ambient concentrations are not a good surrogate for personal exposures,
because the weak associations could have resulted from the day-to-day variation in the
nonambient component of total personal exposure. The type of correlation analysis can have a
substantial effect on the value of the resultant correlation coefficient.
Mage (1999) showed that very low correlations between personal exposure and ambient
concentrations could be obtained when people with very different nonambient exposures are
pooled, even though their individual longitudinal correlations are high.
Xf-v v)f" ")
{n-\)sxsa
where "r" is the longitudinal correlation coefficient between personal exposure and ambient concentration, "a" represents
the ambient concentration, "x" represents exposure, "i" represents the ith subject, "j" represents the jth measurement (with the averaging time
ranging from two days to two weeks for S02 measurement), "s" represents the standard deviation, and "n" in the longitudinal studies is the number
of measurements for each subject. The ambient concentration aj could be measured by one ambient monitor or the average of several ambient
monitors.
May 2008
2-49
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Pooled correlations1 are calculated when a study involves one or only a few measurements
per subject and when different subjects are studied on subsequent days. Pooled correlations
combine individual-subject/individual-day data for the calculation of correlations. Pooled
correlations describe the relationship between daily personal NO2 exposure and daily ambient
S02 concentration across all subjects in the study (e.g., Brauer et al., 1989; Sarnat et al., 2006).
Daily-average correlations2 are calculated by averaging exposure across subjects for each
day. Daily-average correlations then describe the relationship between the daily average
exposure and daily ambient pollutant concentration. This type of correlation (i.e., the association
between community average exposures (ambient component) and ambient concentrations) is
more directly relevant to community time-series and long-term cohort epidemiologic studies, in
which ambient concentrations are used as a surrogate for community average exposure to
pollutants of ambient origin. However, exposure of the population to S02 of ambient origin has
not been reported in any of the studies examined.
Not only does the exposure study design determine the meaning of the correlation
coefficients in the context of exposure assessment in epidemiologic studies, but it also affects the
strength of the association between personal exposures and ambient concentrations. The strength
of the association between personal exposures and ambient and/or outdoor concentrations for a
population is determined by variations in several physical factors: indoor or other local sources,
air exchange rate, penetration, and decay rate of the pollutant in different microenvironments and
the time people spend in different microenvironments with different pollutant concentrations. For
different types of correlation coefficients, the components of the variance of these physical
factors are different, and therefore the strength of different types of correlation coefficients is
different. Longitudinal correlation coefficients reflect the inter-personal variations of these
physical factors; pooled correlation coefficient reflect both inter- and intra- personal variations
of these physical factors; and for the association between community average exposures and
ambient concentrations, inter-personal variations of these physical factors are reduced by
Xf-v v)f" ")
^ =—(
. 1 ^2 J K1 $
V / x a where "n" is the number of paired measurements of exposure and ambient concentration, and all other symbols are
defined the same way as those in the longitudinal correlation coefficient.
Tfc-xlaj-a)
r-=-L
ax (n -1)5—.sa
2 "J , where n is the number of measurement period, during each of which the exposure for all subjects are measured, and all
other symbols are defined the same way as those in the longitudinal correlation coefficient.
May 2008
2-50
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
averaging personal exposures across a community. Therefore, the strength of the associations
between personal exposures and ambient concentrations may not be comparable directly,
although these associations are determined by the same set of physical factors (but affected in
different ways).
Since correlations are standardized quantities that depend on multiple features of the data,
in a correlation, not only is the linear "relatedness" (covariance) of the two quantities important,
but so is the variability of each, which can be affected by exposure factors in various ways. In the
following assessments, the effects of these physical factors on the strength of correlation
coefficients are primarily examined within a study, and the purpose of the inter-study comparison
is to examine the consistency of the effects across different types of studies.
The strength of the associations between personal exposures and ambient concentrations
could also be affected by the quality of the data collected during the exposure studies. There are
at least six aspects associated with the quality of the data: method precision, method accuracy
(compared with FRM), percent of data above method detection limits (based on field blanks),
completeness of the data collection, sample size, and soundness of the quality assurance/quality
control procedures. Unfortunately, not all studies reported the SIX aspects of the data quality
issue. The fraction of data below the detection limit might be a concern for some studies (see
Sarnat et al., 2000; 2001; 2005). Correlation coefficients would be biased low if data used in
their calculation are below detection limits. Sampling interferences associated with both ambient
(see Section 2.3) and personal sampling (see Section 2.5.2) could also affect data quality.
Therefore, caution must be exercised when interpreting the results in Table 2-9 and Table 2-10.
Sarnat et al. (2001; 2005; 2006) examined the associations between ambient S02 concentrations
and ambient or personal co-pollutant concentrations. Sarnat et al. (2001) reported that during the
winter of 1999, ambient S02 was significantly associated (at 5% significance level) with personal
exposure to fine particulate matter (PM2.s) (slope = - 0.24), personal exposure to SO42
(slope = - 0.03), and personal exposure to PM2.5 of ambient origin (slope = - 0.16). However, it
should be noted that all the slopes are negative perhaps as the result of measurement error. Sarnat
et al. (2005) reported that significant associations between ambient S02 and either personal
exposures or ambient concentrations of other pollutants were found for personal SO42 (winter,
slope = 0.06), personal SO42 (summer, slope = 0.39), personal PM2.5 (summer, slope = 1.68),
ambient SO42 (winter, slope = 0.19), and ambient PM2.5 (winter, slope = 0.80). In Sarnat et al.
May 2008
2-51
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
(2006), ambient S02 was observed to be significantly associated with ambient PM2.5, ambient
SO42 and ambient EC during the fall (R2 = 0.22, 0.33, and 0.34 respectively), and was
significantly associated with personal PM2.5 during the summer, personal S042 and personal EC
during the fall (R2 = 0.07, 0.06, and 0.05 respectively).
Of significant concern is the ability of currently available techniques for monitoring either
personal exposures or ambient concentrations to measure S02 concentrations that are typically
found in most urban environments. In some studies, most data, especially data for monitoring
personal exposure and indoor concentrations, might be beneath detection limits. Indeed, in one
study (Chang et al., 2000), the investigators had to discard data for S02, because the values were
mostly beneath detection limits. In the study of Kindzierski and Ranganathan (2006), all indoor
concentration data were beneath detection limits. In Sarnat et al. (2000), -70% of personal
measurements were beneath detection limits, and -33% of personal measurements returned
apparent negative concentration values. In such situations, associations between ambient
concentrations and personal exposure are inadequately characterized. When personal exposure
concentrations are above detection limits, a reasonably strong association is observed between
personal exposures and ambient concentrations.
2.5.4. Exposure Measurement Errors in Epidemiological Studies
For the purposes of this draft, the effects of exposure error on epidemiological study results
refers to changes in the health effects estimate expressed as the relative risk factor, p, and in the
related standard error that results from using the ambient concentration of an air pollutant as an
exposure indicator rather than using the actual personal exposure in the epidemiological
statistical analysis. There are many assumptions made in going from the available measurement
of a pollution indicator to an estimate of the personal exposure. The importance of these
assumptions and their effect on p depend on the type of epidemiological study.
The considerations of exposure error for S02 are simplified compared to those for NO2 and
PM. The only experimental measure available is the ambient concentration of S02. In addition,
indoor and other non-ambient sources of S02 are not thought to be important in population
studies, lessening concerns about the possible influence of exposures other than to ambient S02.
May 2008
2-52
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
2.5.4.1. Community Time-Series Studies
This section applies primarily to studies on the association of daily average S02
concentrations with daily measures of mortality or morbidity in a community. The following
three exposure issues are of primary concern with respect to S02 time-series epidemiological
analysis: (1) the relationship of the measured concentration of S02 to the true concentration; (2)
the relationship of day-to-day variations in the concentrations of S02, as measured at a central
monitoring site, with the corresponding variations in the average concentration of S02 over the
geographic area from which the health measurements are drawn; and (3) the relationship of the
community average concentration of S02 to the average personal exposure to ambient S02. These
three issues are described below.
2.5.4.1.1. Relationship of Measured so2 to the True Concentration
Since there is always a random component to instrumental measurement error, the
correlation of the measured S02 with the true S02, on either a 24-h or 1-h basis, will be less
than 1. Sheppard et al. (2005) indicate that instrument error in the individual or daily average
concentrations have "the effect of attenuating the estimate of a." Zeger et al. (2000) suggest that
instrument error has both Berkson and non-Berkson error components; however, the authors state
that the "instrument error in the ambient levels is close to the Berkson type," and in order for this
error to cause substantial bias in p, the error term (the difference between the true concentrations
and the measured concentrations) must be strongly correlated with the measured concentrations.
Zeger et al. (2000) suggest that, "further investigations of this correlation in cities with many
monitors are warranted." Averaging across multiple unbiased ambient monitors in a region
should reduce the instrument measurement error (Sheppard et al., 2005; Wilson and Brauer,
2006; Zeger et al., 2000). There are concerns about the precision and accuracy of the ambient
concentration measurements, because S02 concentrations are much lower now than when the S02
standards were first promulgated. Typical ambient concentrations of S02 in the contiguous
United States are nearly all at or beneath the detection limit of the monitors currently used in the
regulatory network. Thus, greater relative error is most often observed at the lower ambient
concentrations compared to the less frequent higher concentration exposures, as might occur
because of plume downwash near local point sources or entrainment of plumes downwind from
May 2008
2-53
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
large power plants or smelters. It is unclear how uncertainties in the true concentrations of S02,
i.e., instrument measurement error, will change p.
2.5.4.1.2. Relationship of Day-to-day Variations in the Ambient Concentration of
so2 to Variations in the Community Average
There has been little analysis of the spatial variation of S02 across communities. S02
emissions arise mainly from coal fired power plants (see Annex Table B-4). Newer power plants
and smelters in the United States are no longer located within urban areas. However, some older
power plants and industrial facilities are located in many urban areas, especially in the Midwest
and Northeast. Downwash from the plumes emitted from these facilities can contribute to
elevated levels of S02 at the surface in these cities. However, it is anticipated that S02 will
behave largely as a regional pollutant in most areas. Site-to-site correlations of S02
concentrations, as shown for several cities in Table 2-3 vary from very low to very high values.
This suggests the concentration of S02, measured at any given monitoring site, may not be highly
correlated with the average community concentration in some areas. There are a number of
possible reasons for these findings: local sources that cause the S02 to be unevenly distributed
spatially; a monitoring site being chosen to represent a nearby source; terrain features that divide
the community into several sub-communities that differ in the temporal pattern of pollution; and
errors in the measurement of the low concentrations of S02 present at most sites. To the extent
that the correlation of the ambient concentration with the community average concentration is
< 1, P will be reduced. Similarly, p will be reduced if there are subareas of the community where
the correlation of the subarea average concentrations with the concentrations measured at the
ambient monitoring site is < 1. If concentrations in an area of a community impacted by plumes
from local S02 sources might be higher than, and not well-correlated with, the concentrations at
the ambient monitor, and if such high concentrations affected a sizable portion of the population
affected by a local source, that community might not be suitable for time-series epidemiological
analyses. On the other hand, if the plume impacts the ambient monitor, the high concentration of
S02 not accompanied by a corresponding high effect in the entire community will bias P toward
the null.
May 2008
2-54
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
2.5.4.1.3. Relationship of Community Average Concentration of so2 to Average
Personal Exposure to Ambient so2
People spend much of their time indoors and, in the absence of indoor sources, indoor
concentrations are lower than outdoor concentrations. This is very likely the case with S02, since
the only known significant indoor source of S02 in the United States is the use of kerosene
heaters, not thought to be widespread enough to influence population studies. Differences in
infiltration factors among homes can also result in differences among individuals' personal
exposures. It is necessary to consider how this difference between the ambient concentration,
which is used in epidemiological analyses, and the personal ambient source exposure
concentration (which includes exposure to the full outdoor concentration while outdoors, and
exposure of only a fraction of the outdoor concentrations while indoors) will affect the calculated
p. The contribution of the ambient concentration of S02 to the personal exposure to ambient S02
is given by Ea = a • Ca where Ea is exposure to ambient S02, a is the ambient exposure factor
with values between 0 and 1, and Ca is the ambient S02 concentration as measured at a
community monitoring site. Zeger et al. (2000) noted that for community time-series
epidemiology, which analyzes the association between health effects and potential causal factors
at the community scale rather than the individual scale, it is the correlation of the daily average
ambient concentrations with the daily community average personal exposures that is important,
not the correlation between the daily average ambient concentrations and the individual personal
exposures. Thus, as mentioned in Section 2.5.3, the low correlation between daily average
ambient concentrations and individual personal exposures, as frequently found in pooled panel
exposure studies, is not relevant to community time-series epidemiological analysis.
Unfortunately, no studies provide adequate information about the community average personal
exposure to S02.
There has also been concern with the variation of a. Zeger et al. (2000) suggested (for PM)
that variations in the individual daily values of a would be a Berkson error and would not change
the point estimate of p. Sheppard et al. (2005) used simulations to confirm this for nonreactive
pollutants. However, such variations increase the standard error. Day-to-day variations in the
population average fraction of ambient exposure will not change the point estimate of P unless
the population average fraction of ambient exposure is correlated with seasonal trends in ambient
concentration, according to Sheppard et al.
May 2008
2-55
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Both Zeger et al. (2000) and Sheppard et al. (2005) show that if PA is the health effect
parameter that would be obtained with a time-series analysis using the ambient exposure and PC
is the health effect parameter that would be obtained with a time-series analysis using the
ambient concentration, then PC = a • fiA. Thus, time-series studies yield different parameters
depending on whether they use concentration or exposure. However, the two parameters are
related by a. Overestimation of exposure by substitution of the ambient concentration for the
ambient exposure leads to underestimation of the effect estimate, or generally bias toward the
null.
2.5.4.2. Short-Term Panel Studies
Panel epidemiology refers to time series studies that follow a relatively smaller number of
subjects for a relatively short time. Each subject must be considered individually. Panel studies
typically examine the association between symptoms or health outcomes and either ambient
concentrations or personal exposures. Personal exposures to S02 are not measured; rather,
ambient concentrations are used in panel studies. Similar types of exposure error as discussed for
community time series apply to panel studies.
The ambient exposure factor (a) may differ for each person and each day leading to error
in the exposure estimate. If a panel is composed of subjects who live in similar housing and have
similar activity pattern, and the study is limited to a single season, the variation in a over time
and individual subjects may be small. However, if the panels are composed of more diverse
subjects or extend or more than one season, values of a may be quite variable. Such variability
will cause error in the estimate of exposure for each subject.
2.5.4.3. Long-Term Cohort Studies
For long-term exposure epidemiological studies, concentrations are integrated over time
periods of a year or more, and usually for spatial areas the size of a city, county, or metropolitan
statistical area (MSA), although integration over smaller areas may be feasible. Health effects are
then regressed, in a statistical model, against the average concentrations in the series of cities (or
other areas). In time-series studies, a constant difference between the measured and the true
concentration (instrument offset) will not affect P, nor will variations in the daily average a or
May 2008
2-56
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
the daily average nonambient exposure, unless the variations are correlated with the daily
variations in concentrations. However, in long-term exposure epidemiological studies, if
instrument measurement errors, long-term average values of a, or long-term averages of
nonambient exposure differ for different cities (or other areas used in the analysis), the city-to-
city long-term ambient S02 concentrations will not be perfectly correlated with the long-term
average exposure to either ambient or total S02. This lack of correlation would be expected to
bias the point estimate of p.
2.5.4.4. Summary of Evaluation of Exposure Measurement Error in
Epidemiological Studies
Exposure error caused by using ambient concentrations of S02 as a surrogate for exposure
to ambient S02 affect P in different ways, dependent upon the type of epidemiological study. In
community time-series and short-term panel epidemiological studies, in general, the nonambient
source component of personal exposure and the variation in the ambient exposure factor caused
by building ventilation practices and personal behaviors, will not change the estimate of P; but
the spatial variation of S02 or the representativeness of the ambient monitor might bias the
estimate of P toward null. Therefore, P observed in S02 community time-series or panel
epidemiological studies would be stronger and less uncertain if exposure errors had been
adjusted and/or controlled for. In long-term cohort epidemiological studies, instrument
measurement errors, factors that influence exposure to ambient S02, or long-term averages of
nonambient exposure may differ for different cities, which may bias the estimate of P, but the
extent and direction of this bias is unclear.
2.6. Dosimetry of Inhaled Sulfur Oxides
This section is intended to present an overview of general concepts related to the dosimetry
of S02 in the respiratory tract. Dosimetry of S02 refers to the measurement or estimation of the
amount of S02 or its reaction products reaching and persisting at specific respiratory tract and
systemic sites after exposure. One of the principal effects of inhaled S02 is that it stimulates
bronchial epithelial irritant receptors and initiates a reflexive contraction of smooth muscles in
the bronchial airway. The compound most directly responsible for health effects may be the
inhaled S02, or perhaps its chemical reaction products. Complete identification of the causative
May 2008
2-57
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
agents and their integration into S02 dosimetry is a complex issue that has not been thoroughly
evaluated. Few studies have investigated S02 dosimetry in the interval since the 1982 AQCD and
the 1986 Second Addendum.
2.6.1. Gas Deposition
The major factors affecting the transport and fate of gases and aerosols in the respiratory
tract are: the morphology of the respiratory tract; the physiochemical properties of the mucous
and surfactant layers; respiratory functional parameters such as tidal volume, flow rate, and route
of breathing; physicochemical properties of the gas; and the physical processes that govern gas
transport. Physicochemical properties of S02 relevant to respiratory tract uptake include its
solubility and diffusivity in epithelial lining fluid (ELF), as well as its reaction-rate with ELF
constituents. Henry's law relates the gas phase and liquid phase interfacial concentrations at
equilibrium, and is a function of temperature and pressure. Henry's law shows that the amount of
S02 in the aqueous phase is directly proportional to the partial pressure or concentration of S02 in
the gas phase. Although the solubility of most gases in mucus and surfactant is not known, the
Henry's law constant is known for many gases in water. The Henry's law constant for S02 is
0.048 (mole/liter) air / (mole/liter) water at 37° C and 1 atm; for comparison, the value for 03 is
6.4 under the same conditions (Kimbell and Miller, 1999). In general, the more soluble a gas is in
biological fluids, the more rapid, and proximal its absorption will be in the respiratory tract.
When the partial pressure of S02 on mucosal surfaces exceeds that of the gas phase, such as
during expiration, some desorption of S02 from the ELF may be expected.
Because S02 is highly soluble in water, it is expected to be almost completely absorbed in
the nasal passages of both humans and laboratory animals under resting conditions. The
dosimetry of S02 can be contrasted with the lower solubility gas, O3, for which the predicted
tissue doses (03 flux to liquid-tissue interface) are very low in the trachea and increase to a
maximum in the terminal bronchioles or first airway generation in the pulmonary region (see
Chapter 4, EPA, 2006c). Similar to O3, the nasal passages remove S02 more efficiently than the
oral pathway (Brain, 1970). With exercise, the pattern of S02 absorption shifts from the upper
airway to the tracheobronchial airway in conjunction with a shift from nasal to oronasal
breathing and increased ventilatory rates. Due to its effect on delivery and uptake, mode of
breathing is also recognized as an important determinant of the severity of S02-induced
May 2008
2-58
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
bronchoconstriction, with the greatest responses occurring during oral breathing followed by
oronasal breathing and the smallest responses observed during nasal breathing.
Melville (1970) measured the absorption of S02 (1.5 to 3.4 ppm) during nasal and oral
breathing in 12 healthy volunteers. Total respiratory tract absorption of S02 was significantly
greater (p < 0.01) during nasal than oral breathing (85 versus 70%, respectively) and was
independent of the inspired concentration. Respired flows were NR. Andersen et al. (1974)
measured the nasal absorption of S02 (25 ppm) in 7 volunteers at an average inspired flow of 23
L/min (i.e., eucapnic hyperpnea [presumably] to simulate light exertion). These investigators
reported that the oropharyngeal S02 concentration was below their limit of detection (0.25 ppm),
implying that at least 99% of S02 was absorbed in the nose of subjects during inspiration.
Speizer and Frank (1966) also measured the absorption of S02 (16.1 ppm) in 7 healthy subjects
at an average ventilation of 8.5 L/min (i.e., at rest). They reported that 14% of the inhaled S02
was absorbed within the first 2 cm into nose. The concentration of S02 reaching the pharynx was
below the limit of detection, suggesting that at least 99% was absorbed during inspiration.
Frank et al. (1969) and Brain (1970) investigated the oral and nasal absorption of S02 in
the surgically isolated upper respiratory tract of anesthetized dogs. Radiolabeled S02 (35S02) at
the concentrations of 1, 10, and 50 ppm was passed separately through the nose and mouth at the
steady flows of 3.5 and 35 L/min for 5 min. The nasal absorption of S02 (1 ppm) was 99.9% at
3.5 L/min and 96.8% at 35 L/min. The oral absorption of S02 (1 ppm) was 99.56% at 3.5 L/min,
but only 34% at 35 L/min. The nasal absorption of S02 at 3.5 L/min increased with concentration
at 1, 10, and 50 ppm and was reported to be 99.9, 99.99, and 99.999%, respectively. This
increase in absorption with concentration was hypothesized to be due to increased mucous
secretion and increased nasal resistance at the higher S02 concentrations. The increased mucus
was thought to provide a larger reservoir for S02 uptake. The increased nasal resistance may
increase turbulence in the airflow and, thereby, decrease the boundary layer between the gas and
liquid phases. Dissimilar to the nose, S02 absorption in the mouth decreased from 99.56 to
96.3% when the concentration was increased from 1 to 10 ppm at 3.5 L/min. Frank et al. (1969)
noted that the aperture of the mouth may vary considerably, and that this variation may affect
S02 uptake in the mouth. Although S02 absorption was dependent on inhaled concentration, the
rate and route of flow had a greater effect on the magnitude of S02 absorption in the upper
airway.
May 2008
2-59
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Strandberg (1964) studied the uptake of S02 in the respiratory tract of rabbits. A tracheal
cannula with two outlets was utilized to allow sampling of inspired and expired air, and S02
absorption was observed to depend on inhaled concentration. The absorption during maximal
inspiration was 95% at high concentrations (100 to 700 ppm), reflecting an increased S02
removal in the extrathoracic (ET) airway, whereas it was only 40% at low concentrations (0.05 to
0.1 ppm). On expiration, the total S02 absorbed (i.e., inspiratory removal in the ET airway plus
removal in the lower airway) was 98% at high concentrations and only 80% at the lower
concentrations.
Amdur (1966) examined changes in airway resistance in guinea pigs due to S02 exposure.
Guinea pigs were exposed for 1-h to 0.1- to 800 ppm S02 during natural unencumbered breathing
or to 0.4 to 100 ppm while breathing through a tracheal cannula. At concentrations of 0.4- to 0.5
ppm S02, route of administration did not affect the airway resistance response, whereas at
concentrations of > 2 ppm, the responses were greater in animals exposed by tracheal cannula.
Based on the concentration-dependent absorption of S02 in the ET airway observed by
Strandberg (1964), Amdur (1966) concluded that the airway resistance responses at low-
exposure concentrations were independent of method of administration, because the lung
received nearly the same concentration with or without the cannula as evidenced by minimal ET
absorption.
More recently, Ben-Jebria et al. (1990) investigated the absorption of S02 in excised
porcine tracheae. Absorption was monitored over a 30-min period following the introduction of
S02 (0.1 to 0.6 ppm, inlet concentration) at a constant flow (2.7 to 11 L/min). The data were
analyzed using diffusion-reactor theory. An overall mass transfer coefficient (KS02) was
determined and separated into its contributions due to gas (convection and diffusion) and tissue
phase (diffusivity, solubility, and reaction rates) resistances. S02 in the liquid phase was assumed
to form HSO3 rapidly, in proportion with the gas phase S02 concentration, HSO3 then diffused
down the concentration gradient into the tissues where it reacted irreversibly with biochemical
substrates. Initially, KS02 was limited only by gas phase resistance, but decreased exponentially
over the first 5 to 10 min of S02 exposure to a smaller steady-state value because of tissue
resistance to S02 absorption. The initial and steady-state KS02 values were found to be
independent of inlet S02 concentration, i.e., for a given flow, the fractional absorption of S02 did
not depend on S02 concentration. An increased KS02 (initial and steady-state) was observed with
May 2008
2-60
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
an increasing flow that was thought to be due to a decrease in the boundary layer near the walls
of the trachea for radial S02 transport. This is in agreement with Aharonson et al. (1974), who
also reported that the transfer rate coefficient for S02 increases with increasing flow. However,
the initial molar flux of S02 across the gas-tissue interface appears to increase purely as a
function of the increase in mass transport occurring with increasing flow (see Figure 5 in Ben-
Jebria et al., 1990). Given that the steady-state KS02 remained stable during the 10 to 30 min of
exposure and that no S02 leakage through the tissue was identified, the authors concluded that
there was an irreversible sink for S02 within the tissue.
Mathematical modeling specific to the regional respiratory uptake of S02 is unavailable for
humans and laboratory animals. More generally, the influence of age on gas dosimetry in humans
during light activity (on average) was examined by Ginsberg et al. (2005) using the U.S. EPA
reference concentration methodology (EPA, 1994a). For a highly soluble gas, such as S02, they
predicted that the majority of gas uptake would occur in the extrathoracic airway and that uptake
in these airways would be modestly greater in a 3-month-old infant than an adult. The rate of gas
uptake per surface, however, in the extrathoracic airway and large bronchial airway was not
markedly different between infants and adults. The smaller bronchial airway of adults were
predicted to receive a greater dose (i.e., uptake per unit time and surface area) relative to infants,
although the majority of the inhaled S02 would be removed proximal to these airways.
In summary, inhaled S02 is readily absorbed in the upper airway of both humans and
laboratory animals. During nasal breathing, the majority of available data suggests 95% or
greater S02 absorption occurs in the nasal passages, even under ventilation levels comparable to
exercise. Somewhat less S02 is absorbed in the oral passage than in the nasal passages. The
difference in S02 absorption between the mouth and the nose is highly dependent on respired
flow rates. With an increase in flow from 3.5 to 35 L/min, nasal absorption is relatively
unaffected, whereas, oral absorption is reduced from 100 to 34%. Thus, the rate and route of
breathing have a great effect on the magnitude of S02 absorption in the upper airway and so the
penetration of S02 to the lower airway. Overall, the available data clearly show that the pattern of
S02 absorption which shifts from the upper airway to the tracheobronchial airway in conjunction
a shift from nasal to oronasal breathing and associated increased ventilatory rates in exercising
humans. Mode of breathing is also recognized as an important determinant of the severity of
May 2008
2-61
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
S02-induced bronchoconstriction, with the greatest responses occurring during oral breathing
followed by oronasal breathing and the smallest responses observed during nasal breathing.
2.6.2. Particles and Sulfur Oxide Mixtures
As already discussed, inhaled S02 is readily absorbed in the upper airway, particularly
during nasal breathing. It has been suggested that sulfur oxides may become absorbed to
particles and subsequently transported to more distal lung regions. Depending on atmospheric
conditions, S02 can be transformed to secondary sulfate particles and acid aerosols (H2SO4) and
can adsorb onto particulate matter. Jakab et al. (1996) observed that the conversion of S02 to
SO42 on the surface of carbon black aerosols was dependent on high relative humidity (85%)
and S02 concentration. These investigators suggested that fine carbon black particles can be an
effective vector for delivery of SO42 to the peripheral lung. Other studies investigating the
effects of S02 coated aerosols are briefly discussed in Section 3.1.5.
Sulfate aerosols are hygroscopic and grow in the respiratory tract. The implications of
hygroscopic growth on deposition have been reviewed extensively by Morrow (1986) and Hiller
(1991). In general, compared to nonhygroscopic particles of the same initial size, the deposition
of hygroscopic aerosols in different regions of the lung may be higher or lower, depending on the
initial size. For particles with initial sizes larger than 0.5 [j,m (aerodynamic diameter), the
influence of hygroscopicity would be to increase total deposition with a shift in regional
deposition from the distal to larger proximal airway; for smaller particles deposition would tend
to be decreased. A thorough review of respiratory deposition and clearance of particulate matter
is available elsewhere (EPA, 2004; 2006b). The intent herein was to briefly mention some issues
specific to sulfur oxides.
2.6.3. Distribution and Elimination of Sulfur Oxides
When S02 contacts the fluids lining the airway, it dissolves into the aqueous fluid and
forms hydrogen (H+) ions and bisulfite (HSO3 ) and sulfite (SC>32~) anions (Bascom et al., 1996).
The majority of anions are expected to be present as HS03 at a concentration proportional to the
gas phase concentration of S02 (Ben-Jebria et al., 1990). Because of the chemical reactivity of
these anions, various reactions are possible, leading to the oxidation of S032~ to S042 (see
Section 12.2.1, EPA, 1982). Clearance of S032- from the respiratory tract may involve several
May 2008
2-62
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
intermediate chemical reactions and transformations. Gunnison and Benton (1971) identified S-
sulfonate in blood as a reaction product of inhaled S02. Following inhalation of S02, the
clearance half-time of 4.1 days for ^-sulfonate in rabbits has been reported (Gunnison and
Palmes, 1973).
Some S02 is also removed by desorption of from the respiratory tract. Desorption is
expected when the partial pressure of S02 in airway lining fluids exceeds that of the air. Speizer
and Frank (1966) found that on expiration, 12% of the S02 absorbed during inspiration was
desorbed into the expired air. During the first 15 min after the 25- to 30-min S02 exposure,
another 3% was desorbed. In total, 15% of the amount originally inspired and absorbed S02 was
desorbed from the nasal mucosa. Frank et al. (1969) reported that up to 18% of the S02 was
desorbed within -10 min after exposure.
May 2008
2-63
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Chapter 3. Integrated Health Effects
This integrated discussion is structured to provide a coherent framework for the assessment
of health risks associated with human exposure to ambient S02 in the United States. The main
goals of this chapter are: (1) to integrate newly available epidemiological, human clinical, and
animal toxicological evidence with consideration of key findings and conclusions from the 1982
AQCD for Sulfur Oxides and First Addendum (EPA, 1982), 1986 Second Addendum (EPA,
1986c), and 1994 Supplement to the Second Addendum, (EPA, 1994c); and (2) to draw
conclusions about the causal nature of S02 in relation to a variety of health effects. These causal
determinations utilize the framework outlined in Chapter 1.
This chapter is organized to present morbidity and mortality associated with short-term
exposures to S02, followed by morbidity and mortality associated with long-term exposures.
Human clinical studies examining the effect of peak exposures (less than 1-h, generally 5-10
min) of S02 on respiratory symptoms and lung function are discussed first. Later sections
describe the findings of epidemiological studies that examine the association between short-term
(generally 24-h avg) and long-term (generally months to years) ambient S02 exposure and heath
outcomes, such as respiratory symptoms in children and asthmatics, emergency department (ED)
visits and hospital admissions for respiratory and cardiovascular diseases, and premature
mortality. The human clinical and epidemiological evidence are presented with relevant animal
toxicological data, when available.
Considerations in the Interpretation of Health Evidence
Human clinical studies are conducted in a controlled laboratory setting using fixed
concentrations of air pollutants under carefully regulated environmental conditions and subject
activity levels. Results of human clinical studies provide evidence of potential mechanisms for
observed effects and a direct quantitative assessment of the S02 exposure-health response
relationship among asthmatic individuals. Observed effects in these studies may underestimate
the response in certain sensitive subpopulations for a number of reasons. First, study subjects
must either be healthy, or have a level of illness which does not preclude them from participating
in the study. Second, asthmatics who are unable to withhold the use of bronchodilators for at
May 2008
3-1
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
least 6 hours prior to exposure and subjects with a recent history of upper respiratory tract
infections are typically excluded from clinical studies of exposure to S02. While human clinical
studies provide important information on the biological plausibility of associations observed
between S02 exposure and health outcomes in epidemiological studies, the concentration-
response relationships cannot necessarily be directly extrapolated to concentrations below those
administered in the laboratory. Further, human clinical studies are normally conducted on a
relatively small number of subjects, which reduces the power of the study to detect significant
differences in the health outcomes of interest between exposure to varying concentrations of S02
and clean air.
Epidemiological studies provide important information on the associations between health
effects and exposure of human populations to ambient levels of S02. These studies also help to
identify susceptible subgroups and associated risk factors. However, associations observed
between specific air pollutants and health outcomes in epidemiological studies may be
confounded by copollutants or meteorological conditions, and influenced by model
specifications in the analytical methods. Extensive discussion of these issues is provided in the
2004 AQCD for PM (EPA, 2004) and the 2006 AQCD for O3 and Related Photochemical
Oxidants (EPA, 2006c), and therefore presented only briefly below.
The use of multipollutant regression models has been the prevailing approach for
controlling potential confounding by copollutants in air pollution health effects studies. Finding
the likely causal pollutant from multipollutant regression models is made difficult by the
possibility that one or more air pollutants may be acting as a surrogate for an unmeasured or
poorly-measured pollutant or for a particular mixture of pollutants. S02 presents an especially
interesting test of multipollutant effects models: correlations with sulfate, the principal
atmospheric oxidation product of S02, show temporal and spatial discongruities that can
influence exposures and health effects. Short-term, mostly time-series epidemiological studies
generally use intracity ambient concentration data which show very little or no correlation
between emitted S02 and transformed sulfate. In contrast, long-term epidemiological studies
using intercity data can show correlations between S02 and sulfate on the order of 0.8 or higher.
In these studies the fine-scale spatiotemporal variations in the intracity data are significantly
reduced, since sulfate has sufficient time for production from S02, dispersed over a wide spatial
area, and mixed down to ground level. Layered over these spatial and fine-scale temporal
May 2008
3-2
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
differences are seasonal and regional dissimilarities driven by cities' various S02 emissions
profiles and differing available time and sunlight conditions for oxidation. Thus, attempts to
distinguish gaseous and particle effects related to S02 using multipollutant epidemiological
models must be interpreted with caution. Despite these limitations, the use of multipollutant
models is still the prevailing approach employed in most studies of S02 and health effects, and
may provide some insight into the potential for confounding or interaction among pollutants.
Model specification and model selection also need to be considered in the interpretation of
the epidemiological evidence. The studies presented in this chapter investigated the association
between various measures of S02 (e.g., multiple lags and different exposure metrics) and various
health outcomes using different model specifications. The summary of health effects in this
chapter is vulnerable to the errors of publication bias and multiple testing. Efforts have been
made to reduce the impact of multiple testing errors. For example, although many studies
examined multiple single-day lag models, priority was given to effects observed at 0- or 1-day
lags, rather than at longer lags. Additional focus was placed on results from distributed and
moving average lags as they are able to take into consideration multiday effects. Both single- and
multiple-pollutant models were considered and examined for robustness of results. Additional
analyses of multiple model specifications for adjustment of temporal or meteorological trends are
considered to be sensitivity analyses.
In addition to issues related to confounding by copollutants and model selection, the
evaluation of the epidemiological evidence also considers study population and sample size, with
particular emphasis placed on multicity studies. Other factors considered are study location
(North America versus other regions), meaningfulness and reliability of the health endpoint
measurements, and appropriateness of the statistical analyses methods used. These
considerations in the interpretation of the epidemiological evidence lead to emphasis of certain
studies in the chapter text, tables, and figures.
Animal toxicological studies may provide further evidence for the potential mechanism of
an observed effect; however, most of these studies have been conducted at concentrations vastly
exceeding current ambient conditions. In discussing the mechanisms of SOx toxicity, studies
conducted under atmospherically relevant conditions are emphasized whenever possible; studies
at higher levels are also considered, due to species-to-species differences and potential
differences in sensitivity between study subjects and especially susceptible human populations.
May 2008
3-3
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
This chapter focuses on important new scientific studies, with emphasis on those
conducted at or near current ambient concentrations. Given their respective strengths and
limitations, evidence from human clinical, epidemiological and animal toxicological studies are
considered in order to evaluate the causality of SOx-health effects associations. The annexes
supplement the information included here by presenting a more details of the literature.
3.1. Respiratory Morbidity Associated with Short-Term
Exposure
3.1.1. Summary of Findings from the Previous Review
The majority of the S02 human clinical studies in the 1982 AQCD for Sulfur Oxides
evaluated respiratory effects of S02 exposure in healthy adults, with some limited data from
clinical studies of adults with asthma. S02-related respiratory effects such as increased airway
resistance and decreased forced expiratory volume in 1 s (FEVi) were observed in healthy
individuals at concentrations > 1.0-5.0 ppm, and in asthmatics at concentrations <1.0 ppm. The
1986 Second Addendum (EPA) and 1994 Supplement to the Second Addendum (EPA) reviewed
several additional controlled studies involving both healthy and asthmatic individuals. In general,
these studies found no pulmonary effects of S02 exposure in healthy subjects exposed to
concentrations <1.0 ppm (Bedi et al., 1984; Folinsbee et al., 1985; Kulle et al., 1984; Stacy et
al., 1983). However, in exposures of asthmatic adults, respiratory effects were observed
following short-term exposures (5-10 min) to levels <1.0 ppm (Balmes et al., 1987; Horstman et
al., 1988; Linn et al., 1987).
Only a few epidemiological studies reviewed in the 1982 AQCD were useful in
determining the concentration-response relationship of respiratory health effects from short-term
exposure to S02. The most notable study was by Lawther (1970), which examined the
association between air pollution and worsening health status in bronchitic patients residing in
London, England. It was concluded in the 1982 AQCD that worsening of health status among
chronic bronchitic patients was associated with daily black smoke (BS) levels of 250-500 |ig/m3
in the presence of S02 levels in the range of 191-229 ppb. In the 1986 Second Addendum,
additional studies investigated morbidity associated with short-term exposure to S02. The most
relevant study was by Dockery (1982), which examined pulmonary function in school children in
May 2008
3-4
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Steubenville, OH, as part of the Harvard Six Cities Study. This study found that small but
statistically significant reversible decrements in forced vital capacity (FVC) and forced
expiratory volume in 0.75 s (FEV0.75) were associated with increases in 24-h avg concentrations
of total suspended particles (TSP) at levels ranging up to 220-420 |ig/m3 and S02 at levels
ranging up to 107-176 ppb. However, it was impossible to separate the relative contributions of
TSP and S02, and no threshold level for the observed effects could be discerned from the wide
range of exposure levels.
Epidemiological evidence for an association between S02 and respiratory morbidity, as
indicated by increased use of ED facilities or increased hospital admissions for respiratory
diseases, was also reported in the 1982 AQCD. Overall, these results suggested increased upper
respiratory tract morbidity during episodic marked elevations of PM or S02 (400-500 ppb),
especially among older adults. The 1982 AQCD further concluded that the studies reviewed
provided essentially no evidence for an association between asthma attacks and acute exposures
at typical ambient PM or S02 levels in the United States (the mean annual average S02
concentrations from 1972 to 1977 was approximately 6 ppb, with 90th percentile values ranging
from 15 to 20 ppb).
The 1982 AQCD for Sulfur Oxides (EPA, 1982) reported numerous effects on the
respiratory system in animals exposed to S02. Effects were generally observed at levels
exceeding those found in the ambient environment, and included morphological changes, altered
pulmonary function, lipid peroxidation, and changes in host lung defenses. The immediate effect
of acute S02 exposure in animals was increased pulmonary resistance to airflow, a measure of
bronchoconstriction. Bronchoconstriction was reported to be the most sensitive indicator of lung
function effects in acute S02 exposure.
Collectively, the human clinical, epidemiological and animal toxicological, studies
provided biological plausibility and coherent evidence of an adverse effect of ambient S02 on
respiratory health. Since the 1982 AQCD, 1986 Second Addendum, and 1994 Supplement to the
Second Addendum, additional studies have been conducted to determine the relationship
between short-term exposures to ambient S02 and adverse respiratory health effects, including
respiratory symptoms, lung function, airway inflammation, airway hyperresponsiveness, lung
host defenses, and ED visits and hospitalizations for respiratory causes. The epidemiological,
human clinical, and animal toxicological evidence on the effects of S02 on these various
May 2008
3-5
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
endpoints are discussed below. The finding of the previous review are integrated below with the
current literature.
3.1.2. Potential Mode of Action for Respiratory Health Effects
The 1982 AQCD (EPA, 1982) gave background information on the biochemistry of S02,
chemical reactions of bisulfite (HSO3 ), metabolism of S02, and the activating or inhibiting
effects of bisulfite on various enzymes. S02 readily dissolves in water, rapidly becoming
hydrated to form sulfurous acid, which at physiological pH substantially dissociates to form
bisulfite and sulfite (S032 ) ions. Studies in vitro have shown that S02 and/or bisulfite readily
react with nucleic acids, proteins, lipids, and other classes of biomolecules. Bisulfite participates
in three important types of reactions with biomolecules: sulfonation (sulfitolysis), autooxidation
with generation of free radicals, and addition to cytosine. Products of sulfonation reactions have
been shown to be long-lived in vivo and may be highly reactive. Products of autooxidation may
be responsible for the initiation of lipid peroxidation, which, among other effects, could damage
plasma membranes. In contrast, studies have shown that bisulfite can react with nucleic acids to
convert cytosine to uracil, thus resulting in mutational events. A principal mechanism of
detoxification of S02 (and sulfite/bisulfite) occurs through the enzymatic activity of sulfite
oxidase, resulting in the production of sulfate. Sulfite oxidase is a molybdenum-containing
enzyme, and the 1982 AQCD noted that depleting its activity through a low-molybdenum diet
supplemented with the competitive inhibitor tungsten resulted in a significant lowering of the
LD50 for intraperitoneally injected bisulfite. It was also noted that while in vitro exposure to S02
or sulfite/bi sulfite had been shown to either activate or inhibit a variety of enzymes, no such
effects had yet been demonstrated for in vivo exposure.
As discussed in the 1982 AQCD, the immediate effect of acute S02 exposure in animals
was bronchoconstriction. Reactions of S02 with respiratory tract fluids can result in the
production of bisulfite, sulfite, and a lowering of the pH, which may be involved in the
bronchoconstrictive response. It is now widely appreciated that bronchoconstriction following
S02 exposure is mediated by chemosensitive receptors in the tracheobronchial tree. Rapidly
activating receptors (RARs) and sensory C-fiber receptors found at all levels of the respiratory
tract are sensitive to irritant gases such as S02 (Coleridge and Coleridge, 1994; Widdicombe,
2006). Activation of these vagal afferents causes central nervous system reflexes resulting in
May 2008
3-6
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
bronchoconstriction, mucus secretion, mucosal vasodilation, cough, or apnea, followed by rapid
shallow breathing and effects on the cardiovascular system such as bradycardia and hypotension
or hypertension (Coleridge and Coleridge, 1994; Widdicombe and Lee, 2001; Widdicombe,
2003).
Early experiments demonstrated that S02-induced reflexes were mediated by cholinergic
parasympathetic pathways involving the vagus nerve and inhibited by atropine (Grunstein et al.,
1977; Nadel et al., 1965a; 1965b). Bronchoconstriction was found to involve smooth muscle
contraction since P-adrenergic agonists such as isoproterenol reversed the effects (Nadel et al.,
1965a; 1965b). Acetylcholine and histamine were also thought to be involved in S02-induced
bronchoconstriction (EPA, 1982).
More recent experiments in animal models conducted since 1982 have demonstrated that
both cholinergic and noncholinergic mechanisms may be involved in S02-induced effects. In two
studies utilizing bilateral vagotomy, vagal afferents were found to mediate the immediate
ventilatory responses to S02 (Wang et al., 1996), but not the prolonged bronchoconstrictor
response (Barthelemy et al., 1988). Other studies showed that atropine failed to block S02-
induced bronchoconstriction, and that a local axon reflex resulting in C-fiber secretion of
neuropeptides (i.e.., neurogenic inflammation) was responsible for the effect (Atzori et al.,
1992a; Hajj et al., 1996). Neurogenic inflammation has been shown to play a key role in animal
models of airway inflammatory disease (Groneberg et al., 2004).
In humans, the mechanisms responsible for S02-induced bronchoconstriction are not fully
understood. In non-asthmatics, near complete attenuation of bronchoconstriction has been
demonstrated using the anticholinergic agents atropine and ipratropium bromide (Snashall and
Baldwin, 1982; Tan et al., 1982; Yildirim et al., 2005). However, in asthmatics, these same
anticholinergic agents (Field et al., 1996; Myers et al., 1986a), as well as short- and long-acting
/^-adrenergic agonists (Gong et al., 1996; Linn et al., 1988), theophylline (Koenig et al., 1992),
cromolyn sodium (Myers et al., 1986), nedocromil sodium (Bigby and Boushey, 1993) and
leukotriene receptor antagonists (Gong et al., 2001; Lazarus et al., 1997) only partially blocked
S02-induced bronchoconstriction. That none of these therapies have been shown to completely
attenuate the effects of S02 implies the involvement of both parasympathetic pathways and
inflammatory mediators in asthmatics. Strong evidence of this is borne out in a study by Myers
et al. (1986), in which asthmatic adults were exposed to S02 following pretreatment with
May 2008
3-7
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
cromolyn sodium (a mast cell stabilizer), atropine (a muscarinic receptor antagonist), and the two
medications together. While both treatments individually provided some protection against the
bronchoconstrictive effects of S02, there was a much stronger and statistically significant effect
following concurrent administration of the two medications.
It has been proposed that inflammation contributes to the enhanced sensitivity to S02 seen
in asthmatics by altering autonomic responses (Tunnicliffe et al., 2001), enhancing mediator
release (Tan et al., 1982) and/or sensitizing C-fibers and RARs (Widdicombe and Lee, 2001).
Whether local axon reflexes also play a role in S02-induced bronchoconstriction in asthmatics is
not known (Widdicombe and Lee, 2001; Widdicombe, 2003; Groneberg et al., 2004). However,
differences in respiratory tract innervation between rodents and humans suggest that C-fiber
mediated neurogenic inflammation may be unimportant in humans (Groneberg et al., 2004;
Widdicombe and Lee, 2001; Widdicombe, 2003).
3.1.3. Respiratory Effects Associated with Peak Exposure
S02-induced respiratory effects among exercising asthmatics are well-documented, and
have been consistently observed following peak exposures (defined here as 5-10 min exposures
to relatively higher concentrations, e.g., 0.4-1.0 ppm) (Balmes et al., 1987; Bethel et al., 1985;
Horstman et al., 1986; 1988; Linn et al., 1984b; 1987; 1990; Schachter et al., 1984; Sheppard et
al., 1981). Similar respiratory effects have been observed in some sensitive asthmatics at
concentrations as low as 0.2-0.3 ppm; however, these effects have not reached statistical
significance (Horstman et al., 1986; Linn et al., 1987; 1988; 1990). Since the publication of the
1994 Supplement, several additional human clinical studies have been published that provide
supportive evidence of S02-induced decrements in lung function and increases in respiratory
symptoms among exercising asthmatics (see Annex Table D-2). Descriptions of older studies are
presented in the 1994 Supplement, and will not be described in great detail in this document.
However, based on recent guidance from the American Thoracic Society (ATS) regarding what
constitutes an adverse health effect of air pollution (ATS, 2000a), some key older studies were
reviewed and analyzed along with studies published since 1994. In its official statement, the ATS
recommended that transient loss in lung function with accompanying respiratory symptoms
attributable to air pollution should be considered adverse. In addition, ATS concluded that a
decrease in health-related quality of life, which refers to an individual's perception of well being,
May 2008
3-8
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
should also be considered to represent an adverse effect of air pollution. Therefore, whereas the
conclusions in the 1994 Supplement were based on SO2 exposure concentrations which resulted
in large decrements in lung function along with moderate to severe respiratory symptoms, the
current review of data from human clinical studies focuses on moderate to large SCh-induced
decrements in lung function combined with respiratory symptoms ranging from mild (perceptible
wheeze or chest tightness) to severe (breathing distress requiring the use of a bronchodilator).
3.1.3.1. Respiratory Symptoms
The 1994 Supplement to the Second Addendum described in detail several studies that
evaluated respiratory symptoms following controlled human exposures to SO2. Briefly, following
5-min exposures to 0, 0.2, 0.4, and 0.6 ppm S02 during moderate to heavy levels of exercise (48
L/min), Linn et al. (1983) reported that the severity of respiratory symptoms (i.e., cough, chest
tightness, throat irritation) among asthmatics increased with increasing SO2 concentration.
Relative to clean air exposures, exposures to SO2 resulted in statistically significant increases in
respiratory symptoms at concentrations of 0.4 and 0.6 ppm. In a subsequent study, Linn et al.
(1987) observed a significant effect of SO2 on respiratory symptoms in asthmatics who were
engaged in slightly lower levels of exercise (40 L/min) for a duration of 10 min. Clear increases
in respiratory symptoms were observed at concentrations of 0.6 ppm, with 43% of subjects
experiencing S02-induced symptoms. Some evidence of S02-induced increases in respiratory
symptoms was also demonstrated at concentrations as low as 0.4 ppm, with 15% of subjects
experiencing symptoms (Smith, 1994). It was also observed that these symptoms abated < 1 h
after exposure. Balmes et al. reported that 7 out of 8 asthmatic adults developed respiratory
symptoms, including wheezing and chest tightness, following 3-min exposures to 0.5 ppm SO2
during eucapnic hyperpnea (yE = 60 L/min).
Additional human clinical studies published since the 1994 Supplement to the Second
Addendum have provided support for previous conclusions regarding the effect of peak
exposures to SO2 on respiratory symptoms. In a human clinical study with SCVsensitive
asthmatics, Gong et al. (1995) reported that respiratory symptoms (i.e., shortness of breath,
wheeze, and chest tightness) increased with increasing SO2 concentration (0, 0.5, and 1.0 ppm
S02) following exposures of 10 min with varying levels of exercise. It was also observed that
exposure to 0.5 ppm S02 during light exercise evoked a more severe symptomatic response than
May 2008
3-9
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
heavy exercise in clean air. Trenga et al. (1999) observed a significant correlation between
decreases in FEVi and increases in respiratory symptoms following 10 min exposures to
0.5 ppm S02.
3.1.3.2. Lung Function
In controlled exposures of healthy human subjects to SO2, respiratory effects including
increased respiration rates, decrements in peak flow, bronchoconstriction, and increased airway
resistance have been observed at concentrations > 1 ppm (Abe, 1967; Amdur et al., 1953;
Andersen et al., 1974; Frank et al., 1962; Lawther, 1955; Lawther et al., 1975; Sim and Pattle,
1957; Snell and Luchsinger, 1969). SCVinduced decrements in lung function can be potentiated
by increasing ventilation rate, either through eucapnic hyperpnea or by performing exercise
during exposure. This effect is likely due to an increased uptake of SO2 resulting from both the
increase inyE as well as a shift from nasal breathing to oronasal breathing.
It has been clearly established that subjects with asthma are more sensitive to the
respiratory effects of SO2 exposure than healthy individuals without asthma. Asthmatic
individuals exposed to SO2 concentrations as low as 0.4-0.6 ppm for 5-10 min during exercise
have been shown to experience moderate or greater bronchoconstriction, measured as an increase
in sRaw (> 100%) or decrease in FEVi (> 15%) after correction for exercise-induced responses
in clean air (Linn et al., 1983; 1984; 1987; 1988; 1990; Magnussen et al., 1990; Roger et al.,
1985). Asthmatic subjects who are most sensitive to the respiratory effects of SO2 have been
observed to experience significant decrements in lung function following exposure to SO2 at
concentrations < 0.3 ppm (Horstman et al., 1986; Sheppard et al., 1981). In some cases,
bronchoconstrictive responses to SO2 can occur in as little as 2 min after the start of exposure
(Balmes et al., 1987; Horstman et al., 1988). Gong et al. (1995) demonstrated an exposure-
response relationship between SO2 and lung function by exposing 14 unmedicated, SCVsensitive
asthmatics to 0, 0.5, and 1 ppm S02 under 3 different levels of exercise. It was shown that
increasing SO2 concentration had a greater effect on sRaw and FEVi than increasing exercise
level. Trenga et al. (1999) observed that 25 out of 47 adult asthmatics experienced a drop in
FEVi versus baseline of between 8 and 44% (mean = 17.2%) following a 10 min exposure to 0.5
ppm S02 during moderate exercise.
May 2008
3-10
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Since some of the studies involving asthmatic subjects have used change in sRaw as the
endpoint of interest while others measured changes in FEVi or both, a comparison of FEVi and
sRaw based on data from Linn et al. (1987; 1990) was provided in the 1994 Supplement to the
Second Addendum. Based on simple linear interpolation of the data from these two studies, a
100% increase in sRaw corresponded to a 12 to 15% decrease in FEVi and a 200% increase in
sRaw corresponded to a 25 to 30% decrease in FEVi.
One of the aims of the Linn et al. (1987) study was to determine how the intensity of
response varied with asthma severity or status. In this study, 24 normal, 21 atopic (but not
asthmatic), 16 mild asthmatic, and 24 moderate/severe asthmatic subjects were exposed to SO2
concentrations between 0 and 0.6 ppm. While the moderate/severe asthmatics were more
responsive than mild asthmatics following exposure to clean air during exercise, their increases
in response to increasing SO2 concentrations were similar to those of the mild asthmatic group.
Thus, it was concluded that SO2 response was not strongly dependent on the clinical severity of
asthma. However, the apparent lack of correlation between SO2 response and asthma severity
should be interpreted with caution, since the S02 response may have been attenuated by
medication usage or its persistence. Three of the moderate/severe asthmatics were unable to
withhold medication usage during the exposure period. Conversely, a few of the asthmatics,
including some in the moderate/severe group, did not react to 0.6 ppm SO2.
One of the key studies discussed in the 1994 Supplement to the Second Addendum was by
Horstman et al. (1986). In this study, 27 asthmatic subjects were exposed to concentrations of
SO2 between 0- and 2 ppm SO2 for 10 min on different days under exercising conditions (yE =
42 L/min). The authors reported that for 22% of the subjects, the concentration of S02 needed to
produce a doubling of sRaw compared to clean air exposure [PC(S02)] was <0.5 ppm, with 2
subjects (7.4%>) experiencing moderate decrements in lung function following exposure to
concentrations of SO2 at or below 0.3 ppm (see Figure 3-1). For approximately 15% of the
subjects, the PC(S02) was > 2 ppm, with approximately 35% of asthmatic subjects experiencing
a doubling in sRaw versus clean air at < 0.6-ppm SO2.
May 2008
3-11
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
0,75 1.0
PC(S02) (ppm)
Figure 3-1. Distribution of individual airway sensitivity to S02. Each data point represents the
value of PC(S02) for an individual subject. PC(S02) is defined as the concentration of S02 which
resulted in a doubling of sRaw compared to clean air exposure.
Source: Horstman et al. (1986).
It is important to note that a transient decrement in lung function following exposure to an
air pollutant is not automatically considered to represent an adverse effect. However,
S02-induced decrements in lung function (increased sRaw and decreased FEVi) have frequently
been associated with increases in respiratory symptoms among asthmatics (Balmes et al., 1987;
Gong et al., 1995; Linn et al., 1987; 1988; 1990; 1983; Roger et al., 1985), which together does
constitute an adverse effect under the ATS guidelines. Linn et al. (1987) exposed 40 mild and
moderate asthmatics during 10 min periods of exercise to 0, 0.2, 0.4, and 0.6 ppm SO2. The
effect of SO2 on lung function and respiratory symptoms was assessed immediately following
exposure, and the individual-specific results have been made available to the U.S. EPA by the
study authors (Smith, 1994). Following exposure to 0.6 ppm S02 and after adjusting for effects
of exercise in clean air, 21 of the 40 subjects demonstrated moderate or greater decrements in
lung function, defined as a >15% decrease in FEVi, a >100% increase in sRaw, or both. Of these
May 2008
3-12
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
21 responders, 14 (67%) also experienced mild to severe respiratory symptoms (6 mild, 6
moderate, and 2 severe). In the same study, 14 asthmatics experienced moderate or greater
decrements in lung function at 0.4 ppm S02, 5 of whom (36%) also experienced mild to
moderate respiratory symptoms (2 mild, 3 moderate). Five asthmatics experienced moderate or
greater decrements in lung function at the lowest SO2 concentration tested (0.2 ppm), with 1 of
the 5 (20%) also experiencing mild respiratory symptoms.
It has been proposed that, as in asthmatics, individuals with COPD may also be more
susceptible to SCVinduced respiratory health effects. However, this group has not been
extensively studied in human clinical studies. Among a group of older adults with physician-
diagnosed COPD, Linn et al. (1985) reported no significant effect on lung function following 15
min exposures to S02 at concentrations of 0.4 and 0.8 ppm. While it was concluded that older
adults with COPD appear to be less sensitive to SO2 when compared with younger adult
asthmatics, the authors suggested that the lack of response may have been due in part to the very
low levels of exercise used in the study (yE = 18 L/min), which would result in a lower dose of
SO2 reaching the lower airway. In contrast to studies with asthmatics, most of the subjects in this
study regularly used bronchodilators and were permitted their use up to 4 h prior to the study.
In summary, S02-induced decrements in lung function have been observed following peak
exposures in humans. These effects are particularly evident in exercising asthmatic individuals,
with significant decreases in sRaw and increases in FEVi consistently demonstrated following
5-10 min exposures to 0.4-0.6 ppm SO2. S02-induced decrements in lung function have
frequently been associated with respiratory symptoms, and with increasing SO2 exposure
concentration from 0.2-1.0 ppm, both the magnitude of response among asthmatics and the
percentage of asthmatics significantly affected have been shown to increase.
3.1.3.3. Airway Inflammation
Avery limited number of human clinical studies have investigated the role of airway
inflammation in the asthmatic response following peak exposure to SO2. Gong et al. (2001)
observed an S02-induced increase in sputum eosinophil counts in exercising asthmatics 2 h after
a 10 min exposure to 0.75 ppm SO2. The results of this study provide some evidence that SO2
may elicit an allergic inflammatory response in the airways of asthmatics which extends beyond
the short time period typically associated with S02 effects.
May 2008
3-13
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
3.1.3.4. Evidence of the Effect of Peak Exposure from Animal Studies
In addition to the findings of human clinical studies involving asthmatics, S02-induced
decrements in lung function have been demonstrated following peak exposures to SO2 in
laboratory animals. The 1982 AQCD reported bronchoconstriction, as indicated by increased
pulmonary resistance, as the most sensitive indicator of lung function effects of acute SO2
exposure based on the observations of increased pulmonary resistance in guinea pigs that were
acutely exposed to 0.16 ppm SO2. Since 1982, a few new animal toxicological studies have
demonstrated acute changes in lung function following SO2 exposures of 45 min or less. These
studies are summarized below and in Annex Table E-l.
Lewis and Kirchner (1984) measured lung function in dogs exposed for 5 min to two doses
of SO2 via an endotracheal tube. Increased pulmonary resistance and decreased compliance were
observed in conscious dogs exposed to 30 ppm SO2, but not to 10 ppm SO2.
All other studies focused on the role of local nervous system reflexes and/or C-fiber
receptors in mediating responses to S02. Barthelemy et al. (1988) measured lung function in
anesthetized rabbits exposed for 45 min by endotracheal tube to two doses of SO2. Airway
resistance increased 16% and 50% following 0.5 and 5 ppm SO2, respectively. Bivagal vagotomy
had little effect on the response to 5 ppm, indicating that the prolonged bronchoconstriction
response did not result from a vagal reflex. This study did not rule out the possibility that vagal
reflexes were involved in immediate bronchoconstriction following S02 exposure.
In another study, Atzori et al. (1992a) demonstrated bronchoconstriction, as measured by
changes in dynamic lung compliance and airway conductance, within the first 5 min following
exposure of isolated and perfused guinea pig lungs to 100 and 250 ppm SO2 via an endotracheal
tube. This response was found to be due to a local nervous system reflex. However, this result
does not preclude involvement of central nervous system reflexes in SCVinduced
bronchoconstriction under conditions of an intact vagus nerve. Furthermore, the formation of
sulfite was observed in perfusate following SO2 exposure. Using the same model, Atzori et al.
(1992b) found that S02-induced bronchoconstriction was associated with the release of a sensory
neuropeptide and was inhibited when C-fiber receptors were blocked.
Other animal toxicological studies examined immediate respiratory effects from exposure
to very high SO2 concentrations. Hajj et al. (1996) exposed anesthetized guinea pigs to six tidal
breaths of 500-2,000 ppm S02. Increased total pulmonary resistance, decreased dynamic
May 2008
3-14
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
compliance, and systemic hypotension were observed within seconds. Tachykinin antagonists
blocked the changes in lung airway function, but not the changes in blood pressure in this model
system. Atropine failed to block the airway response. These results suggest that a local nervous
system reflex involving tachykinin release is an important mediator of bronchoconstriction
following high concentrations of SO2. Wang et al. (1996) exposed anesthetized rats to two tidal
breaths of 0.5% SO2 via an endotracheal tube. Immediate and transient bradypnea and
bradycardia were observed. Selective block of the C-fiber receptors and bilateral vagotomy
eliminated the SCVmediated effect on ventilation.
3.1.3.5. Summary of Evidence on the Effect of Peak Exposure on Respiratory
Health
Collectively, evidence from earlier studies considered in the previous review, along with a
limited number of new human clinical studies, consistently indicates that with elevated
ventilation rates, asthmatic individuals experience moderate or greater decrements in lung
function, as well as increased respiratory symptoms, following peak exposures to SO2 at
concentrations as low as 0.4-0.6 ppm (Balmes et al., 1987; Gong et al., 1995; Horstman et al.,
1986; Linn et al., 1987; Linn et al., 1983). These findings are consistent with our understanding
of the potential modes of action for respiratory health as described in Section 3.1.2. Some
sensitive asthmatics have been shown to experience moderate decrements in lung function at
concentrations below 0.3 ppm (Balmes et al., 1987; Linn et al., 1987; Sheppard et al., 1981),
although there is limited evidence of a significant increase in respiratory symptoms at these
exposure concentrations. Among asthmatics, both the magnitude of SCVinduced decrements in
lung function and the percent of individuals affected have consistently been shown to increase
with increasing exposure to SO2 concentrations between 0.2 and 1.0 ppm. This is summarized in
Table 3-1 along with supporting evidence of S02-induced increases in respiratory symptoms at
various exposure concentrations. The table includes data from all studies where individual data
are presented or have been made available by the authors (Smith, 1994). Although the vast
majority of human clinical studies involving controlled exposure to SO2 have been conducted in
adult asthmatics, there is a relatively strong body of evidence to suggest that adolescents may
experience many of the same respiratory effects at similar SO2 exposure concentrations (Koenig
et al., 1981; 1983; 1987; 1988; 1990; 1992). It should be noted, however, that in all of these
May 2008
3-15
DRAFT—DO NOT QUOTE OR CITE
-------
studies involving adolescents, SO2 was administered via inhalation through a mouthpiece rather
than an exposure chamber. This exposure technique bypasses nasal absorption of SO2, likely
resulting in a relative increase of pulmonary S02 uptake (see Section 2.6.1).
Table 3-1. Percentage of asthmatic individuals in controlled human exposures experiencing
S02-induced decrements in lung function.
so2
EXPOSURE
DURATION
NO.
SUBJ
VENTILATION
(L/MIN)
LUNG
FUNCT
CUMULATIVE PERCENTAGE
OF RESPONDERS
(NUMBER OF SUBJECTS)1
RESPIRATORY
SYMPTOMS:
SUPPORTING
STUDIES
CONC
(ppm)
2 100% t
sRaw
2 200% t
2 300% t
REFERENCE
2 15% 4,
FEV,
2 20% si/
2 30% 4-
0.2
10 min
40
-40
sRaw
5% (2)
0
0
Linn et al. (1987)2
10 min
40
-40
FEV,
13% (5)
5% (2)
3% (1)
Linn et al. (1987)
Some evidence of
S02-induced
increases in
respiratory
symptoms in the
most sensitive
individuals: Linn
et al. (1987; 1988;
1990; 1984a; 1983),
Schacter et al.
(1984)
0.25
5 min
19
-50-60
sRaw
32% (6)
16% (3)
0
Bethel et al. (1985)
5 min
9
-80-90
sRaw
22% (2)
0
0
10 min
28
-40
sRaw
4% (1)
0
0
Roger et al. (1985)
0.3
10 min
20
-50
sRaw
10% (2)
5% (1)
5% (1)
Linn et al. (1988)3
10 min
21
-50
sRaw
33% (7)
10% (2)
0
Linn et al. (1990)3
10 min
20
-50
FEV,
15% (3)
0
0
Linn et al. (1988)
10 min
21
-50
FEV,
24% (5)
14% (3)
10% (2)
Linn et al. (1990)
0.4
10 min
40
-40
sRaw
23% (9)
8% (3)
3% (1)
Linn et al. (1987)
Stronger evidence
with some statisti-
cally significant
increases in respi-
ratory symptoms:
Balmes et al.
(1987)4, Gong etal.
(1995), Linn et al.
(1987; 1983), Roger
et al. (1985)
10 min
40
-40
FEV,
30% (12)
23% (9)
13% (5)
Linn et al. (1987)
0.5
5 min
10
-50-60
sRaw
60% (6)
40% (4)
20% (2)
Bethel et al. (1983)
10 min
28
-40
sRaw
21% (6)
4% (1)
4% (1)
Roger et al. (1985)
10 min
45
-30
sRaw
36% (16)
16% (7)
13% (6)
Magnussen et al.
(1990)4
0.6
10 min
40
-40
sRaw
35% (14)
28% (11)
18% (7)
Linn et al. (1987)
10 min
20
-50
sRaw
60% (12)
35% (7)
10% (2)
Linn et al. (1988)
Clear and consistent
increases in SO2-
induced respiratory
symptoms: Linn
et al.(1987; 1988;
1984a; 1990), Gong
et al. (1995),
Horstman et al.
(1988)
10 min
21
-50
sRaw
57% (12)
33% (7)
14% (3)
Linn et al. (1990)
10 min
40
-40
FEV,
53% (21)
45% (18)
20% (8)
Linn et al. (1987)
10 min
20
-50
FEV,
55% (11)
55% (11)
5% (1)
Linn et al. (1988)
10 min
21
-50
FEV,
45% (9)
35% (7)
19% (4)
Linn et al. (1990)
1.0
10 min
28
-40
sRaw
54% (15)
25% (7)
14% (4)
Roger et al. (1985)
10 min
10
-40
sRaw
60% (6)
20% (2)
0
Kehrl et al. (1987)
Data presented from all references from which individual data were available. Percentage of individuals who experienced greater than or equal to a 100, 200, or 300% increase
in specific airway resistance (sRaw), or a 15, 20, or 30% decrease in FEV-|. Lung function decrements are adjusted for effects of exercise in clean air.
2Responses of mild and moderate asthmatics reported in Linn etal. (1987) have been combined.
3Analysis includes data from only mild (1988) and moderate (1990) asthmatics who were not receiving supplemental medication,
indicates studies in which exposures were conducted using a mouthpiece rather than a chamber.
May 2008
3-16
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
In laboratory animals, SCVinduced decrements in lung function were observed following
peak exposures in several studies conducted since the last review. Most of these experiments
were designed to evaluate the mode of action underlying S02-mediated bronchoconstriction.
They used high concentrations of administered SO2, which in many cases were delivered using
an endotracheal tube. As a result, these studies are of limited usefulness in understanding the
effects of SO2 at or near ambient levels or under conditions of nasal breathing.
3.1.4. Respiratory Effects Associated with Short-Term (> 1 h) Exposure
3.1.4.1. Respiratory Symptoms
Consideration of the mode of action suggests that SO2 may contribute to respiratory
symptoms by stimulating mucus secretion and cough through activating central nervous system
reflexes. Recent studies in vitro have demonstrated increased expression of a gene encoding
mucin protein, MUC5 AC, in human bronchial epithelial cells following exposure to the SO2
derivatives sulfite and bisulfite at concentrations of 1-10 |iM (Li and Meng, 2007). Increased
levels of MUC5AC protein were also reported. Sulfite and bisulfite were used, since SO2
dissolves into the aqueous fluid and forms hydrogen ions and bisulfite and sulfite anions when it
contacts the fluids lining the airway. These same investigators conducted a related in vivo study
in which rats were exposed by inhalation to 2 ppm SO2 for 1 h per day for 7 days. Rats which
were sensitized and challenged with ovalbumin, as well as exposed to SO2, had increased
MUC5AC mRNA and protein levels compared with animals treated with ovalbumin or S02
alone (Li et al., 2007b). Further studies are required to determine the relevance of mucin gene
expression to mucous secretion and respiratory symptoms in allergic and non-allergic animals at
ambient levels of SO2. However, evidence from toxicological studies such as these may provide
biological plausibility for the effects of S02 on respiratory symptoms in humans.
Epidemiological studies have examined the association between ambient SO2
concentrations and respiratory symptoms in both adults and children. In air pollution field
studies, respiratory symptoms are usually assessed using questionnaire forms (or "daily diaries")
completed by study subjects. Questions address the daily experience of coughing, wheezing,
shortness of breath (or difficulty breathing), production of phlegm, and others.
May 2008
3-17
DRAFT—DO NOT QUOTE OR CITE
-------
3.1.4.1.1. Children
1 Epidemiological studies on respiratory symptoms published since the last review are
2 summarized in Annex Table F-l; key studies are discussed in detail below.
Locations Pollutants
All 8 urban areas S02 only
S02 adjusting for 03
7 urban areas S02 only
S02 adjusting for 03 and N02
3 urban areas S02 only
S02 adjusting for 03 N02 and PM,0
0.90 1.00 1.10 1.20 1.30 1.40 1.50 1.80 1.70 1.80
Odds Ratio
Figure 3-2. Odds ratios (95% CI) for incidence of morning asthma symptoms of 846 asthmatic
children from the National Cooperative Inner-City Asthma Study. Effects associated with a 20 ppb
increase in 3-h avg S02 with a lag of 1-2 day moving average are presented. S02 effect estimates
from single- and multipollutant models are shown.
Source: Mortimer et al. (2002).
3 The strongest epidemiological evidence for an association between respiratory symptoms
4 and exposure to ambient SO2 comes from two large U.S. multicity studies (Mortimer et al., 2002;
5 Schildcrout et al., 2006). Mortimer et al. examined 846 asthmatic children from eight U.S. urban
6 areas in the National Cooperative Inner-City Asthma Study (NCICAS) for summertime air
7 pollution-related respiratory symptoms. Median 3-h avg SO2 (8 to 11 a.m.) levels ranged from 17
8 ppb in Detroit, MI to 37 ppb in East Harlem, NY. Morning symptoms were found to be most
9 strongly associated with an average of a 1- to 2-day lag of SO2 concentrations. In multipollutant
10 models with 03 and N02 (measured in seven cities), the S02 association remained robust (see
May 2008
3-18
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Figure 3-2). When particulate matter with an aerodynamic diameter of < 10 |im (PMio) was also
included in the multipollutant models, the SO2 effect estimate decreased only slightly; however,
it became nonsignificant, possibly due to reduced statistical power (only three of eight cities
were included in this analysis) or collinearity resulting from adjustment of multiple pollutants. A
similar decline was observed in the effect estimate for PM10 in the multipollutant model
compared to the single-pollutant model.
In the Childhood Asthma Management Program (CAMP) study, the association between
ambient air pollution and asthma exacerbations in children (n = 990) from eight North American
cities was investigated (Schildcrout et al., 2006). SO2 measurements were available in seven of
the eight cities. The median 24-h avg SO2 concentrations ranged from 2.2 ppb (interquartile
range [IQR]: 1.7, 3.1) in San Diego, CAto 7.4 ppb (IQR: 5.3, 10.7) in St. Louis, MO. Results for
the associations between asthma symptoms and all pollutants are shown in Figure 3-3. Analyses
indicate that although SO2 was positively related to increased risk of asthma symptoms at all
lags, only the 3-day moving average was statistically significant. No associations were observed
between S02 and rescue inhaler use. Stronger associations were observed for CO and N02. The
effect estimates appear to be slightly larger in joint-pollutant models with CO or NO2,
particularly at a 2-d lag, but did not change much when PM10 was jointly considered.
A longitudinal study of 1,844 schoolchildren during the summer from the Harvard Six
Cities Study suggested that the association between SO2 and respiratory symptoms could be
confounded by PMi0 (Schwartz et al., 1994). The median 24-h avg S02 concentration during this
period was 4.1 ppb (10th-90th percentile: 0.8, 17.9; max 81.9). SO2 concentrations were found
to be associated with cough incidence and lower respiratory tract symptoms. Of the pollutants
examined, PM10 had the strongest associations with respiratory symptoms. In two-pollutant
models, the effect of PMi0 was found to be robust to adjustment for other copollutants, while the
effect of SO2 was substantially reduced after adjustment for PMi0. Because the PM10
concentrations were correlated strongly to S02-derived sulfate particles (r = 0.80), the diminution
of the SO2 effect estimate may indicate that for PM10 dominated by fine sulfate particles, PM10
has a slightly stronger association than S02. This study further investigated the concentration-
response function and observed a nonlinear relationship between SO2 concentrations and
respiratory symptoms. Though an increasing trend was observed at concentrations as low as
May 2008
3-19
DRAFT—DO NOT QUOTE OR CITE
-------
1 10 ppb, no statistically significant increase in the incidence of lower respiratory tract symptoms
2 was seen until concentration exceeded a 24-h avg SO2 of 22 ppb.
Pollutants
SO,
S02 and CO
S02 and N02
S02 and PM
Lao
0
1
2
3-day moving sum
0
1
2
3-day moving sum
0
1
2
3-day moving sum
0
1
2
3-day moving sum
1,00 1,10
Odds Ratio
120
Figure 3-3. Odds ratios (95% CI) for daily asthma symptoms of 990 asthmatic children from the
Childhood Asthma Management Program Study. Effects associated with a 10 ppb increase in
within-subject concentrations of 24-h avg S02 are presented. Data collected from November 1993
to September 1995 were used. Results from single- and joint-pollutant models are shown.
Source: Schildcrout et al. (2006).
May 2008
3-20
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
In the Pollution Effects on Asthmatic Children in Europe (PEACE) study, a multicenter
study of 14 cities across Europe, the effects of acute exposure to various pollutants including SO2
on the respiratory health of children with chronic respiratory symptoms (n = 2,010) was
examined during the winter of 1993-1994 (Roemer et al., 1998). Mean 24-h avg SO2
concentrations ranged from 1 ppb in the urban area of Umea, Sweden, to 43 ppb in the urban
area of Prague, Czech Republic. No associations were observed between SO2 and daily
prevalence of respiratory symptoms or bronchodilator use at any of the single- and multiday lags
considered. In addition, no associations were observed for any of the other pollutants examined.
It should be noted that during the study period, there were only two major air pollution episodes,
at the beginning and end of the study period. In the epidemiological model, the control for time
trend was accomplished through the use of linear and quadratic terms. Given the timing of the air
pollution episodes, the quadratic trend term would have removed most of the air pollution effect.
Other studies that participated in the PEACE study and analyzed results for longer periods of
time have observed statistically significant associations between SO2 and respiratory symptoms
in children (van der Zee et al., 1999, presented below).
Additional studies have examined the relationship between respiratory symptoms and
ambient SO2 concentrations and generally found positive associations, including two U.S. studies
(Delfino et al., 2003; Neas et al., 1995) and several European studies (Hoek and Brunekreef,
1994; Peters et al., 1996; Roemer et al., 1993; Segala et al., 1998; Timonen and Pekkanen, 1997;
van der Zee et al., 1999). However, some did not find a consistent association between
respiratory symptoms and SO2 concentrations (e.g., Hoek and Brunekreef, 1993; 1995; Romieu
et al., 1996). Only one of these studies examined possible confounding of the SO2 effect by
copollutants. Van der Zee et al. (1999) looked at the association between respiratory symptoms
and S02 in 7- to 11-year-old children (n = 633) with and without chronic respiratory symptoms
in the Netherlands. Significant associations with lower respiratory tract symptoms and increased
bronchodilator use were observed for SO2, as well as PM10, BS, and sulfate, in symptomatic
children living in urban areas (n = 142). In a two-pollutant model with PM10, the results were
robust for bronchodilator use, but slightly reduced for lower respiratory tract symptoms. A
subgroup analysis of this cohort examining S02-related respiratory symptoms in children with
airway hyperresponsiveness and atopy (Boezen et al., 1999) is discussed in Section 3.1.4.4.
May 2008
3-21
DRAFT—DO NOT QUOTE OR CITE
-------
Mmsi
Schwartz et al. (1994)
Neasetal, {1995)
Ward et al. (2002a,b)
Loc«tkxi
8 U.S. cities
Unionlown, PA
Birmingham awl
SandweB, U.K.
•VanderZeeetat. (1999) 5 urban areas
in the Netherlands
*Roemeretal.{1993) The Netherlands
Population
Children {n = 300)
Non-asftmatics (*> = 38)
Children (n=162)
With chronic respiratory
symptoms (n=142)
With chronic respiratory
symptoms (n = ?3)
How and Bmnekreef (1994) The Netherlands Children (n = 1079)
Ward et al. {2002a,b) Birmingham and
SandweB, O.K.
Segala et al. (1338) Paris, France
Romieu et al, ("996) N, Mexico City,
Mexico
Children (n=182)
LIS
0
1-4
0
1
0-8
0
1
0-4
0
1
0-6
0
1
0-6
Mild asthmatics (n =43) 3
Mild asthmatics (n = 71) Not stated
Summer
0.25
0.5 0.75 1.0 1.25 1.5 1.75 2,02.252.5
Odds Ratio
Figure 3-4. Odds ratios (95% CI) for incidence of cough among children, grouped by season. For
single-day lag models, current day and/or previous day S02 effects are shown, except for Segala
et al. (1998), which only presented results for a 3-day lag. Multiday lag models represent the effect
of the mean concentration from the range of days noted. Risk estimates are standardized per
10 ppb increase in 24-h avg S02 level. The size of the box of the central estimate represents the
relative weight of that estimate based on the width of the 95% CI.
*Note that van der Zee et al. (1999) and Roemer et al.(1998) presented results for prevalence of cough.
1 Figure 3-4 and Figure 3-5 present the odds ratios for SCVrelated cough and lower
2 respiratory tract or asthma symptoms, respectively, from several epidemiological studies with
3 relevant data. The results for cough are somewhat variable with wide confidence intervals, as
4 shown in Figure 3-4. The studies conducted in the summer generally indicate increased risk of
5 cough from exposure to S02. A more consistent effect of S02 is observed on lower respiratory
6 tract or asthma symptoms (Figure 3-5). Although there is some variability in the individual effect
7 estimates, the majority of the odds ratios appear to be greater than one. As was the case with
May 2008
3-22
DRAFT—DO NOT QUOTE OR CITE
-------
1 cough, stronger associations with lower respiratory tract or asthma symptoms were observed in
2 the summer, as opposed to the winter. There was some variability among the different lags of
3 exposure; however, effects were generally observed with current day or previous day exposure
4 and, in some cases, with a distributed lag of 2 to 3 days.
SMttmsa
Mortimer ef at. (2002)
Schwartz el at. (1994)
Location
8 U.S. cities
8 U.S. cites
•Van (tor Zee at ai, (1933) 5 urban areas
in the Netherlands
Asthmatics (n = 848)
Children (n = 300)
With cliron c respiratory
symptoms (n = 142)
•Roemerelal. (1993} Tta Nettertands With chronic respiratory
symptoms (n = 73)
Hoe* and Brunekreet (1394) The Ntthsrlands Children (n = 1079)
Sdgaia et al. {1398) Paris, France (Aid asthmatics (n =43)
'Scttldcrout et al. (2008) 7 U.S. Cities Asthmatics (n = 881)
Us
1-2
1
0
1
04
0
1
0-6
0
1
0
1
0-2
Romieuetal. (1996)
N. Mexico City,
Mtxrco
Mild asthmatics (n = 71) Not stated
Summer
Al year
0.8
10
1.2 14 1.6
Odds Ratio
1.8 2.0 2.2 2.4 2.6
Figure 3-5. Odds ratios (95% CI) for the incidence of lower respiratory tract or asthma symptoms
among children, grouped by season. Risk estimates are standardized per 10 ppb increase in
24-h avg S02 level. For single-day lag models, current day and/or previous day S02 effects are
shown. Multiday lag models represent the effect of the moving average from the range of days
noted. The size of the box of the central estimate represents the relative weight of that estimate
based on the width of the 95% CI.
*Note that van derZee et al. (1999), Roemeret al. (1998) and Schildcrout etal. (2006) presented results for
prevalence of cough.
5 Overall, recent epidemiological studies provide evidence for an association between
6 ambient S02 exposures and increased respiratory symptoms in children, particularly those with
7 asthma or chronic respiratory symptoms. Recent U.S. multicity studies observed significant
May 2008
3-23
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
associations between SO2 and respiratory symptoms at a median range of 17 to 37 ppb (75th
percentile: -25 to 50) across cities for 3-h avg SO2 (NCICAS, Mortimer et al., 2002) and 2.2 to
7.4 ppb (90th percentile: 4.4 to 14.2) for 24-h avg S02 (CAMP, Schildcrout et al., 2006).
However, an earlier study that examined the concentration-response function found that a
statistically significant increase in the incidence of lower respiratory tract symptoms was not
observed until concentrations exceeded a 24-h avg SO2 of 22 ppb, though an increasing trend
was observed at concentrations as low as 10 ppb (Harvard Six Cities Study, Schwartz et al.,
1994). In the limited number of studies that examined potential confounding by copollutants
through multipollutant models, the SO2 effect was generally found to be robust after adjusting
for PM and other copollutants. More details of the literature published since the last review are
found in Annex Table F-l.
3.1.4.1.2. Adults
Compared to the number of studies conducted with children, fewer epidemiological studies
were performed that examined the effect of ambient S02 exposure on respiratory symptoms in
adults. Most of these studies focused on potentially susceptible populations, i.e., those with
asthma or COPD. One of the larger studies was conducted by van der Zee et al. (2000) in 50- to
70-year-old adults, with (n = 266) and without (n = 223) chronic respiratory symptoms in the
Netherlands. In adults both with and without chronic respiratory symptoms, no consistent
associations were observed between S02 levels and respiratory symptoms or medication use. A
subgroup analysis of this cohort examining SCVrelated respiratory symptoms in individuals with
airway hyperresponsiveness and atopy (Boezen et al., 2005) is discussed in Section 3.1.4.4.
Studies by Desqueyroux et al. (2002b; 2002a) examined the association between air
pollution and respiratory symptoms in other potentially susceptible populations, i.e., those with
severe asthma (n = 60, mean age 55 years) and COPD (n = 39, mean age 67 years), in Paris,
France. The mean 24-h avg SO2 concentration was 3 ppb (range: 1, 10) in the summer and 7 ppb
(range: 1, 31) in the winter. No associations were observed between SO2 concentrations and the
incidence of asthma attacks or episodes of symptom exacerbation in severe asthmatics or
individuals with COPD. O3 was found to have the strongest effect in these studies.
Several other European studies did observe an association between ambient SO2
concentrations and respiratory symptoms in adults with asthma or chronic bronchitis (Higgins et
May 2008
3-24
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
al., 1995; Neukirch et al., 1998; Peters et al., 1996; Taggart et al., 1996). Only one of these
studies examined possible confounding of the association by copollutants. Higgins et al.
examined the effect of summertime air pollutant exposure on respiratory symptoms in 62 adults
with either asthma, COPD, or both. The max 24-h avg SO2 level was 45 ppb. An association was
observed between SO2 and symptoms of wheeze, and it remained robust after adjustment for O3
and NO2. The effects of PM were not examined in this study.
Results from the epidemiological studies examining the association between S02 and
respiratory symptoms in adults are generally mixed, with some showing positive associations
and others finding no relationship at current ambient levels. The overall epidemiological
evidence that 24-h avg SO2 exposures at or near ambient concentrations has an effect on adults is
inconclusive. However, as discussed in Section 3.1.3.1, human clinical studies have observed an
effect of peak exposures to SO2 on respiratory symptoms, particularly among SCVsensitive
asthmatics, with 10 min exposures to SO2 concentrations as low as 0.4-0.6 ppm under exercise
conditions. These effects in clinical studies are at levels that have sometimes been measured in
ambient air for similarly short-time durations.
3.1.4.2. Lung Function
The 1982 AQCD reported bronchoconstriction, indicated by increased pulmonary
resistance, as the most sensitive indicator of lung function effects of acute S02 exposure, based
on the observations of increased pulmonary resistance in guinea pigs that were acutely exposed
to 0.16 ppm SO2. Since then, only a few new animal toxicological studies have measured lung
function at or near ambient levels of SO2. These studies, and those using higher concentrations of
SO2, are summarized in Annex Table E-4. Increased pulmonary resistance and decreased
dynamic compliance were observed in conscious guinea pigs exposed to 1 ppm S02 for 1 h
(Amdur et al., 1983). Effects were seen immediately after exposure and were not present 1 h
post-exposure. No changes in tidal volume, minute volume or breathing frequency were found.
These same investigators also exposed guinea pigs to 1 ppm SO2 for 3 h/day for 6 days (Conner
et al., 1985). No changes were observed in pulmonary function or respiratory parameters, i.e.,
diffusing capacity for carbon monoxide, functional reserve capacity, vital capacity, total lung
capacity, respiratory frequency, tidal volume, pulmonary resistance or pulmonary compliance. In
another study, Barthelemy et al. (1988) demonstrated a 16% increase in airway resistance
May 2008
3-25
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
following a 45-min exposure of anesthetized rabbits to 0.5 ppm SO2 via an endotracheal tube.
This latter exposure is more relevant to oronasal than nasal breathing.
3.1.4.2.1. Children
Most epidemiological studies discussed in the previous section on respiratory symptoms
also examined lung function. In these studies self-administered PEF meters were primarily used
to assess lung function. PEF follows a circadian rhythm, with the highest values found during the
afternoon and lowest values during the night and early morning (Borsboom et al., 1999).
Therefore, these studies generally analyze PEF data stratified by time of day. The
epidemiological studies on lung function are summarized in Annex Table F-l.
Mortimer et al. (2002) examined 846 asthmatic children from eight U.S. urban areas in the
NCICAS for changes in PEF related to air pollution. The mean 3-h avg SO2 was 22 ppb across
the eight cities during the study period of June through August, 1993. No associations were
observed between SO2 concentrations and morning or evening PEF. Of all the pollutants
examined, including PMi0, 03, and N02, only 03 was associated with changes in morning PEF.
In another U.S. study (Neas et al., 1995), 83 children from Uniontown, PA reported twice-
daily PEF measurements during the summer of 1990. The mean daytime 12-h avg SO2
concentration was 14.5 ppb (max 44.9). No associations were observed between daytime 12-h
avg S02 concentrations and mean deviation in evening PEF, even after concentrations were
weighted by the proportion of hours spent outdoors during the prior 12-h. Statistically significant
associations were observed for O3, total sulfate particles, and particle-strong acidity.
A study by van der Zee et al. (1999) observed associations between ambient SO2
concentrations and daily PEF measurements in 7- to 11-year-old children (n = 142) with chronic
respiratory symptoms living in urban areas of the Netherlands. The OR for a > 10% decrement in
evening PEF per 10 ppb increase in 24-h avg SO2 was 1.20 (95% CI: 0.97, 1.47) with same-day
exposure. A greater effect was observed at a 2-day lag, OR = 1.40 (95% CI: 1.18, 1.67), and this
effect remained robust in a two-pollutant model with PM10, OR = 1.34 (95% CI: 1.08, 1.64).
Multipollutant analyses also were conducted in a study by Chen et al. (1999), which
examined the effects of short-term exposure to air pollution on the pulmonary function of
895 children, ages 8 to 13 years, in three communities in Taiwan. The daytime 1-h max SO2 the
day before spirometry ranged from 0 to 72.4 ppb. In a single-pollutant model, 1-h max SO2
May 2008
3-26
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
concentration at a 2-day lag was significantly associated with FVC, -50.80 mL (95% CI: -97.06,
-4.54), or a 2.6% decline, per 40 ppb 1-h max SO2. However, in multipollutant models, authors
noted that only 03 remained significantly associated with FVC and FEVi. Effect estimates for
SO2 in multipollutant models were not provided.
While additional studies have observed associations between ambient SO2 concentrations
and changes in lung function in children (Hoek and Brunekreef, 1993; Roemer et al., 1993;
Peters et al., 1996; Segala et al., 1998; Timonen and Pekkanen, 1997), several other studies did
not find a significant association between SO2 and lung function parameters. In addition, in
studies that did observe an association, the correlations between SO2 and other pollutants,
particularly PM indices, were high [for example, r = 0.8-0.9 in Peters et al. (1996), making it
difficult to separate the contributions of individual pollutants.
In conclusion, while some epidemiological studies observed a positive association between
short-term SO2 exposure and lung function in children, several others, including a large U.S.
multicity study, did not observe such an association. The limited evaluation of potential
confounding by copollutants also indicated mixed results. Overall, the evidence is insufficient to
conclude that short-term exposure to ambient SO2 has an independent effect on lung function in
children.
3.1.4.2.2. Adults
Only a limited number of epidemiological studies have been conducted examining the
association between ambient SO2 concentrations and lung function in adults, as in the case of
respiratory symptoms. In a cross-sectional survey, Xu et al. (1991) investigated the effects of
indoor and outdoor air pollutants on the respiratory health of 1,140 adults (aged 40 to 69 years)
living in residential, industrial, and suburban areas of Beijing, China. The annual mean
concentrations of SO2 in residential, industrial, and suburban areas from 1981 to 1985 were 49
ppb, 22 ppb, and 7 ppb, respectively. Log-transformed SO2 and TSP were significantly
associated with reductions in FEVi and FVC. The authors cautioned that since SO2 and TSP
concentrations were strongly correlated, the effect of S02 could not be separated.
Van der Zee et al. (2000) observed an association between SO2 and morning PEF in 50- to
70-year-old adults (n = 138) with chronic respiratory symptoms living in urban areas of the
Netherlands. No associations were observed with evening PEF. The OR for a > 20% decrement
May 2008
3-27
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
in PEF was 1.21 (95% CI: 0.76, 1.92) per 10 ppb increase in 24-h avg SO2 with same-day
exposure and 1.56 (95% CI: 1.02, 2.39) at a 1-day lag. No associations were observed for a
> 10%) decrement in PEF. The authors hypothesized that while S02 level did not have much
effect on PEF in most subjects, a small subgroup of individuals experienced fairly large PEF
decrements when SO2 levels were high. No multipollutant analyses were conducted.
Higgins et al. (1995) examined the association between pulmonary function and air
pollution in 75 adults with either asthma, COPD, or both. Exposure to S02 was associated with
increased variation in PEF, but not with mean or minimum PEF. The SO2 effects on PEF
variation were robust to adjustment for O3 and NO2. Effects of PM were not considered.
Neukirch et al. (1998) also observed associations between lung function and SO2 concentrations
in a study of asthmatic adults in Paris, France; however, significant associations were found for
all pollutants examined, including BS, PM13, andNC>2. Other epidemiological studies observed
only weak relationships between ambient SO2 concentrations and lung function in adults (Peters
et al., 1996; Taggart et al., 1996).
Evidence from human clinical studies clearly indicates that asthmatic individuals
experience moderate or greater decrements in lung function, as well as increased respiratory
symptoms, following peak exposure (5-10 min) to SO2 (Balmes et al., 1987; Gong et al., 1995;
Horstman et al., 1986; Linn et al., 1987; 1983) These effects were seen at peak concentrations as
low as 0.4-0.6 ppm. However, in a human clinical study by Tunnicliffe et al. (2003) that
evaluated the effect of 1-h exposures to 0.2 ppm S02 in resting healthy and asthmatic subjects,
no significant changes were observed in lung function as measured by FEVi, FVC, and maximal
midexpiratory flow (MMEF).
In summary, the epidemiological studies examining adults do not provide strong evidence
for an association between short-term exposure to ambient S02 and lung function. While some
studies did observe associations between SO2 exposure and decrements in lung function
parameters, the strong correlation between SO2 and various copollutants in most studies, and the
lack of evidence evaluating potential confounding by copollutants, limit interpretation of
independent effects of S02 on lung function.
May 2008
3-28
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
3.1.4.3. Airway Inflammation
The animal toxicological studies on airway inflammation are summarized in Annex Table
E-l. In one study, guinea pigs were exposed to 1 ppm SO2 for 3 h/day for 5 days and
bronchoalveolar lavage was performed daily (Conner et al., 1985) No change in numbers of total
cells or neutrophils was observed. However, in two models of allergic sensitization, SO2
exposure increased airway inflammation. In one study (Park et al., 2001), guinea pigs were
exposed to 0.1 ppm SO2 for 5 h/day for 5 days and sensitized with 0.1% ovalbumin aerosols for
45 min on days 3-5. One week later, animals were subjected to bronchial challenge with 1.0%
ovalbumin and bronchoalveolar lavage and histopathologic examination were performed 24 h
later. Results demonstrated increased numbers of eosinophils in lavage fluid, and an infiltration
of inflammatory cells, bronchiolar epithelial cell damage and plugging of the airway lumen with
mucus and cells in the bronchial tissues of animals treated with both SO2 and ovalbumin, but not
in animals treated with ovalbumin or SO2 alone.
In a second study, rats which were sensitized and challenged with ovalbumin and exposed
to 2 ppm SO2 for 1 h/day for 7 days had an increased number of inflammatory cells in
bronchoalveolar lavage fluid and an enhanced histopathological response compared with those
treated with ovalbumin or SO2 alone (Li et al., 2007a). Similar responses were noted for ICAM-
1, a protein involved in regulating inflammation. Further experiments are required to determine
whether near ambient S02 also enhance inflammatory responses in non-allergic and allergic rats.
Taken together, these animal experiments suggest that near-ambient levels of SO2 may play a
role in exacerbating allergic responses.
In a human clinical study, Tunnicliffe et al. (2003) measured levels of exhaled NO (eNO)
in asthmatic and healthy adult subjects, before and after 1-h exposure to 0.2 ppm S02 under
resting conditions. While eNO concentrations were higher in the asthmatic than in healthy
subjects, no significant difference was observed between pre- and postexposure in either group.
One epidemiological study by Adamkiewicz et al. (2004) examined eNO as a biological
marker for inflammation in 29 older adults (median age 70.7 years) in Steubenville, OH. The
mean 24-h avg SO2 concentration was 12.5 ppb (IQR 11.5). The authors reported that, while
significant and robust associations were observed between increased daily levels of fine PM
(PM2.5) and increased eNO, no associations were observed with any of the other pollutants
examined, including S02, N02, and 03.
May 2008
3-29
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Overall, the very limited human clinical and epidemiological evidence does not indicate
that exposure to SO2 at current ambient concentrations is associated with inflammation in the
airway. However, toxicological studies suggest that repeated exposures to S02, at concentrations
as low as 0.1 ppm in guinea pigs, may exacerbate inflammatory responses in allergic animals.
3.1.4.4. Airway Hyperresponsiveness and Allergy
The toxicological studies describing S02-induced effects on airway obstruction,
hypersensitivity and/or allergy in guinea pigs and sheep are summarized in Annex Table E-3. In
one study, Amdur et al. (1988) exposed guinea pigs for 1 h to 1 ppm SO2 and measured airway
responsiveness to acetylcholine 2 h later. No airway hyperresponsiveness (AHR) was observed.
In a second study, Douglas et al., (1994) found no AHR following a histamine challenge 24 h
after exposure of rabbits to 5 ppm SO2 for 2 h. In a third study, exposure of sheep for 4 h to 5
ppm SO2 failed to result in AHR following carbachol (Gong et al., 2001). In a fourth study, a 5
min exposure to 30 ppm but not to 10 ppm SO2 resulted in AHR in horses challenged with
methacholine (Lewis and Kirchner, 1984). Collectively, these results show that a single exposure
to SO2 at a concentration of 10 ppm or less failed to induce AHR following challenge in 4
different animal models.
However, two other studies demonstrated increased airway responsiveness in guinea pigs
exposed repeatedly to S02 and allergen. Riedel et al. (1988) studied the effect of S02 exposure
on local bronchial sensitization to inhaled antigen. Guinea pigs were exposed by inhalation to
0.1, 4.3 and 16.6 ppm SO2 for 8 h/d for 5 days. During the last 3 days, SO2 exposure was
followed by exposure to nebulized ovalbumin for 45 min. Following bronchial provocation with
inhaled ovalbumin (0.1%) one week later, airway obstruction was measured by whole body
plethysmography. In addition, specific antibodies against ovalbumin were measured in serum
and bronchaolveolar fluids. Results show significantly higher bronchial obstruction in animals
exposed to SO2 (at all concentration levels) with ovalbumin compared with animals exposed
only to ovalbumin. In addition, significant increases in anti-ovalbumin IgG antibodies were
detected in bronchoalveolar lavage fluid of animals exposed to 0.1, 4.3 and 16.6 ppm S02 and in
serum from animals exposed to 4.3 and 16.6 ppm SO2 compared with controls exposed only to
ovalbumin. These results demonstrate that repeated exposure to SO2 can enhance allergic
sensitization in the guinea pig at a concentration as low as 0.1 ppm. In a second study, guinea
May 2008
3-30
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
pigs were exposed to 0.1 ppm SO2 for 5 h/day for 5 days and sensitized with 0.1% ovalbumin
aerosols for 45 min on days 3 to 5 (Park et al., 2001). One week later, animals were subjected to
bronchial challenge with 1.0% ovalbumin and lung function was evaluated 24 h later by whole
body plethysmography. Results demonstrated a significant increase in enhanced pause (Penh), a
measure of airway obstruction, in animals exposed to SO2 with ovalbumin but not in animals
treated with ovalbumin or SO2 alone. These experiments also indicate that near ambient levels of
S02 may play a role in exacerbating allergic responses in the guinea pig.
In a human clinical study evaluating SCVinduced AHR to an inhaled allergen (house dust
mite), Devalia et al. (1994) found that neither SO2 (0.2 ppm) nor NO2 (0.4 ppm) enhanced
sensitization to the allergen in asthmatic individuals. However, following concurrent exposure
(6 h) to S02 and N02 while at rest, subjects did exhibit increased sensitivity to the inhaled
allergen. In a subsequent study, Rusznak et al. (1996) confirmed these findings and observed that
the combination of SO2 and NO2 enhanced sensitization to house dust mite antigen up to 48
hours post-exposure.
A limited number of epidemiological studies also examined the association between S02
and AHR. Other studies considered individuals with AHR and atopy as a subgroup potentially
susceptible to S02-related health effects. These studies are summarized in Annex Table F-l.
S0yseth et al. (1995) investigated the effect of short-term exposure to SO2 and fluoride on the
number of capillary blood eosinophils, and the prevalence of AHR in schoolchildren, ages 7 to
13 years, (n = 620) from two regions in Norway, a valley containing an S02-emitting aluminum
smelter and a similar but nonindustrialized valley. The median 24-h avg SO2 concentration was
8 ppb (10th-90th percentile: 1, 33) in the exposed area and 1 ppb (10th-90th percentile: 0, 4) in
the nonindustrialized valley. The mean number of eosinophils was significantly greater in
children living near the aluminum smelter compared to the nonindustrialized area. However,
within children in the exposed area, a negative concentration-response relationship was observed
between mean eosinophils and previous-day 24-h avg SO2. The observed association between
SO2 and eosinophils was limited to atopic children. In children living in the exposed area, a
statistically significant positive association was observed between prevalence of AHR and
previous-day 24-h avg SO2 concentrations. Similar associations were observed for fluoride. The
authors hypothesized that recent exposure to SO2 may have induced changes in the airway
May 2008
3-31
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
leading to AHR, in addition to recruitment of eosinophils to the airways in atopic subjects.
Exposure to PM was not assessed in this study.
A study by Taggart et al. (1996) examined the effect of summertime air pollution levels in
northwestern England on AHR in nonsmoking, asthmatic subjects (n = 38) aged 18 to 80 years
who were determined to be methacholine (MCh) reactors. Subjects were tested multiple times,
for a total of 109 evaluable challenge tests, with a range of two to four tests per subject. The max
24-h avg S02 concentration during the study period was 40 ppb. This study reported that
24-h avg SO2 levels were marginally associated with a decreased dose of MCh required for a
20% drop in the postsaline FEVi (PD20FEVi).
Other epidemiological studies investigated the effect of exposure to SO2 on children and
adults with AHR and atopy. Boezen et al. (1999) examined 7- to 11-year-old children (n = 459)
in the Netherlands and tested them for AHR and atopy. These children were a subset of a larger
cohort examined in van der Zee et al. (1999). It was hypothesized that children with AHR, as
measured using a MCh challenge, and atopy, indicated by raised serum total IgE (> 60 kU/L, the
median value), may be susceptible to the effects of air pollution. One of the strengths of this
study was the use of AHR and serum IgE concentration to indicate susceptibility; these
measurements would be less prone to error than self-reported chronic respiratory symptoms. A
total of 121 children were found to have AHR and relatively high serum total IgE; 67 had AHR
and relatively low serum total IgE, 104 had no AHR but had a relatively high serum total IgE
concentration, and 167 were found to have neither AHR nor relatively high serum total IgE. For
the subset of children with relatively low serum total IgE with or without AHR, no associations
were observed between SO2 and any respiratory symptoms. However, for children with relatively
high serum total IgE either with or without AHR, the prevalence of lower respiratory tract
symptoms increased with increasing S02 concentrations. For children with AHR and relatively
high serum total IgE, the OR for the prevalence of lower respiratory tract symptoms was 1.70
(95% CI: 1.26, 2.29) with a 5-day moving average for every 10 ppb increase in SO2. For children
without AHR but with relatively high serum total IgE, the OR was 1.82 (95% CI: 1.33, 2.50)
with a 5-day moving average.
Boezen et al. (2005) conducted a similar study in 50- to 70-year-old adults (n = 327) in the
Netherlands. The subgroup of individuals with elevated serum total IgE, both with (n = 48) and
without (n = 112) AHR, were found to be more susceptible to air pollutants when contrasted with
May 2008
3-32
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
those who did not have elevated serum total IgE (n = 167). Significant associations were ob-
served between previous-day 24-h avg SO2 concentrations and the prevalence of upper respira-
tory tract symptoms in those with elevated serum total IgE. Stratified analyses by gender indi-
cated that, among those with AHR and elevated IgE, only males (n = 25) were at a higher risk for
respiratory symptoms. The OR for these males was 3.54 (95% CI: 1.79, 7.07) increase in
24-h avg SO2 for a 5-day moving average, compared with 1.05 (95% CI: 0.59, 1.91) for the
females.
In summary, the animal toxicological evidence suggests that repeated exposures to SO2 at
concentrations as low as 0.1 ppm in guinea pigs can exacerbate airway responsiveness following
allergic sensitization. Two new human clinical studies have demonstrated an increase in sensitiv-
ity to an inhaled allergen in asthmatic subjects following exposures to a combination of 0.2-ppm
SO2 and 0.4-ppm NO2. These findings are consistent with the very limited epidemiological evi-
dence that suggests that exposure to SO2 may lead to AHR in atopic individuals.
3.1.4.5. Respiratory Illness-Related Absences
An additional concern has been the potential for SO2 exposure to enhance susceptibility to,
or the severity of illness resulting from respiratory infections, especially in children. School
absenteeism is an indicator of morbidity in children caused by acute conditions. Respiratory
conditions are the most frequent cause, particularly influenza and the common childhood infec-
tious diseases. Park et al. (2002) examined the association between air pollution and school
absenteeism in 1,264 first- to sixth-grade students attending school in Seoul, Korea. The study
period extended from March 1996 to December 1999, with a mean 24-h avg SO2 concentration
of 9.19 ppb (SD 4.61). Note that analyses were performed using Poisson Generalized Additive
Model (GAM) with default convergence criteria. Same-day S02 concentrations were positively
associated with illness-related absences (16% excess risk [95% CI: 13, 22] per 10 ppb increase in
24-h avg SO2), but inversely associated with non-illness-related absences (9% decrease [95%
CI: 2, 15]). PM10 and O3 concentrations also were positively associated with illness-related ab-
sences. In two-pollutant models containing S02 and either PMi0 or 03, the S02 estimates were
robust.
A study by Ponka (1990) observed results that were consistent with those from the Park et
al. (2002) study. Ponka found that absenteeism due to febrile illnesses among children in day
May 2008
3-33
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
care centers and schools and in adults was significantly higher on days of higher SO2 concentra-
tions (> 8.1 ppb weekly mean of 1-h avg), compared to days of lower SO2 concentrations. In
addition, on days of higher S02 concentrations, the mean weekly number of cases of upper
respiratory tract infections and tonsillitis reported from health centers increased. Temperature,
but not NO2, was also found to be associated with febrile illnesses and respiratory tract infec-
tions. From these epidemiological studies, it is unknown whether SO2 increases susceptibility to
infection or whether its presence exacerbates preexisting morbidity following infection.
Pino et al. (2004) examined the association between air pollution and respiratory illnesses
in a cohort of 504 infants recruited at 4 months of age from primary health care units in
southeastern Santiago, Chile. The infants were followed through the first year of life. The mean
24-h avg S02 concentration was 11.6 ppb (5th-95th percentile: 3.0, 29.0). The most frequent
diagnosis during follow-up was wheezing bronchitis. No associations were observed between
current-day or previous-day SO2 and wheezing bronchitis, but with a 7-day lag, a 21% (95% CI:
8, 39) excess risk in wheezing bronchitis was observed per 10 ppb increase in 24-h avg SO2.
However, it should be noted that stronger associations were observed with PM2.5, which was
well-correlated with SO2 (r = 0.73). These epidemiologic studies are summarized in Annex
Table F-l.
To summarize, very few studies have examined the association between ambient SO2
concentrations and absences from school or work as a result of respiratory illnesses. The limited
evidence suggests a possible association between exposure to S02 concentrations and increased
respiratory illnesses, particularly among young children; however, this association was also seen
with PM, which was correlated with SO2.
3.1.4.6. Emergency Department Visits and Hospitalizations for Respiratory
Diseases
Total respiratory causes for ED visits and hospital admissions typically include asthma,
bronchitis and emphysema (collectively referred to as COPD), upper and lower respiratory tract
infections, pneumonia, and other minor categories. Temporal associations between ED visits or
hospital admissions for respiratory diseases and the ambient concentrations of SO2 have been the
subject of more than fifty peer-reviewed research publications since 1994. In addition to
considerable statistical and analytical refinements, recent studies have examined responses of
May 2008
3-34
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
morbidity in different age groups, the effect of seasons on ED and hospital usage, and
multipollutant models to characterize the effects of copollutant mixtures. The epidemiological
studies of ED visits and hospital admissions for respiratory causes are summarized in Annex
Table F-2.
3.1.4.6.1. All Respiratory Diseases
There are relatively few studies of ED visits for all respiratory causes in contrast to the
quantity of studies that examine hospital admissions for all respiratory causes. Collectively,
studies of ED visits and hospitalizations provide suggestive evidence of an association between
ambient SO2 levels and ED visits and hospitalizations for all respiratory causes. When analyses
were restricted by age, the results among children (0-14 years) and older adults (65+ years) were
mainly positive, though not all statisticallly significant. The studies that examined the association
of these outcomes and SO2 levels among adults (15-64 years) reported a mix of positive and
negative results. When all age groups were combined, the results of ED and hospitalization
studies were mainly positive; however, the excess risk estimates were generally smaller
compared to the children and older adults groups. It is possible that the effects observed in the
combined age groups were driven by increases in the very young or older adult subpopulations.
The results from the hospitalization and ED studies, separated by analyses among all ages and
age-specific analyses, are shown in Figure 3-6. Overall, the effect estimates in this figure range
from a -5% to 20% excess risk in ED visits or hospital admissions for respiratory causes per 10
ppb increase in 24-h avg SO2, with the large majority of studies suggesting an increase in risk.
Wilson et al. (2005) examined ED visits for all respiratory causes in Portland, ME from
1996-2000 and in Manchester, NH from 1998-2000. The mean 1-h max SO2 concentration in
Portland was 11.1 ppb (SD 9.1), and was higher during the winter months (mean 17.1 ppb (SD
12.0]) and lower in the summer (mean 9.1 ppb [SD 8.0]). In Manchester, the mean 1-h max SO2
concentration was 16.5 ppb (SD 14.7 ppb), and was higher in the winter months (mean 25.7 ppb
[SD 15.8]) as opposed to the summer months (mean 10.6 ppb [SD 15.1]). Though the authors
reported the 1-h max S02 concentrations, they used the 24-h avg S02 concentrations in their
analyses. When all ages where included in analyses, Wilson et al. found positive associations
between ED visits and SO2, with an 8% (95% CI: 3.0, 11) and 11% (95% CI: 0.0, 20.0) excess
May 2008
3-35
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
risk per 10 ppb increase in 24-h avg SO2 at a 0-d lag in Portland, ME and Manchester, NH,
respectively.
Peel et al. (2005) investigated ED visits for all respiratory causes in Atlanta, GAfrom
1993-2000. This study included 484,830 ED visits. The mean 1-h max SO2 concentration was
16.5 ppb (SD 17.1). The researchers found a weak positive relationship between ED visits and
SO2, though the increased risk was not statistically significant (1.6% [95% CI: -0.6, 3.8] excess
risk per 40 ppb increase in 1-h max S02). Tolbert et al. (2007) recently reanalyzed these data
with four additional years of data and found similar results. An analysis by Dab et al. (1996)
examined the association between SO2 and hospital admissions for all respiratory causes in Paris,
France, using both the 24-h avg and 1-h max. It should be noted that these researchers observed
similar effect estimates for both exposure metrics; however, only the estimate using 24-h avg
was statistically significant (1.1% [95% CI: 0.1, 2.0] excess risk per 10 ppb increase in 24-h avg
SO2 versus 1.9% [95% CI: -1.3, 5.0]) per 40 ppb increase in 1-h max SO2).
When analyses were stratified to include only children (0-14 years), evidence of a modest
association between S02 and ED visits or hospitalizations for all respiratory causes in children
was reported in several Australian (Barnett et al., 2005; Petroeschevsky et al., 2001) and
European (Anderson et al., 2001; Atkinson et al., 1999a; 1999b) studies. Excess risks ranging
from 3% to 22% per 10 ppb increase in 24-h avg SO2 were reported by these studies. In a
multicity study spanning Australia and New Zealand, Barnett et al. (2005) compared hospital
admission data collected from 1998-2001 with ambient S02 concentrations, where the mean
24-h avg SO2 concentration ranged from 0.9 to 4.8 ppb. The authors found a 5% (95% CI: 1, 9)
excess risk per 10 ppb increment in 24-h avg SO2 among children (1-4 years) in these cities.
However, some additional U.S. (Wilson et al., 2005), European (Fusco et al., 2001; Ponce de
Leon et al., 1996), and Latin American (Braga et al., 1999; 2001) studies did not find statistically
significant associations between ambient SO2 concentrations and hospitalizations for all
respiratory causes among children.
May 2008
3-36
DRAFT—DO NOT QUOTE OR CITE
-------
Wit-on Ptd f2005>
Wil on htal{?105l
Fe^l Pt sl i2DGbi
, Jffpp), Ql a\
A.th! et 4 \T^a9nj
Wtiion et alu'OUS)
Wi»%n ft ril
A^ktrsr mtdl
WiKrn ft dl f^OuSj
Wit^i n ^ 4 i^DG5^
Aft if * eh Ml 1 yT-b i
Wilson el si (200b)
Wilson st af (2005)
Atkinson et al, {1999b}
Lijginaah et d': {2005}
Petroeschevsky el al, (2001)
Hagen eia! (2000)
Oftedaietai, (2003)
Anderson el al. {2001}
Atkinson et al. {1999a)
Walters etai. {1994}
Ponce de Leon el al. (1996)
Dab et al, M996)
Uorca et si. (2006)
Fu^o et al. (200!:)
Bunetietal. (2001)
Wongetal {1399}
Gouveta and Fletcher [2000}
Baine'ieta!- (2035)
Lugin&a'n et a* (2005)
Petroeschevsfcy &. $ (20G1J
Anderson etas {2001}
Atkinson et al, (1999a)
Porsee de Leor e; ai (19981
r-ic^ et ?\ £200^
tn
"A i^qyUl |1 ^
L jqn f j >/0"i
PetroestH *» . (2001)
Schoutef f -i h
Srtfutes ct 3i i h
Sp^e!t |1fllb)
Ar-ae^oi ct a (2UQt)
Atkirson et -al {1999a)
Pniae de Leon et ai (1996)
Vfgoth pt ai J1&96)
Schwartz et a! (1995)
Schwartz ef al (1995)
Fung Bi a: (2006)
Vang et ai (2003)
Luginaab et a? {2005)
Schoulenslal. (*996)
Spfcelal. (1998)
Andersonetai. (2G01)
Atkinson et ai. {1999a)
Ponce de to eial (19&8)
Vigottiel al, {1098)
Wong e? al. (1999)
Locafeon
Portland, ME
Manchester, NH
Atlanta, GA
Multoty, Canada
London UK
Portland. Mt
Wanciester, toH
uidun Mr
Dcrtla^d ML
Marv'tetter MB
Ljr.iion Vb
c'< rtlatd Ml
Manchester NH
London UK
Windsor, ON
Brisbane, Australia
Drammen, Norway
Drammen, Norway
West Midlands, UK
London, UK
Birmingham, UK
London, UK
Paris Ranee
"orreSavsga, Spain
Rone Italy
Toronto ON
Horg Kong. Chif-a
3 m Pa-jfo, Brazil
Muteicsty, Alalia
Wsndsor, ON
Brisbane. Ausvalia
We$f Midlands. UK
London UK
London, UK
Hnne lay
V!ultr ty Amelia
Hsiry Kung Curs
Windsor ON
Lrsbdifs A strata
Arf^ wjrffl
"Miriam
Multii v taope
'/vest Midlands, UK
London, UK
London, UK
Milan italy
.New Haven CT
Tacoma. WA
Vancouver BC
Vancouver, BC
Windsor, ON
Other
Uh max
Multicity, Europe
West Midlands, UK
London, UK
London. UK
Milan. Italy
rior-g Kong, Cnina
Females
Mates
Summer
Winter
1-h max
Femates
Mates
Females
Msites
Females
Mates
ED Visits
j Older Adults {65
All ages
Hospital
Admissions
Older Adults (65+
I 75 2 00
i r
1 25 1 50
Relative risk
Figure 3-6. Relative risks (95% CI) of S02-associated emergency department visits and
hospitalizations for all respiratory causes among all ages and separated by age group. Risk
estimates are standardized per 10 ppb increase in 24-h avg S02 concentrations or 40 ppb increase
in 1-h max S02. The size of the box of the central estimate represents the relative weight of that
estimate based on the width of the 95% CI.
May 2008
3-37
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Wilson et al. (2005) found a positive association between ED visits and SO2, with a 16%
(95% CI: 8.0, 25.0) excess risk per 10 ppb increase in 24-h avg SO2 at a 0-d lag, and no
association in Manchester, NH when only older adults (65+ years) were considered. In another
two-city study, Schwartz (1995) compared 13,740 hospital admission in New Haven, CT and
Tacoma, WAfrom 1988-1990 with ambient SO2 concentrations. The mean 24-h avg SO2
concentration was 29.8 ppb (90th percentile: 159) in New Haven and 16.8 ppb (90th percentile:
74) in Tacoma. Schwartz found positive associations between hospitalizations and S02, with a
2% (95% CI: 1.0, 3.0) excess risk at a 2-d lag in New Haven and 3% (95% CI: 1.0, 6.0) excess
risk at a 0-d lag in Tacoma per 10 ppb increase in 24-h avg SO2. In two-pollutant models, the
SO2 effect estimate from New Haven, but not Tacoma, was found to be robust to adjustment for
PM10. Here, the term robust is used to indicate that there was little change in the magnitude of
the central estimate, though statistical significance may have been lost. In Vancouver, BC, both
Fung et al. (2006a) and Yang et al. (2003) also found positive associations between
hospitalizations and SO2. In a multipollutant model including coefficient of haze (CoH), NO2,
03, and CO, the S02 effect estimate diminished slightly (Jaffe et al., 2003).
Additional evidence of a positive association between ED visits or hospitalizations for all
respiratory causes among older adults and SO2 comes from several European (Spix et al., 1998;
Sunyer et al., 2003; Vigotti et al., 1996) and Australian (Petroeschevsky et al., 2001) studies.
Excess risks ranging from 1% to 12% per 10 ppb increase in 24-h avg SO2 were reported by
these studies. Petroeschevsky et al. (2001) examined 33,710 hospital admissions in Brisbane,
Australia from 1987-1994. The mean 24-h avg SO2 concentration was 4.1 ppb, and was highest
in the winter months (4.8 ppb) and lowest in the spring (3.7 ppb). Petroeschevsky et al. found a
12%) (95% CI: 2.0, 23.0) excess risk per 10 ppb increase in 24-h SO2 at 0-d lag. Additional
European studies did not find statistically significant associations between ambient S02
concentrations and hospitalizations for all respiratory causes among older adults (Anderson et al.,
2001; Atkinson et al., 1999b; Ponce de Leon et al., 1996; Schouten et al., 1996).
In summary, many studies show a small, positive, though not statistically significant
association between ambient S02 concentrations and ED visits and hospitalizations, particularly
among children and older adults (65+ years). The positive evidence from these studies is
supported by the results of panel, human clinical, and limited toxicological studies that also
found a positive relationship between SO2 levels and adverse respiratory outcomes.
May 2008
3-38
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
3.1.4.6.2. Asthma
Studies of ED visits and hospitalizations provide suggestive evidence of an association
between ambient SO2 levels and ED visits and hospitalizations for asthma. The results from the
hospitalization and ED studies, separated by analyses among all ages and age-specific analyses,
are shown in Figure 3-7. Overall, central effect estimates in the figure range from a -10% to 40%
excess risk in ED visits and hospitalizations for asthma per 10 ppb increase in 24-h avg S02.
Most of the effect estimates are positive (suggesting an association with SO2 and ED visits and
hospitalizations for asthma), though few are statistically significant at the 95% confidence level.
When all ages were included in the analyses, Wilson et al. (2005) found a positive association
between ED visits and S02, with a 10% (95% CI: 2.0, 20.0) excess risk per 10 ppb increase in
24-h avg SO2 at a 0-d lag in Portland, ME and a positive, though not statistically significant
association in Manchester, NH. Ito et al. (2003) found a 36% (95% CI: 1.23, 1.51) excess risk in
asthma ED visits per 10 ppb increase in 24-h avg SO2, though this association was diminished
once N02 was included in the model. A study conducted in (NY Dept of Health, 2006) found a
11%) (95%) CI: 6, 17) excess risk in asthma hospital admissions per 10 ppb increase in 24-h avg
SO2 for Bronx residents, but a null association for the residents of Manhattan. A study conducted
in Atlanta (Peel et al., 2005) found a null relationship between asthma ED visits and 1-h max
S02.
A study by Jaffe et al. (2003) examined the association between S02 and ED visits for
asthma in three cities in Ohio - Cincinnati, Cleveland, and Columbus - in asthmatics aged 5 to
34 years. The mean 24-h avg SO2 concentrations were 14 ppb (range: 1-50) in Cincinnati,
15 ppb (range: 1-64) in Cleveland, and 4 ppb (range: 0-22) in Columbus. A positive association
was observed in the multicity analysis, with a 6.1% (95% CI: 0.5, 11.5) excess risk in asthma
visits observed per 10 ppb increase in 24-h avg SO2. In the city-stratified analyses, significant
associations were only observed for Cincinnati (17.0% [95% CI: 4.6, 30.8]).
When analyses were stratified to include children (0-14 years) only, Wilson et al. (2005)
found positive, but not statistically significant associations between ED visits and S02 in
Portland, ME or Manchester, NH. Similarly, Lin et al. (2005) observed a weak positive
association between hospitalizations for asthma and SO2 among girls, and a null association for
boys (Toronto, ON; mean 24-h avg SO2 of 5.36 ppb [SD 5.90]). Stronger evidence comes from a
study of childhood asthma hospitalizations conducted in Bronx County, New York (Lin et al.,
May 2008
3-39
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
2003b). In this study, the authors conducted a case-control study of children aged 0-14 years and
examined the association of daily ambient SO2 concentrations (categorized into quartiles of both
average and max levels) and cases admitted to the hospital for asthma or controls who were
admitted for reasons other than asthma. The mean 24-h avg SO2 was below 17 ppb for both cases
and controls across all lag days examined. The authors found that cases were exposed to higher
24-h avg SO2 than controls. When the highest exposure quartile was compared with the lowest,
the ORs were strongest when a 3-day lag was employed (OR 2.16 [95% CI: 1.77, 2.65] for 24-h
avg SO2; OR 1.86 [95% CI: 1.52, 2.27] for 1-h max SO2). The results were positive and
statistically significant for all lag days examined. These results suggest a consistent positive
association between SO2 exposure and hospitalizations for childhood asthma.
Additional evidence of a positive association between ED visits or hospitalizations for
asthma and SO2 comes from several European (Anderson et al., 1998; Atkinson et al., 1999a;
Hajat et al., 1999; Sunyer et al., 1997; 2003; Thompson et al., 2001) and Asian (Lee et al., 2002)
studies. Excess risks ranging from 2% to 10% per 10 ppb increase in 24-h avg SO2 were reported
by these studies. Several of these studies observed that the S02 effect estimate was robust to
adjustment for BS and NO2 (Anderson et al., 1998; Sunyer et al., 1997), but one study observed
that the SO2 effect diminished considerably with adjustment for PM10 and benzene (Thompson et
al., 2001). Atkinson et al. (1999b) compared 165,032 hospital admissions in London from 1992-
1994 with ambient SO2 levels (mean 24-h avg of 7.2 ppb [SD 4.7]). They found a 10% (95% CI:
4.0, 16.0) excess risk per 10 ppb increase in 24-h avg S02 at 1-d lag. Additional European (Fusco
et al., 2001), Australian (Barnett et al., 2005; Petroeschevsky et al., 2001), Asian (Ko et al., 2007;
Lee et al., 2006) and Latin American (Gouveia and Fletcher, 2000) studies did not find
statistically significant associations between ambient SO2 concentrations and hospitalizations for
all respiratory causes among children.
In summary, small, positive associations were observed between ambient SO2
concentrations and ED visits and asthma hospitalizations. Evidence from these studies is further
supported by the results of panel and human clinical studies that have also found S02-related
respiratory effects in asthmatics.
May 2008
3-40
DRAFT—DO NOT QUOTE OR CITE
-------
RtftniEt
{,asa|i?n
Other
Lag
Wilson et al 12005)
Portland WE
2
Wilson etal I2O05)
Mcnchestef NH
2
Peel et a> (20GE!
Atlanta GA
1-h ma*
0-2
toe;a' i£0C7i
MevtYort NY
0-1
NY DOH w006t
Bum N>
0-4
Manhattan, NY
Atkinson ei al. (1999b)
London UK
1
Hajat et al. (1893)
Lonaon UK
GP Visits
0-2
Galan e! ai, (2003)
Madrid, Span
0
Jaffe et al. (2003)
Cinciftnali, OH
2
Me et al. (2003)
Cleveland. OH
2
Jaffe et at. (2003)
Columbus, OH
3
Jaffe et al. (2003)
Multeity, OH
KIR
Boutin-Forzano ei ai. (2004)
Marseille, France
0
Wilson it al, (2005)
Portland, ME
2
Wilson et al. (2005)
Manchester, WH
2
Sunyer et al, (1997)
Mufcctjr, Europe
0-3
Atkinson et al, |l999b)
London, UK
1
Hajat at al, (1999)
London. UK
GP Visits
0-3
CasHsague el ai. (19951
Barcelona, Spain
Summer
2
Winter
1
Tertias ef al, (19985
Valencia. Spain
>14
0
Wilson etal. I2005)
Portland. ME
2
Wilson et at. (2005!
Manchester; NH
2
Hajat etal (1999)
London, UK
GP Vis lis
0-1
Schouten et al, (1996)
Amsterdam
~o-7
Anderson et a! (1998}
London, UK
0-3
Atkinson et al. (1999a)
London, UK
1
Walters etal. (1994)
Birmingham, UK
Summer
0
Winter
Dab etal. (1998)
Paris. France
2
Fuscoetal (2001)
Rome, Italy
0
Tsai el si, (2008)
Kaohsiung, Taiwan
>25 C
0-2
<25 C
Wong etal. (1999)
Hong Kong, Chins
0
Ko et al. (2007a)
Hong Kong, China
0-3
Petroesehevsky et al. (2001)
BriStant Austral a
04
Lin et al. (2004)
Bron NY
10
Andereo'* e; al H998i
LnnA.n UK
0-3
Atkin-on eta1 <1 Wai
Lenowi UK
1
Fuscoetal <20011
Rent ital>
0
Barstl et al (2005)
M Jt.citv, Australia
0-1
Gouvtta ard f icicKer (2000)
Sae Pamo Bra/'l
2
BamrttetaUiOOSj
Multiuty Australia
0-1
Petroeschevsky et al. (2001)
Brisbane Austral a
1-h max
0-4
Lin etal (2003)
Toronto, ON
Boys
0
Gills
Lee et al. (2006)
Hong Kong, China
0
Peiroesctavsky et al. (2001)
Snsbane, Australia
24-h avg
1
Anderson etal. (1998)
London, UK
0-2
AtSmson et ai (19998)
Lonaon, UK
3
Koetal. |2007»)
Hong Kong. China
0-3
Sheppard etal. (1399)
Seattle- WA
0
Petroeschewkyetal [2001)
Brshare Australia
24-h avg
0
Anderson c! al 119981
London UK
0-3
Atkinson ei 31(1999ai
Lcn?cn UK
2
ED Visits
Hospital
Admissions
[A# ageSj
[Adults]
| Older Adults (65^71
rATajeil
[Children!
I Adults 1
OWefAduB*
(S5».S 1
0,75
1 00 1,25
Relative risk
1,50
Figure 3-7. Relative risks (95% CI) of S02-associated emergency department visits and
hospitalizations for asthma among all ages and age-specific groups. Risk estimates are
standardized per 10 ppb increase in 24-h avg S02 concentrations or 40 ppb increase in 1-h max
S02. The size of the box of the central estimate represents the relative weight of that estimate
based on the width of the 95% CI.
May 2008 3-41 DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
3.1.4.6.3. Chronic Obstructive Pulmonary Disease
There are relatively few studies that have examined the association of ED visits and
hospitalizations for COPD and ambient SO2 levels, and very little evidence that an association
exists. A recent study (Ko et al., 2007) found a significant association between hospital
admissions for COPD (not including asthma) in Hong Kong (1.8% [95% CI: 0.3, 3.8]) excess
risk per 10 ppb increase in 24-h avg S02 concentration). Three additional studies reported
positive and statistically significant results for COPD and SO2; all three studies included asthma
in their diagnostic definition of COPD (Anderson et al., 2001; Moolgavkar, 2003; Sunyer et al.,
2003). Anderson et al. (2001) reported a 12% (95% CI: 5.0, 20.0) excess risk per 10 ppb increase
in 24-h avg S02 among children, while Moolgavkar (2003) and Sunyer et al. (2003) found 5%
and 2% excess risks per 10 ppb increase in 24-h avg SO2 among older adults populations,
respectively. Other studies examining COPD did not find statistically significant results
(Atkinson et al., 1999b; Burnett et al., 1999; Michaud et al., 2004).
Overall, this limited and inconsistent evidence does not support a relationship between ED
visits and hospitalizations for COPD and ambient SO2 levels.
3.1.4.6.4. Respiratory Diseases Other than Asthma or COPD
Studies of ED visits or hospital admissions for other respiratory diseases looked at several
other specific outcomes. There are limited studies with mixed results for upper respiratory tract
infections (Burnett et al., 1999; Hajat et al., 2002; Lin et al., 2005; Peel et al., 2005), pneumonia
(Barnett et al., 2005; Moolgavkar et al., 1997; Peel et al., 2005), bronchitis (Barnett et al., 2005;
Michaud et al., 2004), and allergic rhinitis (Hajat et al., 1999; Villeneuve et al., 2006). The
limited evidence is suggestive of an association between S02 levels and ED visits for lower
respiratory tract diseases (Atkinson et al., 1999b; Farhat et al., 2005; Hajat et al., 1999; Lin et al.,
1999; Martins et al., 2002). All of the studies that characterized this relationship found a positive
and statistically significant excess risk associated with increases in SO2. Excess risks ranging
from 3% to 33% per 10 ppb increase in 24-h avg S02 were reported by these studies.
In summary, few studies provide results with mixed respiratory health outcomes other than
asthma and COPD. This makes it difficult to draw conclusions about the effects of SO2 on these
diseases. Limited evidence does exist to support a suggestive association between ambient SO2
levels and ED visits for lower respiratory tract diseases.
May 2008
3-42
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
3.1.4.6.5. Summary of Evidence on Emergency Department Visits and
Hospitalizations for Respiratory Diseases
Small, positive associations exist between ambient SO2 concentrations and ED visits and
hospitalizations for all respiratory causes, particularly among children and older adults (65+
years), and for asthma, though not always statistically significant. The S02-related changes in
ED visits or hospital admissions for respiratory causes ranged from -5% to 20% excess risk, with
the large majority of studies suggesting an increase in risk. No association was observed between
SO2 levels and ED visits and hospitalizations for COPD. Given the limited number of studies
with mixed results, it is difficult to draw conclusions about the effect of S02 on other respiratory
diseases, though studies of lower respiratory tract diseases are somewhat suggestive of an
association.
Multipollutant regression analyses indicate that SO2 risk estimates, in general, are not
sensitive to the inclusion of copollutants, including 03 (Anderson et al., 1998; Hajat et al., 1999;
Yang et al., 2003; 2005), PM (Hagen et al., 2000; Lin et al., 2003; 2005; Schwartz, 1995) CO
(Farhat et al., 2005) and NO2 (Anderson et al., 1998; Lin et al., 2004; Sunyer et al., 1997). Figure
3-8 presents SO2 excess risk estimates with and without adjustment for various copollutants. PM
and N02 are the main foci, since these pollutants have been found to be highly-correlated with
SO2 in epidemiological studies and have known respiratory health effects. Although the studies
showed that copollutant adjustment had varying degrees of influence on the SO2 effect estimates,
the effect of SO2 on respiratory health outcomes appears to be generally robust and independent
of the effects of ambient particles or other gaseous copollutants.
The results of several studies (Anderson et al., 1998; Hajat et al., 1999; Schouten et al.,
1996; Spix et al., 1998; Wong et al., 1999) have demonstrated a greater increase in ED visits and
hospitalizations for respiratory illnesses during the summer months, despite the fact that the
average concentrations for SO2 in some of areas studied were greatest in winter. In contrast,
some studies found the associations between ED visits and hospital admissions and respiratory
disease with similar increases in SO2 to be greater in winter than summer (Vigotti et al., 1996;
Walters et al., 1994). Other studies were unable to discern a seasonal difference in ED visits and
hospitalizations for respiratory causes (Castellsague et al., 1995; Tenias et al., 1998; Wong et al.,
2002). These effects were not consistent across age groups. Warmer months were more likely to
show evidence of an association with adverse respiratory outcomes in children, while older
May 2008
3-43
DRAFT—DO NOT QUOTE OR CITE
-------
1 adults appeared more likely to be affected during the cooler months. These seasonal associations
2 remain somewhat uncertain and require additional investigation.
Reference
location
Age
Outcome Lag
Pollutants
Galan et at (2003)
Madrid, Spain
All
Asthma
0
SOj
S02 ~ PMt0
Hajatetal, (1999)
London, UK
0-14
Asthma
1
S02
S02 + PM10
Hajatetal, (1999)
London, UK-
0-14
Asthma
1
sg2
so2+no2
Hajat etal, (1999)
London, UK
0-14
Asthma
1
SOj
SG2 + 03
Schwartz (1995)
New Haven, CT
65+
All Resp
0-1
sg2
SOj + PM1£i
Schwartz (1935)
Tacoma, WA
85+
All Resp
0-1
SO,
SQ2 + PM10
Anderson et al,
(1398)
London. UK
0-14 Asthma 1
Anderson et al, London, UK
(1998)
Sunyer et al. (1997) it*% - Europe 0-14 Asthma 1
Sunyeret ai, (1997) Multeity - Europe 0-14 Asthma 1
0-14 Asthma 1
All
Burnett et al. (2001)* Toronto, ON <2 AIResp 3
Yang etal. (2003) Vancouver, BC <3 AIResp 2
85+ 0
0-14 Asthma 1
Anderson et al, London, UK
(1398)
All
ED visits
Hospital
Ad missions
SO, + NO
SOj + 03
so, +0
so,+0
• Single pollutant
o Gopoliutant
so2+o
S02 + 03
1,00
Relative risk
Figure 3-8. Relative risks (95% CI) of S02-associated emergency department visits and
hospitalizations for all respiratory causes and asthma, with and without copollutant adjustment.
Risk estimates are standardized per 10 ppb increase in 24-h avg S02 concentrations or 40 ppb
increase in 1-h max S02. In Burnett et al. (2001), analyses were performed using default
convergence criteria for Poisson GAM with a nonparametric LOESS prefilter applied to air
pollution and hospitalization data.
May 2008
3-44
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
In conclusion, a large number of epidemiologic studies provide evidence of an association
between ambient SO2 concentrations and ED visits and hospitalizations for all respiratory causes,
in particular among children and older adults (65+ years), and for asthma. The findings are
generally robust when additional copollutants are included in the model. These associations are
supported by panel studies that observed SCVrelated increases in asthma and other respiratory
symptoms in children, and human clinical and animal toxicological studies that found a positive
relationship between S02 exposure and various respiratory outcomes.
3.1.4.7. Summary of Evidence on the Effect of Short-Term (> 1 h) Exposure on
Respiratory Health
Numerous epidemiological studies have observed associations between short-term (> 1-h,
generally 24-h avg) exposure to SO2 and respiratory health effects, ranging from respiratory
symptoms to ED visits and hospital admissions for respiratory causes. The associations between
ambient S02 concentrations and several respiratory outcomes were generally consistent, with the
large majority of studies showing positive associations, and multicity studies, as well as several
single-city studies, indicating statistically significant findings. The respiratory effects related to
short-term exposure to SO2 found in animal toxicological studies, and to a more limited extent
the human clinical studies, provide coherence and biological plausibility for the observed
epidemiological associations. The causal effects of peak exposure to SO2 on respiratory health
found in the human clinical studies (see Section 3.1.3.5) provide further evidence of biological
plausibility for the effects associated with short-term exposure to SO2.
Two recent multicity studies (Mortimer et al., 2002; Schildcrout et al., 2006) and several
other studies (Delfino et al., 2003; Neas et al., 1995; van der Zee et al., 1999) have found an
association between short-term ambient SO2 concentrations and respiratory symptoms in
children. In the limited number of studies that assessed potential confounding by copollutants
using multipollutant models, the SO2 effect on respiratory symptoms was generally found to be
robust to adjustment for copollutants. These findings provide supportive evidence for an
association between short-term exposure to ambient SO2 exposure and respiratory symptoms in
children, particularly those with asthma. Several recent studies (Desqueyroux et al., 2002a;
2002b; van der Zee et al., 2000) found no association between ambient SO2 levels and
respiratory symptoms in adults, though there was limited epidemiological evidence which
May 2008
3-45
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
suggested that atopic adults as well as children may be at increased risk for SCVinduced
respiratory symptoms (Boezen et al., 1999; 2005).
Animal toxicological studies in guinea pigs showed changes in lung function immediately
following 1 ppm SO2 exposure (Amdur et al., 1983). Guinea pigs, as a species, are typically more
sensitive to air pollution than other laboratory animals and, thus, may provide a better model for
characterizing the effects of air pollutants on lung function. Epidemiological studies do not
provide strong evidence of an association between short-term ambient S02 exposure and lung
function in either children (Mortimer et al., 2002; Roemer et al., 1998) or adults (e.g., Peters et
al., 1996; Taggart et al., 1996). Several other studies reported positive results; however, the
generally mixed findings, as well as the relative lack of evidence available to evaluate potential
confounding by copollutants, limits the causal interpretation of ambient S02 on lung function.
Only one epidemiological study (Adamkiewicz et al., 2004) evaluated inflammation, as
indexed by eNO, and found no association with SO2 exposure. Animal toxicological studies
found that repeated exposure to near ambient levels of SO2 leads to increased airway
inflammation in two models involving animals which were sensitized to an antigen (Park et al.,
2001; Li et al., 2007). Studies of other ambient pollutants indicate that influx of macrophages
and other inflammatory cells, with the related release of inflammatory cytokines, is a common
response to — and may further contribute to — injury.
Effects of short-term exposure to SO2 on AHR have been observed. In two animal
toxicological studies, repeated exposure to 0.1 ppm S02 led to AHR in guinea pigs sensitized to
an antigen (Riedel et al., 1988; Park et al., 2001). Human clinical studies by Devalia et al. (1994)
and Rusznak et al. (1996) demonstrated increased sensitivity to an inhaled allergen in asthmatic
subjects following exposure to a combination of SO2 (0.2 ppm) and NO2 (0.4 ppm). This effect
was not observed following exposure to either S02 or N02 alone. These findings of increased
pulmonary resistance are in concordance with the limited epidemiological findings of
S02-induced AHR (Taggart et al., 1996).
Epidemiological studies provide suggestive evidence for an association between ambient
S02 levels and ED visits and hospitalizations for all respiratory diseases in two susceptible
populations: children (Dab et al., 1996; Petroeschevsky et al., 2001; Walters et al., 1994) and
older adults (65+ years) (Fung et al., 2006; Schwartz, 1995; Spix et al., 1998; Wong et al., 1999).
Evidence for an association between ambient SO2 levels and these outcomes in adults was less
May 2008
3-46
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
consistent. A modest association between ambient SO2 and ED visits and hospitalizations for
asthma was also suggested. SO2 effect estimates were generally robust to the inclusion of
copollutants, including PM, 03, CO and N02. indicating that the observed effects of S02 on
respiratory endpoints is independent of the effects of other ambient air pollutants.
3.1.5. Mixtures and Interactive Effects
3.1.5.1. Evidence from Human Clinical Studies
The interaction of S02 with other common air pollutants or the sequential exposure of S02
after prior exposure to another pollutant can potentially modify SCVinduced respiratory effects.
However, only a few human clinical studies have looked at the interactive effects of coexisting
ambient air pollutants. In a human clinical study designed to simulate an ambient "acid summer
haze," Linn et al. (1997) exposed healthy and asthmatic children (9-12 years of age) for 4 h with
intermittent exercise to a mixture of S02 (0.1 ppm), H2SO4 (100 |^g/m3), and O3 (0.1 ppm).
Compared with exposure to filtered air, exposure to the air pollution mixture did not result in
significant changes in lung function or respiratory symptoms. These findings are in agreement
with a series of similar studies conducted by Kleinman et al. (1981; 1984; 1985).
In a human clinical study of asthmatic adolescents (12- to 16-years-old), Koenig et al.
(1983) evaluated changes in FEVi following a 10-min exposure during moderate exercise to
1-mg/m3 NaCl alone and in combination with 0.5 and 1.0 ppm SO2. Significant decreases of 15
and 23% were reported in FEVi following exposure to 1 mg/m3 NaCl in combination with 0.5-
and 1.0-ppm SO2, respectively. No significant changes in FEVi were observed between pre- and
post-exposure to 1-mg/m3 NaCl without SO2. The effect observed in this study may be the result
of the presence of hygroscopic particles that can carry SO2 deeper into the lung.
Koenig et al. (1990) also examined the effect of 15-min exposures to 0.1 ppm S02 in
adolescent asthmatics engaged in moderate levels of exercise. Immediately preceding this
exposure, subjects were exposed for 45 min to 0.12 ppm O3 during intermittent moderate
exercise. Subjects also underwent two additional exposure sequences with the same exercise
regimen: 15-min exposure to 0.1 ppm S02 following a 45-min exposure to clean air, and 15-min
exposure to 0.12 ppm O3 following a 45-min exposure to 0.12 ppm O3. The authors found that
the change in FEVi compared to baseline was significantly different following the O3-SO2
May 2008
3-47
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
exposure (8% decrease) when compared to the change following the air-SC>2 or O3-O3 exposures
(decreases of 3 and 2%, respectively). In a more recent study, Trenga et al. (2001) reported that
among adult asthmatics, exposure to 03 (0.12 ppm for 45 min) resulted in a slight increase in
bronchial responsiveness to SO2 at a concentration of 0.25 ppm (6.5% decrease in FEVi with
pre-exposure to O3, compared with a 3.4% decrease in FEVi with pre-exposure to filtered air).
Hazucha and Bates (1975) demonstrated a synergistic effect of concurrent exposure to SO2
(0.37 ppm) and 03 (0.37 ppm) on lung function in healthy asthmatics; however, no such effect
was observed in a similar study conducted by Bedi et al. (1979).
Jorres and Magnussen (1990) and Rubinstein et al. (1990) investigated the effects of a
prior NO2 exposure on SCVinduced bronchoconstriction in asthmatic adults. While Jorres and
Magnussen suggested that prior exposure to N02 increased the responsiveness to S02,
Rubinstein et al. did not find that NO2 exacerbated the effects of SO2. Linn et al. (1980) reported
no difference in lung function or respiratory symptoms among a group of exercising asthmatics
exposed to both clean air and a combination of NO2 (0.5 ppm) and SO2 (0.3 ppm).
In summary, although findings from some human clinical studies suggest that respiratory
effects of exposure to SO2 may be enhanced when preceded by, or occurring concomitant with
exposure to other air pollutants, this evidence is quite limited and inconsistent.
3.1.5.2. Evidence from Animal Toxicological Studies
As discussed earlier, S02 is a component of complex air pollution mixtures that vary
geographically and temporally (e.g., by hour, week, and season). Depending on atmospheric
conditions, SO2 can be transformed to secondary sulfate particles and acid aerosols (H2SO4) and
can adsorb onto particulate matter. Since SO2, H2SO4 and PM share a common source—fossil
fuels—health effects of fossil-fuel derived air pollution mixtures may be determined by
interactions among individual components. Although epidemiological studies provide
information on real-world exposure, it is difficult to evaluate causative factors and quantitative
relationships from such studies. Animal studies are therefore useful in evaluating health effects
of mixtures. The studies discussed below demonstrate important interactions between S02 and
other air pollution components.
An informative study of complex air pollutants was conducted in dogs and addressed in the
1982 AQCD (EPA, 1982). In dogs that were exposed to SO2 and H2SO4, with or without
May 2008
3-48
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
irradiated or non-irradiated auto exhaust concentrations relevant to urban exposures, functional
lung changes were observed at 61 months of exposure and at 2 years after exposures ended.
Morphological and biochemical changes were observed at 2.5-3 years after exposure.
Since then, studies have demonstrated respiratory responses following inhalation of SO2
which had been layered onto metal or carbon particles. The resulting particles were submicron in
size; they would be expected to deposit in the lower respiratory tract. As discussed in Section
3.2.2, chemosensitive receptors are present at all levels of the respiratory tract and are known to
activate reflexes involving the respiratory and cardiovascular systems. It has been postulated that
bronchial C-fiber receptors are more sensitive to chemical irritants than pulmonary C-fibers
receptors, but that more intense cardiovascular responses are triggered by the pulmonary
receptors (Coleridge and Coleridge, 1994; Widdicombe and Lee, 2001). Some of the work
involving SO2 layered onto particles has been reviewed in the PM AQCD (EPA, 1996; 2004).
Important studies are described briefly to show biological plausibility for the health effects of
SO2 which rarely, if at all, exists in nature in the absence of PM. This work is summarized in
Annex Tables F-15 through 5-1.
3.1.5.2.1. Effects of SO2 Layered on Metallic Particles
Studies examining interactions between SO2 and metallic or carbonaceous particles are
summarized in Annex Table E-14. Metal oxides may be released into the atmosphere with S02
during combustion of fossil fuels or by smelting operations (Lam et al., 1982). The 1982 AQCD
noted that sorption of SO2 onto liquid or solid particles, which may act as carriers, tended to
increase its potency, but the mechanism for the effect was not known. The studies discussed
below strongly suggest that SO2 adsorbed to particles penetrates to more distal regions of the
lung, compared with gaseous S02. In addition, S02 which is adsorbed to particles can transform
to sulfite, sulfate, H2SO4 and sulfur trioxide.
In an early study, guinea pigs were exposed to submicron zinc oxide aerosols alone or to
0.8-6 mg/m3 zinc oxide in combination with 1-2 ppm SO2 for 3 h (Lam et al., 1982). The higher
concentrations of zinc oxide served to carry more sulfur in the aerosol (Amdur et al., 1988).
Animal exposure to zinc oxide alone resulted in a decrease in functional residual capacity;
animal exposure to the combination of zinc oxide and SO2 resulted in dose-dependent decreases
in total lung capacity, vital capacity, functional residual capacity, residual volume, diffusion
May 2008
3-49
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
capacity for carbon monoxide and alveolar volume (Amdur et al., 1988; Lam et al., 1982).
Another study exposed guinea pigs to ~1 ppm SO2 and 1-2 mg/m3 zinc oxide for 1 h (Amdur et
al., 1983). Results demonstrated the formation of sulfite, sulfate, and sulfur trioxide under
conditions of high humidity and temperature, and greater than additive decrements in pulmonary
function, compared with exposures to SO2 or zinc oxide alone or with a mixture of SO2 and zinc
oxide where no transformation had taken place. Additional studies in guinea pigs involved 1 h
exposures to ~1 ppm S02 and 1-3 mg/m3 copper oxide (Chen et al., 1991) and -lppm S02 and
1-3 mg/m3 zinc oxide (Chen et al., 1992). In the copper oxide-SC>2 study, components were
mixed at either 37°C or 1,411°C prior to exposure (Chen et al., 1991). Results demonstrated
increased pulmonary resistance when the compounds were mixed at low temperature, leading to
the formation of sulfite; when components were mixed at high temperature, leading to the
formation of sulfate, no increase was found. This suggested that sulfite has greater biological
effect than sulfate. The zinc oxide-SC>2 study found a synergistic interaction between zinc oxide
and SO2. Co-exposure, but not exposure to either component alone, led to airway
hyperresponsiveness following an acetylcholine challenge (Chen et al., 1991). A further study by
these same investigators determined that H2SO4 was the predominant species of sulfur associated
with the zinc oxide particles mixed with SO2 under high temperature and humidity conditions
(Amdur et al., 1988). H2SO4 has a valence of S(VI), unlike that of SO2 and sulfite, which have a
valence of S(IV). This study also correlated increased bronchial sensitivity to acetylcholine with
the formation of H2S04 at 21 and 30 |ig/m3 in animals exposed for 1 h to this mixture. The
authors concluded that the response might have been enhanced by H2SO4 being carried on the
metal particles. Furthermore, the H2S04-coated zinc oxide particle was five- to tenfold more
potent in eliciting a response than H2SO4 alone.
Several subacute studies were also conducted. One involved the exposure of guinea pigs
for 3 h/day for 6 consecutive days to 6 mg/m3 submicron zinc oxide particles generated in a
humid furnace mixed with 1 ppm SO2 or to SO2 alone (Conner et al., 1985). Exposure to the zinc
oxide/SC>2 particles resulted in significant decreases in total lung capacity, vital capacity and
functional residual capacity compared with controls or exposure to S02 alone. These decreases
were maintained for at least 72 h following the last exposure. Decreases in diffusing capacity for
carbon monoxide and alveolar volume, as well as increased alveolar-duct inflammation, were
observed in response to zinc oxide-S02but not to SO2 alone. Because the changes noted in
May 2008
3-50
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
response to zinc oxide-S02 were identical to those seen in a previous study using zinc oxide, the
authors concluded that SO2 played a minor role in these responses. Subsequent studies employed
a lower concentration of zinc oxide (1 and 2.5 mg/m3) in order to unmask the effects of the
S02-zinc oxide interaction (Amdur et al., 1988; Conner et al., 1989). In one of these, significant
changes were observed in numbers of cells and other lavage fluid components in guinea pigs
exposed for 3 h/day for 5 days to SCVzinc oxide compared with those exposed only to SO2 or
zinc oxide. As previously described, it was determined that the mixture of zinc oxide and S02 in
a high temperature furnace with sufficient humidity resulted in the formation of a zinc oxide/
H2SO4 aerosol. Exposure of guinea pigs for 3 h/day for 5 days to this H2S04-coated ultrafine
aerosol at a concentration of 7-11 |ig/m3 S(VI) resulted in a significant decrease in diffusing
capacity for carbon monoxide, an effect not seen with exposure to zinc oxide or S02 alone. The
authors concluded that the response was due to the H2SO4 associated with the aerosol and
delivered to the lower respiratory tract (Amdur et al., 1988).
3.1.5.2.2. Effects of SO2 Layered on Carbon Particles
Effects of S02-containing mixtures on host defenses were examined in several studies.
Host defense results will not be discussed since high concentrations (10 ppm) of SO2 were used.
However, these experiments were important since they demonstrated that mixing SO2 and
submicron particles of carbon black in the presence of 85% relative humidity led to significant
adsorption of S02 onto the carbon black and oxidation of S02 to acid sulfate. No high
temperature furnace was employed in these studies (Jakab et al., 1996; Clarke et al., 2000).
3.1.5.2.3. Effects of Sulfite Aerosols
Several studies used sulfite particles as a surrogate for S02 adsorbed onto carbonaceous or
metallic particulates. These are discussed below and in Annex Table E-15. Sulfite and SO2, both
S(IV) species, were expected to have similar chemical reactivities. An acute 1-h exposure of
guinea pigs to submicron aerosols of sodium sulfite (204-1,152 |ig/m3 of sulfite) resulted in
significant effects on pulmonary function (Chen et al., 1987). A 50% increase in resistance and
19% decrease in compliance were observed using 972 |ig/m3 of sulfite, while dose-dependent
decreases in total lung capacity, vital capacity, functional residual capacity, residual volume,
diffusion capacity for carbon monoxide and increases in lung wet weight at concentrations of
May 2008
3-51
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
204 |ig/m3 and above. The authors noted that aerosols of S(IV) were 6 times more potent than
gaseous S(IV) in terms of bronchoconstriction and attributed these effects to greater pulmonary
deposition of the S(IV) aerosol. An earlier study by the same group found that 1-h exposure of
rabbits to greater than 1,200 |ig/m3 of sulfite led to accelerated clearance of a tracer aerosol from
the bronchial tree (Chen and Schlesinger, 1983). Chronic studies involving sulfite aerosols are
discussed in Section 3.4.2.5.
3.1.5.2.4. Other Mixtures
In addition to examining the interaction of SO2 and particles, animal studies performed
since the publication of the 1982 AQCD evaluated the effects of binary mixtures, laboratory-
generated complex mixtures (e.g., simulation of regional air pollution), or actual ambient air
mixtures. Annex Tables E-17 through E-20 summarize results from short-term studies on
possible toxicity relationships between SO2 and O3, and between SO2 and sulfates as well as the
effects of complex air pollution mixtures in healthy animals and disease models. Possible
interactions between S02 and cold air were also examined (Annex Table E-20). Generally, most
studies with ambient or laboratory-generated complex mixtures did not include a SCVonly
exposure group, making it difficult to determine the contribution of SOx. No definitive
conclusions can be made from these studies.
3.1.5.2.5. Summary of Evidence on SO2 Interactions with PM and Other Mixtures
The key findings by Amdur and others discussed above demonstrate that the effects of SO2
may be enhanced when aerosol particles act as carriers and deliver SO2 to the lower respiratory
tract. Interaction of SO2 and PM may also lead to transformation of SO2 to other forms of SOx
which may have more potent biological effects than S02. Studies discussed above reported
transformation of SO2 adsorbed onto metal oxide or carbon particles to sulfite, sulfate, sulfur
trioxide, and H2SO4 depending on conditions of temperature and relative humidity.
3.1.6. Evidence of the Effect of S02 on Respiratory Morbidity from
Intervention Studies
Many epidemiological studies have examined the association of short-term S02
concentrations and various respiratory morbidity outcomes. These studies collectively suggest
May 2008
3-52
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
that increased ambient SO2 concentrations are associated with increased risk of respiratory
outcomes, ranging from respiratory symptoms to ED visits and hospitalizations. Further
contributing to the evidence base are intervention studies that reported decreases in respiratory
morbidity following improvements in air quality, particularly reductions in SO2 concentrations.
In Hong Kong, a sudden change in regulation in July 1990 required all power plants and
road vehicles to use fuel oil with a sulfur content of < 0.5% by weight. These regulations were
enforced quickly, and provided opportunities to observe changes in morbidity before and after
the intervention. Peters et al. (1996) followed 3,521 children (mean age 9.5 years) residing in
two districts with good and poor air quality before the intervention from 1989 to 1991. The
intervention resulted in large reductions in SO2 (up to 80% in polluted district), along with a
modest reduction in sulfate (38% in polluted district). Only a small change in TSP levels was
observed after the intervention (15% decline in polluted district). In 1989 and 1990, an excess
risk of respiratory symptoms was observed in the polluted district. After the intervention, there
was a greater decline in reported symptoms of cough, sore throat, phlegm, and wheezing in the
polluted compared with the unpolluted district. For example, the OR for cough, comparing the
polluted to the unpolluted district, was 1.22 (95% CI: 1.05, 1.42) in 1989 and 1990, and
decreased to 0.92 (95% CI: 0.73, 1.15) in 1991.
A study by Keles et al. (1999) evaluated the prevalence of chronic rhinitis among high
school students before and after installation of a natural gas network for domestic heating and
industrial works, in a polluted area of Istanbul, Turkey. Concentrations of CO, N02, and
hydrocarbons were relatively low compared to SO2 and TSP in this area. After the intervention,
the annual mean TSP concentration declined by 23% from 89.7 |ig/m3 to 68.8 |ig/m3. An even
greater decline (46%) was observed for SO2, from an annual mean of 70.8 ppb to 38.2 ppb. The
prevalence of rhinitis decreased significantly from 62.5% to 51% of the student population (p <
0.05) following the installation of the natural gas network. Symptoms of rhinitis were associated
with air pollution levels, but not with any of the other factors considered, including sex,
household crowding, heating source, and smoking status. Although the effects from TSP could
not be separated from S02 effects, this study demonstrated that reductions in both pollutants
(with greater declines in SO2) resulted in significant reductions in the prevalence of chronic
rhinitis in a highly polluted area.
May 2008
3-53
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Another study in Germany observed that reductions in air pollutant levels were associated
with improvement in reported respiratory symptoms. Heinrich et al. (2002) examined the
influence of reduced air pollution levels on respiratory symptoms in children aged 5 to 14 years
(n = 7,632). Questionnaires were collected from the children during 1992-1993, 1995-1996, and
1998-1999 in three study areas. During the study period, SO2 concentrations decreased by more
than 90% and TSP concentrations decreased by approximately 60%. Concentrations of
nucleation-mode particles (10-30 nm) increased during this time period. For most respiratory
outcomes, the prevalence continued to decline in each of the three surveys. The temporal
changes followed similar trends in all three study areas. Stronger effects between SO2 and
prevalence of respiratory symptoms were observed among children without indoor exposures.
For those without indoor exposures, ORs of 1.21 (95% CI: 1.11, 1.32) were observed for
prevalence of bronchitis and 1.11 (95% CI: 1.02, 1.22) for frequent colds per 5-ppb increase in
the annual mean of SO2. Frye et al. (2003) reported changes in lung function parameters
associated with declines in SO2 concentrations in 2,493 children during this period as well. The
researchers observed a 0.6% (95% CI: 0.1, 1.2) increase in FVC and a 0.4% (95% CI: -0.1, 0.9)
increase in FEVi per 5-ppb decrease in the annual mean of SO2. They concluded that the
decreasing prevalence of respiratory symptoms and the increase in lung function following
decreases in air pollution levels might indicate the reversibility of adverse health effects in
children.
In summary, these studies observed that improvements in air quality, in particular large
decreases in SO2 concentrations, were associated with improvements in respiratory symptoms
and lung function. However, the decreased respiratory morbidity following large reductions in
ambient SO2 concentrations does not preclude the possibility that other constituents of the
pollution mixture that share the same source as S02 are also responsible for adverse effects. In
the German and Turkey studies, both SO2 and TSP concentrations decreased dramatically.
Although PM10 levels before and after the intervention were stable in Hong Kong, large
reductions in ambient nickel and vanadium were observed concomitantly with reductions of
sulfur after the intervention (Hedley et al., 2006). As discussed in Section 3.1.5, interactions of
SO2 and PM may lead to transformation of SO2 to other forms of SOx which have more potent
biological effects; thus the improvements in respiratory health may also be attributable to both
declines in SO2 and PM. Nonetheless, considered collectively with the larger body of evidence
May 2008
3-54
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
from epidemiological, human clinical, and animal toxicological studies, these studies are
supportive of SCVrelated effects on respiratory morbidity.
3.1.7. Summary of Evidence of the Effect of Short-Term S02 Exposure
on Respiratory Health
Evaluation of the health evidence led to the conclusion that it is sufficient to infer a causal
relationship between respiratory morbidity and short-term exposure to SO2. This conclusion is
supported by the consistency, coherence, and plausibility of findings observed in human clinical
studies with 5-10 min exposures, epidemiological studies using largely 24-h avg exposures and
animal toxicological studies using exposures of minutes to hours.
The strongest evidence for this causal relationship comes from human clinical studies
reporting respiratory symptoms and decreased lung function following peak exposures of 5-10
min duration to SO2. These effects have been observed consistently across studies involving
exercising mild to moderate asthmatics. Statistically significant decrements in lung function
accompanied by respiratory symptoms including wheeze and chest tightness have been clearly
demonstrated following exposure to 0.4-0.6 ppm SO2. Although studies have not reported
statistically significant respiratory effects following exposure to 0.2-0.3 ppm SO2, some
asthmatic subjects (5-20%) have been shown to experience moderate to large decrements in lung
function at these exposure concentrations. A larger body of evidence supporting this
determination of causality comes from numerous epidemiological studies reporting associations
with respiratory symptoms, ED visits, and hospital admissions with short-term SO2 exposures,
generally of 24-h avg. Important new multicity studies and several other studies have found an
association between 24-h avg ambient SO2 concentrations and respiratory symptoms in children,
particularly those with asthma. Furthermore, limited epidemiological evidence indicates that
atopic children and adults may be at increased risk for SCVinduced respiratory symptoms.
Generally consistent and robust associations also were observed between ambient S02
concentrations and ED visits and hospitalizations for all respiratory causes, particularly among
children and older adults (65+ years), and for asthma. Results of experiments in laboratory
animals support these observations; studies in animals sensitized with antigen demonstrate that
repeated exposure to near ambient S02 levels (as low as 0.1 ppm in guinea pigs) can exacerbate
May 2008
3-55
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
allergic responses including mucin production, airway inflammation and airway
hyperresponsiveness.
3.2. Other Morbidity Associated with Short-Term S02
Exposure
3.2.1. Summary of Findings from the Previous Review
The studies reviewed in the 1982 AQCD primarily investigated respiratory health
outcomes. There were no key animal toxicological or human clinical studies available at the last
review to address effects of S02 exposure on the cardiovascular system. The only report was a
study in dogs exposed to air pollutant mixtures (SO2 + H2SO4 with or without nonirradiated or
irradiated auto exhaust). No changes were observed in cardiovascular function at the end of 3
years of exposure and 3 years after exposure. No epidemiological studies linking exposure to
S02 with cardiovascular physiological endpoints or ED visits or hospital admissions for
cardiovascular causes were examined in the last review. Other organ systems in addition to
cardiovascular were not addressed in the 1982 AQCD.
3.2.2. Cardiovascular Effects Associated with Short-Term Exposure
The biological basis for SCVrelated cardiovascular health effects may lie in the stimulation
of chemosensitive receptors found in the respiratory tract which respond to irritants like S02.
Vagally-mediated responses may affect the cardiovascular system by inducing bradycardia and
either hypotension or hypertension, as discussed in Section 3.1.2. Alternatively oxidation
reactions mediated by the SO2 metabolites sulfite and bisulfite which have been absorbed into
the systemic circulation may potentially alter cardiovascular function. In general, vagally-
mediated responses have been observed at lower concentrations of SO2 than oxidative injury.
Since 1982, several animal toxicological studies have addressed the effects of SO2 on
cardiovascular endpoints. These are summarized below and in Annex Table E-5. In addition,
there is one noteworthy study examining the hematological effects of short-term S02 exposure
(Annex Table E-8). Acute exposure of rats to 0.87 ppm SO2 for 24 h resulted in increased
hematocrit, sulfhemoglobin and osmotic fragility as well as decreased whole blood and packed
cell viscosities (Baskurt, 1988). These results indicate a systemic effect of inhaled SO2 at
May 2008
3-56
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
concentrations near ambient levels and are consistent with an oxidative injury to red blood cells.
Only one study since 1982 measured systemic levels of sulfite or bisulfite following SO2
inhalation (Gunnison et al., 1987; Annex Table E-22). Further studies are required to confirm that
inhalation exposures of SO2 at or near ambient levels increase blood sulfite and bisulfite levels
sufficient for oxidative injury to blood cells or other tissues.
Recent epidemiological studies have examined the association between air pollution and
cardiovascular effects, including increased heart rate (HR), reduced heart rate variability (HRV),
incidence of ventricular arrhythmias, changes in blood pressure, incidence of myocardial
infarctions (MI), and ED visits and hospitalizations due to cardiovascular causes. The results of
these cardiovascular studies are summarized in Annex Tables F-3 and F-4.
3.2.2.1. Heart Rate and Heart Rate Variability
Heart rate variability (HRV) is generally determined by analyses of time (e.g., standard
deviation of normal R-R intervals [SDNN]) and frequency domains (e.g., low frequency [LF] /
high frequency [HF] ratio by power spectral analysis, reflecting autonomic balance) measured
during 24 h of electrocardiography (ECG). Brook et al. (2004) stated that HRV, resting heart rate,
and blood pressure are modulated by a balance between the two determinants of autonomic tone
(the sympathetic and parasympathetic nervous systems). An imbalance of cardiac autonomic
control may predispose susceptible people to greater risk of ventricular arrhythmias and
consequent cardiac deaths (Brook et al., 2004; Liao et al., 2004).
A limited number of human clinical studies examined the effect of SO2 on HRV. During a
controlled exposure of 12 healthy subjects and 12 subjects with asthma to 0.2 ppm SO2 for 1 h
under resting conditions, Tunnicliffe et al. (2001) reported that HF power, LF power, and total
power were higher with S02 exposures compared to air exposure in the healthy subjects, but that
these indices were reduced during SO2 exposure in the subjects with asthma. The LF/HF ratios
were unchanged in both groups. The authors postulated two autonomic pathways for SO2-
mediated bronchoconstriction. In healthy subjects, the dominant pathway was proposed to be the
rapidly adapting receptor/C-fiber route, which results in a central nervous system reflex with an
increase in vagal tone. In the asthmatic subjects, proximal airway narrowing was proposed as the
dominant response, possibly through neurogenic inflammation. This likely causes a
compensatory central nervous system-mediated reduction in vagal tone, resulting in
May 2008
3-57
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
bronchodilation of the distal airway. While there were no detectable changes in symptoms or
lung function in either of the groups, this study provides some evidence that exposure to SO2
may elicit systemic responses at these low levels (0.2 ppm).
In a similar study, Routledge et al. (2006) exposed patients with stable angina as well as
healthy subjects to 50 |ig/m3 carbon particles, 0.2 ppm SO2, alone and in combination, for 1 h
under resting conditions. HRV, C-reactive protein, and coagulation markers were measured. The
authors reported that for the healthy subjects, S02 exposure was associated with a decrease in
HRV markers of cardiac vagal control 4 h after exposure. However, it should be noted that there
was no apparent difference in the absolute value of the root mean square of successive RR
interval differences (r-MSSD) at 4 h postexposure between the control, SO2, carbon, and
carbon/S02 groups. The significant difference reported in the change in r-MSSD from baseline to
4 h postexposure with SO2 appears to be due to a higher baseline value of r-MSSD preceding the
SO2 exposure compared to the baseline value of r-MSSD preceding the air exposure. There were
no changes in HRV among the patients with stable angina. The authors noted that this lack of
response in the heart patients may be due to a drug treatment effect rather than decreased
susceptibility; a large portion of the angina patients were taking beta blockers, which are known
to increase indices of cardiac vagal control.
In an epidemiological study, Liao et al. (2004) investigated short-term associations
between ambient pollutants and cardiac autonomic control from the fourth cohort examination
(1996 through 1998) of the population-based Atherosclerosis Risk in Communities (ARIC) study
using a cross-sectional study design. Men and women aged 45 to 64 years (n = 6,784) from three
U.S. study centers in North Carolina, Minnesota, and Mississippi were examined. Resting,
supine, and 5-min beat-to-beat R-R interval data were collected. The mean 24-h avg SO2 level
measured 1 day prior to the HRV measurement was 4 ppb (SD 4). In addition to S02, the
potential effects of PM10, O3, CO, and NO2 were evaluated. Previous-day SO2 concentrations
were positively associated with HR and inversely associated with SDNN and LF power.
Consistently more pronounced associations were suggested between SO2 and HRV among
persons with a history of coronary heart disease. Significant associations with HRV indices also
were observed for PM10 and the other gaseous pollutants. When the regression coefficients for
each individual pollutant model were compared, the effects of PM10 on HRV were considerably
May 2008
3-58
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
larger than the effects for the gaseous pollutants, including SO2. No multipollutant analyses were
conducted.
Gold et al. (2000; reanalysis 2003) examined the effect of short-term changes in air
pollution on HRV in a panel study of 21 older adults (aged 53 to 87 years) in Boston, MA. The
study participants were observed up to 12 times from June to September 1997. The mean
24-h avg SO2 concentration was 3.2 ppb (range: 0, 12.6). The 24-h avg SO2 concentration was
associated with decreased HR in the first 5-min rest period, but not in the overall 25-min study
protocol. The effect estimate for SO2 slightly diminished but remained marginally significant in a
two-pollutant model with PM2.5 The inverse association between SO2 and HR observed in this
study are in contrast to the SCVrelated increases in HR observed by Liao et al. (2004) and Peters
et al. (1999). No associations were observed between HRV and S02. The strongest associations
with HRV were observed for PM2.5 and O3.
Another study of air pollutants and HRV was conducted in Boston, MA on 497 men from
the Normative Aging Study (Park et al., 2005). The best 4-consecutive-min interval from a 7-min
sample was used for the HRV calculations. For the exposure variable, 4-, 24-, and 48 h moving
averages matched on the time of the ECG measurement for each subject were considered. The
mean 24-h avg SO2 concentration was 4.9 ppb (range: 0.95, 24.7). Associations with measures of
HRV were reported for PM2.5 and O3, but not with SO2 for any of the averaging periods. In
another study conducted in Boston, MA, Schwartz et al. (2005) found significant effects of
increases in PM2.5 on measures of HRV, while no associations with S02 were observed. Other
studies examined the relationship of SO2 with HRV (Chan et al., 2005; de Paula Santos et al.,
2005; Holguin et al., 2003; Luttmann-Gibson et al., 2006). Most of these studies, with the
exception of de Paula Santos et al., did not observe associations with SO2.
In the limited number of epidemiological studies that examined a possible effect of S02 on
HRV, there were some suggestive findings; however, results reported from the human clinical
studies were inconsistent. The overall evidence does not support the conclusion that SO2 affects
cardiac autonomic control.
May 2008
3-59
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
3.2.2.2. Repolarization Changes
In addition to the role played by the autonomic nervous system in arrhythmogenic
conditions, myocardial vulnerability and repolarization abnormalities are believed to be key
factors contributing to the mechanism of such diseases.
Two in vitro studies (Nie and Meng, 2005) conducted with a 1:3 molar:molar mixture of
the S02 derivatives bisulfite and sulfite demonstrated effects of a 10-jam bisulfite:sulfite mixture
on sodium and L-type calcium currents (which included changes in inactivation and/or
activation, recovery from inactivation, and inactivation/activation time constants) in ventricular
myocytes. These in vitro observations suggested a potential role for L-type calcium current in
cardiac injury following S02 exposure. However, in vivo cardiovascular effects were observed
only at high SO2 concentrations (10 ppm and higher). Additional toxicological studies are
necessary to evaluate repolarization changes at ambient levels of SO2.
In an epidemiological study, Henneberger et al. (2005) examined the association of
repolarization parameters (QT duration, T-wave complexity, variability of T-wave complexity,
and T-wave amplitude) with air pollutants in patients with preexisting coronary heart disease (n =
56, all males) in East Germany. The patients were examined repeatedly once every 2 weeks for 6
months, for a total of 12 ECG recordings. The mean 24-h avg SO2 concentration was 2 ppb
(range: 1, 4). Ambient S02 concentrations during the 24-h preceding the ECG were associated
with the QT interval duration, but not with any other repolarization parameters. Stronger
associations were observed between PM indices and QT interval duration, T-wave amplitude,
and T-wave complexity.
To summarize, the evidence, while suggestive, is too limited to draw conclusions on the
association of S02 exposure and repolarization changes at this time.
3.2.2.3. Cardiac Arrhythmias
One toxicological study examined the effects of PM, ultrafine carbon, and S02 on
spontaneous arrhythmia frequency in 18-month-old rats (Nadziejko et al., 2004). The rats were
exposed to 1 ppm SO2 for 4 h. No significant change in the frequency of spontaneous
arrhythmias was found with SO2 and ultrafine carbon exposure. However, rats exposed to
May 2008
3-60
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
concentrated ambient PM had a significantly greater increase in the frequency of delayed beats
than rats exposed to air.
In a panel study of 100 patients with implanted cardioverter defibrillators (ICDs) in
Eastern Massachusetts, Peters et al (2000) tested the hypothesis that patients with ICDs would
experience life-threatening arrhythmias after an air pollution episode. The mean 24-h avg SO2
concentration measured at two sites in Boston during the study period was 7 ppb (5th-95th
percentile: 1, 19). ICDs monitor ECG abnormalities, and treat ventricular fibrillation or
ventricular tachycardias by administering shock therapy to restore the normal cardiac rhythm.
The ICD device also stores information on each tachyarrhythmia and shock. There was no
association between SO2 and defibrillator discharges in the 33 subjects who had any defibrillator
discharges during the follow-up period or in the 6 subjects who had at least 10 discharges. There
was some evidence that NO2 was associated with increased defibrillatory interventions in the
subjects with any defibrillator discharges. Among the patients with at least 10 events, NO2, CO,
and PM2.5 were found to be associated with defibrillator discharges.
In a follow-up study designed to confirm the findings of Peters et al. (2000), Dockery et al.
(2005) used a larger sample of ICD patients in Boston (n = 203) with a longer follow-up period.
The median concentration of 48-h avg SO2 averaged across multiple sites in Boston was 4.9 ppb
(IQR 4.1). No significant associations were found between ventricular arrhythmic episode days
and any of the air pollutants. However, when the analysis was stratified by recent arrhythmias
(i.e., within 3 days), there was evidence of an excess risk of ventricular arrhythmia with S02,
PM2.5, black carbon, NO2, and CO. Since PM2.5, black carbon, NO2, and CO were correlated with
each other and with SO2, the authors noted that differentiating the independent effects of the
pollutants would be difficult. A case-crossover analysis of the same data by Rich et al. (2005)
also observed associations with 48-h avg S02, but the S02 effect was not found to be robust to
adjustment by PM2.5. In a similar study conducted in St. Louis, MO, an excess risk was
associated with SO2 concentrations in the 24 h prior to an arrhythmia, but not with PM2.5 and O3
(Rich et al., 2006). In this study, none of the other measured pollutants (PM, elemental carbon,
03, CO, N02) were correlated with S02. The authors suggested that the different effects observed
in St. Louis and Boston may be due to differences in the source or mix of air pollutants in these
cities. Finally, findings from a time series study of tachyarrythmic events among 518 patients
May 2008
3-61
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
over a 10 year period in Atlanta do not indicate an association with SO2, nor with the other
pollutants studied including PM2.5 and its components (Metzger et al., 2007).
Additional studies have examined the relationship of S02 with arrhythmias in Vancouver,
and observed associations at very low ambient SO2 concentrations (mean 24-h avg SO2 of -2.5
ppb with a max of 8.1 ppb). Vedal et al. (2004) stated that of all pollutants examined, the only
one with somewhat consistent positive associations with arrhythmia events was SO2. In season-
stratified analyses, S02 was positively associated with arrhythmias in the winter, while in the
summer the association was negative. On the other hand, in the Rich et al. (2004) study, positive
associations were observed in the summer but not in the winter. The authors stated that it was
difficult to interpret these findings.
Collectively, the epidemiological evidence for an association between short-term exposure
to SO2 and arrhythmias is inconsistent. The limited toxicological evidence does not provide
biological plausiblity for an effect.
3.2.2.4. Blood Pressure
Two animal toxicological studies examined the effect of SO2 on blood pressure (Annex
Table E-5). Halinen et al. (2000) examined blood pressure changes in guinea pigs. The animals
were hyperventilated to simulate exercise, and exposed to 1-, 2.5-, and 5 ppm SO2 in cold, dry
air. After 10-min exposures to each S02 concentration, separated by 15-min exposures to clean,
warm, humid air, a transient increase in blood pressure was observed during 5 ppm S02 exposure
in cold, dry air. In a second study (Halinen et al., 2000b), hyperventilated guinea pigs were
exposed to cold, dry air alone or to 1 ppm SO2 in cold, dry air for 60 min. The study reported
similar increases in blood pressure and HR with exposure to cold, dry air or to SO2 in cold, dry
air. The increase in HR was gradual, while increases in blood pressure generally occurred during
the first 10 to 20 min of exposure. Similar effects were observed with exposure to cold, dry air or
to SO2 in cold, dry air, suggesting that effects were associated with cold, dry air rather than with
S02.
Ibald-Mulli et al. (2001) examined the association between blood pressure and S02 using
survey data from the MONICA (Monitoring Trends and Determinants in Cardiovascular Disease)
Project. Blood pressure measurements were taken from 2,607 men and women. The mean
24-h avg SO2 concentration was 23 ppb (range: 5, 91). An increase in systolic blood pressure was
May 2008
3-62
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
associated with 24-h avg SO2 and TSP. However, in a two-pollutant model with TSP, the effect of
SO2 on blood pressure was substantially reduced and became nonsignificant while the effect of
TSP was robust.
In a study by de Paula Santos et al. (2005), changes in blood pressure in association with
SO2 were investigated in vehicular traffic controllers (n = 48) aged 31 to 55 years living in Sao
Paulo, Brazil, where vehicles are the primary source of air pollution. The mean 24-h avg SO2
level, measured at six different stations around the city, was 7 ppb (SD 3). Blood pressure was
measured every 10 min when subjects were awake (6 a.m. to 11 p.m.) and every 20 min during
sleep (11 p.m. to 6 a.m.). Results indicated that SO2, as well as CO, were associated with
increases in systolic and diastolic blood pressure. However, when a two-pollutant model was
used to test the robustness of the associations, only the CO effect remained statistically
significant.
Very few studies have examined the effects of short-term SO2 exposure on blood pressure.
Collectively, the limited toxicological and epidemiological evidence does not suggest that
exposure to S02 has effects on blood pressure.
3.2.2.5. Blood Markers of Cardiovascular Risk
Folsom et al. (1997) demonstrated that elevated levels of fibrinogen, white blood cell
count, factor VIII coagulant activity (factor VIII-C), and von Willebrand factor were associated
with risk of cardiovascular disease. Schwartz (200 ^investigated the association between various
blood markers of cardiovascular risk and air pollution among subjects in the Third National
Health and Nutrition Examination Survey (NHANES III) in the United States conducted between
1989 and 1994 across 44 counties. The NHANES III is a random sample of the U.S. population
with oversampling for minorities (30% of NHANES sample) and the elderly (20% of the
sample). The mean SO2 concentration was 17.2 ppb (IQR 17) across the 25 counties where data
were available. This study looked at fibrinogen levels, platelet counts, and white blood cell
counts. After controlling for age, ethnicity, gender, body mass index, and smoking status and
number of cigarettes per day, S02 was found to be positively associated with white blood cell
counts. PM10 was associated with all blood markers. In two-pollutant models, PM10 remained a
significant predictor of white blood cell counts after controlling for SO2, but not vice versa.
May 2008
3-63
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
A recent cross-sectional study by (Liao et al., 2005) investigated the effects of air pollution
on plasma hemostatic and inflammatory markers in the ARIC study (n = 10,208). The authors
hypothesized that short-term exposure to air pollutants was associated with increased levels of
inflammatory markers and lower levels of albumin, as serum albumin is inversely associated
with inflammation. The mean 24-h avg SO2 concentration was 5 ppb (SD 4). Significant
curvilinear relationships were observed between SO2 and factor VIII-C, white blood cell counts,
and serum albumin. The authors noted that since no biological explanation could be offered for
the "U"-shaped curve between SO2 and factor VIII-C and the "inverse U"-shape between SO2
and albumin, generalization of the association should be exercised with caution. No associations
were observed between SO2 and fibrinogen or von Willebrand factor.
In another large cross-sectional study of 7,205 office workers in London, Pekkanen et al
(2000) examined the association between plasma fibrinogen and ambient air pollutants. The
mean 24-h avg SO2 was 9 ppb (10th-90th percentile: 5, 19). Associations with fibrinogen were
observed for all pollutants examined, either in all-year or summer-only analyses, except for SO2
and 03.
Taken together, results from the limited number of studies do not suggest that SO2 is
associated with various blood markers of cardiovascular risk.
3.2.2.6. Acute Myocardial Infarction
The association between air pollution and the incidence of MI was examined in a small
number of studies. As part of the Determinants of Myocardial Infarction Onset Study, Peters
et al. (2001) examined 772 patients with MI living in greater Boston, MA. A case-crossover
design was used to assess changes in the risk of acute MI after exposure to potential triggers. The
mean 24-h avg S02 was 7 ppb (range: 1, 20) during the study period. Similarly, the mean 1-h avg
SO2 was 7 ppb (range: 0, 23). In an analysis that considered both the 2-h avg (between 60 and
180 min before the onset of symptoms) and 24-h avg (between 24 and 48 h before the onset)
concentrations jointly, the study found no significant association between risk of MI and SO2. Of
all the pollutants considered, only PM2.5 and PMi0 were found to be associated with an excess
risk of MI. In a study of 5793 confirmed cases of acute MI in King County Washington, Sullivan
et al. (2005) also used a case-crossover design to investigate the association of ambient levels of
May 2008
3-64
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
several air pollutants 1, 2, 4 and 24 h before the MI onset. No association with SO2 (or with
PM2.5) was observed. The mean SO2 level was 9 ppb (range: 0-39 ppb) at the time of the study.
In the MONICA Project, the effect of air pollution on acute MI was studied in Toulouse,
France, using a case-crossover study design (Ruidavets et al., 2005). The mean 24-h avg SO2
level was 3 ppb (5th-95th percentile: 1, 5). A total of 399 cases of acute MI were recorded during
the study period. O3, but not SO2 or NO2, was found to be associated with the incidence of acute
MI. Exposure to PM was not considered in this study.
Only a limited number of studies examined the association between ambient SO2
concentrations and incidence of acute MI. These studies provide no evidence that exposure to
SO2 increases the risk of MI.
3.2.2.7. Emergency Department Visits and Hospitalizations for Cardiovascular
Diseases
The current review includes more than 30 peer-reviewed studies that address the effect of
SO2 exposure on ED visits or hospitalizations for cardiovascular diseases. These studies are
discussed briefly in this section and further summarized in Annex Table F-4.
3.2.2.7.1. All Cardiovascular Diseases
The disease grouping of all cardiovascular diseases typically includes all diseases of the
circulatory system (e.g., heart diseases and cerebrovascular diseases, ICD9 Codes 390-459). A
summary of the associations reported for ambient SO2 concentrations with all cardiovascular
diseases are presented in Figure 3-9.
In a study of 11 cities in Spain, an excess risk of 3.6% (95% CI: 0.6, 6.7) per 10 ppb
increase in 24-h avg SO2 at a 0-1 day lag was observed for all cardiovascular disease admissions
(Ballester et al., 2006). The mean 24-h avg SO2 level in the cities studied was 6.6 ppb. In
addition, time-series data linking SO2 with hospital admissions for cardiovascular diseases in
three metropolitan areas in the United States (i.e., Cook, Maricopa, Los Angeles Counties) was
conducted (Moolgavkar, 2000; reanalysis, 2003). Among older adults (65+ years) in Los Angeles
County, a 13.7% (95% CI: 11.3, 16.1) excess risk in admissions was observed per 10 ppb
increase in 24-h avg SO2 at lag 0 day, using a Generalized Linear Model (GLM) and natural
splines to adjust for temporal trends rather than GAM. The median 24-h avg S02 level for Los
May 2008
3-65
DRAFT—DO NOT QUOTE OR CITE
-------
1 Angeles County was 2 ppb during the study period. Results for Maricopa and Cook counties
2 were not presented in the reanalysis.
Reference
Location
Lag
'Metzger etai. (2004)
Atlanta, GA
0-2
Balester et al. (2008)
14 Spanish cities
0-1
Atkinson et al. (1393)
London, England
0
Poloniecki et al. (1997)
London, England
1
Ballesteretal. (2001)
Valencia, Spain
2
Llorca et al. (2005)
Toirelavega, Spain
0
Petroeschevsky et al. (2001)
Brisbane, Australia
0
Chang etal. (2005)
Taipei, Taiwan
220°C, 0-2
Chang et al. (2005)
Taipei, Taiwan
<20°C, 0-2
Petroeschevsky et al. (2001)
Brisbane, Australia
0
Moolgavkar (2003)
Los Angeles, CA
0
Atkinson et al. (1399)
London, England
0
•Jalaludin etal ..(2008)
Sydney, Australia
0
Petroeschevsky et al. (2001)
Brisbane, Australia
1
All ages
15-64 years
65+ years
B
JL
Tr
.9 1 1.1
Relative Risk
1.2 1.3 1.4
Figure 3-9. Relative risks (95% CI) of S02-associated emergency department visits (*) and
hospitalizations for all cardiovascular causes, arranged by age group. Risk estimates are
standardized per 10 ppb increase in 24-h avg S02 concentrations or 40 ppb increase in 1-h max
S02. The size of the box of the central estimate represents the relative weight of that estimate
based on the width of the 95% CI.
3 In a large single city study Metzger et al. (2004) examined approximately 4.4 million
4 hospital visits to 31 hospitals from 1993 to 2000 in Atlanta, GA and reported a 1.4% (95% CI: "
5 1.5, 4.4) excess risk in ED visits for cardiovascular causes per 40-ppb increase in 1-h max SO2.
6 Peel et al. (2006) conducted analyses using the same dataset to compare results obtained with a
7 case-crossover design to the Metzger et al. (2007) results, which were obtained using a time
8 series approach. Peel et al. and Metzger et al. report similar findings. The median 1-h max SO2
9 level in Atlanta during the study period was 11 ppb (10th-90th percentile: 2-39). Results from
10 several single-city studies in Europe, Australia, and Taiwan indicated positive associations with
May 2008
3-66
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
SO2 (Atkinson et al., 1999a; Ballester et al., 2001; Jalaludin et al., 2006; Petroeschevsky et al.,
2001; Poloniecki et al., 1997), though others observed negative associations (Chang et al., 2005;
Llorca et al., 2005) (see Figure 3-9).
3.2.2.7.2. Specific Cardiovascular Diseases
Several studies examined the effect of ambient SO2 on hospital admissions for cardiac
disease (ICD9 Codes 390-429), ischemic heart disease (IHD, ICD9 Codes 410-414),
dysrhythmia (ICD9 Code 427), congestive heart failure (CHF, ICD9 Code 428), MI (410) or
cerebrovascular diseases (ICD9 Codes 430-438). AEuropean multicity study reported
statistically significant positive associations with cardiac disease admissions (Ballester et al.,
2006). However, adjustment for PMi0 and CO in two-pollutant models diminished the
association reported by Ballester et al. by approximately 50%. Findings for cardiac disease
admissions reported in several additional single city studies conducted in the United States,
Canada, Australia and Europe were inconsistent (Fung et al., 2005; Jalaludin et al., 2006; Llorca
et al., 2005; Michaud et al., 2004).
Analyses restricted to diagnoses of IHD (Jalaludin et al., 2006; Lee et al., 2003; Lin et al.,
2003a; Metzger et al., 2004; Peel et al., 2007), CHF (Koken et al., 2003; Metzger et al., 2004;
Morris et al., 1995; Peel et al., 2007; Wellenius et al., 2005b), dysrhythmia (Koken et al., 2003;
Metzger et al., 2004; Peel et al., 2007), MI (Koken et al., 2003; Lin et al., 2003a), and angina
pectoris (Hosseinpoor et al., 2005) were also conducted. Metzger et al. (2004) observed weak
nonsignificant or negative associations of 1-h max SO2 with IHD, CHF, and dysrhythmia. Using
the same dataset, Peel et al. (2007) investigated effect modification of cardiovascular disease
outcomes across comorbid disease status categories, including hypertension, diabetes, COPD,
dysrhythmia, and CHF. Authors observed only weak nonsignificant or negative associations for
IHD, CHF, and dysrhythmia with ambient 1-h max SO2 level in any comorbid disease category.
Both increases in admissions or ED visits (Jalaludin et al., 2006; Koken et al., 2003; Wellenius et
al., 2005a) and weak or negative associations (Hosseinpoor et al., 2005; Lee et al., 2003b; Lin et
al., 2003a) were reported in other studies.
Studies of the effect of SO2 on cerebrovascular admissions were also considered. Positive
associations were reported for ischemic stroke (Villeneuve et al., 2006; Wellenius et al., 2005b;
Wellenius et al., 2005a). However Wellenius et al. (2005b) reported stronger associations for
May 2008
3-67
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
NO2 and CO than SO2, and the association reported by Villenueve et al. (2006) was diminished
in two-pollutant models. No meaningful positive associations of ambient SO2 with
cerebrovascular diseases were observed in several other studies (Henrotin et al., 2007; Jalaludin
et al., 2006; Metzger et al., 2004; Peel et al., 2007; Tsai et al., 2003).
3.2.2.7.3. Summary of Evidence on Emergency Department Visits and
Hospitalizations from Cardiovascular Diseases
Several studies have observed positive associations between ambient SO2 concentrations
and ED visits or hospital admissions for cardiovascular diseases (e.g., all cardiovascular diseases,
cardiac diseases, cerebrovascular diseases) particularly among individuals 65+ years of age, but
results are not consistent across studies. The strongest evidence comes from a large multicity
study conducted in Spain (Ballester et al., 2006) that observed statistically significant positive
associations between ambient SO2 and cardiovascular disease admissions; however, the SO2
effect was found to diminish by half with PMi0 and CO adjustment. Only a limited number of
studies assessed potential confounding by copollutants despite the moderate to strong correlation
between SO2 and various copollutants in most studies. While some studies suggest that the
association of SO2 with cardiovascular hospitalizations were generally robust to adjustment for
BS and PMi0 (Ballester et al., 2001; Fung et al., 2005), several other studies, including that by
Balleseter et al. (2006), observed that the effect of SO2 on cardiovascular ED visits and
hospitalizations may be confounded by copollutant exposures. Jalaludin et al. (2006) reported a
3% excess risk in cardiovascular disease hospital admissions per 0.75 ppb incremental change in
24-h avg SO2 in single-pollutant models, which was reduced to null when CO was included.
Chang et al. (2005) noted that the observed negative association of S02 with all cardiovascular
disease hospitalizations they observed was strengthened after adjusting for NO2, PM10, and CO
in two-pollutant models. The authors attributed this finding to possible collinearity problems
between SO2 and copollutants. None of the epidemiological studies examined effects of possible
interactions among copollutants.
3.2.2.8. Summary of Evidence on the Effect of Short-Term SO2 Exposure on
Cardiovascular Health
The overall evidence on the effect of short-term exposure to SO2 on cardiovascular health
effects is inadequate to infer the presence or absence of a causal relationship. Epidemiological
May 2008
3-68
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
studies of HRV, cardiac repolarization changes, and cardiac rhythm disorders provide limited
evidence of associations with short-term exposure to SO2. There was some suggestive evidence
of an association between S02 exposure and HRV in the epidemiological studies, but the
evidence from two human clinical studies were weak and inconsistent. Similarly, several studies
observed positive associations between ambient SO2 concentrations and ED visits or hospital
admissions for cardiovascular diseases, but results were not consistent across studies and specific
cardiovascular disease outcomes. In general, most epidemiological studies observed that these
cardiac outcomes were associated more strongly with PM compared to SO2. In the limited
studies that examined potential confounding by copollutants using multipollutant models, the
SO2 effect was generally found to diminish with adjustment for PM indices or CO.
Given the lack of coherence among the cardiovascular outcomes examined and the limited
evidence available to evaluate potential confounding and interaction by copollutants, the overall
evidence for cardiovascular health effects following exposure to ambient SO2 is weak and
insufficient to make a causal determination.
3.2.3. Other Effects Associated with Short-Term S02 Exposure
The short-term effects of S02 on other organ systems were not examined in the previous
review. A review of animal toxicological studies published since the 1982 AQCD indicates a
limited number of research inquiries addressing the systemic effects of short-term SO2 exposure
in various other organs. The most recent studies on these are summarized in Annex Tables E-6
through E-9 and E-22 through E-24.
Of note are three ex vivo acute exposure studies using SO2 derivatives (sulfite and
bisulfite) on hippocampal or dorsal root ganglion neurons isolated from Wistar rats (Du and
Meng, 2004a; b; 2006). Perturbations were observed in potassium-, sodium-, and calcium-gated
channels at concentrations of 0.01-100 |iM. These authors speculated that such effects might
correlate with the neurotoxicity that has been associated with SO2 inhalation. However effects on
the nervous system have generally been studied using chronic exposures > 5 ppm SO2. Effects
observed at these levels are of questionable significance in evaluating the health effects at
ambient levels. These studies are summarized in Annex Table E-6.
May 2008
3-69
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
3.3. Mortality Associated with Short-Term S02
Exposure
3.3.1. Summary of Findings from the Previous Review
The studies available to review in the 1982 AQCD were mostly from historical data
including London, England, and New York City air pollution episodes. Effects of SOx (mainly
SO2) were investigated along with PM indices because they shared a common source, coal
burning, and separating their associations with mortality was a challenge that many of the earlier
episodic studies could not resolve. The S02 levels observed in these air pollution episodes were
several orders of magnitude higher than the current average levels observed in U.S. cities (e.g., in
the 1962 New York City episode, SO2 in Manhattan peaked at 400 to 500 ppb). Some of these
London and New York City studies suggested that PM, not SO2, was associated with observed
mortality, but the 1982 AQCD could not resolve the relative roles of these two pollutants and
suggested that the clearest mortality associations were seen when both pollutants were at high
levels (24-h avg values of both BS and SO2 exceeding 1000 |ig/m3 [-400 ppb for SO2]) and less
so at lower ranges although the review of the studies and reanalyses found no clear evidence of a
threshold for S02.
The 1986 Second Addendum to the 1982 AQCD reviewed more reanalyses of the London
data and analyses of New York City, Pittsburgh, and Athens data. While these reanalyses and
some new analyses confirmed earlier findings (and suggested stronger evidence of BS effects
than of the S02 effects), given the remaining uncertainties with exposure error and statistical
modeling, there was not sufficient information to quantitatively determine concentration-
response relationships at lower concentrations of either PM or SO2.
A series of short-term mortality effects studies in the late 1980s and early 1990s (Pope,
1989; Fairley, 1990; Dockery et al., 1992; Pope et al., 1992; Schwartz and Dockery, 1992)
showed associations between mortality and PM indices at relatively low levels. Since then, a
large number of epidemiological studies have investigated the adverse health effects of air
pollution with hypotheses mainly focused on PM, and SO2 was often analyzed as one of the
potential confounders in these studies.
May 2008
3-70
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
3.3.2. Associations of Mortality and Short-Term S02 Exposure in
Multicity Studies and Meta-Analyses
In reviewing the range of SO2 mortality effect estimates, multicity studies provide
especially useful information because they analyze data from multiple cities using a consistent
method, avoiding potential publication bias. There have been several multicity studies from the
United States, Canada, and Europe, some of which will be discussed in the sections below. Meta-
analysis studies also provide useful information on describing heterogeneity of effect estimates
across studies; however, in contrast to multicity studies, the observed heterogeneity may reflect
the variation in analytical approaches across studies. In addition, the effect estimate from a meta-
analysis may be subject to publication bias, unless the analysis specifically examines such bias
and adjusts for it. These studies, as well as many other single-city studies, are summarized in
Annex Table F-5.
3.3.2.1.1. Multicity Studies
3.3.2.1.2. National Morbidity, Mortality, and Air Pollution Study
The time-series analysis of the largest 90 U.S. cities (Samet et al., 2000; reanalysis
Dominici et al., 2003) in the National Morbidity, Mortality, and Air Pollution Study (NMMAPS)
is by far the largest multicity study conducted to date to investigate the mortality effects of air
pollution, but its primary focus was PMi0. It should also be noted that, according to the table of
mean pollution levels in the original report (Samet et al., 2000), S02 data were missing in 28 of
90 cities. Annual 24-h avg mean SO2 levels ranged from 0.4 ppb (Riverside, CA) to 14.2 ppb
(Pittsburgh, PA), with a mean of 5.9 ppb during the study period of 1987 to 1994. The analysis in
the original report used GAM models with default convergence criteria. Dominici et al. (2003)
reanalyzed the data using GAM with stringent convergence criteria as well as using GLM. It
should be noted that this model's adjustment for weather effects employs more terms than other
time-series studies in the literature, suggesting that the model adjusts for potential confounders
more aggressively than the models in other studies.
Figure 3-10 shows the all-cause mortality excess risk estimates for S02 from Dominici
et al. (2003). The mortality excess risk estimate with a 1-day lag was 0.60% (95% CI: 0.26, 0.95)
per 10 ppb increase in 24-h avg SO2. PM10 and O3 (in summer) appeared to be more strongly
associated with mortality compared to the other gaseous pollutants. The model with PM10 and
May 2008
3-71
DRAFT—DO NOT QUOTE OR CITE
-------
1 NO2 resulted in an appreciably reduced SO2 excess risk estimate, 0.38% (95% CI: -0.62, 1.38)
2 per 10 ppb increase in 24-h avg SO2. These results suggest that the observed SCVmortality
3 association could be confounded by PMi0 and N02. The authors stated that the results did not
4 indicate associations of SO2, NO2, and CO with all-cause mortality.
1.5.
£ ,0
I 0.5
W
0 ¦
0 -0.5-
X
m
5? -1.0-
-1.5
Lag 0
Lag 1
Lag 2
A = SOj alone
8 = S02 + PMW
C = SO* + PM1(l + 03
d = so2 + pm]0+no2
E = SO, + PM.. + CO
B
BCD
Models
"I
D
Figure 3-10. All cause mortality excess risk estimates for S02 from the National Morbidity,
Mortality, and Air Pollution Study. Posterior means and 95% posterior intervals of national
average estimates of S02 effects on all-cause mortality from non-external causes per 10 ppb
increase in 24-h avg S02 at 0,1, and 2-day lags within sets of the 62 cities with pollutant data
available.
Source: Dominici et al. (2003).
3.3.2.1.3. Canadian Multicity Studies
5 There have been three Canadian multicity studies conducted by the same group of
6 investigators examining the association between mortality and short-term exposure to air
7 pollutants (Burnett et al., 1998; 2000b; 2004). This section focuses on Burnett et al. (2004) as
8 this study is the most extensive Canadian multicity study, both in terms of the length and
9 coverage of cities. The discussion in this study focused on NO2, because NO2 was the best
May 2008
3-72
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
predictor of short-term mortality fluctuations among the pollutants. This was also the case in the
Burnett et al. (1998) study of the gaseous pollutants in 11 Canadian cities. The mean 24-h avg
S02 levels across the 12 cities was 5.8 ppb, with city means ranging from 1 ppb in Winnipeg to
10 ppb in Halifax. The population-weighted average was 5 ppb. The mean SO2 levels in this
study were similar to those in the NMMAPS (mean 24-h avg SO2 levels across the 62 NMMAPS
cities was 5.9 ppb).
All-cause (nonaccidental), cardiovascular, and respiratory mortality were analyzed in
Burnett et al. (2004). For SO2, PM2.5, PM10-2.5, PM10 (arithmetic addition of PM2.5 and PM10-2.5),
CoH, and CO, the strongest mortality association was found at a 1-day lag, whereas for NO2, it
was the 3-day moving average (i.e., average of 0, 1, and 2-day lags), and for O3, it was the 2-day
moving average. The daily 24-h avg values showed stronger associations than the daily 1-h max
values for all the gaseous pollutants and CoH except for O3. The SO2 all-cause mortality excess
risk estimate was 0.74% (95% CI: 0.29, 1.19) per 10 ppb increase in the 24-h avg SO2 with a 1-
day lag. After adjusting for NO2, the SO2 effect estimate was reduced to 0.42% (95% CI: 0.01,
0.84), while the N02 effect estimate was only slightly affected. In this analysis, no regression
analysis using both SO2 and PM was conducted. The Burnett et al. (2000) analysis observed that
the simultaneous inclusion of SO2 and PM2.5 in the model reduced the SO2 effect estimate by
half, whereas the PM2.5 estimate was only slightly reduced. Overall, these results suggest that
SO2 was not an important predictor of daily mortality in the Canadian cities and that its mortality
associations could be confounded with N02 or PM.
3.3.2.1.4. Air Pollution and Health: A European Approach
Several Air Pollution and Health: a European Approach (APHEA) analyses have reported
S02 mortality excess risk estimates. Katsouyanni et al. (1997) examined the association of PMi0,
BS, and SO2 with all-cause mortality in 12 European cities using the standard APHEA (GLM)
approach. The same data set was reanalyzed to adjust for the seasonal cycles (Samoli et al., 2001;
2003). The reanalysis by Samoli et al. (2003) produced results that were similar to those in the
original analysis by Katsouyanni et al. (1997). Since the original analysis presented more results,
including multipollutant model results, discussion will focus on this analysis.
The study by Katsouyanni et al. includes seven western European cities (Athens,
Barcelona, Cologne, London, Lyon, Milan, and Paris) and five central eastern European cities
May 2008
3-73
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
(Bratislava, Kracow, Lodz, Poznan, and Wroclaw). The data covered at least 5 consecutive years
for each city within the years 1980 through 1992. The SO2 levels in these cities were generally
higher than in the United States or Canada, with the median 24-h avg S02 ranging from 5 ppb in
Bratislava to 28 ppb in Kracow. Analysis was restricted to days when PM and SO2
concentrations did not exceed 200 |ig/m3 (76 ppb for SO2) to evaluate the effects of moderate to
low exposures. The data were analyzed by each center separately following a standardized
method, but the lag for the "best" model was allowed to vary in these cities from 0 to 3 days. The
city-specific effect estimates were then examined in the second stage for source of heterogeneity
using city-specific variables such as mean pollution and weather variables, accuracy of the air
pollution measurements, health of the population, smoking prevalence, and geographical
differences.
The city-specific estimates were found to be heterogeneous and, among the explanatory
variables, only the separation between western and central eastern European cities resulted in
more homogeneous groups. The all-cause mortality excess risk estimates were 1.14% (95% CI:
0.88, 1.39), 1.99% (95% CI: 1.15, 2.83), and 0.46% (95% CI: -0.23, 1.15) for all the 12 cities
combined, western cities, and central eastern cities, respectively, per 10 ppb increase in the
24-h avg SO2 at variable single-day lags. Zmirou et al. (1998) analyzed cardiovascular and
respiratory mortality in 10 of the 12 APHEA cities and observed that the cause-specific mortality
excess risk estimates were higher than those for all-cause mortality. As in the analyses of all-
cause mortality, S02 effect estimates for these cause specific deaths were higher in western
European cities than in central eastern European cities.
Seasonal analyses indicated that the summer estimate was slightly higher than the winter
estimate in the western cities, but the difference was not statistically significant. The results for
the two-pollutant model with S02 and BS were presented for the western cities, with a similar
extent (-30%) of reductions in the estimates of both pollutants (1.31% [95% CI: 0.40, 2.23] for
SO2). Furthermore, for western cities, they estimated effects for SO2 for days with high or low
BS levels and the corresponding BS effects for days with high or low SO2 levels and found that
the effects were similar in the stratified data. From these results, Katsuoyanni et al. (1997)
suggested that the effects of the two pollutants were independent.
Overall, the APHEA studies provide some suggestive evidence that the effect of short-term
exposure to SO2 on mortality is independent of PM. This is somewhat in contrast to the U.S. and
May 2008
3-74
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Canadian studies. The SO2 levels were much higher in the European cities, but the type of PM
constituents also might be different.
3.3.2.1.5. The Netherlands Study
In the Netherlands studies by Hoek et al. (Hoek et al., 2000; 2001; reanalysis, Hoek, 2003),
the association between air pollutants and mortality were examined in a large population (14.8
million for the entire country) over the period of 1986 through 1994. The Netherlands were not
part of the APHEA analysis. The median 24-h avg SO2 level in the Netherlands was 4 ppb (6 ppb
for the four major cities). All the pollutants examined, including PM10, BS, O3, NO2, SO2, CO,
sulfate, and nitrate, were associated with all-cause mortality, and for single-day models, a 1-day
lag showed the strongest associations for all the pollutants. The following effect estimates are all
from the GLM models with natural splines for smoothing functions. The SO2 excess risk
estimate in a single-pollutant model was 1.31% (95% CI: 0.69, 1.93) per 10 ppb increase in
24-h avg SO2 at a 1-day lag and 1.78% (95% CI: 0.86, 2.70) at an average of 0- to 6-day lag.
Seasonal analyses showed slightly greater effect estimates during the summer compared to the
winter. SO2 was most highly correlated with BS (r = 0.70). The simultaneous inclusion of SO2
and BS reduced the effect estimates for both pollutants (SO2 effect estimate was 1.07% [95% CI:
-0.27, 2.42] per 10 ppb increase with an average of 0- to 6-day lag of 24-h avg SO2). PM10 was
less correlated with S02 (r = 0.65), and the simultaneous inclusion of these pollutants resulted in
an increase in the S02 effect estimate. These results from the analysis of the Netherlands data
suggested some indication of confounding between SO2 and BS.
Cause specific analysis showed larger excess risk estimates for COPD (3.61% [95% CI: -
0.29, 7.67] per 10 ppb increase in the average of 0- through 6-day lags of 24-h avg SO2) and
pneumonia (6.56% [95% CI: 1.16, 12.24]) deaths compared to that from all causes, but because
essentially all of the pollutants showed larger effect estimates for these sub-categories, it is
difficult to interpret these estimates as effects of SO2 alone. Similarly, the effect estimates for
heart failure (7.1% [95% CI: 2.6, 11.7]) and thrombosis-related deaths (9.6% [95% CI: 3.1,
16.6]) were larger than that for total cardiovascular (2.7% [95% CI: 1.3, 4.1]) causes. Since the
same pattern was seen for other pollutants as well, it is difficult to interpret these cause-specific
effect estimates due to SO2 alone or any one of the pollutants analyzed.
May 2008
3-75
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
3.3.2.1.6. Other European Multicity Studies
Other European multicity studies were conducted in 8 Italian cities (Biggeri et al., 2005), 9
French cities (Le Tertre et al., 2002), and 13 Spanish cities (Ballester et al., 2002). The studies by
Le Tertre et al. and Ballester et al. were conducted using GAM methods with the default
convergence setting.
Biggeri et al. analyzed eight Italian cities (Turin, Milan, Verona, Ravenna, Bologna,
Florence, Rome, and Palermo) for mortality and hospital admissions (mortality data were not
available for Ravenna and Verona). The study period varied from city to city between 1990 and
1999. Only single-pollutant models were examined in this study. The SO2 excess risk estimates
were 4.14% (95% CI: 1.05, 7.33), 4.94% (95% CI: 0.41, 9.67), and 7.37% (95% CI: -3.58,
19.57) per 10 ppb increase with an average of 0- and 1-day lag of 24-h avg SO2 for all-cause,
cardiovascular, and respiratory deaths, respectively. Since all the pollutants showed positive
associations with these mortality categories and the correlations among the pollutants were not
presented, it is not clear how much of the observed associations are shared or confounded. The
mortality excess risk estimates were not heterogeneous across cities for all the gaseous
pollutants. It should be noted that in Turin, Milan, and Rome, the mean SO2 values declined by
50% from the first half to the second half of the study period, while the levels of other pollutants
declined by smaller fractions. This also complicates the interpretation of S02 effect estimates in
this study, which are much higher than those from the APHEA studies.
The Le Tertre et al. study of nine French cities examined BS, SO2, NO2, and O3 by
generally following the APHEA protocol, but using GAM with default convergence criteria and
using the average of lags 0 and 1 day for combined estimates. SO2 data were not available in one
of the nine cities (Toulouse). All four pollutants were positively associated with mortality
outcomes. The study did not report descriptions of correlation among the pollutants or conduct
multipollutant models, and therefore, it is difficult to assess the potential extent of confounding
among these pollutants. The SO2 effect estimates were homogeneous across cities, with the
exception of Bordeaux, which was the only city that used strong acidity as a proxy for S02.
The Spanish Multicentre Study on Air Pollution and Mortality (EMECAM) examined the
association of PM indices (i.e., PM10, TSP, BS) and SO2 with mortality in 13 cities (Ballester et
al., 2002). These studies followed the APHEA protocol, but using the GAM approach. The daily
mean 24-h avg S02 concentrations ranged from 3 to 17 ppb. In the seven cities where 1-h max
May 2008
3-76
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
SO2 data were also available, mean concentrations ranged from 21 to 43 ppb. The combined
effect estimates for all-cause and respiratory mortality were statistically significant for both
24-h avg S02 and 1-h max S02. Controlling for PM indices substantially diminished the effect
estimates for 24-h avg SO2, but not for 1-h max SO2. The authors reported that these results
could indicate an independent impact of peak values of SO2 more than an effect due to a longer
exposure.
3.3.2.2. Meta-Analyses of Air Pollution-Related Mortality Studies
3.3.2.2.1. Meta-Analysis of All Criteria Pollutants
Stieb et al. (2002) reviewed time-series mortality studies published between 1985 and
2000, and conducted a meta-analysis to estimate combined effects for PM10, CO, NO2, O3, and
SO2. Since many of the studies reviewed in that analysis used GAM with default convergence
parameters, Stieb et al. (2003) updated the estimates by separating the GAM versus non-GAM
studies. In addition, separate combined estimates were presented for single- and multipollutant
models. There were more GAM estimates than non-GAM estimates for all the pollutants except
for SO2. For SO2, there were 29 non-GAM estimates from single-pollutant models and 10
estimates from multipollutant models. The lags and multiday averaging used in these estimates
varied. The combined estimate for all-cause mortality was 0.95% (95% CI: 0.64, 1.27) per
10 ppb increase in 24-h avg SO2 from the single-pollutant models and 0.85% (95% CI: 0.32,
1.39) from the multipollutant models. Because these estimates are not from an identical set of
studies, the difference (or lack of a difference, as in this case) between the two estimates may not
necessarily be due to the effect of adding a copollutant in the model. Note that the data extraction
procedure of this meta-analysis for the multipollutant models was to include from each study the
multipollutant model that resulted in the greatest reduction in effect estimates compared with that
observed in single-pollutant models. It should also be noted that all the multicity studies whose
combined estimates have been discussed in the previous section were published after this meta-
analysis.
3.3.2.2.2. Health Effects Institute Review of Air Pollution Studies in Asia
The Health Effects Institute (HEI) conducted a comprehensive review of air pollution
health effects studies (HEI, 2004). They summarized the results from mortality and hospital
May 2008
3-77
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
admission studies of the health effects of ambient air pollution in Asia (East, South, and
Southeast) published in peer-reviewed scientific literature from 1980 through 2003. Of the 138
papers the report identified, most were studies conducted in East Asia (mainland China, Taipei,
Hong Kong, South Korea, and Japan). The levels of SO2 in these Asian cities were generally
higher than in U.S. or Canadian cities, with more than half of these studies reporting mean
24-h avg SO2 levels of > 10 ppb. Based on a comparison of the reported mean SO2 levels from
the same cities in different time periods, it is clear that the S02 levels declined significantly in
the 1990s. The meta-analysis used the most recent estimate for each city to reflect recent
pollution levels. Based on the criteria of having at least one year of data, model adjustment for
major time-varying confounders, and reporting effect estimates per unit increase in air pollution,
the meta-analysis included 28 time-series studies (11 from South Korea, 6 from mainland China,
6 from Hong Kong, and 1 each from Taipei, India, Singapore, Thailand, and Japan). The lags
selected to compute combined estimates were inevitably variable; a systematic approach was
used to favor the a priori lag stated in the study, followed by the most significant lag, and then
the largest effect estimate.
Among the health outcomes examined in the meta-analysis, all-cause mortality was
addressed in the largest number of studies (13 studies) and SO2 was the most frequently studied
pollutant (11 studies). The report generally focused on the results of single-pollutant models, as
there were too few studies with results of comparable multipollutant models to allow meaningful
analysis. The S02 mortality effect estimates showed evidence of heterogeneity. The combined
estimate for all-cause mortality was 1.49% (95% CI: 0.86, 2.13) per 10 ppb increase in 24-h avg
SO2. One of the limitations noted in the report was that some degree of publication bias was
present in these studies.
3.3.3. Evidence of the Effect of S02 on Mortality from an Intervention
Study
Many time-series studies provide estimates of excess risk of mortality, but a question
remains as to the likelihood of a reduction in deaths when SO2 levels are actually reduced. A
sudden change in regulation in Hong Kong in July 1990 required the conversion to fuel oil with
low sulfur content. The reduction in respiratory symptoms among children living in the polluted
district in Hong Kong after the intervention were previously discussed in Section 3.1.6. Hedley
May 2008
3-78
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
et al. (2002) examined changes in mortality rates following the intervention. The SO2 levels after
the intervention declined about 50% (from about 17 ppb to 8 ppb), but the levels for PM10, NO2,
and sulfate did not change and 03 levels slightly increased. The seasonal mortality analysis
results showed that the apparent reduction in seasonal death rate occurred only during the first
winter, and this was followed by a rebound (i.e., higher than expected death rate) in the
following winter, then returned to the expected pattern three to five years after the intervention.
Using Poisson regression of the monthly deaths, the average annual trend in death rate
significantly declined after the intervention for all causes (2.1%), respiratory causes (3.9%), and
cardiovascular causes (2.0%), but not from other causes. These results seem to suggest that a
reduction in SO2 leads to an immediate reduction in deaths and a continuing decline in the annual
trend in death rates. Hedley et al. estimated that the expected average gain in life expectancy per
year due to the lower SO2 levels was 20 days for females and 41 days for males.
Interpreting these results is somewhat complicated by an upward trend in mortality across
the intervention point, which the authors noted was due to increased population size and aging.
The results suggest that such an upward trend is less steep after the introduction of low sulfur
fuel. While the Poisson regression model of monthly deaths does adjust for trend and seasonal
cycles, the regression model does not specifically address the influence of influenza epidemics,
which can vary from year to year. This issue also applies to the analysis of warm to cool season
change in death rates. The most prominent feature of the time-series plot (or the fitted annual
cycle of monthly deaths) presented in this study is the lack of a winter peak for respiratory and
all-cause mortality during the year immediately following the intervention. Much could be made
of this lack of a winter peak, but no discussion of the potential impact of (or a lack of) influenza
epidemics is provided. These issues complicate the interpretation of the estimated decline in
upward trend of mortality rate or the apparent lack of winter peak.
The decline in mortality following the intervention does not preclude the possibility that
other constituents of the pollution mixture that share the same source as SO2 are responsible for
the adverse effects. Even though PM10 levels before and after the intervention were stable in
Hong Kong, it is possible that constituents that do not explain a major fraction of PM may have
declined. As also noted previously in Section 3.1.6, Hedley et al. (2006) noted large reductions in
ambient nickel and vanadium concomitantly with reductions of sulfur after the intervention. SO2
also may be serving as a modifier of the effect of respirable particles. Thus, while the Hong
May 2008
3-79
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Kong data are supportive of SCVmortality effects, the possibility remains that mortality effects
may be caused by constituents of S02-associated sources.
3.3.4. Summary of Evidence on the Effect of Short-Term S02 Exposure
on Mortality
The epidemiological evidence on the effect of short-term exposure to SO2 on all-cause
(nonaccidental) and cardiopulmonary mortality is suggestive but not sufficient to infer a causal
relationship at ambient concentrations. The epidemiological studies are generally consistent in
reporting positive associations between SO2 and mortality; however, there was a lack of
robustness of the observed associations to adjustment for copollutants.
Figure 3-11 presents all-cause SO2 mortality excess risk estimates from the multicity
studies and meta-analyses. The mortality effect estimates from single-pollutant models range
from 0.6% (NMMAPS) to 4.1% (Italian 8-cities study) per 10 ppb increase in 24-h avg SO2
concentrations, but given the large confidence band in the Italian study, a more stable range may
be 0.6 to 2%). It is noteworthy that the SO2 effect estimates for the NMMAPS and Canadian 12-
city studies are quite comparable (0.6 and 0.7%, respectively), considering the differences in the
modeling approach. The heterogeneity of estimates in the multicity studies and meta-analyses
may be due to several factors, including the differences in model specifications, averaging/lag
time, SO2 levels, and effect-modifying factors. Only the APHEA study examined possible
sources of heterogeneity for S02-related mortality. They examined several potential effect
modifiers such as the mean levels of pollution and weather variables, accuracy of the air
pollution measurements, health of the population, smoking prevalence, and geographical
differences. The only variable that could explain the heterogeneity of city-specific effect
estimates was the geographic separation (western versus central eastern European cities) for both
SO2 and BS, but heterogeneity in the SO2 effect estimates remained within the western cities.
May 2008
3-80
DRAFT—DO NOT QUOTE OR CITE
-------
Relative Risk
Reference Study Lag
Dominici et al. (2003) NMMAPS 90 cities study 1
1.00 1.01 1.02 1.03 1.04
l i i i i
i I • Single pollutant
: jo Multipollutant
< —*—
:
i—o {With PM10 and N02)
Burnett et al. (2004) Canadian 12 cities study 0-2
! —
!
) o (With NOg)
Katsouyanni et al. {1997) AFHEA1 {12 European cities) Variable
i ,
7 W. European cities
!
i •
¦ O (With BS)
5 Central E. European cities
=|—
Biggeri et al. (2005) Italian 8 cities study 0-1
Hoek (2003) The Netherlands study 1
!
0-6
I •
;
—i O {With BS)
I
Stieb et al. (2003) Meta-analysis, international Variable
! .
1
: O (With various copolkitants)
[
Health Effects Institute {2004} Meta-analysis, Asian cities Variable
;
; *
9
Figure 3-11. Relative risks (95% CI) of S02-associated all-cause (nonaccidental) mortality, with and
without copollutant adjustment, from multicity and meta-analysis studies. Effect estimates are
standardized per 10 ppb increase in 24-h avg S02 concentrations. For multipollutant models,
results from the models that resulted in the greatest reduction in the S02 effects are shown.
(NMMAPS = National Morbidity, Mortality, and Air Pollution Study; APHEA = Air Pollution and
Health: a European Approach)
May 2008
3-81
DRAFT—DO NOT QUOTE OR CITE
-------
Lcsatieii Stwty
Katsouyanni et al. (1997) APHEA1 {7 W. European cities)
Zmirou et al. (1998) APHEA1 (5 W. European cities)
Biggeri et al. (2005)
Hoek (2003)
Italian 8 cities study
'Le Tertre et al. (2002) French 9 cities study
*Batester et al. (2002) Spanish 13 cities study
The Netherlands study
0.95 1.00
Relative Risk
1.05 1.10 1,15
1.20
m
Variable
Variable
0-1
0-1
0-1
0-6
*-
X All cause
• Respiratory
O Cardiovascular
-m (COPD)
• (Pneumonia)
Figure 3-12. Relative risks (95% CI) of S02-associated mortality for all (nonaccidental), respiratory,
and cardiovascular causes from multicity studies. Effect estimates are standardized per 10 ppb
increase in 24-h avg S02 concentrations. (APHEA = Air Pollution and Health: a European
Approach)
*Note: Le Tertre et alF. (2002) and Ballester et al. (2002) performed analyses using Poisson GAM with default
convergence criteria.
May 2008
3-82
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Several multicity studies provided effect estimates for broad cause-specific categories,
typically respiratory and cardiovascular mortality. A summary of these effect estimates, along
with the all-cause mortality estimates for comparison, are presented in Figure 3-12. These results
from multicity studies suggest that the mortality effect estimates for cardiovascular and
respiratory causes were generally larger than that for all-cause mortality, though in some cases
the effects were not statistically significant, possibly because of reduced statistical power by
which to examine cause-specific associations. In these studies, the effect estimates for respiratory
mortality were also found to be larger than the cardiovascular mortality effect estimates,
suggesting a stronger association of SO2 with respiratory mortality compared to cardiovascular
mortality. This suggestive finding is consistent with the observed greater effects of SO2 on
respiratory morbidity compared to cardiovascular morbidity.
As shown previously in Figure 3-11, the mortality effect estimates from the multipollutant
models in the multicity studies suggest some extent of confounding between SO2 and PM and/or
results from multicity studies generally suggest some evidence of confounding, in the sense of
instability of effect estimates in multipollutant models. It should be noted, however, that
interpretation of the single- versus multipollutant model results are complicated by potential
interaction among copollutants and differing degrees of measurement error for correlated
pollutants.
Very few studies specifically examined possible interactions among the copollutants.
Katsouyanni et al. (1997) examined the effect estimates for S02 and BS in seven western
European cities for subsets stratified by high and low levels of the other pollutant and found that
the estimates were similar for days with low or high levels of the other pollutant. From these
results, Katsuoyanni et al. suggested that the effects of SO2 and BS were independent.
Other multi- or single-city studies did not consider examination of possible interaction
effects between SO2 and copollutants.
In summary, recent epidemiological studies have reported associations between mortality
and SO2, often at mean 24-h avg levels of < 10 ppb. The range of the excess risk estimates for
S02 on all-cause mortality is 0.4 to 2% per 10 ppb increase in 24-h avg S02 in several multicity
studies and meta-analyses. The effect estimates for more specific categories may be larger. The
larger European study suggests that the observed heterogeneity in SO2 effect estimates is at least
in part regional. The intervention study from Hong Kong supports the idea that a reduction in
May 2008
3-83
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
SO2 levels results in a reduction in deaths, but this does not preclude the possibility that the
causal agent is not SO2 but rather something else that is associated with SO2 sources. Results
from the multicity studies suggest that S02-mortality excess risk estimates may be confounded to
some extent by copollutants, making a definitive distribution of effects among the pollutants
difficult. However, the interpretation of multipollutant model results also requires caution
because of possible interaction among the copollutants and influence of varying measurement
error. Very limited information was available to determine possible interaction effects between
SO2 and PM or other copollutants. Overall, the evidence that SO2 is causally related to mortality
at current ambient levels is suggestive, but limited by potential confounding and lack of
understanding regarding the interaction of SO2 with copollutants in the epidemiological data.
3.4. Morbidity Associated with Long-Term S02
Exposure
3.4.1. Summary of Findings from the Previous Review
The 1982 AQCD addressed some effects of long-term SO2 exposure. It was reported that
bronchoconstriction resulted from chronic exposure to 5.1 ppm SO2 in dogs but not in monkeys.
This increased pulmonary resistance was thought to occur as a result of morphological changes
in the airway or hypersecretion of mucus leading to airway narrowing. However, there were no
remarkable pulmonary pathological findings in monkeys and dogs in these studies. This could
have been due to the conventional light microscopic examination applied, which could not detect
alterations in surface membranes or subtle changes in cilia.
It was also noted that repeated exposures of rats > 50 ppm S02 produced a chronic
bronchitis similar to that seen in humans although there was no evidence to suggest that
bronchitis developed in humans at ambient levels of SO2. Furthermore, nasal mucosal alterations
were observed in mice exposed to 10 ppm SO2 for 72 h by inhalation. Lack of data on
morphological effects of SO2 at near ambient concentrations was noted. In addition, some
alterations in lung host defenses were discussed with chronic exposure to S02 at doses exceeding
ambient concentrations.
In the 1982 AQCD, only a few epidemiological studies provided sufficient quantitative
evidence relating respiratory symptoms or pulmonary functions changes to long-term exposure
May 2008
3-84
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
to SO2. Briefly, a study by Lunn et al. (1967) in Sheffield, England, provided the strongest
evidence of an association between pulmonary function decrements and increased frequency of
lower respiratory tract symptoms in 5- to 6-year-old children chronically exposed to ambient BS
(annual level of 230 to 301 |ig/m3) and SO2 levels (69 to 105 ppb). A follow-up study in 1968 by
Lunn (1970) found no effect with much lower levels of BS (range: 48, 169 |ig/m3) and SO2
(range: 36, 97 ppb); it was suggested that this might be due to insufficient power to detect small
health effect changes.
The 1986 Second Addendum presented three additional studies that examined the effects of
long-term exposure on respiratory health. A study by Ware et al. (1986) reported that respiratory
symptoms were associated with annual average TSP in the range of -30 to 150 |ig/m3 in children
(n = 8,380) from six U.S. studies. Only cough was found to be significantly associated with S02.
Although the increase in symptoms did not appear concomitantly with any decrements in lung
function, this may indicate different mechanisms of effect. Other studies by Chapman et al.
(1985) and Dodge et al. (1985) also observed increased prevalence of cough among children and
young adults living in areas of higher S02 concentrations; however, it was noted that the
observed effects might have been due to intermittent high SO2 peak concentrations.
In addition to respiratory effects from long-term exposure to SO2, the potential
carcinogenicity of SO2 or other SOx was also examined in the previous review. The 1982 AQCD
concluded that little or no clear epidemiological evidence substantiated the hypothesized links
between S02 or other SOx and cancer, though there was some animal toxicological evidence that
led to the conclusion that SO2 may be considered a suspect carcinogen/cocarcinogen. There was
very limited consideration of the effects of long-term exposure to SO2 on other organ systems.
Since the 1982 AQCD and the 1986 Second Addendum, a number of animal toxicological
and epidemiological studies have investigated the effect of long-term exposure to S02 on
respiratory morbidity, including asthma, bronchitis and respiratory symptoms, lung
function, morphological effects, and lung host defense. Additional studies have examined the
effect of long-term SO2 exposure on genotoxic and carcinogenic effects, cardiovascular effects,
and prenatal and neonatal outcomes, which are also briefly discussed in this section.
May 2008
3-85
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
3.4.2. Respiratory Effects Associated with Long-Term Exposure to S02
3.4.2.1. Asthma, Bronchitis, and Respiratory Symptoms
Several epidemiological studies have examined the association between long-term
exposure to SO2 and other air pollutants on asthma, bronchitis, and a variety of respiratory
symptoms. These studies are summarized in Annex Table F-l. In the Six Cities Study of Air
Pollution and Health, cross-sectional associations between air pollutants and respiratory
symptoms were examined in 5,422 white children aged 10 to 12 years old from Watertown, MA,
St. Louis, MO, Portage, WI, Kingston-Harriman, TN, Steubenville, OH, and Topeka, KS
(Dockery et al., 1989). Annual means of 24 h avg SO2 concentrations ranged from 3.5 ppb in
Topeka to 27.8 ppb in Steubenville. Except for O3, the correlations among pairs of pollution
measures varied between 0.53 and 0.98. No associations were observed between S02 and a
variety of respiratory symptoms, including bronchitis, chronic cough, chest illness, persistent
wheeze, and asthma. Stronger associations were observed for PM indices.
Dockery et al. (1996) examined the respiratory health effects of acid aerosols in 13,369
white children aged 8 to 12 years old from 24 communities in the United States and Canada
between 1988 and 1991. The city-specific annual mean SO2 concentration was 4.8 ppb, with a
range of 0.2 to 12.9 ppb. With the exception of the gaseous acids, nitrous and nitric acid, none of
the particulate or gaseous pollutants, including SO2, were associated with increased asthma or
any asthmatic symptoms. Stronger associations with particulate pollutants were observed for
bronchitis and bronchitic symptoms. For SO2, the only significant association found was with
chronic phlegm, with an OR of 1.19 (95% CI: 1.00, 1.40) per 5 ppb increase in SO2.
Herbarth et al. (2001) performed a meta-analysis of three cross-sectional surveys
conducted in East Germany investigating the relationship between lifetime exposure (from birth
to completion of questionnaire survey) to SO2 and TSP in children and the prevalence of chronic
bronchitis. Using a logistic model that included variables on parental predisposition (mother or
father with bronchitis) and environmental tobacco smoke exposure, the authors reported that the
OR for bronchitis due to a lifetime exposure to S02 was 3.51 (95% CI: 2.56, 4.82) (the
concentration change for which the OR was based was not presented). No associations were
found between TSP and the prevalence of bronchitis in children.
May 2008
3-86
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
As part of the international SAVIAH (Small-Area Variation in Air Pollution and Health)
study, Pikhart et al. (2001) examined the respiratory health effects from long-term exposure to
S02 in children (n = 6,959) from two central European cities with high pollution levels (Prague,
Czech Republic, and Poznan, Poland). A novel technique was used to estimate the outdoor
concentrations of SO2 at a small-area level. Outdoor SO2 was measured by passive samplers at
130 sites in the two cities during 2-week periods. Concentrations of SO2 at each location in the
study areas were estimated from these data by modeling using a geographic information system
(GIS). The estimated mean exposure to outdoor SO2 was 32 ppb, (range: 25, 37) in Prague and
31 ppb, (range: 17, 53) in Poznan. The prevalence of wheezing or whistling in the past 12
months was associated with SO2 (OR of 1.08 [95% CI: 1.03, 1.13] per 5 ppb increase in SO2).
Moreover, the lifetime prevalence of wheezing or whistling (OR 1.03 [95% CI: 1.00, 1.07]) and
lifetime prevalence of physician-diagnosed asthma (OR 1.09 [95% CI: 1.00, 1.19]) also were
associated with SO2 levels. In the SAVIAH study, the only other pollutant considered in relation
to health outcomes was NO2. An earlier publication by Pikhart et al. (2000) presented
preliminary results of the Prague data and indicated that the observed associations between N02
and respiratory symptoms were generally similar to that of SO2.
The International Study of Asthma and Allergies in Children (ISAAC) included thousands
of children in several European countries and Taiwan (Hirsch et al., 1999; Hwang et al., 2005;
Penard-Morand et al., 2005; Ramadour et al., 2000; Studnicka et al., 1997). Penard-Morand et al.
examined the effect of long-term exposures to air pollution and prevalence of exercise-induced
bronchial reactivity (EIB), flexural dermatitis, asthma, allergic rhinitis, and atopic dermatitis in
9,615 children aged 9 to 11 years in six French communities. Using 3-year averaged
concentrations of SO2, the investigators reported that the prevalence of exercise-induced
bronchial reactivity, lifetime asthma, and allergic rhinitis were significantly associated with
increases in SO2 exposure. The estimated 3-year averaged concentration of SO2 was 2 ppb in the
low-exposure schools and 4 ppb in the high-exposure schools. In a single-pollutant model, the
ORs were 2.37 (95% CI: 1.44, 3.77) for EIB and 1.58 (95% CI: 1.00, 2.46) for lifetime asthma
per 5 ppb increase in S02. In this study, S02 was correlated with PMi0 (r = 0.76) but not with 03
(r = -0.02). Using a two-pollutant model that included PM10, the associations of SO2 with EIB
and lifetime asthma were fairly robust (< 5% change).
May 2008
3-87
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
In a German study of 5,421 children, the annual mean SO2 concentration was associated
with morning cough reported in the last 12 months, but not bronchitis (Hirsch et al., 1999). This
study further observed that the association of S02 and other air pollutants with respiratory
symptoms were stronger in nonatopic than in atopic children. The authors noted that these
findings were in line with the hypothesis that these air pollutants induce nonspecific irritative
rather than allergic inflammatory changes in the airway mucosa, as irritative effects would affect
the clinical course in nonatopic children more strongly than in atopics whose symptoms are also
determined by allergen exposure.
In contrast to the studies noted above, other studies using the ISAAC protocol did not
observe an association between long-term exposure to SO2 and respiratory symptoms. In France,
(Ramadour et al., 2000) performed an epidemiological survey of 2,445 children aged 13 to 14
years living in communities with contrasting levels of air pollution to determine the relationship
between long-term exposure to gaseous air pollutants and prevalence rate of rhinitis, asthma, and
asthma symptoms. The average SO2 concentrations during the 2-month survey period ranged
from 7 ppb to 22 ppb across the seven communities. This study found no relationship between
the mean levels of SO2, NO2, or O3 and the above-mentioned symptoms. Another study of 843
children from eight nonurban communities in Austria did not observe consistent associations
between SO2 and prevalence of asthma and symptoms (Studnicka et al., 1997). Compared to the
lowest SO2 concentration category, the ORs in the higher SO2 concentration categories (third and
fourth quartiles) did not exceed one for any of the symptoms examined (wheeze, cough,
bronchitis, and asthma).
A cohort study was conducted by (Goss et al., 2004) to examine the effect of air pollutants
on a potentially susceptible population, patients with cystic fibrosis. Study participants included
11,484 patients (mean age 18.4 years) enrolled in the Cystic Fibrosis Foundation National Patient
Registry in 1999-2000. Exposure was assessed by linking air pollution values from ambient
monitors with the patient's home ZIP code. During the study period, the mean SO2 concentration
was 4.9 ppb (SD 2.6, IQR: 2.7, 5.9). This study found no association between SO2 and the odds
of having two or more pulmonary exacerbations. One of the limitations addressed by the authors
was the lack of information regarding tobacco use or environmental tobacco smoke, an important
risk factor for pulmonary exacerbations.
May 2008
3-88
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Several studies examined the effects of long-term exposure to SO2 on asthma, bronchitis,
and respiratory symptoms. The studies reported positive associations in children; the notable
exception was the Harvard Six Cities Study. However, there were inconsistencies in the results
observed: some found effects on bronchitic but not asthmatic symptoms; others found the
converse. A major limitation was that some subjects were asked to recall prevalence of
symptoms in the last 12 months or in a lifetime; such long recall periods may have caused
significant recall bias. Another concern is the high correlation of long-term average S02 and
copollutant concentrations, particularly PM, and the very limited evaluation of potential
confounding in these studies. Overall, while the evidence is suggestive, the variety of outcomes
examined and the inconsistencies in the observed results make it difficult to assess the direct
impact of long-term exposure of S02 on asthma, bronchitis, or respiratory symptoms.
3.4.2.2. Lung Function
Only a few new animal toxicological studies involving longer-term inhalation exposures to
S02 were conducted since the last review. These studies are summarized here and in Annex Table
E-l. Rabbits that were neonatally immunized to Alternaria tenuis and exposed to 5 ppm SO2 for
13 weeks beginning in the neonatal period (Douglas et al., 1994) did not demonstrate alterations
in lung resistance, dynamic compliance, trans-pulmonary pressure, tidal volume, respiration rate
or minute volume. Similarly, no changes in physiological function were noted in dogs exposed to
15 ppm S02 for 2 h/day and 4-5 days/week for 5 months (Scanlon et al., 1987), although changes
were noted at 50 ppm. However, Smith et al. (1989) found decreased residual volume and
quasistatic compliance in rats at 4 months of exposure to 1 ppm SO2 for 5 h/day and 5
days/week.
Only a limited number of epidemiological studies examined the association between long-
term exposure to SO2 and changes in lung function. The Harvard Six Cities Study by Dockery
et al. (1989) reported that no associations were observed between lung function and long-term
exposure to air pollution, including SO2, in a cohort of more than 5,000 children. An analysis of
NHANES II data by Schwartz (1989), which included information on children and youths from
44 cities but was limited by a cross-sectional study design, also did not observe an association
with SO2, though inverse associations of FVC and FEVi with annual concentrations of TSP, NO2
and O3 were found. Additional studies conducted in Europe observed mixed results.
May 2008
3-89
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
In a longitudinal cohort study of 1,150 children in nine communities in Austria, Frischer
et al. (1999) examined the effect of long-term exposure to air pollutants on lung function. Lung
function was measured in the spring and fall over a 3-year period from 1994 through 1996.
Annual mean SO2 concentrations ranged from 2 to 6 ppb across the nine communities. The
authors reported no consistent associations between SO2, PM10, or NO2 and lung function. For
SO2, a negative parameter estimate was observed during the summer, but a positive estimate was
found during the winter period. Horak et al. (2002a; b) extended the study of Frischer et al.
(1999) with an additional year of data. The mean SO2 concentration was 6 ppb in the winter and
3 ppb in the summer. This study found a positive association between wintertime SO2
concentrations and changes in FVC, which became null with PM10 in a two-pollutant model.
Jedrychowski et al. (1999) conducted a prospective cohort study of 1,001 preadolescent
children from two areas of Krakow, Poland, that differed in ambient air pollutants. In the city
center, which had higher pollution area, the mean annual level of SO2 was 16.7 ppb (SD 12.5). In
comparison, the mean annual SO2 level in the control area was 12.1 ppb (SD 8.4). A similar
difference in TSP levels was observed between the city center and control area. The adjusted
ORs comparing the city center to the control area for the occurrence of slower lung function
growth over a two-year period were 2.10 (95% CI: 1.27, 3.46) for FVC and 2.10 (95% CI: 1.27,
3.48) for FEVi in boys. The adjusted ORs for girls were 1.54 (95% CI: 0.89, 2.64) for FVC and
1.51 (95% CI: 0.90, 2.53) for FEVi. However, as both TSP and SO2 levels were higher in the city
center, the observed effects on lung function growth cannot be specifically attributable to S02.
One notable study examined the potential effect of long-term exposure to air pollution on
lung function in adults. The study by Ackermann-Liebrich et al. (1997) included 9,651 adults
aged 18 to 60 years old residing in eight different areas in Switzerland (Study on Air Pollution
and Lung Diseases in Adults [SAPALDIA]). They observed a 0.1% decrease in FEVi per 5 ppb
increase in SO2 for adults. Significant associations also were observed for PM10 and NO2. The
limited number of study areas and high intercorrelation between the pollutants made it difficult
to assess the effect of an individual pollutant. The authors concluded that air pollution from fossil
fuel combustion, which was the main source of air pollution for S02, N02, and PMi0 in
Switzerland, was associated with decrements in lung function parameters in this study.
May 2008
3-90
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Collectively, the results from the limited number of animal toxicological and
epidemiological studies do not give support to long-term exposure to ambient SO2 having a
detrimental effect on lung function.
3.4.2.3. Morphological Effects
Three animal toxicological studies of morphological effects resulting from subacute to
chronic S02 exposures have been published since the 1982 AQCD. These studies are
summarized in Annex Table E-ll. No alveolar lesions (including electron microscopic
evaluation) or changes in numbers of tracheal secretory cells were observed in guinea pigs
exposed to 1 ppm SO2 for 3 h/day for 6 days (Conner et al., 1985). No pulmonary or nasal
lesions were observed in rats exposed to 5 ppm S02 for 5 days/week for 4 weeks (Wolff et al.,
1989). A weakness of the latter study is that histopathological methods were not reported.
However, a third study reported histopathological changes in the respiratory system involving
lesions in the bronchioles. Smith et al. (1989) exposed rats for 4 to 8 months to 1 ppm SO2 for
5 h/day and 5 days/week and observed increased incidence of bronchiolar epithelial hyperplasia
and a small increase (12%) in numbers of nonciliated epithelial cells in terminal respiratory
bronchioles at 4 but not 8 months of exposure. A limitation of the study was the examination of a
single concentration, which does not allow for concentration-response assessment or
identification of a no-effect-level.
In summary, results from these animal toxicological studies do not support an association
between long-term exposure to ambient SO2 and prolonged effects on lung morphology.
3.4.2.4. Lung Host Defense
The 1982 AQCD reported some detrimental effects of S02 on lung host defenses that
generally occurred at concentrations exceeding ambient exposure concentrations. In rats exposed
to 0.1 ppm SO2 for ~2 to 3 weeks, clearance of labeled particles from the lung was accelerated at
10 and 23 days following exposure. In rats exposed to 1 ppm for ~2 to 3 weeks, clearance was
accelerated at 10 days and slowed down at 25 days. Tracheal mucus flow was decreased with a
1-year exposure of dogs to 1 ppm SO2, but was unaffected by a 30-minute exposure of donkeys
to 25 ppm SO2. Studies in mice suggested no effect on susceptibility to bacterial infection with
May 2008
3-91
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
exposure to SO2 concentrations of < 5 ppm for 3 months. Antiviral defenses were impaired in
mice exposed to 7-10 ppm SO2 for 7 days. No alterations in pulmonary immune system were
reported with chronic exposure of mice to 2 ppm S02.
Several studies on lung host defense have been conducted since the last review and are
summarized in Annex Table E-4. Only one study published after the last review evaluated
mucociliary clearance in rats after exposure to SO2. In this subchronic study, no effect on
clearance of radiolabeled particles from the lung was observed in rats exposed to 5 ppm S02 for
2 h/day for 4 weeks (Wolff et al., 1989). These findings are in contrast to the altered clearance
reported in the 1982 AQCD. Three other recent studies were conducted evaluating the effects of
10 ppm SO2 on immune responses.
In summary, animal toxicological studies do not provide much evidence for long-term
exposure to ambient SO2 having detrimental effects on lung host defense.
3.4.2.5. S02 Interactions with PM and Other Mixtures
An elegant series of experiments was conducted in dogs exposed to 0.31 mg/m3 neutral
sulfite aerosol for 22.5 h/day for 290 days (Heyder et al., 1992). The aerosol particles were
submicron in size. These studies are summarized in Annex Table E-14. Although sulfite particles
are not usually found in nature, they were engineered and used in these studies for the purpose of
delivering S02-like reactivity to the lower respiratory tract. It should be noted that the reactivity
of S02 is due to the IV-valent sulfur, a feature shared by sulfite but not sulfate which has VI-
valent sulfur. The concentration of sulfite particles used in these studies was comparable to
ambient levels of SO2 on smog-alert days in Germany (i.e., 0.25 ppm). Important findings from
these studies included a significant decrease in specific lung compliance and increase in alveolar-
capillary permeability in sulfite-exposed dogs compared with controls (Maier et al., 1992; Schulz
et al., 1992). In addition, macrophage respiratory burst activity and phagocytic capacity were
significantly decreased while intrapulmonary particle transport to the larynx was increased
(Maier et al., 1992; Kreyling et al., 1992). Morphological effects included hyperplastic changes
in the respiratory mucosa of the nasal cavity and a moderate mononuclear cell infiltration. Loss
of cilia in larynx and trachea was also noted. Some of the dogs also exhibited changes in the
larynx and trachea (Takenaka et al., 1992). The authors concluded that chronic exposure to a low
dose of sulfur (IV) aerosols can initiate a pathophysiological response.
May 2008
3-92
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
A second set of studies was conducted by these same investigators in dogs exposed to
sulfite and sulfate aerosols for 13 months (Heyder et al., 1999). These are summarized in Annex
Table E-14. This protocol involved daily exposures of 16.5 h neutral sulfite aerosol at the same
concentration used in the previous study followed by 6 hrs of an acidic sulfate aerosol at a
concentration of 15.2 |imol/m3 hydrogen ions. Both aerosols were about 1 |im MMAD in size.
The authors stated that the dose received by each dog in 13 months was equivalent to what a
person living for 70 years in an urban environment would receive. Results of these experiments
demonstrated no change in lung compliance or other measure of lung function in dogs exposed
consecutively to sulfite and sulfate each day (Schulz et al., 1999). Alveolar-capillary
permeability was no different than in controls (Maier et al., 1999). Intrapulmonary particle
transport to the larynx was decreased while transport to the tracheobronchial lymph nodes was
increased in dogs exposed to both sulfite and sulfate (Kreyling et al., 1999). No alteration in the
surfactant system was observed (Griese et al., 1999). Slight morphological effects were observed
in the proximal alveolar region but not in the nasal cavity, larynx or trachea (Takenaka et al.,
1999). The authors attributed this milder response to a modulating effect of the acidic sulfate
aerosol. They concluded that inhalation of low levels of sulfite and hydrogen ion is not likely to
constitute a health risk. These results are somewhat surprising given the pathophysiologic
response to sulfite alone found by these same authors in a similar model. Possibly they indicate
an antagonistic interaction between sulfate and sulfite.
In addition to studies examining the interaction of SO2 and particles, other animal studies
performed since the 1982 AQCD involved binary mixtures, laboratory-generated complex
mixtures (e.g., simulation of regional air pollution), or actual ambient air mixtures (Annex Tables
E-18 through E-20). Generally, most studies with ambient or laboratory-generated complex
mixtures did not include an S02-only exposure group, making it difficult to determine the
contribution of SOx. No definitive conclusions can be made from these studies.
3.4.2.6. Summary of Evidence on the Effect of Long-Term Exposure on
Respiratory Health
The overall epidemiological evidence on the respiratory effects of long-term exposure to
SO2 is inadequate to infer the presence or absence of a causal relationship. Studies that
examined the effects of long-term exposure to S02 on asthma, bronchitis, and respiratory
May 2008
3-93
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
symptoms observed positive associations in children. However, the variety of outcomes
examined and the inconsistencies in the observed results make it difficult to assess the impact of
long-term exposure of S02 on respiratory symptoms. In the limited number of studies examining
the SO2 associations with lung function, results were generally mixed. A major consideration in
evaluating SCVrelated health effects in these epidemiological studies of long-term exposure is
the high correlation among the pollutant levels observed, particularly between long-term average
S02 and PM concentrations. The lack of evidence available to evaluate potential confounding by
copollutants limits the ability to make a causal determination based on these studies.
A limited number of animal toxicological have examined the effect of long-term exposure
to SO2 on lung function. Results from these studies do not provide strong biological plausibility
for effects of long-term exposure to S02 on respiratory morbidity. These studies observed no
effects on physiological lung function at S02 concentrations < 5 ppm in rabbits and dogs;
however, one study found decreased residual volume and quasistatic compliance at 1 ppm S02 in
rats. In addition, no morphological changes were found in guinea pigs exposed subacutely to
1 ppm SO2, or in rats exposed subchronically to 5 ppm SO2. While mild, bronchiolar epithelial
hyperplasia was observed in rats exposed to 1 ppm for 4 months, this change was not apparent at
8 months. Furthermore, animal toxicological studies provide no evidence for decrements in lung
host defense at or near ambient levels of SO2.
Overall, results from the generally limited number of epidemiological and animal
toxicological studies do not give support to respiratory effects from long-term exposure to SO2 at
ambient concentrations. However, chronic studies in dogs exposed to sulfite particles at
concentrations equivalent to near ambient levels of SO2 demonstrated a mild pathophysiologic
response, suggesting that deposition of SOx in the lower respiratory tract may lead to more
profound effects on the respiratory system than those observed with gaseous SO2 alone. These
changes were modulated and in some cases reversed by sequential exposure to sulfate particles,
suggesting an antagonistic interaction among the different PM in the mixture.
3.4.3. Carcinogenic Effects Associated with Long-Term Exposure
The 1982 AQCD concluded that little or no clear epidemiological evidence substantiated
the hypothesized links between SO2 or other SOx and cancer. From the toxicological studies, it
was noted that while there were some indications of carcinogenicity for both S02 and S02 +
May 2008
3-94
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
benzo[c/]pyrene (B[a]P), complex exposure regimens, problematic dose determinations, and/or
inadequately reported experimental details led to the conclusion that SO2 could only be
considered a suspect carcinogen/cocarcinogen.
Since the last review, numerous studies have examined the genotoxic effects of SO2. These
are summarized in Annex Table E-22. SO2 and its metabolite sulfite were found not to be
mutagenic or to induce DNA damage in vitro (Pool et al., 1988; Pool-Zobel et al., 1990).
However, inhalation studies demonstrated increased mouse bone marrow micronucleated
polychromatic erythrocytes and DNA damage in cells isolated from various organs when mice
were exposed for 4-6 h/day for 7 days to 5-30 ppm SO2 (Meng et al., 2002; 2005; Ruan et al.,
2003). These in vivo studies suggest that inhaled SO2 may have systemic effects at high
concentrations, but they are of questionable significance in evaluating the effects of S02 at
ambient levels.
The carcinogenic potential of SO2 was examined in animal toxicological studies which are
summarized in Annex Table E-l 1. Gunnison et al. (1988) conducted a two-part study in which
rats were exposed either for 21 weeks (6 h/day, 5 days/week) by inhalation to 0, 10, or 30 ppm
SO2, or for 21 weeks to two tungsten-supplemented, molybdenum-deficient diets. This latter
regimen induced a condition of sulfite oxidase deficiency, resulting in elevated systemic levels of
sulfite:bisulfite relative to control values (e.g., in plasma, from 0 to 44 |iM; and in tracheal tissue,
from 33 to 69 or 550 nmol/g wet weight). Beginning with week 4, some groups from each
regimen received weekly tracheal installations of 1-mg B[a]P for 15 weeks. Overall results
indicated that squamous cell carcinoma was not induced, or in the B[a]P groups coinduced or
promoted, by SO2 inhalation or elevated systemic sulfite:bisulfite. Researchers found a very high
incidence of animals with tumors in the groups exposed to only B[a]P (128/144). As a result,
carcinogenicity or cocarcinogenicity of S02 or sulfite:bisulfite could only have been detected as
a shortening of tumor induction time or an increase in rate of tumor appearance, and neither was
observed. As noted by the authors, these findings do not support the conclusion that SO2
exposure enhances the carcinogenicity of B[a]P . It was proposed that SO2 exposure, by
elevating systemic sulfite:bisulfite, would generate glutathione-»Y-sulfonates, which in turn could
inhibit glutathione ^'-transferase (GST) and reduce intracellular GSH and, thus, interfere with a
major detoxication pathway for B[a]P. See Annex Table E-21 for further discussion from the
work of Menzel et al., (1986).
May 2008
3-95
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Two similar studies were published that investigated the ability of 10 to 11 months of
exposure (16 h/day) to 4 ppm SO2, 6 ppm NO2, or their combination to affect the carcinogenicity
of either urban suspended PM (SPM) (Ito et al., 1997) or diesel exhaust particle (DEP) (Ohyama
et al., 1999) extract-coated carbon particles. The former study found that, while exposure to SPM
extract-coated carbon particles significantly increased pulmonary endocrine cell (PEC)
hyperplasia, coexposure to SO2, NO2, or their combination was without additional affect. Also,
irrespective of gas coexposure, SPM extract-coated carbon particles demonstrated a few PEC
papillomas versus control frequencies of zero.
Using Syrian golden hamsters, Heinrich et al. (1989) investigated whether coexposure to
10 ppm SO2 and 5 ppm NO2 for 6 to 8 months (5 days/week, 19 hours/day) could enhance
tumorigenesis induced by a single subcutaneous injection of diethylnitrosamine (DEN) during
week 2. The combined gas exposure did not affect body weight gain and only minimally
shortened survival times. Compared to the DEN groups, serial sacrifices of gas-exposed animals
demonstrated progressively increasing numbers of tracheal mucosal cells and aberrant tracheal
cell cilia. In the lung, effects related to gas mixtures were largely limited to a progressive type of
alveolar lesion that involved the lining of bronchiolar epithelium and the appearance of pigment-
containing AM and to a mild, diffuse thickening of the alveolar septa. Exposure to the combined
gases by itself did not induce tumors of the upper respiratory tract, nor did it enhance the
induction of such tumors by DEN.
In addition to the animal toxicological studies that examined the genotoxic and
carcinogenic potential of SO2, a limited number of recent epidemiologic studies have
investigated the relationship between long-term exposure to SO2 and lung cancer incidence and
mortality. These studies are summarized in Annex Table F-7. Nyberg et al. (2000) conducted a
case-control study of men aged 40 to 75 years with (n = 1,042) and without (n = 2,364) lung
cancer in Stockholm County, Sweden. They mapped residence addresses to a GIS database to
assign individual exposures to SO2 from defined emission sources (mainly local oil-fueled
residential heating). Available SO2 measurement data were used to calibrate the model. In this
study, S02 was considered an indicator of air pollution from residential heating. Exposure to
NO2, considered to be a marker of traffic pollution, also was evaluated in this study. The 90th
percentile 30-year average SO2 level was 30 ppb. After adjusting for potential confounders (e.g.,
smoking, occupational exposures), long-term average heating-related SO2 exposure was not
May 2008
3-96
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
associated with an increase in risk of lung cancer incidence. A weak association for the 30-year
average traffic-related NO2 exposure was observed.
Very similar results were reported in a Norwegian study by Nafstad et al. (2003). The study
population is a cohort of 16,209 men who enrolled in a study of cardiovascular diseases in 1972.
The Norwegian cancer registry identified 422 incident cases of lung cancer. SO2 exposure data
were modeled based on residence using data for observed concentrations and emission from
point sources (e.g., industry and heating of buildings and private homes) and traffic. Once again,
no association was observed between long-term exposure to SO2 and lung cancer incidence.
Three additional European cohort studies examined the associations between long-term
exposure to air pollution and lung cancer mortality (Beelen et al., 2008; Filleul et al., 2005;
Nafstad et al., 2004) in cohorts ranging in size from 14,284 to 120,852 subjects, who were
followed for 9 to > 20 years. Consistent with the results for lung cancer incidence, none of these
studies observed an association between long-term SO2 exposure and lung cancer mortality.
These studies are discussed in further detail in Section 3.5.2.2.
Similar to the European cohort studies, studies conducted in the United States generally did
not observe an association between long-term exposure to SO2 and lung cancer mortality. In the
reanalysis of the Harvard Six Cities study, Krewski et al. (2000) estimated a RR of 1.03 (95% CI:
0.91, 1.16) per 5 ppb increase in average SO2 over the study period, while Pope et al. observed a
positive but not statistically significant (RR -1.04 per 5 ppb increase in average SO2 from 1982
to 1998) association in the extended analysis of the American Cancer Society (ACS) cohort. The
California Seventh-day Adventists study by Abbey et al. (1999) did observe a statistically
significant association between lung cancer mortality and SO2 (and most of the pollutants
examined including PM10, sulfate, O3, and NO2), but the number of lung cancer deaths in this
cohort was very small (12 for female, 18 for male) and, therefore, it is difficult to interpret these
estimates. More detailed discussions of these studies are provided in Section 3.5.2.2.
In conclusion, the toxicological studies indicate that SO2 at high concentrations may cause
DNA damage but fails to induce carcinogenesis, cocarcinogenesis, or tumor promotion.
Furthermore, the epidemiological studies did not provide evidence that long-term exposure to
SO2 is associated with an excess risk of lung cancer.
May 2008
3-97
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
3.4.4. Cardiovascular Effects Associated with Long-Term Exposure
The effects of SO2 on the cardiovascular system were not addressed in the 1982 AQCD.
Since then, animal toxicological studies have reported oxidation (Meng et al., 2003) and
glutathione (GSH) depletion (Langley-Evans et al., 1996; Meng et al., 2003; Wu and Meng,
2003) in the hearts of rodents which were exposed by inhalation to SO2. However, as
concentrations of S02used in these studies were 5 ppm and above, the oxidative injury observed
is probably not relevant to cardiovascular effects seen at ambient levels of SO2. These and other
animal toxicology studies measuring cardiovascular endpoints are summarized in Annex Table
E-5.
A recent epidemiological study examined the association between long-term exposure to
air pollution, including SO2, and one or more fatal or nonfatal cardiovascular events. In the
Women's Health Initiative cohort study, Miller et al. (2007) studied 65,893 postmenopausal
women between the ages of 50 and 79 years without previous cardiovascular disease in 36 U.S.
metropolitan areas from 1994 to 1998. Subjects' exposures to air pollution were estimated using
residents' five-digit ZIP code, assigning the annual mean levels of air pollutants measured at the
nearest monitor. A total of 1,816 women had one or more fatal or nonfatal cardiovascular events,
including 261 deaths from cardiovascular causes. Hazard ratios for the first cardiovascular event
were estimated. The results for models that only included subjects with non-missing exposure
data for all pollutants (n = 28,402 subjects, resulting in 879 cardiovascular events) are described
here. In the single-pollutant models, PM2.5 showed the strongest associations with cardiovascular
events among the pollutants (Hazard Ratios = 1.24 [95% CI: 1.04, 1.48] per 10 |ig/m3 increase in
annual average), followed by SO2 (1.07 [95% CI: 0.95, 1.20] per 5 ppb increase in the annual
average). In the multipollutant model where all the pollutants (i.e., PM2.5, PM10-2.5, CO, SO2,
N02, 03) were included in the model, the PM2.5 association with overall cardiovascular events
was even stronger (1.53 [95% CI: 1.21, 1.94]). The association with SO2 also became stronger
(1.13 [95% CI: 0.98, 1.30]). Correlations among these pollutants were not described and,
therefore, the extent of confounding among these pollutants in these associations could not be
examined, but among all the air pollutants considered, PM2.5 was clearly the best predictor of
cardiovascular events.
The available toxicological and epidemiological evidence to assess the effect of long-term
exposure to SO2 on cardiovascular health is too limited to make any conclusions at this time.
May 2008
3-98
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
3.4.5. Prenatal and Neonatal Outcomes Associated with Long-Term
Exposure
Several animal toxicological studies examined developmental effects of SO2 and are
summarized in Anne Table E-7. No changes in birth weight or neurobehavioral development
were noted in mouse pups prenatally exposed to 5-30 ppm SO2 (1996), while some behavioral
modifications were seen in adults exposed prenatally to these same levels (Fiore et al.). However,
effects observed at such high concentrations of SO2 are of questionable relevance.
In recent years, the effects of prenatal and neonatal exposure to air pollution have been
examined in epidemiologic studies by several investigators. These studies are summarized in
Annex Table F-8. The most common endpoints studied are low birth weight, preterm delivery,
and measures of intrauterine growth. Preterm birth and low birth weight may result in serious
long-term health outcomes for the infant. Preterm birth is the leading cause of infant mortality
and is a major determinant of a variety of adverse neurodevelopmental outcomes and chronic
adverse respiratory effects (Berkowitz and Papiernik, 1993). Low birth weight has also been
linked with increased risk of infant mortality and morbidity. Other studies have examined
associations between maternal exposure to ambient air pollution and sudden infant death
syndrome (SIDS) and neonatal hospitalizations.
These studies analyzed air pollution data and birth certificates from a given area. In
evaluating the results of these studies, it is important to consider the limitations of these data. For
example, the reliability and validity of birth certificate data have been reviewed (Buescher et al.,
1993; Piper et al., 1993) and have been found to vary in degrees of reliability by specific
variables. The variables considered the most reliable include birth weight, maternal age, race,
and insurance status. Whereas gestational age, parity and delivery type (vaginal vs. cesarean)
were reasonably reliable, obstetrical complications and maternal lifestyle factors such as
smoking and alcohol consumption were not reliable. Another concern in these studies regards
adequate control for potential confounders. While most of these studies adequately controlled for
maternal education, parity, age, and sex of child, many did not adjust for socioeconomic status,
occupational exposures, indoor pollution levels, maternal smoking, alcohol use, prenatal care, or
concurrent temperature exposures as fetal growth is associated with all of these factors. This
makes overall comparisons across studies a difficult task.
May 2008
3-99
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
While most studies analyzed average SO2 exposure for the whole pregnancy, many also
considered exposure during specific trimesters, or other time periods (e.g., first and last months
of gestation). Different exposure periods have been examined because the biological mechanisms
and timing of critical exposures that link air pollution to adverse birth outcomes are yet to be
determined. For example, fetal growth is much more variable during the third trimester;
therefore, exposure during the third trimester would have the greatest likelihood of an
association. However, insufficient placentation during the first trimester may be associated with
early environmental insult, whereby subsequent fetal growth is hindered. Similarly, it is possible
that preterm delivery is associated with insufficient placentation resulting from early exposure.
Furthermore, preterm delivery may be the result of acute exposures just prior to delivery.
Epidemiological studies examining the effects of air pollutants on low birth weight are
summarized in Figure 3-13. Maisonet et al. (2001) examined the association between air
pollution and low birth weight in six northeastern cities: Boston, MA; Hartford, CT;
Philadelphia, PA; Pittsburgh, PA; Springfield, MA; and Washington, DC. The study population
consisted of 89,557 singleton, full-term, live births (37-44 weeks of gestation) born between
January 1994 and December 1996. Low birth weight was classified as < 2,500 g (5.5 lbs.). This
study observed an association between low birth weight and SO2 concentrations during each
trimester among Caucasians; however, the association was not consistent in other races and
ethnicities.
An excess risk for low birth weight associated with ambient S02 concentrations was
reported by Dugandzic et al. (2006) in a large cohort study of 74,284 women with full-term,
singleton births from 1988-2000 in Nova Scotia, Canada. The mean 24-h avg SO2 concentration
over the study period was 10 ppb (IQR 7). These investigators found that exposure only during
the first trimester was associated with increased risk of low birth weight. The RR was 1.14 (95%
CI: 1.04, 1.26) per 5 ppb increase in SO2 level.
May 2008
3-100
DRAFT—DO NOT QUOTE OR CITE
-------
Reference
Maisonetetai, (2001)
Liiiet al.(2O03)
Bobak(2000)
Maisoiitetal. {2»1}
Dugandocetai. (2006)
Bobai( el ai (2S00)
Maisonetetai, (2001)
Liu et al, (2003)
Bobak(20Q0)
Viang et al. (199?)
Location
6 Northeastern cities, U.S.
Ail
White
African American
Hispanic
Vancouver, Canada
Dugartdzie et al. (2006) Nova Scotia, Canada
Czech Republic
8 Northeastern cities, U.S.
White
African American
Hispanic
Nora Scotia, Canada
C«h Republic
8 Northeastern cities, U.S. All
White
African American
Hispanic
Vancouver, Canada
Dugamjzie et al, (2008) Nova Scotia, Canada
Czech Republic
Beijing, China
_J.
1st trimester
+
&
2nd trimester
3rd trimester
0.8 0.9 1 1.1 1,2 1.3
Relative risk
Figure 3-13. Relative risks (95% CI) for low birth weight, grouped by trimester of S02 exposure.
Risk estimates are standardized per 5 ppb increase in S02 concentrations. The size of the box of
the central estimate represents the relative weight of that estimate based on the width of the 95%
CI.
1 Liu et al. (2003) found similar results in a study of pregnancy outcomes and air pollution in
2 Vancouver, Canada. The mean 24-h avg SO2 concentration was 4.9 ppb (IQR 7.7) from 1985 to
3 1998. Maternal exposure during the first month was associated with an increased risk of low
May 2008
3-101
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
birth weight (OR 1.11 [95% CI: 1.01, 1.22]). Additional studies from the United States, Europe,
Latin America and Asia have reported positive associations between low birth weight and
maternal exposure to S02 during the first (Bell et al., 2007; Bobak, 2000; Ha et al., 2001;
Mohorovic, 2004; Yang et al., 2003a), second (Bobak, 2000; Gouveia et al., 2003; Lee et al.,
2003a) and third (Bobak, 2000; Lin et al., 2004; Wang et al., 1997) trimesters.
Preterm delivery, intrauterine growth retardation (IUGR), and birth defects are additional
adverse birth outcomes that have been associated with ambient S02 levels. In a time-series
analysis using data from four Pennsylvania counties, (Sagiv et al., 2005) reported that the mean
6-week SO2 exposure prior to birth was associated with increased risk of preterm birth, with a
RR of 1.05 (95% CI: 1.00, 1.10) per 5 ppb increase in SO2. A 5 ppb increase in SO2
concentrations three days before birth was associated with a RR of 1.02 (95% CI: 0.99, 1.05).
The authors discussed two plausible mechanisms for the effects of air pollution on preterm birth:
(1) changes in blood viscosity due to inflammation as a result of air pollution (citing Peters et al.,
1997); and (2) maternal infection during pregnancy as a consequence of impaired immunity from
air pollution exposure. Liu et al. (2003) reported that S02 exposure during the last month of
pregnancy was associated with preterm birth, with an OR of 1.09 (95% CI: 1.01, 1.19) for a
5 ppb increase in SO2, in Vancouver, Canada. Similar results were found for studies conducted in
the Czech Republic (Bobak, 2000), Korea (Leem et al., 2006), and Beijing (Xu et al., 1995).
Liu et al. (2003) further reported that SO2 exposure during the last month of pregnancy
was associated with IUGR (OR 1.07 [95% CI: 1.01, 1.13]). However, in a later study in the
Canadian cities of Calgary, Edmonton and Montreal, (Liu et al., 2007) did not observe
associations between maternal exposure to SO2 and increased risk of IUGR.
Two Brazilian studies examined exposure to SO2 and neonatal deaths. Pereira et al. (1998)
found a positive association between S02 and intrauterine mortality in Sao Paulo during a 2-year
period, though the effect was sensitive to model specifications and did not support a
concentration-response relationship. The most robust association was observed for an index of
three gaseous pollutants (NO2, SO2, CO) with mortality. Lin et al. (2004) found that a 5 ppb
increase in S02 was associated with an increase of 8.8% (95% CI: 5.8, 11.8). A similar
relationship was found for PMi0. The creation of an index containing both PM10 and SO2 allowed
the observation of their cumulative effects on daily death counts. The result of this analysis was
May 2008
3-102
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
similar in magnitude to the effect of SO2 alone. An ecologic cohort study of infant mortality in
the U.S. found no association with annual averages of SO2 concentration (Lipfert et al., 2000a).
Gilboa et al. (2005) conducted a population-based case-control study to investigate the
association between maternal exposure and air pollutant exposure during weeks 3-8 of
pregnancy, the risk of selected cardiac birth defects and oral clefts in live births, and fetal deaths
between 1997 and 2000 in seven Texas counties. When the highest quartile of exposure was
compared to the lowest, the authors observed a positive association between S02 and isolated
ventricular septal defects (OR 2.16 [95% CI: 1.51, 3.09]). Although this is the only study to have
examined the effect of maternal exposure to SO2 on birth defects, it supports the notion that the
developing embryo and growing fetus is susceptible to maternal air pollution exposure.
Several studies examined adverse health outcomes in relation to S02 concentrations during
the neonatal period. Dales et al. (2006) evaluated hospitalizations for respiratory disorders in
neonates < 4 weeks of age from hospitals in 11 large Canadian cities during a 15-year study
period (population-weighted average 24-h avg SO2 of 4.3 ppb). The researchers observed a 5.5%
(95%) CI: 2.8, 8.3) excess risk in respiratory hospitalizations associated with a 10 ppb increase in
24-h avg SO2 concentrations with a 2-d lag. This effect was slightly attenuated after adjusting for
PM10 and gaseous copollutants. To investigate the influence of ambient SO2 concentrations on
SIDS, Dales et al. (2004) conducted a time-series analysis comparing daily rates of SIDS and
daily SO2 concentrations from 12 large, Canadian cities during a 16-year period. The mean
24-h avg S02 level across the 12 cities was 5.51 ppb (IQR 4.92). There was an 18.0%> (95%> CI:
4.4, 33.4) excess risk in SIDS incidence for a 10 ppb increase in 24-h avg SO2 levels. The
authors concluded that the effect of SO2 was independent of sociodemographic factors, temporal
trends, and weather.
In summary, epidemiological studies on birth outcomes have found suggestive positive
associations between SO2 exposure and low birth weight; however, toxicological studies provide
very little biological plausibility for reproductive outcomes related to SO2 exposure. The
inconsistent results across trimesters of pregnancy and the lack of evidence regarding
confounding by copollutants further limit the interpretation of these studies. The limited number
of studies addressing preterm delivery, IUGR, birth defects, neonatal hospitalizations, and infant
mortality make it difficult to draw conclusions regarding the effect of SO2 on these outcomes.
May 2008
3-103
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
3.4.6. Other Organ System Effects Associated with Long-Term
Exposure
The 1982 AQCD presented only one chronic exposure study which was relevant to nervous
system effects. Dogs were exposed for 68 months to a mixture of S02 and H2S04. No effects on
visual evoked brain potentials during or immediately after exposure to the SOx mixture were
observed. Since then, numerous studies have examined brain lipid content, lipid peroxidation and
glutathione content and antioxidant enzymes following inhalation exposure of rodents to SO2 at
concentrations of 10 ppm or higher. Concentrations of 5 ppm or higher SO2 were used in studies
examining neurobehavior and neurodevelopment in mice. These studies are summarized in
Annex Table E-6.
In the past 25 years, numerous animal toxicological studies have evaluated the effects of
long-term SO2 exposure on other organ systems such as reproductive, hematological,
gastrointestinal, renal, lymphatic, and endocrine systems. Most of these studies used
concentrations of SO2 of 5 ppm or higher. Many of these studies examined alteration profiles of
lipid peroxidation and antioxidant levels (Langley-Evans et al., 1996; Meng and Bai, 2004;
Meng et al., 2003b) and are summarized in Annex Table E-7 through E-9 and E-22 through E-24.
3.5. Mortality Associated with Long-Term S02
Exposure
3.5.1. Summary of Findings from the Previous Review
At the time of the 1982 AQCD, the available studies on the effects of long-term exposure
to SO2 on mortality were all ecological cross-sectional studies. This study design could not take
into consideration such confounders as cigarette smoking, occupational exposures, and social
status. In addition, there were questions regarding how representative the aerometric data used
were for community exposure. Therefore, it was concluded that the epidemiological studies did
not provide valid quantitative data relating respiratory disease or other types of mortality to long-
term (annual average) exposures to S02 or PM.
The 1986 Secondary Addendum reviewed more studies of this type, with information on
more detailed components of PM (inhalable and fine particles, and particulate sulfate). While
some studies suggested importance of the size of PM, the fundamental problem of the study
May 2008
3-104
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
design made it difficult to interpret the effect estimates. The 1986 Secondary Addendum also
reviewed a Japanese study in which the death rates from asthma and chronic bronchitis in a
highly polluted section of Yokkaichi, an industrial city with large S02 emissions from the largest
oil-fired power plant in Japan, were compared with those in a less polluted area of the same city.
SOx levels (measured using the lead peroxide method) averaged across several monitoring sites
in the polluted harbor area ranged from around 1.0 to 2.0 mg/day (annual average) during 1964
through 1972 and then steadily declined to less than 0.5 mg/day in 1982. This is in contrast to
levels consistently < 0.3 mg/day in the low pollution areas throughout 1967 through 1982.
Annual average levels for other pollutants (i.e., NO2, TSP, oxidants) monitored in the high
pollution area were consistently low from 1974 through 1982. The results indicated elevated
rates of chronic bronchitis mortality in the highly polluted area compared to the less polluted
area, but the 1986 Secondary Addendum could not conclude that this was due to SO2 alone,
because sulfate or other particulate SOx such as H2SO4 could have been responsible.
Several, more recent studies have examined long-term exposure effects of air pollution,
including S02, on mortality. These studies are summarized in Annex Table F-9. As with short-
term exposure studies, the focus of most of these studies was mainly on PM though some
focused on traffic-related air pollution. They all used Cox proportional hazards regression
models with adjustment for potential confounders. The designs of these studies were better than
earlier cross-sectional studies as the outcome and most of the potential confounders (e.g.,
smoking history, occupational exposure) were measured on an individual basis. However, the
geographic scale and method for exposure estimates varied across these studies.
3.5.2. Associations of Mortality and Long-Term Exposure in Key
Studies
3.5.2.1. U.S. Cohort Studies
3.5.2.1.1. Harvard Six Cities Studies
Dockery et al. (1993) conducted a prospective cohort study to study the effects of air
pollution with the main focus on PM components in six U.S. cities. These cities were chosen
based on levels of air pollution, with Portage, WI and Topeka, KS representing the least polluted
cities and Steubenville, OH representing the most polluted city. Mean SO2 levels ranged from
May 2008
3-105
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
1.6 ppb in Topeka to 24.0 ppb in Steubenville from 1977 to 1985. Cox proportional hazards
regression was conducted with data from a 14- to 16-year follow-up study of 8,111 adults in the
six cities. Dockery et al. reported that lung cancer and cardiopulmonary mortality were more
strongly associated with the concentrations of inhalable and fine PM, and sulfate particles, than
with the levels of TSP, SO2, NO2, or acidity of the aerosol.
Krewski et al. (2000) conducted a sensitivity analysis of the Harvard Six Cities study and
examined associations between gaseous pollutants (i.e., 03, N02, S02, and CO) and mortality.
SO2 showed positive associations with total mortality (RR = 1.05 [95% CI: 1.02, 1.09] per 5 ppb
increase in average SO2 over the study period) and cardiopulmonary deaths (1.05 [95% CI: 1.00,
1.10]), but in this dataset SO2 was highly correlated with PM2.5 (r = 0.85), sulfate (r = 0.85), and
N02 (r = 0.84).
3.5.2.1.2. American Cancer Society Cohort Studies
Pope et al. (1995) investigated associations between long-term exposure to PM and the
mortality outcomes in the ACS cohort. Ambient air pollution data from 151 U.S. metropolitan
areas in 1981 were linked with individual risk factors in 552,138 adults who resided in these
areas when enrolled in the prospective study in 1982. Death outcomes were ascertained through
1989. PM2.5 and sulfate were associated with total, cardiopulmonary, and lung cancer mortality,
but not with mortality for all other causes. Gaseous pollutants were not analyzed in this study.
Krewski and co-investigators (Jerrett et al., 2003; Krewski et al., 2000) conducted an
extensive sensitivity analysis of the Pope et al. (1995) ACS data, augmented with additional
gaseous pollutants data. The mean SO2 concentrations were 7.18 ppb in the warm season (April
to September) and 11.24 ppb in the cool season (October to March). Among the gaseous
pollutants examined, only S02 showed positive associations with mortality. The RR for total
mortality was 1.06 (95% CI: 1.05, 1.07) per 5 ppb increase in the annual average SO2. Analysis
using SO2 measured in different seasons produced a somewhat higher estimate for the warm
season than that for the cool season (7% compared to 5% excess risk per 5 ppb increase).
Although the subjects in the ACS cohort came from all regions of the United States, the majority
of the 151 cities were located in the eastern United States, where both SO2 and sulfate tend to be
higher. PM2.5 levels are also higher in the east. To address the influence of these spatial patterns,
which may confound associations between mortality and these pollutants, Krewski et al. (2000)
May 2008
3-106
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
conducted extensive two-stage regression modeling. In these models, the association between
SO2 and mortality was diminished but persisted after adjusting for sulfate, PM2.5, and other
variables. For example, in the spatial filtering model (which resulted in the largest reduction of
the SO2 effect estimate when sulfate was included), the SO2 total mortality RR estimate was 1.07
(95% CI: 1.03, 1.11) in the single-pollutant model and 1.04 (95% CI: 1.02, 1.06) with sulfate in
the two-pollutant model. The effect estimates for PM2.5 and sulfate also were diminished when
S02 was included in the models. The result further showed that S02 effect estimates were
generally insensitive to adjustment for spatial correlation. Thus, these results suggest that the
association between SO2 and mortality may be confounded with PM, but the association cannot
be accounted for by PM2.5 or sulfate alone. Krewski et al. (2000) noted that their reanalysis of the
ACS and Harvard Six Cities studies suggested that mortality might be attributed to more than
one component of the complex mixture of ambient air pollutants in urban areas in the United
States.
The original Pope et al. (1995) study and the Krewski et al. (2000) reanalysis both used the
air pollution exposure estimates that are based on the average over the Metropolitan Statistical
Area (MSA), which consists of multiple counties. To investigate the effects of geographic scale
over which the air pollution exposures are averaged, Willis et al. (2003) reanalyzed the ACS
cohort data using the exposure estimates averaged over the county scale, and compared the
results with those based on the MSA-scale average exposure. Less than half of the cohort used in
the MSA-based study was used in the county-scale based analysis, because of the limited
availability of sulfate monitors and the reduced sample size due to the loss of subjects when
using the five-digit ZIP codes. The mean (9.3 ppb versus 10.7 ppb) and range (0.0 to 29.3 ppb
versus 0.0 to 27.2 ppb) of the MSA- and county-level SO2 data sets were similar. In the analysis
comparing the two-pollutant model with sulfate and S02, they found that the inclusion of S02
reduced sulfate effect estimates substantially (> 25%) in the MSA-scale model but not
substantially (< 25%) in the county-scale model. In the MSA-level analysis (with 113 MSAs),
the SO2 RR estimate was 1.04 (95% CI: 1.02, 1.06) per 5 ppb increase, with sulfate in the model.
In the county-level analysis (91 counties) with sulfate in the model, the corresponding estimate
was smaller (1.02 [95% CI: 1.00, 1.05]). It should also be noted that the correlation between
covariates were different between the MSA-level data and county-level data. The correlation
between SO2 and sulfate was 0.48 in the MSA-level data, but it was 0.56 in the county-level
May 2008
3-107
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
data. The correlation between poverty rate and SO2 was -0.16 in the MSA-level data, but it was
0.15 in the county-level data. Thus, the extent of confounding between SO2 and PM components
as well as among other covariates in the model can be affected by the geographic scale of
aggregation of exposure estimates. It is not clear, however, if the smaller geographic scale
increases or decreases exposure characterization error for SO2, because a certain extent of
smoothing (averaging) over distance may reduce very local concentration peaks that are not
relevant to the city-wide population.
Pope et al. (2002) extended analysis of the ACS cohort with double the follow-up time (to
1998) and triple the number of deaths compared to the original Pope et al. (1995) study. In
addition to PM2.5, all the gaseous pollutant data were retrieved for the extended period and
analyzed for their associations with death outcomes. As in the 1995 analysis, the air pollution
exposure estimates were based on the MSA-level averages. PM2.5 was associated with total,
cardiopulmonary, and lung cancer mortality but not with deaths for all other causes. SO2 was
associated with all the mortality outcomes, including all other causes of deaths. The SO2 RR
estimate for total mortality was 1.03 (95% CI: 1.02, 1.05) per 5 ppb increase (1982 to 1998
average). The association of SO2 with mortality for all other causes (sulfate also showed this
pattern) makes it difficult to interpret the effect estimates. This lack of specificity for SO2 (in
contrast to PM) is not consistent with causal inference.
3.5.2.1.3. The EPRI-Washington University Veterans' Cohort Mortality Studies
Lipfert et al. (2000b) conducted an analysis of a national cohort of-70,000 male U.S.
military veterans who were diagnosed as hypertensive in the mid 1970s and were followed up for
about 21 years (up to 1996). This cohort was 35% black and 57% were current smokers (81% of
the cohort had been smokers at one time). PM2.5, PMi0, PM10-2.5, TSP, sulfate, CO, 03, N02, S02,
and lead were examined in this analysis. No mean or median level of SO2 was reported. The
county of residence at the time of entry to the study was used to estimate exposures. Four
exposure periods (from 1960 to 1996) were defined, and deaths during each of the three most
recent exposure periods were considered. The results for S02 were presented only qualitatively
as part of their preliminary screening regression results. Lipfert et al. (2000b) noted that lead and
SO2 were not found to be associated with mortality, thus were not considered further. They also
noted that the pollution effect estimates were sensitive to the regression model specification,
May 2008
3-108
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
exposure periods, and the inclusion of ecological and individual variables. The authors reported
that indications of concurrent mortality risks were found for NO2 and peak O3.
Lipfert et al. (2006b) examined associations between traffic density and mortality in the
same cohort, whose follow-up period was extended to 2001. As in their 2000 study, four
exposure periods were considered but included more recent years. The 95th percentiles of daily
average in each of the exposure periods were considered for SO2. For the 1997-2001 data period,
the estimated mortality RR for S02 was 0.99 (95% CI: 0.97, 1.01) per 5 ppb increase in a single-
pollutant model. They reported that traffic density was a better predictor of mortality than
ambient air pollution variables with the possible exception of O3. The log-transformed traffic
density variable was only weakly correlated with SO2 (r = 0.32) and PM2.5 (r = 0.50) in this data
set.
Lipfert et al. (2006a) further extended analysis of the veterans' cohort data to include the
EPA's Speciation Trends Network (STN) data, which collected chemical components of PM2.5.
They analyzed the STN data for year 2002, again using county-level averages. PM2.5 and gaseous
pollutants data for 1999 through 2001 were also analyzed. As in the previous Lipfert et al. (2006)
study, traffic density was the most important predictor of mortality, but associations were also
seen for elemental carbon, vanadium, nickel, and nitrate. O3, NO2, and PM10 also showed
positive but weaker associations. Once again, no associations were observed between long-term
exposure to SO2 and mortality.
3.5.2.1.4. Seventh-day Adventist Study
Abbey et al. (1999) investigated associations between long-term ambient concentrations of
PM10, sulfate, SO2, O3, and NO2 (1973 through 1992) and mortality (1977 through 1992) in a
cohort of 6,338 nonsmoking California Seventh-day Adventists. Monthly indices of ambient air
pollutant concentrations at 348 monitoring stations throughout California were interpolated to
ZIP codes according to home or work location of study participants, cumulated, and then
averaged over time. They reported associations between PM10 and total mortality for males and
nonmalignant respiratory mortality for both sexes. S02 was not associated with total mortality
(RR 1.07 [95% CI: 0.92, 1.24] for males and 1.00 [95% CI: 0.88, 1.14] for females per 5 ppb
increase in multiyear average SO2), cardiopulmonary deaths, or respiratory mortality for either
gender.
May 2008
3-109
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
3.5.2.2. European Cohort Studies
A study by Beelen et al. (2008) examined associations between traffic-related air pollution
and mortality. They analyzed data from the Netherlands Cohort Study on Diet and Cancer with
120,852 subjects who were followed from 1987 to 1996. BS, NO2, SO2, PM2 5, and four types of
traffic-exposure estimates were analyzed. While the local traffic component was estimated for
BS, N02, and PM2.5, no such attempt was made for S02, because there was "virtually no traffic
contributions to this pollutant." Thus, only "background" SO2 levels were reflected in the
exposure estimates. Traffic intensity on the nearest road was associated with all-cause mortality
and a larger RR was observed for respiratory mortality. Results were similar for BS, NO2 and
PM2.5, but no associations were found for S02 (RR = 0.98 [95% CI: 0.93, 1.03] per 5 ppb
increase in multiyear average SO2).
Nafstad et al. (2004) investigated the association between mortality and long-term
exposure to air pollution exposure in a cohort of Norwegian men followed from 1972-1973
through 1998. Data from 16,209 males (aged 0 to 49 years) living in Oslo, Norway, in 1972-
1973 were linked with data from the Norwegian Death Register and with estimates of the
average annual air pollution levels at the participants' home addresses. PM was not considered in
this study because measurement methods changed during the study period. Exposure estimates
for nitrogen oxides (NOx) and S02 were constructed using models based on subject addresses,
emission data for industry, heating, and traffic, and measured concentrations. While NOx was
associated with total, respiratory, lung cancer, and ischemic heart disease deaths, SO2 did not
show any associations with mortality. The authors noted that the SO2 levels were reduced by a
factor of 7 during the study period (from 5.6 ppb in 1974 to 0.8 ppb in 1995), whereas NOx did
not show any clear downward trend.
Filleul et al. (2005) investigated long-term effects of air pollution on mortality in 14,284
adults who resided in 24 areas from seven French cities when enrolled in the Air Pollution and
Chronic Respiratory Diseases (PAARC) survey in 1974. Daily measurements of SO2, TSP, BS,
N02, and NO were made in the 24 areas for 3 years (1974 through 1976). Models were run
before and after exclusion of six area monitors influenced by local traffic as determined by a
NO:N02 ratio of > 3. Before exclusion of the six areas, none of the air pollutants was associated
with mortality outcomes. After exclusion of these areas, analyses showed associations between
total mortality and TSP, BS, N02, and NO but not S02 (RR = 1.01 [95% CI: 0.97, 1.06] per
May 2008
3-110
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
5 ppb multiyear average) or acidimetric measurements. It should be noted that SO2 levels in
these French cities declined markedly between the 1974 through 1976 period and the 1990
through 1997 period by a factor of 2 to 3, depending on the city. The changes in air pollution
levels over the study period complicate interpretation of reported effect estimates.
3.5.2.3. Cross-Sectional Analysis Using Small Geographic Scale
Elliott et al. (2007) examined associations of BS and S02 with mortality in Great Britain
using a cross-sectional analysis. However, unlike the earlier ecological cross-sectional mortality
analyses in the United States in which mortality rates and air pollution levels were compared
using large geographic boundaries (i.e., MSAs or counties), in the Elliot et al. analysis, the
mortality rates and air pollution were compared using a much smaller geographic unit, the
electoral ward, with a mean area of 7.4 km2 and a mean population of 5,301 per electoral ward.
Death rates were computed for four successive 4-year periods from 1982 to 1994 and associated
with 4-year exposure periods from 1966 to 1994. The number of deaths from all causes in the
10,520 wards was 420,776. Of note, S02 levels declined from 41.4 ppb in the 1966 to 1970
period to 12.2 ppb in 1990 to 1994. This type of analysis does not allow adjustments for
individual risk factors, but the study did adjust for socioeconomic status data available for each
ward from the 1991 census. Social deprivation and air pollution were more highly correlated in
the earlier exposure windows. They observed associations for both BS and S02 and mortality
outcomes. The estimated effects were stronger for respiratory illness than other causes of
mortality for the most recent exposure period and most recent mortality period (when pollution
levels were lower). The adjustment for social deprivation reduced the effect estimates for both
pollutants. The adjusted mortality RRs for S02 for the pooled mortality periods using the most
recent exposure windows were 1.021 (95% CI: 1.018, 1.024) for all causes, 1.015 (95% CI:
1.011, 1.019) for cardiovascular, and 1.064 (95% CI: 1.056, 1.072) for respiratory causes per
5 ppb increase in SO2. The effect estimates for the most recent mortality period using the most
recent exposure windows were larger. Simultaneous inclusion of BS and SO2 reduced effect
estimates for BS but not S02. Elliott et al. (2007) noted that the results were consistent with
those reported in the Krewski et al. (2000) reanalysis of the ACS study. This analysis was
ecological, but the exposure estimates in the smaller area compared to that in the U.S. cohort
studies may have resulted in less exposure misclassification error, and the large underlying
May 2008
3-111
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
population appears to be reflected in the narrow confidence bands of effect estimates. The results
from this study suggest an association between long-term exposures (especially in recent years)
to S02 and mortality.
3.5.3. Summary of Evidence on the Effect of Long-Term Exposure on
Mortality
The available epidemiological evidence on the effect of long-term exposure to SO2 on
mortality is inadequate to infer the presence or absence of a causal relationship at this time. The
ecological cross-sectional studies examined in the 1982 AQCD and 1986 Secondary Addendum
found suggestive relationships between long-term exposure to S02 and mortality. However, there
were concerns as to whether the observed association was due to SO2 alone, because sulfate or
other particulate SOx such as H2SO4 could have been responsible. In the more recent longitudinal
cohort studies, once again, positive associations have been observed between long-term exposure
to S02 and mortality; however, several issues affect the interpretation of these results.
Figure 3-14 presents all-cause mortality RR estimates associated with long-term exposure
to SO2 from the U.S. and European cohort studies. The overall range of RRs spans 0.97 to 1.07
per 5 ppb increase in the annual (or longer period) average SO2. The analyses of the Harvard Six
Cities and the ACS cohort data, which likely provide effect estimates that are most useful for
evaluating possible health effects in the United States, observed RRs of 1.02 to 1.07. Note that
each of the U.S. cohort data has its own advantages and limitations. The Harvard Six Cities data
have a small number of exposure estimates, but the location of the monitors were chosen
carefully for epidemiological purposes. The ACS cohort had far more subjects, but the
population was more highly educated than the representative U.S. population. Since educational
status appeared to be an important effect modifier of air pollution effects in both studies, the
overall effect estimate for the ACS cohort may underestimate that for the more general
population. However, it should also be noted that several other U.S. and European studies did not
observe an association between long-term exposure to SO2 and mortality.
May 2008
3-112
DRAFT—DO NOT QUOTE OR CITE
-------
AH Cause Mortality
Relative Risk Estimates
Reference
Krewski et al, (2000)
Krewski et al, (2000);
Jerrett et al, (2003)
Willis et al, (2003)
Pope et al, (2002)
Lipfert et al, (2008a)
Abbey etal, (1999)
Beeten et al, (2008)
Nafstad et al. (2004)
Filled etal. (2005)
Study Other
Harvard Six Cities Study
American Cancer Society Study
Spatial filtering
Spatial filtering
American Cancer Society Study MSA-level
County-level
American Cancer Society Study
(extended follow-up)
Veterans Cohort Study
Seventh-day Adventist Study
The Netherlands Cohort Study
Norwegian Cohort Study
French Cohort Study
Male
Male
Female
Male
• SO, only
O SO, With sulfate
Figure 3-14. Relative risks (95% CI) of S02-associated all-cause (nonaccidental) mortality, with and
without adjustment for sulfate, from longitudinal cohort studies. Effect estimates are
standardized per 5 ppb increase in S02 concentrations. The exposure estimates for Krewski et al.
(2000) and Pope et al. (2002) are based on MSA (Metropolitan Statistical Area)-level averaging;
Lipfert et al. (2006a) used county-level averaging.
1 The geographic scale of analysis appears to influence S02 effect estimates and exposure
2 error. In a reanalysis of the ACS data, the county-level analysis showed a smaller SO2 effect
3 estimate than MSA-level analysis. For sulfate, the opposite pattern was found. Thus, the impact
4 of the geographic scale of analysis may also depend on the spatial distribution of air pollutants.
5 The cross-sectional analysis in Great Britain using small-scale electoral wards observed an effect
May 2008
3-113
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
estimate similar to the lower end of the range of effect estimates for all-cause mortality from
U.S. cohort studies, though it is not clear if the effect estimates from this cross-sectional study
are directly comparable to those from cohort studies.
Another important issue that these studies could not resolve was the possible confounding
and/or interaction among PM indices and SO2. The possibility that the observed effects may not
be due to SO2, but other constituents that come from the same source as SO2, or that PM may be
more toxic in the presence of S02 or other components associated with S02, cannot be ruled out.
For example, the ACS cohort came from all regions of the United States, but a major fraction of
the ACS cities were located in the eastern United States, where both SO2 and sulfate levels tend
to be higher. Therefore, even with sophisticated spatial modeling, separating possible
confounding of S02 effects by PM is challenging. Future and on-going studies that take into
consideration within- versus between-city variation of these pollutants may help elucidate this
issue.
Overall, the results from two major U.S. epidemiological studies observe an association
between long-term exposure to S02 or sulfur-containing particulate air pollution and mortality.
However, several other U.S. and European cohort studies did not observe an association. The
lack of consistency across studies, inability to distinguish potential confounding by copollutants,
and uncertainties regarding the geographic scale of analysis limit the interpretation of a causal
relationship.
May 2008
3-114
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Chapter 4. Public Health Impact
This chapter addresses several issues relating to the broader public health impact from
exposure to ambient SO2. First, the shape of the concentration-response relationship for SO2 is
discussed, with consideration of interindividual variability in responses and evaluation of the
limited evidence available to assess a population threshold value for health effects. The next
section identifies characteristics of subpopulations which may experience increased risks from
SO2 exposures, through either enhanced susceptibility (e.g., as a result of pre-existing disease,
genetic factors, age) and/or differential vulnerability associated with increased exposure (e.g.,
close proximity to sources, activities). The final section discusses the potential public health
impact from adverse health effects associated with SO2 by examining the prevalence of
susceptible individuals in the U.S. population.
4.1. Assessment of Concentration-Response Function
and Potential Thresholds
An important consideration in characterizing the public health impacts associated with SO2
exposure is whether the concentration-response relationship is linear across the full concentration
range, or if there are concentration ranges where there are departures from linearity (i.e.,
nonlinearity). Of particular interest is the shape of the concentration-response curve at and below
the level of the current SO2 NAAQS level of a 24-h avg level of 0.14 ppm or the annual average
of 0.03 ppm.
Some human clinical studies provide individual-level response data in relation to different
levels of S02 exposure; this allows evaluation of both the percentage of individuals showing
responses across the range of exposures as well as the concentration at which an individual
begins to indicate a response. In epidemiological studies, rather than identifying interindividual
differences in response, most studies evaluate whether there is a population-level threshold,
which is the concentration of S02 that must be exceeded to elicit a health response in the study
population. Low data density in the lower concentration range, measurement error in the
response, and exposure measurement error are some of the factors that complicate the ability to
determine the shape of the concentration-response curve, including the presence of any
May 2008
4-1
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
threshold. Biological characteristics that tend to linearize concentration-response relationships
include individual biological differences in susceptibility to air pollution health effects, additivity
of S02-induced effects to a naturally occurring background level, and additivity of effects from
other pollutant exposures. Epidemiological and human clinical studies that examined the shape
of the concentration-response function for different averaging times or exposure durations are
presented below. The discussion focuses on respiratory morbidity effects associated with short-
term exposure to S02, for which the strongest causal evidence exists.
4.1.1. Evidence from Human Clinical Studies
In human clinical studies of exercising asthmatics, moderate SCVinduced decrements in
lung function have been observed at the lowest levels tested (i.e., 0.2 to 0.3 ppm, 5 to 10 min
exposures) in the most sensitive individuals (approximately 5-20% of subjects). Statistically
significant respiratory effects have been consistently observed at concentrations of 0.4-0.6 ppm,
with 20-60% of asthmatics experiencing moderate to large decrements in lung function following
5-10 min exposures (see Table 3-1). Smaller, yet statistically significant decrements in lung
function have also been demonstrated at SO2 concentrations < 0.2 ppm when preceded by
exposure to 03 (see Section 3.1.5.1).
With increasing exposure concentration between 0.2 and 1.0 ppm, there is a clear increase
both in magnitude of respiratory effect and percent of asthmatics affected (Table 3-1). A subset of
the data presented in this table was taken from a series of studies conducted by Linn et al. (1987;
1988; 1990) and is presented graphically in Figure 4-1 through Figure 4-3. In these studies, mild
and moderate asthmatics were exposed for 10 min to SO2 concentrations between 0 and 0.6 ppm
during moderate to heavy exercise. These particular studies were selected for inclusion in this
meta-analysis owing to similarities between exposure protocols, with all subjects being exposed
to multiple concentrations of S02. In the 1987 study, subjects were exposed to S02
concentrations of 0, 0.2, 0.4, and 0.6 ppm, while in the 1988 and 1990 studies, subjects were
exposed to concentrations of 0, 0.3, and 0.6 ppm. The percent of asthmatics experiencing
moderate or greater SCVinduced decrements in lung function (increase in sRaw > 100% or
decrease in FEVi > 15%) is shown in Figure 4-1. At 0.2 ppm, between 5 and 13% of subjects are
affected, and this fraction increases with increasing concentration, with approximately 50% of
subjects experiencing respiratory effects at a concentration of 0.6 ppm.
May 2008
4-2
DRAFT—DO NOT QUOTE OR CITE
-------
c
c
§L
o
*3
£
CO
<
c
m
2
CD
0.
70
60
50
40
30
z 20
10
~ sRaw
¦ FEV,
13%
5%
E
30%
22%
23%
II
51%
47% 1
0.10 0.20 0.30 0.40 0.50
S02 Concentration (ppm)
0.60
Figure 4-1. Percent of mild and moderate asthmatics (vE = 40-50 L/min) experiencing an S02-
induced increase in sRaw of > 100% or a decrease in FEV1 of > 15%, adjusted for effects of
moderate to heavy exercise in clean air. The data represents lung function measurements from
40, 41, 40, and 81 subjects at concentrations of 0.2, 0.3, 0.4, and 0.6 ppm, respectively.
Source: Data taken from Linn et al. (1987; 1988; 1990)
1 Figure 4-2 and Figure 4-3 present the concentration-response relationship between S02 and
2 decrements in lung function in SCVsensitive asthmatics, i.e., those asthmatics experiencing
3 significant decrements in lung function at the highest exposure concentration (0.6 ppm) used in
4 the Linn et al. studies (1987; 1988; 1990). This analysis demonstrates a clear increase in the
5 magnitude of respiratory effects with increasing exposure concentration, with more marked
6 effects observed at SO2 concentrations greater than 0.3 ppm. The results of a study by Gong et al.
7 (1995) support this conclusion: the authors observed a linear relationship between SO2
8 concentration (0, 0.5, and 1.0 ppm) and both lung function (decrease in FEVi, and increase in
9 sRaw) and respiratory symptoms.
May 2008
4-3
DRAFT—DO NOT QUOTE OR CITE
-------
Csl
° Cg
CO
? to
5 °
O 'J=
£ <2
< E
5 £
TO (0
a: <
w a)
a> a)
w "a
re o
2
c 2
¦o
a>
o
3
¦D
C
(0
c
cm a)
O W
W
0.3 0.4 0.:
S02 Concentration (ppm)
Figure 4-2. S02-induced increase in sRaw among S02-sensitive mild and moderate asthmatics
(n=38) following 10-min exposures during moderate to heavy exercise (VE = 40-50 L/min). Only
S02-sensitive asthmatics, defined here as asthmatics experiencing > 100% S02-induced increase
in sRaw at 0.6 ppm, were included in this analysis. The analysis includes data from 14 S02-
sensitive subjects from Linn et al. (1987) exposed to concentrations of 0.2, 0.4, and 0.6 ppm, as
well as 24 S02-sensitive subjects from Linn et al. 1988 and 1990 exposed to 0.3 and 0.6 ppm. Error
bars = ± 1 SE.
4.1.2. Evidence from Epidemiological Studies
1 Although there are numerous epidemiological studies that examined the association
2 between SO2 and various health effects, only a few of these studies attempted to evaluate the
3 concentration-response function. Most studies assumed a linear or log-linear relationship
4 between ambient SO2 concentrations and the health outcome in their evaluations.
5 Epidemiological studies have examined the concentration-response relationship for SO2
6 using various statistical methods, including the comparison of effect estimates in increasing
7 quartiles or quintiles, plotting the risk observed against increasing S02 concentrations, and using
8 nonparametric smoothed curves to assess the nonlinearity of the SCVeffect relationship. Most of
9 the epidemiological studies that examined the concentration-response function between SO2
10 exposure and respiratory morbidity observed that the relationship was linear across the entire
11 concentration range.
May 2008
4-4
DRAFT—DO NOT QUOTE OR CITE
-------
5
o
i
-30
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
SO2 Concentration (ppm)
Figure 4-3. S02-induced decrease in FEV1 among S02-sensitive mild and moderate asthmatics
(n=41) following 10 min exposures during moderate to heavy exercise (VE = 40-50 L/min). Only
S02-sensitive asthmatics, defined here as asthmatics experiencing > 15% S02-induced decrease
in FEV1 at 0.6 ppm, were included in this analysis. The analysis includes data from 21 S02-
sensitive subjects from Linn et al. (1987) exposed to concentrations of 0.2, 0.4, and 0.6 ppm, as
well as 20 S02-sensitive subjects from Linn et al. 1988 and 1990 exposed to 0.3 and 0.6 ppm. Error
bars = ± 1 SE.
1 Only one epidemiological study investigated the concentration-response function of peak
2 SO2 exposures. The association between asthma hospitalizations and ambient 1-h max SO2
3 concentrations was examined in a case-control study of children in Bronx County, NY (Lin et al.,
4 2004). The 1-h max concentration ranged from 2.9 to 66.4 ppb. The authors categorized 1-h max
5 S02 concentrations and estimated ORs for each category using the lowest exposure group as the
6 reference (2.9 to 9.2 ppb). They observed an increasing linear trend across the range of
7 concentrations, with more marked effects observed at 1-h max SO2 concentrations greater than
8 40 ppb Figure 4-4. A 1-h max SO2 concentration of 40 ppb falls between the 90th and 95th
9 percentiles of the ambient regulatory data for the years 2003-2005; during these years 24-h avg
10 SO2 concentrations for the 90th and 95th percentiles were 10 and 13 ppb, respectively.
May 2008 4-5 DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
-~ Same day
—¦ 2-d iag
¦•A 3-d lag
¦ o 4-d lag
Figure 4-4. Adjusted odds ratios of asthma hospitalizations by groupings of 1-h max S02
concentrations in Bronx County, New York. All groups were compared with the lowest exposure
group (2.9-9.2 ppb). ORs for 1-h max S02 concentrations on the same day, as well as from a 2-day,
3-day, and 4-day moving average lag are presented.
Source: Lin et al. (2004)
Most epidemiological studies investigating the concentration-response function examined
the effects of short-term 24-h avg exposures to S02. The Harvard Six Cities study by Schwartz
et al. (1994) investigated the concentration-response function and observed a nonlinear
relationship between SO2 concentrations and respiratory symptoms. A figure plotting the relative
odds of incidence of lower respiratory tract symptoms against SO2 concentrations lagged 1 day
indicated that no statistically significant increase in the incidence of lower respiratory tract
symptoms was seen until concentrations exceeded a 24-h avg SO2 of 22 ppb though an
increasing trend was observed at concentrations as low as 10 ppb (see Figure 4-5). In a study of
respiratory hospitalizations, Ponce de Leon et al. (1996) found that a weak relationship with SO2
was only observable at 24-h avg S02 concentrations above 23 ppb. In both this study and the
study by Schwartz et al. (1994), a statistically significant increased risk was observable only at
24-h avg SO2 concentrations that were above the 90th percentile. The nonlinearity observed in
May 2008
4-6
DRAFT—DO NOT QUOTE OR CITE
-------
1 these concentration-response functions is dependent on only a few influential observations; thus,
2 the results should be viewed with caution.
1.6 -
o
»
o
0
10
20
30
40
24-h avg S02 (ppb)
Figure 4-5. Relative odds ratio of incidence of lower respiratory tract symptoms smoothed
against 24-h avg S02 concentrations on the previous day, controlling for temperature, city, and
day of week.
Source: Schwartz et al. (1994).
3 A study by Jaffe et al. (2003) examined the association between SO2 and emergency
4 department (ED) visits for asthma in three cities in Ohio, and found significant associations only
5 in Cincinnati using Poisson regression analysis. To examine the concentration-response function,
6 they also conducted quintile analyses. In Cincinnati, an increasing linear trend in risk was
7 observed across the range of concentrations. Wong et al. (2002; using GAM with default
8 convergence criteria) constructed a plot of risk against 24-h avg SO2 concentrations to examine
9 the concentration-response relationship in Hong Kong and London. In general, a linear
10 relationship between risk of respiratory hospitalizations and S02 was observed across the range
May 2008
4-7
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
of SO2 concentrations in Hong Kong, but not in London. Several other studies that examined the
concentration-response relationship found that the association between respiratory
hospitalizations and S02 did not deviate from linearity (Atkinson et al., 1999a; Burnett et al.,
1997a; 1997b; 1999; Hajat et al., 2002).
4.1.3. Summary of Evidence on Concentration-Response Functions
and Thresholds
In the previous two sections, evidence from human clinical and epidemiological studies on
the concentration-response function that may inform identification of any potential population
threshold was presented. Evidence from human clinical studies indicates a clear increase in the
magnitude of respiratory effects with increasing exposure concentration between 0.2 and 1.0
ppm with 5-10 min SO2 exposures. Epidemiological studies also have observed generally
increasing trends across the entire range of SO2 exposures; however, in a limited number of
studies a marked increase in effect was observed only at the higher concentrations (above 90th
percentile values).
Discerning a possible population-level threshold for air pollution-related effects in
epidemiological studies is quite challenging. Using PM2.5 as an example, Brauer et al. (2002)
examined the relationship between ambient concentrations and mortality risk in a simulated
population with specified common individual threshold levels. They found that no population
threshold was detectable when a low threshold level was specified. Even at high-specified
individual threshold levels, the apparent threshold at the population level was much lower than
specified. Brauer et al. (2002) concluded that the use of surrogate measures of exposure (i.e.,
those from centrally located ambient monitors) that were not highly correlated with personal
exposures made it difficult to discern population-level thresholds in epidemiological studies even
if a common threshold exists for individuals within the population.
The wide interindividual variability in sensitivity to S02 exposure further hinders the
ability to discern a potential threshold level in population studies. Human clinical studies have
shown that asthmatics experience greater increases in sRaw following peak SO2 exposures
compared to healthy individuals (Linn et al., 1987). Among asthmatics, interindividual
differences in response also have been noted, with some asthmatics experiencing S02-related
effects at much lower levels than others (Horstman et al., 1986).
May 2008
4-8
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Another factor that complicates the identification of a possible threshold of effects is that
currently deployed ambient monitors may be inadequate for accurate and precise measurements
at lower 24-h avg S02 levels. Ambient concentrations of S02 have been declining since the
1980s and are now at or very near the limit of detection of the ambient monitors in the regulatory
network. The mean 24-h avg SO2 concentration across the metropolitan statistical areas (MSAs)
from 2003 through 2005 was 4 ppb (5th-95th percentile: 1, 13). Thus, there is greater uncertainty
at the lower concentration range compared to the higher concentrations, which likely limits the
ability to detect any clear-cut potential threshold.
In conclusion, evidence from human clinical studies indicates wide interindividual
variability in response to SO2 exposures, with peak (5 to 10 min) exposures at levels as low as
0.2-0.3 ppm eliciting respiratory responses in some asthmatic individuals. Several
epidemiological studies that examined the concentration-response function between short-term
(24-h avg or 1-h max) exposure to SO2 and respiratory morbidity observed that the relationship
was linear across the entire concentration range, suggesting a lack of a threshold in effect.
However, given the various limitations in observing a possible threshold in population studies,
the lack of evidence does not necessarily indicate that there is indeed no threshold in SO2 health
effects. Some epidemiological studies did report that though there was generally an increasing
trend at the lower SO2 concentrations, a marked increase in SCVrelated respiratory health effects
was observed at higher concentrations. However, as these observations were based on a few
potentially influential data points (24-h avg S02 concentrations above the 90th percentile), the
results should be interpreted with caution. The overall limited evidence from epidemiological
studies examining the concentration-response function of SO2 health effects is inconclusive
regarding the presence of an effect threshold at current ambient levels.
4.2. Susceptible and Vulnerable Populations
Interindividual variation in human responses to air pollutants indicates that not all
individuals exposed to pollutants respond similarly. That is, some subpopulations are at increased
risk to the detrimental effects of pollutant exposure. The NAAQS are intended to provide an
adequate margin of safety for both general populations and sensitive subpopulations, or those
subgroups potentially at increased risk for ambient air pollution health effects. For the purposes
of this document, a susceptible population is defined as one that might exhibit an adverse health
May 2008
4-9
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
effect to a pollutant at concentrations lower than those needed to elicit the same response in the
general population, and a vulnerable population as one that might be differentially exposed to
higher concentrations of a pollutant than the general population, regardless of health outcome.
The previous review of the SO2 NAAQS identified certain groups within the population that may
be more susceptible to the effects of SO2 exposure, including asthmatics, individuals not
diagnosed as asthmatic but with atopic disorders (e.g., allergies), and individuals with COPD or
cardiovascular disease. Other subgroups considered to be somewhat sensitive in this document
include children and older adults; people with other respiratory disease; genetic factors;
socioeconomic status (SES); and populations experiencing heightened exposure levels (e.g.,
those living near roadways or other "hot spots" or engaged in outdoor work or exercise). Also of
concern are individuals who generally may not be susceptible to S02-related health effects but
may experience transient airways reactivity to respiratory irritants such as SO2 following a recent
viral respiratory infection (Stempel and Boucher, 1981). These groups comprise a large fraction
of the U.S. population. Given the likely heterogeneity of individual responses to air pollution, the
severity of health effects experienced by a susceptible subgroup may be much greater than that
experienced by the population at large (Zanobetti et al., 2000).
4.2.1. Preexisting Disease as a Potential Risk Factor
Several researchers have investigated the effect of air pollution among potentially
susceptible groups with preexisting medical conditions. A recent report of the National Research
Council emphasized the need to evaluate the effect of air pollution on susceptible groups,
including those with respiratory illnesses and cardiovascular diseases (National Research
Council., 2004). Generally, asthma, COPD, conduction disorders, congestive heart failure (CHF),
diabetes, and myoardial infarction (MI) are conditions believed to put persons at greater risk of
adverse events associated with air pollution. Asthmatics are known to be one of the most S02-
responsive subgroups in the population; the evidence related to respiratory illness, including
asthma and other factors, is discussed in further detail below.
4.2.1.1. Individuals with Respiratory Diseases
The 1982 AQCD concluded that asthmatics are more susceptible to respiratory effects from
SO2 exposures than the general public. This conclusion was primarily drawn from the strong
May 2008
4-10
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
human clinical evidence. Recent epidemiological studies have strengthened this conclusion,
reporting associations between a range of health outcomes with both short-term and long-term
S02 exposures in subjects with respiratory disease.
In human clinical studies, asthmatics have been shown to be more responsive to respiratory
effects of SO2 exposures than healthy non-asthmatics. While SCVattributable decrements in lung
function generally have not been demonstrated at concentrations <1.0 ppm in non-asthmatics
(Lawther et al., 1975; Linn et al., 1987; Schachter et al., 1984), statistically significant increases
in respiratory symptoms and decreases in lung function have been observed in exercising
asthmatics following peak (5 to 10 min) SO2 exposures to concentrations as low as 0.4-0.6 ppm
(Gong et al., 1995; Horstman et al., 1986; Linn et al., 1983). Respiratory effects have been
observed in some sensitive asthmatics at concentrations as low as 0.2-0.3 ppm (Horstman et al.,
1986; Linn et al., 1987). There is no evidence that individuals with COPD have increased
susceptibility to S02-induced respiratory effects.
A number of epidemiological studies reported increased respiratory morbidity associated
with S02 exposures in asthmatics and atopic individuals. Notably, two U.S. multicity studies
observed associations between ambient SO2 concentrations and respiratory symptoms in
asthmatic children (Mortimer et al., 2002; Schildcrout et al., 2006). Additional studies also have
generally indicated positive associations for asthma among children and included a U.S. study
(Delfino et al., 2003) and several European studies (Higgins et al., 1995; Neukirch et al., 1998;
Peters et al., 1996; Roemer et al., 1993; Segala et al., 1998; Taggart et al., 1996; Timonen and
Pekkanen, 1997; van der Zee et al., 1999). Studies of adults found no consistent association
between respiratory symptoms among asthmatics and SO2 concentrations (Desqueyroux et al.,
2002a; 2002b; Romieu et al., 1996; van der Zee et al., 2000).
A suggestive association between ambient S02 concentrations and ED visits and
hospitalizations provides further evidence that asthmatics are susceptible to the effects of SO2.
The associations between ambient concentrations of 24-h avg SO2 and ED visits and
hospitalizations for asthma in the United States are generally positive (Jaffe et al., 2003; Lin et
al., 2004; Michaud et al., 2004; Wilson et al., 2005), though a large time-series study conducted
in Atlanta, GA did not find an association between ambient 1-h max SO2 levels and ED visits
(Peel et al., 2005). Studies conducted outside the United States (Atkinson et al., 1999a; Hajat et
al., 1999; Sunyer et al., 1997; Thompson et al., 2001) also generally found positive results.
May 2008
4-11
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
In summary, substantial evidence from epidemiological studies suggests that individuals
with preexisting respiratory diseases, particularly asthma, are more susceptible to respiratory
health effects, though not mortality, from S02 exposures than the general public. The
observations from human clinical studies indicating increased sensitivity to SO2 exposures in
asthmatic subjects compared to healthy subjects provide coherence and biological plausibility for
these observations in epidemiological studies.
4.2.1.2. Individuals with Cardiovascular Diseases
The evidence available to evaluate the susceptibility of individuals with cardiovascular
disease for SCVrelated health effects is very limited. One human clinical study observed no
evidence to suggest that patients with stable angina were more susceptible to S02-related health
effects compared with healthy subjects (Routledge et al., 2006). The authors noted that this lack
of response in the heart patients may be due to a drug treatment effect rather than decreased
susceptibility, as a large portion of the angina patients were taking beta blockers, which are
known to increase indices of cardiac vagal control.
Liao et al. (2004) investigated short-term associations between ambient pollutants and
cardiac autonomic control and observed that consistently more pronounced associations were
found between SO2 and heart rate variability among persons with a history of coronary heart
disease. In another epidemiological study, Henneberger et al. (2005) examined the association of
repolarization parameters with air pollutants in East German men with preexisting coronary heart
disease. Ambient SO2 concentrations during the 24-h preceding the ECG were associated with
the QT interval duration, but not with any other repolarization parameters.
Evidence is inconsistent in studies analyzing the associations between ambient levels of air
pollutants and ED visits or hospitalizations for cardiovascular diseases. A recent epidemiological
study investigated the association of SO2 with cardiac-related hospital admissions among persons
with preexisting cardiopulmonary conditions and observed no associations with ambient 1-h max
SO2 level for any cardiac disease investigated (i.e., ischemic heart disease [IHD], CHF, and
dysrhythmia) across strata of comorbid disease status, including hypertension, diabetes, and
COPD (Peel et al., 2007).
Goldberg et al. (2003) compared the risk estimates for death with the underlying cause of
CHF and those deaths classified as having CHF 1 year before death and did not find associations
May 2008
4-12
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
between air pollution and those with CHF as an underlying cause of death. The authors found
associations between some of the air pollutants examined (coefficient of haze [CoH], SO2, and
N02) and the deaths that were classified as having CHF 1 year before death, but the association
with the specific cause of death was not unique to SO2. This pattern of association, including but
not specific to SO2, with specific causes of death also was observed in an additional cohort of
patients with CHF (Kwon et al., 2001).
In conclusion, the very limited evidence examining the susceptibility of individuals with
preexisting cardiovascular disease to adverse health effects from ambient SO2 exposures is
inconclusive.
4.2.2. Genetic Factors for Oxidant and Inflammatory Damage from Air
Pollutants
A consensus now exists among scientists that genetic factors related to health outcomes
and ambient pollutant exposures merit serious consideration (Gilliland et al., 1999; Kauffmann,
2004). Several criteria must be satisfied in selecting and establishing useful links between
polymorphisms in candidate genes and adverse respiratory effects. First, the product of the
candidate gene must be significantly involved in the pathogenesis of the effect of interest, which
is often a complex trait with many determinants. Second, polymorphisms in the gene must
produce a functional change in either the protein product or in the level of expression of the
protein. Third, in epidemiological studies, the issue of confounding by other genes or
environmental exposures must be carefully considered.
Several glutathione S-transferase (GST) families have common, functionally important
polymorphic alleles (e.g., homozygosity for the null allele at the GSTM1 and GSTT1 loci,
homozygosity for the A105G allele at the GSTP1 locus) that significantly reduce expression of
enzyme function in the lung. Exposure to radicals and oxidants from air pollution induces
decreases in GSH that increase GST transcription. Individuals with genotypes that result in
enzymes with reduced or absent glutathione peroxidase activity are likely to have reduced
oxidant defenses and increased susceptibility to inhaled oxidants and radicals.
Gilliland et al. (2002) examined effects of GSTM1, GSTT1, and GSTP1 genotypes and
acute respiratory illness, specifically respiratory illness-related absences from school. The goal
was to examine potential susceptibilities on this basis, but not specifically to air pollutants. They
May 2008
4-13
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
concluded that fourth grade schoolchildren who inherited a GSTP1 Val-105 variant allele had a
decreased risk of respiratory illness-related school absences, indicating that GSTP1 genotype
influences the risk and/or severity of acute respiratory infections in school-aged children.
Lee et al. (2004) studied ninth grade schoolchildren with asthma in Taiwan for a gene-
environmental interaction between GSTP1-105 genotypes and outdoor pollution. They examined
general district air pollution levels of low (mean SO2 level of 3.6 ppb from 1994 to 2001),
moderate (mean S02 of 6.2 ppb), and high (mean S02 of 8.6 ppb) and found that compared with
individuals with any Val-105 allele in the low air pollution district, Ile-105 homozygotes in the
high air pollution district had a significantly increased risk of asthma.
Gauderman et al. (2007) describe a study method that uses principal components analysis
computed on single nucleotide polymorphism (SNP) markers to test for an association between a
disease and a candidate gene. For example, they evaluated the association between respiratory
symptoms in children and four SNPs in the GSTP1 locus, using data from the Southern
California Children's Health Study (CHS). The authors observed stronger evidence of an
association using the principal components approach (p = 0.044) than using either a genotype-
based (p = 0.13) or haplotype-based (p = 0.052) approach. This method may be applied to
relationships in this and other databases to evaluate aspects of air pollutants such as SO2.
Winterton et al. (2001) attempted to identify a genetic biomarker for susceptibility to SO2.
They screened 62 asthmatic subjects for SO2 responsiveness using an inhalation challenge and
collected genetic material via buccal swabs to test for associations between S02 sensitivity and
specific gene polymorphisms. Subjects inhaled 0.5 ppm SO2 by mouthpiece for 10 min while
wearing noseclips during moderate exercise on a treadmill. Subjects were defined as SO2-
sensitive if FEVi was decreased 12%. Genetic polymorphisms as biomarkers of susceptibility
were evaluated in five regions coding for the p2-adrenergic receptor, the a subunit of the
interleukin-4 (IL-4) receptor, the Clara cell secretory protein (CC16), tumor necrosis factor-a
(TNF-a), and lymphotoxin-a (also known as TNF-P). The authors found a significant association
between response to SO2 and the homozygous wild-type allele of TNF-a. All of the SO2-
sensitive subjects had the homozygous wild-type allele for TNF-a, while 61% of the
nonresponders had this genotype. Homozygosity for the TNF-1 allele was associated with a 5-
fold increased risk of physician-diagnosed asthma relative to other genotypes. None of the other
polymorphisms showed significant trends.
May 2008
4-14
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
In summary, the differential effects of air pollution among genetically diverse
subpopulations have been examined for a number of GST genes and other genotypes. The
limited number of studies may provide some insight into susceptible groups and a potential
genetic role in such. Only one of these studies specifically examined SO2 as the exposure of
interest, and it found a significant association with the homozygous wild-type allele for TNF-a.
Khoury et al. (2005) states that while genomics is still in its infancy, opportunities exist for
developing, testing, and applying its tools to public health research of outcomes with possible
environmental causes. At this time, there are insufficient data on which to base a conclusion
regarding the effect of SO2 exposure on genetically distinct subpopulations.
4.2.3. Age-Related Susceptibility
The American Academy of Pediatrics (2004) notes that children and infants are among the
most susceptible to many air pollutants, including S02. Eighty percent of alveoli are formed
postnatally and changes in the lung continue through adolescence; furthermore, the developing
lung is highly susceptible to damage from exposure to environmental toxicants (Dietert et al.,
2000). Children also have increased vulnerability as they spend more time outdoors, are highly
active, and have high minute ventilation, which collectively increase the dose they receive
(Plunkett et al., 1992; Wiley, 1991a; 1991b). In addition to children, the elderly are frequently
classified as being particularly susceptible to air pollution. The basis of the increased sensitivity
in the elderly is not known, but one possibility is that it may be related to changes in the
respiratory tract lining fluid antioxidant defense network (Kelly and Mudway, 2003) or a general
reduction in immune competence.
Adverse respiratory effects have been observed in adolescents following SO2 exposure in a
laboratory setting (Koenig et al., 1981; 1983; 1987; 1988; 1990). However, there is no evidence
from human clinical studies to suggest that the respiratory effects in adolescents are more severe
than those observed in adults. Similarly, a number of epidemiological studies have observed
increased respiratory symptoms in children associated with increasing SO2 exposures (Mortimer
et al., 2002; Schildcrout et al., 2006; Schwartz et al., 1994), though there is no evidence from a
limited number of studies suggesting this same effect in adults (Desqueyroux et al., 2002a;
Romieu et al., 1996; van der Zee et al., 2000).
May 2008
4-15
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
A number of studies, investigating the association between ambient SO2 levels and ED
visits or hospital admissions for all respiratory causes or asthma, stratified their analyses by age
group. Figure 4-6 summarizes the evidence of age-specific associations between S02 and acute
respiratory ED visits and hospitalizations. Several studies demonstrated that the excess risk of
ED visits or hospitalizations for all respiratory causes or asthma was higher for children
(Anderson et al., 2001; Atkinson et al., 1999a; Atkinson et al., 1999b; Petroeschevsky et al.,
2001) and older adults (Anderson et al., 1998; Petroeschevsky et al., 2001; Ponce de Leon et al.,
1996; Wilson et al., 2005) when compared to the risk for all ages together. Increased risks for
children and older adults were more prevalent in the studies of all respiratory disease than those
considering asthma as the outcome.
Cakmak et al. (2007) reported that among seven Chilean urban centers, the percent
increase in nonaccidental mortality associated with a 10 ppb increase in 24-h avg SO2 was 3.4%
(95% CI: 0.7, 6.1) for those < 65 years of age and 5.6% (95% CI: 2.2, 9.1) for those > 85 years
of age. The authors concluded that the elderly are particularly susceptible to dying from air
pollution, and suggested that concentrations deemed acceptable for the general population may
not adequately protect the very elderly.
There is limited epidemiologic evidence to suggest that children and older adults (65+
years) are more susceptible to the adverse respiratory effects associated with ambient SO2
concentrations when compared to the general population.
4.2.4. Other Potentially Susceptible Populations
There are a number of other potentially susceptible groups that, while not included here
due to a lack of data specific to SO2 exposures, deserve mention in this document. These include
obese individuals, individuals in a chronic pro-inflammatory state (e.g., diabetics), and children
born prematurely or with low birth weight.
Enhanced susceptibility for air pollution-related cardiovascular events has been shown for
older individuals and persons with conditions associated with chronic inflammation, such as
May 2008
4-16
DRAFT—DO NOT QUOTE OR CITE
-------
Reference
Wilson et al. (2005)
Wilson et al. (2005)
Atkinson elal. (1999a)
Atkinson etal. (1999b)
Anderson etal. (2001)
Wilson etal. (2005)
Wilson et al. (2005)
Atkinson et al. (1999a)
Location
Portland, ME
Manchester, NH
London, UK
London, UK
Ponce de Leon et al. (1996) London, UK
West Midlands, UK
Petroeschevsky et al. (2001) Brisbane, Australia
Portland, ME
Manchester, NH
London, UK
Atkinson etal. (1999b) London, UK
Anderson et al. (1998) London, UK
Petroeschevsky etal. (2001) Brisbane, Australia
0.8
0.9
I
Relative risk
1.0
1.1
i
1.2
1.3
All Respiratory
Asthma
• All ages
° 0-14 years
¦ 65+ years
Figure 4-6. Relative risks (95% CI) of age-specific associations between short-term exposure to
S02 and respiratory ED visits and hospitalizations. Risk estimates are standardized per 10 ppb
increase in 24-h avg S02 concentrations or 40 ppb increase in 1-h max S02.
1 diabetes, coronary artery disease, and past myocardial infarctions (Bateson and Schwartz, 2004;
2 Goldberg et al., 2001; Zanobetti and Schwartz, 2002). Dubowsky et al. (2006) observed that
3 individuals with conditions associated with both chronic inflammation and increased cardiac risk
4 were more vulnerable to the short-term pro-inflammatory effects of air pollution. This included
May 2008
4-17
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
individuals with diabetes; obesity; and concurrent diabetes, obesity and hypertension. Zanobetti
and Schwartz (2001) reported more than twice the risk for hospital admissions for heart disease
in persons with diabetes than in persons without diabetes associated with exposure to ambient air
pollution, indicating that persons with diabetes are an important at-risk group. Data from the
Third National Health and Nutrition Examination Survey indicate that 5.1% of the U.S.
population older than 20 years of age have diagnosed diabetes and an additional 2.7% have
undiagnosed diabetes (Harris et al., 1998). Moreover, another study found that subjects with
impaired glucose tolerance without type II diabetes also had reduced heart rate variability
(Schwartz, 2001). This suggests the at-risk population may be even larger.
Mortimer et al. (2000) reported that among asthmatic children, birth characteristics
continue to be associated with increased susceptibility to air pollution later in life, demonstrating
that air pollution-induced asthma symptoms are more severe in children born prematurely or of
low birth weight. Specifically, the authors revealed asthmatic children born more than three
weeks prematurely or weighing less than 2,500 grams (5.5 pounds) had a six-fold decrease in
breathing capacity associated with air pollution compared to full-weight, full-term children. The
low birth weight and premature children also reported a five-fold greater incidence of symptoms
like wheezing, coughing and tightness in the chest.
4.2.5. Factors that Potentially Increase Vulnerability to S02
A limited amount of information exists on exposures to SO2 among vulnerable populations.
Because indoor and personal S02 concentrations are generally much lower than outdoor or
ambient measurements, individuals that spend most of their time indoors, such as older adults,
are not anticipated to be vulnerable to high SO2 exposures, though in some cases they may be
more susceptible to the effects of these exposures than the general population due to preexisting
health factors.
Other individuals with increased vulnerability include those who spend a lot of time
outdoors at increased exertion levels, for example outdoor workers and individuals who exercise
or play outdoor sports. Exercise may cause an increase in uptake of SO2 resulting from an
increase in ventilation rate and accompanying shift from nasal to oronasal breathing. Children,
who generally spend more time playing outdoors, may qualify as both a susceptible population
May 2008
4-18
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
due to their developing physiology and as a vulnerable population since ambient SO2
concentrations are several-fold higher than indoor concentrations.
Residential location is not as strong of a predictor of exposure vulnerability for S02 as for
traffic-related pollutants, because meteorological conditions have a greater impact on pollutant
plume direction from primary point sources such as coal-fired power plants.
Social economic status (SES) is a known determinant of health and there is evidence that
SES modifies the effects of air pollution (Makri and Stilianakis, 2007; O'Neill et al., 2003). Both
higher exposures to air pollution and greater susceptibility to its effects may contribute to a
complex pattern of risk among those with lower SES. Conceptual frameworks have been
proposed to explain the relationship between SES, susceptibility and exposure to air pollution.
Common to these frameworks is the consideration of the broader social context in which people
live and its effect on health in general (Gee and Payne-Sturges, 2004; O'Neill et al., 2003), as
well as on maternal and child health (Morello-Frosch and Shenassa, 2006), and asthma (Wright
et al., 2007) specifically. Multilevel modeling approaches that allow parameterization of
community level stressors such as increased life stress, as well as individual risk factors, are
considered by these authors. In addition, statistical methods that allow for temporal and spatial
variability in exposure and susceptibility, have been discussed in the recent literature (Jerrett and
Finkelstein, 2005; Kunzli, 2005).
Most studies to date have examined modification by SES indicators on the association
between mortality and PM (Finkelstein et al., 2003; Jerrett et al., 2004; Martins et al., 2004;
O'Neill et al., 2003; Romieu et al., 2004) or other indices such as traffic density, distance to
roadway or a general air pollution index (Finkelstein et al., 2005; Ponce et al., 2005; Woodruff et
al., 2003). However, modification of SO2 associations has been examined in a few studies. For
example, in a study conducted in 10 large Canadian cities, living in communities in which
individuals have lower household education and income levels increased the individual's
vulnerability to air pollution (Cakmak et al., 2006). These effects were statistically significant for
several gaseous criteria pollutants, but not for SO2. In addition, Finkelstein et al. (2003)
evaluated neighborhood levels of income and air pollution in southern Ontario, Canada. They
found that both income and SO2 levels were associated with mortality differences. Specifically,
among people with below-median income, the relative risk for those with above-median
exposure to SO2 was 1.18 (95% CI: 1.11, 1.26); the corresponding relative risk among subjects
May 2008
4-19
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
with above-median income was 1.03 (95% CI: 0.83, 1.28). Overall, there is very limited
evidence available from which conclusions on the human health effects from the interaction
between SES and S02 can be drawn.
4.3. Potential Public Health Impacts
Exposure to ambient SO2 is associated with a variety of outcomes including increases in
respiratory symptoms, particularly among asthmatic children, and ED visits and hospital
admissions for respiratory diseases among children and older adults (65+ years). In protecting
public health, a distinction must be made between health effects that are considered "adverse"
and those that are not. What constitutes an adverse health effect varies for different population
groups. Some changes in healthy individuals are not viewed as adverse while those of similar
type and magnitude in other susceptible individuals with preexisting disease are.
4.3.1. Concepts Related to Defining Adverse Health Effects
The American Thoracic Society (ATS) published an official statement titled "What
Constitutes an Adverse Health Effect of Air Pollution?" (ATS, 2000a). This statement updated
the guidance for defining adverse respiratory health effects that had been published 15 years
earlier (Society, 1985), taking into account new investigative approaches used to identify the
effects of air pollution and reflecting concern for impacts of air pollution on specific susceptible
groups. In the 2000 update, there was an increased focus on quality of life measures as indicators
of adversity and a more specific consideration of population risk. Exposure to air pollution that
increases the risk of an adverse effect to the entire population is viewed as adverse, even though
it may not increase the risk of any identifiable individual to an unacceptable level. For example,
a population of asthmatics could have a distribution of lung function such that no identifiable
single individual has a level associated with significant impairment. Exposure to air pollution
could shift the distribution to lower levels that still do not bring any identifiable individual to a
level that is associated with clinically relevant effects. However, this shift to a lower level of lung
function would be considered adverse because individuals within the population would have
diminished reserve function and, therefore, would be at increased risk if affected by another
agent.
May 2008
4-20
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Reflecting new investigative approaches, the ATS statement also describes the potential
usefulness of research into the genetic basis for disease, including responses to environmental
agents that provide insights into the mechanistic basis for susceptibility and provide markers of
risk status. Likewise, biomarkers that are indicators of exposure, effect, or susceptibility may
someday be useful in defining the point at which one or an array of responses should be
considered an adverse effect.
The 2006 Ozone AQCD (EPA, 2006) provided information helpful in defining adverse
health effects associated with ambient O3 exposure by describing the gradation of severity and
adversity of respiratory-related O3 effects. The definitions that relate to responses in impaired
individuals are reproduced and presented here in Table 4-1. The severity of effects described in
the tables and the approaches taken to define their relative adversity are valid and reasonable in
the context of the new ATS (2000b) statement.
Table 4-1. Gradation of individual responses to short-term S02 exposure in individuals with
impaired respiratory systems.
FUNCTIONAL
RESPONSE
NONE
SMALL
MODERATE
LARGE
FEV1 change
Decrements of < 3%
Decrements of 3 -10%
Decrements of 10 - 20%
Decrements of > 20%
Nonspecific bronchial
responsiveness3
Within normal range
Increases of < 100%
Increase of < 300%
Increases of > 300%
Airway resistance
(sRaw)
Within normal range
(±20%)
sRaw increased < 100%
sRaw increased up to 200% or
up to 15 cm H2O • s
sRaw increased > 200% or
more than 15 cm h^O- s
Duration of response
None
< 4 h
4 h - 24 h
> 24 h
SYMPTOMATIC
RESPONSE
NORMAL
MILD
MODERATE
SEVERE
Wheeze
None
With otherwise normal breathing
With shortness of breath
Persistent with
shortness of breath
Cough
Infrequent cough
Cough with deep breath
Frequent spontaneous cough
Persistent
uncontrollable cough
Chest pain
None
Discomfort just noticeable on
exercise or deep breath
Marked discomfort on
exercise or deep breath
Severe discomfort on
exercise or deep breath
Duration of response
None
< 4 h
4 h - 24 h
> 24 h
IMPACT OF
RESPONSES
NORMAL
MILD
MODERATE
SEVERE
Interference with
normal activity
None
Few individuals
choose to limit activity
Many individuals
choose to limit activity
Most individuals
choose to limit activity
Medical treatment
No change
Normal medication as needed
Increased frequency
of medication use or additional
medication
Physician or emergency
room visit
aAn increase in nonspecific bronchial responsiveness of 100% is equivalent to a 50% decrease in PD20 or PD100.
This table is adapted from the 2006 Ozone AQCD (Table 8-3, page 8-68) (EPA, 26).
May 2008
4-21
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
As assessed in detail in earlier chapters of this document and briefly recapitulated in
preceding sections of this chapter, exposures to a range of SO2 concentrations have been reported
to be associated with increasing severity of health effects, ranging from respiratory symptoms to
ED visits and hospital admission for respiratory causes. Respiratory effects associated with
short-term S02 exposures have been by far the most extensively studied and most clearly shown
to be causally related to SO2 exposure. Such effects are observed among children, older adults,
and persons with respiratory impairments.
4.3.2. Estimation of Potential Numbers of Persons in At-Risk
Susceptible Population Groups in the United States
Although S02-related health risk estimates may appear to be small, they may be significant
from an overall public health perspective due to the large numbers of individuals in the potential
risk groups. Several subpopulations have been identified as possibly having increased
susceptibility or vulnerability to adverse health effects from SO2, including children, older adults,
and individuals with preexisting pulmonary diseases. One consideration in the assessment of
potential public health impacts is the size of various population groups that may be at increased
risk for health effects associated with SCVrelated air pollution exposure. Table 4-2 summarizes
information on the prevalence of chronic respiratory conditions in the U.S. population in 2004
and 2005 (NHIS, 2006a, 2006b).
Of most concern are those individuals with preexisting respiratory conditions, with
approximately 10% of adults and 13% of children having been diagnosed with asthma and 6% of
adults with COPD (chronic bronchitis and/or emphysema). As summarized in Section 3.1.3.5,
human clinical studies indicate that a significant fraction (20-60%) of asthmatic individuals
experience moderate or greater decrements in lung function as well as increased respiratory
symptoms following peak (5-10 min) SO2 exposures to concentrations of as low as 0.4-0.6 ppm
(Table 3-1). Some sensitive asthmatics (5-20%) have been shown to experience moderate
decrements in lung function at concentrations between 0.2 and 0.3 ppm. Among asthmatics, both
the magnitude of S02-induced decrements in lung function as well as the percent of individuals
affected have been shown to increase with increasing exposure concentrations between 0.2 and
1.0 ppm.
May 2008
4-22
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
In addition, subpopulations based on age group also comprise substantial segments of the
population that may be potentially at risk for SCVrelated health impacts. Based on U.S. Census
data from 2000, about 72.3 million (26%) of the U.S. population are under 18 years of age,
18.3 million (7.4%) are under 5 years of age, and 35 million (12%) are 65 years of age or older.
Hence, large proportions of the U.S. population are included in age groups that are considered
likely to have increased susceptibility and vulnerability for health effects from ambient SO2
exposure. For example, Figure 4-6 demonstrates that the S02-related excess risk for asthma is,
on average, 50% higher among children when compared to risk estimates that include all ages
with a 10 ppb increase in 24-h avg SO2 concentration.
Table 4-2. Prevalence of selected respiratory disorders by age group in the United States (2004
[U.S. adults] and 2005 [U.S. children] National Health Interview Survey).
CHRONIC CONDITION/DISEASE
ADULTS (18+ YEARS)
AGE (YEARS)
ALL ADULTS
18-44
45-64
65-74
75+
CASES (x 106)
%
%
%
%
%
Respiratory Conditions: Asthma
14.4
6.7
6.4
7.0
7.5
6.6
COPD: Chronic Bronchitis
8.6
4.2
3.2
4.9
6.1
6.3
COPD: Emphysema
3.5
1.7
0.3
2
4.9
6.0
CHRONIC CONDITION/DISEASE
CHILDREN (<18 YEARS)
ALL CHILDREN
0-4
5-11
12-17
CASES (x 106)
%
%
%
%
Respiratory Conditions
6.5
8.9
6.8
9.9
9.6
Source: National Center for Health Statistics (2006a,b)
Evidence indicates that several groups are potentially at increased risks from SO2
exposures compared to the average population. Susceptible subgroups include individuals with
preexisting disease, especially asthma, and children and older adults. Other individuals with
potentially increased vulnerability include those who spend a lot of time outdoors at increased
exertion levels (e.g., outdoor workers, children, individuals who exercise or play sports) and
those in proximity to large uncontrolled or poorly controlled sources. The considerable size of
the population groups at risk indicates that exposure to ambient SO2 could have a potentially
significant impact on public health in the United States, with the greatest public health risks for
the smaller subset of susceptible individuals exposed to relatively high peak SO2 concentrations.
May 2008
4-23
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Chapter 5. Summary and Conclusions
Previous chapters present the most policy-relevant information related to the review of the
NAAQS for SOx, which are 0.14 ppm averaged over a 24-hour period not to be exceeded more
than once per year, and 0.030 ppm annual arithmetic mean, with SO2 as the indicator. This
chapter summarizes and integrates key findings from atmospheric sciences, ambient air data
analyses, exposure assessment, dosimetry, and health evidence. The EPA framework for causal
determinations described in Chapter 1 has been applied to the body of evidence in order to judge
the scientific data about exposure to SOx and health effects in a two-step process. This
framework draws from similar efforts across the Federal government and the wider scientific
community, especially from the recent NAS Institute of Medicine document Improving the
Presumptive Disability Decision-Making Process for Veterans (IOM, 2007). The first step is to
determine the weight of evidence in support of causation at relevant pollutant exposures and
characterize the strength of any resulting causal classification. The EPA framework applied here
employs a five-level hierarchy for causal determination to be consistent with the Guidelines for
Carcinogen Risk Assessment (EPA, 2005):
¦ Sufficient to infer a causal relationship
¦ Sufficient to infer a likely causal relationship (i.e., more likely than not)
¦ Suggestive but not sufficient to infer a causal relationship
¦ Inadequate to infer the presence or absence of a causal relationship
¦ Suggestive of no causal relationship
The second step evaluates the quantitative evidence regarding the concentration-response
relationships including levels and exposure durations at which effects are observed. These two
steps characterize the health effects of SOx and levels at which effects may occur.
5.1. Emissions and Ambient Concentrations of S02
Anthropogenic SO2 is emitted mainly by fossil fuel combustion (chiefly coal and oil) and
metal smelting. The largest source of emissions is from elevated point sources such as the stacks
of power plants and industrial facilities. Since 1990, in response to controls applied under the
Acid Rain Program (EPA, 2006), S02 emissions from these sources have declined substantially.
May 2008
5-1
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Emissions demonstrate a strong gradient increasing from west to east, owing to the high
concentration of SCVemitting electric generating utilities in the Ohio River Valley and regions to
the south. PRB levels of S02 are estimated to be in the range of a few hundredths of a ppb (< 1%
of typical ambient levels) across most of the United States, though much higher values are found
in areas affected by volcanic or geothermal activity or in areas affected by episodic transport of
high concentration plumes from Asia and Europe.
The levels of the current primary NAAQS for SOx are 0.14 ppm for 24-h avg S02
concentrations and 0.03 ppm for an annual avg SO2 concentration. Exceedances in recent years
have become rare, as the mean 24-h and annual avg SO2 concentrations in the United States for
the years 2003 to 2005 were ~4 ppb, with 99th percentile values of -25 ppb for the 24-h avg, and
-15 ppb for both the 99th percentile and max values of the annual avg. Mean 1-h max
concentrations in these years were -13 ppb, with a 99th percentile value of -120 ppb and
maximum value of -700 ppb. The large differences between 99th percentile and maximum values
for the shorter term averages suggest that the maxima are strongly limited spatially and
temporally and are not a major determinant of the mean values. The nonuniform spatiotemporal
distribution of 5-min data, which are voluntarily supplied from a very few monitors without a
specific regulatory mandate makes them very difficult to use quantitatively for determining
concentrations and exposures at this very short time duration.
5.2. Health Effects of S02
Evaluation of the health evidence, with consideration of issues related to atmospheric
sciences, exposure assessment, and dosimetry, led to the conclusion that it is sufficient to infer a
causal relationship between respiratory morbidity and short-term exposure to SO2. This
conclusion is supported by the consistency, coherence, and plausibility of findings observed in
human clinical studies with 5-10 min exposures, epidemiological studies mostly using 24-h avg
exposures,7 and animal toxicological studies using exposures of minutes to hours.
The respiratory health effects of SO2 are consistent with the mode of action of SO2 as it is
currently understood. The immediate effect of SO2 on the respiratory system is
bronchoconstriction. This response is mediated by chemosensitive receptors in the
tracheobronchial tree. These receptors trigger reflexes at the central nervous system level
resulting in bronchoconstriction, mucus secretion, mucosal vasodilation, cough, and apnea
May 2008
5-2
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
followed by rapid shallow breathing. In some cases, local nervous system reflexes also may be
involved. Asthmatics are more sensitive to the effects of SO2 likely resulting from preexisting
inflammation associated with this disease. This inflammation may lead to enhanced release of
mediators, alterations in the autonomic nervous system and/or sensitization of the
chemosensitive receptors. These biological processes are likely to underlie the respiratory
symptoms; exacerbations of airways inflammation, reactivity, and responsiveness; and decreased
lung function observed in response to S02 exposure.
The strongest evidence for this causal relationship comes from human clinical studies
reporting respiratory symptoms and decreased lung functions following peak exposures of 5-
10 min duration to SO2 at concentrations which have sometimes been measured in ambient air
for similarly short-time durations. These effects are particularly evident among exercising
asthmatics, with some sensitive asthmatics (5-20%) experiencing moderate or greater decrements
in lung function at SO2 concentrations as low as 0.2-0.3 ppm (see Table 5-1). At concentrations
Table 5-1. Key health effects of short-term exposure to S02 observed in human clinical studies.
CONCENTRATION
EXPOSURE
EFFECTS
STUDIES
0.2-0.3 ppm
5-10 min
Moderate to large reductions in FEV1 and increases in specific airway
resistance (sRaw) observed among some asthmatic adults (5-20%) during
moderate to heavy exercise. Bronchial responsiveness to SO2 may be
enhanced when preceded by exposure to 03. Limited evidence of S02-
induced increases in respiratory symptoms.
Bethel et al. (1985); Horstman
et al. (1986); Koenig et al. (1990);
Linnet al. (1983, 1987; 1988;
1990); Schachter et al. (1984);
Sheppard et al. (1981); Trenga et
al. (2001)
1-6 h
Enhanced sensitivity to an inhaled allergen following exposure to SO2 with
NO2 in resting asthmatics. No evidence of respiratory symptoms or decre-
ments in lung function in resting asthmatics or healthy adults. Some weak
and inconsistent evidence to suggest that SO2 exposure may lead to
changes in heart rate variability.
Devalia et al. (1994) ; Routledge
et al. (2006); Rusznak et al.
(1996); Tunnicliffe et al. (2001,
2003)
0.4-0.5 ppm
1-10 min
Decrements in lung function clearly demonstrated in asthmatics during
exercise with significant interindividual variability in response (approximately
30% of asthmatics experienced moderate or greater decrements in lung
function). Effects observed within 1-5 min of exposure generally not
enhanced by increasing exposure duration. Respiratory symptoms (e.g.,
wheezing, chest tightness) observed at concentrations as low as 0.4 ppm
and have been shown to increase with increasing exposure concentrations.
Balmes et al. (1987); Gong et al.
(1995); Horstman et al. (1986);
Koenig et al. (1983); Linn et al.
(1983, 1987); Magnussen et al.
(1990); Schachter et al. (1984);
Sheppard et al. (1981); Trenga et
al. (1999)
~1-h
Decrements in lung function among asthmatics following 10 min of exercise
at the end of a 60-75 min exposure are statistically significant, but less
severe than effects observed following a 10 min period of exercise at the
start of the exposure.
Linn et al. (1987); Roger et al.
(1985)
0.6-1.0 ppm
1-10 min
Clear and consistent SCVinduced increases in respiratory symptoms ob-
served among exercising asthmatics. Moderate to large decrements in lung
function demonstrated in 35-60% of asthmatics. Respiratory effects
attributed to SO2 among asthmatics during exercise may be diminished after
cessation of exercise, even with continued SO2 exposure. No respiratory
effects reported in healthy, non-asthmatics.
Balmes et al. (1987); Gong et al.
(1995); Hackney et al. (1984);
Horstman et al. (1986, 1988);
Koenig et al. (1983); Linn et al.
(1987; 1988; 1990); Roger et al.
(1985); Schachter et al. (1984)
1-6 h
Decrements in lung function among asthmatics following 5-10 min of exer-
cise at the end of a 1-6 h exposure are statistically significant, but less
severe than effects observed following a 5-10 min period of exercise at the
start of the exposure.
Linn et al. (1984 ; 1987); Hackney
et al. (1984) ; Roger et al. (1985)
May 2008
5-3
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
> 0.4 ppm, a greater percentage (20-60%) of asthmatics experience SCVinduced decrements in
lung function, which are frequently accompanied by respiratory symptoms. A clear
concentration-response relationship has been demonstrated following exposures to S02 at
concentrations between 0.2 and 1.0 ppm, both in terms of severity of effect and percentage of
asthmatics adversely affected. Animal toxicological studies have also reported
bronchoconstiction with short-term exposures of 0.5 to 1 ppm SO2 (see Table 5-2).
A larger body of evidence supporting this determination of causality comes from numerous
epidemiological studies reporting associations with respiratory symptoms, ED visits, and hospital
admissions with short-term SO2 exposures, generally of 24-h avg. Almost all of these studies
were conducted in areas where the maximum ambient 24-h avg SO2 concentration was consis-
tently below the current 24-h avg NAAQS level of 0.14 ppm. Important new multicity studies
and several other studies have found an association between 24-h avg ambient SO2
concentrations and respiratory symptoms in children, particularly those with asthma.
Furthermore, limited epidemiological evidence indicates that atopic children and adults may be
at increased risk for S02-induced respiratory symptoms. Generally consistent associations also
were observed between ambient SO2 concentrations and ED visits and hospitalizations for all
respiratory causes, particularly among children and older adults (> 65 years), and for asthma.
The S02-related changes in ED visits or hospital admissions for respiratory causes ranged from -
5% to 20% excess risk. Results of experiments in laboratory animals support these observations.
Studies in animals sensitized with antigen demonstrated that repeated exposure to S02 levels as
low as 0.1 ppm exacerbated allergic responses including mucin production, airway inflammation
and airway hyperresponsiveness. These responses are consistent with exacerbation of asthma in
humans.
The consistency and internal coherence of the epidemiological evidence for respiratory
effects associated with short-term exposure to SO2 are illustrated in Figures 5-1 and 5-2, which
present effect estimates for respiratory symptoms, ED visits, and hospitalizations in children.
Associations between short-term ambient SO2 concentrations and respiratory symptoms, ED
visits, and hospitalizations are largely positive, with several of the more precise effect estimates
(suggestive of greater study power) indicating statistical significance. The epidemiological
findings of asthma symptoms with 24-h avg SO2 exposures are generally coherent with increases
May 2008
5-4
DRAFT—DO NOT QUOTE OR CITE
-------
Table 5-2. Key respiratory health effects of exposure to S02 in animal toxicological studies.
REFERENCE
EXPOSURE
SPECIES
EFFECTS
LUNG FUNCTION
Amdur et al.
(1983)
1 ppm SC^for 1-h
Male Hartley guinea pigs
An 11% increase in pulmonary resistance and 12% decrease in dynamic
compliance were observed. Neither effect persisted into the 1-h period
following exposure. No effects were observed for breathing frequency, tidal
volume, or min volume.
Conner et al.
(1985)
1 ppm (2.62 mg/m3); nose
only; 3-h/day for 6 days;
animals evaluated for up to
48-h following exposure
Hartley guinea pig,
male, age not reported,
250-320 g, n= <18
group/time point
No effect was observed on residual volume, functional reserve capacity, vital
capacity, total lung capacity, respiratory frequency, tidal volume, pulmonary
resistance, pulmonary compliance, diffusing capacity for carbon monoxide or
alveolar volume at 1 - or 48-h after last exposure.
Barthelemy
et al. (1988)
0.5 or 5 ppm
(1.3 or 13.1 mg/m3); intratra-
cheal; 45 min
Rabbit, sex not reported,
adult, mean 2.0 kg,
n = 5-9/group; rabbits
were mechanically
ventilated
Lung resistance increased by 16% and 50% in response to 0.5 and 5 ppm
SO2, respectively. Bivagotomy had no effect on 5 ppm S02-induced in-
creases in lung resistance (54% increase before vagotomy and 56% in-
crease after vagotomy). Reflex bronchoconstrictive response to phenyldigua-
nide (intravenously administered) was eliminated by exposure to SO2 but
SO2 had no effect on lung resistance induced by intravenously-administered
histamine. The study authors concluded that (1) vagal reflex is not
responsible for S02-induced increase in lung resistance at 45 min and (2)
transient alteration in tracheobronchial wall following SO2 exposure may
have reduced accessibility of airway nervous receptors to phenyldiguanide.
LUNG INJURY, INFLAMMATION AND MORPHOLOGY
Conner et al.
(1989)
1 ppm (2.62 mg/m3); nose
only 3-h/day for 5 days; bron-
choalveolar lavage
performed daily
Hartley guinea pig,
male, age not reported,
250-320 g, n = 4
No change in numbers of total cells and neutrophils, protein levels or enzyme
activity in lavage fluid following SO2 exposure.
Park et al.
(2001)
0.1 ppm (0.26 mg/m3); whole
body; with and without expo-
sure to ovalbumin, 5-h/day
for 5 days
Dunkin-Hartley guinea
pig, male, age not re-
ported, 250-350 g,
n = 7-12/group
After bronchial challenge, the ovalbumin/S02-exposed group had signifi-
cantly increased eosinophil counts in BAL fluids compared with all other
groups, including the S02-only group. The bronchial and lung tissue of the
ovalbumin/S02-exposed group showed infiltration of inflammatory cells,
bronchiolar epithelial damage, and mucus and cell plug in the lumen.
Li et al. (2007b)
2 ppm (5.24 mg/m3), with
and without exposure to
ovalbumin, 1-h/dayfor7
days
Wistar rats, male, age
not reported
Increased number of inflammatory cells in BALfluid, increased levels of
MUC5AC and ICAM-1 and an enhanced histopathological response com-
pared with those treated with ovalbumin or SO2 alone
Conner et al.
(1985)
1 ppm, 3-h/day/6 day. Evalu-
ated up to 72-h postexpo-
sure
Male Hartely guinea pigs
No alveolar lesions.
Smith et al.
(1989)
1 ppm, 5-h/day, 5 day/wk up
to 4 and 8 mos
Male Sprague-Dawley
rats
Increased bronchial epithelial hyperplasia and number of nonciliated epithe-
lial cells observed at 4 mos.
AIRWAY HYPERRESPONSIVENESS AND ALLERGY
Riedel et al.
(1988)
0.1, 4.3, or 16.6 ppm (0,
0.26, 11.3, or 43.5 mg/m3);
whole body; 8-h/day for 5
days; animals were sensi-
tized to ovalbumin on the last
3 days of exposure
Perlbright-White Guinea
pig, female, age not
reported, 300-350 g,
n = 5 or 6/group (14
controls)
Bronchial provocation with ovalbumin was conducted every other day for 2
wks, starting at 1 wk after the last exposure. Numbers of animals displaying
symptoms of bronchial obstruction after ovalbumin provocation was in-
creased in all SO2 groups compared to air-exposed groups. Anti-ovalbumin
antibodies (IgG total and lgG1) were increased in BALfluid and serum of
S02-exposed compared to air-exposed controls, with statistical significance
obtained for IgG total in BAL fluid at >4.3 ppm SO2 and in serum at all SO2
concentrations. Results indicate that in this model, subacute exposure to
even low concentrations of S02 can potentiate allergic sensitization of the
airway.
Park et al.
(2001)
0.1 ppm (0.26 mg/m3); whole
body; with and without expo-
sure to ovalbumin; 5-h/day
for 5 days
Dunkin-Hartley guinea
pig, male, age not re-
ported, 250-350 g,
n = 7-12/group
After bronchial challenge, the ovalbumin/S02-exposed group had signifi-
cantly increased enhanced pause (indicator of airway obstruction) compared
with all other groups, including the SO2 group. Study authors concluded that
low level SO2 may enhance the development of ovalbumin-induced
asthmatic reactions in guinea pigs.
May 2008
5-5
DRAFT—DO NOT QUOTE OR CITE
-------
1 in symptoms reported in asthmatics in human clinical studies with 5-10 min exposures; it is
2 possible that these epidemiological associations are determined in large part by peak exposures
3 within a 24-h period. The effects of S02 on respiratory symptoms, lung function, and airway
4 inflammation observed in the human clinical studies using peak exposures further provides a
5 basis for a progression of respiratory morbidity resulting in increased ED visits and hospital
6 admissions. Collectively, these findings provide biological plausibility for the observed
7 associations between ambient S02 levels and ED visits and hospitalizations for all respiratory
8 diseases and asthma, notably in children and older adults (> 65 years).
Odds Ratios for Respiratory Symptoms
0.40 0,80 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40
Reenter et at. (1993)
The Netherlands
SchiWcrout et ai. (2006)
7 US Cities
Romieu et al. (1996)
Mexico City, Mexico
Mortimer et al. (2002)
8 US Cities
Schwartz et al. (1994)
6 US Cities
van tier Zee et ai. (1899)
The Netherlands
Neasetal. (1995)
Oritorrtown, PA
Hoetc & Brunekreef (1993)
The Netherlands
Ward et al (2002a)
Birmingham and Sandweli UK
Pikbart et al. (1999)
Czech Republic
Segaia et al. (1998)
Paris, France
SOB
Shortness of breath
c
Cough
p
Phlegm
w
Wheeze
AS
Asthma symptoms
RN
Runny Nose
Inhaler Use
IRS
Lower respiratory
symptom
URS
Upper respiratory
symptoms
Rl
Respiratory Infection
~ SOB
Figure 5-1. Odds ratios (95% CI) for the association between short-term exposures to ambient
S02 and respiratory symptoms in children. Odds ratios are standardized per 10-ppb increase in
May 2008
5-6
DRAFT—DO NOT QUOTE OR CITE
-------
24-h avg S02 level. Studies are generally presented in the order of increasing width of the 95% CI.
Relative Risk for Respiratory ED Visits and Hospitalizations
0.80
0.80
Ponce de Leon et ai. (1896)
London, UK
Burnett etal, (2001)
Toronto, ON
Anderson et al. (2001)
Wsst Midlands. UK
Wfong et al. (1898)
Horg Kong, China
Atkinson et al. (1999a)
London, UK
Atkinson etal. (1999b)
London, UK
Luginaah et ai {2005)*
Windsor ON
Luginaah et al (2005)-*
Windsor. ON
Gouveia and Fletcher (2000)
Sao Paulo, Brazil
Wilson et al (2005)
Portland, ME
Petroeschevsky et al. (2001)
Brisbane, Australia
Fusco etal. (2001)
Rome Italy
Wilson et at (ZOOS)
Manchester, NH
Barnettetat (ZOOS)
Multeity, Australia
Bamettetai (2005)
Multeity, Australia
Anderson et al. (1998)
London, UK
Lee et al. (2006)
Hong Kong, China
Atkinson et al (1999a)
London, UK
Lin et al, (2004a)
Bronx, NY
Lin et al (2003)*
Toronto, ON
Lin et al (2003)™
Toronto, ON
Gouveia and Fletcher (2000)
Sao Paulo, Brazil
Petroeschevsky et al. (2001)
Brisbane, Australia
Fusco etal. (2001)
Kome, Italy
Petroeschevsky et al. (2001)
Brisbane, Australia
Bamett et al. (2005)
Multeity, Australia
Barnetl et ai. (2005)
Multicity, Australia
1,00 1.20 1.40 1.80 1.80 2.00
J_ J
"N
¦ ED
£
X)
3
I
o
•ED
-ED
J
J
Figure 5-2. Relative risks (95% CI) for the association between short-term exposures to ambient
S02 and emergency department (ED) visits/hospitalizations for all respiratory diseases and
asthma in children. Relative risks are standardized per 10-ppb increase in 24-h avg S02 level. The
May 2008
5-7
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
studies are generally presented in the order of increasing width of the 95% CI. For Luginaah et al.
(2005) and Lin et al. (2003), risk estimates for males (*) and females (**) are shown separately.
Overall, the epidemiological evidence for respiratory morbidity is consistent, with
associations reported in studies conducted in numerous locations using a variety of
methodological approaches. In the epidemiological studies that assessed potential confounding
by copollutants using multipollutant models, SO2 effect estimates were generally robust to the
inclusion of copollutants, including PM, O3, CO, and NO2, suggesting that the observed effects
of SO2 on respiratory endpoints occur independent of the effects of other ambient air pollutants.
Intervention studies provide additional evidence that supports a causal relationship
between S02 exposure and respiratory health effects. The proposition that intervention studies
can provide strong support for causal inferences was emphasized by Hill (1965). Two notable
studies conducted in several cities in Germany and in Hong Kong reported that decreases in SO2
concentrations were associated with improvements in respiratory symptoms. In eastern Germany,
a decrease in the prevalence of respiratory symptoms was correlated with a steep decline in
ambient SO2 concentrations of more than 90% from 1992-1993 to 1998-999. During this study
period, decreases in other ambient air pollutants, including -60% lower TSP concentrations, also
occurred in these cities. In Hong Kong, respiratory health improved with similarly large
reductions in S02 of up to 80% in the polluted district but with much smaller reductions in TSP
(less than 20%) compared with those in the cities in eastern Germany. The possibility remains
that these health improvements may be partially attributable to declining concentrations of air
pollutants other than SO2, most notably PM or constituents of PM. Animal toxicological studies
have reported that interactions of S02 and PM may lead to transformation of S02 to other sulfur-
containing compounds which may have more potent biological effects; thus the improvements in
respiratory health may be jointly attributable to declines in both SO2 and PM.
The draft ISA also evaluates the evidence of other health outcomes and exposure durations.
For short-term exposure to S02 and mortality, the evidence was found to be suggestive but not
sufficient to infer a casual relationship. Recent epidemiological studies have consistently
reported positive associations between mortality and SO2, with slightly larger effect estimates
observed for respiratory mortality compared to cardiovascular mortality. However, the SO2 effect
estimates were generally reduced after adjusting for copollutants in the regression models,
indicating some extent of confounding among these pollutants. The evidence between short-term
May 2008
5-8
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
SO2 exposure and cardiovascular effects, and morbidity and mortality with long-term SO2
exposures is inadequate to infer a causal relationship. The key conclusions on the health effects
of S02 exposure are briefly summarized in Section 5.5.
5.3. Interpretation of the Epidemiological Evidence
This section highlights some key considerations for the evaluation of epidemiological
evidence in this draft ISA. As discussed above, clinical studies provide the strongest evidence
that short-term S02 exposure is associated with respiratory morbidity. Numerous
epidemiological studies report associations for a broader range of respiratory health outcomes, at
lower concentrations than the clinical studies, at levels below the current standard. There is,
however, uncertainty about the magnitude of the epidemiological effects estimates. Several
sources of uncertainty and the implications for risk assessment are discussed below.
Although the numerous epidemiological studies provide supportive evidence in making a
causal determination for the effect of SO2 on respiratory health, much uncertainty remains in the
magnitude of the effect estimates related to ambient SO2 exposures. Exposure measurement error
is a key source of this uncertainty as there are questions about the extent to which concentrations
measured by the regulatory ambient monitoring network typically used in epidemiological
studies can accurately represent an individual's exposure to SO2 of ambient origin. Factors
contributing to exposure measurement error include the spatial variation in ambient SO2,
variation in time-activity patterns and the infiltration characteristics of microenvironments, as
well as instrument error in the ambient and personal monitors.
SO2 monitors currently deployed in the regulatory monitoring networks are adequate to
determine compliance with current standards, since both the 24-h avg and annual standards are
substantially above the operating limit of detection of these monitors. However, these monitors
are inadequate for accurate and precise measurements at or near the current ambient mean 24-h
avg SO2 levels of ~4 ppb. Also, typical 24-h avg personal SO2 exposures are often below the
detection limit of commonly deployed passive SO2 monitors. Therefore, the association between
ambient concentrations and personal exposure may be inadequately characterized in recent
studies at lower ambient concentrations.
For community time-series and short-term panel epidemiological studies using daily SO2
concentrations from ambient monitors, these exposure and analytical measurement errors would
May 2008
5-9
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
tend to bias the effect estimate towards the null, leading to uncertainty in accurately quantifying
the magnitude of the effect. In long-term exposure studies, the variable ambient measurement
and exposure error could also result in bias, but the extent and direction of this bias is unclear.
Another factor that contributes to uncertainty in estimating the SCVrelated effect from
epidemiological studies is that SO2 is one component of a complex air pollution mixture
including various other components, e.g., PM and NO2, known to affect respiratory health.
As a consequence of these uncertainties, the epidemiological observations of S02 health
effects can be interpreted in several ways which are not mutually exclusive. First, the reported
SO2 effect estimates in epidemiological studies may reflect independent SO2 effects on
respiratory health. This is supported by evidence from human clinical studies which indicate that
peak exposures (5-10 min) to S02 at levels as low as 0.2-0.3 ppm are capable of eliciting
respiratory responses in some sensitive asthmatics. Because pure SO2 does not appear alone in
real-world ambient conditions but rather is part of a pollutant mixture, it is difficult to relate
these peak exposures in the human clinical studies unequivocally to the 24-h avg SO2
concentrations typically assessed in epidemiological studies. It is possible that higher, shorter-
term concentrations of SO2 may be driving the observed associations in epidemiological studies.
Among the limited number of epidemiological studies evaluating the concentration-response
function, several reported a linear relationship across the entire range of concentrations,
suggesting the lack of a population effects threshold. However, other studies found that a marked
increase in S02-related respiratory health effects was only observed at higher concentrations
(above 90th percentile values).
A second interpretation is that ambient SO2 may be serving as an indicator of complex
ambient air pollution mixtures sharing the same source as SO2 (i.e., combustion of sulfur-
containing fuels or metal smelting). Other components of mixed emissions from these sources
include trace elements such as vanadium, nickel, selenium, and arsenic. Distinguishing effects of
individual pollutants in multipollutant regression models is made difficult by the possibility that
a given air pollutant may be acting as a surrogate for a less-well-measured or unmeasured
pollutant, or that several pollutants may all be acting as surrogates for the same mixtures of
pollutants. Therefore, reported SCVrelated effects may represent those of the overall mixture or
other chemical components within the mixture. However, analysis of ambient data compiled
monthly for the years 2003 to 2005 showed that SO2 concentrations were uncorrected with SO42"
May 2008
5-10
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
in 12 CMS As having multiple monitors. Moreover, in multipollutant models adjusted for PM
indices, SO2 effect estimates were generally found to be robust.
A third interpretation is that in situations of complex pollution mixtures, copollutants may
enhance the toxic capability of SO2 or that SO2 may influence the toxicity of copollutants.
Findings from animal toxicological studies demonstrate that the effects of SO2 may be
exacerbated when aerosol particles act as carriers and deliver sulfur-containing compounds more
effectively to the lower respiratory tract. The synergism observed with combined exposure to
SO2 and PM in the animal toxicological studies provides supportive evidence for the SCVrelated
respiratory effects observed under ambient conditions in the epidemiological studies.
5.4. Susceptible and Vulnerable Populations
Evidence from epidemiological and human clinical studies has indicated that certain
subgroups within the population are more susceptible and/or vulnerable to the effects of SO2
exposure. There is substantial evidence from epidemiological and human clinical studies
indicating that asthmatics are more susceptible to respiratory health effects from SO2 exposures
than the general public. Limited epidemiological evidence further suggests that children and
older adults (> 65 years) are more susceptible to the adverse respiratory effects associated with
ambient SO2 concentrations when compared to the general population. A number of potentially
susceptible groups, including obese individuals, individuals in a chronic pro-inflammatory state
like diabetics, and children born prematurely or with low birth weight (< 2,500 grams), may
experience increased adverse effects associated with exposure to air pollution, but these
relationships have not been examined specifically in relation to SO2. The differential effects of
air pollution among genetically diverse subpopulations have been examined for a number of GST
genes and other genotypes. While limited in number, these studies provide some insight into a
potential genetic role in the determination of susceptibility to air pollution.
Human clinical studies have clearly shown that exercising asthmatics are at greatest risk of
experiencing adverse respiratory effects related to SO2 exposure. Oronasal breathing during
exercise increases vulnerability as it allows a larger fraction of inhaled SO2 to reach the lower
airways. Therefore, individuals with increased vulnerability for S02-related respiratory health
effects include those who spend time outdoors at increased exertion levels, for example children,
outdoor workers, and individuals who exercise or play sports.
May 2008
5-11
DRAFT—DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
5.5. Conclusions
The important findings of this draft ISA on the health effects of SO2 exposure, including
the levels at which effects are observed, are briefly summarized in Table 5-3. Also summarized
are conclusions drawn in the previous review for comparison.
Collectively, the epidemiological, human clinical, and animal toxicological data support
the finding of a causal relationship between short-term exposure to SO2 and respiratory
morbidity. Observed associations between S02 exposure and an array of respiratory outcomes,
including respiratory symptoms, lung function, airway inflammation, airway
hyperresponsiveness, and ED visits and hospitalizations from the human clinical, animal
toxicological, and epidemiological studies, in combination, provide clear and convincing
evidence of consistency, specificity, temporal and biologic gradients, biological plausibility, and
coherence.
Human clinical studies provide strong evidence of respiratory morbidity among asthmatics
following peak exposures (5-10 min) to SO2 concentrations > 0.4 ppm, with some evidence of
respiratory effects at concentrations as low as 0.2 ppm in the most sensitive asthmatics. In the
epidemiological studies, the SCVrelated respiratory effects were consistently observed in areas
where the maximum ambient 24-h avg SO2 concentration was below the current 24-h avg
NAAQS level of 0.14 ppm (Tables 5-4 and 5-5). Potentially susceptible and vulnerable
subgroups include asthmatics, children, older adults, and individuals who spend a lot of time
outdoors at increased exertion levels.
In addition to respiratory morbidity related to short-term exposure to SO2, studies of other
health outcomes and exposure durations were also evaluated in this draft ISA. The evidence is
suggestive but not sufficient to infer a causal relationship between short-term exposure to SO2
and mortality. The evidence linking short-term S02 exposure and cardiovascular effects, and
morbidity and mortality with long-term exposures to SO2 is inadequate to infer a causal
relationship.
May 2008
5-12
DRAFT—DO NOT QUOTE OR CITE
-------
Table 5-3. Key findings on the health effects of S02 exposure
Short-Term Exposure: RESPIRATORY MORBIDITY
Sufficient to infer a causal relationship
RESPIRATORY SYMPTOMS
Previous Conclusion: Among exercising asthmatics, there is a clear, statistically significant increase in
respiratory symptoms following peak exposures (5-10 min) to 0.6-1.0 ppm S02. Significant, but less
severe symptoms are associated with peak SO2 exposures at concentrations of 0.4-0.5 ppm in human
clinical studies.
In the epidemiological studies, an association with aggravation of bronchitis is consistently obsserved
at 24-h avg S02 levels of 0.19 to 0.23 ppm and in some cases at levels below 0.19 ppm.
Current Conclusion. Recent human clinical studies provide additional evidence of respiratory symptoms
in asthmatics following peak exposures (5-10 min) with exercise to 0.5 ppm SO2. Statistically signifi-
cant increases in respiratory symptoms are observed at SO2 concentrations of as low as 0.4 ppm, with
the severity of symptoms shown to increase with increasing concentration between 0.4 and 0.6 ppm.
Epidemiological studies provide consistent evidence of an association between ambient SO2 exposure
and increased respiratory symptoms in children, particularly those with asthma or chronic respiratory
symptoms. Multicity studies have observed these associations at a median range of 17 to 37 ppb
(75th percentile: -25 to 50) across cities for 3-h avg S02 and 2.2 to 7.4 ppb (90th percentile: 4.4 to
14.2) for 24-h avg S02.
In contrast, the epidemiological evidence on the association between SO2 and respiratory symptoms
in adults are generally mixed, with some showing positive associations and others finding no
relationship at current ambient levels.
LUNG FUNCTION
Previous Conclusion: Bronchoconstriction has been found to be the most sensitive indicator of lung
function effects following acute exposure to SO2. Guinea pigs were found to be the most sensitive
species, with bronchoconstriction observed using 0.16 ppm SO2. In human clinical studies, < 10-20%
of exercising asthmatic individuals experience large decrements in lung function (i.e., sRaw increases
> 200% or FEV1 decreases > 20%) following 5-10 min exposures to S02 concentrations of 0.2-0.5
ppm. At 0.6-1.0 ppm SO2, ^ 20-25% of exercising asthmatics are similarly affected.
Small, reversible declines in lung function in children are observed in epidemiological studies at levels
of 0.10 to 0.18 ppm but not at levels of 0.04 to 0.08 ppm.
Current Conclusion: Evaluation of the human clinical evidence focused on moderate or greater
decrements in lung function in exercising asthmatics. SC>2-induced increases in sRaw (> 100%) or de-
creases in FEV1 (> 15%) following 5-10 min exposures are observed in 5-10% of individuals at 0.2
ppm, 10-20% of individuals at 0.3 ppm, and 20-60% of individuals at 0.4-1.0 ppm.
The results are inconsistent for the association between 24-h avg SO2 and lung function in children
and adults in the epidemiological studies.
AIRWAYS INFLAMMATION
Previous Conclusion: No overall conclusion.
Current Conclusion. A limited number of health studies have evaluated the effect of SO2 on airway
inflammation. One human clinical study observed an SC>2-induced increase in sputum eosinophil
counts in exercising asthmatics 2 h after a 10 min exposure to 0.75 ppm S02. The results of this study
provide some evidence that SO2 may elicit an allergic inflammatory response in the airways of
asthmatics which extends beyond the short time period typically associated with SO2 effects.
Animal toxicological studies suggest that repeated exposures to SO2, at concentrations as low as 0.1
ppm in guinea pigs, may exacerbate inflammatory responses in allergic animals.
May 2008
5-13
DRAFT—DO NOT QUOTE OR CITE
-------
AIRWAYS RESPONSIVENESS
Previous Conclusion: No conclusions in the previous review.
Current Conclusion. Animal toxicological evidence suggests that repeated exposures to SO2, at
concentrations as low as 0.1 ppm in guinea pigs, can exacerbate airway responsiveness following
allergic sensitization. In a human clinical study, concurrent exposure (6 h) to 0.2 ppm S02 and 0.4
ppm NO2 has been observed to enhance airway responsiveness to an inhaled antigen among resting
asthmatics. These findings are consistent with the very limited epidemiological evidence that suggests
that exposure to SO2 may lead to airway hyperresponsiveness in atopic individuals.
ED VISITS/HOSPITALIZATIONS
Previous Conclusion: No conclusions in the previous review.
Current Conclusion: Epidemiological studies provide evidence of an association between ambient SO2
concentrations and ED visits and hospitalizations for all respiratory causes, particularly among
children and older adults (age 65+ years), and for asthma. The SO2 effect estimates ranged from a 5%
decreased risk to a 20% excess risk per 10- ppb increase in 24-h avg S02, with the large majority of
studies suggesting an increase in risk. These effects were observed in studies with mean 24-h avg
concentrations as low as 4 ppb, but two studies evaluating the concentration-response function
observed that a marked increase in SC>2-related effects was only observed higher concentrations
(above 90th percentile values).
Short-Term Exposure: CARDIOVASCULAR MORBIDITY
Inadequate to infer the presence or absence of a causal relationship
CARDIOVASCULAR EFFECTS; ED VISITS/ HOSPITALIZATIONS
Previous Conclusion: No conclusions in the previous review.
Current Conclusion.There was some suggestive evidence of an association between 24-h avg SO2
exposure and heart rate variability in the epidemiological studies, but the evidence from two human
clinical studies with were weak and inconsistent. Some epidemiological studies have observed
positive associations between ambient SO2 concentrations and ED visits or hospital admissions for
cardiovascular diseases, but results are not consistent across studies and the SO2 effect estimate was
generally not robust to copollutant adjustment.
Short-Term Exposure: MORTALITY
Suggestive but not sufficient to infer a casual relationship
NONACCIDENTAL AND CARDIOPULMONARY MORTALITY
Previous Conclusion: Epidemiological studies based on historical air pollution episodes observed the
clearest mortality associations when both black smoke (BS) and SO2 concentrations were at high
levels (24-h avg values exceeding 1,000 |jg/m3 [-400 ppb for SO2]). Later studies observed that an in-
creased risk of mortality was associated with exposure to BS and SO2 levels in the range 0.19 to
0.38 ppm. Because of the high colinearity between BS and SO2 levels, it is difficult to readily separate
the effects of these pollutants on mortality.
Current Conclusion: Recent epidemiological studies have consistently reported positive associations
between mortality and SO2, often at mean 24-h avg levels < 10 ppb. The range of SO2 excess risk
estimates for nonaccidental mortality is 0.4 to 2% per 10 ppb increase in 24-h avg SO2 in several
multicity studies and meta-analyses. SO2 effect estimates for respiratory mortality were generally
larger than the cardiovascular mortality estimates, suggesting a stronger association of SO2 with
respiratory mortality compared to cardiovascular mortality. The S02 effect estimates were generally
reduced when copollutants were added in the model, indicating some extent of confounding among
these pollutants.
May 2008
5-14
DRAFT—DO NOT QUOTE OR CITE
-------
Long-Term Exposure: RESPIRATORY MORBIDITY
Inadequate to infer the presence or absence of a causal relationship
RESPIRATORY SYMPTOMS AND LUNG FUNCTION
Previous Conclusion: The limited available epidemiological data indicated associations between
respiratory illnesses and symptoms and persistent exposures to SO2 in areas with long-term averages
exceeding 0.04 ppm.
Current Conclusion: Several epidemiological studies that examined the effects of long-term exposure to
SO2 on asthma, bronchitis, and respiratory symptoms observed positive associations in children.
While the evidence is suggestive, the variety of outcomes examined and the inconsistencies in the
observed results make it difficult to assess the direct impact of long-term exposure of SO2 on
respiratory symptoms. The epidemiological and animal toxicological evidence generally do not indicate
that long-term exposure to SO2 has a detrimental effect on lung function.
Long-Term Exposure: OTHER MORBIDITY
Inadequate to infer the presence or absence of a causal relationship
CARCINOGENIC EFFECTS
Previous Conclusion: Epidemiological evidence did not substantiate the hypothesized links between SO2
or other SOx and cancer, though there was some animal toxicological evidence that led to the
conclusion that SO2 may be considered a suspect carcinogen/cocarcinogen.
Current Conclusion: Animal toxicological studies indicate that SO2 at high concentrations may cause
DNA damage but fails to induce carcinogenesis, cocarcinogenesis, or tumor promotion. Epide-
miological studies did not provide evidence that long-term exposure to SO2 is associated with an
increased incidence of or mortality from lung cancer.
PRENATAL AND NEONATAL OUTCOMES
Previous Conclusion: No conclusions in the previous review.
Current Conclusion: Epidemiological studies on birth outcomes have found suggestive positive
associations between SO2 exposure and low birth weight. However, the inconsistent results across tri-
mesters of pregnancy and the lack of evidence to evaluate confounding by copollutants limit the
interpretation of these studies.
Long-Term Exposure: MORTALITY
Inadequate to infer the presence or absence of a causal relationship
NONACCIDENTAL AND CARDIOPULMONARY MORTALITY
Previous Conclusion: The available studies on the effects of long-term exposure to S02 on mortality
were all ecological cross-sectional studies which did not take into consideration potential confounders.
In addition, it was concluded that effects from PM and SO2 could not be distinguished in these studies.
Current Conclusion: Two major U.S. epidemiological studies observed associations between long-term
exposure to SO2 and mortality, but several other U.S. and European cohort studies did not observe an
association. The relative risks ranged from 0.97 to 1.07 per 5-ppb increase in the long-term average
SO2. Evaluation of these studies is further limited by the inability to distinguish potential confounding
by copollutants and uncertainties regarding the geographic scale of analysis.
May 2008
5-15
DRAFT—DO NOT QUOTE OR CITE
-------
Table 5-4. Effects of short-term exposure to S02 on respiratory symptoms among children.
STUDY
POPULATION
MEAN CONCENTRATION
S02
(PPb)
98TH%
so2
(PPb)
99TH%
so2
(PPb)
RANGE
S02
(PPb)
UPPER
%TILE
STANDARDIZED
ODDS RATIO
(95% Cl)a
United States
Schildcrout et al. (2006)
Multicity, North America
Seattle, WA; Baltimore, MD;
St. Louis, MS (Nov 1993-
Aug 1995); Denver, CO;
San Diego, CA (Nov 1993-
Jul 1995); Toronto, ON (Dec
1993-Jul 1995); Boston,
MA (Jan 1994-Sep 1995)
No SO2 data available in
Albuquerque, NM
Asthmatic
children
(n = 990)
24-h avg:
2.2-7.4
(range of city-specific medians)
NR
NR
NR
75th: 3.1,
10.7
90th: 4.4,
14.2
(range in
city
specific
estimates)
Asthma symptoms:
S02 alone: 1.04 (1.00,
1.08), 3-day sum
S02& N02: 1.04 (1.00,
1.09), 3-day sum
S02& PM10: 1.04 (0.99,
1.08), 3-day sum
Schwartz et al. (1994)
Multicity, United States
Watertown, MA (Apr-Aug
1985); Kingston-Harriman,
TN; St. Louis, MO (Apr-Aug
1986); Steubenville, OH;
Portage, Wl (Apr-Aug
1987); Topeka, KS (Apr-
Aug 1988)
Children in
grades 2-5
(n = 1,844)
24-h avg:
4.1 (median)
NR
NR
NR
75th: 8.2
90th: 17.9
Max: 81.9
Cough incidence:
S02 alone: 1.15(1.02-
1.31), 4-day avg
S02, adjusting for PMi0:
1.08 (0.93, 1.25), 4-day
avg
S02, adjusting for N02:
1.09 (0.94, 1.30), 4-day
avg
Neas et al. (1995)
Uniontown, PA
Summer 1990
Children in
grades 4-5
(n = 83)
12-h avg: 0.2
5.9 overnight
14.5 daytime
NR
NR
IQR:
11.1
Max: 44.9
Evening cough:
1.19 (1.00, 1.42), lag 12-h
Mortimer et al. (2002)
Multicity, United States
Bronx, NY; East Harlem,
NY; Baltimore, MD;
Washington, DC; Detroit,
Ml; Cleveland, OH;
Chicago, IL; St. Louis, MO
(Jun-Aug 1993)
Asthmatic
children, aged
4-9
(n = 846)
3-h avg:
22 (shown in figure)
NR
NR
0-75 ppb
(shown
in graph)
NR
Asthma symptoms:
S02 alone (8 cities):
1.19 (1.06, 1.35), lag 1-2
S02, adjusting for 03 &
N02 (7 cities): 1.19(1.04,
1.37), lag 1-2
S02, adjusting for03, N02
& PM10 (3 cities):1.23
(0.94, 1.62), lag 1-2
Europe
Timonen and Pekkanen
(1997)
Kuopio (urban and
suburban) Finland
Winter 1994
Children
7-12 yrs with
asthma or
cough
symptoms
(n = 169)
24-h avg: 2.3
NR
NR
NR
75th: 2.7
Max: 12.3
Upper respiratory
symptoms:
2.71 (1.19, 6.17), lag 0
3.17 (1.21, 8.78), lag 1
Ward et al. (2002)
Birmingham and Sandwell,
UK
Jan-Mar 1997
May-Jul 1997
Children, age at
enrollment 9 yrs
(n = 162)
24-h avg:
Median
5.4, Winter
4.7, Summer
NR
NR
2, 18
Winter
2, 10
Summer
NR
Cough:
Winter: 0.59 (0.25, 1.40),
Summer: 0.90 (0.49, 1.66)
Shortness of breath:
Winter: 0.59 (0.32, 1.09),
Summer: 0.81 (0.30, 2.17)
Wheeze:
Winter: 0.79 (0.38, 1.63),
Summer: 0.77 (0.28, 2.08)
(7-day avg lag for above)
May 2008
5-16
DRAFT—DO NOT QUOTE OR CITE
-------
STUDY
POPULATION
MEAN CONCENTRATION
S02
(PPb)
98TH%
so2
(PPb)
99TH%
so2
(PPb)
RANGE
so2
(PPb)
UPPER
%TILE
STANDARDIZED
ODDS RATIO
(95% Cl)a
Segala et al. (1998)
Paris, France
Nov 1992-May 1993
Children
7-15 yrs with
physician-
diagnosed
asthma
(n = 84)
24-h avg:
8.3 (5.2)
NR
NR
1.7-32.2
NR
Prevalent asthma:
1.32 (1.08, 1.62), lag 0
1.26 (0.93, 1.71), lag 1
Prevalent shortness of
breath:
1.17 (0.53, 2.62), lag 0
1.21 (0.99, 1.49) lag 1
Incident asthma
1.73 (1.15, 2.60), lag 0
1.60 (1.01, 2.53), lag 1
Incident wheeze
1.22 (0.95, 1.58), lag 0
1.13 (0.68, 1.88), lag 1
Boezen et al. (1998)
Amsterdam and Meppel
(urban and rural),
the Netherlands
Winter 1993-1994
Children 7-11
yrs, with and
w/o BHR and
high serum
concentrations
of total IgE
(n = 632)
24-h avg: Means: 1.7, 8.7;
Medians: 1.4, 8.3 (range in city-
specific estimates)
NR
NR
1.9, 23.6
NR
Among children with BHR
and relatively high serum
total IgE - lower
respiratory symptoms:
1.27 (1.09, 1.49), lag 0
1.25 (1.06, 1.48), lag 1
1.69 (1.26, 2.28), 5-day
avg
Roemer et al. (1993)
Wageningen, the
Netherlands
Winter 1990-1991
Children 6-12
yrs with chronic
respiratory
conditions
(n = 73)
24-h avg
1-h max
NR
NR
0, 40.4
(24-h
avg)
Max: 56.5
(1-h max)
Asthma attack: 1.79 (1.35,
2.38). 7-day avg Wheeze:
1.97 (1.42, 2.72), 7-day
avg
Waken with symptoms:
1.79 (1.12, 2.87), 7-day
avg
Shortness of breath: 1.48
(1.06, 2.07), 7-day avg
Cough: 1.97 (1.03, 3.77),
7-day avg
Hoek and Brunekreff (1993)
Wageningen, the
Netherlands
Winter 1991
Children 7-11
yrs, nonurban
area
(n = 112)
24-h avg
NR
NR
NR
Max:
40.4
Cough:
1.22 (0.20, 7.39), lag 0
0.25 (0.04, 1.65), lag 1
3.67 (0.002, 7.331.974),
7-day avg
Lower resp symptoms:
1.82 (0.14, 24.3), lag 0
0.33 (0.02, 6.05), lag 1
0.005 (0.0, 44.7), 5-day
avg
May 2008
5-17
DRAFT—DO NOT QUOTE OR CITE
-------
STUDY
POPULATION
MEAN CONCENTRATION
S02
(PPb)
98TH%
so2
(PPb)
99TH%
so2
(PPb)
RANGE
so2
(PPb)
UPPER
%TILE
STANDARDIZED
ODDS RATIO
(95% Cl)a
Van derZee et al. (1999)
Urban and nonurban areas
The Netherlands
3 winters, 1992-1995
Children 7-11
yrs, with and
without chronic
respiratory
symptoms
(n = 633)
24-h avg:
1.4,8.8
(range in city-specific medians)
NR
NR
NR
Max: 6.5,
58.5
(range in
city-
specific
maxi-
mums)
Lower respiratory symp-
toms, urban, S02 alone:
1.22 (1.01, 1.46), lag 0
1.14 (0.95, 1.38), lag 1
1.34 (0.98, 1.82), 5-day
avg
Lower respiratory
symptoms, urban, S02,
adjusting for PM10:
1.18 (0.96, 1.45), lag 0
1.03 (0.83, 1.27), lag 1
1.08 (0.72, 1.63), 5-day
avg
Lower respiratory
symptoms, nonurban:
0.94 (0.79, 1.12), lag 0
0.94 (0.78, 1.13), lag 1
1.10 (0.75, 1.63), 5-day
avg
Cough, urban:
0.93 (0.84, 1.03), lag 0
1.08 (0.98, 1.19), lag 1
1.08 (0.89, 1.30) 5-day
avg
Cough, nonurban:
1.05 (0.96, 1.15), lag 0
0.98 (0.90, 1.08), lag 1
1.04 (0.83, 1.30), 5-day
avg
a24-h avg S02 and 12-h avg S02 standardized to 10- ppb incremental change; 3-h avg S02 standardized to 20-ppb incremental change; and 1-h max S02 standardized to
40-ppb incremental change. NR = Not Reported BHR = Bronchial Hyperresponsiveness
Table 5-5. Effects of short-term S02 exposure on emergency department visits and hospital
admissions for respiratory outcomes.
STUDY
POPULATION
MEAN
CONCENTRATION
so2
(PPb)
98TH%
so2
(PPb)
99TH%
S02 (ppb)
RANGE
so2
(PPb)
UPPER
%TILE
STANDARDIZED
ODDS RATIO
(95% Cl)a
EMERGENCY DEPARTMENT VISITS - ALL RESPIRATORY
UNITED STATES
Wilson et al. (2005)
Portland, ME
Jan 1998-Dec 2000
Manchester, NH
Jan 1996-Dec 2000
= 84,000 ED visits
1-h max:
Portland:
11.1 (9.1)
Manchester:
16.5 (14.7)
NR
NR
NR
NR
Portland:
All ages: 8% (3, 11)
0-14: -2.6% (-10.3,
2.7)
15-64: 11% (5.4,
13.9)
65+: 16.8% (8.2,
25.8)
Manchester:
All ages: 0% (-3, 5)
0-14: 0% (8, 8)
15-64: 0% (-3, 5)
65+: 8% (6, 23)
Tolbert et al. (2007)
Atlanta, GA
Jan 1993-Dec 2004
> 1,000,000 ED visits for
all respiratory causes
1-h max:
14.9
NR
NR
1.0-149.0
75th: 20.0
90th: 35.0
0.75% (-0.75, 2.3)
May 2008
5-18
DRAFT—DO NOT QUOTE OR CITE
-------
STUDY
POPULATION
MEAN
CONCENTRATION
so2
(PPb)
98TH%
so2
(PPb)
99TH%
S02 (ppb)
RANGE
so2
(PPb)
UPPER
%TILE
STANDARDIZED
ODDS RATIO
(95% Cl)a
Peel et al. (2005)
Atlanta, GA
Jan 1993-Aug 2000
484,830 ED visits, all
ages from 31 hospitals
1-h max:
16.5 (17.1)
NR
NR
NR
90th: 39.0
1.6% (-0.6, 3.8)
EUROPE
Atkinson et al. (1999a)
London, UK
Jan 1992-Dec 1994
98,685 ED visits from 12
hospitals
24-h avg:
8.0 (2.9)
NR
NR
2.8, 30.9
50th: 7.4
90th: 11.7
All Ages:
4.2% (1.1, 7.4)
0-14: 9.0% (4.4, 13.8)
15-64:4.0% (-0.3,
8.5)
65+: -2.7% (-5.4, 3.3)
EMERGENCY DEPARTMENT VISITS - ASTHMA
UNITED STATES
Ito et al. (2007)
New York, NY
Jan 1999-Dec 2002
Asthma ED visits, all
ages from 11 hospitals
24-h avg:
7.8 (4.6)
NR
NR
NR
75th: 10
95th: 17
35% (23%, 51%)
NY Department of Health
(2006)
Bronx & Manhattan, NY
Jan 1999-Dec 2000
Asthma ED visits among
children from 22
hospitals
24-h avg :
11 (7.2)
NR
NR
NR
NR
5-day moving
average:
Manhattan: -1% (-12,
12)
Bronx: 11% (6, 17)
Jaffe et al. (2003)
Cincinnati, OH
Cleveland, OH
Columbus, OH
Jul 1991-Jun 1996
4,416 ED visits for
asthma, age 5-34
24-h avg:
Cincinnati:
13.5 (9.4)
Cleveland:
14.7 (9.5)
Columbus:
4.2 (3.2)
NR
NR
Cincinnati:
0.6, 49.6
Cleveland:
0.98, 62.8
Columbus
0, 21.4
NR
Cincinnati: 17.3%
(4.7, 30.8)
Cleveland: 3.1% (-
3.8, 10.7)
Columbus: 13.1% (-
14.2, 48.6)
All Cities: 6.2% (0.5,
11.6)
Wilson et al. (2005)
Portland, ME
Jan 1998-Dec 2000
Manchester, NH
Jan 1996-Dec 2000
= 84,000 ED visits
1-h max:
Portland:
11.1 (9.1)
Manchester:
16.5 (14.7)
NR
NR
NR
NR
Portland:
All ages:
11.0% (0.0, 19.7)
0-14: 5.4% (-12.8,
25.8)
15-64: 11% (0, 22.7)
65+: 11.0% (-15.2,
48.4)
Manchester:
All ages:
5.4% (-2.6, 16.8)
0-14: 19.7% (-2.6,
51.8)
15-64: 2.7% (-7.8,
13.9)
65+: 11.0% (-28.8,
77.2)
Peel et al. (2005)
Atlanta, GA
Jan 1993-Aug 2000
Asthma ED visits, all
ages and 2-18 yrs from
31 hospitals
1-h max: 16.5 (17.1)
NR
NR
NR
90th: 39.0
0.2% (-3.2, 3.4)
May 2008
5-19
DRAFT—DO NOT QUOTE OR CITE
-------
STUDY
POPULATION
MEAN
CONCENTRATION
so2
(PPb)
98TH%
so2
(PPb)
99TH%
S02 (ppb)
RANGE
so2
(PPb)
UPPER
%TILE
STANDARDIZED
ODDS RATIO
(95% Cl)a
EUROPE
Atkinson et al. (1999a)
London, UK
Jan 1992-Dec 1994
98,685 ED visits from 12
hospitals
24-h avg:
8.0 (2.9)
NR
NR
2.8, 30.9
50th: 7.4
90th: 11.6
All ages: 7.4% (2.3,
12.8)
0-14: 15.0% (7.1,
23.5)
15-64: 6.3% (-0.8,
13.8)
Hajat et al. (1999)
London, UK
Jan 1992-Dec 1994
General practitioner
visits for asthma
24-h avg:
All yr: 8.0 (2.9)
Warm: 7.7(2.4)
Cool: 8.3 (3.4)
NR
NR
NR
All yr:
90th: 11.6
Warm:
90th: 10.7
Cool:
90th: 12.4
All ages: 6.6% (1.3,
11.9)
0-14: 6.6% (-1.0,
14.7)
15-64: 5.2% (-1.5,
12.3)
65+: 7.2% (-4.3,
20.1)
Boutin-Forzano et al.
(2004)
Marseille, France
Apr 1997-Mar 1998
549 ED visits for asthma
24-h avg: 8.5
NR
NR
0.0, 35.3
NR
3-49 yrs: 0.6% (-1.4,
2.7)
Galan et al. (2003)
Madrid, Spain
Jan 1995-Dec 1998
4,827 ED visits for
asthma
24-h avg: 8.9 (5.8)
NR
NR
1.9, 45.6
50th: 7.0
75th: 11.8
90th: 16.5
All ages:
4.9% (-4.2, 15.0)
Tenias et al. (1998)
Valencia, Spain
Jan 1993-Dec 1995
734 ED visits for asthma
24-h avg: 10.0
Cold: 11.9
Warm: 8.2
1-h max: 21.2
Cold: 24.3
Warm: 18.1
NR
NR
NR
24-h avg:
50th: 9.8
75th: 12.9
95th: 16.0
1-h max:
50th: 19.6
75th: 27.1
95th: 35.8
> 14 yrs:
13.9% (-7.0, 39.4)
Sunyer et al. (1997)
Multicity, Europe
Barcelona, Spain;
Helsinki, Finland; Paris,
France; London, UK
Jan 1986-Dec 1992
All ED visits for asthma
24-h avg:
Barcelona: 15.4
Helsinki: 6.0
London: 11.6
Paris: 8.6
NR
NR
Barcelona:
0.8, 60.2
Helsinki:
1.1, 35.7
London:
3.4, 37.6
Paris:
0.4, 82.3
NR
0-14 yrs: 3.2% (-0.2,
6.8)
15-64: 0.2% (-2.2,
2.6)
Castellsague et al. (1995)
Barcelona, Spain
Jan 1986-Dec 1989
ED visits for asthma
from 4 hospitals
24-h avg:
Summer: 15.3|
Winter: 19.5
NR
NR
NR
Summer:
50th: 13.5
75th: 20.3
95th: 30.8
Winter:
50th: 18.4
75th: 25.2
95th: 35.3
15-64 yrs, summer:
5.5% (-2.1, 13.8)
15-64 yrs, winter:
2.1% (-4.2, 9.0)
HOSPITAL ADMISSIONS - ALL RESPIRATORY
UNITED STATES
May 2008
5-20
DRAFT—DO NOT QUOTE OR CITE
-------
STUDY
POPULATION
MEAN
CONCENTRATION
so2
(PPb)
98TH%
so2
(PPb)
99TH%
S02 (ppb)
RANGE
so2
(PPb)
UPPER
%TILE
STANDARDIZED
ODDS RATIO
(95% Cl)a
Schwartz (1995)
New Haven, CT
Tacoma, WA
Jan 1988-Dec 1990
= 13,470 Hospital
admissions, ages 65+
24-h avg:
New Haven: 29.8
Tacoma: 16.8
NR
NR
NR
New
Haven:
75th: 37.6
90th: 59.8
Tacoma:
75th: 21.1
90th: 27.8
New Haven: 1.6%
(1.1, 2.6)
Tacoma: 3.2%
(0.5, 6.2)
CANADA
Fung et al. (2006b)
Vancouver, BC
Jun1995-Mar1999
= 41,000 respiratory
admissions for elderly
(65+ yrs)
24-h avg:
3.46 (1.82)
NR
NR
0.0, 12.5
NR
12.6% (4.1, 22.0)
Cakmak et al. (2006)
Multicity, Canada
Calgary, Edmonton,
Halifax, London, Ottawa,
Saint John, Toronto,
Vancouver, Windsor,
Winnipeg
Jan 1993-Dec 2000
> 200,000 hospital
admissions for all
respiratory causes
24-h avg: 4.6
NR
NR
2.8, 10.2
NR
2.4% (1.1, 3.9)
Yang et al. (2003b)
Vancouver, BC
Jan 1986-Dec 1998
Respiratory hospital
admissions among
young children (< 3 yrs)
and elderly (>65 yrs)
24-h avg:
4.84 (2.84)
NR
NR
NR
75th: 6.25
100th:
24.00
< 3 yrs: 3% (-6, 15)
65+ yrs: 5.8% (0.0,
11.9)
*Burnett et al. (2001)
Toronto, ON
Jan 1980-Dec 1994
All respiratory
admissions for young
children (< 2 yrs)
1-h max: 11.8
NR
55
NR
75th: 15
95th: 32
100th:
110
11% (-0.3, 23.6)
Luginaah et al. (2005)
Windsor, ON
Apr 1995-Dec 2000
All respiratory
admissions ages
0-14, 15-64, and 65+
from 4 hospitals
1-h max:
27.5 (16.5)
NR
NR
0, 129
NR
All ages, female:
2.1% (-0.7, 5.0)
All ages, male:
2.5% (-5.3, 0.5)
0-14, female: 5.6%
(0.6, 10.9)
0-14, male: -2.5%
(-6.8, 1.9)
15-64, female:
1.6% (-3.7, 7.2)
15-64, male:
4.5% (-8.4, 5.8)
65+, female: 1.5%
(-2.6, 5.8)
65+, male: -3.1%
(-7.5, 1.5)
May 2008
5-21
DRAFT—DO NOT QUOTE OR CITE
-------
STUDY
POPULATION
MEAN
CONCENTRATION
so2
(PPb)
98TH%
so2
(PPb)
99TH%
S02 (ppb)
RANGE
so2
(PPb)
UPPER
%TILE
STANDARDIZED
ODDS RATIO
(95% Cl)a
AUSTRALIA
Barnett et al. (2005)
Multicity, Australia/New
Zealand
Auckland, Brisbane,
Canberra, Christchurch,
Melbourne, Perth,
Sydney
Jan 1998-Dec 2001
All respiratory hospital
admissions
24-h avg:
Auckland: 4.3
Brisbane: 1.8
Christchurch: 2.8
Sydney: 0.9
1-h max: Brisbane: 7.6
Christchurch: 10.1
Sydney: 3.7
NA in Auckland,
Canberra, Melbourne,
and Perth
NR
NR
24-h avg:
Auckland:
0, 24.3
Brisbane:
0, 8.2
Christchurch:
0, 11.9
Sydney:
0, 3.9
1-h max
Brisbane:
0, 46.5
Christchurch:
0.1, 42.1
Sydney:
0.1, 20.2
1-4 yrs: 5.1% (0.0,
9.1)
5-14: 3.7% (-9.9,
19.5)
Petroeschevsky et al.
(2001)
Brisbane, Australia
Jan 1987-Dec 1994
33,710 hospital
admissions
24-h avg: 4.1
1-h max: 9.2
NR
NR
NR
NR
All ages: -5.9% (-
12.4, 1.1)
0-14: 8.0% (-2.9,
20.1)
15-64: -21.6% (-34.4,
-6.2)
EUROPE
Oftedal et al. (2003)
Drammen, Norway
Jan 1994-Dec 2000
All respiratory hospital
admissions
24-h avg:
1.1 (0.8)
NR
NR
NR
NR
All ages:
71.8% (15.5, 152.7)
Fusco et al. (2001)
Rome, Italy
Jan 1995-Oct 1997
All respiratory hospital
admissions
24-h avg:
3.4 (2.2)
NR
NR
NR
50th: 3.0
75th: 4.5
All age: 1.6% (-4.9,
8.8)
0-14: -2.7% (-4.6,
10.8)
Llorca et al. (2005)
Torrelavega, Spain
Jan 1992-Dec 1995
Hospital admissions
from one hospital
24-h avg:
5.0 (6.3)
NR
NR
NR
NR
All ages: 1.0% (-2.8,
4.7)
Anderson et al. (2001)
West Midlands
conurbation, UK
Oct 1994-Dec 1996
Hospital admissions
stratified by age
24-h avg:
7.2 (4.7)
NR
NR
1.9, 59.8
90th: 12.3
All ages: 1.4% (-0.8,
3.8)
0-14: 5.1% (1.6, 8.7)
15-64: -1.0% (-5.3,
3.7)
65+: -2.2% (-5.4, 1.2)
Atkinson et al. (1999a)
London, UK
Jan 1992-Dec 1994
165,032 hospital
admissions
24-h avg:
8.0 (2.9)
NR
NR
2.8, 30.9
50th: 7.4
90th: 11.7
All ages: 3.0% (0.4,
5.6)
0-14: 7.7% (3.8, 11.7)
15-64: 2.8% (-1.2,
7.0)
65+: 3.3% (-0.1, 6.9)
Schouten et al. (1996)
Multicity, The Netherlands
Amsterdam, Rotterdam
Apr 1977-Sep 1989
All respiratory hospital
admissions
24-h avg:
Amsterdam: 10.5
Rotterdam: 15.0
1-h max:
Amsterdam: 24.4
Rotterdam: 37.2
NR
NR
NR
NR
Amsterdam:
15-64: -2.3% (-5.5,
0.9)
65+: 0.2% (-2.8, 3.3)
Rotterdam:
15-64: -2.9% (-6.2,
0.5)
May 2008
5-22
DRAFT—DO NOT QUOTE OR CITE
-------
STUDY
POPULATION
MEAN
CONCENTRATION
so2
(PPb)
98TH%
so2
(PPb)
99TH%
S02 (ppb)
RANGE
so2
(PPb)
UPPER
%TILE
STANDARDIZED
ODDS RATIO
(95% Cl)a
Spix et al. (1998)
Multicity, Europe
London, UK; Amsterdam
& Rotterdam, the
Netherlands; Paris,
France; Milan, Italy
Jan 1977-Dec 1991
All respiratory hospital
admissions
24-h avg:
London: 10.9
Amsterdam: 7.9
Rotterdam: 9.4
Paris: 8.6
Milan: 24.8
NR
NR
NR
NR
15-64 yrs: 0.5% (-0.4,
1.3)
65+: 1.1% (0.3, 2.4)
Dab et al. (1996)
Paris, France
Jan 1987-Sep 1992
Hospital admissions
from 27 hospitals
All yr:
24-h avg: 11.2
1-h max: 22.5
Warm season
24-h avg: 7.6
1-h max: 16.1
Cold season
24-h avg: 15.1
1-h max: 29.4
NR
NR
NR
All yr:
24-h avg:
99th: 50.0
1-h max:
99th: 87.5
Warm
season
99th: 18.5
1-h max:
99th: 50.3
Cold
season
24-h avg:
99th: 56.0
1-h max:
99th:
100.9
All ages: 1.1% (0.1,
2.1)
a24-h avg S02 and 12-h avg S02 standardized to 10- ppb incremental change; 3-h avg S02 standardized to 20-ppb incremental change; and 1-h max S02 standardized to
40-ppb incremental change.
May 2008
5-23
DRAFT—DO NOT QUOTE OR CITE
-------
References
Abbey DE, Nishino N, McDonnell WF, Burchette RJ, 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.
Ackermann-Liebrich U, Leuenberger P, Schwartz J, Schindler C, Monn C, Bolognini B, Bongard JP,
Brandli O, Domenighetti G, Elsasser S et al. 1997. Lung function and long term exposure to air
pollutants in Switzerland. Am. J. Respir. Crit. Care Med. 155:122-129.
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.
Aharonson EF, Menkes H, Gurtner G, Swift DL, Proctor DF. 1974. Effect of respiratory airflow rate on
removal of soluble vapors by the nose. J Appl Physiol 37:654-7.
Amdur MO, Chen LC, Guty J, Lam HF, Miller PD. 1988. Speciation and pulmonary effects of acidic SOx
formed on the surface of ultrafine zinc oxide aerosols. Atmospheric Environment 22:557-560.
Amdur MO, McCarthy JF, Gill MW. 1983. Effect of mixing conditions on irritant potency of zinc oxide and
sulfur dioxide. Am Ind Hyg Assoc J 44:7-13.
Amdur MO. 1966. Respiratory absorption data and S02 dose-response curves. Arch Environ Health
12:729-32.
American Academy of Pediatrics, Committee on Environmental Health. 2004. Ambient air pollution: health
hazards to children. Pediatrics 114:1699-1707.
American Thoracic Society. 1985. Guidelines as to what constitutes an adverse respiratory health effect,
with special reference to epidemiologic studies of air pollution. Am. Rev. Respir. Dis. 131: 666-
668.
American Thoracic Society. 2000). What constitutes an adverse health effect of air pollution? Am. J.
Respir. Crit. Care Med. 161: 665-673.
Andersen IB, Lundqvist GR, Jensen PL, Proctor DF. 1974. Human response to controlled levels of sulfur
dioxide. Arch Environ Health 28:31-9.
Anderson HR, Bremner SA, Atkinson RW, Harrison RM, Walters S. 2001. Particulate matter and daily
mortality and hospital admissions in the west midlands conurbation of the United Kingdom:
associations with fine and coarse particles, black smoke and sulphate. Occup. Environ. Med.
58:504-510.
Atkinson R. 1991. Kinetics and mechanisms of the gas-phase reactions of the N03 radical with organic
compounds. Journal of Physical and Chemical Reference Data 20:459-507.
Atkinson RW, Anderson HR, Strachan 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 13:257-65.
Atkinson RW, Bremner SA, Anderson HR, Strachan DP, Bland JM, Ponce de Leon A. 1999a. Short-term
associations between emergency hospital admissions for respiratory and cardiovascular disease
and outdoor air pollution in London. Arch. Environ. Health 54:398-411.
Atzori L, Bannenberg G, Corriga A, Lou Y, Lundberg J, RyrfeldtA, Moldeus P. 1992a. Sulfur dioxide-
induced bronchoconstriction via ruthenium red-sensitive activation of sensory nerves. Respiration
59:272-278.
Atzori L, Bannenberg G, Corriga A, Moldeus P, RyrfeldtA. 1992b. Sulfur dioxide-induced
bronchoconstriction in the isolated perfused and ventilated guinea-pig lung. Respiration 59:16-21.
May 2008
R-1
DRAFT—DO NOT QUOTE OR CITE
-------
Ballester F, Rodriguez P, Iniguez C, Saez M, Daponte A, Galan I, Taracido M, Arribas F, Bellido J, Cirarda
FB et al. 2006. Air pollution and cardiovascular admisisons association in Spain: results within the
EMECAS project. Journal of epidemiology and community health 60:328-336.
Balmes JR, Fine JM, Sheppard D. 1987. Symptomatic bronchoconstriction after short-term inhalation of
sulfur dioxide. Am Rev Respir Dis 136:1117-21.
BarnettAG, Williams GM, Schwartz J, NellerAH, BestTL, Petroeschevsky AL, Simpson RW. 2005. Air
pollution and child respiratory health: a case-crossover study in Australia and New Zealand. Am J
Respir Crit Care Med 171:1272-8.
Barthelemy P, Badier M, Jammes Y. 1988. Interaction between S02 and cold-induced bronchospasm in
anesthetized rabbits. Respir Physiol 71:1-10.
Bascom R, Kesavanathan J, Permutt T, Fitzgerald TK, Sauder L, Swift DL. 1996. Tobacco smoke upper
respiratory response relationships in healthy nonsmokers. Fundam Appl Toxicol 29:86-93.
Baskurt OK. 1988. Acute hematologic and hemorheologic effects of sulfur dioxide inhalation. Arch Environ
Health 43:344-8.
Bateson TF, Schwartz J. 2004. Who is sensitive to the effects of particulate air pollution on mortality? A
case-crossover analysis of effect modifiers. Epidemiology 15:143-9.
Bedi JF, Folinsbee LJ, Horvath SM, Ebenstein RS. 1979. Human exposure to sulfur dioxide and ozone:
absence of a synergistic effect. Arch Environ Health 34:233-9.
Beelen R, Hoek G, van den Brandt PA, Goldbohm RA, Fischer P, Schouten LJ, Jerrett M, Hughes E,
Armstrong B, Brunekreef B. 2008. Long-term effects of traffic-related air pollution on mortality in a
Dutch cohort (NLCS-AIR Study). Environ Health Perspect 116:196-202.
Bell ML, Ebisu K, Belanger K. 2007. Ambient air pollution and low birth weight in Connecticut and
Massachusetts. Environ Health Perspect 115:1118-24.
Ben-Jebria A, FullAP, DeMaria DD, Ball BA, Ultman JS. 1990. Dynamics of sulfur dioxide absorption in
excised porcine tracheae. Environmental Research. 53:119-34.
Berkowitz GS, Papiernik E. 1993. Epidemiology of preterm birth. Epidemiol Rev 15:414-43.
Berresheim H, Eisele FL, Tanner DJ, Mclnnes LM, Ramsey-Bell DC, Covert DS. 1993. Atmospheric sulfur
chemistry and cloud condensation nuclei (CCN) concentrations over the northeastern pacific
coast. J Geophys Res 98: 12701-12711.
Berresheim H; Wine PH; Davis DD. 1995 Sulfur in the atmosphere. In: Singh, H. B., ed. Composition,
chemistry, and climate of the atmosphere. New York, NY: Van Nostrand Reinhold; pp. 251-307.
Bethel RA, Epstein J, Sheppard D, Nadel JA, Boushey HA. 1983. Sulfur dioxide-induced
bronchoconstriction in freely breathing, exercising, asthmatic subjects. Am Rev Respir Dis
128:987-90.
Bethel RA, Sheppard D, Geffroy B, Tam E, Nadel JA, Boushey HA. 1985. Effect of 0.25 ppm sulfur
dioxide on airway resistance in freely breathing, heavily exercising, asthmatic subjects. Am Rev
Respir Dis 131:659-61.
Biersteker K, de Graaf H, Nass CAG. 1965. Indoor air pollution in Rotterdam homes. Int J Air Water Pollut
9:343-350.
Bigby B, Boushey H. 1993. Effects of nedocromil sodium on the bronchomotor response to sulfur dioxide
in asthmatic patients. J Allergy Clin Immunol 92:195-7.
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 of Occup Environ Health 11:107-122.
Bobak M. 2000. Outdoor air pollution, low birth weight, and prematurity. Environ Health Perspect 108:173-
6.
May 2008
R-2
DRAFT—DO NOT QUOTE OR CITE
-------
Boezen HM, van der Zee SC, Postma DS, Vonk JM, Gerritsen J, Hoek G, Brunekreef B, Rijcken B,
Schouten JP. 1999. Effects of ambient air pollution on upper and lower respiratory symptoms and
peak expiratory flow in children. Lancet 353:859-60.
Boezen M, Schouten J, Rijcken B, Vonk J, Gerritsen J, van der 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 RespirCrit Care Med 158:1848-54.
Borsboom GJ, van Pelt W, van Houwelingen HC, van Vianen BG, Schouten JP, Quanjer PH. 1999.
Diurnal variation in lung function in subgroups from two Dutch populations: consequences for
longitudinal analysis. Am J Respir Crit Care Med 159:1163-71.
Boutin-Forzano S, Adel N, Gratecos L, Jullian H, Gamier JM, RamadourM, LanteaumeA, Hamon M,
Lafay V, Charpin D. 2004. Visits to the emergency room for asthma attacks and short-term
variations in air pollution. A case-crossover study. Respiration 71:134-7.
Brain JD. 1970. Free cells in the lungs. Some aspects of their role, quantitation, and regulation. Arch
Intern Med 126:477-87.
Brauer M, Henderson S, Kirkham T, Lee KS, Rich K, Teschke K. 2002. Review of the health risks
associated with nitrogen dioxide and sulfur dioxide in indoor air. Vancouver, British Columbia,
Canada: University of British Columbia, School of Occupational and Environmental Hygiene.
Report to Health Canada.
Brauer M, Hirtle RD, Hall AC, Yip TR. 1999. Monitoring personal fine particle exposure with a particle
counter. 9:228-236.
Brauer M, Koutrakis P, KeelerGJ, Spengler JD. 1991. Indoor and outdoor concentrations of inorganic
acidic aerosols and gases. J Air Waste Manage Assoc 41:171 -81.
Brauer M, Koutrakis P, Spengler JD. 1989. Personal exposures to acidic aerosols and gases. Env Sci
Tech 23:1408-1412.
Brook RD, Franklin B, Cascio W, Hong Y, Howard G, Lipsett M, Luepker R, Mittleman M, Samet J, Smith
SC, Jr. et al. 2004. Air pollution and cardiovascular disease. A statement for healthcare
professionals from the Expert Panel on Population and Prevention Science of the American Heart
Association. Circulation 109:2655-2671.
Buescher PA, Taylor KP, Davis MH, Bowling JM. 1993. The quality of the new birth certificate data: a
validation study in North Carolina. Am J Public Health 83:1163-5.
Burnett RT, Brook J, Dann T, Delocla C, Philips O, Cakmak S, Vincent R, Goldberg MS, Krewski D. 2000.
Association between particulate- and gas-phase components of urban air pollution and daily
mortality in eight Canadian cities. Inhal Toxicol 12 Suppl 4:15-39.
Burnett RT, Brook JR, Yung WT, Dales RE, Krewski D. 1997a. Association between ozone and
hospitalization for respiratory diseases in 16 Canadian cities. Environ Research 72:24-31.
Burnett RT, Cakmak S, Brook JR, Krewski D. 1997b. The role of particulate size and chemistry in the
association between summertime ambient air pollution and hospitalization for cardiorespiratory
diseases. Environ Health Perspect 105:614-20.
Burnett RT, Smith-Doiron M, Stieb D, Raizenne ME, Brook JR, Dales RE, Leech JA, Cakmak S, Krewski
D. 2001. Association between ozone and hospitalization for acute respiratory diseases in children
less than 2 years of age. Am J Epidemiol 153:444-52.
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
Health 59:228-236.
Cakmak S, Dales RE, Judek S. 2006. Do gender, education, and income modify the effect of air pollution
gases on cardiac disease? J Occup Environ Med 48:89-94.
May 2008
R-3
DRAFT—DO NOT QUOTE OR CITE
-------
Cakmak S, Dales RE, Judek S. 2007. Respiratory health effects of air pollution gases: Modification by
education and income. Arch Environ Occup Health 61:5-10.
Calvert JG, Su F, Bottenheim JW, Strausz OP. 1978. Mechanism of the homogeneous oxidation of sulfur
dioxide in the troposphere. Atmos. Environ. 12:197-226.
Castellsague J, Sunyer J, Saez M, Anto JM. 1995. Short-term association between air pollution and
emergency room visits for asthma in Barcelona. Thorax 50:1051-1056.
Chang CC, Tsai SS, Ho SC, Yang CY. 2005. Air pollution and hospital admissions for cardiovascular
disease in Taipei, Taiwan. Environ Res 98:114-9.
Chang MC, Sioutas C, Kim S, Gong H, Jr., Linn WS. 2000. Reduction of nitrate losses from filter and
impactor samplers by means of concentration enrichment. Atmos Environ 34:85-98.
Chao CY. 2001. Comparison between indoor and outdoor air contaminant levels in residential buildings
from passive sampler study. Building and Environment 36:999-1007.
Chapman RS, Calafiore DC, Hasselblad V. 1985. Prevalence of persistent cough and phlegm in young
adults in relation to long-term ambient sulfur oxide exposure. Am Rev RespirDis 132:261-7.
Chen LC, Lam HF, Ainsworth D, Guty J, Amdur MO. 1987. Functional changes in the lungs of guinea pigs
exposed to sodium sulfite aerosols. Toxicol Appl Pharmacol 89:1-8.
Chen LC, Miller PD, Amdur MO, Gordon T. 1992. Airway hyperresponsiveness in guinea pigs exposed to
acid-coated ultrafine particles. Toxicology environ health 35:165-174.
Chen LC, Peoples SM, Amdur MO. 1991. Pulmonary effects of sulfur oxides on the surface of copper
oxide aerosol. Occup environ hyg 52:187-191.
Chen LC, Schlesinger RB. 1983. Response of the bronchial mucociliary clearance system in rabbits to
inhaled sulfite and sulfuric acid aerosols. Toxicol Appl Pharmacol 71:123-31.
Chen PC, Lai YM, Chan CC, Hwang JS, Yang CY, Wang JD. 1999. Short-term effect of ozone on the
pulmonary function of children in primary school. Environ Health Perspect 107:921.
Code of Federal Regulations. 2006. Lower detectable limit—(1) technical definition. C.F.R. 40:§53.23c.
Coleridge HM, Coleridge JC. 1994. Pulmonary reflexes: neural mechanisms of pulmonary defense. Annu
Rev Physiol 56:69-91.
Conner MW, Flood WH, Rogers AE, Amdur MO. 1989. Changes in pulmonary lavage fluid of guinea pigs
exposed to ultrafine zinc oxide with adsorbed sulfuric acid. J Toxicol Environ Health 26:223-34.
Conner MW, Lam HF, Rogers AE, Fitzgerald S, Amdur MO. 1985. Lung injury in guinea pigs caused by
multiple exposures to submicron zinc oxide mixed with sulfur dioxide in a humidified furnace. J
Toxicol Environ Health 16:101-14.
Crutzen PJ. 1976. Upper limits on atmospheric ozone reductions following increased application of fixed
nitrogen to the soil. Geophysical Research Letters 3:169-172.
Dab W, Medina S, Quenel P, Le MoullecY, Le Tertre A, Thelot B, Monteil C, Lameloise P, Pirard P, Momas
I et al. 1996. Short term respiratory health effects of ambient air pollution: results of the APHEA
project in Paris. J Epidemiol Community Health 50 Suppl 1:s42-6.
Dabdub D, Vutukuru S. 2008. Air Quality Impacts of Ship Emissions in the South Coast Air Basin of
California. Irvine, CA: State of California Air Resources Board.
Dales R, Burnett RT, Smith-Doiron M, Stieb DM, Brook JR. 2004. Air pollution and sudden infant death
syndrome. Pediatrics 113:e628-31.
Dales RE, Cakmak S, Doiron MS. 2006. Gaseous air pollutants and hospitalization for respiratory disease
in the neonatal period. Environ Health Perspect 114:1751-4.
May 2008
R-4
DRAFT—DO NOT QUOTE OR CITE
-------
Delfino RJ, Gong Jr 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.
Demokritou P, Kavouras IG, Ferguson ST, Koutrakis P. 2001. Development and laboratory performance
evaluation of a multipollutant sampler for simultaneous measurements for particulate and
gaseous pollutants. Aerosol Science and Technology 35:741-752.
Dentener F, Drevet J, Lamarque JF, Bey I, Eickhout B, Fiore AM, Hauglustaine D, Horowitz LW, Krol M,
Kulshrestha UC et al. 2006. Nitrogen and sulfur deposition on regional and global scales: a multi-
model evaluation. Global Biogeochemical Cycles 20:CitelD GB4003.
Desqueyroux H, Pujet JC, Prosper M, Le Moullec Y, Momas I. 2002a. Effects of air pollution on adults with
chronic obstructive pulmonary disease. Arch Environ Health 57:554-60.
Desqueyroux H, Pujet JC, Prosper M, Squinazi F, Momas I. 2002b. Short-term effects of low-level air
pollution on respiratory health of adults suffering from moderate to severe asthma. Environ Res
89:29-37.
Devalia JL, Rusznak C, Herdman MJ, Trigg CJ, Tarraf H, Davies RJ. 1994. Effect of nitrogen dioxide and
sulphur dioxide on airway response of mild asthmatic patients to allergen inhalation. Lancet
344:1668-71.
Dietert RR, Etzel RA, Chen D, Halonen M, Holladay SD, JarabekAM, Landreth K, Peden DB, Pinkerton
K, Smialowicz RJ et al. 2000. Workshop to identify critical window of exposure for children's
health: immune and respiratory systems work group summary. Environ Health Perspect 108:483-
490.
Dockery DW, Cunningham J, Damokosh Al, Neas LM, Spengler JD, Koutrakis P, Ware JH, Raizenne M,
Speizer FE. 1996. Health effects of acid aerosols on North American children: respiratory
symptoms. Environ Health Perspect 104:500-5.
Dockery DW, Pope CA, 3rd, Xu X, Spengler JD, Ware JH, Fay ME, Ferris BG, Jr., Speizer FE. 1993. An
association between air pollution and mortality in six U.S. cities. N Engl J Med 329:1753-9.
Dockery DW, Speizer FE, Stram DO, Ware JH, Spengler JD, Ferris BG, Jr. 1989. Effects of inhalable
particles on respiratory health of children. Am Rev RespirDis 139:587-94.
Dockery DW, Ware JH, Ferris BG, Jr., Speizer FE, Cook NR, Herman SM. 1982. Change in pulmonary
function in children associated with air pollution episodes. J Air Pollut Control Assoc 32:937-42.
Dodge R, Solomon P, Moyers J, Hayes C. 1985. A longitudinal study of children exposed to sulfur oxides.
Am J Epidemiol 121:720-36.
Dominici F; McDermott A.; Daniels M; Zeger SL; Samet JM. 2003. Mortality among residents of 90 cities.
In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA:
Health Effects Institute; pp. 9-24. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf [12
May, 2004],
Douglas GJ, Price JF, Page CP. 1994. A method for the long-term exposure of rabbits to environmental
pollutant gases. EurRespir J 7:1516-1526.
Dubowsky SD, Suh H, Schwartz J, Coull BA, Gold DR. 2006. Diabetes, obesity, and hypertension may
enhance associations between air pollution and markers of systemic inflammation. Environ
Health Perspect 114:992-8.
Dugandzic R, Dodds L, Stieb D, Smith-Doiron M. 2006. The association between low level exposures to
ambient air pollution and term low birth weight: a retrospective cohort study. Environ Health 5:3.
Eggleston HS. 1993. Uncertainties in the estimates of emissions of VOCs from motor cars. Proceedings
of the TNO/EURASAP workshop on the reliability of VOC emission databases, Delft, The
Netherlands.
May 2008
R-5
DRAFT—DO NOT QUOTE OR CITE
-------
Elliott P, Shaddick G, Wakefield JC, de Hoogh C, Briggs DJ. 2007. Long-term associations of outdoor air
pollution with mortality in Great Britain. Thorax 62:1088-94.
Farhat SCL, Paulo RLP, Shimoda TM, Conceigao GMS, Lin CA, Braga ALF, Warth MPN, Saldiva PHN.
2005. Effect of air pollution on pediatric respiratory emergency room visits and hospital
admissions. Brazilian Journal of Medical and Biological Research 38:227-235.
Federal Register. 1988. Proposed decision not to revise the national ambient air quality standards for
sulfur oxides (sulfur dioxide). F.R.April 26 53:14926-14952
Federal Register. 1994. National ambient air quality standards for sulfur oxides (sulfur dioxide)-
reproposal. F.R. November 15. 59: 58958-58980
Federal Register. 1996. National ambient air quality standards for sulfur oxides (sulfur dioxide)~final
decision. F.R. May 22. 61:25566-25580.
Federal Register. 2006. Science assessment for sulfur oxides. F.R. 71. May 15: 28023-28024.
Filleul L, Rondeau V, Vandentorren S, Le Moual N, Cantagrel A, Annesi-Maesano I, Charpin D, Declercq
C, Neukirch F, Paris C et al. 2005. Twenty five year mortality and air pollution: results from the
French PAARC survey. Occup Environ Med 62:453-60.
Finkelstein MM, Jerrett M, DeLuca P, Finkelstein N, Verma DK, Chapman K, Sears MR. 2003. Relation
between income, air pollution and mortality: a cohort study. CMAJ 169:397-402.
Finkelstein MM, Jerrett M, Sears MR. 2005. Environmental inequality and circulatory disease mortality
gradients. J Epidemiol Community Health 59:481-7.
Finlayson-Pitts BJ, Pitts JN, Jr. 1999. Chemistry of the upper and lower atmosphere: theory, experiments
and applications.
Fiore M, Petruzzi S, DeN'Omo G, Alleva E. 1998. Prenatal sulfur dioxide exposure induces changes in the
behavior of adult male mice during agonistic encounters. Neurotoxicol Teratol 20:543-8.
Folsom AR, Wu KK, Rosamond WD, SharrettAR, Chambless LE. 1997. Prospective study of hemostatic
factors and incidence of coronary heart disease: the Atherosclerosis Risk in Communities (ARIC)
Study. Circulation 96:1102-8.
Frank NR, Yoder RE, Brain JD, Yokoyama E. 1969. S02 (35S labeled) absorption by the nose and mouth
under conditions of varying concentration and flow. Arch Environ Health 18:315-22.
FrischerT, Studnicka M, Gartner C, TauberE, Horak F, VeiterA, Spengler J, Kuhr J, Urbanek R. 1999.
Lung function growth and ambient ozone: a three-year population study in school children. Am J
Respir Crit Care Med 160:390-6.
Frye C; Hoelscher B; Cyrys J; Wjst M; Wchmann HE; Heinrich J. 2003. Association of lung function with
declining ambient air pollution. Environ. Health Perspect. 111: 383-387.
Fung KY, Khan S, Krewski D, Chen Y. 2006b. Association between air pollution and multiple respiratory
hospitalizations among the elderly in Vancouver, Canada. Inhal Toxicol 18:1005-11.
Fusco D, Forastiere F, Michelozzi P, Spadea T, Ostro B, Area M, Perucci CA. 2001. Air pollution and
hospital admissions for respiratory conditions in Rome, Italy. Eur Respir J 17:1143-50.
Galan I, Tobias A, Banegas JR, Aranguez E. 2003. Short-term effects of air pollution on daily asthma
emergency room admissions. Eur Respir J 22:802-8.
Gauderman WJ, Murcray C, Gilliland F, Conti DV. 2007. Testing association between disease and multiple
SNPs in a candidate gene. Genet Epidemiol 31:383-95.
Gauderman WJ, Murcray C, Gilliland F, Conti DV. 2007. Testing association between disease and multiple
SNPs in a candidate gene. Genet Epidemiol 31:383-95. Gee GC, Payne-Sturges DC. 2004.
Environmental health disparities: a framework integrating psychosocial and environmental
concepts. Environ Health Perspect 112:1645-1653.
May 2008
R-6
DRAFT—DO NOT QUOTE OR CITE
-------
Gee GC, Payne-Sturges DC. 2004. Environmental health disparities: a framework integrating
psychosocial and environmental concepts. Environ Health Perspect 112:1645-1653.
Gilboa SM, Mendola P, Olshan AF, Langlois PH, Savitz DA, Loomis D, Herring AH, Fixler DE. 2005.
Relation between ambient air quality and selected birth defects, seven county study, Texas, 1997-
2000. Am J Epidemiol 162:238-52.
Gilliland FD, McConnell R, Peters J, Gong H, Jr. 1999. A theoretical basis for investigation ambient air
pollution and children's respiratory health. Environ Health Perspect 3:403-407.
Gilliland FD, Rappaport EB, Berhane K, Islam T, Dubeau L, Gauderman WJ, McConnell R. 2002. Effects
of glutathione S-transferase P1, M1, and T1 on acute respiratory illness in school children. Am J
Respir Crit Care Med 166:346-51.
Ginsberg GL, Foos BP, Firestone MP. 2005. Review and analysis of inhalation dosimetry methods for
application to children's risk assessment. J Toxicol Environ Health A 68:573-615.
Goldberg MS, Burnett RT, Bailar JC, 3rd, Tamblyn R, Ernst P, Flegel K, Brook J, Bonvalot Y, Singh R,
Valois MF et al. 2001. Identification of persons with cardiorespiratory conditions who are at risk of
dying from the acute effects of ambient air particles. Environ Health Perspect 109 Suppl 4:487-94.
Goldberg MS; Burnett RT. 2003 Revised analysis of the Montreal time-series study. In: Revised analyses
of time-series studies of air pollution and health. Special report. Boston, MA: Health Effects
Institute; pp. 113-132. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf [18 October,
2004],
Gong H, Jr., Lachenbruch PA, Harber P, Linn WS. 1995. Comparative short-term health responses to
sulfur dioxide exposure and other common stresses in a panel of asthmatics. Toxicol Ind Health
11:467-87.
Gong H, Jr., Lachenbruch PA, Harber P, Linn WS. 1995. Comparative short-term health responses to
sulfur dioxide exposure and other common stresses in a panel of asthmatics. Toxicol Ind Health
11:467-87.
Gong H, Jr., Linn WS, Terrell SL, Anderson KR, Clark KW. 2001. Anti-inflammatory and lung function
effects of montelukast in asthmatic volunteers exposed to sulfur dioxide. Chest 119:402-8.
Goss CH, Newsom SA, Schildcrout JS, Sheppard L, Kaufman JD. 2004. Effect of ambient air pollution on
pulmonary exacerbations and lung function in cystic fibrosis. Am J Respir Crit Care Med 169:816-
21.
Gouveia N, Fletcher T. 2000. Respiratory diseases in children and outdoor air pollution in Sao Paulo,
Brazil: a time series analysis. Occup Environ Med 57:477-83.
Gouveia N, Hajat S, Armstrong B. 2003. Socioeconomic differentials in the temperature-mortality
relationship in Sao Paulo, Brazil. Int J Epidemiol 32:390-7.
Griese M, WinzingerG, SchamsA, Josten M, ZiesenisA, MaierK. 1999. Health effects of sulfur-related
environmental air pollution: the pulmonary surfactant system is not disturbed by exposure to
acidic sulfate and neutral sulfite aerosols. J Aerosol Med 12:37-44.
Groneberg DA, Quarcoo D, Frossard N, Fischer A. 2004. Neurogenic mechanisms in bronchial
inflammatory diseases. Allergy 59:1139-52.
Grunstein MM, Hazucha M, Sorli MH, Milic-Emili J. 1977. Effect of S02 on control of breathing in
anesthetized cats. J Appl Physiol 43:844-51.
Gunnison AF, Benton AW. 1971. Sulfur dioxide: Sulfite. Interaction with mammalian serum and plasma.
Arch Environ Health 22:381-8.
Gunnison AF, Palmes ED. 1973. Persistence of plasma S-sulfonates following exposure of rabbits to
sulfite and sulfur dioxide. Toxicol Appl Pharmacol 24:266-78.
May 2008
R-7
DRAFT—DO NOT QUOTE OR CITE
-------
Gunnison AF, SellakumarA, Currie D, Snyder EA. 1987. Distribution, metabolism and toxicity of inhaled
sulfur dioxide and endogenously generated sulfite in the respiratory tract of normal and sulfite
oxidase-deficient rats. J Toxicol Environ Health 21:141-62.
Gunnison AF, SellakumarA, Snyder EA, Currie D. 1988. The effect of inhaled sulfur dioxide and systemic
sulfite on the induction of lung carcinoma in rats by benzo[a]pyrene. Environ Res 46:59-73.
Ha EH, Hong YC, Lee BE, Woo BH, Schwartz J, Christiani DC. 2001. Is air pollution a risk factor for low
birth weight in Seoul? Epidemiology 12:643-8.
Hajat S, Anderson HR, Atkinson RW, Haines A. 2002. Effects of air pollution on general practitioner
consultations for upper respiratory diseases in London. Occup Environ Med 59:294-9.
Hajat S, Haines A, Goubet SA, Atkinson RW, Anderson HR. 1999. Association of air pollution with daily
GP consultations for asthma and other lower respiratory conditions in London. Thorax 54:597-
605.
HajjAM, Burki NK, Lee LY. 1996. Role of tachykinins in sulfur dioxide-induced bronchoconstriction in
anesthetized guinea pigs. J Appl Physiol 80:2044-50.
Harris Ml, Flegal KM, Cowie CC, Eberhardt MS, Goldstein DE, Little RR, Wiedmeyer HM, Byrd-Holt DD.
1998. Prevalence of diabetes, impaired fasting glucose, and impaired glucose tolerance in U.S.
adults. The Third National Health and Nutrition Examination Survey, 1988-1994. Diabetes Care
21:518-24.
Hazucha M, Bates DV. 1975. Combined effect of ozone and sulphur dioxide on human pulmonary
function. Nature 257:50-1.
Health Effects Institute. 2004. Health effects of outdoor air pollution in developing countries of Asia: a
literature review. Boston, MA: HEI International Scientific Oversight Committee; HEI Public Health
and Air Pollution in Asia Program; special report 15.
Hedley AJ, McGhee SM, Wong CM, Barron B, Chau P, Chau J, Thach TQ, Wong TW, Loh C. 2006. Air
Pollution: Costs and Paths to a Solution. Hong Kong: Civic Exchange.
Hedley AJ; Wong CM; Thach TQ; Ma S; Lam TH; Anderson HR. 2002. Cardiorespiratory and all-cause
mortality after restrictions on sulphur content of fuel in Hong Kong: an intervention study. Lancet
360: 1646-1652.
Heinrich J, Hoelscher B, Frye C, Meyer I, Pitz M, Cyrys J, Wjst M, Neas L, Wchmann H-E. 2002.
Improved air quality in reunified Germany and decreases in respiratory symptoms. Epidemiology
13:394-401.
Heinrich U, Mohr U, Fuhst R, Brockmeyer C. 1989. Investigation of a potential cotumorigenic effect of the
dioxides of nitrogen and sulfur, and of diesel-engine exhaust, on the respiratory tract of Syrian
golden hamsters. Res Rep Health Eff lnst:1 -27.
Henneberger A, Zareba W, Ibald-Mulli A, Ruckerl R, Cyrys J, Couderc JP, Mykins B, Woelke G,
Wchmann HE, Peters A. 2005. Repolarization changes induced by air pollution in ischemic heart
disease patients. Environ Health Perspect 113:440-6.
Herbarth O, Fritz G, Krumbiegel P, Diez U, Franck U, Richter M. 2001. Effect of sulfur dioxide and
particulate pollutants on bronchitis in children-a risk analysis. Environ Toxicol 16:269-76.
Heyder J, Beck-Speier I, Ferron GA, Heilmann P, Karg E, Kreyling WG, Lenz AG, Maier K, Schulz H,
Takenaka S et al. 1992. Early response of the canine respiratory tract following long-term
exposure to a sulfur(IV) aerosol at low concentration. I. Rationale, design, methodology, and
summary. Inhal Toxicol 4:159-174.
Higgins BG, Francis HC, Yates CJ, Warburton CJ, Fletcher AM, Reid JA, Pickering CA, WoodcockAA.
1995. Effects of air pollution on symptoms and peak expiratory flow measurements in subjects
with obstructive airways disease. Thorax 50:149-155.
Hill AB. 1965. The Environment and Disease: Association or Causation? Proc R Soc Med 58:295-300.
May 2008
R-8
DRAFT—DO NOT QUOTE OR CITE
-------
Hiller FC. 1991. Health implications of hygroscopic particle growth in the human respiratory tract. J.
Aerosol Med 4:23.
Hoek G, Brunekreef B, Fischer P, van Wijnen J. 2001. The association between air pollution and heart
failure, arrhythmia, embolism, thrombosis, and other cardiovascular causes of death in a time
series study. Epidemiology 12:355-7.
Hoek G, Brunekreef B, Verhoeff A, van Wijnen J, Fischer P. 2000. Daily mortality and air pollution in The
Netherlands. J Air Waste Manag Assoc 50:1380-9.
Hoek G, Brunekreef B. 1993. Acute effects of a winter air pollution episode on pulmonary function and
respiratory symptoms of children. Arch Environ Health 48:328-35.
Hoek G, Brunekreef B. 1995. Effect of photochemical air pollution on acute respiratory symptoms in
children. Am J Respir Crit Care Med 151:27-32.
Hoek G. 2003. Daily mortality and air pollution in The Netherlands. In: Revised analyses of time-series
studies of air pollution and health. Special report. Boston, MA: Health Effects Institute; pp. 133-
141. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf.
Holland HD. 1978. The chemistry of the atmosphere and oceans.
Hoppel WA, Caffrey PF. 2005. Oxidation of S(IV) in sea-salt aerosol at high pH: ozone versus aerobic
reaction. Journal of Geophysical Research 110D23202.
Horowitz LW, Walters S, Mauzerall DL, Emmons LK, Rasch PJ, Granier C, Tie X, Lamarque J-F, Schultz
MG, Tyndall GS et al. 2003. A global simulation of tropospheric ozone and related tracers:
description and evaluation of MOZART, version 2. Journal of Geophysical Research 108D24.
Horstman D, Roger LJ, Kehrl H, Hazucha M. 1986. Airway sensitivity of asthmatics to sulfur dioxide.
Toxicol Ind Health 2:289-98.
Horstman DH, Seal E, Jr., Folinsbee LJ, Ives P, Roger LJ. 1988. The relationship between exposure
duration and sulfur dioxide-induced bronchoconstriction in asthmatic subjects. Am Ind Hyg Assoc
J 49:38-47.
Hosseinpoor AR, Forouzanfar MH, Yunesian M, Asghari F, Naieni KH, Farhood D. 2005. Air pollution and
hospitalization due to angina pectoris in Tehran, Iran: a time-series study. Environ Res 99:126-31.
IARC. 2006. Report of the advisory group to review the amended preamble to the IARC
monographs.(IARC Int Rep No 06/001).
Ibald-Mulli A, Stieber J, Wchmann HE, Koenig W, Peters A. 2001. Effects of air pollution on blood
pressure: a population-based approach. Am J Public Health 91:571-7.
IOM. 2007. Improving the Presumptive Disability Decision-Making Process for Veterans. In: Samet James
M, Bodurow CC, editors: Institute of Medicine of the National Academies.
Ito A, Ito A, Akimoto H. 2007. Seasonal and interannual variations in CO and BC emissions from open
biomass burning in Southern Africa during 1998-2005. Global Biogeochemical Cycles 21 :[np],
Ito T, Ohyama K, Kusano T, Usuda Y, Nozawa A, Hayashi H, Ohji H, Kitamura H, Kanisawa M. 1997.
Pulmonary endocrine cell hyperplasia and papilloma in rats induced by intratracheal injections of
extract from particulate air pollutants. Exp Toxicol Pathol 49:65-70.
Jacob DJ. 1999. Introduction to atmospheric chemistry. Princeton, New Jersey: Princeton University
Press.
Jacobson MZ. 2002. Atmospheric pollution: history, science, and regulation.
Jaffe DH, Singer ME, Rimm AA. 2003. Air pollution and emergency department visits for asthma among
Ohio Medicaid recipients, 1991-1996. Environ Res 91:21-8.
May 2008
R-9
DRAFT—DO NOT QUOTE OR CITE
-------
Jakab GJ, Clarke RW, Hemenway DR, Longphre MV, Kleeberger SR, Frank R. 1996. Inhalation of acid
coated carbon black particles impairs alveolar macrophage phagocytosis. Toxicology letters
88:243-248.
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. J Expo Sci Environ Epidemiol 16:225-37.
Jedrychowski W. 1999. Ambient air pollution and respiratory health in the east Baltic region. Scand J
Work Environ Health 25 Suppl 3:5-16.
Jerrett M, Burnett RT, Brook J, Kanaroglou P, Giovis C, Finkelstein N, Hutchison B. 2004. Do
socioeconomic characteristics modify the short term association between air pollution and
mortality? Evidence from a zonal time series in Hamilton, Canada. J Epidemiol Community Health
58:31-40.
Jerrett M, Burnett R, Goldberg M, Sears M, Krewski D, Catalan R, Kanaroglou P, Giovis C, Finkelstein N.
2003. Spatial Analysis for Environmental Health Research: Concepts, Methods, and Examples.
Journal of Toxicology and Environmental Health Part A 66:1783-1810. Jerrett M, Finkelstein M.
2005. Geographies of risk in studies linking chronic air pollution exposure to health outcomes. J
Toxicol Environ Health A 68:1207-42.
Jet Propulsion Laboratory (JPL). 2003. Chemical kinetics and photochemical data for use in atmospheric
studies. Pasadena, CA: California Institute of Technology; JPL publication no. 02-25. Available:
http://jpldataeval.jpl.nasa.gov/pdf/JPL_02-25_1_intro_rev0.pdf.
Jorres R, Magnussen H. 1990. Airways response of asthmatics after a 30 min exposure, at resting
ventilation, to 0.25 ppm N02 or 0.5 ppm S02. EurRespir J 3:132-137.
Kasper-Giebl A, Krenn S, Puxbaum H. 1999. Laboratory and field measurements of a badge type passive
sampler for the determination of ambient sulfur dioxide concentrations. Analytical and
bioanalytical chemistry 363:73-76.
Katsouyanni K, Touloumi G, Spix C, Schwartz J, Balducci F, Medina S, Rossi G, Wojtyniak B, Sunyer J,
Bacharova L et al. 1997. Short term effects of ambient sulphur dioxide and particulate matter on
mortality in 12 European cities: results from time series data from the APHEA project. British
Medical Journal 314:1658-1663.
Kauffmann F. 2004. Post-genome respiratory epidemiology: a multidisciplinary challenge. Eur Respir J
24:471-80.
Kehrl HR, Roger LJ, Hazucha MJ, Horstman DH. 1987. Differing response of asthmatics to sulfurdioxide
exposure with continuous and intermittent exercise. Am Rev Respir Dis 135:350-5.
Kele§ N; llicali C; Deger K. 1999. The effects of different levels of air pollution on atopy and symptoms of
allergic rhinitis. Am. J. Rhinol. 13: 185-190.
Kelly FJ, Mudway IS. 2003. Protein oxidation at the air-lung interface. Amino Acids 25:375-96.
Khoury MJ, Davis R, Gwinn M, Lindegren ML, Yoon P. 2005. Do we need genomic research for the
prevention of common diseases with environmental causes? Am J Epidemiol 161:799-805.
Kimbell JS, Miller FJ. 1999. Regional respiratory-tract absorption of inhaled reactive gases: a modeling
approach.
Kindzierski WB, Ranganathan HKS. 2006. Indoor and outdoor S02 in a community near oil sand
extraction and production facilities in northern Alberta. Journal of Environmental Engineering and
Science 5:S121-S129.
Kindzierski WB, Sembaluk S. 2001. Indoor outdoor relationship of S02 concentrations in a rural and an
urban community of Alberta. Canadian Journal of Civil Engineering 28:163-169.
May 2008
R-10
DRAFT—DO NOT QUOTE OR CITE
-------
Klepeis NE, Nelson WC, R. OW, Robinson JPT, A. M., Switzer P, Behar JV, Hern SC, Engelmann WH.
2001. The National Human Activity Pattern Survey (NHAPS): a resource for assessing exposure
to environmental pollutants. Journal of Exposure Science and Environmental Epidemiology
11:231-252.
Koenig JQ, Covert DS, Hanley QS, van Belle G, Pierson WE. 1990. Prior exposure to ozone potentiates
subsequent response to sulfur dioxide in adolescent asthmatic subjects. Am Rev Respir Dis
141:377-80.
Koenig JQ, Dumler K, Rebolledo V, Williams PV, Pierson WE. 1992. Theophylline mitigates the
bronchoconstrictor effects of sulfur dioxide in subjects with asthma. J Allergy Clin Immunol
89:789-94.
Koenig JQ, Marshall SG, Horike M, Shapiro GG, Furukawa CT, Bierman CW, Pierson WE. 1987. The
effects of albuterol on sulfur dioxide-induced bronchoconstriction in allergic adolescents. The
Journal of allergy and clinical immunology 79:54-8.
Koenig JQ, Marshall SG, van Belle G, McManus MS, Bierman CW, Shapiro GG, Furukawa CT, Pierson
WE. 1988. Therapeutic range cromolyn dose-response inhibition and complete obliteration of
S02-induced bronchoconstriction in atopic adolescents. J Allergy Clin Immunol 81:897-901.
Koenig JQ, Pierson WE, Horike M, Frank R. 1981. Effects of S02 plus NaCI aerosol combined with
moderate exercise on pulmonary function in asthmatic adolescents. Environ Res 25:340-8.
Koenig JQ, Pierson WE, Horike M, Frank R. 1983. A comparison of the pulmonary effects of 0.5 ppm
versus 1.0 ppm sulfur dioxide plus sodium chloride droplets in asthmatic adolescents. J Toxicol
Environ Health 11:129-39.
Koken PJ, Piver WT, Ye F, ElixhauserA, Olsen LM, PortierCJ. 2003. Temperature, air pollution, and
hospitalization for cardiovascular diseases among elderly people in Denver. Environ Health
Perspect 111:1312-7.
Koutrakis P, Wolfson JM, Spengler JD. 1988. An improved method for measuring aerosol strong acidity:
results from a nine-month study in St. Louis, Missouri and Kingston, Tennessee. Atmospheric
Environment 22:157-162.
Koutrakis P; Suh HH; Sarnat JA; Brown KW; Coull BA; Schwartz J. 2005. Characterization of particulate
and gas exposures of sensitive populations living in Baltimore and Boston. Boston, MA: Health
Effects Institute; research report no. 131.
Krewski D, Burnett RT, Goldberg MS, Hoover K, Siemiatycki J, Jerrett M, Abrahamowicz M, White WH.
2000 Reanalysis of the Harvard Six Cities study and the American Cancer Society study of
particulate air pollution and mortality: a special report of the Institute's Particle Epidemiology
Reanalysis Project. Cambridge, MA: Health Effects Institute. Available:
http://pubs.healtheffects.org/view.php?id=6 [6 March, 2007],
Kunzli N. 2005. Unifying susceptibility, exposure, and time: discussion of unifying analytic approaches and
future directions. J Toxicol Environ Health A 68:1263-71.
Kwon HJ, Cho SH, Nyberg F, Pershagen G. 2001. Effects of ambient air pollution on daily mortality in a
cohort of patients with congestive heart failure. Epidemiology 12:413-9.
Lam HF, Peisch R, AmdurMO. 1982. Changes in lung volumes and diffusing capacity in guinea pigs
exposed to a combination of sulfur dioxide and submicron zinc oxide mixed in a humidified
furnace. Toxicology and Applied Pharmacology 66:427-433.
Langley-Evans SC, Phillips GJ, Jackson AA. 1996. Sulphur dioxide: a potent glutathione depleting agent.
Comp Biochem Physiol C Pharmacol Toxicol Endocrinol 114:89-98.
Lawther PJ, Macfarlane AJ, Waller RE, Brooks AG. 1975. Pulmonary function and sulphur dioxide, some
preliminary findings. Environ Res 10:355-67.
Lawther PJ, Waller RE, Henderson M. 1970. Air pollution and exacerbations of bronchitis. Thorax 25:525-
39.
May 2008
R-11
DRAFT—DO NOT QUOTE OR CITE
-------
Lee BE, Ha EH, Park HS, Kim YJ, Hong YC, Kim H, Lee JT. 2003a. Exposure to air pollution during
different gestational phases contributes to risks of low birth weight. Hum Reprod 18:638-43.
Lee JT, Kim H, Cho YS, Hong YC, Ha EH, Park H. 2003b. Air pollution and hospital admissions for
ischemic heart diseases among individuals 64+ years of age residing in Seoul, Korea. Arch
Environ Health 58:617-23.
Lee JT, Kim H, Song H, Hong YC, Cho YS, Shin SY, Hyun YJ, Kim YS. 2002. Air pollution and asthma
among children in Seoul, Korea. Epidemiology 13:481-4.
Lee J-T, Shin D, Chung Y. 1999. Air pollution and daily mortality in Seoul and Ulsan, Korea. Environ
Health Perspect 107:149-154.
Lee YL, Lin YC, Lee YC, Wang JY, Hsiue TR, Guo YL. 2004. Glutathione S-transferase P1 gene
polymorphism and air pollution as interactive risk factors for childhood asthma. Clin Exp Allergy
34:1707-13.
Leem JH, Kaplan BM, Shim YK, Pohl HR, Gotway CA, Bullard SM, Rogers JF, Smith MM, Tylenda CA.
2006. Exposures to air pollutants during pregnancy and preterm delivery. Environ Health Perspect
114:905-10.
Leung FT, A. J. Colussi, M. R. Hoffmann, and G. C. Toon. 2002. Isotopic fractionation of carbonyl sulfide
in the atmosphere: implications for the source of background stratospheric sulfate aerosol.
Geophysical Research Letters 29.
Levine JS, Pinto JP. 1998. The production of CO by biomass burning.
Lewis AJ, KirchnerT. 1984. Modulation of sulfur dioxide-induced airways hyperresponsiveness in the
conscious dog. IntArch Allergy Appl Immunol 75:188-90.
Li R, Meng Z, Xie J. 2007a. Effects of sulfur dioxide on the expressions of MUC5AC and ICAM-1 in
airway of asthmatic rats. Regulatory Toxicology and Pharmacology 48:284-291.
Li R, Meng Z, Xie J. 2007b. Effects of sulfur dioxide on the expressions of MUC5AC and ICAM-1 in
airway of asthmatic rats. Regul Toxicol Pharmacol 48:284-91.
Li R, Meng Z. 2007. Effects of S02 derivatives on expressions of MUC5AC and IL-13 in human bronchial
epithelial cells. Arch Toxicol 81:867-74.
Li YQ, Harrison RM. 1990. Comparison of indoor and outdoor concentrations of acid gases, ammonia and
their associated salts. Environ. Technol.
Liao D, Duan Y, Whitsel EA, Zheng ZJ, Heiss G, Chinchilli VM, Lin HM. 2004. Association of higher levels
of ambient criteria pollutants with impaired cardiac autonomic control: a population-based study.
Am J Epidemiol 159:768-77.
Lin CA, Amador Pereira LA, de Souza Conceicao GM, Kishi HS, Milani R, Jr., Ferreira Braga AL,
Nascimento Saldiva PH. 2003a. Association between air pollution and ischemic cardiovascular
emergency room visits. Environ Res 92:57-63.
Lin CM, Li CY, Yang GY, Mao IF. 2004. Association between maternal exposure to elevated ambient sulfur
dioxide during pregnancy and term low birth weight. Environ Res 96:41-50.
Lin M, Chen Y, Burnett RT, Villeneuve PJ, Krewski D. 2003b. Effect of short-term exposure to gaseous
pollution on asthma hospitalisation in children: a bi-directional case-crossover analysis. J
Epidemiol Community Health 57:50-5.
Lin M, Stieb DM, Chen Y. 2005. Coarse Particulate Matter and Hospitalization for Respiratory Infections in
Children Younger Than 15 Years in Toronto: A Case-Crossover Analysis. Pediatrics 116:e235-240.
Lin SC, Tai CC, Chan CC, Wang JD. 1994 The indoor/outdoor relationship of acid aerosols in Taipei. Sci
Total Environ.
Linn WS, Avol EL, Peng RC, Shamoo DA, Hackney JD. 1987. Replicated dose-response study of sulfur
dioxide effects in normal, atopic, and asthmatic volunteers. Am Rev RespirDis 136:1127-34.
May 2008
R-12
DRAFT—DO NOT QUOTE OR CITE
-------
Linn WS, Avol EL, Shamoo DA, Peng RC, Spier CE, Smith MN, Hackney JD. 1988. Effect of
metaproterenol sulfate on mild asthmatics' response to sulfur dioxide exposure and exercise.
Arch Environ Health 43:399-406.
Linn WS, Avol EL, Shamoo DA, Venet TG, Anderson KR, Whynot JD, Hackney JD. 1984. Asthmatics'
responses to 6-hr sulfur dioxide exposures on two successive days. Arch Environ Health 39:313-
9.
Linn WS, Fischer DA, Shamoo DA, Spier CE, Valencia LM, Anzar UT, Hackney JD. 1985. Controlled
exposures of volunteers with chronic obstructive pulmonary disease to sulfur dioxide. Environ
Res 37:445-51.
Linn WS, Gong H, Jr., Shamoo DA, Anderson KR, Avol EL. 1997. Chamber exposures of children to
mixed ozone, sulfur dioxide, and sulfuric acid. Arch Environ Health 52:179-87.
Linn WS, Jones MP, Bailey RM, Kleinman MT, Spier CE, Fischer DA, Hackney JD. 1980. Respiratory
effects of mixed nitrogen dioxide and sulfur dioxide in human volunteers under simulated ambient
exposure conditions. Environ Res 22:431-8.
Linn WS, Shamoo DA, Peng RC, Clark KW, Avol EL, Hackney JD. 1990. Responses to sulfur dioxide and
exercise by medication-dependent asthmatics: effect of varying medication levels. Arch Environ
Health 45:24-30.
Linn WS, Venet TG, Shamoo DA, Valencia LM, Anzar UT, Spier CE, Hackney JD. 1983. Respiratory
effects of sulfur dioxide in heavily exercising asthmatics. A dose-response study. Am Rev Respir
Dis 127:278-83.
Lipfert FW, Baty 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 18:645-57.
Lipfert FW, Morris SC, Wyzga RE. 2000a. Daily mortality in the Philadelphia metropolitan area and size-
classified particulate matter. J Air Waste Manag Assoc 50:1501-13.
Lipfert FW, Perry HM, Jr., Miller JP, Baty JD, Wyzga RE, Carmody SE. 2000b. The Washington
University-EPRI Veterans' Cohort Mortality Study: preliminary results. Inhal Toxicol 12 Suppl 4:41-
73.
Lipfert FW, Wyzga RE, Baty 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.
Atmosph Environment 40:154-169.
Lippmann M, Ito K, Nadas A, Burnett RT. 2000. Association of particulate matter components with daily
mortality and morbidity in urban populations. Res Rep Health Eff lnst:5-72, discussion 73-82.
Liu Q, Wong-Riley MT. 2003. Postnatal changes in cytochrome oxidase expressions in brain stem nuclei
of rats: implications for sensitive periods. J Appl Physiol 95:2285-91.
Liu S, Krewski D, Shi Y, Chen Y, Burnett RT. 2007. Association between maternal exposure to ambient air
pollutants during pregnancy and fetal growth restriction. J Expo Sci Environ Epidemiol 17:426-32.
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 68:30-5.
Luginaah IN, Fung KY, Gorey KM, Webster G, Wlls 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.
Luke WT. 1997. Evaluation of a commercial pulsed fluorescence detector for the measurement of low-
level S02 concentrations during the Gas-Phase Sulfur Intercomparison Experiment. Journal of
Geophysical Research 102:16,255-16,265.
Lunn JE, Knowelden J, Handyside AJ. 1967. Patterns of respiratory illness in Sheffield infant
schoolchildren. Br J Prev Soc Med 21:7-16.
May 2008
R-13
DRAFT—DO NOT QUOTE OR CITE
-------
Lunn JE, Knowelden J, Roe JW. 1970. Patterns of respiratory illness in Sheffield junior schoolchildren. A
follow-up study. Br J Prev Soc Med 24:223-8.
Mage D, Wilson W, Hasselblad V, Grant L. 1999. Assessment of human exposure to ambient particulate
matter. J Air Waste Manag Assoc 49:1280-91.
Magnussen H, Jorres R, Wagner HM, von Nieding G. 1990. Relationship between the airway response to
inhaled sulfur dioxide, isocapnic hyperventilation, and histamine in asthmatic subjects. IntArch
Occup Environ Health 62:485-91.
Maier K, Beck-Speier I, Dayal N, Heilmann P, Hinze H, Lenz A-G, Leuschel L, Matejkova E, Miaskowski U,
J. H. 1992. Early response of the canine respiratory tract following long-term exposure to a sulfur
(IV) aerosol at low concentration. II. Biochemistry and cell biology of lung lavage fluid. Inhal
Toxicol 4:175-195.
Maisonet M, Bush TJ, Correa A, Jaakkola JJ. 2001. Relation between ambient air pollution and low birth
weight in the Northeastern United States. Environ Health Perspect 109 Suppl 3:351-6.
Makri A, Stilianakis Nl. 2007. Vulnerability to air pollution health effects. Int J Hyg Environ Health.
Martins MC, Fatigati FL, Vespoli TC, Martins LC, Pereira LA, Martins MA, Saldiva PH, Braga AL. 2004.
Influence of socioeconomic conditions on air pollution adverse health effects in elderly people: an
analysis of six regions in Sao Paulo, Brazil. J Epidemiol Community Health 58:41-6.
McCurdy T, Glen G, Smith L, Lakkadi Y. 2000. The national exposure research laboratory's consolidated
human activity database. J Expo Anal Environ Epidemiol 10:566-78.
Melville GN. 1970. Changes in specific airway conductance in healthy volunteers following nasal and oral
inhalation of S02. West Indian Med J 19:231-5.
Meng Z, Geng H, Bai J, Yan G. 2003. Blood pressure of rats lowered by sulfur dioxide and its derivatives.
Inhal Toxicol 15:951-9.
Meng Z, Qin G, Zhang B. 2005. DNA damage in mice treated with sulfur dioxide by inhalation.
Environmental and Molecular Mutagenesis 46:150-155.
Meng Z, Zhang B, Ruan A, Sang N, Zhang J. 2002. Micronuclei induced by sulfur dioxide inhalation in
mouse bone-marrow cells in vivo. Inhal Toxicol 14:303-309.
Menzel DB, Keller DA, Leung K-H. 1986. Covalent reactions in the toxicity of S02 and sulfite.
Meranger JC, Brule D, Smith-Palmer T, Wentzell B, Donini JC. 1987. Indoor-outdoor levels of nitrogen
and sulphur species and their relation to air flow in Antigonish, Nova Scotia. International Journal
of Environmental Analytical Chemistry 29:61-72.
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-
92.
Metzger KB, Tolbert PE, Klein M, Peel JL, Flanders WD, Todd K, Mulholland JA, Ryan PB, Frumkin H.
2004. Ambient air pollution and cardiovascular emergency department visits. Epidemiology 15:46-
56.
Michaud J-P, Grove JS, Krupitsky D. 2004. Emergency department visits and "vog"-related air quality in
Hilo, Hawai'i. Environmental Research 95:11-19.
Miller CA, Lemieux PM. 2007. Emissions from the burning of vegetative debris in air curtain destructors. J
Air Waste Manag Assoc 57:959-67.
Mohorovic L. 2004. First two months of pregnancy - critical time for preterm delivery and low birthweight
caused by adverse effects of coal combustion toxics. Early Human Development. Elsevier
Science Ltd, Oxford, UK: 2004. 80: 2, 115-123.
May 2008
R-14
DRAFT—DO NOT QUOTE OR CITE
-------
Moolgavkar SH. 2003. Air pollution and daily deaths and hospital admissions in Los Angeles and Cook
counties. Revised Analyses of Time-Series Studies of Air Pollution and Health. Special Report.
Boston: Health Effects lnstitute:183—198.
Morello-Frosch R, Shenassa ED. 2006. The Environmental "Riskscape" and Social Inequality:
Implications for Explaining Maternal and Child Health Disparities. Environ Health Perspect
114:1150.
Morrow PE. 1986. Factors determining hygroscopic aerosol deposition in airways. Physiol Rev 66:330-76.
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.
Mortimer KM, Tager IB, Dockery DW, Neas LM, Redline S. 2000. The effect of ozone on inner-city
children with asthma: identification of susceptible subgroups. Am J Respir Crit Care Med
162:1838-45.
Myers DJ, Bigby BG, Calvayrac P, Sheppard D, Boushey HA. 1986. Interaction of cromolyn and a
muscarinic antagonist in inhibiting bronchial reactivity to sulfur dioxide and to eucapnic hyperpnea
alone. Am Rev Respir Dis 133:1154-8.
Myhre G, Berglen TF, Myhre CE, Isaksen ISA. 2004. The radiative effect of the anthropogenic influence
on the stratospheric sulfate aerosol layer. Tellus Ser B 56:294-299.
Nadel JA, Salem H, Tamplin B, Tokiwa Y. 1965a. Mechanism of Bronchoconstriction during Inhalation of
Sulfur Dioxide. J Appl Physiol 20:164-7.
Nadel JA, Tamplin B, Tokiwa Y. 1965b. Mechanism of Bronchoconstriction. During Inhalation of Sulfur
Dioxide; Reflex Involving Vagus Nerves. Arch Environ Health 10:175-8.
Nadziejko C, Fang K, Narciso S, Zhong M, Su WC, Gordon T, Nadas A, Chen LC. 2004. Effect of
particulate and gaseous pollutants on spontaneous arrhythmias in aged rats. Inhal Toxicol
16:373-80.
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 58:1071-6.
Nafstad P, Haheim LL, WisloffT, 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 112:610-5.
NAPAP Oversight Review Board. 1991. The experience and legacy of NAPAP. Report of the Oversight
Review Board of the National Acid Precipitation Assessment Program, 40 pp.
National Center for Health Statistics. 2006. Summary health statistics for U.S. adults: National Health
Interview Survey, 2004. Hyattsville, MD: U.S. Department of Health & Human Services, Centers
for Disease Control and Prevention. (Vital and Health Statistics, Series 10, no. 228). Available:
http://www.cdc.gov/nchs/data/series/sr_10/sr10_228.pdf.
National Research Council. 2004. Research priorities for airborne particulate matter. IV. Continuing
research progress. Washington, DC: National Academies Press. Available:
http://www.nap.edu/catalog.php?record_id=10957 [1 August, 2007],
Neas LM, Dockery DW, Koutrakis P, Tollerud DJ, Speizer FE. 1995. The association of ambient air
pollution with twice daily peak expiratory flow rate measurements in children. Am J Epidemiol
141:111-22.
Neukirch F, Segala C, Le Moullec Y, Korobaeff M, Aubier M. 1998. Short-term effects of low-level winter
pollution on respiratory health of asthmatic adults. Arch Environ Health 53:320-8.
Nicks DKB, R. L. 2001. Subminute measurements of S02 and low parts per trillion by volume mixing
ratios in the atmosphere. J. Geophys. Res. [Atmos.] 106:27,769-27,776.
May 2008
R-15
DRAFT—DO NOT QUOTE OR CITE
-------
O'Neill MS, Jerrett M, Kawachi I, Levy Jl, Cohen AJ, Gouveia N, Wilkinson P, Fletcher T, Cifuentes L,
Schwartz J. 2003. Health, wealth, and air pollution: advancing theory and methods. Environ
Health Perspect 111:1861-70.
Oftedal B, Nafstad P, Magnus P, Bjorkly S, Skrondal A. 2003. Traffic related air pollution and acute
hospital admission for respiratory diseases in Drammen, Norway 1995-2000. European Journal
of Epidemiology 18:671-676.'
Park H, Lee B, Ha EH, Lee JT, Kim H, Hong YC. 2002. Association of air pollution with school
absenteeism due to illness. Arch Pediatr Adolesc Med 156:1235-9.
Park JK, Kim YK, Lee SR, Cho SH, Min KU, Kim YY. 2001. Repeated exposure to low levels of sulfur
dioxide (S02) enhances the development of ovalbumin-induced asthmatic reactions in guinea
pigs. Ann Allergy Asthma Immunol 86:62-7.
Patterson E, Eatough DJ. 2000. Indoor/outdoor relationships for ambient PM2.5 and associated
pollutants: epidemiological implications in Lindon, Utah. J Air Waste Manag Assoc 50:103-10.
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
165:625-33.
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-
74.
Pekkanen J, Brunner EJ, Anderson HR, Tiittanen P, Atkinson RW. 2000. Daily concentrations of air
pollution and plasma fibrinogen in London. Occup Environ Med 57:818-22.
Pereira LAA, Loomis D, Conceicao GMS, Braga ALF, Areas RM, Kishi HS, Singer JM, Bohm GM, Saldiva
PHN. 1998. Association between air pollution and intrauterine mortality in Sao Paulo, Brazil.
Environ Health Perspect 106:325-329.
Peters A, Dockery DW, Muller JE, Mittleman MA. 2001. Increased particulate air pollution and the
triggering of myocardial infarction. Circulation 103:2810-5.
Peters A, Doring A, Wchmann HE, Koenig W. 1997. Increased plasma viscosity during an air pollution
episode: a link to mortality? Lancet 349:1582-7.
Peters A, Goldstein IF, Beyer U, Franke K, Heinrich J, Dockery DW, Spengler JD, Wichmann HE. 1996.
Acute health effects of exposure to high levels of air pollution in eastern Europe. Am J Epidemiol
144:570-81.
Peters A, Perz S, Doring A, Stieber J, Koenig W, Wchmann HE. 1999. Increases in heart rate during an
air pollution episode. Am J Epidemiol 150:1094-8.
Petroeschevsky A, Simpson RW, Thalib L, Rutherford S. 2001. Associations between outdoor air pollution
and hospital admissions in Brisbane, Australia. Arch Environ Health 56:37-52.
Petruzzi S, DeN'Omo G, Fiore M, Chiarotti F, Bignami G, Alleva E. 1996. Behavioural disturbances in adult
CD-1 mice and absence of effects on their offspring upon S02 exposure. Archives of Toxicology
70:757-766.
Pikhart H, Bobak M, Gorynski P, Wojtyniak B, Danova J, Celko MA, Kriz B, Briggs D, Elliot P. 2001.
Outdoor sulphur dioxide and respiratory symptoms in Czech and Polish school children: a small-
area study (SAVIAH). International Archives of Occupational and Environmental Health 74:574-
578.
Pikhart H, Bobak M, Kriz B, Danova J, Celko MA, Prikazsky V, Pryl K, Briggs D, Elliott P. 2000. Outdoor
air concentrations of nitrogen dioxide and sulfur dioxide and prevalence of wheezing in school
children. Epidemiology 11:153-60.
Pino P, Walter T, Oyarzun M, Villegas R, Romieu I. 2004. Fine particulate matter and wheezing illnesses
in the first year of life. Epidemiology 15:702-8.
May 2008
R-16
DRAFT—DO NOT QUOTE OR CITE
-------
Piper JM, Mitchel EF, Jr., Snowden M, Hall C, Adams M, Taylor P. 1993. Validation of 1989 Tennessee
birth certificates using maternal and newborn hospital records. American Journal of Epidemiology
137:758-768.
Plunkett LM, Turnbull D, Rodricks JV. 1992. Differences between adults and children affecting exposure
assessment. In Guzelian, P. S.; Henry, D. J.; Olin, S. S., eds. Similarities and differences between
children and adults: implications for risk assessment. Washington, DC: ILSI Press:79-96.
Ponce de Leon A, Anderson HR, Bland JM, Strachan DP, Bower J. 1996. Effects of air pollution on daily
hospital admissions for respiratory disease in London between 1987-88 and 1991-92. In: The
APHEA project. Short term effects of air pollution on health: a European approach using
epidemiological time series data. J. Epidemiol. Community Health 50:S63-S70.
Ponce NA, Hoggatt KJ, Wilhelm M, Ritz B. 2005. Preterm Birth: The Interaction of Traffic-related Air
Pollution with Economic Hardship in Los Angeles Neighborhoods. Am. J. Epidemiol. 162:140-148.
Ponka A. 1990. Absenteeism and respiratory disease among children and adults in Helsinki in relation to
low-level air pollution and temperature. Environ Res 52:34-46.
Pool BL, BrendlerS, Klein RG, Monarca S, Pasquini R, Schmezer P, Zeller WJ. 1988. Effects of S02 or
NOx on toxic and genotoxic properties of chemical carcinogens. II. Short term in vivo studies.
Carcinogenesis 9:1247-1252.
Pool-Zobel BL, Schmezer P, Zeller WJ, Klein RG. 1990. In vitro and ex vivo effects of the air pollutants
S02 and NOx on benzo(a)pyrene activating enzymes of the rat liver. Exp Pathol 39:207-212.
Pope CA, 3rd, Burnett RT, Thun MJ, Calle EE, Krewski D, Ito K, Thurston GD. 2002. Lung cancer,
cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA
287:1132-41.
Pope CA, 3rd, Thun MJ, Namboodiri MM, Dockery DW, Evans JS, Speizer FE, Heath CW, Jr. 1995.
Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am. J.
Respir. Crit. Care Med. 151:669-674.
Ramadour M, Burel C, Lanteaume A, Vervloet D, Charpin D, Brisse F, Dutau H. 2000. Prevalence of
asthma and rhinitis in relation to long-term exposure to gaseous air pollutants. Allergy 55:1163-9.
Rich DQ, Schwartz J, Mittleman MA, Link M, Luttmann-Gibson H, Catalano PJ, Speizer FE, Dockery DW.
2005. Association of short-term ambient air pollution concentrations and ventricular arrhythmias.
Am J Epidemiol 161:1123-32.
Rich DQ, Schwartz J, Mittleman MA, Link M, Luttmann-Gibson H, Catalano PJ, Speizer FE, Dockery DW.
2005. Association of short-term ambient air pollution concentrations and ventricular arrhythmias.
Am J Epidemiol 161:1123-32.
Rich KE, Petkau J, Vedal S, Brauer M. 2004. A case-crossover analysis of particulate air pollution and
cardiac arrhythmia in patients with implantable cardioverter defibrillators. Inhal Toxicol 16:363-72.
Riedel F, Kramer M, Scheibenbogen C, Rieger CH. 1988. Effects of S02 exposure on allergic
sensitization in the guinea pig. J Allergy Clin Immunol 82:527-34.
Robinson JM. 1989. On uncertainty in the computation of global emissions from biomass burning.
Climatic Change 14:243-261.
RoemerW, HoekG, Brunekreef B, Haluszka J, KalandidiA, Pekkanen J. 1998. Daily variations in air
pollution and respiratory health in a multicentre study: the PEACE project. Eur Respir 12:1354-
1361.
Roemer W, Hoek G, Brunekreef B. 1993. Effect of ambient winter air pollution on respiratory health of
children with chronic respiratory symptoms. Am Rev Respir Dis 147:118-24.
Roger LJ, Kehrl HR, Hazucha M, Horstman DH. 1985. Bronchoconstriction in asthmatics exposed to
sulfur dioxide during repeated exercise. J Appl Physiol 59:784-91.
May 2008
R-17
DRAFT—DO NOT QUOTE OR CITE
-------
Romieu I, Meneses F, Ruiz S, Sienra JJ, Huerta J, White MC, Etzel RA. 1996. Effects of air pollution on
the respiratory health of asthmatic children living in Mexico City. Am. J. Respir. Crit. Care Med.
154:300-307.
Romieu I, Ramirez-Aguilar M, Moreno-Macias H, Barraza-Villarreal A, Miller P, Hernandez-Cadena L,
Carbajal-Arroyo LA, Hernandez-Avila M. 2004. Infant mortality and air pollution: modifying effect
by social class. J Occup Environ Med 46:1210-6.
Rothman KJ, Greenland S. 1998. Modern epidemiology. Rothman KJ, Greenland S, editors. Philadelphia,
PA: Lippincott-Raven Publishers.
Ruan A, Min H, Meng Z, Lu Z. 2003. Protective effects of seabuckthorn seed oil on mouse injury induced
by sulfur dioxide inhalation. Inhal Toxicol 15:1053-1058.
Rusznak C, Devalia JL, Davies RJ. 1996. Airway response of asthmatic subjects to inhaled allergen after
exposure to pollutants. Thorax 51:1105-8.
SagivSK, Mendola P, Loomis D, Herring AH, Neas LM, Savitz DA, Poole C. 2005. A time-series analysis
of air pollution and preterm birth in Pennsylvania, 1997-2001. Environ Health Perspect 113:602-6.
Samet JM, Dominici F, Zeger SL, Schwartz J, Dockery DW. 2000. National morbidity, mortality, and air
pollution study. Part I: methods and methodologic issues. Cambridge, MA: Health Effects
Institute; research report no. 94.
Samoli E, Schwartz J; Analitis A; Petasakis Y; Wojtyniak B; Touloumi G; Spix C; Balducci F; Medina S;
Rossi G; Sunyer J; Anderson HR; Katsouyanni K. 2003. Sensitivity analyses of regional
differences in short-term effects of air pollution on daily mortality in APHEA cities. In: Revised
analyses of time-series studies of air pollution and health. Special report. Boston, MA: Health
Effects Institute; pp. 205-210. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf.
Sarnat JA, Brown KW, Schwartz J, Coull BA, Koutrakis P. 2005. Ambient gas concentrations and personal
particulate matter exposures: implications for studying the health effects of particles.
Epidemiology 16:385-95.
Sarnat JA, Koutrakis P, Suh HH. 2000. Assessing the relationship between personal particulate and
gaseous exposures of senior citizens living in Baltimore, MD. J Air Waste Manag Assoc 50:1184-
98.
Sarnat JA, Schwartz J, Catalano PJ, Suh HH. 2001. Gaseous pollutants in particulate matter
epidemiology: confounders or surrogates? Environ Health Perspect 109:1053-61.
Sarnat SE, Coull BA, Schwartz J, Gold DR, Suh HH. 2006. Factors affecting the association between
ambient concentrations and personal exposures to particles and gases. Environ Health Perspect
114:649-654.
Scanlon PD, Seltzer J, Ingram RH, Jr., Reid L, Drazen JM. 1987. Chronic exposure to sulfur dioxide.
Physiologic and histologic evaluation of dogs exposed to 50 or 15 ppm. Am Rev Respir Dis
135:831-9.
Schachter EN, WitekTJ, Jr., Beck GJ, Hosein HB, Colice G, Leaderer BP, Cain W. 1984. Airway effects of
low concentrations of sulfur dioxide: dose-response characteristics. Arch Environ Health 39:34-
42.
Schildcrout JS, Sheppard L, Lumley T, Slaughter JC, Koenig JQ, Shapiro GG. 2006. Ambient Air Pollution
and Asthma Exacerbations in Children: An Eight-City Analysis. Am. J. Epidemiol. 164:505-517.
Schouten JP, Vonk JM, de GraafA. 1996. Short term effects of air pollution on emergency hospital
admissions for respiratory disease: results of the APHEA project in two major cities in The
Netherlands, 1977-89. In: The APHEA project. Short term effects of air pollution on health: a
European approach using epidemiological time series data. J. Epidemiol. Community Health
50:S22-S29.
Schulz H, EderG, Heilmann P, Karg E, Meyer T, SchulzA, Ziesenis A, Heyder J. 1999. Health effects of
sulfur-related environmental air pollution. IV. Respiratory lung functions. Inhal Toxicol 11:423-438.
May 2008
R-18
DRAFT—DO NOT QUOTE OR CITE
-------
Schulz H, EderG, Heilmann P, Ruprecht L, Schumann G, Takenaka S, J. H. 1992. Early response of the
canine respiratory tract following long-term exposure to a sulfur (IV) aerosol at low concentration.
IV. Respiratory lung function. Inhal Toxicol 4:235-246.
Schwartz J, Dockery DW, Neas LM, Wypij 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
RespirCrit Care Med 150:1234-42.
Schwartz J. 1989. Lung function and chronic exposure to air pollution: a cross-sectional analysis of
NHANES II. Environ Res 50:309-21.
Schwartz J. 1995. Short term fluctuations in air pollution and hospital admissions of the elderly for
respiratory disease. Thorax 50:531-538.
Schwartz J. 2001. Air pollution and blood markers of cardiovascular risk. Environ Health Perspect 109
Suppl 3:405-9.
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. EurRespir J 11:677-685.
Seinfeld JH, Pandis SN. 1998. Atmospheric chemistry and physics: from air pollution to climate change.
New York, NY: John Wiley & Sons, Inc.
Sheppard L, Slaughter JC, Schildcrout J, Liu LJ, Lumley T. 2005. Exposure and measurement
contributions to estimates of acute air pollution effects. J Expo Anal Environ Epidemiol 15:366-76.
Smith E. 1994. Subject data supplied by the researchers for the recent controlled human studies analyzed
in the staff paper supplement and accompanying memorandum. Memorandum to Docket No. A-
84-25, Item IV-B-5.
Smith LG, Busch RH, Buschbom RL, Cannon WC, Loscutoff SM, Morris JE. 1989. Effects of sulfur dioxide
or ammonium sulfate exposure, alone or combined, for 4 or 8 months on normal and elastase-
impaired rats. Environ Res 49:60-78.
Snashall PD, Baldwin C. 1982. Mechanisms of sulphur dioxide induced bronchoconstriction in normal and
asthmatic man. British Medical Journal 37:118.
Soyseth V, Kongerud J, Broen P, Lilleng P, Boe J. 1995. Bronchial responsiveness, eosinophilia, and
short term exposure to air pollution. Arch Dis Child 73:418-22.
Speizer FE, Frank NR. 1966. The uptake and release of S02 by the human nose. Arch Environ Health
12:725-8.
Spengler JD, Ferris BG, Dockery DW, Speizer FE. 1979. Sulfur dioxide and nitrogen dioxide levels inside
and outside homes and the implications on health effects research. Environ. Sci. Technol. ;
Vol/lssue: 13:10:Pages: 1276-1280.
Spix C, Anderson HR, Schwartz J, Vigotti MA, LeTertre A, Vonk JM, Touloumi G, Balducci F, Piekarski T,
Bacharova L et al. 1998. Short-term effects of air pollution on hospital admissions of respiratory
diseases in Europe: a quantitative summary of APHEA study results. Archives of Environmental
and Occupational Health 53:54-64.
Stempel DA, Boucher RC. 1981. Respiratory infection and airway reactivity. Med Clin North Am 65:1045-
53.
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 52:470-84.
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 53:258-61.
May 2008
R-19
DRAFT—DO NOT QUOTE OR CITE
-------
Stock TH, Kotchmar DJ, Contant CF, Buffler PA, Holguin AH, Gehan BM, Noel LM. 1985. The estimation
of personal exposures to air pollutants for a community-based study of health effects in
asthmatics-design and results of air monitoring. Journal of the Air and Waste Management
Association 35:1266-1273.
Strandberg LG. 1964. S02 absorption in the respiratory tract: studies on the absorption in rabbit, its
dependence on concentration and breathing phase. Arch. Environ. Health 9:160-166.
Studnicka M, Hackl E, Pischinger J, FangmeyerC, Haschke N, Kuhr J, Urbanek R, Neumann M, Frischer
T. 1997. Traffic-related N02 and the prevalence of asthma and respiratory symptoms in seven
year olds. EurRespir J 10:2275-8.
Sullivan J, Sheppard L, Schreuder A, Ishikawa N, Siscovick D, Kaufman J. 2005. Relation between short-
term fine-particulate matter exposure and onset of myocardial infarction. Epidemiology 16:41-8.
Sunyer J, SpixC, Quenel P, Ponce-de-Leon A, Ponka A, Barumandzadeh T, Touloumi G, Bacharova L,
Wojtyniak B, Vonk J et al. 1997. Urban air pollution and emergency admissions for asthma in four
European cities: the APHEA Project. Thorax 52:760-5.
Surgeon General USS. 2004. The health consequences of smoking: a report of the Surgeon General.
Rockville, MD: US Department of Health and Human Services, Office of the Surgeon General.
Taggart SCO, Custovic A, Francis HC, Faragher EB, Yates CJ, Higgins BG, Woodcock A. 1996. Asthmatic
bronchial hyperresponsiveness varies with ambient levels of summertime air pollution. Eur Respir
9:1146-1154.
Takenaka S, Furst G, Heilman P, Heini A, Heinzmann U, Kreyling WG, Murray AB, Schulz H, Heyder J.
1992. Early response of the canine respiratory tract following long-term exposure to a sulfur (IV)
aerosol at low concentration. V. Morphology and morphometry. Inhal Toxicol 4:247-272.
Takenaka S, Godleski JJ, Heini A, Karg E, Kreyling WG, RitterB, Schulz H, ZiesenisA, Heyder J. 1999.
Health effects of sulfur-related environmental air pollution. V. Lung structure. Inhal Toxicol 11:439-
54.
Tan WC, Cripps E, Douglas N, Sudlow MF. 1982. Protective effect of drugs on bronchoconstriction
induced by sulphur dioxide. Thorax 37:671-5.
Tenias JM, Ballester F, Rivera ML. 1998. Association between hospital emergency visits for asthma and
air pollution in Valencia, Spain. Occup Environ Med 55:541-7.
Thompson AJ, Shields MD, Patterson CC. 2001. Acute asthma exacerbations and air pollutants in
children living in Belfast, Northern Ireland. Arch Environ Health 56:234-41.
Thornton DCB, A. R. 1999. Sulfur dioxide distribution over the Pacific Ocean 1991 -1996. J. Geophys.
Res. [Atmos.] 104:5845-5854.
Thornton DCB, A. R.; Tu, F. H.; Blomquist, B. W.; Mitchell, G. M.; Nadler, W.; Lenschow, D. H. 2002. Fast
airborne sulfur dioxide measurements by atmospheric pressure ionization mass spectrometry
(APIMS). J. Geophys. Res. [Atmos.] 107.
Timonen KL, Pekkanen J. 1997. Air pollution and respiratory health among children with asthmatic or
cough symptoms. American Journal of Respiratory & Critical Care Medicine 156:546-552.
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
Suppl 2:S29-35.
Trenga CA, Koenig JQ, Wlliams PV. 1999. Sulphur dioxide sensitivity and plasma antioxidants in adult
subjects with asthma. Occup Environ Med 56:544-7.
Trenga CA, Koenig JQ, Wlliams PV. 2001. Dietary antioxidants and ozone-induced bronchial
hyperresponsiveness in adults with asthma. Arch Environ Health 56:242-9.
May 2008
R-20
DRAFT—DO NOT QUOTE OR CITE
-------
Triche EW, Belanger K, Bracken MB, Beckett WS, Holford TR, Gent JF, McSharry J-E, Leaderer BP.
2005. Indoor heating sources and respiratory symptoms in nonsmoking women. Epidemiology
16:377-384.
Tunnicliffe WS, Hilton MF, Harrison RM, Ayres JG. 2001. The effect of sulphur dioxide exposure on
indices of heart rate variability in normal and asthmatic adults. Eur Respir 17:604-608.
U.S. Code. 2003a. Clean Air Act, §109, national ambient air quality standards. U.S.C. 42: §7408.
U.S. Code. 2003b. Clean Air Act, §108, air quality criteria and control techniques. U.S.C. 42: §7409.
U.S. Court of Appeals for the District of Columbia. 1980. Lead Industries v. U.S. Environmental Protection
Agency. 647 F2d 1130, 1154 (DC Cir. 1980).
U.S. Court of Appeals for the District of Columbia. 1981. American Petroleum Institute v. Costle. 665 F2d
1176, 1186 (DC Cir. 1981).
U.S. Environmental Protection Agency. 2004. Air quality criteria for particulate matter. In: Agency USEP,
editor. 2v ed. Research Triangle Park, NC: National Center for Environmental Assessment.
U.S. Environmental Protection Agency. 2006b. Technology Transfer Network clearinghouse for
inventories and emission factors. In: Agency USEP, editor.
U.S. Environmental Protection Agency. 2006d. Acid rain program (ARP) 2005 progress report. In: Agency
USEP, editor. Washington, DC: Office of Air and Radiation, Clean Air Markets Division.
U.S. Environmental Protection Agency. 1982. Air quality criteria for particulate matter and sulfur oxides. In:
Agency USEP, editor. Research Triangle Park, NC: Office of Health and Environmental
Assessment,
U.S. Environmental Protection Agency. 1986a.eview of the national ambient air quality standards for
sulfur oxides: updated assessment of scientific and technical information, addendum to the 1982
OAQPS staff paper. Research Triangle Park, NC: Office of Air Quality Planning and Standards,
Strategies and Air Standards Division; report no. EPA/450/5-86-13. Available from: NTIS,
Springfield, VA; PB87-200259/XAB.
U.S. Environmental Protection Agency. 1986b. Second addendum to air quality criteria for particulate
matter and sulfur oxides (1982): assessment of newly available health effects information. In:
Agency USEP, editor. Research Triangle Park, NC: Office of Health and Environmental
Assessment,
U.S. Environmental Protection Agency. 1994a. Supplement to the second addendum (1986) to air quality
criteria for particulate matter and sulfur oxides (1982). In: Agency USEP, editor, report no. EPA-
600/FP-93/002 ed. Research Triangle Park, NC: Office of Health and Environmental Assessment,
Environmental Criteria and Assessment Office.
U.S. Environmental Protection Agency. 1994b. Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry. In: Environmental Criteria and Assessment
Office OoHaEA, Office of Research and Development, editor. Research Triangle Park, NC.
U.S. Environmental Protection Agency. 1994c. Review of the national ambient air quality standards for
sulfur oxides: assessment of scientific and technical information, supplement to the 1986 OAQPS
staff paper addendum. Research Triangle Park, NC: Office of Air Quality Planning and Standards,
Air Quality Management Division; report no. EPA/452/R-94/013.
U.S. Environmental Protection Agency. 1996. Air quality criteria for particulate matter. In: Agency USEP,
editor. 3v ed. Research Triangle Park, NC: National Center for Environmental Assessment.
U.S. Environmental Protection Agency. 2005 Technical assistance document (TAD) for precursor gas
measurements in the NCore multi-pollutant monitoring network. Version 4. Research Triangle
Park, NC: Office of Air Quality Planning and Standards; report no. EPA-454/R-05-003.
May 2008
R-21
DRAFT—DO NOT QUOTE OR CITE
-------
U.S. Environmental Protection Agency. 2006a. Air quality criteria for ozone and related photochemical
oxidants. In: Agency USEP, editor. 3v ed. Research Triangle Park, NC: National Center for
Environmental Assessment.
U.S. Environmental Protection Agency. 2006c. Report on air quality in nonattainment areas for 2003-2005
covering ozone, particulate matter, carbon monoxide, sulfur dioxide, nitrogen dioxide, and lead.
In: Agency USEP, editor. Research Triangle Park, NC: Office of Air Quality Planning and
Standards, Air Quality Assessment Division, Air Quality Analysis Group.
U.S. Environmental Protection Agency. 2007. S02 NAAQS Review Plan - Final. In: Agency USEP, editor.
Research Triangle Park, NC: National Center for Environmental Assessment.
U.S. EPA 2005. Guidelines for Carcinogen Risk Assessment. EPA/630/P-03/001F. Washington, DC:Risk
Assessment Forum, U.S. Environmental Protection Agency. Available:
http://oaspub.epa.gov/eims/eimscomm.getfile?p_download_id=439797/.
U.S. EPA. 2005. Guidelines for Carcinogen Risk Assessment. EPA/630/P-03/001F. Washington, D.C.
U.S. EPA. 2006. Acid rain program (ARP) 2005 progress report. In: Office of Air and Radiation. CAMD,
editor. Washington, DC.
U.S. Senate. 1970.National Air Quality Standards Act of 1970: report of the Committee on Public Works,
United States Senate together with individual views to accompany S. 4358. Washington, DC:
Committee on Public Works; report no. CONG/91-1196.
U.S. Supreme Court. 2001. Whitman v.American Trucking Association. 531 U.S. 457 (nos. 99-1257 and
99-1426).
Van Aardenne JA, Dentener FJ, Olivier JGJ, Klein Goldewijk CGM, Lelieveld J. 2001. A 1 degree x
1 degree resolution data set of historical anthropogenic trace gas emissions for the period 1980-
1990. Global Biogeochem Cycles 15:909-928.
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.
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 15:700-9.
van Noije TPC, Eskes HJ, Dentener FJ, Stevenson DS, Ellingsen K, Schultz MG, Wild O, Amann M,
Atherton CS, Bergmann DJ et al. 2006. Multi-model ensemble simulations of tropospheric N02
compared with GOME retrievals for the year 2000. Atmosph and Phys 6:2943-2979.
Vedal SR, K.; Brauer, M.; White, R.; Petkau, J. 2004. Air pollution and cardiac arrhythmias in patients with
implantable cardiovascular defibrillators. Inhal Toxicol. 16:353-362.
Vutukuru S, Dabdub D. 2008. Modeling the effects of ship emissions on coastal air quality: A case study
of southern California. Atmosph Environ In Press, Corrected Proof.
Wang AL, Blackford TL, Lee LY. 1996. Vagal bronchopulmonary C-fibers and acute ventilatory response
to inhaled irritants. Respir Physiol 104:231-9.
Wang C, Corbett JJ, Firestone J. 2007. Modeling energy use and emissions from north american
shipping: application of the ship traffic, energy, and environment model. Environ Sci Technol
41:3226-3232.
Wang X, Ding H, Ryan L, Xu X. 1997. Association between air pollution and low birth weight: a
community-based study. Environ Health Perspect 105:514-20.
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.
May 2008
R-22
DRAFT—DO NOT QUOTE OR CITE
-------
Ware JH, Ferris BG, Jr., Dockery DW, Spengler JD, Stram DO, Speizer FE. 1986. Effects of ambient sulfur
oxides and suspended particles on respiratory health of preadolescent children. Am Rev Respir
Dis 133:834-42.
Warneck P. 1999. The relative importance of various pathways for the oxidation of sulfur dioxide and
nitrogen dioxide in sunlit continental fair weather clouds. Phys. Chem. Chem. Phys. 1:5471 -5483.
Wellenius GA, Bateson TF, Mittleman MA, Schwartz J. 2005a. Particulate air pollution and the rate of
hospitalization for congestive heart failure among medicare beneficiaries in Pittsburgh,
Pennsylvania. Am J Epidemiol 161:1030-6.
Wellenius GA, Schwartz J, Mittleman MA. 2005b. Air pollution and hospital admissions for ischemic and
hemorrhagic stroke among medicare beneficiaries. Stroke 36:2549-53.
Wddicombe J, Lee LY. 2001. Airway reflexes, autonomic function, and cardiovascular responses. Environ
Health Perspect 109 Suppl 4:579-84.
Wddicombe JG. 2003. Overview of neural pathways in allergy and asthma. Pulm Pharmacol Ther 16:23-
30.
Wley JAR, J. P.; Cheng, Y.-T.; Piazza, T.; Stork, L.; Pladsen, K. 1991b. Study of children's activity
patterns: final report. In: Board CAR, editor. Sacramento, CA.
Wley JAR, J. P.; Piazza, T.; Garrett, K.; Cirksena, K.; Cheng, Y.-T.; Martin, G. 1991a. Activity patterns of
California residents. Final report. In: Board CAR, editor. Sacramento, CA: Available from: NTIS,
Springfield, VA.
Wilis A, Jerrett M, Burnett RT, Krewski D. 2003. The association between sulfate air pollution and
mortality at the county scale: an exploration of the impact of scale on a long-term exposure study.
J Toxicol Environ Health A 66:1605-24.
Wlson AM, Wake CP, Kelly T, Salloway JC. 2005. Air pollution, weather, and respiratory emergency room
visits in two northern New England cities: an ecological time-series study. Environ Res 97:312-21.
Wlson WE, Brauer M. 2006. Estimation of ambient and non-ambient components of particulate matter
exposure from a personal monitoring panel study. J Expo Sci Environ Epidemiol 16:264-74.
Wnterton DL, Kaufman J, Keener CV, Quigley S, Farin FM, Williams PV, Koenig JQ. 2001. Genetic
polymorphisms as biomarkers of sensitivity to inhaled sulfur dioxide in subjects with asthma. Ann.
Allergy Asthma Immunol. 86:232-238.
Wong C-M, Atkinson RW, Anderson HR, Hedley AJ, Ma S, Chau PY-K, Lam T-H. 2002. A tale of two
cities: effects of air pollution on hospital admissions in Hong Kong and London compared. Environ
Health Perspect 110:67-77.
Woodruff TJ, Parker JD, Kyle AD, Schoendorf KC. 2003. Disparities in exposure to air pollution during
pregnancy. Environ Health Perspect 111:942-6.
Wright AL, Hons FM, Lemon RG, McFarland ML, Nichols RL. 2007. Stratification of nutrients in soil for
different tillage regimes and cotton rotations. Soil & tillage research. 96:19-27.
Wu D, Meng Z. 2003. Effect of sulfur dioxide inhalation on the glutathione redox system in mice and
protective role of sea buckthorn seed oil. Arch Environ Contam Toxicol 45:423-8.
Xu X, Ding H, Wang X. 1995. Acute effects of total suspended particles and sulfur dioxides on preterm
delivery: a community-based cohort study. Arch Environ Health 50:407-15.
Xu XP, Dockery DW, Wang LH. 1991. Effects of air pollution on adult pulmonary function. Arch Environ
Health 46:198-206.
Yang CY, Tseng YT, Chang CC. 2003a. Effects of air pollution on birth weight among children born
between 1995 and 1997 in Kaohsiung, Taiwan. J Toxicol Environ Health A 66:807-16.
May 2008
R-23
DRAFT—DO NOT QUOTE OR CITE
-------
Yang Q, Chen Y, Shi Y, Burnett RT, McGrail KM, Krewski D. 2003b. Association between ozone and
respiratory admissions among children and the elderly in Vancouver, Canada. Inhal Toxicol
15:1297-308.
Yildirim Z, KilicT, Koksal N, Kotuk M. 2005. Protective effect of ipratropium bromide on
bronchoconstriction induced by sulfur dioxide exposure during apricot sufurization processes that
causes asthma-like syndrome in agricultural environment. Pharmacol Res 51:479-82.
Zanobetti A, Schwartz J, Gold D. 2000. Are there sensitive subgroups for the effects of airborne particles?
Environ Health Perspect 108:841-5.
Zanobetti A, Schwartz J. 2001. Are diabetics more susceptible to the health effects of airborne particles?
Am J Respir Crit Care Med 164:831-3.
Zanobetti A, Schwartz J. 2002. Cardiovascular damage by airborne particles: are diabetics more
susceptible? Epidemiology 13:588-92.
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? Official statement of the American
Thoracic Society: concepts and consequences. Environ Health Perspect. 108:877-685.
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 108:419-26.
Zhang JZ, Millero FJ. 1991. The rate of sulfite oxidation in seawater. Geochimica et Cosmochimica Acta
55:677-685.
Zmirou D, Schwartz J, Saez M, Zanobetti A, Wojtyniak B, Touloumi G, Spix C, Ponce de Leon A, Le
Moullec Y, Bacharova L et al. 1998. Time-series analysis of air pollution and cause-specific
mortality. Epidemiology 9:495-503.
May 2008
R-24
DRAFT—DO NOT QUOTE OR CITE
------- |