£EPA
United States September 2007
Environmental Protection EPA/600/R-07/108
Integrated Science Assessment
for Sulfur Oxides -
Health Criteria
(First External Review Draft)
-------
September 2007
EPA/600/R-07/108
Integrated Science Assessment
for Sulfur Oxides - Health Criteria
National Center for Environmental Assessment-RTF Division
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
DISCLAIMER
This document is a first 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.
PREFACE
Legislative Requirements
Two sections of the Clean Air Act (CAA) govern the establishment and revision of the
national ambient air quality standards (NAAQS). Section 108 (U.S. Code, 2003a) directs the
Administrator to identify and list "air pollutants" that "in his judgment, may reasonably be
anticipated to endanger public health and welfare" and whose "presence ... in the ambient air
results from numerous or diverse mobile or stationary sources" and to issue air quality criteria
for those that are listed. Air quality criteria are intended to "accurately reflect the latest scientific
knowledge useful in indicating the kind and extent of identifiable effects on public health or
welfare which may be expected from the presence of [a] pollutant in ambient air ...."
Section 109 (U.S. Code, 2003b) directs the Administrator to propose and promulgate
"primary" and "secondary" NAAQS for pollutants listed under section 108. Section 109(b)(l)
defines a primary standard as one "the attainment and maintenance of which in the judgment of
the Administrator, based on such criteria and allowing an adequate margin of safety, are requisite
to protect the public health."1 A secondary standard, as defined in section 109(b)(2), must
"specify a level of air quality the attainment and maintenance of which, in the judgment of the
Administrator, based on such criteria, is 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 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" [U.S. Senate (1970)].
Welfare effects as defined in section 302(h) [U.S. Code, 2005] 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."
September 2007 ii DRAFT-DO NOT QUOTE OR CITE
-------
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. See 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 (1981). 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 U.S. Environmental Protection Agency (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.
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,
465-472, 475-76 (2001).
Section 109(d)(l) requires that "not later than December 31, 1980, and at 5-year
intervals thereafter, the Administrator shall complete a thorough review of the criteria
published under section 108 and the national ambient air quality standards ... and shall make
such revisions in such criteria and standards and promulgate such new standards as may be
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
September 2007 iii DRAFT-DO NOT QUOTE OR CITE
-------
independent review function has been performed by the Clean Air Scientific Advisory
Committee (CASAC) of EPA's Science Advisory Board.
History of Reviews of the Primary NAAQS for Sulfur Dioxide
On April 30, 1971, the EPA promulgated primary NAAQS for sulfur dioxide (SO2).
These primary standards, which were based on the findings outlined in the original 1969 Air
Quality Criteria (hereafter "AQCD") for Sulfur Oxides (U.S. DHEW, 1969), 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. In 1982, EPA published the AQCD for
Particulate Matter (PM) and Sulfur Oxides along with an addendum of newly published
controlled human exposure studies (U.S. Environmental Protection Agency, 1982), 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 (U.S. Environmental Protection Agency, 1986). In 1988,
EPA reviewed and revised the health criteria upon which the SO2 standards were based. As a
result of that review, EPA published a proposed decision not to revise the existing standards
(Federal Register, 1988). However, EPA specifically requested public comment on the
alternative of revising the current standards and adding a new 1-h 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 (Federal Register,
1994). The 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 SO2 exposures in asthmatics
(U.S. Environmental Protection Agency, 1994). As in the 1988 proposal, EPA proposed to
retain the existing 24-h and annual standards. The EPA also solicited comment on three
regulatory alternatives to further reduce the health risk posed by exposure to high 5-min peaks of
SO2 if additional protection were judged to be necessary. The three alternatives included (1)
revising the existing primary SO2 NAAQS by adding a new 5-min standard of 0.60 ppm SO2; (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 SO2; and (3) augmenting
implementation of existing standards by focusing on those sources or source types likely to
produce high 5-min peak concentrations of SO2. On May 22, 1996, EPA's final decision, that
revisions of the NAAQS for sulfur oxides were not appropriate at that time, was announced in
September 2007 iv DRAFT-DO NOT QUOTE OR CITE
-------
the Federal Register (Federal Register, 1996). 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 SO2.
September 2007 v DRAFT-DO NOT QUOTE OR CITE
-------
TABLE OF CONTENTS
Page
List of Tables ix
List of Figures xi
Authors, Contributors, and Reviewers xiii
U.S. Environmental Protection Agency Project Team xviii
U.S. Environmental Protection Agency Science Advisory Board (SAB)
Staff Office Clean Air Scientific Advisory Committee (CASAC) xxi
Abbreviations and Acronyms xxiv
1. INTRODUCTION 1-1
1.1 DOCUMENT DEVELOPMENT 1-2
1.2 ORGANIZATION OF THE DOCUMENT 1-5
2. SOURCE-TO-TISSUEDOSE 2-1
2.1 ATMOSPHERIC CHEMISTRY 2-1
2.2 SOURCES OF SULFUR OXIDES 2-3
2.3 MEASUREMENT METHODS AND ASSOCIATED ISSUES 2-5
2.4 ENVIRONMENTAL CONCENTRATIONS OF SULFUR OXIDES 2-8
2.4.1 Ambient Air Quality Data for Sulfur Dioxide and
Other Sulfur Oxides 2-8
2.4.2 Spatial and Temporal Variability of Ambient Sulfur Dioxide
Concentrations 2-12
2.4.3 Policy Relevant Background Concentrations of Sulfur Dioxide 2-24
2.5 ISSUES ASSOCIATED WITH EVALUATING EXPOSURES
TO SULFUR OXIDES 2-27
2.5.1 General Considerations for Personal Exposures 2-27
2.5.2 Methods Used for Monitoring Personal Exposure to SO2 2-31
2.5.3 Relationships between Personal Exposures and Ambient
Concentrations 2-33
2.5.4 Exposure Measurement Errors in Epidemiological Studies 2-37
2.6 DOSIMETRY OF INHALED SO2 2-42
3. INTEGRATED HEALTH EFFECTS OF EXPOSURE TO
SULFUR DIOXIDE 3-1
3.1 MORBIDITY ASSOCIATED WITH SHORT-TERM SO2
EXPOSURE 3-3
3.1.1 Respiratory Effects Associated with Short-Term
Exposure to SO2 3-3
3.1.2 Cardiovascular Effects Associated with Short-Term
SO2 Exposure 3-45
3.1.3 Other Systemic Effects Associated with Short-Term
SO2 Exposure 3-58
September 2007 vi DRAFT-DO NOT QUOTE OR CITE
-------
TABLE OF CONTENTS
(cont'd)
rage
3.2 MORTALITY ASSOCIATED WITH SHORT-TERM SO2
EXPOSURE 3-63
3.2.1 Associations of Mortality and Short-Term SO2 Exposure
inMulticity Studies and Meta-Analyses 3-64
3.2.2 Cause-Specific Mortality Associated with Short-Term
SO2 Exposure 3-75
3.2.3 Evidence from an Intervention Study 3-78
3.2.4 Summary of Effects of Short-Term SO2 Exposure on
Mortality 3-79
3.3 MORBIDITY ASSOCIATED WITH LONG-TERM SO2
EXPOSURE 3-80
3.3.1 Respiratory Effects Associated with Long-Term
Exposure to SO2 3-80
3.3.2 Carcinogenic Effects Associated with Long-Term
Exposure to SO2 3-86
3.3.3 Prenatal and Neonatal Outcomes Associated with
Long-Term SO2 Exposure 3-88
3.4 MORTALITY ASSOCIATED WITH LONG-TERM SO2
EXPOSURE 3-93
3.4.1 Associations of Mortality and Long-Term SO2 Exposure
in Key Studies 3-94
3.4.2 Summary of Effects of Long-Term SO2 Exposure
on Mortality 3-103
4. PUBLIC HEALTH IMPACT 4-1
4.1 ASSESSMENT OF CONCENTRATION-RESPONSE FUNCTION
AND POTENTIAL THRESHOLDS 4-1
4.2 SUSCEPTIBLE AND VULNERABLE POPULATIONS 4-6
4.2.1 Exposure of Susceptible and Vulnerable Populations to SO2 4-7
4.2.2 Preexisting Disease as a Potential Risk Factor 4-7
4.2.3 Age-Related Variations in Susceptibility 4-12
4.2.4 Genetic Factors for Oxidant and Inflammatory Damage
from Air Pollutants 4-14
4.3 POTENTIAL PUBLIC HEALTH IMPACTS 4-16
4.3.1 Concepts Related to Defining Adverse Health Effects 4-16
4.3.2 Estimation of Potential Numbers of Persons in At-Risk
Susceptible Population Groups in the United States 4-17
5. KEY FINDINGS AND CONCLUSIONS 5-1
September 2007 vii DRAFT-DO NOT QUOTE OR CITE
-------
TABLE OF CONTENTS
(cont'd)
rage
5.1 SUMMARY OF KEY FINDINGS RELATED TO THE
SOURCE-TO-DOSE RELATIONSHIP 5-1
5.1.1 Emission Sources, Atmospheric Science, and Ambient
Monitoring Methods 5-1
5.1.2 Ambient Concentrations 5-2
5.1.3 Exposure Assessment 5-3
5.1.4 Dosimetry 5-4
5.2 SUMMARY OF KEY HEALTH EFFECTS FINDINGS 5-4
5.2.1 Findings from the Previous Review of the NAAQS for SO2 5-4
5.2.2 New Findings on the Health Effects of Exposure to SO2 5-6
5.3 CONCLUSIONS 5-15
APPENDIX 5A. SUMMARY OF NEW ANIMAL TOXICOLOGICAL,
HUMAN CLINICAL, AND EPIDEMIOLOGICAL STUDIES
OF HEALTH EFFECTS ASSOCIATED WITH EXPOSURES
TO SULFUR DIOXIDE 5A-1
6. REFERENCES 6-1
September 2007 viii DRAFT-DO NOT QUOTE OR CITE
-------
LIST OF TABLES
Number Page
2.4-1. Regional Distribution of 862 and SC>42 Ambient Concentrations, Averaged
for 2003-2005 2-48
2.4-2. Distributions of Temporal Averaging Inside and Outside CMSAs 2-49
2.4-3. Range of mean 862 concentrations and Pearson correlation coefficients in
urban areas having at least four monitors 2-50
2.5-1. Relationships of Indoor to Outdoor 862 Concentrations 2-51
3.1-1. SO2 Effects on Guinea Pig and Rat Brain 3-61
4.3-1. Gradation of Individual Responses to Short-Term SC>2 Exposure in
Individuals with Impaired Respiratory Systems 4-20
4.3-2. Prevalence of Selected Respiratory Disorders by Age Group and by
Geographic Region in the United States (2004 [U.S. Adults] and 2005
[U.S. Children] National Health Interview Survey) 4-21
5A-1. Key Respiratory Health Effects of Exposure to Sulfur Dioxide Observed in
Animal Toxicological Studies 5A-2
5A-2. Key Human Health Effects of Peak Exposure to Sulfur Dioxide Observed in
Clinical Studies 5A-4
5A-3. Effects Of Short-Term Exposure To Sulfur Dioxide On Respiratory Symptoms
Among Children 5A-5
5A-4. Effects of Short-Term SO2 Exposure On Emergency Department Visits
And Hospital Admissions For Respiratory Outcomes 5A-10
September 2007 ix DRAFT-DO NOT QUOTE OR CITE
-------
LIST OF FIGURES
Number Page
2.4-1. State-level SO2 emissions, 1990-2005 2-9
2.4-2. Annual mean ambient 862 concentration, (a) 1989 through 1991, and (b)
2003 through 2005 2-10
2.4-3. Annual mean ambient SC>42 Concentration, (a) 1989 through 1991, and
(b) 2003 through 2005 2-11
2.4-4. Annual SC>2 Emissions in 2006 for Acid Rain Program Cooperating
Facilities 2-12
2.4-5. Boxplot of hourly 862 concentrations across all cities in focus 2-13
2.4-6(a). Monthly mean, minimum, and maximum SC>2 concentrations at Steubenville,
OH for the years 2003 through 2005 2-15
2.4-6(b). Monthly mean, minimum, and maximum SC>42 concentrations at Steubenville,
OH for the years 2003 through 2005 2-16
2.4-6(c). Monthly mean SO42 concentrations as a function of SO2 concentrations at
Steubenville, OH for the years 2003 through 2005 2-16
2.4-7(a). Monthly mean, minimum, and maximum SO2 concentrations at Philadelphia,
PA for the years 2003 through 2005 2-17
2.4-7(b). Monthly mean, minimum, and maximum SO42 concentrations at Philadelphia,
PA for the years 2003 through 2005 2-17
2.4-7(c). Monthly mean SO42 concentrations as a function of SO2 concentrations at
Philadelphia, PA for the years 2003 through 2005 2-18
2.4-8(a). Monthly mean, minimum, and maximum SO2 concentrations at Los Angeles,
CA for the years 2003 through 2005 2-18
2.4-8(b). Monthly mean, minimum, and maximum SO42 concentrations at
Los Angeles, CA for the years 2003 through 2005 2-19
2.4-8(c). Monthly mean SO42 concentrations as a function of SO2 concentrations at
Los Angeles, CA for the years 2003 through 2005 2-19
2.4-9(a). Monthly mean, minimum, and maximum SO2 concentrations at Riverside,
CA for the years 2003 through 2005 2-20
2.4-9(b). Monthly mean, minimum, and maximum SO42 concentrations at Riverside,
CA for the years 2003 through 2005 2-20
2.4-9(c). Monthly mean SO42 concentrations as a function of SO2 concentrations at
Riverside, CA for the years 2003 through 2005 2-21
2.4-10(a). Monthly mean, minimum, and maximum SO2 concentrations at Phoenix,
AZ for the years 2003 through 2005 2-21
September 2007 x DRAFT-DO NOT QUOTE OR CITE
-------
LIST OF FIGURES
(cont'd)
Number Page
2.4-10(b). Monthly mean, minimum, and maximum SC>42 concentrations at Phoenix,
AZ for the years 2003 through 2005 2-22
2.4-10(c). Monthly mean SC>42 concentrations as a function of SC>2 concentrations at
Phoenix, AZ for the years 2003 through 2005 2-22
2.4-11. Annual mean model-predicted concentrations of SO2 (ppb) in surface air
over the United States in the present-day (upper panel) and policy relevant
background (middle panel) MOZART-2 simulations 2-26
2.5-1. Percentage of time people spend in different environments in the United
States as determined by the National Human Activity Pattern Survey
(NHAPS) 2-28
2.5-2. Average annual indoor and outdoor SC>2 concentrations for each of the six
cities included in the analysis 2-34
3.1-1. Odds ratios for daily asthma symptoms associated with a 10-ppb
increase in within-subject concentrations of 24-h average
SO2, using data collected from November 1993 to September 1995 3-7
3.1-2. Relative odds ratio of incidence of lower respiratory symptoms
smoothed against 24-h average 862 concentrations on the previous
day, controlling for temperature, city, and day of week 3-8
3.1-3. Odds ratios (95% CI) for the incidence of cough among children,
grouped by season 3-10
3.1-4. Odds ratios (95% CI) for the incidence of lower respiratory or
asthma symptoms among children, grouped by season 3-11
3.1-5. Specific airways resistance (sRaw) of 16 mild and 24 moderate
asthmatic subjects exposed to 0-, 0.4-, and 0.6-ppm SO2 3-20
3.1-6. Distribution of individual airway sensitivity to SO2 3-21
3.1-7. Relative risks (95% CI) of SO2-associated emergency department
visits (*) and hospitalizations for all respiratory causes among all ages 3-32
3.1-9. Relative risks (95% CI) of SO2-associated emergency department
visits (*) and hospitalizations for asthma among all ages 3-36
3.1-10. Relative risks (95% CI) of SO2-associated emergency department
visits (*) and hospitalizations for asthma, stratified by age groups 3-37
3.1-12. Relative risks (95% CI) of SO2-associated emergency department
visits and hospitalizations for all cardiovascular causes 3-54
September 2007 xi DRAFT-DO NOT QUOTE OR CITE
-------
LIST OF FIGURES
(cont'd)
Number Page
3.1-13. Relative risks (95% CI) of SO2-associated emergency department
visits (*) and hospitalizations for cardiovascular causes, with and
without copollutant adjustment 3-59
3.2-1. Posterior means and 95% posterior intervals of national average
estimates of 862 effects on total mortality from non-external causes
per 10-ppb increase in 24-h average SC>2 at 0-, 1-, and 2-day lags
within sets of the 62 cities with pollutant data available 3-66
3.2-2. All cause (nonaccidental) SC>2 mortality risk estimates (95% CI)
from multicity and meta-analysis studies 3-73
3.2-3. All-cause (nonaccidental) and broad cause-specific (respiratory and
cardiovascular) SC>2 mortality risk estimates (95% CI) from
multicity studies 3-76
3.3-1. Relative risks (95% CI) for low birth weight, grouped by trimester
of SC>2 exposure 3-90
3.4-1. Total SO2-mortality relative risk estimates (95% CI) from
longitudinal cohort studies 3-102
4.2-1. Relative risks (95% CI) of age-specific associations between short-term
exposure to SC>2 and respiratory ED visits and hospitalizations 4-13
September 2007 xii DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
Authors
Dr. Jee Young Kim (SOX Team Leader)—National Center for Environmental Assessment (B243-
01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Jeffrey Arnold—National Center for Environmental Assessment (B243-01), U.S.
Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. James S. Brown—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Douglas Bryant—Intrinsik Science, 1900 Minnesota Court, Mississauga, Ontario L8S IPS
Dr. Ila Cote—National Center for Environmental Assessment (B243-01), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Dr. Arlene Fiore—Geophysical Fluid Dynamics Laboratory/National Oceanographic &
Atmospheric Administration, 201 Forrestal Rd., Princeton, NJ 08542-0308
Dr. Panos Georgopoulos—Computational Chemodynamics Laboratory, EOHSI Room 308, 170
Frelinghuysen Road, Piscataway, New Jersey 08854
Dr. Brett Grover—National Exposure Research Laboratory (D205-03), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Dr. Vic Hasselblad—Duke University Medical Center, Box 17969, Durham, NC 27715
Dr. Larry Horowitz—Geophysical Fluid Dynamics Lab oratory/National Oceanographic &
Atmospheric Administration, Princeton University Forrestal Campus, 201 Forrestal Road,
Princeton, NJ 08540-5063
Dr. Annette lanucci—Sciences International, 1800 Diagonal Road, Suite 500, Alexandria, VA
22314
Dr. Kazuhiko Ito—New York University School of Medicine, 57 Old Forge Road, Tuxedo, NY
10987
Dr. Doug Johns—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Ellen Kirrane—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
September 2007 xiii DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Authors
Dr. Jane Koenig—University of Washington, Department of Environmental and Occupational
Health Sciences, Box 357234, Seattle, WA 98195-7234
Dr. Thomas Long—National Center for Environmental Assessment (B243-01), U.S.
Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Thomas Luben—National Center for Environmental Assessment (B243-01), U.S.
Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Therese Mar—University of Washington, Department of Environmental and Occupational
Health Sciences, Box 357234, Seattle, WA 98195-7234
Dr. Qingyu Meng—National Center for Environmental Assessment (B243-01), U.S.
Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Anu Mudipalli—National Center for Environmental Assessment (B243-01), U.S.
Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Joseph Pinto—National Center for Environmental Assessment (B243-01), U.S.
Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Mary Ross—National Center for Environmental Assessment (B243-01), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Dr. David Svendsgaard—National Center for Environmental Assessment (B243-01), U.S.
Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lori White—National Center for Environmental Assessment (B243-01), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Dr. William Wilson—National Center for Environmental Assessment (B243-01), U.S.
Environmental Protection Agency, Research Triangle Park, NC 27711
September 2007 xiv DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors
Dr. Dale Allen, University of Maryland, College Park, MD
Ms. Louise Camalier, U.S. EPA, OAQPS, Research Triangle Park, NC
Dr. Russell Dickerson, University of Maryland, College Park, MD
Dr. Tina Fan, EOHSI/UMDNJ, Piscataway, NJ
Dr. William Keene, University of Virginia, Charlottesville, VA
Dr. Randall Martin, Dalhousie University, Halifax, Nova Scotia
Dr. Maria Morandi, University of Texas, Houston, TX
Dr. William Munger, Harvard University, Cambridge, MA
Mr. Charles Piety, University of Maryland, College Park, MD
Dr. Sandy Sillman, University of Michigan, Ann Arbor, MI
Dr. Helen Suh, Harvard University, Boston, MA
Dr. Charles Wechsler, EOHSI/UMDNJ, Piscataway, NJ
Dr. Clifford Weisel, EOHSI/UMDNJ, Piscataway, NJ
Dr. Jim Zhang, EOHSI/UMDNJ, Piscataway, NJ
Reviewers
Dr. Tina Bahadori—American Chemistry Council, 1300 Wilson Boulevard, Arlington, VA
22209
Dr. Tim Benner—Office of Science Policy, Office of Research and Development, Washington,
DC 20004
Dr. Daniel Costa—National Program Director for Air, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
September 2007 xv DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Reviewers
Dr. Robert Devlin—National Health and Environmental Effects Research Laboratory, Office of
Research and Development, U.S. Environmental Protection Agency, Chapel Hill, NC
Dr. Judy Graham—American Chemistry Council, LRI, 1300 Wilson Boulevard, Arlington, VA
22209
Dr. Stephen Graham—Office of Air Quality Planning and Standards (C504-06), Office of Air
and Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Beth Hassett-Sipple—Office of Air Quality Planning and Standards (C504-06), Office of
Air and Radiation, U.S. Environmental Protection Agency (C504-06), Research Triangle Park,
NC27711
Dr. Gary Hatch—National Health and Environmental Effects Research Laboratory, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
27711
Dr. Scott Jenkins—Office of Air Quality Planning and Standards (C504-06), Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. David Kryak—National Exposure Research Laboratory, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. John Langstaff—Office of Air Quality Planning and Standards (C504-06), Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Morton Lippmann—NYU School of Medicine, 57 Old Forge Road, Tuxedo, NY 10987
Dr. Karen Martin—Office of Air Quality Planning and Standards (C504-06), Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. William McDonnell—William F. McDonnell Consulting, 1207 Hillview Road, Chapel Hill,
NC27514
Dr. Dave McKee—Office of Air Quality Planning and Standards (C504-06), Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lucas Neas—National Health and Environmental Effects Research Laboratory, Office of
Research and Development, U.S. Environmental Protection Agency, Chapel Hill, NC 27711
September 2007 xvi DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Reviewers
Dr. Russell Owen—National Health and Environmental Effects Research Laboratory, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
27711
Dr. Haluk Ozkaynak—National Exposure Research Laboratory, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Jennifer Peel—Colorado State University, 1681 Campus Delivery, Fort Collins, CO 80523-
1681
Mr. Harvey Richmond—Office of Air Quality Planning and Standards (C504-06), Office of Air
and Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Steven Silverman—Office of General Counsel, U.S. Environmental Protection Agency,
Washington, DC 20460
Dr. Michael Stewart—Office of Air Quality Planning and Standards (C504-06), Office of Air
and Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Susan Stone—Office of Air Quality Planning and Standards (C504-06), Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Chris Trent—Office of Air Quality Planning and Standards (C504-06), Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. John Vandenberg—National Center for Environmental Assessment (B243-01), Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
27711
Dr. Alan Vette—National Exposure Research Laboratory, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Ron Williams—National Exposure Research Laboratory, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
September 2007 xvii DRAFT-DO NOT QUOTE OR CITE
-------
U.S. Environmental Protection Agency Project Team
for Development of Integrated Scientific Assessment
for Sulfur Oxide
Executive Direction
Dr. Ila Cote (Acting Director)—National Center for Environmental Assessment-RTF Division,
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Scientific Staff
Dr. Jee Young Kim (SOX Team Leader)—National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Jeff Arnold—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. James S. Brown—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Ellen Kirrane—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Dennis Kotchmar—National Center for Environmental Assessment (B243-01),
Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Tom Long—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Thomas Luben—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Quingyu Meng—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Joseph Pinto—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Paul Reinhart—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
September 2007 xviii DRAFT-DO NOT QUOTE OR CITE
-------
U.S. Environmental Protection Agency Project Team
for Development of Integrated Scientific Assessment
for Sulfur Oxide
(cont'd)
Scientific Staff
(cont'd)
Dr. Mary Ross—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. David Svendsgaard—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lori White—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. William Wilson—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Technical Support Staff
Ms. Emily R. Lee—Management Analyst, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Christine Searles—Management Analyst, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Debra Walsh—Program Analyst, National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Richard Wilson—Clerk, National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Document Production Staff
Ms. Barbra H. Schwartz—Task Order Manager, Computer Sciences Corporation,
2803 Slater Road, Suite 220, Morrisville, NC 27560
Mr. John A. Bennett—Technical Information Specialist, Library Associates of Maryland,
11820 Parklawn Drive, Suite 400, Rockville, MD 20852
September 2007 xix DRAFT-DO NOT QUOTE OR CITE
-------
U.S. Environmental Protection Agency Project Team
for Development of Integrated Scientific Assessment
for Sulfur Oxide
(cont'd)
Document Production Staff
(cont'd)
Mrs. Melissa Cesar—Publication/Graphics Specialist, Computer Sciences Corporation,
2803 Slater Road, Suite 220, Morrisville, NC 27560
Mrs. Rebecca Early—Publication/Graphics Specialist, TekSystems, 1201 Edwards Mill Road,
Suite 201, Raleigh, NC 27607
Mr. Eric Ellis—Records Management Technician, InfoPro, Inc., 8200 Greensboro Drive, Suite
1450, McLean, VA 22102
Ms. Stephanie Harper—Publication/Graphics Specialist, TekSystems, 1201 Edwards Mill Road,
Suite 201, Raleigh, NC 27607
Ms. Sandra L. Hughey—Technical Information Specialist, Library Associates of Maryland,
11820 Parklawn Drive, Suite 400, Rockville, MD 20852
Dr. Barbara Liljequist—Technical Editor, Computer Sciences Corporation, 2803 Slater Road,
Suite 220, Morrisville, NC 27560
Ms. Molly Windsor—Graphic Artist, Computer Sciences Corporation, 2803 Slater Road,
Suite 220, Morrisville, NC 27560
September 2007 xx DRAFT-DO NOT QUOTE OR CITE
-------
U.S. Environmental Protection Agency
Science Advisory Board (SAB)
Staff Office Clean Air Scientific Advisory Committee (CASAC)
CASAC NOX and SOX Primary NAAQS Review Panel
Chair
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 [M.D.]*, 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
September 2007 xxi DRAFT-DO NOT QUOTE OR CITE
-------
U.S. Environmental Protection Agency
Science Advisory Board (SAB)
Staff Office Clean Air Scientific Advisory Committee (CASAC)
CASAC NOX and SOX Primary NAAQS Review Panel
(cont'd)
Members
(cont'd)
Dr. Kent Pinkerton, Professor, Regents of the University of California, Center for Health and
the Environment, University of California, Davis, CA
Mr. Richard L. Poirot*, Environmental Analyst, Air Pollution Control Division, Department of
Environmental Conservation, Vermont Agency of Natural Resources, Waterbury, VT
Dr. Edward Postlethwait, Professor and Chair, Department of Environmental Health Sciences,
School of Public Health, University of Alabama at Birmingham, Birmingham, AL
Dr. Armistead (Ted) Russell*, Georgia Power Distinguished Professor of Environmental
Engineering, Environmental Engineering Group, School of Civil and Environmental
Engineering, Georgia Institute of Technology, Atlanta, GA
Dr. Richard Schlesinger, Associate Dean, Department of Biology, Dyson College, Pace
University, New York, NY
Dr. Christian Seigneur, Vice President, Atmospheric and Environmental Research, Inc., San
Ramon, CA
Dr. Elizabeth A. (Lianne) Sheppard, Research Professor, Biostatistics and Environmental &
Occupational Health Sciences, Public Health and Community Medicine, University of
Washington, Seattle, WA
Dr. Frank Speizer [M.D.]*, Edward Kass Professor of Medicine, Channing Laboratory,
Harvard Medical School, Boston, MA
Dr. George Thurston, Associate Professor, Environmental Medicine, NYU School of Medicine,
New York University, Tuxedo, NY
Dr. James Ultman, Professor, Chemical Engineering, Bioengineering Program, Pennsylvania
State University, University Park, PA
Dr. Ronald Wyzga, Technical Executive, Air Quality Health and Risk, Electric Power Research
Institute, P.O. Box 10412, Palo Alto, CA
September 2007 xxii DRAFT-DO NOT QUOTE OR CITE
-------
U.S. Environmental Protection Agency
Science Advisory Board (SAB)
Staff Office Clean Air Scientific Advisory Committee (CASAC)
CASAC NOX and SOX Primary NAAQS Review Panel
(cont'd)
SCIENCE ADVISORY BOARD STAFF
Mr. Fred Butterfield, CASAC Designated Federal Officer, 1200 Pennsylvania Avenue, N.W.,
Washington, DC, 20460, Phone: 202-343-9994, Fax: 202-233-0643 fbutterfield.fred@epa.gov)
(Physical/Courier/FedEx Address: Fred A. Butterfield, III, EPA Science Advisory Board Staff
Office (Mail Code 1400F), Woodies Building, 1025 F Street, N.W., Room 3604, Washington,
DC 20004, Telephone: 202-343-9994)
* Members of the statutory Clean Air Scientific Advisory Committee (CASAC) appointed by the EPA
Administrator
September 2007 xxiii DRAFT-DO NOT QUOTE OR CITE
-------
a
ACS
ADS
AHR
AM
APHEA
APEX
APIMS
ARIC
ARP
AQCD
asl
atm
P
B[or]P
BHR
BS
CAA
CAMP
CASAC
CASTNet
CDC
CHAD
CHF
CHS
CH3-S-H
CH3-S-S-CH3
CI
CMSA
CO
CoH
CONUS
COPD
CS2
CVD
DEN
DEP
DL
DMS
Abbreviations and Acronyms
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
Clean Air Act
Childhood Asthma Management Program
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
detection limit
dimethyl sulfide
September 2007
xxiv
DRAFT-DO NOT QUOTE OR CITE
-------
ED
ECG
EIB
ELF
EMECAM
EPA
eNO
ET
FEMs
FEVo.75
FEVi
FPD
FPD-TA
FRM
FVC
GAM
GIS
GLM
GSH
GST
H+
HEADS
HEI
HF
HNO2
HNO3
HO2
H202
HR
HRV
H2S
HSO3
HSO4
H2SO4
hv
ICD9
ICDs
Ig
fflD
IL
emergency department
el ectrocardi ography; el ectrocardi ogram
exercise-induced bronchial reactivity
epithelial lining fluid
Spanish Multicentre Study on Air Pollution and Mortality
U.S. Environmental Protection Agency
exhaled nitric oxide
extrathoracic
Federal Equivalent Methods
forced expiratory volume in 0.75 second
forced expiratory volume in 1 second
flame photometric detection
flame photometric detection-thermal analysis
Federal Reference Method
forced vital capacity
Generalized Additive Model(s)
Geographic Information System
Generalized Linear Model(s)
glutathione; reduced glutathione
glutathione S-transferase (e.g., GSTM1, GSTP1, GSTT1)
hydrogen ion
Harvard-EPA Annular Denuder System
Health Effects Institute
high frequency
nitrous acid
nitric acid
hydroperoxyl; hydroperoxy radical
hydrogen peroxide
hazard ratio
heart rate variability
hydrogen sulfide
hydrogen sulfite, bisulfite
bisulfate ion
sulfuric acid
solar ultraviolet photon
International Classification of Diseases, Ninth Revision
implanted cardioverter defibrillators
immunoglobulin (e.g., IgA, IgE, IgG)
ischemic heart disease
interleukin (e.g., IL-4, IL-6, IL-8)
September 2007
xxv
DRAFT-DO NOT QUOTE OR CITE
-------
IQR
ISA
IUGR
K
LF
LOD
LRD
MCh
MENTOR
MI
MEF50o/0
MMEF
MONICA
MOZART-2
MSA
N, n
NAAQS
NaCl
NaCO3
NADP
NAMS
NAS
NCICAS
NCore
NERL
NH4+
NHAPS
NHANES
NMMAPS
NO
N02
NO3
NO3
NOX
NR
NTN
02
03
ocs
OH
interquartile range
Integrated Science Assessment
intrauterine growth retardation
mass transfer coefficient
low frequency
limit of detection
lower respiratory disease
methacholine
Modeling Environment for Total Risk for One-Atmosphere studies
myocardial infarction
maximal midexpiratory flow at 50% of forced vital capacity
maximal midexpiratory flow
Monitoring Trend and Determinants in Cardiovascular Disease
(registry)
Model for Ozone and Related Chemical Tracers, version 2
metropolitan statistical area
number of observations
National Ambient Air Quality Standards
sodium chloride
sodium carbonate
National Atmospheric Deposition Program
National Air Monitoring Stations
Normative Aging Study
National Cooperative Inner-City Asthma Study
National Core Monitoring Network
National Exposure Research Laboratory
ammonium ion
National Human Activity Pattern Survey
National Health and Nutrition Examination Survey
National Morbidity, Mortality, and Air Pollution Study
nitric oxide
nitrogen dioxide
nitrate radical
nitrate ion
oxides of nitrogen
not reported
National Trends Network
molecular oxygen, diatomic oxygen
ozone
carbonyl sulfide
hydroxyl radical
September 2007
xxvi
DRAFT-DO NOT QUOTE OR CITE
-------
OR
P,P
PAARC
PAH
PC(SO2)
PD20FEVi
PD20
PD100
PEACE
PEC
PEF
PEMs
PF
PM
PM2.5
PM
iQ-2.5
PMT
ppb
ppbv
ppm
pptv
PRB
PS
r
R2
RAS
Raw
RH
r-MSSD
RR
s2-
SAPALDIA
SAVIAH
odds ratio
probability value
Air Pollution and Chronic Respiratory Diseases (study)
polycyclic aromatic hydrocarbon
provocative concentration of SC>2 that produces a 100% increase in
specific airways resistance
20% decrease in forced expiratory volume in 1 second
provocative dose that produces a 20% decrease in FEVi
provocative dose that produces a 100% increase in sRAW
Pollution Effects on Asthmatic Children in Europe (study)
pulmonary endocrine cell
peak expiratory flow
personal exposure monitors
pulsed fluorescence
particulate matter
particulate matter with 50% upper cut point aerodynamic diameter
of 2.5 |im for sample collection; surrogate for fine PM
particulate matter with 50% upper cut point aerodynamic diameter
of 10 |im for sample collection
particulate matter with 10 jim as upper cut point aerodynamic
diameter and 2.5 jim as lower cut point for sample collection;
surrogate for thoracic coarse PM (does not include fine PM)
particulate matter with 50% upper cut point aerodynamic diameter
of 13 |im for sample collection
photomultiplier tube
parts per billion
parts per billion by volume
parts per million
parts per trillion by volume
policy relevant background
passive sample
correlation coefficient
multiple correlation coefficient
roll-around system
airways resistance
relative humidity
root mean square of successive differences in R-R intervals.
rate ratio; relative risk
sulfur radical
Study of Air Pollution and Lung Diseases in Adults
Small-Area Variation in Air Pollution and Health (study)
September 2007
xxvn
DRAFT-DO NOT QUOTE OR CITE
-------
SD
SDNN
SES
SHEDS
SIDS
SNP
35§
SLAMS
SO
S02
SO3
SO32
SO42
SOX
S20
SPM
sRaw
STN
T
TEARS
TEA
TNF
TSP
URI
UV
VE
standard deviation
standard deviation of normal R-R intervals
socioeconomic status
Simulation of Human Exposure and Dose System
sudden infant death syndrome
single nucleotide polymorphism
sulfur-35 radionuclide
State and Local Air Monitoring Stations
sulfur monoxide
sulfur dioxide
sulfur trioxide
sulfite ion
sulfate ion
sulfur oxides
di sulfur monoxide
suspended particulate matter
specific airways resistance
Speciation Trends Network
tau; atmospheric lifetime
thiobarbituric acid reactive substances
triethanolamine
tumor necrosis factor (e.g., TNF-a)
total suspended particles
upper respiratory infections
ultraviolet
minute ventilation
September 2007
xxvin
DRAFT-DO NOT QUOTE OR CITE
-------
i 1. INTRODUCTION
2
O
4 The draft Integrated Science Assessment (ISA) presents a concise synthesis of the most
5 policy-relevant science to form the scientific foundation for the review of the primary (health-
6 based) National Ambient Air Quality Standards (NAAQS) for sulfur dioxide (862).l The draft
7 ISA is intended to "accurately reflect the latest scientific knowledge useful in indicating the kind
8 and extent of identifiable effects on public health which may be expected from the presence of
9 [a] pollutant in ambient air" (Clean Air Act, Section 108 (42 U.S.C. 7408)).2 The draft ISA
10 contains the key information and judgments formerly found in the Air Quality Criteria Document
11 (AQCD) for Sulfur Oxides (SOX), and a series of Annexes to the draft ISA provide more
12 extensive and detailed summaries of the most pertinent scientific literature. The draft ISA thus
13 serves to update and revise the information included in the AQCD published by the U.S.
14 Environmental Protection Agency (EPA) in 1982 (U.S. Environmental Protection Agency,
15 1982).
16 The draft ISA for this review of the SC>2 NAAQS critically evaluates and integrates
17 scientific information on the health effects associated with exposure to sulfur oxides in the
18 ambient air. It focuses on scientific information that has become available since the last review
19 and reflects the current state of knowledge on the most relevant issues pertinent to the review of
20 the primary 862 NAAQS. The ISA is supported by a more detailed and comprehensive
21 assessment of the scientific literature, which will be compiled into a series of annexes. Together,
22 the ISA and annexes replace the AQCD that was prepared in previous NAAQS reviews.
23 SC>2 is one of a group of substances known as sulfur oxides, which include multiple
24 gaseous (e.g., SC>2, sulfur trioxide [SOs], particulate (e.g., sulfate [SC>42 ], or sulfuric acid
25 [H2SO4]) species. For the current review, multiple species of sulfur oxides are considered as
26 appropriate and as allowed by the available data. For example, descriptions of the atmospheric
27 chemistry of sulfur oxides include both gaseous and particulate species, because a meaningful
28 analysis would not be possible otherwise. In addition, the health effects of gaseous sulfur oxides
29 other than 862 are considered when information on these other species is available. Finally, the
1 Information on legislative requirements and history of SO2 NAAQS reviews are presented in the Preface.
2 The secondary NAAQS for SO2 is being reviewed independently, in conjunction with the review of the secondary
NAAQS for nitrogen dioxide (NO2). A review of the primary NAAQS for NO2 is also underway.
September 2007 1-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 possible influence of other atmospheric pollutants on the interpretation of the role of 862 in
2 health effects studies is considered, including interactions of 862 with other pollutants that co-
3 occur in the environment (e.g., nitrogen oxides, carbon monoxide [CO], ozone [O3], particulate
4 matter [PM]).
5 As discussed in the Draft Integrated Plan for the Review of the Primary NAAQS for
6 Sulfur Dioxide (U.S. Environmental Protection Agency, 2007), a series of policy-relevant
7 questions frames this review of the scientific evidence to provide a scientific basis for a decision
8 on whether the current primary NAAQS for 862 (0.030 parts per million [ppm], annual average;
9 0.14 ppm, 24-h average) should be retained or revised. The draft ISA focuses on evaluation of
10 the newly available scientific evidence to best inform consideration of these framing questions,
11 including the following:
12 • Has new information altered/substantiated the scientific support for the
13 occurrence of health effects following short- and/or long-term exposure to levels
14 of SOX found in the ambient air?
15 • Does new information impact conclusions from the previous review regarding the
16 effects of SOX on susceptible populations?
17 • At what levels of SOX exposure do health effects of concern occur?
18 • Has new information altered conclusions from previous reviews regarding the
19 plausibility of adverse health effects caused by SOX exposure?
20 • To what extent have important uncertainties identified in the last review been
21 reduced and/or have new uncertainties emerged?
22 • What are the air quality relationships between short-term and longer-term
23 exposures to SOX?
24
25
26 1.1 DOCUMENT DEVELOPMENT
27 EPA formally initiated the current review of the SC>2 NAAQS by announcing the
28 commencement of the review in the Federal Register with a call for information in May 2006
29 (Federal Register, 2006). In addition to the call for information, publications are identified
30 through an ongoing literature search process that includes searching MEDLINE and other
31 databases using as key words the terms: sulfur oxides, sulfur dioxide, SOX, SC>2, and reduced
September 2007 1 -2 DRAFT-DO NOT QUOTE OR CITE
-------
1 sulfur gases. This search strategy is periodically reexamined and modified to enhance
2 identification of pertinent published papers. Additional papers are identified for inclusion in the
3 publication base in several ways. First, EPA staff reviews pre-publication tables of contents for
4 journals in which relevant papers may be published. Second, expert chapter authors are charged
5 with independently identifying relevant literature. Finally, additional publications that may be
6 pertinent are identified by both the public and the Clean Air Scientific Advisory Committee
7 (CASAC) during the external review process. The studies identified include research published
8 or accepted for publication by a date determined to be as inclusive as possible given the relevant
9 target dates in the NAAQS review schedule. Some additional studies, published after that date,
10 may also be included if they provide new information that impacts one or more key scientific
11 issues. The combination of these approaches should produce a comprehensive collection of
12 pertinent studies to form the basis of the ISA and to appear summarized in the next ISA annexes.
13 The following sections briefly summarize criteria for selection of studies for this draft
14 ISA. Consideration of these issues informs our judgments on the relative quality of individual
15 studies and allows us to focus the assessment on the most pertinent studies.
16
17 Criteria for Selecting Epidemiological Studies
18 In selecting epidemiological studies for the present assessment, EPA considers whether a
19 given study contains information on (1) short- or long-term exposures at or near ambient levels
20 of SOX; (2) health effects of specific SOX species or indicators related to 862 sources; (3) health
21 endpoints that repeat or extend findings from earlier assessments as well as those not previously
22 researched; (4) susceptible and vulnerable populations to SOX exposure; (5) multiple pollutant
23 analyses and other approaches to address issues related to potential interactions (e.g., synergistic
24 effects of SOX with other pollutants), confounding (e.g., SOX associations with health endpoints
25 independent of copollutants), and effect modification (e.g., copollutant modification of SOX
26 effects on health endpoints); and/or (6) important methodological issues (e.g., lag of effects,
27 model specifications, thresholds, mortality displacement). Among the epidemiological studies,
28 particular emphasis is focused on those relevant to standard setting in the United States.
29 Specifically, studies conducted in the United States or Canada will be generally accorded more
30 text discussion than those from other geographic regions, as the potential impacts of different
31 health care systems and the underlying health status of populations need to be accounted for in
32 the assessment. In addition, emphasis in the text is placed on discussion of (1) new, multicity
September 2007 1 -3 DRAFT-DO NOT QUOTE OR CITE
-------
1 studies that employ standardized methodological analyses for evaluating SOX effects, provide
2 overall estimates for effects based on combined analyses of information pooled across cities, and
3 examine results for consistency across cities; (2) new studies that provide quantitative effect
4 estimates for populations of interest; and (3) studies that regard SOX as a component of a
5 complex mixture of air pollutants and, thus, give consideration to the levels of other copollutants,
6 correlations of SOX with these copollutants, and conduct multipollutant analyses.
7
8 Criteria for Selecting Experimental Studies
9 A set of explicit criteria is also used to select experimental studies for discussion. The
10 selection of research evaluating controlled exposures of laboratory animals focuses primarily on
11 those studies conducted at or near ambient SOX concentrations and those studies that
12 approximate expected human exposure conditions in terms of concentration and duration which
13 will depend on the toxicokinetics and biological sensitivity of the particular laboratory animal
14 examined. In discussing the mechanisms of SOX toxicity, studies conducted under
15 atmospherically relevant conditions are emphasized whenever possible. However, studies at
16 higher levels are also considered to allow for species-to-species differences and potential
17 differences in sensitivity between study subjects and especially susceptible human populations.
18 For research evaluating controlled human exposures to SOX, emphasis is placed on studies that
19 (1) investigate effects on potentially susceptible populations such as asthmatics, particularly
20 studies where subjects serve as their own control to compare responses following SOX exposure
21 and sham exposure and where responses in susceptible individuals are compared with those in
22 age-matched healthy controls; (2) address issues such as dose-response or time-course of
23 responses; (3) investigate exposure to SOX separately and in combination with other pollutants;
24 (4) include controlled exposures to filtered air; and (5) have sufficient sample size to adequately
25 assess findings.
26 In assessing the scientific quality and relevance of epidemiological, animal lexicological,
27 and human controlled exposure studies, the following considerations are taken into account:
28 (1) where ambient air measurements are used, to what extent are the data of adequate quality and
29 sufficiently representative to serve as credible exposure indicators; (2) were the study
30 populations adequately selected and are they sufficiently well-defined to allow for meaningful
31 comparisons between study groups; (3) are the health endpoint measurements meaningful and
32 reliable; (4) are the statistical analyses appropriate, properly performed, and properly interpreted;
September 2007 1 -4 DRAFT-DO NOT QUOTE OR CITE
-------
1 (5) are likely covariates (i.e., potential confounders or effect modifiers) adequately controlled or
2 taken into account in the study design and statistical analyses; and (6) are the reported findings
3 internally consistent. Consideration of these issues informs our judgments on the relative quality
4 of individual studies and will allow us to focus the assessment on the most pertinent studies.
5
6
7 1.2 ORGANIZATION OF THE DOCUMENT
8 This draft ISA includes five chapters. This introductory chapter (Chapter 1) presents
9 background information on the purpose of the document and characterizes how policy-relevant
10 scientific studies are identified. Chapter 2 highlights key concepts or issues relevant to
11 understanding the atmospheric chemistry, sources, exposure, and dosimetry of sulfur oxides,
12 following a "source-to-dose" paradigm. Chapter 3 evaluates and integrates health information
13 relevant to the review of the primary NAAQS for SC>2. In this chapter, findings from
14 epidemiological, lexicological, and human clinical studies are integrated in an assessment of the
15 relationships between exposure to ambient SOX and health outcomes. The focus of this chapter
16 is on the strength of underlying epidemiological or toxicological evidence and the coherence and
17 plausibility of the body of evidence for effects on the respiratory, cardiovascular, or other
18 system. Chapter 4 provides information relevant to the public health impact of exposure to
19 ambient SOX, including potential susceptible population groups. Finally, Chapter 5 summarizes
20 key findings and conclusions from the atmospheric sciences, ambient air data analyses, exposure
21 assessment, dosimetry, and health effects in consideration of the review of the NAAQS for SC>2.
22 The draft ISA is supplemented by a series of annexes, which are focused on
23 accomplishing two goals. The first goal is to identify scientific research that is relevant to
24 informing key policy issues. The second goal is to produce a base of evidence containing all of
25 the publications relevant to the SC>2 NAAQS review. The annexes provide information on
26 (1) the atmospheric chemistry of SOX as well as the sampling/analytic methods for measurement
27 of SOX3, (2) environmental concentrations and human exposure to SOX, (3) dosimetry;
28 (4) toxicological studies of SOX health effects in laboratory animals, and (5) epidemiological
This section will also provide information on NO2 in order to support the reviews of the primary and secondary
NAAQS for both SO2 and NO2. The atmospheric chemistry of NOX and SOX are intricately linked. Therefore,
discussion of their combined chemistry is more effective and more efficient than a separate discussion of each
pollutant.
September 2007 1 -5 DRAFT-DO NOT QUOTE OR CITE
-------
1 studies of health effects from short- and long-term exposure to SOX. More detailed information
2 on various methods and results for the health studies is summarized in tabular form in the annex.
3 These tables are generally organized to include information about (1) concentrations of SOX and
4 averaging times, (2) description of study methods employed, (3) results and comments, and
5 (4) quantitative outcomes for SOX effect estimates.
September 2007 1 -6 DRAFT-DO NOT QUOTE OR CITE
-------
i 2. SOURCE-TO-TISSUE DOSE
2
O
4 This chapter provides basic information about concepts and findings that relate to
5 considerations in atmospheric science, human exposure assessment, and human dosimetry. It is
6 meant to serve as a prologue for detailed discussions of evidence on health effects that will
7 follow in Chapters 3 and 4. Section 2.1 provides an overview of the atmospheric chemistry
8 processes involved in the oxidation of SC>2 and those involved in the production of SC>2 from
9 reduced sulfur gases in the atmosphere. Sources of SC>2 are presented in Section 2.2. A
10 description of the methods for measuring 862 and issues associated with its measurement are
11 presented in Section 2.3. Data for ambient 862 concentrations are characterized in Section 2.4.
12 Policy Relevant Background concentrations of 862, i.e., those concentrations that basically
13 define uncontrollable levels are also presented in Section 2.4. Factors governing personal
14 exposures to SC>2 and associated issues are discussed in Section 2.5. Finally, the dosimetry of
15 SC>2 in the respiratory tract is discussed in Section 2.6. The order of topics in this chapter
16 follows in large measure that given in the National Research Council paradigm for integrating air
17 pollutant research (National Research Council, 1998).
18
19
20 2.1 ATMOSPHERIC CHEMISTRY
21 The only forms of monomeric sulfur oxides of interest in tropospheric chemistry are
22 sulfur dioxide (802) and sulfur trioxide (SOs). SOs can be emitted from the stacks of power
23 plants and factories; however, it reacts extremely rapidly with H^O in the stacks or immediately
24 after release into the atmosphere to form sulfuric acid (H^SC^) which then partitions into the
25 aqueous phase of particles. Thus, only 862 is present in the tropospheric boundary layer at
26 concentrations significant for atmospheric chemistry and human exposures.
27 SC>2 is oxidized either in the gas phase or in the aqueous phase in cloud drops because it
28 is highly water-soluble. The gas phase oxidation of SO2 proceeds through the reaction:
29 SO2 +OH + M-* HSO3 + M (2_1}
30 followed by:
September 2007 2-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 HSO3 + O2 ->• SO3 + HO 2 (2_2)
2 SO3 + H2O^> H2SO4 (2_3)
3 Because H2SO4 is extremely soluble, it will be removed rapidly by transfer to the aqueous phase
4 of aerosol particles and cloud drops. Rate coefficients for the reactions of 862 with either the
5 hydroperoxyl (HO2) or nitrate radical (NOs) are too low to be significant (JPL, 2003).
6 The major sulfur species in clouds are hydrogen sulfite (HSO3 ) and the sulfite ion
7 (SOs2 ) (both of which are derived from the dissolution of SC>2 in water and are referred to as
8 S(IV)), and bisulfate ion (HSO4 ) and sulfate (SO42 ) (which are referred to as S(VI)). The chief
9 species capable of oxidizing S(IV) to S(VI) in cloud water are ozone (Os), peroxides (either
10 hydrogen peroxide [H2O2] or organic peroxides), hydroxyl (OH) radicals, and ions of transition
1 1 metals such as Fe and Cu that can catalyze the oxidation of S(IV) to S(VI) by 62.
12 The basic mechanism of the aqueous phase oxidation of 862 has long been studied and
13 can be found in numerous texts on atmospheric chemistry, e.g., Seinfeld and Pandis (1998),
14 Jacob (1999), and Jacobson (2002). Following Jacobson (2002), the steps involved in the
15 aqueous phase oxidation of SO2 can be summarized as follows:
16 Dissolution of SO2
17 22 (2.4)
18 The formation and dissociation
19
SO,(aq) + H7O(aq)^H?SO3^H+ + HSO
20 In the pH range commonly found in rainwater (2 to 6), the most important reaction converting
21 S(IV) to S(VI) is:
22 HSO3- + H2 O2 + H + <=> SO42~ + H2O + 2H+ (2 g
23 as SOs2 is much less abundant than HSOs .
24 For pH up to about 5.3, H2O2 is the dominant oxidant, while at pH > 5.3, Os followed by
25 Fe(III) become dominant. Higher pH levels are expected to be found mainly in marine aerosols.
26 However, in marine aerosols, the chloride-catalyzed oxidation of S(IV) may be more important
September 2007 2-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 (Zhang and Millero, 1991; Hoppel and Caffrey, 2005). Because the ammonium ion (NH4+) is so
2 effective in neutralizing acidity, when present, it affects the rate of oxidation of S(IV) to S(VI)
3 and the rate of dissolution of SO2 in particles and cloud drops.
4 A comparison of the relative rates of oxidation by gas and aqueous phase reactions by
5 Warneck (1999) indicates that only about 20% of SO2 is oxidized by gas phase reactions. Thus,
6 SO2 is oxidized mainly by aqueous phase reactions. SO2 is also removed from the atmosphere
7 by dry deposition to moist surfaces, resulting in an atmospheric lifetime (T) with respect to
8 deposition on the order of 1 week, depending on humidity. The rate of oxidation of SO2 to SO42
9 ranges from 0.5 to -2% IT1 as measured in power plant plumes (Pueschel and van Valin, 1978),
10 resulting in an atmospheric lifetime ranging from about 2 days to about a week, with respect to
11 this process. These two processes, oxidation and deposition, lead to an overall lifetime of SO2 in
12 the atmosphere of a few days.
13
14
15 2.2 SOURCES OF SULFUR OXIDES
16 Anthropogenic emissions of SO2 are mainly from combustion of fossil fuels by electrical
17 utilities (-66 %) and industry (-29%), with transportation-related sources making only a minor
18 contribution (-5%) in 2002 (U.S. Environmental Protection Agency, 2006a). Thus, most SO2
19 emissions originate from point sources. Since sulfur is a volatile component of fuels, it is almost
20 quantitatively released during combustion. Hence, sulfur emissions can be calculated on the
21 basis of sulfur content in fuel stocks to greater accuracy than can be done for other pollutants like
22 nitrogen oxides or primary particulate matter (PM). However, the estimates given above are
23 nationwide averages and may not accurately reflect the contribution of specific local sources
24 determining a person's exposures to SO2 at any given location and time. For example, shipping
25 and associated in-port activities may be a significant source of SO2 in some coastal cities (Wang
26 et al., 2007).
27 The largest natural sources of SO2 are volcanoes and biomass burning. Even so, SO2
28 constitutes a relatively minor fraction (0.005% by volume) of total volcanic emissions (Holland,
29 1978). Volcanic sources of SO2 in the United States are limited to the Pacific Northwest,
30 Alaska, and Hawaii. Emissions of SO2 from burning vegetation are generally in the range of 1 to
31 2% of the biomass burned (see e.g., Levine and Pinto, 1998). Sulfur is bound in amino acids in
September 2007 2-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 vegetation and is released during combustion. Gaseous sulfur emissions from this source are
2 mainly in the form of SC>2.
3 In addition to its role as an emitted primary pollutant, SO2 is also produced by the
4 photochemical oxidation of reduced sulfur compounds such as dimethyl sulfide, or DMS,
5 (CH3-S-CH3), hydrogen sulfide (H2S), carbon disulfide (CS2), carbonyl sulfide (OCS), methyl
6 mercaptan (CHrS-H), and dimethyl disulfide (CHrS-S-CH?). The sources for these compounds
7 are mainly biogenic and are discussed in Annex Section 2.5. Emissions of reduced sulfur species
8 are associated typically with marine organisms living either in pelagic or coastal zones and with
9 anaerobic bacteria in marshes and estuaries. Emissions of dimethyl sulfide (DMS) from marine
10 plankton represent the largest single source of reduced sulfur species to the atmosphere (e.g.,
11 Berresheim et al., 1995). Except for OCS, which is lost mainly by photolysis (T ~ 6 months), all
12 the other species are lost mainly by reaction with OH and NOs radicals and are relatively short-
13 lived, having lifetimes of the order of a few hours to a few days (see Annex Section 2.3).
14 Reaction with NOs radicals at night most likely represents the major loss process for DMS and
15 methyl mercaptan. Although the mechanisms for the oxidation of DMS are still not completely
16 understood, excess sulfate in marine aerosol appears related mainly to the production of SO2
17 from the oxidation of DMS. Emissions of sulfur from natural sources are small compared to
18 anthropogenic emissions within the United States. However, important exceptions occur locally
19 as the result of volcanic activity, wildfires and in certain coastal zones as described here.
20 Because OCS is relatively long lived, it can survive oxidation in the troposphere and be
21 transported upwards into the stratosphere. Crutzen (1976) proposed that its oxidation to sulfate
22 in the stratosphere serves as the major source of mass in the stratospheric aerosol layer.
23 However, Myhre et al. (2004) proposed that SO2 transported upwards from the troposphere by
24 deep convection is the most likely source, as the flux of OCS is too small. In addition, in-situ
25 measurements of the isotopic composition of sulfur in stratospheric sulfate do not match those of
26 OCS (Leung et al., 2002). Thus, anthropogenic SO2 emissions could be important precursors to
27 the formation of the stratospheric aerosol layer.
28
29
September 2007 2-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.3 MEASUREMENT METHODS AND ASSOCIATED ISSUES
2 Currently, ambient 862 is measured using instruments based on pulsed ultraviolet (UV)
3 fluorescence. The UV fluorescence monitoring method for atmospheric 862 was developed to
4 improve upon the flame photometric detection (FPD) method, which in turn had displaced the
5 pararosaniline wet chemical method. The pararosaniline method is still the EPA Federal
6 Reference Method (FRM) for atmospheric SC>2, but it is rarely used because of its complexity
7 and slow response, even in its automated forms. Both the UV fluorescence and FPD methods are
8 designated as Federal Equivalent Methods (FEMs) by EPA, but UV fluorescence has largely
9 supplanted the FPD approach because of the UV method's inherent linearity, sensitivity, and the
10 need for consumable hydrogen gas for the FPD method.
11 In the UV fluorescence method, 862 molecules absorb UV light at one wavelength and
12 emit UV light at longer wavelengths in the process known as fluorescence through excitation of
13 the SC>2 molecule to a higher energy (singlet) electronic state. Once excited, the molecule decays
14 nonradiatively to a lower-energy electronic state from which it then decays to the original, or
15 ground, electronic state by emitting a photon of light at a longer wavelength (i.e., a lower-energy
16 photon) than the original, incident photon. The intensity of the emitted light is thus proportional
17 to the number of 862 molecules in the sample gas.
18 In commercial analyzers, light from a high-intensity UV lamp passes through a
19 bandwidth filter, allowing only photons with wavelengths around the 862 absorption peak (near
20 214 nm) to enter the optical chamber. The light passing through the source bandwidth filter is
21 collimated using a UV lens and passes through the optical chamber, where it is detected on the
22 opposite side of the chamber by the reference detector. A photomultiplier tube (PMT) is offset
23 from and placed perpendicular to the light path to detect the SC>2 fluorescence. Since the SC>2
24 fluorescence at 330 nm is different from its excitation wavelength, an optical bandwidth filter is
25 placed in front of the PMT to filter out any stray light from the UV lamp. A lens is located
26 between the filter and the PMT to focus the fluorescence onto the active area of the detector and
27 optimize the fluorescence signal. The limit of detection (LOD) for a non-trace level 862
28 analyzer is 10 ppb (CFR, 2006). However, most commercial analyzers have detection limits of
29 about 3 ppb. The EPA through its NCore initiative (U.S. Environmental Protection Agency,
30 2005) is engaged in a program to install and operate newer trace-level SC>2 instruments that will
31 increase the accuracy and precision of measurements at much lower levels.
September 2007 2-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 Sources of Positive Interference
2 The most common source of interference to the UV fluorescence method for SC>2 is from
3 other gases that fluoresce in a similar fashion to SO2 when exposed to UV radiation of that
4 wavelength. The most significant of these are poly cyclic aromatic hydrocarbons (PAHs), of
5 which naphthalene is a prominent example. Xylene is another common hydrocarbon that can
6 cause fluorescent interference. Consequently, any such aromatic hydrocarbons that are in the
7 optical chamber can act as a positive interference. To remove this source of interference, high-
8 sensitivity SC>2 analyzers like those to be used in the NCore network (U.S. Environmental
9 Protection Agency, 2005), have hydrocarbon scrubbers to remove these compounds from the
10 sample stream before the sample air enters the optical chamber.
11 Luke (1997) reported the positive artifacts of a modified pulsed fluorescence detector
12 generated by the coexistence of nitric oxide (NO), €82, and a number of highly fluorescent
13 aromatic hydrocarbons such as benzene, toluene, o-xylene, m-xylene, />-xylene, m-ethyltoluene,
14 ethylbenzene, and 1,2,4-trimethylbenzene. The positive artifacts could be reduced by using a
15 hydrocarbon "kicker" membrane. At a flow rate of 300 standard cc min"1 and a pressure drop of
16 645 torr across the membrane, the interference from ppm levels of many aromatic hydrocarbons
17 was eliminated entirely. NO fluoresces in a spectral region close to that of SO2. However, in
18 high-sensitivity SO2 analyzers, the bandpass filter in front of the PMT is designed to prevent NO
19 fluorescence from being detected at the PMT. Care must be exercised when using
20 multicomponent calibration gases containing both NO and SO2 so that the NO rejection ratio of
21 the SO2 analyzer is sufficient to prevent NO interference.
22 The most common source of positive bias (as contrasted with positive spectral
23 interference) in high-sensitivity SO2 monitoring is stray light reaching the optical chamber.
24 Since SO2 can be electronically excited by a broad range of UV wavelengths, any stray light with
25 an appropriate wavelength that enters the optical chamber can excite SO2 in the sample and
26 increase the fluorescence signal. Furthermore, stray light at the wavelength of the SO2
27 fluorescence that enters the optical chamber may impinge on the PMT and increase the
28 fluorescence signal. Several design features are incorporated to minimize the stray light that
29 enters the chamber. These features include the use of light filters, dark surfaces, and opaque
30 tubing to prevent light from entering the chamber.
September 2007 2-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 Nicks and Benner (2001) reported a sensitive 862 chemiluminescence detector based on
2 a differential measurement where response from ambient SCh is determined by the difference
3 between air containing SO2 and air scrubbed of SO2 when both air samples contain other
4 detectable sulfur species. Assuming monotonic efficiency of the sulfur scrubber, all positive
5 artifacts should also be reduced with this technique.
6
1 Sources of Negative Interference
8 Nonradiative deactivation (quenching) of excited SO2 molecules can occur from
9 collisions with common molecules in air, including nitrogen, oxygen, and water. During
10 collisional quenching, the excited 862 molecule transfers energy, kinetically allowing the 862
11 molecule to return to the original lower energy state without emitting a photon. Collisional
12 quenching results in a decrease in the 862 fluorescence and, hence, an underestimation of 862
13 concentration in the air sample. Of particular concern is the variable water vapor content of air.
14 Luke (1997) reported that the response of the detector could be reduced by about 7 and 15% at
15 water vapor mixing ratios of 1 and 1.5 mole percent (relative humidity [RH] = 35 to 50% at 20 to
16 25°C and 1 atm for a modified pulsed fluorescence detector (Thermo Environmental
17 Instruments, Model 43s). Condensation of water vapor in sampling lines must be avoided, as
18 water on the inlet surfaces can absorb 862 from the sample air. The simplest approach to avoid
19 condensation is to heat sampling lines to a temperature above the expected dewpoint and to
20 within a few degrees of the controlled optical bench temperature. At very high 862
21 concentrations, reactions between electronically excited 862 and ground state 862 might occur,
22 forming SO3 and SO (Calvert et al., 1978). However, the possibility that this artifact might be
23 affecting measurements at very high SO2 levels has not been examined.
24
25 Other Techniques for Measuring SO2
26 More sensitive techniques for measuring SO2 are available, but most of these systems are
27 too complex and expensive for routine monitoring applications. However, techniques such as
28 those described by Luke (1997) can be used to improve the sensitivity of ambient
29 SO2 monitors by eliminating sources of common interference.
30
31
September 2007 2-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.4 ENVIRONMENTAL CONCENTRATIONS OF SULFUR OXIDES
2
3 2.4.1 Ambient Air Quality Data for Sulfur Dioxide and Other
4 Sulfur Oxides
5 SC>2 data collected from the State and Local Air Monitoring Stations (SLAMS) and
6 National Air Monitoring Stations (NAMS) networks show that the decline in SC>2 emissions
7 from electric generating utilities has improved air quality. There has not been a single monitored
8 exceedance of the SC>2 annual ambient air quality standard in the United States since 2000,
9 according to the U.S. Environmental Protection Agency Acid Rain Program (ARP) 2005
10 Progress Report (U.S. Environmental Protection Agency, 2006b). EPA's trends data
11 (www.epa.gov/airtrends) reveal that the national composite average 862 annual mean ambient
12 concentration decreased by 48% from 1990 to 2005, with the largest single-year reduction
13 coming in 1994-1995, the ARP's first operating year (U.S. Environmental Protection Agency,
14 2006b). Figure 2.4-1 depicts data for 862 emissions in the continental United States (CONUS)
15 in these years that reflect this reduction with individual state-level totals.
16 These emissions data trends are consistent with the trends in the observed ambient
17 concentrations from the Clean Air Status and Trends Network (CASTNet). Following
18 implementation of the Phase I controls on ARP sources between 1995 and 2000, significant
19 reductions in SC>2 and ambient SC>42 concentrations were observed at CASTNet sites throughout
20 the eastern United States. The mean annual concentrations of SC>2 and SC>42 from CASTNet's
21 long-term monitoring sites can be compared using two 3-year periods (1989 through 1991 and
22 2003 through 2005) in Figures 2.4-2a and 2.4-2b for SO2, and Figures 2.4-3a and
23 2.4-3bforSO42~.
24 From 1989 through 1991, that is, in the years prior to implementation of the ARP Phase I,
25 the highest ambient mean concentrations of 862 and SC>42 were observed in western
26 Pennsylvania and along the Ohio River Valley: >20 jig nT3 (~8 ppb) SO2 and >15 jig nT3
27 SC>42 . As with SC>2, in the years since the ARP controls were enacted, both the magnitude of
28 SO42 concentrations and their areal extent have been significantly reduced, with the largest
29 decreases again coming along the Ohio River Valley.
September 2007 2-8 DRAFT-DO NOT QUOTE OR CITE
-------
• S02 Emissions in 1990
I I SO2 Emissions in 1995
I I SO2 Emissions in 2000
I I SO2 Emissions in 2005
Scale: Largest bar equals
2.2 million tons of SO2
emissions in Ohio, 1990
Figure 2.4-1.
State-level SO2 emissions, 1990-2005.
Source: Environmental Protection Agency Clean Air Markets Division (www.epa.gov/airmarkets/index.html).
September 2007
2-9
DRAFT-DO NOT QUOTE OR CITE
-------
Figure 2.4-2. Annual mean ambient SOi concentration, (a) 1989 through 1991, and
(b) 2003 through 2005.
* Dots on all maps represent monitoring sites. Lack of shading for Southern Florida indicates lack of monitoring
coverage.
Source: Environmental Protection Agency, CASTNet (www.epa.gov/castnet/).
September 2007
2-10
DRAFT-DO NOT QUOTE OR CITE
-------
(b)
-50
.„
-7.0
->8.0
Figure 2.4-3.
2-
Annual mean ambient SC>4 Concentration, (a) 1989 through 1991, and
(b) 2003 through 2005.
* Dots on all maps represent monitoring sites. Lack of shading for Southern Florida indicates lack of monitoring
coverage.
Source: Environmental Protection Agency, CASTNet (www.epa.gov/castnet/).
1 Figure 2.4-4 depicts for the CONUS the magnitude and spatial distribution of 862
2 emissions in 2006 from sources in the ARP. This depiction shows clearly the continuing
3 overrepresentation of 862 sources in the United States east of the Mississippi River as compared
4 to west of it, a trend even stronger in the central Ohio River Valley and which was evident in the
5 smoothed concentration plots in Figures 2.4-2a and 2.4-2b. As shown in Table 2.4-1, regional
6 distributions of SC>2 and SC>42 concentrations averaged for the 3 years 2003 through 2005 reflect
7 this geospatial emissions source difference as well.
September 2007
2-11
DRAFT-DO NOT QUOTE OR CITE
-------
'. - •;"* .•".. / "•'
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 tons
Figure 2.4-4. Annual SOi Emissions in 2006 for Acid Rain Program Cooperating
Facilities.
Source: Environmental Protection Agency, Clean Air Markets Division (www.epa.gov/airmarkets/index.html).
1 2.4.2 Spatial and Temporal Variability of Ambient Sulfur Dioxide
2 Concentrations
3 SC>2 concentrations have been falling throughout all regions of the United States as
4 demonstrated by the CASTNet data reviewed above. In and around most individual
5 Consolidated Metropolitan Statistical Areas (CMSAs), the trends are also toward lower 862
6 levels. Table 2.4-2 shows that many annual and even 1-h mean concentrations for the years 2003
7 through 2005 were consistently at or below the operating LOD of ~3 ppb for the standard SO2
8 monitor deployed in the regulatory networks, while the aggregate mean value over all 3 years
9 and all sites in and around the CMSAs was just above the LOD at ~4 ppb, and identical to the
September 2007
2-12
DRAFT-DO NOT QUOTE OR CITE
-------
1 1-h and 24-h means. Hence, it appears reasonable to aggregate up in time from available 1-h
2 samples to daily and even annual exposure estimates.
3 Figure 2.4-5 shows the composite diurnal variation in hourly SO2 concentrations in
4 boxplot form from all monitors reporting SC>2 data into the Air Quality System (AQS) database.
5 The AQS contains measurements of air pollutant concentrations in the 50 states, plus the District
6 of Columbia, Puerto Rico, and the Virgin Islands for the six criteria air pollutants (SC>2, NC>2,
7 PM, CO, Pb, Os) and hazardous air pollutants. The same data were used to construct Table 2.4-2
8 and to configure Figure 2.4-5. As can be seen from Figure 2.4-5, concentrations beneath the
9 95th percentile level are indistinguishable from each other, but are typically in the range of only
10 a few ppb. However, the peaks in the distribution at any hour of the day can be a factor of 10 or
11 more higher than values in the bulk of the concentration distribution. Overall, there is some
12 indication that the highest values are reached either at midday or during the middle of the night.
13 Daytime peaks could result from down-mixing of air aloft due to convective activity, as SO2 is
14 emitted mainly by elevated sources. Nighttime peaks are more likely due to trapping of local
15 emissions beneath a shallow nocturnal boundary layer.
800
700-
600-
a
£ 400 H
CM
o
« 300 H
200-
100-
0-
TT
I 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 2.4-5. Boxplot of hourly SOi concentrations across all cities in focus.
September 2007
2-13
DRAFT-DO NOT QUOTE OR CITE
-------
1 To be sure, the maximum 1-h concentration observed at some sites in and around some
2 CMS As did still exceed the mean by a large margin, with maximum 1-h values in Table 2.4-2 of
3 >600 ppb. However, the 50th percentile maximum value outside CMSAs, 5 ppb, was only
4 slightly greater than the 1-h, 24-h, and annual mean value, 4 ppb. The 50th percentile maximum
5 value inside CMSAs, 7 ppb, was 75% greater than these longer-term averages, reflecting
6 heterogeneity in source strength and location. In addition, even with 1-h maximum values of
7 >600 ppb, the maximum annualized mean value for all CMSAs was still <16 ppb, and, hence,
8 below the current annual primary 862 NAAQS.
9 The strong east-to-west gradient in 862 emissions described above is well replicated in
10 the observed concentrations in individual CMSAs. Thus, for example, values in Table 2.4-3
11 represent the mean annual concentrations in the years 2003 through 2005 for the 12 CMSAs with
12 four or more SC>2 regulatory monitors, ranging from a reported low of ~1 ppb in Riverside, CA
13 and San Francisco, CA to a high of-12 ppb in Pittsburgh, PA and Steubenville, OH in the
14 highest SC>2 source region.
15 The Pearson correlation coefficients (r) for multiple monitors in these CMSAs (see also
16 Table 2.4-3) were generally very low for all cities, especially at the lower end of the observed
17 concentration ranges, and are even negative at the very lowest levels on the West Coast. This
18 reflects strong heterogeneity in 862 ambient concentrations even within any one CMS A and,
19 therefore, indicates possibly different exposures of spatially distinct subgroups of humans in
20 these CMSAs to these very low concentrations of SO2. At higher concentrations, the r values
21 were also higher. In some CMSAs, this heterogenerity may result from meteorological effects
22 whereby a generally well-mixed subsiding air mass containing one or more relatively high
23 concentration SC>2 plumes would be spread more nearly uniformly across an area than would
24 faster-moving plumes with lower SC>2 concentrations. However, because the highest r values,
25 i.e., those >0.7, correspond to the highest 862 concentrations, i.e., >6 and >10 ppb, instrument
26 error may also play a role. Since the lowest 862 concentrations are at or below the operating
27 LOD and demonstrate the lowest correlation across monitors that share at least some air mass
28 characteristics most of the year, the unbiased instrument error in this range may be confounding
29 interpretation of any possible correlation. This could be because the same actual ambient value
30 would be reported by different monitors (with different error profiles) in the CMS A as different
31 values in this lowest concentration range.
September 2007 2-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 To better characterize the extent and spatiotemporal variance of 862 concentrations
2 within each of the CMS As having more than four SC>2 monitors (listed in Table 2.4-3), the
3 means, minima, and maxima were computed from daily mean data across all available monitors
4 for each month for the years 2003 through 2005. Because many of these CMS As with SC>2
5 monitors also reported SC>42 , it is possible to compute the degree of correlation between SC>2,
6 the emitted species, and SC>42 , the most prominent oxidized product from 862. SC>42 values,
7 however, while averaged over all available data, are generally available at their monitoring sites
8 on a schedule of only 1 in 3 days or 1 in 6 days. Furthermore, 862 and SO42 monitors are not
9 collocated throughout the CMS As. For each of five example CMS As, Figures 2.4-6 through
10 2.4-10 depict monthly values aggregated from daily means of (a) the monthly mean, minimum,
11 and maximum SC>2 concentrations; (b) the monthly mean, minimum and maximum SC>42
12 concentrations; and (c) a scatterplot of SC>2 versus SC>42 concentrations.
JD
Q.
-S
CM
o
co
Figure 2.4-6(a). Monthly mean, minimum, and maximum SOi concentrations at
Steubenville, OH for the years 2003 through 2005.
September 2007
2-15
DRAFT-DO NOT QUOTE OR CITE
-------
40 •
36..
32
28
CT)
3: 20
i
16' •
12- •
8 - •
4 ••
0 •
\
<&
2-
Figure 2.4-6(b). Monthly mean, minimum, and maximum SC>4 concentrations at
Steubenville, OH for the years 2003 through 2005.
o-1-
0
S02 (ppb)
Figure 2.4-6(c).
2-
Monthly mean SO4 concentrations as a function of SO2 concentrations
at Steubenville, OH for the years 2003 through 2005.
September 2007
2-16
DRAFT-DO NOT QUOTE OR CITE
-------
d
CO
45-
40-
35-
30 -
25
20 -
15"
10 ••
5 ••
o-L
Figure 2.4-7(a). Monthly mean, minimum, and maximum SOi concentrations at
Philadelphia, PA for the years 2003 through 2005.
70 T
60-
w -
-r
g 40 -
1
20
10 •
n
!•
:
•1
lfltjf
i
. •
•Hi!
Ill*
2-
Figure 2.4-7(b). Monthly mean, minimum, and maximum SC>4 concentrations at
Philadelphia, PA for the years 2003 through 2005.
September 2007
2-17
DRAFT-DO NOT QUOTE OR CITE
-------
C-
0
* 9
'•
;•
6
S02 (ppb)
10
12
2-
Figure 2.4-7(c). Monthly mean SO4 concentrations as a function of SOi concentrations
at Philadelphia, PA for the years 2003 through 2005.
16--
14-
12-
10 •
1
Q.
I
*
Figure 2.4-8(a).
Monthly mean, minimum, and maximum SOi concentrations at Los
Angeles, CA for the years 2003 through 2005.
September 2007
2-18
DRAFT-DO NOT QUOTE OR CITE
-------
45
«
35 -
4 concentrations at
Los Angeles, CA for the years 2003 through 2005.
20--
IS-
16--
14 - -
~ 12-
s
3 10 - •
o
01 e-
23
SO, (pp
2-
Figure 2.4-8(c). Monthly mean SC>4 concentrations as a function of SOi concentrations
at Los Angeles, CA for the years 2003 through 2005.
September 2007
2-19
DRAFT-DO NOT QUOTE OR CITE
-------
16 -r
12 •
10 •
SI
a.
O
V)
311
II
II
111
II
jf
Figure 2.4-9(a).
Monthly mean, minimum, and maximum SOi concentrations at
Riverside, CA for the years 2003 through 2005.
i
g
?
M
•1
10
?!
6
2
0
Figure 2.4-9(b).
2-
Monthly mean, minimum, and maximum SC>4 concentrations at
Riverside, CA for the years 2003 through 2005.
September 2007
2-20
DRAFT-DO NOT QUOTE OR CITE
-------
£
O)
O
CO
V
»
1 2
S02 (ppb)
2-
Figure 2.4-9(c). Monthly mean SC>4 concentrations as a function of SOi concentrations
at Riverside, CA for the years 2003 through 2005.
10 T
o J-
II
Iff
Figure 2.4-10(a). Monthly mean, minimum, and maximum SOi concentrations at
Phoenix, AZ for the years 2003 through 2005.
September 2007
2-21
DRAFT-DO NOT QUOTE OR CITE
-------
0 -1
* ,£• *
2-
Figure 2.4-10(b). Monthly mean, minimum, and maximum SC>4 concentrations at
Phoenix, AZ for the years 2003 through 2005.
4 T
3 +
•7
3. 2
O
a
i -
0.0 0.5 1.0 1.5
2.0 2.5 3.0
SOjlppb)
3.5 4.0 4.5 5.0
Figure 2.4-10(c). Monthly mean SO4 concentrations as a function of SO2
concentrations at Phoenix, AZ for the years 2003 through 2005.
September 2007
2-22
DRAFT-DO NOT QUOTE OR CITE
-------
1 Moving in order from the region of highest to lowest SC>2 concentrations, consider
2 Steubenville, OH (Figure 2.4-6), where the SC>2 concentrations were highest of all 12 CMSAs
3 with more than four monitors. Even here, however, all monthly mean SC>2 concentrations
4 (Figure 2.4-6a) were substantially <30 ppb, though maximum daily means in some months were
5 often >60 ppb, or even >90 ppb. Sulfate data at Steubenville (Figure 2.4-6b) were insufficient to
6 make meaningful comparisons, though the 12 months of available SC>42 data suggest no
7 correlation with SC>2 (see Figure 2.4-6c).
8 Next, consider Philadelphia, PA. 862 in Philadelphia, PA (Figure 2.4-7a) is present at
9 roughly one-half the monthly mean concentrations in Steubenville, OH (compare Figures 2.4-6a
10 and 2.4-7a), and demonstrates a strong seasonality with 862 concentrations peaking in winter.
11 By contrast, SC>42 concentrations (Figure 2.4-7b) in Philadelphia peak in the three summer
12 seasons, with pronounced wintertime minima. This seasonal anticorrelation still contains
13 considerable monthly scatter, however, as Figure 2.4-7c makes clear.
14 Los Angeles, CA (see Figure 2.4-8a-c) presents a special case since its size and power
15 requirements place a larger number of SC>2 emitters near it than would otherwise be expected on
16 the West Coast. Concentrations of 862 (Figure 2.4-8a) demonstrate weak seasonality in these
17 3 years, with summertime means of ~3 to 4 ppb, and maxima generally higher than wintertime
18 ones, though the highest means and maxima occur during the winter of 2004-2005. SC>42 at Los
19 Angeles (Figure 2.4-8b) shows stronger seasonality, most likely because the longer summer days
20 of sunny weather allow for additional oxidation of SC>2 to SC>42 than would be available in
21 winter. Weak seasonal effects in SC>2 likely explain the complete lack of correlation between
22 SO2 and SO42 here, as Figure 2.4-8c shows.
23 The Riverside, CA CMSA (see Figure 2.4-9a-c) presents the strongest example among
24 the 12 examined for this study of correlation between SO2 and SO42 (Figure 2.4-9c), though
25 even here the R2 value is merely 0.3. Seasonal peaks are obvious in summertime for 862 and
26 SC>42 , both at roughly half the ambient concentrations seen in Los Angeles (compare Figures
27 2.4-8a and 2.4-8b to Figures 2.4-9a and 2.4-9b). This is very likely due to Riverside's
28 geographic location just downwind of the regionally large sources near Los Angeles and the
29 prevailing westerly winds in summer. Again, as with Los Angeles, the summertime peaks in
30 SC>42 are most likely due to the combination of peaking 862 and favorable meteorological
31 conditions allowing more complete oxidation.
September 2007 2-23 DRAFT-DO NOT QUOTE OR CITE
-------
1 Phoenix, AZ was the CMSA with the lowest monthly mean SC>2 and SC>42
2 concentrations examined here (see Figures 2.4-10a and b). In Phoenix, nearly all monthly mean
3 SC>2 values were at or below the regulatory monitors' operating LOD of ~3 ppb. SC>42
4 concentrations were equivalently low, roughly one-half the concentrations seen in Riverside, CA,
5 for example. The monthly mean data in Figures 2.4-10a and 2.4-1 Ob show strong summertime
6 peaks for even these very low-level SC>42 observations, which, at ~1 to 3 jig nT3, were generally
7 one-half those in Philadelphia (compare Figure 2.4-7b). Figure 2.4-10a suggests some
8 seasonality in SO2, though anticorrelated with SO42 ; however, the trend is very weak, as the
9 correlation scatterplot (Figure 2.4-10c) shows.
10
11 2.4.3 Policy Relevant Background Concentrations of Sulfur Dioxide
12 Background concentrations of 862 used for purposes of informing decisions about
13 NAAQS are referred to as Policy Relevant Background (PRB) concentrations. PRB
14 concentrations are those concentrations that would occur in the United States in the absence of
15 anthropogenic emissions in continental North America (defined here as the United States,
16 Canada, and Mexico). PRB concentrations include contributions from natural sources
17 everywhere in the world and from anthropogenic sources outside these three countries.
18 Background levels so defined facilitate separation of cases where pollution levels can be
19 controlled by U.S. regulations (or through international agreements with neighboring countries)
20 from cases where pollution is generally uncontrollable by the United States. EPA assesses risks
21 to human health and environmental effects from SC>2 levels in excess of PRB concentrations.
22 Contributions to PRB concentrations include natural emissions of SO2 and photochemical
23 reactions involving reduced sulfur compounds of natural origin as well as their long-range
24 transport from outside of North America from whatever source. As an example, transport of SC>2
25 from Eurasia across the Pacific Ocean or the Arctic Ocean would carry PRB SO2 into the U.S. A
26 schematic diagram showing the major photochemical processes involved in the sulfur cycle
27 including natural sources of reduced sulfur species from anaerobic microbial activity in wetlands
28 and volcanic activity appears in Annex 2. Volcanoes and biomass burning are the major natural
29 source of SO2. Biogenic emissions from agricultural activities are not considered in the
30 formation of PRB concentrations. Discussions of the sources and estimates of emissions are
31 given in Annex Section 2.6.2.
September 2007 2-24 DRAFT-DO NOT QUOTE OR CITE
-------
1 Analysis ofPRB Contributions to Sulfur Oxide Concentrations and Deposition over the
2 United States
3 The MOZART-2 global model of tropospheric chemistry (Horowitz et al., 2003) is used
4 to estimate the PRB contribution to nitrogen and sulfur oxide concentrations, as well as to total
5 (wet plus dry) deposition. The model setup for the present-day simulation, i.e., including all
6 sources in the U.S. Canada and Mexico, was published in a series of papers from a recent model
7 intercomparison (Dentener et al., 2006a,b; Shindell et al., 2006; Stevenson et al., 2006; van Noije
8 et al., 2006). MOZART-2 is driven by the National Oceanic and Atmospheric Administration's
9 National Center for Environmental Prediction (NOAA/NCEP) meteorological fields and the
10 International Institute for Applied Systems Analysis (HASA) 2000 emissions at a resolution of
11 1.9° x 1.9° with 28 o (sigma) levels in the vertical, and includes gas- and aerosol-phase
12 chemistry. Results shown in Figure 2.4-11 are for the meteorological year 2001. An additional
13 PRB simulation was conducted in which continental North American anthropogenic emissions
14 were set to zero.
15 The role of PRB in contributing to SO2 concentrations in surface air is examined first.
16 Figure 2.4-11 shows the annual mean predicted SO2 concentrations in surface air in the
17 simulation including all sources, or the "base case" (top panel); the PRB simulation (middle
18 panel); and the percentage contribution of the background to the total base case SO2 (bottom
19 panel). Maximum concentrations in the base case simulation, >5 ppb, occur along the Ohio
20 River Valley (upper panel Figure 2.4-11). Background SO2 concentrations are orders of
21 magnitude smaller, below 10 parts per trillion (ppt) over much of the United States (middle
22 panel; of Figure 2.4-11). Maximum PRB concentrations of SO2 are 30 ppt. In the Northwest
23 where there are geothermal sources of SO2, the contribution of PRB to total SO2 is 70 to 80%.
24 However, with the exception of the West Coast where volcanic SO2 emissions cause high PRB
25 concentrations, the PRB contributes <1% to present-day SO2 concentrations in surface air
26 (bottom panel Figure 2.4-11).
27 When estimating background concentrations it is instructive to consider also
28 measurements of SO2 at relatively remote monitoring sites, i.e., sites located in sparsely
29 populated areas not subject to obvious local sources of pollution. Berresheim et al. (1993) used a
30 type of atmospheric pressure ionization mass spectrometer (APIMS) at Cheeka Peak, WA
31 (48.30°N 124.62°W, 480 m asl), in April 1991 during a field study for dimethyl sulfide (DMS)
September 2007 2-25 DRAFT-DO NOT QUOTE OR CITE
-------
Total
100°W
80°W
< 0.01 1.21 2.41 3.60
Background
4.80
B.OO
ppb
120°W
100°W
30°W
< 0.001 O.OOB 0.011 0.015 0.020
Percent Background Contribution
0.025
ppb
120°W
100°W
so°w
10
15
20
25
Figure 2.4-11.
Annual mean model-predicted concentrations of SOi (ppb) in surface
air over the United States in the present-day (upper panel) and policy
relevant background (middle panel) MOZART-2 simulations. The
bottom panel shows the percentage contribution of the background to
the present-day concentrations.
September 2007
2-26
DRAFT-DO NOT QUOTE OR CITE
-------
1 oxidation products. 862 concentrations ranged between 20 and 40 ppt. Thornton et al. (2002)
2 have also used an APIMS with an isotopically
3 labeled internal standard to determine background SO2 levels. SO2 concentrations of 25 to
4 40 ppt were observed in northwestern Nebraska in October 1999 at 150 m above ground using
5 the National Center for Atmospheric Research (NCAR)'s C-130 research aircraft. These data are
6 comparable to remote central South Pacific convective boundary layer SO2 data (Thornton et al.,
7 1999).
8 As noted earlier in Section 2.4.2, volcanic sources of 862 in the United States are found
9 in the Pacific Northwest, Alaska, and Hawaii. The most serious impact in the United States from
10 volcanic 862 occurs on the island of Hawaii. Nearly continuous venting of 862 from Mauna
11 Loa and Kilauea produces SO2 in such large amounts that as far as > 100 km downwind of the
12 island SC>2 concentrations can exceed 30 ppb (Thornton and Bandy, 1993). Depending on the
13 wind direction, the west coast of Hawaii (Kona region) has had significant impacts from SC>2 and
14 acidic SC>42 aerosols for the past decade. Indeed, 862 levels in Volcanoes National Park, HI
15 exceeded both the secondary 3-h and the primary 24-h average (24-h avg) NAAQS in 2004-
16 2005. Since 1980, the Mount St. Helens volcano in Washington Cascade Range (46.20°N,
17 122.18°W, summit 2549 m asl) has been a variable source of SC>2. Its major effects came in the
18 explosive eruptions of 1980, which primarily affected the northern part of the mountain west of
19 the United States. The Augustine volcano near the mouth of the Cook Inlet in southwestern
20 Alaska (59.363 °N, 153.43 °W, summit 1252 m asl) has emitted variable quantities of SO2 since
21 its last major eruptions in 1986. Volcanoes in the Kamchatka peninsula in far eastern Siberia do
22 not particularly affect the surface concentrations in northwestern North America.
23 Overall, the background contribution to SOX over the United States is relatively small,
24 with a maximum PRB of 0.030 ppb SC>2, except for areas with volcanic activity.
25
26
27 2.5 ISSUES ASSOCIATED WITH EVALUATING EXPOSURES TO
28 SULFUR OXIDES
29
30 2.5.1 General Considerations for Personal Exposures
31 Human exposure to an airborne pollutant consists of contact between the human and the
32 pollutant at a specific concentration for a specified period of time. People spend various
September 2007 2-27 DRAFT-DO NOT QUOTE OR CITE
-------
1 amounts of time in different microenvironments (Figure 2.5-1) characterized by different
2 pollutant concentrations. The figure represents a composite average across the United States
3 across all age groups. Different cohorts, e.g., the elderly, might be expected to exhibit different
4 activity patterns. The integrated exposure of a person to a given pollutant is the sum of the
5 exposures over all time intervals for all microenvironments in which the individual spent time.
NHAPS - Nation, Percentage Time Spent
Total n = s>.ll>6
TOTAL TIME SPENT
IN A RESIDENCE (Mi.7%( ^- -^ I I INDOORS
-------
1 where Et is the time-weighted personal exposure concentration over a certain period of time, n is
2 the total number of microenvironments that a person encounters,/ is the fraction of time spent in
3 the rth microenvironment, and Ct is the average concentration in the rth microenvironment during
4 the time fraction,/. The types of exposure a person experiences can be characterized as an
5 instantaneous exposure, a peak exposure, an average exposure, or an integrated exposure over all
6 the environments a person encounters. These distinctions are important because health effects
7 caused by long-term, low-level exposures may differ from those caused by short-term, peak
8 exposures.
9 An individual's total exposure (Et) can also be represented by the following equation:
22
ET= Ea + Ena = {y() + ^yf [PM/fa + k,)} }Ca + Ena = {y0 + I jyF/w/:} Ca + Ena
10 ' / ' (2-8)
1 1 subj ect to the constraint
y
12 / (2-9)
13 where Ea is the person's exposure to pollutants of ambient origin; Ena is the person's exposure to
14 pollutants that are not of ambient origin;^ is the fraction of time people spend outdoors and^ is
15 the fraction of time they spend in the rth microenvironment; Finfh Pt, at, and kt are the infiltration
16 factor, penetration coefficient, air exchange rate, and decay rate for a pollutant in the rth
17 microenvironment. In this equation, it is assumed that each microenvironment is well mixed
18 (i.e., concentrations are homogeneous) and that air exchange occurs with ambient air only, not
19 between microenvironments.
20 In the case where microenvironmental exposures occur mainly in one microenvironment,
21 Equation 2-8 may be approximated by
= {y + (l-y)[Pa/(a + k)]}Ca + Ena = aCa+ E
na
23 where y is the fraction of time people spend outdoors, a is the ratio of a person's exposure to a
24 pollutant of ambient origin to the pollutant' s ambient concentration. Other symbols have the
25 same definitions in Equation 2-8 and 2-9. If microenvironmental concentrations are considered,
26 then Equation 2-10 can be recast as
September 2007 2-29 DRAFT-DO NOT QUOTE OR CITE
-------
! Cme = Ca + Cna = [Pa /(a + k)]Ca + S/\V(a + k)] ^ 1}
2 where Cme is the concentration in a microenvironment; Ca and Cna are the contributions to Cme
3 from ambient and nonambient sources; S is the microenvironmental source strength; Fis the
4 volume of the microenvironment, and the symbols in brackets have the same meaning as in
5 Equation 2-10.
6 Microenvironments in which people are exposed to air pollutants typically include
7 residential indoor environments, other indoor locations, outdoor environments, and in vehicles,
8 as shown in Figure 2.5-1. Indoor combustion sources such as kerosene space heaters need to be
9 considered when evaluating exposures to SO2. Exposure misclassification may result when total
10 human exposure is not disaggregated between various microenvironments, and this may obscure
11 the true relationship between ambient air pollutant exposures and health outcomes.
12 In a given microenvironment, the ambient component of a person's microenvironmental
13 exposure to a pollutant is determined by the following physical factors:
14 • Ambient concentration,
15 • The air exchange rate,
16 • The pollutant specific penetration coefficient,
17 • The pollutant specific decay rate, and
18 • The fraction of time an individual spends in the microenvironment.
19
20 These factors are in turn determined by the following potential exposure factors:
21 • Environmental conditions, such as weather and season;
22 • Dwelling conditions, such as the location of the house which determines proximity to
23 sources and geographical features that can modify transport from sources; the amount of
24 natural ventilation (e.g., open windows and doors, and the "draftiness" of the dwelling)
25 and ventilation system (e.g., filtration efficiency and operation cycle);
26 • Personal activities (e.g., time spent cooking or commuting);
27 • Socioeconomic status (e.g., level of education and the income level);
28 • Demographic factors (e.g., age and gender);
29 • Indoor sources and sinks of a pollutant; and
30 • Microenvironmental line and point sources (e.g., lawn equipment).
September 2007 2-30 DRAFT-DO NOT QUOTE OR CITE
-------
1 In general, the relationship between personal exposures and ambient concentrations can
2 be modified by microenvironments in the following ways. (1) Ambient pollutants can be lost
3 through chemical and physical loss processes during infiltration, and therefore, the ambient
4 component of a pollutant's concentration in a microenvironment is not the same as its ambient
5 concentration. Instead it is the product of the ambient concentration and the infiltration factor
6 (Finfor a if people spend 100% of their time indoors) and (2) exposure to nonambient,
7 microenvironmental sources.
8 Time activity diaries, completed by study participants, are often used in exposure models
9 and assessments. The EPA's National Exposure Research Laboratory (NERL) has consolidated
10 the majority of the most significant human activity databases into one comprehensive database
11 called the Consolidated Human Research Laboratory Database (CHAD). Eleven different
12 human activity pattern studies were united to obtain over 22,000 person-days of 24-h human
13 activities in CHAD (McCurdy et al., 2000). These data can be useful in assembling population
14 cohorts for exposure modeling and analysis and determining inhalation rates for dosimetry
15 calculations.
16 In practice, it is extremely difficult to characterize community exposures by
17 measurements of each individual's personal exposures. Instead, the distribution of personal
18 exposures in a community, or the population exposure can be characterized by extrapolating
19 measurements of personal exposure using various techniques or by stochastic, deterministic, or
20 hybrid exposure modeling approaches such as APEX, SHEDS, and MENTOR (see Annex AX3
21 for a description of these modeling methods). Variations in community-level personal exposures
22 are determined by cross-community variations in ambient pollutant concentrations and the
23 physical and exposure factors mentioned above. These factors also determine the strength of the
24 association between population exposure to SO2 of ambient origin and ambient SC>2
25 concentrations.
26
27 2.5.2 Methods Used for Monitoring Personal Exposure to SO2
28 Three basic methods of analysis have been used as personal exposure monitors (PEMs) to
29 measure personal exposure to 862. The Harvard-EPA annular denuder system (HEADS) was
30 initially developed to measure particles and acid gases simultaneously (Koutrakis et al, 1988).
31 The aerosol is initially sampled at 10 L/min through an impactor that is attached to an annular
September 2007 2-31 DRAFT-DO NOT QUOTE OR CITE
-------
1 denuder to remove particles. Subsequently, the aerosol is sampled through an annular denuder
2 coated with sodium carbonate (Na2CO3). This denuder is used to trap SO2, nitric acid (HNOs),
3 and nitrous acid (HNO2). Following sampling, the denuder is extracted with ultrapure water and
4 analyzed by ion chromatography. Collection efficiencies of SO2 in the denuder are typically
5 around 0.993, which compares well with predicted values. Because the HEADS system is not
6 easily converted for use as a PEM, other personal monitoring systems have been employed more
7 recently in exposure monitoring studies.
8 For a study conducted in Baltimore, MD, Chang et al., (2000) developed and employed a
9 personal roll-around system (RAS, an active sampling system designed to measure short-term
10 exposure) to measure personal exposure concentrations of several atmospherically relevant
11 species, including SO2. For the measurement of SO2, the RAS employed an NO2/SO2 sorbent
12 denuder worn on a vest by the study participant. The hollow glass denuder, incased in an
13 aluminum jacket, is coated with triethanolamine (TEA) for the collection of SO2 and NO2, and
14 aerosol is sampled through the denuder at 100 cc/min. Following sampling, the denuder can be
15 extracted and analyzed for SO2 concentrations by ion chromatography. The detection limit for
16 1-h sampling of SO2 was reported to be 66 ppb.
17 The most commonly employed SO2 PEM method for personal exposure studies is the
18 passive badge sampler. A personal multipollutant sampler has been developed to measure
19 particulate and gaseous pollutants simultaneously (Demokritou et al, 2001). A single elutriator,
20 operating at 5.2 L/min, is employed to sample parti culate pollutants. A passive SO2 badge is
21 attached diametrically to the elutriator, which has been coated with Teflon to minimize reactive
22 gas losses. The passive badge sample is coated with TEA for the collection of SO2 and NO2.
23 Because wind speed can affect the collection rate of the passive badge sampler, this system
24 employs a constant face velocity across the passive badge sampler. For 24-h sampling times, the
25 estimated limit of detection (LOD) for SO2 is 5 ppb.
26 Currently, limits exist in using PEM systems to measure personal exposure to SO2.
27 Because SO2 concentrations have been declining annually in the United States, little focus has
28 been placed on improving methods of analysis for SO2. LODs for SO2 PEMs (~5 ppb) are often
29 greater than the concentrations of SO2 that are typically observed in urban ambient
30 environments. Personal exposure monitoring studies often suffer from many of the SO2 samples
31 (30 to 70%) being collected being below the sampler's LOD.
September 2007 2-32 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.5.3 Relationships between Personal Exposures and Ambient
2 Concentrations
3 Relationships between personal, indoor, outdoor, and ambient concentrations are
4 examined in this section. Because SC>2 concentrations have declined markedly over the past few
5 decades, relatively few studies have focused on 862 since the last AQCD for 862 was published.
6 Another consideration is that currently, indoor and outdoor levels in many areas are often
7 beneath detection personal monitor limits for 862.
8
9 2.5.3.1 Indoor versus Outdoor SOi Concentrations
10 Several studies in the United States, Canada, Europe, and Asia have examined the
11 relationships of indoor, outdoor, and personal concentrations of SO2 to ambient SO2
12 concentrations. Perhaps the most comprehensive set of indoor-outdoor data was obtained by
13 Spengler et al. (1979) during the Harvard Six Cities Study. These data are shown in Figure
14 2.5-2. Twenty-four-hour ambient and indoor SC>2 concentrations were measured every sixth day
15 for 1 year in a minimum of 10 homes or public facilities for each of the cities studied. One-year
16 average concentrations for indoor and outdoor concentrations of SO2 for each city studied are
17 shown in Figure 2.5-2.
18 A summary of ratios of indoor to outdoor concentrations found in this and other studies is
19 given in Table 2.5-1. As can be seen from Table 2.5-1, a wide range is found in the ratio of
20 indoor to outdoor concentrations among the different studies. These differences among studies
21 could be due in part to differences in building characteristics (e.g., residences versus schools or
22 other public buildings), in activities affecting air exchange rates, and in analytical capabilities.
23 In several studies, high values for R2 were found, suggesting that indoor levels were largely
24 driven by outdoor levels. A few studies found higher levels of SO2 indoors than outdoors in
25 some samples. This situation could have arisen if there were indoor sources or because of
26 analytical measurement issues. One would expect to find lower concentrations indoors than
27 outdoors, because SO2 is consumed by reactions on indoor surfaces, especially those that are
28 moist. Chao (2001) acknowledged this point but could not account for the findings of this study.
29 It was noted that two samples had unusually high indoor to outdoor ratios and that the mean
30 ratios would have been much lower otherwise. Winter-summer differences in the indoor:outdoor
31 ratio are consistent with seasonal differences in air exchange rates, as noted by Brauer et al.
32 (1991).
September 2007 2-33 DRAFT-DO NOT QUOTE OR CITE
-------
50
?T 40
E
"5)
,3 30
61
<° 20
10
Outdoor
Indoor
PORT TOPE
KING
WAT
STL
STEU
Figure 2.5-2. Average annual indoor and outdoor SOi concentrations for each of the six
cities included in the analysis. PORT = Portage, WI; TOPE = Topeka, KS;
KING = Kingston, TN; WAT = Watertown, MA; STL = St. Louis, MO;
STEU = Steubenville, OH.
Source: Adapted from Spengler et al. (1979).
1 Indoor, or nonambient, sources of SC>2 could complicate associations between personal
2 exposure to ambient SC>2 and ambient SC>2. Possible sources of indoor SC>2 are associated with
3 the use of sulfur-containing fuels, with higher levels expected when emissions are poorly vented.
4 Brauer et al. (2002) noted that only one study (Biersteker et al., 1965) conducted inferential
5 analyses of potential determinants of exposure to indoor 862 levels. In the Biersteker et al.
6 study, conducted in the Netherlands, indoor levels increased with oil, coal, and gas heating and
7 smoking in homes and with increased outdoor levels.
8 Triche et al. (2005) measured SC>2 levels in homes in which secondary heating sources
9 (fireplaces, kerosene heaters, gas space heaters, and wood stoves) were used. They found
10 elevated indoor levels of SC>2 when kerosene heaters were in use. Median levels of SC>2 when
11 kerosene heaters were used (6.4 ppb) were much higher than when they were not in use
12 (0.22 ppb). The maximum 862 level associated with kerosene heater use was 90.5 ppb. They
13 did not find elevated 862 levels when the other secondary heating sources were in use.
September 2007
2-34
DRAFT-DO NOT QUOTE OR CITE
-------
1 2.5.3.2 Relationship of Personal Exposure to Ambient Concentrations
2 A few studies evaluated the association of personal exposure to SO2 to ambient
3 concentrations (Brauer et al., 1989; Chang et al., 2000; Sarnat et al., 2000, 2001, 2005, 2006).
4 Some of these studies fall under the umbrella of the Health Effects Institute's Characterization of
5 Particulate and Gas Exposures of Sensitive Subpopulations Living in Baltimore and Boston
6 research plan (Koutrakis et al., 2005). However, the focus of many of these studies has been
7 exposure to particles, with acid gases included to evaluate confounder or surrogate issues.
8 Brauer et al. (1989) determined the slope of the regression line between personal and
9 ambient concentrations to be 0.13 ± 0.02, R2 = 0.43, based on 44 measurements made in Boston,
10 MA during the summer of 1988. Most if not all of the data points obtained using the HEADS
11 appeared to be above analytical detection limits based on the use of laboratory blanks and ion
12 chromatography instrument sensitivity instead of field blanks, which are used in most other
13 studies to calculate the overall method detection limit. Note that calculating detection limits in
14 this way could result in lower detection limits than if field blanks are used. The authors reported
15 significance at the p<0.001 level, but the intercept was not significant at the p<0.001 level.
16 Since the stationary monitoring site was located at an elevation of 250 m above street level, the
17 use of data from this ambient monitoring site will overestimate personal exposure, as the
18 concentration of SO2 increases with height because it is emitted mainly by elevated point
19 sources. Indeed, the ambient concentrations are about a factor of two higher than the outdoor
20 concentrations.
21 A few personal exposure studies were conducted in Baltimore, MD (Chang et al., 2000;
22 Sarnat et al., 2000, 2001). Chang et al. (2000) tested a new personal active sampling device
23 (a RAS with a TEA-based denuder) on volunteer participants to measure hourly personal
24 exposure to 862. However, the method detection limit was too high for SCh (62 ppb for
25 1-h sampling) to generate a robust 862 exposure dataset to perform further analysis, and so the
26 authors did not use the 862 data for this purpose. Sarnat et al. (2000) reported a longitudinal
27 exposure study with older adults as participants. Twenty-four-hour averaged personal SO2
28 exposures were measured with TEA-based passive sampler badges. The authors reported that
29 70% of the personal exposure concentrations of SC>2 were below the method detection limit
30 (6.5 ppb for 24-h sampling). The mean ambient and personal exposure concentrations were
31 reported as 8.9 and 0.0 ppb, respectively, during February and March of 1999. The maximum,
September 2007 2-3 5 DRAFT-DO NOT QUOTE OR CITE
-------
1 minimum, and median Spearman rank correlation coefficients between personal exposure and
2 ambient concentrations over 12 days for 14 participants were 0.65, -0.75, and 0.02, respectively.
3 Sarnat et al. (2001) reported another (8- to 12-day) longitudinal exposure study with a cohort that
4 was similar to that used in Sarnat et al. (2000) except for including children and patients with
5 chronic obstructive pulmonary disease (COPD). Data quality was not specifically discussed in
6 Sarnat et al., (2001), but the readers were referred to Sarnat (2000) and Chang et al. (2000) for
7 information about precision, accuracy, and method-detection limits. During the study, the
8 median ambient and personal exposures were 8 and 1 ppb, respectively (estimated from Figure
9 1 in Sarnat et al., 2001). The authors reported that during the winter of 1999, ambient 862 was a
10 significant predictor (at 5% significance level) of personal exposure to SO2 (slope = -0.05),
11 personal exposure to fine particulate matter (PM^.s) (slope = -0.24), personal exposure to SC>42
12 (slope = -0.03), and personal exposure to PM2.5 of ambient origin (slope = -0.16). However, it
13 should be noted that all the slopes are negative.
14 Sarnat et al. (2005) conducted a longitudinal 12-day exposure study on 43 children and
15 older adults in Boston, MA during the summer of 1999 and the following winter (1999-2000).
16 They reported that 95.4 and 96.5% of the 862 concentrations were below detection limits
17 (3.2 and 2.3 ppb, respectively, for winter and summer 24-h sampling). The absolute and relative
18 sampling precisions for SO2 were 0.8 ppb and 69.5%, respectively. The authors reported that the
19 mean ambient concentrations ranged from 2.8 to 10.7 ppb during the study, while the mean
20 personal exposure concentrations were <1.9 ppb. Associations between ambient SC>2 and either
21 personal exposures or ambient concentrations of other pollutants were found for personal SC>42
22 (winter, slope = 0.06), personal SC>42 (summer, slope = 0.39), personal PM2.5 (summer,
23 slope = 1.68), ambient SC>42 (winter, slope = 0.19), and ambient PM2.5 (winter, slope = 0.80).
24 Sarnat et al. (2006) reported the results of a personal exposure study in Steubenville, OH.
25 The authors reported that 36.5 and 33.8% of ambient 862 were below the detection limit during
26 the summer (5.5 ppb) and fall (3.8 ppb), and 53.5 and 36.1% of personal concentrations of 862
27 were below the detection limit during the summer (5.5 ppb) and fall (3.8 ppb), respectively. On
28 average, personal exposures were lower than the ambient concentrations (1.5 ppb for personal
29 and 2.7 for ambient during the summer; 0.7 ppb for personal and 5.4 ppb for ambient during the
30 fall); however, the maximum personal exposure could be higher than the ambient concentration
31 (30.4 ppb for personal and 21.9 ppb for ambient during the summer). Ambient SC>2 was
September 2007 2-36 DRAFT-DO NOT QUOTE OR CITE
-------
1 observed to be significantly associated with personal 862 exposures during the fall (slope = 0.08
2 for overall population, 0.07 for subjects in buildings with low ventilation rates, and 0.13 for
3 subjects in buildings with high ventilation rates).
4 Of significant concern is the ability of currently available techniques for monitoring
5 either personal exposures or ambient concentrations to measure SC>2 concentrations that are
6 typically found in most urban environments. In some studies, most data, especially data for
7 monitoring personal exposure and indoor concentrations, might be beneath detection limits.
8 Indeed, in one study (Chang et al., 2000), the investigators had to discard data for 862, because
9 the values were mostly beneath detection limits. In the study of Kindzierski and Ranganathan
10 (2006), all indoor concentration data were beneath detection limits. In Sarnat et al. (2000),
11 -70% of personal measurements were beneath detection limits, and -33% of personal
12 measurements returned apparent negative concentration values. In such situations, associations
13 between ambient concentrations and personal exposure are inadequately characterized. When
14 personal exposure concentrations are above detection limits, a reasonably strong association is
15 observed between personal exposures and ambient concentrations.
16
17 2.5.4 Exposure Measurement Errors in Epidemiological Studies
18 For the purposes of the draft Integrated Science Assessment (ISA), the effects of
19 exposure error on epidemiological study results refers to changes in the point estimate and in the
20 standard error of the calculated health effect estimate, P, that result from using the concentration
21 of an air pollutant as an exposure indicator rather than using the actual personal exposure to the
22 causal factor in the epidemiological statistical analysis. There are many assumptions made in
23 going from the available experimental measurement of a pollution indicator to an estimate of the
24 personal exposure to the causal factor. The importance of these assumptions and their effect on
25 P depend on the type of epidemiological study.
26 The considerations of exposure error for SC>2 are simplified compared to those for NC>2
27 and PM. The only experimental measure available is the ambient concentration of SC>2. In
28 addition, indoor and other nonambient sources of 862 are not thought to be important in
29 population studies, lessening concerns about the possible influence of exposures other than to
30 ambient 862. The only known significant indoor source of 862 in the United States is the use of
31 kerosene heaters, which is not thought to be widespread enough to influence population studies.
September 2007 2-37 DRAFT-DO NOT QUOTE OR CITE
-------
1 In addition, as is the case with other air pollutants, exposure to nonambient 862 would not affect
2 P in time-series studies using ambient concentrations as the exposure surrogate unless the
3 nonambient exposures were correlated with the ambient concentrations.
4
5 2.5.4.1 Community Time-Series Studies
6 This section applies primarily to studies of the association of daily average SO2
7 concentrations with daily measures of mortality or morbidity. With SCh time-series
8 epidemiological analysis, the following four exposure issues are of primary concern: (1) the
9 relationship of the experimental measurement of SC>2 to the true concentration of SO2; (2) the
10 relationship of day-to-day variations of the concentration of the indicator, as measured at a
11 central monitoring site, with the corresponding variations in the average concentration of the
12 indicator over the geographic area from which the health measurements are drawn; (3) the
13 relationship of the community average concentration of SO2 to the average personal exposure to
14 ambient SC>2; and (4) the relationship of SC>2 to the true causal factor. These four issues are
15 described below.
16
17 2.5.4.1.1 Relationship of Experimental Measurement of SO2 to the True Concentration
18 Since there is always some instrumental measurement error, the correlation of the
19 measured SC>2 with the true SC>2, on either a 24-h or 1-h basis, will be less than 1. Sheppard
20 et al. (2005) indicate that instrument error in the individual or daily average concentrations have
21 "the effect of attenuating the estimate of a." However, Zeger et al. (2000) state that the
22 "instrument error in the ambient levels is close to the Berkson type" and in order for this error to
23 cause substantial bias in PC, the error term (the difference between the true concentrations and
24 the measured concentrations) must be strongly correlated with the measured concentrations.
25 Zeger et al. (2000) suggest that, "Further investigations of this correlation in cities with many
26 monitors are warranted." Averaging across multiple unbiased ambient monitors in a region
27 should reduce the instrument measurement error (Sheppard et al., 2005; Wilson and Brauer,
28 2006; Zeger et al., 2000). There are concerns about the precision and accuracy of the ambient
29 concentration measurements, because 862 concentrations are much lower now than when the
30 SC>2 standards were first promulgated. Current ambient concentrations of 862 in the United
31 States are nearly all at or very near the detection limit of the monitors currently used in the
32 regulatory network. Thus, greater uncertainty is most often observed at the lower ambient
September 2007 2-3 8 DRAFT-DO NOT QUOTE OR CITE
-------
1 concentrations as compared with the less frequent higher concentration exposures because of the
2 plume downwash near local sources or entrainment of plumes downwind from large power
3 plants or smelters. It is unclear how uncertainties in the true concentrations of SO2, i.e.,
4 instrument measurement error, will change p. Zeger et al. (2000) suggest that instrument error
5 has both Berkson and non-Berkson error components.
6
7 2.5.4.1.2 Relationship of Day-to-day Variations of the Concentration of the Indicator
8 There has been little analysis of the spatial variation of SO2 across communities. SO2 is
9 thought to come primarily from power plants or smelters. New power plants and smelters in the
10 United States generally have 862 emission controls and are no longer located within urban areas.
11 However, older sources may still be located within urban areas and may not have as effective
12 SC>2 emission controls. Therefore, it is not clear whether 862 will act as a regional or local
13 pollutant and whether its spatial behavior might differ in different cities. Site-to-site correlations
14 of SC>2 concentrations, as shown for several cities in Table 2.4-3 include some very low values.
15 This suggests the concentration of SC>2, measured at any given monitoring site, may not be
16 highly correlated with the average community concentration. This could be due to local sources
17 that cause the SC>2 to be unevenly distributed spatially, to a monitoring site being chosen to
18 represent a nearby source, or to terrain features, source, or sink locations that divide the
19 community into several subcommunities that differ in the temporal pattern of pollution. It is also
20 possible that errors in the measurement of the low concentrations of 862 present at most sites
21 contribute to the lack of high correlations between monitors. To the extent that the correlation of
22 the ambient concentration with the community average concentration is <1, P will be reduced if
23 the single pollutant model is the true model. Similarly, P will be reduced if there are subareas of
24 the community where the correlation of the subarea average concentrations with the
25 concentrations measured at the ambient monitoring site is <1. Concentrations in an area of a
26 community impacted by plumes from local SC>2 sources or a large power plant or smelter might
27 be higher than, and not well-correlated with, the concentrations measured at the community
28 measurement site. If such high concentrations affected a sizable portion of the population, that
29 community might not be suitable for time-series epidemiological analyses.
30
September 2007 2-39 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.5.4.1.3 Relationship of Community Average Concentration of SO2 to Average Personal
2 Exposure to Ambient SO2
3 People spend much of their time indoors and, in the absence of indoor sources, indoor
4 concentrations are lower than outdoor concentrations. It is necessary to consider how this
5 difference between the ambient concentration, which is used in epidemiological analyses, and the
6 personal exposure to the ambient concentration (which includes exposure to the full outdoor
7 concentration while outdoors and exposure of only a fraction of the outdoor concentrations while
8 indoors) will affect the calculated p. The contribution of the ambient concentration of SO2 to the
9 personal exposure to ambient SC>2 is given by EA = a • C where EA is exposure to ambient SC>2, a
10 is the exposure factor (or more correctly the ambient exposure factor) with value between 0 and
11 1 as defined in Equation 2-10, and C is the ambient 862 concentration as measured at a
12 community monitoring site. Zeger et al. (2000) made a major contribution to our understanding
13 of exposure error by pointing out that for community time-series epidemiology, which analyzes
14 the association between health effects and potential causal factors at the community scale rather
15 than the individual scale, it is the correlation of the daily community average personal exposure
16 to the ambient concentration, XA, with daily community average concentration, Ct, that is
17 important, not the correlation of each individual's exposure JG/4 with Ct. Thus, the low
18 correlation of XitA with Cf, as frequently found in pooled panel exposure studies, is not relevant to
19 error in community time-series epidemiological analysis. Unfortunately, few experimental
20 studies provide adequate information to calculate the community average exposure. Most
21 exposure panel studies measure one or a few subjects on 1 day, and another one or a few subjects
22 on the next day, etc. (i.e., a pooled study design). A few studies have measured one subject for
23 several days and another subject for a different several days (i.e., a longitudinal study design).
24 However, in order to use experimental data to calculate a community average ambient exposure,
25 it is necessary to measure the personal exposure of every subject on every day and to have
26 sufficient information to estimate the ambient exposure from the measured total personal
27 exposure. Such information is available from one study of combined coarse and fine PM (PMio)
28 and shows that the correlation of Xf with Ct is much greater than the correlation of XitA with Ct
29 (U.S. Environmental Protection Agency, 2004). The Research Triangle Park PM Panel Study
30 found similar effects in the relationship of outdoor and personal PM2.s concentrations (Williams
31 et al., 2003). Ott et al. (2000) have provided a statistical argument that such an increase in the
32 correlation of the daily average over the individual values should be expected.
September 2007 2-40 DRAFT-DO NOT QUOTE OR CITE
-------
1 There has also been concern with the variation of a. Zeger et al. (2000) have stated (for
2 PM) and Sheppard et al. (2005) have used simulations (for PM or other nonreactive pollutant
3 such as CO) to show that the variations in individual daily values of a# around the daily average
4 at is a Berkson error and will not change the point estimate of P, although it may increase the
5 standard error. Sheppard et al. (2005) have shown that day-to-day variations in the average a
6 will not change the point estimate unless at is correlated with Ct. (Since most time-series
7 epidemiology uses 24-h concentrations, no anaylsis is available for shorter time periods.)
8 Both Zeger et al. (2000) and Sheppard et al. (2005) show that if $A is the health effect
9 parameter that would be obtained with an epidemiological analysis using the ambient exposure
10 and PC is the health effect parameter that would be obtained with an epidemiological analysis
11 using the ambient concentration, Ct, then pc = a • P^. Thus, an epidemiological analysis using
12 the ambient concentration, Ct, yields not P^, but a • P^. Overestimation of exposure by
13 substitution of the ambient concentration for the ambient exposure leads to underestimation of
14 the effect estimate, or bias toward the null.
15
16 2.5.4.1.4 Relationship of SO2 to the True Causal Factor
17 The remaining and most critical assumption is whether SC>2 is the causal factor (pollutant
18 that causes the examined health effect) or whether 862 is a surrogate for some other pollutant,
19 mixture of other pollutants, or mixture of pollutants including 862 that is the true causal factor.
20 For example, depending on the source of 862, 862 might be a surrogate for vanadium and nickel
21 from oil-fired power plants; selenium, arsenic, and mercury from coal-fired power plants; and/or
22 nickel and copper from smelters. The current data do not permit a quantitative assessment of the
23 relative contribution of SC>2 and correlated pollutants to the observed P value.
24
25 2.5.4.2 Long- Term Cohort Studies
26 For long-term exposure epidemiologic studies, concentrations are integrated over time
27 periods of a year or more, and usually for spatial areas the size of a city, county, or metropolitan
28 statistical area (MSA), although integration over smaller areas may be feasible. Health effects
29 are then regressed, in a statistical model, against the average concentrations in the series of cities
30 (or other areas). In time-series studies, a constant difference between the measured and the true
31 concentration (instrument offset) will not affect P, nor will variations in the daily average a or
32 the daily average nonambient exposure, unless the variations are correlated with the daily
September 2007 2-41 DRAFT-DO NOT QUOTE OR CITE
-------
1 variations in concentrations. However, in long-term exposure epidemiologic studies, if
2 instrument measurement errors, long-term average values of a, or long-term averages of
3 nonambient exposure differ for different cities (or other areas used in the analysis), the city-to-
4 city long-term ambient SC>2 concentrations will not be perfectly correlated with the long-term
5 average exposure to either ambient or total SC>2. This lack of correlation would be expected to
6 lead to a lowering of the point estimate of p.
7 In summary, the use of ambient concentrations of SC>2 as a surrogate for exposure to
8 ambient 862 is not generally expected to change the principal conclusions from 862
9 epidemiological studies, because the errors and uncertainties would be expected to reduce rather
10 to increase p. However, 862 may not be the causal agent, or the sole causal agent, but may be
11 serving as a surrogate for some other pollutant, or mix of pollutants, whose concentration is
12 correlated with that of SC>2. This may be particularly relevant for SC>2 because of atmospheric
13 chemistry linking it to its oxidation products SC>42" and to fine particulate matter. Therefore,
14 while population health risk estimates derived using ambient SC>2 levels are useful, evidence
15 from clinical and animal toxicological studies also needs to be considered in attempting to
16 understand the potential effects of 862 on human health.
17
18
19 2.6 DOSIMETRY OF INHALED SO2
20 This section is intended to present an overview of general concepts related to the
21 dosimetry of SC>2 in the respiratory tract. Dosimetry of SC>2 refers to the measurement or
22 estimation of the amount of SC>2 or its reaction products reaching and persisting at specific
23 respiratory tract sites after exposure. One of the principal effects of inhaled 862 is that it
24 stimulates bronchial epithelial receptors and initiates a reflexive contraction of smooth muscles
25 in the bronchial airways. The compound most directly responsible for health effects may be the
26 inhaled SO2 or perhaps its chemical reaction products. Complete identification of the causative
27 agents and their integration into SC>2 dosimetry is a complex issue that has not been thoroughly
28 evaluated. Few studies have investigated SC>2 dosimetry since the 1982 AQCD and the 1986
29 Second Addendum.
30 The major factors affecting the transport and fate of aerosols and gases in the respiratory
31 tract are the morphology of the respiratory tract; the physiochemical properties of the mucous
32 and surfactant layers; tidal volume, flow rate, and route of breathing; physicochemical properties
September 2007 2-42 DRAFT-DO NOT QUOTE OR CITE
-------
1 of the gas; and the physical processes that govern gas transport. When 862 contacts the fluids
2 lining the airways, it dissolves into the aqueous fluid and forms hydrogen (H+) ions and bisulfite
3 (HSO3 ) and sulfite (SO32~) anions (Bascom et al., 1996). The majority of anions are expected to
4 be present as HSO3 at a concentration proportional to the gas phase concentration of 862
5 (Ben-Jebria et al., 1990). Because of the chemical reactivity of these anions, various reactions
6 are possible, leading to the oxidation of SO32~ to SO42 (see Section 12.2.1, U.S. Environmental
7 Protection Agency, 1982). Clearance of SO32~ from the respiratory tract may involve several
8 intermediate chemical reactions and transformations (see Section 12.2.1.2, U.S. Environmental
9 Protection Agency, 1982). Gunnison and Benton (1971) identified ^-sulfonate in blood as a
10 reaction product of inhaled SC>2.
11 Physicochemical properties of SC>2 relevant to respiratory tract uptake include its
12 solubility and diffusivity in epithelial lining fluid (ELF), as well as its reaction-rate with ELF
13 constituents. Henry's law relates the gas phase and liquid phase interfacial concentrations at
14 equilibrium and is a function of temperature and pressure. Henry's law shows that the amount of
15 SC>2 in the aqueous phase is directly proportion to the partial pressure or concentration of 862 in
16 the gas phase. Although the solubility of most gases in mucus and surfactant is not known, the
17 Henry's law constant is known for many gases in water. The Henry's law constant for SC>2 is
18 0.048 (mole/liter)air / (mole/liter)water at 37 °C and 1 atm; for comparison, the value for O3 is
19 6.4 under the same conditions (Kimbell and Miller, 1999). In general, the more soluble a gas is
20 in biological fluids, the sooner, and more proximally, it is absorbed in the respiratory tract.
21 When the partial pressure of 862 on mucosal surfaces exceeds that of the gas phase, such as
22 during expiration, some desorption of 862 from the ELF may be expected.
23 Because 862 is highly soluble in water, it is expected to be almost completely absorbed
24 in the nasal passages of subjects at rest. The dosimetry of SO2 can be contrasted with the lower
25 solubility gas, O3, for which the predicted tissue doses (O3 flux to liquid-tissue interface) are
26 very low in the trachea and increase to a maximum in the terminal bronchioles or first airway
27 generation in the pulmonary region (see Chapter 4, U.S. Environmental Protection Agency,
28 2006b). Similar to O3, the nasal passages remove SC>2 more efficiently than the oral pathway
29 (Brain, 1970; Melville, 1970; Nodelman and Ultman, 1999). With exercise, the pattern of SO2
30 absorption shifts from the upper airways to the tracheobronchial airways in conjunction with a
31 shift from nasal to oronasal breathing and increased ventilatory rates. Due to its effect on
September 2007 2-43 DRAFT-DO NOT QUOTE OR CITE
-------
1 delivery and uptake, mode of breathing is also recognized as an important determinant of the
2 severity of SCVinduced bronchoconstriction, with the greatest responses occurring during oral
3 breathing followed by oronasal breathing and the smallest responses observed during nasal
4 breathing.
5 Melville (1970) measured the absorption of SO2 (1.5 to 3.4 ppm) during nasal and oral
6 breathing in 12 healthy volunteers. Total respiratory tract absorption of SC>2 was significantly
7 greater (p < 0.01) during nasal than oral breathing (85 versus 70%, respectively) and was
8 independent of the inspired concentration. Respired flows were not reported. Andersen et al.
9 (1974) measured the nasal absorption of 862 (25 ppm) in 7 volunteers at an average inspired
10 flow of 23 L/min (i.e., eucapnic hyperpnea [presumably] to simulate light exertion). These
11 investigators reported that the oropharyngeal SO2 concentration was below their limit of
12 detection (0.25 ppm), implying that at least 99% of SC>2 was absorbed in the nose of subjects
13 during inspiration. Speizer and Frank (1966) also measured the absorption of SC>2 (16.1 ppm) in
14 7 healthy subjects at an average ventilation of 8.5 L/min (i.e., at rest). They reported that 14% of
15 the inhaled SC>2 was absorbed within the first 2 cm into nose. The concentration of SC>2 reaching
16 the pharynx was below the limit of detection, suggesting that at least 99% was absorbed during
17 inspiration. On expiration, 12% of the 862 absorbed during inspiration was desorbed into the
18 expired air. During the first 15 min after the 25- to 30-min 862 exposure, another 3% was
19 desorbed. In total, 15% of the amount originally inspired and absorbed 862 was desorbed from
20 the nasal mucosa.
21 Frank et al. (1969) and Brain (1970) investigated the oral and nasal absorption of SC>2 in
22 the surgically isolated upper respiratory tract of anesthetized dogs. Radiolabeled SC>2 (35SC>2) at
23 the concentrations of 1, 10, and 50 ppm was passed separately through the nose and mouth at the
24 steady flows of 3.5 and 35 L/min for 5 min. The nasal absorption of SC>2 (1 ppm) was 99.9% at
25 3.5 L/min and 96.8% at 35 L/min. The oral absorption of SO2 (1 ppm) was 99.56% at 3.5 L/min,
26 but only 34% at 35 L/min. The nasal absorption of SC>2 at 3.5 L/min increased with
27 concentration at 1, 10, and 50 ppm and was reported to be 99.9, 99.99, and 99.999%,
28 respectively. This increase in absorption with concentration was hypothesized to be due to
29 increased mucous secretion and increased nasal resistance at the higher SC>2 concentrations. The
30 increased mucus was thought to provide a larger reservoir for SC>2 uptake. The increased nasal
31 resistance may increase turbulence in the airflow and, thereby, decrease the boundary layer
September 2007 2-44 DRAFT-DO NOT QUOTE OR CITE
-------
1 between the gas and liquid phases. Dissimilar to the nose, 862 absorption in the mouth
2 decreased from 99.56 to 96.3% when the concentration was increased from 1 to 10 ppm at
3 3.5 L/min. Frank et al. (1969) reported that up to 18% of the SO2 was desorbed within -10 min
4 after exposure. The authors noted that the aperture of the mouth may vary considerably, and that
5 this variation may affect SC>2 uptake in the mouth. Although SC>2 absorption was dependent on
6 inhaled concentration, the rate and route of flow had a greater effect on the magnitude of SO2
7 absorption in the upper airways.
8 Strandberg (1964) studied the uptake of 862 in the respiratory tract of rabbits. A tracheal
9 cannula with two outlets was utilized to allow sampling of inspired and expired air, and 862
10 absorption was observed to depend on inhaled concentration. The absorption during maximal
11 inspiration was 95% at high concentrations (100 to 700 ppm), reflecting an increased SO2
12 removal in the extrathoracic (ET) airways, whereas it was only 40% at low concentrations (0.05
13 to 0.1 ppm). On expiration, the total SC>2 absorbed (i.e., inspiratory removal in the ET airways
14 plus removal in the lower airways) was 98% at high concentrations and only 80% at the lower
15 concentrations.
16 Amdur (1966) examined changes in airways resistance in guinea pigs due to 862
17 exposure. Guinea pigs were exposed for 1 h to 0.1- to 800-ppm 862 during natural
18 unencumbered breathing or to 0.4 to 100 ppm while breathing through a tracheal cannula.
19 At concentrations of 0.4- to 0.5-ppm 862, route of administration did not affect the airway
20 resistance response, whereas at concentrations of >2 ppm, the responses were greater in animals
21 exposed by tracheal cannula. Based on the concentration-dependent absorption of SC>2 in the ET
22 airways observed by Strandberg (1964), Amdur (1966) concluded that the airway resistance
23 responses at low-exposure concentrations were independent of method of administration,
24 because the lung received nearly the same concentration with or without the cannula as
25 evidenced by minimal ET absorption.
26 More recently, Ben-Jebria et al. (1990) investigated the absorption of 862 in excised
27 porcine tracheae. Absorption was monitored over a 30-min period following the introduction of
28 SO2 (0.1 to 0.6 ppm, inlet concentration) at a constant flow (2.7 to 11 L/min). The data were
29 analyzed using diffusion-reactor theory. An overall mass transfer coefficient (Kso2) was
30 determined and separated into its contributions due to gas (convection and diffusion) and tissue
31 phase (diffusivity, solubility, and reaction rates) resistances. SC>2 in the liquid phase was
September 2007 2-45 DRAFT-DO NOT QUOTE OR CITE
-------
1 assumed to form HSOs rapidly, in proportion with the gas phase 862 concentration, HSOs then
2 diffused down the concentration gradient into the tissues where it reacted irreversibly with
3 biochemical substrates. Initially, KSo2 was limited only by gas phase resistance, but decreased
4 exponentially over the first 5 to 10 min of SC>2 exposure to a smaller steady-state value because
5 of tissue resistance to SC>2 absorption. The initial and steady-state Kso2 values were found to be
6 independent of inlet SC>2 concentration, i.e., for a given flow, the fractional absorption of SC>2 did
7 not depend on SC>2 concentration. An increased KSo2 (initial and steady-state) was observed with
8 an increasing flow that was thought to be due to a decrease in the boundary layer near the walls
9 of the trachea for radial 862 transport. This is in agreement with Aharonson et al. (1974), who
10 also reported that the transfer rate coefficient for SC>2 increases with increasing flow. However,
11 the initial molar flux of SO2 across the gas-tissue interface appears to increase purely as a
12 function of the increase in mass transport occurring with increasing flow (see Figure 5 in Ben-
13 Jebria, 1990). Given that the steady-state Kso2 remained stable during the 10 to 30 min of
14 exposure and that no SC>2 leakage through the tissue was identified, the authors concluded that
15 there was an irreversible sink for SC>2 within the tissue.
16 In summary, inhaled 862 is readily absorbed in the upper airways. During nasal
17 breathing, the majority of available data suggests 95% or greater 862 absorption occurs in the
18 nasal passages, even under ventilation levels comparable to exercise. One study, however,
19 reported only 85% nasal absorption of 862 in humans. Somewhat less 862 is absorbed in the
20 oral passage than in the nasal passages. The difference in SO2 absorption between the mouth and
21 the nose is highly dependent on respired flow rates. In one study, for example, with an increase
22 in flow from 3.5 to 35 L/min, nasal absorption was reduced from 100 to 97%; whereas, oral
23 absorption was reduced from 100 to 34%. Several in vivo studies have reported greater
24 respiratory tract absorption of SC>2 at high versus low SC>2 concentrations. However, the ex vivo
25 uptake of 862 is not related to 862 concentration. It has been postulated that increased mucous
26 secretion and/or increased nasal resistance at high 862 concentrations may account for the
27 increased absorption efficiency observed in vivo. Although 862 absorption may depend on
28 inhaled SO2 concentration, the rate and route of breathing have a greater effect on the magnitude
29 of SC>2 absorption in the upper airways. In exercising humans, the pattern of SC>2 absorption
30 should be expected to shift from the upper airways to the tracheobronchial airways in
September 2007 2-46 DRAFT-DO NOT QUOTE OR CITE
-------
1 conjunction with a shift from nasal to oronasal breathing and associated increased ventilatory
2 rates.
September 2007 2-47 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 2.4-1. REGIONAL DISTRIBUTION OF SO2 AND SO42 AMBIENT
CONCENTRATIONS, AVERAGED FOR 2003-2005
Concentration
Region SO2 (ppb) SO42' (jig m"3)
Mid-Atlantic 3.3 4.5
Midwest 2.3 3.8
Northeast 1.2 2.5
Southeast 1.3 4.1
September 2007 2-48 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 2.4-2. DISTRIBUTIONS OF TEMPORAL AVERAGING INSIDE AND OUTSIDE CMSAS
If
3
^ '
CD
^
to
O
o
to
VO
o
^
^rj
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Percentiles
Averaging Time
Monitor Locations n Mean 1 5 10 25 30 50 70 75 90 95
1-h Maximum Concentration
Inside CMSAs 332405 13 1 1 1 3 4 7 13 16 30 45
Outside CMSAs 53417 13 1 1 1 1 2 5 10 13 31 51
1-h Average Concentration
Inside CMSAs 7408145 41111 1 2 4 5 10 15
Outside CMSAs 1197179 4 1 1 1 1 1 2 3 3 7 13
24-h Average Concentration
Inside CMSAs 327918 41111 2 3 5 6 10 13
Outside CMSAs 52871 4 1 1 1 1 1 2 3 4 8 12
Annual Average Concentration
Inside CMSAs 898 4 1 1 1 1 2 4 5 6 8 10
Outside CMSAs 143 4 1 1 1 1 2 3 4 5 8 9
Aggregate 3-yr Average
Concentration, 2003-2005
Inside CMSAs 283 4 1 1 1 2 3 3 5 5 8 10
Outside CMSAs 42 41112234589
* Values are ppb
** CMSA = Consolidated Metropolitan Statistical Area
99 Max
92 714
116 636
34 714
36 636
23 148
25 123
12 15
13 14
12 14
13 13
-------
TABLE 2.4-3. RANGE OF MEAN SO2 CONCENTRATIONS AND PEARSON
CORRELATION COEFFICIENTS IN URBAN AREAS HAVING AT LEAST
FOUR MONITORS
Metropolitan Area Mean SO2 Concentration
(Number of Monitors) (ppb) Pearson Correlation Coefficient
Philadelphia, PA (10) 3.6-5.9 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 0.01-0.25
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 -0.16-0.31
September 2007 2-50 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 2.5-1. RELATIONSHIPS OF INDOOR TO OUTDOOR
SO2 CONCENTRATIONS
Reference
Spengleretal. (1979)
Stock etal. (1985)
Meranger and Brule
(1987)
Brauer etal. (1989)
Li and Harrison (1990)
Brauer etal. (1991)
Chan etal. (1994)
Lee etal. (1999)
Patterson and Eatough
(2000)
Kindzierski and Sembaluk
(2001)
Chao (2001)
Kindzierski and
Ranganathan (2006)
Location
Portage, WI
Topeka, KS
Kingston, TN
Watertown, MA
St. Louis. MO
Steubenville, OH
Houston, TX
Antigonish, NS,
Canada
Boston, MA
Essex, UK
Boston, MA
Taipei, Taiwan
Hong Kong
Lindon, UT
Boyle, Alberta,
Canada
Sherwood Park,
Alberta, Canada
Hong Kong
Fort McKay, Alberta,
Canada
Indoor to Outdoor Ratio
(number of samples)
0.67 (349)
0.50 (389)
0.08 (425)
0.33 (486)
0.31(543)
0.39 (499)
0.54 (2425)
0.84 (8)
0.23 (24)
0.22
0.39 (geom. mean) (29),
R2 = 0.89
0.05 (geom. mean) (23),
R2 = 0.73
0.24 (15)
0.23 (37)
0.92, R2 = 0.56
0.027 ± 0.0023, R2 = 0.73
0.12(12)
0.14(13)
1.01 ±0.78 (10)
0.35 (30)
Notes
One year during
Harvard Six Cities
study. West-Gaeke
method.
May to October,
continuous FRM for
indoor and outdoor.
Early spring, 1 wk avg
in 1 house with oil
furnace, FPD-TA
Summer, HEADS
Summer
Summer, HEADS
Winter, HEADS
Summer, PS
Winter, PS
Winter, PF
Winter, ADS, all
samples
Late Fall, PS
Late Fall, PS
Summer. Windows
mainly kept closed, PS
Fall. All indoor levels
-------
i 3. INTEGRATED HEALTH EFFECTS OF EXPOSURE TO
2 SULFUR DIOXIDE
o
4
5 This integrated discussion is structured to provide a coherent framework for the
6 assessment of health risks associated with human exposure to ambient sulfur dioxide (802) in the
7 United States. The main goal of this chapter is to integrate newly available epidemiological,
8 human clinical, and animal toxicological evidence with consideration of key findings and
9 conclusions from the 1982 Air Quality Criteria Document (AQCD) for Sulfur Oxides (U.S.
10 Environmental Protection Agency, 1982), 1986 Second Addendum (U.S. Environmental
11 Protection Agency, 1986b), and 1994 Supplement to the Second Addendum, (U.S.
12 Environmental Protection Agency, 1994a), so as to address issues central to the U.S.
13 Environmental Protection Agency (EPA)'s assessment of evidence needed to support the current
14 review of the primary 862 National Ambient Air Quality Standards (NAAQS).
15 This chapter is organized to present morbidity and mortality associated with short-term
16 exposures to SC>2, followed by morbidity and mortality associated with long-term exposures.
17 These sections describe the findings of epidemiological studies that have examined the
18 association between short-term (generally 24-h average) and long-term (generally months to
19 years) ambient 862 exposure and heath outcomes such as increases in respiratory symptoms in
20 asthmatics; increases in emergency department (ED) visits and hospital admissions for
21 respiratory and cardiovascular diseases (CVDs); and increased risk of premature mortality.
22 Human clinical studies examining the effect of peak (1 h or less, generally 5-15 min) exposures
23 of SC>2 on respiratory symptoms and lung function are also discussed in this chapter. These
24 outcomes are presented with relevant animal toxicological data to assess coherence, biological
25 plausibility, and potential mechanistic evidence.
26 The epidemiological studies constitute important information on associations between
27 health effects and exposures of human populations to ambient levels of 862 and also help to
28 identify susceptible subgroups and associated risk factors. However, associations observed
29 between specific air pollutants and health outcomes in epidemiological studies may be
30 confounded by copollutants and/or meteorological conditions and influenced by model
31 specifications in the analytical methods. Extensive discussion of issues related to confounding
32 effects among air pollutants in epidemiological studies are provided in the 2004 AQCD for
September 2007 3-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 Particulate Matter (PM) and therefore not reported here. Briefly, the use of multipollutant
2 regression models has been the prevailing approach for controlling potential confounding by
3 copollutants in air pollution health effects studies. A specific concern is that a given pollutant
4 may act as a surrogate for other unmeasured or poorly measured pollutants. In the event that one
5 or more pollutants act as surrogates for an unmeasured component of a mixture actually
6 responsible for the observed association, the strongest predictor in a multipollutant model could
7 simply indicate which measured pollutant is the best surrogate for the unmeasured pollutant of
8 interest. Particularly in the case of 862, atmospheric chemistry links 862 to SCVderived fine
9 sulfate (SO42 ) particles. Since SO2 and SO42 particles coexist in most ambient situations,
10 observational epidemiologic studies have little ability to distinguish between the adverse health
11 effects of pure gaseous SC>2 with SO42 or other parti culate matter (PM) indices. Attempts to
12 distinguish the gaseous and particle effects related to SC>2 using multipollutant epidemiologic
13 models must be interpreted with caution. Despite the limitations, the use of multipollutant
14 models is still the prevailing approach employed in most studies of SC>2 health effects and serves
15 as an important tool in addressing the issue of confounding by copollutants.
16 Model specification and model selection also are factors to consider in the interpretation
17 of the epidemiological evidence. Epidemiological studies investigated the association between
18 various measures of 862 (e.g., multiple lags, different exposure metrics) and various health
19 outcomes using different model specifications (for further discussion, see 2006 AQCD for Ozone
20 [Os] and Related Photochemical Oxidants). The summary of health effects in this chapter is
21 vulnerable to the errors of publication bias and multiple testing. Efforts have been made to
22 reduce the impact of multiple testing errors on the conclusions in this document. For example,
23 although many studies examined multiple single-day lag models, priority was given to effects
24 observed at 0- or 1-day lags rather than at longer lags. Both single- and multiple-pollutant
25 models that include 862 were considered and examined for robustness of results. Analyses of
26 multiple model specifications for adjustment of temporal or meteorological trends will be
27 considered sensitivity analyses.
28 In addition to evaluating available evidence from epidemiologic studies, this chapter also
29 examined human clinical studies. Human clinical studies conducted in controlled exposure
30 chambers use fixed concentrations of air pollutants under carefully regulated environmental
31 conditions and subject activity levels to minimize possible confounding of the health associations
September 2007 3-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 by other factors. While human clinical studies do in fact provide a direct quantitative assessment
2 of the SO2 exposure-health response relationship, such studies have a number of limitations.
3 First, study subjects must be either healthy individuals or individuals whose level of illness does
4 not preclude them from participating in the study. Subjects with a recent history of upper
5 respiratory tract infections are typically excluded from clinical studies of exposure to SC>2, as are
6 asthmatics who are unable to withhold the use of brochodilators for at least 6 hours prior to
7 exposure. Therefore, the results of human clinical studies may underestimate the health effects
8 of exposure to certain sensitive subpopulations. In addition, studies of controlled exposure to
9 SO2 have typically used peak concentrations for shorter durations (5-15 min). While these
10 studies provide important information on the biological plausibility of associations observed
11 between SO2 exposure and health outcomes in epidemiological studies, the concentration-
12 response relationships cannot be directly extrapolated to concentrations below those
13 administered in the laboratory. Finally, human clinical studies are normally conducted on a
14 relatively small number of subjects, which reduces the power of the study to detect significant
15 differences in the health outcomes of interest between exposure to varying concentrations of SC>2
16 and clean air.
17 The chapter discussion focuses on the important new scientific studies, with emphasis on
18 those conducted at or near current ambient concentrations. The attached annexes include a broad
19 survey of the epidemiology and toxicology literature to supplement the information presented
20 here.
21
22
23 3.1 MORBIDITY ASSOCIATED WITH SHORT-TERM SO2
24 EXPOSURE
25
26 3.1.1 Respiratory Effects Associated with Short-Term Exposure to SOi
27 In the 1982 AQCD for Sulfur Oxides, only a few epidemiological studies were useful in
28 determining the concentration-response relationship of respiratory health effects from short-term
29 exposure to SC>2. The most notable study was by Lawther et al. (1970), which examined the
30 association between air pollution and worsening health status in bronchitic patients. It was
31 concluded in the 1982 AQCD that worsening of health status among chronic bronchitic patients
32 was associated with daily black smoke (BS) levels of 250 to 500 |ig/m3 in the presence of 862
September 2007 3-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 levels in the range of 500 to 600 |ig/m3 (191 to 229 ppb). In the 1986 Second Addendum,
2 additional studies investigated morbidity associated with short-term exposure to 862. The most
3 relevant study was by Dockery et al. (1982), which examined pulmonary function in school
4 children in Steubenville, OH as part of the Harvard Six Cities Study. This study found that small
5 but statistically significant reversible decrements in forced vital capacity (FVC) and forced
6 expiratory volume in 0.75 s (FEVo.ys) were associated with increases in 24-h average
7 concentrations of total suspended particles (TSP) at levels ranging up to 220 to 420 |ig/m3 and
8 SO2 at levels ranging up to 280 to 460 |ig/m3 (107 to 176 ppb). However, it was impossible to
9 separate the relative contributions of TSP and 862, and no threshold level for the observed
10 effects could be discerned from the wide range of exposure levels.
11 Epidemiological evidence for an association between SO2 and morbidity as indicated by
12 increased use of ED facilities or increased hospital admissions for respiratory disease outcomes
13 were also reported in the 1982 AQCD. Overall, these results suggested increased upper
14 respiratory tract morbidity, especially among older adults, during episodic marked elevations of
15 PM or SO2 (0.4 to 0.5 ppm). The 1982 AQCD further concluded that the reviewed studies
16 provided essentially no evidence for an association between asthma attacks and acute exposures
17 at typical ambient 24-h average PM or 862 levels in the United States.
18 The majority of the 862 human clinical studies reviewed in the 1982 AQCD evaluated
19 respiratory effects of 862 exposure in healthy adults, with some limited data from clinical studies
20 of adults with asthma. Respiratory effects from SO2 exposure such as increased airway
21 resistance and decreased forced expiratory volume in 1 s (FEVi) were well documented. The
22 1986 Second Addendum and 1994 Supplement to the Second Addendum reviewed several
23 additional controlled studies involving both healthy and asthmatic individuals. In general, these
24 studies found no pulmonary effects of SC>2 exposure in healthy subjects exposed to
25 concentrations of <1.0 ppm (Bedi et al., 1984; Folinsbee et al., 1985; Kulle et al., 1984; Stacy
26 et al., 1983). However, in exposures of asthmatic adults, respiratory effects have been observed
27 following short-term exposures (<5 min) to levels of <1.0 ppm (Balmes et al., 1987; Horstman
28 et al., 1988). Decreases in lung function have consistently been demonstrated in relatively
29 healthy, exercising asthmatic adults following 5-15 minute exposures to 0.5-1.0 ppm SC>2.
30 The 1982 AQCD also noted that numerous effects on the respiratory system were
31 observed in animals exposed to SC>2. Effects were generally observed at levels exceeding
September 2007 3-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 ambient exposure levels and included morphological changes, altered pulmonary function, lipid
2 peroxidation, and changes in host lung defenses. The immediate effect of acute 862 exposure in
3 animals was observed to be increased pulmonary resistance to airflow, a measure of
4 bronchoconstriction. It was postulated that increased pulmonary resistance is mediated through
5 bronchial epithelial receptors that activate an autonomic reflex arc through the vagus nerve, a
6 process that is also believed to occur in humans. Because the reflex was blocked by atropine, it
7 was determined to be cholinergic. SO2-induced bronchoconstriction was hypothesized to involve
8 smooth muscle contraction, because it was reversed by p-adrenergic agonists such as
9 isoproterenol. Acetylcholine and histamine were also thought to be involved in SCVinduced
10 bronchoconstriction. The 1982 AQCD reported some effects of SC>2 on lung defenses that
11 usually occurred at concentrations exceeding ambient exposure concentrations. Alterations in
12 the antiviral defense system and pulmonary immune system and slowed mucociliary clearance
13 were reported in mice exposed to 2- to 10-ppm SC>2.
14 Collectively, the epidemiological, human clinical, and animal toxicological studies
15 provided biological plausibility and coherent evidence of an adverse effect of ambient SC>2 on
16 respiratory health. Since the 1982 AQCD, 1986 Second Addendum, and 1994 Supplement to the
17 Second Addendum, additional studies have been conducted on the relationship between short-
18 term exposures to ambient 862 and adverse respiratory health effects, including respiratory
19 symptoms, lung function, airways inflammation, airways hyperresponsiveness, lung host
20 defenses, and ED visits and hospitalizations for respiratory causes. The epidemiological, human
21 clinical, and animal toxicological evidence on the effects of SC>2 on these various endpoints are
22 discussed below.
23
24 3.1.1.1 Respiratory Symptoms
25 Respiratory symptoms in air pollution field studies are usually measured using
26 questionnaire forms (or "daily diaries") that are filled out by study subjects. Questions address
27 the daily experience of coughing, wheezing, shortness of breath (or difficulty breathing),
28 production of phlegm, and others. In this section, the effects of short-term exposure to 862 on
29 respiratory symptoms in children and adults will be discussed separately. Epidemiological
30 studies on respiratory symptoms published since the last review are summarized in Annex Table
31 AX5-1 with key studies discussed in further detail below.
September 2007 3-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 Children
2 The strongest epidemiological evidence for an association between respiratory symptoms
3 and exposure to ambient SO2 comes from two large U.S. multicity studies (Mortimer et al., 2002;
4 Schildcrout et al., 2006). Mortimer et al. (2002) examined 846 asthmatic children from eight
5 U.S. urban areas in the National Cooperative Inner-City Asthma Study (NCICAS) for
6 summertime air pollution-related respiratory symptoms. Median 3-h average SC>2 (8 to 11 a.m.)
7 levels ranged from 17 ppb in Detroit to 37 ppb in East Harlem. Morning symptoms were found
8 to be most strongly associated with an average of a 1- to 2-day lag of 862 concentrations. In
9 two-pollutant models with 63 and nitrogen dioxide (NO2) (measured in seven cities), the 862
10 association remained robust. When particulate matter with an aerodynamic diameter of < 10|i
11 (PMio) was also included in the multipollutant models using data from three cities, the effect
12 estimate remained similar, but became nonsignificant likely due to reduced statistical power.
13 In the Childhood Asthma Management Program (CAMP) study, the association between
14 ambient air pollution and asthma exacerbations in children (n = 990) from eight North American
15 cities was investigated (Schildcrout et al., 2006). 862 measurements were available in seven of
16 the eight cities. The median 24-h average 862 concentrations ranged from 2.2 ppb (interquartile
17 range [IQR]: 1.7, 3.1) in San Diego to 7.4 ppb (IQR: 5.3, 10.7) in St. Louis. Results for the
18 associations between asthma symptoms and all pollutants are shown in Figure 3.1-1. Analyses
19 indicated that, although SC>2 lags were positively related to increased risk of asthma symptoms,
20 only the 3-day moving average was statistically significant. Stronger associations were observed
21 for carbon monoxide (CO) and NC>2. In two-pollutant models with CO, NO2, and PMio, the
22 effect estimate and 95% confidence interval (CI) remained consistent (Figure 3.1-1).
23 A longitudinal study of 1,844 schoolchildren during the summer from the Harvard Six
24 Cities Study suggested that the association between SO2 and respiratory symptoms could be
25 confounded by PMio (Schwartz et al., 1994). The median 24-h average SO2 concentration
26 during this period was 4.1 ppb (10th-90th percentile: 0.8, 17.9; maximum 81.9). SO2
27 concentrations were found to be associated with cough incidence and lower respiratory
28 symptoms. Of the pollutants examined, PMio had the strongest associations with respiratory
29 symptoms. In two-pollutant models, the effect of PMio was found to be robust to adjustment for
30 other copollutants, while the effect of SO2 was substantially reduced after adjustment for
September 2007 3-6 DRAFT-DO NOT QUOTE OR CITE
-------
Figure 3.1-1.
S02
1 an ft n QQ-
I in 1 fl 0^
I an 9 0 QQ.
3-day moving sum 1.00
S02 and CO
1 an H ft QQ-
I an 1 fl OR
Lag 2 1.
3-day moving sum 1 .0(
1.06
1.05
1.06
_l_04_1
1.07
1.06
L 1.
105
SOa and N02
LagO 093 1-06
I in 1 0 0^
Lag 2 1.(
3-day moving sum 1.0C
Lag 0 0 98
I -»n 1 f> OR
I an 7 n QR
3-day moving sum 0.99
1.02
11^
,1 -t c
,.,.,1 1">
08
1 10
10 „
1,09
1 15
4 44
! 1.09
1.04
105
1.04
1.05
104
— . — 1
1.09
1 13
1 id
1 1"1
.08
0.80 0.90 1.00 1.10 1.00
Odds Ratio
Odds ratios for daily asthma symptoms associated with a 10-ppb
increase in within-subject concentrations of 24-h average SOi, using
data collected from November 1993 to September 1995. All city-
specific estimates of pollutant effects were included in calculations of
study-wide effects except SOi in Albuquerque, NM and NOi in
Seattle, WA.
Source: Schildcrout et al. (2006).
2-
1 As the PMio concentrations were correlated strongly to SO2-derived SC>4 particles (r = 0.80),
2 the diminution of the SC>2 effect estimate may indicate that for PMio dominated by fine SC>42
3 particles, PMio has a slightly stronger association than SC>2. This study further investigated the
4 concentration-response function and observed a nonlinear relationship between SC>2
5 concentrations and respiratory symptoms. A figure plotting the relative odds of incidence of
6 lower respiratory symptoms against SC>2 concentrations lagged 1 day indicated that no
7 statistically significant increase in the incidence of lower respiratory symptoms was seen until
September 2007
3-7
DRAFT-DO NOT QUOTE OR CITE
-------
1.6 -
1.4 -
re
G£
w
TJ
§ 1-2
1.0 -
10 20 30
24-h avg S02 (ppb)
40
Figure 3.1-2.
Relative odds ratio of incidence of lower respiratory symptoms
smoothed against 24-h average SOi concentrations on the previous
day, controlling for temperature, city, and day of week.
Source: Schwartz etal. (1994).
1 concentrations exceeded a 24-h average SO2 of 22 ppb though an increasing trend was observed
2 at concentrations as low as 10 ppb (Figure 3.1-2).
3 In the Pollution Effects on Asthmatic Children in Europe (PEACE) study, a multicenter
4 study of 14 cities across Europe, the effects of acute exposure to various pollutants including 862
5 on the respiratory health of children with chronic respiratory symptoms (n = 2,010) was
6 examined during the winter of 1993-1994 (Roemer et al., 1998). Mean 24-h average SC>2
7 concentrations ranged from 2 |ig/m3 (1 ppb) in the urban area of Umea, Sweden, to 113.9 |ig/m3
8 (43 ppb) in the urban area of Prague, Czech Republic. No associations were observed between
9 SC>2 and daily prevalence of respiratory symptoms or bronchodilator use at any of the single- and
September 2007
3-8
DRAFT-DO NOT QUOTE OR CITE
-------
1 multiday lags considered. In addition, no associations were observed for any of the other
2 pollutants examined. It should be noted that during the study period, there were only two major
3 air pollution episodes, one at the beginning and one at the end of the study period. In the
4 epidemiologic model, the control for time trend was accomplished through the use of linear and
5 quadratic terms. Given the timing of the air pollution episodes, the quadratic trend term would
6 have removed most of the air pollution effect. Other studies that participated in the PEACE
7 study and analyzed results for longer periods of times have observed statistically significant
8 associations between 862 and respiratory symptoms in children (for example, see van der Zee
9 et al., 1999, presented below).
10 Other studies have examined the relationship between respiratory symptoms and ambient
11 SO2 concentrations. These studies generally indicated positive associations, including two U.S.
12 studies (Delfino et al., 2003; Neas et al., 1995) and several European studies (Hoek and
13 Brunekreef, 1994; Peters et al., 1996; Roemer et al., 1993; Segala et al., 1998; Timonen and
14 Pekkanen, 1997; van der Zee et al., 1999). However, some studies found no consistent
15 association (e.g., Hoek and Brunekreef, 1993, 1995; Romieu et al., 1996) between respiratory
16 symptoms and 862 concentrations. Given the high correlations among the air pollutants,
17 particularly with PM indices or sulfate (SO42 ), it is possible that SO2 might be an indicator for
18 particulate air pollution characterized by PMio or SC>42 or it might also be a surrogate for other
19 unmeasured combustion products. Only one of these studies examined possible confounding of
20 the SO2 effect by copollutants. Van der Zee et al. (1999) studied the association between
21 respiratory symptoms and SO2 in 7- to 11-year-old children (n = 633) with and without chronic
22 respiratory symptoms in the Netherlands. Significant associations with lower respiratory
23 symptoms and increased bronchodilator use were observed for SO2, as well as PMio, BS, and
24 SO42 , in symptomatic children living in urban areas (n = 142). In a two-pollutant model with
25 PMio, the results were robust for bronchodilator use, but slightly reduced for lower respiratory
26 symptoms.
27 Figures 3.1-3 and 3.1-4 present the odds ratios for SO2-related cough and lower
28 respiratory or asthma symptoms, respectively, from several epidemiological studies with relevant
29 data. The results for cough are somewhat variable with wide confidence intervals, as shown in
30 Figure 3.1-3. The studies conducted in the summer generally indicate increased risk of cough
31 from exposure to SO2. A more consistent effect of SO2 is observed on lower respiratory or
September 2007 3-9 DRAFT-DO NOT QUOTE OR CITE
-------
Reference Location Population
Schwartz el al. (1994) 6 U.S. cities Children (n = 300)
Neas et al. (1995) Uniontown, PA Non-asthmatics [n = 98)
Ward et al, (2002a,b) Birmingham and Children (n=162)
Sandwell, U.K.
'Van der Zee et al. (1999) 5 urban areas With chronic respiratory
in the Netherlands symptoms (n=142)
'Roemer et al. (1993) The Netherlands With chronic respiratory
symptoms (n = 73)
Hoek and Brunekreef (1994) The Netherlands Children (n = 1079)
Ward et al. (2002a,b) Birmingham and Children (n=162)
Sandwell, U.K.
Segalaetal.(1998)
Romieu et al. (1996)
Paris. France
N. Mexico City,
Mexico
Mild asthmatics (n =43)
Mild asthmatics (n = 71) Not stated -
Ujg
0 -
1-4 -
0 -
0 -
1 .
f\ C _
0 -
1 -
M -
0 -
1 -
0 -
4 ,
1 _
ated -
Summer
— *
Winter — t-
All year "+
I I
—
— •
1
•
«
•
1 1 1 1 1 1
0.25
0.5 0.75 1.0 1.25 1.5 1.752.02.252.5
Odds Ratio
Figure 3.1-3.
Odds ratios (95% CI) for the incidence of cough among children,
grouped by season. For single-day lag models, current day and/or
previous day SOi effects are shown, except for Segala et al. (1998),
which only presented results for a 3-day lag. Risk estimates are
standardized per 10-ppb increase in 24-h average SOi 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. (1993) presented results for prevalence of cough.
1 asthma symptoms (Figure 3.1-4). Although there is some variability in the individual effect
2 estimates, the majority of the odds ratios appear to be >1. Similar to cough, stronger associations
3 with lower respiratory or asthma symptoms were observed in the summer compared to the
4 winter. There was some variability among the different lags of exposure; however, effects were
September 2007
3-10
DRAFT-DO NOT QUOTE OR CITE
-------
Reference Location
Mortimer etal. (2002) 8 U.S. cities
Schwartz etal. (1994) 6 U.S. cities
Asthmatics (n = 846)
Children (n = 300)
*Van der Zee et al. (1999) 5 urban areas With chronic respiratory
in the Netherlands symptoms (n = 142)
*Roemer et a I. (1993) TheNetherlands With chronic respiratory
symptoms (n = 73)
Hoek and Brunekreef (1994) The Netherlands Children (n = 1079)
Segala et al. (1998) Paris, France Mild asthmatics (n =43)
•Schildcrout et al. (2006) 7 U.S. Cities Asthmatics (n = 881)
Romieu etal.(1996)
N. Mexico City.
Mexico
Mild asthmatics (n = 71) Not stated -
Lag
1-2 -
1 _
0 -
1 -
0-4 -
0 -
1 -
0-6 -
0 -
1 -
1 -
0 -
1 -
0-2 -
ated -
-t
I
0.8 1
Summer
~*~
i Winter
-i
6
—> — All year
— i
-B-
I I I I I I I I
0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.f
Figure 3.1-4.
Odds Ratio
Odds ratios (95% CI) for the incidence of lower respiratory or asthma
symptoms among children, grouped by season. For single-day lag
models, current day and/or previous day SOi effects are shown. Risk
estimates are standardized per 10-ppb increase in 24-h average SOi
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), Roemer et al. (1993), and Schildcrout et al. (2006) presented results for
prevalence of symptoms.
1 generally observed with current day or previous day exposure and, in some cases, with a
2 distributed lag of 2 to 3 days.
3 The 1982 AQCD concluded that there was insufficient evidence on the effect of 862 and
4 PM on asthma attacks but that exposure to these pollutants was associated with increases in the
5 occurrence of upper respiratory symptoms, including exacerbation of preexisting chronic
6 bronchitis. A study by Keles et al. (1999) evaluated the prevalence of chronic rhinitis among
7 high school students before and after installation of a natural gas network for domestic heating
September 2007
3-11
DRAFT-DO NOT QUOTE OR CITE
-------
1 and industrial works in a polluted area of Istanbul, Turkey. Concentrations of CO, NC>2, and
2 hydrocarbons were relatively low compared to 862 and TSP in this area. After the intervention,
3 the annual mean TSP concentration declined by 23% from 89.7 |ig/m3 to 68.8 |ig/m3. An even
4 greater decline (46%) was observed for SO2, from an annual mean of 185.4 |ig/m3 (70.8 ppb) to
5 100.0 |ig/m3 (38.2 ppb). The prevalence of rhinitis decreased significantly from 62.5 to 51% of
6 the student population (p < 0.05) following the installation of the natural gas network.
7 Symptoms of rhinitis were associated with air pollution levels but not with any of the other
8 factors considered, including sex, household crowding, heating source, and smoking status.
9 Although the effects from TSP could not be separated from 862, this study demonstrated that
10 reductions in both pollutants (with greater declines in 862) resulted in significant reductions in
11 the prevalence of chronic rhinitis in a highly polluted area.
12 Overall, recent epidemiological studies provide evidence for an association between
13 ambient SO2 exposure and increased respiratory symptoms in children, particularly those with
14 asthma or chronic respiratory symptoms. Recent U.S. multicity studies observed significant
15 associations between SC>2 and respiratory symptoms at a median range of 17 to 37 ppb
16 (75th percentile: -25 to 50) across cities for 3-h average SO2 (NCICAS, Mortimer et al., 2002)
17 and 2.2 to 7.4 ppb (90th percentile: 4.4 to 14.2) for 24-h average SO2 (CAMP, Schildcrout et al.,
18 2006). However, an earlier study that examined the concentration-response function found that a
19 statistically significant increase in the incidence of lower respiratory symptoms was not observed
20 until concentrations exceeded a 24-h average SO2 of 22 ppb, though an increasing trend was
21 observed at concentrations as low as 10 ppb (Harvard Six Cities Study, Schwartz et al., 1994).
22 In the limited number of studies that examined potential confounding by copollutants through
23 multipollutant models, the SC>2 effect was generally found to be robust after adjusting for PM
24 and other copollutants.
25
26 Epidemiological Studies of Adults
27 Compared to the number of studies conducted with children, fewer studies were
28 performed that examined the effect of ambient SC>2 exposure on respiratory symptoms in adults.
29 Most of these studies focused on potentially susceptible populations, i.e., those with asthma or
30 chronic obstructive pulmonary disease (COPD). One of the larger studies was conducted by van
31 der Zee et al. (2000) in 50- to 70-year-old adults, with (n = 266) and without (n = 223) chronic
32 respiratory symptoms in the Netherlands. In adults both with and without chronic respiratory
September 2007 3-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 symptoms, no consistent associations were observed between 862 levels and respiratory
2 symptoms or medication use.
3 Studies by Desqueyroux et al. (2002a,b) examined the association between air pollution
4 and respiratory symptoms in other potentially susceptible populations, i.e., those with severe
5 asthma (n = 60, mean age 55 years) and COPD (n = 39, mean age 67 years), in Paris, France.
6 The mean 24-h average SO2 concentration was 7 |ig/m3 (3 ppb, range: 1, 10) in the summer and
7 19 |ig/m3 (7 ppb, range: 1, 31) in the winter. No associations were observed between SC>2
8 concentrations and the incidence of asthma attacks or episodes of symptom exacerbation in the
9 severe asthmatics or individuals with COPD. 63 was found to have the strongest effect in these
10 studies.
11 Several other European studies did observe an association between ambient SO2
12 concentrations and respiratory symptoms in adults with asthma or chronic bronchitis (Higgins
13 et al., 1995; Neukirch et al., 1998; Peters et al., 1996; Taggart et al., 1996). However, only one
14 of these studies examined possible confounding of the association by copollutants. Higgins et al.
15 (1995) examined the effect of summertime air pollutant exposure on respiratory symptoms in 62
16 adults with either asthma, COPD, or both. The maximum 24-h average 862 level was 117 |ig/m3
17 (45 ppb). An association was observed between 862 and symptoms of wheeze, and it remained
18 robust to adjustment for 63 and NC>2. The effects of PM were not examined in this study.
19 Results from the epidemiological studies examining the association between 862 and
20 respiratory symptoms in adults are generally mixed, with some showing positive associations
21 and others finding no relationship at current ambient levels.
22
23 Human Clinical Studies of Adults
24 The 1994 Supplement to the Second Addendum described in detail several studies that
25 evaluated respiratory symptoms following controlled human exposures to SC>2. Briefly,
26 following 1-h exposures to 0-, 0.2-, 0.4-, and 0.6-ppm SC>2, Linn et al. (1987) reported that the
27 severity of respiratory symptoms (i.e., cough, chest tightness, throat irritation) increased relative
28 to air exposures only in moderate/severe asthmatics who were exposed at the highest exposure
29 concentration (0.6 ppm). It was also observed that these symptoms abated within <1 h after
30 exposure. Balmes et al. (1987) reported that 7/8 asthmatic adults developed respiratory
September 2007 3-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 symptoms including wheezing and chest tightness following 3-min exposures to 0.5-ppm
2 during eucapnic hyperpnea (minute ventilation [yE] = 60 L/min).
3 Since the publication of the 1994 Supplement to the Second Addendum, several
4 additional publications have evaluated the effect of SO2 exposure on respiratory symptoms in a
5 laboratory setting. In a human clinical study with SO2-sensitive asthmatics, Gong et al. (1995)
6 reported that respiratory symptoms (i.e., shortness of breath, wheeze, and chest tightness)
7 increased with increasing 862 concentration (0-, 0.5-, and 1.0-ppm 862) following exposures of
8 10 min with varying levels of exercise. It was also observed that exposure to 0.5-ppm 862
9 during light exercise evoked a more severe symptomatic response than heavy exercise in clean
10 air. In a more recent study, Tunnicliffe et al. (2003) found no association between respiratory
11 symptoms (i.e., throat irritation, cough, wheeze) and 1-h exposures at rest to 0.2-ppm SC>2 in
12 either asthmatics or healthy adults.
13 Collectively, evidence from the previous review along with a limited number of new
14 human clinical studies indicate increased respiratory symptoms with peak (5-15 min) SC>2
15 exposures as low as 0.5 ppm in asthmatic subjects.
16
17 3.1.1.2 Lung Function
18 Most of the studies discussed in the previous section for effects of 862 on respiratory
19 symptoms also examined lung function. In studies assessing the relationship between acute
20 exposure to air pollution and lung function, self-administered PEF meters were primarily used.
21 Since PEF follows a circadian rhythm, with the highest values found during the afternoon and
22 lowest values during the night and early morning (Borsboom et al., 1999), these studies generally
23 have analyzed PEF data stratified by time of day. The epidemiological studies on lung function
24 are summarized in Annex Table AX5-1.
25
26 Children
27 Mortimer et al. (2002) examined 846 asthmatic children from eight U.S. urban areas in
28 the NCICAS for changes in PEF related to air pollution. The mean 3-h average 862 was 22 ppb
29 across the eight cities during the study period of June through August 1993. No associations
30 were observed between SO2 concentrations and morning or evening PEF. Of all the pollutants
31 examined, including PMi0, O3, and NO2, only O3 was associated with changes in morning PEF.
September 2007 3-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 In another U.S. study (Neas et al., 1995), 83 children from Uniontown, PA reported
2 twice-daily PEF measurements during the summer of 1990. The mean daytime 12-h average
3 SC>2 concentration was 14.5 ppb (maximum 44.9). No associations were observed between
4 daytime 12-h average SO2 concentrations and mean deviation in evening PEF, even after
5 concentrations were weighted by the proportion of hours spent outdoors during the prior 12 h.
6 Statistically significant associations were observed for 63, total SC>42 particles, and particle-
7 strong acidity.
8 A study by van der Zee et al. (1999) observed associations between ambient 862
9 concentrations and daily PEF measurements in 7- to 11-year-old children (n = 142) with chronic
10 respiratory symptoms living in urban areas of the Netherlands (van der Zee et al., 1999). The
11 odds ratio (OR) for a >10% decrement in evening PEF per 10-ppb increase in 24-h average SO2
12 was 1.20 (95% CI: 0.97, 1.47) with same-day exposure. A greater effect was observed at a
13 2-day lag, OR = 1.40 (95% CI: 1.18, 1.67), and this effect remained robust in a two-pollutant
14 model with PMio, OR = 1.34 (95% CI: 1.08, 1.64).
15 Multipollutant analyses also were conducted in a study by Chen et al. (1999), which
16 examined the effects of short-term exposure to air pollution on the pulmonary function of
17 895 children (age 8 to 13 years) in three communities in Taiwan. The daytime 1-h max SO2 the
18 day before spirometry ranged from 0 to 72.4 ppb. In a single-pollutant model, 1-h max SO2
19 concentration at a 2-day lag was significantly associated with FVC, -50.80 mL (95% CI:
20 -97.06, -4.54), or a 2.6% decline, per 40-ppb 1-h max SO2. However, in multipollutant models,
21 only Os remained significantly associated with FVC and FEVi.
22 While additional studies have observed associations between ambient SO2 concentrations
23 and changes in lung function in children (e.g., Hoek and Brunekreef, 1993; Peters et al., 1996;
24 Roemer et al., 1993; Segala et al., 1998; Timonen and Pekkanen, 1997), several other studies did
25 not find a significant association between SO2 and lung function parameters (e.g., Delfmo et al.,
26 2003; Peacock et al., 2003; Romieu et al., 1996).
27 In a human clinical study of asthmatic adolescents (12 to 16 years old), Koenig et al.
28 (1983) evaluated changes in FEVi following a 10-min exposure during moderate exercise to
29 0.5- and 1.0-ppm SO2 + 1-mg/m3 NaCl. Significant decreases of 15 and 23% were reported in
30 FEVi following exposure to 0.5- and 1.0-ppm SO2, respectively. No significant changes in FEVi
31 were observed between pre- and postexposure to 1-mg/m3 NaCl without SO2.
September 2007 3-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 The mixed results observed in epidemiological studies, along with the high to moderate
2 correlation between 862 levels and other copollutants, most notably PM, reported in most studies
3 generally suggest that short-term exposure to ambient SO2 does not have an independent effect
4 on lung function in children. One human clinical study provided evidence that during exercise,
5 peak exposures (10 min) to SC>2 at concentrations of as low as 0.5 ppm in the presence of
6 hygroscopic particles that can carry SC>2 deeper into the lung can elicit significant changes in
7 pulmonary function in asthmatic adolescents.
8
9 Epidemiological Studies of Adults
10 Van der Zee et al. (2000) observed an association between 862 and morning PEF in
11 50- to 70-year-old adults (n = 138) with chronic respiratory symptoms living in urban areas of
12 the Netherlands. No associations were observed with evening PEF. The OR for a >20%
13 decrement in PEF was 1.21 (95% CI: 0.76, 1.92) per 10-ppb increase in 24-h average SO2 with
14 same-day exposure and 1.56 (95% CI: 1.02, 2.39) at a 1-day lag. No associations were observed
15 for a >10% decrement in PEF. The authors hypothesized that while SO2 level did not have much
16 effect on PEF in most subjects, there was a small subgroup of individuals who experienced fairly
17 large PEF decrements when SO2 levels were high. No multipollutant analyses were conducted.
18 Higgins et al. (1995) examined the association between pulmonary function and air
19 pollution in 75 adults with either asthma, COPD, or both. Exposure to 862 was associated with
20 increased variation in PEF but not with mean or minimum PEF. The SC>2 effects on PEF
21 variation were robust to adjustment for 63 and NC>2. Effects of PM were not considered.
22 Neukirch et al. (1998) also observed associations between lung function and SO2 concentrations
23 in a study of asthmatic adults in Paris, France, but significant associations were found for all
24 pollutants examined, including BS, PMo, and NC>2.
25 In a cross-sectional survey, Xu et al. (1991) investigated the effects of indoor and outdoor
26 air pollutants on the respiratory health of 1,140 adults (aged 40 to 69 years) living in residential,
27 industrial, and suburban areas of Beijing, China. The annual mean concentrations of 862 in
28 residential, industrial, and suburban areas from 1981 to 1985 were 128 |ig/m3 (49 ppb), 57 |ig/m3
29 (22 ppb), and 18 |ig/m3 (7 ppb), respectively. Log-transformed 862 and TSP were significantly
30 associated with reductions in FEVi and FVC. The authors cautioned that since SO2 and TSP
31 concentrations were strongly correlated, the effect of SC>2 could not be separated from that
32 of TSP.
September 2007 3-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 Others observed no relationship between ambient 862 concentrations and lung function
2 in adults (Peters et al., 1996; Taggart et al., 1996). Similar to the results observed for children,
3 the epidemiological studies examining adults do not provide strong evidence for an association
4 between short-term exposure to ambient SO2 and lung function. While some studies did observe
5 significant associations between SO2 exposure and decrements in lung function parameters, the
6 results were not consistent across studies. In addition, the strong correlation between SC>2 and
7 various copollutants in most studies limits interpretation of independent effects of SC>2 on lung
8 function.
9
10 Human Clinical Studies of Adults
11
12 Healthy Individuals
13 In controlled 862 exposures of healthy human subjects under resting conditions,
14 respiratory effects including increased respiration rates, decrements in peak flow,
15 bronchoconstriction, and increased airway resistance have been observed. Most of these studies
16 report effects at concentrations of >5 ppm (Abe, 1967; Andersen et al., 1974; Frank et al., 1962;
17 Lawther, 1955; Sim and Pattle, 1957), with only a few studies reporting significant health effects
18 at concentrations as low as 1 ppm. Snell and Luchsinger (1969) observed a significant decrease
19 in maximum expiratory flow at 50% of forced vital capacity (MEF5oo/0) in healthy resting adult
20 subjects following 15-min inhalation exposures through a mouthpiece to 1-ppm SC>2. Amdur
21 et al. (1953) reported an increase in respiration rate and a decrease in tidal volume at 1-ppm 862;
22 however, this may be considered to be an irritant response rather than an adverse health effect of
23 exposure.
24 The respiratory effects of SO2 can be potentiated by increasing ventilation rate either
25 through eucapnic hyperpnea or by performing light exercise during exposure. This effect is
26 likely due to an increased uptake of SC>2 because of both the increase invE as well as a shift from
27 nasal breathing to oronasal breathing. Lawther et al. (1975) found that deep breathing of 1-ppm
28 SC>2 by mouth resulted in an increase in specific airways resistance (sRaw) compared to
29 breathing air alone. Stacy et al. (1981) exposed 16 healthy males to 0.75-ppm 862 for 2 h with a
30 15-min period of exercise at the end of the first hour of exposure (yE ~ 60 L/min). A separate
31 group of 15 healthy males were exposed to clean air for 2 h and served as the control for this
32 study. In the SC>2-exposed group, airways resistance (Raw) decreased by 2 to 55% compared to
September 2007 3-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 baseline after the 15 min of exercise, but then returned to the baseline value by the end of the 2-h
2 exposure. However, in the control group, Raw decreased throughout the 2-h exposure, resulting
3 in statistically significant differences between the two groups in the change in Raw occurring
4 between both baseline and post-exercise and between baseline and postexposure.
5
6 Asthmatic Individuals
1 During the last review, it was established that subjects with asthma are more sensitive to
8 the effects of SO2 exposure than healthy individuals without asthma. In fact, it has been
9 demonstrated that asthmatic individuals exposed to 2 while performing moderate to
10 heavy exercise for 5 min suffer significant bronchoconstriction or increases in sRaw (Bethel
11 et al., 1983; Linn et al., 1983, 1984). Gong et al. (1995) was able to show an exposure-response
12 relationship between 862 and respiratory effects by exposing 14 unmedicated, SCVsensitive
13 asthmatics to 0-, 0.5-, and 1-ppm SO2 under 3 different levels of exercise. It was shown that
14 increasing SC>2 concentration had a greater effect on sRaw and FEVi than increasing exercise
15 level. Tunnicliffe et al. (2003) evaluated the effect of a lower exposure concentration of SC>2 in
16 resting healthy and asthmatic subjects. No significant changes in lung function as measured by
17 FEVi, FVC, and maximal midexpiratory flow (MMEF) were observed following 1-h exposure to
18 0.2-ppm SC>2. The authors reported a small but statistically significant increase in respiratory
19 rate in the asthmatic group after 862 exposure compared to placebo (958.9 breaths/h with 862
20 compared to 906.8 breaths/h with air). However, this effect was counterbalanced by a reduction
21 in tidal volume, resulting in no net change in volume breathed during exposure.
22 Since some of the studies involving asthmatic subjects have used change in sRaw as the
23 endpoint of interest while others have measured changes in FEVi or both, a comparison of FEVi
24 and sRaw based on data from Linn et al. (1987, 1990) were provided in the 1994 Supplement to
25 the Second Addendum. Based on simple linear interpolation of the data from these two studies
26 (Linn et al., 1987, 1990), a 100% increase in sRaw corresponded to a 12 to 15% decrease in
27 FEVi and a 200% increase in sRaw corresponded to a 25 to 30% decrease in FEVi.
28 One of the aims of the Linn et al. (1987) study was to determine how the intensity of
29 response varied with asthma severity or status. In this study, 24 normal, 21 atopic (but not
30 asthmatic), 16 mild asthmatic, and 24 moderate/severe asthmatic subjects were exposed to
31 0-, 0.2-, 0.4-, and 0.6-ppm SC>2. The exposure protocol consisted of 1-h exposures that included
September 2007 3-18 DRAFT-DO NOT QUOTE OR CITE
-------
1 three 10-min exercise periods (VE ~ 40 L/min). Physiological responses were measured at
2 approximately 15- and 55-min of exposure. Pooling data from both the mild and
3 moderate/severe asthmatic groups (n = 40) and using only measurements made at 15 min into the
4 exposure, the group mean sRaw was doubled with 0.6-ppm SO2 exposure. In the project report
5 (Hackney et al., 1987) upon which the Linn et al. (1987) article was based, individual data were
6 presented that showed that 15/40 moderate/severe subjects (37.5%) had a doubling of the sRaw
7 at concentrations of <0.6-ppm 862.
8 Linn et al. (1987) demonstrated that moderate and severe asthmatics had the most severe
9 physiological and symptom responses. While the moderate/severe asthmatics were more
10 responsive than mild asthmatics following exposure to clean air during exercise, their increases
11 in response with increasing SC>2 concentrations were similar to the mild asthmatic group. Thus,
12 it was concluded that SC>2 response was not strongly dependent on the clinical severity of
13 asthma. Figure 3.1-5 illustrates the effect of varying concentrations of SC>2 on sRaw for the mild
14 and moderate/severe asthmatics groups after adjusting for the effect of exercise. The apparent
15 lack of correlation between 862 response and asthma severity should be interpreted with caution,
16 since the 862 response may have been attenuated by medication usage or its persistence. Three
17 of the moderate/severe asthmatics were unable to withhold medication usage during the exposure
18 period. It was also suggested that individual SO2 response could not be predicted by severity of
19 asthma or asthma status, since a few of the atopic individuals who were not asthmatic nor had
20 exercise-induced bronchoconstriction were reactive to SC>2. On the other extreme, a few of the
21 asthmatics, including some in the moderate/severe group, did not react to 0.6-ppm SC>2.
22 Nevertheless, the largest sRaw increases and most substantial decrements in FEVi occurred in
23 the moderate/severe asthmatic group.
24 One of the key studies discussed in the 1986 Second Addendum was by Horstman et al.
25 (1986) who exposed 27 asthmatic subjects for 10 min on different days to concentrations of SO2
26 between 0- and 2-ppm SO2 under exercising conditions (VE = 42 L/min). These authors reported
27 that for 25% of the subjects, the concentration of SO2 needed to produce a doubling of the sRaw
28 (PC(SO2)) was <0.5 ppm, and for about 20% of the subjects the PC(SO2) was >1.95 ppm, with a
29 median PC(SO2) of 0.75 ppm. Based on a cumulative frequency plot of PC(SO2) versus SO2
30 concentration (Figure 3.1-6), approximately 35% of asthmatic subjects in the Horstman et al.
September 2007 3-19 DRAFT-DO NOT QUOTE OR CITE
-------
CM
I
o
1
25
20
10
Figure 3.1-5.
SO2 Increment
J Exercise Increment
I Baseline sRaw
Mild
o.o
0.4
0.6
0.0
0.4
0.6
S02 (ppm)
Specific airways resistance (sRaw) of 16 mild and 24 moderate
asthmatic subjects exposed to 0-, 0.4-, and 0.6-ppm SOi. The exercise
increment represents the increase in sRaw following exercise with
exposure to clean air. Redrawn from the 1994 Supplement to the
Second Addendum (U.S. Environmental Protection Agency, 1994).
Source: Linn etal. (1987).
1 study (1986) reached PC(SC>2) at <0.6-ppm 862. This is consistent with the 37.5% incidence of
2 PC(SO2) at concentrations <0.6 ppm observed by Hackney et al. (1987).
3 Though Hackney et al. (1987) demonstrated the distribution of bronchial sensitivity of
4 asthmatics to 862, the authors cautioned against expressing 862 response in terms of PC(SC>2).
5 Hackney et al. (1987) noted several limitations to using PC(SO2) analysis for risk assessment
6 purposes. First, the choice of a 100% increase in sRaw is arbitrary and may not necessarily have
7 any health significance. For example, as noted by the authors, an increase in sRaw from 2 to 4
8 would meet the 100% criterion but may not be of clinical significance. However, an increase
9 from 12 to 22, while not meeting the criterion, would be of clinical significance. Second, there
September 2007
3-20
DRAFT-DO NOT QUOTE OR CITE
-------
100-
*"?
8^ 75-
u
c
0)
3
CT
0)
ul 50-
2) for an individual subject. For each
subject, PC(SC>2) is determined by plotting change in sRaw, corrected
for exercise-induced bronchoconstriction, against SOi concentration.
The SOi concentration that caused a 100% increase in sRaw is
determined by linear interpolation.
Source: Horstman et al. (1986).
1 may be loss of information from the rest of the exposure-response curve other than the chosen
2 point. For example, two subjects may have similar values of PC(SC>2) but substantially different
3 overall risk because of differences in threshold levels and slopes. Finally, PC(SC>2) based on the
4 Hackney et al. (1987) study was not necessarily a stable and reproducible measurement. In some
5 cases, the sRaw change exceeded 100% at low concentrations but not at high concentrations.
6 Two key studies have shown that a bronchoconstrictive response to SC>2 can occur in as
7 little as 2 min in asthmatic subjects. Horstman et al. (1988) exposed 12 SO2-sensitive asthmatic
8 subjects to 1.0-ppm SO2 with exercise (VE = 40 L/min). Correcting for exercise-induced
9 responses, sRaw was shown to increase by 121% after a 2-min exposure and by 307% after a
September 2007
3-21
DRAFT-DO NOT QUOTE OR CITE
-------
1 5-min exposure. Balmes et al. (1987) exposed 8 asthmatic subjects to 0.5- and 1.0-ppm 862
2 during eucapnic hyperpnea (60 L/min) by mouthpiece on separate days for 1-, 3- and 5-min
3 durations. The magnitude of bronchoconstriction increased progressively over the three time
4 periods. At 0.5-ppm SO2, sRaw increased by 34, 173, and 234% compared to baseline at 1, 3,
5 and 5 min of exposure, respectively. For the 1.0-ppm SC>2 exposure, sRaw increased by 93, 395,
6 and 580% compared to baseline at 1, 3, and 5 min of exposure, respectively.
7 The interaction of SC>2 with other common air pollutants or the sequential exposure of
8 SC>2 after prior exposure to another pollutant can modify the SCVinduced respiratory effects.
9 However, only a few studies have looked at the interactive effects of coexisting ambient air
10 pollutants. These few studies have been well summarized in the 1994 Supplement to the Second
11 Addendum. In brief, Koenig et al. (1990) examined the effect of 15-min exposures to 0.1-ppm
12 SC>2 in adolescent asthmatics engaged in moderate levels of exercise. Immediately preceding
13 this exposure, subjects were exposed for 45 min to 0.12-ppm Os during intermittent moderate
14 exercise. In this study, subjects also underwent two additional exposure sequences with the same
15 exercise regimen: 15-min exposure to 0.1-ppm SC>2 following a 45-min exposure to clean air,
16 and 15-min exposure to 0.12-ppm Os following a 45-min exposure to 0.12-ppm Os. The authors
17 found that the change in FEVi compared to baseline was significantly different following the
18 03-862 exposure (8% decrease) when compared to the change following the air-SC>2 or 03-63
19 exposures (decreases of 3 and 2%, respectively). Torres and Magnussen (1990) and Rubinstein
20 et al. (1990) investigated the effects of a prior NO2 exposure on SO2-induced
21 bronchoconstriction in asthmatic adults. While Torres and Magnussen (1990) suggested that
22 prior exposure to NO2 increased the responsiveness to SO2, Rubinstein et al. (1990) did not find
23 that NO2 exacerbated the effects of 862.
24
25 Individuals with Chronic Obstructive Pulmonary Disease
26 Linn et al. (1985) examined the respiratory effects of 862 exposure on subjects with
27 COPD. In this controlled laboratory study, 24 subjects with COPD were exposed for 1 h to 0-,
28 0.4-, and 0.8-ppm 862 with two 15-min periods of light exercise (yE = 18 L/min). In contrast to
29 studies with asthmatics, most of the subjects in this study regularly used bronchodilators and
30 were permitted their use up to 4 h prior to the study. The authors reported no SC>2 effects on
31 sRaw, spirometric measures, or arterial oxygen saturation. While it was concluded that older
September 2007 3-22 DRAFT-DO NOT QUOTE OR CITE
-------
1 adults with COPD seem less reactive to 862 compared to heavily exercising young adult
2 asthmatics, it was thought that this may be due to differences in medication usage as well as to
3 the lower ventilation rate observed in subjects with COPD, which would itself result in a
4 reduction in the pulmonary uptake of SC>2.
5
6 Summary of Human Clinical Studies on Lung Function in Adults
1 Results from human clinical studies have consistently demonstrated decreases in lung
8 function (e.g., decreased forced expiratory volume in 1 s [FEVi] and increased specific airways
9 resistance [sRaw]) following peak exposures (5 to 15 min) to SC>2. These effects have clearly
10 and consistently been shown to be exacerbated among individuals with asthma, with asthmatics
11 exhibiting significant decrements in lung function following 5- to 15-min exposures to SC>2
12 concentrations of as low as 0.5 ppm while performing moderate levels of exercise (e.g., Gong
13 et al., 1995; Horstman et al., 1986; Linn et al., 1987; Sheppard et al., 1981). The effect of peak
14 SC>2 exposure on lung function has been shown to increase in magnitude with increasing SC>2
15 concentrations above 0.5 ppm. Studies have further observed significant decrements in lung
16 function in some sensitive asthmatics following 5-15 min exposures to SC>2 concentrations of as
17 low as 0.25 ppm while performing moderate levels of exercise (Horstman et al., 1986; Sheppard
18 et al., 1981). Thus, the observations of increased bronchoconstriction and airway resistance in
19 human clinical studies provide clear evidence for 862 effects with peak exposure.
20
21 Animal Toxicological Studies
22 The 1982 AQCD reported bronchoconstriction (as indicated by increased pulmonary
23 resistance) as the most sensitive indicator of lung function effects of acute SC>2 exposure based
24 on the observations of increased pulmonary resistance in guinea pigs that were acutely exposed
25 to 0.16-ppm SC>2. Some of the new animal toxicological studies are consistent with these
26 observations. These studies on lung function are summarized in Annex Table AX4-1.
27 Increases in pulmonary resistance and decreased dynamic compliance were the most
28 frequently observed effects in conscious guinea pigs exposed to 1-ppm 862 for 1 h (Amdur et al.,
29 1983). Studies to understand the potential role of neuronal component in SO2-induced
30 pulmonary resistance used the anesthetics ketamine in guinea pigs exposed to 1-ppm SC>2 for
31 3 h/day for 6 days (Conner et al., 1985), carbamate in rabbits exposed to 5-ppm SC>2 for 45 min
32 (Barthelemy et al., 1988), or surgical manipulation (bivagotomy). These studies indicted that
September 2007 3-23 DRAFT-DO NOT QUOTE OR CITE
-------
1 pulmonary resistance was increased in ethyl carbamate-anesthetized rabbits exposed to 862 but
2 not in ketamine-anesthetized guinea pigs and that the SCVinduced increase in lung resistance
3 was not mediated by the vagus nerve in rabbits. Further, observations of the elimination of
4 reflex bronchoconstrictor response by phenyldiguanide in rabbits exposed to 5-ppm SO2, but not
5 the lung resistance induced by histamine, suggested that SO2-induced bronchoconstriction in
6 rabbits is not mediated through the vagus nerve. Though these results provided some
7 understanding on the mechanisms involved in the development of SO2-induced
8 bronchoconstriction, these studies were carried out using only one 862 exposure dose and
9 precluded assessment of concentration-response relationships and identification of a no-effect
10 level.
11 In summary, animal studies have shown that guinea pigs exposed to 0.16- to 1-ppm and
12 rabbits exposed to 5-ppm SC>2 have increased pulmonary resistance that is not mediated through
13 the vagus nerve.
14
15 3.1.1.3 Airway Inflammation
16 One epidemiological study by Adamkiewicz et al. (2004) examined exhaled nitric oxide
17 (eNO) as a biological marker for inflammation in 29 older adults (median age 70.7 years) in
18 Steubenville, OH. The mean 24-h average SO2 concentration was 12.5 ppb (IQR 11.5). The
19 authors reported that, while significant and robust associations were observed between increased
20 daily levels of fine PM (PM2.5) and increased eNO, no associations were observed with any of
21 the other pollutants examined, including SO2, NO2, and Os.
22 In a controlled-exposure, time-response study, Sandstrom et al. (1989) exposed 22
23 healthy male subjects for 20 min to 8-ppm SO2 under light exercising conditions.
24 Bronchoalveolar lavage was performed in all subjects at least 2 weeks prior to exposure, as well
25 as at 4, 8, 24, and 72 h after exposure in 8/22 subjects. The authors found that as early as 4 h
26 after exposure to SO2, lysozyme-positive macrophages, lymphocytes, and mast cells were
27 significantly increased compared to baseline. Twenty-four hours after exposure, these markers
28 of inflammation, as well as the total alveolar macrophages (AM) and total cell number, were at
29 peak levels. This study demonstrated that SO2-induced inflammation may extend beyond the
30 short time period often associated with SO2 effects. A limitation of this study, however, is that
31 the levels of exposure used are well above air pollution levels normally encountered. Tunnicliffe
32 et al. (2003) measured levels of eNO in asthmatic and healthy adult subjects before and after 1-h
September 2007 3-24 DRAFT-DO NOT QUOTE OR CITE
-------
1 exposure to 0.2-ppm SC>2 under resting conditions. While eNO concentrations were higher in the
2 asthmatic versus healthy subjects, no significant difference was observed between pre- and
3 postexposure in either group.
4 Two recent studies that examined inflammatory responses in animals exposed to SC>2
5 report characteristic responses such as leukocyte influx and changes in enzyme levels or
6 activities in the lung at high SC>2 concentrations. In brief, Meng et al. (2005) observed elevated
7 levels of pro-inflammatory cytokines interleukin-6 and tumor necrosis factor-a in lung tissue of
8 mice exposed to SO2 concentrations of 5.35 and 10.7 ppm. The levels of anti-inflammatory
9 cytokine transforming growth factor-pi were not affected at any exposure level. For example, in
10 rats exposed to 5, 50, or 100 ppm of 862 for 5 h/day for 28 days, increased leukocyte numbers in
11 bronchoalveolar lavage fluid was observed at 100 ppm, but no such infiltration of leukocytes was
12 observed in rats exposed to 5 or 50 ppm (Langley-Evans et al., 1996). The animal toxicological
13 studies on airway inflammation are summarized in Annex Table AX4-2.
14 Overall, the limited epidemiological, human clinical, and toxicological evidence does not
15 indicate that exposure to SC>2 at current ambient concentrations is associated with inflammation
16 in the airways.
17
18 3.1.1.4 Airway Hyperresponsiveness and Allergy
19 A limited number of epidemiological studies have examined the association between 862
20 and airway hyperresponsiveness (AHR). Other studies have also considered individuals with
21 AHR and atopy as a potentially susceptible subgroup to SO2-related health effects. These studies
22 are summarized in Annex Table AX5-1. S0yseth et al. (1995) investigated the effect of short-
23 term exposure to SO2 and fluoride on the number of capillary blood eosinophils and the
24 prevalence of bronchial hyperresponsiveness (BHR) in schoolchildren aged 7 to 13 years
25 (n = 620) from two regions in Norway, a valley containing an SO2-emitting aluminum smelter
26 and a similar but nonindustrialized valley. The median 24-h average SO2 concentration was
27 22.2 |ig/m3 (8 ppb, 10th-90th percentile: 1, 33) in the exposed area and 2.5 |ig/m3 (1 ppb,
28 10th-90th percentile: 0, 4). The mean number of eosinophils was significantly greater in
29 children living near the aluminum smelter compared to the nonindustrialized area. However,
30 within children in the exposed area, a negative concentration-response relationship was observed
31 between mean eosinophils and previous-day 24-h average SO2. The observed association
September 2007 3-25 DRAFT-DO NOT QUOTE OR CITE
-------
1 between SO2 and eosinophils was limited to atopic children. In children living in the exposed
2 area, a statistically significant positive association was observed between prevalence of BHR and
3 previous-day 24-h average SO2 concentrations. Similar associations were observed for fluoride.
4 The authors hypothesized that recent exposure to SC>2 may have induced changes in the airways
5 leading to BHR, in addition to recruitment of eosinophils to the airways in atopic subjects.
6 Exposure to PM was not assessed in this study.
7 A study by Taggart et al. (1996) examined the effect of summertime air pollution levels
8 in northwestern England on BHR in nonsmoking, asthmatic subjects (n = 38) aged 18 to 80 years
9 who were determined to be methacholine (MCh) reactors. Subjects were tested multiple times,
10 for a total of 109 evaluable challenge tests. The maximum 24-h average 862 concentration
11 during the study period was 103.7 |ig/m3 (40 ppb). This study reported that 24-h average SO2
12 levels were marginally associated with a decreased dose of MCh required for a 20% drop in the
13 postsaline FEVi (PD20FEVi).
14 Other epidemiological studies investigated the effect of exposure to SO2 on children and
15 adults with BHR and atopy. Boezen et al. (1999) examined children (n = 459) aged 7 to 11 years
16 old in the Netherlands and tested them for BHR using MCh and relatively high serum
17 concentrations of total IgE (>60 kU/L, the median value). These children were a subset of a
18 larger cohort examined in van der Zee et al. (1999). It was hypothesized that children with BHR
19 and atopy, indicated by raised serum total IgE, may be susceptible to the effects of air pollution.
20 One of the strengths of this study was that the use of BHR and serum IgE concentration as a
21 marker for susceptibility was less prone to error than self-reported chronic respiratory symptoms.
22 A total of 121 children were found to have BHR and relatively high serum total IgE, 67 had
23 BHR and relatively low serum total IgE, 104 had no BHR but had a relatively high serum total
24 IgE concentration, and 167 were found to have neither BHR nor relatively high serum total IgE.
25 In the subset of children with relatively low serum total IgE with or without BHR, no
26 associations were observed between SO2 and any respiratory symptoms. However, for children
27 with relatively high serum total IgE either with or without BHR, the prevalence of lower
28 respiratory symptoms increased with increasing SO2 concentrations. For children with BHR and
29 relatively high serum total IgE, the OR for the prevalence of lower respiratory symptoms was
30 1.70 (95% CI: 1.26, 2.29) with a 5-day moving average for every 10-ppb increase in SO2. For
September 2007 3-26 DRAFT-DO NOT QUOTE OR CITE
-------
1 children without BHR but with relatively high serum total IgE the OR was 1.82 (95% CI: 1.33,
2 2.50) with a 5-day moving average.
3 Boezen et al. (2005) did a similar study in 50- to 70-year-old adults (n = 327) in the
4 Netherlands. Subjects underwent spirometry and MCh challenges to determine AHR. The
5 subgroup of individuals with elevated serum total IgE, both with (n = 48) and without (n = 112)
6 AHR were found to be more susceptible to air pollutants compared to those who did not have
7 elevated serum total IgE (n = 167). Significant associations were observed between previous-
8 day 24-h average SO2 concentrations and the prevalence of upper respiratory symptoms in those
9 with elevated serum total IgE. Stratified analyses by gender indicated that, among those with
10 AHR and elevated IgE, only males (n = 25) were at a higher risk for respiratory symptoms. The
11 OR for these males was 3.54 (95% CI: 1.79, 7.07) increase in 24-h average SO2 for a 5-day
12 moving average, compared to 1.05 (95% CI: 0.59, 1.91) for the females.
13 One human clinical study investigated the relationship between hyperresponsiveness to
14 SO2 and AHR to MCh (Nowak et al., 1997). Responsiveness to both MCh and SO2 were tested
15 on 790 subjects between the ages of 20 and 44. The authors reported that among subjects with
16 AHR to MCh, 22.4% were hyperresponsive to SO2, whereas among the MCh-nonresponsives
17 only 0.4% were hyperresponsive to SO2. Using a logistic regression model, they also determined
18 that a positive skin test (p < 0.05), a positive history of respiratory symptoms (p < 0.05), and
19 hyperresponsiveness to MCh (p < 0.0001) were significant predictors of a positive SO2 response.
20 A limited number of animal studies also suggest acute SO2-induced increases in airway
21 obstruction and hypersensitivity in allergen-sensitized guinea pigs and sheep. These
22 toxicological studies are summarized in Annex Table AX4-3. Bronchial responses (pulmonary
23 resistance or reduced dynamic compliance to agonists (i.e., histamine, MCh,
24 5-hydroxytryptamine) are examined after exposure to evaluate toxic effects of pulmonary
25 toxicants. Exposure of rabbits to 5-ppm SO2 for 2 h had no effect on airway responsiveness to
26 histamine (Douglas et al., 1994). Even at higher concentrations of 10-ppm SO2 for 5 min,
27 hyperresponsiveness and hyperreactivity effects to aerosolized MCh or 5-hydroxytryptamine
28 were not observed in dogs (Lewis and Kirchner, 1984), but positive responses were observed at
29 the higher concentration of 30 ppm. Studies with chronic exposure of dogs suggest no increased
30 sensitivity to agonists at SO2 concentrations of > 15 ppm (Scanlon et al., 1987).
September 2007 3-27 DRAFT-DO NOT QUOTE OR CITE
-------
1 Riedel et al. (1988) studied the effect of 862 exposure in ovalbumin-sensitized guinea
2 pigs exposed to 862 at 0.1, 4.3, or 16.6 ppm for 8 h/day for 5 days. On bronchial provocation,
3 they observed increased bronchial obstruction in animals exposed to 0.1-ppm SO2 compared to
4 air-exposed animals. In addition, increased amounts of anti-ovalbumin IgG antibodies were
5 detected in bronchoalveolar lavage fluid of animals exposed to >4.3-ppm SC>2 and in the serum
6 of animals exposed to > 0.1-ppm 862.
7 Similar findings were observed in studies in which guinea pigs were exposed to a single
8 SC>2 concentration. Airway obstruction induced by an ovalbumin challenge was higher in
9 ovalbumin-sensitized guinea pigs exposed to 0.1-ppm 862 for 5 h/day for 5 days compared to
10 sensitized guinea pigs that were not exposed to 862 (Park et al., 2001). In guinea pigs sensitized
11 with Candida albicans, exposure to 5-ppm SC>2 for 4 h/day on 5 days/week for 6 weeks resulted
12 in an increased number of animals displaying prolonged expiration or inspiration after an
13 inhalation challenge with C. albicam (Kitabatake et al., 1992, 1995).
14 The effect of SC>2 on antigen-induced sensitivity reactions was assessed in sheep. A 4-h
15 exposure to 5-ppm SC>2 increased airway reactivity in response to carbachol in sheep that had
16 been sensitized to Ascaris suum antigen 24-h postexposure, but increased sensitivity was not
17 observed in nonsensitized sheep (Abraham et al., 1981).
18 Limited epidemiological evidence suggests that exposure to 862 may lead to AHR in
19 atopic individuals. Toxicological studies that observed increased airway obstruction and
20 hypersensitivity in allergen-sensitized animals provide biological plausibility. The
21 epidemiological evidence further indicates that atopic individuals may be at increased risk for
22 SO2-induced respiratory symptoms.
23
24 3.1.1.5 Lung Host Defense
25 An additional concern has been the potential for SC>2 exposure to enhance susceptibility
26 to, or the severity of illness resulting from, respiratory infections, especially in children. School
27 absenteeism is an indicator of morbidity in children resulting from acute conditions. Respiratory
28 conditions are the most frequent cause, particularly influenza and the common childhood
29 infectious diseases. Park et al. (2002) examined the association between air pollution and school
30 absenteeism in 1,264 first- to sixth-grade students attending school in Seoul, Korea. The study
31 period extended from March 1996 to December 1999, with a mean 24-h average SC>2
September 2007 3-28 DRAFT-DO NOT QUOTE OR CITE
-------
1 concentration of 9.19 ppb (SD 4.61). Note that analyses were performed using Poisson
2 Generalized Additive Model (GAM) with default convergence criteria. Same-day 862
3 concentrations were positively associated with illness-related absences (9% increase [95% CI: 7,
4 12] per 5.68-ppb increase in 24-h average 802), but inversely associated with non-illness-related
5 absences (5% decrease [95% CI: 1, 8]). PMio and Os concentrations also were positively
6 associated with illness-related absences. In two-pollutant models containing SC>2 and either
7 PMio or Os, the SC>2 estimates were robust. These results are consistent with those of Ponka
8 (1990), who observed that absenteeism due to febrile illnesses among children in day care
9 centers and schools and in adults was significantly higher on days of higher 862 concentrations
3
10 (>21.1 |ig/m [8.1 ppb] weekly mean of 1-h average) compared to days of lower SO
2
11 concentrations. In addition, on days of higher SO2 concentrations, the mean weekly number of
12 cases of upper respiratory infections and tonsillitis reported from health centers increased.
13 Temperature, but not NC>2, was also found to be associated with febrile illnesses and respiratory
14 tract infections. From these epidemiological studies, it is unknown whether SC>2 increases
15 susceptibility to infection or whether they exacerbate preexisting morbidity following infection.
16 Pino et al. (2004) examined the association between air pollution and respiratory illnesses
17 in a cohort of 504 infants recruited at 4 months of age from primary health care units in
18 southeastern Santiago, Chile. The infants were followed through the first year of life. The mean
19 24-h average 862 concentration was 11.6 ppb (5th-95th percentile: 3.0, 29.0). The most
20 frequent diagnosis during follow-up was wheezing bronchitis. No associations were observed
21 between current-day or previous-day SC>2 and wheezing bronchitis, but with a 7-day lag, a 21%
22 (95% CI: 8, 39) increased risk in wheezing bronchitis was observed per 10-ppb increase in 24-h
23 average SC>2. However, it should be noted that stronger associations were observed with PM2.5,
24 which was well correlated with SC>2 (r = 0.73). These epidemiological studies are summarized in
25 Annex Table AX5-1.
26 The animal toxicological studies reviewed in the 1982 AQCD on the effects of SC>2 on
27 lung defenses reported concentration and species-specific differential effects. In rats exposed to
28 0.1-ppm SO2 for ~2 to 3 weeks, clearance of labeled particles from the lung was accelerated at
29 10 and 23 days following exposure. While this clearance was accelerated at 10 days, it slowed
30 down at 25 days in rats exposed to 1 ppm for ~2 to 3 weeks. No difference in macrophage-
31 containing particles was observed in rats chronically-exposed to up to 3-ppm SC>2. Only one
September 2007 3-29 DRAFT-DO NOT QUOTE OR CITE
-------
1 study published after the last review evaluated mucociliary clearance in rats after exposure to
2 SO2. In this subchronic study, no effect on clearance of radiolabeled particles from the lung was
3 observed in rats exposed to 5-ppm SC>2 for 2 h/day for 4 weeks (Wolff et al., 1989), which is in
4 contrast to the altered clearance reported in the 1982 AQCD. The current studies are
5 summarized in Annex Table AX4-4.
6 There was only limited data available in the 1982 AQCD from animal toxicological
7 studies on effects of SC>2 on immune and macrophage function. The studies reviewed there
8 indicated no effect on susceptibility to bacterial infection with exposure to 862 at <5 ppm for
9 3 months and alterations in pulmonary immune system were reported with chronic exposure of
10 mice to 2-ppm 862. At high-dose exposures to 7- to 10-ppm 862 for 7 days, impairment of
11 antiviral defenses was observed in mice.
12 Two recent studies using a 10-ppm SC>2 exposure regimen in mice found no effect on
13 bactericidal activity toward Staphylococcus aureus following acute (4 h) exposure (Clarke et al.,
14 2000; Jakab et al., 1996). However, increased mortality rate and decreased survival time were
15 observed in mice that were exposed to the same dose for 1 day or 1, 2, or 3 weeks and then
16 challenged with an aerosol of Klebsiellapneumoniae (Azoulay-Dupuis et al., 1982). No effects
17 on macrophage phagocytosis of red blood cells were observed in mice exposed to 10-ppm 862
18 for 4 h (Clarke et al., 2000; Jakab et al., 1996).
19 Although the limited epidemiologicical evidence weakly suggests a possible association
20 between ambient SO2 concentrations and increased respiratory illnesses, there is little
21 toxicological evidence to support this observed relationship.
22
23 3.1.1.6 Emergency Department Visits and Hospitalizations for Respiratory Diseases
24 Total respiratory causes for ED visits typically include asthma, pneumonia, bronchitis,
25 and emphysema (collectively referred to as COPD), upper and lower respiratory infections, and
26 other minor categories (U.S. Environmental Protection Agency, 2006d). Temporal associations
27 between ED visits or hospital admissions for respiratory diseases and the ambient concentrations
28 of SC>2 have been the subject of >50 well-conducted research publications since 1994. In
29 addition to considerable statistical and analytical refinements, the more recent studies have
30 examined responses of morbidity in different age groups, the effect of seasons on ED and
September 2007 3-30 DRAFT-DO NOT QUOTE OR CITE
-------
1 hospital usage, and multipollutant models to evaluate potential confounding effects of
2 copollutants.
3
4 All Respiratory Diseases
5 Relatively few studies of ED visits for all respiratory causes were conducted in
6 comparison with studies that examined hospital admissions for all respiratory causes as the
7 outcome. Collectively, studies of ED visits and hospitalizations provide suggestive evidence of
8 an association between ambient SO2 levels and ED visits and hospitalizations for all respiratory
9 causes among children (0 to 14 years) and older adults (65+ years). The studies that examined
10 the association of these outcomes and 862 levels among adults (15 to 64 years) overwhelmingly
11 reported null results. When all age groups were combined, the results of ED and hospitalization
12 studies were mixed; it is likely that any significant effect estimates found in these studies were
13 driven by increases in the very young and/or older adult subpopulations. The epidemiological
14 studies of ED visits and hospital admissions for respiratory causes are summarized in Annex
15 Tables AX5-2.
16 The results from the hospitalization and ED studies, separated by analyses among all ages
17 or age-specific analyses, are shown in Figures 3.1-7 and 3.1-8. Wilson et al. (2005) examined
18 ED visits for all respiratory causes in Portland, ME from 1996 through 2000 and in Manchester,
19 NH from 1998 through 2000. The mean 1-h max SC>2 concentration in Portland was 11.1 ppb
20 (SD 9.1), and it was higher during the winter months (mean 17.1 ppb (SD 12.0]) and lower in the
21 summer months (mean 9.1 ppb [SD 8.0]). In Manchester, the mean 1-h max SC>2 concentration
22 was 16.5 ppb (SD 14.7 ppb), and it was higher in the winter months (mean 25.7 ppb [SD 15.8])
23 compared to the summer months (mean 10.6 ppb [SD 15.1]). When all ages where included in
24 analyses, Wilson et al. (2005) found positive associations between ED visits and SC>2, with an
25 8% (95% CI: 3.0, 11) and 11% (95% CI: 0.0, 20.0) increased risk per 10-ppb increase in 24-h
26 average SC>2 at a 0-day lag in Portland, ME and Manchester, NH, respectively.
27 Peel et al. (2005) investigated ED visits for all respiratory causes in Atlanta, GA from
28 1993 through 2000. This study included 484,830 ED visits. The mean 1-h max SO2
29 concentration during the study period was 16.5 ppb (SD 17.1). Peel et al. (2005) found a weak
30 positive relationship between ED visits and SO2, though the increased risk was not statistically
31 significant (1.6% [95% CI: -0.6, 3.8]). Tolbert et al. (2007 in press) recently reanalyzed these
September 2007 3-31 DRAFT-DO NOT QUOTE OR CITE
-------
Reference Location Age Other Lag
Wilson et al. (2005) Portland, ME All 0 -
Manchester, NH
Peel etal. (2005) Atlanta, GA All 1-hmaxO-2 -
Atkinson etal.(1999b) London, UK All 1 -
Luginaah et al. (2005) Windsor, ON All Females 1 -
Males
Petroeschevsky etal. (2001) Brisbane, Australia All 2
Andersonetal.(2001) West Midlands, UK All NR -
Atkinson etal. (1999a) London, UK All 1 -
Waltersetal.(1994) Birmingham, UK All Summer 0 -
Winter 0 -
Ponce de Leon etal. (1996) London, UK All 1 -
Dab etal. (1996) Paris, France All 0-2 -
Llorca etal. (2005) Torrelavega, Spain All NR -
Fuscoetal.(2001} Rome, Italy All 0 -
-B-
i
[-[ ED Visits
D
B-
-B-
Hospital
PI Admissions
~H
B-
*
]
-B-
3
—
B—
l l I
0.8 0.9 1.0 1.1 1.2 1.3
Relative Risk
Figure 3.1-7. Relative risks (95% CI) of SOi-associated emergency department
visits (*) and hospitalizations for all respiratory causes among all
ages. Risk estimates are standardized per 10-ppb increase in 24-h
average SOi concentrations or 40-ppb increase in 1-h max SOi. The
size of the box of the central estimate represents the relative weight of
that estimate based on the width of the 95% CI.
September 2007
3-32
DRAFT-DO NOT QUOTE OR CITE
-------
Wilson et al (2005)
Atkinson etal. (1999b)
Wilson etal (2005)
Atkinson el al.(1999b)
Wilson et al (2005)
Atkinson etal. (1999b)
Wong etal. (1999)
Gouveia and Fletcher (2000)
Barnett et al. (2005)
Luginaah et al (2005)
Petroeschevsky etal. (2001)
Anderson et al. (2001)
Atkinson etal.(1999a)
Ponce de Leon etal. (1996)
FuscoeEai. (2001)
Barnett et al. (2005)
Wong etal. (1999)
Luginaah etal (2005)
Location
Portland, ME
Manchester, NH
London, UK
Portland, ME
Manchester, NH
London, UK
Portland, ME
Manchester, NH
London, UK
Hong Kong, China
Sao Paulo, Brazil
Mullicity, Australia
Windsor, OM
Brisbane, Australia
West Midlands, UK
London, UK
London, UK
Rome, Italy
Mullicity, Australia
Hong Kong, China
Windsor, ON
Other
0-14
0-14
15-64
15-64
65+
65+
0-4
<5
1-4
0-14
0-14
0-14
0-14
0-14
0-14
5-14
5-64
15-64
Females
Males
Females
Males
Schouten etal. (1996)
Spix etal. (1998)
Anderson etal. (2001)
Atkinson et al. (1999a)
Ponce de Leon etal. (1996)
Vigotti etal. (1996)
Schwartz (1995)
Fung et al (2006)
Yang et al. (2003a)
Luginaah et al. (2005)
Schouten et al. (1996)
Spix etal. (1998)
Anderson etal. (2001)
Atkinson et al. (1999 a)
Ponce de Leon etal. (1996)
Vigotti etal. (1996)
Wong etal. (1999)
Amsterdam
Rotterdam
Mullicity, Europe
West Midlands, UK
London, UK
London, UK
Milan, Italy
New Haven, CT
Tacoma, WA
Vancouver, BC
Vancouver, BC
Windsor, ON
Amsterdam
Mullicity, Europe
West Midlands, UK
London, UK
London, UK
Milan, Italy
Hong Kong, China
15-64
15-64
15-64
15-64
15-64
15-64
65+
65+
65+
65+ Females
Males
65+
65+
65+
65+
65+
65+
65+
Lag
0
2
0
3
0
3
0
1
0-1
1
0
0-1
0
1
0
0-1
0
1
0-3
0-2
NR
0-1
3
1
0
2
0
0-?
0
1
0-3
NR
0-1
3
2
0
0
0.
•
1 —
• —
1 —
1
8 0.9 1
ED Visits
•
•
-• Hospital
Admissions
•B
— •
}
1
-«
•B-
— •
1
fl-
-i —
ft-
1 1 1
.0 1.1 1.2 1.3
Relative Risk
Figure 3.1-8. Relative risks (95% CI) of SOi-associated emergency department visits and
hospitalizations for all respiratory causes, stratified by age groups. Risk
estimates are standardized per 10-ppb increase in 24-h average SOi
concentrations or 40-ppb increase in 1-h max SOi. The size of the box of the
central estimate represents the relative weight of that estimate based on the
width of the 95% CI.
September 2007
3-33
DRAFT-DO NOT QUOTE OR CITE
-------
1 data with 4 additional years of data and found the same results. An analysis by Dab et al. (1996)
2 examined the association between 862 and hospital admissions for all respiratory causes using
3 both the 24-h average and 1-h max. It should be noted that they observed similar effect estimates
4 for both exposure metrics, but only the estimate using 24-h average was statistically significant
5 (1.1%[95%CI: 0.1, 2.0] per 10-ppb increase in 24-h average SO2 versus 1.9% [95% CI: -1.3,
6 5.0]) per 40-ppb increase in 1-h max 802).
7 When analyses were stratified to include only children (0 to 14 years), Wilson et al.
8 (2005) did not find statistically significant associations between ED visits and 862 in Portland,
9 ME or Manchester, NH. Additional evidence of a modest association between 862 and ED visits
10 or hospitalizations for all respiratory causes in children from several Australian (Barnett et al.,
11 2005; Petroeschevsky et al., 2001) and European (Anderson et al., 2001; Atkinson et al.,
12 1999a,b) studies. Increased risks ranging from 3 to 22% per 10-ppb increase in 24-h average
13 SC>2 were reported by these studies. In a multicity study spanning Australia and New Zealand,
14 Barnett et al. (2005) compared hospital admission data collected from 1998 through 2001 with
15 ambient SC>2 concentrations, where the mean 24-h average SC>2 concentration ranged from 0.9 to
16 4.8ppb. The authors found a 5% (95% CI: 1, 9) increased risk per 10-ppb increase in 24-h
17 average 862 among children (1 to 4 years) in these cities. However, some additional European
18 (Fusco et al., 2001; Ponce de Leon et al., 1996) and Latin American (Braga et al., 1999, 2001)
19 studies did not find statistically significant associations between ambient SO2 concentrations and
20 hospitalizations for all respiratory causes among children.
21 Wilson et al. (2005) found a positive association between ED visits and SC>2, with a 16%
22 (95% CI: 8.0, 25.0) increased risk per 10-ppb increase in 24-h average SC>2 at a 0-day lag in
23 Portland, ME and a null association in Manchester, NH when only older adults (65+ years) were
24 considered. In another two-city study, Schwartz (1995) compared 13,740 hospital admissions in
25 New Haven, CT and Tacoma, WA from 1988 through 1990 with ambient SC>2 concentrations.
26 The mean 24-h average SC>2 concentration was 29.8 ppb (90th percentile: 159) in New Haven
27 and 16.8 ppb (90th percentile: 74) in Tacoma. Schwartz found positive associations between
28 hospitalizations and SO2, with a 2% (95% CI: 1.0, 3.0) and 3% (95% CI: 1.0, 6.0) increased risk
29 per 10-ppb increase in 24-h average SC>2 at a 0-day lag in New Haven and Tacoma, respectively.
30 In two-pollutant models, the SC>2 effect estimate from New Haven, but not Tacoma, was found to
31 be robust to adjustment for PMio. Here, the term robust is used to indicate that there was little
September 2007 3-34 DRAFT-DO NOT QUOTE OR CITE
-------
1 change in the magnitude of the central estimate, though statistical significance may have been
2 lost. In Vancouver, BC, both Fung et al. (2006) and Yang et al. (2003a) also found positive
3 associations between hospitalizations and SO2. In a multipollutant model including coefficient
4 of haze (CoH), NO2, 63, and CO, the 862 effect estimate diminished slightly (Yang et al.,
5 2003a).
6 Additional evidence of a positive association between ED visits or hospitalizations for
7 all respiratory causes among older adults and SC>2 comes from several European (Spix et al.,
8 1998; Sunyer et al., 2003a; Vigotti et al., 1996) and Australian (Petroeschevsky et al., 2001)
9 studies. Increased risks ranging from 1 to 12% per 10-ppb increase in 24-h average 862 were
10 reported by these studies. Petroeschevsky et al. (2001) examined 33,710 hospital admissions
11 in Brisbane, Australia from 1987 through 1994. The mean 24-h average SC>2 concentration was
12 4.1 ppb and was highest in the winter months (4.8 ppb) and lowest in the spring months
13 (3.7 ppb). Petroeschevsky et al. found a 12% (95% CI: 2.0, 23.0) increased risk per 10-ppb
14 increase in 24-h SC>2 at 0-day lag. Additional European studies did not find statistically
15 significant associations between ambient SC>2 concentrations and hospitalizations for all
16 respiratory causes among older adults (Schouten et al., 1996; Anderson et al., 2001; Atkinson
17 et al., 1999a; Ponce de Leon et al., 1996).
18 In summary, many studies have observed positive, though not statistically significant
19 associations between ambient 862 concentrations and ED visits and hospitalizations, particularly
20 among children and older adults (age 65+ years).
21
22 Asthma
23 Studies of ED visits and hospitalizations provide suggestive evidence of an association
24 between ambient SC>2 levels and ED visits and hospitalizations for asthma among children (0 to
25 14 years). The studies that examined the association of these outcomes and SC>2 levels among
26 adults (15 to 64 years) and older adults (65+ years) overwhelmingly reported null results. When
27 all age groups were combined, the results of ED and hospitalization studies were mixed, and it is
28 likely that any significant effect estimates found in these studies were driven by increases in the
29 young subpopulations.
30 The results from the hospitalization and ED studies, separated by analyses among all ages
31 and age-specific analyses, are shown in Figures 3.1-9 and 3.1-10. When all ages were included
32 in analyses, Wilson et al. (2005) found a positive association between ED visits and SC>2, with a
September 2007 3-35 DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Wilson et al. (2005)
Peel etal. (2005)
Atkinson etal. (1999b)
Galan et al. (2003)
Petroeschevsky etal. (2001)
Schouten etal. (1996)
Anderson etal. (1998)
Atkinson etal. (1999a)
Walters etal. (1994)
Dab etal. (1996)
Fusco etal. (2001)
Tsai etal. (2003)
Wong etal. (1999)
Location
Portland, ME
Manchester, NH
Atlanta, GA
London
Madrid, Spain
Brisbane, AU
Amsterdam
London
London
Birmingham, UK
Paris, France
Rome, Italy
Kaohsiung, Taiwan
Hong Kong, China
Other Lag
2
1-hmax 0-2 _
1
0 _
1 _
0-3 _
0-3 _
1
Summer 0 _
Winter
2 _
0 _
>25C 0-2 _
<25
0 _
0
|
I
I I
7 0.8 0.9 1
R
, ED Visits
J ,
Hospital
Admissions
u
~*~
I
+-
\ \ 1 1 1 1 1
0 1.1 1.2 1.3 1.4 1.5 1.6 1.7
slative Risk
Figure 3.1-9.
Relative risks (95% CI) of SOi-associated emergency department
visits (*) and hospitalizations for asthma among all ages. Risk
estimates are standardized per 10-ppb increase in 24-h average
concentrations or 40-ppb increase in 1-h max SOi. The size of the box
of the central estimate represents the relative weight of that estimate
based on the width of the 95% CI.
September 2007
3-36
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Wilson et al. (2005)
Sunyeretal. (1997)
Atkinson etal. (1999b)
Jaffeetal. (2003)
Boutin-Forzano et al. (2004)
Wilson et al. (2005)
Sunyeretal. (1997)
Atkinson etal. (1999b)
Castellsague etal. (1995)
Teniasetal. (1998)
Wilson et al. (2005)
Petroeschevsky et al. (2001)
Anderson etal. (1998)
Atkinson etal. (1999a)
Fuscoetal. (2001)
Barnett et al. (2005)
Gouveia and Fletcher (2000)
Barnett et al. (2005)
Petroeschevsky et al. (2001)
Lin et al. (2003b)
Lee et al. (2006)
Petroeschevsky et al. (2001)
Anderson etal. (1998)
Atkinson etal. (1999a)
Sheppard et al. (1999)
Petroeschevsky et al. (2001)
Anderson etal. (1998)
Atkinson etal. (1999a)
Location
Portland, ME
Manchester, NH
Multicity, Europe
London, UK
Cincinnati, OH
Cleveland, OH
Columbus, OH
Multicity, OH
Marseille, France
Portland, ME
Manchester, NH
Multicity, Europe
London, UK
Barcelona, Spain
Valencia, Spain
Portland, ME
Manchester, NH
Brisbane, Australia
London, UK
London, UK
Rome, Italy
Multicity, Australia
Sao Paulo, Brazil
Multicity, Australia
Brisbane, Australia
Toronto, ON
Hong Kong, China
Brisbane, Australia
London, UK
London, UK
Seattle, WA
Brisbane, Australia
London, UK
London, UK
Ages
0-14
0-14
0-14
5-34
349
15-64
15-64
15-64
15-64
>14
65+
04
0-14
0-14
0-14
14
<5
5-14
5-14
6-12
<18
5-64
15-64
15-64
<65
65+
65+
65+
Other Lag
2
2-3
1
2
2
3
NR
0
2
0-3
1
Summer 2
Winter 1
0
2
04
0-3
1
0
0-1
2
0-1
1-h max 04
Boys 0
Girls
0
1
0-2
3
0
0
0-3
2
0
H
ED Visits
^_
h
• i
«
Hospital
, ' Admissions
I
-t —
III II II
6 0.7 0.8 0.9 1.0 1.1 1.3 1.5 1,7 1.9
Relative Risk
Figure 3.1-10.
Relative risks (95% CI) of SOi-associated emergency department
visits (*) and hospitalizations for asthma, stratified by age groups.
Risk estimates are standardized per 10-ppb increase in 24-h average
SOi concentrations or 40-ppb increase in 1-h max SOi. The size of the
box of the central estimate represents the relative weight of that
estimate based on the width of the 95% CI.
September 2007
3-37
DRAFT-DO NOT QUOTE OR CITE
-------
1 10% (95% CI: 2.0, 20.0) increased risk per 10-ppb increase in 24-h average SO2 at a 0-day lag
2 in Portland, ME and a null association in Manchester, NH. Peel et al. (2005) found a null
3 relationship between asthma ED visits and 1-h max SC>2.
4 When analyses were stratified to include only children (0 to 14 years), Wilson et al.
5 (2005) found positive, but not statistically significant, associations between ED visits and SC>2 in
6 Portland, ME or Manchester, NH. Similarly, Lin et al. (2003a) (Toronto, ON; mean 24-h
7 average SC>2 of 5.36 ppb [SD 5.90]) observed a weak positive association between
8 hospitalizations for asthma and 862 among girls and a null association for boys.
9 A study by Jaffe et al. (2003) examined the association between 862 and ED visits for
10 asthma in three cities in Ohio, i.e., Cincinnati, Cleveland, and Columbus, in asthmatics aged 5 to
11 34 years. The mean 24-h average SO2 concentrations were 14 ppb (range: 1, 50) in Cincinnati,
12 15 ppb (range: 1, 64) in Cleveland, and 4 ppb (range: 0, 22) in Columbus. A positive
13 association was observed in the multicity analysis, with a 6.1% (95% CI: 0.5, 11.5) increase in
14 asthma visits observed per 10-ppb increase in 24-h average SC>2. In the city-stratified analyses,
15 significant associations were only observed for Cincinnati (17.0% [95% CI: 4.6, 30.8]).
16 Stronger evidence of a positive association between ED visits or hospitalizations for
17 asthma and 862 comes from several European (Anderson et al., 1998; Atkinson et al., 1999a,b;
18 Hajat et al., 1999; Sunyer et al., 1997, 2003b; Thompson et al., 2001) and Asian (Lee et al.,
19 2002) studies. Increased risks ranging from 2 to 10% per 10-ppb increase in 24-h average 862
20 were reported by these studies. Several of these studies observed that the SO2 effect estimate
21 was robust to adjustment for BS and NC>2 (Anderson et al., 1998; Sunyer et al., 1997), but one
22 study observed that the SC>2 effect diminished considerably with adjustment for PMio and
23 benzene (Thompson et al., 2001). Atkinson et al. (1999a) compared 165,032 hospital admissions
24 in London from 1992 through 1994 with ambient SC>2 levels (mean 24-h average of 7.2 ppb [SD
25 4.7]). They found a 10% (95% CI: 4.0, 16.0) increased risk per 10-ppb increase in 24-h average
26 SC>2 at 1-day lag. Additional European (Fusco et al., 2001), Australian (Barnett et al., 2005;
27 Petroeschevsky et al., 2001), Asian (Lee et al., 2006), and Latin American (Gouveia and Fletcher
28 2000) studies did not find statistically significant associations between ambient SO2
29 concentrations and hospitalizations for all respiratory causes among children.
September 2007 3-38 DRAFT-DO NOT QUOTE OR CITE
-------
1 In general, positive associations were observed between ambient 862 concentrations and
2 ED visits and asthma hospitalizations, particularly among children, in various epidemiologic
3 studies conducted in different study locations and during varying time periods.
4
5 Chronic Obstructive Pulmonary Disease
6 Relatively few studies have examined the association of ED visits and hospitalizations for
7 COPD and ambient SO2 levels, and very little evidence exists for an association. Only three
8 studies reported positive and statistically significant results for COPD and SO2, and all three of
9 these studies included asthma in their diagnostic definition of COPD (Anderson et al., 2001;
10 Moolgavkar 2003; Sunyer et al., 2003b). Anderson et al. (2001) reported a 12% (95% CI: 5.0,
11 20.0) increased risk per 10-ppb increase in 24-h average SO2 among children, while Moolgavkar
12 (2003) and Sunyer et al. (2003b) found a 5 and 2% increased risk per 10-ppb increase in 24-h
13 average SO2 among older adults populations, respectively. All of the other studies examining
14 this outcome reported null results (Atkinson et al., 1999a; Burnett et al., 1999; Michaud et al.,
15 2004; Peel et al., 2005; Tenias et al., 2002).
16 Overall, this limited evidence does not support a relationship between ED visits and
17 hospitalizations for COPD and ambient SO2 levels.
18
19 Respiratory Diseases Other than Asthma or COPD
20 Emergency visits or hospital admissions for respiratory diseases include upper respiratory
21 infections (URIs), pneumonia, bronchitis, allergic rhinitis, and lower respiratory disease (LRD).
22 There are limited studies with mixed results for URIs (Burnett et al., 1999; Hajat et al., 2002; Lin
23 et al., 2005; Peel et al., 2005), pneumonia (Barnett et al., 2005; Moolgavkar et al., 1997; Peel
24 et al., 2005), bronchitis (Barnett et al., 2005; Michaud et al., 2004), and allergic rhinitis (Hajat
25 et al., 2001; Villeneuve et al., 2006). The evidence for an association between SO2 levels and
26 ED visits for LRD, though limited, is suggestive of an effect. All of the studies that
27 characterized this relationship found a positive and statistically significant increase in risk
28 associated with increases in SO2 (Farhat et al. 2005, Martins et al., 2002; Lin et al., 1999; Hajat
29 et al., 1999; Atkinson et al., 1999a). Increased risks ranging from 3 to 33% per 10-ppb increase
30 in 24-h average SO2 were reported in these studies.
31 In summary, there were limited studies providing mixed results for many of the health
32 outcomes other than asthma and COPD, making it difficult draw conclusions about the effects of
September 2007 3-39 DRAFT-DO NOT QUOTE OR CITE
-------
1 SO2 on these diseases. Limited evidence does exist to support a suggestive association between
2 ambient SC>2 levels and ED visits for LRD.
3
4 Potential Confounding by Copollutants
5 Multipollutant regression analyses indicated that SO2 risk estimates for respiratory ED
6 visits and hospitalizations, in general, were not sensitive to the inclusion of copollutants,
7 including O3 (Anderson et al., 1998; Hajat et al., 1999; Yang et al., 2003a, 2005), PM (Lin et al.,
8 2003a, 2005; Hagen et al., 2000; Schwartz, 1995), CO (Farhat et al., 2005), and NO2 (Anderson
9 et al., 1998; Lin et al., 2004a; Sunyer et al., 1997). There is limited evidence that the inclusion of
10 benzene in copollutant models attenuates 862 risk estimates (Hagen et al., 2000; Thompson
11 et al., 2001). Figure 3.1-11 presents 862 risk estimates with and without adjustment for various
12 copollutants, with a focus on PM and NC>2 as these pollutants tend to be moderately to highly
13 correlated with SO2 and have known respiratory health effects. Although the studies show that
14 copollutant adjustment had varying degrees of influence on the SC>2 effect estimates, among the
15 studies with tighter confidence intervals (an indicator of study power), the effect of SC>2 on
16 respiratory health outcomes appears to be generally robust and independent of the effects of
17 ambient particles or other gaseous copollutants.
18
19 Seasonal Effects ofSO2
20 The results of several studies (Anderson et al., 1998; Hajat et al., 1999; Schouten et al.,
21 1996; Spix et al., 1998; Wong et al., 1999) have demonstrated a greater increase in ED visits and
22 hospitalizations for respiratory illnesses during the summer months despite the fact that the
23 average concentrations for SC>2 in some of the areas were greater in the winter months (Anderson
24 et al., 1998; Schouten et al., 1996; Wong et al., 1999). In contrast, some studies found the
25 associations between ED visits and hospital admission and respiratory disease with similar
26 increases in SC>2 to be greater in winter than in summer months (Vigotti et al. 1996; Walters
27 et al., 1994). Additional studies were unable to discern a seasonal difference in ED visits and
28 hospitalizations for respiratory causes (Castellsague et al., 1995; Tenias et al., 1998; Wong et al.,
29 2002). These effects were not consistent across age groups. Warmer months were more likely to
30 show evidence of an association with adverse respiratory outcomes in children, while older
31 adults appeared to be more likely to be affected during the cooler months. These seasonal
32 associations remain somewhat uncertain and require additional investigation.
September 2007 3-40 DRAFT-DO NOT QUOTE OR CITE
-------
Relative Risk
0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4
Reference
Schwartz (1995)
Schwartz (1995)
Yangetal.(2003a)
Llorcaetal.(2005)
'Lin etal. (1999)
Linetal.(2003a)
Sunyer et al.
(1997)
Anderson et al.
(1998)
Anderson et al.
(1998)
"Thompson et al.
(2001)
•Galan etal. (2003)
Location Age
New Haven, 65+
CT
Tacoma, 65+
WA
Vancouver, <3
BC, Canada
65+
Torrelavega, All
Spain
Sao Paulo, <13
Brazil
Toronto, ON, 6-12
Canada (males)
6-12
(females)
Europe 0-14
London, 0-14
England
London, All
England
Belfast, <18
Ireland
Madrid, All
Spain
Lag Pollutants
0-1 S02 _
so2+PM10 _
0-1 S02 _
so2+PM10 _
2 S02 _
S02+COH+N02+03+CO
0 S02 _
S02+COH+N02+03+CO _
Not S02 _
reported S02+TSP+N02+NO+H2S _
0-5 S02 _
S02+PM10+N02+03+CO _
0-2 S02 _
S02+PM25+PM10.25 _
0-2 S02 _
S02+PM25+PMm25 _
1 S02 _
0-1 S02+BS _
1 S02+N02 _
1 S02 _
S02+BS _
S02+N02 _
1 S02 _
S02+BS _
S02+N02 _
0-1 S02 _
S02+PM10+Benzene _
0 S02 _
S02+PM10 _
i i
^
i i i
[All Respiratory
•o
—
» —
Asthma
«•
•0-
0-
-*•
0
t>
O' [ II + *
o Multipolkilanl
Figure 3.1-11.
Relative risks (95% CI) of SOi-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 average SOi concentrations or 40-ppb
increase in 1-h max
September 2007
3-41
DRAFT-DO NOT QUOTE OR CITE
-------
1 Summary of ED Visits and Hospitalizations for Respiratory Diseases
2 A large number of epidemiologic studies provide evidence of positive, but not always
3 statistically significant, associations between ambient SO2 concentrations and ED visits and
4 hospitalizations for all respiratory causes and asthma, particularly among children and older
5 adults. These findings are generally robust when additional copollutants are included in the
6 model. Biologic plausibility for these findings of increased ED visits and hospitalizations is
7 found in the epidemiologic and human clinical studies that observed increased respiratory
8 symptoms and decreased lung function, and the animal toxicological studies that observed SCV
9 induced altered lung host defenses. Season may modify the effect of 862 on ED visits and
10 hospitalizations for children in warmer months, and for older adults in cooler months.
11
12 3.1.1.7 Integration of Respiratory Effects Associated with Short-Term SOi Exposure
13 The previous reviews examining adverse health effects associated with short-term
14 exposures of SO2 have shown some biological plausibility and coherent evidence in the
15 epidemiological, human clinical, and animal toxicological studies completed to that time for a
16 limited number of respiratory effects. New studies of associations between SC>2 exposure and
17 respiratory symptoms, lung function, airway inflammation, AHR, lung host defenses, and ED
18 visits/hospitalizations have added modestly to this evidence base.
19 Respiratory symptoms. Two important new multicity studies (Mortimer et al., 2002;
20 Schildcrout et al., 2006) and several other studies (e.g., Delfmo et al., 2003; Neas et al., 1995)
21 have shown an association between short-term (24-h average) ambient SO2 concentrations and
22 respiratory symptoms in children. However, some other studies (e.g., Hoek and Brunekreef,
23 1993; Romieu et al., 1996) found no consistent association. Several new studies (e.g.,
24 Desqueyroux et al., 2002a,b; van der Zee et al., 2000) found no association between SC>2 levels
25 and respiratory symptoms in adults. These findings suggest supportive evidence for an
26 association between short-term (24-h average) exposure to ambient 862 exposure and respiratory
27 symptoms in children, particularly those with asthma, but not in adults. Evidence from the
28 previous review along with a limited number of new human clinical studies indicate increased
29 respiratory symptoms with peak (5-15 min) SO2 exposures as low as 0.5 ppm in asthmatic
30 subjects.
31 Lung function. Epidemiological studies do not provide strong evidence of associations
32 between short-term (24-h average) ambient SC>2 exposures and lung function in either children
September 2007 3-42 DRAFT-DO NOT QUOTE OR CITE
-------
1 (e.g., Mortimer et al., 2002; Roemer et al., 1998) or adults (e.g., Peters et al., 1996; Taggart et al.,
2 1996). Though several other studies reported positive findings, the mixed results and correlation
3 between SO2 levels and other ambient copollutants suggests a lack of independent effects on
4 lung function. In human clinical studies of lung function in healthy resting adults, a few studies
5 reported effects at 1 ppm, but most effects were observed at concentrations of >5 ppm. In
6 asthmatic adults, significant bronchoconstriction and increases in sRaw have been observed with
7 5- to 15-min peak (5-15 min) exposures to < 1-ppm SC>2, with some studies reporting a
8 bronchoconstrictive response to 862 within minutes of the start of exposure (Balmes et al., 1987;
9 Horstman et al., 1988). Increasing SC>2 levels from 0 to 0.5 ppm has been shown to have a
10 greater effect on sRaw and FEVi than increasing the level of exercise (Gong et al., 1995).
11 Moderate to severe asthmatics have greater exercise-induced sRaw increases and FEVi
12 decrements compared to normal and mild asthmatics; however, respiratory response with
13 increasing SC>2 concentration has not been shown to differ significantly between mild and
14 moderate/severe asthmatics (Linn et al., 1987). Lung function has been shown to be unaffected
15 by SC>2 exposures up to 0.8 ppm in individuals with COPD (Linn et al., 1985). Thus, the
16 observations of increased bronchoconstriction and airway resistance in human clinical studies
17 provide biological plausibility for 862 effects with peak exposure.
18 Airway inflammation. Only one epidemiological study (Adamkiewicz et al., 2004)
19 evaluated inflammation, as indexed by eNO, and found no association with 862 exposure. One
20 human clinical study observed increased markers of inflammation (i.e., increased macrophages,
21 lymphocytes, mast cells), but only at a concentration of 8-ppm SO2 in healthy adults (Sandstrom
22 et al., 1989). A study at more environmentally relevant levels (0.2 ppm) found no effects in
23 either healthy or asthmatic adults (Tunnicliffe et al., 2003). One animal study found increases in
24 inflammatory cytokines at 5.35 ppm but may not be relevant due to the inherent limitations of
25 high-concentration studies. Thus, the limited epidemiological, human clinical, and toxicological
26 evidence does not suggest that exposure to SO2 at environmentally relevant concentrations is
27 associated with inflammation in the airways. However, studies of other ambient pollutants
28 indicate that influx of macrophages and other inflammatory cells, with the related release of
29 inflammatory cytokines, is a common mechanism of injury.
30 Airway hyperresponsiveness. Only a limited number of epidemiological studies have
31 found an association between SO2 exposure and AHR. S0yseth et al. (1995) observed an
September 2007 3-43 DRAFT-DO NOT QUOTE OR CITE
-------
1 association between low (8 ppb) ambient 862 levels and eosinophil numbers in atopic children.
2 Taggart et al. (1996) found a marginal association between 40-ppb 862 concentrations and
3 decreased responses to MCh challenge in adult asthmatics. Boezen et al. (1999, 2005) reported
4 complex associations between SC>2 concentrations, BHR, and serum IgE levels in both children
5 and adults. As with other respiratory endpoints, a limited toxicological database provides some
6 biological plausibility for these findings. Bronchial responses were not observed in rabbits at
7 5-ppm SO2 (Douglas et al., 1994) or in dogs at 10-ppm SO2 (Lewis and Kirchner, 1984).
8 Ovalbumin-sensitized guinea pigs demonstrated increased bronchial obstruction following
9 exposure to 0.1-ppm 862 (Park et al., 2001; Riedel et al., 1988). Guinea pigs, as a species, are
10 typically more sensitive to air pollution than other laboratory animals (U.S. Environmental
11 Protection Agency, 2006d) and, thus, may provide a better model for characterizing the effects of
12 air pollutants on AHR. The finding of increased pulmonary resistance in this species is in
13 concordance with the limited epidemiological findings of SC>2-induced AHR.
14 Lung host defenses. Two epidemiological studies (Park et al., 2002; Ponka, 1990)
15 provide limited evidence of an association between school absences due to respiratory illness and
16 ambient 862. Scant animal evidence, typically at levels much higher than ambient, provides
17 weak biological plausibility for these epidemiological findings. SCVinduced modulation of
18 clearance and macrophage function were found in some subchronic and chronic studies but does
19 little to inform the mechanism(s) of action occurring in humans with short-term exposures.
20 ED visits/hospitalizations. Epidemiological studies provide suggestive evidence for an
21 association between ambient SC>2 levels and ED visits and hospitalizations for all respiratory
22 diseases, particularly among children and older adults (65+ years of age). A modest association
23 between ambient SC>2 and ED visits and hospitalizations for asthma particularly among children
24 <14 years old is also suggested. No relationship is apparent in the limited number of studies
25 evaluating ED visits and hospitalizations for COPD or other respiratory diseases, though there is
26 a somewhat suggestive association between ambient 862 levels and ED visits for LRD. Overall,
27 SC>2 risk estimates were not sensitive to the inclusion of copollutants, including PM, 63, CO, and
28 NO2, indicating that the observed effects of SO2 on respiratory endpoints is independent of the
29 effects of other ambient air pollutants. Biologic plausibility for these findings of increased ED
30 visits and hospitalizations is found in the epidemiologic and human clinical studies that observed
September 2007 3-44 DRAFT-DO NOT QUOTE OR CITE
-------
1 increased respiratory symptoms and decreased lung function, and the animal toxicological
2 studies that observed SCVinduced altered lung host defenses.
3
4 3.1.2 Cardiovascular Effects Associated with Short-Term SO2 Exposure
5 The studies reviewed in the 1982 AQCD primarily investigated respiratory health
6 outcomes. No epidemiological studies linking exposure to 862 with cardiovascular
7 physiological endpoints or CVD ED visits or hospital admissions were examined in the last
8 review. There were also key human clinical and animal toxicological studies available at the last
9 review to address effects of SO2 exposure on the cardiovascular system. The only report from a
10 study in dogs exposed to air pollutant mixtures (SC>2 + sulfuric acid [H^SC^], with or without
11 nonirradiated or irradiated auto exhaust) reported no changes in cardiovascular function at the
12 end of 3 years of exposure and 3 years after exposure.
13 A few recent animal toxicological studies have investigated the potential effects of 862
14 exposure on physiological and biochemical parameters of cardiovascular effects and reported
15 oxidation (Meng et al., 2003a) and glutathione (GSH) depletion (Langley-Evans et al., 1996;
16 Meng et al., 2003a; Wu and Meng, 2003) in the hearts of rodents (see Annex Table AX4-5).
17 Several recent epidemiological studies also have examined the association between air pollution
18 and cardiovascular effects, including increased heart rate (HR), reduced heart rate variability
19 (HRV), incidence of ventricular arrhythmias, changes in blood pressure, incidence of myocardial
20 infarctions (MI), and ED visits and hospitalizations due to cardiovascular causes. The results of
21 these cardiovascular studies are summarized in Annex Tables AX5-3 and AX5-4.
22
23 3.1.2.1 HR and HRV
24 HRV is generally determined by analyses of time (e.g., standard deviation of normal R-R
25 intervals [SDNN]) and frequency domains (e.g., low frequency [LF] / high frequency [HF] ratio
26 by power spectral analysis, reflecting autonomic balance) measured during 24 h of
27 electrocardiography (ECG). Brook et al. (2004) state that FIRV, resting FIR, and blood pressure
28 are modulated by a balance between the two determinants of autonomic tone (the sympathetic
29 and parasympathetic nervous systems). They note that decreased HRV predicts an increased risk
30 of cardiovascular morbidity and mortality in older adults and those with significant heart disease.
31 Liao et al. (2004) investigated short-term associations between ambient pollutants and
32 cardiac autonomic control from the fourth cohort examination (1996 through 1998) of the
September 2007 3-45 DRAFT-DO NOT QUOTE OR CITE
-------
1 population-based Atherosclerosis Risk in Communities (ARIC) study using a cross-sectional
2 study design. Men and women aged 45 to 64 years (n = 6,784) from three U.S. study centers in
3 North Carolina, Minnesota, and Mississippi were examined. Resting, supine, 5-min beat-to-beat
4 R-R interval data were collected. The mean 24-h average SO2 level measured 1 day prior to the
5 HRV measurement was 4 ppb (SD 4). In addition to 862, the potential effects of PMio, Os, CO,
6 and NO2 were evaluated. Previous-day SC>2 concentrations were positively associated with HR
7 and inversely associated with SDNN and LF power. Consistently more pronounced associations
8 were suggested between 862 and HRV among persons with a history of coronary heart disease.
9 Significant associations with HRV indices also were observed for PMio and the other gaseous
10 pollutants. When the regression coefficients for each individual pollutant model were compared,
11 the effects of PMio on HRV were considerably larger than the effects for the gaseous pollutants,
12 including SC>2. No multipollutant analyses were conducted.
13 Gold et al. (2000; reanalysis Gold et al., 2003) examined the effect of short-term changes
14 in air pollution on HRV in a panel study of 21 older adults (aged 53 to 87 years) in Boston, MA.
15 The study participants were observed up to 12 times from June to September 1997. The mean
16 24-h average 862 concentration was 3.2 ppb (range: 0,12.6). The 24-h average 862
17 concentration was associated with decreased HR in the first 5-min rest period, but not in the
18 overall 25-min study protocol. The effect estimate for 862 slightly diminished but remained
19 marginally significant in a two-pollutant model with PM2.5. The inverse association between
20 SO2 and HR observed in this study are in contrast to the SO2-related increases in HR observed by
21 Liao et al. (2004) and Peters et al. (1999). No associations were observed between HRV and
22 SC>2. The strongest associations with HRV were observed for PM2.5 and Os.
23 Another study of air pollutants and HRV was conducted in Boston, MA on 497 men from
24 the VA Normative Aging Study (NAS) (Park et al., 2005). The best 4-consecutive-min interval
25 from a 7-min sample was used for the HRV calculations. For the exposure variable, 4-, 24-, and
26 48-h moving averages matched on the time of the ECG measurement for each subject were
27 considered. The mean 24-h average 862 concentration was 4.9 ppb (range: 0.95, 24.7).
28 Associations with measures of HRV were reported for PM2.s and O3, but not with SO2 for any of
29 the averaging periods. In another study conducted in Boston, MA, Schwartz et al. (2005) found
30 significant effects of increases in PM2.5 on measures of HRV, while no associations with SC>2
31 were observed. Other studies have examined the relationship of SC>2 with HRV (Chan et al.,
September 2007 3-46 DRAFT-DO NOT QUOTE OR CITE
-------
1 2005; de Paula Santos et al., 2005; Holguin et al., 2003; Luttmann-Gibson et al., 2006). Most of
2 these studies, with the exception of de Paula Santos et al. (2005), did not observe associations
3 with SO2.
4 A limited number of human clinical studies examined the effect of SC>2 on HRV. During
5 a controlled exposure of 12 healthy subjects and 12 subjects with asthma to 0.2-ppm SO2 for 1 h
6 under resting conditions, Tunnicliffe et al. (2001) reported that HF power, LF power, and total
7 power were higher with SC>2 exposures compared to air exposure in the healthy subjects, but that
8 these indices were reduced during 862 exposure in the subjects with asthma. The LF/HF ratios
9 were unchanged in both groups. The authors postulated that these results suggest that there are
10 two autonomic pathways for SCVmediated bronchoconstriction. The investigators proposed that
11 in healthy subjects, the dominant pathway was via the rapidly adapting receptor/C-fiber route,
12 which results in a central nervous system reflex with an increase in vagal tone. In the asthmatic
13 subjects, proximal airway narrowing was proposed as the dominant response, possibly through
14 neurogenic inflammation. This likely causes a compensatory central nervous system-mediated
15 reduction in vagal tone, resulting in bronchodilation of the distal airways. While there were no
16 detectable changes in symptoms or lung function in either of the groups, this study suggests that
17 exposure to 862 can provoke autonomic responses at these low levels (0.2 ppm).
18 In a similar study, Routledge et al. (2006) exposed patients with stable angina as well as
19 healthy subjects to 50-|ig/m3 carbon particles and to 0.2-ppm 862, alone and in combination, for
20 1 h under resting conditions. HRV, C-reactive protein, and markers of coagulation markers were
21 measured. These authors reported that in the healthy subjects, SC>2 exposure was associated with
22 a decrease in HRV markers of cardiac vagal control 4 h after exposure. However, it should be
23 noted that there was no apparent difference in the absolute value of the root mean square of
24 successive RR interval differences (r-MSSD) at 4 h postexposure between the control, SC>2,
25 carbon, and carbon/SC>2 groups. The significant difference reported in the change in r-MSSD
26 from baseline to 4 h postexposure with 862 appears to be due to a higher baseline value of r-
27 MSSD preceding the 862 exposure compared to the baseline value of r-MSSD preceding the air
28 exposure. There were no changes in HRV among the patients with stable angina. It was noted
29 by the authors that this lack of response in the heart patients may be due to a drug treatment
30 effect rather than decreased susceptibility; a large portion of the angina patients were taking P-
31 blockers, which are known to increase indices of cardiac vagal control.
September 2007 3-47 DRAFT-DO NOT QUOTE OR CITE
-------
1 In the limited number of epidemiological and human clinical studies that examined a
2 possible effect of 862 on HRV, there are some suggestive findings; however, the overall
3 evidence that SO2 affects cardiac autonomic control is weak and inconsistent.
4
5 3.1.2.2 Repolarization Changes
6 In addition to the role played by the autonomic nervous system in arrhythmogenic
7 conditions, myocardial vulnerability and repolarization abnormalities are believed to be key
8 factors contributing to the mechanism of such diseases. Measures of repolarization include QT
9 duration, T-wave complexity, variability of T-wave complexity, and T-wave amplitude.
10 Henneberger et al. (2005) examined the association of repolarization parameters with air
11 pollutants in patients with preexisting coronary heart disease (n = 56, all males) in East
12 Germany. The patients were examined repeatedly once every 2 weeks for 6 months, for a total
13 of 12 ECG recordings. The mean 24-h average SO2 concentration was 4.1 |ig/m3 (2 ppb [range:
14 1, 4]). Ambient SC>2 concentrations during the 24-h preceding the ECG were associated with the
15 QT interval duration, but not with any other repolarization parameters. Stronger associations
16 were observed between PM indices and QT interval duration, T-wave amplitude, and T-wave
17 complexity.
18 Two in vitro studies (Me and Meng, 2005, 2006) conducted with a 1:3 molarmolar
19 mixture of the 862 derivatives bisulfite (HSCV) and sulfite (SOs2 ) demonstrated effects of a
20 10-|im bisulfite:sulfite mixture on sodium and L-type calcium currents (which included changes
21 in inactivation and/or activation, recovery from inactivation, and inactivation/activation time
22 constants) in ventricular myocytes. These in vitro observations suggest a potential role for
23 L-type calcium current in cardiac injury following SC>2 exposure; however, in vivo
24 cardiovascular effects were observed only at high SC>2 concentrations (10 ppm and higher).
25 Additional epidemiological and toxicological studies are necessary to evaluate the evidence of an
26 association between 862 and repolarization changes.
27
28 3.1.2.3 Cardiac Arrhythmias
29 In a panel study of 100 patients with implanted cardioverter defibrillators (ICDs) in
30 Eastern Massachusetts, Peters et al (2000) tested the hypothesis that patients with ICDs would
31 experience life-threatening arrhythmias after an air pollution episode. The mean 24-h average
32 SO2 concentration measured at two sites in Boston during the study period was 7 ppb (5th-95th
September 2007 3-48 DRAFT-DO NOT QUOTE OR CITE
-------
1 percentile: 1,19). ICDs monitor ECG abnormalities and treat ventricular fibrillation or
2 ventricular tachycardias by administering shock therapy to restore the normal cardiac rhythm.
3 The ICD device also stores information on each tachyarrhythmia and shock. There was no
4 association between SO2 and defibrillator discharges in the 33 subjects who had any defibrillator
5 discharges during the follow-up period or in the 6 subjects who had at least 10 discharges. There
6 was some evidence that NO2 was associated with increased defibrillatory interventions in the
7 subjects with any defibrillator discharges. Among the patients with at least 10 events, NO2, CO,
8 and PM2.5 was found to be associated with defibrillator discharges.
9 In a follow-up study designed to confirm the findings of Peters et al. (2000), Dockery
10 et al. (2005) used a larger sample of ICD patients in Boston (n = 203) with a longer follow-up
11 period. The median concentration of 48-h average SO2 averaged across multiple sites in Boston
12 was 4.9 ppb (IQR 4.1). No significant associations were found between ventricular arrhythmic
13 episode days and any of the air pollutants. However, when the analysis was stratified by recent
14 arrhythmias (i.e., within 3 days), there was evidence of an increased risk of ventricular
15 arrhythmia with SO2, PM2.5, black carbon, NO2, and CO. Since PM2.5, black carbon, NO2, and
16 CO were correlated with each other and SO2, the authors noted that differentiating the
17 independent effects of the pollutants would be difficult. A case-crossover analysis of the same
18 data by Rich et al. (2005) also observed associations with 48-h average SO2, but the SO2 effect
19 was not found to be robust to adjustment by PM2.5. In a similar study conducted in St. Louis,
20 MO, an increased risk was associated with SO2 concentrations in the 24 h prior to an arrhythmia,
21 but not with PM2 5 and Os (Rich et al., 2006). In this study, none of the other measured
22 pollutants (PM, elemental carbon, Os, CO, NO2) were correlated with SO2. The authors
23 suggested that the different effects observed in St. Louis and Boston may be due to differences in
24 the source or mix of air pollutants in these cities.
25 Additional studies have examined the relationship of SO2 with arrhythmias in Vancouver,
26 Canada (Rich et al., 2004; Vedal et al., 2004) and observed associations at very low ambient SO2
27 concentrations (mean 24-h average SO2 of-2.5 ppb with a maximum of 8.1 ppb). Vedal et al.
28 (2004) stated that of all pollutants examined, the only one with somewhat consistent positive
29 associations with arrhythmia events was SO2. In season-stratified analyses, SO2 was positively
30 associated with arrhythmias in the winter, while in the summer the association was negative. On
September 2007 3-49 DRAFT-DO NOT QUOTE OR CITE
-------
1 the other hand, in the Rich et al. (2004) study, positive associations were observed in the summer
2 but not in the winter. The authors stated that it was difficult to interpret these findings.
3 One toxicological study examined the effects of PM, ultrafine carbon, and SO2 on
4 spontaneous arrhythmia frequency in 18-month-old rats (Nadziejko et al., 2004). The rats were
5 exposed to 1-ppm SO2 for 4 h. No significant change in the frequency of spontaneous
6 arrhythmias was found with SC>2 and ultrafine carbon exposure. However, rats exposed to
7 concentrated ambient PM had a significantly greater increase in the frequency of delayed beats
8 than rats exposed to air.
9 Collectively, the epidemiological evidence for an association between short-term
10 exposure to 862 and arrhythmias is inconsistent. The limited toxicological evidence did not
11 provide biological plausiblity of an effect of SO2 on arrhythmias.
12
13 3.1.2.4 Blood Pressure
14 Ibald-Mulli et al. (2001) examined the association between blood pressure and SC>2 using
15 survey data from the MONICA (Monitoring Trends and Determinants in Cardiovascular
16 Disease) Project. Blood pressure measurements were taken from 2,607 men and women. The
17 mean 24-h average SO2 concentration was 60.2 |ig/m3 (23 ppb [range: 5,91]). An increase in
18 systolic blood pressure was associated with 24-h average 862 and TSP. However, in a two-
19 pollutant model with TSP, the effect of 862 on blood pressure was substantially reduced and
20 became nonsignificant while the effect of TSP was robust.
21 In a study by de Paula Santos et al. (2005), changes in blood pressure in association with
22 SO2 were investigated in vehicular traffic controllers (n = 48) aged 31 to 55 years living in Sao
23 Paulo, Brazil, where vehicles are the primary source of air pollution. The mean 24-h average
24 SC>2 level, measured at six different stations around the city, was 17.1 |ig/m3 (7 ppb [SD 3]).
25 Blood pressure was measured every 10 min when subjects were awake (6 a.m. to 11 p.m.) and
26 every 20 min during sleep (11 p.m. to 6 a.m.). Results indicated that SC>2, as well as CO, were
27 associated with increases in systolic and diastolic blood pressure. However, when a two-
28 pollutant model was used to test the robustness of the associations, only the CO effect remained
29 statistically significant.
30 Several animal toxicological studies examined the effect of SO2 on blood pressure.
31 Halinen et al. (2000a) examined blood pressure changes in guinea pigs that were exposed to 1-,
32 2.5-, or 5-ppm SO2 in cold, dry air while being hyperventilated to simulate exercise. Animals
September 2007 3-50 DRAFT-DO NOT QUOTE OR CITE
-------
1 received 10-min exposures to each 862 concentration that were separated by 15-min exposures
2 to clean warm, humid air. A transient increase in blood pressure was observed during exposure
3 to 5-ppm SO2 in cold, dry air. In a second study (Halinen et al., 2000b), guinea pigs were
4 exposed to cold, dry air alone or 1-ppm SC>2 in cold, dry air for 60 min while being
5 hyperventilated. The study reported similar increases in blood pressure and HR with exposure to
6 cold, dry air or cold, dry air plus SC>2. The increase in HR was gradual, while increases in blood
7 pressure generally occurred during the first 10 to 20 min of exposure. Similar effects were
8 observed with exposure to cold, dry air or 862 in cold, dry air, suggesting that effects were
9 associated with cold, dry air rather than 862. Opposite effects (a transient decrease in blood
10 pressure) was observed when rats were exposed to a higher dose (10-ppm 862) in air that was
11 presumably at room temperature for 3 days (Meng et al., 2003b).
12 Collectively, the limited epidemiological and toxicological evidence does not suggest that
13 short-term exposure to SC>2 has effects on blood pressure.
14
15 3.1.2.5 Blood Markers of Cardiovascular Risk
16 Folsom et al. (1997) demonstrated that elevated levels of fibrinogen, white blood cell
17 count, factor VIII coagulant activity (factor VIII-C), and von Willebrand factor were associated
18 with risk of CVD. Schwartz (2001) investigated the association between various blood markers
19 of cardiovascular risk and air pollution among subjects in the Third National Health and
20 Nutrition Examination Survey (NHANES III) in the United States conducted between 1989 and
21 1994 across 44 counties. The NHANES III is a random sample of the U.S. population with
22 oversampling for minorities (30% of NHANES sample) and the elderly (20% of the sample).
23 The mean SC>2 concentration was 17.2 ppb (IQR 17) across the 25 counties where data were
24 available. This study looked at fibrinogen levels, platelet counts, and white blood cell counts.
25 After controlling for age, ethnicity, gender, body mass index, and smoking status and number of
26 cigarettes per day, SC>2 was found to be positively associated with white blood cell counts. PMi0
27 was associated with all blood markers. In two-pollutant models, PMio remained a significant
28 predictor of white blood cell counts after controlling for 862, but not vice versa.
29 A recent cross-sectional study by Liao et al. (2005) investigated the effects of air
30 pollution on plasma hemostatic and inflammatory markers in the ARIC study (n = 10,208). The
31 authors hypothesized that short-term exposure to air pollutants was associated with increased
32 levels of inflammatory markers and lower levels of albumin, as serum albumin is inversely
September 2007 3-51 DRAFT-DO NOT QUOTE OR CITE
-------
1 associated with inflammation. The mean 24-h average 862 concentration was 5 ppb (SD 4).
2 Significant curvilinear relationships were observed between 862 and factor VIII-C, white blood
3 cell counts, and serum albumin. The authors noted that since no biological explanation could be
4 offered for the "U"-shaped curve between SC>2 and factor VIII-C and the "inverse U"-shape
5 between SC>2 and albumin, generalization of the association should be exercised with caution.
6 No associations were observed between SC>2 and fibrinogen or von Willebrand factor.
7 In another large cross-sectional study of 7,205 office workers in London, Pekkanen et al.
8 (2000) examined the association between plasma fibrinogen and ambient air pollutants. The
9 mean 24-h average 862 was 23.2 |ig/m3 (9 ppb [10th-90th percentile: 5, 19]). Associations with
10 fibrinogen were observed for all pollutants examined, either in all-year or summer-only analyses,
11 except for SO2 and O3. Taken together, results from the limited number of studies do not suggest
12 that SC>2 is associated with various blood markers of cardiovascular risk.
13
14 3.1.2.6 Acute Myocardial Infarctions
15 The association between air pollution and the incidence of MI was examined in a small
16 number of studies. As part of the Determinants of Myocardial Infarction Onset Study, Peters
17 et al. (2001) examined 772 patients with MI living in greater Boston, MA. A case-crossover
18 design was used to assess changes in the risk of acute MI after exposure to potential triggers.
19 The mean 24-h average 862 was 7 ppb (range: 1, 20) during the study period. Similarly, the
20 mean 1-h average 862 was 7 ppb (range: 0, 23). In an analysis that considered both the 2-h
21 average (between 60 and 180 min before the onset of symptoms) and 24-h average (between 24
22 and 48 h before the onset) concentrations jointly, the study found no significant association
23 between risk of MI and SC>2. Of all the pollutants considered, only PM2.5 and PMio were found
24 to be associated with an increased risk of MI.
25 In the MONICA Project, the effect of air pollution on acute MI was studied in Toulouse,
26 France, using a case-crossover study design (Ruidavets et al., 2005). The mean 24-h average
27 SO2 level was 8.3 |ig/m3 (3 ppb [5th-95th percentile: 1, 5]). A total of 399 cases of acute MI
28 were recorded during the study period. Os, but not SO2 nor NO2, was found to be associated
29 with the incidence of acute MI. Exposure to PM was not considered in this study.
30 Only a limited number of studies examined the association between ambient SO2
31 concentrations and incidence of acute MI. These studies provide no evidence that exposure to
32 SO2 increases the risk of MI.
September 2007 3-52 DRAFT-DO NOT QUOTE OR CITE
-------
1 3.1.2.7 ED Visits and Hospitalizations for CVD
2 The current review includes more than 30 studies that address the effect of sulfur oxides
3 (SOX) exposure on ED visits or hospitalizations for CVD. Cases of CVD are typically identified
4 using ICD codes recorded on hospital discharge records. However, counts of hospital or ED
5 admissions are also used. Studies of ED visits include cases that may be less severe than those
6 requiring hospitalization and may be subject to greater misclassification compared to studies that
7 rely on confirmed doctors diagnoses coded on discharge records. Studies of hospital admissions
8 and ED visits are clearly distinguished on figures and in the Annex Tables AX5-4.
9
10 All CVD
11 The disease grouping "All CVD" typically includes all diseases of the circulatory system
12 (e.g., heart diseases and cerebrovascular diseases, ICD9 Codes 390-459). A summary of the
13 results are presented in Figure 3.1-12.
14 In a study of 11 cities in Spain, an increase of 3.6% (95% CI: 0.6, 6.7) per 10-ppb
15 increase in 24-h average SO2 at a 0-1 day lag was observed for all CVD admissions (Ballester
16 et al., 2006). The mean 24-h average SC>2 level in the cities studied was 6.6 ppb. In addition,
17 time-series data linking 862 with hospital admissions for CVD in three metropolitan areas in the
18 United States (i.e., Cook, Maricopa, Los Angeles Counties) was conducted (Moolgavkar, 2000;
19 reanalysis, Moolgavkar, 2003). A 13.7% (95% CI: 11.3, 16.1) increase in admissions per
20 10-ppb increase in 24-h average 862 at lag 0 day, using Generalized Linear Model(s) (GLM) and
21 natural splines to adjust for temporal trends, was observed among older adults (65+ years) in Los
22 Angeles County. The median 24-h average SC>2 level for Los Angeles County was 2 ppb during
23 the study period. Results for Maricopa and Cook counties were not presented in the reanalysis.
24 However, in previous GAM analyses, increases of 4.1% (95% CI: 2.7, 5.3) and 7.5% (95% CI:
25 4.1, 10.8) were reported for Cook and Maricopa Counties, respectively (Moolgavkar, 2000), per
26 10-ppb increase in 24-h average 862 level. The author indicates that the use of stringent
27 convergence criteria did not appreciably change results (but increased smoothing did diminish
28 effect estimates) (Moolgavkar, 2003).
29 Metzger et al. (2004) examined approximately 4.4 million hospital visits to 31 hospitals
30 from 1993 to 2000 in Atlanta, GA and reported null associations between SC>2 and ED visits for
31 all CVD. A 1.4% (95% CI: -1.5, 4.4) increase in admissions per 40-ppb increase in 1-h max
September 2007 3-53 DRAFT-DO NOT QUOTE OR CITE
-------
Reference
*Metzger etal. (2004)
Bailester et al. (2006)
Atkinson etal. (1999a)
Potoniecki el al. (1997)
Bailester etal. (2001)
Llorca et a). (2005)
Petroeschevsky et al. (2001)
Chang etal. (2005)
Chang et al. (2005)
Petroeschevsky et al. (2001)
Moolgavkar (2003)
Atkinson etal. (1999a)
•Jalaludin etal. (2006)
Petroeschevsky etal. (2001)
Location
Atlanta, GA
14 Spanish cities
London, England
London, England
Valencia, Spain
Torrelavega, Spain
Brisbane, Australia
Taipei, Taiwan
Taipei, Taiwan
Brisbane, Australia
Los Angeles, CA
London, England
Sydney, Australia
Brisbane, Australia
Lag
0-2
0-1
0
1
2
0
0
>2Q C, 0-2
<20 C, 0-2
0
0
0
0
1
All ages
— B-
1 5-b4 years
65+ years
1 1 1
0.7 0.8 0.9 1
f-
-B-
B
-i —
— i —
1 1 1 I
0 1.1 1.2 1.3 1.4
Figure 3.1-12.
Relative Risk
Relative risks (95% CI) of SOi-associated emergency department
visits and hospitalizations for all cardiovascular causes. Risk
estimates are standardized per 10-ppb increase in 24-h average SOi
concentrations or 40-ppb increase in 1-h max SOi. 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 SO2 level was observed. The median 1-h max 862 level in Atlanta during the study period was
2 11 ppb (10th-90th percentile: 2, 39).
3 Results from single-city studies in Europe, Australia, and Taiwan are inconsistent.
4 Atkinson et al. (1999a) reported a significant increase in CVD admissions in London (2.3%
5 [95% CI: 0.3, 4.3] per 10-ppb increase in 24-h average SO2), while Llorca et al. (2005) reported
6 a null association in Torrelavega, Spain. A time-series analysis conducted in Sydney, Australia,
7 reported an increase in all CVD admissions of 19.3% (95% CI: 3.3, 38) per 10-ppb increase in
8 24-h average SC>2 at lag 0 day among those 65+ years of age (Jalaludin et al., 2006). The mean
9 24-h average 862 level in Sydney during the study period was 1.07 ppb (IQR 0.75) (the authors'
September 2007
3-54
DRAFT-DO NOT QUOTE OR CITE
-------
1 estimates related the percent increase in admissions to an incremental increase in 862 equivalent
2 to the IQR [1.33%, 95% CI: 0.24, 2.43]). A study conducted in Brisbane reported a
3 nonsignificant increase of 3.8% (95% CI: -1.2, 9.1) at lag 1-day for all CVD per 10-ppb
4 increase in 24-h average SO2, among those 65+ years (Petroeschevsky et al., 2001). In a study
5 conducted in Taipei, Taiwan, Chang et al. (2005) reported a significant decrease in CVD
6 admissions of -11.5% (95% CI: -20.2, -1.8) per 10-ppb increase in 24-h average 862 at a lag
7 of 0 to 2 days, among all ages, when the temperature was greater than 20 °C. A nonsignificant
8 increase of 5.6% (95% CI: -12.4, 27.2) was reported for cooler days. The mean 24-average 862
9 level in Taiwan during the study period was 4.3 ppb.
10 Some studies have observed positive associations between ambient 862 concentrations
11 and ED visits and hospital admissions for all CVD, particularly among individuals 65+ years of
12 age. Given the limited number of studies that assessed potential confounding by copollutants for
13 this outcome, which is of concern given the moderate to strong correlation between SO2 and
14 various copollutants in most studies, and the lack of supportive data from panel/field studies and
15 human clinical studies on cardiovascular health effects, the collective evidence that ambient SO2
16 has an effect of CVD ED visits and hospitalizations is weak.
17
18 Specific Cardiac Diseases
19 Cardiac disease (ICD9 Codes 390-429) is defined to exclude diseases of the
20 cerebrovascular system and is further restricted in some studies to include only ischemic heart
21 disease (fflD, ICD9 Codes 410-414), dysrhythmia (ICD9 Code 427), congestive heart failure
22 (CHF, ICD9 Code 428) or MI (410).
23 In a study of seven European cities (Milan, Paris, Rome, London, Birmingham, the
24 Netherlands, and Stockholm), an increase of 1.9% (95% CI: 0.8, 2.9) per 10-ppb increase in 24-h
25 average SC>2 lagged 0-1 day, was observed for cardiac disease hospital admissions (Sunyer et al.,
26 2003b; used GAM with default convergence criteria). The mean 24-h average SC>2 level in the
27 cities studied was 5.2 ppb. Ballester et al. (2006) reported a 4.6% (95% CI: 1.3, 8.0) increased
28 risk of cardiac disease admissions per 10-ppb increase in 24-h average SC>2 at lag 0-1 day, pooled
29 across 14 Spanish cities. Adjustment for PMi0 and CO in two-pollutant models diminished the
30 effect estimate by approximately half.
September 2007 3-55 DRAFT-DO NOT QUOTE OR CITE
-------
1 In a time-series study of cardiac disease and 862 in Windsor, Ontario, similar results
2 were observed for those aged <65 years (4.5% [95% CI: -3.7, 14.1) per 40-ppb increase in 1-h
3 max SO2] and 65+ years (5.5% [95% CI: 0.0, 11.3]) (Fung et al., 2005). These results were
4 found to be generally robust to adjustment for PMio. The mean 1-h max SC>2 level in Windsor
5 during the study period was 27.5 ppb (range: 0, 129). Michaud et al. (2004) conducted a study
6 of hospital visits for cardiac disease in Hilo, HI, where volcanic eruptions contribute to ambient
7 SO2 levels. A-5.0% (95% CI: -13.5, 4.4) change in hospital visits for cardiac disease was
8 observed per 10-ppb increase in 862 (averaging time 12:00 p.m. to 6:00 a.m.). The mean daily
9 SC>2 level in Hilo during the study period was 1.97 ppb (range: 0, 108.5). In Sydney, Australia,
10 an increase in cardiac admissions among those 65+ years of age of 1.6% (95% CI: 0.33, 2.93)
11 was reported per 0.75-ppb increase (an IQR change) in 24-h average SO2 level (Jalaludin et al.,
12 2006). Standardized to a 10-ppb increase in 24-h average SC>2, the increased risk is 23.9% (95%
13 CI: 4.5,46.9). The mean 24-h average SC>2 level in Sydney during the study period was
14 1.07 ppb (range: 0.09,3.94). Llorca et al. (2005) reported a null association for cardiac disease
15 hospital admissions and SC>2 in Torrelavega, Spain.
16 Analyses restricted to diagnoses of IHD (Jalaludin et al., 2006; Lee et al., 2003a; Lin
17 et al., 2003b; Metzger et al., 2004; Peel et al., 2007), CHF (Koken et al., 2003; Metzger et al.,
18 2004; Peel et al., 2007; Wellenius et al., 2005a), dysrhythmia (Koken et al., 2003; Metzger et al.,
19 2004; Peel et al., 2007), MI (Koken et al., 2003; Lin et al., 2003b), and angina pectoris
20 (Hosseinpoor et al., 2005) were conducted. Two studies conducted in Atlanta, GA reported no
21 significant associations between SC>2 and admissions for specific cardiac outcomes (Metzger
22 et al. 2004; Peel et al. 2007). Metzger et al. observed null associations of 1-h max SC>2 with
23 IHD, CHF, and dysrhythmia. Using the same dataset, Peel et al. (2007) investigated effect
24 modification of CVD outcomes across comorbid disease status categories, including
25 hypertension, diabetes, COPD, dysrhythmia, and CHF. Authors observed no significant
26 associations for any cardiac disease outcome studied (i.e., IHD, CHF, dysrhythmia) with ambient
27 1-h max SO2 level in any comorbid disease category.
28 SC>2-associated increases in admissions for CHF, IHD, and dysrhythmia were reported in
29 a limited number of studies (Jalaludin et al., 2006; Koken et al., 2003; Wellenius et al., 2005a).
30 Results from other analyses of specific cardiac disease endpoints were null (Hosseinpoor et al.,
31 2005; Lee et al., 2003a; Lin et al., 2003b).
September 2007 3-56 DRAFT-DO NOT QUOTE OR CITE
-------
1 In conclusion, the strongest evidence comes from a large multicity study conducted in
2 Spain (Ballester et al., 2006) that observed statistically significant positive associations between
3 ambient SO2 and cardiac disease; however, the SO2 effect was found to diminish by half with
4 PMio and CO adjustment. Overall, findings on the relationship between ambient SO2 and
5 cardiac disease are generally inconsistent.
6
7 Cerebrovascular Disease and Stroke
8 Cerebrovascular diseases include diseases of the blood vessels supplying the brain (ICD9
9 Codes 430-438). Separate analyses for ischemic stroke (ICD9 434-436), hemorrhagic stroke
10 (ICD9 Codes 431-432), and transient ischemic attack (ICD9 435) are often conducted.
11 Positive findings were reported for ischemic stroke and SO2 in a study of nine U.S. cities
12 (Wellenius et al., 2005a). This study examined time-series data including more than 155,500
13 ischemic stroke hospitalizations between 1986 and 1999 in the cities of Birmingham, AL,
14 Chicago, IL, Cleveland, OH, Detroit, MI, New Haven, CT, Pittsburgh, PA, Salt Lake, UT, and
15 Seattle, WA. The median 24-h average SO2 level in these cities was 6.2 ppb (10th, 90th
16 percentile: 2.17,16.17). The study reported a 1.2% (95% CI: 0.1, 2.4) increase of ischemic
17 stroke hospitalizations per 10-ppb increase in 24-h average SO2 level at lag 0-2 days. Wellenius
18 et al. did not analyze multipollutant models, but the authors noted that other pollutants studied
19 (i.e., NO2, CO) were more strongly were associated with increased admissions for ischemic
20 stroke. In a study in Edmonton, Canada, Villeneuve et al. (2006) found significantly increased
21 risk of ischemic stroke during the warm season among older adults and a significant association
22 between transient ischemic attacks and SO2 among older adults in the warm season and all year.
23 The positive results were diminished in multipollutant models.
24 By contrast, Metzger et al. (2004) reported a null increase for all peripheral and
25 Cerebrovascular diseases of 0.2% (95% CI: -4.4, 5.0) per 40-ppb increase in 1-h max SO2. Peel
26 et al. (2007) also observed null results for this Atlanta population across comorbid disease status
27 categories. Similarly, Jalaludin et al. (2006) observed a null association between Cerebrovascular
28 admissions and SO2 in Sydney. Furthermore, primary intracerebral hemorrhage and ischemic
29 stroke were not found to be significantly associated with SO2 in a study of admissions records
30 from 63 hospitals in Taiwan (Tsai et al., 2003).
September 2007 3-57 DRAFT-DO NOT QUOTE OR CITE
-------
1 A limited number of studies have examined the effect of ambient 862 on cerebrovascular
2 disease and stroke. In general, findings relating ambient 862 level to these outcomes have been
3 inconsistent.
4
5 Potential Confounding by Copollutants
6 Studies of all CVD or cardiac diseases that report multipollutant results are summarized in
7 Figure 3.1-13. Overall, effects for all CVD, cardiac diseases, and specific cardiac outcomes
8 were diminished in multipollutant models (Ballester et al., 2001, 2006; Morris et al., 1995;
9 Wellenius et al., 2005b). In addition, Jalaludin et al. (2006) reported a 3% increase in CVD
10 hospital admissions per 0.75-ppb incremental change in 24-h average 862 in single-pollutant
11 models, which was reduced to null when CO was included. This study was not included in
12 Figure 3.1-13, because the range of 862 concentrations was far below the 10-ppb increment to
13 which other effect sizes were standardized. A study by Chang et al. (2005) examined the effect
14 of SO2 on all CVD hospitalizations by season and observed a nonsignificant negative association
15 in single-pollutant models for the cool season in Taiwan. After adjusting for NC>2, PMio, and CO
16 in two-pollutant models, this negative association strengthened and achieved significance. The
17 authors attributed this finding to possible collinearity problems between SO2 and copollutants.
18 Collectively, these results suggest that the effect of SO2 on cardiovascular ED visits and
19 hospitalizations is likely confounded by copollutant exposures.
20
21 3.1.3 Other Systemic Effects Associated with Short-Term SO2 Exposure
22 The effects of SO2 on the nervous system and other organ systems were not examined in
23 the previous review. The 1982 AQCD presented only one chronic exposure study (68 months),
24 in which dogs were exposed to a mixture of SO2 and H2SO4. This study reported no effects on
25 visual evoked brain potentials during or immediately after exposure to the SOX mixture. In the
26 past 25 years, an increased number of animal toxicological studies evaluated the effects of SO2
27 exposure on neurophysiological, biochemical, and neurobehavior as well as on other organ
28 systems in adult and developing animals. The most recent studies on SO2 effects on various
29 organ systems are summarized in Annex Tables AX4-6 through AX4-9.
30
31 3.1.3.1 Nervous System Effects Associated with Short-Term SOi Exposure
September 2007 3-58 DRAFT-DO NOT QUOTE OR CITE
-------
Relative Risk
Reference
Ballesteretal.(2001)
Fung eta!. (2005)
Ballesteretal. (2006)
Ballesleretal.(2001)
Llorcaetal.(2005)
Morris etal. (1995)
Morris etal. (1995)
Morris etal. (1995)
Morris etal. (1995)
Morris etal. (1995)
Morris etal. (1995)
Morris etal. (1995)
Welleniusetal. (2005a)
Location Age Lag
Valencia, Spain All 2
Windsor, OH, 65+ 0
Canada
Multicily Spain All 0-1
Valencia. Spain All 2
Torrelavega, Spain All 0
SO,
Los Angeles, CA 65+ Not
reported
Chicago, IL 65+ Not
reported
Philadelphia, PA 65+ Not
reported
New York, NY 65+ Not
reported
Detroit, Ml 65+ Not
reported
Houston, TX 65+ Not
reported
Milwaukee, Wl 65+ Not
reported
Allegheny County, PA 65+ 0
0
Pollutants
S0?_
S02+BS _
S02+C0
S02_
S02+PM10_
S02_
S02+PMto _
S02+C0
SO.+NO,
SO,
S02+BS
S02+C0 _
so2_
+TSP+N02+NO+H2S _
so2_
SOj+CO+NOj+Oj
so2_
S02+CO+N02+03 _
S02_
S02+CO+NO;+03 _
S02_
S02+CO+N02+03 _
S02_
S02+CO+N02+03 _
S0;_
S02+CO+N02+03 _
S02_
S02+CO+N02+03 _
S02_
S02+PM1t| _
S0/N02 _
9 1.0 11 1.;
i
•—
•4
— (
^
^K]
— '
€ I Cardiovascular]
m | Cardiac Disease |
o(p<0.5)
Congestive Heart Failure]
"~~
i —
•*•
>—
•
— o
, 'Single pollutant
oMjIti pollutant
Figure 3.1-13.
Relative risks (95% CI) of SOi-associated emergency department
visits (*) and hospitalizations for cardiovascular causes, with and
without copollutant adjustment. Risk estimates are standardized
per 10-ppb increase in 24-h average SOi concentrations or 40-ppb
increase in 1-h max
September 2007
3-59
DRAFT-DO NOT QUOTE OR CITE
-------
1 Effects of Sulfur Oxides on Neurotransmitters, Receptors, Voltage-Gated Channels, and Other
2 Neurophysiological and Biochemical Components
3 The effects of 862 exposure (10-ppm 862, 1 h/day) on lipids, lipid peroxidation, and
4 lipase activity in different regions of the brain were investigated in guinea pigs exposed for 21 to
5 24 days (Haider et al., 1981) and in rats exposed for 30 days (Haider et al., 1982). As
6 summarized in Table 3.1-1, exposure to SC>2 resulted in altered lipid profiles in both species that
7 were qualitatively and/or quantitatively brain-region specific. While levels of total lipids and
8 free fatty acids were generally lowered, the effects on phospholipids, cholesterol, esterified fatty
9 acids, and gangliosides were variable. Lipase activity and lipid peroxidation (as measured by
10 malonaldehyde content) were elevated in brain tissue due to 862 exposure. These studies
11 suggest that subacute exposure to 10-ppm 862 can lead to degradation of brain lipids. Similar
12 findings were observed in a study in which guinea pigs were exposed to 10-ppm 862 alternated
13 daily with 20-ppm of H2S (i.e., 15 daily 1-h exposures to each gas by itself) for 1 h/day for 30
14 days (Haider and Hasan, 1984). No lower concentrations were examined to determine possible
15 concentration-response relationships or a no-effect level, and effects observed at these higher
16 levels may be due to mechanisms not induced at more environmentally relevant concentrations.
17 The effect of SC>2 exposure on neuronal GSH level, antioxidant status, and antioxidant
18 enzymes was investigated in mice and rats. Wu and Meng (2003) did not observe any exposure-
19 induced changes in GSH level or related enzyme activity in brain at the lowest concentration
20 (8.4 ppm) studied. Studies that investigated oxidant status (thiobarbituric acid reactive
21 substances [TEARS] levels) in brain regions and retina in rats exposed to 10-ppm 862 for
22 1 h/day, 7 days/week, for 6 weeks also included effect of age (Kilic, 2003; Yargi9oglu et al.,
23 1999) and experimentally induced diabetes (Agar et al., 2000; Kii9iikatay et al., 2003). These
24 studies reported consistent increases in TEARS levels in brain regions in both normal and
25 diabetic rats, but results from the retina were not consistent.
26 SC>2-induced changes in neurophysiological endpoints (i.e., somatosensory-evoked
27 potentials, peak-to-peak amplitudes, visual-evoked potentials) were also investigated. SCV
28 induced changes in somatosensory-evoked potentials and peak-to-peak amplitudes were
29 observed in young (3 months), but not in older (24 months), rats. The effects of 862 exposure on
30 visual-evoked potential in experimental diabetic rats were found to be additive.
September 2007 3-60 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3.1-1. SO2 EFFECTS ON GUINEA PIG AND RAT BRAIN
Responses" in Different Brain Regions
Parameter Cerebral Hemisphere Cerebellum Brain Stem
Total Lipids -0- * (-0-) * -0- *
Free Fatty Acids -0- — -0- — -0- —
Phospholipids ft ** -0- •§•<=>«»
Cholesterol ft * -0- * -0- O)
Esterified Fatty Acids -0- — ft — -0- —
Gangliosides — * — -f- — -f-
Lipid Peroxidation ft -f- t> -f- t> (•§•)
Lipase Activity ft + ft — ft —
a Open symbols = guinea pig, closed symbols = rat; vertical arrows = significant changes (p < 0.001-0.05), vertical arrows in parentheses =
statistically nonsignificant changes > 10%, horizontal arrows = statistically nonsignificant changes < 10%, dashes = parameter not
measured.
Source: Haider etal. (1981, 1982).
1 Three ex vivo acute exposure studies using SO2 derivatives on hippocampal or dorsal
2 root ganglion neurons isolated from Wistar rats (Du and Meng, 2004a,b, 2006) observed
3 perturbations in potassium-, sodium-, and calcium-gated channels. These authors speculated that
4 such effects might correlate with the neurotoxicity that has been associated with 862 inhalation.
5 Details about all the above studies are presented in Table AX4-6.
6
7 Neurodevelopmental and Neurobehavioral Effects
8 Three studies conducted in rodents provide some information on possible
9 neurodevelopmental effects. In offspring of mice exposed to >5-ppm SC>2 from 9 days before
10 mating through the 12th to 14th day of gestation, there were no effects on somatic and
11 neurobehavioral development (e.g., eyelid and ear opening, incisor eruption, reflex development)
12 or passive avoidance testing of adult males (Petruzzi et al., 1996). A second study reported
13 delayed righting and negative geotaxis reflexes in offspring of mice exposed to >32-ppm 862 on
14 gestation days 7 through 18 (Singh, 1989).
September 2007 3-61 DRAFT-DO NOT QUOTE OR CITE
-------
1 Neurobehavioral responses were examined in adult male offspring of mice exposed to
2 >5-ppm SC>2 from 9 days before mating through gestation day 14 (Fiore et al., 1998). Compared
3 to controls, SO2-exposed male offspring displayed an increased duration of self-grooming
4 (5-ppm group), decreased frequency and duration of tail rattling (>5-ppm groups), and decreased
5 duration of defensive postures in response to an intruder mouse (> 12-ppm groups).
6 Studying the influence of age and diabetes on SO2-induced lipid peroxidation, antioxidant
7 enzyme status, and active avoidance learning in rats, Yargi9oglu et al. (2001) and Kii9iikatay
8 et al. (2007) reported that TEARS levels (indicative of lipid peroxidation) was significantly
9 increased and antioxidant status was altered in all the experimental groups studied. The authors
10 also concluded that 862 exposure induces impairments in learning in young (3 month old)
11 animals and potentiates diabetes-induced learning impairments in rats.
12 Behavioral effects in adult animals were examined in male and female mice exposed to
13 >5-ppm SC>2 from 9 days before mating through gestation days 12 through 14 (Petruzzi et al.,
14 1996). No effects were observed at concentrations of <30 ppm.
15
16 Summary of Nervous System Effects
17 In a limited number of toxicological studies, exposure to 862 has been shown to affect
18 certain neurodevelopmental and cognitive effects. There was suggestive evidence that young
19 animals and those with preexisting conditions such as diabetes were more susceptible to these
20 effects. These effects were observed only at high concentrations of 862.
21
22 3.1.3.2 Other Organ System Effects Associated with Short-Term SO2 Exposure
23 A review of animal toxicological studies published since the 1982 AQCD indicates a
24 limited number of research inquiries were conducted into the systemic effects of SC>2 exposure in
25 various other organ systems such as reproductive, hematological, gastrointestinal, renal,
26 lymphatic, and endocrine systems. The majority of these studies examined alteration profiles of
27 lipid peroxidation and antioxidant levels (Langley-Evans et al., 1996; Meng and Bai, 2004;
28 Meng et al., 2003c).
29 Though limited, the overall animal toxicological database on 862 exposure suggests no
30 adverse effects on development or reproduction. Acute exposure to 862 (0.87 ppm) in rats has
31 been found to induce hematological alterations such as increased hematocrit and decreased
32 whole blood and packed cell viscosities (Baskurt, 1988).
September 2007 3-62 DRAFT-DO NOT QUOTE OR CITE
-------
1 No overt pathological changes were observed in the liver and gastrointestinal system of
2 rats in acute or subchronic exposure studies (Gunnison et al., 1987; Langley-Evans et al., 1996).
3 Decreases in the expression of certain cytochrome P450s (CYP1A2 and CYP1 Al) in liver were
4 reported at higher concentrations (Qin and Meng, 2005). Smith et al. (1989) did not find any
5 significant effects on spleen weight or mitogen-induced activation of peripheral blood
6 lymphocytes or spleen cells in Sprague-Dawley rats exposed to 1-ppm SC>2 for 5 h/day,
7 5 days/week for 4 months. Two studies that examined the effects of SC>2 exposure in rodents
8 (Langley-Evans et al., 1996; Wu and Meng, 2003) reported alterations in GSH levels or GSH-
9 related enzymes.
10 The available studies that examined the effects of SCh exposures on the endocrine system
11 evaluated insulin-related parameters in diabetic rats that were fed a standard diet (normal), a high
12 cholesterol diet, or treated with streptozotocin to induce diabetes (Lovati et al., 1996). Exposure
13 to >5-ppm SC>2 had been found to lower plasma insulin levels in normal and
14 hypercholesterolemic rats and to result in a nonsignificant increase in plasma insulin levels in
15 diabetic rats.
16
17
18 3.2 MORTALITY ASSOCIATED WITH SHORT-TERM SO2
19 EXPOSURE
20 The studies available to review in the 1982 AQCD were mostly from historical data
21 including London, England, and New York City air pollution episodes. Effects of SOX (mainly
22 862) were investigated along with PM indices because they shared a common source, coal
23 burning, and separating their associations with mortality was a challenge that many of the earlier
24 episodic studies could not necessarily resolve. The SC>2 levels observed in these air pollution
25 episodes were several tens of times higher than the current average levels observed in U.S. cities
26 (e.g., in the 1962 New York City episode, 862 in Manhattan peaked at 400 to 500 ppb). Some of
27 these London and New York City studies suggested that PM, not SC>2, was associated with
28 observed mortality, but the 1982 AQCD could not resolve the relative roles of these two
29 pollutants and suggested that the clearest mortality associations were seen when both pollutants
30 were at high levels (24-h average values of both BS and 862 exceeding 1000 |ig/m3 [~ 400 ppb
31 for SO2]) and less so at lower ranges although the review of the studies and reanalyses found no
32 clear evidence of a threshold for SC>2.
September 2007 3-63 DRAFT-DO NOT QUOTE OR CITE
-------
1 The 1986 Second Addendum to the 1982 AQCD reviewed more reanalyses of the
2 London data and analyses of New York City, Pittsburgh, and Athens data. While these
3 reanalyses and some new analyses confirmed earlier findings (and suggested stronger evidence
4 of BS effects than of the SO2 effects), given the remaining uncertainties with exposure error and
5 statistical modeling, there was not sufficient information to quantitatively determine
6 concentration-response relationships at lower concentrations of either PM or SC>2. In the analysis
7 of nonepisodic London data, there was an indication that mortality effects were seen at BS levels
8 as low as 150 to 200 |ig/m3.
9 A series of short-term mortality effects studies in the late 1980s and early 1990s (e.g.,
10 Pope, 1989; Fairley, 1990; Dockery et al., 1992; Pope et al., 1992; Schwartz and Dockery,
11 1992a,b) showed associations between mortality and PM indices at relatively low levels. Since
12 then, a large number of epidemiological studies have investigated the adverse health effects of
13 air pollution with hypotheses mainly focused on PM, and SC>2 was often analyzed as one of the
14 potential confounders in these studies.
15
16 3.2.1 Associations of Mortality and Short-Term SO2 Exposure in Multicity
17 Studies and Meta-Analyses
18 In reviewing the range of 862 mortality risk estimates, multicity studies provide
19 especially useful information, because they analyze data from multiple cities using a consistent
20 method, avoiding potential publication bias. There have been several multicity studies from the
21 United States, Canada, and Europe, some of which will be discussed in the sections below.
22 Meta-analysis studies also provide useful information on describing heterogeneity of risk
23 estimates across studies; however, unlike multicity studies, the heterogeneity of risk estimates
24 seen in meta-analysis may reflect the variation in analytical approaches across studies. These
25 studies, as well as many other single-city studies, are summarized in Annex Table AX5-5.
26
27 3.2.1.1 Multicity Studies
28
29 National Morbidity, Mortality, and Air Pollution Study of 90 U.S. Cities
30 The time-series analysis of the largest 90 U.S. cities (Samet et al., 2000; Dominici et al.,
31 2003) in the National Morbidity, Mortality, and Air Pollution Study (NMMAPS) is by far the
32 largest multicity study conducted to date to investigate the mortality effects of air pollution, but
September 2007 3-64 DRAFT-DO NOT QUOTE OR CITE
-------
1 its primary interest was PMi0. It should also be noted that, according to the table of mean
2 pollution levels in the original report (Samet et al., 2000), SO2 was missing in 28/90 cities.
3 Annual 24-h average mean SO2 levels ranged from 0.4 ppb (Riverside, CA) to 14.2 ppb
4 (Pittsburgh, PA), with a mean of 5.9 ppb during the study period of 1987 to 1994. The analysis
5 in the original report used GAM models with default convergence criteria. Dominici et al.
6 (2003) reanalyzed the data using GAM with stringent convergence criteria as well as using
7 GLM. It should be noted that this model's adjustment for weather effects employs more terms
8 than other time-series studies in the literature, suggesting that the model adjusts for potential
9 confounders more aggressively than the models in other studies.
10 PMio and 63 (in summer) appeared to be more strongly associated with mortality than the
11 other gaseous pollutants. The authors stated that the results did not indicate associations of SO2,
12 NO2, and CO with total mortality. However, as with PMio, the gaseous pollutants SO2, NO2, and
13 CO each showed the strongest association at a 1-day lag (for Os, a 0-day lag). In contrast to
14 PMio and NO2, the inclusion of copollutants in the regression models generally resulted in
15 reduced SO2 risk estimates. Figure 3.2-1 shows the total mortality risk estimates for SO2 from
16 Dominici et al. (2003). The mortality risk estimate with a 1-day lag was 0.60% (95% CI: 0.26,
17 0.95) per 10-ppb increase in 24-h average SO2. The model with PMio and NO2 resulted in an
18 appreciably reduced SO2 risk estimate, 0.38% (95% CI: -0.62, 1.38) per 10-ppb increase in 24-h
19 average SO2. These results suggest that the observed SO2-mortality association could be
20 confounded by PMio and NO2.
21
22 Canadian Multicity Studies
23 There have been three Canadian multicity studies examining the association between
24 mortality and short-term exposure to air pollutants: (1) an analysis of gaseous pollutants in 11
25 cities from 1980 to 1991 (Burnett et al., 1998); (2) an analysis of PM2.5, coarse PM (PMi0.2.5),
26 and gaseous pollutants in 8 cities from 1986 to 1996 (Burnett et al., 2000); and (3) an analysis of
27 PM2.5, PMio-2.5, and gaseous pollutants in 12 cities from 1981 to 1999 (Burnett et al., 2004). The
28 first two studies utilized GAM with default convergence criteria. Only the PM indices were
29 reanalyzed for the Burnett et al. (2000) study by Burnett and Goldberg (2003).
30 Burnett et al. (2004) is the most extensive Canadian multicity study, both in terms of the
31 length and coverage of cities. The discussion in this study focused on NO2, because NO2 was the
September 2007 3-65 DRAFT-DO NOT QUOTE OR CITE
-------
1.5-
>, 1'°"
Jo,-
•E 0-
o>
U)
JS -0.5-
0
55
-1.0-
-1.5-
LagO
I
i
i
i
Lagl
<
i
,
i
Lag 2
i
» '
' <
*
i
B
D
BCD
Models
B
D
Figure 3.2-1.
Posterior means and 95% posterior intervals of national average
estimates of SO2 effects on total mortality from non-external causes
per 10-ppb increase in 24-h average SO2 at 0-, 1-, and 2-day lags
within sets of the 62 cities with pollutant data available. Models A =
SO2 alone; B = SO2 + PMi0; C = SO2 + PMio + O3; D = SO2 +
NO2; E = SO2 + PM10 + CO.
Source: Dominici et al. (2003).
1 best predictor of short-term mortality fluctuations among the pollutants. This was also the case
2 in the Burnett et al. (1998) study of the gaseous pollutants in 1 1 Canadian cities. The mean 24-h
3 average SO2 levels across the 12 cities was 5.8 ppb, with city means ranging from 1 ppb in
4 Winnipeg to 10 ppb in Halifax. The population-weighted average was 5 ppb. The mean SO2
5 levels in this study were similar to those in the NMMAPS (mean 24-h average SO2 levels across
6 the 62 NMMAPS cities was 5.9 ppb).
7 Total (nonaccidental), cardiovascular, and respiratory mortality were analyzed in Burnett
8 et al. (2004). For SO2, PM2.5, PMi0.2.5, PMio (arithmetic addition of PM2.5 and PMio-2.5), CoH,
9 and CO, the strongest mortality association was found at a 1-day lag, whereas for NO2, it was the
10 3-day moving average (i.e., average of 0-, 1-, and 2-day lags), and for Os, it was the 2-day
1 1 moving average. The daily 24-h average values showed stronger associations than the daily 1-h
September 2007
3-66
DRAFT-DO NOT QUOTE OR CITE
-------
1 max values for all the gaseous pollutants and CoH except for 63. The SO2 total mortality risk
2 estimate was 0.74% (95% CI: 0.29, 1.19) per 10-ppb increase in the 24-h average SO2 with a 1-
3 day lag. After adjusting for NO2, the SO2 risk estimate was reduced to 0.42% (95% CI: 0.01,
4 0.84), while the NO2 risk estimate was only slightly affected. In this analysis, no regression
5 analysis using both SO2 and PM was conducted. The Burnett et al. (2000) analysis observed that
6 the simultaneous inclusion of SO2 and PM2 5 in the model reduced the SO2 risk estimate by half,
7 whereas the PM2 5 estimate was only slightly reduced. Overall, these results suggest that SO2
8 was not an important predictor of daily mortality in the Canadian cities and that its mortality
9 associations could be confounded by NO2 or PM.
10
11 Air Pollution and Health: A European Approach, Studies 1 and 2
12 Several Air Pollution and Health: a European Approach (APHEA) analyses have reported
13 SO2 mortality risk estimates. Katsouyanni et al. (1997) examined the association of PMi0, BS,
14 and SO2 with total mortality in 12 European cities using the standard APHEA1 (GLM) approach.
15 The same data set was reanalyzed using nonparametric smooth functions in GAM models with
16 default convergence criteria to adjust for the seasonal cycles (Samoli et al., 2001) and using
17 GAM with more stringent convergence criteria as well as a parametric smoother in GLM
18 (Samoli et al., 2003). An analysis of cardiovascular and respiratory mortality in 10/12 APHEA
19 cities was conducted by Zmirou et al. (1998). The reanalysis by Samoli et al. (2003) produced
20 results that were similar to those in the original analysis by Katsouyanni et al. (1997). Since the
21 original analysis presented more results, including multipollutant model results, discussion will
22 focus on this analysis.
23 The study by Katsouyanni et al. (1997) includes seven western European cities (Athens,
24 Barcelona, Cologne, London, Lyon, Milan, and Paris) and five central eastern European cities
25 (Bratislava, Kracow, Lodz, Poznan, and Wroclaw). The data covered at least 5 consecutive
26 years for each city within the years 1980 through 1992. The SO2 levels in these cities were
27 generally higher than in the United States or Canada, with the median 24-h average SO2 ranging
28 from 13 |ig/m3 (5 ppb) in Bratislava to 74 |ig/m3 (28 ppb) in Kracow. Analysis was restricted to
29 days when PM and SO2 concentrations did not exceed 200 |ig/m3 (76 ppb for SO2). The data
30 were analyzed by each center separately following a standardized method, but the lag for the
31 "best" model was allowed to vary in these cities from 0 to 3 days. The city-specific risk
32 estimates were then examined in the second stage for source of heterogeneity using city-specific
September 2007 3-67 DRAFT-DO NOT QUOTE OR CITE
-------
1 variables such as mean pollution and weather variables, accuracy of the air pollution
2 measurements, health of the population, smoking prevalence, and geographical differences.
3 The city-specific estimates were found to be heterogeneous and, among the explanatory
4 variables, only the separation between western and central eastern European cities resulted in
5 more homogeneous groups. The total mortality risk estimates were 1.14% (95% CI: 0.88,1.39),
6 1.99% (95% CI: 1.15, 2.83), and 0.46% (95% CI: -0.23, 1.15) for all the 12 cities combined,
7 western cities, and central eastern cities, respectively, per 10-ppb increase in the 24-h average
8 SO2 at variable single-day lags. Seasonal analyses indicated that the summer estimate was
9 slightly higher than the winter estimate in the western cities, but the difference was not
10 statistically significant. The results for the two-pollutant model with 862 and BS were presented
11 for the western cities, with a similar extent (-30%) of reductions in the estimates of both
12 pollutants (1.31% [95% CI: 0.40, 2.23] for SO2). Furthermore, for western cities, they estimated
13 effects for SO2 for days with high or low BS levels and the corresponding BS effects for days
14 with high or low SC>2 levels and found that their effects were similar for days with low or high
15 levels of the other pollutant. From these results, Katsuoyanni et al. (1997) suggested that the
16 effects of the two pollutants were independent.
17 Overall, the APHEA studies provide some suggestive evidence that the effect of short-
18 term exposure to SC>2 on mortality is independent of PM. This is somewhat in contrast to the
19 U.S. and Canadian studies. The SO2 levels were much higher in the European cities, but the type
20 of PM constituents also might be different.
21
22 The Netherlands Study
23 In the Netherlands studies by Hoek et al. (2000, 2001; reanalysis Hoek, 2003), the
24 association between air pollutants and mortality were examined in a large population (14.8
25 million for the entire country) over the period of 1986 through 1994. The Netherlands were not
26 part of the APHEA SC>2 analysis. The median 24-h average SC>2 level in the Netherlands was 4
27 ppb (6 ppb for the four major cities). All the pollutants examined, including PMio, BS, 63, NC>2,
28 SO2, CO, SO42 , and nitrate, were associated with total mortality, and for single-day models, a
29 1-day lag showed the strongest associations for all the pollutants. The following risk estimates
30 are all from the GLM models with natural splines for smoothing functions. The SO2 risk
31 estimate in a single-pollutant model was 1.31% (95% CI: 0.69, 1.93) per 10-ppb increase in 24-h
September 2007 3-68 DRAFT-DO NOT QUOTE OR CITE
-------
1 average SO2 at a 1-day lag and 1.78% (95% CI: 0.86, 2.70) at an average of 0- to 6-day lag.
2 Seasonal analyses showed slightly greater effect estimates during the summer compared to the
3 winter. SC>2 was most highly correlated with BS (r = 0.70). The correlation pattern of SC>2 with
4 other pollutants was similar to that for NO2, but weaker (e.g., correlation between NO2 and BS
5 was 0.87). The simultaneous inclusion of SC>2 and BS reduced the risk estimates for both
6 pollutants (SO2 risk estimate was 1.07% [95% CI: -0.27, 2.42] per 10-ppb increase with an
7 average of 0- to 6-day lag of 24-h average 802). PMi0 was less correlated with SC>2 (r = 0.65),
8 and the simultaneous inclusion of these pollutants resulted in an increase in the 862 risk
9 estimate. These results from the analysis of the Netherlands data suggested some indication of
10 confounding between 862 and BS. Generally, the SCVmortality associations resembled the
11 pattern for NO2-mortality associations, but weaker.
12
13 Other European Multicity Studies
14 Other European multicity studies were conducted in 8 Italian cities (Biggeri et al., 2005),
15 9 French cities (Le Tertre et al., 2002), and 13 Spanish cities (Ballester et al., 2002). The studies
16 by Le Tertre et al. (2002) and Ballester et al. (2002) were conducted using GAM methods with
17 the default convergence setting.
18 Biggeri et al. (2005) analyzed eight Italian cities (Turin, Milan, Verona, Ravenna,
19 Bologna, Florence, Rome, and Palermo) for mortality and hospital admissions (mortality data
20 were not available for Ravenna and Verona). The study period varied from city to city between
21 1990 and 1999. Only single-pollutant models were examined in this study. The SO2 risk
22 estimates were 4.14% (95% CI: 1.05, 7.33), 4.94% (95% CI: 0.41, 9.67), and 7.37% (95% CI:
23 -3.58, 19.57) per 10-ppb increase with an average of 0-1-day lag of 24-h average SO2 for total,
24 cardiovascular, and respiratory deaths, respectively. Since all the pollutants showed positive
25 associations with these mortality categories and the correlations among the pollutants were not
26 presented, it is not clear how much of the observed associations are shared or confounded. The
27 mortality risk estimates were not heterogeneous across cities for all the gaseous pollutants. It
28 should be noted that in Turin, Milan, and Rome, the mean SO2 values declined by 50% from the
29 first half to the second half of the study period, while the levels of other pollutants declined by
30 smaller fractions. This also complicates the interpretation of SO2 risk estimates in this study,
31 which are much higher than those from the APHEA studies.
September 2007 3-69 DRAFT-DO NOT QUOTE OR CITE
-------
1 The French nine cities study by Le Tertre et al. (2002) examined BS, 862, NO2, and 63
2 by generally following the APHEA protocol, but using GAM with default convergence criteria
3 and using the average of lags 0 and 1 day for combined estimates. SO2 data were not available in
4 one of the nine cities (Toulouse). All four pollutants were positively associated with mortality
5 outcomes. The study did not report descriptions of correlation among the pollutants or conduct
6 multipollutant models, and therefore, it is difficult to assess the potential extent of confounding
7 among these pollutants. The SC>2 risk estimates were homogeneous across cities, with the
8 exception of Bordeaux, which was the only city that used strong acidity as a proxy for 862.
9 The Spanish Multicentre Study on Air Pollution and Mortality (EMECAM) examined the
10 association of PM indices (i.e., PMio, TSP, BS) and 862 with mortality in 13 cities (Ballester
11 etal., 2002). These studies followed the APHEA protocol, but using the GAM approach. The
12 daily mean 24-h average SC>2 concentrations ranged from 8.1 to 44.5 |ig/m3 (3 to 17 ppb). In the
13 seven cities where 1-h max SC>2 data were also available, mean concentrations ranged from 54.9
14 to 113.2 |ig/m3 (21 to 43 ppb). The combined effect estimates for total and respiratory mortality
15 were statistically significant for both 24-h average SC>2 and 1-h max SC>2. Controlling for PM
16 indices substantially diminished the risk estimates for 24-h average 862, but not for 1-h max
17 862. The authors reported that these results could indicate an independent impact of peak values
18 of 862 more than an effect due to a longer exposure.
19
20 3.2.1.2 Meta-Analyses of Air Pollution-Related Mortality Studies
21
22 Meta-Anafysis of All Criteria Pollutants (1985 to 2000)
23 Stieb et al. (2002) reviewed time-series mortality studies published between 1985 and
24 2000, and conducted a meta-analysis to estimate combined effects for PMi0, CO, NO2, O3, and
25 SO2. Since many of the studies reviewed in that analysis used GAM wth default convergence
26 parameters, Stieb et al. (2003) updated the estimates by separating the GAM versus non-GAM
27 studies. In addition, separate combined estimates were presented for single- and multipollutant
28 models. There were more GAM estimates than non-GAM estimates for all the pollutants except
29 for SC>2. For SC>2, there were 29 non-GAM estimates from single-pollutant models and 10
30 estimates from multipollutant models. The lags and multiday averaging used in these estimates
31 varied. The combined estimate for total mortality was 0.95% (95% CI: 0.64, 1.27) per 10-ppb
32 increase in the daily average SO2 from the single-pollutant models and 0.85% (95% CI: 0.32,
September 2007 3-70 DRAFT-DO NOT QUOTE OR CITE
-------
1 1.39) from the multipollutant models. Because these estimates are not from an identical set of
2 studies, the difference (or lack of a difference, as in this case) between the two estimates may not
3 necessarily be due to the effect of adding a copollutant in the model. Note that the data
4 extraction procedure of this meta-analysis for the multipollutant models was to include from
5 each study the multipollutant model that resulted in the greatest reduction in risk estimates
6 compared with that observed in single-pollutant models. It should also be noted that all the
7 multicity studies whose combined estimates have been discussed in the previous section were
8 published after this meta-analysis.
9
10 Health Effects Institute Review of Air Pollution Studies in Asia
11 The Health Effects Institute (HEI) conducted a comprehensive review of air pollution
12 health effects studies (HEI, 2004). They summarized the results from mortality and hospital
13 admission studies of the health effects of ambient air pollution in Asia (East, South, and
14 Southeast) published in peer-reviewed scientific literature from 1980 through 2003. Of the 138
15 papers the report identified, most were studies conducted in East Asia (mainland China, Taipei,
16 Hong Kong, South Korea, and Japan). The levels of SO2 in these Asian cities were generally
17 higher than in U.S. or Canadian cities, with more than half of these studies reporting mean 24-h
18 average 862 levels of >10 ppb. Based on a comparison of the reported mean 862 levels from the
19 same cities in different time periods, it is clear that the 862 levels declined significantly in the
20 1990s. However, the meta-analysis used the most recent estimate for each city to reflect recent
21 pollution levels. Based on the criteria of having at least 1 year of data, model adjustment for
22 major time-varying confounders, and reporting risk estimates per unit increase in air pollution,
23 the meta-analysis included 28 time-series studies (11 from South Korea, 6 from mainland China,
24 6 from Hong Kong, and 1 each from Taipei, India, Singapore, Thailand, and Japan). The lags
25 selected to compute combined estimates were inevitably variable; a systematic approach was
26 used to favor the a priori lag stated in the study, followed by the most significant lag, and then
27 the largest effect estimate. Eleven mortality risk estimates were used to compute a combined
28 estimate for 862. In general, the report focused on the results of single-pollutant models only, as
29 there were too few studies with results of comparable multipollutant models to allow meaningful
30 analysis. The SO2 mortality risk estimates were found to be heterogeneous. The publication bias
31 test suggested some indication of bias. The combined estimate for total mortality was 1.49%
32 (95% CI: 0.86, 2.13) per 10-ppb increase in 24-h average SO2. The report mentioned that the
September 2007 3-71 DRAFT-DO NOT QUOTE OR CITE
-------
1 resulting combined risk estimates for PM and SC>2 were similar to those found in Western
2 countries.
3
4 3.2.1.3 Summary of Risk Estimates from Multicity Studies and Meta-Analyses
5 Figure 3.2-2 shows combined estimates for total mortality per the standardized
6 increments (10-ppb increase for 24-h average 802) from the multicity studies and meta-analyses
7 discussed above. The mortality risk estimates from single-pollutant models range from 0.6%
8 (the NMMAPS) to 4.1% (the Italian 8-cities study), but given the large confidence band in the
9 Italian study, a more stable range may be 0.6 to 2%. The heterogeneity of estimates in these
10 studies may be due to several factors, including the differences in model specifications,
11 averaging/lag time, 862 levels, and effect-modifying factors. However, given the variability of
12 SC>2 and copollutants concentrations and differences in other effect-modifying factors, the range
13 of SO2 risk estimates appear to be rather narrow. It is noteworthy that the SO2 risk estimates for
14 the NMMAPS and Canadian 12-city studies are quite comparable (0.6 and 0.7%, respectively),
15 considering the difference in the modeling approach. This is in contrast to the pattern for the
16 PMio (U.S. Environmental Protection Agency, 2004) and NC>2 (U.S. Environmental Protection
17 Agency, draft, 2007) mortality risk estimates, in which the risk estimates for NMMAPS tended
18 to be smaller than those from the Canadian or other multicity studies.
19 There was not enough evidence to suggest a difference in risk estimates due to lag or
20 averaging time. In the Netherlands study, the estimate for the average of 0 to 6 days (1.8%) was
21 larger than that for the 1-day lag (1.3%). In the APHEA1 study, the estimate for "cumulative
22 effects" (2.3%, for the average of 2 to 4 consecutive days including the current day) for the
23 western cities was only slightly larger than that for the single-day lag estimate (2%). Thus, while
24 the risk estimates for multiday effects may be larger than the single-day estimates, the evidence
25 so far indicates that the magnitude of such multiday effects is not substantial.
26 Only the APHEA study examined possible source of effect modifications for SC>2 in
27 multicity or meta-analyses. They examined several potential effect modifiers such as the mean
28 levels of pollution and weather variables, accuracy of the air pollution measurements, health of
29 the population, smoking prevalence, and geographical differences. The only variable that could
30 explain the heterogeneity of city-specific risk estimates was the geographic separation (western
September 2007 3-72 DRAFT-DO NOT QUOTE OR CITE
-------
% Change in Mortality
246
Dominici et al. (2003)—U.S. 90 cities study
Lag 1 day
With PM10 and NO2
Burnett et al. (2004)—Canadian 12 cities study
Average of lag 0-2 days
With NO2
Katsouyanni et al. (1997)—European 12 cities study
All cities, variable single-day lags
Western European cities
Western European cities, with BS
Central eastern European cities
Bigger! et al. (2005)—Italian 8 cities study
Average of lag 0-1 days
Hoek (2003)—the Netherlands study
Lag 1 day
Average of lag 0-6 days
With BS
Stieb et al. (2003)—Meta-analysis, international
Variable lags
With copollutants that showed largest reduction
HEI (2004)—Meta-analysis, Asian cities
Variable lags
• Single pollutant
O Multipollutant
Figure 3.2-2. All cause (nonaccidental) SOi mortality risk estimates (95% CI) from
multicity and meta-analysis studies. Risk estimates are standardized
per 10-ppb increase in 24-h average SOi concentrations. For
multipollutant models, results from the models that resulted in the
greatest reduction in SOi risk estimates are shown.
September 2007
3-73
DRAFT-DO NOT QUOTE OR CITE
-------
1 versus central eastern European cities) for both 862 and BS, but heterogeneity in the 862 risk
2 estimates remained within the western cities.
3 In summary, the range of SO2 total mortality risk estimates is 0.4 to 2% per 10-ppb
4 increase in 24-h average SC>2. There was some suggestion of confounding between SC>2 and PM
5 and/or NC>2. The extent of multiday effects, if they exist, is not substantial. There is no clear
6 effect modifier, but the larger European study suggested that the observed heterogeneity in SC>2
7 risk estimates is at least in part regional.
8
9 3.2.1.4 Potential Confounding by Copollutants of the Association of Mortality and
10 Short-Term SO2 Exposure
11 As shown in Figure 3.2-2, the mortality risk estimates from the multipollutant models in
12 the multicity studies suggest some extent of confounding between SC>2 and PM and/or NC>2, as
13 indicated by the reduced magnitude of the 862 risk estimates. NMMAPS and the Canadian
14 study showed a similar extent of reductions in the 862 risk estimates in the multipollutant
15 models (from 0.6 to 0.4% in the NMMAPS and from 0.7 to 0.4% in the Canadian study). In both
16 the European APHEA1 analysis and the Netherlands analysis, the SO2 mortality associations
17 were reduced (though not eliminated) when BS was added to the model. The meta-analysis by
18 Stieb et al. (2003) does not suggest confounding of SC>2 by copollutants, but this was not a direct
19 comparison of estimates from the same set of studies (29 studies for single pollutant models and
20 10 studies for multipollutant models). Thus, the results from multicity studies suggest some
21 evidence of confounding, in the sense of instability of risk estimates in multipollutant models.
22 Additional single-city studies have also examined potential confounding of the 862 effect
23 on mortality by copollutants through multipollutant analyses. The studies that examined 862 and
24 PM indices and did not find substantial (i.e., more than 50%) reductions in 862 risk estimates
25 after adjustment for PM include analyses of data from Philadelphia, PA, with TSP (Kelsall et al.,
26 1997, using GAM with default convergence criteria; Moolgavkar et al., 1995); Cook County, IL,
27 with PMio (Moolgavkar, 2003, using GAM with default convergence criteria); and Los Angeles,
28 CA, with PMio or PM2.5 (Moolgavkar, 2003, using GAM with default convergence criteria).
29 Other studies that analyzed SC>2 and PM indices and did find major reductions in SC>2 risk
30 estimates after adjustment for PM include analyses of data from Philadelphia, PA, with TSP
31 (Schwartz, 2000); New York City, NY, with PMio (De Leon et al., 2003); and Santiago, Chile,
32 with PM2.s (Cifuentes et al., 2000). It is difficult to find a consistent pattern of evidence of
September 2007 3-74 DRAFT-DO NOT QUOTE OR CITE
-------
1 confounding with PM in these single-city results. It is also possible that the constituents of PM
2 (e.g., relative contribution of traffic-related pollution to PM mass) vary from city to city, and
3 hence correlations of PM with SO2 vary, contributing to apparently inconsistent results.
4 Fewer single-city studies examined multipollutant models with SC>2 and other gaseous
5 pollutants. Most studies observed that adjusting for other gaseous pollutants generally did not
6 substantially influence the SC>2 risk estimate (Bremner et al., 1999; Kelsall et al., 1997; Kwon
7 et al., 2001; Wong et al., 2001). However, one study by Cifuentes et al. (2000) did find that the
8 SO2 risk estimate was reduced substantially by adding any of CO, 63, or NC>2 in the two-
9 pollutant model in Santiago, Chile. Again, the results from these single-city studies are too
10 limited to exhibit a consistent pattern.
11 In summary, because of the lack of consistency in the way multipollutants were examined
12 (e.g., lags examined, combination of pollutants examined, model specification) and because of
13 the limited statistical power in individual cities, it is difficult to extract information that help
14 elucidate a pattern of confounding between SC>2 and other pollutants from these single-city
15 studies. The multipollutant results from multicity studies provide more useful information on
16 this issue. As noted before, the results from the multicity studies from the United States, Canada,
17 and Europe generally suggest that 862 mortality risk estimates may be confounded by
18 copollutants.
19
20 3.2.2 Cause-Specific Mortality Associated with Short-Term SO2 Exposure
21 Assessing cause-specific mortality is complicated by the lack of clarifying information on
22 contributing causes of death. That is, attribution to one or the other of the more specific
23 cardiopulmonary causes may underplay contributions of chronic CVD to respiratory-related
24 deaths (e.g., a heart attack victim succumbing to acute pneumonia) or vice versa. Several
25 multicity studies provided risk estimates for broad cause-specific categories, typically respiratory
26 and cardiovascular mortality. A summary of these risk estimates, along with the all-cause
27 mortality estimates for comparison, are presented in Figure 3.2-3. These results from multicity
28 studies suggest that the mortality risk estimates for cardiovascular and respiratory causes were
29 generally larger than that for all-cause mortality, though in some cases the effects were not
30 statistically significant, possibly because of reduced statistical power by which to examine cause-
31 specific associations. In these studies, the effect estimates for respiratory mortality were also
September 2007 3-75 DRAFT-DO NOT QUOTE OR CITE
-------
Katsouyannietal. (1997)
Zmirouetal. (1998)
Biggerietal. (2005)
*Le Tertre et al. (2002)
'Ballesteretal. (2002)
Hoek(2003)
Figure 3.2-3.
-5
Study
APHEA1 (7 W. European cities)
APHEA1 (5 W. European cities)
Italian 8 cities study
Lag
Variable
Variable
0-1
French 9 cities study
0-1
Spanish 13 cities study
0-1
The Netherlands study
0-6
% Change in Mortality
o
i
10
I
15
I
20
(COPD)
-• (Pneumonia)
All-cause (nonaccidental) and broad cause-specific (respiratory and
cardiovascular) SOi mortality risk estimates (95% CI) from multicity
studies. Risk estimates are standardized per 10-ppb increase in 24-h
average SOi concentrations.
*Note: Le Tertre et al. (2002) and Ballester et al. (2002) performed analyses using Poisson GAM with default
convergence criteria.
September 2007
3-76
DRAFT-DO NOT QUOTE OR CITE
-------
1 found to be larger than the cardiovascular mortality risk estimates, suggesting a stronger
2 association of SO2 with respiratory mortality compared to cardiovascular mortality. However,
3 this pattern was not unique to SO2; other pollutants often showed similar patterns. There were
4 numerous single-city studies that also examined broad specific causes (cardiovascular and
5 respiratory), but the patterns were not always consistent, likely due to smaller sample size, or the
6 lags reported were not consistent across the specific causes examined.
7 Some studies examined more specific causes within cardiovascular or respiratory causes.
8 In the Netherlands study (Hoek et al. 2001; reanalysis Hoek, 2003), the risk estimates for heart
9 failure (7.1% [95% CI: 2.6, 11.7] per 10-ppb increase in the average of 0- through 6-day lags of
10 24-h average SO2) and thrombosis-related deaths (9.6% [95% CI: 3.1, 16.6]) were larger than
11 that for total cardiovascular (2.7% [95% CI: 1.3, 4.1]) causes. However, a similar pattern was
12 seen for PMio, CO, and NO2 as well. In the analysis by Goldberg et al. (2003) of Montreal data,
13 the risk estimates for death with underlying cause of CHF and those deaths classified as having
14 CHF 1 ear before death were compared. They did not find associations between air pollution
15 and those with underlying cause of CHF (e.g., SO2 risk estimate was -0.1% [95% CI: -8.9, 9.6]
16 per 10-ppb increase in 24-h average SO2 with a 1-day lag), but they found associations between
17 some of the air pollutants examined (i.e., CoH, SO2, NO2) and the deaths that were classified as
18 having CHF 1 year before death (SO2 risk estimate was 5.4% [95% CI: 1.3, 9.5]). Again, the
19 association with the specific cause of death was not unique to SO2. This pattern of association
20 between multiple pollutants (including, but not specific to, SO2) and specific causes of deaths
21 was seen for an asthma mortality (Saez et al., 1999) cohort with severe asthma (Sunyer et al.
22 2002), a cohort of patients with intrauterine mortality (Pereira et al., 1998), and a cohort with
23 CHF (Kwon et al., 2001).
24 In summary, both cardiovascular and respiratory causes, as well as more specific causes
25 or categories of death, have been shown to be associated with ambient SO2 concentrations.
26 However, since other pollutants also showed similar associations with these causes or categories,
27 the possibility of confounding by these copollutants remains. While SO2 may have contributed
28 to these associations as part of the mixture of pollutants or as a surrogate index, it is difficult to
29 evaluate the specificity of SO2 effects on these specific causes of death.
30
September 2007 3-77 DRAFT-DO NOT QUOTE OR CITE
-------
1 3.2.3 Evidence from an Intervention Study
2 Many time-series studies provide estimates of excess risk of mortality, but a question
3 remains as to the likelihood of a reduction in deaths when 862 levels are actually reduced.
4 Hedley et al. (2002) took advantage of a sudden change in regulation in Hong Kong in July 1990
5 that required all power plants and road vehicles to use fuel oil with a sulfur content of <0.5% by
6 weight. The SC>2 levels after the intervention declined about 50% (from about 17 ppb to 8 ppb),
7 but the levels for PMio, NO2, and SC>42 did not change and 63 levels slightly increased. The
8 seasonal mortality analysis results showed that the apparent reduction in seasonal death rate
9 occurred only during the first winter, and this was followed by a rebound (i.e., higher than
10 expected death rate) in the following winter. Using Poisson regression of the monthly deaths,
11 the average annual trend in death rate significantly declined after the intervention for all causes
12 (2.1%), respiratory causes (3.9%), and cardiovascular causes (2.0%), but not from other causes.
13 These results seem to suggest that a reduction in SC>2 leads to an immediate reduction in deaths.
14 Hedley et al. (2002) estimated that the expected average gain in life expectancy per year due to
15 the lower SC>2 levels was 20 days for females and 41 days for males.
16 Interpreting these results is somewhat complicated by an upward trend in mortality across
17 the intervention point, which the authors noted was due to increased population size and aging.
18 The results suggest that such an upward trend is less steep after the introduction of low sulfur
19 fuel. While the Poisson regression model of monthly deaths does adjust for trend and seasonal
20 cycles, the regression model does not specifically address the influence of influenza epidemics,
21 which can vary from year to year. This issue also applies to the analysis of warm to cool season
22 change in death rates. The most prominent feature of the time-series plot (or the fitted annual
23 cycle of monthly deaths) presented in this study is the lack of a winter peak for respiratory and
24 all-cause mortality during the year immediately following the intervention. Much could be made
25 of this lack of a winter peak, but no discussion of the potential impact of (or a lack of) influenza
26 epidemics is provided. These issues make the interpretation of the estimated decline in upward
27 trend of mortality rate or the apparent lack of winter peak difficult.
28 Further, the decline in mortality following the intervention does not preclude the
29 possibility that other constituents of the pollution mixture that share the same source as SC>2 is
30 responsible for the adverse effects. Even though the PMio levels before and after the
31 intervention were stable in Hong Kong, it is possible that constituents that do not explain a major
September 2007 3-78 DRAFT-DO NOT QUOTE OR CITE
-------
1 fraction of PM may have declined. Lippmann et al. (2006) mentioned that unpublished data
2 from Hedley and coworkers reported large reductions in nickel and vanadium but not in other
3 metals in Hong Kong after the intervention. SO2 also may be serving as a modifier of the effect
4 of respirable particles. Thus, while the Hong Kong intervention data are supportive of SC>2
5 mortality effects, the possibility of mortality effects by other constituents that are associated with
6 SC>2 sources remains.
7
8 3.2.4 Summary of Effects of Short-Term SO2 Exposure on Mortality
9 The 1982 AQCD could not resolve the relative effects of short-term exposure to PM and
10 SC>2 on mortality and suggested that the clearest mortality associations were seen when both
11 pollutants were at high levels (24-h average values of both BS and SC>2 exceeding 1000 |ig/m3
12 [-400 ppb for 862]), and less so at lower ranges. The 1986 Secondary Addendum reviewed
13 more reanalyses of the London data and analyses of New York City, Pittsburgh, and Athens data,
14 but it concluded that there was not sufficient information to quantitatively determine
15 concentration-response relationships at lower concentrations of either PM or SC>2. However, in
16 the analysis of nonepisodic London data, there was an indication that mortality effects were seen
17 at BS levels as low as 150 to 200 |ig/m3.
18 Recent epidemiological studies have reported associations between mortality and SC>2,
19 often at mean 24-h average levels of <10 ppb. The range of SC>2 all cause (nonaccidental)
20 mortality risk estimates is 0.4 to 2% per 10-ppb increase in 24-h average 862 in several large
21 multicity studies and meta-analyses. Limited information suggests that the extent of multiday
22 effects, if present, is not substantial. The risk estimates for more specific categories may be
23 larger. In the large multicity time-series studies, the SO2 risk estimates were generally reduced
24 when copollutants, either PM indices and/or NC>2, were added in the model. Thus, some extent
25 of confounding among these pollutants is suggested.
26 The APHEA analysis of 12 European cities sought possible sources of heterogeneity in
27 the city-specific risk estimates, but the only important effect modifier was the geographical area,
28 with western cities showing larger 862 risk estimates than central eastern cities. However, this
29 pattern was also seen for BS. Both 862 and BS showed slightly larger estimates in the warm
30 season.
September 2007 3-79 DRAFT-DO NOT QUOTE OR CITE
-------
1 The intervention study from Hong Kong supports the idea that a reduction in 862 levels
2 results in a reduction in deaths, but this does not preclude the possibility that the causal agent is
3 not SO2 but rather something else that is associated with SO2 sources. Overall, the evidence that
4 SC>2 is causally related to mortality at current ambient levels is suggestive but limited by
5 potential confounding in the epidemiological data and the absence of strong biological
6 plausibility.
7
8
9 3.3 MORBIDITY ASSOCIATED WITH LONG-TERM SO2 EXPOSURE
10
11 3.3.1 Respiratory Effects Associated with Long-Term Exposure to SO2
12 In the 1982 AQCD, only a few studies provided sufficient quantitative evidence relating
13 respiratory symptoms or pulmonary functions changes to long-term exposure to 862. Briefly, a
14 study by Lunn et al. (1967) in Sheffield, England, provided the strongest evidence of an
15 association between pulmonary function decrements and increased frequency of lower
16 respiratory symptoms in 5- to 6-year-old children chronically exposed to ambient BS (annual
17 level of 230 to 301 |ig/m3) and SO2 levels (181 to 275 |ig/m3 [69 to 105 ppb]). A follow-up
18 study in 1968 by Lunn found no effect with much lower levels of BS (range: 48, 169 |ig/m3) and
19 SC>2 (range: 94, 253 |ig/m3 [36, 97 ppb]); it was suggested that this might be due to insufficient
20 power to detect small health effect changes.
21 The 1986 Second Addendum presented three additional studies that examined the effects
22 of long-term exposure on respiratory health. A study by Ware et al. (1986) reported that
23 respiratory symptoms were associated with annual average TSP in the range of-30 to 150 |ig/m3
24 in children (n = 8,380) from six U.S. studies. Only cough was found to be significantly
25 associated with SO2. Although the increase in symptoms did not appear concomitantly with any
26 decrements in lung function, this may indicate different mechanisms of effect. Other studies by
27 Chapman et al. (1985) and Dodge et al. (1985) also observed increased prevalence of cough
28 among children and young adults living in areas of higher SO2 concentrations; however, it was
29 noted the observed effects might have been due to intermittent high SO2 peak concentrations.
30 The 1982 AQCD noted no remarkable pulmonary pathological findings in animals
31 (monkeys and dogs) following chronic exposures to SO2 at <5.1 ppm; however, this could have
32 been due to the conventional light microscopic examination applied, which could not detect
September 2007 3-80 DRAFT-DO NOT QUOTE OR CITE
-------
1 alterations in surface membranes or cilia. Nasal mucosal alterations were observed in mice
2 exposed to 10-ppm 862 for 72 h by inhalation. Lack of data on morphological effects of 862 at
3 near ambient concentrations was noted.
4 Since the 1982 AQCD and the 1986 Second Addendum, long-term exposure studies
5 of SC>2 have investigated effects on asthma, bronchitis and respiratory symptoms, lung
6 function, and morphological effects. The epidemiological studies are summarized in Annex
7 Table AX5-6.
8
9 3.3.1.1 Asthma, Bronchitis, and Respiratory Symptoms
10 In the Six Cities Study of Air Pollution and Health, cross-sectional associations between
11 air pollutants and respiratory symptoms were examined in 5,422 white children aged 10 to 12
12 years old from Watertown, MA, St. Louis, MO, Portage, WI, Kingston-Harriman, TN,
13 Steubenville, OH, and Topeka, KS (Dockery et al., 1989). Annuals means of 24-h average SO2
14 concentrations ranged from 3.5 ppb in Topeka to 27.8 ppb in Steubenville. Except for Os, the
15 correlations among pairs of pollution measures varied between 0.53 and 0.98. No associations
16 were observed between SO2 and a variety of respiratory symptoms, including bronchitis, chronic
17 cough, chest illness, persistent wheeze, and asthma. Stronger associations were observed for PM
18 indices.
19 Dockery et al. (1996) examined the respiratory health effects of acid aerosols in 13,369
20 white children aged 8 to 12 years old from 24 communities in the United States and Canada
21 between 1988 and 1991. The city-specific annual mean SO2 concentration was 4.8 ppb, with a
22 range of 0.2 to 12.9 ppb. With the exception of the gaseous acids, nitrous and nitric acid, none of
23 the particulate or gaseous pollutants, including SO2, were associated with increased asthma or
24 any asthmatic symptoms. Stronger associations with particulate pollutants were observed for
25 bronchitis and bronchitic symptoms. For SO2, the only significant association found was with
26 chronic phlegm, with an OR of 1.19 (95% CI: 1.00, 1.40) per 5-ppb increase in SO2.
27 As part of the international SAVIAH (Small-Area Variation in Air Pollution and Health)
28 study, Pikhart et al. (2001) examined the respiratory health effects from long-term exposure to
29 SO2 in children (n = 6,959) from two central European cities with high pollution levels (Prague,
30 Czech Republic, and Poznan, Poland). A novel technique was used to estimate the outdoor
31 concentrations of SO2 at a small-area level. Outdoor SO2 was measured by passive samplers at
32 130 sites in the two cities during 2-week periods. Concentrations of SO2 at each location in the
September 2007 3-81 DRAFT-DO NOT QUOTE OR CITE
-------
1 study areas were estimated from these data by modeling using a geographic information system
2 (GIS). The estimated mean exposure to outdoor SO2 was 84 |ig/m3 (32 ppb), with a range of 66
3 to 97 |ig/m3 (25, 37 ppb), in Prague and 80 |ig/m3 (31 ppb), with a range of 44 to 140 |ig/m3 (17,
4 53 ppb) in Poznan. The prevalence of wheezing or whistling in the past 12 months was
5 associated with SO2 (OR of 1.08 [95% CI: 1.03, 1.13] per 5-ppb increase in SO2). Moreover,
6 the lifetime prevalence of wheezing or whistling (OR 1.03 [95% CI: 1.00, 1.07]) and lifetime
7 prevalence of physician-diagnosed asthma (OR 1.09 [95% CI: 1.00, 1.19]) also were associated
8 with SO2 levels.
9 Penard-Morand et al. (2005) examined the effect of long-term exposures to air pollution
10 and prevalence of exercise-induced bronchial reactivity (EIB), flexural dermatitis, asthma,
11 allergic rhinitis, and atopic dermatitis in 9,615 children aged 9 to 11 years in six French
12 communities. Using 3-year averaged concentrations of SO2, the investigators reported that the
13 prevalence of exercise-induced bronchial reactivity, lifetime asthma, and allergic rhinitis were
14 significantly associated with increases in SO2 exposure. The estimated 3-year averaged
15 concentration of SO2 was 4.6 |ig/m3 (2 ppb) in the low-exposure schools and 9.6 |ig/m3 (4 ppb)
16 in the high-exposure schools. In a single-pollutant model, the ORs were 2.37 (95% CI: 1.44,
17 3.77) for EIB and 1.58(95% CI: 1.00, 2.46) for lifetime asthma per 5-ppb increase in SO2. In
18 this study, SO2 was moderately correlated with PMio (r = 0.76) but not with Os (r = -0.02).
19 Using a two-pollutant model that included PMi0, the associations of SO2 with EIB and lifetime
20 asthma were fairly robust (<5% change).
21 Herbarth et al. (2001) performed a meta-analysis of three cross-sectional surveys
22 conducted in East Germany investigating the relationship between lifetime exposure (from birth
23 to completion of questionnaire survey) to SO2 and TSP in children and the prevalence of chronic
24 bronchitis. Using a logistic model that included variables on parental predisposition (mother or
25 father with bronchitis) and environmental tobacco smoke exposure, the authors reported that the
26 OR for bronchitis due to a lifetime exposure to SO2 was 3.51 (95% CI: 2.56, 4.82) (the
27 concentration change for which the OR was based was not presented). No associations were
28 found between TSP and the prevalence of bronchitis in children.
29 In a German study of 5,421 children, the annual mean SO2 concentration was associated
30 with morning cough over the last 12 months, but not bronchitis (Hirsch et al., 1999). This study
31 further observed that the association of SO2 and other air pollutants with respiratory symptoms
September 2007 3-82 DRAFT-DO NOT QUOTE OR CITE
-------
1 were stronger in nonatopic than in atopic children. The authors noted that these findings were in
2 line with the hypothesis that these air pollutants induce nonspecific irritative rather than allergic
3 inflammatory changes in the airway mucosa, as irritative effects would affect the clinical course
4 in nonatopic children more strongly than in atopies whose symptoms are also determined by
5 allergen exposure.
6 In a cross-sectional analysis, Heinrich et al. (2002) examined the influence of decreased
7 air pollution levels on respiratory symptoms in children aged 5 to 14 years (n = 7,632) in the
8 reunified Germany. Questionnaires were collected from the children during 1992-1993, 1995-
9 1996, and 1998-1999 in three study areas. Improvements in air quality were associated with
10 decreasing prevalence of nonallergic respiratory symptoms. The effect estimates were stronger
11 among children without indoor exposures. For those without indoor exposures, ORs of 1.21
12 (95% CI: 1.11, 1.32) were observed for prevalence of bronchitis and 1.11 (95% CI: 1.02, 1.22)
13 for frequent colds per 5-ppb increase in the annual mean of SC>2. The authors concluded that the
14 decreasing prevalence of respiratory symptoms following decreases in air pollution levels might
15 indicate the reversibility of adverse health effects in children.
16 In France, Ramadour et al. (2000) performed a cross-sectional epidemiological survey of
17 2,445 children aged 13 to 14 years living in communities with contrasting levels of air pollution
18 to determine the relationship between long-term exposure to gaseous air pollutants and
19 prevalence rate of rhinitis, asthma, and asthma symptoms. The average 862 concentrations
20 during the 2-month survey period ranged from 17.3 |ig/m3 (7 ppb) to 57.4 |ig/m3 (22 ppb) across
21 the seven communities. This study found no relationship between the mean levels of SC>2, NC>2,
22 or Os and the above-mentioned symptoms. Another study conducted in eight nonurban
23 communities in Austria observed no consistent associations between SC>2 and prevalence of
24 asthma and symptoms (Studnicka et al., 1997).
25 In California, Euler et al. (1987) studied the effects of long-term cumulative exposure to
26 TSP and SC>2 on COPD symptoms in 7,445 nonsmoking non-Hispanic white adult participants in
27 the Adventist Health Study. Using indices of cumulative exposure, this study reported that
28 cumulative exposure levels of SO2 above 400 ppb (the former California 24-h standard) resulted
29 in an increased risk of chronic obstructive pulmonary disease symptoms. Exposure levels
30 >400 ppb for 500 h/year resulted in an 18% increased risk of having COPD symptoms,
31 250 h/year, a 9% increased risk, and 100 h/year, a 3% risk.
September 2007 3-83 DRAFT-DO NOT QUOTE OR CITE
-------
1 Goss et al. (2004) conducted a cohort study to examine the effect of air pollutants on
2 11,484 patients (mean age 18.4 years) with cystic fibrosis. Study participants were enrolled in
3 the Cystic Fibrosis Foundation National Patient Registry in 1999-2000. Exposure was assessed
4 by linking air pollution values from ambient monitors with the patient's home ZIP code. During
5 the study period, the mean SO2 concentration was 4.9 ppb (SD 2.6, IQR: 2.7,5.9). This study
6 found no association between SO2 and the odds of having two or more pulmonary exacerbations.
7 One of the limitations addressed by the authors was the lack of information regarding tobacco
8 use or environmental tobacco smoke, an important risk factor for pulmonary exacerbations.
9 Several studies that examined the effects of long-term exposure to 862 on asthma,
10 bronchitis, and respiratory symptoms observed positive associations in children, with the notable
11 exception of the Harvard Six Cities study. However, there are inconsistencies in the findings
12 observed, with some finding effects on bronchitic but not asthma symptoms and vice versa. A
13 major limitation of some studies is that subjects were asked to recall prevalence of symptoms in
14 the last 12 months or in a lifetime; such long recall periods may result in significant recall bias.
15 Overall, while the evidence is suggestive, the variety of outcomes examined and the
16 inconsistencies in the observed results make it difficult to assess the impact of long-term
17 exposure of 862 on respiratory health.
18
19 3.3.1.2 Lung Function
20 Two major U.S. studies, the Harvard Six Cities Study by Dockery et al. (1989) and a
21 cross-sectional analysis of NHANES II data by Schwartz (1989), reported that no associations
22 were observed between long-term exposure to SO2 and lung function. Additional studies
23 conducted in Europe observed mixed results.
24 In a longitudinal cohort study of 1,150 children in nine communities in Austria, Frischer
25 et al. (1999) examined the effect of long-term exposure to air pollutants on lung function. Lung
26 function was measured in the spring and fall over a 3-year period from 1994 through 1996.
27 Annual mean SC>2 concentrations ranged from 2 to 6 ppb across the nine communities. The
28 authors reported no consistent associations between 862, PMio, or NC>2 and lung function.
29 Horak et al. (2002a,b) extended the study of Frischer et al. (1999) with an additional year of data.
30 The mean SO2 concentration was 16.8 |ig/m3 (6 ppb) in the winter and 6.9 |ig/m3 (3 ppb) in the
31 summer. This study found a positive association between wintertime SO2 concentrations and
32 changes in FVC, which became null with PMio in a two-pollutant model.
September 2007 3-84 DRAFT-DO NOT QUOTE OR CITE
-------
1 Frye et al. (2003) observed changes in lung function parameters associated with declines
2 in SO2 concentrations in a cross-sectional study of children (n = 2,493) conducted in East
3 Germany. During the period from 1992-1993 to 1998-1999, the annual mean SO2 level
4 dramatically declined from 113 |ig/m3 (42 ppb) to 6 |ig/m3 (2 ppb) and corresponding increases
5 in FVC and FEVi were observed. The annual mean of TSP declined from 79 |ig/m3 to 25 |ig/m3
6 as well. This study reported a 4.9% (95% CI: 0.7, 9.3) increase in FVC and a 3.0% (95% CI:
7 -1.1, 7.2) increase in FEVi per 100-|ig/m3 (38 ppb) decrease in the annual mean of 862. Results
8 from this study indicated that a reduction of air pollution in a short time period may improve
9 children's lung function; however, the observed increases in lung function parameters were
10 likely not solely attributable to decreases in 862.
11 Ackermann-Liebrich et al. (1997) examined the effect of long-term exposure to air
12 pollutants in a cross-sectional population-based sample of adults aged 18 to 60 years old
13 (n = 9,651) residing in eight different areas in Switzerland (Study on Air Pollution and Lung
14 Diseases in Adults [SAPALDIA]). They observed a 1.2% decrease in FEVi per 109-|ig/m3
15 (42 ppb) increase in SC>2 for adults. Significant associations also were observed for PMi0 and
16 NC>2. The limited number of study areas and high intercorrelation between the pollutants made it
17 difficult to assess the effect of an individual pollutant. The authors concluded that air pollution
18 from fossil fuel combustion, which was the main source of air pollution for 862, NC>2, and PMio
19 in Switzerland, was associated with decrements in lung function parameters in this study.
20 An animal toxicological study in rabbits that were exposed to 5-ppm SC>2 for 13 weeks
21 beginning in the neonatal period (Douglas et al., 1994) did not observe any alterations in
22 pulmonary function or respiratory parameters, i.e., lung resistance, dynamic compliance, trans-
23 pulmonary pressure, tidal volume, respiration rate, minute volume. These results, taken together
24 with the epidemiological evidence, do not indicate that long-term exposure to 862 has a
25 detrimental effect on lung function.
26
27 3.3.1.3 Morphological Effects
28 Three animal toxicological studies published since the 1982 AQCD reported some
29 histopathological changes in the respiratory system following acute (<24 h) to chronic
30 (>6 month) exposures and lesions were primarily observed in airways. No alveolar lesions
31 (including electron microscopic evaluation) were observed in guinea pigs exposed to 1-ppm SC>2
September 2007 3-85 DRAFT-DO NOT QUOTE OR CITE
-------
1 for 3 h/day for 6 days (Conner et al., 1985). No pulmonary or nasal lesions were observed in rats
2 exposed to 5-ppm 862 for 5 days/week for 4 weeks (Wolff et al., 1989). A weakness of the
3 study is that histopathological methods were not reported. Smith et al. (1989) exposed rats for 4
4 to 8 months to 1-ppm SC>2 and observed increased incidence of bronchiolar epithelial hyperplasia
5 and a small increase (12%) in numbers of nonciliated epithelial cells in terminal respiratory
6 bronchioles at 4 but not 8 months of exposure. A limitation of the study was the examination of
7 a single concentration, which does not allow for concentration-response assessment or
8 identification of a no-effect-level. The studies on the morphological effects are summarized in
9 Annex Table AX4-10.
10
11 3.3.2 Carcinogenic Effects Associated with Long-Term Exposure to SO2
12 The 1982 AQCD concluded that little or no clear epidemiological evidence substantiated
13 the hypothesized links between 862 or other SOX and cancer. From the toxicological studies, it
14 was noted that while there were some indications of carcinogenicity for both 862 and 862 +
15 benzo[a]pyrene (B[a]P), complex exposure regimens, problematic dose determinations, and/or
16 inadequately reported experimental details led to the conclusion that SC>2 could only be
17 considered a suspect carcinogen/cocarcinogen. More recent studies on SO2-related
18 carcinogenicity are summarized in Annex Tables AX5-7 (epidemiological studies) and AX4-11
19 (toxicological studies).
20 A limited number of recent epidemiological studies have investigated the relationship
21 between long-term exposure to 862 and lung cancer incidence. Nyberg et al. (2000) conducted a
22 case-control study of men aged 40 to 75 years with (n = 1,042) and without (n = 2,364) lung
23 cancer in Stockholm County, Sweden. They mapped residence addresses to a GIS database to
24 assign individual exposures to SC>2 from defined emission sources (mainly local oil-fueled
25 residential heating). Available SC>2 measurement data were used to calibrate the model. In this
26 study, SC>2 was considered an indicator of air pollution from residential heating. Exposure to
27 NC>2, considered to be a marker of traffic pollution, also was evaluated in this study. The 90th
28 percentile 30-year average 862 level was 78.20 |ig/m3 (30 ppb). After adjusting for potential
29 confounders (e.g., smoking, occupational exposures), long-term average heating-related 862
30 exposure was not associated with an increase in risk of lung cancer. A weak association for the
31 30-year average traffic-related NC>2 exposure was observed.
September 2007 3-86 DRAFT-DO NOT QUOTE OR CITE
-------
1 Very similar results were reported in a Norwegian study by Nafstad et al. (2003). The
2 study population is a cohort of 16,209 men who enrolled in a study of CVD in 1972. The
3 Norwegian cancer registry identified 422 incident cases of lung cancer. SO2 exposure data were
4 modeled based on residence using data for observed concentrations and emission from point
5 sources (e.g., industry and heating of buildings and private homes) and traffic. Once again, no
6 association was observed between long-term exposure to SO2 and lung cancer incidence.
7 The carcinogenic potential of SC>2 was examined more extensively in animal
8 toxicological studies. Gunnison et al. (1988) conducted a two-part study in which rats were
9 exposed either for 21 weeks (6 h/day, 5 days/week) to 0-, 10-, or 30-ppm 862, or for 21 weeks to
10 two tungsten-supplemented, molybdenum-deficient diets. This latter regimen induces a
11 condition of sulfite oxidase deficiency, resulting in elevated systemic levels of sulfite:bisulfite
12 relative to control values (e.g., in plasma, from 0 to 44 jiM; and in tracheal tissue, from 33 to 69
13 or 550 nmol/g wet wt). Beginning with week 4, some groups from each regimen received
14 weekly tracheal installations of 1-mg B[a]P for 15 weeks. Overall results indicated that
15 squamous cell carcinoma was not induced, or in the B[a]P groups coinduced or promoted, by
16 862 inhalation or elevated systemic sulfite:bisulfite. Due to the very high incidences of animals
17 with tumors in the groups exposed to only B[a]P (65/27, 63/72), carcinogenicity or
18 cocarcinogenicity of 862 or sulfite:bisulfite could only have been detected as a shortening of
19 tumor induction time and/or an increase in rate of tumor appearance, and neither was observed.
20 As noted by the authors, these findings do not support the hypothesis that SO2 exposure might
21 enhance the carcinogenicity of B[a]P by elevating systemic sulfite:bisulfite that could generate
22 glutathione-^-sulfonates, which in turn could inhibit glutathione ^-transferase (GST) and reduce
23 intracellular GSH and, thus, interfere with a major detoxication pathway.
24 Two similar studies were published that investigated the ability of 10 to 11 months of
25 exposure (16 h/day) to 4-ppm 862, 6-ppm NC>2, or their combination to affect the carcinogenicity
26 of either urban suspended PM (SPM) (Ito et al., 1997) or diesel exhaust particle (DEP) (Ohyama
27 et al., 1999) extract-coated carbon particles. The former study found that, while exposure to
28 SPM extract-coated carbon particles significantly increased pulmonary endocrine cell (PEC)
29 hyperplasia, coexposure to SC>2, NC>2, or their combination was without additional affect. Also,
30 irrespective of gas coexposure, SPM extract-coated carbon particles demonstrated a few PEC
31 papillomas versus control frequencies of zero.
September 2007 3-87 DRAFT-DO NOT QUOTE OR CITE
-------
1 Using Syrian golden hamsters, Heinrich et al. (1989) investigated whether coexposure to
2 10-ppm SO2 and 5-ppm NC>2 for 6 to 8 months (5 days/week, 19 hours/day) could enhance
3 tumorigenesis induced by a single subcutaneous injection of diethylnitrosamine (DEN) during
4 week 2. The combined gas exposure did not affect body weight gain and only minimally
5 shortened survival times. Compared to the DEN groups, serial sacrifices of gas-exposed animals
6 demonstrated progressively increasing numbers of tracheal mucosal cells and aberrant tracheal
7 cell cilia. In the lung, gas-mixture-related effects were largely limited to a progressing type of
8 alveolar lesion that involved a lining of bronchiolar epithelium and the appearance of pigment-
9 containing AM and to a mild, diffuse thickening of the alveolar septa. Exposure to the combined
10 gases by itself did not induce tumors of the upper respiratory tract, nor did it enhance the
11 induction of such tumors by DEN.
12 In conclusion, the epidemiological studies did not provide any evidence that long-term
13 exposure to SO2 is associated with an increased risk of lung cancer. The toxicological studies
14 indicate that any potential pathways of SOX to induce carcinogenesis, cocarcinogenesis, or tumor
15 promotion appear complex and may be highly situational. SC>2 and its derivatives appear
16 unlikely to have significant carcinogenic potential.
17
18 3.3.3 Prenatal and Neonatal Outcomes Associated with Long-Term SO2
19 Exposure
20 In recent years, the effects of prenatal and neonatal exposure to air pollution have been
21 examined by several investigators. The most common endpoints studied are low birth weight,
22 preterm delivery, and measures of intrauterine growth. Preterm birth and low birth weight may
23 result in serious long-term health outcomes for the infant. Preterm birth is the leading cause of
24 infant mortality and is a major determinant of a variety of adverse neurodevelopmental outcomes
25 and chronic respiratory effects (Berkowitz and Papiernik, 1993). Low birth weight has also been
26 linked with increased risk of infant mortality and morbidity. Additional studies have examined
27 sudden infant death syndrome (SIDS) and neonatal hospitalizations. Epidemiological studies of
28 ambient SO2 effects on prenatal and neonatal exposure are summarized in Annex Table AX5-8.
29 Usually, these studies have used routinely collected air pollution data and birth
30 certificates from a given area for their analysis. In evaluating the results of these studies, the
31 limitations of both fixed site monitoring for routine air pollution and the limitations of data on
32 birth certificates must be kept in mind. The reliability and validity of birth certificate data has
September 2007 3-88 DRAFT-DO NOT QUOTE OR CITE
-------
1 been reviewed (Buescher et al., 1993; Piper et al., 1993) and found to vary in degrees of
2 reliability by specific variables. Variables rated the most reliable included birth weight, maternal
3 age, race, and insurance status. Gestational age, parity, and delivery type (vaginal versus
4 cesarean) were reasonably reliable, while obstetrical complications and personal exposures such
5 as smoking and alcohol consumption, were not.
6 While most studies analyzed average SO2 exposure for the whole pregnancy, many also
7 considered exposure during specific trimesters or other time periods. Fetal growth, for example,
8 is much more variable during the third trimester. Thus, studies of fetal growth might anticipate
9 that exposure during the third trimester would have the greatest likelihood of an association.
10 However, growth can also be affected through placentation, which occurs in the first trimester.
11 Similarly, preterm delivery might be expected to be related to exposure early in pregnancy
12 affecting placentation, or through acute effects occurring just prior to delivery.
13 Epidemiological studies examining the effects of air pollutants on low birth weight are
14 summarized in Figure 3.3-1. Maisonet et al. (2001) examined the association between air
15 pollution and low birth weight in six northeastern cities of the United States, i.e., Boston, MA,
16 Hartford, CT, Philadelphia, PA, Pittsburgh, PA, Springfield, MA, and Washington, DC. The
17 study population consisted of 89,557 singleton, term live births (37 to 44 weeks of gestation)
18 born between January 1994 and December 1996. Low birth weight was classified as <2500 g.
19 This large multicity study observed an association between low birth weight and 862
20 concentrations among whites during each trimester. This association was not robust to the
21 inclusion of all races and ethnicities. A consistent concentration-response relationship was not
22 observed.
23 An increased risk for low birth weight associated with ambient SC>2 concentrations was
24 reported by Dugandzic et al. (2006) in a large cohort study of 74,284 women with term,
25 singleton births from 1988 through 2000 in Nova Scotia, Canada. The mean 24-h average SC>2
26 concentration over the study period was 10 ppb (IQR 7). These investigators found that 862
27 concentrations during the first trimester, but not the other two trimesters, were associated with
28 increased risk of low birth weight. The effect estimate was 1.14 (95% CI: 1.04, 1.26) per 5-ppb
29 increase in SC>2 level.
September 2007 3-89 DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Maisonetetal. (2001)
Liu etal. (2003)
Bobak (2000)
Maisonetetai. (2001)
Dugandzic et at. (2006)
Bobak et al. (2000)
Maisonetetal. (2001)
Liu etal. (2003)
Bobak (2000)
Wang etal. (1997)
Location
6 Northeastern cities, U.S.
All
White
African American
Hispanic
Vancouver, Canada
Dugandzic et al. (2006) Nova Sootia, Canada
Czech Republic
6 Northeastern cities, U.S.
White
African American
Hispanic
Nova Scotia, Canada
Czech Republic
6 Northeastern cities, U.S.
Vancouver, Canada
All
White
African American
Hispanic
Dugandzic et al. (2006) Nova Sootia, Canada
Czech Republic
Beijing, China
1 1st" trimester]
[ 2nd trimester
fl
3rd trimester
0.8 0.9 1.0 1.1 1.2
Relative Risk
1.3
Figure 3.3-1.
Relative risks (95% CI) for low birth weight, grouped by trimester of
exposure. Risk estimates are standardized per 5-ppb increase in
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.
September 2007
3-90
DRAFT-DO NOT QUOTE OR CITE
-------
1 Liu et al. (2003) found similar results in a study of pregnancy outcomes and air pollution
2 in Vancouver, Canada. The mean 24-h average 862 concentration was 4.9 ppb (IQR 7.7) from
3 1985 to 1998. Maternal exposure during the first month was associated with an increased risk of
4 low birth weight (OR 1.11 [95% CI: 1.01,1.22]). Additional studies from the United States,
5 Europe, Latin America, and Asia have reported positive associations between low birth weight
6 and maternal exposure during the first (Bell et al., 2007; Bobak, 2000; Ha et al., 2001;
7 Mohorovic, 2004; Yang et al., 2003b), second (Bobak, 2000; Gouveia et al., 2004; Lee et al.,
8 2003b), and third (Bobak, 2000; Lin et al., 2004b; Wang et al., 1997) trimesters.
9 Preterm delivery, intrauterine growth retardation (IUGR), and birth defects are additional
10 adverse birth outcomes that have been associated with ambient 862 levels. In a time-series
11 analysis using data from four Pennsylvania counties, Sagiv et al. (2005) reported that the mean
12 6-week SC>2 exposure prior to birth was associated with increased risk of preterm birth with a
13 relative risk (RR) of 1.05 (95% CI: 1.00, 1.10) per 5-ppb increase in SO2. A 5-ppb increase in
14 SO2 concentrations 3 days before birth was associated with an RR of 1.02 (95% CI: 0.99, 1.05).
15 The authors discussed two plausible mechanisms for the effects of air pollution on preterm birth:
16 (1) changes in blood viscosity due to inflammation as a result of air pollution (citing Peters et al.,
17 1997), and (2) maternal infection during pregnancy as a consequence of impaired immunity from
18 air pollution exposure. Liu et al. (2003) reported that 862 exposure during the last month of
19 pregnancy was associated with preterm birth with an OR of 1.09 (95% CI: 1.01,1.19) for a
20 5-ppb increase in SO2, in Vancouver, Canada. Similar results were found for studies conducted
21 in the Czech Republic (Bobak, 2000), Korea (Leem et al., 2006), and Beijing (Xu et al., 1995).
22 Liu et al. (2003) further reported that SC>2 exposure during the last month of pregnancy
23 was associated with IUGR (OR = 1.07 [95% CI: 1.01,1.13]). However, in a later study in the
24 Canadian cities of Calgary, Edmonton, and Montreal, Liu et al. (2006) did not observe
25 associations between maternal exposure to SO2 and increased risk of IUGR.
26 Pereira et al. (1998) found a positive association between SO2 and intrauterine mortality
27 in Sao Paulo, Brazil, during a 2-year period, though the effect was sensitive to model
28 specifications and did not support a concentration-response relationship. The most robust
29 association was observed for an index of three gaseous pollutants (i.e., NO2, SO2, CO) with
30 mortality.
September 2007 3-91 DRAFT-DO NOT QUOTE OR CITE
-------
1 Gilboa et al. (2005) conducted a population-based case-control study to investigate the
2 association between maternal exposure to air pollutants during weeks 3 through 8 of pregnancy
3 and the risk of selected cardiac birth defects and oral clefts in live births and fetal deaths between
4 1997 and 2000 in seven Texas counties. When the highest quartile of exposure was compared to
5 the lowest, the authors observed a positive association between SO2 and isolated ventricular
6 septaldefects(OR = 2.16 [95% CI: 1.51,3.09]). This study supports the notion that the
7 developing embryo and growing fetus constitute a subpopulation susceptible to air pollution
8 exposure.
9 Several studies have examined adverse health outcomes in relation to 862 concentrations
10 during the neonatal period. Dales et al. (2006) evaluated hospitalizations for respiratory
11 disorders in neonates <4 weeks of age from hospitals in 11 large Canadian cities during a 15-year
12 study period (population-weighted average 24-h average SO2 of 4.3 ppb). They observed a 5.5%
13 (95% CI: 2.8, 8.3) increase in respiratory hospitalizations associated with a 10-ppb increase in
14 24-h average SC>2 concentrations with a 2-day lag. This effect was slightly attenuated after
15 adjusting for PMi0 and gaseous copollutants. To investigate the influence of ambient SC>2
16 concentrations on SIDS, Dales et al. (2004) conducted a time-series analysis comparing daily
17 rates of SIDS and daily 862 concentrations from 12 large Canadian cities during a 16-year
18 period. The mean 24-h average 862 level across the 12 cities was 5.51 ppb (IQR 4.92). There
19 was an 18.0% (95% CI: 4.4, 33.4) increase in SIDS incidence for a 10-ppb increase in 24-h
20 average SO2 levels. The authors concluded that the effect of SO2 was independent of
21 sociodemographic factors, temporal trends, and weather.
22 In summary, studies on birth outcomes have found suggestive positive associations
23 between SC>2 exposure and low birth weight. While most of these studies adequately controlled
24 for maternal education, parity, age, and sex of child, many could not adjust for socioeconomic
25 status, occupational exposures, indoor pollution levels, maternal smoking, alcohol use, or
26 prenatal care. This may make comparisons across studies difficult to interpret. Additional
27 limitations affecting the interpretation of these studies is a lack of evidence for biological
28 plausibility of an effect, inconsistencies across trimesters of pregnancy, and a lack of evidence to
29 evaluate confounding by copollutants. The limited number of studies addressing preterm
30 delivery, IUGR, birth defects, neonatal hospitalizations, and infant mortality make it difficult to
31 draw conclusions regarding these outcomes.
September 2007 3-92 DRAFT-DO NOT QUOTE OR CITE
-------
1 3.4 MORTALITY ASSOCIATED WITH LONG-TERM SO2 EXPOSURE
2 At the time of the 1982 AQCD, the available studies on the effects of long-term exposure
3 to SO2 on mortality were all ecological cross-sectional studies. This study design could not take
4 into consideration such confounders as cigarette smoking, occupational exposures, and social
5 status. In addition, there were questions regarding how representative the aerometric data used
6 were for community exposure. Therefore, it was concluded that the epidemiological studies did
7 not provide valid quantitative data relating respiratory disease or other types of mortality to long-
8 term (annual average) exposures to SC>2 or PM.
9 The 1986 Secondary Addendum reviewed more studies of this type, with information on
10 more detailed components of PM (inhalable and fine particles and particulate SO42 ). While
11 some studies suggested importance of the size of PM, the fundamental problem of the study
12 design made it difficult to interpret the risk estimates. The 1986 Secondary Addendum also
13 reviewed a Japanese study in which the death rates from asthma and chronic bronchitis in a
14 highly polluted section of Yokkaichi, an industrial city with large SC>2 emissions from the largest
15 oil-fired power plant in Japan, were compared with those in a less polluted area of the same city.
16 SO levels in the polluted harbor area ranged from around 1.0 to 2.0 mg/day (annual average)
17 during 1964 through 1972 and then steadily declined to less than 0.5 mg/day in 1982. This is in
18 contrast to levels consistently <0.3 mg/day in the low pollution areas throughout 1967 through
19 1982. Annual average levels for other pollutants (i.e., NO2, TSP, oxidants) monitored in the high
20 pollution area were consistently low from 1974 through 1982. The results indicated elevated
21 rates of chronic bronchitis mortality in the highly polluted area compared to the less polluted
22 area, but the 1986 Secondary Addendum could not conclude that this was due to SC>2 alone,
23 because SC>42 or other sulfur agents such as H2SO4 could have been responsible.
24 Several, more recent studies have examined long-term exposure effects of air pollution,
25 including SC>2, on mortality. These studies are summarized in Annex Table AX5-9. As with
26 short-term exposure studies, the focus of most of these studies was mainly on PM though some
27 focused on traffic-related air pollution. They all used Cox-proportional hazards regression
28 models with adjustment for potential confounders. The designs of these studies are
29 epidemiologically better than earlier cross-sectional studies in that the outcome and most of the
30 potential confounders (e.g., smoking history, occupational exposure) are measured on an
September 2007 3-93 DRAFT-DO NOT QUOTE OR CITE
-------
1 individual basis. However, the geographic scale and method for exposure estimates varied
2 across these studies.
3
4 3.4.1 Associations of Mortality and Long-Term SO2 Exposure in Key
5 Studies
6
7 3.4.1.1 U.S. Cohort Studies
8
9 Harvard Six Cities Studies
10 Dockery et al. (1993) conducted a prospective cohort study to study the effects of air
11 pollution with the main focus on PM components in six U.S. cities. These cities were chosen
12 based on the levels of air pollution, with Portage, WI and Topeka, KS representing the less
13 polluted cities and Steubenville, OH representing the most polluted city. Mean SO2 levels
14 ranged from 1.6 ppb in Topeka to 24.0 ppb in Steubenville from 1977 to 1985. Cox proportional
15 hazards regression was conducted with data from a 14- to 16-year follow-up of 8,111 adults in
16 the six cities. Dockery et al. reported that lung cancer and cardiopulmonary mortality were more
17 strongly associated with the levels of inhalable and fine PM and SC>42 particles than with the
18 levels of TSP, SO2, NO2, or acidity of the aerosol.
19 Krewski et al. (2000) conducted a sensitivity analysis of the Harvard Six Cities study and
20 examined associations between gaseous pollutants (i.e., Os, NO2, SO2, and CO) and mortality.
21 SO2 showed positive associations with total (RR = 1.05 [95% CI: 1.02, 1.09] per 5-ppb increase
22 in the average SO2 over the study period) and cardiopulmonary (1.05 [95% CI: 1.00, 1.10])
23 deaths, but in this dataset SO2 was highly correlated with PM2.5 (r = 0.85), SO42 (r = 0.85), and
24 NO2(r = 0.84).
25
26 American Cancer Society Cohort Studies
27 Pope et al. (1995) investigated associations between long-term exposure to PM and the
28 mortality outcomes in the American Cancer Society (ACS) cohort. Ambient air pollution data
29 from 151 U.S. metropolitan areas in 1981 were linked with individual risk factors in 552,138
30 adults who resided in these areas when enrolled in the prospective study in 1982. Death
31 outcomes were ascertained through 1989. PM2.5 and SO42 were associated with total,
32 cardiopulmonary, and lung cancer mortality, but not with mortality for all other causes. Gaseous
33 pollutants were not analyzed in the 1995 Pope et al. study. Krewski and co-investigators
September 2007 3-94 DRAFT-DO NOT QUOTE OR CITE
-------
1 (Krewski et al., 2000; Jerrett et al., 2003) conducted an extensive sensitivity analysis of the Pope
2 et al. (1995) ACS data, augmented with additional gaseous pollutants data. The mean 862
3 concentrations were 7.18 ppb in the warm season (April to September) and 11.24 ppb in the cool
4 season (October to March). Among the gaseous pollutants examined, only SC>2 showed positive
5 associations with mortality. The relative risk estimates for total mortality was 1.06 (95% CI:
6 1.05, 1.07) per 5-ppb increase in the annual average SC>2. Analysis using SC>2 measured in
7 different seasons produced a somewhat higher estimate for the warm season than that for the
8 cool season (7% compared to 5% per 5-ppb increase). Although the subjects in the ACS cohort
9 came from all regions of the United States, the majority of the cities fall in the eastern United
10 States, where both SC>2 and SC>42 tend to be higher. PM2.5 levels are also higher in the East. To
11 address the influence of these spatial patterns, which may confound associations between
12 mortality and these pollutants, Krewski et al. (2000) conducted extensive two-stage regression
13 modeling. In these models, the association between SC>2 and mortality persisted after adjusting
14 for SO42 , PM2.s, and other variables. For example, in the spatial filtering model (which resulted
15 in the largest reduction of 862 risk estimate when SC>42 was included), the 862 total mortality
16 RR estimate was 1.07 (95% CI: 1.03, 1.11) in the single-pollutant model and 1.04 (95% CI:
17 1.02, 1.06) with SC>42 in the two-pollutant model. The risk estimates for PM2.5 and SO42 were
18 diminished when SC>2 was included in the models. The results also showed that SC>2 risk
19 estimates were generally insensitive to adjustment for spatial correlation. Thus, these results
20 suggest that the association between SC>2 and mortality may be confounded with PM, but the
21 association cannot be accounted for by PM2.5 or SC>42 alone. Krewski et al. (2000) noted that
22 their reanalysis of the ACS and Harvard Six Cities studies suggested that mortality might be
23 attributed to more than one component of the complex mixture of ambient air pollutants in urban
24 areas in the United States.
25 The original Pope et al. (1995) study and the Krewski et al. (2000) reanalysis both used
26 the air pollution exposure estimates that are based on the average over the Metropolitan
27 Statistical Area (MSA), which consists of multiple counties. To investigate the effects of
28 geographic scale over which the air pollution exposures are averaged, Willis et al. (2003)
29 reanalyzed the ACS cohort data using the exposure estimates averaged over the county scale, and
30 compared the results with those based on the MSA-scale average exposure. Less than half of the
31 cohort used in the MSA-based study was used in the county-scale based analysis, because of the
September 2007 3-95 DRAFT-DO NOT QUOTE OR CITE
-------
1 limited availability of SO42 monitors and because of the loss of subjects with the use of five-
2 digit ZIP codes. The mean (9.3 ppb versus 10.7 ppb) and range (0.0 to 29.3 ppb versus 0.0 to
3 27.2 ppb) of the MSA- and county-level SO2 data sets were similar. In the analysis comparing
4 the two-pollutant model with SC>42 and SO2, they found that the inclusion of SC>2 reduced SC>42
5 risk estimates substantially (>25%) in the MSA-scale model but not substantially (<25%) in the
6 county-scale model. In the MSA-level analysis (with 113 MSAs), the SO2 RR estimate was 1.04
7 (95% CI: 1.02, 1.06) per 5-ppb increase, with SC>42 in the model. In the county-level analysis
8 (91 counties) with SO42 in the model, the corresponding estimate was smaller (1.02 [95% CI:
9 1.00, 1.05]). It should also be noted that the correlation between covariates are different between
10 the MSA-level data and county-level data. The correlation between SO2 and SC>42 was 0.48 in
11 the MSA-level data, but it was 0.56 in the county-level data. The correlation between poverty
12 rate and SC>2 was -0.16 in the MSA-level data, but it was 0.15 in the county-level data. Thus,
13 the extent of confounding between SC>2 and PM components as well as among other covariates in
14 the model can be affected by the geographic scale of aggregation of exposure estimates. It is not
15 clear, however, if the smaller geographic scale increases or decreases exposure characterization
16 error for 862, because a certain extent of smoothing (averaging) over distance may reduce very
17 local concentration peaks that are not relevant to the city-wide population.
18 Pope et al. (2002) extended analysis of the ACS cohort with double the follow-up time
19 (to 1998) and triple the number of deaths compared to the original Pope et al. (1995) study. In
20 addition to PM2.5, all the gaseous pollutants were retrieved for the extended period and analyzed
21 for their associations with death outcomes. As in the 1995 analysis, the air pollution exposure
22 estimates were based on the MSA-level averages. PM2.s was associated with total,
23 cardiopulmonary, and lung cancer mortality but not with deaths for all other causes. 862 was
24 associated with all the mortality outcomes, including all other causes of deaths. The 862 RR
25 estimate for total mortality was 1.03 (95% CI: 1.02, 1.05) per 5-ppb increase (1982 to 1998
26 average). The association of SO2 with mortality for all other causes (SC>42 also showed this
27 pattern) makes it difficult to interpret the risk estimates. The sensitivity analysis by Krewski
28 et al. (2000) did not provide SO2 risk estimates for all other causes, and it is not clear whether
29 this pattern is found in other data sets.
30
September 2007 3-96 DRAFT-DO NOT QUOTE OR CITE
-------
1 Women's Health Initiative Cohort Study
2 Miller et al. (2007) studied 65,893 postmenopausal women between the ages of 50 and
3 79 years without previous CVD in 36 U.S. metropolitan areas from 1994 to 1998. They
4 examined the association between one or more fatal or nonfatal cardiovascular events and the air
5 pollutant concentrations. Subjects' exposures to air pollution were estimated by assigning the
6 annual mean levels of air pollutants measured at the nearest monitor to the location of residence
7 on the basis of its five-digit ZIP code centroid, which allowed estimation of effects due to both
8 within-city and between-city variation of air pollution (this was only done for PM2.5).
9 A total of 1,816 women had one or more fatal or nonfatal cardiovascular events,
10 including 261 deaths from cardiovascular causes. Hazard ratios (HR) for the first cardiovascular
11 event were estimated. The results for models that only included subjects with non-missing
12 exposure data for all pollutants (n = 28,402 subjects, resulting in 879 CVD events) are described
13 here. In the single-pollutant models, PM2.5 showed the strongest associations with the
14 cardiovascular events by far among the pollutants (HR = 1.24 [95% CI: 1.04, 1.48] per
15 10-|ig/m3 increase in annual average), followed by SO2 (1.07 [95% CI: 0.95, 1.20] per 5-ppb
16 increase in the annual average). In the multipollutant model where all the pollutants (i.e., PM2.5,
17 PMio-2.5, CO, SO2, NO2, 63) were included in the model, the PM2.5 association with overall
18 cadiovascular events was even stronger (1.53 [95% CI: 1.21,1.94]). The association with SO2
19 also became stronger (1.13 [95% CI: 0.98,1.30]). Correlations among these pollutants were not
20 described and, therefore, the extent of confounding among these pollutants in these associations
21 could not be examined, but PM2 5 clearly was the best predictor of cardiovascular events. Miller
22 et al. (2007) did not report the associations between SO2 with cardiovascular mortality.
23 However, because the PM2 5 HR for cardiovascular deaths was even larger than that for the
24 overall cardiovascular events, it also seems possible that this may be the case for SO2, though the
25 concern for potential confounding remains.
26
27 The EPRI-Washington University Veterans' Cohort Mortality Studies
28 Lipfert et al. (2000) conducted an analysis of a national cohort of-70,000 male U.S.
29 military veterans who were diagnosed as hypertensive in the mid 1970s and were followed up for
30 about 21 years (up to 1996). This cohort was 35% black and 57% were current smokers (81% of
31 the cohort had been smokers at one time). PM2.5, PMi0, PMi0.2.5, TSP, SO42 , CO, O3, NO2, SO2,
32 and lead were examined in this analysis. No mean or median level of SO2 was reported. The
September 2007 3-97 DRAFT-DO NOT QUOTE OR CITE
-------
1 county of residence at the time of entry to the study was used to estimate exposures. Four
2 exposure periods (1960-1974, 1975-1981, 1982-1988, and 1989-1996) were defined, and deaths
3 during each of the three most recent exposure periods were considered. The results for SO2 were
4 presented only qualitatively as part of their preliminary screening regression results. Lipfert
5 et al. (2000) noted that lead and SO2 were consistently negative and were not considered further.
6 They also noted that the pollution risk estimates were sensitive to the regression model
7 specification, exposure periods, and the inclusion of ecological and individual variables. The
8 authors reported that indications of concurrent mortality risks were found for NC>2 and peak 63.
9 Lipfert et al. (2006a) examined associations between traffic density and mortality in the
10 same cohort, whose follow-up period was extended to 2001. As in their 2000 study, four
11 exposure periods were considered but including more recent years. The 95th percentiles of daily
12 average in each of the exposure periods were considered for SC>2. They reported that traffic
13 density was a better predictor of mortality than ambient air pollution variables with the possible
14 exception of Os. The log-transformed traffic density variable was only weakly correlated with
15 SO2 (r = 0.32) and PM2.5 (r = 0.50) in this data set. For the 1997-2001 data period (apparently
16 this was the only period in which 862 was considered), the estimated mortality relative risk for
17 SO2 was 0.99 (95%CI: 0.97, 1.01) per 5-ppb increase in a single-pollutant model. Thetwo-
18 pollutant model with the traffic density variable did not affect 862 risk estimate. Interestingly,
19 as the investigators pointed out, the risk estimates due to traffic density did not vary appreciably
20 across these four periods. They speculated that other environmental factors such as tire particles,
21 traffic noise, and spatial gradients in socioeconomic status might have been involved.
22 Lipfert et al. (2006b) further extended analysis of the veterans' cohort data to include the
23 EPA's Speciation Trends Network (STN) data, which collected chemical components of PM2.5.
24 They analyzed the STN data for year 2002, again using county-level averages. PM2.5 and
25 gaseous pollutants data for 1999 through 2001 were also analyzed. As in the previous Lipfert
26 et al. (2006a) study, traffic density was the most important predictor of mortality, but
27 associations were also seen for elemental carbon, vanadium, nickel, and nitrate. 63, NC>2, and
28 PMio also showed positive but weaker associations. The risk estimate for SO2 was essentially
29 the same as that reported in the Lipfert et al. (2006a) analysis (RR = 0.99 [95% CI: 0.96, 1.01]
30 per 5 ppb) in a single-pollutant model. Multipollutant model results were not presented for SC>2.
31
September 2007 3-98 DRAFT-DO NOT QUOTE OR CITE
-------
1 Seventh-day Adventist Study
2 Abbey et al. (1999) investigated associations between long-term ambient concentrations
3 of PMio, O3, NO2, SO2, and CO (1973 through 1992) and mortality (1977 through 1992) in a
4 cohort of 6,338 nonsmoking California Seventh-day Adventists. Monthly indices of ambient air
5 pollutant concentrations at 348 monitoring stations throughout California were interpolated to
6 ZIP code centroids according to home or work location histories of study participants,
7 cumulated, and then averaged over time. They reported associations between PMio and total
8 mortality for males and nonmalignant respiratory mortality for both sexes. 862 was not
9 associated with total (RR = 1.07 [95% CI: 0.92, 1.24] for male and 1.00 [95% CI: 0.88, 1.14]
10 for female per 5-ppb increase in multiyear average SO2), cardiopulmonary, or respiratory
11 mortality for either sex. Lung cancer mortality showed large risk estimates for most of the
12 pollutants in either or both sexes, but the number of lung cancer deaths in this cohort was very
13 small (12 for female, 18 for male) and, therefore, it is difficult to interpret these estimates.
14
15 3.4.1.2 European Cohort Studies
16 Nafstad et al. (2004) investigated the association between mortality and long-term
17 exposure to air pollution exposure in a cohort of Norwegian men followed from 1972-1973
18 through 1998. Data from 16,209 men (aged 0 to 49 years) living in Oslo, Norway, in 1972-1973
19 were linked with data from the Norwegian Death Register and with estimates of the average
20 annual air pollution levels at the participants' home addresses. PM was not considered in this
21 study because measurement methods changed during the study period. Exposure estimates for
22 nitrogen oxides (NOX) and SO2 were constructed using models based on subject addresses,
23 emission data for industry, heating, and traffic, and measured concentrations. While NOX was
24 associated with total, respiratory, lung cancer, and ischemic heart disease deaths, SO2 did not
25 show any associations with mortality. The authors noted that the SO2 levels were reduced by a
26 factor of 7 during the study period (from 5.6 ppb in 1974 to 0.8 ppb in 1995), whereas NOX did
27 not show any clear downward trend. The very low levels of 862 may be related to the lack of
28 association in this data set.
29 Filleul et al. (2005) investigated long-term effects of air pollution on mortality in 14,284
30 adults who resided in 24 areas from seven French cities when enrolled in the PAARC (Air
31 Pollution and Chronic Respiratory Diseases) survey in 1974. Daily measurements of SC>2, TSP,
32 BS, NC>2, and NO were made in the 24 areas for 3 years (1974 through 1976). Models were run
September 2007 3-99 DRAFT-DO NOT QUOTE OR CITE
-------
1 before and after exclusion of six area monitors influenced by local traffic as determined by a
2 NO:NO2 ratio of > 3. Before exclusion of the six areas, none of the air pollutants was associated
3 with mortality outcomes. After exclusion of these areas, analyses showed associations between
4 total mortality and TSP, BS, NO2, and NO but not SO2 (1.01 [95% CI: 0.97, 1.06] per 5-ppb
5 multiyear average) or acidimetric measurements. From these results, the authors noted that
6 inclusion of air monitoring data from stations directly influenced by local traffic could
7 overestimate the mean population exposure and bias the results. It should be noted that SO2
8 levels in these French cities declined markedly between the 1974 through 1976 period and the
9 1990 through 1997 period by a factor of 2 to 3, depending on the city. The changes in air
10 pollution levels over the study period complicate interpretation of reported risk estimates.
11
12 3.4.1.3 Cross-Sectional Analysis Using Small Geographic Scale
13 Elliott et al. (2007) examined associations of BS and SO2 with mortality in Great Britain
14 using a cross-sectional analysis. However, unlike the earlier ecological cross-sectional mortality
15 analyses in the United States in which mortality rates and air pollution levels were compared
16 using large geographic boundaries (i.e., MSAs or counties), in the Elliot et al. analysis, the
17 mortality rates and air pollution were compared using a much smaller geographic unit, the
18 electoral ward, with a mean area of 7.4 km2 and a mean population of 5,301 per electoral ward.
19 Death rates were computed for four successive 4-year periods from 1982 to 1994 and associated
20 with 4-year exposure periods starting in 1966 to 1970 and ending in 1990 to 1994. The number
21 of deaths from all causes in the 10,520 wards from 1982 to 1994 was 420,776. Of note, SO2
22 levels declined from 41.4 ppb in the 1966 to 1970 period to 12.2 ppb in 1990 to 1994. This type
23 of analysis does not allow adjustments for individual risk factors, but the study did adjust for
24 socioeconomic status data available for each ward from the 1991 census. Social deprivation and
25 air pollution were more highly correlated in the earlier exposure windows. They observed
26 associations for both BS and SO2 and mortality outcomes. The estimated effects were stronger
27 for respiratory illness than other causes of mortality for the most recent exposure period and
28 most recent mortality period (when pollution levels were lower). The adjustment for social
29 deprivation reduced the risk estimates for both pollutants. The adjusted mortality risk estimates
30 for SO2 for the pooled mortality periods using the most recent exposure windows were 1.021
31 (95% CI: 1.018, 1.024) for all causes, 1.015 (95% CI: 1.011, 1.019) for cardiovascular, and
32 1.064 (95% CI: 1.056, 1.072) for respiratory causes per 5-ppb increase in SO2. The effect
September 2007 3-100 DRAFT-DO NOT QUOTE OR CITE
-------
1 estimates for the most recent mortality period using the most recent exposure windows were
2 larger. Simultaneous inclusion of BS and 862 reduced risk estimates for BS but not 862. Elliott
3 et al. (2007) noted that the results were consistent with those reported in the Krewski et al.
4 (2000) reanalysis of the ACS study. This analysis was ecological, but the exposure estimates in
5 the smaller area compared to that in the U.S. cohort studies may have resulted in less exposure
6 misclassification error, and the large underlying population appears to be reflected in the narrow
7 confidence bands of risk estimates. The results from this study suggest an association between
8 long-term exposures (especially in recent years) to 862 and mortality.
9
10 3.4.1.4 Summary of Risk Estimates from Long-Term Exposure Studies
11 Figure 3.4-1 summarizes the 862 risk estimates per 5-ppb increase in the annual (or
12 longer period) average 862 for total mortality in the studies reviewed above. The overall range
13 of RRs spans 0.97 to 1.07, but considering the precision of estimates (width of confidence
14 bands), relevance to the U.S. setting, representativeness of study population, and the extent of
15 sensitivity analyses conducted of the data, the analyses of the Harvard Six Cities and the ACS
16 cohort data likely provide the risk estimates that are most useful for evaluating possible health
17 effects in the United States. This narrows the range of RRs to 1.02 to 1.07. Note that each of the
18 U.S. cohort data has its own advantages and limitations. The Harvard Six Cities data have a
19 small number of exposure estimates, but the location of the monitors were chosen carefully for
20 epidemiological purposes. The ACS cohort had far more subjects, but the population was more
21 highly educated than the representative U.S. population. Since educational status appeared to be
22 an important effect modifier of air pollution effects in both studies, the overall effect estimate for
23 the ACS cohort may underestimate that for the more general population.
24 Another important issue that these studies could not resolve was the possible confounding
25 among PM indices and SC>2. The possibility that the observed effects may not be due to SC>2, but
26 other constituents that come from the same source as SC>2, cannot be ruled out. In these long-
27 term exposure studies, the strong correlation observed between 862 and PM is mainly because
28 both SO2, which tends to be locally impacted, and SO42 , which is regionally distributed, tend to
29 be higher in the eastern United States. Despite the geographic variation in the studies, most
30 concentrated on major cities in the eastern United States, even the nationwide ACS cohort.
September 2007 3-101 DRAFT-DO NOT QUOTE OR CITE
-------
0.8
Krewski et al. (2000)—Harvard six cities -
Krewski et al. (2000); Jerrett et al. (2003)—ACS
SO2 only
SO2 only, spatial filtering
SO2 with sulfate, spatial filtering
Willis et al. (2003)—ACS
SO2 with sulfate, MSA-level
SO2 with sulfate, County-level
Pope et al. (2002)—ACS -
Lipfert et al. (2006a)—Veterans
Male
Abbey et al. (1999)—Seventh-day Adventist
Male
Female -
Nafstad et al. (2004)—Norweigian
Male
Filleul et al. (2005)—French PAAC survey -
Relative Risk
0.9 1.0 1.1
1.2
1.3
SO2 only
o SO2 with sulfate
Figure 3.4-1.
Total SOi-mortality relative risk estimates (95% CI) from
longitudinal cohort studies. Risk estimates are standardized per
5-ppb increase in SOi concentrations. The exposure estimates for the
ACS analyses in the Krewski et al. (2000) and Pope et al. (2002)
studies are based on MSA-level averaging; Lipfert et al. (2006a) used
county-level averaging.
September 2007
3-102
DRAFT-DO NOT QUOTE OR CITE
-------
1 Therefore, even with sophisticated spatial modeling, separating possible confounding of 862
2 effects by PM is challenging.
3 Finally, the extent of uncertainty related to the geographic scale used to aggregate air
4 pollution exposure estimates is not clear at this point. Willis et al. (2003) showed that the SO2
5 risk estimate based on the county-level analysis (1.02) was smaller than that from the MSA-level
6 analysis (1.04). For SC>42 , the opposite pattern was found. Thus, the impact of the geographic
7 scale of analysis may also depend on the spatial distribution of air pollutants. These
8 complications must be considered when interpretating 862 risk estimates.
9
10 3.4.2 Summary of Effects of Long-Term SOi Exposure on Mortality
11 The ecological cross-sectional studies examined in the 1982 AQCD and 1986 Secondary
12 Addendum found suggestive relationships between long-term exposure to 862 and mortality.
13 However, there were concerns as to whether the observed association was due to 862 alone,
14 because SC>42 or other sulfur agents such as H^SC^ could have been responsible. In the more
15 recent longitudinal cohort studies, once again, positive associations have been observed between
16 long-term exposure to SC>2 and mortality; however, several issues affect the interpretation of
17 these results.
18 In the limited number of available studies, the risk estimates for total mortality that are
19 most relevant to the current U.S. population ranged from 2 to 7% per 5-ppb increase in annual or
20 longer averages of 862. However, it should also be noted that several other U.S. and European
21 studies did not observe an association between long-term exposure to 862 and mortality. The
22 geographic scale of analysis appears to affect 862 risk estimates and also likely affects exposure
23 error in the analysis. In a reanalysis of the ACS data, the county-level analysis showed a smaller
24 SC>2 risk estimate than MSA-level analysis. The cross-sectional analysis in Great Britain using
25 small-scale electoral wards observed a risk estimate similar to the lower end of the range of risk
26 estimates for all-cause mortality from U.S. cohort studies, though it is not clear if the risk
27 estimates from this cross-sectional study are directly comparable to those from cohort studies.
28 In the long-term studies of the ACS cohort, the peculiar spatial pattern of the
29 concentration of major cities and 862 sources in the eastern United States dominated the
30 overall mortality associations with PM2.5, SC>42 , and SC>2, making it difficult to separate out
31 the potential confounding or mixture effects. Future and on-going studies that take into
September 2007 3-103 DRAFT-DO NOT QUOTE OR CITE
-------
1 consideration within- versus between-city variation of these air pollutants may help resolve
2 this issue.
3 Overall, the results from two major U.S. epidemiological studies suggest an association
4 between long-term exposure to SC>2 or sulfur-containing particulate air pollution. However, it
5 should be noted that authors of the reanalyses of these studies concluded that in the absence of a
6 plausible toxicological mechanism by which SC>2 could lead to increased mortality suggested that
7 SC>2 might be acting as a marker for other mortality-associated pollutants (Krewski et al., 2000).
8 The inability to distinguish potential confounding by copollutants, uncertainties regarding the
9 geographic scale of analysis and copollutant confounding limit the interpretation of a causal
10 relationship.
September 2007 3-104 DRAFT-DO NOT QUOTE OR CITE
-------
i 4. PUBLIC HEALTH IMPACT
2
O
4 This chapter addresses several issues relating to the broader public health impact from
5 exposure to ambient sulfur oxides (SOX). First, the shape of the concentration-response
6 relationship for sulfur dioxide (802) is discussed and the evidence for a threshold value for
7 health effects is evaluated. The next section identifies characteristics of subpopulations which
8 may experience increased risks from SC>2 exposures, through either enhanced susceptibility (e.g.,
9 as a result of age, pre-existing disease, genetic factors) and/or differential vulnerability
10 associated with increased exposure owing to close proximity to sources, for example. The final
11 section defines adverse health effects associated with 862 and classifies them according to
12 severity for individuals with impaired respiratory systems. The prevalence of such respiratory
13 disorders in the U.S. population is considered to assess the impact of ambient SO2 exposure on
14 public health.
15
16
17 4.1 ASSESSMENT OF CONCENTRATION-RESPONSE FUNCTION
18 AND POTENTIAL THRESHOLDS
19 An important consideration in characterizing the public health impacts associated with
20 SC>2 exposure is whether the concentration-response relationship is linear across the full
21 concentration range that is encountered or if there are concentration ranges where there are
22 departures from linearity (i.e., nonlinearity). Of particular interest is the shape of the
23 concentration-response curve at and below the level of the current SC>2 NAAQS level of a 24-h
24 average level of 0.14 ppm or the annual average of 0.03 ppm.
25 Identifying possible "thresholds" in air pollution epidemiological studies is challenging.
26 A threshold in this case would be defined as the level of SC>2 that must be exceeded to elicit a
27 health response. Low data density in the lower concentration range, measurement error in the
28 response, exposure measurement error, and a shallow slope near the threshold are some of the
29 error sources that complicate determining the shape of the concentration-response curve.
30 Biological characteristics that tend to linearize concentration-response relationships include
31 individual differences in susceptibility to air pollution health effects additivity of SCVinduced
32 effects to a naturally occurring background level and additivity of effects from other pollutant
September 2007 4-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 exposures. With consideration of these limitations, epidemiological and human clinical studies
2 that examined the shape of the concentration-response function for different averaging times or
3 exposure durations are presented below. The discussion focuses on respiratory morbidity effects
4 associated with short-term exposure to SO2, for which the strongest causal evidence exists.
5 Evidence from human clinical studies provides insights into the discussion of the
6 concentration-response function and possible thresholds of SC>2 health effects. In human clinical
7 studies, significant effects have been observed at the lowest levels tested (i.e., 0.2 to 0.25 ppm, 5-
8 to 15-min exposures) in some sensitive individuals; however, there was great interindividual
9 variability in the observed SCVrelated responses. Human clinical studies largely examined the
10 effects of peak exposures (< 1 h, typically 5 or 15 min) to SC>2 on respiratory health. The
11 majority of these studies have involved short-term exposures of asthmatic adults to varying
12 concentrations of SC>2 while they perform light to heavy levels of exercise. Sheppard et al.
13 (1981) reported a significant SC>2-induced increase in specific airways resistance (sRaw)
14 compared to filtered-air exposures among asthmatic adults following 10-min exposures to SC>2
15 with moderate exercise at concentrations as low as 0.25 ppm. Doubling the exposure
16 concentration of 862 (0.5 ppm) increased sRaw by approximately 75% compared to the average
17 value of sRaw following exposure to 0.25-ppm 862. In a similar study, Linn et al. (1982) found
18 no significant SCVattributable increase in sRaw following 1-h exposures to 0.25- and 0.5-ppm
19 SO2 among asthmatics engaged in 10-min intervals of moderate levels of exercise. The authors
20 suggested that the apparent contrast between their findings and those of Sheppard et al. (1981)
21 may be explained by differences in the exposure methods used. Linn et al. (1982) conducted
22 exposures in a chamber, allowing normal oronasal breathing, while Sheppard et al. (1981)
23 conducted exposures through a mouthpiece, which likely resulted in a relative increase of
24 pulmonary 862 uptake.
25 Linn et al. (1983) later evaluated the concentration-response relationship between 862
26 and respiratory effects following 5-min exposures to 0-, 0.2-, 0.4-, and 0.6-ppm 862 during
27 heavy exercise (minute ventilation [VE] ~ 48 L/min). The results appeared to demonstrate an
28 increase in sRaw attributable to increasing SC>2 concentrations. However, only exposures to
29 0.4 and 0.6 ppm were found to significantly differ from the control (0-ppm 802). Schachter
30 et al. (1984) found the SO2-induced respiratory response to be highly variable among asthmatic
31 adults. These investigators exposed asthmatic and healthy, non-asthmatic subjects to 0-, 0.25-,
September 2007 4-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 0.5-, 0.75-, and 1.0-ppm 862 for 10 min during moderate levels of exercise. No SCVassociated
2 respiratory effects were observed in the healthy, non-asthmatics at any of the exposure
3 concentrations. While some asthmatic subjects exhibited a decrease in forced expiratory volume
4 in 1 s (FEVi) beginning at concentrations as low as 0.25 ppm, a consistent and significant
5 reduction in FEVi compared to baseline was not observed at levels <0.75-ppm SC>2. Finally, in a
6 study involving SO2-sensitive asthmatics, Gong et al. (1995) observed a linear relationship
7 between SC>2 concentration (0-, 0.5-, and 1.0-ppm) and both lung function (decrease in FEVi,
8 and increase in sRaw) and respiratory symptoms. The evidence from human clinical studies
9 demonstrates consistent SCVinduced respiratory effects following 5-to 15-min exposures of 862
10 at levels between 0.5 and 1.0 ppm, with weaker evidence of effects at concentrations as low as
11 0.25 ppm in some sensitive asthmatics.
12 Epidemiological studies have examined the concentration-response relationship for SC>2
13 using various statistical methods, including the comparison of effect estimates in increasing
14 quartiles or quintiles, plotting the risk observed against increasing SC>2 concentrations, and using
15 nonparametric smoothed curves to assess the nonlinearity of the SO2-effect relationship. Most of
16 the epidemiological studies that examined the concentration-response function between 862
17 exposure and respiratory morbidity observed that the relationship was linear across the entire
18 concentration range, as discussed below.
19 Only one epidemiological study investigated the concentration-response function of peak
20 SO2 exposures. The association between asthma hospitalizations and ambient 1-h maximum
21 (1-h max) SC>2 concentrations was examined in a case-control study of children in Bronx County,
22 NY (Lin et al., 2004). The 1-h max concentration ranged from 2.9 to 66.4 ppb. Lin et al.
23 categorized 1-h max SC>2 concentrations and estimated odds ratios (ORs) for each category using
24 the lowest exposure group as the reference (2.9 to 9.2 ppb). They observed an increasing linear
25 trend across the range of concentrations, suggesting that there was no threshold in the observed
26 association between asthma hospitalizations and 1-h max 862 concentrations.
27 Most epidemiological studies investigating the concentration-response function examined
28 the effects of short-term 24-h average exposures to SO2. A study by Jaffe et al. (2003) examined
29 the association between SC>2 and emergency department (ED) visits for asthma in three cities in
30 Ohio, i.e., Cincinnati, Cleveland, and Columbus. The mean 24-h average SO2 concentrations
31 were 14 ppb (range: 1, 50) in Cincinnati, 15 ppb (range: 1, 64) in Cleveland, and 4 ppb (range:
September 2007 4-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 0, 22) in Columbus. Significant associations were observed only in Cincinnati using the Poisson
2 regression analysis. To examine the concentration-response function, they also conducted
3 quintile analyses. In Cincinnati, an increasing linear trend in risk was observed across the range
4 of concentrations.
5 Wong et al. (2002; using Generalized Additive Model(s) (GAM) with default
6 convergence criteria) observed that ambient SO2 concentrations were associated with hospital
7 admissions for respiratory causes in adults aged 65+ years in Hong Kong (mean 24-h average
8 SO2 of 7 ppb [range: 0, 34]), but not London (mean 24-h average SCh of 9 ppb [range: 2, 43]).
9 A plot of risk against 24-h average 862 concentrations was constructed to examine the
10 concentration-response relationship in these cities. In general, a linear relationship between risk
11 of respiratory hospitalizations and SO2 was observed across the range of SO2 concentrations in
12 Hong Kong.
13 Burnett et al. (1997a,b) examined the relationship between adjusted hospital admission
14 rates for respiratory diseases and ambient SC>2 concentrations for nonlinearity. A nonparametric
15 smoothed curve using locally estimated smoothing splines (LOESS) was applied in the Toronto
16 study (mean 1-h max 862 of 7.9 ppb [range: 0, 26]), while cubic polynomials and quadratic
17 polynomials were fitted to the data in the 16 Canadian cities study (mean 1-h max 862 of
18 14.4 ppb [90th percentile: 97]. In no case did the results suggest that the association between
19 respiratory hospitalizations and 862 deviated from linearity.
20 Additional European studies by Atkinson et al. (1999a), Hajat et al. (1999) and Hajat
21 et al. (2002; using GAM with default convergence criteria) also did not find a threshold in the
22 response. In Atkinson et al. (1999a), the bubble plot indicated that the relationship between
23 hospital admission for respiratory causes and SO2 concentrations was approximately linear.
24 Hajat et al. (1999, 2002) reported a generally linear association between SO2 concentrations and
25 general practitioner visits for lower and upper respiratory conditions, using a bubble plot and a
26 concentration-response plot across the range of 862 concentrations.
27 However, some studies did report a nonlinear relationship between 862 and respiratory
28 health effects. The Harvard Six Cities study by Schwartz et al. (1994) examined the effects of
29 summertime air pollution on the respiratory health of 1,844 schoolchildren. The median 24-h
30 average SC>2 concentration during the study period was 4.1 ppb (10th-90th percentile: 0.8, 17.9,
31 maximum 81.9). While SC>2 concentrations were found to be associated with increased cough
September 2007 4-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 incidence and lower respiratory symptoms, the relationship was nonlinear. A figure plotting the
2 relative odds of incidence of lower respiratory symptoms against 862 concentrations lagged
3 1 day indicated that no statistically significant increase in the incidence of lower respiratory
4 symptoms was seen until concentrations exceeded a 24-h average SCh of 22 ppb (see Figure 3.1-
5 2 in Section 3.1), though an increasing trend was observed at concentrations as low as a 24-h
6 average SO2 of 10 ppb.
7 Using time-series data, Ponce de Leon et al. (1996) studied the association between
8 hospitalizations for respiratory causes and ambient 862 concentrations in London. The mean
9 24-h average 862 concentration was 32.2 ppb (5th-95th percentile: 6,21). Bubble plots of
10 adjusted residuals of log admission counts sorted by 862 level indicated that a weak relationship
11 with SO2 was only observable at 24-h average SO2 concentrations above 23 ppb. In both this
12 study and the study by Schwartz et al. (1994), a statistically significant increased risk was
13 observable only at 24-h average SC>2 concentrations that were above the 90th percentile. These
14 possible threshold values are dependent on only a few influential observations; so the results
15 should be viewed with caution.
16 As discussed earlier in this section, many factors may obscure the presence of thresholds
17 in epidemiological studies at the population level. Using fine particulate matter (PM2.5) as an
18 example, Brauer et al. (2002) examined the relationship between ambient concentrations and
19 mortality risk in a simulated population with specified common individual threshold levels.
20 They found that no population threshold was detectable when a low threshold level was
21 specified. Even at high-specified individual threshold levels, the apparent threshold at the
22 population level was much lower than specified. Brauer et al. (2002) concluded that the use of
23 surrogate measures of exposure (i.e., those from centrally located ambient monitors) that were
24 not highly correlated with personal exposures obscured the presence of thresholds in
25 epidemiological studies at the population level even if a common threshold exists for individuals
26 within the population.
27 The wide interindividual variability in sensitivity to SC>2 exposure further hinders the
28 ability to find a threshold level in population studies. Human clinical studies have shown that
29 asthmatics experience greater increases in sRaw following peak SC>2 exposures compared to
30 healthy individuals (Linn et al., 1987). Amongst asthmatics, interindividual differences in
September 2007 4-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 response also have been noted, with some asthmatics experiencing SCVrelated effects at much
2 lower levels than others (Horstman et al., 1986).
3 Another factor that complicates the identification of a possible threshold of effects is that
4 currently deployed ambient monitors may be inadequate for accurate and precise measurements
5 at lower 24-h average SO2 levels. Ambient concentrations of SO2 have been declining since the
6 1980s and are now at or very near the limit of detection of the ambient monitors in the regulatory
7 network. The mean 24-h average SC>2 concentration across the metropolitan statistical areas
8 (MSAs) from 2003 through 2005 was 4 ppb (5th-95th percentile: 1, 13). Thus, there is greater
9 uncertainty at the lower concentration range compared to the higher concentrations, which likely
10 limits the ability to detect a threshold.
11 In conclusion, evidence from human clinical studies indicated wide interindividual
12 variability in response to SC>2 exposures, with peak (5 to 10 min) exposures at levels as low as
13 0.25 ppm eliciting respiratory responses in some asthmatic individuals. Several epidemiological
14 studies that examined the concentration-response function between short-term (24-h average)
15 exposure to SC>2 and respiratory morbidity observed that the relationship was linear across the
16 entire concentration range, suggesting a lack of a threshold in effect. However, given the various
17 limitations in observing a possible threshold in population studies, the lack of evidence does not
18 necessarily indicate that there is indeed no threshold in 862 health effects. Two epidemiological
19 studies did report a possible threshold level of 22 to 23 ppb (24-h average) at which no
20 statistically significant SO2-related respiratory health effect was observed. However, as these
21 observations were based on only a few influential data points (24-h average SC>2 concentrations
22 above the 90th percentile), the results should be viewed with caution. The overall limited
23 evidence from epidemiological studies examining the concentration-response function of SC>2
24 health effects is inconclusive regarding the presence of an effect threshold.
25
26
27 4.2 SUSCEPTIBLE AND VULNERABLE POPULATIONS
28 The previous review of the SC>2 NAAQS identified certain groups within the population
29 that may be more susceptible to the effects of SC>2 exposure, including asthmatics, individuals
30 not diagnosed as asthmatic but with atopic disorders (e.g., allergies), and individuals with
31 chronic obstructive pulmonary disease (COPD) or cardiovascular disease (CVD). Other
32 subgroups that were considered to be somewhat sensitive included children and the elderly.
September 2007 4-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 Many factors such as age, preexisting disease, gender, nutritional status, smoking, and
2 genetic variability may contribute to the interindividual variability in responses to environmental
3 pollutants, including SO2. Individuals in potentially sensitive groups are of concern, as they may
4 be affected by lower levels of SO2 than the general population or experience a greater impact
5 with the same level of exposure. This section focuses on vulnerable groups that may be exposed
6 to SC>2 levels above the community average and differential effects among susceptible groups,
7 including subpopulations with asthma as well as genetic factors and age-related variability.
8
9 4.2.1 Exposure of Susceptible and Vulnerable Populations to SO2
10 A limited amount of information exists on exposures of susceptible and vulnerable
11 populations to SC>2. Indoor and personal SC>2 concentrations are generally much lower than
12 outdoor or ambient measurements and occur near or below the detection limit of the passive
13 samplers used in most studies (Kindzierski and Ranganathan, 2006; Sarnat et al., 2000, 2005,
14 2006). Infiltration of 862 into residences is limited (Brauer et al., 1989), partially due to its
15 reactivity with the building envelope and indoor surfaces. Contributions of indoor sources to
16 personal SC>2 exposures are low, with the possible exception of SC>2 emitted from indoor
17 kerosene or gas heaters (Triche et al., 2005). Hence, individuals that spend most of their time
18 indoors, such as older adults, are not anticipated to be vulnerable to high SC>2 exposures, though
19 in some cases they may be more susceptible to the effects of these exposures than the general
20 population. Other individuals with increased vulnerability include those who spend a lot of time
21 outdoors at increased exertion levels, for example outdoor workers and individuals who exercise
22 or play sports. Children, who generally spend more time playing outdoors, may qualify as both a
23 susceptible population due to their developing physiology and as a vulnerable population since
24 ambient SC>2 concentrations are several-fold higher than indoor concentrations.
25 Residential location is not as strong of a predictor of exposure vulnerability for SC>2 as for
26 traffic-related pollutants, because meteorological conditions have a greater impact on pollutant
27 plume direction from primary point sources such as coal-fired power plants.
28
29 4.2.2 Preexisting Disease as a Potential Risk Factor
30 Several researchers have investigated the effect of air pollution among potentially
31 susceptible groups with preexisting medical conditions. A recent report of the National Research
32 Council emphasized the need to evaluate the effect of air pollution on susceptible groups,
September 2007 4-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 including those with respiratory illnesses and CVD (NRC 2004). Generally, asthma, COPD,
2 conduction disorders, congestive heart failure (CHF), diabetes, and MI are conditions believed to
3 put persons at greater risk of adverse events associated with air pollution. Asthmatics are known
4 to be one of the most SO2-responsive subgroups in the population; the evidence related to
5 respiratory illness, including asthma, is discussed in further detail below.
6
7 4.2.2.1 Individuals with Respiratory Diseases
8 The 1982 Air Quality Criteria Document (AQCD) concluded that asthmatics are likely
9 more susceptible to effects from SC>2 exposures than the general public. Recent epidemiological
10 studies have strengthened this conclusion, reporting associations between a range of health
11 outcomes with both short-term and long-term 862 exposures in subjects with respiratory disease.
12 In controlled human exposure studies, asthmatics have been shown to be more responsive
13 to respiratory effects of SO2 exposures than healthy, non-asthmatics. While SO2-attributable
14 decrements in lung function have not generally been demonstrated at concentrations <1.0 ppm in
15 non-asthmatics (Lawther et al., 1975; Linn et al., 1987; Schachter et al., 1984), increases in
16 respiratory symptoms and decreases in lung function have been observed in some asthmatics
17 following peak (5 to 15 min) 862 exposures to concentrations <0.5 ppm (Gong et al., 1995;
18 Horstman et al., 1986; Linn et al., 1983).
19 A number of epidemiological studies reported increased respiratory symptoms associated
20 with SC>2 exposures in asthmatics and atopic individuals. In contrast, lung function generally
21 was not positively associated with ambient SO2 in epidemiological studies of asthmatic children
22 (Mortimer et al., 2002) or among adults with asthma or COPD (Higgins et al., 1995; Neukirch
23 et al., 1998; van der Zee et al., 2000). A series of epidemiological studies from the Netherlands
24 has investigated the effect of exposure to SC>2 and other air pollutants on children and adults with
25 airways hyperreactiveness (AHR) and atopy. In 1998, Boezen et al. found that adults with
26 airway lability (defined as the presence of AHR or an increase in peak expiratory flow [PEF]
27 variability) had a significantly increased prevalence of respiratory symptoms, including lower
28 and upper respiratory symptoms, cough, and phlegm, with increasing levels of 862. In
29 subsequent analyses, the authors examined whether children with AHR and elevated levels of
30 IgE were vulnerable to the effects of SO2 (Boezen et al., 1999). In a panel study of children aged
31 7 to 11 years, the authors found no associations between SC>2 and any respiratory symptoms in
September 2007 4-8 DRAFT-DO NOT QUOTE OR CITE
-------
1 the subset of children with relatively low serum total IgE with or without AHR. However, for
2 children with relatively high serum total IgE either with or without AHR, the prevalence of lower
3 respiratory symptoms increased in relation to increasing SO2 concentrations. In a similar study
4 conducted among older adults aged 50 to 70 years, Boezen et al. (2005) found that the subgroup
5 of individuals with elevated serum total IgE, both with and without AHR, to be more susceptible
6 to air pollutants compared to those who did not have elevated serum total IgE. Significant
7 associations were observed between previous-day 24-h average SC>2 concentrations and the
8 prevalence of upper respiratory symptoms in those with elevated serum total IgE. Stratified
9 analyses by gender indicated that among those with AHR and elevated IgE, only males were at a
10 higher risk for respiratory symptoms.
11 In a German study of 5,421 children, the annual mean SO2 concentration was associated
12 with morning cough over the previous 12 months, but not bronchitis (Hirsch et al., 1999). In
13 contrast to the results reported by Boezen et al. (1999, 2005), this study observed that the
14 association of SO2 and other air pollutants with respiratory symptoms were stronger in nonatopic
15 than in atopic children. The authors noted that these findings were in line with the hypothesis
16 that these air pollutants induce nonspecific irritative rather than allergic inflammatory changes in
17 the airway mucosa, as irritative effects would affect the clinical course in nonatopic children
18 more strongly than in atopies whose symptoms are also determined by allergen exposure.
19 U.S. multicity studies of ambient 862 exposure also examined respiratory symptoms in
20 asthmatic children (Mortimer et al., 2002; Schildcrout et al., 2006). In the National Cooperative
21 Inner-City Asthma Study (NCICAS; Mortimer et al., 2002), the greatest effect was seen for
22 morning symptoms, i.e., cough, wheeze, shortness of breath (range of median 3-h average SC>2
23 across eight cities: 17to37ppb). In the Childhood Asthma Management Program (CAMP)
24 study (Schildcrout et al., 2006), the strongest association between SC>2 and increased asthma
25 symptoms was found for a 3-day moving average lag (range of median 24-h average 862 across
26 seven cities: 2.2, 7.4 ppb [90th percentile range: 4.4, 14.2]). The Harvard Six Cities Study
27 found an association between 862 concentration and cough incidence and lower respiratory
28 symptoms among healthy children, but suggested that the relationship was nonlinear, with
29 increased risk only observed at levels >20 ppb (Schwartz et al., 1994). These studies indicate
30 that SO2 effects on respiratory symptoms were observed in asthmatics at lower ambient levels of
31 SC>2 compared to healthy children.
September 2007 4-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 Other studies have examined the relationship between respiratory symptoms among
2 subpopulations with asthma and/or COPD and ambient 862 concentrations. These studies
3 generally indicated positive associations for asthma among children and included a U.S. study
4 (Delfino et al., 2003) and several European studies (Higgins et al., 1995; Neukirch et al., 1998;
5 Peters et al., 1996; Roemer et al., 1993; Segala et al., 1998; Taggart et al., 1996; Timonen and
6 Pekkanen, 1997; van der Zee et al., 1999). Studies of asthma among adults found no consistent
7 association between respiratory symptoms among asthmatics and SC>2 concentrations
8 (Desqueyroux et al., 2002a,b; Romieu et al., 1996; van der Zee et al., 2000).
9 A suggestive association between ambient 862 concentrations and ED visits and
10 hospitalizations among children and the elderly provides evidence that asthmatics are susceptible
11 to the effects of SO2. The associations between ambient concentrations of 24-h average SO2 and
12 ED visits and hospitalizations for asthma in the United States are generally positive (Jaffe et al.,
13 2003; Lin et al., 2004a; Michaud et al., 2004; Wilson et al., 2005), though a large time-series
14 study conducted in Atlanta, GA did not find an association between ambient 1-h max SC>2 levels
15 and ED visits (Peel et al., 2005). Studies conducted outside the United States (Atkinson et al.,
16 1999b; Hajat et al., 1999; Sunyer et al., 1997; Thompson et al., 2001) also generally found
17 positive results.
18 There was no association between 862 levels and asthma mortality in the general
19 population (Saez et al., 1999) or among patients previously diagnosed with severe asthma
20 (Sunyer et al., 2002).
21 In summary, there is substantial evidence from epidemiological studies that suggests that
22 individuals with preexisting respiratory diseases, particularly asthma, are more susceptible to
23 respiratory health effects, though not mortality, from SC>2 exposures than the general public. The
24 observations in human clinical studies of increased sensitivity to SC>2 exposures in asthmatic
25 subjects compared to healthy subjects provide coherence and biological plausibility for these
26 observations in epidemiological studies.
27
28 4.2.2.2 Individuals with CVDs
29 Routledge et al. (2006) exposed patients with stable angina as well as healthy subjects to
30 50-|ig/m3 carbon particles and 0.2-ppm SO2, alone and in combination for 1 h. Heart rate
31 variability (HRV), C-reactive protein, and markers of coagulation were measured. There was no
32 evidence to suggest that patients with stable angina were more susceptible to SC>2-related health
September 2007 4-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 effects than healthy subjects. The authors noted that this lack of response in the heart patients
2 may be due to a drug treatment effect rather than decreased susceptibility, as a large portion of
3 the angina patients were taking beta blockers, which are known to increase indices of cardiac
4 vagal control.
5 Liao et al. (2004) investigated short-term associations between ambient pollutants and
6 cardiac autonomic control. Resting, supine, 5-min beat-to-beat R-R interval data were collected.
7 Previous-day SC>2 concentrations were positively associated with heart rate and inversely
8 associated with the standard deviation of normal R-R intervals (SDNN) and low frequency (LF)
9 power. Consistently more pronounced associations were suggested between 862 and HRV
10 among persons with a history of coronary heart disease.
11 Henneberger et al. (2005) examined the association of repolarization parameters with air
12 pollutants in men with preexisting coronary heart disease in East Germany. The patients were
13 examined repeatedly once every 2 weeks for 6 months, for a total of 12 electrocardiogram (ECG)
14 recordings. The mean 24-h average SC>2 concentration was 4.1 |ig/m3 (2 ppb [range: 1, 4]).
15 Ambient SC>2 concentrations during the 24-h preceding the ECG were associated with the QT
16 interval duration, but not with any other repolarization parameters. Stronger associations were
17 observed between PM indices and QT interval duration, T-wave amplitude, and T-wave
18 complexity.
19 Evidence from ED visit and hospitalizations studies of the association between ambient
20 levels of air pollutants and CVD is inconsistent. A recent epidemiological study investigated the
21 association of SC>2 with cardiac hospital admissions among persons with preexisting
22 cardiopulmonary conditions and observed no associations with ambient 1-h max SC>2 level for
23 any cardiac disease investigated (i.e., ischemic heart disease [IHD], CHF, and dysrhythmia)
24 across strata of comorbid disease status, including hypertension, diabetes, and COPD (Peel
25 et al., 2007).
26 Goldberg et al. (2003) compared the risk estimates for death with the underlying cause of
27 CHF and those deaths classified as having CHF 1 year before death and did not find associations
28 between air pollution and those with CHF as an underlying cause of death. The authors found
29 associations between some of the air pollutants examined (coefficient of haze [CoH], SC>2, and
30 NC>2) and the deaths that were classified as having CHF 1 year before death, but the association
31 with the specific cause of death was not unique to SC>2. This pattern of association, including but
September 2007 4-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 not specific to 862, with specific causes of death also was observed in an additional cohort of
2 patients with CHF (Kwon et al., 2001).
3 In summary, there is weak evidence from a small number of panel studies that suggests
4 that individuals with preexisting CVD may be more susceptible to adverse health effects from
5 ambient SC>2 exposures than the general public. The evidence from one human clinical study
6 does not support these conclusions. Additional research is necessary to assess whether
7 individuals with preexisting CVD constitute a susceptible group for SC>2 health effects.
8
9 4.2.3 Age-Related Variations in Susceptibility
10 The American Academy of Pediatrics (2004) notes that children and infants are among
11 the most susceptible to many air pollutants, including SC>2. Eighty percent of alveoli are formed
12 postnatal and changes in the lung continue through adolescence; furthermore, the developing
13 lung is highly susceptible to damage from exposure to environmental toxicants (Dietert et al.,
14 2000). Children also have increased vulnerability as they spend more time outdoors, are highly
15 active, and have high minute ventilation, which collectively increase the dose they receive
16 (Plunkett et al., 1992; Wiley et al., 1991a,b). In addition to children, the elderly are frequently
17 classified as being particularly susceptible to air pollution. The basis of the increased sensitivity
18 in the elderly is not known, but one hypothesis is that it may be related to changes in the
19 respiratory tract lining fluid antioxidant defense network (Kelly et al., 2003).
20 A number of studies investigating the association between ambient 862 levels and ED
21 visits or hospital admissions for all respiratory causes or asthma stratified their analyses by age
22 group. Figure 4.2-1 summarizes the evidence of age-specific associations between 862 and
23 acute respiratory ED visits and hospitalizations. Several studies demonstrated that the risk of ED
24 visits or hospitalizations for all respiratory causes or asthma associated with a 10-ppb increase in
25 24-h average SC>2 levels was higher for children (Anderson et al., 2001; Atkinson et al., 1999a,b;
26 Petroeschevsky et al., 2001; Ponce de Leon et al., 1996) and older adults (Anderson et al., 1998;
27 Petroeschevsky et al., 2001; Ponce de Leon et al., 1996; Wilson et al., 2005) when compared to
28 the risk for all ages together. Increased risks for children and older adults were more prevalent in
29 the studies of all respiratory disease than those considering asthma as the outcome. Two studies
30 investigated the association between ambient 862 levels and ED visits or hospital admissions for
31 all cardio vascular causes, with analyses stratified by age (Atkinson et al., 1999a,b; Sunyer et al.,
September 2007 4-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 2003b). Neither of these studies found a difference in effect estimates when analyses were
2 stratified to ages 65+ years, compared to when all ages were included in the analyses.
Relative Risk
0.8 0.9 1,0 1.1 1.2 1.3
Reference Location
Wilson et al. (2005) Manchester, NH
Atkinson et al. (1999a) London, UK
Atkinson et al. (1999 b) London UK
Ponce de Leon et al. (1996) London, UK
Anderson et al. (2001 ) West Midlands, UK
Petroeschevs ky et al. (200 1 ) Bri s ba ne, Austra I ia
Wilson et al (2005) Portland, ME
Wilson etal (2005) Manchester, NH
Atkinson et al (1999a) London, UK
Atkinson etal. (1999b) London, UK
Anderson et al, { 1 998 ) London, UK
Petroeschevsky et al. [2001 ) Brisbane, Australia
i
0
i i
| All Respiratory |
•
. — • —
-+-
•u —
-o
-• —
°~
„ 1 Asthma 1
i 1
*
0
J£
• All ages
0 0-14 years
0 • 65+ years
Figure 4.2-1.
Relative risks (95% CI) of age-specific associations between short-
term exposure to SO2 and respiratory ED visits and hospitalizations.
Risk estimates are standardized per 10-ppb increase in 24-h average
concentrations or 40-ppb increase in 1-h max
September 2007
4-13
DRAFT-DO NOT QUOTE OR CITE
-------
1 Cakmak et al. (2007) reported that among seven Chilean urban centers, the percent
2 increase in nonaccidental mortality associated with a 10-ppb increase in 24-h average SC>2 was
3 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
4 years of age. The authors concluded that the elderly are particularly susceptible to dying from
5 air pollution, and suggested that concentrations deemed acceptable for the general population
6 may not adequately protect the very elderly.
7 There is limited epidemiological evidence to suggest that children and older adults
8 (65+ years) are more susceptible to the adverse respiratory effects associated with ambient SCh
9 concentrations when compared to the general population. The few studies that conducted age-
10 stratified analyses when examining cardiovascular outcomes did not find any difference in
11 outcomes when analyses were stratified by age.
12
13 4.2.4 Genetic Factors for Oxidant and Inflammatory Damage from Air
14 Pollutants
15 A consensus now exists among scientists that genetic factors related to health outcomes
16 and ambient pollutant exposures merit serious consideration (Gilliland et al., 1999; Kauffmann
17 et al., 2004). Several criteria must be satisfied in selecting and establishing useful links between
18 polymorphisms in candidate genes and adverse respiratory effects. First, the product of the
19 candidate gene must be significantly involved in the pathogenesis of the adverse effect of
20 interest, which is often a complex trait with many determinants. Second, polymorphisms in the
21 gene must produce a functional change in either the protein product or in the level of expression
22 of the protein. Third, in epidemiological studies, the issue of confounding by other genes or
23 environmental exposures must be carefully considered.
24 Several glutathione S-transferase (GST) families have common, functionally important
25 polymorphic alleles (e.g., homozygosity for the null allele at the GSTM1 and GSTT1 loci,
26 homozygosity for the A105G allele at the GSTP1 locus) that significantly reduce expression of
27 function in the lung. Exposure to radicals and oxidants in air pollution induces decreases in GSH
28 that increase GST transcription. Individuals with genotypes that result in enzymes with reduced
29 or absent glutathione peroxidase activity are likely to have reduced oxidant defenses and
30 increased susceptibility to inhaled oxidants and radicals.
September 2007 4-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 Gilliland et al. (2002) examined effects of GSTM1, GSTT1, and GSTP1 genotypes and
2 acute respiratory illness, specifically respiratory illness-related absences from school. The goal
3 was to examine potential susceptibilities on this basis, but not specifically to air pollutants. They
4 concluded that fourth grade schoolchildren who inherited a GSTP1 Val-105 variant allele had a
5 decreased risk of respiratory illness-related school absences, indicating that GSTP1 genotype
6 influences the risk and/or severity of acute respiratory infections in school-aged children.
7 Lee et al. (2004) studied ninth grade schoolchildren with asthma in Taiwan for a gene-
8 environmental interaction between GSTP1-105 genotypes and outdoor pollution. They
9 examined general district air pollution levels of low (mean SO2 level of 3.6 ppb from 1994 to
10 2001), moderate (mean 862 of 6.2 ppb), and high (mean 862 of 8.6 ppb) in the analysis and
11 found that compared with individuals with any Val-105 allele in the low air pollution district,
12 Ile-105 homozygotes in the high air pollution district had a significantly increased risk of
13 asthma.
14 Gauderman et al. (2007) describe a study method that uses principal components analysis
15 computed on single nucleotide polymorphism (SNP) markers to test for an association between a
16 disease and a candidate gene. For example, they evaluated the association between respiratory
17 symptoms in children and four SNPs in the GSTP1 locus, using data from the Southern
18 California Children's Health Study (CHS). The authors observed stronger evidence of an
19 association using the principal components approach (p = 0.044) than using either a genotype-
20 based (p = 0.13) or haplotype-based (p = 0.052) approach. This method may be applied to
21 relationships in this and other databases to evaluate aspects of air pollutants such as SC>2.
22 In 2001, Winterton et al. (2001) attempted to identify a genetic biomarker for
23 susceptibility to SC>2. They screened 62 asthmatic subjects for SC>2 responsiveness using an
24 inhalation challenge and collected genetic material via buccal swabs to test for associations
25 between 862 sensitivity and specific gene polymorphisms. Subjects inhaled 0.5-ppm 862 by
26 mouthpiece for 10 min while wearing noseclips during moderate exercise on a treadmill.
27 Subjects were defined as SO2-sensitive if FEVi dropped > 12%. Genetic polymorphisms as
28 biomarkers of susceptibility were evaluated in five regions coding for the p2-adrenergic receptor,
29 the a subunit of the interleukin-4 (IL-4) receptor, the Clara cell secretory protein (CC16), tumor
30 necrosis factor-a (TNF-a), and lymphotoxin-a (also known as TNF-P). The authors found a
31 significant association between response to SC>2 and the homozygous wild-type allele of TNF-a.
September 2007 4-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 All of the SCVsensitive subjects had the homozygous wild-type allele for TNF-a, while 61% of
2 the nonresponders had this genotype. Homozygosity for the TNF-1 allele was associated with a
3 5-fold increased risk of physician-diagnosed asthma relative to other genotypes. None of the
4 other polymorphisms showed significant trends.
5 In summary, the differential effects of air pollution among genetically diverse
6 subpopulations have been examined for a number of GST genes and other genotypes. The
7 limited number of studies may provide some insight into susceptible groups and a potential
8 genetic role in such. Only one of these studies specifically examined 862 as the exposure of
9 interest, and it found a significant association with the homozygous wild-type allele for TNF-a.
10 Khoury et al. (2005) states that while genomics is still in its infancy, opportunities exist for
11 developing, testing, and applying its tools to public health research of outcomes with possible
12 environmental causes. At this time, there are only very limited data on which to base a
13 conclusion regarding the effect of SC>2 exposure on genetically distinct subpopulations.
14
15
16 4.3 POTENTIAL PUBLIC HEALTH IMPACTS
17 Exposure to ambient 862 is associated with a variety of outcomes including increases in
18 respiratory symptoms, particularly among asthmatic children, and ED visits and hospital
19 admissions for respiratory diseases among children and older adults (65+ years). In protecting
20 public health, a distinction must be made between health effects that are considered "adverse"
21 and those that are not. What constitutes an adverse health effect varies for different population
22 groups. Some changes in healthy individuals are not viewed as adverse while those of similar
23 type and magnitude in other susceptible individuals with preexisting disease are.
24
25 4.3.1 Concepts Related to Defining Adverse Health Effects
26 The American Thoracic Society (ATS) published an official statement titled "What
27 Constitutes an Adverse Health Effect of Air Pollution?" (ATS, 2000). This statement updated
28 the guidance for defining adverse respiratory health effects that had been published 15 years
29 earlier (ATS, 1985), taking into account new investigative approaches used to identify the effects
30 of air pollution and reflecting concern for impacts of air pollution on specific susceptible groups.
31 In the 2000 update, there was an increased focus on quality of life measures as indicators of
32 adversity and a more specific consideration of population risk. Exposure to air pollution that
September 2007 4-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 increases the risk of an adverse effect to the entire population is viewed as adverse, even though
2 it may not increase the risk of any identifiable individual to an unacceptable level. For example,
3 a population of asthmatics could have a distribution of lung function such that no identifiable
4 single individual has a level associated with significant impairment, and exposure to air pollution
5 could shift the distribution to lower levels that still do not bring any identifiable individual to a
6 level that is associated with clinically relevant effects. However, this shift to a lower level would
7 be considered adverse because individuals within the population would have diminished reserve
8 function and, therefore, would be at increased risk if affected by another agent.
9 Reflecting new investigative approaches, the ATS statement also describes the potential
10 usefulness of research into the genetic basis for disease, including responses to environmental
11 agents that provide insights into the mechanistic basis for susceptibility and provide markers of
12 risk status. Likewise, biomarkers that are indicators of exposure, effect, or susceptibility may
13 someday be useful in defining the point at which one or an array of responses should be
14 considered an adverse effect.
15 The 2006 O3 AQCD (U.S. Environmental Protection Agency, 2006) provided
16 information useful in helping to define adverse health effects associated with ambient 63
17 exposure by describing the gradation of severity and adversity of respiratory-related 63 effects.
18 The definitions that relate to responses in impaired individuals are reproduced and presented here
19 in Table 4.3-1. The severity of effects described in the tables and the approaches taken to define
20 their relative adversity are valid and reasonable in the context of the new ATS (2000) statement.
21 As assessed in detail in earlier chapters of this document and briefly recapitulated in
22 preceding sections of this chapter, exposures to a range of SO2 concentrations have been reported
23 to be associated with increasing severity of health effects, ranging from respiratory symptoms to
24 ED visits and hospital admission for respiratory causes. Respiratory effects associated with
25 short-term 862 exposures have been by far the most extensively studied and most clearly shown
26 to be causally related to 862 exposure. Such effects are observed among children, older adults,
27 and persons with respiratory impairments.
28
29 4.3.2 Estimation of Potential Numbers of Persons in At-Risk Susceptible
30 Population Groups in the United States
31 Although SC>2-related health risk estimates may appear to be small, they may well be
32 significant from an overall public health perspective due to the large numbers of individuals in
September 2007 4-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 the potential risk groups. Several subpopulations have been identified as possibly having
2 increased susceptibility or vulnerability to adverse health effects from 862, including children,
3 older adults, and individuals with preexisting pulmonary diseases. One consideration in the
4 assessment of potential public health impacts is the size of various population groups that may be
5 at increased risk for health effects associated with SO2-related air pollution exposure. Table
6 4.3-2 summarizes information on the prevalence of chronic respiratory conditions in the U.S.
7 population in 2004 and 2005 (NHIS, 2006a,b). Individuals with preexisting cardiopulmonary
8 disease constitute a fairly large proportion of the population, with tens of millions of people
9 included in each disease category. Of most concern are those individuals with preexisting
10 respiratory conditions, with approximately 10% of adults and 13% of children having been
11 diagnosed with asthma and 6% of adults with COPD (chronic bronchitis and/or emphysema).
12 In addition, subpopulations based on age group also comprise substantial segments of the
13 population that may be potentially at risk for SO2-related health impacts. Based on U.S. census
14 data from 2000, about 72.3 million (26%) of the U.S. population are under 18 years of age,
15 18.3 million (7.4%) are under 5 years of age, and 35 million (12%) are 65 years of age or older.
16 Hence, large proportions of the U.S. population are included in age groups that are considered
17 likely to have increased susceptibility and vulnerability for health effects from ambient SO2
18 exposure.
19 The prevalence and number of people affected for selected respiratory disorders by age
20 group are summarized in Table 4.3-2. In addition to their high prevalence, these diseases may be
21 severe, resulting in deaths or hospitalizations. There are approximately 2.5 millions deaths from
22 all causes per year in the U.S. population, with about 100,000 deaths from chronic lower
23 respiratory diseases (Kochanek et al., 2004) and 4,000 from asthma (NCHS Health E Stats). For
24 respiratory health diseases, there are nearly 4 million hospital discharges per year (DeFrances
25 et al., 2005), 14 million ED visits (McCaig and Burt, 2005), 112 million ambulatory care visits
26 (Woodwell and Cherry, 2004), and an estimated 700 million restricted activity days per year due
27 to respiratory conditions (Adams et al., 1999). Of the total number of visits for respiratory
28 disease, 1.8 million annual ED visits are reported for asthma, including more than 750,000 visits
29 by children. In addition, nearly 500,000 annual hospitalizations for asthma are reported (NCHS
30 Health E Stats summarizing 2005 NHIS data).
September 2007 4-18 DRAFT-DO NOT QUOTE OR CITE
-------
1 Centers for Disease Control and Prevention (CDC) analyses have shown that the burden
2 of asthma has increased over the past two decades (NCHS Health E Stats 2005 NHIS data for
3 both adults and children). In 2005, approximately 22.2 million (7.7% of the population)
4 currently had asthma. The incidence was higher among children (8.9% of children) compared to
5 adults (7.2%) (Note: 2004 data is shown in Table 4.3-2, with a prevalence of 6.7%). In addition,
6 prevalence and severity is higher among certain ethnic or racial groups, such as Puerto Ricans,
7 American Indians, Alaskan Natives, and African Americans. The asthma hospitalization rate for
8 black people was 240% higher than that for white people. Puerto Ricans were reported to have
9 the highest asthma death rate (360% higher than non-Hispanic white people) and non-Hispanic
10 black people had an asthma death rate 200% higher than non-Hispanic white people. Gender and
11 age is also a determinant of prevalence and severity, with adult females having a 40% higher
12 prevalence than adult males, while boys had a 30% higher rate than girls. Overall, females had a
13 hospitalization rate about 35% higher than males.
14 Evidence indicates that several groups are at increased risks from SC>2 exposures
15 compared to the average population. Susceptible subgroups include individuals with preexisting
16 disease, especially asthma, and children and older adults. Other individuals with increased
17 vulnerability include those who spend a lot of time outdoors at increased exertion levels (e.g.,
18 outdoor workers, children, individuals who exercise or play sports) and those in proximity to
19 large uncontrolled or poorly controlled sources. The considerable size of the population groups
20 at risk indicate that exposure to ambient SO2 could have a significant impact on public health in
21 the United States.
September 2007 4-19 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4.3-1. GRADATION OF INDIVIDUAL RESPONSES TO SHORT-TERM SO2
EXPOSURE IN INDIVIDUALS WITH IMPAIRED RESPIRATORY SYSTEMS
Functional
Response
None
Small
Moderate
Large
FEVi change
Decrements of Decrements of 3 to
Within normal Increases of <100%
range
Decrements of > 10
but <20%
Increase of < 300%
Decrements of
>20%
Increases of >300%
Nonspecific
bronchial
responsiveness3
Airways resistance Within normal sRaw increased <100% sRaw increased up to sRaw increased
(sRaw) range (±20%) 200% or up to >200% or more than
Duration of
response
Symptomatic
Response
Wheeze
Cough
Chest pain
Duration of
response
Impact of
Responses
Interference with
normal activity
Medical treatment
None
Normal
None
Infrequent
cough
None
None
Normal
None
No change
<4h
Mild
With otherwise normal
breathing
Cough with deep breath
Discomfort just
noticeable on exercise
or deep breath
<4h
Mild
Few individuals choose
to limit activity
Normal medication as
needed
15 cmH2Os
>4hbut <24h
Moderate
With shortness of
breath
Frequent spontaneous
cough
Marked discomfort on
exercise or deep
breath
>4 h, but <24 h
Moderate
Many individuals
choose to limit
activity
Increased frequency
of medication use or
additional medication
15cmH2Os
>24h
Severe
Persistent with
shortness of breath
Persistent
uncontrollable
cough
Severe discomfort
on exercise or deep
breath
>24h
Severe
Most individuals
choose to limit
activity
Physician or
emergency room
visit
aAn increase in nonspecific bronchial responsiveness of 100% is equivalent to a 50% decrease in PD20 or PD100.
Source: This table is reproduced from the 1996 O3 AQCD (Table 9-2, page 9-25) (U.S. Environmental Protection
Agency, 1996).
September 2007
4-20
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4.3-2. PREVALENCE OF SELECTED RESPIRATORY DISORDERS BY AGE GROUP AND BY GEOGRAPHIC
t^) '
CD
^
O
O
-^
K>
O
H
6
o
0
H
O
o
H
W
O
O
H
W
INTERVIEW SURVEY)
Age (Years)
Adults
(18+ Years) 18-44 45-64 65-74 75+
Cases
Chronic Condition/Disease (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
Emphysema 3.5 1.7 0.3 2 4.9 6.0
Age (Years)
Children
(<18 years) 0-4 5-11 12-17
Cases
Chronic Condition/Disease (x 106) % % % %
Respiratory Conditions
Asthma 6.5 8.9 6.8 9.9 9.6
Source: National Center for Health Statistics (2006a,b).
Region
Northeast Midwest South West
% % % %
6.8 6.8 6.0 7.5
4.0 4.7 4.4 3.5
1.5 1.7 2.0 1.1
Region
Northeast Midwest South West
% % % %
10.1 8.5 9.3 7.9
-------
i 5. KEY FINDINGS AND CONCLUSIONS
2
O
4 The previous chapters have presented the most policy-relevant information related to the
5 atmospheric chemistry and exposures to sulfur dioxide (SO2) and have discussed the health effects
6 of SC>2 exposure. This chapter provides concise summaries of key findings and reports conclusions
7 drawn from atmospheric sciences, ambient air data analyses, exposure assessment, dosimetry, and
8 health evidence in consideration of the review of the National Ambient Air Quality Standards
9 (NAAQS) for SO2.
10
11
12 5.1 SUMMARY OF KEY FINDINGS RELATED TO THE
13 SOURCE-TO-DOSE RELATIONSHIP
14 Key elements linking sources to health effects are: emission source identification,
15 atmospheric chemistry and transport of a pollutant, techniques for ambient measurement, spatial
16 and temporal patterns in concentrations, correlations to other relevant chemical species, and patterns
17 of human exposures to ambient pollutants.
18
19 5.1.1 Emission Sources, Atmospheric Science, and Ambient Monitoring
20 Methods
21 The characteristics of anthropogenic sources and atmospheric chemistry and monitoring
22 methods for SC>2 are relatively well known.
23 • Anthropogenic 862 is emitted mainly by fossil fuel combustion (chiefly coal and oil) and
24 metal smelting, with its largest emissions coming from elevated point sources like the stacks
25 of power plants and industrial facilities.
26 • Anthropogenic SC>2 emissions from electric generating utilities and smelters have declined
27 substantially since 1990
28 • SC>2 is a soluble gas that is oxidized mainly in the aqueous phase in cloud drops with gas
29 phase oxidation being of secondary importance. Both pathways lead quantitatively to
30 sulfate formation in cloud drops and/or in particles. Sulfur dioxide and sulfate are removed
31 from the atmosphere by wet and dry deposition.
32 • Ambient SC>2 is most commonly monitored in regulatory networks using the pulsed
33 fluorescence technique and reported with a 1-h frequency, although finer time-scales are
September 2007 5-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 sometimes available at selected sites. More sensitive techniques for measuring 862 are
2 available, but most of these systems are too complex and expensive for routine monitoring
3 applications.
4
5 5.1.2 Ambient Concentrations
6 The decline in SC>2 emissions from electric generating utilities and smelters since 1990 has
7 lowered ambient SO2 concentrations and improved air quality dramatically, as demonstrated in the
8 data collected from the State and Local Air Monitoring Stations (SLAMS) and National Air
9 Monitoring Stations (NAMS) networks.
10 • Measured annual mean SO2 values have not been observed to exceed the annual primary
11 NAAQS (0.03 ppm) since 2000. Ambient concentrations decreased 48% between 1990 and
12 2005 owing to controls administered by EPA's Acid Rain Program and Clean Air Markets
13 Division. In addition, means and maxima of the 24-h concentrations in the 12 consolidated
14 metropolitan statistical areas (CMSAs) with >4 monitors in the years 2003 through 2005
15 were never in excess of the 24-h primary SC>2 NAAQS (0.14 ppm). The ambient monitors
16 currently deployed in the regulatory networks are fully adequate to determine compliance
17 with these standards.
18 • Monitors deployed in the current regulatory network are adequate to detect SO2
19 concentrations above 3 ppb. But their detection technique is inadequate for accurate and
20 precise measurements at or near the current mean 24-h SC>2 levels (~3 ppb). The U.S.
21 Environmental Protection Agency (EPA) through its National Core Monitoring Network
22 (NCore) initiative is engaged in a program to install and operate trace-level SC>2 instruments
23 with lower limits of detection that will increase the accuracy and precision of observations at
24 current low ambient levels.
25 • Ambient annual average concentrations reported in the regulatory monitoring networks of
26 the continental U.S. (CONUS) over the years 2003 to 2005 ranged from a low of ~1 ppb
27 (reported) on the West Coast to a high of ~3 ppb (reported) in the Mid-Atlantic region where
28 SO2 emissions remain highest. Both emissions and ambient concentrations demonstrate a
29 strong east-to-west gradient, owing to the overrepresentation of SO2-emitting electric
30 generating units in the Ohio River Valley and upper South regions.
September 2007 5-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 • SO2 concentrations demonstrate no correlation to concentrations of sulfate (SC>42 ) at the
2 12 CMSAs with more than four SC>2 monitors except at Riverside, CA. This exception
3 likely arises from Riverside's geographic location downwind of the regionally important
4 SC>2 sources near Los Angeles, and the strongly correlated seasonality of SO2 with SC>42 ,
5 each showing peaks in summer when SC>2 oxidation would be maximized during transport to
6 Riverside.
7 • Policy relevant background levels of SC>2 are estimated to be <1% of typical ambient levels
8 (or in the range of a few hundredths of a ppb) across most of the United States. However,
9 much higher values are found in areas affected by volcanic or geothermal activity as in
10 Hawaii (>30 ppb) or in the Pacific Northwest, or in areas affected by trans-Pacific or trans-
11 Arctic transport from Eurasia.
12
13 5.1.3 Exposure Assessment
14 The amount of time a person spends in different microenvironments and the infiltration
15 characteristics of these microenvironments are strong determinants of the association between
16 ambient 862 concentrations and human exposures. Relatively few studies have been conducted
17 since the last review examining the relationships among personal exposures, and indoor, outdoor,
18 and ambient concentrations of SO2.
19 • In studies in which personal exposure concentrations were above detection limits, or in
20 studies using active denuder systems, reasonably strong associations were found between
21 personal exposure and ambient SC>2 concentrations.
22 • Passive badges used to monitor personal exposures generally cannot accurately measure
23 concentrations of SC>2 over commonly used sampling periods because concentrations
24 typically encountered by the subjects wearing them are often well below limits of detection.
25 Hence, associations between ambient concentrations and personal exposure, using
26 commonly deployed passive methods, can be incorrectly or inadequately characterized.
27 • The main source of SC>2 in indoor environments is infiltration from outdoors, as evidenced
28 by relatively high correlations between indoor and outdoor values and lower values indoors
29 than outdoors. However, a wide range of indoor-outdoor ratios was reported in the studies
30 examined here: from 0.03 to 1.01. A number of factors including instrument measurement
31 error contribute to these results.
September 2007 5-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 • In addition to ambient 862, people could also be exposed to 862 produced by indoor heating
2 sources. Exposures of this sort would be limited, though, because the chief identified source
3 activity, kerosene space heater use, is not widespread. However, disparities in
4 socioeconomic status and behavior, and any differential susceptibilities related to such
5 disparities may result in increased exposure of selected groups.
6 • The effect of exposure error on community time-series epidemiology studies has been
7 investigated in a limited number of studies, although not specifically for SC>2. Variations in
8 non-ambient exposure and in the fractional contribution of ambient pollutants to exposure
9 will not influence the observed health effect estimate, unless they are correlated with the
10 ambient concentration.
11
12 5.1.4 Dosimetry
13 Dosimetry of SC>2 is the measurement or estimation of the amount of SC>2 or its reaction
14 products reaching and persisting at specific sites in the respiratory tract following exposures.
15 • Due to its high solubility, SC>2 is readily removed in the moist surfaces of the nose and other
16 respiratory passages. With quiescent nasal breathing, almost all inhaled SC>2 is removed in
17 the extrathoracic (head) region. This limits the potential for direct effects on the more
18 sensitive thoracic regions of the respiratory tract. Factors that can increase penetration of
19 SC>2 to these regions include oral and oronasal breathing, increased ventilation rates and the
20 presence of particles or fog droplets that may act as carriers for SO2.
21
22
23 5.2 SUMMARY OF KEY HEALTH EFFECTS FINDINGS
24
25 5.2.1 Findings from the Previous Review of the NAAQS for SO2
26 In the previous review of the NAAQS for SC>2, the following conclusions were reached
27 regarding health effects of SO2 (U.S. Environmental Protection Agency, 1982b, 1994b):
28 • Although SC>2 may produce effects through several mechanisms, the most striking acute
29 effects observed appear to result from stimulation of receptors in the tracheobronchial
30 region, leading to a neurally mediated reflex bronchoconstriction.
31 • The major effects categories of concern associated with high exposures to 862 include
32 sensory and other nonrespiratory responses, effects on respiratory mechanics and symptoms,
September 2007 5-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 aggravation of existing respiratory and cardiovascular disease, effects on clearance and other
2 host defense mechanisms, and mortality.
3 • The major subgroups of the population that appear likely to be most sensitive to the effects
4 of SO2 include asthmatics, individuals not diagnosed as asthmatic but with atopic disorders
5 (e.g., allergies), and individuals with chronic obstructive pulmonary or cardiovascular
6 disease. Other subgroups that may be somewhat sensitive include the elderly and children.
7 • The major effects observed in human clinical studies following peak exposures (1 h or less,
8 generally 5 to 15 min) are increases in airway resistance and decreases in other functional
9 measures indicative of significant bronchoconstriction in relatively healthy asthmatic or
10 atopic subjects. At 0.4-0.6 ppm SO2, changes in functional measures were accompanied by
11 mild increases in perceptible symptoms such as wheezing, chest tightness, and coughing. At
12 higher concentrations, effects were more pronounced and the fraction of asthmatic subjects
13 who responded increased, with clearer indications of clinically or physiologically significant
14 effects at 0.6 to 0.75 ppm and above.
15 • A substantial percentage (25 to 50 percent) of mild to moderate asthmatic individuals
16 exposed for 5 to 15 minutes to 0.6 to 1.0 ppm SO2 during moderate exercise would be
17 expected to have respiratory function changes. The effects observed after exposure to 0.6 to
18 1.0 ppm SC>2 are relatively transient (not lasting more than a few hours) and are not likely to
19 worsen or to reoccur with the same magnitude of response if re-exposure to another SC>2
20 peak occurred within the next several hours after the initial exposure. At SC>2 concentrations
21 at or below 0.5 ppm with moderate exercise, only a relatively small percentage (>10 to 20
22 percent) of mild and moderate asthmatic individuals are likely to experience lung function
23 changes distinctly larger than those they typically experience. Furthermore, compared to the
24 response at 0.6 to 1.0 ppm SO2, the response at or below 0.5 ppm SO2 is less likely to be
25 perceptible and of immediate health concern.
26 • In the epidemiological studies, an association of short-term (generally hours to days) SC>2
27 exposure with daily mortality was likely at levels of 0.19 to 0.38 ppm, an association with
28 aggravation of bronchitis was likely at levels of 0.19 to 0.23 ppm and possible at levels
29 below 0.19 ppm, and small, reversible declines in lung function in children were possible at
30 0.10 to 0.18 ppm.
September 2007 5-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Although the possibility of effects from long-term (generally months to years) lower level
2 exposures to 862 could not be ruled out, no quantitative rationale could be offered to
3 support a specific range of interest for an annual standard. The limited available
4 epidemiological data indicated associations between respiratory illnesses and symptoms and
5 persistent exposures to SO2 in areas with long-term averages exceeding 0.04 ppm.
6
1 5.2.2 New Findings on the Health Effects of Exposure to SOi
8 New evidence developed since the previous NAAQS review for SC>2 has confirmed and
9 extended the conclusions articulated in the 1982 Air Quality Criteria Document (AQCD), 1986
10 Second Addendum, and 1994 Supplement to the Second Addendum. In the time since the previous
11 review, the epidemiological evidence has grown substantially, including new field or panel studies
12 on respiratory health outcomes, numerous time-series epidemiological studies of effects including
13 emergency department (ED) visits and hospital admissions, and a substantial number of studies
14 evaluating mortality risk with short-term (generally 24-h average) 862 exposures. Several new
15 studies have reported findings from prospective cohort studies on respiratory health effects and
16 mortality with long-term (generally months to years) 862 exposure. While not as marked as the
17 growth in epidemiological literature, a number of new human clinical and animal toxicological
18 studies provide some additional biological plausibility for the observed relationships between SC>2
19 exposure and health effects in epidemiological studies.
20 The key findings of this draft Integrated Science Assessment (ISA) on the health effects of
21 SC>2 exposure are presented below. Here, we build upon the discussions in Chapter 3 to draw
22 conclusions regarding the overall strength of the body of evidence and the extent to which causal
23 inferences may be made. Strong evidence from human clinical studies can lead to a conclusion of a
24 "causal" relationship between exposure and adverse health effects. Where the epidemiological
25 evidence is strong and there is coherent and plausible clinical or toxicological evidence, we have
26 concluded that the relationship is "likely causal." Where the epidemiological findings are generally
27 strong and consistent, but the available experimental evidence is too limited to draw conclusions
28 regarding coherence, mechanism(s) of action, or plausibility of the results, we have concluded that
29 this relationship is "suggestive." In some situations, the evidence from epidemiological and
30 experimental studies is not found to be strong or consistent (sometimes with very limited available
31 evidence) and there is limited support for coherence and plausibility; these relationships we judge to
September 2007 5-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 be "inconclusive." Where possible, we have also included observations about the general levels or
2 ranges of concentrations at which effects have been observed. A series of tables with information
3 supporting these observations are presented in the appendix following this chapter. Table 5A-1
4 summarizes key animal toxicological studies and the lowest levels at which effects have been seen
5 for a series of effect categories. Table 5A-2 summarizes the key findings of human clinical studies,
6 and the exposure levels at which those effects have been observed. The results of key new
7 epidemiological studies on respiratory health effects are presented in Tables 5A-3 (panel studies of
8 respiratory symptoms) and 5A-4 (population studies of ED visits and hospital admissions for
9 respiratory causes) and include information about the distribution of 862 levels (generally provided
10 as mean and range) as presented in the study.
11
12 5.2.2.1 Peak (5-15 min) Exposure to SO2 and Respiratory Health Effects
13 We conclude that there is a causal relationship between peak (1-h or less, typically 5 to 15
14 min) exposure to SC>2 and effects on the respiratory system, based on evidence from human clinical
15 studies. Human clinical studies provide clear evidence that peak exposures to SO2 at levels of 0.5 to
16 1.0 ppm cause effects on the respiratory system, namely decrements in lung function and increases
17 in respiratory symptoms in exercising asthmatic adults.
18
19 Respiratory Symptoms
20 • The human clinical studies have reported increased respiratory symptoms with SC>2
21 concentrations of as low as 0.5 ppm in asthmatic subjects (Section 3.1.1.1). One human
22 clinical study with SCVsensitive asthmatics reported that respiratory symptoms (i.e.,
23 shortness of breath, wheeze, and chest tightness) increased with increasing 862
24 concentration (0-, 0.5-, and 1.0-ppm 802) following exposures of 10 min with varying levels
25 of exercise (Gong et al., 1995). It was also observed that exposure to 0.5-ppm SO2 during
26 light exercise evoked a more severe symptomatic response than heavy exercise in clean air.
27
2 8 Lung Function
29 • Human clinical studies have consistently demonstrated decreases in lung function (e.g.,
30 decreased forced expiratory volume in 1 s [FEVi] and increased specific airways resistance
31 [sRaw]) following peak exposures (5 to 15 min) to SC>2 (Section 3.1.1.2). These effects
32 have clearly and consistently been shown to be exacerbated among individuals with asthma,
September 2007 5-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 with asthmatics exhibiting significant decrements in lung function following 5- to 15-min
2 exposures to 862 concentrations of as low as 0.5 ppm while performing moderate levels of
3 exercise (e.g., Gong et al., 1995; Horstman et al., 1986; Linn et al., 1987; Sheppard et al.,
4 1981). The effect of peak SO2 exposure on lung function has been shown to increase in
5 magnitude with increasing SC>2 concentrations above 0.5 ppm. Studies have further
6 observed significant decrements in lung function in some sensitive asthmatics following 5-
7 15 min exposures to SC>2 concentrations of as low as 0.25 ppm while performing moderate
8 levels of exercise (Horstman et al., 1986; Sheppard et al., 1981). Thus, the observations of
9 increased bronchoconstriction and airway resistance in human clinical studies provide clear
10 evidence for 862 effects with peak exposure.
11
12 5.2.2.2 Short-Term (24-h average) Exposure to SO2 and Respiratory Health Effects
13 We conclude that there is a likely causal relationship between short-term exposure to SC>2 at
14 ambient concentrations and effects on the respiratory system, based on consideration of all the data.
15 Numerous new epidemiological studies, supported by evidence from toxicological and human
16 clinical studies provide evidence of a relationship between short-term (24-h average) exposures to
17 SC>2 and respiratory health effects, ranging from respiratory symptoms and increasing in severity to
18 ED visits and hospital admissions for respiratory causes. These effects were observed particularly
19 in individuals with preexisting respiratory diseases, children, and older adults (65+ years).
20 Associations between short-term exposure to 862 and respiratory morbidity were generally robust
21 to adjustment for potential confounding by copollutants, as assessed using multipollutant models.
22 As shown in Tables 5A-3 and 5A-4, almost all of the epidemiologic studies have been conducted in
23 areas where the maximum ambient 24-h average SC>2 concentration was below the current 24-h
24 average NAAQS level of 0.14 ppm. Evidence related to specific types of respiratory effects is
25 highlighted below.
26
27 Respiratory Symptoms
28 • Recent epidemiological studies provide evidence for an association between ambient SC>2
29 exposure and increased respiratory symptoms in children, particularly those with asthma or
30 chronic respiratory symptoms (Section 3.1.1.1, see Figures 3.1-3 and 3.1-4). Recent U.S.
31 multicity studies observed significant associations between 862 and respiratory symptoms at
32 a median range of 17 to 37 ppb (75th percentile: ~ 25 to 50) across cities for 3-h average
September 2007 5-8 DRAFT-DO NOT QUOTE OR CITE
-------
1 SO2 (National Cooperative Inner-City Asthma Study [NCICAS], Mortimer et al., 2002) and
2 2.2 to 7.4 ppb (90th percentile: 4.4 to 14.2) for 24-h average SO2 (Childhood Asthma
3 Management Program [CAMP], Schildcrout et al., 2006). The SO2 effect was generally
4 found to be robust after adjusting for particulate matter (PM) and other copollutants.
5 • Results from the epidemiological studies examining the association between SO2 and
6 respiratory symptoms in adults are generally mixed, with some showing positive
7 associations and others finding no relationship at current ambient levels (Section 3.1.1.1).
8
9 Lung Function
10 • Epidemiological studies observed mixed results for the association between 24-h average
11 ambient SO2 and lung function in children and adults (Section 3.1.1.2). A limited number of
12 animal studies and human clinical studies of >l-h exposures provide some degree of
13 biologic plausibility and no concentration-response information to allow an understanding of
14 the inconclusive epidemiological findings.
15
16 A irway Hyperresponsiveness
17 • Very limited epidemiological evidence suggests that exposure to SO2 may lead to airway
18 hyperresponsiveness in atopic individuals (Section 3.1.1.4). Toxicological studies that
19 observed increased airway obstruction and hypersensitivity at low levels (0.1 ppm) in
20 allergen-sensitized animals provide biological plausibility for these findings. The
21 epidemiological evidence further observed that atopic individuals may be at increased risk
22 for SO2-induced respiratory symptoms.
23
24 Inflammation
25 • The limited epidemiological, human clinical, and toxicological evidence does not suggest
26 that exposure to SO2 at current ambient concentrations is associated with inflammation in
27 the airways (Section 3.1.1.3).
28
29 Respiratory ED Visits and Hospitalizations
30 • A large number of epidemiologic studies provide evidence of positive, but not always
31 statistically significant, associations between ambient SO2 concentrations and ED visits and
32 hospitalizations for all respiratory causes and asthma, particularly amount children and older
September 2007 5-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 adults (Section 3.1.1.6, see Figures 3.1-7 through 3.1-10). These findings are generally
2 robust when additional copollutants are included in the model (Figure 3.1-11). Biologic
3 plausibility for these findings of increased ED visits and hospitalizations is found in the
4 epidemiologic and human clinical studies that observed increased respiratory symptoms and
5 decreased lung function, and the animal toxicological studies that observed SO2-induced
6 altered lung host defenses (Section 3.1.1.5).
7
8 5.2.2.3 Short-Term Exposure to SO2 and Cardiovascular Health Effects
9 The collective evidence with regard to the effect of SC>2 on the cardiovascular system is
10 inconclusive.
11 • Evidence from epidemiological studies of heart rate variability (HRV), cardiac
12 repolarization changes, and cardiac rhythm disorders provide limited evidence of
13 associations with 862 exposure (Section 3.1.2.1 to 3.1.2.3). The parameters measured in
14 these studies were associated most strongly with PM compared to other ambient pollutants,
15 so the effects observed for SO2 may have been confounded. Two human clinical studies
16 provided weak and inconsistent evidence for an effect of SC>2 on FtRV, while one animal
17 toxicological study did not provide support for an effect on spontaneous arrhythmias.
18 Overall, evidence that SC>2 affects cardiac autonomic control and cardiac rhythm is
19 inconclusive.
20 • Some studies have observed positive associations between ambient SC>2 concentrations and
21 ED visits and hospital admissions for all cardiovascular diseases (CVDs), particularly
22 among individuals 65 years or greater (Section 3.1.2.7, see Figure 3.1-12). Given the
23 limited number of studies that assessed potential confounding by copollutants for this
24 outcome (Figure 3.1-13), which is of concern because of the moderate to strong correlation
25 between SO2 and various copollutants in most studies, and the lack of supportive data from
26 panel studies and human clinical studies on cardiovascular health effects, the collective
27 evidence that ambient SC>2 has an effect of CVD ED visits and hospitalizations is weak.
28
29 5.2.2.4 Short-Term Exposure to SO2 and Other Systemic Effects
30 The limited toxicological evidence for SC>2-related effects on the nervous system and other
31 organ systems is inconclusive.
September 2007 5-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 • In a limited number of toxicological studies, exposure to SO2 has been shown to affect
2 certain neurodevelopmental and cognitive effects (Section 3.1.3.1). There was suggestive
3 evidence that young animals and those with preexisting conditions such as diabetes were
4 more susceptible to these effects. These effects were observed only at high concentrations
5 ofSO2.
6 • Though limited, the overall animal toxicology database on SO2 exposure suggests no overt
7 adverse effects on the reproductive, hematological, gastrointestinal, renal, lymphatic, and
8 endocrine systems (Section 3.1.3.2).
9
10 5.2.2.5 Effects of Short-Term Exposure to SO2 on Mortality
11 Epidemiological evidence is suggestive of associations between SO2 and nonaccidental all-
12 cause and cardiopulmonary-related mortality, but additional research is needed to more fully
13 establish underlying mechanisms by which such effects occur.
14 • Recent epidemiological studies have reported associations between mortality and SO2, often
15 at mean 24-h average levels below 10 ppb (Section 3.2.1, see Figure 3.2-2). The range of
16 SO2 all cause (nonaccidental) mortality risk estimates is 0.4 to 2% per 10-ppb increase in
17 24-h average SO2 in several large multeity studies and meta-analyses. In the large multicity
18 time-series studies, the SO2 risk estimates were generally reduced when copollutants, either
19 PM indices and/or nitrogen dioxide (NO2), were added in the model. Thus, some extent of
20 confounding among these pollutants is suggested.
21 • Results from multicity studies indicate that the SO2 effect estimates for respiratory mortality
22 were generally larger than the cardiovascular mortality risk estimates, suggesting a stronger
23 association of SO2 with respiratory mortality compared to cardiovascular mortality;
24 however, similar associations were observed for other pollutants, including PM and NO2
25 (Section 3.2.2, See Figure 3.2-3). There is some biological plausibility for the stronger
26 associations observed between ambient SO2 and respiratory mortality given the likely causal
27 relationship between SO2 and respiratory morbidity outcomes.
28 • An intervention study from Hong Kong (Hedley et al., 2002) supports the notion that a
29 reduction in SO2 levels results in a reduction in deaths, but this does not preclude the
30 possibility that the causal agent is not SO2 but rather something else that is emitted along
31 with SO2, such as the trace metals vanadium and nickel (Section 3.2.3). Overall, the
September 2007 5-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 evidence that 862 is causally related to mortality at current ambient levels is suggestive but
2 limited by potential confounding in the epidemiological data and the absence of strong
3 biological plausibility.
4
5 5.2.2.6 Effects of Long-Term Exposure to SO2 on Morbidity
6 The epidemiological findings, along with the very limited toxicological findings, provide
7 inconclusive evidence that long-term exposure to SC>2 has adverse health effects.
8 • Several epidemiological studies that examined the effects of long-term exposure to SO2 on
9 asthma, bronchitis, and respiratory symptoms observed positive associations in children
10 (Section 3.3.1.1). However, there are inconsistencies in the findings observed, with some
11 finding effects on bronchitic symptoms but not asthma symptoms and vice versa. Overall,
12 while the evidence is suggestive, the variety of outcomes examined and the inconsistencies
13 in the observed results make it difficult to assess the impact of long-term exposure of 862 on
14 respiratory health.
15 • The epidemiological evidence reported mixed results on the effect of long-term exposure on
16 lung function (Section 3.3.1.2). An animal toxicological study in rabbits that were exposed
17 to 5-ppm SO2 for 13 weeks did not observe any alterations in pulmonary function or
18 respiratory parameters. These results, collectively, do not indicate that long-term exposure
19 to SC>2 has a detrimental effect on lung function.
20 • A very limited number of animal toxicological studies examined histopathological changes
21 in the respiratory system following exposure to SC>2 (Section 3.3.1.3). In one study, rats
22 were exposed for 4 to 8 months to 1-ppm SC>2 and an increased incidence of bronchiolar
23 epithelial hyperplasia and a small increase (12%) in numbers of nonciliated epithelial cells
24 in terminal respiratory bronchioles were observed at 4 but not at 8 months of exposure. Two
25 other toxicological studies with shorter exposure periods (6 days and 4 weeks) did not report
26 any alveolar or other pulmonary lesions.
27 • The epidemiological studies did not provide any evidence that long-term exposure to 862 is
28 associated with an increased risk of lung cancer (Section 3.3.2). The toxicological studies
29 indicate that any potential pathways for sulfur oxides (SOX) to induce carcinogenesis,
30 cocarcinogenesis, or tumor promotion appear to be complex and may be highly situational.
31 SC>2 and its derivatives appear unlikely to have significant carcinogenic potential.
September 2007 5-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Epidemiological studies on birth outcomes have found suggestive positive associations
2 between 862 exposure and low birth weight (Section 3.3.3, see Figure 3.3-1). One concern,
3 however, is that many of these studies could not adjust for potential confounding factors.
4 Additional limitations affecting the interpretation of these studies is a lack of evidence for
5 biological plausibility of an effect, inconsistencies across trimesters of pregnancy, and a lack
6 of evidence to evaluate confounding by copollutants.
7
8 5.2.2.7 Effects of Long-Term Exposure to SO2 on Mortality
9 Results from the limited number of epidemiological studies are inconclusive regarding the
10 association between long-term exposure to SO2 and mortality.
11 • The results from two major U.S. epidemiological studies (Harvard Six Cities Study
12 [Dockery et al., 1993; reanalysis, Krewski et al., 2000] and the American Cancer Study
13 [ACS] [Pope et al., 1995; reanalysis, Krewski et al., 2000]) observed associations between
14 long-term exposure to 862 and mortality (Section 3.4.1, see Figure 3.4-1). However,
15 Krewski et al. concluded that in the absence of a plausible toxicological mechanism by
16 which SC>2 could lead to increased mortality suggested that 862 might be acting as a marker
17 for other mortality-associated pollutants. The inability to distinguish potential confounding
18 by copollutants, inconsistent observations across the various U.S. and European studies and
19 the remaining uncertainties regarding the geographic scale of analysis and copollutant
20 confounding limit the interpretation of a causal relationship.
21
22 5.2.2.8 Concentration-Response Function and Potential Thresholds
23 The limited evidence from epidemiological studies examining the concentration-response
24 function of 862 health effects is inconclusive regarding the presence of an effect threshold
25 (Section 4.1).
26 • Evidence from human clinical studies indicated wide interindividual variability in response
27 to SO2 exposures (Horstman et al., 1986; Linn et al., 1987). The evidence from human
28 clinical studies demonstrates consistent SCVinduced respiratory effects following 5 to 15
29 min exposures of SO2 at levels between 0.5 and 1.0 ppm, with weaker evidence of effects at
30 concentrations as low as 0.25 ppm in some sensitive asthmatics.
31 • Several epidemiological studies that examined the concentration-response function between
32 short-term (24-h average) exposure to SC>2 and respiratory morbidity observed a linear
September 2007 5-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 relationship across the entire concentration range, suggesting a lack of a threshold in effect.
2 However, given the various limitations in observing a possible threshold in population
3 studies, the lack of evidence for a threshold does not necessarily indicate that there is indeed
4 no threshold for SC>2 health effects. Two epidemiological studies did report a possible
5 threshold level of 22 to 23 ppb (24-h average) at which no statistically significant SO2-
6 related respiratory health effect was observed. However, as these observations were based
7 on only a few influential data points (24-h average SC>2 concentrations above the 90th
8 percentile), the results should be viewed with caution.
9 • In considering the factors that influence the dosimetry of 862, a mechanistic argument for
10 individual thresholds in SCVrelated health effects can be made. The individual thresholds
11 for response may not necessarily translate to a detectable population threshold. Additivity
12 of SCVinduced responses to a background level of response and interindividual differences
13 in susceptibility to SO2-related health effects will tend to linearize the concentration-
14 response relations and obscure any population threshold that exists.
15
16 5.2.2.9 Susceptible and Vulnerable Populations
17 Certain subgroups within the population have been found to be more susceptible or
18 vulnerable to the effects of SC>2 exposure, including individuals with preexisting respiratory
19 diseases, children, and older adults (65+ years) (Section 4.2). It should be further noted that other
20 individuals who may not generally be susceptible to SCVrelated health effects may experience
21 transient airways reactivity to respiratory irritants such as 862 following a recent viral respiratory
22 infection (Stempel and Boucher, 1981).
23 • Substantial evidence from epidemiological studies suggests that subjects with respiratory
24 illnesses, particularly asthma, are more susceptible to respiratory health effects from 862
25 exposures than the general public (Section 4.2.2.1). The observations in human clinical
26 studies of increased sensitivity to SC>2 exposures in asthmatic subjects compared to healthy
27 subjects provide coherence and biological plausibility for these observations in
28 epidemiological studies.
29 • There is weak evidence from a small number of panel studies that suggests that individuals
30 with preexisting CVD may be more susceptible to adverse health effects from ambient SC>2
31 exposures than the general public (Section 4.2.2.2). Additional research is necessary to
September 2007 5-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 assess whether individuals with preexisting CVD constitutes a susceptible group for 862
2 health effects.
3 • Limited epidemiological evidence suggests that children and older adults (65+ years) are
4 more susceptible to the adverse respiratory effects associated with ambient 862
5 concentrations when compared to the general population (Section 4.2.3, see Figure 4.2-1).
6 The few studies that conducted age-stratified analyses when examining cardiovascular
7 outcomes did not find any difference in outcomes when analyses were stratified by age.
8 • Differential effects of air pollution among genetically diverse subpopulations have been
9 examined for a number of glutathione S-transferase (GST) genes and other genotypes in a
10 limited number of studies (Section 4.2.4). Only one of these studies specifically examined
11 SC>2 as the exposure of interest, and it found a significant association with the homozygous
12 wild-type allele for tumor necrosis factor-a (TNF-a). At this time there are only very
13 limited data on which to base a conclusion regarding the effect of 862 exposure on
14 genetically distinct subpopulations.
15
16
17 5.3 CONCLUSIONS
18 This draft ISA focused on scientific information that has become available since the last
19 review and reflects the current state of knowledge on the most relevant issues pertinent to the
20 review of the primary NAAQS for SC>2. The current primary SC>2 NAAQS has two parts - a 24-h
21 average of 0.14 ppm, not to be exceeded more than once per year, and an annual average of
22 0.03 ppm. Exceedances in recent years have become rare, as the mean 24-h average and annual
23 average SC>2 concentrations in the United States for the years 2003 to 2005 were ~4 ppb, with
24 maximum values of >120 ppb for the 24-h average and -14-15 ppb for the annual average. For the
25 monitors reporting a 1-h max in these years, the mean concentration was -13 ppb, with a maximum
26 value of >600 ppb.
27 In the review of the scientific literature for SC>2, evidence from the various disciplines of
28 atmospheric sciences, exposure assessment, dosimetry, human and animal toxicology, and
29 epidemiology was integrated and collectively considered in formulating conclusions. Overall, we
30 conclude that there is a causal relationship between peak (1 h or less, typically 5 to 15 min)
31 exposure to SO2 and effects on the respiratory system, based on evidence from human clinical
32 studies. Human clinical studies provide clear and consistent evidence of a causal relationship
September 2007 5-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 between peak exposures to 862 at levels of 0.5 to 1.0 ppm and effects on the respiratory system,
2 namely decrements in lung function in exercising asthmatic adults. We further conclude that there
3 is a likely causal relationship between short-term (generally 24-h average) SC>2 exposure at ambient
4 levels and respiratory health effects, mostly based on the epidemiological studies. A large body of
5 new epidemiological studies provides evidence of consistent and robust associations between short-
6 term exposure to ambient SC>2 and respiratory health endpoints, ranging from increased respiratory
7 symptoms in children with asthma or chronic respiratory symptoms, and increasing in severity to
8 ED visits and hospital admissions for respiratory causes particularly in children and older adults
9 (65+ years of age). The public health impact of ambient 862 exposures may be large, owing to the
10 fact that these potentially susceptible subgroups constitute a large part of the general population.
11 Associations of health effects with ambient SO2 exposure have been reported in locations where the
12 maximum 24-h average SC>2 concentration was below the levels of the current NAAQS (see Tables
13 5A-3 and 5A-4).
14 However, much uncertainty remains in the interpretation of the health evidence related to
15 ambient SC>2 exposures. Exposure error is one key source of uncertainty, as typical indoor 24-h
16 average 862 concentrations are often below the detection limit of personal exposure monitors, and
17 ambient SC>2 concentrations in locations with low levels may be at or below the detection limit of
18 existing monitors in the regulatory networks. Other sources of uncertainty include the magnitude of
19 SC>2 risk estimates and the shape of concentration-health response relationships. Together, these
20 uncertainties complicate our ability to attribute observed health effects to SO2 directly.
21 The epidemiological observations of SC>2 health effects can be interpreted in several ways
22 that are not mutually exclusive. First, the reported SC>2 effect estimates in epidemiological studies
23 may be attributable to SC>2 per se, reflecting independent SC>2 effects on respiratory health.
24 Available human clinical and animal toxicological studies are conducted at higher than average
25 ambient 862 exposures, and do not examine the most susceptible populations. Due to its high
26 solubility, 862 is readily removed in the moist surfaces of the nose and other respiratory passages,
27 limiting the potential for direct effects on the more sensitive thoracic regions of the respiratory tract
28 during nasal breathing. Factors that can increase penetration of SO2 to these regions, including oral
29 and oronasal breathing, increased ventilation rates, and the presence of high levels of particles or
30 fog droplets that may act as carriers for SC>2. Evidence from human clinical studies indicate that
31 peak exposures (5 to 15 min) to SC>2 at levels as low as 0.5 ppm have been associated with
September 2007 5-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 increased respiratory symptoms and decreased lung function in exercising asthmatics, with levels as
2 low as 0.25 ppm eliciting respiratory responses in some sensitive individuals. These findings
3 provide supportive evidence that peak concentrations of SO2 may be driving the observed
4 associations in epidemiological studies. SO2 at levels such as these are found in only a very few
5 areas in the United States and under specific meteorological conditions.
6 Second, ambient SO2 may be serving as an indicator of complex ambient air pollution
7 mixtures that share the same source as SC>2 (i.e., combustion of sulfur-containing fuels or metal
8 smelting). Other components of mixed emissions from these sources include trace elements such as
9 vanadium, nickel, selenium, and arsenic. It should be noted that paniculate SO42 was found not to
10 be correlated with 862 in ambient data for 12 CMSAs with multiple monitors. In multipollutant
11 models adjusting for PM indices, SO2 effect estimates generally were found to be robust. However,
12 in the event that one or more pollutants act as surrogates for an unmeasured component of a mixture
13 actually responsible for the observed association, the strongest predictor in a multipollutant model
14 could indicate simply which measured pollutant is the best surrogate for the unmeasured pollutant
15 of interest. Therefore, reported SCVrelated effects may represent those of the overall mixture.
16 Third, in the presence of complex pollution mixtures, copollutants may enhance the toxic
17 capability of 862 or 862 may influence the toxicity of copollutants. For example, water-soluble
18 gases such as 862 that are usually largely removed by deposition to wet surfaces in the upper
19 portion of the respiratory tract could be dissolved in particle-bound water and, thereby, be carried
20 into the lower regions of the respiratory tract. In turn, SC>2 can acidify particles, increasing the
21 bioavailability of soluble transition metals capable of inducing lung injury.
22 Assessment of the health effects directly attributable to SC>2 at current average ambient
23 concentrations is difficult at present, particularly due to the uncertainties related to exposure
24 characterization in epidemiological studies using ambient 862 concentration data and the inability
25 to discern the shape of the concentration-response function in the available epidemiology studies.
26 Lack of clear mechanistic understanding for low level exposures increases the difficulty with which
27 available findings can be integrated in assessing the coherence of SCVrelated evidence. Despite
28 these difficulties, the epidemiological evidence, along with limited toxicological and human clinical
29 information, indicates a likely causal association between short-term exposure to SC>2 and
30 respiratory health outcomes. Whether SC>2 has a direct effect, SC>2 is a surrogate for pollution
31 mixtures with the same source, and/or the toxicity of SC>2 is influencing or influenced by the
September 2007 5-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 presence of copollutants, reduction of ambient 862 concentrations will result in decreased
2 frequency and severity of SCVrelated respiratory health effects.
September 2007 5-18 DRAFT-DO NOT QUOTE OR CITE
-------
i APPENDIX 5A.
2
3 SUMMARY OF NEW ANIMAL TOXICOLOGICAL,
4 HUMAN CLINICAL, AND EPIDEMIOLOGICAL STUDIES
5 OF HEALTH EFFECTS ASSOCIATED WITH
6 EXPOSURES TO SULFUR DIOXIDE
September 2007 5A-1 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5A-1. KEY RESPIRATORY HEALTH EFFECTS OF EXPOSURE TO SULFUR
DIOXIDE OBSERVED IN ANIMAL TOXICOLOGICAL STUDIES
SO2 Concentration &
Exposure Duration
MORPHOLOGY
Species
Observed Effects
References
1 ppm, 3 h/day/6 day
Evaluated up to 72 h
postexposure
Male Hartely No alveolar lesions.
guinea pigs
Conner etal. (1985)
5 ppm, 2 h/day,
5 day/wk/4 wk
Male and female
F344 rats
No nasal or pulmonary lesions. No
effect on mucociliary clearance of
radiolabeled aluminosilicate particles.
Wolff etal. (1989)
LUNG INJURY AND INFLAMMATION
1 ppm, 5 h/day, 5 day/wk up
to 4 and 8 mos
Male Sprague- Increased bronchial epithelial
Dawley rats hyperplasia and number of nonciliated
epithelial cells observed at 4 mos.
Smith etal. (1989)
5-21 ppm, 4 h/dayII day
Effects observed at as low
as 5 ppm
Male Kunming Elevated levels of pro-inflammatory
albino mice cytokines IL-6 and TNF-a in lung and
TNF-a in serum.
Meng et al. (2005a)
5, 50, and 100 ppm,
5 h/day/28 day
Male Wistar rats
No evidence of lung injury and lung
epithelial permeability. Significant
elevation in neutrophil number of
5-ppm group at day 14.
Langley-Evans et al.
(1996)
AIRWAY HYPERRESPONSIVENESS AND ALLERGY
0.1, 4.3, and 16.6 ppm
8 h/day/5 day
With ovalbumin challenge
in the last 3 days
0.1 ppm, 5 h/day/5 day
With or without ovalbumin
exposure
Perlbright-female
white guinea pigs
Male, Dunkin-
Hartley guinea
pigs
Bronchial obstruction with ovalbumin
challenge in all the SO2 groups. SO2-
induced potentiation of allergic
sensitization of airway.
Enhanced eosinophil count in SO2-
exposed, and SO2 + ovalbumin-
exposed group of animals. Infiltration
of inflammatory cells. SO2
potentiates ovalbumin-induced
asthmatic reaction in guinea pigs.
Riedeletal. (1988)
Park etal. (2001)
5 ppm, whole body
4 h/day/5 day/6 wk
Sensitization with Candida
albicans after 2 wks of
exposure to SO2
Male Hartley The number of SO2-exposed animals
guinea pigs with prolonged expiration and
inspiration increased after 15 h of
challenge with the antigen. SO2
exposure increases dyspneic
symptoms in guinea pigs.
Kitabatake etal. (1995)
5 ppm SO2 for 4 h Sheep
Sensitized to Ascaris suum
SO2 exposure significantly increased
airway reactivity in allergic sheep.
Abraham etal. (1981)
September 2007
5A-2
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5A-1 (cont'd). KEY RESPIRATORY HEALTH EFFECTS OF EXPOSURE
TO SULFUR DIOXIDE OBSERVED IN ANIMAL TOXICOLOGICAL STUDIES
SO2 Concentration &
Exposure Duration Species
Observed Effects
References
LUNG FUNCTION
1 ppm SO2 for 1 h
Male Hartley Increase in pulmonary resistance and
guinea pigs decrease in dynamic compliance up to
1 h following exposure. No effect of
SO2 exposure on breathing frequency,
tidal volume or minute volume.
Amduretal. (1983)
1 ppm, nose only
3 h/day/6 day
Analyses up to 48 h
following exposure
Ketamine- No effect of SO2 exposure on residual
anesthetized volume, functional reserve capacity,
male Hartley vital capacity, total lung capacity,
guinea pigs respiratory frequency, tidal volume,
pulmonary resistance, or pulmonary
compliance.
Conner etal. (1985)
5 ppm for 45 min
Adult rabbits SO2 exposure results in increased lung
resistance. Bivagotomy had no effect
on this phenomenon, indicating the
noninvolvement of vagal reflex in this
process. SO2 had no effect on the lung
resistance induced by intravenously
administered histamine.
Barthelemy et al. (1988)
5 ppm, 2 h/day
for 13 wks
New Zealand SO2 exposure had no effect on lung
white male and resistance, dynamic compliance,
female rabbits transpulmonary pressure, tidal volume,
respiration rate, or minute volume.
Douglas etal. (1994)
HOST DEFENSE
10 ppm for 4 h,
nose only
White Swiss No effect on red blood cell Fc-receptor
mice mediated phagocytosis or bactericidal
activity.
Jakab etal. (1996)
10 ppm for 4 h,
nose only
Male Wistar No effect of SO2 exposure on alveolar
rats macrophage phagocytosis or
bactericidal activity to Staphylococcus
aureus.
Clarke et al. (2000)
10 ppm for 24 h,
1, 2, and 3 wks
OF1 mice Respiratory challenge withKlebsiella
pneumoniae resulted in increased
mortality and decreased survival time
in SO2-exposed animals.
Azoulay-Dupuis et al. (1982)
September 2007
5A-3
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5A-2. KEY HUMAN HEALTH EFFECTS OF PEAK EXPOSURE TO SULFUR DIOXIDE OBSERVED IN
CLINICAL STUDIES
SO2 Concentration
(ppm)
Exposure
Duration
Observed Effects
References
to
o
o
0.2-0.4 5 min-1 h Significant reductions in FEVi and increases in specific airways
resistance (sRaw) observed among some asthmatic adults. Some
weak and inconsistent evidence to suggest that SO2 exposure may
lead to changes in heart rate variability.
Bethel et al. (1985); Horstman et al. (1986); Linn
et al. (1982, 1983, 1987); Routledge et al. (2006);
Schachter et al. (1984); Sheppard et al. (1981);
Tunnicliffe et al. (2001, 2003)
0.4-0.6 1 min- 2 h Decrements in lung function observed between 0.4- and 0.6-ppm
SO2 in asthmatic adults and adolescents during exercise.
Significant interindividual variability in response has been
consistently demonstrated. Effects observed within 1-5 min of
exposure are generally not enhanced by increasing exposure
duration. Respiratory symptoms (e.g., wheezing and chest
tightness) increase with increasing exposure concentrations above
0.4 ppm. No respiratory effects reported in healthy, non-
asthmatics.
Balmes et al. (1987); Bedi et al. (1979); Gong et al.
(1995); Horstman et al. (1986); Koenig et al.
(1983); Linnetal. (1982, 1983, 1987); Magnussen
etal. (1990); Schachter et al. (1984); Sheppard
etal. (1981)
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
0.6-1.0 lmin-2h Specific airway resistance shown to double following 10-min
exposures to SO2 concentrations between 0.25 and 0.75 ppm with
moderate exercise in 50% of asthmatics tested. Some evidence of
an increase in airway resistance in healthy, non-asthmatic subjects
exposed to SO2 concentrations of as low as 0.75 ppm during heavy
exercise. Respiratory effects attributed to SO2 among asthmatics
during exercise may be diminished after cessation of exercise, even
with continued SO2 exposure.
> 1.0 3 min-1 h Among healthy adults, SO2-attributed decrements in lung function
generally occur at concentrations above 1 ppm during exercise and
above 5 ppm at rest. Markers of airway inflammation are
significantly elevated at 4 h postexposure, reaching peak levels
24 h postexposure.
Balmes et al. (1987); Gong et al. (1995); Hackney
et al. (1984); Horstman et al. (1986, 1988); Koenig
et al. (1983); Linn et al. (1985, 1987); Schachter
etal. (1984); Stacy etal. (1981)
Amdur et al. (1953); Kreisman et al. (1976);
Lawtheretal. (1955, 1975); Sandstrom et al.
(1989); Sim and Pattle (1957); Snell and
Luchsinger (1969)
-------
TABLE 5A-3. EFFECTS OF SHORT-TERM EXPOSURE TO SULFUR DIOXIDE ON RESPIRATORY
SYMPTOMS AMONG CHILDREN
to
o
o
>
l^ft
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Reference, Study
Location, and Period
Study Population
Averaging Time,
Mean(SD) SO2
Levels (ppb)
Statistics for SO2
Air Quality Data (ppb)
98th % 99th % Range Upper Percentile
Standardized Odds Ratio
(95% CI)a
UNITED STATES
Schildcrout et al.
(2006)
Eight North American
Cities
1993-1995
Schwartz etal. (1994)
Six cities, U.S.
Apr-Aug 1985, 1986,
1987 (depends on the
city)
Neas etal. (1995)
Uniontown, PA
Summer 1990
Mortimer et al. (2002)
Eight urban areas, U.S.
Jun-Aug 1993
Asthmatic children
(n = 990)
Children in grades
2-5 (n= 1,844)
Children in grades
4-5(n=83)
Asthmatic
children, aged 4-9
(n = 846)
24-havg: 2.2-7.4
(range of city-
specific medians)
24-havg: 4.1
(median)
12-havg: 10.2
5.9 overnight
14.5 daytime
3-havg: 22
(shown in figure)
NR NR NR 75th: 3.1, 10.7
90th: 4.4, 14.2
(range in city
specific
estimates)
NR NR NR 75th: 8.2
90th: 17.9
Max: 81.9
NR NR IQR: Max: 44.9
11.1
NR NR 0-75 ppb NR
(shown
in graph)
Asthma symptoms: SO2 alone:
1.04 (1.00, 1.08), 3-day sum
S02 + N02:
1.04 (1.00, 1.09), 3-day sum
SO2 + PM10:
1.04 (0.99, 1.08), 3-day sum
Cough incidence : SO2 alone :
1.15(1.02-1.31), 4-day avg
SO2 + PM10:
1.08(0.93, 1.25), 4-day avg
SO2 + NO2:
1.09(0.94, 1.30), 4-day avg
Evening cough:
1.19(1.00, 1.42), lag 12 h
Asthma symptoms:
SO2 alone (8 cities):
1.19(1.06, 1.35), lag 1-2
SO2 + O3+NO2 (7 cities):
1.19(1.04, 1.37), lag 1-2
SO2 + O3 + NO2 + PM10
(3 cities): 1.23 (0.94, 1.62), lag
1-2
-------
to
o
o
TABLE 5A-3 (cont'd). EFFECTS OF SHORT-TERM EXPOSURE TO SULFUR DIOXIDE ON RESPIRATORY
SYMPTOMS AMONG CHILDREN
Statistics for SO2
Reference, Study Location,
and Period
Study
Population
Averaging Time. Air ^uaiit;
Mean(SD) SO2
Levels (ppb) 98th % 99th %
y iiata (ppuj
Range
I
Upper
Percentile
Standardized Odds Ratio
(95% CI)a
EUROPE
Timonen and Pekkanen
(1997)
Kuopio (urban and suburban),
Finland
1994
Children
7-12 yrs with
asthma or cough
symptoms
(n = 169)
24-h avg: 2.3 NR NR NR 75th: 2.7 Upper respiratory symptoms:
Max: 12.3 2.71 (1.19, 6.17), lag 0
3.17(1.21,8.78), lag 1
Ward et al. (2002)
Birmingham and Sandwell,
England
Jan-Mar 1997
May-Jul 1997
Children, age at 24-h avg:
enrollment 9 yrs Median
(n=162) 5.4, Winter
4.7, Summer
NR
NR
2, 18
Winter
2, 10
Summer
NR
Cough:
0.59 (0.25, 1.40), Winter
0.90 (0.49, 1.66), Summer
Shortness of breath:
0.59 (0.32, 1.09), Winter
0.81(0.30, 2.17), Summer
Wheeze:
0.79 (0.38, 1.63), Winter
0.77 (0.28, 2.08), Summer
(7-day avg lag for above results)
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Segalaetal. (1998)
Paris, France
Nov 1992-May 1993
Children
7-15 yrs with
physician-
diagnosed
asthma (n = 84)
24-h avg:
(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
-------
TABLE 5A-3 (cont'd). EFFECTS OF SHORT-TERM EXPOSURE TO SULFUR DIOXIDE ON RESPIRATORY
SYMPTOMS AMONG CHILDREN
to
o
o
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Statistics for SO2
Air Quality Data (ppb)
Reference, Study
Location, and Period Study Population
Averaging lime,
Mean(SD) SO2
Levels (ppb) 98th % 99th %
Upper
Range Percentile
Standardized Odds Ratio
(95% CI)a
EUROPE (cont'd)
Boezenetal. (1998)
Amsterdam and
Meppel (urban and
rural), Netherlands
Winter 1993-1994
Roemeretal. (1993)
Wageningen, the
Netherlands
Winter 1990-1991
Children 7-1 lyrs,
with and without
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)
Children with 24-h avg
chronic respiratory 1-h max
conditions 6-12 yrs
(n = 73)
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
NR NR 0,40.4 Max: 56.5 Asthma attack:
(24-h avg) (1-h max) 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,
Netherlands
Winter 1991
Children 7-1 1 yrs, 24-h avg
nonurban area
(n=112)
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 respiratory 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
-------
to
o
o
TABLE 5A-3 (cont'd). EFFECTS OF SHORT-TERM EXPOSURE TO SULFUR DIOXIDE ON RESPIRATORY
SYMPTOMS AMONG CHILDREN
Reference, Study
Location, and
Period
Study Population
Averaging Time,
Mean(SD) SO2
Levels (ppb)
Statistics for SO2
Air Quality Data (ppb)
98th % 99th % Range Upper Percentile
Standardized Odds Ratio
(95% CI)a
EUROPE (cont'd)
oo
Van der Zee et al.
(1999)
Urban and nonurban
areas, the Netherlands
3 winters, 1992-1995
Children 7-llyrs,
with and without
chronic respiratory
symptoms (n= 633)
24-havg: 1.4,8.8
(range in city-
specific medians)
NR
NR
NR
Max:
6.5, 58.5 (range in
city-specific
maximums)
Lower respiratory
symptoms:
Urban:
SO2 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
SO2 + 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
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
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
-------
to
o
o
TABLE 5A-3 (cont'd). EFFECTS OF SHORT-TERM EXPOSURE TO SULFUR DIOXIDE ON RESPIRATORY
SYMPTOMS AMONG CHILDREN
Statistics for SO2
Reference, Study
Location, and Period
Study Population
Averaging Time. Air ^ual1
Mean(SD) SO2
Levels (ppb) 98th % 99th %
ity iiata tpp
Range
0)
Upper
Percentile
Standardized Odds Ratio
(95% CI)a
EUROPE (cont'd)
Van der Zee etal. (1999)
(cont'd)
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
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
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
a24-h avg SO2 and 12-h avg SO2 standardized to 10-ppb incremental change; 3-h avg SO2 standardized to 20-ppb incremental change; and 1-h max SO2 standardized to 40-ppb incremental change.
NR = Not Reported
BHR = Bronchial Hyperresponsiveness
NR = Not Reported
BHR = Bronchial Hyperresponsiveness
-------
TABLE 5A-4. EFFECTS OF SHORT-TERM SO2 EXPOSURE ON EMERGENCY DEPARTMENT VISITS
AND HOSPITAL ADMISSIONS FOR RESPIRATORY OUTCOMES
to
o
o
>
o
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Reference, Study
Location, and Period Study Population
EMERGENCY DEPARTMENT VISITS - ALL
Statistics for SO2
Averaging Time. Air Quality Data (ppb)
Mean (SD) SO2 Upper
Levels (ppb) 98th % 99th % Range Percentile
RESPIRATORY
Standardized Percent
Excess Risk (95% CI)
UNITED STATES
Wilson et al. (2005) ~ 84,000 ED visits
Portland, ME
1998-2000
Manchester, NH
1996-2000
Peel et al. (2005) 484,830 ED visits, all
Atlanta, GA ages from 3 1 hospitals
Jan 1 993 -Aug 2000
1-hmax: NR NR NR NR
Portland:
11.1(9.1)
Manchester:
16.5 (14.7)
1-hmax: 16.5 NR NR NR 90th: 39.0
(17.1)
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: 6%
(1, 12)
0-14: 5.4% (-12.8, 25.8)
15-64: 11.0% (0.0, 22.7)
65+: 11.0% (-15.2, 48.4)
1.6% (-0.6, 3.8)
EUROPE
Atkinson et al. 98,685 ED visits from
(1999b) 12 hospitals
London, United
Kingdom
Jan 1992-Dec 1994
24-havg: 8.0 NR NR 2.8,30.9 50th: 7.4
(2.9) 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)
-------
TABLE 5A-4 (cont'd). EFFECTS OF SHORT-TERM SO2 EXPOSURE ON EMERGENCY DEPARTMENT VISITS
AND HOSPITAL ADMISSIONS FOR RESPIRATORY OUTCOMES
to
o
o
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Reference, Study
Location, and
Period Study Population
Averaging Time,
Mean (SD) SO2
Levels (ppb)
Statistics for SO2
Air Quality Data (ppb)
Upper Standardized Percent
98th % 99th % Range Percentile Excess Risk (95% CI)
EMERGENCY DEPARTMENT VISITS - ASTHMA
UNITED STATES
Jaffe et al. (2003) 4,4 16 ED visits for
Cincinnati, Cleveland, asthma, age 5-34
Columbus, OH
Jul 1991-Jun 1996
Wilson et al. (2005) ~ 84,000 ED visits
Portland, ME
1998-2000
Manchester, NH
1996-2000
Peel et al. (2005) Asthma ED visits, all
Atlanta, GA ages and 2-18 yrs
Jan 1993-Aug 2000 from 31 hospitals
24-h avg:
Cincinnati:
13.5 (9.4)
Cleveland:
14.7 (9.5)
Columbus:
4.2 (3.2)
1-hmax:
Portland:
11.1(9.1)
Manchester:
16.5 (14.7)
1-hmax: 16.5
(17.1)
NR NR Cincinnati: NR Cincinnati: 17.3% (4.7, 30.8)
0.6,49.6 Cleveland: 3.1% (-3.8, 10.7)
Cleveland: Columbus: 13.1% (-14.2, 48.6)
0.98,62.8 All Cities: 6.2% (0.5, 11.6)
Columbus:
0,21.4
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)
NR NR NR 90th: 39.0 0.2% (-3.2, 3.4)
-------
TABLE 5A-4 (cont'd). EFFECTS OF SHORT-TERM SO2 EXPOSURE ON EMERGENCY DEPARTMENT VISITS
AND HOSPITAL ADMISSIONS FOR RESPIRATORY OUTCOMES
{V
t^) '
CD
^
to
O
o
>
1
to
Reference, Study
Location, and
Period
EUROPE
Atkinson et al.
(1999b)
London, United
Kingdom
Jan 1992-Dec 1994
Hajatetal. (1999)
London, United
Kingdom
1992-1994
Statistics for SO2
Averaging Time. Air Quality Data (ppb)
Mean (SD) SO2
Study Population Levels (ppb) 98th % 99th % Range
98,685 ED visits from 24-havg: 8.0(2.9) NR NR 2.8,30.9
12 hospitals
General practitioner Allyr: NR NR NR
visits for asthma 24-h avg: 8.0 (2.9)
Warm:
24-havg: 7.7(2.4)
Cool:
24-havg: 8.3(3.4)
Upper
Percentile
50th: 7.4
90th: 11.6
All yr:
90th: 11.6
Warm:
90th: 10.7
Cool:
90th: 12.4
Standardized Percent
Excess Risk (95% CI)
All ages: 7.4%
(2.3, 12.8)
0-14: 15.0% (7.1, 23. 5)
15-64: 6.3%
(-0.8, 13.8)
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)
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Boutin-Forzano et al.
(2004)
Marseille, France
Apr 1997-Mar 1998
Galan et al. (2003)
Madrid, Spain
1995-1998
549 ED visits for
asthma
4,827 ED visits for
asthma
24-havg: 8.5
NR
NR
0.0,35.3
NR
24-havg: 8.9(5.8) NR
NR
1.9,45.6
50th: 7.0
75th: 11.8
90th: 16.5
3-49: 0.6% (-1.4, 2.7)
All ages: 4.9%
(-4.2, 15.0)
-------
TABLE 5A-4 (cont'd). EFFECTS OF SHORT-TERM SO2 EXPOSURE ON EMERGENCY DEPARTMENT VISITS
AND HOSPITAL ADMISSIONS FOR RESPIRATORY OUTCOMES
to
o
o
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Reference, Study
Location, and
Period Study Population
Averaging Time,
Mean (SD) SO2
Levels (ppb)
Statistics for SO2
Air Quality Data (ppb)
Upper
98th % 99th % Range Percentile
Standardized Percent
Excess Risk (95% CI)
EUROPE (cont'd)
Tenias et al. (1998) 734 ED visits for
Valencia, Spain asthma
1993-1995
Sunyer et al. (1997) All ED visits for
Multicity, Europe asthma
(Barcelona, Helsinki,
Paris, London)
1986-1992
Castellsague et al. ED visits for asthma
(1995) from 4 hospitals
Barcelona, Spain
1986-1989
24-havg: 10.0
Cold: 11.9
Warm: 8.2
1-hmax: 21.2
Cold: 24.3
Warm: 18.1
24-h median:
Barcelona: 15.4
Helsinki: 6.0
London: 11.6
Paris: 8.6
24-h avg:
Summer: 15.3
Winter: 19.5
NR NR NR 24-havg:
50th: 9.8
75th: 12.9
95th: 16.0
1-hmax:
50th: 19.6
75th: 27.1
95th: 35.8
NR NR Barcelona: 0.8, NR
60.2
Helsinki: 1.1,35.7
London: 3.4, 37.6
Paris: 0.4,82.3
NR NR NR Summer:
50th: 13.5
75th: 20.3
95th: 30.8
Winter:
50th: 18.4
75th: 25.2
95th: 35.3
>14yrs: 13.9%
(-7.0,39.4)
0-14: 3.2% (-0.2, 6.8)
15-64: 0.2%
(-2.2,2.6)
Summer:
15-64: 5.5%
(-2.1, 13.8)
Winter:
15-64: 2.1%
(-4.2,9.0)
-------
to
o
o
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
TABLE 5A-4 (cont'd). EFFECTS OF SHORT-TERM SO2 EXPOSURE ON EMERGENCY DEPARTMENT VISITS
AND HOSPITAL ADMISSIONS FOR RESPIRATORY OUTCOMES
Reference, Study
Location, and Period
Study Population
Averaging
Time, Mean
(SD) SO2
Levels (ppb)
Statistics for SO2
Air Quality Data (ppb)
Upper
98th % 99th % Range Percentile
Standardized Percent
Excess Risk (95% CI)
EMERGENCY DEPARTMENT VISITS - COPD
UNITED STATES
Peel et al. (2005)
Atlanta, GA
Jan 1 993 -Aug 2000
COPD ED visits, all
ages from 3 1 hospitals
1-hmax:
16.5(17.1)
NR NR NR 90th: 39.0
3.2% (-3, 10)
HOSPITAL ADMISSIONS - ALL RESPIRATORY
UNITED STATES
Schwartz (1995)
New Haven, CT
Tacoma, WA
1988-1990
~ 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. (2006)
Vancouver, BC
Jun 1995-Mar 1999
~ 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)
5% (- 1, 12)
Yang et al. (2003)
Vancouver, BC
1986-1998
*Burnett et al. (2001)
Toronto, ON
1980-1994
Respiratory hospital 24-h avg: 4.84 NR
admissions among (2.84)
young children (<3 yrs)
and elderly (>65 yrs)
NR NR
75th: 6.25 <3 yrs: 3% (-6, 15)
100th: 24.00 >65yrs: 5.8% (0.0,
11.9)
All respiratory
admissions for young
children (<2 yrs)
1-hmax: 11.8 NR
55
NR
75th: 15
95th: 32
100th: 110
11% (-0.3, 23.6)
-------
TABLE 5A-4 (cont'd). EFFECTS OF SHORT-TERM SO2 EXPOSURE ON EMERGENCY DEPARTMENT VISITS
AND HOSPITAL ADMISSIONS FOR RESPIRATORY OUTCOMES
to
o
o
>
(^
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Reference, Study
Location, and
Period
Averaging Time,
Mean (SD) SO2
Study Population Levels (ppb)
Statistics for SO2
Air Quality Data (ppb)
98th 99th Upper Standardized Percent
% % Range Percentile Excess Risk (95% CI)
CANADA (cont'd)
Luginaah et al.
(2005) Windsor, ON
Apr 1995-Dec 2000
All respiratory 1-h max: 27.5(16.5)
admissions ages
0-14, 15-64, and
65+ from
4 hospitals
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)
AUSTRALIA
Barnett et al. (2005)
Multicity,
Australia/New
All respiratory 24-h avg:
hospital Auckland: 4.3
admissions Brisbane: 1.8
NR NR 24-havg: NR 1-4 yrs: 5.1% (0.0, 9.1)
Auckland: 0,24.3 5-14: 3.7% (-9.9, 19.5)
Brisbane: 0, 8.2
Zealand; (Auckland,
Brisbane, Canberra,
Christchurch,
Melbourne, Perth,
Sydney)
1998-2001
Christchurch: 2.8
Sydney: 0.9
NA in Canberra,
Melbourne, and Perth
1-h max:
Brisbane: 7.6
Christchurch: 10.1
Sydney: 3.7
NA in Auckland,
Canberra, Melbourne,
and Perth
Christchurch: 0,
11.9
Sydney: 0,3.9
1-h max
Christchurch: 0.1,
42.1
Sydney: 0.1,20.2
-------
TABLE 5A-4 (cont'd). EFFECTS OF SHORT-TERM SO2 EXPOSURE ON EMERGENCY DEPARTMENT VISITS
AND HOSPITAL ADMISSIONS FOR RESPIRATORY OUTCOMES
V,V
(J '
CD
i— i
K>
O
O
1
a(
O
r^
H
i
0
o
H
O
c|
o
H
W
O
O
HH
H
W
Reference, Study
Location, and
Period
AUSTRALIA (cont'd)
Petroeschevsky et al.
(2001)
Brisbane, Australia
1987-1994
EUROPE
Oftedal et al. (2003)
Drammen, Norway
1994-2000
Fuscoetal. (2001)
Rome, Italy
Period of study:
1/1995-10/1997
Llorca et al. (2005)
Torrelavega, Spain
1992-1995
Anderson et al.
(2001)
West Midlands
conurbation, United
Kingdom
Oct 1994-Dec 1996
Study Population
33,710 hospital
admissions
All respiratory
hospital
admissions
All respiratory
hospital
admissions
Hospital
admissions from
one hospital
Hospital
admissions
stratified by age
Statistics for SO2
Averaging Time. Air Quality Data (ppb)
Mean (SD) SO2
Levels (ppb) 98th % 99th % Range
24-havg: 4.1 NR NR NR
1-hmax: 9.2
24-havg: 1.1 NR NR NR
(0.8)
24-havg: 3.4 NR NR NR
(2.2)
24-havg: 5.0 NR NR NR
(6.3)
24-havg: 7.2 NR NR 1.9,59.8
(4.7)
Upper Standardized Percent
Percentile Excess Risk (95 % CI)
NR All ages: -5.9% (-12.4, 1.
0-14: 8.0% (-2.9, 20.1)
15-64: -21.6% (-34.4, -6.
NR All ages: 71.8% (15.5, 152
50th: 3.0 All age: 1.6% (-4.9, 8.8)
75th: 4.5 0-14: -2.7% (-4.6, 10.8)
NR All ages: 1.0% (-2.8, 4.7)
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)
1)
2)
.7)
-------
TABLE 5A-4 (cont'd). EFFECTS OF SHORT-TERM SO2 EXPOSURE ON EMERGENCY DEPARTMENT VISITS
AND HOSPITAL ADMISSIONS FOR RESPIRATORY OUTCOMES
to
o
o
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Reference, Study
Location, and
Period
Study Population
Statistics for SO2
Averaging Time. Air Quality Data (ppb)
Mean (SD) SO2
Levels (ppb) 98th % 99th % Range
Upper Standardized Percent
Percentile Excess Risk (95% CI)
EUROPE (cont'd)
Atkinson et al.
(1999a)
London, England
1992-1994
Schouten et al. (1996)
Multicity, The
Netherlands
(Amsterdam,
Rotterdam)
Period of study:
Apr 1977-Sep 1989
Spixetal. (1998)
Multicity (London,
Amsterdam,
Rotterdam, Paris,
Milan), Europe
1977-1991
165,032 hospital
admissions
All respiratory
hospital
admissions
All respiratory
hospital
admissions
24-havg: 8.0 NR NR 2.8,30.9
(2.9)
24-havg: NR NR NR
Amsterdam: 10.5
Rotterdam: 15.0
1-hmax:
Amsterdam: 24.4
Rotterdam: 37.2
24-havg: NR NR NR
London: 10.9
Amsterdam: 7.9
Rotterdam: 9.4
Paris: 8.6
Milan: 24.8
50th: 7.4 All ages: 3.0% (0.4, 5.6)
90th: 11.7 0-14: 7.7% (3.8, 11.7)
15-64: 2.8% (-1.2, 7.0)
65+: 3. 3% (-0.1, 6.9)
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)
NR 15-64: 0.5% (-0.4, 1.3)
65+: 1.1% (0.3, 2.4)
-------
TABLE 5A-4 (cont'd). EFFECTS OF SHORT-TERM SO2 EXPOSURE ON EMERGENCY DEPARTMENT VISITS
AND HOSPITAL ADMISSIONS FOR RESPIRATORY OUTCOMES
to
o
o
oo
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Reference, Study
Location, and
Period
Averaging Time,
Mean (SD) SO2
Study Population Levels (ppb) 98th %
Statistics for SO2
Air Quality Data (ppb)
Upper
99th % Range Percentile
Standardized Percent
Excess Risk (95% CI)
EUROPE (cont'd)
Dab etal. (1996)
Paris, France
Period of study:
1/1/87-9/30/92
Ponce de Leon et al.
(1996)
London, England
1987-1988
1991-1992
Hospital All year: NR
admissions from 24-havg: 11.2
27 hospitals 1-hmax: 22.5
Warm season
24-havg: 7.6
1-hmax: 16.1
Cold season
24-havg: 15.1
1-hmax: 29.4
19,901 hospital 24-havg: 12.1 NR
admissions (4.7)
NR NR All year:
24-h avg:
99th: 50.0
1-hmax:
99th: 87.5
Warm season
99th: 18.5
1-hmax:
99th: 50.3
Cold season
24-h avg:
99th: 56.0
1-hmax:
99th: 100.9
NR NR 50th: 11.7
75th: 14.7
90th: 17.7
95th: 20.3
All ages: 1.1% (0.1, 2.1)
All ages: 0.8 (-0.7, 2.4)
0-14: 0.9 (-1.5, 3.3)
15-64: 2.0% (-0.5, 4.7)
65+: 2.0% (-0.3, 4.4)
-------
TABLE 5A-4 (cont'd). EFFECTS OF SHORT-TERM SO2 EXPOSURE ON EMERGENCY DEPARTMENT VISITS
AND HOSPITAL ADMISSIONS FOR RESPIRATORY OUTCOMES
to
o
o
VO
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Reference, Study
Location, and
Period
Study
Population
Averaging Time,
Mean (SD) SO2
Levels (ppb) 98th %
Statistics for SO2
Air Quality Data (ppb)
99th % Range Upper Percentile
Standardized Percent
Excess Risk (95% CI)
EUROPE (cont'd)
Walters etal. 1994
Birmingham, United
Kingdom
1988-1990
Hagen et al. (2000)
Drammen, Norway
1994-1997
All respiratory
hospital
admissions
Hospital
admissions for
all respiratory
outcomes
24-h avg: NR
All year: 14.7
Spring: 16.1
Summer: 14.2
Autumn: 15.4
Winter: 12.9
24-h avg: NR
Winter: 21
Spring: 18
Summer: 15
Autumn: 19
Number of
monitors: 1
NR NR Max: 47.5
NR Winter: 11, NR
33
Spring: 13,
29
Summer: 5,
24
Autumn: 16,
23
All ages:
Summer: 1.5% (0.3, 2.7)
Winter: 4.5% (2.3, 6.5)
All ages: 92.8% (16.8,
218.8)
LATIN AMERICA
Gouveia and Fletcher
(2000)
Sao Paulo, Brazil
Nov 1992-Sep 1994
All respiratory
hospital
admissions
24-h avg: 6.9(3.4) NR
NR 1.2,22.9 50th: 6.2
75th: 8.3
95th: 13.5
<5yrs: 3.7% (-1.7, 9.4)
ASIA
Wong etal. (1999)
Hong Kong
1994-1995
Hospital
admissions from
12 hospitals
24-h avg: 6.4 NR
NR 1.0,25.7 75th: 9.4
0-4 yrs: 1.3% (-2.4, 4.9)
5-64: 2.1% (-1.1, 5.7)
65+: 6.2% (3.2, 9.9)
-------
TABLE 5A-4 (cont'd). EFFECTS OF SHORT-TERM SO2 EXPOSURE ON EMERGENCY DEPARTMENT VISITS
AND HOSPITAL ADMISSIONS FOR RESPIRATORY OUTCOMES
to
o
o
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Reference, Study
Location, and
Period
Study
Population
Averaging Time,
Mean (SD) SO2
Levels (ppb)
Statistics for SO2
Air Quality Data (ppb)
98th % 99th % Range Upper Percentile
Standardized Percent
Excess Risk (95% CI)
HOSPITAL ADMISSIONS - ASTHMA
UNITED STATES
Sheppard et al.
(1999; reanalysis
2003)
Seattle, WA
1987-1994
7,837 asthma
hospital
admissions for
patients <65 yrs
24-havg: 8
NR NR NR 75th: 10.0
90th: 13.0
<65yrs: 4.0% (-4.0, 10.3)
CANADA
*Burnett et al. (1999)
Toronto, ON
1980-1994
Lin et al. (2003)
Toronto, ON
1981-1993
Asthma hospital
admissions
7,3 19 asthma
hospital
admissions
among 6-12 yr
olds
24-havg: 5.35
24-havg: 5.36
(5.90)
NR NR NR 75th: 8
95th: 17
100th: 57
NR NR 0,57.0 75th: 8.00
1.9% (-0.2, 4.0)
Boys: 0% (-7. 1,7.2)
Girls: 5. 8% (-4.3, 16.1)
-------
to
o
o
TABLE 5A-4 (cont'd). EFFECTS OF SHORT-TERM SO2 EXPOSURE ON EMERGENCY DEPARTMENT VISITS
AND HOSPITAL ADMISSIONS FOR RESPIRATORY OUTCOMES
Reference, Study
Location, and
Period
Study
Population
Averaging Time,
Mean (SD) SO2
Levels (ppb)
Statistics for SO2
Air Quality Data (ppb)
98th % 99th % Range
Upper
Percentile
Standardized Percent
Excess Risk (95% CI)
AUSTRALIA
Barnett et al. (2005)
Multicity,
Australia/New
All respiratory
hospital
admissions
24-h avg:
Auckland: 4.3
Brisbane: 1.8
NR NR 24-h avg:
Auckland: 0,24.3
Brisbane: 0, 8.2
NR
1-4 yrs: 6.4% (-7.8, 22.5)
5-14: 6.2% (-10.1, 25.4)
Zealand; (Auckland,
Brisbane, Canberra,
Christchurch,
Melbourne, Perth,
Sydney)
Period of study:
1998-2001
Christchurch: 2.8
Sydney: 0.9
NA in Canberra,
Melbourne, and Perth
1-hmax:
Brisbane: 7.6
Christchurch: 10.1
Sydney: 3.7
NA in Auckland,
Canberra, Melbourne,
and Perth
Christchurch: 0,
11.9
Sydney: 0,3.9
1-hmax:
Brisbane: 0, 46.5
Christchurch: 0.1,
42.1
Sydney: 0.1,20.2
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Petroeschevsky et al.
(2001)
Brisbane, Australia
1987-1994
33,710 hospital
admissions
24-h avg: 4.1
1-hmax: 9.2
NR
NR
NR
NR All ages: 8.0% (3.0, 13.1)
0-4: 22.4% (8.7, 37.7)
5-14: 21.1%(-5.5, 55.1)
15-64: 3.3% (-10.5, 11.8)
65+: 12.1% (1.9, 23.4)
EUROPE
Fuscoetal. (2001)
Rome, Italy
Jan 1995-Oct 1997
All respiratory
hospital
admissions
24-h avg: 3.4(2.2)
NR
NR
NR
50th: 3.0 All ages: -5.7% (-23.2,
75th: 4.5 15.9)
0-14: -9.7% (-34.6, 25.2)
-------
TABLE 5A-4 (cont'd). EFFECTS OF SHORT-TERM SO2 EXPOSURE ON EMERGENCY DEPARTMENT VISITS
AND HOSPITAL ADMISSIONS FOR RESPIRATORY OUTCOMES
to
o
o
to
to
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Reference, Study
Location, and Period
Study
Population
Averaging Time,
Mean (SD) SO2
Levels (ppb)
Statistics for SO2
Air Quality Data (ppb)
Upper Standardized Percent
98th % 99th % Range Percentile Excess Risk (95% CI)
EUROPE (cont'd)
Atkinson et al. (1999a)
London, England
1992-1994
Schouten et al. (1996)
Multicity, the
Netherlands
(Amsterdam,
Rotterdam)
Period of study:
Apr 1977-Sep 1989
Dab etal. (1996)
Paris, France
Jan 1987-Sep 1992
165,032
hospital
admissions
All respiratory
hospital
admissions
Hospital
admissions
from 27
hospitals
24-havg: 8.0(2.9)
24-h avg:
Amsterdam: 10.5
Rotterdam: 15.0
1-hmax:
Amsterdam: 24.4
Rotterdam: 37.2
All year:
24-havg: 11.2
1-hmax: 22.5
Warm season
24-havg: 7.6
1-hmax: 16.1
Cold season
24-havg: 15.1
1-hmax: 29.4
NR NR 2.8,30.9 50th: 7.4 All ages: 5.0% (0.6, 9.6)
90th: 11.7 0-14: 10.1% (4.3, 16.2)
15-64: 6.8% (-0.3, 14.5)
65+: 9.5% (-2.3, 22.7)
NR NR NR NR Amsterdam:
All ages: -6.0% (-10.7, -1.1)
NR NR NR All year: All ages: 1.8% (0.1, 3.6)
24 h avg:
99th: 50.0
1-hmax:
99th: 87.5
Warm season
99th: 18.5
1-hmax:
99th: 50.3
Cold season
24-h avg:
99th: 56.0
1-hmax:
99th: 100.9
-------
TABLE 5A-4 (cont'd). EFFECTS OF SHORT-TERM SO2 EXPOSURE ON EMERGENCY DEPARTMENT VISITS
1 1 1
<*->
i— i
K>
O
O
0,
1
UJ
>
H
I
0
o
0
H
O
o
H
W
O
O
1 — I
H
W
AND HOSPITAL ADMISSIONS FOR RESPIRATORY OUTCOMES
Reference, Study
Location, and
Period
EUROPE (cont'd)
Anderson et al.
(1998)
London, England
Apr 1987-Feb 1992
Walters etal. (1994)
Birmingham,
United Kingdom
1988-1990
LATIN AMERICA
Gouveia and Fletcher
(2000)
Sao Paulo, Brazil
Nov 1992-Sep 1994
ASIA
Wong etal. (1999)
Hong Kong, China
1994-1995
Lee et al. (2006)
Hong Kong, China
1997-2002
Study Population
All hospital
admissions for
asthma
All respiratory
hospital admissions
All respiratory
hospital admissions
Hospital admissions
from 12 hospitals
26,663 hospital
admissions for
asthma
Averaging Time,
Statistics for SO2
Air Quality Data (ppb)
Mean (SD) SO2 Upper
Levels (ppb) 98th % 99th % Range Percentile
24-havg: 12.0(4.4) NR
24-havg: NR
All year: 14.7
Spring: 16.1
Summer: 14.2
Autumn: 15.4
Winter: 12.9
24-havg: 6.9(3.4) NR
24-havg: 6.4 NR
24-havg: 6.6(4.0) NR
NR 3.4,37.6 50th: 11.6
75th: 14.3
90th: 17.3
95th: 19.5
NR NR Max: 47.5
NR 1.2,22.9 50th: 6.2
75th: 8.3
95th: 13.5
NR 1.0,25.7 75th: 9.4
NR NR 50th: 5.7
75th: 8.2
Standardized Percent
Excess Risk (95% CI)
All ages
0-14: 0.
15-64:
: 2.8% (1.2, 4.3)
5% (0.1, 1.0)
-0.7% (-2.7, 1.3)
65+: 3.1% (-0.7, 7.0)
Summer:
All ages
Winter:
All ages
<5 yrs:
All ages
<18yrs:
: 0.4% (-2.8, 9.2)
: 0.7% (-2.2, 1.6)
10.4% (-1.9, 24.2)
: 4.6% (-0.5, 9.9)
-3. 7% (-6.7, -0.6)
-------
TABLE 5A-4 (cont'd). EFFECTS OF SHORT-TERM SO2 EXPOSURE ON EMERGENCY DEPARTMENT VISITS
AND HOSPITAL ADMISSIONS FOR RESPIRATORY OUTCOMES
u
^
to
o
o
1
4^
O
H
6
o
Z,
O
H
O
O
H
W
O
O
HH
H
W
Reference, Study
Location, and
Period Study Population
HOSPITAL ADMISSIONS - COPD
UNITED STATES
Moolgavkar (2000; Hospital admissions
reanalysis 2003)
Chicago, Los
Angeles, Phoenix,
1987-1995
CANADA
Yang (2005) COPD admissions
Vancouver, BC among elderly (65+)
1994-1998
Burnett et al. ( 1 999) COPD hospital
Toronto, ON admissions
1980-1994
a 24-h avg SO2 standardized to 10 ppb incremental change;
Statistics for SO2
Averaging Time. Air Quality Data (ppb)
Mean (SD) SO2 Upper
Levels (ppb) 98th % 99th % Range Percentile
24-havg: NR NR Chicago: 0.5, Chicago:
Chicago: 6; 36 75th: 8
LA: 2; LA: 0, 16 LA: 75th: 4
Phoenix: 2 Phoenix: 0, 14 Phoenix:
75th: 4
24-havg: 3.79 NR NR 0.75,22.67 NR
(2.12)
24-havg: 5.35 NR NR NR 75th: 8
95th: 17
100th: 57
l-h max SO2 standardized to 40 ppb incremental change.
Standardized Percent
Excess Risk (95% CI)
Chicago: 5% (1.9, 8.2)
0.3% (-26, 15)
7.3% (-7, 23)
15% (-3.9, 31.6)
0.1% (-2.1, 2.3)
* Analyses using Poisson GAM with default convergence criteria.
NA: Not Available
NR: Not Reported
-------
6. REFERENCES
4 Abbey, D. E.; Nishino, N.; McDonnell, W. F.; Burchette, R. 1; Knutsen, S. F.; Beeson, W. L.;
5 Yang, J. X. (1999) Long-term inhalable particles and other air pollutants related to
6 mortality in nonsmokers. Am. J. Respir. Crit. Care Med. 159: 373-382.
7 Abe, M. (1967) Effects of mixed NO2-SO2 gas on human pulmonary functions: effects of air
8 pollution on the human body. Bull. Tokyo Med. Dent. Univ. 14: 415-433.
9 Abraham, W. M.; Oliver, W., Jr.; Welker, M. J.; King, M. M.; Wanner, A.; Sackner, M. A.
10 (1981) Differences in airway reactivity in normal and allergic sheep after exposure to
11 sulfur dioxide. J. Appl. Physiol. 51: 1651-1656.
12 Ackermann-Liebrich, U.; Leuenberger, P.; Schwartz, J.; Schindler, C.; Monn, C.; Bolognini, B.;
13 Bongard, J. P.; Brandli, O.; Domenighetti, G.; Elsasser, S.; Grize, L.; Karrer, W.; Keller,
14 R.; Keller-Wossidlo, H.; Kunzli, N.; Martin, B. W.; Medici, T. C.; Perruchoud, A. P.;
15 Schoni, M. H.; Tschopp, J. M.; Villiger, B.; Wuthrich, B.; Zellweger, J. P.; Zemp, E.
16 (1997) Lung function and long term exposure to air pollutants in Switzerland. Am. J.
17 Respir. Crit. Care Med. 155: 122-129.
18 Adamkiewicz, G.; Ebelt, S.; Syring, M.; Slater, J.; Speizer, F. E.; Schwartz, J.; Suh, H.; Gold, D.
19 R. (2004) Association between air pollution exposure and exhaled nitric oxide in an
20 elderly population. Thorax 59: 204-209.
21 Adams, P. F.; Hendershot, G. E.; Marano, M. A. (1999) Current estimates from the National
22 Health Interview Survey, 1996. Hyattsville, MD: U.S. Department of Health and Human
23 Services, Public Health Service, National Center for Health Statistics; DHHS publication
24 no. (PHS) 99-1528. (Vital and health statistics: v. 10, data from the National Health
25 Survey, no. 200). Available:
26 http://www.cdc.gov/nchs/products/pubs/pubd/series/srlO/pre-200/pre-200.htm [12
27 March, 2001].
28 Agar, A.; Kii9iikatay, V.; Yargi9oglu, P.; Aktekin, B.; Kipmen-Korgun, S.; Gumus.lu, D.;
29 Apaydin, C. (2000) The effect of sulfur dioxide inhalation on visual evoked potentials,
30 antioxidant status, and lipid peroxidation in alloxan-induced diabetic rats. Arch. Environ.
31 Contam. Toxicol. 39: 257-264.
32 Aharonson, E. F.; Menkes, H.; Gurtner, G.; Swift, D. L.; Proctor, D. F. (1974) Effect of
33 respiratory airflow rate on removal of soluble vapors by the nose. J. Appl. Physiol. 37:
34 654-657.
35 Amdur, M. O. (1966) Respiratory absorption data and SC>2 dose-response curves. Arch. Environ.
36 Health 12: 729-732.
37 Amdur, M. O.; Melvin, W. W., Jr.; Drinker, P. (1953) Effects of inhalation of sulphur dioxide by
38 man. Lancet 2: 758-759.
39 Amdur, M. O.; McCarthy, J. F.; Gill, M. W. (1983) Effect of mixing conditions on irritant
40 potency of zinc oxide and sulfur dioxide. Am. Ind. Hyg. Assoc. J. 44: 7-13.
September 2007 6-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 American Academy of Pediatrics, Committee on Environmental Health. (2004) Ambient air
2 pollution: health hazards to children. Pediatrics 114: 1699-1707.
3 American Thoracic Society. (1985) Guidelines as to what constitutes an adverse respiratory
4 health effect, with special reference to epidemiologic studies of air pollution. Am. Rev.
5 Respir. Dis. 131: 666-668.
6 American Thoracic Society. (2000) What constitutes an adverse health effect of air pollution?
7 Am. J. Respir. Crit. Care Med. 161: 665-673.
8 Andersen, I; Lundqvist, G. R.; Jensen, P. L.; Proctor, D. F. (1974) Human response to controlled
9 levels of sulfur dioxide. Arch. Environ. Health 28: 31-39.
10 Anderson, H. R.; Ponce de Leon, A.; Bland, J. M.; Bower, J. S.; Emberlin, J.; Strachen, D. P.
11 (1998) Air pollution, pollens, and daily admissions for asthma in London 1987-92.
12 Thorax 53: 842-848.
13 Anderson, H. R.; Bremner, S. A.; Atkinson, R. W.; Harrison, R. M.; Walters, S. (2001)
14 Particulate matter and daily mortality and hospital admissions in the west midlands
15 conurbation of the United Kingdom: associations with fine and coarse particles, black
16 smoke and sulphate. Occup. Environ. Med. 58: 504-510.
17 Atkinson, R. W.; Bremner, S. A.; Anderson, H. R.; Strachan, D. P.; Bland, J. M.; Ponce de Leon,
18 A. (1999a) Short-term associations between emergency hospital admissions for
19 respiratory and cardiovascular disease and outdoor air pollution in London. Arch.
20 Environ. Health 54: 398-411.
21 Atkinson, R. W.; Anderson, H. R.; Strachan, D. P.; Bland, J. M.; Bremner, S. A.; Ponce de Leon,
22 A. (1999b) Short-term associations between outdoor air pollution and visits to accident
23 and emergency departments in London for respiratory complaints. Eur. Respir. J. 13:
24 257-265.
25 Azoulay-Dupuis, E.; Bouley, G.; Blayo, M. C. (1982) Effects of sulfur dioxide on resistance to
26 bacterial infection in mice. Environ. Res. 29: 312-319.
27 Ballester, F.; Tenias, J. M.; Perez-Hoyos, S. (2001) Air pollution and emergency hospital
28 admissions for cardiovascular diseases in Valencia, Spain. J. Epidemiol. Community
29 Health 55: 57-65.
30 Ballester, F.; Saez, M.; Perez-Hoyos, S.; Ifiiguez, C.; Gandarillas, A.; Tobias, A.; Bellido, J.;
31 Taracido, M.; Arribas, F.; Daponte, A.; Alonso, E.; Canada, A.; Guillen-Grima, F.;
32 Cirera, L.; Perez-Boillos, M. J.; Saurina, C.; Gomez, F.; Tenias, J. M. (2002) The
33 EMECAM project: a multicentre study on air pollution and mortality in Spain: combined
34 results for particulates and for sulfur dioxide. Occup. Environ. Med. 59: 300-308.
35 Ballester, F.; Rodriguez, P.; Ifiiguez, C.; Saez, M.; Daponte, A.; Galan, L; Taracido, M.; Arribas,
36 F.; Bellido, J.; Cirarda, F. B.; Canada, A.; Guillen, J. J.; Guillen-Grima, F.; Lopez, E.;
37 Perez-Hoyos, S.; Lertxundi, A.; Toro, S. (2006) Air pollution and cardiovascular
38 admisisons association in Spain: results within the EMECAS project. J. Epidemiol.
39 Community Health 60: 328-336.
40 Balmes, J. R.; Fine, J. M.; Sheppard, D. (1987) Symptomatic bronchoconstriction after short-
41 term inhalation of sulfur dioxide. Am. Rev. Respir. Dis. 136: 1117-1121.
September 2007 6-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 Barnett, A. G.; Williams, G. M.; Schwartz, J.; Neller, A. H.; Best, T. L.; Petroeschevsky, A. L.;
2 Simpson, R. W. (2005) Air pollution and child respiratory health: a case-crossover study
3 in Australia and New Zealand. Am. J. Respir. Crit. Care Med. 171: 1272-1278.
4 Barthelemy, P.; Badier, M.; Jammes, Y. (1988) Interaction between SC>2 and cold-induced
5 bronchospasm in anesthetized rabbits. Respir. Physiol. 71: 1-10.
6 Bascom, R.; Bromberg, P. A.; Costa, D. A.; Devlin, R.; Dockery, D. W.; Frampton, M. W.;
7 Lambert, W.; Samet, J. M.; Speizer, F. E.; Utell, M. (1996) Health effects of outdoor air
8 pollution. Am. J. Respir. Crit. Care Med. 153: 3-50.
9 Baskurt, O. K. (1988) Acute hematologic and hemorheologic effects of sulfur dioxide inhalation.
10 Arch. Environ. Health 43: 344-348.
11 Bedi, J. F.; Folinsbee, L. J.; Horvath, S. M.; Ebenstein, R. S. (1979) Human exposure to sulfur
12 dioxide and ozone: absence of a synergistic effect. Arch. Environ. Health 34: 233-239.
13 Bedi, J. F.; Folinsbee, L. J.; Horvath, S. M. (1984) Pulmonary function effects of 1.0 and 2.0
14 ppm sulfur dioxide exposure in active young male non-smokers. J. Air Pollut. Control
15 Assoc. 34: 1117-1121.
16 Bell, M. L.; Ebisu, K.; Belanger, K. (2007) Ambient air pollution and low birth weight in
17 Connecticut and Massachusetts. Environ. Health Perspect. 115: 1118-1125.
18 Ben-Jebria, A.; Full, A. P.; DeMaria, D. D.; Ball, B. A.; Ultman, J. S. (1990) Dynamics of sulfur
19 dioxide absorption in excised porcine tracheae. Environ. Res. 53: 119-134.
20 Berkowitz, G. S.; Papiernik, E. (1993) Epidemiology of preterm birth. Epidemiol. Rev. 15: 414-
21 443.
22 Berresheim, H.; Eisele, F. L.; Tanner, D. J.; Mclnnes, L. M.; Ramsey-Bell, D. C.; Covert, D. S.
23 (1993) Atmospheric sulfur chemistry and cloud condensation nuclei (CCN)
24 concentrations over the northeastern Pacific coast. J. Geophys. Res. [Atmos.] 98: 12,701-
25 12,711.
26 Berresheim, H.; Wine, P. H.; Davis, D. D. (1995) Sulfur in the atmosphere. In: Singh, H. B., ed.
27 Composition, chemistry, and climate of the atmosphere. New York, NY: Van Nostrand
28 Reinhold; pp. 251-307.
29 Bethel, R. A.; Epstein, J.; Sheppard, D.; Nadel, J. A.; Boushey, H. A. (1983) Sulfur dioxide-
30 induced bronchoconstriction in freely breathing, exercising, asthmatic subjects. Am. Rev.
31 Respir. Dis. 128: 987-990.
32 Bethel, R. A.; Sheppard, D.; Geffroy, B.; Tarn, E.; Nadel, J. A.; Boushey, H. A. (1985) Effect of
33 0.25 ppm sulfur dioxide on airway resistance in freely breathing, heavily exercising,
34 asthmatic subjects. Am. Rev. Respir. Dis. 131: 659-661.
35 Biersteker, K.; de Graaf, H.; Nass, C. A. G. (1965) Indoor air pollution in Rotterdam homes. Int.
36 J. Air Water Pollut. 9: 343-350.
37 Biggeri, A.; Baccini, M.; Bellini, P.; Terracini, B. (2005) Meta-analysis of the Italian studies of
38 short-term effects of air pollution (MISA), 1990-1999. Int. J. Occup. Environ. Health 11:
39 107-122.
September 2007 6-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 Bobak, M. (2000) Outdoor air pollution, low birth weight, and prematurity. Environ. Health
2 Perspect. 108: 173-176.
3 Boezen, M.; Schouten, J.; Rijcken, B.; Vonk, J.; Gerritsen, J.; Van Der Zee, S.; Hoek, G.;
4 Brunekreef, B.; Postma, D. (1998) Peak expiratory flow variability, bronchial
5 responsiveness, and susceptibility to ambient air pollution in adults. Am. J. Respir. Crit.
6 Care Med. 158: 1848-1854.
7 Boezen, H. M.; Van Der Zee, S. C.; Postma, D. S.; Vonk, J. M.; Gerritsen, J.; Hoek, G.;
8 Brunekreef, B.; Rijcken, B.; Schouten, J. P. (1999) Effects of ambient air pollution on
9 upper and lower respiratory symptoms and peak expiratory flow in children. Lancet 353:
10 874-878.
11 Boezen, H. M.; Vonk, J. M.; Van Der Zee, S. C.; Gerritsen, J.; Hoek, G.; Brunekreef, B.;
12 Schouten, J. P.; Postma, D. S. (2005) Susceptibility to air pollution in elderly males and
13 females. Eur. Respir. J. 25: 1018-1024.
14 Borsboom, G. J. J. M.; Van Pelt, W.; Van Houwelingen, H. C.; Van Vianen, E.G.; Schouten, J.
15 P.; Quanjer, P. H. (1999) Diurnal variation in lung function in subgroups from two Dutch
16 populations. Consequences for longitudinal analysis. Am. J. Respir. Crit. Care Med. 159:
17 1163-1171.
18 Boutin-Forzano, S.; Adel, N.; Gratecos, L.; Jullian, H.; Gamier, J. M.; Ramadour, M.;
19 Lanteaume, A.; Hamon, M.; Lafay, V.; Charpin, D. (2004) Visits to the emergency room
20 for asthma attacks and short-term variations in air pollution. A case-crossover study.
21 Respiration 71: 134-137.
22 Braga, A. L. F.; Concei9ao, G. M. S.; Pereira, L. A. A.; Kishi, H. S.; Pereira, J. C. R.; Andrade,
23 M. F.; Gon9alves, F. L. T.; Saldiva, P. H. N.; Latorre, M. R. D. O. (1999) Air pollution
24 and pediatric respiratory hospital admissions in Sao Paulo, Brazil. J. Environ. Med. 1: 95-
25 102.
26 Braga, A. L. F.; Saldiva, P. H. N.; Pereira, L. A. A.; Menezes, J. J. C.; Concei9ao, G. M. C.; Lin,
27 C. A.; Zanobetti, A.; Schwartz, J.; Dockery, D. W. (2001) Health effects of air pollution
28 exposure on children and adolescents in Sao Paulo, Brazil. Pediatr. Pulmonol. 31: 106-
29 113.
30 Brain, J. D. (1970) Free cells in the lungs: some aspects of their role, quantitation and regulation.
31 Arch. Intern. Med. 126: 477-487.
32 Brauer, M.; Koutrakis, P.; Spengler, J. D. (1989) Personal exposures to acidic aerosols and gases.
33 Environ. Sci. Technol. 23: 1408-1412.
34 Brauer, M.; Koutrakis, P.; Keeler, G. J.; Spengler, J. D. (1991) Indoor and outdoor
35 concentrations of inorganic acidic aerosols and gases. J. Air Waste Manage. Assoc. 41:
36 171-181.
37 Brauer, M.; Henderson, S.; Kirkham, T.; Lee, K. S.; Rich, K.; Teschke, K. (2002) Review of the
38 health risks associated with nitrogen dioxide and sulfur dioxide in indoor air. Vancouver,
39 British Columbia, Canada: University of British Columbia, School of Occupational and
40 Environmental Hygiene. Report to Health Canada.
September 2007 6-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 Bremner, S. A.; Anderson, H. R.; Atkinson, R. W.; McMichael, A. J.; Strachan, D. P.; Bland, J.
2 M.; Bower, J. S. (1999) Short term associations between outdoor air pollution and
3 mortality in London 1992-4. Occup. Environ. Med. 56: 237-244.
4 Brook, R. D.; Franklin, B.; Cascio, W.; Hong, Y.; Howard, G.; Lipsett, M.; Luepker, R.;
5 Mittleman, M.; Samet, J.; Smith, S. C., Jr.; Tager, I. (2004) Air pollution and
6 cardiovascular disease. A statement for healthcare professionals from the Expert Panel on
7 Population and Prevention Science of the American Heart Association. Circulation 109:
8 2655-2671.
9 Buescher, P. A.; Taylor K. P.; Davis, M. H.; Bowling, J. M. (1993) The quality of the new birth
10 certificate data: a validation study in North Carolina. Am. J. Public Health 83: 1163-
11 1165.
12 Burnett, R. T.; Goldberg, M. S. (2003) Size-fractionated particulate mass and daily mortality in
13 eight Canadian cities. In: Revised analyses of time-series studies of air pollution and
14 health. Special report. Boston, MA: Health Effects Institute; pp. 85-89. Available:
15 http://www.healtheffects.org/news.htm [16 May, 2003].
16 Burnett, R. T.; Cakmak, S.; Brook, J. R.; Krewski, D. (1997a) The role of particulate size and
17 chemistry in the association between summertime ambient air pollution and
18 hospitalization for cardiorespiratory diseases. Environ. Health Perspect. 105: 614-620.
19 Burnett, R. T.; Brook, J. R.; Yung, W. T.; Dales, R. E.; Krewski, D. (1997b) Association
20 between ozone and hospitalization for respiratory diseases in 16 Canadian cities. Environ.
21 Res. 72:24-31.
22 Burnett, R. T.; Cakmak, S.; Brook, J. R. (1998) The effect of the urban ambient air pollution mix
23 on daily mortality rates in 11 Canadian cities. Can. J. Public Health 89: 152-156.
24 Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Cakmak, S.; Brook, J. R. (1999) Effects of
25 particulate and gaseous air pollution on cardiorespiratory hospitalizations. Arch. Environ.
26 Health 54: 130-139.
27 Burnett, R. T.; Brook, J.; Dann, T.; Delocla, C.; Philips, O.; Cakmak, S.; Vincent, R.; Goldberg,
28 M. S.; Krewski, D. (2000) Association between particulate- and gas-phase components of
29 urban air pollution and daily mortality in eight Canadian cities. In: Grant, L. D., ed.
30 PM2000: particulate matter and health. Inhalation Toxicol. 12(suppl. 4): 15-39.
31 Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Raizenne, M. E.; Brook, J. R.; Dales, R. E.; Leech,
32 J. A.; Cakmak, S.; Krewski, D. (2001) Association between ozone and hospitalization for
33 acute respiratory diseases in children less than 2 years of age. Am. J. Epidemiol. 153:
34 444-452.
35 Burnett, R. T.; Stieb, D.; Brook, J. R.; Cakmak, S.; Dales, R.; Raizenne, M.; Vincent, R.; Dann,
36 T. (2004) Associations between short-term changes in nitrogen dioxide and mortality in
37 Canadian cities. Arch. Environ. Health 59: 228-236.
38 Cakmak, S.; Dales, R. E.; Vidal, C. B. (2007) Air pollution and mortality in Chile: susceptibility
39 among the elderly. Environ. Health Perspect. 115: 524-527.
40 Calvert, J. G.; Su, F.; Bottenheim, J. W.; Strausz, O. P. (1978) Mechanism of the homogeneous
41 oxidation of sulfur dioxide in the troposphere. In: Sulfur in the atmosphere: proceedings
September 2007 6-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 of the international symposium; September 1977; Dubrovnik, Yugoslavia. Atmos.
2 Environ. 12: 197-226.
3 Castellsague, J.; Sunyer, J.; Saez, M.; Anto, J. M. (1995) Short-term association between air
4 pollution and emergency room visits for asthma in Barcelona. Thorax 50: 1051-1056.
5 Chan, C.-C.; Chuang, K.-J.; Su, T.-C.; Lin, L.-Y. (2005) Association between nitrogen dioxide
6 and heart rate variability in a susceptible population. Eur. J. Cardiovasc. Prev. Rehabil.
7 12: 580-586.
8 Chang, L.-T.; Koutrakis, P.; Catalano, P. J.; Suh, H. H. (2000) Hourly personal exposures to fine
9 particles and gaseous pollutants—results from Baltimore, Maryland. J. Air Waste
10 Manage. Assoc. 50: 1223-1235.
11 Chang, C.-C.; Tsai, S.-S.; Ho, S.-C.; Yang, C.-Y. (2005) Air pollution and hospital admissions
12 for cardiovascular disease in Taipei, Taiwan. Environ. Res. 98: 114-119.
13 Chao, C. Y. (2001) Comparison between indoor and outdoor air contaminant levels in residential
14 buildings from passive sampler study. Build. Environ. 36: 999-1007.
15 Chapman, R. S.; Calafiore, D. C.; Hasselblad, V. (1985) Prevalence of persistent cough and
16 phlegm in young adults in relation to long-term ambient sulfur oxide exposure. Am. Rev.
17 Respir. Dis. 132: 261-267.
18 Chen, P.-C.; Lai, Y.-M.; Chan, C.-C.; Hwang, J.-S.; Yang, C.-Y.; Wang, J.-D. (1999) Short-term
19 effect of ozone on the pulmonary function of children in primary school. Environ. Health
20 Perspect. 107: 921-925.
21 Cifuentes, L. A.; Vega, J.; Kopfer, K.; Lave, L. B. (2000) Effect of the fine fraction of parti culate
22 matter versus the coarse mass and other pollutants on daily mortality in Santiago, Chile.
23 J. Air Waste Manage. Assoc. 50: 1287-1298.
24 Clarke, R. W.; Antonini, J. M.; Hemenway, D. R.; Frank, R.; Kleeberger, S. R.; Jakab, G. J.
25 (2000) Inhaled particle-bound sulfate: effects on pulmonary inflammatory responses and
26 alveolar macrophage function. Inhalation Toxicol. 12: 169-186.
27 Code of Federal Regulations. (2006) Lower detectable limit—(1) technical definition. C. F. R.
28 40: §53.23c.
29 Conner, M. W.; Lam, H. F.; Rogers, A. E.; Fitzgerald, S.; Amdur, M. O. (1985) Lung injury in
30 guinea pigs caused by multiple exposures to submicron zinc oxide mixed with sulfur
31 dioxide in a humidified furnace. J. Toxicol. Environ. Health 16: 101-114.
32 Crutzen, P. J. (1976) Upper limits on atmospheric ozone reductions following increased
33 application of fixed nitrogen to the soil. Geophys. Res. Lett. 3: 169-172.
34 Dab, W.; Medina, S.; Quenel, P.; Le Moullec, Y.; Le Tertre, A.; Thelot, B.; Monteil, C.;
35 Lameloise, P.; Pirard, P.; Momas, I; Ferry, R.; Festy, B. (1996) Short term respiratory
36 health effects of ambient air pollution: results of the APHEA project in Paris. In: St
37 Leger, S., ed. The APHEA project. Short term effects of air pollution on health: a
38 European approach using epidemiological time series data. J. Epidemiol. Commun.
39 Health 50(suppl. 1): S42-S46.
September 2007 6-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 Dales, R.; Burnett, R. T.; Smith-Doiron, M.; Stieb, D. M.; Brook, J. R. (2004) Air pollution and
2 sudden infant death syndrome. Pediatrics 113: 628-631.
3 Dales, R. E.; Cakmak, S.; Doiron, M. S. (2006) Gaseous air pollutants and hospitalization for
4 respiratory disease in the neonatal period. Environ. Health Perspect. 114: 1751-1754.
5 DeFrances, C. J.; Hall, M. J.; Podgornik, M. N. (2005) 2003 National Hospital Discharge
6 Survey. Hyattsville, MD: National Center for Health Statistics; DHHS publication no.
7 (PHS) 2004-1250. (Advance data from vital and health statistics; no. 359). Available:
8 http://www.cdc.gov/nchs/data/ad/ad359.pdff3 August, 2005].
9 De Leon, S. F.; Thurston, G. D.; Ito, K. (2003) Contribution of respiratory disease to
10 nonrespiratory mortality associations with air pollution. Am. J. Respir. Crit. Care Med.
11 167:1117-1123.
12 Delfino, R. J.; Gone, H.; Linn, W. S.; Pellizzari, E. D.; Hu, Y. (2003a) Asthma symptoms in
13 Hispanic children and daily ambient exposures to toxic and criteria air pollutants.
14 Environ. Health Perspect. Ill: 647-656.
15 Demokritou, P.; Kavouras, I. G.; Ferguson, S. T.; Koutrakis, P. (2001) Development and
16 laboratory performance evaluation of a multipollutant sampler for simultaneous
17 measurements for particulate and gaseous pollutants. Aerosol Sci. Technol. 35: 741-752.
18 Dentener, F.; Stevenson, D.; Ellingsen, K.; VanNoije, T.; Schultz, M.; Amann, M.; Atherton, C.;
19 Bell, N.; Bergmann, D.; Bey, I; Bouwman, L.; Butler, T.; Cofala, J.; Collins, B.; Drevet,
20 J.; Doherty, R.; Eickhout, B.; Eskes, H.; Fiore, A.; Gauss, M.; Hauglustaine, D.;
21 Horowitz, L.; Isaksen, I. S. A.; Josse, B.; Lawrence, M.; Krol, M.; Lamarque, J. F.;
22 Montanaro, V.; Muller, J. F.; Peuch, V. H.; Pitari, G.; Pyle, J.; Rast, S.; Rodriguez, J.;
23 Sanderson, M.; Savage, N. H.; Shindell, D.; Strahan, S.; Szopa, S.; Sudo, K.; Van
24 Dingenen, R.; Wild, O.; Zeng, G. (2006a) The global atmospheric environment for the
25 next generation. Environ. Sci. Technol. 40: 3586-3594.
26 Dentener, F.; Drevet, J.; Lamarque, J. F.; Bey, L; Eickhout, B.; Fiore, A. M.; Hauglustaine, D.;
27 Horowitz, L. W.; Krol, M.; Kulshrestha, U. C.; Lawrence, M.; Galy-Lacaux, C.; Rast, S.;
28 Shindell, D.; Stevenson, D.; Van Noije, T.; Atherton, C.; Bell, N.; Bergman, D.; Butler,
29 T.; Cofala, J.; Collins, B.; Doherty, R.; Ellingsen, K.; Galloway, J.; Gauss, M.;
30 Montanaro, V.; Muller, J. F.; Pitari, G.; Rodriguez, J.; Sanderson, M.; Solman, F.;
31 Strahan, S.; Schultz, M.; Sudo, K.; Szopa, S.; Wild, O. (2006b) Nitrogen and sulfur
32 deposition on regional and global scales: a multi-model evaluation. Global Biogeochem.
33 Cycles: 20(GB4003): 10.1029/2005GB002672.
34 De Paula Santos, U.; Braga, A. L. F.; Giorgi, D. M. A.; Pereira, L. A. A.; Grupi, C. J.; Lin, C. A.;
35 Bussacos, M. A.; Zanetta, D. M.; Saldiva, P. H. do N.; Terra Filho, M. (2005) Effects of
36 air pollution on blood pressure and heart rate variability: a panel study of vehicular traffic
37 controllers in the city of Sao Paulo, Brazil. Eur. Heart J. 26: 193-200.
38 Desqueyroux, H.; Pujet, J.-C.; Prosper, M.; Squinazi, F.; Momas, I. (2002a) Short-term effects of
39 low-level air pollution on respiratory health of adults suffering from moderate to severe
40 asthma. Environ. Res. A 89: 29-37.
September 2007 6-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 Desqueyroux, H.; Pujet, J.-C.; Prosper, M.; Le Moullec, Y.; Momas, I. (2002b) Effects of air
2 pollution on adults with chronic obstructive pulmonary disease. Arch. Environ. Health
3 57: 554-560.
4 Dietert, R. R.; Etzel, R. A.; Chen, D.; Halonen, M.; Holladay, S. D.; Jarabek, A. M.; Landreth,
5 K.; Peden, D. B.; Pinkerton, K.; Smialowicz, R. 1; Zoetis, T. (2000) Workshop to
6 identify critical window of exposure for children's health: immune and respiratory
7 systems work group summary. Environ. Health Perspect. Suppl. 108(3): 483-490.
8 Dockery, D. W.; Ware, J. H.; Ferris, B. G., Jr.; Speizer, F. E.; Cook, N. R.; Herman, S. M.
9 (1982) Change in pulmonary function in children associated with air pollution episodes.
10 J. Air Pollut. Control Assoc. 32: 937-942.
11 Dockery, D. W.; Speizer, F. E.; Stram, D. O.; Ware, J. H.; Spengler, J. D.; Ferris, B. G., Jr.
12 (1989) Effects of inhalable particles on respiratory health of children. Am. Rev. Respir.
13 Dis. 139: 587-594.
14 Dockery, D. W.; Schwartz, J.; Spengler, J. D. (1992) Air pollution and daily mortality:
15 associations with particulates and acid aerosols. Environ. Res. 59: 362-373.
16 Dockery, D. W.; Pope, C. A., Ill; Xu, X.; Spengler, J. D.; Ware, J. H.; Fay, M. E.; Ferris, B. G,
17 Jr.; Speizer, F. E. (1993) An association between air pollution and mortality in six U.S.
18 cities. N. Engl. J. Med. 329: 1753-1759.
19 Dockery, D. W.; Cunningham, J.; Damokosh, A. I; Neas, L. M.; Spengler, J. D.; Koutrakis, P.;
20 Ware, J. H.; Raizenne, M.; Speizer, F. E. (1996) Health effects of acid aerosols on North
21 American children: respiratory symptoms. Environ. Health Perspect. 104: 500-505.
22 Dockery, D. W.; Luttmann-Gibson, H.; Rich, D. Q.; Link, M. S.; Mittleman, M. A.; Gold, D. R.;
23 Koutrakis, P.; Schwartz, J. D.; Verrier, R. L. (2005) Association of air pollution with
24 increased incidence of ventricular tachyarrhythmias recorded by implanted cardioverter
25 defibrillators. Environ. Health Perspect. 113: 670-674.
26 Dodge, R.; Solomon, P.; Moyers, J.; Hayes, C. (1985) A longitudinal study of children exposed
27 to sulfur oxides. Am. J. Epidemiol. 121: 720-736.
28 Dominici, F.; McDermott, A.; Daniels, M.; Zeger, S. L.; Samet, J. M. (2003) Mortality among
29 residents of 90 cities. In: Revised analyses of time-series studies of air pollution and
30 health. Special report. Boston, MA: Health Effects Institute; pp. 9-24. Available:
31 http://www.healtheffects.org/Pubs/TimeSeries.pdf [12 May, 2004].
32 Douglas, G. J.; Price, J. F.; Page, C. P. (1994) A method for the long-term exposure of rabbits to
33 environmental pollutant gases. Eur. Respir. J. 7: 1516-1526.
34 Du, Z.; Meng, Z. (2004a) Effects of derivatives of sulfur dioxide on transient outward potassium
35 currents in acutely isolated hippocampal neurons. Food Chem. Toxicol. 42: 1211-1216.
36 Du, Z.; Meng, Z. (2004b) Modulation of sodium currents in rat dorsal root ganglion neurons by
37 sulfur dioxide derivatives. Brain Res. 1010: 127-133.
38 Du, Z.; Meng, Z. (2006) Sulfur dioxide derivatives modulation of high-threshold calcium
39 currents in rat dorsal root ganglion neurons. Neurosci. Lett. 405: 147-152.
September 2007 6-8 DRAFT-DO NOT QUOTE OR CITE
-------
1 Dugandzic, R. Dodds, L.; Stieb, D.; Smith-Doiron, M. (2006) The association between low level
2 exposures to ambient air pollution and term low birth weight: a retrospective cohort
3 study. Environ. Health 5:3.
4 Elliott, P.; Shaddick, G.; Wakefield, J. C.; Hoogh, C. de; Briggs, D. J. (2007) Long-term
5 associations of outdoor air pollution with mortality in Great Britain. Thorax
6 10.1136/thx.2006.076851.
7 Euler, G. L.; Abbey, D. E.; Magie, A. R.; Hodgkin, J. E. (1987) Chronic obstructive pulmonary
8 disease symptom effects of long-term cumulative exposure to ambient levels of total
9 suspended particulates and sulfur dioxide in California Seventh-Day Adventist residents.
10 Arch. Environ. Health 42: 213-222.
11 Fairley, D. (1990) The relationship of daily mortality to suspended particulates in Santa Clara
12 county, 1980-86. Environ. Health Perspect. 89: 159-168.
13 Farhat, S. C. L.; Paulo, R. L. P.; Shimoda, T. M.; Conceicao, G. M. S.; Lin, C. A.; Braga, A. L.
14 F.; Warm, M. P. N.; Saldiva, P. H. N. (2005) Effect of air pollution on pediatric
15 respiratory emergency room visits and hospital admissions. Braz. J. Med. Biol. Res. 38:
16 227-235.
17 Federal Register. (1988) Proposed decision not to revise the national ambient air quality
18 standards for sulfur oxides (sulfur dioxide). F. R. (April 26) 53: 14,926-14,952.
19 Federal Register. (1994) National ambient air quality standards for sulfur oxides (sulfur dioxide)-
20 -reproposal. F. R. (November 15) 59: 58958-58980.
21 Federal Register. (1996) National ambient air quality standards for sulfur oxides (sulfur
22 dioxide)—final decision. F. R. (May 22) 61: 25566-25580.
23 Federal Register. (2006) Science assessment for sulfur oxides. F. R. 71 (May 15): 28023-28024.
24 Filleul, L.; Rondeau, V.; Vandentorren, S.; Le Moual, N.; Cantagrel, A.; Annesi-Maesano, L;
25 Charpin, D.; Declercq, C.; Neukirch, F.; Paris, C.; Vervloet, D.; Brochard, P.; Tessier, J.
26 F.; Kauffmann, F.; Baldi, I. (2005) Twenty five year mortality and air pollution: results
27 from the French PAARC survey. Occup. Environ. Med. 62: 453-460.
28 Fiore, M.; Petruzzi, S.; Dell'Omo, G.; Alleva, E. (1998) Prenatal sulfur dioxide exposure induces
29 changes in the behavior of adult male mice during agonistic encounters. Neurotoxicol.
30 Teratol. 20: 543-548.
31 Folinsbee, L. J.; Bedi, J. F.; Horvath, S. M. (1985) Pulmonary response to threshold levels of
32 sulfur dioxide (1.0 ppm) and ozone (0.3 ppm). J. Appl. Physiol. 58: 1783-1787.
33 Folsom, A. R.; Wu, K. K.; Rosamond, W. D.; Sharrett, A. R.; Chambless, L. E. (1997)
34 Prospective study of hemostatic factors and incidence of coronary heart disease: the
35 Atherosclerosis Risk in Communities (ARIC) study. Circulation 96: 1102-1108.
36 Frank, N. R.; Amdur, M. O.; Worcester, J.; Whittenberger, J. L. (1962) Effects of acute
37 controlled exposure to SC>2 on respiratory mechanics in healthy male adults. J. Appl.
38 Physiol. 17: 252-258.
September 2007 6-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 Frank, N. R.; Yoder, R. E.; Brain, J. D.; Yokoyama, E. (1969) SO2 (35S labeled) absorption by
2 the nose and mouth under conditions of varying concentration and flow. Arch. Environ.
3 Health 18:315-322.
4 Frischer, T.; Studnicka, M.; Gartner, C.; Tauber, E.; Horak, F.; Veiter, A.; Spengler, J.; Kiihr, J.;
5 Urbanek, R. (1999) Lung function growth and ambient ozone: a three-year population
6 study in school children. Am. J. Respir. Crit. Care Med. 160: 390-396.
7 Frye, C.; Hoelscher, B.; Cyrys, J.; Wjst, M.; Wichmann, H. E.; Heinrich, J. (2003) Association of
8 lung function with declining ambient air pollution. Environ. Health Perspect. Ill: 383-
9 387.
10 Fung, K. Y.; Luginaah, I; Gorey, K. M.; Webster, G. (2005) Air pollution and daily hospital
11 admissions for cardiovascular diseases in Windsor, Ontario. Can. J. Public Health 96: 29-
12 33.
13 Fung, K. Y.; Khan, S.; Krewski, D.; Chen, Y. (2006) Association between air pollution and
14 multiple respiratory hospitalizations among the elderly in Vancouver, Canada. Inhalation
15 Toxicol. 18: 1005-1011.
16 Fusco, D.; Forastiere, F.; Michelozzi, P.; Spadea, T.; Ostro, B.; Area, M.; Perucci, C. A. (2001)
17 Air pollution and hospital admissions for respiratory conditions in Rome, Italy. Eur.
18 Respir. J. 17: 1143-1150.
19 Galan, I.; Tobias, A.; Banegas, J. R.; Aranguez, E. (2003) Short-term effects of air pollution on
20 daily asthma emergency room admissions. Eur. Respir. J. 22: 802-808.
21 Gauderman, W. J.; Murcray, C.; Gilliland, F.; Conti, D. V. (2007) Testing association between
22 disease and multiple SNPs in a candidate gene. Genet. Epidemiol.: 10.1002/gepi.20219.
23 Gilboa, S. M.; Mendola, P.; Olshan, A. F.; Langlois, P. H.; Savitz, D. A.; Loomis, D.; Herring,
24 A. H.; Fixler, D. E. (2005) Relation between ambient air quality and selected birth
25 defects, seven county study, Texas, 1997-2000. Am. J. Epidemiol. 162: 238-252.
26 Gilliland, F. D.; McConnell, R.; Peters, J.; Gong, H, Jr. (1999) A theoretical basis for
27 investigation ambient air pollution and children's respiratory health. Environ. Health
28 Perspect. 107(suppl. 3): 403-407.
29 Gilliland, F. D.; Rappaport, E. B.; Berhane, K.; Islam, T.; Dubeau, L.; Gauderman, W. J.;
30 McConnell, R. (2002) Effects of glutathione S-Transferase PI, Ml, and Tl on acute
31 respiratory illness in school children. Am. J. Respir. Crit. Care Med. 166: 346-351.
32 Gold, D. R.; Litonjua, A.; Schwartz, J.; Lovett, E.; Larson, A.; Nearing, B.; Allen, G.; Verrier,
33 M.; Cherry, R.; Verrier, R. (2000) Ambient pollution and heart rate variability.
34 Circulation 101: 1267-1273.
35 Gold, D. R.; Schwartz, J.; Litonjua, A.; Verrier, R.; Zanobetti, A. (2003) Ambient pollution and
36 reduced heart rate variability. In: Revised analyses of time-series studies of air pollution
37 and health. Special report. Boston, MA: Health Effects Institute; pp. 107-112. Available:
38 http://www.healtheffects.org/Pubs/TimeSeries.pdf [18 October, 2004].
39 Goldberg, M. S.; Burnett, R. T. (2003) Revised analysis of the Montreal time-series study. In:
40 Revised analyses of time-series studies of air pollution and health. Special report. Boston,
September 2007 6-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 MA: Health Effects Institute; pp. 113-132. Available:
2 http://www.healtheffects.org/Pubs/TimeSeries.pdf [18 October, 2004].
3 Gong, H., Jr.; Lachenbruch, P. A.; Harber, P.; Linn, W. S. (1995) Comparative short-term health
4 responses to sulfur dioxide exposure and other common stresses in a panel of asthmatics.
5 Toxicol. Ind. Health 11: 467-487.
6 Goss, C. H.; Newsom, S. A.; Schildcrout, J. S.; Sheppard, L.; Kaufman, J. D. (2004) Effect of
7 ambient air pollution on pulmonary exacerbations and lung function in cystic fibrosis.
8 Am. J. Respir. Crit. Care Med. 169: 816-821.
9 Gouveia, N.; Fletcher, T. (2000) Respiratory diseases in children and outdoor air pollution in Sao
10 Paulo, Brazil: a time series analysis. Occup. Environ. Med. 57: 477-483.
11 Gouveia, N.; Bremner, S. A.; Novaes, H. M. D. (2004) Association between ambient air
12 pollution and birth weight in Sao Paulo, Brazil. J. Epidemiol. Community Health 58: 11-
13 17.
14 Gunnison, A. F.; Benton, A. W. (1971) Sulfur dioxide: sulfite. Interaction with mammalian
15 serum and plasma. Arch. Environ. Health 22: 381-388.
16 Gunnison, A. F.; Sellakumar, A.; Currie, D.; Snyder, E. A. (1987) Distribution, metabolism and
17 toxicity of inhaled sulfur dioxide and endogenously generated sulfite in the respiratory
18 tract of normal and sulfite oxidase-deficient rats. J. Toxicol. Environ. Health 21: 141-162.
19 Gunnison, A. F.; Sellakumar, A.; Snyder, E. A.; Currie, D. (1988) The effect of inhaled sulfur
20 dioxide and systemic sulfite on the induction of lung carcinoma in rats by
21 benzo[a]pyrene. Environ. Res. 46: 59-73.
22 Ha, E.-H.; Hong, Y.-C.; Lee, B.-E.; Woo, B.-H.; Schwartz, J.; Christian!, D. C. (2001) Is air
23 pollution a risk factor for low birth weight in Seoul? Epidemiology 12: 643-648.
24 Hackney, J. D.; Linn, W. S.; Bailey, R. M.; Spier, C. E.; Valencia, L. M. (1984) Time course of
25 exercise-induced bronchoconstriction in asthmatics exposed to sulfur dioxide. Environ.
26 Res. 34: 321-327.
27 Hackney, J. D.; Linn, W. S.; Avol, E. L. (1987) Replicated dose-response study of sulfur dioxide
28 effects in normal, atopic, and asthmatic volunteers: interim special report. Palo Alto, CA:
29 Electric Power Research Institute; research project 1225-2.
30 Hagen, J. A.; Nafstad, P.; Skrondal, A.; Bj0rkly, S.; Magnus, P. (2000) Associations between
31 outdoor air pollutants and hospitalization for respiratory diseases. Epidemiology 11: 136-
32 140.
33 Haider, S. S.; Hasan, M. (1984) Neurochemical changes by inhalation of environmental
34 pollutants sulfur dioxide and hydrogen sulfide: degradation of total lipids, elevation of
35 lipid peroxidation and enzyme activity in discrete regions of the guinea pig brain and
36 spinal cord. Ind. Health 22: 23-31.
37 Haider, S. S.; Hasan, M.; Hasan, S. N.; Khan, S. R.; Ali, S. F. (1981) Regional effects of sulfur
38 dioxide exposure on the guinea pig brain lipids, lipid peroxidation and lipase activity.
39 Neurotoxicology 2: 443-450.
September 2007 6-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 Haider, S. S.; Hasan, M.; Khan, N. H. (1982) Air pollutant sulfur dioxide-induced alterations on
2 the levels of lipids, lipid peroxidation and lipase activity in various regions of the rat
3 brain. ActaPharmacol. Toxicol. 51: 45-50.
4 Hajat, S.; Haines, A.; Goubet, S. A.; Atkinson, R. W.; Anderson, H. R. (1999) Association of air
5 pollution with daily GP consultations for asthma and other lower respiratory conditions in
6 London. Thorax 54: 597-605.
7 Hajat, S.; Haines, A.; Atkinson, R. W.; Bremner, S. A.; Anderson, H. R.; Emberlin, J. (2001)
8 Association between air pollution and daily consultations with general practitioners for
9 allergic rhinitis in London, United Kingdom. Am. J. Epidemiol. 153: 704-714.
10 Hajat, S.; Anderson, H. R.; Atkinson, R. W.; Haines, A. (2002) Effects of air pollution on
11 general practitioner consultations for upper respiratory diseases in London. Occup.
12 Environ. Med. 59: 294-299.
13 Halinen, A. I; Salonen, R. O.; Pennanen, A. S.; Kosma, V. M. (2000a) Combined respiratory
14 effects of cold air with SC>2 orNC>2 in repeated 10-minute exposures of hyperventilating
15 guinea pigs. Inhalation Toxicol. 12: 671-691.
16 Halinen, A. I.; Salonen, R. O.; Pennanen, A. S. (2000b) Combined respiratory effects of cold air
17 with SC>2 or NC>2 in single 1-hour exposures of hyperventilating guinea pigs. Inhalation
18 Toxicol. 12: 693-713.
19 Health Effects Institute. (2004) Health effects of outdoor air pollution in developing countries of
20 Asia: a literature review. Boston, MA: HEI International Scientific Oversight Committee;
21 HEI Public Health and Air Pollution in Asia Program; special report 15.
22 Hedley, A. J.; Wong, C.-M.; Thach, T. Q.; Ma, S.; Lam, T.-H.; Anderson, H. R. (2002)
23 Cardiorespiratory and all-cause mortality after restrictions on sulphur content of fuel in
24 Hong Kong: an intervention study. Lancet 360: 1646-1652.
25 Heinrich, U.; Mohr, U.; Fuhst, R.; Brockmeyer, C. (1989) Investigation of a potential
26 cotumorigenic effect of the dioxides of nitrogen and sulfur, and of diesel-engine exhaust,
27 on the respiratory tract of Syrian golden hamsters. Cambridge, MA: Health Effects
28 Institute; research report no. 26. Available from: NTIS, Springfield, VA; PB90-111147.
29 Heinrich, J.; Hoelscher, B.; Frye, C.; Meyer, I; Pitz, M.; Cyrys, J.; Wjst, M.; Neas, L.;
30 Wichmann, H.-E. (2002) Improved air quality in reunified Germany and decreases in
31 respiratory symptoms. Epidemiology 13: 394-401.
32 Henneberger, A.; Zareba, W.; Ibald-Mulli, A.; Riickerl, R.; Cyrys, J.; Couderc, J.-P.; Mykins, B.;
33 Woelke, G.; Wichmann, H.-E.; Peters, A. (2005) Repolarization changes induced by air
34 pollution in ischemic heart disease patients. Environ. Health Perspect. 113: 440-446.
35 Herbarth, O.; Fritz, G.; Krumbiegel, P.; Diez, U.; Franck, U.; Richter, M. (2001) Effect of sulfur
36 dioxide and particulate pollutants on bronchitis in children—a risk analysis. Environ.
37 Toxicol. 16: 269-276.
38 Higgins, B. G.; Francis, H. C.; Yates, C. J.; Warburton, C. J.; Fletcher, A. M.; Reid, J. A.;
39 Pickering, C. A. C.; Woodcock, A. A. (1995) Effects of air pollution on symptoms and
40 peak expiratory flow measurements in subjects with obstructive airways disease. Thorax
41 50: 149-155.
September 2007 6-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 Hirsch, T.; Weiland, S. K.; Von Mutius, E.; Safeca, A. F.; Grafe, H.; Csaplovics, E.; Duhme, H.;
2 Keil, U.; Leupold, W. (1999) Inner city air pollution and respiratory health and atopy in
3 children. Eur. Respir. J. 14: 669-677.
4 Hoek, G. (2003) Daily mortality and air pollution in The Netherlands. In: Revised analyses of
5 time-series studies of air pollution and health. Special report. Boston, MA: Health Effects
6 Institute; pp. 133-141. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdffl2
7 May, 2004].
8 Hoek, G.; Brunekreef, B. (1993) Acute effects of a winter air pollution episode on pulmonary
9 function and respiratory symptoms of children. Arch. Environ. Health 48: 328-335.
10 Hoek, G.; Brunekreef, B. (1994) Effects of low-level winter air pollution concentrations on
11 respiratory health of Dutch children. Environ. Res. 64: 136-150.
12 Hoek, G.; Brunekreef, B. (1995) Effect of photochemical air pollution on acute respiratory
13 symptoms in children. Am. J. Respir. Crit. Care Med. 151: 27-32.
14 Hoek, G.; Brunekreef, B.; Verhoeff, A.; Van Wijnen, J.; Fischer, P. (2000) Daily mortality and
15 air pollution in the Netherlands. J. Air Waste Manage. Assoc. 50: 1380-1389.
16 Hoek, G.; Brunekreef, B.; Fischer, P.; Van Wijnen, J. (2001) The association between air
17 pollution and heart failure, arrhythmia, embolism, thrombosis, and other cardiovascular
18 causes of death in a time series study. Epidemiology 12: 355-357.
19 Holguin, F.; Tellez-Rojo, M. M.; Hernandez, M.; Cortez, M.; Chow, J. C.; Watson, J. G.;
20 Mannino, D.; Romieu, I. (2003) Air pollution and heart rate variability among the elderly
21 in Mexico City. Epidemiology 14: 521-527.
22 Holland, H. D. (1978) The chemistry of the atmosphere and oceans. New York, NY: Wiley.
23 Hoppel, W. A.; Caffrey, P. F. (2005) Oxidation of S(IV) in sea-salt aerosol at high pH: ozone
24 versus aerobic reaction. J. Geophys. Res. [Atmos.] 110(D23202):
25 10.1029/2005JD006239.
26 Horak, F., Jr.; Studnicka, M.; Gartner, C.; Spengler, J. D.; Tauber, E.; Urbanek, R.; Veiter, A.;
27 Frischer, T. (2002a) Particulate matter and lung function growth in children: a 3-yr
28 follow-up study in Austrian schoolchildren. Eur. Respir. J. 19: 838-845.
29 Horak, F., Jr.; Studnicka, M.; Gartner, C.; Spengler, J. D.; Tauber, E.; Urbanek, R.; Veiter, A.;
30 Frischer, T. (2002b) Particulate matter and lung function growth in children: a 3-yr
31 follow-up study in Austrian schoolchildren [author response]. Eur. Respir. J. 20: 1355.
32 Horowitz, L. W.; Walters, S.; Mauzerall, D. L.; Emmons, L. K.; Rasch, P. J.; Granier, C.; Tie,
33 X.; Lamarque, J.-F.; Schultz, M. G; Tyndall, G. S.; Orlando, J. J.; Brasseur, G. P. (2003)
34 A global simulation of tropospheric ozone and related tracers: description and evaluation
35 of MOZART, version 2. J. Geophys. Res. [Atmos.] 108(D24): 10.1029/2002JD002853.
36 Horstman, D.; Roger, L. J.; Kehrl, H.; Hazucha, M. (1986) Airway sensitivity of asthmatics to
37 sulfur dioxide. Toxicol. Ind. Health 2: 289-298.
38 Horstman, D. H.; Seal, E., Jr.; Folinsbee, L. J.; Ives, P.; Roger, L. J. (1988) The relationship
39 between exposure duration and sulfur dioxide-induced bronchoconstriction in asthmatic
40 subjects. Am. Ind. Hyg. Assoc. J. 49: 38-47.
September 2007 6-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 Hosseinpoor, A. R.; Forouzanfar, M. H.; Yunesian, M.; Asghari, F.; Naieni, K. H.; Farhood, D.
2 (2005) Air pollution and hospitalization due to angina pectoris in Tehran, Iran: a time-
3 series study. Environ. Res. 99: 126-131.
4 Ibald-Mulli, A.; Stieber, I; Wichmann, H.-E.; Koenig, W.; Peters, A. (2001) Effects of air
5 pollution on blood pressure: a population-based approach. Am. J. Public Health 91: 571-
6 577.
7 Ito, T.; Ohyama, K.-L; Kusano, T.; Usuda, Y.; Nozawa, A.; Hayashi, H.; Ohji, H.; Kitamura, H.;
8 Kanisawa, M. (1997) Pulmonary endocrine cell hyperplasia and papilloma in rats induced
9 by intratracheal injections of extract from particulate air pollutants. Exp. Toxicol. Pathol.
10 49: 65-70.
11 Jacob, D. J. (1999) Introduction to atmospheric chemistry. Princeton, New Jersey: Princeton
12 University Press.
13 Jacobson, M. Z. (2002) Atmospheric pollution: history, science, and regulation. New York, NY:
14 Cambridge University Press.
15 Jaffe, D. H.; Singer, M. E.; Rimm, A. A. (2003) Air pollution and emergency department visits
16 for asthma among Ohio Medicaid recipients, 1991-1996. Environ. Res. 91: 21-28.
17 Jakab, G. J.; Clarke, R. W.; Hemenway, D. R.; Longphre, M. V.; Kleeberger, S. R.; Frank, R.
18 (1996) Inhalation of acid coated carbon black particles impairs alveolar macrophage
19 phagocytosis. Toxicol. Lett. 88: 243-248.
20 Jalaludin, B.; Morgan, G.; Lincoln, D.; Sheppeard, V.; Simpson, R.; Corbett, S. (2006)
21 Associations between ambient air pollution and daily emergency department attendances
22 for cardiovascular disease in the elderly (65+ years), Sydney, Australia. J. Exposure Sci.
23 Environ. Epidemiol. 16: 225-237.
24 Jerrett, M.; Burnett, R. T.; Willis, A.; Krewski, D.; Goldberg, M. S.; DeLuca, P.; Finkelstein, N.
25 (2003) Spatial analysis of the air pollution-mortality relationship in the context of
26 ecologic confounders. J. Toxicol. Environ. Health A 66: 1735-1777.
27 Jet Propulsion Laboratory (JPL). (2003) Chemical kinetics and photochemical data for use in
28 atmospheric studies. Pasadena, CA: California Institute of Technology; JPL publication
29 no. 02-25. Available: http://jpldataeval.jpl.nasa.gov/pdf/JPL_02-25_l_intro_rev0.pdf (18
30 December, 2003).
31 Jorres, R.; Magnussen, H. (1990) Airways response of asthmatics after a 30 min exposure, at
32 resting ventilation, to 0.25 ppm NC>2 or 0.5 ppm SC>2. Eur. Respir. J. 3: 132-137.
33 Katsouyanni, K.; Touloumi, G.; Spix, C.; Schwartz, J.; Balducci, F.; Medina, S.; Rossi, G.;
34 Wojtyniak, B.; Sunyer, J.; Bacharova, L.; Schouten, J. P.; Ponka, A.; Anderson, H. R.
35 (1997) Short term effects of ambient sulphur dioxide and particulate matter on mortality
36 in 12 European cities: results from time series data from the APHEA project. Br. Med. J.
37 314: 1658-1663.
38 Kauffmann, F.; Post Genome Respiratory Epidemiology Group. (2004) Post-genome respiratory
39 epidemiology: a multidisciplinary challenge. Eur. Respir. J. 24: 471-480.
40 Keles., N.; Ilicali, C.; Deger, K. (1999) The effects of different levels of air pollution on atopy
41 and symptoms of allergic rhinitis. Am. J. Rhinol. 13: 185-190.
September 2007 6-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 Kelly, F. J.; Dunster, C.; Mudway, I. (2003) Air pollution and the elderly: oxidant/antioxidant
2 issues worth consideration. Eur. Respir. J. Suppl. 40: 70S-75S.
3 Kelsall, J. E.; Samet, J. M.; Zeger, S. L.; Xu, J. (1997) Air pollution and mortality in
4 Philadelphia, 1974-1988. Am. J. Epidemiol. 146: 750-762.
5 Khoury, M. J.; Davis, R.; Gwinn, M.; Lindegren, M. L.; Yoon, P. (2005) Do we need genomic
6 research for the prevention of common diseases with environmental causes? Am. J.
7 Epidemiol. 161: 799-805.
8 Kilic, D. (2003) The effects of ageing and sulfur dioxide inhalation exposure on visual-evoked
9 potentials, antioxidant enzyme systems, and lipid-peroxidation levels of the brain and
10 eye. Neurotoxicol. Teratol. 25: 587-598.
11 Kimbell, J. S.; Miller, F. J. (1999) Regional respiratory-tract absorption of inhaled reactive
12 gases: a modeling approach. In: Gardner, D. E.; Crapo, J. D.; McClellan, R. O., eds.
13 Toxicology of the Lung. 3rd ed. New York, NY: Taylor and Francis; pp. 557-598.
14 Kindzierski, W. B.; Ranganathan, H. K. S. (2006) Indoor and outdoor SO2 in a community near
15 oil sand extraction and production facilities in northern Alberta. J. Environ. Eng. Sci. 5:
16 S121-S129.
17 Kitabatake, M.; Yoshida, K.; Kasama, K.; Murase, S.; Yuan, P. F.; Manjurul, H.; Yamauchi, T.
18 (1992) Procedure for evaluating changes in respiratory symptoms of experimentally
19 asthma-induced guinea pigs by a personal computer. J. Toxicol. Environ. Health 37: 265-
20 275.
21 Kitabatake, M.; Yamamoto, H.; Yuan, P. F.; Manjurul, H.; Murase, S.; Yamauchi, T. (1995)
22 Effects of exposure to NO2 or SO2 on bronchopulmonary reaction induced by Candida
23 albicans in guinea pigs. J. Toxicol. Environ. Health 45: 75-82.
24 Klepeis, N. E.; Nelson, W. C.; Ott. W. R.; Robinson, J. P. Tsang, A. M.; Switzer, P.; Behar, J.
25 V.; Hern, S. C.; Engelmann, W. H. (2001) The National Human Activity Pattern Survey
26 (NHAPS): a resource for assessing exposure to environmental pollutants. J. Exposure
27 Anal. Environ. Epidemiol. 11: 231-252.
28 Kochanek, K. D.; Murphy, S. L.; Anderson, R. N.; Scott, C. (2004) Deaths: final data for 2002.
29 Hyattsville, MD: U.S. Department of Health & Human Services, National Center for
30 Health Statistics; DHHS publication no. (PHS) 2005-1120. (National vital statistics
31 reports: v. 53, no. 5). Available:
32 http://www.cdc.gov/nchs/data/nvsr/nvsr53/nvsr53_05.pdf [3 August, 2005].
33 Koenig, J. Q.; Pierson, W. E.; Horike, M.; Frank, R. (1983) A comparison of the pulmonary
34 effects of 0.5 ppm versus 1.0 ppm sulfur dioxide plus sodium chloride droplets in
35 asthmatic adolescents. J. Toxicol. Environ. Health 11: 129-139.
36 Koenig, J. Q.; Covert, D. S.; Hanley, Q. S.; Van Belle, G.; Pierson, W. E. (1990) Prior exposure
37 to ozone potentiates subsequent response to sulfur dioxide in adolescent asthmatic
38 subjects. Am. Rev. Respir. Dis. 141: 377-380.
39 Koken, P. J. M.; Piver, W. T.; Ye, F.; Elixhauser, A.; Olsen, L. M.; Portier, C. J. (2003)
40 Temperature, air pollution, and hospitalization for cardiovascular diseases among elderly
41 people in Denver. Environ. Health Perspect. Ill: 1312-1317.
September 2007 6-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 Koutrakis, P.; Wolfson, J. M.; Slater, J. L.; Brauer, M.; Spengler, J. D.; Stevens, R. K.; Stone, C.
2 L. (1988) Evaluation of an annular denuder/filter pack system to collect acidic aerosols
3 and gases. Environ. Sci. Technol. 22: 1463-1468.
4 Koutrakis, P.; Suh, H. H.; Sarnat, J. A.; Brown, K. W.; Coull, B. A.; Schwartz, J. (2005)
5 Characterization of particulate and gas exposures of sensitive populations living in
6 Baltimore and Boston. Boston, MA: Health Effects Institute; research report no. 131.
7 Kreisman, H.; Mitchell, C. A.; Hosein, H. R.; Bouhuys, A. (1976) Effect of low concentrations
8 of sulfur dioxide on respiratory function in man. Lung 154: 25-34.
9 Krewski, D.; Burnett, R. T.; Goldberg, M. S.; Hoover, K.; Siemiatycki, J.; Jerrett, M.;
10 Abrahamowicz, M.; White, W. H. (2000) Reanalysis of the Harvard Six Cities study and
11 the American Cancer Society study of particulate air pollution and mortality: a special
12 report of the Institute's Particle Epidemiology Reanalysis Project. Cambridge, MA:
13 Health Effects Institute. Available: http://pubs.healtheffects.org/view.php?id=6 [6 March,
14 2007].
15 Kii9iikatay, V.; Agar, A.; Yargi9oglu, P.; Gumus.lu, S.; Aktekin, B. (2003) Changes in
16 somatosensory evoked potentials, lipid peroxidation, and antioxidant enzymes in
17 experimental diabetes: effect of sulfur dioxide. Arch. Environ. Health 58: 14-22.
18 Kii9iikatay, V.; Bor-Kii9iikatay, M.; Atsak, P.; Agar, A. (2007) Effect of ingested sulfite on
19 hippocampus antioxidant enzyme activities in sulfite oxidase competent and deficient
20 rats. Int. J. Neurosci. 117: 971-983.
21 Kulle, T. J.; Sauder, L. R.; Shanty, F.; Kerr, H. D.; Farrell, B. P.; Miller, W. R.; Milman, J. H.
22 (1984) Sulfur dioxide and ammonium sulfate effects on pulmonary function and
23 bronchial reactivity in human subjects. Am. Ind. Hyg. Assoc. J. 45: 156-161.
24 Kwon, H.-J.; Cho, S.-H.; Nyberg, F.; Pershagen, G. (2001) Effects of ambient air pollution on
25 daily mortality in a cohort of patients with congestive heart failure. Epidemiology 12:
26 413-419.
27 Langley-Evans, S. C.; Phillips, G. J.; Jackson, A. A. (1996) Sulphur dioxide: a potent glutathione
28 depleting agent. Comp. Biochem. Physiol. Part C: Pharmacol. Toxicol. Endocrinol. 114:
29 89-98.
30 Lawther, P. J. (1955) Effects of inhalation of sulfur dioxide on respiration and pulse-rate in
31 normal subjects. Lancet 2: 745-748.
32 Lawther, P. J.; Waller, R. E.; Henderson, M. (1970) Air pollution and exacerbations of
33 bronchitis. Thorax 25: 525-539.
34 Lawther, P. J.; MacFarlane, A. J.; Waller, R. E; Brooks, A. G. F. (1975) Pulmonary function and
35 sulphur dioxide, some preliminary findings. Environ. Res. 10: 355-367.
36 Lee, J.-T.; Kim, H.; Song, H.; Hong, Y.-C.; Cho, Y.-S.; Shin, S.-Y.; Hyun, Y.-J.; Kim, Y.-S.
37 (2002) Air pollution and asthma among children in Seoul, Korea. Epidemiology 13: 481-
38 484.
39 Lee, J.-T.; Kim, H.; Cho, Y.-S.; Hong, Y.-C.; Ha, E.-H.; Park, H. (2003a) Air pollution and
40 hospital admissions for ischemic heart diseases among individuals 64+ years of age
41 residing in Seoul, Korea. Arch. Environ. Health 58: 617-623.
September 2007 6-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 Lee, B. E.; Ha, E. H.; Park, H. S.; Kim, Y. 1; Hong, Y. C.; Kim, H.; Lee, J. T. (2003b) Exposure
2 to air pollution during different gestational phases contributes to risks of low birth
3 weight. Hum. Reprod. 18: 638-643.
4 Lee, Y.-L.; Lin, Y.-C.; Lee, Y.-C.; Wang, J.-Y.; Hsiue, T.-R.; Guo, Y. L. (2004) Glutathione S-
5 transferase PI gene polymorphism and air pollution as interactive risk factors for
6 childhood asthma. Clin. Exp. Allergy 34: 1707-1713.
7 Lee, S. L.; Wong, W. H. S.; Lau, Y. L. (2006) Association between air pollution and asthma
8 admission among children in Hong Kong. Clin. Exp. Allergy 36: 1138-1146.
9 Leem, J.-H.; Kaplan, B. M.; Shim, Y. K.; Pohl, H. R.; Gotway, C. A.; Bullard, S. M.; Rogers, J.
10 F.; Smith, M. M.; Tylenda, C. A. (2006) Exposures to air pollutants during pregnancy
11 and preterm delivery. Environ. Health Perspect. 114: 905-910.
12 Le Tertre, A.; Quenel, P.; Eilstein, D.; Medina, S.; Prouvost, H.; Pascal, L.; Boumghar, A.;
13 Saviuc, P.; Zeghnoun, A.; Filleul, L.; Declercq, C.; Cassadou, S.; Le Goaster, C. (2002)
14 Short-term effects of air pollution on mortality in nine French cities: a quantitative
15 summary. Arch. Environ. Health 57: 311-319.
16 Leung, F.-Y. T.; Colussi, A. J.; Hoffmann, M. R. (2002) Isotopic fractionation of carbonyl
17 sulfide in the atmosphere: implications for the source of background stratospheric sulfate
18 aerosol. Geophys. Res. Lett. 29: 10.1029/2001GL013955.
19 Levine, J. S.; Pinto, J. P. (1998) The production of CO by biomass burning. In: Khalil, M. A. K.;
20 Pinto, J. P.; Shearer, M. J., eds. Atmospheric carbon monoxide and its environmental
21 effects: proceedings of the international conference; December 1997; Portland, Oregon.
22 Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research
23 and Development; report no. EPA/600/R-98/047; pp. 251-256.
24 Lewis, A. J.; Kirchner, T. (1984) Modulation of sulfur dioxide-induced airways
25 hyperresponsiveness in the conscious dog. Int. Arch. Allergy Appl. Immunol. 75: 188-
26 190.
27 Liao, D.; Duan, Y.; Whitsel, E. A.; Zheng, Z.-J.; Heiss, G; Chinchilli, V. M.; Lin, H.-M. (2004)
28 Association of higher levels of ambient criteria pollutants with impaired cardiac
29 autonomic control: a population-based study. Am. J. Epidemiol. 159: 768-777.
30 Liao, D.; Heiss, G; Chinchilli, V. M.; Duan, Y.; Folsom, A. R.; Lin, H. M.; Salomaa, V. (2005)
31 Association of criteria pollutants with plasma hemostatic/inflammatory markers: a
32 population-based study. J. Exposure Anal. Environ. Epidemiol. 15: 319-328.
33 Lin, C. A.; Martins, M. A.; Farhat, S. C. L.; Pope, C. A., Ill; Concei9ao, G. M. S.; Anastacio, V.
34 M.; Hatanaka, M.; Andrade, W. C.; Hamaue, W. R.; Bohm, G. M.; Saldiva, P. H. N.
35 (1999) Air pollution and respiratory illness of children in Sao Paulo, Brazil. Paediatr.
36 Perinat. Epidemiol. 13: 475-488.
37 Lin, M.; Chen, Y.; Burnett, R. T.; Villeneuve, P. J.; Krewski, D. (2003a) Effect of short-term
38 exposure to gaseous pollution on asthma hospitalisation in children: a bi-directional case-
39 crossover analysis. J. Epidemiol. Community Health 57: 50-55.
September 2007 6-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 Lin, C. A.; Amador Pereira, L. A.; Souza Conceicao, G. M. de; Kishi, H. S.; Milani, R., Jr.;
2 Ferreira Braga, A. L.; Nascimento Sal diva, P. H. (2003b) Association between air
3 pollution and ischemic cardiovascular emergency room visits. Environ. Res. 92: 57-63.
4 Lin, M.; Chen, Y.; Villeneuve, P. J.; Burnett, R. T.; Lemyre, L.; Hertzman, C.; McGrail, K. M.;
5 Krewski, D. (2004a) Gaseous air pollutants and asthma hospitalization of children with
6 low household income in Vancouver, British Columbia, Canada. Am. J. Epidemiol. 159:
7 294-303.
8 Lin, C.-M.; Li, C.-Y.; Yang, G.-Y.; Mao, I-F. (2004b) Association between maternal exposure to
9 elevated ambient sulfur dioxide during pregnancy and term low birth weight. Environ.
10 Res. 96: 41-50.
11 Lin, S.; Hwang, S.-A.; Pantea, C.; Kielb, C.; Fitzgerald, E. (2004c) Childhood asthma
12 hospitalizations and ambient air sulfur dioxide concentrations in Bronx County, New
13 York. Arch. Environ. Health 59: 266-275.
14 Lin, M.; Stieb, D. M.; Chen, Y. (2005) Coarse particulate matter and hospitalization for
15 respiratory infections in children younger than 15 years in Toronto: a case-crossover
16 analysis. Pediatrics 116: 235-240.
17 Linn, W. S.; Bailey, R. M.; Shamoo, D. A.; Venet, T. G.; Wightman, L. H.; Hackney, J. D.
18 (1982) Respiratory responses of young adult asthmatics to sulfur dioxide exposure under
19 simulated ambient conditions. Environ. Res. 29: 220-232.
20 Linn, W. S.; Venet, T. G.; Shamoo, D. A.; Valencia, L. M.; Anzar, U. T.; Spier, C. E.; Hackney,
21 J. D. (1983) Respiratory effects of sulfur dioxide in heavily exercising asthmatics: a dose-
22 response study. Am. Rev. Respir. Dis. 127: 278-283.
23 Linn, W. S.; Shamoo, D. A.; Venet, T. G.; Bailey, R. M.; Wightman, L. H.; Hackney, J. D.
24 (1984) Comparative effects of sulfur dioxide exposures at 5° C and 22° C in exercising
25 asthmatics. Am. Rev. Respir. Dis. 129: 234-239.
26 Linn, W. S.; Fischer, D. A.; Shamoo, D. A.; Spier, C. E.; Valencia, L. M.; Anzar, U. T.;
27 Hackney, J. D. (1985) Controlled exposures of volunteers with chronic obstructive
28 pulmonary disease to sulfur dioxide. Environ. Res. 37: 445-451.
29 Linn, W. S.; Avol, E. L.; Peng, R.-C.; Shamoo, D. A.; Hackney, J. D. (1987) Replicated dose-
30 response study of sulfur dioxide effects in normal, atopic, and asthmatic volunteers. Am.
31 Rev. Respir. Dis. 136: 1127-1134.
32 Linn, W. S.; Shamoo, D. A.; Peng, R.-C.; Clark, K. W.; Avol. E. L.; Hackney, J. D. (1990)
33 Responses to sulfur dioxide and exercise by medication-dependent asthmatics: effect of
34 varying medication levels. Arch. Environ. Health 45: 24-30.
35 Lipfert, F. W.; Perry, H. M., Jr.; Miller, J. P.; Baty, J. D.; Wyzga, R. E.; Carmody, S. E. (2000)
36 The Washington University-EPRI veterans' cohort mortality study: preliminary results.
37 In: Grant, L. D., ed. PM2000: particulate matter and health. Inhalation Toxicol. 12(suppl.
38 4): 41-73.
39 Lipfert, F. W.; Wyzga, R. E.; Baty, J. D.; Miller, J. P. (2006a) Traffic density as a surrogate
40 measure of environmental exposures in studies of air pollution health effects: long-term
41 mortality in a cohort of US veterans. Atmos. Environ. 40: 154-169.
September 2007 6-18 DRAFT-DO NOT QUOTE OR CITE
-------
1 Lipfert, F. W.; Baty, J. D.; Miller, J. P.; Wyzga, R. E. (2006b) PM2.5 constituents and related air
2 quality variables as predictors of survival in a cohort of U.S. military veterans. Inhalation
3 Toxicol. 18: 645-657.
4 Lippmann, M.; Ito, K.; Hwang, J. S.; Maciejczyk, P.; Chen, L. C. (2006) Cardiovascular effects
5 of nickel in ambient air. Environ. Health Perspect. 114: 1662-1669.
6 Liu, S.; Krewski, D.; Shi, Y.; Chen, Y.; Burnett, R. T. (2003) Association between gaseous
7 ambient air pollutants and adverse pregnancy outcomes in Vancouver, Canada. Environ.
8 Health Perspect. Ill: 1773-1778.
9 Liu, S.; Krewski, D.; Shi, Y.; Chen, Y.; Burnett, R. (2006) Association between maternal
10 exposure to ambient air pollutants during pregnancy and fetal growth restriction. J.
11 Exposure Sci. Environ. Epidemiol.: 10.1038/sj.jes.7500503.
12 Llorca, J.; Salas, A.; Prieto-Salceda, D.; Chinchon-Bengoechea, V.; Delgado-Rodriguez, M.
13 (2005) Nitrogen dioxide increases cardiorespiratory admissions in Torrelavega (Spain). J.
14 Environ. Health 68: 30-35.
15 Lovati, M. R.; Manzoni, C.; Daldossi, M.; Spolti, S.; Sirtori, C. R. (1996) Effects of sub-chronic
16 exposure to SC>2 on lipid and carbohydrate metabolism in rats. Arch. Toxicol. 70: 164-
17 173.
18 Luginaah, I. N.; Fung, K. Y.; Gorey, K. M.; Webster, G.; Wills, C. (2005) Association of
19 ambient air pollution with respiratory hospitalization in a government designated "area of
20 concern": the case of Windsor, Ontario. Environ. Health Perspect. 113: 290-296.
21 Luke, W. T. (1997) Evaluation of a commercial pulsed fluorescence detector for the
22 measurement of low-level SO2 concentrations during the Gas-Phase Sulfur
23 Intercomparison Experiment. J. Geophys. Res. [Atmos.] 102: 16,255-16,265.
24 Lunn, J. E.; Knowelden, J.; Handy side, A. J. (1967) Patterns of respiratory illness in Sheffield
25 infant schoolchildren. Br. J. Prev. Soc. Med. 21: 7-16.
26 Luttmann-Gibson, H.; Suh, H. H.; Coull, B. A.; Dockery, D. W.; Sarnet, S. E.; Schwartz, J.;
27 Stone, P. H.; Gold, D. R. (2006) Short-term effects of air pollution on heart rate
28 variability in senior adults in Steubenville, Ohio. J. Occup. Environ. Med. 48: 780-788.
29 Magnussen, H.; Torres, R.; Wagner, H. M.; Von Nieding, G. (1990) Relationship between the
30 airway response to inhaled sulfur dioxide, isocapnic hyperventilation, and histamine in
31 asthmatic subjects. Int. Arch. Occup. Environ. Health 62: 485-491.
32 Maisonet, M.; Bush, T. J.; Correa, A.; Jaakkola, J. J. K. (2001) Relation between ambient air
33 pollution and low birth weight in the northeastern United States. Environ. Health
34 Perspect. Suppl. 109(3): 351-356.
35 Martins, L. C.; Latorre, M. R. D. O.; Saldiva, P. H. N.; Braga, A. L. F. (2002) Air pollution and
36 emergency room visits due to chronic lower respiratory diseases in the elderly: an
37 ecological time-series study in Sao Paulo, Brazil. J. Occup. Environ. Med. 44: 622-627.
38 McCaig, L. F.; Burt, C. W. (2005) National Hospital Ambulatory Medical Care Survey: 2003
39 Emergency Department Summary. Hyattsville, MD: National Center for Health
40 Statistics; DHHS publication no. (PHS) 2005-1250. (Advance data from vital and health
September 2007 6-19 DRAFT-DO NOT QUOTE OR CITE
-------
1 statistics; no. 358). Available: http://www.cdc.gov/nchs/data/ad/ad358.pdff3 August,
2 2005].
3 McCurdy, T.; Glen, G.; Smith, L.; Lakkadi, Y. (2000) The National Exposure Research
4 Laboratory's Consolidated Human Activity Database. J. Exposure Anal. Environ.
5 Epidemiol. 10: 566-578.
6 Melville, G. N. (1970) Changes in specific airway conductance in healthy volunteers following
7 nasal and oral inhalation of SO2. West Indian Med. J. 19: 231-235.
8 Meng, Z.; Bai, W. (2004) Oxidation damage of sulfur dioxide on testicles of mice. Environ. Res.
9 96:298-304.
10 Meng, Z.; Qin, G.; Zhang, B.; Geng, H.; Bai, Q.; Bai, W.; Liu, C. (2003a) Oxidative damage of
11 sulfur dioxide inhalation on lungs and hearts of mice. Environ. Res. 93: 285-292.
12 Meng, Z.; Geng, H.; Bai, J.; Yan, G. (2003b) Blood pressure of rats lowered by sulfur dioxide
13 and its derivatives. Inhalation Toxicol. 15: 951-959.
14 Meng, Z.; Zhang, B.; Bai, J.; Geng, H.; Liu, C. (2003c) Oxidative damage of sulfur dioxide
15 inhalation on stomachs and intestines of mice. Inhalation Toxicol. 15: 397-410.
16 Meng, Z.; Liu, Y.; Wu, D. (2005) Effect of sulfur dioxide inhalation on cytokine levels in lungs
17 and serum of mice. Inhalation Toxicol. 17: 303-307.
18 Metzger, K. B.; Tolbert, P. E.; Klein, M.; Peel, J. L.; Flanders, W. D.; Todd, K. H.; Mulholland,
19 J. A.; Ryan, P. B.; Frumkin , H. (2004) Ambient air pollution and cardiovascular
20 emergency department visits. Epidemiology 15: 46-56.
21 Michaud, J.-P.; Grove, J. S.; Krupitsky, D. (2004) Emergency department visits and "vog"-
22 related air quality in Hilo, Hawai'i. Environ. Res. 95: 11-19.
23 Miller, K. A.; Siscovick, D. S.; Sheppard, L.; Shepherd, K.; Sullivan J. H.; Anderson, G. L.;
24 Kaufman, J. D. (2007) Long-term exposure to air pollution and incidence of
25 cardiovascular events in women. N. Engl. J. Med. 356: 447-458.
26 Mohorovic, L. (2004) First two months of pregnancy—critical time for preterm delivery and low
27 birthweight caused by adverse effects of coal combustion toxics. Early Human Dev. 80:
28 115-123.
29 Moolgavkar, S. H. (2000) Air pollution and hospital admissions for diseases of the circulatory
30 system in three U.S. metropolitan areas. J. Air Waste Manage. Assoc. 50: 1199-1206.
31 Moolgavkar, S. H. (2003) Air pollution and daily deaths and hospital admissions in Los Angeles
32 and Cook counties. In: Revised analyses of time-series studies of air pollution and health.
33 Special report. Boston, MA: Health Effects Institute; pp. 183-198. Available:
34 http://www.healtheffects.org/news.htm [16 May, 2003].
35 Moolgavkar, S. H.; Luebeck, E. G.; Hall, T. A.; Anderson, E. L. (1995) Air pollution and daily
36 mortality in Philadelphia. Epidemiology 6: 476-484.
37 Moolgavkar, S. H.; Luebeck, E. G.; Anderson, E. L. (1997) Air pollution and hospital
38 admissions for respiratory causes in Minneapolis-St. Paul and Birmingham.
39 Epidemiology 8: 364-370.
September 2007 6-20 DRAFT-DO NOT QUOTE OR CITE
-------
1 Morris, R. D.; Naumova, E. N.; Munasinghe, R. L. (1995) Ambient air pollution and
2 hospitalization for congestive heart failure among elderly people in seven large US cities.
3 Am. J. Public Health 85: 1361-1365.
4 Mortimer, K. M.; Neas, L. M.; Dockery, D. W.; Redline, S.; Tager, I. B. (2002) The effect of air
5 pollution on inner-city children with asthma. Eur. Respir. J. 19: 699-705.
6 Myhre, G.; Berglen, T. F.; Myhre, C. E.; Isaksen, I. S. A. (2004) The radiative effect of the
7 anthropogenic influence on the stratospheric sulfate aerosol layer. Tellus Ser. B 56: 294-
8 299.
9 Nadziejko, C.; Fang, K.; Narciso, S.; Zhong, M.; Su, W. C.; Gordon, T.; Nadas, A.; Chen, L. C.
10 (2004) Effect of particulate and gaseous pollutants on spontaneous arrhythmias in aged
11 rats. Inhalation Toxicol. 16: 373-380.
12 Nafstad, P.; Haheim, L. L.; Oftedal, B.; Gram, F.; Holme, L; Hjermann, L; Leren, P. (2003)
13 Lung cancer and air pollution: a 27 year follow up of 16,209 Norwegian men. Thorax 58:
14 1071-1076.
15 Nafstad, P.; Haheim, L. L.; Wisloff, T.; Gram, F.; Oftedal, B.; Holme, L; Hjermann, L; Leren, P.
16 (2004) Urban air pollution and mortality in a cohort of Norwegian men. Environ. Health
17 Perspect. 112:610-605.
18 National Center for Health Statistics. (2006a) Summary health statistics for U.S. adults: National
19 Health Interview Survey, 2004. Hyattsville, MD: U.S. Department of Health & Human
20 Services, Centers for Disease Control and Prevention. (Vital and Health Statistics, Series
21 10, no. 228). Available: http://www.cdc.gov/nchs/data/series/sr_10/srlO_228.pdf (14
22 September, 2007).
23 National Center for Health Statistics. (2006b) Summary health statistics for U.S. children:
24 National Health Interview Survey, —2005. Hyattsville, MD: U.S. Department of Health
25 & Human Services, Centers for Disease Control and Prevention. (Vital and Health
26 Statistics, Series 10, no. 231). Available:
27 http://www.cdc.gov/nchs/data/series/sr_10/srlO_231 .pdf (14 September, 2007).
28 National Research Council. (1998) Research priorities for airborne parti culate matter. I.
29 Immediate priorities and a long-range research portfolio. Washington, DC: National
30 Academy Press. Available: http://www.nap.edu/catalog/6131.html (4 June 2003).
31 National Research Council. (2004) Research priorities for airborne parti culate matter. IV.
32 Continuing research progress. Washington, DC: National Academies Press. Available:
33 http://www.nap.edu/catalog.php?record_id=10957 [1 August, 2007].
34 Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Tollerud, D. J.; Speizer, F. E. (1995) The
35 association of ambient air pollution with twice daily peak expiratory flow rate
36 measurements in children. Am. J. Epidemiol. 141: 111-122.
37 Neukirch, F.; Segala, C.; Le Moullec, Y.; Korobaeff, M.; Aubier, M. (1998) Short-term effects of
38 low-level winter pollution on respiratory health of asthmatic adults. Arch. Environ.
39 Health 53: 320-328.
40 Nicks, D. K.; Benner, R. L. (2001) Subminute measurements of 862 and low parts per trillion by
41 volume mixing ratios in the atmosphere. J. Geophys. Res. [Atmos.] 106: 27,769-27,776.
September 2007 6-21 DRAFT-DO NOT QUOTE OR CITE
-------
1 Me, A.; Meng, Z. (2005) Study of the interaction of sulfur dioxide derivative with cardiac
2 sodium channel. Biochim. Biophys. Acta 1718: 67-73.
3 Me, A.; Meng, Z. (2006) Modulation of L-type calcium current in rat cardiac myocytes by sulfur
4 dioxide derivatives. Food Chem. Toxicol. 44: 355-363.
5 Nodelman, V.; Ultman, J. S. (1999) Longitudinal distribution of chlorine absorption in human
6 airways: a comparison to ozone absorption. J. Appl. Physiol. 87: 2073-2080.
7 Nowak, D.; Torres, R.; Berger, J.; Claussen, M.; Magnussen, H. (1997) Airway responsiveness to
8 sulfur dioxide in an adult population sample. Am. J. Respir. Crit. Care Med. 156: 1151-
9 1156.
10 Nyberg, F.; Gustavsson, P.; Jarup, L.; Bellander, T.; Berglind, N.; Jakobsson, R.; Pershagen, G.
11 (2000) Urban air pollution and lung cancer in Stockholm. Epidemiology 11: 487-495.
12 Oftedal, B.; Nafstad, P.; Magnus, P.; Bj0rkly, S.; Skrondal, A. (2003) Traffic related air pollution
13 and acute hospital admission for respiratory diseases in Drammen, Norway 1995-2000.
14 Eur. J. Epidemiol. 18: 671-675.
15 Ohyama, K.; Ito, T.; Kanisawa, M. (1999) The roles of diesel exhaust particle extracts and the
16 promotive effects of NO2 and/or SO2 exposure on rat lung tumorigenesis. Cancer Lett.
17 139: 189-197.
18 Ott, W.; Wallace, L.; Mage, D. (2000) Predicting particulate (PMi0) personal exposure
19 distributions using a random component superposition statistical model. J. Air Waste
20 Manage. Assoc. 50: 1390-1406.
21 Park, J.-K.; Kim, Y.-K.; Lee, S.-R.; Cho, S.-H.; Min, K.-U.; Kim, Y.-Y. (2001) Repeated
22 exposure to low levels of sulfur dioxide (SO2) enhances the development of ovalbumin-
23 induced asthmatic reactions in guinea pigs. Ann. Allergy Asthma Immunol. 86: 62-67.
24 Park, H.; Lee, B.; Ha, E.-H.; Lee, J.-T.; Kim, H.; Hong, Y.-C. (2002) Association of air pollution
25 with school absenteeism due to illness. Arch. Pediatr. Adolesc. Med. 156: 1235-1239.
26 Park, J. W.; Lim, Y. H.; Kyung, S. Y.; An, C. H.; Lee, S. P.; Jeong, S. H.; Ju, S.-Y. (2005)
27 Effects of ambient particulate matter on peak expiratory flow rates and respiratory
28 symptoms of asthmatics during Asian dust periods in Korea. Respirology 10: 470-476.
29 Peacock, J. L.; Symonds, P.; Jackson, P.; Bremner, S. A.; Scarlett, J. F.; Strachan, D. P.;
30 Anderson, H. R. (2003) Acute effects of winter air pollution on respiratory function in
31 schoolchildren in southern England. Occup. Environ. Med. 60: 82-89.
32 Peel, J. L.; Tolbert, P. E.; Klein, M.; Metzger, K. B.; Flanders, W. D.; Knox, T.; Mulholland, J.
33 A.; Ryan, P. B.; Frumkin, H. (2005) Ambient air pollution and respiratory emergency
34 department visits. Epidemiology 16: 164-174.
35 Peel, J. L.; Metzger, K. B.; Klein, M.; Flanders, W. D.; Mulholland, J. A.; Tolbert, P. E. (2007)
36 Ambient air pollution and cardiovascular emergency department visits in potentially
37 sensitive groups. Am. J. Epidemiol. 165: 625-633.
38 Pekkanen, J.; Brunner, E. J.; Anderson, H. R.; Tiittanen, P.; Atkinson, R. W. (2000) Daily
39 concentrations of air pollution and plasma fibrinogen in London. Occup. Environ. Med.
40 57: 818-822.
September 2007 6-22 DRAFT-DO NOT QUOTE OR CITE
-------
1 Penard-Morand, C.; Charpin, D.; Raherison, C.; Kopferschmitt, C.; Caillaud, D.; Lavaud, F.;
2 Annesi-Maesano, I. (2005) Long-term exposure to background air pollution related to
3 respiratory and allergic health in schoolchildren. Clin. Exp. Allergy 35: 1279-1287.
4 Pereira, L. A. A.; Loomis, D.; Conceicao, G. M. S.; Braga, A. L. F.; Areas, R. M.; Kishi, H. S.;
5 Singer, J. M.; Bohm, G. M.; Saldiva, P. H. N. (1998) Association between air pollution
6 and intrauterine mortality in Sao Paulo, Brazil. Environ. Health Perspect. 106: 325-329.
7 Peters, A.; Goldstein, I. F.; Beyer, U.; Franke, K.; Heinrich, J.; Dockery, D. W.; Spengler, J. D.;
8 Wichmann, H.-E. (1996) Acute health effects of exposure to high levels of air pollution
9 in eastern Europe. Am. J. Epidemiol. 144: 570-581.
10 Peters, A.; Doling, A.; Wichmann, H.-E.; Koenig, W. (1997) Increased plasma viscosity during
11 an air pollution episode: a link to mortality? Lancet 349: 1582-1587.
12 Peters, A.; Perz, S.; Doring, A.; Stieber, J.; Koenig, W.; Wichmann, H.-E. (1999) Increases in
13 heart rate during an air pollution episode. Am. J. Epidemiol. 150: 1094-1098.
14 Peters, A.; Liu, E.; Verrier, R. L.; Schwartz, J.; Gold, D. R.; Mittleman, M.; Baliff, J.; Oh, J. A.;
15 Allen, G.; Monahan, K.; Dockery, D. W. (2000) Air pollution and incidence of cardiac
16 arrhythmia. Epidemiology 11: 11-17.
17 Peters, A.; Dockery, D. W.; Muller, J. E.; Mittleman, M. A. (2001) Increased particulate air
18 pollution and the triggering of myocardial infarction. Circulation 103: 2810-2815.
19 Petroeschevsky, A.; Simpson, R. W.; Thalib, L.; Rutherford, S. (2001) Associations between
20 outdoor air pollution and hospital admissions in Brisbane, Australia. Arch. Environ.
21 Health 56: 37-52.
22 Petruzzi, S.; Dell'Omo, G.; Fiore, M.; Chiarotti, F.; Bignami, G.; Alleva, E. (1996) Behavioural
23 disturbances in adult CD-I mice and absence of effects on their offspring upon SC>2
24 exposure. Arch. Toxicol. 70: 757-766.
25 Pikhart, H.; Bobak, M.; Gorynski, P.; Wojtyniak, B.; Danova, J.; Celko, M. A.; Kriz, B.; Briggs,
26 D.; Elliot, P. (2001) Outdoor sulphur dioxide and respiratory symptoms in Czech and
27 Polish school children: a small-area study (SAVIAH). Int. Arch. Occup. Environ. Health
28 74: 574-578.
29 Pino, P.; Walter, T.; Oyarzun, M.; Villegas, R.; Romieu, I. (2004) Fine particulate matter and
30 wheezing illnesses in the first year of life. Epidemiology 15: 702-708.
31 Piper, J. M.; Mitchel, E. F., Jr.; Snowden, M.; Hall, C.; Adams, M.; Taylor, P. (1993) Validation
32 of 1989 Tennessee birth certificates using maternal and newborn hospital records. Am. J.
33 Epidemiol. 137: 758-768.
34 Plunkett, L. M.; Turnbull, D.; Rodricks, J. V. (1992) Differences between adults and children
35 affecting exposure assessment. In: Guzelian, P. S.; Henry, D. J.; Olin, S. S., eds.
36 Similarities and differences between children and adults: implications for risk
37 assessment. Washington, DC: ILSI Press, pp. 79-96.
38 Poloniecki, J. D.; Atkinson, R. W.; Ponce de Leon, A.; Anderson, H. R. (1997) Daily time series
39 for cardiovascular hospital admissions and previous day's air pollution in London, UK.
40 Occup. Environ. Med. 54: 535-540.
September 2007 6-23 DRAFT-DO NOT QUOTE OR CITE
-------
1 Ponce de Leon, A.; Anderson, H. R.; Bland, J. M.; Strachan, D. P.; Bower, J. (1996) Effects of
2 air pollution on daily hospital admissions for respiratory disease in London between
3 1987-88 and 1991-92. In: St Leger, S., ed. The APHEA project. Short term effects of air
4 pollution on health: a European approach using epidemiological time series data. J.
5 Epidemiol. Community Health 50(suppl. 1): S63-S70.
6 Ponka, A. (1990) Absenteeism and respiratory disease among children and adults in Helsinki in
7 relation to low-level air pollution and temperature. Environ. Res. 52: 34-46.
8 Pope, C. A., III. (1989) Respiratory disease associated with community air pollution and a steel
9 mill, Utah Valley. Am. J. Public Health 79: 623-628.
10 Pope, C. A., Ill; Schwartz, J.; Ransom, M. R. (1992) Daily mortality and PMi0 pollution in Utah
11 valley. Arch. Environ. Health 47: Ill-Ill.
12 Pope, C. A., Ill; Thun, M. J.; Namboodiri, M. M.; Dockery, D. W.; Evans, J. S.; Speizer, F. E.;
13 Heath, C. W., Jr. (1995) Particulate air pollution as a predictor of mortality in a
14 prospective study of U.S. adults. Am. J. Respir. Crit. Care Med. 151: 669-674.
15 Pope, C. A., Ill; Burnett, R. T.; Thun, M. J.; Calle, E. E.; Krewski, D.; Ito, K.; Thurston, G. D.
16 (2002) Lung cancer, cardiopulmonary mortality, and long-term exposure to fine
17 particulate air pollution. JAMA J. Am. Med. Assoc. 287: 1132-1141.
18 Pueschel, R. F.; Van Valin, C. C. (1978) Cloud nucleus formation in a power plant plume.
19 Atmos. Environ. 12: 307-312.
20 Qin, G.; Meng, Z. (2005) Effect of sulfur dioxide inhalation on CYP1 Al and CYP1A2 in rat
21 liver and lung. Toxicol. Lett. 160: 34-42.
22 Ramadour, M.; Burel, C.; Lanteaume, A.; Vervloet, D.; Charpin, D.; Brisse, F.; Dutau, H.;
23 Charpin, D. (2000) Prevalence of asthma and rhinitis in relation to long-term exposure to
24 gaseous air pollutants. Allergy (Copenhagen) 55: 1163-1169.
25 Rich, K. E.; Petkau, J.; Vedal, S.; Brauer, M. (2004) A case-crossover analysis of particulate air
26 pollution and cardiac arrhythmia in patients with implantable cardioverter defibrillators.
27 Inhalation Toxicol. 16: 363-372.
28 Rich, D. Q.; Schwartz, J.; Mittleman, M. A.; Link, M.; Luttmann-Gibson, H.; Catalano, P. J.;
29 Speizer, F. E.; Dockery, D. W. (2005) Association of short-term ambient air pollution
30 concentrations and ventricular arrhythmias. Am. J. Epidemiol. 161: 1123-1132.
31 Rich, D. Q.; Kim, M. H.; Turner, J. R.; Mittleman, M. A.; Schwartz, J.; Catalano, P. J.; Dockery,
32 D. W. (2006) Association of ventricular arrhythmias detected by implantable cardioverter
33 defibrillator and ambient air pollutants in the St Louis, Missouri metropolitan area.
34 Occup. Environ. Med. 63: 591-596.
35 Riedel, F.; Kramer, M.; Scheibenbogen, C.; Rieger, C. H. L. (1988) Effects of SC>2 exposure on
36 allergic sensitization in the guinea pig. J. Allergy Clin. Immunol. 82: 527-534.
37 Roemer, W.; Hoek, G.; Brunekreef, B. (1993) Effect of ambient winter air pollution on
38 respiratory health of children with chronic respiratory symptoms. Am. Rev. Respir. Dis.
39 147:118-124.
September 2007 6-24 DRAFT-DO NOT QUOTE OR CITE
-------
1 Roemer, W.; Hoek, G.; Brunekreef, B.; Haluszka, J.; Kalandidi, A.; Pekkanen, J. (1998) Daily
2 variations in air pollution and respiratory health in a multicentre study: the PEACE
3 project. Eur. Respir. J. 12: 1354-1361.
4 Romieu, I; Meneses, F.; Ruiz, S.; Sienra, J. J.; Huerta, J.; White, M. C.; Etzel, R. A. (1996)
5 Effects of air pollution on the respiratory health of asthmatic children living in Mexico
6 City. Am. J. Respir. Crit. Care Med. 154: 300-307.
7 Routledge, H. C.; Manney, S.; Harrison, R. M.; Ayres, J. G.; Townend, J. N. (2006) Effect of
8 inhaled sulphur dioxide and carbon particles on heart rate variability and markers of
9 inflammation and coagulation in human subjects. Heart 92: 220-227.
10 Rubinstein, I; Bigby, B. G.; Reiss, T. F.; Boushey, H. A., Jr. (1990) Short-term exposure to 0.3
11 ppm nitrogen dioxide does not potentiate airway responsiveness to sulfur dioxide in
12 asthmatic subjects. Am. Rev. Respir. Dis. 141: 381-385.
13 Ruidavets, J.-B.; Cassadou, S.; Cournot, M.; Bataille, V.; Meybeck, M.; Ferrieres, J. (2005)
14 Increased resting heart rate with pollutants in a population based study. J. Epidemiol.
15 Community Health 59: 685-693.
16 Saez, M.; Tobias, A.; Mufioz, P.; Campbell, M. J. (1999) A GEE moving average analysis of the
17 relationship between air pollution and mortality for asthma in Barcelona, Spain. Stat.
18 Med. 18: 2077-2086.
19 Sagiv, S. K.; Mendola, P.; Loomis, D.; Herring, A. H.; Neas, L. M.; Savitz, D. A.; Poole, C.
20 (2005) A time-series analysis of air pollution and preterm birth in Pennsylvania, 1997-
21 2001. Environ. Health Perspect. 113: 602-606.
22 Samet, J. M.; Dominici, F.; Zeger, S. L.; Schwartz, J.; Dockery, D. W. (2000) National
23 morbidity, mortality, and air pollution study. Part I: methods and methodologic issues.
24 Cambridge, MA: Health Effects Institute; research report no. 94.
25 Samoli, E.; Schwartz, J.; Wojtyniak, B.; Touloumi, G.; Spix, C.; Balducci, F.; Medina, S.; Rossi,
26 G.; Sunyer, J.; Bacharova, L.; Anderson, H. R.; Katsouyanni, K. (2001) Investigating
27 regional differences in short-term effects of air pollution on daily mortality in the
28 APHEA project: a sensitivity analysis for controlling long-term trends and seasonality.
29 Environ. Health Perspect. 109: 349-353.
30 Samoli, E.; Schwartz, J.; Analitis, A.; Petasakis, Y.; Wojtyniak, B.; Touloumi, G.; Spix, C.;
31 Balducci, F.; Medina, S.; Rossi, G.; Sunyer, J.; Anderson, H. R.; Katsouyanni, K. (2003)
32 Sensitivity analyses of regional differences in short-term effects of air pollution on daily
33 mortality in APHEA cities. In: Revised analyses of time-series studies of air pollution
34 and health. Special report. Boston, MA: Health Effects Institute; pp. 205-210. Available:
35 http://www.healtheffects.org/Pubs/TimeSeries.pdf [18 October, 2004].
36 Sandstrom T.; Stjernberg, N.; Andersson, M. D.; Kolmodin-Hedman, B.; Lindstrom, K.;
37 Rosenhall, L. (1989) Cell response in bronchoalveolar lavage fluid after sulfur dioxide
38 exposure. Scand. J. Work Environ. Health 15: 142-146.
39 Sarnat, J. A.; Koutrakis, P.; Suh, H. H. (2000) Assessing the relationship between personal
40 particulate and gaseous exposures of senior citizens living in Baltimore, MD. J. Air
41 Waste Manage. Assoc. 50: 1184-1198.
September 2007 6-25 DRAFT-DO NOT QUOTE OR CITE
-------
1 Sarnat, J. A.; Schwartz, J.; Catalano, P. J.; Suh, H. H. (2001) Gaseous pollutants in paniculate
2 matter epidemiology: confounders or surrogates? Environ. Health Perspect. 109: 1053-
3 1061.
4 Sarnat, J. A.; Brown, K. W.; Schwartz, J.; Coull, B. A.; Koutrakis, P. (2005) Ambient gas
5 concentrations and personal particulate matter exposures: implications for studying the
6 health effects of particles. Epidemiology 16: 385-395.
7 Sarnat, S. E.; Coull, B. A.; Schwartz, J.; Gold, D. R.; Suh, H. H. (2006) Factors affecting the
8 association between ambient concentrations and personal exposures to particles and
9 gases. Environ. Health Perspect. 114: 649-654.
10 Scanlon, P. D.; Seltzer, J.; Ingram, R. H., Jr.; Reid, L.; Drazen, J. M. (1987) Chronic exposure to
11 sulfur dioxide: physiologic and histologic evaluation of dogs exposed to 50 or 15 ppm.
12 Am. Rev. Respir. Dis. 135: 831-839.
13 Schachter, E. N.; Witek, T. J., Jr.; Beck, G. J.; Hosein, H. R.; Colice, G.; Leaderer, B. P.; Cain,
14 W. (1984) Airway effects of low concentrations of sulfur dioxide: dose-response
15 characteristics. Arch. Environ. Health 39: 34-42.
16 Schildcrout, J. S.; Sheppard, L.; Lumley, T.; Slaughter, J. C.; Koenig, J. Q.; Shapiro, G. G.
17 (2006) Ambient air pollution and asthma exacerbations in children: an eight-city analysis.
18 Am. J. Epidemiol. 164: 505-517.
19 Schouten, J. P.; Vonk, J. M.; de Graaf, A. (1996) Short term effects of air pollution on
20 emergency hospital admissions for respiratory disease: results of the APHEA project in
21 two major cities in The Netherlands, 1977-89. In: St Leger, S., ed. The APHEA project.
22 Short term effects of air pollution on health: a European approach using epidemiological
23 time series data. J. Epidemiol. Community Health 50(suppl. 1): S22-S29.
24 Schwartz, J. (1989) Lung function and chronic exposure to air pollution: a cross-sectional
25 analysis of NHANES II. Environ. Res. 50: 309-321.
26 Schwartz, J. (1995) Short term fluctuations in air pollution and hospital admissions of the elderly
27 for respiratory disease. Thorax 50: 531-538.
28 Schwartz, J. (2000) Daily deaths are associated with combustion particles rather than SC>2 in
29 Philadelphia. Occup. Environ. Med. 57: 692-697.
30 Schwartz, J. (2001) Air pollution and blood markers of cardiovascular risk. Environ. Health
31 Perspect. Suppl. 109(3): 405-409.
32 Schwartz, J.; Dockery, D. W. (1992a) Increased mortality in Philadelphia associated with daily
33 air pollution concentrations. Am. Rev. Respir. Dis. 145: 600-604.
34 Schwartz, J.; Dockery, D. W. (1992b) Particulate air pollution and daily mortality in
35 Steubenville, Ohio. Am. J. Epidemiol. 135: 12-19.
36 Schwartz, J.; Dockery, D. W.; Neas, L. M.; Wypij, D.; Ware, J. H.; Spengler, J. D.; Koutrakis,
37 P.; Speizer, F. E.; Ferris, B. G., Jr. (1994) Acute effects of summer air pollution on
38 respiratory symptom reporting in children. Am. J. Respir. Crit. Care Med. 150: 1234-
39 1242.
September 2007 6-26 DRAFT-DO NOT QUOTE OR CITE
-------
1 Schwartz, J.; Litonjua, A.; Suh, H.; Verrier, M.; Zanobetti, A.; Syring, M.; Nearing, B.; Verrier,
2 R.; Stone, P.; MacCallum, G.; Speizer, F. E.; Gold, D. R. (2005) Traffic related pollution
3 and heart rate variability in a panel of elderly subjects. Thorax 60: 455-461.
4 Segala, C.; Fauroux, B.; Just, J.; Pascual, L.; Grimfeld, A.; Neukirch, F. (1998) Short-term effect
5 of winter air pollution on respiratory health of asthmatic children in Paris. Eur. Respir. J.
6 11:677-685.
7 Seinfeld, J. H.; Pandis, S. N. (1998) Atmospheric chemistry and physics: from air pollution to
8 climate change. New York, NY: John Wiley & Sons, Inc.
9 Sheppard, D.; Saisho, A.; Nadel, J. A.; Boushey, H. A. (1981) Exercise increases sulfur dioxide-
10 induced bronchoconstriction in asthmatic subjects. Am. Rev. Respir. Dis. 123: 486-491.
11 Sheppard, L.; Levy, D.; Norris, G.; Larson, T. V.; Koenig, J. Q. (1999) Effects of ambient air
12 pollution on nonelderly asthma hospital admissions in Seattle, Washington, 1987-1994.
13 Epidemiology 10: 23-30.
14 Sheppard, L.; Slaughter, J. C.; Schildcrout, J.; Liu, L.-J. S.; Lumley, T. (2005) Exposure and
15 measurement contributions to estimates of acute air pollution effects. J. Exposure Anal.
16 Environ. Epidemiol. 15: 366-376.
17 Shindell, D. T.; Faluvegi, G.; Stevenson, D. S.; Krol, M. C.; Emmons, L. K.; Lamarque, J.-F.;
18 Petron, G.; Dentener, F. J.; Ellingsen, K.; Schultz, M. G.; Wild, O.; Amann, M.;
19 Atherton, C. S.; Bergmann, D. J.; Bey, L; Butler, T.; Cofala, J.; Collins, W. J.; Derwent,
20 R. G.; Doherty, R. M.; Drevet, J.; Eskes, H. J.; Fiore, A. M.; Gauss, M.; Hauglustaine, D.
21 A.; Horowitz, L. W.; Isaksen, I. S. A.; Lawrence, M. G.; Montanaro, V.; Miiller, J.-F.;
22 Pitari, G.; Prather, M. J.; Pyle, J. A.; Rast, S.; Rodriguez, J. M.; Sanderson, M. G.;
23 Savage, N. H.; Strahan, S. E.; Sudo, K.; Szopa, S.; Unger, N.; Van Noije, T. P. C.; Zeng,
24 G. (2006) Multimodel simulations of carbon monoxide: comparison with observations
25 and projected near-future changes. J. Geophys. Res. [Atmos.] 111(D19306):
26 10.1029/2006JD007100.
27 Sim, V. M.; Pattle, R. E. (1957) Effect of possible smog irritants on human subjects. JAMA J.
28 Am. Med. Assoc. 165: 1908-1913.
29 Singh, J. (1989) Neonatal development altered by maternal sulfur dioxide exposure.
30 NeuroToxicology 10: 523-527.
31 Smith, L. G.; Busch, R. H.; Buschbom, R. L.; Cannon, W. C.; Loscutoff, S. M.; Morris. J. E.
32 (1989) Effects of sulfur dioxide or ammonium sulfate exposure, alone or combined, for 4
33 or 8 months on normal and elastase-impaired rats. Environ. Res. 49: 60-78.
34 Snell, R. E.; Luchsinger, P. C. (1969) Effects of sulfur dioxide on expiratory flow rates and total
35 respiratory resistance in normal human subjects. Arch. Environ. Health 18: 693-698.
36 S0yseth, V.; Kongerud, J.; Broen, P.; Lilleng, P.; Boe, J. (1995) Bronchial responsiveness,
37 eosinophilia, and short term exposure to air pollution. Arch. Dis. Child. 73: 418-422.
38 Speizer, F. E.; Frank, N. R. (1966) The uptake and release of SC>2 by the human nose. Arch.
39 Environ. Health 12: 725-728.
September 2007 6-27 DRAFT-DO NOT QUOTE OR CITE
-------
1 Spengler, J. D.; Ferris, B. G., Jr.; Dockery, D. W.; Speizer, F. E. (1979) Sulfur dioxide and
2 nitrogen dioxide levels inside and outside homes and the implications on health effects
3 research. Environ. Sci. Technol. 13: 1276-1280.
4 Spix, C.; Anderson, H. R.; Schwartz, J.; Vigotti, M. A.; LeTertre, A.; Vonk, J. M.; Touloumi, G.;
5 Balducci, F.; Piekarski, T.; Bacharova, L.; Tobias, A.; Ponka, A.; Katsouyanni, K. (1998)
6 Short-term effects of air pollution on hospital admissions of respiratory diseases in
7 Europe: a quantitative summary of APHEA study results. Arch. Environ. Health 53: 54-
8 64.
9 Stacy, R. W.; House, D. E.; Friedman, M.; Hazucha, M.; Green, J.; Raggio, L.; Roger, L. J.
10 (1981) Effects of 0.75ppm sulfur dioxide on pulmonary function parameters of normal
11 human subjects. Arch. Environ. Health 36: 172-178.
12 Stacy, R. W.; Seal, E., Jr.; House, D. E.; Green, J.; Roger, L. J.; Raggio, L. (1983) A survey of
13 effects of gaseous and aerosol pollutants on pulmonary function of normal males. Arch.
14 Environ. Health 38: 104-115.
15 Stempel, D. A.; Boucher, R. C. (1981) Respiratory infection and airway reactivity. Med. Clin.
16 North Am. 65: 1045-1053.
17 Stevenson, D.; Dentener, F. J.; Schultz, M. G.; Ellingsen, K.; Van Noije, T. P. C.; Wild, O.;
18 Zeng, G.; Amann, M.; Atherton, C. S.; Bell, N.; Bergmann, D. J.; Bey, L; Butler, T.;
19 Cofala, J.; Collins, W. J.; Derwent, R. G.; Doherty, R. M.; Drevet, J.; Eskes, H. J.; Fiore,
20 A. M.; Gauss, M.; Hauglustaine, D. A.; Horowitz, L. W.; Isaksen, I. S. A.; Krol, M. C.;
21 Lamarque, J.-F.; Lawrence, M. G.; Montanaro, V.; Miiller, J.-F.; Pitari, G.; Prather, M. J.;
22 Pyle, J. A.; Rast, S.; Rodriguez, J. M.; Sanderson, M. G. (2006) Multimodel ensemble
23 simulations of present-day and near-future tropospheric ozone. J. Geophys. Res. [Atmos.]
24 111(D08301): 10.1029/2005JD006338.
25 Stieb, D. M.; Judek, S.; Burnett, R. T. (2002) Meta-analysis of time-series studies of air pollution
26 and mortality: effects of gases and particles and the influence of cause of death, age, and
27 season. J. Air Waste Manage. Assoc. 52: 470-484.
28 Stieb, D. M.; Judek, S.; Burnett, R. T. (2003) Meta-analysis of time-series studies of air pollution
29 and mortality: update in relation to the use of generalized additive models. J. Air Waste
30 Manage. 53:258-261.
31 Strandberg, L. G. (1964) SC>2 absorption in the respiratory tract: studies on the absorption in
32 rabbit, its dependence on concentration and breathing phase. Arch. Environ. Health 9:
33 160-166.
34 Studnicka, M.; Hackl, E.; Pischinger, J.; Fangmeyer, C.; Haschke, N.; Kuhr, J.; Urbanek, R.;
35 Neumann, M.; Frischer, T. (1997) Traffic-related NC>2 and the prevalence of asthma and
36 respiratory symptoms in seven year olds. Eur. Respir. J. 10: 2275-2278.
37 Sunyer, J.; Spix, C.; Quenel, P.; Ponce-de-Leon, A.; Ponka, A.; Barumandzadeh, T.; Touloumi,
38 G.; Bacharova, L.; Wojtyniak, B.; Vonk, J.; Bisanti, L.; Schwartz, J.; Katsouyanni, K.
39 (1997) Urban air pollution and emergency admissions for asthma in four European cities:
40 the APHEA project. Thorax 52: 760-765.
41 Sunyer, J.; Basagafia, X.; Belmonte, J.; Anto, J. M. (2002) Effect of nitrogen dioxide and ozone
42 on the risk of dying in patients with severe asthma. Thorax 57: 687-693.
September 2007 6-28 DRAFT-DO NOT QUOTE OR CITE
-------
1 Sunyer, J.; Atkinson, R.; Ballester, F.; Le Tertre, A.; Ayres, J. G.; Forastiere, F.; Forsberg, B.;
2 Vonk, J. M.; Bisanti, L.; Anderson, R. H.; Schwartz, J.; Katsouyanni, K. (2003a)
3 Respiratory effects of sulphur dioxide: a hierarchical multicity analysis in the APFIEA 2
4 study. Occup. Environ. Med. 60: e2.
5 Sunyer, J.; Ballester, F.; Le Tertre, A.; Atkinson, R.; Ayres, J. G.; Forastiere, F.; Forsberg, B.;
6 Vonk, J. M.; Bisanti, L.; Tenias, J. M.; Medina, S.; Schwartz, J.; Katsouyanni, K. (2003b)
7 The association of daily sulfur dioxide air pollution levels with hospital admissions for
8 cardiovascular diseases in Europe (The Aphea-II study). Eur. Heart J. 24: 752-760.
9 Taggart, S. C. O.; Custovic, A.; Francis, H. C.; Faragher, E. B.; Yates, C. J.; Higgins, B. G.;
10 Woodcock, A. (1996) Asthmatic bronchial hyperresponsiveness varies with ambient
11 levels of summertime air pollution. Eur. Respir. J. 9: 1146-1154.
12 Tenias, J. M.; Ballester, F.; Rivera, M. L. (1998) Association between hospital emergency visits
13 for asthma and air pollution in Valencia, Spain. Occup. Environ. Med. 55: 541-547.
14 Tenias, J. M.; Ballester, F.; Perez-Hoyos, S.; Rivera, M. L. (2002) Air pollution and hospital
15 emergency room admissions for chronic obstructive pulmonary disease in Valencia,
16 Spain. Arch. Environ. Health 57: 41-47.
17 Thompson, A. J.; Shields, M. D.; Patterson, C. C. (2001) Acute asthma exacerbations and air
18 pollutants in children living in Belfast, Northern Ireland. Arch. Environ. Health 56: 234-
19 241.
20 Thornton, D. C.; Bandy, A. R. (1993) Sulfur dioxide and dimethyl sulfide in the central Pacific
21 troposphere. J. Atmos. Chem. 17: 1-13.
22 Thornton, D. C.; Bandy, A. R.; Blomquist, B. W.; Driedger, A. R.; Wade, T. P. (1999) Sulfur
23 dioxide distribution over the Pacific Ocean 1991-1996. J. Geophys. Res. [Atmos.] 104:
24 5845-5854.
25 Thornton, D. C.; Bandy, A. R.; Tu, F. H.; Blomquist, B. W.; Mitchell, G. M.; Nadler, W.;
26 Lenschow, D. H. (2002) Fast airborne sulfur dioxide measurements by atmospheric
27 pressure ionization mass spectrometry (APIMS). J. Geophys. Res. [Atmos.] 107(D22):
28 10.1029/2002JD02289.
29 Timonen, K. L.; Pekkanen, J. (1997) Air pollution and respiratory health among children with
30 asthmatic or cough symptoms. Am. J. Respir. Crit. Care Med. 156: 546-552.
31 Tolbert, P. E.; Klein, M.; Peel, J. L.; Sarnat, S. E.; Sarnat, J. A. (2007) Multipollutant modeling
32 issues in a study of ambient air quality and emergency department visits in Atlanta. J.
33 Exposure Sci. Environ. Epidemiol.: in press.
34 Triche, E. W.; Belanger, K.; Bracken, M. B.; Beckett, W. S.; Holford, T. R.; Gent, J. F.;
35 McSharry, J.-E.; Leaderer, B. P. (2005) Indoor heating sources and respiratory symptoms
36 in nonsmoking women. Epidemiology 16: 377-384.
37 Tsai, S.-S.; Goggins, W. B.; Chiu, H.-F.; Yang, C.-Y. (2003) Evidence for an association
38 between air pollution and daily stroke admissions in Kaohsiung, Taiwan. Stroke 34:
39 2612-2616.
September 2007 6-29 DRAFT-DO NOT QUOTE OR CITE
-------
1 Tunnicliffe, W. S.; Hilton, M. F.; Harrison, R. M.; Ayres, J. G. (2001) The effect of sulphur
2 dioxide exposure on indices of heart rate variability in normal and asthmatic adults. Eur.
3 Respir. J. 17: 604-608.
4 Tunnicliffe, W. S.; Harrison, R. M.; Kelly, F. J.; Dunster, C.; Ayres, J. G. (2003) The effect of
5 sulphurous air pollutant exposures on symptoms, lung function, exhaled nitric oxide, and
6 nasal epithelial lining fluid antioxidant concentrations in normal and asthmatic adults.
7 Occup. Environ. Med. 60: e!5.
8 U.S. Code. (2003a) Clean Air Act, §108, air quality criteria and control techniques. U. S. C. 42:
9 §7408.
10 U.S. Code. (2003b) Clean Air Act, §109, national ambient air quality standards. U. S. C. 42:
11 §7409.
12 U.S. Code. (2005) Clean Air Act, §302, definitions. U. S. C. 42: §7602(h).
13 U.S. Court of Appeals for the District of Columbia. (1980) Lead Industries v. U.S.
14 Environmental Protection Agency. 647 F2d 1130, 1154 (DC Cir. 1980).
15 U.S. Court of Appeals for the District of Columbia. (1981) American Petroleum Institute v.
16 Costle. 665 F2d 1176, 1186 (DC Cir. 1981).
17 U. S. Department of Health, Education, and Welfare (DHEW). (1969) Air quality criteria for
18 sulfur oxides. Washington, DC: National Air Pollution Control Administration; report no.
19 AP-50. Available from: NTIS, Springfield, VA; PB-190252.
20 U.S. Environmental Protection Agency. (1982) Air quality criteria for particulate matter and
21 sulfur oxides. Research Triangle Park, NC: Office of Health and Environmental
22 Assessment, Environmental Criteria and Assessment Office; EPA report no. EPA-600/8-
23 82-029aF-cF. 3v. Available from: NTIS, Springfield, VA; PB84-156777.
24 U.S. Environmental Protection Agency. (1986a) Second addendum to air quality criteria for
25 particulate matter and sulfur oxides (1982): assessment of newly available health effects
26 information. Research Triangle Park, NC: Office of Health and Environmental
27 Assessment, Environmental Criteria and Assessment Office; EPA report no. EPA-600/8-
28 86-020F. Available from: NTIS, Springfield, VA; PB87-176574.
29 U.S. Environmental Protection Agency. (1986b) Review of the national ambient air quality
30 standards for sulfur oxides: updated assessment of scientific and technical information,
31 addendum to the 1982 OAQPS staff paper. Research Triangle Park, NC: Office of Air
32 Quality Planning and Standards, Strategies and Air Standards Division; report no.
33 EPA/450/5-86-13. Available from: NTIS, Springfield, VA; PB87-200259/XAB.
34 U.S. Environmental Protection Agency. (1994a) Supplement to the second addendum (1986) to
35 air quality criteria for particulate matter and sulfur oxides (1982): assessment of new
36 findings on sulfur dioxide acute exposure health effects in asthmatic individuals.
37 Research Triangle Park, NC: Office of Health and Environmental Assessment,
38 Environmental Criteria and Assessment Office; report no. EPA-600/FP-93/002. Available
39 from: NTIS, Springfield, VA; PB98-132384.
40 U.S. Environmental Protection Agency. (1994b) Review of the national ambient air quality
41 standards for sulfur oxides: assessment of scientific and technical information,
September 2007 6-30 DRAFT-DO NOT QUOTE OR CITE
-------
1 supplement to the 1986 OAQPS staff paper addendum. Research Triangle Park, NC:
2 Office of Air Quality Planning and Standards, Air Quality Management Division; report
3 no. EPA/452/R-94/013.
4 U.S. Environmental Protection Agency. (1996) Air quality criteria for ozone and related
5 photochemical oxidants. Research Triangle Park, NC: Office of Research and
6 Development; report nos. EPA/600/AP-93/004aF-cF. 3v. Available from: NTIS,
7 Springfield, VA; PB96-185582, PB96-185590, and PB96-185608. Available:
8 http://cfpub2.epa.gov/ncea/.
9 U.S. Environmental Protection Agency. (2004) Air quality criteria for particulate matter.
10 Research Triangle Park, NC: National Center for Environmental Assessment; report no.
11 EPA/600/P-99/002aF-bF. 2v. Available: http://cfpub.epa.gov/ncea/ [9 November, 2004].
12 U.S. Environmental Protection Agency. (2005) Technical assistance document (TAD) for
13 precursor gas measurements in the NCore multi-pollutant monitoring network. Version 4.
14 Research Triangle Park, NC: Office of Air Quality Planning and Standards; report no.
15 EPA-454/R-05-003.
16 U.S. Environmental Protection Agency. (2006a) Technology Transfer Network clearinghouse for
17 inventories and emission factors. Available:
18 http://www.epa.gov/ttn/chief/net/2002inventory.html [19 September, 2007].
19 U.S. Environmental Protection Agency. (2006b) Report on air quality in nonattainment areas for
20 2003-2005 covering ozone, parti culate matter, carbon monoxide, sulfur dioxide, nitrogen
21 dioxide, and lead. Research Triangle Park, NC: Office of Air Quality Planning and
22 Standards, Air Quality Assessment Division, Air Quality Analysis Group.
23 U.S. Environmental Protection Agency. (2006c) Acid rain program (ARP) 2005 progress report.
24 Washington, DC: Office of Air and Radiation, Clean Air Markets Division; EPA-430-R-
25 06-015. Available: http://www.epa.gov/airmarkets/progress/docs/2005report.pdffl2
26 March, 2007].
27 U.S. Environmental Protection Agency. (2006d) Air quality criteria for ozone and related
28 photochemical oxidants. Research Triangle Park, NC: National Center for Environmental
29 Assessment; report no. EPA/600/R-05/004aF-cF. 3v. Available:
30 http://cfpub.epa.gov/ncea/ [24 March, 2006].
31 U.S. Environmental Protection Agency. (2007) Draft integrated plan for review of the primary
32 national ambient air quality standard for sulfur dioxide. Research Triangle Park, NC:
33 National Center for Environmental Assessment. Available:
34 http://nsdi.epa.gov/ttn/naaqs/standards/so2/data/so2_review_plan_draft_4-17-07.pdf [11
35 September, 2007].
36 U.S. Senate. (1970) National Air Quality Standards Act of 1970: report of the Committee on
37 Public Works, United States Senate together with individual views to accompany S.
38 4358. Washington, DC: Committee on Public Works; report no. CONG/91-1196.
39 U.S. Supreme Court. (2001) Whitman v. American Trucking Association. 531 U.S. 457 (nos. 99-
40 1257 and 99-1426).
September 2007 6-31 DRAFT-DO NOT QUOTE OR CITE
-------
1 Van Der Zee, S. C.; Hoek, G.; Boezen, H. M.; Schouten, J. P.; Van Wijnen, J. H.; Brunekreef, B.
2 (1999) Acute effects of urban air pollution on respiratory health of children with and
3 without chronic respiratory symptoms. Occup. Environ. Med. 56: 802-813.
4 Van Der Zee, S. C.; Hoek, G.; Boezen, M. H.; Schouten, J. P.; Van Wijnen, J. H.; Brunekreef, B.
5 (2000) Acute effects of air pollution on respiratory health of 50-70 yr old adults. Eur.
6 Respir. J. 15: 700-709.
7 Van Noije, T. P. C.; Eskes, H. J.; Dentener, F. J.; Stevenson, D. S.; Ellingsen, K.; Schultz, M. G.;
8 Wild, O.; Amann, M.; Atherton, C. S.; Bergmann, D. J.; Bey, I; Boersma, K. F.; Butler,
9 T.; Cofala, J.; Drevet, J.; Fiore, A. M.; Gauss, M.; Hauglustaine, D. A.; Horowitz, L. W.;
10 Isaksen, I. S. A.; Krol, M. C.; Lamarque, J.-F.; Lawrence, M. G.; Martin, R. V.;
11 Montanaro, V.; Muller, J.-F.; Pitari, G.; Prather, M. J.; Pyle, J. A.; Richter, A.;
12 Rodriguez, J. M.; Savage, N. H.; Strahan, S. E.; Sudo, K., Szopa, S.; Van Roozendael, M.
13 (2006) Multi-model ensemble simulations of tropospheric NC>2 compared with GOME
14 retrievals for the year 2000. Atmos. Chem. Phys. Discuss. 6: 2965-3047.
15 Vedal, S.; Rich, K.; Brauer, M.; White, R.; Petkau, J. (2004) Air pollution and cardiac
16 arrhythmias in patients with implantable cardiovascular defibrillators. Inhalation Toxicol.
17 16:353-362.
18 Vigotti, M. A.; Rossi, G.; Bisanti, L.; Zanobetti, A.; Schwartz, J. (1996) Short term effects of
19 urban air pollution on respiratory health in Milan, Italy, 1980-89. In: Leger, S., ed. The
20 APHEA project. Short term effects of air pollution on health: a European approach using
21 epidemiological time series data. J. Epidemiol. Community Health 50(suppl. 1): S71-S75.
22 Villeneuve, P. J.; Chen, L.; Stieb, D.; Rowe, B. H. (2006) Associations between outdoor air
23 pollution and emergency department visits for stroke in Edmonton, Canada. Eur. J.
24 Epidemiol. 21: 689-700.
25 Walters, S.; Griffiths, R. K.; Ayres, J. G. (1994) Temporal association between hospital
26 admissions for asthma in Birmingham and ambient levels of sulphur dioxide and smoke.
27 Thorax 49: 133-140.
28 Wang, C.; Corbett, J. J.; Firestone, J. (2007) Modeling energy use and emissions from North
29 American shipping: application of the ship traffic, energy, and environment model.
30 Environ. Sci. Technol. 41: 3226-3232.
31 Wang, X.; Ding, H.; Ryan, L.; Xu, X. (1997) Association between air pollution and low birth
32 weight: a community-based study. Environ. Health Perspect. 105: 514-520.
33 Ward, D. J.; Roberts, K. T.; Jones, N.; Harrison, R. M.; Ayres, J. G.; Hussain, S.; Walters, S.
34 (2002a) Effects of daily variation in outdoor particulates and ambient acid species in
35 normal and asthmatic children. Thorax 57: 489-502.
36 Ward, D. J.; Roberts, K. T.; Jones, N.; Harrison, R. M.; Ayres, J. G.; Hussain, S.; Walters, S.
37 (2002b) Effects of daily variation in outdoor particulates and ambient acid species in
38 normal and asthmatic children (correction). Thorax 57: 752.
39 Ware, J. H.; Ferris, B. G., Jr.; Dockery, D. W.; Spengler, J. D.; Stram, D. O.; Speizer, F. E.
40 (1986) Effects of ambient sulfur oxides and suspended particles on respiratory health of
41 preadolescent children. Am. Rev. Respir. Dis. 133: 834-842.
September 2007 6-32 DRAFT-DO NOT QUOTE OR CITE
-------
1 Warneck, P. (1999) The relative importance of various pathways for the oxidation of sulfur
2 dioxide and nitrogen dioxide in sunlit continental fair weather clouds. Phys. Chem.
3 Chem. Phys. 1: 5471-5483.
4 Wellenius, G. A.; Schwartz, J.; Mittleman, M. A. (2005a) Air pollution and hospital admissions
5 for ischemic and hemorrhagic stroke among medicare beneficiaries. Stroke 36: 2549-
6 2553.
7 Wellenius, G. A.; Bateson, T. F.; Mittleman, M. A.; Schwartz, J. (2005b) Particulate air pollution
8 and the rate of hospitalization for congestive heart failure among medicare beneficiaries
9 in Pittsburgh, Pennsylvania. Am. J. Epidemiol. 161: 1030-1036.
10 Wiley, J. A.; Robinson, J. P.; Piazza, T.; Garrett, K.; Cirksena, K.; Cheng, Y.-T.; Martin, G.
11 (1991a) Activity patterns of California residents. Final report. Sacramento, CA:
12 California Air Resources Board; report no. ARB/R93/487. Available from: NTIS,
13 Springfield, VA.; PB94-108719.
14 Wiley, J. A.; Robinson, J. P.; Cheng, Y.-T.; Piazza, T.; Stork, L.; Pladsen, K. (1991b) Study of
15 children's activity patterns: final report. Sacramento, CA: California Air Resources
16 Board; report no. ARB-R-93/489.
17 Williams, R.; Suggs, J.; Creason, J.; Rodes, C.; Lawless, P.; Kwok, R.; Zweidinger, R.; Sheldon,
18 L. (2000) The 1998 Baltimore Particulate Matter Epidemiology-Exposure Study: part 2.
19 Personal exposure assessment associated with an elderly study population. J. Exposure
20 Anal. Environ. Epidemiol. 10: 533-543.
21 Williams, R.; Suggs, J.; Rea, A.; Sheldon, L.; Rodes, C.; Thornburg, J. (2003). The Research
22 Triangle Park particulate matter panel study: modeling ambient source contribution to
23 personal and residential PM concentrations. Atmos. Environ. 37: 5365-5378.
24 Willis, A.; Jerrett, M.; Burnett, R. T.; Krewski, D. (2003) The association between sulfate air
25 pollution and mortality at the county scale: an exploration of the impact of scale on a
26 long-term exposure study. J. Toxicol. Environ. Health Part A 66: 1605-1624.
27 Wilson, W. E.; Brauer, M. (2006) Estimation of ambient and non-ambient components of
28 particulate matter exposure from a personal monitoring panel study. J. Exposure Sci.
29 Environ. Epidemiol. 16: 264-274.
30 Wilson, A. M.; Wake, C. P.; Kelly, T.; Salloway, J. C. (2005) Air pollution, weather, and
31 respiratory emergency room visits in two northern New England cities: an ecological
32 time-series study. Environ. Res. 97: 312-321.
33 Winterton, D. L.; Kaufman, J.; Keener, C. V.; Quigley, S.; Farin, F. M.; Williams, P. V.; Koenig,
34 J. Q. (2001) Genetic polymorphisms as biomarkers of sensitivity to inhaled sulfur dioxide
35 in subjects with asthma. Ann. Allergy Asthma Immunol. 86: 232-238.
36 Wolff, R. K.; Griffith, W. C.; Henderson, R. F.; Hahn, F. F.; Harkema, J. R.; Rebar, A. H.;
37 Eidson, A. F.; McClellan, R. O. (1989) Effects of repeated inhalation exposures to 1-
38 nitropyrene, benzo(a)pyrene, Ga2Os particles, and 862 alone and in combinations on
39 particle clearance, bronchoalveolar lavage fluid composition, and histopathology. J.
40 Toxicol. Environ. Health 27: 123-138.
September 2007 6-33 DRAFT-DO NOT QUOTE OR CITE
-------
1 Wong, T. W.; Lau, T. S.; Yu, T. S.; Neller, A.; Wong, S. L.; Tarn, W.; Pang, S. W. (1999) Air
2 pollution and hospital admissions for respiratory and cardiovascular diseases in Hong
3 Kong. Occup. Environ. Med. 56: 679-683.
4 Wong, C.-M.; Ma, S.; Hedley, A. I; Lam, T.-H. (2001) Effect of air pollution on daily mortality
5 in Hong Kong. Environ. Health Perspect. 109: 335-340.
6 Wong, C.-M.; Atkinson, R. W.; Anderson, H. R.; Hedley, A. I; Ma, S.; Chau, P. Y.-K.; Lam, T.-
7 H. (2002) A tale of two cities: effects of air pollution on hospital admissions in Hong
8 Kong and London compared. Environ. Health Perspect. 110: 61-11.
9 Woodwell, D. A.; Cherry, D. K. (2004) National Ambulatory Medical Care Survey: 2002
10 summary. Hyattsville, MD: National Center for Health Statistics; DHHS publication no.
11 (PHS) 2004-1250. (Advance data from vital and health statistics; no. 346). Available:
12 http://www.cdc.gov/nchs/data/ad/ad346.pdff3 August, 2005].
13 Wu, D.; Meng, Z. (2003) Effect of sulfur dioxide inhalation on the glutathione redox system in
14 mice and protective role of sea buckthorn seed oil. Arch. Environ. Contam. Toxicol. 45:
15 423-428.
16 Xu, X.; Dockery, D. W.; Wang, L. (1991) Effects of air pollution on adult pulmonary function.
17 Arch. Environ. Health 46: 198-206.
18 Xu, X.; Ding, H.; Wang, X. (1995) Acute effects of total suspended particles and sulfur dioxides
19 on preterm delivery: a community-based cohort study. Arch. Environ. Health 50: 407-
20 415.
21 Yang, Q.; Chen, Y.; Shi, Y.; Burnett, R. T.; McGrail, K. M.; Krewski, D. (2003a) Association
22 between ozone and respiratory admissions among children and the elderly in Vancouver,
23 Canada. Inhalation Toxicol. 15: 1297-1308.
24 Yang, C.-Y.; Tseng, Y.-T.; Chang, C.-C. (2003b) Effects of air pollution on birthweight among
25 children born between 1995 and 1997 in Kaohsiung, Taiwan. J. Toxicol. Environ. Health
26 Part A 66: 807-816.
27 Yang, Q.; Chen, Y.; Krewski, D.; Burnett, R. T.; Shi, Y.; McGrail, K. M. (2005) Effect of short-
28 term exposure to low levels of gaseous pollutants on chronic obstructive pulmonary
29 disease hospitalizations. Environ. Res. 99: 99-105.
30 Yargi9oglu, A.; Agar, S.; Gumus.lu, S.; Bilmen, S.; Oguz, Y. (1999) Age-related alterations in
31 antioxidant enzymes, lipid peroxide levels, and somatosensory-evoked potentials: effect
32 of sulfur dioxide. Arch. Environ. Contam. Toxicol. 37: 554-560.
33 Yargi90glu, P.; Gumusluoriob, S.; Agar, A.; Korgun, D. K.; Ku9ukatay, V. (2001) Effect of
34 sulfur dioxide inhalation on erythrocyte antioxidant status, food intake, and lipid
35 peroxidation during aging. Arch. Environ. Health 56: 53-57.
36 Zeger, S. L.; Thomas, D.; Dominici, F.; Samet, J. M.; Schwartz, J.; Dockery, D.; Cohen, A.
37 (2000) Exposure measurement error in time-series studies of air pollution: concepts and
38 consequences. Environ. Health Perspect. 108: 419-426.
39 Zhang, J.-Z.; Millero, F. J. (1991) The rate of sulfite oxidation in seawater. Geochim.
40 Cosmochim. Acta 55: 677-685.
September 2007 6-34 DRAFT-DO NOT QUOTE OR CITE
-------
1 Zmirou, D.; Schwartz, J.; Saez, M.; Zanobetti, A.; Wojtyniak, B.; Touloumi, G.; Spix, C.; Ponce
2 de Leon, A.; Le Moullec, Y.; Bacharova, L.; Schouten, J.; Ponka, A.; Katsouyanni, K.
3 (1998) Time-series analysis of air pollution and cause-specific mortality. Epidemiology
4 9:495-503.
September 2007 6-3 5 DRAFT-DO NOT QUOTE OR CITE
-------
4>EPA
Please make all necessary changes in the below label, PRESORTED STANDARD
detach copy or copy, and return to the address in the upper POSTAGE & FEES PAID
left-hand corner. EPA
United States DCDMIT M^ r* Q^
Environmental Protection _ PERMIT Na Q-35
Agency If you do not wish to receive these reports CHECK HERE LJ;
detach copy or copy, and return to the address in the upper
left-hand corner.
National Center for
Environmental Assessment
Research Triangle Park, NC 27711
Official Business
Penalty for Private Use
$300
EPA/600/R-07/108
September 2007
------- |