National Ambient Air Quality Standards for Ozone

[Federal Register: July 11, 2007 (Volume 72, Number 132)]
[Proposed Rules]
[Page 37817-37866]
From the Federal Register Online via GPO Access [wais.access.gpo.gov]
[DOCID:fr11jy07-19]
[[Page 37818]]

-----------------------------------------------------------------------

ENVIRONMENTAL PROTECTION AGENCY
40 CFR Part 50
[EPA-HQ-OAR-2005-0172; FRL-8331-5]
RIN 2060-AN24

AGENCY: Environmental Protection Agency (EPA).
ACTION: Proposed rule.

-----------------------------------------------------------------------

SUMMARY: Based on its review of the air quality criteria for ozone
(O₃) and related photochemical oxidants and national ambient
air quality standards (NAAQS) for O₃, EPA proposes to make
revisions to the primary and secondary NAAQS for O₃ to
provide requisite protection of public health and welfare,
respectively, and to make corresponding revisions in data handling
conventions for O₃.
With regard to the primary standard for O₃, EPA proposes
to revise the level of the 8-hour standard to a level within the range
of 0.070 to 0.075 parts per million (ppm), to provide increased
protection for children and other ``at risk'' populations against an
array of O₃-related adverse health effects that range from
decreased lung function and increased respiratory symptoms to serious
indicators of respiratory morbidity including emergency department
visits and hospital admissions for respiratory causes, and possibly
cardiovascular-related morbidity as well as total nonaccidental and
cardiopulmonary mortality. The EPA also proposes to specify the level
of the primary standard to the nearest thousandth ppm. The EPA solicits
comment on alternative levels down to 0.060 ppm and up to and including
retaining the current 8-hour standard of 0.08 ppm (effectively 0.084
ppm using current data rounding conventions).
With regard to the secondary standard for O₃, EPA
proposes to revise the current 8-hour standard with one of two options
to provide increased protection against O₃-related adverse
impacts on vegetation and forested ecosystems. One option is to replace
the current standard with a cumulative, seasonal standard expressed as
an index of the annual sum of weighted hourly concentrations, cumulated
over 12 hours per day (8 a.m. to 8:00 p.m.) during the consecutive 3-
month period within the O₃ season with the maximum index
value, set at a level within the range of 7 to 21 ppm-hours. The other
option is to make the secondary standard identical to the proposed
primary 8-hour standard. The EPA solicits comment on specifying a
cumulative, seasonal standard in terms of a 3-year average of the
annual sums of weighted hourly concentrations; on the range of
alternative 8-hour standard levels for which comment is being solicited
for the primary standard, including retaining the current secondary
standard, which is identical to the current primary standard; and on an
alternative approach to setting a cumulative, seasonal secondary
standard(s).

DATES: Written comments on this proposed rule must be received by
October 9, 2007.

ADDRESSES: Submit your comments, identified by Docket ID No. EPA-HQ-
OAR-2005-0172, by one of the following methods:
• www.regulations.gov: Follow the on-line instructions for
submitting comments.
• E-mail: a-and-r-Docket@epa.gov.
• Fax: 202-566-1741.
• Mail: Docket No. EPA-HQ-OAR-2005-0172, Environmental
Protection Agency, Mail code 6102T, 1200 Pennsylvania Ave., NW.,
Washington, DC 20460. Please include a total of two copies.
• Hand Delivery: Docket No. EPA-HQ-OAR-2005-0172,
Environmental Protection Agency, EPA West, Room 3334, 1301 Constitution
Ave., NW., Washington, DC. Such deliveries are only accepted during the
Docket's normal hours of operation, and special arrangements should be
made for deliveries of boxed information.
Instructions: Direct your comments to Docket ID No. EPA-HQ-OAR-
2005-0172. The EPA's policy is that all comments received will be
included in the public docket without change and may be made available
online at www.regulations.gov, including any personal information
provided, unless the comment includes information claimed to be
Confidential Business Information (CBI) or other information whose
disclosure is restricted by statute. Do not submit information that you
consider to be CBI or otherwise protected through www.regulations.gov
or e-mail. The www.regulations.gov Web site is an ``anonymous access''
system, which means EPA will not know your identity or contact
information unless you provide it in the body of your comment. If you
send an e-mail comment directly to EPA without going through
www.regulations.gov, your e-mail address will be automatically captured
and included as part of the comment that is placed in the public docket
and made available on the Internet. If you submit an electronic
comment, EPA recommends that you include your name and other contact
information in the body of your comment and with any disk or CD-ROM you
submit. If EPA cannot read your comment due to technical difficulties
and cannot contact you for clarification, EPA may not be able to
consider your comment. Electronic files should avoid the use of special
characters, any form of encryption, and be free of any defects or
viruses. For additional information about EPA's public docket, visit
the EPA Docket Center homepage at http://www.epa.gov/epahome/
dockets.htm.
Docket: All documents in the docket are listed in the
www.regulations.gov index. Although listed in the index, some
information is not publicly available, e.g., CBI or other information
whose disclosure is restricted by statute. Certain other material, such
as copyrighted material, will be publicly available only in hard copy.
Publicly available docket materials are available either electronically
in www.regulations.gov or in hard copy at the Air and Radiation Docket
and Information Center, EPA/DC, EPA West, Room 3334, 1301 Constitution
Ave., NW., Washington, DC. The Public Reading Room is open from 8:30
a.m. to 4:30 p.m., Monday through Friday, excluding legal holidays. The
telephone number for the Public Reading Room is (202) 566-1744 and the
telephone number for the Air and Radiation Docket and Information
Center is (202) 566-1742.
Public Hearings: The EPA intends to hold public hearings around the
end of August to early September in several cities across the country,
and will announce in a separate Federal Register notice the dates,
times, and addresses of the public hearings on this proposed rule.

FOR FURTHER INFORMATION CONTACT: Dr. David J. McKee, Health and
Environmental Impacts Division, Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency, Mail code C504-06,
Research Triangle Park, NC 27711; telephone: 919-541-5288; fax: 919-
541-0237; e-mail: mckee.dave@epa.gov.

SUPPLEMENTARY INFORMATION:

General Information

What Should I Consider as I Prepare My Comments for EPA?

1. Submitting CBI. Do not submit this information to EPA through
www.regulations.gov or e-mail. Clearly mark the part or all of the
information that you claim to be CBI. For CBI information in a disk or
CD ROM that

[[Page 37819]]

you mail to EPA, mark the outside of the disk or CD ROM as CBI and then
identify electronically within the disk or CD ROM the specific
information that is claimed as CBI. In addition to one complete version
of the comment that includes information claimed as CBI, a copy of the
comment that does not contain the information claimed as CBI must be
submitted for inclusion in the public docket. Information so marked
will not be disclosed except in accordance with procedures set forth in
40 CFR part 2.
2. Tips for Preparing Your Comments. When submitting comments,
remember to:
• Identify the rulemaking by docket number and other
identifying information (subject heading, Federal Register date and
page number).
• Follow directions--The Agency may ask you to respond to
specific questions or organize comments by referencing a Code of
Federal Regulations (CFR) part or section number.
• Explain why you agree or disagree, suggest alternatives,
and substitute language for your requested changes.
• Describe any assumptions and provide any technical
information and/or data that you used.
• If you estimate potential costs or burdens, explain how you arrived
at your estimate in sufficient detail to allow for it to be reproduced.
• Provide specific examples to illustrate your concerns, and
suggest alternatives.
• Explain your views as clearly as possible, avoiding the
use of profanity or personal threats.
• Make sure to submit your comments by the comment period
deadline identified.

Availability of Related Information

A number of documents relevant to this rulemaking are available on
EPA Web sites. The Air Quality Criteria for Ozone and Related
Photochemical Oxidants (Criteria Document) (two volumes, EPA/ and EPA/,
date) is available on EPA's National Center for Environmental
Assessment Web site. To obtain this document, go to http://www.epa.gov/
ncea, and click on ``Ozone.'' The Staff Paper, human exposure and
health risk assessments, vegetation exposure and impact assessment, and
other related technical documents are available on EPA's Office of Air
Quality Planning and Standards (OAQPS) Technology Transfer Network
(TTN) Web site. The Staff Paper is available at http://www.epa.gov/
ttn/naaqs/standards/ozone/s_o3_cr_sp.html, and the exposure and risk
assessments and other related technical documents are available at
http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_cr_td.html. EPA
will be making available corrected versions of the final Staff Paper
and human exposure and health risk assessment technical support
documents on these same EPA Web sites on or around July 16, 2007. These
and other related documents are also available for inspection and
copying in the EPA docket identified above.

Table of Contents

The following topics are discussed in this preamble:
I. Background
A. Legislative Requirements
B. Related Control Requirements
C. Review of Air Quality Criteria and Standards for O₃
II. Rationale for Proposed Decision on the Primary Standard
A. Health Effects Information
1. Mechanisms
2. Nature of Effects
3. Interpretation and Integration of the Health Evidence
4. O₃-Related Impacts on Public Health
B. Human Exposure and Health Risk Assessments
1. Exposure Analyses
2. Quantitative Health Risk Assessment
C. Conclusions on the Adequacy of the Current Primary Standard
1. Background
2. Evidence- and Exposure/Risk-Based Considerations
3. CASAC Views
4. Administrator's Proposed Conclusions Concerning Adequacy of
Current Standard
D. Conclusions on the Elements of the Primary Standard
1. Indicator
2. Averaging Time
3. Form
4. Level
E. Proposed Decision on the Primary Standard
III. Communication of Public Health Information
IV. Rationale for Proposed Decision on the Secondary Standard
A. Vegetation Effects Information
1. Mechanisms Governing Plant Response to Ozone
2. Nature of Effects
3. Adversity of Effects
B. Biologically Relevant Exposure Indices
C. Vegetation Exposure and Impact Assessment
1. Exposure Characterization
2. Assessment of Risks to Vegetation
D. Conclusions on the Adequacy of the Current Standard
1. Background
2. Evidence- and Exposure/Risk-Based Considerations
3. CASAC Views
4. Administrator's Proposed Conclusions Concerning Adequacy of
Current Standard
E. Conclusions on the Elements of the Secondary Standard
1. Indicator
2. Cumulative, Seasonal Standard
3. 8-Hour Average Standard
F. Proposed Decision on the Secondary Standard
V. Creation of Appendix P--Interpretation of the NAAQS for Ozone
A. Data Completeness
B. Data Handling and Rounding O₃ Conventions
VI. Ambient Monitoring Related to Proposed Revised Standards
VII. Statutory and Executive Order Reviews
References

I. Background

A. Legislative Requirements

Two sections of the Clean Air Act (CAA) govern the establishment
and revision of the NAAQS. Section 108 (42 U.S.C. 7408) 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 (42 U.S.C. 7409) directs the Administrator to propose
and promulgate ``primary'' and ``secondary'' NAAQS for pollutants
listed under section 108. Section 109(b)(1) defines a primary standard
as one ``the attainment and maintenance of which in the judgment of the
Administrator, based on such criteria and allowing an adequate margin
of safety, are requisite to protect the public health.'' \1\ A
secondary standard, as defined in section 109(b)(2), must ``specify a
level of air quality the attainment and maintenance of which, in the
judgment of the Administrator, based on such criteria, is requisite to
protect the public welfare from any known or anticipated adverse
effects associated with the presence of [the]
pollutant in the ambient
air.'' \2\
---------------------------------------------------------------------------

\1\ The legislative history of section 109 indicates that a
primary standard is to be set at ``the maximum permissible ambient
air level * * * which will protect the health of any [sensitive]
group of the population,'' and that for this purpose ``reference
should be made to a representative sample of persons comprising the
sensitive group rather than to a single person in such a group'' [S.
Rep. No. 91-1196, 91st Cong., 2d Sess. 10 (1970)].
\2\ Welfare effects as defined in section 302(h) (42 U.S.C.
7602(h)) include, but are not limited to, ``effects on soils, water,
crops, vegetation, man-made materials, animals, wildlife, weather,
visibility and climate, damage to and deterioration of property, and
hazards to transportation, as well as effects on economic values and
on personal comfort and well-being.''

---------------------------------------------------------------------------

[[Page 37820]]

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. Lead Industries Association v. EPA, 647 F.2d 1130, 1154 (DC
Cir 1980), cert. denied, 449 U.S. 1042 (1980); American Petroleum
Institute v. Costle, 665 F.2d 1176, 1186 (D.C. Cir. 1981), cert.
denied, 455 U.S. 1034 (1982). Both kinds of uncertainties are
components of the risk associated with pollution at levels below those
at which human health effects can be said to occur with reasonable
scientific certainty. Thus, in selecting primary standards that include
an adequate margin of safety, the Administrator is seeking not only to
prevent pollution levels that have been demonstrated to be harmful but
also to prevent lower pollutant levels that may pose an unacceptable
risk of harm, even if the risk is not precisely identified as to nature
or degree. The CAA does not require the Administrator to establish a
primary NAAQS at a zero-risk level or at background concentration
levels, see Lead Industries Association v. EPA, 647 F.2d at 1156 n. 51,
but rather at a level that reduces risk sufficiently so as to protect
public health with an adequate margin of safety.
In addressing the requirement for an adequate margin of safety, EPA
considers such factors as the nature and severity of the health effects
involved, the size of the 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. Lead
Industries Association v. EPA, 647 F.2d at 1161-62; Whitman v. American
Trucking Associations, 531 U.S. 457, 495 (2001) (Breyer, J., concurring
in part and concurring in judgment).
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. Whitman v. American Trucking
Associations, 531 U.S. 457, 473. In establishing ``requisite'' primary
and secondary standards, EPA may not consider the costs of implementing
the standards. Id. at 471. As discussed by Justice Breyer in Whitman v.
American Trucking Associations, however, ``this interpretation of Sec.
109 does not require the EPA to eliminate every health risk, however
slight, at any economic cost, however great, to the point of
``hurtling'' industry over ``the brink of ruin,'' or even forcing
``deindustrialization.'' Id. at 494 (Breyer J., concurring in part and
concurring in judgment) (citations omitted). Rather, as Justice Breyer
explained:

The statute, by its express terms, does not compel the
elimination of all risk; and it grants the Administrator sufficient
flexibility to avoid setting ambient air quality standards ruinous
to industry.
Section 109(b)(1) directs the Administrator to set standards
that are ``requisite to protect the public health'' with ``an
adequate margin of safety.'' But these words do not describe a world
that is free of all risk--an impossible and undesirable objective.
(citation omitted). Nor are the words ``requisite'' and ``public
health'' to be understood independent of context. We consider
football equipment ``safe'' even if its use entails a level of risk
that would make drinking water ``unsafe'' for consumption. And what
counts as ``requisite'' to protecting the public health will
similarly vary with background circumstances, such as the public's
ordinary tolerance of the particular health risk in the particular
context at issue. The Administrator can consider such background
circumstances when ``deciding what risks are acceptable in the world
in which we live.'' (citation omitted).
The statute also permits the Administrator to take account of
comparative health risks. That is to say, she may consider whether a
proposed rule promotes safety overall. A rule likely to cause more
harm to health than it prevents is not a rule that is ``requisite to
protect the public health.'' For example, as the Court of Appeals
held and the parties do not contest, the Administrator has the
authority to determine to what extent possible health risks stemming
from reductions in tropospheric ozone (which, it is claimed, helps
prevent cataracts and skin cancer) should be taken into account in
setting the ambient air quality standard for ozone. (citation omitted).
The statute ultimately specifies that the standard set must be
``requisite to protect the public health'' ``in the judgment of the
Administrator,'' Sec. 109(b)(1), 84 Stat. 1680 (emphasis added), a
phrase that grants the Administrator considerable discretionary
standard-setting authority.
The statute's words, then, authorize the Administrator to
consider the severity of a pollutant's potential adverse health
effects, the number of those likely to be affected, the distribution
of the adverse effects, and the uncertainties surrounding each
estimate. (citation omitted). They permit the Administrator to take
account of comparative health consequences. They allow her to take
account of context when determining the acceptability of small risks
to health. And they give her considerable discretion when she does so.
This discretion would seem sufficient to avoid the extreme
results that some of the industry parties fear. After all, the EPA,
in setting standards that ``protect the public health'' with ``an
adequate margin of safety,'' retains discretionary authority to
avoid regulating risks that it reasonably concludes are trivial in
context. Nor need regulation lead to deindustrialization.
Preindustrial society was not a very healthy society; hence a
standard demanding the return of the Stone Age would not prove
``requisite to protect the public health.''
Although I rely more heavily than does the Court upon
legislative history and alternative sources of statutory
flexibility, I reach the same ultimate conclusion. Section 109 does
not delegate to the EPA authority to base the national ambient air quality
standards, in whole or in part, upon the economic costs of compliance.

Id. at 494-496.
Section 109(d)(1) of the CAA 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 * *
*.'' This independent review function is performed by the Clean Air
Scientific Advisory Committee (CASAC) of EPA's Science Advisory Board.

B. Related Control Requirements

States have primary responsibility for ensuring attainment and
maintenance of ambient air quality standards once EPA has established
them. Under section 110 of the Act (42 U.S.C. 7410) and related
provisions, States are to submit, for EPA approval, State
implementation plans (SIPs) that provide for the attainment and
maintenance of such standards through control programs directed to
emission sources. The majority of man-made NO_X and VOC
emissions that contribute to O₃ formation in the United
States come from three types of sources: mobile sources, industrial
processes (which include consumer and commercial products), and the
electric

[[Page 37821]]

power industry.\3\ Mobile sources and the electric power industry were
responsible for 78 percent of annual NO_X emissions in 2004.
That same year, 99 percent of man-made VOC emissions came from
industrial processes (including solvents) and mobile sources. Emissions
from natural sources, such as trees, may also comprise a significant
portion of total VOC emissions in certain regions of the country,
especially during the O₃ season, which are considered
natural background emissions.
---------------------------------------------------------------------------

\3\ See EPA report, Evaluating Ozone Control Programs in the
Eastern United States: Focus on the NOX Budget Trading Program, 2004.
---------------------------------------------------------------------------

EPA has developed new emissions standards for many types of
stationary sources and for nearly every class of mobile sources in the
last decade to reduce O₃ by decreasing emissions of
NO_X and VOC. These programs complement State and local
efforts to improve O₃ air quality and meet current national
standards. Under the Federal Motor Vehicle Control Program (FMVCP, see
title II of the CAA, 42 U.S.C. 7521-7574), EPA has established new
emissions standards for nearly every type of automobile, truck, bus,
motorcycle, earth mover, and aircraft engine, and for the fuels used to
power these engines. EPA also established new standards for the smaller
engines used in small watercraft, lawn and garden equipment. Recently
EPA proposed new standards for locomotive and marine diesel engines.
Benefits from engine standards increase modestly each year as older,
more-polluting vehicles and engines are replaced with newer, cleaner
models. In time, these programs will yield substantial emission
reductions. Benefits from fuel programs generally begin as soon as a
new fuel is available.
The reduction of VOC emissions from industrial processes has been
achieved either directly or indirectly through implementation of
control technology standards, including maximum achievable control
technology, reasonably available control technology, and best available
control technology standards; or are anticipated due to proposed or
upcoming proposals based on generally available control technology or
best available controls under provisions related to consumer and
commercial products. These standards have resulted in VOC emission
reductions of almost a million tons per year accumulated starting in
1997 from a variety of sources including combustion sources, coating
categories, and chemical manufacturing. The EPA is currently working to
finalize new federal rules, or amendments to existing rules, that will
establish new nationwide VOC content limits for several categories of
consumer and commercial products, including aerosol coatings,
architectural and industrial maintenance coatings, and household and
institutional commercial products. These rules will take effect in
2009, and will yield significant new reductions in nationwide VOC
emissions--about 200,000 tons per year. Additionally, in O₃
nonattainment areas, we anticipate reductions of an additional 25,000
tons per year following completion of control technique recommendations
for 3 additional consumer and commercial product categories. These
emission reductions primarily result from solvent controls and
typically occur where and when the solvent is used, such as during
manufacturing processes.
The power industry is one of the largest emitters of NO_X
in the United States. Power industry emission sources include large
electric generating units and some large industrial boilers and
turbines. The EPA's landmark Clean Air Interstate Rule (CAIR), issued
on March 10, 2005, permanently caps power industry emissions of
NO_X in the eastern United States. The first phase of the cap
begins in 2009, and a lower second phase cap begins in 2015. By 2015,
EPA projects that the CAIR and other programs in the Eastern U.S. will
reduce power industry O₃ season NO_X emissions in
that region by about 50 percent and annual NO_X emissions by
about 60 percent from 2003 levels.
With respect to agricultural sources, the U.S. Department of
Agriculture (USDA) has approved conservation systems and activities
that reduce agricultural emissions of NO_X and VOC. Current
practices that may reduce emissions of NO_X and VOC include
engine replacement programs, diesel retrofit programs, manipulation of
pesticide applications including timing of applications, and animal
feeding operations waste management techniques. The EPA recognizes that
USDA has been working with the agricultural community to develop
conservation systems and activities to control emissions of
O₃ precursors.
These conservation activities are voluntarily adopted through the
use of incentives provided to the agricultural producer. In cases where
the States need these measures to attain the standard, the measures
could be adopted. The EPA will continue to work with USDA on these
activities with efforts to identify and/or improve the control
efficiencies, prioritize the adoption of these conservation systems and
activities, and ensure that appropriate criteria are used for
identifying the most effective application of conservation systems and
activities.
The EPA will work together with USDA and with States to identify
appropriate measures to meet the primary and secondary standards,
including site-specific conservation systems and activities. Based on
prior experience identifying conservation measures and practices to
meet the PM NAAQS requirements, the EPA will use a similar process to
identify measures that could meet the O₃ requirements. The
EPA anticipates that certain USDA-approved conservation systems and
activities that reduce agricultural emissions of NO_X and VOC
may be able to satisfy the requirements for applicable sources to
implement reasonably available control measures for purposes of
attaining the primary and secondary O₃ NAAQS.

C. Review of Air Quality Criteria and Standards for O3

Tropospheric (ground-level) O₃ is formed from biogenic
and anthropogenic precursor emissions. Naturally occurring
O₃ in the troposphere can result from biogenic organic
precursors reacting with naturally occurring nitrogen oxides
(NO_X) and by stratospheric O₃ intrusion into the
troposphere. Anthropogenic precursors of O₃, specifically
NO_X and volatile organic compounds (VOC), originate from a
wide variety of stationary and mobile sources. Ambient O₃
concentrations produced by these emissions are directly affected by
temperature, solar radiation, wind speed and other meteorological factors.
The last review of the O₃ NAAQS was completed on July
18, 1997, based on the 1996 O₃ CD (U.S. EPA, 1996a) and 1996
O₃ Staff Paper (U.S. EPA, 1996b). EPA revised the primary
and secondary O₃ standards on the basis of the then latest
scientific evidence linking exposures to ambient O₃ to
adverse health and welfare effects at levels allowed by the 1-hour
average standards (62 FR 38856). The O₃ standards were
revised by replacing the existing primary 1-hour average standard with
an 8-hour average O₃ standard set at a level of 0.08 ppm,
which is equivalent to 0.084 ppm using the standard rounding
conventions. The form of the primary standard was changed to the annual
fourth-highest daily maximum 8-hour average concentration, averaged
over three years. The secondary O₃ standard was changed by
making it identical in all respects to the revised primary standard.
Following promulgation of the revised O₃ NAAQS,
petitions for review were

[[Page 37822]]

filed addressing a broad range of issues. In May 1999, in response to
those challenges, the U.S. Court of Appeals for the District of
Columbia Circuit held that EPA's approach to establishing the level of
the standards in 1997, both for the O₃ and for the
particulate matter (PM) NAAQS promulgated on the same day, effected
``an unconstitutional delegation of legislative authority.'' American
Trucking Associations v. EPA, 175 F.3d 1027 (DC Cir., 1999). Although
the D.C. Circuit stated that ``factors EPA uses in determining the
degree of public health concern associated with different levels of
O₃ and PM are reasonable,'' it remanded the rule to EPA,
stating that when EPA considers these factors for potential non-
threshold pollutants ``what EPA lacks is any determinate criterion for
drawing lines'' to determine where the standards should be set. Id. at
1034. Consistent with EPA's long-standing interpretation and DC Circuit
precedent, the court also reaffirmed prior rulings holding that in
setting the NAAQS, it is ``not permitted to consider the cost of
implementing those standards.'' Id. at 1040-41. The DC Circuit further
directed EPA to consider on remand the potential indirect beneficial
health effects of O₃ pollution in shielding the public from
the effects of solar ultraviolet (UV) radiation, as well as the direct
adverse health effects of O₃ pollution.
Both sides filed cross appeals on the constitutional and cost
issues to the United States Supreme Court, and the Court granted
certiorari. On February 27, 2001, the U.S. Supreme Court issued a
unanimous decision upholding the EPA's position on both the
constitutional and the cost issues. Whitman v. American Trucking
Associations, 531 U.S. at 464, 475-76. On the constitutional issue, the
Court held that the statutory requirement that NAAQS be ``requisite''
to protect public health with an adequate margin of safety sufficiently
guided EPA's discretion, affirming EPA's approach of setting standards
that are neither more nor less stringent than necessary. The Supreme
Court remanded the case to the D.C. Circuit for resolution of any
remaining issues that had not been addressed by that Court's earlier
decisions. Id. at 475-76. On March 26, 2002, the D.C. Circuit Court
rejected all remaining challenges to the NAAQS, holding under
traditional standard of review that EPA ``engaged in reasoned decision-
making'' in setting the 1997 O₃ NAAQS. Whitman v. American
Trucking Associations, 283 F.3d 355 (DC Cir. 2002).
In response to the DC Circuit Court's remand to consider the
potential indirect beneficial health effects of O₃ in
shielding the public from the effects of solar (UV) radiation, on
November 14, 2001, EPA proposed to leave the 1997 8-hour NAAQS
unchanged (66 FR 57267). After considering public comment on the
proposed decision, EPA reaffirmed the 8-hour O₃ NAAQS set in
1997 (68 FR 614). Finally, on April 30, 2004, EPA issued an 8-hour
implementation rule that, among other things, provided that the 1-hour
O₃ NAAQS would no longer apply to areas one year after the
effective date of the designation of those areas for the 8-hour NAAQS
(69 FR 23966).\4\ For most areas, the date that the 1-hour NAAQS no
longer applied was June 15, 2005. (See 40 CFR 50.9 for details.)
---------------------------------------------------------------------------

\4\ On December 22, 2006, the D.C. Circuit vacated the April 30,
2004 implementation rule. South Coast Air Quality Management
District v. EPA, 472 F.3d 882. In March 2007, EPA requested the
Court to reconsider its decision.
---------------------------------------------------------------------------

The EPA initiated this current review in September 2000 with a call
for information (65 FR 57810) for the development of a revised Air
Quality Criteria Document for O₃ and Other Photochemical
Oxidants (henceforth the ``Criteria Document''). A project work plan
(U.S. EPA, 2002) for the preparation of the Criteria Document was
released in November 2002 for CASAC and public review. EPA held a
series of workshops in mid-2003 on several draft chapters of the
Criteria Document to obtain broad input from the relevant scientific
communities. These workshops helped to inform the preparation of the
first draft Criteria Document (EPA, 2005a), which was released for
CASAC and public review on January 31, 2005; a CASAC meeting was held
on May 4-5, 2005 to review the first draft Criteria Document. A second
draft Criteria Document (EPA, 2005b) was released for CASAC and public
review on August 31, 2005, and was discussed along with a first draft
Staff Paper (EPA, 2005c) at a CASAC meeting held on December 6-8, 2005.
In a February 16, 2006 letter to the Administrator, the CASAC offered
final comments on all chapters of the Criteria Document (Henderson,
2006a), and the final Criteria Document (EPA, 2006a) was released on
March 21, 2006. In a June 8, 2006 letter (Henderson, 2006b) to the
Administrator, the CASAC offered additional advice to the Agency
concerning chapter 8 of the final Criteria Document (Integrative
Synthesis) to help inform the second draft Staff Paper.
A second draft Staff Paper (EPA, 2006b) was released on July 17,
2006 and reviewed by CASAC on August 24 and 25, 2006. In an October 24,
2006 letter to the Administrator, CASAC provided advice and
recommendations to the Agency concerning the second draft Staff Paper
(Henderson, 2006c). A final Staff Paper (EPA, 2007) was released on
January 31, 2007. Around the time of the release of the final Staff
Paper in January 2007, EPA discovered a small error in the exposure
model that when corrected resulted in slight increases in the human
exposure estimates. Since the exposure estimates are an input to the
lung function portion of the health risk assessment, this correction
also resulted in slight increases in the lung function risk estimates
as well. The exposure and risk estimates discussed in this notice
reflect the corrected estimates, and thus are slightly different than
the exposure and risk estimates cited in the January 31, 2007 Staff
Paper.\5\ In a March 26, 2007 letter (Henderson, 2007), CASAC offered
additional advice to the Administrator with regard to recommendations
and revisions to the primary and secondary O₃ NAAQS.
---------------------------------------------------------------------------

\5\ EPA plans to make available corrected versions of the final
Staff Paper and the human exposure and health risk assessment
technical support documents on or around July 16, 2007 on the EPA
web site listed in the Availability of Related Information section
of this notice.
---------------------------------------------------------------------------

The schedule for completion of this review is governed by a consent
decree resolving a lawsuit filed in March 2003 by a group of plaintiffs
representing national environmental and public health organizations,
alleging that EPA had failed to complete the current review within the
period provided by statute.\6\ The modified consent decree that governs
this review, entered by the court on December 16, 2004, provides that
EPA sign for publication notices of proposed and final rulemaking
concerning its review of the O₃ NAAQS no later than March
28, 2007 and December 19, 2007, respectively. This consent decree was
further modified in October 2006 to change these proposed and final
rulemaking dates to no later than May 30, 2007 and February 20, 2008,
respectively. These dates for signing the publication notices of
proposed and final rulemaking were further extended to no later than
June 20, 2007 and March 12, 2008, respectively.
---------------------------------------------------------------------------

\6\ American Lung Association v. Whitman (No. 1:03CV00778,
D.D.C. 2003).
---------------------------------------------------------------------------

This action presents the Administrator's proposed decisions on the
review of the current primary and secondary O₃ standards.
Throughout this preamble a number of conclusions, findings, and
determinations proposed by the Administrator are noted. While

[[Page 37823]]

they identify the reasoning that supports this proposal, they are not
intended to be final or conclusive in nature. The EPA invites general,
specific, and/or technical comments on all issues involved with this
proposal, including all such proposed judgments, conclusions, findings,
and determinations.

II. Rationale for Proposed Decision on the Primary Standard

This section presents the rationale for the Administrator's
proposed decision to revise the existing 8-hour O₃ primary
standard by lowering the level of the standard to within a range from
0.070 to 0.075 ppm, and to specify the standard to the nearest
thousandth ppm (i.e., to the nearest parts per billion). As discussed
more fully below, this rationale is based on a thorough review, in the
Criteria Document, of the latest scientific information on human health
effects associated with the presence of O₃ in the ambient
air. This rationale also takes into account and is consistent with: (1)
Staff assessments of the most policy-relevant information in the
Criteria Document and staff analyses of air quality, human exposure,
and health risks, presented in the Staff Paper, upon which staff
recommendations for revisions to the primary O₃ standard are
based; (2) CASAC advice and recommendations, as reflected in
discussions of drafts of the Criteria Document and Staff Paper at
public meetings, in separate written comments, and in CASAC's letters
to the Administrator; and (3) public comments received during the
development of these documents, either in connection with CASAC
meetings or separately.
In developing this rationale, EPA has drawn upon an integrative
synthesis of the entire body of evidence, published through early 2006,
on human health effects associated with the presence of O₃
in the ambient air. As discussed below in section II.A, this body of
evidence addresses a broad range of health endpoints associated with
exposure to ambient levels of O₃ (EPA, 2006a, chapter 8),
and includes over one hundred epidemiologic studies conducted in the
U.S., Canada, and many countries around the world.\7\ In considering
this evidence, EPA focuses on those health endpoints that have been
demonstrated to be caused by exposure to O₃, or for which
the Criteria Document judges associations with O₃ to be
causal, likely causal, or for which the evidence is highly suggestive
that O₃ contributes to the reported effects. This rationale
also draws upon the results of quantitative exposure and risk
assessments, discussed below in section II.B. Evidence- and exposure/
risk-based considerations that form the basis for the Administrator's
proposed decisions on the adequacy of the current standard and on the
elements of the range of proposed alternative standards are discussed
below in sections II.C and II.D, respectively.
---------------------------------------------------------------------------

\7\ In its assessment of the epidemiological evidence judged to
be most relevant to making decisions on the level of the
O₃ primary standard, EPA has placed greater weight on
U.S. and Canadian epidemiologic studies, since studies conducted in
other countries may well reflect different demographic and air
pollution characteristics.
---------------------------------------------------------------------------

Judgments made in the Criteria Document and Staff Paper about the
extent to which relationships between various health endpoints and
short-term exposures to ambient O₃ are likely causal have
been informed by several factors. As discussed below in section II.A,
these factors include the nature of the evidence (i.e., controlled
human exposure, epidemiological, and/or toxicological studies) and the
weight of evidence, which takes into account such considerations as
biological plausibility, coherence of evidence, strength of
association, and consistency of evidence.
In assessing the health effects data base for O₃, it is
clear that human studies provide the most directly applicable
information for determining causality because they are not limited by
the uncertainties of dosimetry differences and species sensitivity
differences, which would need to be addressed in extrapolating animal
toxicology data to human health effects. Controlled human exposure
studies provide data with the highest level of confidence since they
provide human effects data under closely monitored conditions and can
provide exposure-response relationships. Epidemiological data provide
evidence of associations between ambient O₃ levels and more
serious acute and chronic health effects (e.g., hospital admissions and
mortality) that cannot be assessed in controlled human exposure
studies. For these studies the degree of uncertainty introduced by
confounding variables (e.g., other pollutants, temperature) and other
factors affects the level of confidence that the health effects being
investigated are attributable to O₃ exposures, alone and in
combination with other copollutants.
In using a weight of evidence approach to inform judgments about
the degree of confidence that various health effects are likely to be
caused by exposure to O₃, confidence increases as the number
of studies consistently reporting a particular health endpoint grows
and as other factors, such as biological plausibility and strength,
consistency, and coherence of evidence, increase. Conclusions regarding
biological plausibility, consistency, and coherence of evidence of
O₃-related health effects are drawn from the integration of
epidemiological studies with mechanistic information from controlled
human exposure studies and animal toxicological studies. As discussed
below, this type of mechanistic linkage has been firmly established for
several respiratory endpoints (e.g., lung function decrements, lung
inflammation) but remains far more equivocal for cardiovascular
endpoints (e.g., cardiovascular-related hospital admissions). For
epidemiological studies, strength of association refers to the
magnitude of the association and its statistical strength, which
includes assessment of both effects estimate size and precision. In
general, when associations yield large relative risk estimates, it is
less likely that the association could be completely accounted for by a
potential confounder or some other bias. Consistency refers to the
persistent finding of an association between exposure and outcome in
multiple studies of adequate power in different persons, places,
circumstances and times. For example, the magnitude of effect estimates
is relatively consistent across recent studies showing association
between short-term, but not long-term, O₃ exposure and mortality.
Based on the information discussed below in sections II.A.1-II.A.3,
judgments concerning the extent to which relationships between various
health endpoints and ambient O₃ exposures are likely causal
are summarized below in section II.A.3.c. These judgments reflect the
nature of the evidence and the overall weight of the evidence, and are
taken into consideration in the quantitative exposure and risk
assessments, discussed below in Section II.B.
To put judgments about health effects that have been demonstrated
to be caused by exposure to O₃, or for which the Criteria
Document judges associations with O₃ to be causal, likely
causal, or for which the evidence is highly suggestive that
O₃ contributes to the reported effects into a broader public
health context, EPA has drawn upon the results of the quantitative
exposure and risk assessments. These assessments provide estimates of
the likelihood that individuals in particular population groups that
are at risk for various O₃-related physiological health effects
would experience ``exposures of concern'' and specific health endpoints

[[Page 37824]]

under varying air quality scenarios (e.g., just meeting the current or
alternative standards), as well as characterizations of the kind and
degree of uncertainties inherent in such estimates.
In this review, the term ``exposures of concern'' is defined as
personal exposures while at moderate or greater exertion to 8-hour
average ambient O₃ levels at and above specific benchmark
levels which represent exposure levels at which O₃-related
health effects are known or can reasonably be inferred to occur in some
individuals, as discussed below in section II.B.1.\8\ EPA emphasizes
that although the analysis of ``exposures of concern'' was conducted
using three discrete benchmark levels (i.e., 0.080, 0.070, and 0.060
ppm), the concept is more appropriately viewed as a continuum with
greater confidence and less uncertainty about the existence of health
effects at the upper end and less confidence and greater uncertainty as
one considers increasingly lower O₃ exposure levels. EPA
recognizes that there is no sharp breakpoint within the continuum
ranging from at and above 0.080 ppm down to 0.060 ppm. In considering
the concept of exposures of concern, it is important to balance
concerns about the potential for health effects and their severity with
the increasing uncertainty associated with our understanding of the
likelihood of such effects at lower O₃ levels.
---------------------------------------------------------------------------

\8\ Exposures of concern were also considered in the last review
of the O₃ NAAQS, and were judged by EPA to be an
important indicator of the public health impacts of those
O₃-related effects for which information was too limited
to develop quantitative estimates of risk but which had been
observed in humans at and above the benchmark level of 0.08 ppm for
6-to 8-hour exposures * * * including increased nonspecific
bronchial responsiveness (for example, aggravation of asthma),
decreased pulmonary defense mechanisms (suggestive of increased
susceptibility to respiratory infection), and indicators of
pulmonary inflammation (related to potential aggravation of chronic
bronchitis or long-term damage to the lungs). (62 FR 38868)
---------------------------------------------------------------------------

Within the context of this continuum, estimates of exposures of
concern at discrete benchmark levels provide some perspective on the
public health impacts of O₃-related health effects that have
been demonstrated in human clinical and toxicological studies but
cannot be evaluated in quantitative risk assessments, such as lung
inflammation, increased airway responsiveness, and changes in host
defenses. They also help in understanding the extent to which such
impacts have the potential to be reduced by meeting the current and
alternative standards. These O₃-related physiological
effects are plausibly linked to the increased morbidity seen in
epidemiological studies (e.g., as indicated by increased medication use
in asthmatics, school absences in all children, and emergency
department visits and hospital admissions in people with lung disease).
Estimates of the number of people likely to experience exposures of
concern cannot be directly translated into quantitative estimates of
the number of people likely to experience specific health effects,
since sufficient information to draw such comparisons is not
available--if such information were available, these health outcomes
would have been included in the quantitative risk assessment. Due to
individual variability in responsiveness, only a subset of individuals
who have exposures at and above a specific benchmark level can be
expected to experience such adverse health effects, and susceptible
subpopulations such as those with asthma are expected to be affected
more by such exposures than healthy individuals. The amount of weight
to place on the estimates of exposures of concern at any of these
benchmark levels depends in part on the weight of the scientific
evidence concerning health effects associated with O₃
exposures at and above that benchmark level. It also depends on
judgments about the importance from a public health perspective of the
health effects that are known or can reasonably be inferred to occur as
a result of exposures at and above the benchmark level. Such public
health policy judgments are embodied in the NAAQS standard setting
criteria (i.e., standards that, in the judgment of the Administrator,
are requisite to protect public health with an adequate margin of safety).
As discussed below in section II.B.2, the quantitative health risk
assessment conducted as part of this review includes estimates of risks
of lung function decrements in asthmatic and all school age children,
respiratory symptoms in asthmatic children, respiratory-related
hospital admissions, and non-accidental and cardiorespiratory-related
mortality associated with recent ambient O₃ levels, as well
as risk reductions and remaining risks associated with just meeting the
current and various alternative O₃ standards in a number of
example urban areas. There were two parts to this risk assessment: one
part was based on combining information from controlled human exposure
studies with modeled population exposure, and the other part was based
on combining information from community epidemiological studies with
either monitored or adjusted ambient concentrations levels. This
assessment not only provided estimates of the potential magnitude of
O₃-related health effects, as well as a characterization of
the uncertainties and variability inherent in such estimates. This
assessment also provided insights into the distribution of risks and
patterns of risk reductions associated with meeting alternative
O₃ standards.
As discussed below, a substantial amount of new research has been
conducted since the last review of the O₃ NAAQS, with
important new information coming from epidemiologic studies as well as
from controlled human exposure, toxicological, and dosimetric studies.
The newly available research studies evaluated in the Criteria Document
and the exposure and risk assessments presented in the Staff Paper have
undergone intensive scrutiny through multiple layers of peer review and
many opportunities for public review and comment. While important
uncertainties remain in the qualitative and quantitative
characterizations of health effects attributable to exposure to ambient
O₃, the review of this information has been extensive and
deliberate. In the judgment of the Administrator, this intensive
evaluation of the scientific evidence has provided an adequate basis
for regulatory decision making. This review also provides important
input to EPA's research plan for improving our future understanding of
the effects of ambient O₃ at lower levels, especially in at-
risk population groups.

A. Health Effects Information

This section outlines key information contained in the Criteria
Document (chapters 4-8) and in the Staff Paper (chapter 3) on known or
potential effects on public health which may be expected from the
presence of O₃ in ambient air. The information highlighted
here summarizes: (1) New information available on potential mechanisms
for health effects associated with exposure to O₃; (2) the
nature of effects that have been associated directly with exposure to
O₃ and indirectly with the presence of O₃ in
ambient air; (3) an integrative interpretation of the evidence,
focusing on the biological plausibility and coherence of the evidence;
and (4) considerations in characterizing the public health impact of
O₃, including the identification of ``at risk'' subpopulations.
The decision in the last review focused primarily on evidence from
short-term (e.g., 1 to 3 hours) and prolonged ( 6 to 8 hours)
controlled-exposure studies reporting lung function decrements,
respiratory symptoms, and respiratory inflammation in humans, as well
as epidemiology studies reporting excess

[[Page 37825]]

hospital admissions and emergency department (ED) visits for
respiratory causes. The Criteria Document prepared for this review
emphasizes a large number of epidemiological studies published since
the last review with these and additional health endpoints, including
the effects of acute (short-term and prolonged) and chronic exposures
to O₃ on lung function decrements and enhanced respiratory
symptoms in asthmatic individuals, school absences, and premature
mortality. It also emphasizes important new information from
toxicology, dosimetry, and controlled human exposure studies.
Highlights of the evidence include:
(1) Two new controlled human-exposure studies are now available
that examine respiratory effects associated with prolonged
O₃ exposures at levels below 0.080 ppm, which was the lowest
exposure level that had been examined in the last review.
(2) Numerous controlled human-exposure studies have examined
indicators of O₃-induced inflammatory response in both the
upper respiratory tract (URT) and lower respiratory tract (LRT), while
other studies have examined changes in host defense capability
following O₃ exposure of healthy young adults and increased
airway responsiveness to allergens in subjects with allergic asthma and
allergic rhinitis exposed to O₃.
(3) Animal toxicology studies provide new information regarding
mechanisms of action, increased susceptibility to respiratory
infection, and the biological plausibility of acute effects and
chronic, irreversible respiratory damage.
(4) Numerous acute exposure epidemiological studies published
during the past decade offer added evidence of ambient O₃-
related lung function decrements and respiratory symptoms in physically
active healthy subjects and asthmatic subjects, as well as evidence on
new health endpoints, such as the relationships between ambient
O₃ concentrations and school absenteeism and between ambient
O₃ and cardiac-related physiological endpoints.
(5) Several additional studies have been published over the last
decade examining the temporal associations between O₃
exposures and emergency department visits for respiratory diseases and
on respiratory-related hospital admissions.
(6) A large number of newly available epidemiological studies have
examined the effects of acute exposure to PM and O₃ on
mortality, notably including large multicity studies that provide much
more robust and credible information than was available in the last
review, as well as recent meta-analyses that have evaluated potential
sources of heterogeneity in O₃-mortality associations.
1. Overview of Mechanisms
Evidence on possible mechanisms by which exposure to O₃
may result in acute and chronic health effects is discussed in chapters
5 and 6 of the Criteria Document.\9\ Evidence from dosimetry,
toxicology, and human exposure studies has contributed to an
understanding of the mechanisms that help to explain the biological
plausibility and coherence of evidence for O₃-induced
respiratory health effects reported in epidemiological studies. More
detailed information about the physiological mechanisms related to the
respiratory effects of short- and long-term exposure to O₃
can be found in section II.A.3.b.i and II.A.3.b.iii, respectively. In
the past, however, little information was available to help explain
potential biological mechanisms which linked O₃ exposure to
premature mortality or cardiovascular effects. As discussed more fully
in section II.A.3.b.ii below, since the last review an emerging body of
animal toxicology and human clinical evidence is beginning to suggest
mechanisms that may mediate acute O₃ cardiovascular effects.
While much is known about mechanisms that play a role in O₃-
related respiratory effects, additional research is needed to more
clearly understand the role that O₃ may have in contributing
to cardiovascular effects.
---------------------------------------------------------------------------

\9\ While most of the available evidence addresses mechanisms
for O₃, O₃ clearly serves as an indicator for
the total photochemical oxidant mixture found in the ambient air.
Some effects may be caused by one or more components in the overall
pollutant mix, either separately or in combination with O₃.
---------------------------------------------------------------------------

With regard to the mechanisms related to short-term respiratory
effects, scientific evidence discussed in the Criteria Document
(section 5.2) indicates that reactions of O₃ with lipids and
antioxidants in the epithelial lining fluid and the epithelial cell
membranes of the lung can be the initial step in mediating deleterious
health effects of O₃. This initial step activates a cascade
of events that lead to oxidative stress, injury, inflammation, airway
epithelial damage and increased alveolar permeability to vascular
fluids. Inflammation can be accompanied by increased airway
responsiveness, which is an increased bronchoconstrictive response to
airway irritants and allergens. Continued respiratory inflammation also
can alter the ability to respond to infectious agents, allergens and
toxins. Acute inflammatory responses to O₃ in some healthy
people are well documented, and precursors to lung injury can become
apparent within 3 hours after exposure in humans. Repeated respiratory
inflammation can lead to a chronic inflammatory state with altered lung
structure and lung function and may lead to chronic respiratory
diseases such as fibrosis and emphysema (EPA, 2006a, section 8.6.2).
The severity of symptoms and magnitude of response to acute exposures
depend on inhaled dose, as well as individual susceptibility to
O₃, as discussed below. At the same O₃ dose,
individuals who are more susceptible to O₃ will have a
larger response than those who are less susceptible; among individuals
with similar susceptibility, those who receive a larger dose will have
a larger response to O₃.
The inhaled dose is the product of O₃ concentration (C),
minute ventilation or ventilation rate, and duration of exposure (T),
or (C x ventilation rate x T). A large body of data regarding the
interdependent effect of these components of inhaled dose on pulmonary
responses was assessed in the 1986 and 1996 O₃ Criteria
Documents. In an attempt to describe O₃ dose-response
characteristics, acute responses were modeled as a function of total
inhaled O₃ dose which was generally found to be a better
predictor of response than O₃ concentration, ventilation
rate, or duration of exposure, alone, or as a combination of any two of
these factors (EPA 2006a, section 6.2). Predicted O₃-induced
decrements in lung function have been shown to be a function of
exposure concentration, duration and exercise level for healthy, young
adults (McDonnell et al., 1997). A meta-analysis of 21 studies (Mudway
and Kelly, 2004) showed that markers of inflammation and increased
cellular permeability in healthy subjects are associated with total
O₃ dose.
The Criteria Document summarizes information on potentially
susceptible and vulnerable groups in section 8.7. As described there,
the term susceptibility refers to innate (e.g., genetic or
developmental) or acquired (e.g., personal risk factors, age) factors
that make individuals more likely to experience effects with exposure
to pollutants. A number of population groups have been identified as
potentially susceptible to health effects as a result of O₃
exposure, including people with existing lung diseases, including
asthma, children and older adults, and people who have larger than
normal lung function responses that may be due to genetic
susceptibility. In addition, some population groups have been
identified as having increased

[[Page 37826]]

vulnerability to O₃-related effects due to increased
likelihood of exposure while at elevated ventilation rates, including
healthy children and adults who are active outdoors, for example,
outdoor workers, and joggers. Taken together, the susceptible and
vulnerable groups are more commonly referred to as ``at-risk'' groups
\10\, as discussed more fully below in section II.A.4.b.
---------------------------------------------------------------------------

\10\ In previous Staff Papers and Federal Register notices
announcing proposed and final decisions on the O₃ and
other NAAQS, EPA has used the phrase ``sensitive population groups''
to include both population groups that are at increased risk because
they are more susceptible and population groups that are at
increased risk due to increased vulnerability or exposure. In this
notice, we use the phrase, ``at risk'' populations to include both
types of population groups.
---------------------------------------------------------------------------

Based on new evidence from animal, human clinical and
epidemiological studies the Criteria Document concludes that people
with preexisting pulmonary disease are likely to be among those at
increased risk from O₃ exposure. Altered physiological,
morphological and biochemical states typical of respiratory diseases
like asthma, COPD and chronic bronchitis may render people sensitive to
additional oxidative burden induced by O₃ exposure (EPA
2006a, section 8.7). Children and adults with asthma are the group that
has been studied most extensively. Evidence from controlled human
exposure studies indicates that asthmatics may exhibit larger lung
function decrements in response to O₃ exposure than healthy
controls. As discussed more fully in section II.A.4.b.ii below,
asthmatics present a differential response profile for cellular,
molecular, and biochemical parameters (CD, Figure 8-1) that are altered
in response to acute O₃ exposure. They can have larger
inflammatory responses, as manifested by larger increases in markers of
inflammation such as white bloods cells (e.g., PMNs) or inflammatory
cytokines. Asthmatics, and people with allergic rhinitis, are more
likely to mount an allergic-type response upon exposure to
O₃, as manifested by increases in white blood cells
associated with allergy (i.e., eosinophils) and related molecules,
which increase inflammation in the airways. The increased inflammatory
and allergic responses also may be associated with the larger late-
phase responses that asthmatics can experience, which can include
increased bronchoconstrictor responses to irritant substances or
allergens and additional inflammation. These more serious responses in
asthmatics and others with lung disease provide biological plausibility
for the respiratory morbidity effects observed in epidemiological studies.
Children with and without asthma were found to be particularly
susceptible to O₃ effects on lung function and generally
have greater lung function responses than older people. The American
Academy of Pediatrics (2004) notes that children and infants are among
the population groups most susceptible to many air pollutants,
including O₃. This is in part because their lungs are still
developing. For example, eighty percent of alveoli are formed after
birth, and changes in lung development continue through adolescence
(Dietert et al., 2000). Moreover, children have high minute ventilation
rates and relatively high levels of physical activity which also
increases their O₃ dose (Plunkett et al., 1992). Thus,
children are at risk due to both their susceptibility and vulnerability.
Looking more broadly at age-related differences in susceptibility,
several mortality studies have investigated age-related differences in
O₃ effects (EPA, 2006a, section 7.6.7.2), primarily in the
older adult population. Among the studies that observed positive
associations between O₃ and mortality, a comparison of all
age or younger age (65 years of age) O₃-mortality effect
estimates to that of the elderly population (>65 years) indicates that,
in general, the elderly population is more susceptible to O₃
mortality effects. There is supporting evidence of age-related
differences in susceptibility to O₃ lung function effects.
The Criteria Document concludes that the elderly population (>65 years
of age) appears to be at greater risk of O₃-related
mortality and hospitalizations compared to all ages or younger
populations, and children (<18 years of age) experience other
potentially adverse respiratory health outcomes with increased
O₃ exposure (EPA, 2006a, section 7.6.7.2).
Controlled human exposure studies have also indicated a high degree
of interindividual variability in some of the pulmonary physiological
parameters, such as lung function decrements. The variable effects in
individuals have been found to be reproducible, in other words, a
person who has a large lung function response after exposure to
O₃ will likely have about the same response if exposed again
to the same dose of O₃ (EPA 2006a, p. 6-2). In human
clinical studies, group mean responses are not representative of this
segment of the population that has much larger than average responses
to O₃. Recent studies, discussed in section II.A.4.iv below,
reported a role for genetic polymorphism (i.e., the occurrence together
in the same population of more than one allele or genetic marker at the
same locus with the least frequent allele or marker occurring more
frequently than can be accounted for by mutation alone) in observed
differences in antioxidant enzymes and genes involved in inflammation
to modulate pulmonary function and inflammatory responses to
O₃ exposure. These observations suggest a potential role for
these markers in the innate susceptibility to O₃, however,
the validity of these markers and their relevance in the context of
prediction to population studies needs additional experimentation.
Clinical studies that provide information about mechanisms of the
initial response to O₃ (e.g., lung function decrements,
inflammation, and injury to the lung) also inform the selection of
appropriate lag times to analyze in epidemiological studies through
elucidation of the time course of these responses (EPA 2006a, section
8.4.3). Based on the results of these studies, it would be reasonable
to expect that lung function decrements could be detected
epidemiologically within lags of 0 (same day) or 1 to 2 days following
O₃ exposure, given the rapid onset of lung function changes
and their persistence for 24 to 48 hours among more responsive human
subjects in clinical studies. Other responses take longer to develop
and can persist for longer periods of time. For example, although
asthmatic individuals may begin to experience symptoms soon after
O₃ exposure, it may take anywhere from 1 to 3 days after
exposure for these subjects to seek medical attention as a result of
increased airway responsiveness or inflammation that may persist for 2
to 3 days. This may be reflected by epidemiologic observations of
significantly increased risk for asthma-related emergency department
visits or hospital admissions with 1- to 3-day lags, or, perhaps,
enhanced distributed lag risks (combined across 3 days) for such
morbidity indicators. Analogously, one might project increased
mortality within 0 to 3 day lags as a possible consequence of
O₃-induced increases in clotting agents arising from the
cascade of events, starting with cell injury described above, occurring
within 12 to 24 hours of O₃ exposure. The time course for
many of these initial responses to O₃ is highly variable.
Moreover these observations pertain only to the initial response to
O₃. Consequent responses can follow. For example,
J[ouml]rres et al., (1996) found that in subjects with

[[Page 37827]]

asthma and allergic rhinitis, a maximum percent fall in FEV₁
of 27.9% and 7.8%, respectively, occurred 3 days after O₃
exposure when they were challenged with the highest common dose of
allergen.
2. Nature of Effects
The Criteria Document provides new evidence that notably enhances
our understanding of short-term and prolonged exposure effects,
including effects on lung function, symptoms, and inflammatory effects
reported in controlled exposure studies. These studies support and
extend the findings of the previous Criteria Document. There is also a
significant body of new epidemiological evidence of associations
between short-term and prolonged exposure to O₃ and effects
such as premature mortality, hospital admissions and emergency
department visits for respiratory (e.g., asthma) causes. Key
epidemiological and controlled human exposure studies are summarized
below and discussed in chapter 3 of the Staff Paper, which is based on
scientific evidence critically reviewed in chapters 5, 6, and 7 of the
Criteria Document, as well as the Criteria Document's integration of
scientific evidence contained in chapter 8.\11\ Conclusions drawn about
O₃-related health effects are based upon the full body of
evidence from controlled human exposure, epidemiological and
toxicological data contained in the Criteria Document.
---------------------------------------------------------------------------

\11\ Health effects discussions are also drawn from the more detailed
information and tables presented in the Criteria Document's annexes.
---------------------------------------------------------------------------

a. Morbidity
This section summarizes scientific information on the effects of
inhalation of O₃, including public health effects of short-
term, prolonged, and long-term exposures on respiratory morbidity and
cardiovascular system effects, as discussed in chapters 6, 7 and 8 of
the Criteria Document and chapter 3 of the Staff Paper. This section
also summarizes the uncertainty about the potential indirect effects on
public health associated with changes due to increases in UV-B
radiation exposure, such as UV-B radiation-related skin cancers, that
may be associated with reductions in ambient levels of ground-level
O₃, as discussed in chapter 10 of the Criteria Document and
chapter 3 of the Staff Paper.
i. Effects on the Respiratory System From Short-Term and Prolonged
O₃ Exposures
Controlled human exposure studies have shown that O₃
induces a variety of health effects, including: lung function
decrements, respiratory symptoms, increased airway responsiveness,
respiratory inflammation and permeability, increased susceptibility to
respiratory infection, and acute morphological effects. Epidemiology
studies have reported associations between O₃ exposures
(i.e., 1-hour, 8-hour and 24-hour) and a wide range of respiratory-
related health effects including: Pulmonary function decrements;
respiratory symptoms; increased asthma medication use; increased school
absences; increased emergency department visits and hospital admissions.
(a) Pulmonary Function Decrements, Respiratory Symptoms, and Asthma
Medication Use
(i) Results From Controlled Human Exposure Studies
A large number of studies published prior to 1996 that investigated
short-term O₃ exposure health effects on the respiratory
system from short-term O₃ exposures were reviewed in the
1986 and 1996 Criteria Documents (EPA, 1986, 1996). In the last review,
0.50 ppm was the lowest O₃ concentration at which
statistically significant reductions in forced vital capacity (FVC) and
forced expiratory volume in 1 second (FEV₁) were reported in
sedentary subjects. During exercise, spirometric (lung function) and
symptomatic responses were observed at much lower O₃
exposures. When minute ventilation was considerably increased by
continuous exercise (CE) during O₃ exposures lasting 2 hour
or less at >= 0.12 ppm, healthy subjects generally experienced
decreases in FEV₁, FVC, and other measures of lung function;
increases in specific airway resistance (sRaw), breathing frequency,
and airway responsiveness; and symptoms such as cough, pain on deep
inspiration, shortness of breath, throat irritation, and wheezing. When
exposures were increased to 4 to 8 hours in duration, statistically
significant lung function and symptom responses were reported at
O₃ concentrations as low as 0.08 ppm and at lower minute
ventilation (i.e., moderate rather than high level exercise) than the
shorter duration studies.
The most important observations drawn from studies reviewed in the
1996 Criteria Document were that: (1) Young healthy adults exposed to
O₃ concentrations >= 0.080 ppm develop significant,
reversible, transient decrements in pulmonary function if minute
ventilation or duration of exposure is increased sufficiently; (2)
children experience similar lung function responses but report lesser
symptoms from O₃ exposure relative to young adults; (3)
O₃-induced lung function responses are decreased in the
elderly relative to young adults; (4) there is a large degree of
intersubject variability in physiological and symptomatic responses to
O₃, but responses tend to be reproducible within a given
individual over a period of several months; (5) subjects exposed
repeatedly to O₃ for several days show an attenuation of
response upon successive exposures, but this attenuation is lost after
about a week without exposure; and (6) acute O₃ exposure
initiates an inflammatory response which may persist for at least 18 to
24 hours post exposure.
The development of these respiratory effects is time-dependent
during both exposure and recovery periods, with great overlap for
development and disappearance of the effects. In healthy human subjects
exposed to typical ambient O₃ levels near 0.120 ppm, lung
function responses largely resolve within 4 to 6 hours post-exposure,
but cellular effects persist for about 24 hours. In these healthy
subjects, small residual lung function effects are almost completely
gone within 24 hours, while in hyperresponsive subjects, recovery can
take as much as 48 hours to return to baseline. The majority of these
responses are attenuated after repeated consecutive exposures, but such
attenuation to O₃ is lost one week post-exposure.
Since 1996, there have been a number of studies published
investigating lung function and symptomatic responses that generally
support the observations previously drawn. Recent studies for acute
exposures of 1 to 2 hours and 6 to 8 hours in duration are compiled in
the Staff Paper (Appendix 3C). As summarized in more detail in the
Staff Paper (section 3.3.1.1), among the more important of the recent
studies that examined changes in FEV₁ in large numbers of
subjects over a range of 1-2 hours at exposure levels of 0.080 to 0.40
ppm were studies by McDonnell et al. (1997) and Ultman et al. (2004).
These studies observed considerable intersubject variability in
FEV₁ decrements, which was consistent with findings in the
1996 Criteria Document.
For prolonged exposures (4 to 8 hours) in the range of 0.080 to
0.160 ppm O₃ using moderate intermittent exercise and
typically using square-

[[Page 37828]]

wave exposure patterns (i.e., a constant exposure level during time of
exposure), several pre- and post-1996 studies (Folinsbee et al., 1988,
1994; Horstman et al., 1990; Adams, 2002, 2003a, 2006) have reported
statistically significant lung function responses and increased
symptoms in healthy adults with increasing duration of exposure,
O₃ concentration, and minute ventilation. Studies that
employed triangular exposure patterns (i.e., integrated exposures that
begin at a low level, rise to a peak, and return to a low level during
the exposure) (Hazucha et al., 1992; Adams 2003a, 2006) suggest that
the triangular exposure pattern can potentially lead to greater
FEV₁ decrements and respiratory symptoms than square-wave
exposures (when the overall O₃ doses are equal). These
results suggest that peak exposures, reflective of the pattern of
ambient O₃ concentrations in some locations, are important
in terms of O₃ toxicology.
McDonnell (1996) used data from a series of studies to investigate
the frequency distributions of FEV₁ decrements following 6.6
hour exposures and found statistically significant but relatively small
group mean decreases in average FEV₁ responses (between 5
and 10 percent) at 0.080 ppm O₃.\12\ Notably, about 26
percent of the 60 exposed subjects had lung function decrements >10
percent, including about 8 percent of the subjects that experienced
large decrements (>20 percent) (EPA, 2007, Figure 3-1A). These results
(which were not corrected for exercise in filtered air responses)
demonstrate that while average responses may be relatively small at the
0.080 ppm exposure level, some individuals experience more severe
effects that may be clinically significant. Similar results at the
0.080 ppm exposure level (for 6.6 hours during intermittent exercise)
were seen in more recent studies of 30 healthy young adults by Adams
(2002, 2006).\13\ In these studies, relatively small but statistically
significant lung function decrements and respiratory symptom responses
were found (for both square-wave and triangular exposure patterns),
with 17 percent of the subjects (5 of 30) experiencing >= 10 percent
FEV₁ decrements (comparing pre- and post-exposures) when the
results were not corrected for the effects of exercise alone in
filtered air (EPA, 2007, Figure 3-1B) and with 23 percent of subjects
(7 of 30) experiencing such effects when the results were corrected
(EPA, 2007, p. 3-6).\14\
---------------------------------------------------------------------------

\12\ This study and other studies (Folinsbee et al., 1988;
Horstman et al., 1990; and McDonnell et al., 1991), conducted in
EPA's clinical research facility in Chapel Hill, NC, measured ozone
concentrations to within +/-5 percent or +/-0.004 ppm at the 0.080
ppm exposure level.
\13\ These studies, conducted at a facility at the University of
California, in Davis, CA, reported O₃ concentrations to
be accurate within +/-0.003 ppm over the range of concentrations
included in these studies.
\14\ These distributional results presented in the Criteria
Document and Staff Paper for the Adams studies are based on study
data that were not included in the publication but were obtained
from the author.
---------------------------------------------------------------------------

These studies by Adams (2002, 2006) are notable in that they are
the only available controlled exposure human studies that examine
respiratory effects associated with prolonged O₃ exposures
at levels below 0.080 ppm, which was the lowest exposure level that had
been examined in the last review. The Adams (2006) study investigated a
range of exposure levels (0.000, 0.040, 0.060, and 0.080 ppm
O₃) using square-wave and triangular exposure patterns. The
study was designed to examine multiple comparisons of pulmonary
function (FEV₁) and respiratory symptom responses (total
subjective symptoms (TSS) and pain on deep inspiration (PDI)) between
these various exposure protocols at six different time points within
the exposure periods. At the 0.060 ppm exposure level, the author
reported no statistically significant differences for FEV₁
decrements nor for most respiratory symptoms responses; statistically
significant responses were reported only for TSS for the triangular
exposure pattern toward the end of the exposure period, with the PDI
responses being noted as following a closely similar pattern (Adams,
2006, p. 131-132). EPA's reanalysis of the data from the Adams (2006)
study, comparing FEV₁ responses pre- and post-exposure at
the 0.060 ppm exposure level, found small group mean differences from
responses to filtered air that were statistically significant.\15\
Notably, these studies report a small percentage of subjects
experiencing lung function decrement (>= 10 percent) at the 0.060 ppm
exposure level.\16\
---------------------------------------------------------------------------

\15\ Brown, J.S. (2007). EPA Office of Research and Development
memorandum to Ozone NAAQS Review Docket (OAR-2005-0172); Subject:
The effects of ozone on lung function at 0.06 ppm in healthy adults,
June 14, 2007.
\16\ Based on study data (Adams, 2006) provided by the author, 7
percent of the subjects (2 of 30 subjects) experienced notable
FEV₁ decrements >= 10 percent) with the square wave
exposure pattern at the 0.060 ppm exposure level (comparing pre- and
post-exposures) when the results were corrected for the effects of
exercise alone in filtered air (EPA, 2007, p. 3-6).
---------------------------------------------------------------------------

(ii) Results of Epidemiological and Field Studies
A relatively large number of field studies investigating the
effects of ambient O₃ concentrations, in combination with
other air pollutants, on lung function decrements and respiratory
symptoms have been published over the last decade that support the
major findings of the 1996 Criteria Document that lung function
changes, as measured by decrements in FEV₁ or peak
expiratory flow (PEF), and respiratory symptoms in healthy adults and
asthmatic children are closely correlated to ambient O₃
concentrations. Pre-1996 field studies focused primarily on children
attending summer camps and found O₃-related impacts on
measures of lung function, but not respiratory symptoms, in healthy
children. The newer studies have expanded to evaluate O₃-
related effects on outdoor workers, athletes, the elderly, hikers,
school children, and asthmatics. Collectively, these studies confirm
and extend clinical observations that prolonged (i.e., 6-8 hour)
exposure periods, combined with elevated levels of exertion or
exercise, increase the dose of O₃ to the lungs at a given
ambient exposure level and result in larger lung function effects. The
results of one large study of hikers (Korrick et al., 1998), which
reported outcome measures stratified by several factors (e.g., gender,
age, smoking status, presence of asthma) within a population capable of
more than normal exertion, provide useful insight. In this study, lung
function was measured before and after hiking, and individual
O₃ exposures were estimated by averaging hourly
O₃ concentrations from ambient monitors located at the base
and summit. The mean 8-hour average O₃ concentration was
0.040 ppm (8-hour average concentration range of 0.021 ppm to 0.074 ppm
O₃). Decreased lung function was associated with
O₃ exposure, with the greatest effect estimates reported for
the subgroup that reported having asthma or wheezing, and for those who
hiked for longer periods of time.
Asthma panel studies conducted both in the U.S. and in other
countries have reported that decrements in PEF are associated with
routine O₃ exposures among asthmatic and healthy persons.
One large U.S. multicity study, the National Cooperative Inner City
Asthma Study or NCICAS, (Mortimer et al., 2002) examined O₃-
related changes in PEF in 846 asthmatic children from 8 urban areas and
reported that the incidence of >= 10 percent decrements in morning PEF
are associated with increases in 8-hour average O₃ for a 5-
day cumulative lag, suggesting that O₃ exposure may be
associated with clinically significant changes in PEF in

[[Page 37829]]

asthmatic children; however, no associations were reported with evening
PEF. The mean 8-hour average O₃ was 0.048 ppm across the 8
cities. Excluding days when 8-hour average O₃ was greater
than 0.080 ppm (less than 5 percent of days), the associations with
morning PEF remained statistically significant. Mortimer et al. (2002)
discussed potential biological mechanisms for delayed effects on
pulmonary function in asthma, which included increased nonspecific
airway responsiveness secondary to airway inflammation due to
O₃ exposure. Two other panel studies (Romieu et al., 1996,
1997) carried out simultaneously in northern and southwestern Mexico
City with mildly asthmatic school children reported statistically
significant O₃-related reductions in PEF, with variations in
effect depending on lag time and time of day. Mean 1-hour maximum
O₃ concentrations in these locations ranged from 0.190 ppm
(SD 80) in northern Mexico City to 0.196 ppm (SD 78) in southwestern
Mexico City. While several studies report statistically significant
associations between O₃ exposure and reduced PEF in
asthmatics, other studies did not, possibly due to low levels of
O₃ exposure. EPA concludes that these studies collectively
indicate that O₃ may be associated with short-term declines
in lung function in asthmatic individuals and that the Mortimer et al.
(2002) study showed statistically significant effect at concentrations
in the range below 0.080 ppm O₃.
Most of the panel studies which have investigated associations
between O₃ exposure and respiratory symptoms or increased
use of asthma medication are focused on asthmatic children. Two large
U.S. studies (Mortimer et al., 2002; Gent et al., 2003) have reported
associations between ambient O₃ concentrations and daily
symptoms/asthma medication use, even after adjustment for copollutants.
Results were more mixed, meaning that a greater proportion of studies
were not both positive and statistically significant, across smaller
U.S. and international studies that focused on these health endpoints.
The NCICAS reported morning symptoms in 846 asthmatic children from
8 U.S. urban areas to be most strongly associated with a cumulative 1-
to 4-day lag of O₃ concentrations (Mortimer et al., 2002).
The NCICAS used standard protocols that included instructing caretakers
of the subjects to record symptoms (including cough, chest tightness,
and wheeze) in the daily diary by observing or asking the child. While
these associations were not statistically significant in several
cities, when the individual data are pooled from all eight cities,
statistically significant effects were observed for the incidence of
symptoms. The authors also reported that the odds ratios remained
essentially the same and statistically significant for the incidence of
morning symptoms when days with 8-hour O₃ concentrations
above 0.080 ppm were excluded. These days represented less than 5
percent of days in the study.
Gent and colleagues (2003) followed 271 asthmatic children under
age 12 and living in southern New England for 6 months (April through
September) using a daily symptom diary. They found that mean 1-hour max
O₃ and 8-hour max O₃ concentrations were 0.0586
ppm (SD 19.0) and 0.0513 ppm (SD 15.5), respectively. The data were
analyzed for two separate groups of subjects, those who used
maintenance asthma medications during the follow-up period and those
who did not. The need for regular medication was considered to be a
proxy for more severe asthma. Not taking any medication on a regular
basis and not needing to use a bronchodilator would suggest the
presence of very mild asthma. Statistically significant effects of 1-
day lag O₃ were observed on a variety of respiratory
symptoms only in the medication user group. Both daily 1-hour max and
8-hour max O₃ concentrations were similarly related to
symptoms such as chest tightness and shortness of breath. Effects of
O₃, but not PM_2.5, remained significant and even
increased in magnitude in two-pollutant models. Some of the
associations were noted at 1-hour max O₃ levels below 0.060
ppm. In contrast, no effects were observed among asthmatics not using
maintenance medication. In terms of person days of follow-up, this is
one of the larger studies currently available that address symptom
outcomes in relation to O₃, and provides supportive evidence
for effects of O₃ independent of PM_2.5. Study
limitations include the post-hoc nature of the population
stratification by medication use. Also, the study did not account for
all of the important meteorological factors that might influence these
results, such as relative humidity or dew point.
The multicity study by Mortimer et al. (2002), which provides an
asthmatic population representative of the United States, and several
single-city studies indicate a robust association of O₃
concentrations with respiratory symptoms and increased medication use
in asthmatics. While there are a number of well-conducted, albeit
relatively smaller, U.S. studies which showed only limited or a lack of
evidence for symptom increases associated with O₃ exposure,
these studies had less statistical power and/or were conducted in areas
with relatively low 1-hour maximum average O₃ levels, in the
range of 0.03 to 0.09 ppm. Even so, the evidence has continued to
expand since 1996 and now is considered to be much stronger than in the
previous review. The Criteria Document concludes that the asthma panel
studies, as a group, and the NCICAS in particular, indicate a positive
association between ambient concentrations and respiratory symptoms and
increased medication use in asthmatics. The evidence has continued to
expand since 1996 and now is considered to be much stronger than in the
previous review of the O₃ primary standard.
School absenteeism is another potential surrogate for the health
implications of O₃ exposure in children. The association
between school absenteeism and ambient O₃ concentrations was
assessed in two relatively large field studies. Chen et al. (2000)
examined total daily school absenteeism in about 28,000 elementary
school students in Nevada over a 2-year period (after adjusting for
PM₁₀ and CO concentrations) and found that ambient
O₃ concentrations with a distributed lag of 14 days were
statistically significantly associated with an increased rate of school
absences. Gilliland et al. (2001) studied O₃-related
absences among about 2,000 4th grade students in 12 southern California
communities and found statistically significant associations between 8-
hour average O₃ concentrations (with a distributed lag out
to 30 days) and all absence categories, and particularly for
respiratory causes. Neither PM₁₀ nor NO₂ were
associated with any respiratory or nonrespiratory illness-related
absences in single pollutant models. The Criteria Document concludes
that these studies of school absences suggest that ambient
O₃ concentrations, accumulated over two to four weeks, may
be associated with school absenteeism, and particularly illness-related
absences, but further replication is needed before firm conclusions can
be reached regarding the effect of O₃ on school absences. In
addition, more research is needed to help shed light on the
implications of variation in the duration of the lag structures (i.e.,
1 day, 5 days, 14 days, and 30 days) found both across studies and
within data sets by health endpoint and exposure metric.

[[Page 37830]]

(b) Increased Airway Responsiveness
As discussed in more detail in the Criteria Document (section 6.8)
and Staff Paper (section 3.3.1.1.2), increased airway responsiveness,
also known as airway hyperresponsiveness (AHR) or bronchial
hyperreactivity, refers to a condition in which the propensity for the
airways to bronchoconstrict due to a variety of stimuli (e.g., exposure
to cold air, allergens, or exercise) becomes augmented. This condition
is typically quantified by measuring the decrement in pulmonary
function after inhalation exposure to specific (e.g., antigen,
allergen) or nonspecific (e.g., methacholine, histamine)
bronchoconstrictor stimuli. Exposure to O₃ causes an
increase in airway responsiveness as indicated by a reduction in the
concentration of stimuli required to produce a given reduction in
FEV₁ or airway obstruction. Increased airway responsiveness
is an important consequence of exposure to O₃ because its
presence means that the airways are predisposed to narrowing on
inhalation of various stimuli, such as specific allergens, cold air or
SO₂. Statistically significant and clinically relevant
decreases in pulmonary function have been observed in early phase
allergen response in subjects with allergic rhinitis after consecutive
(4-day) 3-hour exposures to 0.125 ppm O₃ (Holz et al.,
2002). Similar increased airway responsiveness in asthmatics to house
dust mite antigen 16 to 18 hours after exposure to a single dose of
O₃ (0.160 ppm for 7.6 hours) was observed. These
observations, based on O₃ exposures to levels much higher
than the current standard level suggest that O₃ exposure may
be a clinically important factor that can exacerbate the response to
ambient bronchoconstrictor substances in individuals with preexisting
allergic asthma or rhinitis. Further, O₃ may have an
immediate impact on the lung function of asthmatics as well as
contribute to effects that persist for longer periods.
Kreit et al. (1989) found that O₃ can induce increased
airway responsiveness in asthmatic subjects to O₃, who
typically have increased airway responsiveness at baseline. A
subsequent study (J[ouml]rres et al., 1996) suggested an increase in
specific (i.e., allergen-induced) airway reactivity in subjects with
allergic asthma, and to a lesser extent in subjects with allergic
rhinitis after short-term exposure to higher O₃ levels;
other studies reported similar results. According to one study
(Folinsbee and Hazucha, 2000), changes in airway responsiveness after
O₃ exposure resolve more slowly than changes in
FEV₁ or respiratory symptoms. Other studies of repeated
exposure to O₃ suggest that changes in airway responsiveness
tend to be somewhat less affected by attenuation with consecutive
exposures than changes in FEV₁ (EPA, 2006a, p. 6-31).
The Criteria Document (section 6.8) concludes that O₃
exposure is linked with increased airway responsiveness. Both human and
animal studies indicate that increased airway responsiveness is not
mechanistically associated with inflammation, and does not appear to be
strongly associated with initial decrements in lung function or
increases in symptoms. As a result of increased airway responsiveness
induced by O₃ exposure, human airways may be more
susceptible to a variety of stimuli, including antigens, chemicals, and
particles. Because asthmatic subjects typically have increased airway
responsiveness at baseline, enhanced bronchial response to antigens in
asthmatics raises potential public health concerns as they could lead
to increased morbidity (e.g., medication usage, school absences,
emergency room visits, hospital admissions) or to more persistent
alterations in airway responsiveness (Criteria Document, p. 8-21). As
such, increased airway responsiveness after O₃ exposure
represents a plausible link between O₃ exposure and
increased hospital admissions.
(c) Respiratory Inflammation and Increased Permeability
Based on evidence from the previous review, acute inflammatory
responses in the lung have been observed subsequent to 6.6 hour
O₃ exposures to the lowest tested level--0.080 ppm--in
healthy adults engaged in moderately high exercise (section 6.9 of the
Criteria Document and section 3.3.1.3 of the Staff Paper). Some of
these prior studies suggest that inflammatory responses may be detected
in some individuals following O₃ exposures in the absence of
O₃-induced pulmonary decrements in those subjects. These
studies also demonstrate that short-term exposures to O₃
also can cause increased permeability in the lungs of humans and
experimental animals. Inflammatory responses and epithelial
permeability have been seen to be independent of spirometric responses.
Not only are the newer lung inflammation and increased cellular
permeability findings discussed in the Criteria Document (pp. 8-21 to
8-24) consistent with the previous review, but they provide better
characterization of the physiological mechanisms by which O₃
causes these effects.
Lung inflammation and increased permeability, which are distinct
events controlled by different mechanisms, are two commonly observed
effects of O₃ exposure observed in all of the species
studied. Increased cellular permeability is a disruption of the lung
barrier that leads to leakage of serum proteins, influx of
polymorphonuclear leukocytes (neutrophils or PMNs), release of
bioactive mediators, and movement of compounds from the airspaces into
the blood.
A number of controlled human exposure studies have analyzed
bronchoalveolar lavage (BAL) and nasal lavage (NL)\17\ fluids and cells
for markers of inflammation and lung damage (EPA, 2006a, Annex AX6).
Increased lung inflammation is demonstrated by the presence of
neutrophils found in BAL fluid in the lungs, which has long been
accepted as a hallmark of inflammation. It is apparent, however, that
inflammation within airway tissues may persist beyond the point that
inflammatory cells are found in the BAL fluid. Soluble mediators of
inflammation, such as cytokines and arachidonic acid metabolites have
been measured in the BAL fluid of humans exposed to O₃. In
addition to their role in inflammation, many of these compounds have
bronchoconstrictive properties and may be involved in increased airway
responsiveness following O₃ exposure. An in vitro study of
epithelial cells from nonatopic and atopic asthmatics exposed to 0.010
to 0.100 ppm O₃ showed significantly increased permeability
compared to cells from normal persons. This indicates a potentially
inherent susceptibility of cells from asthmatic individuals for
O₃-induced permeability.
---------------------------------------------------------------------------

\17\ Graham and Koren (1990) compared inflammatory mediators
present in NL and BAL fluids of humans exposed to 0.4 ppm
O₃ for 2 hours and found similar increases in PMNs in
both fluids, suggesting a qualitative correlation between
inflammatory changes in the lower airways (BAL) and upper
respiratory tract (NL).
---------------------------------------------------------------------------

In the 1996 Criteria Document, assessment of controlled human
exposure studies indicated that a single, acute (1 to 4 hours)
O₃ exposure (>= 0.080 to 0.100 ppm) of subjects engaged in
moderate to heavy exercise could induce a number of cellular and
biochemical changes suggestive of pulmonary inflammation and lung
permeability (EPA, 2006a, p. 8-22). These changes persisted for at
least 18 hours. Markers from BAL fluid following both 2-hour and 4-hour
O₃ exposures repeated up to 5 days indicate that there is
ongoing cellular damage irrespective of attenuation of

[[Page 37831]]

some cellular inflammatory responses of the airways, pulmonary
function, and symptom scores (EPA, 2006a, p. 8-22). Acute airway
inflammation was shown in Devlin et al. (1990) to occur among adults
exposed to 0.080 ppm O₃ for 6.6 hours with exercise. McBride
et al. (1994) reported that asthmatic subjects were more sensitive than
non-asthmatics to upper airway inflammation for O₃ exposures
that did not affect pulmonary function (EPA, 2006a, p. 6-33). However,
the public health significance of these changes is not entirely clear.
The studies reporting inflammatory responses and markers of lung
injury have clearly demonstrated that there is significant variation in
response of subjects exposed, especially to 6.6 hours O₃
exposures at 0.080 and 0.100 ppm. To provide some perspective on the
public health impact for these effects, the Staff Paper (section
3.3.1.1.3) notes that one study (Devlin et al., 1991) showed that
roughly 10 to 50 percent of the 18 young healthy adult subjects
experienced notable increases (i.e., >= 2 fold increase) in most of the
inflammatory and cellular injury indicators analyzed, associated with
6.6-hour exposures at 0.080 ppm. Similar, although in some cases
higher, fractions of the population of 10 healthy adults tested saw > 2
fold increases associated with 6.6-hour exposures to 0.100 ppm. The
authors of this study expressed the view that ``susceptible
subpopulations such as the very young, elderly, and people with
pulmonary impairment or disease may be even more affected'' (Devlin et
al., 1991).
Since 1996, a substantial number of human exposure studies have
been published which have provided important new information on lung
inflammation and epithelial permeability. Mudway and Kelly (2004)
examined O₃-induced inflammatory responses and epithelial
permeability with a meta-analysis of 21 controlled human exposure
studies and showed that an influx in neutrophils and protein in healthy
subjects is associated with total O₃ dose (product of
O₃ concentration, exposure duration, and minute ventilation)
(EPA, 2006a, p. 6-34). Results of the analysis suggest that the time
course for inflammatory responses (including recruitment of neutrophils
and other soluble mediators) is not clearly established, but there is
evidence that attenuation profiles for many of these parameters are
different (EPA, 2006a, p. 8-22).
The Criteria Document (chapter 8) concludes that interaction of
O₃ with lipid constituents of epithelial lining fluid (ELF)
and cell membranes and the induction of oxidative stress is implicated
in injury and inflammation. Alterations in the expression of cytokines,
chemokines, and adhesion molecules, indicative of an ongoing oxidative
stress response, as well as injury repair and regeneration processes,
have been reported in animal toxicology and human in vitro studies
evaluating biochemical mediators implicated in injury and inflammation.
While antioxidants in ELF confer some protection, O₃
reactivity is not eliminated at environmentally relevant exposures
(Criteria Document, p. 8-24). Further, antioxidant reactivity with
O₃ is both species-specific and dose-dependent.
(d) Increased Susceptibility to Respiratory Infection
As discussed in more detail in the Criteria Document (sections
5.2.2, 6.9.6, and 8.4.2), short-term exposures to O₃ have
been shown to impair physiological defense capabilities in experimental
animals by depressing alveolar macrophage (AM) functions and by
altering the mucociliary clearance of inhaled particles and microbes
resulting in increased susceptibility to respiratory infection. Short-
term O₃ exposures also interfere with the clearance process
by accelerating clearance for low doses and slowing clearance for high
doses. Animal toxicological studies have reported that acute
O₃ exposures suppress alveolar phagocytosis and immune
system functions. Dysfunction of host defenses and subsequent increased
susceptibility to bacterial lung infection in laboratory animals has
been induced by short-term exposures to O₃ levels as low as
0.080 ppm.
A single controlled human exposure study reviewed in the 1996
Criteria Document reported that exposure to 0.080 to 0.100 ppm
O₃ for 6.6 hours (with moderate exercise) induced decrements
in the ability of AMs to phagocytose microorganisms (EPA, 2006a, p. 8-
26). Integrating the recent animal study results with human exposure
evidence available in the 1996 Criteria Document, the Criteria Document
concludes that available evidence indicates that short-term
O₃ exposures have the potential to impair host defenses in
humans, primarily by interfering with AM function. Any impairment in AM
function may lead to decreased clearance of microorganisms or nonviable
particles. Compromised AM functions in asthmatics may increase their
susceptibility to other O₃ effects, the effects of
particles, and respiratory infections (EPA, 2006a, p. 8-26).
(e) Morphological Effects
The 1996 Criteria Document found that short-term O₃
exposures cause similar alterations in lung morphology in all
laboratory animal species studied, including primates. As discussed in
the Staff Paper (section 3.3.1.1.5), cells in the centriacinar region
(CAR) of the lung (the segment between the last conducting airway and
the gas exchange region) have been recognized as a primary target of
O₃-induced damage (epithelial cell necrosis and remodeling
of respiratory bronchioles), possibly because epithelium in this region
receives the greatest dose of O₃ delivered to the lower
respiratory tract. Following chronic O₃ exposure, structural
changes have been observed in the CAR, the region typically affected in
most chronic airway diseases of the human lung (EPA, 2006a, p. 8-24).
Ciliated cells in the nasal cavity and airways, as well as Type I
cells in the gas-exchange region, are also identified as targets. While
short-term O₃ exposures can cause epithelial cell
proliferation and fibrolitic changes in the CAR, these changes appear
to be transient with recovery time after exposure, depending on species
and O₃ dose. The potential impacts of repeated short-term
and chronic morphological effects of O₃ exposure are
discussed below in the section on effects from long-term exposures.
Long-term or prolonged exposure has been found to cause chronic lesions
similar to early lesions of respiratory bronchiolitis, which have the
potential to progress to fibrotic lung disease (Criteria Document, p. 8-25).
Recent studies continue to show that short-term and sub-chronic
exposures to O₃ cause similar alterations in lung structure
in a variety of experimental animal species. For example, a series of
new studies that used infant rhesus monkeys and simulated seasonal
ambient exposure (0.5 ppm 8 hours/day for 5 days, every 14 days for 11
episodes) reported remodeling in the distal airways; abnormalities in
tracheal basement membrane; eosinophil accumulation in conducting
airways; and decrements in airway innervation (Criteria Document, p. 8-
25). Based on evidence from animal toxicological studies, short-term
and sub-chronic exposures to O₃ can cause morphological
changes in the respiratory systems, particularly in the CAR, of a
number of laboratory animal species (EPA, 2006a, section 5.2.4).

[[Page 37832]]

(f) Emergency Department Visits/Hospital Admissions for Respiratory Causes
Increased summertime emergency department visits and hospital
admissions for respiratory causes have been associated with ambient
exposures to O₃. As discussed in section 3.3.1.1.6 of the
Staff Paper, numerous studies conducted in various locations in the
U.S. and Canada consistently have shown a relationship between ambient
O₃ levels and increased incidence of emergency department
visits and hospital admissions for respiratory causes, even after
controlling for modifying factors, such as weather and copollutants.
Such associations between elevated ambient O₃ during summer
months and increased hospital admissions have a plausible biological
basis in the human and animal evidence of functional, symptomatic, and
physiologic effects discussed above and in the increased susceptibility
to respiratory infections observed in laboratory animals.
In the last review of the O₃ NAAQS, the Criteria
Document evaluated emergency department visits and hospital admissions
as possible outcomes following exposure to O₃ (EPA, 2006a,
section 7.3). The evidence was limited for emergency department visits,
but results of several studies generally indicated that short-term
exposures to O₃ were associated with respiratory emergency
department visits. The strongest and most consistent evidence, at both
lower levels (i.e., below 0.120 ppm 1-hour max O₃) and at
higher levels (above 0.120 ppm 1-hour max O₃), was found in
the group of studies which investigated summertime\18\ daily hospital
admissions for respiratory causes in different eastern North American
cities. These studies consistently demonstrated that ambient
O₃ levels were associated with increased hospital admissions
and accounted for about one to three excess respiratory hospital
admissions per million persons with each 0.100 ppm increase in 1-hour
max O₃, after adjustment for possible confounding effects of
temperature and copollutants. Overall, the 1996 Criteria Document
concluded that there was strong evidence that ambient O₃
exposures can cause significant exacerbations of preexisting
respiratory disease in the general public. Excess respiratory-related
hospital admissions associated with O₃ exposures for the New
York City area (based on Thurston et al., 1992) were included in the
quantitative risk assessment in the prior review and are included in
the current assessment along with estimates for respiratory-related
hospital admissions in Cleveland, Detroit, and Los Angeles based on
more recent studies (Staff Paper, chapter 5). Significant uncertainties
and the difficulty of obtaining reliable baseline incidence numbers
resulted in emergency department visits not being used in the
quantitative risk assessment in either the last or the current
O₃ NAAQS review.
---------------------------------------------------------------------------

\18\ Discussion of the reasons for focusing on warm season
studies is found in the section 2.A.3.a below.
---------------------------------------------------------------------------

In the past decade, a number of studies have examined the temporal
pattern associations between O₃ exposures and emergency
department visits for respiratory causes (EPA, 2006a, section 7.3.2).
These studies are summarized in the Criteria Document (chapter 7 Annex)
and some are shown in Figure 1 (in section II.A.3). Respiratory causes
for emergency department visits include asthma, bronchitis, emphysema,
pneumonia, and other upper and lower respiratory infections, such as
influenza, but asthma visits typically dominate the daily incidence
counts. Most studies report positive associations. Among studies with
adequate controls for seasonal patterns, many reported at least one
significant positive association involving O₃.
In reviewing evidence for associations between emergency department
visits for asthma and short-term O₃ exposures, the Criteria
Document notes that in general, O₃ effect estimates from
summer only analyses tended to be positive and larger compared to
results from cool season or all year analyses (Figure 7-8, EPA, 2006a,
p. 7-68). Several of the studies reported significant associations
between O₃ concentrations and emergency department visits
for respiratory causes, in particular asthma. However, inconsistencies
were observed which were at least partially attributable to differences
in model specifications and analysis approach among various studies.
For example, ambient O₃ concentrations, length of the study
period, and statistical methods used to control confounding by seasonal
patterns and copollutants appear to affect the observed O₃
effect on emergency department visits. Thus, the Criteria Document has
concluded that stratified analyses by season generally supported a
positive association between O₃ concentrations and emergency
department visits for asthma in the warm season.
Hospital admissions studies focus specifically on unscheduled
admissions because unscheduled hospital admissions occur in response to
unanticipated disease exacerbations and are more likely than scheduled
admissions to be affected by variations in environmental factors, such
as daily O₃ levels. Results of a fairly large number of
these studies published during the past decade are summarized in
Criteria Document (chapter 7 Annex), and results of U.S. and Canadian
studies are shown in Figure 1 below (in section II.A.3). As a group,
these hospital admissions studies tend to be larger geographically and
temporally than the emergency department visit studies and provide
results that are generally more consistent. The strongest associations
of respiratory hospital admissions with O₃ concentrations
were observed using short lag periods, in particular for a 0-day lag
(same day exposure) and a 1-day lag (previous day exposure). Most
studies in the United States and Canada indicated positive,
statistically significant associations between ambient O₃
concentrations and respiratory hospital admissions in the warm season.
However, not all studies found a statistically significant relationship
with O₃, possibly because of very low ambient O₃
levels. Analyses for confounding using multipollutant regression models
suggest that copollutants generally do not confound the association
between O₃ and respiratory hospitalizations. Ozone effect
estimates were robust to PM adjustment in all-year and warm-season only
data.
Overall, the Criteria Document concludes that positive and robust
associations were found between ambient O₃ concentrations
and various respiratory disease hospitalization outcomes, when focusing
particularly on results of warm-season analyses. Recent studies also
generally indicate a positive association between O₃
concentrations and emergency department visits for asthma during the
warm season (EPA, 2006a, p. 7-175). These positive and robust
associations are supported by the human clinical, animal toxicological,
and epidemiological evidence for lung function decrements, increased
respiratory symptoms, airway inflammation, and increased airway
responsiveness. Taken together, the overall evidence supports a causal
relationship between acute ambient O₃ exposures and
increased respiratory morbidity outcomes resulting in increased
emergency department visits and hospitalizations during the warm season
(EPA, 2006a, p. 8-77).

[[Page 37833]]

ii. Effects on the Respiratory System of Long-Term O₃ Exposures
The 1996 Criteria Document concluded that there was insufficient
evidence from the limited number of studies to determine whether long-
term O₃ exposures resulted in chronic health effects at
ambient levels observed in the U.S. However, the aggregate evidence
suggested that O₃ exposure, along with other environmental
factors, could be responsible for health effects in exposed
populations. Animal toxicological studies carried out in the 1980's and
1990's demonstrated that long-term exposures can result in a variety of
morphological effects, including permanent changes in the small airways
of the lungs, including remodeling of the distal airways and CAR and
deposition of collagen, possibly representing fibrotic changes. These
changes result from the damage and repair processes that occur with
repeated exposure. Fibrotic changes were also found to persist after
months of exposure providing a potential pathophysiologic basis for
changes in airway function observed in children in some recent
epidemiological studies. It appears that variable seasonal ambient
patterns of exposure may be of greater concern than continuous daily
exposures.
Several studies published since 1996 have investigated lung
function changes over seasonal time periods (EPA, 2006a, section
7.5.3). The Criteria Document (p. 7-114) summarizes these studies
collectively indicate that seasonal O₃ exposure is
associated with smaller growth-related increases in lung function in
children than they would have experienced living in areas with lower
O₃ levels and that there is some limited, as yet uncertain,
evidence that seasonal O₃ also may affect lung function in
young adults, although the uncertainty about the role of copollutants
makes it difficult to attribute the effects to O₃ alone.
Lung capacity grows during childhood and adolescence as body size
increases, reaches a maximum during the twenties, and then begins to
decline steadily and progressively with age. Long-term exposure to air
pollution has long been thought to contribute to slower growth in lung
capacity, diminished maximally attained capacity, and/or more rapid
decline in lung capacity with age (EPA, 2006a, section 7.5.4).
Toxicological findings evaluated in the 1996 Criteria Document
demonstrated that repeated daily exposure of rats to an episodic
profile of O₃ caused small, but significant, decrements in
growth-related lung function that were consistent with early indicators
of focal fibrogenesis in the proximal alveolar region, without overt
fibrosis. Because O₃ at sufficient concentrations is a
strong respiratory irritant and has been shown to cause inflammation
and restructuring of the respiratory airways, it is plausible that
long-term O₃ exposures might have a negative impact on
baseline lung function, particularly during childhood when these
exposures might have long-term risks.
Several epidemiological studies published since 1996 have examined
the relationship between lung function development and long-term
O₃ exposure. The most extensive and robust study of
respiratory effects in relation to long-term air pollution exposures
among children in the U.S. is the Children's Health Study carried out
in 12 communities of southern California starting in 1993. One analysis
(Peters et al., 1999a) examined the relationship between long-term
O₃ exposures and self-reports of respiratory symptoms and
asthma in a cross sectional analysis and found a limited relationship
between outcomes of current asthma, bronchitis, cough and wheeze and a
0.040 ppm increase in 1-hour max O₃ (EPA, 2006a, p. 7-115).
Another analysis (Peters et al., 1999b) examined the relationship
between lung function at baseline and levels of air pollution in the
community. They reported evidence that annual mean O₃ levels
were associated with decreases in FVC, FEV₁, PEF and forced
expiratory flow (FEF_25-75) (the latter two being
statistically significant) among females but not males. In a separate
analysis (Gauderman et al., 2000) of 4th, 7th, and 10th grade students,
a longitudinal analysis of lung function development over four years
found no association with O₃ exposure. The Children's Health
Study enrolled a second cohort of more than 1500 fourth graders in 1996
(Gauderman et al., 2002). While the strongest associations with
negative lung function growth were observed with acid vapors in this
cohort, children from communities with higher 4-year average
O₃ levels also experienced smaller increases in various lung
function parameters. The strongest relationship with O₃ was
with PEF. Specifically, children from the least-polluted community had
a small but statistically significant increase in PEF as compared to
those from the most-polluted communities. In two-pollutant models, only
8-hour average O₃ and NO₂ were significant joint
predictors of FEV₁ and maximal midexpiratory flow (MMEF).
Although results from the second cohort of children are supportive of a
weak association, the definitive 8-year follow-up analysis of the first
cohort (Gauderman et al., 2004a) provides little evidence that long-
term exposure to ambient O₃ at current levels is associated
with significant deficits in the growth rate of lung function in
children. Avol et al. (2001) examined children who had moved away from
participating communities in southern California to other states with
improved air quality. They found that a negative, but not statistically
significant, association was observed between O₃ and lung
function parameters. Collectively, the results of these reports from
the children's health cohorts provide little evidence to support an
impact of long-term O₃ exposures on lung function development.
Evidence for a significant relationship between long-term
O₃ exposures and decrements in maximally attained lung
function was reported in a nationwide study of first year Yale students
(Kinney et al., 1998; Galizia and Kinney, 1999) (EPA, 2006a, p. 7-120).
Males had much larger effect estimates than females, which might
reflect higher outdoor activity levels and correspondingly higher
O₃ exposures during childhood. A similar study of college
freshmen at University of California at Berkeley also reported
significant effects of long-term O₃ exposures on lung
function (K[uuml]nzli et al., 1997; Tager et al., 1998). In a
comparison of students whose city of origin was either Los Angeles or
San Francisco, long-term O₃ exposures were associated with
significant changes in mid- and end-expiratory flow measures, which
could be considered early indicators for pathologic changes that might
progress to COPD.
There have been a few studies that investigated associations
between long-term O₃ exposures and the onset of new cases of
asthma (EPA, 2006a, section 7.5.6). The Adventist Health and Smog
(AHSMOG) study cohort of about 4,000 was drawn from nonsmoking, non-
Hispanic white adult Seventh Day Adventists living in California (Greer
et al., 1993; McDonnell et al., 1999). During the ten-year follow-up in
1987, a statistically significant increased relative risk of asthma
development was observed in males, compared to a nonsignificant
relative risk in females (Greer et al., 1993). In the 15-year follow-up
in 1992, it was reported that for males, there was a statistically
significant increased relative risk of developing asthma associated
with 8-hour average O₃ exposures, but there was no evidence
of an association in females. Consistency of results in the two studies
with different follow-up

[[Page 37834]]

times provides supportive evidence of the potential for an association
between long-term O₃ exposure and asthma incidence in adult
males; however, representativeness of this cohort to the general U.S.
population may be limited (EPA, 2006a, p. 7-125).
In a similar study (McConnell et al., 2002) of incident asthma
among children (ages 9 to 16 at enrollment), annual surveys of 3,535
children initially without asthma were used to identify new-onset
asthma cases as part of the Children's Health Study. Six high-
O₃ and six low-O₃ communities were identified
where the children resided. There were 265 children who reported new-
onset asthma during the follow-up period. Although asthma risk was no
higher for all residents of the six high-O₃ communities
versus the six low-O₃ communities, asthma risk was 3.3 times
greater for children who played three or more sports as compared with
children who played no sports within the high-O₃
communities. This association was absent in the communities with lower
O₃ concentrations. No other pollutants were found to be
associated with new-onset asthma (EPA, 2006a, p. 7-125). Playing sports
may result in extended outdoor activity and exposure occurring during
periods when O₃ levels are higher. It should be noted,
however, that the results of the Children's Health Study were based on
a small number of new-onset asthma cases among children who played
three or more sports. Future replication of these findings in other cohorts
would help determine whether a causal interpretation is appropriate.
In animal toxicology studies, the progression of morphological
effects reported during and after a chronic exposure in the range of
0.50 to 1.00 ppm O₃ is complex, with inflammation peaking
over the first few days of exposure, then dropping, then plateauing,
and finally, largely disappearing (EPA, 2006a, section 5.2.4.4). By
contrast, fibrotic changes in the tissue increase very slowly over
months of exposure, and, after exposure ceases, the changes sometimes
persist or increase. Epithelial hyperplasia peaks soon after the
inflammatory response but is usually maintained in both the nose and
lungs with continuous exposure; it also does not return to pre-exposure
levels after the end of exposure. Patterns of exposure in this same
concentration range determine effects, with 18 months of daily
exposure, causing less morphologic damage than exposures on alternating
months. This is important as environmental O₃ exposure is
typically seasonal. Long-term studies by Plopper and colleagues (Evans
et al., 2003; Schelegle et al., 2003; Chen et al., 2003; Plopper and
Fanucchi, 2000) investigated infant rhesus monkeys exposed to
simulated, seasonal O₃ and demonstrated: (1) Remodeling in
the distal airways, (2) abnormalities in tracheal basement membrane;
(3) eosinophil accumulation in conducting airways; and (4) decrements
in airway innervation (EPA, 2006a, p. 5-45). These findings provide
additional information regarding possible injury-repair processes
occurring with long-term O₃ exposures suggesting that these
processes are only partially reversible and may progress following
cessation of O₃ exposure. Further, these processes may lead
to nonreversible structural damage to lung tissue; however, there is
still too much uncertainty to characterize the significance of these
findings to human exposure profiles and effect levels (EPA, 2006a, p. 8-25).
In summary, in the past decade, important new longitudinal studies
have examined the effect of chronic O₃ exposure on
respiratory health outcomes. Limited evidence from recent long-term
morbidity studies have suggested in some cases that chronic exposure to
O₃ may be associated with seasonal declines in lung function
or reduced lung function development, increases in inflammation, and
development of asthma in children and adults. Seasonal decrements or
smaller increases in lung function measures have been reported in
several studies; however, the extent to which these changes are
transient remains uncertain. While there is supportive evidence from
animal studies involving effects from chronic exposures, large
uncertainties still remain as to whether current ambient levels and
exposure patterns might cause these same effects in human populations.
The Criteria Document concludes that epidemiological studies of new
asthma development and longer-term lung function declines remain
inconclusive at present (EPA, 2006a, p. 7-134).
iii. Effects on the Cardiovascular System of O₃ Exposure
At the time of the 1997 review, the possibility of O₃-
induced cardiovascular effects was largely unrecognized. Since then, a
very limited body of evidence from animal, controlled human exposure
and epidemiologic studies has emerged that provides evidence for some
potential plausible mechanisms for how O₃ exposures might
exert cardiovascular system effects, however much needs to be done to
substantiate these potential mechanisms. Possible mechanisms may
involve O₃-induced secretions of vasoconstrictive substances
and/or effects on neuronal reflexes that may result in increased
arterial blood pressure and/or altered electrophysiologic control of
heart rate or rhythm. Some animal toxicology studies have shown
O₃-induced decreases in heart rate, mean arterial pressure,
and core temperature. One controlled human exposure study that
evaluated effects of O₃ exposure on cardiovascular health
outcomes found no significant O₃-induced differences in ECG
or blood pressure in healthy or hypertensive subjects but did observe a
significant O₃-induced increase the alveolar-to-arterial
PO₂ gradient and heart rate in both groups resulting in an
overall increase in myocardial work and impairment in pulmonary gas
exchange (Gong et al., 1998). In another controlled human exposure
study, inhalation of a mixture of PM_2.5 and O₃ by
healthy subjects increased brachial artery vasoconstriction and
reactivity (Brook et al., 2002).
The evidence from a few animal studies also includes potential
direct effects such as O₃-induced release from lung
epithelial cells of platelet activating factor (PAF) that may
contribute to blood clot formation that would have the potential to
increase the risk of serious cardiovascular outcomes (e.g., heart
attack, stroke, mortality). Also, interactions of O₃ with
surfactant components in epithelial lining fluid of the lung may result
in production of oxysterols and reactive oxygen species that may
exhibit PAF-like activity contributing to clotting and also may exert
cytotoxic effects on lung and heart muscle cells.
Epidemiologic panel and field studies that examined associations
between O₃ and various cardiac physiologic endpoints have
yielded limited evidence suggestive of a potential association between
acute O₃ exposure and altered heart rate variability,
ventricular arrhythmias, and incidence of heart attacks. A number of
epidemiological studies have also reported associations between short-
term exposures and hospitalization for cardiovascular diseases. As
shown in Figure 7-13 of the Criteria Document, many of the studies
reported negative or inconsistent associations. Some other studies,
especially those that examined the relationship when O₃
exposures were higher, have found robust positive associations between
O₃ and cardiovascular hospital admissions (EPA, 2006a, p. 7-
82). For example, one study reported a positive association between
O₃ and cardiovascular hospital admissions in Toronto, Canada
in a summer-only analysis (Burnett et al.,

[[Page 37835]]

1997b). The results were robust to adjustment for various PM indices,
whereas the PM effects diminished when adjusting for gaseous
pollutants. Other studies stratified their analysis by temperature,
i.e., by warm days versus cool days. Several analyses using warm season
days consistently produced positive associations.
The epidemiologic evidence for cardiovascular morbidity is much
weaker than for respiratory morbidity, with only one of several U.S./
Canadian studies showing statistically significant positive
associations of cardiovascular hospitalizations with warm-season
O₃ concentrations. Most of the available European and
Australian studies, all of which conducted all-year O₃
analyses, did not find an association between short-term O₃
concentrations and cardiovascular hospitalizations. Overall, the
currently available evidence is inconclusive regarding an association
between cardiovascular hospital admissions and ambient O₃
exposure (EPA, 2006a, p. 7-83).
In summary, based on the evidence from animal toxicology, human
controlled exposure, and epidemiologic studies, from the Criteria
Document concludes that this generally limited body of evidence is
suggestive that O₃ can directly and/or indirectly contribute
to cardiovascular-related morbidity, but that much needs to be done to
more fully integrate links between ambient O₃ exposures and
adverse cardiovascular outcomes (EPA, 2006a, p. 8-77).
b. Mortality
i. Mortality and Short-Term O₃ Exposure
The 1996 Criteria Document concluded that an association between
daily mortality and O₃ concentration for areas with high
O₃ levels (e.g., Los Angeles) was suggested. However, due to
a very limited number of studies available at that time, there was
insufficient evidence to conclude that the observed association was
likely causal.
The current Criteria Document includes results from numerous
epidemiological analyses of the relationship between O₃ and
mortality. Additional single city analyses have also been conducted
since 1996, however, the most pivotal studies in EPA's (and CASAC's)
finding of increased support for the relationship between premature
mortality and O₃ is in part related to differences in study
design--limiting analyses to warm seasons, better control for
copollutants, particularly PM, and use of multicity designs (both time
series and meta-analytic designs). Key findings are available from
multi-city time-series studies that report associations between
O₃ and mortality. These studies include analyses using data
from 90 U.S. cities in the National Mortality, Morbidity and Air
Pollution (NMMAPS) study (Dominici et al., 2003) and from 95 U.S.
communities in an extension to the NMMAPS analyses (Bell et al., 2004).
The original 90-city NMMAPS analysis, with data from 1987 to 1994,
was primarily focused on investigating effects of PM₁₀ on
mortality. A significant association was reported between mortality and
24-hour average O₃ concentrations in analyses using all
available data as well as in the warm season only analyses (Dominici et
al., 2003). The estimate using all available data was about half that
for the summer-only data at a lag of 1-day. The extended NMMAPS
analysis included data from 95 U.S. cities and included an additional 6
years of data, from 1987-2000 (Bell et al., 2004). Significant
associations were reported between O₃ and mortality in
analyses using all available data. The effect estimate for increased
mortality was approximately 0.5 percent per 0.020 ppm change in 24-hour
average O₃ measured on the same day, and approximately 1.04
percent per 0.020 ppm change in 24-hour average O₃ in a 7-
day distributed lag model (EPA, 2006a, p. 7-88). In analyses using only
data from the warm season, the results were not significantly different
from the full-year results. The authors also report that O₃-
mortality associations were robust to adjustment for PM (EPA, 2006a, p.
7-100). Using a subset of the NMMAPS data set, Huang et al. (2005)
focused on associations between cardiopulmonary mortality and
O₃ exposure (24-hour average) during the summer season only.
The authors report an approximate 1.47 percent increase per 0.020 ppm
change in O₃ concentration measured on the same day and an
approximate 2.52 percent increase per 0.020 ppm change in O₃
concentration using a 7-day distributed lag model. These findings
suggest that the effect of O₃ on mortality is immediate but
also persists for several days.
As discussed below in section II.A.3.a, confounding by weather,
especially temperature, is complicated by the fact that higher
temperatures are associated with the increased photochemical activities
that are important for O₃ formation. Using a case-crossover
study design, Schwartz (2005) assessed associations between daily
maximum concentrations and mortality, matching case and control periods
by temperature, and using data only from the warm season. The reported
effect estimate of approximately 0.92 percent change in mortality per
0.040 ppm O₃ (1-hour maximum) was similar to time-series
analysis results with adjustment for temperature (approximately 0.76
percent per 0.040 ppm O₃), suggesting that associations
between O₃ and mortality were robust to the different
adjustment methods for temperature.
An initial publication from APHEA, a European multi-city study,
reported statistically significant associations between daily maximum
O₃ concentrations and mortality in four cities in a full
year analysis (Toulomi et al., 1997). An extended analysis was done
using data from 23 cities throughout Europe (Gryparis et al., 2004). In
this report, a positive but not statistically significant association
was found between mortality and 1-hour daily maximum O₃ in a
full year analysis. Gryparis et al. (2004) noted that there was a
considerable seasonal difference in the O₃ effect on
mortality; thus, the small effect for the all-year data might be
attributable to inadequate adjustment for confounding by seasonality.
Focusing on analyses using summer measurements, the authors report
statistically significant associations with total mortality,
cardiovascular mortality and with respiratory mortality (EPA, 2006a, p.
7-93, 7-99).
Numerous single-city analyses have also reported associations
between mortality and short-term O₃ exposure, especially for
those analyses using warm season data. As shown in Figure 7-21 of the
Criteria Document, the results of recent publications show a pattern of
positive, often statistically significant associations between short-
term O₃ exposure and mortality during the warm season. In
considering results from year-round analyses, there remains a pattern
of positive results but the findings are less consistent. In most
single-city analyses, effect estimates were not substantially changed
with adjustment for PM (EPA, 2006a, Figure 7-22).
In addition, several meta-analyses have been conducted on the
relationship between O₃ and mortality. As described in
section 7.4.4 of the Criteria Document, these analyses reported fairly
consistent and positive combined effect estimates ranging from
approximately 1.5 to 2.5 percent increase in mortality for a
standardized change in O₃ (EPA, 2006a, Figure 7-20). Three
recent meta-analyses evaluated potential sources of heterogeneity in
O₃-mortality associations (Bell et al., 2005; Ito et al.,
2005; Levy et al., 2005). The

[[Page 37836]]

Criteria Document (p. 7-96) observes common findings across all three
analyses, in that all reported that effect estimates were larger in
warm season analyses, reanalysis of results using default convergence
criteria in generalized additive models (GAM) did not change the effect
estimates, and there was no strong evidence of confounding by PM. Bell
et al. (2005) and Ito et al. (2005) both provided suggestive evidence
of publication bias, but O₃-mortality associations remained
after accounting for that potential bias. The Criteria Document
concludes that the ``positive O₃ effects estimates, along
with the sensitivity analyses in these three meta-analyses, provide
evidence of a robust association between ambient O₃ and
mortality'' (EPA, 2006a, p. 7-97).
Most of the single-pollutant model estimates from single-city
studies range from 0.5 to 5 percent excess deaths per standardized
increments. Corresponding summary estimates in large U.S. multi-city
studies ranged between 0.5 to 1 percent with some studies noting
heterogeneity across cities and studies (EPA, 2006a, p. 7-110).
Finally, from those studies that included assessment of
associations with specific causes of death, it appears that effect
estimates for associations with cardiovascular mortality are larger
than those for total mortality. The meta-analysis by Bell et al. (2005)
observed a slightly larger effect estimate for cardiovascular mortality
compared to mortality from all causes. The effect estimate for
respiratory mortality was approximately one-half that of cardiovascular
mortality in the meta-analysis. However, other studies have observed
larger effect estimates for respiratory mortality compared to
cardiovascular mortality. The apparent inconsistency regarding the
effect size of O₃-related respiratory mortality may be due
to reduced statistical power in this subcategory of mortality (EPA,
2006a, p. 7-108).
In summary, many single- and multi-city studies observed positive
associations of ambient O₃ concentrations with total
nonaccidental and cardiopulmonary mortality. The Criteria Document
finds that the results from U.S. multi-city time-series studies provide
the strongest evidence to date for O₃ effects on acute
mortality. Recent meta-analyses also indicate positive risk estimates
that are unlikely to be confounded by PM; however, future work is
needed to better understand the influence of model specifications on
the risk coefficient (EPA, 2006a, p. 7-175). A meta-analysis that
examined specific causes of mortality found that the cardiovascular
mortality risk estimates were higher than those for total mortality.
For cardiovascular mortality, the Criteria Document (Figure 7-25, p. 7-
106) suggests that effect estimates are consistently positive and more
likely to be larger and statistically significant in warm season
analyses. The findings regarding the effect size for respiratory
mortality have been less consistent, possibly because of lower
statistical power in this subcategory of mortality. The Criteria
Document (p. 8-78) concludes that these findings are highly suggestive
that short-term O₃ exposure directly or indirectly
contribute to non-accidental and cardiopulmonary-related mortality, but
additional research is needed to more fully establish underlying
mechanisms by which such effects occur.\19\
---------------------------------------------------------------------------

\19\ In commenting on the Criteria Document, the CASAC Ozone
Panel raised questions about the implications of these time-series
results in a policy context, emphasizing that ``* * * while the
time-series study design is a powerful tool to detect very small
effects that could not be detected using other designs, it is also a
blunt tool'' (Henderson, 2006b). They note that ``* * * not only is
the interpretation of these associations complicated by the fact
that the day-to-day variation in concentrations of these pollutants
is, to a varying degree, determined by meteorology, the pollutants
are often part of a large and highly correlated mix of pollutants,
only a very few of which are measured'' (Henderson, 2006b). Even
with these uncertainties, the CASAC Ozone Panel, in its review of
the Staff Paper, found ``* * * premature total non-accidental and
cardiorespiratory mortality for inclusion in the quantitative risk
assessment to be appropriate.'' (Henderson, 2006b).
---------------------------------------------------------------------------

ii. Mortality and Long-Term O₃ Exposure
Little evidence was available in the last review on the potential
for associations between mortality and long-term exposure to
O₃. In the Harvard Six City prospective cohort analysis, the
authors report that mortality was not associated with long-term
exposure to O₃ (Dockery et al., 1993). The authors note that
the range of O₃ concentrations across the six cities was
small, which may have limited the power of the study to detect associations
between mortality and O₃ levels (EPA, 2006a, p. 7-127).
As discussed in section 7.5.8 of the Criteria Document, in this
review there are results available from three prospective cohort
studies: the American Cancer Society (ACS) study (Pope et al., 2002),
the Adventist Health and Smog (AHSMOG) study (Beeson et al., 1998;
Abbey et al., 1999), and the U.S. Veterans Cohort study (Lipfert et
al., 2000, 2003). In addition, a major reanalysis report includes
evaluation of data from the Harvard Six City cohort study (Krewski et
al., 2000).\20\ This reanalysis also includes additional evaluation of
data from the initial ACS cohort study report that had only reported
results of associations between mortality and long-term exposure to
fine particles and sulfates (Pope et al., 1995). This reanalysis was
discussed in the Staff Paper (section 3.3.2.2) but not in the Criteria
Document.
---------------------------------------------------------------------------

\20\ This reanalysis report and the original prospective cohort
study findings are discussed in more detail in section 8.2.3 of the
Air Quality Criteria for Particulate Matter (EPA, 2004).
---------------------------------------------------------------------------

In this reanalysis of data from the previous Harvard Six City
prospective cohort study, the investigators replicated and validated
the findings of the original studies, and the report included
additional quantitative results beyond those available in the original
report (Krewski et al., 2000). In the reanalysis of data from the
Harvard Six Cities study, the effect estimate for the association
between long-term O₃ concentrations and mortality was
negative and nearly statistically significant (relative risk = 0.87, 95
percent CI: 0.76, 1.00).
The ACS study is based on health data from a large prospective
cohort of approximately 500,000 adults and air quality data from about
150 U.S. cities. The initial report (Pope et al., 1995) focused on
associations with fine particles and sulfates, for which significant
associations had been reported in the earlier Harvard Six Cities study
(Dockery et al., 1993). As part of the major reanalysis of these data,
results for associations with other air pollutants were also reported,
and the authors report that no significant associations were found
between O₃ and all-cause mortality. However, a significant
association was reported for cardiopulmonary mortality in the warm
season (Krewski et al., 2000). The ACS II study (Pope et al., 2002)
reported results of associations with an extended data base; the
mortality records for the cohort had been updated to include 16 years
of follow-up (compared with 8 years in the first report) and more
recent air quality data were included in the analyses. Similar to the
earlier reanalysis, a marginally significant association was observed
between long-term exposure to O₃ and cardiopulmonary
mortality in the warm season. No other associations with mortality were
observed in both the full-year and warm season analyses.
The Adventist Health and Smog (AHSMOG) cohort includes about 6,000
adults living in California. In two studies from this cohort, a
significant association has been reported between long-term
O₃ exposure and increased risk of lung cancer mortality
among males only (Beeson et al., 1998; Abbey

[[Page 37837]]

et al., 1999). No significant associations were reported between long-
term O₃ exposure and mortality from all causes or
cardiopulmonary causes. Due to the small numbers of lung cancer deaths
(12 for males, 18 for females) and the precision of the effect estimate
(i.e., the wide confidence intervals), the Criteria Document discussed
concerns about the plausibility of the reported association with lung
cancer (EPA, 2006a, p. 7-130).
The U.S. Veterans Cohort study (Lipfert et al., 2000, 2003) of
approximately 50,000 middle-aged males diagnosed with hypertension,
reported some positive associations between mortality and peak
O₃ exposures (95th percentile level for several years of
data). The study included numerous analyses using subsets of exposure
and mortality follow-up periods which spanned the years 1960 to 1996.
In the results of analyses using deaths and O₃ exposure
estimates concurrently across the study period, there were positive,
statistically significant associations between peak O₃ and
mortality (EPA, 2006a, p. 7-129).
Overall, the Criteria Document concludes that consistent
associations have not been reported between long-term O₃
exposure and all-cause, cardiopulmonary or lung cancer mortality (EPA,
2006a, p. 7-130).
c. Role of Ground-Level O₃ in Solar Radiation-Related Human
Health Effects
Beyond the direct health effects attributable to inhalation
exposure to O₃ in the ambient air discussed above, the
Criteria Document also assesses potential indirect effects related to
the presence of O₃ in the ambient air by considering the
role of ground-level O₃ in mediating human health effects
that may be directly attributable to exposure to solar ultraviolet
radiation (UV-B). The Criteria Document (chapter 10) focuses this
assessment on three key factors, including those factors that govern
(1) UV-B radiation flux at the earth's surface, (2) human exposure to
UV-B radiation, and (3) human health effects due to UV-B radiation. In
so doing, the Criteria Document provides a thorough analysis of the
current understanding of the relationship between reducing ground-level
O₃ concentrations and the potential impact these reductions
might have on increasing UV-B surface fluxes and indirectly
contributing to UV-B related health effects.
There are many factors that influence UV-B radiation penetration to
the earth's surface, including latitude, altitude, cloud cover, surface
albedo, PM concentration and composition, and gas phase pollution. Of
these, only latitude and altitude can be defined with small uncertainty
in any effort to assess the changes in UV-B flux that may be
attributable to any changes in tropospheric O₃ as a result
of any revision to the O₃ NAAQS. Such an assessment of UV-B
related health effects would also need to take into account human
habits, such as outdoor activities (including age- and occupation-
related exposure patterns), dress and skin care to adequately estimate
UV-B exposure levels. However, little is known about the impact of
these factors on individual exposure to UV-B.
Moreover, detailed information does not exist regarding other
factors that are relevant to assessing changes in disease incidence,
including: Type (e.g., peak or cumulative) and time period (e.g.,
childhood, lifetime, current) of exposures related to various adverse
health outcomes (e.g., damage to the skin, including skin cancer;
damage to the eye, such as cataracts; and immune system suppression);
wavelength dependency of biological responses; and interindividual
variability in UV-B resistance to such health outcomes. Beyond these
well recognized adverse health effects associated with various
wavelengths of UV radiation, the Criteria Document (section 10.2.3.6)
also discusses protective effects of UV-B radiation. Recent reports
indicate the necessity of UV-B in producing vitamin D, and that vitamin
D deficiency can cause metabolic bone disease among children and
adults, and may also increase the risk of many common chronic diseases
(e.g., type I diabetes and rheumatoid arthritis) as well as the risk of
various types of cancers. Thus, the Criteria Document concludes that
any assessment that attempts to quantify the consequences of increased
UV-B exposure on humans due to reduced ground-level O₃ must
include consideration of both negative and positive effects. However,
as with other impacts of UV-B on human health, this beneficial effect
of UV-B radiation has not been studied in sufficient detail to allow
for a credible health benefits or risk assessment. In conclusion, the
effect of changes in surface-level O₃ concentrations on UV-
induced health outcomes cannot yet be critically assessed within
reasonable uncertainty (Criteria Document, p. 10-36).
The Agency last considered indirect effects of O₃ in the
ambient air in its 2003 final response to a remand of the Agency's 1997
decision to revise the O₃ NAAQS. In so doing, based on the
available information in the last review, the Administrator determined
that the information linking (a) Changes in patterns of ground-level
O₃ concentrations likely to occur as a result of programs
implemented to attain the 1997 O₃ NAAQS to (b) changes in
relevant exposures to UV-B radiation of concern to public health was
too uncertain at that time to warrant any relaxation in the level of
public health protection previously determined to be requisite to
protect against the demonstrated direct adverse respiratory effects of
exposure to O₃ in the ambient air (68 FR 614). At that time,
the more recent information on protective effects of UV-B radiation was
not available, such that only adverse UV-B-related effects could be
considered. Taking into consideration the more recent information
available in this review, the Criteria Document and Staff Paper
conclude that the effect of changes in ground-level O₃
concentrations, likely to occur as a result of revising the
O₃ NAAQS, on UV-induced health outcomes, including whether
these changes would ultimately result in increased or decreased
incidence of UV-B-related diseases, cannot yet be critically assessed.
EPA requests comment on available studies or data that would be
relevant to conducting a critical assessment with reasonable certainty
of UV-induced health outcomes and how evidence of UV-induced health
outcomes might inform the Agency's review of the primary O₃
standard.
3. Interpretation and Integration of Health Evidence
As discussed below, in assessing the new health evidence, the
Criteria Document integrates findings from experimental (e.g.,
toxicological, dosimetric and controlled human exposure) and
epidemiological studies, to make judgments about the extent to which
causal inferences can be made about observed associations between
health endpoints and exposure to O₃. In evaluating the
evidence from epidemiological studies, the EPA focuses on well-
recognized criteria, including: The strength of reported associations,
including the magnitude and precision of reported effect estimates and
their statistical significance; the robustness of reported
associations, or stability in the effect estimates after considering
factors such as alternative models and model specification, potential
confounding by co-pollutants, and issues related to the consequences of
exposure measurement error; potential aggregation bias in pooling data;
and the consistency of the effects associations as observed by looking
across results of multiple- and

[[Page 37838]]

single-city studies conducted by different investigators in different
places and times. Consideration is also given to evaluating
concentration-response relationships observed in epidemiological
studies to inform judgments about the potential for threshold levels
for O₃-related effects. Integrating more broadly across
epidemiological and experimental evidence, the Criteria Document also
focuses on the coherence and plausibility of observed O₃-
related health effects to reach judgments about the extent to which
causal inferences can be made about observed associations between
health endpoints and exposure to O₃ in the ambient air.
a. Assessment of Evidence From Epidemiological Studies
Key elements of the evaluation of epidemiological studies are
briefly summarized below.
(1) The strength of associations most directly refers to the
magnitude of the reported relative risk estimates. Taking a broader
view, the Criteria Document draws upon the criteria summarized in a
recent report from the U.S. Surgeon General, which define strength of
an association as ``the magnitude of the association and its
statistical strength'' which includes assessment of both effect
estimate size and precision, which is related to the statistical power
of the study (CDC, 2004). In general, when associations are strong in
terms of yielding large relative risk estimates, it is less likely that
the association could be completely accounted for by a potential
confounder or some other source of bias, whereas with associations that
yield small relative risk estimates it is especially important to
consider potential confounding and other factors in assessing
causality. Effect estimates between O₃ and some of the
health outcomes are generally small in size and could thus be
characterized as weak. For example, effect estimates for associations
with mortality generally range from 0.5 to 5 percent increases per
0.040 ppm increase in 1-hour maximum O₃ or equivalent,
whereas associations for hospitalization range up to 50 percent
increases per standardized O₃ increment. However, the
Criteria Document notes that there are large multicity studies that
find small associations between short-term O₃ exposure and
mortality or morbidity and have done so with great precision due to the
statistical power of the studies (EPA, 2006a, p. 8-40). That is, the
power of the studies allows the authors to reliably distinguish even
weak relationships from the null hypothesis with statistical confidence.
(2) In evaluating the robustness of associations, the Criteria
Document (sections 7.1.3 and 8.4.4.3) and Staff Paper (section 3.4.2)
have primarily considered the impact of exposure error, potential
confounding by copollutants, and alternative models and model
specifications.
In time-series and panel studies, the temporal (e.g., daily or
hourly) changes in ambient O₃ concentrations measured at
centrally-located ambient monitoring stations are generally used to
represent a community's exposure to ambient O₃. In
prospective cohort or cross-sectional studies, air quality data
averaged over a period of months to years are used as indicators of a
community's long-term exposure to ambient O₃ and other
pollutants. In both types of analyses, exposure error is an important
consideration, as actual exposures to individuals in the population
will vary across the community.
Ozone concentrations measured at central ambient monitoring sites
may explain, at least partially, the variance in individual exposures
to ambient O₃; however, this relationship is influenced by
various factors related to building ventilation practices and personal
behaviors. Further, the pattern of exposure misclassification error and
the influence of confounders may differ across the outcomes of interest
as well as in susceptible populations. As discussed in the Criteria
Document (section 3.9), only a limited number of studies have examined
the relationship between ambient O₃ concentrations and
personal exposures to ambient O₃. One of the strongest
predictors of the relationship between ambient concentrations and
personal exposures appears to be time spent outdoors. The strongest
relationships were observed in outdoor workers (Brauer and Brook, 1995,
1997; O'Neill et al., 2004). Statistically significant correlations
between ambient concentrations and personal exposures were also
observed for children, who likely spend more time outdoors in the warm
season (Linn et al., 1996; Xue et al., 2005). There is some concern
about the extent to which ambient concentrations are representative of
personal O₃ exposures of another particularly susceptible
group of individuals, the debilitated elderly, since those who suffer
from chronic cardiovascular or respiratory conditions may tend to
protect themselves more than healthy individuals from environmental
threats by reducing their exposure to both O₃ and its
confounders, such as high temperature and PM. Studies by Sarnat et al.
(2001, 2005) that included this susceptible group reported mixed
results for associations between ambient O₃ concentrations
and personal exposures to O₃. Collectively, these studies
observed that the daily averaged personal O₃ exposures tend
to be well correlated with ambient O₃ concentrations despite
the substantial variability that existed among the personal
measurements. These studies provide supportive evidence that ambient
O₃ concentrations from central monitors may serve as valid
surrogate measures for mean personal exposures experienced by the
population, which is of most relevance for time-series studies. A
better understanding of the relationship between ambient concentrations
and personal exposures, as well as of the other factors that affect
relationship will improve the interpretation of concentration-
population health response associations observed.
The Criteria Document (section 7.1.3.1) also discusses the
potential influence of exposure error on epidemiologic study results.
Zeger et al. (2000) outlined the components to exposure measurement
error, finding that ambient exposure can be assumed to be the product
of the ambient concentration and an attenuation factor (i.e., building
filter) and that panel studies and time-series studies that use ambient
concentrations instead of personal exposure measurements will estimate
a health risk that is attenuated by that factor. Navidi et al. (1999)
used data from a children's cohort study to compare effect estimates
from a simulated ``true'' exposure level to results of analyses from
O₃ exposures determined by several methods, finding that
O₃ exposures based on the use of ambient monitoring data
overestimate the individual's O₃ exposure and thus generally
result in O₃ effect estimates that are biased downward (EPA,
2006a, p. 7-8). Similarly, in a reanalysis of a study by Burnett et al.
(1994) on the acute respiratory effects of ambient air pollution, Zidek
et al. (1998) reported that accounting for measurement error, as well
as making a few additional changes to the analysis, resulted in
qualitatively similar conclusions, but the effects estimates were
considerably larger in magnitude (EPA, 2006a, p. 7-8). A simulation
study by Sheppard et al. (2005) also considered attenuation of the risk
based on personal behavior, their microenvironment, and the qualities
of the pollutant in time-series studies. Of particular interest is
their finding that risk estimates were not further attenuated in time-
series studies even when the correlations between personal exposures
and ambient

[[Page 37839]]

concentrations were weak. In addition to overestimation of exposure and
the resulting underestimation of effects, the use of ambient
O₃ concentrations may obscure the presence of thresholds in
epidemiologic studies (EPA, 2006a, p. 7-9).
As discussed in the Criteria Document (section 3.9), using ambient
concentrations to determine exposure generally overestimates true
personal O₃ exposures by approximately 2- to 4-fold in
available studies, resulting in attenuated risk estimates. The
implication is that the effects being estimated occur at fairly low
exposures and the potency of O₃ is greater than these
effects estimates indicate. As very few studies evaluating
O₃ health effects with personal O₃ exposure
measurements exist in the literature, effect estimates determined from
ambient O₃ concentrations must be evaluated and used with
caution to assess the health risks of O₃. In the absence of
available data on personal O₃ exposure, the use of routinely
monitored ambient O₃ concentrations as a surrogate for
personal exposures is not generally expected to change the principal
conclusions from O₃ epidemiologic studies. Therefore,
population health risk estimates derived using ambient O₃
levels from currently available observational studies, with appropriate
caveats about personal exposure considerations, remain useful. The
Criteria Document recommends caution in the quantitative use of effect
estimates calculated using ambient O₃ concentrations as they
may lead to underestimation of the potency of O₃. However,
the Staff Paper observes that the use of these risk estimates for
comparing relative risk reductions between alternative ambient
O₃ standards considered in the risk assessment (discussed
below in section II.B.2) is less likely to suffer from this concern.
Confounding occurs when a health effect that is caused by one risk
factor is attributed to another variable that is correlated with the
causal risk factor; epidemiological analyses attempt to adjust or
control for potential confounders. Copollutants (e.g., PM, CO,
SO₂ and NO₂) can meet the criteria for potential
confounding in O₃-health associations if they are potential
risk factors for the health effect under study and are correlated with
O₃. Effect modifiers include variables that may influence
the health response to the pollutant exposure (e.g., co-pollutants,
individual susceptibility, smoking or age). Both are important
considerations for evaluating effects in a mixture of pollutants, but
for confounding, the emphasis is on controlling or adjusting for
potential confounders in estimating the effects of one pollutant, while
the emphasis for effect modification is on identifying and assessing
the effects for different modifiers. The Criteria Document (p. 7-148)
observes that O₃ is generally not highly correlated with
other criteria pollutants (e.g., PM₁₀, CO, SO₂
and NO₂), but may be more highly correlated with secondary
fine particles, especially during the summer months, and that the
degree of correlation between O₃ and other pollutants may
vary across seasons. For example, positive associations are observed
between O₃ and pollutants such as fine particles during the
warmer months, but negative correlations may be observed during the
cooler months (EPA, 2006a, p. 7-17). Thus, the Criteria Document
(section 7.6.4) pays particular attention to the results of season-
specific analyses and studies that assess effects of PM in potential
confounding of O₃-health relationships. The Criteria
Document also discussed the limitations of commonly used multipollutant
models that include the difficulty in interpreting results where the
copollutants are highly colinear, or where correlations between
pollutants change by season (EPA, 2006a, p. 7-150). This is
particularly the situation where O₃ and a copollutant, such
as sulfates, are formed under the same atmospheric condition; in such
cases multipollutant models would produce unstable and possibly
misleading results (EPA, 2006a, p. 7-152).
For mortality, the results from numerous multi-city and single-city
studies indicate that O₃-mortality associations do not
appear to be substantially changed in multipollutant models including
PM₁₀ or PM_2.5 (EPA, 2006a, p. 7-101; Figure 7-
22). Focusing on results of warm season analyses, effect estimates for
O₃-mortality associations are fairly robust to adjustment
for PM in multipollutant models (EPA, 2006a, p. 7-102; Figure 7-23).
The Criteria Document concludes that in the few multipollutant analyses
conducted for these endpoints, copollutants generally do not confound
the relationship between O₃ and respiratory hospitalization
(EPA, 2006a, p. 7-79 to 7-80; Figure 7-12). Multipollutant models were
not used as commonly in studies of relationships between respiratory
symptoms or lung function with O₃, but the Criteria Document
reports that results of available analyses indicate that such
associations generally were robust to adjustment for PM_2.5
(EPA, 2006a, p. 7-154). For example, in a large multi-city study of
asthmatic children (Mortimer et al., 2002), the O₃ effect
was attenuated, but there was still a positive association; in Gent et
al. (2003), effects of O₃, but not PM_2.5,
remained statistically significant and even increased in magnitude in
two-pollutant models (EPA, 2006a, p. 7-53). Considering this body of
studies, the Criteria Document concludes: ``Multipollultant regression
analyses indicated that O₃ risk estimates, in general, were
not sensitive to the inclusion of copollutants, including
PM_2.5 and sulfate. These results suggest that the effects of
O₃ on respiratory health outcomes appear to be robust and
independent of the effects of other copollutants (EPA, 2006a, p. 7-154).''
The Criteria Document observes that another challenge of time-
series epidemiological analysis is assessing the relationship between
O₃ and health outcomes while avoiding bias due to
confounding by other time-varying factors, particularly seasonal trends
and weather variables (EPA, 2006a, p. 7-14). These variables are of
particular interest because O₃ concentrations have a well-
characterized seasonal pattern and are also highly correlated with
changes in temperature, such that it can be difficult to distinguish
whether effects are associated with O₃ or with seasonal or
weather variables in statistical analyses.
The Criteria Document (section 7.1.3.4) discusses statistical
modeling approaches that have been used to adjust for time-varying
factors, highlighting a series of analyses that were done in a Health
Effects Institute-funded reanalysis of numerous time-series studies.
While the focus of these reanalyses was on associations with PM, a
number of investigators also examined the sensitivity of O₃
coefficients to the extent of adjustment for temporal trends and
weather factors. In addition, several recent studies, including U.S.
multi-city studies (Bell et al., 2005; Huang et al., 2005; Schwartz et
al., 2005) and a meta-analysis study (Ito et al., 2005), evaluated the
effect of model specification on O₃-mortality associations.
As discussed in the Criteria Document (section 7.6.3.1), these studies
generally report that associations reported with O₃ are not
substantially changed with alternative modeling strategies for
adjusting for temporal trends and meteorologic effects. In the meta-
analysis by Ito et al. (2005), a separate multi-city analysis was
presented that found that alternative adjustments for weather resulted
in up to 2-fold difference in the O₃ effect estimate.
Significant confounding can occur when strong seasonal cycles are
present, suggesting

[[Page 37840]]

that season-specific results are more generally robust than year-round
results in such cases. A number of epidemiological studies have
conducted season-specific analyses, and have generally reported
stronger and more precise effect estimates for O₃
associations in the warm season than in analyses conducted in the cool
seasons or over the full year.
(3) Consistency refers to the persistent finding of an association
between exposure and outcome in multiple studies of adequate power in
different persons, places, circumstances and times (CDC, 2004). In
considering results from multi-city studies and single-city studies in
different areas, the Criteria Document (p. 8-41) observes general
consistency in effects of short-term O₃ exposure on
mortality, respiratory hospitalization and other respiratory health
outcomes. The variations in effects that are observed may be
attributable to differences in relative personal exposure to
O₃, as well as varying concentrations and composition of
copollutants present in different regions. Thus, the Criteria Document
(p. 8-41) concludes that ``consideration of consistency or
heterogeneity of effects is appropriately understood as an evaluation
of the similarity or general concordance of results, rather than an
expectation of finding quantitative results with a very narrow range.''
(4) The Staff Paper recognizes that it is likely that there are
biological thresholds for different health effects in individuals or
groups of individuals with similar innate characteristics and health
status. For O₃ exposure, individual thresholds would
presumably vary substantially from person to person due to individual
differences in genetic susceptibility, pre-existing disease conditions
and possibly individual risk factors such as diet or exercise levels
(and could even vary from one time to another for a given person).
Thus, it would be difficult to detect a distinct threshold at the
population level below which no individual would experience a given
effect, especially if some members of a population are unusually
sensitive even down to very low concentrations (EPA, 2004, p. 9-43, 9-44).
Some studies have tested associations between O₃ and
health outcomes after removal of days with higher O₃ levels
from the data set; such analyses do not necessarily indicate the
presence or absence of a threshold, but provide some information on
whether the relationship is found using only lower-concentration data.
For example, using data from 95 U.S. cities, Bell et al. (2004) found
that the effect estimate for an association between short-term
O₃ exposure and mortality was little changed when days
exceeding 0.060 ppm (24-hour average) were excluded in the analysis.
Bell et al. (2006) found no difference in estimated effect even when
all days with 24-hour O₃ concentrations <0.020 ppm were
excluded (EPA, 2006a, p. 8-43). Using data from 8 U.S. cities, Mortimer
and colleagues (2002) also reported that associations between
O₃ and both lung function and respiratory symptoms remained
statistically significant and of the same or greater magnitude in
effect size when concentrations greater than 0.080 ppm (8-hour average)
were excluded (EPA, 2006a, p. 7-46). Several single-city studies also
report similar findings of associations that remain or are increased in
magnitude and statistical significance when data at the upper end of
the concentration range are removed (EPA, 2006a, section 7.6.5).
Other time-series epidemiological studies have used statistical
modeling approaches to evaluate whether thresholds exist in
associations between short-term O₃ exposure and mortality.
As discussed in section 7.6.5 of the Criteria Document, one European
multi-city study included evaluation of the shape of the concentration-
response curve, and observed no deviation from a linear function across
the range of O₃ measurements from the study (Gryparis et
al., 2004; EPA, 2006a, p. 7-154). Several single-city studies also
observed a monotonic increase in associations between O₃ and
morbidity that suggest that no population threshold exists (EPA, 2006a,
p. 7-159).
On the other hand, a study in Korea used several different modeling
approaches and reported that a threshold model provided the best fit
for the data. The results suggested a potential threshold level of
about 0.045 ppm (1-hour maximum concentration; <0.035 ppm, 8-hour
average) for an association between mortality and short-term
O₃ exposure during the summer months (Kim et al., 2004; EPA,
2006a, p. 8-43). The authors reported larger effect estimates for the
association for data above the potential threshold level, suggesting
that an O₃-mortality association might be underestimated in
the non-threshold model. A threshold analysis recently reported by Bell
et al. (2006) for 98 U.S. communities, including the same 95
communities in Bell et al. (2004), indicated that if a population
threshold existed for mortality, it would likely fall below a 24-hour
average O₃ concentration of 0.015 ppm (<0.025 ppm, 8-hour
average). In addition, Burnett and colleagues (1997a,b) plotted the
relationships between air pollutant concentrations and both respiratory
and cardiovascular hospitalization, and it appears in these results
that the associations with O₃ are found in the concentration
range above about 0.030 ppm (1-hour maximum; <0.025 ppm, 8-hour
average). Vedal and colleagues (2003) reported a significant
association between O₃ and mortality in British Columbia
where O₃ concentrations were quite low (mean 1-hour maximum
concentration of 0.0273 ppm). The authors did not specifically test for
threshold levels, but the fact that the association was found in an
area with such low O₃ concentrations suggests that any
potential threshold level would be quite low in this data set.
In summary, the Criteria Document finds that, taken together, the
available evidence from clinical and epidemiological studies suggests
that no clear conclusion can now be reached with regard to possible
threshold levels for O₃-related effects (EPA, 2006a, p. 8-
44). Thus, the available epidemiological evidence neither supports nor
refutes the existence of thresholds at the population level for effects
such as increased hospital admissions and premature mortality. There
are limitations in epidemiological studies that make discerning
thresholds in populations difficult, including low data density in the
lower concentration ranges, the possible influence of exposure
measurement error, and interindividual differences in susceptibility to
O₃-related effects in populations. There is the possibility
that thresholds for individuals may exist in reported associations at
fairly low levels within the range of air quality observed in the
studies but not be detectable as population thresholds in
epidemiological analyses.
b. Biological Plausibility and Coherence of Evidence
The body of epidemiological studies discussed in the Staff Paper
emphasizes the role of O₃ in association with a variety of
adverse respiratory and cardiovascular effects. While recognizing a
variety of plausible mechanisms, there exists a general consensus
suggesting that O₃ could, either directly or through
initiation, interfere with basic cellular oxidation processes
responsible for inflammation, reduced antioxidant capacity,
atherosclerosis and other effects. Reasoning that O₃
influences cellular chemistry through basic oxidative properties (as
opposed to a unique chemical interaction), other reactive oxidizing
species (ROS) in the

[[Page 37841]]

atmosphere acting either independently or in combination with
O₃ may also contribute to a number of adverse respiratory
and cardiovascular health effects. Consequently, the role of
O₃ should be considered more broadly as O₃
behaves as a generator of numerous oxidative species in the atmosphere.
In considering the biological plausibility of reported
O₃-related effects, the Staff Paper (section 3.4.6)
considers this broader question of health effects of pollutant mixtures
containing O₃. The potential for O₃-related
enhancements of PM formation, particle uptake, and exacerbation of PM-
induced cardiovascular effects underscores the importance of
considering contributions of O₃ interactions with other
often co-occurring air pollutants to health effects due to
O₃-containing pollutant mixes. The Staff Paper summarizes
some examples of important pollutant mixture effects from studies that
evaluate interactions of O₃ with other co-occurring
pollutants, as discussed in chapters 4, 5, and 6 of the Criteria Document.
All of the types of interactive effects of O₃ with other
co-occurring gaseous and nongaseous viable and nonviable PM components
of ambient air mixes noted above argue that O₃ acts not only
alone but that O₃ also is a surrogate indicator for air
pollution mixes which may enhance the risk of adverse effects due to
O₃ acting in combination with other pollutants. Viewed from
this perspective, those epidemiologic findings of morbidity and
mortality associations, with ambient O₃ concentrations
extending to quite low levels in many cases, become more understandable
and plausible.
The Criteria Document integrates epidemiological studies with
mechanistic information from controlled human exposure studies and
animal toxicological studies to draw conclusions regarding the
coherence of evidence and biological plausibility of O₃-
related health effects to reach judgments about the causal nature of
observed associations. As summarized below, coherence and biological
plausibility are discussed for each of the following types of
O₃-related effects: short-term effects on the respiratory
system, effects on the cardiovascular system, effects related to long-term
O₃ exposure, and short-term mortality-related health endpoints.
i. Coherence and Plausibility of Short-Term Effects on the Respiratory
System
Acute respiratory morbidity effects that have been associated with
short-term exposure to O₃ include such health endpoints as
decrements in lung function, increased airway responsiveness, airway
inflammation, increased permeability related to epithelial injury,
immune system effects, emergency department visits for respiratory
diseases, and hospitalization due to respiratory illness.
Recent epidemiological studies have supported evidence available in
the previous O₃ NAAQS review on associations between ambient
O₃ exposure and decline in lung function for children. The
Criteria Document (p. 8-34) concludes that exposure to ambient
O₃ has a significant effect on lung function and is
associated with increased respiratory symptoms and medication use,
particularly in asthmatics. Short-term exposure to O₃ has
also been associated with more severe morbidity endpoints, such as
emergency department visits and hospital admissions for respiratory
cases, including specific respiratory illness (e.g., asthma) (EPA,
2006a, sections 7.3.2 and 7.3.3). In addition, a few epidemiological
studies have reported positive associations between short-term
O₃ exposure and respiratory mortality, though the associations
are not generally statistically significant (EPA, 2006a, p. 7-108).
Considering the evidence from epidemiological studies, the results
described above provide evidence for coherence in O₃-related
effects on the respiratory system. Effect estimates from U.S. and
Canadian studies are shown in Figure 1, where it can be seen that
mostly positive associations have been reported with respiratory
effects ranging from respiratory symptoms, such as cough or wheeze, to
hospitalization for various respiratory diseases, and there is
suggestive evidence for associations with respiratory mortality. Many
of the reported associations are statistically significant,
particularly in the warm season. In Figure 1, the central effect
estimate is indicated by a square for each result, with the vertical
bar representing the 95 percent confidence interval around the
estimate. In the discussions that follow, an individual study result is
considered to be statistically significant if the 95 percent confidence
interval does not include zero.\21\ Positive effect estimates indicate
increases in the health outcome with O₃ exposure. In
considering these results as a whole, it is important to consider not
only whether statistical significance at the 95 percent confidence
level is reported in individual studies but also the general pattern of
results, focusing in particular on studies with greater statistical
power that report relatively more precise results.
---------------------------------------------------------------------------

\21\ Results for studies of respiratory symptoms are presented
as odds ratios; an odds ratio of 1.0 is equivalent to no effect, and
thus is presented as equivalent to the zero effect estimate line.
---------------------------------------------------------------------------

BILLING CODE 6560-50-P

[[Page 37842]]
[GRAPHIC]
[TIFF OMITTED] TP11JY07.000

BILLING CODE 6560-50-C
Considering also evidence from toxicological, chamber, and field
studies, the Criteria Document (section 8.6) discusses biological
plausibility and coherence of evidence for acute O₃-induced
respiratory health effects. Inhalation of O₃ for several
hours while subjects are physically active can elicit both acute
adverse pathophysiological changes and subjective respiratory tract
symptoms (EPA, 2006a, section 8.4.2). Acute pulmonary responses
observed in healthy humans exposed to O₃ at ambient
concentrations include: decreased inspiratory capacity; mild
bronchoconstriction; rapid, shallow breathing during exercise;
subjective symptoms of tracheobronchial airway irritation, including
cough and pain on deep inspiration; decreases in measures of lung
function; and increased airway resistance. The severity of symptoms and
magnitude of response depends on inhaled dose, individual O₃
sensitivity, and the degree of attenuation or enhancement of response
resulting from previous O₃ exposures. Lung function studies
of several animal species acutely exposed to relatively low
O₃ levels (0.25 to 0.4 ppm) show responses similar to those
observed in humans, including increased breathing frequency, decreased
tidal volume, increased resistance, and decreased FVC. Alterations in
breathing pattern return to normal within hours of exposure, and
attenuation in functional responses

[[Page 37843]]

following repeated O₃ exposures is similar to those observed
in humans.
Physiological and biochemical alterations investigated in
controlled human exposure and animal toxicology studies tend to support
certain hypotheses of underlying pathological mechanisms which lead to
the development of respiratory-related effects reported in epidemiology
studies (e.g., increased hospitalization and medication use). Some of
these are: (a) Decrements in lung function, (b) bronchoconstriction,
(c) increased airway responsiveness, (d) airway inflammation, (e)
epithelial injury, (f) immune system activation, (g) host defense
impairment, and (h) sensitivity of individuals, which depends on at
least a person's age, disease status, genetic susceptibility, and the
degree of attenuation present due to prior exposures. The time
sequence, magnitude, and overlap of these complex events, both in terms
of development and recovery, illustrate the inherent difficulty of
interpreting the biological plausibility of O₃-induced
cardiopulmonary health effects (EPA, 2006a, p. 8-48).
The interaction of O₃ with airway epithelial cell
membranes and ELF to form lipid ozonation products and ROS is supported
by numerous human, animal and in vitro studies. Ozonation products and
ROS initiate a cascade of events that lead to oxidative stress, injury,
inflammation, airway epithelial damage and increased epithelial damage
and increased alveolar permeability to vascular fluids. Repeated
respiratory inflammation can lead to a chronic inflammatory state with
altered lung structure and lung function and may lead to chronic
respiratory diseases such as fibrosis and emphysema (EPA, 2006a,
section 8.6.2). Continued respiratory inflammation also can alter the
ability to respond to infectious agents, allergens and toxins. Acute
inflammatory responses to O₃ are well documented, and lung
injury can become apparent within 3 hours after exposure in humans.
Taken together, the Criteria Document concludes that the evidence
from experimental human and animal toxicology studies indicates that
acute O₃ exposure is causally associated with respiratory
system effects, including O₃-induced pulmonary function
decrements, respiratory symptoms, lung inflammation, and increased lung
permeability, airway hyperresponsiveness, increased uptake of nonviable
and viable particles, and consequent increased susceptibility to PM-
related toxic effects and respiratory infections (EPA, 2006a, p. 8-48).
ii. Coherence and Plausibility of Effects on the Cardiovascular System
There is very limited experimental evidence of animals and humans
that has evaluated possible mechanisms or physiological pathways by
which acute O₃ exposures may induce cardiovascular system
effects. Ozone induces lung injury, inflammation, and impaired
mucociliary clearance, with a host of associated biochemical changes
all leading to increased lung epithelial permeability. As noted above
in section II.A.2.b, the generation of lipid ozonation products and ROS
in lung tissues can influence pulmonary hemodynamics, and ultimately
the cardiovascular system. Other potential mechanisms by which
O₃ exposure may be associated with cardiovascular disease
outcomes have been described. Laboratory animals exposed to relatively
high O₃ concentrations (>=0.5 ppm) demonstrate tissue edema
in the heart and lungs. Ozone-induced changes in heart rate, edema of
heart tissue, and increased tissue and serum levels of ANF found with
8-hour 0.5 ppm O₃ exposure in animal toxicology studies
(Vesely et al., 1994a, b, c) also raise the possibility of potential
cardiovascular effects of acute ambient O₃ exposures.
Animal toxicology studies have found both transient and persistent
ventilatory responses with and without progressive decreases in heart
rate (Arito et al., 1997). Observations of O₃-induced
vasoconstriction in a controlled human exposure study by Brook et al.
(2002) suggests another possible mechanism for O₃-related
exacerbations of preexisting cardiovascular disease. One controlled
human study (Gong et al., 1998) evaluated potential cardiovascular
health effects of O₃ exposure. The overall results did not
indicate acute cardiovascular effects of O₃ in either the
hypertensive or control subjects. The authors observed an increase in
rate-pressure product and heart rate, a decrement for FEV₁,
and a >10 mm Hg increase in the alveolar/arterial pressure difference
for O₂ following O₃ exposure. Foster et al.
(1993) demonstrated that even in relatively young healthy adults,
O₃ exposure can cause ventilation to shift away from the
well-perfused basal lung. This effect of O₃ on ventilation
distribution may persist beyond 24-hours post-exposure (Foster et al.,
1997). These findings suggest that O₃ may exert
cardiovascular effects indirectly by impairing alveolar-arterial
O₂ transfer and potentially reducing O₂ supply to
the myocardium. Ozone exposure may increase myocardial work and impair
pulmonary gas exchange to a degree that could perhaps be clinically
important in persons with significant preexisting cardiovascular impairment.
As noted above in section II.A.2.b, a limited number of new
epidemiological studies have reported associations between short-term
O₃ exposure and effects on the cardiovascular system. Among
these studies, three were population-based and involved relatively
large cohorts; two of these studies evaluated associations between
O₃ and heart rate variability (HRV) and the other study
evaluated the association between O₃ levels and the relative
risk of myocardial infarction (MI). Such studies may offer more
informative results based on their large subject-pool and design.
Results from these three studies were suggestive of an association
between O₃ exposure and the cardiovascular endpoints
studied. In other recent studies on the incidence of MI and some more
subtle cardiovascular health endpoints, such as changes in HRV or
cardiac arrhythmia, some but not all studies reported associations with
short-term exposure to O₃ (EPA, 2006a, section 7.2.7.1).
From these studies, the Criteria Document concludes that the ``current
evidence is rather limited but suggestive of a potential effect on HRV,
ventricular arrhythmias, and MI incidence'' (EPA, 2006a, p. 7-65).
An increasing number of studies have evaluated the association
between O₃ exposure and cardiovascular hospital admissions.
As discussed in section 7.3.4 of the Criteria Document, many reported
negative or inconsistent associations, whereas other studies,
especially those that examined the relationship when O₃
exposures were higher, have found positive and robust associations
between O₃ and cardiovascular hospital admissions. The
Criteria Document finds that the overall evidence from these studies
remains inconclusive regarding the effect of O₃ on
cardiovascular hospitalizations (EPA, 2006a, p. 7-83).
The Criteria Document notes that the suggestive positive
epidemiologic findings of O₃ exposure on cardiac autonomic
control, including effects on HRV, ventricular arrhythmias and MI, and
reported associations between O₃ exposure and cardiovascular
hospitalizations generally in the warm season gain credibility and
scientific support from the results of experimental animal toxicology
and human clinical studies, which are indicative of plausible pathways
by which O₃ may exert cardiovascular effects (EPA, 2006a,
section 8.6.1).

[[Page 37844]]

iii. Coherence and Plausibility of Effects Related to Long-Term
O₃ Exposure
Human chamber studies can not evaluate effects of long-term
exposures to O₃; there is some evidence available from
toxicological studies. While early animal toxicology studies of long-
term O₃ exposures were conducted using continuous exposures,
more recent studies have focused on exposures which mimic diurnal and
seasonal patterns and more realistic O₃ exposure levels
(EPA, 2006a, p. 8-50). Studies of monkeys that compared these two
exposure scenarios found increased airway pathology only with the
latter design. Persistent and irreversible effects reported in chronic
animal toxicology studies suggest that additional complementary human
data are needed from epidemiologic studies (EPA, 2006a, p. 8-50).
There is limited evidence from human studies for long-term
O₃-induced effects on lung function. As discussed in section
8.6.2 of the Criteria Document, previous epidemiological studies have
provided only inconclusive evidence for either mortality or morbidity
effects of long-term O₃ exposure. The Criteria Document
observes that the inconsistency in findings may be due to a lack of
precise exposure information, the possibility of selection bias, and
the difficulty of controlling for confounders (EPA, 2006a, p. 8-50).
Several new longitudinal epidemiology studies have evaluated
associations between long-term O₃ exposures and morbidity
and mortality and suggest that these long-term exposures may be related
to changes in lung function in children; however, little evidence is
available to support a relationship between chronic O₃
exposure and mortality or lung cancer incidence (EPA, 2006a, p. 8-50).
The Criteria Document (p. 8-51) concludes that evidence from animal
toxicology studies strongly suggests that chronic O₃
exposure is capable of damaging the distal airways and proximal
alveoli, resulting in lung tissue remodeling leading to apparent
irreversible changes. Such structural changes and compromised pulmonary
function caused by persistent inflammation may exacerbate the
progression and development of chronic lung disease. Together with the
limited evidence available from epidemiological studies, these findings
offer some insight into potential biological mechanisms for suggested
associations between long-term or seasonal exposures to O₃
and reduced lung function development in children which have been
observed in epidemiologic studies (EPA, 2006a, p. 8-51).
iv. Coherence and Plausibility of Short-Term Mortality-Related Health
Endpoints
An extensive epidemiological literature on air pollution related
mortality risk estimates from the U.S., Canada, and Europe is discussed
in the Criteria Document (sections 7.4 and 8.6.3). These single- and
multi-city mortality studies coupled with meta-analyses generally
indicate associations between acute O₃ exposure and elevated
risk for all-cause mortality, even after adjustment for the influence
of season and PM. Several single-city studies that specifically
evaluated the relationship between O₃ exposure and
cardiopulmonary mortality also reported results suggestive of a
positive association (EPA, 2006a, p. 8-51). These mortality studies
suggest a pattern of effects for causality that have biologically
plausible explanations, but our knowledge regarding potential
underlying mechanisms is very limited at this time and requires further
research. Most of the physiological and biochemical parameters
investigated in human and animal studies suggest that O₃-
induced biochemical effects are relatively transient and attenuate over
time. The Criteria Document (p. 8-52) hypothesizes a generic pathway of
O₃-induced lung damage, potentially involving oxidative lung
damage with subsequent inflammation and/or decline in lung function
leading to respiratory distress in some sensitive population groups
(e.g., asthmatics), or other plausible pathways noted below that may
lead to O₃-related contributions to cardiovascular effects
that ultimately increase risk of mortality.
The third National Health and Nutrition Examination Follow-up data
analysis indicates that about 20 percent of the adult population has
reduced FEV₁ values, suggesting impaired lung function in
some portion of the population. Most of these individuals have COPD,
asthma or fibrotic lung disease (Manino et al., 2003), which are
associated with persistent low-grade inflammation. Furthermore,
patients with COPD are at increased risk for cardiovascular disease.
Also, lung disease with underlying inflammation may be linked to low-
grade systemic inflammation associated with atherosclerosis,
independent of cigarette smoking (EPA, 2006a, p. 8-52). Lung function
decrements in persons with cardiopulmonary disease have been associated
with inflammatory markers, such as C-reactive protein (CRP) in the
blood. At a population level it has been found that individuals with
the lowest FEV₁ values have the highest levels of CRP, and
those with the highest FEV₁ values have the lowest CRP
levels (Manino et al., 2003; Sin and Man, 2003). This complex series of
physiological and biochemical reactions following O₃
exposure may tilt the biological homeostasis mechanisms which could
lead to adverse health effects in people with compromised
cardiopulmonary systems.
Of much interest are several other types of newly available data
that support reasonable hypotheses that may help to explain the
findings of O₃-related increases in cardiovascular mortality
observed in some epidemiological studies. These include the direct
effect of O₃ on increasing PAF in lung tissue that can then
enter the general circulation and possibly contribute to increased risk
of blood clot formation and the consequent increased risk of MI,
cerebrovascular events (stroke), or associated cardiovascular-related
mortality. Ozone reactions with cholesterol in lung surfactant to form
epoxides and oxysterols that are cytotoxic to lung and heart muscles
and that contribute to atherosclerotic plaque formation in arterial
walls represent another potential pathway. Stimulation of airway
irritant receptors may lead to increases in tissue and serum levels of
ANF, changes in heart rate, and edema of heart tissue. A few new field
and panel studies of human adults have reported associations between
ambient O₃ concentrations and changes in cardiac autonomic
control (e.g., HRV, ventricular arrhythmias, and MI). These represent
plausible pathways that may lead to O₃-related contributions
to cardiovascular effects that ultimately increase the risk of mortality.
In addition, O₃-induced increases in lung permeability
allow more ready entry for inhaled PM into the blood stream, and
O₃ exposure may increase the risk of PM-related
cardiovascular effects. Furthermore, increased ambient O₃
levels contribute to ultrafine PM formation in the ambient air and
indoor environments. Thus, the contributions of elevated ambient
O₃ concentrations to ultrafine PM formation and human
exposure, along with the enhanced uptake of inhaled fine particles,
consequently may contribute to exacerbation of PM-induced
cardiovascular effects in addition to those more directly induced by
O₃ (EPA, 2006a, p. 8-53).
c. Summary
Judgments concerning the extent to which relationships between
various health endpoints and ambient O₃

[[Page 37845]]

exposures are likely causal are informed by the conclusions and
discussion in the Criteria Document as discussed above and summarized
in section 3.7.5 of the Staff Paper. These judgments reflect the nature
of the evidence and overall weight of the evidence, and are taken into
consideration in the quantitative risk assessment discussed below in
section II.B.2.
For example, there is a very high level of confidence that
O₃ induces lung function decrements in healthy adults and
children due in part to the dozens of controlled human exposure and
epidemiological studies consistently showing such effects. The Criteria
Document (p. 8-74) states that these studies provide clear evidence of
causality for associations between short-term O₃ exposures
and statistically significant declines in lung function in children,
asthmatics and adults who exercise outdoors. An increase in respiratory
symptoms (e.g., cough, shortness of breath) has been observed in
controlled human exposure studies of short-term O₃
exposures, and significant associations between ambient O₃
exposures and a wide variety of symptoms have been reported in
epidemiology studies (EPA, 2006a, p. 8-75). Aggregate population time-
series studies showing robust associations with respiratory hospital
admissions and emergency department visits are strongly supported by
human clinical, animal toxicologic, and epidemiologic evidence for
O₃-related lung function decrements, respiratory symptoms,
airway inflammation, and airway hyperreactivity. The Criteria Document
(p. 8-77) concludes that, taken together, the overall evidence supports
the inference of a causal relationship between acute ambient
O₃ exposures and increased respiratory morbidity outcomes
resulting in increased emergency department visits and hospitalizations
during the warm season. Further, recent epidemiologic evidence has been
characterized in the Criteria Document (p. 8-78) as highly suggestive
that O₃ directly or indirectly contributes to non-accidental
and cardiopulmonary-related mortality.
4. O₃-Related Impacts on Public Health
The following discussion draws from chapters 6 and 7 and section
8.7 of the Criteria Document and section 3.6 of the Staff Paper to
characterize factors which modify responsiveness to O₃,
subpopulations potentially at risk for O₃-related health
effects, the adversity of O₃-related effects, and the size
of the at-risk subpopulations in the U.S. These considerations are all
important elements in characterizing the potential public health
impacts associated with exposure to ambient O₃.
a. Factors That Modify Responsiveness to Ozone
There are numerous factors that can modify individual
responsiveness to O₃. These include: influence of physical
activity; age; gender and hormonal influences; racial, ethnic and
socioeconomic status (SES) factors; environmental factors; and oxidant-
antioxidant balance. These factors are discussed in more detail in
section 6.5 of the Criteria Document.
It is well established that physical activity increases an
individual's minute ventilation and will thus increase the dose of
O₃ inhaled (EPA, 2006a, section 6.5.4). Increased physical
activity results in deeper penetration of O₃ into more
distal regions of the lungs, which are more sensitive to acute
O₃ response and injury. This will result in greater lung
function decrements for acute exposures of individuals during increased
physical activity. Research has shown that respiratory effects are
observed at lower O₃ concentrations if the level of exertion
is increased and/or duration of exposure and exertion are extended.
Predicted O₃-induced decrements in lung function have been
shown to be a function of exposure concentration, duration and exercise
level for healthy, young adults (McDonnell et al., 1997).
Most of the studies investigating the influence of age have used
lung function decrements and symptoms as measures of response. For
healthy adults, lung function and symptom responses to O₃
decline as age increases. The rate of decline in O₃
responsiveness appears greater in those 18 to 35 years old compared to
those 35 to 55 years old, while there is very little change after age
55. In one study (Seal et al., 1996) analyzing a large data set, a 5.4%
decrement in FEV1 was estimated for 20 year old individuals exposed to
0.12 ppm O₃, whereas similar exposure of 35 year old
individuals were estimated to have a 2.6% decrement. While healthy
children tend not to report respiratory symptoms when exposed to low
levels of O₃, for subjects 18 to 36 years old symptom
responses induced by O₃ tend to decrease with increasing age
(McDonnell et al., 1999).
Limited evidence of gender differences in response to O₃
exposure has suggested that females may be predisposed to a greater
susceptibility to O₃. Lower plasma and NL fluid levels of
the most prevalent antioxidant, uric acid, in females relative to males
may be a contributing factor. Consequently, reduced removal of
O₃ in the upper airways may promote deeper penetration.
However, most of the evidence on gender differences appears to be
equivocal, with one study (Hazucha et al., 2003) suggesting that
physiological responses of young healthy males and females may be
comparable (EPA, 2006a, section 6.5.2).
A few studies have suggested that ethnic minorities might be more
responsive to O₃ than Caucasian population groups (EPA,
2006a, section 6.5.3). This may be more the result of a lack of
adequate health care and socioeconomic status (SES) than any
differences in sensitivity to O₃. The limited data
available, which have investigated the influence of race, ethnic or
other related factors on responsiveness to O₃, prevent
drawing any clear conclusions at this time.
Few human studies have examined the potential influence of
environmental factors such as the sensitivity of individuals who
voluntarily smoke tobacco (i.e., smokers) and the effect of high
temperatures. New controlled human exposure studies have confirmed that
smokers are less responsive to O₃ than nonsmokers; however,
time course of development and recovery of these effects, as well as
reproducibility, was not different from nonsmokers (EPA, 2006a, section
6.5.5). Influence of ambient temperature on pulmonary effects induced
by O₃ has been studied very little, but additive effects of
heat and O₃ exposure have been reported.
Antioxidants, which scavenge free radicals and limit lipid
peroxidation in the ELF, are the first line of defense against
oxidative stress. Ozone exposure leads to absorption of O₃
in the ELF with subsequent depletion of antioxidant in the nasal ELF,
but concentration and antioxidant enzyme activity in ELF or plasma do
not appear related to O₃ responsiveness (EPA 2006a, section
6.5.6). Controlled studies of dietary antioxidant supplements have
shown some protective effects on lung function decrements but not on
symptoms and airway inflammatory responses. Dietary antioxidant
supplements have provided some protection to asthmatics by attenuating
post-exposure airway hyperresponsiveness. Animal studies have also
supported the protective effects of ELF antioxidants.
b. At-Risk Subgroups for O₃-Related Effects
Several characteristics may increase the extent to which a
population group shows increased susceptibility or vulnerability.
Information on potentially susceptible and vulnerable groups is
summarized in section 8.7 of the

[[Page 37846]]

Criteria Document. As described there, the term susceptibility refers
to innate (e.g., genetic or developmental) or acquired (e.g., personal
risk factors, age) factors that make individuals more likely to
experience effects with exposure to pollutants. A number of population
groups have been identified as potentially susceptible to health
effects as a result of O₃ exposure, including people with
existing lung diseases, including asthma, children and older adults,
and people who have larger than normal lung function responses that may
be due to genetic susceptibility. In addition, some population groups
have been identified as having increased vulnerability to
O₃-related effects due to increased likelihood of exposure
while at elevated ventilation rates, including healthy children and
adults who are active outdoors, for example, outdoor workers, and
joggers. Taken together, the susceptible and vulnerable groups make up
``at-risk'' groups.\22\
---------------------------------------------------------------------------

\22\ In the Staff Paper and documents from previous
O₃ NAAQS reviews, ``at-risk'' groups have also been
called ``sensitive'' groups, to mean both groups with greater
inherent susceptibility and those more likely to be exposed.
---------------------------------------------------------------------------

i. Active People
A large group of individuals at risk from O₃ exposure
consists of outdoor workers and children, adolescents, and adults who
engage in outdoor activities involving exertion or exercise during
summer daylight hours when ambient O₃ concentrations tend to
be higher. This conclusion is based on a large number of controlled-
human exposure studies and several epidemiologic field/panel studies
which have been conducted with healthy children and adults and those
with preexisting respiratory diseases (EPA 2006a, sections 6.2, 6.3,
7.2, and 8.4.4). The controlled human exposure studies show a clear
O₃ exposure-response relationship with increasing
spirometric and symptomatic response as exercise level increases.
Furthermore, O₃-induced response increases as time of
exposure increases. Studies of outdoor workers and others who
participate in outdoor activities indicate that extended exposures to
O₃ at elevated exertion levels can produce marked effects on
lung function, as discussed above in section IIA.2 (Brauer et al.,
1996; H[ouml]ppe et al., 1995; Korrick et al., 1998; McConnell et al.,
2002).
These field studies with subjects at elevated exertion levels
support the extensive evidence derived from controlled human exposure
studies. The majority of human chamber studies have examined the
effects of O₃ exposure in subjects performing continuous or
intermittent exercise for variable periods of time. Significant
O₃-induced respiratory responses have been observed in
clinical studies of exercising individuals. The epidemiologic studies
discussed above also indicate that prolonged exposure periods, combined
with elevated levels of exertion or exercise, may magnify O₃
effects on lung function. Thus, outdoor workers and others who
participate in higher exertion activities outdoors during the time of
day when high peak O₃ concentrations occur appear to be
particularly vulnerable to O₃ effects on respiratory health.
Although these studies show a wide variability of response and
sensitivity among subjects and the factors contributing to this
variability continue to be incompletely understood, the effect of
increased exertion is consistent. It should be noted that this wide
variability of response and sensitivity among subjects may be in part
due to the wide range of other highly reactive photochemical oxidants
coexisting with O₃ in the ambient air.
ii. People With Lung Disease
People with preexisting pulmonary disease are likely to be among
those at increased risk from O₃ exposure. Altered
physiological, morphological and biochemical states typical of
respiratory diseases like asthma, COPD and chronic bronchitis may
render people sensitive to additional oxidative burden induced by
O₃ exposure. At the time of the last review, it was
concluded that this group was at greater risk because the impact of
O₃-induced responses on already-compromised respiratory
systems would noticeably impair an individual's ability to engage in
normal activity or would be more likely to result in increased self-
medication or medical treatment. At that time there was little evidence
that people with pre-existing disease were more responsive than healthy
individuals in terms of the magnitude of pulmonary function decrements
or symptomatic responses. The new results from controlled exposure and
epidemiologic studies continue to indicate that individuals with
preexisting pulmonary disease are a sensitive subpopulation for
O₃ health effects.
Several clinical studies reviewed in the 1996 Criteria Document on
atopic and asthmatic subjects had suggested but not clearly
demonstrated enhanced responsiveness to acute O₃ exposure
compared to healthy subjects. The majority of the newer studies
reviewed in Chapter 6 of the Criteria Document indicate that asthmatics
are as sensitive as, if not more sensitive than, normal subjects in
manifesting O₃-induced pulmonary function decrements. In one
key study (Horstman et al., 1995), the FEV1 decrement observed in the
asthmatics was significantly larger than in the healthy subjects (19%
versus 10%, respectively). There was also a notable tendency for a
greater O₃-induced decrease in FEF_25-75 in
asthmatics relative to the healthy subjects (24% versus 15%,
respectively). A significant positive correlation in asthmatics was
also reported between O₃-induced spirometric responses and
baseline lung function, i.e., responses increased with severity of disease.
Asthmatics present a differential response profile for cellular,
molecular, and biochemical parameters (Criteria Document, Figure 8-1)
that are altered in response to acute O₃ exposure. Ozone-
induced increases in neutrophils, IL-8 and protein were found to be
significantly higher in the BAL fluid from asthmatics compared to
healthy subjects, suggesting mechanisms for the increased sensitivity
of asthmatics (Basha et al., 1994; McBride et al., 1994; Scannell et
al., 1996; Hiltermann et al., 1999; Holz et al., 1999; Bosson et al.,
2003). Neutrophils, or PMNs, are the white blood cell most associated
with inflammation. IL-8 is an inflammatory cytokine with a number of
biological effects, primarily on neutrophils. The major role of this
cytokine is to attract and activate neutrophils. Protein in the airways
is leaked from the circulatory system, and is a marker for increased
cellular permeability.
Bronchial constriction following provocation with O₃
and/or allergens presents a two-phase response. The early response is
mediated by release of histamine and leukotrienes that leads to
contraction of smooth muscle cells in the bronchi, narrowing the lumen
and decreasing the airflow. In people with allergic airway disease,
including people with rhinitis and asthma, these mediators also cause
accumulation of eosinophils in the airways (Bascom et al., 1990; Jorres
et al., 1996; Peden et al., 1995 and 1997; Frampton et al., 1997a;
Michelson et al., 1999; Hiltermann et al., 1999; Holz et al., 2002;
Vagaggini et al., 2002). In asthma, the eosinophil, which increases
inflammation and allergic responses, is the cell most frequently
associated with exacerbations of the disease. A study by Bosson et al.
(2003) evaluated the difference in O₃-induced bronchial
epithelial cytokine expression between healthy and asthmatic subjects.
After O₃ exposure the epithelial expression of IL-5 and GM-
CSF increased significantly in

[[Page 37847]]

asthmatics, compared to healthy subjects. Asthma is associated with
Th2-related airway response (allergic response), and IL-5 is an
important Th2-related cytokine. The O₃-induced increase in
IL-5, and also in GM-CSF, which affects the growth, activation and
survival of eosinophils, may indicate an effect on the Th2-related
airway response and on airway eosinophils. The authors reported that
the O₃-induced Th2-related cytokine responses that were
found within the asthmatic group may indicate a worsening of their
asthmatic airway inflammation and thus suggest a plausible link to
epidemiological data indicating O₃-associated increases in
bronchial reactivity and hospital admissions.
The accumulation of eosinophils in the airways of asthmatics is
followed by production of mucus and a late-phase bronchial constriction
and reduced airflow. In a study of 16 intermittent asthmatics,
Hiltermann et al. (1999) found that there was a significant inverse
correlation between the O₃-induced change in the percentage
of eosinophils in induced sputum and the change in PC₂₀, the
concentration of methacholine causing a 20% decrease in
FEV₁. Characteristic O₃-induced inflammatory
airway neutrophilia at one time was considered a leading mechanism of
airway hyperresponsiveness. However, Hiltermann et al. (1999)
determined that the O₃-induced change in percentage
neutrophils in sputum was not significantly related to the change in
PC₂₀. These results are consistent with the results of Zhang
et al. (1995), which found neutrophilia in a murine model to be only
coincidentally associated with airway hyperresponsiveness, i.e., there
was no cause and effect relationship. (Criteria Document, AX 6-26).
Hiltermann et al. (1999) concluded that the results point to the role
of eosinophils in O₃-induced airway hyperresponsiveness.
Increases in O₃-induced nonspecific airway responsiveness
incidence and duration could have important clinical implications for
asthmatics.
Two studies (J[ouml]rres et al., 1996; Holz et al., 2002) observed
increased airway responsiveness to O₃ exposure with
bronchial allergen challenge in subjects with preexisting allergic
airway disease. J[ouml]rres et al. (1996) found that O₃
causes an increased response to bronchial allergen challenge in
subjects with allergic rhinitis and mild allergic asthma. The subjects
were exposed to 0.25 ppm O₃ for 3 hours with IE. Airway
responsiveness to methacholine was determined 1 hour before and after
exposure; responsiveness to allergen was determined 3 hours after
exposure. Statistically significant decreases in FEV₁
occurred in subjects with allergic rhinitis (13.8%) and allergic asthma
(10.6%), and in healthy controls (7.3%). Methacholine responsiveness
was statistically increased in asthmatics, but not in subjects with
allergic rhinitis or healthy controls. Airway responsiveness to an
individual's historical allergen (either grass and birch pollen, house
dust mite, or animal dander) was significantly increased after
O₃ exposure when compared to FA exposure. In subjects with
asthma and allergic rhinitis, a maximum percent fall in FEV₁
of 27.9% and 7.8%, respectively, occurred 3 days after O₃
exposure when they were challenged with of the highest common dose of
allergen. The authors concluded that subjects with asthma or allergic
rhinitis, without asthma, could be at risk if a high O₃
exposure is followed by a high dose of allergen. Holz et al. (2002)
reported an early phase lung function response in subjects with
rhinitis after a consecutive 4-day exposure to 0.125 ppm O₃
that resulted in a clinically relevant (>20%) decrease in FEV1. Ozone-
induced exacerbation of airway responsiveness persists longer and
attenuates more slowly than O₃-induced lung function
decrements and respiratory symptom responses and can have important
clinical implications for asthmatics.
A small number of in vitro studies corroborate the differences in
the responses of asthmatic and healthy subject generally found in
controlled human exposure studies. In vitro studies (Schierhorn et al.,
1999) of nasal mucosal biopsies from atopic and nonatopic subjects
exposed to 0.1 ppm O₃ found significant differences in
release of IL-4, IL-6, IL-8, and TNF-[alpha]. Another study by
Schierhorn et al. (2002) found significant differences in the
O₃-induced release of the neuropeptides neurokinin A and
substance P for allergic patients in comparison to nonallergic
controls, suggesting increased activation of sensory nerves by
O₃ in the allergic tissues. Another study by Bayram et al.
(2002) using in vitro culture of bronchial epithelial cells recovered
from atopic and nonatopic asthmatics also found significant increases
in epithelial permeability in response to O₃ exposure.
The new data on airway responsiveness, inflammation, and various
molecular markers of inflammation and bronchoconstriction indicate that
people with asthma and allergic rhinitis (with or without asthma)
comprise susceptible groups for O₃-induced adverse effects.
This body of evidence indicates that human clinical and epidemiological
panel studies of lung function decrements and respiratory symptoms that
evaluate only healthy, non-asthmatic subjects likely underestimate the
effects of O₃ exposure on asthmatics and other susceptible
populations. The effects of O₃ on lung function,
inflammation, and increased airway responsiveness demonstrated in
subjects with asthma and other allergic airway diseases, provide
plausible mechanisms underlying the more serious respiratory morbidity
effects, such as emergency department visits and hospital admissions,
and respiratory mortality effects.
A number of epidemiological studies have been conducted using
asthmatic study populations. The majority of epidemiological panel
studies that evaluated respiratory symptoms and medication use related
to O₃ exposures focused on children. These studies suggest
that O₃ exposure may be associated with increased
respiratory symptoms and medication use in children with asthma. Other
reported effects include respiratory symptoms, lung function
decrements, and emergency department visits, as discussed in the
Criteria Document (section 7.6.7.1). Strong evidence from a large
multi-city study (Mortimer et al., 2002), along with support from
several single-city studies suggest that O₃ exposure may be
associated with increased respiratory symptoms and medication use in
children with asthma. With regard to ambient O₃ levels and
increased hospital admissions and emergency department visits for
asthma and other respiratory causes, strong and consistent evidence
establishes a correlation between O₃ exposure and increased
exacerbations of preexisting respiratory disease for 1-hour maximum
O₃ concentrations <0.12 ppm. As discussed in the Criteria
Document, section 7.3, several hospital admission and emergency
department visit studies in the U.S., Canada, and Europe have reported
positive associations between increase in O₃ and increased
risk of emergency department visits and hospital admissions for asthma
and other respiratory diseases, especially during the warm season.
Finally, from epidemiological studies that included assessment of
associations with specific causes of death, some studies have observed
larger effects estimates for respiratory mortality and others have
observed larger effects estimates for cardiovascular mortality. The
apparent inconsistency regarding the effect size of O₃-
related respiratory mortality may be due to reduced statistical power
in this

[[Page 37848]]

subcategory of mortality (EPA, 2006a, p. 7-108).
Newly available reports from controlled human exposure studies (see
chapter 6 in the Criteria Document) utilized subjects with preexisting
cardiopulmonary diseases such as COPD, asthma, allergic rhinitis, and
hypertension. The data generated from these studies that evaluated
changes in spirometry did not find clear differences between filtered
air and O₃ exposure in COPD subjects. However, the new data
on airway responsiveness, inflammation, and various molecular markers
of inflammation and bronchoconstriction indicate that people with
atopic asthma and allergic rhinitis comprise susceptible groups for
O₃-induced adverse health effects.
Although controlled human exposure studies have not found evidence
of larger spirometric changes in people with COPD relative to healthy
subjects, this may be due to the fact that most people with COPD are
older adults who would not be expected to have such changes based on
their age. However, in section 8.7.1, the Criteria Document notes that
new epidemiological evidence indicates that people with COPD may be
more likely to experience other effects, including emergency room
visits, hospital admissions, or premature mortality. For example,
results from an analysis of five European cities indicated strong and
consistent O₃ effects on unscheduled respiratory hospital
admissions, including COPD (Anderson et al., 1997). Also, an analysis
of a 9-year data set for the whole population of the Netherlands
provided risk estimates for more specific causes of mortality,
including COPD (Hoek et al., 2000, 2001; reanalysis, Hoek, 2003); a
positive, but nonsignificant, excess risk of COPD-related mortality was
found to be associated with short-term O₃ concentrations.
Moreover, as indicated by Gong et al. (1998), the effects of
O₃ exposure on alveolar-arterial oxygen gradients may be
more pronounced in patients with preexisting obstructive lung diseases.
Relative to healthy elderly subjects, COPD patients have reduced gas
exchange and low SaO2. Any inflammatory or edematous responses due to
O₃ delivered to the well-ventilated regions of the COPD lung
could further inhibit gas exchange and reduce oxygen saturation. In
addition, O₃-induced vasoconstriction could also acutely
induce pulmonary hypertension. Inducing pulmonary vasoconstriction and
hypertension in these patients would perhaps worsen their condition,
especially if their right ventricular function was already compromised
(EPA, 2006a, section 6.10).
iii. Children and Older Adults
Supporting evidence exists for heterogeneity in the effects of
O₃ by age. As discussed in section 6.5.1 of the Criteria
Document, children, adolescents, and young adults (<18 yrs of age)
appear, on average, to have nearly equivalent spirometric responses to
O₃, but have greater responses than middle-aged and older
adults when exposed to comparable O₃ doses. Symptomatic
responses to O₃ exposure, however, do not appear to occur in
healthy children, but are observed in asthmatic children, particularly
those who use maintenance medications. For adults (>17 yrs of age)
symptoms gradually decrease with increasing age. In contrast to young
adults, the diminished symptomatic responses in children and the
diminished symptomatic and spirometric responses in older adults
increases the likelihood that these groups continue outdoor activities
leading to greater O₃ exposure and dose.
As described in the section 7.6.7.2 of the Criteria Document, many
epidemiological field studies focused on the effect of O₃ on
the respiratory health of school children. In general, children
experienced decrements in pulmonary function parameters, including PEF,
FEV₁, and FVC. Increases in respiratory symptoms and asthma
medication use were also observed in asthmatic children. In one German
study, children with and without asthma were found to be particularly
susceptible to O₃ effects on lung function. Approximately
20% of the children, both with and without asthma, experienced a
greater than 10% change in FEV₁, compared to only 5% of the
elderly population and athletes (Hoppe et al., 2003).
The American Academy of Pediatrics (2004) notes that children and
infants are among the population groups most susceptible to many air
pollutants, including O₃. This is in part because their
lungs are still developing. For example, eighty percent of alveoli are
formed after birth, and changes in lung development continue through
adolescence (Dietert et al., 2000). Children are also likely to spend
more time outdoors than adults, which results in increased exposure to
air pollutants (Wiley et al., 1991a,b). Moreover, children have high
minute ventilation rates and high levels of physical activity which
also increases their dose (Plunkett et al., 1992).
Several mortality studies have investigated age-related differences
in O₃ effects (EPA, 2006a, section 7.6.7.2). Older adults
are also often classified as being particularly susceptible to air
pollution. The basis for increased O₃ sensitivity among the
elderly is not known, but one hypothesis is that it may be related to
changes in the respiratory tract lining fluid antioxidant defense
network (Kelly et al., 2003). (EPA 2006a, p. 8-60) Older adults have
lower baseline lung function than younger people, and are also more
likely to have preexisting lung and heart disease. Increased
susceptibility of older adults to O₃ health effects is most
clearly indicated in the newer mortality studies. Among the studies
that observed positive associations between O₃ and
mortality, a comparison of all age or younger age (<=65 years of age)
O₃-mortality effect estimates to that of the elderly
population (>65 years) indicates that, in general, the elderly
population is more susceptible to O₃ mortality effects. The
meta-analysis by Bell et al. (2005) found a larger mortality effect
estimate for the elderly than for all ages. In the large U.S. 95
communities study (Bell et al., 2004), mortality effect estimates were
slightly higher for those aged 65 to 74 years, compared to individuals
less than 65 years and 75 years or greater. The absolute effect of
O₃ on premature mortality may be substantially greater in
the elderly population because of higher rates of preexisting
respiratory and cardiac diseases. The Criteria Document concludes that
the elderly population (>65 years of age) appear to be at greater risk
of O₃-related mortality and hospitalizations compared to all
ages or younger populations (EPA, 2006a, p. 7-177).
The Criteria Document notes that, collectively, there is supporting
evidence of age-related differences in susceptibility to O₃
lung function effects. The elderly population (>65 years of age) appear
to be at increased risk of O₃-related mortality and
hospitalizations, and children (<18 years of age) experience other
potentially adverse respiratory health outcomes with increased
O₃ exposure (EPA, 2006a, section 7.6.7.2).
iv. People With Increased Responsiveness to Ozone
New animal toxicology studies using various strains of mice and
rats have identified O₃-sensitive and resistant strains and
illustrated the importance of genetic background in determining
O₃ susceptibility (EPA, 2006a, section 8.7.4). Controlled
human exposure studies have also indicated a high degree of variability
in some of the pulmonary physiological parameters. The variable effects
in individuals have

[[Page 37849]]

been found to be reproducible, in other words, a person who has a large
lung function response after exposure to O₃ will likely have
about the same response if exposed again to the same dose of
O₃. In human clinical studies, group mean responses are not
representative of this segment of the population that has much larger
than average responses to O₃. Recent studies of asthmatics
by David et al. (2003) and Romieu et al. (2004) reported a role for
genetic polymorphism in observed differences in antioxidant enzymes and
genes involved in inflammation to modulate pulmonary function and
inflammatory responses to O₃ exposure.\23\
---------------------------------------------------------------------------

\23\ Similar to animal toxicology studies referred above, a
polymorphism in a specific proinflammatory cytokine gene has been
implicated in O₃-induced lung function changes in
healthy, mild asthmatics and individuals with rhinitis. These
observations suggest a potential role for these markers in the
innate susceptibility to O₃, however, the validity of
these markers and their relevance in the context of prediction to
population studies needs additional experimentation.
---------------------------------------------------------------------------

Biochemical and molecular parameters extensively evaluated in these
experiments were used to identify specific loci on chromosomes and, in
some cases, to relate the differential expression of specific genes to
biochemical and physiological differences observed among these species.
Utilizing O₃-sensitive and O₃-resistant species,
it has been possible to identify the involvement of increased airway
reactivity and inflammation processes in O₃ susceptibility.
However, most of these studies were carried out using relatively high
doses of O₃, making the relevance of these studies
questionable in human health effects assessment. The genes and genetic
loci identified in these studies may serve as useful biomarkers and,
ultimately, can likely be integrated with epidemiological studies.
v. Other Population Groups
There is limited, new evidence supporting associations between
short-term O₃ exposures and a range of effects on the
cardiovascular system. Some but not all, epidemiological studies have
reported associations between short-term O₃ exposures and
the incidence of MI and more subtle cardiovascular health endpoints,
such as changes in HRV and cardiac arrhythmia. Others have reported
associations with hospitalization or emergency department visits for
cardiovascular diseases, although the results across the studies are
not consistent. Studies also report associations between short-term
O₃ exposure and mortality from cardiovascular or
cardiopulmonary causes. The Criteria Document concludes that current
cardiovascular effects evidence from some field studies is rather
limited but supportive of a potential effect of short-term
O₃ exposure and HRV, cardiac arrhythmia, and MI incidence
(EPA, 2006a, p. 7-65). In the Criteria Document's evaluation of studies
of hospital admissions for cardiovascular disease (EPA 2006a, section
7.3.4), it is concluded that evidence from this growing group of
studies is generally inconclusive regarding an association with
O₃ in studies conducted during the warm season (EPA 2006a,
p. 7-83). This body of evidence suggests that people with heart disease
may be at increased risk from short-term exposures to O₃;
however, more evidence is needed to conclude that people with heart
disease are a susceptible population.
Other groups that might have enhanced sensitivity to O₃,
but for which there is currently very little evidence, include groups
based on race, gender and SES, and those with nutritional deficiencies,
which presents factors which modify responsiveness to O₃.
c. Adversity of Effects
In making judgments as to when various O₃-related
effects become regarded as adverse to the health of individuals, the
Administrator has looked to guidelines published by the American
Thoracic Society (ATS) and the advice of CASAC. While recognizing that
perceptions of ``medical significance'' and ``normal activity'' may
differ among physicians, lung physiologists and experimental subjects,
the ATS (1985) \24\ defined adverse respiratory health effects as
``medically significant physiologic changes generally evidenced by one
or more of the following: (1) Interference with the normal activity of
the affected person or persons, (2) episodic respiratory illness, (3)
incapacitating illness, (4) permanent respiratory injury, and/or (5)
progressive respiratory dysfunction.'' During the 1997 review, it was
concluded that there was evidence of causal associations from
controlled human exposure studies for effects in the first of these
five ATS-defined categories, evidence of statistically significant
associations from epidemiological studies for effects in the second and
third categories, and evidence from animal toxicology studies, which
could be extrapolated to humans only with a significant degree of
uncertainty, for the last two categories.
---------------------------------------------------------------------------

\24\ In 2000, the American Thoracic Society (ATS) published an
official statement on ``What Constitutes an Adverse Health Effect of
Air Pollution?'' (ATS, 2000), which updated its earlier guidance
(ATS, 1985). Overall, the new guidance does not fundamentally change
the approach previously taken to define adversity, nor does it
suggest a need at this time to change the structure or content of
the tables describing gradation of severity and adversity of effects
described below.
---------------------------------------------------------------------------

For ethical reasons, clear causal evidence from controlled human
exposure studies still covers only effects in the first category.
However, for this review there are results from epidemiological
studies, upon which to base judgments about adversity, for effects in
all of the categories. Statistically significant and robust
associations have been reported in epidemiology studies falling into
the second and third categories. These more serious effects include
respiratory events (e.g., triggering asthma attacks) that may require
medication (e.g., asthma), but not necessarily hospitalization, as well
as respiratory hospital admissions and emergency department visits for
respiratory causes. Less conclusive, but still positive associations
have been reported for school absences and cardiovascular hospital
admissions. Human health effects for which associations have been
suggested through evidence from epidemiological and animal toxicology
studies, but have not been conclusively demonstrated still fall
primarily into the last two categories. In the last review of the
O₃ standard, evidence for these more serious effects came
from studies of effects in laboratory animals. Evidence from animal
studies evaluated in this Criteria Document strongly suggests that
O₃ is capable of damaging the distal airways and proximal
alveoli, resulting in lung tissue remodeling leading to apparently
irreversible changes. Recent advancements of dosimetry modeling also
provide a better basis for extrapolation from animals to humans.
Information from epidemiological studies provides supporting, but
limited evidence of irreversible respiratory effects in humans than was
available in the prior review. Moreover, the findings from single-city
and multi-city time-series epidemiology studies and meta-analyses of
these epidemiology studies are highly suggestive of an association
between short-term O₃ exposure and mortality particularly in
the warm season.
While O₃ has been associated with effects that are
clearly adverse, application of these guidelines, in particular to the
least serious category of effects related to ambient O₃
exposures, involves judgments about which medical experts on the CASAC
panel and public commenters have expressed diverse views in the past.
To help frame such judgments, EPA staff have defined specific ranges of
functional responses

[[Page 37850]]

(e.g., decrements in FEV₁ and airway responsiveness) and
symptomatic responses (e.g., cough, chest pain, wheeze), together with
judgments as to the potential impact on individuals experiencing
varying degrees of severity of these responses, that have been used in
previous NAAQS reviews. These ranges of pulmonary responses and their
associated potential impacts are summarized in Tables 3-2 and 3-3 of
the Staff Paper.
For active healthy people, moderate levels of functional responses
(e.g., FEV₁ decrements of >=10% but <20%, lasting up to 24
hours) and/or moderate symptomatic responses (e.g., frequent
spontaneous cough, marked discomfort on exercise or deep breath,
lasting up to 24 hours) would likely interfere with normal activity for
relatively few responsive individuals. On the other hand, EPA staff
determined that large functional responses (e.g., FEV₁
decrements >=20%, lasting longer than 24 hours) and/or severe
symptomatic responses (e.g., persistent uncontrollable cough, severe
discomfort on exercise or deep breath, lasting longer than 24 hours)
would likely interfere with normal activities for many responsive
individuals. EPA staff determined that these would be considered
adverse under ATS guidelines. In the context of standard setting, CASAC
indicated that a focus on the mid to upper end of the range of moderate
levels of functional responses (e.g., FEV₁ decrements >=15%
but <20%) is appropriate for estimating potentially adverse lung
function decrements in active healthy people. However, for people with
lung disease, even moderate functional (e.g., FEV₁
decrements >=10% but <20%, lasting up to 24 hours) or symptomatic
responses (e.g., frequent spontaneous cough, marked discomfort on
exercise or with deep breath, wheeze accompanied by shortness of
breath, lasting up to 24 hours) would likely interfere with normal
activity for many individuals, and would likely result in more frequent
use of medication. For people with lung disease, large functional
responses (e.g., FEV₁ decrements >=20%, lasting longer than
24 hours) and/or severe symptomatic responses (e.g., persistent
uncontrollable cough, severe discomfort on exercise or deep breath,
persistent wheeze accompanied by shortness of breath, lasting longer
than 24 hours) would likely interfere with normal activity for most
individuals and would increase the likelihood that these individuals
would seek medical treatment. In the context of standard setting, the
CASAC indicated (Henderson, 2006c) that a focus on the lower end of the
range of moderate levels of functional responses (e.g., FEV₁
decrements >=10%) is most appropriate for estimating potentially
adverse lung function decrements in active healthy people.
In judging the extent to which these impacts represent effects that
should be regarded as adverse to the health status of individuals, an
additional factor that has been considered in previous NAAQS reviews is
whether such effects are experienced repeatedly during the course of a
year or only on a single occasion. While some experts would judge
single occurrences of moderate responses to be a ``nuisance,''
especially for healthy individuals, a more general consensus view of
the adversity of such moderate responses emerges as the frequency of
occurrence increases.
The new guidance builds upon and expands the 1985 definition of
adversity in several ways. There is an increased focus on quality of
life measures as indicators of adversity. There is also a more specific
consideration of population risk. Exposure to air pollution that
increases the risk of an adverse effect to the entire population is
adverse, even though it may not increase the risk of any individual to
an unacceptable level. For example, a population of asthmatics could
have a distribution of lung function such that no individual has a
level associated with significant impairment. Exposure to air pollution
could shift the distribution to lower levels that still do not bring
any individual to a level that is associated with clinically relevant
effects. However, this would be considered to be adverse because
individuals within the population would have diminished reserve
function, and therefore would be at increased risk if affected by
another agent.
Of the various effects of O₃ exposure that have been
studied, many would meet the ATS definition of adversity. Such effects
include, for example, any detectible level of permanent lung function
loss attributable to air pollution, including both reductions in lung
growth or acceleration of the age-related decline of lung function;
exacerbations of disease in individuals with chronic cardiopulmonary
diseases; reversible loss of lung function in combination with the
presence of symptoms; as well as more serious effects such as those
requiring medical care including hospitalization and, obviously, mortality.
d. Size of At-Risk Subpopulations
Although O₃-related health risk estimates may appear to
be small, their significance from an overall public health perspective
is determined by the large numbers of individuals in the subpopulations
potentially at-risk for O₃-related health effects discussed
above. For example, a population of concern includes people with
respiratory disease, including approximately 11 percent of U.S. adults
and 13 percent of children who have been diagnosed with asthma and 6
percent of adults with chronic obstructive pulmonary disease (chronic
bronchitis and/or emphysema) in 2002 and 2003 (Table 8-4 in the
Criteria Document, section 8.7.5.2). More broadly, individuals with
preexisting cardiopulmonary disease may constitute an additional
population of concern, with potentially tens of millions of people
included in each disease category. In addition, subpopulations based on
age group also comprise substantial segments of the population that may
be potentially at risk for O₃-related health impacts. Based
on U.S. census data from 2003, about 26 percent of the U.S. population
are under 18 years of age and 12 percent are 65 years of age or older.
Hence, large proportions of the U.S. population are included in age
groups include those most likely to have increased susceptibility to
the health effects of O₃ and or those with the highest
ambient O₃ exposures.
The Criteria Document (section 8.7.5.2) notes that the health
statistics data illustrate what is known as the ``pyramid'' of effects.
At the top of the pyramid, there are approximately 2.5 millions deaths
from all causes per year in the U.S. population, with about 100,000
deaths from chronic lower respiratory diseases. For respiratory health
diseases, there are nearly 4 million hospital discharges per year, 14
million emergency department visits, 112 million ambulatory care
visits, and an estimated 700 million restricted activity days per year
due to respiratory conditions from all causes per year. Applying small
risk estimates for the O₃-related contribution to such
health effects with relatively large baseline levels of health outcomes
can result in quite large public health impacts related to ambient
O₃ exposure. Thus, even a small percentage reduction in
O₃ health impacts on cardiopulmonary diseases would reflect
a large number of avoided cases. In considering this information
together with the concentration-response relationships that have been
observed between exposure to O₃ and various health
endpoints, the Criteria

[[Page 37851]]

Document (section 8.7.5.2) concludes that exposure to ambient O₃
likely has a significant impact on public health in the U.S.

B. Human Exposure and Health Risk Assessments

To put judgments about health effects that are adverse for
individuals into a broader public health context, EPA has developed and
applied models to estimate human exposures and health risks. This
broader context includes consideration of the size of particular
population groups at risk for various effects, the likelihood that
exposures of concern will occur for individuals in such groups under
varying air quality scenarios, estimates of the number of people likely
to experience O₃-related effects, the variability in
estimated exposures and risks, and the kind and degree of uncertainties
inherent in assessing the exposures and risks involved.
As discussed below there are a number of important uncertainties
that affect the exposure and health risk estimates. It is also
important to note that there have been significant improvements in both
the exposure and health risk model. CASAC expressed the view that the
exposure analysis represents a state-of-the-art modeling approach and
that the health risk assessment was ``well done, balanced and
reasonably communicated'' (Henderson, 2006c). While recognizing and
considering the kind and degree of uncertainties in both the exposure
and health risk estimates, the Staff Paper judged that the quality of
the estimates is such that they are suitable to be used as an input to
the Administrator's decisions on the O₃ primary standard
(Staff Paper, p. 6-20--6-21).
In modeling exposures and health risks associated with just meeting
the current and alternative O₃ standards, EPA has simulated
air quality to represent conditions just meeting these standards based
on O₃ air quality patterns in several recent years and on
how the shape of the O₃ air quality distribution has changed
over time based on historical trends in monitored O₃ air
quality data. As described in the Staff Paper (section 4.5.8) and
discussed below, recent O₃ air quality distributions have
been statistically adjusted to simulate just meeting the current and
selected alternative standards. These simulations do not reflect any
consideration of specific control programs or strategies designed to
achieve the reductions in emissions required to meet the specified
standards. Further, these simulations do not represent predictions of
when, whether, or how areas might meet the specified standards.\25\
---------------------------------------------------------------------------

\25\ Modeling that projects whether and how areas might attain
alternative standards in a future year is presented in the Regulatory
Impact Analysis being prepared in connection with this rulemaking.
---------------------------------------------------------------------------

As noted in section I.C above, around the time of the release of
the final Staff Paper in January 2007, EPA discovered a small error in
the exposure model that when corrected resulted in slight increases in
the simulated exposures. Since the exposure estimates are an input to
the lung function portion of the health risk assessment, this
correction also resulted in slight increases in the lung function risk
estimates as well. The exposure and risk estimates discussed in this
notice reflect the corrected estimates, and thus are slightly different
than the exposure and risk estimates cited in the January 31, 2007
Staff Paper.\26\
---------------------------------------------------------------------------

\26\ EPA plans to make available corrected versions of the final
Staff Paper, and human exposure and health risk assessment technical
support documents on or around July 16, 2007 on the EPA web site
listed in the Availability of Related Information section of this notice.
---------------------------------------------------------------------------

1. Exposure Analyses
a. Overview
The EPA conducted exposure analyses using a simulation model to
estimate O₃ exposures for the general population, school age
children (ages 5-18), and school age children with asthma living in 12
U.S. metropolitan areas representing different regions of the country
where the current 8-hour O₃ standard is not met. The
emphasis on children reflects the finding of the last O₃
NAAQS review that children are an important at-risk group. The 12
modeled areas combined represent a significant fraction of the U.S.
urban population, 89 million people, including 18 million school age
children of whom approximately 2.6 million have asthma. The selection
of urban areas to include in the exposure analysis took into
consideration the location of O₃ epidemiological studies,
the availability of ambient O₃ data, and the desire to
represent a range of geographic areas, population demographics, and
O₃ climatology. These selection criteria are discussed
further in chapter 5 of the Staff Paper. The geographic extent of each
modeled area consists of the census tracts in the combined statistical
area (CSA) as defined by OMB (OMB, 2005).\27\
---------------------------------------------------------------------------

\27\ The 12 CSAs modeled are: Atlanta-Sandy Springs-Gainesville,
GA-AL; Boston-Worcester-Manchester, MA-NH; Chicago-Naperville-
Michigan City, IL-IN-WI; Cleveland-Akron-Elyria, OH; Detroit-Warren-
Flint, MI; Houston-Baytown-Huntsville, TX; Los Angeles-Long Beach-
Riverside, CA; New York-Newark-Bridgeport, NY-NJ-CT-PA;
Philadelphia-Camden-Vineland, PA-NJ-DE-MD; Sacramento-Arden-Arcade-
Truckee, CA-NV; St. Louis-St. Charles-Farmington, MO-IL; Washington-
Baltimore-N. Virginia, DC-MD-VA-WV.
---------------------------------------------------------------------------

Exposure estimates were developed using a probabilistic exposure
model that is designed to explicitly model the numerous sources of
variability that affect people's exposures. As discussed below, the
model estimates population exposures by simulating human activity
patterns, air conditioning prevalence, air exchange rates, and other
factors. The modeled exposure estimates were developed for three recent
years of ambient O₃ concentrations (2002, 2003, and 2004),
as well as for O₃ concentrations adjusted to simulate
conditions associated with just meeting the current NAAQS and various
alternative 8-hour standards based on the three year period 2002-
2004.\28\ This exposure assessment is more fully described and
presented in the Staff Paper and in a technical support document, Ozone
Population Exposure Analysis for Selected Urban Areas (US EPA, 2006b;
hereafter Exposure Analysis TSD). The scope and methodology for this
exposure assessment were developed over the last few years with
considerable input from the CASAC Ozone Panel and the public.\29\
---------------------------------------------------------------------------

\28\ All 12 of the CSAs modeled did not meet the current
O₃ NAAQS for the three year period examined.
\29\ The general approach used in the current exposure
assessment was described in the draft Health Assessment Plan (EPA,
2005a) that was released to the CASAC and general public in April
2005 and was the subject of a consultation with the CASAC
O₃ Panel on May 5, 2005. In October 2005, OAQPS released
the first draft of the Staff Paper containing a chapter discussing
the exposure analyses and first draft of the Exposure Analyses TSD
for CASAC consultation and public review on December 8, 2005. In
July 2006, OAQPS released the second draft of the Staff Paper and
second draft of the Exposure Analyses TSD for CASAC review and
public comment which was held by the CASAC O₃ Panel on
August 24-25, 2006.
---------------------------------------------------------------------------

The goals of the O₃ exposure assessment were: (1) To
provide estimates of the size of at-risk populations exposed to various
levels associated with recent O₃ concentrations, and with
just meeting the current O₃ NAAQS and alternative
O₃ standards, in specific urban areas; (2) to provide
distributions of exposure estimates over the entire range of ambient
O₃ concentrations as an important input to the lung function
risk assessment summarized below in section II.B.2; (3) to develop a
better understanding of the influence of various inputs and assumptions
on the exposure estimates; and (4) to gain insight into the
distribution of exposures and patterns of exposure

[[Page 37852]]

reductions associated with meeting alternative O₃ standards.
EPA recognizes that there are many sources of variability and
uncertainty inherent in the inputs to this assessment and that there is
uncertainty in the resulting O₃ exposure estimates. With
respect to variability, the exposure modeling approach accounts for
variability in ambient O₃ levels, demographic
characteristics, physiological attributes, activity patterns, and
factors affecting microenvironmental (e.g., indoor) concentrations. In
EPA's judgment, the most important uncertainties affecting the exposure
estimates are related to the modeling of human activity patterns over
an O₃ season, the modeling of variations in ambient
concentrations near roadways, and the modeling of air exchange rates
that affect the amount of O₃ that penetrates indoors.
Another important uncertainty that affects the estimation of how many
exposures are associated with moderate or greater exertion, is the
characterization of energy expenditure for children engaged in various
activities. As discussed in more detail in the Staff Paper (section
4.3.4.7), the uncertainty in energy expenditure values carries over to
the uncertainty of the modeled breathing rates, which are important
since they are used to classify exposures occurring at moderate or
greater exertion which are the relevant exposures since O₃-
related effects observed in clinical studies only are observed when
individuals are engaged in some form of exercise. The uncertainties in
the exposure model inputs and the estimated exposures have been
assessed using quantitative uncertainty and sensitivity analyses.
Details are discussed in the Staff Paper (section 4.6) and in a
technical memorandum describing the exposure modeling uncertainty
analysis (Langstaff, 2007).
b. Scope and Key Components
Population exposures to O₃ are primarily driven by
ambient outdoor concentrations, which vary by time of day, location,
and peoples' activities. Outdoor O₃ concentration estimates
used in the exposure assessment are provided by measurements and
statistical adjustments to the measured concentrations. The current
exposure analysis allows comparisons of population exposures to
O₃ within each urban area, associated with current
O₃ levels and with O₃ levels just meeting several
potential alternative air quality standards or scenarios. Human
exposure, regardless of the pollutant, depends on where individuals are
located and what they are doing. Inhalation exposure models are useful
in realistically estimating personal exposures to O₃ based
on activity-specific breathing rates, particularly when recognizing
that large scale population exposure measurement studies have not been
conducted that are representative of the overall population or at-risk
subpopulations.
The model EPA used to simulate O₃ population exposure is
the Air Pollutants Exposure Model (APEX), the human inhalation exposure
model within the Total Risk Integrated Methodology (TRIM) framework
(EPA, 2006c,d). APEX is conceptually based on the probabilistic NAAQS
exposure model for O₃ (pNEM/O₃) used in the last
O₃ NAAQS review. Since that time, the model has been
restructured, improved, and expanded to reflect conceptual advances in
the science of exposure modeling and newer input data available for the
model. Key improvements to algorithms include replacement of the cohort
approach with a probabilistic sampling approach focused on individuals,
accounting for fatigue and oxygen debt after exercise in the
calculation of breathing rates, and a new approach for construction of
longitudinal activity patterns for simulated persons. Major
improvements to data input to the model include updated air exchange
rates, more recent census and commuting data, and a greatly expanded
daily time-activities database.
APEX is a probabilistic model designed to explicitly model the
numerous sources of variability that affect people's exposures. APEX
simulates the movement of individuals through time and space and
estimates their exposures to O₃ in indoor, outdoor, and in-
vehicle microenvironments. The exposure model takes into account the
most significant factors contributing to total human O₃
exposure, including the temporal and spatial distribution of people and
O₃ concentrations throughout an urban area, the variation of
O₃ levels within each microenvironment, and the effects of
exertion on breathing rate in exposed individuals. A more detailed
description of APEX and its application is presented in chapter 4 of
the Staff Paper and associated technical documents (EPA, 2006b, c, d).
Several methods have been used to evaluate the APEX model and to
characterize the uncertainty of the model estimates. These include
conducting model evaluation, sensitivity analyses, and a detailed
uncertainty analysis for one urban area. These are discussed fully in
the Staff Paper (section 4.6) and in Langstaff (2007). The uncertainty
of model structure was judged to be of lesser importance than the
uncertainties of the model inputs and parameters. Model structure
refers to the algorithms in APEX designed to simulate the processes
that result in people's exposures, for example, the way that APEX
models exposures to individuals when they are near roads. The
uncertainties in the model input data (e.g., measurement error, ambient
concentrations, air exchange rates, and activity pattern data) have
been assessed individually, and their impact on the uncertainty in the
modeled exposure estimates was assessed in a unified quantitative
analysis with results expressed in the form of estimated confidence
ranges around the estimated measures of exposure. This uncertainty
analysis was conducted for one urban area (Boston) using the observed
2002 O₃ concentrations and 2002 concentrations adjusted to
simulate just meeting the current standard, with the expectation that
the results would be similar for other cities and years. One
significant source of uncertainty, due to limitations in the database
used to model peoples' daily activities, was not included in the
unified analysis, and was assessed through separate sensitivity
analyses. This analysis indicates that the uncertainty of the exposure
results is relatively small. For example, 95 percent uncertainty
intervals were calculated for the APEX estimates of the percent of
children or asthmatic children with exposures above 0.060, 0.070, or
0.080 ppm under moderate exertion, for two air quality scenarios
(current 2002 and 2002 adjusted to simulate just meeting the current
standard) in Boston (Tables 26 and 27 in Langstaff, 2007). The 95
percent uncertainty intervals for this set of 12 exposure estimates
indicate the possibility of underpredictions of the exposure estimates
ranging from 3 to 25 percent of the modeled estimates, and
overpredictions ranging from 4 to 11 percent of the estimates. For
example, APEX estimates the percent of asthmatic children with
exposures above 0.070 ppm under moderate exertion to be 24 percent, for
Boston 2002 O₃ concentrations adjusted to simulate just
meeting the current standard. The 95 percent uncertainty interval for
this estimate is 23-30 percent, or -4 to +25 percent of the estimate.
These uncertainty intervals do not include the uncertainty engendered
by limitations of the activity database, which is in the range of one
to ten percent.
The exposure periods modeled here are the O₃ seasons in
2002, 2003, and

[[Page 37853]]

2004. The O₃ season in each area includes the period of the
year where elevated O₃ levels tend to be observed and for
which routine hourly O₃ monitoring data are available.
Typically this period spans from March or April through September or
October, or in some areas, spanning the entire year. Three years were
modeled to reflect the substantial year-to-year variability that occurs
in O₃ levels and related meteorological conditions, and
because the standard is specified in terms of a three-year period. The
year-to-year variability observed in O₃ levels is due to a
combination of different weather patterns and the variation in
emissions of O₃ precursors. Nationally, 2002 was a
relatively high year with respect to the 4th highest daily maximum 8-
hour O₃ levels observed in urban areas across the U.S. (EPA,
2007, Figure 2-16), with the mean of the distribution of O₃
levels for the urban monitors being in the upper third among the years
1990 through 2006. In contrast, on a national basis, 2004 is the lowest
year on record through 2006 for this same air quality statistic, and 8-
hour daily maximum O₃ levels observed in most, but not all
of the 12 urban areas included in the exposure and risk analyses were
relatively low compared to other recent years. The 4th highest daily
maximum 8-hour O₃ levels observed in 2003 in the 12 urban
areas and nationally generally were between those observed in 2002 and
2004.
Regulatory scenarios examined include the current 0.08 ppm, average
of the 4th daily maximum 8-hour averages over a three year period
standard; standards with the same form but with alternative levels of
0.080, 0.074, 0.070, and 0.064 ppm; standards specified as the average
of the 3rd highest daily maximum 8-hour averages over a three year
period with alternative levels of 0.084 and 0.074 ppm; and a standard
specified as the average of the 5th highest daily maximum 8-hour
averages over a three year period with a level of 0.074 ppm.\30\ The
current standard uses a rounding convention that allows areas to have
an average of the 4th daily maximum 8-hour averages as high as 0.084
ppm and still meet the standard. All alternative standards analyzed
were intended to reflect improved precision in the measurement of
ambient concentrations, where the precision would extend to three
instead of two decimal places (in ppm).
---------------------------------------------------------------------------

\30\ The current O₃ standard is 0.08 ppm, but the
current rounding convention specifies that the average of the 4th
daily maximum 8-hour average concentrations over a three-year period
must be at 0.084 ppm or lower to be in attainment of the standard.
When EPA staff selected alternative standards to analyze, it was
presumed that the same type of rounding convention would be used,
and thus alternative standards of 0.084, 0.074, 0.064 ppm were chosen.
---------------------------------------------------------------------------

The current standard and all alternative standards were modeled
using a quadratic rollback approach to adjust the hourly concentrations
observed in 2002-2004 to yield a design value \31\ corresponding to the
standard being analyzed. The quadratic rollback technique reduces
higher concentrations more than lower concentrations near ambient
background levels.\32\ This procedure was considered in a sensitivity
analysis in the last review of the O₃ standard and has been
shown to be more realistic than a linear, proportional rollback method,
where all of the ambient concentrations are reduced by the same factor.
---------------------------------------------------------------------------

\31\ A design value is a statistic that describes the air
quality status of a given area relative to the level of the NAAQS.
Design values are often based on multiple years of data, consistent
with specification of the NAAQS in Part 50 of the CFR. For the
current O₃ NAAQS, the 3-year average of the annual 4th-
highest daily maximum 8-hour average concentrations, based on the
monitor within (or downwind of) an urban area yielding the highest
3-year average, is the design value.
\32\ The quadratic rollback approach and evaluation of this
approach are described by Johnson (1997), Duff et al. (1998) and
Rizzo (2005, 2006).
---------------------------------------------------------------------------

c. Exposure Estimates and Key Observations
The exposure assessment, which provides estimates of the number of
people exposed to different levels of ambient O₃ while at
specified exertion levels \33\ serve two purposes. First, the entire
range of modeled personal exposures to ambient O₃ is an
essential input to the portion of the health risk assessment based on
exposure-response functions from controlled human exposure studies,
discussed in the next section. Second, estimates of personal exposures
to ambient O₃ concentrations at and above specific benchmark
levels provide some perspective on the public health impacts of health
effects that we cannot currently evaluate in quantitative risk
assessments that may occur at current air quality levels, and the
extent to which such impacts might be reduced by meeting the current
and alternative standards. This is especially true when there are
exposure levels at which we know or can reasonably infer that specific
O₃-related health effects are occurring. We refer to exposures
at and above these benchmark concentrations as ``exposures of concern.''
---------------------------------------------------------------------------

\33\ As discussed above in Section II.A., O₃ health
responses observed in human clinical studies are associated with
exposures while engaged in moderate or greater exertion and,
therefore, these are the exposure measures of interest. The level of
exertion of individuals engaged in particular activities is measured
by an equivalent ventilation rate (EVR), ventilation normalized by
body surface area (BSA, in m\2\), which is calculated as VE/BSA,
where VE is the ventilation rate (liters/minute). Moderate and
greater exertion levels were defined as EVR > 13 liters/min-m\2\
(Whitfield et al., 1996) to correspond to the exertion levels
measured in most subjects studied in the controlled human exposure
studies that reported health effects associated with 6.6 hour
O₃ exposures.
---------------------------------------------------------------------------

EPA emphasizes that, although the analysis of ``exposures of
concern'' was conducted using three discrete benchmark levels (i.e.,
0.080, 0.070, and 0.060 ppm), the concept is more appropriately viewed
as a continuum with greater confidence and less uncertainty about the
existence of health effects at the upper end and less confidence and
greater uncertainty as one considers increasingly lower O₃
exposure levels. EPA recognizes that there is no sharp breakpoint
within the continuum ranging from at and above 0.080 ppm down to 0.060
ppm. In considering the concept of exposures of concern, it is
important to balance concerns about the potential for health effects
and their severity with the increasing uncertainty associated with our
understanding of the likelihood of such effects at lower O₃
levels.
Within the context of this continuum, estimates of exposures of
concern at discrete benchmark levels provide some perspective on the
public health impacts of O₃-related health effects that have
been demonstrated in human clinical and toxicological studies but
cannot be evaluated in quantitative risk assessments, such as lung
inflammation, increased airway responsiveness, and changes in host
defenses. They also help in understanding the extent to which such
impacts have the potential to be reduced by meeting the current and
alternative standards. In the selection of specific benchmark
concentrations for this analysis, we first considered the exposure
level of 0.080 ppm, at which there is a substantial amount of clinical
evidence demonstrating a range of O₃-related health effects
including lung inflammation and airway responsiveness in healthy
individuals. Thus, as in the last review, this level was selected as a
benchmark level for this assessment of exposures of concern. Evidence
newly available in this review is the basis for identifying additional,
lower benchmark levels of 0.070 and 0.060 ppm for this assessment.
More specifically, as discussed above in section II.A.2, evidence
available from controlled human exposure and epidemiology studies
indicates that people with asthma have larger and more serious effects
than healthy individuals, including lung function, respiratory
symptoms, increased airway

[[Page 37854]]

responsiveness, and pulmonary inflammation, which has been shown to be
a more sensitive marker than lung function responses. Further, a
substantial new body of evidence from epidemiology studies shows
associations with serious respiratory morbidity and cardiopulmonary
mortality effects at O₃ levels that extend below 0.080 ppm.
Additional, but very limited new evidence from controlled human
exposure studies shows lung function decrements and respiratory
symptoms in healthy subjects at an O₃ exposure level of
0.060 ppm. The selected benchmark level of 0.070 ppm reflects the new
information that asthmatics have larger and more serious effects than
healthy people and therefore controlled human exposure studies done
with healthy subjects may underestimate effects in this group, as well
as the substantial body of epidemiological evidence of associations
with O₃ levels below 0.080 ppm. The selected benchmark level
of 0.060 ppm additionally reflects the very limited new evidence from
controlled human exposure studies that show lung function decrements
and respiratory symptoms in some healthy subjects at the 0.060 ppm
exposure level, recognizing that asthmatics are likely to have more
serious responses and that lung function is not likely to be as
sensitive a marker for O₃ effects as is lung inflammation.
The estimates of exposures of concern were reported in terms of
both ``people exposed'' (the number and percent of people who
experience a given level of O₃ concentrations, or higher, at
least one time during the O₃ season in a given year) and
``occurrences of exposure'' (the number of times a given level of
pollution is experienced by the population of interest, expressed in
terms of person-days of occurrences). Estimating exposures of concern
is important because it provides some indication of the potential
public health impacts of a range of O₃-related health
outcomes, such as lung inflammation, increased airway responsiveness,
and changes in host defenses. These particular health effects have been
demonstrated in controlled human exposure studies of healthy
individuals to occur at levels as low as 0.080 ppm O₃, but
have not been evaluated at lower levels in controlled human exposure
studies. EPA has not included these effects in the quantitative risk
assessment due to a lack of adequate information on the exposure-
response relationships.
The 1997 O₃ NAAQS review estimated exposures associated
with 1-hour heavy exertion, 1-hour moderate exertion, and 8-hour
moderate exertion for children, outdoor workers, and the general
population. EPA's analysis in the 1997 Staff Paper showed that exposure
estimates based on the 8-hour moderate exertion scenario for children
yielded the largest number of children experiencing exposures at or
above exposures of concern. Consequently, EPA has chosen to focus on
the 8-hour moderate and greater exertion exposures in all and asthmatic
school age children in the current exposure assessment. While outdoor
workers and other adults who engage in moderate or greater exertion for
prolonged durations while outdoors during the day in areas experiencing
elevated O₃ concentrations also are at risk for experiencing
exposures associated with O₃-related health effects, EPA did
not focus on quantitative estimates for these populations due to the
lack of information about the number of individuals who regularly work
or exercise outdoors. Thus, the exposure estimates presented here and
in the Staff Paper are most useful for making relative comparisons
across alternative air quality scenarios and do not represent the total
exposures in all children or other groups within the general population
associated with the air quality scenarios.
Population exposures to O₃ were estimated in 12 urban
areas for 2002, 2003, and 2004 air quality, and also using
O₃ concentrations adjusted to just meet the current and
several alternative standards. The estimates of 8-hour exposures of
concern at and above benchmark levels of 0.080, 0.070, and 0.060 ppm
aggregated across all 12 areas are shown in Table 1 for air quality
scenarios just meeting the current and four alternative 8-hour average
standards.\34\ Table 1 provides estimates of the number and percent of
school age children and asthmatic school age children exposed, with
daily 8-hour maximum exposures at or above each O₃ benchmark
level of exposures of concern, while at intermittent moderate or
greater exertion and based on O₃ concentrations observed in
2002 and 2004. Table 1 summarizes estimates for 2002 and 2004, because
these years reflect years that bracket relatively higher and lower
O₃ levels, with year 2003 generally containing O₃
levels in between when considering the 12 urban areas modeled. This
table also reports the percent change in the number of persons exposed
when a given alternative standard is compared with the current standard.
---------------------------------------------------------------------------

\34\ The full range of quantitative exposure estimates
associated with just meeting the current and alternative
O₃ standards are presented in chapter 4 and Appendix 4A
of the Staff Paper.
---------------------------------------------------------------------------

Key observations important in comparing exposure estimates
associated with just meeting the current NAAQS and alternative
standards under consideration include:
(1) As shown in Table 6-1 of the Staff Paper, the patterns of
exposure in terms of percentages of the population exceeding a given
exposure level are very similar for the general population and for
asthmatic and all school age (5-18) children, although children are
about twice as likely to be exposed, based on the percent of the
population exposed, at any given level.
(2) As shown in Table 1 below, the number and percentage of
asthmatic and all school-age children aggregated across the 12 urban
areas estimated to experience 1 or more exposures of concern decline
from simulations of just meeting the current standard to simulations of
alternative 8-hour standards by varying amounts depending on the
benchmark level, the population subgroup considered, and the year
chosen. For example, the estimated percentage of school age children
experiencing one or more exposures >= 0.070 ppm, while engaged in
moderate or greater exertion, during an O₃ season is about
18 percent of this population when the current standard is met using
the 2002 simulation; this is reduced to about 12, 4, 1, and 0.2 percent
of children upon meeting alternative standards of 0.080, 0.074, 0.070,
and 0.064 ppm, respectively (all specified in terms of the 4th highest
daily maximum 8-hour average), using the 2002 simulation.

[[Page 37855]]

Table 1.--Number and Percent of All and Asthmatic School Age Children in 12 Urban Areas Estimated to Experience
8-Hour Ozone Exposures Above 0.080, 0.070, and 0.060 ppm While at Moderate or Greater Exertion, One or More
Times Per Season and the Number of Occurrences Associated with Just Meeting Alternative 8-Hour Standards Based
on Adjusting 2002 and 2004 Air Quality Data\1, 2\
----------------------------------------------------------------------------------------------------------------
All children, ages 5-18 aggregate for Asthmatic children, ages 5-18
8-Hour air 12 urban areas, number of children Aggregate for 12 urban areas, number
Benchmark levels of quality exposed (% of all) [%reduction from of children exposed (% of group) [%
exposures of standards\3\ current standard]
reduction from current standard]
concern (ppm) (ppm) ----------------------------------------------------------------------------
2002 2004 2002 2004
----------------------------------------------------------------------------------------------------------------
0.080.............. 0.084 700,000 (4%)...... 30,000 (0%)...... 110,000 (4%)..... 0 (0%)
0.080 290,000 (2%) [70%]
10,000 (0%) [67%] 50,000 (2%) [54%] 0 (0%)
0.074 60,000 (0%) [91%]. 0 (0%) [100%].... 10,000 (0%) [91%]
0 (0%)
0.070 10,000 (0%) [98%]. 0 (0%) [100%].... 0 (0%) [100%].... 0 (0%)
0.064 0 (0%) [100%]..... 0 (0%) [100%].... 0 (0%) [100%].... 0 (0%)
----------------------------------------------------------------------------------------------------------------
0.070.............. 0.084 3,340,000 (18%)... 260,000 (1%)..... 520,000 (20%).... 40,000 (1%)
0.080 2,160,000 (12%) 100,000 (1%) 330,000 (13%) 10,000 (0%) [75%]
[35%]. [62%]. [36%].
0.074 770,000 (4%) [77%]
20,000 (0%) [92%] 120,000 (5%) 0 (0%) [100%]
[77%].
0.070 270,000 (1%) [92%]
0 (0%) [100%].... 50,000 (2%) [90%] 0 (0%) [100%]
0.064 30,000 (0.2%) 0 (0%) [100%].... 10,000 (0.2%) 0 (0%) [100%]
[99%]. [98% ].
----------------------------------------------------------------------------------------------------------------
0.060.............. 0.084 7,970,000 (44%)... 1,800,000 (10%).. 1,210,000 (47%).. 270,000 (11%)
0.080 6,730,000 (37%) 1,050,000 (6%) 1,020,000 (40%) 150,000 (6%)
[16%]. [42%]. [16%]. [44%]
0.074 4,550,000 (25%) 350,000 (2%) 700,000 (27%) 50,000 (2%) [81%]
[43%]. [80%]. [42%].
0.070 3,000,000 (16%) 110,000 (1%) 460,000 (18%) 10,000 (1%) [96%]
[62%]. [94%]. [62%].
0.064 950,000 (5%) [88%]
10,000 (0%) [99%] 150,000 (6%) 0 (0%) [100%]
[88%].
----------------------------------------------------------------------------------------------------------------
\1\ Moderate or greater exertion is defined as having an 8-hour average equivalent ventilation rate >= 13 l-min/
m\2\.
\2\ Estimates are the aggregate results based on 12 combined statistical areas (Atlanta, Boston, Chicago,
Cleveland, Detroit, Houston, Los Angeles, New York, Philadelphia, Sacramento, St. Louis, and Washington, DC).
Estimates are for the ozone season which is all year in Houston, Los Angeles and Sacramento and March or April
to September or October for the remaining urban areas.
\3\ All standards summarized here have the same form as the current 8-hour standard which is specified as the 3-
year average of the annual 4th highest daily maximum 8-hour average concentrations must be at or below the
concentration level specified. As described in the Staff Paper (section 4.5.8), recent O3 air quality
distributions have been statistically adjusted to simulate just meeting the current and selected alternative
standards. These simulations do not represent predictions of when, whether, or how areas might meet the
specified standards.

(3) Substantial year-to-year variability in exposure estimates is
observed over the three-year modeling period. For example, the
estimated number of school age children experiencing one or more
exposures >=0.070 ppm during an O₃ season when the current
standard is met in the 12 urban areas included in the analysis is 3.3,
1.0, or 0.3 million for the 2002, 2003, and 2004 simulations,
respectively.
(4) There is substantial variability observed across the 12 urban
areas in the percent of the population subgroups estimated to
experience exposures of concern. For example, when 2002 O₃
concentrations are simulated to just meet the current standard, the
aggregate 12 urban area estimate is 18 percent of all school age
children are estimated to experience O₃ exposures (>=0.070
ppm (Table 1 below), while the range of exposure estimates in the 12
urban areas considered separately for all children range from 1 to 38
percent (EPA, 2007, Exhibit 2, p. 4-48). There was also variability in
exposure estimates among the modeled areas when using the 2004 air
quality simulation for the same scenario; however it was reduced and
ranged from 0 to 7 percent in the 12 urban areas (EPA, 2007, Exhibit 8,
p. 4-60).
(5) Of particular note, as discussed above in section II.A. of this
notice, high inter-individual variability in responsiveness means that
only a subset of individuals in these groups who are exposed at and
above a given benchmark level would actually be expected to experience
such adverse health effects.
(6) In considering these observations, it is important to take into
account the variability, uncertainties, and limitations associated with
this assessment, including the degree of uncertainty associated with a
number of model inputs and uncertainty in the model itself, as
discussed above.
2. Quantitative Health Risk Assessment
This section discusses the approach used to develop quantitative
health risk estimates associated with exposures to O₃
building upon a more limited risk assessment that was conducted during
the last review.\35\ As part of the last review, EPA conducted a health
risk assessment that produced risk estimates for the number and percent
of children and outdoor workers experiencing lung function and
respiratory symptoms associated with O₃ exposures for 9
urban areas.\36\ The risk assessment for the last review also included
risk estimates for excess respiratory-related hospital admissions
related to O₃ concentrations for New York City. In the last
review, the risk estimates played a significant role in both the staff
recommendations and in the proposed and final decisions to revise the
O₃ standards. The health risk assessment conducted for the
current review builds upon the methodology and lessons learned from the
prior review.
---------------------------------------------------------------------------

\35\ The methodology, scope, and results from the risk
assessment conducted in the last review are described in Chapter 6
of the 1996 Staff Paper (EPA, 1996) and in several technical reports
(Whitfield et al., 1996; Whitfield, 1997) and publication (Whitfield
et al., 1998).
\36\ The 9 urban study areas included in the exposure and risk
analyses conducted during the last review were: Chicago, Denver,
Houston, Los Angeles, Miami, New York City, Philadephia, St. Louis,
and Washington, DC.
---------------------------------------------------------------------------

a. Overview
The updated health risk assessment conducted as part of this review
includes estimates of (1) Risks of lung function decrements in all and
asthmatic school age children, respiratory symptoms in asthmatic
children, respiratory-related hospital admissions, and non-accidental
and cardiorespiratory-related mortality associated with recent ambient
O₃ levels; (2) risk reductions and remaining

[[Page 37856]]

risks associated with just meeting the current 8-hour O₃
NAAQS; and (3) risk reductions and remaining risks associated with just
meeting various alternative 8-hour O₃ NAAQS in a number of
example urban areas. This risk assessment is more fully described and
presented in the Staff Paper (EPA, 2007, chapter 5) and in a technical
support document (TSD), Ozone Health Risk Assessment for Selected Urban
Areas (Abt Associates, 2006, hereafter referred to as ``Risk Assessment
TSD''). The scope and methodology for this risk assessment were
developed over the last few years with considerable input from the
CASAC O₃ Panel and the public.\37\ The information contained
in these documents included specific criteria for the selection of
health endpoints, studies, and locations to include in the assessment.
In a peer review letter sent by CASAC to the Administrator documenting
its advice in October 2006 (Henderson, 2006c), the CASAC O₃
Panel concluded that the risk assessment was ``well done, balanced, and
reasonably communicated'' and that the selection of health endpoints
for inclusion in the quantitative risk assessment was appropriate.
---------------------------------------------------------------------------

\37\ The general approach used in the current risk assessment
was described in the draft Health Assessment Plan (EPA, 2005a) that
was released to the CASAC and general public in April 2005 and was
the subject of a consultation with the CASAC O₃ Panel on
May 5, 2005. In October 2005, OAQPS released the first draft of the
Staff Paper containing a chapter discussing the risk assessment and
first draft of the Risk Assessment TSD for CASAC consultation and
public review on December 8, 2005. In July 2006, OAQPS released the
second draft of the Staff Paper and second draft of the Risk
Assessment TSD for CASAC review and public comment which was held by
the CASAC O₃ Panel on August 24-25, 2006.
---------------------------------------------------------------------------

The goals of the risk assessment are: (1) To provide estimates of
the potential magnitude of several morbidity effects and mortality
associated with current O₃ levels, and with meeting the
current and alternative 8-hour O₃ standards in specific
urban areas; (2) to develop a better understanding of the influence of
various inputs and assumptions on the risk estimates; and (3) to gain
insights into the distribution of risks and patterns of risk reductions
associated with meeting alternative O₃ standards. The health
risk assessment is intended to be dependent on and reflect the overall
weight and nature of the health effects evidence discussed above in
section II.A and in more detail in the Criteria Document and Staff
Paper. While not independent of the overall evaluation of the health
effects evidence, the quantitative health risk assessment provides
additional insights regarding the relative public health implications
associated with just meeting the current and several alternative 8-hour
standards.
The risk assessment covers a variety of health effects for which
there is adequate information to develop quantitative risk estimates.
However, as noted by CASAC (Henderson, 2007) and in the Staff Paper,
there are a number of health endpoints (e.g., increased lung
inflammation, increased airway responsiveness, impaired host defenses,
increased medication usage for asthmatics, increased emergency
department visits for respiratory causes, and increased school
absences) for which there currently is insufficient information to
develop quantitative risk estimates, but which are important to
consider in assessing the overall public health impacts associated with
exposures to O₃. These additional health endpoints are
discussed above in section II.A.2 and are also taken into account in
considering the level of exposures of concern in populations
particularly at risk, discussed above in this notice.
There are two parts to the health risk assessment: one based on
combining information from controlled human exposure studies with
modeled population exposure and the other based on combining
information from community epidemiological studies with either
monitored or adjusted ambient concentrations levels. Both parts of the
risk assessment were implemented within a new probabilistic version of
TRIM.Risk, the component of EPA's Total Risk Integrated Methodology
(TRIM) model framework that estimates human health risks.
EPA recognizes that there are many sources of uncertainty and
variability in the inputs to this assessment and that there is
significant variability and uncertainty in the resulting O₃
risk estimates. As discussed in chapters 2, 5, and 6 of the Staff
Paper, there is significant year-to-year and city-to-city variability
related to the air quality data that affects both the controlled human
exposure studies-based and epidemiological studies-based parts of the
risk assessment. There are also uncertainties associated with the air
quality adjustment procedure used to simulate just meeting the current
and selected alternative standards In the prior review, different
statistical approaches using alternative functional forms (i.e.,
quadratic, proportional, Weibull) were used to reflect how
O₃ air quality concentrations have historically changed.
Based on sensitivity analyses conducted in the prior review, the choice
of alternative air quality adjustment procedures had only a modest
impact on the risk estimates (EPA, 2007, p. 6-20). With respect to
uncertainties about estimated background concentrations, as discussed
below and in the Staff Paper (EPA 2007b, section 5.4.3), alternative
assumptions about background levels have a variable impact depending on
the location, standard, and health endpoint analyzed.
With respect to the lung function part of the health risk
assessment, key uncertainties include uncertainties in the exposure
estimates, discussed above, and uncertainties associated with the shape
of the exposure-response relationship, especially at levels below 0.08
ppm, 8-hour average, where only very limited data are available down to
0.04 ppm and there is an absence of data below 0.04 ppm (EPA, 2007, pp.
6-20--6-21). Concerning the part of the risk assessment based on
effects reported in epidemiological studies, important uncertainties
include uncertainties (1) Surrounding estimates of the O₃
coefficients for concentration-response relationships used in the
assessment, (2) involving the shape of the concentration-response
relationship and whether or not a population threshold or non-linear
relationship exists within the range of concentrations examined in the
studies, (3) related to the extent to which concentration-response
relationships derived from studies in a given location and time when
O₃ levels were higher or behavior and /or housing conditions
were different provide accurate representations of the relationships
for the same locations with lower air quality distributions and/or
different behavior and/or housing conditions, and (4) concerning the
possible role of co-pollutants which also may have varied between the
time of the studies and the current assessment period. An important
additional uncertainty for the mortality risk estimates is the extent
to which the associations reported between O₃ and non-
accidental and cardiorespiratory mortality actually reflect causal
relationships.
As discussed below, some of these uncertainties have been addressed
quantitatively in the form of estimated confidence ranges around
central risk estimates; others are addressed through separate
sensitivity analyses (e.g., the influence of alternative estimates for
policy-relevant background levels) or are characterized qualitatively.
For both parts of the health risk assessment, statistical uncertainty
due to sampling error has been characterized and is expressed in terms
of 95 percent credible intervals. EPA recognizes that these credible
intervals do not reflect all of the uncertainties noted above.

[[Page 37857]]

b. Scope and Key Components
The current health risk assessment is based on the information
evaluated in the final Criteria Document. The risk assessment includes
several categories of health effects and estimates risks associated
with just meeting the current and alternative 8-hour O₃
NAAQS and with several individual recent years of air quality (i.e.,
2002, 2003, and 2004). The risk assessment considers the same
alternative air quality scenarios that were examined in the human
exposure analyses described above. Risk estimates were developed for up
to 12 urban areas selected to illustrate the public health impacts
associated with these air quality scenarios.\38\ As discussed above in
section II.B.1, the selection of urban areas was largely determined by
identifying areas in the U.S. which represented a range of geographic
areas, population demographics, and climatology; with an emphasis on
areas that do not meet the current 8-hour O₃ NAAQS and which
included the largest areas with O₃ nonattainment problems.
The selection criteria also included whether or not there were
acceptable epidemiological studies available that reported
concentration-response relationships for the health endpoints selected
for inclusion in the assessment.
---------------------------------------------------------------------------

\38\ The 12 urban areas are the same urban areas evaluated in
the exposure analysis discussed in the prior section. However, for
most of the health endpoints based on findings from epidemiological
studies, the geographic areas and populations examined in the health
risk assessment were limited to those counties included in the
original epidemiological studies that served as the basis for the
concentration-response relationships.
---------------------------------------------------------------------------

The short-term exposure related health endpoints selected for
inclusion in the quantitative risk assessment include those for which
the final Criteria Document and or Staff Paper concluded that the
evidence as a whole supports the general conclusion that O₃,
acting alone and/or in combination with other components in the ambient
air pollution mix, is either clearly causal or is judged to be likely
causal. Some health effects met this criterion of likely causality, but
were not included in the risk assessment for other reasons, such as
insufficient exposure-response data or lack of baseline incidence data.
As discussed in the section above describing the exposure analysis,
in order to estimate the health risks associated with just meeting the
current and alternative 8-hour O₃ NAAQS, it is necessary to
estimate the distribution of hourly O₃ concentrations that
would occur under any given standard. Since compliance is based on a 3-
year average, the amount of control has been applied to each year of
data (i.e., 2002 to 2004) to estimate risks for a single O₃
season or single warm O₃ season, depending on the health
effect, based on a simulation that adjusted each of these individual
years so that the three year period would just meet the specified
standard.
Consistent with the risk assessment approach used in the last
review, the risk estimates developed for both recent air quality levels
and just meeting the current and selected alternative 8-hour standards
represent risks associated with O₃ levels attributable to
anthropogenic sources and activities (i.e., risk associated with
concentrations above ``policy-relevant background''). Policy-relevant
background O₃ concentrations used in the O₃ risk
assessment were defined in chapter 2 of the Staff Paper (EPA, 2007, pp.
2-48--2-55) as the O₃ concentrations that would be observed
in the U.S. in the absence of anthropogenic emissions of precursors
(e.g., VOC, NO_X, and CO) in the U.S., Canada, and Mexico.
The results of a global tropospheric O₃ model (GEOS-CHEM)
have been used to estimate monthly background daily diurnal profiles
for each of the 12 urban areas for each month of the O₃
season using meteorology for the year 2001. Based on the results of the
GEOS-CHEM model, the Criteria Document indicates that background
O₃ concentrations are generally predicted to be in the range
of 0.015 to 0.035 ppm in the afternoon, and they are generally lower
under conditions conducive to man-made O₃ episodes.\39\
---------------------------------------------------------------------------

\39\ EPA notes that the estimated level of policy-relevant
background O₃ used in the prior risk assessment was a
single concentration of 0.04 ppm, which was the midpoint of the
range of levels for policy-relevant background that was provided in
the 1996 Criteria Document.
---------------------------------------------------------------------------

This approach of estimating risks in excess of background is judged
to be more relevant to policy decisions regarding ambient air quality
standards than risk estimates that include effects potentially
attributable to uncontrollable background O₃ concentrations.
Sensitivity analyses examining the impact of alternative estimates for
background on lung function and mortality risk estimates have been
developed and are included in the Staff Paper and Risk Assessment TSD
and key observations are discussed below. Further, CASAC noted the
difficulties and complexities associated with available approaches to
estimating policy-relevant background concentrations (Henderson, 2007).
Recognizing these complexities, EPA requests comments on the new
approach used in this review for estimating these levels as an input to
the health risk assessment.\40\
---------------------------------------------------------------------------

\40\ Recognizing the importance of this issue, EPA intends to
conduct additional sensitivity analyses related to policy-relevant
background and its implications for the risk assessment.
---------------------------------------------------------------------------

In the first part of the current risk assessment, lung function
decrement, as measured by FEV₁, is the only health response
that is based on data from controlled human exposure studies. As
discussed above, there is clear evidence of a causal relationship
between lung function decrements and O₃ exposures for school
age children engaged in moderate exertion based on numerous controlled
human exposure and summer camp field studies conducted by various
investigators. Risk estimates have been developed for O₃-
related lung function decrements (measured as changes in
FEV₁) for all school age children (ages 5 to 18) and a
subset of this group, asthmatic school age children (ages 5 to 18),
whose average exertion over an 8-hour period was moderate or greater.
The exposure period and exertion level were chosen to generally match
the exposure period and exertion level used in the controlled human
exposure studies that were the basis for the exposure-response
relationships. A combined data set including individual level data from
the Folinsbee et al. (1988), Horstman et al. (1990), and McDonnell et
al. (1991) studies, used in the previous risk assessment, and more
recent data from Adams (2002, 2003, 2006) have been used to estimate
probabilistic exposure-response relationships for 8-hour exposures
under different definitions of lung function response (i.e., >=10, 15,
and 20 percent decrements in FEV₁). As discussed in the
Staff Paper (EPA, 2007, p. 5-27), while these specific controlled human
exposure studies only included healthy adults aged 18-35, findings from
other controlled human exposure studies and summer camp field studies
involving school age children in at least six different locations in
the northeastern United States, Canada, and Southern California
indicated changes in lung function in healthy children similar to those
observed in healthy adults exposed to O₃ under controlled
chamber conditions.
Consistent with advice from CASAC (Henderson, 2006c), EPA has
considered both linear and logistic functional forms in estimating the
probabilistic exposure-response relationships for lung function
responses. A Bayesian Markov Chain Monte Carlo approach, described in
more detail in the Risk Assessment TSD, has been used that incorporates
both model uncertainty and uncertainty due

[[Page 37858]]

to sample size in the combined data set that served as the basis for
the assessment. EPA has chosen a model reflecting a 90 percent
weighting on a logistic form and a 10 percent weighting on a linear
form as the base case for the current risk assessment. The basis for
this choice is that the logistic form provides a very good fit to the
combined data set, but a linear model cannot be entirely ruled out
since there are only very limited data (i.e., 30 subjects) at the two
lowest exposure levels (i.e., 0.040 and 0.060 ppm). EPA has conducted a
sensitivity analysis which examines the impact on the lung function
risk estimates of two alternative choices, an 80 percent logistic/20
percent linear split and a 50 percent logistic/50 percent linear split.
As noted above, risk estimates have been developed for three
measures of lung function response (i.e., >=10, 15, and 20 percent
decrements in FEV₁). However, the Staff Paper and risk
estimates summarized below focus on FEV₁ decrements >=15
percent for all school age children and >=10 percent for asthmatic
school age children, consistent with the advice from CASAC (Henderson,
2006c) that these levels of response represent indicators of adverse
health effects in these populations. The Risk Assessment TSD and Staff
Paper present the broader range of risk estimates including all three
measures of lung function response.
Developing risk estimates for lung function decrements involved
combining probabilistic exposure-response relationships based on the
combined data set from several controlled human exposure studies with
population exposure distributions for all and asthmatic school age
children associated with recent air quality and air quality simulated
to just meet the current and alternative 8-hour O₃ NAAQS
based on the results from the exposure analysis described in the
previous section. The risk estimates have been developed for 12 large
urban areas for the O₃ season.\41\ These 12 urban areas
include approximately 18.3 million school age children, of which 2.6
million are asthmatic school age children.\42\
---------------------------------------------------------------------------

\41\ As discussed above in section II.B.1, the urban areas were
defined using the consolidated statistical areas definition and the
total population residing in the 12 urban areas was approximately
88.5 million people.
\42\ For 9 of the 12 urban areas, the O₃ season is
defined as a period running from March or April to September or
October. In 3 of the urban areas (Houston, Los Angeles, and
Sacramento), the O₃ season is defined as the entire year.
---------------------------------------------------------------------------

In addition to uncertainties arising from sample size
considerations, which are quantitatively characterized and presented as
95 percentile credible intervals, there are additional uncertainties
and caveats associated with the lung function risk estimates. These
include uncertainties about the shape of the exposure-response
relationship, particularly at levels below 0.080 ppm, and about policy-
relevant background levels, for which sensitivity analyses have been
conducted. Additional important caveats and uncertainties concerning
the lung function portion of the health risk assessment include: (1)
The uncertainties and limitations associated with the exposure
estimates discussed above and (2) the inability to account for some
factors which are known to affect the exposure-response relationships
(e.g., assigning healthy and asthmatic children the same responses as
observed in healthy adult subjects and not adjusting response rates to
reflect the increase and attenuation of responses that have been
observed in studies of lung function responses upon repeated
exposures). A more complete discussion of assumptions and uncertainties
is contained in chapter 5 of the Staff Paper and in the Risk Assessment
TSD (Abt Associates, 2006).
The second part of the risk assessment is based on health effects
observed in epidemiological studies. Based on a review of the evidence
evaluated in the Criteria Document and Staff Paper, as well as the
criteria discussed in chapter 5 of the Staff Paper, the following
categories of health endpoints associated with short-term exposures to
ambient O₃ concentrations were included in the risk
assessment: respiratory symptoms in moderate to severe asthmatic
children, hospital admissions for respiratory causes, and non-
accidental and cardiorespiratory mortality. As discussed above, there
is strong evidence of a causal relationship for the respiratory
morbidity endpoints included in the current risk assessment. With
respect to nonaccidental and cardiorespiratory mortality, the Criteria
Document concludes that there is strong evidence which is highly
suggestive of a causal relationship between nonaccidental and
cardiorespiratory-related mortality and O₃ exposures during
the warm O₃ season. As discussed in the Staff Paper (chapter
5), EPA also recognizes that for some of the effects observed in
epidemiological studies, such as increased respiratory-related hospital
admissions and nonaccidental and cardiorespiratory mortality,
O₃ may be serving as an indicator for reactive oxidant
species in the overall photochemical oxidant mix and that these other
constituents may be responsible in whole or part for the observed
effects.
Risk estimates for each health endpoint category were only
developed for areas that were the same or close to the location where
at least one concentration-response function for the health endpoint
had been estimated.\43\ Thus, for respiratory symptoms in moderate to
severe asthmatic children only the Boston urban area was included and
four urban areas were included for respiratory-related hospital
admissions. Nonaccidental mortality risk estimates were developed for
12 urban areas and 8 urban areas were included for cardiorespiratory
mortality.
---------------------------------------------------------------------------

\43\ The geographic boundaries for the urban areas included in
this portion of the risk assessment were generally matched to the
geographic boundaries used in the epidemiological studies that
served as the basis for the concentration-response functions. In
most cases, the urban areas were defined as either a single county
or a few counties for this portion of the risk assessment.
---------------------------------------------------------------------------

The concentration-response relationships used in the assessment are
based on findings from human epidemiological studies that have relied
on fixed-site ambient monitors as a surrogate for actual ambient
O₃ exposures. In order to estimate the incidence of a
particular health effect associated with recent air quality in a
specific county or set of counties attributable to ambient
O₃ exposures in excess of background, as well as the change
in incidence corresponding to a given change in O₃ levels
resulting from just meeting the current or alternative 8-hour
O₃ standards, three elements are required for this part of
the risk assessment. These elements are: (1) Air quality information
(including recent air quality data for O₃ from ambient
monitors for the selected location, estimates of background
O₃ concentrations appropriate for that location, and a
method for adjusting the recent data to reflect patterns of air quality
estimated to occur when the area just meets a given O₃
standard); (2) relative risk-based concentration-response functions
that provide an estimate of the relationship between the health
endpoints of interest and ambient O₃ concentration; and (3)
annual or seasonal baseline health effects incidence rates and
population data, which are needed to provide an estimate of the
seasonal baseline incidence of health effects in an area before any
changes in O₃ air quality.
A key component in the portion of the risk assessment based on
epidemiological studies is the set of concentration-response functions
which provide estimates of the relationships

[[Page 37859]]

between each health endpoint of interest and changes in ambient
O₃ concentrations. Studies often report more than one
estimated concentration-response function for the same location and
health endpoint. Sometimes models include different sets of co-
pollutants and/or different lag periods between the ambient
concentrations and reported health responses. For some health
endpoints, there are studies that estimated multi-city and single-city
O₃ concentration-response functions. While the Risk
Assessment TSD and chapter 5 of the Staff Paper present a more
comprehensive set of risk estimates, EPA has focused on estimates based
on multi-city studies where available. The advantages of relying more
heavily on concentration-response functions based on multi-city studies
include: (1) More precise effect estimates due to larger data sets,
reducing the uncertainty around the estimated coefficient; (2) greater
consistency in data handling and model specification that can eliminate
city-to-city variation due to study design; and (3) less likelihood of
publication bias or exclusion of reporting of negative or
nonsignificant findings. Where studies reported different effect
estimates for varying lag periods, consistent with the Criteria
Document, single day lag periods of 0 to 1 days were used for
associations with respiratory hospital admissions and mortality. For
mortality associated with exposure to O₃ which may result
over a several day period after exposure, distributed lag models, which
take into account the contribution to mortality effects over several
days, were used where available.
One of the most important elements affecting uncertainties in the
epidemiological-based portion of the risk assessment is the
concentration-response relationships used in the assessment. The
uncertainty resulting from the statistical uncertainty associated with
the estimate of the O₃ coefficient in the concentration-
response function was characterized either by confidence intervals or
by Bayesian credible intervals around the corresponding point estimates
of risk. Confidence and credible intervals express the range within
which the true risk is likely to fall if the only uncertainty
surrounding the O₃ coefficient involved sampling error.
Other uncertainties, such as differences in study location, time period
(i.e., the years in which the study was conducted), and model
uncertainties are not represented by the confidence or credible
intervals presented, but were addressed by presenting estimates for
different urban areas, by including risk estimates based on studies
using different time periods and models, where available, and/or are
discussed throughout section 5.3 of the Staff Paper. Because
O₃ effects observed in the epidemiological studies have been
more clearly and consistently shown for warm season analyses, all
analyses for this portion of the risk assessment were carried out for
the same time period, April through September.
The Criteria Document finds that no definitive conclusion can be
reached with regard to the existence of population thresholds in
epidemiological studies (Criteria Document, pp. 8-44). EPA recognizes,
however, the possibility that thresholds for individuals may exist for
reported associations at fairly low levels within the range of air
quality observed in the studies, but not be detectable as population
thresholds in epidemiological analyses. Based on the Criteria
Document's conclusions, EPA judged and CASAC concurred, that there is
insufficient evidence to support use of potential population threshold
levels in the quantitative risk assessment. However, EPA recognizes
that there is increasing uncertainty about the concentration-response
relationship at lower concentrations which is not captured by the
characterization of the statistical uncertainty due to sampling error.
Therefore, the risk estimates for respiratory symptoms in moderate to
severe asthmatic children, respiratory-related hospital admissions, and
premature mortality associated with exposure to O₃ must be
considered in light of uncertainties about whether or not these
O₃-related effects occur in these populations at very low
O₃ concentrations.
With respect to variability within this portion of the risk
assessment, there is variability among concentration-response functions
describing the relation between O₃ and both respiratory-
related hospital admissions and nonaccidental and cardiorespiratory
mortality across urban areas. This variability is likely due to
differences in population (e.g., age distribution), population
activities that affect exposure to O₃ (e.g., use of air
conditioning), levels and composition of co-pollutants, baseline
incidence rates, and/or other factors that vary across urban areas. The
current risk assessment incorporates some of the variability in key
inputs to the analysis by using location-specific inputs (e.g.,
location-specific concentration-response functions, baseline incidence
rates, and air quality data). Although spatial variability in these key
inputs across all U.S. locations has not been fully characterized,
variability across the selected locations is imbedded in the analysis
by using, to the extent possible, inputs specific to each urban area.
c. Risk Estimates and Key Observations
The Staff Paper (chapter 5) and Risk Assessment TSD present risk
estimates associated with just meeting the current and several
alternative 8-hour standards, as well as three recent years of air
quality as represented by 2002, 2003, and 2004 monitoring data. As
discussed in the exposure analysis section above, there is considerable
city-to-city and year-to-year variability in the O₃ levels
during this period, which results in significant variability in both
portions of the health risk assessment.
In the 1997 risk assessment, risks for lung function decrements
associated with 1-hour heavy exertion, 1-hour moderate exertion, and 8-
hour moderate exertion exposures were estimated. Since the 8-hour
moderate exertion exposure scenario for children clearly resulted in
the greatest health risks in terms of lung function decrements, EPA has
chosen to include only the 8-hour moderate exertion exposures in the
current risk assessment for this health endpoint. Thus, the risk
estimates presented here and in the Staff Paper are most useful for
making relative comparisons across alternative air quality scenarios
and do not represent the total risks for lung function decrements in
children or other groups within the general population associated with
any of the air quality scenarios. Thus, some outdoor workers and adults
engaged in moderate exertion over multi-hour periods (e.g., 6-8-hour
exposures) also would be expected to experience similar lung function
decrements. However, the percentage of each of these other
subpopulations expected to experience these effects is expected to be
smaller than all school age children who tend to spend more hours
outdoors while active based on the exposure analyses conducted during
the prior review.
Table 2 presents a summary of the risk estimates for lung function
decrements for the current standard and several alternative 8-hour
standard levels with the same form as the current 8-hour standard. The
estimates are for the aggregate number and percent of all school age
children across 12 urban areas and the aggregate number and percent of
asthmatic school age children

[[Page 37860]]

across 5 urban areas \44\ who are estimated to have at least 1 moderate
or greater lung function response (defined as FEV₁ >=15
percent in all children and >=10 percent in asthmatic children)
associated with 8-hour exposures to O₃ while engaged in
moderate or greater exertion on average over the 8-hour period. The
lung function risk estimates summarized in Table 2 illustrate the year-
to-year variability in both remaining risk associated with a relatively
high year (i.e., based on adjusting 2002 O₃ air quality
data) and relatively low year (based on adjusting 2004 O₃
air quality data) as well as the year-to-year variability in the risk
reduction estimated to occur associated with various alternative
standards relative to just meeting the current standard. For example,
it is estimated that about 610,000 school age children (3.2 percent of
school age children) would experience 1 or more moderate lung function
decrements for the 12 urban areas associated with O₃ levels
just meeting the current standard based on 2002 air quality data
compared to 230,000 (1.2 percent of children) associated with just
meeting the current standard based on 2004 air quality data.
---------------------------------------------------------------------------

\44\ Due to time constraints, lung function risk estimates for
asthmatic school age children were developed for only 5 of the 12
urban areas, and the areas were selected to represent different
geographic regions. The 5 areas were: Atlanta, Chicago, Houston, Los
Angeles, and New York City.

Table 2.--Number and Percent of All and Asthmatic School Age Children in Several Urban Areas Estimated To
Experience Moderate or Greater Lung Function Responses 1 or More Times per Season Associated With 8-Hour Ozone
Exposures Associated With Just Meeting Alternative 8-Hour Standards Based on Adjusting 2002 and 2004 Air Quality
Data \1,\ \2\
----------------------------------------------------------------------------------------------------------------
All children, ages 5-18, FEV1 >=15 Asthmatic children, ages 5-18, FEV1
percent, aggregate for 12 urban areas, >=10 percent, aggregate for 5 urban
number of children affected (% of all) areas, number of children affected (%
8-Hour air quality standards \3\ [% reduction from current standard]
of group) [% reduction from current
---------------------------------------- standard]
---------------------------------------
2002 2004 2002 2004]
----------------------------------------------------------------------------------------------------------------
0.084 ppm (Current standard).... 610,000 (3.3%).... 230,000 (1.2%).... 130,000 (7.8%).... 70,000 (4.2%).
0.080 ppm....................... 490,000 (2.7%) 180,000 (1.0%) NA \4\............ NA.
[20% reduction]. [22% reduction].
0.074 ppm....................... 340,000 (1.9%) 130,000 (0.7%) 90,000 (5.0%) [31 40,000 (2.7%) [43%
[44% reduction]. [43% reduction]. % reduction]. reduction].
0.070 ppm....................... 260,000 (1.5%) 100,000 (0.5%) NA................ NA.
[57% reduction]. [57% reduction].
0.064 ppm....................... 180,000 (1.0%) 70,000 (0.4%) [70% 50,000 (3.0%) [62% 20,000 (1.5%) [71%
[70% reduction]. reduction]. reduction]. reduction].
----------------------------------------------------------------------------------------------------------------
\1\ Associated with exposures while engaged in moderate or greater exertion which is defined as having an 8-hour
average equivalent ventilation rate >=13 l-min/m \2\.
\2\ Estimates are the aggregate central tendency results based on either 12 urban areas (Atlanta, Boston,
Chicago, Cleveland, Detroit, Houston, Los Angeles, New York, Philadelphia, Sacramento, St. Louis, and
Washington, DC) or 5 urban areas (Atlanta, Chicago, Houston, Los Angeles, New York). Estimates are for the O3
season which is all year in Houston, Los Angeles and Sacramento and March or April to September or October for
the remaining urban areas.
\3\ All standards summarized here have the same form as the current 8-hour standard which is specified as the 3-
year average of the annual 4th highest daily maximum 8-hour average concentrations must be at or below the
stated concentration level. As described in the Staff Paper (section 4.5.8), recent O3 air quality
distributions have been statistically adjusted to simulate just meeting the current and selected alternative
standards. These simulations do not represent predictions of when, whether, or how areas might meet the
specified standards
\4\ NA (not available) indicates that EPA did not develop risk estimates for these scenarios for the asthmatic
school age children population.

As discussed in the Staff Paper, a child may experience multiple
occurrences of a lung function response during the O₃
season. For example, upon meeting the current 8-hour standard, the
median estimates are that about 610,000 children would experience a
moderate or greater lung function response 1 or more times for the
aggregate of the 12 urban areas over a single O₃ season
(based on the 2002 simulation), and that there would be almost 3.2
million total occurrences. Thus, on average it is estimated that there
would be about 5 occurrences per O₃ season per responding
child for air quality just meeting the current 8-hour standard across
the 12 urban areas. While the estimated number of occurrences per
O₃ season is lower when based on the 2004 simulation than
for the 2002 simulation, the estimated number of occurrences per
responding child is similar. EPA recognizes that some children in the
population might have only 1 or 2 occurrences while others may have 6
or more occurrences per O₃ season. Risk estimates based on
adjusting 2003 air quality to simulate just meeting the current and
alternative 8-hour standards are intermediate to the estimates
presented in Table 2 above in this notice and are presented in the
Staff Paper (chapter 5) and Risk Assessment TSD.
For just meeting the current 8-hour standard, Table 5-8 in the
Staff Paper shows that median estimates across the 12 urban areas for
all school age children experiencing 1 or more moderate lung function
decrements ranges from 0.9 to 5.4 percent based on the 2002 simulation
and from 0.8 to 2.2 percent based on the 2004 simulation. Risk
estimates for each urban area included in the assessment, for each of
the three years analyzed, and for additional alternative standards are
presented in chapter 5 of the Staff Paper and in the Risk Assessment TSD.
For just meeting the current 8-hour standard, the median estimates
across the 5 urban areas for asthmatic school age children range from
3.4 to 10.9 percent based on the 2002 simulation and from 3.2 to 6.9
percent based on the 2004 simulation.
Key observations important in comparing estimated lung function
risks associated with attainment of the current NAAQS and alternative
standards under consideration include:
(1) As discussed above, there is significant year to year
variability in the range of median estimates of the number of school
age children (ages 5-18) estimated to experience at least one
FEV₁ decrement >=15 percent due to 8-hour O₃
exposures across the 12 urban

[[Page 37861]]

areas analyzed, and similarly across the 5 urban areas analyzed for
asthmatic school age children (ages 5-18) estimated to experience at
least one FEV₁ decrement >=10 percent, when the current and
alternative 8-hour standards are just met.
(2) For asthmatic school age children, the median estimates of
occurrences of FEV₁ decrements >=10% range from 52,000 to
nearly 510,000 responses associated with just meeting the current
standard (based on the 2002 simulation) and range from 61,000 to about
240,000 occurrences (based on the 2004 simulation). These risk
estimates would be reduced to a range of 14,000 to about 275,000
occurrences (2002 simulation) and to about 18,000 to nearly 125,000
occurrences (2004 simulation) upon just meeting the most stringent
alternative 8-hour standard (0.064 ppm, 4th highest). The average
number of occurrences per asthmatic child in an O₃ season
ranged from about 6 to 11 associated with just meeting the current
standard (2002 simulation). The average number of occurrences per
asthmatic child ranged from 4 to 12 upon meeting the most stringent
alternative examined (0.064 ppm, 4th-highest) based on the 2002
simulation. The number of occurrences per asthmatic child is similar
for the scenarios based on the 2004 simulation.
As discussed above, several epidemiological studies have reported
increased respiratory morbidity outcomes (e.g., respiratory symptoms in
moderate to severe asthmatic children, respiratory-related hospital
admissions) and increased nonaccidental and cardiorespiratory mortality
associated with exposure to ambient O₃ concentrations. The
results and key observations from this portion of the risk assessment
are presented below:
(1) Estimates for increased respiratory symptoms (i.e., chest
tightness, shortness of breath, and wheeze) in moderate/severe
asthmatic children (ages 0-12) were developed for the Boston urban area
only. The median estimated number of days involving chest tightness
(using the concentration-response relationship with only O₃
in the model) is about 6,100 (based on the 2002 simulation) and about
4,500 (based on the 2004 simulation) upon meeting the current 8-hour
standard and this is reduced to about 4,600 days (2002 simulation) and
3,100 days (2004 simulation) upon meeting the most stringent
alternative examined (0.064 ppm, 4th-highest daily maximum 8-hour
average). This corresponds to 11 percent (2002 simulation) and 8
percent (2004 simulation) of total incidence of chest tightness upon
meeting the current 8-hour standard and to about 8 percent (2002
simulation) and 5.5 percent (2004 simulation) of total incidence of
chest tightness upon meeting a 0.064 ppm, 4th-highest daily maximum 8-
hour average standard. Similar patterns of effects and reductions in
effects are observed for each of the respiratory symptoms examined.
(2) The Staff Paper and Risk Assessment TSD present unscheduled
hospital admission risk estimates for respiratory illness and asthma in
New York City associated with short-term exposures to O₃
concentrations in excess of background levels from April through
September for several recent years (2002, 2003, and 2004) and upon just
meeting the current and alternative 8-hour standards based on
simulating O₃ levels using 2002-2004 O₃ air
quality data. For total respiratory illness, EPA estimates about 6.4
cases per 100,000 relevant population (2002 simulation) and about 4.6
cases per 100,000 relevant population (2004 simulation), which
represents 1.5 percent (2002 simulation) and 1.0 percent (2004
simulation) of total incidence or about 510 cases (2002 simulation) and
about 370 cases (2004 simulation) upon just meeting the current 8-hour
standard. For asthma-related hospital admissions, which are a subset of
total respiratory illness admissions, the estimates are about 5.5 cases
per 100,000 relevant population (2002 simulation) and about 3.9 cases
per 100,000 relevant population (2004 simulation), which represents
about 3.3 percent (2002 simulation) and 2.4 percent (2004 simulation)
of total incidence or about 440 cases (2002) and about 310 cases (2004)
for this same air quality scenario.
For increasingly more stringent alternative 8-hour standards, there
is a gradual reduction in respiratory illness cases per 100,000
relevant population from 6.4 cases per 100,000 upon just meeting the
current 8-hour standard to 4.6 cases per 100,000 under the most
stringent 8-hour standard (i.e., 0.064 ppm, average 4th-highest daily
maximum) analyzed based on the 2002 simulation. Similarly, based on the
2004 simulation there is a gradual reduction from 4.6 cases per 100,000
relevant population upon just meeting the current 8-hour standard to
3.0 cases per 100,000 under the 0.064 ppm, average 4th-highest daily
maximum standard.
Additional respiratory-related hospital admission estimates for
three other locations are provided in the Risk Assessment TSD. EPA
notes that the concentration-response functions for each of these
locations examined different outcomes in different age groups (e.g., >
age 30 in Los Angeles, > age 64 in Cleveland and Detroit, vs. all ages
in New York City), making comparison of the risk estimates across the
areas very difficult.
(3) Based on the median estimates for incidence for nonaccidental
mortality (based on the Bell et al. (2004) 95 cities concentration-
response function), meeting the most stringent standard (0.064 ppm) is
estimated to reduce mortality by 40 percent of what it would be
associated with just meeting the current standard (based on the 2002
simulation). The patterns for cardiorespiratory mortality are similar.
The aggregate O₃-related cardiorespiratory mortality upon
just meeting the most stringent standard shown is estimated to be about
42 percent of what it would be upon just meeting the current standard,
using simulated O₃ concentrations that just meet the current
and alternative 8-hour standards based on the 2002 simulation. Using
the 2004 simulation, the corresponding reductions show a similar
pattern but are somewhat greater.
(4) Much of the contribution to the risk estimates for non-
accidental and cardiorespiratory mortality upon just meeting the
current 8-hour standard is associated with 24-hour O₃
concentrations between background and 0.040 ppm. Based on examining
relationships between 24-hour concentrations averaged across the
monitors within an urban area and 8-hour daily maximum concentrations,
8-hour daily maximum levels at the highest monitor in an urban area
associated with these averaged 24-hour levels are generally about twice
as high as the 24-hour levels. Thus, most O₃-related
nonaccidental mortality is estimated to occur when O₃
concentrations are between background and when the highest monitor in
the urban area is at or below 0.080 ppm, 8-hour average concentration.
The discussion below highlights additional observations and
insights from the O₃ risk assessment, together with
important uncertainties and limitations.
(1) As discussed in the Staff Paper (section 5.4.5) EPA has greater
confidence in relative comparisons in risk estimates between
alternative standards than in the absolute magnitude of risk estimates
associated with any particular standard.
(2) Significant year-to-year variability in O₃
concentrations combined with the use of a 3-year design value to
determine the amount of air quality adjustment to be applied to each
year analyzed, results in significant year-to-year variability in the
annual health risk

[[Page 37862]]

estimates upon just meeting the current and potential alternative 8-
hour standards.
(3) There is noticeable city-to-city variability in estimated
O₃-related incidence of morbidity and mortality across the
12 urban areas analyzed for both recent years of air quality and for
air quality adjusted to simulate just meeting the current and selected
potential alternative standards. This variability is likely due to
differences in air quality distributions, differences in exposure
related to many factors including varying activity patterns and air
exchange rates, differences in baseline incidence rates, and
differences in susceptible populations and age distributions across the
12 urban areas.
(4) With respect to the uncertainties about estimated policy-
relevant background concentrations, as discussed in the Staff Paper
(section 5.4.3), alternative assumptions about background levels had a
variable impact depending on the health effect considered and the
location and standard analyzed in terms of the absolute magnitude and
relative changes in the risk estimates. There was relatively little
impact on either absolute magnitude or relative changes in lung
function risk estimates due to alternative assumptions about background
levels. With respect to O₃-related non-accidental mortality,
while notable differences (i.e., greater than 50 percent)\45\ were
observed for nonaccidental mortality in some areas, particularly for
more stringent standards, the overall pattern of estimated reductions,
expressed in terms of percentage reduction relative to the current
standard, was significantly less impacted.
---------------------------------------------------------------------------

\45\ For example, assuming lower background levels resulted in
increased estimates of non-accidental mortality incidence per
100,000 that were often 50 to 100 percent greater than the base case
estimates; assuming higher background levels resulted in decreased
estimates of non-accidental mortality incidence per 100,000 that
were less than the base case estimates by 50 percent or more in many
of the areas.
---------------------------------------------------------------------------

C. Conclusions on the Adequacy of the Current Primary Standard

1. Background
The initial issue to be addressed in the current review of the
primary O₃ standard is whether, in view of the advances in
scientific knowledge and additional information, the existing standard
should be revised. In evaluating whether it is appropriate to retain or
revise the current standard, the Administrator builds upon the last
review and reflects the broader body of evidence and information now
available. The Administrator has taken into account both evidence-based
and quantitative exposure- and risk-based considerations in developing
conclusions on the adequacy of the current primary O₃
standard. Evidence-based considerations include the assessment of
evidence from controlled human exposure, animal toxicological, field,
and epidemiological studies for a variety of health endpoints. For
those endpoints based on epidemiological studies, greater weight has
been placed on associations with health endpoints that are causal or
likely causal based on an integrative synthesis of the entire body of
evidence, including not only all available epidemiological evidence but
also evidence from animal toxicological and controlled human exposure
studies. Less weight has been placed on evidence of associations that
were judged to be only suggestive of possible causal relationships.
Consideration of quantitative exposure- and risk-based information
draws from the results of the exposure and risk assessments described
above. More specifically, estimates of the magnitude of O₃-
related exposures and risks associated with recent air quality levels,
as well as the exposure and risk reductions likely to be associated
with just meeting the current 8-hour primary O₃ NAAQS, have
been considered.
In this review, a series of general questions frames the approach
to reaching a decision on the adequacy of the current standard, such as
the following: (1) To what extent does newly available information
reinforce or call into question evidence of associations of
O₃ exposures with effects identified in the last review?;
(2) to what extent has evidence of new effects and/or at-risk
populations become available since the last review?; (3) to what extent
have important uncertainties identified in the last review been reduced
and have new uncertainties emerged?; (4) to what extent does newly
available information reinforce or call into question any of the basic
elements of the current standards?
The question of whether the available evidence supports
consideration of a standard that is more protective than the current
standard includes consideration of: (1) Whether there is evidence that
associations, especially likely causal associations, extend to ambient
O₃ concentration levels that are as low as or lower than had
previously been observed, and the important uncertainties associated
with that evidence; (2) the extent to which exposures of concern and
health risks are estimated to occur in areas upon meeting the current
standard and the important uncertainties associated with the estimated
exposures and risks; and (3) the extent to which the O₃-
related health effects indicated by the evidence and the exposure and
risk assessments are considered important from a public health
perspective, taking into account the nature and severity of the health
effects, the size of the at-risk populations, and the kind and degree
of the uncertainties associated with these considerations.
The current primary O₃ standard is an 8-hour standard,
which was set at a level of 0.08 ppm,\46\ with a form of the annual
fourth-highest daily maximum 8-hour average concentration, averaged
over three years. This standard was chosen to provide protection to the
public, especially children and other at-risk populations, against a
wide range of O₃-induced health effects. As an introduction
to this discussion of the adequacy of the current O₃
standard, it is useful to summarize the key factors that formed the
basis of the decision in the last review to revise the averaging time,
level, and form of the then current 1-hour standard.
---------------------------------------------------------------------------

\46\ If the standard were to be specified to the nearest
thousandth ppm, the current 0.08 ppm 8-hour standard would be
equivalent to a standard set at 0.084 ppm, reflecting the data
rounding conventions that are part of the definition of the current
8-hour standard.
---------------------------------------------------------------------------

In the last review, the key factor in deciding to revise the
averaging time of the primary standard was evidence from controlled
human exposure studies of healthy young adult subjects exposed for 1 to
8 hours to O₃. The best documented health endpoints in these
studies were decrements in indices of lung function, such as forced
expiratory volume in 1 second (FEV₁), and respiratory
symptoms, such as cough and chest pain on deep inspiration. For short-
term exposures of 1 to 3 hours, group mean FEV1 decrements were
statistically significant for O₃ concentrations only at and
above 0.12 ppm, and only when subjects engaged in very heavy exertion.
By contrast, evidence available in the prior review showed that
prolonged exposures of 6 to 8 hours produced statistically significant
group mean FEV₁ decrements at the lowest O₃
concentrations evaluated in those studies, 0.080 ppm, even when
experimental subjects were engaged in more realistic intermittent
moderate exertion levels. The health significance of this newer
evidence led to the conclusion in the 1997 final decision that the 8-
hour averaging time is more directly associated with health effects of
concern at lower O₃ concentrations than is the 1-hour
averaging time.

[[Page 37863]]

Based on the available evidence of O₃-related health
effects, the following factors were of particular importance in the
last review in informing the selection of the level and form of a new
8-hour standard: (1) Quantitative estimates of O₃-related
risks to active children, who were judged to be an at-risk subgroup of
concern, in terms of transient and reversible respiratory effects
judged to be adverse, including moderate to large decreases in lung
function and moderate to severe pain on deep inspiration, and the
uncertainty and variability in such estimates; (2) consideration of
both the estimated percentages, total numbers of children, and number
of times they were likely to experience such effects; (3)
epidemiological evidence of associations between ambient O₃
and increased respiratory hospital admissions, and quantitative
estimates of percentages and total numbers of asthma-related admissions
in one example urban area that were judged to be indicative of a
pyramid of much larger effects, including respiratory-related hospital
admissions, emergency department visits, doctor visits, and asthma
attacks and related increased medication use; (4) quantitative
estimates of the number of ``exposures of concern\47\'' (defined as
exposures >= 0.080 ppm for 6 to 8 hour) that active children are likely
to experience, and the uncertainty and variability in such estimates;
(5) the judgment that such exposures are an important indicator of
public health impacts of O₃-related effects for which
information is too limited to develop quantitative risk estimates,
including increased nonspecific bronchial responsiveness (e.g., related
to aggravation of asthma), decreased pulmonary defense mechanisms
(suggestive of increased susceptibility to respiratory infection), and
indicators of pulmonary inflammation (related to potential aggravation
of chronic bronchitis or long-term damage to the lungs); (6) the
broader public health perspective of the number of people living in
areas that would breathe cleaner air as a result of the revised
standard; (7) consideration of the relative seriousness of various
health effects and the relative degree of certainty in both the
likelihood that people will experience various health effects and their
medical significance; (8) the relationship of a standard level to
estimated ``background'' levels associated with nonanthropogenic
sources of O₃; and (9) CASAC's advice and recommendations.
Additional factors considered in selecting the form of the standard
included balancing the public health implications of the estimated
number of times in an O₃ season that the standard level
might be exceeded in an area that is in attainment with the standard
with the year-to-year stability of the air quality statistic, which can
be particularly affected by years with unusual meteorology. A more
stable air quality statistic serves to avoid disruptions to ongoing
control programs that could result from moving into and out of
attainment, thereby interrupting the public health protection afforded
by such control programs.
---------------------------------------------------------------------------

\47\ In the last review, ``exposures of concern'' referred to
exposures at and above 0.08 ppm, 8-hour average, at which a range of
health effects have been observed in controlled human studies, but
for which data were too limited to allow for quantitative risk
assessment. (62 FR 38860, July 18, 1997).
---------------------------------------------------------------------------

In reaching a final decision in the last review, the Administrator
was mindful that O₃ exhibits a continuum of effects, such
that there is no discernible threshold above which public health
protection requires that no exposures be allowed or below which all
risks to public health can be avoided. The final decision reflected a
recognition that important uncertainties remained, for example with
regard to interpreting the role of other pollutants co-occurring with
O₃ in observed associations, understanding biological
mechanisms of O₃-related health effects, and estimating
human exposures and quantitative risks to at-risk populations for these
health effects.
2. Evidence- and Exposure/Risk-Based Considerations in the Staff Paper
The Staff Paper (section 6.3.1) considers the evidence presented in
the Criteria Document as discussed above in section II.A as a basis for
evaluating the adequacy of the current O₃ standard,
recognizing that important uncertainties remain. The extensive body of
human clinical, toxicological, and epidemiological evidence serves as
the basis for the judgments about O₃-related health effects
discussed above, including judgments about causal relationships with a
range of respiratory morbidity effects, including lung function
decrements, increased respiratory symptoms, airway inflammation,
increased airway responsiveness, and respiratory-related
hospitalizations and emergency department visits in the warm season,
and about the evidence being highly suggestive that O₃
directly or indirectly contributes to non-accidental and
cardiopulmonary-related mortality.
These judgments take into account important uncertainties that
remain in interpreting this evidence. For example, with regard to the
utility of time-series epidemiological studies to inform judgments
about a NAAQS for an individual pollutant, such as O₃,
within a mix of highly correlated pollutants, such as the mix of
oxidants produced in photochemical reactions in the atmosphere, the
Staff Paper notes that there are limitations especially at ambient
O₃ concentrations below levels at which O₃-
related effects have been observed in controlled human exposure
studies. The Staff Paper (section 3.4.5) also recognizes that the
available epidemiological evidence neither supports nor refutes the
existence of thresholds at the population level for effects such as
increased hospital admissions and premature mortality. There are
limitations in epidemiological studies that make discerning thresholds
in populations difficult, including low data density in the lower
concentration ranges, the possible influence of exposure measurement
error, and variability in susceptibility to O₃-related
effects in populations.
While noting these limitations in the interpretation of the
findings from the epidemiological studies, the Staff Paper (section
3.4.5) concludes that if a population threshold level does exist, it
would likely be well below the level of the current O₃
standard and possibly within the range of background levels. As
discussed above in section II.A.3.a, this conclusion is supported by
several epidemiological studies that have explored the question of
potential thresholds directly, either using a statistical curve-fitting
approach to evaluate whether linear or non-linear models fit the data
better using sub-sets of the data, where days over or under a specific
cutpoint (e.g., 0.080 ppm or even lower O₃ levels) were
excluded and then evaluating the association for statistical
significance. In addition to direct consideration of the
epidemiological studies, findings from controlled human exposure
studies discussed above in section II.A.2.a.i(a)(i) indicate that
prolonged exposures produced statistically significant group mean
FEV₁ decrements and symptoms in healthy adult subjects at
levels down to at least 0.060 ppm, with a small percentage of subjects
experiencing notable effects (e.g., >10 percent FEV₁
decrement, pain on deep inspiration). Controlled human exposure studies
evaluated in the last review also found significant responses in
indicators of lung inflammation and cell injury at 0.080 ppm in healthy
adult subjects. The effects in these controlled human exposure studies
were observed in healthy young adult subjects, and it is likely that
more serious responses, and

[[Page 37864]]

responses at lower levels, would occur in people with asthma and other
respiratory diseases. These physiological effects have been linked to
aggravation of asthma and increased susceptibility to respiratory
infection, potentially leading to increased medication use, increased
school and work absences, increased visits to doctors' offices and
emergency departments, and increased hospital admissions. The
observations provide additional support for the conclusion in the Staff
Paper that the associations observed in the epidemiological studies,
particularly for respiratory-related effects and potentially for
cardiovascular effects, extend down to O₃ levels well below
the current standard (i.e., 0.084 ppm) (EPA, 2007, p. 6-7).
As discussed above in section II.A and in the Staff Paper (section
3.7), the newly available information reinforces the judgments about
the likelihood of causal relationships between O₃ exposure
and respiratory effects observed in the last review and broadens the
evidence of O₃-related associations to include additional
respiratory-related endpoints, newly identified cardiovascular-related
health endpoints, and mortality. Newly available evidence also has
shown that people with asthma are likely to experience more serious
effects than people who do not have asthma (section II.A.4.b.ii above).
The Staff Paper also concludes that substantial progress has been made
since the last review in advancing the understanding of potential
mechanisms by which ambient O₃, alone and in combination
with other pollutants, is causally linked to a range of respiratory-
related health endpoints, and may be causally linked to a range of
cardiovascular-related health endpoints. Thus, the Staff Paper (section
6.3.6) finds strong support in the evidence developed since the last
review, for consideration of an O₃ standard that is at least
as protective as the current standard and finds no support for
consideration of an O₃ standard that is less protective than
the current standard. This conclusion is consistent with the advice and
recommendations of CASAC and with the views expressed by all interested
parties who provided comments on drafts of the Staff Paper. While CASAC
and some commenters supported revising the current standard to provide
increased public health protection and other commenters supported
retaining the current standard, no one who provided comments supported
a standard that would be less protective than the current standard.
a. Evidence-Based Considerations
In looking more specifically at the controlled human exposure and
epidemiological evidence (which is summarized in chapter 3 and Appendix
3B of the Staff Paper), the Staff Paper first notes that controlled
human exposure studies provide the clearest and most compelling
evidence for an array of human health effects that are directly
attributable to acute exposures to O₃ per se. Evidence from
such human studies, together with animal toxicological studies, help to
provide biological plausibility for health effects observed in
epidemiological studies. In considering the available evidence, the
Staff Paper focuses on studies that examined health effects that have
been demonstrated to be caused by exposure to O₃, or for
which the Criteria Document judges associations with O₃ to
be causal or likely causal, or for which the evidence is highly
suggestive that O₃ contributes to the reported effects. In
considering the epidemiological evidence as a basis for reaching
conclusions about the adequacy of the current standard, the Staff Paper
focuses on studies reporting effects in the warm season, for which the
effect estimates are more consistently positive and statistically
significant than those from all-year studies. The Staff Paper (section
6.3.1.1) considers the extent to which such studies provide evidence of
associations that extend down to ambient O₃ concentrations
below the level of the current standard, which would thereby call into
question the adequacy of the current standard. In so doing, the Staff
Paper notes, as discussed above, that if a population threshold level
does exist for an effect observed in such studies, it would likely be
at a level well below the level of the current standard. The Staff
Paper (section 6.3.1.1) also attempts to characterize whether the area
in which a study was conducted likely would or would not have met the
current standard during the time of the study, although it recognizes
that the confidence that would appropriately be placed on the
associations observed in any given study, or on the extent to which the
association would likely extend down to relatively low O₃
concentrations, is not dependent on this distinction. Further, the
Staff Paper considered studies that examined subsets of data that
include only days with ambient O₃ concentrations below the
level of the current O₃ standard, or below even lower
O₃ concentrations, and continue to report statistically
significant associations. The Staff Paper (section 6.3.1.1) judges that
such studies are directly relevant to considering the adequacy of the
current standard, particularly in light of reported responses to
O₃ at levels below the current standard found in controlled
human exposure studies.
i. Lung Function, Respiratory Symptoms, and Other Respiratory Effects
Health effects for which the Criteria Document continues to find
clear evidence of causal associations with short-term O₃
exposures include lung function decrements, respiratory symptoms,
pulmonary inflammation, and increased airway responsiveness. In the
last review, these O₃-induced effects were demonstrated with
statistical significance down to the lowest level tested in controlled
human exposure studies at that time (i.e., 0.080 ppm). As discussed in
chapter 3 of the Staff Paper, and in section II.A.2.a.i.(a)(i) above,
two new studies are notable in that they are the only controlled human
exposure studies that examined respiratory effects, including lung
function decrements and respiratory symptoms, in healthy adults at
lower exposure levels than had previously been examined. EPA's
reanalysis of the data from the most recent study shows small group
mean decrements in lung function responses to be statistically
significant at the 0.060 ppm exposure level, while the author's
analysis did not yield statistically significant lung function
responses but did yield some statistically significant respiratory
symptom responses toward the end of the exposure period. Notably, these
studies report a small percentage of subjects experiencing lung
function decrements (>= 10 percent) at the 0.060 ppm exposure level.
These studies provide very limited evidence of O₃-related
lung function decrements and respiratory symptoms at this lower
exposure level.
The Staff Paper (section 3.3.1.1.1) notes that evidence from
controlled human exposures studies indicates that people with moderate-
to-severe asthma have somewhat larger decreases in lung function in
response to O₃ relative to healthy individuals and that lung
function responses in people with asthma appear to be affected by
baseline lung function (i.e., magnitude of responses increases with
increasing disease severity). As discussed in the Criteria Document
(p.8-80), this newer information expands our understanding of the
physiological basis for increased sensitivity in people with asthma and
other airway diseases, recognizing that

[[Page 37865]]

people with asthma present a different response profile for cellular,
molecular, and biochemical responses than people who do not have
asthma. New evidence indicates that some people with asthma have
increased occurrence and duration of nonspecific airway responsiveness,
which is an increased bronchoconstrictive response to airway irritants.
Controlled human exposure studies also indicate that some people with
allergic asthma and rhinitis have increased airway responsiveness to
allergens following O₃ exposure. Exposures to O₃
exacerbated lung function decrements in people with pre-existing
allergic airway disease, with and without asthma. Ozone-induced
exacerbation of airway responsiveness persists longer and attenuates
more slowly than O₃-induced lung function decrements and
respiratory symptom responses and can have important clinical
implications for asthmatics.
The Staff Paper (p.6-10) also concludes that newly available human
exposure studies suggest that some people with asthma also have
increased inflammatory responses, relative to non-asthmatic subjects,
and that this inflammation may take longer to resolve. The new data on
airway responsiveness, inflammation, and various molecular markers of
inflammation and bronchoconstriction indicate that people with asthma
and allergic rhinitis (with or without asthma) comprise susceptible
groups for O₃-induced adverse effects. This body of evidence
qualitatively informs the Staff Paper's (pp.6-10 to 6-11) evaluation of
the adequacy of the current O₃ standard in that it indicates
that human clinical and epidemiological panel studies of lung function
decrements and respiratory symptoms that evaluate only healthy, non-
asthmatic subjects likely underestimate the effects of O₃
exposure on asthmatics and other susceptible populations.
The Staff Paper (p.6-11) notes that in addition to the experimental
evidence of lung function decrements, respiratory symptoms, and other
respiratory effects in healthy and asthmatic populations discussed
above, epidemiological studies have reported associations of lung
function decrements and respiratory symptoms in several locations
(Appendix 3B; also Figure 3-4 for respiratory symptoms). As discussed
in the Staff Paper (section 3.3.1.1.1) and above, two large U.S. panel
studies which together followed over 1000 asthmatic children on a daily
basis (Mortimer et al., 2002, the National Cooperative Inner-City
Asthma Study, or NCICAS; and Gent et al., 2003), as well as several
smaller U.S. and international studies, have reported robust
associations between ambient O₃ concentrations and measures
of lung function and daily symptoms (e.g., chest tightness, wheeze,
shortness of breath) in children with moderate to severe asthma and
between O₃ and increased asthma medication use. Overall, the
multi-city NCICAS (2002), Gent et al. (2003), and several other single-
city studies indicate a robust positive association between ambient
O₃ concentrations and increased respiratory symptoms and
increased medication use in asthmatics.
In considering the large number of single-city epidemiological
studies reporting lung function or respiratory symptoms in healthy or
asthmatic populations (Staff Paper, Appendix 3B), the Staff Paper (p.6-
11) notes that most such studies that reported positive and often
statistically significant associations in the warm season were
conducted in areas that likely would not have met the current standard.
In considering the large multi-city NCICAS (Mortimer et al., 2002), the
Staff Paper notes that the 98th percentile 8-hour daily maximum
O₃ concentrations at the monitor reporting the highest
O₃ concentrations in each of the study areas ranged from
0.084 ppm to >0.10 ppm. However, the authors indicate that less than 5
percent of the days in the eight urban areas had 8-hour daily
O₃ concentrations exceeding 0.080 ppm. Moreover, the authors
observed that when days with 8-hour average O₃ levels
greater than 0.080 ppm were excluded, similar effect estimates were
seen compared to estimates which included all of the days. There are
also a few other studies in which the relevant air quality statistics
provide some indication that lung function and respiratory symptom
effects may be occurring in areas that likely would have met the
current standard (EPA, 2007, p.6-12).
ii. Respiratory Hospital Admissions and Emergency Department Visits
At the time of the last review, many time-series studies indicated
positive associations between ambient O₃ and increased
respiratory hospital admissions and emergency room visits, providing
strong evidence for a relationship between O₃ exposure and
increased exacerbations of preexisting lung disease at O₃
levels below the level of the then current 1-hour standard (EPA 2007,
section 3.3.1.1.6). Analyses of data from studies conducted in the
northeastern U.S. indicated that O₃ air pollution was
consistently and strongly associated with summertime respiratory
hospital admissions.
Since the last review, new epidemiological studies have evaluated
the association between short-term exposures to O₃ and
unscheduled hospital admissions for respiratory causes. Large multi-
city studies, as well as many studies from individual cities, have
reported positive and often statistically significant O₃
associations with total respiratory hospitalizations as well as asthma-
and COPD-related hospitalizations, especially in studies analyzing the
O₃ effect during the summer or warm season. Analyses using
multipollutant regression models generally indicate that copollutants
do not confound the association between O₃ and respiratory
hospitalizations and that the O₃ effect estimates were
robust to PM adjustment in all-year and warm-season only data. The
Criteria Document (p.8-77) concludes that the evidence supports a
causal relationship between acute O₃ exposures and increased
respiratory-related hospitalizations during the warm season.
In looking specifically at U.S. and Canadian respiratory
hospitalization studies that reported positive and often statistically
significant associations (and that either did not use GAM or were
reanalyzed to address GAM-related problems), the Staff Paper (p.6-12)
notes that many such studies were conducted in areas that likely would
not have met the current O₃ standard, with many providing
only all-year effect estimates, and with some reporting a statistically
significant association in the warm season. Of the studies that provide
some indication that O₃-related respiratory hospitalizations
may be occurring in areas that likely would have met the current
standard, the Staff Paper notes that some are all-year studies, whereas
others reported statistically significant warm-season associations.
Emergency department visits for respiratory causes have been the
focus of a number of new studies that have examined visits related to
asthma, COPD, bronchitis, pneumonia, and other upper and lower
respiratory infections, such as influenza, with asthma visits typically
dominating the daily incidence counts. Among studies with adequate
controls for seasonal patterns, many reported at least one significant
positive association involving O₃. However, inconsistencies
were observed which were at least partially attributable to differences
in model specifications and analysis approach among various studies. In
general, O₃ effect estimates from summer-only analyses
tended to be positive and larger compared to results from cool season
or all-year analyses. Almost all of the studies that reported

[[Page 37866]]

statistically significant effect estimates were conducted in areas that
likely would not have met the current standard. The Criteria Document
(section 7.3.2) concluded that analyses stratified by season generally
supported a positive association between O₃ concentrations
and emergency department visits for asthma in the warm season. These
studies provide evidence of effects in areas that likely would not have
met the current standard and evidence of associations that likely
extend down to relatively low ambient O₃ concentrations.
iii. Mortality
The 1996 Criteria Document concluded that an association between
daily mortality and O₃ concentrations for areas with high
O₃ levels (e.g., Los Angeles) was suggested. However, due to
a very limited number of studies available at that time, there was
insufficient evidence to conclude that the observed association was
likely causal, and thus the possibility that O₃ exposure may
be associated with mortality was not relied upon in the 1997 decision
on the O₃ primary standard.
Since the last review, as described above, the body of evidence
with regard to O₃-related health effects has been expanded
by animal, human clinical, and epidemiological studies and now includes
biologically plausible mechanisms by which O₃ may affect the
cardiovascular system. In addition, there is stronger information
linking O₃ to serious morbidity outcomes, such as
hospitalization, that are associated with increased mortality. Thus,
there is now a coherent body of evidence that describes a range of
health outcomes from lung function decrements to hospitalization and
premature mortality.
Newly available large multi-city studies (Bell et al., 2004; Huang
et al.,2005; and Schwartz 2005) designed specifically to examine the
effect of O₃ and other pollutants on mortality have provided
much more robust and credible information. Together these studies have
reported significant associations between O₃ and mortality
that were robust to adjustment for PM and different adjustment methods
for temperature and suggest that the effect of O₃ on
mortality is immediate but also persists for several days. One recent
multi-city study (Bell et al., 2006) examined the shape of the
concentration-response function for the O₃-mortality
relationship in 98 U.S. urban communities for the period 1987 to 2000
specifically to evaluate whether a ``safe'' threshold level exists.
Results from various analytic methods all indicated that any threshold,
if it exists, would likely occur at very low concentrations, far below
the level of the current O₃ NAAQS and nearing background
levels.
New data are also available from several single-city studies
conducted world-wide, as well as from several meta-analyses that have
combined information from multiple studies. Three recent meta-analyses
evaluated potential sources of heterogeneity in O₃-mortality
associations. All three analyses reported common findings, including
effect estimates that were statistically significant and larger in warm
season analyses. Reanalysis of results using default GAM criteria did
not change the effect estimates, and there was no strong evidence of
confounding by PM. The Criteria Document (p.7-175) finds that the
majority of these studies suggest that there is an elevated risk of
total nonaccidental mortality associated with acute exposure to
O₃, especially in the summer or warm season when
O₃ levels are typically high, with somewhat larger effect
estimate sizes for associations with cardiovascular mortality.
Overall, the Criteria Document (p.8-78) finds that the results from
U.S. multi-city time-series studies, along with the meta-analyses,
provide relatively strong evidence for associations between short-term
O₃ exposure and all-cause mortality even after adjustment
for the influence of season and PM. The results of these analyses
indicate that copollutants generally do not appear to substantially
confound the association between O₃ and mortality. In
addition, several single-city studies observed positive associations of
ambient O₃ concentrations with total nonaccidental and
cardiopulmonary mortality.
Finally, from those studies that included assessment of
associations with specific causes of death, it appears that effect
estimates for associations with cardiovascular mortality are larger
than those for total mortality; effect estimates for respiratory
mortality are less consistent in size, possibly due to reduced
statistical power in this subcategory of mortality. For cardiovascular
mortality, the Criteria Document (p.7-106) suggests that effect
estimates are consistently positive and more likely to be larger and
statistically significant in warm season analyses. The Criteria
Document (p.8-78) concludes that these findings are highly suggestive
that short-term O₃ exposure directly or indirectly
contributes to nonaccidental and cardiopulmonary-related mortality, but
additional research is needed to more fully establish underlying
mechanisms by which such effects occur.
b. Exposure- and Risk-Based Considerations
As discussed above in section II.B, the Staff Paper also estimated
quantitative exposures and health risks associated with recent air
quality levels and with air quality that meets the current standard to
help inform judgments about whether or not the current standard
provides adequate protection of public health. In so doing, it
presented the important uncertainties and limitations associated with
the exposure and risk assessments (discussed above in section II.B and
more fully in chapters 4 and 5 of the Staff Paper).
The Staff Paper (and the CASAC) also recognized that the exposure
and risk analyses could not provide a full picture of the O₃
exposures and O₃-related health risks posed nationally. The
Staff Paper did not have sufficient information to evaluate all
relevant at-risk groups (e.g., outdoor workers) or all O₃-
related health outcomes (e.g., increased medication use, school
absences, and emergency department visits that are part of the broader
pyramid of effects discussed above in section II.A.4.d), and the scope
of the Staff Paper analyses was generally limited to estimating
exposures and risks in 12 urban areas across the U.S., and to only five
or just one area for some health effects included in the risk
assessment. Thus, national-scale public health impacts of ambient
O₃ exposures are clearly much larger than the quantitative
estimates of O₃-related incidences of adverse health effects
and the numbers of children likely to experience exposures of concern
associated with recent air quality or air quality that just meets the
current or alternative standards. On the other hand, inter-individual
variability in responsiveness means that only a subset of individuals
in each group estimated to experience exposures exceeding a given
benchmark exposure of concern level would actually be expected to
experience such adverse health effects.
As described above in section II.B, the Staff Paper estimated
exposures and risks for the three most recent years (2002-2004) for
which data were available at the time of the analyses. Within this 3-
year period, 2002 was a year with relatively higher O₃
levels in most, but not all, areas and simulation of just meeting the
current standard based on 2002 air quality data provides a generally
more upper-end estimate of exposures and risks, while 2004 was a year
with relatively lower O₃ levels in

[[Continued on page 37867]]

Notices For
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994

National Ambient Air Quality Standards for Ozone

Local Navigation