National Ambient Air Quality Standards for Particulate Matter
Federal RegisterJan 17, 2006
Ask Donna
What actually matters in this document.
Text
ENVIRONMENTAL PROTECTION AGENCY
40 CFR Part 50
[OAR-2001-0017; FRL-8015-8]
RIN 2060-AI44
National Ambient Air Quality Standards for Particulate Matter
AGENCY:
Environmental Protection Agency (EPA).
ACTION:
Proposed rule.
SUMMARY:
Based on its review of the air quality criteria and national ambient air quality standards (NAAQS) for particulate matter (PM), EPA proposes to make revisions to the primary and secondary NAAQS for PM to provide requisite protection of public health and welfare, respectively, and to make corresponding revisions in monitoring reference methods and data handling conventions for PM.
With regard to primary standards for fine particles (particles generally less than or equal to 2.5 micrometers (μm) in diameter, PM
2.5
), EPA proposes to revise the level of the 24-hour PM
2.5
standard to 35 micrograms per cubic meter (μg/m
3
), providing increased protection against health effects associated with short-term exposure (including premature mortality and increased hospital admissions and emergency room visits) and to retain the level of the annual PM
2.5
standard at 15 μg/m
3
, continuing protection against health effects associated with long-term exposure (including premature mortality and development of chronic respiratory disease). The EPA solicits comment on alternative levels of the 24-hour PM
2.5
standard (down to 25 μg/m
3
and up to 65 μg/m
3
) and the annual PM
2.5
standard (down to 12 μg/m
3
), and on alternative approaches for selecting the standard levels.
With regard to primary standards for particles generally less than or equal to 10 μm in diameter (PM
10
), EPA proposes to revise the 24-hour PM
10
standard in part by establishing a new indicator for thoracic coarse particles (particles generally between 2.5 and 10 μm in diameter, PM
10-2.5
), qualified so as to include any ambient mix of PM
10-2.5
that is dominated by resuspended dust from high-density traffic on paved roads and PM generated by industrial sources and construction sources, and excludes any ambient mix of PM
10-2.5
that is dominated by rural windblown dust and soils and PM generated by agricultural and mining sources. The EPA proposes to set the new PM
10-2.5
standard at a level of 70 μg/m
3
, continuing to provide a generally equivalent level of protection against health effects associated with short-term exposure (including hospital admissions for cardiopulmonary diseases, increased respiratory symptoms and possibly premature mortality). Also, EPA proposes to revoke, upon finalization of a primary 24-hour standard for PM
10-2.5
, the current 24-hour PM
10
standard in all areas of the country except in areas where there is at least one monitor located in an urbanized area (as defined by the U.S. Bureau of the Census) with a minimum population of 100,000 that violates the current 24-hour PM
10
standard based on the most recent three years of data. In addition, EPA proposes to revoke the current annual PM
10
standard upon promulgation of this rule. The EPA solicits comment on alternative approaches for selecting the level of a 24-hour PM
10-2.5
standard, on alternative approaches based on retaining the current 24-hour PM
10
standard, and on revoking and not replacing the 24-hour PM
10
standard.
With regard to secondary PM standards, EPA proposes to revise the current standards by making them identical to the suite of proposed primary standards for fine and coarse particles, providing protection against PM-related public welfare effects including visibility impairment, effects on vegetation and ecosystems, and materials damage and soiling. Also, EPA solicits comment on adding a new sub-daily PM
2.5
standard to address visibility impairment.
DATES:
Written comments on this proposed decision must be received by April 17, 2006.
ADDRESSES:
Submit your comments, identified by Docket ID No. EPA-HQ-OAR-2001-0017 by one of the following methods:
•
http://www.regulations.gov:
Follow the on-line instructions for submitting comments.
• E-mail:
a-and-r-Docket@epa.gov.
• Fax: 202-566-1749.
• Mail: Docket ID No. EPA-HQ-OAR-2001-0017, Environmental Protection Agency, Mailcode: 6102T, 1200 Pennsylvania Avenue, NW., Washington, DC 20460. Please include a total of two copies.
• Hand Delivery: Environmental Protection Agency, EPA West Building, Room B102, 1301 Constitution Avenue, 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-2001-0017. 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
http://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
http://www.regulations.gov
or e-mail. The
http://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
http://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
http://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
http://www.regulations.gov
or in hard copy at the Air and Radiation Docket and Information Center, EPA/DC, EPA West, Room B102, 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
February in Philadelphia, Chicago, and San Francisco, and will announce in a separate
Federal Register
notice the date, time, and address of the public hearings on this proposed decision.
FOR FURTHER INFORMATION CONTACT:
Dr. Erika Sasser, mail code C539-01, Air Quality Strategies and Standards Division, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711, telephone: (919) 541-3889, e-mail:
sasser.erika@epa.gov.
SUPPLEMENTARY INFORMATION:
General Information
A. What Should I Consider As I Prepare My Comments for EPA?
1.
Submitting CBI
. Do not submit this information to EPA through
http://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 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 are available on EPA Web sites. The Air Quality Criteria for Particulate Matter (Criteria Document) (two volumes, EPA/600/P-99/002aF and EPA/600/P-99/002bF, October 2004) 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 “Particulate Matter”. The Staff Paper, human health risk assessment, and several 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/pm/s_pm_cr_sp.html,
and the risk assessment and technical documents are available at
http://www.epa.gov/ttn/naaqs/standards/pm/s_pm_cr_td.html.
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 today's preamble:
I. Background
A. Legislative Requirements
B. Review of Air Quality Criteria and Standards for PM
C. Related Control Programs to Implement PM Standards
D. Overview of Current PM NAAQS Review
II. Rationale for Proposed Decisions on Primary PM
2.5
Standards
A. Health Effects Related to Exposure to Fine Particles
1. Mechanisms
2. Nature of Effects
3. Integration and Interpretation of the Health Evidence
4. Sensitive Subgroups for PM
2.5
-Related Effects
5. PM
2.5
-Related Impacts on Public Health
B. Quantitative Risk Assessment
1. Overview
2. Scope and Key Components
3. Risk Estimates and Key Observations
C. Need for Revision of the Current Primary PM
2.5
Standards
D. Indicator of Fine Particles
E. Averaging Time of Primary PM
2.5
Standards
F. Form of Primary PM
2.5
Standards
1. 24-Hour PM
2.5
Standard
2. Annual PM
2.5
Standard
G. Level of Primary PM
2.5
Standards
1. 24-Hour PM
2.5
Standard
2. Annual PM
2.5
Standard
H. Proposed Decisions on Primary PM
2.5
Standards
III. Rationale for Proposed Decisions on the Primary PM
10
Standards
A. Health Effects Related to Exposure to Thoracic Coarse Particles
1. Mechanisms
2. Nature of Effects
3. Integration and Interpretation of the Health Evidence
4. Sensitive Subgroups for Effects of Thoracic Coarse Particle Exposure
5. Impacts on Public Health from Thoracic Coarse Particle Exposure
B. Quantitative Risk Assessment
C. Need for Revision of the Current Primary PM
10
Standards
D. Indicator of Thoracic Coarse Particles
E. Averaging Time of Primary PM
10-2.5
Standard
F. Form of Primary PM
10-2.5
Standard
G. Level of Primary PM
10-2.5
Standard
H. Proposed Decisions on Primary PM
10-2.5
Standard
IV. Rationale for Proposed Decisions on Secondary PM Standards
A. Visibility Impairment
1. Visibility Impairment Related to Ambient PM
2. Need for Revision of the Current Secondary PM Standards for Visibility Protection
3. Indicator of PM for Secondary Standard to Address Visibility Impairment
4. Averaging Time of a Secondary PM
2.5
Standard for Visibility Protection
5. Elements of a Secondary PM
2.5
Standard for Visibility Protection
B. Other PM-related Welfare Effects
1. Nature of Effects
2. Need for Revision of Current Secondary PM Standards to Address Other PM-related Welfare Effects
C. Proposed Decision on Secondary PM Standards
V. Interpretation of the NAAQS for PM
A. Proposed Amendments to Appendix N—Interpretation of the National Ambient Air Quality Standards for PM
2.5
1. General
2. PM
2.5
Monitoring and Data Reporting Considerations
3. PM
2.5
Computations and Data Handling Conventions
4. Secondary Standard
5. Conforming Revisions
B. Proposed Appendix P—Interpretation of the National Ambient Air Quality Standards for PM
10-2.5
1. General
2. PM
2.5
Data Reporting Considerations
3. PM
10-2.5
Computations and Data Handling Conventions
4. Exceptional Events
VI. Reference Methods for the Determination of Particulate Matter as PM
2.5
and PM
10-2.5
A. Proposed Appendix O: Reference Method for the Determination of Coarse Particulate Matter (as PM
10-2.5
) in the Atmosphere
1. Purpose of the New Reference Method
2. Rationale for Selection of the New Reference Method
3. Consideration of Other Methods for the Federal Reference Method
4. Consideration of Automated Method
5. Relationship of Proposed FRM to Transportation Equity Act Requirements
6. Use of the Proposed Federal Reference Method
7. Basic Requirements of the Proposed Federal Reference Method Sampler
8. Other Important Aspects of the Proposed Federal Reference Method Sampler
B. Proposed Amendments to Appendix L—Reference Method for the Determination of Fine Particulate Matter (as PM
2.5
) in the Atmosphere
VIII. Statutory and Executive Order Reviews
A. Executive Order 12866: Regulatory Planning and Review
B. Paperwork Reduction Act
C. Regulatory Flexibility Act
D. Unfunded Mandates Reform Act
E. Executive Order 13132: Federalism
F. Executive Order 13175: Consultation and Coordination With Indian Tribal Governments
G. Executive Order 13045: Protection of Children from Environmental Health and Safety Risks
H. Executive Order 13211: Actions That Significantly Affect Energy Supply, Distribution or Use
I. National Technology Transfer Advancement Act
J. Executive Order 12898: Federal Actions To Address Environmental Justice in Minority Populations and Low-Income Populations
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.”
In setting standards that are “requisite” to protect public health and welfare, as provided in section 109(b), EPA's task is to establish standards that are neither more nor less stringent than necessary for these purposes. In establishing “requisite” primary and secondary standards, EPA may not consider the costs of implementing the standards. See generally
Whitman
v.
American Trucking Associations,
531 U.S. 457, 465-472, 475-76 (2001).
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 (D.C. Cir 1980),
cert. denied,
449 U.S. 1042 (1980);
American Petroleum Institute
v.
Costle,
665 F.2d 1176, 1186 (D.C. Cir. 1981),
cert. denied,
455 U.S. 1034 (1982). Both kinds of uncertainties are components of the risk associated with pollution at levels below those at which human health effects can be said to occur with reasonable scientific certainty. Thus, in selecting primary standards that include an adequate margin of safety, the Administrator is seeking not only to prevent pollution levels that have been demonstrated to be harmful but also to prevent lower pollutant levels that may pose an unacceptable risk of harm, even if the risk is not precisely identified as to nature or degree. 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, supra,
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 sensitive population(s) at risk, and the kind and degree of the uncertainties that must be addressed. The selection of any particular approach to providing an adequate margin of safety is a policy choice left specifically to the Administrator's judgment.
Lead Industries Association
v.
EPA, supra,
647 F.2d at 1161-62.
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. Review of Air Quality Criteria and Standards for PM
Particulate matter is the generic term for a broad class of chemically and physically diverse substances that exist as discrete particles (liquid droplets or solids) over a wide range of sizes. Particles originate from a variety of anthropogenic stationary and mobile sources as well as from natural sources. Particles may be emitted directly or formed in the atmosphere by transformations of gaseous emissions such as sulfur oxides (SO
X
), nitrogen oxides (NO
X
), and volatile organic compounds (VOC). The chemical and physical properties of PM vary greatly with time, region, meteorology, and source category, thus complicating the assessment of health and welfare effects.
The last review of PM air quality criteria and standards was completed in July 1997 with notice of a final decision to revise the existing standards (62 FR 38652, July 18, 1997). In that decision, EPA revised the PM NAAQS in several respects. While EPA determined that the PM NAAQS should continue to focus on particles less than or equal to 10 μm in
diameter (PM
10
), EPA also determined that the fine and coarse fractions of PM
10
should be considered separately. The EPA added new standards, using PM
2.5
as the indicator for fine particles (with PM
2.5
referring to particles with a nominal mean aerodynamic diameter less than or equal to 2.5 μm), and retained PM
10
standards for the purpose of regulating the coarse fraction of PM
10
(referred to as thoracic coarse particles or coarse-fraction particles; generally including particles with a nominal mean aerodynamic diameter greater than 2.5 μm and less than or equal to 10 μm, or PM
10-2.5
). The EPA established two new PM
2.5
standards: an annual standard of 15 μg/m
3
, based on the 3-year average of annual arithmetic mean PM
2.5
concentrations from single or multiple community-oriented monitors; and a 24-hour standard of 65 μg/m
3
, based on the 3-year average of the 98th percentile of 24-hour PM
2.5
concentrations at each population-oriented monitor within an area. Also, EPA established a new reference method for the measurement of PM
2.5
in the ambient air and adopted rules for determining attainment of the new standards. To continue to address thoracic coarse particles, EPA retained the annual PM
10
standard, while revising the form, but not the level, of the 24-hour PM
10
standard to be based on the 99th percentile of 24-hour PM
10
concentrations at each monitor in an area. The EPA revised the secondary standards by making them identical in all respects to the primary standards.
Following promulgation of the revised PM NAAQS, petitions for review were filed by a large number of parties, addressing a broad range of issues. In May 1999, a three-judge panel of the U.S. Court of Appeals for the District of Columbia Circuit issued an initial decision that upheld EPA's decision to establish fine particle standards, holding that “the growing empirical evidence demonstrating a relationship between fine particle pollution and adverse health effects amply justifies establishment of new fine particle standards.”
American Trucking Associations
v.
EPA,
175 F.3d 1027, 1055-56 (D.C. Cir. 1999) (rehearing granted in part and denied in part, 195 F.3d 4 (D.C. Cir. 1999), affirmed in part and reversed in part,
Whitman
v.
American Trucking Associations,
531 U.S. 457 (2001). The Panel also found “ample support” for EPA's decision to regulate coarse particle pollution, but vacated the 1997 PM
10
standards, concluding in part that PM
10
is a “poorly matched indicator for coarse particulate pollution” because it includes fine particles.
Id.
at 1053-55. Pursuant to the court's decision, EPA removed the vacated 1997 PM
10
standards from the Code of Federal Regulations (CFR) (69 FR 45592, July 30, 2004) and deleted the regulatory provision (at 40 CFR 50.6(d)) that controlled the transition from the pre-existing 1987 PM
10
standards to the 1997 PM
10
standards (65 FR 80776, December 22, 2000). The pre-existing 1987 PM
10
standards remained in place.
Id.
at 80777.
More generally, the three-judge panel held (with one dissenting opinion) that EPA's approach to establishing the level of the standards in 1997, both for PM and for ozone NAAQS promulgated on the same day, effected “an unconstitutional delegation of legislative authority.”
Id.
at 1034-40. Although the panel stated that “the factors EPA uses in determining the degree of public health concern associated with different levels of ozone 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. Consistent with EPA's long-standing interpretation, the panel also reaffirmed prior rulings holding that in setting NAAQS EPA is “not permitted to consider the cost of implementing those standards.”
Id.
at 1040-41.
Both sides filed cross appeals on these issues to the United States Supreme Court, and the Court granted
certiorari.
In February 2001, the Supreme Court issued a unanimous decision upholding EPA's position on both the constitutional and cost issues.
Whitman
v.
American Trucking Associations,
531 U.S. 457, 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 Court of Appeals for resolution of any remaining issues that had not been addressed in that court's earlier rulings.
Id.
at 475-76. In March 2002, the Court of Appeals rejected all remaining challenges to the standards, holding under the traditional standard of judicial review that EPA's PM
2.5
standards were reasonably supported by the administrative record and were not “arbitrary and capricious.”
American Trucking Associations
v.
EPA,
283 F.3d 355, 369-72 (D.C. Cir. 2002).
In October 1997, EPA published its plans for the current periodic review of the PM criteria and NAAQS (62 FR 55201, October 23, 1997), including the 1997 PM
2.5
standards and the 1987 PM
10
standards. As part of the process of preparing an updated Air Quality Criteria Document for Particulate Matter (henceforth, the “Criteria Document”), EPA's National Center for Environmental Assessment (NCEA) hosted a peer review workshop in April 1999 on drafts of key Criteria Document chapters. The first external review draft Criteria Document was reviewed by CASAC and the public at a meeting held in December 1999. Based on CASAC and public comment, NCEA revised the draft Criteria Document and released a second draft in March 2001 for review by CASAC and the public at a meeting held in July 2001. A preliminary draft of a staff paper, Review of the National Ambient Air Quality Standards for Particulate Matter: Assessment of Scientific and Technical Information (henceforth, the “Staff Paper”) prepared by EPA's Office of Air Quality Planning and Standards (OAQPS) was released in June 2001 for public comment and for consultation with CASAC at the same public meeting. Taking into account CASAC and public comments, a third draft Criteria Document was released in May 2002 for review at a meeting held in July 2002.
Shortly after the release of the third draft Criteria Document, the Health Effects Institute (HEI)
3
announced that researchers at Johns Hopkins University had discovered problems with applications of statistical software used in a number of important epidemiological studies that had been discussed in that draft Criteria Document. In response to this significant issue, EPA took steps in consultation with CASAC to encourage researchers to reanalyze affected studies and to submit them expeditiously for peer review by a special expert panel convened at EPA's request by HEI. The results of this reanalysis and peer-review process were subsequently incorporated into a fourth draft Criteria Document, which was released in June 2003 and reviewed by CASAC and the public at a meeting held in August 2003.
3
The HEI is an independent research institute, jointly sponsored by EPA and a group of U.S. manufacturers and marketers of motor vehicles and engines, that conducts health effects research on major air pollutants related to motor vehicle emissions.
The first draft Staff Paper, based on the fourth draft Criteria Document, was released at the end of August 2003, and was reviewed by CASAC and the public at a meeting held in November 2003.
During that meeting, EPA also consulted with CASAC on a new framework for the final chapter (integrative synthesis) of the Criteria Document and on ongoing revisions to other Criteria Document chapters to address previous CASAC comments. The EPA held additional consultations with CASAC at public meetings held in February, July, and September 2004, leading to publication of the final Criteria Document in October 2004. The second draft Staff Paper, based on the final Criteria Document, was released at the end of January 2005, and was reviewed by CASAC and the public at a meeting held in April 2005. The CASAC's advice and recommendations to the Administrator, based on its review of the second draft Staff Paper, were further discussed during a public teleconference held in May 2005 and are provided in a June 6, 2005 letter to the Administrator (Henderson, 2005a). The final Staff Paper, issued in June, 2005, takes into account the advice and recommendations of CASAC and public comments received on the earlier drafts of this document. The Administrator subsequently received additional advice and recommendations from the CASAC, specifically on potential standards for thoracic coarse particles in a teleconference on August 11, 2005, and in a letter to the Administrator dated September 15, 2005 (Henderson, 2005b).
4
4
The EPA has posted on its Web site (
http://www.epa.gov/ttn/naaqs/standards/pm/s_pm_index.html
) a second edition of the Staff Paper which was prepared for the purpose of including as an attachment this September 2005 letter from CASAC.
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 organizations. The lawsuit alleged that EPA had failed to perform its mandatory duty, under section 109(d)(1), of completing the current review within the period provided by statute.
American Lung Association
v.
Whitman
(No. 1:03CV00778, D.D.C. 2003). An initial consent decree was entered by the court in July 2003 after an opportunity for public comment. The consent decree, as modified by the court, provides that EPA will sign for publication notices of proposed and final rulemaking concerning its review of the PM NAAQS no later than December 20, 2005 and September 27, 2006, respectively.
C. Related Control Programs to Implement PM Standards
States are primarily responsible for ensuring attainment and maintenance of ambient air quality standards once EPA has established them. Under section 110 of the CAA (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 sources of the pollutants involved. The States, in conjunction with EPA, also administer the prevention of significant deterioration (PSD) program (42 U.S.C. 7470-7479) for these pollutants. In addition, Federal programs provide for nationwide reductions in emissions of these and other air pollutants through the Federal Mobile Source Control Program under title II of the CAA (42 U.S.C. 7521-7574), which involves controls for automobile, truck, bus, motorcycle, nonroad or off-highway, and aircraft emissions; the new source performance standards under section 111 (42 U.S.C. 7411); and the national emission standards for hazardous air pollutants under section 112 (42 U.S.C. 7412).
As described in a recent EPA report, The Particle Pollution Report: Current Understanding of Air Quality and Emissions through 2003 (EPA, 2004b), State and Federal programs have made substantial progress in reducing ambient concentrations of PM
10
and PM
2.5
. For example, PM
10
concentrations have decreased 31 percent nationally since 1988. Regionally, PM
10
concentrations decreased most in areas with historically higher concentrations—the Northwest (39 percent decline), the Southwest (33 percent decline), and southern California (35 percent decline). Direct emissions of PM
10
have decreased approximately 25 percent nationally since 1988.
Programs aimed at reducing direct emissions of particles have played an important role in reducing PM
10
concentrations, particularly in western areas. Some examples of PM
10
controls include paving unpaved roads and using best management practices for agricultural sources of resuspended soil. Additionally, EPA's Acid Rain Program has substantially reduced sulfur dioxide (SO
2
) emissions from power plants since 1995 in the eastern United States, contributing to lower PM concentrations. Of the 87 areas that were designated nonattainment for PM
10
in the early 1990s, 64 now meet those standards. In cities that have not attained the PM
10
standards, the number of days above the standards is down significantly.
Nationally, PM
2.5
concentrations have declined by 10 percent from 1999 to 2003. Generally, PM
2.5
concentrations have also declined the most in regions with the highest concentrations—the Southeast (20 percent decline), southern California (16 percent decline), and the Industrial Midwest (9 percent decline). With the exception of the Northeast, the remaining regions posted modest declines in PM
2.5
concentrations from 1999 to 2003. Direct emissions of PM
2.5
have decreased by 5 percent nationally over the past 5 years.
National programs that affect regional emissions have contributed to lower sulfate concentrations and, consequently, to lower PM
2.5
concentrations, particularly in the Industrial Midwest and Southeast. National ozone-reduction programs designed to reduce emissions of volatile organic compounds (VOCs) and nitrogen oxides (NO
X
) also have helped reduce carbon and nitrates, both of which are components of PM
2.5
. Nationally, SO
2
emissions have declined 9 percent, NO
X
emissions have declined 9 percent, and VOC emissions have declined by 12 percent from 1999 to 2003. In eastern States affected by the Acid Rain Program, sulfates decreased 7 percent over the same period.
Over the next 10 to 20 years, national and regional regulations will make major reductions in ambient PM
2.5
levels. The Clean Air Interstate Rule (CAIR) and the NO
X
SIP Call will reduce SO
2
and NO
X
emissions from electric generating units and industrial boilers across the eastern half of the U.S., regulations to implement the current ambient air quality standards for PM
2.5
will require direct PM
2.5
and PM
2.5
precursor controls in nonattainment areas, and new national mobile source regulations affecting heavy-duty diesel engines, highway vehicles, and other mobile sources will reduce emissions of NO
X
, direct PM
2.5
, SO
2
, and VOCs. The EPA estimates that these regulations for stationary and mobile sources will cut SO
2
emissions by 6 million tons annually in 2015 from 2001 levels. Emissions of NO
X
will be cut by 9 million tons annually in 2015 from 2001 levels. Emissions of VOCs will drop by 3 million tons, and direct PM
2.5
emissions will be cut by 200,000 tons in 2015, compared to 2001 levels.
Modeling done by EPA indicates that by 2010, 18 of the 39 areas currently not attaining the PM
2.5
standards will come into attainment just based on regulatory programs already in place, including CAIR, the Clean Diesel Rules, and other Federal measures. Four more PM
2.5
areas are projected to attain the standards by 2015 based on the implementation of these programs. All areas in the eastern U.S. will have lower PM
2.5
concentrations in 2015 relative to present-day conditions. In most cases,
the predicted improvement in PM
2.5
ranges from 10 percent to 20 percent.
D. Overview of Current PM NAAQS Review
This action presents the Administrator's proposed decisions on the review of the current primary and secondary PM
2.5
and PM
10
standards. Primary standards for fine particles and for thoracic coarse particles are addressed separately below in sections II and III, respectively, consistent with the decision made by EPA in the last review and with the conclusions in the Criteria Document and Staff Paper that fine and thoracic coarse particles should continue to be considered as separate subclasses of PM pollution. Thus, the principal focus of this current review of the air quality criteria and primary standards for PM is on evidence of health effects and risks related to exposures to fine particles and to thoracic coarse particles. Secondary standards for fine and coarse-fraction particles are addressed below in section IV.
Past and current decisions to address fine particles and thoracic coarse particles separately are based in part on long-established information on differences in sources, properties, and atmospheric behavior between fine and coarse particles (EPA, 2005a, section 2.2). Fine particles are produced chiefly by combustion processes and by atmospheric reactions of various gaseous pollutants, whereas thoracic coarse particles are generally emitted directly as particles as a result of mechanical processes that crush or grind larger particles or the resuspension of dusts. Sources of fine particles include, for example, motor vehicles, power generation, combustion sources at industrial facilities, and residential fuel burning. Sources of thoracic coarse particles include, for example, resuspension of traffic-related emissions such as tire and brake lining materials, direct emissions from industrial operations, construction and demolition activities, and agricultural and mining operations. Fine particles can remain suspended in the atmosphere for days to weeks and can be transported thousands of kilometers, whereas thoracic coarse particles generally deposit rapidly on the ground or other surfaces and are not readily transported across urban or broader areas. The approach in this review to continue to address fine and thoracic coarse particles separately is reinforced by new information that advances our understanding of differences in human exposure relationships and dosimetric patterns characteristic of these two subclasses of PM pollution, as well as the apparent independence of health effects that have been associated with them in epidemiologic studies (EPA, 2004, section 3.2.3). See also
American Trucking Associations
v.
EPA
, 175 F. 3d at 1053-54, 1055-56 (EPA justified in establishing separate standards for fine and thoracic coarse particles).
Today's proposed decisions separately addressing fine and coarse particles are based on a thorough review in the Criteria Document of the latest scientific information on known and potential human health and welfare effects associated with exposure to these subclasses of PM at levels typically found in the ambient air. These proposed decisions also take into account: (1) Staff assessments in the Staff Paper of the most policy-relevant information in the Criteria Document and as well as a quantitative risk assessment; (2) CASAC advice and recommendations, as reflected in the CASAC's letters to the Administrator, discussions of drafts of the Criteria Document and Staff Paper at public meetings, and separate written comments prepared by individual members of the CASAC PM Review Panel
5
(henceforth, “CASAC Panel”), and (3) public comments received during the development of these documents, either in connection with CASAC meetings or separately.
5
The CASAC PM Review Panel is comprised of the seven members of the chartered CASAC, supplemented by fifteen subject-matter experts appointed by the Administrator to provide the types of scientific expertise relevant to this review of the PM NAAQS.
The EPA is aware that a number of new scientific studies on the health effects of PM have been published since the 2002 cutoff date for inclusion in the Criteria Document. As in the last PM NAAQS review, EPA intends to conduct a review and assessment of any significant new studies published since the close of the Criteria Document, including studies submitted during the public comment period in order to ensure that, before making a final decision, the Administrator is fully aware of the new science that has developed since 2002. In this assessment, EPA will examine these new studies in light of the literature evaluated in the Criteria Document. This assessment and a summary of the key conclusions will be placed in the rulemaking docket. A preliminary list of potentially significant new studies identified to date has been compiled and placed in the rulemaking docket for this proposal, and EPA solicits comment on other relevant studies that may be added to this list. This list includes a wide array of different types of studies that are potentially relevant to various issues discussed in the following sections, including issues related to the elements of the standards under review.
Throughout this preamble a number of conclusions, findings, and determinations by the Administrator are noted. It should be understood that these are all provisional and proposed in nature. While they identify the reasoning that supports this proposal, they are not intended to be final or conclusive in nature. The EPA invites comments on all issues involved with this proposal, including all such proposed judgments, conclusions, findings, and determinations.
II. Rationale for Proposed Decisions on Primary PM
2.5
Standards
As discussed more fully below, the rationale for the proposed revisions of the primary PM
2.5
NAAQS includes consideration of: (1) Evidence of health effects related to short- and long-term exposures to fine particles; (2) insights gained from a quantitative risk assessment; and (3) specific conclusions regarding the need for revisions to the current standards and the elements of PM
2.5
standards (i.e., indicator, averaging time, form, and level) that, taken together, would be requisite to protect public health with an adequate margin of safety.
In developing this rationale, EPA has drawn upon an integrative synthesis of the entire body of evidence of associations between exposure to ambient fine particles and a broad range of health endpoints (EPA, 2004, Chapter 9), focusing on those health endpoints for which the Criteria Document concludes that the associations are likely to be causal. This body of evidence includes hundreds of studies conducted in many countries around the world, using various indicators of fine particles. In its assessment of the evidence judged to be most relevant to making decisions on elements of the primary PM
2.5
standards, EPA has placed greater weight on U.S. and Canadian studies using PM
2.5
measurements, since studies conducted in other countries may well reflect different demographic and air pollution characteristics.
As with virtually any policy-relevant scientific research, there is uncertainty in the characterization of health effects attributable to exposure to ambient fine particles. As discussed below, however, an unprecedented amount of new research has been conducted since the last review, with important new information coming from epidemiologic, toxicologic, controlled human exposure,
and dosimetric studies. Moreover, the newly available research studies evaluated in the Criteria Document have undergone intensive scrutiny through multiple layers of peer review and extended opportunities for public review and comment. While important uncertainties remain, the review of the health effects 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 at this time. This review also provides important input to EPA's research plan for improving our future understanding of the relationships between exposures to ambient fine particles and health effects.
A. Heath Effects Related to Exposure to Fine Particles
This section outlines key information contained in the Criteria Document (Chapters 6-9 and the Staff Paper (Chapter 3) on known or potential effects associated with exposure to fine particles and their major constituents. The information highlighted here summarizes: (1) New information available on potential mechanisms for health effects associated with exposure to fine particles and constituents; (2) the nature of the effects that have been associated with ambient fine particles or fine particle constituents; (3) an integrative assessment of the evidence on fine particle-related health effects; (4) subpopulations that appear to be sensitive to effects of exposure to fine particles; and (5) the public health impact of exposure to ambient fine particles.
As was true in the last review, evidence from epidemiologic studies plays a key role in the Criteria Document's evaluation of the scientific evidence. Some highlights of the new epidemiologic evidence include:
(1) New multi-city studies that use uniform methodologies to investigate the effects of various indicators of PM on health with data from multiple locations with varying climate and air pollution mixes, contributing to increased understanding of the role of various potential confounders, including gaseous co-pollutants, on observed associations with fine particles. These studies provide more precise estimates of the magnitude of an effect of exposure to PM, including fine particles, than most smaller-scale individual city studies.
(2) More studies of various health endpoints evaluating associations between effects and fine particles and thoracic coarse particles (discussed below in section III), as well as ultrafine particles or specific components (e.g., sulfates, nitrates, metals, organic compounds, and elemental carbon) of fine particles.
(3) Numerous new studies of cardiovascular endpoints, with particular emphasis on assessment of cardiovascular risk factors or physiological changes.
(4) Studies relating population exposure to fine particles and other pollutants measured at centrally located monitors to estimates of exposure to ambient pollutants at the individual level. Such studies have led to a better understanding of the relationship between ambient fine particles levels and personal exposures to fine particles of ambient origin.
(5) New analyses and approaches to addressing issues related to potential confounding by gaseous co-pollutants, possible thresholds for effects, and measurement error and exposure misclassification.
6
6
“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; epidemiologic analyses attempt to adjust or control for potential confounders (EPA, 2004, section 8.1.3.2; EPA, 2005a, section 3.6.4). A “threshold” is a concentration below which it is expected that effects are not observed (EPA, 2004, section 8.4.7; EPA, 2005a, section 3.6.6). “Gaseous co-pollutants” generally refer to other commonly-occuring air pollutants, specifically O
3
, CO, SO
2
and NO
2
. “Measurement error” refers to uncertainty in the air quality measurements, while “exposure misclassification” includes uncertainty in the use of ambient pollutant measurements in characterizing population exposures to PM (EPA, 2004, section 8.4.5; EPA, 2005a, section 3.6.2)
(6) Preliminary attempts to evaluate the effects of fine particles from different sources (e.g., motor vehicles, coal combustion, vegetative burning, crustal
7
), using factor analysis or source apportionment methods with fine particle speciation data.
7
“Crustal” is used here to describe particles of geologic origin, which can be found in both fine- and coarse-fraction PM.
(7) Several new “intervention studies” providing evidence for improvements in respiratory or cardiovascular health with reductions in ambient concentrations of particles and gaseous co-pollutants.
In addition, the body of evidence on PM-related effects has greatly expanded with findings from studies on potential mechanisms or pathways by which particles may result in the effects identified in the epidemiologic studies. These studies include important new dosimetry, toxicologic and controlled human exposure studies, as highlighted below:
(8) Animal and controlled human exposure studies using concentrated ambient particles (CAPs), new indicators of response (e.g., C-reactive protein and cytokine levels, heart rate variability), and animal models simulating sensitive human subpopulations. The results of these studies are relevant to evaluation of plausibility of the epidemiologic evidence and provide insights into potential mechanisms for PM-related effects.
(9) Dosimetry studies using new modeling methods that provide increased understanding of the dosimetry of different particle size classes and in members of potentially sensitive subpopulations, such as people with chronic respiratory disease.
1. Mechanisms
In the last review, EPA considered the lack of demonstrated biologic mechanisms for the varying effects observed in epidemiologic studies to be an important caution in its integrated assessment of the health evidence. Much new evidence is now available on potential mechanisms or pathways for PM-related effects, ranging from effects on the respiratory system to indicators of cardiovascular response; these new findings are discussed in depth in Chapter 7 of the Criteria Document. While questions remain, the new findings have advanced our understanding of the complex and different patterns of particle deposition and clearance in the respiratory tract and provide insights into potential mechanisms for PM-related effects and support the plausibility of the findings of epidemiologic studies.
Although there are differences among the size fractions of particles, fine particles, including accumulation mode and ultrafine particles, and thoracic coarse particles can all penetrate into and be deposited in the tracheobronchial and alveolar regions of the respiratory tract (i.e., the “thoracic” regions).
8
Penetration into the tracheobronchial and alveolar regions is greater for accumulation mode particles than for coarse or ultrafine particles, since coarse and ultrafine particles are more efficiently removed from the air in the extrathoracic region than are accumulation-mode fine particles; the evidence from dosimetric studies is
reviewed in detail in Chapter 6 of the Criteria Document.
8
Particles are often classified in modes based on their distribution by characteristics such as mass, surface area, and particle number. “Coarse mode” particles are those with diameters mostly greater than the minimum in the particle mass distribution, which generally occurs between about 1 and 3 μm. “Accumulation mode” particles are those with diameters from about 0.1 μm to between about 1 and 3 μm. Ultrafine particles are generally those with diameters below about 0.1 μm (EPA, 2004, pages 2-14).
Fine particles have varying physical or chemical characteristics that may influence health responses. Physical characteristics that may be of importance are solubility or physical state of the particles (e.g., solid, liquid). Fine particle components include metals, acids, organic compounds, biogenic constituents, sulfate and nitrate salts, elemental carbon, and reactive components such as peroxides; size and surface area of the particles can also influence health responses. By way of illustration, Mauderly et al. (1998) discussed particle components or characteristics hypothesized to contribute to health, producing an illustrative list of 11 components or characteristics of interest for which some evidence existed. The list included: (1) Particle mass concentration, (2) particle size/surface area, (3) ultrafine particles, (4) metals, (5) acids, (6) organic compounds, (7) biogenic particles, (8) sulfate and nitrate salts, (9) peroxides, (10) soot, and (11) co-factors, including effects modification or confounding by co-occurring gases and meteorology. The authors stressed that this list is neither definitive nor exhaustive, and note that “it is generally accepted as most likely that multiple toxic species act by several mechanistic pathways to cause the range of health effects that have been observed” (Mauderly et al., 1998). The range of health outcomes linked with fine particle exposures is also broad, including effects on the cardiovascular and respiratory systems, and potential links with developmental effects in children (e.g., low birth weight) and death from lung cancer. It appears unlikely that the complex mixes of particles that are present in ambient air would act alone through any single pathway of response. Accordingly, it is plausible that several physiological responses might occur in concert to produce reported health endpoints.
As discussed in section 7.10 of the Criteria Document, the potential pathways for direct effects on the respiratory system include lung injury and inflammation, increased airway reactivity and asthma exacerbation, and increased susceptibility to respiratory infections. New toxicologic or controlled human exposure studies have reported some evidence of inflammatory responses in animals, as well as increased susceptibility to infections. Toxicologic studies also report evidence of lung injury, inflammation, or altered host defenses with exposure to ambient particles or particle constituents. Some toxicologic evidence, particularly from results of studies using diesel exhaust particle exposures, also indicates that PM can aggravate asthmatic symptoms or increase airway reactivity.
Potential pathways for fine particle-related effects also include systemic effects that are secondary to effects in the respiratory system. These include impairment of lung function leading to cardiac effects, pulmonary inflammation and cytokine production leading to systemic hemodynamic effects, lung inflammation leading to increased blood coagulability, and lung inflammation leading to hematopoiesis effects. While more limited than for direct pulmonary effects, some new toxicologic studies suggest that injury or inflammation in the respiratory system can lead to changes in heart rhythm, reduced oxygenation of the blood, changes in blood cell counts, and changes in the blood that can increase the risk of blood clot formation, a risk factor for heart attacks and strokes. In addition, health studies have suggested potential pathways for effects on the heart that include effects related to uptake of particles or particle constituents in the blood, and effects on the autonomic control of the heart and circulatory system. In the last review, little or no evidence was available from toxicologic studies on potential cardiovascular effects. More recent studies have provided some initial evidence that particles can have direct cardiovascular effects. Particle deposition in the respiratory system also could lead to cardiovascular effects, such as fine particle-induced pulmonary reflexes resulting in changes in the autonomic nervous system that then could affect heart rhythm. Also, inhaled fine particles could affect the heart or other organs if particles or particle constituents are released into the circulatory system from the lungs; some new evidence indicates that the smaller ultrafine particles or their soluble constituents can move directly from the lungs into systemic circulation.
The potential mechanisms and/or general pathways for effects discussed above are primarily effects related to short-term rather than long-term exposure to fine particles; for the most part, air pollution toxicologic studies are not designed to assess long-term exposure effects. While repeated occurrences of some short-term insults, such as inflammation, might contribute to long-term effects, it is likely that wholly different mechanisms are involved in the development of chronic health responses. Some mechanistic evidence is available, however, for potential carcinogenic or genotoxic effects of ambient fine particles and combustion products of coal, wood, diesel, and gasoline (discussed in section 7.8 of the Criteria Document).
Overall, the findings indicate that different health responses are linked with different particle characteristics and that both individual components and complex particle mixtures appear to be responsible for many biologic responses relevant to fine particle exposures. In evaluating the new body of evidence, the Criteria Document states: “Thus, there appear to be multiple biologic mechanisms that may be responsible for observed morbidity/mortality due to exposure to ambient PM. It also appears that many biologic responses are produced by PM whether it is composed of a single component or a complex mixture” (EPA, 2004, p. 7-206).
2. Nature of Effects
In the last review, evidence from health studies indicated that exposure to PM (using various indicators) was associated with premature mortality and indices of morbidity including respiratory hospital admissions and emergency room visits, school absences, work loss days, restricted activity days, effects on lung function and symptoms, morphological changes, and altered host defense mechanisms.
9
As reviewed in Chapter 8 of the Criteria Document, recent epidemiologic studies have continued to report associations between short-term exposure to fine particles or fine particle indicators, and effects such as premature mortality, hospital admissions or emergency department visits for respiratory disease, and effects on lung function and symptoms. In addition, recent epidemiologic studies have provided some new evidence linking short-term fine particle exposures to effects on the cardivascular system, including cardiovascular hospital admissions and more subtle indicators of cardiovascular health. Long-term exposure to PM
2.5
and sulfates has also been associated with mortality from cardiopulmonary diseases and lung cancer, and effects on the respiratory system such as decreased lung function or the development of chronic respiratory disease. The
evidence for such effects is summarized below.
9
Historical reports of dramatic pollution episodes, considered in the 1987 review of the PM NAAQS, provided clear evidence of mortality associated with high levels of PM and other pollutants, such as the air pollution episode that occurred in London in 1952 (EPA, 1996a, pp. 12-28 to 12-31).
BILLING CODE 6560-50-P
EP17JA06.048
BILLING CODE 6560-50-C
a. Effects Associated With Short-Term Exposure to Fine Particles
Numerous epidemiologic studies have demonstrated statistical associations between short-term exposure to fine particles and health outcomes ranging from total mortality to respiratory symptoms, as discussed below. Figure 1 summarizes results from both multi-city and single-city epidemiologic studies using short-term exposures to PM
2.5
, including all U.S. and Canadian studies that used direct measurements of PM
2.5
and for which effect estimates and confidence intervals were reported.
10
The central effect estimate is indicated by a diamond for each study 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. Positive effect estimates indicate increases in the health outcome with PM
2.5
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.
10
In the following discussion of specific studies, results from single-pollutant models are referred to, as shown in Figure 1, unless otherwise noted.
i. Mortality
Since the last review, a large number of new time-series studies of the relationship between short-term exposure to PM, including PM
2.5
, and mortality have been published, including several multi-city studies that are responsive to the recommendations from the last review. As discussed in section 8.2 of the Criteria Document, these include studies that have been conducted in single cities or locations in the U.S. or Canada, as well as Mexico City and locations in Europe, South America, Asia, and Australia.
Several recent multi-city studies have been published since the last review that are of particular relevance for this review. The results of multi-city studies on associations between PM
10
and mortality across 90 U.S. cities (Dominici, 2003) and across ten U.S. cities (Schwartz, 2003b), while not specifically on fine particles, have provided important new information to help address uncertainties regarding a number of issues, including model specification, potential confounding by co-pollutants and the form of concentration-response functions (EPA, 2004, section 8.2.2.3). Two multi-city studies have included measurements of PM
2.5
; one was conducted in six U.S. cities (Schwartz et al., 2003a; Klemm and Mason, 2003) and the other in eight Canadian cities (Burnett and Goldberg, 2003). In the last review, results from one multi-city study (the Six Cities study) were available, in which the authors reported significant associations for total mortality with PM
2.5
and PM
10
, but not with PM
10-2.5
. Reanalyses of Six Cities data have reported results consistent with the findings of the original study, with statistically significant increases for total mortality with short-term exposure to PM
2.5
(Schwartz, 2003a; Klemm and Mason, 2003). In a study using data from the eight largest Canadian cities, positive associations were reported for PM
2.5
, PM
10
, and PM
10-2.5
with mortality, and the association with PM
2.5
was statistically significant (Burnett and Goldberg, 2003).
Single-city studies of mortality associations with short-term exposures to fine particles have also been conducted in areas across U.S. and Canada as well as in Europe, Australia and Mexico (some using fine particle indicators such as British Smoke). In general, it can be seen in Figure 1 that the effect estimates for associations between mortality and short-term exposure to PM
2.5
are positive and a number are statistically significant, particularly when focusing on the results of studies with greater precision. For total nonaccidental mortality, the effect estimates from the multi-city and single-city studies with greater precision generally fall in a range of 2 to 6 percent increases per 25 μg/m
3
PM
2.5
.
11
Somewhat larger effect estimates have been reported for associations with cardiovascular or respiratory mortality than with total nonaccidental mortality although the confidence intervals may also be larger, especially for respiratory mortality since respiratory deaths comprise only a small proportion of total deaths (EPA, 2005a, p. 3-15). Some studies evaluated seasonal variation in effects, and there is no consistent pattern in results. The Criteria Document concludes that the results of recent epidemiologic studies are generally consistent with findings available in the previous review (EPA, 2004, p. 8-305).
11
In general, the results of studies conducted over shorter time periods and/or smaller areas have a broader range or effect estimates with larger standard errors, as shown in Figure 1.
In addition, associations have been reported between mortality and short-term exposure to a number of fine particle components, including sulfates, nitrates, metals, organic compounds and elemental carbon (EPA, 2004, Section 8.2.2.5.2), as well as gaseous precursors such as SO
2
and NO
2
and other gaseous pollutants such as CO. Further, three recent studies have used PM
2.5
speciation data to evaluate the effects of air pollutant combinations or mixtures using factor analysis or source apportionment methods to evaluate potential associations between mortality and PM
2.5
from different source categories. These studies reported that short-term exposures to fine particles from combustion sources, including motor vehicle emissions, coal combustion, oil burning and vegetative burning, were associated with increased mortality (EPA, 2004, Section 8.2.2.5.3). However, different patterns of associations between various components or source categories of fine particles and total or cardiovascular mortality are seen in different studies (EPA, 2004, p. 8-70, Tables 8-3, 8-4).
ii. Respiratory Morbidity
As discussed in Section 8.4.6.4 of the Criteria Document, recent epidemiologic studies have provided further evidence for fine particle effects on morbidity, including effects such as hospital admissions or emergency department for respiratory diseases, respiratory symptoms and lung function changes.
(a) Hospital Admissions or Emergency Department Visits for Respiratory Diseases
In the last review, results were available from one study that reported associations between PM
2.5
and hospitalization for respiratory diseases; these findings were also supported by a number of studies using other fine particle indicators. Numerous studies had also reported statistically significant associations between hospital admissions or emergency department visits for respiratory diseases short-term exposures with various indicators ambient PM, especially PM
10
, in areas where fine particles are the predominant fraction of PM
10
, such as locations in the Eastern U.S. and in Ontario, Canada (EPA, 1996a, p. 13-39).
The body of evidence has been expanded with numerous new studies in the U.S. and other countries that have reported associations between PM
2.5
and hospitalization or emergency department visits (discussed more fully in Section 8.3.2 of the Criteria Document). As shown in Figure 1, all U.S. and Canadian studies report
associations between PM
2.5
and hospitalization for all respiratory causes that are positive and statistically significant. A number of studies have also reported findings for hospital admissions for individual disease categories (COPD, pneumonia, and asthma) that are positive, but not always statistically significant, perhaps due to smaller sample sizes for the specific respiratory diseases. The effect estimates for respiratory hospital admissions tend to fall in the range of 5 to 15 percent per 25 μg/m
3
PM
2.5
.
12
In addition, several studies have reported positive, statistically significant associations between exposure to PM
2.5
and emergency department visits for respiratory diseases. The effect estimates for these associations range up to about 25 percent per 25 μg/m
3
PM
2.5
(EPA, 2005a, pp. 3-20, 3-21).
12
Some studies have evaluated seasonal variation in effects, and no consistent pattern is apparent in the results. For example, stronger associations were reported between PM
2.5
and asthma hospitalization in the warmer season in Seattle (Sheppard et al., 2003) but in the cooler season in Los Angeles (Nauenberg and Basu, 1999).
(b) Respiratory Symptoms and Lung Function Changes
Associations between short-term exposure to PM
2.5
and symptoms in U.S. and Canadian studies are presented in Figure 1. As discussed in Section 8.3.3 of the Criteria Document, a number of new studies have reported significant associations between short-term exposure to PM and increased respiratory symptoms (e.g., cough, wheeze, shortness of breath) and decreased lung function in people with asthma. In studies of nonasthmatic subjects, there were generally positive associations between short-term PM
2.5
exposures and respiratory symptoms that often were not statistically significant and the results for changes in lung function were somewhat inconsistent. The Criteria Document concludes that the findings of these studies suggest associations with fine PM in reduced lung function and increased respiratory symptoms. For example, significant associations were reported between ambient PM
2.5
and lower respiratory symptoms in children in a number of U.S. cities (Schwartz and Neas, 2000), and significant associations were found with reduced lung function in Philadelphia (Neas et al., 1999). These findings are supported by results from numerous studies conducted in Europe and Central and South America. The Criteria Document finds that the recent epidemiologic findings are consistent with those of the previous review in showing associations with both respiratory symptom incidence and decreased lung function (EPA, 2004, Section 8.4.6.4).
iii. Cardiovascular Morbidity
In the last review, none of the available studies had evaluated associations between exposure to PM and cardiovascular morbidity, though some studies had reported associations with cardiopulmonary morbidity. In this area, the evidence on PM-related effects has been greatly expanded. Numerous recent studies, including multi-city analyses, have reported significant associations between short-term exposures to PM and health endpoints related to cardiovascular morbidity, including hospitalization or emergency department visits for cardiovascular diseases, incidence of myocardial infarction, cardiac arrhythmia, changes in heart rate or heart rate variability and changes in cardiac health indicators such as fibrinogen or C-reactive protein (EPA, 2004, section 9.2.3.2.1).
(a) Hospital Admissions and Emergency Department Visits for Cardiovascular Diseases
Several recent studies, including multi-city analyses, have reported significant associations between short-term exposures to various PM indicators and hospital admissions or emergency department visits for cardiovascular diseases. Among the studies using PM
2.5
measurements are a number of single-city analyses of hospitalization or emergency department visits for cardiovascular diseases. As shown in Figure 1, studies conducted in Los Angeles, Toronto and Detroit have reported associations with hospital admissions or emergency department visits for all cardiovascular diseases that are positive and statistically significant or nearly so (Burnett et al., 1997; Ito, 2003; Moolgavkar, 2003). As was true for respiratory diseases, the results for specific diseases (ischemic heart disease, dysrhythmia, congestive heart disease or heart failure, and stroke) are positive but often not statistically significant. The effect estimates reported for associations with hospitalization for cardiovascular diseases range from about 1 to 10 percent per 25 μg/m
3
PM
2.5
(EPA, 2004, p. 8-310); effect estimates reported for associations with emergency department visits are generally somewhat larger.
(b) Cardiovascular Health Indicators
In addition to the greatly expanded body of evidence on hospitalization or emergency department visits for cardiovascular diseases, new epidemiologic studies have also reported associations with more subtle physiological changes in the cardiovascular system with short-term exposures to PM, particularly PM
10
and PM
2.5
(EPA, 2004, p. 9-67). Associations between short-term exposures to ambient PM (often using PM
10
) have been reported with measures of changes in cardiac function such as arrhythmia, alterations in electrocardiogram (ECG) patterns, heart rate or heart rate variability changes, although the Criteria Document urges caution in drawing conclusions regarding the effects of PM on heart rhythm, recognizing the need for further research to more firmly establish and understand links between particles and these more subtle endpoints. Recent studies have also reported increases in blood components or biomarkers such as increased levels of C-reactive protein and fibrinogen. Several of these studies report significant associations between various cardiovascular endpoints and short-term PM
2.5
exposures, including one in which statistically significant associations were reported between onset of myocardial infarction and short-term PM
2.5
exposures averaged over 2 and 24 hours (EPA, 2004, p. 8-165; Peters et al., 2001). In this study, the effect estimates for the two averaging periods are quite similar in magnitude suggesting that for certain health outcomes very short-term fine particle concentration fluctuations are important (EPA, 2004, p. 9-42; Peters et al., 2001). These new epidemiologic findings provide important insight into potential biologic mechanisms that could underlie associations between short-term PM exposure and cardiovascular mortality and hospitalization that have been reported previously.
b. Effects Associated With Long-Term Exposure to Fine Particles
In the last review, results were available from several cohort studies that suggested associations between long-term exposure to PM (using various indicators) and both mortality and respiratory morbidity. Two studies of adult populations (the Six Cities and ACS studies) reported associations between increases in mortality and long-term exposure to PM
2.5
, and results of a 24-city study indicated that long-term exposure to fine particles was associated with increased respiratory illness in children.
As discussed below, the new evidence available in the current review includes an extensive reanalysis of data from the Six Cities and ACS studies, new analyses using updated data from the ACS and California Seventh Day
Adventist (AHSMOG) studies, and a new analysis using data from a cohort of veterans. In addition, new studies have been published on the association between long-term exposure to fine particles and respiratory morbidity using data from a cohort of schoolchildren in Southern California. In general, the newly available evidence has supported earlier findings, and the results of reanalyses have increased confidence in the associations reported in previous prospective cohort studies.
i. Mortality
In the 1996 Criteria Document, statistically significant associations between long-term exposure to both PM
2.5
and sulfates and mortality were reported in studies from the Six Cities and ACS cohorts (Dockery et al., 1993; Pope et al., 1995). These studies reported effect estimates of 6.6 percent (95 percent CI: 3.5, 9.8) increases in total mortality per 10 μg/m
3
PM
2.5
in the ACS study and 13 percent (95 percent CI: 4.2, 23) increases in total mortality per 10 μg/m
3
PM
2.5
in the Six Cities study, with somewhat larger effect estimates reported for cardiopulmonary mortality (EPA, 2004, p. 8-117). A number of reviewers raised questions about the adequacy of adjustments for potential confounders and other issues (61 FR 65642, December 13, 1996). Subsequently, as discussed in more detail in Section 8.2.3 of the Criteria Document, the Health Effects Institute conducted a major reanalysis of the data from the Six Cities and ACS studies by a group of independent investigators to address questions and uncertainties raised about these prospective cohort studies. The reanalysis included two major components, a replication and validation study and a sensitivity analysis. In the first part of the reanalysis, the investigators validated the data used by the original investigators in both studies, and they were able to replicate the original results. The results confirmed the original investigators' findings of associations with both total and cardiorespiratory mortality, and the authors reported that the results were not dependent on the computer programs used in the original analyses (EPA, 2004, p. 8-91; Krewski et al., 2000, p. 91).
The second component of the reanalysis project evaluated an array of different models and variables to determine whether the original results would remain robust to different analytic assumptions. This included controlling for other individual level variables, such as cigarette smoking, alcohol consumption, obesity and occupational exposures to dusts or other pollutants, and evaluation of the sensitivity of results to the addition of a range of additional city-level variables such as population change, income, education levels, and access to health care. The sensitivity analysis included assessment of effects in different subgroups of the population. The investigators also evaluated the sensitivity of the results to the inclusion of gaseous co-pollutants, and tested the effects of different statistical modeling approaches, including methods to adjust for spatial patterns, such as the correlation in pollutant levels between cities.
The authors found that adjustment for individual-level variables did not alter the results for the association between long-term PM
2.5
or sulfate exposure and mortality (Krewski et al., 2000, p. 218). In addition, in most (but not all) cases the associations between mortality and long-term exposure to PM
2.5
and sulfates were unchanged when additional city-level variables were added to the models (Krewski et al., 2000, p. 233). Analyses to assess the potential modification of effects in different subgroups of the population found, for the most part, little difference in effects for different subgroups. However, education level was found to modify the estimated effect of fine particles, in that associations were statistically significant for those subgroups with lower education levels, whereas the effect estimates from associations for the subgroup with better than high school education were appreciably smaller and were statistically insignificant. The authors suggest that educational attainment may be a marker for lower socioeconomic status and thus greater vulnerability to fine particle-related effects (EPA, 2004, p. 8-94; Krewski et al., 2000, p. 232).
13
13
In multivariate models, the association found between mortality and long-term PM
2.5
exposure was little changed with addition of education level to the model (Krewski et al., 2000, p. 184). This indicates that education level was not a confounder in the relationship between fine particles and mortality, but the relationship between fine particles and mortality is larger in the population subsets with lower education in this study and not statistically significant in the population subset with the highest education (EPA, 2004, p. 8-100).
In single-pollutant models, none of the gaseous co-pollutants was significantly associated with mortality except SO
2
. Further reanalysis included multi-pollutant models with the gaseous pollutants, and the associations between mortality and both fine particles and sulfates were unchanged in these models, except when SO
2
was included, which decreased the size of the effect estimates for PM
2.5
to one-sixth of its original value and for sulfates to less than one-third of its original value (EPA, 2004, p. 8-136; Krewski et al., 2000, pp. 183-184).
14
However, the regional association of SO
2
and PM
2.5
was relatively high, such that the effects of the separate pollutants could not be distinguished. The authors conclude that these findings support the notion that increased mortality may be attributable to more than one component of ambient air pollution, and that throughout the reanalyses, fine particles, sulfates, and SO
2
demonstrated positive associations with mortality (Krewski et al., 2000, p. 233-234). As discussed more generally in the Criteria Document, this result may be reflecting the relatively high correlation between PM
2.5
levels and SO
2
levels that would be expected in cities across the industrial Midwest and northeastern states, the role that SO
2
has as a precursor to sulfate components in the mix of PM
2.5
, and/or the likelihood that SO
2
is part of the causal pathway linking exposure to PM
2.5
to adverse health outcomes (EPA, 2004, section 8.1.3.2).
14
For a 24.5 μg/m
3
change in PM
2.5
, the relative risk for the association between mortality and PM
2.5
alone was 1.20 (95 percent CI: 1.11-1.29), and after adjustment for SO
2
it was 1.03 (95 percent CI: 0.95-1.13). The relative risk for SO
2
alone was 1.49 (95 percent CI: 1.36-1.64) and after adjustment for PM
2.5
was 1.46 (95 percent CI: 1.32-1.63) (Krewski et al., 2000, p. 184). The relative risk for sulfates alone was 1.28 (95 percent CI: 1.18-1.40) and after adjustment for SO
2
it was 1.14 (95 percent CI: 1.04-1.25) (Krewski et al., 2000, p. 184). These relative risks for PM
2.5
are equivalent to effect estimates of 7.5 percent and 1.2 percent increases in mortality per 10 μg/m
3
, in single-pollutant and two-pollutant models, respectively.
Finally, Krewski and colleagues used several methods to address spatial patterns in the data; for example, concentrations of air pollutants may be correlated between cities within a region. These analyses were primarily based on sulfate concentrations, since more cities had data for sulfates than for fine particles. Addressing spatial patterns in the data generally reduced the size of the association between sulfates and mortality, but the models all continued to show associations between mortality risk and long-term sulfate exposures, although not all were statistically significant (Krewski et al., 2000, p. 228). Overall, considering the results of the extensive set of replication and sensitivity analyses, the authors report that the reanalysis confirmed the association between mortality and fine particle and sulfate exposures (EPA, 2004, p. 8-95; Krewski et al., 2000).
In addition, extended analyses were conducted for the ACS cohort study that included follow-up health data and air quality data from the new fine particle
monitoring network for 1999-2000. In this study of the expanded ACS cohort, significant associations were reported between long-term exposure to fine particles (using various averaging periods for air quality concentrations) and premature mortality from all causes, cardiopulmonary diseases, and lung cancer (Pope et al., 2002; EPA, 2004, 8-102). This extended analysis included the use of more recent data on fine particle concentrations, as well as data on gaseous co-pollutant concentrations, though no multi-pollutant model results are presented. Further evaluation of the influence of other covariates (e.g., dietary intake data, occupational exposure) used methods similar to those in the reanalysis described above, and new statistical approaches were used for modeling the PM-mortality relationship as well as adjusting for spatial correlation (EPA, 2004, section 8.2.3.2.2). The investigators reported that the associations found with fine particle and sulfate concentrations were not markedly affected by adjustment for numerous socioeconomic variables, demographic factors, environmental variables, indicators of access to health services or personal health variables (e.g., dietary factors, alcohol consumption, body mass index). Similar to the results of Krewski et al. (2000), education level was found to be a modifier in the relationship between fine particles and mortality, in that associations were statistically significant for those subgroups with lower education levels, whereas effect estimates from associations for those with better than a high school education were close to zero and were statistically insignificant.
There are also new analyses using updated data from the AHSMOG cohort. These include estimated PM
2.5
concentrations from visibility data, along with new health information from continued follow-up of the Seventh Day Adventist cohort. Positive associations were reported for mortality with PM
2.5
in males, but the estimates were generally not statistically significant (Abbey et al., 1999; McDonnell et al., 2000; EPA, 2004, pp. 8-110 and 8-117). In addition, one new set of analyses was done using subsets of PM exposure and mortality time periods and data from a Veterans Administration (VA) cohort of hypertensive men. The investigators report inconsistent and largely nonsignificant associations between PM exposure (including, depending on availability, TSP, PM
10
, PM
2.5
, PM
15
and PM
15-2.5
) and mortality (EPA, 2004, pp. 8-110 to 8-111; Lipfert et al., 2000b).
The Criteria Document and Staff Paper place greatest weight on the findings of the Six Cities and ACS studies (including reanalyses and extended analyses) that include measured fine particle data (in contrast with AHSMOG effect estimates based on TSP or visibility measurements), have study populations more similar to the general population than the VA study cohort, and have been replicated and examined through exhaustive reanalysis (EPA, 2005a, at 5-22; see also EPA, 2004, at 8.2.3.2.5.). In these studies, effect estimates for deaths from all causes fall in a range of 6 to 13 percent increased risk per 10 μg/m
3
PM
2.5
, while effect estimates for deaths from cardiopulmonary causes fall in a range of 6 to 19 percent per 10 μg/m
3
PM
2.5
. For lung cancer mortality, the effect estimate was a 13 percent increase per 10 μg/m
3
PM
2.5
in the results of the extended analysis from the ACS cohort (Pope et al., 2002; CD, Table 8-12).
The prospective cohort studies have used air quality measurements averaged over long periods of time, such as several years, to characterize the long-term ambient levels in the community. The exposure comparisons are basically cross-sectional in nature, and do not provide evidence concerning any temporal relationship between exposure and effect (EPA, 2004, p. 9-42). As discussed in the Criteria Document, it is not easy to differentiate the role of historic exposures from more recent exposures, leading to potential exposure measurement error that is increased if average PM concentrations change over time differentially between areas (EPA, 2004, p. 5-118). Several new studies have used different air quality periods for estimating long-term exposure and tested associations with mortality for the different exposure periods. As discussed in section 3.6.5.4 of the Staff Paper, these analyses indicate that averaging PM concentrations over a longer time period results in stronger associations, and that the longer series of data is likely a better indicator of cumulative exposure. Thus, in evaluating these findings, EPA has focused on the results of analyses using fine particle or sulfate measurements for the longer exposure periods in the studies.
ii. Respiratory Morbidity
In the last review, several studies had reported that long-term PM exposure was linked with increased respiratory disease and decreased lung function. One study, using data from 24 U.S. and Canadian cities (“24 Cities” study), reported associations with these effects and long-term exposure to fine particles or acidic particles, but not with PM
10
exposure (Dockery et al., 1996; Raizenne et al., 1996). More specifically, statistically significant associations were reported between long-term exposure to fine particles and decreases in several measures of lung function evaluated at a single point in time (Raizenne et al., 1996). In addition, positive but not statistically significant associations were reported between long-term exposure to fine particles and prevalence of a range of respiratory conditions (e.g., asthma, bronchitis, chronic cough) (Dockery et al., 1996).
In the current review, new studies conducted in the U.S. have been based on data from cohorts of schoolchildren in 12 Southern California Communities and an adult cohort of Seventh Day Adventists (AHSMOG) (EPA, 2004, section 8.3.3.2). Information specifically on associations with long-term PM
2.5
exposures are available from the Southern California children's cohort study. Early findings from cross-sectional analyses done at the beginning of the study suggested associations between long-term PM
2.5
exposures and respiratory morbidity, but the findings were generally not statistically significant.
15
Later publications from this cohort have reported associations with lung function growth in children over four-year follow-up periods. In a study of a cohort of children followed from 4th to 7th grade, some measures of decreases in lung function growth were statistically significantly associated with increasing exposure to PM
2.5
, whereas in a second cohort of 4th graders, the associations generally did not reach statistical significance (Gauderman et al., 2002). Decreases in measures of lung function growth were also reported for cohorts of older children, but the associations did not reach statistical significance (Gauderman et al., 2000). The Criteria Document finds that these studies “provide the best evidence” on effects of long-term fine particle exposure (EPA, 2004, p. 8-314). However, this is the only cohort study to have evaluated associations with decreases in lung function growth in children over time. Considered together, the Criteria Document finds that the evidence from these studies indicates that long-term PM
2.5
exposures may
result in chronic respiratory effects (EPA, 2004, p. 8-314).
15
In an initial report on the prevalence of respiratory illnesses reported at the beginning of the study, positive associations, though not statistically significant, were reported between long-term PM
2.5
exposure and risk of bronchitis and cough only in the subset of children with asthma (McConnell et al., 1999), and no significant associations with long-term PM
2.5
exposure were reported for the full cohort (Peters et al., 1999a). In addition, long-term PM
2.5
exposure was associated with decreases in some lung function measurements made at that time, but the associations were only statistically significant for females (Peters et al., 1999b).
3. Integration and Interpretation of the Health Evidence
In evaluating the evidence from epidemiologic studies, the Criteria Document and Staff Paper focused on well-recognized criteria, including the strength of associations; robustness of reported associations to the use of alternative model specifications, potential confounding by co-pollutants, and exposure misclassification related to measurement error; consistency of findings in multiple studies of adequate power, and in different persons, places, circumstances and times; the nature of concentration-response relationships; and information from so-called natural experiments or intervention studies. These evaluations addressed key methodological issues that are relevant to interpretation of evidence from epidemiologic studies. Further, findings from epidemiologic studies were integrated with experimental (e.g., dosimetric and toxicologic) studies, in considering the extent of coherence and biological plausibility of effects observed in epidemiologic studies. This integrative assessment provided the basis for the judgments made in the Criteria Document and Staff Paper about the extent to which causal inferences can be made about observed associations between health endpoints and PM
2.5
(as well as other indicators or constituents of ambient PM), acting alone and/or in combination with other pollutants. Key elements of these evaluations are briefly summarized below.
(1) For short-term exposures to fine particles, in considering the magnitude and statistical strength of the associations, there is a pattern of positive and often statistically significant associations for cardiovascular and respiratory health outcomes with short-term exposure to PM
10
and PM
2.5
. Of particular note are several multi-city studies that have yielded relative risk estimates for associations between short-term exposure to various indices of PM and mortality or morbidity. Although small in size, the effect estimates from multi-city studies have great precision due to the statistical power of the studies. New analyses of pre-existing cohorts with studies of long-term exposure to fine particles are available that confirm and strengthen conclusions from the previous review, although the effect estimates are sensitive to education level, co-pollutant effects of SO
2
, and spatial correlation, as discussed above.
(2) The Criteria Document and Staff Paper have evaluated the robustness of epidemiologic associations in part by considering the effect of differences in statistical model specification, potential confounding by co-pollutants and exposure error on PM-health associations (EPA, 2004, section 9.2.2.2; EPA, 2005a, sections 3.4.2 and 3.6).
As discussed in section 8.4.2 of the Criteria Document and section 3.6.3 of the Staff Paper, the influence of alternative modeling strategies on epidemiologic study results was assessed, with a particular focus on the recent set of analyses to address statistical modeling questions in epidemiologic studies for short-term PM exposures. Numerous recent studies used a certain type of statistical method (i.e., generalized additive methods (GAM)) in widely used statistical software (Splus), and it was discovered that the default program settings could potentially result in biased effect estimates for associations between pollutants and health outcomes. Results from a number of epidemiologic studies were reanalyzed to address this problem. These reanalyses also more broadly included the use of alternative statistical models and alternative methods of control for time-varying effects, such as weather or season (HEI, 2003). In general, the results of the reanalyses to address the use of default program settings in the Splus software showed little change in effect estimates for some studies; in others the effect estimates were reduced in size, though it was observed that the reductions were often not substantial (EPA, 2004, p. 9-35). For example, in comparing results for numerous studies of mortality associations with PM
10
, the Criteria Document found that the extent of reduction in effect estimates resulting from reanalysis was smaller than the variation in effect estimate size across studies (EPA, 2004, p. 8-229 and Figure 8-15). A review panel commentary on the set of reanalysis studies (using various PM indicators) notes that most studies were considered to show “little or no change” in results with initial reanalyses to address questions about the use of modeling specifications in the statistical software package (HEI, 2003, pp. 258-259).
In addition, the reanalyses also refocused attention in general on the control for relationships between health effects and weather variables in time-series epidemiologic studies; such issues had been also discussed at length in the 1996 Criteria Document (EPA, 2004, section 8.4.3.5). The reanalysis results showed greater sensitivity to the modeling approach used to account for temporal effects and weather variables than to correcting the initial problem with default settings in the use of GAM in Splus software (EPA, 2004, p. 8-236). For example, in the review panel commentary, sixteen of the reanalyzed studies were considered to have “little or no change” in results of initial reanalyses, while only two studies showed “substantial” changes (Goldberg and Burnett, 2003; some results in Ito, 2003; HEI, 2003, pp. 258-259). In contrast, four of the eight studies that were reanalyzed with additional methods to adjust for time-related variables were considered to show “substantial” changes in effect estimate size (HEI, 2003, p. 262).
The recent time-series epidemiologic studies evaluated in the Criteria Document have included some degree of control for variations in weather and seasonal variables. As summarized in the HEI review panel commentary, selecting a level of control to adjust for time-varying factors, such as temperature, in time-series epidemiologic studies involves a trade-off. For example, if the model does not sufficiently adjust for the relationship between the health outcome and temperature, some effects of temperature could be falsely ascribed to the pollution variable. Conversely, if an overly aggressive approach is used to control for temperature, the result would possibly underestimate the pollution-related effect and compromise the ability to detect a small but true pollution effect (EPA, 2004, p. 8-236; HEI, 2003, p. 266). The selection of approaches to address such variables depends in part on prior knowledge and judgments made by the investigators, for example, about weather patterns in the study area and expected relationships between weather and other time-varying factors and health outcomes considered in the study. While recognizing the need for further exploration of alternative modeling approaches for time-series analyses, the Criteria Document found that the studies included in this part of the reanalysis in general continued to demonstrate associations between PM and mortality and morbidity beyond those attributable to weather variables alone (EPA, 2004, pp. 8-340, 8-341). Further, considering the full set of reanalyses, the Criteria Document concludes that associations between short-term exposure to PM and various health outcomes are generally robust to the use of alternative modeling strategies, again recognizing that further evaluation of alternative modeling strategies was warranted (EPA, 2004, p. 9-48).
For long-term exposure to fine particles, the reanalysis and extended analyses of data from prospective cohort studies, discussed above in section II.A.2, have shown that reported associations between mortality and long-term exposure to fine particles are robust to alternative modeling strategies (Krewski et al., 2000). As stated in the reanalysis report, “The risk estimates reported by the Original Investigators were remarkably robust to alternative specifications of the underlying risk models, thereby strengthening confidence in the original findings” (Krewski et al., 2000, p. 232). In extended analysis, Krewski et al. (2000) identified model sensitivities related to education level and spatial correlation, as well as to co-pollutant effects of SO
2
, as discussed below.
The Criteria Document also included extensive evaluation of the sensitivity of PM-health responses to confounding by gaseous co-pollutants (EPA, 2004, section 8.4.3, Figures 8-16 to 8-19). Results of new multi-city short-term exposure studies, that combine data from locations with different mixes of pollutants, provide important new results. Using PM
10
, the NMMAPS results indicated that associations with mortality were not confounded by co-pollutant concentrations across 90 U.S. cities (Dominici, 2003),
16
and a similar lack of confounding was observed in a mortality study across 10 U.S. cities (Schwartz, 2003b) (EPA, 2004, Figure 8-16). That is, in these studies, the size of the effect estimates are little changed and the associations remain statistically significant in multi-pollutant models including one or more of the gaseous co-pollutants. Similar results are seen in some single-city studies using PM
2.5
for some health outcomes in which the single-pollutant model association was statistically significant (EPA, 2004, Figures 8-16 to 8-18), including the association with mortality in Santa Clara County, CA (Fairley, 2003); associations with hospital admissions in Detroit (for heart failure and pneumonia in Ito, 2003) and Seattle (for asthma in Sheppard et al., 2003); and associations with cardiovascular-related biomarkers in Boston (Gold et al., 2000). The size of the effect estimates were little changed in other studies as well in which the single-pollutant model associations were not statistically significant (e.g., for some health outcomes in Ito, 2003; for mortality in Chock et al., 2000). In yet other studies, however, for some combinations of pollutants in some areas, substantial reductions in the size of the effect estimates for PM
2.5
were observed; notably, Moolgavkar (2003) reports substantial reductions in effect estimates when CO is included in models for mortality and hospitalization in Los Angeles, and Thurston et al. (1994) and Delfino et al. (1998) report substantial reductions when O
3
is included in models for hospital admissions in Toronto and emergency department visits in Montreal, respectively.
17
It is recognized that collinearity between co-pollutants can make interpretation of such multi-pollutant model results difficult (EPA, 2004, p. 8-253). Further, associations between long-term exposure to PM
2.5
and mortality were not generally sensitive to inclusion of co-pollutants, with the notable exception of the inclusion of SO
2
in multipollutant models used in the reanalysis of the ACS study, as discussed above in section II.A.2 (EPA, 2004, p. 8-136). Overall, the Criteria Document concluded that these studies indicate that effect estimates for associations between mortality and morbidity and various PM indices are generally robust to confounding by co-pollutants, while recognizing that disentangling the effects attributable to various pollutants within an air pollution mixture is challenging (EPA, 2004, p. 9-37).
16
In the HEI Review Panel commentary on the results of the NMMAPS multi-city analyses, the Panel stated that the results did not show a confounding effect of other pollutants, observing that the PM
10
effects on mortality were not changed by addition of either O
3
, SO
2
, NO
2
or CO to the models (HEI, 2000, p. 77).
17
The correlation coefficients between concentrations of PM
2.5
and the noted co-pollutants in these studies were high; the coefficient with CO in Los Angeles was 0.58, and the coefficients with O
3
were 0.58 and 0.72 in Montreal and Toronto, respectively.
Finally, as discussed in section 3.6.2, a number of recent studies have evaluated the influence of exposure error on PM-health associations. This includes both consideration of error in measurements of PM and other co-pollutants, and the degree to which measurements from an individual monitor reflect exposures to the surrounding community. As further discussed in section 3.6.2, several studies have shown that fairly extreme conditions (e.g., very high correlation between pollutants and no measurement error in the “false” pollutant) are needed for complete “transfer of causality” of effects from one pollutant to another (EPA, 2004, p. 9-38). In comparing fine and thoracic coarse particles, the Criteria Document observes that exposure error is likely to be more important for associations with PM
10-2.5
than with PM
2.5
, since there is generally greater error in PM
10-2.5
measurements, PM
10-2.5
concentrations are less evenly distributed across a community, and less likely to penetrate into buildings (EPA, 2004, p. 9-38). Therefore, while the Criteria Document concludes that associations reported with PM
10
, PM
2.5
and PM
10-2.5
are generally robust, it recognizes that factors related to exposure error may result in reduced precision for epidemiologic associations with PM
10-2.5
(EPA, 2004, p. 9-46).
(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). The 1996 Criteria Document reported associations between short-term PM exposure and mortality or morbidity from studies conducted in locations across the U.S. as well as in other countries, and concluded that the epidemiologic data base had “general internal consistency” (EPA, 1996a, p. 13-30). New multi-city studies have allowed evaluation of consistency in effect estimates across geographic locations, using uniform statistical modeling approaches; the results suggest that effect estimates differ from one location to another, but the extent of variation is not clear. For example, the Canadian 8-city study reported no evidence of heterogeneity in city-specific results in the initial study findings; however, in the reanalysis to address model specification issues, the findings suggested more evidence of heterogeneity in associations between mortality and short-term PM
2.5
exposure (Burnett and Goldberg, 2003; EPA, 2004, p. 9-39). The Criteria Document discussed a number of factors that would be likely to cause variation in PM-health outcomes in different populations and geographic areas in section 9.2.2.3, including indicators of exposure to traffic-related pollution, population characteristics that affect susceptibility or exposure differences, distribution of PM sources, or geographic features that would affect the spatial distribution of PM (EPA, 2004, p. 9-41). In addition, the use of data collected on a 1-in-6 or 1-in-3 day schedule results in reduced statistical power, resulting in less precision for estimated effect estimates for the individual cities and increased potential variability in results (EPA, 2004, p. 9-40). Overall, the Criteria document concluded that “[f]ocusing on the studies with the most precision, it can be concluded that there is much consistency in epidemiologic evidence regarding associations between short-term and long-term exposures to fine
particles and cardiopulmonary mortality and morbidity.” (EPA, 2004, p. 9-47).
(4) The form of concentration-response relationships (e.g., linear, sigmoid) and the potential existence of thresholds was one of the important research questions remaining in the previous review. The Criteria Document recognized that it is reasonable to expect that there likely are biologic thresholds for different health effects in individuals or groups of individuals with similar innate characteristics and health status (EPA, 2004, Section 9.2.2.5). Individual thresholds would presumably vary substantially from person to person due to individual differences in genetic-level susceptibility and pre-existing disease conditions (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. The person-to-person difference in the relationship between personal exposure to PM of ambient origin and the concentration observed at a monitor may also add to the variability in observed concentration-response relationships, further obscuring potential population thresholds within the range of observed concentrations (CD, p. 9-43, 9-44).
Several new epidemiologic studies have used different modeling methods to address this question, and most have been unable to detect threshold levels in the relationship between short-term PM exposure (generally using PM
10
) and mortality; in fact, one single-city analysis suggests that statistical methods would allow detection of a threshold in the epidemiologic data if a clear threshold existed. However, a few analyses in individual cities have provided suggestions of some potential threshold levels, generally at fairly low ambient concentrations. One single-city study used PM
2.5
and PM
10-2.5
measurements in Phoenix and reported that there was suggestive evidence of a threshold for the association between mortality and short-term exposure to PM
2.5
in the range of 20-25 μg/m
3
(Smith et al., 2000; EPA, 2004, p. 8-322).
The shape of the concentration-response function for long-term exposure to PM
2.5
with mortality was evaluated using data from the ACS cohort. In the ACS reanalysis, the authors report that the concentration-response functions for PM
2.5
and all-cause and cardiopulmonary mortality demonstrate near-linear increasing trends through the range of particle levels observed in the fine particle cohort (Krewski, p. 160). However, the HEI Review Committee concluded that these results show no clear evidence either for or against overall linearity (Krewski, p. 265). In the extended ACS study, the authors reported that the associations for all-cause, cardiovascular and lung cancer mortality “were not significantly different from linear associations” (Pope, et al., 2002).
Thus, evaluation of the health effects data summarized in the Criteria Document provides no evidence to support selecting any particular population threshold for PM
2.5
. The Staff Paper also recognized, however, that it is reasonable to expect that, for individuals, there may be thresholds for specific health responses and that it is possible that such thresholds exist toward the lower end of these ranges (or below these ranges) but cannot be detected due to variability in susceptibility across a population. Even in those few studies with suggestive evidence of such thresholds, the potential thresholds are at fairly low concentrations (EPA, 2004, sections 8.4.7 and 9.2.2.5).
(5) Few studies are available that assess the extent to which reductions in ambient PM actually lead to reductions in health effects attributable to PM. As discussed in sections 8.2.3.4 and 9.2.2.6 of the Criteria Document, several epidemiologic studies were done in the Utah Valley area over a time period when a major source of PM was closed, resulting in markedly decreased PM
10
concentrations. An epidemiologic study reported that respiratory hospital admissions decreased during the plant closure time period (EPA, 2004, p. 8-131; Pope et al., 1989). Newly available controlled human exposure and animal toxicology studies, using particles extracted from stored PM
10
sampling filters from the Utah Valley, have shown inflammatory responses that are greater with extracts of particles collected during the time period of source operation than when the source was closed, suggesting that the PM from the steel mill was more harmful than other ambient PM on an equal mass basis (EPA, 2004, p. 9-73). Epidemiologic studies in Dublin, Ireland and Hong Kong also provides evidence for reduced relative risks for mortality when PM (measured as BS or PM
10
) and SO
2
were reduced as the result of interventions aimed at reducing air pollution. The Criteria Document concluded that this small group of studies add further support to the results of the hundreds of other epidemiologic studies linking ambient PM exposure to an array of health effects, and provide strong evidence that reducing emissions of PM and gaseous pollutants has beneficial public health impacts (EPA, 2004, p. 9-45 to 9-46).
(6) Several issues related to fine particle exposure time periods were assessed in the Criteria Document, as summarized in section 3.6.5 of the Staff Paper. As discussed above in this section, these include the exposure time periods used in long-term exposure studies as well as health outcome associations with very short time periods (e.g., 2-hour average). An additional issue is the time period (“lag”) between fine particle exposure and health outcome that is reported in short-term exposure study results. In these epidemiologic studies, associations are often tested for a range of lag periods, for example, with PM concentrations from the same day as the effect, and one or more days preceding the effect. In evaluating these results, it is important to consider the pattern of results that is seen across the series of lag periods. If there is an apparent pattern of results across the different lags, with positive associations reported for a series of consecutive lag periods, then selecting the single-day lag with the largest effect from a series of positive associations is likely to underestimate the overall effect size, since single-day lag effect estimates do not fully capture the risk that may be distributed over adjacent or other days (EPA, 2004, sections 8.4.4 and 9.2.2.4). For many epidemiologic studies, the authors have reported just such a pattern of associations across several consecutive lag periods (EPA, 2004, p. 8-279). However, if there is no apparent pattern or reported effects vary across lag days, any result for a single day may well be biased (CD, p. 9-42).
Some new studies have used a “distributed lag” model approach, that captures an effect of PM over a series of days following exposure.
18
Where effects are found for a series of lag periods, a distributed lag model will more accurately characterize the effect estimate size. A number of recent studies that have investigated associations with distributed lags provide effect estimates for health responses that persist over a period of time (days to weeks) after the exposure period. Effect estimates from distributed lag models are thus often, but not always, larger in size that those for single-day lag periods (EPA, 2004, p. 8-281).
18
The available studies have generally used PM
10
, but not PM
2.5
or PM
10-2.5
.
The Criteria Document concludes that it is likely that the most appropriate lag period for a study will vary depending on the health outcome and the specific pollutant under study. For example, for a health outcome such as a delayed asthma response, the lag period of a day or several days might be expected between exposure and outcome; however, some cardiovascular responses might be expected to occur within a very short time period (e.g., an hour) after exposure (EPA, 2004, p. 8-279). As shown in Figures 8-24 to 8-28, the Criteria Document notes a pattern of stronger associations between PM
10
and mortality or cardiovascular hospitalization with shorter lag periods (e.g., same-day or 1-day lagged PM
10
). For other effects, however, such as respiratory symptoms, asthma emergency department visits or hospitalization, stronger effects were reported with PM concentrations averaged over several days (EPA, 2004, pp. 8-273 to 8-279). Thus, the Criteria Document concludes that one would expect to see different best-fitting lags for different health effects, based on potentially different biological mechanisms as well as individual variability in responses (EPA, 2004, p. 8-342). For some health outcomes, it is reasonable to expect associations to be observed with PM exposures on the same day or with very short lag periods, but not longer lag periods. In other cases, multi-day average exposure periods or distributed lag models would more appropriately estimate potential PM-related health risks.
(7) Looking more broadly to integrate epidemiologic evidence with that from exposure-related, dosimetric and toxicologic studies, EPA has considered the coherence of the evidence and the extent to which the new evidence provides insights into mechanisms by which PM, especially fine particles, may be affecting human health. Progress made in gaining insights into potential mechanisms lends support to the biologic plausibility of results observed in epidemiologic studies. For cardiovascular effects, the convergence of important new epidemiologic and toxicologic evidence (especially from studies using concentrated ambient particles) builds support for the plausibility of causal associations, especially between fine particles and physiological endpoints indicative of increased risk of cardiovascular disease and changes in cardiac rhythm. This finding is supported by new cardiovascular effects research focused on fine particles that has notably advanced our understanding of potential mechanisms by which PM
2.5
exposure, especially in susceptible individuals, could result in changes in cardiac function or blood parameters that are risk factors for cardiovascular disease. For respiratory effects, toxicologic studies have provided evidence that supports plausible biologic pathways for fine particles, including inflammatory responses, increased airway responsiveness, or altered responses to infectious agents. Further, coherence across a broad range of cardiovascular and respiratory health outcomes is supported by evidence from epidemiologic and toxicologic studies done in the same location, for example, in the series of studies conducted in or evaluating ambient PM from Boston and the Utah Valley (EPA, 2004, 7-42 to 43, 7-46 to 47, and 9-45). Toxicologic studies have suggested that some combustion-related particles, including particles from wood burning and diesel engine exhaust, but not others such as coal fly ash, may have carcinogenic effects (EPA, 2004, Section 7.8.4). This evidence supports the plausibility of the observed relationship between fine particles and lung cancer mortality. Evidence for PM-related infant mortality and developmental effects poses an emerging concern, but the current information is still very limited in support of the plausibility of potential ambient PM relationships. More generally, toxicologic animal studies often test effects of exposures to individual chemical components, and thus the physical and chemical characteristics may differ from those of particles in ambient air to which humans are exposed. These and other differences in toxicologic and epidemiologic study designs complicate the assessment of coherence in results from across disciplines (EPA, 2004, section 9.2.3.1; Schlesinger and Cassee, 2003).
Overall, the Criteria Document finds that much more evidence is now available related to the coherence and plausibility of effects than in the last review. For short-term exposures, integration of evidence from epidemiologic and toxicologic studies indicates both coherence and plausibility of effects on the cardiovascular and respiratory systems, especially for fine particles (EPA, 2004, p. 9-79). There is evidence supporting coherence and plausibility for the observed associations between long-term exposures to fine particles and lung cancer mortality (EPA, 2004, p. 9-78).
(8) In summary, as discussed in the Staff Paper (section 3.5) and the Criteria Document (section 9.2.2), the extensive body of epidemiologic evidence now available continues to support likely causal associations between PM
2.5
and a broad range of mortality and morbidity health outcomes based on an assessment of the strength of the evidence, including the strength and robustness of reported associations and the consistency of the results. While the limitations and uncertainties in the available evidence suggest caution in interpreting the epidemiologic studies at the lower levels of air quality observed in the studies, the evidence now available provides strong support that both short-term and long-term exposures to fine particles are plausibly associated with a broad range of effects on the respiratory and cardiovascular systems. The Criteria Document concludes: “the epidemiological evidence continues to support likely causal associations between PM
2.5
and PM
10
and both mortality and morbidity from cardiovascular and respiratory diseases, based on an assessment of strength, robustness, and consistency in results.” (EPA, 2004, p. 9-48). In its integrative assessment, the Criteria Document finds that health evidence from various disciplines provides a strong and coherent basis for concluding that both short-term and long-term exposure to fine particles is associated with health effects ranging from subtle changes in lung function to premature mortality.
4. Sensitive Subgroups for PM
2.5
-Related Effects
As described in the PM Criteria Document, 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 subgroups have been identified as potentially susceptible to health effects as a result of PM exposure, including people with existing heart and lung diseases, including diabetes, and older adults and children. In addition, new attention has been paid to the concept of some population groups having increased vulnerability to pollution-related effects due to factors such as socioeconomic status or factors that result in particularly elevated exposure levels, such as residence near sources such as roadways (EPA, 2004, p. 9-81).
A good deal of evidence indicates that people with existing heart or lung diseases are more susceptible to PM-related effects. In addition, new studies have suggested that people with diabetes, who are at risk for cardiovascular disease, may have
increased susceptibility to PM exposures. As discussed in Section 9.2.4.1 of the Criteria Document, this body of evidence includes findings from epidemiologic studies that associations with mortality or morbidity are greater in those with preexisting conditions, as well as evidence from toxicologic studies using animal models of cardiopulmonary disease. In addition, dosimetric evidence indicates that deposition of particles is increased, and can be focused in “hot spots” in the respiratory tract, in people with chronic respiratory diseases.
Two age groups, older adults and the very young, are also potentially at greater risk for PM-related effects. Epidemiologic studies have generally not shown striking differences between adult age groups. However, some epidemiologic studies have suggested that serious health effects, such as premature mortality, are greater among older populations (EPA, 2005a, p. 8-328). In addition, preexisting respiratory or cardiovascular conditions are more prevalent in older adults than younger age groups; thus there is some overlap between potentially susceptible groups of older adults and people with heart or lung diseases.
Epidemiologic evidence has reported associations with emergency hospital admissions for respiratory illness and asthma-related symptoms in children. Several factors may make children susceptible to PM-related effects, including the greater ventilation rate per kilogram body weight in children, greater prevalence of chronic asthma, and the fact that children are more likely to be active outdoors and thus have greater exposures. In addition, there is a more limited body of new evidence from epidemiologic studies for potential PM-related health effects in infants, using various PM indicators. Results from this body of evidence, though mixed, are suggestive of possible effects; more research is needed to further elucidate the potential risks of PM exposure for these health outcomes (EPA, 2004, p. 8-222).
In summary, there are several population groups that may be especially susceptible or vulnerable to PM-related effects. These groups include those with preexisting heart and lung diseases, older adults and children. Emerging evidence indicates that people from lower socioeconomic strata or who have particularly elevated exposures may be more vulnerable to PM-related effects.
5. PM
2.5
-Related Impacts on Public Health
As just discussed, there are several population groups that may be especially susceptible or vulnerable to effects from exposure to PM. These population subgroups, such as young children or older adults, and people with pre-existing heart or lung diseases, constitute a large portion of the U.S. population. For example, approximately 22 million people, or 11 percent of the U.S. population, have received a diagnosis of heart disease, about 20 percent of the population has hypertension and about 9 percent of adults and 11 percent of children in the U.S. have been diagnosed with asthma. In addition, about 26 percent of the U.S. population is under 18 years of age,
19
and about 12 percent is 65 years of age or older (EPA, 2004, Table 9-4). EPA recognizes that combining fairly small risk estimates and small changes in PM concentrations with large groups of the U.S. population would result in large public health impacts.
19
Health studies that have suggested that children are susceptible to PM-related effects include varying age ranges, for example, for hospital admissions in children up to 18 years of age, or respiratory symptoms in panels of 4th and 5th grade children.
One issue that is important for interpreting the public health implications of the associations reported between mortality and short-term exposure to PM is whether mortality is occurring only in very frail individuals (sometimes referred to as “harvesting”), resulting in loss of just a few days of life expectancy. A number of new analyses assess the likelihood of such “harvesting” occurring in the short-term exposure studies. Overall, the Criteria Document concludes from the time-series studies that there appears to be no strong evidence to suggest that short-term exposure to PM is only shortening life by a few days (EPA, 2004, Section 8.4.10). In addition to the evidence from short-term exposure studies discussed above, one new report used the mortality risk estimates from the ACS prospective cohort study to estimate potential loss of life expectancy from PM-related mortality in a population. The authors estimated that the loss of population life expectancy associated with long-term exposure to PM
2.5
was on the order of a year or so (EPA, 2004, p. 8-334). The Criteria Document recognizes that these calculations were based on studies in adult populations, and potential population life shortening would be increased if the new, albeit limited, evidence from infant mortality studies was considered (EPA, 2004, p. 8-335). The Criteria Document also observes that the risk estimates reported for long-term fine particle exposures and lung cancer mortality are in about the same range as the risk seen for a nonsmoker living with a smoker (EPA, 2004, p. 9-94).
Large subgroups of the U.S. population are included in subpopulations considered to be potentially sensitive to effects related to fine particle exposures (EPA, 2004, section 9.2.5.1). While individual epidemiologic effect estimates may be small in size, the public health impact of the mortality and morbidity associations can be quite large. In addition, it appears that mortality risks are not limited to the very frail. Taken together, these results suggest that exposure to ambient PM, especially PM
2.5
, can have substantial public health impacts (EPA, 2004, p. 9-93).
B. Quantitative Risk Assessment
This section discusses the approach used to develop quantitative risk estimates associated with exposures to PM
2.5
building upon a more limited risk assessment that was conducted during the last review.
20
At that time, EPA conducted a very limited risk assessment covering a portion of two cities (i.e., Philadelphia County and Southeast Los Angeles County) for which ambient PM
2.5
data were available. For short-term exposure mortality and morbidity health effects, the prior assessment relied on either pooled analyses that combined the results from several studies of individual cities or individual single- and multi-city studies, none of which included the two urban counties for which risks were estimated, to estimate concentration-response relationships for these two cities. EPA recognized that the lack of city-specific relative risks introduced substantial uncertainties in the risk estimates due to inherent differences (e.g., different population characteristics, PM size distributions) that might influence the concentration-response relationships. For long-term exposure mortality, the prior assessment relied on the concentration-response relationship reported in the original ACS study (Pope et al., 1995). Additional important uncertainties noted at the time of that assessment with respect to all health effects included: (1) The absence of clear evidence regarding mechanisms of
action for the various effects of interest, (2) uncertainties about the shape of the concentration-response relationships; and (3) concern about whether the use of ambient PM
2.5
fixed-site monitoring data adequately reflected the relevant population exposures to PM that are responsible for the reported health effects (61 FR 65650).
20
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, 1996b) and in several technical reports (Abt Associates, 1996; Abt Associates, 1997a,b) and publications (Post et al., 2000; Deck et al., 2001).
In light of the substantial uncertainties in the prior risk estimates, EPA placed greater weight on the overall conclusions derived from the health effect studies—that ambient PM was likely causing or contributing to significant adverse effects at levels below those permitted by the then-existing PM
10
standards—than on the specific concentration-response functions and quantitative risk estimates derived from them. Nevertheless, EPA judged that the assessment provided reasonable estimates as to the possible extent of risk for those effects given the available information (62 FR at 38656).
1. Overview
The updated risk assessment conducted as part of this review includes estimates of (1) risks of mortality, morbidity, and symptoms associated with recent ambient PM
2.5
levels; (2) risk reductions and remaining risks associated with just meeting the current suite of PM
2.5
NAAQS; and (3) risk reductions and remaining risks associated with just meeting various alternative PM
2.5
standards in a number of example urban areas. This risk assessment is more fully described and presented in the Staff Paper (EPA, 2005a, Chapter 4) and in a technical support document,
Particulate Matter Health Risk Assessment for Selected Urban Areas
(Abt Associates, 2005a). The scope and methodology for this risk assessment were developed over the last few years with considerable input from the CASAC PM Panel and the public.
21
The information presented in these documents included specific criteria for the selection of health endpoints and studies to include in the assessment. It also addressed which alternative statistical models (e.g., for control of time-varying factors such as weather and for various lags) to include in the assessment, recognizing that some of the health studies presented results from a large number of alternative models. In an advisory letter sent by CASAC to the Administrator documenting its advice in May 2002 (Hopke, 2002), CASAC concluded that the general methodology and framework to be used in the assessment were appropriate.
21
In June 2001, OAQPS released a draft document, PM NAAQS Risk Analysis Scoping Plan (EPA, 2001), for CASAC consultation and public comment, which described staff's general plan for this assessment. In January 2002, OAQPS released a more detailed draft document, Proposed Methodology for Particulate Matter Risk Analyses for Selected Urban Areas (Abt Associates, 2002), for CASAC review and public comment, which described staff's plans to assess (a) PM
2.5
-related risks for several health endpoints, including mortality, hospital admissions, and respiratory symptoms and (b) PM
10-2.5
-related risks for hospital admissions and respiratory symptoms (as discussed below in Section III.B).
The goals of the PM
2.5
risk assessment were: (1) To provide estimates of the potential magnitude of mortality and morbidity effects associated with current PM
2.5
levels, and with meeting the current suite of PM
2.5
NAAQS and alternative PM
2.5
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 suites of PM
2.5
standards. EPA recognizes that there are many sources of uncertainty and variability inherent in the inputs to this assessment and that there is a high degree of uncertainty in the resulting PM
2.5
risk estimates. While some of these uncertainties have been addressed quantitatively in the form of estimated confidence ranges around central risk estimates, other uncertainties and the variability in key inputs are not reflected in these confidence ranges, but rather have been addressed through separate sensitivity analyses or characterized qualitatively.
2. Scope and Key Components
The risk assessment estimates risks of various health effects associated with exposure to ambient PM
2.5
in nine urban areas selected to illustrate the public health impacts associated with a recent year of air quality and potential reductions in risk associated with just meeting the current suite of PM
2.5
standards and alternative suites of standards. The selection of urban areas was largely determined by identifying areas in the U.S. for which acceptable epidemiological studies were available that estimated concentration-response relationships for PM
2.5
, which were then used in assessing the risks. Thus, unlike the prior risk assessment, the current risk assessment for short-term exposure mortality and morbidity health effects used concentration-response relationships reported in studies that included the urban areas for which risks were estimated. Based on a review of the evidence evaluated in the Criteria Document and Staff Paper, as well as the criteria discussed in Chapter 4 of the Staff Paper, the following broad categories of health endpoints were included in the risk assessment for PM
2.5
associated with short-term exposure: Total (non-accidental), cardiovascular, and respiratory mortality; hospital admissions for cardiovascular and respiratory causes; and respiratory symptoms not requiring hospitalization. Also included in the PM
2.5
risk assessment were total, cardiopulmonary, and lung cancer mortality associated with long-term exposure.
The available long-term exposure mortality concentration-response functions are all based on cohort studies, in which a cohort of individuals is followed over time. Based on the evaluation contained in the Criteria Document and EPA's assessment of the complete data base addressing mortality associated with long-term exposure to PM
2.5
, studies based on the following two cohorts were identified as being particularly relevant for the PM
2.5
risk assessment: (1) The Six Cities study cohort (referred to as Krewski et al. (2000)—Six Cities) and (2) the ACS cohort (referred to Krewski et al. (2000)—ACS), which includes a much larger number of individuals from many more cities. In addition, Pope et al. (2002) extended the follow-up period for the ACS cohort to sixteen years and published findings on the relation of long-term exposure to PM
2.5
and all-cause mortality as well as cardiopulmonary and lung cancer mortality (referred to as Pope et al. (2002)—ACS extended).
22
22
The use of these particular cohort studies to estimate health risks associated with long-term exposure to PM
2.5
is consistent with the views expressed in the National Academy of Sciences (2002) report, “Estimating the Public Health Benefits of Proposed Air Pollution Regulations,” and the Science Advisory Board Clean Air Act Compliance Council review of the proposed methodology to estimate the health benefits associated with the Clean Air Act (SAB, 2004).
The available short-term exposure morbidity and mortality concentration-response functions used in the risk assessment are all from time series studies. The risk assessment included only those health endpoints for which the the Criteria Document concluded that there is likely to be a causal relationship with short-term exposure to PM
2.5
based on the overall weight of the evidence from the collective body of available studies. Also, given the large number of endpoints and studies addressing PM
2.5
-related effects, the assessment only included the more severe and better understood (in terms of health consequences) health effects. As noted above, in contrast to the prior risk assessment, the concentration-response functions used in this assessment for each urban area are
based on results of studies for that specific area or from a multi-city study that included that specific area.
The concentration-response relationships used in the assessment were based on findings from human epidemiological studies that have relied on fixed-site, population-oriented, ambient monitors as a surrogate for actual ambient PM
2.5
exposures. The risk assessment addresses risks attributable to anthropogenic sources and activities (i.e., risk associated with concentrations above policy-relevant background
23
or above various selected higher cutpoints intended as surrogates for alternative assumed population thresholds). This approach of estimating risks in excess of background was judged to be more relevant to policy decisions regarding ambient air quality standards than risk estimates that include effects potentially attributable to uncontrollable background PM concentrations. For the base case analyses, an estimate of the annual average background level was used, rather than a maximum 24-hour value, since estimated risks were aggregated for each day throughout the year.
23
Background PM concentrations used in the PM risk assessment were defined in Chapter 2 of the Staff Paper as the PM concentrations that would be observed in the U.S. in the absence of anthropogenic emissions of PM and its precursors in the U.S., Canada, and Mexico. For the initial base case risk estimates, the midpoints of the appropriate ranges of annual average estimates for PM
2.5
background presented in the Staff Paper were used (i.e., eastern values were used for eastern study locations and western values were used for western study locations). Estimated policy-relevant background concentrations are 3.5 μg/m
3
in eastern cities, and 2.5 μg/m
3
in western cities.
In order to estimate the incidence of a particular health effect associated with recent conditions in a specific county or set of counties attributable to ambient PM
2.5
exposures in excess of background or various alternative cutpoints, as well as the change in incidence corresponding to a given change in PM
2.5
levels resulting from just meeting a specified set of alternative PM
2.5
standards, three elements are required. These elements are: (1) Air quality information (including recent air quality data for PM
2.5
from ambient monitors for the selected location, estimates of background PM
2.5
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 set of PM
2.5
standards); (2) relative risk-based concentration-response functions that provide an estimate of the relationship between the health endpoints of interest and ambient PM concentrations; and (3) annual or seasonal baseline health effects incidence rates and population data, which are needed to provide an estimate of the annual or seasonal baseline incidence of health effects in an area before any changes in PM air quality.
The risk assessment for PM
2.5
included a series of base case analyses that characterized the uncertainty associated with the form of the concentration-response relationship drawn from the studies used in the assessment—this uncertainty had by far the greatest impact on estimated risks. Other uncertainties addressed in various sensitivity analyses (e.g., the use of single-versus multi-pollutant models, single-versus multi-city models, use of a distributed lag model, alternative assumptions about the relevant air quality for long-term exposure mortality, and alternative constant or varying background levels) all have a more moderate and often variable impact on the risk estimates in some or all of the cities.
In estimating health risks remaining upon just meeting the current and alternative PM
2.5
standards, the assessment includes a series of base cases, while noting that the confidence ranges in the estimates do not reflect all the identified uncertainties. As discussed above in section II.A.3, additional uncertainty for short-term exposure mortality is related to the use of alternative statistical models and methods to control for time-varying effects, such as weather or season, and to address alternative lag structures. To provide a consistent basis for comparison across studies and locations, the risk assessment used concentration-response functions based on the most common type of analysis (“generalized additive methods”) and on lag structures judged to be most appropriate for each specific health endpoint, as discussed in the Staff Paper (EPA, 2005a, p. 4-24). The risk assessment included a sensitivity analysis for one location where a wide array of statistical models and lags was reported in the health study for that location (Los Angeles, as reported in Moolgavkar, 2003). EPA recognizes that there is additional uncertainty associated with choices about appropriate modeling strategy (EPA, 2004, 8.4.2) and that this uncertainty is not included in the confidence ranges presented for the risk estimates.
As noted earlier, EPA recognizes that while there are likely biological thresholds in individuals for specific health endpoints, the available epidemiologic studies do not support or refute the existence of thresholds at the population level for either long-term or short-term PM
2.5
exposures within the range of air quality observed in the studies (EPA, 2004, 9.2.2.5). Thus, base case risks were estimated using not only the linear or log-linear concentration-response functions reported in the studies, but also using a series of modified linear functions, as discussed below, as surrogates for assumed non-linear functions that would reflect the possibility that thresholds may exist in the reported associations within the range of air quality observed in the studies.
For short-term exposure mortality and morbidity outcomes associated with PM
2.5
, the initial base case includes linear or log-linear concentration-response models reported in the epidemiology studies which are applied down to the estimated policy-relevant background concentration level. Generally, the lowest measured concentrations in the short-term exposure studies were relatively near or below the estimated policy-relevant background levels such that little or no extrapolation was required beyond the range of data in the studies. In the case of the long-term exposure mortality studies for PM
2.5
that have been included in the risk assessment, the lowest measured levels were in the range 7.5 to 11 μg/m
3
. For the initial base case scenario for this endpoint, the reported linear models were applied down to 7.5 μg/m
3
, which is the lowest measured level reported in the long-term studies. Going down to an estimated policy-relevant background level for short-term exposure studies and to 7.5 μg/m
3
for long-term studies provides a consistent framework which facilitates comparison of risk estimates across urban locations within each group of studies and avoids significant extrapolation beyond the range of concentrations included in these studies.
Additional base case scenarios for both short- and long-term exposure health endpoints involved the use of alternative concentration-response functions that incorporated a modified linear slope with an imposed cutpoint (i.e., an assumed threshold). For mortality associated with short-term exposure, the base case analyses included risk estimates associated with cutpoints of 10, 15, and 20 μg/m. For mortality associated with long-term PM
2.5
exposure, cutpoints of 10 and 12 μg/m
3
were included. For the base case scenarios involving alternative cutpoints, the approach used to develop alternative functions incorporates a modified linear slope with an imposed cutpoint (i.e., an assumed population threshold) that is intended to reflect a
hypothetical inflection point in a typical non-linear, “hockey-stick” shaped function, below which there is little or no population response. More specifically, the slope of the concentration-response relationship has been adjusted assuming that the upward-sloping portion of the “hockey stick” would be the slope estimated in the original epidemiologic study adjusted by the inverse of the proportion of the range of PM levels observed in the study that was above the cutpoint. The Staff Paper concludes that this simple slope adjustment approach represents a reasonable approach to illustrating the potential impact of possible non-linear concentration-response relationships. In its review of the Staff Paper and risk assessment, the CASAC PM Panel commented that for the purpose of estimating public health impacts, it “favored the primary use of an assumed threshold of 10 μg/m
3
” and that “a major research need is for more work to determine the existence and level of any thresholds that may exist or the shape of nonlinear concentration-response curves at low levels of exposure that may exist” (Henderson, 2005a).
3. Risk Estimates and Key Observations
In focusing on the five study areas that do not meet the current PM
2.5
standards based on 2001-2003 air quality data (Detroit, Los Angeles, Philadelphia, Pittsburgh, and St. Louis), the total mortality risk estimates associated with simulating air quality reductions to just meet the current PM
2.5
standards (based on associations with long-term PM
2.5
exposure, and using the lowest cutpoint of 7.5 μg/m
3
) range from several hundred to over 1500 deaths per year, which translate into an incidence rate of approximately 16 to 35 deaths per year per hundred thousand population.
24
These estimated risks associated with long-term exposure represent approximately 2.6 to 3.2 percent of total mortality in those areas. Estimated risks associated with long-term exposure based on an assumed cutpoint of 10 μg/m
3
are roughly half as large as the estimates based on a cutpoint of 7.5 μg/m
3
. In the same five areas, the estimates of mortality risk associated with short-term PM
2.5
exposure, based on a cutpoint equal to policy-relevant background or 10 μg/m, range from less than 20 percent to over 50 percent of the estimates associated with long-term exposure.
25
24
The full range of quantitative risk estimates associated with just meeting the current PM
2.5
standards are presented in Tables 4-9, 4-10, 4-12, and 4-13 in Chapter 4 of the Staff Paper.
25
In some areas, the 95 percent confidence ranges associated with the risk estimates for short-term exposure (but not long-term exposure) extend to below zero, reflecting appreciably more uncertainty in estimates based on positive but not statistically significant associations.
Reductions in risk associated with simulating air quality to just meet a range of lower alternative annual and 24-hour PM
2.5
standards were also estimated in this assessment. The estimated risk reductions are depicted graphically in the Staff Paper (EPA, 2005a, Figures 5-1 and 5-2 and Figures 5A-1 and 5A-2), showing patterns of estimated risk reductions associated with alternative suites of standards for all the various assumed cutpoints. As would be expected, patterns of increasing estimated risk reductions are observed as either the annual or 24-hour standard, or both, are reduced over the range considered in this assessment, and the estimated percentage reductions in risk are strongly influenced by the assumed cutpoint level.
The discussion below highlights additional observations and insights from this PM
2.5
risk assessment, together with important caveats and limitations.
(1) With respect to short-term exposure mortality and morbidity, this risk assessment provides the basis for greater confidence in the results as compared to the prior assessment, given that studies are now available using PM
2.5
as the indicator in a much greater number of locations, and the assessment is able to use city-specific functions that are matched to the locations for which risks are estimated. This contrasts with the use of pooled concentration-response functions in the prior assessment which did not include studies for the specific cities included in that assessment. However, EPA recognizes that the confidence ranges, which only reflect uncertainty associated with the precision of the study (related to the population size and duration of the study), may be larger for the current risk estimates due to the use of concentration-response functions from smaller, city-specific studies now versus the use of concentration-response functions from pooled sets of studies that have greater statistical precision. Comparing the risk estimates for the only two specific locations that were included in both the prior and current assessments, the magnitude of the estimates associated with just meeting the current annual standard, in terms of percentage of total incidence, is similar in one of the locations (Philadelphia) and the current estimate is lower in the other location (Los Angeles).
(2) With respect to long-term exposure mortality risk estimates, the prior risk assessment focused on the estimates based on the original ACS study (Pope et al., 1995). Since that time additional cohort analyses have been published and evaluated in the Criteria Document. EPA has greater confidence in the current risk estimates for long-term exposure mortality, given the extensive review of these studies and the extension of the ACS study to additional years of data, as well as improvements in the statistical approach. However, ACS-based risk estimates remain sensitive to plausible changes in statistical model specifications. The choice of studies and concentration-response functions to use for the base case risk estimates is discussed in the Staff Paper (EPA, 2005a, p. 4-25) and risk assessment report (Abt Associates, 2005, pp.49-50) and is consistent with the advice provided by both the National Academy of Sciences and the Science Advisory board Clean Air Act Compliance Council (see footnote 22). At the same time, EPA recognizes that alternative statistical models were examined in the reanalysis of the ACS and Six-Cities studies, and that the uncertainty associated with model selection (such as multipollutant models and different effect estimates associated with different educational levels) is not reflected in the confidence ranges presented in this assessment. Thus, for long-term exposure mortality risk estimates there are additional unquantified uncertainties associated with a lack of understanding as to which statistical model best represents the actual concentration-response function. The relative risk estimates used in the current risk assessment from the ACS extended study are only slightly smaller (1.06 with 95 percent confidence interval of 1.02-1.11) compared to the original ACS study (1.07 with 95 percent confidence interval 1.04-1.10) used in the prior assessment. In terms of the magnitude of the risk estimates, the estimates in terms of percentage of total incidence are very similar for the two specific locations included in both the prior and current assessments.
(3) A fairly wide range of risk estimates are observed for PM
2.5
-related morbidity and mortality risk associated with recent air quality across the urban areas analyzed. The impact of adding additional co-pollutants to the models was variable; sometimes there was relatively little difference, while in other cases there were larger differences. The wide variability in risk estimates associated with a recent year of air
quality is to be expected given the wide range of PM
2.5
levels across the urban areas analyzed and the variation observed in the concentration-response relationships obtained from the original epidemiologic studies. Among other factors, this variability may reflect differences in the mixture of components or sources of fine particles, populations, exposure considerations (e.g., degree of air conditioning use), differences in co-pollutants and/or other stressors, differences in study design, and differences related to exposure and monitor measurement error.
(4) The single most important factor influencing the quantitative estimates of risk is which of the alternative concentration-response functions included in this assessment are considered to best represent the unknown “true” concentration-response relationships. In comparison, the following uncertainties have only a moderate impact on the risk estimates in some or all of the cities: choice of an alternative estimated constant background level, use of a distributed lag model, and alternative assumptions about the relevant air quality for estimating exposure levels for long-term exposure mortality. Use of a distribution of daily background concentrations had very little impact on the risk estimates.
The overall pattern of risk associated with short-term PM
2.5
exposures across the distribution of PM
2.5
air quality, as typically observed in urban areas, is similar to that observed in the last review. That is, on an annual basis, the very highest days (which pose the greatest risk in terms of deaths per day) contribute less to the total annual health risk associated with short-term exposures than the middle of the distribution, due to the much greater number of days that occur in this part of the air quality distribution.
(5) Risk estimates associated with just meeting the current suite of PM
2.5
standards in five urban areas that do not meet the current PM
2.5
standards showed a wide range of PM
2.5
-related risk estimates for short-term exposure mortality and morbidity. This is likely due, in large part, to differences in concentration-response relationships among single-location short-term exposure studies, differences in baseline incidence rates, and varying population sizes. Results of a sensitivity analysis which applied one multi-city concentration-response function to all five urban areas analyzed narrowed considerably the range of risk estimates when a risk metric was used that normalized for different population sizes. However, it is still unknown whether the wider range of estimates observed using single-city concentration-response functions reflect methodological differences between studies and/or real city-to-city differences related to exposure, population, composition of the particles, or other factors.
(6) For the risk estimates associated with just meeting the current suite of PM
2.5
standards and alternative suites of standards, the single most important factor influencing the short- and long-term exposure mortality and morbidity estimates is again which of the alternative concentration-response functions included in this assessment are considered to best represent the unknown “true” concentration-response relationships. Several additional sources of uncertainty are introduced into this portion of the risk assessment, including: (1) Uncertainty in the degree to which the pattern of air quality concentration reductions estimated for the risk assessment cities represents the distribution of actual PM concentration changes that would be observed in a given area (“rollback uncertainty”) and (2) uncertainty concerning the degree to which current PM risk coefficients may reflect contributions from other pollutants, or uncertainty concerning whether all of the constituents of PM
2.5
would be reduced in similar proportion to the reduction in PM
2.5
as a whole, and, if not, what impact this would have on estimated reductions in risk. For areas where the current annual standard is the controlling standard, one alternative approach to rolling back the distribution of daily PM
2.5
concentrations, in which the upper end of the distributions of concentrations was reduced by a greater amount than the rest of the distribution, had little impact on the risk estimates. This approach or alternative approaches to rolling back the distribution of daily concentrations may have a greater impact on the risk estimates in areas where the daily standard is the controlling standard.
(7) For the risk estimates associated with just meeting the current or alternative suites of PM
2.5
standards, there is a significant decrease in the mortality risk estimates based on short-term PM
2.5
exposure remaining as one considers alternative higher cutpoints. There also is a significant increase observed in the percent reduction in estimated risk upon just meeting alternative standards with higher alternative cutpoints. These findings are even more pronounced for the mortality risk estimates associated with long-term PM
2.5
exposure as higher alternative cutpoint levels are considered.
C. Need for Revision of the Current Primary PM
2.5
Standards
The initial issue to be addressed in the current review of the primary PM
2.5
standards is whether, in view of the advances in scientific knowledge reflected in the Criteria Document and Staff Paper, the existing standards should be revised. Based on the information and conclusions presented in the Criteria Document, summarized above in section II.A., the Staff Paper concludes that the newly available information generally reinforces the associations between PM
2.5
and mortality and morbidity effects observed in the last review. While important uncertainties and research questions remain, much progress has been made in reducing some key uncertainties since the last review. The examination of specific components, properties, and sources of fine particles that are linked with health effects remains an important research need. Other important research needs include better characterizing the shap
This text is long and has been trimmed here. Open the source document for the complete record.
This is a copy of a public record, reproduced as it was published. It is not legal advice, and it may not be the version a court would rely on. Check the official source before you cite it.