National Ambient Air Quality Standards for Particulate Matter

110 Pages·2012·1.62 MB·English

r
Friday
July 18, 1997
e
t
s
i
g
e
r
Part II

l
Environmental
a
Protection Agency
40 CFR Part 50
r
National Ambient Air Quality Standards
for Particulate Matter; Final Rule
e
d
e
f
38651
38652 Federal Register / Vol. 62, No. 138 / Friday, July 18, 1997 / Rules and Regulations
ENVIRONMENTAL PROTECTION of the PM primary and secondary inspection and copying in the EPA
AGENCY standards (Docket No. A–95–54) is docket at the address under
available for public inspection in the ‘‘ADDRESSES,’’ at the beginning of this
40 CFR Part 50 Central Docket Section of the U.S. document.
Environmental Protection Agency,
[AD–FRL–5725–2] Electronic Availability
South Conference Center, Rm. 4, 401 M
RIN 2060–AE66 St., SW., Washington, DC. This docket The Staff Paper and human health
incorporates the docket established for risk assessment support documents are
National Ambient Air Quality
the air quality Criteria Document available on the Agency’s Office of Air
Standards for Particulate Matter
(Docket No. ECAO–CD–92–0671). The Quality Planning and Standards’
AGENCY: Environmental Protection docket may be inspected between 8 a.m. (OAQPS) Technology Transfer Network
Agency (EPA). and 3 p.m., Monday through Friday, (TTN) Bulletin Board System (BBS) in
ACTION: Final rule. except legal holidays, and a reasonable the Clean Air Act Amendments area,
SUMMARY: This document describes fee may be charged for copying. The under Title I, Policy/Guidance
EPA’s decision to revise the national information in the docket constitutes Documents. To access the bulletin
ambient air quality standards (NAAQS) the complete basis for the decision board, a modem and communications
for particulate matter (PM) based on its announced in this document. For the software are necessary. To dial up, set
review of the available scientific availability of related information, see your communications software to 8 data
evidence linking exposures to ambient ‘‘SUPPLEMENTARY INFORMATION.’’ bits, no parity and one stop bit. Dial
PM to adverse health and welfare effects FORFURTHERINFORMATIONCONTACT: John (919) 541–5742 and follow the on-
at levels allowed by the current PM H. Haines, MD–15, Air Quality screen instructions to register for access.
standards. The current primary PM Strategies and Standards Division, After registering, proceed to choice
standards are revised in several Office of Air Quality Planning and ‘‘<T> Gateway to TTN Technical
respects: Two new PM2.5standards are Standards, U.S. Environmental Areas’’, then choose ‘‘<E> CAAA BBS’’.
added, set at 15 m g/m3, based on the 3- Protection Agency, Research Triangle From the main menu, choose ‘‘<1> Title
year average of annual arithmetic mean Park, NC 27711; telephone: (919) 541– I: Attain/Maint of NAAQS’’, then ‘‘<P>
PM2.5concentrations from single or 5533; e-mail: Policy Guidance Documents.’’ To access
multiple community-oriented monitors, [email protected]. these documents through the World
and 65 m g/m 3, based on the 3-year SUPPLEMENTARYINFORMATION: Wide Web, click on ‘‘TTN BBSWeb’’,
average of the 98thpercentile of 24-hour then proceed to the Gateway to TTN
Related Final Rules on PM Monitoring
PM2.5concentrations at each Technical areas, as above. If assistance
population-oriented monitor within an In a separate document published is needed in accessing the system, call
area; and the current 24-hour PM10 elsewhere in this issue of the Federal the help desk at (919) 541–5384 in
standard is revised to be based on the Register, EPA is amending its ambient Research Triangle Park, NC.
99thpercentile of 24-hour PM air quality surveillance requirements (40
10
Implementation Strategy For Revised
concentrations at each monitor within CFR part 58) and its ambient air
Air Quality Standards
an area. The new suite of primary monitoring reference and equivalent
standards will provide increased methods (40 CFR part 53) for PM. On Wednesday, July 16, 1997,
protection against a wide range of PM- President Clinton signed a
Availability of Related Information
related health effects, including memorandum to the Administrator
premature mortality and increased Certain documents are available from specifying his goals for the
hospital admissions and emergency the U.S. Department of Commerce, implementation of the O and PM
3
room visits, primarily in the elderly and National Technical Information Service, standards. Attached to the President’s
individuals with cardiopulmonary 5285 Port Royal Road, Springfield, VA memorandum is a strategy prepared by
disease; increased respiratory symptoms 22161. Available documents include: an interagency Administration group
and disease, in children and individuals (1) Air Quality Criteria for Particulate outlining the next steps that would be
with cardiopulmonary disease such as Matter (Criteria Document) (three necessary for implementing these
asthma; decreased lung function, volumes, EPA/600/P–95–001aF thru standards. The EPA will prepare
particularly in children and individuals EPA/600/P–95–001cF, April 1996, NTIS guidance and proposed rules consistent
with asthma; and alterations in lung #PB–96–168224, $234.00 paper copy). with the President’s memorandum.
tissue and structure and in respiratory (2) Review of the National Ambient Copies of the Presidential document are
tract defense mechanisms. The current Air Quality Standards for Particulate available in paper copy by contacting
secondary standards are revised by Matter: Policy Assessment of Scientific the U.S. Environmental Protection
making them identical in all respects to and Technical Information (Staff Paper) Agency Library at the address under
the new suite of primary standards. The (EPA–452/R–96–013, July 1996, NTIS ‘‘Availability of Related Information’’
new secondary standards, in #PB–97–115406, $47.00 paper copy and and in electronic form as discussed
conjunction with a regional haze $19.50 microfiche). (Add a $3.00 above in ‘‘Electronic Availability.’’
program, will provide appropriate handling charge per order.) The following topics are discussed in
protection against PM-related public A limited number of copies of other this preamble:
welfare effects including soiling, documents generated in connection I. Background
material damage, and visibility with this standard review, such as A. Legislative Requirements
impairment. In conjunction with the technical support documents pertaining B. Related Control Requirements
new PM standards, a new reference to air quality, monitoring, and health C. Review of Air Quality Criteria and
method 2h.5as been specified for risk assessment, can be obtained from: Standards for PM
D. Summary of Proposed Revisions to the
monitoring PM as PM . Environmental Protection Agency
2.5 PM Standards
EFFECTIVEDATE: This action is effective Library (MD–35), Research Triangle II. Rationale for the Primary PM Standards
September 16, 1997. Park, NC 27711, telephone (919) 541– A. Introduction
ADDRESSES: A docket containing 2777. These and other related B. Need for Revision of the Current
information relating to the EPA’s review documents are also available for Primary PM Standards
Federal Register / Vol. 62, No. 138 / Friday, July 18, 1997 / Rules and Regulations 38653
C. Indicators of PM primary standard as one ‘‘the attainment criteria and standards, as appropriate.
D. Averaging Time of PM2.5Standards and maintenance of which in the The committee established under
E. Form of PM2.5Standards judgment of the Administrator, based on section 109(d)(2) of the Act is known as
F. Levels for the Annual and 24-Hour PM2.5 [the] criteria and allowing an adequate the Clean Air Scientific Advisory
Standards
margin of safety, are requisite to protect Committee (CASAC), a standing
G. SCtoanndclaursdisons Regarding the Current PM10 the public health.’’ The margin of safety committee of EPA’s Science Advisory
H. Final Decisions on Primary PM requirement was intended to address Board.
Standards uncertainties associated with
III. Rationale for the Secondary Standards inconclusive scientific and technical B. Related Control Requirements
A. Need for Revision of the Current information available at the time of
States are primarily responsible for
SecondaryStandards standard setting, as well as to provide a
B. Decision on the Secondary Standards reasonable degree of protection against ensuring attainment and maintenance of
IV. Other Issues ambient air quality standards once EPA
hazards that research has not yet
A. Consideration of Costs has established them. Under section 110
identified. Both kinds of uncertainties
B. Margin of Safety of the Act (42 U.S.C. 7410) and related
are components of the risk associated
C. Data Availability provisions, States are to submit, for EPA
D. 1990 Amendments with pollution at levels below those at approval, State implementation plans
V. Revisions to 40 CFR Part 50, Appendix which human health effects can be said
(SIP’s) that provide for the attainment
K—Interpretation of the PM NAAQS to occur with reasonable scientific
and maintenance of such standards
A. PM2.5Computations and Data Handling certainty. Thus, by selecting primary
Conventions standards that provide an adequate through control programs directed to
B. PM10Computations and Data Handling margin of safety, the Administrator is sources of the pollutants involved. The
Conventions States, in conjunction with EPA, also
seeking not only to prevent pollution
C. Changes that Apply to Both PM2.5and levels that have been demonstrated to be administer the prevention of significant
PM10Computations harmful but also to prevent lower deterioration program (42 U.S.C. 7470–
VI. Reference Methods for the Determination 7479) for these pollutants. In addition,
of Particulate Matter as PM10and PM2.5 pollutant levels that she finds may pose Federal programs provide for
in the Atmosphere an unacceptable risk of harm, even if the
nationwide reductions in emissions of
A. Revisions to 40 CFR Part 50, Appendix risk is not precisely identified as to
these and other air pollutants through
J—Reference Method for PM10 nature or degree. The Act does not
B. 40 CFR Part 50, Appendix L—New require the Administrator to establish a the Federal Motor Vehicle Control
Reference Method for PM2.5 primary NAAQS at a zero-risk level, but Program under Title II of the Act (42
VII. Effective Date of the Revised PM rather at a level that reduces risk U.S.C. 7521–7574), which involves
Standards and Applicability of the controls for automobile, truck, bus,
sufficiently so as to protect public
Existing PM10Standards health with an adequate margin of motorcycle, nonroad engine, and aircraft
VIII. Regulatory and Environmental Impact safety. The selection of any particular emissions; the new source performance
Analyses
A. Executive Order 12866 approach to providing an adequate standards under section 111 of the Act
B. Regulatory Flexibility Analysis margin of safety is a policy choice left (42 U.S.C. 7411); and the national
C. Impact on Reporting Requirements specifically to the Administrator’s emission standards for hazardous air
D. Unfunded Mandates Reform Act judgment. Lead Industries Ass’n v. EPA, pollutants under section 112 of the Act
E. Environmental Justice 647 F.2d 1130, 1161–1162 (D.C. (42 U.S.C. 7412).
F. Submission to Congress and the Cir.1980).
C. Review of Air Quality Criteria and
Comptroller General A secondary standard, as defined in
IX. Response to Petition for Administrator section 109 (b)(2) of the Act, must Standards for PM
Browner’s Recusal
‘‘specify a level of air quality the Particulate matter is the generic term
X. References
attainment and maintenance of which in for a broad class of chemically and
I. Background the judgment of the Administrator, physically diverse substances that exist
A. Legislative Requirements based on [the] criteria, [are] requisite to as discrete particles (liquid droplets or
protect the public welfare from any solids) over a wide range of sizes.
Two sections of the Clean Air Act known or anticipated adverse effects Particles originate from a variety of
(Act) govern the establishment, review, associated with the presence of [the] anthropogenic stationary and mobile
and revision of NAAQS. Section 108 of pollutant in the ambient air.’’ Welfare sources as well as from natural sources.
the Act (42 U.S.C. 7408) directs the effects as defined in section 302(h) of Particles may be emitted directly or
Administrator to identify certain the Act (42 U.S.C. 7602(h)) include, but formed in the atmosphere by
pollutants which ‘‘may reasonably be are not limited to, ‘‘effects on soils, transformations of gaseous emissions
anticipated to endanger public health water, crops, vegetation, manmade such as sulfur oxides (SO ), nitrogen
and welfare’’ and to issue air quality materials, animals, wildlife, weather, oxides (NO ), and volatilex organic
criteria for them. These air quality visibility, and climate, damage to and x
compounds (VOC). The chemical and
criteria are to ‘‘accurately reflect the deterioration of property, and hazards to
physical properties of PM vary greatly
latest scientific knowledge useful in transportation, as well as effects on
with time, region, meteorology, and
indicating the kind and extent of all economic values and on personal
source category, thus complicating the
identifiable effects on public health or comfort and well-being.’’
assessment of health and welfare effects.
welfare which may be expected from the Section 109(d)(1) of the Act requires
presence of [a] pollutant in the ambient periodic review and, if appropriate, The last review of PM air quality
air * * *.’’ revision of existing air quality criteria criteria and standards was completed in
Section 109 of the Act (42 U.S.C. and NAAQS. Section 109(d)(2) of the July 1987 with notice of a final decision
7409) directs the Administrator to Act requires appointment of an to revise the existing standards
propose and promulgate ‘‘primary’’ and independent scientific review published in the Federal Register (52
‘‘secondary’’ NAAQS for pollutants committee to review criteria and FR 24854, July 1, 1987). In that decision,
identified under section 108 of the Act. standards and recommend new EPA changed the indicator for PM from
Section 109(b)(1) of the Act defines a standards or revisions of existing total suspended particles (TSP) to
38654 Federal Register / Vol. 62, No. 138 / Friday, July 18, 1997 / Rules and Regulations
PM .1Identical primary and secondary and May 1996.4The CASAC came to City, UT. A fourth public hearing,
10
PM10standards were set for two closure in its review of the Staff Paper, which focused primarily on PM
averaging times: 50 m g/m3, expected advising the Administrator in a June 13, monitoring issues, was held in Durham,
annual arithmetic mean, averaged over 1996 closure letter (Wolff, 1996b) that NC on January 14, 1997. Over 400
3 years, and 150 m g/m3, 24-hour average, ‘‘the Staff Paper, when revised, will citizens and organizations testified
with no more than one expected provide an adequate summary of our during these public hearings. EPA also
exceedance per year.2 present understanding of the scientific held two national satellite telecasts to
The EPA initiated this current review basis for making regulatory decisions answer questions on the standards and
of the air quality criteria and standards concerning PM standards.’’ CASAC and
participated in meetings sponsored by
for PM in April 1994 by announcing its public comments from these meetings,
the Air and Waste Management
intention to develop a revised Air subsequent written comments, and the
Association on the proposed revisions
Quality Criteria Document for CASAC closure letter were incorporated
to the standards at more than 10
Particulate Matter (henceforth, the as appropriate in the final Staff Paper
locations across the country. Beyond
‘‘Criteria Document’’). Thereafter, the (U.S. EPA, 1996b).
that, several EPA regional offices held
EPA presented its plans for review of On November 27, 1996, EPA
public meetings and workshops and
the criteria and standards for PM under announced its proposed decision to
participated in hearings that States and
a highly accelerated, court-ordered revise the NAAQS for PM (61 FR 65638,
schedule3at a public meeting of the December 13, 1996) (hereafter cities held around the country.
CASAC in December 1994. Several ‘‘proposal’’) as well as its proposed As a result of this intensive effort to
workshops were held by EPA’s National decision to revise the NAAQS for ozone solicit public input, over 50,000 written
Center for Environmental Assessment (O3)(61 FR 65716, December 13, 1996). and oral comments were received on the
(NCEA) to discuss important new health In the proposal, EPA identified proposed revisions to the PM NAAQS
effects information in November 1994 proposed revisions, based on the air by the close of the public comment
and January 1995. External review drafts quality criteria for PM, and solicited period on March 12, 1997. Major issues
of the Criteria Document were made public comments on alternative primary raised in the comments are discussed
available for public comment and were standards and on the proposed forms of throughout the preamble of this final
reviewed by CASAC at public meetings the standards. decision. A comprehensive summary of
To ensure the broadest possible
held in August and December 1995 and all significant comments, along with
public input on the PM and O
February 1996. The CASAC came to 3 EPA’s response to such comments
proposals, EPA took extensive and
closure in its review of the Criteria (hereafter ‘‘Response to Comments’’),
unprecedented steps to facilitate the
Document, advising the Administrator can be found in the docket for this
public comment process beyond the
in a March 15, 1996 closure letter rulemaking (Docket No. A–95–54).
normal process of providing an
(Wolff, 1996a) that ‘‘although our
opportunity to request a hearing and The principal focus of this current
understanding of the health effects of
receiving written comments submitted review of the air quality criteria and
PM is far from complete, a revised
to the rulemaking docket. The EPA standards for PM is on recent
Criteria Document which incorporates
established a national toll-free epidemiological evidence reporting
the Panel’s latest comments will provide
telephone hotline to facilitate public associations between ambient
an adequate review of the available
comments on the proposed revisions to concentrations of PM and a range of
scientific data and relevant studies of
PM.’’ CASAC and public comments the PM and O3NAAQS, and on related serious health effects. Particular
from these meetings, and from notices dealing with the implementation attention has been given to several size-
subsequent written comments and the of revised PM and O3standards, as well specific classes of particles, including
closure letter, were incorporated as as a system for the public to submit PM10and the principal fractions of
appropriate in the final Criteria comments on the proposals PM10, referred to as the fine (PM2.5)5and
Document (U.S. EPA, 1996a). electronically via the Internet. Over coarse (PM10–2.5)6fractions. As
External review drafts of a Staff Paper 14,000 calls and over 4,000 electronic discussed in the Criteria Document, fine
prepared by the Office of Air Quality mail messages were received through and coarse fraction particles can be
Planning and Standards (OAQPS), these channels. The public could also differentiated by their sources and
Review of the National Ambient Air access key supporting documents formation processes and their chemical
(including the Criteria Document, Staff
Quality Standards for Particulate Matter: and physical properties, including
Paper, related technical documents and
Assessment of Scientific and Technical behavior in the atmosphere. Detailed
fact sheets) via the Internet.
Information (henceforth, the ‘‘Staff discussions of atmospheric formation,
The EPA also held several public
Paper’’), were made available for public ambient concentrations, and health and
hearings and meetings across the
comment and were reviewed by CASAC welfare effects of PM, as well as
country to provide direct opportunities
at public meetings in December 1995 quantitative estimates of human health
for public comment on the proposed
risks associated with exposure to PM,
revisions to the PM and O NAAQS and
dia1mPMete10r rleefsesr tsh taon p oarr teiqclueasl wtoi tah n aonm aienraold 1y0namic to disseminate information3 to the public caannd bine tfhoue nSdta ifnf Pthaep eCrr.iteria Document
micrometers. Technical details further specifying about the proposed standard revisions.
the measurement of PM10are contained in 40 CFR On January 14 and 15, 1997, EPA held
part 50, Appendices J and M. concurrent, 2-day public hearings in 5PM2.5refers to particles with an aerodynamic
2A more complete history of the PM NAAQS is diameter less than or equal to a nominal 2.5
Boston, MA, Chicago, IL, and Salt Lake
presented in section II.B of the OAQPS Staff Paper, micrometers, as further specified in 40 CFR part 50,
Review of National Ambient Air Quality Standards Appendix L in this document.
for Particulate Matter: Assessment of Scientific and 4The Staff Paper evaluates policy implications of 6PM10–2.5refers to those particles with an
Technical Information (U.S. EPA, 1996b). the key studies and scientific information in the aerodynamic diameter less than or equal to a
3A court order entered in American Lung Criteria Document, identifies critical elements that nominal 10 micrometers but greater than 2.5
Association v. Browner, CIV–93–643–TUC–ACM (D. EPA staff believes should be considered, and micrometers. In other words, it refers to the
Ariz.,October 6, 1994), as subsequently modified, presents staff conclusions and recommendations of inhalable particles that remain if fine (PM2.5)
requires publication of EPA’s final decision on the suggested options for the Administrator’s particles are removed from a sample of PM10
review of the PM NAAQS by July 19, 1997. consideration. particles.
Federal Register / Vol. 62, No. 138 / Friday, July 18, 1997 / Rules and Regulations 38655
D. Summary of Proposed Revisions to in the ambient air. These decisions also unusually high degree of scrutiny and
the PM Standards take into account: reanalysis over the past several years,
(1) Staff Paper assessments of the beginning with a series of workshops
For reasons discussed in the proposal,
most policy-relevant information in the held early in the review process to
the Administrator proposed to revise the
Criteria Document, upon which staff discuss important new information. A
current primary standards for PM (as
recommendations for new and revised number of opportunities were provided
indicated by PM ), by adding two new
10 primary standards are based. for public comment on successive drafts
primary PM2.5standards set at 15 m g/m3, (2) CASAC advice and of the Criteria Document and Staff
annual mean, and 50 m g/m3, 24-hour recommendations, as reflected in Paper, as well as for intensive peer
average. The proposed annual PM2.5 discussions of drafts of the Criteria review of these documents by CASAC at
standard would be based on the 3-year Document and Staff Paper at public several public meetings attended by
average of the annual arithmetic mean meetings, in separate written comments, many knowledgeable individuals and
PM2.5concentrations, spatially averaged and in the CASAC’s closure letters to representatives of interested
across an area. The proposed 24-hour the Administrator. organizations. In addition, there have
PM2.5standard would be based on the (3) Public comments received during been a number of important scientific
3-year average of the 98thpercentile of the development of these documents, conferences, symposia, and colloquia on
24-hour PM2.5concentrations at each either in connection with CASAC PM issues, sponsored by the EPA and
population-oriented monitor within an meetings or separately. others, in the U.S. and abroad, during
area. The proposal solicited comment (4) Extensive public comments this period. While significant
on two alternative approaches for received on the proposed decisions uncertainties exist, the review of the
selecting the levels of PM2.5standards. regarding the primary PM standards. health effects information has been
The Administrator also proposed to After taking this information and thorough and deliberate. In the
revise the current 24-hour primary PM10 comments into account, and for the judgment of the Administrator, this
standard of 150 m g/m3by replacing the reasons discussed below in this unit, the intensive evaluation of the scientific
1-expected-exceedance form with a 98th Administrator concludes that revisions evidence has provided an adequate
percentile form, averaged over 3 years at to the current primary standards to basis for regulatory decision making at
each monitor within an area, solicited provide increased public health this time, as well as for the
comment on an alternative proposal to protection against a variety of health comprehensive research needs
revoke the 24-hour PM10standard, and risks are appropriate. More specifically, document recently developed by EPA,
proposed to retain the current annual the Administrator has determined that it and reviewed by CASAC and others, for
primary PM10standard of 50 m g/m3. The is appropriate to establish new annual improving our future understanding of
proposal also solicited comment on and 24-hour PM2.5standards, to revise the relationships between ambient PM
proposed revisions to 40 CFR part 50, the current 24-hour PM10standard, and exposures and health effects.
Appendix K to establish new data to retain the current annual PM10 The health effects information and
handling conventions for calculating standard. As discussed more fully below human risk assessment were
98thpercentile values and spatial in this unit, the rationale for the final summarized in the proposal and are
averages, revisions to 40 CFR part 50, decisions regarding the PM primary only briefly outlined below in this unit.
Appendix J to modify the reference NAAQS includes consideration of: Subsequent units provide a more
method for monitoring PM as PM , and (1) Health effects information, and complete discussion of the
10
a proposed new reference method for alternative views on the appropriate Administrator’s rationale, in light of key
monitoring PM as PM (40 CFR part 50, interpretation and use of the issues raised in public comments, for
2.5
Appendix L). information, as the basis for judgments concluding that it is appropriate to
With regard to the secondary about the risks to public health revise the current primary standards
standards, the Administrator proposed presented by population exposures to (Unit II.B. of this preamble) and to
to revise the current secondary ambient PM. revise the specific elements of the
standards by making them identical to (2) Insights gained from a quantitative standards including indicator (Unit II.C.
the suite of proposed primary standards, risk assessment conducted to provide a of this preamble); averaging time, form,
in conjunction with the establishment of broader perspective for judgments about and level of new PM2.5standards (Units
a regional haze program under section protecting public health from the risks II.D., II.E., and II.F. of this preamble);
169A of the Act. associated with PM exposures. and averaging time, form, and level of
(3) Specific conclusions regarding the revised PM standards (Unit II.G. of
10
II. Rationale for the Primary Standards need for revisions to the current this preamble).
standards and the elements of PM 2. Summary of the health effects
A. Introduction
standards (i.e., indicator, averaging evidence. In brief, since the last review
1. Overview. This document presents time, form, and level) that, taken of the PM criteria and standards, the
the Administrator’s final decisions together, would be appropriate to most significant new evidence on the
regarding the need to revise the current protect public health with an adequate health effects of PM is the greatly
primary ambient air quality standards margin of safety. expanded body of community
for PM, and, more specifically, As with virtually any policy-relevant epidemiological studies. The Criteria
regarding the establishment of new scientific research, there is uncertainty Document stated that these recent
annual and 24-hour PM2.5primary in the characterization of health effects studies provide ‘‘evidence that serious
standards and revisions to the form of attributable to exposure to ambient PM. health effects (mortality, exacerbation of
the current 24-hour PM10primary As discussed in the proposal, however, chronic disease, increased hospital
NAAQS. These decisions are based on there is now a greatly expanded body of admissions, etc.) are associated with
a thorough review, in the Criteria health effects information as compared exposures to ambient levels of PM
Document, of the latest scientific with that available during the last found in contemporary U.S. urban
information on known and potential review of the PM standards. Moreover, airsheds even at concentrations below
human health effects associated with the recent evidence on PM-related current U.S. PM standard’’ (U.S. EPA,
exposure to PM at levels typically found health effects has undergone an 1996a; p. 13-1). Although a variety of
38656 Federal Register / Vol. 62, No. 138 / Friday, July 18, 1997 / Rules and Regulations
responses to constituents of ambient PM with: existing PM air quality levels, order of hundreds of premature deaths
have been hypothesized to contribute to projected PM air quality levels that each year, hundreds to thousands of
the reported health effects, the relevant would occur upon attainment of the respiratory-related hospital admissions,
toxicological and controlled human current PM10standards, and projected and tens of thousands of additional
studies published to date have not PM air quality levels that would occur respiratory related symptoms in
identified any accepted mechanism(s) upon attainment of alternative PM2.5 children.
that would explain how such relatively standards. The risk assessment is (2) Based on the results from the
low concentrations of ambient PM intended as an aid to the Administrator sensitivity analyses of key uncertainties
might cause the health effects reported in judging which alternative PM and the integrated uncertainty analyses,
in the epidemiological literature. NAAQS would reduce risks sufficiently the single most important factor
Unit II.A. of the proposal further to protect public health with an influencing the uncertainty associated
outlines key information contained in adequate margin of safety, recognizing with the risk estimates is whether or not
the Criteria Document, Chapters 10-13, that such standards will not be risk-free. a threshold concentration exists below
and the Staff Paper, Chapter V, on the The risk assessment is described more which PM-associated health risks are
known and potential health effects fully in the Staff Paper and summarized not likely to occur.
associated with airborne PM, alone and in the proposal. Related technical (3) Over the course of a year, the few
in combination with other pollutants reports and updates7have been placed peak 24-hour PM2.5concentrations
that are routinely present in the ambient in the docket (Abt Associates, 1996a,b; appear to contribute a relatively small
air. The information highlighted there 1997a,b). amount to the total health risk posed by
summarizes: EPA emphasizes that it places greater the entire air quality distribution as
(1) The nature of the effects that have weight on the overall conclusions compared to the aggregated risks
been reported to be associated with derived from the studies—that PM air associated with the low to mid-range
ambient PM, which include premature pollution is likely causing or concentrations.
mortality, aggravation of respiratory and contributing to significant adverse (4) There is greater uncertainty about
cardiovascular disease (as indicated by effects at levels below those permitted both the existence and the magnitude of
increased hospital admissions and by the current standards—than on the estimated excess mortality and other
emergency room visits, school absences, specific concentration-response effects associated with PM exposures as
work loss days, and restricted activity functions and quantitative risk estimates one considers increasingly lower
days), changes in lung function and derived from them. These quantitative concentrations approaching background
increased respiratory symptoms, risk estimates include significant levels.
changes to lung tissues and structure, uncertainty and, therefore, should not
B. Need for Revision of the Current
and altered respiratory defense be viewed as demonstrated health
Primary PM Standards
mechanisms. impacts. EPA believes, however, that
(2) Sensitive subpopulations that they do represent reasonable estimates 1. Introduction. The overarching issue
appear to be at greater risk to such as to the possible extent of risk for these in the present review of the primary
effects, specifically individuals with effects given the available information. NAAQS is whether, in view of the
respiratory disease and cardiovascular Keeping in mind the important advances in scientific knowledge
disease and the elderly (premature uncertainties inherent in any such reflected in the Criteria Document and
mortality and hospitalization), children analyses, the key insights from the risk Staff Paper, the existing PM standards
(increased respiratory symptoms and assessment that are most pertinent to should be revised and, if so, what
decreased lung function), and asthmatic the current decision include: revised or new standards would be
children and adults (aggravation of (1) Fairly wide ranges of estimates of appropriate. The concluding section of
symptoms). the incidence of PM-related mortality the integrative synthesis of health
(3) An integrated evaluation of the and morbidity effects and risk effects information in the Criteria
health effects evidence, with an reductions associated with attainment of Document, which CASAC characterized
emphasis on the key issues raised in alternative standards were calculated for as EPA’s ‘‘best ever example of a true
assessing community epidemiological the two locations analyzed when the integrative summary of the state of
studies, including alternative effects of key uncertainties and knowledge about the health effects of
interpretations of the evidence, both for alternative assumptions were airborne PM,’’ (Wolff, 1996b) provides
individual studies and for the evidence considered. Significantly, the combined the following summary of the science
as a whole. analysis for these two cities alone found with respect to this issue:
(4) The PM fractions of greatest that the risk remaining after attaining The evidence for PM-related effects from
concern to health. the current PM10standards was on the epidemiological studies is fairly strong, with
most studies showing increases in mortality,
The summary in the proposal will not
7The risk assessment results that appear in the hospital admissions, respiratory symptoms,
be repeated here. EPA emphasizes that
Staff Paper and are summarized in the proposal and pulmonary function decrements
the final decisions on these standards have been updated to include analyses of the associated with several PM indices. These
take into account the more particular forms of standard alternatives contained epidemiological findings cannot be wholly
comprehensive and detailed discussions in the proposal and to correct estimates for one attributed to inappropriate or incorrect
effects category (mortality from long-term exposure)
of the scientific information on these to reflect the actual statistics used in the study upon statistical methods, misspecification of
issues contained in the Criteria which they were based (Pope et al., 1995). The concentration-effect models, biases in study
Document and Staff Paper, which were corrections, which cumulatively reduce estimates of design or implementation, measurement
reviewed by the CASAC and the public. mortality associated with long-term exposures by 20 errors in health endpoint, pollution
to 35%, have no effect on risk estimates for exposure, weather, or other variables, nor
3. Key insights from the risk
mortality associated with short-term exposures or confounding of PM effects with effects of
assessment. The Staff Paper presents the the estimates for any other effects. Because the key
other factors. While the results of the
results of a quantitative assessment of sensitivity analyses that provide additional insights
epidemiological studies should be
health risks for two example cities, raengda rrdeliantge dth irsessuheosl dinsv, ocolvpeodl ltuhtea nsthso, ratv-teerramgi nexgp toimsuere interpreted cautiously, they nonetheless
including risk estimates for several studies, none of these results are affected by provide ample reason to be concerned that
categories of health effects associated changes to the long-term exposure risk estimates. there are detectable health effects attributable
Federal Register / Vol. 62, No. 138 / Friday, July 18, 1997 / Rules and Regulations 38657
to PM at levels below the current NAAQS. locations and the coherent nature of the states supporting even stronger
[U.S. EPA, 1996a, p. 13-92] observed effects9are suggestive of a standards acknowledged the lack of
Given the nature of the health effects likely causal role of ambient PM in demonstrated mechanism(s) and other
in question, this finding, which is based contributing to the reported effects. uncertainties but stressed the strength of
on a large number of studies that used 2. Comments on scientific basis for the other evidence in urging EPA to set
PM measurements, as well as studies revision. A majority of the public protective standards.
10
using other indicators of PM, clearly comments received on the proposal Many comments were also received
indicates that revision of the current PM agreed that, based on the available from representatives of environmental
NAAQS is appropriate. Quite apart from scientific information, the current PM10 or community health organizations that
the issue of whether PM should be the standards are not of themselves supported the adoption of air quality
10
sole indicator for the PM NAAQS, the sufficient to protect public health and it standards for PM2.5. These commenters
extensive PM epidemiological data base would be appropriate to revise them. agreed with EPA’s finding that a large
provides evidence of serious health Included in those calling for revisions to body of compelling evidence
effects (e.g., mortality, exacerbation of the current standards are many public demonstrates that exposure to
chronic disease, increased hospital health professionals, including particulate matter pollution, in general,
admissions) in sensitive populations numerous medical doctors and is associated with premature death,
(e.g., the elderly, individuals with academic researchers. For example, a aggravation of heart and lung diseases,
cardiopulmonary disease), as well as group of 27 members of the scientific increased respiratory illness and
significant adverse health effects (e.g., and medical community recognized as reduced lung function. They agreed
increased respiratory symptoms, school having substantial expertise in with EPA that these studies present a
absences, and lung function conducting research on the health consistent and coherent relationship
decrements) in children. Moreover, effects of air pollution stated: between exposure to PM and both
these effects associations are observed Health studies conducted in the U.S. and mortality and various measures of
in areas or at times when the levels of around the world have demonstrated that morbidity. However, the majority of
the current PM standards are met. levels of particulate and ozone air pollution these commenters argued that EPA’s
10
Although the increase in relative risk is below the current U.S. National Air Quality proposed standards for PM2.5were
small for the most serious outcomes, Standards exacerbate serious respiratory inadequate and recommended adoption
EPA believes it is significant from an disease and contribute to early death. A large of more stringent levels of the 24-hour
overall public health perspective, body of scientific and medical evidence and/or annual air quality standards for
clearly indicates that the current NAAQS are
because of the large number of not sufficiently protective of public health. PM2.5. Many of these commenters also
individuals in sensitive populations that [Thurston, 1997] urged EPA to revise the NAAQS for
are exposed to ambient PM, as well as PM10to be more protective of public
Similar conclusions were reached in a
the significance of the health effects health. These commenters based their
letter signed by more than 1,000
involved (U.S. EPA, 1996a, p. 1-21). The recommendations on the findings of the
scientists, clinicians, researchers, and
results of the two-city PM risk studies that were reviewed in the
other health care professionals (Dickey,
assessment reinforce these conclusions preparation of the Criteria Document
1997). The cosigners to this letter argued
regarding the significance of the public and Staff Paper. One commenter used
that tens of thousands of hospital visits
health risk—even under a scenario in results from five of these studies as the
and premature deaths could be
awtthaiicnhe dth.e current PM10standards are prevented with the proposed air quality obfa s1i0s mfogr/ mre3co(amnmnueanld) ianngd P 1M8 2m.5gs/tman3d(2a4rd-s
standard revisions. In fact, these
While the lack of demonstrated hour) (Dockery et al., 1993; Pope et al.,
commenters argued that even stronger
mechanisms that explain the extensive 1995; Schwartz et al., 1996; Schwartz et
standards than those proposed by EPA
body of epidemiological findings is an al., 1994; Thurston et al., 1994). The
are needed to protect the health of the
important caution, which presents commenters agreed with EPA on the
most vulnerable residents of our
difficulties in providing an integrated significance of these studies’ results and
communities.
assessment of PM health effects the need to revise the PM standards,
A number of State and local
research, a number of potential while differing with EPA’s
government authorities also submitted
mechanisms have been hypothesized in interpretation of the findings for
comments in support of adopting new
the recent literature (U.S. EPA, 1996b; p. purposes of developing the proposed
air quality standards for fine particulate
V-5 to V-8; appendix D). Moreover, PM standards.
matter. The commenters concurred with
qualitative information from laboratory Several commenters made reference
conclusions reached through the EPA’s
studies of the effects of particle to the conclusions of a number of
peer review process that the PM
components at high concentrations and international scientific panels regarding
standards should be revised to protect
dosimetry considerations suggest that the health effects of exposure to
public health. A number of these
the kinds of effects observed in airborne particulate matter—the British
commenters suggested that the
community studies (e.g., respiratory- Expert Panel on Air Quality Standards,
standards proposed by EPA should be
and cardiovascular-related responses) the British Committee on the Medical
even stronger, while several other State
are at least plausibly related to Effects of Air Pollutants, the World
agencies recommended that EPA adopt
inhalation of PM.8Indeed, as discussed Health Organization, the Canadian
PM standards, but at less stringent
in the Criteria Document and section 2.5 Ministry of Environment, Lands and
levels. A number of the comments from
V.E of the Staff Paper, the consistency Parks, and the Health Council of the
of the results of the epidemiological Netherlands -- and argued that all these
studies from a large number of different 9As noted in the proposal, the kinds of effects panels found that PM concentrations
observed in the epidemiological studies are
equivalent to the current U.S. standards
logically related. For example, the association of
8As discussed more fully below in this unit, PM with mortality is mainly linked to respiratory for PM10are not protective of human
epidemiological studies alone cannot be used to and cardiovascular causes, which is coherent with health and made recommendations for
demonstrate mechanisms of action, but they can observed PM associations with respiratory and greater protection. One commenter
provide evidence useful in making inferences with cardiovascular hospital admissions and respiratory
noted that the findings of the British
regard to causal relationships (U.S. EPA, 1996b, p. symptoms. Further, similar categories of effects are
V-9). seen in long- and short-term exposure studies. Health Panel have resulted in a British
38658 Federal Register / Vol. 62, No. 138 / Friday, July 18, 1997 / Rules and Regulations
proposal to adopt a 24-hour PM Comments, certain key points are EPA clearly specified the key criteria
10
standard of 50 m g/m3, which is one-third summarized below in this unit. by which it evaluated the available
the level of the current U.S. NAAQS. a. General comments on the use of epidemiological studies in section
In these comments, some epidemiological studies. The first 12.1.2 of the Criteria Document, with
toxicological studies were cited as category of comments was largely substantial reliance on those specified
providing evidence for toxicity of derived from ad hoc panels of by Hill (1965). In rejecting results with
particulate pollution. These commenters occupational and other epidemiological relative risks less than 1.5 to 2 as
disagreed with arguments that PM experts, consulting groups, and meaningful absent demonstrated
standards cannot be adopted due to a individual consultants. Most of these biological mechanisms, the commenters
lack of a sufficient understanding of the individuals and groups commented on fail to note that Hill and other expert
biological mechanism of injury. The the use of epidemiology in reaching groups (U.S. DHEW, 1964) have
commenters argued that there is scientific and policy conclusions emphasized that no one criterion is
sufficient evidence that particulate primarily from an occupational or definitive by itself, nor is it necessary
pollution is associated with adverse hazard assessment perspective, in that all be met in order to support a
health effects to make it inappropriate to contrast to the perspective of the review determination of causality (U.S. EPA,
delay the establishment of standards of ambient PM criteria and standards, 1996a, p. 12-3).
while further studies are undertaken. With respect to biological plausibility,
where the use of community air
This group of commenters was also Hill noted that ‘‘this is a feature I am
pollution epidemiological studies are
critical of arguments against the convinced we cannot demand. What is
central. Citing accepted criteria used in
establishment of additional PM biologically plausible depends upon the
evaluating epidemiological studies to
standards based on the possibility of biological knowledge of the day’’ (Hill,
assess the likelihood of causality (most
confounding by other pollutants, and 1965). This statement is clearly
notably those of Sir Austin Bradford
urged that more attention be paid pertinent to the toxicological and
Hill, 1965), these commenters argued
instead to the possible additive or mechanistic understanding of the effects
that in the absence of a demonstrated
synergistic effects of multiple pollutant of PM and associated air pollutants,
biological mechanism, the relative risks
exposures. especially at lower concentrations. It is
of effects in the PM epidemiological
In general, the EPA agrees with these also important to stress that while the
studies are too low (less than values
commenters’ arguments regarding the mechanistic evidence published as of
variously cited as 1.5 to 2.0) to reach
need to revise the PM standards. The the time the Criteria Document closed
any conclusions regarding causality or
scientific studies cited by these does not provide quantitative support
to form the basis for regulations. In
commenters were the same studies used for the epidemiological results, neither
general, the commenters applied these
in the development of the Criteria can such limited evidence refute these
criteria to a subset of studies evaluated
Document and the Staff Paper, and the findings. It is also important to stress
in the Criteria Document, including as
EPA agrees that there is a sufficient that our understanding of biological
few as two long-term exposure studies
body of evidence that the current mechanisms for PM pollution effects is
(EOP Group) (API, 1997), a group of 9
NAAQS for PM are not adequately not sufficient to explain the effects
selected studies (Greenland panel) (API,
protective of the public health. For observed at much higher concentrations
1997), those studies cited in the
reasons detailed in Unit II.F. of this in air pollution episodes, for which
proposal (AIHC, 1997), or as many as 23
preamble and in the Response to causality is generally accepted.
selected short-term exposure studies
Comments, EPA disagrees with aspects Moreover, the toxicological literature
examined in a recently published
of these commenters’ views on the level has only recently begun to examine
review paper (Gamble and Lewis, 1996).
of protection that is appropriate and animal models (or controlled human
supported by the available scientific Based on a careful review of these studies) that might reflect the sensitive
information. comments, EPA notes a number of populations in question (the elderly,
Another body of commenters, limitations in these commenters’ individuals with chronic respiratory
including almost all commenters evaluations of the epidemiological and cardiovascular disease) or that
representing businesses and industry studies that they considered, as adequately reproduce all of the physico-
associations, many local governmental discussed in detail in the Response to chemical properties of particles in the
groups and private citizens, and some Comments. In summary, EPA notes that ambient atmosphere. In short, the
States opposed revising the standards. these commenters provided scientific absence of evidence of a particular
Many of these commenters argued that advice and conclusions that are in mechanism is hardly proof that there are
the available scientific evidence does substantial disagreement with the no mechanisms that could explain the
not provide an adequate basis for conclusions of the review reflected in effects observed so consistently in the
revising the current standards. The the Criteria Document and Staff Paper. epidemiological studies. The absence of
central arguments made by these EPA stands behind the scientific biological mechanisms did not deter
commenters can be divided into two conclusions reached in these documents CASAC from recommending revisions
categories: (1) General comments on the regarding the appropriate use of the to the PM standards in 1982, 1986, and
appropriateness of relying on the available community epidemiological again in 1996.
epidemiological evidence for making studies. These documents were the While Hill appropriately emphasized
regulatory decisions, and (2) more product of an extended public process the strength of the association as
specific comments challenging EPA’s that included conducting public important (e.g., size of the relative risk),
appraisal of the consistency and workshops involving the leading he also pointed out that ‘‘We must not
coherence of the available information, researchers in the field, drafts of the be too ready to dismiss a cause-and-
EPA’s conclusions regarding causality, Criteria Document and Staff Paper effect hypothesis merely on the ground
and the use of these studies for risk providing opportunities for public that the observed association appears to
assessment and decisions on whether to scrutiny and comment on, and, not be slight. There are many occasions in
revise the standards. While EPA has least, receiving the advice of an medicine when this in truth is so’’ (Hill,
included comprehensive responses to independent panel of air pollution 1965). EPA believes that the effects of
these comments in the Response to experts, including epidemiologists. air pollution containing PM is such a
Federal Register / Vol. 62, No. 138 / Friday, July 18, 1997 / Rules and Regulations 38659
case. Unlike the ‘‘textbook’’ examples of different conclusions about the pollution, it is not realistic to require
unlikely significant associations consistency of the observed associations personal monitors in air pollution
provided by some commenters (e.g., ice because of their assumptions that all studies of daily mortality, which require
cream consumption correlated with heat model building strategies by all authors urban scale population data over a
stroke), the abundant epidemiological are equally valid. Even the most period of years. Furthermore, the use of
literature on combustion particles thorough of these treatments (Gamble community monitoring-based
documents numerous occasions in and Lewis, 1996) shared this flaw, epidemiological studies as a basis for
which single short-term episodes of particularly in the discussion of the establishing standards and guidelines
high air pollution produced series of Philadelphia mortality studies has a long history in air pollution,
unequivocally elevated relative risks. and in the discussion of modeling including the British authorities’
For the week of the well documented approaches. The authors’ treatment of response to the London episodes and
1952 London air pollution episode, for modeling and confounding issues was the establishment of the original U.S.
example, the relative risk of mortality further limited because they did not NAAQS in 1971. Rejecting the use of the
for all causes was 2.6, while the relative include the most recent Philadelphia vast array of such studies on this basis
risk for bronchitis mortality was as high results (Samet et al., 1996a,b) sponsored alone would also go against the advice
as 9.3 (Ministry of Health, 1954). by the Health Effects Institute (HEI, of the independent scientific experts on
Hospital admissions also increased by 1997). One of the important functions of every CASAC panel that has addressed
more than a factor of two. British the Criteria Document is to evaluate the the subject of PM pollution through the
epidemiologists in the 1950s concluded strengths and limitations of various years, each of which has recommended
that increased mortality was likely studies. As discussed more fully below general PM standards based primarily
when PM (as mass calibrated British in this unit, the Criteria Document on the results of community
Smoke <4.5 m m in aerodynamic found that some of the studies cited by epidemiological studies (Friedlander,
diameter) exceeded 500 m g/m3(Martin commenters as suggesting a lack of 1982; Lippmann, 1986; Wolff, 1996b).
and Bradley, 1960). This is only about consistency had important limitations. As noted above in this unit, EPA has
a factor of 3 higher than that allowed by In general, these commenters’ analyses included a more detailed discussion of
the current PM standard. Unlike the suffered by ignoring the much more its responses to these comments in the
‘‘textbook’’ and other unlikely statistical thorough critical review of these studies Response to Comments.
associations noted by some commenters, and issues contained in the Criteria b. Specific comments on
where the only evidence is for low Document, notably that in section 12.6 epidemiologic studies. The second
relative risk, clear and convincing links on alternative modeling approaches. group of commenters noted above made
between high-level PM concentrations
EPA also rejects the notion advanced more specific challenges to EPA’s
and mortality and morbidity buttress the
by these commenters that assessment of the epidemiological
findings of similar associations at much
epidemiological studies must use studies. These comments, although
lower PM concentrations as suggested in
personal exposure monitoring to be overlapping some of those made by the
the more recent epidemiological
considered for regulatory purposes. In first group, were generally made by
literature.
particular, commenters ignore the commenters who have taken a more
These commenters also appear to
significant strengths of the time-series active role in the review of the Criteria
ignore several epidemiological studies
studies and prospective cohort studies Document and Staff Paper. These
conducted at low PM concentrations in
relied on by EPA as compared to cross- commenters asserted that the
U.S. and European cities, including both
sectional epidemiological studies. Time- epidemiological evidence on PM is not
short- and long-term exposures to PM
series studies, such as the daily as consistent and coherent as EPA has
air pollution, that find statistically
mortality studies, look at changes in claimed, and, in particular, charged that
significant relative risks of respiratory
response rate in relation to changes in EPA ignored or downplayed a number
symptom categories in children in the
range of 1.5 to 5 (Schwartz et al., 1994; weather and air pollution over time of studies that the commenters argue
Pope and Dockery, 1992; Braun- intervals of a few days. This controls for contradict the evidence the Agency
Fahrlander et al., 1992; Dockery et al., other factors such as smoking and cited as supporting the consistency and
1989; Dockery et al., 1996). socioeconomic status, which are little coherence of PM effects. The studies, all
Concentrations in these studies extend changed during such short intervals. of which commenters contend do a
from moderately above to well below Prospective cohort studies (e.g., Pope et better job of addressing one or more key
those permitted by the current PM al., 1995; Raizenne et al., 1996), on the issues, such as confounding pollutants,
10
standards. While, as noted in the other hand, look at changes in health weather, exposure misclassification, and
proposal, most of the recent status in a selected cohort of model specification, than earlier
epidemiological studies of mortality and individuals, which allows direct studies, include several that were
hospital admissions report adjustment for smoking status, available during preparation of the
comparatively small relative risks, the socioeconomic status, and other subject- Criteria Document, and a number that
findings of relative risks well in excess specific factors. The commenters also appeared after the Criteria Document
of the 1.5 to 2 criterion noted by ignore the Criteria Document and Staff Paper were completed.
commenters for earlier studies of high conclusions on how properly conducted Because the status of the later studies
PM episodes, as well as the relative monitoring can provide an adequate differ from that of the earlier ones for
risks of 1.5 to 5 reported in more recent index of population exposure to purposes of decisions under section 109
studies of less serious, but still ambient air pollution in such studies of the Act, the two categories are
important effects categories, lend that, as detailed below, is more relevant discussed separately below in this unit.
credibility to EPA’s interpretation of the to establishing ambient air quality Additional responses to comments
results. standards (U.S. EPA 1996a, chapter 7). relating to both sets of studies have been
In addition to basing their Although personal monitoring may be included in the Response to Comments.
conclusions primarily on their own practical for some occupational and In addition to the inclusion of specific
assessment of a limited set of studies, epidemiological studies, and has been studies, commenters also raised other
this group of commenters reached employed in some past studies of air issues regarding the limitations of the
38660 Federal Register / Vol. 62, No. 138 / Friday, July 18, 1997 / Rules and Regulations
epidemiological information and the use et al., 1995)11, with extended analyses other investigators (Li and Roth, 1995;
of these studies in EPA’s two-city risk for Philadelphia in Phase I.B (Samet et Wyzga and Lipfert, 1995). Further
assessment. Both of these topics are also al., 1996a,b). The most important analytical studies of the Philadelphia
discussed below in this unit. finding in the HEI Phase I.A reanalyses data set were carried out by HEI (Samet
(i) Studies available for inclusion in of the six areas is ‘‘the confirmation of et al., 1996a,b) and have largely resolved
the criteria review. With some the numerical results of the earlier many of the uncertainties in the earlier
exceptions, most of the above analyses of all six data sets’’ (HEI, analyses; in EPA’s opinion, these
commenters cited somewhat similar 1995)12. After replicating the original studies supersede the results of the
lists of ‘‘negative’’ studies that they investigators’ analyses, Samet et al. original investigators (Schwartz and
argue EPA ignored or downplayed in (1995) also found similar results Dockery, 1992a) and the several earlier
arriving at conclusions on consistency analyzing the data using an improved reanalyses, including Moolgavkar
and coherence. Of the most commonly statistical model. The HEI Oversight (1995a), Moolgavkar and Luebeck
cited studies, the following were Committee found (1996), Li and Roth (1995), Wyzga and
available for inclusion in the Criteria Lipfert (1995), and Samet et al. (1995).
[I]t is reasonable to conclude that, in these
Document: daily mortality studies by six data sets, daily mortality from all causes Even though TSP is not the best PM
Styer et al. (1995), Lyon et al. (1995), Li combined, and from cardiovascular and indicator for health effects, since it
and Roth (1995), Moolgavkar (1995a,b), respiratory causes in particular, increases as includes a substantial fraction of non-
Wyzga and Lipfert (1995), Lipfert and levels of particulate air pollution indexes thoracic particles, the extended Criteria
Wyzga (1995), and Samet et al. (1995, increase. [HEI, 1995] Document assessment (U.S. EPA, 1996a,
1996a,b); the long-term exposure It is important to note that these pp. 12-291 to -299; 12-327) of the Phase
mortality study by Abbey et al. (1991); reanalyses by respected independent I.B HEI analyses in Philadelphia (Samet
and the re-examination of the Six-City scientists confirm the reliability and et al., 1996a,b) serves to support the
mortality results (Dockery et al., 1993) reproducibility of the work of the following findings:
by Lipfert (1995). original investigators, particularly in (1) The mortality effects estimates for
The written record of EPA’s view of the concerns some commenters TSP do not depend heavily on statistical
evaluations of these studies effectively have expressed about EPA’s reliance on methods when appropriate models are
refutes the claim that the Agency a number of PM studies published by used.
ignored any of these studies and these authors. (2) Estimated PM effects are not
supports the treatment the Agency The Phase I.A HEI results for highly sensitive to appropriate methods
accorded to each of them. All of the Philadelphia also found that it was for adjusting for time trends and for
studies available to EPA at the time of difficult to separate the effects of PM weather.
CASAC closure on the PM Criteria from those of co-occurring SO2, in (3) Air pollution has significant health
Document (March 1996) were examined agreement with the Moolgavkar et effects above and beyond those of
for inclusion in the Criteria Document al.(1995a) analysis. Subsequent HEI weather.
and Staff Paper, which form the basis work, and several of the other so-called (4) Copollutants such as ozone, CO,
for the PM proposal. ‘‘Negative’’10 ‘‘negative’’ studies cited above in this and NO2may be important predictors of
studies were evaluated in detail along unit, further examined this issue in mortality, but their effects can be
with ‘‘positive’’ studies when they were terms of confounding or effects substantially separated from those of
found to have no critical methodological modification by one or more co- TSP and SO2.
(5) The health effects of TSP in
deficiencies, or to point out strengths occurring gaseous pollutants or weather.
Philadelphia cannot be completely
and limitations. Studies that had more Contrary to commenters’ claims, this
separated from SO , which is itself a
serious problems were generally issue and these studies received 2
precursor of fine particles, based solely
discussed in less detail, whether considerable attention in the Criteria
on the epidemiological analyses in this
positive or negative, than studies with Document and Staff Paper, and the
single city.
fewer or small deficiencies. The EPA overall implications and conclusions
The most recent HEI Oversight
assessments were evaluated by peer from these assessments were
Committee comments on these studies
reviewers, by CASAC, and by the summarized in the proposal. In
(HEI, 1997), which were submitted to
public. particular, the so-called ‘‘negative’’ and
the docket by HEI, state that:
Most of the short-term exposure other findings of Moolgalvkar et al.
studies cited above in this unit are (1995a,b) in their Philadelphia and Although individual air pollutants (TSP,
reanalyses and extensions of PM/ Steubenville studies were discussed in SO2, and ozone) are associated with
mortality studies that had been great detail in section 12.6 of the PM increased daily mortality in these data, the
limitations of the Philadelphia data make it
published by other investigators. In Critera Document and compared to
impossible to establish that particulate air
general, the Criteria Document those of the original investigators
pollution alone is responsible for the widely
concluded that the most comprehensive (Schwartz and Dockery, 1992a,b) and observed associations between increased
and thorough reanalyses were those in mortality and air pollution in that city. All
the series conducted for the HEI, which 11Data sets were those used in the original we can conclude is that it appears to play a
reanalyzed data sets used in studies studies by Dockery et al. (1992) for St. Louis and role. [HEI, 1997; p.38.]
Eastern Tennessee; Pope et al. (1992) for Utah
from six urban areas in Phase I.A (Samet Valley; Schwartz and Dockery (1992a) for While recognizing the limitations in the
Philadelphia; Schwartz (1993) for Birmingham; and conclusions that can be made based on
10The term ‘‘negative’’ studies, as used in these a portion of the Santa Clara data from Fairley studies in a single city, the Oversight
comments, should not be construed to mean those (1990). The data set from the Moolgavkar et al. Committee endorses the approach taken
in which there is a negative effects estimate (either (1995a) Philadelphia reanalysis was also included
significant or non-significant) for the nominal (Samet et al., 1995). by EPA in evaluating a broader set of
cause. As used by these commenters, the term also 12The HEI Board of Directors appointed an eight epidemiological studies:
includes statistically non-significant positive effect member Oversight Committee consisting of leading
Consistent and repeated observations in
estimates. In other words, the commenters define scientists in several disciplines relevant to air
‘‘positive’’ studies as including only those in which pollution epidemiology to oversee key aspects of locales with different air pollution profiles
the effect estimate is both positive and statistically the project and to prepare HEI’s assessment of the can provide the most convincing
significant. results. epidemiological evidence to support