vvEPA
450/05 86-012
        REVIEW OF THE NATIONAL AMBIENT AIR
    QUALITY STANDARDS  FOR PARTICULATE MATTER

        UPDATED ASSESSMENT  OF SCIENTIFIC
            AND TECHNICAL INFORMATION


    ADDENDUM TO THE 1982 OAQPS  STAFF PAPER
MASS
60
5O-
M 4O
g
i 3c>
*• 20-
10-
»' 3-

IS7
si zw r
-n n «

Inl



f
X
X
X
X


x1
X
X
X
X
X

2-


1-

CALCIUM
p


^^
n
k
D ^ J n ^





n

PTKWSL'S P T K W SL» S
r-
6
5-
"c 4-
^»
* 3-
4 J
2-
I
SULFUR
^
1
IT1

1
I

3-

2-
1-
IRON





r

.
aDanUl
P T K * 3L* S ? - K W SL* S
3-

2


1-

SILICA
-
_
,_.

_








^ '/,
'•?!?. L ft £
P T K W SL* S
LEAD

"

3-


G^
lip
aJlJJ
|
p r K w SL*S
                Strategies and Air Standards Division
             Office of Air Quality Planning and Standards
                U.S. Environmental Protection Agency
                 Research Triangle Park.  N.C. 27711
                       December 1986

-------
Cover Illustration.   Mean  concentration  of  components  of  participate
matter from dichotomous sampling  (1978-79)  in  Portage,  WI (P),  Topeka,
KS (T), Kingston, TN (K),  Watertown,  MA  (W), St.  Louis, MO (SL)  *(1976),
and Steubenville, OH (S).   Shading  represents  fine  fraction  (<  2.5 pm),
remainder is coarse  fraction up to  a  nominal 15 pm.  These communities
are the subject of the Harvard  "Six Cities  Study" of the  health  effects
of air pollution.  The cities were  chosen to reflect a  gradient  in PM
and SOX air pollution.  Although  a  major component  of  this study—
reflecting longitudinal analyses—has not been completed, the results
of cross-sectional analyses (Ware et  al., 1986) and a  series  of
episode studies (Dockery et al.,  1982) have been  identified  as
being among the more important  recent publications  for  examining
the health effects of particulate matter.   The data in  the figure
illustrate the variations  in particle mass  and composition among
these cities during  the period when these studies were  being  conducted.

Reference.  Spengler, et al. (1980).   Fine  Particle Measurements in
Six U.S. Cities.  In Proceedings  of the  Technical Basis for  a Size
Specific Particulate Standard Speciality Conference.   March  1980.

-------
REVIEW OF THE NATIONAL AMBIENT AIR QUALITY STANDARDS FOR PARTICIPATE MATTER:

        UPDATED ASSESSMENT OF SCIENTIFIC AND TECHNICAL INFORMATION



                  ADDENDUM TO THE 1982 OAQPS STAFF PAPER
                  Strategies and Air Standards Division
               Office of Air Quality Planning and Standards
                   U.S. Environmental Protection Agency
                   Research Triangle Park, N.C.  27711
                              December 1986

-------
                                    11
                             Acknowledgments

     This staff paper is  the  product of  the Office of Air Quality
Planning and Standards.   The  principal authors  include John Bachmann
and Jeff Cohen.  The report  incorporates comments from OAOPS,  the Office
of Research and Development,  the Office  of Policy, Planning,  and Evaluation,
and the Office of General  Counsel within EPA  and was formally  reviewed  hy
the Clean Air Scientific  Advisory Committee.
     Helpful comments and suggestions were also submitted by  a number of
independent scientists, by officials from the California Air  Resources.
Board, and by environmental  and  industry groups including the  National
Resources Defense Council, the American  Lung  Association, the  American  Iron
and Steel Institute, the  American Mining Congress, the Utility Air  Regulatory
Group, Consolidation Coal  Company,  the Mining and Reclamation  Council of
America, the Indiana Coal  Council,  Phelps Dodge Corporation,  and Middle
South Services.
     The authors wish to  thank Teresa demons and Tricia Holland for word
processing, and Dick Atherton for graphics assistance.

-------
                                    1.1 i


                             TABLE OF CONTENTS

                                                                        Page

 List  of Figures	    iv

 List  of Tables	    iv

 Executive Summary 	     v

   I.   Introduction  	     1

       A. Purpose	     1

       B. Background	     1

       C.,Approach  	     4

 II.   Air Quality Considerations	  .      5

      A. Current PM^g Concentrations   	      6

      B. Historical Trends in Six Cities   	      9

III.   Critical  Elements  in the Review of  the Primary Standards   ...     12

      A.  Mechanisms	     12

      B.  Concentration/Response Information	     16

 IV.   Factors to be Considered in Selecting Primary Standards  for

      Particles	     32

      A.  Pollutant Indicator	     32

      8.  Level of the Standards	     37

      C.  Summary of Staff Conclusions  and Recommendations   .....     60

 Appendix A.  Summary of Recent Epidemiological  Studies  on
              Particulate Matter  	     A-l

 Appendix B.  Calculation of PM^g/TSP Relationships 	     8-1

 Appendix C.  CASAC Closure Memorandum  . . .	     C-l

 References

-------
                                  IV.
Number

2-1


2-2

3-1

4-1

4-2


4-3
4-4



Number

1


2


2-1

3-1



3-2



4-1


4-2


A-l


A-2


A-3
               LIST OF FIGURES


     Concentrations from Dichotomous Samplers
in EPA IP Network

Trends in Seasonal  Particle Fractions in Steubenville

Regional  Deposition of Monodisperse Aerosols

Estimates of Thoracic Deposition of Particles

Mean Daily Mortality vs. Mean British Smoke for
Days with BS < 500 ug/m^ During 14 London Winters

Mean Change in FVC Compared to Baseline for
Children in Relation to Occurrence of Pollution
Episodes in Steubenville, Ohio, and the Ijmond
Area of the Netherlands

Adjusted Frequency of Cough for Children Living in 6
U.S. Cities vs. 4-year Average Estimated PM^o Levels

                LIST OF TABLES
Updated Staff Assessment of Short-Term Epidemiological
Studies

Updated Staff Assessment of Long-Term Epidemiological
Studies

Estimated Counties Exceeding Proposed Standard Limits

Summary of Recent (1982-86) Epidemiological Studies
Providing Most Useful  Concentration-Response
Information for Acute Particle Exposures

Summary of Epidemiological  Study Providing Most
Useful Concentration-Response Information for
Long-Term Particle Exposures (1982-86)

Updated Staff Assessment of Short-Term
Epidemiological Studies

Updated Staff Assessment of Long-Term
Epidemiological Studies

Epidemiological Studies (1982-86) on Short-Term
Changes in Mortality and Exposure to Particles

Epidemiological Studies (1982-86) of Effects
on Mortality Due to Long-Term Exposures to Particles

Epidemiological Studies (1982-86) of Effects
on Morbidity Due to Long-Term Exposures to Particles
Page

  7


 10

 13

 35

 42


 46
 56



Page


 ix


 xi

  9

 18



 29



 50


 58


 A-2


 A-4


 A-5

-------
                             EXECUTIVE SUMMARY





     This paper evaluates and interprets the updated scientific and



technical information that the EPA staff believes is most relevant to



decision making on revised primary (health) national ambient air quality



standards (NAAQS) for particulate matter and is an addendum to the 1982



particulate matter staff paper.  The paper assesses the factors the staff



believes should be considered in selecting the pollutant indicator and



level for the primary particulate matter standards, updating and supplementing



previous staff conclusions and recommendations in these areas to incorporate



more recent information.  This assessment is intended to help bridge tne



gap between the scientific review contained in the EPA criteria document



addendum "Second Addendum to Air Quality Criteria for Particulate Matter



and Sulfur Oxides (1982):  Assessment of Newly Available Health Effects



Information" and the judgments required of the Administrator in making



final decisions on revisions to the primary NAAQS for particulate matter



that were proposed in March 1984 (49 FR 10408).  The staff paper and this



addendum are, therefore, important elements in the standards review process



and provide an opportunity for public comment on proposed staff recommenda-



tions before they are presented to the Administrator.



     Particulate matter represents a broad class of chemically and physically



diverse substances that exist as discrete particles (liquid droplets or



solids) ranging in size from molecular clusters of 0.005 micrometers (um)



to coarse particles on the order of 100 um.  The major chemical and physical



properties of particulate matter vary greatly with time, region, meteorology



and source category, complicating the assessment of health and welfare



effects as related to various indicators of particulate pollution.  The

-------
                                     vi
original  measurement method for  the  particulate matter  NAAQS  was  the "hi
volume" sampler, which collects  particles  of sizes  up to  a  nominal  25-45 urn
(so called "Total  Suspended Particulate"  or  TSP).   EPA  has  proposed to
replace this particulate matter  indicator  with one  that includes  only
particles with aerodynamic  diameters smaller than a nominal  10 urn,  termed
"PMig".  Although a large number of  PM^g  monitors are now in  place, reliable
and consistent data are, at present, limited.  Data from  39  sites in EPA's
IP network show long-term urban  PM^g levels  range between 25  and  75 ug/nr and
maximum 24-hour values range from 50 to 175  ug/m3.   Higher  values are
likely as more data become  available.  Both  fine  (<2.5  urn)  and coarse
(>2.5 urn) particles are substantial  components of PM^g  mass,  with a tendency
for higher coarse contributions  in western US locations with  higher concentra-
tions.  National estimates  of PM^g levels  are derived from  applying measured
PM10/TSP ratios to the wider TSP data  set.  This  analysis (for 1983-85 data)
estimated that 193 counties exceeded the  lower bound of the  ranges  proposed
                            .-                o
for PM-^Q standards (150 ug/m  24 hour, 50  .ug/m  annual) while 136 counties had
sites that exceeded the current  primary TSP  standards.
Particle Indicator
     Based on an examination of  air  quality  composition,  respiratory tract
deposition, and health effects and related considerations,  the 1982 staff
paper recommended adoption  of the size specific indicator (PMig)  proposed
in 1984.   The present staff assessment of  the more  recent informaton on
respiratory tract deposition contained in  the criteria  document addendum
reinforces the conclusions  reached in the  original  staff  assessment in
1982.  The staff finds that the  recent data  do not  support  alternative
indicators that have been suggested, which exclude  all  particles  larger
than 10 urn.  The PM^g indicator  is generally conservative over the  range  of
tracheobronchial deposition.

-------
                                    vn


Recent information suggesting enhanced tracheobronchial particle deposition

for children relative to adults provides an additional reason for an indicator

that includes particles capable of such penetration.  Given these considerations

and its earlier conclusions, the staff reaffirms its recommendation to

replace TSP as the particle indicator for the primary standards with a new

indicator that includes only those particles smaller than a nominal 10 urn

in aerodynamic diameter (PM^o).  The previously developed effectiveness

criteria for samplers are acceptable for regulatory purposes.

Level  of Standards

     The major scientific basis for selecting PM standards that have an
                                                      4
adequate margin of safety remains community epidemiological research, with

mechanistic support from toxicological and controlled human investigations.

The limitations of epidemiological studies for these purposes must, however,

be recognized.  Such studies, while representing real world conditions, can

only provide associations between a complex pollutant mix measured at

specific locations and times and a particular set of observable health

points.  Difficulties in conducting and interpreting^ epidemilogical studies

limit  the reliance that can be placed on the results of any single study.

None of the available studies have used PM^g as a direct measure of pollution,

requi ring—where appropriate—further conversion of results to estimated

PM^Q units.

     The 1982 criteria document and the criteria document addendum identify

a limited set of epidemiological studies most useful for developing quanti-

tative conclusions regarding the effects of particulate matter.  This

updated staff assessment incorporates the previous evaluation of the earlier

studies as well as'the present assessment of more recent studies.

-------
                                    viii



     The updated staff assessment  of.the  short-term epidemiological  data


is summarized in Table 1;  levels are expressed  in  Doth  the  original


(British smoke—"BS" or TSP)  and PMi0 units.  The  "effects  likely"  row denotes


concentration ranges derived  from  the criteria  document  and its  addendum at'


or above which a consensus judgment  suggests  greatest certainty  that some


effects would occur, at least under  the conditions that  obtained in  the


original studies.  The data do not,  however,  show  evidence  of  clear  population


thresholds but suggest a continuum of response  with both the risk of effects


occuring and the magnitude of any  potential effect decreasing  with  concen-


tration.  This is particularly true  for the statistical  analyses of  daily
                                           4

mortality in London.  Substantial  agreement exists that  wintertime  pollution


episodes produced premature mortality in  elderly and ill  populations, but


the range and nature of association  provide no  clear basis  for distinguishing


any particular lowest "effects likely" levels or for defining  a  concentration


below which no association remains.   The  recent lung function  studies in


children suggest that effects are  possible in the  range  listed in Table 1,


but the relationships are  not certain enough  to derive  "effects  likely"


levels for PMio  They do  suggest  levels  below  which detectable  functional


changes are unlikely to occur.


     Based on this staff assessment  of the short-term epidemiological data


the range of 24-hour PM^Q  levels of  interest  are 140 to  250 ug/m .   The


up'per end of the range reflects the  judgment  of the Administrator with


regard to the maximum level proposed in 1984  for a 24-hour  standard, based


on his consideration of the earlier  criteria  and assessments.  Although the


recent information provides additional support  for the  possibility of


effects at lower levels, it 'does not demonstrate that adverse  effects would

-------
                         TABLE 1.  UPDATED STAFF ASSESSMENT OF SHORT-TERM EPlUEfUOLOGICAL  STUDIES
Effects/Study
Effects Likely
Effects Possible
No Significant
Effects Noted
o
Measured British Smoke Levels (as ug/m )
(24-hr, avg.)
Daily Mortality
in London*
1000
J
?
-
Aggravation of
Bronchitis2
250*-500*
< 250*
-
Combined
Range
250-500
<250
-
o
Measured TSP Levels (ug/nr)
(24-hr, avg.)
Small, reversible declines
in lung function in children3*
-
220*-4203
200-2504
125*4-1603
Equivalent PMi0
Levels (gg/m3)
Combined
Range5
350-600
140-350
<125
*Indicates levels used for upper and lower bound of range,

^Various analyses of daily mortality encompassing the London winter of 1958-59,  14 winters  from 1958-72,  in  aggregate
 and individually.  Early winters dominated by high smoke and S02 from coal  combustion  with frequent  fogs.   From  1982  CD:
 Martin and Bradley (1960); Ware et al., (1981); Mazumdar et al. (1981).   From 1986 CD  Addendum:   Mazumdar et  al.  (1982);
 Ostro (1984); Shumway et al., (1983); Schwartz and Marcus (1986).  Later studies show  association across entire  range of
 smoke, with no clear delineation of "likely" effects or threshold of response possible.
2Study of symptoms reported by bronchitis patients in London, mid-50's to early  70's; Lawther  et  al .  (1970).
3Study of pollution "episodes" in Steubenville, Ohio, 1978-80; Dockery et al.  (1982).
Jstudy of 1985 pollution episo'de in Ijmond, The Netherlands; Dassen et al. (1986).
^a) Conversion of BS readings to PM^j levels:  Assumes for London conditions and BS readings in the range 100-500 ug/m ,
 BS 
-------
                                                        •3
occur with certainty at  a PM^Q  concentration  of  250  ug/m .   This level,


therefore, remains an appropriate  upper  bound.   The  recent  data suggest


that the range of levels under  consideration  of  alternative .standards can


be reduced to 140 ug/m3, although  the  original lower bound  of 150 ug/rn3


is within the range of uncertainty associated with expressing the data as


PMiQ.  Neither the studies used to derive  this range nor the  more qualitative


studies of effects in other sensitive  population groups  (e.g., asthmatics)


or effects in controlled human  or  animal  studies provide convincing scientific


support for health risks of consequence  below 140 ug/m3  in  current U.S.


atmospheres.  These qualitative data,  as well as factors such as aerosol


composition and exposure characteristics,  should also be considered in


evaluating margins of safety associated  with  alternative standards in the


range of 140 ug/m3 to 250 ug/m3.


     The amended staff assessment  of the more quantitative  long-term


epidemiological data is  summarized in  Table 2.   Long-term studies "are


subject to additional confounding  variables that reduce  their sensitivity


and make interpretation  more .difficalt.   The  most important new study shows


a gradient of responses  in children among  six U.S. cities that follows the


measured gradient in particulate matter,  but  response comparisons for


locations with somewhat  smaller pollution  gradients  within  some of


these cities do not follow the  same patterns.  The results  of a separate


series of studies on long and intermediate term  (2-6 weeks) exposures in  a


number of U.S. cities (Ostro, 1983, 1987;  Hausman et. al, 1984) is more


supportive of the possibility of within  city  effects at  comparable U.S.


exposure levels.  Thus some risk of effects is possible  at  levels somewhat


below those suggested by the 1982  assessment, but it is  uncertain given the


potential for confounding present  in these more  recent studies.

-------
                         TABLE 2.  UPDATED STAFF ASSESSMENT OF'LONG-TERM EPIDEMIOLOGICAL  STUDIES

Effects/Study
Effects Likely
Effects Possible
No Significant6
Effects Noted
Measured BS
Levels (as ug/m3
Increased
Respiratory
Disease, Reduced
Lung Function
in Children1
230-300 BS
<230 BS
'
Measured TSP Levels (ug/m3)
Increased Respiratory
Disease Symtpoms,
Small Reduction in
Lung Function in
Adults^
180*
130-180*
80-130
Increased
Respiratory
Symptoms
in Adults-*
-
60-150(110)

Increased
Res'pi ratory
Symptoms and
Illnesses in
Children4
-
60*-114
-
Reduced
Lung
Function
in
Children4
-
-
40-114
Combined
Range
>_180
60-180
<60
Equivalent
PM^Q Levels
(ug/m3)
Combined
Range**
80-90
40-90
• <40

*Indicates levels used for upper and lower bound of range.

*Study conducted in 1963-65 in Sheffield, England (Lunn et al., 1967).  BS levels (as ug/m3)  uncertain.
2Studies conducted in 1961-73 in Berlin, N.H. (Ferris et al., 1973, 1976).  Effects level  (180 ug/m3)
 based on uncertain 2-month average.  Effects in lung function were relatively small.
"3Study conducted in 1973 in two Connecticut towns. (Bouhuys et al. 1973).  Exposure estimates  reflect  1965-73 data  in
 Anson.  Median value (110 ug/m3) used to indicate long-term concentration.  No effects  on lung function,  but some
 suggesstion of effects on respiratory symptoms.
4Study conducted in 1976-1980 in 6 U.S. cities (Ware et al., 1986).  Exposure estimates  reflect 4-year averages  across
 cities.  Comparable pollution/effects gradients not noted within cities.
 Conversion of TSP to PMjM equivalents for Berlin, Ansonia studies based on estimated ratio of PMjy/TSP  for  current
 U.S. atmospheres (Pace, 1983).  The estimated ratio ranged between 0.45 and U.5.  Conversion for  six-city study
 based on site-specific analysis of particle size data (Spengler et al., 1986).
^Ranges reflect gradients in which no significant effects were detected for categories at  top.  Combined range
 reflects all columns.

-------
                                    XII




     Based on this updated assessment of the long-term epidemiological data,


the staff recommends that the range of annual PM^Q levels of interest be


40 to 66 ug/m3.  The upper-end of the range reflects the judgment of the


Administrator with regard to the maximum level  proposed "for an annual


standard, based on his consideration of the earlier criteria and assessment.


The staff concludes that this level remains a useful upper bound.  The


recent data prompt consideration of a standard  level below the previous


lower bound (50 ug/m3) to values as low as 40 pg/m3.  Uncertain data from


one recent study of six cities suggest that at  this level some risk may


remain of respiratory effects in children, but  no detectable increases in
  i

pulmonary function are expected in children or  adults.


     When evaluating margins of safety for an annual standard, it is


particularly important to examine the results of qualitative data from a


number of epidemiological, animal, and air quality studies.  These suggest


concern for effects not directly evaluated in the studies used to develop -


the ranges.  Such effects include damage to lung tissues contributing to


^chronic respiratory disease, cancer, and premature mortality.  The available


scientific data do not suggest major risks for  these effects categories at


current ambient particle levels in most U.S. areas.  Nevertheless, the risk


that both fine and coarse particles may produce these responses supports


the need to limit long-term levels of PM^Q for  a variety of aerosol


compositions.


     When selecting final standard levels, consideration should be given to


the combined protection .afforded by the 24-hour and annual standards taken


together.  For example, a 24-hour standard at 150 ug/m3 would substantially


reduce annual levels in a number of areas below 50 ug/m3 adding to the

-------
                                    xiii

protection afforded by an annual standard in areas with higher 24-hour
peak to annual mean ratios.
     Because of different form, averaging procedures, size range, and
limited PM^g data, precise comparison between the above ranges of PM^Q
standards and the current primary TSP standards is not possible.  A staff
analysis of Pl^g/ISP ratios applied to recent TSP data shows that the
revised lower bounds,  taken together, would result in standards clearly
more stringent that the current standards.  In various analyses, standards
at the lower bound of  the previous range (150,50) have appeared to range
from more stringent to approximately comparable to the-present primary
standards.  Standards  at the upper end of the range could, however, result
in about a four-fold decrease in the number of areas exceeding the primary
standards.

-------
XIV

-------
REVIEW OF THE NATIONAL AMBIENT AIR QUALITY STANDARDS FOR PARTICIPATE MATTER:
        UPDATED ASSESSMENT OF SCIENTIFIC AND TECHNICAL INFORMATION

                  ADDENDUM TO THE 198Z OAQPS STAFF PAPER


I.  INTRODUCTION

A.  Purpose

        This paper evaluates and interprets the most relevant scientific

and technical information reviewed in the EPA document, Second Addendum

to Air Quality Criteria for Particulate Matter and Sulfur Oxides (1982):

Assessment of Newly Available Health Effects Information (EPA, 1986) and

represents an update of the 1982 particulate matter staff paper (EPA, 1982a).

This staff paper addendum is intended to help bridge the gap between the

scientific review of recent health effects information contained in the

criteria document addendum and the judgments required of the Administrator

in making final decisions on the proposed revisions to the primary national

ambient air quality standards (NAAQS) for particulate matter (49 FR 10408).

As such, particular emphasis in this paper is placed on conclusions,

recommendations, and uncertain-ties regarding the pollutant indicator and

levels for the primary standards.  While the paper should be of use to all

parties interested in the standards review, it is written for those decision

makers, scientists, and staff who have some familiarity with the technical

discussions contained in the criteria document addendum.

8.  Background

     1.  Legislative Requirements

     Since 1970 the Clean Air Act as amended has provided authority and

guidance for the listing of certain ambient air pollutants which may endanger

public health or welfare and the setting and revising of NAAQS for those

pollutants.  Primary standards must be based on health effects criteria and

provide an adequate margin of safety to ensure protection of public health.

-------
                                    2



As several  recent judicial  decisions have  made clear,  the  economic and



technological  feasibility of attaining  primary standards are not to be



considered  in  setting them,-although such  factors may  be considered to a



degree in the  development of state  plans to  implement  the  standards (D.C.



Cir., 1980, 1981).  Further guidance provided in the legislative history



of the Act  indicates that the standards should be set  at "the maximum



permissible ambient air level  .  .  . which  will protect the health of any



(sensitive) group of the population." Also,  margins of safety are to be



provided such  that the standards will afford "a reasonable degree of



protection  . . .  against hazards which  research has not yet identified."



(Committee  on  Public Works, 1974).   In  the final analysis, the EPA



Administrator  must make a policy decision  in setting primary standards,



based on his judgment regarding the  implications of all the health effects



evidence and the  requirement that  an adequate margin of safety be provided.



     2.  Original PM Standards and  Proposed  Revisions



     The current  primary standards  for  particulate matter  (to protect



public health) are 75 micrograms per cubic meter (ug/m^) annual  geometric



mean, and 260  ug/m^, maximum 24 hour concentration not to  be exceeded more



than once per  year.  The reference  method  for measuring attainment of the



primary standards is the "hi-volume" sampler (40 CFR Part  50, Appendix



B), which effectively collects particles in  the range  of up to 25-45



micrometers (urn)  in diameter (so-called "total suspended particulate," or



"TSP").  Thus, TSP is the current  indicator  for the particulate  matter



standards.



     On March  20, 1984, EPA proposed changes in the standards (49 FR 10408)



based on the Agency's review and revision  of the health and welfare criteria.



The proposed changes to the primary  standards included:

-------
                                    3



     1)  replacing TSP as the indicator for particulate matter with a new



indicator that includes only those particles with an aerodynamic diameter



smaller than or equal  to a nominal 10 urn (PMig);



     2)  changing the  level  of the 24-hour standard to a value to be



selected from a range  of 150 to 250 gg/m^ and replacing the deterministic



form of the standard with a statistical form that permits one expected



exceedance of the standard level  per year; and



     3)  changing the  level  and form of the annual standard to a value to



be selected from a range of 50 to 65 ug/m^, expected annual arithmetic mean.



     Given the precautionary nature of the Act, the Administrator stated



an inclination-to select the primary standards from the lower portions of the,



above ranges.  The proposal  notice (49 FR 10408) sets forth the rationale



for these and other proposed revisions of the particulate matter NAAQS



and background information related to the proposal.



     3.  Developments  Subsequent  to Proposal



     After the close of the public comment period on the proposed standards



provisions, the Clean  Air Scientific Advisory Committee (CASAC) met on



December 16-17, 1985 to review the proposal and to discuss the relevance



of certain new scientific studies on the health effects of particulate



matter that had emerged since the Committee completed its review of the



criteria document and  staff paper in January 1982.  Based on its preliminary



review of these new studies, the  Committee recommended that the Agency



prepare addenda to the criteria document and staff paper to evaluate the



relevant new studies and consider their potential implications for standard



setting.  The Agency announced its decision to prepare these addenda on



April 1, 1986 (51 FR. 11058).



     A preliminary draft of this  paper was reviewed by the CASAC in October 1986*



This final product incorporates the suggestions and recommendations of the

-------
                                    4
CASAC as well  as other  appropriate  comments  received on  the  initial  draft.
The CASAC closure memorandum (Lippmann,  1986)  is  reprinted in  Appendix C.
C.  Approach
        The approach in this paper  is  to  address  the newly available health
effects information in  the criteria document addendum  (CD addendum or CDA;
EPA, 1986a) in the context of those critical elements  which  the  staff believes
have implications for the proposed  revisions to the primary  particulate
matter standards.  Particular attention  is drawn  to judgments, related to the
proposed indicator for  the primary  standards (i.e., PM^Q), and the proposed
ranges of interest for  the level of the  primary standards.   Previous staff
conclusions and recommendations  related  to the secondary standards wi 11 not
be addressed here.
     Sections II and III review  important recent  scientific  and  technical
information relevant to standard setting. Section II  provides a brief
update of aspects of current and historical  air quality  information on
particulate matter to support discussions of the  standards.   Section III
addresses those essential elements  of  the health  effects information that
require re-examination  in light  of  the new information in the  CD addendum,
which include:
    1) respiratory tract deposition and  clearance of inhaled particles; and
    2) concentration/response relationships  for both acute and long-term
exposures to particulate matter  derived  from community epidemiological
studies.
    Drawing from the discussion  in  Sections  II and III,  Section  IV
identifies and assesses the factors the  staff  believes should  be considered
in selecting the particulate pollutant indicator  and level of  primary
standards.  Staff conclusions and recommendations on policy  alternatives
are updated and supplemented to  incorporate  the more recent  information.

-------
II.  AIR QUALITY CONSIDERATIONS

     More than any other criteria pollutant, "participate matter" represents

a broad class of chemically and physically diverse substances.  Their

principal common feature is existence as discrete particles in the condensed

(liquid or solid) phase ranging in size from molecular clusters of 0.005 urn

to coarse particles on the order of 100 urn.*  The major chemical  and

physical properties of particulate matter vary greatly with time, region,

meteorology, and source category.  It is to be expected, then, that the

effects of given quantities of particles on public health and welfare also

will vary.  This variable composition complicates the evaluation  of the"

applicability of specific particle health and welfare studies for establish-

ing national ambient air quality standards.  The 1982 staff paper ("SP,"

1982) (Section IV) summarized some key features of our understanding of

historical and current particulate matter composition to provide  perspective

for interpretation of-the effects studies derived-from the 1982 criteria

document ("CD," 1982b).  This section of the addendum updates the original

work in two areas:  1) an overview of recent measured and estimated PM^Q

concentrations and potential  exposures, and 2) a summary of historical

particle size relationships associated with six U.S. cities that  are the

subject of the most important new epidemiological studies.
*Where not otherwise specified, particle sizes reported in this  paper
 reflect aerodynamic equivalent diameter (AED).  A number of terms  (e.g.,
 fine, coarse, inhalable, thoracic, TSP) are used to describe various fractions
 of particulate matter.  Many of these terms are defined by the  instruments
 used for measurement.  The major particle indicators discussed  in  this
 paper are defined in Appendix D of the 1982 staff paper.

-------
                                      6
 A.  Current PM^g Concentrations
      Since the original  staff and  CASAC  recommendations  for  a  10 urn cut-
 point for the primary particulate  matter standards, a  number of different
 sampling devices capable of measuring this  fraction (termed  PM^g)  have
 been developed.  Several hundred PM^g instruments are  now  being operated
 in the field by state and local agencies, industries,  and  researchers.
 Data from these sources  are, however, fragmentary due  in part  to start-up
 and reporting limitations, and the available  results have  not  yet  been
 adequately screened.  Thus, it is  not yet possible to  provide  an adequate
• national assessment of PM^g concentrations.
      Some idea of PM^g levels and  composition  across the country can,
 however, be derived from later years  of  EPA's  "Inhalable Particle" (IP)
 network (Pace, 1986).  Beginning in 1982, the  39 sites in  this network with
 dichotomous samplers were retrofitted with  PM^g inlets.  These sites were
 chosen'to represent areas with the highest  particulate matter  concentrations
 in the original 163 site network.   As a  group, they have considerably
 higher TSP levels than most sites  in  EPA's  "SAROAD" data base.  Because  of
 the limited number and duration, however, it  is virtually  certain  that
 other locations in the nation will  record similarly high or  even higher
 concentrations.  The PM^g samplers came  on  line at various times in 1983 and
 1984 and were operated on a 1 in 6 day sampling schedule.  Thirty-eight
 of the sites provide useful data during  this  period, with  a  total  collection
 of 11 to 113 readings per site.
      With these limitations in mind,  Figure 2-1 presents the annual  and
 maximum 24-hour values from this network in 1983-84.   These  data suggest
 that both fine and coarse particles are  major  contributors to  PM^g mass
 across all sites, with a tendency  for sites with higher  concentrations to

-------
        MAX  24-HOUR  CONCENTRATIONS   (83-84)
                                 38 SITES IN EPA IP NETWORK
     200
                                                      FP
COARSE
ANNUAL   MEAN  PMIO   CONCENTRATIONS    (83-84)
m
  E
 \
  0)
 Z
 O
 H
 I-
 <
 d
 H-

 UJ
 O

 O
 U
                         38  SITES IN EPA IP NETWORK
                                                                        COARSE
         Figure 2-1.   PMio  Concentrations  from Dichotomous Samplers  In EPA IP Network,
         ordered by concentration,   a) Maximum 24-hour PMio  values with  associated fine
         mass.   Due to limited  sampling  frequency,  these data probably understate the
         actual maxima,  b)  Annual  means of PKjo  and fine particles.  Fine mass is a
         substantial  fraction of  PMjo mass particularly in eastern sites; coarse
         particles tend to  constitute a  large fraction at higher concentrations and
         at western sites.   (Pace,  1986)

-------
                                     8
have a higher coarse fraction.   This  tendency  holds for both  maximum 24-hour
and annual  data over all  sites  on the days with highest PM^g  concentration,
the fine fraction average about 60% of  PM^g mass.
     The data in Figure 2-1 suggest that  no more than eight sites  would exceed
the lower bounds of the ranges  for PM^g standards proposed in 1984.   As noted
above, however, these results are likely  to understate the extent  of higher
concentrations across the country.  To  provide some sense of  the  nature of
such concentrations as well as  potential  human exposures to them,  EPA
staff have developed an approach for  estimating the probability of exceeding
particular PM^g values using available  TSP measurements (Pace and  Frank,
1984).  The approach is based on a detailed examination of size fractionated
data (PMig, PMis, and TSP)  across the nation  (Pollack et al., 1985).  To
provide a best estimate of the  number of  areas that would exceed  particular
PM^g values, staff applied PM^g/TSP relationships associated  with  a  50%
probability of exceeding the specified  limits  to the national  TSP  data
set for the years 1982-84.   The results shown  in Table 2-1 represent the
estimated number of counties (and population  residing therein) that  would
exceed combined PM^g standards  set at the extreme upper and lower  bounds
of the proposed ranges.  For comparison purposes, the effect  of adding
a secondary annual TSP standard of 90 ug/m^ and the counties  exceeding the
current primary TSP standards are also  shown.  These estimates are highly
uncertain, but give some perspective  on the nature of current PM^g air
quality and potential exposures with  respect  to the proposed  standards.
More definitive data from actual PM^g monitoring will be available in
the near future.

-------
                                      9
     TABLE 2-1.  ESTIMATED COUNTIES EXCEEDING PROPOSED STANDARD LIMITS*
Standards No. Counties
(24 hr/annual , ug/m3) Exceeding Limit
Upper bound of P
(250/65)
Upper PM10 + TSP
Lower bound of P
(150/50)
Lower PM^Q + TSP
Current primary
(260/75)
MIQ ranges
secondary**
MIQ ranges
secondary**
TSP standards
36
73
173
176
155
Population in Counties
12 mil lion
-
60 mil lion
-
50 mil lion
 *Based on 1982-84 TSP data, counties with probability of exceeding standards
 (probex) _>. 0«5, 1980 census data.  Geographical  area exceeding limits may, in
 many cases, be much smaller than county size.  Accordingly, populations in
 the vicinity of such concentrations are lower than the total  county populations,
**90 ug/m3 annual arithmetic mean.
 8.  Historical Trends in Six Cities
      The draft criteria document addendum indicates that two of the more
 important recent publications on the effects of particulate matter are  derived
 from the Harvard "Six Cities" study.  To aid in the assessment  of these
 studies, EPA commissioned an examination of the relationships between  TSP
 and size fractionated particle mass measurements in these cities (Spengler
 et al., 1986).  The results, which span some seven years of size specific
 data, are useful both in examining trends and in permitting improved estimates
 of historical PM^Q levels from TSP measurements.  Details on the analysis,
 methodology, and relationship to health study sites are contained in a
 separate report (Spengler et al., 1986).
      The results from Steubenville in Figure 2-2 are illustrative.  Over
 the six year period of record, concentrations of all particle fractions

-------
                                      10
     100-


      90"


      80-


      7O-
4O"
3O '
2O-
io-
•AW \w
v V--

i 1 . . , 1 . . , 1
1979 198O
. "\Vy 'YJ \Vy
\ . \ •— '*N
a
1981 1982 1903 • 1981
1.2-
1.1-
i.o-
cu
H^ O.8-
in
^ 0.7-
cu
V 0.6-
Uj
O O.5-
^ 0.4'
ct
0.3-
0.2
0.1

-

1 ill 1
Jii I u 6 1 nuaiy/iiii 1

0 "VTvOuTTyPTin.
v n • T y v i T T . v r
T ' T




0.0 | ... | ...,...,...,

















i

•
'









i
i-

i

-


i r

T






                1979
1980
1981
1982
198
Figure 2-2.  Trends in Seasonal  Particle  Fractions  at  Steubenville (Spengler
et al., 1986).  a)  From 1979 on,  all  size fractions show  similar seasonal
trends with a general  decline in all  measures,   b)  The ratio of PMis/TSP as
measured by dichotomous sampler  is,  however  reasonably stable over the same
period.  The ratios of these fractions  at other six city  sites also do not
show clear trends,  but in some cases  the  decrease  in coarse particles
(> 2.5 pm) is more  pronounced than  that for  fine,  suggesting that the
historical ratios of PMio/TSP were  somewhat  lower.

-------
                                    11
generally declined as source emissions declined, with some suggestion of
an increase in 1984.  The PM^/TSP ratio remained reasonably stable through
this period.  This suggests that the PMnj/TSP ratio of U.5 measured in
1984 would be reasonably representative of the recent historical  past.
Size fraction ratios at the other five sites also showed little in the
way of trends, but the examination of trends in particle mass suggested
that the PM^g/TSP ratio in three cities (St. Louis, Watertown and Topeka)
were likely to have been somewhat lower in earlier years when TSP levels
were higher (Spengler et al., 1986).  Recent (1984) PMio/TSP ratios for these
cities are Portage, WI (0.64), Topeka, KS (0.46), -St.  Louis, MO (0.62),
Harriman, TN (0.66), and Watertown, MA (0.54).  The report notes that the.
PM^g and some other aerometric data for certain years were obtained by use of
Beta-Gauge measurement rather than gravimetric mass.  Based on an examination
of trends, no perceptible difference is noted between these two measures,
at least for determining longer-term averages.
     The ratios derived from annual averages do not necessarily apply to any
particular 24-hour period.  Data presented in the Spengler et al. (1986)
report (Figures IV-3 to IV-8) also include size fraction and ratio data
that encompass the 1979 and 1980 episode studies reported in Dockery et
al. (1982).  On various days during the three study periods, the PM^/TSP
ratios measured by dichotomous samplers ranged between 0.4 and 0.8.  Based
on the overall ratios among fractions in Steubenville (Tables V-5,6 in the
Spengler et al. report), the PMjQ/TSP ratios would be expected to be a
factor of about 0.8 to 0.9 of these PMi5/TSP ratios (see Appendix 8).

-------
                                    12

III.  CRITICAL ELEMENTS IN THE  REVIEW OF  THE  PRIMARY  STANDARDS
     This section summarizes  recent  information on particle  deposition
in the respiratory tract and  on concentration-response  relationships  from
community studies.  A comprehensive  discussion of these and  other  critical
elements, including mechanisms  of toxicity, effects of  concern,  and sensitive
populations, is contained in  Section V of the 1982 staff paper  (1982a).
The present summary provides  a  basis for  later discussions of the  implications
of the more recent studies for  selecting  the  particle indicator  and examining
concentration/response relationships.
A.   Mechanisms:   Particle Deposition
     The major relevant new information reviewed in the CD^adde.ndum concerning
mechanisms related to penetration and deposition of particles in the  respiratory
tract falls into the following  categories:  (1) extension of  experimental  data
on deposition and clearance of  large (> 10  urn) particles;  (2) assessment of
particle deposition during oronasal  breathing; and (3)  information on variations
in deposition and clearance for children  and  individuals with respiratory
illness, as well  as for altered breathing patterns.   Each of these areas  is
briefly discussed below.
     1.   Thoracic Deposition of Large (> 10  urn) Particles
     Figure 3-1 updates the range of available experimental  data on alveolar
and tracheobronchial  particle deposition  presented in the 1982 CD  (Figure 2,
CDA).  The recent experimental  deposition data on larger particles (> 10
urn) from three laboratories are represented as the individual points  shown.
The CD addendum notes that the  data  of Svartengren (1986) reflect  an  atypical
inhalation pattern; accordingly, less emphasis should be placed on those
data.  Nevertheless,  the major  thrust of  the new results, taken together,
is to substantiate the original extrapolation of the upper bound of the

-------
                                             13
o
er
u.
o
&
ui
O
0.9


0.8


0.7


0.6


0.5


0.4


0.3


0.2


0.1
      —r—i—i  i   i  i
       RANGE OF ALVEOLAR DEPOSITION.
       MOUTH BREATHING

       ESTIMATE OF ALVEOLAR DEPOSITION. NOSE BREATHING

       RANGE OF TRACHEOBRONCHIAL DEPOSITION.
       MOUTH BREATHING

 —— EXTRAPOLATION OF ABOVE TO POINT ( Q > PREDICTED
       BY MILLER « «!.. (19791

 o«
 OB

-O+ SVARTENGHEN (1986)
 OPEN SYMBOLS: TRACHEOBRONCHIAL DEPOSITION
 SOLID SYMBOLS: ALVEOLAR DEPOSITION
                                                1—I—I   I   I  I  I  lllllll
                0.2
        PHYSICAL DIAMETER,
                                            2.0   3.0  4.0 5.0
                                                                     10 121416 20
                                    AERODYNAMIC DIAMETER,
     Figure 3-1.  Regional deposition  of  monodisperse aerosols by  indicated
     particle diameter for mouthpiece  breathing (alveolar, tracheobronchial)  and
     nose breathing (alveolar)  (CDA, Figure 2).  The alveolar band  indicates
     the range of results found by  different investigators using different
     subjects and flow parameters for  alveolar deposition following mouth
     breathing.  Variability  is also expected following nasal inhalation.
     The tracheobronchial band  indicates  intersubject variability  in  deposition
     over the range of sizes  as measured  by Chan and Lippmann (1980).   Deposition
     is expressed as fraction of particles  entering the mouth (or  nose).  Also
     shown is an extrapolation  of the  upper bound of the TB curve  to  the  point
     predicted by Miller et al. (1979).   The extrapolation illustrates  the
     likely shape of the curve  in this size range but is uncertain.   However,
     the data of Emmett et al.  (1982), Heyder (1986), and Svartengren  (1986)
     tend to substantiate this  extrapolation.  In the Svartengren  (1986)
     studies, subjects took maximally  deep  inhalations at a flow of
     500 cm3 s1.

-------
                                    14
tracheobronchial  deposition curve  in  Figure  3-1  to  the  point  predicted by
Miller et al. (1979).   With the  exception  of the Svartengren  results, the
newer data are also reasonably consistent  with the  range  of alveolar
deposition illustrated; taken together,  however, the  added  points suggest
slightly higher alveolar deposition  for  larger particles  than did the
previous data.
    2.  Assessment of  Deposition During  Uronasal  Breathing
    The experimental  results depicted in Figure  3-1 were  obtained from
studies in which the subjects inhaled through a  mouthpiece.   Such results
tend to overstate particle penetration under more natural oronasal  breathing
conditions.  Swift and Proctor (1982) attempted  to  quantify this overstatement
and simulate  deposition under natural oronasal and  oral breathing for
ventilation rates corresponding  to light activity.  Based on  their results,
the authors predicted  that little  thoracic deposition would occur for
particles larger than  10 pro with natural  breathing  conditions'.  The
CD addendum points out that this conclusion  does not  appear to be
consistent with the information  available  in 1982;  moreover,  the analysis
itself has been superceded by improved simulations  using  more recent
experimental  data (Miller et al.,  1984,  1986).
    As indicated in the CD addendum,  these latter analyses  provide
significantly improved fits of the deposition data  and  extend both the
particle size range and ventilation  patterns simulated.   These results
show that the Swift and Proctor  simulation and related  predictions
understate thoracic deposition for particles larger than  10 urn under all
conditions and understate deposition  of  particles larger  than 6 urn for
individuals who habitually breathe oronasally (mouth  breathers) at  light
activity levels.  The  CD addendum  concludes  that the more recent

-------
                                    15





deposition data shown in Figure 3-1 "are relevant to examining the



potential of particles to penetrate to the lower respiratory tract and



pose a potentially increased risk.  Increased risk may be due to increased



localized dose or to the exceedingly long half-times for clearance of



larger particles (Gerrity et al., 1983)" (p. 2-18).



     3.  Variations in Deposition and Clearance for Children and Other Groups



     Experimental deposition data discussed above are restricted to



adults.  The epiderniological evidence, however, indicates increased risk



to young children exposed to ambient particulate matter.  Phalen et al.



(1985) have modeled tracheobronchial deposition of particles.  Although



not accounting for prior extrathoracic removal, the results suggest a



tendency towards increased particle deposition efficiencies for the range



of particle sizes modeled (0.5 to 10 urn) in smaller (younyer) individuals



(CDA, Figure 4).  Attempts to quantify age-dependent differences in



deposition will' require improved information on differences in children



related to alveolar and extrathoracic deposition, deposition over the



entire breathing cycle, and clearance patterns.



     Subject characteristics, disease states, and other factors can also



alter the deposition and clearance of particles from more typically



observed ranges.  Recent work by Heyder (1982) examined biological  variability



of particle deposition in adults and found very small  intrasubject



variability mainly due to daily variations in breathing cycle and flow



rate.  The more extensive variability of deposition rate between subjects.



breathing the same aerosol was found to be less determined by the morphological



constitution of the respiratory tract than by individual ventilatory



patterns.

-------
                                    16
     Several  new studies  on  clearance mechanisms further  support  previous
conclusions in the 1982 CD and  Staff Paper  regarding  the  consequences
of retarded mucous transport when  impaired  by  disease or  other  insults on
residence times of inhaled particles, long-term clearance times of
insoluble particle from the  alveolar region, and the  regional deposition
of inhaled particles (Svartengren  et al.,  1986; Levandowski  et  al.,  1985,
Garrard et al., 1985;  Bailey et  al., 1982;  Bohning et al.,  1982;  Philipson
et al., 1985; Gerrity  et  al., 1983).
B.   Concentration-Response  Information
     As discussed in the  1982 Staff  Paper,  associations between air
pollution and both acute  and chronic effects have been demonstrated  in
many countries and different population  groups, supported by controlled
laboratory exposures of animals  and  humans  to  various components  of
particulate matter (SP, section  V.A,B,C; Appendix B). • Assessing
the precise level of particulate pollution  associated with  observed
effects on health, however,  has  many problems.  Suspended particulate
matter is not a uniquely  defined entity.   The  comprehensive physical  and
chemical  characteristics  are not only hard  to  measure and relate  to
health effects, but vary  with monitoring device, geography,  and time.
This variability increases the  uncertainty  of  any extrapolations  from one
set of circumstances to another, and greatly limits the utility of laboratory
studies of single substances for quantifying health risks.
     Epidemiological  studies can provide strong evidence  for the  existence
of pollutant effects,  but are more limited  for identifying  accurate  effects
levels for specific pollutants  or  pollutant classes.   Among  the more
important limitations  of  epidemiology as discussed in the 1982  CD are:
1)  inadequate and inconsistent  measurement of the exposure  burden of

-------
                                     17
individuals; 2) variability in the measurement of health endpoints (e.g.,
lung function, hospital admissions, frequency of symptoms) and in the
sensitivity of populations studied; 3) failure, especially in cross-sectional
studies, to control fully for confounding or covarying factors, such as
cigarette smoking and socioeconomic status; 4) difficulty in distinguishing
particles from other pollutants; and 5) inability to establish a causal
relationship, or negate one, based only on statistical associations.
     Recognizing these limitations, epidemiological  studies still
form the principal basis for developing concentration response assessments
for particulate pollution..  The key concentration-response information
derived from the 1982 CD is discussed in the 1982 Staff Paper and in the
1984 PM proposal notice (49 FR 10408).  The following review summarizes the
recent epidemiological studies cited by the CD addendum as providing the
most reliable exposure-response information on mortality and morbidity
effects associated with acute and'chronic exposures  to particulate
matter.  Other recent studies that may provide reasonable evidence of
exposure-response relationships or qualitative insights are summarized in
tabular form in Appendix A.  Further assessment of the epidemiological
studies as applied to selecting alternative levels for air quality standards
is presented in Section IV.
     1.   Acute Exposures
     a)   Mortality
     Table 3-1 summarizes recent epidemiological studies providing the most
useful concentration response information for assessing acute exposures to
particulate matter.  The initial entry reflects the  newer CD addendum conclusions
regarding reanalyses of daily London mortality in relation to short-term
(24-hour) exposures to PM and S02 (Mazumdar et al.,  1982; Ostro, 1984 Shumway

-------
                     TABLE 3-1.  SUMMARY OF RECENT (1982-86) EPIDEMIOLOGICAL  STUDIES  PROVIDING MOST USEFUL
                                CONCENTRATION/RESPONSE INFORMATION FOR ACUTE  PARTICLE EXPOSURES
Jbserved
 ffects
            Observed Concentration Range
Time
                                                                                 Study
increases in
daily mortality
in metropolitan
.ondon
ihort-term
•eductions in
 ung function
in 330 school
:hildren,
iteubenville,
iH (330 total)
 hort-term
 eduction in
 ung function
 n 179 school
 hildren in
 he
 etherlands
 Ijmond)
1958-1972
winters
<500 BS*
>500
       24-hr averages
Four
separate
study
periods of
3 weeks
following
pollution
"episodes"
in 1978-1980
Before,
during, af-
ter pollu-
tion epi-
sode Nov.
1984-Feb.
1985
1) 420 TSP        280
2) 270 TSP        460
3) 220 TSP        170
4) 160 TSP        190
  (max 24 averages for
   "alert" or "sham"
   episode)
 200-250 TSP      200-250
 and RSP
 (D5n < 3.5
  urn)
      24-hr averages
                                                                                                               Shumway et al.
                                                                                                               1983,  Schwartz
                                                                                                               and Marcus,
                                                                                                               1986
Recently published studies reinforce 1982 CD,      Mazumdar et
SP conclusions regarding likelihood of increased   1982, 1983;
mortality at 500 to 1000 ug/m3 for BS and SOo,     Ostro 1984
with no clearly defined threshold for BS in the
range of 150 to 500 ug/m^.  Year-by-year analyses
indicate significant BS-mortality associations
in most to all winters.  Nature of relationships
vary significantly with model.  Suggestion of
surrogate behavior.
Recent unpublished analyses confirm major
findings of the published studies with advanced
statistical techniques accounting for auto-
correlation and temperature effects.  Schwartz
and Marcus findings suggest significant
association for BS at lowest levels (<100
BS), but not for S02 below about 500 ug/m .

First 3 episodes: small (2%-3%) but significant
reversible declines in FVC up to 2-3 weeks after
peak.  Less consistent results for FEV.  No
significant effects after 4th "sham" episode.
Baseline measurements for 1st, 4th taken on days
with high pollution.  Linear regression of pooled
data for 330 children indicate significantly more
negative slopes in functions vs. TSP and SO;? across
ranges (10-270 ug/m^, 0-280 pg/m^, respectively).
Higher response in some children suggests sensitive
subgroup.

Small (3-5%) reversible declines in several        Dassen et
measures of airway function (FVC, FEVj, MEF)       al., 1986
during episode and 5 days later.  No effect
after 26 days or shortly after a day when TSP,
RSP and S02 levels all averaged 100-150 ug/m .
Separate sub-groups of children tested on each
day.  Peak TSP levels possibly understated.
al,
                                                               Dockery et al,
                                                               1982
    00
 Britsh Smoke   s  is a pseudo-mass indicator related to small  particle (size less than a nominal  4.5 \im)  darkness
 CD, pp. 1-88 to 1-90).

-------
                                     19

et al., 1983, Schwartz and Marcus, 1986).  Among the important unresolved

issues raised regarding these London data are identification of a practical

threshold for PM-mortality associations, separating effects of PM and S02,

the changes in coefficients obtained with different subsets of data sets

and models, the effects of unmeasured variables such as other outdoor

pollutants, demographic changes over time and indoor air pollution, and the

appropriate statistical methods to account for long-term seasonal trends in

mortality (Roth et al., 1986).  When considering the available evidence,

the CD Addendum finds that:
                                                         o

     "the following conclusions appear-to be warranted based on the earlier
     criteria 'review (U.S. EPA, 1982a) and present evaluation of newly
     available analyses of the London mortality experience:  (1)  Markedly
     increased mortality occurred, mainly among the elderly and chronically
     ill, in association with BS and S02 concentrations above 1000 pg/m ,
     especially during episodes when such pollutant elevations occurred for
     several consecutive days; (2)  During such episodes coincident high
     humidity or fog was also likely important, possibly by providing
     conditions leading to formation of ^$04 or other acidic aerosols;
     (3)  Increased risk of mortality is associated with exposure to 6S and
     S02 levels; i'n the'range of 500 to 1000 ug/m , -for S02 most clearly at
     concentrations   700 ug/m3; and (4) Convincing evidence indicates that
     relatively small but statistically significant increases in the risk of
     mortality exist at 8S (but not SOo) levels below 500 ug/m , with no
     indications of any specific threshold level having been demonstrated at
     lower concentrations of BS (e.g., at _<_ 150 ug/rn3).  However, precise
     quantitative specification of the lower PM levels associated with mortality
     is not possible, nor can one rule out potential contributions of other
     possible confounding variables at these low PM levels" (CDA, p. 3-9).

     Analyses of deviations in daily mortality from 15-day moving means for

each of the 14 winters individually in two publications found that the

mortality-BS relationship was significant in most to all of the years

(Ostro, 1984, Mazumdar et al., 1982).  In separate regressions involving a

linear model of S02 and BS jointly, a linear model of BS alone, and a quadratic

analysis of BS alone, Mazumdar et al. found that the BS-mortality relationship

was significant in 7, 14, and 13 winters respectively.  In a linear regression

of year by year data for days when BS was below 150 ug/m3, Ostro found

-------
                                     20
significant regression coefficients i.n 7 of 12 winters with a substantial
number of days with BS < 150 ug/m3, including 6 of the most recent 7 winters.
Both Mazumdar et al. and Ostro found a tendency for the regression coefficients
"to increase in later years in smoke only regressions; a trend is not apparent
in the joint smoke-S02 regression.
     From a methodological perspective, the recent report by Shumway et al.
(1983) represents a significant addition to the London mortality analyses.
Their investigations developed a complex time series structure that accounted
for long-term trends in mortality as well  as auto-correlation in the data.
No attempt was made to separate the effects of BS and S0£ and the effects
of the two pollutants were found to be nearly identical.  Total, cardiovascular,
and respiratory mortality all increased with BS (or S02) concentration across
the range of concentrations with no discernible threshold.  Slopes decreased
at higher concentrations similar to findings of Mazumdar et al. and Ostro
for smoke aTone.  Temperature was also a-significant predictor with the
greatest impact when both current and 2-day lag temperature was used.
Based on analyses of alternative time-lagged models, the authors concluded
that (1) the mechanism by which these factors influence mortality has pollution
acting strongly and instantaneously, and (2) the largest fraction of variance
in daily mortality could be attributed to cyclical patterns in temperature
and pollution that had 7-21 day periods.  Taken together, these conclusions
suggest that although relatively small elevations of pollution may influence
daily mortality, larger effects are more likely when the elevated concentrations
occur as part of a multi-day cycle than after short duration episodes.
     In order to delineate further the degree of reliance that can be
placed on the more recent analyses outlined above (Ostro, 1984; Mazumda-r
et al., 1982; Shumway et al., 1983), EPA conducted a reanalysis of the 14

-------
                                     21
winter London mortality data set (Schwartz and Marcus, 1986).*  Schwartz
and Marcus controlled for the effects of autocorrelation in separate time
series regressions of daily mortality that incorporated various combinations
of temperature, humidity, SU2, and BS.  They found that both "crude" (or
absolute) daily mortality as well as deviations in daily mortality from a
15-day moving mean were positively and significantly correlated with increases
in BS.  Significant linear correlations of crude mortality with BS were
observed for 13 of 14 winters (deviations were significant in all  14 winters),
including 6 of the last 7 winters, during which the maximum daily  BS levels
were well below 500 gg/m3.  The overall effect of accounting for autocorrelation
was to increase the strength of the associations.  When compared to the
previous published analyses, the magnitude of the regression coefficients
for each year were comparable to those found by Mazumdar et al, (1982)  and
Ostro (1984).  As in the earlier studies, Schwartz and Marcus found a
tendency for the overall regression coefficients to increase in the later
years with lower concentrations.  This is also evidenced in an  apparently
concave concentration-response relationships when the data for  all  winters
were grouped and plotted.  When only days with BS < 200 ug/m3 were included in
the regression, however, the regression coefficients were more  stable,  with
no clear tendency to increase with time.  In essence, the BS/mortality
relationship across and within individual winters appears to be concave,
with no apparent threshold at various BS levels tested in this  and earlier
analyses (500 gg/m3, 250 ug/m3, 200 ug/m3, 150 gg/m3).  The Schwartz and
*This paper and a summary memorandum (Marcus and Schwartz, 1986),  are
reprinted in full as Appendix A to the Criteria Document Addendum.   Although
not published, the paper was presented to the CASAC and the public  for
review at the October 15-16, 1986 meeting.  Copies were made available to
the public at the time of the meeting.  Subsequently, EPA received  and
considered comments on this study from industry and environmental  groups
and from members of the scientific community.

-------
                                     22



Marcus results thus  reinforce the  findings  of  Ostro  (1984)  regarding the



absence of an apparent threshold and  both Ostro  and  Mazumdar  et  al., (1982)



with respect to the  magnitude of the  regression  coefficients.   The  suggestion



by Mazumdar of a "quadratic"  concentration-response  relationship with a



threshold at 300 ug/m3 is  not supported  by  the reanalyses.



     Schwartz and Marcus also examined further the suggestion  raised by



Mazumdar et al. (1982) that the effects  of  smoke are separable from  those



of S02«  In regressions involving  both pollutants, the  col linearity  between



the two tended to deflate  the apparent significance  of  both.   However, the



overall results for  all years combined and  for those individual  years with



lower correlations between BS and  SOg (r <  0.9)  show that  the  mortality



effects of BS remain significant and  relatively  large even  when  S02  is



included in the model, while the inclusion  of  BS in  the model  reduces the



S02 coefficients to  insignificant  values.   Thus, while  an  independent



effect of S02 cannot be excluded,  particularly at higher concentrations,



these analyses add weight  to previous suggestions that  BS  is  significantly



correlated with mortality  independent of S02.



     Based on the various  studies  discussed above, it is currently  not



possible to derive an appropriately quantitative model  for  a  gravimetric



particulate matter/daily mortality relationship  across  the  range of  concentrations



observed in London or to specify a concentration below  which  no  association



remains.  It is even more  problematic to apply such  relations  to locations



other than London.  However,  the results of Mazumdar et al. (1982)  provide



some perspective on  the relative magnitude  of  any effects  during various



winters.  Using a linear model with coefficients comparable to those found



in other studies, these investigators found the  mean effects  of  smoke

-------
                                     23



accounted for on the order of 4 to 9% of daily mortality in London during



the early winters and about 2 to 3% in later winters.



     Other recent studies discussed in the CD addendum and Appendix B of this



document examined pollutant/mortality relationships in more contemporary



atmospheres in New York City, Pittsburgh, and Athens, Greece.   The Ozkaynak



et al. (1986) reanalysis of 14 years of N.Y.C. data (1963-1976) found



significant associations between excess daily mortality and PM, S02 and



temperature using time-series methods to control  for autocorrelation.



Differences in the rate of change of S02 and PM indicators during the study



period allowed estimation of their separate effects.  In joint regression.



analysis across all years, PM indicators (coefficient of haze  and visibility



extinction coefficient) together accounted for significantly greater excess



mortality than did S02.  Although their findings  are considered preliminary



for risk assessment purposes, these results are of particular  interest



given the'possibility that fairly contemporaneous particulate  air pollution



in a U.S. urban area could be contributing to mortality (CDA p. 3-10 to



3-12).



     The work of Mazumdar and Sussman (1983) in -Pittsburgh and that of



Hatzakis et al. (1986) in Athens, however, found  conflicting results.  The



first found significant association between particulate matter and excess



deaths in Pittsburgh, but no effect of S02, while the Athens study found an



association with SU2 but not with smoke measurements.  The CD' addendum



points out that limitations in both studies with  respect to measuring



particulate matter as well as methodological difficulties prevent meaningful



conclusions from these studies with respect to the effects of  particulate



matter and S02.

-------
                                     24



     b)  Morbidity



     Previous conclusions regarding concentration-response relationships for



morbidity effects of daily PM/S02  exposures  were based primarily on studies



of bronchitic subjects in London during the  1950's  through early 1970's.



Results more relevant to contemporary U.S. conditions  are presented by



Dockery et al. (1982) and summarized in Table 3-1 along with a comparable



recent study from the Netherlands  (Oassen  et al., 1986).



     The CD addendum concludes  that the repeated measurements of lung



function by Dockery et al. (1982)  showed statistically significant but



physiologically small and apparently reversible group  mean declines in



Forced Vital Capacity (FVC) and Forced Expiratory Volume at 0.75 seconds



(FEVg.75) associated with short-term increases in PM and S02 air pollution



(p. 3-16).  The small, reversible  decrements appear to persist for up to



3-4 weeks after episodic exposures to these  pollutants.



     The data were analyzed for each episode separately and also for pooled



results for all four study periods.  Taken individually, statistically



significant declines in FVC (2-3%) were seen consistently during the first



three study periods while FEV declines were  significant only for the second



and third.  This suggests that  significant effects  on  lung function occurred



in these children for those episodes with  maximum 24-hour TSP levels of



220 to 422 ug/rn3.  The possibility of effects below 220 ^g/rn^ can not be



dismissed, but the absence of effects on either FVC or FEV in the fourth



study period (Fall 1980) suggests  that the peak TSP level  measured during



that period (160 ug/m^ 24-hour  maximum) might be considered as a practical



no effects level.



     The interpretation of the  episode results is,  however, complicated by the



frequent moderate peaks in pollution that  occurred  at  various times through

-------
                                     25



each episode.  TSP levels in excess of 150 ug/m^ occurred during three of



the baseline measurements in the 1978 episode, potentially diminishing the



apparent significance of any declines as measured following the subsequent



alert.  Similarly, TSP levels during some "baseline" periods in the Fall 1980



study approached or exceeded those during the rest of the study.  Moreover,



the presence of intermediate peaks following an alert can cloud interpretation



of the time to recovery from the functional depressions.  In these respects,



the Fall 1979 study and to a lesser extent the Spring 1980 study, both with



relatively low pollution during baseline measurements, offer the clearest



results.  The Spring study, however, had intermediate TSP peaks that at a



second site reached about 240 ug/m^ (Spengler et al., 1986) at about



the time of the second follow-up measurement.  Since this suggests exposures



at or above those following the "sham," no firm conclusions regarding the



effects of the-first peak can be -drawn from this follow-up.



     In contrast to the episode studies, the pooled  regression analysis



assumed that functional response resulted from the previous day pollution



levels across the range of measured concentrations.   The authors concluded



that because a significantly greater number of subjects had negative regression



coefficients for both lung function measures vs. TSP and S$2> ^uny function



might be altered across the full range of TSP and SO^ levels.  As the



authors note, however, a non-linear threshold model  cannot be precluded,



especially given the absence of pulmonary function effects in the Fall 1980



study.  The CD addendum also notes that the regression analysis apparently



included a large number of subjects with data only from the first study



with the highest pollution and largest FVC changes.   This might have unduly



affected the regression results.  Ancillary regression data showing a



significant negative slope for the testing days in the Spring 1980 study

-------
                                     26
(Ferris et al., 1983) suggest,  however,  that  excluding  the  1978 data would
not change the conclusions.
     Although the group mean changes  in  lung  function during  individual
episodes were small  (generally  2  to 4%), the  pooled  data  suggests  the
possibility that some children  showed enhanced  responses.   The  CO  addendum
notes that the predicted changes  in FVC  per unit TSP for  the  upper quartile
of children was -0.386 ml/pg/m^,  or 5 times higher than that  for the group
mean.  The distribution of individual  regression slopes (Figure 3  in Dockery
et al.) indicates that approximately  5%  of the  children had negative slopes
of 1 ml/ug/m^ TSP or more.  Some  of the  larger  negative slopes  are likely
to be due to chance  or non-pollution  factors  such as reduced  effort in
follow-up functional measurements.  Some of those children, however, may
have been substantially more sensitive to pollution  than  the  group mean.
     A study of episodic exposures of children  to particulate matter and
S02 conducted in the Netherlands  by Dassen et al . (1986)  produced  results
similar to the episode component  of Uockery et  al .   Pulmonary function
values measured during an air pollution  episode in which  24-hour average
measurements of TSP, RSP* and S02 at  a 6 station network  all  reached a
range of 200-250 ug/m^, were significantly  lower (3-5%) than  baseline
values measured 1-2  months earlier for the same subgroup  of children.  Lung
function parameters  that showed significant declines on the second day of
the episode included FVC and FEV, as  well as  measures of  small  airway
function (e.g., maximum mid-expiratory flow,  maximum flow at  50% of vital
capacity).  Declines from baseline were  observed 16  days  after  the episode
in a different subset of children, but not after 25  days  in yet a  third
subgroup.  Shortly before the last set of measurements, 24-hour average  TSP,
*Respirable Suspended Particles,  reportedly  D^Q £ 3.5  by  cyclone  sampler.

-------
                                     27


RSP and S02 jointly reached 100-150 ug/m , suggesting that these levels


were not associated with observable functional effects (CDA, p. 3-17).


     The authors note that TSP values may be somewhat low, but partially


overlapping measurements at a local network suggest they were unlikely to be


underestimated by more than 10 to 20%.  Overall, collocated RSP measurements


were 0.79 to 0.94 of TSP, with network averages actually exceeding TSP


during the episode.  The authors indicate rain, north winds, and snow may


have accounted for the apparent low levels of coarse particles during this


period.


     In comparison with the Steubenville episodes, the pattern of pollution
                                                                t

is much less problematic.  Baseline and intermediate concentrations, with


one exception, were low.  Thus, the finding of a similar time course of


response (two to three weeks for recovery) provides additional support for


an extended depression in function following a single episode.  The absolute


magnitude of functional  changes appears somewhat greater in the Dutch


episode, but much of the difference is due to the fact that the latter


results were adjusted for lung function growth over the course> of the study


while the Steubenville results were not.  A confounding aspect of the Dutch


study is the use of different subgroups during follow-up measurements.


     The findings of these recent episode studies are consistent with those


of other, more qualitative, community studies identified in the 1982 staff


paper reporting pulmonary function changes in children and adults exposed to


high short-term levels of particles alone (Lebowitz et al., 1974) or in


combination with SOg (Van der Lende et al., 1975; Stebbings et al. 1979;


Saric et al., 1981).


     Other recent studies on'the relationship between short-term exposures


to particles and acute morbidity effects are characterized in the CD

-------
                                     28


the CD addendum as allowing no definitive  interpretations  at  this  time


(Mazumdar and Sussman,  1983; Perry  et  al.,  1983;  Bates  and Sizto,  1983,


1985).


     2.  Long-Term Exposures


     Recent cross sectional studies of the  association  between  long-term


particulate matter concentrations and  mortality are  summarized  in  Appendix


B.  While these may be  of qualitative  interest in  supplementing prior


analyses, at present there is no basis by which to derive  exposure-response


information given their unstable results,  inadequate exposure characterization,


and internal inconsistencies.
                                                   i

     A number of newly  available studies have examined  the long-term effects


of exposures to particles (with and without  $62)  on  respiratory mechanics,


symptoms, and illness (Table A-3).   The CD  addendum  identifies  only  the


Ware et al. (1986) paper (summarized in Table 3-2) as possibly  providing


results by which to derive quantitative conclusions  concerning  exposure-effect


relationships.on morbidity.  The remainder  are either too  preliminary to


interpret definitively  (van der Lende  et al., 1986)  or  are subject to


significant uncertainties regarding the nature of  any gradients in PM


exposure levels (e.g.,  Pengelly et  al., 1985; CEC, 1983).


       Ware et al. (1986) found significant, positive associations between


some respiratory symptoms and illness  in children  and concentrations of


TSP, and the sulfate fraction of TSP (TSCty), and  between one  symptom and


SO;?.  However, an examination of somewhat  smaller  pollution variance within


two of the cities did not produce the  expected gradient in response, with


the exception of illness before age two.   Pulmonary  function  parameters


were not associated with pollutant  concentrations within the  observed


ranges.  The authors note that the  between-city results may represent

-------
     TABLE 3-2.  SUMMARY OF EPIOEMIOLOGICAL'STUDY PROVIDING MOST USEFUL
CONCENTRATION/RESPONSE INFORMATION FOR LONG-TERM PARTICLE EXPOSURES (1982-86)
Jbserved
ffects Population
'ossible 10,000
increased rates 6-9 year olds
if cough, in 6 U.S.
>ronchitis, lower cities
espiratory
il Iness
to differernce
in lung function
"


City_
Portage
Topeka
Watertown
Kingston/
Harriman
St. Louis
Steubenville


Within City
Gradients:
Time
76-79
77-80
74-77
75-78

75-78
76-79




City Mean Pollution •
(in ug/m3)
TSP SO? TSO/i Comments
39
63
•46
62

94
114




12
3
18
25

68
61




5.4
5.4
8.4
9.5

11
19




Well des
cross-set
ongoing 1
Symptom,
on parenl
of elevai
vs. fall
quality i
(1 per y«
cohorts 1
adjusted
             Steubenville
              -Valley
              -Ridge
             St. Louis
              -Carondelet
              -Remainder
133
 95

116
 73
80
54

98
38
                                                                                         Study
         gned.  Preliminary     Ware et
         tional results from    1986
         ongitudinal  study.
         illness data based
         al recall, suggestion
         ed response in spring
         surveys.  Nine air
         egions with 3 cohorts
         ar) generated 27
         or analysis.  Effects
         for 1) age,  sex,
parental education and smoking
and 2) random city, region, year
variability.  S02 associated
significantly only with cough.
Within city  results  not consistent
with inter-city findings.  Pollution
gradient maintained when adjusted
by city specific PMifl ratios
(Spengler et al., 1986).
                                                                                                 al
ro

-------
                                     3U

differences unrelated to pollution  exposure  such as cultural  factors,

persistent differences between  cities  in  illness or reporting rates,  or

better recall  of illnesses  in more  polluted  cities.  These  cities  tended to

be visited in  the spring, while some  of the  cities with  lower pollution

were visited in the fall  when past  winter illnesses were more remote.

     The CD addendum concludes  that the Ware et al. (1986)  study:

     "provides evidence of  respiratory symptoms in children  being  associated
     with particulate matter exposures in contemporary U.S.  cities without
     evident threshold across a range  of  TSP levels from 30  to 150 ug/m^
     with more marked effects notable  in  the 60-150 ug/ro^ range in comparison
     to lower  levels...  The medical  significance of the observed  increase
     in symptoms unaccompanied  by decrements in lung function remains  to be
     fully evaluated but is of  likely  health concern.  Caution is  warranted,
     however,  in using these findings  for risk assessment purposes in  view
     of the lack of significant associations for the same variables when
     assessed  from data within  individual  cities included in  the Ware  et
     al. (1986) study" (p.  3-49).

     The CO addendum further notes  that:

     "the reported stronger associations  between TSL)4 levels  and other
     measures  of ambient air FP concentrations are highly suggestive  of
     possible  associations  between  health effects observed .in the  Ware
     et al. (1986) study and exposure  to  small particles in  contemporary
     U.S.  atmospheres....   However,  full  interpretation of  the strength
     and significance of these  findings is difficult at this  point, in
     light of  further follow-up of  these  children still being in progress
     and the expectation that longitudinal analyses will later be  carried
     out which will relate  health data to more extensive aerometric data
     (including such data collected in later years)" (p. 3-37).

     A series  of studies by Ostro and  coworkers (Ostro,  1983, 1987; Hausman

et al., 1984)  provide qualitative indication of morbidity in  adults in

U.S. cities with particulate matter concentrations overlapping those

found in the six cities study.   The series of investigations  encompassed

both annual and shorter term (2 week)  exposures.  The most  recent  work

(Ostro, 1987;  Hausman et al., 1984) examined Health Interview Survey  (HIS)

data and yielded "associations between  particulate pollution  and increases

in restricted  activity days (RAD),  respiratory related RAD,  and work  loss

days, as well  as other, even more generalized health indicators.  The  most

-------
                                     31


consistently significant correlations were for effects and average exposure

occurring 2 to 4 week previously (2 period lag).  This somewhat puzzling

result raises some questions about the mechanism of action.  It is, however,

consistent with the kind of delayed response suggested by the Steubenvilie •

and Netherlands episode studies.  Additional questions raised in the CD

addendum include the nature of the HIS data, the statistical modeling used,

and the estimates of fine particle concentrations based on airport visibility

data.  Only limited pollution data are provided in the published reports,

but in 1976 annual TSP levels ranged between 40 to 133 ug/m^ and the mean

estimated FP level for these cities was 22 ug/m^.  Other issues include the
                     i
degree for which the fixed effects model  accounts for city specific effects,

the role of ozone and other pollutants not included in the regressions and

consistency among other examinations of the HIS data (Portney and Mull any,

1986).  The results from further analyses that address many of these

issues are expected in the near future (Ostro, 1986).  At present, however,

the CD addendum concludes that these analyses:

     "have found consistent associations  between PM and morbidity measures
     for adults tha't are reasonably consistent between and within contemporary
     American cities.  As such, the results tend to reinforce the plausibility
     of the Ware et al. (1986) findings of associations between morbidity
     measures in children and PM concentrations found in contemporaneous
     American urban air sheds.  However,  the Ostro analyses do not allow
     for the estimation of quantitative relationships between morbidity
     effects and more usual 24-hr or annual average direct gravimetric
     measures of participate matter air pollution (e.g., TSP, PM^Q, etc.)"
     (p. 3-40).

-------
                                     32



IV.  FACTORS  TO BE  CONSinERED  IN SELECTING  PRIMARY STANDARDS  FOR  PARTICLES



     This section,  drawing  upon the  previous  summary  of  newly available



scientific information,  enumerates key factors  that should  be considered



by the Administrator in  making decisions  on the proposed revisions  to



the primary standards  for participate matter.   The staff conclusions  and



recommendations on  the most appropriate  policy  options  presented  update



and supplement those made in the 1982 staff assessment.   Where the  original



conclusions and recommendations and  supporting  rationale are  unchanged



by the newly available information,  they  are  summarized  without restating



the supporting discussions.  Particular  emphasis is placed  on aspects of
            t


the new information that amend or  revise  the  original  assessment.  The key



standard components discussed are  the pollutant indicator,  averaging  time,



and levels for the  primary  standards.



A.  Pollutant Indicator



     Based on the re-evaluation of  available  scientific  information,  the



staff finds that the following conclusions  reached  in the 1982 assessment



remain valjd:



     1) A separate  general  particulate matter standard  (as  opposed  to a



combination standard for particulate matter and SO;?)  remains  a reasonable



public health policy choice.



     2) Given current scientific knowledge  and  uncertainties, a size-specific



(rather than chemical-specific) indicator should be used.



     3)  Health risks posed by inhaled particles are  influenced both by the



penetration and deposition  of particles  in  the  various  regions of the



respiratory tract,  and by the biological  responses  to these deposited



materials.

-------
                                     33
     4)  The risks of adverse health effects associated with deposition of
ambient fine and coarse particles in the thorax (tracheobronchial  and
alveolar regions of the respiratory tract) are markedly greater than for
deposition in the extrathoracic (head) region.  Maximum particle penetration
to the thoracic region occurs during oronasal  or mouth breathing.
     5)  The risks of adverse health effects from extrathoracic deposition
of general ambient particulate matter are sufficiently low that particles
which deposit only in that region can safely be excluded from the  standard
indicator.
     6)  The size-specific indicator for primary standards should  represent
those particles capable of penetrating to the  thoracic region, including
both the tracheobronchial  and alveolar regions.
     Considering these conclusions in light of data on air quality composition,
respiratory tract deposition and health effects, the need to provide protection
for sensitive individuals  who may breathe by month and/or oronasally, and '
the similar convention on  particles penetrating the thoracic region recently
adopted by the International Standards Organization (ISO, 1981), the staff,  •
recommended that the size-specific indicator include particles less than or
equal to a nominal 10 urn "cut point."*  This indicator, referred to as
"thoracic particles" in the 1982 staff paper,  has been termed "PMio"
regulatory purposes.
*The more precise term is 50% cut point or 50% diameter (059).   This
is the aerodynamic particle diameter for which the efficiency of particle
collection is 50%.  Larger particles are collected with substantially
lower efficiency and smaller particles with greater (up to 100%) efficiency.
In practical usage, acceptable ambient samplers with this cut point provide a
reliable estimate of the total mass of suspended particulate matter of aerodynamic
size less than or "equal  to 10 urn.  See additional  discussion regarding the
Federal Reference Method in the notice of proposed revisions (49 FR 10408).

-------
                                     34
       Figure 4-1 summarizes  many  of  the more  relevant  aspects  of the recent
particle deposition studies contained in the CD addendum  and  discussed in
Section III.A of this paper.   The  figure represents  thoracic  deposition of
particles under nasal and  oronasal  breathing as estimated by  Miller et al.
(1986).  Superimposed on the  figure are the estimates of  the  band of thoracic
deposition by Swift and Proctor  (1982).  The latter  analysis  has  been used
to support recommendations for an  alternative  particle  size indicator, which
would have a "D0" of 10 urn and a 050  of approximately 6 urn.   The  figure
shows that such an indicator  would omit the non-trivial fraction  of thoracic
deposition contributed by  particles larger than 10 pin for all  breathing'
conditions and would also  understate  deposition of particles  larger than
6 urn for "mouth" breathers.  •
     The sampler effectiveness curves for two  prototype PM^g  inlets also
plotted in Figure 4-1 illustrate the  generally conservative nature of the
PM^o indicator"when compared  to these data.'  The  samplers reach'100% efficiency
for particles of 7 ym and  smaller, while the respiratory  tract  deposition
data do not quite reach 50%  (in effectiveness  terms).   Practical, satnplers
could not, of course, realistically match this performance.   Thus, a
better way to compare the  deposition  data with the sampler effectiveness
is to scale the data such  that the maximum deposition point represents "1," or
100%.  Viewed from this perspective,  the maximum  point  for the  distribution
illustrated in Figure 4-1  generally lies between  3 and  5  urn and the 50%
point tends to be in the vicinity  of  10 urn.  In this relative  sense, the
PM^Q indicator, follows the "inlet" portion of  respiratory tract penetration
pattern, but substantially overcollects fine particles  smaller  than 3 to 5
um relative to lung deposition.  The  figure indicates that most fine mass
is not deposited in the respiratory tract, while  PMjg samplers  would collect

-------
                                        35
  0.5
  0.4
  0.3
ui
O
  0.2
ac
O
  0.1
   I  I  II UK    I  I  i I  II U.
    — — — MOUTH BREATHERS
           NORMAL AUGMENTERS —
    0 -L
I   I  I  I IIlll
I   I  I  I I  I H-
                                            B
                                                                 T   I  I  I IIU.
                                                     Inlet 2
                                                                   Inlet 1
I	I
     1.0
           10.0             100 1.0

                  AERODYNAMIC DIAMETER,
                               10.0
                             100
   Figure 4-1.  Estimates of thoracic deposition  of particles  between  1  and
   15 urn by Miller et  al. (1986) for normal  augmenters  (solid  lines) and  mouth
   breathers  (broken lines) are.shown for minute  ventilation  (Ve)  exceeding
   the switch point of 35 L m1nl(A) and for  lower Vp  (B).   Normal  augmenters
   are individuals who normal
   airflow when V. exceeds about
                   /  MI ivi i w i  i wrv«» i  » 0 \ u / o  i *w i HI a i
                  ly use  oronasal breathing to augment
                  out 35  L min ,  while mouth breather
                                      respiratory
                                     refers to
   those individuals who  habitually breathe oronasal ly  (Niinimaa et al.,  1981).
   The shaded area  (B)  is a composite of the computed bands  of  thoracic
   deposition of particles less  than 8  urn by Swift  and  Proctor  (1982)  for Ve
   of approximately 24.6  and 15  L min.1  Also  plotted are  the sampler
   effectiveness curves for two  representative PM^g Inlets.

-------
                                     36
100% of this fraction.   As  stated  in  the  1982  staff  paper,  given  the larger
surface area in the fine mode  as well  as  other concerns,  the  greater weight
given fine vs. coarse particles by  a  10 gni  indicator  remains  prudent and
appropriate.
     In summary, the staff  assessment  of  more  recent  information  on respiratory
tract deposition contained  in  the  criteria  document  addendum  reinforces  the
conclusions reached in  the  original staff assessment  in  1982.   In particular,
the staff finds that:
     1) the recent data do  not provide support for an  indicator that excludes
all particles greater than  1U  gm;
     2) the analysis used to specify  an alternative  indicator  with a nominal
size cut of 6 urn (Swift and Proctor,  1982)  can significantly  understate
thoracic deposition of  particles larger than 6 gm under  natural breathing
conditions;
     3) the PM}Q indicator  appears  somewhat less conservative  than previously
thought with respect to large  (> 10 pm) particle deposition under conditions
of natural mouthbreathing.   Nevertheless, this indicator  is generally
conservative for tracheobronchial  deposition;  and
     4) recent information  suggesting  enhanced tracheobronchial particle
deposition for children relative to adults  provides  an additional  reason
for an indicator that includes particles  capable of  such  penetration (Section
III).
     Given these considerations and the earlier conclusions, the  staff
reaffirms its recommendation to replace TSP as the particle indicator for
the primary standards with  a new indicator  that includes  only  those particles
less than a nominal  10  gm (PM^o).
       In the previous  assessment,  the staff also made recommendations with
respect to the shape of sampler effectiveness  curves.  Analysis of the

-------
                                     37

influence of outpoint and effectiveness curves under various simulated

ambient conditions have tended to show that (1) the 059 of the inlet has

the major influence, and (2) for a fixed cutpoint, the mass collected does

not vary greatly with the shape of the effectiveness curve (Rodes et al.,

1981; Van der Meulen, 1986).  For this reason, and because of the difficulty

in precisely matching the most recent respiratory tract deposition estimates,

the staff concludes that, for regulatory purposes, the effectiveness criteria

developed based on the 1982 CO remain acceptable.

B.  Level of the Standards

    1.  General Considerations

    This treatment of the implications of more recent studies follows the

framework and maintains the underlying philosphy of the 1982 staff paper

as discussed therein (SP, pp. 83-89).  The following general considerations

are drawn from that more complete discussion.

    The major scientific basis for selecting PM standards that have an adequate

margin of safety remains community epidemiological research, with mechanistic

support from toxicological and controlled human investigations.  The
                                    «
limitations of epidemiological stu'dies for quantitative evaluation of the

health risks of particulate matter under current U.S. conditions are

detailed in the 1982 criteria document (EPA, 1982b) and its addendum (EPA,

1986) as well as in the 1982 PM staff paper (pp. 83-86).  Such studies,

while representing real world conditions, can  only provide associations

between a complex pollutant mix measured at specific locations and times

and a particular set of observable health points.  Difficulties in

conducting and interpreting epidemiological studies limit the reliance

that can be placed on the results of any single study.  Furthermore, even

the best studies often provide no clear evidence of population "thresholds."

Thus the approach of identifying specific "lowest demonstrated effects" levels

-------
                                     38



for current U.S.  exposures  and  adding  margin  of safety  considerations is



less appropriate  in this  case.   Instead, the  approach followed  in  the 1982



staff paper and here is to  assess  the  nature  of health  risks  along a continuum



of exposure using the full  range of  available information.   It  follows



that, although the scientific literature provides substantial information



on the potential  health risks associated with various mixes  and levels of



particles, selection of any general  particulate standard  remains  largely a



public health policy judgment.



     Because particulate  matter is a pollutant class with variable composition,



and none of the published studies  have used the proposed  PM^Q indicator, the



range of aerosol  composition and size  indices must be considered  in using the



relevant epidemiological  studies for developing standards.   For example, in



order to translate the re-suits  of  historical  British studies  into  terms



useful for setting U.S. standards, general relationships  between  British



smoke, readings and particle mass units (i.e., PM^Q), estimated  in  the 1982



staff paper are used here.   Those  relationships were based on available



calibration data  from the study periods, incorporating  reasonable  assumptions
                         «


concerning pollution composition,  relative role of particles, and  the



nature of U.S. vs. British  exposure  regimes (SP, pp. 7-13, 96-100).



Conversions are also made between  TSP  concentrations measured in the U.S.



studies and corresponding PM^Q  levels, in some cases using more detailed



site-specific data.



     The following sections present  a  brief staff assessment  of the



concentration/response relationships suggested by the most significant



epidemiological studies in  the  CD  addendum.   This assessment  supplements



the quantitative  information in the  1982 staff paper and  indicates

-------
                                     39


how these studies may be applied in developing ranges for final  decision-making


on standards for particulate matter, as indicated by PM^Q.  The presentation


also outlines a qualitative assessment of the key factors that affect the


margins of safety (risk) associated with the concentration-response relationships


derived from these studies, as translated to contemporary U.S. exposures.


The margins of safety associated with the levels of interest for PM^Q


derived from the quantitative studies should also be evaluated with respect to


any potential effects that may reasonably be anticipated from qualitative


human and animal health studies summarized in the 1982 staff paper.  Short-


and long term exposure are discussed separately.


    2.  Short-term Exposures


    a.  Derivation of Range of Interest from Epidemiological Studies


        i) Concentration-Response Relationships


           The 1982 CD indicates that the epidemiological studies most


useful for developing quantitative conclusions regarding the effects of


short-term exposures to particulate matter include a series of studies and


analysis of daily mortality in London (Martin and Bradley, 1960; Martin,


1964; Ware et al., 1981; Mazumdar, et al., 1981) and studies of bronchitis


patients, also in London (Lawther et al., 1970).


       The assessment of the earlier London mortality studies in the 1982 CO


concluded that 1) clear increases in excess daily mortality occur at BS

                                    o
and S02 levels at or above 1000 ug/m , and 2) some indication of likely


increase in excess mortality exists in the range of 500 to 1000 ug/m^ BS


and S02, with greatest certainty of increases occurring when both pollutants


exceed 750 ug/m^ (CuA, Table 1).  These estimates represent judgments with


respect to the most scientifically reliable "demonstrated effects likely


levels" for daily smoke (and SOg) and mortality at least in the context of

historical London pollution exposures.

-------
                                     40
     Because of the severity  of the health endpoints  in these  studies,  and
the need to provide an  adequate margin of safety in standard setting, the 1982
CD and staff paper also examined these studies to determine whether  the data
support the possibility of health  risks  at lower BS levels.  This  assessment
concluded that data from the  earlier  London  studies do not provide clear
evidence of absolute population thresholds,  and suggest instead  a  continuum
of response, with both  the likelihood and extent of any effects  occurring
decreasing with concentration.  Thus, based  on these  earlier studies,
effects were judged to  be "possible"  at  levels below  500  to 1000 ug/m3
smoke down to a practical  lower bound of 150 ug/m3  (as BS) derived from
the Martin and Bradley  (1960) study.  The analysis  stressed that because
evidence is less clear, the nature and extent of risks at lower  levels  are
much more uncertain.
     The more recent analyses of London  mortality during  the winters between
1958 and 1972 cited,in  the CD addendum include Mazumdar et al.  (1982),
Ostro (1984), Shumway et al.  (1983),  and Schwartz and Marcus (1986).
In essence, these analyses add to  the evidence for the possibility that
participate pollution accounted for a small  but statistically  significant
portion of daily mortality at levels  extending well below 500  ug/m3  8S
(24-hour avg.), with no discernible threshold.  Considering the  findings of
these more recent studies, the staff  amends  its earlier assessment of the
London mortality data (SP, pp. 89-9b) with the following  conclusions:
       1) The finding of significant  associations between BS and mortality in
the majority of the 14  winters by  different  investigators in published
(Mazumdar et al., 1982; 1983; Ostro,  1984) and unpublished (Schwartz and
Marcus, 1986) analyses  using  several  approaches strengthens the  plausibility
of the associations. The findings of significant associations in  later

-------
                                     41


"non-episodic" years when particle composition and levels began to approach


U.S. conditions is of particular significance.


       2) The finding in some analyses of a trend towards increased regression


coefficients with decreased concentration and the concave shape apparent


across the range of mortality-BS data as plotted by Schwartz and Marcus


(See Figure 4-2) raises questions regarding whether the statistical association


reflects a causal relationship.  The possibility that smoke may be acting


as a "surrogate" for unmeasured factor(s) at lower BS levels, as suggested


by Mazumdar et al. (1982) cannot be precluded.  Non-pollution factors such
                                                                         •

as weather, demographic shifts and indoor pollution exposures have been


advanced as possible alternatives (Roth et al., 1986).  To date, however,


smoke/mortality relationships have retained (or even increased in) significance


when meteorological factors (temperature and humidity) are included and the


year-to-year consistency of association, particularly for BS < 200 ug/m^,


argue against the observed effect being explained by changing indoor-heating


practices in London or by long-term demographic shifts (CDA, p. 3-7).


Moreover, as Mazumdar et al. points out, BS might be a surrogate for other


particulate components rather than some as yet unanalyzed non-pollution


variable.  Schwartz and Marcus (1986) suggest that the decreasing response


with higher pollution may result from the effect of higher pollution in


earlier winters being blunted by public awareness (and hence reduced exposure)


or by a tendency for the most susceptible individuals to succumb on the


earliest day of very high pollution in a multi-day episode.  Some of the


curvilinear shape between BS and mortality might also be due to the non-linear


relationship between BS and gravimetric mass at lower BS levels.  In a


qualitative sense, adjusting for this relationship would make the corresponding


particle mass/response relationship more linear.

-------
                                       42
                       Daily Mortality - British Smoka (All Day* Undar 600 po/m* BS)
                               London Wmtara 1968-69 to 1971-72
  360
  326
1300

1
  276
  260
            60       100      160     200     260      300      36O      4OO      46O     6OO


                                    BrHWi Smoka. ng/m>
    Figure 4-2.   Mean  daily  mortality  vs.  mean British Smoke  (BS)  for days with
    BS <  500  ug/m3  during  14 London  winters (1958-72) (Schwartz  and  Marcus, 1986).
    Each  point represents  the mean  "crude" daily mortality and BS  for 20 adjacent
    values of BS.   Grouping  data  points  in this fashion (c.f. Ware et al.,
    1981; SP, Figure 6-2)  reduces scatter  and reveals an apparently  concave
    relationship  extending to the lowest observed BS levels with a decreasing
    slope at  higher concentrations.  This  is consistent with  the findings of
    higher regression  slopes in years  with lower average concentrations
    (Mazumdar et  al.,  1982;  Ostro,  1984).   The concave shape  may,  however,
    be an artifact.  Some  possible explanations include a non-linear relationship
    between BS and  gravimetric mass, reduced population exposures  during
    publicized high pollution episodes and correlations of BS with unmeasured
    non-pollution variables  that  are causally related to mortality (see  text).

-------
                                     43



       3) The approach of Shumway et al., (1983) is an important addition



to the literature that, while reinforcing the above findings,  suggest



additional complexity in the potential  concentration response  function,



particularly with respect to the influence of temperature and  nature of



temporal patterns in pollution.  Temperature appears to exert  a same-day



positive effect, with higher daily mortality associated with an increase



in temperature.  The lag result in Shumway et al., however,  also suggests



that reduced mortality is likely during a cold spell after a dip in  temperature.



This phenomenon could be explained by increased outdoor-related activities



on warmer days in the winter.  The Schwartz and Marcus analyses subsumes



any such lag effect of temperature in the autoregressive model.  Further



analyses are desirable to examine possible interactions or non-linear



responses involving temperature, humidity, fog, and wind.speed.



       4) While it is still  difficult to separate the effects  of S02 and 6S



on mortality, the preliminary findings  of Schwartz and Marcus  (1986)



support the suggestion (Mazumdar et al., 1982) that at lower S02 values



mortality effects may be associated with particulate matter  alone.



       5) Taken together, the analyses  to date do not permit identification of



a clear "no effects" level.   The lower  bound derived from earlier analyses



is no longer appropriate.  The individual regression analyses,  however, provide



some suggestion that effects do not always achieve significance in the  last two



winters when mean smoke levels were below 75 to 100 ug/m-^ (Ostro, 1984;



Schwartz and Marcus, 1986).



       As the earlier assessment noted, the London data—even  in more



recent winters—have inherent limitations when applied to assessing



effects in U.S. atmospheres.  The pollution composition, meteorological



patterns, indoor sources, population characteristics, and other factors in

-------
                                     44


London may have been uniquely  responsible  for  the  observed  results.  Thus,


the findings from recent  reanalyses  of  daily mortality  during 14 years in


New York City (1963-1976)  (Ozkaynak  and Spengler,  1985, Ozkaynak et al.,


1986) are of particular interest.  The  results,  although preliminary,


showing associations between mortality  and  particle  concentrations (indicated


by coefficient of haze, or CoHs, and atmospheric visibility readings) add


to the evidence for a more general association between  elevated particulate


matter levels and increased mortality.   This recent  work reinforces earlier


qualitative findings of PM/mortality associations  in New York City (Schimmel


and Murawski, 1976; Schimmel,  1978).


       The 1982 CD evaluation  of the Lawther et  al.  (1958,  1970) studies


concluded that a worsening of  health status of chronic  bronchitic patients


could occur on days wi,th  BS _>  250-500 gg/m3 and  S02  > 500-600 gg/m3.  The


1982 CD also noted that associations between pollution  and  health status


persisted at lower levels  in selected,  more sensitive individuals, although


over the vigorous objection of the lead investigator (Lawther, 1986).  Better


evidence for effects on morbidity at lower  concentrations  is  provided by


the two recent studies of  U.S. and Dutch children  exposed during pollution


episodes with elevated 24-hour TSP and  S02  levels  (Dockery  et al., 1982;


Oassen et al., 1986).


       The U.S. study found evidence of small  but  significant (2-3%)


reductions in lung function (FVC, FEV^yij)  in  Steubenville  following periods


when 24-hour TSP levels reached 220  to  420  ug/m3 and S02 reached 280 to 45b


gg/m3, but no significant  changes following 24-hour  TSP and S02 maxima of 160


and 190 gg/m3, respectively.   The Dutch study  found  comparable functional


reductions during and following an episode when  concentrations of TSP, RSP,

                                                o
and S02 were each in the  range of 200 to 250 gg/m  (based on  six monitoring

-------
                                     45



sites) for 2 to 4 days, with no significant reduction shortly after a more



modest 24-hour pollution increase when levels of all  3 pollutants averaged



100 to 150 ug/m3.



       Taken together, these studies suggest that functional  declines associated



with episodic exposures occur rapidly and persist for up to 2 to 3 weeks



before recovery, with a tendency for larger declines  to occur following



episodes with higher concentrations of smaller size particles..  This is



illustrated in Figure 4-2, which compares the Dutch findings  with the



Steubenville episode (Fall 1979)° that has the most comparable air quality



patterns (see Section III.B).  In both studies, functional  measurements



show a substantial decline, as measured a day or two  into the episode, that



persists for 16 to 18 days.  Given the lack of decline in the Steubenville



"sham" (2 weeks after baseline) and the fact only two test  days  (episode



and 1st follow-up) showed declines in the Netherlands, it seems  unlikely



that lack of interest in follow-up tests could account for  the pollution



related results.



     Comparison of the magnitude of response between  two different investigations



with children of overlapping but different ages should be viewed with caution.



Although the results may suggest slightly larger functional changes during



the Dutch episode, it is not clear whether any differences  would be



significant.  With both TSP and RSP levels at 200 to  250 Mg/m3,  it is



reasonable to assume intermediate to small particle indicators (PM^5 or PM^j)



levels were in the same range in Ijmond.  Based on size specific measurements



of Steubenville during the Fall 1979 episode (Spengler et al., 1986) maximum



concentrations of small particles were somewhat lower in Steubenville.



Applying factors appropriate for that episode (Section II), peak PM^g



levels were on the order of 150 to 170 ug/m3.  An earlier Steubenville

-------
                                      46
  Fill 1979
  Steubenville
                                                 Ijmond   t   =
                                                                     'Unadjusted for lung
                                                                      function growth for
                                                                      comparison with
                                                                      Steubenville
                                   1 week          2 weeks


                    Time Relative to Onset of Pollution Episode
Episode/
Alert
                                                               3 weeks
                                                                             4 weeks
Figure 4-3.  Mean change in FVC compared to baseline  for children in relation
to occurrence of pollution episodes in Steubenville,  Ohio (Dockery et al.,  1982)
and the Ijmond area of the Netherlands (Dassen et  al.,  1986).  Inserts:  Air
quality during each study period.  Steubenville  study:   Fall, 1979 episode,
184 3rd and 4th grade children with 69 tested during  alert,  all  tested
during follow up.  Netherlands study:  Winter, 1985 episode, 179 children
aged 7 to 11 years, with each follow up reflecting a  different group; FVC
adjusted for growth (light triangles).  Arrows show results  unadjusted for
growth (Brunkreff, 1986) for direct comparison with Steubenville results,
which are also not adjusted.  The patterns in response  for the two studies
show a remarkable similarity.  The maximum unadjusted mean changes for the
Netherlands episode are comparable to slightly larger than Steubenville
(~1 t.o 2%).  Although maximum TSP levels are similar  for each episode (see
inserts), based on corollary measurements of PM^  (Spengler  et al., 1986)
in Steubenville and RSP in Ijmond, concentrations  of  smaller sized particles
(as PMi$, PM^j, or PM3.5) were higher in Ijmond.   SO^ peaks  were", however,
higher in Steubenville.

-------
                                       47



  episode  (Fall  1978) with levels of small particles potentially approaching



  those of the Dutch episode  (Fall, 1978) found comparable to larger unadjusted



  maximal  declines  in FVC  ( ~ 50 ml).



       Although  it  is difficult to separate the effects of particles from



  S02, peak  levels  of SOg were higher in the 1979 Steubenville study (455



  ug/m3),  and lower in both the Dutch (200 to 250 ug/m3) and Fall 1978 (280



  ug/m3) study that had comparable to larger changes in FVC.  Other pollutants



  possibly associated with functional changes (03, N02) were unlikely to



  confound these  studies.  The effects of seasonal patterns temperature or



.  other meteorological factors cannot' be ruled out, but neither study found



  any significant correspondence between lung function and temperature.



       Other important aspects of these two studies are as follows:



       1.  It appears reasonable to expect that short-term changes in lung



  function in children following acute exposures to particulate matter is, in



  most cases, a more sensitive response than premature mortality or worsening



  of bronchi tic  symptoms.



       2.  Fairly contemporary atmospheric conditions were studied and,



  particularly in the case of the Steubenville study, particle composition is



  fairly representative of contemporary U.S. cities, significantly increasing



  •the applicability of the results to current standard setting.



      3.   The observed lung function declines beneath baseline never exceeded



  3 to 5%  on average, and recovery apparently began after 2-3 weeks.  It is



  difficult  to assess the significance of such reductions.  Dockery et al.



  note that  they  might be associated with aggravation of respiratory symptoms



  in children with  pre-existing illness.  Long-term examination of Steubenville



  children suggest  higher rates of respiratory illnesses and symptoms compared



  to other U.S. cities with lower PM levels, but no evidence for any persistent



  reductions in lung function (Ware et al., 198b).

-------
                                     48
     The extent to which  some  children may  be  considered  to  be  "responders"
has not yet been formally examined.   Oockery et al.,  however, also show
that the upper quartile (25%)  of  children included  in  their  pooled analysis
of lung function vs.  TSP  had individual  regression  coefficients  of FVC and
FEVg.75 5 and 17 times  the median,  respectively,  suggesting  a correspondingly
greater than average  decline in lung  function  across  the  range  of pollution
levels.  As noted in  Section III.B.,  a smaller subset  representing the
upper 5% of these children showed even more substantial negative regression
coefficients (for FVC,  £ -1 ml/gg/m3  TSP).  Assuming  a linear response with
no threshold across the range  of  TSP  concentrations observed in  the regression
study (11 to 272 ug/m3),  this  group would have a  predicted decline in FVC on
the order of 10 to 15%.  Such  calculations  almost certainly  overstate
the percentage of potentially  sensitive  children  because  some or all  of the
larger responses may  be due to a  random  distribution  of results.  (Section
III.B.).  Nevertheless, this assessment  suggests  that  functional  changes of
potential concern—even on a transient basis—might occur in some small
sensitive subgroups of  children.
     4.  Although the data suggest that  decrements  in  lung function may
occur during or immediately following a  single day  of  high particulate matter
levels, it is not clear whether or not multi-day  episodes are required
to produce more prolonged (2 to 3 weeks) decrements.   Results from controlled
human and animal toxicologies! studies provide support for mechanisms by
which short or longer term functional declines could  result  from particle
exposures (Table 5-2, 1982 SP).   Little  evidence  exists,  however, that would
support prolonged declines from single short-term SOg  exposures  at these
concentrations.
     5.  Although questions can be raised regarding potential variability
in lung function testing  throughout the  study  period  (especially given the

-------
                                      49



youth of the subjects), it appears that state-of-the-art measurement



procedures were used as well as appropriate controls for inter-observer



bias and extreme or missing values.



     6.  The results of these more recent studies are consistent with



earlier more qualitative acute studies in Pittsburgh (Stebbings et al., 1975),



Tuscon (Lebowitz et al., 1974), and the Netherlands (van der Lende et al.,



1975).



    ii)  Translation to PMip Indicator



         Table 4-1 summarizes the updated staff assessment of the more



recent and earlier available quantitative short-term epidemiological  studies.



Following the approach in the 1982 staff paper, the assessment incorporates



available data and assumptions necessary to express results obtained using



different particle indicators in terms of the recommended particle indicator.



     The "effects likely" row in -Table 4-1, based on the 1982 CD,



reflects the previous staff assessment and underlying rationale (SP, pp.



96-100).  As discussed therein and above, effects are possible at



concentrations below those consensus "effects likely levels."  In light



of the assessment of the more recent London mortality studies outlined above,



no lower bound smoke concentration is indicated in the "effects possible"



column.  Due to the lack of a clear threshold in these studies, the uncertainty



in translating these results to contemporary U.S. atmospheres, and the



availability of more recent U.S. data involving potentially more sensitive



effects the staff has chosen not to attempt to derive any lower bound PMu



concentration from the London data.  The approach used previously to bound



8S/PM10 relationships does not apply to lower values (< 100 ug/m3 BS), and it



is also unclear to what extent BS readings were calibrated to mass measurements



in the later years when smoke levels had declined appreciably.  Thus the



lowest pollutant levels of interest in the remaining short-term studies are

-------
                         TABLE 4-1.  UPDATED STAFF ASSESSMENT OF SHORT-TERM EPIOEMIOLOGICAL  STUDIES
Effects/Study
Effects Likely
Effects Possible
No Significant
Effects Noted
o
Measured British Smoke Levels (as ug/m )
(24-hr, avg.)
Daily Mortality
in London*
1000
1
?
-
Aggravation of
Bronchitis2
250*-500*
< 250*
-
Combined
Range
250-500
<250
-
Measured TSP Levels (ug/m3)
(24-hr, avg.)
Small, reversible declines
in lung function in children3*
-
220*-4203
200-2504
125*4-1603
Equivalent PMiQ
Levels (ug/m3)
Combined
Range5
350-600
140-350
<125
Mndicates levels used for upper and lower bound of range.

Various analyses of daily mortality encompassing the London  winter  of  1958-59,  14  winters  from  1958-72,  in  aggregate
 and individually.  Early winters dominated by high smoke and S02 from  coal  combustion with frequent  fogs.   From  1982 CD:
 Martin and Bradley (1960); Ware et al., (1981); Mazumdar et  al.  (1981).   From  1986 CD Addendum:   Mazumdar et  al.  (1982);
 Ostro (1984); Shurnway et al., (1983); Schwartz and Marcus  (1986).   Later  studies show association across entire  range  of
 smoke, with no clear delineation of "likely" effects or threshold of  response  possible.
2Study of symptoms reported by bronchitis patients in London, mid-50's  to  early  70's; Lawther  et  al.  (1970).
3Study of pollution "episodes" in Steubenville, Ohio, 1978-80; Dockery  et  al.  (1982).
4Study of 1985 pollution episode in Ijmond, The Netherlands;  Dassen  et  al. (1986).
 a) Conversion of BS readings to PM^Q levels:  Assumes for  London conditions  and BS readings in  the range 100-500  ug/m  ,
 BS 
-------
                                     51





250 ug/m3 (BS) and SOU gg/nr S02 (based on the earlier bronchitic studies)



and 200 to 420 (TSP) and 190 to 455 ug/m3 (S02) (based on the recent



studies of lung function in children).  The recent studies provide some



suggestion of "no observed effects" levels with TSP concentrations of 100 to 160



ug/m3.  The relative importance of S02 in these studies cannot be specified,



but collectively the data suggest a greater role for particles.  Thus the



conservative assumption (for particles) is made that the response might have



occurred without substantial amounts of S02 present.



     Conversion of the British data to PM^g equivalents is particularly uncertain,



and the approach is discussed in the earl'ier assessment (pp. 98-101).  The



original upper bound of a range of interest for a 24-hour standard derived



from the Lawther study was 350 Mg/m3 as PM1Q (SP, p. 97-99).  Because this



level  contained little or no margin of safety, staff and CASAC recommended



that consideration of standard levels begin at lower concentrations.  Accordingly,



the Administrator, considering this advice', as well as other factors, proposed



250 pg/m3 as the upper bound of the range of levels for a possible 24-hour



standard (49 FR 10408).  Thus, the upper bound for the range of interest is



250 ug/m3, as PM1(J.  The translation of the Steubenville and Netherlands



results to PMio is summarized in Table 4-1.  Based on these results, the lowest



PMlO level of interest derived from the short-term studies can be reduced to



140 ug/m^, although the original lower bound of 150 pg/m^ is within the range



of uncertainty of the conversion.  A level of 140 ug/m3 contains a large



margin of safety against exposures clearly associated with the more serious



effects of particulate matter and is at the lower end where reversible,



physiological responses of uncertain health significance may be observed.



However, the original lower-bound recommended by staff and CASAC also contains



a substantial margin of safety.

-------
                                     52
     b) Additional  Factors  to  be Considered in Evaluating Margins  of Safety
        and Risks - Short-term Exposure?
     The 1982 staff paper  identified  a number of factors to  be  considered
in developing a standard with  a margin of  safety.   In applying  the results
of the more recent  studies  to  determine  the margin  of safety for  sensitive
populations provided by alternative PM^Q standards  in the above range,  the
following additional factors should be considered:
     (i)  Aerosol Composition
     1.  As noted above, the likelihood  of high ozone levels during the
U.S. and Dutch episodes seems  low.  Where  high photochemical  smog  levels
are present, the observed  effects of  ozone on lung  function  (e.g., McDonnell
et al., 1983) suggest the  possibility of interactive responses  not accounted
for by these or the British studies.
     2.  When particle components differ substantially  from  those  in the
communities studied, risk  will  vary.  The  variability of composition
(e.g., relative fraction of sulfate,  nitrate, secondary organics,  carbonaceous
material, and coarse particles) is high.   Accordingly,  the risks  associated
with PM^g will vary among  U.S.  cities.
     (i i)  Exposure
           Although the assessment of the  mortality studies  suggest any risk
of premature mortality to  sensitive individuals may be  small  at lower
concentrations, the number  of  people  exposed to lower concentrations is
substantially larger than  the  number  exposed to higher  levels.  Table 2-1
shows that at present, the  total U.S. population living in counties with
PM1Q levels in excess of 250 ug/m3 is on the order  of the size  of  the London
population.  The number in  counties in excess of 150 ug/m3 is estimated at
six times larger.  The increased number  of sensitive individuals exposed
increases the risk  that some effects  will  occur in  the total  population
exposed.

-------
                                     b3



     Relative exposures and indoor/outdoor pollution relationships



are an important consideration in interpreting the British studies; these



•are discussed elsewhere (SP, p. 101).  With respect to the more recent



morbidity studies, for comparable outdoor concentrations, the overall  exposures



to maximum 24-hour outdoor pollution in Steubenville and in the Netherlands



was likely as high as typically occurs in contemporary U.S. exposure situations



during the fall through spring seasons.  Summertime exposures, would,



however, tend to be greater in many areas.



     (iii)  Risks For Other Sensitive Groups. Effects Not Evaluated



     Consistent with evidence from toxicalogical, controlled human and.



quantitative epidemiological data, the studies used to derive ranges  of



interest identify a number of groups and effects  as particularly susceptible



to ambient particles:  (1) premature mortality in very sensitive individuals



with chronic respiratory and cardiovascular diseases, individuals with



influenza, and'the elderly, (2) aggravation of"disease in bronchitic patients,



and (3) lung function declines in children.



     While other groups may be affected by ambient particle exposures,



such as asthmatics or even younger children, the  previous assessment found



no data to support the existence of significant effects below the suggested



range  (SP, pp. 102-103). The most significant new information in this



regard is the finding of restricted activity in adults associated with



earlier particle exposures of 2 week or longer durations (Ostro, 1987).  At



present, the results cannot be interpreted as demonstrating such effects



occur at 24-hour levels below the range of interest.



     Additional short-term effects of particulate matter sugyested by



qualitative evidence, such as altered respiratory clearance, possibly



resulting in infections, are identified in the 1982 staff paper (p. 103).

-------
                                     54
     3.  Long-Term Exposures
     a)  Derivation of Range of  Interest.from  Epidemiological  Studies
     Earlier cross sectional and  longitudinal  studies  useful  in establishing
ranges of interest for long-term  (annual)  PM^o standards  are  identified in
and discussed in the 1982 staff paper (pp.  57-63;  103-107)  and CD.   Of the
newly available studies,  the CO addendum cites the Ware et  al. (1986)
cross-sectional study of  children  in  6 U.S.  cities as  providing potentially
useful results for examining quantitative  relationships.
     In interpreting the  Ware et  al.  (1986)  study  the  CO  addendum concludes
that there is evidence of respiratory symptoms in  children  associated with
particulate matter exposures without  apparent  threshold across the  range of
measured TSP levels (CDA, p. 5-6).  As for  all  cross sectional studies, however,
these results—though adjusted for  a  number  of confounding  factors—
should be viewed with caution. A particular concern is the apparent
absence of the expected gradient  in response for-results  within cities.
The lack of within city effects does  not necessarily negate the results
among cities.  Possible explanations  include movement  of  the  population
throughout the area (reducing the  within city  gradient),  the  presence of a
lesser gradient in smaller sized  particles,  or a tendency for  hyperresponders
to move to cleaner areas  of the city  (consistent with  the negative  and
significant within city gradient  for  wheeze).   Results of a separate
series of studies of long- and intermediate-term (2 to 6  weeks) exposures
indicate consistent associations  between respiratory-related  restrictions
in adults and PM gradients within,  as well  as  among, a number  of U.S.
cities (Hausman et al., 1984; Ostro,  1983,  1987).  While  these results cannot
be used to estimate quantitative  relationships between morbidity effects
and PMio» the CD addendum indicates they do  provide qualitative support

-------
                                     55

for the possibility of within-city effects related to comparable U.S.

exposure levels.  Nevertheless, until  further results are available,

the within-city anomaly of the Ware et al. results as well  as. other

uncertainties (e.g., parental  recall  in spring vs. fall)  noted in Section

III.B. caution against interpreting the results as demonstrating "effects

likely" levels.  Considering the assessment in the CD addendum, however,

the six city results do suggest the possibility of effects,  at least in the

more polluted areas.

     In deriving a range of possible  effects levels from  this study, it

is useful to examine the gradient in  PM^Q terms, based on applying the

ratios developed by Spengler et al. (1986) to the data.  Because the

staff previously recommended an expected annual mean PMio standard and

because Ware et al. (1986) found that  long-term (life-time)  TSP exposures

were also significantly associated with respiratory effects, the PMiu

levels are estimated in terms  of multi-year averages.  Figure 4-3 plots

the relationship between long-term averages in frequency  of  cough and

estimated PM^Q levels across the six  cities.  This is the same effect plotted

against annual TSP levels in the Ware  et al. paper (Figure  5, COA), originally

chosen because it was most consistently associated with TSP  levels across

cities.

     As illustrated in Figure  4-3, three "cleaner" cities—Portage, Watertown,

and Topeka—consistently had the lowest frequencies of respiratory illnesses

and symptoms.  The highest symptom prevalence rates were  consistently found

in Kingston/Harriman, St. Louis and Steubenville.  Based  on  this very

qualitative break-down, the staff concludes that the most convincing evidence

for the possibility of effects is for  the latter three cities, with long-term
                                          o
average TSP levels between 60 and 114  ug/nr and corresponding PM^Q values

between 40 and 60 pg/m3.

-------
                                         56
  120 ••
o
o
o
  100 •
UJ
3  80
o
u

u.
o

UJ
u


I  60 t

                        P—I
   20
      10
                    20
30
40
50
60
              ESTIMATED  LONG  TERM PM10 CONCENTRATION  
-------
                                     57
     Table 4-2 amends the previous staff assessment of the the most useful


long-term epidemiological data to reflect this newer information; particulate


matter levels are expressed in both the original  and converted PM^Q units.


The "effects likely" row reflects the earlier assessment based on the


pre-1982 studies.  In adding the six cities results to the "effects possible"


row, it is interesting to note some consistency among the U.S. studies


with respect to concentration at which functional  and symptomatic effects


occur.  The Ferris et al. work (1973, 1976) suggests functional  effects may


occur down to levels of 130 ug/m3 TSP, but none of these studies find such


effects at lower concentrations.  The finding of  symptomatic responses' in


children with no change in lung function in the range of 60 to 114 ug/m3


(as TSP) (Ware et al. 1986) is consistent with similar findings  in adults


for a long term mean of 110 ug/m3 (60 to 150 ug/m^) TSP from the Bouhuys et


al. (1973) study.


     The conversio'n of the earlier studies to PM^Q in Table 4-2  reflects more


recent information.  In considering the earlier assessment and recommendations,


the Administrator proposed that the level of the  annual standard be no


higher than 65 ug/m3 PM1Q (49 FR 10408).  This, therefore, is the upper bound


of the present range of interest.  The lower bound is lowered from the


previous assessment to 40 ug/m3 as PM1(J, based on  the recent results of


Ware and coworkers.  The staff, therefore, recommends a range of interest

                      O
between 40 and 65 ug/nr for decision making on an  annual standard for PM^Q.


     The results of the original studies, assessment of risks at lower


levels, and conversion to a common indicator all  are subject to  considerable


uncertainties.  Furthermore, effects are not demonstrated within the ranges


listed above; the lower bounds represent conservative estimates  where some


risk of effect is not ruled out by the data.

-------
                         Table 4-2.  UPDATED STAFF ASSESSMENT OF  LONG-TERM  EPIDEMIOLOGICAL  STUDIES

Effects/Study
Effects Likely
Effects Possible
No Significant6
Effects Noted
Measured BS
Levels (as ug/m3
Increased
Respiratory
Disease, Reduced
Lung Function
in Children1
230-300 BS
<230 BS
-
Measured TSP Levels (ug/m3)
Increased Respiratory
Disease Symtpoms,
Small Reduction in
Lung Function in
Adults2
180*
130-180*
80-130
Increased
Respiratory
Symptoms
in Adults3

60-150(110)
-
Increased
Respiratory
Symptoms and
Illnesses in
Children4
-
60*-114
V
Reduced
Lung
Function
in
Children4
-
-
40-114
Combined
Range
_>180
60-180 •
<60 -
Fquivalent
PMjo Levels
(ug/m3)
Combined
Range;*
80-90
40-90
<40
'
                                                                                                                          en
                                                                                                                          oo
*Indicates levels used for upper and lower bound of range..

!study conducted in 1963-65 in Sheffield, England (Lunn et al.,  1967).   BS  levels  (as  ug/m3)  uncertain.
2Studies conducted in 1961-73 in Berlin, N.H. (Ferris et al.,  1973,  1976).   Effects  level  (180  ug/m3)
 based on uncertain 2-month average.  Effects in lung function were  relatively  small.
3Study conducted in 1973 in two Connecticut towns. (Bouhuys  et al.  1973).  Exposure estimates  reflect  1965-73  data  in
 Anson.  Median value (110 ug/m3) used to indicate long-term concentration.   No effects  on lung function,  but some
 suggesstion of effects on respiratory symptoms.
4Study conducted in 1976-1980 in 6 U.S. cities (Ware et al.,  1986).   Exposure estimates  reflect 4-year  averages  across
^cities.  Comparable pollution/effects gradients not noted within  cities.
 Conversion of TSP to PMjn equivalents for Berlin, Ansonia studies  based on  estimated  ratio of  PM^g/TSP  for current
 U.S. atmospheres (Pace, 1983).  The estimated ratio ranged  between  0.45 and 0.5.  Conversion for  six-city study
 based on site-specific analysis of particle size data (Spengler et  al., 1986).
^Ranges reflect gradients in which no significant effects  were detected for  categories  at  top.   Combined range
 reflects all columns.

-------
                                     59

     b)  Additional Factors to be Considered In Evaluating Margins of Safety
         and Risks - Long-term Exposures

When evaluating margins of safety (TISKS) in this range, additional factors

identified in the  1982 assessment (SP, p. 106-111) should be considered.

     (i) Aerosol Composition

     1.  S02 levels in the six cities studied by Ware et al. generally

covaried with TSP.  Where high S02 levels co-exist with PM].o» the above

range would appear to be protective.

     2.  The six cities study is directly relevant to current U.S. atmospheres

with periodic elevations of ozone.

     3.  The risks of lung function and respiratory illness noted in

these long-term studies can be expected to vary with particle composition

among different regions.  Although reliable comparisons of relative aerosol

toxicity on a unit mass basis are not available, the potential  impact of

such variability is reduced by the fact that Ware et al. (1986) compared

cities of distinct pollution and geographic characteristics.

  (ii) Risk for Other Sensitive Groups, Effects Not Evaluated

       Because of the limited scope and number of long-term quantitative

studies, it is important to examine tne results of qualitative  data from

epidemiological and animal studies.  These studies justify concern for

other sensitive groups (asthmatics, bronchitic subjects, the elderly,

individuals with cardiopulmonary disease), and for serious effects (damage to

lung tissue from acid aerosols and mineral dusts, cancer, premature mortality)

not directly evaluated.  Available data do not suggest major risks for

these effects categories or populations at current ambient particle levels

in most U.S. areas.  Nevertheless, the risk that both fine and  coarse mode

particles may produce these responses adds to the need to limit long-term

levels of PM^Q for a variety of aerosol compositions.

-------
                                     60

 C.   Summary  of  Staff Conclusions and Recommendations
      The  major  updated staff conclusions and recommendations made in
 Section  IV.A,B  are briefly summarized below.
      1.   The staff reaffirms its recommendation to replace TSP as the
 particle  indicator for the primary standards with a new indicator that
 includes  only those particles less than or equal to a nominal 10 urn, termed
 PMlQ.  The previously developed effectiveness criteria for samplers are
 acceptable for  regulatory purposes.
      2.   Based  on an updated staff assessment of the short-term epidemiological
 data the  range  of 24-hour PM^g levels of interest is 140 to 250 ug/m .
 The  upper end of the range reflects the judgment of the Administrator with
 regard to the maximum level proposed for a 24-hour standard, based on his
 consideration of the earlier criteria and assessments.  Although the recent
'information  provides additional -support for the possibility of effects at
 lower levels, it does not demonstrate that adverse effects would occur with
 certainty at a  PM^Q concentration of 250 ug/m .  This level, therefore,
 remains  an appropriate upper bound.  The recent data suggest that the range  of
 levels under consideration -of alternative standards can be reduced to 140
 ug/m-*, although the original lower bound of 150 ug/m3 is within the range
 of uncertainty  associated with expressing the data as PMio«  Neither the
 studies  used to derive this range nor the more qualitative studies of effects
 in other  sensitive population groups (e.g., asthmatics) or effects in controlled
 human or  animal studies provide convincing scientific support for health
 risks of  consequence below 140 ug/m3 in current U.S. atmospheres.  These
 qualitative  data as well as factors such as aerosol composition and exposure
 characteristics should also be considered in evaluating margins of 'safety
 associated with alternative standards in the range of 140 ug/m3 to 250 ug/m3.

-------
                                      61



      3.   Based  on  an  updated  staff  assessment  of  the  long-term  epidemiological



 data, 'the range of annual  PM^  levels of  interest is  40 to 65 ug/m  .   The



 upper end of  the range  reflects  the judgment of the Administrator with



 regard to the maximum level proposed  for  an annual standard, based  on  his



 consideration of the  earlier  criteria and assessment.  The staff concludes



 that  this level  remains  a  useful  upper bound.  The recent data  prompt



 consideration of a standard level below the previous  lower bound (50 ug/m3)



 to  values as  low as 40  ug/m3.   Uncertain,  data  from one recent study suggest



 that  at  this  level  some  risk  may  remain of respiratory effects  in children,



 but no detectable  increases in  pulmonary  function are expected  in children



 or  adults.



      When evaluating  margins  of  safety for an  annual  standard,  it is particularly



 important to  examine  the results  of qualitative data  from a number of



•epidemiologica-1, animal, and  air-quality  studies.  These suggest concern



 for effects not  directly evaluated  in the studies used to develop the



 ranges.   Such effects include damage  to lung tissues  contributing to chronic



 respiratory disease,  cancer,  and  premature mortality.  The available scientific



 data  do  not suggest major  risks  for these effects categories at current



 ambient  particle levels  in most  U.S.  areas.  Nevertheless, the  risk that



 both  fine and coarse  particles may  produce these  responses supports the



 need  to  limit long-term  levels of PM^Q for a variety  of aerosol compositions.



      4.   When selecting  final standard levels, consideration should be



 given to the  combined protection  afforded by the  24-hour and annual standards



 taken together.  For  example, a  24-hour standard  at 150 gg/m3 would



 substantially reduce  annual levels  in a number of areas below 50 ug/m3



 adding to the protection afforded by  an annual standards in areas with



 higher 24-hour  peak to  annual mean  ratios.

-------
                                     62



     Because of different  form,  averaging  procedures,  size  range,  and



limited PM^Q data,  precise comparisons between  the  above  ranges  of PM^y



standards and the current  primary  TSP standards are not possible.   A staff



analysis of PMig/TSP ratios applied  to recent TSP data shows  that  the



revised lower bounds, taken together, would  result  in  standards  clearly



more stringent than the  current  standards.   In  various analyses, standards



at the lower bound  of the  previous range  (150,50) have appeared  to range



from somewhat more  stringent to  approximately comparable  to the  present



primary standards.   Standards at the upper end  of the  range could, however,



result in about a four-fold decrease in the  areas exceeding the  primary



standards.

-------
APPENDIX A.  SUMMARY OF RECENT EPIDEMIOLOGICAL STUDIES ON PARTICULATE MATTER







A.I  INTRODUCTION



     This appendix presents a tabular summary and assessment of the community



epidemiological studies of particulate matter published since closure on



the 1982 criteria document and included in the CD addendum and not



summarized individually in Tables 3-1 or 3-2.  It is intended to support



discussions in Sections III and IV of this paper.  The tables follow the



organization of the criteria document and begin with studies of mortality



(Table A-l) associated with short-term exposures and are followed by and a



tabular summary of mortality (Table A-2) and morbidity (Table A-3) associated



w-ith long term exposures.

-------
        TABLE A-l.  EPIOEMIOLOGICAL STUDIES (1982-1986) ON SHORT-TERM CHANGES IN MORTALITY AND EXPOSURE TO PARTICLES
Data Base
Observed Effects/Comments
                                                                                Study
Daily fluctuations in total
London mortality and pollu-
tion during 14 winters
(1958 - 1972) over a period
when PM (BS) and SO-, levels
declined by 80% and 50%,
respectively.
Reanalysis of above data.
Report detailing reanalysis
of above data.  Mortality
data stratified by cause.
Unpublished reanalysis of
above data.
Regression coefficients averaged over 14 individual  winters; 25.1% change
in mortality per mg/m3 BS vs. 1.2 % per mg/m3 SOo.  Stratified quartile
analysis also indicates association primarily with BS but pollutant col-
linearity cannot be eliminated.  Subset of highest pollution days shows mor-
tality increases with BS across range; with S02 only >700 gg/m3.  Data cor-
rected for temperature, humidi-ty, day of week, annual, seasonal  trends.
Possible over-control for temperature.  Linear and quadratic models
using all BS data fit equally well.  Below 300 gg/rn3, smoke explains <0.2% of
mortality variation in quadratic model, * 10% in linear model.  Authors
hypothesize threshold and that quadratic model more plausible.  Higher
smoke coefficients in later, less polluted years possibly explained by
statistical model, surrogate behavior.

Significant effect of BS on mortality deviations for days of BS < 150 gg/m3
in 9 of 12 winters, and in 5 of 6 later winters with no BS levels > 500
In first 6 winters, BS significantly associated with mortality on days
with BS > 150 gg/m3; not for later years (with very few observations).
                                                                                Mazumdar
                                                                                et al.,  1982
                                                                                Ostro, 1984
Temperature, humidity controlled;
with higher BS as in above study.
                                                               SO^ not included.   Lower coefficients
Multiple time series analysis of detrended data controlled for autocorre-
lation.  Effect of temperature significant, greatest with 2-day lag.  BS and
S0£ predict mortality equally.well.  Log-linear relation holds for all  years -
no evident threshold or lag for pollution effect.  Strongest associations
with pollution and temperature cycles of 7-21 days; shorter cycles had  small
effect.  Pollutants more important than temperature in predicting overall
and respiratory deaths; temperature more important in cardiovascular mortality.
Use of log pollution levels not comparable with other analyses.  PM and
effects were not separated.
                                                                                Shumway et al.,
                                                                                1983
Autoregression models to control for time series effects (e.g., day of week,
epidemics), weather.  Accounting for temperature, humidity enhances signi-
ficance of pollution in explaining mortality.  BS consistently associated,
with mortality with or without S0£ included down to levels < 150-200 gg/nr
higher slopes in later years and at lower levels.  S02 significant only ir
2 years with high levels.  Diagnostic plots, regression results suggest
concave mortality/BS relationship.
                                                                                Schwartz and
                                                                                Marcus,  1986
                                                                                                     -3.

-------
         TABLE A-l.  EPIDEMIOLOGICAL STUDIES ON (1982-1986) SHORT-TERM CHANGES IN MORTALITY AND EXPOSURE TO PARTICLE
                                                          (cont'd)
Data Base
Observed Effects/Comments
Study
Daily fluctuations in mor-
tality and pollution during
5 winters (1972-1977) in
Pittsburgh, Pa., characteri-
zed by variations in PM
(CoH) and S02 across 3
monitoring sites.

Daily fluctuations in
mortality (sum of circu-
latory, respiratory,
cancer).  CoH, SOg, and
airport visibility during
14 years in N.Y.C. (1963-
1976) over a period when
S02 levels declined
by = 75% and CoH declined
slightly.
Daily fluctuations in
total mortality, SOg and
BS during 8 years in
Athens (197b-1982) over
a period when SOo and BS
levels declined by - 75%.
5 monitoring sites with
considerable variation in
BS, but not S02 levels.
Significant association between total and heart disease mortality
and PM, not S02, at high pollution site (CoHs * 1.25).  Non-significant, in-
consistent associations at other sites.  Seasonal  trends, day-of-week, weather
controlled.  Data filtered to account for autocorrelation.  Possible over-
control for temperature.  Only same-day effects considered.
Preliminary time series analysis, controlling for non-linear time trends,
found significant associations with PM indicators, S02, temperature.
Elevated CoH levels typical of period associated with 1.2-2% increase in
daily mortality.  Lower end of range corresponds to coefficients reported
by Schimmel (1978) in prior analysis of same data.  S02 associated with
0.3-1.5% increase in mortality.  On days of regional  visibility
deterioration, visibility derived extinction coefficient (surrogate for fine
particles) accounted for ~ 1% increase in mortality.   S02 and PM declined
at different rates in NYC such that effects of pollutants less confounded
than in other studies.  Unlike Schimmel (1978), S02 significantly correlated
with mortality.  CoH and S02 measured at only 1 site; mortality estimates
somewhat sensitive to location of visibility reading  (3 airports).

Unlike other studies, significant association between mortality and S02, but
not smoke, after controlling for temperature, secular, seasonal, monthly,
weekly variations and their interactions.  Regression coefficient for S02
unaltered after sequential removal  of days with values in excess of 500 down
to 150 ug/in .  Authors inferred threshold slightly below 150 ug/m  SO?.
Temporal trends controlled by subtracting expected mortality from 1956-1958
may introduce bias.  Reliability of smoke measurements unclear.
Mazumdar and
Sussman, 1983
Ozkaynak and
Spengler, 1985
Ozkaynak et
al., 1986
                                                                                                                            CO
Hatzakis
et al ., 1986

-------
       TABLE A-2.  EPIDEMIOLUGICAL STUDIES (1982-1986) OF EFFECTS ON MUKTALITY DUE TO LONG-TERM EXPOSURES TO PARTICLES
Date Base
Results/Comment
Study
Age-and sex-specific
1969-1970 mortality in 112 U.S.
SMSAs related to annual TSP, 504
(and ozone for subset of 69 SMSAs
based on 1975 levels)
1980 U.S. mortality in U.S. SMSAs
along with annual average fine
particle (FP; < 2.5 urn), inhalable
particle (IP < 15 urn), TSP, and
504 from central monitors.

1968-1972 mortality among 45-54 year
old whites in U.S. counties and aggre-
gated in Public Use Samples (PUS).
TSP, S02, N02 values derived from
1974-1976 data.
Asthma and bronchitis mortality during
1963-1983 in Japanese industrial city
(Yokkaichi) over a period when SOX
levels, dominated by petroleum emissions,
increased up to 1967 and declined after
1970 (50% reduction by 1973; 75% by
1982).  Levels of N02, TSP, oxidants
consistently low; only data from 1974-
1982 presented.  Comparisons with clean,
control areas.

1969-1973 mortality among 45-74 year
olds in England and Wales.  Smoke and
S02 data from 1971 and estimated his-
torical pollution exposure based on
coal consumption rates.
Attempts to improve Lave and Seskin (1978)  model  with additional         Lipfert,
variables (diet, drinking water, residential  heating fuels,  migration,   1984
SMSA growth) and to evaluate collinearity in  pollutant levels.   As
variables added, pollution lost significance.  TSP and 504 coefficients
unstable (elasticities between zero and ~6%);  neither significant  in
joint regression.  TSP coefficient more often significant across  data
sets, typically ~ 0.7 deaths/year/100,000 per ug/m3.  Sulfate coef-
ficent less robust, more often non-significant.  03 coefficient fairly
stable, when significant ~ 1.3.  Effects of individual pollutants
difficult to separate.

Regression analysis with control variables  for  % elderly, race,         Ozkaynak  &
population density, college education, poverty  (not smoking).  TSP  and   Spengler,
IP coefficients most often not significant.  Mean 504 or FP most         1985
consistent predictors of mortality.  Preliminary analysis, more
complex models await testing.

Regression analysis with 17 socio-economic  (SES) and 4 weather          Selvin et
control variables (no control for smoking,  occupational  exposure).       al., 1985
Inconsistent associations; S02 most often significant, TSP and
N02 mostly negative coefficients.  Interpretation of results
limited by use of retrospective exposure estimates, geographic
aggregations by groups of counties, possible  overcbntrol of SES.

Asthma mortality for persons >^ 60 years old rose (in 1967) and fell    Imai  et
(after 1971) significantly with SOX levels.  Significant increase      al.,  1986
in mortality due to bronchitis in persons >^ 60; after 1967,
continued increase following SOX declines in  1971; then decreased  •
from 1976 on, 5 years after pollution reductions.  Measurements of
SOX using lead peroxide methods limits quantification of acid
sulfate versus S02 effects.
Unlike comparable analyses of data from 1948-1964,  no association
between pollution and mortality except between S02  and chronic
bronchitis, hypertensive disease and all  causes of  death in females.
Positive 562 results inconsistent with other studies.  Suggests de-
clining effect of pollution on mortality  since mid-1950's.  No account
for occupational  exposure, migration.  BS readings  not calibrated to
local mass readings; limits reliability.
Chinn et
al., 1981

-------
          TABLE 'A-3.  EPIDEMIOLOGICAL STUUIES (1982-1986) ON MORBIDITY EFFECTS OF  LONG-TERM  EXPOSURES  TO*PARTICLES
Population/Exposure
                 Results/Comment
                                                                        Study
Parental questionnaire and lung
function tests for = 3500 2nd-4th
graders in 4 sections of Hamilton,
Ontario during 1979-1982.  Within-city
gradients in PM levels; multiple TSP
and size-selective particle monitors.
= 700 men and women (15-64 years)
examined between 1965 and 1984 in 3
year interval, living in 2 areas of
contasting pollution levels in the
Netherlands:
          Vlaardingen   Vlagtwedde
          Annual means  (24-hr, max)
          BS  S02
1965  40(185) 200(1030)
1974  40(165)  85( 362)
1984  37(124)  45(124)
  BS
"very low"
 "low" 16(158)
14(105) 15(117)
 « 22,000 children (6-11 years old)
from 19 geographic regions in 6
European countries in 1975.  Numerous
pollution monitors using different
measurement methods - collocated
monitors used to standardize
readings.  Range for adjusted annual
Black Smoke 5-57 ug/nr; annual  SO?
19-326 ug/m3.
No significant association between cough or episodes of bronchitis  and
pollution indices after adjusting for maternal  smoking, SES,  gas
stove use.  Peak flow and MEFy5, an index of small  airway function,
significantly associated with "fine" particles.  No association
between FEV and PM.  Possible biases in Cascade impactor readings,
modest gradients, limit conclusions (likely included much larger
particles).

After 1st 4 follow-ups (9 years), significantly greater (= 6050
decline over time of vital capacity and FEV^ in Vlaardingen compared
with rural area. Preliminary results indicate that  over 15 years,
difference in FEVj decline not significant—mainly  due to lower
FEV values than expected in last 2 exams in rural area.  Prevalence
of chronic phlegm and breathlessness always greater in Vlaardingen
but much smaller differences in later exams (1976-1984) when  symptom
prevalence increased in rural town (consistent  with lung function
chanyes)--possibly due to short episodes (5-10  days) of peak
pollution observed in rural area (24 hr. S02 =  50-80 ug/m ) during
last 2 exams
expected
More thorough data analyses and further testing
                 Across  countries  no significant  differences  in  respiratory  symptoms
                 related to smoke  or S02.   When systematic  differences  in  health
                 between countries accounted  for,  strong  associations in  Italy  and
                 Ireland between  smoke and  wheezing,  breathlessness, cough,  and non-
                 specific chronic  lung disease.   Annual levels in  both  those countries
                 between = 7-38 ug/m3 (winter means  24-54 ug/m3).   No smoke  effects
                 in  other countries.  Significant  S02 effects in some countries.
                 Limited application to current U.S.  PM exposures  given sampling
                 differences,  lack of calibration.
                                                                                         Pengelly
                                                                                         et  al.,
                                                                                         1986
                                                                                         van
                                                                                         der  Lende
                                                                                         et a I.,
                                                                                        1986
                                                                        CEC, 1983

-------
          TABLE A-3.  EPIOEMIULOGICAL STUDIES ON (1902-1986)  MORBIDITY EFFECTS  OF  LONG-TERM  EXPOSURES  TO  PARTICLES
                                                          (cont'd)
Population/Exposure
Children < 4 years from residential
areas of contrasting SOo and dustfall
levels reporting to a clinic in
Duisburg, W. Germany
Results/Comment
Clear distinction in incidence of croup and obstructive bronchitis
between high and low pollution areas (S02 < 300 ug/nr >, dustfall
< 0.35 g/mZ.day >). Other factors (e.g., infections, distance
to clinic, degree of crowding) accounted for but effects of S02
and PM cannot be separated. Limited application to assessing
effects of U.S. particles.
Study
Muhling
et al .,
1985
3,088 residents (aged 19-70 years)  of
Cracow, Poland, surveyed in 1968 and
1973, living in high pollution area
in city center (mean suspended
particle concentration = 118 ug/m^,
SOo = 114 ug/nr) versus those in
other areas (mean SP = 109 ug/m3,
S02 = 53
Air pollution by itself not a significant predictor of
bronchitis.  However, effects of occupational  exposure to
hazards (i.e., chemicals, irritating gases,  high temperature and
humidity) on prevalence of bronchitis, chronic cough/phlegm,
reduced FEV much greater in men living in high pollution areas.
Among men with persistent cough/phlegm, more frequent exacerbations
of symptoms in high pollution areas.  Little effect of pollution  on
women.  Attempts to test interactions among  multiple variables
suggest marked pollution effects only in combination with other
factors.  PM measurements not directly applicable to U.S.
levels.  Cannot separate PM from
Wojtyniak
et al.,
1986

-------
              APPENDIX B.  ESTIMATION OF PM10 LEVELS ASSOCIATED
                         WITH KEY STEUBENVILLE ALERTS

     This appendix details the calculations involved in converting TSP

measurements from the Steubenville alert studies into PM^g units.  Two of

the four Steubenville alert studies from Dockery et al . (1982) were used in

deriving the "possible" and "no significant effects noted" levels in

Table 4-1.  These were the Spring and Fall  1980 studies.  The peak 24-hour

TSP levels measured in these studies were,  respectively 220 ug/m^ and

160 ug/m^.  Relevant information on TSP and other particle size fraction

ratios in Steubenville is summarized in the report of Spengler et al .

(1986).  PM10/TSP ratias were available only for 1984-85, whereas PM15/TSP
                                                     i
ratios, as measured by dichotomous samplers, were available for portions of the

actual alert periods in Steubenville.  Therefore, the general approach taken

was to estimate a range of PMio/OClb ratios (where DC^ is PM^5 as measured

by dichotomous sample) expected for Steubenville and then multiply this by

the DCi5/TSP ratio observed during the relevant alert period -to estimate the

         ratio for the period, i.e.,
        PM10/DC15         x       DC15/TSP       =*      PM10/TSP         (8-1)
(estimated, 1979-85 data)   (observed for alert)   (estimated for alert)


     The range of PMio/DCl5 ratios was estimated by using 1984-85 measurements of

PM^O/TSP and PM^o/IP]^ ratios (where IPi5 is PM^5 as measured by size

selective high volume sampler) together with measurements of IPi5/DCi5 and

TSP/DCi5 made the 1979-84 period.  The two approaches can be written as:
    PM10/IP15         x        IPl5/DCi5       =      PMio/OC15           (6-2)
(measured 84-85)            (measured 79-84)         (estimated)

and;

    PM10/TSP          x        TSP/OC15        =      PM10/DC15           (B-3)
(measured 84-85)            (measured 79-84)         (estimated).

-------
                                   R-2



     The median PMio/IPi5  ratio  for S.teubenville  in 1984  is 0.76  (Table  V-6,


Spengler et al., 1986).  The median IPis/DC-is  ratio for 1980-84  (5-yr  average)


in Steubenville is 1/0.82  or 1.22  (Table V-5,  Spengler et al., 1986).  The


PMjo/0^15 estimated from equation  B-2  is,  therefore:


     0.76 x 1.22 = 0.93.


A somewhat lower figure  (0.87) can be  obtained by using the 1985  PMjQ/IPis


ratio while a comparable to  somewhat higher  estimate can  be derived  from


using the IPi5/DCi5 ratio  from 1980 (the year  in which the episodes  occurred)


in place of the longer term  value. Given  the  absence of  clear trends  in the


ratios, the longer term  estimate is preferred  in  this case.   Given the
                                            t

uncertainties in the estimates  a rounded PMio/DCis ratio  of 0.9  is derived from


this procedure.


     Similarly, a lower  bound estimate for PM^o/DC-is can  be derived  from


equation B-3.  The median  1984 PMio/TSP ratio  for Steubenville is 0.5  while


that for 1985 is 0.51 (Table V-6,  Spengler et  al., 1986). The median


TSP/DC15 ratio for 1979-84 (6 yr.  average) is  1/0.66 or 1.5.  From equation


A-3, the P^io/DCis ratio is  0.76,  which rounds to 0.8.  From  the  above


calculations, therefore, the range of  estimated  median PMio/0^15  ratios  for


Steubenville is 0.8 to 0.9.


     The PMio/TSP ratios for specific  alert  periods can be estimated using


these average ratios calculated  above  (0.8 to  0.9) together with  specific


PMJ5/TSP measurements made during  the  alert  studies.  In  the  spring  1980


study, the peak TSP level  of 220 ug/m^ occurred  on or about April 21 (Figure  1,


Dockery et al., 1982).  Although no particle ratios were  available on  that


day, the DCjs/TSP ratios for the higher concentrations days occurring  just


before and after that date are approximately 0.8  (Figure  IV-6, Spengler  et.


al., 1986).  Thus, by equation B-l, the range  of  estimated PMjo/TSP  ratios

-------
                                    B-3
for that alert period are:



    (0.8 to 0.9) x 0.8 = 0.64 to 0.72.



To estimate the lower bound of the "possible" effects  range,  the lower •



PM10/TSP ratio (0.64) is used and the estimated PM1Q level  is 2?0 ug/m3



x 0.64 = 140 ug/m3.  Although not useful  for deriving  a lower bound,  the



upper bound estimate would  yield 220 ug/m3 x 0.72 = 160 ug/m3.  These upper



and lower estimates bound the original lower bound (150 ug/m3) of the range



proposed for a 24-hour PM^g standard in 1984.



     In the Fall 1980 study, in which significant effects  were not observed,



the peak TSP concentration  of 160 ug/m3 was reported during the baseline



measurement period on or about October 17 (Figure 1, Dockery  et al.,  1982).



Corollary measures indicate somewhat higher concentrations  (  200 ug/m3 as



TSP) on this date as well as on a subsequent date (about October 29)  (Figure



IV-7, Spengler et al., 1986).  DCjs/TSP ratios for this study period  were



typically in the vicinity of 0.5 for most of the study period.  Thus,  the



range of estimated PM^o/TSP ratios from equation B-l is:



     (0.8 to 0.9) x 0.5 = 0.4 to 0.45.  '   ;  '



The range of estimated peak PMjg levels asociated with this study period  is



therefore:



     (160 to 200 ug/m3) x (0.4 to 0.45) = 65 to 90 ug/m3.

-------
     APPENDIX C





CASAC Closure Letter

-------
                                                          SAB-CASAC-87-010
             UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                           WASHINGTON. Q.C.  20460


                           December 16, 1986
                                                                   OFFICE OF
_.         ,,  ,   _.                                            THE ADMINISTB A TOW
The Honorable Lee Thomas
Administrator
U.S. Environmental Protection
  Agency
Washington, DC  20460

Dear Mr. Thomas:

     The Clean Air Scientific Advisory Committee (CASAC) has completed
its review of the 1986 Addendum to the 1982 Staff Paper on Particulate
Matter  (Review of the NAAQS for Particulate Matter;  Assessment of
Scientific and Technical Information) prepared by the Agency's Office
of Air Quality Planning and Standards (OAQPS).

     The Committee unanimously concludes that this document is consistent
in all significant respects with the scientific evidence presented and
interpreted in the combined Air Quality Criteria Document for Particulate
Matter/Sulfur Oxides and its 1986 Addendum, on which the CASAC recently
issued its closure letter.  The Committee believes that this document
provides you with the Jcind and amount of technical guidance that will
be needed to make appropriate revisions to the standards.  The Committee's
major findings and conclusions concerning the various scientific issues
and studies discussed in the Staff Paper Addendum are contained in the
attached report.

     Thank you for the opportunity to present the Committee's views on
this important public health issue.
                                   Sincerely,
                                   Morton Lippmann/Ph.D.
                                   Chairman
                                   Clean Air Scientific Advisory
                                      Committee
cc:   A. James Barnes
      Gerald Emison
      Vaun Newill
      John O'Connor
      Craig Potter
      Terry Yosie

-------
                                                                     SAB-CASAC-87-010
               SUMMARY OF MAJOR SCIENTIFIC  ISSUES AND CASAC
                 .CONCLUSIONS ON THE  1986 DRAFT ADDENDUM
                TO THE 1982 PARTICULATE MATTER STAFF PAPER
     The Committee found the technical discussions contained  in the Staff
Paper Addendum to be acceptable with minor revisions.

Particle Size Indicator

     The CASAC reaffirms its January 29,  1982  recommendation  that a particle
size indicator that includes only those particles less  than or equal to a
nominal 10 um aerodynamic diameter,  termed PM^o, is  appropriate for regulation
of particulate concentrations.   This judgment  is based  on analysis of the
earlier available data, and the analysis  of the recent  scientific studies
discussed in the 1986 Addendum to the Air Quality Criteria for Particulate
Matter/Sulfur Oxides and the 1986 Addendum to  the Particulate Matter Staff
Paper.

Implications of London Mortality Studies

     Further analyses of the London  mortality  studies,  including recent
analysis by Agency staff, suggest that:

     1) the data provide no evidence for  a threshold for the  association
        between airborne particles and daily mortality  or a change of
        coefficient with changes in  particle composition;

     2) mortality effects can be associated with PM  alone (with or
        without sulfur oxides);

     3) there is no reliable quantitative basis  for  converting
        Rritish Smoke (BS) readings  to PM^Q gravimetric mass
        at low (<100-200 ug/m3) BS levels, and hence the mortality
        data are not readily useful  for establishing a  lower bound for
        24-hour PM^o NAAQS, although the  suggestion  of  mortality at
        relatively low PM levels must be  given serious  consideration
        in selecting a margin of safety.

Interpretation of Lung Function Studies  for 24-hour  Standard

     Although the lung function decrements observed  in  children during and
after air pollution episodes are of  uncertain  health significance, the two
episodic lung function studies (Dockery et al.,  1986? Dassen et al., 1986)
are consistent with each other and the earlier work  of  Stebbings.  They
provide a relatively sensitive indication of possible short term physiological
responses.  Given the difficulty in  deriving a lower limit from the mortality
studies, these lung function studies can be useful  in determining lower
bounds for a 24-hour PM^g standard.

-------
                                     -2-
Interptetaticn of the six Cities Study for Annual Standard

     In general, the Committee felt that the six cities data are useful in
establishing the Tower bound of the range for the annual standard.  In
addition, the following are suggested by the data:

      1)  Cough and bronchitis, as defined in this study, are about twice
          as prevalent in children living in cities with PM^Q in t^6
          range of 40-60 ug/m3, in comparison to cities with 20-30 ug/m3;

      2)  Because factors other than particulate matter may affect the
          inter-city differences, it is difficult to determine whether
          these associations should be designated as "likely" health
          effects;

      3)  The results are consistent with the Ostro studies in terms of
          morbidity responses at long-term average particulate matter
          exposures within current particulate ambient air quality
          standards; and

      4)  The results are consistent with the Bouhuys study in terms
          of symptoms without changes in pulmonary function.

Ranges for 24—hour and Annual Standards for PM]p

     In its January 2, L986 letter to the Administrator, the CASAC noted
that its preliminary analyses of the jnore recent data do not indicate the
need for fundamental changes in the structure of the proposed particle
standards; however, the Committee pointed out that these new data suggest
the need to focus consideration on standards at or perhaps below the low
ends of the ranges proposed in the March 20, 1984 Federal Register Notice.
The ranges of interest then proposed were 150-250 ug/m3 for 24-hour standard,
and 50-65 ug/m3 for annual standard.

     Since then, EPA staff have proposed updated ranges of  interest for
both the 24-hour standard (140-250 ug/m3), and the annual standard (40-65
ug/m3), based on short-term and long-term epidemiological data, respectively.
The Committee finds these ranges of interest reasonable, given the scientific
data and related uncertainties; however, a final decision should also weigh
evidence from clinical and toxicolcgical studies as well.  The Committee
agrees with EPA staff that selection of final standards must include
consideration of the combined protection afforded by the 24-hour and annual
standards taken together.

     The Committee recommends that you consider setting the revised standards
at the lower ends of the proposed ranges for both the 24-hour and annual
standards.  The Committee recognizes that the exact levels  to be  chosen
for the 24-hour and annual standards represent a policy choice, influenced
by the need to include a margin of safety.  Given the uncertainty in  the
supporting scientific data, the Committee cannot distinguish the 'health
effects that may be observed at different levels near the  lower bound,
such as the health significance of setting the 24-hour standard at 140
ug/m3 compared to 150 ug/m3.

-------
                                 REFERENCES


Bailey, M. R.;  Fry,  F.  A.;  James, A. C.  (1982) The long-term  clearance  kinetics
  of insoluble  particles  frqm the human  lung. Ann. Occup.  Hyg.  26:  273-290.

Bates, 0. V.; Sizto, R. (1986) A  study of  hospital admissions and  air
  pollutants in southern  Ontario. In: Lee, S. D.; Schneider,  T.; Grant,  L.
  D.; Verkerk,  P.,  eds. Aerosols: research,  risk assessment,  control
  strategies, proceedings of  the  2nd US-Dutch international symposium;  May
  1985; Williamsburg, VA. Chelsea, MI: Lewis Publishers,  Inc.;  pp.  753-766.

Bates, D. V.; Sizto, R. V.  (1983) Relationships between air pollutant  levels
  and hospital  admissions in  southern Ontario. Can. J. Public Health 74:
  117-122.

Bohning, D. E.; Atkins, H.  L.; Cohn, S.  H. (1982) Long-term particle clearance
  in man: normal  and impaired. Ann. Occup. Hyg. 26: 259-271.

Bouhuys, A. G.; Beck, G.  J.;  Schoenberg, J.  B. (1978) Do  present levels  of  air
  pollution outdoors affect respiratory  health? Nature 276: 466-471.

Brunekreef, B.  (1986)  Personal communication to John Bachmann, Ambient  Standards
  Branch.  October  15,  1986.

CEC [Commission of  the  European Communities] (1983) Report on the  EC epidemiology
  survey on the relationship  between air pollution and respiratory  health  in
  primary school  children.  Brussels, Belgium: Environmental Research Programme.

Chann, T.L.; Lippmann,  M. (1980).  Experimental measurements  and empirical
  modeling of the regional  deposition of inhaled particles in human.s.
  Am. Ind. Hyg. Assoc.  41:'  399-409.

Chinn, S.; Florey,  C. du  V.;  Baldwin, I.G.,; Gorgol, M. (1981)  The relation
  of mortality  in England and Uales 1969-73  to measurements of  air  pollution.
  J. Epidemiol. Commun. Health 35: 174-179.

Committee on Public  Works,  U.S. Senate (1974)  A Legislative  History of  the
  Clean Air Amendments.  Volume 1.  Serial No. 93-18.  U.S. Government
  Printing Office,  Washington, O.C.  Prepared oy the Environmental  Policy
  Division of the Congressional Research Service of the Library of  Congress.

Uassen, W.; Brunekreef, B.;  Hoek, G.; Hofschreuder, P.; Staatsen,  B.;
  deGrout, H.;  Schouten,  E.;  Biersteker,. K.  (1986).  Decline  in children's
  pulmonary function during an air pollution episode.  J. Air.  Poll. Control
  Assoc.  36:1223-1227.

D.C. Cir. (1980)  Lead Industries  Association, Inc. v. EPA, F.2d, 14 ERC  1906
  (O.C. Cir.) Cert.  Denied  49 U.S.L.W. 3428  December 8, 1980.

D.C. Cir (1981) American  Petroleum Institute v. Costle, Nos.  79.1104
  ejt._aj_. (D.C. Cir.) September 3, 1981.

Dockery, D. W.; Ware, J.  H.;  Ferris, B.  G.,  Jr.; Speizer, F.  E.; Cook,  N. R.;
  Herman, S. M. (1982)  Change in  pulmonary function in children associated
  with air pollution episodes. J. Air Pollut. Control Assoc.  32: 937-942.

-------
 Emmet,  P.C.;  Aitken,  R.J.;  Hannan, W.J.  (1982).  Measurements of the total
   and  regional  deposition of  inhaled particles in the human respiratory
.   tract.   J.  Aerosol  Sci. 13:   549-560.-

 Ferris, B. G.,  Jr.; Mahoney,  J. R.; Patterson, R. M.; First, M. VI. (1973) Air
   quality, Berlin,  New Hampshire, March  1966 to December 1967. Am. Rev.
   ReSpir.  Dis.  108: 77-84. •

 Ferris, B. G.,  Jr.; Chen, H.; Puleo, S.; Murphy, R. L. H., Jr. (1976) Chronic
   non-specific  respiratory  disease in Berlin, New Hampshire, 1967-1973. A
   further  follow-up study.  Am.  Rev. Respir. Dis. 113: 475-485.

 Ferris, B.G., Jr.;  Dockery, D.W.; Ware,  J.H.; Speizer, F.E.; Spiro, R. Ill
   (1983)   The Six-City study:   examples  of problems in analysis of the data.
   Environ. Health  Perspect. 52: 115-123.

 Ferris, B.G., Jr.;  Speizer, F.E.; Ware,  J.H.; Spengler, J.D.; Dockery, D.
   W.  (1986)   The Harvard six  cities study.  In:  Lee, S.D.; Schneider, T.;
   Grant, L.D.;  Verkerk,  P.J., eds.  Aerosols, research, risk assessment
   and  control strategies.   Proceedings of the second U.S.-Dutch international
   symposium:  aerosols;  May 1985; Williamsburg, VA; Chelsea, MI; Lewis Publishers,
   Inc.; pp. 721-730.

 Garrard, C. S.; Levandowski,  R. A.; Gerrity, T. R.; Yeates, D. 8.; Klein, E.
   (1985) The  effects  of  acute respiratory virus infection upon trachea!
   mucous transport. Arch. Environ. Health. 40: 322-325.

 Gerrity, T. R.; Garrard, C. S.; Yeates,  D. 8. (1983) A mathematical model of
   particle retention  in  the air-spaces of human lungs. 8r. J. In'd. Med. 40:
   121-130.

 Hatzakis,  A.; Katsouyanni,  K.;  Kalandidi, A.; Day, N.; Trichopoulos, 0. (1986)
   Short-term  effects  of  air pollution on mortality in Athens. Int. J.
   Epidemiol.  15: 73-81.

 Hausman, J.A.;  Ostro, B.D.; Wise, D.A. (1984)  Air pollution and lost work.
   Cambridge,  MA: National Bureau of Economic Research; N8ER Working paper
   no.  1263.

 Heyder, J.; Gebhart,  J.; Stahlhofen, W.; Stuck, B. (1982) Biological variabil-
   ity  of particle  deposition  in the human respiratory tract during con-
   trolled  and spontaneous mouth-breathing. In: Walton, W. H.; Critchlow, A.;
   Coppock, S. M.,  eds. Inhaled  particles V: proceedings of an international
   symposium organized by the  British Occupational Hygiene Society; September
   1980; Cardiff, United  Kingdom. Ann. Occup. Hyg. 26: 137-147.

 Imai,  M.;  Yoshida,  K.; Kitabatake, M. (1986) Mortality from asthma and chronic
   bronchitis  associated  with  changes in  sulfur oxides air pollution. Arch.
   Environ. Health  41: 29-35.

 ISO  [International  Standards  Organization] (1981) Size-definitions for particle
   sampling.   Am. Ind. Hyg.  Assoc. I. 42:64-68a.

-------
Lawther, P.O. (1986)  Letter  to  John  Bachmann, Office of Air  Quality  Planning
  and Standards,  U.S. Environmental  Protection Agency, Research  Triangle Park,
  N.C.  August 22, 1986.

Lawther, P. J.; Waller,  R.  £.;  Henderson, M.  (1970) Air pollution  and
  exacerbations of bronchitis.  Thorax  25: 525-539.

Lebowitz, M. D.;  Bendheim,  P.;  Cn'stea,  6.; Markovitz, 0.; Misiaszek,  J.;
  Staniec, M.; Van Wyck,  D.  (1974)   The  effect of air pollution  and  weather
  on lung function in exercising children and adolescents.   Am.  Rev.
  Respir. Dis. 109:  262-273.

Levandowski, R. A.;  Gerrity, T. R.;  Garrard, C.  S.  (1985) Modifications  of lung
  clearance mechanisms by acute influenza A infection. J. Lab. Clin. Med.
  106: 428-432.

Lipfert, F. W. (1984) Air pollution  and  mortality:  specification searches using
  SMSA-based data. J. Environ.  Econ. Manage.  11: 208-243.

Lippmann, M. (1986)   CASAC  Review and  Closure on the OAQPS Particulate Matter
  Staff Paper Addendum.   Letter to Lee M. Thomas. December 16,  1986.

Lunn, J. E.; Knowelden,  J.;  Handyside, A. J.  (1967) Patterns of  respiratory
  illness in Sheffield infant schoolchildren. Br. J. Prev. Soc.  Med. 21:
  7-16.

Marcus, A.; Schwartz, J.  (1986).  Statistical Reanalyses of  Data Relating
  Mortality to Air Pollution During  14 London Winters.  Memorandum to
  Bruce Jordan, Chief, Ambient  Standards Branch  a'nd Les Grant, Director",
  Environmental Criteria  and Assessment  Office,  U.S. Environmental Protection
  Agency, Research Triangle  Park, N.C.  27711

Martin, A. E. (1964)  Mortality  and morbidity  statistics and  air  pollution.
  Proc. R. Soc. Med.  57:  969-975.

Martin, A. E.; Bradley,  W.  H. (1960) Mortality,  fog and atmospheric  pollution--
  An investigation during the winter of  1958-59. Mon. Bull.  Minist.  Health
  Public Health Lab.  Serv.  19:  56-72.

Mazumdar, S.; Sussman, N. (1981) Relationships of air pollution  to health:
  Results from the Pittsburgh Study. Presented at:  74th annual meeting;  of
  the Air Pollution  Control  Association; June, Philadelphia, PA. Pittsburgh,
  PA; Air Pollution  Control  Association.

Mazumdar, S.; Sussman, N. (1983) Relationships of air pollution  to health:
  results from the Pittsburgh study. Arch. Environ. Health 38: 17-24.

Mazumdar, S.; Schimmel,  H.;  Higgins, I.  T. T. (1982) Relation of daily
  mortality to air pollution: an analysis of  14  London winters,
  1958/59-1971/72. Arch.  Environ. Health 37:  213-220.

Mazumdar, S.; Schimmel H.;  Higgins,  I.T.T. (1983)   Response  to letter  to
  the Editor.  Arch.  Env. Health 38: 123-126.

-------
McDonnell, W.F.; Horstman, D.H.; Hazucha, M.J.; Seal,  E.,  Jr.;  Haak,  E.D.,
  Salaam, S.; House, !) = £, (1983) Pulmonary effects of  ozone exosure during
  exercise:  dose-response characteristics.  J. Appl.  Physiol.  54:   1345-1352.

Miller, F. J.; Grady, M. A.; Martonen, T. B. (1984) Coarse mode aerosol
  behavior in man: theory and experiment. In: Liu, 8.  Y.  H.; Piu,  D.  Y.  H.;
  Fissan, H., eds. Aerosols: science, technology, and  industrial  applications
  of airborne particles. New York, NY: Elsevier Science Pub. Co.;  pp.
  999-1002.

Miller, F. J.; Martonen, T. 8.; Menache, M. G.; Spektor,  0. M.; Lippmann, M.
  (1986) Influence of breathing mode and activity level  on the  regional
  deposition of inhaled  particles and implications for regulatory  standards.
  Cambridge, United Kingdom:  Inhaled particles VI: accepted for  publication.

Miller, F. J.; Gardner,  D. E.; Graham, J. A.; Lee, R.  E.,  Jr.;  Wilson,  W. E.;
  Bachmann, J. D. (1979) Size considerations for establishing a standard for
  inhalable particles. J. Air Pollut. Control Assoc.  29:  612-615.

Muhlin'g, P.; Bory, J.; Haupt, H. (1985) Einfluss der  Luftbelastung  auf
  Atemwegserkrankungen.  Untersuchungen bei  Saeuglingen und Kleinkindern
  [The influence of air  pollution on respiratory diseases: studies  of babies
  and infants]. Staub Reinhalt. Luft 45: 35-38.

Ostro, B. D. (1983) The  effects of air pollution on work  loss and  morbidity.
  J. Environ. Econ. Manag. 10: 371-382.

Ostro, 8. (1984) A search for a threshold in the relationship of  air  pollution
  to mortality: a reanalysis of data on London winters.  EHP Environ.  Health
  Perspect. 58: 397-399.

Ostro, B. (1986).  Letter to Les Grant, Environmental  Criteria  and  Assessment
  Office.  August 15, 1986.

Ostro, B. D. (1987) Air  pollution and morbidity- revisited: a specification
  test. J. Environ. Econ. Manage.: in press.

Ozkaynak, H.; Spengler,  J. U.; Garsd, A.; Thruston, G.O.  (1986) Assessment  of
  population health risks resulting from exposures to  airborne  particles.
  In:  Lee, S.O.; Schneider, T.; Grant, L.D.; Verkerk, P.J., eds.   Aerosols:
  research, risk assessment and control strategies, proceedings of  the
  Second U. S.-Dutch international symposium on aerosols;  May 1985;
  Williamsburg, VA.  Chelsea, MI: Lewis Publishers, Inc.;  pp. 1067-1080.

Ozkaynak, H.; Spengler,  J. D. (1985) Analysis of health effects resulting from
  population exposures to acid precipitation precursors.  EHP Environ. Health
  Perspect. 63: 45-55.

Pace, T.G. (1983) The-Use of TSP Data to Estimate PM10 Concentrations.
  Monitoring and Data Analysis Division.  Technical Memorandum  to  John  Bachmann,
  Strategies and Air Standards Division, Office of Air Quality  Planning  and
  Standards, U.S. EPA, Research Triangle Park, N.C.

-------
Pace, T.G.  (1986)  Review of  IP  Network PMjQ  Data.  Air Management  Technology
  Branch.   Memorandum to Bruce  Jordan, -Chief, Ambient Standards  Branch,  Office
  of Air Quality Planning and Standards, Research Triangle  Park,  N.C.

Pace, T.G.; Frank, N.H.  (1984)  Procedures for Estimating  Probability of
  Non-attainment of a PM^g NAAQS Using Total Suspended Particulate or  Inhalable
  Particulate Data (Draft).  Monitoring and  Data Analysis Division, Office
  of Air Quality Planning and Standards, Research Triangle  Park,  N.C.  27711

Pengelly,  L. D.; Goldsmith,  C.  H.; Kerigan,  A. T.; Furlong, W.;  Toplack,  S. A.
  (1986) The Hamilton study: effect of particle size on respiratory health
  in children. In: Lee,  S. D.;  Schneider, T.; Grant L. D.;  Verkerk, P.,  eds.
  Aerosols: research, risk assessment, control strategies,  proceedings of
  the 2nd  US-Dutch international symposium;  May 1985; Wi1liamsburg, VA.
  Chelsea,  MI: Lewis Publishers, Inc.; pp. 753-766.

Perry, G.  B.; Chai, H.;  Dickey, D. W.; Jones, R. H.; Kinsman,  R.  A.; Morrill,
  C. G.; Spector,  S. L.; Weiser, R. C. (1983) Effects of  particulate air
  pollution on asthmatics. Am.  J. Public Health 73: 50-56.

Phalen, R.  F.; Oldham, M. J.; Beaucage, C. B.; Crocker, T.  T.; Mortensen, J. D.
  (1985) Postnatal enlargement  of human tracheobronchial  airways  and
  implications for particle  deposition. Anat. Rec. 212: 368-380.

Philipson,  K.; Falk, R.; Camner, P. (1985) Long-term clearance in  humans
  studied  with teflon particles labeled with chromium-51. Exp. Lung Res.  9:
  31-42.

Pollack, A.K.; Hudischewskj, A.B.; Thrall, A.D. (1985) An Examination  of
  1982-1983 Particulate  Matter  Ratios and Their Use in the  Estimation  of
  PM]_Q NAAQS Attainment  Status.  Systems Application, Inc., San  Rafael,  CA.
  U.S. EPA, Office of Air Quality Planning and Standards, Research Triangle
  Park, N.C., EPA-450/4-85-010.

Portney, P.R.; Mullahy,  J. (1986) Urban air  quality and acute  respiratory
  illness.   J. Urban Econ. 20:  21-38-.

Rodes, C.E.; Rahme, K.A.; Purdue, L.J. (1981) Particle Collection  Criteria
  for 10 Micron Samplers.  U.S. Environmental Protection  Agency,  Environmental
  Monitoring Systems Laboratory, Research Triangle Park,  N.C.  27711

Roth, H.D.; Wyzga, R.E.; Hayter, A.J. (1986) Methods and  Problems  in Estimating
  Health Risks from Particulates In Aerosols (S.D. Lee, T.  Schneider,
  L.D.  Grant, P.J. Verkerk, eds).  Lewis Publishers, Chelsea, MI,  pp. 837-957.

Saric, M.;  Fugas,  M.; Hrustic,  0. (1981) Effects of urban air  pollution  on
  school age children.  Arch. Environ. Health 36:  101-108.

Schimmel,  H-. (1978) Evidence for possible acute health effects of  ambient
  air pollution from time series analysis:   methodological  questions and
  some new results based on  New York City daily mortality,  1963-1976.
  Bull N.Y. Acad.  54: 1052-1108.

-------
Schimmel, H.; Murawski,  T.d. (1976) The relation of air pollution to
  mortality,  J, Occup,  Med. 18:  316-333,

Schwartz, J.; Marcus, A. (1986).  Statistical  reanalysis of data relating
  mortality to air pollution during London winters 1958-1972.  U.S.  EPA,
  Office of Policy Planning and Evaluation, Office of Air Quality Planning
  and Standards, Research Triangle Park, N.C.  27711.

Selvin, S.; Merrill, 0.; Wong,  L.; Sacks, S. T. (1984) Ecological  regression
  analysis and the study of the influence of air quality on mortality.  EHP
  Environ. Health Perspect. 54: 333-340.

Shumway, R. H.; Tai, R.  Y.; Tai, L. P.; Pawitan, Y. (1983)  Statistical  analysis
  of daily London mortality and associated weather and pollution effects.
  Sacramento, CA: California Air Resources Board;  contract  no.  Al-154-33.

Spengler, J. D.; Thurston,  G. D. (1983) Mass and elemental  composition  of fine
  and coarse particles  in six U.S. cities. J.  Air Pollut. Control  Assoc.  33:
  1162-1171.

Spengler, J.D.; liriggs,  S.L.K.; Ozkaynak, H. (1986) Relationships Between
  TSP Measurements and  Size-Fractionated Particle Mass Measurements  in
  Six Cities Participating  in the Harvard Air  Pollution Health  Study.   Prepared
  for Office of Air Quality Planning and Standa-rds, Office  of Policy Analysis,
  U.S. Environmental Protection Agency, Oecember 5, 1986.

Stebbings, J. H., Jr.;  Fogleman, D. G.  (1979)  Identifying a susceptible
  subgroup: effects of  the  Pittsburgh air pollution episode upon school
  ch.ildren. Am. -J.' Epid'emiol. 110: 27-40."          '  '                         '

Svartengren, M.; Hassler, E.; Philipson, K.; Camner, P. (1986)  Spirometric data
  and penetration of particles  to the alveoli. Br. J. Indus. Med.  43:
  188-191.

Svartengren, M. (1986)  Lung deposition  and clearance of particles  in healthy
  persons and patients with bronchiectasis. Stockholm, Sweden.

Swift, D. L.; Proctor,  0. F. (1982) Human respiratory deposition of  particles
  during oronasal breathing. Atmos. Environ. 16: 2279-2282.

U. S. Environmental  Protection  Agency.  (1982a) Review of the National Ambient
  Air Quality Standards  for Particulate Matter: Assessment  of Scientific  and
  Technical Information  (OAQPS  staff paper). Research Triangle  Park, NC.
  Office of Air Quality  Planning and Standards, EPA report  no.  EPA-450/5-
  82-001. Available from: NTIS, Springfield, VA; PB82-177874.

U. S. Environmental  Protection  Agency.  (1982b) Air Quality  Criteria  for Particulate
  Matter and Sulfur Oxides:  Research Triangle Park, NC: Environmental  Criteria
  and Assessment Office, EPA report no. EPA-600/8-82-029cF.  Available  from:
  NTIS,  Springfield, VA; PBS4-156801/REB.

-------
van der Lende,  R.;  Schouten,  J.P.; Rijcken, 8.; van der Meulen,  A.  (1986)
  Longitudinal  epidemiologic  studies on effects of air pollution  in  the
  Netherlands.  In:   Lee,  S.D.;  Schneider, T.; Grant, L.O.; Verkerk,  P.J.,
  eds.  Aerosols:   research,  risk assessment and control strategies,
  proceedings of the second U.S.-Dutch international symposium;  May,  1985;
  Williamsburg, VA. Chelsea,  MI.  Lewis Publishers, Inc.;,pp.  731-742.

van der Meulen  A.  (1986)  The  relationship between PM^g and thoracic  particle
  sampling. J.  Air  Pollut. Control Assoc. 35: 383-387.

Ware, J. H.; Thibodeau,  L. A.;  Speizer, F. E.; Colome, S.; Ferris,  B. G., Jr.
  (1981) Assessment of the health effects of atmospheric sulfur  oxides and
  particulate matter:  Evidence  from observational studies. Environ.
  Health Perspect.  41: 255-276.

Ware, J. H.; Ferris, B.  G., Jr.; Dockery, D. W.; Spengler, J.  0.; Stram, D. 0;
  Speizer, F. E. (1986)  Effects of ambient sulfur oxides and suspended
  particles on  respiratory health of preadolescent children. Am.  Rev.
 • Respir. Dis.  133: 834-842.

Wojtyniak, B.;  Krzyzanowski,  M.; Jedrychowski, W. (1984) Importance  of urban
  air pollution in  chronic respiratory problems = Die Bedeutung  staedtischer
  Luftverunreinigung fuer chronische Atemwegsbeschwerden. Z. Erkr.
  Atmungsorgane 163: 274-284.

-------
                                   TECHNICAL REPORT DATA
                            (Please read Insmictions on the reverse before completing)
1. REPORT NO.-
 EPA  450/05 86-012
                                                           3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE         ...    . .   „   , . L   ,*.   j  j
 Review  of the National Ambient Air  Quality Standards.
 for  Particulate Matter:  Updated Assessment of
 Scientific and Technical Information   Addendum to the
                     Staff Paper	
             5. REPORT DATE
              December  1986
             6. PERFORMING ORGANIZATION CODE
                                                           8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Office of Air and Radiation
  Office of Air Quality Planning and Standards
  U.S.  Environmental Protection Agency
  Research Triangle Park, North Carolina  27711
                                                           10. PROGRAM ELEMENT NO.
             11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
                                                           13. TYPE OF REPORT AND PERIOD COVERED
                                                            Final
                                                           14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
  This paper evaluates and  interprets the updated scientific and  technical  information
  that the EPA staff believes  is  most relevant to decision making on  revised primary
  (health) national ambient air quality standards (NAAQS) for  particulate matter and is
  an addendum to the 1982 particulate matter staff paper.  This assessment is intended
  to help bridge the gap between  the scientific review contained  in the EPA criteria
  document addendum and the judgments required of the Administrator in making final
  decisions on revisions to the primary NAAQS for particulate  matter that were proposed
  in March 1984 (49 FR 10408).  The major recommendations of this addendum include the
  following:
       1.  The staff reaffirms its recommendation to replace TSP  as the particle indica-
  tor for the primary standards-with a new indicator that includes only those particles
  less than or equal to a nominal  10 um, termed PM,Q.
       2.  Based on an updated staff assessment of the short-ternuepidemiological data,
  the range of 24-hour PM,Q levels of interest is 140 to 250 ug/m .
       3.  Based on an updated staff assessment of the long-term  epidemiological data,
  the range of. annual PM,Q  levels of interest is 40 to 65 ug/m3.
       4.  When selecting final standard levels, consideration should be given to the
  combined protection afforded by the 24-hour and annual standards taken together.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b.IDENTIFIERS/OPEN ENDED TERMS
                           c.  COS AT I  I ield/Group
  Particulate Matter
  Aerosols
  Air Pollution
  Sulfur Oxides
Air Quality  Standards
18. DISTRIBUTION STATEMENT

  Release to  public
19. SECURITY CLASS (This Report I
Unclassified
21. NO. OF PAGES
       100
                                              20. SECURITY CLASS (THis page;
                                                                         22. PRIC H
EPA Form 2220-1 (Rev. 4-77)
                      PREVIOUS EDITION IS OBSOLETE

-------