-------
o
tr
NJ
O
O
TABLE 6-23 (cont'd). LONG-TERM PARTICULATE MATTER EXPOSURE: RESPIRATORY SYMPTOM, LUNG
FUNCTION, AND BIOMARKER EFFECTS
Reference citation, location, duration, type of
study, sample size, pollutants measured,
summary of values
Health outcomes measured, analysis design,
covanates included, analysis problems
Results and Comments
Effects of co-pollutants
Effect estimates as reported by study
authors. Negative coefficients for lung
function and ORs greater than 1 for other
endpomts suggest effects of PM
ON
O
o
2
o
H
O
H
o
HH
H
Europe (cont'd)
Turnovska and Kostiranev (1999)
Dimitrovgrad, Bulgaria, May 1996
Total suspended paniculate matter (TSPM)
mean levels were 520 ±161 jug/m3 in 1986
and 187 ± 9 ^ug/m3 in 1996. S02, H2S, and
NO2 also measured.
Jedrychowski et al. (1999)
In Krakow, Poland in 1995 and 1997
Spacial distributions for BS and SO2 derived
from network of 17 air monitoring stations.
BS 52.6 Mg/m ± 53.98 in high area and
33.23 ±35.99 in low area.
Jedrychowski and Flak (1998)
In Kracow Poland, in 1991-1995
Daily 24 h concentration of SPM (black
smoke) measured at 17 air monitoring
stations.
High areas had 52.6 //g/m3 mean compared to
low areas at 33.2
Respiratory function of 97 schoolchildren
(mean age 10.4 ± 0.6 yr) measured in May
1996 as a sample of 12% of all four-graders in
Dimitrovgrad. The obtained results were
compared with reference values for Bulgarian
children aged 7 to 14 yr, calculated in the same
laboratory in 1986 and published (Gerginova
etal., 1989; Kostiranevetal., 1994). Variation
analysis technique were used to treat the data.
Effects on lung function growth studied in
preadolescent children. Lung function growth
rate measured by gain in FVC and FEV, and
occurrence of slow lung function growth
(SLFG) over the 2 yr period defined as lowest
quintile of the distribution of a given test in
gender group. 1129 children age 9 participated
in first year and 1001 in follow-up 2 years
later. ATS standard questionnaire and PFT
methods used. Initially univariate descriptive
statistics of pulmonary function indices and
SLFG were established, followed by
multivariate linear regression analyses
including gender, ETS, parental education,
home heating system and mold. SO2 also
analyzed.
Respiratory health survey of 1,129 school
children (aged 9 yr). Respiratory outcomes
included chronic cough, chronic phlegm,
wheezing, difficulty breathing and asthma.
Multi-variable logistic regression used to
calculate prevalence OR for symptoms
adjusted for potential confounding.
Vital capacity and FEV, were
significantly lower (mean value. =
88.54% and 82.5% respectfully)
comparing values between 1986 and
1996. TSPM pollution had
decreased by 2.74 times to levels
still higher than Bulgarian and
WHO standards.
Statistically significant negative
association between air pollution
level and lung function growth
(FVC and FEV,) over the follow up
in both gender groups. SLFG was
significantly higher in the more
polluted areas only among boys.
In girls there was consistency in the
direction of the effect, but not stat.
significant. Could not separate BS
and SO2 effects on lung function
growth. Excluding asthma subjects
subsample (size 917) provided
similar results.
The comparison of adjusted effect
estimates revealed chronic phlegm
as unique symptom related neither
to allergy nor to indoor variable but
was associated significantly with
outdoor air pollution category
(APL). No potential confounding
variable had major effect.
Boys
SLFG (FVC)
OR = 2.15(CI 1.25 -3.69)
SLFG (FEV,)
OR=1.90(CI 1.12-3.25)
Girls
FVC OR = 1.50 (CI 0.84 - 2.68)
FEV1 OR=1.39(CI0.78- 2.44)
It was not possible to assess separately the
contribution of the different sources of air
pollutants to the occurrence of respiratory
symptoms. ETS and household heating
(coal vs. gas vs. central heating) appeared
to be of minimal importance.
-------
M
g.
N>
O
o
TABLE 6-23 (cont'd). LONG-TERM PARTICULATE MATTER EXPOSURE: RESPIRATORY SYMPTOM, LUNG
FUNCTION, AND BIOMARKER EFFECTS
Effect estimates as reported by study
authors. Negative coefficients for lung
function and ORs greater than 1 for other
endpomts suggest effects of PM
Reference citation, location, duration, type of
study, sample size, pollutants measured,
summary of values
Health outcomes measured, analysis design,
covariates included, analysis problems
Results and Comments
Effects of co-pollutants
Tl
H
6
o
z
o
H
/O
G
O
H
Latin America
Calderon-Garciduenas et al. (2000)
Southwest Metropolitan Mexico City
(SWMMC) winter of 1997 and summer of
1998.
Australia
Lewis etal. (1998)
Summary measures of PM]0 and SO2
estimated for each of 10 areas in steel cities of
New South Wales.
Asia
Wong etal. (1999)
Hong Kong, 1989 to 1991
Sulfate concentrations in respirable particles
fell by 38% after implementing legislation
reducing fuel sulfur levels.
Study of 59 SWMMC children to evaluate
relationship between exposure to ambient
pollutants (O3 and PM1()) and chest x-ray
abnormalities. Fishers exact test used to
determine significance in a 2x2 task between
hyperinflation and exposure to SWMMC
pollutant atmosphere and to control, low-
pollutant city atmosphere.
Cross-sectional survey of children's health and
home environment between Oct 1993 and Dec
1993 evaluated frequency of respiratory
symptoms (night cough, chest colds, wheeze,
and diagnosed asthma). Covariates included
parental education and smoking, unflued gas
heating, indoor cats, age, sex, and maternal
allergy. Logistic regression analysis used
allowing for clustering by GEE methods.
3405 nonsmoking, women (mean age 36.5 yr;
SD ± 3.0) in a polluted district and a less
polluted district were studied for six
respiratory symptoms via self-completed
questionnaires. Binary latent variable
modeling used.
Bilateral symmetric mild lung
hyperinflation was significantly
associated with exposure to the
SWMMC air pollution mixture
(p>0.0004). This raises concern for
development of chronic disease
outcome in developing lungs.
SO2 was not related to differences in
symptom rates, but adult indoor
smoking was.
Night cough OR 1.34(1.18, 1.53)
Chest colds OR 1.43 (1.12, 1.82)
Wheeze OR 1.13 (0.93, 1.38)
Comparison was by district; no PM
measurements reported. Results
suggest control regulation may have
had some (but not statistically
significant) impact.
o
s
-------
O
o
TABLE 6-23 (cont'd). LONG-TERM PARTICULATE MATTER EXPOSURE: RESPIRATORY SYMPTOM, LUNG
===^======^== FUNCTION, AND BIOMARKER EFFECTS
Effect estimates as reported by study
authors. Negative coefficients for lung
function and ORs greater than 1 for other
endpomts suggest effects of PM
Reference citation, location, duration, type of
study, sample size, pollutants measured,
summary of values
Health outcomes measured, analysis design,
covanates included, analysis problems
Results and Comments
Effects of co-pollutants
ON
H
6
o
2!
O
H
O
G
O
H
tfl
O
*J
O
H—(
H
M
(cont'd)
Wang etal. (1999)
Kaohsiung and Panting, Taiwan
October 1995 to June 1996
TSP measured at 11 stations, PMM, at 16
stations. PM10 annual mean ranged from 19.4
to 112.81 ,ug/m3 (median = 91.00 ^g/m3)
TSP ranged from 112.81 to 237.82 ,ug/m3
(median = 181.00). CO, NO2, SO2,
hydrocarbons and O3 also measured.
Quo etal. (1999)
Taiwan, October 1955 and May 1996
PM,0 measured by beta-gauge.
Also monitoring for SO2, NO2> O3, CO.
Wang etal. (1999)
Chongqumg, China
April to July 1995
Dichot samplers used to measure PM2 s.
Mean PM2 5 level high in both urban
(143 /ug/m3) and suburban (139 //g/m3) area.
SO2 also measured
Relationship between asthma and air pollution
examined in cross-sectional study among
165,173 high school students (11- 16 yr).
Evaluated wheeze, cough and asthma
diagnosed by doctor. Video determined if
student displayed signs of asthma. Only
155,283 students met all requirements for
study analyses and, of these, 117,080 were
covered by air monitoring stations. Multiple
logistic regression analysis used to determine
independent effects of risk factors for asthma
after adjusting for age, gender, ETS, parents
education, area resident, and home incense use.
Study of asthma prevalence and air pollutants.
Survey for respiratory disease and symptoms in
middle-school students age < 13 to > 15 yr.
Total of 1,018,031 (89.3%) students and their
parents responded satisfactorily to the
questionnaire. Schools located with 2 km of
55 monitoring sites. Logistic regression
analysis conducted, controlling for age, hx
eczema, parents education.
Study examined relationship between PFT and
air pollution. Pulmonary function testing
performed on 1,075 adults (35 - 60 yr) who
had never smoked and did not use coal stoves
for cooking. Generalized additive model used
to estimate difference, between two areas for
FEV,, FVC, and FEV,/FVC% with adjustment
for confounding factors (gender; age, height,
education, passive smoking, and occupational
exposures).
Asthma significantly related to high
levels of TSP, NO2, CO, O3 and
airborne dust. However PMHI and
SO2 not associated with asthma.
The lifetime prevalence of asthma
was 18.5% and the 1-year
prevalence was 12.5%.
Because of close correlation among
air pollutants, not possible to
separate effects of individual ones.
Factor analysis used to group into
two classes (traffic-related and
stationary fossil fuel-related). No
association found between lifetime
asthma prevalence and nontraffic
related air pollutants (SO2, PM10).
Mean SO2 concentration in the
urban and suburban area highly
statistically significant different
(213 and 103 ^g/m3 respectfully).
PM2 5 difference was small, while
levels high in both areas. Estimated
effects on FEV1 statistically
different between the two areas.
Adjusted OR
PM,0
1.00(0.96-1.05)
TSP
1.29(1.24-1.34)
Difference between urban and suburban
area excluding occupational exposures:
FEV,
B - 119.79
SE 28.17
t - 4.25
p<0.01
FVC
B - 57.89
SE 30.80
t- 1.88
p < 0.05
-------
2
fa
K)
O
O
TABLE 6-23 (cont'd). LONG-TERM PARTICULATE MATTER EXPOSURE: RESPIRATORY SYMPTOM, LUNG
FUNCTION, AND BIOMARKER EFFECTS
Effect estimates as reported by study
authors. Negative coefficients for lung
function and ORs greater than 1 for other
endpomts suggest effects of PM
Reference citation, location, duration, type of
study, sample size, pollutants measured,
summary of values
Health outcomes measured, analysis design,
covariates included, analysis problems
Results and Comments
Effects of co-pollutants
K>
£
Tl
H
6
o
2
O
H
O
c
o
Asia (cont'd)
Zhang etal. (1999)
4 areas of 3 Chinese Cities (1985 - 1988)
TSP levels ranged from an annual arithmetic
mean 137 /ug/m3 to 1250 ,ug/m3 using
gravimetric methods.
Qianetal. (2000)
4 China cities
The 4 year average PM means were 191, 296,
406, and 1067 Mg/m3. SO2 and NO2
measurements were also available.
A pilot study of 4 districts of 3 Chinese cities
in for the years 1985-1988, TSP levels and
respiratory health outcomes studied. 4,108
adults (< 49 yrs) examined by questionnaires
for couth, phlegm, wheeze, asthma, and
bronchitis. Categorical logistic—regression
model used to calculate odds ratio. SO2 and
NO2 were also examined. Other potential
confounding factors (age, education level,
indoor ventilation, and occupation) examined
in the multiple logistic regression model.
Pilot cross-sectional survey of 2789 elementary
school children in four Chinese communities
chosen for their PM gradient. Frequency of
respiratory symptoms (cough, phlegm, wheeze,
and diagnosed asthma, bronchitis, or
pneumonia) assessed by questionnaire.
Covariates included parental occupation,
education and smoking. The analysis used
logistic regression, controlling for age, sex,
parental smoking, use of coal in home, and
home ventilation.
Results suggested that the OR's for
cough, phlegm, persistent cough
and phlegm and wheeze increased
as outdoor TSP concentrations did. .
Wheeze produced largest OR for both
mothers and fathers in all locations.
Results not directly related to
pollution levels, but symptom rates
were highest in highest pollution
area for cough, phlegm,
hospitalization for respiratory
disease, bronchitis, and pneumonia.
No gradient correlating with
pollution levels found for the three
lower exposure communities.
o
>— i
H
m
-------
1 decreased between surveys as did the prevalence of all respiratory symptoms (including
2 bronchitis). Also, Kramer et al. (1999) reported a study in six East and West Germany
3 communities, which found yearly decreasing TSP levels to be related to ever-diagnosed
4 bronchitis from 1991-1995. Lastly, Jedrychowski et al. (1999) reported an association between
5 both BS and SO2 levels in various areas of Krakow, Poland, and slowed lung function growth
6 (FVC and FEV,).
7
8 6.3.3.2.3 Summary of Long-Term Particulate Matter Exposure Respiratory Effects
9 The methodology used in the long-term studies varies much more than the methodology in
10 the short-term studies. Some studies reported highly significant results (related to PM) while
11 others reported no significant results. The cross-sectional studies are often confounded, in part,
12 by unexplained differences between geographic regions. The studies that looked for a time trend
13 are also confounded by other conditions that were changing over time. Probably the most
14 credible cross-sectional study remains that described by Dockery et al. (1996) and Raizenne et al.
15 (1996). This study, reported in the previous 1996 PM AQCD, found differences in peak flow
16 and bronchitis rates associated with fine particle strong acidity. Whereas most studies included
17 only two to six communities, this study included 24 communities. The effective sample size for
18 a cross sectional analysis is the number of communities, so that six or fewer communities allow
19 many fewer degrees of freedom by which to test hypotheses about various pollutant effects.
20 Newly available studies since the 1996 PM AQCD, overall, provide evidence consistent
21 with the findings from the above 24-City Study. Most notably several U.S. and European studies
22 report associations between PM measures and bronchitis rates and/or lung function decrements
23 or slowed lung function growth. One also provided evidence of PM effects on immune function
24 in school children, with stronger associations for fine particle indicators than for ambient coarse
25 particles.
26
27
March 2001 6-215 DRAFT-DO NOT QUOTE OR CITE
-------
1 6.4 INTERPRETIVE ASSESSMENT OF EPIDEMIOLOGIC DATABASE
2 ON HEALTH EFFECTS OF AMBIENT PARTICIPATE MATTER
3 6.4.1 Introduction
4 As noted at the outset of this chapter, numerous PM epidemiology studies assessed in the
5 1996 PM AQCD implicated ambient PM as a likely key contributor to mortality and morbidity
6 effects observed epidemiologically to be associated with ambient air pollution exposures. Since
7 preparation of the last previous PM AQCD in 1996, the epidemiologic evidence concerning
8 ambient PM-related health effects has expanded greatly. The most important types of additions
9 to the database beyond that assessed in the 1996 PM AQCD, as evaluated above are:
10 • New multi-city studies on a variety of endpoints providing more precise effects estimates than
11 most smaller-scale individual city studies;
12 • More studies of various health endpoints using ambient PM10 and/or closely related mass
13 concentration indices (e.g., PM13 and PM7), which substantially lessen the need to rely on
14 non-gravimetric indices (e.g., BS or COH);
15 • New studies on a variety of endpoints for which information on the ambient PM coarse fraction
16 (PM(10_2 5)), the ambient fine-particle fraction (PM2 5), and even ambient ultrafine particle mass
17 measures (PM0 [ and smaller) were observed and/or estimated from site-specific calibrations;
18 • A few new studies in which the relationship of some health endpoints to ambient particle
19 number concentrations were evaluated;
20 • Many new studies which evaluated the sensitivity of estimated PM effects to the inclusion of
21 gaseous co-pollutants in the model;
22 • Preliminary attempts to evaluate the effects of air pollutant combinations or mixtures including
23 PM components, based on empirical combinations (e.g., factor analysis) or source profiles;
24 • Numerous new studies of cardiovascular endpoints, with particular emphasis on assessment of
25 cardiovascular risk factors as well as symptoms;
26 • Additional new studies on asthma and other respiratory conditions potentially exacerbated by
27 PM exposure;
28 • New studies of infants and children as a potentially susceptible population.
29 The vast majority of the new PM epidemiology studies, both of short-term and long-term PM
30 exposure, continue to show statistically significant excess mortality risk and/or morbidity
March 2001 6-216 DRAFT-DO NOT QUOTE OR CITE
-------
1 endpoints to be associated with ambient PM indexed by a variety of ambient community
2 monitoring methods in many U.S. cities and elsewhere.
3 Several methodological issues, discussed in the 1996 PM AQCD, are still of much
4 importance in assessing and interpreting the overall PM epidemiology database and its
5 implications for estimating risks associated with exposure to ambient PM concentrations in the
6 United States. The fundamental issue essentially subsuming all of the other modeling issues is
7 the selection of an appropriate statistical model in the absence of any strong prior hypotheses or
8 information about underlying relationships between health outcomes and ambient PM, other
9 copollutants, or other important factors such as seasonal variations and/or demographic
10 characteristics of study populations. These critical methodological issues are: (1) potential
11 confounding of PM effects by co-pollutants (especially major gaseous pollutants such as O3, CO,
12 NO2, SO2); (2) the attribution of PM effects to specific PM components (e.g., PM,0, PM10_25,
13 PM2 5, ultrafmes, sulfates, metals, etc.) or source-oriented indicators (motor vehicle emissions,
14 vegetative burning, etc.); (3) the temporal relationship between exposure and effect (lags,
15 /mortality displacement, etc.); (4) the general shape of exposure-response relationship(s) between
16 PM and/or other pollutants and observed health effects (e.g., potential indications of thresholds
17 for PM effects); and (5) the consequences of measurement error. In addition, the newer multi-
18 city study results, e.g, the NMMAPS analysis of the 90 largest U.S. cities (Samet et al., 2000a,b)
19 show evidence of more geographical heterogeneity in the estimated PM risks across cities and
20 regions than had been seen in the studies assessed in the 1996 PM AQCD. Thus, the issue of
21 geographical heterogeneity in PM effects estimates also warrants further evaluation here.
22 Assessing the above issue(s) in relation to the PM epidemiology data base remains quite a
23 challenge. The basic issue is that there are an extremely large number of possible models, any of
24 which may turn out to give the best statistical "fit" of a given set of data, and only some of which
25 can be dismissed a priori as biologically or physically illogical or impossible, except that
26 putative cause clearly cannot follow effect in time. Most of these models are fitted in a stepwise
27 manner, first by removing effects known almost certainly to be present, including general
28 changes in death rates or hospital admissions rates over long time intervals and across season, by
29 day of week, and attributable to weather and climate. Many of the temporal and weather variable
30 models have been fitted to data using semi-parametric methods such as spline functions or local
31 regression smoothers (loess). The goodness of fit of these base models has been evaluated by
March 2001 6-217 DRAFT-DO NOT QUOTE OR CITE
-------
1 criteria suitable for generalized linear models with Poisson or hyper-Poisson responses (number
2 of events) with a log link function, particularly the Akaike Information Criterion (AIC) and the
3 more conservative Bayes or Schwarz information criterion (BIC), that adjust for the number of
4 parameters estimated from the data. The Poisson over-dispersion index and the auto-correlation
5 of residuals are also often used. Once a best-fitting baseline model is selected, the specification
6 of variables in the base model is often held fixed while a better model is selected using one or
7 more PM indices (e.g., fine and coarse) and/or one or more gaseous co-pollutants. In general,
8 one would expect that the best-fitting models with PM would be models with the largest and
9 most significant PM indices. If PM effects are confounded with those of other pollutants, then a
10 large positive estimated PM effect might be associated with a non-biological estimated negative
11 effect for one or more other criteria pollutants, as found by some analyses for NO2 in a joint
12 pollutant model (most likely a statistical artifact). Also, if high correlations between PM and one
13 or more gaseous pollutants emitted from a common source (e.g., motor vehicles) exist in a given
14 area, then disentangling their relative individual partial contributions to observed health effects
15 associations becomes very difficult. Unfortunately, there have been very few attempts at broad,
16 systematic investigations of the model selection issue and little reporting in published reports of
17 goodness-of-fit criteria among competing models that provide a better basis by which to better
18 assess or compare models.
19 Given the now extremely large number of published epidemiologic studies of ambient PM
20 associations with health effects in human populations and the considerably wide diversity in
21 applications of even similar statistical approaches (e.g., "time-series analyses" for short-term PM
22 exposure effects), it is neither feasible nor useful here to try to evaluate the methodological
23 soundness of every individual study. Rather, two feasible approaches are likely to yield useful
24 evaluative information: (1) an overall characterization of evident general commonalities (or
25 notable marked differences) among findings from across the body of studies dealing with
26 particular PM exposure indices and types of health outcomes; and (2) more thorough, critical
27 assessment of key newly published multi-city analyses of PM effects, given that greater scientific
28 weight is likely ascribable to their results than those of smaller sized studies (often of individual
29 cities) yielding presumably less precise effects estimates.
30
31
March 2001 6-218 DRAFT-DO NOT QUOTE OR CITE
-------
1 6.4.2 New Assessments of Confounding
2 As discussed previously, the issue of potential confounding by weather was extensively
3 examined in two studies as reviewed in the 1996 PM AQCD, and was considered essentially
4 resolved. Potential confounding by co-pollutants, however, was nevertheless still suggested by
5 several studies reviewed in the 1996 AQCD. Therefore, discussion of confounding in this
6 section is focused on potential confounding among PM and other major gaseous air pollutants as
7 evaluated in newly available studies.
8
9 6.4.2.1 Assessment of Copollutant Confounding
10 Analyses of one city's data by different researchers may produce conflicting results. For
11 example, Moolgavkar and Luebeck (1996) and Samet et al. (1996) or Kelsall et al. (1997),
12 (which presented essentially the same results) analyzed Philadelphia mortality data for nearly the
13 same period (1973-1988 and 1974-1988, respectively), but produced somewhat different results
14 and interpretations. The notable differences in findings in these studies were: (a) NO2 in the
15 Samet et al. 's study was mostly negatively associated (except summer) with mortality, while in
16 the Moolgavkar-Luebeck study, NO2 was mostly positively associated (except winter); and (b) O3
17 in Samet et al.'s study was positively associated with mortality across seasons (weakest in the
18 summer), while in the Moolgavkar-Luebeck study, O3 was positively associated with mortality
19 only in the summer. The differences may have been due to the difference in the optimum lags
20 chosen for pollutants (in Samet et al., concurrent day levels were used for all the pollutants
21 except CO; whereas, in the Moolgavkar-Luebeck study, one-day lag was used for all pollutants
22 except NO2). Moolgavkar-Luebeck concluded that "..it is not possible with the present evidence
23 to show a convincing correlation between particulate air pollution and mortality", while Samet's
24 group concluded "...These analyses confirm the association between TSP and mortality found in
25 previous studies in Philadelphia and the association is robust to consideration of other
26 pollutants".
27 Such discrepancies could, in part, result from instability of regression coefficients due to
28 collinearity of co-pollutants, as well as model specification choice. The collinearity problem may
29 be further complicated by different seasonal patterns of concentrations for each pollutant, which
30 also vary from city to city. Thus, evaluation of apparently inconsistent results from one or a few
31 cities analyzed using different model specifications, without quantitative information on city
March 2001 6-219 DRAFT-DO NOT QUOTE OR CITE
-------
1 specific characteristics, is unlikely to yield useful information by which to resolve the issue of
2 confounding. By analyzing multiple cities' data, a more consistent pattern may emerge, although
3 differences in approach may still result in inconsistent multi-city results by different researchers.
4 Several studies have examined the issue of confounding using multi-city analyses. Basic
5 descriptions of these studies were provided in earlier text and in Table 6-1; some of their more
6 salient results regarding confounding by co-pollutants are discussed below.
7 Samet and co-workers (2000a,b) reported PM,0 RR estimates for PM10-only and multiple
8 pollutant models that also included O3 as the only gaseous pollutant or O3 and another gaseous
9 pollutant, in both 20-cities analysis and 90-cities analysis. The effects of adding gaseous
10 pollutants in the model on PM10 coefficients were similar in these two data sets, in that adding O3
11 did not change PM10 coefficients, but additional inclusions of another gaseous pollutant reduced
12 PM10 coefficients somewhat. Figure 6-10 shows the posterior probability results for the 90-cities
13 analysis. It can be seen that the PM10 coefficient reduced from about 0.47 to 0.35 when another
14 gaseous pollutant was included in the model besides O3. Importantly, however, the posterior
15 probabilities that the overall effects are greater than 0 remain 1.0 in all these models. It should
16 also be noted that the results shown in the figure are for PM10 at lag 1 day (of the 0-, 1-, and
17 2-day lags examined, the 1-day lag was most significant). The lags for the gaseous pollutants
18 included in these models were also apparently 1-day lags. This choice of the same lags seems
19 reasonable, as the air pollution variables are generally highly correlated with no lag. However,
20 using the most significant lags for gaseous pollutants might have produced somewhat different
21 results. That is, even though air pollution variables may be highly correlated, or not, at 0 lag,
22 various health effects possibly due to different pollutants may occur with different lag times.
23 The HEI Health Committee Review Panel commentary on the NMMAPS analyses stated
24 that an important consideration in assessing the validity of the observed PM10 effects is whether
25 they are due to PM10 itself or due to another air pollutant that is correlated with PM10. That is, do
26 effects of other pollutants confound the observed PM10 effect? The NMMAPS investigators took
27 a commonly used approach to address this issue in the mortality analysis: does the addition of
28 other air pollutant concentrations to the PM,0 regression models result in any substantial change
29 in the estimated PM10 effect? If the PM10 effect does not change, the other pollutants presumably
30 have not confounded the observed PM10 effect. The Panel identified a few issues related to
31 possible confounding effects by co-pollutants, but concluded that the probable impact of any of
March 2001 6-220 DRAFT-DO NOT QUOTE OR CITE
-------
0.0 0.2 0.4
% Change in Mortality per 10 pg/m3 Increase in PM10
Figure 6-10. Marginal posterior distributions for effect of PM10 on total mortality at lag 1
with and without control for other pollutants, for the 90 cities. The numbers
in the upper right legend are the posterior probabilities that the overall
effects are greater than 0.
Source: Samet et al. (2000a,b).
1 these was not considered to be sufficiently large to alter the observed PM10 effect. For example,
2 when the investigators controlled for co-pollutants, they assumed the co-pollutants effect in the
3 model to be linear.
4 Another consideration is the impact of limiting assessment of the possible confounding
5 effect to the relevant season for pollutants that have seasonal patterns. This assessment is
6 complicated in these data because the seasonal effect of ozone, for example, is assumed to be
7 somewhat different across the cities. Finally, a pollutant (e.g., sulfate or acid aerosols) for which
8 only inadequate data are available in the AIRS database and which, therefore, could not be
March 2001
6-221
DRAFT-DO NOT QUOTE OR CITE
-------
1 analyzed, might be more clearly delineated as responsible for the effects attributed to PM10,
2 per se.
3 Even given the above considerations, the HEI Review Panel nevertheless agreed that, in the
4 NMMAPS 20 cities analysis of potential copollutant confounding, no convincing evidence was
5 found for PM10 effects on mortality being changed by addition of either O3, SO2, NO2, or CO to
6 the models, suggesting that none of the other pollutants is responsible for the observed PM10
7 effects. Subsequent analyses by the investigators, that appear to use similar statistical techniques,
8 controlled for gaseous pollutants in the 90 cities and also did not show a confounding copollutant
9 effects.
10 In the morbidity analysis, based on assessment of the likelihood of confounding by other
11 pollutants in stage 2 of the modeling for 14 of the NMMAPS cities, there was evidence that the
12 PM10 effect on each diagnosis was not confounded, similar to the finding in the mortality analysis
13 (but differences in the approach make it difficult to assess whether morbidity findings are as
14 robust). While the approach used in the morbidity analysis is novel (comparing the PM10
15 regression coefficient with the regression coefficient between PM10 and the co-pollutants), the
16 question arises as to the adequacy of statistical power for performing these analyses. Power may
17 be low because the regression is fit to only 14 locations, and in some cases 12 locations, and
18 when the regression coefficients between PM10 and the potentially confounding co-pollutants are
19 similar across cities.
20 The HEI commentary further noted that although NMMAPS focuses on the effects of PM10,
21 examination of the independent effects of other pollutants is also warranted. Effects on daily
22 mortality were found for most of the gaseous pollutants (SO2, CO, NO2 ) in the 20 cities,
23 although these effects were generally diminished when the model controlled for PM10 and other
24 pollutants. In contrast, the PM10 effect did not appear to be affected by other pollutants in this
25 model. An effect of each pollutant except ozone on mortality in the 90 cities is shown in the
26 NMMAPS II Report. A relatively strong effect appears to be present for each of those gaseous
27 pollutants in 90 cities in analyses that assess the effect of each pollutant alone. Thus, findings on
28 independent effects of the gaseous pollutants based on the 20 cities should be viewed as
29 preliminary until a 90 city analysis specifically controlling for PM10 and other pollutants is
30 available. Evaluation of independent effects of the gaseous pollutants on hospitalizations would
31 also be useful in follow-up analyses.
March 2001 6-222 DRAFT-DO NOT QUOTE OR CITE
-------
1 Schwartz (2000a), in his analysis of 10 U.S. cities' data (see New Multi-City Studies and
2 Table 6-1 for basic study description), took an approach that is different from the usual
3 simultaneous inclusion of co-pollutants in the model to address confounding. He postulated that,
4 if the PM10 effect is really due to confounding by another pollutant, one would expect a larger
5 PM10 effect in cities or seasons where PM10 represents more of that other pollutant (i.e., where
6 the slope of the confounder to PM,0 is larger). This approach relied on the large variability in the
7 relationship between PM10 and possibly confounding gaseous pollutants across the 10 cities.
8 Schwartz first illustrated this idea with a simulated example. In the analysis of the real 10 cities'
9 data, the PM,0 coefficients obtained from city-specific analyses were regressed on the regression
10 coefficients relating the gaseous pollutant to PM10 in each city. If the PM)0 effects were due to
11 confounding only, according to the model, then such regression would result in zero intercept.
12 To accommodate greater differences in the gaseous pollutants mean level between the indoor
13 heating season and the warm season, the city-specific regressions were conducted by season,
14 producing 20 city-specific coefficients. The results indicated that the resulting intercepts (i.e.,
15 PM10 effects after controlling for confounders) were not substantially different from that without
16 the gaseous pollutants (0.57, 0.90, and 0.69, for confounding by SO2, CO, and O3, respectively,
17 vs. 0.67 percent excess mortality deaths per 10 /ug/m3 increase in PM10). While this approach
18 appears to be reasonable, it is not certain if the data had sufficient and relevant signals to reflect
19 actual difference in exposures to PM10 vs. gaseous co-pollutants across cities and seasons. For
20 example, a high SO2 to PM10 slope in winter may not be as relevant to a high SO2 to PM10 slope
21 in summer, because of the lower air exchange rate in winter. Such air exchange rate would also
22 vary from city to city, possibly further blurring the relevant exposure picture. Also, the gaseous
23 pollutant's slope on PM,0 can be influenced by error in both PM]0 and the gaseous pollutants.
24 While such slopes may be more accurately estimated for spatially more uniform pollutants such
25 as O3 (and fine component of PM in the summer in northeast), for primary pollutants such as CO
26 and SO2, local source impacts may have contributed to less precision for their slopes on PM10.
27 In Schwartz's analysis of 10 U.S. cities noted earlier, the new approach was not used to
28 examine the changes in the gaseous pollutants' mortality effects coefficients. Such an analysis
29 would have been useful in providing an overall assessment of possible confounding among the
30 air pollutants. However, such an analysis was conducted in Schwartz's analysis (2000b) of
31 Philadelphia data for 1974-1988. In this analysis, the same approach to test confounding was
March 2001 6-223 DRAFT-DO NOT QUOTE OR CITE
-------
1 applied for both TSP and SO2. Instead of using the variability in PM]0 to gaseous pollutants
2 relationships across cities, this Philadelphia data analysis used the changing relationship between
3 TSP and SO2 over the 15-year period. The mortality data were thus analyzed for warm and cold
4 seasons of each year, yielding 30 regression coefficients for both TSP and SO2. Regression
5 coefficients for TSP on SO2, as well as SO2 on TSP, were also obtained for each period. In the
6 second stage regression, the 30 TSP mortality coefficients were regressed on the regression
7 coefficients of SO2 on TSP, and vice versa. In addition, visual range-derived extinction
8 coefficient (an indicator of fine particles) was analyzed as a confounder for TSP in the same
9 manner. The results indicated that the RR for SO2 was substantially reduced (from 1.12 to
10 1.02 per 50 ppb SO2 increase) by controlling for TSP, whereas TSP RR was not reduced, but
11 rather increased, by controlling for SO2 (1.09 to 1.21 per 100 /ug/m3 TSP increase). However, the
12 TSP RR was reduced (1.09 to 1.01) when the extinction coefficient was included in the model.
13 Therefore, the author concluded that the association between air pollution and daily deaths in
14 Philadelphia was due to fine combustion particles.
15 A very different approach to co-pollutant modeling was used by Schwartz in the NMMAPS
16 Part II morbidity analyses, and in a recently published paper on mortality (Schwartz, 2000a).
17 The method attempts to identify total or partial confounding of a nominal causal pollutant such
18 as PM10 with a co-pollutant or other confounder, based on the relationship between a regression
19 of the health effect on the nominal causal pollutant, as a function of the regression coefficient of
20 the nominal causal pollutant against the designated co-pollutant. If the relationship has zero
21 intercept, then one might infer that the two pollutants are totally confounded, with no direct
22 effect of the causal pollutant on the health endpoint that is not mediated by the co-pollutant.
23 If the relationship has a non-zero intercept, then the causal pollutant and the co-pollutant are
24 partially confounded, with the causal pollutant having a direct effect as well as an effect mediated
25 by the co-pollutant. A non-zero intercept and no relation to the co-pollutant slope implies that
26 only a direct health effect exists with the causal pollutant, with no confounding by the
27 co-pollutant.
28 Figures 32 and 33 [not shown here] in NMMAPS II (pp. 40-41) appear to show the
29 expected outcome described above. The vertical axes on both of these figures show the risk
30 estimates for cardiovascular disease, COPD, or pneumonia in a single-pollutant model in 12 to
31 14 cities, with a causal pollutant Z = PMI0. There is no statistically significant relationship
March 2001 6-224 DRAFT-DO NOT QUOTE OR CITE
-------
1 between the estimated PM10 effect on health, and the slopes between Z = PM10 and X = one of
2 the covariables temperature, relative humidity, SO2, or O3. Visual examination of these figures
3 suggests that the high-risk estimates for pneumonia vs. the SO2 or O3 slopes in Figure 33 occur in
4 Colorado Springs, a relatively small city with very little correlation between co-pollutants and
5 PM. Similarly, the high-risk estimates for COPD vs. the PM-O3 slope in Figure 33 occur in
6 Boulder, another relatively small Colorado city with very little correlation between co-pollutants
7 and PM. Absent these three points, there is no relationship between the estimated PM]0 and the
8 regression slopes between PM10 and one of temperature, relative humidity, SO2 or O3, and
9 certainly not a linear relationship, which implies only a direct relationship with PM10.
10 The analogous mortality study (Schwartz, 2000a) does not provide as much detail as the
11 morbidity study in NMMAPS II, 2000. Schwartz (p. 566) notes: "For all three cooccurring
12 pollutants, the effect size after controlling for confounding was not substantially (or statistically
13 significantly) different from the baseline result. This is illustrated in Figure 3." Figure 4 [not
14 shown here] plots each of eighteen city-season pairs, showing little or no relationship. The lack
15 of a relationship does not, however, necessarily confirm that there is no confounding. A more
16 complete implementation of this intriguing approach would be of interest. Until that time,
17 however, the potential effects of confounding should be examined by several different
18 approaches, included the estimated correlation matrix among all of the estimated regression
19 coefficients.
20 In summary, the above results from several multi-city studies using different approaches
21 suggest that possible confounding influences of gaseous pollutants on PM indices are not
22 substantial. However, the interpretation of the relative impact of the gaseous co-pollutants as
23 putative causative agents requires caution and warrants further, more detailed evaluation.
24
25 6.4.2.2 Simulation Analysis of Confounding
26 Since no single model specification can a priori be designated as "correct" in addressing
27 confounding effects of co-pollutants, discrepancies in results among studies, even for the same
28 dataset, are expected. While any assessment of relative "adequacy" of these alternative model
29 specifications is difficult with observational data, the implication of "inadequate" model
30 specifications may be studied through simulations using synthetic data in which the "correct"
31 model is known. Chen et al. (1999) conducted such simulations using a synthetic data set in
March 2001 6-225 DRAFT-DO NOT QUOTE OR CITE
-------
1 which the causal variables are known, and the effects of model misspecification were studied in
2 the presence of two variables (x, and x2), with varying level of correlation, in a Poisson model.
3 They considered three situations: (1) model under/it, in which mortality was generated with both
4 x, and x2, but regressed only on x,; (2) model over/it, in which mortality was generated with only
5 x,, but regressed on both x, and x2; (3) model misfit, in which mortality was generated with either
6 x, or x2 but regressed on the other variable. They observed that the confounding of covariates in
7 an overfitted model does not bias the estimated coefficients but does reduce their significance;
8 and that the effect of model underfit or misfit leads to not only erroneous estimated coefficients
9 but also to erroneous significance. Chen et al., based on these observations, suggested that
10 "models which use only one or two air quality variables (such as PM10 and SO2) are probably
11 unreliable, and that models containing several correlated and toxic or potentially toxic air quality
12 variables should also be investigated...". While conceptually useful, this simulation study
13 ignored one factor that is crucial in evaluating the implication of confounding, the relative error.
14 For example, including several correlated pollutants in a regression model may lead to erroneous
15 inferences unless one considers the relative error associated with each of the pollutants. Several
16 simulation studies that considered such relative errors are discussed below.
17
18 6.4.2.3 Alternative Approaches to Deal with Confounding
19 In time-series analyses of the acute effects of PM, the usual approach to deal with gaseous
20 co-pollutants is to treat them as confounders and to simply include them simultaneously in
21 regression models. There has even been a suggestion (as mentioned above) based on a
22 simulation analysis of synthetic data, that "several" correlated pollutants should be included in
23 regression models (Chen et al., 1999). This prevailing approach can not only lead to misleading
24 conclusions in "identifying" a specific "causal" pollutant (e.g., when pollutants have a varying
25 extent of exposure error), but also ignores the potential combined effects of PM and gaseous
26 co-pollutants (e.g., when PM absorbs SO2 and carries it deeper in the airways, as shown by
27 Amdur and Chen, 1989).
28 Another potential problem of the simultaneous inclusion of PM and gaseous pollutants is
29 that the gaseous pollutant in question may be coming from the same source, or that the PM
30 constituent may be derived from the gaseous pollutants. For example, SO2 can be converted to
31 sulfate, which is a PM constituent. Since a confounder cannot be an intermediate step in the
March 2001 6-226 DRAFT-DO NOT QUOTE OR CITE
-------
1 causal pathway (Rothman and Greenland, 1998), strictly speaking, SO2 does not qualify as a
2 confounder of PM, except in a situation where the PM is known to be solely of secondary origin
3 (transported aerosols) and SO2 is solely from local origin. Furthermore, any reduction in
4 emission of a gaseous pollutant may also affect the level of PM. In such a case, the inference
5 drawn from the results of simultaneous regression may be misleading, because the relative risk
6 for PM is based on the assumption that the covariates could be kept unchanged while the PM
7 level changes.
8 Alternative approaches are needed to address the above noted weakness in the general
9 practice of effect estimation using simple simultaneous regressions. Several alternative
10 approaches have been tried in recent years to estimate the effects of air pollution. The studies
11 include: (1) Ozkaynak et al.'s (1996) analysis of Toronto, CN data; (2) Laden et al.'s (2000)
12 analysis of the Harvard Six Cities PM2 5 data; (3) Mar et al.'s (2000) analysis of Phoenix, AZ
13 PM25data; and, (4) Tsai et al.'s (2000) analysis of 3 New Jersey cities (Newark, Camden, and
14 Elizabeth) data. These studies, as previously described in this chapter, utilized factor analysis to
15 identify underlying factors that could be characterized in terms of source types. This approach
16 thus greatly lessens or prevents inclusion of correlated individual variables in the regression
17 model (depending on the rotation approach used) and also allows source-oriented evaluation of
18 health impacts of PM components (as discussed more specifically below in Section 6.4.3).
19 Factor analysis has been routinely used in the air pollution source apportionment field, but
20 its application to evaluation of PM health effects is relatively new. It may be a useful alternative
21 approach for a source-oriented evaluation of the combined effects of fine particles and gaseous
22 co-pollutants. The advantages of the factor analysis approach include: (a) it allows an
23 examination of association between a health outcome and a group(s) of pollutants that vary
24 together (due to the same source type); (b) it reduces multi-collinearity in a regression model; and
25 (c) it may reduce error associated with individual variables. On the other hand, factor analyses
26 can also be vulnerable to several problems: (a) the factors are sometimes quite sensitive to the
27 inclusion or exclusion of variables from the initial correlation matrix; (b) the minor factors are
28 very sensitive to the number of factors considered in the analysis; and (c) the inclusion of
29 variables with other sources of variation (measurement error, other artifacts, or physical
30 properties) can have a major impact on the selection of factors. With respect to the latter
31 problem, there are valid concerns about studies that include both numerous PM elements
March 2001 6-227 DRAFT-DO NOT QUOTE OR CITE
-------
1 determined by XRF and gaseous pollutants (CO, NO2, O3) in the initial correlation matrix. There
2 are also additional issues in assessing results from factor analysis studies, including the
3 "interpretability" of resulting factors and technical issues (such as approach used for rotation of
4 factors). Thus, there are still some issues that need to be further investigated.
5
6 6.4.3 Role of Participate Matter Components
7 In the 1996 PM AQCD, extensive epidemiologic evidence substantiated very well positive
8 associations between ambient PM10 concentrations and various health indicators, e.g., mortality,
9 hospital admissions, respiratory symptoms, pulmonary function decrements, etc.. A somewhat
10 more limited number of studies were then available which substantiated mortality and morbidity
11 associations with various fine particle indicators (e.g., PM2 5, sulfate, H+, etc.); and only one, the
12 Harvard Six Cities analysis by Schwartz et al. (1996a), evaluated relative contributions of the
13 fine PM2 5 versus coarse (PM10.2 5) fraction of PM10, with PM2 5 appearing to be associated more
14 strongly with mortality effects than PM]0.2 5. Lastly, only a very few studies seemed to be
15 indicative of possible coarse particle effects, e.g., increased asthma risks associated with quite
16 high PM10 concentrations in a few locations where coarse particles strongly dominated the
17 ambient PM10 mix.
18
19 6.4.3.1 Fine-and Coarse-Particle Effects on Mortality
20 A greatly enlarged and still rapidly growing number of new studies published since the
21 1996 PM AQCD provide much new evidence further substantiating ambient PM associations
22 with increased human mortality and morbidity. As indicated in Table 6-1, most newly reported
23 analyses, with few exceptions, continue to show statistically significant associations between
24 short-term (24-h) PM concentrations and increases in daily mortality in many U.S. and Canadian
25 cities, as well as elsewhere. Also, the reanalyses of Harvard Six City and ACS study data
26 substantiate the original investigator's findings of long-term PM exposure associations with
27 increased mortality as well.
28
29 6.4.3.1.1 Effects on Total Mortality
30 The effects estimates from the newly reported studies generally comport well with those
31 derived from the earlier 1996 PM AQCD assessment, which reported risk estimates for excess
March 2001 6-228 DRAFT-DO NOT QUOTE OR CITE
-------
1 total (nonaccidental) deaths associated with short-term PM exposures as generally falling within
2 the range of ca. 1.5 to 8.5% per 50 /ug/m3 PM10 (24-h) increment and ca. 2.5 to 5.5% increase per
3 25 //g/m3 PM2 5 (24-h) increment.
4 Several new PM epidemiology studies which conducted time-series analyses in multiple
5 cities were noted to be of particular interest, in that they provide evidence of effects across
6 various geographic locations (using standardized methodologies) and more precise pooled effect
7 size estimates with narrow confidence bounds reflecting the typically much stronger power of
8 such multi-city studies over individual-city analyses. Based on pooled analyses across multiple
9 cities, the percent total (non-accidental) excess deaths per 50 Aig/m3 PM10 increment were
10 estimated in different multi-city analyses to be: (1) 2.3% in the 90 largest U.S. cities; (2) 3.4% in
11 10 U.S. cities; (3) 3.5% in the 8 largest Canadian cities; and (4) 2.0% in Western European cities.
12 Many new individual-city studies found positive associations (most statistically significant
13 at p < 0.05) for the PM2 5 fraction, with effect size estimates typically ranging from ca. 2.0 to ca.
14 8.5% per 25 Aig/m3 PM2 5 for U.S. and Canadian cities. Of the 10 or so new analyses that not
15 only evaluated PM10 effects but also made an effort to compare fine versus coarse fraction
16 contributions to total mortality, only two are multi-city analyses yielding pooled effects
17 estimates: (1) the Klemm and Mason (2000) recomputation of Harvard Six Cities data,
18 confirming the original published findings by Schwartz et al. (1996a); and (2) the Burnett et al.
19 (2000) study of the 8 largest Canadian cities. Both of these studies found roughly comparable,
20 statistically significant excess risk estimates for PM2 5, i.e., ca. 3% increased total mortality per
21 25 //g/m3 PM25 increment.
22 With regard to possible coarse particle short-term exposure effects on mortality, in those
23 new studies which evaluated PM]0_2 5 effects as well as PM2 5 effects, the coarse particle (PM10_2 5)
24 fraction was also consistently positively associated with increased total mortality, albeit the
25 coarse effect size estimates were generally less precise than those for PM2 5 and statistically
26 significant at p < 0.05 in only a few studies. Still, the overall picture tends to suggest that excess
27 total mortality risks may well reflect actual coarse particle effects, in at least some locations.
28 This may be most consistently the case in arid areas (e.g., Southern California, the Phoenix area,
29 Mexico City, and Santiago, Chile) and during summer months (perhaps reflecting, in part,
30 stronger contributions of biogenic materials to coarse fraction PM10_2 5 particles during warmer
31 weather). On the other hand, significant (or nearly significant) elevations in coarse PM-related
March 2001 6-229 DRAFT-DO NOT QUOTE OR CITE
-------
1 total mortality risks elsewhere (e.g., eastern U.S. urban areas in the Harvard Six City Study, the
2 8 largest Canadian cities, and Detroit, MI) may either reflect (a) typically moderate correlations
3 there between PM10_2 5 and PM2 5 or, possibly, (b) true PM coarse fraction toxic potency. Excess
4 total mortality risks associated with short-term (24-h) exposures to coarse fraction particles
5 capable of depositing in the lower respiratory tract generally fall in the range of 0.5 to 6.0% per
6 25 Acg/m3 PM10.25 increment for U.S. and Canadian cities.
7 Three new papers provide particularly interesting new information on relationships between
8 short-term and coarse particle exposures and total elderly mortality (age 65 and older) using
9 exposure TEOM data from the EPA ORD NERL monitoring site in Phoenix, AZ. Each used
10 quite different models but each reported statistically significant relationships between mortality
11 and coarse PM, specifically PM,0.2 5, an indicator for the thoracic fraction of coarse-mode PM.
12 Smith et al. (2000) using as the exposure metric a three-day running average performed
13 linear regression of the square root of daily mortality on the long-term trend, meteorological and
14 PM-based variables. Two mortality variables were used, total (non-accidental) deaths for the city
15 of Phoenix and the same for a larger, regional area. Using a linear analysis, effects based on
16 coarse PM were statistically significant for both regions, whereas effects based on fine PM
17 (PM2 5) were not. However, when the possibility of a nonlinear response was taken into account,
18 no evidence was found for a nonlinear effect for coarse PM, but fine PM was found to have a
19 statistically significant effect for concentration thresholds of 20 and 25 yUg/m3. There was no
20 evidence of confounding between fine and coarse PM, suggesting that fine and coarse PM are
21 "essentially separate pollutants having distinct effects". Smith et al. (2000) also observed a
22 seasonal effect for coarse PM, the effect being statistically significant only during spring and
23 summer. Based on a principal component analysis of elemental concentrations, crustal elements
24 are highest in spring and summer and anthropogenic elements lowest, but Smith et al. (2000) felt
25 that the implication that crustal, rather than anthropogenic elements, were responsible for the PM
26 mortality was counterintuitive.
27 Clyde et al. (2000) used a more conventional model, a Poisson regression of log deaths on
28 linear PM variables; but they employed Bayesian model averaging to consider a wide variety of
29 variations in the basic model. They considered three regions, the Phoenix metropolitan area, a
30 small subset of zip code to give a region presumably with uniform PM2 5, and a still smaller zip
31 code region surrounding the monitoring site, thought to be uniform as to PM10 concentrations.
March 2001 6-230 DRAFT-DO NOT QUOTE OR CITE
-------
1 The models considered lags of 0, 1, 2, or 3 days but only for single day PM variables (no running
2 averages as used by Smith et al., 2000). A PM effect with a reasonable probability was found
3 only in the uniform PM2 5 region and only for coarse PM.
4 Mar et al. (2000) used conventional Poisson regression methods and limited their analyses
5 to the smallest area (called Uniform PM,0 by Clyde et al). They reported modeling data for lag
6 days 0 to 4. Coarse PM was marginally significant on lag day 0. No direct fine particle measures
7 were statistically significant on day 0. A regional sulfate factor determined from source
8 apportionment, however, was statistically significant. No correlations were reported for the
9 source apportionment factors but the correlation coefficient between sulfur (S) in PM2 5 (as
10 measured by XRF) with coarse PM was only 0. 13, suggesting separate and distinct effects for
1 1 regional sulfate and coarse PM.
12 The above three studies of PM- total mortality relationships in Phoenix tend to suggest a
13 statistical association of coarse PM with total elderly mortality in addition to and different from
14 any relationship with fine PM, fine PM components, or source factors for fine PM.
1 5 With regard to long-term PM exposure effects on total (non-accidental) mortality, the
1 6 newly available evidence from the HEI Reanalyses of Harvard Six Cities and ACS data (and
1 7 extensions, thereof), substantiate well associations attributable to chronic exposures to inhalable
1 8 thoracic particles (indexed by PM,5 or PM10) and the fine fraction of such particles (indexed by
1 9 PM2 5 and/or sulfates). Statistically significant excess risk for total mortality was shown by the
20 reanalyses to fall in the range of 4-1 8% per 20 //g/m3 PM15/10 increment and 14-28% per
21 20 /ug/m3 PM2 5 increase, thus suggesting likely stronger associations with fine versus coarse
22 fraction particles. Significant fine PM associations with total mortality were also found in the
23 latest reported AHSMOG results for males, but not in females.
24 Other recent studies on the relation of mortality to particle composition and source (Laden
25 et al., 2000; Mar et al., 2000; Ozkaynak et al., 1996; Tsai et al., 2000) suggest that particles from
26 certain sources may have much higher potential for adverse health effects than others, as
27 delineated by source-oriented evaluations involving factor analyses. Laden et al. (2000)
28 conducted factor analyses of the elemental composition of PM2 5 for Harvard Six Cities study
29 data for 1 979-1 988. In the analysis for all six cities combined, the excess risk for daily mortality
30 was estimated to be 3.4% (CI, 1.7 to 5.2) per 10 ,ug/m3 increment in a mobile source factor; 1.1%
3 1 (CI, 0.3 to 2.0) per 10 ^g/m3 for a coal source factor, and -2.3% (CI, -5.8 to 1 .2) per 10
March 200 1 6-23 1 DRAFT-DO NOT QUOTE OR CITE
-------
1 for a crustal factor. There was large variation among the cities and some suggestion of an
2 association with a fuel oil factor identified by V or Mn, but it was not statistically significant.
3 Mar et al. (2000) applied factor analysis to evaluate mortality in relation to 1995-1997 fine
4 particle elemental components and gaseous pollutants (CO, NO2, SO2) in an area of Phoenix, AZ,
5 close to the air pollution monitors. The PM2 5 constituents included sulfur, Zn, Pb, soil-corrected
6 potassium, organic and elemental carbon, and a soil component estimated from oxides of Al, Si,
7 Ca, Fe, and It. Based on models fitted using one pollutant at a time, statistically significant
8 associations were found between total mortality and PMIO, CO (lags 0 and 1), NO2 (lags 0, 1, 3,
9 4), S (negative), and soil (negative). Statistically significant associations were also found
10 between cardiovascular mortality and CO (lags 0 to 4), NO2 (lags 1 and 4), SO2 (lags 3 and 4),
11 PM2 5 (lags 1,3,4), PM10 (lag 0), PMi0_2 5 (lag 0), and elemental, organic, or total carbon.
12 Cardiovascular mortality was significantly related to a vegetative burning factor (high loadings
13 on organic carbon and soil-corrected potassium), motor vehicle exhaust/resuspended road dust
14 factor (with high loadings on Mn, Fe, Zn, Pb, OC, EC, CO, and NO2), and a regional sulfate
15 factor (with a high loading on S). However, total mortality was negatively associated with a soil
16 factor (high loadings on Al, Fe, Si) and a local SO2 source factor, but was positively associated
17 with the regional sulfate factor.
18 Tsai et al. (2000) analyzed daily time series of total and cardiorespiratory deaths, using
19 short periods of 1981 -1983 data for Newark, Elizabeth, and Camden, NJ. In addition to
20 inhalable particle mass (PM15) and fine particle mass (PM2 5), the study evaluated data on metals
21 Pb, Mn, Fe, Cd, V, Ni, Zn, Cu, and three fractions of extractable organic matter. Factor analyses
22 were carried out using the metals, CO, and sulfates. The most significant sources or factors
23 identified as predictors of daily mortality were oil burning (targets V, Ni), Zn and Cd processing,
24 and sulfates. Other factors (dust, motor vehicles targeted by Pb and CO, industrial Cu or Fe
25 processing) were not significant predictors. In Newark, oil burning sources and sulfates were
26 positive predictors, and Zn/Cd a negative predictor for total mortality. In Camden oil burning
27 and motor vehicle emissions predicted total mortality, but copper showed a marginal negative
28 association. Oil burning, motor vehicle emissions, and sulfates were predictors of
29 cardiorespiratory mortality in Camden. In Elizabeth, resuspended dust indexed by Fe and Mn
30 showed marginal negative associations with mortality, as did industrial sources traced by Cu.
March 2001 6-232 DRAFT-DO NOT QUOTE OR CITE
-------
1 The set of results from the above factor analyses studies do not yet allow one to identify
2 with great certainty a clear set of specific high-risk chemical components of PM. Nevertheless,
3 some commonalities across the studies seem to highlight the likely importance of mobile source
4 and other fuel combustion emissions (and apparent lesser importance of crustal particles) as
5 contributing to increased total or cardiorespiratory mortality.
6
7 6.4.3.1.2 Effects on Cause-Specific Mortality
8 Numerous new studies have evaluated PM-related effects on cause-specific mortality.
9 Most all report positive, often statistically significant (at p < 0.05), short-term (24-h) PM
10 exposure associations with cardiovascular (CVD)- and respiratory-related deaths. Cause-specific
11 effects estimates appear to mainly fall in the range of 3.0 to 7.0% per 25 jug/m3 24-h PM2 5 for
12 cardiovascular or combined cardiorespiratory mortality and 2.0 to 7.0% per 25 /wg/ni3 24-h PM2 5
13 for respiratory mortality in U.S. cities. Effect size estimates for the coarse fraction (PM10.2 5) for
14 cause-specific mortality generally fall in the range of ca. 3.0 to 8.0% for cardiovascular and ca.
15 3.0 to 16.0% for respiratory causes per 25 jUg/m3 increase in PM]0_2 5.
16 Also of particular interest, the above noted study by Mar et al. examined the associations of
17 a variety of PM indicators with cardiovascular mortality (for age >65), again in the zip code area
18 near the Phoenix monitoring site. For this end point, coarse PM was statistically significant on
19 lag day 0 but not on subsequent lag days. PM2 5 and a number of fine PM indicators were
20 statistically significant on lag day 1 but not on lag day 0. This suggests a distinct and separate
21 relationship of PM2 5 and PM10_2 5. As in the case of total mortality, the only fine PM indicator
22 found to be statistically significant on lag day 0 was regional sulfate. However, the low
23 correlation coefficient between S in PM2 5 and PM10.2 5 (r = 0.13) suggests that the two
24 relationships represent different sets of deaths. Thus, there is some evidence suggesting that the
25 risk of cardiovascular mortality , as well as that of total mortality, may be statistically associated
26 with PM10_2 5 and that this relationship may be independent of any relationships with fine particle
27 indicators.
28
29 6.4.3.2 Fine- and Coarse-Particulate Matter Effects on Morbidity
30 At the time of the 1996 PM AQCD, fine particle morbidity studies were mostly limited to
31 Schwartz et al. (1994), Neas et al. (1995, 1994); Koenig et al. (1993); Dockery et al. (1996); and
March 2001 6-233 DRAFT-DO NOT QUOTE OR CITE
-------
1 Raizenne et al. (1996); and discussion of coarse particles morbidity effects was also limited to
2 only a few studies (Gordian et al., 1996; Hefflin et al., 1991) which implicated PM10_2 5 a possible
3 important fraction of PM10. Since the 1996 PM AQCD, several new studies have been published
4 in which newly available size-fractionated PM data allowed investigation of the effects of both
5 fine (PM2 5) and coarse (PM10_2 5) particles. Fine (FP) and coarse (CP) particle results are
6 presented below for studies by morbidity outcome areas, as follows: cardiovascular disease
7 (CVD) hospital admissions (HA's), respiratory medical visits and hospital admissions, and
8 respiratory symptoms and pulmonary function changes.
9 As discussed in Section 6.3.1 (on cardiovascular effects associated with acute ambient PM
10 exposure), extensive evidence for significant PM,0 effects on cardiovascular-related hospital
11 admissions and visits has recently been provided by several new multi-city studies (Schwartz,
12 1999; Samet et al., 2000a,b; Zanobetti et al., 2000b) that yield pooled estimates of PM-CVD
13 effects across numerous U.S. cities and regions. These studies found not only significant PM
14 associations, but also associations with other gaseous pollutants as well, thus hinting at likely
15 independent effects of certain gases (O3, CO, NO2, SO2) and/or interactive effects with PM.
16 These and other individual-city studies generally appear to confirm likely excess risk of
17 CVD-related hospital admission for U.S. cities in the range of 3-10% per 50 pig/in3 PM10,
18 especially among the elderly (> 65 yr).
19 In addition to the PM10 studies, several new U.S. and Canadian studies evaluated fine-mode
20 PM effects on cardiovascular outcomes. Moolgavkar (2000a) reported PM2 5 to be significantly
21 associated with CVD HA for lag 0 and 1 in Los Angeles. Burnett et al. (1997b) reported that fine
22 particles were significantly associated with CVD HA in a single pollutant model, but not when
23 gases were included in multipollutant models for the 8 largest Canadian city data. Stieb et al.
24 (2000) reported both PM10 and PM2 5 to be associated with CVD emergency department (ED)
25 visits in single pollutant, but not multipollutant models. Similarly, Morgan et al. (1998) reported
26 that PM2 5 measured by nepholonetry was associated with CVD HA for all ages and 65+ yr, but
27 not in the multipollutant model. Tolbert et al. (2000a) reported that coarse particles were
28 significantly associated with dysrhythmias, whereas PM2 5 was not. Other studies (e.g., Liao
29 et al., 1999, Pope et al., 1999b,c) reported associations between increases in PM25 and several
30 measures of decreased heart rate variability, but Gold et al. (1998, 2000) reported a negative
31 association of PM2 5 with heart rate and decreased variability in r-MSSD (one heart rate
March 2001 6-234 DRAFT-DO NOT QUOTE OR CITE
-------
1 variability measure). Overall, these studies appear to implicate fine particles, as well as possibly
2 some gaseous copollutants, in cardiovascular morbidity, but the relative contributions of fine
3 particles acting alone or in combination with gases such as O3, CO, NO2 or SO2 remain to be
4 more clearly delineated and quantified. The most difficult issue relates to interpretation of
5 reduced PM effect size and /or statistical significance when copollutants derived from the same
6 source(s) as PM are included in multipollutant models.
7 Section 6.3.1 also discussed U.S. and Canadian studies that present analysis of coarse
8 particles (CP) relationships to CVD outcomes. Lippmann et al. (2000) found significant positive
9 associations of coarse particles (PMIO_2 5) with ischemic heart disease hospital admissions in
10 Detroit (RR = 1.10, CI 1.026, 1.18). Tolbert et al. (2000a) reported significant positive
11 associations of heart dysrhythmias with CP (p = 0.04) as well as for elemental carbon (p =
12 0.004), but these preliminary results must be interpreted with caution until more complete
13 analyses are carried out and reported. Burnett et al. (1997b) noted that CP was the most robust of
14 the particle metrics examined to inclusion of gaseous covariates for cardiovascular
15 hospitalization, but concluded that particle mass and chemistry could not be identified as an
16 independent risk factor for exacerbation of cardiorespiratory disease in this study. Based on
17 another Canadian study, Burnett et al. (1999), reported statistically significant associations for
18 CP in univariate models but not in multipollutant models; but the use of estimated rather than
19 measured PM exposures limits the interpretation of the PM results reported.
20 The collective evidence reviewed above, in general, appears to suggest excess risks for
21 CVD-related hospital admissions of ca. 4.0 to 10% per 25 /ug/m3 PM2 5 or PM10.2 5 increment.
22 Section 6.3.2 also discussed new studies of effects of short-term PM exposure on the
23 incidence of respiratory hospital admissions and medical visits. Several new U.S. and Canadian
24 studies have yielded particularly interesting results suggestive of roles of both fine and coarse
25 particles respiratory-related hospital admissions. In an analysis of Detroit data, Lippmann et al.
26 (2000) found comparable effect size estimates for PM2 5 and PM10_2 5. That is, the excess risk for
27 pneumonia hospital admissions (in no copollutant model) was 13% (CI 3.7, 22) per 25 /ug/m3
28 PM2 5 and 12% (CI 0.8, 24) per 25 ^g/m3 PM10.2 5. Because PM2 5 and PM10.2 5 were not highly
29 correlated, the observed association between coarse particles and health outcomes were possibly
30 not confounded by smaller particles. Despite the greater measurement error associated with
31 PMio-2.5 than with either PM25 and PMIO, this indicator of the coarse particles within the thoracic
March 2001 6-235 DRAFT-DO NOT QUOTE OR CITE
-------
1 fraction was associated with some of the outcome measures. The interesting result is that PMi0.2 5
2 appeared to be a separate factor from other PM metrics, especially given the effect estimates of
3 PM10_2 5 with pneumonia hospital admissions, (lag 1; RR = 1.11, 95% CI: 1.006, 1.233). Burnett
4 et al. (1997b) also reported PM (PM10, PM2 5, and PM10_2 5) associations with respiratory hospital
5 admissions, even with O3 in the model. Notably, the PM10_2 5 association was significant (RR =
6 1.13 for 25 yUg/m3; CI = 1.05 - 1.20); and inclusion of ozone still yielded a significant coarse
7 mass RR = 1.11 (CI = 1.04 -1.19). Moolgavkar et al. (2000) showed the most consistent
8 association for PM10 across lags (0-4d), while PM2 5 yielded the strongest positive PM metric
9 association at lag 3 days. Also, Moolgavkar (2000a) reported that, in Los Angeles, both PM10
10 and PM2 5 yielded both positive and negative associations at different lags for single pollutant
11 models but not in two pollutant models. Delfino et al. (1997) reported that both PM2 5 and PMIO
12 are positively associated with ED visits for respiratory disease. Morgan et al. (1998) reported
13 that PM2 5 estimated from nephelometry yielded a PM2 5 association with COPD HA for 1-hr max
14 PM that was more positive than 24-h average PM2 5.
15 Some new studies appear to substantiate PM associations with asthma-related hospital
16 admissions. For example, Norris et al (1999) reported associations of emergency department
17 visits for asthma in children with both PM2 5 and PM10.2 5. Two other studies presented uniquely
18 different analyses of hospital admissions in the Seattle, Washington area. Sheppard et al. (1999)
19 studied relationships between PM metrics that included PM10.2 5 and non-elderly adult hospital
20 admissions for asthma in the greater Seattle area and reported significant relative rates for PMto,
21 PM25andPM10,25(lagged 1 day). For PMI025, the relative risk was 1.04 (95% CI 1.01, 1.07).
22 In a different analysis, Lumley and Heagerty (1999) examined PM( and PM1(M in the King
23 County, WA (Seattle) area during the same time period but for hospital admissions for overall
24 respiratory disease. Since only a significant hospital admission association was found with PM, 0
25 and not PM10.,, a dominant role by sub-micron particles in PM2 5 - asthma HA association was
26 suggested, but this may not be an appropriate conclusion based on several differences between
27 the study analysis methods and differences between asthma versus respiratory outcome measures
28 used in the two Seattle studies.
29 Several other studies (Chen et al. 2000; Choudhury et al., 1997; Moolgavlar 2000a;
30 Lippsett et al., 1997) report results for areas (e.g., Reno-Sparks, NV; Anchorage, AK; Phoenix,
31 AZ; Santa Clara, CA) where coarse particles tend to constituent a large fraction of PM,0 but no
March 2001 6-236 DRAFT-DO NOT QUOTE OR CITE
-------
1 measures of PMIO_25 were available. These studies showing significant PM10 effects on
2 respiratory hospital admissions provide additional data suggestive of likely coarse particle effects
3 on respiratory morbidity.
4 Thus, although PM10 mass has most often been implicated as the PM pollution index
5 affecting respiratory hospital admissions, the overall collection of new studies reviewed in
6 Section 6.3.2 appear to suggest relative roles for both fine and coarse PM mass fractions, such as
7 PM2 5 and PM10.2 5.
8 Section 6.3.3 assessed relationships between PM exposure on lung function and respiratory
9 symptoms. While most data examine PM10 effects, several studies also examined fine and coarse
10 fraction effects. Schwartz and Neas (2000) report that cough was the only response in which
11 coarse particles appeared to provide an independent contribution to explaining the increased
12 incidence. The correlation between CM and PM25 was moderate (0.41). Coarse particles had
13 little association with evening peak flow. Tiittanen et al. (1999) also reported a significant effect
14 of PM10.2 5 for cough. Thus, cough may be an appropriate outcome related to coarse particle
15 effects. However, the limited data base suggests that further study is appropriate. The report by
16 Zhang, et al. (2000) of an association between coarse particles and the indicator "runny nose" is
17 noted also.
18 For respiratory symptoms and PFT changes, several new asthma studies report relationships
19 with ambient PM measures. The peak flow analyses results for asthmatics tend to show small
20 decrements for both PM10 and PM2 5. Several studies included PM2 5 and PM10 independently in
21 their analyses of peak flow. Of these, Gold et al. (1999), Naeher et al. (1999), Tiittanen et al.
22 (1999), Pekkanen et al. (1997), and Romieu et al. (1996) all found comparable results for PM2 5
23 and PMI0. The study of Peters et al. (1997b) found slightly larger effects for PM2 5. The study of
24 Schwartz and Neas (2000) found larger effects for PM2 5 than for coarse mode particles. Three
25 studies included both PMIO and PM2 5 in their analyses of respiratory symptoms. The studies of
26 Peters et al. (1997b) and Tiittanen et al. (1999) found similar effects for the two PM measures.
27 Only the Romieu et al. (1996) study found slightly larger effects for PM2 5.
28 For non-asthmatics, several studies evaluated PM2 5 effects. Naeher et al. (1999) reported
29 similar AM PEF decrements for both PM2 5 and PM10. Neas et al. (1996) reported a
30 nonsignificant negative association for PEF and PM2 „ and Neas et al. (1999) also reported
31 negative but nonsignificant PEF results. Schwartz and Neas (2000) reported a significantly PM
March 2001 6-237 DRAFT-DO NOT QUOTE OR CITE
-------
1 PEF association with PM2 5, and Tiittanen et al. (1999) also reported negative but nonsignificant
2 association for PEF and PM2 5. Gold et al. (1999) reported significantly PEF results. Schwartz
3 and Neas (2000) reported significant PM2 5 effects relative to lower respiratory symptoms.
4 Tiittanen et al. (1999) showed significant effects for cough and PM2 5 for a 4-day average.
5 Another study, Peters et al. (1997b) in Erfurt in 1992 is unique for two reasons: (1) they
6 studied the size distribution in the range 0.01 to 2.5 ^m and (2) examined the number of
7 particles. They report that the health effects of 5 day means of the number count (NC) for
8 ultrafine particles were larger than those related to the mass of the fine particles. For NC 0.01 -
9 0.1, cough was significant for the same day and the five day mean.
10 In a chronic respiratory disease study of 22-24 North American communities evaluated in
11 the 1996 PM AQCD, Raizenne et al. (1996) found PM2, to be related to a statistically significant
12 FVC deficit of-3.21% (-4.98, -1.41). Dockery et al. (1996) also reported PM2, associations
13 with increased bronchitis; odds ratio = 1.50 (95% CI = 0.91, 2.47).
14 The above new studies offer much more information than was available in 1996. Effects
15 were noted for several morbidity endpoints: cardiovascular hospital admissions, respiratory
16 hospital admissions and cough. Still insufficient data exists from these relatively limited studies
17 to allow strong conclusions at this time as to which size-related ambient PM components may be
18 most strongly related to one or another morbidity endpoints. Very preliminarily, however, fine
19 particles appear to be more strongly implicated in cardiovascular outcomes than are coarse ones,
20 whereas both seem to impact respiratory endpoints.
21
22 6.4.4 The Question of Lags
23 In most of the past air pollution health effects time-series studies, after the basic model (the
24 best model with weather and seasonal cycles as covariates) was developed, several pollution lags
25 (usually 0 to 3 or 4 days) were individually introduced and the most significant lag(s) chosen for
26 the RR calculation. While this practice may bias the chance of finding a significant association,
27 without a firm biological reason to establish a fixed pre-determined lag, it appears reasonable.
28 Due to likely individual variability in response to air pollution, the apparent lags of effects
29 observed for aggregated population counts are expected to be "distributed" (i.e., symmetric or
30 skewed bell-shape). The "most significant lag" in such distributed lags is also expected to
31 fluctuate statistically. The "vote-counting" of the most significant lags reported in the past
March 2001 6-238 DRAFT-DO NOT QUOTE OR CITE
-------
1 PM-mortality studies shows that 0 and 1 day lags are, in that order, the most frequently reported
2 "optimal" lags, but such estimates may be biased because these lags are also likely the most
3 frequently examined ones. Thus, a more systematic approach across different data sets was
4 needed to investigate this issue.
5 The recent Samet et al. (2000b) analysis of the 90 largest U.S. cities provides particularly
6 useful information on this matter. Figure 6-11 depicts Samet et al.'s overall pooled results,
7 showing the posterior distribution of PM10 effects for the 90 cities for lag 0, 1, and 2 days. It can
8 be seen that the effect size estimate for lag Iday is about twice that for lag 0 or lag 2 days, though
9 their distributions overlap. However, a careful examination of Figures 6 and 7 in NMMAPS I
10 suggests that the maximum PM10 effect may occur in different cities with somewhat different lag
11 relationships. In terms of the magnitude of the estimated PMIO effects, Table 6-24, based on
12 NMMAPS I Figure 7 (posterior bivariate distribution for each county; PM10 effect adjusted for
13 O3), suggests that somewhat different patterns may apply in different locations. These data
14 suggest that while lag 1 effects are typically the largest, there may be some situations in which
15 lag 0 or lag 2 effects are larger.
16 The NMMAPS mortality and morbidity analyses, and another HEI-sponsored study on PM
17 components (Lippmann et al., 2000) illustrate three different ways to deal with temporal
18 structure: (1) assume all sites have the same lag, e.g., 1 day, for a given effect; (2) use the lag or
19 moving average giving the largest or most significant effect; and (3) use a flexible distributed lag
20 model, with parameters adjusted to each site.
21 The NMMAPS mortality analyses used the first approach. This approach introduces a
22 consistent response model across all locations. However, since the cardiovascular, respiratory, or
23 other causes of acute mortality usually associated with PM are not at all specific, there is little
24 a priori reason to believe that they must have the same relation to current or previous PM
25 exposures at different sites. The imposed consistency in lag that maximizes the aggregate effect
26 of lag 1 across all cities, in Figure 15-18 and 24 of NMMAPS II, may obscure important regional
27 or local differences for lags other than 1 day.
28 The NMMAPS morbidity studies evaluate 0- and 1-day lags, the moving average of 0 and
29 1-day lags, polynomial distributed lag models, and unrestricted distributed lag models. The
30 first-stage models for each city in the study were fitted for each city, with no restriction as to a
31 consistent model across all cities, and combined across all 14 cities in the second stage as shown
March 2001 6-239 DRAFT-DO NOT QUOTE OR CITE
-------
__
....
— —
LAGO
LAG1
LAG 2
0.99
1.00
0.98
I
-0.2
I
0.0
0.2 0.4 0.6 0.8 1.0
% Change in Mortality per 10 |jg/m3 Increase in PM10
Figure 6-11. Marginal posterior distribution for effects of PM10 on all cause mortality at
lag 0,1, and 2 for the 90 cities. From Samet et al. (2000a,b). The numbers in
the upper right legend are posterior probabilities that overall effects are
greater than 0.
1 in Table 14 and Figure 23 of NMMAPS II. A comparison of the data tabulated in the NMMAPS
2 Report Appendices shows large differences across cities in the apparent magnitude of the PM,0
3 effect, depending on how the PM concentration data over the preceding few days are used.
4 The approach used in Lippmann et al. (2000) and many other studies is to use the model
5 that maximizes some global model goodness-of-fit criterion. This leads to selection of different
6 models at different sites, as might be expected. However, the best-fitting model (for lags, for
7 example) is often the model with the largest or most significant PM10 coefficient. All models for
8 the pollutant(s) of interest are usually compared among themselves only after a preliminary
9 baseline model has been fitted. The baseline model takes into account most of the other
March 2001
6-240
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 6-24. COMPARISON
ANALYSES FOR 0,1, AND
OF PM10 EFFECT SIZES ESTIMATED BY NMMAPS
2 DAY LAGS FOR THE 20 LARGEST U.S. CITIES
County
Ordered PMm effect sizes
10
Los Angeles
New York
Chicago
Dallas/Fort Worth
Houston
San Diego
Santa Ana /Anaheim
Phoenix
Detroit
Miami
Philadelphia
Seattle
San Jose
Cleveland
San Bernardino
Pittsburgh
Oakland
San Antonio
Riverside
Lag 0 < lag 1 « lag 2
Lag 0 = lag 1 » lag 2
Unreadable
Lag 0> lag l,lag 1< lag 2
Lag 0< lag l,lag 1 > lag 2
Lag 0 = lag 1 > lag 2
Lag 0 > lag 1 > lag 2
Lag 0 = lag 1 < lag 2
Lag 0< lag l,lag 1 > lag 2
Lag 0 < lag 1 = lag 2
Lag 0< lag l,lag 1 > lag 2
Lag 0< lag 1, lag 1 > lag 2
Lag 0 > lag 1 = lag 2
Lag 0> lag l,lag 1 < lag 2
Lag 0 > lag 1 = lag 2
Lag 0< lag l,lag 1 > lag 2
Lag 0 < lag 1 = lag 2
Lag 0 = lag 1 < lag 2
Lag 0< lag l,lag 1 > lag 2
1 variables with which PM10 could be plausibly associated, so that the remaining variation in
2 morbidity or mortality that can be explained by including PM,0 indicators with different temporal
3 structures is nearly "orthogonal" or independent of the baseline model. The restriction to the
4 same lag day at all sites certainly increases the precision of that estimate, but possibly at the cost
5 of obscuring different relationships between time of exposure and health effect at other sites.
March 2001
6-241
DRAFT-DO NOT QUOTE OR CITE
-------
1 An additional complication in assessing the shape of a distributed lag is that the apparent
2 spread of the distributed lag may depend on the pattern of persistence of air pollution (i.e.,
3 episodes may persist for a few days), which may vary from city to city and from pollutant to
4 pollutant. If this is the case, fixing the lag across cities or across pollutants may not be ideal, and
5 may tend to obscure important nuances of lag structures that may provide important clues to
6 possible different lags between PM exposures and different cause-specific effects.
7 Thus, it is possible that the extent of lag and its spread may vary depending on the cause of
8 death. For example, Rossi et al. (1999) report that, in their analysis of TSP-cause specific
9 mortality in Milan, Italy, the lags varied for different cause of death (i.e., same day for respiratory
10 infections and heart failure; 3-4 days for myocardial infarction and COPD). Thus, the lag for
11 total mortality may exhibit mixed lags (weighted by the frequency of deaths in each cause).
12 Another example was reported for a recent Mexico City study (Borja-Aburto et al., 1998), in
13 which they found significant PM2 5-total mortality associations for same day and 4-day lag, but
14 not for the intervening 2 to 3 days (percent increases per 25 Aig/m3 were 3.38, -4.00, 1.03, 1.08,
15 3.43, 2.49, for 0 through 5 day lags, respectively). The authors state: "This phenomenon is
16 consistent with both a harvesting of highly susceptible persons on the day of exposure to high
17 pollution levels and a lagged increase in mortality due to delayed effects of reduction of
18 pulmonary defenses, cardiovascular complications, or other homeostatic changes among
19 less-compromised individuals". It is interesting to note that Wichmann et al. (2000) also
20 reported that the most predictive single day effects on mortality for mass concentrations of
21 0.01-2.5 /2 particles were either immediate (0-1 d lag) or delayed (4-5 d lag) for their data from
22 Erfurt, Germany.
23 It should also be noted that if one chooses the most significant single lag day only, and if
24 more than one lag day shows positive (significant or otherwise) associations with mortality, then
25 reporting a RR for only one lag would also underestimate the pollution effects. Schwartz
26 (2000b) investigated this issue, using the 10 U.S. cities data where daily PM10 values were
27 available for 1986-1993. Daily total (non-accidental) deaths of persons 65 years of age and older
28 were analyzed. For each city, a GAM Poisson model adjusting for temperature, dewpoint,
29 barometric pressure, day-of-week, season, and time was fitted. Effects of distributed lag were
30 examined using four models: 1-day mean at lag 0 day; 2-day mean at lag 0 and 1 day; second-
31 degree distributed lag model using lags 0 through 5 days; unconstrained distributed lag model
March 2001 6-242 DRAFT-DO NOT QUOTE OR CITE
-------
1 using lags 0 through 5 days. The inverse variance weighted averages of the ten cities' estimates
2 were used to combine results. The results indicated that the effect size estimates for the
3 quadratic distributed model and unconstrained distributed lag model were similar. Both
4 distributed lag models resulted in substantially larger effect size estimates (7.25% and 6.62%,
5 respectively, as percent excess total death per 50 /wg/m3 increase in PM10) than the single day lag
6 (3.29%) and moderately larger effect size estimates than the two-day average models (5.36%).
7 Samet et al. (2000a,b) also applied 7- and 14-day unconstrained distributed lag models to
8 Chicago, Minneapolis/St. Paul, and Pittsburgh data, and reported that the sum of the 7-day
9 distributed lag coefficients was greater than the estimates based on a single day's value, but the
10 14-day estimate was substantially lower than the 7-day estimate in Chicago and Minneapolis/
11 St. Paul. Thus, it is possible that the usual RR estimate using one lag day may underestimate PM
12 effects.
13
14 6.4.5 New Assessments of Mortality Displacement
15 There have been a few studies that investigated the question of "harvesting", a phenomenon
16 in which a deficit in mortality occurs following days with (pollution-caused) elevated mortality,
17 due to depletion of the susceptible population pool. This issue is very important in interpreting
18 the public health implication of the reported short-term PM mortality effects. The 1996 PM
19 AQCD discussed suggestive evidence observed by Spix et al. (1993) during a period when air
20 pollution levels were relatively high. Recent studies, however, generally typically used data from
21 areas with lower, non-episodic pollution levels.
22 Schwartz (2000c) separated time-series air pollution, weather, and mortality data from
23 Boston, MA, into three components: (1) seasonal and longer fluctuations; (2) "intermediate"
24 fluctuations; (3) "short-term" fluctuations. By varying the cut-off between the intermediate and
25 short term, evidence of harvesting was sought. The idea is, for example, if the extent of
26 harvesting were a matter of a few days, associations between weekly average values of mortality
27 and air pollution (controlling for seasonal cycles) would not be seen. For COPD, Schwartz
28 (2000c) reported evidence indicating that most of the mortality was only displaced by a few
29 weeks; for pneumonia, heart attacks, and all cause mortality, the effect size increased as longer
30 time scales were included. The percent increase in deaths associated with a 25 ^g/m3 increase in
31 PM2 5 increased from 5.3% (95%CI: 6.8, 9.0) to 9.64% (95%CI: 8.2, 11.1).
March 2001 6-243 DRAFT-DO NOT QUOTE OR CITE
-------
1 Schwartz and Zanobetti (2000) used the same approach described above to analyze a larger
2 data set from Chicago, IL for 1988-1993. Total (non-accidental), in-hospital, out-of-hospital
3 deaths, as well as heart disease, COPD, and pneumonia elderly hospital admissions were
4 analyzed to investigate possible PM10"harvesting" effects. GAM Poisson models adjusting for
5 temperature, relative humidity, day-of-week, and season were applied in baseline models using
6 the average of the same day and previous day's PM,0. Seasonal and trend decomposition
7 techniques called STL were applied to the health outcome and exposure data to decompose them
8 into different time-scales (i.e., short-term to long-term), excluding long seasonal cycles (120 day
9 window). The associations were examined with smoothing windows of 15, 30, 45, and 60 days.
10 The effect size estimate for deaths outside hospital was larger than for deaths inside hospital.
11 All cause mortality showed an increase in effect size at longer time scales. The effect size for
12 deaths outside hospital increases more steeply with increasing time scale than that for inside
13 hospital deaths.
14 Zanobetti et al. (2000a) used GAM distributed lag models to help quantify mortality
15 displacement in Milan, Italy, 1980-1989. Non-accidental total deaths were regressed on smooth
16 functions of TSP distributed over the same day and the previous 45 days using penalized splines
17 for the smooth terms and seasonal cycles, temperature, humidity, day-of-week, holidays, and
18 influenza epidemics. The mortality displacement was modeled as the initial positive increase,
19 negative rebound (due to depletion), followed by another positive coefficients period, and the
20 sum of the three phases were considered as the total cumulative effect. TSP was positively
21 associated with mortality up to 13 days, followed by nearly zero coefficients between 14 and
22 20 days, and then followed by smaller but positive coefficients up to the 45th day (maximum
23 examined). The sum of these coefficients was over three times larger than that for the single-day
24 estimate.
25 Zeger et al. (1999) first illustrated, through simulation, the implication of harvesting for PM
26 regression coefficients (i.e., mortality relative risk) as observed in frequency domain. Three
27 levels of harvesting, 3 days, 30 days, and 300 days were simulated. As expected, the shorter the
28 harvesting, the larger the PM coefficient in the higher frequency range. However, in the real data
29 from Philadelphia, the regression coefficients increased toward the lower frequency range,
30 suggesting that the extent of harvesting, if it exists, is not in the short-term range. Zeger
31 suggested that "harvesting-resistant" regression coefficients could be obtained by excluding the
March 2001 6-244 DRAFT-DO NOT QUOTE OR CITE
-------
1 coefficients in the very high frequency range (to eliminate short-term harvesting) and in the very
2 low frequency range (to eliminate seasonal confounding). Since the observed frequency domain
3 coefficients in the very high frequency range were smaller than those in the mid frequency range,
4 eliminating the "short-term harvesting" effects would only increase the average of those
5 coefficients in the rest of the frequency range.
6 Frequency domain analyses are rarely performed in air pollution health effects studies,
7 except perhaps the spectra analysis (variance decomposition by frequency) to identify seasonal
8 cycles. Examinations of the correlation by frequency (coherence) and the regression coefficients
9 by frequency (gain) may be useful in evaluating the potentially frequency-dependent
10 relationships among multiple time series. A few past examples in air pollution health effects
11 studies include: (1) Shumway et al.'s (1983) analysis of London mortality analysis, in which
12 they observed that significant coherence occurred beyond two week periodicity (they interpreted
13 this as "pollution has to persist to affect mortality"); (2) Shumway et al.'s (1988) analysis of
14 Los Angeles mortality data, in which they also found larger coherence in the lower frequency;
15 (3) Ito's (1990) analysis of London mortality data in which he observed relatively constant gain
16 (regression coefficient) for pollutants across the frequency range, except the annual cycle. These
17 results also suggest that associations and effect size, at least, are not concentrated in the very high
18 frequency range.
19 Schwartz (2000c), Zanobetti et al. (2000a), and Zeger et al.'s (1999) results all suggest that
20 the extent of harvesting, if any, is not a matter of only a few days. Other past studies that used
21 frequency domain analyses are also at least qualitatively in agreement with the evidence against
22 the short-term only harvesting. Since very long wave cycles (> 6 months) need to be controlled
23 in time-series analyses to avoid seasonal confounding, the extent of harvesting beyond 6 months
24 periodicity is not possible in time-series study design. While these studies suggest that observed
25 short-term associations are not simply due to short-term harvesting, more data are needed to
26 obtain quantitative estimates of the extent of prematurity of deaths.
27
28 6.4.6 New Assessment of Threshold in Concentration-Response
29 Relationships
30 In the 1996 PM AQCD, the limitations of identifying 'threshold' in the concentration-
31 response relationships in observational studies were discussed including the low data density in
March 2001 6-245 DRAFT-DO NOT QUOTE OR CITE
-------
1 the lower PM concentration range, the small number of quantile indicators often used, and the
2 possible influence of measurement error. Also, a threshold for a population, as opposed to a
3 threshold for an individual, has some conceptual issues that need to be noted. For example,
4 Schwartz (1999) discussed that, since individual thresholds would vary from person to person
5 due to individual differences in genetic level susceptibility and pre-existing disease conditions, it
6 would be almost mathematically impossible for a threshold to exist in the population. The
7 person-to-person difference in the relationship between personal exposure and the concentration
8 observed at a monitor would also add to the variability. Because one cannot directly measure but
9 can only compute or estimate a population threshold, it would be difficult to interpret an
10 observed threshold, if any, biologically. Despite these issues, several studies have attempted to
11 address the question of threshold by analyzing large databases, or by conducting simulations.
12 Cakmak et al. (1999) investigated methods to detect and estimate threshold levels in time
13 series studies. Based on the realistic range of error observed from actual Toronto pollution data
14 (average site-to-site correlation: 0.90 for O3; 0.76 for COH; 0.69 for TSP; 0.59 for SO2; 0.58 for
15 NO2; and 0.44 for CO), pollution levels were generated with multiplicative error for six levels of
16 exposure error (1.0, 0.9, 0.8, 0.72, 0.6, 0.4, site-to-site correlation). Mortality series were
17 generated with three PMIO threshold levels (12.8 /^g/m3, 24.6 /^g/m3, and 34.4 Mg/m3). LOESS
18 with a 60% span was used to observe the exposure-response curves for these 18 combinations of
19 exposure-response relationships with error. A parameter threshold model was also fit using non-
20 linear least squares. Graphical presentations indicate that LOESS adequately detects threshold
21 under no error, but the thresholds were "smoothed out" under the extreme error scenario. Use of
22 a parametric threshold model was adequate to give "nearly unbiased" estimates of threshold
23 concentrations even under the conditions of extreme measurement error, but the uncertainty in
24 the threshold estimates increased with the degree of error. They concluded, "if threshold exists,
25 it is highly likely that standard statistical analysis can detect it".
26 Schwartz and Zanobetti (2000) investigated the presence of threshold by simulation and
27 actual data analysis of 10 U.S. cities: New Haven, CT; Pittsburgh, PA; Birmingham, AL;
28 Detroit, MI; Canton, OH; Chicago, IL; Minneapolis-St. Paul, MN; Colorado Springs, CO;
29 Spokane, WA; and Seattle, WA, where daily PM10 were available for years 1986-1993. First, a
30 simulation was conducted to show that the combining smoothed curves across cities (the authors
31 called this approach "meta-smoothing") could produce unbiased exposure-response curves.
March 2001 6-246 DRAFT-DO NOT QUOTE OR CITE
-------
1 Three hypothetical curves (linear, piecewise linear, and logarithmic curves) were used to generate
2 mortality series in the 10 cities, and GAM Poisson models were used to estimate respective
3 exposure-response curves. Effects of measurement errors were also simulated. In the analysis of
4 actual 10 cities data, GAM Poisson models were fitted, adjusting for temperature, dewpoint, and
5 barometric pressure, and day-of-week. Smooth function of PM10 with the same span (0.7) was
6 used in each of the cities. The predicted values of the log relative risks were computed for
7 2 Mg/m3 increments between 5.5 /ug/m3 and 69.5 /ug/m3 of PM10 levels. Then, the predicted
8 values were combined across cities using inverse-variance weighting. The simulation results
9 indicated that the "meta-smoothing" approach did not bias the underlying relationships for the
10 linear and threshold models, but did result in a slight downward bias for the logarithmic model.
11 Measurement error (additive or multiplicative) in the simulations did not cause upward bias in
12 the relationship below threshold. The threshold detection in the simulation was not very
13 sensitive to the choice of span in smoothing. In the analysis of real data from 10 cities, the
14 combined curve did not show evidence of a threshold in the PMi0-mortality associations.
15 The Smith et al. (2000) study of associations between daily total mortality and PM2 5 and
16 PM,0_25 in Phoenix, AZ (during 1995-1997) also investigated the possibility of a threshold.
17 In the linear model, the authors found that mortality was significantly associated with PM10_2 5,
18 but not with PM2 5. In modeling possible thresholds, they applied: (1) a piecewise linear model
19 in which several possible thresholds were specified; and (2) a B-spline (spline with cubic
20 polynomials) model with 4 knots. Using the piecewise model, there was no indication that there
21 was a threshold for PM10_2 5. However, for PM2 5, the piecewise model resulted in suggestive
22 evidence for a threshold, around 20 to 25 /ug/m3. The B-spline results also showed no evidence
23 of threshold for PM10_2 5, but for PM2 5, a non-linear curve showed a change in the slope around
24 20 /xtg/m3. A further Bayesian analysis for threshold selection suggested a clear peak in the
25 posterior density around 22 /ug/m3. These results, if they in fact reflect reality, make it difficult to
26 evaluate the relative roles of different PM components (in this case, PM2 5 vs. PM,0_2 5).
27 However, the concentration-response curve for PM2 5 presented in this publication suggests more
28 of a U- or V-shaped relationship than the usual "hockey stick" relationship. Such a relationship
29 is, unlike temperature-mortality relationship, difficult to interpret biologically. Because the
30 sample size of this data (=3 years) is relatively small, further investigation of this issue using
31 similar methods but a larger data set is warranted.
March 2001 6-247 DRAFT-DO NOT QUOTE OR CITE
-------
1 Daniels et al. (2000) examined the presence of threshold using the largest 20 U.S. cities for
2 1987-1994. The authors compared three log-linear GAM regression models: (1) using a linear
3 PM10 term; (2) using a cubic spline of PM10 with knots at 30 and 60 ^g/m3 (corresponding
4 approximately to 25 and 75 percentile of the distribution); and, (3) using a threshold model with
5 a grid search in the range between 5 and 200 /wg/m3 with 5 //g/m3 increment. The covariates
6 included in these models are similar to those used by the same research group previously (Kelsall
7 et al., 1997; Samet et al., 2000a,b), including the smoothing function of time, temperature and
8 dewpoint, and day-of-week indicators. Total, cardiorespiratory, and other mortality series were
9 analyzed. These models were fit for each city separately, and for model (1) and (2), the
10 combined estimates across cities were obtained by using inverse variance weighting if there was
11 no heterogeneity across cities, or by using a two-level hierarchical model if there was
12 heterogeneity. The best fit among the models, within each city and over all cities, were also
13 determined using the Akaike's Information Criterion (AIC). The results using the spline model
14 showed that, for total and cardiorespiratory mortality, the spline curves were roughly linear,
15 consistent with the lack of a threshold. For mortality from other causes, however, the curve did
16 not increase until PM10 concentrations exceeded 50 A^g/m3. While the test of heterogeneity
17 indicated that there was considerable heterogeneity in these curves across cities, the shapes of the
18 curves were similar across cities, with no indication of one city unduly influencing the overall
19 estimate of the curves. The hypothesis of linearity was examined by comparing the AIC values
20 across models. The results suggested that the linear model was preferred over the spline and the
21 threshold models. Thus, these results suggest that linear models without a threshold may well be
22 appropriate for estimating the effects of PM,0 on the types of mortality of main interest.
23
24 6.4.7 New Theoretical Assessments of Consequences of Measurement Error
25 Since the 1996 PM AQCD, there have been some advances in conceptual framework
26 development to investigate the effects of measurement error on PM health effects estimated in
27 time-series studies. Several new studies evaluated the extent of bias caused by measurement
28 errors under a number of scenarios with varying extent of error variance and covariance structure
29 between co-pollutants.
30 Zidek et al. (1996) investigated, through simulation, the joint effects of multi-collinearity
31 and measurement error in Poisson regression model, with two covariates with varying extent of
March 2001 6-248 DRAFT-DO NOT QUOTE OR CITE
-------
1 relative errors and correlation. Their error model was of classical error form (W=X+U, where W
2 and X are surrogate and true measurements, respectively, and the error U is normally distributed).
3 The results illustrated the transfer of effects from the "causal" variable to the confounder.
4 However, for the confounder to have larger coefficients than the true predictor, the correlation
5 between the two covariates had to be large (r = 0.9), with moderate error (a > 0.5) for the true
6 predictor, and no error for the confounder in their scenarios. The transfer-of-causality effect was
7 mitigated when the confounder also became subject to error. Another interesting finding that
8 Zidek et al, reported is the behavior of the standard errors of these coefficients: when the
9 correlation between the covariates was high (r = 0.9) and both covariates had no error, the
10 standard errors for both coefficients were inflated by factor of 2; however, this phenomenon
11 disappeared when the confounder had error. Thus, multi-collinearity influences the significance
12 of the coefficient of the causal variable only when the confounder is accurately measured.
13 Zeger et al. (2000) also conducted a mathematical analysis of PM mortality effects in
14 ordinary least square model (OLS) with the classical error model, under varying extent of error
15 variance and correlation between two predictor variables. The error described here was
16 analytical error (e.g., discrepancy between the co-located monitors). In general, they found that
17 positive regression coefficients are only attenuated, but null predictors (zero coefficient) or weak
18 predictors are only able to appear stronger than true positive predictors under unusual conditions:
19 (1) true predictors must have very large positive or negative correlation (i.e., ]r| > 0.9); (2)
20 measurement error must be substantial (i.e., error variance ~ signal variance); and (3)
21 measurement errors must have a large negative correlation. They concluded that estimated FP
22 health effects are likely underestimated, although the magnitude of bias due to the analytical
23 measurement error is not very large.
24 Zeger et al. (1999) illustrated the implication of the classical error model and the Berkson
25 error model (i.e., X = W + U) in the context of time-series study design. Their simulation of the
26 classical error model with two predictors, with various combinations of error variance and
27 correlation between the predictors/error terms, showed results similar to those reported by Zidek
28 et al. (1996). Most notably, for the transfer of the effects of one variable to the other (i.e., error-
29 induced confounding) to be large, the two predictors or their errors need to be substantially
30 correlated. Also, for the spurious association of a null predictor to be more significant than the
March 2001 6-249 DRAFT-DO NOT QUOTE OR CITE
-------
1 true predictor, their measurement errors have to be extremely negatively correlated—a condition
2 not yet demonstrated as occurring in actual air pollution data sets.
3 Zeger et al. also laid out a comprehensive framework for evaluating the effects of exposure
4 measurement error on estimates of air pollution mortality relative risks in time-series studies.
5 The error, the difference between personal exposure and the central station's measurement of
6 ambient concentration was decomposed into three components: (1) the error due to having
7 aggregate rather than individual exposure; (2) the difference between the average personal
8 exposure and the true ambient concentration level; and, (3) the difference between the true and
9 measured ambient concentration level. By aggregating individual risks to obtain expected
10 number of deaths, they showed that the first component of error (the aggregate rather than
11 individual) is a Berkson error, and, therefore is not a significant contributor to bias in the
12 estimated risk. The second error component is a classical error and can introduce bias if there are
13 short-term associations between indoor source contributions and ambient concentration levels.
14 Recent analysis, however, both using experimental data (Mage et al., 1999; Wilson et al., 2000)
15 and theoretical interpretations and models (Ott et al., 2000) indicate that there is no relationship
16 between the ambient concentration and the nonambient components of personal exposure to PM.
17 However, a bias can arise due to the difference between the personal exposure to ambient PM
18 (indoors plus outdoors) and the ambient concentration. The third error component is the
19 difference between the true and the measured ambient concentration. According to Zeger et al.
20 the final term is largely of the Berkson type if the average of the available monitors is an
21 unbiased estimate of the true spatially averaged ambient level.
22 Using this framework, Zeger et al. (2000) then used PTEAM Riverside, CA data to
23 estimate the second error component and its influence on estimated risks. The correlation
24 coefficient between the error (the average population PM10 total exposure minus the ambient
25 PM10 concentration) and the ambient PM10 concentration was estimated to be -0.63. Since this
26 correlation is negative, the flz (the estimated value of the pollution-mortality relative risk in the
27 regression of mortality on z,, the daily ambient concentration) will tend to underestimate the
A
28 coefficient fix that would be obtained in the regression of mortality on xt, the daily average total
29 personal exposure, in a single-pollutant analysis. Zeger et al. (2000) then proceed to assess the
30 size of the bias that will result from this exposure misclassification, using daily ambient
31 concentration, zr As shown in Equation 9, the daily average total personal exposure, xt, can be
March 2001 6-250 DRAFT-DO NOT QUOTE OR CITE
-------
1 separated into a variable component, 0, zp dependent on the daily ambient concentration, z,, and
2 a constant component, 00, independent of the ambient concentration.
3
4 xt = 90 + G.z, + et [9]
5 where £t is an error term.
6
7 If the nonambient component of the total personal exposure is independent of the ambient
8 concentration, as appears to be the case, Equation 9 from Zeger et al. (2000) becomes the
9 regression analysis equation familiar to exposure analysts (Dockery and Spengler, 1981; Ott
10 et al., 2000; Wilson et al., 2000). In this case, 00 gives the average nonambient component of the
11 total personal exposure and 0, gives the ratio of the ambient component of personal exposure to
12 the ambient concentration. (The ambient component of personal exposure includes exposure to
13 ambient PM while outdoors and, while indoors, exposure to ambient PM that has infiltrated
14 indoors.) In this well-known approach to adjust for exposure measurement error, called
i- yv ^
15 regression calibration (Carroll et al., 1995), the estimate of j3x has the simple form f}x = Pz/9\.
16 Thus, for the regression calibration, the value of fix (based on the total personal exposure) does
17 not depend on the total personal exposure but is given by /?z, based on the ambient concentration,
18 times 015 the ratio of the ambient component of personal exposure to the ambient concentration.
19 A regression analysis of the PTEAM data gave an estimate 0, = 0.60.
20 Zeger et al. (2000) use Equation 9, with 60 = 59.95 and 0, = 0.60, estimated from the
21 PTEAM data, to simulate values of daily average personal exposure, x*t, from the ambient
22 concentrations, zt, for PM10 in Riverside, CA, 1987-1994. They then compare the mean of the
vv
23 simulated J3X s, obtained by the series of log-linear regressions of mortality on the simulated x*t,
24 with the normal approximation of the likelihood function for the coefficient Pz from the
25 log-linear regression of mortality directly on zt. The resulting /3>z / (3X = 0.59, is very close to
26 0, = 0.60. Dominici et al. (2000) provide a more complete analysis of the bias in /3Z as an
27 estimate of f3x using the PTEAM Study and four other data sets and a more complete statistical
^ /\
28 model. Their findings were qualitatively similar in that flx was close to j9z/0,. Thus, it appears
March 2001 6-251 DRAFT-DO NOT QUOTE OR CITE
-------
1 that the bias is very close to 0, which depends not on the total personal exposure but only on the
2 ratio of the ambient component of personal exposure to the ambient concentration.
3 Zeger et al. (2000), in the analyses described above, also suggested that the error due to the
4 difference between the average personal exposure and the ambient level (the second error type
5 described above) is likely the largest source of bias in estimated relative risk. This suggestion at
6 least partly comes from the comparison of PTEAM data and site-to-site correlation (the third type
7 of error described above) for PM10 and O3 in 8 US cities. While PM10 and O3 both showed
8 relatively high site-to-site correlation (=0.6-0.9), a similar extent of site-to-site correlation for
9 other pollutants is not necessarily expected. Ito et al. (1998) estimated site-to-site correlations
10 (after adjusted for seasonal cycles) for PM10, O3, SO2, NO2, CO, temperature, dewpoint
11 temperature, and relative humidity, using multiple stations' data from seven central and eastern
12 states (IL, IN, MI, OH, PA, WV, WI), and found that, in a geographic scale of less 100 miles,
13 these variables could be categorized into three groups in terms of the extent of correlation:
14 weather variables (r > 0.9); O3, PM10, NO2 (r: 0.6 - 0.8); CO and SO2 (r < 0.5). These results
15 suggest that the contribution from the third component of error, as described in Zeger et al.
16 (2000), would vary among pollution and weather variables. Furthermore, the contribution from
17 the second component of error would also vary among pollutants; i.e., the ratio of ambient
18 exposure to ambient concentration, called the attenuation coefficient, is expected to be different
19 for each pollutant. Some of the ongoing studies are expected to shed some light on this issue.
20 However, more information is needed on attenuation coefficients for a variety of pollutants.
21 With regard to the PM exposure, longitudinal studies (Wallace, 2000; Mage et al., 1999),
22 show reasonably good correlation (r = 0.6 to 0.9) between ambient PM concentrations and
23 average population PM exposure, lending support for the use of ambient data as a surrogate for
24 personal exposure to ambient PM in time-series mortality or morbidity studies. Furthermore,
25 fine particles are expected to show even better site-to-site correlation than PM10. Wilson and Sun
26 (1997) examined site-to-site correlation of PM10, PM2 5, and PM10.2 5 in Philadelphia and
27 St. Louis, and found that site-to-site correlations were high (r = 0.9) for PM2 5 but low for PM10_2 5
28 (r ~ 0.4), indicating that fine particles have smaller errors in representing community-wide
29 exposures. This finding supports Lipfert and Wyzga's (1997) speculation that the stronger
30 mortality associations for fine particles than coarse particles found in the Schwartz et al. (1996a)
31 study may be due in part to larger measurement error for coarse particles.
March 2001 6-252 DRAFT-DO NOT QUOTE OR CITE
-------
1 However, as Lipfert and Wyzga (1997) suggested, the issue is not whether the fine particle
2 association with mortality is a "false positive", but rather, whether the weaker mortality
3 association with coarse particles is a "false negative". Carrothers and Evans (2000) also
4 investigated the joint effects of correlation and relative error, but they specifically addressed the
5 issue of fine (FP) vs. coarse particle (CP) effect, by assuming three levels of relative toxicity of
6 fine vs. coarse particles (ppp / PCP =1,3, and 10) and, then, evaluating the bias, (B = {E[PF] /
7 E[PC.]} / (PF/ Pc}, as a function of FP-CP correlation and relative error associated with FP and
8 CP. Their results indicate: (1) if the FP and CP have the same toxicity, there is no bias (i.e.,
9 B=l) as long as FP and CP are measured with equal precision, but, if, for example, FP is
10 measured more precisely than CP, then FP will appear to be more toxic than CP (i.e., B > 1);
11 (2) when FP is more toxic than CP (i.e., Ppp / PCP = 3 and 10), however, the equal precision of FP
12 and CP results in downward bias of FP (B < 1), implying a relative overestimation of the less
13 toxic CP. That is, to achieve non-bias, FP must be measured more precisely than CP, even more
14 so as the correlation between FP and CP increases. They also applied this model to real data
15 from the Harvard Six Cities Study, in particular, the data from Boston and Knoxville. Estimation
16 of spatial variability for Boston was based on external data and a range of spatial variability for
17 Knoxville (since there was no spatial data available for this city). For Boston, where the
18 estimated FP-CP correlation was low (r = 0.28), estimated error was smaller for FP than for CP
19 (0.85 vs. 0.65, as correlation between true vs. error-added series), and the observed FP to CP
20 coefficient ratio was high (11), the calculated FP to CP coefficient ratio was even larger (26)-thus
21 providing evidence against the hypothesis that FP is absorbing some of the coefficient of CP.
22 For Knoxville, where FP-CP correlation was moderate (0.54), the error for FP was smaller than
23 for CP (0.9 vs. 0.75), and the observed FP to CP coefficient ratio was 1.4, the calculated true FP
24 to CP coefficient ratio was smaller (0.9) than the observed value, indicating that the coefficient
25 was overestimated for the better-measured FP, while the coefficient was underestimated for the
26 worse-measured CP. Since the amount (and the direction) of bias depended on several variables
27 (i.e., correlation between FP and CP; the relative error for FP and CP; and, the underlying true
28 ratio of the FP toxicity to CP toxicity), the authors concluded "...for instance, it is inadequate to
29 state that differences in measurement error among fine and coarse particles will lead to false
30 negative findings for coarse particles".
March 2001 6-253 DRAFT-DO NOT QUOTE OR CITE
-------
1 Fung and Krewski (1999) conducted a simulation study of measurement error adjustment
2 methods for Poisson models, using scenarios similar to those used in the simulation studies that
3 investigated implication of joint effects of correlated covariates with measurement error. The
4 measurement error adjustment methods employed were the Regression Calibration (RCAL)
5 method (Carroll et al., 1995) and the Simulation Extrapolation (SIMEX) method (Cook and
6 Stefanski, 1994). Briefly, RCAL algorithm consists of: (1) estimation of the regression of X on
7 W (observed version of X, with error) and Z (covariate without error); (2) replacement of X by
8 its estimate from (1), and conducting the standard analysis (i.e., regression); and (3) adjustment
9 of the resulting standard error of coefficient to account for the calibration modeling. SIMEX
10 algorithm consists of: (1) addition of successively larger amount of error to the original data;
11 (2) obtaining naive regression coefficients for each of the error added data sets; and, (3) back
12 extrapolation of the obtained coefficients to the error-free case using a quadratic or other
13 function. Fung and Krewski examined the cases for: (1) Px = 0.25; Pz = 0.25; (2) Px = 0.0;
14 pz = 0.25; (3) px = 0.25; pz = O.O., all with varying level of correlation (-0.8 to 0.8) with and
15 without classical additive error, and also considering Berkson type error. The behaviors of naive
16 estimates were essentially similar to other simulation studies. In most cases with the classical
17 error, RCAL performed better than SIMEX (which performed comparably when X-Z correlation
18 was small), recovering underlying coefficients. In the presence of Berkson type error, however,
19 even RCAL did not recover the underlying coefficients when X-Z correlation was large (> 0.5).
20 This is the first study to examine the performance of available error adjustment methods that can
21 be applied to time-series Poisson regression. The authors recommend RCAL over SIMEX.
22 Possible reasons why RCAL performed better than SIMEX in these scenarios were not discussed,
23 nor are they clear from the information given in the publication. There has not been a study to
24 apply these error adjustment methods in real time-series health effects studies. These
25 methodologies require either replicate measurements or some knowledge on the nature of error
26 (i.e., distributional properties, correlation, etc.). Since the information regarding the nature of
27 error is still being collected at this time, it may take some time before applications of these
28 methods become practical.
29 Another issue that measurement error may affect is the detection of threshold in time-series
30 studies. Lipfert and Wyzga (1996) suggested that measurement error may obscure the true shape
31 of the exposure-response curve, and that such error could make the exposure-response curve to
March 2001 6-254 DRAFT-DO NOT QUOTE OR CITE
-------
1 appear linear even when a threshold may exist. However, based on a simulation with realistic
2 range of exposure error (due to site-to-site correlation), Cakmak et al. (1999) illustrated that the
3 modern smoothing approach, LOESS, can adequately detect threshold levels (12.8 ,ug/m3,
4 24.6 Aig/m3, and 34.4 /^g/m3) even with the presence of exposure error (see also Section 6.4.6
5 above).
6 Other issues related to exposure error that have not been investigated include potential
7 differential error among subpopulations. If the exposure errors are different between susceptible
8 population groups (e.g., people with COPD) and the rest of the population, the estimation of bias
9 may need to take such differences into account. Also, the exposure errors may vary from season
10 to season, due to seasonal differences in the use of indoor emission sources and air exchange
11 rates due to air conditioning and heating. This may possibly explain reported season-specific
12 effects of PM and other pollutants. Such season-specific contributions of errors from indoor and
13 outdoor sources are also expected to be different from pollutant to pollutant.
14 In summary, the studies that examined joint effects of correlation and error suggest that PM
15 effects are likely underestimated, and that spurious PM effects (i.e., qualitative bias such as
16 change in the sign of coefficient) due to transferring of effects from other covariates require
17 extreme conditions and are, therefore, unlikely. Also, one simulation study suggests that, under
18 the likely range of error for PM, it is unlikely that a threshold is ignored by common smoothing
19 methods. More data are needed to examine the exposure errors for other pollutants, since their
20 relative error contributions will influence their relative significance in relative risk estimates.
21
22 6.4.8 New Assessment of Methodological Issues
23 6.4.8.1 Time Series Model Specification
24 Methodological issues in time-series analyses of air pollution-mortality association were
25 discussed extensively in the 1996 PM AQCD. Since then, increasing numbers of researchers
26 have been utilizing essentially the same Poisson regression approach: (1) model seasonal cycles
27 and other temporal trends using smoothing functions of time; (2) model weather effects using
28 smoothing functions of temperature, humidity, and/or their interaction at various lags; (3) after
29 adjustment for these confounding factors, enter various lags (and averaging periods) of air
30 pollutant, and report results for all the lags, and/or report results for the lags that resulted in the
31 highest significance; (4) repeat (3) with other pollutants in the model; (5) conduct sensitivity
March 2001 6-255 DRAFT-DO NOT QUOTE OR CITE
-------
1 analyses using alternative weather model specifications. Seasonal cycles and weather effects are
2 often modeled using Generalized Additive Models (GAM). As the modeling of temporal trends
3 became more efficient using the GAM models, it became clearer that the residual over-dispersion
4 and autocorrelation could be essentially eliminated. Also, more researchers appear to rely on
5 Akaike's Information Criteria (AIC) or on the more conservative Bayes (Schwarz, 1978)
6 Information Criterion (BIC) to choose between models when epidemiological reasoning does not
7 favor one over the other. While these techniques do not necessarily eliminate inadequate model
8 specifications, they do help "standardize" the approaches that researchers can take, reducing the
9 inconsistency in model specification among studies.
10 A few remaining inconsistencies in approach among studies include: (1) choice of the
11 range of lags and averaging periods of pollution included; (2) smoothing spans used for modeling
12 temporal trends and weather effects; (3) the increment used to calculate relative risks; and,
13 (4) choice to detrend pollution variables. The choice of lag can lead to inconsistent results even
14 for the same data. The choice of the combination of lags multiplies as the number of
15 co-pollutants in the model increases. In the case of temperature effects, it has been repeatedly
16 observed that the heat effects tend to be immediate (0 or 1 day lag), while cold effects tend to lag
17 longer (2 to 4 days). For pollutants, however, reported lags are less consistent. The smoothing
18 span for temporal trends can be determined based on epidemiological reasons (i.e., eliminate
19 influenza epidemics), but the span for weather effects may be determined through data
20 exploration. Using the inter-quartile range for all the co-pollutants may be problematic when
21 co-pollutants have inconsistent distributional characteristics. While these issues may appear
22 rather minor, in practice, they may make substantial differences in reported effects and
23 interpretations.
24
25 6.4.8.2 Case-Crossover Study Design
26 Navidi et al. (1999) proposed the use of "bi-directional" controls in applying the case-
27 crossover design to study acute effects of air pollution. In the original case-crossover studies in
28 which risk factors were behavior-related (e.g., coffee consumption), the control period was
29 chosen prior to the case period (i.e., retrospective uni-directional) because choosing the control
30 period after the event would interfere with behavioral modification associated with the risk
31 factor, possibly resulting in bias. In the case of environmental exposures such as ambient air
March 2001 6-256 DRAFT-DO NOT QUOTE OR CITE
-------
1 pollution, however, the event is unlikely to modify future exposure. Furthermore, in the case of
2 observational air pollution study, the bi-directional control periods would be necessary to avoid
3 confounding due to temporal trends in both events (e.g., influenza-related mortality or morbidity)
4 and exposure (natural seasonal trends). Navidi conducted simulations to illustrate that the
5 relative risk estimates are resistant to confounding by time-trend.
6 Bateson and Schwartz (1999) also conducted a simulation study to compare five case-
7 crossover control sampling strategies including the matched pair, a symmetric bi-directional, a
8 total history approach, and the two approaches that Navidi proposed. The symmetric
9 bi-directional approach using 1-week lag estimated the true relative risks correctly in the
10 presence of confounding seasonal trends, whereas the other four approaches failed to control for
11 the confounding trends. They concluded that the bi-directional case-crossover design could
12 control for confounding by design, though it is not as efficient as Poisson time-series analysis.
13 There have been several studies that applied the case-crossover design to analyze air
14 pollution - mortality associations, as described below.
15 Neas et. al. (1999) analyzed Philadelphia TSP data for 1973-1980. Total, age over 65,
16 cancer, and cardiovascular deaths were analyzed for their association with TSP. A conditional
17 logistic regression analysis with a case-crossover design was conducted using the control periods
18 of 7, 14, and 21 days before and after the case period. Other covariates included temperature on
19 the previous day, dewpoint on the same day, an indicator for hot days (> 80°F), an indicator for
20 humid days (dewpoint > 66°F), and interaction between the same-day temperature and winter
21 season. In each set of the six control periods, TSP was associated with total mortality. A model
22 with four symmetric reference periods 7 and 14 days around the case period produced a similar
23 result. A model with only two symmetric reference periods of 7 days around the case produced a
24 larger estimate. A larger effect was seen for deaths in persons > 65 years of age and for deaths
25 due to pneumonia and to cardiovascular disease. Thus, this study basically confirmed the
26 original findings by Schwartz and Dockery (1992) for this city.
27 Sunyer et al. (2000) analyzed Barcelona, Spain BS data for 1990-1995. Those who were
28 over age 35 and had sought emergency room services for COPD exacerbation between 1985 and
29 1989, and had died during 1990-1995 were included in analysis. Total, respiratory, and
30 cardiovascular deaths were analyzed using a conditional logistic regression analysis with a case-
31 crossover design, adjusting for temperature, relative humidity, and influenza epidemics.
March 2001 6-257 DRAFT-DO NOT QUOTE OR CITE
-------
1 Bi-directional control period at 7 days was used. The average of the same and previous 2 days
2 was used for pollution exposure period. Data were also stratified by potential effect modifiers
3 (e.g., age, gender, severity of ER visits, number of ER visits, etc.) and were analyzed. BS levels
4 were associated with all cause deaths. The association was stronger for respiratory causes. Older
5 women, patients admitted to intensive care units, and patients with a higher rate of ER visits were
6 at greater risk of deaths associated with BS.
7 Lee and Schwartz (1999) analyzed data from Seoul, Korea for 1991 -1995. Total deaths
8 were analyzed for their association with TSP, SO2, and O3. A conditional logistic regression
9 analysis with a case-crossover design was conducted. Three-day moving average values (current
10 and two past days) of TSP and SO2, and 1-hr max O3 were analyzed separately. The control
11 periods are 7 and 14 days before and/or after the case period. Both unidirectional and
12 bi-directional controls (7 or 7 and 14 days) were examined, resulting in six sets of control
13 selection schemes. Other covariates included temperature and relative humidity. Among the six
14 control periods, the two unidirectional retrospective control schemes resulted in odds ratios less
15 than 1; the two unidirectional prospective control schemes resulted in larger odds ratios (e.g.,
16 1.4 for 50 ppb increase in SO2); and bi-directional control schemes resulted in odds ratios
17 between those for uni-directional schemes. SO2 was more significantly associated with mortality
18 than TSP. These results suggested that risk estimates were rather sensitive to the choice of the
19 control periods.
20 These analyses suggest that the overall findings are not very sensitive to these analytic
21 choices; thus we can have more confidence in the mortality results. The sensitivity analyses are
22 not as extensive for examining the PM10 effect on morbidity, and the investigators used a
23 different time window across the 14 cities to control for temporal effects. Future analyses of
24 both the mortality and morbidity data might include a seasonally stratified analysis (given the
25 seasonal variability in pollutant concentrations, outcome measures, and potential confounding
26 factors). Loss of statistical power due to the shorter periods of observation in any season should
27 be only a minor issue, at least in the mortality data set.
28
29 6.4.9 Heterogeneity of Particulate Matter Effects Estimates
30 Approximately 35 then-available acute PM exposure community epidemiologic studies
31 were assessed in the 1996 PM AQCD as collectively demonstrating increased risks of mortality
March 2001 6-258 DRAFT-DO NOT QUOTE OR CITE
-------
1 being associated with short-term (24-h) PM exposures indexed by various ambient PM
2 measurement indices (e.g., PM10, PM2 5, BS, COH, sulfates, etc.) in many different cities in the
3 United States and internationally. Much homogeneity appeared to exist across various
4 geographic locations, with many studies suggesting, for example, increased relative risk (RR)
5 estimates for total nonaccidental mortality on the order of 1.025 to 1.05 (or 2.5 to 5.0% excess
6 deaths) per 50 yUg/m3 increase in 24-h PM10, with statistically significant results extending more
7 broadly in the range of 1.5 to 8.0%. The elderly >65 yrs. old and those with preexisting
8 cardiopulmonary conditions had somewhat higher excess risks. One study, the Harvard Six City
9 Study, also provided estimates of increased RR for total mortality falling in the range of 1.02 to
10 1.056 (2.0 to 5.6% excess deaths) per 25 ywg/m3 24-h PM2 5 increment.
11 Now, more than 70 new time-series PM-mortality studies assessed earlier in this chapter
12 provide extensive additional evidence which, qualitatively, largely substantiates significant
13 ambient PM-mortality relationships, again based on 24-h exposures indexed by a wide variety of
14 PM metrics in many different cities of the United States, in Canada, in Mexico, and elsewhere (in
15 South America, Europe, Asia, etc.). The newly available effect size estimates from such studies
16 are reasonably consistent with the ranges derived from the earlier studies reviewed in the 1996
17 PM AQCD. For example, newly estimated PM,0 effects generally fall in the range of 1.0 to 8.0%
18 excess deaths per 50 Aig/m3 PM10 increment in 24-h concentration; whereas new PM2 5 excess
19 estimates for short-term exposures generally fall in the range of 2 to 8% per 25 Mg/m3 increment
20 in 24-h PM2 5 concentration.
21 However, somewhat greater spatial heterogeneity appears to exist across newly reported
22 study results, both with regard to PM-mortality and morbidity effects. The newly apparent
23 heterogeneity of findings across locations is perhaps most notable in relation to reports based on
24 multiple-city studies in which investigators used the same analytical strategies and models
25 adjusted for the same or similar co-pollutants and meteorological conditions, raising the
26 possibility of different findings reflecting real location-specific differences in exposure-response
27 relationships rather than potential differences in models used, pollutants measured and included
28 in the models, etc. Some examples of newly reported and well-conducted multiple-city studies
29 include: the NMMAPS analyses of mortality and morbidity in 20 and 90 U.S. cities (Samet et al.,
30 2000a,b; Dominici et al., 2000); the Schwartz (2000b,c) analyses of 10 U.S. cities; the study of
31 eight largest Canadian cities (Burnett et al., 2000); the study of hospital admissions in eight U.S.
March 2001 6-259 DRAFT-DO NOT QUOTE OR CITE
-------
1 counties (Schwartz, 1999); and the APHEA studies of mortality and morbidity in several
2 European cities (Katsouyanni et al., 1997; Zmirou et al., 1998). The recently completed large
3 NMMAPS studies of morbidity and mortality in U.S. cities add especially useful and important
4 information about potential U.S. within- and between-region heterogeneity.
5
6 6.4.9.1 Evaluation of Heterogeneity of Participate Matter Mortality Effect Estimates
7 In all of the U.S. multi-city analyses, the heterogeneity in the PM estimates across cities
8 was not explained by city-specific characteristics in the 2nd stage model. The heterogeneity of
9 effects estimates across cities in the multi-city analyses may be due to chance alone, to
10 mis-specification of covariate effects in small cities, or to real differences from location to
11 location in effects of different location-specific ambient PM mixes, for which no mechanistic
12 explanations are yet known. Or, the apparent heterogeneity may simply reflect imprecise PM
13 effect estimates derived from smaller-sized analyses of less extensive available air pollution data
14 or numbers of deaths in some cities tending to obscure more precise effects estimates from
15 larger-size analyses for other locations, which tend to be consistently more positive and
16 statistically significant.
17 Some of these possibilities can be evaluated by using data from the NMMAPS study
18 (Samet et al., 2000b). Data in Figure 6-1 were optically scanned and digitized, producing
19 reasonably accurate estimates by comparison with the 20 largest U.S. cities in their Table A-2.
20 The cities were divided among 7 regions, and excess risk with 95% confidence intervals plotted
21 against the total number of effective observations, measured by the number of days of PM10 data
22 times the mean number of daily deaths in the community. This provides a useful measure of the
23 weight that might be assigned to the results, since the uncertainty of the RR estimate based on a
24 Poisson mean is roughly inversely proportional to this product. That is, the expected pattern
25 typically shows less spread of estimated excess risk with increasing death-days of data. A more
26 refined weight index would also include the spread in the distribution of PM concentrations. The
27 results are plotted in Figure 6-12 for all cities and Figure 6-13 for each of the 7 regions.
28 Figure 6-12 for all cities suggests some relationship between precision of the effects
29 estimates and study weight, overall. That is, the more the mortality-days observations, the
30 narrower the 95% confidence intervals and the more precise the effects estimates (with nearly all
31
March 2001 6-260 DRAFT-DO NOT QUOTE OR CITE
-------
All Cities
6 7 8 9 10 11 12 13
Natural Log of Mortality - Days
Figure 6-12. The EPA-derived plot showing relationship of PMi0 total mortality effects
estimates and 95% confidence intervals for all cities in the Samet et al.
(2000a,b) NMMAPS 90-cities analyses in relation to study size (i.e., the
natural logarithm of numbers of deaths times days of PM observations). Note
generally narrower confidence intervals for more homogeneously positive
effects estimates as study size increases beyond about the log 9 value (i.e.,
beyond about 8,000 deaths-days of observation). The dashed line depicts the
overall nationwide effect estimate (grand mean) of approximately 0.5% per
10 (tg/m3 PM10.
March 2001
6-261
DRAFT-DO NOT QUOTE OR CITE
-------
5.4
Q.
1
0. 00
3?
»" -1.8
ct
to
U)
1
.
-'
I 1
Northwest
r I ~"Tr T t
I IF J
i i
•
-
- J
•
j.
•
Industrial Midwest
iflllTT '
• •• | ij-fTi"I""3r
1 •LIi
Northeast
I.
7 8 9 10 11 7 8 9 10 11 12 13 6 7 8 9 10 11 12
Natural Log of Mortality - Days Natura| Log Qf Morta|jty. Dgys Natura| Log Qf Morta|i(y. Dgys
o «•**
0.
o>
a 1.8
o
o> n n
£
to
(A
R
X
"-1 C 4
Southern California
'II I
11
-
[II
1
-
-
- .
i i i i i
Other Midwest
i i
'
7
-
i i
-
J-
i i i
Southeast
r
: •
_
i_
_
-LI
• • ,
n
••
I]
1
"
-JH
tl .
i i
8 9 10 11 12 6 7 8 9 10 11 12 7 8 9 10 11
Natural Log of Mortality - Days Natura| Log of Morta|ity. Days Natura| Log of Morta|ity. Days
5.4
:>
Q.
3.
O
Of.
to
1.8 -
-
-
1 1
Southwest
1 1
i-i
i i
'8.0 8.5 9.0 9.5 10.0
Natural Log of Mortality - Days
Figure 6-13. The EPA-derived plots showing relationships of PMi0-mortality (total,
nonaccidental) effects estimates and 95% confidence intervals to study size
(defined as in Figure 6-10) for cities broken out by regions as per the
NMMAPS regional analyses of Samet et al. (2000a,b). Dashed line on each
plate depicts overall nationwide effect estimate (grand mean) of
approximately 0.5% per 10 /ug/m3 PM,
MO'
March 2001
6-262
DRAFT-DO NOT QUOTE OR CITE
-------
1 these for cities with > log 9 mortality-days being positive and many statistically significant at
2 p <0.05).
3 The Figure 6-13 depiction for each of the 7 regions is also informative. In the Northeast,
4 there is considerable homogeneity (not heterogeneity) of effect size for larger study-size cities,
5 even with moderately wide confidence intervals for those with log mortality-days = 8 to 9, and all
6 clearly exceed the overall nationwide grand mean indicated by the dashed line. On the other
7 hand, the smaller study-size Northeast cities (with much wider confidence intervals at log < 8)
8 show much greater heterogeneity of effects estimates and less precision. Also, most of the
9 estimates for larger study-size (log > 9) cities in the industrial midwest are positive and several
10 statistically significant, so that an overall significant regional risk is plausible there as well.
11 There may even be some tendency for relatively large risks for some cities with small study sizes
12 and wide confidence intervals in the industrial midwest, and further investigation of that would
13 be of interest. The plot for Southern California in Figure 6-13 clearly shows a rather consistent
14 estimate of effect size and width of the confidence intervals across cities of varying study-size.
15 All risk estimates are positive and most are significant at p s 0.05 or nearly so for the Southern
16 California cities. For Northwestern cities plotted in Figure 6-13, the value for Oakland, CA (at
17 ca. log 9.5) is notable (it being very positive and significant), whereas many but not all of the
18 other cities have positive effect estimates not too far off the nationwide grand mean, but with
19 sufficiently wide confidence intervals so as not to be statistically significant at p < 0.05. The
20 Southwestern cities (except for 2 cities), too, mostly appear to have effect sizes near the
21 nationwide mean, but with confidence intervals too wide to be significant at p < 0.05. The
22 "Other" (non-industrial or "Upper", as per NMMAPS) Midwest cities and the Southeastern cities
23 in Figure 6-13 show more heterogeneity, although most of the larger study size cities (log > 9.0)
24 tend to be positive and not far off the nationwide mean (even though not significant at p < 0.05).
25 Given the wide range of effects estimates and confidence intervals seen for Southeastern cities,
26 further splitting of the region might be informative.
27 In fact, closer reexamination of results for each of the regions may reveal interesting new
28 insights into what factors may account for any apparent disparities among the cities within a
29 given region or across regions. Several possibilities readily come to mind. First, cursory
30 inspection of the mean PM10 levels shown for each city in Appendix 6A-2 suggests that many of
31 the cities showing low effects estimates and wide confidence intervals tend to be among those
March 2001 6-263 DRAFT-DO NOT QUOTE OR CITE
-------
1 having the lowest mean PM10 levels and, therefore, likely the smallest range of PM10 values
2 across which to distinguish any PM-related effect, if present. It may also be possible that those
3 areas with higher PM2 5 proportions of PM10 mass (i.e., larger percentages of fine particles) may
4 show higher effects estimates (e.g., in Northeastern cities) than those with higher coarse-mode
5 fractions (e.g., as would be more typical of Southwestern cities). Also, more industrialized cities
6 with greater fine-particle emissions from coal combustion (e.g., in the industrial Midwest) and/or
7 those with high fine-particle emissions from heavy motor vehicle emissions (e.g., typical of
8 Southern California cities) may show larger PM10 effects estimates than other cities. Lastly, the
9 extent of air-conditioning use may also account for some of the differences, with greater use in
10 many Southeastern and Southwestern cities perhaps decreasing actual human exposure to
11 ambient particles present versus higher personal exposure to ambient PM (including indoors) in
12 those areas where less air-conditioning is used (e.g., the Northeast and industrial Midwest).
13
14 6.4.9.2 Comparison of Spatial Relationships in the NMMAPS and Cohort
15 Reanalyses Studies
16 Both the NMMAPS and HEI Cohort Reanalyses studies had a sufficiently large number of
17 U.S. cities to allow considerable resolution of regional PM effects within the "lower 48" states,
18 but an attempt was made to take this approach to a much more detailed level in the Cohort
19 Reanalysis studies than in NMMAPS. There were: 88 cities with PM10 effect size estimates in
20 NMMAPS; 50 cities with PM25 and 151 cities with sulfates in the original Pope et al. (1995)
21 ACS analyses and in the HEI reanalyses using the original data; and 63 cities with PM2 5 data and
22 144 cities with sulfate data in the additional analyses done by the HEI Cohort Reanalysis team.
23 The relatively large number of data points utilized in the HEI reanalyses effort and additional
24 analyses allowed estimation of surfaces for elevated long-term concentrations of PM2 5, sulfates,
25 and SO2 with resolution on a scale of a few tens to hundreds of kilometers.
26 The patterns for PM2 5 and sulfates are similar, but not identical, hi particular, the modeled
27 PM2 5 surface (Krewski et al., 2000; Figure 18) has peak levels around Chicago - Gary, in the
28 eastern Kentucky - Cleveland region, and around Birmingham AL, with elevated but lower PM2 5
29 almost everywhere east of the Mississippi, as well as southern California. This is similar to the
30 modeled sulfate surface (Krewski et al., 2000; Figure 16), with the absence of a peak in
31 Birmingham and an emerging sulfate peak in Atlanta. The only area with markedly elevated SO2
March 2001 6-264 DRAFT-DO NOT QUOTE OR CITE
-------
1 concentrations is the Cleveland - Pittsburgh region. A preliminary evaluation is that secondary
2 sulfates in particles derived from local SO2 are more likely to be important in the industrial
3 midwest, south from the Chicago - Gary region into Ohio, northeastern Kentucky, West Virginia,
4 and southwest Pennsylvania, possibly related to combustion of high-sulfur fuels.
5 The overlay of mortality with air pollution patterns is also of much interest. The spatial
6 overlay of long-term PM2 5 and mortality (Krewski et al., 2000; Figure 21) is highest from
7 southern Ohio to northeastern Kentucky/West Virginia, but also includes a significant association
8 over most of the industrial midwest from Illinois to the eastern non-coastal parts of North
9 Carolina, Virginia, Pennsylvania, and New York. This is reflected, in diminished form, by the
10 sulfates and SO2 maps (Krewski et al., 2000; Figures 19 and 20), where there appears to be a
11 somewhat tighter focus of elevated risk in the upper Ohio River Valley area. This suggests that,
12 while SO2 may be an important precursor of sulfates in this region, there may also be some other
13 (non-sulfur) contributors to associations between PM2 5 and long-term mortality, embracing a
14 wide area of the Northcentral Midwest and non-coastal Mid-Atlantic region.
15 It should be noticed that, while a variety of spatial modeling approaches were discussed in
16 the NMMAPS methodology report (NMMAPS Part I, pp. 66-71 [Samet et al., 2000aj), the
17 primary spatial analyses in the 90-city study (NMMAPS, Part II [Samet et al., 2000b]) were
18 based on a simpler seven-region breakdown of the contiguous 48 states. The 20-city results
19 reported for the spatial model in NMMAPS I show a much smaller posterior probability of a
20 PM10 excess risk of short-term mortality, with a spatial posterior probability vs. a non-spatial
21 probability of a PM10 effect of 0.89 vs. 0.98 at lag 0, of 0.92 vs. 0.99 at lag 1, and of 0.85 vs. 0.97
22 at lag 2. The evidence that PM10 is associated with an excess short-term mortality risk is still
23 moderately strong with a spatial model, but less strong than with a non-spatial model.
24 Even so, there is a considerable degree of coherence between the short-term and long-term
25 mortality findings of the two studies, with strong evidence of a modest but significant short-term
26 PMIO effect and a large long-term fine particle (PM2 5 in general or sulfate) effect in the industrial
27 Midwest. The short-term PM10 effects are large in the Northeast and in Southern California
28 (though less certain there), whereas long-term PM2 5 effects seem to be moderate to high in these
29 areas as well. This may tend to suggest that at least some of the more notable PM10 effects found
30 in the NMMAPS regional analyses may coincide with the presence of higher proportions of fine
31 versus coarse particles in the PM10 mix.
March 2001 6-265 DRAFT-DO NOT QUOTE OR CITE
-------
1 The apparently substantial differences in PM10 and/or PM2 5 effect sizes across different
2 regions should not be attributed merely to possible variations in measurement error or other
3 statistical artifact(s). Some of these differences may reflect: real regional differences in particle
4 composition or co-pollutant mix; differences in relative human exposures to ambient particles or
5 other gaseous pollutants; sociodemographic differences (e.g., percent of infants or elderly in
6 regional population); or other important, as of yet unidentified PM effect modifiers.
7
8
9 6.5 KEY FINDINGS AND CONCLUSIONS DERIVED FROM
10 PARTICULATE MATTER EPIDEMIOLOGY STUDIES
11 It is not possible to assign any absolute measure of certainty to conclusions based on the
12 findings of the epidemiology studies discussed in this chapter. However, these observational
13 study findings would be further enhanced by supportive findings of causal studies from other
14 scientific disciplines (dosimetry, toxicology, etc.), as discussed in Chapters 7 to 9. The most
15 salient conclusions derived from the PM epidemiology studies include:
16 (1) A very large and sufficiently convincing body of epidemiology evidence substantiates
17 strong associations between short- and long-term ambient PM10 exposures (inferred from
18 stationary air monitor measures) and mortality/morbidity effects to conclude that PM10 (or
19 one or more PM10 components) is a probable contributory cause of human health effects.
20 (2) It is likely that there is meaningful heterogeneity in the city-specific excess risk estimates
21 for the relationships between short-term ambient PM10 concentrations and acute health
22 effects. The reasons for such variation in effects estimates are not well understood at this
23 time, but do not negate ambient PM's likely causative contribution to observed PM-
24 mortality and/or morbidity associations in many locations.
25 (3) A smaller (but growing) body of epidemiology evidence is sufficiently indicative of
26 associations between short- and long-term ambient PM2 5 exposures (inferred from
27 stationary air monitor measures) and health effects to conclude that PM2 5 (or one or more
28 PM2 5 components) is a probable contributing cause of observed PM-associated health
29 effects. Some new epidemiology findings also suggest that health effects are associated
30 with mass or number concentrations of ultrafine (nuclei-mode) particles, but not necessarily
31 more so than for other ambient fine PM components.
March 2001 6-266 DRAFT-DO NOT QUOTE OR CITE
-------
1 (4) An even smaller body of evidence exists which appears to support an association between
2 short-term ambient coarse-fraction (PM10_2 5) exposures (inferred from stationary air
3 monitor measures) and short-term health effects in epidemiology studies. This suggests
4 that PM10_2 5, or some constituent component(s) of PM10_2 5, may be a contributory cause of
5 health effects in some locations. Reasons for differences among findings on coarse-particle
6 health effects reported for different cities are still poorly understood, but several of the
7 locations where significant PM10.2 5 effects have been observed (Phoenix, Mexico City,
8 Santiago) tend to be in drier climates and may have contributions to observed effects due to
9 higher levels of organic particles from biogenic processes (endotoxins, molds, etc.) during
10 warm months. Other studies suggest that coarse fraction (PM10.2 5) particles of crustal
11 origin are unlikely to exert notable health effects under most ambient exposure conditions.
12 (5) Long-term PM exposure durations, on the order of months to years, as well as on the order
13 of a few days, are likely associated with serious human health effects (indexed by mortality,
14 hospital admissions/medical visits, etc.). More chronic PM exposures, on the order of
15 years or decades, appear to be associated with life shortening beyond that accounted for by
16 the simple accumulation of the more acute effects of short-term PM exposures (on the order
17 of a few days). While the few studies of this relationship were generally well conducted,
18 notable uncertainties remain regarding the meaning, magnitude, and mechanisms for more
19 chronic health effects of long-term PM exposures. New findings of associations between
20 ambient PM exposures (indexed by various measures) during early pregnancy and/or early
21 post-natally and slowed fetal growth or infant mortality, respectively, suggest potentially
22 much larger life-shortening impacts of PM than previously estimated.
23 (6) Considerable coherence exists among effect size estimates for ambient PM health effects.
24 For example, results derived from several multi-city studies, based on pooled analyses of
25 data combined across multiple cities (thought to yield the most precise effect size
26 estimates), show the percent excess total (non-accidental) deaths estimated per 50 /ug/m3
27 increase in 24-h PM10 to be: 2.3% in the 90 largest U.S. cities (4.5% in the Northeast U.S.
28 region); 3.4% in 10 U.S. cities; 3.5% in the 8 largest Canadian cities; and 2.0% in western
29 European cities (using PM10 = TSP*0.55). These combined estimates are consistent with
30 the range of PM,0 estimates previously reported in the 1996 PM AQCD. These and excess
31 risk estimates from many other individual-city studies, generally falling in the range of ca.
March 2001 6-267 DRAFT-DO NOT QUOTE OR CITE
-------
1 1.5 to 8.0% per 50 /ug/m3 24-h PM10 increment, also comport well with numerous new
2 studies confirming increased cause-specific cardiovascular- and respiratory-related
3 mortality. They are also coherent with larger effect sizes reported for cardiovascular (in the
4 range of ca. 3.0 to 10.0% per 50 Aig/m3 24-h PM10 increment) and respiratory (in the range
5 of ca. 5 to 25% per 50 /wg/m3 24-h PM10) hospital admissions/visits, as would be expected
6 for these morbidity endpoints versus those for PM10-related mortality.
7 (7) Several independent panel studies (but not all) that evaluated temporal associations
8 between PM exposures and measures of heart beat rhythm in elderly subjects provide
9 generally consistent indications of decreased heart rate variability (HRV) being associated
10 with ambient PM exposure (decreased HRV being an indicator of increased risk for serious
11 cardiovascular outcomes, e.g., heart attacks). Other studies point toward changes in blood
12 characteristics (e.g., C-reactive protein levels) related to increased risk of ischemic heart
13 disease also being associated with ambient PM exposures. However, these heart rhythm
14 and blood characteristics findings should currently be viewed as providing only limited or
15 preliminary support for PM-related cardiovascular effects.
16 (8) Notable new evidence now exists which substantiates positive associations between
17 ambient PM concentrations and increased respiratory-related hospital admissions,
18 emergency department, and other medical visits, particularly in relation to PM10 levels.
19 Of much interest are new, but limited, findings tending to implicate not only fine particle
20 components but also coarse (e.g., PMIO_2 5) particles as likely contributing to exacerbation of
21 asthma conditions. Also of much interest are emerging new findings indicative of likely
22 increased occurrence of chronic bronchitis in association with (especially chronic) PM
23 exposure.
24 (9) One major methodological issue affecting epidemiology studies of both short-term and
25 long-term PM exposure effects is that ambient PM of varying size ranges is typically found
26 in association with other air pollutants, including gaseous criteria pollutants (e.g, O3, NO2,
27 SO2, CO), air toxics, and/or bioaerosols. Available statistical methods for assessing
28 potential confounding arising from these associations may not yet be fully adequate. The
29 inclusion of multiple pollutants often produces statistically unstable estimates. Omission of
30 other pollutants may incorrectly attribute their independent effects to PM. Much progress
31 in sorting out relative contributions of ambient PM components versus other copollutants is
March 2001 6-268 DRAFT-DO NOT QUOTE OR CITE
-------
1 nevertheless being made and, overall, tends to substantiate that observed PM effects are at
2 least partly due to ambient PM acting alone or in the presence of other covarying gaseous
3 pollutants.
4 (10) It is likely that differences in observed health effects will be found to depend as much on
5 site-specific differences in chemical and physical composition characteristics of ambient
6 particles as on differences in PM mass concentration. For example, the Utah Valley study
7 (Dockery et al., 1999; Pope et al., 1991, 1999b) showed that PMIO particles, known to be
8 richer in metals during exposure periods while the steel mill was operating, were more
9 highly associated with adverse health effects than was PM10 during the PM exposure
10 reduction while the steel mill was closed. In contrast, PM10 or PM2 5 was relatively higher
11 in crustal particles during windblown dust episodes in Spokane and in three central Utah
12 sites than at other times, but was not associated with higher total mortality. These
13 differences require more research that may become more feasible as the PM2 5 sampling
14 network produces air quality data related to speciated samples.
15 (11) The above reasons suggest it is inadvisable to pool epidemiology studies at different
16 locations, different time periods, with different population sub-groups, or different health
17 endpoints, without assessing the consequences of these differences. Published multi-city
18 analyses using common data bases, measurement devices, and analytical strategies such as
19 those carried out in the APHEA and NMMAPS studies are likely to be useful after careful
20 evaluation. Pooled analyses of more diverse collections of independent studies of different
21 cities, using varying methodology and/or data quality or representativeness, are likely less
22 credible and should not, in general, be used without careful assessment of their underlying
23 scientific comparability.
24 (12) It may be possible that different PM components may produce effects which appear at
25 different lags or that different preexisting conditions may lead to different delays between
26 exposure and effect. Thus, although maximum effect sizes for PM effects have often been
27 reported for 0-1 day lags, evidence is also beginning to suggest that more consideration
28 should be given to lags of several days. Also, if it is considered that all health effects
29 occurring at different lag days are all real effects, so that the risks for each lag day should
30 be additive, then higher overall risks may exist than implied by maximum estimates for any
31 particular single or two-day lags.
March 2001 6-269 DRAFT-DO NOT QUOTE OR CITE
-------
1 (13) Certain classes of ambient particles may be distinctly less toxic than others and may not
2 exert human health effects at typical ambient exposure concentrations or only under special
3 circumstances. For example, particles of crustal origin may be relatively non-toxic under
4 most circumstances compared to those of combustion origin. However, crustal particles
5 contaminated with pesticides or herbicides (as may occur in agricultural situations) or with
6 emissions from vehicles, smelters, or other industrial operations may be sufficiently toxic
7 to cause human health effects under some exposure conditions. More research is needed to
8 identify conditions under which one or another class of particles cause little or no adverse
9 health effects, as well as conditions under which particles cause notable effects.
10 (14) Certain epidemiology evidence suggests that reducing ambient PM,0 concentrations may
11 reduce a variety of health effects on a time scale from a few days to a few months. This has
12 been found in epidemiology studies of "natural experiments" such as in the Utah Valley,
13 and by supporting toxicology studies using the particles from ambient community sampling
14 filters from the Utah Valley. Recent studies in Germany and in the Czech Republic also
15 support a hypothesis that reductions in air pollution are associated with reductions in the
16 incidence of adverse health effects, but these studies cannot unambiguously attribute
17 improved health to reduced PM alone.
18 (15) Adverse health effects in children are emerging as a more important area of concern than in
19 the 1996 PM AQCD. Unfortunately, relatively little is known about the relationship of PM
20 to the most serious health endpoints (low birth weight, preterm birth, neonatal and infant
21 mortality, emergency hospital admissions and mortality in older children). Also, little is yet
22 known about involvement of PM exposure in the progression from less serious childhood
23 conditions, such as asthma and respiratory symptoms, to more serious disease endpoints
24 later in life. This is an important health issue because childhood illness or death may cost a
25 very large number of productive life-years. Lastly, new epidemiologic studies of ambient
26 PM associations with increased non-hospital medical visits (physician visits) and asthma
27 effects suggest likely much larger health impacts and costs to society due to ambient PM
28 than just those indexed by mortality and/or hospital admissions/visits.
29
March 2001 6-270 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Abbey, D. E.; Mills, P. K.; Petersen, F. F.; Beeson, W. L. (1991) Long-term ambient concentrations of total
3 suspended participates and oxidants as related to incidence of chronic disease in California Seventh-Day
4 Adventists. Environ. Health Perspect. 94: 43-50.
5 Abbey, D. E.; Lebowitz, M. D.; Mills, P. K.; Petersen, F. F.; Beeson, W. L.; Burchette, R. J. (1995a) Long-term
6 ambient concentrations of particulates and oxidants and development of chronic disease in a cohort of
7 nonsmoking California residents. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on
8 particulate air pollution and human mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol.
9 7: 19-34.
10 Abbey, D. E.; Ostro, B. E.; Fraser, G.; Vancuren, T.; Burchette, R. J. (1995b) Estimating fine particulates less that
11 2.5 microns in aerodynamic diameter (PM2 5) from airport visibility data in California. J. Exposure Anal.
12 Environ. Epidemiol. 5: 161-180.
13 Abbey, D. E.; Burchette, R. J.; Knutsen, S. F.; McDonnell, W. F.; Lebowitz, M. D.; Enright, P. L. (1998) Long-term
14 particulate and other air pollutants and lung function in nonsmokers. Am. J. Respir. Crit. Care Med.
15 158:289-298.
16 Abbey, D. E.; Nishino, N.; McDonnell, W. F.; Burchette, R. J.; Knutsen, S. F.; Beeson, W. L.; Yang, J. X. (1999)
17 Long-term inhalable particles and other air pollutants related to mortality in nonsmokers. Am. J. Respir. Crit.
18 Care Med. 159:373-382.
19 Ackermann-Liebrich, U.; Leuenberger, P.; Schwartz, J.; Schindler, C.; Monn, C.; Bolognini, B.; Bongard, J. P.;
20 Brandli, O.; Domenighetti, G.; Elsasser, S.; Gnze, L.; Karrer, W.; Keller, R.; Keller-Wossidlo, H.;
21 Kiinzli, N.; Martin, B. W.; Medici, T. C.; Perruchoud, A. P.; Schoni, M. H.; Tschopp, J. M.; Villiger, B.;
22 Wiithrich, B.; Zellweger, J. P.; Zemp, E. (1997) Lung function and long term exposure to air pollutants in
23 Switzerland. Am. J. Respir. Crit. Care Med. 155: 122-129.
24 Adams, P. F.; Hendershot, G. E.; Marano, M. A. (1999) Current estimates from the National Health Interview
25 Survey, 1996. Hyattsville, MD: U.S. Department of Health and Human Services, Public Health Service,
26 National Center for Health Statistics; publication no. 99-1528. (Vital and health statistics: v. 10, no. 200).
27 Available: http://www.cdc.gov/nchs/products/pubs/pubd/senes/srlO/pre-200/pre-200.htm [12 March, 2001].
28 Agocs, M. M.; White, M. C.; Ursicz, G.; Olson, D. R.; Vamos, A. (1997) A longitudinal study of ambient air
29 pollutants and the lung peak expiratory flow rates among asthmatic children in Hungary. Int. J. Epidemiol.
30 26: 1272-1280.
31 Alberdi Odriozola, J. C.; Diaz Jimenez, J.; Montero Rubio, J. C.; Miron Perez, I. J.; Pajares Ortiz, M. S.; Ribera
32 Rodrigues, P. (1998) Air pollution and mortality in Madrid, Spain: a time-series analysis. Int. Arch. Occup.
33 Environ. Health 71: 543-549.
34 Amdur, M. O.; Chen, L. C. (1989) Furnace-generated acid aerosols: speciation and pulmonary effects.
35 In: Symposium on the health effects of acid aerosols; October 1987; Research Triangle Park, NC.
36 Environ. Health Perspect. 79: 147-150.
37 Anderson, H. R.; Ponce de Leon, A.; Bland, J. M.; Bower, J. S.; Strachan, D. P. (1996) Air pollution and daily
38 mortality in London: 1987-92. Br. Med. J. 312: 665-669.
39 Anderson, H. R.; Spix, C.; Medina, S.; Schouten, J. P.; Castellsague, J.; Rossi, G.; Zmirou, D.; Touloumi, G.;
40 Wojtyniak, B.; Ponka, A.; Bacharova, L.; Schwartz, J.; Katsouyanni, K. (1997) Air pollution and daily
41 admissions for chronic obstructive pulmonary disease in 6 European cities: results from the APHEA project.
42 Eur. Respir. J. 10: 1064-1071.
43 Anderson, H. R.; Ponce de Leon, A.; Bland, J. M.; Bower, J. S.; Emberlin, J.; Strachen, D. P. (1998) Air pollution,
44 pollens, and daily admissions for asthma in London 1987-92. Thorax 53: 842-848.
45 Atkinson, R. W.; Anderson, H. R.; Strachan, D. P.; Bland, J. M.; Bremner, S. A.; Ponce de Leon, A. (1999a)
46 Short-term associations between outdoor air pollution and visits to accident and emergency departments in
47 London for respiratory complaints. Eur. Respir. J. 13: 257-265.
48 Atkinson, R. W.; Bremner, S. A.; Anderson, H. R.; Strachan, D. P.; Bland, J. M.; Ponce de Leon, A. (1999b)
49 Short-term associations between emergency hospital admissions for respiratory and cardiovascular disease
50 and outdoor air pollution in London. Arch. Environ. Health 54: 398-411.
51 Awasthi, S.; Glick, H. A.; Fletcher, R. H.; Ahmed, N. (1996) Ambient air pollution & respiratory symptoms
52 complex in preschool children. Indian J. Med. Res. 104: 257-262.
53 Bailey, D. L. R.; Clayton, P. (1982) The measurement of suspended particle and total carbon concentrations in the
54 atmosphere using standard smoke shade methods. Atmos. Environ. 16: 2683-2690.
March 2001 6-271 DRAFT-DO NOT QUOTE OR CITE
-------
1 Baldi, I.; Tessier, J. F.; Kauffmann, F.; Jacqmin-Gadda, H.; Nejjari, C.; Salamon, R. (1999) Prevalence of asthma
2 and mean levels of air pollution: results from the French PAARC survey. Eur. Respir. J. 14: 132-138.
3 Bateson, T. F.; Schwartz, J. (1999) Control for seasonal variation and time trend in case-crossover studies of acute
4 effects of environmental exposures. Epidemiology 10: 539-544.
5 Beeson, W. L.; Abbey, D. E.; Knutsen, S. F. (1998) Long-term concentrations of ambient air pollutants and incident
6 lung cancer in California adults: results from the AHSMOG study. Environ. Health Perspect. 106: 813-823.
7 Berglund, D. J.; Abbey, D. E.; Lebowitz, M. D.; Knutsen, S. F.; McDonnell, W. F. (1999) Respiratory symptoms
8 and pulmonary function in an elderly nonsmoking population. Chest 115: 49-59.
9 Beyer, U.; Franke, K.; Cyrys, J.; Peters, A.; Heinrich, J.; Wichmann, H. E.; Brunekreef, B. (1998) Air pollution and
10 respiratory health of children: the PEACE panel study in Hettstedt and Zerbst, Eastern Germany. Eur. Respir.
11 Rev. 8:61-69.
12 Bobak, M.; Leon, D. A. (1992) Air pollution and infant mortality in the Czech Republic, 1986-1988.
13 Lancet (8826): 1010-1014.
14 Bobak, M.; Leon, D. (1998) Air pollution and infant mortality: the effects are specific for respiratory causes in
15 postneonatal period. Epidemiology 9: S58.
16 Bobak, M.; Leon, D. A. (1999) Pregnancy outcomes and outdoor air pollution: an ecological study in districts of the
17 Czech Republic 1986-8. Occup. Environ. Med. 56: 539-543.
18 Bobak, M.; Roberts, A. (1997) Heterogeneity of air pollution effects is related to average temperature [letter].
19 Br. Med. J. 315: 1161.
20 Boezen, M.; Schouten, J.; Rijcken, B.; Vonk, J.; Gemtsen, J.; Van Der Zee, S.; Hoek, G.; Brunekreef, B.;
21 Postma, D. (1998) Peak expiratory flow variability, bronchial responsiveness, and susceptibility to ambient
22 air pollution in adults. Am. J. Respir. Crit. Care Med. 158: 1848-1854.
23 Boezen, H. M.; Van Der Zee, S. C.; Postma, D. S.; Vonk, J. M.; Gerritsen, J.; Hoek, G.; Brunekreef, B.;
24 Rijcken, B.; Schouten, J. P. (1999) Effects of ambient air pollution on upper and lower respiratory symptoms
25 and peak expiratory flow in children. Lancet 353: 874-878.
26 Borja-Aburto, V. H.; Loomis, D. P.; Bangdiwala, S. I.; Shy, C. M.; Rascon-Pacheco, R. A. (1997) Ozone,
27 suspended particulates, and daily mortality in Mexico City. Am. J. Epidemiol. 145: 258-268.
28 Borja-Aburto, V. H.; Castillejos, M.; Gold, D. R.; Bierzwinski, S.; Loomis, D. (1998) Mortality and ambient fine
29 particles in southwest Mexico City, 1993-1995. Environ. Health Perspect. 106: 849-855.
30 Braga, A. L. F.; Conceicao, G. M. S.; Pereira, L. A. A.; Kishi, H. S.; Pereira, J. C. R.; Andrade, M. F.; Goncalves,
31 F. L. T.; Saldiva, P. H. N.; Latorre, M. R. D. O. (1999) Air pollution and pediatric respiratory hospital
32 admissions in Sao Paulo, Brazil. J. Environ. Med. 1: 95-102.
33 Braga, A. L. F.; Zanobetti, A.; Schwartz, J. (2000) Do respiratory epidemics confound the association between air
34 pollution and daily deaths? Eur. Respir. J. 16: 723-728.
35 Braun-Fahrlander, C.; Vuille, J. C.; Sennhauser, F. H.; Neu, U.; Kiinzle, T.; Grize, L.; Gassner, M.; Minder, C.;
36 Schindler, C.; Varonier, H. S.; Wiithnch, B.; SCARPOL team. (1997) Respiratory health and long-term
37 exposure to air pollutants in Swiss schoolchildren. Am. J. Respir. Crit. Care Med. 155: 1042-1049.
38 Bremner, S. A.; Anderson, H. R.; Atkinson, R. W.; McMichael, A. J.; Strachan, D. P.; Bland, J. M.; Bower, J. S.
39 (1999) Short term associations between outdoor air pollution and mortality in London 1992-4. Occup.
40 Environ. Med. 56: 237-244.
41 Brook, J. R.; Wiebe, A. H.; Woodhouse, S. A.; Audette, C. V.; Dann, T. F.; Callaghan, S.; Piechowski, M.;
42 Dabek-Zlotorzynska, E.; Dloughy, J. F. (1997) Temporal and spatial relationships in fine particle strong
43 acidity, sulphate, PM10, and PM25 across multiple Canadian locations. Atmos. Environ. 31: 4223-4236.
44 Brunekreef, B. (1997) Air pollution and life expectancy: is there a relation? Occup. Environ. Med. 54: 781-784.
45 Brunekreef, B.; Janssen, N. A. H.; Van Vliet, P. H. N.; Aarts, F. J. H. (2000) Paniculate matter concentrations in
46 relation to degree of urbanization and proximity to highways in The Netherlands. Presented at: PM2000:
47 particulate matter and health—the scientific basis for regulatory decision-making, specialty conference &
48 exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management Association.
49 Burnett, R. T.; Dales, R. E.; Raizenne, M. E.; Krewski, D.; Summers, P. W.; Roberts, G. R.; Raad-Young, M.;
50 Dann, T.; Brook, J. (1994) Effects of low ambient levels of ozone and sulfates on the frequency of
51 respiratory admissions to Ontario hospitals. Environ. Res. 65: 172-194.
52 Burnett, R. T.; Dales, R.; Krewski, D.; Vincent, R.; Dann, T.; Brook, J. R. (1995) Associations between ambient
53 particulate sulfate and admissions to Ontario hospitals for cardiac and respiratory diseases. Am. J. Epidemiol.
54 142: 15-22.
March 2001 6-272 DRAFT-DO NOT QUOTE OR CITE
-------
1 Burnett, R. T.; Dales, R. E.; Brook, J. R.; Raizenne, M. E.; Krewski, D. (1997a) Association between ambient
2 carbon monoxide levels and hospitalizations for congestive heart failure in the elderly in 10 Canadian cities.
3 Epidemiology 8: 162-167.
4 Burnett, R. T.; Cakmak, S.; Brook, J. R.; Krewski, D. (1997b) The role of paniculate size and chemistry in the
5 association between summertime ambient air pollution and hospitalization for cardiorespiratory diseases.
6 Environ. Health Perspect. 105: 614-620.
7 Burnett, R. T.; Brook, J. R.; Yung, W. T.; Dales, R. E.; Krewski, D. (1997c) Association between ozone and
8 hospitalization for respiratory diseases in 16 Canadian cities. Environ. Res. 72: 24-31.
9 Burnett, R. T.; Cakmak, S.; Brook, J. R. (1998a) The effect of the urban ambient air pollution mix on daily mortality
10 rates in 11 Canadian cities. Can. J. Public Health 89: 152-156.
11 Burnett, R. T.; Cakmak, S.; Raizenne, M. E.; Stieb, D.; Vincent, R.; Krewski, D.; Brook, J. R.; Philips, O.;
12 Ozkaynak, H. (1998b) The association between ambient carbon monoxide levels and daily mortality in
13 Toronto, Canada. J. Air Waste Manage. Assoc. 48: 689-700.
14 Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Cakmak, S.; Brook, J. R. (1999) Effects of particulate and gaseous air
15 pollution on cardiorespiratory hospitalizations. Arch. Environ. Health 54:130-139.
16 Burnett, R. T.; Brook, J., Dann, T.; Delocla, C.; Philips, O.; Cakmak, S.; Vincent, R.; Goldberg, M. S.; Krewski, D.
17 (2000) Association between particulate- and gas-phase components of urban air pollution and daily mortality
18 in eight Canadian cities. In: Grant, L. D., ed. PM2000: particulate matter and health. Inhalation Toxicol.
19 12(suppl.4): 15-39.
20 Cakmak, S.; Burnett, R. T.; Krewski, D. (1999) Methods for detecting and estimating population threshold
21 concentrations for air pollution-related mortality with exposure measurement error. Risk Anal. 19: 487-496.
22 Calderon-Garciduenas, L.; Mora-Tiscareno, A.; Chung, C. J.; Valencia, G.; Fordham, L. A.; Garcia, R.; Osnaya, N.;
23 Romero, L.; Acufia, H.; Villarreal-Calderon, A.; Devlin, R. B.; Koren, H. S. (2000) Exposure to air pollution
24 is associated with lung hyperinflation in healthy children and adolescents in southwest Mexico City: a pilot
25 study. Inhalation Toxicol. 12: 537-561.
26 Carroll, R. J.; Ruppert, D.; Stefanski, L. A. (1995) Measurement error in nonlinear models. London, United
27 Kingdom: Chapman & Hall. (Cox, D. R.; Hinkley, D. V.; Keiding, N.; Reid, N.; Rubin, D. B.; Silverman,
28 B. W., eds. Monographs on statistics and applied probability: v. 63).
29 Carrothers, T. J.; Evans, J. S. (2000) Assessing the impact of differential measurement error on estimates of fine
30 particle mortality. J. Air Waste Manage. Assoc. 50: 65-74.
31 Cassino, C.; Ito, K.; Bader, I.; Ciotoli, C.; Thurston, G.; Reibman, J. (1999) Cigarette smoking and ozone-associated
32 emergency department use for asthma by adults in New York City. Am. J. Respir. Crit. Care Med.
33 159: 1773-1779.
34 Castillejos, M.; Borja-Aburto, V. H.; Dockery, D. W.; Gold, D. R.; Loomis, D. (2000) Airborne coarse particles and
35 mortality. In: Inhalation Toxicology: proceedings of the third colloquium on particulate air pollution and
36 human health; June, 1999; Durham, NC. Inhalation Toxicology 12(suppl. 1): 61-72.
37 Checkoway, H.; Levy, D.; Sheppard, L.; Kaufman, J.; Koenig, J.; Siscovick, D. (2000) A case-crossover analysis of
38 fine particulate matter air pollution and out-of-hospital sudden cardiac arrest. Cambridge, MA: Health Effects
39 Institute; research report 99. Available: http://www.healtheffects.org/pubs-recent.htm [19 March, 2001].
40 Chen, C.; Chock, D. P.; Winkler, S. L. (1999) A simulation study of confounding in generalized linear models for
41 air pollution epidemiology. Environ. Health Perspect. 107: 217-222.
42 Chen, L.; Yang, W.; Jennison, B. L.; Omaye, S. T. (2000) Air particulate pollution and hospital admissions for
43 chronic obstructive pulmonary disease in Reno, Nevada. Inhalation Toxicol. 12: 281-298.
44 Chew, F. T.; Goh, D. Y. T.; Ooi, B. C.; Saharom, R.; Hui, J. K. S.; Lee, B. W. (1999) Association of ambient
45 air-pollution levels with acute asthma exacerbation among children in Singapore. Allergy (Copenhagen)
46 54:320-329.
47 Chock, D. P.; Winkler, S.; Chen, C. (2000) A study of the association between daily mortality and ambient air
48 pollutant concentrations in Pittsburgh, Pennsylvania. J. Air Waste Manage. Assoc. 50: 1481-1500.
49 Choudhury, A. H.; Gordian, M. E.; Morris, S. S. (1997) Associations between respiratory illness and PM,0 air
50 pollution. Arch. Environ. Health 52: 113-117.
51 Cifuentes, L. A.; Vega, J.; Kopfer, K.; Lave, L. B. (2000) Effect of the fine fraction of particulate matter versus the
52 coarse mass and other pollutants on daily mortality in Santiago, Chile. J. Air Waste Manage. Assoc.
53 50: 1287-1298.
54
55
March 2001 6-273 DRAFT-DO NOT QUOTE OR CITE
-------
1 Clench-Aas, J.; Bartonova, A.; Skj0nsberg, O. H.; Leegaard, J.; Hagen, L. O.; Giasver, P.; Moseng, J.; Roemer, W.
2 (1998) Air pollution and respiratory health of children: the PEACE study in Oslo, Norway. Eur. Respir. Rev.
3 8:36-43.
4 Clyde, M. A.; Guttorp, P.; Sullivan, E. (2000) Effects of ambient fine and coarse particles on mortality in Phoenix,
5 Arizona. J. Exposure Anal. Environ. Epidemiol.: submitted.
6 Cook, J. R.; Stefanski, L. A. (1994) Simulation-extrapolation estimation in parametric measurement error models.
7 J. Am. Stat. Assoc. 89: 1314-1328.
8 Coutant, R. W. (1977) Effect of environmental variables on collection of atmospheric sulfate. Environ. Sci.
9 Technol. 11:873-878.
10 Cropper, M. L.; Simon, N. B.; Alberini, A.; Arora, S.; Sharma, P. K. (1997) The health benefits of air pollution
11 control in Delhi. Am. J. Agric. Econ. 79: 1625-1629.
12 Cuijpers, C. E. J.; Swaen, G. M. H.; Wesseling, G.; Wouters, E. F. M. (1994) Acute respiratory effects of summer
13 smog in primary school children. Toxicol. Lett. 72: 227-235.
14 Dab, W.; Medina, S.; Quenel, P.; Le Moullec, Y.; Le Tertre, A.; Thelot, B.; Monteil, C.; Lameloise, P.; Pirard, P.;
15 Momas, I.; Ferry, R.; Festy, B. (1996) Short term respiratory health effects of ambient air pollution: results of
16 the APHEA project in Paris. In: St Leger, S., ed. The APHEA project. Short term effects of air pollution on
17 health: a European approach using epidemiological time series data. J. Epidemiol. Community Health
18 50(suppl. 1):S42-S46.
19 Damia, A. D.; Fabregas, M. L.; Tordera, M. P.; Torrero, L. C. (1999) Effects of air pollution and weather conditions
20 on asthma exacerbation. Respiration 66: 52-58.
21 Daniels, M.; Dominici, F.; Samet, J. M.; Zeger, S. L. (2000) Estimating particulate matter-mortality dose-response
22 curves and threshold levels: an analysis of daily time-series for the 20 largest US cities. Am. J. Epidemiol.
23 152:397-406.
24 Delfino, R. J.; Coate, B. D.; Zeiger, R. S.; Seltzer, J. M.; Street, D. H.; Koutrakis, P. (1996) Daily asthma severity in
25 relation to personal ozone exposure and outdoor fungal spores. Am. J. Respir. Crit. Care Med. 154: 633-641.
26 Delfino, R. J.; Murphy-Moulton, A. M.; Burnett, R. T.; Brook, J. R.; Becklake, M. R. (1997) Effects of air pollution
27 on emergency room visits for respiratory illnesses in Montreal, Quebec. Am. J. Respir. Crit. Care Med.
28 155:568-576.
29 Delfino, R. J.; Murphy-Moulton, A. M.; Becklake, M. R. (1998) Emergency room visits for respiratory illnesses
30 among the elderly in Montreal: association with low level ozone exposure. Environ. Res. 76: 67-77.
31 Dejmek, J.; Selevan, S. G.; Benes, L; Solansky, I.; Sram, R. J. (1999) Fetal growth and maternal exposure to
32 particulate matter during pregnancy. Environ. Health Perspect. 107: 475-480.
33 Diaz, J.; Garcia, R.; Ribera, P.; Alberdi, J. C.; Hernandez, E.; Pajares, M. S.; Otero, A. (1999) Modeling of air
34 pollution and its relationship with mortality and morbidity in Madrid, Spain. Int. Arch. Occup. Environ.
35 Health 72: 366-376.
36 Dockery, D. W.; Spengler, J. D. (1981) Personal exposure to respirable particulates and sulfates. J. Air Pollut.
37 Control Assoc. 31: 153-159.
38 Dockery, D. W.; Spengler, J. D.; Neas, L. M.; Speizer, F. E.; Ferris, B. G., Jr.; Ware, J. H.; Brunekreef, B. (1989)
39 An epidemiologic study of respiratory health status and indicators of indoor air pollution from combustion
40 sources. In: Harper, J. P., ed. Combustion processes and the quality of the indoor environment: transactions
41 of an international specialty conference; September 1988; Niagara Falls, NY. Pittsburgh, PA: Air & Waste
42 Management Association; pp. 262-271. (A&WMA transactions series: TR-15).
43 Dockery, D. W.; Schwartz, J.; Spengler, J. D. (1992) Air pollution and daily mortality: associations with particulates
44 and acid aerosols. Environ. Res. 59: 362-373.
45 Dockery, D. W.; Pope, C. A, III; Xu, X.; Spengler, J. D.; Ware, J. H.; Fay, M. E.; Ferris, B. G., Jr.; Speizer, F. E.
46 (1993) An association between air pollution and mortality in six U.S. cities. N. Engl. J. Med.
47 329: 1753-1759.
48 Dockery, D. W.; Cunningham, J.; Damokosh, A. I.; Neas, L. M.; Spengler, J. D.; Koutrakis, P.; Ware, J. H.;
49 Raizenne, M.; Speizer, F. E. (1996) Health effects of acid aerosols on North American children: respiratory
50 symptoms. Environ. Health Perspect. 104: 500-505.
51 Dockery, D. W.; Pope, C. A., Ill; Kanner, R. E.; Villegas, G. M.; Schwartz, J. (1999) Daily changes in oxygen
52 saturation and pulse rate associated with particulate air pollution and barometric pressure. Cambridge, MA:
53 Health Effects Institute; research report no. 83.
54 Dominici, F.; Zeger, S. L.; Samet, J. (2000) A measurement error model for time-series studies of air pollution and
55 mortality. Biostatistics 1: 157-175.
March 2001 6-274 DRAFT-DO NOT QUOTE OR CITE
-------
1 Fairley, D. (1990) The relationship of daily mortality to suspended participates in Santa Clara county, 1980-86.
2 Environ. Health Perspect. 89: 159-168.
3 Fairley, D. (1999) Daily mortality and air pollution in Santa Clara County, California: 1989-1996. Environ. Health
4 Perspect. 107:637-641.
5 Fanucchi, M. V.; Plopper, C. G. (1997) Pulmonary developmental responses to toxicants. In: Roth, R. A., ed.
6 Toxicology of the respiratory system, v. 8 of comprehensive toxicology. New York, NY: Pergamon, p.
7 203-220.
8 Fanucchi, M. V.; Wong, V. J.; Hinds, D.; Tarkington, B. K.; Van Winkle, L. S.; Evans, M. J.; Plopper, C. G. (2000)
9 Repeated episodes of exposure to ozone alters postnatal development of distal conducting airways in infant
10 rhesus monkeys. Am. J. Respir. Crit. Care Med. 161: A615.
11 Forsberg, B.; Segerstedt, B.; Stjernberg, N.; Roemer, W. (1998) Air pollution and respiratory health of children:
12 the PEACE panel study in Umea, Sweden. Eur. Respir. Rev. 8: 12-19.
13 Frischer, T.; Studmcka, M.; Gartner, C.; Tauber, E.; Horak, F.; Veiter, A.; Spengler, J,; Kiihr, J.; Urbanek, R.
14 (1999) Lung function growth and ambient ozone: a three-year population study in school children. Am. J.
15 Respir. Crit. Care Med. 160: 390-396.
16 Fung, K. Y.; Krewski, D. (1999) On measurement error adjustment methods in Poisson regression. Environmetrics
17 10:213-224.
18 Gamble, J. L. (1998) Effects of ambient air pollution on daily mortality: a time series analysis of Dallas, Texas,
19 1990-1994. Presented at: 91st annual meeting and exhibition of the Air & Waste Management Association;
20 June; San Diego, CA. Pittsburgh, PA: Air & Waste Management Association; paper no. 98-MP26.03.
21 Garcia-Aymerich, J.; Tobias, A.; Anto, J. M.; Sunyer, J. (2000) Air pollution and mortality in a cohort of patients
22 with chronic obstructive pulmonary disease: a time series analysis. J. Epidemiol. Community Health
23 54:73-74.
24 Garty, B. Z.; Kosman, E.; Ganor, E.; Berger, V.; Garty, L.; Wietzen, T.; Waisman, Y.; Mimouni, M.; Waisel, Y.
25 (1998) Emergency room visits of asthmatic children, relation to air pollution, weather, and airborne allergens.
26 Ann. Allergy Asthma Immunol. 81: 563-570.
27 Gauderman, W. J.; Mcconnell, R.; Gilliland, F.; London, S.; Thomas, D.; Avol, E.; Vora, H.; Berhane, K.;
28 Rappaport, E. B.; Lurmann, F.; Margolis, H. G.; Peters, J. (2000) Association between air pollution and lung
29 function growth in southern California children. Am. J. Respir. Crit. Care Med. 162: 1383-1390.
30 Gerginova, M.; Kostianev, S.; Ivanova, M. (1989) Reference values of dynamic pulmonary functional indices in
31 boys between 7 and 14 years of age. Pediatria 28: 52-58.
32 Gielen, M. H.; Van Der Zee, S. C.; Van Wijnen, J. H.; Van Steen, C. J.; Brunekreef, B. (1997) Acute effects of
33 summer air pollution on respiratory health of asthmatic children. Am. J. Respir. Crit. Care Med.
34 155:2105-2108.
35 Gilmour, P. S.; Brown, D. M.; Lindsay, T. G.; Beswick, P. H.; MacNee, W.; Donaldson, K. (1996) Adverse health
36 effects of PM,0 particles: involvement of iron in generation of hydroxyl radical. Occup. Environ. Med.
37 53: 817-822.
38 Gold, A.; Litonius, J.; Schwartz, M.; Verrier, R.; Milstein, A.; Larson, E.; Lovett, B. (1998) Cardiovascular
39 vulnerability to particulate pollution. Presented at: 1998 International conference [American Thoracic
40 Society]; April; Chicago, IL. Am. J. Respir. Crit. Care Med. 157: A261.
41 Gold, D. R.; Damokosh, A. L; Pope, C. A., Ill; Dockery, D. W.; McDonnell, W. F.; Serrano, P.; Retama, A.;
42 Castillejos, M. (1999) Particulate and ozone pollutant effects on the respiratory function of children in
43 southwest Mexico City. Epidemiology 10: 8-16.
44 Gold, D. R.; Litonjua, A.; Schwartz, J.; Lovett, E.; Larson, A.; Nearing, B.; Allen, G.; Verrier, M.; Cherry, R.;
45 Verrier, R. (2000) Ambient pollution and heart rate variability. Circulation 101: 1267-1273.
46 Goldberg, M. S.; Bailar, J. C., Ill; Burnett, R. T.; Brook, J. R.; Tamblyn, R.; Bonvalot, Y.; Ernst, P.; Flegel, K. M.;
47 Singh, R. K.; Valois, M.-F. (2000) Identifying subgroups of the general population that may be susceptible to
48 short-term increases in particulate air pollution: a time-series study in Montreal, Quebec. Cambridge, MA:
49 Health Effects Institute; research report 97. Available: http://www.healtheffects.org/pubs-research.htm
50 [15 February, 2001].
51 Gordian, M. E.; Ozkaynak, H.; Xue, J.; Morris, S. S.; Spengler, J. D. (1996) Particulate air pollution and respiratory
52 disease in Anchorage, Alaska. Environ. Health Perspect. 104: 290-297.
53 Gouveia, N.; Fletcher, T. (2000) Respiratory diseases in children and outdoor air pollution in Sao Paulo, Brazil:
54 a time series analysis. Occup. Environ. Med. 57: 477-483.
March 2001 6-275 DRAFT-DO NOT QUOTE OR CITE
-------
1 Grievink, L.; Van der Zee, S. C.; Hoek, G.; Boezen, H. M.; Van't Veer, P.; Brunekreef, B. (1999) Modulation of the
2 acute respiratory effects of winter air pollution by serum and dietary antioxidants: a panel study. Eur. Respir.
3 J. 13: 1439-1446.
4 Giintzel, O.; Bollag, U.; Helfenstein, U. (1996) Asthma and exacerbation of chronic bronchitis: sentinel and
5 environmental data in a time series analysis. Zentralbl. Hyg. Umweltmed. 198: 383-393.
6 Guo, Y. L.; Lin, Y.-C.; Sung, F.-C.; Huang, S.-L.; Ko, Y.-C.; Lai, J.-S.; Su, H.-J.; Shaw, C.-K.; Lin, R.-S.; Dockery,
7 D. W. (1999) Climate, traffic-related air pollutants, and asthma prevalence in middle-school children in
8 Taiwan. Environ. Health Perspect. 107: 1001 -1006.
9 Gwynn, R. C.; Burnett, R. T.; Thurston, G. D. (2000) A time-series analysis of acidic paniculate matter and daily
10 mortality and morbidity in the Buffalo, New York, region. Environ. Health Perspect. 108: 125-133.
11 Hagen, J. A.; Nafstad, P.; Skrondal, A.; Bjorkly, S.; Magnus, P. (2000) Associations between outdoor air pollutants
12 and hospitalization for respiratory diseases. Epidemiology 11: 136-140.
13 Hajat, S.; Haines, A.; Goubet, S. A.; Atkinson, R. W.; Anderson, H. R. (1999) Association of air pollution with
14 daily GP consultations for asthma and other lower respiratory conditions in London. Thorax 54: 597-605.
15 Haluszka, J.; Pisiewicz, K.; Miczynski, J.; Roemer, W.; Tomalak, W. (1998) Air pollution and respiratory health in
16 children: the PEACE panel study in Krakow, Poland. Eur. Respir. Rev. 8: 94-100.
17 Harre, E. S. M.; Price, P. D.; Ayrey, R. B.; Toop, L. J.; Martin, I. R.; Town, G. I. (1997) Respiratory effects of air
18 pollution in chronic obstructive pulmonary disease: a three month prospective study. Thorax 52: 1040-1044.
19 Harris, R. J. (1975) A primer of multivariate statistics. New York, NY: Academic Press; pp. 155-224.
20 Hefflin, B. J.; Jalaludin, B.; McClure, E.; Cobb, N.; Johnson, C. A.; Jecha, L.; Etzel, R. A. (1994) Surveillance for
21 dust storms and respiratory diseases in Washington State, 1991. Arch. Environ. Health 49: 170-174.
22 Heinrich, J.; Hoelscher, B.; Wjst, M.; Ritz, B.; Cyrys, J.; Wichmann, H.-E. (1999) Respiratory diseases and
23 allergies in two polluted areas in East Germany. Environ. Health Perspect. 107: 53-62.
24 Heinrich, J.; Hoelscher, B.; Wichmann, H. E. (2000) Decline of ambient air pollution and respiratory symptoms in
25 children. Am. J. Respir. Crit. Care Med. 161: 1930-1936.
26 Hiltermann, T. J. N.; de Bruijne, C. R.; Stolk, J.; Zwinderman, A. H.; Spieksma, F. Th. M.; Roemer, W.;
27 Steerenberg, P. A.; Fischer, P. H.; van Bree, L.; Hiemstra, P. S. (1997) Effects of photochemical air pollution
28 and allergen exposure on upper respiratory tract inflammation in asthmatics. Am. J. Respir. Crit. Care Med.
29 156: 1765-1772.
30 Hiltermann, T. J. N.; Stolk, J.; Van der Zee, S. C.; Brunekreef, B.; De Bruijne, C. R.; Fischer, P. H.; Ameling,
31 C. B.; Sterk, P. J.; Hiemstra, P. S.; Van Bree, L. (1998) Asthma severity and susceptibility to air pollution.
32 Eur. Respir. J. 11:686-693.
33 Hoek, G.; Brunekreef, B. (1994) Effects of low-level winter air pollution concentrations on respiratory health of
34 Dutch children. Environ. Res. 64: 136-150.
35 Hoek, G.; Schwartz, J. D.; Groot, B.; Eilers, P. (1997) Effects of ambient particulate matter and ozone on daily
36 mortality in Rotterdam, the Netherlands. Arch. Environ. Health 52: 455-463.
37 Hoek, G.; Dockery, D. W.; Pope, A.; Neas, L.; Roemer, W.; Brunekreef, B. (1998) Association between PM10 and
38 decrements in peak expiratory flow rates in children: reanalysis of data from five panel studies. Eur. Respir.
39 J. 11: 1307-1311.
40 Hoek, G.; Brunekreef, B.; Verhoeff, A.; van, Wijnen, J.; Fischer, P. (2000) Daily mortality and air pollution in the
41 Netherlands. J. Air Waste Manage. Assoc. 50: 1380-1389.
42 Hodgkin, J. E.; Abbey, D. E.; Euler, G. L.; Magie, A. R. (1984) COPD prevalence in nonsmokers in high and low
43 photochemical air pollution areas. Chest 86: 830-838.
44 Hong, Y.-C.; Leem, J.-H.; Ha, E.-H.; Christiani, D. C. (1999) PM10 exposure, gaseous pollutants, and daily
45 mortality in Inchon, South Korea. Environ. Health Perspect. 107: 873-878.
46 Ito, K. (1990) An examination of the role of aerosol acidity in historical London, England daily mortality
47 [dissertation]. Syracuse, NY: New York University. Available from: University Microfilms International,
48 Ann Arbor, MI; AAD91-13012.
49 Ito, K.; Burnett, R. T.; Lippmann, M. (1998) Particulate matter components and respiratory hospital admissions in
50 the elderly in Detroit, Ml. In: Abstracts: 1998 International conference; April; Chicago, IL, American
51 Thoracic Society. Am. J. Respir. Crit. Care Med. 157(3 suppl.): A879.
52 Jacobs, J.; Kreutzer, R.; Smith, D. (1997) Rice burning and asthma hospitalizations, Butte County, California,
53 1983-1992. Environ. Health Perspect. 105: 980-985.
54 Jamason, P. F.; Kalkstein, L. S.; Gergen, P. J. (1997) A synoptic evaluation of asthma hospital admissions in
55 New York City. Am. J. Respir. Crit. Care Med. 156: 1781-1788.
March 2001 6-276 DRAFT-DO NOT QUOTE OR CITE
-------
1 Jedrychowski, W.; Flak, E. (1998) Separate and combined effects of the outdoor and indoor air quality on chronic
2 respiratory symptoms adjusted for allergy among preadolescent children. Int. J. Occup. Med. Environ. Health
3 11:19-35.
4 Jedrychowski, W.; Flak, E.; Mroz, E. (1999) The adverse effect of low levels of ambient air pollutants on lung
5 function growth in preadolescent children. Environ. Health Perspect. 107: 669-674.
6 Kalandidi, A.; Gratziou, C.; Katsouyanni, K.; Manalis, N.; Tzala, L.; Pantazopoulou, A.; Efthimiou, M.;
7 Roussos, C.; Roemer, W. (1998) Air pollution and respiratory health of children: the PEACE panel study in
8 Athens, Greece. Eur. Respir. Rev. 8: 117-124.
9 Katsouyanni, K.; Touloumi, G. (1998) Causes of regional differences in air pollution effects are being studied
10 further [letter]. Br. Med. J. 316: 1982.
11 Katsouyanni, K.; Schwartz, J.; Spix, C.; Touloumi, G.; Zmirou, D.; Zanobetti, A.; Wojtyniak, B.; Vonk, J. M.;
12 Tobias, A.; Ponka, A.; Medina, S.; Bacharova, L.; Andersen, H. R. (1996) Short term effects of air pollution
13 on health: a European approach using epidemiology time series data: the APHEA protocol. In: St Leger, S.,
14 ed. The APHEA project. Short term effects of air pollution on health: a European approach using
15 epidemiological time series data. J. Epidemiol. Community Health 50(suppl. 1): S12-S18.
16 Katsouyanni, K.; Touloumi, G.; Spix, C.; Schwartz, J.; Balducci, F.; Medina, S.; Rossi, G.; Wojtyniak, B.;
17 Sunyer, J.; Bacharova, L.; Schouten, J. P.; Ponka, A.; Anderson, H. R. (1997) Short term effects of ambient
18 sulphur dioxide and particulate matter on mortality in 12 European cities: results from time series data from
19 the APHEA project. Br. Med. J. 314: 1658-1663.
20 Keles, N.; Ilicali, O. C.; Deger, K. (1999) Impact of air pollution on prevalence of rhinitis in Istanbul. Arch.
21 Environ. Health 54: 48-51.
22 Kelsall, J. E.; Samet, J. M.; Zeger, S. L.; Xu, J. (1997) Air pollution and mortality in Philadelphia, 1974-1988.
23 Am. J. Epidemiol. 146: 750-762.
24 Klemm, R. J.; Mason, R. M., Jr. (2000) Aerosol research and inhalation epidemiological study (ARIES): air quality
25 and daily mortality statistical modeling—interim results. J. Air. Waste Manage. Assoc. 50: 1433-1439.
26 Klemm, R. J.; Mason, R. M., Jr.; Heilig, C. M.; Neas, L. M.; Dockery, D. W. (2000) Is daily mortality associated
27 specifically with fine particles? Data reconstruction and replication of analyses. J. Air Waste Manage. Assoc.
28 50: 1215-1222.
29 Koenig, J. Q,; Larson, T. V.; Hanley, Q. S.; Rebolledo, V.; Dumler, K.; Checkoway, H.; Wang, S.-Z.; Lin, D.;
30 Pierson, W. E. (1993) Pulmonary function changes in children associated with fine particulate matter.
31 Environ. Res. 63: 26-38.
32 Kontos, A. S.; Fassois, S. D.; Deli, M. F. (1999) Short-term effects of air pollution on childhood respiratory illness
33 in Piraeus, Greece, 1987-1992: nonparametric stochastic dynamic analysis. Environ. Res. 81: 275-296.
34 Korrick, S. A.; Neas, L. M.; Dockery, D. W.; Gold, D. R.; Allen, G. A.; Hill, L. B.; Kimball, K. D.; Rosner, B. A.;
35 Speizer, F. E. (1998) Effects of ozone and other pollutants on the pulmonary function of adult hikers.
36 Environ. Health Perspect. 106: 93-99.
37 Kostianev, S.; Gerginova, M.; Ivanova, M. (1994) Reference values of lung function parameters in Bulgarian girls
38 aged 7 to 14 years. Pediatria 33: 30-33.
39 Kotesovec, F.; Vitnerova, N.; Leixner, M.; Benes, I.; Skorkovsky, J.; Roemer, W. (1998) Air pollution and
40 respiratory health of children: the PEACE panel study in Teplice, Czech Republic. Eur. Respir. Rev.
41 8:70-77.
42 Kotesovec, F.; Skorkovsky, J.; Brynda, J.; Peters, A.; Heinrich, J. (2000) Daily mortality and air pollution in
43 northern Bohemia; different effects for men and women. Cent. Eur. J. Public Health 8: 120-127.
44 Kramer, U.; Behrendt, H.; Dolgner, R.; Ranft, U.; Ring, J.; Wilier, H.; Schlipkoter, H.-W. (1999) Airway diseases
45 and allergies in East and West German children during the first 5 years after reunification: time trends and
46 the impact of sulphur dioxide and total suspended particles. Int. J. Epidemiol. 28: 865-873.
47 Kramer, U.; Koch, T.; Ranft, U.; Ring, J.; Behrendt, H. (2000) Traffic-related air pollution is associated with atopy
48 in children living in urban areas. Epidemiology 11: 64-70.
49 Krewski, D.; Burnett, R. T.; Goldberg, M. S.; Hoover, K.; Siemiatycki, J.; Jerrett, M.; Abrahamowicz, M.; White,
50 W. H. (2000) Reanallysis of the Harvard Six Cities study and the American Cancer Society study of
51 particulate air pollution and mortality. A special report of the Institute's Particle Epidemiology Reanalysis
52 Project. Cambridge, MA: Health Effects Institute.
53 Kiinzli, N.; Tager, I. B. (1997) The semi-individual study in air pollution epidemiology: a valid design as compared
54 to ecologic studies. Environ. Health Perspect. 105: 1078-1083.
March 2001 6-277 DRAFT-DO NOT QUOTE OR CITE
-------
1 Kiinzli, N.; Ackermann-Liebrich, U.; Brandli, O.; Tschopp, J. M.; Schindler, C.; Leuenberger, P.; SAPALDIA
2 Team. (2000) Clinically "small" effects of air pollution on FVC have a large public health impact. Eur.
3 Respir. J. 15: 131-136.
4 Laden, F.; Neas, L. M.; Dockery, D. W.; Schwartz, J. (2000) Association of fine particulate matter from different
5 sources with daily mortality in six U.S. cities. Environ. Health Perspect. 108: 941-947.
6 Lave, L. B.; Seskin, E. P. (1977) Air pollution and human health. Baltimore, MD: The Johns Hopkins University
7 Press.
8 Lebowitz, M. D.; Collins, L.; Holberg, C. J. (1987) Time series analyses of respiratory responses to indoor and
9 outdoor environmental phenomena. Environ. Res. 43: 332-341.
10 Lee, J.-T.; Schwartz, J. (1999) Reanalysis of the effects of air pollution on daily mortality in Seoul, Korea:
11 a case-crossover design. Environ. Health Perspect. 107: 633-636.
12 Lee, J.-T.; Shy, C. M. (1999) Respiratory function as measured by peak expiratory flow rate and PM,0:
13 six communities study. J. Exposure Anal. Environ. Epidemiol. 9: 293-299.
14 Lee, R. E., Jr.; Caldwell, J. S.; Morgan, G. B. (1972) The evaluation of methods for measuring suspended
15 particulates in air. Atmos. Environ. 6: 593-622.
16 Lee, J.-T.; Shin, D.; Chung, Y. (1999) Air pollution and daily mortality in Seoul and Ulsan, Korea. Environ. Health
17 Perspect. 107: 149-154.
18 Leonard), G. S.; Houthuijs, D.; Steerenberg, P. A.; Fletcher, T.; Armstrong, B.; Antova, T. (2000) Immune
19 biomarkers in relation to exposure to particulate matter: a cross-sectional survey in 17 cities of central
20 Europe. In: Grant, L. D., ed. PM2000: particulate matter and health. Inhalation Toxicol. 12(suppl. 4): 1-14.
21 Levy, D. (1998) Fine particulate air pollution and out-of-hospital mortality in King County, Washington. In: Vostal,
22 J. J., ed. Health effects of particulate matter in ambient air. Proceedings of an international conference; 1997;
23 Prague, Czech Republic. Pittsburgh, PA: Air & Waste Management Association; pp. 262-271. (A&WMA
24 publication VIP-80).
25 Lewis, P. R.; Hensley, M. J.; Wlodarczyk, J.; Toneguzzi, R. C.; Westley-Wise, V. J.; Dunn, T.; Calvert, D. (1998)
26 Outdoor air pollution and children's respiratory symptoms in the steel cities of New South Wales. Med. J.
27 Aust. 169: 459-463.
28 Liao, D.; Creason, J.; Shy, C.; Williams, R.; Watts, R.; Zweidinger, R. (1999) Daily variation of particulate air
29 pollution and poor cardiac autonomic control in the elderly. Environ. Health Perspect. 107: 521-525.
30 Lin, C. A.; Martins, M. A.; Farhat, S. C. L.; Pope, C. A., Ill; Conceicao, G. M. S.; Anastacio, V. M.; Hatanaka, M.;
31 Andrade, W. C.; Hamaue, W. R.; Bohm, G. M.; Saldiva, P. H. N. (1999) Air pollution and respiratory illness
32 of children in Sao Paulo, Brazil. Paediatr. Perinat. Epidemiol. 13: 475-488.
33 Linn, W. S.; Shamoo, D. A.; Anderson, K. R.; Peng, R.-C.; Avol, E. L.; Hackney, J. D.; Gong, H., Jr. (1996)
34 Short-term air pollution exposures and responses in Los Angeles area schoolchildren. J. Exposure Anal.
35 Environ. Epidemiol. 6: 449-472.
36 Linn, W. S.; Szlachcic, Y.; Gong, H., Jr.; Kinney, P. L.; Berhane, K. T. (2000) Air pollution and daily hospital
37 admissions in metropolitan Los Angeles. Environ. Health Perspect. 108: 427-434.
38 Lipfert, F. W.; Wyzga, R. E. (1996) The effects of exposure error on environmental epidemiology. In: Lee, J.;
39 Phalen, R., eds. Proceedings of the 2nd colloquium on particulate air pollution and human health; May;
40 Park City, UT; pp. 4-295 - 4-302.
41 Lipfert, F. W.; Wyzga, R. E. (1997) Air pollution and mortality: the implications of uncertainties in regression
42 modeling and exposure measurement. J. Air Waste Manage. Assoc. 47: 517-523.
43 Lipfert, F. W.; Morris, S. C.; Wyzga, R. E. (2000a) Daily mortality in the Philadelphia metropolitan area and
44 size-classified particulate matter. J. Air Waste Manage. Assoc.: 1501-1513.
45 Lipfert, F. W.; Perry, H. M., Jr.; Miller, J. P.; Baty, J. D.; Wyzga, R. E.; Carmody, S. E. (2000b) The Washington
46 University-EPRI veterans' cohort mortality study: preliminary results. In: Grant, L. D., ed. PM2000:
47 particulate matter and health. Inhalation Toxicol. 12(suppl. 4): 41-73.
48 Lipfert, F. W.; Zhang, J.; Wyzga, R. E. (2000c) Infant mortality and air pollution: a comprehensive analysis of U.S.
49 data for 1990. J. Air Waste Manage. Assoc. 50: 1350-1366.
50 Lippmann, M.; Thurston, G. D. (1996) Sulfate concentrations as an indicator of ambient particulate matter air
51 pollution for health risk evaluations. J. Exposure Anal. Environ. Epidemiol. 6: 123-146.
52 Lippmann, M.; Ito, K.; Nadas, A.; Burnett, R. T. (2000) Association of particulate matter components with daily
53 mortality and morbidity in urban populations. Cambridge, MA: Health Effects Institute; research report no.
54 95.
March 2001 6-278 DRAFT-DO NOT QUOTE OR CITE
-------
1 Lipsett, M.; Hurley, S.; Ostro, B. (1997) Air pollution and emergency room visits for asthma in Santa Clara County,
2 California. Environ. Health Perspect. 105: 216-222.
3 Liu, L.-J. S.; Olson, M. P., Ill; Allen, G. A.; Koutrakis, P.; McDonnell, W. F.; Gerrity, T. R. (1994) Evaluation of
4 the Harvard ozone passive sampler on human subjects indoors. Environ. Sci. Technol. 28: 915-923.
5 Long, W.; Tate, R. B.; Neuman, M.; Manfreda, J.; Becker, A. B.; Anthonisen, N. R. (1998) Respiratory symptoms
6 in a susceptible population due to burning of agricultural residue. Chest 113: 351-357.
7 Loomis, D.; Castillejos, M.; Gold, D. R.; McDonnell, W.; Borja-Aburto, V. H. (1999) Air pollution and infant
8 mortality in Mexico City. Epidemiology 10: 118-123.
9 Lumley, T.; Heagerty, P. (1999) Weighted empirical adaptive variance estimators for correlated data regression.
10 J. R. Stat. Soc. B 61 (part 2): 459-477.
11 Mage, D.; Wilson, W.; Hasselblad, V.; Grant, L. (1999) Assessment of human exposure to ambient paniculate
12 matter. J. Air Waste Manage. Assoc. 49: 174-185.
13 Mar, T. F.; Morris, G. A.; Koenig, J. Q.; Larson, T. V. (2000) Associations between air pollution and mortality in
14 Phoenix, 1995-1997. Environ. Health Perspect. 108: 347-353.
15 Marcus, A. H.; Chapman, R. (1998) Estimating the health effects of fine particles from epidemiology studies:
16 how serious are problems of measurement error, correlation, and confounding? In: Chow, J.; Koutrakis, P.,
17 eds. PM2 5: a fine particle standard. Volume II: proceedings of an international specialty conference.; January;
18 Long Beach, CA. Pittsburgh, PA: Air & Waste Management Association; pp. 899-919.
19 McConnell, R.; Berhane, K.; Gilliland, F.; London, S. J.; Vora, H.; Avol, E.; Gauderman, W. J.; Margolis, H. G.;
20 Lurmann, F.; Thomas, D. C.; Peters, J. M. (1999) Air pollution and bronchitic symptoms in southern
21 California children with asthma. Environ. Health Perspect. 107: 757-760.
22 McDonnell, W. F.; Nishino-Ishikawa, N.; Petersen, F. F.; Chen, L. H.; Abbey, D. E. (2000) Relationships of
23 mortality with the fine and coarse fractions of long-term ambient PM10 concentrations in nonsmokers.
24 J. Exposure Anal. Environ. Epidemiol. 10: 427-436.
25 McGregor, G. R.; Walters, S.; Wordley, J. (1999) Daily hospital respiratory admissions and winter air mass types,
26 Birmingham, UK. Int. J. Biometeorol. 43: 21-30.
27 Medina, S.; Le Tertre, A.; Quenel, P.; Le Moullec, Y.; Lameloise, P.; Guzzo, J. C.; Festy, B.; Ferry, R.; Dab, W.
28 (1997) Air pollution and doctors' house calls: results from the ERPURS system for monitoring the effects of
29 air pollution on public health in greater Paris, France, 1991-1995. Environ. Res. 75: 73-84.
30 Michelozzi, P.; Forastiere, F.; Fusco, D.; Perucci, C. A.; Ostro, B.; Ancona, C.; Pallotti, G. (1998) Air pollution and
31 daily mortality in Rome, Italy. Occup. Environ. Med. 55: 605-610.
32 Mills, P. K.; Abbey, D.; Beeson, W. L.; Petersen, F. (1991) Ambient air pollution and cancer in California
33 Seventh-day Adventists. Arch. Environ. Health 46: 271 -280.
34 Moolgavkar, S. H. (2000a) Air Pollution and Mortality in Three U.S. Counties. Environ. Health perspective
35 108:777-784.
36 Moolgavkar, S. H. (2000b) Air pollution and hospital admissions for diseases of the circulatory system in three U.S.
37 metropolitan areas. J. Air Waste Manage Assoc. 50: 1199-1206.
38 Moolgavkar, S. H. (2000c) Air pollution and hospital admissions for chronic obstructive pulmonary disease in three
39 metropolitan areas in the United States. In: Grant, L. D., ed. PM2000: particulate matter and health.
40 Inhalation Toxicol. 12(suppl. 4): 75-90.
41 Moolgavkar, S. H.; Luebeck, E. G. (1996) A critical review of the evidence on particulate air pollution and
42 mortality. Epidemiology 7: 420-428.
43 Moolgavkar, S. H.; Luebeck, E. G.; Anderson, E. L. (1997) Air pollution and hospital admissions for respiratory
44 causes in Minneapolis-St. Paul and Birmingham. Epidemiology 8: 364-370.
45 Moolgavkar, S. H.; Hazelton, W.; Luebeck, G.; Levy, D.; Sheppard, L. (2000) Air pollution, pollens, and
46 admissions for chronic respiratory disease in King County, Washington. In: Inhalation toxicology:
47 proceedings of the third colloquium on particulate air pollution and human health; June, 1999; Durham, NC.
48 Inhalation Toxicology 12(suppl. 1): 157-171.
49 Morgan, G.; Corbett, S.; Wlodarczyk, J.; Lewis, P. (1998) Air pollution and daily mortality in Sydney, Australia,
50 1989 through 1993. Am. J. Public Health 88: 759-764.
51 Morris, R. D.; Naumova, E. N. (1998) Carbon monoxide and hospital admissions for congestive heart failure:
52 evidence of an increased effect at low temperatures. Environ. Health Perspect. 106: 649-653.
53 Morris, R. D.; Naumova, E. N.; Munasinghe, R. L. (1995) Ambient air pollution and hospitalization for congestive
54 heart failure among elderly people in seven large US cities. Am. J. Public Health 85: 1361-1365.
March 2001 6-279 DRAFT-DO NOT QUOTE OR CITE
-------
1 Naeher, L. P.; Holford, T. R.; Beckett, W. S.; Belanger, K.; Triche, E. W.; Bracken, M. B.; Leaderer, B. P. (1999)
2 Healthy women's PEF variations with ambient summer concentrations of PM,0, PM2 5, SO42", H+, and O3.
3 Am. J. Respir. Crit. Care Med. 160: 117-125.
4 Nauenberg, E.; Basu, K. (1999) Effect of insurance coverage on the relationship between asthma hospitalizations
5 and exposure to air pollution. Public Health Rep. 114: 135-148.
6 Navidi, W.; Thomas, D.; Langholz, B.; Stram, D. (1999) Statistical methods for epidemiologic studies of the health
7 effects of air pollution. Cambridge, MA: Health Effects Institute; research report no. 86.
8 Neas, L. M.; Dockery, D. W.; Ware, J. H.; Spengler, J. D.; Ferris, B. G., Jr.; Speizer, F. E. (1994) Concentration of
9 indoor paniculate matter as a determinant of respiratory health in children. Am. J. Epidemiol.
10 139: 1088-1099.
11 Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Tollerud, D. J.; Speizer, F. E. (1995) The association of ambient air
12 pollution with twice daily peak expiratory flow rate measurements in children. Am. J. Epidemiol.
13 141:111-122.
14 Neas, L. M.; Dockery, D. W.; Burge, H.; Koutrakis, P.; Speizer, F. E. (1996) Fungus spores, air pollutants, and
15 other determinants of peak expiratory flow rate in children. Am. J. Epidemiol. 143: 797-807.
16 Neas, L. M.; Schwartz, J.; Dockery, D. (1999) A case-crossover analysis of air pollution and mortality in
17 Philadelphia. Environ. Health Perspect. 107: 629-631.
18 Neukirch, F.; Segala, C.; Le Moullec, Y.; Korobaeff, M.; Aubier, M. (1998) Short-term effects of low-level winter
19 pollution on respiratory health of asthmatic adults. Arch. Environ. Health 53: 320-328.
20 Norris, G.; Young-Pong, S. N.; Koenig, J. Q.; Larson, T. V.; Sheppard, L.; Stout, J. W. (1999) An association
21 between fine particles and asthma emergency department visits for children in Seattle. Environ. Health
22 Perspect. 107:489-493.
23 Norris, G.; Larson, T.; Koenig, J.; Claiborn, C.; Sheppard, L.; Finn, D. (2000) Asthma aggravation, combustion, and
24 stagnant air. Thorax 55: 466-470.
25 Ostro, B. (1995) Fine particulate air pollution and mortality in two Southern California counties. Environ. Res.
26 70:98-104.
27 Ostro, B. D.; Lipsett, M. J.; Wiener, M. B.; Seiner, J. C. (1991) Asthmatic responses to airborne acid aerosols.
28 Am. J. Public Health 81: 694-702.
29 Ostro, B. D.; Lipsett, M. J.; Mann, J. K.; Braxton-Owens, H.; White, M. C. (1995) Air pollution and asthma
30 exacerbations among African-American children in Los Angeles. In: Phalen, R. F.; Bates, D. V., eds.
31 Proceedings of the colloquium on particulate air pollution and human mortality and morbidity, part II;
32 January 1994; Irvine, CA. Inhalation Toxicol. 7: 711-722.
33 Ostro, B.; Chestnut, L.; Vichit-Vadakan, N.; Laixuthai, A. (1998) The impact of fine particulate matter in Bangkok,
34 Thailand. In: Chow, J.; Koutrakis, P., eds. PM2.5: a fine particle standard. Volume II: proceedings of an
35 international specialty conference; January; Long Beach, CA. Pittsburgh, PA: Air & Waste Management
36 Association; pp. 939-949. (A&WMA publication VIP-81).
37 Ostro, B. D.; Hurley, S.; Lipsett, M. J. (1999a) Air pollution and daily mortality in the Coachella Valley, California:
38 a study of PMIO dominated by coarse particles. Environ. Res. 81: 231-238.
39 Ostro, B. D.; Eskeland, G. S.; Sanchez, J. M.; Feyzioglu, T. (1999b) Air pollution and health effects: a study of
40 medical visits among children in Santiago, Chile. Environ. Health Perspect. 107: 69-73.
41 Ostro, B. D.; Broadwin, R.; Lipsett, M. J. (2000) Coarse and fine particles and daily mortality in the Coachella
42 Valley, CA: a follow-up study. J. Exposure Anal. Environ. Epidemiol.: 10: 412-419.
43 Ott, W.; Wallace, L.; Mage, D. (2000) Predicting particulate (PM,0) personal exposure distributions using a random
44 component superposition statistical model. J. Air Waste Manage. Assoc. 50: 1390-1406.
45 Ozkaynak, H.; Thurston, G. D. (1987) Associations between 1980 U.S. mortality rates and alternative measures of
46 airborne particle concentration. Risk Anal. 7: 449-461.
47 Ozkaynak, H.; Xue, J.; Zhou, H.; Raizenne, M. (1996) Associations between daily mortality and motor vehicle
48 pollution in Toronto, Canada. Boston, MA: Harvard University School of Public Health, Department of
49 Environmental Health; March 25.
50 Pantazopoulou, A.; Katsouyanni, K.; Kourea-Kremastinou, J.; Trichopoulos, D. (1995) Short-term effects of air
51 pollution on hospital emergency outpatient visits and admissions in the greater Athens, Greece area. Environ.
52 Res. 69:31-36.
53 Pekkanen, J.; Timonen, K. L.; Ruuskanen, J.; Reponen, A.; Mirme, A. (1997) Effects of ultrafine and fine particles
54 in urban air on peak expiratory flow among children with asthmatic symptoms. Environ. Res. 74: 24-33.
March 2001 6-280 DRAFT-DO NOT QUOTE OR CITE
-------
1 Pereira, L. A. A.; Loomis, D.; Conceicao, G. M. S.; Braga, A. L. F.; Areas, R. M.; Kishi, H. S.; Singer, J. M.;
2 Bohm, G. M.; Saldiva, P. H. N. (1998) Association between air pollution and intrauterine mortality in
3 Sao Paulo, Brazil. Environ. Health Perspect. 106: 325-329.
4 Perry, H. M., Jr.; Schnaper, H. W.; Meyer, G.; Swatzell, R. (1982) Clinical program for screening and treatment of
5 hypertension in veterans. J. Natl. Med. Assoc. 74: 433-444.
6 Peters, A.; Goldstein, I. F.; Beyer, U.; Franke, K.; Heinrich, J.; Dockery, D. W.; Spengler, J. D.; Wichmann, H.-E.
7 (1996) Acute health effects of exposure to high levels of air pollution in eastern Europe. Am. J. Epidemiol.
8 144:570-581.
9 Peters, A.; Doring, A.; Wichmann, H.-E.; Koenig, W. (1997a) Increased plasma viscosity during an air pollution
10 episode: a link to mortality? Lancet 349: 1582-1587.
11 Peters, A.; Wichmann, H. E.; Tuch, T.; Heinrich, J.; Heyder, J. (1997b) Respiratory effects are associated with the
12 number of ultrafine particles. Am. J. Respir. Crit. Care Med. 155: 1376-1383.
13 Peters, A.; Dockery, D. W.; Heinrich, J.; Wichmann, H. E. (1997c) Short-term effects of paniculate air pollution on
14 respiratory morbidity in asthmatic children. Eur. Respir. J. 10: 872-879.
15 Peters, A.; Kotesovec, F.; Skorkovsky, J.; Brynda, J.; Heinrich, J. (1999a) Akute Auswirkung der
16 Schwebstaubkonzentrationen in der Aufienluft auf die Mortahtat - Vergleichsstudie Nordost-Bayern /
17 Nordbohmen. AbschluBbericht [Acute effects of suspended particle concentrations in the atmosphere on
18 mortality - a study comparing northeast Bavaria and north Bohemia. Final report]. Bavaria, Federal Republic
19 of Germany: Institut fur Epidemiologie; report no. GSF-EP S 1/99.
20 Peters, J. M.; Avol, E.; Navidi, W.; London, S. J.; Gauderman, W. J.; Lurmann, F.; Linn, W. S.; Margolis, H.;
21 Rappaport, E.; Gong, H., Jr.; Thomas, D. C. (1999b) A study of twelve southern California communities with
22 differing levels and types of air pollution. I. Prevalence of respiratory morbidity. Am. J. Respir. Crit. Care
23 Med. 159:760-767.
24 Peters, J. M.; Avol, E.; Gauderman, W. J.; Linn, W. S.; Navidi, W.; London, S. J.; Margolis, H.; Rappaport, E.;
25 Vora, H.; Gong, H., Jr.; Thomas, D. C. (1999c) A study of twelve southern California communities with
26 differing levels and types of air pollution. II. Effects on pulmonary function. Am. J. Respir. Crit. Care Med.
27 159: 768-775.
28 Peters, A.; Liu, E.; Verrier, R. L.; Schwartz, J.; Gold, D. R.; Mittleman, M.; Baliff, J.; Oh, J. A.; Allen, G.;
29 Monahan, K.; Dockery, D. W. (2000a) Air pollution and incidence of cardiac arrhythmia. Epidemiology
30 11:11-17.
31 Peters, A.; Frohlich, M.; Doring, A.; Immervoll, T.; Wichmann, H.-E.; Hutchinson, W. L.; Pepys, M. B.; Koenig,
32 W. (2000b) Particulate air pollution is associated with an acute phase response in men: results from the
33 MONICA-Augsburg Study. Eur. Heart J.: in press.
34 Ponce de Leon, A.; Anderson, H. R.; Bland, J. M.; Strachan, D. P.; Bower, J. (1996) Effects of air pollution on
35 daily hospital admissions for respiratory disease in London between 1987-88 and 1991-92. In: St Leger, S.,
36 ed. The APHEA project. Short term effects of air pollution on health: a European approach using
37 epidemiological time series data. J. Epidemiol. Community Health 50(suppl. 1): S63-S70.
38 Ponka, A.; Savela, M.; Virtanen, M. (1998) Mortality and air pollution in Helsinki. Arch. Environ. Health
39 53:281-286.
40 Pope, C. A., III. (1991) Respiratory hospital admissions associated with PMIO pollution in Utah, Salt Lake, and
41 Cache Valleys. Arch. Environ. Health 46: 90-97.
42 Pope, C. A., Ill; Kalkstein, L S. (1996) Synoptic weather modeling and estimates of the exposure-response
43 relationship between daily mortality and particulate air pollution. Environ. Health Perspect. 104: 414-420.
44 Pope, C. A., Ill; Dockery, D. W.; Spengler, J. D.; Raizenne, M. E. (1991) Respiratory health and PM)0 pollution:
45 a daily time series analysis. Am. Rev. Respir. Dis. 144: 668-674.
46 Pope, C. A., Ill; Thun, M. J.; Namboodiri, M. M.; Dockery, D. W.; Evans, J. S.; Speizer, F. E.; Heath, C. W., Jr.
47 (1995) Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am. J.
48 Respir. Crit. Care Med. 151: 669-674.
49 Pope, C. A, III; Hill, R. W.; Villegas, G. M. (1999a) Particulate air pollution and daily mortality on Utah's Wasatch
50 Front. Environ. Health Perspect.: 107: 567-573.
51 Pope, C. A., Ill; Dockery, D. W.; Kanner, R. E.; Vollegas, G. M.; Schwartz, J. (1999b) Oxygen saturation, pulse
52 rate, and particulate air pollution: a daily time-series panel study. Am. J. Respir. Crit. Care Med.
53 159:365-372.
March 2001 6-281 DRAFT-DO NOT QUOTE OR CITE
-------
1 Pope, C. A., Ill; Verrier, R. L.; Lovett, E. G.; Larson, A. C.; Raizenne, M. E.; Kanner, R. E.; Schwartz, J.; Villegas,
2 G. M.; Gold, D. R.; Dockery, D. W. (1999c) Heart rate variability associated with paniculate air pollution.
3 Am. Heart J. 138:890-899.
4 Prescott, G. J.; Cohen, G. R.; Elton, R. A.; Fowkes, F. G. R.; Agius, R. M. (1998) Urban air pollution and
5 cardiopulmonary ill health: a 14.5 year time series study. Occup. Environ. Med. 55: 697-704.
6 Prescott, E.; Lange, P.; Vestbo, J.; Copenhagen City Heart Study Group. (1999) Socioeconomic status, lung
7 function and admission to hospital for COPD: results from the Copenhagen City Heart Study. Eur. Respir. J.
8 13:1109-1114.
9 Qian, Z. Chapman, R. S.; Tian, Q.; Chen, Y.; Lioy, P. J.; Zhang, J. (2000) Effects of air pollution on children's
10 respiratory health in three Chinese cities. Arch. Environ. Health 55: 126-133.
11 Rahlenbeck, S. I.; Kahl, H. (1996) Air pollution and mortality in East Berlin during the winters of 1981-1989. Int. J.
12 Epidemiol. 25: 1220-1226.
13 Raizenne, M.; Neas, L. M.; Damokosh, A. I.; Dockery, D. W.; Spengler, J. D.; Koutrakis, P.; Ware, J. H.; Speizer,
14 F. E. (1996) Health effects of acid aerosols on North American children: pulmonary function. Environ.
15 Health Perspect. 104: 506-514.
16 Roemer, W.; Hoek, G.; Brunekreef, B. (1993) Effect of ambient winter air pollution on respiratory health of
17 children with chronic respiratory symptoms. Am. Rev. Respir. Dis. 147: 118-124.
18 Roemer, W.; Hoek, G.; Brunekreef, B.; Haluszka, J.; Kalandidi, A.; Pekkanen, J. (1998) Daily variations in air
19 pollution and respiratory health in a multicentre study: the PEACE project. Eur. Respir. J. 12: 1354-1361.
20 Roemer, W.; Hoek, G.; Brunekreef, B.; Clench-Aas, J.; Forsberg, B.; Pekkanen, J.; Schutz, A. (2000) PM,0
21 elemental composition and acute respiratory health effects in European children (PEACE project). Eur.
22 Respir. J. 15:553-559.
23 Rojas-Bracho, L.; Suh, H. H.; Koutrakis, P. (2000) The relationship between personal exposures to PM25 and PM10
24 and their corresponding indoor and outdoor concentrations in Boston, MA. Environ. Health Perspect.:
25 submitted.
26 Romieu, I.; Meneses, F.; Ruiz, S.; Sienra, J. J.; Huerta, J.; White, M. C.; Etzel, R. A. (1996) Effects of air pollution
27 on the respiratory health of asthmatic children living in Mexico City. Am. J. Respir. Crit. Care Med.
28 154:300-307.
29 Romieu, I.; Meneses, F.; Ruiz, S.; Huerta, J.; Sienra, J. J.; White, M.; Etzel, R.; Hernandez, M. (1997) Effects of
30 intermittent ozone exposure on peak expiratory flow and respiratory symptoms among asthmatic children in
31 Mexico City. Arch. Environ. Health 52: 368-376.
32 Rooney, C.; McMichael, A. J.; Kovats, R. S.; Coleman, M. P. (1998) Excess mortality in England and Wales, and in
33 greater London, during the 1995 heatwave. J. Epidemiol. Community Health 52: 482-486.
34 Rosas, I.; McCartney, H. A.; Payne, R. W.; Calderon, C.; Lacey, J.; Chapela, R.; Ruiz-Velazco, S. (1998) Analysis
35 of the relationships between environmental factors (aeroallergens, air pollution, and weather) and asthma
36 emergency admissions to a hospital in Mexico City. Allergy (Copenhagen) 53: 394-401.
37 Rossi, G.; Vigotti, M. A.; Zanobetti, A.; Repetto, F.; Gianelle, V.; and Schwartz, J. (1999) Air pollution and
38 cause-specific mortality in Milan, Italy, 1980-1989. Arch. Environ. Health 54: 158-164.
39 Rothman, K. J.; Greenland, S., eds. (1998) Modern epidemiology. 2nd ed. Philadelphia, PA: Lippincott-Raven
40 Publishers.
41 Rutherford, S.; Clark, E.; McTainsh, G.; Simpson, R.; Mitchell, C. (1999) Characteristics of rural dust events shown
42 to impact on asthma severity in Brisbane, Australia. Int. J. Biometeorol. 42: 217-225.
43 Samet, J. M.; Zeger, S. L.; Kelsall, J. E.; Xu, J.; Kalkstein, L. S. (1996) Weather, air pollution and mortality in
44 Philadelphia, 1973-1980, report to the Health Effects Institute on phase IB, Particle Epidemiology Evaluation
45 Project. Cambridge, MA: Health Effects Institute; review draft.
46 Samet, J. M.; Zeger, S. L.; Kelsall, J. E.; Xu, J.; Kalkstein, L. S. (1997) Paniculate air pollution and daily mortality:
47 analysis of the effects of weather and multiple air pollutants. The phase I.B report of the Particle
48 Epidemiology Evaluation Project. Cambridge, MA: Health effects Institute; HEI special report.
49 Samet, J.; Zeger, S.; Kelsall, J.; Xu, J.; Kalkstein, L. (1998) Does weather confound of modify the association of
50 particulate air pollution with mortality? An analysis of the Philadelphia data, 1973-1980. Environ. Res.
51 77:9-19.
52 Samet, J. M.; Dominici, F.; Zeger, S. L.; Schwartz, J.; Dockery, D. W. (2000a) National morbidity, mortality, and
53 air pollution study. Part I: methods and methodologic issues. Cambridge, MA: Health Effects Institute;
54 research report no. 94.
March 2001 6-282 DRAFT-DO NOT QUOTE OR CITE
-------
1 Samet, J. M.; Zeger, S. L.; Dominici, F.; Curriero, F.; Coursac, I.; Dockery, D. W.; Schwartz, J.; Zanobetti, A.
2 (2000b) The national morbidity, mortality, and air pollution study. Part II: morbidity, mortality, and air
3 pollution in the United States. Cambridge, MA: Health Effects Institute; research report no. 94.
4 Scarlett, J. F.; Abbott, K. J.; Peacock, J. L.; Strachan, D. P.; Anderson, H. R. (1996) Acute effects of summer air
5 pollution on respiratory function in primary school children in southern England. Thorax 51:1109-1114.
6 Schouten, J. P.; Vonk, J. M.; de Graaf, A. (1996) Short term effects of air pollution on emergency hospital
7 admissions for respiratory disease: results of the APHEA project in two major cities in The Netherlands,
8 1977-89. In: St Leger, S., ed. The APHEA project. Short term effects of air pollution on health: a European
9 approach using epidemiological time series data. J. Epidemiol. Community Health 50(suppl. 1): S22-S29.
10 Schwartz, J. (1994a) What are people dying of on high air pollution days? Environ. Res. 64: 26-35.
11 Schwartz, J. (1994b) PM,0, ozone, and hospital admissions for the elderly in Minneapolis, MN. Arch. Environ.
12 Health 49: 366-374.
13 Schwartz, J. (1995) Short term fluctuations in air pollution and hospital admissions of the elderly for respiratory
14 disease. Thorax 50: 531-538.
15 Schwartz, J. (1997) Air pollution and hospital admissions for cardiovascular disease in Tucson. Epidemiology
16 8:371-377.
17 Schwartz, J. (1999) Air pollution and hospital admissions for heart disease in eight U.S. counties. Epidemiology
18 10:17-22.
19 Schwartz, J. (2000a) Assessing confounding, effect modification, and thresholds in the association between ambient
20 particles and daily deaths. Environ. Health Perspect. 108: 563-568.
21 Schwartz, J. (2000b) The distributed lag between air pollution and daily deaths. Epidemiology 11: 320-326.
22 Schwartz, J. (2000c) Harvesting and long term exposure effects in the relation between air pollution and mortality.
23 Am. J. Epidemiol. 151: 440-448.
24 Schwartz, J. (2000d) Daily deaths are associated with combustion particles rather than SO2 in Philadelphia.
25 Occup. Environ. Med. 57: 692-697.
26 Schwartz, J.; Dockery, D. W. (1992) Increased mortality in Philadelphia associated with daily air pollution
27 concentrations. Am. Rev. Respir. Dis. 145: 600-604.
28 Schwartz, J.; Morris, R. (1995) Air pollution and hospital admissions for cardiovascular disease in Detroit,
29 Michigan. Am. J. Epidemiol. 142: 23-35.
30 Schwartz, J.; Neas, L. M. (2000) Fine particles are more strongly associated than coarse particles with acute
31 respiratory health effects in schoolchildren. Epidemiology. 11:6-10.
32 Schwartz, J.; Zanobetti, A. (2000) Using meta-smoothing to estimate dose-response trends across multiple studies,
33 with application to air pollution and daily death. Epidemiology 11: 666-672.
34 Schwartz, J.; Dockery, D. W.; Neas, L. M.; Wypij, D.; Ware, J. H.; Spengler, J. D.; Koutrakis, P.; Speizer, F. E.;
35 Ferris, B. G., Jr. (1994) Acute effects of summer air pollution on respiratory symptom reporting in children.
36 Am. J. Respir. Crit. Care Med. 150: 1234-1242.
37 Schwartz, J.; Dockery, D. W.; Neas, L. M. (1996a) Is daily mortality associated specifically with fine particles?
38 J. Air Waste Manage. Assoc. 46: 927-939.
39 Schwartz, J.; Spix, C.; Touloumi, G.; Bacharova, L.; Barumamdzadeh, T.; le Tertre, A.; Piekarksi, T.; Ponce de
40 Leon, A.; Ponka, A.; Rossi, G.; Saez, M.; Schouten, J. P. (1996b) Methodological issues in studies of air
41 pollution and daily counts of deaths or hospital admissions. In: St Leger, S., ed. The APHEA project. Short
42 term effects of air pollution on health: a European approach using epidemiological time series data.
43 J. Epidemiol. Community Health 50(suppl. 1): S3-S11.
44 Schwartz, J.; Norris, G.; Larson, T.; Sheppard, L.; Claiborne, C.; Koenig, J. (1999) Episodes of high coarse particle
45 concentrations are not associated with increased mortality. Environ. Health Perspect. 107: 339-342.
46 Schwarz, G. (1978) Estimating the dimension of a model. Ann. Stat. 6: 461-464.
47 Seaton, A.; Soutar, A.; Crawford, V.; Elton, R.; McNerlan, S.; Cherrie, J.; Watt, M.; Agius, R.; Stout, R. (1999)
48 Particulate air pollution and the blood. Thorax 54: 1027-1032.
49 Segala, C.; Fauroux, B.; Just, J.; Pascual, L.; Grimfeld, A.; Neukirch, F. (1998) Short-term effect of winter air
50 pollution on respiratory health of asthmatic children in Paris. Eur. Respir. J. 11: 677-685.
51 Sheppard, L.; Levy, D.; Norris, G.; Larson, T. V.; Koenig, J. Q. (1999) Effects of ambient air pollution on
52 nonelderly asthma hospital admissions in Seattle, Washington, 1987-1994. Epidemiology 10: 23-30.
53 Shumway, R. H.; Tai, R. Y.; Tai, L. P.; Pawitan, Y. (1983) Statistical analysis of daily London mortality and
54 associated weather and pollution effects. Sacramento, CA: California Air Resources Board; contract no.
55 Al-154-33.
March 2001 6-283 DRAFT-DO NOT QUOTE OR CITE
-------
1 Shumway, R. H.; Azari, A. S.; Pawitan, Y. (1988) Modeling mortality fluctuations in Los Angeles as functions of
2 pollution and weather effects. Environ. Res. 45: 224-241.
3 Simpson, R. W.; Williams, G.; Petroeschevsky, A.; Morgan, G.; Rutherford, S. (1997) Associations between
4 outdoor air pollution and daily mortality in Brisbane, Australia. Arch. Environ. Health 52: 442-454.
5 Smith, M. A.; Jalaludin, B.; Byles, J. E.; Lim, L.; Leeder, S. R. (1996) Asthma presentations to emergency
6 departments in western Sydney during the January 2194 bushfires. Int. J. Epidemiol. 25: 1227-1236.
7 Smith, R. L.; Spitzner, D.; Kim, Y.; Fuentes, M. (2000) Threshold dependence of mortality effects for fine and
8 coarse particles in Phoenix, Arizona. J. Air Waste Manage. Assoc. 50: 1367-1379.
9 Spektor, D. M.; Hofmeister, V. A.; Artaxo, P.; Brague, J. A. P.; Echelar, F.; Nogueira, D. P.; Hayes, C.; Thurston,
10 G. D.; Lippmann, M. (1991) Effects of heavy industrial pollution on respiratory function in the children of
11 Cubatao, Brazil: a preliminary report. Environ. Health Perspect. 94: 51-54.
12 Spix, C.; Heinrich, J.; Dockery, D.; Schwartz, J.; Volksch, G.; Schwinkowski, K.; Collen, C.; Wichmann, H. E.
13 (1993) Air pollution and daily mortality in Erfurt, East Germany, 1980-1989. Environ. Health Perspect.
14 101:518-526.
15 Spix, C.; Anderson, H. R.; Schwartz, J.; Vigotti, M. A.; LeTertre, A.; Vonk, J. M.; Touloumi, G.; Balducci, F.;
16 Piekarski, T.; Bacharova, L.; Tobias, A.; Ponka, A.; Katsouyanni, K. (1998) Short-term effects of air
17 pollution on hospital admissions of respiratory diseases in Europe: a quantitative summary of APHEA study
18 results. Arch. Environ. Health 53: 54-64.
19 Stieb, D. M.; Burnett, R. T.; Beveridge, R. C.; Brook, J. R. (1996) Association between ozone and asthma
20 emergency department visits in Saint John, New Brunswick, Canada. Environ. Health Perspect.
21 104: 1354-1360.
22 Stieb, D. M.; Beveridge, R. C.; Rowe, B. H.; Walter, S. D.; Judek, S. (1998a) Assessing diagnostic classification in
23 an emergency department: implications for daily time series studies of air pollution. Am. J. Epidemiol.
24 148:666-670.
25 Stieb, D. M.; Brook, J. R.; Broder, I.; Judek, S.; Burnett, R. T.; Beveridge, R. C. (1998b) Personal exposure of
26 adults with cardiorespiratory disease to particulate acid and sulfate in Saint John, New Brunswick, Canada.
27 Appl. Occup. Environ. Hyg. 13: 461-468.
28 Stieb, D. M.; Beveridge, R. C.; Brook, J. R.; Smith-Doiron, M.; Burnett, R. T.; Dales, R. E.; Beaulieu, S.; Judek, S.;
29 Mamedov, A. (2000) Air pollution, aeroallergens and cardiorespiratory emergency department visits in Saint
30 John, Canada. J. Exposure Anal. Environ. Epidemiol.: 10: 461-477.
31 Studnicka, M. J.; Frischer, T.; Meinert, R.; Studnicka-Benke, A.; Hajek, K.; Spengler, J. D.; Neumann, M. G.
32 (1995) Acidic particles and lung function in children: a summer camp study in the Austrian Alps. Am. J.
3 3 Respir. Crit. Care Med. 151: 423-430.
34 Sunyer, J.; Spix, C.; Quenel, P.; Ponce-de-Leon, A.; Ponka, A.; Barumandzadeh, T.; Touloumi, G.; Bacharova, L.;
35 Wojtyniak, B.; Vonk, J.; Bisanti, L.; Schwartz, J.; Katsouyanni, K. (1997) Urban air pollution and emergency
36 admissions for asthma in four European cities: the APHEA project. Thorax 52: 760-765.
37 Sunyer, J.; Schwartz, J.; Tobias, A.; Macfarlane, D.; Garcia, J.; Anto, J. M. (2000) Patients with chronic obstructive
38 pulmonary disease are at increased risk of death associated with urban particle air pollution: a case-crossover
39 analysis. Am. J. Epidemiol. 151: 50-56.
40 Taggart, S. C. O.; Custovic, A.; Francis, H. C.; Faragher, E. B.; Yates, C. J.; Higgins, B. G.; Woodcock, A. (1996)
41 Asthmatic bronchial hyperresponsiveness varies with ambient levels of summertime air pollution. Eur.
42 Respir. J. 9: 1146-1154.
43 Tanaka, H.; Honma, S.; Nishi, M.; Igarashi, T.; Teramoto, S.; Nishio, F.; Abe, S. (1998) Acid fog and hospital visits
44 for asthma: an epidemiological study. Eur. Respir. J. 11: 1301-1306.
45 Tenias, J. M.; Ballester, F.; Rivera, M. L. (1998) Association between hospital emergency visits for asthma and air
46 pollution in Valencia, Spain. Occup. Environ. Med. 55: 541-547.
47 Thurston, G. D.; Lippmann, M.; Scott, M. B.; Fine, J. M. (1997) Summertime haze air pollution and children with
48 asthma. Am. J. Respir. Crit. Care Med. 155: 654-660.
49 Tiittanen, P.; Timonen, K. L.; Ruuskanen, J.; Mirme, A.; Pekkanen, J. (1999) Fine particulate air pollution,
50 resuspended road dust and respiratory health among symptomatic children. Eur. Respir. J. 13: 266-273.
51 Timonen, K. L.; Pekkanen, J. (1997) Air pollution and respiratory health among children with asthmatic or cough
52 symptoms. Am. J. Respir. Crit. Care Med. 156: 546-552.
53 Tobias, A.; Campbell, M. J. (1999) Modelling influenza epidemics in the relation between black smoke and total
54 mortality. A sensitivity analysis. J. Epidemiol. Community Health 53: 583-584.
March 2001 6-284 DRAFT-DO NOT QUOTE OR CITE
-------
1 Tolbert, P. E.; Klein, M.; Metzger, K. B.; Flanders, W. D.; Todd, K.; Mulholland, J. A.; Ryan, P. B.; Frumkin, H.
2 (2000a) Interim results of the study of particulates and health in Atlanta (SOPHIA). J. Exposure Anal.
3 Environ. Epidemiol. 10: 446-460.
4 Tolbert, P. E.; Mulholland, J. A.; Macintosh, D. L.; Xu, F.; Daniels, D.; Devine, O. J.; Carlin, B. P.; Klein, M.;
5 Dorley, J.; Butler, A. J.; Nordenberg, D. F.; Frumkin, H.; Ryan, P. B.; White, M. C. (2000b) Air quality and
6 pediatric emergency room visits for asthma in Atlanta, Georgia. Am. J. Epidemiol. 151: 798-810.
7 Touloumi, G.; Katsouyanni, K.; Zmirou, D.; Schwartz, J.; Spix, C.; Ponce de Leon, A.; Tobias, A.; Quennel, P.;
8 Rabczenko, D.; Bacharova, L.; Bisanti, L.; Vonk, J. M.; Ponka, A. (1997) Short-term effects of ambient
9 oxidant exposure on mortality: a combined analysis within the APHEA project. Am. J. Epidemiol.
10 146: 177-185.
11 Tsai, F. C.; Apte, M. G.; Daisey, J. M. (2000) An exploratory analysis of the relationship between mortality and the
12 chemical composition of airborne particulate matter. Inhalation Toxicol. 12(suppl.): 121-135.
13 Turnovska, T.; Kostianev, S. (1999) Effects of reduced air pollution on children's pulmonary function. Cent. Eur. J.
14 Public Health 7: 77-79.
15 U.S. Environmental Protection Agency. (1986) Second addendum to air quality criteria for particulate matter and
16 sulfur oxides (1982): assessment of newly available health effects information. Research Triangle Park, NC:
17 Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; EPA report
18 no. EPA-600/8-86-020F. Available from: NTIS, Springfield, VA; PB87-176574.
19 U.S. Environmental Protection Agency. (1996) Air quality criteria for particulate matter. Research Triangle Park,
20 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
21 Van Der Zee, S. C.; Hoek, G.; Boezen, H. M.; Schouten, J. P.; Van Wijnen, J. H.; Brunekreef, B. (1999) Acute
22 effects of urban air pollution on respiratory health of children with and without chronic respiratory
23 symptoms. Occup. Environ. Med. 56: 802-813.
24 Van Der Zee, S. C.; Hoek, G.; Boezen, M. H.; Schouten, J. P.; Van Wijnen, J. H.; Brunekreef, B. (2000) Acute
25 effects of air pollution on respiratory health of 50-70 yr old adults. Eur. Respir. J. 15: 700-709.
26 Vedal, S.; Petkau, J.; White, R.; Blair, J. (1998) Acute effects of ambient inhalable particles in asthmatic and
27 nonasthmatic children. Am. J. Respir. Crit. Care Med. 157: 1034-1043.
28 Veterans Administration Cooperative Study Group on Antihypertensive Agents. (1967) Effects of treatment on
29 morbidity in hypertension. Results in patients with diastolic blood pressures averaging 115 through 129 mm
30 Hg. JAMA J. Am. Med. Assoc. 202: 1028-1034.
31 Veterans Administration Cooperative Study Group on Antihypertensive Agents. (1970) Effects of treatment on
32 morbidity in hypertension. II. Results in patients with diastolic blood pressure averaging 90 through 114 mm
33 Hg. JAMA J. Am. Med. Assoc. 213: 1143-1152.
34 Vigotti, M. A.; Rossi, G.; Bisanti, L.; Zanobetti, A.; Schwartz, J. (1996) Short term effects of urban air pollution on
35 respiratory health in Milan, Italy, 1980-89. In: Leger, S., ed. The APHEA project. Short term effects of air
36 pollution on health: a European approach using epidemiological time series data. J. Epidemiol. Community
37 Health 50(suppl. 1): S71-S75.
38 Wallace, L. (1996) Indoor particles: a review. J. Air Waste Manage. Assoc. 46: 98-126.
39 Wallace, L. (2000) Correlations of personal exposure to particles with outdoor air measurements: a review of recent
40 studies. Aerosol Sci. Technol. 32: 15-25.
41 Wang, B.; Peng, Z.; Zhang, X.; Xu, Y.; Wang, H.; Allen, G.; Wang, L.; Xu, X. (1999) Particulate matter, sulfur
42 dioxide, and pulmonary function in never-smoking adults in Chongqing, China. Int. J. Occup. Environ
43 Healths: 14-19.
44 Ward, D. J.; Miller, M. R.; Walters, S.; Harrison, R. M.; Ayres, J. G. (2000) Impact of correcting peak flow for
45 nonlinear errors on air pollutant effect estimates from a panel study. Eur. Respir. J. 15: 137-140.
46 Ware, J. H.; Ferris, B. G., Jr.; Dockery, D. W.; Spengler, J. D.; Stram, D. O.; Speizer, F. E. (1986) Effects of
47 ambient sulfur oxides and suspended particles on respiratory health of preadolescent children. Am. Rev.
48 Respir. Dis. 133: 834-842.
49 Wichmann, H.-E.; Spix, C.; Tuch, T.; Wolke, G.; Peters, A.; Heinrich, J.; Kreyling, W. G.; Heyder, J. (2000) Daily
50 mortality and fine and ultrafine particles in Erfurt, Germany. Part I: role of particle number and particle mass.
51 Cambridge, MA: Health Effects Institute; Research Report no. 98.
52 Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
53 epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
March 2001 6-285 DRAFT-DO NOT QUOTE OR CITE
-------
1 Wilson, W. E.; Mage, D. T.; Grant, L. D. (2000) Estimating separately personal exposure to ambient and
2 nonambient participate matter for epidemiology and risk assessment: why and how. J. Air Waste Manage.
3 Assoc. 50: 1167-1183.
4 Wolff, G. T.; Stroup, C. M.; Stroup, D. P. (1983) The coefficient of haze as a measure of particulate elemental
5 carbon. J. Air Pollut. Control Assoc. 33: 746-750.
6 Wong, C. M.; Lam, T. H.; Peters, J.; Hedley, A. J.; Ong, S. G.; Tarn, A. Y. C.; Liu, J.; Spiegelhalter, D. J. (1998)
7 Comparison between two districts of the effects of an air pollution intervention on bronchial responsiveness
8 in primary school children in Hong Kong. J. Epidemiol. Community Health 52: 571-578.
9 Wong, C. M.; Hu, Z. G.; Lam, T. H.; Hedley, A. J.; Peters, J. (1999) Effects of ambient air pollution and
10 environmental tobacco smoke on respiratory health of non-smoking women in Hong Kong. Int. J. Epidemiol.
11 28: 859-864.
12 Woodruff, T. J.; Grille, J.; Schoendorf, K. C. (1997) The relationship between selected causes of postneonatal
13 infant mortality and particulate air pollution in the United States. Environ. Health Perspect. 105: 608-612.
14 Wordley, J.; Walters, S.; Ayres, J. G. (1997) Short term variations in hospital admissions and mortality and
15 particulate air pollution. Occup. Environ. Med. 54: 108-116.
16 Xu, Z.; Yu, D.; Jing, L.; Xu, X. (2000) Air pollution and daily mortality in Shenyang, China. Arch. Environ. Health
17 55:115-120.
18 Yang, W.; Jennison, B. L.; Omaye, S. T. (1997) Air pollution and asthma emergency room visits in Reno, Nevada.
19 Inhalation Toxicol. 9: 15-29.
20 Ye, S.-H.; Zhou, W.; Song, J.; Peng, B.-C.; Yuan, D.; Lu, Y.-M.; Qi, P.-P. (2000) Toxicity and health effects of
21 vehicle emissions in Shanghai. Atmos. Environ. 34: 419-429.
22 Zanobetti, A.; Schwartz, J. (2000) Race, gender, and social status as modifiers of the effects of PM10 on mortality.
23 J. Occup. Environ. Med. 42: 469-474.
24 Zanobetti, A.; Wand, M. P.; Schwartz, J.; Ryan, L. M. (2000a) Generalized additive distributed lag models:
25 quantifying mortality displacement. Biostatistics 1: 279-292.
26 Zanobetti, A.; Schwartz, J.; Dockery, D. W. (2000b) Airborne particles are a risk factor for hospital admissions for
27 heart and lung disease. Environ. Health Perspect. 108: 1071-1077.
28 Zeger, S. L.; Dominici, F.; Samet, J. (1999) Harvesting-resistant estimates of air pollution effects on mortality.
29 Epidemiology 10: 171-175.
30 Zeger, S. L.; Thomas, D.; Dominici, F.; Samet, J. M.; Schwartz, J.; Dockery, D.; Cohen, A. (2000) Exposure
31 measurement error in time-series studies of air pollution: concepts and consequences. Environ. Health
32 Perspect. 108: 419-426.
33 Zeghnoun, A.; Beaudeau, P.; Carrat, F.; Delmas, V.; Boudhabhay, O.; Gayon, F.; Guincetre, D.; Czernichow, P.
34 (1999) Air pollution and respiratory drug sales in the city of Le Havre, France, 1993-1996. Environ. Res.
35 81:224-230.
36 Zemp, E.; Elsasser, S.; Schindler, C.; Kiinzli, N.; Perruchoud, A. P.; Domenighetti, G.; Medici, T.;
37 Ackermann-Liebrich, U.; Leuenberger, P.; Monn, C.; Bolognini, G.; Bongard, J.-P.; Brandli, O.; Karrer, W.;
38 Keller, R.; Schoni, M. H.; Tschopp, J.-M.; Villiger, B.; Zellweger, J.-P.; SAPALDIA Team. (1999)
39 Long-term ambient air pollution and respiratory symptoms in adults (SAPALDIA study). Am. J. Respir. Crit.
40 Care Med. 159: 1257-1266.
41 Zhang, J.; Qian, Z.; Kong, L.; Zhou, L.; Yan, L.; Chapman, R. S. (1999) Effects of air pollution on respiratory
42 health of adults in three Chinese cities. Arch. Environ. Health 54: 373-381.
43 Zhang, H.; Triche, E.; Leaderer, B. (2000) Model for the analysis of binary time series of respiratory symptoms.
44 Am.J. Epidemiol. 151: 1206-1215.
45 Zidek, J. V.; Wong, H.; Le, N. D.; Burnett, R. (1996) Causality, measurement error and multicollinearity in
46 epidemiology. Environmetrics 7: 441-451.
47 Zmirou, D.; Schwartz, J.; Saez, M.; Zanobetti, A.; Wojtyniak, B.; Touloumi, G.; Spix, C.; Ponce de Leon, A.;
48 Le Moullec, Y.; Bacharova, L.; Schouten, J.; Ponka, A.; Katsouyanni, K. (1998) Time-series analysis of air
49 pollution and cause-specific mortality. Epidemiology 9: 495-503.
March 2001 6-286 DRAFT-DO NOT QUOTE OR CITE
-------
i APPENDIX 6A
2
3
4 Demographic and Pollution Data for 90-City Analysis
of NMMAPS Project
6
7
March 2001 6A-1 DRAFT-DO NOT QUOTE OR CITE
-------
2
P
3
£3"
O
O
ON
K>
O
>
Tl
H
1
O
o
^
0
H
c
o
H
W
O
J*3
H- -t
H
m
TABLE 6A-1. THE 90 CITIES AND THEIR INCLUDED COUNTIES BY POPULATION SIZE WITH MEAN DAILY
NUMBER OF DEATHS BY CATEGORY (1987-1994). Evaluated in NMMAPS 90-Cities Analyses, Samet et al. (2000a,b).
CVD/Respiratory
City
Los Angeles
New York
Chicago
Dallas/Fort Worth
Houston
San Diego
Santa Ana/ Anaheim
Phoenix
Detroit
Miami
Philadelphia
Minneapolis/St. Paul
Seattle
San Jose
Cleveland
San Bernardino
Pittsburgh
Oakland
Atlanta
San Antonio
Riverside
Denver
Sacramento
St. Louis
Buffalo
Abbreviation
la
ny
chic
dlft
hous
sand
staa
phoe
del
miam
phil
minn
seat
sanj
clev
sanb
pitt
oakl
atla
sana
rive
denv
sacr
stlo
buff
County
Los Angeles
Bronx, Kings, New York,
Richmond, Queens, Westchester
Cook
Collin, Dallas, Rockwall, Tarrant
Harris
San Diego
Orange
Maricopa
Wayne
Dade
Philadelphia
Hennepin, Ramsey
King
Santa Clara
Cuyahoga
San Bernardino
Allegheny
Alameda
Fulton, De Kalb
Bexar
Riverside
Denver, Adams, Arapahoe
Sacramento
St. Louis City
Erie
State
CA
NY
IL
TX
TX
CA
CA
AZ
MI
FL
PA
MN
WA
CA
OH
CA
PA
CA
GA
TX
CA
CO
CA
MO
NY
Population
8,863,164
8,197,430
5,105,067
3,312,553
2,818,199
2,498,016
2,410,556
2,122,101
2,111,687
1,937,094
1,585,577
1,518,196
1,507,319
1,497,577
1,412,140
1,418,380
1,336,449
1,279,182
1,194,788
1,185,394
1,170,413
1,124,159
1,041,219
993,529
968,532
Total
148.1
190.9
113.9
47.9
39.9
41.6
32.4
38.4
46.9
43.8
42.3
26.3
25.6
19.7
36.5
20.6
37.6
22.2
17.5
20.1
20.1
9.1
17.2
10.7
25.2
Disease
87.0
108.3
62.0
26.0
20.0
22.6
18.7
20.9
26.5
23.6
21.5
13.9
13.4
10.7
20.1
12.1
21.0
12.2
8.8
10.5
12.4
5.0
9.5
6.0
14.8
Other
61.1
82.6
51.9
21.9
19.8
19.0
13.6
17.5
20.4
20.2
20.8
12.4
12.2
9.0
16.4
8.5
16.9
10.0
8.7
9.6
7.7
4.1
7.7
4.7
10.3
-------
P3
cr
0
0
>
1
a
£>
"Tl
H
a
o
2;
o
H
O
c
o
H
W
O
n
DAILY NUMBER OF DEATHS BY CATEGORY (1987-1994). Evaluated in NMMAPS 90-Cities Analyses,
Samet et al. (2000a,b).
City
Columbus
Cincinnati
St. Petersburg
Kansas City
Honolulu
Tampa
Memphis
Indianapolis
Newark
Baltimore
Salt Lake City
Rochester
Worcester
Orlando
Jacksonville
Fresno
Louisville
Boston
Birmingham
Washington
Oklahoma City
Providence
El Paso
Tacoma
Austin
Abbreviation County
clmo
cine
stpe
kan
hono
tamp
memp
indi
nwk
bait
salt
roch
wor
orla
jckv
fres
loui
bost
birm
dc
okla
prov
elpa
taco
aust
Franklin
Hamilton
Pinellas
Clay, Jackson, Platte
Honolulu
Hillsborough
Shelby
Marion
Essex
Baltimore City
Salt Lake
Monroe
Worcester
Orange
Duval
Fresno
Jefferson
Suffolk
Jefferson
Washington DC
Oklahoma
Providence
El Paso
Pierce
Travis
State
OH
OH
FL
MO
HI
FL
TN
IN
NJ
MD
UT
NY
MA
FL
FL
CA
KY
MA
AL
DC
OK
RI
TX
WA
TX
Population
961,437
866,228
851,659
844,510
836,231
834,054
826,330
797,159
778,206
736,014
725,956
713,968
709,705
677,491
672,971
667,490
664,937
663,906
651,525
606,900
599,611
596,270
591,610
586,203
576,407
CVD/Respiratory
Total Disease
16.8
19.9
29.3
16.7
11.9
16.9
17.5
16.9
18.4
20.2
9.3
14.6
15.2
11.0
13.0
11.1
16.3
13.2
16.2
15.5
12.9
14.6
7.7
10.0
7.0
8.9
11.0
17.7
9.3
6.4
9.1
9.7
9.0
8.7
9.8
4.9
7.9
8.2
5.8
7.0
6.2
8.8
6.5
8.5
7.0
7.3
7.9
3.8
5.7
3.4
Other
7.9
8.9
11.6
7.5
5.5
7.8
7.7
8.0
9.7
10.4
4.4
6.7
6.9
5.2
6.0
4.9
7.5
6.7
7.7
8.5
5.6
6.7
3.9
4.3
3.6
-------
p
o
TABLE 6A-1 (cont'd). THE
DAILY NUMBER OF
90 CITIES AND THEIR INCLUDED COUNTIES BY POPULATION SIZE WITH MEAN
DEATHS BY CATEGORY (1987-1994). Evaluated in NMMAPS 90-Cities Analyses,
bo
o
I— t
ON
•^
O
^
H
6
o
z
0
H
O
G
0
H
W
O
o
H
w
aanu
;i ei ai. (zuuua,c
>;•
CVD/Respiratory
City
Dayton
Jersey City
Bakersfield
Akron
Charlotte
Nashville
Tulsa
Grand Rapids
New Orleans
Stockton
Albuquerque
Syracuse
Toledo
Raleigh
Wichita
Colorado Springs
Baton Rouge
Modesto
Madison
Spokane
Little Rock
Greensboro
Knoxville
Shreveport
Des Moines
Abbreviation
dayt
jers
bake
akr
char
nash
tuls
gdrp
no
stoc
albu
syra
tole
ral
wich
colo
batr
mode
madi
spok
Itrk
grnb
knox
shr
desm
County
Montgomery
Hudson
Kern
Summit
Mecklenburg
Davidson
Tulsa
Kent
Orleans
San Joaquin
Bernalillo
Onondaga
Lucas
Wake
Sedwick
El Paso
East Baton Rouge
Stanislaus
Dane
Spokane
Pulaski
Guilford
Knox
Bossier, Caddo
Polk
State
OH
NJ
CA
OH
NC
TN
OK
MI
LA
CA
NM
NY
OH
NC
KS
CO
LA
CA
WI
WA
AR
NC
TN
LA
IA
Population
573,809
553,099
543,477
514,990
511,433
510,784
503,341
500,631
496,938
480,628
480,577
468,973
462,361
423,380
403,662
397,014
380,105
370,522
367,085
361,364
349,660
347,420
335,749
334,341
327,140
Total
11.9
11.5
8.6
10.7
8.5
11.0
10.0
8.7
12.0
8.5
7.6
9.7
10.8
5.6
7.2
5.0
6.3
6.6
5.3
7.8
7.0
6.9
6.7
6.8
6.1
Disease
6.5
5.9
5.0
5.8
4.3
6.0
5.8
4.9
5.9
4.8
3.8
5.4
6.3
2.9
4.0
2.8
3.4
3.8
2.9
4.5
3.7
3.8
3.5
3.7
3.4
Other
5.4
5.6
3.6
4.9
4.2
5.0
4.2
3.8
6.1
3.6
3.8
4.3
4.5
2.7
3.3
2.3
3.0
2.8
2.4
3.3
3.3
3.1
3.1
3.1
2.6
-------
o
TABLE 6A-1 (cont'd). THE 90 CITIES AND THEIR INCLUDED COUNTIES BY POPULATION SIZE WITH MEAN
DAILY NUMBER OF DEATHS BY CATEGORY (1987-1994). Evaluated in NMMAPS 90-Cities Analyses,
Samet et al. (2000a,b).
'•w'
O
ON
i
L/1
O
£
H
1
D
0
O
H
0
d
0
H
W
O
7*
o
H
m
CVD/Respiratory
City
Fort Wayne
Corpus Chrisit
Norfolk
Jackson
Hunts ville
Anchorage
Lexington
Lubbock
Richmond
Arlington
Kingston
Evansville
Kansas City
Olympia
Topeka
Abbreviation
ftwa
corp
nor
jcks
hunt
anch
lex
lubb
rich
arlv
king
evan
kans
olym
tope
County
Allen
Nueces
Norfolk
Hinds
Madison
Anchorage
Fayette
Lubbock
Richmond City
Arlington
Ulster
Vanderburgh
Wyandotte
Thurston
Shawnee
State
IN
TX
VA
MS
AL
AK
KY
TX
VA
VA
NY
IN
KS
WA
KS
Population
300,836
291,145
261,229
254,441
238,912
226,338
225,366
222,636
203,056
170,936
165,304
165,058
161,993
161,238
160,976
Total
5.9
4.9
4.8
5.3
3.9
1.9
4.1
3.9
5.1
2.4
3.0
4.4
3.2
2.8
3.6
Disease
3.4
2.5
2.6
3.0
2.2
0.8
2.1
2.3
2.7
1.3
1.8
2.5
1.8
1.5
2.0
Other
2.5
2.4
2.2
2.3
1.7
1.1
2.0
1.6
2.4
1.2
1.2
1.9
1.4
1.3
1.6
-------
TABLE 6A-2. MEAN DAILY POLLUTION LEVELS BY CITY (1987-1994)
Evaluated in NMMAPS 90-Cities Analyses (Samet et al., 2000a,b)
City
Los Angeles
New York
Chicago
Dallas/Ft. Worth
Houston
San Diego
Santa Ana/Anaheim
Phoenix
Detroit
Miami
Philadelphia
Minneapolis/St. Paul
Seattle
San Jose
Cleveland
San Bernardino
Pittsburgh
Oakland
Atlanta
San Antonio
Riverside
Denver
Sacramento
St. Louis
Buffalo
Columbus
Cincinnati
St. Petersburg
Kansas City
Honolulu
Tampa
Memphis
Indianapolis
Newark
Baltimore
PM,o
46.0
28.8
35.6
23.8
30.0
33.6
37.4
40.3
40.9
25.7
35.4
26.9
25.3
30.4
45.1
37.0
31.6
26.3
36.1
23.8
52.0
29.6
33.3
30.1
21.7
29.0
34.2
23.5
25.9
15.3
28.3
30.3
32.0
32.9
32.9
03
ppb
22.8
19.6
18.6
25.3
20.5
31.6
23.0
22.5
22.6
25.9
20.5
NA
19.4
17.9
27.4
35.9
20.7
17.2
25.1
22.2
33.4
21.4
26.7
22.8
22.9
26.0
25.8
24.6
27.6
18.9
23.5
29.0
31.9
15.2
21.2
NO2
39.4
38.9
24.3
13.8
18.8
22.9
35.1
16.6
21.3
11.0
32.2
17.6
NA
25.1
25.2
27.9
27.6
21.2
26.0
NA
25.0
27.9
16.3
22.5
19.0
NA
26.7
11.8
9.2
NA
21.2
26.8
20.2
33.6
32.9
SO2
1.9
12.8
4.6
1.1
2.8
1.7
1.3
3.5
6.4
NA
9.9
2.6
NA
NA
10.3
0.7
14.2
NA
6.0
NA
0.4
5.5
NA
11.3
8.6
5.9
11.9
NA
2.4
NA
7.8
6.8
7.7
9.6
8.4
CO
ppm
15.1
20.4
7.9
7.4
8.9
11.0
12.3
12.7
6.6
10.6
11.8
11.8
17.8
9.4
8.5
10.3
12.2
9.1
8.9
10.1
11.2
10.3
9.4
10.5
7.3
7.6
10.0
7.1
6.2
8.3
7.8
11.9
9.0
8.7
9.2
March 2001 6A-6 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 6A-2 (cont'd). MEAN DAILY POLLUTION LEVELS BY CITY (1987-1994)
Evaluated in NMMAPS 90-Cities Analyses (Samet et al., 2000a,b)
City
Salt Lake City
Rochester
Worcester
Orlando
Jacksonville
Fresno
Louisville
Boston
Birmingham
Washington DC
Oklahoma City
Providence
El Paso
Tacoma
Austin
Dayton
Jersey City
Bakersfield
Akron
Charlotte
Nashville
Tulsa
Grand Rapids
New Orleans
Stockton
Albuquerque
Syracuse
Toledo
Raleigh
Wichita
Colorado Springs
Baton Rouge
Modesto
Madison
Spokane
PM,o
/•ig/m3
32.9
21.9
22.2
22.7
29.9
43.4
30.8
26.0
31.2
28.2
25.0
30.9
41.2
28.0
21.1
27.4
30.5
53.2
22.4
30.7
32.4
26.6
22.8
29.0
39.0
16.9
24.5
25.6
25.6
25.6
26.3
27.3
41.7
19.9
36.0
03
ppb
23.0
22.7
30.0
24.1
28.2
29.4
19.8
17.9
22.4
17.5
28.4
25.4
24.4
23.8
25.5
26.6
19.7
33.3
30.5
29.3
16.2
31.4
27.7
20.5
22.6
25.8
23.7
27.1
35.4
24.2
24.3
21.2
26.1
29.7
32.6
NO2
yUg/m3
29.6
NA
25.2
11.4
14.8
21.7
22.4
29.9
NA
25.6
13.9
21.9
23.6
NA
NA
NA
28.7
19.4
NA
16.2
NA
16.6
NA
21.3
24.2
NA
NA
NA
12.7
NA
NA
16.6
24.2
NA
NA
SO2
/"g/m3
4.4
10.4
6.7
1.5
2.2
1.9
8.4
10.0
6.6
11.2
NA
9.5
9.1
6.5
NA
NA
10.7
3.0
12.0
NA
11.6
6.9
3.0
NA
1.7
NA
3.6
5.9
NA
4.8
NA
5.2
1.9
3.3
NA
CO
ppm
13.5
6.3
8.9
9.3
9.2
6.8
11.2
11.3
10.5
12.3
7.1
10.0
12.5
16.6
NA
8.2
20.1
10.5
7.0
11.1
11.2
6.5
5.7
9.4
8.2
7.9
11.7
10.3
16.1
6.5
10.9
4.3
9.1
10.4
21.9
March 2001 6A-7 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 6A-2 (cont'd). MEAN DAILY POLLUTION LEVELS BY CITY (1987-1994)
Evaluated in NMMAPS 90-Cities Analyses (Samet et al., 2000a,b)
City
Little Rock
Greensboro
Knoxville
Shreveport
Des Moines
Fort Wayne
Corpus Christi
Norfolk
Jackson
Huntsville
Anchorage
Lexington
Lubbock
Richmond
Arlington
Kingston
Evansville
Kansas City
Olympia
Topeka
PMIO
Aig/m3
25.8
27.5
36.3
24.7
35.5
23.2
24.7
26.0
26.4
26.0
23.0
24.5
25.1
25.4
22.0
20.4
32.4
43.4
22.7
29.0
03
ppb
27.7
NA
29.6
28.2
11.8
32.1
23.9
NA
23.9
30.4
NA
32.8
NA
NA
29.0
NA
NA
18.5
NA
NA
NO2
/"g/iri3
9.3
NA
NA
NA
NA
NA
NA
19.6
NA
12.9
NA
16.4
NA
23.7
25.5
NA
NA
17.6
NA
NA
S02
/•ig/m3
2.6
4.2
NA
2.3
NA
4.0
1.0
6.7
NA
NA
NA
6.2
NA
5.8
NA
NA
NA
4.7
NA
NA
CO
ppm
NA
12.2
13.6
NA
8.6
14.4
NA
7.3
7.9
6.3
16.1
8.8
NA
6.6
6.6
NA
NA
8.2
12.7
NA
March 200 1
6A-8
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 6A-3. NUMBER OF DAYS FOR WHICH MONITORING WAS AVAILABLE
BY POLLUTANT FOR CITIES (1987-1994). Evaluated in 90-Cities NMMAPS
Analyses of Sametet al. (2000b)
City
Los Angeles
New York
Chicago
Dallas/Fort Worth
Houston
San Diego
Santa Ana/Anaheim
Phoenix
Detroit
Miami
Philadelphia
Minneapolis/St. Paul
Seattle
San Jose
Cleveland
San Bernardino
Pittsburgh
Oakland
Atlanta
San Antonio
Riverside
Denver
Sacramento
St. Louis
Buffalo
Columbus
Cincinnati
St. Petersburg
Kansas City
Honolulu
PM,0
580
489
2,683
624
793
521
480
376
1,348
484
495
2,764
2,205
945
1,298
538
2,899
508
482
670
545
1,645
488
487
489
1,564
1,705
367
670
415
03
2,922
2,922
2,922
2,922
2,922
2,922
2,922
2,554
1,861
2,882
2,901
0
1,820
2,922
1,712
2,922
2,883
2,922
2,200
2,918
2,922
2,922
2,922
1,731
2,884
1,494
1,712
2,920
2,856
1,681
NO2
2,922
2,493
2,922
2,557
2,557
2,922
2,922
740
2,686
2,863
2,554
2,725
0
1,957
2,555
2,922
2,537
2,921
2,922
0
2,904
2,484
2,916
2,919
2,522
0
2,554
2,235
2,922
0
SO2
2,922
2,920
1,409
2,908
2,922
2,922
2,922
1,272
2,922
0
2,919
2,914
0
0
2,922
2,922
2,922
0
2,918
0
2,908
2,860
0
2,919
2,922
964
2,905
0
1,094
0
CO
2,922
2,920
2,922
2,922
2,922
2,922
2,922
2,554
2,922
2,919
2,919
2,918
2,922
2,922
2,897
2,922
2,920
2,922
2,839
2,891
2,921
2,922
2,922
2,920
2,921
2,557
2,922
2,922
2,922
2,919
March 2001
6A-9
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 6A-3 (cont'd). NUMBER OF DAYS FOR WHICH MONITORING WAS
AVAILABLE BY POLLUTANT FOR CITIES (1987-1994). Evaluated in 90-Cities
NMMAPS Analyses of Samet et al. (2000b)
City
Tampa
Memphis
Indianapolis
Newark
Baltimore
Salt Lake City
Rochester
Worcester
Orlando
Jacksonville
Fresno
Louisville
Boston
Birmingham
Washington
Oklahoma City
Providence
El Paso
Tacoma
Austin
Dayton
Jersey City
Bakersfield
Akron
Charlotte
Nashville
Tulsa
Grand Rapids
New Orleans
Stockton
Albuquerque
Syracuse
Toledo
Raleigh
Wichita
PMIO
508
480
1,269
484
1,220
1,356
486
450
421
555
517
485
631
900
417
563
485
2,587
482
646
461
1,367
550
1,495
454
1,989
411
111
531
488
1,200
485
416
480
366
03
2,922
1,707
1,588
2,726
2,063
2,409
2,886
1,763
2,920
2,791
2,922
2,603
2,882
2,200
2,847
2,832
1,634
2,922
1,601
2,909
1,696
2,843
2,557
1,677
1,936
2,861
2,834
1,615
2,889
2,475
2,922
2,864
1,711
1,267
2,913
NO2
941
2,254
2,874
2,882
2,843
1,903
0
2,864
2,024
2,727
2,922
1,604
2,922
0
2,842
2,295
2,441
2,472
0
0
0
2,496
2,557
0
1,593
0
2,462
0
2,879
2,379
0
0
0
1,219
0
SO2
1,818
2,823
2,922
2,896
2,912
2,739
2,921
2,452
2,878
2,738
2,398
2,841
2,922
1,916
2,286
0
2,922
2,906
2,756
0
0
2,918
2,557
2,827
0
2,619
2,426
2,907
0
867
0
2,857
2,921
0
1,423
CO
2,922
2,922
2,922
2,894
2,865
2,922
2,921
2,899
2,921
2,922
2,922
2,922
2,922
2,922
2,341
2,909
2,921
2,922
2,766
0
2,922
2,883
2,659
2,922
2,922
2,771
2,836
2,903
2,922
2,906
2,922
2,908
2,897
2,160
2,922
March 2001
6A-10
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 6A-3 (cont'd). NUMBER OF DAYS FOR WHICH MONITORING WAS
AVAILABLE BY POLLUTANT FOR CITIES (1987-1994). Evaluated in 90-Cities
NMMAPS Analyses of Samet et al. (2000b)
City
Colorado Springs
Baton Rouge
Modesto
Madison
Spokane
Little Rock
Greensboro
Knoxville
Shreveport
Des Moines
Fort Wayne
Corpus Christi
Norfolk
Jackson
Hunts ville
Anchorage
Lexington
Lubbock
Richmond
Arlington
Kingston
Evansville
Kansas City
Olympia
Topeka
PM10
481
474
199
338
2,393
516
445
577
349
1,334
336
613
474
508
1,382
2,379
816
1,306
474
313
323
404
551
1,135
269
03
2,920
2,922
2,496
1,698
974
2,922
0
1,679
2,922
2,782
1,587
2,919
0
2,191
2,173
0
1,709
0
0
1,705
0
0
2,890
0
0
NO2
0
2,880
2,449
0
0
2,921
0
0
0
0
0
0
1,787
0
1,090
0
2,871
0
2,537
2,306
0
0
324
0
0
SO2
0
2,891
845
2,432
0
2,908
1,077
0
2,881
0
1,219
2,920
2,148
0
0
0
2,906
0
2,907
0
0
0
2,909
0
0
CO
2,922
2,888
2,892
2,709
2,922
0
1,855
2,511
0
2,825
1,822
0
2,921
2,574
2,532
1,488
2,865
0
2,922
2,896
0
0
2,775
950
0
March 2001
6A-11
DRAFT-DO NOT QUOTE OR CITE
-------
REFERENCES
Samet, J. M.; Dominici, F.; Zeger, S. L.; Schwartz, J.; Dockery, D. W. (2000a) National morbidity, mortality, and
air pollution study. Part I: methods and methodologic issues. Cambridge, MA: Health Effects Institute;
research report no. 94.
Samet, J. M.; Zeger, S. L.; Dominici, F.; Curriero, F.; Coursac, I.; Dockery, D. W.; Schwartz, J.; Zanobetti, A.
(2000b) The national morbidity, mortality, and air pollution study. Part II: morbidity, mortality, and air
pollution in the United States. Cambridge, MA: Health Effects Institute; research report no. 94.
March 2001 6A-12 DRAFT-DO NOT QUOTE OR CITE
-------
APPENDIX 6B
Heart Rate Variability as a Predictor of Serious Cardiac Outcomes
March 2001 6B-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 As an adjunct to discussion of newly emerging literature evaluating relationships between
2 ambient PM and heart rate variability (HRV), factors affecting HR are reviewed briefly and
3 summarized here. More detail description of HRV, and its measurement and interpretation can
4 be found elsewhere (1996 Task Force of the European Society of Cardiology and the North
5 American Society of Pacing and Electrophysiology).
6
7 Factors Affecting Heart Rate. The heart has a spontaneous rhythm of approximately 100
8 beats/mm in the absence of extrinsic influences, because the electrical signal triggering heartbeat
9 originates in and spreads throughout the heart via a specialized conduction system. The tissue
10 structures that comprise the conduction system of the heart include the sinoatrial (SA) node, the
11 internodal pathways, the atrioventricular (AV) node, the bundle of His and its branches, and the
12 Purkinje system. Although all parts of the conduction system are capable of spontaneous
13 electrical discharge and heartbeat initiation, it is the SA node (with its higher rate of discharge)
14 that is the normal cardiac pacemaker in the healthy heart. The spontaneous discharge rate of the
15 SA node, and therefore heartbeat, is modulated by nervous impulses and by circulating
16 substances, such as epinephrine that originate outside the heart. One category of modulating
17 input to the heart is through the sympathetic and parasympathetic divisions of the autonomic
18 nervous system via numerous nerve fibers that innervate the heart.
19 Stimulation of the heart via parasympathetic nerve fibers decreases the rate of discharge of
20 the SA node, thereby decreasing HR (bradycardia), and decreases the excitability of the AV
21 junctional fibers between the atrial musculature and the AV node, thereby slowing transmission
22 of the impulse into the ventricles. Stimulation of the heart via sympathetic fibers increases the
23 rate of discharge of the SA node, thereby increasing HR (tachycardia) and increasing the
24 excitability of the AV node and increasing transmission of the cardiac impulse into the ventricle.
25 During the resting state parasympathetic input to the heart predominates, so the normal resting
26 HR is well below the inherent rate of 100 beats/min. The HR along with stroke volume
27 determines cardiac output, which interacts with peripheral resistance to determine blood pressure.
28 The autonomic control of HR is modulated by the vasomotor center located in the brain in
29 the reticular substance of the medulla and lower third of the pons. Impulses sent forth from the
30 vasomotor center through the parasympathetic and sympathetic neurons regulate HR and
31 vasomotor tone. The medial portion of the vasomotor center transmits inhibitory impulses that
March 2001 6B-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 decrease HR through the parasympathetic nerve fibers (vagus nerve). The lateral portions of the
2 vasomotor center transmit excitatory impulses that increase both HR and contractility through
3 sympathetic nerve fibers to the heart. In this way, the vasomotor center can either increase or
4 decrease HR, as well as vasomotor tone. The vasomotor center, in turn, is influenced by
5 impulses arising in higher centers of the brain.
6 Thus, HR is the resultant of the intrinsic rate of the heart modified by various internal and
7 external factors. Most important of these is the output of the vasomotor center delivered via the
8 autonomic nervous system. Other factors affecting HRV include exercise and changes in
9 ambient temperature and oxygen tension.
10
11 Measures of Heart Rate Variability. Heart rate variability is being used increasingly in
12 applications from basic research to clinical practice (Berntson et al., 1997). Meaningful analysis
13 of HRV is dependent on fidelity of the basic cardiac input signal that is derived from the
14 electrocardiogram (ECG). This signal is digitized and a series of intervals between successive
15 R (R-R) waves are determined. The population of R-R intervals or pairs of R-R intervals are
16 treated as if they were a set of temporarily unordered data. The variability of these measures is
17 expressed either by conventional statistical measures (Malik, 1995) or other analytical methods,
18 whereby specific patterns of HRV may be related to specific physiological processes and
19 mechanisms.
20 A wide variety of estimates of HRV have been described. The Task Force of the European
21 Society of Cardiology and the North American Society of Pacing and Electrophysiology (1996)
22 has recommended standard time domain measures that index overall heart rate variability, short-
23 term heart rate variability, and long-term heart rate variability (Table 9-9). A measure
24 recommended for HRV is the standard deviation of all normal-to-normal (N-N), also designated
25 as R-R, heart beat intervals (SDNN). The recommended estimate of short-term variability is the
26 root mean square of the successive beat differences (rMSSD) and that for longer-term variability
27 is the standard deviation of the mean N-N interval for each 5-min segment of recording
28 (SDANN).
29 Periodic components of HRV tend to aggregate within several frequency domains (see
30 Table 6B-1). In young healthy individuals at rest, the most conspicuous frequency band is at the
31 normal respiratory frequency of 0.15 to 0.4 Hz and is termed high-frequency (HF) domain. The
March 2001 6B-3 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 6B-1. TERMS USED IN EXPRESSING HEART RATE VARIABILITY
Abbreviation
HR
HRV
SDNN
rMSSD
SPANN
HF Domain
LF Domain
VLF Domain
Definition
Heart rate (beats/min)
Heart rate variability
Standard deviation of all normal to normal heart
beat intervals
Root mean square of successive beat differences
Standard deviation of the mean normal to normal
(N-N) interval for each 5 min
0.15 -0.4 HZ
0.05 -0.1 5 HZ
0.003 - 0.05 HZ
Domain
—
—
Time
Time
Time
Frequency
Frequency
Frequency
1 band from 0.05 to 0.15 Hz is termed low-frequency (LF) domain. Other domains have been
2 described including very low frequencies (VLF), 0.003 to 0.05 Hz, and ultra-low frequencies
3 (ULF) that include circadian rhythms. Thus, HRV is quantitated by both time domain metrics
4 (NN, SDNN, rMSSD, and SDANN) and frequency domain metrics (HF, LF, and ULF).
5
6 Factors Affecting Heart Rate Variability. Heart rate variability after a myocardial infarction is
7 associated with increased mortality (Kleiger et al., 1987). Aging and gender also are associated
8 with depressed HRV (Umetani et al., 1998). Reardon and Malik (1996) examined the affect of
9 aging in healthy subjects (age range 40 to 102 years; 39 women) with normal resting ECGs.
10 In all subjects, 24-h Holter recordings were performed and used to measure HRV. The HRV
11 triangular index decreased significantly with age, whereas rMSSD did not change. There was a
12 significant difference in HRV index in subjects >70 years compared with those <70 years. There
13 was no significant difference in rMSSD between the two age groups. The authors conclude that
14 aging reduces HRV and decreased HRV may reflect reduced responsiveness of autonomic
15 activity to external environmental stimuli with age.
16
March 2001
6B-4
DRAFT-DO NOT QUOTE OR CITE
-------
1 Umetani et al. (1998) studied the effects of age and gender on 24-h HR and HRV in healthy
2 subjects (10 to 99 years old; 112 male and 148 female). The authors conclude that (1) HRV in
3 healthy subjects declines with aging; (2) HRV of healthy subjects, particularly those >65 years
4 old, may decrease to below levels associated with increased risk of mortality; (3) gender
5 influences HRV (gender differences in HRV are age and measure dependent); and (4) age and
6 gender also affect HRV.
7 Tsuji et al. (1994) studied HRV in the original subjects of the Framingham Heart Study.
8 Subjects with transient or persistent nonsinus rhythm, 50% of recorded time, and those taking
9 antiarrhythmic medications were excluded. The associations between HRV measures and all-
10 cause mortality during 4 years of follow-up were assessed. A 1-SD decrement in low-frequency
11 power was associated with 1.70 times greater hazard for all-cause mortality (95% confidence
12 interval of 1.37 to 2.09). The authors concluded that estimation of HRV offers prognostic
13 information beyond that provided by the evaluation of traditional risk factors.
14
15
March 2001 6B-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Berntson, G. G.; Bigger, J. T., Jr.; Eckberg, D. L.; Grossman, P.; Kaufmann, P. G.; Malik, M.; Nagaraja, H. N.;
3 Forges, S. W.; Saul, J. P.; Stone, P. H.; Van Der Molen, M. W. (1997) Heart rate variability: origins,
4 methods, and interpretive caveats. Psychophysiology 34: 623-648.
5 Bigger, J. T., Jr.; Fleiss, J. L.; Steinman, R. C.; Rolnitzky, L. M.; Kleiger, R. E.; Rottman, J. N. (1992) Frequency
6 domain measures of heart period variability and mortality after myocardial infarction. Circulation
7 85: 164-171.
8 Burnett, R. T.; Dales, R.; Krewski, D.; Vincent, R.; Dann, T.; Brook, J. R. (1995) Associations between ambient
9 particulate sulfate and admissions to Ontario hospitals for cardiac and respiratory diseases. Am. J. Epidemiol.
10 142: 15-22.
11 Godleski, J. J.; Sioutas, C.; Katler, M.; Koutrakis, P. (1996) Death from inhalation of concentrated ambient air
12 particles in animal models of pulmonary disease. Am. J. Respir. Crit. Care Med. 153: A15.
13 Godleski et al. (1998) Increased cardiac vulnerability during exposure to inhaled environmental particles [abstract].
14 Presented at: Health Effects Institute annual meeting; May; Boston, MA. Boston, MA: Health Effects
15 Institute.
16 Gold, D. R.; Litonjua, A.; Schwartz, J.; Lovett, E.; Larson, A.; Nearing, B.; Allen, G.; Verrier, M.; Cherry, R.;
17 Verrier, R. (2000) Ambient pollution and heart rate variability. Circulation 101:1267-1273.
18 Hayano, J.; Sakakibara, Y.; Yamada, M.; Ohte, N.; Fujinami, T.; Yokoyama, K.; Watanabe, Y.; Takata, K. (1990)
19 Decreased magnitude of heart rate spectral components in coronary artery disease. Its relation to
20 angiographic severity. Circulation 81: 1217-1224.
21 Killingsworth, C. R.; Alessandrini, F.; Krishna Murthy, G. G.; Catalano, P. J.; Paulauskis, J. D.; Godleski, J. J.
22 (1997) Inflammation, chemokine expression, and death in monocrotaline-treated rats following fuel oil fly
23 ash inhalation. Inhalation Toxicol. 9: 541-565.
24 Kleiger, R. E.; Miller, J. P.; Bigger J. T. Jr.; Moss, A. J.; Multicenter Post-infarction Research Group. (1987)
25 Decreased heart rate variability and its association with increased mortality after acute myocardial infarction.
26 Am. J. Cardiol. 59: 256-262.
27 Liao, D.; Creason, J.; Shy, C.; Williams, R.; Watts, R.; Zweidinger, R. (1999) Daily variation of particulate air
28 pollution and poor cardiac autonomic control in the elderly. Environ. Health Perspect. 107: 521-525.
29 Malik, M. (1995) Graphical representation of circadian patterns of heart rate variability components. Pacing Clin.
30 Electrophysiol. 18: 1575-1580.
31 Martin, G. J.; Magid, N. M.; Myers, G.; Barnett, P. S.; Schaad, J. W.; Weiss, J. S.; Lesch, M.; Singer, D. H. (1987)
32 Heart rate variability and sudden death secondary to coronary artery disease during ambulatory
33 electrocardiographic monitoring. Am. J. Cardiol. 60: 86-89.
34 Morris, R. D.; Naumova, E. N.; Munasinghe, R. L. (1995) Ambient air pollution and hospitalization for congestive
35 heart failure among elderly people in seven large US cities. Am. J. Public Health 85: 1361-1365.
36 Pope, C. A., Ill; Verrier, R. L.; Lovett, E. G.; Larson, A. C.; Raizenne, M. E.; Kanner, R. E.; Schwartz, J.; Villegas,
37 G. M.; Gold, D. R.; Dockery, D. W. (1999) Heart rate variability associated with particulate air pollution.
38 Am. Heart J. 138: 890-899.
39 Reardon, M.; Malik, M. (1996) Changes in heart rate variability with age. Pacing Clin. Electrophysiol.
40 19: 1863-1866.
41 Schwartz, J. (1997) Air pollution and hospital admissions for cardiovascular disease in Tucson. Epidemiology
42 8:371-377.
43 Schwartz, J.; Morris, R. (1995) Air pollution and hospital admissions for cardiovascular disease in Detroit,
44 Michigan. Am. J. Epidemiol. 142: 23-35.
45 Singer, D. H.; Martin, G. J.; Magid, N.; Weiss, J. S.; Schaad, J. W.; Kehoe, R.; Zheutlin, T.; Fintel, D. J.;
46 Hsieh, A.-M.; Lesch, M. (1988) Low heart rate variability and sudden cardiac death. J. Electrocardiol.
47 21(suppl.):S46-S55.
48 Task Force of the European Society of Cardiology and the North American Society of Pacing and
49 Electrophysiology. (1996) Heart rate variability: standards of measurement, physiological interpretation and
50 clinical use. Circulation 93: 1043-1065.
51 Tsuji, H.; Venditti, F. J.; Manders, E. S.; Evans, J. C.; Larson, M. G.; Feldman, C. L.; Levy, D. (1994) Reduced
52 heart rate variability and mortality risk in an elderly cohort: The Framingham Heart Study. Circulation
53 90: 878-883.
March 2001 6B-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 Umetani, K.; Singer, D. H.; McCraty, R.; Atkinson, M. (1998) Twenty-four hour time domain heart rate variability
2 and heart rate: relations to age and gender over nine decades. J. Am. Coll. Cardiol. 31: 593-601.
3 Watkinson, W. P.; Campen, M. J.; Costa, D. L. (1998) Cardiac arrhythmia induction after exposure to residual oil
4 fly ash particles in a rodent model of pulmonary hypertension. Toxicol. Sci. 41: 209-216.
March 2001 6B-7 DRAFT-DO NOT QUOTE OR CITE
-------
-------
i 7. DOSIMETRY OF PARTICULATE MATTER
2
3
4 7.1 INTRODUCTION
5 A basic principle in health effects evaluation is that the dose delivered to the target site of
6 concern, rather than the external exposure, is the proximal cause of any biological response.
7 Characterization of the exposure-dose-response continuum for particulate matter (PM), a
8 fundamental objective of any dose-response assessment for evaluation of health effects, requires
9 the elucidation and understanding of the mechanistic determinants of inhaled particle dose,
10 which is dependent initially on the deposition of particles within the respiratory tract. Particle
11 deposition refers to the removal of particles from their airborne state because of their
12 aerodynamic or thermodynamic behavior in air. Once particles have deposited onto the surfaces
13 of the respiratory tract, they subsequently will be subjected to either absorptive or nonabsorptive
14 particulate removal processes. This may result in their removal from airway surfaces, as well as
15 their removal to various degrees from the respiratory tract. Particulate matter translocated from
16 initial deposition sites is said to have undergone clearance. Clearance of deposited particles
17 depends upon the initial site of deposition and upon the physicochemical properties of the
18 particles, both of which impact upon specific translocation mechanisms. Retained particle
19 burdens are determined by the dynamic relationship between deposition and clearance
20 mechanisms.
21 This chapter is concerned with particle dosimetry, the study of the deposition, translocation,
22 clearance and retention of particles within the respiratory tract and extrapulmonary tissues.
23 It summarizes basic concepts as presented in the 1996 EPA document, Air Quality Criteria for
24 Particulate Matter or "PM AQCD" (U.S. Environmental Protection Agency, 1996), specifically
25 in Chapter 10; and it updates the state of the science based upon new literature on particle
26 deposition, clearance and retention appearing since publication of the 1996 PM AQCD.
27 Although the basic mechanisms governing deposition and clearance of inhaled particles have not
28 changed, there has been significant additional information on the role of certain biological
29 determinants of the deposition/clearance process, such as gender and age. Also, the
March 2001 7-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 understanding of regional dosimetry and the particle size range over which this has been
2 evaluated has been expanded.
3 The dose from inhaled particles deposited and retained in the respiratory tract is governed
4 by a number of factors. These include exposure concentration and exposure duration, respiratory
5 tract anatomy and ventilatory parameters, and by physicochemical properties of the particles
6 themselves (e.g., particle size, hygroscopicity, solubility). The basic characteristics of particles
7 as they relate to deposition and retention, as well as anatomical and physiological factors
8 influencing particle deposition and retention, were discussed in depth in the 1996 PM AQCD.
9 Thus, in this current chapter, only an overview of basic information related to one critical factor
10 in deposition, namely particle size, is provided (Section 7.1.1), so as to allow the reader to
11 understand the different terms used in the remainder of this chapter and subsequent ones dealing
12 with health effects. This is followed, in Section 7.1.2, by a basic overview of respiratory tract
13 structure as it relates to deposition evaluation. The ensuing major sections of this chapter then
14 provide updated information on particle deposition, clearance, and retention in the respiratory
15 tract of humans, as well as laboratory animals, which are useful in the evaluation of PM health
16 effects. Issues related to the phenomenon of particle overload as it may apply to human exposure
17 and the use of instillation as an exposure technique to evaluate PM health effects also are
18 discussed. The final sections of the chapter deal with mathematical models of particle
19 disposition in the respiratory tract.
20 It must be emphasized that any dissection into discrete topics of factors that control dose
21 from inhaled particles tends to mask the dynamic and interdependent nature of the intact
22 respiratory system. For example, although deposition is discussed separately from clearance
23 mechanisms, retention (i.e., the actual amount of particles found in the respiratory tract at any
24 point in time) is determined by the relative rates of both deposition and clearance. Thus,
25 assessment of overall dosimetry requires integration of these various components of the overall
26 process.
27
28 7.1.1 Size Characterization of Inhaled Particles
29 Information about particle size distribution is important in the evaluation of effective
30 inhaled dose. This section summarizes particle attributes requiring characterization and provides
31 general definitions important in understanding particle fate within the respiratory tract.
March 2001 7-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 Particles exist in the atmosphere as aerosols, which are airborne suspensions of finely
2 dispersed solid or liquid particles. Because aerosols can consist of almost any material, their
3 description in simple geometric terms can be misleading unless important factors relating to
4 constituent particle size, shape, and density are considered. Although the size of particles within
5 aerosols can be described based on actual physical measurements (such as those obtained with a
6 microscope), in many cases it is better to use some equivalent diameter in place of the physical
7 diameter. The most commonly used metric is aerodynamic equivalent diameter (AED), whereby
8 particles of differing geometric size, shape and density are compared in terms of aerodynamic
9 behavior (i.e., terminal setting velocity) of particles that are unit density (1 gm/cm3) spheres. The
10 aerodynamic behavior of unit density spherical particles constitutes a useful standard by which
11 many types of particles can be compared in terms of certain deposition mechanisms.
12 It is important to note that aerosols present in natural and work environments have
13 polydisperse size distributions. This means that the constituent particles within an aerosol have a
14 range of sizes and are more appropriately described in terms of a size distribution parameter.
15 The lognormal distribution (i.e., the situation whereby the logarithms of particle diameter are
16 distributed normally) can be used for describing size distributions of most aerosols. In linear
17 form, the logarithmic mean is the median of the distribution, and the metric of variability around
18 this central tendency is the geometric standard deviation (og). The og, a dimensionless term, is
19 the ratio of the 84th (or 16th) percentile particle size to the 50th percentile size. Thus, the only
20 two parameters needed to describe a log normal distribution of aerosol particle sizes are the
21 median diameter and the geometric standard deviation. However, the actual size distribution
22 may be obtained in various ways. For example, when a distribution is described by counting
23 particles, the median is called the count median diameter (CMD). On the other hand, the median
24 of a distribution based on particle mass in an aerosol is the mass median diameter (MMD).
25 When using aerodynamic diameters, a term that is encountered frequently is mass median
26 aerodynamic diameter (MMAD), which refers to the median of the distribution of mass with
27 respect to aerodynamic equivalent diameter. Most of the present discussion will focus on
28 MMAD because it is the most commonly used measure of aerosol distribution. However,
29 alternative distributions should be used for particles with actual physical size below about
30 0.5 yum, because, for these, aerodynamic properties become less important. One such metric is
March 2001 7-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 thermodynamic-equivalent size, which is the diameter of a spherical particle that has the same
2 diffusion coefficient in air as the particle of interest.
3
4 7.1.2 Structure of the Respiratory Tract
5 Detailed discussion of respiratory tract structure was provided in the 1996 PM AQCD (U.S.
6 Environmental Protection Agency, 1996), and only a brief synopsis is presented here.
7 For dosimetry purposes, the respiratory tract can be divided into three regions: (1) extrathoracic
8 (ET), (2) tracheobronchial (TB), and (3) alveolar (A). The ET region consists of head airways
9 (i.e., nasal or oral passages) through the larynx and represents the areas through which inhaled air
10 first passes. In humans, inhalation can occur through the nose or mouth (or both, known as
11 oronasal breathing). However, most laboratory animals commonly used in respiratory
12 toxicological studies are obligate nose breathers.
13 From the ET region, inspired air enters the TB region at the trachea. From the level of the
14 trachea, the conducting airways then undergo branching for a number of generations. The
15 terminal bronchiole is the most peripheral of the distal conducting airways and these lead,
16 in humans, to the respiratory bronchioles, alveolar ducts, alveolar sacs and alveoli (all of which
17 comprise the A region). All of the conducting airways, except the trachea and portions of the
18 mainstem bronchi, are surrounded by parenchymal tissue. This is composed primarily of the
19 alveolated structures of the A region and associated blood and lymphatic vessels. It should be
20 noted that these respiratory tract regions are comprised of various cell types, and that there are
21 distinct differences in the cellular composition of the ET, TB, and A regions. Although a
22 discussion of cellular structure of the respiratory tract is beyond the scope of this section, details
23 may be found in a number of sources (e.g., Crystal et al., 1997).
24
25
26 7.2 PARTICLE DEPOSITION
27 This section discusses the deposition of particles in the respiratory tract. It begins with an
28 overview of the basic physical mechanisms that govern deposition. This is followed by an
29 update on both total respiratory tract and regional deposition patterns in humans. Some critical
March 2001 7-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 biological factors that may modulate deposition are then presented. The section ends with a
2 discussion of new information related to interspecies patterns of particle deposition.
3
4 7.2.1 Mechanisms of Deposition
5 Particles may deposit within the respiratory tract by five mechanisms: (1) inertial
6 impaction, (2) sedimentation, (3) diffusion, (4) electrostatic precipitation, and (5) interception.
7 Sudden changes in airstream direction and velocity cause particles to fail to follow the
8 streamlines of airflow. As a consequence, the particles contact, or impact, onto airway surfaces.
9 The ET and upper TB airways are characterized by high air velocities and sharp directional
10 changes and, thus, dominate as sites of inertial impaction. Impaction is a significant deposition
11 mechanism for particles larger than 1 /^m AED.
12 All aerosol particles are continuously influenced by gravity, but particles with an AED
13 >0.5 /urn are affected to the greatest extent. A particle will acquire a terminal settling velocity
14 when a balance is achieved between the acceleration of gravity acting on the particle and the
15 viscous resistance of the air, and it is this settling out of the airstream that takes it into contact
16 with airway surfaces. Both sedimentation and inertial impaction can influence the deposition of
17 particles within the same size range. These deposition processes act together in the ET and TB
18 regions, with inertial impaction dominating in the upper airways and gravitational settling
19 becoming increasingly dominant in the lower conducting airways, especially for the largest
20 particles, which can penetrate into the smaller bronchi.
21 Particles having actual physical diameters <1 //m are subjected increasingly to diffusive
22 deposition because of random bombardment by air molecules, which results in contact with
23 airway surfaces. The root mean square displacement that a particle experiences in a unit of time
24 along a given cartesian coordinate is a measure of its diffusivity. The density of a particle is
25 unimportant in determining particle diffusivity. Thus, instead of having an aerodynamic
26 equivalent size, diffusive particles of different shapes can be related to the diffusivity of a
27 thermodynamic equivalent size based on spherical particles.
28 The particle size region around 0.3 to 0.5 ^m frequently is described as consisting of
29 particles that are small enough to be minimally influenced by impaction or sedimentation and
30 large enough to be minimally influenced by diffusion. Such particles are the most persistent in
31 inhaled air and undergo the lowest extent of deposition in the respiratory tract.
March 2001 7-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 Interception is deposition by physical contact with airway surfaces. The interception
2 potential of any particle depends on its physical size, and fibers are the chief concern in relation
3 to the interception process. Their aerodynamic size is determined predominantly by their
4 diameter, rather than their length.
5 Electrostatic precipitation is deposition related to particle charge. The minimum charge an
6 aerosol particle can have is zero when it is electrically neutral. This condition rarely is achieved
7 because of the random charging of aerosol particles by air ions. Aerosol particles will acquire
8 charges from these ions by collisions with them because of their random thermal motion.
9 Furthermore, many laboratory generated aerosols are charged. Such aerosols will lose their
10 charge slowly as they attract oppositely charged ions. An equilibrium state of these competing
11 processes eventually is achieved. This Boltzmann equilibrium represents the charge distribution
12 of an aerosol in charge equilibrium with bipolar ions. The minimum amount of charge is very
13 small, with a statistical probability that some particles within the aerosol will have no charge, and
14 others will have one or more charges.
15 The electrical charge on some particles may result in an enhanced deposition over what
16 would be expected from size alone. This results from image charges induced on the surface of
17 the airway by these particles or to space-charge effects, whereby repulsion of particles containing
18 like charges results in increased migration toward the airway wall. The effect of charge on
19 deposition is inversely proportional to particle size and airflow rate. This type of deposition is
20 probably small compared to the effects of turbulence and other deposition mechanisms, and
21 generally has been considered to be a minor contributor to overall particle deposition. However,
22 a recent study (Cohen et al., 1998) employing hollow airway casts of the human tracheobronchial
23 tree assessed deposition of ultrafine (0.02 jum) and fine (0.125 /um) particles; the deposition of
24 singly charged particles was found to be 5 to 6 times that of particles having no charge and 2 to
25 3 times that of particles at Boltzmann equilibrium. This suggests that electrostatic precipitation
26 may, in fact, be a significant deposition mechanism for ultrafine, and some fine, particles within
27 the TB region.
28
29 7.2.2 Deposition Patterns in the Human Respiratory Tract
30 Knowledge of sites where particles of different sizes deposit in the respiratory tract and the
31 amount of deposition is necessary for understanding and interpreting the health effects associated
March 2001 7-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 with exposure to particles. Particles deposited in the various respiratory tract regions are
2 subjected to large differences in clearance mechanisms and pathways and, consequently,
3 retention times. This section summarizes concepts of particle deposition in humans and
4 laboratory animals as reported in U.S. Environmental Protection Agency (1996), and provides
5 additional information based on studies published since the release of that earlier document.
6 The ambient air often contains particles that are too massive to be inhaled. The descriptor
7 "inhalability" is used to denote the overall spectrum of particle sizes that potentially are capable
8 of entering the respiratory tract. Inhalability is defined as the ratio of the number concentration
9 of particles of a certain aerodynamic diameter that are inspired through the nose or mouth to the
10 number concentration of the same diameter particle present in an inspired volume of ambient air
11 (International Commission on Radiological Protection, 1994). In general, for humans, unit
12 density particles >100-yam diameter have a low probability of entering the mouth or nose in still
13 air. However, there is no sharp cutoff to zero probability. Furthermore, there is no lower limit to
14 inhalability as long as the particle exceeds a critical size where the aggregation of atomic or
15 molecular units is stable enough to endow it with "particulate" properties, in contrast to those of
16 free ions or gas molecules.
17
18 7.2.2.1 Total Respiratory Tract Deposition
19 Total human respiratory tract deposition, as a function of particle size, is depicted in
20 Figure 7-1. These data were obtained by various investigators using different sizes of spherical
21 test particles in healthy male adults under different ventilation conditions; the large standard
22 deviations reflect interindividual and breathing pattern-related variability of deposition
23 efficiencies. Deposition with nose breathing is generally higher than that with mouth breathing
24 because of the superior filtration capabilities of the nasal passages. For particles with
25 aerodynamic diameters greater than 1 //m, deposition is governed by impaction and
26 sedimentation, and it increases with increasing AED. When AED is >10 //m, almost all inhaled
27 particles are deposited. As the particle size decreases from ~0.5 /j.m, diffusional deposition
28 becomes dominant and total deposition depends more on the actual physical diameter of the
29 particle, with decreasing particle diameter leading to an increase in total deposition. Total
30 deposition shows a minimum for particle diameters in the range of 0.3 to 0.5 /^m where, as noted
31 above, neither sedimentation, impaction or diffusion deposition are very effective.
March 2001 7-7 DRAFT-DO NOT QUOTE OR CITE
-------
c
o
o
100
80
60
40
20
.2 o
I
O Human (Oral)
• Human (Nasal)
1
I
0.01
0.1 1.0
Particle Diameter (urn)
10
Figure 7-1. Total deposition data (percentage deposition of amount inhaled) in humans
as a function of particle size. All values are means with standard deviations
when available. Particle diameters are aerodynamic (MMAD) for those
Source: Modified from Schlesinger (1988).
1 Besides particle size, breathing pattern is the most important factor affecting lung
2 deposition. Recently, Kim (2000) reported total lung deposition values in healthy adults for a
3 wide range of breathing patterns; tidal volume 375 to 1500 mL, flow rate 150 to 1000 mL/s, and
4 respiratory time 2 to 12s. Total lung deposition increased with increasing tidal volume at a
5 given flow rate and increased with increasing flow rate at a given respiratory time. Various
6 deposition values were correlated with a single composite parameter consisting of particle size,
7 flow rate, and tidal volume.
March 2001
DRAFT-DO NOT QUOTE OR CITE
-------
1 One of the specific size modes of the ambient aerosol that is being evaluated in terms of
2 potential toxicity is the ultrafme mode (i.e., particles having diameters <0.1 yum [CMD]). There
3 is little information on total respiratory tract deposition of such particles. Frampton et al. (2000)
4 exposed healthy adult females to 26.7-nm diameter carbon particles (at 10 /^g/m3) for 2 h. The
5 inspired and expired particle number concentration and size distributions were evaluated. Total
6 respiratory tract deposition fraction was determined for six particle size fractions, ranging from
7 7.5 to 133.4 nm. They found an overall total lung deposition fraction of 0.66 (by particle
8 number) or 0.58 (by particle mass), indicating that exhaled mean particle diameter was slightly
9' larger than inhaled diameter. The deposition fraction decreased with increasing particle size
10 within the ultrafme range, from 0.76 at the smallest size to 0. 47 at the largest. Jaques and Kim
11 (2000) found the greatest deposition fraction for smaller particles and for breathing patterns with
12 longer residence times (i.e., low flow and higher tidal volume) consistent with deposition by
13 diffusion.
14 A property of some ambient particulate species that affects deposition is hygroscopicity, the
15 propensity of a material for taking up and retaining moisture under certain conditions of humidity
16 and temperature. Such particles can increase in size in the humid air within the respiratory tract
17 and, when inhaled, will deposit according to their hydrated size rather than their initial size. The
18 implications of hygroscopic growth on deposition has been reviewed extensively by Morrow
19 (1986) and Hiller (1991), whereas the complications of studying lung deposition of hygroscopic
20 aerosols have been reviewed recently by Kim (2000). In general, compared to nonhygroscopic
21 particles of the same initial size, the deposition of hygroscopic aerosols in different regions of the
22 lung may be higher or lower, depending on the initial size. Thus, for particles with initial sizes
23 larger than =0.5 yum, the influence of hygroscopicity is to increase total deposition, whereas for
24 smaller ones total deposition is decreased.
25
26 7.2.2.2 Deposition in the Extrathoracic Region
27 The fraction of inhaled particles depositing in the ET region is quite variable, depending on
28 particle size, flow rate, breathing frequency and whether breathing is through the nose or the
29 mouth. Mouth breathing bypasses much of the filtration capabilities of the nasal airways, leading
30 to increased deposition in the lungs (TB and A regions). The ET region is clearly the site of first
31 contact with particles in the inhaled air, and essentially acts as a "prefilter" for the lungs.
March 2001 7-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 Since release of the 1996 PM AQCD, a number of studies have explored ET deposition
2 with in vivo studies, as well as in both physical and mathematical model systems. In one study,
3 the relative distribution of particle deposition between the oral and nasal passages was assessed
4 during "inhalation" by use of a physical model (silicone rubber) of the human upper respiratory
5 system, extending from the nostrils and mouth through the main bronchi (Lennon et al., 1998).
6 Monodisperse particles ranging in size from 0.3 to 2.5 /um were used at various flow rates
7 ranging from 15 to 50 L/min. Total deposition was assessed, as was regional deposition in the
8 oral passages, lower oropharynx-trachea, nasal passages, and nasopharynx-trachea. Deposition
9 within the nasal passages was found to agree with available data obtained from a human
10 inhalation study (Heyder and Rudolf, 1977), being proportional to particle size, density, and
11 inspiratory flow rate. It also was found that for oral inhalation, the relative distribution between
12 the oral cavity and the oropharynx-trachea was similar, whereas for nasal inhalation, the nasal
13 passages contained most of the particles deposited in the model, with only about 10% depositing
14 in the nasopharynx-trachea section. Furthermore, the deposition efficiency of the
15 nasopharynx-trachea region was greater than that of the oropharynx-trachea region.
16 For simulated oronasal breathing, deposition in the ET region depended primarily on particle size
17 rather than flow rate. For all flows and for all breathing modes, total deposition in the ET region
18 increased as particle diameter increased. Such information on deposition patterns in the ET
19 region is useful in refining empirical deposition models.
20 Deposition within the nasal passages was further evaluated by Kesavanathan and Swift
21 (1998), who examined the deposition of 1- to 10-yum particles in the nasal passages of normal
22 adults under an inhalation regime in which the particles were drawn through the nose and out
23 through the mouth at flow rates ranging from 15 to 35 L/min. At any particle size, deposition
24 increased with increasing flow rate; whereas, at any flow rate, deposition increased with
25 increasing particle size. In addition, as was shown experimentally by Lennon et al. (1998) under
26 oronasal breathing conditions, deposition of 0.3- to 2.5-yum particles within the nasal passages
27 was significantly greater than within the oral passages, and nasal inhalation resulted in greater
28 total deposition in the model than did oral inhalation. These results are consistent with other
29 studies discussed in the 1996 PM AQCD and with the known dominance of impaction deposition
30 within the ET region.
March 2001 7-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 For ultrafme particles (d < 0.1 /mi), deposition in the ET region is controlled by diffusion,
2 which depends only on the particle's geometric diameter. Prior to 1996, ET deposition for this
3 particle size range had not been studied extensively in humans, and this remains the case. In the
4 earlier document, the only data available for ET deposition of ultrafme particles were from cast
5 studies. More recently, deposition in the ET region was examined using mathematical modeling.
6 Three dimensional numerical simulations of flow and particle diffusion in the human upper
7 respiratory tract, which included the nasal region, oral region, larynx, and first two generations of
8 bronchi, were performed by Yu et al. (1998). Deposition of particles ranging from 0.001 to
9 0.1 /mi in these different regions was calculated under inspiratory and expiratory flow conditions.
10 Deposition efficiencies in the total model were lower on expiration than inspiration, although the
11 former were quite high. Nasal deposition of ultrafme particles can also be quite high. For
12 example, nasal deposition accounted for up to 54% of total deposition in the model system for
13 0.001-/on particles. The total deposition efficiency in the model was 75% (of the amount
14 entering), for this size particle. With oral breathing, deposition efficiency was estimated at 48%
15 (of amount entering) (Yu et al., 1998).
16 Swift and Strong (1996) examined the deposition of ultrafine particles, ranging in size from
17 0.053 to 0.062 /mi, in the nasal passages of normal adults during constant inspiratory flows of
18 6 to 22 L/min. The results are consistent with results noted in studies above, namely that the
19 nasal passages are highly efficient collectors for ultrafine particles. In this case, fractional
20 deposition ranged from 94 to 99% (of amount inhaled). There was found to be only a weak
21 dependence of deposition on flow rate, which contrasts with results noted above (i.e., Lennon et
22 al., 1998) for particles >0.3 //m, but is consistent with diffusion as the deposition mechanism.
23 Cheng et al. (1997) examined oral airway deposition in a replicate cast of the human nasal
24 cavity, oral cavity, and laryngeal-tracheal sections. Particle sizes ranged from 0.005 to 0.150 /mi,
25 and constant inspiratory and expiratory flow rates of 7.5 to 30 L/min were used. They noted that
26 the deposition fractions within the oral cavity were essentially the same as that in the
27 laryngeal-tracheal sections for all particle sizes and flowrates. They ascribed this to the balance
28 between flow turbulence and residence time in these two regions. Svartengren et al. (1995)
29 examined the effect of changes in external resistance on oropharyngeal deposition of 3.6-/mi
30 particles in asthmatics. Under control mouthpiece breathing conditions (flow rate 0.5 L/s), the
31 median deposition as a percentage of inhaled particles in the mouth and throat was 20%
March 2001 7-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 (mean = 33%; range 12 to 84%). Although the mean deposition fell to 22% with added
2 resistance, the median value remained at 20% (range 13 to 47%). Fiberoptic examination of the
3 larynx revealed that there was a trend for increased mouth and throat deposition associated with
4 laryngeal narrowing. Katz et al. (1999) indicate, on the basis of mathematical model
5 calculations, that turbulence plays a key role in enhancing particle deposition in the larynx and
6 trachea.
7 The results of all of the above studies support the previously known ability of the ET
8 region, and especially the nasal passages, to act as an efficient filter for inhaled particles. Even
9 ultrafine particles have significant deposition within the ET region, and this region, therefore,
10 serves as an important filter for such particles as well as for larger ones, potentially reducing the
11 amount of particles within a wide range that are available for deposition in the TB and A regions.
12
13 7.2.2.3 Deposition in the Tracheobronchial and Alveolar Regions
14 Particles that do not deposit in the ET region enter the lung, but their regional deposition in
15 the lung cannot be precisely measured. Much of the available regional deposition data have been
16 obtained from experiments with radioactive labeled poorly soluble particles. These have been
17 described previously (U.S. Environmental Protection Agency, 1996). Although there are no new
18 regional data obtained by means of the radioactive aerosol method since the publication of that
19 document, a novel serial bolus delivery method has been introduced. Using this bolus technique,
20 regional deposition has been measured for fine and coarse aerosols (Kim et al., 1996) and for
21 ultrafine aerosols (Kim and Jacques, 2000). The serial bolus method uses nonradioactive
22 aerosols and can measure regional deposition in virtually an unlimited number of lung
23 compartments. The Kim and Jaques studies cited above measured regional deposition in
24 10 serial compartments of the lung, and obtained tracheobronchial and alveolar deposition for
25 particles ranging from 0.04 to 5.0 //m in diameter. TB and alveolar deposition also have been
26 measured in men and women using this method (Kim and Hu, 1998).
27
28 7.2.2.4 Local Distribution of Deposition
29 Airway structure and its associated air flow patterns are exceedingly complex and
30 ventilation distribution of air in different parts of the lung is uneven. Thus, it is expected that
31 particle deposition patterns within the ET, TB, and A regions would be highly nonuniform, with
March 2001 7-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 some sites exhibiting deposition that is much greater than average levels within these regions.
2 This was discussed in detail previously in the 1996 PM AQCD. Basically, using deposition data
3 from living subjects as well as from mathematical and physical models, enhanced deposition has
4 been shown to occur in the nasal passages and trachea and at branching points in the TB and
5 A regions. Recently, Churg and Vedal (1996) examined retention of particles on carinal ridges
6 and tubular sections of airways from lungs obtained at necropsy. Results indicated significant
7 enhancement of particle retention on carinal ridges through the segmental bronchi; the ratios
8 were similar in all airway generations examined.
9 Deposition "hot spots" at airway bifurcations have undergone additional analyses using
10 mathematical modeling techniques. Using calculated deposition sites, a number of studies
11 showed a strong correlation between secondary flow patterns and deposition sites and density for
12 large (10 ^m) particles, as well as for ultrafine particles (0.01 /urn) (Heistracher and Hofmann,
13 1997; Hofmann et al., 1996). This supports experimental work, noted in U.S. Environmental
14 Protection Agency (1996), indicating that, like larger particles, ultrafine particles also show
15 enhanced deposition at airway branch points, even in the upper tracheobronchial tree.
16 The pattern of particle distribution on a more regional scale was evaluated by Kim et al.
17 (1996). Deposition patterns were measured in situ in healthy nonsmoking young adult males,
18 using an aerosol bolus technique that delivered 1-, 3-, or 5-//m particles into specific volumetric
19 depths within the lungs. The distribution of particle deposition was uneven, and it was noted that
20 sites of peak deposition shifted from distal to proximal regions of the lungs with increasing
21 particle size. Furthermore, the surface dose was found to be greater in the conducting airways
22 than in the alveolar region for all of the particle sizes evaluated. Within the conducting airways,
23 the largest airway regions (i.e., 50 to 100 mL volume) received the greatest surface doses.
24 Kim and Fisher (1999) studied local deposition efficiencies and deposition patterns of
25 aerosol particles (2.9 to 6.7 jwm) in sequential double bifurcation tube models with two different
26 branching geometries: one with in-plane (A) and another with out of plane (B) bifurcation. The
27 deposition efficiencies (DE) in each bifurcation increased with increasing Stokes number (Stk).
28 With symmetric flow conditions, DE was somewhat smaller in the second than the first
29 bifurcation in both models. DE was greater in the second bifurcation in model B than in model
30 A. With asymmetric flows, DE was greater in the low-flow side compared to the high-flow side
31 and this was consistent in both models. Deposition pattern analysis showed highly localized
March 2001 7-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 deposition on and in the immediate vicinity of each bifurcation ridge, regardless of branching
2 pattern and flow pattern.
3 Comer et al. (2000) used the same three-dimensional computer simulation technique to
4 measure local deposition patterns in sequentially bifurcating, airway models. The simulation was
5 for 3-, 5-, and 7-^m particles and assumed steady, laminar, constant property air flow with
6 symmetry about the first bifurcation. The overall trend of the particle deposition efficiency (i.e.,
7 an exponential increase with Stokes number) was similar for all bifurcations. Local deposition
8 patterns consistently showed that the majority of the deposition occurred in the carinal region.
9 Kim and Jaques (2000) used the respiratory bolus technique to measure the respiratory dose
10 of fine particles (0.04, 0.06, 0.08, and 0.1 /^m) in young adults. Under normal breathing
11 conditions (tidal volume 500 mL, respiratory flow rate 250 mL/s), bolus aerosols were delivered
12 sequentially to a lung depth ranging from 50 to 500 mL in 50-mL increments. The results
13 indicate that regional deposition varies widely along the depth of the lung regardless of the
14 particle sizes used. Peak deposition occurred in the lung regions situated between 150 and
15 200 mL from the mouth and sites of peak deposition shifted proximally with a decrease in
16 particle size. Deposition dose per unit surface area was greatest in the proximal lung regions and
17 decreased rapidly with increased lung depth. Peak surface dose was 5 to 7 times greater than the
18 average lung dose. These results indicate that local enhancement of dose occurs in healthy lungs,
19 and dose enhancement could be an important factor in eliciting pathophysiological effects.
20
21 7.2.2.5 Deposition of Specific Size Modes of Ambient Aerosol
22 The studies described in previous sections generally evaluated deposition using individual
23 particle sizes within certain ranges, without consideration of specific relevant ambient size
24 ranges. Some recent studies, however, have considered the deposition profiles of particle modes
25 that exist in ambient air, so as to provide information on dosimetry of these "real world" particle
26 size fractions. One such study (Venkataraman and Kao, 1999) examined the contribution of two
27 specific size modes of the PM,0 ambient aerosol, namely the fine mode (defined as particles with
28 diameters up to 2.5 ,um) and the coarse mode (defined as particles with diameters 2.5 to 10 /^m),
29 to total lung and regional lung doses (i.e., a daily dose expressed as jUg/day, and a surface dose
30 expressed a yug/cm2/day) resulting from a 24-h exposure to a particle concentration of 150 /ug/m3.
31 The study also evaluated deposition in terms of two metrics, namely mass dose and number dose.
March 2001 7-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 Deposition was calculated using a mathematical model for a healthy human lung under both
2 moderate exertion and vigorous exertion. Regional deposition values were obtained for the
3 nasopharyngeal region (NP), the tracheobronchial tree (TB), and the pulmonary airways (A).
4 The daily mass dose from exposure to PM10 for three breathing cycles resulted in 36% of
5 the inhaled coarse particle mass deposited in the lung and 30% in the NP, 4% in TB, and 2% in
6 A. About 9% of the fine particle mass was deposited in the lungs, 1.5% in NP and TB and 6% in
7 A. The daily mass dose peaked in the A airways (generation 20) for all breathing patterns,
8 whereas that for the coarse fractions was comparable in the TB and A regions. The mass per unit
9 surface area of various airways from the fine and coarse fractions was larger in the trachea and
10 first few generations of bronchi (gen 3 to 5). It was suggested that these large surface doses may
11 be related to aggravation of upper respiratory tract illness in geographical areas where coarse
12 particles were present.
13 The daily number dose from exposure to PM,0 resulted in 18% of the inhaled coarse
14 particles being deposited in the lungs, 13% in the NP, 2% in the TB, and 3% in A. About 11% of
15 inhaled fine particle number was deposited in the lungs, 0.06% in NP, 2% in TB, and 9% in A.
16 Daily number dose was different for fine and coarse fractions in all lung airways, with the dose
17 from the fine fraction higher by about 100 times in the NP and about 105 times in internal lung
18 airways. The surface number dose (particles/cm2/day) was 103 to 105 times higher for fine than
19 for coarse particles in all lung airways, indicating the larger number of fine particles depositing.
20 Particle number doses did not follow trends in mass doses and are much higher for fine than
21 coarse particles and are higher for different breathing patterns. It also was concluded that the fine
22 fraction contributes 10,000 times greater particle number per alveolar macrophage than the
23 coarse fraction particles. These results must be viewed with caution because they were obtained
24 using a pure mathematical model that must be validated.
25 Another evaluation of deposition that included consideration of size mode of the ambient
26 aerosol was that of Broday and Georgopoulos (2000). In this case, a mathematical model was
27 used to account for particle hygroscopic growth, transport, and deposition in tracking the changes
28 in the size distribution of inhaled aerosols. It was concluded that different rates of particle
29 growth in the inspired air resulted in a change in the size distribution of the aerosol, such that
30 increased mass and number fractions of inspired fine particles are found in the size range
31 between 0.1 to 1 yum and, therefore, deposit to a lesser extent due to a decrease in diffusion
March 2001 7-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 deposition. On the other hand, particles that were originally in the 0.1 - to 1 -//m size range when
2 inhaled will undergo enhanced deposition because of their increase in size resulting from
3 hygroscopic growth. Thus, the speciation of the inhaled polydisperse aerosol and its initial size
4 distribution affect the evolution of size distribution once inhaled and, thus, its deposition profile
5 in the respiratory tract. Hygroscopicity of respirable particles must be considered for accurate
6 predictions of deposition. Because different fractions likely have different chemical
7 composition, such changes in deposition patterns will affect dosimetry and biological responses.
8
9 7.2.3 Biological Factors Modulating Deposition
10 Experimental deposition data in humans are commonly derived using healthy adult
11 Caucasian males. Various factors can act to alter deposition patterns from those obtained in this
12 group. Evaluation of these factors is important to help understand potentially susceptible
13 subpopulations, because differences in biological response following pollutant exposure may be
14 caused by dosimetry differences as well as by differences in innate sensitivity. The effects of
15 different biological factors on deposition were discussed in U.S. Environmental Protection
16 Agency (1996) and are summarized below, with additional information obtained from more
17 recent studies.
18
19 7.2.3.1 Gender
20 Males and females differ in body size and ventilatory parameters; so, it is expected that
21 there would be gender differences in deposition. Using particles in the 2.5- to 7.5-,um size range
22 Pritchard et al. (1986) indicated that, for comparable particle sizes and inspiratory flow rates,
23 females had higher ET and TB deposition and smaller A deposition than did males. The ratio of
24 A deposition to total thoracic deposition in females also was found to be smaller. These
25 differences were attributed to gender differences in airway size.
26 In a recent study (Bennett et al., 1996), the total respiratory tract deposition of 2-/um
27 particles was examined in adult males and females aged 18 to 80 years who breathed with a
28 normal resting pattern. Deposition was assessed in terms of a deposition fraction, which was the
29 difference between the amount of particles inhaled and exhaled during oral breathing. Although
30 there was a tendency for a greater deposition fraction in females compared to males, because
March 2001 7-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 males had greater minute ventilation, the deposition rate (i.e., deposition per unit time) was
2 greater in males than in females.
3 Kim and Hu (1998) assessed regional deposition patterns in healthy adult males and
4 females using aerosols with median aerodynamic sizes of 1, 3, and 5 /j.m and a bolus delivery
5 technique, which involved controlled breathing. The total deposition in the lungs was similar for
6 both genders with the smaller particle, but was greater in women for the 3- and 5-/um particles,
7 regardless of the inhalation flow rate used; this difference ranged from 9 to 31 %, with higher
8 values associated with higher flow rates. The pattern of deposition was similar for both genders,
9 although females showed enhanced deposition peaks for all three particle sizes. The volumetric
10 depth location of these peaks was found to shift from peripheral (increased volumetric depth) to
11 proximal (shallow volumetric depth) regions of the lung with increasing particle size, but the
12 shift was greater in females than in males. Thus, deposition appeared to be more localized in the
13 lungs of females compared to those of males. These differences were attributed to a smaller size
14 of the upper airways in females than in males (particularly of the laryngeal structure). Local
15 deposition of l-/u.m particles was somewhat flow dependent, but for larger (5-^m) particles was
16 largely independent of flow (flows did not include those that would be typical of exercise).
17 In a related study, Kim et al. (2000) evaluated differences in deposition between males and
18 females related to exercise levels of ventilation and breathing patterns. Using particles at the
19 same size noted above and a number of breathing conditions, total lung deposition was
20 comparable between men and women for l-/um particles but was greater in women than men for
21 3- and 5-^m particles with all breathing patterns. The gender difference was about 15% at rest,
22 and variable during exercise, depending on particle size. However, total lung deposition rate
23 (deposition per unit time) was found to be 3 to 4 times greater during moderate exercise than
24 during rest for all particle sizes. Thus, it was concluded that exercise may increase the health risk
25 from particles because of increased deposition, and that women may be more susceptible to this
26 exercise-induced change.
27 Jaques and Kim (2000) and Kim and Jaques (2000) expanded the evaluation of deposition
28 in males and females to particles <1 /^m. They measured total lung deposition in healthy adults
29 using sizes in the ultrafine mode (0.04 to 0.1 /wm), in addition to those having diameters of 1 and
30 5 ,um. Total lung deposition was greater in females than in males for 0.04- and 0.06-^m
31 particles. The difference was negligible for 0.08- and 0.1 -/um particles. Therefore, the gender
March 2001 7-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 effect was particle-size dependent, showing a greater deposition in females for very small
2 ultrafine and large coarse particles, but not for fine particles ranging from 0.08 to 1 //m. A local
3 deposition fraction was determined in each volumetric compartment of the lung to which
4 particles are injected based on the inhalation procedure (Kim and Jaques, 2000). The deposition
5 fraction was found to increase with increasing lung depth from the mouth, reach a peak value and
6 then decrease with further increase in lung volumetric depth. The height of the peak and its
7 depth did vary with particle size and breathing pattern. Peak deposition for the 5-/um particles
8 was more proximal than that for the \-jj.m particles, whereas that for the ultrafine particles
9 occurred between these two peaks. For the ultrafine particles, the peak deposition became more
10 proximal as particle size decreased. Although this pattern of deposition distribution was similar
11 for both men and women, the region of peak deposition was shifted closer to the mouth and peak
12 height was slightly greater for women than for men for all exposure conditions.
13
14 7.2.3.2 Age
15 Airway structure and respiratory conditions vary with age, and these variations may alter
16 the deposition pattern of inhaled particles. The limited experimental studies reported in the
17 earlier PM AQCD (U. S. Environmental Protection Agency, 1996) indicated results ranging from
18 no clear dependence of total deposition on age to slightly higher deposition in children than
19 adults. Potential regional deposition differences between children and adults were assessed to a
20 greater extent using mathematical models. These indicated that if the entire respiratory tract and
21 a complete breathing cycle at normal rate are considered, that ET deposition in children generally
22 would be higher than that in adults, but that TB and A regional deposition in children may be
23 either higher or lower than the adult, depending on particle size (Xu and Yu, 1986). Enhanced
24 deposition in the TB region would occur for particles <5 ,um in children (Xu and Yu, 1986;
25 Hofmann et al, 1989a).
26 An age dependent theoretical model to predict regional particle deposition in childrens'
27 lungs, and that incorporates breathing parameters and morphology of the growing lung, was
28 developed by Musante and Martonen (1999). The model was used to compare deposition, at rest,
29 of monodisperse aerosols, ranging from 0.25 to 5 /urn, in the lungs of children (aged 7, 22, 48,
30 and 98 mo) to that in adults (aged 30 years). Compared to adults, A deposition was highest in the
31 48- and 98-mo subjects for all particle sizes, TB deposition was found to be a monotonically
March 2001 7-18 DRAFT-DO NOT QUOTE OR CITE
-------
1 decreasing function of age for all sizes; and total lung deposition (i.e., TB+A) was generally
2 higher in children than adults, with children of all ages showing similar total deposition fractions.
3 This model was used by Musante and Martonen (2000a) to evaluate the deposition of a
4 polydisperse aerosol that has been extensively used in toxicological studies, namely residual oil
5 fly ash (ROFA) having an MMAD of 1.95 ^m. Deposition was evaluated under resting
6 breathing conditions. The mass based deposition fraction of the particles was found to decrease
7 with age from 7 mo to adulthood, but the mass deposition per unit surface area in the lungs of
8 children could be significantly greater than that in the adult.
9 Cheng et al. (1995) examined deposition of ultrafme particles in replica casts of the nasal
10 airways of children aged 1.5 to 4 years. Particle sizes ranged from 0.0046 to 0.2 /^m, and both
11 inspiratory and expiratory flowrates were used (3 to 16 L/min). Deposition efficiency was found
12 to decrease with increasing age for a given particle size and flowrate.
13 Oldham et al. (1997) examined the deposition of monodisperse particles, having diameters
14 of 1, 5, 10, and 15 ^m, in hollow airway models that were designed to represent the trachea and
15 the first few bronchial airway generations of an adult, a 7-year-old child, and a 4-year-old child.
16 They noted that in most cases, the total deposition efficiency was greater in the child-size models
17 than in the adult model.
18 Bennett et al. (1997a) analyzed the regional deposition of 4.5 /urn, poorly soluble particles
19 in children and in adults with mild cystic fibrosis (CF), but who likely had normal upper airway
20 anatomy, such that intra- and extrathoracic deposition would be similar to that in healthy adults.
21 The mean age of the children was 13.8 years and adults were 29.1 years. ET deposition, as a
22 percentage of total respiratory tract deposition, was higher by about 50% in children compared to
23 CF and healthy adults (30.7%, 20.1%, and 16.0%, respectively). There was an age dependence
24 of ET deposition in the children, in that the percentage ET deposition tended to be higher at a
25 younger age; the younger group (<14 years) had almost twice the percentage ET deposition of the
26 older group (>14 years). Additional analyses showed an inverse correlation of extrathoracic
27 deposition with body height. There was no significant difference in lung or total respiratory tract
28 deposition between the children and adults. Because ET deposition was age dependent and total
29 deposition was not, this suggests that the ET region does a more effective job in children of
30 filtering out the particles that would otherwise reach the TB region. However, because the lungs
March 2001 7-19 DRAFT-DO NOT QUOTE OR CITE
-------
1 of children are smaller than those of adults, children may still have comparable deposition per
2 unit surface area as would adults.
3 Bennett and Zeman (1998) measured the deposition of monodisperse 2 //m (MMAD)
4 particles in children aged 7 to 14 years and adolescents aged 14 to 18 years for comparison to
5 that in adults (19 to 35 years). Each subject inhaled the particles by following their previously
6 determined individual spontaneous resting breathing pattern. Deposition was assessed by
7 measuring the amount of particles inhaled and exhaled. There was no age-related difference in
8 deposition within the children group. There was also no significant difference in deposition
9 between the children and adolescents, between the children and adults, or between the
10 adolescents and adults. However, the investigators noted that, because the children had smaller
11 lungs and higher minute volumes relative to lung size, they likely would receive greater doses of
12 particles per lung surface area compared to adults. Furthermore, deposition in children did vary
13 with tidal volume, increasing with increasing volume to a greater extent than was seen in adults.
14 These additional studies still do not provide unequivocal evidence for significant differences in
15 deposition between adults and children, even when considering differences in lung surface area.
16 However, it should be noted that differences in levels of activity between adults and children are
17 likely to play a fairly large role in age-related differences in deposition patterns of ambient
18 particles. Children generally have higher activity levels during the day, and higher associated
19 minute ventilation per lung size, which can contribute to a greater size-specific dose of particles.
20 Activity levels in relationship to exposure are discussed more fully in Chapter 5.
21 Another subpopulation of potential concern related to susceptibility to inhaled particles is
22 the elderly. In the study of Bennett et al. (1996), in which the total respiratory tract deposition of
23 2-^m particles was examined in people aged 18 to 80 years, the deposition fraction in the lungs
24 of people with normal lung function was found to be independent of age, depending solely on
25 breathing pattern and airway resistance.
26
27 7.2.3.3 Respiratory Tract Disease
28 The presence of respiratory tract disease can affect airway structure and ventilatory
29 parameters, thus altering deposition compared to that in healthy individuals. The effect of airway
30 diseases on deposition has been studied extensively, as described in the earlier PM AQCD (U.S.
31 Environmental Protection Agency, 1996). Studies described therein had shown that people with
March 2001 7-20 DRAFT-DO NOT QUOTE OR CITE
-------
1 chronic obstructive pulmonary disease (COPD) had very heterogeneous deposition patterns, with
2 differences in regional deposition compared to normals. People with asthma and obstructive
3 pulmonary disease tended to have greater TB deposition than did healthy people. Furthermore,
4 there tended to be an inverse relationship between bronchconstriction and the extent of
5 deposition in the A region, whereas total respiratory tract deposition generally increased with
6 increasing level of airway obstruction. The described studies were performed during controlled
7 breathing, where all subjects breathed with the same tidal volume and respiratory rate. However,
8 although resting tidal volume is similar or elevated in people with COPD compared to normals,
9 the former tend to breathe at a faster rate, resulting in higher than normal tidal peak flow and
10 resting minute ventilation. Thus, some of the reported differences in the deposition of particles
11 could have been caused by increased fractional deposition with each breath. Although the extent
12 to which lung deposition may change with respect to particle size, breathing pattern, and disease
13 status in people with COPD is still unclear, some recent studies have attempted to provide
14 additional insight into this issue.
15 Bennett et al. (1997b) measured the fractional deposition of insoluble 2-^m particles in
16 people with severe to moderate COPD (mix of emphysema and chronic bronchitis, mean age
17 62 years) and compared this to healthy older adults (mean age 67 years) under conditions where
18 the subjects breathed using their individual resting breathing pattern, as well as a controlled
19 breathing pattern. People with COPD tended to breathe with elevated tidal volume and at a
20 faster rate than people with healthy lungs, resulting in about 50% higher resting minute
21 ventilation. Total respiratory tract deposition was assessed in terms of deposition fraction, a
22 measure of the amount deposited based on measures of aerosol inhaled and amount exhaled, and
23 deposition rate, the particles deposited per unit time. Under typical breathing conditions, people
24 with COPD had about 50% greater deposition fraction than did age-matched healthy adults.
25 Because of the elevation in minute ventilation, people with COPD had average deposition rates
26 about 2.5 times that of healthy adults. Similar to previously reviewed studies (U.S.
27 Environmental Protection Agency, 1996), these investigators observed an increase in deposition
28 with an increase in airway resistance, suggesting that, at rest, COPD resulted in increased
29 deposition of fine particles in proportion to the severity of airway disease. The investigators also
30 reported a decrease in deposition with increasing mean effective airspace diameter; this
31 suggested that the enhanced deposition was associated more with the chronic bronchitic
March 2001 7-21 DRAFT-DO NOT QUOTE OR CITE
-------
1 component of COPD than with the emphysematous component of the disease. Greater
2 deposition was noted with natural breathing compared to the fixed pattern.
3 Kim and Kang (1997) measured lung deposition of l-yum particles inhaled via the mouth by
4 healthy adults (mean age 27 years) and by those with various degrees of airway obstruction,
5 namely smokers (mean age 27 years), smokers with small airway disease (SAD; mean age
6 37 years), asthmatics (mean age 48 years), and patients with COPD (mean age 61 years)
7 breathing under the same controlled pattern. Deposition fraction was obtained by measuring the
8 number of particles inhaled and exhaled, breath by breath. There was a marked increase in
9 deposition in people with COPD. Deposition was 16%, 49%, 59%, and 103% greater in
10 smokers, smokers with SAD, asthmatics and people with COPD, respectively, than healthy
11 adults. Deposition in COPD patients was significantly greater than that associated with either
12 SAD or asthma; there was no significant difference in deposition between people with SAD and
13 asthma. Deposition fraction was found to be correlated with percent predicted forced expiratory
14 volume (FEV,) and forced expiratory flow (FEF25-75%). Airway resistance was not correlated
15 strongly with total lung deposition. Kohlhaufl et al. (1999) also showed increased deposition of
16 fine particles (0.9 //m) in women with bronchial hyperresponsiveness.
17 Segal et al. (2000a) developed a mathematical model for airflow and particle motion in the
18 lungs that was used to evaluate how lung cancer affects deposition patterns in the lungs of
19 children. It was noted that the presence of airway tumors could affect deposition, by increasing
20 probability of inertial deposition and diffusion. The former would occur on the upstream
21 surfaces of tumors, whereas the latter would occur on downstream surfaces. It was concluded
22 that particle deposition is affected by the presence of airway disease, but that effects may be
23 systematic and could be predicted and incorporated into dosimetry models.
24 Thus, the database related to particle deposition and lung disease suggests that total lung
25 deposition generally is increased with obstructed airways, regardless of deposition distribution
26 between the TB and A regions. Airflow distribution is very uneven in COPD because of the
27 irregular pattern of obstruction, and there can be closure of small airways. In this situation, a part
28 of the lung is inaccessible, and particles can penetrate deeper into other better ventilated regions.
29 Thus, deposition can be enhanced locally in regions of active ventilation, particularly in the
30 A region. The relationships between lung deposition and airway obstruction or ventilation
31 distribution were previously studied in vivo in animal models (Kim, 1989; Kim et al., 1989).
March 2001 7-22 DRAFT-DO NOT QUOTE OR CITE
-------
1 7.2.3.4 Anatomical Variability
2 As indicated above, variations in anatomical parameters between genders and between
3 healthy people and those with obstructive lung disease can affect deposition patterns. However,
4 previous analyses generally have overlooked the effect on deposition of normal interindividual
5 variability in airway structure in healthy individuals. This is an important consideration in
6 dosimetry modeling, which often is based on a single idealized structure. Studies available since
7 1996 have attempted to assess the influence of such variation in respiratory tract structure on
8 deposition patterns.
9 The ET region is the first to contact inhaled particles and, therefore, deposition within this
10 region would reduce the amount of particles available for deposition in the lungs. Variations in
11 relative deposition within the ET region will, therefore, propagate through the rest of the
12 respiratory tract, creating differences in calculated doses from individual to individual.
13 A number of studies have examined the influence of variations in airway geometry on deposition
14 in the ET region.
15 Cheng et al. (1996) examined nasal airway deposition in healthy adults using particles
16 ranging in size from 0.004 to 0.15 //m at two constant inspiratory flow rates, 167 and 33 mL/s.
17 Deposition was evaluated in relation to measures of nasal geometry as determined by magnetic
18 resonance imaging and acoustic rhinometry. They noted that interindividual variability in
19 deposition was correlated with the wide variation of nasal dimensions, in that greater surface
20 area, smaller cross-sectional area and increasing complexity of airway shape were all associated
21 with enhanced deposition.
22 Using a regression analysis of data on nasal airway deposition derived from Cheng et al.
23 (1996), Guilmette et al. (1997) noted that the deposition efficiency within this region was highly
24 correlated with both nasal airway surface area and volume; this indicated that airway size and
25 shape factors were important in explaining intraindividual variability noted in experimental
26 studies of human nasal airway aerosol deposition. Thus, much of the variability in measured
27 deposition among people resulted from differences in the size and shape of airway regions.
28 Kesavanathan and Swift (1998) also evaluated the influence of geometry in affecting
29 deposition in the nasal passages of normal adults from two ethnic groups. Mathematical
30 modeling of the results indicated that the shape of the nostril affected particle deposition in the
31 nasal passages, but that there still remained large intersubject variations in deposition when this
March 2001 7-23 DRAFT-DO NOT QUOTE OR CITE
-------
1 was accounted for, and that likely was caused by geometric variability in the mid and posterior
2 regions of the nasal passages.
3 Bennett et al. (1998) studied the role of anatomic dead space (ADS) in particle deposition
4 and retention in bronchial airways using an aerosol bolus technique. They found that the
5 fractional deposition was dependant on the subject's ADS, and that a significant number of
6 particles were retained beyond 24 h. This finding of prolonged retention of insoluble particles in
7 the airways is consistent with the findings of Scheuch et al. (1995) and Stahlhofen et al. (1986a).
8 Bennett et al. (1999) also found a lung volume-dependent asymmetric distribution of particles
9 between the left and right lung; the left:right ratio was increased at increased percentage of total
10 lung capacity (e.g., at 70% TLC, L:R was 1.60).
11 From the analysis of detailed deposition patterns measured by a serial bolus delivery
12 method, Kim and Hu (1998) and Kim and Jaques (2000) found a marked enhancement in
13 deposition in the very shallow region of the lungs in females. The enhanced local deposition for
14 both ultrafine and coarse particles was attributed to a smaller size of the upper airways,
15 particularly of the laryngeal structure.
16 Hofmann et al. (2000) examined the role of heterogeneity of airway structure in the rat
17 acinar region in affecting deposition patterns within this area of the lungs. By the use of different
18 morphometric models, they showed that substantial variability in predicted particle deposition
19 would result.
20
21 7.2.4 Interspecies Patterns of Deposition
22 The primary purpose of this document is to assess the health effects of particles in humans.
23 As such, human dosimetry studies have been stressed. Such studies avoid uncertainties
24 associated with extrapolation of dosimetry from laboratory animals to humans. Nevertheless,
25 animal models have been and are currently being used in evaluations of health effects from
26 particulate matter, because there are ethical limits to the types of studies that can be performed on
27 human subjects. Because of this, there is considerable need to understand dosimetry in animals,
28 and to understand dosimetric differences between animals and humans, hi this regard, there has
29 been a number of new studies that were designed to assess particle dosimetry in commonly used
30 animals and to relate this to dosimetry in humans.
March 2001 7-24 DRAFT-DO NOT QUOTE OR CITE
-------
1 The various species used in inhalation toxicology studies that serve as the basis for
2 dose-response assessment may not receive identical doses in a comparable respiratory tract
3 region (i.e., ET, TB, or A) when exposed to the same aerosol at the same inhaled concentration.
4 Such interspecies differences are important, because any adverse toxic effect is often related to
5 the quantitative pattern of deposition within the respiratory tract as well as to the exposure
6 concentration; this pattern determines not only the initial respiratory tract tissue dose, but also the
7 specific pathways by which deposited material is cleared and redistributed (Schlesinger, 1985).
8 Differences in patterns of deposition between humans and animals were summarized previously
9 in the earlier PM AQCD (U.S. Environmental Protection Agency, 1996; Schlesinger et al., 1997).
10 Such differences in initial deposition must be considered when relating biological responses
11 obtained in laboratory animal studies to effects in humans.
12 One of the issues that must be considered in interspecies comparisons of hazards from
13 inhaled particles is inhalability of the aerosol in the atmosphere of concern. Although this may
14 not be an issue for humans per se as far as exposure to ambient particles are concerned, it can be
15 an important issue when attempting to relate results of studies using animal species employed in
16 inhalation toxicological studies (Miller et al., 1995). For example, differences in inhalability
17 between rat and human become very pronounced for particles >5 /um, and some differences are
18 also evident for particles as small as 1 /urn.
19 Several recent studies have addressed various aspects of interspecies differences in
20 deposition using mathematical modeling approaches. Hofmann et al. (1996) compared
21 deposition between rat and human lungs using three-dimensional asymmetric bifurcation models
22 and mathematical procedures for obtaining air flow and particle trajectories. Deposition in
23 segmental bronchi and terminal bronchioles was evaluated under both inspiration and expiration,
24 at particle sizes of 0.01, 1, and 10 //m (which covered the range of deposition mechanisms from
25 diffusion to impaction). Total deposition efficiencies of all particles in the upper and lower
26 airway bifurcations were comparable in magnitude for both rat and human. However, the
27 investigators noted that penetration probabilities from preceding airways must be considered.
28 When considering the higher penetration probability in the human lung, the resulting bronchial
29 deposition fractions were generally higher in human than rat. For all particle sizes, deposition at
30 rat bronchial bifurcations was less enhanced on the carinas compared to that found in human
31 airways.
March 2001 7-25 DRAFT-DO NOT QUOTE OR CITE
-------
1 Hoftnann et al. (1996) attempted to account for interspecies differences in branching
2 patterns in deposition analyses. Numerical simulations of three-dimensional particle deposition
3 patterns within selected (species-specific) bronchial bifurcations indicated that morphologic
4 asymmetry was a major determinant of the heterogeneity of local deposition patterns. They noted
5 that many interspecies deposition calculations used morphometry that was described by
6 deterministic lung models (i.e., the number of airways in each airway generation adopts a
7 constant value, and all airways in a given generation have identical lengths and diameters). Such
8 models cannot account for variability and branching asymmetry of airways in the lungs. Thus,
9 their study employed computations that used stochastic morphometric models of human and rat
10 lungs (Koblinger and Hofmann, 1985, 1988; Hofmann et al., 1989b) and evaluated regional and
11 local particle deposition. Stochastic models of lung structure describe, in mathematical terms,
12 the inherent asymmetry and variability of the airway system, including diameter, length and
13 angle. They are based on statistical analyses of actual morphometric analyses of lungs. The
14 model also incorporated breathing patterns for humans and rats. The dependence of deposition
15 on particle size was found to be similar in both rats and humans, with deposition minima in the
16 size range of 0.1 to 1 /um for both total deposition and deposition within the TB region. This was
17 not found to occur in the A region, where a deposition maximum occurred at about 0.02 to
18 0.03 /urn in both species followed by a decline, and then another maximum between 3 and 5 /urn.
19 The deposition decrease in the A region at the smallest and largest sizes resulted from the
20 filtering efficiency of upstream airways. Although deposition patterns were qualitatively similar
21 in rat and human, total respiratory tract and TB deposition in the human lung appeared to be
22 consistently higher than in the rat. Alveolar region deposition fraction in humans was lower than
23 in the rat over the size range of 0.001 to 10 /^m. Furthermore, both species showed a similar
24 pattern of dependence of deposition on flow rate.
25 The above model also assessed local deposition. In both human and rat, deposition of
26 0.001- and 10-^m particles was highest in the upper bronchial airways, whereas 0.1- and l-/um
27 particles showed higher deposition in more peripheral airways, namely the bronchiolar airways
28 in rat and the respiratory bronchioles in humans. Deposition was variable within any branching
29 generation because of differences in airway dimensions, and regional and total deposition also
30 exhibited intrasubject variations. Airway geometric differences between rats and humans were
31 reflected in deposition. Because of the greater branching asymmetry in rats, prior to about
March 2001 7-26 DRAFT-DO NOT QUOTE OR CITE
-------
1 generation 12, each generation showed deposition maxima at two particle sizes, reflecting
2 deposition in major and minor daughters. These geometric differences became reduced with
3 depth into the lung; beyond generation 12, these two maxima were no longer seen. A later
4 analysis (Hofmann and Bergmann, 1998), using a stochastic morphometric model of human and
5 rat lungs to compare regional and local particle deposition in the human and rat lungs over a wide
6 range of particle sizes (1 to 10 //m) and flow rates, noted that, although there were quantitative
7 differences in the deposition patterns within the lungs of these two species, the dependence of
8 deposition on particle size and flow rate was qualitatively similar. This indicates that the
9 dependence of deposition on physical factors is similar for all species.
10 Another comparison of deposition in lungs of humans and rats was performed by Musante
11 and Martonen (2000b). An interspecies mathematical dosimetry model was used to determine
12 the deposition of residual oil fly ash (ROFA) in the lungs under sedentary and light activity
13 breathing patterns. This latter was mimicked in the rat by increasing the CO2 level in the
14 exposure system. The MMAD of the aerosol was 1.95 jum. They noted that physiologically
15 comparable respiratory intensity levels did not necessarily correspond to comparable dose
16 distribution in the lungs. Because of this, the resting rat may not be a good model for the resting
17 human. The ratio of aerosol mass deposited in the TB region to that in the A region for the
18 human at rest was 0.961, indicating fairly uniform deposition throughout the lungs. On the other
19 hand, in the resting rat, the ratio was 2.24, indicating greater deposition in the TB region than in
20 the A region. However, by mimicking light activity in the rat, the ratio was reduced to 0.97,
21 similar to the human. This suggests that ventilatory characteristics in animal models may have to
22 be adjusted to provide for comparable regional deposition to that in humans.
23 The relative distribution of particles deposited in the bronchial and alveolar region airways
24 may differ in the lungs of animals and humans, for the same total amount of deposited matter,
25 because of structural differences. The effect of such structural difference between rat and human
26 airways on particle deposition patterns was examined by Hofmann et al. (1999) in an attempt to
27 find the most appropriate morphometric parameter to characterize local particle deposition for
28 extrapolation modeling purposes. Particle deposition patterns were evaluated as functions of
29 three morphometric parameters, namely (1) airway generation, (2) airway diameter, and
30 (3) cumulative path length. It was noted that airway diameter was a more appropriate
March 2001 7-27 DRAFT-DO NOT QUOTE OR CITE
-------
1 morphometric parameter for comparison of particle deposition patterns in human and rat lungs
2 than was airway generation.
3 The influence of exposure concentration on the pattern of particle retention in rats (exposed
4 to diesel soot) and humans (exposed to coal dust) was examined by Nikula et al. (2000) using
5 histological lung sections obtained from both species. The exposure concentrations for diesel
6 soot were 0.35, 3.5, or 7.0 mg/m3, and exposure duration was 7 h/day, 5 days/week for 24 mo.
7 The human lung sections were obtained from nonsmoking nonminers, nonsmoking coal miners
8 exposed to levels <2 mg dust/m3 for 3 to 20 years, or nonsmoking miners exposed to <10 mg/m3
9 for 33 to 50 years. In both species, the volume density of deposition increased with increasing
10 dose (which is related to exposure duration and concentration). In rats, the diesel exhaust
11 particles were found to be primarily in the lumens of the alveolar duct and alveoli, whereas, in
12 humans, retained dust was found primarily in the interstitial tissue. Thus, different lung cells
13 contact retained particles in the two species and may result in different biological responses with
14 chronic dust exposure.
15 The manner in which particle dose is expressed, that is, the specific dose metric, may
16 impact on relative differences in deposition between humans and other animal species.
17 For example, although deposition when expressed on a mass per unit alveolar surface area basis
18 may not be different between rats and humans, dose metrics based on particle number per various
19 anatomical parameters (e.g., per alveolus or alveolar macrophage) can differ between rats and
20 humans, especially for particles around 0.1 to 0.3 yum (Miller et al., 1995). Furthermore, in
21 humans with lung disease such as asthma or COPD, differences between rat and human can be
22 even more pronounced.
23 The probability of any biological effect in humans or animals depends on deposition and
24 retention of particles, as well as the underlying dose-response relationship. Interspecies
25 dosimetric extrapolation must consider differences in deposition, clearance, and dose response.
26 Thus, even similar deposition patterns may not result in similar effects in different species
27 because dose also is affected by clearance mechanisms and species sensitivity. In addition, the
28 total number of particles deposited in the lung may not be the most relevant dose metric to
29 compare species. For example, it may be the number of deposited particles per unit surface area
30 that determines response. More specifically, even if deposition is similar in rat and human, there
31 would be a higher deposition density in the rat because of the smaller surface area of rat lung.
March 2001 7-28 DRAFT-DO NOT QUOTE OR CITE
-------
1 Thus, species-specific differences in deposition density should be considered when health effects
2 observed in laboratory animals are being evaluated in terms of the human situation.
3
4
5 7.3 PARTICLE CLEARANCE AND TRANSLOCATION
6 This section discusses the clearance and translocation of particles that have deposited in the
7 respiratory tract. A basic overview of biological mechanisms and pathways of clearance in the
8 various region of the respiratory tract is presented first. This is followed by an update on
9 regional kinetics of particle clearance. Interspecies patterns of clearance are then addressed,
10 followed by new information on biological factors that may modulate clearance.
11
12 7.3.1 Mechanisms and Pathways of Clearance
13 Particles that deposit on airway surfaces may be cleared from the respiratory tract
14 completely, or may be translocated to other sites within this system, by various regionally distinct
15 processes. These clearance mechanisms, which are outlined in Table 7-1, can be categorized as
16 either absorptive (i.e., dissolution) or nonabsorptive (i.e., transport of intact particles) and may
17 occur simultaneously or with temporal variations. It should be mentioned that particle solubility
18 in terms of clearance refers to solubility within the respiratory tract fluids and cells. Thus, a
19 poorly soluble particle is considered to be one whose rate of clearance by dissolution is
20 insignificant compared to its rate of clearance as an intact particle. For the most part, all
21 deposited particles are subject to clearance by the same mechanisms, with their ultimate fate a
22 function of deposition site, physicochemical properties (including solubility and any toxicity),
23 and sometimes deposited mass or number concentration. Clearance routes from the various
24 regions of the respiratory tract have been discussed previously in detail (U.S. Environmental
25 Protection Agency, 1996; Schlesinger et al., 1997). They are schematically shown in Figure 7-2
26 (for extrathoracic and tracheobronchial regions) and in Figure 7-3 (for poorly soluble particle
27 clearance from the alveolar region) and are reviewed only briefly below.
28
29
March 2001 7-29 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 7-1. OVERVIEW OF RESPIRATORY TRACT PARTICLE CLEARANCE
AND TRANSLOCATION MECHANISMS
Extrathoracic region (ET)
Mucociliary transport
Sneezing
Nose wiping and blowing
Dissolution and absorption into blood
Tracheobronchial region (TB)
Mucociliary transport
Endocytosis by macrophages/epithelial cells
Coughing
Dissolution and absorption into blood/lymph
Alveolar region (A)
Macrophages, epithelial cells
Interstitial
Dissolution and absorption into blood/lymph
Source: Schlesinger (1995).
( Nasal Passages
Blood
)
Dissolution
Posterior
Extrinsic Clearance
Mucociliary
Transport
Pharynx
Tracheobronchial Tree
Figure 7-2. Major clearance pathways for particles deposited in the extrathoracic region
and tracheobronchial tree.
Source: Adapted from Schlesinger et al. (1997).
March 2001
7-30
DRAFT-DO NOT QUOTE OR CITE
-------
L'CjyL'tS^f It^U f df ll\fl^
T
Phagocytosis by ^j
Alveolar Macrophages I
1 \
T p
Movement within _^ .
Alveolar Lumen
1
Bronchiolar / Bronchial ^
Lurncn ^
*
Mucociliary Blanket
i
Gl Tract
Endocyt
^ lype i A
Epithelic
T
assage Through
Iveolar Epithelium ^"
i
Interstitium ^
i
i
mphatic Channels *^~
v
1 ^ f mi^h NI/^/Hoc*
osis by
Iveolar
il Pnllr- Rlnnrl ^
11 L/cllo DIUUU -^
A
Passage through
Pulmonary Capillary
Endothelium
A A
i
] Phagocytosis by \
^ Interstitial
n \^ Macrophages J
Figure 7-3. Diagram of known and suspected clearance pathways for poorly soluble
particles depositing in the alveolar region.
Source: Modified from Schlesinger et al. (1997).
1 7.3.1.1 Extra thoracic Region
2 The clearance of poorly soluble particles deposited in the posterior portions of the nasal
3 passages occurs via mucociliary transport, with the general flow of mucus towards the
4 nasopharynx. Mucus flow in the most anterior portion of the nasal passages is forward, clearing
5 deposited particles to the vestibular region where removal is by sneezing, wiping, or blowing.
6 Soluble material deposited on the nasal epithelium is accessible to underlying cells via
7 diffusion through the mucus. Dissolved substances may be translocated subsequently into the
8 bloodstream. The nasal passages have a rich vasculature, and uptake into the blood from this
9 region may occur rapidly.
10
March 2001
7-31
DRAFT-DO NOT QUOTE OR CITE
-------
1 Clearance of poorly soluble particles deposited in the oral passages is by coughing and
2 expectoration or by swallowing into the gastrointestinal tract. Soluble particles are likely to be
3 rapidly absorbed after deposition.
4
5 7.3.1.2 Tracheobronchial Region
6 Poorly soluble particles deposited within the TB region are cleared by mucociliary
7 transport towards the oropharynx, followed by swallowing. Poorly soluble particles also may
8 traverse the epithelium by endocytotic processes, entering the peribronchial region, be engulfed
9 via phagocytosis by airway macrophages, which can then move cephalad on the mucociliary
10 blanket, or enter the airway lumen from the bronchial or bronchiolar mucosa. Soluble particles
11 may be absorbed through the epithelium into the blood. There is, however, evidence that even
12 some soluble particles may be cleared by mucociliary transport (Bennett and Ilowite, 1989;
13 Matsuietal., 1998).
14
15 7.3.1.3 Alveolar Region
16 Clearance from the A region occurs via a number of mechanisms and pathways. Particle
17 removal by macrophages comprises the main nonabsorptive clearance process in this region.
18 These cells, which reside on the epithelium, phagocytize and transport deposited material that
19 they contact by random motion or via directed migration under the influence of chemotactic
20 factors.
21 Although alveolar macrophages normally comprise up to about 5% of the total alveolar
22 cells in healthy, nonsmoking humans and other mammals, the actual cell count may be altered by
23 particle loading. The magnitude of any increase in cell number is related to the number of
24 deposited particles rather than to total deposition by weight. Thus, equivalent masses of an
25 identically deposited substance would not produce the same response if particle sizes differed,
26 and the deposition of smaller particles would tend to result in a greater elevation in macrophage
27 number than would deposition of larger particles.
28 Particle-laden macrophages may be cleared from the A region along a number of pathways.
29 As noted in Figure 7-3, this includes cephalad transport via the mucociliary system after the cells
30 reach the distal terminus of the mucus blanket; movement within the interstitium to a lymphatic
31 channel; or perhaps traversing of the alveolar-capillary endothelium, directly entering the
March 2001 7-32 DRAFT-DO NOT QUOTE OR CITE
-------
1 bloodstream. Particles within the lymphatic system may be translocated to tracheobronchial
2 lymph nodes, which can become reservoirs of retained material. Particles subsequently reaching
3 the postnodal lymphatic circulation will enter the blood. Once in the systemic circulation, these
4 particles, or transmigrated macrophages, can travel to extrapulmonary organs. Deposited
5 particles that are not ingested by alveolar macrophages may enter the interstitium, where they are
6 subject to phagocytosis by resident interstitial macrophages, and may travel to perivenous,
7 peribronchiolar or subpleural sites, where they become trapped, increasing particle burden. The
8 migration and grouping of particles and macrophages within the lungs can lead to the
9 redistribution of initially diffuse deposits into focal aggregates. Some particles or components
10 can bind to epithelial cell membranes or macromolecules, or other cell components, delaying
11 clearance from the lungs.
12 Churg and Brauer (1997) examined lung autopsy tissue from 10 never-smokers from
13 Vancouver, Canada. They noted that the geometric mean particle diameter (GMPD) in lung
14 parenchymal tissue was 0.38 ^m (og = 2.4). Ultrafmes were less than 5% of the total retained
15 particulate matter. Metal particles had a GMPD of 0.17 //m and silicates 0.49 f^m. Ninety-six
16 percent of retained PM was less than 2.5 //m.
17 Clearance by the absorptive mechanism involves dissolution in the alveolar surface fluid,
18 followed by transport through the epithelium and into the interstitium, and diffusion into the
19 lymph or blood. Although factors affecting the dissolution of deposited particles are poorly
20 understood, solubility is influenced by the particle's surface to volume ratio and other properties,
21 such as hydrophilicity and lipophilicity (Mercer, 1967; Morrow, 1973; Patten, 1996). Thus, as
22 noted, materials generally considered to be relatively insoluble still may have high dissolution
23 rates and short dissolution half-times if the particle size is small.
24 Some deposited particles may undergo dissolution in the acidic milieu of the
25 phagolysosomes after ingestion by macrophages. Intracellular dissolution may be the initial step
26 in translocation from the lungs for these particles and for material associated with these particles
27 (Kreyling, 1992; Lundborg et al., 1985). Following dissolution, the material can be absorbed
28 into the blood. Dissolved materials may then leave the lungs at rates that are more rapid than
29 would be expected based on an "expected" normal dissolution rate in lung fluid.
30
31
March 2001 7-33 DRAFT-DO NOT QUOTE OR CITE
-------
1 7.3.2 Clearance Kinetics
2 The kinetics of clearance has been reviewed in U.S. Environmental Protection Agency
3 (1996) and in a number of monographs (e.g., Schlesinger et al., 1997) and is discussed only
4 briefly here. The actual time frame over which clearance occurs affects the cumulative dose
5 delivered to the respiratory tract, as well as that delivered to extrapulmonary organs.
6
7 7.3.2.1 Extrathoracic Region
8 Mucus flow rates in the posterior nasal passages are highly nonuniform, but the median rate
9 in a healthy adult human is about 5 mm/min, resulting in a mean anterior to posterior transport
10 time of about 10 to 20 min for poorly soluble particles (Rutland and Cole, 1981; Stanley et al.,
11 1985). Particles deposited in the anterior portion of the nasal passages are cleared more slowly
12 by mucus transport, and are usually more effectively removed by sneezing, wiping, or nose
13 blowing (Fry and Black, 1973; Morrow, 1977).
14
15 7.3.2.2 Tracheobronchial Region
16 Mucus transport in the tracheobronchial tree occurs at different rates in different local
17 regions; the velocity of movement is fastest in the trachea, and it becomes progressively slower
18 in more distal airways. In healthy nonsmoking humans, using noninvasive procedures and no
19 anesthesia, average tracheal mucus transport rates have been measured at 4.3 to 5.7 mm/min
20 (Yeates et al., 1975, 1981; Foster et al., 1980; Leikauf et al., 1981, 1984), whereas that in the
21 main bronchi has been measured at -2.4 mm/min (Foster et al., 1980). Estimates for human
22 medium bronchi range between 0.2 to 1.3 mm/min, whereas those in the most distal ciliated
23 airways range down to 0.001 mm/min (Morrow et al., 1967; Cuddihy and Yeh, 1988; Yeates and
24 Aspin, 1978).
25 The total duration of bronchial clearance, or some other time parameter, often is used as an
26 index of mucociliary kinetics. Although clearance from the TB region is generally rapid, there is
27 experimental evidence, discussed in U.S. Environmental Protection Agency (1996), that a
28 fraction of material deposited in the TB region is retained much longer than the 24 h commonly
29 used as the outer range of clearance time for particles within this region (Stahlhofen et al.,
30 1986a,b; Scheuch and Stahlhofen, 1988; Smaldone et al., 1988). Some recent studies described
March 2001 7-34 DRAFT-DO NOT QUOTE OR CITE
-------
1 below continue to support the concept that TB regional clearance consists of both a fast and a
2 slow component.
3 Falk et al. (1997) studied clearance in healthy adults using monodisperse Teflon particles
4 (6.2 ^m) inhaled at two flow rates. A considerable fraction (about 50%) of particles deposited in
5 small airways had not cleared within 24 h following exposure. These particles cleared with a
6 half time of 50 days. Although the deposition sites of the particles were not confirmed
7 experimentally, calculations suggested these to be in the smaller ciliated airways. Camner et al.
8 (1997) also noted that clearance from the TB region was incomplete by 24 h postexposure, and
9 suggested that this may be caused by incomplete clearance from bronchioles. Healthy adults
10 inhaled teflon particles (6, 8, and 10 //m) under low flow rates to maximize deposition in the
11 small ciliated airways. The investigators noted a decrease in 24-h retention with increasing
12 particle size, indicating a shift with increasing size toward either a smaller retained fraction,
13 deposition more proximally in the respiratory tract, or both. They calculated that a large fraction,
14 perhaps as high as 75%, of particles depositing in generations 12 through 16 was still retained at
15 24 h postexposure.
16 In a study to examine retention kinetics in the tracheobronchial tree (Falk et al., 1999),
17 normal nonsmoking adults inhaled radioactively tagged 6.1 -/^m particles at both a normal flow
18 rate and slow flow rate designed to deposit particles preferentially in the small ciliated airways.
19 Lung retention was measured from 24 h to 6 mo after exposure. Following the normal
20 inhalation, 14% of the particles retained at 24 h cleared with a half time of 3.7 days, and 86%
21 with a half time of 217 days. Following the slow inhalation, 35% of the particles retained at 24 h
22 cleared with a half time of 3.6 days, and 65% with a half time of 170 days. Deposition calculated
23 using a number of mathematical models indicated higher deposition in the bronchiolar region
24 (generations 9 through 15) with the slow rate inhalation compared to the normal rate. The
25 experimental data and predictions of the deposition modeling indicated that 40% of the particles
26 deposited in the conducting airways during the slow inhalation were retained after 24 h. The
27 particles that cleared with the shorter half time were mainly deposited in the bronchiolar region,
28 but only about 25% of the particles deposited in this region cleared in this phase. This study
29 provided additional confirmation for a phase of slow clearance from the bronchial tree.
30 The underlying sites and mechanisms of long-term TB retention in the smaller airways are
31 not known. Some proposals were presented in the earlier PM AQCD (U.S. Environmental
March 2001 7-35 DRAFT-DO NOT QUOTE OR CITE
-------
1 Protection Agency, 1996). This slow clearing tracheobronchial compartment likely is associated
2 with bronchioles <1 mm in diameter (Lay et al., 1995; Kreyling et al., 1999; Falk et al., 1999).
3 Based on a study in which an adrenergic agonist was used to stimulate mucus flow, so as to
4 examine the role of mucociliary transport in the bronchioles, it was found that clearance from the
5 smaller airways was not influenced by the drug, suggesting to the investigators that mucociliary
6 transport was not as an effective clearance mechanism from this region as in larger airways
7 (Svartengren et al., 1998, 1999). Although slower or less effective mucus transport may result in
8 longer retention times in these small airways, other factors may account for long-term TB
9 retention. One such proposal is the movement of particles into the gel phase because of surface
10 tension forces in the liquid lining of the small airways (Gehr et al., 1990, 1991). The issue of
11 particle retention in the tracheobronchial tree certainly is not resolved.
12 Long-term TB retention patterns are not uniform. There is an enhancement at bifurcation
13 regions (Radford and Martell, 1977; Henshaw and Fews, 1984; Cohen et al., 1988), the likely
14 result of both greater deposition and less effective mucus clearance within these areas. Thus,
15 doses calculated based on uniform surface retention density may be misleading, especially if the
16 material is, toxicologically, slow acting.
17
18 7.3.2.3 Alveolar Region
19 Particles deposited in the A region generally are retained longer than those deposited in
20 airways cleared by mucociliary transport. There are limited data on alveolar clearance rates in
21 humans. Within any species, reported clearance rates vary widely because, in part, of different
22 properties of the particles used in the various studies. Furthermore, some chronic experimental
23 studies have employed high concentrations of poorly soluble particles, which may have interfered
24 with normal clearance mechanisms, resulting in clearance rates different from those that would
25 typically occur at lower exposure levels. Prolonged exposure to high particle concentrations is
26 associated with what is termed particle "overload". This is discussed later in greater detail in
27 Section 7.4.
28 There are numerous pathways of A region clearance, and the utilization of these may
29 depend on the nature of the particles being cleared. Little is known concerning relative rates
30 along specific pathways. Thus, generalizations about clearance kinetics are difficult to make.
31 Nevertheless, A region clearance is usually described as a multiphasic process, each phase
March 2001 7-36 DRAFT-DO NOT QUOTE OR CITE
-------
1 considered to represent removal by a different mechanism or pathway, and often characterized by
2 increased retention half times following exposure.
3 The initial uptake of deposited particles by alveolar macrophages is very rapid and
4 generally occurs within 24 h of deposition (Lehnert and Morrow, 1985; Naumann and
5 Schlesinger, 1986; Lay et al., 1998). The time for clearance of particle-laden alveolar
6 macrophages via the mucociliary system depends on the site of uptake relative to the distal
7 terminus of the mucus blanket at the bronchiolar level. Furthermore, clearance pathways, and
8 subsequent kinetics, may depend to some extent on particle size. For example, some smaller
9 ultrafme particles (perhaps <0.02 /urn) may be less effectively phagocytosed than larger ones
10 (Oberdorster, 1993).
11 Uningested particles may penetrate into the interstitium within a few hours following
12 deposition. This transepithelial passage seems to increase as particle loading increases,
13 especially to that level above which macrophage numbers increase (Ferin, 1977; Perm et al.,
14 1992; Adamson and Bowden, 1981). It also maybe particle size dependent, because insoluble
15 ultrafme particles (<0.1 //m diameter) of low intrinsic toxicity show increased access to the
16 interstitum and greater lymphatic uptake than do larger particles of the same material
17 (Oberdorster et al., 1992; Ferin et al., 1992). However, ultrafme particles of different materials
18 may not enter the interstitium to the same extent. Similarly, a depression of phagocytic activity,
19 a reduction in macrophage ability to migrate to sites of deposition (Madl et al., 1998), or the
20 deposition of large numbers of ultrafme particles may increase the number of free particles in the
21 alveoli, perhaps enhancing removal by other routes. In any case, free particles may reach the
22 lymph nodes, perhaps within a few days after deposition (Lehnert et al., 1988; Harmsen et al.,
23 1985), although this route is not certain and may be species dependent.
24 The extent of lymphatic uptake of particles may depend on the effectiveness of other
25 clearance pathways, in that lymphatic translocation probably increases when phagocytic activity
26 of alveolar macrophages is decreased. This may be a factor in lung overload. However, it seems
27 that the deposited mass or number of particles must exceed some threshold below which
28 increases in loading do not affect translocation rate to the lymph nodes (Ferin and Feldstein,
29 1978; LaBelle and Brieger, 1961). In addition, the rate of translocation to the lymphatic system
30 may be somewhat particle size dependent. Although no human data are available, translocation
31 of latex particles to the lymph nodes of rats was greater for 0.5- to 2-^m particles than for 5- and
March 2001 7-37 DRAFT-DO NOT QUOTE OR CITE
-------
1 9-yUm particles (Takahashi et al., 1992), and smaller particles within the 3- to 15-yum size range
2 were found to be translocated at faster rates than were larger sizes (Snipes and Clem, 1981).
3 On the other hand, translocation to the lymph nodes was similar for both 0.4-,um barium sulfate
4 or 0.02-,um gold colloid particles (Takahashi et al., 1987). It seems that particles <2 ,wm clear to
5 the lymphatic system at a rate independent of size, and it is particles of this size, rather than those
6 > 5 /u.m, that would have significant deposition within the A region following inhalation, hi any
7 case, the normal rate of translocation to the lymphatic system is quite slow, and elimination from
8 the lymph nodes is even slower, with half times estimated in tens of years (Roy, 1989).
9 Soluble particles depositing in the A region may be cleared rapidly via absorption through
10 the epithelial surface into the blood. Actual rates depend on the size of the particle (i.e., solute
11 size), with smaller molecular weight solutes clearing faster than larger ones. Absorption may be
12 considered as a two stage process, with the first stage being dissociation of the deposited
13 particles into material that can be absorbed into the circulation (i.e., dissolution), and the second
14 stage being uptake of this material. Each of these stages may be time dependent. The rate of
15 dissolution depends on a number of factors, including particle surface area and chemical
16 structure. A portion of the dissolved material may be absorbed more slowly because of binding
17 to respiratory tract components. Accordingly, there is a very wide range for absorption rates,
18 depending on the physicochemical properties of the material deposited.
19
20 7.3.3 Interspecies Patterns of Clearance
21 The inability to study the retention of certain materials in humans for direct risk assessment
22 requires use of laboratory animals. Because dosimetry depends on clearance rates and routes,
23 adequate toxicologic assessment necessitates that clearance kinetics in these animals be related to
24 those in humans. The basic mechanisms and overall patterns of clearance from the respiratory
25 tract are similar in humans and most other mammals. However, regional clearance rates can
26 show substantial variation between species, even for similar particles deposited under
27 comparable exposure conditions, as extensively reviewed elsewhere (U.S. Environmental
28 Protection Agency, 1996; Schlesinger et al., 1997; Snipes et al., 1989).
29 In general, there are species-dependent rate constants for various clearance pathways.
30 Differences in regional and total clearance rates between some species are a reflection of
31 differences in mechanical clearance processes. For example, the relative proportion of particles
March 2001 7-38 DRAFT-DO NOT QUOTE OR CITE
-------
1 cleared from the A region in the short and longer term phases differs between laboratory rodents
2 and larger mammals, with a greater percentage cleared in the faster phase in rodents. A recent
3 study (Oberdorster et al., 1997) showed interstrain differences in mice and rats in the handling of
4 particles by alveolar macrophages. Macrophages of B6C3F1 mice could not phagocytize 10-//m
5 particles, but those of CSV black/6Jmice did. In addition, the nonphagocytized lQ-/um particles
6 were efficiently eliminated from the alveolar region, whereas previous work in rats found that
7 these large particles, after uptake by macrophages, were retained persistently (Snipes and Clem,
8 1981; Oberdorster et al., 1992). The end result of interspecies differences in clearance for
9 consideration in assessing particle dosimetry is that the retention of deposited particles can differ
10 between species, and this may result in differences in response to similar particulate exposure
11 atmospheres.
12 Hsieh and Yu (1998) summarized the existing data on pulmonary clearance of inhaled,
13 poorly soluble particles in the rat, mouse, guinea pig, dog, monkey, and human. Clearance at
14 different initial lung burdens, ranging from 0.001 to 10 mg particles/g lung, was analyzed using a
15 two-phase exponential decay function. Two clearance phases in the alveolar region, namely fast
16 and slow, were associated with mechanical clearance along two pathways, the former with the
17 mucociliary system and the latter with the lymph nodes. Rats and mice were noted to be fast
18 clearers compared to the other species. Increasing the initial lung burden resulted in an
19 increasing mass fraction of particles cleared by the slower phase. As lung burden increased
20 beyond 1 mg particles/g lung, the fraction cleared by the slow phase increased to almost 100%
21 for all species. However, the rate for the fast phase was similar in all species and did not change
22 with increasing lung burden of particles, while the rate for the slow phase decreased with
23 increasing lung burden. At elevated burdens, the "overload" effect on clearance rate was greater
24 in rats than in humans, an observation consistent with previous findings (Snipes, 1989).
25
26 7.3.4 Biological Factors Modulating Clearance
27 A number of factors have been assessed in terms of modulation of normal clearance
28 patterns. These include aging, gender, workload, disease, and irritant inhalation, and have been
29 discussed in detail previously (U.S. Environmental Protection Agency, 1996).
30
31
March 2001 7-39 DRAFT-DO NOT QUOTE OR CITE
-------
1 7.3.4.1 Age
2 Studies previously described (U.S. Environmental Protection Agency, 1996) indicated that
3 there appeared to be no clear evidence for any age-related differences in clearance from the
4 respiratory tract, either from child to adult, or young adult to elderly. Studies of mucociliary
5 function have shown either no changes or some slowing in mucous clearance function with age
6 after maturity, but at a rate that would be unlikely to significantly affect overall clearance
7 kinetics.
8
9 7.3.4.2 Gender
10 Previous studies (U.S. Environmental Protection Agency, 1996) indicated no gender related
11 differences in nasal mucociliary clearance rates in children (Passali and Bianchini Ciampoli,
12 1985) nor in tracheal transport rates in adults (Yeates et al., 1975).
13
14 7.3.4.3 Physical Activity
15 The effect of increased physical activity on mucociliary clearance is unresolved, with
16 previously discussed studies (U.S. Environmental Protection Agency, 1996) indicating either no
17 effect or an increased clearance rate with exercise. However, it is possible to have an enhanced
18 mucus transport by nonmucociliary mechanisms such as a two-phase gas-liquid interaction.
19 During exercise, breathing patterns become similar to "huffing", fast expiration compared to
20 inspiration. With this breathing mode, effective mucus transport has been demonstrated in
21 simulated airway models (Kim et al., 1987). There are no data concerning changes in A region
22 clearance with increased activity levels. Breathing with an increased tidal volume was noted to
23 increase the rate of particle clearance from the A region, and this was suggested to result from
24 distension-related evacuation of surfactant into proximal airways, resulting in a facilitated
25 movement of particle-laden macrophages or uningested particles because of the accelerated
26 motion of the alveolar fluid film (John et al., 1994).
27
28 7.3 A A Respiratory Tract Disease
29 Various respiratory tract diseases are associated with clearance alterations. The
30 examination of clearance in individuals with lung disease requires careful interpretation of results
31 because differences in deposition of particles used to assess clearance function may occur
March 2001 7-40 DRAFT-DO NOT QUOTE OR CITE
-------
1 between normal individuals and those with respiratory disease; this would impact directly on the
2 measured clearance rates, especially in the tracheobronchial tree. Earlier studies reported in U.S.
3 Environmental Protection Agency (1996) noted findings of slower nasal mucociliary clearance in
4 humans with chronic sinusitis, bronchiectasis, rhinitis, or cystic fibrosis and slowed bronchial
5 mucus transport associated with bronchial carcinoma, chronic bronchitis, asthma, and various
6 acute respiratory infections. However, a recent study by Svartengren et al. (1996a) concluded,
7 based on deposition and clearance patterns, that particles cleared equally effectively from the
8 small ciliated airways of healthy humans and those with mild to moderate asthma. However, this
9 similarity was ascribed to effective therapy for the asthmatics.
10 In another study, Svartengren et al. (1996b) examined clearance from the TB region in
11 adults with chronic bronchitis who inhaled 6-/zm Teflon particles. Based on calculations,
12 particle deposition was assumed to be in small ciliated airways at low flow and in larger airways
13 at higher flow. The results were compared to that obtained in healthy subjects from other
14 studies. At low flow, a larger fraction of particles was retained over 72 h in people with chronic
15 bronchitis compared to healthy subjects, indicating that clearance resulting from spontaneous
16 cough could not fully compensate for impaired mucociliary transport in small airways. For larger
17 airways, patients with chronic bronchitis cleared a larger fraction of the deposited particles over
18 72 h than did healthy subjects, but this was reportedly because of differences in deposition
19 resulting from airway obstruction.
20 An important mechanism of clearance from the tracheobronchial region, under some
21 circumstances, is cough. Although cough is generally a reaction to an inhaled stimulus, in some
22 individuals with respiratory disease, spontaneous coughing also serves to clear the upper
23 bronchial airways of deposited substances by dislodging mucus from the airway surface. Recent
24 studies confirm that this mechanism likely plays a significant role in clearance for people with
25 mucus hypersecretion, at least for the upper bronchial tree, and for a wide range of deposited
26 particle sizes (0.5 to 5 /wm) (Toms et al., 1997; Groth et al., 1997). There appears to be a general
27 trend towards an association between the extent (i.e., number) of spontaneous coughs and the rate
28 of particle clearance, with faster clearance associated with a greater number of coughs (Groth
29 et al., 1997). Thus, recent evidence continues to support cough as an adjunct to mucociliary
30 movement in the removal of particles from the lungs of individuals with COPD. However, some
31 recent evidence suggests that, like mucociliary function, cough-induced clearance may become
March 2001 7-41 DRAFT-DO NOT QUOTE OR CITE
-------
1 depressed with worsening airway disease. Noone et al. (1999) found that the efficacy of
2 clearance via cough in patients with primary ciliary dyskinesia, who rely on coughing for
3 clearance because of immotile cilia, correlated with lung function (FEV1), in that decreased
4 cough clearance was associated with decreased percentage of predicted FEV1.
5 Earlier reported studies (U.S. Environmental Protection Agency, 1996) indicated that rates
6 of A region particle clearance were reduced in humans with chronic obstructive lung disease and
7 in laboratory animals with viral infections, whereas the viability and functional activity of
8 macrophages was impaired in human asthmatics and in animals with viral induced lung
9 infections. However, any modification of functional properties of macrophages appears to be
10 injury specific, in that they reflect the nature and anatomic pattern of disease.
11 A factor that may affect clearance of particles is the integrity of the epithelial surface lining
12 of the lungs. Damage or injury to the epithelium may result from disease or from the inhalation
13 of chemical irritants. Earlier studies performed with particle instillation had shown that alveolar
14 epithelial damage at the time of deposition in mice resulted in increased translocation of inert
15 carbon to pulmonary interstitial macrophages (Adamson and Hedgecock, 1995). A similar
16 response was observed in a more recent assessment (Adamson and Prieditis, 1998), whereby
17 silica (<0.3 /urn) was instilled into a lung having alveolar epithelial damage, as evidenced by
18 increased permeability, and particles were noted to reach the interstitium and lymph nodes.
19
20
21 7.4 PARTICLE OVERLOAD
22 Experimental studies using some laboratory rodents have employed high exposure
23 concentrations of relatively nontoxic, poorly soluble particles. These particle loads interfered
24 with normal clearance mechanisms, producing clearance rates different from those that would
25 occur at lower exposure levels. Prolonged exposure to high particle concentrations is associated
26 with a phenomenon that has been termed particle "overload", defined as the overwhelming of
27 macrophage-mediated clearance by the deposition of particles at a rate that exceeds the capacity
28 of that clearance pathway. It has been hypothesized that in the rat, overload will begin when
29 deposition approaches 1 mg particles/g lung tissue (Morrow, 1988). Overload is a nonspecific
30 effect noted in experimental studies using many different kinds of poorly soluble particles and
31 results in A region clearance slowing or stasis, with an associated chronic inflammation and
March 2001 7-42 DRAFT-DO NOT QUOTE OR CITE
-------
1 aggregation of macrophages in the lungs and increased translocation of particles into the
2 interstitium.
3 The relevance of lung overload to humans exposed to poorly soluble, nonfibrous particles
4 remains unclear. Although it is likely to be of little relevance for most "real world" ambient
5 exposures, it may be of concern in interpreting some long-term experimental exposure data and,
6 perhaps, also for occupational exposures. For example, it has been suggested that a condition
7 called progressive massive fibrosis, which is unique to humans, has features indicating that dust
8 overload is a factor in its pathogenesis (Green, 2000). This condition is associated with
9 cumulative dust exposure and impaired clearance, and can occur following high exposure
10 concentrations associated with occupational situations. In addition, the relevance to humans is
11 clouded by the suggestion that macrophage-mediated clearance is normally slower and perhaps of
12 less relative importance in overall clearance in humans than in rats (Morrow, 1994), and that
13 there can be significant differences in macrophage loading between species. On the other hand,
14 overload may be a factor in individuals with compromised lungs under normal exposure
15 conditions. Thus, it has been hypothesized (Miller et al., 1995) that localized overload of particle
16 clearance mechanisms in people with compromised lung status may occur, whereby these
17 mechanisms are overwhelmed, resulting in morbidity or mortality from particle exposure.
18
19
20 7.5 COMPARISON OF DEPOSITION AND CLEARANCE PATTERNS OF
21 PARTICLES ADMINISTERED BY INHALATION AND
22 INTRATRACHEAL INSTILLATION
23 The most relevant exposure route to evaluate the toxicity of particulate matter is inhalation.
24 However, many studies delivered particles by intratracheal instillation. This latter technique has
25 been used because it is easy to perform; requires significantly less effort, cost, and amount of test
26 material than does inhalation; and can deliver a known, exact dose of a toxicant to the lungs.
27 Because particle disposition is a determinant of dose, it is important to compare deposition and
28 clearance of particles delivered by these two routes. However, in most instillation studies, the
29 effect of this route of administration on particle deposition and clearance per se was not
30 examined. Although these parameters were evaluated in some studies, it has been very difficult
31 to compare particle deposition/clearance between different inhalation and instillation studies
March 2001 7-43 DRAFT-DO NOT QUOTE OR CITE
-------
1 because of differences in experimental procedures and in the manner by which particle
2 deposition/clearance was quantitated. A recent paper provided a detailed evaluation of the role
3 of instillation in respiratory tract dosimetry and toxicology studies (Driscoll et al., 2000), and a
4 short summary derived from this paper is provided in this section.
5 The pattern of initial regional deposition is strongly influenced by the exposure technique
6 used. Furthermore, the patterns within specific respiratory tract regions also are influenced in
7 this regard. Depending on particle size, inhalation results in varying degrees of deposition within
8 the ET airways, a region that is completely bypassed by instillation. Thus, differences in amount
9 of particles deposited in the lower airways will occur between the two procedures.
10 The exposure technique also influences the intrapulmonary distribution of particles, which
11 potentially would affect routes and rates of ultimate clearance from the lungs and dose delivered
12 to specific sites within the respiratory tract or to extrapulmonary organs. Intratracheal instillation
13 tends to disperse particles fairly evenly within the tracheobronchial tree, but can result in
14 heterogeneous distribution in the alveolar region, whereas inhalation tends to produce a more
15 homogeneous distribution throughout the major conducting airways as well as the alveolar region
16 for the same particles. Thus, inhalation results in a randomized distribution of particles within
17 the lungs, whereas intratracheal instillation produces an heterogeneous distribution in that the
18 periphery of the lung receives little particle load and most of the instilled particles are found in
19 regions that have a short path length from the major airways. Furthermore, inhalation results in
20 greater deposition in apical areas of the lungs and less in basal areas, whereas intratracheal
21 instillation results in less apical than basal deposition.
22 Comparison of the kinetics of clearance of particles administered by instillation or
23 inhalation have shown similarities, as well as differences, in rates for different clearance phases,
24 dependent on the exposure technique used. However, some of the differences in kinetics may be
25 explained by differences in the initial sites of deposition.
26 One of the major pathways of clearance involves particle uptake and removal via
27 pulmonary macrophages. Domes and Valberg (1992) noted that inhalation resulted in a lower
28 percentage of particles recovered in lavaged cells and a more even distribution of particles among
29 macrophages. More individual cells received measurable amounts of particles via inhalation than
30 via intratracheal instillation, whereas with the latter, many cells received little or no particles,
31 although others received very high burdens. Furthermore, with intratracheal instillation,
March 2001 7-44 DRAFT-DO NOT QUOTE OR CITE
-------
1 macrophages at the lung periphery contained few, if any, particles, whereas cells in the regions of
2 highest deposition were overloaded, reflecting the heterogeneity of particle distribution when
3 particles are administered via instillation. Thus, the route of exposure influences the particle
4 distribution in the macrophage population and could, by assumption, influence clearance
5 pathways and clearance kinetics.
6 In conclusion, inhalation may result in deposition within the ET region, the extent of which
7 depends on the size of the particles used. Of course, intratracheal instillation bypasses this
8 portion of the respiratory tract and delivers particles directly to the tracheobronchial tree.
9 Although some studies indicate that short (0 to 2 days) and long (100 to 300 days postexposure)
10 phases of clearance of insoluble particles delivered either by inhalation or intratracheal
11 instillation are similar, other studies indicate that the percentage retention of particles delivered
12 by instillation is greater than that for inhalation, at least up to 30 days postexposure. Thus, there
13 is some inconsistency. Perhaps the most consistent conclusion regarding differences between
14 inhalation and intratracheal instillation is related to the intrapulmonary distribution of particles.
15 Inhalation generally results in a fairly homogeneous distribution of particles throughout the
16 lungs. On the other hand, instillation results in a heterogeneous distribution, especially within
17 the alveolar region, and focally high concentrations of particles. The bulk of instilled material
18 penetrates beyond the major tracheobronchial airways, but the lung periphery is often virtually
19 devoid of particles. This difference is reflected in particle burdens within macrophages, with
20 those from animals inhaling particles being burdened more homogeneously and those from
21 animals with instilled particles showing some populations of cells with no particles and others
22 with heavy burdens. This difference reflects on clearance pathways, dose to cells and tissues,
23 and systemic absorption. Exposure method, thus, clearly influences dose distribution.
24
25
26 7.6 MODELING THE DISPOSITION OF PARTICLES IN THE
27 RESPIRATORY TRACT
28 7.6.1 Modeling Deposition and Clearance
29 The biologic effects of inhaled particles are a function of their disposition. This, in turn,
30 depends on their patterns of both deposition and clearance. Removal of deposited materials
31 involves the competing processes of macrophage-mediated clearance and dissolution-absorption.
March 2001 7-45 DRAFT-DO NOT QUOTE OR CITE
-------
1 Over the years, mathematical models for predicting deposition, clearance and, ultimately,
2 retention of particles in the respiratory tract have been developed. Such models help interpret
3 experimental data and can be used to make predictions of deposition for cases where data are not
4 available.
5 A review of various mathematical deposition models was given by Morrow and Yu (1993)
6 and in U.S. Environmental Protection Agency (1996). There are three major elements involved
7 in mathematical modeling. First, a structural model of the airways must be specified in
8 mathematical terms. Second, deposition efficiency in each airway must be derived for each of
9 the various deposition mechanisms. Finally, a computational procedure must be developed to
10 account for the transport and deposition of the particles in the airways. As noted earlier, most
11 models are deterministic, in that particle deposition probabilities are calculated using anatomical
12 and airflow information on an airway generation by airway generation basis. Other models are
13 stochastic, whereby modeling is performed using individual particle trajectories and finite
14 element simulations of airflow.
15 Recent reports involve modeling the deposition of ultrafine particles and deposition at
16 airway bifurcations. Zhang and Martonen (1997) used a mathematical model to simulate
17 diffusion deposition of ultrafine particles in the human upper tracheobronchial tree and compared
18 the results to those in a hollow cast obtained by Cohen et al. (1990). The model was in good
19 agreement with experimental data. Zhang and Martonen (1997) studied the inertial deposition of
20 particles in symmetric three-dimensional models of airway bifurcations, mathematically
21 examining effects of geometry and flow. They developed equations for use in predicting
22 deposition based on Stokes numbers, Reynolds numbers, and bifurcation angles for specific
23 inflows.
24 Models for deposition, clearance, and dosimetry of the respiratory tract of humans have
25 been available for the past four decades. The International Commission on Radiological
26 Protection (ICRP) has recommended three different mathematical models during this time period
27 (International Commission on Radiological Protection, 1960, 1979, 1994). These models make
28 it possible to calculate the mass deposition and retention in different parts of the respiratory tract
29 and provide, if needed, mathematical descriptions of the translocation of portions of the
30 deposited material to other organs and tissues beyond the respiratory tract.
March 2001 7-46 DRAFT-DO NOT QUOTE OR CITE
-------
1 A morphological model based on laboratory data from planar gamma camera and single-
2 photon emission tomography images has been developed (Martonen et al., 2000; Segal et al.,
3 2000b). This model defines the parenchymal wall in mathematical terms, divides the lung into
4 distinct left and right components, derives a set of branching angles from experimental
5 measurements, and confines the branching network within the left and right components (so there
6 is no overlapping of airways). The authors conclude that this more physiologically realistic
7 model can be used to calculate PM deposition patterns for risk assessment.
8 Musante and Martonen (2000c) developed an age-dependent theoretical model to predict
9 dosimetry in the lungs of children. The model comprises dimensions of individual airways and
10 geometry of branching airway networks within developing lungs and breathing parameters as a
11 function of age. The model suggests that particle size, age, and activity level markedly affect
12 deposition patterns of inhaled particles. Simulations thus far predict a lung deposition fraction of
13 38% in an adult and 73%, nearly twice as high, in a 7-mo-old for 2/2-particles inhaled during
14 heavy breathing. The authors conclude that use of this model will be useful for estimating dose
15 delivered to sensitive subpopulations, such as children.
16 Segal et al. (2000a) developed a computer model for airflow and particle motion in the
17 lungs of children to study how airway disease, specifically cancer, affects inhaled PM deposition.
18 The model considers how tumor characteristics (size and location) and ventilatory parameters
19 (breathing rates and tidal volumes) influence particle trajectories and deposition patterns. The
20 findings indicate that PM may be deposited on the upstream surfaces of tumors because of
21 enhanced efficiency of inertial impaction. Also, submicron particles and larger particles,
22 respectively, may be deposited on the downstream surfaces of tumors because of enhanced
23 efficiency of diffusion and sedimentation. The mechanisms of diffusion and sedimentation are
24 functions of the particle residence times in airways. Eddies downstream of tumors would trap
25 particles and allow more time for deposition to occur by diffusion and sedimentation. The
26 authors conclude that particle deposition is complicated by the presence of airway disease but
27 that the effects are systematic and predictable.
28 Broday and Georgopoulos (2000) recently have presented a model that solves a variant of
29 the general dynamic equation for size evolution of respirable particles within human
30 tracheobronchial airways. The model considers polydisperse aerosols with respect to size and
31 heterosperse with respect to thermodynamic state and chemical composition. The aerosols have
March 2001 7-47 DRAFT-DO NOT QUOTE OR CITE
-------
1 an initial bimodal lognormal size distribution that evolves with time in response to condensation-
2 evaporation and deposition processes. Simulations reveal that submicron size particles grow
3 rapidly and cause increased number and mass fractions of the particle population to be found in
4 the intermediate size range. Because deposition by diffusion decreases with increasing size, fine
5 hygroscopic particles persist longer in the inspired air than nonhygroscopic particles of
6 comparable initial size distribution. In contrast, the enhanced deposition fraction of hygroscopic
7 particles, initially from the intermediate size range, increases their deposition fraction in the
8 airways. The model demonstrates that the combined effect of growth and deposition tends to
9 decrease the size nonuniformiry of persistent particles in the airways and form an aerosol that is
10 characterized by a smaller variance; these factors also alter the deposition profile along airways.
11 Another respiratory tract dosimetry model was developed, concurrently with the new ICRP
12 model, by the National Council on Radiation Protection and Measurements (NCRP) (1997).
13 As with the ICRP model (International Commission on Radiological Protection, 1994), the new
14 NCRP model addresses inhalability of particles, revised subregions of the respiratory tract,
15 dissolution-absorption as an important aspect of the model, and body size and age. The NCRP
16 model defines the respiratory tract in terms of a naso-oro-pharyngo-laryngeal (NOPL) region, a
17 tracheobronchial (TB) region, a pulmonary (P) region, and lung-associated lymph nodes (LN).
18 Deposition and clearance are calculated separately for each of these regions. As with the 1994
19 ICRP model, inhalability of aerosol particles is considered, and deposition in the various regions
20 of the respiratory tract is modeled using methods that relate to mechanisms of inertial impaction,
21 sedimentation, and diffusion.
22 Fractional deposition in the NOPL region was developed from empirical relationships
23 between particle diameter and air flow rate. Deposition in the TB and P regions were projected
24 from model calculations based on geometric or aerodynamic particle diameter and physical
25 deposition mechanisms such as impaction, sedimentation, diffusion, and interception.
26 Deposition in the TB and P regions used the lung model of Yeh and Schum (1980), with a
27 method of calculation similar to that of Findeisen (1935) and Landahl (1950). This method was
28 modified to accomodate an adjustment of lung volume and substitution of realistic deposition
29 equations. These calculations were based on air flow information and idealized morphometry,
30 using a typical pathway model. Comparison of regional deposition fraction predictions between
31 the NCRP and ICRP models was provided in U.S. Environmental Protection Agency (1996).
March 2001 7-48 DRAFT-DO NOT QUOTE OR CITE
-------
1 Inhalability was defined as per the American Conference of Governmental Industrial Hygenists
2 (1985) definition. Breathing frequency, tidal volume, and functional residual capacity are the
3 ventilatory factors used to model deposition. These were related to body weight and to three
4 levels of physical activity, namely low activity, light exertion and heavy exertion.
5 Clearance from all regions of the respiratory tract was considered to result from
6 competitive mechanical and absorptive mechanisms. Mechanical clearance in the NOPL and TB
7 regions was considered to result from mucociliary transport. This was represented in the model
8 as a series of escalators moving towards the glottis and where each airway had an effective
9 clearance velocity. Clearance from the P region was represented by fractional daily clearance
10 rates to the TB region, the pulmonary LN region, and the blood. A fundamental assumption in
11 the model was that the rates for absorption into blood were the same in all regions of the
12 respiratory tract; the rates of dissolution-absorption of particles and their constituents were
13 derived from clearance data primarily from laboratory animals. The effect of body growth on
14 particle deposition also was considered in the model, but particle clearance rates were assumed to
15 be independent of age. Some consideration for compromised individuals was incorporated into
16 the model by altering rates (compared to normal) for the NOPL and TB regions.
17 Mathematical deposition models for deposition in a number of nonhuman species have
18 been developed and discussed previously (U.S. Environmental Protection Agency, 1996).
19 Despite difficulties, modeling studies in laboratory animals remain a useful step in extrapolating
20 exposure-dose-response relationships from laboratory animals to human. Some additional work
21 on modeling deposition in animals has been reported, but it merely expands on work and
22 approaches already noted in the 1996 PM AQCD (U.S. Environmental Protection Agency, 1996).
23 Respiratory-tract clearance begins immediately on deposition of inhaled particles. Given
24 sufficient time, the deposited particles may be removed completely by these clearance processes.
25 However, single inhalation exposures may be the exception rather than the rule. It generally is
26 accepted that repeated or chronic exposures are common for environmental aerosols. As a result
27 of such exposures, accumulation of particles may occur. Chronic exposures produce respiratory
28 tract burdens of inhaled particles that continue to increase with time until the rate of deposition is
29 balanced by the rate of clearance. This is defined as the "equilibrium respiratory tract burden".
30 It is important to evaluate these accumulation patterns, especially when assessing ambient
31 chronic exposures, because they dictate what the equilibrium respiratory tract burdens of inhaled
March 2001 7-49 DRAFT-DO NOT QUOTE OR CITE
-------
1 particles will be for a specified exposure atmosphere. Equivalent concentrations can be defined
2 as "species-dependent concentrations of airborne particles which, when chronically inhaled,
3 produce equal lung deposits of inhaled particles per gram of lung during a specified exposure
4 period" (Schlesinger et al., 1997). Available data and approaches to evaluate exposure
5 atmospheres that produce similar respiratory tract burdens in laboratory animals and humans
6 have been discussed in detail in the previous criteria document.
7 Several laboratory animal models have been developed to help interpret results from
8 specific studies that involved chronic inhalation exposures to nonradioactive particles (Wolff
9 et al., 1987; Strom et al., 1988; Stober et al., 1994). These models were adapted to data from
10 studies involving high level chronic inhalation exposures in which massive lung burdens of low
11 toxicity, poorly soluble particles were accumulated, but the models have not been adapted to
12 chronic exposures to low concentrations of aerosols in which particle overload does not occur.
13 Asgharian et al. (2000) described a method for calculating a deposited fraction for a
14 specific size distribution based on a summary of published data on regional deposition of
15 different size particles. The method is based on constructing nomograms that are used to
16 estimate alveolar deposition fractions for three species (human, monkey, and rat). The data is
17 then incorporated into a regression model that calculates more exact deposition fractions. The
18 model is somewhat constrained at present because of limitations in the underlying deposition
19 database.
20 Hofmann et al. (2000) used three different morphometric models of the rat lung to compute
21 particle deposition in the acinar airways: the multipath lung model (MPL), with a fixed airway
22 geometry; the stochastic lung (SL) model, with a randomly selected branching structure; and a
23 hybrid of the MPL and SL models. They calculated total and regional deposition for a range of
24 particle sizes during quiet and heavy breathing. Although the total bronchial and acinar
25 deposition fractions were similar for the three models, the SL and the hybrid models predicted a
26 substantial variation in particle deposition among different acini. Acinar deposition variances in
27 the MPL model were consistently smaller than in the SL and the hybrid lung models. The
28 authors conclude that the similarity of acinar deposition variations in the latter two models and
29 their independence of the breathing pattern suggest the heterogeneity of the acinar airway
30 structure is primarily responsible for the heterogeneity of acinar particle deposition.
31
March 2001 7-50 DRAFT-DO NOT QUOTE OR CITE
-------
1 7.6.2 Models To Estimate Retained Dose
2 Models have been used routinely to express retained dose in terms of temporal patterns for
3 alveolar retention of acutely inhaled materials. Available information for a variety of
4 mammalian species and humans can be used to predict deposition patterns in the respiratory tract
5 for inhalable aerosols with reasonable degrees of accuracy. Additionally, alveolar clearance data
6 for mammalian species commonly used in inhalation studies are available from numerous
7 experiments that involved small amounts of inhaled radioactive particles.
8 An important factor in using models to predict retention patterns in laboratory animals or
9 humans is the dissolution-absorption rate of the inhaled material. Factors that affect the
10 dissolution of materials or the leaching of their constituents in physiological fluids, and the
11 subsequent absorption of these constituents, are not fully understood. Solubility is known to be
12 influenced by the surface-to-volume ratio and other surface properties of particles (Mercer, 1967;
13 Morrow, 1973). The rates at which dissolution and absorption processes occur are influenced by
14 factors that include the chemical composition of the material. Temperature history of materials is
15 an important consideration for some metal oxides. For example, in controlled laboratory
16 environments, the solubility of oxides usually decreases when the oxides are produced at high
17 temperatures, which generally results in compact particles having small surface-to-volume ratios.
18 It is sometimes possible to accurately predict dissolution-absorption characteristics of materials
19 based on physical/chemical considerations; but, predictions for in vivo dissolution-absorption
20 rates for most materials, especially if they contain multivalent cations or anions, should be
21 confirmed experimentally.
22 Phagocytic cells, primarily macrophages, clearly play a role in dissolution-absorption of
23 particles retained in the respiratory tract (Kreyling, 1992). Some particles dissolve within the
24 phagosomes because of the acidic milieu in those organelles (Lundborg et al., 1984, 1985), but
25 the dissolved material may remain associated with the phagosomes or other organelles in the
26 macrophage rather than diffuse out of the macrophage to be absorbed and transported elsewhere
27 (Cuddihy, 1984). This same phenomenon has been reported for organic materials. For example,
28 covalent binding of benzo[a]pyrene or metabolites to cellular macromolecules resulted in an
29 increased alveolar retention time for that compound after inhalation exposures of rats (Medinsky
30 and Kampcik, 1985). Understanding these phenomena and recognizing species similarities and
March 2001 7-51 DRAFT-DO NOT QUOTE OR CITE
-------
1 differences are important for evaluating alveolar retention and clearance processes and
2 interpreting results of inhalation studies.
3 Dissolution-absorption of materials in the respiratory tract is clearly dependent on the
4 chemical and physical attributes of the material. Although it is possible to predict rates of
5 dissolution-absorption, it is prudent to experimentally determine this important clearance
6 parameter. It is important to understand the impact of this clearance process for the lung, TLNs,
7 and other body organs that might receive particles, or their constituents that enter the circulatory
8 system from the lung.
9 Insufficient data were available to adequately model long-term retention of particles
10 deposited in the conducting airways of any mammalian species at the time of the previous
11 document, and this remains the case. Additional research must be done to provide the
12 information needed to properly evaluate retention of particles in conducting airways.
13 However, a number of earlier studies discussed in the previous document and in
14 Section 7.2.2.2 herein noted that some particles were retained for relatively long times in the
15 upper respiratory tract and tracheobronchial regions, effectively contradicting the general
16 conclusion that almost all inhaled particles that deposit in the TB region clear within hours or
17 days. These studies have demonstrated that variable portions of the particles that deposit in, or
18 are cleared through, the TB region are retained with half times on the order of weeks or months.
19 Long-term retention and clearance patterns for particles that deposit in the head airways and TB
20 region must continue to be thoroughly evaluated because of the implications of this information
21 for respiratory tract dosimetry and risk assessment.
22 Model projections are possible for the A region using the cumulative information in the
23 scientific literature relevant to deposition, retention, and clearance of inhaled particles.
24 Clearance parameters for six laboratory animal species were summarized in U.S. Environmental
25 Protection Agency (1996). Recently, Nikula et al. (1997) evaluated results in rats exposed to
26 high levels of either diesel soot or coal dust. Although the amount of retained material was
27 similar in both species, the rats retained a greater portion in the lumens of the alveolar ducts and
28 alveoli than did monkeys, whereas the monkeys retained a greater portion of the material in the
29 interstitium than did rats. The investigators concluded that intrapulmonary retention patterns in
30 one species may not be predictive of those in another species at high levels of exposure, but this
31 may not be the case at lower levels.
March 2001 7-52 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Adamson, I. Y. R.; Bowden, D. H. (1981) Dose response of the pulmonary macrophagic system to various
3 particulates and its relationship to transepithelial passage of free particles. Exp. Lung Res. 2: 165-175.
4 Adamson, I. Y. R.; Hedgecock, C. (1995) Patterns of particle deposition and retention after instillation to mouse
5 lung during acute injury and fibrotic repair. Exp. Lung Res. 21: 695-709.
6 Adamson, I. Y. R.; Prieditis, H. (1998) Silica deposition in the lung during epithelial injury potentiates fibrosis and
7 increases particle translocation to lymph nodes. Exp. Lung Res. 24: 293-306.
8 American Conference of Governmental Industrial Hygienists (ACG1H). (1985) Particle size-selective sampling in
9 the workplace: report of the ACGIH technical committee on air sampling procedures. Cincinnati, OH:
10 American Conference of Governmental Industrial Hygienists.
11 Asgharian, B.; Wood, R.; Schlesinger, R. B. (1995) Empirical modeling of particle deposition in the alveolar region
12 of the lungs: a basis for interspecies extrapolation. Fundam. Appl. Toxicol. 27: 232-238.
13 Bennett, W. D.; Ilowite, J. S. (1989) Dual pathway clearance of 9"Tc-DTPA from the bronchial mucosa. Am. Rev.
14 Respir. Dis. 139: 1132-1138.
15 Bennett, W. D.; Zeman, K. L. (1998) Deposition of fine particles in children spontaneously breathing at rest.
16 Inhalation Toxicol. 10: 831-842.
17 Bennett, W. D.; Zeman, K. L.; Kim, C. (1996) Variability of fine particle deposition in healthy adults: effect of age
18 and gender. Am. J. Respir. Crit. Care Med. 153: 1641-1647.
19 Bennett, W. D.; Zeman, K. L.; Kang, C. W.; Schechter, M. S. (1997a) Extrathoracic deposition of inhaled, coarse
20 particles (4.Sum) in children vs adults. Ann. Occup. Hyg. 41(suppl.l): 497-502.
21 Bennett, W. D.; Zeman, K. L.; Kim, C.; Mascarella, J. (1997b) Enhanced deposition of fine particles in COPD
22 patients spontaneously breathing at rest. Inhalation Toxicol. 9: 1-14.
23 Bennett, W. D.; Scheuch, G.; Zeman, K. L.; Brown, J. S.; Kim, C.; Heyder, J.; Stahlhofen, W. (1998) Bronchial
24 airway deposition and retention of particles in inhaled boluses: effect of anatomic dead space. J. Appl.
25 Physiol. 85: 685-694.
26 Bennett, W. D.; Scheuch, G.; Zeman, K. L.; Brown, J. S.; Kim, C.; Heyder, J.; Stahlhofen, W. (1999) Regional
27 deposition and retention of particles in shallow, inhaled boluses: effect if lung volume. J. Appl. Physiol.
28 86: 168-173.
29 Broday, D. M.; Georgopoulos, P. G. (2000) Growth and deposition of hygroscopic particulate matter in the human
30 lungs. Aerosol Sci. Technol.: submitted.
31 Camner, P.; Anderson, M.; Philipson, K.; Bailey, A.; Hashish, A.; Jarvis, N.; Bailey, M.; Svartengren, M. (1997)
32 Human bronchiolar deposition and retention of 6-, 8-, and 10-u particles. Exp. Lung Res. 23: 517-535.
33 Cheng, Y.-S.; Smith, S. M.; Yeh, H.-C.; Kim, D.-B.; Cheng, K.-H.; Swift, D. L. (1995) Deposition of ultrafine
34 aerosols and thoron progeny in replicas of nasal airways of young children. Aerosol. Sci. Technol.
35 23:541-552.
36 Cheng, K.-H.; Cheng, Y.-S.; Yeh, H.-C.; Guilmette, R. A.; Simpson, S. Q.; Yang, Y.-H.; Swift, D. L. (1996) In vivo
37 measurements of nasal airway dimensions and ultrafine aerosol deposition in the human nasal and oral
38 airways. J. Aerosol Sci. 27: 785-801.
39 Cheng, K.-H.; Cheng, Y.-S.; Yeh, H.-C.; Swift, D. L. (1997) An experimental method for measuring aerosol
40 deposition efficiency in the human oral airway. Am. Ind. Hyg. Assoc. J. 58: 207-213.
41 Churg, A.; Brauer, M. (1997) Human lung parenchyma retains PM25. Am. J. Respir. Crit. Care Med.
42 155:2109-2111.
43 Churg, A.; Vedal, S. (1996) Carinal and tubular airway particle concentrations in the large airways of non-smokers
44 in the general population: evidence for high particle concentration at airway carinas. Occup. Environ. Med.
45 53:553-558.
46 Cohen, B. S.; Harley, N. H.; Schlesinger, R. B.; Lippmann, M. (1988) Nonuniform particle deposition on
47 tracheobronchial airways: implications for lung dosimetry. In: Dodgson, J.; McCallum, R. I.; Bailey, M. R.;
48 Fisher, D. R., eds. Inhaled particles VI: proceedings of an international symposium and workshop on lung
49 dosimetry; September 1985; Cambridge, United Kingdom. Ann. Occup. Hyg. 32 (suppl. 1): 1045-1053.
50 Cohen, B. S.; Sussman, R. G.; Lippmann, M. (1990) Ultrafine particle deposition in a human tracheobronchial cast.
51 Aerosol Sci. Technol. 12: 1082-1091.
52 Cohen, B. S.; Xiong, J. Q.; Fang, C.-P.; Li, W. (1998) Deposition of charged particles on lung airways. Health
53 Phys. 74: 554-560.
March 2001 7-53 DRAFT-DO NOT QUOTE OR CITE
-------
1 Comer, J.K., Kleinstreuer, C., Hyun, S. and Kim, C. S. (2000) Aerosol transport and deposition in sequentially
2 bifurcating airways. J. Biomech. Eng. 122:152-158.
3 Crystal, R. G., West, J. B.; Barnes, P. J.; Weibel, E. R., eds. (1997) The lung: scientific foundations. Volume 1.
4 Section III: major components. 2nd ed. Philadelphia, PA: Lippincott-Raven; chapters 30-67, pp. 445-991.
5 Cuddihy, R. G. (1984) Mathematical models for predicting clearance of inhaled radioactive materials. In: Smith, H.;
6 Gerber, G., eds. Lung modelling for inhalation of radioactive materials: proceedings of a meeting jointly
7 organized by the Commission of the European Communities and the National Radiological Protection Board;
8 March; Oxford, United Kingdom. Luxembourg: Commission of the European Communities; pp. 167-179;
9 report no. EUR 9384 EN.
10 Cuddihy, R. G.; Yeh, H. C. (1988) Respiratory tract clearance of particles and substances dissociated from particles.
11 In: Mohr, U.; Dungworth, D.; Kimmerle, G.; Lewkowski, J.; McClellan, R.; Stober, W., eds. Inhalation
12 toxicology: the design and interpretation of inhalation studies and their use in risk assessment. New York,
13 NY: Springer-Verlag; pp. 169-193.
14 Dorries, A. M.; Valberg, P. A. (1992) Heterogeneity of phagocytosis for inhaled versus instilled material. Am. Rev.
15 Respir. Dis. 146:831-837.
16 Driscoll, K. E.; Costa, D. L.; Hatch, G.; Henderson, R.; Oberdorster, G.; Salem, H.; Schlesinger, R. B. (2000)
17 Intratracheal instillation as an exposure technique for the evaluation of respiratory tract toxicity: uses and
18 limitations. Toxicol. Sci. 55: 24-35.
19 Falk, R.; Philipson, K.; Svartengren, M.; Jarvis, N.; Bailey, M.; Camner, P. (1997) Clearance of particles from small
20 ciliated airways. Exp. Lung Res. 23: 495-515.
21 Falk, R.; Philipson, K.; Svartengren, M.; Bergmann, R.; Hofmann, W.; Jarvis, N.; Bailey, M.; Camner, P. (1999)
22 Assessment of long-term bronchiolar clearance of particles from measurements of lung retention and
23 theoretical estimates of regional deposition. Exp. Lung Res. 25: 495-516.
24 Ferin, J. (1977) Effect of particle content of lung on clearance pathways. In: Sanders, C. L.; Schneider, R. P.; Dagle,
25 G. E.; Ragan, H. A., eds. Pulmonary macrophages and epithelial cells: proceedings of the sixteenth annual
26 Hanford biology symposium; September 1976; Richland, WA. Oak Ridge, TN: Energy Research and
27 Development Administration; pp. 414-423. Available from: NTIS, Springfield, VA; CONF-760927.
28 (ERDA symposium series 43).
29 Ferin, J.; Feldstein, M. L. (1978) Pulmonary clearance and hilar lymph node content in rats after particle exposure.
30 Environ. Res. 16: 342-352.
31 Ferin, J.; Oberdorster, G.; Penney, D. P. (1992) Pulmonary retention of ultrafine and fine particles in rats. Am. J.
32 Respir. Cell Mol. Biol. 6: 535-542.
33 Findeisen, W. (1935) Ober das Absetzen kleiner, in der Luft suspendierter Teilchen in der menschlichen Lunge bei
34 der Atmung [The deposition of small airborne particles in the human lung during respiration]. Pfluegers
35 Arch. Gesamte Physiol. Menschen Tiere 236: 367-379.
36 Foster, W. M.; Langenback, E.; Bergofsky, E. H. (1980) Measurement of tracheal and bronchial mucus velocities in
37 man: relation to lung clearance. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 48: 965-971.
38 Frampton, M. W.; Chalupa, D.; Morrow, P. E.; Gibb, F. R.; Oberdorster, G.; Speers, D. M.; Zareba, W.; Utell, M. J.
39 (2000) Deposition and effects of inhaled ultrafine carbon particles in healthy subjects at rest. Presented at:
40 PM2000: particulate matter and health—the scientific basis for regulatory decision-making, specialty
41 conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management Association.
42 Fry, F. A.; Black, A. (1973) Regional deposition and clearance of particles in the human nose. J. Aerosol Sci.
43 4:113-124.
44 Gehr, P.; Schiirch, S.; Berthaiume, Y.; Im Hof, V.; Geiser, M. (1990) Particle retention in airways by surfactant.
45 J. Aerosol Med. 3: 27-43.
46 Gehr, P.; Im Hof, V.; Geiser, M.; Schurch, S. (1991) The fate of particles deposited in the intrapulmonary
47 conducting airways. J. Aerosol Med. 4: 349-362.
48 Green, F. H. Y. (2000) Pulmonary responses to inhaled poorly soluble particulate in the human. In: Gardner, D. E.,
49 ed. ILSI Risk Science Institute Workshop: The Relevance of the Rat Lung Response to Particle Overload for
50 Human Risk Assessment; March, 1998. Inhalation Toxicol. 12: 59-95.
51 Groth, M. L.; Macri, K.; Foster, W. M. (1997) Cough and mucociliary transport of airway particulate in chronic
52 obstructive lung disease. In: Cherry, N.; Ogden, T., eds. Inhaled Particles VIII: proceedings of an
53 international symposium on inhaled particles organised by the British Occupational Hygiene Society;
54 August 1996; Cambridge, UK. Ann. Occup. Hyg. 41(suppl.): 515-521.
March 2001 7-54 DRAFT-DO NOT QUOTE OR CITE
-------
1 Guilmette, R. A.; Cheng, Y. S.; Griffith, W. C. (1997) Characterising the variability in adult human nasal airway
2 dimensions. In: Cherry, N.; Ogden, T., eds. Inhaled Particles VIII: proceedings of an international
3 symposium on inhaled particles organised by the British Occupational Hygiene Society; August 1996;
4 Cambridge, UK. Ann. Occup. Hyg. 41(suppl. 1): 491-496.
5 Harmsen, A. G.; Muggenburg, B. A.; Snipes, M. B.; Bice, D. E. (1985) The role of macrophages in particle
6 translocation from lungs to lymph nodes. Science (Washington, DC) 230: 1277-1280.
7 Heistracher, T.; Hofmann, W. (1997) Flow and deposition patterns in successive airway bifurcations. In: Cherry, N.;
8 Ogden, T., eds. Inhaled Particles VIII: proceedings of an international symposium on inhaled particles
9 organised by the British Occupational Hygiene Society; August 1996; Cambridge, UK. Ann. Occup. Hyg.
10 41(suppl.): 537-542.
11 Henshaw, D. L.; Fews, A. P. (1984) The microdistribution of alpha emitting particles in the human lung.
12 In: Smith, H.; Gerber, G., eds. Lung modelling for inhalation of radioactive materials: proceedings of a
13 meeting jointly organized by the Commission of the European Communities and the National Radiological
14 Protection Board; March; Oxford, United Kingdom. Luxembourg: Commission of the European
15 Communities; pp. 199-218; report no. EUR 9384 EN.
16 Heyder, J.; Rudolf, G. (1977) Deposition of aerosol particles in the human nose. In: Walton, W. H.; McGovern, B.,
17 eds. Inhaled particles IV: proceedings of an international symposium, part 1; September 1975; Edinburgh,
18 United Kingdom. Oxford, United Kingdom: Pergamon Press, Ltd.; pp. 107-126.
19 Killer, F. C. (1991) Health implications of hygroscopic particle growth in the human respiratory tract. J. Aerosol
20 Med. 4: 1-23.
21 Hofmann, W.; Bergmann, R. (1998) Predictions of particle deposition patterns in human and rat airways. Inhalation
22 Toxicol. 10: 557-583.
23 Hofmann, W.; Martonen, T. B.; Graham, R. C. (1989a) Predicted deposition of nonhygroscopic aerosols in the
24 human lung as a function of subject age. J. Aerosol Med. 2: 49-68.
25 Hofmann, W.; Koblmger, L.; Martonen, T. B. (1989b) Structural differences between human and rat lungs:
26 implications for Monte Carlo modeling of aerosol deposition. In: Mahaffey, J. A., ed. 26th Hanford life
27 sciences symposium, modeling for scaling to man: biology, dosimetry, and response. Health Phys.
28 57(suppl. 1): 41-47.
29 Hofmann, W.; Balashazy, I.; Heistracher, T.; Koblinger, L. (1996) The significance of particle deposition patterns in
30 bronchial airway bifurcations for extrapolation modeling. Aerosol Sci. Technol. 25: 305-327.
31 Hofmann, W.; Bergmann, R.; Koblinger, L. (1999) Characterization of local particle deposition patterns in human
32 and rat lungs by different morphometric parameters. J. Aerosol Sci. 30: 651-667.
33 Hofmann, W.; Asgharian, B.; Bergmann, R.; Anjilvel, S.; Miller, F. J. (2000) The effect of heterogeneity of lung
34 structure on particle deposition in the rat lung. Toxicol. Sci. 53: 430-437.
35 Hsieh, T. H.; Yu, C. P. (1998) Two-phase pulmonary clearance of insoluble particles in mammalian species.
36 Inhalation Toxicol. 10: 121-130.
37 International Commission on Radiological Protection. (1960) Report of Committee II on permissible dose for
38 internal radiation (1959). Health Phys. 3.
39 International Commission on Radiological Protection. (1979) Limits for intakes of radionuclides by workers.
40 Oxford, United Kingdom: Pergamon Press; ICRP publication 30, part 1.
41 International Commission on Radiological Protection. (1994) Human respiratory tract model for radiological
42 protection: a report of a task group of the International Commission on Radiological Protection. Oxford,
43 United Kingdom: Elsevier Science Ltd. (ICRP publication 66; Annals of the ICRP: v. 24, nos. 1-3).
44 John, J.; Wollmer, P.; Dahlback, M.; Luts, A.; Jonson, B. (1994) Tidal volume and alveolar clearance of insoluble
45 particles. J. Appl. Physiol. 76: 584-588.
46 Jaques, P. A.; Kim, C. S. (2000) Measurement of total lung deposition of inhaled ultrafine particles in healthy men
47 and women. Inhalation Toxicol. 12: 715-731.
48 Katz, I. M.; Davis, B. M.; Martonen, T. B. (1999) A numerical study of particle motion within the human larynx
49 and trachea. J. Aerosol Sci. 30: 173-183.
50 Kesavanathan, J.; Swift, D. L. (1998) Human nasal passage particle deposition: the effect of particle size, flow rate,
51 and anatomical factors. Aerosol Sci. Technol. 28: 457-463.
52 Kim C. S. (1989) Aerosol deposition in the lung with obstructed airways. J. Aerosol Med. 2:111-120.
53 Kim, C. S. (2000) Methods of calculating lung delivery and deposition of aerosol particles. Respir. Care
54 45:695-711.
March 2001 7-55 DRAFT-DO NOT QUOTE OR CITE
-------
1 Kim C. S. and Fisher, D.M. (1999) Deposition characteristics of aerosol particles in sequentially bifurcating airway
2 models. Aerosol Sci. Technol. 31:198-220.
3 Kim, C. S.; Hu, 5. C. (1998) Regional deposition of inhaled particles in human lungs: comparison between men and
4 women. J. Appl. Physiol. 84: 1834-1844.
5 Kim, C. S. and Jaques, P. A. (2000) Respiratory dose of inhaled ultrafine particles in healthy adults. Phil. Trans.
6 Roy. Soc. London A 358:2693-2705.
7 Kim, C. S.; Kang, T. C. (1997) Comparative measurement of lung deposition of inhaled fine particles in normal
8 subjects and patients with obstructive airway disease. Am. J. Respir. Crit. Care Med. 155: 899-905.
9 Kim, C. S., Iglesias, A. J., and Sackner, M.A. (1987) Mucus clearance by two-phase gas-liquid flow mechanism:
10 Asymmetric periodic flow model. J. Appl. Physiol. 62:959-971.
11 Kim C. S., Eldridge, M. A., Garcia, L. and Wanner A. (1989) Aerosol deposition in the lung with asymmetric
12 airways obstruction: in vivo observation. J. Appl. Physiol. 67:2579-2585.
13 Kim, C. S.; Hu, S. C.; DeWitt, P.; Gerrity, T. R. (1996) Assessment of regional deposition of inhaled particles in
14 human lungs by serial bolus delivery method. J. Appl. Physiol. 81: 2203-2213.
15 Kim, C. S.; Hu, S. C.; DeWitt, P. (2000) Variation of lung deposition dose of inhaled particles with breathing
16 pattern at rest and during moderate exercise. Presented at: PM2000: particulate matter and health—the
17 scientific basis for regulatory decision-making, specialty conference & exhibition; January; Charleston, SC.
18 Pittsburgh, PA: Air & Waste Management Association.
19 Koblinger, L.; Hofmann, W. (1985) Analysis of human lung morphometric data for stochastic aerosol deposition
20 calculations. Phys. Med. Biol. 30: 541-556.
21 Koblinger, L.; Hofmann, W. (1988) Stochastic morphological model of the rat lung. Anat. Rec. 221: 533-539.
22 Kohlhaufl, M.; Brand, P.; Scheuch, G.; Meyer, T. S.; Schulz, H.; Haussinger, K.; Heyder, J. (1999) Increased fine
23 particle deposition in women with asymptomatic nonspecific airway hyperresponsiveness. Am. J. Respir.
24 Crit. Care Med. 159: 902-906.
25 Kreyling, W. G. (1992) Intracellular particle dissolution in alveolar macrophages. Environ. Health Perspect.
26 97:121-126.
27 Kreyling, W. G.; Blanchard, J. D.; Godleski, J. J.; Haeussermann, S.; Heyder, J.; Hutzler, P.; Schulz, H.; Sweeney,
28 T. D.; Takenaka, S.; Ziesenis, A. (1999) Anatomic localization of 24- and 96-h particle retention in canine
29 airways. J. Appl. Physiol. 87: 269-284.
30 LaBelle, C. W.; Brieger, H. (1961) Patterns and mechanisms in the elimination of dust from the lung. In: Davies,
31 C. N., ed. Inhaled particles and vapours: proceedings of an international symposium; March-April 1960;
32 Oxford, United Kingdom. New York, NY: Pergamon Press; pp. 356-368.
33 Landahl, H. D. (1950) On the removal of air-borne droplets by the human respiratory tract: I. the lung. Bull. Math.
34 Biophys. 12: 43-56.
35 Lay, J. C.; Berry, C. R.; Chong, S. K.; Bennett, W. D. (1995) Retention of insoluble particles after local
36 intrabronchial deposition in dogs. J. Appl. Physiol. 79: 1921-1929.
37 Lay, J. C.; Bennett, W. D.; Kim, C. S.; Devlin, R. B.; Bromberg, P. A. (1998) Retention and intracellular
38 distribution of instilled iron oxide particles in human alveolar macrophages. Am. J. Respir. Cell Mol. Biol.
39 18:687-695.
40 Lehnert, B. E.; Morrow, P. E. (1985) Association of 59iron oxide with alveolar macrophages during alveolar
41 clearance. Exp. Lung Res. 9: 1-16.
42 Lehnert, B. E.; Valdez, Y. E.; Bomalaski, S. H. (1988) Analyses of particles in the lung free cell, tracheobronchial
43 lymph nodal, and pleural space compartments following their deposition in the lung as related to lung
44 clearance mechanisms. In: Dodgson, J.; McCallum, R. I.; Bailey, M. R.; Fisher, D. R., eds. Inhaled particles
45 VI: proceedings of an international symposium and workshop on lung dosimetry; September 1985;
46 Cambridge, United Kingdom. Ann. Occup. Hyg. 32 (suppl. 1): 125-140.
47 Leikauf, G.; Yeates, D. B.; Wales, K. A.; Spektor, D.; Albert, R. E.; Lippmann, M. (1981) Effects of sulfuric acid
48 aerosol on respiratory mechanics and mucociliary particle clearance in healthy nonsmoking adults. Am. Ind.
49 Hyg. Assoc. J. 42: 273-282.
50 Leikauf, G. D.; Spektor, D. M.; Albert, R. E.; Lippmann, M. (1984) Dose-dependent effects of submicrometer
51 sulfuric acid aerosol on particle clearance from ciliated human lung airways. Am. Ind. Hyg. Assoc. J.
52 45:285-292.
53 Lennon, S.; Zhang, Z.; Lessmann, R.; Webster, S. (1998) Experiments on particle deposition in the human upper
54 respiratory system. Aerosol Sci. Technol. 28: 464-474.
March 2001 7-56 DRAFT-DO NOT QUOTE OR CITE
-------
1 Lundborg, M.; Lind, B.; Camner, P. (1984) Ability of rabbit alveolar macrophages to dissolve metals. Exp. Lung
2 Res. 7:11-22.
3 Lundborg, M.; Eklund, A.; Lind, B.; Camner, P. (1985) Dissolution of metals by human and rabbit alveolar
4 macrophages. Br. J. Ind. Med. 42: 642-645.
5 Madl, A. K.; Wilson, D. W.; Segall, H. J.; Pinkerton, K. E. (1998) Alteration in lung particle translocation,
6 macrophage function, and microfilament arrangement in monocrotaline-treated rats. Toxicol. Appl.
7 Pharmacol. 153:28-38.
8 Martonen, T. B.; Schroeter, J. D.; Hwang, D.; Fleming, J. S.; Conway, J. H. (2000) Human lung morphology
9 models for particle deposition studies. In: Grant, L. D., ed. PM2000: particulate matter and health.
10 Inhalation Toxicol. 12(suppl. 4): 109-121.
11 Matsui, H.; Randell, S. H.; Peretti, S. W.; Davis, C. W.; Boucher, R. C. (1998) Coordinated clearance of periciliary
12 liquid and mucus from airway surfaces. J. Clin. Invest. 102: 1125-1131.
13 Medinsky, M. A.; Kampcik, S. J. (1985) Pulmonary retention of [14C]benzo/a7pyrene in rats as influenced by the
14 amount instilled. Toxicology 35: 327-336.
15 Mercer, T. T. (1967) On the role of particle size in the dissolution of lung burdens. Health Phys. 13: 1211-1221.
16 Miller, F. J.; Anjilvel, S.; Menache, M. G.; Asgharian, B.; Gerrity, T. R.. (1995) Dosimetric issues relating to
17 particulate toxicity. Inhalation Toxicol. 7: 615-632.
18 Morrow, P. E. (1973) Alveolar clearance of aerosols. Arch. Intern. Med. 131: 101-108.
19 Morrow, P. E. (1977) Clearance kinetics of inhaled particles. In: Brain, J. D.; Proctor, D. F.; Reid, L. M., eds.
20 Respiratory defense mechanisms (in two parts), part II. New York, NY: Marcel Dekker, Inc.; pp. 491-543.
21 (Lenfant, C., ed. Lung biology in health and disease: v. 5).
22 Morrow, P. E. (1986) Factors determining hygroscopic aerosol deposition in airways. Physiol. Rev. 66: 330-376.
23 Morrow, P. E. (1988) Possible mechanisms to explain dust overloading of the lungs. Fundam. Appl. Toxicol.
24 10:369-384.
25 Morrow, P. E. (1994) Mechanisms and significance of "particle overload." In: Mohr, U.; Dungworth, D. L.;
26 Mauderly, J. L.; Oberdorster, G., eds. Toxic and carcinogenic effects of solid particles in the respiratory tract:
27 [proceedings of the 4th international inhalation symposium]; March 1993; Hannover, Germany. Washington,
28 DC: International Life Sciences Institute Press; pp. 17-25.
29 Morrow, P. E.; Yu, C. P. (1993) Models of aerosol behavior in airways and alveoli. In: Moren, F.; Dolovich, M. B.;
30 Newhouse, M. T.; Newman, S. P., eds. Aerosols in medicine: principles, diagnosis and therapy. 2nd rev. ed.
31 Amsterdam, The Netherlands: Elsevier; pp. 157-193.
32 Morrow, P. E.; Gibb, F. R.; Gazioglu, K. M. (1967) A study of particulate clearance from the human lungs.
33 Am. Rev. Respir. Dis. 96: 1209-1221.
34 Musante, C. J.; Martonen, T. B. (1999) Predicted deposition patterns of ambient particulate air pollutants in
35 children's lungs under resting conditions. In: Proceedings of the third colloquium on particulate air pollution
36 and human health; June; Durham, NC. Irvine, CA: University of California, Air Pollution Health Effects
3 7 Laboratory, p. 7-15 - 7-20.
38 Musante, C. J.; Martonen, T. B. (2000a) Computer simulations of particle deposition in the developing human lung.
39 J. Air Waste Manage. Assoc. 50: 1426-1432.
40 Musante, C. J.; Martonen, T. B. (2000b) An extrapolation model to aid in toxicological studies of particulate air
41 pollutants. Presented at: PM2000: particulate matter and health—the scientific basis for regulatory
42 decision-making, specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste
43 Management Association.
44 Musante, C. J.; Martonen, T. B. (2000c) Particulate matter deposition in the lungs of children and adults:
45 predictions of an age-dependent computer model. Presented at: PM2000: particulate matter and health—the
46 scientific basis for regulatory decision-making, specialty conference & exhibition; January; Charleston, SC.
47 Pittsburgh, PA: Air & Waste Management Association.
48 National Council on Radiation Protection and Measurements. (1997) Deposition, retention and dosimetry of inhaled
49 radioactive substances. Bethesda, MD: National Council on Radiation Protection and Measurements; report
50 no. 125.
51 Naumann, B. D.; Schlesinger, R. B. (1986) Assessment of early alveolar particle clearance and macrophage
52 function following an acute inhalation of sulfuric acid mist. Exp. Lung Res. 11: 13-33.
53 Nikula, K. J.; Avila, K. J.; Griffith, W. C.; Mauderly, J. L. (1997) Lung tissue responses and sites of particle
54 retention differ between rats and cynomolgus monkeys exposed chronically to diesel exhaust and coal dust.
55 Fundam. Appl. Toxicol. 37: 37-53.
March 2001 7-57 DRAFT-DO NOT QUOTE OR CITE
-------
1 Nikula, K. J.; Vallyathan, V.; Green, F. H. Y.; Hahn, F. F. (2000) Influence of dose on the distribution of retained
2 particulate material in rat and human lungs. Presented at: PM2000: particulate matter and health—the
3 scientific basis for regulatory decision-making, specialty conference & exhibition; January; Charleston, SC.
4 Pittsburgh, PA: Air & Waste Management Association.
5 Noone, P. G.; Bennett, W. D.; Regnis, J. A.; Zeman, K. L; Carson, J. L.; King, M.; Boucher, R. C.; Knowles, M. R.
6 (1999) Effect of aerosolized uridine-5'-triphosphate on airway clearance with cough in patients with primary
7 ciliary dyskinesia. Am. J. Respir. Crit. Care Med.: 160: 144-149.
8 Oberdorster, G. (1993) Lung dosimerry: pulmonary clearance of inhaled particles. Aerosol Sci. Technol.
9 18:279-289.
10 Oberdorster, G.; Ferin, J.; Gelein, R.; Soderholm, S. C.; Finkelstein, J. (1992) Role of the alveolar macrophage in
11 lung injury: studies with ultrafine particles. Environ. Health Perspect. 97: 193-199.
12 Oberdorster, G.; Cox, C.; Gelein, R. (1997) Intratracheal instillation versus intratracheal inhalation of tracer
13 particles for measuring lung clearance function. Exp. Lung Res. 23: 17-34.
14 Oldham, M. J.; Mannix, R. C.; Phalen, R. F. (1997) Deposition of monodisperse particles in hollow models
15 representing adult and child-size tracheobronchial airways. Health Phys. 72: 827-834.
16 Passali, D.; Bianchini Ciampoli, M. (1985) Normal values of mucociliary transport time in young subjects. Int. J.
17 Pediatr. Otorhinolaryngol. 9: 151-156.
18 Patton, J. S. (1996) Mechanisms of macromolecule absorption by the lungs. Adv. Drug Delivery Rev. 19: 3-36.
19 Pritchard, J. N.; Jefferies, S. J.; Black, A. (1986) Sex differences in the regional deposition of inhaled particles in
20 the 2.5—7.5 urn size range. J. Aerosol Sci. 17: 385-389.
21 Radford, E. P.; Martell, E. A. (1977) Polonium-210: lead-210 ratios as an index of residence time of insoluble
22 particles from cigarette smoke in bronchial epithelium. In: Walton, W. H.; McGovern, B., eds. Inhaled
23 particles IV: proceedings of an international symposium, part 2; September 1975; Edinburgh, United
24 Kingdom. Oxford, United Kingdom: Pergamon Press, Ltd.; pp. 567-581.
25 Roy, M. (3989) Lung clearance modeling on the basis of physiological and biological parameters. Health Phys.
26 57(suppl. 1): 255-262.
27 Rutland, J.; Cole, P. J. (1981) Nasal mucociliary clearance and ciliary beat frequency in cystic fibrosis compared
28 with sinusitis and bronchiectasis. Thorax 36: 654-658.
29 Scheuch, G.; Stahlhofen, W. (1988) Particle deposition of inhaled aerosol boluses in the upper human airways.
30 J. Aerosol Med. 1:29-36.
31 Scheuch, G.; Philipson, K.; Falk, R.; Svartengren, M.; Stahlhofen, W.; Camner, P. (1995) Retention of particles
32 inhaled in boli with and without induced bronchoconstriction. Exp. Lung Res. 21: 901-916.
33 Schlesinger, R. B. (1985) Clearance from the respiratory tract. Fundam. Appl. Toxicol. 5: 435-450.
34 Schlesinger, R. B. (1988) Biological disposition of airborne particles: basic principles and application to vehicular
35 emissions. In: Watson, A. Y.; Bates, R. R.; Kennedy, D., eds. Air pollution, the automobile, and public
36 health. Washington, DC: National Academy Press; pp. 239-298.
37 Schlesinger, R. B. (1995) Deposition and clearance of inhaled particles. In: McClellan, R. O.; Henderson, R. F.,
38 eds. Concepts in inhalation toxicology. 2nd ed. Washington, DC: Taylor & Francis; pp. 191-224.
39 Schlesinger, R. B.; Ben-Jebria, A.; Dahl, A. R.; Snipes, M. B.; Ultman, J. (1997) Disposition of inhaled toxicants.
40 In: Massaro, E. J., ed. Handbook of human toxicology. Boca Raton, FL: CRC Press; pp. 493-550.
41 Segal, R. A.; Martonen, T. B.; Shearer, M.; (2000a) Particle trajectories in the cancerous lungs of children.
42 Presented at: PM2000: particulate matter and health—the scientific basis for regulatory decision-making,
43 specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management
44 Association.
45 Segal, R. A.; Martonen, T. B.; Kim, C. S. (2000b) Comparison of computer simulations to total lung deposition to
46 human subject data in healthy test subjects. J. Air Waste Manage. Assoc. 50: 1262-1268.
47 Smaldone, G. C.; Perry, R. J.; Bennett, W. D.; Messina, M. S.; Zwang, J.; Ilowite, J. (1988) Interpretation of
48 "24 hour lung retention" in studies of mucociliary clearance. J. Aerosol Med. 1:11 -20.
49 Snipes, M. B. (1989) Long-term retention and clearance of particles inhaled by mammalian species. CRC Crit. Rev.
50 Toxicol. 20: 175-211.
51 Snipes, M. B.; Clem, M. F. (1981) Retention of microspheres in the rat lung after intratracheal instillation.
52 Environ. Res. 24:33-41.
53 Snipes, M. B.; McClellan, R. O.; Mauderly, J. L.; Wolff, R. K. (1989) Retention patterns for inhaled particles in the
54 lung: comparisons between laboratory animals and humans for chronic exposures. Health Phys.
55 57(suppl. 1): 69-78.
March 2001 7-58 DRAFT-DO NOT QUOTE OR CITE
-------
1 Stahlhofen, W.; Gebhart, J.; Rudolf, G.; Scheuch, G.; Philipson, K. (1986a) Clearance from the human airways of
2 particles of different sizes deposited from inhaled aerosol boli. In: Aerosols: formation and reactivity,
3 proceedings of the second international aerosol conference; September; Berlin, Federal Republic of
4 Germany. Oxford, United Kingdom: Pergamon Press; pp. 192-196.
5 Stahlhofen, W.; Gebhart, J.; Rudolf, G.; Scheuch, G. (1986b) Measurement of lung clearance with pulses of
6 radioactively-labelled aerosols. J. Aerosol Sci. 17: 333-336.
7 Stanley, P. J.; Wilson, R.; Greenstone, M. A.; Mackay, I. S.; Cole, P. J. (1985) Abnormal nasal mucociliary
8 clearance in patients with rhinitis and its relationship to concomitant chest disease. Br. J. Dis. Chest
9 79:77-82.
10 Stober, W.; Morrow, P. E.; Koch, W.; Morawietz, G. (1994) Alveolar clearance and retention of inhaled insoluble
11 particles in rats simulated by a model inferring macrophage particle load distributions. J. Aerosol Sci.
12 25:975-1002.
13 Strom, K. A.; Chan, T. L.; Johnson, J. T. (1988) Pulmonary retention of inhaled submicron particles in rats: diesel
14 exhaust exposures and lung retention model. In: Dodgson, J.; McCallum, R. I.; Bailey, M. R.; Fischer, D. R.,
15 eds. Inhaled particles VI: proceedings of an international symposium and workshop on lung dosimetry;
16 September 1985; Cambridge, United Kingdom. Ann. Occup. Hyg. 32(suppl. 1): 645-657.
17 Svartengren, K.; Lindestad, P.; Svartengren, M.; Philipson, K.; Bylin, G.; Camner, P. (1995) Added external
18 resistance reduces oropharyngeal deposition and increases lung deposition of aerosol particles in asthmatics.
19 Am. J. Respir. Crit. Care Med. 152: 32-37.
20 Svartengren, K.; Philipson, K.; Svartengren, M.; Anderson, M.; Camner, P. (1996a) Tracheobronchial deposition
21 and clearance in small airways in asthmatic subjects. Eur. Respir. J. 9: 1123-1129.
22 Svartengren, K.; Ericsson, C. H.; Svartengren, M.; Mossberg, B.; Philipson, K.; Camner, P. (1996b) Deposition and
23 clearance in large and small airways in chronic bronchitis. Exp. Lung Res. 22: 555-576.
24 Svartengren, K.; Philipson, K.; Svartengren, M.; Camner, P. (1998) Effect of adrenergic stimulation on clearance
25 from small ciliated airways in healthy subjects. Exp. Lung Res. 24: 149-158.
26 Svartengren, M.; Svartengren, K.; Aghaie, F.; Philipson, K.; Camner, P. (1999) Lung deposition and extremely slow
27 inhalations of particles. Limited effect of induced airway obstruction. Exp. Lung Res. 25: 353-366.
28 Swift, D. L.; Strong, J. C. (1996) Nasal deposition of ultrafine 2l8Po aerosols in human subjects. J. Aerosol Sci.
29 27:1125-1132.
30 Takahashi, S.; Asaho, S.; Kubota, Y.; Sato, H.; Matsuoka, O. (1987) Distribution of l98Au and 133Ba m thoracic and
31 cervical lymph nodes of the rat following the intratracheal instillation of mAu-colloid and 1MBaSO4.
32 J. Radiat. Res. 28: 227-231.
33 Takahashi, S.; Kubota, Y.; Hatsuno, H. (1992) Effect of size on the movement of latex particles in the respiratory
34 tract following local administration. Inhalation Toxicol. 4: 113-123.
35 Toms, N.; Hasani, A.; Pavia, D.; Clarke, S. W.; Agnew, J. E. (1997) Effect of mucus hypersecretion on initial time
36 course of inert particle clearance from the lung. In: Cherry, N.; Ogden, T., eds. Inhaled Particles VIII:
37 proceedings of an international symposium on inhaled particles organised by the British Occupational
38 Hygiene Society; August 1996; Cambridge, UK. Ann. Occup. Hyg. 41(suppl.): 509-514.
39 U.S. Environmental Protection Agency. (1996) Air quality criteria for particulate matter. Research Triangle Park,
40 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
41 Venkataraman, C.; Kao, A. S. (1999) Comparison of particle lung doses from the fine and coarse fractions of urban
42 PM-10 aerosols. Inhalation Toxicol. 11: 151-169.
43 Wolff, R. K.; Henderson, R. F.; Snipes, M. B.; Griffith, W. C.; Mauderly, J. L.; Cuddihy, R. G.; McClellan, R. O.
44 (1987) Alterations in particle accumulation and clearance in lungs of rats chronically exposed to diesel
45 exhaust. Fundam. Appl. Toxicol. 9: 154-166.
46 Xu, G. B.; Yu, C. P. (1986) Effects of age on deposition of inhaled aerosols in the human lung. Aerosol Sci.
47 Technol. 5: 349-357.
48 Yeates, D. B.; Aspin, M. (1978) A mathematical description of the airways of the human lungs, Respir. Physiol.
49 32:91-104.
50 Yeates, D. B.; Aspin, N.; Levison, H.; Jones, M. T.; Bryan, A. C. (1975) Mucociliary tracheal transport rates in
51 man. J. Appl. Physiol. 19: 487-495.
52 Yeates, D. B.; Pitt, B. R.; Spektor, D. M.; Karron, G. A.; Albert, R. E. (1981) Coordination of mucociliary transport
53 in human trachea and intrapulmonary airways. J. Appl. Physiol.: Respir. Environ. Exercise Physiol.
54 51:1057-1064.
March 2001 7-59 DRAFT-DO NOT QUOTE OR CITE
-------
1 Yeh, H.-C.; Schum, G. M. (1980) Models of human lung airways and their application to inhaled particle
2 deposition. Bull. Math. Biol. 42: 461-480.
3 Yu, G.; Zhang, Z.; Lessmann, R. (1998) Fluid flow and particle diffusion in the human upper respiratory system.
4 Aerosol Sci. Technol. 28: 146-158.
5 Zhang, Z.; Martonen, T. (1997) Deposition of ultrafine aerosols in human tracheobronchial airways. Inhalation
6 Toxicol.9:99-110.
March 2001 7-60 DRAFT-DO NOT QUOTE OR CITE
-------
i 8. TOXICOLOGY OF PARTICULATE MATTER
2
3
4 8.1 INTRODUCTION
5 Toxicological research on ambient particulate matter (PM) is used to address several
6 related questions, including (1) what causal mechanisms may be involved in the toxicological
7 response to PM exposures, (2) what factors affect individual or subpopulation susceptibility to
8 the effects of PM exposures, (3) what characteristics of PM (e.g., size, composition) are
9 producing observed toxicity, and (4) what are the combined effects of PM and gaseous
10 co-pollutants in producing toxic responses? A variety of research approaches are used to address
11 these questions, including in vivo studies of human volunteers to controlled exposures; in vivo
12 studies of animals such as nonhuman primates, dogs and rodent species; and in vitro studies of
13 tissue, cellular, genetic, and biochemical systems. Similarly, a variety of exposure conditions are
14 employed, including whole body and nose-only inhalation exposures to artificially generated PM
15 or concentrated ambient air, pulmonary instillation, and in vitro exposure to test materials in
16 solution. The various research designs are targeted to test hypotheses and, ultimately, provide a
17 scientific basis for an improved understanding of the role of PM in producing health effects
18 identified by epidemiological studies.
19 Because of the sparsity of toxicological data on ambient PM at the time the previous PM
20 Air Quality Criteria Document or "PM AQCD" (U.S. Environmental Protection Agency, 1996a)
21 was completed, the discussion of respiratory effects of PM were organized into specific chemical
22 components of ambient PM or model "surrogate" particles (e.g., acid aerosols, metals, ultrafine
23 particles, bioaerosols, "other particle matter"). In this chapter, the conclusions of the 1996 PM
24 AQCD are summarized for each of these components. Since completion of the previous
25 document, there are many new studies demonstrating the potentially adverse effects of
26 combustion-related particles. The main reason for this increased interest in combustion particles
27 is that these particles are typically the dominant sources represented in the fine fraction of
28 ambient PM.
29 This chapter is organized as follows. The respiratory effects of specific components of
30 ambient PM or surrogate particles delivered by in vivo exposures of both humans and laboratory
March 2001 8-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 animals are discussed first (Section 8.2), followed by systemic and cardiovascular effects of
2 particles (Section 8.3) and effects in laboratory animal models that mimic human disease
3 (Section 8.4). The in vitro exposure studies are discussed next (Section 8.5) because they
4 provide valuable information on potential hazardous constituents and mechanisms of PM injury.
5 The remaining section on exposure studies examines the health effects of mixtures of ambient
6 PM or PM surrogates with gaseous pollutants (Section 8.6). This organization provides the
7 underlying data for evaluation in the final section of this chapter (Section 8.7), but it may fail to
8 adequately convey the extensive and intricate linkages among the pulmonary, cardiac, and
9 nervous systems, all of which may be involved individually and in concert to represent the effects
10 of exposure to PM.
11
12
13 8.2 RESPIRATORY EFFECTS OF PARTICULATE MATTER IN
14 HEALTHY HUMANS AND LABORATORY ANIMALS: IN VIVO
15 EXPOSURES
16 The following sections assess the results of human exposure to various types of PM and
17 also discuss controlled animal toxicology studies, as well as in vitro studies using animal or
18 human respiratory cells. The discussion focuses on those studies published since the previous
19 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a).
20 The biological responses occurring in the respiratory tract following controlled PM
21 inhalation encompass a continuum of changes, including changes in pulmonary inflammation,
22 pulmonary function, and systemic effects. The observed responses are dependent on the
23 physicochemical characteristics of the PM, the total exposure, and the health status of the host.
24 Many of the responses are usually seen only at higher level exposures characteristic of
25 occupational and laboratory animal studies and not at (typically much lower) ambient particle
26 concentrations; however, there are substantial differences in the inhalability and deposition
27 profiles of PM in humans and rodents (see Chapter 7 for details). Observed responses and
28 dose-response relationships also are very dependant on the variable being measured.
29 Paniculate matter is a broad term that encompasses thousands of chemical species, many
30 of which have not been investigated in controlled laboratory animal or human studies. However,
31 a full discussion of all types of particles that have been studied is beyond the scope of this
March 2001 8-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 chapter. Thus, specific criteria were used to select topics for presentation. High priority was
2 placed on studies that may (1) elucidate health effects of major common constituents of ambient
3 PM or (2) contribute to enhanced understanding of the epidemiological studies (e.g., use of
4 ambient particles, "surrogate" particles, or particles with low inherent toxicity that may cause
5 effects because of their physicochemical characteristics, such as their size and composition).
6 Most studies, therefore, have been designed to address the question of biologic plausibility,
7 rather than providing dose-response or risk assessment quantitation.
8 Diesel exhaust particles (DPM) generally fit the criteria; but, because they are described
9 elsewhere in great detail (U. S. Environmental Protection Agency, 1999; Health Effects Institute,
10 1995), they are not covered extensively in this chapter except in the discussions of their
11 immunological effects. Particles with high inherent toxicity, such as silica and asbestos, that are
12 of concern primarily because of occupational exposure, also are excluded from this chapter and
13 are discussed in detail elsewhere (U.S. Environmental Protection Agency, 1996b; Gift and Faust,
14 1997). Most of the laboratory animal studies summarized here have used high particulate mass
15 concentrations administered by inhalation, compared to ambient levels, even when laboratory
16 animal-to-human dosimetric differences or high doses by intratracheal instillation are considered.
17 More research on particle dose extrapolation is needed, therefore, to determine species
18 differences and the importance of exercise and other factors influencing particle deposition in
19 humans that together can account for a 50-fold or more difference in dose.
20 As mentioned earlier, the data available in the previous 1996 PM AQCD were from studies
21 that investigated the respiratory effects of specific components of ambient PM or surrogate
22 particles. More recently, pulmonary effects of controlled exposures to ambient PM have been
23 investigated by the use of aerosol concentrators (Sioutas et al., 1995; Gordon et al., 1998). These
24 concentrators are capable of exposing animals or humans to PM concentrations that are up to
25 90-fold higher than ambient PM levels and have been used to investigate the effects of ambient
26 PM in normal and compromised animals and humans.
27
28 8.2.1 Acid Aerosols
29 There have been extensive studies of the effects of controlled exposures to aqueous acid
30 aerosols on various aspects of lung function in humans and laboratory animals. Many of these
31 studies were reviewed in the previous criteria document (U.S. Environmental Protection Agency
March 2001 8-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 1996a) and in the Acid Aerosol Issue Paper (U.S. Environmental Protection Agency, 1989); more
2 recent studies are summarized in the present document (see Table 8-1). Methodology and
3 measurement methods for controlled human exposure studies have been reviewed elsewhere
4 (Folinsbeeetal., 1997).
5 These studies illustrate that aqueous acidic aerosols have minimal effects on symptoms and
6 mechanical lung function in young healthy adult volunteers at concentrations as high as
7 2000 Aig/m3. The findings include minimal changes in lung function accompanied by only mild
8 lower respiratory symptoms. However at concentrations as low as 100 /wg/m3, acid aerosols can
9 alter mucociliary clearance. Brief exposures (< 1 h) to low concentrations (=100 /wg/m3) may
10 accelerate clearance while longer (multihour) exposures to higher concentrations (>100 /ug/rn3)
11 can depress clearance. Asthmatic subjects appear to be more sensitive to the effects of acidic
12 aerosols on mechanical lung function. Responses have been reported in adolescent asthmatics at
13 concentrations as low as 68 /wg/m3 and modest bronchoconstriction has been seen in adult
14 asthmatics exposed to concentrations >400 //g/m3, but the available data are not consistent.
15 A previously described acid aerosol exposure in humans (1000 Mg/m3) did not result in
16 airway inflammation (Frampton et al., 1992), and there was no evidence of altered macrophage
17 host defenses. More recently, Zelikoff et al. (1997) compared the responses of rabbits and
18 humans exposed to similar concentrations of acid aerosol. For both rabbits and humans, there
19 was no evidence of PMN infiltration into the lung and no change in BAL protein level, although
20 there was an increase in LDH in rabbits but not in humans. Macrophages showed less
21 antimicrobial activity in rabbits; insufficient data were available for humans. Macrophage
22 phagocytic activity was slightly reduced in rabbits but not in humans. Superoxide production by
23 macrophages was somewhat depressed in both species. No respiratory effects of long-term
24 exposure to acid aerosol were found in dogs (Heyder et al., 1999).
25
26 8.2.2 Metal Particles, Fumes, and Smoke
27 Data from occupational and laboratory animal studies reviewed in the previous criteria
28 document (U. S. Environmental Protection Agency, 1996a) indicated that acute exposures to very
29 high levels (hundreds of /ug/m3 or more) or chronic exposures to lower levels (up to 15 /ug/m3,
30 albeit high compared to ambient levels) of metallic particles could have an effect on the
31 respiratory tract. However, it was concluded on the basis of available data that the metals at
March 2001 8-4 DRAFT-DO NOT QUOTE OR CITE
-------
§
tr
O
O
TABLE 8-1. RESPIRATORY EFFECTS OF ACID AEROSOLS IN HUMANS AND LABORATORY ANIMALS
Species, Gender, Strain
Age, etc.
Particle
Exposure
Technique
Concentration
Particle Size
Exposure
Duration
Effects of Particles
Reference
oo
T1
H
6
o
o
H
O
c
o
Healthy beagle dogs; Neutral sulfite Inhalation 1.5 mg/m3
n = 16 aerosol
1.0/^mMMAD 165h/day Long-term exposure to particle-associated sulfur Heyderet al (1999)
6g=22 for 13 mo and hydrogen ions at concentrations close to
ambient levels caused only subtle respiratory
responses and no change in lung pathology.
Asthmatic subjects;
13 M, 11 F
Female Fischer 344 rats
Female Hartley Guinea
Pigs
Healthy nonsmokers;
10M, 2 F, 2 1-37 years
old
Acidic sulfate
aerosol
H,SO4 aerosol
NHVSO'5
aerosol
H2SO4 aerosol
H,SO4 aerosol
Inhalation 5 7 mg/m3
Inhalation by 500 ^g/m3
face mask
Inhalation 94 mg/m3
43 mg/m3
Inhalation 1 ,000 f^g/m3
1.1 jumMMAD
6g = 2.0
9 nm MMAD
7 urn MMAD
0.80 ± 1 89 6g
0.93 ±2 11 8g
0.8-0.9 nm
MMAD
6 h/day for
13 mo
1 h
4h
3h
Exposure to simulated natural acid fog did not
induce bronchoconstnction and did not change
bronchial responsiveness in asthmatics
Acid aerosol increased surfactant film
compressibility in guinea pigs
No inflammatory responses, LDH activity in BAL
was elevated Effect on bacterial killing by
macrophages was inconclusive; latex particle
phagocytosis was reduced 28%
Leducetai (1995)
Leeetal (1999)
Zehkoffetal. (1997)
BAL - Bronchoalveolar lavage
LDH - Lactate dehydrogenase
MMAD - Mass median aerodynamic diameter
MMD - Mass median diameter
5g - Geometnc standard deviation
n
t— H
H
W
-------
1 concentrations present in the ambient atmosphere (1 to 14 /ug/m3) were not likely to have a
2 significant acute effect in healthy individuals. These metals include arsenic, cadmium, copper,
3 vanadium, iron, and zinc. Other metals found at concentrations less than 0.5 //g/m3 were not
4 reviewed in the previous criteria document. However, published data added to the existing PM
5 data base demonstrate that particle-associated metals are plausible causal components of PM.
6 Only limited controlled human exposure studies have been performed with particles other
7 than acid aerosols (see Table 8-2). Controlled inhalation exposure studies to high concentrations
8 of two different metal fumes, MgO and ZnO, demonstrate the differences in response based on
9 particle metal composition (Kuschner et al., 1997). Up to 6400 mg/m3* min cumulative dose of
10 MgO had no effect on lung function (spirometry, DLCO), symptoms of metal fume fever, or
11 changes in inflammatory mediators or cells recovered by BAL. However, lower concentrations
12 of ZnO fume (165 to 1110 mg/m3* min) induced a neutrophilic inflammatory response in the
13 airways 20 h postexposure. Lavage fluid PMNs, TNF-cc, and IL-8 were increased by ZnO
14 exposure. However, the concentrations used in these exposure studies exceed ambient levels by
15 more than 1000-fold. The absence of a response to an almost 10-fold higher concentration of
16 MgO compared with ZnO indicates that metal composition may be more important than particle
17 size (ultrafine/fine) when considering observed health responses to inhaled PM. Fine et al.
18 (1997) have shown elevated body temperature (metal fume fever) and increased levels of plasma
19 IL-6 (from 2.9 to 6.4 pg/mL) in naive subjects exposed to the 8-h TLV concentration of ZnO of
20 5 mg/m3 for 2 h.
21 Several metals have been shown to stimulate cytokine release in cultured human pulmonary
22 cells including zinc, chromium, cobalt, and vanadium. Boiler makers, exposed occupationally to
23 approximately 400 to 500 yUg/m3 of fuel oil ash, showed acute nasal inflammatory responses
24 characterized by increased PMNs and elevated IL-8 that were associated with vanadium levels
25 (increased about ninefold) in the upper airway (Woodin et al., 1998). Irsigler et al. (1999)
26 reported that V2O5 can induce asthma and bronchial hyperreactiviry in exposed workers.
27 A comparison of autopsy cases in Mexico City from the 1950s versus the 1980s indicated
28 substantially higher levels of (5- to 20-fold) Cd, Co, Cu, Ni, and Pb in lung tissue from the 1980s
29 (Fortoul et al., 1996). Similar studies have examined metal content in human blood and lung
30 tissue (Tsuchiyama et al., 1997; Osman et al., 1998). The autopsy data suggest that chronic
31 exposures to urban air pollution leads to an increased deposition of metals in human tissues.
March 2001 8-6 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 8-2. RESPIRATORY EFFECTS OF METAL PARTICLES, FUMES, AND SMOKE IN HUMANS AND
o
cr
to
O
Species, Gender,
Strain, Age, etc. Particle
Naive subjects; ZnO
8 M, 5 F; 30.8±
7.7 years old
Healthy Colloidal iron
nonsmokers; oxide
12 M, 4F; 18-35
years old
Vanadium plant V2OS
workers; 40 M,
19-60 years old
Exposure
Technique
Inhalation by
face mask
Bronchial
instillation
Ambient air
JLABUKAJH
Concentration Particle Size
2.5 mg/m3 0.3 Mm MMD
5 mg/m3
SmgmlOmL 2. 6 Mm
<0.05-1.53 N/A
mg/m3
JKY APMMA
Exposure
Duration
2h
1, 2, and 4 days
after instillation
Variable
LS
Effect of Particles
Increased oral temperature after 2.5- and 5.0-mg/m3
exposure. Elevated IL-6 after exposure to 5 mg/m3.
Symptoms (myalgia, cough, fatigue) peaked 9 h after
5 mg/m3 exposure.
L-femtm increased after iron oxide particle exposure;
transfemn was decreased Both lactoferrin and
transfemn receptors were increased.
1 2/40 workers had bronchial hyperreactivity that
persisted in some for up to 23 mo.
Reference
Fine etal. (1997)
Ohio et al
(1998a)
Irsigler et al.
(1999)
oo
Healthy MgO
nonsmokers; 4 M, ZnO
2 F; 21-43 years
old
Inhalation 100-200 mg/m3
99% < 1.8 Mm
29% < 0.1 ^m
45 mm
No significant differences in BAL inflammatory cell Kuschner et al.
concentrations, BAL mterleukms (IL-1, IL-6, IL-8), (1997)
tumor necrosis factor, pulmonary function, or peripheral
blood neutrophils.
O
5
>
T1
1
o
2,
O
H
0
G
O
H
W
O
Hrl
?3
O
H— 1
H
M
Healthy Fe,O3
nonsmokers;
27 M, 7 F, 20-36
years old
Fischer 344 rats Fe2O,
(250 g)
NMRI mice; MnO2
Mouse peritoneal
macrophage
7-week-old CdO Fume
WISTAR Furth
rats;
C57BL6 and
DBA3NCR mice
Rat, M, F344, TiO2
1 75-225 g
Intrapulmonary
instillation
Intratracheal
instillation
Intratracheal
instillation;
in vitro
Nose-only
Inhalation
Intratracheal
inhalation and
Intratrachea)
instillation
3x 10s
microspheres in
10 mL saline.
7.7 x 107
microspheres in
5 mL saline
0.037,0.12,
0.75,
2.5 mg/amma!
1 04 mg/m3
Rats dose =
18.72Mg
Mouse dose =
4.59 Mg
Inhalation at
125 Mg/m3
Instillation at
500 Mg for fine,
750 Mg for
ultrafine
2.6 Mm
2. 6 Mm
surface area of
0.16,05, 17,
62 m2/g
CMD = 0 008 Mm
8g= 1.1
Fine: 250 nm
Ultrafine: 2 1 nm
N/A
N/A
Sacrificed at
5 days
1 x3h
Inhalation
exposure, 2 h;
sacrificed at 0,
1, 3, and 7 days
postexposure for
both techniques
Transient inflammation induced initially (neutrophils,
protein, LDH, IL-8) was resolved by 4 days
postmstillation.
Transient inflammation at I day postmstillation.
LDH, protein and cellular recruitment increased with
increasing surface area; freshly ground particles had
enhanced cytotoxicity.
Mice created more metallothionem than rats, which may
be protective of tumor formation
Inflammation produced by mtratracheal inhalation (both
seventy and persistence) was less than that produced by
instillation; ultrafine particles produced greater
inflammatory response than fine particles for both dosing
methods.
Lay etal (1998)
Lay etal. (1998)
Lison et al.
(1997)
McKenna et al.
(1998)
Osier and
Oberdorster
(1997)
-------
fu
s*
oo
6
o
2
o
H
O
O
H
m
O
w
o
h—H
H
TABLE 8-2 (cont'd). RESPIRATORY EFFECTS OF METAL PARTICLES, FUMES, AND SMOKE IN HUMANS AND
LABORATORY ANIMALS
Species, Gender,
Strain, Age, etc. Particle
Rat, M. F344, TiO2
1 75-225 g
Rats NaVO,
VOSO,
VA
Boilermakers VA
(18 M), 26-61
years old, and
utility worker
controls (1 1 M),
30-55 years old
Exposure
Technique
Intratracheal
inhalation and
Intratracheal
instillation
Intratracheal
instillation
Inhalation of
fuel-oil ash
Concentration Particle Size
Inhalation at Fine. 250 nm
125,ug/m' Ultrafine: 21 nm
Instillation at
500 us for fine,
750 jj.% for
ultrafine
21or210//g N/A
V/kg (NaVOj,
VOSO, soluble)
42 or 420 ,ug
V/kg (V A) 'ess
soluble
0.4-0.47 mg/m3 lO^m
0 l-O.I3mg/m3
Exposure
Duration
Inhalation
exposure, 2 h;
sacrificed at 0,
1, 3, and 7 days
postexposure for
both techniques
1 h or 1 0 days
following
instillation
6 weeks
Effect of Particles
MIP-2 increased m lavage cells but not in supernatant m
those groups with increased PMN (more in instillation
than in inhalation, more in ultrafine than in fine); TNF-ct
levels had no correlation with either particle size or
dosing methods
PMN influx was greatest following VOSO4, lowest for
VA, VOSO4 induced inflammation persisted longest;
MIP-2 and KC (CXC chemokmes) were rapidly induced
as early as I h postmstillation and persisted for 48 h;
Soluble V induced greater chemokine mRNA expression
than insoluble V; AMs have the highest expression level
Exposure to fuel-oil ash resulted in acute upper airway
inflammation, possibly mediated by increased IL-8 and
PMNs.
Reference
Osier et al.
(1997)
Pierce et al.
(1996)
Woodm et al.
(1998)
BAL - Bronchoalveolar lavage
CMD - Count median diameter
1L - Interleukin
LDH - Lactate dehydrogenase
MIP-2 - Macrophage inflammatory protem-2
mRNA - Messenger RNA (nbonucleic acid)
N/A - Data not available
-------
1 Iron is the most abundant of the elements that are capable of catalyzing oxidant generation
2 and also present in ambient urban particles. Lay et al. (1998) and Ohio et al. (1998a) tested the
3 hypothesis that the human respiratory tract will attempt to diminish the added, iron-generated
4 oxidative stress. They examined the cellular and biochemical response of human subjects
5 instilled with iron (III) oxide via the intrapulmonary route. Saline alone and iron-containing
6 particles suspended in saline were instilled into separate lung segments of human subjects.
7 Subjects underwent bronchoalveolar lavage at 1 to 91 days after instillation of 2.6-yUm diameter
8 iron oxide agglomerates. Lay and colleagues found the greatest iron oxide-induced inflammatory
9 response in the alveolar fraction of the lavage fluid, although a significant increase in
10 macrophages also was observed in the bronchial fraction. The peak response for all cellular and
11 biochemical changes occurred at 1 day postinstillation. Lung lavage within 1 day of exposure
12 revealed decreased transferrin concentrations and increased ferritin and lactoferrin
13 concentrations, consistent with a host-generated decrease in the availability of catalytically
14 reactive iron (Ohio et al., 1998a). Normal iron homeostasis returned within 4 days of the iron
15 particle instillation. The same iron oxide preparation, which contained a small amount of soluble
16 iron, produced similar pulmonary changes in rats. Instillation of rats with two iron oxide
17 preparations that contained no soluble iron did not produce injury or inflammation (see Section
18 8.2.2), thus suggesting that soluble iron was responsible for the observed intrapulmonary
19 changes. Although only a small amount of the iron instilled in human subjects was "active", it is
20 not clear if the total dose of iron oxide delivered acutely to the lung segments (approximately
21 5 mg or 2.1 x 10s particles) would be relevant to deposition of iron oxide particles at the
22 concentrations of iron present in ambient urban air (generally less than 1 yUg/m3).
23
24 8.2.3 Ambient Combustion-Related and Surrogate Particulate Matter
25 The majority of the in vivo exposures to ambient particles have utilized intratracheal
26 instillation techniques in laboratory animals. Discussions on the pros and cons of this technique
27 are covered in Chapter 7 (Section 7.5), and these issues have also been reviewed elsewhere
28 (Driscoll et al., 2000; Oberdorster et al., 1997; Osier and Oberdorster, 1997). In most of these
29 studies, PM samples were collected on filters, resuspended in a vehicle (usually saline), and a
30 small volume of the suspension was instilled intratracheally into the animals. The
31 physiochemcial characteristics of PM are altered by deposition on a filter and resuspension in an
March 2001 8-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 aqueous medium. In addition, the doses used in these instillation studies are generally high
2 compared to ambient concentrations, even when laboratory animal-to-human dosimetric
3 differences are considered. Therefore, in terms of direct extrapolation to humans in ambient
4 exposure scenarios, greater importance should be place on inhalation studies. However, delivery
5 of PM by instillation has the advantage that much less material is needed and the delivered dose
6 can be determined directly without extrapolating from estimates of lung deposition. Instillation
7 studies have proven valuable in comparing the effects of different types of PM and for
8 investigating the mechanisms by which particulates cause lung injury and inflammation.
9 Tables 8-3, 8-4, and 8-5 outline studies in which various biological endpoints were measured
10 following exposures to ambient PM, complex combustion-related PM, or laboratory-derived
11 surrogate PM, respectively.
12 There were only limited data available from human studies or laboratory animal studies on
13 ultrafme aerosols at the time of the release of the previous criteria document (U.S. Environmental
14 Protection Agency, 1996a). In vitro studies have shown that ultrafme particles have the capacity
15 to cause injury to cells of the respiratory tract. High levels of ultrafme particles, as metal or
16 polymer "fume," are associated with toxic respiratory responses in humans and other mammals.
17 Such exposures are associated with cough, dyspnea, pulmonary edema, and acute inflammation.
18 At concentrations less than 50 ywg/m3, freshly generated insoluble ultrafme teflon polymer fume
19 particles can be severely toxic to the lung. However, it was not clear what role in the observed
20 effects was played by fume gases which adhered to the particles. Thus, it was not clear at the
21 time of the previous review what role, if any, ambient ultrafine particles may play in PM-induced
22 mortality and morbidity. Newer data from clinical exposures have demonstrated that
23 composition and not particle size was responsible for the adverse health effects associated with
24 exposures to metal fumes containing both ultrafine and fine particles (Kuschner et al., 1997).
25 Toxicologic studies of other particulate matter species also were discussed in the previous
26 criteria document. These studies included exposures to fly ash, volcanic ash, coal dust, carbon
27 black, TiO2, and miscellaneous other particles, either alone or in mixture. Some of the particles
28 discussed were considered to be models of "nuisance" or "inert" dusts (i.e., those having low
29 intrinsic toxicity) and were used in instillation studies to delineate nonspecific particle effects
30 from effects of known toxicants. A number of studies on "other PM" examined effects of up to
31 50,000 /ug/m3 of respirable particles with inherently low toxicity. Although there was no
March 2001 8-10 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 8-3. RESPIRATORY EFFECTS OF AMBIENT PARTICULATE MATTER
3
tr
o
o
Species, Gender,
Strain, Age, etc
Male S-D rats
200-225 g,
control and
SCytreated
S-D rats
60 days
Particle
Concentrated
ambient particles
(Boston) (CAP)
Provo, UT,
TSP filters
(10 years old),
soluble and
insoluble extracts
Exposure
Technique
Harvard/EPA
fine particle
concentrator
animals restrained
in chamber
Intratracheal
instillation
Concentration Particle Size Exposure Duration
206,733, 607 vg/m3 for 0 1 8 ^m 5 h/day for 3 days
Days 1-3; 29 °C, 6g = 2.9
59% RH
100-1000/^gofPM N/A 24 h
extract in 0.5 mL saline
Effect of Particles
PEF and TV increased in CAPS exposed animals.
Increased protein and % neutrophlls and
lymphocytes in lavage fluid after CAPS exposure.
Responses were greater in SO2-bronchitis animals.
No changes in LDH. No deaths occurred.
Inflammation (PMN) and lavage fluid protein was
greater with the soluble fraction containing more
metal (Zn, Fe, Cu).
Reference
Clarke et al.
(1999)
Ohio et al
(1999a)
oo
Healthy
nonsmokers;
18 to 40 yr old
Mongral dogs,
some with balloon
occluded LAD
coronary artery
n= 14
CAP
CAP
Male F 344 rats,
monocrotahne
treated
CAP
Inhalation 23.1 to 311.1 ^g/m3 0.65 ^m 2 h; analysis at 18 h Increased BAL neutrophils m both bronchial and Ghioetal
6g = 2 35 alveolar fractions (2000a)
Inhalation via 69-828 ,ug/m3 0 23 to 0 34 ^m 6 h/day * 3 days Decreased respiratory rate and increased lavage Godleski et al.
tracheostomy 6g = 0 2 to 2 9 fluid neutrophlls in normal dogs (2000)
Inhalation 132 to 919 /^g/m3 0.2 to 1.2 ,um I*3hor3x6h No inflammatory responses, no cell damage Gordon et al.
6g = 0 2 to 3 9 responses, no PFT changes. (2000)
H
b
o
2
o
H
O
M
O
?o
o
HH
H
m
8-mo-old Bi TO-2
male hamsters
Rats
CAP
Nose-only
inhalation
S-D rats TSP collected in Intratracheal
Human Bronchial Provo instillation
Epithelial
(BEAS-2B) cells
11 0-350 Mg/m3
TSP filter samples (36 5
mg/mL) agitated m
deionized H,O2 for 96 h,
centnfuged at 1200g for
30 mm, lyophyhzed and
resuspended m deionized
H202 or saline
N/A 3 h Increased penpheral blood neutrophlls and Gordon et al.
decreased lymphocytes. (1998)
N/A (TSP samples, Sacrificed at 24 h Provo particles caused cytokme-mduced Kennedy et al.
comprised 50 to neutrophil-chemoattractant-dependent (1998)
60% PMK)) inflammation of rat lungs; Provo particles
stimulated IL-6 and IL-8 production, increased
IL-8 mRNA and ICAM-1 in BEAS-2B cells, and
stimulated IL-8 secretion m primary cultures of
BEAS-2B cells; cytokine secretion was preceded
by activation of NF-KB and was reduced by SOD,
DEF, or NAC; quantities of Cu2+ found in Provo
particles replicated the effects
-------
|
|
H
NT PARTICULA
S
5
0
i/5
H
U
U
fa
W
X
0
H
K
^^
PH
fT1
.
?
C
O
u
f)
00
td
-1
ca
•^
H
o
CD
W
"o
Effect of Part
« c
3 2
|1
w
S
c/5
JD
£
a.
_.
0
2
*^
g
g
0
U
S S
3 CT*
o c
X gj
ju
«
Q-
w"
*O 4)
W u"
O w>
|.s
8.S
t« VI
cd
w
c
ra ''reT
iS §.
s |
l|
h; (S
;ased BALF protein and neul
chitic rats; responses were \;
isure regimens.
lit
x
tt
IS 17
vo <~J
m
I
O
sO
.2
'•^
—
w
£
cu
U
rt -*-•
>-Q g_
-S . -£ ^
o 2 1 gj
p^
0\
~-
g;
c.
ra
^
"5
'1 1 1 1
X m » t-o
;ased PMN, protein, and LD
ter response with ultrafme C
eased GSH level in BAL; fre
lete supercoil DNA), leukoc:
lals produced greater NO am
C a g a. c
£ ob-o 2- §
M
•a
o
0 j-
t/2 ^0
E
s?s
=> II c cQ
SmgS
S 0 S D
c
00
CN) C
"T
0 ^
iri o
S1""*
c
-c o
03 -^
*3 — -"
2 s
i: vi
£ S
oa"
S u u
&• -fi >~
§ 3 ^ - — •
— -^ *2 CO
•^ PJ O ^— '
~^r
X
!/)
It
f* t8
March 2001
8-12
DRAFT-DO NOT QUOTE OR CITE
-------
ta
§•
K>
o
o
TABLE 8-4. RESPIRATORY EFFECTS OF COMPLEX COMBUSTION-RELATED PARTICULATE MATTER
OO
I
U)
H
6
o
2
o
H
O
d
o
H
W
Species, Gender, Strain,
Age, etc.
M Syrian golden
hamsters 90- 125g
NMRI mouse
Rats, male, S-D, 60 days
old
MCT (60 mg/kg), ip
Exposure
Particle Technique
Kuwaiti oil fire Intratracheal
particles instillation
Urban particles
from St. Louis,
MO
CFA Intratracheal
CMP instillation
WC
Emission Intratracheal
source PM instillation
Ambient
airshed PM
ROFA
Concentration
0.15, 0.75, and 3.75
mg/lOOg
CMP: 20 //g
arsenic/kg, or CMP:
lOOmg
particles/kg,
WC alone
(100 mg/kg), CFA
alone (100 mg/kg
[i.e., 20 Mg
arsenic/kg]), CMP
mixed with WC
(CMP, 13.6 mg/kg
[(i.e., 20 ng
arsenic/kg]), WC
(86.4 mg/kg) and
Ca3(AsO4)2 mixed
with WC (20 ^g
arsenic/kg), WC
(100 mg/kg)
Total mass:
2.5 mg/rat
Total transition
metal: 46 ^g/rat
Particle Size
Oil fire particles:
<3.5 /J.m, 10 days of
24-h samples (April
30 to May 9, 1991),
in Ahmadi, Kuwait
N/A
Emission PM:
1 78-4.1 7 ^m
Ambient PM-
3.27-4 09 ^m
Exposure
Duration
Sacrificed 1 and
7 days post
instillation
1,5, 30 days
posttreatment,
lavage for total
protein content,
inflammatory
cell number and
type, and TNF-
a production
particle
retention
Analysis at 24
and 96 h
following
instillation
Effect of Particles
Increases in PMN, AM, albumin, LDH,
myeloperoxidase, and
(3-N-acetylglucosaminidase;
acute toxicity of the particles found in the
smoke from the Kuwaiti oil fires is
comparable to that of urban particles.
Mild inflammation for WC; Ca3(AsO,)2
caused significant inflammation;
CMP caused severe but transient
inflammation; CFA caused persistent
alveolitis; cytokine production was
upregulated in WC- and Ca3(AsO4) treated
animals after 6 and 30 days, respectively;
a 90% inhibition of TNF-a production still
was still observed at Day 30 after
administration of CMP and CFA;
a significant fraction persisted (10-15% of
the arsenic administered) in the lung of
CMP- and CFA-treated mice at Day 30
Suppression of TNF-a production is
dependent on the slow elimination of the
particles and their metal content from the
lung
Increases in PMNs, albumin, LDH, PMN,
and eosmophils following exposure to
emission and ambient particles;
induction of injury by emission and
ambient PM samples is determined
primarily by constituent metals and their
bioavailability;
MCT-ROFA show enhanced neutrophihc
inflammation and an increase in mortality.
Reference
Brain et al. (1998)
Broeckaert et al.
(1997)
Costa and Dreher
(1997)
WISTAR male rats
Bor: WISW strain
Coal oil fly ash Inhalation
(chamber)
0,11,32, and 1.9-2.6 ^m 6 h/day, At the highest concentration, type II cell
103mg/m3 6g=1.6-18 5days/week, proliferation and mild fibrosis occurred and
4 weeks increased penvascular lymphocytes were
seen. The mam changes at the lowest
concentration were particle accumulation in
AM and mediastmal lymph nodes.
Lymphoid hyperplasia observed at all
concentrations. Effects increased with
exposure duration
Dormans et al.
(1999)
-------
TABLE 8-4 (cont'd). RESPIRATORY EFFECTS OF COMPLEX COMBUSTION-RELATED PARTICULATE MATTER
g.
K>
O
O
oo
*•
a
j>
3
>— j
i
o
o
o
H
O
c
w
n
NV/
o
H
m
Species, Gender, Strain,
Age, etc.
Rat, male, S-D, 60 days
old
Male S-D rats 60 days
old
5-day-old male S-D rats
Rat, male, S-D, 60 days
old
Female
Balb/cJ
mice 7- 1 5 weeks
7-week-oId Female
Balb/cJ mice (16-21 g)
Rat, male, S-D
Mice, normal and Hp,
105 days old
2-day-old BALB/C mice
sensitized to ovalbumin
(OVA)
Particle
ROFA
#6 ROFA,
volcanic ash
LowS
#6 ROFA,
volcanic ash
saline
Two ROFA
samples
Rl had 2*
salme-
leachable
sulfate, Ni, and
V and 40* Fe
as R2; R2 had
3 1 * higher Zn
ROFA
ROFA lo-S
residual oil
ROFA
ROFA
Aerosolized
ROFA leachate
Exposure
Technique
Intratracheal
instillation
Instillation in NaCl
solution
Intratracheal
Intratracheal
instillation
Intratracheal
instillation
Inhalation and
instillation
challenge with
OVA
Intratracheal
instillation
Intratracheal
instillation
Nose-only
inhalation
Concentration
8 33 mg/mL
0.3 mL/rat
03, 1.7
8 3 mg/mL
8 3 mg/mL
0 3, 1 7,
8 3 mg/mL
in saline
8 3 mg/kg BW
1 mL/kg BW
2 5 mg in 0 3 mL
60 Mg in 50 ML
(dose 3mg/kg)
1 58 ± 3 mg/m1
500 Mg/ammal
50 Mg
50 mg/mL
Particle Size
1 95 Mm MMAD
1 .95 Mm
6g = 2.19
1 4 Mm
1 .95 Mm
6g = 1 95
1 4 Mm
Rl: 1 88 Mm,
MMAD
R2. 2.03 Mm,
MMAD
<25
PM,5 sample
3. 6 Mm
1 95 Mm
Exposure
Duration
Analysis at
24 and 96 h
24 h
24 h
Analysis at
4 days
N/A
1,3,8, 15 days
after
instillation
Analyzed
4 and 96 h
postexposure
Analysis at
24 h
30 mm
Effect of Particles
Increased PMNs, protein, LDH at both time
points.
Plasma flbnnogen elevated after ROFA
instillation but not volcanic ash
Increased WBC count in ROFA-exposed rats
plasma flbnnogen increased 86% in ROFA
rats at highest concentration.
Four of the 24 animals treated with R2 or R2s
(supernatant) died; none in R 1 s treated
animals; more AM, PMN, eosmophils protein,
and LDH in R2 and R2s animals; more focal
alveolar lesions, thickened alveolar septae,
hyperplasia of type II cells, alveolar fibrosis in
R2 and R2s animals; baseline pulmonary
function and airway hyperreactwity were
worse in R2 and R2s groups
ROFA caused increases in eosmophils, IL-4
and IL-5 and airway responsiveness in
ovalbumm-sensitized and challenged mice.
Increased BAL protein and LDH at 1 and
3 days but not at 1 5 days postexposure.
Combined OVA and ROFA challenge
increased all damage markers and enhanced
allergen sensitization. Increased methacholme
response after ROFA.
Femtm and transfemn were elevated, greatest
increase in femtm, lactofemn, transfemn
occurred 24 h postexposure.
Diminished lung injury (e g., decreased lavage
fluid ascorbate, protein, lactate
dehydrogenase, inflammatory cells, cytokmes)
in Hp mice lacking transfemn; associated
with increased metal storage and transport
proteins
Increased airway response to methylchohne
and to OVA in ROFA exposed mice,
increased airway inflammation also.
Reference
Dreheretal. (1997)
Gardner et al. (2000)
Gardner et al. (2000)
Gavettetal (1997)
Gavettetal (1999)
Gavettetal (1999)
Ghioetal. (1998b)
Ghioetal (2000b)
Hamadaetal (1999)
-------
s
to
3
cr
K)
O
O
TABLE 8-4 (cont'd). RESPIRATORY EFFECTS OF COMPLEX COMBUSTION-RELATED PARTICULATE MATTER
oo
H
6
o
2
o
H
0
G
O
H
M
O
5*
n
H
Species, Gender, Strain,
Age, etc. Particle
Rat, male, S-D, ROFA
60 days old
Rats FOFA
MCT
Exposure
Technique
Intratracheal
instillation
Inhalation
Concentration
1 0 mg in 0.5 mL
saline
580 ± 110^g/m3
Particle Size
1 .95 nm
2.06 //m MMAD
8g=157
Exposure
Duration
Analysis at
24 h
6 h/day for
3 days
Effect of Particles
Increased PMNs, protein.
Death occurred only in MCT rats exposed to
ROFA. Neutrophils in lavage fluid were
increased significantly in MCT rats exposed
to ROFA versus filtered air. MIP-2 mRNA
expression in lavage cells was induced in
normal animals exposed to fly ash.
Reference
Kadiiska et al.
(1997)
Killingsworth et al.
(1997)
Male S-D and F-344 rats ROFA
(60 days old)
Male S-D, WIS, and
F-344 rats (60 days old)
ROFA
Intratracheal 8.3 mg/kg 1 95 /j.m Sacrificed at Increase in neutrophils in both S-D and F-344
instillation 6g = 2.14 24 h rats; a time-dependent increase in eosmophils
occurred in S-D rats but not in F-344 rats.
Intratracheal 8 3 mg/kg 1.95 t*m Sacrificed at 6, Inflammatory cell infiltration, as well as
instillation 6g = 2.14 24,48, and alveolar, airway, and interstitial thickening in
72 h; 1,3, and all three rat strains; a sporadic incidence of
12 weeks focal alveolar fibrosis in S-D rats, but not in
WIS and F-344 rats; cellular fibronectm (cFn)
mRNA isoforms EIIIA(+) were up-regulated
in S-D and WIS rats but not in F-344 rats. Fn
mRNA expression by macrophage and
alveolar and airway epithelium and within
fibrotic areas m S-D rats; increased presence
of Fn EIIIA(+) protein in the areas of fibrotic
injury and basally to the airway epithelium.
Kodavanti et al.
(1996)
Kodavanti et al
(1997a)
Male S-D Rats,
60 days old
Male S-D rats,
60 days old
ROFA Intratracheal
instillation
Fe2(S04)3,
VSO4,
NiSO4
10 ROFA Intratracheal
compositionally instillation
different
particles from a
Boston power
plant
8.33 mg/kg 1.95^m
6g = 2.14
ROFA-equivalent
dose of metals
0.833,333,8.3 1. 99-2.59 laa
mg/kg MMAD
Analysis at 3,
24, and 96 h,
postmstillation
Sacrificed at
24 h
ROFA-induced pathology lesions were as
severe as those caused by Ni. Metal mixture
caused less injury than ROFA or Ni alone; Fe
was less pathogenic. Cytokme and adhesion
molecule gene expression occurred as early as
3 h after exposure V-mduced gene expression
was transient but Ni caused persistent
expression and injury.
ROFA-induced increases in BAL protein and
LDH, but not PMN, were associated with
water-leachable total metal, Ni, Fe, and S;
BALF neutrophilic inflammation was
correlated with V but not Ni or S.
Chemiluminescence signals m vitro (AM)
were greatest with ROFA containing soluble
V and less with Ni plus V.
Kodavanti et al
(1997b)
Kodavanti et al.
(1998a)
-------
td
H
^
^?
W
H
ON-RELATED PARTICULA
H
OJ
^^
CQ
5
O
U
X
td
•J
o.
O
U
ts.
O
C/5
H
U
W
ta
fe
U
OS
o
H
S
NH
0.
!/3
S
^"H
a
^
"^^
G
O
Q
^
X
N
Cfl
H
u
c
3J
(H
&
Effect of Particles
1 I
x 5
uj D
u
N
"2
•^
£
£
c
o
S
Concent
S 3
§ 5
tl
£
-H
rt
a.
c"
i
C/3
i."
•o
C
G
S s
«
4J
i_
w c^
1§
Both IT and IN rats showed inflammatory
responses (1L-6, MIP-2, inflammatory cells,
etc.). 58% of IT rats exposed to ROFA died
within 96 h. No mortality occurred in the IN
rats. ROFA exacerbated lung lesions
(edema, inflammatory cells, alveolar
thickening) in MCT rats.
Z
L.~JS
^ 1 —
S 5 i?Q
^ ^- ^j
fN ^O m 03
^.
E — •
in II
°! oo
— . ^
x: 5 ^ §
2 Js 9 "S
S 3 g j3
£ £ Z £
.
u.
O
.
— U
rt ^
111?
V A S
T Q « Q
O i P i^
«
^
—
Iff
« 2
•gs
£ ci
Increased BALF protein and LDH alveotis
with macrophage accumulation m alveoli;
increased neutrophils in BAL. Increased
S
„ cfl
a1 Z
~° -3 j=
2 § s
•a-
E -
m II
o^ M
— «o
si
oo ~oi)
o £
S™
c
IS
2.2
2 3
£ ^
^
O
a!.
,
S
l-
tf\
•a 3
c o
M 2
"c5 j_
pulmonary protein leakage and inflammation
in SH rats. Effects of metal constituents of
ROFA were strain specific, vanadium caused
pulmonary injury only in SH rats; nickel was
toxic in both SH and WKY rats.
00
"o
~~ '
o
- 4
£°s
> Z S
"ra
a
!„
M """
|I
More puhnonary injury in SH rats. Increased
RBCsm BALF of SH rats. ROFA increased
airway reactivity to Ach in both SH and
WKY rats. Increased protein, albumin, and
LDH in BALF after ROFA exposure
(SH>WKY). Increased oxidative stress in
SH rats. SH rats failed to increase
glutathlone.
& 0
X «
a Z
^ c °c
^.
£ ~
m II
<*! oo
— <0
"E
E
in
~~"
~e .3
o •£
+
u,
O
„
1fl
•-
£fi
T3 2
ni o
^" "S
^ u
i?
Ji -
"(5 '
s -
^^
ON
"rt
4J
•g
_C
1
ROFA enhanced the response to house dust
mite (HDM) antigen challenge. Eosmophil
numbers, LDH, BAL protein, and IL-10 were
increased with ROFA + HDM versus HDM
alone.
„(.
Z
<
Z
OO 00
a. =i
§8
S n
11
2 "5
i: u5
_c c
.,
O
oi
15
£•
td
S
1
g
m
^
2
^r
"c3
tj
g
-o
T3
cd
Production of acetaldehyde increased at 2 h
postinstillation
0
e
1 .c
— (N
E
OO
O
-H
Ov
— •
U OJQ
ji e
|l-if
o^2 —
"S
1
ROFA increased PGE2 via cycloxygenase
expression
H
^j
O
G.
r-j
<
z
E
^
o
o
o
"*
S c
•5 2
si
2 ~
£ .S
^
tu
O
OL
tn
e
Q
or)
jj
1
2
3
1
O
0
o
o
CM
*e3
^,
(U
C
O
,0
No inflammation until 24 h.
^=
^"
o
— :
r-i
H
^4
§1
— V3
S c
"S 2
2 i
2 "5
J=5 H
^
O
"» S
Jj ^
i2
2
9
(/)
,u
1
"3
ra
-o
0
March 2001
8-16
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 8-4 (cont'd). RESPIRATORY EFFECTS OF COMPLEX COMBUSTION-RELATED PARTICULATE MATTER
o
to
I
-J
"fl
H
6
o
1
o
H
O
H
W
O
&
n
si
Species, Gender, Strain,
Age, etc.
Particle
Exposure
Technique
Concentration
Particle Size
Exposure
Duration
Effect of Particles
Reference
60-day-old male S-D rats Ottawa dust, Intratracheal Dose 0, 0.25, 1.95 /j.m 6 h/day for IT ROFA caused acute and dose-related Watkmson et al
and 60-day male ROFA, and instillation, nose- 1.0, and 3-day increase in pulmonary inflammation; no effect (2000)
Wistar-Kyoto rats, volcanic ash only inhalation 2.5 mg/rat inhalation; of volcanic ash
60-day-old male SH rats, instillation -
some cold-stressed, 96 h post-IT
some ozone-exposed,
some MCT-treated
-------
TABLE 8-5. RESPIRATORY EFFECTS OF SURROGATE PARTICULATE MATTER
1-1
o
O
O
oo
I
oo
H
6
o
2
o
H
O
O
H
W
O
?=»
O
H^
H
m
Species, Gender, Strain,
Age, etc
Syrian golden hamsters
900 M
900 F
C57B1/6J mice
Rat
Mice,C57BL/6J,
8 weeks and 8 mo old
MCT-treated S-D rats
Male S-D rat
(200g)
Male mice, 6-8 weeks
old (AJ, AKRJ, Sulfate,
B6C3F1), BALBcJ
strains raised in a
pathogen free laboratory
Swiss- Webster mice
Particle
Toner
TiO2
Silica
PTFE
TiO2
PTFE Fumes
PTFE Fumes
Fluorescent
microspheres
Diesel,
SiO2,
carbon black
Carbon black
Regal 660
SO4"
Exposure
Technique
Nose-only
inhalation
Inhalation
Whole body
inhalation
Whole body
inhalation
Inhalation
Intratracheal
instillation
Nose only
inhalation
Concentration
1 5, 6.0, or
24 mg/m3
40 mg/m3
3 mg/m3
PTFE:
1.25, 25, or
5x 105
particles/cc
TiO2-F 10 mg/m3
NiO. 5 mg/m3
Ni3S,: 0 5 mg/m3
1,2.5, or 5 x 10s
particles/cm3
1,2.5, or 5 * 10s
particles/cm3
3.85 ±0.81
Mg/1"3
1 mg in 0 4 mL.
10 Mg/m3
285 Mg/m3
Particle Size
4.0 Mm
1 1 Mm
1.4 Mm
PTFE: 18nm
TiCyF: 200 nm
TiO2-D. lOnm
18nm
18nm
1.0 Mm
1 38 ±0.10 Mm
DEP Collected as
TSP - disaggregated
in solution by
somcation (20 nm);
SiO2 (7 nm),
carbon black
0.29 Mm
±2. 7 Mm
Exposure
Duration
3,9, 15 mo
6h/day
5days/week
30 min or
6 h/day,
5days/week,
6 mo
1 5 mm,
analysis 4 h
postexposure
30-mm
exposure,
analysis 6 h
following
exposure
3 h/day
x 3 days
Sacrificed at
2,7,21,42,
and 84 days
postmstillation
4h
Effect of Particles
Retention increased with increased exposure.
Effects on the epithelium caused by direct
interactions with particles, not a result of
macrophage-derived mediators, and suggest
a more significant role in the overall
pulmonary response than previously
suspected; type II cell growth factor
production may be significant in the
pathogenesis of pulmonary fibrosis.
Increased PMN, mRNA of MnSOD and MT,
IL-la, IL-1P, IL-6, MIP-2, TNF-a mRNA of
MT and IL-6 expressed around all airways and
interstitial regions; PMN expressed IL-6, MT,
and TNF-a; AM and epithelial cells were
actively involved.
Increased PMN, lymphocytes, and protein
levels in old mice over young mice; increased
TNF-a mRNA in old mice over young mice;
no difference in LDH and P-Glucuronidase
Monocrotahne-treated animals contained
fewer microspheres in their macrophages,
probably because of impaired chemotaxis.
Amorphous SiO, increased permeability, and
neutrophilhc inflammation. Carbon black
and DEP translocated to mterstltum and
lymph nodes by 1 2 weeks.
Differences in inflammatory responses
(PMN) across strains Appears to be genetic
component to the susceptibility
Reference
Creutzenberg et al
(1998)
Fmkelstem et al.
(1997)
Johnston et al. (1996)
Johnston et al. (1998)
Madletal (1998)
Murphy et al. (1998)
Ohtsuka et al. 2000
-------
1 mortality, some mild pulmonary function changes after exposure to 5,000 to 10,000 /ug/m3 of
2 inert particles were observed in rats and guinea pigs. Lung morphology studies revealed focal
3 inflammatory responses, some epithelial hyperplasia, and fibrotic responses after exposure to
4 >5,000 /ug/m3. Changes in macrophage clearance after exposure to >10,000 /ug/m3 were
5 equivocal (no infectivity effects). In studies of mixtures of particles and other pollutants, effects
6 were variable depending on the toxicity of the associated pollutant, hi humans, co-exposure to
7 carbon particles appeared to increase responses to formaldehyde but not to acid aerosol. None of
8 the "other" particles mentioned above are present in ambient air in more than trace quantities.
9 Thus, it was concluded that the relevance of any of these studies to standard setting for ambient
10 PM may be extremely limited.
11 Recent studies that examined the acute effects of intratracheal instillation of ambient PM
12 have shown clearly that PM obtained from various sources can cause lung inflammation and
13 injury. Costa and Dreher (1997) showed that instillation of PM samples from three emission
14 sources (two oil and one coal fly ash) and four ambient airsheds (St Louis, MO; Washington,
15 DC; Dusseldorf, Germany; and Ottawa, Canada) resulted in increases in lung PMN and
16 eosinophils in rats 24 h after instillation. Biomarkers of permeability (total protein and albumin)
17 and cellular injury (LDH) also were increased. This study demonstrated that the lung dose of
18 bioavailable transition metal, not instilled PM mass, was the primary determinant of the acute
19 inflammatory response. Kennedy et al. (1998) reported a similar dose-dependent inflammation
20 (i.e., increase in protein and PMN in lavage fluid, proliferation of bronchiolar epithelium, and
21 intraalveolar hemorrhage) in rats instilled with water-extracted particles (TSP) collected in
22 Provo, UT. This study also indicated that the metal constituent, in this case PM-associated Cu,
23 was a plausible cause of the outcome. Likewise, instillation of ambient PM10 collected in
24 Edinburgh, Scotland, also caused pulmonary injury and inflammation in rats (Li et al., 1996,
25 1997). Brain et al. (1998) examined the effects of instillation of particles that resulted from the
26 Kuwaiti oil fires in 1991 compared to urban particulate matter collected in St. Louis (NITS SRM
27 1648, collected in a bag house in early 1980s) and showed that on an equal mass basis, the acute
28 toxicity of the Kuwaiti oil fire particles was similar to that of urban particles collected in the
29 United States.
30 The fact that instillation of ambient PM collected from different geographical areas and
31 from a variety of emission sources consistently caused pulmonary inflammation and injury tends
March 2001 8-19 DRAFT-DO NOT QUOTE OR CITE
-------
1 to corroborate epidemiological studies that report increased respiratory morbidity and mortality
2 associated with PM in many different geographical areas and climates. However, high-dose
3 instillation studies may produce different effects on the lung than inhalation exposures done at
4 more relevant concentrations. This concern is somewhat diminished by the results of an
5 inhalation study of concentrated PM in healthy nonsmokers. Ghio et al. (2000a) exposed 38
6 volunteers exercising intermittently at moderate levels of exertion for 2 h to either filtered air or
7 particles concentrated from the air in Chapel Hill, NC (23 to 311 ug/m2). Analysis of cells and
8 fluid obtained 18 h after exposure showed a mild increase in neutrophils in the bronchial and
9 alveolar fractions. No respiratory symptoms or decrements in pulmonary function were found
10 after exposure to CAP.
11 Because emission sources contribute to the overall ambient air particulate burden (Spengler
12 and Thurston, 1983), many studies investigating the response of laboratory animals to particle
13 exposures have used fly ash as a useful source of particle for exposure (see Table 8-3). The
14 residual oil fly ash (ROFA) samples used in toxicological studies have been collected from a
15 variety of sources such as boilers, bag houses used to control emissions from power plants, and
16 from the fine particles that are emitted downstream of the collection devices.
17 ROFA has a high content of water soluble sulfate and metals, accounting for 82 to 92% of
18 water-soluble mass, while the water-soluble mass fraction in ambient air varies from low teens to
19 more than 60% (Costa and Dreher, 1997; Prahalad et al., 1999). More than 90% of the metals in
20 ROFA are transition metals, whereas these metals are only a small subfraction of the total
21 ambient PM mass. Thus, the dose of bioavailable metal that is delivered to the lung when ROFA
22 is instilled into a laboratory animal can be orders of magnitude greater than a ambient PM dose,
23 even under a worst-case scenario.
24 Intratracheal instillation of various doses of ROFA suspension has been shown to produce
25 severe inflammation, an indicator of pulmonary injury that includes recruitment of neutrophils,
26 eosinophils, and monocytes into the airway. The biological effects of ROFA in rats have been
27 shown to depend on aqueous leachable chemical constituents of the particles (Dreher et al., 1997;
28 Kodavanti et al., 1997b). A leachate prepared from ROFA, containing predominantly Fe, Ni, V,
29 Ca, Mg, and sulfate, produced similar lung injury to that induced by the complete ROFA
30 suspension (Dreher et al., 1997). Depletion of Fe, Ni, and V from the ROFA leachate eliminated
31 its pulmonary toxicity. Correspondingly, minimal lung injury was observed in animals exposed
March 2001 8-20 DRAFT-DO NOT QUOTE OR CITE
-------
1 to saline-washed ROFA particles. A surrogate transition metal sulfate solution containing Fe, V,
2 and Ni largely reproduced the lung injury induced by ROFA. Interestingly, ferric sulfate and
3 vanadium sulfate antogonized the pulmonary toxicity of nickel sulfate. Interactions between
4 different metals and the acidity of PM were found to influence the severity and kinetics of lung
5 injury induced by ROFA and its soluble transition metals.
6 To further investigate the response to ROFA with differing metal and sulfate composition,
7 male Sprague Dawley rats (60 days old) were exposed to ROFA or metal sulfates (iron,
8 vanadium, and nickel, individually or in combination) (Kodavanti et al., 1997b). Transition
9 metal sulfate mixtures caused less injury than ROFA or Ni alone, suggesting metal interactions.
10 In addition, this study showed that V-induced effects were less severe than that of Ni and were
11 transient. Ferric sulfate was least pathogenic. Cytokine gene expression was induced prior to the
12 pathology changes in the lung and the kinetics of gene expression suggested persistent injury by
13 nickel sulfate. Another study by the same investigators was performed using 10 different ROFA
14 samples collected at various sites within a power plant burning residual oil firing chamber
15 (Kodavanti et al., 1998a). Animals received intratracheal instillations of either saline, or a saline
16 suspension of whole ROFA (<3.0 yum MMAD) at three concentrations (0.833, 3.33, or
17 8.33 mg/kg). This study showed that ROFA-induced PMN influx appeared to be associated with
18 its water-leachable V content; however, protein leakage appeared to be associated with water-
19 leachable Ni content. ROFA-induced in vitro activation of AM was highest with ROFA
20 containing leachable V but not with Ni plus V, suggesting that the potency and the mechanism of
21 pulmonary injury may differ between emissions containing bioavailable V and Ni.
22 Other studies have shown that soluble metal components play an important role in the
23 toxicity of emission source particles. Gavett et al. (1997) investigated the effects of two ROFA
24 samples of equivalent diameters, but having different metal and sulfate content, on pulmonary
25 responses in Sprague-Dawley rats. ROFA sample 1 (Rl) (the same emission particles used by
26 Dreher et al. [1997]) had approximately twice as much saline-leachable sulfate, nickel, and
27 vanadium, and 40 times as much iron as ROFA sample 2 (R2); whereas R2 had a 31 -fold higher
28 zinc content. Rats were instilled with suspensions of 2.5 mg R2 in 0.3 mL saline, the supernatant
29 of R2 (R2s), the supernatant of 2.5 mg Rl (Rls), or saline only. By 4 days after instillation, 4 of
30 24 rats treated with R2s or R2 had died. None of those treated with Rls or saline died.
31 Pathological indices, such as alveolitis, early fibrotic changes and perivascular edema, were
March 2001 8-21 DRAFT-DO NOT QUOTE OR CITE
-------
1 greater in both R2 groups, hi surviving rats, baseline pulmonary function parameters and airway
2 hyperreactivity to acetylcholine were significantly worse in R2 and R2s groups than in the Rls
3 groups. Other than BAL neutrophils, which were significantly higher in the R2 and R2s groups,
4 no other inflammatory cells (macrophages, eosinophils, or lymphocytes) or biochemical
5 parameters of lung injury were significantly different between the R2 and R2s groups and the
6 Rls group. Although soluble forms of zinc had been found in guinea pigs to produce a greater
7 pulmonary response than other sulfated metals (Amdur et al., 1978), and, although the level of
8 zinc was 30-fold greater in R2 than Rl, the precise mechanisms by which zinc may induce such
9 responses are unknown. Nevertheless, these results show that the composition of soluble metals
10 and sulfate leached from ROFA, a type of emission source particle, is critical in the development
11 of airway hyperractivity and lung injury.
12 It has been shown that reactive oxygen species play an important role in the in vivo toxicity
13 of ROFA, Dye et al. (1997) pretreated rats with an intraperitoneal injection of saline or
14 dimethylthiourea (DMTU) (500 mg/kg), followed 30 min later by intratracheal instillation of
15 either acidic saline (pH = 3.3) or an acidified suspension of ROFA (500 /ug). The systemic
16 administration of DMTU impeded development of the cellular inflammatory response to ROFA,
17 but did not ameliorate biochemical alterations in BAL fluid. In a subsequent study, these
18 investigators determined that oxidant generation, possibly induced by soluble vanadium
19 compounds in ROFA, are responsible for the subsequent rat tracheal epithelial cells gene
20 expression, inflammatory cytokine productions (MIP-2 and IL-6), and cytotoxicity (Dye et al.,
21 1999).
22 In addition to transition metals, other components in fly ash also may cause lung injury.
23 The effects of arsenic compound in coal fly ash or copper smelter dust on the lung integrity and
24 on the ex vivo release of TNFa by alveolar phagocytes were investigated by Broeckaert et al.
25 (1997). Female Naval Medical Research Institute (NMRI) mice were instilled with different
26 particles normalized for the arsenic content (20 Mg/kg body weight [i.e., 600 ng arsenic/mouse])
27 and the particle load (100 mg/kg body weight [i.e., 3 mg/mouse]). Mice received tungsten
28 carbide (WC) alone, coal fly ash (CFA) alone, copper smelter dust (CMP) mixed with WC, and
29 Ca3(AsO4)2 mixed with WC (see Table 8-2 for concentration details). Copper smelter dust
30 caused a severe but transient inflammatory reaction, whereas a persisting alveolitis (30 days
31 postexposure) was observed after treatment with coal fly ash. hi addition, TNFa production in
March 2001 8-22 DRAFT-DO NOT QUOTE OR CITE
-------
1 response to LPS by alveolar phagocytes were significantly inhibited at Day 1 and still was
2 observed at 30 days after administration of CMP and CFA. Although arsenic was cleared from
3 the lung tissue 6 days after Ca3(AsO4)2 administration, a significant fraction persisted (10 to 15%
4 of the arsenic administered) in the lung of CMP- and CFA-treated mice at Day 30. It is possible
5 that suppression of TNF-a production is dependent upon the slow elimination of the particles and
6 their metal content from the lung.
7 In summary, intratracheally injected ROFA produced acute lung injury and inflammation.
8 The water soluble metals in ROFA appear to play a key role in the acute effects of instilled
9 ROFA. Although studies done with ROFA clearly show that combustion generated particles
10 with a high metal content can cause substantial lung injury, there is still insufficient data to
11 extrapolate these effects to the low levels of particle associated transition metals in ambient PM.
12
13 8.2.4 Ambient Bioaerosols
14 Ambient bioaerosols include fungal spores, pollen, bacteria, viruses, endotoxins, and plant
15 and animal debris. Such biological aerosols can produce various health effects including:
16 infections, hypersensitivity, and toxicoses. Bioaerosols present in the ambient environment have
17 the potential to cause disease in humans under certain conditions. However, it was concluded in
18 the previous criteria document (U.S. Environmental Protection Agency, 1996a) that bioaerosols,
19 at the concentrations present in the ambient environment, would not account for the observed
20 effects of particulate matter on human mortality and morbidity reported in PM epidemiological
21 studies. Moreover, bioaerosols generally represent a rather small fraction of the measured urban
22 ambient PM mass and are typically present even at lower concentrations during the winter
23 months when notable ambient PM effects have been demonstrated. Bioaerosols tend to be in the
24 coarse fraction of PM, but some bioaerosols are found in the fine fraction.
25 More recent studies on ambient bioaerosols are summarized in Table 8-6. Endotoxin
26 exposure in pig farmers is associated with a large annual decline in FEV, (mean of 73 ml/year),
27 which is about 2 to 3 times more rapid than in healthy adults (Vogelzang et al., 1998). Michel
28 et al. (1997) examined the dose-response relationship to inhaled lipopolysaccharide (LPS: the
29 purified derivative of endotoxin) in normal healthy volunteers exposed to 0, 0.5, 5, and 50 /ug of
30 LPS. Inhalation of 5 or 50 yUg of LPS resulted in increased PMNs in blood and sputum samples.
31 At the higher concentration, a slight (3%) but not significant decrease in FEV, was observed.
March 2001 8-23 DRAFT-DO NOT QUOTE OR CITE
-------
C/J
0
CNT BIOAEROS
HH
S
fe
0
t/3
H
U
U,
M
^,
o:
o
H
tf
£"H
OH
C/3
tt
vd
00
W
J
^
H
u
g
CU
a£
Z
•c
Effect of Pa;
u
1 -I
w Q
to
QJ
"0
(f
c
_0
"s
"c
c
o
a §
3 O1
0 J
UJ "
"y
c^
CU
CJ O
"O "Jj
D c^
O M
tn" <
-- C
0. -fa
CO CO
s
s ^^
c ^
o ^
U C.
fc-
0 (U
S 1 "°
-c c: in
^
.g ^ "o
|^ |
111
Rcantly reduced FEV,, incr
ho-responsiveness, and inc
ror all exposures. Effects w
of exposure.
5 c « £
a g = 3
In
<
Z
<
Z
J= S
^ *- c
ills
.3 5-sl
5 -o § "3
H § 1 8
CJ)
0) £
S 2
S 3
00 .0
S'S
0 £
IJ
•^ i ^
•^ vi O
H O f*"l DO
-C C (N CQ
•§
U4
|
ion of LPS and O, on mflamm:
Significant interact
o
S
(N
CO
CL.
J
> s
o"
o
rats. Oj and UF-C interacted •
o produce greater PMN respon
t on lung inflammatory respon:
responses in young
"priming" by LPS t
has a priming effec
and UF-C
u
B m PMN lummol-enhanced
1
•o
1
1C
1
o^
0
C
1
0
c
oo'
cu
00
o
_c
1
chemllummescence
id increase in sputum PMNs,
CRP and PMNs, an
e
£
1
'O with 5.0 ,ug LPS, increase n
PMNs, blood and unne CRP, i
monocytes, and MF
'1
lymphocytes, TNFa, and ECP
temperature, blood
PMNs, monocytes,
50 Mg LPS.
w
0>
o
at
o
s of pool-associated granulom!
Recurring outbreak
oo
Ov
£
"3
f- <~
3); case patients had higher cui
sis indicated increased levels o
pneumonitis (n = 3.
work hours. Analyi
ir and water
endotoxin m pool a
E
o
Q
f~ UI 00
O (N eO
S
"Ss
3.
o •£
c u.
, o
o
<
Z
E
3.
00
I;
U. O
'B'S
-o S
vi
S S •*
IIS
S u u
CJ = o
>l 8.
Q -S '
UJ es
r~ II
« £
e 2
, O o,
M
O
-
•2 ~ I,
-
March 2001
8-24
DRAFT-DO NOT QUOTE OR CITE
-------
1 Cormier et al. (1998) reported an approximate 10% decline in FEV, and an increase in
2 methacholine airway responsiveness after a 5-h exposure inside a swine containment building.
3 This exposure induced significant neutrophilic inflammation in both the nose and the lung.
4 Although these exposures are massive compared to endotoxin levels in ambient PM in U.S.
5 cities, these studies serve to illustrate the effects of endotoxin and associated bioaerosol material
6 in healthy nonsensitized individuals.
7 Some health effects have been observed after occupational exposure to complex aerosols
8 containing endotoxin at concentrations relevant to ambient levels. Zock et al. (1998) reported a
9 decline in FEV, (=3%) across a shift in a potato processing plant with up to 56 endotoxin units
10 (EU)/m3 in the air. Rose et al. (1998) reported a high incidence (65%) of BAL lymphocytes in
11 lifeguards working at a swimming pool where endotoxin levels in the air were on the order of
12 28 EU/m3. Although these latter two studies may point towards pulmonary changes at low
13 concentrations of airborne endotoxin, it is not possible to rule out the contribution of other agents
14 in these complex organic aerosols.
15
16
17 8.3 SYSTEMIC EFFECTS OF PARTICULATE MATTER IN HEALTHY
18 HUMANS AND LABORATORY ANIMALS: IN VIVO EXPOSURES
19 A small number of epidemiology studies have demonstrated that increases in cardiac-
20 related deaths are associated with exposure to PM (U.S. Environmental Protection Agency,
21 1996a), and that PM-related cardiac deaths appear to be as great or greater than those attributed
22 to respiratory causes (see Chapter 6). The toxicological consequences of inhaled particles on the
23 cardiovascular system had not been extensively investigated prior to 1996. Since then (see
24 Table 8-7), Costa and colleagues (Costa and Dreher, 1997) have demonstrated that intratracheal
25 instillation of high levels of ambient particles can increase or accelerate death related to
26 monocrotaline administration in rats. These deaths did not occur with all types of ambient
27 particles tested. Some dusts, such as volcanic ash from Mount Saint Helens, were relatively
28 inert, whereas other ambient dusts, including those from urban sites, were toxic. These early
29 observations suggested that particle composition plays an important role in the adverse health
30 effects associated with episodic exposure to ambient PM, despite the "general particle" effect
31 attributed to the epidemiological associations of ambient PM exposure and increased mortality in
March 2001 8-25 DRAFT-DO NOT QUOTE OR CITE
-------
2
1
to
0
0
~
oo
KJ
ON
O
>
H
1
a
0
2
0
H
XD
r^^
O
H
W
O
0
HH
H- 1
D
W
TABLE 8-7. CARDIOVASCULAR EFFECTS AND OTHER SYSTEMIC
Species, Gender,
Strain Age, or
Body Weight Particle
Fischer 344 rats, OTT
male, 200-250 g
Rats, S-D, male, ROFA
60 days old,
MCT and healthy,
n = 64
Female mongrel CAP
dogs, 14 to 17 kg
Rats, male, S-D, Emission
60 days old, source PM
MCT (60 mg/kg), Ambient
ip and healthy airshed PM
ROFA
Rats, S-D, male, ROFA
60 days old
Healthy CAP
nonsmokers, 1 8 to
40 years old
Mongrel dogs, CAP
some with balloon
occluded LAD
coronary artery,
n= 14
F-344 rat, male, CAP
MCT-treated
Hamster, 6-8 mo
old, Bio TO-2
Exposure Mass
Technique Concentration
Nose-only 40 mg/m'
Inhalation
Instillation 0 0, 0.25, 1.0, and
2.5 mg/rat
Inhalation via 3-360 //g/m3
tracheostomy
Instillation Total mass:
2.5 mg/rat
Total transition
metal 46 Mg/rat
Instillation 0 3, 1 7, or
8 3 mg/kg
Inhalation 23 1 to
3111 Mg/m3
Inhalation via 69-828 Mg/m'
tracheostomy
Inhalation 132-91 9 Mg/m!
Particle
Size
4 to 5 Mm MMAD
1 95 Mm
0.2 to 0 3 Mm
Emission PM
1 78-4. 17 Mm
Ambient PM-
3. 27-4 09 Mm
1.95 Mm
6g = 219
065 Mm
8g = 235
0 23 to 0.34 Mm
6g = 02to2.9
02-1.2 Mm
5g = 0 2-3 9
Exposure
Duration
4h
Analysis at
96 h
6 h/day for 3 days
Analysis at 24
and 96 h
following
instillation
Analysis at
24 h
2 h, analysis at
18h
6 h/day for
3 days
3 h, evaluated at
3 and 24 h
EFFECTS OF PARTICULATE MATTER
Cardiovascular Effects
Increased plasma levels of endothehn- 1 .
No acute lung injury; however, lung NO
production decreased and macrophage
inflammatory protem-2 from lung lavage cells
increased after exposure.
Dose-related hypothermia and bradycardia in
healthy rats, potentiated by compromised models
Peripheral blood parameters were related to
specific particle constituents. Factor analysis
from paired and crossover expenments showed
that hematologic changes were not associated
with increases in total CAP mass concentration
ROFA alone induced some mild arrhythumas;
MCT-ROFA showed enhanced neutrophihc
inflammation;
MCT-ROFA animals showed more numerous
arrhythmias including S-T segment inversions
and A-V block.
Increased plasma fibrinogen at 8 3 mg/kg only.
Increased blood fibnnogen.
Decreased time to ST segment elevation and
increased magnitude in compromised dogs.
Decreased heart and respiratory rate and
increased lavage fluid neutrophils in normal
dogs.
No increase in cardiac arrhythmias;
PM associated increases in HR and blood cell
differential counts, and atnal conduction time
of rats were inconsistent. No adverse cardiac
or pulmonary effects in hamsters
Reference
Bouthilher et al
(1998)
Campen et al. (2000)
Clarke et al (2000)
Costa and Dreher
(1997)
Gardner et al. (2000)
Ghioetal (2000a)
Godleski et al. (2000)
Gordon et al (2000)
-------
00
H
6
O
2
O
H
/O
O
H
Species, Gender,
Strain Age, or
Body Weight Particle
Rats, S-D, MCT FOFA
(50 mg/kg SC),
250 g
12 to 13 -week-old ROFA
male WKY and SH
rats
Hartley guinea pig, DEP
male, 890 g
Healthy 10.5-year- ROFA
old beagles,
n = 4
Rabbit, New Colloidal
Zealand White, carbon
female, 1.8 to
2.4kg
Rat, S-D male, ROFA
MCT
Exposure Mass Particle Exposure
Technique Concentration Size Duration
Inhalation 580 ± 1 10 ^g/m' 2 06 ^m MMAD 6 h/day for 3 days
5g=157
Nose-only 1 5 mg/m' N/A 6 h/day for
inhalation 3 days
Intravenous 500 mg/mL 0 34 ,um 10% solution
solution every 5 mm
Oral inhalation 3 mg/m! 2 22 ^m MMAD 3 h/day for
6g = 27I 3 days
Instillation 2 mL of 1 % < 1 ^m Examined for 24
colloidal carbon to 1 92 h after
(20 mg) instillation
Instillation 0, 250, 1000, or 1 .95 ^m MMAD Monitored for
2500 Mg in 0.3 mL 8g = 2.19 96 h after
saline instillation of
ROFA particles
Cardiovascular Effects
Increased expression of the proinflammatory
chemokme MP-2 in the lung and heart of
MCT-treated rats, less in healthy rats
Significant mortality only in MCT-treated rats.
Cardiomyopathy and monocytic cell infiltration,
along with increased cytokine expression, was
found in left ventricle of SH rats because of
underlying cardiovascular disease ECG showed
exacerbated ST segment depression caused by
ROFA
DMSO extract of DEP solution induced
arrhythmias and deaths by AV block; thus,
water-soluble fractions of DEP may be
responsible for cardiotoxicity
No consistent changes in ST segment, the form
or amplitude of the T wave, or arrhythmias,
slight bradycardia during exposure.
Colloidal carbon stimulated the release of
BRDU-labeled PMNs from the bone marrow
The supernatant of alveolar macrophages treated
with colloidal carbon in vitro also stimulated the
release of PMNs from bone marrow, likely via
cytokmes
Dose-related increases in the incidence and
duration of serious arrhythmic events in normal
rats. Incidence and seventy of arrythmias were
increased greatly in the MCT rats. Deaths were
seen at each instillation level in MCT rats only
(6/12 died after MCT + ROFA)
Reference
Killmgsworth et al.
(1997)
Kodavanti et al.
(2000b)
Minami et al.
(1999)
Muggenberg et al.
(2000)
Terashima et al
(1997)
Watkmson et al.
(1998)
O
HH
H
m
-------
TABLE 8-7 (cont'd). CARDIOVASCULAR EFFECTS AND OTHER SYSTEMIC EFFECTS OF PARTICULATE MATTER
*-«
•-(
o
to
o
1—
oo
to
00
Species, Gender,
Strain Age, or
Body Weight
(1) Healthy S-D
rats and cold-
stressed, ozone-
treated, and
MCT-treated rats
(2) Heathy and
MCT-treated S-D
rats, SH rats,
WKY rats
(3) 15-mo-oldSH
rats
(4) MCT-treated
S-D rats
Particle
ROFA
ROFA
OTT
ROFA
MSH
Fe2(S04)3
VS04
NiSO,
Exposure
Technique
Intratracheal
instillation
Inhalation
Intratracheal
instillation
Intratracheal
instillation
Mass
Concentration Particle Size
00, 025, 10, or 1.95 ^m
2 5 mg/rat
15 mg/m3 1.95 f^m
2 5 mg I 95 Atm
0.5 mg
25mg
105/ug 1.95 Mm
245 Aig
262 5 A-g
Exposure
Duration
Monitored for
96 h after
instillation
6 h/day for
3 days
Monitored for
96 h after
instillation
Monitored for
96 h after
instillation
Cardiovascular Effects
(I) Healthy rats exposed IT to ROFA
demonstrated dose-related hypothermia,
bradycardia, and increased arrhythmias.
Compromised rats demonstrated exaggerated
hypothermia and cardiac responses to IT ROFA
Mortality was seen only in the MCT-treated rats
exposed to ROFA by IT. (2) Pulmonary
hypertensive (MCT-treated S-D) and
systemically hypertensive (SH) rats exposed to
ROFA by inhalation demonstrated similar
effects, but of diminished amplitude. There were
no lethalities by the inhalation route. (3) Older
rats exposed IT to OTT showed a pronounced
biphasic hypothermia and a severe drop in HR
accompanied by increased arrhythmias; exposure
to ROFA caused less pronounced, but similar
effects No cardiac effects were seen with
exposure to MSH. (4) Ni and V showed the
greatest toxicity, Fe-exposed rats did not differ
from controls
Reference
Watkinson et al
(2000)
Tl
H
6
o
2
3
O
d
o
H
w
g
o
H—I
H
W
-------
1 many regions of the United States (i.e., regions with varying particle composition). Work that
2 examines the role of inherent susceptibility to the adverse effects of PM in compromised animal
3 models provides a potentially important link to epidemiological observations.
4 To date, studies examining the systemic and cardiovascular effects of particles have used a
5 number of compromised animal models, largely rodents, to mimic human disease. Two studies
6 in normal or compromised dogs (Godleski et al., 2000; Muggenberg et al., 2000) also have been
7 published as well as the preliminary results from human exposure studies (see Section 8.4.1).
8 The following discussion of the systemic effects of PM first describes studies performed using
9 metal-laden ROFA as a source particle and then compares these findings with studies using
10 concentrated ambient PM.
11 Killingsworth and colleagues (1997) used a fuel oil fly ash to examine the adverse effects
12 of a model urban particle in an animal model (monocrotaline-MCT) of cardiorespiratory disease;
13 MCT causes pulmonary vascular inflammation and hypertension. They observed 42% mortality
14 in MCT rats exposed to approximately 580 /ug/m3 fly ash for 6 h/day for 3 consecutive days.
15 Deaths did not occur in MCT rats exposed to filtered air or in saline-treated rats exposed to fly
16 ash. The increase in deaths in the MCT/fly ash group was accompanied by an increase in
17 neutrophils in lavage fluid and an increased immunostaining of MIP-2 in the heart and lungs of
18 the MCT/fly ash animals. Cardiac immunohistochemical analysis indicated increased MIP-2 in
19 cardiac macrophages. The fly ash-induced deaths did not result from a change in pulmonary
20 arterial pressure; the cause of death was not identified.
21 In a similar experimental model, Watkinson et al. (1998) examined the effects of
22 intratracheally instilled ROFA (0.0, 0.25, 1.0, 2.5 mg in 3 mL saline) on ECG measurements in
23 control and MCT rats. They observed a dose-related increase in the incidence and duration of
24 serious arrhythmic events in control animals exposed to ROFA particles and these effects were
25 clearly exacerbated in the MCT animals. Similar to the results of Killingsworth et al. (1997),
26 health animals treated with ROFA suffered no deaths, but MCT rats had 1, 2, and 3 deaths in the
27 low-, medium-, and high-dose groups, respectively. This study suggests that ROFA PM may be
28 implicated in conductive and hypoxemic arrhythmias associated with the cardiac-related deaths.
29 Kodavanti et al. (1999) exposed MCT rats to ROFA by either intratracheal instillation
30 (0.83 or 3.33 mg/kg) or nose-only inhalation (15 mg/m3, 6 h/day for 3 consecutive days). Similar
31 to Watkinson et al. (1998), intratracheal instillation of ROFA in MCT rats resulted in 58%
March 2001 8-29 DRAFT-DO NOT QUOTE OR CITE
-------
1 mortality, whereas no mortality occurred in MCT rats exposed to ROFA by inhalation exposure.
2 No mortality occurred in healthy rats exposed to ROFA or in MCT rats exposed to clean air.
3 Despite the fact that mortality was not associated with ROFA inhalation exposure of MCT rats,
4 exacerbation of lung lesions and pulmonary inflammatory cytokine gene expression, as well as
5 ECG abnormalities, clearly were evident.
6 Watkinson and colleagues further examined the effect of instilled ROFA in two additional
7 rodent models of compromised health (Watkinson et al., 2000; Campen et al., 2000). The effect
8 of ozone-induced pulmonary inflammation (preexposure to 1 ppm ozone for 6 h) or housing in
9 the cold (10 °C) on the response to ROFA in rats was similar to that produced by MCT.
10 Bradycardia, arrhythmias, and hypothermic changes were consistently observed in the ozone and
11 hypothermic animals treated with ROFA, although, unlike in the MCT animals, no deaths
12 occurred. Thus, in three rodent models of cardiopulmonary disease/stress, instillation of 0.25 mg
13 or more of ROFA can produce systemic changes that can be considered adverse health effects
14 and address potential mechanisms of toxicity consistent with the epidemiology and panel studies
15 showing cardiac effects in humans.
16 Watkinson and colleagues (2000) also sought to examine the relative toxicity of different
17 particles on the cardiovascular system of spontaneously hypertensive rats. They instilled 2.5 mg
18 of representative particles from ambient (Ottawa) or natural (Mount Saint Helens volcanic ash)
19 sources and compared the response to 0.5 mg ROFA. Instilled particles were either mass
20 equivalent dose or adjusted to produce equivalent metal dose. They observed adverse changes in
21 ECG, heart rate, and arrhythmia incidence that were much greater in the Ottawa- and ROFA-
22 treated rats than in the Mount Saint Helens-treated rats. The cardiovascular changes observed
23 with the Ottawa particles were actually greater than with the ROFA particles. These series of
24 experiments by Watkinson and colleagues clearly demonstrate that instillation of particles, albeit
25 at a very high concentration, can produce cardiovascular effects. They also demonstrate that PM
26 exposures of equal mass dose did not produce the same cardiovascular effects, suggesting that
27 PM composition was responsible for the observed effects and that PM metal content was a better
28 indicator than PM mass.
29 Because of concerns regarding the relevance of particles administered by intratracheal
30 instillation, investigators also have examined the cardiovascular effects of ROFA particles using
31 more realistic inhalation exposure protocols. Kodavanti et al. (2000b) found that exposure to a
March 2001 8-30 DRAFT-DO NOT QUOTE OR CITE
-------
1 high concentration of ROFA (15 mg/m3 for 6 h/day for 3 days) produced alterations in the ECG
2 waveform of spontaneously hypertensive (SH) but not normotensive rats. Although the ST
3 segment area of the ECG was depressed in the SH rats exposed to air, further depressions in the
4 ST segment were observed at the end of the 6-h exposure to ROFA on Days 1 and 2. The
5 enhanced ST segment depression was not observed on the third day of exposure, suggesting that
6 adaptation to the response had occurred. Thus, exposure to a high concentration of ROFA
7 exacerbated a defect in the electroconductivity pattern of the heart in an animal model of
8 hypertension. This ROFA-induced alteration in the ECG waveform was not accompanied by an
9 enhancement in the monocytic cell infiltration and cardiomyopathy that also develop in SH rats.
10 Further work is necessary to determine the relevance of this ROFA study to PM at concentrations
11 relevant to ambient exposures.
12 Godleski and colleagues (2000) have performed a series of important experiments
13 examining the cardiopulmonary effects of inhaled concentrated ambient PM on normal mongrel
14 dogs and on dogs undergoing coronary artery occlusion. Dogs were exposed to concentrated
15 ambient PM for 6 h/day for 3 consecutive days. The investigators found little evidence of
16 pulmonary inflammation or injury in normal dogs exposed to PM (daily range of mean
17 concentrations was approximately 100 to 1000 //g/m3). A greater than twofold increase in
18 percent neutrophils (p < 0.05) was the only lavage parameter that was significantly different from
19 sham-exposed animals. Despite the absence of major pulmonary effects, a significant increase in
20 heart rate variability (an indice of cardiac autonomic activity), a decrease in heart rate, and an
21 increase in T alternans (an indice of vulnerability to ventricular fibrillation) were observed. The
22 significance of these effects is not yet clear as the effects did not occur on all exposure days.
23 For example, the change in heart rate variability was observed on 10 of the 23 exposure days.
24 In support of the "general particle" theory, exposure assessment of particle composition produced
25 no specific components of the particles that were correlated with the day-to-day variability in
26 response. Moreover, whereas the heart rate variability change suggests a proarrhythmic response,
27 the increase in T alternans suggests an anti-arrhythmic effect of inhaled concentrated ambient
28 PM.
29 The most important finding in the experiments of Godleski and colleagues (2000) was the
30 observation of a potential increase in ischemic stress of the cardiac tissue from repeated exposure
31 to concentrated ambient PM. During coronary occlusion in four dogs exposed to PM, they
March 2001 8-31 DRAFT-DO NOT QUOTE OR CITE
-------
1 observed a significantly more rapid development of ST elevation of the ECG waveform. In
2 addition, the peak ST-segment elevation was greater after PM exposure. Together, these changes
3 suggest that concentrated ambient PM can augment the ischemia associated with coronary artery
4 occlusion in this dog model. Additional work in more dogs as well as other species is necessary
5 to determine the significance of these findings to the human response to ambient PM.
6 Contrary to the adverse effects of inhaled concentrated ambient PM reported by Godleski
7 and colleagues in a peer-reviewed publication on ambient PM (Godleski et al., 2000).
8 Muggenberg and colleagues (2000) have found that exposure to high concentrations of ROFA
9 produces no consistent changes in amplitude of the ST-segment, form of the T wave, or
10 arrhythmias in dogs. In their studies, four beagle dogs were exposed to 3 mg/m3 ROFA particles
11 generated for 3 h/day for 3 consecutive days with a Wright dust feeder. They did note that there
12 was a slight but variable decrease in heart rate, but the changes were not statistically or
13 biologically significant. The ROFA was collected from the same power plant as the Godleski
14 study but at a later time point. The transition metal content of the ROFA used by Muggenberg
15 was approximately 15% by mass, a value that is on the order of a magnitude higher than that
16 found in ambient urban PM samples. Although the study did not specifically address the effect
17 of metals, it suggests that inhalation of high concentrations of metals may have little effect on the
18 cardiovascular system of a healthy individual. Therefore, the different findings between the dog
19 studies illustrate the difficulties in extrapolating animal toxicology data to human health effects.
20 In a series of studies, Gordon, Nadziejko, and colleagues examined the response of the
21 rodent cardiovascular system to concentrated ambient PM derived from New York city air
22 (Gordon et al., 2000). Particles of 0.2 to 2.5 /^m in diameter were concentrated up to 10 times
23 their levels in ambient air (-150 to 900 ,ug/m3) to maximize possible differences in effects
24 between normal and cardiopulmonary-compromised laboratory animals. ECG changes were not
25 detected in normal Fischer 344 rats or hamsters exposed to concentrated ambient PM for 1 to 3
26 days. Similarly, no deaths or ECG changes were observed in MCT rats or cardiomyopathic
27 hamsters exposed to PM. Contrary to the decrease in heart rate observed in dogs exposed to
28 concentrated ambient PM (Godleski et al., 2000), heart rate was increased in both normal and
29 MCT rats exposed to PM. The increase was approximately 5% and was not observed on all
30 exposure days. Thus, extrapolation of the heart rate changes in these animal studies to human
March 2001 8-32 DRAFT-DO NOT QUOTE OR CITE
-------
1 health effects is difficult, although the increase in heart rate in rats is similar to that observed in
2 human population studies (see Chapter 6).
3 Gordon and colleagues (1998) have reported other cardiovascular effects in animals
4 exposed to inhaled CAP. Increases in peripheral blood platelets and neutrophils were observed
5 in control and MCT rats at 3 h, but not 24 h, after exposure to 150 to 400 /ug/m3 concentrated
6 ambient PM (CAP). This neutrophil effect, likely a result of vascular demargination, did not
7 appear to be dose related and did not occur on all exposure days, thus, suggesting that day-to-day
8 changes in particle composition may play an important role in the systemic effects of inhaled
9 particles. Terashima et al. (1997) also examined the effect of particles on circulating neutrophils.
10 They instilled rabbits with 20 mg colloidal carbon, a relatively inert particle (<1 /urn), and
11 observed a stimulation of the release of 5'-bromo-2'deoxyuridine (BrdU)-labeled PMNs from the
12 bone marrow at 2 to 3 days after instillation. Because the instilled supernatant from rabbit AMs
13 treated in vitro with colloidal carbon also stimulated the release of PMNs from the bone marrow,
14 they hypothesized that cytokines released from activated macrophages could be responsible for
15 this systemic effect.
16 The results of epidemiology studies have suggested that homeostatic changes in the
17 vascular system can occur after episodic exposure to ambient PM. Ohio et al. (2000a) have
18 shown that inhalation of concentrated PM in healthy nonsmokers causes increased levels of
19 blood fibrinogen. They exposed 38 volunteers exercising intermittently at moderate levels of
20 exertion for 2 h to either filtered air or particles concentrated from the air in Chapel Hill, NC (23
21 to 311 ug/m2). Blood obtained 18 h after exposure contained significantly more fibrinogen than
22 blood obtained before exposure. The observed effects in blood may associated with the mild
23 inflammation also found 18 h after exposure to CAP (see Section 8.2.3).
24 Gardner et al. (2000) examined whether the instillation of particles would alter blood
25 coagulability factors in laboratory animals. Sprague-Dawley rats were instilled with 0.3, 1.7, or
26 8.3 mg/kg of ROFA or 8.3 mg/kg Mount Saint Helens volcanic ash. They observed an increase
27 in plasma fibrinogen in healthy rats. Because fibrinogen is a known risk factor for ischemic heart
28 disease and stroke, the authors suggested that this alteration in the coagulation pathway could
29 take part in the triggering of cardiovascular events in susceptible individuals. Elevations in
30 plasma fibrinogen, however, were observed in healthy rats only at the highest treatment dose, and
31 no other changes in clotting function were noted. Because the lower treatment doses are known
March 2001 8-33 DRAFT-DO NOT QUOTE OR CITE
-------
1 to cause pulmonary injury and inflammation, albeit to a lower extent, the absence of plasma
2 fibrinogen changes at these lower doses suggests that only high levels of pulmonary injury are
3 able to produce an effect in healthy test animals.
4 In summary, controlled animal studies have provided initial evidence that high
5 concentrations of inhaled or instilled particles can have systemic, especially cardiovascular,
6 effects, hi the case of MCT rats, these effects can be lethal. Understanding the pathways by
7 which very small concentrations of inhaled ambient PM can produce systemic, life-threatening
8 changes, however, is far from clear. Among the hypotheses that have been proposed to account
9 for the nonpulmonary effects of PM are activation of neural reflexes, cytokine effects on heart
10 tissue (Killingsworth et al., 1997), alterations in coagulability (Seaton et al., 1995; Sjogren,
11 1997), and perturbations in homeostatic processes such as heart rate or heart rate variability
12 (Watkinson et al., 1998). A great deal of research using controlled exposures of animal and
13 human subjects to PM will be necessary to test mechanistic hypotheses generated to date, as well
14 as those that are likely to be proposed in the future.
15
16
17 8.4 SUSCEPTIBILITY TO THE EFFECTS OF PARTICIPATE
18 MATTER EXPOSURE
19 Susceptibility of an individual to adverse health effects of PM can vary depending on a
20 variety of host factors such as age, nutritional status, physiological activity profile, genetic
21 predisposition, or preexistent disease. The potential for preexistent disease to alter adverse
22 response to toxicant exposure is widely acknowledged but poorly understood. Because of
23 inherent variability (necessitating large numbers of subjects) and ethical concerns associated with
24 using diseased subjects in clinical research studies, a solid database on human susceptibilities is
25 lacking. For more control over both host and environmental variables, animal models often are
26 used. However, care must be taken in extrapolation from animal models of human disease to
27 humans. Rodent models of human disease, their use in toxicology and the criteria for judging
28 their appropriateness as well as their limitations must be considered (Kodavanti et al., 1998b;
29 Kodavanti and Costa, 1999).
30
March 2001 8-34 DRAFT-DO NOT QUOTE OR CITE
-------
1 8.4.1 Effects of Particulate Matter on Cardiopulmonary Compromised Hosts
2 Epidemiological studies suggest there may be subsegments of the population that are
3 especially susceptible to effects from inhaled particles (see Chapter 6). The elderly with chronic
4 cardiopulmonary disease, those with pneumonia and possibly other lung infections, and those
5 with asthma (at any age) appear to be at higher risk than healthy people of similar age.
6 Unfortunately, most toxicology studies have used healthy adult animals. An increasing number
7 of newer studies have examined effects of ambient particles in compromised host models. Costa
8 and Dreher (1997) used a rat model of cardiopulmonary disease to explore the question of
9 susceptibility and the possible mechanisms by which PM effects are potentiated. Rats with
10 advanced monocrotaline (MCT)-induced pulmonary vasculitis/hypertension were given
11 intratracheal instillations of ROFA (0, 0.25, 1.0, and 2.5 mg/rat). The MCT animals had a
12 marked neutrophilic inflammation. In the context of this inflammation, ROFA induced a four- to
13 fivefold increase in BAL PMNs. There was increased mortality at 96 h that was ROFA-dose
14 dependent. The results of this study indicate that particles, albeit at a high concentration,
15 enhanced the neutrophilic inflammation and mortality in MCT animals.
16 Kodavanti et al. (1999) also studied PM effects in the MCT rat model of pulmonary
17 disease. Rats treated with 60 mg/kg MCT were exposed to 0, 0.83. or 3.3 mg/kg ROFA by
18 intratracheal instillation and to 15 mg/kg ROFA by inhalation. Both methods of exposure caused
19 inflammatory lung responses and ROFA exacerbated the lung lesions, as shown by increased
20 lung edema, inflammatory cells, and alveolar thickening.
21 The manner in which MCT can alter the response of rats to inhaled particles was examined
22 by Madl and colleagues (1998). Rats were exposed to fluorescent colored microspheres (1 //m)
23 2 weeks after treatment with MCT. In vivo phagocytosis of the microspheres was altered in the
24 MCT rats in comparison with control animals. Fewer microspheres were phagocytized in vivo
25 by alveolar macrophages and there was a concomitant increase in free microspheres overlaying
26 the epithelium at airway bifurcations. The decrease in in vivo phagocytosis was not accompanied
27 by a similar decrease in vitro. Macrophage chemotaxis, however, was impaired significantly in
28 MCT rats compared with control rats. Thus, MCT appeared to impair particle clearance from the
29 lungs via inhibition of macrophage chemotaxis.
30 The sulfur dioxide (SO2)-induced model of chronic bronchitis has also been used to
31 examine the potential interaction of PM with preexisting lung disease. Clarke and colleagues
March 2001 8-35 DRAFT-DO NOT QUOTE OR CITE
-------
1 pretreated Sprague Dawley rats for 6 weeks with air or 170 ppm SO2 for 5 h/day and 5 days/week
2 (Clarke et al., 1999). Exposure to concentrated air particles for 5 h/day for 3 days at an average
3 concentration of 515 Mg/m3 produced changes in pulmonary function as evidenced by significant
4 increases in tidal volume in both air- and SO2-pretreated rats. Exposure to concentrated ambient
5 PM also produced significant changes in both cellular and biochemical markers in lavage fluid.
6 In comparison to control animal values, protein was increased approximately threefold in SO2-
7 pretreated animals exposed to concentrated ambient PM. Lavage fluid neutrophils and
8 lymphocytes were increased significantly in both pretreatment groups of rats exposed to
9 concentrated ambient PM, with greater increases in both cell types in the SO2-pretreated rats.
10 Thus, exposure to concentrated ambient PM produced adverse changes in the respiratory system,
11 but no deaths, in both normal rats and in a rat model of chronic bronchitis.
12 Clarke et al. (2000) next examined the effect of concentrated ambient PM in normal rats of
13 different ages. Unlike the earlier study that used Sprague-Dawley rats, 4- and 20-mo-old Fischer
14 344 were examined after 3 days of exposure to concentrated ambient PM. They found that
15 exposure to daily mean concentrations of 80, 170, and 50 /ug/m3 PM produced statistically
16 significant increases in total neutrophil counts (up over 10-fold) in lavage fluid of the young, but
17 not the old, rats. Thus, repeated exposure to relatively low concentrations of ambient PM
18 produced an inflammatory response, although the actual percent neutrophils in the concentrated
19 ambient PM-exposed young rats was low (approximately 3%). On the other hand, Gordon and
20 colleagues found no evidence of neutrophil influx in the lungs of normal and monocrotaline-
21 treated Fischer 344 rats exposed in nine separate experiments to concentrated ambient PM
22 (Gordon et al., 2000) as high as 400 /u,g/m3 for a 6-h exposure or 192 /ug/m3 for three daily 6-h
23 exposures. Similarly, normal and cardiomyopathic hamsters showed no evidence of pulmonary
24 inflammation or injury after a single exposure to concentrated ambient PM. Gordon and
25 colleagues did report a statistically significant doubling in protein concentration in lavage fluid in
26 monocrotaline-treated rats exposed for 6 h to 400 /ug/m3 concentrated ambient PM. Because of
27 the disparity in findings in the response of normal Fischer 344 rats to concentrated ambient PM
28 between these two labs, it is important that the reproducibility of these experiments be examined.
29 Kodavanti and colleagues (1998b) also have examined the effect of concentrated ambient
30 PM in normal rats and rats with sulfur dioxide-induced chronic bronchitis. In four separate
31 exposures to PM, there was a significant increase in lavage fluid protein in bronchitic rats from
March 2001 8-36 DRAFT-DO NOT QUOTE OR CITE
-------
1 only one exposure protocol in which the rats were exposed to 444 and 843 /ug/m3 PM on
2 2 consecutive days (6 h/day). Neutrophil counts were increased in bronchitic rats exposed to
3 concentrated ambient PM in three of the four exposure protocols, but was decreased in the fourth
4 protocol. No other changes in normal or bronchitic rats were observed, even in the exposure
5 protocols with higher PM concentrations. Thus, rodent studies have demonstrated that
6 inflammatory changes can be produced in normal and compromised animals exposed to
7 concentrated ambient PM. These findings are important because only a limited number of
8 studies have used real-time inhalation exposures to actual ambient urban PM.
9 Pulmonary function measurements are often less invasive than other means to assess the
10 effects of inhaled air pollutants on the mammalian lung. Although the publication of the 1996
11 PM AQCD, a number of investigators have examined the response of rodents and dogs to inhaled
12 ambient particles. In general, these investigators have demonstrated that ambient PM has
13 minimal effects on pulmonary function tests. Gordon et al. (2000) exposed normal and
14 monocrotaline-treated rats to filtered air or 181 //g/m3 concentrated ambient PM for 3 h.
15 For both normal and monocrotaline-treated rats, no differences in lung volume measurements or
16 diffusion capacity for carbon monoxide were observed between the air or PM exposed animals at
17 3 or 24 h after exposure. Similarly, in cardiomyopathic hamsters, concentrated ambient PM had
18 no effect on these same pulmonary function measurements.
19 In an examination of the effect of concentrated ambient PM on airway responsiveness in
20 mice, Goldsmith and colleagues (1999) exposed control and ovalbumin-sensitized mice to an
21 average concentration of 787 /^g/m3 PM for 6 h/day for 3 days. Although ovalbumin
22 sensitization itself produced an increase in the nonspecific airway responsiveness to inhaled
23 methylcholine, concentrated ambient PM did not change the response to methylcholine in
24 ovalbumin-sensitized or control mice. For comparison, these investigators examined the effect
25 of inhalation of an aerosol of the active soluble fraction of ROFA on control and ovalbumin-
26 sensitized mice and demonstrated that ROFA could produce nonspecific airway
27 hyperresponsiveness to methylcholine in both control and ovalbumin-sensitized mice. Similar
28 increases in airway responsiveness have been observed after exposure to ROFA in normal and
29 ovalbumin-sensitized rodents (Gavett et al., 1997, 1999; Hamada et al., 1999, 2000). Other
30 pulmonary function endpoints have been studied in animals exposed to concentrated ambient
31 PM. Clarke et al. (1999) observed that tidal volume was increased slightly in both control rats
March 2001 8-37 DRAFT-DO NOT QUOTE OR CITE
-------
1 and rats with sulfur dioxide-induced chronic bronchitis exposed to 206 to 733 /ug/m3 PM on
2 3 consecutive days. No changes in peak expiratory flow, respiratory frequency, or minute
3 volume were observed after exposure to concentrated ambient PM. In the series of dog studies
4 by Godleski et al. (2000) (also see Section 8.3), no signficant changes in pulmonary functions
5 were observed in normal mongrel dogs exposed to concentrated ambient PM, although a 20%
6 decrease in respiratory frequency was observed in dogs that underwent coronary artery occlusion
7 and were exposed to PM. Thus, studies using normal and compromised animal models exposed
8 to concentrated ambient PM have found minimal biological effects of ambient PM on pulmonary
9 function.
10 In studying the influence of age on susceptibility to PM, Johnston et al. (1998) exposed
11 8-week-old mice (young) and 18-mo-old mice (old) to polytetrafluoroethylene fumes (PTFE)
12 (0, 10, 25, and 50 yUg/m3) for 30 min. Lung lavage endpoints (PMN, protein, LDH, and
13 p-glucuronidase) as well as lung tissue mRNA levels for various cytokines, metallothionein and
14 for Mn superoxide dismutase were measured 6 h following exposure. Protein, lymphocyte,
15 PMN, and TNF-a mRNA levels were increased in older mice when compared to younger mice.
16 These findings suggest that the inflammatory response to PTFE fumes is altered with age, being
17 greater in the older animals. Although Teflon particles are not a valid surrogate for ambient
18 ultrafme particles (Oberdorster et al., 1992), this study did provide evidence to support the
19 hypothesis that particle-induced pulmonary inflammation is different between young and old
20 organisms.
21 Kodavanti et al. (2000b; 2001) used genetically predisposed spontaneously hypertensive
22 (SH) rats as a model of cardiovascular disease to study PM-related susceptibility. The SH rats
23 were found to be more susceptible to acute pulmonary injury from intratracheal ROFA exposure
24 than normotensive control Wistar Kyoto (WKY) rats (Kodavanti et al., 2001). The primary
25 metal constituents of ROFA, V and Ni, caused differential species-specific effects. Vanadium,
26 which was less toxic than Ni in both strains, caused inflammatory responses only in WKY rats,
27 whereas Ni was injurious to both WKY and SH rats (SH > WKY). This differential
28 responsiveness of V and Ni was correlated with their specificity for airway and parenchymal
29 injury, discussed in another study (Kodavanti et al., 1998b). When exposed to the same ROFA
30 by inhalation, SH rats were more sensitive than WKY rats in regards to vascular leakage
31 (Kodavanti et al., 2000b). The SH rats exhibited a hemorrhagic response to ROFA. Oxidative
March 2001 8-38 DRAFT-DO NOT QUOTE OR CITE
-------
1 stress was much higher in ROFA exposed SH rats than matching WKY rats. Also, SH rats,
2 unlike WKY rats, showed a compromised ability to increase BALF glutathione in response to
3 ROFA, suggesting a potential link to increased susceptibility. Cardiovascular effects were
4 characterized by ST-segment area depression of the ECG in ROFA-exposed SH but not WKY
5 rats. These studies demonstrate the potential utility of cardiovascular disease models for the
6 study of PM health effects and show that genetic predisposition to oxidative stress and
7 cardiovascular disease may play a role in sensitivity to increased PM-related cardiopulmonary
8 injury.
9 In summary, although these studies are just emerging and are only now being replicated or
10 followed more thoroughly to investigate the mechanisms, they do provide evidence of enhanced
11 susceptibility to inhaled PM in "compromised" hosts.
12
13 8.4.2 Genetic Susceptibility to Inhaled Particles
14 A key question in understanding the adverse health effects of inhaled PM is which
15 individuals are susceptible to PM. Although factors such as age and health status have been
16 studied in both epidemiology and toxicology studies, a number of investigators have begun to
17 examine the importance of genetic susceptibility in the response to inhaled particles because of
18 considerable evidence that genetic factors play a role in the response to inhaled pollutant gases.
19 To accomplish this goal, investigators typically have studied the interstrain response to particles
20 in rodents. The response to ROFA instillation in different strains of rats has been investigated by
21 Kodavanti et al. (1996, 1997a). In the first study, male Sprague Dawley (SD) and Fischer-344
22 (F-344) rats were instilled intratracheally with saline or ROFA particles. ROFA instillation
23 produced an increase in lavage fluid neutrophils in both SD and F-344 rats, whereas a time-
24 dependent increase in eosinophils occurred only in SD rats. In a subsequent study (Kodavanti
25 et al., 1997a), SD, Wistar (WIS), and F-344 rats (60 days old) were exposed to saline or ROFA
26 (8.3 mg/kg) by intratracheal instillation and examined for up to 12 weeks. Histology indicated
27 focal areas of lung damage showing inflammatory cell infiltration as well as alveolar, airway, and
28 interstitial thickening in all three rat strains during the week following exposure. Trichrome
29 staining for fibrotic changes indicated a sporadic incidence of focal alveolar fibrosis at 1, 3, and
30 12 weeks in SD rats, whereas WIS and F-344 rats showed only a modest increase in trichrome
31 staining in the septal areas. One of the isoforms of fibronectin mRNA was upregulated in
March 2001 8-39 DRAFT-DO NOT QUOTE OR CITE
-------
1 ROFA-exposed SD and WIS rats, but not in F-344 rats. Thus, there appeared to be a rat strain-
2 dependent variability in the fibrotic response to instilled ROFA.
3 Kleeberger and colleagues have examined closely the role that genetic susceptibility plays
4 in the effect of inhaled acid-coated particles on macrophage function. Nine inbred strains of
5 mice were exposed nose-only to acid-coated particles (10 mg/m3 with 285 yUg/m3 sulfate) for 4 h
6 (Yoshinori et al., 2000). Significant inter-strain differences in Fc-receptor-mediated macrophage
7 phagocytosis were observed, with C57BL/6J mice being the most sensitive. Although neutrophil
8 counts were increased more in C3H/HeOuJ and C3H/HeJ strains of mice than in the other
9 strains, the overall magnitude of change was small and not correlated with the changes in
10 macrophage phagocytosis. In follow-up studies, Ohtsuka et al. (2000a,b) performed a genome-
11 wide scan with a intercross cohort derived from C57BL/6J and C3H/HeJ mice. Analyses of
12 macrophage dysfunction phenotypes of segregant and nonsegregant populations derived from
13 these two strains indicate that two unlinked genes control susceptibility. They identified a
14 3-centiMorgan segment on mouse chromosome 17 that contains an acid-coated particle
15 susceptibility loci. Interestingly, this quantitative trait loci overlaps with those described for
16 ozone-induced inflammation (Kleeberger et al., 1997) and acute lung injury (Prows et al., 1997)
17 and contains several promising candidate genes that may be responsible for the observed genetic
18 susceptibility for macrophage dysfunction in mice exposed to acid-coated particles.
19 Only one study has examined the interstrain susceptibility to ambient particles. C57BL/6J
20 and C3H/HeJ mice were exposed to 250 /^g/m3 concentrated ambient PM for 6 h and examined
21 at 0 and 24 h after exposure for changes in lavage fluid parameters and cytokine mRNA
22 expression in lung tissue (Shukla et al., 2000). No interstrain differences in response were
23 observed. Surprisingly, although no indices of pulmonary inflammation or injury were increased
24 over control values in the lavage fluid, increases in cytokine mRNA expression were observed in
25 both murine strains exposed to PM. Although the increase in cytokine mRNA expression was
26 generally small (approximately twofold), the effect on IL-6, TNF-cc, TGF-P2, and y-interferon
27 was consistent and replication of this study is necessary.
28 Thus, a handful of studies have begun to demonstrate that genetic susceptibility can play a
29 role in the response to inhaled particles. Similar strain differences in response to inhaled metal
30 particles have been observed by other investigators (McKenna et al., 1998; Wesselkamper et al.,
31 2000), although the concentration of metals used in these studies is more relevant to occupational
March 2001 8-40 DRAFT-DO NOT QUOTE OR CITE
-------
1 rather than environmental exposure levels. It remains to be determined whether genetic
2 susceptibility plays as significant a role in the adverse effects of ambient PM as does age or
3 health status.
4
5 8.4.3 Effect of Particulate Matter on Allergic Hosts
6 Relatively little is known about the effects of inhaled particles on humoral (antibody) or
7 cell-mediated immunity. Alterations in the response to a specific antigenic challenge have been
8 observed in animal models at high concentrations of acid sulfate aerosols (above 1,000 Aig/m3)
9 (Pinto et al., 1979; Kitabatake et al., 1979; Fujimaki et al., 1992). Several studies have reported
10 an enhanced response to nonspecific bronchoprovocation agents, such as acetylcholine and
11 histamine, after exposure to inhaled particles. This nonspecific airway hyperresponsiveness,
12 a central feature of asthma, occurs in animals and human subjects exposed to sulfuric acid under
13 controlled conditions (Gearhart and Schlesinger, 1986; Utell et al., 1983). Although, its
14 relevance to specific allergic responses in the airways of atopic individuals is unclear, it
15 demonstrates that the airways of asthmatics may become sensitized to either specific or
16 nonspecific triggers that could result in increases in asthma severity and asthma-related hospital
17 admissions (Peters et al., 1997; Jacobs et al, 1997; Lipsett et al., 1997).
18 Nel et al. (1998) have suggested that the rise in the U.S. prevalence rate for allergic rhinitis
19 (5% in the 1950s to about 20% in the 1980s) may be related to increased diesel particulate matter
20 (DPM), in addition to other combustion related PM. Combustion particles also may serve as
21 carrier particles for allergens (Knox et al., 1997).
22 A number of in vivo and in vitro studies have demonstrated that DPM can alter the immune
23 response to challenge with specific antigens and suggest that DPM may act as an adjuvant.
24 These studies have shown that treatment with DPM enhances the secretion of antigen-specific
25 IgE in mice (Takano et al., 1997) and in the nasal cavity of human subjects (Diaz-Sanchez et al.,
26 1996, 1997; Ohtoshi et al., 1998). Because IgE levels play a major role in allergic asthma
27 (Wheatley and Platts-Mills, 1996), upregulation of its production could lead to an increased
28 response to inhaled antigen in particle-exposed individuals.
29 Only a small number of studies have examined the mechanisms underlying the
30 enhancement of allergic asthma by ambient urban particles. Ohtoshi et al. (1998) reported that a
31 coarse size-fraction of resuspended ambient PM, collected in Tokyo, induced the production of
March 2001 8-41 DRAFT-DO NOT QUOTE OR CITE
-------
1 granulocyte macrophage colony stimulating factor (GMCSF), an upregulator of dendritic cell
2 maturation and lymphocyte function, in human airway epithelial cells in vitro. In addition to
3 increased GMCSF, epithelial cell supernatants contained increased IL-8 levels when incubated
4 with DPM, a principal component of ambient particles collected in Tokyo. Although the sizes of
5 the two types of particles used in this study were not comparable, the results suggest that ambient
6 PM, or at least the DPM component of ambient PM, can upregulate the immune response to
7 inhaled antigen through GMCSF production. Similarly, Takano et al. (1998) has reported airway
8 inflammation, airway hyperresponsiveness, and increased GM-GSF and IL-5 in mice exposed to
9 diesel exhaust.
10 Gavett et al. (1999) have investigated the effects of ROFA (intratracheal instillation) in
11 ovalbumin (OVA) sensitized and challenged mice. Instillation of 3 mg/kg (approximately 60 /ug)
12 ROFA induced inflammatory and physiological responses in the OVA mice that were related to
13 increases in Th2 cytokines (IL-4, IL-5). ROFA induced greater than additive increases in
14 eosinophil numbers and in airway responsiveness to methylcholine.
15 Hamada et al. (1999, 2000) have examined the effect of a ROFA leachate aerosol in a
16 neonatal mouse model of allergic asthma. In the first study, neonatal mice sensitized by ip
17 injection with OVA developed airway hyperresponsiveness, eosinophilia, and elevated serum
18 anti-ovalbumin IgE after a challenge with inhaled OVA. Exposure to the ROFA leachate aerosol
19 had no marked effect on the airway responsiveness to inhaled methacholine in nonsensitized
20 mice, but did enhance the airway hyperresponsiveness to methylcholine produced in
21 OVA-sensitized mice. No other interactive effects of ROFA exposure with OVA were observed.
22 In a subsequent study, Hamada et al. clearly demonstrated that, whereas inhaled OVA alone was
23 not sufficient to sensitize mice to a subsequent inhaled OVA challenge, pretreatment with a
24 ROFA leachate aerosol prior to the initial exposure to aerosolized OVA resulted in an allergic
25 response to the inhaled OVA challenge. Thus, exposure to a ROFA leachate aerosol can alter the
26 immune response to inhaled OVA both at the sensitization stage at an early age and at the
27 challenge stage.
28 Lambert et al. (1999) also examined the effect of ROFA on a rodent model of pulmonary
29 allergy. Rats were instilled intratracheally with 200 or 1,000 f^g ROFA 3 days prior to
30 sensitization with house dust mite antigen. HDM sensitization after 1000 /^g ROFA produced
31 increased eosinophils, LDH, BAL protein, and IL-10 relative to HDM alone. The immediate
March 2001 8-42 DRAFT-DO NOT QUOTE OR CITE
-------
1 bronchoconstrictive and associated antigen-specific IgE response to a subsequent antigen
2 challenge was increased in the ROFA-treated group in comparison with the control group.
3 Together, these studies suggest the components of ROFA can augment the immune response to
4 antigen. Evidence that metals are responsible for the ROFA-enhancement of an allergic
5 sensitization was demonstrated by Lambert et al. (2000). In this follow-up study, Brown Norway
6 rats were instilled with 1 mg ROFA or the three main metal components of ROFA (iron,
7 vanadium, or nickel) prior to sensitization with instilled house dust mite. The three individual
8 metals were found to augment different aspects of the immune response to house dust mite.
9 Nickel and vanadium produced an enhanced immune response to the antigen as seen by higher
10 house dust mite-specific IgE serum levels after an antigen challenge at 14 days after sensitization.
11 Nickel and vanadium also produced an increase in the lymphocyte proliferative response to
12 antigen in vitro. In addition, the antigen-induced bronchoconstrictive response was greater only
13 in nickel-treated rats. Thus, instillation of metals at concentrations equivalent to those present in
14 the ROFA leachate mimicked the response to ROFA, suggesting that the metal components of
15 ROFA are responsible for the increased allergic sensitization observed in ROFA-treated animals.
16 Goldsmith et al. (1999) have compared the effect of inhalation of concentrated ambient PM
17 for 6 h/day for 3 days versus the effect of a single exposure to a ROFA leachate aerosol on the
18 airway responsiveness to methylcholine in OVA-sensitized mice. Daily exposure to ROFA
19 leachate aerosols significantly enhanced the airway hyperresponsiveness in OVA-sensitized
20 mice. Importantly, exposure to concentrated ambient PM (average concentration of 787 ,wg/m3)
21 had no effect on airway responsiveness in six separate experiments. Thus, the effect of the
22 ROFA leachate aerosols on the induction of airway hyperresponsiveness in allergic mice was
23 significantly different than that of a high concentration of concentrated ambient PM. Although
24 airway responsiveness was examined at only one postexposure time point, these findings do
25 suggest that a great deal of caution should be used in interpreting the results of studies using
26 ROFA particles or leachates in the attempt to investigate the biologic plausibility of the adverse
27 health effects of PM.
28 Several other studies have examined in greater detail the contribution to allergic asthma of
29 the particle component and the organic fraction of DPM. Tsien et al. (1997) treated transformed
30 IgE-producing human B lymphocytes in vitro with the organic extract of DPM. The organic
31 phase extraction had no effect on cytokine production but did increase IgE production.
March 2001 8-43 DRAFT-DO NOT QUOTE OR CITE
-------
1 Moreover, these experiments determined that DPM appeared to be acting on cells already
2 committed to IgE production, thus suggesting a mechanism by which the organic fraction of
3 combustion particles can directly affect B cells and influence human allergic asthma.
4 Cultured epithelial cells from atopic asthmatics show a greater response to DPM exposure
5 when compared with cells from nonatopic nonasthmatics. IL-8, GM-CSF, and soluble ICAM-1
6 increased in response to DPM at a concentration of 10 //g/mL DPM (Bayram et al., 1998a,b).
7 This study suggests that particles could modulate airway disease through their actions on airway
8 epithelial cells. This study also suggests that bronchial epithelial cells from asthmatics are
9 different from those of nonasthmatics in regard to their mediator release in response to DPM.
10 Sagai and colleagues (1996) repeatedly instilled mice with DPM for up to 16 weeks and
11 found increased numbers of eosinophils, goblet cell hyperplasia, and nonspecific airway
12 hyperresponsiveness, changes which are central features of chronic asthma (National Institutes of
13 Health, 1997). Takano et al. (1997) extended this line of research and examined the effect of
14 repeated instillation of DPM on the specific response to antigen (OVA) in mice. They observed
15 that antigen-specific IgE and IgG levels were significantly greater in mice repeatedly instilled
16 with both DPM and OVA. Because this upregulation in antigen-specific immunoglobulin
17 production was not accompanied by an increase in inflammatory cells or cytokines in lavage
18 fluid, it would suggest that, in vivo, DPM may act directly on immune system cells, as described
19 in the work by Tsien et al. (1997). Animal studies have confirmed that the adjuvant activity of
20 DPM also applies to the sensitization of Brown Norway rats to timothy grass pollen (Steerenberg
21 etal., 1999).
22 Diaz-Sanchez and colleagues (1996) have continued to study the mechanism of DPM-
23 induced upregulation of allergic response in the nasal cavity of human subjects. In one study, a
24 200 juL aerosol bolus containing 0.15 mg of DPM was delivered into each naris of subjects with
25 or without seasonal allergies, hi addition to increases in IgE in nasal lavage fluid (NAL), they
26 found an enhanced production of IL-4, IL-6, and IL-13, cytokines known to be B cell
27 proliferation factors. The levels of several other cytokines also were increased, suggesting a
28 general inflammatory response to a nasal challenge with DPM. In a following study, these
29 investigators delivered ragweed antigen, alone or in combination with DPM, on two occasions, to
30 human subjects with both allergic rhinitis and positive skin tests to ragweed (Diaz-Sanchez et al.
31 1997). They found that the combined challenge with ragweed antigen and DPM produced
March 2001 8-44 DRAFT-DO NOT QUOTE OR CITE
-------
1 significantly greater antigen-specific IgE and IgG4 in NAL. A peak response was seen at 96 h
2 postexposure. The combined treatment also induced expression of IL-4, IL-5, IL-10, and IL-13,
3 with a concomitant decrease in expression of Thl-type cytokines. Although the treatments were
4 not randomized (antigen alone was given first to each subject), the investigators reported that
5 pilot work showed no interactive effect of repeated antigen challenge on cellular and biochemical
6 markers in NAL. DPM also resulted in the nasal influx of eosinophils, granulocytes, monocytes,
7 and lymphocytes, as well as the production of various inflammatory mediators. The combined
8 DPM plus ragweed exposure did not increase the rhinitis symptoms beyond those of ragweed
9 alone.
10 Blomberg et al. (1998) observed a 10-fold increase in NAL fluid ascorbate concentration
11 after a 1-h exposure to diluted diesel exhaust (300 /ug/m3 particles and 1.6 ppm NO2). However,
12 there were no effects on BAL ascorbate levels. Rudell et al. (1990) had previously shown
13 increased BAL neutrophils in nonsmoking subjects exposed to 100 /^g/m3 of DPM in diesel
14 exhaust (gases were present). Thus, diesel exhaust (particles and gases) can produce an enhanced
15 response to antigenic material in the nasal cavity. Extrapolation of these findings, of enhanced
16 allergic response in the nose, to the lung, would suggest that ambient combustion particles
17 containing DPM may have significant effects on allergic asthma. These studies provide
18 biological plausibility for the exacerbation of allergic asthma associated with episodic exposure
19 to PM. Although DPM may make up only a fraction of the mass of urban PM, because of their
20 small size, DPM may represent a significant fraction of the ultrafine particle mode in urban air,
21 especially in cities and countries that rely heavily on diesel-powered vehicles. It must be noted
22 that the potential contribution of DPM to the rising prevalence in asthma is complicated by the
23 fact that DPM levels have been decreasing over the last decade (CALEPA report). The reported
24 decrease in DPM levels is a result of the increased combustion efficiency of diesel engines. This
25 improvement in diesel engine design also has brought about a significant decrease in the particle
26 size of diesel emissions. Thus, the balance between a decrease in diesel emissions and the
27 production of a potentially more toxic particle size needs further exploration.
28
29 8.4.4 Resistance to Infectious Disease
30 The development of an infectious disease requires both the presence of the appropriate
31 pathogen, as well as host susceptibility to the pathogen. There are numerous specific and
March 2001 8-45 DRAFT-DO NOT QUOTE OR CITE
-------
1 nonspecific anti-microbial host defenses against microbes, and the ability of inhaled particles to
2 modify resistance to bacterial infection could result from a decreased ability to clear or kill
3 microbes. Rodent infectivity models frequently have been used to examine the effect of inhaled
4 particles on host defense and infectivity. Mice or rats are challenged with a bacterial or viral load
5 either before or after exposure to the particles (or gas) of interest; mortality rate, survival time, or
6 bacterial clearance are then examined. A number of studies that have used the infectivity model
7 to assess the effect of inhaled PM were discussed previously (U.S. Environmental Protection
8 Agency, 1982, 1989, 1996a). In general, acute exposure to sulfuric acid aerosols at
9 concentrations up to 5,000 /ug/m3 were not very effective in enhancing mortality in a bacterially
10 mediated murine model. In rabbits, however, sulfuric acid aerosols altered anti-microbial
11 defenses after exposure for 2 h/day for 4 days to 750 //g/m3 (Zelikoff et al., 1994). Acute or
12 short-term repeated exposures to high concentrations of relatively inert particles have produced
13 conflicting results. Carbon black (10,000 /wg/m3) was found to have no effect on susceptibility to
14 bacterial infection (Jakab, 1993), whereas a very high concentration of TiO2 decreased the
15 clearance of microbes and the bacterial response of lymphocytes isolated from mediastinal lymph
16 nodes (Gilmour et al., 1989a,b). In addition, exposure to DPM has been shown to enhance the
17 susceptibility of mice to the lethal effects of some, but not all, microbial agents (Hatch et al.,
18 1985; Hahon et al., 1985). Thus, the pulmonary response to microbial agents has been shown to
19 be altered at relatively high particle concentrations in animal models. Moreover, these effects
20 appear to be highly dependent on the microbial challenge and the test animal studied. Pritchard
21 et al. (1996) observed in rats exposed to particles with a high concentration of metals (e.g.,
22 ROFA), that the increased mortality rate after streptococcus infection was associated with the
23 amount of metal in the PM.
24 Despite the reported association between ambient PM and deaths caused by pneumonia
25 (Schwartz, 1994), there are few recent studies that have examined the mechanisms that may be
26 responsible for the effect of PM on infectivity. In one study, Cohen and colleagues (1997)
27 examined the effect of inhaled vanadium (V) on immunocompetence. Healthy rats were
28 repeatedly exposed to 2 mg/m3 V, as ammonium metavanadate, and then instilled with
29 polyinosinic-polycytidilic acid (poly I:C), a double-stranded polyribonucleotide that acts as a
30 potent immunomodulator. Induction of increases in lavage fluid protein and neutrophils was
31 greater in animals preexposed to V. Similarly, IL-6 and interferon-gamma were increased in
March 2001 8-46 DRAFT-DO NOT QUOTE OR CITE
-------
1 V-exposed animals. Alveolar macrophage function, as determined by zymosan-stimulated
2 superoxide anion production and by phagocytosis of latex particles, was depressed to a greater
3 degree after poly I:C instillation in V-exposed rats as compared to filtered air-exposed rats.
4 These findings provide evidence that inhaled V, a trace metal found in combustion particles and
5 shown to be toxic in vivo in studies using instilled or inhaled ROFA (Dreher et al., 1997;
6 Kodavanti et al., 1997b, 1999), has the potential to inhibit the pulmonary response to microbial
7 agents. It must be taken into consideration that these effects were found at very high
8 occupational exposure concentrations of V, and as with many studies, care must be taken in
9 extrapolating the results to the ambient exposure of healthy individuals or those with preexisting
10 cardiopulmonary disease to trace concentrations (approximately 3 orders of magnitude lower
11 concentration) of metals in ambient PM.
12
13
14 8.5 MECHANISMS OF PARTICIPATE MATTER TOXICITY AND
15 PATHOPHYSIOLOGY: IN VITRO EXPOSURES
16 8.5.1 Introduction
17 The mechanisms that underlie injury from PM exposure are unclear. Section 8.5.2
18 discusses the more recently published in vitro studies that provide an approach toward
19 identifying potential mechanisms by which PM mediates health effects. The remaining sections
20 discuss potential mechanisms in relation to PM characteristics based on these available data.
21
22 8.5.2 Experimental Exposure Data
23 In vitro exposure is a useful technique to provide information on potential hazardous PM
24 constituents and mechanisms of PM injury, especially when only limited quantities of the test
25 material are available. Exposing respiratory cells to particles in vitro not only reduces the
26 amount of material needed for the experiments but also provides an opportunity to investigate the
27 mechanisms of particle toxicity. In addition, in vitro exposure allows the examination of the
28 response to particles in only one or two cell types. Limitations of in vitro studies include
29 difficulty in extrapolating dose-response relationships and from in vitro to in vivo biological
30 response and mechanistic extrapolations. In addition to alterations in physiochemcial
31 characteristics of PM because of the collection and resuspension processes, these exposure
March 2001 8-47 DRAFT-DO NOT QUOTE OR CITE
-------
1 conditions do not simulate the air-cell interface that actually exists within the lungs, and, thus,
2 the exact dosage delivered to target cells is not known. Furthermore, unless an in vitro exposure
3 system that is capable of delivering particles uniformly to monolayers of airway epithelial cells
4 cultured in an air-liquid interface system is used (Chen et al., 1993), the conventional incubation
5 system alters the microenvironment surrounding the cells and may alter the mechanisms of
6 cellular injury induced by these agents.
7 Even with these limitations, in vitro studies do provide an approach to identify potential
8 cellular and molecular mechanisms by which PM mediates health effects. These mechanisms
9 can then be evaluated in vivo. In vitro studies are summarized in Table 8-8.
10
11 8.5.2.1 Ambient Particles
12 Several studies have exposed airway epithelial cells, alveolar macrophages, or blood
13 monocytes to ambient PM to investigate cellular processes such as oxidant generation and
14 cytokine production that may contribute to the pathophysiological response seen in vivo. Among
15 the ambient PM being examined were samples collected from Boston (Goldsmith et al., 1998),
16 North Provo, UT (Ohio et al., 1999a,b), St. Louis, MO (SRM 1648, Dong et al., 1996; Becker
17 and Soukup, 1998), Washington, DC (SRM 1649, Becker and Soukup, 1998), Ottawa, Canada
18 (EHC-93, Becker and Soukup, 1998), Dusseldorf and Duisburg, Germany (Hitzfeld et al., 1997),
19 Mexico City (Bonner et al., 1998), and Terni, Italy (Fabiani et al., 1997).
20 Because soluble metals of ROFA have been shown to be associated with biological effect
21 and toxicity, several studies have investigated whether the soluble components in ambient PM
22 may have the same biological activities. Ambient PM samples collected from North Provo, UT,
23 (during 1981 and 1982) were used to test whether the soluble components or ionizable metals,
24 which accounted for approximately 0.1% of the mass, are responsible for the biological activity
25 of ambient PM. The oxidant generation (thiobarbituric acid reactive products), release of IL-8
26 from BEAS-2B cells, and PMN influx in rats exposed to these samples correlated with sulfate
27 content and the ionizable fraction of these PM samples (Ohio et al., 1999a,b). In addition, these
28 particles stimulated IL-6 and IL-8 production as well as increased IL-8 mRNA and enhanced
29 expression of intercellular adhesion molecule-1 (ICAM-1) in BEAS-2B cells (Kennedy et al.,
30 1998). Cytokine secretion was preceded by activation of nuclear factor kappa B (NF-KB) and
31 was reduced by treatment with superoxide dismutase (SOD), Deferoxamine (DBF), or
March 2001 8-48 DRAFT-DO NOT QUOTE OR CITE
-------
65
g.
O
O
TABLE 8-8. IN VITRO EFFECTS OF PARTICULATE MATTER AND PARTICULATE MATTER CONSTITUENTS
oo
Tl
H
O
O
g
H
O
o
H
m
o
&
o
t-H
H
M
Species, Cell type,
etc.
Human bronchial
epithelial cells, asthmatic
(ASTH) nonasthmatic
(NONA)
Human bronchial
epithelial cells
(smokers)
Human and
rat alveolar
macrophages
Human AM and blood
monocytes
Rat alveolar
macrophages
Particle or
Constituent
DPM
DPM
Four Urban air
particles.
ROFA
DPM
Volcanic ash
Silica
Urban air
particles,
St. Louis SRM
1648;
Washington, DC,
SRM 1649,
Ottawa, Canada,
EHC-93
PMI(,
Mexico City
1993; volcanic
ash (MSHA)
Exposure
Technique Concentration
In vitro 10-100Mg/mL
In vitro 10-lOO^g/mL
In vitro exposure, Urban and DPM;
2x 105 cells exposed 12,27, 111,333, or
for 2 h 1000 Mg/mL
SiO2 and Ti02:
4, 12, 35, or
167^g/mL
Fe2O3 1:1,3.1,
10:1 particles/cell
ratio
In vitro 33 or 100 Mg/mL
In vitro 1-lOOMg/mL
Particle Size
0 4 //m
0 4//m
Urban particles:
0.3-0.4 A^m
DPM: 0.3 Mm
ROFA: 0 5 Mm
Volcanic ash: 1 8 Mm
Silica: 05- 10 Mm
TiO2: <5 Mm
Latex: 3 8 Mm
0 2 to 0.7 Mm
<10,um
Exposure Duration Effect of Particles Reference
2, 4, 6, 24 h DPM caused no gross cellular Bayram et al.
damage. Ciliary beat frequency was ( 1 998a)
attentuated at all doses. DPM
caused IL-8 release at lower dose m
ASTH than NONA. Higher
concentrations of DPM suppressed
IL-8, GM-CSF, and RANTES m
ASTH cells.
24 h DPM attenuated ciliary beating. Bayram et al
Release of IL-8, protein, GM-CSF, ( 1 998b)
and SICAM-1 increased after DPM
exposure.
2 h for cytotoxicity, 16-18 h UAP-induced cytokme production Becker et al.
for cytokine assay; (TNF, IL-6) m AM of both species (1996)
chemiluminescence at that is not related to respiratory
30 minutes burst or transition metals but may be
related to LPS (blocked by
polymyxin B but not DEF)
ROFA induced strong
chemiluminescence but had weak
effects on TNF production.
3, 6, or 18-20 h Phagocytosis was inhibited by UAP Becker and
at 1 8 h. UAP caused decreased Soukup (1 998)
expression of p2-mtegnns involved
m antigen presentation and
phagocytosis.
24 h PM 10 stimulated alveolar Bonner et al.
macrophages to induce up- (1998)
regulation of PDGF « receptor on
myofiboroblasts. Endotoxin and
NHBE cells
ROFA
In vitro
0, 50, or 200 Mg/mL
Analysis at 2 and 24 h
postexposure
metal components of PM,0 stimulate
release of IL-p. This is a possible
mechanism for PM10-induced airway
remodeling.
Increase in expression of the Carter et al
cytokines IL-6, IL-8, and TNF-a; (1997)
inhibition by DMTU or
deferoxamme.
-------
oo
o
6
O
z
o
H
O
c
o
H
W
O
?o
O
H
W
TABLE 8-8 (cont'd). IN VITRO EFFECTS OF PARTICULATE MATTER AND PARTICULATE
MATTER CONSTITUENTS
o
o
k— t
Species, Cell type,
etc
Supercoiled
DNA
Particle or Constituent
PM,,| from Edinburgh,
Scotland
Exposure
Technique
In vitro
Concentration Particle Size
996.2 ±181 8 PM,0
//g/filterin 100//L
Exposure Duration
8h
Effect of Particles
PMK) caused damage to DNA, mediated by
hydroxyl radicals (inhibited by mannitol) and
Reference
Donaldson et
al. (1997)
Rat AM
UAP
DPM
In vitro 50 to 200 //g/mL
Primary cultures of ROFA
RTE
In vitro 5, 10, or 20 p;g/cm2
DPM- 1.1 - 1.3 pirn
UAP: St Louis,
between 1974 and
1976 in a baghouse,
sieved through
200-mesh(125/jm)
Same as Dreher et al.
(1997)
2 h exposure;
supernatant
collected 18 h
postexposure
Analysis at 6 and
24 h
Peripheral blood Organic extract of TSP, In vitro 42.5 ^g extract/m1 N/A, collected from 2h
monocytes Italy (acetone) high-volume sampler
(60 mVh)
iron (DEF). Clear supernatant has all of the
suspension activity. Free radical activity is
derived either from a fraction that is not
centrifugeable on a bench centrifuge or that the
radical generating system is released into
solution.
Dose dependent increase in TNF-a, IL-6, CINC, Dong et al.
MIP-2 gene expression by urban particles but (1996)
not with DPM; cytokme production were not
related to ROS, cytokme production can be
inhibited by polymyxin B; LPS was detected on
UAP but not DPM; endotoxm is responsible for
the cytokme gene expression induced by UAP in
AM..
Particle induced epithelial-cell detachment and Dye et al.
lytic cell injury; alterations in the permeability of (1997)
the cultured RTE cell layer; increase in LDH, G-
6-PDH, gluathione reductase, glutathione S-
transferase; mechanism of ROFA-mduced RTE
cytotoxicity and pulmonary cellular
inflammation involves the development of an
oxidative burden
Superoxide anion generation was inhibited at a Fabiani et al
paniculate concentration of 0.17 mg/mL when (1997)
stimulated with PMA, 50% increase in LDH;
disintegration of plasma membrane.
Rat AM
ROFA, iron sulfate,
nickel sulfate, vanadyl
sulfate
Latex particles with
metal complexed on the
surface
In vitro 001-1 0 mg/mL 3.6^mMMAD
(0.7 x 10'
cells/mL)
Up to 400 mm Increase chemilummescence, inhibited by DEF
and hydroxyl radical scavengers; solutions of
metal sulfates and metal-complexed latex
particles similarly elevated chemilummescence
in a dose-and time-dependent manner.
Ohio et al
(1997a)
-------
TABLE 8-8 (cont'd). IN VITRO EFFECTS OF PARTICULATE MATTER AND PARTICULATE
o
0
0
Species, Cell type,
etc.
NHBE
BEAS-2B
Particle or Constituent
ROFA
MA1TJ
Exposure
Technique Concentration
In vitro 5-200 /j.g/mL
tK UUINSlil UEIN IS
Exposure
Particle Size Duration
3 6 Aim 2 and 24 h
Effect of Particles
mRNA for ferritm did not change, ferritm protein
increase; mRNA for transfemn receptor decreased,
mRNA for lactofemn increased; transfemn
decreased whereas lactofemn increased;
deferoxamine alone increased lactofemn mRNA.
Reference
Ghio et al.
(1998c)
BEAS-2B
respiratory
epithelial cells
BEAS-2B
0X174 RF1 DNA
Oil fly ash
Provo
TSP soluble and
insoluble extract
PM |(1 from Edinburgh,
Scotland
In vitro lOOi/g/mL N/A
In vitro 500 A*g/mL TSP
In vitro 3.7 or 7.5 Aig/mL PM,,,
= Ih
24 h
8 h
Lactofemn binding with PM metal occurred within Ghio et al.
5 mm. V and Fe "">, but not Ni, bound to the (1999b)
lactofemn receptor.
Water soluble fraction caused greater release of IL-8 Ghio et al.
than insoluble fraction The effect was blocked by (1999a)
deferoxamine and presumably because of metals (Fe,
Cu, Zn, Pb)
Significant free radical activity on degrading
supercoiled DNA; mainly because of hydroxyl
radicals (inhibited by manmtol); Fe involvement
(DEF-B conferred protection), more Fe3* was
released compared to Fe2*, especially at pH 4.6 than
at 7.2.
Gilmour et al.
(1996)
D
§
K>
H -
1
O
0
2
o
H
O
O
H
M
o
***^
o
H
Hamster AM ROFA or CAP In vitro 0, 25, 50, 100, or
200 fig/mL
Hamster AM CAP, ROFA, and their In vitro 0-200 mg/mL
water-soluble and
particulate fractions
AMs from female Vanadyl chloride sodium In vitro 10-1000 A
-------
p
TABLE 8-8
IN VITRO EFFECTS OF PARTICULATE MATTER AND PARTICULATE
MATTER CONSTITUENTS
h-*
o
o
~
OO
t-/i
ho
o
>
H
6
o
z;
0
H
O
c
o
H
W
O
?3
n
HH
H
m
Species, Cell type,
etc.
Human PMN
Human AM
Rat AM
BEAS-2B, airway
epithelial cells
Male (Wistar) rat
lung macrophages
Human blood
monocytes and
neutrophils (PMN)
Human airway
epithelium-derived
cell lines BEAS-2B
(S6-subclone)
Human airway
epithelium-derived
cell line BEAS 2B
Particle or
Constituent
Aqueous and organic
extracts of TSP in
Dusseldorf and
Duisburg, Germany
UAP
(#1648, 1649)
Volcanic ash
ROFA
ROFA, 10 samples
with differing metal
composition
ROFA
Urban dust SRM
1 649, TiO,, quartz
Ambient air
particles, carbon
black, oil fly ash,
coal fly ash
ROFA
ROFA
Exposure Exposure
Technique Concentration Particle Size Duration
In vitro 042-0. 78 mg Collected by high Upto35min
dust/mL volume sampler, 90%
<5 Aim, 50% < 1 Aim,
maximum at
0.3-0.45 Aim
Extracted using water
and then
dichloromethane to
yield aqueous and
organic extracts
In vitro 0,25, 100, or Volume median 24 h
200 /ig/mL diameter:
ROFA 1 1 Aim
#1648: 1.4 Aim
#1649- 1.1 //m
volcanic ash 2.3 Aim
In vitro Oor50Aig/mL 1.99-255^m l-6h
MMAD
In vitro 0,0 5, or 2.0 mg in N/A Ih
lOmL
In vitro 0-100 A(g in I mL N/A 18 h
In vitro 100 (tgm N/A 40 mm
02mL
In vitro 0, 6, 1 2, 25, or 1 96 Aim 1 and 24 h
50 A^g/mL
In vitro 2, 20, or 60 Aig/cm2 1 .96 Aim 24-h exposure
Effect of Particles
PM extract alone significantly stimulated the production
and release of ROS in resting but not in zymosan-
stimulated PMN The effects of the PM extracts were
inhibited by SOD, catalase and sodium azide (NaN3);
Zymosan-mduced LCL is inhibited by both types of
extracts, but aqueous extracts have a stronger inhibitory
effect.
ROFA highly toxic, urban PM toxic at 200Aig/mL;
ROFA produced significant apoptosis as low as
25 Aig/mL; UAP produced apoptosis at 100 Aig/mL;
UAP and ROFA also affect AM phenotype:
increased immune stimulatory, whereas decreased
immune suppressor phenotype
Macrophage activation, as determined by
chemilummescence was maximal with the V-rich
particles as opposed to V plus Ni-nch particles.
ROFA induced production of acetaidehyde in dose-
dependant fashion
Cytotoxicity ranking was quartz > SRM 1 649 > TiO,,
based on cellular ATP decrease and LDH, acid
phosphatase, and p-glucuromdase release.
ROS generation, measured by LCL increased in PMN,
was correlated with Si, Fe, Mn, Ti, and Co content but
not V, Cr, Ni, and Cu Deferoxamme, a metal lon-
chelator, and did not affect LCL in PMN, suggesting that
metal ions are not related to the induction of LCL.
Activation of IL-6 gene by NF-KB activation and
binding to specific sequences in promoter of IL-6 gene;
inhibition ofNF-icB activation by DEF and NAC,
increase in PGE,, IL-6, TNF, and IL-8, activation NF-KB
may be a critical first step in the inflammatory cascade
following exposure to ROFA particles.
Epithelial cells exposed to ROFA for 24 h secreted
substantially increased amounts of the PHS products
prostaglandms E2 and F,n; ROFA-mduced increase in
prostaglandin synthesis was correlated with a marked
increase in PHS activity.
Reference
Hitzfeld et al
(1997)
Hohan et al
(1998)
Kodavanti et al
(1998a)
Madden et al
(1999)
Nadeau et al.
(1996)
Prahalad et al
(1999)
Quay et al.
(1998)
Samet et al.
(1996)
-------
O
o
oo
I!A
U)
D
a
o
2
o
H
O
o
H
TABLE 8-8 (cont'd). IN VITRO EFFECTS OF PARTICULATE MATTER AND PARTICULATE
MATTER CONSTITUENTS
Species, Cell type,
etc.
Human airway
epithelium-derived
cell line BEAS
Human airway
epithelium-derived
Particle or
Constituent
ROFA
Synthetic ROFA
(soluble Ni, Fe,
and V)
Particle
components As, Cr,
Exposure
Technique
In vitro
In vitro
Concentration
ROFA: 0-200
,ug/mL
Synthetic ROFA
(lOOyug/mL):
Ni, 64 A*M
Fe, 63 ^M
V, 370mM
500 ^M of As, F,
Cr (III), Cu, V, Zn
Particle Size
ROFA- 1.96 Aim
Synthetic ROFA- N/A
(soluble)
N/A (soluble)
Exposure
Duration
Up to 24 h
20 min and
6 and 24 h
Effect of Particles
Tyrosme phosphatase activity, which was known to be
inhibited by vanadium ions, was markedly diminished
after ROFA treatment; ROFA exposure induces
vanadium ion-mediated inhibition of tyrosine
phosphatase activity, leading to accumulation of protein
phosphotyrosmes in BEAS cells.
Noncytotoxic concentrations of As, V, and Zn induced
a rapid phosphorylation of MAPK in BEAS cells;
Reference
Samet et al.
(1997)
Samet et al.
(1998)
cell lines BEAS-2B Cu, Fe, Ni, V, and
Zn
activity assays confirmed marked activation of ERK,
JNK, and P38 in BEAS cells exposed to As, V, and Zn
Cr and Cu exposure resulted in a relatively small
activation of MAPK, whereas Fe and Ni did not activate
MAPK under these conditions; the transcription factors
c-Jun and ATF-2, substrates of JNK and P38,
respectively, were markedly phosphorylated in BEAS
cells treated with As, Cr, Cu, V, and Zn; acute exposure
to As, V, or Zn that activated MAPK was sufficient to
induce a subsequent increase in IL-8 protein expression
in BEAS cells.
A549
0X174RF1DNA
Rat (Wistar) AM
RAM cells (a rat AM
cell line)
A549
Urban particles
SRM 1648,
St Louis
SRM 1649,
Washington, DC
TiO2
ROFA, a-quartz,
TiO2
In vitro 1 mg/mL for Fe
mobilization assay
In vitro 20, 50, or 80 ^g/mL
In vitro 1 mg/mL
SRM 1648: 50% <
10 //m
SRM 1649: 30% <
10 pirn
N/A
N/A
Up to 25 h Single-strand breaks in DNA were induced by PM only
in the presence of ascorbate, and correlated with amount
of Fe that can be mobilized; ferritm m A549 cells was
increased with treatment of PM suggesting mobilization
of Fe in the cultured cells.
4 h Opsomzation of TiO2 with surfactant components
resulted in a modest increase in AM uptake compared
with that of unopsomzed TiO2; surfactant components
increase AM phagocytosis of particles.
60 mm Exposure of A549 cells to ROFA, a-quartz, but not TiO2,
caused increased IL-8 production in TNF-ct primed cells
in a concentration-dependent manner.
Smith and
Aust(1997)
Stringer and
Kobzik
(1996)
Stnnger and
Kobzik
(1998)
O
h-H
H
W
-------
2 TABLE 8-8 (cont'd). IN VITRO EFFECTS OF PARTICULATE MATTER AND PARTICULATE
1 MATTER CONSTITUENTS
£^ Species, Cell type,
O £tc-
A549
RLE-6TN cells
(type II like cell line)
rat, Long Evans
epithelial cells
°° BEAS-2B human
V| bronchial epithelial
cells
Particle or
Constituent
TiO2, Fe:O3, CAP,
and the fibrogenic
particle a-quartz
PM, 5, Burlington,
VT;"
Fme/ultrafine TiO,
CFA
PFA
a-quartz
ROFA
Birmingham, AL.
188mg/gofVO
Exposure
Technique Concentration Particle Size
In vitro Ti02 [40 Mg/mL], N/A
Fe20, [lOO^g/mL],
a-quartz
[200 Mg/mL], or
CAP [40 Mg/mL]
In vitro 1,2.5,5, or PM25:39nm
10Mg/mL FmeTiOj 159nm
UFTiO2-37nm
17 7 Mm
2 5 //m
In vitro lOO^g/mL N/A
Exposure
Duration Effect of Particles
24 h TiO2 > Fe2Oj > a-quartz > CAP in particle binding;
binding of particle was found to be calcium-dependent
for TiO2 and Fe,O5, while a-quartz binding was
calcium-independent, scavenger receptor, mediate
paniculate binding; a-quartz, but not TiO2 or CAP,
caused a dose-dependent production of IL-8
24 and 48 h Increases in c-Jun kmase activity, levels of
exposure phosphorylated c-Jun immunoreactive protein, and
transcnptional activation of activator protein- 1-
dependent gene expression; elevation in number of
cells incorporating 5'-bromodeoxyundme.
3 h CFA produced highest level of hydroxyl radicals, iron
content is more important than quartz content.
2-6 h ROFA caused increased intracellular Ca", IL-6, IL-8,
and TNF-a through activation of capsicm- and
pH-sensitive receptors.
Reference
Stringer et al
(1996)
Timblin et al
(1998)
Van Maanen
etal. (1999)
Veronesi et al.
(1999)
"fl
H
6
o
z
o
0
G
O
H
W
O
70
O
H-H
H
m
-------
1 N-acetylcysteine. The addition of similar quantities of Cu2+ as found in the Provo extract
2 replicated the biological effects observed with particles alone. When normal constituents of
3 airway lining fluid (mucin or ceruloplasmin) were added to BEAS cells, particulate-induced
4 secretion of IL-8 was modified. Mucin reduced IL-8 secretion, whereas ceruloplasmin
5 significantly increased IL-8 secretion and activation of NF-KB. The authors suggest that copper
6 ions may cause some of the biologic effects of inhaled PM in the Provo region and may provide
7 an explanation for the sensitivity of asthmatics to Provo PM seen in epidemiologic studies.
8 There are regional as well as daily variations in the composition of ambient PM and, hence,
9 its biological activities. For example, concentrated ambient PM (CAP, from Boston urban air)
10 has substantial day-to-day variability in its composition and oxidant effects (Goldsmith et al.,
11 1998). Similar to Utah PM, the water-soluble component of Boston CAPs significantly
12 increased AM oxidant production and inflammatory cytokine (MIP2 and TNFa) production over
13 negative control values. These effects can be blocked by metal chelators or antioxidants. The
14 regional difference in biological activity of ambient PM has been shown by Becker and Soukup
15 (1998). The oxidant generation, phagocytosis, as well as the expressions of receptors important
16 for phagocytosis in human alveolar macrophage and blood monocyte were reduced significantly
17 by PM exposure.
18 Becker and Soukup (1998) and others (Dong et al., 1996, Becker et al., 1996) have
19 suggested that the biological activity of the ambient PM may result from the presence of
20 endotoxin on the particles rather than metal-associated oxidant generation. Using the same urban
21 particles (SRM 1648), cytokine production (TNF-a, IL-1,11-6, CINC, and MIP-2) was increased
22 in macrophages following treatment with 50 to 200 /^g/mL of urban PM (Dong et al., 1996). The
23 urban particle-induced TNF-a secretion was abrogated completely by treatment with polymyxin
24 B, an antibiotic that blocks LPS-associated activities, but not with antioxidants. Although it is
25 possible that LPS may be responsible for ambient PM induced cytokine gene expression,
26 extrapolation of these in vitro results to a potential role for endotoxin in the adverse effects of
27 ambient PM must be done with caution because the investigators could not exclude the
28 possibility that the presence of endotoxin with the PM was caused by inadvertent contamination
29 during the year-long collection process or from the handling of the particles.
30 The involvement of endotoxin, at least partially, in PM induced biological effects was
31 supported more recently by Bonner et al. (1998). Urban PM10 collected from north, south, and
March 2001 8-55 DRAFT-DO NOT QUOTE OR CITE
-------
1 central regions of Mexico City was used with SD rat AM to examine PM effects on platelet
2 derived growth factor (PDGF) receptors on lung myofibroblasts (Bonner et al., 1998).
3 Mexico City PM10 (but not volcanic ash) stimulated secretion of upregulatory factors for the
4 PDGF a receptor, possibly via IL-lp. In the presence of an endotoxin-neutralizing protein, the
5 Mexico City PM10 effect on PDGF was blocked partially, suggesting that LPS was responsible
6 partially for the effect of the PM10 on macrophages. In addition, both LPS and vanadium (both
7 present in the PM10) acted directly on lung myofibroblasts. However, the V levels in Mexico
8 City PM10 were probably not high enough to exert an independent effect. The authors concluded
9 that PM10 exposure could lead to airway remodeling by enhancing myofibroblast replication and
10 chemotaxis.
11 The effects of water soluble as well as organic components (extracted in dichloromethane)
12 of ambient PM were investigated by exposing human PMN to PM extracts (Hitzfeld et al., 1997).
13 PM was collected with high-volume samplers in two German cities, Dusseldorf and Duisburg;
14 these sites have high traffic and high industrial emissions, respectively. Organic, but not
15 aqueous, extracts of PM alone significantly stimulated the production and release of ROS in
16 resting human PMN. The effects of the PM extracts were inhibited by SOD, catalase, and
17 sodium azide (NaN3). Similarly, the organic fraction (extractable by acetone) of ambient PM
18 from Terni, Italy, had been shown to produce cytotoxicity, superoxide release in response to
19 PMA and zymosan in peripheral monocytes (Fabiani et al., 1997).
20
21 8.5.2.2 Residual Oil Fly Ash
22 In a series of studies using the same ROFA samples, several experiments have investigated
23 the biochemical and molecular mechanisms involved in ROFA induced cellular injury.
24 Prostaglandin metabolism in cultured human airway epithelial cells (BEAS-2B and NHBE)
25 exposed to ROFA was investigated by Samet et al. (1996). Epithelial cells exposed to ROFA for
26 24 h secreted substantially increased amounts of prostaglandins E2 and F2 a. The ROFA-
27 induced increase in prostaglandin synthesis was correlated with a marked increase in activity of
28 the PHS-2 form of prostaglandin H synthase as well as mRNA coded for this enzyme.
29 In contrast, expression of the PHS1 form of the enzyme was not affected by ROFA treatment of
30 airway epithelial cells. These investigators further demonstrated that noncytotoxic levels of
31 ROFA induced a significant dose- and time-dependent increase in protein tyrosine phosphate, an
March 2001 8-56 DRAFT-DO NOT QUOTE OR CITE
-------
1 important regulator of signal transduction leading to cell growth and proliferation. ROFA-
2 induced increases in protein phosphotyrosines were associated with its soluble fraction and were
3 mimicked by V-containing solutions but not iron or nickel solutions (Samet et al., 1997).
4 ROFA also stimulates respiratory cells to secret inflammatory cytokines such as IL-6, IL-8,
5 and TNF. Normal human bronchial epithelial (NHBE) cells exposed to ROFA produced
6 significant amounts of IL-8, IL-6, and TNF, as well as mRNAs coding for these cytokines (Carter
7 et al., 1997). Increases in cytokine production, but not m-RNA expression, were dose-dependent.
8 The cytokine production was inhibited by the addition of metal chelator, DBF, or the free radical
9 scavenger, DMTU. Similar to the data of Samet et al. (1997), V but not Fe or Ni compounds
10 were responsible for these effects. Cytotoxicity, decreased cellular glutathione levels in primary
11 cultures of rat tracheal epithelial (RTE) cells exposed to suspensions of ROFA indicated that
12 respiratory cells exposed to ROFA were under oxidative stress. Treatment with buthionine
13 sulfoxamine (an inhibitor of y-glutamyl cysteine synthetase) augmented ROFA-induced
14 cytotoxicity, whereas treatment with DMTU inhibited ROFA-induced cytoxicity further
15 suggested that ROFA-induced cell injury may be mediated by hydroxyl-radical-like ROS (Dye
16 et al., 1997). Using BEAS-2B cells, a time- and dose-dependent increase in IL-6 mRNA induced
17 by ROFA was shown to precede by the activation of nuclear proteins NF-kB (Quay et al., 1998).
18 Taking together, ROFA exposure increases oxidative stress, perturbs protein tyrosine phosphate
19 homeostasis, activates NF-kB, and up-regulates inflammatory cytokine and prostaglandin
20 synthesis and secretion to produce lung injury.
21 Stringer and Kobzik (1998) observed that "primed" lung epithelial cells exhibited enhanced
22 cytokine responses to PM. Compared to normal cells, exposure of TNF-cc-primed A549 cells to
23 ROFA or a -quartz caused increased IL-8 production in a concentration-dependent manner for
24 particle concentrations ranging from 0-200 /^g/mL. Addition of the antioxidant NAC (1.0 mM)
25 decreased ROFA and a -quartz-mediated IL-8 production by approximately 50% in both normal
26 and TNF- a -primed A549 cells. Exposure of A549 cells to ROFA caused an increase in oxidant
27 levels that could be inhibited by NAC. These data suggest that (1) lung epithelial cells primed by
28 inflammatory mediators show increased cytokine production after exposure to PM, and
29 (2) oxidant stress is an important mechanism for this response.
30 In summary, exposure of lung cells to ambient PM or ROFA leads to increased production
31 of cytokines and the effects may be mediated, at least in part, through production of ROS.
March 2001 8-57 DRAFT-DO NOT QUOTE OR CITE
-------
1 Day-to-day variations in the components of PM, such as soluble transition metals, which may be
2 critical to eliciting the response, are suggested. The involvement of organic components in
3 ambient PM also was suggested in some studies.
4
5 8.5.3 Potential Cellular and Molecular Mechanisms
6 8.5.3.1 Reactive Oxygen Species
7 Ambient particulate matter contains transition metals, such as iron (most abundant),
8 copper, nickel, vanadium, and cobalt. These metals are capable of catalyzing the one-electron
9 reductions of molecular oxygen necessary to generate reactive oxygen species (ROS). These
10 reactions can be demonstrated by the iron-catalyzed Haber-Weiss reactions that follow.
11
12
13 Reductant" + Fe(III) -> Reductantn+1 + Fe(II) (1)
14 Fe(II) + 0~2^> Fe(III) + O2 (2)
15 HO2+O~+H+^ O2+H2O2 (3)
16 Fe(II) + H2O2 -> Fe(III)+*OH + HO"(Fenton Reaction) (4)
17
18 Iron will continue to participate in the redox cycle in the above reactions as long as there is
19 sufficient O2 or H2O2 and reductants.
20 Soluble metals from inhaled PM dissolved into the fluid lining of the airway lumen can
21 react directly with biological molecules (acting as a reductant in the above reactions) to produce
22 ROS. For example, ascorbic acid in the human lung epithelial lining fluid can react with Fe(III)
23 from inhaled PM to cause single strand breaks in supercoiled plasmid DNA, (j>X174 RFI (Smith
24 and Aust, 1997). The DNA damage caused by a PM10 suspension can be inhibited by mannitol,
25 an hydroxyl radical scavenger, further confirming the involvement of free radicals in these
26 reactions (Gilmour et al., 1996; Donaldson et al., 1997; Li et al., 1997). Because the clear
27 supernatant of the centrifuged PM10 suspension contained all of the suspension activity, the free
28 radical activity is derived either from a fraction that is not centrifugable (10 min at 13,000 rpm
March 2001 8-58 DRAFT-DO NOT QUOTE OR CITE
-------
1 on a bench centrifuge) or the radical generating system is released into solution (Gilmour et al.,
2 1996; Donaldson et al., 1997; Li et al., 1997).
3 In addition to measuring the interactions of ROS and biomolecules directly, the role of
4 ROS in PM-induced lung injury also can be assessed by measuring the electron spin resonance
5 (ESR) spectrum of radical adducts or fluorescent intensity of dichlorofluorescin (DCFH), an
6 intracellular dye that fluoresces on oxidation by ROS. Alternatively, ROS can be inhibited using
7 free radical scavengers, such as dimethylthiourea (DMTU); antioxidants, such as glutathione or
8 N-acetylcysteine (NAC); or antioxidant enzymes, such as superoxide dismutase (SOD). The
9 diminished response to PM after treatment with these antioxidants indicates the involvement of
10 ROS.
11 As described earlier, Kadiiska et al. (1997) used the ESR spectra of 4-POBN [a-(4-pyridyl
12 l-oxide)-N-tert-butylnitrone] adducts to measure ROS in rats instilled with ROFA and
13 demonstrated the association between ROS production within the lung and soluble metals in
14 ROFA. Using DMTU to inhibit ROS production, Dye et al. (1997) had shown that systemic
15 administration of DMTU impeded development of the cellular inflammatory response to ROFA,
16 but did not ameliorate biochemical alterations in BAL fluid. Goldsmith et al. (1998), as
17 described earlier, showed that ROFA and CAPs caused increases in ROS production in AMs.
18 The water-soluble component of both CAPs and ROFA significantly increased AM oxidant
19 production over negative control values. In addition, increased PM-induced cytokine production
20 was inhibited by NAC. Li et al. (1996, 1997) instilled rats with PM10 particles (collected on
21 filters from an Edinburgh, Scotland, monitoring station). Six hours after intratracheal instillation
22 of PM10, they observed a decrease in glutathione (GSH) levels in the BAL fluid. Although this
23 study does not describe the composition of the PM10, the authors suggest that changes in GSH, an
24 important lung antioxidant, support the contention that the free radical activity of PM10 is
25 responsible for its biological activity in vivo.
26 In addition to ROS generated directly by PM, resident or newly recruited AMs or PMNs
27 also are capable of producing these reactive species on stimulation. The ROS produced during
28 the oxidative burst can be measured using a chemiluminescence (CL) assay. With this assay,
29 AM CL signals in vitro have been shown to be greatest with ROFA containing primarily soluble
30 V and were less with ROFA containing Ni plus V (Kodavanti et al., 1998a). As described
31 earlier, exposures to Dusseldorf and Duisburg PM increased the resting ROS production in
March 2001 8-59 DRAFT-DO NOT QUOTE OR CITE
-------
1 PMNs, which could be inhibited by SOD, catalase, and sodium azide (Hitzfeld et al., 1997).
2 Stringer and Kobzik (1998) showed that addition of NAC (1.0 mM) decreased ROFA-mediated
3 IL-8 production by approximately 50% in normal and TNF-a-primed A549 cells. In addition,
4 exposures of A549 cells to ROFA caused a substantial (and NAC inhibitable) increase in oxidant
5 levels as measured by DCFH oxidation. In human AMs, Becker et al. (1996) found a CL
6 response for ROFA, but not urban air particles (Ottawa and Dusseldorf) or volcanic ash.
7 Metal compounds of PM are the most probable species capable of catalyzing ROS
8 generation on exposure to PM. To determine elemental content and solubility in relation to their
9 ability to generate ROS, PMN or monocytes were exposed to a wide range of ambient air
10 particles from divergent sources (one natural dust, two types of oil fly ash, two types of coal fly
11 ash, five different ambient air samples, and one carbon black sample) (Prahalad et al., 1999), and
12 CL production was measured over a 20-min period postexposure. Percent of sample mass
13 accounted for by XRF detectable elements was 1.2% (carbon black); 22 to 29% (natural dust and
14 ambient air particles); 13 to 22% (oil fly ash particles); and 28 to 49% (coal fly ash particles).
15 The major proportion of elements in most of these particles were aluminosilicates and insoluble
16 iron, except oil derived fly ash particles in which soluble vanadium and nickel were in highest
17 concentration, consistent with particle acidity as measured in the supernatants. All particles
18 induced CL response in cells, except carbon black. The CL response of PMNs in general
19 increased with all washed particles, with oil fly ash and one urban air particle showing statistical
20 differences between deionized water washed and unwashed particles. These CL activities were
21 significantly correlated with the insoluble Si, Fe, Mn, Ti, and Co content of the particles.
22 No relationship was found between CL and soluble transition metals such as V, Cr, Ni, and Cu.
23 Pretreatment of the particles with a metal ion chelator, deferoxamine, did not affect CL activities.
24 Particle sulfate content and acidity of the particle suspension did not correlated with CL activity.
25 Soluble metals can be mobilized into the epithelial cells or AMs to produce ROS
26 intracellularly. Size fractionated coal fly ash particles (2.5, 2.5 to 10, and <10 yum) of bituminous
27 b (Utah coal), c (Illinois coal), and lignite (Dakota coal) were used to compare the amount of iron
28 mobilization in A549 cells and by citrate (I mM) in cell-free suspensions (Smith et al., 1998).
29 Iron was mobilized by citrate from all three size fractions of all three coal types. More iron, in
30 Fe(III) form, was mobilized by citrate from the <2.5-p:m fraction than from the >2.5-p:m
31 fractions. In addition, the amount of iron mobilized was dependent on the type of coal used to
March 2001 8-60 DRAFT-DO NOT QUOTE OR CITE
-------
1 generate the fly ash (Utah coal > Illinois coal = Dakota coal) but not related to the total amount
2 of iron present in the particles. Ferritin (an iron storage protein) levels in A549 cells increased by
3 as much as 11.9-fold in cells treated with coal fly ash (Utah coal > Illinois coal > Dakota coal).
4 More ferritin was induced in cells treated with the <2.5-/^m fraction than with the >2.5-/um
5 fractions. Mossbauer spectroscopy of a fly ash sample showed that the bioavailable iron was
6 assocated with the glassy aluminosilicate fraction of the particles (Ball et al., 2000). As with the
s
7 bioavailability of iron, there was an inverse correlation between the production of IL-8 and fly
8 ash particle size with the Utah coal fly ash being the most potent.
9 Using ROFA and colloidal iron oxide, Ohio et al. (1997b; 1998a,b,c; 1999c; 2000b) have
10 shown that exposures to these particles disrupted iron homeostasis and induced the production of
11 ROS in vivo and in vitro. Treatment of animals or cells with metal-chelating agents such as DBF
12 with an associated decrease in response has been used to infer the involvement of metal in PM-
13 induced lung injury. Metal chelation by DEF (1 mM) caused significant inhibition of particulate-
14 induced AM oxidant production, as measured using DCFH (Goldsmith et al., 1998). DEF
15 treatment also reduced NF-KB activation and cytokine secretion in BEAS-2B cells exposed to
16 Provo PM (Kennedy et al., 1998). However, treatment of ROFA suspension with DEF was not
17 effective in blocking leachable metal induced acute lung injury (Dreher et al., 1997). Dreher
18 et al. (1997) indicated that DEF could chelate Fe(III) and V(II), but not Ni(II), suggesting that metal
19 interactions played a significant role in ROFA-induced lung injury.
20 Other than Fe, several V compounds have been shown to increase mRNA levels for
21 selected cytokines in BAL cells and also to induce pulmonary inflammation (Pierce et al., 1996).
22 NaVO3 and VOSO4, highly soluble forms of V, tended to induce pulmonary inflammation and
23 inflammatory cytokine mRNA expression more rapidly and more intensely than the less soluble
24 form, V2O5, in rats. Neutrophil influx was greatest following exposure to VOSO4 and lowest
25 following exposure to V2O5. However, metal components of fly ash have not been shown to
26 consistently increase ROS production from bovine AM treated with combustion particles
\
27 (Schluter et al., 1995). For example, As(III), Ni(II), and Ce(III), which are major components of
28 fly ash, had been shown to inhibit the secretion of superoxide anions (O2~) and hydrogen
29 peroxide. In the same study, O2~ were lowered by Mn(II) and Fe(II), whereas V(IV) increased O2~
30 and H2O2. In contrast, Fe(III) increase O2" productions, demonstrating that the oxidation state of
March 2001 8-61 DRAFT-DO NOT QUOTE OR CITE
-------
1 metal may influence its oxidant generating properties. Other components of fly ash, such as
2 Cd(II), Cr(III), and V(V), had no effects on ROS.
3 It is likely that a combination of several components rather than a single metal in PM is
4 responsible for the PM induced cellular response. For example, V and Ni+V but not Fe or Ni
5 alone (in saline with the final pH at 3.0) resulted in increased epithelial permeability, decreased
6 cellular glutathione, cell detachment, and lytic cell injury in rat tracheal epithelial cells exposed
7 to soluble salts of these metals at equivalent concentrations found in ROFA (Dye et al., 1999).
8 Treatment of V-exposed cells with buthionine sulfoximine further increased cytotoxicity.
9 Conversely, treatment with radical scavenger dimethyl thiourea inhibited the effects in a dose-
10 dependent manner. These results showed that soluble metal or combinations of several metals in
11 ROFA are responsible for these effects.
12 Similar to combustion particles such as ROFA, the biological response to exposure to
13 ambient PM also appear to depend on the metal content of the particles. Human subject were
14 instilled with 500 fj.g (in 20 mL sterile saline) of Utah Valley dust (UVD1, 2, 3, collected during
15 3 successive years) on the left segmental bronchus and on the right side with sterile saline as
16 control. Twenty-four-hour postinstillation, a second bronchoscopy was performed and
17 phagocytic cells were obtained on both side of the segmental bronchus. AM from subjects
18 instilled with UVD, obtained by bronchoaveolar lavage 24 h postinstillation, were incubated with
19 fluoresceinated yeast (Saccharomyces cerevisiae) to assess their phagocytic ability. Although the
20 same proportion of AMs were exposed to UVD phagocytized yeast, AMs exposed to UVD1,
21 which were collected while a local steel mill was open, took up significantly less particles than
22 AMs exposed to other extracts (UVD2 when the steel mill was closed and UVD3 when the plant
23 reopened). AMs exposed to UVD1 also exhibited a small decrease in oxidant activity (using
24 dihydrorhodamine-123, DHR). AMs from healthy volunteers were incubated in vitro with the
25 various UVD extracts to assess whether similar effects on human AMs function could be
26 observed to those seen following in vivo exposure. The percentage of AMs that engulfed yeast
27 particles was significantly decreased by exposure to UVD1 at 100 //g/mL, but not at 25 //g/mL.
28 However, the amount of particles engulfed was the same following exposure to all three UVD
29 extracts. AMs also demonstrated increased oxidant stress (using chemiluminescence) after in
30 vitro exposure to UVD1 and this effect was not abolished with pretreatment of the extract with
31 the metal chelator deferoxamine. As with the AMs exposed to UVD in vivo, AM exposed to
March 2001 8-62 DRAFT-DO NOT QUOTE OR CITE
-------
1 UVD in vitro had a decreased oxidant activity (DHR assay). UVD1 contains 61 times and
2 2 times the amount of Zn compared to UVD 2 and UVD3, respectively, whereas UVD3
3 contained 5 times more Fe than UVD1. Ni and V were present only in trace amounts. Using the
4 same particles, Frampton et al. (1999) exposed BEAS-2B cells for 2 and 24 h. Similar results
5 were observed for oxidant generation in these cells (i.e., UVD 2, which contains the lowest
6 concentrations of soluble iron, copper, and zinc, produced the least response). Only
7 UVD 3 produced cytotoxicity at a dose of 500 //g/mL. UVD 1 and 3, but not 2, induced
8 expression of IL-6 and 8 in a dose-dependent fashion. Taken together, these data showed that
9 biological response to ambient particles exposure is heavily dependent on the source and, hence,
10 the chemical composition of PM.
11
12 8.5.3.2 Intracellular Signaling Mechanisms
13 In has been shown that the intracellular redox state of the cell modulates the activity of
14 several transcription factors, including NF-KB, a critical step in the induction of a variety of
15 proinflammatory cytokine and adhesion-molecule genes. NF-KB is a heterodimeric protein
16 complex that in most cells resides in an inactive state in the cell cytoplasm by binding to
17 inhibitory kappa B alpha (IkBcc). On appropriate stimulation by cytokines or ROS, hcBcc is
18 phosphorylated and subsequently degraded by proteolysis. The dissociation of iKBa from NF-KB
19 allows the latter to translocate into the nucleus and bind to appropriate sites in the DNA to
20 initiate transcription of various genes. Two studies in vitro have shown the involvement of
21 NF-KB in particulate-induced cytokine and intercellular adhesion molecule-1 (ICAM-1)
22 production in human airway epithelial cells (BEAS-2B) (Quay et al., 1998; Kennedy et al.,
23 1998). Cytokine secretion was preceded by activation of NF-KB and was reduced by treatment
24 with antioxidants or metal chelators. These results suggest that metal-induced oxidative stress
25 may play a significant role in the initiation phase of the inflammatory cascade following
26 particulate exposure.
27 A second well-characterized human transcription factor, AP-1, also responds to the
28 intracellular ROS concentration. AP-1 exists in two forms, either in a homodimer of c-jun
29 protein or a heterodimer consisting of c-jun and c-fos. Small amounts of AP-1 already exist in
30 the cytoplasm in an inactive form, mainly as phosphorylated c-jun homodimer. Many different
31 oxidative stress-inducing stimuli, such as UV light and IL-1, can activate AP-1. Exposure of rat
March 2001 8-63 DRAFT-DO NOT QUOTE OR CITE
-------
1 lung epithelial cells to ambient PM in vitro resulted in increases in c-jun kinase activity, levels of
2 phosphroylated c-jun immunoreactive protein, and transcriptional activation of AP-1 -dependent
3 gene expression (Timblin et al., 1998). This study demonstrated that interaction of ambient
4 particles with lung epithelial cells initiates a cell signaling cascade related to aberrant cell
5 proliferation.
6 Early response gene transactivation has been linked to the development of apoptosis, a
7 unique type of programmed cell injury and a potential mechanism to account for PM-induced
8 changes in cellular response. Apoptosis of human AMs exposed to ROFA (25 //g/mL) or urban
9 PM was observed by Holian et al. (1998). In addition, both ROFA and urban PM upregulated the
10 expression of the RFD1 + AM phenotype, whereas only ROFA decreased the RFDl+7f phenotype.
11 It has been suggested that an increase in the AM phenotype ratio of RFDl+/RFDl+7f may be
12 related to disease progression in patients with inflammatory diseases. These data showed that
13 ROFA and urban PM can induce apoptosis of human AMs and increase the ratio of AM
14 phenotypes toward a higher immune active state and may contribute to or exacerbate lung
15 inflammation.
16 Another intracellular signaling pathway that can lead to diverse cellular responses such as
17 cell growth, differentiation, proliferation, apoptosis, and stress responses to environmental
18 stimuli, is the phosphorylation-dependent, mitogen-activated protein kinase (MAPK).
19 Noncytotoxic levels of ROFA have been shown to induce significant dose- and time-dependent
20 increases in protein tyrosine phosphate levels in BEAS cells (Samet et al., 1997). In a
21 subsequent study, the effects of As, Cr, Cu, Fe, Ni, V, and Zn on the MAPK, extracellular
22 receptor kinase (ERK), c-jun N-terminal kinase (JNK), and P38 in BEAS cells were investigated
23 (Samet et al., 1998). Noncytotoxic concentrations of As, V, and Zn induced a rapid
24 phosphorylation of MAPK in BEAS cells. Activity assays confirmed marked activation of ERK,
25 JNK, and P38 in BEAS cells exposed to As, V, and Zn. Cr and Cu exposure resulted in a
26 relatively small activation of MAPK, whereas Fe and Ni did not activate MAPK. Similarly, the
27 transcription factors c-Jun and ATF-2, substrates of JNK and P38, respectively, were markedly
28 phosphorylated in BEAS cells treated with As, Cr, Cu, V, and Zn. The same acute exposure to
29 As, V, or Zn that activated MAPK was sufficient to induce a subsequent increase in IL-8 protein
30 expression in BEAS cells. These data suggest that MAPK may mediate metal-induced
31 expression of inflammatory proteins in human bronchial epithelial cells. The ability of ROFA to
March 2001 8-64 DRAFT-DO NOT QUOTE OR CITE
-------
1 induce activation of MAPKs in vitro was demonstrated by Silbajoris et al. (2000). In addition,
2 Gerchen et al. (1996) showed that the ROS production induced by PM was markedly decreased
3 by the inhibition of protein kinase C as well as phospholipase A2.
4 The major cellular response downstream of ROS and the cell signaling pathways described
5 above is the production of inflammatory cytokines or other reactive mediators. In an effort to
6 determine the contribution of cyclooxygenase to the pulmonary responses to ROFA exposure
7 in vivo, Samet et al. (2000) intratracheally instilled Sprague Dawley rats with ROFA (200 or
8 500 fj.g in 0.5 mL saline). These animals were pretreated, intraperitoneally, with 1 mg/kg ROFA
9 (in 20% ethanol in saline) 30 min prior to the intratracheal exposure. At 12 h after intratracheal
10 instillations, intraperitoneal injections (1 mL) were repeated. ROFA treatment induced a marked
11 increase in the level of PGE2 recovered in the BALF, which was effectively decreased by
12 pretreating the animals with specific prostaglandin H synthase 2 (COX2) inhibitor NS398.
13 Immunohistochemical analyses of rat airway showed concomitant expression of COX2 in the
14 proximal airway epithelium of rats treated with soluble fraction of ROFA. This study further
15 showed that, although COX2 products participated in ROFA induced lung inflammation, the
16 COX metabolites are not involved in IL-6 expression nor the influx of PMN influx into the
17 airway. However, the rationale for the use of intraperitoneal challenge was not elaborated.
18 The production of cytokines and mediators also has been shown to depend on the type of
19 PM used in the experiments. A549 cells (a human airway epithelial cell line) were exposed to
20 several PM, carbon black (CB, Elftex-12, Cabot Corp.), diesel soot (ND from NIST, LD
21 produced from General Motors LH 6.2 V8 engine at light duty cycle), ROFA (from the heat
22 exchange section of the Boston Edison), OAA (Ottowa ambient air PM, EHC-93), SiO2, and
23 Ni3S2 at lmg/cm2 (Seagrave and Nikula, 2000). Results indicated that (1) SiO2 and Ni3S2 caused
24 dose dependent acute toxicity and apototic changes; (2) ROFA and LD, ND were significant only
25 at the highest concentrations, (3) SiO2 and Ni3S2 increased IL-8 (three and eight times over the
26 control, respectively) at low concentrations but suppressed IL-8 at high concentrations, (4) OAA
27 and ROFA also induced IL-8 but lower than SiO2 and Ni3S2, and (5) both diesel soots suppressed
28 IL-8 production. The order of potency in alkaline phasphatase production is OAA > LD =
29 ND > ROFA » SiO2 = Ni3S2. These results demonstrated that not only the type of particle used
30 but also the exposure-dose influence the biological response.
March 2001 8-65 DRAFT-DO NOT QUOTE OR CITE
-------
1 Expression of MIP-2 and IL-6 genes was significantly upregulated as early as 6 h
2 post-ROFA-exposure in rat tracheal epithelial cells, whereas gene expression of iNOS was
3 maximally increased 24 h postexposure. V but not Ni appeared to be mediating the effects of
4 ROFA on gene expression. Treatment with dimethylthiourea inhibited both ROFA and V
5 induced gene expression in a dose-dependent manner (Dye et al., 1999).
6 It appears that many biological responses are produced by PM whether it is composed of a
7 single component or a complex mixture. A technical approach is to use the newly developed
8 gene array to monitor the expressions of many mediator genes, which regulate complex and
9 coordinated cellular events involved in tissue injury and repair, in a single assay. Using an array
10 consisting of 27 rat genes representing inflammatory and anti-inflammatory cytokines, growth
11 factors, adhesion molecules, stress proteins, transcription factors, and antioxidant enzymes,
12 Nadadur et al. (2000) measured the expressions of these genes in rats intratracheally instilled
13 with ROFA (3.3 mg/kg), NiSO4 (1.3 /umol/kg), and VSO4 (2.2 //mol/kg). Their data revealed a
14 twofold induction of IL-6 and TIMP-1 at 24 h post-ROFA or Ni exposure. The expression of
15 cellular fibronectin (cFn-EIIIA), ICAM-1, IL-lb, and iNOS gene also were increased 24 h
16 post-ROFA, V, or Ni exposure. This study demonstrated that gene array may provide a tool for
17 screening the expression profile of tissue specific markers following exposure to PM.
18 To investigate the interaction between respiratory cells and PM, Kobzik (1995) showed that
19 scavenger receptors are responsible for AM binding of unopsonized PM and that different
20 mechanisms mediate binding of carbonaceous dusts such as DPM. In addition, surfactant
21 components can increase AM phagocytosis of environmental particulates in vitro, but only
22 slightly relative to the already avid AM uptake of unopsonized particles (Stringer and Kobzik,
23 1996). Respiratory tract epithelial cells are also capable of binding with PM to secrete cytokine
24 IL-8. Using a respiratory epithelial cell line (A549), Stringer et al. (1996) found that binding of
25 particles to epithelial cells was calcium-dependent for TiO2 and Fe2O3, while cc-quartz binding
26 was not calcium dependent. In addition, as observed in AMs, PM binding by A549 cells also
27 was mediated by scavenger receptors, albeit those distinct from the heparin-insensitive
28 acetylated-LDL receptor. Furthermore, a-quartz, but not TiO2 or CAPs, caused a dose-dependent
29 production of IL-8 (range 1 to 6 ng/mL), demonstrating a particle-specific spectrum of epithelial
30 cell cytokine (IL-8) response.
31
March 2001 8-66 DRAFT-DO NOT QUOTE OR CITE
-------
1 8.5.3.3 Other Potential Cellular and Molecular Mechanisms
2 In addition to inducing cytokine mediated inflammation, PM also may affect the alveolar
3 surfactant's ability to reduce both the tendency of alveoli to collapse at the end of expiration and
4 the transudation of fluid from the capillaries to the airspace. Lee et al. (1999) exposed guinea
5 pigs and rats to high concentrations of sulfuric acid aerosol (43 and 94 nig/m3, 0.9 /urn MMAD)
6 and investigated the effects of this aerosol on the surface properties of reconstituted phospholipid
7 using a captive bubble surfactometer. The acid exposure significantly increased the surface
8 tension of guinea pig but not rat BAL. The most sensitive index of surfactant inhibition was
9 found to be the maximum film compressibility of the compression isotherm. The index was
10 119 times greater for the acid exposed guinea pigs compared to control animals. These results
11 were associated with an increase in protein and PMN in the BAL. Although unusually high
12 concentrations of acid aerosols were used in this study, the results may explain the lack of
13 response in the rat to acid aerosol exposures.
14 The potential mechanism involving in the alteration of surface tension may be related to
15 changes in the expression of matrix metalloproteinases (MMPs), such as pulmonary matrilysin
16 and gelatinase A and B, and tissue inhibitor of metalloproteinase (TIMP) (Su et al., 2000a,b).
17 Sprague Dawley rats exposed to ROFA by intratracheal injection (2.5 mg/rat) had increased
18 mRNA levels of matrilysin, gelatinase A, and TIMP-1, Gelatinase B, not expressed in control
19 animals, was increased significantly from 6 to 24 h following ROFA exposure. Alveolar
20 macrophages, epithelial cells, and inflammatory cells were major cellular sources for the
21 pulmonary MMP expression. The expression of Gelatinase B in rats exposed to the same dose of
22 ambient PM (<1.7 /^m and 1.7 to 3.7 /^m) collected from Washington, DC, was significantly
23 increased as compared to saline control, whereas the expression of TIMP-2 was suppressed.
24 Ambient PM between 3.7 and 20 ^m also increased the Gelatinase B expression. Increases in
25 MMPs, which degrade most of the extracellular matrix, suggest that ROFA and ambient PM can
26 similarly increase the total pool of proteolytic activity to the lung and contribute in the
27 pathogenesis of PM-induced lung injury.
28 Sensory nerves originating from trigeminal, nodos, and dorsal root ganglion neurons
29 (DRGs) extend their terminals into the nasal and/or pulmonary epithelium. These nerve
30 terminals together with sensory irritant receptors (capsaicin and acid sensitive receptors) found
31 on the cell bodies can be triggered by irritants such as ambient PM or its components. The
March 2001 8-67 DRAFT-DO NOT QUOTE OR CITE
-------
1 activation of these receptors and nerve terminals can result in the release of inflammatory
2 cytokines leading to airway disorders. Intracellular calcium levels increased immediately in
3 BEAS-2B cells exposed to ROFA, which is followed by increased in IL-6, IL-8, and TNFoc gene
4 expressions (Veronesi et al., 1999). Furthermore, acidic media alone or soluble components of
5 ROFA produced similar effects. These responses were reduced by pretreating cells with
6 neuropeptide antagonists. However, treating cells with capsaicin antagonist, a pH receptor
7 antagonist, or exposing cells to ROFA in Ca2+ free media inhibited both intracellular Ca2+ as well
8 as cytokine release. Using synthetic polymer microspheres (SPMs) resembling ROFA particles
9 in size (2 and 6 ^m in diameter) and surface potential (zeta potential -29 mV) but lacking
10 confounding factors such as metals or biologies, Oortgiesen et al. (2000) demonstrated that
11 BEAS-2B and DRGs responded to both ROFA and charged SPMs with an increase in
12 intracellular Ca2+ ([Ca2+],) concentration and the release of IL-6, whereas neutral SPMs bound
13 with polyethylene glycol (0 mV zeta potential) were relatively ineffective. In DRGs, the SPM-
14 induced increases in [Ca2+], were correlated with the presence of acid- or capsaicin-sensitive
15 pathways. By this pathway, soluble components of ROFA, which is acidic, and other acidic PM
16 may initiate or exacerbate symptoms of airway inflammation. These data not only demonstrated
17 that the surface chemistry of the particles determines whether cells are activated but also that
18 direct contact of the particle with the target cells and their receptors is necessary for particles to
19 evoke a response.
20
21 8.5.4 Specific Particle Size and Surface Area Effects
22 Most particles used in laboratory animal toxicology and occupational studies are greater
23 than 0.1 /zm in size. However, the enormous number and huge surface area of the ultrafine
24 particles demonstrate the importance of considering the size of the particle in assessing response.
25 Ultrafine particles with a diameter of 20 nm when inhaled at the same mass concentration have a
26 number concentration that is approximately 6 orders of magnitude higher than for a 2.5-/urn
27 diameter particle; particle surface area is also greatly increased (Table 8-9).
28 Many studies summarized in U.S. Environmental Protection Agency (1996a), as well as in
29 this document, suggest that the surface of particles or substances that are released from the
30 surface (e.g., transition metals) interact with the biological system, and that surface-associated
31 free radicals or free radical-generating systems may be responsible for toxicity. Thus, if ultrafine
March 2001 8-68 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 8-9. NUMBERS AND SURFACE AREAS OF MONODISPERSE
PARTICLES OF UNIT DENSITY OF DIFFERENT SIZES AT A MASS
CONCENTRATION OF 10
Particle Diameter
Cum)
0.02
0.1
0.5
1.0
2.5
Particle Number
(per cm3 air)
2,400,000
19,100
153
19
1.2
Particle Surface Area
(/^m2 per cm3 air)
3,016
600
120
60
24
Source: Oberdorster et al. (1995).
1 particles were to cause toxicity by a transition metal-mediated mechanism, for example, then the
2 relatively large surface area for a given mass of ultrafine particles would mean high
3 concentrations of transition metals being available to cause oxidative stress to cells.
4 Two groups have examined the toxic differences between fine and ultrafine particles, with
5 the general finding that the ultrafine particles show a significantly greater response at similar
6 mass doses (Oberdorster et al., 1992; Li et al., 1996, 1997, 1999). However, only a few studies
7 have investigated the ability of ultrafine particles to generate a greater oxidative stress when
8 compared to fine particles of the same material. Studies by Gilmour et al. (1996) have shown
9 that at equal mass, ultrafine TiO2 caused more plasmid DNA strand breaks than fine TiO2. This
1 0 effect could be inhibited with mannitol. Osier and Oberdorster (1 997) compared the response of
1 1 rats (F344) exposed by intratracheal inhalation to "fine" (approximately 250 nm) and "ultrafine"
12 (approximately 2 1 nm) TiO2 particles with rats exposed to similar doses by intratracheal
13 instillation. Animals receiving particles through inhalation showed a smaller pulmonary
14 response, measured by BAL parameters, in both severity and persistence, when compared with
1 5 those animals receiving particles through instillation. These results demonstrate a difference in
16 pulmonary response to an inhaled versus an instilled dose, which may result from differences in
17 dose rate, particle distribution, or altered clearance between the two methods. Consistent with
1 8 these in vivo studies, Finkelstein et al. (1997) has shown that exposing primary cultures of rat
19 Type II cells to 10 ^g/mL ultrafine TiO2 (20 nm) causes increased TNF and IL-1 release
March 2001 8-69 DRAFT-DO NOT QUOTE OR CITE
-------
1 throughout the entire 48-h incubation period. In contrast, fine TiO2 (200 nm) had no effect.
2 In addition, ultrafine polystyrene carboxylate-modified microspheres (UFP, fluorospheres,
3 molecular probes 44 ± 5 nm) have been shown induce a significant enhancement of both
4 substance P and histamine release after administration of capsaicin (10"4 M), to stimulate C-fiber,
5 and carbachol (10"4 M), a cholinergic agonist in rabbit intratracheally instilled with UFP
6 (Nemmar et al, 1999). A significant increase in histamine release also was recorded in the
7 UFP-instilled group following the administration of both Substance P (10~6 M) plus thiorpan
8 (10~5 M) and compound 48/80 (C48/80, 10"3 M) to stimulate mast cells. BAL analysis showed an
9 influx of PMN, an increase in total protein concentration, and an increase in lung wet weight/dry
10 weight ratio. Electron microscopy showed that both epithelial and endothelial injuries were
11 observed. The pretreatment of rabbits in vivo with a mixture of either SR 140333 and SR 48368,
12 a tachykinin NK, and NK2 receptor antagonist, or a mixture of terfenadine and cimetidine,
13 a histamine H, and H2 receptor antagonist, prevented UFP-induced PMN influx and increased
14 protein and lung WW/DW ratio.
15 As discussed earlier, it is believed that ultrafine particles caused greater cellular injury
16 because of the relatively large surface area for a given mass. However, in a study that compared
17 the response to carbon black particles of two different sizes, Li et al. (1999) demonstrated that in
18 the instillation model, a localized dose of particle over a certain level causes the particle mass to
19 dominate the response, rather than the surface area. Ultrafine carbon black (ufCB, Printex 90),
20 14 nm in diameter, and fine carbon black (CB, Huber 990), 260 nm in diameter, were instilled
21 intratracheally in rats and BAL profile at 6 h was assessed. At mass of 125 /^g or below, ufCB
22 generated a greater response (increase LDH, epithelial permeability, decrease in GSH, TNF, and
23 NO productions) than fine CB at various time postexposure. However, higher dose of CB caused
24 more PMN influx than the ufCB. In contrast to the effect of CB, which showed dose-related
25 increasing inflammatory response, ufCB at the highest dose caused less of a neutrophil influx
26 than at the lower dose. Moreover, when the PMN influx was expressed as a function of surface
27 area, CB produced greater response than ufCB at all doses used in this study. Although particle
28 insterstitialization with a consequent change in the chemotatic gradient for PMN was offered as
29 an explanation, these results need further scrutinization.
30 Oberdorster et al. (2000) recently completed a series of studies in rats and mice using
31 ultrafine particles of various chemical compositions (PTFE, TiO2, C, Fe, Fe2O3, Pt, V, and V2O5).
March 2001 8-70 DRAFT-DO NOT QUOTE OR CITE
-------
1 In old rats sensitized with endotoxin and exposed to ozone plus ultrafine carbon particles, they
2 found a ninefold greater release of reactive oxygen species in old rats than in similarly treated
3 young rats. Exposure to ultrafine PM alone in sensitized old rats also caused an inflammatory
4 response.
5 Although the exact mechanism of ultrafine-induced lung injury remains unclear, it is likely
6 that ultrafine particles, because of their small size, can easily penetrate the airway epithelium and
7 cause cellular damage. Using electron microscopy to examine rat tracheal explants treated with
8 fine (0.12 /urn) and ultrafine (0.021 ^trn) TiO2 partilces for 3 or 7 days, Churg et al. (1998) found
9 both size particles in the epithelium at both time points, but in the subepithelial tissues, they were
10 found only at Day 7. The volume proportion (the volume of TiO2 over the entire volume of
11 epithelium or subepithelium area) of both fine and ultrafine particles in the epithelium increased .
12 from 3 to 7 days. It was greater for ultrafine at 3 days but was greater for fine at 7 days. The
13 volume proportion of particles in the subepithelium at day 7 was equal for both particles, but the
14 ratio of epithelial to subepithelial volume proportion was 2:1 for fine and 1:1 for ultrafine.
15 Ultrafine particles persist in the tissue as relatively large aggregates, whereas the size of fine
16 particle aggregates becomes smaller over time. Ultrafine particles appear to enter the epithelium
17 faster and, once in the epithelium, a greater proportion of them is translocated to the subepithelial
18 space compared to fine particles. However, if it is assumed that the volume proportion is
19 representative of particle number, the number of particles reaching the interstitial space is
20 directly proportional to the number applied (i.e., there is no preferential transport from lumen to
21 interstitium by size). These data are in direct contrast to the results of instillation or inhalation of
22 fine and ultrafine TIO2 particles reported earlier (Ferin et al., 1990, 1992). Free of inflammatory
23 cells, possibility of overloading of the explants with dust, and the use of liquid suspension for
24 exposure were among the possible reasons cited for the observed effects.
25 Only two studies examined the influence of specific surface area on biological activity
26 (Lison et al., 1997; Oettinger et al., 1999). The biological responses to various MnO2 dusts with
27 different specific surface area (0.16, 0.5, 17, and 62 m2/g) were compared in vitro and in vivo
28 (Lison et al., 1997). In both systems, the results show that the amplitude of the response is
29 dependent on the total surface area that is in contact with the biological system, indicating that
30 surface chemistry phenomena are involved in the biological reactivity. Freshly ground particles
31 with a specific surface area of 5 m2/g also were examined in vitro. These particles exhibited an
March 2001 8-71 DRAFT-DO NOT QUOTE OR CITE
-------
1 enhanced cytotoxic activity, which was almost equivalent to that of particles with a specific
2 surface area of 62 m2/g, indicating that undefined reactive sites produced at the particle surface
3 by mechanical cleavage also may contribute to the toxicity of insoluble particles. In another
4 study, two types of carbon black particles, Printex 90 (P90, Degussa, Germany, formed by
5 controlled combustion, consists of defined granules with specific surface area of 300 m2/g and
6 particle size of 14 nm) and FR 101 (Degussa, Germany, with specific surface area of 20 m2/g and
7 particle size of <95 nm, has a coarse structure, and the ability to adsorb polycyclic and other
8 carbons) were used in the study (Oettinger et al., 1999). Exposure of AMs to 100 fj.g/106 cells of
9 FR 101 and P90 resulted in a 1.4- and 2.1-fold increase in ROS release. These exposures also
10 caused a fourfold up-regulation of NF-kB gene expression. These studies indicated that PM of
11 single component with larger surface properties produce greater biological response than similar
12 particles with smaller surface area. By exposing bovine AMs to metal oxide coated silica
13 particles, Schluter et al. (1995) showed that most of the metal coatings (Li, Cr, Fe, Mn, Ni, Ph,
14 and V) had no effect on ROS production by these cells. However, coating with CuO markedly
15 lowered the O2" and H2O2, whereas V(IV) increases both ROI. This study demonstrated that, in
16 addition to specific area, chemical composition of the particle surface also influence its cellular
17 response.
18
19 8.5.5 Pathophysiological Mechanisms for the Effects of Low Concentrations
20 of Particulate Air Pollution
21 The pathophysiological mechanisms involved in PM-associated cardiovascular and
22 respiratory health effects still are not elucidated fully, but progress has been made since the 1996
23 PM AQCD (U. S. Environmental Protection Agency, 1996a) was prepared. This section
24 summarizes current hypotheses and reviews the toxicological evidence for these potential
25 pathophysiological mechanisms.
26
27 8.5.5.1 Direct Pulmonary Effects
28 When the 1996 PM AQCD (U. S. Environmental Protection Agency, 1996a) was written,
29 the lung was thought to be the primary organ to affected by particulate air pollution. There is
30 growing toxicological and epidemiological evidence that the cardiovascular system is affected as
31 well. Nonetheless, understanding how particulate air pollution causes or exacerbates respiratory
March 2001 8-72 DRAFT-DO NOT QUOTE OR CITE
-------
1 disease remains an important goal. There is some toxicological evidence for the following three
2 mechanisms for direct pulmonary effects.
3
4 Paniculate Air Pollution Causes Lung Injury and Inflammation
5 In the last few years, numerous studies have shown that instilled and inhaled ROFA, a
6 product of fossil fuel combustion, can cause substantial lung injury and inflammation. The toxic
7 effects of ROFA are largely caused by its high content of soluble metals, and the pulmonary
8 effects of ROFA can be reproduced by equivalent exposures to soluble metal salts. In contrast,
9 controlled exposures of animals to sulfuric acid aerosols, acid coated carbon, and sulfate salts
10 cause little lung injury or inflammation, even at high concentrations. Inhalation of concentrated
11 ambient PM (which contains only small amounts of metals) by laboratory animals at
12 concentrations in the range of 100 to 1000 /^g/m3 have been shown in some (but not all) studies
13 to cause mild pulmonary injury and inflammation. Rats with SO2-induced bronchitis and
14 monocrotaline-treated rats have been reported to have a greater inflammatory response to
15 concentrated ambient PM than normal rats. These studies suggest that exacerbation of
16 respiratory disease by ambient PM may be caused in part by lung injury and inflammation.
17
18 Particulate Air Pollution Causes Increased Susceptibility to Respiratory Infections
19 At this time there are no newly published studies on the effects of inhaled concentrated
20 ambient PM on host susceptibility to infectious agents. Ohtsuka et al. (2000a,b) have shown that
21 in vivo exposure of mice to acid-coated carbon particles at a mass concentration of 10,000 /ug/m3
22 causes decreased phagocytic activity of alveolar macrophages, even in the absence of lung injury.
23
24 Particulate Air Pollution Increases Airway Reactivity and Exacerbates Asthma
25 The strongest evidence supporting this hypothesis is from studies on diesel particulate
26 matter (DPM). DPM has been shown to increase production of antigen-specific IgE in mice and
27 humans (summarized in Section 8.2.4.2). In vitro studies have suggested that the organic
28 fraction of DPM is involved in the increased IgE production. ROFA leachate also has been
29 shown to enhance antigen-specific airway reactivity in mice (Goldsmith et al., 1999) indicating
30 that soluble metals can also enhance an allergic response. However, in this same study, exposure
31 of mice to concentrated ambient PM did not affect antigen-specific airway reactivity. It is
March 2001 8-73 DRAFT-DO NOT QUOTE OR CITE
-------
1 premature to conclude from this one experiment that concentrated ambient PM does not
2 exacerbate allergic airways disease because the chemical composition of the PM (as indicated by
3 studies with DPM and ROFA) may be more important than the mass concentration.
4
5 8.5.5.2 Systemic Effects Secondary to Lung Injury
6 When the 1996 PM AQCD was written, it was thought that cardiovascular-related
7 morbidity and mortality most likely would be secondary to impairment of oxygenation or some
8 other consequence of lung injury and inflammation. Newly available toxicologic studies provide
9 some additional evidence regarding such possibilities.
10
11 Lung Injury from Inhaled Paniculate Matter Causes Impairment of Oxygenation and
12 Increased Work of Breathing That Adversely Affects the Heart
14 Instillation of ROFA has been shown to cause a 50% mortality rate in monocrotaline-
15 treated rats (Watkinson et al., 2000). Although blood oxygen levels were not measured in this
16 study, there were ECG abnormalities consistent with severe hypoxemia in about half of the rats
17 that subsequently died. Given the severe inflammatory effects of instilled ROFA and the fact
18 that monocrotaline-treated rats have increased lung permeability as well as pulmonary
19 hypertension, it is plausible that instilled ROFA can cause severe hypoxemia leading to death in
20 this rat model. Results from studies in which animals (normal and compromised) were exposed
21 to concentrated ambient PM (at concentrations many times higher than would be encountered in
22 the United States) indicate that ambient PM is unlikely to cause severe disturbances in
23 oxygenation or pulmonary function. However, even a modest decrease in oxygenation can have
24 serious consequences in individuals with ischemic heart disease. Kleinman et al. (1998) has
25 shown that a reduction in arterial blood saturation from 98 to 94% by either mild hypoxia or by
26 exposure to 100 ppm CO significantly reduced the time to onset of angina in exercising
27 volunteers. Thus, information is needed on the effects of PM on arterial blood gases and
28 pulmonary function to fully address the above hypothesis.
29
30 Lung Inflammation and Cytokine Production Cause Adverse Systemic Hemodynamic Effects
31 It has been suggested that systemic effects of particulate air pollution may result from
32 activation of cytokine production in the lung (Li et al., 1997). In support of this idea,
March 2001 8-74 DRAFT-DO NOT QUOTE OR CITE
-------
1 monocrotaline-treated rats exposed to inhaled ROFA (15,000 /ug/m3, 6 h/day for 3 days) showed
2 increased pulmonary cytokine gene expression, bradycardia, hypothermia, and increased
3 arrhythmias (Watkinson et al., 2000). However, spontaneously hypertensive rats had a similar
4 cardiovascular response to inhaled ROFA (except that they also developed ST segment
5 depression) with no increase in pulmonary cytokine gene expression. Studies in dogs exposed to
6 concentrated ambient PM showed minimal pulmonary inflammation and no positive staining for
7 IL-8, IL-1, or TNF in airway biopsies. However, there was a significant decrease in the time of
8 onset of ischemic ECG changes following coronary artery occlusion in PM-exposed dogs
9 compared to controls (Godleski et al., 2000). Thus, there is not a clear-cut link between changes
10 in cardiovascular function and production of cytokines in the lung. Because human and animal
11 exposure studies of ambient PM are using increasingly sophisticated and sensitive measures of
12 cardiac function, basic information on the effects of mild pulmonary injury on these cardiac
13 endpoints is needed to understand the mechanisms of how inhaled PM affects the heart.
14
15 Lung Inflammation from Inhaled Particulate Matter Causes Increased Blood Coagulability
16 That Increases the Risk of Heart Attacks and Strokes
18 There is abundant evidence linking risk of heart attacks and strokes to small prothrombotic
19 changes in the blood coagulation system. However, the published toxicological evidence that
20 moderate lung inflammation causes increased blood coagulability is inconsistent. Ohio et al.
21 (2000) have shown that inhalation of concentrated ambient PM in healthy nonsmokers causes
22 increased levels of blood fibrinogen. Gardner et al. (2000) have shown that a high dose
23 (8,300 /wg/kg) of instilled ROFA in rats causes increased levels of fibrinogen, but no effect was
24 seen at lower doses. Exposure of dogs to concentrated ambient PM had no effect on fibrinogen
25 levels (Godleski et al., 2000). The coagulation system is as multifaceted and complex as the
26 immune system, and there are many other sensitive and clinically significant parameters that
27 should be examined in addition to fibrinogen. Thus, it is premature to draw any conclusions on
28 the relationship between PM and blood coagulation.
29
30 Interaction of Particulate Matter with the Lung Affects Hematopoiesis
31 Terashima et al. (1997) found that instillation of fine carbon particles (20,000 //g/rabbit)
32 stimulated release of PMNs from the bone marrow. In further support of this hypothesis, Gordon
March 2001 8-75 DRAFT-DO NOT QUOTE OR CITE
-------
1 and colleagues reported that the percentage of PMNs in the peripheral blood increased in rats
2 exposed to ambient PM in some but not all exposures. On the other hand, Godleski et al. (2000)
3 found no changes in peripheral blood counts of dogs exposed to concentrated ambient PM.
4 Thus, direct evidence that PM ambient concentrations can affect hematopoiesis remains to be
5 demonstrated.
6
7 8.5.5.3 Direct Effects on the Heart
8 Changes in heart rate, heart rate variability, and conductance associated with ambient PM
9 exposure have been reported in animal studies (Godleski et al., 2000; Gordon et al., 2000;
10 Watkinson et al., 2000), in several human panel studies (described in Chapter 6), and in a
11 reanalysis of data from the MONICA study (Peters et al., 1997). Some of these studies included
12 endpoints related to respiratory effects but few significant adverse respiratory changes were
13 detected. This raises the possibility that ambient PM may have effects on the heart that are
14 independent of adverse changes in the lung. There is certainly precedent for this idea.
15 For example, tobacco smoke (which is a mixture of combustion-generated gases and PM) causes
16 cardiovascular disease by mechanisms that are independent of its effect on the lung. Two types
17 of hypothesized direct effects of PM on the heart are noted below.
18
19 Inhaled Paniculate Matter Affects the Heart by Uptake of Particles into the Circulation or
20 Release of a Soluble Substances into the Circulation.
21
22 Drugs can be rapidly and efficiently delivered to the systemic circulation by inhalation.
23 This implies that the pulmonary vasculature absorbs inhaled materials, including charged
24 substances such as small proteins and peptides. Cigarettes are a widely used method for
25 delivering nicotine to the blood stream. It is likely that soluble materials absorbed onto airborne
26 particles find their way into the blood stream, but it is not clear whether the particles themselves
27 enter the blood. It is anticipated that more information will be available on this important
28 question in the next few years.
29
30 Inhaled Particulate Matter Affects Autonomic Control of the Heart and Cardiovascular
31 System
33 There is growing evidence for this idea as described above. This raises the question of how
34 inhaled particles could affect the autonomic nervous system. Activation of neural receptors in
March 2001 8-76 DRAFT-DO NOT QUOTE OR CITE
-------
1 the lung is a logical area to investigate. Studies in conscious rats have shown that inhalation of
2 wood smoke causes marked changes in sympathetic and parasympathetic input to the
3 cardiovascular system that are mediated by neural reflexes (Nakamura and Hayashida, 1992).
4 Although research on airway neural receptors and neural-mediated reflexes is a well established
5 discipline, the cardiovascular effects of stimulating airway receptors continue to receive less
6 attention than the pulmonary effects. Previous studies of airway reflex-mediated cardiac effects
7 usually employed very high doses of chemical irritants, and the results may not be applicable to
8 air pollutants. There is a need for basic physiological studies to examine effects on
9 cardiovascular system when airway and alveolar neural receptors are stimulated in a manner
10 relevant to air pollutants.
11
12
13 8.6 RESPONSES TO PARTICULATE MATTER AND GASEOUS
14 POLLUTANT MIXTURES
15 Ambient PM itself is a mixture of particles of varying size and composition. The following
16 discussion examines effects of mixtures of ambient PM, or PM surrogates, with gaseous
17 pollutants. Ambient PM co-exists in indoor and outdoor air with a number of co-pollutant gases,
18 including ozone, sulfur dioxide, oxides of nitrogen, and carbon monoxide. Toxicological
19 interactions between PM and gaseous co-pollutants may be antagonistic, additive, or synergistic
20 (Mauderly, 1993). The presence and nature of any interaction appears to depend on the size and
21 concentration of pollutants in the mixture, exposure duration, and the endpoint being examined.
22 It is not possible to predict a priori from the presence of certain pollutants whether any
23 interaction will occur and, if there is interaction, whether it will be synergistic, additive, or
24 antagonistic (Table 8-10).
25 Mechanisms responsible for the various forms of interaction are speculative. In terms of
26 potential health effects, the greatest hazard from pollutant interaction is the possibility of synergy
27 between particles and gases, especially if effects occur at concentrations at which no effects
28 occur when individual constituents are inhaled. Various physical and chemical mechanisms may
29 underlie synergism. For example, physical adsorption or absorption of some material on a
30 particle could result in transport to more sensitive sites, or sites where this material would not
31 normally be deposited in toxic amounts. This physical process may explain the interaction found
March 2001 8-77 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 8-10. RESPIRATORY AND CARDIOVASCULAR EFFECTS OF MIXTURES
C- Species,
^ Gender, Strain
O Age, or Body
2 Weight
Rats, Fischer
NNia, male,
22 to 24 mo
old
Rats
Gases and Exposure
Particles Technique
Carbon, Inhalation
ammonium
bisulfate,
and O3
O3 and Ottawa Inhalation
urban dust
Mass Particle
Concentration Size
50 A
-------
TABLE 8-10 (cont'd). RESPIRATORY AND CARDIOVASCULAR EFFECTS OF MIXTURES
o
N)
0
o
H— *
oo
~Lj
^o
O
i
H
6
o
z
o
H
O
a,
o
H
M
o
o
H
W
Species,
Gender, Strain
Age, or Body Gases and
Weight Particles
Rats H,SO4 and O3
Healthy and H2SO4,
asthmatic SO,, and 03
children
Pigeons Ambient gases
(Columba and particles
livia)
Canine Ambient gases
and particles
Rats 03 and
resuspended
urban PM
Exposure Particle
Technique Mass Concentration Size
Inhalation, 20 to 1 50 Aig/W 0.4 to 0 8 ^m
whole body H,SO4 and 0. 1 2 or
0.2 ppm O3
Inhalation 60 to 1 40 pig/m3 06^mIl,SO4
H2SO4, 0.1 ppm SO,,
and 0. 1 ppm O3
Natural 24-h
exposure in
urban and
rural areas
Natural 24-h
exposure in
four urban
areas of
Mexico City
and one
rural area
Inhalation, 0 8 ppm O3 and
whole-body 5,000 or
50,000 ^g/m3 PM
Exposure
Duration
Intermittent
(12h/day)or
continuous
exposure for up
to 90 days
Single 4-h
exposure with
intermittent
exercise
Continuous
ambient exposure
Continuous
ambient exposure
Single 4-h
exposure
Respiratory Effects of Inhaled Particles on Markers
m Lavage Fluid
No interactive effect of H,SO4 and O3 on
biochemical and morphometnc endpoints.
A positive association between acid concentration
and symptoms, but not spirometry, in asthmatic
children. No changes in healthy children.
Increased number of AMs and decreased number of
lamellar bodies in type II epithelial cells in urban
pigeons.
No significant differences in AMs or total cell
counts in lavage from dogs studied among the
five regions. A significant increase in lavage fluid
neutrophils and lymphocytes in the southwest
region, where the highest O, levels were recorded,
compared to the two industnal regions with the
highest PM levels.
PM alone caused no change in cell proliferation in
bronchioles or parenchyma Co-exposure with O3
greatly potentiated the prohferative changes
induced by O3 alone. These changes were greatest
in the epithelium of the terminal bronchioles and
alveolar ducts
Reference
Last and Pmkerton
(1997)
Linn et al.
(1997)
Lorz and Lopez
(1997)
Vanda et al.
(1998)
Vincent et al.
(1997)
-------
1 in studies of mixtures of carbon black and formaldehyde or of carbon black and acrolein (Jakab,
2 1992, 1993).
3 Chemical interactions between particles and gases can occur on particle surfaces, thus,
4 forming secondary products that may be more active lexicologically than the primary materials
5 and that can then be carried to a sensitive site. The hypothesis of such chemical interactions has
6 been examined in the gas and particle exposure studies by Amdur and colleagues (Amdur and
7 Chen, 1989; Chen et al., 1992) and Jakab and colleagues (Jakab and Hemenway, 1993; Jakab
8 et al., 1996). These investigators have demonstrated that synergism occurs as secondary
9 chemical species are produced, especially under conditions of increased temperature and relative
10 humidity.
1 1 Another potential mechanism of gas-particle interaction may involve a pollutant-induced
12 change in the local microenvironment of the lung, enhancing the effects of the co-pollutant.
13 For example, Last et al. (1984) suggested that the observed synergism between ozone and acid
14 sulfates in rats was due to a decrease in the local microenvironmental pH of the lung following
1 5 deposition of acid, enhancing the effects of ozone by producing a change in the reactivity or
1 6 residence time of reactants, such as radicals, involved in ozone-induced tissue injury.
1 7 As noted in U.S. Environmental Protection Agency (1996a), the toxicology database for
1 8 mixtures containing PM other than acid sulfates is still quite sparse. Vincent et al. (1997)
19 exposed rats to 0.8 ppm ozone in combination with 5 or 50 mg/m3 of resuspended urban particles
20 for 4 h. Although PM alone caused no change in cell proliferation (3H-thymidine labeling),
2 1 co-exposure to either concentration of resuspended PM with ozone greatly potentiated the
22 proliferative effects of exposure to ozone alone. These interactive changes occurred in epithelial
23 cells of the terminal bronchioles and the alveolar ducts. These findings using resuspended dusts,
24 although at high concentrations, are consistent with studies demonstrating interaction between
25 sulfuric acid (H2SO4) aerosols and ozone. Kimmel and colleagues (1997) examined the effect of
26 acute co-exposure to ozone and fine or ultrafine H2SO4 aerosols on rat lung morphology. They
27 determined morphometrically that alveolar septal volume was increased in animals co-exposed to
28 ozone and ultrafine, but not fine, H2SO4. Interestingly, cell labeling, an index of proliferative cell
29 changes, was increased only in animals co-exposed to fine H2SO4 and ozone, as compared to
30 animals exposed to ozone alone. Importantly, Last and Pinkerton (1997) extended their previous
3 1 work and found that subchronic exposure to acid aerosols (20 to 1 50 A^g/m3 H2SO4) had no
March 2001 8-80 DRAFT-DO NOT QUOTE OR CITE
-------
1 interactive effect on the biochemical and morphometric changes produced by either intermittent
2 or continuous ozone exposure (0.12 to 0.2 ppm). Thus, the interactive effects of ozone and acid
3 aerosol co-exposure in the lung disappeared during the long-term exposure.
4 Kleinman et al. (1999) examined the effects of ozone plus fine, H2SO4-coated, carbon
5 particles (MMAD = 0.26 /^m) for 1 or 5 days. They found the inflammatory response with the
6 ozone-particle mixture was greater after 5 days (4 h/day) than after Day 1. This contrasted with
7 ozone exposure alone (0.4 ppm), which caused marked inflammation on acute exposure, but no
8 inflammation after 5 consecutive days of exposure.
9 Studies have examined interaction between carbon particles and gaseous co-pollutants.
10 Jakab et al. (1996) challenged mice with a single 4-h exposure to a high concentration of carbon
11 (10 mg/m3) in the presence of SO2 at low and high relative humidities. Macrophage phagocytosis
12 was depressed significantly only in mice exposed to the combined pollutants under high relative
13 humidity conditions. This study demonstrates that fine carbon particles can serve as an effective
14 carrier for acidic sulfates where chemical conversion of adsorbed SO2 to acid sulfate species
15 occurred. Interestingly, the depression in macrophage function was present as late as 7 days
16 postexposure. Bolarin et al. (1997) exposed rats to only 50 or 100 /ug/m3 carbon particles in
17 combination with ammonium bisulfate and ozone. Despite 4 weeks of exposure, they observed
18 no changes in protein concentration in lavage fluid or blood prolyl 4-hydroxylase, an enzyme
19 involved in collagen metabolism. Slight decreases in plasma fibronectin were present in animals
20 exposed to the combined pollutants versus ozone alone. Thus as, previously noted, the potential
21 for adverse effects in the lungs of animals challenged with a combined exposure to particles and
22 gaseous pollutants is dependent on numerous factors, including the gaseous co-pollutant,
23 concentration, and time.
24 In a complex series of exposures, Oberdorster and colleagues examined the interaction of
25 ultrafine carbon particles (100 /wg/m3) and ozone (1 ppm) in young and old Fischer 344 rats that
26 were pretreated with aerosolized endotoxin (Elder et al., 2000). In old rats, exposure to carbon
27 and ozone produced an interaction that resulted in a greater influx in neutrophils than that
28 produced by either agent alone. This interaction was not seen in young rats. Oxidant release
29 from lavage fluid cells was also assessed and the combination of endotoxin, carbon particles, and
30 ozone produced an increase in oxidant release in old rats. This combination produced the
31 opposite response in the cells recovered from the lungs of the young rats, indicating that the
March 2001 8-81 DRAFT-DO NOT QUOTE OR CITE
-------
1 lungs of the aged animals underwent greater oxidative stress in response to this complex
2 pollutant mix of particles, ozone, and a biogenic agent.
3 Linn and colleagues (1997) examined the effect of a single exposure to 60 to 140 /wg/m3
4 H2SO4, 0.1 ppm SO2, and 0.1 ppm ozone in healthy and asthmatic children. The children
5 performed intermittent exercise during the 4-h exposure to increase the inhaled dose of the
6 pollutants. An overall effect on the combined group of healthy and asthmatic children was not
7 observed. A positive association between acid concentration and symptoms was seen, however,
8 in the subgroup of asthmatic children. The combined pollutant exposure had no effect on
9 spirometry in asthmatic children, and no changes in symptoms or spirometry were observed in
10 healthy children. Thus, the effect of combined exposure to PM and gaseous co-pollutants
11 appeared to have less effect on asthmatic children exposed under controlled laboratory conditions
12 in comparison with field studies of children attending summer camp (Thurston et al., 1997).
13 However, prior exposure to H2SO4 aerosol may enhance the subsequent response to ozone
14 exposure (Linn et al., 1994; Frampton et al., 1995); the timing and sequence of the exposures
15 may be important.
16 Three unique animal field studies have examined the adverse respiratory effects of complex
17 mixtures in urban and rural environments. These studies have taken advantage of the differences
18 in pollutant makeup of urban and rural environments and studied animals under natural,
19 continuous exposure conditions. Gulisano et al. (1997) examined the morphologic changes
20 produced by continuous ambient exposure to air pollutants in lambs raised for 3 mo in rural
21 (n = 2) or urban (n = 10) environments. Compared to the lungs of the rural lambs, irritation, as
22 characterized by mucus hypersecretion and morphological changes in the epithelial cells lining
23 the nasopharyngeal region, was present in the lambs exposed to urban air pollution. Lorz and
24 Lopez (1997) performed a similar study using pigeons as the test animal. They observed an
25 increase in the number of AMs and a decrease in the number of lamellar bodies in Type II
26 epithelial cells in the lungs of urban pigeons. Extrapolation of these studies is hampered by an
27 incomplete characterization of the exposure atmospheres. A more thorough examination of the
28 ambient level of pollutants was performed in the study by Vanda et al. (1998), who studied the
29 effect of pollutant exposure in dogs raised in four urban regions of Mexico City and one nearby
30 rural area. They found no significant differences in AM number or total cell counts in lavage
31 fluid from the dogs among the five regions. A significant increase in lavage fluid neutrophils and
March 2001 8-82 DRAFT-DO NOT QUOTE OR CITE
-------
1 lymphocytes was found in dogs from the urban region with the highest ozone levels in
2 comparison to the regions with the highest PM levels. Thus, the effect of ozone on cellular
3 parameters in lavage fluid appeared to be greater than that for PM. In summary, each of these
4 three animal field studies provides evidence that urban air pollutants can produce greater lung
5 changes than would occur from exposure to rural pollution. However, extrapolation of these
6 results is severely hampered by the uncontrolled exposure conditions, small sample size,
7 behavior patterns, and nutritional factors. Thus, in these field studies, it is difficult to assign a
8 role to PM in the observed adverse pulmonary effects.
9
10
11 8.7 SUMMARY
12 8.7.1 Biological Plausibility
13 Toxicological studies can play an integral role in answering the following two key
14 questions regarding biological plausibility of PM health effects.
15 (1) What component (or components) of ambient PM cause health effects?
16 (2) Are the statistical associations between PM and health effects biologically plausible?
17 This summary focuses on the progress that toxicological studies have made towards answering
18 these questions.
19
20 8.7.1.1 Link Between Specific Participate Matter Components and Health Effects
21 Key to the validity of the biological plausibility is the need to identify the components of
22 airborne PM responsible for the adverse effects and the individuals at risk. The plausibility of
23 the association between PM and increases in morbidity and mortality has been questioned
24 because the adverse cardiopulmonary effects have been observed at very low PM concentrations,
25 often below the current NAAQS for PM10. To date, toxicology studies on PM have provided
26 only very limited evidence for specific PM components being responsible for observed
27 cardiopulmonary effects of ambient PM. Studies have shown that some components of particles
28 are more toxic than others. For example, high concentrations of ROFA and associated soluble
29 metals have produced clinically significant effects (including death) in compromised animals.
30 The relevance of these findings to understanding the adverse effects of PM components is
March 2001 8-83 DRAFT-DO NOT QUOTE OR CITE
-------
1 tempered, however, by the large difference between metal concentrations delivered to the test
2 animals and metal concentrations present in the ambient urban environment. Such comparisons
3 must be applied to the interpretation of all studies that examine the individual components of
4 ambient urban PM. A summary of potential contributions of individual physical/chemical factors
5 of particles to cardiopulmonary effects is given below.
6
7 Acid Aerosols
8 There is relatively little new information on the effects of acid aerosols, and the conclusions
9 of the 1996 PM AQCD are unchanged. It was previously concluded that acid aerosols cause
10 little or no change in pulmonary function in healthy subjects, but asthmatics may develop small
11 changes in pulmonary function. This conclusion is supported by the recent study of Linn and
12 colleagues (1997) in which children (26 children with allergy or asthma and 15 healthy children)
13 were exposed to sulfuric acid aerosol (100 //g/m3) for 4 h. There were no significant effects on
14 symptoms or pulmonary function when data from the entire group was analyzed, but the allergy
15 group had a significant increase in symptoms after the acid aerosol exposure.
16 Although pulmonary effects of acid aerosols have been the subject of extensive research in
17 past decades, the cardiovascular effects of acid aerosols have received little attention. Zhang
18 et al. (1997) reported that inhalation of acetic acid fumes caused reflex mediated increases in
19 blood pressure in normal and spontaneously hypertensive rats. Thus, acid components should
20 not be ruled out totally as possible mediators of PM health effects. In particular, the
21 cardiovascular effects of acid aerosols at realistic concentrations need further investigation.
22
23 Metals
24 The previous PM AQCD (U.S. Environmental Protection Agency, 1996a) mainly relied on
25 data related to occupational exposures to evaluate the potential toxicity of metals in particulate
26 air pollution. Since that time, in vivo and in vitro studies using ROFA or soluble transition
27 metals have contributed substantial new information on the health effects of particle-associated
28 soluble metals. Although there are some uncertainties about differential effects of one transition
29 metal versus another, water soluble metals leached from ROFA have been shown consistently
30 (albeit at high concentrations) to cause cell injury and inflammatory changes in vitro and in vivo.
March 2001 8-84 DRAFT-DO NOT QUOTE OR CITE
-------
1 Even though it is clear that combustion particles that have a high content of soluble metals
2 can cause lung injury and even death in compromised animals, it has not been established that the
3 small quantities of metals associated with ambient PM are sufficient to cause health effects.
4 Moreover, it cannot be assumed that metals are the primary toxic component of ambient PM.
5 In studies in which various ambient and emission source particulates were instilled into rats, the
6 soluble metal content did appear to be the primary determinant of lung injury (Costa and Dreher,
7 1997). However, one published study has compared the effects of inhaled ROFA (at 1 mg/m3) to
8 concentrated ambient PM (four experiments, at mean concentrations of 475 to 900 yUg/m3) in
9 normal and SO2- induced bronchitic rats. A statistically significant increase in at least one lung
10 injury marker was seen in bronchitic rats with only one out of four of the concentrated ambient
11 exposures, whereas inhaled ROFA had no effect even though the content of soluble iron,
12 vanadium, and nickel was much higher in the ROFA sample than in the concentrated ambient
13 PM. Although the role of metals in contributing to health effects of ambient PM is not
14 established, the recent studies based on ROFA have important implications.
15
16 Ultraflne Particles
17 When this subject was reviewed in the 1996 PM AQCD (U. S. Environmental Protection
18 Agency, 1996a), it was not known whether the pulmonary toxicity of freshly generated ultrafine
19 teflon particles was due to particle size or a result of absorbed fumes. Subsequent studies with
20 other types of ultrafine particles have shown that the chemical constituents of ultrafines
21 substantially modulate their toxicity. For example, Kuschner et al. (1997) have established that
22 inhalation of MgO particles produces far fewer respiratory effects than does ZnO. Also,
23 inhalation exposure of normal rats to ultrafine carbon particles generated by electric arc discharge
24 (100 //g/m3 for 6 h) caused minimal lung inflammation (Elder et al., 2000), compared to ultrafine
25 Teflon or metal particles. On the other hand, instillation of 125 ^g of ultrafine carbon black
26 (20 nm) caused substantially more inflammation than did the same dose of fine particles of
27 carbon black (200 to 250 nm), suggesting that ultrafine particles may cause more inflammation
28 than larger particles (Li et al., 1997). However, the chemical constituents of the two sizes of
29 carbon black used in this study were not analyzed, and it cannot be assumed that the chemical
30 composition was the same for the two sizes. Thus, there is still insufficient toxicological
31 evidence to conclude that ambient concentrations of ultrafine particles contribute to the health
March 2001 8-85 DRAFT-DO NOT QUOTE OR CITE
-------
1 effects of particulate air pollution. However, with acid aerosols, studies of ultrafine particles
2 have focused largely on effects in the lung, and it is possible that inhaled ultrafine particles may
3 have systemic effects that are independent of effects on the lung.
4
5 Bioaerosols
6 Recent studies support the conclusion of the 1996 PM AQCD (U. S. Environmental
7 Protection Agency, 1996a), which stated that bioaerosols, at concentrations present in the
8 ambient environment, would not account for the reported health effects of ambient PM.
9 Dose-response studies in healthy volunteers exposed to 0.55 and 50 //g endotoxin, by the
10 inhalation route, showed a threshold for pulmonary and systemic effects for endotoxin between
11 0.5 and 5.0 //g (Michel et al., 1997). Monn and Becker (1999) examined effects of size
12 fractionated outdoor PM on human monocytes and found cytokine induction characteristic of
13 endotoxin activity in the coarse-size fraction but not in the fine fraction. Available information
14 suggests that ambient concentrations of endotoxin are very low and do not exceed 0.5 ng/m3.
15
16 Diesel Exhaust Particles
17 As described in Section 8.2.4.2, there is growing toxicological evidence that diesel PM
18 exacerbates the allergic response to inhaled antigens. The organic fraction of diesel exhaust has
19 been linked to eosinophil degranulation and induction of cytokine production, suggesting that the
20 organic constituents of diesel PM is responsible part for the immune effects. It is not known
21 whether the adjuvant-like activity of diesel PM is unique or whether other combustion particles
22 have similar effects. It is important to compare the immune effects of other source-specific
23 emissions, as well as concentrated ambient PM, to diesel PM to determine the extent to which
24 exposure to diesel exhaust may contribute to the incidence and severity of allergic rhinitis and
25 asthma.
26
27 Organic Compounds
28 Published research on the acute effects of particle-associated organic carbon constituents is
29 conspicuous by its relative absence, except for diesel exhaust particles. Like metals, organics are
30 common constituents of combustion-generated particles and have been found in ambient PM
31 samples over a wide geographical range. Organic carbon constituents comprise a substantial
March 2001 8-86 DRAFT-DO NOT QUOTE OR CITE
-------
1 portion of the mass of ambient PM (10 to 60% of the total dry mass [Turpin, 1999]). The
2 organic fraction of ambient PM has been evaluated for its mutagenic effects. Although the
3 organic fraction of ambient PM is a poorly characterized heterogeneous mixture of an unknown
4 number of different compounds, strategies have been proposed for examining the health effects
5 of this potentially important constituent (Turpin, 1999).
6
7 Ambient Particle Studies
8 Ambient particle studies should be the most relevant in understanding the susceptibility of
9 individuals to PM and the underlying mechanisms. Studies have used collected urban PM for
10 intratracheal administration to healthy and compromised animals. Despite the difficulties in
11 extrapolating from the bolus delivery used in such studies, they have provided strong evidence
12 that the chemical composition of ambient particles has a major influence on toxicity. More
13 recent work with inhaled concentrated ambient PM has observed cardiopulmonary changes in
14 rodents and dogs at high concentrations of fine PM. No comparative studies to examine the
15 effects of ultrafine and coarse ambient PM have been done, although a new ambient particle
16 concentrator developed by Sioutas and colleagues should permit the direct toxicological
17 comparison of various ambient particle sizes. Importantly, it has become evident that, although
18 the concentrated ambient PM studies can provide important dose-response information, identify
19 susceptibility factors in animal models, and permit examination of mechanisms related to PM
20 toxicity, they are not particularly well suited, however, for the identification of toxic components
21 in urban PM. Because only a limited number of exposures using concentrated ambient PM can
22 be reasonably conducted by a given laboratory in a particular urban environment, there may be
23 insufficient information to conduct a factor analysis on an exposure/response matrix. This may
24 also hinder principal component analysis techniques that are useful in identifying particle
25 components responsible for adverse outcomes.
26
27 8.7.1.2 Susceptibility
28 Progress has been made in understanding the role of individual susceptibility to ambient
29 PM effects. Studies have consistently shown that animals with compromised health, either
30 genetic or induced, are more susceptible to instilled or inhaled particles, although the increased
31 animal-to-animal variability in these models has created problems. Moreover, because PM
March 2001 8-87 DRAFT-DO NOT QUOTE OR CITE
-------
1 seems to affect broad categories of disease states, ranging from cardiac arrhythmias to pulmonary
2 infection, it can be difficult to know what disease models to use in understanding the biological
3 plausibility of the adverse health effects of PM. Thus, the identification of susceptible animal
4 models has been somewhat slow, but overall it represents solid progress when one considers that
5 data from millions of people are necessary in epidemiology studies to develop the statistical
6 power to detect small increases in PM-related morbidity and mortality.
7
8 8.7.2 Mechanisms of Action
9 The mechanisms that underlie the biological responses to ambient PM are not clear.
10 Various toxicologic studies using particulate matter having diverse physicochemical
11 characteristics have shown that these characteristics have a great impact on the specific response
12 that is observed. Thus, there may, in fact, be multiple biological mechanisms that may be
13 responsible for observed morbidity/mortality because of exposure to ambient PM, and these
14 mechanisms may be highly dependent on the type of particle in the exposure atmosphere.
15 However, it should be noted that many controlled exposure studies used particle concentrations
16 much higher than those typically occurring in ambient air. Thus, some of the mechanisms
17 elicited may not occur with exposure to lower levels. Clearly, controlled exposure studies have
18 not as yet been able to unequivocally determine the particle characteristics and the toxicological
19 mechanisms by which ambient PM may affect biological systems.
20
March 2001 8-88 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Amdur, M. O.; Chen, L. C. (1989) Furnace-generated acid aerosols: speciation and pulmonary effects.
3 In: Symposium on the health effects of acid aerosols; October 1987; Research Triangle Park, NC. Environ.
4 Health Perspect. 79: 147-150.
5 Amdur, M. O.; Dubriel, M.; Creasia, D. A. (1978) Respiratory response of guinea pigs to low levels of sulfuric acid.
6 Environ. Res. 15:418-423.
7 Ball, B. R.; Smith, K. R.; Veranth, J. M.; Aust, A. E. (2000) Bioavailability of iron from coal fly ash: mechanisms
8 of mobilization and of biological effects. In: Grant, L. D., ed. PM2000: particulate matter and health.
9 Inhalation Toxicol. 12(suppl. 4): 209-225.
10 Bayram, H.; Devalia, J. L.; Khair, O. A.; Abdelaziz, M. M.; Sapsford, R. J.; Sagai, M.; Davies, R. J. (1998a)
11 Comparison of ciliary activity and inflammatory mediator release from bronchial epithelial cells of nonatopic
12 nonasthmatic subjects and atopic asthmatic patients and the effect of diesel exhaust particles in vitro.
13 J. Allergy Clin. Immunol. 102: 771-782.
14 Bayram, H.; Devalia, J. L.; Sapsford, R. J.; Ohtoshi, T.; Miyabara, Y.; Sagai, M.; Davies, R. J. (1998b) The effect
15 of diesel exhaust particles on cell function and release of inflammatory mediators from human bronchial
16 epithelial cells in vitro. Am. J. Respir. Cell Mol. Biol. 18: 441-448.
17 Becker, S.; Soukup, J. M. (1998) Decreased CD1 IB expression, phagocytosis, and oxidative burst in urban
18 particulate pollution-exposed human monocytes and alveolar macrophages. J. Toxicol. Environ. Health Part
19 A 55: 455-477.
20 Becker, S.; Soukup, J. M.; Gilmour, M. I.; Devlin, R. B. (1996) Stimulation of human and rat alveolar macrophages
21 by urban air particulates: effects on oxidant radical generation and cytokine production. Toxicol. Appl.
22 Pharmacol. 141:637-648.
23 Blomberg, A.; Sainsbury, C.; Rudell, B.; Frew, A. J.; Holgate, S. T.; Sandstrom, T.; Kelly, F. J. (1998) Nasal cavity
24 lining fluid ascorbic acid concentration increases in healthy human volunteers following short term exposure
25 to diesel exhaust. Free Radical Res. 28: 59-67.
26 Bolarin, D. M.; Bhalla, D. K.; Kleinman, M. T. (1997) Effects of repeated exposures of geriatric rats to ozone and
27 particle-containing atmospheres: an analysis of bronchoalveolar lavage and plasma proteins. Inhalation
28 Toxicol. 9: 423-434.
29 Bonner, J. C.; Rice, A. B.; Lindroos, P. M.; O'Brien, P. O.; Dreher, K. L.; Rosas, I.; Alfaro-Moreno, E.;
30 Osornio-Vargas, A. R. (1998) Induction of the lung myofibroblast PDGF receptor system by urban ambient
31 particles from Mexico City. Am. J. Respir. Cell Mol. Biol. 19: 672-680.
32 Bouthillier, L.; Vincent, R.; Goegan, P.; Adamson, I. Y. R.; Bjarnason, S.; Stewart, M.; Guenette, J.; Potvin, M.;
33 Kumarathasan, P. (1998) Acute effects of inhaled urban particles and ozone: lung morphology, macrophage
34 activity, and plasma endothelin-1. Am. J. Pathol. 153: 1873-1884.
35 Brain, J. D.; Long, N. C.; Wolfthal, S. F.; Dumyahn, T.; Dockery, D. W. (1998) Pulmonary toxicity in hamsters of
36 smoke particles from Kuwaiti oil fires. Environ. Health Perspect. 106: 141-146.
37 Broeckaert, F.; Buchet, J. P.; Huaux, F.; Lardot, C.; Lison, D.; Yager, J. W. (1997) Reduction of the ex vivo
3 8 production of tumor necrosis factor alpha by alveolar phagocytes after administration of coal fly ash and
39 copper smelter dust. J. Toxicol. Environ. Health 51:189-202.
40 Campen, M. J.; Costa, D. L.; Watkinson, W. P. (2000) Cardiac and thermoregulatory toxicity of residual oil fly ash
41 in cardiopulmonary-compromised rats. Inhalation Toxicol. 12(suppl. 2): 7-22.
42 Carter, J. D.; Ohio, A. J.; Samet, J. M.; Devlin, R. B. (1997) Cytokine production by human airway epithelial cells
43 after exposure to an air pollution particle is metal-dependent. Toxicol. Appl. Pharmacol. 146: 180-188.
44 Chen, L. C.; Fine, J. M.; Qu, Q.-S.; Amdur, M. O.; Gordon, T. (1992) Effects of fine and ultrafine sulfuric acid
45 aerosols in guinea pigs: alterations in alveolar macrophage function and intracellular pH. Toxicol. Appl.
46 Pharmacol. 113: 109-117.
47 Chen, L. C.; Fang, C. P.; Qu, Q. S.; Fine, J. M.; Schlesinger, R. B. (1993) A novel system for the in vitro exposure
48 of pulmonary cells to acid sulfate aerosols. Fundam. Appl. Toxicol. 20: 170-176.
49 Churg, A.; Stevens, B.; Wright, J. L. (1998) Comparison of the uptake of fine and ultrafine TiO2 in a tracheal
50 explant system. Am. J. Physiol. 274: L81-L86.
51 Clarke, R. W.; Catalano, P. J.; Koutrakis, P.; Krishna Murthy, G. G.; Sioutas, C.; Paulauskis, J.; Coull, B.;
52 Ferguson, S.; Godleski, J. J. (1999) Urban air particulate inhalation alters pulmonary function and induces
53 pulmonary inflammation in a rodent model of chronic bronchitis. Inhalation Toxicol. 11: 637-656.
March 2001 8-89 DRAFT-DO NOT QUOTE OR CITE
-------
1 Clarke, R. W.; Catalano, P.; Coull, B.; Koutrakis, P.; Krishna Murthy, G. G.; Rice, T.; Godleski, J. J. (2000)
2 Age-related responses in rats to concentrated urban air particles (CAPs). In: Inhalation Toxicology:
3 proceedings of the third colloquium on particulate air pollution and human health; June, 1999; Durham, NC.
4 Inhalation Toxicology 129(suppl. 1): 73-84.
5 Cohen, M. D.; Becker, S.; Devlin, R.; Schlesinger, R. B.; Zelikoff, J. T. (1997) Effects of vanadium upon
6 polyl:C-induced responses in rat lung and alveolar macrophages. J. Toxicol. Environ. Health 51: 591-608.
7 Cormier, Y.; Laviolette, M.; Bedard, G.; Dosman, J.; Israel-Assayag, E. (1998) Effect of route of breathing on
8 response to exposure in a swine confinement building. Am. J. Respir. Crit. Care Med. 157: 1512-1521.
9 Costa, D. L.; Dreher, K. L. (1997) Bioavailable transition metals in particulate matter mediate cardiopulmonary
10 injury in healthy and compromised animal models. In: Driscoll, K. E.; Oberdorster, G., eds. Proceedings of
11 the sixth international meeting on the toxicology of natural and man-made fibrous and non-fibrous particles;
12 September 1996; Lake Placid, NY. Environ. Health Perspect. Suppl. 105(5): 1053-1060.
13 Creutzenberg, O.; Bellmann, B.; Muhle, H.; Dasenbrock, C.; Morrow, P.; Mermelstein, R. (1998) Lung clearance
14 and retention of toner, TiO2, and crystalline silica, utilizing a tracer technique during chronic inhalation
15 exposure in Syrian Golden hamsters. Inhalation Toxicol. 10: 731-751.
16 Diaz-Sanchez, D.; Tsien, A.; Casillas, A.; Dotson, A. R.; Saxon, A. (1996) Enhanced nasal cytokine production in
17 human beings after in vivo challenge with diesel exhaust particles. J. Allergy Clin. Immunol. 98: 114-123.
18 Diaz-Sanchez, D.; Tsien, A.; Fleming, J.; Saxon, A. (1997) Combined diesel exhaust particulate and ragweed
19 allergen challenge markedly enhances human in vivo nasal ragweed-specific IgE and skews cytokine
20 production to a T helper cell 2-type pattern. J. Immunol. 158: 2406-2413.
21 Donaldson, K.; Brown, D. M.; Mitchell, C.; Dineva, M.; Beswick, P. H.; Gilmour, P.; MacNee, W. (1997) Free
22 radical activity of PM10: iron-mediated generation of hydroxyl radicals. In: Driscoll, K. E.; Oberdorster, G.,
23 eds. Proceedings of the sixth international meeting on the toxicology of natural and man-made fibrous and
24 non-fibrous particles; September 1996; Lake Placid, NY. Environ. Health Perspect. Suppl. 105(5):
25 1285-1289.
26 Dong, W.; Lewtas, J.; Luster, M. I. (1996) Role of endotoxin in tumor necrosis factor ocexpression from alveolar
27 macrophages treated with urban air particles. Exp. Lung Res. 22: 577-592.
28 Dormans, J. A. M. A.; Steerenberg, P. A.; Arts, J. H. E.; Van Bree, L.; De Klerk, A.; Verlaan, A. P. J.; Bruijntjes,
29 J. P.; Beekhof, P.; Van Soolingen, D.; Van Loveren, H. (1999) Pathological and immunological effects of
30 respirable coal fly ash in male wistar rats. Inhalation Toxicol. 11: 51-69.
31 Dreher, K. L.; Jaskot, R. H.; Lehmann, J. R.; Richards, J. H.; McGee, J. K.; Ohio, A. J.; Costa, D. L. (1997)
32 Soluable transition metals mediate residual oil fly ash induced acute lung injury. J. Toxicol. Environ. Health
33 50:285-305.
34 Driscoll, K. E.; Costa, D. L.; Hatch, G.; Henderson, R.; Oberdorster, G.; Salem, H.; Schlesinger, R. B. (2000)
35 Intratracheal instillation as an exposure technique for the evaluation of respiratory tract toxicity: uses and
36 limitations. Toxicol. Sci. 55: 24-35.
37 Dye, J. A.; Adler, K. B.; Richards, J. H.; Dreher, K. L. (1997) Epithelial injury induced by exposure to residual oil
38 fly-ash particles: role of reactive oxygen species? Am. J. Respir. Cell Mol. Biol. 17: 625-633.
39 Dye, J. A.; Adler, K. B.; Richards, J. H.; Dreher, K. L. (1999) Role of soluble metals in oil fly ash-induced airway
40 epithelial injury and cytokine gene expression. Am. J. Physio). 277: L498-L510.
41 Elder, A. C. P.; Gelein, R.; Finkelstein J. N.; Cox, C.; Oberdorster, G. (2000) Endotoxin priming affects the lung
42 response to ultrafine particles and ozone in young and old rats. In: Inhalation Toxicology: proceedings of the
43 third colloquium on particulate air pollution and human health; June, 1999; Durham, NC. Inhalation
44 Toxicology 12(suppl. 1): 85-98.
45 Fabiani, R.; Pampanella, L.; Minelli, A.; Mezzasoma, I.; Vecchiarelli, A.; Morozzi, G. (1997) Effect of airborne
46 particulate extracts on monocyte oxidative metabolism. J. Environ. Pathol. Toxicol. Oncol. 16: 195-199.
47 Ferin, J.; Oberdorster, G.; Penney, D. P.; Soderholm, S. C.; Gelein, R.; Piper, H. C. (1990) Increased pulmonary
48 toxicity of ultrafine particles? I. Particle clearance, translocation, morphology. J. Aerosol Sci. 21: 381-384.
49 Ferin, J.; Oberdorster, G.; Penney, D. P. (1992) Pulmonary retention of ultrafine and fine particles in rats. Am. J.
50 Respir. Cell Mol. Biol. 6: 535-542.
51 Fine, J. M.; Gordon, T.; Chen, L. C.; Kinney, P.; Falcone, G.; Beckett, W. S. (1997) Metal fume fever:
52 characterization of clinical and plasma IL-6 responses in controlled human exposures to zinc oxide fume at
53 and below the threshold limit value. J. Occup. Environ. Med. 39: 722-726.
54
March 2001 8-90 DRAFT-DO NOT QUOTE OR CITE
-------
1 Finkelstein, J. N.; Johnston, C.; Barrett, T.; Oberdorster, G. (1997) Particulate-cell interactions and pulmonary
2 cytokine expression. In: Driscoll, K. E.; Oberdorster, G., eds. Proceedings of the sixth international meeting
3 on the toxicology of natural and man-made fibrous and non-fibrous particles; September 1996; Lake Placid,
4 NY. Environ. Health Perspect. Suppl. 105(5): 1179-1182.
5 Folinsbee, L. J.; Kim, C. S.; Kehrl, H. R.; Prah, J. D.; Devlin, R. B. (1997) Methods in human inhalation toxicology.
6 In: Massaro, E. J., ed. Handbook of human toxicology. Boca Raton, FL: CRC Press LLC; pp. 607-670.
7 Fortoul, T. I.; Osorio, L. S.; Tovar, A. T.; Salazar, D.; Castilla, M. E.; Olaiz-Fernandez, G. (1996) Metals in lung
8 tissue from autopsy cases in Mexico City residents: comparison of cases from the 1950s and the 1980s.
9 Environ. Health Perspect. 104: 630-632.
10 Frampton, M. W.; Voter, K. Z.; Morrow, P. E.; Roberts, N. J., Jr.; Culp, D. J.; Cox, C.; Utell, M. J. (1992) Sulfuric
11 acid aerosol exposure in humans assessed by bronchoalveolar lavage. Am. Rev. Respir. Dis. 146: 626-632.
12 Frampton, M. W.; Morrow, P. E.; Cox, C.; Levy, P. C.; Condemi, J. J.; Speers, D.; Gibb, F. R.; Utell, M. J. (1995)
13 Sulfuric acid aerosol followed by ozone exposure in healthy and asthmatic subjects. Environ. Res. 69: 1-14.
14 Frampton, M. W.; Ohio, A. J.; Samet, J. M.; Carson, J. L.; Carter, J. D.; Devlin, R. B. (1999) Effects of aqueous
15 extracts of PM10 filters from the Utah valley on human airway epithelial cells. Am. J. Physiol.
16 277: L960-L967.
17 Fujimaki, H.; Katayama, N.; Wakamori, K. (1992) Enhanced histamine release from lung mast cells of guinea pigs
18 exposed to sulfuric acid aerosols. Environ. Res. 58: 117-123.
19 Gardner, S. Y.; Lehmann, J. R.; Costa, D. L. (2000) Oil fly ash-induced elevation of plasma fibrinogen levels in
20 rats. Toxicol. Sci. 56: 175-180.
21 Gavett, S. H.; Madison, S. L.; Dreher, K. L.; Winsett, D. W.; McGee, J. K.; Costa, D. L. (1997) Metal and sulfate
22 composition of residual oil fly ash determines airway hyperreactivity and lung injury in rats. Environ. Res.
23 72: 162-172.
24 Gavett, S. H.; Madison, S. L.; Stevens, M. A.; Costa, D. L. (1999) Residual oil fly ash amplifies allergic cytokines,
25 airway responsiveness, and inflammation in mice. Am. J. Respir. Crit. Care Med. 160: 1897-1904.
26 Gearhart, J. M.; Schlesinger, R. B. (1986) Sulfuric acid-induced airway hyperresponsiveness. Fundam. Appl.
27 Toxicol. 7: 681-689.
28 Gerchenetal. (1996)
29 Ohio, A. J.; Meng, Z. H.; Hatch, G. E.; Costa, D. L. (1997a) Luminol-enhanced chemiluminescence after in vitro
30 exposures of rat alveolar macrophages to oil fly ash is metal dependent. Inhalation Toxicol. 9: 255-271.
31 Ohio, A. J.; Piantadosi, C. A.; Crumbliss, A. L. (1997b) Hypothesis: iron chelation plays a vital role in neutrophilic
32 inflammation. BioMetals 10: 135-142.
33 Ghio, A. J.; Carter, J. D.; Richards, J. H.; Brighton, L. E.; Lay, J. C.; Devlin, R. B. (1998a) Disruption of normal
34 iron homeostasis after bronchial instillation of an iron-containing particle. Am. J. Physiol. 274: L396-L403.
35 Ghio, A. J.; Richards, J. H.; Dittrich, K. L.; Samet, J. M. (1998b) Metal storage and transport proteins increase after
36 exposure of the rat lung to an air pollution particle. Toxicol. Pathol. 26: 388-394.
37 Ghio, A. J.; Carter, J. D.; Samet, J. M.; Reed, W.; Quay, J.; Dailey, L. A.; Richards, J. H.; Devlin, R. B. (1998c)
38 Metal-dependent expression of ferritin and lactoferrin by respiratory epithelial cells. Am. J. Physiol.
39 274: L728-L736.
40 Ghio, A. J.; Stonehuerner, J.; Dailey, L. A.; Carter, J. D. (1999a) Metals associated with both the water-soluble and
41 insoluble fractions of an ambient air pollution particle catalyze an oxidative stress. Inhalation Toxicol.
42 11:37-49.
43 Ghio, A. J.; Carter, J. D.; Dailey, L. A.; Devlin, R. B.; Samet, J. M. (1999b) Respiratory epithelial cells demonstrate
44 lactoferrin receptors that increase after metal exposure. Am. J. Physiol. 276: L933-L940.
45 Ghio, A. J.; Stoneheurner, J.; McGee, J. K.; Kinsey, J. S. (1999c) Sulfate content correlates with iron concentrations
46 in ambient air pollution particles. Inhalation Toxicol. 11: 293-307.
47 Ghioetal. (1999d)
48 Ghio, A. J.; Kim, C.; Devlin, R. B. (2000a) Concentrated ambient air particles induce mild pulmonary inflammation
49 in healthy human volunteers. Am. J. Respir. Crit. Care Med. 162: 981-988.
50 Ghio, A. J.; Carter, J. D.; Richards, J. H.; Crissman, K. M.; Bobb, H. H.; Yang, F. (2000b) Diminished injury in
51 hypotransferrinemic mice after exposure to a metal-rich particle. Am. J. Physiol. 278: L1051-LI061.
52 Gift, J. S.; Faust, R. A. (1997) Noncancer inhalation toxicology of crystalline silica: exposure-response assessment.
53 J. Exposure Anal. Environ. Epidemiol. 7: 345-358.
54 Gilmour, M. I.; Taylor, F. G. R.; Baskerville, A.; Wathes, C. M. (1989a) The effect of titanium dioxide inhalation
55 on the pulmonary clearance of Pasteurella haemolytica in the mouse. Environ. Res. 50: 157-172.
March 2001 8-91 DRAFT-DO NOT QUOTE OR CITE
-------
1 Gilmour, M. I.; Taylor, F. G. R.; Wathes, C. M. (1989b) Pulmonary clearance of Pasteurella haemolytica and
2 immune responses in mice following exposure to titanium dioxide. Environ. Res. 50: 184-194.
3 Gilmour, P. S.; Brown, D. M.; Lindsay, T. G.; Beswick, P. H.; MacNee, W.; Donaldson, K. (1996) Adverse health
4 effects of PM10 particles: involvement of iron in generation of hydroxyl radical. Occup. Environ. Med.
5 53:817-822.
6 Godleski, J. J.; Verrier, R. L.; Koutrakis, P.; Catalano, P. (2000) Mechanisms of morbidity and mortality from
7 exposure to ambient air particles. Cambridge, MA: Health Effects Institute; research report no. 91.
8 Goldsmith, C.-A.; Frevert, C.; Imrich, A.; Sioutas, C.; Kobzik, L. (1997) Alveolar macrophage interaction with air
9 pollution particulates. In: Driscoll, K. E.; Oberdorster, G., eds. Proceedings of the sixth international meeting
10 on the toxicology of natural and man-made fibrous and non-fibrous particles; September 1996; Lake Placid,
11 NY. Environ. Health Perspect. Suppl. 105(5): 1191-1195.
12 Goldsmith, C.-A. W.; Imrich, A.; Danaee, H.; Ning, Y. Y.; Kobzik, L. (1998) Analysis of air pollution
13 particulate-mediated oxidant stress in alveolar macrophages. J. Toxicol. Environ. Health Part A 54: 529-545.
14 Goldsmith, C.-A. W.; Hamada, K.; Ning, Y. Y.; Qin, G.; Catalano, P.; Murthy, G. G. K.; Lawrence, J.; Kobzik, L.
15 (1999) The effects of environmental aerosols on airway hyperresponsiveness in a murine model of asthma.
16 Inhalation Toxicol. 11: 981-998.
17 Gordon, T.; Nadziejko, C.; Schlesinger, R.; Chen, L. C. (1998) Pulmonary and cardiovascular effects of acute
18 exposure to concentrated ambient particulate matter in rats. Toxicol. Lett. 96-97: 285-288.
19 Gordon, T.; Nadziejko, C.; Chen, L. C.; Schlesinger, R. (2000) Effects of concentrated ambient particles in rats and
20 hamsters: an exploratory study. Cambridge, MA: Health Effects Institute; research report no. 93.
21 Grabowski, G. M.; Paulauskis, J. D.; Godleski, J. J. (1999) Mediating phosphorylation events in the
22 vanadium-induced respiratory burst of alveolar macrophages. Toxicol. Appl. Pharmacol. 156: 170-178.
23 Gulisano, M.; Marceddu, S.; Barbara, A.; Pacini, A.; Buiatti, E.; Martini, A.; Pacini, P. (1997) Damage to the
24 nasopharyngeal mucosa induced by current levels of urban air pollution: a field study in lambs. Eur. Respir.
25 J. 10:567-572.
26 Hahon, N.; Booth, J. A.; Green, F.; Lewis, T. R. (1985) Influenza virus infection in mice after exposure to coal dust
27 and diesel engine emissions. Environ. Res. 37: 44-60.
28 Hamada, K.; Goldsmith, C.-A.; Kobzik, L. (1999) Increased airway hyperresponsiveness and inflammation in a
29 juvenile mouse model of asthma exposed to air-pollutant aerosol. J. Toxicol. Environ. Health 58: 129-143.
30 Hamada, K.; Goldsmith, C.-A.; Goldman, A.; Kobzik, L. (2000) Resistance of very young mice to inhaled allergen
31 sensitization Is overcome by coexposure to an air-pollutant aerosol. Am. J. Respir. Crit. Care Med.
32 161:1285-1293.
33 Hatch, G. E.; Boykin, E.; Graham, J. A.; Lewtas, J.; Pott, F.; Loud, K.; Mumford, J. L. (1985) Inhalable particles
34 and pulmonary host defense: in vivo and in vitro effects of ambient air and combustion particles. Environ.
35 Res. 36: 67-80.
36 Health Effects Institute. (1995) Diesel exhaust: a critical analysis of emissions, exposure, and health effects:
37 a special report of the Institute's Diesel Working Group. Cambridge, MA: Health Effects Institute.
38 Heyder, J.; Beck-Speier, I.; Busch, B.; Dirscherl, P.; Heilmann, P.; Ferron, G. A.; Josten, M.; Karg, E.; Kreyling,
39 W. G.; Lenz, A.-G.; Maier, K. L.; Miaskowski, U.; Platz, S.; Reitmeir, P.; Schulz, H.; Takenaka, S.;
40 Ziesenis, A. (1999) Health effects of sulfur-related environmental air pollution. I. Executive summary.
41 Inhalation Toxicol. 11: 343-359.
42 Hitzfeld, B.; Friedrichs, K. H.; Ring, J.; Behrendt, H. (1997) Airborne particulate matter modulates the production
43 of reactive oxygen species in human polymorphonuclear granulocytes. Toxicology 120: 185-195.
44 Holian, A.; Hamilton, R. F., Jr.; Morandi, M. T.; Brown, S. D.; Li, L. (1998) Urban particle-induced apoptosis and
45 phenotype shifts in human alveolar macrophages. Environ. Health Perspect. 106: 127-132.
46 Irsigler, G. B.; Visser, P. J.; Spangenberg, P. A. L. (1999) Asthma and chemical bronchitis in vanadium plant
47 workers. Am. J. Ind. Med. 35: 366-374.
48 Jacobs, J.; Kreutzer, R.; Smith, D. (1997) Rice burning and asthma hospitalizations, Butte County, California,
49 1983-1992. Environ. Health Perspect. 105: 980-985.
50 Jakab, G. J. (1992) Relationship between carbon black particulate-bound formaldehyde, pulmonary antibacterial
51 defenses, and alveolar macrophage phagocytosis. Inhalation Toxicol. 4: 325-342.
52 Jakab, G. J. (1993) The toxicologic interactions resulting from inhalation of carbon black and acrolein on
53 pulmonary antibacterial and antiviral defenses. Toxicol. Appl. Pharmacol. 121: 167-175.
March 2001 8-92 DRAFT-DO NOT QUOTE OR CITE
-------
1 Jakab, G. J.; Hemenway, D. R. (1993) Inhalation coexposure to carbon black and acrolein suppresses alveolar
2 macrophage phagocytosis and TNF-arelease and modulates peritoneal macrophage phagocytosis. Inhalation
3 Toxicol. 5: 275-289.
4 Jakab, G. J.; Clarke, R. W.; Hemenway, D. R.; Longphre, M. V.; Kleeberger, S. R.; Frank, R. (1996) Inhalation of
5 acid coated carbon black particles impairs alveolar macrophage phagocytosis. Toxicol. Lett. 88: 243-248.
6 Johnston, C. J.; Finkelstein, J. N.; Gelein, R.; Baggs, R.; Oberdorster, G. (1996) Characterization of the early
7 pulmonary inflammatory response associated with PTFE fume exposure. Toxicol. Appl. Pharmacol.
8 140: 154-163.
9 Johnston, C. J.; Finkelstein, J. N.; Gelein, R.; Oberdorster, G. (1998) Pulmonary inflammatory responses and
10 cytokine and antioxidant mRNA levels in the lungs of young and old C57BL/6 mice after exposure to Teflon
11 fumes. Inhalation Toxicol. 10: 931-953.
12 Kadiiska, M. B.; Mason, R. P.; Dreher, K. L.; Costa, D. L.; Ohio, A. J. (1997) In vivo evidence of free radical
13 formation in the rat lung after exposure to an emission source air pollution particle. Chem. Res. Toxicol.
14 10:1104-1108.
15 Kennedy, T.; Ghio, A. J.; Reed, W.; Samet, J.; Zagorski, J.; Quay, J.; Carter, J.; Dailey, L.; Hoidal, J. R.; Devlin,
16 R. B. (1998) Copper-dependent inflammation and nuclear factor-KB activation by particulate air pollution.
17 Am. J. Respir. Cell Mol. Biol. 19: 366-378.
18 Killingsworth, C. R.; Alessandrini, F.; Krishna Murthy, G. G.; Catalano, P. J.; Paulauskis, J. D.; Godleski, J. J.
19 (1997) Inflammation, chemokine expression, and death in monocrotaline-treated rats following fuel oil fly
20 ash inhalation. Inhalation Toxicol. 9: 541-565.
21 Kimmel, T. A.; Chen, L. C.; Bosland, M. C.; Nadziejko, C. (1997) Influence of acid aerosol droplet size on
22 structural changes in the rat lung caused by acute exposure to sulfuric acid and ozone. Toxicol. Appl.
23 Pharmacol. 144: 348-355.
24 Kitabatake, M.; Imai, M.; Kasama, K.; Kobayashi, I.; Tomita, Y.; Yoshida, K. (1979) Effects of air pollutants on the
25 experimental induction of asthma attacks in guinea pigs: sulfuric acid mist and mixture of the mist and sulfur
26 dioxide. Mie Med. J. 29: 29-36.
27 Kleeberger, S. R.; Levitt, R. C.; Zhang, L.-Y.; Longphre, M.; Harkema, J.; Jedlicka, A.; Eleff, S. M.;
28 DiSilvestre, D.; Holroyd, K. J. (1997) Linkage analysis of susceptibility to ozone-induced lung inflammation
29 in inbred mice. Nat. Genet. 17: 475-478.
30 Kleinman, M. T.; Leaf, D. A.; Kelly, E.; Caiozzo, V.; Osann, K.; O'Niell, T. (1998) Urban angina in the mountains:
31 effects of carbon monoxide and mild hypoxemia on subjects with chronic stable angina. Arch. Environ.
32 Health 53: 388-397.
33 Kleinman, M. T.; Mautz, W. J.; Bjarnason, S. (1999) Adaptive and non-adaptive responses in rats exposed to ozone,
34 alone and in mixtures, with acidic aerosols. Inhalation Toxicol. 11: 249-264.
35 Knox, R. B.; Suphioglu, C.; Taylor, P.; Desai, R.; Watson, H. C.; Peng, J. L.; Bursill, L. A. (1997) Major grass
36 pollen allergen Lol p 1 binds to diesel exhaust particles: implications for asthma and air pollution. Clin. Exp.
37 Allergy 27: 246-251.
38 Kobzik, L. (1995) Lung macrophage uptake of unopsonized environmental particulates: role of scavenger-type
39 receptors. J. Immunol. 155: 367-376.
40 Kodavanti, U.; Costa, D. L. (1999) Animal models to study for pollutant effects. In: Holgate, S. T.; Koren, H. S.;
41 Samet, J. M.; Maynard, R. L., eds. Air pollution and health. London, United Kingdom: Academic Press;
42 pp. 165-197.
43 Kodavanti, U. P.; Jaskot, R. H.; Bonner, J.; Badgett, A.; Dreher, K. L. (1996) Eosmophilic lung inflammation in
44 particulate-induced lung injury: technical consideration in isolating RNA for gene expression studies.
45 Exp. Lung Res. 22: 541-554.
46 Kodavanti, U. P.; Jaskot, R. H.; Su, W. Y.; Costa, D. L.; Ghio, A. J.; Dreher, K. L. (1997a) Genetic variability in
47 combustion particle-induced chronic lung injury. Am. J. Physiol. 272: L521-L532.
48 Kodavanti, U. P.; Jaskot, R. H.; Costa, D. L.; Dreher, K. L. (1997b) Pulmonary proinflammatory gene induction
49 following acute exposure to residual oil fly ash: roles of particle-associated metals. Inhalation Toxicol.
50 9:679-701.
51 Kodavanti, U. P.; Hauser, R.; Christiani, D. C.; Meng, Z. H.; McGee, J.; Ledbetter, A.; Richards, J.; Costa, D. L.
52 (1998a) Pulmonary responses to oil fly ash particles in the rat differ by virtue of their specific soluble metals.
53 Toxicol. Sci. 43: 204-212.
54 Kodavanti, U. P.; Costa, D. L.; Bromberg, P. A. (1998b) Rodent models of cardiopulmonary disease: their potential
55 applicability in studies of air pollutant susceptibility. Environ. Health Perspect. 106(suppl. 1): 111-130.
March 2001 8-93 DRAFT-DO NOT QUOTE OR CITE
-------
1 Kodavanti, U. P.; Jackson, M. C.; Ledbetter, A. D.; Richards, J. R.; Gardner, S. Y.; Watkinson, W. P.; Campen,
2 M. J.; Costa, D. L. (1999) Lung injury from intratracheal and inhalation exposures to residual oil fly ash in a
3 rat model of monocrotaline-induced pulmonary hypertension. J. Toxicol. Environ. Health Part A
4 57:101-121.
5 Kodavanti, U. P.; Mebane, R.; Ledbetter, A.; Krantz, T.; McGee, J.; Jackson, M. C.; Walsh, L.; Hilliard, H.; Chen,
6 B. Y.; Richards, J.; Costa, D. L. (2000a) Variable pulmonary responses from exposure to concentrated
7 ambient air particles in a rat model of bronchitis. Toxicol. Sci. 54: 441-451.
8 Kodavanti, U. P.; Schladweiler, M. C.; Ledbetter, A. D.; Watkinson, W. P.; Campen, M. J.; Winsett, D. W.;
9 Richards, J. R.; Crissman, K. M.; Hatch, G. E.; Costa, D. L. (2000b) The spontaneously hypertensive rat as a
10 model of human cardiovascular disease: evidence of exacerbated cardiopulmonary injury and oxidative stress
11 from inhaled emission particulate matter. Toxicol. Appl. Pharmacol. 164: 250-263.
12 Kodavanti, U. P.; Schladweiler, M. C. J.; Richards, J. R.; Costa, D. L. (2001) Acute lung injury from intratracheal
13 exposure to fugitive residual oil fly ash and its constituent metals in normo- and spontaneously hypertensive
14 rats. Inhalation Toxicol. 13: 37-54.
15 Kuschner, W. G.; Wong, H.; D'Alessandro, A.; Quinlan, P.; Blanc, P. D. (1997) Human pulmonary responses to
16 experimental inhalation of high concentration fine and ultrafine magnesium oxide particles. Environ. Health
17 Perspect. 105: 1234-1237.
18 Lambert, A. L.; Dong, W.; Winsett, D. W.; Selgrade, M. K.; Gilmour, M. I. (1999) Residual oil fly ash exposure
19 enhances allergic sensitization to house dust mite. Toxicol. Appl. Pharmacol. 158: 269-277.
20 Lambert, A. L.; Dong, W.; Selgrade, M. K.; Gilmour, M. I. (2000) Enhanced allergic sensitization by residual oil fly
21 ash particles is mediated by soluble metal constituents. Toxicol. Appl. Pharmacol. 165: 84-93.
22 Last, J. A.; Pinkerton, K. E. (1997) Chronic exposure of rats to ozone and sulfuric acid aerosol: biochemical and
23 structural responses. Toxicology 116: 133-146.
24 Last, J. A.; Hyde, D. M.; Chang, D. P. Y. (1984) A mechanism of synergistic lung damage by ozone and a
25 respirable aerosol. Exp. Lung Res. 7: 223-235.
26 Lay, J. C.; Bennett, W. D.; Kim, C. S.; Devlin, R. B.; Bromberg, P. A. (1998) Retention and intracellular
27 distribution of instilled iron oxide particles in human alveolar macrophages. Am. J. Respir. Cell Mol. Biol.
28 18:687-695.
29 Leduc, D.; Fally, S.; De Vuyst, P.; Wollast, R.; Yernault, J.-C. (1995) Acute exposure to realistic acid fog: effects
30 on respiratory function and airway responsiveness in asthmatics. Environ. Res. 71: 89-98.
31 Lee, J.-T.; Shin, D.; Chung, Y. (1999) Air pollution and daily mortality in Seoul and Ulsan, Korea. Environ. Health
32 Perspect. 107: 149-154.
33 Li, X. Y.; Gilmour, P. S.; Donaldson, K.; MacNee, W. (1996) Free radical activity and pro-inflammatory effects of
34 particulate air pollution (PM10) in vivo and in vitro. Thorax 51: 1216-1222.
35 Li, X. Y.; Gilmour, P. S.; Donaldson, K.; MacNee, W. (1997) In vivo and in vitro proinflammatory effects of
36 particulate air pollution (PM10). In: Driscoll, K. E.; Oberdorster, G., eds. Proceedings of the sixth
37 international meeting on the toxicology of natural and man-made fibrous and non-fibrous particles;
38 September 1996; Lake Placid, NY. Environ. Health Perspect. Suppl. 105(5): 1279-1283.
39 Li, X. Y.; Brown, D.; Smith, S.; MacNee, W.; Donaldson, K. (1999) Short-term inflammatory responses following
40 intratracheal instillation of fine and ultrafine carbon black in rats. Inhalation Toxicol. 11: 709-731.
41 Linn, W. S.; Shamoo, D. A.; Anderson, K. R.; Peng, R.-C.; Avol, E. L.; Hackney, J. D. (1994) Effects of prolonged,
42 repeated exposure to ozone, sulfuric acid, and their combination in healthy and asthmatic volunteers. Am. J.
43 Respir. Crit. Care Med. 150: 431 -440.
44 Linn, W. S.; Gong, H., Jr.; Shamoo, D. A.; Anderson, K. R.; Avol, E. L. (1997) Chamber exposures of children to
45 mixed ozone, sulfur dioxide, and sulfuric acid. Arch. Environ. Health 52: 179-187.
46 Lipsett, M.; Hurley, S.; Ostro, B. (1997) Air pollution and emergency room visits for asthma in Santa Clara County,
47 California. Environ. Health Perspect. 105: 216-222.
48 Lison, D.; Lardot, C.; Huaux, F.; Zanetti, G.; Fubini, B. (1997) Influence of particle surface area on the toxicity of
49 insoluble manganese dioxide dusts. Arch. Toxicol. 71: 725-729.
50 Lorz, C.; Lopez, J. (1997) Incidence of air pollution in the pulmonary surfactant system of the pigeon (Columbia
51 livid). Anat. Rec. 249: 206-212.
52 Madden, M. C.; Thomas, M. J.; Ghio, A. J. (1999) Acetaldehyde (CH3CHO) production in rodent lung after
53 exposure to metal-rich particles. Free Radical Biol. Med. 26: 1569-1577.
March 2001 8-94 DRAFT-DO NOT QUOTE OR CITE
-------
1 Madl, A. K.; Wilson, D. W.; Segall, H. J.; Pinkerton, K. E. (1998) Alteration in lung particle translocation,
2 macrophage function, and microfilament arrangement in monocrotaline-treated rats. Toxicol. Appl.
3 Pharmacol. 153:28-38.
4 Mauderly, J. L. (1993) Toxicological approaches to complex mixtures. Environ. Health Perspect. Suppl.
5 101(4): 155-165.
6 McKenna, I. M; Gordon, T.; Chen, L. C.; Anver, M. R.; Waalkes, M. P. (1998) Expression of metallothionein
7 protein in the lungs of Wistar rats and C57 and DBA mice exposed to cadmium oxide fumes. Toxicol. Appl.
8 Pharmacol. 153: 169-178.
9 Michel, O.; Nagy, A.-M.; Schroeven, M.; Duchateau, J.; Neve, J.; Fondu, P.; Sergysels, R. (1997) Dose-response
10 relationship to inhaled endotoxin in normal subjects. Am. J. Respir. Crit. Care Med. 156: 1157-1164.
11 Minami, M.; Endo, T.; Hamaue, N.; Hirafuji, M.; Mori, Y.; Hayashi, H.; Sagai, M.; Suzuki, A. K. (1999)
12 Electrocardiographic changes induced by diesel exhaust particles (DEP) in guinea pigs. Res. Comm. Mol.
13 Pathol. Pharmacol. 105: 67-76.
14 Monn, C.; Becker, S. (1999) Cytotoxicity and induction of proinflammatory cytokines from human monocytes
15 exposed to fine (PM2 5) and coarse particles (PM,0.2 5) in outdoor and indoor air. Toxicol. Appl. Pharmacol.
16 155:245-252.
17 Muggenberg, B. A.; Barr, E. B.; Cheng, Y. S.; Seagrave, J. C.; Tilley, L. P.; Mauderly, J. L. (2000) Effect of
18 inhaled residual oil fly ash on the electrocardiogram of dogs. In: Grant, L. D., ed. PM2000: particulate matter
19 and health. Inhalation Toxicol. 12(suppl. 4): 189-208.
20 Murphy, S. A.; BeruBe, K. A.; Pooley, F. D.; Richards, R. J. (1998) The response of lung epithelium to well
21 characterised fine particles. Life Sci. 62: 1789-1799.
22 Nadadur, S. S.; Schladweiler, M. C. J.; Kodavanti, U. P. (2000) A pulmonary rat gene array for screening altered
23 expression profiles in air pollutant-induced lung injury. Inhalation Toxicol. 12: 1239-1254.
24 Nadeau, D.; Vincent, R.; Kumarathasan, P.; Brook, J.; Dufresne, A. (1996) Cytotoxicity of ambient air particles to
25 rat lung macrophages: comparison of cellular and functional assays. Toxicol. In Vitro 10: 161-172.
26 Nakamura and Hayashida (1992)
27 National Institutes of Health. (1997) Guidelines for the diagnosis and management of asthma: expert panel report 2.
28 Bethesda, MD: U.S. Department of Health and Human Services, National Heart, Lung, and Blood Institute;
29 publication no. 97-4051.
30 Nel, A. E.; Diaz-Sanchez, D.; Ng, D.; Hiura, T.; Saxon, A. (1998) Enhancement of allergic inflammation by the
31 interaction between diesel exhaust particles and the immune system. J. Allergy Clin. Immunol. 102: 539-554.
32 Nemmar, A.; Delaunois, A.; Nemery, B.; Dessy-Doize, C.; Beckers, J.-F.; Sulon, J.; Gustin, P. (1999) Inflammatory
33 effect of intratracheal instillation of ultrafine particles in the rabbit: role of C-fiber and mast cells.
34 Toxicol. Appl. Pharmacol. 160: 250-261.
35 Oberdorster, G.; Ferin, J.; Gelein, R.; Soderholm, S. C.; Finkelstein, J. (1992) Role of the alveolar macrophage in
36 lung injury: studies with ultrafine particles. Environ. Health Perspect. 97: 193-199.
37 Oberdorster etal. (1995)
38 Oberdorster, G.; Cox, C.; Gelein, R. (1997) Intratracheal instillation versus intratracheal inhalation of tracer
39 particles for measuring lung clearance function. Exp. Lung Res. 23: 17-34.
40 Oberdorster, G.; Finkelstein, J. N.; Johnston, C.; Gelein, R.; Cox, C.; Baggs, R.; Elder, A. C. P. (2000) Acute
41 pulmonary effects of ultrafine particles in rats and mice. Cambridge, MA: Health Effects Institute; research
42 report no. 96.
43 Oettinger, R.; Drumm, K.; Knorst, M.; Krinyak, P.; Smolarski, R.; Kienast, K. (1999) Production of reactive oxygen
44 intermediates by human macrophages exposed to soot particles and asbestos fibers and increase in NF-KB
45 P50/pl05mRNA. Lung 177: 343-354.
46 Ohtoshi, T.; Takizawa, H.; Okazaki, H.; Kawasaki, S.; Takeuchi, N.; Ohta, K.; Ito, K. (1998) Diesel exhaust
47 particles stimulate human airway epithelial cells to produce cytokines relevant to airway inflammation in
48 vitro. J. Allergy Clin. Immunol. 101: 778-785.
49 Ohtsuka, Y.; Clarke, R. W.; Mitzner, W.; Brunson, K.; Jakab, G. J.; Kleeberger, S. R. (2000a) Interstrain variation
50 in murine susceptibility to inhaled acid-coated particles. Am. J. Physiol. L469-L476.
51 Ohtsuka, Y.; Brunson, K. J.; Jedlicka, A. E.; Mitzner, W.; Clarke, R. W.; Zhang, L.-Y.; Eleff, S. M.; Kleeberger,
52 S. R. (2000b) Genetic linkage analysis of susceptibility to particle exposure in mice. Am. J. Respir. Cell Mol.
53 Biol. 22:574-581.
54 Oortgiesen, M.; Veronesi, B.; Eichenbaum, G.; Kiser, P. F.; Simon, S. A. (2000) Residual oil fly ash and charged
55 polymers activate epithelial cells and nociceptive sensory neurons. Am. J. Physiol. 278: L683-L695.
March 2001 8-95 DRAFT-DO NOT QUOTE OR CITE
-------
1 Osier, M.; Oberdorster, G. (1997) Intratracheal inhalation vs intratracheal instillation: differences in particle effects.
2 Fundam. Appl. Toxicol. 40: 220-227.
3 Osier, M.; Baggs, R. B.; Oberdorster, G. (1997) Intratracheal instillation versus intratracheal inhalation: influence of
4 cytokines on inflammatory response. In: Driscoll, K. E.; Oberdorster, G., eds. Proceedings of the sixth
5 international meeting on the toxicology of natural and man-made fibrous and non-fibrous particles;
6 September 1996; Lake Placid, NY. Environ. Health Perspect. Suppl. 105(5): 1265-1271.
7 Osman, K.; Zejda, J. E.; Schiitz, A.; Mielzynska, D.; Elinder, C. G.; Vahter, M. (1998) Exposure to lead and other
8 metals in children from Katowice district, Poland. Int. Arch. Occup. Environ. Health 71: 180-186.
9 Peters, A.; Wichmann, H. E.; Tuch, T.; Heinrich, J.; Heyder, J. (1997) Respiratory effects are associated with the
10 number of ultrafme particles. Am. J. Respir. Crit. Care Med. 155: 1376-1383.
11 Pierce, L. M.; Alessandrini, F.; Godleski, J. J.; Paulauskis, J. D. (1996) Vanadium-induced chemokine mRNA
12 expression and pulmonary inflammation. Toxicol. Appl. Pharmacol. 138: 1-11.
13 Pinto, M.; Birnbaum, S. C.; Kadar, T.; Goldberg, G. M. (1979) Lung injury in mice induced by factors acting
14 synergistically with inhaled particulate antigen. Clin. Immunol. Immunopathol. 13: 361-368.
15 Prahalad, A. K.; Soukup, J. M.; Inmon, J.; Willis, R.; Ohio, A. J.; Becker, S.; Gallagher, J. E. (1999) Ambient air
16 particles: effects on cellular oxidant radical generation in relation to particulate elemental chemistry. Toxicol.
17 Appl. Pharmacol. 158:81-91.
18 Pritchard, R. J.; Ohio, A. J.; Lehmann, J. R.; Winsett, D. W.; Tepper, J. S.; Park, P.; Gilmour, M. I.; Dreher, K. L.;
19 Costa, D. L. (1996) Oxidant generation and lung injury after particulate air pollutant exposure increase with
20 the concentrations of associated metals. Inhalation Toxicol. 8: 457-477.
21 Prows, D. R.; Shertzer, H. G.; Daly, M. J.; Sidman, C. L.; Leikauf, G. D. (1997) Genetic analysis of ozone-induced
22 acute lung injury in sensitive and resistant strains of mice. Nat. Genet. 17: 471-474.
23 Quay, J. L.; Reed, W.; Samet, J.; Devlin, R. B. (1998) Air pollution particles induce IL-6 gene expression in human
24 airway epithelial cells via NF-KB activation. Am. J. Respir. Cell Mol. Biol. 19: 98-106.
25 Rose, C. S.; Martyny, J. W.; Newman, L. S.; Milton, D. K.; King, T. E., Jr.; Beebe, J. L.; McCammon, J. B.;
26 Hoffman, R. E.; Kreiss, K. (1998) "Lifeguard lung": endemic granulomatous pneumonitis in an indoor
27 swimming pool. Am. J. Public Health 88: 1795-1800.
28 Rudell, B.; Sandstrom, T.; Stjernberg, N.; Kolmodin-Hedman, B. (1990) Controlled diesel exhaust exposure in an
29 exposure chamber: pulmonary effects investigated with bronchoalveolar lavage. J. Aerosol Sci.
30 21(suppl. 1):S411-S414.
31 Sagai, M.; Furuyama, A.; Ichinose, T. (1996) Biological effects of diesel exhaust particles (DEP). III. Pathogenesis
32 of asthma like symptoms in mice. Free Radical Biol. Med. 21:199-209.
33 Samet, J. M.; Reed, W.; Ohio, A. J.; Devlin, R. B.; Carter, J. D.; Dailey, L. A.; Bromberg, P. A.; Madden, M. C.
34 (1996) Induction of prostaglandin H synthase 2 in human airway epithelial cells exposed to residual oil fly
35 ash. Toxicol. Appl. Pharmacol. 141: 159-168.
36 Samet, J. M.; Stonehuerner, J.; Reed, W.; Devlin, R. B.; Dailey, L. A.; Kennedy, T. P.; Bromberg, P. A.; Ohio, A. J.
37 (1997) Disruption of protein tyrosine phosphate homeostasis in bronchial epithelial cells exposed to oil fly
38 ash. Am. J. Physiol. 272: L426-L432.
39 Samet, J. M.; Graves, L. M.; Quay, J.; Dailey, L. A.; Devlin, R. B.; Ohio, A. J.; Wu, W.; Bromberg, P. A.; Reed, W.
40 (1998) Activation of MAPKs in human bronchial epithelial cells exposed to metals. Am. J. Physiol.
41 275: L551-L558.
42 Samet, J. M.; Ghio, A. J.; Costa, D. L.; Madden, M. C. (2000) Increased expression of cyclooxygenase 2 mediates
43 oil fly ash-induced lung injury. Exp. Lung Res. 26: 57-69.
44 Seagrave, J. C.; Nikula, K. J. (2000) Multiple modes of responses to air pollution particulate materials in
45 A549 alveolar type II cells. In: Grant, L. D., ed. PM2000: particulate matter and health. Inhalation Toxicol.
46 12(suppl. 4): 247-260.
47 Seaton, A.; MacNee, W.; Donaldson, K.; Godden, D. (1995) Particulate air pollution and acute health effects.
48 Lancet (8943): 176-178.
49 Schluter, T.; Berg, I.; Dorger, M.; Gercken, G. (1995) Effect of heavy metal ions on the release of reactive oxygen
50 intermediates by bovine alveolar macrophages. Toxicology 98: 47-55.
51 Shukla et al. (2000)
52 Silbajoris, R.; Ghio, A. J.; Samet, J. M.; Jaskot, R.; Dreher, K. L.; Brighton, L. E. (2000) In vivo and in vitro
53 correlation of pulmonary map kinase activation following metallic exposure. Inhalation Toxicol. 12: 453-468.
54 Sioutas, C.; Koutrakis, P.; Burton, R. M. (1995) A technique to expose animals to concentrated fine ambient
55 aerosols. Environ. Health Perspect. 103: 172-177.
March 2001 8-96 DRAFT-DO NOT QUOTE OR CITE
-------
1 Sjogren, B. (1997) Occupational exposure to dust: inflammation and ischaemic heart disease. Occup. Environ. Med.
2 54:466-469.
3 Smith, K. R.; Aust, A. E. (1997) Mobilization of iron from urban particulates leads to generation of reactive oxygen
4 species in vitro and induction of ferritin synthesis in human lung epithelial cells. Chem. Res. Toxicol.
5 10: 828-834.
6 Smith, K. R.; Veranth, J. M.; Lighty, J. S.; Aust, A. E. (1998) Mobilization of iron from coal fly ash was dependent
7 upon the particle size and the source of coal. Chem. Res. Toxicol. 11: 1494-1500.
8 Spengler, J. D.; Thurston, G. D. (1983) Mass and elemental composition of fine and coarse particles in six U.S.
9 cities. J. AirPollut. Control Assoc. 33: 1162-1171.
10 Steerenberg, P. A.; Dormans, J. A. M. A.; Van Doom, C. C. M.; Middendorp, S.; Vos, J. G.; Van Loveren, H.
11 (1999) A pollen model in the rat for testing adjuvant activity of air pollution components. Inhalation Toxicol.
12 11:1109-1122.
13 Stringer, B.; Kobzik, L. (1996) Alveolar macrophage uptake of the environmental particulate titanium dioxide:
14 role of surfactant components. Am. J. Respir. Cell Mol. Biol. 14: 155-160.
15 Stringer, B.; Kobzik, L. (1998) Environmental particulate-mediated cytokine production in lung epithelial cells
16 (A549): role of preexisting inflammation and oxidant stress. J. Toxicol. Environ. Health Part A 55: 31-44.
17 Stringer, B.; Imrich, A.; Kobzik, L. (1996) Lung epithelial cell (A549) interaction with unopsonized environmental
18 particulates: quantitation of particle-specific binding and IL-8 production. Exp. Lung Res. 22: 495-508.
19 Su, W.-Y.; Jaskot, R. H.; Richards, J.; Abramson, S. R.; Woessner, F. J.; Wu, W.-H.; Dreher, K. L. (2000a)
20 Induction of pulmonary matrilysin expression by combustion and ambient air particles. Am. J. Physiol.
21 279:L152-L160.
22 Su, W.-Y.; Jaskot, R. H.; Dreher, K. L. (2000b) Particulate matter induction of pulmonary gelatinase A, gelatinase
23 B, tissue inhibitor of metaloproteinase expression. Inhalation Toxicol. 12(suppl. 2): 105-119.
24 Takano, H.; Yoshikawa, T.; Ichinose, T.; Miyabara, Y.; Imaoka, K.; Sagai, M. (1997) Diesel exhaust particles
25 enhance antigen-induced airway inflammation and local cytokine expression in mice. Am. J. Respir. Crit.
26 Care Med. 156:36-42.
27 Takano, H.; Ichinose, T.; Miyabara, Y.; Shibuya, T.; Lim, H.-B.; Yoshikawa, T. Sagai, M. (1998) Inhalation of
28 diesel exhaust enhances allergen-related eosinophil recruitment and airway hyperresponsiveness in mice.
29 Toxicol. Appl. Pharmacol. 150: 328-337.
30 Terashima, T.; Wiggs, B.; English, D.; Hogg, J. C.; Van Eeden, S. F. (1997) Phagocytosis of small carbon particles
31 (PMio) by alveolar macrophages stimulates the release of polymorphonuclear leukocytes from bone marrow.
32 Am. J. Respir. Crit. Care Med. 155: 1441-1447.
33 Thurston, G. D.; Lippmann, M.; Scott, M. B.; Fine, J. M. (1997) Summertime haze air pollution and children with
34 asthma. Am. J. Respir. Crit. Care Med. 155: 654-660.
35 Timblin, C.; BeruBe, K.; Chrug, A.; Driscoll, K.; Gordon, T.; Hemenway, D.; Walsh, E.; Cummins, A. B.;
36 Vacek, P.; Mossman, B. (1998) Ambient particulate matter causes activation of the c-jun
37 kinase/stress-activated protein kinase cascade and DNA synthesis in lung epithelial cells. Cancer Res.
38 58:4543-4547.
39 Tsien, A.; Diaz-Sanchez, D.; Ma, J.; Saxon, A. (1997) The organic component of diesel exhaust particles and
40 phenanthrene, a major polyaromatic hydrocarbon constituent, enhances IgE production by IgE-secreting
41 EBV-transformed human B cells in vitro. Toxicol. Appl. Pharmacol. 142: 256-263.
42 Tsuchiyama, F.; Hisanaga, N.; Shibata, E.; Aoki, T.; Takagi, H.; Ando, T.; Takeuchi, Y. (1997) Pulmonary metal
43 distribution in urban dwellers. Int. Arch. Occup. Environ. Health 70: 77-84.
44 Turpin, B. J. (1999) Options for characterizing organic particulate matter. Environ. Sci. Technol. 33: 76A-79A.
45 U.S. Environmental Protection Agency. (1982) Air quality criteria for particulate matter and sulfur oxides. Research
46 Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
47 Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA; PB84-156777.
48 U.S. Environmental Protection Agency. (1989) An acid aerosols issue paper: health effects and aerometrics.
49 Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
50 Assessment Office; EPA report no. EPA-600/8-88-005F. Available from: NTIS, Springfield, VA;
51 PB91-125864.
52 U.S. Environmental Protection Agency. (1996a) Air quality criteria for particulate matter. Research Triangle Park,
53 NC: National Center for Environmental Assessment-RTP Office; report nos. EPA/600/P-95/001aF-cF. 3v.
March 2001 8-97 DRAFT-DO NOT QUOTE OR CITE
-------
1 U.S. Environmental Protection Agency. (1996b) Ambient levels and noncancer health effects of inhaled crystalline
2 and amorphous silica: health issue assessment [draft]. Research Triangle Park, NC: National Center for
3 Environmental Assessment; EPA report no. EPA/600/R-95/115.
4 U.S. Environmental Protection Agency. (1999) Health assessment document for diesel emissions, SAB review draft.
5 Washington, DC: Office of Research and Development; report no. EPA/600/8-90/057C.
6 Utell, M. J.; Morrow, P. E.; Hyde, R. W. (1983) Latent development of airway hyperreactivity in human subjects
7 after sulfuric acid aerosol exposure. J. Aerosol Sci. 14: 202-205.
8 Van Maanen, J. M.; Borm, P. J.; Knaapen, A.; Van Herwijnen, M.; Schilderman, P. A.; Smith, K. R.; Aust, A. E.;
9 Tomatis, M.; Fubini, B. (1999) In vitro effects of coal fly ashes: hydroxyl radical generation, iron release,
10 and DNA damage and toxicity in rat lung epithelial cells. Inhalation Toxicol. 11:1123-1142.
11 Vanda, B.; de Buen, N.; Jasso, R.; Valero, G.; Vargas, M. H.; Olmos, R.; Arreola, J. L.; Santillan, P.; Alonso, P.
12 (1998) Inflammatory cells and ferruginous bodies in bronchoalveolar lavage in urban dogs. Acta Cytol.
13 42:939-944.
14 Veronesi, B.; Oortgiesen, M.; Carter, J. D.; Devlin, R. B. (1999) Particulate matter initiates inflammatory cytokine
15 release by activation of capsaicin and acid receptors in a human bronchial epithelial cell line. Toxicol. Appl.
16 Pharmacol. 154: 106-115.
17 Vincent, R.; Bjarnason, S. G.; Adamson, I. Y. R.; Hedgecock, C.; Kumarathasan, P.; Guenette, J.; Potvin, M.;
18 Goegan, P.; Bouthillier, L. (1997) Acute pulmonary toxicity of urban particulate matter and ozone. Am. J.
19 Pathol. 151: 1563-1570.
20 Vogelzang, P. F. J.; Van Der Gulden, J. W. J.; Folgering, H.; Kolk, J. J.; Heederik, D.; Preller, L.; Tielen. M. J. M.;
21 Van Schayck, C. P. (1998) Endotoxin exposure as a major determinant of lung function decline in pig
22 farmers. Am. J. Respir. Crit. Care Med. 157: 15-18.
23 Watkinson, W. P.; Campen, M. J.; Costa, D. L. (1998) Cardiac arrhythmia induction after exposure to residual oil
24 fly ash particles in a rodent model of pulmonary hypertension. Toxicol. Sci. 41: 209-216.
25 Watkinson, W. P.; Campen, M. J.; Dreher, K. L.; Su, W.-Y.; Kodavanti, U.P.; Highfill, J. W.; Costa, D. L. (2000)
26 Thermoregulatory effects following exposure to particulate matter in healthy and
27 cardiopulmonary-compromised rats. J. Therm. Biol. 25: 131-137.
28 Wesselkamper, S. C.; Prows, D. R.; Biswas, P.; Willeke, K.; Bingham, E.; Leikauf, G. D. (2000) Genetic
29 susceptibility to irritant-induced acute lung injury in mice. Am. J. Physiol. 279: L575-L582.
30 Wheatley, L. M.; Platts-Mills, T. A. E. (1996) Perennial allergens and the asthma epidemic. Sci. Med. 3: 6-13.
31 Woodin, M. A.; Hauser, R.; Liu, Y.; Smith, T. J.; Siegel, P. D.; Lewis, D. M.; Tollerud, D. J.; Christiani, D. C.
32 (1998) Molecular markers of acute upper airway inflammation in workers exposed to fuel-oil ash. Am. J.
33 Respir. Crit. Care Med. 158: 182-187.
34 Yoshinari et al., 2000
35 Zelikoff, J. T.; Sisco, M.; Cohen, M. D.; Frampton, M. W.; Utell, M. J; Schlesinger, R. B. (1994) Interspecies
36 comparison of immunotoxicity of inhaled sulfuric acid. II. New Zealand white rabbits. In: 1994 international
37 conference sponsored by the American Lung Association and the American Thoracic Society; May; Boston,
38 MA. Am. J. Respir. Crit. Care Med. 149: A621.
39 Zelikoff, J. T.; Frampton, M. W.; Cohen, M. D.; Morrow, P. E.; Sisco, M.; Tsai, Y.; Utell, M. J.; Schlesinger, R. B.
40 (1997) Effects of inhaled sulfuric acid aerosols on pulmonary immunocompetence: a comparative study in
41 humans and animals. Inhalation Toxicol. 9: 731-752.
42 Zhang, T.; Huang, C.; Johns, E. J. (1997) Neural regulation of kidney function by the somatosensory system in
43 normotensive and hypertensive rats. Am. J. Physiol. 273: R1749-R1757.
44 Zock, J.-P.; Hollander, A.; Heederik, D.; Douwes, J. (1998) Acute lung function changes and low endotoxin
45 exposures in the potato processing industry. Am. J. Ind. Med. 33: 384-391.
46
March 2001 8-98 DRAFT-DO NOT QUOTE OR CITE
-------
i CHAPTER 9. INTEGRATIVE SYNTHESIS:
2 PARTICULATE MATTER ATMOSPHERIC SCIENCE,
3 AIR QUALITY, HUMAN EXPOSURE, DOSIMETRY,
4 AND HEALTH RISKS
5
6
7 9.1 INTRODUCTION
8 This chapter focuses on integration of key information on exposure-dose-response risk
9 assessment components drawn from the preceding detailed chapters, to provide a coherent
10 framework for assessment of human health risks posed by ambient particulate matter (PM) in the
11 United States. As such, the chapter updates the integrated assessment provided in the 1996
12 Particulate Matter Air Quality Criteria Document (1996 PM AQCD; U.S. Environmental
13 Protection Agency, 1996) of available scientific information regarding ambient PM sources,
14 exposures, and health risks as they pertain to the United States. This assessment must be
15 considered provisional at this time, pending public comment and Clean Air Scientific Advisory
16 Committee (CASAC) review of other earlier, more detailed chapters from which key findings
17 were extracted for discussion and preliminary integration here. More complete integration and
18 conclusions will be incorporated in chapter revisions to be made subsequent to the CASAC
19 review.
20 This chapter first provides background information on key features of atmospheric
21 particles, highlighting important distinctions between fine- and coarse-mode particles with regard
22 to their size, chemical composition, sources, atmospheric behavior, and potential human
23 exposure relationships—distinctions that collectively continue to suggest that fine- and coarse-
24 mode particles should be treated as two distinct subclasses of air pollutants. Information on
25 recent trends in U.S. concentrations of different ambient PM size and composition fractions (e.g.,
26 PM,0, PM2 5, and PM10_2 5) and ranges of variability seen in U.S. regions and urban air sheds also
27 is summarized to place the ensuing health effects discussions in perspective.
28 The chapter next summarizes key points regarding respiratory tract dosimetry, followed by
29 discussion of the extensive PM epidemiologic database that has expanded greatly during recent
30 years. The latter includes numerous new studies of populations throughout the world published
March 2001 9-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 since the 1996 PM AQCD that contain further evidence that serious health effects (mortality,
2 exacerbation of chronic disease, increased hospital admissions, etc.) are associated with
3 exposures to ambient levels of PM found in contemporary U.S. urban air sheds. Evaluations of
4 other possible explanations for the reported PM epidemiology results (e.g., effects of weather,
5 other co-pollutants, choice of models, etc.) also are discussed, ultimately leading to the
6 conclusion that the reported associations of PM exposure and effects are valid. The newer
7 evidence is then discussed that (a) further substantiates associations of such serious health effects
8 with U.S. ambient PM10 levels, (b) also more strongly establishes fine particles (as indexed by
9 various indicators, e.g., PM2 5) as likely being important contributors to the observed human
10 health effects, and (c) now provides additional information on associations between coarse-
11 fraction (PM10_2 5) particles and adverse health impacts. The overall coherence of the newer
12 epidemiologic database also is discussed, which strengthens the 1996 PM AQCD evaluation
13 suggesting a likely causal role of ambient PM in contributing to the reported effects.
14 The nature of the observed effects and the biological mechanisms that might underlie such
15 effects then are discussed. The discussion of potential mechanisms of injury examines ways in
16 which PM could induce health effects. The increased, but still limited, availability of new
17 experimental evidence necessary to evaluate or directly substantiate the viability of hypothesized
18 mechanisms is noted. Information concerning possible contributions of particular classes of
19 specific ambient PM constit aents also is summarized.
20 The chapter also provides information on the identification of population groups at special
21 risk for ambient PM effects and factors placing them at increased risk, which need to be
22 considered in generating risk estimates for the possible occurrence of PM-related health events in
23 the United States.
24
25
26 9.2 ATMOSPHERIC SCIENCE CONSIDERATIONS
27 As discussed in Chapter 2 of this document, airborne PM is not a single pollutant but many
28 classes of pollutants; each class consists of several to many individual chemical species. One
29 classification is based on the natural division of the atmospheric aerosol into fine- and coarse-
30 mode particles. Fine-mode particles, in general, are smaller than coarse-mode particles; they also
31 differ in many other aspects such as formation mechanisms, chemical composition, sources,
March 2001 9-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 physical behavior, human exposure relationships, and control approaches required for risk
2 reduction. Such differences alone are sufficient to justify consideration of fine- and coarse-mode
3 particles as separate pollutants, regardless of the extent or lack of evidence regarding differences
4 in respiratory tract dosimetry or associated health effects in laboratory animals or humans. The
5 various physical and chemical differences between fine- and coarse-mode particles, their sources,
6 ambient concentrations, factors affecting human exposure, and their respiratory tract deposition
7 are summarized concisely below as a prelude to discussion of key health effects associated with
8 ambient PM exposures and other information useful in assessing PM-related public health risks
9 in the United States.
10 Atmospheric particles originate from a variety of sources and possess a range of
11 morphological, chemical, physical, and thermodynamic properties. The composition and
12 behavior of airborne particles are linked with those of surrounding gases. Aerosol may be
13 defined as a suspension of solid or liquid particles in air and includes both the particles and all
14 vapor or gas phase components of air. However, the term aerosol often is used, as is PM, to refer
15 to the suspended particles only. A complete description of the atmospheric aerosol would
16 include an accounting of the size, morphology, and chemical composition of each particle and the
17 relative abundance of each particle type as a function of particle size.
18
19 9.2.1 Ambient Particulate Matter Size Distinctions
20 Atmospheric particles differ in density and are not always spherical. Therefore, their
21 diameters often are described by an "equivalent" diameter. The aerodynamic equivalent diameter
22 (AED), defined as the diameter of a spherical particle with a density of 1 g/cm3 that would have a
23 settling velocity equal to the particle in question, is important for particle transport, collection,
24 and respiratory tract deposition.
25 The distribution of particles with respect to size is an important physical parameter
26 governing their behavior. Because atmospheric particles cover several orders of magnitude in
27 particle size, size distributions often are expressed in terms of the logarithm of the particle
28 diameter (D), on the X-axis, and the differential concentration on the Y-axis. If the differential
29 concentration is plotted on a linear scale, the surface, volume, mass, or number of particles
30 between D and D + AD is proportional to the area under the curve. Atmospheric aerosol size
31 distributions frequently are approximated by a sum of log-normal distributions.
March 2001 9-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 The aerosol community uses various approaches or conventions in the classification of
2 particles by size, including modes, based on the observed size distributions in the atmosphere and
3 formation mechanisms and cut point, usually based on the 50% cut point of the specific sampling
4 device i.e., the particle size at which 50% of the particles enter and 50% of the particles are
5 excluded, as summarized below.
6 Atmospheric size distributions show that most atmospheric particles are quite small, below
7 0.1 jum, whereas most of the particle volume (and therefore most of the mass) and much of the
8 surface area is found in particles greater than 0.1 //m. The surface area peaks around ca. 0.2 ,um.
9 An important feature of the mass or volume size distributions of atmospheric aerosols is their
10 multimodal nature. Volume-size distributions, measured in ambient air in the United States,
11 almost always are found to be bimodal, with a minimum between 1.0 and 3.0 //m (see
12 Figure 9-1). The distribution of particles that are mostly larger than the minimum is termed the
13 coarse mode, whereas the distribution of particles that are mostly smaller than the minimum is
14 termed the fine mode. In the ambient atmosphere, fine-mode particles include both the nuclei
15 mode and the accumulation mode. The nuclei mode, that portion of the fine-particle fraction
16 with diameters below about 0.1 /^m, can be observed as a separate mode in mass or volume
17 distributions only in clean or remote areas or near sources of new particle formation by
18 nucleation. Accumulation-mode particles are that portion of the fine-particle fraction with
19 diameters above about 0.1 /zm. Toxicologists use the term "ultrafme" to refer to particles in the
20 nuclei-mode size range. Aerosol physicists and material scientists tend to use "nanoparticles" to
21 refer to particles in this size range generated in the laboratory.
22 Another set of definitions of particle size fractions arises from considerations related to
23 size-selective sampling (see Figure 9-2). Size-selective sampling refers to the collection of
24 particles below or within a specified aerodynamic size range, usually defined by the upper 50%
25 cut point size, and has arisen in an effort to measure particle size fractions with some special
26 significance (e.g., health, visibility, source apportionment). Dichotomous samplers split the
27 particles into smaller and larger fractions, which may be collected on separate filters. However, a
28 fraction (-10%) of the fine particles are collected with the coarse particle fraction. Cascade
29 impactors use multiple size cuts to obtain a distribution of size cuts for mass or chemical
30 composition measurements.
March 2001 9-4 DRAFT-DO NOT QUOTE OR CITE
-------
o
6 -
5
4 -
O)
I 3
I
2 -
1 -
0.002
Mechanically
Generated
Nuclei Mode
0.1 1
Particle Diameter, Dp, |jm
Accumulation Mode
Coarse Mode
Fine-Mode Particles
Coarse-Mode Particles
Figure 9-1. Volume-size distribution, measured in traffic, showing fine- and coarse-mode
particles and the nuclei and accumulation modes within the fine-particle mode.
DGV (geometric mean diameter by volume, equivalent to volume median
diameter) and ag (geometric standard deviation) are shown for each mode.
Also shown are transformation and growth mechanisms (e.g., nucleation,
condensation, coagulation).
Source: Adapted from Wilson and Suh (1997).
1 Prior to 1987, the indicator for the National Ambient Air Quality Standards (NAAQS) for
2 PM was total suspended particulate matter (TSP). TSP is defined by the design of the High
3 Volume Sampler (hivol), which collects all of the fine particles but only part of the coarse
4 particles. The upper cut off size of the hivol depends on the wind speed and direction and may
5 vary from 25 to 40 ^m. In 1987, the NAAQS for PM were revised to use PM10, rather than TSP,
6 as the indicator for the PM NAAQS (Federal Register, 1987). The use of PM10 as an indicator is
7 an example of size-selective sampling. The selection of PM10 as an indicator was based on
March 2001
9-5
DRAFT-DO NOT QUOTE OR CITE
-------
n
70
60
50
J§ 40
o>
.o
30 -
CO
CO
< 20
10 -
Fine-Mode Particles
Coarse-Mode Particles
TSP
HiVol
WRAC
0.1
i
0.2
I
I
I
0.5 1.0 2 5 10
Aerodynamic Particle Diameter Da,
I
20
I
50 100
Total Suspended Particles (TSP)
PM
10
PM
'25
PM
(10-2.5)
Figure 9-2. An idealized distribution of ambient particulate matter showing fine- and
coarse-mode particles and the fractions collected by size-selective samplers.
WRAC is the Wide Range Aerosol Classifier, which collects the entire coarse
mode (Lundgren and Burton, 1995).
Source: Adapted from Wilson and Suh (1997).
1 dosimetric considerations and was intended to focus regulatory concern on those particles small
2 enough to enter the thoracic region of the human respiratory tract. The PM2 5 indicator
3 promulgated by U. S. Environmental Protection Agency (EPA) in 1997 is also an example of
4 size-selective sampling.
5 An idealized distribution showing the typically observed division of ambient aerosols into
6 fine- and coarse-mode particles and size fractions collected by the WRAC, TSP, PM10, PM2 5, and
March 2001
9-6
DRAFT-DO NOT QUOTE OR CITE
-------
1 PM(10.2 5) samplers, is shown in Figure 9-2. PM10 samplers, as defined in Appendix J to 40 Code
2 of Federal Regulations (CFR) Part 50 (Code of Federal Regulations, 199 la; Federal Register,
3 1987), collect all of the fine particles and part of the coarse particles. The upper cut point is
4 defined as having a 50% collection efficiency at 10 ± 0.5 jum AED. The slope of the collection
5 efficiency curve is defined in amendments to 40 CFR, Part 53, (Code of Federal Regulations,
6 1991b).
7 Over the years, the terms "fine" and "coarse", as applied to particle sizes, have lost the
8 original precise meaning of fine and coarse mode. In any given article, therefore, the meaning of
9 fine and coarse, unless defined, must be inferred from the author's usage. In particular, PM2 5
10 and fine-mode particles are not equivalent. In this chapter and document, the term "mode" is
11 used with fine and coarse when it is desired to specify the distribution of fine- or coarse-mode
12 particles as shown in Figures 9-1 and 9-2.
13
14 9.2.2 Fine- and Coarse-Mode Particle Distinctions vis-a-vis Sources,
15 Formation Mechanisms, and Atmospheric Behavior
16 Table 9-1 summarizes important physical and chemical properties, sources, and
17 atmospheric behavior that distinguish between nuclei-mode (ultrafine) and accumulation-mode
18 components of fine particles, as well as coarse-mode particles.
19 Several processes influence the formation and growth of particles. New particles may be
20 formed by nucleation from gas-phase material. Particles may grow by condensation as gas-phase
21 material condenses onto existing particles. Particles also may grow by coagulation as two
22 particles combine to form one. Gas-phase material condenses preferentially on smaller particles
23 and the rate constant for coagulation of two particles decreases as the particle size increases.
24 Therefore, nuclei-mode particles grow into the accumulation mode, but accumulation mode
25 particles do not grow into the coarse mode under normal atmospheric conditions.
26 As discussed in Chapter 2 of this document, the major constituents of atmospheric PM are
27 sulfate, nitrate, ammonium, and hydrogen ions; particle-bound water; elemental carbon; a great
28 variety of organic compounds; and crustal material. Atmospheric PM contains a large number of
29 elements in various compounds and concentrations and hundreds to thousands of specific organic
30 compounds. Particulate matter can be primary or secondary. Particulate matter is called primary
31 if it is in the same chemical form in which it was emitted into the atmosphere. Particulate matter
March 2001 9-7 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9-1. COMPARISON OF AMBIENT PARTICLES,
FINE (Nuclei Mode Plus Accumulation Mode) AND COARSE MODE
Fine
Nuclei
Accumulation
Coarse
Formed from:
Formed by:
Composed of:
Solubility:
Sources:
Atmospheric
half-life:
Removal
processes:
Combustion, high-temperature
processes, and atmospheric reactions
Nucleation
Condensation
Coagulation
Sulfates
Elemental carbon
Metal compounds
Organic compounds
with very low,
saturation vapor
pressure at ambient
temperature
Probably less
soluble than
accumulation mode
Combustion
Atmospheric
transformation of
SO2 and some
organic compounds
High temperature
processes
Condensation
Coagulation
Evaporation of fog and
cloud droplets in which
gases have dissolved and
reacted
Sulfate, SC-4
Nitrate, NO,
Ammonium, NHJ
Hydrogen ion, H+
Elemental carbon
Large variety of organic
compounds
Metals: compounds of Pb,
Cd, V, Ni, Cu, Zn, Mn, Fe,
etc.
Particle-bound water
Largely soluble,
hygroscopic, and
deliquescent
Combustion of coal, oil,
gasoline, diesel fuel, and
wood
Atmospheric transformation
products of NOX, SO2, and
organic compounds,
including biogenic organic
species (e.g., terpenes)
High-temperature
processes, smelters, steel
mills, etc.
Minutes to hours Days to weeks
Grows into
accumulation mode
Travel distance: <1 to 10s of km
Forms cloud droplets and
rains out
Dry deposition
100s to 1000s of km
Break-up of large solids/droplets
Mechanical disruption (crushing,
grinding, and abrasion of surfaces)
Evaporation of sprays
Suspension of dusts
Reactions of gases in or on particles
Suspended soil or street dust
Fly ash from uncontrolled combustion
of coal, oil, and wood
Nitrates and chlorides from HNO3 and
HC1
Oxides of crustal elements
(Si, Al, Ti, and Fe)
CaCO3, NaCl, and sea salt
Pollen, mold, and fungal spores
Plant and animal fragments
Tire, brake pad, and road wear debris
Largely insoluble and nonhygroscopic
Resuspension of industrial dust and
soil tracked onto roads and streets
Suspension from disturbed soil (e.g.,
farming, mining, unpaved roads)
Construction and demolition
Uncontrolled coal and oil combustion
Ocean spray
Biological sources
Minutes to hours
Dry deposition by fallout
Scavenging by falling rain drops
<1 to 10s of km
(100s to 1000s in dust storms)
Source: Adapted from Wilson and Suh (1997).
March 2001
9-8
DRAFT-DO NOT QUOTE OR CITE
-------
1 is called secondary if it is formed by chemical reactions in the atmosphere. Primary coarse
2 particles usually are formed by mechanical processes. Primary fine particles are emitted from
3 sources either directly as particles or as vapors that rapidly condense to form particles.
4 Most of the sulfate and nitrate and a portion of the organic compounds in atmospheric
5 particles are secondary (i.e., they are formed by chemical reactions in the atmosphere).
6 Secondary aerosol formation depends on numerous factors, including the concentrations of
7 precursors; the concentrations of other gaseous reactive species such as ozone (O3), hydroxyl
8 radical, peroxy radicals, or hydrogen peroxide; atmospheric conditions, including solar radiation
9 and relative humidity; and the interactions of precursors and preexisting particles within cloud or
10 fog droplets or on or in the liquid film on solid particles. As a result of such transformations, it is
11 considerably more difficult to relate ambient concentrations of secondary species to individual
12 sources of precursor emissions than it is to identify the sources of primary particles.
13 The atmospheric lifetimes of particles vary with the aerodynamic diameter of the particle.
14 Coarse particles can settle rapidly from the atmosphere within minutes or hours and normally
15 travel only short distances. However, when mixed high into the atmosphere, as in dust storms,
16 the smaller-sized, coarse-mode particles may have longer atmospheric residence times and travel
17 greater distances. Nuclei-mode particles rapidly grow into accumulation-mode fine particles,
18 which are kept suspended by normal air motions and have very low deposition rates to surfaces.
19 They can be transported thousands of kilometers and remain in the atmosphere for a number of
20 days. Particulate matter can be removed from the atmosphere by wet and dry deposition. Dry
21 deposition rates are expressed in terms of a deposition velocity, which varies with particle size,
22 reaching a minimum between 0.1 and 1.0 yum AED. For small particles, dry deposition is
23 accomplished by impaction on surfaces by turbulent motion. For larger particles (i.e., coarse
24 mode), buoyancy forces are not large enough to overcome the force of gravity, and gravitational
25 settling becomes important. Soluble particles are removed from the atmosphere primarily by
26 incorporation into cloud droplets, which then rain out. Coarse-mode and ultrafine, but not
27 accumulation-mode, particles are removed by impaction with falling rain drops.
28
March 2001 9-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 9.2.3 Particle Size-Related Distinctions vis-a-vis Number, Surface Area,
2 and Mass
3 The distribution of particles in terms of numbers, surface area, and mass in relation to size
4 is gaming more attention as efforts focus on trying to identify specific toxic components of the
5 ambient PM mix that may contribute to observed human health and environmental effects.
6 Examples of averaged atmospheric size distributions are shown in Figures 9-3 and 9-4.
7 Figure 9-3 describes the number of particles as a function of particle diameter for rural,
8 urban-influenced rural, urban, and freeway-influenced urban aerosols. For the urban data, the
9 particle volume distribution is shown in Figure 9-4. The particle diameter is always shown on a
10 logarithmic scale. The particle number frequently is shown on a logarithmic scale to display the
11 wide range in number concentration for different particle sizes and different sites. When shown
12 on an arithmetic scale the volume, surface area, or number of particles in any specified size range
13 is proportional to the corresponding area under the curve (see Figure 9-5). These distributions
14 show that most of the particles are very small (<0.1 /um), whereas most of the particle volume
15 (and therefore most of the mass) and the surface area is found in particles >0.1 /um. Also,
16 particle surface area peaks around 0.2 /um.
17 The number concentrations of coarse particles are usually too small to be seen in arithmetic
18 plots (Figure 9-3b) but can be seen in a logarithmic plot (Figure 9-3a). Whitby and Sverdrup
19 (1980) observed that rural aerosols, not much influenced by nearby sources, have a small
20 accumulation mode and no observable nuclei mode. For urban aerosols, the accumulation- and
21 coarse-particle modes are comparable in volume. However, in urban aerosols, the nuclei-mode
22 can be observed only in volume distributions that are influenced by nearby traffic or other
23 sources of nuclei-mode particles. Still, the nuclei-mode dominates the number distributions of
24 urban aerosols. Whitby's conclusions were based on extensive studies of size distributions in a
25 number of western and midwestern locations during the 1970s (Whitby, 1978; Whitby and
26 Sverdrup, 1980). No size-distribution studies of similar scope have been published since then.
27 Newer data from particle counting techniques and size-segregation impactor studies, including
28 data from Europe (U.S. Environmental Protection Agency, 1996) and Australia (Keywood et al.,
29 1999), show similar results.
30
March 2001 9-10 DRAFT-DO NOT QUOTE OR CITE
-------
1,000,000-
10,000 -
-------
70
65 -
60 -
55
50 -
» 45 -
^40 -
"S
930 -
O)
% 25 H
i 20
15 -
10 -
5 -
0
Clean Rural
Urban Influenced
Rural
South-Central
New Mexico
70
65 -
60 -
55 -
50 -
„ 45 -
J<0 -
^35 H
"S
9 30 -
2 25 -
£ 20 -
15 -
10 -
5 -
0
Average Urban
Urban + Freeway
0.01 0.1 1 10
Particle Diameter, Dp (u.m)
100 0.01 0.1 1 10
Particle Diameter, Dp (urn)
100
Figure 9-4. Particle volume distribution as a function of particle diameter: (a) for the
averaged rural and urban-influenced rural number distributions shown in
Figure 9-3 and a distribution from south central New Mexico, and (b) for the
averaged urban and freeway-influenced urban number distributions shown in
Figure 9-3.
Source: Whitby and Sverdrup (1980) and Kim et al. (1993).
1
2
3
4
5
6
7
(1) Particles containing heavy metals. Nuclei-mode particles of metal oxides or other
metal compounds are generated during metal smelting processes or, more widely, when
metallic impurities in coal or oil are vaporized during combustion and the vapor undergoes
nucleation. Metallic ultrafine particles also may be formed from metals in lubricating oil or
fuel additives that are vaporized during combustion of gasoline or diesel fuels.
(2) Elemental carbon (EC) or soot. Elemental carbon particles are formed primarily by
condensation of C2 molecules generated during combustion processes. Because EC has a
very low equilibrium vapor pressure, ultrafine EC particles can nucleate even at high
March 2001
9-12
DRAFT-DO NOT QUOTE OR CITE
-------
CD
.Q
E
CO
0)
0)
o
1
co
x
IE
o
D)
O
o
0.
Q
CD
O
CO
15-
10 -
5-
600-
400-
200-
30 -1
E ^ 20 H
5 Q
D)
JO
10H
Nn = 7.7x 10
DGNn = 0.013
(a)
Na = 1.3x10
DGNa = 0.069
= 2.03
Nc = 4.2
DGNC = 0.97
agc=2.15
Vn = 0.33
DGVn = 0.031
0.001 0.01
0.1
1.0
10
100
Figure 9-5. Distribution of coarse (c), accumulation (a), and nuclei- or ultrafine (n) -mode
particles by three characteristics, (1) number (N), (2) surface area (S), and
(3) volume (V) for the grand average continental size distribution. DGV =
geometric mean diameter by volume; DCS = geometric mean diameter by
surface area; DGN = geometric mean diameter by number; Dp = geometric
diameter.
Source: Whitby (1978).
March 2001
9-13
DRAFT-DO NOT QUOTE OR CITE
-------
1 temperatures (Kittelson, 1998; Morawska et al., 1998). Thus, substantial amounts of EC
2 can be released into the air as the result of biomass burning (because of agricultural
3 clearing, forest fires, etc.) or combustion of fossil fuels (e.g., gasoline or diesel fuel derived
4 from oil or coal used for power generation, industrial boilers, etc.).
5
6 (3) Sulfates and nitrates. Sulfuric acid (H2SO4), or its neutralization products with
7 ammonia (NH3) (i.e., ammonium sulfate [(NH4)2SO4] or ammonium acid sulfate
8 [NH4HSO4]), are generated in the atmosphere by conversion of sulfur dioxide (SO2) to
9 H2SO4. As H2SO4 is formed, it can either nucleate to form new ultrafine particles, or it can
10 condense onto existing nuclei-mode or accumulation-mode particles (Clark and Whitby,
11 1975; Whitby, 1978). However, the possible formation of ultrafine ammonium nitrate
12 (NH4NO3) by reaction of NH3 and nitric acid (HNO3) apparently has not been investigated.
13
14 (4) Organic carbon (OC). Recent smog chamber studies and indoor experiments show
15 that atmospheric oxidation of certain organic compounds found in the atmosphere can
16 produce highly oxidized organic compounds with an equilibrium vapor pressure low
17 enough to result in nucleation (Kamens et al., 1999; Weschler and Shields, 1999). Organic
18 carbon compounds originate from a wide variety of processes, including biomass burning,
19 fossil fuel combustion, use of various dry cleaning or industrial solvents, and release of
20 naturally occurring substances (e.g., terpenes) from certain terrestrial plant species.
21
22 Ambient concentrations of nuclei-mode particles importantly reflect a balance between
23 formation and removal processes. Nuclei-mode particles are removed mainly by growth into the
24 accumulation mode but also may be removed by dry deposition. Such growth takes place as
25 other low-vapor-pressure material condenses onto the particles or as nuclei-mode particles
26 coagulate with themselves or with accumulation-mode particles. Because the rate of coagulation
27 will vary with the concentration of accumulation-mode particles, it might be expected that
28 atmospheric concentrations of nuclei-mode particles would increase with decreases in
29 accumulation-mode mass. On the other hand, the concentration of particles would be expected to
30 decrease with a decrease in the rate of generation of particles by reduction in emissions of metal
31 and carbon particles or a decrease in the rate of generation of H2SO4 or condensable organic
March 2001 9-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 vapor. The rate of generation of H2SO4 depends on the concentration of SO2 and OH radicals.
2 OH is generated primarily by the photolysis of ozone at wavelengths <320 nm, followed by the
3 reaction of electronically excited oxygen atoms with water vapor.
4 Exposure to ultrafine particles may occur near sources of primary ultrafme particles (e.g., in
5 traffic). Secondary ultrafme particles are generated by photochemistry throughout the boundary
6 layer; so exposure to ultrafine particles is not limited to locations near primary sources. Models
7 exist to predict formation and coagulation rates, but no careful analyses of how rapidly various
8 ultrafine (nuclei-mode) particles may agglomerate or adhere to larger particles as they "age" in
9 the ambient air or how this may impact lung deposition of such particles has been published.
10 Thus, it may be important to monitor particle number and surface area, as well as mass, to further
11 delineate PM exposure-response relationships and to determine the relative effectiveness of
12 strategies for reducing particle mass, surface area, and number to ameliorate PM-related health
13 risks.
14
15
16 9.3 CHARACTERIZATION OF U.S. AMBIENT PARTICULATE MATTER
17 CONCENTRATIONS AND CONTRIBUTING SOURCES AND
18 EMISSIONS
19 9.3.1 Ambient Particulate Matter Measurement Methods
20 The EPA decision to revise the PM standards by adding daily and annual PM2 5 NAAQS
21 has led to renewed interest in the measurement of atmospheric particles and better understanding
22 of problems in obtaining precise and accurate airborne particle measurements.
23 The U.S. Federal Reference Methods (FRJM) for PM2 5 and PM10 provide relatively precise
24 (±10%) methods for determining the mass of material remaining on a Teflon filter after
25 equilibration at 25 °C and 40% relative humidity. However, many uncertainties exist as to
26 relationships between the mass and composition of material remaining on the filter, as measured
27 by the FRM, and the mass and composition of material that exists in the atmosphere as
28 suspended PM. It is currently not possible to characterize accurately the material that exists as a
29 particle in the atmosphere, in part because of difficulties in creating a reference standard for
30 particles suspended in the atmosphere. As a result, EPA defines accuracy for PM measurements
31 in terms of agreement of a candidate sampler with a reference sampler under standardized
March 2001 9-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 conditions for sample collection, storage, and analysis. Therefore, intercomparisons of samplers
2 become very important in determining how well various samplers agree and how various design
3 choices influence what is actually measured. Data from ambient PM monitoring is needed to
4 guide implementation of a standard; to determine whether or not a standard has been attained;
5 and to determine effects on health, ecosystems, visibility, and the transfer of solar ultraviolet and
6 visible radiation.
7 Current filtration-based mass measurements lead to significant evaporative losses, during
8 and possibly after collection, of a variety of semivolatile components (i.e., species that exist in
9 the atmosphere in dynamic equilibrium between the condensed phase and gas phase). Important
10 examples include ammonium nitrate, semivolatile organic compounds, and particle-bound water.
11 In designing an aerosol indicator, choices must be made regarding the treatment of the
12 semivolatile components. Other areas where choices must be made include selection of an upper
13 cut point; separation of fine- and coarse-mode PM; and treatment of pressure, temperature, and
14 relative humidity.
15 It is becoming increasingly apparent that the semivolatile component of PM impacts
16 significantly the quality of the measurement, and leads to both positive and negative sampling
17 artifacts. Negative artifacts, because of the loss of ammonium nitrate and semivolatile organic
18 compounds, occur during sampling, because of temperature, relative humidity, composition of
19 the aerosol, or because of pressure drop across the filter. Negative artifacts also occur during
20 handling and storage because of evaporation. Positive artifacts occur when volatile species
21 adsorb onto, or react with, filter media or collected PM.
22 The loss of particulate nitrate may be determined by comparing nitrate collected on a
23 Teflon filter to that collected on a nylon filter (which absorbs nitrate), preceded by a denuder to
24 remove nitric acid. In two studies, the PM25 mass lost because of volatilization of ammonium
25 nitrate was found to represent 10 to 20% of the total PM2 5 mass and almost a third of the nitrate.
26 Semivolatile organic compounds (SVOC) can similarly be lost from Teflon filters because of
27 volatilization during or after collection. Such losses can cause the PM2 5 mass to be
28 underestimated significantly. The FRM for PM2 5 will suffer loss of particulate nitrates and
29 SVOC, similar to the losses experienced with other single-filter collection systems.
30 It is generally desirable to collect and measure ammonium nitrate and SVOC. However, it
31 is also desirable to remove the particle-bound water before determining the mass. Calculations
March 2001 9-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 and measurements indicate that aerosol water content is strongly dependent on composition, but
2 that liquid water could represent a significant mass fraction of aerosol concentration at relative
3 humidities above 60%.
4 Federal Reference Methods for equilibrated mass have been specified for PM10 and PM2 5.
5 In addition to FRM sampling to determine compliance with PM standards, EPA requires states to
6 conduct speciation sampling to determine contributions from different source categories and to
7 evaluate exposure to trace elements. The current speciation samplers include three filters:
8 (1) a Teflon filter for equilibrated mass and elemental analysis; (2) a nylon filter, preceded by a
9 nitric acid denuder, to collect nitrate; and (3) a quartz fiber filter for elemental and organic
10 carbon (but without any correction for positive or negative artifacts caused by adsorption of
11 volatile organic compounds on the quartz filters or evaporation of semivolatile organic
12 compounds from the collected particles).
13 The EPA expects that more than 200 local agency monitoring sites throughout the United
14 States will operate continuous PM monitors. However, EPA has not yet provided any guidance
15 regarding appropriate continuous monitoring techniques. All currently available continuous
16 measurements of suspended particle mass share the problem of dealing with semivolatile PM
17 components (i.e., so as not to include particle-bound water as part of the mass, the particle-bound
18 water must be removed by heating or dehumidification). However, heating also causes loss of
19 ammonium nitrate and semivolatile organic components. Several candidates for continuous PM
20 mass measurements, which use dehumidification instead of heating to remove particle-bound
21 water, currently are being field tested. In addition to continuous mass measurement, a number of
22 techniques for continuous measurement of sulfate, nitrate, or elements are being tested. Aerosol
23 time-of-flight mass spectroscopy provides a new technique for real-time measurement of
24 correlated size and composition profiles of individual atmospheric aerosol particles.
25 For measurement of the chemical composition of PM collected on a filter, adequate
26 techniques exist for measurement of the heavier elements (Na and higher); sulfate, nitrate,
27 ammonium, and hydrogen ions; and total carbon. The split between elemental carbon and
28 organic carbon is defined operationally and depends on the measurement technique used. The
29 definition of elemental carbon (measured by oxidation to CO2 and quantification of the CO2
30 formed) and black carbon (measured by optical absorption or transmission) is also operational
31 and determined by the methods used. Determination of the mass of organic material (carbon plus
March 2001 9-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 molecularly bound hydrogen and oxygen) remains a problem, as does the identification of the
2 many individual organic compounds. However, measurement techniques for polynuclear
3 aromatic hydrocarbons (PAH) and some other toxic compounds in PM are well developed.
4
5 9.3.2 Patterns and Trends in U.S. Particulate Matter Concentrations
6 Since the 1987 setting of the PM,0 NAAQS, extensive PM10 monitoring has been carried
7 out throughout the United States, allowing for confident characterization of PMIO patterns and
8 trends during the past decade or so. However, only very recently, with the deployment of a
9 nationwide PM2 5 monitoring network during 1998, has it became possible to begin, in a
10 systematic fashion, to characterize PM25 patterns and trends, starting with data for 1999.
11
12 9.3.2.1 PM10 Trends and Concentrations
13 Annual average PM10 mass concentrations throughout the United States, for different
14 regions within the United States, and for most subregions or cities, have generally decreased over
15 the past decade. Nationwide average PM10 concentrations decreased from 31.7 /ug/m3 in 1989 to
16 23.7 Aig/m3 in 1998. Decreases were largest (38%) in the Pacific Northwest and smallest in the
17 Southeast (18%).
18 Annual mean PMIO concentrations in urban areas, found in EPA's Air Information
19 Retrieval System database (Fitz-Simons et al., 2000), generally were greater than about 20 /ug/m3
20 for 1999. Annual average concentrations above 50 //g/m3 are found in several locations in
21 southern and central California, Nevada, Arizona, Texas, South Carolina, and Puerto Rico.
22 At rural sites in national parks, wilderness areas, and national monuments in the western United
23 States, the annual average PM10 concentrations were in the range of 5 to 10 /ug/m3. Higher PM10
24 concentrations have been reported at some rural sites in the eastern United States. The
25 corresponding PM2 5 concentrations in western rural or remote sites were approximately 3 Aig/m3
26 and, in eastern rural or remote sites, were in the range of 5 to 10 yUg/m3.
27 A few attempts to infer various types of "background" levels of PM25 and PM10 have been
28 made. The background levels most relevant to the present criteria document include: (l)an
29 uncontrollable"background" (which includes the "natural background" defined below and
30 anthropogenic sources outside of North America), and (2) a "natural background" (which
31 includes all natural sources but excludes all anthropogenic sources anywhere in the world).
March 2001 9-18 DRAFT-DO NOT QUOTE OR CITE
-------
1 Annual average background levels of PM10 (according to the first definition) have been estimated
2 to range from 4 to 8 /ug/m3 in the western United States and 5 to 11 Aig/m3 in the eastern United
3 States. Corresponding PM2 5 background levels have been estimated to range from 1 to 4 ,ug/m3
4 in the western United States and from 2 to 5 /wg/m3 in the eastern United States.
5
6 9.3.2.2 PM2 5 Trends and Concentrations
7 The recently deployed PM25 FRM network has returned data for a large number of sites
8 across the United States beginning in January of 1999. As of the end of 1999, the network
9 consisted of 1025 monitors. Annual mean PM2 5 concentrations for 1999 ranged from about
10 5 yUg/m3 to more than 20 //g/m3. As might be expected, annual average PM2 5 concentrations
11 towards the low end of the range were found in relatively small, nonindustrialized cities such as
12 Bangor, ME; Fargo, ND; Cheyenne, WY; and Albuquerque, NM. Higher annual averages were
13 found in larger urban areas such as Atlanta, GA, and Los Angeles, CA, as well as in a number of
14 urban areas in the eastern United States. Because FRM measurements of PM25 only began in
15 January 1999, data tend to be limited in many areas, especially for the first quarter. However, a
16 number of observations can be made regarding PM2 s concentrations and the patterns of seasonal
17 variability in urban areas across the United States. Generally, similar patterns of seasonal
18 variability were found at all sites within Metropolitan Statistical Areas (MSAs) sampled
19 nationwide, although there were exceptions at individual sites, which may have been related to
20 contributions from local sources as opposed to contributions from regional background sources.
21 At sites in the eastern United States, highest quarterly mean values and maximum values
22 occurred during the third quarter (summer) of 1999, with exceptions occurring at several
23 locations. For example, at monitoring sites in Miami and Puerto Rico, maximum concentrations
24 occurred during the second quarter and may have been related to the transport of dust from the
25 Sahara Desert. At sites west of the Mississippi River, highest mean values occurred during the
26 first or fourth quarter (winter or autumn) of 1999, and, again, there were exceptions. Because of
27 the limited nature of these data, definitive conclusions regarding long-term patterns of seasonal
28 variability cannot be drawn from these data alone. These findings are generally consistent with
29 those based on longer term data sets such as the Metropolitan Acid Aerosol Characterization
30 Study (MAACS) in the eastern United States and the California Air Resources Board (CARS)
31 network of dichotomous samplers in California. Very limited data sets are available for obtaining
March 2001 9-19 DRAFT-DO NOT QUOTE OR CITE
-------
1 trends in PM2 5 concentrations in urban areas. Data obtained by the CARB indicate that annual
2 average PM2 5 concentrations decreased from 35 to 50% in large urban areas in California from
3 1990 to 1995. Smaller decreases ranging from 2 to 34% were observed as part of the children's
4 health study in southern California. In contrast, the urban IMPROVE site in Washington, DC
5 measured only a 5% decline in PM2 5 concentrations from 1989-1997.
6
7 9.3.2.3 Spatial Variability in PM2.5 Concentrations
8 The 1999 FRM PM2 5 data indicate that, in general, PM2 5 concentrations are highly
9 correlated among sites within several MSAs (Atlanta, GA; Detroit, MI; Phoenix-Mesa, AZ; and
10 Seattle-Bellevue-Everett, WA), although there are some exceptions to this rule. These findings
11 are consistent with those of earlier studies in Philadelphia, PA; and Los Angeles, CA.
12 Concentrations of PM25 also tended to be highly correlated on much larger spatial scales in many
13 areas in the United States, supporting the inference that PM2 5 is a regionally distributed pollutant.
14
15 9.3.2.4 Relationships Among Particulate Matter in Different Size Fractions
16 PM25 to PM10 ratios from the FRM network were generally higher in the eastern (=0.7)
17 than in the central or western (-0.5) United States during 1999. These values are consistent with
18 those found in numerous earlier studies presented in the 1996 PM AQCD.
19 The results of ambient monitoring studies and receptor modeling studies in the eastern
20 United States indicate that PM2 5 is dominated by secondary components. Depending on the
21 origin of OC in ambient samples, PM2 5, on average, also may be dominated by secondary
22 components throughout the rest of the United States. Primary constituents represent smaller but
23 still important components of PM25, on average. Crustal materials constitute the largest fraction
24 of PM(10.2 5) throughout the United States. Crustal materials in the lower tail of the coarse-mode
25 particles also may be present in the PM2 5-size fraction. Data collected in several airsheds,
26 including the Los Angeles Basin, Bakersfield and Fresno, CA; and Philadelphia, PA, suggest that
27 secondary PM components are more uniformly distributed than are primary components.
28 Compositional data obtained at multiple sites in other urban areas are sparse.
29
March 2001 9-20 DRAFT-DO NOT QUOTE OR CITE
-------
1 9.3.2.5 Short-Term Temporal Variability of Particulate Matter Concentrations
2 Hour-to-hour changes in PM2 5 concentrations have been obtained at 31 sites by various
3 continuous monitors. The 1999 nationwide composite circadian variability in PM2 5
4 concentrations obtained by these monitors indicate two typical intra-day peaks. The first peak
5 occurs from about 6 to 9 a.m. and the second peak occurs from about 5 to 10 p.m. The amplitude
6 of these peaks is much smaller than the daily mean concentration. It also should be noted that
7 this pattern may not be apparent in the data obtained by any given monitor on any given day.
8 Although the 98th percentile values for positive and negative excursions in 24-h PM2 5
9 concentrations are typically less than 20 /wg/m3, maximum hour-to-hour excursions may be over
10 200 yag/m3 in some locations.
11 The only data sets from which the long-term, day-to-day variability in PM2 5 and PM10
12 concentrations could be assessed, based on daily filter measurements, were obtained in
13 Philadelphia, PA, from 1992 to 1995 and in Phoenix, AZ, from 1995 through 1997. In the
14 Philadelphia data set, average day-to-day concentration differences obtained were
15 6.8 ± 6.5 Mg/m3 for PM2 5 and 8.6 ± 7.5 Mg/m3 for PM10, whereas maximum day-to-day
16 differences obtained were 54.7 /ig/m3 for PM2 5 and 50.4 //g/m3 for PM10. In the Phoenix, AZ,
17 data set, average day-to-day PM2 5 concentration differences were 2.9 ± 3.0 /wg/m3, and the
18 maximum day-to-day concentration difference was 23 /wg/m3.
19
20 9.3.3 Sources of Particulate Matter
21 As shown in Table 9-1, fine and coarse particles have different types of sources. The major
22 sources of fine and coarse PM are summarized in Table 9-2. Because of the complexity of the
23 composition of ambient PM2 5 and PM(IO_2 5), sources are best discussed in terms of individual
24 constituents of both primary and secondary PM2 5 and PM(IO_2 5). Each of these constituents can
25 have anthropogenic and natural sources, as shown in Table 9-2. The distinction between natural
26 and anthropogenic sources is not always obvious. For example, although windblown dust might
27 seem to be the result of natural processes, highest emission rates are associated with agricultural
28 activities in areas that are susceptible to periodic drought, such as in the dust bowl region of the
29 mid-western United States. Also, most forest fires in the United States could be classified as
30 human in origin, either through prescribed burning, by accident, or through forest management
31 practices, which allow the buildup of combustible material, thereby increasing the likelihood of
March 2001 9-21 DRAFT-DO NOT QUOTE OR CITE
-------
s
p
3
cr
to
o
o
TABLE 9-2. CONSTITUENTS OF ATMOSPHERIC PARTICLES AND THEIR MAJOR SOURCES
o
z
o
H
O
C
O
H
tfl
O
?o
n
HH
H
W
Sources
Pnmary (PM < 2. 5 ^m) Primary (PM > 2.5 ^m) Secondary PM Precursors (PM < 2.5 /urn)
Aerosol
species Natural Anthropogenic Natural
SO4" Sea spray Fossil fuel combustion Sea spray
Sulfate
Anthropogenic Natural
— a Oxidation of reduced sulfur
gases emitted by the oceans
and wetlands and SO2 and
H2S emitted by volcanism
and forest fires
Anthropogenic
Oxidation of SO, emitted
from fossil fuel
combustion1'
Nitrate
Minerals
Oxidation of NOX produced
by soils, forest fires, and
lighting
Oxidation of NO, emitted
from fossil fuel combustion
and in motor vehicle
exhaust
Erosion and
reentramment
Fugitive dust, paved
and unpaved roads, and
agriculture and forestry
Erosion and
reentrainment
Fugitive dust, paved
and unpaved road
dust, and agriculture
and forestry
NH4+
Ammonium
Organic
Carbon (OC)
Elemental
Carbon
(EC)
Metals
Bioaerosols
—
Wildfires
Wildfires
Volcanic
activity
Viruses and
bacteria
—
Prescribed burning,
wood burning, motor
vehicle exhaust, and
cooking
Motor vehicle
exhaust, wood
burning, and cooking
Fossil fuel
combustion, smelting,
and brake wear
—
— —
— Tire and asphalt wear
and paved road dust
— —
Erosion, reentrainment, —
and organic debris
Plant, insect fragments, —
pollen, fungal spores,
and bacterial
agglomerates
Emissions of NH3 from
wild animals and
undisturbed soil
Oxidation of hydrocarbons
emitted by vegetation,
(terpenes, waxes) and
wildfires
—
—
—
Emissions of NH, from
animal husbandry, sewage,
and fertilized land
Oxidation of hydrocarbons
emitted by motor vehicles,
prescribed burning, and
wood burning
—
—
—
"Dash (—) indicates either very minor source or no known source of component.
""Major source of each component shown in boldface type.
-------
1 fire from whatever cause. As seen in Table 9-2, emissions of crustal material (mineral dust),
2 organic debris, and sea spray are concentrated mainly in the coarse fraction of PM10 (>2.5 /u.m
3 AED). A small fraction of this material is in the PM2 5 size range (<2.5 /urn AED). Nevertheless,
4 the concentrations of crustal material can be appreciable, especially during dust events.
5 Emissions from combustion sources (mobile and stationary sources, biomass burning, etc.) are
6 predominantly in the PM2 5 size range.
7 The results of receptor modeling studies throughout the United States indicate that the
8 combustion of fossil and biomass fuels is a major source of PM25. Fugitive dust, found mainly in
9 the PM(10_2 5) range size, represents the largest source of PM10 in many locations in the western
10 United States. Quoted uncertainties in source apportionments of constituents in ambient aerosol
11 samples typically range from 10 to 50%. It is apparent that a relatively small number of source
12 categories, compared to the total number of chemical species that are typically measured in
13 ambient monitoring-source receptor model studies, are needed to account for most of the
14 observed mass of PM in these studies.
15 Although most emphasis in this discussion has been on sources within the United States,
16 it should be remembered that sources outside the United States also contribute to ambient PM
17 levels in the United States that can, at times, exceed the ambient NAAQS level for PM. Perry
18 et al. (1997) have found that the highest concentrations of mineral dust in the PM2 5 fraction are
19 found in the eastern United States during the summer and not in arid areas of the western
20 United States. This dust originates from the Sahara Desert and is then transported across the
21 Atlantic Ocean. Much of the Saharan dust that reaches the United States is in the PM2 5 size
22 range. Large-scale dust storms in the deserts of central Asia also have contributed to PM levels
23 in the Northwest on an episodic basis. In addition, uncontrolled biomass burning in Central
24 America and Mexico occasionally contributes to elevated U.S. ambient PM levels, having led at
25 times to brief exceedances of daily PM NAAQS level in Texas. Wildfires throughout the United
26 States, Canada, Mexico, and Central America all contribute to background concentrations of PM
27 in the United States.
28
29
March 2001 9-23 DRAFT-DO NOT QUOTE OR CITE
-------
1 9.4 HUMAN EXPOSURES TO AMBIENT PARTICULATE MATTER
2 The concentration of PM in the air inhaled by a person is not necessarily the same as that
3 measured at a community ambient-air monitoring station. Personal exposure is defined as the
4 concentration, integrated over a time period, of PM near the breathing zone but not influenced by
5 exhaled breath. Total personal exposure, including ambient and nonambient PM, may be
6 measured by a personal exposure monitor (PEM) carried by the person. There are several
7 reasons why an individual's personal exposure may be different from the ambient concentration.
8 First, the concentration of PM outside a person's home may be different from the concentration
9 measured at a monitoring station. However, for cities where sufficient information is available
10 (e.g., see Section 9.3.2), PM concentrations measured at different pairs of stations (sited to
11 measure community-wide pollution levels rather than individual source contributions) have been
12 found to be highly correlated for PM2 5 and PM10, but not so highly correlated for PM,0_2 5.
13 Although it would be desirable to check the spatial variability of PM indicators in each city
14 where epidemiologic studies are conducted, it seems likely that PM2 5 and PM,0 concentrations
15 are distributed evenly enough so that one site, or the average of several sites, provides an
16 adequate measure of the community average concentration for PM25 and PM10. This may not be
17 the case for PM10.2 5, for specific chemical components, for source contributions, or for sites
18 located near sources.
19 Second, the concentration of ambient PM found indoors is generally less than the
20 concentration of ambient PM outdoors. Ambient air, and the ambient PM it contains, penetrates
21 indoors through open doors and windows and through small openings in the building structure.
22 An equal volume of indoor air moves out of the indoor microenvironment. Unless the air
23 exchange rate is very high, the ambient PM that penetrates indoors will be removed by deposition
24 more rapidly than it can be replaced. The ratio of ambient PM indoors to ambient PM outdoors,
25 called the infiltration factor, depends on the air exchange rate and, also, on the penetration
26 efficiency and deposition or removal rate, both of which vary with particle aerodynamic size.
27 The infiltration factor is a maximum for particles within the accumulation mode (=0.3 to 0.7 /um
28 AED) and decreases for smaller (ultrafine) or larger (coarse-mode) particles. For a given size
29 particle, the relationship between the indoor and outdoor PM concentration, given by the
30 infiltration factor, will vary with the air exchange rate. For a home closed for heating or air-
31 conditioning, the air exchange rate depends on the temperature difference between the indoor and
March 2001 9-24 DRAFT-DO NOT QUOTE OR CITE
-------
1 outdoor air; the greater the difference, the greater the air exchange rate. If windows are opened
2 for ventilation or doors are opened frequently, the air exchange rate will be higher.
3 The relationship between ambient concentration of PM and personal exposure to ambient
4 PM also is modulated by the time spent outdoors, because, while outdoors, a person is exposed to
5 the ambient concentration. The ratio of personal exposure to ambient PM to the ambient PM
6 concentration is called the attenuation factor and is given the symbol a. Both the infiltration
7 factor and the attenuation factor, a, may be estimated by several techniques. The most direct
8 method is to measure the personal exposure and ambient concentration of a chemical species that
9 has no indoor sources and is in the same size range as the PM component of interest. Candidates
10 are sulfate and, in homes with no open combustion, elemental carbon. The ratio of the personal
11 exposure to the tracer to the ambient concentration of the tracer gives a. In turn, a times the
12 ambient concentration of the appropriate PM indicator gives the personal exposure to that
13 component of ambient PM.
14 Personal exposure also contains a component resulting from indoor sources of PM, which
15 tend to produce ultrafme and coarse-mode particles rather than accumulation-mode particles.
16 Important indoor sources are tobacco smoke and other open combustion (ultrafme); cleaning,
17 sweeping, dusting, vacuuming (coarse); oven cooking (ultrafme); and resuspension caused by
18 walking on rugs (coarse). Stove-top cooking produces both ultrafme and coarse-mode particles.
19 Vacuum cleaners may produce ultrafme carbon or copper particles from motor brushes.
20 Another recently identified indoor source involves PM generated by the reaction of ozone (which
21 infiltrates with ambient air) with terpenes from air fresheners or cleaning agents.
22 Indoor-generated ultrafine particles will grow into the accumulation mode unless they are
23 removed first by deposition or air exchange. Indoor-generated PM would be expected to have a
24 lower proportion of transition and toxic metals and highly oxidized and nitrated organic
25 compounds than ambient air.
26 Community time-series epidemiology studies evaluate the daily totals of deaths (or other
27 health outcomes) in the community in relation to concentrations of one or more air pollutants
28 measured at stationary community ambient air monitoring sites (assumed to be representative of
29 the community average). There has been some controversy over whether the ambient
30 concentration should be considered to be a surrogate for total human personal exposure or only
31 for exposure to the ambient-generated component of total personal exposure. Some exposure
March 2001 9-25 DRAFT-DO NOT QUOTE OR CITE
-------
1 analysts feel that ambient concentrations represent a surrogate for total personal exposure (the
2 sum of exposure to ambient-generated pollution plus exposure to nonambient exposure). This
3 view is difficult to reconcile with epidemiologic studies that find statistically significant
4 relationships between ambient concentrations and health outcomes, even though correlations of
5 ambient concentrations with total personal exposures are found to be near zero. On the other
6 hand, certain other scientists have argued that the ambient concentration represents a surrogate
7 only for exposure to the ambient-generated component of total PM exposure.
8 Recent studies of exposure error suggest that, provided ambient-generated and nonambient
9 PM have equal toxicity, the increase in health outcomes per unit increase in concentration,
10 compared to PE, the increase in health outcome per unit increase in exposure, will not be biased
11 by the nonambient component of exposure if the nonambient component is independent of the
12 ambient concentration. Both logic and experiment suggest that nonambient PM exposures are
13 independent of daily ambient concentrations. However, the increase in health outcomes per unit
14 increase in ambient concentration will be biased low compared to the increase in health outcomes
15 per unit increase in exposure. For a constant average ratio of exposure to ambient-generated PM
16 to PM concentration (the attenuation factor, a, discussed earlier), the bias will be given by this
17 ratio which might be expected to vary from a few tenths (0.1 s) to nearly 1.0 depending on air
18 exchange and indoor removal rates. Thus, it seems reasonable to conclude that community
19 time-series epidemiology studies provide information on the statistical association of exposure to
20 ambient-generated pollutants with health outcomes, but do not provide any information on the
21 relationship of nonambient exposure with health outcomes. It is likely that the nonambient
22 component of total personal exposure also has health effects. However, techniques other than
23 community time-series epidemiology must be used to identify relationships between nonambient
24 exposure and health outcomes.
25
26
27 9.5 DOSIMETRY CONSIDERATIONS
28 A basic health effects assessment principle is that dose delivered to the target site, rather
29 than external exposure, is the proximal cause of biological responses. Characterization of an
30 exposure-dose-response continuum for PM (key objective of any dose-response assessment for
31 evaluation of health effects) requires elucidation of mechanistic determinants of inhaled particle
March 2001 9-26 DRAFT-DO NOT QUOTE OR CITE
-------
1 dose, which depend initially on deposition of particles in the respiratory tract. Once deposited on
2 respiratory tract surfaces, particles undergo absorptive or nonabsorptive removal (clearance)
3 processes that may result in their removal from airway surfaces and translocation from the
4 respiratory tract. Clearance depends on initial site of deposition and physicochemical properties
5 of the particles; both impact translocation mechanisms. Retained particle burdens are determined
6 by dynamic relationships between deposition and clearance mechanisms. The dose from inhaled
7 particles deposited and retained in the respiratory tract is governed by many factors (e.g.,
8 exposure concentration and duration, respiratory tract anatomy and ventilatory parameters, and
9 physicochemical properties of the particles (e.g., particle size, hygroscopicity, solubility).
10 Particles exist in the atmosphere as aerosols (i.e., airborne suspensions of finely dispersed
11 solid or liquid particles). As noted in Chapter 2 and Section 9.2.1, the most commonly used
12 metric AED, whereby particles of differing geometric size, shape, and density are compared
13 aerodynamically with the instability behavior (i.e., terminal setting velocity) of particles that are
14 unit density (1 gm/cm3) spheres. Importantly, aerosols present in natural and work environments
15 have polydisperse size distributions (i.e., particles within an aerosol have a range of sizes most
16 appropriately described by a size distribution). Aerosol size distributions are frequently modeled
17 by a sum of lognormal distributions, one for each mode (nuclei, accumulation, and coarse). Two
18 parameters needed to describe a log normal distribution of aerosol particle sizes are the median
19 diameter and the geometric standard deviation. When using aerodynamic diameters, the mass
20 median aerodynamic diameter (MMAD) refers to the median of the distribution of mass with
21 respect to the AED, the most commonly used measure of aerosol distribution.
22 As Chapter 7 notes, for dosimetry purposes, the respiratory tract can be divided into three
23 main regions: (1) extrathoracic (ET), (2) tracheobronchial (TB), and (3) alveolar (A). The ET
24 region consists of head airways (i.e., nasal or oral passages) through the larynx, the areas through
25 which inhaled air first passes. In humans, inhalation can occur via the nose or mouth or both
26 (i.e., oronasal breathing). From the ET region, inspired air enters the TB region at the trachea.
27 From the trachea, the conducting airways then undergo branching for several generations.
28 The terminal bronchioles are the most peripheral of the distal conducting airways and these lead,
29 in humans, to respiratory bronchioles, alveolar ducts, alveolar sacs, and alveoli, all comprising
30 the A region.
31
March 2001 9-27 DRAFT-DO NOT QUOTE OR CITE
-------
1 9.5.1 Particle Deposition in the Respiratory Tract
2 Knowledge of respiratory tract regional deposition patterns for particles of different sizes is
3 important for understanding possible health effects associated with exposure to ambient PM and
4 for extrapolating and interpreting data obtained from studies of laboratory animals. Particles
5 deposited in various respiratory tract regions are subjected to large differences in clearance
6 mechanisms and pathways and, consequently, retention times.
7 Particles deposit in the respiratory tract by five mechanisms: (1) inertial impaction,
8 (2) sedimentation, (3) diffusion, (4) electrostatic precipitation, and (5) interception. Sudden
9 changes in airstream direction and velocity cause inhaled particles to impact onto airway
10 surfaces. The ET and upper TB airways are dominant sites of inertial impaction, a key
11 mechanism for particles with AED >1 /urn. Particles with AED > 0.5 /urn mostly are affected by
12 sedimentation out of the airstream. Both sedimentation and inertial impaction influence
13 deposition of particles in the same size range and occur in the ET and TB regions, with inertial
14 impaction dominating in the upper airways and gravitational settling (sedimentation) increasingly
15 more dominant in lower conducting airways. Particles with actual physical diameters <1 /urn are
16 increasingly subjected to diffusive deposition due to random bombardment by air molecules,
17 resulting in contact with airway surfaces. Particles circa 0.3 to 0.5 //m in size are small enough
18 to be little influenced by impaction or sedimentation and large enough to be minimally
19 influenced by diffusion, and so, they undergo the least respiratory tract deposition. The
20 interception potential of any particle depends on its physical size; fibers are of chief concern for
21 interception, their aerodynamic size being determined mainly by their diameter. Electrostatic
22 precipitation is deposition related to particle charge; effects of charge on deposition are inversely
23 proportional to particle size and airflow rate. This type of deposition is likely small compared to
24 effects of other deposition mechanisms and is generally a minor contributor to overall particle
25 deposition, but one recent study found it to be a significant TB region deposition mechanism for
26 ultrafine, and some fine, particles.
27 Total human respiratory tract deposition, as a function of particle size, is depicted in
28 Figure 9-6 for healthy male adults under different ventilation conditions. The ET region acts as
29 an efficient filter that reduces penetration of inhaled particles to the TB and A regions of the
30 lower respiratory tract. Total respiratory tract deposition increases with particle size for particles
31 >1.0 jj.ro. AED, is at a minimum for particles 0.3 to 0.5 //m, and increases as particle size
March 2001 9-28 DRAFT-DO NOT QUOTE OR CITE
-------
.0
'55
o
Q.
a
100
80
60
40
20
0
^ O
I
O Human (Oral)
• Human (Nasal)
0 T
0.01
0.1 1.0
Particle Diameter (urn)
10
Figure 9-6. Total human respiratory tract deposition (percent deposition of amount
inhaled) as a function of particle size. AH values are means with standard
deviations as available. Particle diameters are aerodynamic (MMAD) for those
^0.5 fj,m.
Source: Modified from Schlesinger (1988).
1 decreases below that range. The ET deposition is higher with nose breathing than for mouth
2 breathing, with increased ventilation rates associated with increasing levels of physical activity or
3 exercise leading to more oronasal breathing and increased delivery of inhaled particles to TB and
4 A regions in the lung.
5 Hygroscopicity, the propensity of a material for taking up and retaining moisture, is a
6 property of some ambient particle species and affects respiratory tract deposition. Such particles
7 can increase in size in humid air in the respiratory tract and, when inhaled, deposit according to
March 2001
9-29
DRAFT-DO NOT QUOTE OR CITE
-------
1 their hydrated size rather than their initial size. Compared to nonhygroscopic particles of the
2 same initial size, deposition of hygroscopic aerosols in different regions varies, depending on
3 initial size: hygroscopicity generally increases total deposition for particles with initial sizes
4 larger than =0.5 /urn, but decreases deposition for smaller ones.
5 Enhanced particle retention occurs on carinal ridges in the trachea and through segmental
6 bronchi; and deposition "hot spots" occur at airway bifurcations or branching points. Peak
7 deposition sites shift from distal to proximal sites as a function of particle size, with greater
8 surface dose in conducting airways than in the A region for all particle sizes. Whereas both fine
9 (<2.5 /^m) and coarse (2.5 to 10 jam) inhalable particles deposit to about the same extent on a
10 percent particle mass basis in the trachea and upper bronchi, a distinctly higher percent of fine
11 particles deposit in the A region. However, surface number dose (particles/cm2/day) is much
12 higher for fine particles than for coarse, indicating much higher numbers of fine particles
13 depositing, with the fine fraction contributing upwards of 10,000 times greater particle number
14 per alveolar macrophage.
15 Ventilation rate, gender, age, and respiratory' disease status are all factors that affect total
16 and regional respiratory tract particle deposition. In general, because of somewhat faster
17 breathing rates and likely smaller airway size, women have somewhat greater deposition of
18 inhaled particles than men in upper TB airways, but somewhat lower A region deposition than
19 for men. Children appear to show four effects: (1) greater total respiratory tract deposition than
20 adults (possibly as much as 50% greater for those <14 years old than for adults >14 years),
21 (2) distinctly enhanced ET region deposition (decreasing with age from 1 year), (3) enhanced TB
22 deposition for particles <5 /u.m, and (4) enhanced A region deposition (also decreasing with age).
23 Overall, given that children have smaller lungs and higher minute volumes relative to lung size,
24 they likely receive greater doses of particles per lung surface area than adults for comparable
25 ambient PM exposures. This and the propensity for young children to generally exhibit higher
26 activity levels and associated higher breathing rates than adults likely contribute to enhanced
27 susceptibility to ambient particle effects resulting from particle dosimetry factors. In contrast,
28 limited available data on respiratory tract deposition across adult age groups (18 to 80 years) with
29 normal lung function do not indicate age-dependent effects (e.g., enhanced deposition in healthy
30 elderly adults). Altered PM deposition patterns resulting from respiratory disease status may put
March 2001 9-30 DRAFT-DO NOT QUOTE OR CITE
-------
1 certain groups of adults (including some elderly), as well as certain groups of children, at greater
2 risk for PM effects.
3 Both information noted in the 1996 PM AQCD and newly published findings indicate that
4 respiratory disease status is an especially important determinant of respiratory tract particle
5 deposition. Of particular importance is the finding that chronic obstructive disease states
6 contribute to more heterogenous deposition patterns and differences in regional deposition. One
7 new study indicates that people with COPD tend to breath faster and deeper than those with
8 normal lungs (i.e., about 50% higher resting ventilation), and had ca. 50% greater deposition than
9 age-matched healthy adults under typical breathing conditions and average deposition rates
10 2.5 times higher under elevated ventilation rates. Enhanced deposition appears to be associated
11 more with the chronic bronchitic than the emphysematous component of COPD. In this and
12 other new studies, fine-particle deposition increased markedly with increased degree of airway
13 obstruction (ranging up to ca. 100% greater with severe COPD). With increasing airway
14 obstruction and uneven airflow because of irregular obstruction patterns, particles tend to
15 penetrate more into remaining better ventilated lung areas, leading to enhanced focal deposition
16 at airway bifurcations and alveoli in those A region areas. In contrast, TB deposition increases
17 with increasingly more severe bronchoconstrictive states, as occur with asthmatic conditions.
18 Differences between humans and animals in deposition patterns were summarized in the
19 1996 PM AQCD and by Schlesinger et al. (1997) and should be considered when relating
20 biological responses obtained in laboratory animal studies to effects in humans. Various species
21 used in inhalation toxicology studies serving as the basis for dose-response assessment may not
22 receive identical doses in a comparable respiratory tract region (i.e., ET, TB, A) when exposed to
23 the same aerosol at the same inhaled concentration.
24 New mathematical modeling studies evaluate interspecies differences in respiratory tract
25 deposition. For example, Hofmann et al. (1996) found total deposition efficiencies for all
26 particles (0.01, 1, and 10 /um) at upper and lower airway bifurcations to be comparable for rats
27 and humans, but when higher penetration probabilities from preceding airways in the human lung
28 were considered, bronchial deposition fractions were mostly higher for humans. For all particle
29 sizes, deposition at rat bronchial bifurcations was less enhanced on the carinas than in human
30 airways. Numerical simulations of three-dimensional particle deposition patterns within selected
31 (species-specific) bronchial bifurcations indicated that interspecies differences in morphologic
March 2001 9-31 DRAFT-DO NOT QUOTE OR CITE
-------
1 asymmetry is a major determinant of local deposition patterns. The dependence of deposition on
2 particle size is similar in rats and humans, with deposition minima in the 0.1- to l-^m size range
3 for both total deposition and deposition in the TB and A regions, but total respiratory tract and
4 TB deposition was consistently higher in the human lung. Alveoli region deposition in humans
5 was lower than in rat for 0.001- to 10-//m particles (deposition of such particles being highest in
6 the upper bronchial airways), whereas it was higher for 0.1 - and 1 -//m particles in more
7 peripheral airways (i.e., bronchiolar airways in rat, respiratory bronchioles in humans). In a new
8 histology study, Nikula et al. (2000) examined particle retention in rats (exposed to diesel soot)
9 and humans (exposed to coal dust). In both, the volume density of deposition increased with
10 increasing dose, hi rats, diesel exhaust particles were found mainly in lumens of the alveolar
11 duct and alveoli, whereas in humans, retained dust was mainly in interstitial tissue. Thus, in the
12 two species, different lung cells appear to contact retained particles and may result in different
13 biological responses with chronic exposure.
14 The probability of any biological effect of PM in humans or animals depends on particle
15 deposition and retention, as well as underlying dose-response relationships. Interspecies
16 dosimetric extrapolation must consider differences in deposition, clearance, and dose-response.
17 Even similar deposition patterns may not result in similar effects in different species, because
18 dose also is affected by clearance mechanisms and species sensitivity. Total number of particles
19 deposited in the lung may not be the most relevant dose metric by which to compare species;
20 rather, the number of deposited particles per unit surface area may determine response. Even if
21 deposition is similar in rats and humans, there would be a higher deposition density in the rat
22 because of the smaller surface area of rat lung. Thus, species-specific differences in deposition
23 density are important when attempting to extrapolate health effects observed in laboratory
24 animals to humans.
25
26 9.5.2 Particle Clearance and Translocation
27 Particles depositing on airway surfaces may be cleared from the respiratory tract completely
28 or translocated to other sites within this system by regionally specific clearance mechanisms, as
29 follow: ET region—mucocialiary transport, sneezing, nose wiping and blowing, and dissolution
30 and absorption into blood; TB region—mucociliary transport, endocytosis by macrophages and
31 epithelial cells, coughing, and dissolution and absorption into blood and lymph;
March 2001 9-32 DRAFT-DO NOT QUOTE OR CITE
-------
1 A region—macrophages, epithelial cells, interstitial, and dissolution and absorption into blood
2 and lymph. Clearance routes from various respiratory tract regions are depicted in Chapter 7
3 (Figures 7-2 and 7-3).
4 Regionally specific clearance defense mechanisms operate to clear deposited particles of
5 varying particle characteristics (size, solubility, etc.) from the ET, TB, and A regions and are
6 variously affected by different disease states. For example, particles are cleared from the ET
7 region by mucociliary transport to the nasopharynx area, dissolution and absorption into the
8 blood, or sneezing, wiping or blowing of the nose, but such clearance is slowed by chronic
9 sinusitis, bronchiectasis, rhinitis, and cystic fibrosis. Also, in the TB region, poorly soluble
10 particles are cleared mainly by upward mucociliary transport or by phagocytosis by airway
11 macrophages that move upward on the mucociliary blanket, followed by swallowing. Soluble
12 particles in the TB region are absorbed mostly into the blood and some by mucociliary transport.
13 Although TB clearance is generally fast and much material is cleared in <24 h, the slow
14 component of TB clearance (likely associated with bronchides 24 h and clearance
16 half-times of about 50 days. Bronchial mucous transport is slowed by bronchial carcinoma,
17 chronic bronchitis, asthma, and various acute respiratory infections; these are disease conditions
18 that logically would be expected to increase retention of deposited particle material and, thereby,
19 increase the probability of toxic effects from inhaled ambient PM components reaching the TB
20 region. Also, spontaneous coughing, an important TB region clearance mechanism, does not
21 appear to fully compensate for impaired mucociliary clearance in small airways and may become
22 depressed with worsening airway disease, as seen in COPD.
23 Clearance of particles from the A region via alveolar macrophages and their mucociliary
24 transport is usually rapid (<24 h). However, penetration of uningested particles into the
25 interstitium increases with increasing particle load and results in increased translocation to lymph
26 nodes. Soluble particles not absorbed quickly into the blood stream and translocated to
27 extrapulmonary organs (e.g., the heart) within minutes also may enter the lymphatic system, with
28 lymphatic translocation probably being increased as other clearance mechanisms (e.g., removal
29 by macrophages) are taxed or overwhelmed under "particle overload" conditions. Particles
30 <2 /urn clear to the lymphatic system at a rate independent of size; particles of this size, more so
31 than those >5.0 /urn, are deposited significantly in the A region. Translocation into the lymphatic
March 2001 9-33 DRAFT-DO NOT QUOTE OR CITE
-------
1 system is quite slow, and elimination from lymph nodes even slower (half-times estimated in
2 decades). Focal accumulations of reservoirs of potentially toxic materials and their slow release
3 for years after initial ambient PM exposure may account partially for the higher relative risks
4 observed in epidemiologic studies to be associated with long-term ambient PM exposure beyond
5 additive effects of acute PM exposures. Alveolar region clearance rates are decreased in human
6 COPD sufferers and slowed by acute respiratory infections, and the viability and functioning of
7 alveolar macrophages are reduced in human asthmatics and in animals with viral lung infections,
8 this suggests that persons with asthma or acute lung infections are likely at increased risk for
9 ambient PM exposure effects.
10 Differences in regional and total clearance rates between some species reflect differences in
11 mechanical clearance processes. The importance of interspecies clearance differences is that
12 retention of deposited particles can differ between species and may result in differences in
13 response to similar PM exposures. Hsieh and Yu (1998) summarize existing data on pulmonary
14 clearance of inhaled, poorly soluble particles in the rat, mouse, guinea pig, dog, monkey, and
15 human. Two clearance phases "fast" and "slow" in the A region are associated with mechanical
16 clearance along two pathways, the former with the mucociliary system and the latter with lymph
17 nodes. Rats and mice are fast clearers, compared to other species. Increasing initial lung burden
18 results in an increasing mass fraction of particles cleared by the slower phase. As lung burden
19 increases beyond 1 mg particle/g lung, the fraction cleared by the slow phase increases to almost
20 100% for all species. The rate for the fast phase is similar in all species, not changing with
21 increasing lung burden, whereas the slow phase rate decreases with increasing lung burden.
22 At elevated burdens, the "overload" effect on clearance rate is greater in rats than in humans.
23
24 9.5.3 Deposition and Clearance Patterns of Particles Administered by
25 Inhalation Versus Intratracheal Instillation
26 Inhalation is the most directly relevant exposure route for evaluating PM toxicity, but many
27 studies deliver particles by intratracheal instillation. Because particle disposition is a determinant
28 of dose, it is important to compare deposition and clearance of particles delivered by instillation
29 versus inhalation. It is difficult to compare particle deposition and clearance among different
30 inhalation and instillation studies because of differences in experimental methods and in
31 quantification of particle deposition and clearance. Key points from a recent detailed evaluation
March 2001 9-34 DRAFT-DO NOT QUOTE OR CITE
-------
1 (Driscoll et al., 2000) of the role of instillation in respiratory tract dosimetry and toxicology
2 studies are informative. In brief, inhalation may result in deposition within the ET region, the
3 extent of which depends on the size of the particles used, but intratracheal instillation bypasses
4 this portion of the respiratory tract and delivers particles directly to the TB tree. Although some
5 studies indicate that short (0 to 2 days) and long (100 to 300 days postexposure) phases of
6 clearance of insoluble particles delivered either by inhalation or intratracheal instillation are
7 similar, others indicate that the percent retention of particles delivered by instillation is greater
8 than for inhalation, at least up to 30 days postexposure. Another salient finding is that inhalation
9 generally results in a fairly homogeneous distribution of particles throughout the lungs, but
10 instillation is typified by heterogeneous distribution (especially in the A region) and high levels
11 of focal particles. Most instilled material penetrates beyond the major tracheobronchial airways,
12 but the lung periphery is often virtually devoid of particles. This difference is reflected in
13 particle burdens within macrophages, those from animals inhaling particles being burdened more
14 homogeneously and those from animals with instilled particles showing some populations of
15 cells with no particles and others with heavy burdens, and is likely to impact clearance pathways,
16 dose to cells and tissues, and systemic absorption. Exposure method, thus, clearly influences
17 dose distribution that argues for caution in interpreting results from instillation studies.
18
19 9.5.4 Inhaled Particles as Potential Carriers of Toxic Agents
20 It has been proposed that particles also may act as carriers to transport toxic gases into the
21 deep lung. Water-soluble gases, which would be removed by deposition to wet surfaces in the
22 upper respiratory system during inhalation, could dissolve in particle-bound water and be carried
23 with the particles into the deep lung. Equilibrium calculations indicate that particles do not
24 increase vapor deposition in human airways. However, these calculations do show that soluble
25 gases are carried to higher generation airways (deeper into the lung) in the presence of particles
26 than in the absence of particles. In addition, species such as SO2 and formaldehyde react in
27 water, reducing the concentration of the dissolved gas-phase species and providing a kinetic
28 resistence to evaporation of the dissolved gas. Thus, the concentration of the dissolved species
29 may be greater than that predicted by the equilibrium calculations. Also, certain other toxic
30 species (e.g., nitric oxide [NO], nitrogen dioxide [NO2], benzene, polycyclic aromatic
31 hydrocarbons [PAH], nitro-PAH, a variety of allergens) may be absorbed onto solid particles and
March 2001 9-35 DRAFT-DO NOT QUOTE OR CITE
-------
1 carried into the lungs. Thus, ambient particles may play important roles not only in inducing
2 direct health impacts of their constituent components but also in facilitating delivery of toxic
3 gaseous pollutants or bioagents into the lung and may, thereby, serve as key mediators of health
4 effects caused by the overall air pollutant mix.
5
6
7 9.6 HEALTH EFFECTS OF AMBIENT PARTICULATE MATTER
8 9.6.1 Introduction
9 This section evaluates available scientific evidence regarding the physiologic and health
10 effects of exposure to ambient PM. The three main objectives of this evaluation are (1) to
11 summarize and evaluate strengths and limitations of available epidemiologic findings; (2) to
12 assess the biomedical coherence of findings across studied endpoints; and (3) to evaluate the
13 biologic plausibility of available evidence in light of (a) linkages between specific PM
14 components and health effects and (b) various dosimetric, mechanistic, and pathophysiologic
15 considerations.
16 Epidemiologic findings are emphasized first because they provide the strongest body of
17 evidence directly relating ambient PM concentrations to biomedical outcomes. Numerous
18 epidemiologic studies have shown statistically significant associations of ambient PM levels with
19 a variety of human health endpoints, including mortality, hospital admissions, emergency
20 department visits, other medical visits, respiratory illness and symptoms measured in community
21 surveys, and physiologic changes in pulmonary function. Associations have been consistently
22 observed between both short- and long-term PM exposure and these endpoints. The general
23 internal consistency of the epidemiologic database and available findings demonstrate well that
24 notable human health effects are associated with exposures to ambient PM at concentrations
25 currently found in many geographic locations across the United States. However, many
26 difficulties still exist with regard to delineating the magnitudes and variabilities of risk estimates
27 for ambient PM, the ability to attribute observed health effects to specific PM constituents, the
28 time intervals over which PM health effects are manifested, the extent to which findings in one
29 location can be generalized to other locations, and the nature and magnitude of the overall public
30 health risk imposed by ambient PM exposure.
March 2001 9-36 DRAFT-DO NOT QUOTE OR CITE
-------
1 The etiology of most air-pollution-related health outcomes is highly multifactorial, and the
2 impact of ambient air pollution exposure on these outcomes is often small in comparison to that
3 of other etiologic factors (e.g., smoking). Also, ambient PM exposure usually is accompanied by
4 exposure to many other pollutants, and PM itself is composed of numerous physical/chemical
5 components. Assessment of the health effects attributable to PM and its constituents within an
6 already-subtle total air pollution effect is difficult even with well-designed studies. Indeed,
7 statistical partitioning of separate pollutant effects may not characterize fully the etiology of
8 effects that actually depend on simultaneous exposure to multiple air pollutants, hi this regard,
9 several viewpoints existed at the time of the 1996 PM AQCD regarding how best to interpret the
10 epidemiology data: one saw the PM exposure indicators as surrogate measures of complex
11 ambient air pollution mixtures, and the reported PM-related effects as representative of those of
12 the overall mixture; another held that reported PM-related effects are attributable to PM
13 components (per se) of the air pollution mixture and reflect independent PM effects, and a third
14 viewpoint holds that PM can be viewed both as a surrogate indicator, as well as a specific cause
15 of health effects.
16 Several other key issues and problems also must be considered when attempting to interpret
17 the data reviewed in this document. For example, although the epidemiology data provide strong
18 support for the associations mentioned above, questions remain regarding potential underlying
19 mechanisms. Although much progress has been made toward identification of anatomic sites at
20 which particles trigger specific health effects and elucidation of biological mechanisms
21 underlying induction of such effects, this area of scientific inquiry is still at an early stage.
22 Nevertheless, compared to the lack of much solid evidence available in the 1996 PM AQCD,
23 there now is a stronger basis for assessing biologic plausibility of the epidemiologic observations
24 given notable improvement in conceptual formulation of reasonable mechanistic hypotheses and
25 evidence bearing on such hypotheses. Several hypotheses are discussed later with regard to
26 possible mechanisms by which ambient PM may exert human health effects, and new evidence is
27 discussed that tends to support a causal relationship between low ambient concentrations of PM
28 and observed increased mortality or morbidity risks. At the same time, much still remains to be
29 done to identify more confidently specific causal agents among typical ambient PM constituents.
30
March 2001 9-37 DRAFT-DO NOT QUOTE OR CITE
-------
1 9.6.2 Community-Health Epidemiologic Evidence for Ambient Particulate
2 Matter Effects
3 In recent years, epidemiologic studies showing associations of ambient air pollution
4 exposure with mortality, exacerbation of preexisting illness, and pathophysiologic changes have
5 increased concern about the extent to which exposure to ambient air pollution exacerbates or
6 causes harmful health outcomes at pollutant concentrations now experienced in the United
7 States. The PM epidemiology studies assessed in the 1996 PM AQCD implicated ambient PM
8 as a likely key contributor to mortality and morbidity effects observed epidemiologically to be
9 associated with ambient air pollution exposures. New studies appearing since the 1996 PM
10 AQCD are important in extending results of earlier studies to many more cities and in confirming
11 earlier findings.
12 In epidemiologic studies of ambient air pollution, small positive estimates of air pollutant
13 health effects have been observed quite consistently, frequently being statistically significant at
14 p < 0.05. If ambient air pollution promotes or produces harmful health effects, relatively small
15 effect estimates from current PM concentrations in the United States and many other countries
16 would generally be expected on biological and epidemiologic grounds. Also, magnitudes and
17 significance levels of observed air pollution-related effects estimates would be expected to vary
18 somewhat from place to place, if the observed epidemiologic associations denote actual effects,
19 because (a) not only would the complex mixture of PM vary from place to place, but also
20 (b) affected populations may differ in characteristics that could affect susceptibility to air
21 pollution health effects. Such characteristics include sociodemographic factors, underlying
22 health status, indoor-outdoor activities, diet, medical care access, exposure to risk factors other
23 than ambient air pollution (such as extreme weather conditions), and variations in factors (e.g.,
24 air-conditioning) affecting human exposures to ambient-generated PM.
25 Although it has been argued by some that the observed effects estimates for ambient air
26 pollution are not sufficiently constant across epidemiologic studies and that epidemiologic
27 studies are trustworthy only if they show relatively large effects estimates (e.g., large relative
28 risks), these arguments have only limited weight in relation to ambient air pollution studies.
29 Also, in any large population exposed to ambient air pollution, even a small relative risk for a
30 widely prevalent health disorder could result in a substantial public health burden attributable to
31 air pollution exposure.
March 2001 9-38 DRAFT-DO NOT QUOTE OR CITE
-------
1 As noted above, small health effects estimates generally have been observed for ambient air
2 pollutants, as would be expected on biological and epidemiologic grounds. In contrast to effects
3 estimates derived for the 1952 London smog episode with relative risk (RR) exceeding 4.0 (i.e.,
4 400% increase over baseline) for extremely high (>2 mg/m3) ambient PM concentrations, effects
5 estimates in most current epidemiology studies at distinctly lower PM concentrations (often
6 < 100 Mg/m3) are relatively small. The statistical estimates (1) are more often subject to small
7 (but proportionately large) differences in estimated effects of PM and other pollutants; (2) may
8 be sensitive to a variety of methodological choices; and (3) sometimes may not be statistically
9 significant, reflecting low statistical power of the study design to detect a small but real effect.
10 The ambient atmosphere contains numerous air pollutants, and it is important to continue to
11 recognize that health effects associated statistically with any single pollutant may actually be
12 mediated by multiple components of the complex ambient mix. Specific attribution of effects to
13 any single pollutant may therefore be overly simplistic. Particulate matter is one of many air
14 pollutants derived from combustion sources, including mobile sources. These pollutants include
15 PM, carbon monoxide (CO), sulfur oxides, nitrogen oxides, and ozone, all of which have been
16 considered in various epidemiologic studies to date. Many volatile organic compounds (VOCs)
17 or semivolatile compounds (SVOCs) also emitted by combustion sources or formed in the
18 atmosphere have not yet been systematically considered in relation to noncancer health outcomes
19 usually associated with exposure to criteria air pollutants. In many newly available
20 epidemiologic studies, harmful health outcomes are often associated with multiple combustion-
21 related or mobile-source-related air pollutants, and some investigators have raised the possibility
22 that PM may be a key surrogate or marker for a larger subset of the overall ambient air pollution
23 mix. This possibility takes on added potential significance to the extent that ambient aerosols
24 indeed may not only exert health effects directly attributable to their constituent components, per
25 se, but also serve as carriers for more efficient delivery of water soluble toxic gases (e.g., O3,
26 NO2, SO2) deeper into lung tissue, as noted earlier in Section 9.5.5. This suggests that airborne
27 particle effects may be enhanced by the presence of other toxic agents or mistakenly attributed to
28 them if their respective concentrations are highly correlated temporally. Thus, although
29 associations of PM with harmful effects continue to be observed consistently across most new
30 studies, the newer findings do not fully resolve issues concerning relative contributions to the
31 observed epidemiologic associations of (1) PM acting alone, (2) PM acting in combination with
March 2001 9-39 DRAFT-DO NOT QUOTE OR CITE
-------
1 gaseous co-pollutants, (3) the gaseous pollutants per se, and (4) the overall ambient pollutant
2 mix.
3 It seems likely that, for pollutants whose concentrations are not highly correlated, effects
4 estimates in multipollutant models would be more biologically and epidemiologically sound than
5 those in single-pollutant models, although it is conceivable that single-pollutant models also
6 might be credible if independent biological plausibility evidence supported designation of PM or
7 some other single pollutant as likely being the key toxicant in the ambient pollutant mix
8 evaluated. However, neither of these possibilities have been demonstrated convincingly, and
9 scientific consensus as to optimal interpretation of modeling outcomes for time series air
10 pollution studies has not yet been achieved. Therefore, the choice of appropriate effects
11 estimates to employ in risk assessments for ambient PM effects remains a difficult issue. Issues
12 related to confounding by co-pollutants, along with issues related to time scales of exposure and
13 response and concentration-response function, importantly apply to new epidemiologic studies
14 relating concentrations of PM or correlated ambient air pollutants to hospital admissions,
15 exacerbation of respiratory symptoms, and asthma in children, to reduced pulmonary function in
16 children and adults, and to changes in heart rate, and heart rate variability in adults.
17 With considerable new experimental evidence also in hand, it is now possible to
18 hypothesize various ways in which ambient exposure to multiple air pollutants (including not
19 only PM acting alone but also in combination with others) could plausibly be involved in the
20 complex chain of biological events leading to harmful health effects in the human population.
21 The newer experimental evidence, therefore, adds considerable support for interpreting the
22 epidemiologic findings discussed below as being indicative of causal relationships between
23 exposures to ambient PM and consequent associated increased morbidity and mortality risks.
24
25 9.6.2.1 Short-Term Particulate Matter Exposure Effects on Mortality
26 This section focuses primarily on discussion of short-term PM exposure effects on
27 mortality, but also highlights some morbidity effects in relation to the mortality findings.
28 Morbidity effects are discussed more fully after discussion of long-term mortality effects in the
29 section following this one.
30
31
March 2001 9-40 DRAFT-DO NOT QUOTE OR CITE
-------
1 9.6,2.1.1 Summary of Previous Findings on Short-Term Paniculate Matter Exposure-
2 Mortality Effects
3 Time series mortality studies reviewed in the 1996 PM AQCD provided strong evidence
4 that ambient PM air pollution is associated with increased daily mortality. The 1996 PM AQCD
5 summarized about 35 PM-mortality time series studies published between 1988 and 1996.
6 Available information from those studies was consistent with the hypothesis that PM is a causal
7 agent in the mortality impacts of air pollution. The PM,0 relative risk estimates derived from the
8 PM,0 studies reviewed in the 1996 PM AQCD suggested that an increase of 50 /ug/m3 in the 24-h
9 average of PM10 is associated with an increased risk of premature total mortality (total deaths
10 minus accidents and injuries) mainly on of the order of relative risk (RR) = 1.025 to 1.05 (i.e.,
11 2.5 to 5.0% excess risk) in the general population, with statistically significant increases being
12 reported more broadly across the range of 1.5 to 8.5% per 50 /ug/nr1 PMI0. Higher relative risks
13 were indicated for the elderly and for those with preexisting respiratory conditions. Also, based
14 on the then recently published Schwartz et al. (1996a) analysis of Harvard Six City data, the 1996
15 PM AQCD found the relative risk for excess total mortality in relation to 24-h fine-particle
16 concentrations to be in the range of RR = 1.026 to 1.055 per 25 yug/m3 PM2 5 (i.e., 2.6 to 5.5%
17 excess risk per 25 /ug/m3 PM2 5). Relative risk estimates for morbidity and mortality effects
18 associated with standard increments in ambient PM10 concentrations and for fine-particle
19 indicators (e.g., PM25, sulfates, etc.) were presented in Chapters 12 and 13 of the 1996 PM
20 AQCD (see Appendix 9A), and those effect estimates are updated below in light of the extensive
21 newly available evidence discussed in Chapter 6 of this document.
22 Although numerous studies reported PM-mortality associations, several important issues
23 needed to be addressed in interpreting those relative risks. The 1996 PM AQCD extensively
24 discussed the following critical issues: (1) seasonal confounding and effect modification,
25 (2) confounding by weather, (3) confounding by co-pollutants, (4) measurement error,
26 (5) functional form and threshold, (6) harvesting and life shortening; and (7) the roles of specific
27 PM components.
28 Season-specific analyses are often not feasible because of small magnitudes of expected
29 effect size or small sample sizes (low power) available for some studies. Some studies had
30 earlier suggested possible season-specific variations in PM coefficients, but it was not clear if
31 these were caused by peak variations in PM effects from season to season, varying extent of PM
March 2001 9-41 DRAFT-DO NOT QUOTE OR CITE
-------
1 correlations with other co-pollutants, or weather factors during different seasons. The likelihood
2 of PM effects being accounted for mainly by weather factors was addressed by various methods
3 that controlled for weather variables in most studies (including some involving sophisticated
4 synoptic weather pattern evaluations), and that possibility was found to be very unlikely.
5 Many early PM studies considered at least one co-pollutant in the mortality regression, and
6 an increasing number have examined multiple pollutants. Usually, when PM indices were
7 significant in single-pollutant models, addition of a co-pollutant diminished the PM effect size
8 somewhat, but did not eliminate PM associations. In multiple-pollutant models performed by
9 season, the PM coefficients became less stable, again possibly because of varying correlations of
10 PM with co-pollutants among seasonal or smaller sample sizes. However, in many studies, PM
11 indices showed the highest significance in both single- and multiple-pollutant models. Thus,
12 PM-mortality associations did not appear to be seriously distorted by co-pollutants.
13 Interpretation of the relative significance of each pollutant in mortality regression in
14 relation to its relative causal strength was difficult, however, because of lack of quantitative
15 information on pertinent exposure measurement errors among the air pollutants. Measurement
16 errors can influence the size and significance of air pollution coefficients in time series
17 regression analyses, an issue also important in assessing confounding among multiple pollutants,
18 because the varying extent of such errors among pollutants may influence corresponding relative
19 significance. The 1996 PM AQCD discussed several types of exposure measurement and
20 characterization errors, including site-to-site variability and site-to-person variability. These
21 errors are thought to bias the estimated PM coefficients downward in most cases, but there was
22 insufficient quantitative information available at the time to allow estimation of such bias.
23 The 1996 PM AQCD also reviewed evidence for threshold and various other functional
24 forms of short-term PM mortality associations. Some studies indicated that associations were
25 seen monotonically to even below the PM standards. It was considered difficult, however, to
26 statistically identify a threshold from available data because of low data density at lower ambient
27 PM concentrations, potential influence of measurement error, and adjustments for other
28 covariates. Thus, use of relative risk (rate ratio) derived from log-linear Poisson models was
29 deemed adequate.
30 The extent of prematurity of death (i.e., mortality displacement [or harvesting]) in observed
31 PM-mortality associations has important public health policy implications. At the time of the
March 2001 9-42 DRAFT-DO NOT QUOTE OR CITE
-------
1 1996 PM AQCD review, only a few studies had investigated this issue. Although one of the
2 studies suggested that the extent of such prematurity might be only a few days, this may not be
3 generalized because this estimate was obtained for identifiable PM episodes. Insufficient
4 evidence then existed to suggest the extent of prematurity for nonepisodic periods, from which
5 most of the recent PM relative risks were derived.
6 Only a few PM-mortality studies had analyzed fine particles and chemically specific
7 components of PM. The Harvard Six Cities Study (Schwartz et al., 1996a) analyzed size-
8 fractionated PM (PM2 5, PM10/15, and PM10/15.2 5) and PM chemical components (sulfates and H+).
9 The results suggested that PM2 5 was associated most significantly with mortality among the PM
10 components. Although FT was not significantly associated with mortality in this and earlier
11 analyses, the smaller sample size for H+ than for other PM components made direct comparison
12 difficult. Also, certain respiratory morbidity studies showed associations between hospital
13 admissions and visits with components of PM in the fine-particle range. Thus, the 1996 PM
14 AQCD concluded that there was adequate evidence to suggest that fine particles play especially
15 important roles in observed PM mortality effects.
16 Overall, then, the outcome of assessment of the above key issues in the 1996 PM AQCD
17 can be thusly summarized: (1) observed PM effects are not likely seriously biased by inadequate
18 statistical modeling (e.g., control for seasonality); (2) observed PM effects are not likely
19 significantly confounded by weather; (3) observed PM effects may be confounded or modified to
20 some extent by co-pollutants, and such extent may vary from season to season; (4) determining
21 the extent of confounding and effect modification by co-pollutants requires knowledge of relative
22 exposure measurement/characterization error among pollutants (there was not sufficient
23 information on this); (5) no clear evidence for any threshold for PM-mortality associations was
24 reported (statistically identifying a threshold from existing data also was considered difficult, if
25 not impossible); (6) some limited evidence for harvesting, a few days of life-shortening, was
26 reported for episodic periods (no study was conducted to investigate harvesting in nonepisodic
27 U.S. data); and (7) only a relatively limited number of studies suggested a causal role of fine
28 particles in PM-mortality associations, but in light of historical data, biological plausibility, and
29 results from morbidity studies, a greater role for fine particles than coarse particles was suggested
30 as being likely.
31
March 2001 9-43 DRAFT-DO NOT QUOTE OR CITE
-------
1 9.6.2.1.2 Updated Epidemiologic Findings for Short-Term Ambient Particulate Matter
2 Exposure Effects on Mortality
3 With regard to updating the assessment of PM effects in light of new epidemiologic
4 information published since the 1996 PM AQCD, the most salient key points on relationships
5 between short-term PM exposure and mortality (drawn from Chapter 6 discussions in this
6 document) can be summarized as follows.
7 Since the 1996 PM AQCD, there have been more than 70 new time-series PM-mortality
8 analyses, several of which investigated multiple cities using consistent data analytical
9 approaches. With only few exceptions, the estimated mortality relative risks in these studies are
10 generally positive, many are statistically significant, and they generally comport well with
11 previously reported PM-mortality effects estimates delineated in the 1996 PM AQCD. There are
12 also now numerous additional studies demonstrating associations between short-term (24-h) PM
13 exposures and various morbidity endpoints.
14 Several new studies conducted time series analyses in multiple cities. The major advantage
15 of these studies over meta-analyses for multiple "independent" studies is the consistency in data
16 handling and model specifications, thus eliminating variation in results attributable to study
17 design. Also, many of the cities included in these studies were ones for which no earlier time
18 series analyses had been conducted. Therefore, unlike regular meta-analysis, they likely do not
19 suffer from omission of negative studies caused by publication bias. Furthermore, any spatial or
20 geographic variability of air pollution effects can be systematically evaluated in such multi-city
21 analyses.
22 PMJO Effect Size Estimates. In the NMMAPS (Samet et al., 2000a,b) analysis of the
23 90 largest U.S. cities, the combined nationwide relative risk estimate was about a 2.3% increase
24 in total mortality per SO-^g/m3 increase in PMI0. The NMMAPS effect size estimates did vary
25 somewhat by U.S. region (see Figures 6-2 and 6-3), with the largest estimate being for the
26 Northeast (4.5% for a 1-day lag, the lag typically showing maximum effect size for most U.S.
27 regions). Various other U.S. multi-city analyses, as well as single-city analyses, obtained PM10
28 effect sizes mainly in the range of 2.5 to 5.0% per 50-^g/m3 increase in PM,0. There is some
29 evidence that, if the effects over multiple days are considered, the effect size may be larger.
30 What heterogeneity existed for the estimated PMIO risks across NMMAPS cities could not be
31 explained with the city-specific explanatory variables (e.g., as the mean levels of pollution and
March 2001 9-44 DRAFT-DO NOT QUOTE OR CITE
-------
1 weather), mortality rate, sociodemographic variables (e.g., median household income),
2 urbanization, or variables related to measurement error.
3 Also, the multi-city APHEA study showed generally consistent associations between
4 mortality and both SO2 and PM indices in western European cities, but not for central and eastern
5 European cities. The pooled estimate of PM10-mortality relative risks for western European cities
6 comport well with estimates derived from U.S. data. The contrast between western and
7 central/eastern Europe results might result from possible differences in representativeness of
8 exposure measures, air pollution mix or resultant toxicity, proportions of sensitive
9 subpopulations, climate, etc.
10 Certain other individual-city studies using similar methodology in analyses for each city
11 (but not generating combined overall pooled effect estimates) also report variations in PM effect
12 size estimates between cities and in their robustness to inclusion of gaseous copollutants in
13 multi-pollutant models. Thus, one cannot entirely rule out that real differences may exist in
14 excess risk levels associated with varying size distributions, number, or mass of the chemical
15 constituents of ambient PM; the combined influences of varying co-pollutants present in the
16 ambient air pollution mix from location to location or season to season; or to variations in the
17 relationship between exposure and ambient PM concentration.
18 Nevertheless, there still appears to be reasonably good consistency among the results
19 derived from those several new multi-city studies providing pooled analyses of data combined
20 across multiple cities (thought to yield the most precise effect size estimates). Such analyses
21 indicate the percent excess total (nonaccidental) deaths estimated per 50 jUg/m3 increase in 24-h
22 PM10 to be 2.3% in the 90 largest U.S. cities (4.5% in the Northeast region); 3.4% in 10 U.S.
23 cities; 3.5% in the eight largest Canadian cities; and about 2.0% in western European cities
24 (using PM,0 = TSP*0.55). These combined estimates are reasonably consistent with the range of
25 PM10 estimates previously reported in the 1996 PM AQCD (i.e., 1.5 to 8.5% per 50 Aig/m3 PM)0).
26 These and other excess risk estimates from many other individual-city studies comport well with
27 a number of new studies confirming increased cause-specific cardiovascular- and respiratory-
28 related mortality, and those noted below as showing ambient PM associations with increased
29 cardiovascular and respiratory hospital admissions and medical visits.
30
March 2001 9-45 DRAFT-DO NOT QUOTE OR CITE
-------
1 Fine and Coarse Particle Effect Size Estimates. Table 9-3 summarizes effects estimates
2 (RR values) for increased mortality and/or morbidity associated with variable increments in
3 short-term (24-h) exposures to ambient fine particles indexed by various fine PM indicators
4 (PM2 5, sulfates, H+, etc.) in U.S. and Canadian cities. Table 9-4 shows analogous effect size
5 estimates for inhalable thoracic fraction coarse particles (i.e., PM^ 5). In both tables, studies
6 that were highlighted in comparable tables in the 1996 PM AQCD are indicated by italics.
7 The effect size estimates derived for PM2 5 as an ambient fine particle indicator (especially
8 those based on directly measured versus estimated PM2 5 levels) generally appear to fall in the
9 range of 2.0 to 8.5% increase in total (nonaccidental) deaths per 25-^g/m3 increment in 24-h
10 PM2 5 for U.S. and Canadian cities. Cause-specific effects estimates appear to fall mainly in the
11 range of 3.0 to 7.0% per 25 /wg/m3 24-h PM2 5 for cardiovascular or combined cardiorespiratory
12 mortality and 2.0 to 7.0% per 25 /ug/m3 24-h PM2 5 for respiratory mortality in U.S. cities.
13 In the 1996 PM AQCD, there was only one study, the Harvard Six Cities study, in which
14 the relative importance of fine and coarse particles was examined. That study suggested that fine
15 particles, but not coarse particles, were associated with daily mortality. Now, more than
16 10 studies have analyzed both PM2 5 and PM,0_2 5 for their associations with mortality (see
17 Figure 9-7). Although some of these studies (e.g., the Santa Clara County, CA, analysis and the
18 eight largest Canadian cities analysis) suggest that PM2 5 is more important than PM,0_2 5 in
19 predicting mortality fluctuations, several others (e.g., the Mexico City and Santiago, Chile
20 studies) seem to suggest that PM10.2 5 may be as important as PM2 5 in certain locations (some
21 shown to date being drier, more arid areas). Seasonal dependence of PM components'
22 associations observed in some of the locations (e.g., higher coarse [PM10_2 5] fraction estimates for
23 summer than winter in Santiago, Chile) hint at possible contributions of biogenic materials (e.g.,
24 molds, endotoxins, etc.) to the observed coarse particle effects in at least some locations.
25 Overall, for U.S. and Canadian cities, effect size estimates for the coarse fraction (PMI0.2 5) of
26 those inhalable thoracic particles capable of depositing in TB and A regions of the respiratory
27 tract generally appear to fall in the range of 0.5 to 6.0% excess total (nonaccidental) deaths per
28 25 Aig/m3 of 24-h PMIO_2 5. Respective increases for cause-specific mortality are 3.0 to 8.0% for
29 cardiovascular and 3.0 to 16.0% for respiratory causes per 25-jUg/m3 increase in 24-h PM10_2 5.
30 Chemical Components of Particulate Matter. Several new studies examined the role of
31 specific chemical components of PM. Studies of U.S. and Canadian cities showed mortality
March 2001 9-46 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9-3. EFFECT ESTIMATES PER VARIABLE INCREMENTS IN 24-HOUR
CONCENTRATIONS OF FINE PARTICLE INDICATORS (PM2 5, SO=, H+)
FROM U.S. AND CANADIAN STUDIES*
Study Location
Indicator
RR(±CI)**per25-A25 ^g/m3) 2.868 (1.126, 7.250)
(<25 ^g/m3) 0.779 (0.610, 0.995)
1 .06 (NS, from figure)
1.003(0.992, 1.015)
1.118(1.013, 1.233)
1.053(1.018, 1.090)
1.043(1.028, 1.059)
1.057(1.001, 1.115)
1.018(0.946, 1.095)
1.030(1.011,1.050)
77.2 (±7.8)
72.2 (±7.4)
15. 7 (±9. 2)
18.7 (±10.5)
20.8 (±9.6)
29.6 (±21.9)
Median 14.7
Means 11. 3-30.5
13(2,105)
61.7(0.78,390.5)
nmol/m3
17.28 (-0.6, 72.6)
18(6,86)
13.0(0,42)
NR
22 (4, 86)
32.5(9.3, 190.1)
16.8(5,48)
15.6 (±9.2)
42.1 (±22.0)
39.9 (±18.0)
37.1 (±19.8)
13.3 (max 86)
March 2001
9-47
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9-3 (cont'd). EFFECT ESTIMATES PER VARIABLE INCREMENTS IN
24-HOUR CONCENTRATIONS OF FINE PARTICLE INDICATORS (PM2 5, SO;, H+)
FROM U.S. AND CANADIAN STUDIES*
Study Location
Toronto, Canada0
Montreal, Canadap
Indicator
Est. PM25
PM25
RR (± CI)** per 25-^g/m3
PM Increase or 1 S-^g/m3
SOJ Increase or 75-nmol/m3 H+ increase
1.048(1.033, 1.064)
1.044(1.025, 1.063)
Reported PM
Levels Mean (Min,
Max)***
18.0(8,90)
17.4(2.2,72.0)
Cause-Specific Mortality
Cardiorespiratory:
Three New Jersey Ctties:M
Newark, NJ
Camden, NJ
Elizabeth, NJ
Total Cardiovascular:
Santa Clara County, CAC
Buffalo, NY)D
Philadelphia, PAF
(seven-county area)
Detroit, MI°
Phoenix, AZH
Los Angeles, CA1
San Bernadino and
Riverside Counties, CAJ
Coachella Valley, CAK
Cerebro vascular:
Los Angeles, CA1
Total Respiratory:
Santa Clara County, CAC
Buffalo, NYD
Philadelphia, PAF
(seven-county area)
Detroit, MIG
San Bernadino and
Riverside Counties, CAJ
PM25
PM25
PM25
PM25
S04=
PM25
PM25
PM25
PM25
Est. PM25
PM25
PM25
PM25
so;
PM25
PM25
Est. PM25
1.051 (1.031, 1.072)
1.062(1.006, 1.121)
1.023(0.950, 1.101)
1.07(p>0.05)
1.040(0.995, 1.088)
1. 028 (p< 0.055)
1.032(0.977,1.089)
1.187(1.057,1.332)
1.266(1.003, 1.048)
1.007(0.997, 1.017)
1.086(0.937,1.258)
1.036(0.994,1.080)
1.13(p>0.05)
1.108(1.007,1.219)
1.014 (p> 0.055)
1.023(0.897, 1.166)
1.021 (0.997, 1.045)
42.1 (±22.0)
39.9 (±18.0)
37.1 (±19.8)
13(2, 105)
61.7(0.78,390.5)
nmol/m3
17.28 (-0.6, 72.6)
18(6,86)
13.0(0,42)
22 (4, 86)
32.5(9.3, 190.1)
16.8(5,48)
22 (4, 86)
13(2,105)
61.7(0.78,390.5)
nmol/m3
17.28 (-0.6, 72.6)
18(6,86)
32.5(9.3, 190.1)
March 2001
9-48
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9-3 (cont'd). EFFECT ESTIMATES PER VARIABLE INCREMENTS IN
24-HOUR CONCENTRATIONS OF FINE PARTICLE INDICATORS (PM2 5, SO;, H+)
FROM U.S. AND CANADIAN STUDIES*
Study Location
COPD:
Los Angeles, CA1
Increased Hospitalization
Ontario, Canada0
Ontario, Canada*
NYC/Buffalo, NY5
Toronto, Canada5
Total Respiratory:
King County, WAT
Toronto, Canada0
Buffalo, NYD
Montreal, Canadav
Montreal, Canada*
St. John, Canada"
Pneumonia:
Detroit, MIF
Respiratory infections:
Toronto, Canadau
COPD:
Atlanta, GAZ
Detroit, MIF
King County WAM
Los Angeles, CABB
Toronto, CanadaY
Indicator
PM25
SO'4
SO",
03
SO"4
H+ (Nmol/m3)
SO°4
PM2S
PM,
PM25
so;
PM25
PM25
PM25
PM25
PM25
PM25
PM25
PM25
PM25
PM25
RR(±CI)**per25-^g/m3
PM Increase or 1 5-//g/m3
SOJ Increase or 75-nmol/m3 H+ increase
1.027(0.966,1.091)
1.03 (1.02, 1.04)
1.03 (1.02, 1.04)
1.03 (1.02, 1.05)
1.05 (1.01, 1.10)
1.16(1.03, 1.30)'
1.12 (1.00, 1.24)
1.15 (1.02, 1.78)
1.058(1.011, 1.110)
1.085(1.034,1.138)
1.082(1.042, 1.128)
1.261 (1.059, 1.503)
1.137(0.998, 1.266)
1.057(1.006, 1.110)
1.125(1.037,1.220)
1.108(1.072,1.145)
1.124(0.921, 1.372)
1.055(0.953,1.168)
1.064(1.009,1.121)
1.051 (1.009,1.094)
1.048(0.998,1.100)
Reported PM
Levels Mean (Min,
Max)***
22 (4, 86)
R = 3.1-8.2
R= 2.0-7.7
NR
28.8 (NR/391)
7. 6 (NR, 48.7)
18.6(NR, 66.0)
NR
16.8(1,66)
61.7(0.78,390.5)
nmol/m3
Summer 93
12.2 (max 31)
18.6(SD9.3)
Summer 93
8.5 (max 53.2)
18(6,86)
18.0 (max 90)
19.4 (±9.35)
18(6,86)
18.1 (3,96)
Median 22 (4, 86)
18.0 (max 90)
March 2001
9-49
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9-3 (cont'd). EFFECT ESTIMATES PER VARIABLE INCREMENTS IN
24-HOUR CONCENTRATIONS OF FINE PARTICLE INDICATORS (PMZ 5, SO=, IT)
FROM U.S. AND CANADIAN STUDIES*
Study Location
Asthma:
Atlanta, GAZ
Seattle, WACC
Seattle, WADD
Toronto, CanadaY
Total Cardiovascular:
Atlanta, GAY
Buffalo, NYD
Los Angeles, CAEE
St. John, Canada"
Toronto, Canada11
Ischemic Heart Disease:
Detroit, MIF
Toronto, CanadaY
Dvsrhvthmias:
Atlanta, GAZ
Detroit, MIF
Toronto, CanadaY
Heart Failure:
Detroit, MIF
Toronto, CanadaY
Cerebrovascular:
Los Angeles, CAEE
Toronto, CanadaY
Peripheral circulation diseases:
Toronto, CanadaY
Indicator
PM25
PM25
Est. PM25
PM25
PM25
so:
PM25
PM25
PM25
PM25
PM25
PM25
PM25
PM25
PM25
PM25
PM25
PM25
PM25
RR(±CI)**per25-^g/m3
PM Increase or 1 5-//g/m3
SC>4 Increase or 75-nmol/m3 H+ increase
1.023(0.852,1.227)
1.087(1.033,1.143)
1.445(1.217, 1.714)
1.064(1.025, 1.106)
1.061(0.969,1.162)
1.015(0.987, 1.043)
(65+) 1.043 (1.025, 1.061)
(<65) 1.035 (1.01 8, 1.053)
1.151 (1.006, 1.110))
1.059(1.018, 1.102)
1.043(0.986, 1.104)
1.080(1.054,1.108)
1.061 (0.874, 1.289)
1.032(0.934, 1.140)
1.061 (1.019, 1.104)
1.091 (1.023,1.162)
1.066(1.025,1.108)
1.015(0.992,1.038)
"NEG" reported
"NEC" reported
Reported PM
Levels Mean (Min,
Max)***
19.4 (±9.35)
16.7(6,32)
4.8(1.2,32.4)
18.0 (max 90)
19.4 (±9.35)
61.7(0.78,390.5)
nmol/m3
Median 22 (4, 86)
Summer 93
8.5 (max 53.2)
16.8(1,66)
18(6,86)
18.0 (max 90)
19.4 (±9.35)
18(6,86)
18.0 (max 90)
18(6,86)
18.0 (max 90)
Median 22 (4, 86)
18.0 (max 90)
1 8.0 (max 90)
March 2001
9-50
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9-3 (cont'd). EFFECT ESTIMATES PER VARIABLE INCREMENTS IN
24-HOUR CONCENTRATIONS OF FINE PARTICLE INDICATORS (PM2 5, SO;, H+)
FROM U.S. AND CANADIAN STUDIES*
Study Location Indicator
Stroke:
Detroit, MIF PM25
Increased Respiratory Symptoms
Southern CaliforniaFF SO°4
SixCitiesCG PM25
(Cough) SO"4
Six Cities00 PM25
(Lower Resp. Symp.) SO°j
Uniontown, PAHH PM2i
(Evening Cough)
Connecticut summer camp" SOJ
State College, PAJJ PM2 ,
(Wheeze)
State College, PAJJ PM2 ,
(Cough)
State College, PAW PM2 ,
(Cold)
Decreased Lung Function
Uniontown, PAHH PM25
Uniontown, PA1* PM25
(Reanalysis)
State College, PA1^ PM25
(Reanalysis)
Connecticut summer camp" SO^
Southwest, VALL PM25
State College, PAJJ PM2 ,
RR (± CI)** per 25-^g/m3
PM Increase or 1 5-,ug/m3
SO; Increase or 75-nmol/m3 H+ increase
1.018(0.947, 1.095)
Odd Ratio (95% CI) per 25-Mg/m3
PM Increase or 15-/ug/m3
SO; Increase or 75-nmol/m3 H+ increase
1.48(1.14, 1.91)
1.24 (1.00, 1.54)
1.86(0.86,4.03)
1.19(0.66,2.15)
1.58(1.18,2.10)
6.82(2.09, 17.35)
1.16(0.10, 13.73)
1.45(1.07, 1.97)
1.71 (1.30,2.25)
1.59(0.94,2.71)
1.61 (1.21,2.17)
1.2.45(1.29,4.64)
PEFR change (L/min) per 25-^g/m3
PM Increase or 1 5-yUg/m3
SO; Increase or 75-nmol/m3 H+ increase
PEFR -1.38 (-2.77, 0.02)
pm PEFR -1 .52, (-2.80, -0.24)
pm PEFR -0.93 (-1.88, 0.01)
PEFR -5.4 (-12.3, 1.52)
am PEFR -1.825 (-3.45, -0.21)
pm PEFR -0.63 (-1.73, 0.44)
Reported PM
Levels Mean (Min,
Max)***
18(6,86)
R = 2-37
18.0 (max 86.0)
2. 5 (max 15.1)
18.1 (max 37 1.1)
nmol/m3
18.0 (max 86.0)
2.5 (max 15.1)
18.1 (max 371.1)
nmol/m3
24. 5 (max 88.1)
7.0(1.1,26.7)
23.5 (max 85.8)
23.5 (max 85.8)
23. 5 (max 85. 8)
24.5 (max 88.1)
24.5 (max 88.1)
23.5 (max 85.8)
7.0(1.1,26.7)
21.62(3.48,59.65)
23.5 (max 85.8)
March 2001
9-51
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9-3 (cont'd). EFFECT ESTIMATES PER VARIABLE INCREMENTS IN
24-HOUR CONCENTRATIONS OF FINE PARTICLE INDICATORS (PM2 5, SO,, H+)
FROM U.S. AND CANADIAN STUDIES*
Study Location
Philadelphia, PA1™
RR(±CI)**per25-^g/m3
PM Increase or 1 5-/ug/m3
Indicator SO^ Increase or 75-nmol/m3 H+ increase
PM25 am PEFR -3. 18 (-6.64, 0.07)
pmPEFR-0.91 (-4.04,2.21)
Reported PM
Levels Mean (Min,
Max)***
22.2 (IQR 16.2)
* Studies highlighted in the 1996 CD are in italics; new studies in plain text.
** Relative Risk (95% Confidence Interval), except for Fairley (1999) and Lipfert et al. (2000), where insufficient
data were available to calculate confidence intervals so p-value is given in parentheses.
*** Min, Max 24-h PM indicator level shown in parentheses unless otherwise noted as (±S.D.), NR = not reported,
or R = range of values from min-max, no mean value reported.
References:
ASchwartz et al. (1996a)
BLaden et al. (2000)
cFairley(1999)
DGwynn et al. (2000)
ELipfert et al. (2000a)
FLippmann et al. (2000)
GMar et al. (2000)
"Smith et al. (2000)
'Moolgavkar (2000a)
JOstro(1995)
KOstro et al. (2000)
LSchwartz (2000a)
MTsai et al. (2000)
NBurnett et al. (2000)
°Burnett et al. (1998a)
pGoldberg et al. (2000)
QBurnettetal.(1994)
RBurnett et al. (1995)
sThurstonetal.(1992, 1994)
TLumley and Heagerty (1999)
"Burnett et al. (1997)
vDelfmo et al. (1997)
wDelfmo et al. (1998)
xStieb et al. (2000)
YBurnett et al. (1999)
zTolbert et al. (2000)
AAMoolgavkar et al. (2000)
BBMoolgavkar (2000b)
ccSheppard et al. (1999)
DDNorris et al. (1999)
EEMoolgavkar (2000c)
FFOstroetal.(1993)
GGSchwartzetal.(1994)
HHNeasetal.(1995)
"Thurston et al. (1997)
"Neasetal. (1996)
KKSchwartz and Neas (2000)
LLNaeher et al. (1999)
^Neasetal. (1999)
March 2001
9-52
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9-4. EFFECT ESTIMATES PER VARIABLE INCREMENTS IN 24-HOUR
CONCENTRATIONS OF COARSE-FRACTION PARTICLES (PM,0.2 5)
FROM U.S. AND CANADIAN STUDIES*
RR(±CI)**per25-^g/m3
Reported PM
Levels Mean (Min,
Study Location
Indicator
Increase
Max)***
Acute Mortality
Six Cities:4
Portage, Wl
Topeka, KS
Boston, MA
St. Louis, MO
Kingston/Knoxville, TN
Steubenville, OH
Overall Six-City Results
Coachella Valley, CAB
Detroit, MIC
Philadelphia, PAD
Phoenix, AZE
Phoenix, AZF
Santa Clara County, CAG
Eight Canadian Cities"
PM,0.2S
PMI0.25
PM!0.2S
P"flO-2 5
PMI0.2S
PM10.2S
PM,0_25
PM10.25
PMI0.25
PM10.25
PM,o.25
PM25
PM10.25
PM10.2,
7.073 (0.970, 1.058)
0.968 (0.920, 1.015)
1.005 (0.985, 1.030)
1.005 (0.983, 1.028)
1.025 (0.985, 1.066)
1.061 (1.013, 1.111)
1.004(0.999, 1.010)
1.013(0.994, 1.032)
1.040(0.988,1.094)
1. 052 (p> 0.055)
1.030(0.995,1.066)
(>25 ^g/m3) 1.185 (1.069, 1.314)
(<25/j.g/m3) 1.020(1.005, 1.035)
1.02(p>0.05))
1.018(0.992, 1.044)
6.6 (±6.8)
14.5 (±12.2)
8.8 (±7.0)
11.9 (±8.5)
11.2 (±7.4)
16.1 (±13.0)
Median 9.0
17.9(0, 149)
13 (4, 50)
6.80 (-20.0, 28.3)
33.5(5, 187)
NR
1 1 (0, 45)
12.9 (max 99)
Cause-Specific Mortality
Total Cardiovascular:
Coachella Valley, CAB
Detroit, MIC
Philadelphia, PAD
(seven-county area)
Phoenix, AZE
Santa Clara County, CAG
Total Respiratory:
Coachella Valley, CAB
Detroit, MI°
Philadelphia, PAD
(seven-county area)
Santa Clara County, CAG
PM10-25
PM10.25
PM.o-25
PM10.25
PM,o.25
PM10.25
PM10.25
PM,0.25
PMIB.,,
1.026(1.006, 1.045)
1.078(1.000, 1.162)
1. 034 (p> 0.055)
1.064(1.014, 1.117)
1.03(p>0.05)
1.026(1.006, 1.045)
1.074(0.910, 1.269)
1. 030 (p> 0.055)
1.16(p>0.05)
17.9(0, 149)
13(4,50)
6.80 (-20.0, 28.3)
33.5(5,187)
1 1 (0, 45)
17.9(0,149)
13(4,50)
6.80 (-20.0, 28.3)
1 1 (0, 45)
March 2001
9-53
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9-4 (cont'd). EFFECT ESTIMATES PER VARIABLE INCREMENTS IN
24-HOUR CONCENTRATIONS OF COARSE-FRACTION PARTICLES (PM1M5)
FROM U.S. AND CANADIAN STUDIES*
Study Location
Indicator
RR (±CI)** per 25-^g/m3
Increase
Reported PM
Levels Mean (Min,
Max)***
Increased Hospitalization
Total Respiratory:
Toronto, Canada1
Pneumonia:
Detroit, MIC
Respiratory infections:
Toronto, Canada1
COPD:
Atlanta, GAK
Detroit, MIC
Toronto, Canada1
Total Cardiovascular:
Atlanta, GAK
Toronto, Canada1
Ischemic Heart Disease:
Detroit, MIC
Toronto, Canada'
Dvsrhvthmias:
Detroit, MIC
Atlanta, GAK
Toronto, Canada1
Heart Failure:
Detroit, MIC
Toronto, Canada1
Stroke:
Detroit, MIC
Cerebrovascular:
Toronto, Canada1
Peripheral Circulation Diseases:
Toronto, Canada1
PM.o.25
PM,0.25
PM10.25
PM10.25
PMI0.25
PM10.25
PM10.25
PM10.25
PM10.25
PM10.25
PM10.25
PMI0.25
PM10.25
PM,0.25
PM10.25
PM10.25
PMI0.25
PM11L,,
1.125(1.052,1.20)
1.119(1.006, 1.244)
1.093(1.046,1.142)
0.770(0.493,1.202)
1.093(0.958, 1.247)
1.128(1.049,1.213)
1.176(0.954, 1.450)
1.205(1.082, 1.341)
1.105(1.027, 1.189
1.037(1.013, 1.062))
1.002(0.877, 1.144)
1.532(1.021,2.30)
1.051 (0.998, 1.108)
1.052(0.967, 1.144)
1.079(1.023,1.138)
1.049(0.953, 1.155)
"NEG" reported
1.056(1.003, 1.112)
11.6(1,56)
13(4,50)
12.2 (max 68)
9.39 (±4.52)
13(4,50)
12. 2 (max 68)
9.39 (±4.52)
11.6(1,56)
13(4,50)
12.2 (max 68)
13(4,50)
9.39 (±4.52)
12.2 (max 68)
13(4,50)
12. 2 (max 68)
13(4,50)
12.2 (max 68)
12.2 (max 68)
March 2001
9-54
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9-4 (cont'd). EFFECT ESTIMATES PER VARIABLE INCREMENTS IN
24-HOUR CONCENTRATIONS OF COARSE-FRACTION PARTICLES (PM10.2 5)
FROM U.S. AND CANADIAN STUDIES*
Study Location Indicator
Asthma:
Seattle, WAL PM,0.25
Toronto, CanadaJ PM,0.2,
Increased Respiratory Symptoms
Six U.S. CitiesM PM10.25
(Lower Respiratory
Symptoms)
Six U.S. Cities" PM10.25
(Cough)
Southwest VirginiaN PM , 0.2 5
(Runny or Stuffy Nose)
Decreased Lung Function
Southwest Virginia0 PM , 0.2 5
Uniontown, PAM PM,0.25
(Reanalysis)
RR(±CI)**per25-/wg/m3
Increase
1.111(1.028,1.201)
1.111(1.058,1.166)
Odds Ratio (95% CI) per 25-Mg/m3
PM Increase
1.51 (0.94,4.87)
1.77(1.24,2.55)
2.62(1.16,5.87)
PEFR change (L/min) per 25-^g/m3
PM Increase
am PEFR 5.3 (2.6, 8.0)
pm PEFR +1.73 (5.67, -2.2)
Reported PM
Levels Mean (Min,
Max)***
16.2(6,29)
12.2 (max 68)
NR
NR
NR
27.07 (4.89, 69.07)
NR
State College, PAM
(Reanalysis)
Philadelphia, PAP
PM,,
PM,
pm PEFR -0.28 (2.86, -3.45)
am PEFR-4.31 (-11.44,2.75)
NR
9.5(IQR5.1)
* Studies highlighted in the 1996 CD are in italics; new studies in plain text.
** Relative Risk (95% Confidence Interval), except for Fairley (1999) and Lipfert et al. (2000), where insufficient
data were available to calculate confidence intervals so p-value is given in parentheses.
*** Min, Max 24-h PM indicator level shown in parentheses unless otherwise noted as (±S.D.), NR = not reported,
or R = range of values from min-max, no mean value reported.
References:
ASchwartz et al., (1996a)
BOstro et al. (2000)
cLippmann et al (2000)
DLipfert et al (2000)
EMar et al. (2000)
FSmith et al. (2000)
°Fairley(1999)
"Burnett et al. (2000)
'Burnett et al. (1997)
JBurnett et al. (1999)
KTolbert et al. (2000)
LSheppardetal.(1999)
MSchwartz and Neas (2000)
NNaeheretal. (1999)
°Zhang et al. (2000)
•"Neasetal. (1999)
March 2001
9-55
DRAFT-DO NOT QUOTE OR CITE
-------
os
3.
O
O
ON
T1
H
6
o
2
o
H
O
G
O
H
W
O
*>
O
HH
H
CD
Percent excess death (total unless otherwise noted) per
25 ug/m3 increase in PM2.s (•) or PM10.2.5 (o).
_
Klemm etal (2000) _
Harvard 6 Cities (recomputed)
Burnett et al (2000) _
8 Canadian Cities
Chock et al (2000)
Pittsburgh, PA
Klemm and Mason (2000)
Atlanta, GA ~
Lipfert et al (2000) _
Philadelphia, PA ~
Lippman et al (2000)
Detroit, Ml ~
Mar et al (2000) _
Phoenix, AZ
Fairley(1999) _
Santa Clara Co
Ostroetal (2000)
Coachella Valley, CA ~
Castillejos et al (2000)
Mexico City, Mexico
Cifuentes et al (2000)
Santiago, Chile
5-4-3-2-10 1 2 34 56 7 8 9 10 11 12 13 14 1
1 I I 1 I 1 1 1 1 1 I 1 1 1 1 1 I 1 I 1
1
1
^ , .^_
^* }aye>75
1 «g 1 rtay ^
cardi - ubr
_— O— " mortal ity
•
Laq5dayMA> • "" " ^
"' Q }Allyear
Laq2dayMA> ^ }Wmtoi
1 V
1
Figure 9-7. Percent excess risks estimated per 25-fj.g/m3 increase in PM2 5 or PM10.2 5 from new studies evaluating both PM2 5
and PM10_2 5 data for multiple years. All lags = 1 day, unless indicated otherwise.
-------
1 associations with one or more of several specific fine particle components of PM, including H+,
2 sulfate, nitrate, as well as COH; but their relative importance varied from city to city, likely
3 depending, in part, on their concentrations (e.g., no clear associations in those cities where H+
4 and sulfate levels were very low [i.e., circa nondetection limits]). Figure 9-8 depicts relatively
5 consistent estimates of total mortality excess risk resulting from a 5-yUg/m3 increase in sulfate,
6 possibly reflecting impacts of sulfate per se or perhaps sulfate serving as a surrogate for fine
7 particles in general. Sulfate effect size estimates generally fall in the range of 1 to 4% excess
8 total mortality per 5-^g/m3 increase for U.S. and Canadian cities.
9
Percent excess death (total mortality, unless otherwise noted)
per 5 |jg/m3 increase in sulfate
Burnett etal. (1998a)
Toronto, Canada ~
Burnett et al. (2000)
8 Largest -
Canadian Cities
Fairley(1999)
Santa Clara, Co.
Gwynn et al. (2000)
Buffalo, NY
Klemm et al. (2000)
Atlanta, GA
Lipfert et al. (2000a)
Philadelphia, PA
Lippman et al. (2000)
Detroit, Ml
Tsai et al. (2000)
3 NJ Cities
-2 0 2 4 6 8 10
1 1 1 1 1 I
,
^ Elizabeth
Figure 9-8. Relative risks estimated per 5-^g/m3 increase in sulfate from U.S. and
Canadian studies in which both PM,S and PMin,* data were available.
March 2001
9-57
DRAFT-DO NOT QUOTE OR CITE
-------
1 A significant factor in some western cities is the occasional occurrence of high levels of
2 windblown crustal particles that constitute the major part of the coarse PM fraction and a
3 substantial fraction of intermodal fine particles (PM2 5.,). The small-size tail of the windblown
4 crustal particles extends into the PM2 5_, size range (intermodal), at times contributing
5 significantly to PM2 5. Claiborn et al. (2000) report that in Spokane, WA, PM2 5 constitutes about
6 30% of PM10 on dust event days, but 48% on days preceding the dust event. The intermodal
7 fraction represents about 51 % of PM2 5 during windblown dust events, about 28% on preceding
8 days. However, PM[ in Spokane often shows little change during dust events, when coarse
9 particles (presumably crustal particles) are transported into the region. The lack of increased
10 mortality during periods of time with high wind speeds and presumably high crustal material
11 concentrations was shown by Schwartz et al. (1999) for Spokane, and by Pope et al. (1999a) for
12 three cities in the Wasatch front region of Utah. Other recent studies suggest that coarse particles
13 also may be associated with excess mortality as well as fine particles in certain U.S. locations
14 e.g., in Phoenix, AZ (Smith et al., 2000; Clyde et al., 2000; Mar et al., 2000) the Coachella
15 Valley of California (Ostro et al., 2000), Mexico City (Castillejos et al., 2000) or Santiago, Chile
16 (Cifuentes et al., 2000). However, the coarse particle association with mortality may not be
17 caused by the crustal components. An important advantage of using source profiles for PM2 5 in
18 western cities is that it allows separation of crustal PM from accumulation-mode PM derived
19 from anthropogenic origins.
20 Several new studies highlighted in Chapter 6 conducted source-oriented evaluations of PM
21 components using factor analysis (see Table 9-5). The results of these studies (Laden et al.,
22 2000; Mar et al., 2000; Tsai et al., 2000; Ozkaynak et al., 1996) generally suggest that a number
23 of combustion-related source-types are associated with excess mortality risk, including: regional
24 sulfate; automobile emissions; coal combustion; oil burning; and vegetative (biomass) burning.
25 In contrast, the crustal factor from fine particles was generally not positively associated with total
26 mortality, with Mar et al. (2000) reporting a negative association between the crustal component
27 of PM2 5 and cardiovascular mortality.
28 However, these source-oriented evaluation results are derived from relatively limited
29 underlying analytic bases and must be viewed with caution at this time. For example, whereas
30 Laden et al. (2000) had 6211 days of every-other-day data from the Harvard Six City Study of
31 eastern/midwest U.S. cities, they had only elements in PM2 5 analyzed by X-ray fluorescence
March 2001 9-58 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9-5. SUMMARY OF SOURCE-ORIENTED EVALUATIONS OF
PARTICULATE MATTER COMPONENTS IN RECENT STUDIES
Author, City
Source Types Identified (or Suggested) and
Associated Tracers
Source Types Associated with Mortality.
Comments.
Laden et. al., (2000)
Harvard Six Cities
1979-1988
Mar et al. (2000).
Phoenix, AZ
1995-1997
Ozkaynak et al.
(1996).
Toronto, Canada.
Tsai et al. (2000).
Newark, Elizabeth,
and Camden, NJ.
1981-1983.
Soil and crustal material: Si
Motor vehicle emissions: Pb
Coal combustion: Se
Fuel oil combustion: V
Salt: Cl
Note: the trace elements are from PM2 5 samples
PM25 (from DFPSS) trace elements.
Motor vehicle emissions and resuspended road dust:
Mn, Fe, Zn, Pb, OC, EC, CO, and NO2
Soil: Al, Si, and Fe
Vegetative burning: OC and Ks (soil-corrected
potassium)
Local SO, sources: SO2
Regional sulfate: S
PM,I>.25 (from dichot) trace elements:
Soil: Al, Si, K, Ca, Mn, Fe, Sr, and Rb
A source of coarse fraction metals: Zn, Pb, and Cu
A marine influence: Cl
Motor vehicle emissions: CO, COH, and NO2
Motor vehicle emissions: Pb and CO
Geological (Soil): Mn and Fe
Oil burning: V and Ni
Industrial: Zn, Cu, and Cd (separately)
Sulfate/secondarv aerosol: Sulfate
Note: The trace elements are from PMI5 samples.
The strongest increase in daily mortality
was associated with the mobile source
factor. The coal combustion factor was
positively associated with mortality in all
metropolitan areas, with the exception of
Topeka. The crustal factor from the fine
particles was not associated with
mortality.
Coal and mobile sources account for the
majority of fine particles in each city.
PM^ < factors results: Soil factor and local
SO2 factor were negatively associated with
total mortality. Regional sulfate was
positively associated with total mortality
on the same day, but negatively associated
on the lag 3 day. Motor vehicle factor,
vegetative burning factor, and regional
sulfate factor were significantly positively
associated with cardiovascular mortality.
Factors from dichot PM10_25 trace elements
were not analyzed for their associations
with mortality because of the small sample
size (every-third-day samples from June
1996).
Motor vehicle factor was a significant
predictor for total, cancer, cardiovascular,
respiratory, and pneumonia deaths.
Oil burning, industry, secondary aerosol,
and motor vehicle factors were associated
with mortality.
1 (XRF) spectroscopy (no organic PM or gases) and they used Pb as a tracer to identify a motor
2 vehicle source, Se to identify a coal combustion source, and Si as a tracer for soil. The "motor
3 vehicle" and "coal combustion" sources were statistically significant for total mortality as well as
March 2001
9-59
DRAFT-DO NOT QUOTE OR CITE
-------
1 mortality resulting from ischemic heart disease and respiratory diseases (COPD plus pneumonia).
2 The crustal component had a negative association with total mortality.
3 The Mar et al. (2000) study had 3 years of pollutant data for Phoenix, AZ. In addition to
4 elements determined by XRF, they had pollutant gases (CO, NO2, SO2, and O3) and total,
5 organic, and elemental carbon. They were able to identify five sources. Motor vehicles (plus
6 resuspended road dust), vegetative burning, and regional sulfate all had statistically significant
7 associations with cardiovascular mortality, but soil (indexed by Si and Al, as crustal markers)
8 had a statistically significant negative association.
9 Tsai et al. (2000) had only 156 days of data and used measurements of CO, sulfate, and
10 some elements; and they did not have Si, Ca, Al, or Mg as soil tracers nor Se as a tracer of coal
11 combustion, although much of the sulfate probably came from coal combustion. They had three
12 fractions of extractable organic matter, but these did not appear to be useful in determining
13 source factors. Nevertheless, they were still able to identify motor vehicles, oil burning, and
14 sulfate as statistically significant (p > 0.05) factors for both total daily deaths and combined
15 cardiovascular and respiratory daily deaths in at least one or another of the three New Jersey
16 cities studied (Newark, Camden, and Elizabeth). Also, an industrial source containing Zn and Cd
17 was statistically significant for total deaths in Newark; and an industrial source containing Cd
18 was marginally statistically significant for cardiorespiratory disease in Elizabeth.
19 Ozkaynak et al. (1996) had only TSP, coefficient of haze (COH), and gases; however, they
20 reported that a factor with COH, CO, and NO2 (considered to be representative of motor vehicle
21 emissions) was associated with mortality in Toronto, Canada.
22 None of these studies had measurements of nitrate or semivolatile organic compounds nor
23 did they use the newest, and most effective, techniques for source apportionment. For example,
24 using positive matrix factorization, Ramadan et al. (2000) were able to determine eight factors
25 using the same data set as Mar et al. (2000). In spite of these deficiencies, all four studies were
26 able to associate one or more types of morality with motor vehicles, several with coal
27 combustion, and three with sulfate.
28 Factor analyses also were described briefly in a report by Lippmann et al. (2000). In that
29 study, neither sulfate nor acid aerosols were related significantly to morbidity or mortality, but
30 the concentrations were extremely low (with about 70% of the acid measurements below
31 detection limit).
March 2001 9-60 DRAFT-DO NOT QUOTE OR CITE
-------
1 It is difficult to compare these source-related assessments. They are based on different
2 regions of the country over different periods of time when the sources of particles and other
3 urban air pollutants were changing greatly. Furthermore, each of these studies constructed
4 factors based on city-specific data. Thus, the factors in each study are based on the
5 idiosyncrasies of the specific data set for each city in the study, so the factors may indeed
6 represent different sources in different locations. Nevertheless, although somewhat limited at
7 this time, the new factor analysis results appear to implicate ambient PM derived from fossil fuel
8 (oil, coal) combustion and vegetative burning, and secondarily formed sulfates as important
9 contributors to observed mortality effects, but not crustal particles.
10 In summary, there is evidence that exposure to particles from several different source
11 categories and, of different composition and size may have independent associations with health
12 outcomes. The excess risks from different types of combustion sources (coal, oil, gasoline,
13 wood, and vegetation) may vary from place to place and from time to time, so that substantial
14 intra-regional and inter-regional heterogeneity would be expected. Likewise, although earlier
15 evaluations in the 1996 PM AQCD seemed to indicate coarse particles and intermodal particles
16 of crustal composition as not likely being associated with adverse health effects, there are now
17 some reasonably credible studies suggesting that coarse particles (although not necessarily those
18 of crustal composition) may sometimes be as associated with excess mortality in at least some
19 locations.
20
21 9.6.2.2 Updated Epidemiologic Findings for Long-Term Particulate Matter Exposure
22 Effects on Mortality
23 The 1996 PM AQCD indicated that past epidemiologic studies of chronic PM exposures
24 collectively indicate increases in mortality to be associated with long-term exposure to airborne
25 particles of ambient origins (see appendix Table 9A-3). The PM effect size estimates for total
26 mortality from these studies also indicated that a substantial portion of these deaths reflected
27 cumulative PM impacts above and beyond those exerted by acute exposure events. Table 9-6
28 shows long-term exposure effects estimates (RR values) per variable increments in ambient PM
29 indicators in U.S. and Canadian cities, including results from newer analyses since the 1996 PM
30 AQCD.
March 2001 9-61 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9-6. EFFECT ESTIMATES PER INCREMENTSA IN LONG-TERM MEAN
LEVELS OF FINE AND INHALABLE PARTICLE INDICATORS FROM U.S. AND
CANADIAN STUDIES
Type of Health
Effect and Location
Increased Total Mortality in
Six City6
ACSStudyc
(151 U.S. SMSA)
Six City Reanalysis0
ACS Study Reanalysis0
Southern CahforniaE
Indicator
Adults
PM,5/IO(20^g/m3)
PM25(20/ug/m3)
SOr.OSvg/m')
PM25(20/ug/m3)
S0:(15^g/m3)
PMl5/10(20//g/m3)
PM25(20,ug/m3)
PM15/lo(20//g/m3)
(SSI)
PM25(20^g/m3)
PM10(50Mg/m3)
PMIO (cutoff =
30 days/year
PM10 (50Mg/m3)
PM,0 (cutoff =
30 days/year
>100,wg/m3)
Increased Bronchitis in Children
Six Cif/
Six City0
24 City"
24 City"
24 City"
24 City"
Southern California'
12 Southern California
communitiesj
(all children)
1 2 Southern California
communitieslc
(children with asthma)
PMlm(50^/m3)
TSP (100/ug/m3)
H+ (100 nmol/m3)
S0:(l5vg/m3)
PM2l(25/ug/m3)
PMlo(50/ug/m3)
S0=(15^g/m3)
PM,0(25Mg/m3)
Acid vapor (1 .7 ppb)
PM10(19,ug/m3)
Acid vapor (1.8 ppb)
Change in Health Indicator per
Increment in PMa
Relative Risk (95% CI)
1.18(1.06-1.32)
1.28(1.09-1.51)
1.46(1.16-2.16)
1.14(1 07-1.21)
1.10(1.06-1.16)
1.19(1.06-1.34)
1.28(1.09-1.51)
1.02(0.99-1.04)
1.14(1.08-1.21)
1.242 (0.955- 1.6 16) (males)
1.082 (1.008- 1.1 62) (males)
0.879 (0.713-1.085) (females)
0.958 (0.899-1.021) (females)
Odds Ratio (95% CI)
3.26(1.13, 10.28)
280(1.17, 7.03)
2.65 (1.22, 5.74)
3.02 (1.28, 7.03)
1.97 (0.85,4.51)
3.29(0.81, 13.62)
1.39(0.99, 1.92)
0.94(0.74, 1.19)
1.16(0.79, 1.68)
1.4(1.1, 1.8)
1.4(0.9,2.3)
1.1 (0.7, 1.6)
Range of City
PM Levels *
Means (//g/m3)
18-47
11-30
5-13
9-34
4-24
18.2-46.5
11.0-29.6
58.7(34-101)
9.0-33.4
51 (±17)
51 (±17)
20-59
39-114
6.2-41.0
18.1-67.3
9 1-173
22.0-28.6
—
28.0-84.9
0.9-3.2 ppb
13.0-70.7
6.7-31.5
1.0-5.0 ppb
March 2001
9-62
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9-6 (cont'd). EFFECT ESTIMATES PER INCREMENTSA IN LONG-TERM
MEAN LEVELS OF FINE AND INHALABLE PARTICLE INDICATORS FROM U.S.
AND CANADIAN STUDIES
Type of Health
Effect and Location
Increased Cough in Children
1 2 Southern California
communitiesj
(all children)
12 Southern California
communitiesK
(children with asthma)
Indicator
PM10 (25 ^g/m3)
Acid vapor (1 .7 ppb)
PM10(19^g/m3)
PM^dS^g/m3)
Acid vapor (1.8 ppb)
Change in Health Indicator per
Increment in PMa
Odds Ratio (95% CI)
1.06(0.93,1.21)
1.13(0.92,1.38)
1.1(0.0.8, 1.7)
1.3(0.7,2.4)
1.4(0.9,2.1)
Range of City
PM Levels *
Means C"g/m3)
28.0-84.9
0.9-3.2 ppb
13.0-70.7
6.7-31.5
1.0-5.0 ppb
Increased Obstruction in Adults
Southern California1"
Decreased Lung Function in
Six Cit/
Six City0
24 City"
24 City"
24 City"
24 City*1
12 Southern California
communities'^
(all children)
12 Southern California
communities1"1
(all children)
12 Southern California
communities0
(4th grade cohort)
12 Southern California
communities0
(4th grade cohort)
PM]0 (cutoff of
42 days/year
>100Aig/m3)
Children
PM15/lo(50^g/m3)
TSP(100jug/m3)
H+ (52 nmoles/m3)
PM2l(15^g/m3)
S0'4(7^g/m3)
PM10(17 Mg/m3)
PM10(25Mg/m3)
Acid vapor (1.7 ppb)
PM10 (25 ^g/m3)
Acid vapor (1 .7 ppb)
PM,0(51.5,ug/m3)
PM25(25.9/ug/m3)
PM10.25(25.6^g/m3)
Acid vapor (4.3 ppb)
PM,0(51.5Aig/m3)
PM25(25.9Mg/m3)
PM10.25(25.6//g/m3)
Acid vapor (4.3 ppb)
1.09(0.92, 1.30)
NS Changes
NS Changes
-3.45% (-4.87, -2.01) FVC
-3.21% (-4.98, -1. 41) FVC
-3.06% (-4.50, -1.60) FVC
-2.42% (-4.30, -.0.51) FVC
-24.9 (-47.2, -2.6) FVC
-24.9 (-65.08, 15.28) FVC
-32.0 (-58.9, -5.1) MMEF
-7.9 (-60.43, 44.63) MMEF
-0.58 (-1.14, -0.02) FVC growth
-0.47 (-0.94, 0.01) FVC growth
-0.57 (-1.20, 0.06) FVC growth
-0.57 (-1.06, -0.07) FVC growth
-1.32 (-2.43, -0.20) MMEF growth
-1.03 (-1.95, -0.09) MMEF growth
-1.37 (-2.57, -0.15) MMEF growth
-1.03 (-2.09, 0.05) MMEF growth
NR
20-59
39-114
6.2-41.0
18.1-67.3
9.1-17.3
22.0-28.6
28.0-84.9
0.9-3.2 ppb
28.0-84.9
0.9-3.2 ppb
NR
NR
March 2001
9-63
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9-6 (cont'd). EFFECT ESTIMATES PER INCREMENTSA IN LONG-TERM
MEAN LEVELS OF FINE AND INHALABLE PARTICLE INDICATORS FROM U.S.
AND CANADIAN STUDIES
Type of Health
Effect and Location
Indicator
Change in Health Indicator per
Increment in PMa
Range of City
PM Levels *
Means (/ug/m3)
Decreased Lung Function in Adults
Southern Californiap
(% predicted FEV,,
females)
PMIO (cutoff of
54.2 days/year
>100,ug/m3)
Southern California1" PM,0 (cutoff of
(% predicted FEV,, males) 54.2 days/year
>100,ug/m3)
Southern California1" PM]0 (cutoff of
(% predicted FEV,, males 54.2 days/year
whose parents had asthma, >100 ,ug/m3)
bronchitis, emphysema)
Southern Californiap SO^ (1.6 /ug/m3)
(% predicted FEV,,
females)
Southern California1" SO^ (1.6 Afg/m3)
(% predicted FEV,, males)
+0.9 % (-0.8, 2.5) FEV,
+0.3 % (-2.2, 2.8) FEV,
-7.2% (-11.5,-2.7) FEV,
Not reported
-1.5% (-2.9,-0.1) FEV,
52.7(21.3,80.6)
54.1 (20.0, 80.6)
54.1 (20.0,80.6)
7.4(2.7, 10.1)
7.3(2.0, 10.1)
*Range of mean PM levels given unless, as indicated, studies reported overall study mean (min, max), or mean
(±SD); NR=not reported.
AResults calculated using PM increment between the high and low levels in cities, or other PM increments given
in parentheses; NS Changes = No significant changes.
References:
"Dockery et al. (1993)
cPopeetal.(1995)
DKrewski et al. (2000)
EAbbeyetal. (1999)
FDockery et al. (1989a)
GWare etal.( 1986)
"Dockeryetal. (1996)
1 Abbey et al. (1995a,b,c)
JPetersetal. (1999b)
KMcConnell et al. (1999)
LBerglund et al. (1999)
MRaizenneetal.(1996)
NPetersetal. (1999a)
°Gauderman et al. (2000)
pAbbeyetal. (1998)
1 One of the most important advances since the 1996 PM AQCD is the substantial
2 verification and extension of the findings of the Six City prospective cohort study (Dockery
3 et al., 1993) and the cohort study relating American Cancer Society (ACS) health data to
4 fine-particle data from 50 cities and sulfate data from 151 cities (Pope et al., 1995). The
5 reanalyses, sponsored by the Health Effects Institute (HEI), included a data audit, replication of
March 2001
9-64
DRAFT-DO NOT QUOTE OR CITE
-------
1 the original investigators' findings, and additional analyses to explore the sensitivity of the
2 original findings to other model specifications. The investigators of the HEI Reanalysis Project
3 (Krewski et al., 2000) first performed a data audit, using random samples to verify the accuracy
4 of the data sets used in the original Six City analyses, including death certificate data, air
5 pollution data, and socioeconomic data. In general, the air pollution data were reproducible and
6 correlated highly with the original aerometric data in Pope et al. (1995).
7 The reanalyses substantially verified the findings of the original investigators, with PM2 5 or
8 sulfate relative risk (RR) estimates for total mortality and for cardiopulmonary mortality differing
9 at most by ±0.02 (±2% excess risk) from the least polluted to the most polluted cities in the
10 study. A larger difference was noted for the PM2 5 lung cancer relative risk in the Six Cities
11 study, 1.37 originally and 1.43 in the reanalysis, neither estimate being statistically significant.
12 The sensitivity analyses for the Six Cities study found generally similar results with other
13 individual covariates included. The time-dependent covariate model for total mortality (taking
14 into account higher postexposures in early years of the study and changes over time to the last
15 years of the study) had a substantially lower RR than the model without time-dependent
16 covariates. Educational level made a large difference, with individuals having less than a high
17 school education at much greater risk for mortality than those with any postsecondary education.
18 Among the ecological covariates, sulfates adjusted for artifact had little effect on the risk
19 estimates for total mortality compared to that without adjustment, but, in the ACS study, the filter
20 adjustment actually increased the relative risk for all causes and cardiopulmonary mortality,
21 while substantially reducing the estimated sulfate effect on lung cancer. Inclusion of SO2 as an
22 additional ecological covariate greatly reduced the estimated PM2 5 and sulfate effects in the ACS
23 study, whereas a spatial model including SO2 effects caused only a modest reduction of the
24 estimated PM2 5 and sulfate effects. However, the SO2 effects were reduced greatly when sulfates
25 were included in the model. Sulfur dioxide and sulfates often are highly correlated, because of
26 the formation of secondary sulfates.
27 Many model selection issues in the prospective cohort studies are analogous to those in the
28 time series analyses. One issue of particular concern is whether the exposure indices used in the
29 analyses adequately characterize the exposure of the participants in the study during the months
30 or years preceding death. This question is particularly conspicuous in regard to the Pope et al.
31 (1995) study, in which PM2 5 and sulfate data were collected in the 1979 to 1982 period from the
March 2001 9-65 DRAFT-DO NOT QUOTE OR CITE
-------
1 EPA AIRS database and the Inhalable Particle Network, largely preceding the collection of the
2 ACS cohort data by only a few years, and so possibly not adequately reflecting exposure to
3 presumably much higher PM concentrations occurring long before the cohort was recruited, nor
4 exposure to presumably lower concentrations during the study. This issue was raised in the 1996
5 PM AQCD. However, the Six Cities Study did have air pollution data and repeated survey data
6 over time, with PM2 5 and sulfate data measured every other day and sometimes daily, and so the
7 new investigators were able to use the information about time-dependent cumulative PM
8 concentrations during the course of the study. Changes in smoking status and body mass index
9 over the 10 to 12 years of the study had little effect on risk estimates, but taking into account the
10 decrease in particle concentrations from the earlier years to the later years reduced the effect size
11 estimate substantially, although it remained statistically significant. Nevertheless, overall, the
12 reanalyses of the ACS and Harvard Six-Cities studies (Krewski et al., 2000) "replicated the
13 original results, and tested those results against alternative risk models and analytic approaches
14 without substantively altering the original findings of an association between indicators of
15 particulate matter air pollution and mortality."
16 The shape of the relationship of concentration to mortality also was explored. Preliminary
17 findings suggest some possible nonlineariry, but further study is needed. Among the most
18 important new findings of the study are spatial relationships between mortality and air pollution,
19 discussed later below.
20 With regard to the role of various PM constituents in the PM-mortality association, past
21 cross-sectional studies generally have found that the fine particle component, as indicated either
22 by PM2 5 or sulfates, was the PM constituent most consistently associated with chronic PM
23 exposure-mortality. Although the relative measurement errors of the various PM constituents
24 must be further evaluated as a possible source of bias in these estimate comparisons, the Harvard
25 Six-Cities study and the latest reported AHSMOG prospective semi-individual study results
26 (Abbey, et al., 1999a,b,c; McConnell et al., 2000) studies are both indicative of the fine mass
27 components of PM likely being associated more strongly with the mortality effects of PM than
28 coarse PM components; and the ACS study, which only evaluated fine particle indicators, further
29 substantiates ambient fine particle effects.
30 Several other new studies report epidemiologic evidence indicating that: (a) PM exposure
31 early in pregnancy (during the first month) may be associated with slowed intrauterine growth
March 2001 9-66 DRAFT-DO NOT QUOTE OR CITE
-------
1 leading to low birth weight events (Dejmek et al., 1999); and (b) early postnatal PM exposures
2 may lead to increased infant mortality (Woodruff et al., 1997; Boback and Leon, 1999; Loomis
3 et al., 1999; Lipfert et al., 2000b).
4 Recent investigations of the public health implications of effect estimates for long-term PM
5 exposures also were reviewed in Chapter 6. Life table calculations by Brunekreef (1997) found
6 that relatively small differences in long-term exposure to airborne PM of ambient origin can have
7 substantial effects on life expectancy. For example, a calculation for the 1969 to 71 life table for
8 U.S. white males indicated that a chronic exposure increase of 10 yUg/m3 PM was associated with
9 a reduction of 1.31 years for the entire population's life expectancy at age 25. The new evidence
10 noted above of infant mortality associations with PM exposure suggests that life shortening in the
11 entire population from long-term PM exposure could well be significantly larger than estimated
12 by Brunekreef (1997).
13
14 9.6.2.3 Relationships of Ambient Participate Matter Concentrations to Morbidity
15 Outcomes
16 New epidemiology studies add greatly to the overall database relating morbidity outcomes
17 to ambient PM levels. These include much additional evidence for cardiovascular and
18 respiratory diseases being related to ambient PM. The newer epidemiology studies expand the
19 evidence on cardiovascular (CVD) disease and are discussed first below, followed by discussion
20 of respiratory disease effects with particular emphasis on newly enhanced evidence for
21 PM-asthma relationships.
22
23 9.6.2.3.1 Cardiovascular Effects of Ambient Paniculate Matter Exposures
24 About 75% of all U.S. deaths occur in persons at least 65 years old, and, of these, nearly
25 40% are for cardiac causes (nearly 45%, if deaths from cerebrovascular causes are also included).
26 Thus, if ambient PM exposure indeed produces increased total mortality in the elderly, it would
27 seem possible that cardiovascular (CVD) deaths may be involved.
28
29 Cardiovascular Hospital Admissions. Just two studies were available for review in the 1996
30 PM AQCD that provided data on acute cardiovascular morbidity outcomes (Schwartz and
31 Morris, 1995; Burnett et al., 1995). Both studies were of ecologic time series design using
March 2001 9-67 DRAFT-DO NOT QUOTE OR CITE
-------
1 standard statistical methods. Analyzing 4 years of data on the > 65-year-old Medicare population
2 in Detroit, MI, Schwartz and Morris (1995) reported significant associations between ischemic
3 heart disease admissions and PM,0, controlling for environmental covariates. Based on an
4 analysis of admissions data from 168 hospitals throughout Ontario, Canada, Burnett and
5 colleagues (1995) reported significant associations between particle sulfate concentrations, as
6 well as other air pollutants, and daily cardiovascular admissions. The relative risk because of
7 sulfate particles was slightly larger for respiratory than for cardiovascular hospital admissions.
8 The 1996 PM AQCD concluded on the basis of these studies that, "There is a suggestion of a
9 relationship to heart disease, but the results are based on only two studies and the estimated
10 effects are smaller than those for other endpoints." The PM AQCD went on to state that acute
11 impacts on CVD admissions had been demonstrated for elderly populations (i.e., >65), but that
12 insufficient data existed to assess relative impacts on younger populations.
13 Although the literature still remains relatively sparse, an important new body of data now
14 exists that both extends the available quantitative information on relationships between ambient
15 PM pollution and hospital CVD admissions, and that, more intriguingly, illuminates some of the
16 physiological changes that may occur on the mechanistic pathway leading from PM exposure to
17 adverse cardiac outcomes. Figure 9-9 depicts excess risk estimates derived from 10 studies of
18 acute PM10 exposure effects on CVD admissions in U.S. cities. Although new studies depicted
19 in Figure 9-9 have reported generally consistent associations between daily hospitalizations for
20 cardiovascular disease and measures of PM, the data not only implicate PM, but also CO and
21 NO2 as well, possibly because of covarying of PM and these other gaseous pollutants derived
22 from common emission sources (e.g., motor vehicles). Taken as a whole, this body of evidence
23 suggests that PM is likely an important risk factor for cardiovascular hospitalizations in the
24 United States.
25 For example, in the recently published NMMAPS 14-city analysis of daily CVD hospital
26 admissions in persons 65 and older in relation to PM10 (Samet et al., 2000a,b). The mean risk
27 estimate (for average 0-1 day lag) was a 8.5% increase in CVD admissions per 50 /ug/m3 PM,0
28 (95% CI: 1.0 to 33.0%). No relationship was observed between city-specific risk estimates and
29 measures of socioeconomic status, including percent living in poverty, percent non-white, and
30 percent with college educations. In another study, remarkably consistent PM10 associations with
31 cardiovascular admissions were observed across eight U.S. metropolitan areas, with a 25 /ug/m3
March 2001 9-68 DRAFT-DO NOT QUOTE OR CITE
-------
Sametetal (2000a,b) -
14US Cities
Schwartz (1999) -
8 US Counties
Moolgavkar (2000c) -
Maricopa, AZ
Moolgavkar (2000c) -
LA.CA
Moolgavkar (2000c) -
Cook County
Linnetal (2000) -
LA.CA
Schwartz (1997) -
Tucson.AZ
Tolbertetal (2000) -
Atlanta
Morris and Naumova (1998) -
Chicago
Lippmann et al.(2000) -
Total CVD
i
-15
Period 1 (MRS Data)
CHF
i » i
Period 2 (Suqersite Data)
-10 -50 5 10
Reconstructed Excess Risk Percentage
50 ug/m3 Increase in PM,0
Figure 9-9. Acute cardiovascular hospitalizations and PM exposure excess risk estimates
derived from selected U.S. PM,0 studies. CVD = cardiovascular disease and
CHF = congestive heart failure.
1 increase in PM10 associated with between 1.8 and 4.2 percent increases in admissions (Schwartz,
2 1999). Also, in a study of Los Angeles data from 1992-1995, PM10, CO, and NO2 were all
3 significantly associated with increased cardiovascular admission in single-pollutant models
4 among persons 30 and older (Linn et al., 2000). Moolgavkar (2000c) analyzed PM10, CO, NO2,
5 O3, and SO2 in relation to daily total cardiovascular (CVD) and total cerebrovascular admissions
6 for persons 65 and older from three urban counties (Cook, IL; Los Angeles, CA; Maricopa, AZ),
7 and found that, in univariate regressions, PM10 (and PM2 5 in LA) was associated with CVD
March 2001
9-69
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
admissions in Cook and LA counties but not in Maricopa county. On the other hand, in
two-pollutant models in Cook and LA counties, the PM risk estimates diminished and/or were
rendered nonsignificant.
The recent NMMAPS study of PM10 concentrations and hospital admissions by persons
65 and older in 14 U.S. cities provides particularly important findings of positive and significant
associations, even when concentrations are below 50 /ug/m3 (Samet et al., 2000a,b). As noted in
Table 9-7, this study indicates PM10 effects similar to other cities, but with narrower confidence
bands, because of its greater power derived by combining multiple cities in the same analysis.
This allows significant associations to be identified, despite the fact that many of the cities
considered have relatively small populations and that each of the 14 cities had mean PM10 below
50
TABLE 9-7. PERCENT INCREASE IN HOSPITAL ADMISSIONS PER 10-Aig/m3
INCREASE IN 24-HOUR PM,,, IN 14 U.S. CITIES
CVD
Constrained Lag Models
One-day mean3
Previous-day mean
Two-day meanb
PM10 <50 Mg.m3
(2-day mean)b
Quadratic distributed lag
Increase
(Fixed Effect
1.07
0.68
1.17
1.47
1.18
(95%
CI)
COPD
Increase
(95% CI)
Pneumonia
Increase
(95% CI)
Estimates)
(0.93,
1.22)
(0.54,0.81)
(1.01,
(1.18,
(0.96,
1.33)
1.76)
1.39)
1
1
1
2
2
.44
.46
.98
.63
.49
(1.00,
(1.03,
(1.49,
(1.71,
(1.78,
1.89)
1.88)
2.47)
3.55)
3.20)
1.57
1.
1.
2.
1
.31
.98
.84
.68
(1-27,
(1.03,
(1.65,
(2.21,
(1.25,
1.87)
1.58)
2.31)
3.48)
2.11)
Unconstrained Distributed Lag
Fixed effects estimate
Random effects estimate
1.19
1.07
(0.97,
(0.67,
1.41)
1.46)
2
2
.45
.88
(1.75,
(0.19,
3.17)
5.64)
1
2
.90
.07
(1.46,
(0.94,
2.34)
3.22)
aLag.
bMean of lag 0 and lag 1.
Source: Samet et al., 2000a,b.
March 2001
9-70
DRAFT-DO NOT QUOTE OR CITE
-------
1 Physiologic Measures of Cardiac Function. Several very recent studies by independent groups
2 of investigators have also reported longitudinal associations between ambient PM concentrations
3 and physiologic measures of cardiovascular function. These studies measure outcomes and most
4 covariates at the individual level, making it possible to draw conclusions regarding individual
5 risks, as well as to explore mechanistic hypotheses. For example, several studies recently have
6 reported temporal associations between PM exposures and various electrocardiogram (ECG)
7 measures of heart beat or rhythm in panels of elderly subjects. Reduced HR variability is a
8 predictor of increased cardiovascular morbidity and mortality risks. Three independent studies
9 reported decreases in HR variability associated with PM in elderly cohorts, although r-MSSD
10 (one measure of high-frequency HR variability) showed elevations with PM in one study.
11 Differences in methods used and results obtained across the studies argue for caution in drawing
12 any strong conclusions yet regarding PM effects from them, especially in light of the complex
13 intercorrelations that exist among measures of cardiac physiology, meteorology, and air pollution
14 (Dockery et al., 1999). Still, the new heart rhythm results, in general, comport well with other
15 findings of cardiovascular mortality and morbidity endpoints being associated with ambient PM.
16 Chapter 5 discusses available exposure studies of elderly subjects with CVD, such as the Sarnat
17 et al. (2000) Baltimore study. Less active groups tend to have lower exposure to nonambient PM
18 because of reduced personal activity. However, Williams et al. (2000a,b,c) report a very high
19 pooled correlation coefficient between PM2 5 personal exposure and outdoor concentrations.
20 These exposure studies tend to enhance the plausibility of panel study findings of impacts on HR
21 variability being caused by exposure to ambient-generated PM.
22
23 Changes in Blood Characteristics. Additional epidemiologic findings (Peters et al., 1997a)
24 also provide new evidence for ambient PM exposure effects on blood characteristics (e.g.,
25 increased c-reactive protein in blood) thought to be associated with increased risk of serious
26 cardiac outcomes (e.g., heart attacks).
27
28 • Key Conclusions Regarding PM-CVD Morbidity. Overall, the newly available studies of
29 PM-CVD relationships appear to support the following conclusions regarding several key
30 issues:
31
March 2001 9-71 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Temporal Patterns of Response. The evidence from recent time series studies of CVD
2 admissions suggests rather strongly that PM effects are likely maximal at lag 0, with some
3 carryover to lag 1.
4
5 • Physical and Chemical Attributes Related to Participate Matter Health Effects. The
6 characterization of ambient PM attributes associated with acute CVD is incomplete.
7 Insufficient data exist from the time series CVD hospital admissions literature or from the
8 emerging individual-level studies to provide clear guidance as to which PM attributes, defined
9 either on the basis of size or composition, determine potency. The epidemiologic studies
10 published to date have been constrained by the limited availability of multiple PM metrics.
11 Where multiple PM metrics exist, they often are of differential quality because of differences in
12 numbers of monitoring sites and in monitoring frequency. Until more extensive and consistent
13 data become available for epidemiologic research, the question of PM size and composition, as
14 they relate to acute CVD impacts, will remain open.
15
16 • Susceptible Subpopulations. Because they lack data on individual subject characteristics,
17 ecologic time series studies provide only limited information on susceptibility factors based on
18 stratified analyses. The relative impact of PM on cardiovascular (and respiratory) admissions
19 reported in ecologic time series studies is generally somewhat higher than those reported for
20 total admissions. This provides some limited support for the hypothesis that acute effects of
21 PM operate via cardiopulmonary pathways or that persons with preexisting cardiopulmonary
22 disease have greater susceptibility to PM, or both. Although there is some data from the
23 ecologic time series studies showing larger relative impacts of PM on cardiovascular
24 admissions in adults 65 and over as compared with younger populations, the differences are
25 neither striking nor consistent. Some individual-level studies of cardiophysiologic function
26 suggest that elderly persons with preexisting cardiopulmonary disease are susceptible to subtle
27 changes in heart rate variability (HRV) in association with PM exposures. However, because
28 younger and healthier populations have not yet been assessed, it is not possible to say at present
29 whether the elderly have clearly increased susceptibility compared to other groups, as indexed
30 by cardiac pathophysiological indices such as HRV.
31
March 2001 9-72 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Role of Other Environmental Factors. The ecologic time series morbidity studies published
2 since 1996 generally have controlled adequately for weather influences. Thus, it is unlikely that
3 residual confounding by weather accounts for the PM associations observed. With one possible
4 exception (Pope et al., 1999b), the roles of meteorological factors have not been analyzed
5 extensively as yet in the individual-level studies of cardiac physiologic function. Thus, the
6 possibility of confounding in such studies as yet cannot be discounted totally or readily.
7 Co-pollutants have been analyzed rather extensively in many of the recent time series studies of
8 hospital admissions and PM. In some studies, PM clearly carries an independent association
9 after controlling for gaseous co-pollutants. In others, the "PM effects" are reduced markedly
10 once co-pollutants are added to the model. Among the gaseous criteria pollutants, CO has
11 emerged as the most consistently associated with cardiovascular (CVD) hospitalizations. The
12 CO effects are generally robust in the multi-pollutant model, sometimes as much so as PM
13 effects. However, the typically low levels of ambient CO concentrations in most such studies
14 and minimal expected impacts on carboxyhemoglobin levels and consequent associated
15 hypoxic effects thought to underlie CO CVD effects complicate interpretation of the CO
16 findings and argue for the possibility that CO may be serving as a general surrogate for
17 combustion products (e.g., PM) in the ambient pollution mix. See the recently completed EPA
18 CO criteria document (U.S. Environmental Protection Agency, 2000a).
19
20 9.6.2.3.2 Respiratory Effects of Ambient Particulate Matter Exposures
21 The number of studies examining hospitalization and emergency department visits for
22 respiratory-related causes and other respiratory morbidity endpoints has increased markedly since
23 the 1996 PM AQCD. In addition to evaluating statistical relationships for PMIO, quite a few new
24 studies also evaluated other PM metrics. Those providing estimates of increased risk in U.S. and
25 Canadian cities for respiratory-related morbidity measures (hospitalizations, respiratory
26 symptoms, etc.) in relation to 24-h increments in ambient fine particles (PM2 5) or coarse fraction
27 (PM10_2 5) of inhalable thoracic particles are included in Tables 9-3 and 9-4, respectively.
28
29 Respiratory-Related Hospital Admission/Visits. PM hospital admissions/ visit studies that
30 evaluated excess risks in relation to PMIO measures are still quite informative. Maximum excess
31 risk estimates for PM10 associations with respiratory-related hospital admissions and visits in
March 2001 9-73 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
U.S. cities are shown in Figure 9-10. Nearly all the studies showed positive, statistically
significant relationships between ambient PM10 and increased risk for respiratory-related doctors'
visits and hospital admissions. Overall, the results substantiate well ambient PM10 impacts on
respiratory-related hospital admissions/visits. The excess risk estimates fall most consistently in
the range of 5 to 25.0% per 50 /^g/m3 PM10 increment, with those for asthma hospital admissions
and doctor's visits being higher than for COPD and pneumonia hospitalization. Other, more
limited, new evidence (not depicted in Figure 9-10) shows excess risk estimates for overall
respiratory-related or COPD hospital admissions falling in the range of 5 to 15.0% per 24-h
25 Mg/m3 increment in PM2 5 or PM10.2 5. Larger estimates are found for asthma admissions or
physician visits, ranging up to ca. 40 to 50% for children <18 yr old in one study.
Tolbert et al. (2000) Atlanta -
Norris et al (2000) Seattle -
Norns et al (2000) Spokane -
Norris etal (1999) Seattle -
Choudhury et al (1997) Anchorage -
Nauenberg and Basu (1999) LA.CA -
Sheppard etal (1999) Seattle -
Zonobetti et al. (2000) Chicago -
Sametelal (2000a,b) 14 US Cities -
Moolgavkar (2000b) Phoenix -
Moolgavkar (2000b) LA.CA -
Moolgavkar (2000b) Chicago -
Moolgavkar et af (2000a) King C -
Moolgavkar etal (1997) Minn-SP -
Moolgavkar et al (1997) Birm -
Chen et al (2000) Reno.NV -
Zanobetti et al (2000) Chicago -
Sametetal. (2000a,b) 14 US Cities -
-25
\
Asthma Visits
, ,
Asthma Hospital Admissions
, |
'
w
* '
.n COPD Hospital Admissions
l-»-H
»H
w Pneumonia Hospital Admissions
25 50 75 100
Excess Risk, %
125
150
Figure 9-10. Maximum excess risk in selected studies of U.S. cities relating PM,0 estimate
of exposure (50 /wg/m3) to respiratory-related hospital admissions and visits.
March 2001
9-74
DRAFT-DO NOT QUOTE OR CITE
-------
1 Of particular note in Figure 9-10, are the large effect size estimates now being reported for
2 asthma hospitalizations and visits. Very importantly, these hospital admission/visit studies and
3 other new studies on respiratory symptoms and lung function decrements in asthmatics are
4 emerging as possibly indicative of ambient PM likely being a notable contributor to exacerbation
5 of asthma. Additional evidence for PM-asthma effects is also emerging from panel studies of
6 lung function and respiratory symptoms, and these are discussed in more detail below.
7 New panel studies of lung function and respiratory symptoms in asthmatic subjects have
8 been conducted by more than 10 research teams in various locations world-wide. As a group, the
9 studies examine health outcome effects that are similar, such as pulmonary peak flow rate
10 (PEFR); and the studies typically characterize the clinical-symptomatic aspects in a sample of
11 mild to moderate asthmatics (mainly children aged 5 to 16 yrs) observed in their natural setting.
12 Their asthma typically is being treated to keep them symptom free (with "normal" pulmonary
13 function rates, and activity levels) and to prevent recurrent exacerbations of asthma. Severity of
14 their asthma is characterized by symptom, pulmonary function, and medication use and would be
15 classified to include mild intermittent to mild persistent asthma suffers (National Institutes of
16 Health, 1997). As a group, they may thusly differ from asthmatics examined in studies of
17 hospitalization or doctor visits for acute asthmatic episodes, who may have more severe asthma.
18 Most studies reported ambient PM10 results, but PM2 5 was examined in two studies. Other
19 ambient PM measures (BS and SO4) also were used. For these studies, mean PM10 levels range
20 from a low of 13 Aig/m3 in Finland to a high of 167 /ug/m3 in Mexico City. The Mexico City
21 level is over three times more than each of the other levels and is unique compared to the others.
22 Related 95% CI for these means or ranges show 1-day maximums above 100 Aig/m3 in four
23 studies, with two of these above 150 jUg/m3. Hence, these studies mainly evaluated different PM
24 metrics indexing PM concentrations in the range found in U.S. cities (see Chapter 3). All the
25 studies controlled for temperature, and several controlled for relative humidity.
26 Many panel studies are analyzed using a design that takes advantage of the repeated
27 measures on the same subject. Study subject number (N) varied from 12 to 164, with most
28 having N >50; and all gathered adequate subject-day data to provide sufficient power for their
29 analyses. Linear models often are used for lung function and logistic models for dichotomous
30 outcomes. Meteorological variables are used as covariates; and medication use is also sometimes
31 evaluated as a dependent variable or treated as an important potential confounder. However,
March 2001 9-75 DRAFT-DO NOT QUOTE OR CITE
-------
1 perhaps the most critical choice in the model is selection of the lag for the pollution variable.
2 Presenting lag periods with only the strongest associations introduces potential bias, because the
3 biological basis for lag structure may be related to effect. No biological bases for pertinent lag
4 periods are known, but some hypotheses can be proposed. Acute asthmatic reactions can occur
5 4 to 6 h after exposure and, thus, 0-day lag may be more appropriate than 1-day lags for that
6 acute reaction. Lag 1 may be more relevant for morning measurement of asthma outcome from
7 PM exposure the day before, and longer term lags (i.e., 2 to 5 days) may represent the outcome of
8 a more prolonged inflammatory mechanism; but too little information is now available to
9 predetermine appropriate lag(s).
10 Chapter 7 noted that people with asthma tend to have greater TB deposition then do healthy
11 people, but this data was not derived from the younger age group studied in most asthma panel
12 studies. The Peters et al. (1997b) study is unique for two reasons: (1) they studied the size
13 distribution of the particles in the range 0.01 to 2.5 ptm and (2) examined the number of particles.
14 They reported that asthma-related health effects of 5-day means of the number of ultrafme
15 particles were larger than those of the mass of the fine particles. In contrast, Pekkanen et al.
16 (1997) also examined a range of PM sizes, but PMIO was more consistently associated with PEF.
17 Delfino et al. (1998) is unique in that they report larger effects for 1- and 8-h maximum PM10
18 than for the 24-h mean.
19 The results for the asthma panels of the peak flow analysis consistently show small
20 decrements for both PM10 and PM2 5. The effects using 2- to 5-day lags averaged about the same
21 as did the 0 to 1 day lags. Stronger relationships often were found with ozone. The analyses
22 were not able to clearly separate co-pollutant effects. The effects on respiratory symptoms in
23 asthmatics also tended to be positive. Most studies showed increases in cough, phlegm,
24 difficulty breathing, and bronchodilator use. The only endpoint more strongly related to longer
25 lag times was bronchodilator use, which was observed in three studies. The peak flow
26 decrements and respiratory symptoms are indicators for asthma episodes.
27 For PMIO, nearly all of the point estimates showed decreases, but most were not statistically
28 significant, as shown in Figure 9-11 as an example of PEF outcomes. Lag 1 may be more
29 relevant for morning measurement of asthma outcome from the previous day. The figure
30 presents studies that provided this data. The results were consistent for both AM and PM peak
31 flow analyses. Similar results were found for the PM2 5 studies, although there were fewer
March 2001 9-76 DRAFT-DO NOT QUOTE OR CITE
-------
Romieuetal. (1996)
(Mexico)
Pekkannen et al. (1997)
(Finland)
Gielenetal. (1997)
(Netherlands)
Romieuetal. (1997) -
(Mexico)
-10 -505
Change in Pulmonary Function, L/min
Figure 9-11. Selected acute pulmonary function change studies of asthmatic children.
Effect of 50 Aig/m3 PM10 on morning Peak flow lagged 1 day.
1 studies. Several studies included PM2 5 and PM10 independently in their analyses of peak flow.
2 Of these, Gold et al. (1999), Naeher et al. (1999), Tiittanen et al. (1999), Pekkanen et al. (1997),
3 and Romieu et al. (1996) all found similar results for PM2 5 and PM10. The study of Peters et al.
4 (1997b) found slightly larger effects for PM2 5. The study of Schwartz and Neas (2000) found
5 larger effects for PM2 5 than for the coarse mode. Naeher et al. (1999) found that H+ was related
6 significantly to a decrease in morning PEF. Thus, there is no evidence here for a stronger effect
7 of PM2 5 when compared to PM10. Also, of studies that provided analyses that attempted to
8 separate out effects of PM10 and PM2 5 from other pollutants, Gold et al. (1999) studied possible
9 interactive effects of PM2 5 and ozone on PEF; they found independent effects of the two
10 pollutants, but the joint effect was slightly less than the sum of the independent effects.
11 The effects on respiratory symptoms in asthmatics also tended to be positive, although
12 much less consistent than the lung function effects. Most studies showed increases in cough,
13 phlegm, difficulty breathing, and bronchodilator use (although generally not statistically
March 2001
9-77
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
significant), as shown in Figure 9-12 for cough as an example. Three studies included both PMIO
and PM2 5 in their analyses. The studies of Peters et al. (1997c) and Tiittanen et al. (1999) found
comparable effects for the two measures. Only the Romieu et al. (1996) found slightly larger
effects for PM2 5. These studies also give no good evidence for a stronger effect of PM2 5 when
compared to PM,0.
Vedaletal. (1998)
(Canada)
Romieu etal. (1997)
(Mexico)
Gielen etal. (1997)
(Netherlands)
Peters etal. (1997c)
(Czech Republic)
M
01234567
Odds Ratios for Cough
Figure 9-12. Odds ratios for cough for a 50-/^g/m3 increase in PM,0 for selected asthmatic
children studies, with lag 0 with 95% CI.
8 The results of PM10 peak flow analyses for nonasthmatic populations were inconsistent.
9 Fewer studies reported results in the same manner as the asthmatic studies. Many of the point
10 estimates showed increases rather than decreases. PM2 5 studies found similar results. The
11 effects on respiratory symptoms in nonasthmatics were similar to those in asthmatics: most
12 studies showed that PM10 increases cough, phlegm, and difficulty breathing, but these increases
13 were generally not statistically significant. Schwartz and Neas (2000) found that PM10.2 5 coarse
March 2001
9-78
DRAFT-DO NOT QUOTE OR CITE
-------
1 particles were significantly related to cough. Tiittanen et al. (1999) found that 1-day lag of
2 PMI0.2 5 was related to morning PEF, but not evening PEF. Neas et al. (1999) found no coarse
3 mode effects of PEF in non-asthmatic subjects.
4
5 9.6.2.3.3 Long-Term Paniculate Matter Exposure Effects on Lung Function and Respiratory
6 Symptoms
1 In the 1996 PM AQCD, the available respiratory disease studies were limited in terms of
8 conclusions that could be drawn. At that time, three studies based on a similar type of
9 questionnaire administered at three different times as part of the Harvard Six-City and 24-City
10 Studies provided data on the relationship of chronic respiratory disease to PM. All three studies
11 suggest a chronic PM exposure effect on respiratory disease. The analysis of chronic cough,
12 chest illness, and bronchitis tended to be significantly positive for the earlier surveys described
13 by Ware et al. (1986) and Dockery et al. (1989). Using a design similar to the earlier one,
14 Dockery et al. (1996) expanded the analyses to include 24 communities in the United States and
15 Canada. Bronchitis was found to be higher (odds ratio = 1.66) in the community with highest
16 exposure of strongly acidic particles when compared with the least polluted community. Fine
17 PM sulfate was also associated with higher reporting of bronchitis (OR = 1.65, 95% CI 1.12,
18 2.42).
19 The studies by Ware et al. (1986), Dockery et al. (1989), and Neas et al. (1994) all had
20 good monitoring data and well-conducted standardized pulmonary function testing over many
21 years, but showed no effect on children of PM pollution indexed by TSP, PM15, PM2 5, or
22 sulfates. In contrast, the latest 24-city analyses reported by Raizenne et al. (1996) found
23 significant associations of effects on FEV, or FVC in U.S. and Canadian children with both
24 acidic particles and other PM indicators. Overall, the available studies provided limited evidence
25 suggestive of pulmonary lung function decrements being associated with chronic exposure to PM
26 indexed by various measures (TSP, PMIO, sulfates, etc.).
27 A number of studies have been published since 1996, which evaluate the effects of
28 long-term PM exposure on lung function and respiratory symptoms, as presented in Chapter 6.
29 The methodology in the long-term studies varies much more than the methodology in the short-
30 term studies. Some studies reported highly significant results (related to PM), whereas others
31 reported no significant results. Of particular note are several studies reporting associations
March 2001 9-79 DRAFT-DO NOT QUOTE OR CITE
-------
1 between long-term PM exposures (indexed by various measures) or changes in such exposures
2 over time and chronic bronchitis rates, consistent with bronchitis results from the Dockery et al.
3 (1996) study noted above.
4 Unfortunately, the cross-sectional studies often are potentially confounded, in part, by
5 unexplained differences in geographic regions; and it is difficult to separate out results consistent
6 with a PM gradient from any other pollutants or factors having the same gradient. The studies
7 that looked for a time trend also are confounded by other conditions that changed over time. The
8 most credible cross-sectional study remains that described by Dockery et al. (1996) and Raizenne
9 et al. (1996). Whereas most studies include two to six communities, this study included 24
10 communities and is considered to provide the most credible estimates of long-term PM exposure
11 effects on lung function and respiratory symptoms.
12
13 9.6.2.4 Methodological Issues
14 Chapter 6 discussed several still important methodological issues related to assessment of
15 the overall PM epidemiologic database. These include, especially, issues related to model
16 specifications and consequent adequacy of control for potentially confounding of PM effects by
17 co-pollutants, evaluations of possible source relationships to pollutant effects that may be useful
18 in sorting out better effects attributable to PM versus other co-pollutants or both, and other issues
19 such as lag structure. Key points are discussed concisely below.
20
21 9.6.2.4.2 Time Series Studies: Confounding by Co-Pollutants in Individual Cities
22 The co-pollutant issue was discussed at length in the 1996 document and still remains an
23 important issue. It must be recognized that there are large differences in concentrations of
24 measured gaseous co-pollutants (and presumably unmeasured pollutants as well) in different
25 parts of the United States, as well as the rest of the world; and the concentrations are often
26 correlated with concentrations of PM and its components because of commonality in source
27 emissions, wind speed and direction, atmospheric processes, and other human activities and
28 meteorological conditions. Large sources in the United States include motor vehicle emissions
29 (gasoline combustion, diesel fuel combustion, evaporation, particles generated by tire wear, etc.),
30 coal combustion, fuel oil combustion, industrial processes, residential wood burning, solid waste
31 combustion, and so on. Thus, one might reasonably expect some large correlations among PM
March 2001 9-80 DRAFT-DO NOT QUOTE OR CITE
-------
1 and co-pollutants, but possibly with substantial differences in relation by season in different
2 cities or regions. Statistical theory suggests that PM and co-pollutant effect size estimates will be
3 highly unstable and often insignificant in multi-pollutant models when collinearity exists. Many
4 recent studies demonstrate this effect, for both hospital admissions (Moolgavkar, 2000b) and
5 mortality (Moolgavkar, 2000a; Chock et al., 2000). Because the problem seems largely insoluble
6 in studies in single cities, the new multi-city studies (Samet et al., 2000a,b; Schwartz, 1999;
7 Schwartz and Zanobetti, 2000) have provided important new insights. See discussions of
8 NMMAPS analysis in Chapter 6 and below for discussion of issues related to control for co-
9 pollutant effects. Overall, although such issues may warrant further evaluation, it now appears
10 unlikely that such confounding accounts for the vast array of effects attributed to ambient PM
11 based on the rapidly expanding PM epidemiology database.
12 Numerous new studies have reported associations not only between PM, but also gaseous
13 pollutants (O3, SO2, NO2, and CO), and mortality. In many of these studies, simultaneous
14 inclusion of one or more gaseous pollutants in regression models did not markedly affect PM
15 effect size estimates, as was generally the case in the NMMAPS analyses for 90 cities (see
16 Figure 9-13). On the other hand, some studies reporting positive and statistically significant
17 effects for gaseous copollutants (e.g., O3, NO2, SO2, CO) found varying degrees of robustness of
18 their effects estimates or those of PM in multipollutant models. Thus, although it is likely that
19 there are independent health effects of PM and gaseous pollutants, there is not yet sufficient
20 evidence by which to confidently separate out fully the relative contributions of PM versus those
21 of other gaseous pollutants or by which to quantitate modifications of PM effects by other co-
22 pollutants, including possible synergistic interactions that may vary seasonally or from location
23 to location. Overall, it appears, however, that ambient PM and O3 can be most clearly separated
24 out as likely having independent effects, their concentrations often not being highly correlated.
25 More difficulty is encountered, at times, in sorting out whether NO2, CO, or SO2 are exerting
26 independent effects in cities where they tend to be highly correlated with ambient PM
27 concentrations, possibly because of derivation of important PM constituents from the same
28 source (e.g., NO2, CO, PM from mobile sources) or a gaseous pollutant (e.g., SO2) serving as a
29 precursor for a significant PM component (e.g., sulfate).
March 2001 9-81 DRAFT-DO NOT QUOTE OR CITE
-------
PM10
PM10
PM10 + O3 + N02
PM10 + O3+S02
PM10
0.0 0.2
% Change in Mortality per 10 gg/m3 Increase in PM
10
Figure 9-13. Marginal posterior distributions for effect of PM10 on total mortality at lag 1,
with and without control for other pollutants, for the 90 cities. The numbers
in the upper right legend are the posterior probabilities that the overall
effects are greater than 0.
Source: Samet et al. (2000a,b).
1 9.6.2.4.3 Time Series Studies: Model Selection for Lags, Moving Averages, and Distributed
2 Lags
3 A number of different approaches have been used to evaluate the temporal dependence of
4 mortality or morbidity on time-lagged PM concentrations, including unweighted moving
5 averages of PM concentrations over one or more days, general weighted moving averages, and
6 polynomial distributed moving averages. Unless there are nearly complete daily data, each
7 different lag will be using a different set of mortality data corresponding to spaced PM
8 measurement; for example, for lag 0 with every-sixth-day PM measurements, the mortality data
March 2001
9-82
DRAFT-DO NOT QUOTE OR CITE
-------
1 are on the same day as the PM data, for lag 1 the mortality data are on the next day after the PM
2 data, and so on. Although this effect is likely to be small, it should nonetheless be kept in mind.
3 The issue of dealing with lag structure, which may not necessarily be the same for all cities
4 or for all regions, can be illustrated by NMMAPS findings. As shown in Table 9-8, the rank
5 ordering of effects by lag days differs somewhat among NMMAPS regions. The combined data
6 set suggests that lag 1 provides the best fit, but with some regional differences. This raises the
7 question as to whether a single lag model should be assumed to characterize a diverse set of
8 regional findings. Because the particle constituents, co-pollutants, susceptible subpopulations,
9 and meteorological covariates are likely to differ substantially from one region to another, the
10 timing of the largest mortality effects also may be presumed to differ in at least some cases. This
11 undoubtedly contributes to the variance of the estimated effects.
12
13
TABLE 9-8. PERCENT INCREASE IN MORTALITY PER 10 yug/m3 PM10 IN SEVEN
U.S. REGIONS (from Figure 23 in NMMAPS II)
Region Rank Order of Effects by Lags
Northwest lag 0 < lag 1 = lag 2
Southwest lag 0 < lag 1 < lag 2
Southern California lag 0 < lag 1, lag 1 > lag 2, lag 0 < lag 2
Upper Midwest lag 0 > lag 1, lag 0 > lag 2, lag 1 < lag 2
Industrial Midwest lag 0 < lag 1, lag 1 > lag 2
Northeast lag 0 < lag 1, lag 1 » lag 2
Southeast lag 0 « lag 1, lag 1 > lag 2
Combined lag 0 < lag 1, lag 1 > lag 2
14 The distributed lag models used in the NMMAPS II morbidity studies are a noteworthy
15 methodological advance. The fitted distributed lag models showed significant heterogeneity
16 across cities for COPD and pneumonia, however (see Table 15 therein), again raising the
March 2001 9-83 DRAFT-DO NOT QUOTE OR CITE
-------
1 question of how heterogeneous effects can best be combined so as not to obscure potentially real
2 city-specific or region-specific differences.
3 Only three cities with nearly complete daily PM10 data were used to evaluate more general
4 multi-day lag models (Chicago, Minneapolis/St. Paul, Pittsburgh), and these show somewhat
5 different patterns of effect, with lag 0 < lag 1 and lag 1 » lag 2 for Chicago, lag 0 = lag 1 > lag 2
6 for Minneapolis, and lag 0 < lag 1 = lag 2 for Pittsburgh. The 7-day distributed lag model is
7 significant for Pittsburgh, but less so in the other cities. The remaining data are limited
8 intrinsically in what they can reveal about temporal structure.
9
10 9.6.2.4.4 Time Series Studies: Model Selection for Concentration-Response Functions
11 Given the number of analyses that needed to be performed, it is not surprising that most of
12 the NMMAPS studies focused on linear concentration-response models. More recent studies
13 (Daniels et al., 2000) for the 20 largest U.S. cities have found posterior mean effects of 2 to 2.7%
14 excess risk of total daily mortality per 50 /ug/m3 24-h PM10 at lags 0, 1, 0+1 days; 2.4 to 3.5%
15 excess risk of cardiovascular and respiratory mortality; and 1.2 to 1.7% for other causes of
16 mortality. The posterior 95% credible regions are all significantly greater than 0. However, the
17 threshold models gave distinctly different estimates of 95% credible regions for the threshold for
18 total mortality (15 /ug/m3 at lag 1, range 10 to 20), cardiovascular and respiratory mortality
19 (15 //g/m3 at lag 0+1, range 0 to 20), and other causes of mortality (65 /ug/m3 at lag 0+1, range
20 50 to 75 /^g/ni3).
21 Another problem is that the shape of the relationship between mortality and PM10 may
22 depend, to some extent, on the associations of PM10 with gaseous co-pollutants. The association
23 is not necessarily linear, and is indeed likely to have both seasonal and secular components that
24 depend on the city location. Thus, further elaborations of these models may be desirable.
25
26
27 9.6.2.4.5 Effects of Exposure Error in Daily Time Series Epidemiology
28 There has been considerable controversy over how to deal with the nonambient component
29 of personal exposure. Recent biostatistical analyses of exposure error have indicated that the
30 nonambient component will not bias the statistically calculated risk in community time-series
31 epidemiology, provided that the nonambient component of personal exposure is independent of
March 2001 9-84 DRAFT-DO NOT QUOTE OR CITE
-------
1 the ambient concentration. Consideration of the random nature of nonambient sources and recent
2 studies, in which estimates of a, ambient-generated PM divided by ambient PM concentrations,
3 have been used to estimate separately the ambient-generated and nonambient components of
4 personal exposure, support the assumption that the nonambient exposure is independent of the
5 ambient concentration. Therefore, it is reasonable to conclude that community time series
6 epidemiology describes statistical associations between health effects and exposure to ambient-
7 generated PM, but does not provide any information on possible health effects resulting from
8 exposure to nonambient PM (e.g., indoor-generated PM).
9 From the point of view of exposure error, it is also significant to note that, although
10 ambient concentrations of a number of gaseous pollutants (O3, NO2, SO2) often are found to be
11 highly correlated with various PM parameters, personal exposures to these gases are not
12 correlated highly with personal exposure to PM indicators. The correlations of the ambient
13 concentrations of these gases also are not correlated highly with the personal exposure to these
14 gases. Therefore, when significant statistical associations are found between these gases and
15 health effects, it could be that these gases may, at times, be serving as surrogates for PM rather
16 than being causal themselves. Pertinent information on CO has not been reported.
17 The attenuation factor, a, is a useful variable. For relatively constant a, the risk because of
18 a personal exposure to 10 //g/m3 of ambient PM is equal I/a times the risk from a concentration
19 of 10 yWg/m3 of ambient PM, where a varies from a low of 0.1 to 0.2 to a maximum of 1.0. (The
20 health risk for an interquartile change in ambient concentration of PM is the same as that for an
21 interquartile change in exposure to ambient PM). Differences in a among cities, reflecting
22 differences in air-exchange rates (e.g., because of variation in seasonal temperatures and in extent
23 of use of air conditioners) and differences in indoor/outdoor time ratios, may, in part, account for
24 any differences in risk estimates based on statical associations between ambient concentrations
25 and health effects for different cities or regions. If a were 0.3 in city A, but 0.6 in city B, and the
26 risks for an increase in personal exposure of 10 //g/m3 were identical, then a regression of health
27 effects on ambient concentrations would yield a health risk for city B that would be twice that
28 obtained for city A.
29 A number of exposure analysts have discussed the PM exposure paradox (i.e., that
30 epidemiology yields statistically significant associations between ambient concentrations and
31 health effects even though there is a near zero correlation between ambient concentrations and
March 2001 9-85 DRAFT-DO NOT QUOTE OR CITE
-------
1 personal exposure in many studies). Several explanations have been advanced to resolve this
2 paradox. First, personal exposure contains both an ambient-generated and a nonambient
3 component. Community time series epidemiology yields information only on the ambient-
4 generated component of exposure. Therefore, the appropriate correlation to investigate is the
5 correlation between ambient concentration and personal exposure to ambient-generated PM, not
6 between ambient concentrations and total personal exposure (i.e., the sum of ambient-generated
7 and nonambient PM). Second, biostatistical analysis of exposure error indicates that if the risk
8 function is linear in the PM indicator, the average of the sum of the individual risks (risk function
9 times individual exposure) may be replaced by the risk function times the community average
10 exposure. Thus, the appropriate correlation (of ambient concentrations and ambient-generated
11 exposure) is not the pooled correlation of different days and different people but the correlation
12 between the daily ambient concentrations and the community average daily personal exposure to
13 ambient-generated PM. Because the nonambient component is not a function of the ambient
14 concentration, its average will tend to be similar each day. Therefore, the correlation coefficient
15 will depend on a but not on the nonambient exposure. These types of correlation yield high
16 correlation coefficients.
17 A few studies have conducted simulation analyses of effects of measurement errors on the
18 estimated PM mortality effects. These studies suggest that ambient PM excess risk effects are
19 more likely underestimated than overestimated, and that spurious PM effects (i.e., qualitative
20 bias such as change in the sign of the coefficient) because of transferring of effects from other
21 covariates require extreme conditions and are therefore very unlikely. The error because the
22 difference between the average personal exposure and the ambient concentration is likely the
23 major source of bias in the estimated relative risk. One study also suggested that apparent linear
24 exposure-response curves are unlikely to be artifacts of measurement error.
25 In conclusion, for time-series epidemiology, ambient concentration is a useful surrogate for
26 personal exposure to ambient-generated PM, although the risk per unit ambient PM
27 concentration is biased low by the factor a compared to the risk per unit exposure to ambient-
28 generated PM. Epidemiologic studies of statistical associations between long-term effects and
29 long term ambient concentrations compare health outcome rates across cities with different
30 ambient concentrations. Ordinarily, PM exposure measurement errors are not expected to
31 influence the interpretation of findings from either the community time-series or long-term
March 2001 9-86 DRAFT-DO NOT QUOTE OR CITE
-------
1 epidemiologic studies that have used ambient concentration data if they include sufficient
2 adjustments for seasonality and key personal and geographic confounders. When individual level
3 health outcomes are measured in small cohorts, to reduce exposure misclassification errors, it is
4 essential that better real-time exposure monitoring techniques be used and that further speciation
5 of indoor-generated, ambient, and personal PM mass be accomplished. This should enable
6 measurement (or estimation) of both ambient and nonambient components of personal exposure
7 and evaluation of the extent to which personal exposure to ambient-generated PM, personal
8 exposure to nonambient PM, or total personal exposure (to ambient-generated plus nonambient
9 PM) contribute to observed health effects.
10
11 9.6.3 Coherence of Reported Epidemiologic Findings
12 Interrelationships Between Health Endpoints. Considerable coherence exists across
13 newly available epidemiologic study findings. For example, it was earlier noted that effects
14 estimates for total (nonaccidental) mortality generally fall in the range of 2.5 to 5.0% excess
15 deaths per 50 /ug/m3 24-h PM10 increment. These estimates comport well with those found for
16 cause-specific cardiovascular- and respiratory-related mortality. Furthermore, larger effect sizes
17 for cardiovascular (in the range of 3 to 6% per 50 /^g/m3 24-h PM10 increment) and respiratory (in
18 the range of 5 to 25% per 50 Aig/m3 24-h PM,0) hospital admissions and visits are found, as
19 would be expected versus those for PM10-related mortality. Also, several independent panel
20 studies, evaluating temporal associations between PM exposures and measures of heart beat
21 rhythm in elderly subjects, provide generally consistent indications of decreased heart rate (HR)
22 variability being associated with ambient PM exposure (decreased HR variability being an
23 indicator of increased risk for serious cardiovascular outcomes, e.g., heart attacks). Other studies
24 point toward changes in blood characteristics (e.g., increased C-reactive protein levels) related to
25 increased risk of ischemic heart disease as also being associated with ambient PM exposures.
26
27 Spatial Interrelationships. Both the NMMAPS and Cohort Reanalyses studies had a
28 sufficiently large number of cities to allow considerable resolution of regional PM effects within
29 the "lower 48" states, but this approach was taken much farther in the Cohort Reanalysis studies
30 than in NMMAPS. There were 88 cities with PM10 effect size estimates in NMMAPS; 50 cities
31 with PM2 5 and 151 cities with sulfates in Pope et al. (1995) and in the reanalyses using the
March 2001 9-87 DRAFT-DO NOT QUOTE OR CITE
-------
1 original data; and, in the additional analyses by the cohort study reanalysis team, 63 cities with
2 PM25 data and 144 cities with sulfate data. The relatively large number of data points allowed
3 estimation of surfaces for elevated long-term concentrations of PM2 5, sulfates, and SO2 with
4 resolution on a scale of a few tens to hundreds of kilometers. Information drawn from the maps
5 presented in Figures 16-21 in Krewski et al. (2000) is summarized below.
6 The patterns are similar, but not identical. In particular, the modeled PM2 5 surface
7 (Krewski, Figure 18) has peak levels in the industrial midwest, including the Chicago and
8 Cleveland areas, the upper Ohio River Valley, and around Birmingham, AL. Lower, but
9 elevated, PM2 5 is found almost everywhere else east of the Mississippi, as well as in southern
10 California. This is rather similar to the modeled sulfate surface (Krewski, Figure 16), with the
11 absence of a peak in Birmingham and an emerging sulfate peak in Atlanta. The only region with
12 elevated SO2 concentrations is the Cleveland-Pittsburgh area. A preliminary evaluation is that
13 secondary sulfates in particles derived from local SO2 is more likely to be important in the
14 industrial midwest, south from the Chicago-Gary region and along the upper Ohio River region.
15 This intriguing pattern may be related to the combustion of high-sulfur fuels in the subject areas.
16 The overlay of mortality and air pollution is also of interest. The spatial overlay of long-
17 term PM2 5 and mortality (Krewksi, Figure 21) is highest for the upper Ohio River region, but
18 also includes a significant association over most of the industrial midwest from Illinois to the
19 eastern noncoastal parts of North Carolina, Virginia, Pennsylvania, and New York. This is
20 reflected, in diminished form, by the sulfates map (Krewski, Figure 19) where the peak sulfate-
21 mortality associations occur somewhat east of the peak PM2 5-mortality associations. The SO2
22 map (Krewski, Figure 20) shows peak associations similar to, but slightly east of, the peak
23 sulfate associations. This suggests that, although SO2 may be an important precursor of sulfates
24 in this region, there may be other considerations (e.g., metals) in the association between PM2 5
25 and long-term mortality, embracing a wide area of the midwest and northeast (especially
26 noncoastal areas).
27 It should be noticed that, although a variety of spatial modeling approaches were discussed
28 in the NMMAPS methodology report (NMMAPS Part I, pp. 66-71), the primary spatial analyses
29 in the 90-city study (NMMAPS, Part II) were based on a simpler seven-region breakdown of the
30 contiguous 48 states. The 20-city results reported for the spatial model in NMMAPS I show a
31 much smaller posterior probability of a PM10 excess risk of short-term mortality, with a spatial
March 2001 9-88 DRAFT-DO NOT QUOTE OR CITE
-------
1 posterior probability versus a nonspatial probability of a PM10 effect of 0.89 versus 0.98 at lag 0,
2 of 0.92 versus 0.99 at lag 1, and of 0.85 versus 0.97 at lag 2. The evidence that PM10 is
3 associated with an excess short-term mortality risk is still moderately strong with a spatial model,
4 but much less strong than with a nonspatial model. In view of the sensitivity of the strength of
5 evidence to the spatial model, the model assumptions warrant additional study. Even so, there is
6 a considerable degree of coherence between the long-term and short-term mortality findings of
7 the studies, with stronger evidence of a modest but significant short-term PM10 effect and a larger
8 long-term fine particle (PM2 5 or sulfate) effect in the industrial midwest. The short-term effects
9 are larger but less certain in southern California and the northeast, whereas the long-term effects
10 seem less certain there. Possible differences should be further explored.
11
12 9.6.4 Toxicologic Insights on Biological Plausibility
13 Toxicological studies can play an integral role in answering key questions regarding
14 biological plausibility of health effects associated with ambient PM. The materials presented
15 below focus on the progress that toxicological studies have made towards answering the
16 following two key questions.
17 (1) What are the potential mechanisms by which PM causes health effects?
18 (2) What specific component or components of ambient PM cause health effects?
19
20 9.6.4.1 Mechanisms of Action
21 Various studies using particulate matter having diverse physicochemical characteristics
22 have shown that these characteristics have a great impact on the specific response that is
23 observed. Thus, there may, in fact, be multiple biological mechanisms responsible for observed
24 morbidity/mortality because of exposure to ambient PM, and these mechanisms may be highly
25 dependent on the type of particle in the exposure atmosphere. However, it should be noted that
26 many controlled exposure studies used concentrations of PM that were much higher than those
27 occurring in ambient air. Thus, some of the effects elicited may not occur with exposure to lower
28 levels. Clearly, controlled exposure studies as yet have not been able to unequivocally determine
29 the particle characteristics and the toxicological mechanisms by which ambient PM may affect
30 biological systems. There is growing toxicological and epidemiological evidence that both the
31 cardiovascular and respiratory systems are affected by ambient PM. Nonetheless, understanding
March 2001 9-89 DRAFT-DO NOT QUOTE OR CITE
-------
1 how participate air pollution causes and exacerbates cardiovascular or respiratory diseases
2 remains an important goal. The pathophysiological mechanisms involved in PM-associated
3 cardiovascular and respiratory health effects remain unclear, but progress has been made since
4 the 1996 PM AQCD was written. This section summarizes current hypotheses and reviews the
5 toxicological evidence for potential pathophysiological mechanisms.
6
7 9.6.4.1.1 Direct Respiratory System Effects
8 Emerging new toxicological evidence for three key mechanisms hypothesized as underlying
9 direct effects of PM on the respiratory system is summarized below.
10
11 Lung Injury and Inflammation. In the last few years, numerous studies have shown that
12 instilled and inhaled ROFA, a product of fossil fuel combustion, can cause substantial lung injury
13 and inflammation. The toxic effects of ROFA largely result from its high content of soluble
14 metals, and the pulmonary effects of ROFA can be reproduced by equivalent exposures to
15 soluble metal salts. In contrast, controlled exposures of animals to sulfuric acid aerosols, acid
16 coated carbon, and sulfate salts cause little lung injury or inflammation even at high
17 concentrations. Inhalation of concentrated ambient PM (which contains only small amounts of
18 metals) by laboratory animals at concentrations in the range of 100 to 1000 /^g/m3 have been
19 shown in some (but not all) studies to cause mild pulmonary injury and inflammation. Rats with
20 SO2-induced bronchitis and monocrotaline-treated rats have a greater inflammatory response to
21 concentrated ambient PM than healthy rats. These studies suggest that exacerbation of
22 respiratory disease by ambient PM may be caused, in part, by lung injury and inflammation.
23
24 Increased Susceptibility to Respiratory Infections. There are no published studies on the effects
25 of inhaled concentrated ambient PM on host susceptibility to infectious agents. In vivo exposure
26 of mice to acid-coated carbon particles at a mass concentration of 10,000 ^g/m3 causes decreased
27 phagocytic activity of alveolar macrophages even in the absence of lung injury (Ohtsuka et al.,
28 2000). More studies are needed on the effects of concentrated ambient PM on the pulmonary
29 immune defense system.
30
March 2001 9-90 DRAFT-DO NOT QUOTE OR CITE
-------
1 Increased Airway Reactivity and Asthma Exacerbation. The strongest toxicologic evidence
2 supporting this hypothesis is from studies on diesel particulate matter (DPM). Diesel particulate
3 matter has been shown to increase production of antigen-specific IgE in mice and humans
4 (summarized in Section 8.2.4.3). In vitro studies have suggested that the organic fraction of
5 DPM is involved in the increased IgE production. The ROFA leachate also enhances antigen-
6 specific airway reactivity in mice (Goldsmith et al., 1999), indicating that soluble metals also can
7 enhance an allergic response. However, in this same study, exposure of mice to concentrated
8 ambient PM did not affect antigen-specific airway reactivity. It is premature to conclude from
9 this one experiment that concentrated ambient PM does not exacerbate allergic airways disease
10 because the chemical composition of the PM (as indicated by studies with DPM and ROFA) may
11 be more important than the mass concentration.
12
13 9.6.4.1.2 Systemic Effects Secondary to Lung Injury
14 When the 1996 PM AQCD was written, it was thought that cardiovascular-related
15 morbidity and mortality most likely would be secondary to impairment of oxygenation or some
16 other consequence of lung injury and inflammation. There is some toxicological evidence for the
17 following mechanisms for adverse systemic effects secondary to lung injury.
18
19 Impairment of Heart Function by Lowering Blood Oxygen Levels and Increasing the Work of
20 Breathing. Instillation of ROFA has been shown to cause a 50% mortality rate in
21 monocrotaline- treated rats (Watkinson et al., 2000). Although blood oxygen levels were not
22 measured in this study, there were ECG abnormalities consistent with severe hypoxemia in about
23 half of the rats that subsequently died. Given the severe inflammatory effects of instilled ROFA
24 and the fact that monocrotaline-treated rats have increased lung permeability as well as
25 pulmonary hypertension, it is plausible that instilled ROFA can cause severe hypoxemia leading
26 to death in this rat model. However, results from studies in which animals (normal and
27 compromised) were exposed to concentrated ambient PM (at concentrations many times higher
28 than would be encountered in the United States) indicate that ambient PM is unlikely to cause
29 severe disturbances in blood oxygenation or pulmonary function. However, even a modest
30 decrease in oxygenation can have serious consequences in individuals with ischemic heart
31 disease. For example, reducing arterial blood saturation from 98 to 94% by either mild hypoxia
March 2001 9-91 DRAFT-DO NOT QUOTE OR CITE
-------
1 or by exposure to 100 ppm CO significantly reduced the time to onset of angina in exercising
2 volunteers (Kleinman et al., 1998). Thus, more information is needed on the effects of PM on
3 arterial blood gases and pulmonary function to fully address the above hypothesis.
4
5 Lung Inflammation and Cytokine Production Leading to Systemic Hemodynamic Effects.
6 It has been suggested that systemic effects of particulate air pollution may be caused by
7 activation of cytokine production in the lung (Li et al., 1997). In support of this idea,
8 monocrotaline-treated rats exposed to inhaled ROFA showed increased pulmonary cytokine gene
9 expression, bradycardia, hypothermia, and increased arrhythmias (Watkinson et al., 2000).
10 However, spontaneously hypertensive rats had a similar cardiovascular response to inhaled
11 ROFA (except they also developed ST segment depression) with no increase in pulmonary
12 cytokine gene expression. Studies in dogs exposed to concentrated ambient PM showed minimal
13 pulmonary inflammation and no positive staining for IL-8, IL-1, or TNF in airway biopsies.
14 However, the time of onset of ischemic ECG changes following coronary artery occlusion
15 decreased significantly (Godleski et al., 2000). Thus, there is not a clear-cut link between
16 changes in cardiovascular function and production of cytokines in the lung. Because human and
17 animal exposure studies of ambient PM are using increasingly sophisticated and sensitive
18 measures of cardiac function, basic information on the effects of mild pulmonary injury on these
19 cardiac endpoints is needed to understand the mechanisms by which inhaled PM may affect the
20 heart.
21
22 Increased Risk of Heart Attacks and Strokes Because of Increasing Blood Coagulability
23 Secondary to Lung Inflammation. There is abundant evidence linking risk of heart attacks and
24 strokes to small prothrombotic changes in the blood coagulation system; and some new
25 epidemiologic evidence (discussed earlier above) indicates that ambient PM may affect blood
26 coagulation and/or other blood characteristics related to increased risk of serious cardiac
27 outcomes. However, there is no published experimental evidence as of yet that moderate lung
28 inflammation increases blood coagulability; a high dose (8,300 Mg/kg) of instilled ROFA did
29 cause increased levels of fibrinogen, but no effect was seen at lower doses (Gardner et al., 2000).
30 Also, exposure of dogs to concentrated ambient PM also had no effect on fibrinogen levels
31 (Godleski et al., 2000). The coagulation system is as multifaceted and complex as the immune
March 2001 9-92 DRAFT-DO NOT QUOTE OR CITE
-------
1 system, and there are many other sensitive and clinically significant parameters that should be
2 examined in addition to fibrinogen. Thus, it is premature to draw any strong conclusions about
3 ambient PM exposure effects on cardiovascular morbidity or mortality being mediated via PM
4 effects on blood coagulability or other blood characteristics.
5
6 Particulate Matter and Lung Interactions Potentially Affecting Hematopoiesis. Instillation of
7 fine carbon particles (20,000 /wg/rabbit) stimulated release of PMNs from the bone marrow
8 (Terashima et al., 1997). In support of this hypothesis, Gordon and colleagues reported that the
9 percentage of PMNs in the peripheral blood increased in rats exposed to ambient PM in some but
10 not all exposures. On the other hand, Godleski et al. (2000) found no changes in peripheral
11 blood counts of dogs exposed to concentrated ambient PM. Thus, direct evidence that ambient
12 concentrations of PM can affect hematopoiesis is still needed.
13
14 9.6.4.1.3 Direct Effects on the Heart
15 Changes in heart rate and heart rate variability associated with ambient PM exposure have
16 been reported in animal studies (Godleski et al., 2000; Gordon et al., 2000), in several human
17 panel studies (described in Chapter 6), and in a reanalysis of data from the MONICA study
18 (Peters et al., 2000). Some of these studies included endpoints related to respiratory effects, but
19 few significant adverse respiratory changes were detected. This raises the possibility that
20 ambient PM may have effects on the heart that are independent of adverse changes in the lung.
21 There is precedent for this idea: tobacco smoke (a mixture of combustion-generated gases and
22 particles) causes cardiovascular disease by mechanisms independent of its lung effects.
23
24 Heart Rate Variability. Epidemiological studies have linked fine particulate air pollution with
25 cardiopulmonary morbidity and mortality (Schwartz and Morris, 1995; Burnett et al., 1995;
26 Morris et al., 1995; Schwartz, 1997), but the underlying biologic mechanisms remain unclear.
27 Recently, attention has focused on possible effects on heart rate (HR) variability as a potential
28 mechanism underlying cardiovascular morbidity and mortality effects associated with ambient
29 PM. During recent decades, a large clinical database has developed describing a significant
30 relationship between autonomic dysfunction and sudden cardiac death. Moreover, low HR
31 variability has been implicated as a marker for a number of pathophysiological conditions
March 2001 9-93 DRAFT-DO NOT QUOTE OR CITE
-------
1 including myocardial infarction (Task Force of the European Society of Cardiology and the
2 North American Society of Pacing and Electrophysiology 1996; Bigger et al., 1992; Hayano
3 et al., 1990; Kleiger et al., 1987; Martin et al., 1987; Singer et al., 1988). This is further
4 elaborated in Appendix 6-B.
5 Some studies (Liao et al., 1999; Pope et al., 1999b) provide new evidence for relationships
6 between ambient PM and decreased HR variability. Pope et al. (1999b) reported an association
7 between particulate air pollution, heart rate, and HR variability. A relationship between PM and
8 HR variability also is supported by laboratory animal studies. Combustion particles instilled into
9 rat lungs produce arrhythmias and a doubling of mortality (Watkinson et al., 1998).
10 Concentrated ambient air particles breathed by dogs elicited electrocardiographic changes,
11 including T-wave alterans and arrhythmias (Godleski et al., 1998).
12
13 Autonomic Control of the Heart and Cardiovascular System. There is growing evidence for
14 the idea that inhaled particles could affect the heart through the autonomic nervous system.
15 Activation of neural receptors in the lung is a logical area to investigate. Studies in conscious
16 rats have shown that inhalation of wood smoke causes marked changes in sympathetic and
17 parasympathetic input to the cardiovascular system that are mediated by neural reflexes
18 (Nakamura and Hayashida, 1992). Although research on airway neural receptors and neural -
19 mediated reflexes is a well-established discipline, the cardiovascular effects of stimulating airway
20 receptors continue to receive less attention than the pulmonary effects. Previous studies of
21 airway reflex-mediated cardiac effects usually have employed very high doses of chemical
22 irritants, and the results may not be applicable to air pollutants. There is a need for basic
23 physiological studies to examine cardiovascular system effects when airway and alveolar neural
24 receptors are stimulated in a manner relevant to air pollutants.
25
26 Uptake of Particles and Distribution of Soluble Substances into the Systemic Circulation.
27 Drugs can be delivered rapidly and efficiently to the systemic circulation by inhalation (as occurs
28 with nicotine from inhaled cigarette smoke). This implies that the pulmonary vasculature absorbs
29 inhaled materials, including charged substances such as small proteins and peptides. It is likely
30 that soluble materials absorbed onto airborne PM find their way into the bloodstream, but it is
March 2001 9-94 DRAFT-DO NOT QUOTE OR CITE
-------
1 not clear whether the particulate materials themselves enter the blood. It is anticipated that more
2 information will be available on this important question in the next few years.
3
4 9.6.4.2 Links Between Specific Particulate Matter Components and Health Effects
5 Key to enhancing confidence in the biological plausibility of ambient PM health effects is
6 the need to identify those components of airborne PM responsible for the health effects and for
7 placing susceptible individuals at risk. The plausibility of epidemiologically demonstrated
8 associations between ambient PM and increases in morbidity and mortality has been questioned
9 because associations with health effects have been observed at very low PM concentrations.
10 To date, toxicology studies on PM have provided only limited evidence for specific PM
11 components being likely responsible for cardiovascular or respiratory effects of ambient PM.
12 The latest available experimental information concerning potential contributions of individual
13 physical and chemical factors of particles to cardiorespiratory effects is summarized below.
14
15 Acid Aerosols. There is relatively little new information on the effects of acid aerosols, and the
16 basic conclusions of the 1996 PM AQCD remain unchanged. It previously was concluded that
17 acid aerosols cause little or no change in pulmonary function in healthy subjects, but asthmatics
18 may experience small decrements in pulmonary function. These conclusions are further
19 supported by a recent study by Linn and colleagues (1997), in which healthy children (and
20 children with allergy or asthma) were exposed to sulfuric acid aerosol (100 /ug/m3) for 4 h. There
21 were no significant effects on symptoms or pulmonary function when the entire group was
22 analyzed, but the allergy group had a significant increase in symptoms after the acid aerosol
23 exposure (albeit to distinctly higher than typical ambient acid concentrations).
24 Although pulmonary effects of acid aerosols have been the subject of extensive research,
25 the cardiovascular effects of acid aerosols have received much less attention. However,
26 inhalation of acetic acid fumes has been reported to cause reflex mediated increases in blood
27 pressure in normal and spontaneously hypertensive rats (Zhang et al., 1997). Thus, acid
28 components should not be ruled out as possible mediators of PM health effects. In particular, the
29 cardiovascular effects of acid aerosols (at realistic concentrations) need further investigation.
30
March 2001 9-95 DRAFT-DO NOT QUOTE OR CITE
-------
1 Metals. The previous 1996 PM AQCD mainly relied on data related to occupational exposures
2 to evaluate the potential toxicity of metals in particulate air pollution. Since that time, in vivo
3 and in vitro studies using ROFA or soluble transition metals have contributed substantial new
4 information on the health effects of particle-associated soluble metals. Although there are some
5 uncertainties about differential effects of one transition metal versus another, water soluble
6 metals leached from ROFA, albeit at high concentrations, consistently have been shown to cause
7 cell injury and inflammatory changes in vitro and in vivo.
8 Even though it is clear that combustion particles that have a high content of soluble metals
9 can cause lung injury and even death in compromised animals, it has not been established that the
10 small quantities of metals associated with relatively low concentrations of ambient particles are
11 sufficient to cause health effects. In studies in which various ambient and emission source
12 particulates were instilled into rats, the soluble metal content did appear to be the primary
13 determinant of lung injury. However, one published study compared the effects of inhaled
14 ROFA (at 1 mg/m3) to concentrated ambient PM (of 475 to 900 Aig/m3) in normal and
15 SO2-induced bronchitic rats. A statistically significant increase in at least one lung injury marker
16 was seen in bronchitic rats in only one out of four of the concentrated ambient PM exposures,
17 and inhaled ROFA had no effect, even though the content of soluble iron, vanadium, and nickel
18 was much higher in the ROFA sample. Thus, the potential roles of metals in contributing to
19 health effects of ambient PM remains to be more clearly established. There has been increasing
20 attention focused in recent years on the possibility of ultrafme particles playing a major role in
21 observed ambient PM health effects due to large absolute number counts and/or surface area of
22 ultrafine particles deposited in the lung.
23
24 Ultrafine Particles. When this subject was reviewed in the 1996 PM AQCD, it was not known
25 whether the pulmonary toxicity of freshly generated ultrafine Teflon particles was because of
26 particle size or a result of absorbed fumes. Subsequent studies with other types of ultrafine
27 particles have shown that the chemical constituents of ultrafmes substantially modulate their
28 toxicity. Inhalation of MgO particles, for example, produces far fewer respiratory effects than
29 does ZnO (Kuschner et al., 1997). Also, inhalation exposure of normal rats to ultrafine carbon
30 particles generated by electric arc discharge caused minimal lung inflammation (Elder et al.,
31 2000) compared to ultrafine Teflon or metal particles. On the other hand, instillation of ultrafine
March 2001 9-96 DRAFT-DO NOT QUOTE OR CITE
-------
1 carbon black caused substantially more inflammation than did the same dose of fine particles of
2 carbon black, suggesting that ultrafme particles cause more inflammation than larger particles (Li
3 et al., 1997). However, the chemical constituents of the two carbon black sizes were not
4 analyzed and it is uncertain that the chemical composition was the same. As with acid aerosols,
5 studies of ultrafme particles have focused largely on effects in the lung, but it is possible that
6 inhaled ultrafme particles may have systemic effects that are independent of lung effects. It is
7 also important to note that at least one very recent new epidemiology study (Wichmann et al.,
8 2000) provides interesting new evidence implicating both ultrafine (nuclei-mode) and
9 accumulation-mode fine particles in PM-mortality relationships.
10
11 Diesel Exhaust Particulate Matter. As described in Section 8.2.4.2, there is growing
12 toxicological evidence that diesel exhaust particulate matter (DPM) exacerbates the allergic
13 response to inhaled antigens. The organic fraction of diesel exhaust has been linked to
14 eosinophil degranulation and induction of cytokine production suggesting that the organic
15 constituents of DPM are responsible for the immune effects. It is not known whether the
16 adjuvant-like activity of DPM is unique or whether other combustion-related particles have
17 similar effects. It is important to compare the immune effects of other source-specific emissions,
18 as well as concentrated ambient PM, to DPM to determine the extent to which exposure to diesel
19 exhaust may contribute to the incidence and severity of allergic rhinitis and asthma. Other types
20 of noncancer and carcinogenic (especially lung cancer) effects are of concern with regard to
21 DPM exposures, as discussed in a separate EPA Health Assessment Document for Diesel
22 Exhaust (U.S. Environmental Protection Agency, 2000b).
23
24 Organic Compounds. Published research on the acute effects of particle-associated organic
25 carbon constituents is conspicuous by its relative absence, except for diesel exhaust particles.
26 Like metals, organics are common constituents of combustion-generated particles and are found
27 in ambient PM samples over a wide geographical range. Organic carbon constituents comprise a
28 substantial portion of the mass of ambient PM (10 to 60% of the total dry mass [Turpin, 1999]).
29 The organic fraction of ambient PM has been evaluated for its mutagenic effects. Although the
30 organic fraction of particulate matter is a poorly characterized heterogeneous mixture of a widely
March 2001 9-97 DRAFT-DO NOT QUOTE OR CITE
-------
1 varying number of different compounds, strategies have been proposed for examining the health
2 effects of potentially important organic constituents (Turpin, 1999).
3
4 Bioaerosols. Recent studies support the conclusion of the 1996 PM AQCD that bioaerosols, at
5 the concentrations present in the ambient environment, do not likely account for the health
6 effects of ambient PM. Dose response studies in healthy volunteers exposed to 0.55 and 50 /u.g
1 endotoxin, by inhalation, showed the threshold for pulmonary and systemic effects for endotoxin
8 to be between 0.5 and 5.0 yUg (Michel et al., 1997). Available information suggests that ambient
9 concentrations of endotoxin are very low and do not exceed 0.5 ng/m3. Also, Monn and Becker
10 (1999) found cytokine induction by human monocytes, characteristic of endotoxin activity, in the
11 coarse size fraction of outdoor PM but not in the fine fraction.
12
13 Concentrated Ambient Particle Studies (CAPS). Ambient particle studies are potentially among
14 the most relevant in improving our understanding of the susceptibility of individuals to PM and
15 underlying mechanisms of toxicity. New studies have used collected urban PM for intratracheal
16 administration to healthy and compromised animals, and some recent work with inhaled
17 concentrated ambient PM has reported cardiopulmonary changes in rodents and dogs at high
18 concentrations of fine PM. Thus, despite difficulties in extrapolating from the bolus delivery
19 used in such studies, they are contributing some new evidence enhancing the plausibility of
20 health effects of fine particles observed in epidemiologic studies.
21
22 Animal Models of Susceptibility. Progress has been made in understanding the role of
23 individual susceptibility to ambient PM effects. Studies have shown consistently that animals
24 with compromised health, either genetic or induced, are more susceptible to instilled or inhaled
25 particles, although the increased animal-to-animal variability in these models has created
26 problems. Moreover, because PM seems to affect broad categories of disease states ranging from
27 altered cardiac rhythms to pulmonary infection, it can be difficult to know what disease models
28 to use in understanding the biological plausibility of the adverse health effects of PM. Thus, the
29 identification of susceptible animal models has been slow, but, overall, it represents solid
30 progress when one considers the numbers of people necessary in epidemiology studies to develop
31 the statistical power to detect small increases in morbidity and mortality.
March 2001 9-98 DRAFT-DO NOT QUOTE OR CITE
-------
1 9.7 RISK FACTORS AND POTENTIALLY SUSCEPTIBLE POPULATION
2 GROUPS
3 9.7.1 Introduction
4 The 1996 PM AQCD identified several population groups as likely being at increased risk
5 for experiencing health impacts of ambient PM exposure. Elderly individuals (>65 years) were
6 most clearly identified, along with those having preexisting cardiovascular or respiratory disease
7 conditions. The latter likely include smokers and ex-smokers as individuals comprising large
8 percentages of cardiovascular and respiratory disease (e.g., COPD) sufferers. Individuals with
9 asthma also were, albeit more tentatively, identified as a likely susceptible population group as
10 well, as were children. The new studies appearing since the 1996 PM AQCD, as assessed earlier
11 in this document and chapter, provide considerable additional evidence substantiating all of the
12 above named groups as likely being at increased risk for ambient PM-related morbidity or
13 mortality effects. Information related to factors contributing to such increased susceptibility or
14 useful in placing the potential public health impacts in perspective is presented below.
15
16 9.7.2 Preexisting Disease as a Risk Factor for Particulate Matter Health
17 Effects
18 Earlier available information reviewed in the 1996 PM AQCD has now been extensively
19 augmented by new studies that substantiate well that preexisting disease conditions are among
20 the most important key risk factors for ambient PM health effects. Cardiovascular- and
21 respiratory-related diseases have been shown to be of greatest concern, thus far, in relation to
22 increasing risk for PM mortality and morbidity. Table 9-9 shows the numbers of U.S. cases
23 reported for COPD, asthma, heart disease, and hypertension.
24
25 9.7.2.1 Ambient Particulate Matter Exacerbation of Cardiovascular Disease Conditions
26 Exacerbation of heart disease has been epidemiologically associated, not only with ambient
27 PM, but also with other combustion-related ambient pollutants such as CO. Thus, while leaving
28 little doubt that ambient PM exposures importantly affect CVD mortality and morbidity, the
29 quantitation of the proportion of risk for such exacerbation specifically attributable to ambient
30 PM exposure is difficult. Recent studies (e.g., concentrated ambient particle studies [CAPS])
31 have demonstrated cardiovascular effects in response to ambient particle exposures, and
March 2001 9-99 DRAFT-DO NOT QUOTE OR CITE
-------
tr
TABLE 9-9. INCIDENCE OF SELECTED CARDIORESPIRATORY DISORDERS BY AGE AND
BY GEOGRAPHIC REGION, 1996
(reported as incidence per thousand population and as number of cases in thousands)
0
o
1
0
o
a
!>
H
1
a
o
o
H
o
a
o
H
tn
O
73
o
H
W
Chronic Condition/Disease
COPD*
Incidence/ 1,000 persons
No. cases x 1,000
Asthma
Incidence/1,000 persons
No. cases x 1,000
Heart Disease
Incidence/ 1,000 persons
No. cases x ] ,000
HD-ischemic
Incidence/1,000 persons
No. cases x 1,000
HD-rhythmic
Incidence/ 1 ,000 persons
No. cases x 1,000
Hypertension
Incidence/ 1,000 persons
No. cases x 1,000
All Ages
60.4
15,971
55.2
14,596
78.2
20,653
29
7,672
33
8,716
107.1
28,314
Under 45
50.6
9,081
58.9
10,570
33.1
5,934
2.5
453
24.3
4,358
30.1
5,391
Age
45-64
72.3
3,843
48.6
2,581
116.4
6,184
51.6
2,743
40.7
2,164
214.1
11,376
Over 65
95.9
3,047
45.5
1,445
268.7
8,535
140.9
4,476
69.1
2,195
363.5
11,547
Over 75
99.9
1,334
48.0
641
310.7
4,151
154.6
2,065
73.1
977
373.8
4,994
Regional
NE MW S W
57.8 67.6 59.4 56.6
61.8 56.6 51.8 52.9
88.5 78.0 77.0 70.4
28.9 30.0 30.7 25.0
40.2 34.0 28.1 32.9
109.3 108.2 113.5 93.7
'Total chronic bronchitis and emphysema.
Source: Adams et al. (1999).
-------
1 studies utilizing other techniques also have produced various results suggesting some plausible
2 mechanisms for cardiovascular effects. However, much remains to be resolved with regard to
3 delineation of dose-response relationships for the induction of such effects and the extrapolation
4 of such to estimate effective human equivalent exposures to ambient PM (or specific constituent)
5 concentrations.
6 Schwartz (1999) has argued that independent effects of both PM and other pollutants are
7 biologically plausible. In the case of PM10, Schwartz's plausibility argument draws on the
8 emerging literature, which has demonstrated effects of ambient PM on pulmonary inflammation
9 in laboratory animals and human volunteers (Gilmour et al., 1996; Salvi et al., 1999), toxicity of
10 transition metals carried by combustion- generated particles (Costa and Dreher, 1999), effects on
11 cardiac dysfunction in animals with preexisting cardiopulmonary disease (Godleski et al., 1996;
12 Watkinson et al., 1998), and new epidemiologic evidence of associations between ambient PM
13 and physiologic changes in cardiac function (Pope et al., 1999a,b; Liao et al., 1999; Peters et al.,
14 1999b; Gold et al., 1998, 2000) and plasma viscosity (Peters et al., 1997a) in humans. For CO,
15 his argument is based on well-established effects of CO on oxygen transport by hemoglobin,
16 although such an impact typically is observed only at much higher CO concentrations than those
17 seen in these ambient studies. Although much more research is needed to clarify and confirm the
18 hypothesized linkages among these new findings, these arguments provide an initial framework
19 for such a linkage.
20 One very recently published HEI report on an epidemiologic study conducted by Goldberg
21 et al. (2000) in Montreal, Canada, provides especially interesting new information regarding
22 types of medical conditions potentially putting susceptible individuals at increased risk for PM-
23 associated mortality effects, and obtained results suggestive of other diseases with cardiovascular
24 complications being affected by ambient PM. First, the immediate causes of death, as listed on
25 death certificates, were evaluated in relation to various ambient PM indices (TSP, PM10
26 estimated PM2 5, COH, sulfates, and extinction coefficients) lagged for 0 to 4 days. Significant
27 associations were seen between each of the PM measures and total nonaccidental deaths,
28 respiratory diseases, and diabetes, with an approximate 2% increase in excess nonaccidental
29 mortality being observed per 9.5 /^g/m3 interquartile increase in 3-day mean estimated PM2 5
30 exposure. When underlying clinical conditions identified in the decedents' medical records were
31 then evaluated in relation to ambient PM measures, all three measures (COH, sulfate, and
March 2001 9-101 DRAFT-DO NOT QUOTE OR CITE
-------
1 estimated PM2 5) were associated with acute lower respiratory disease, congestive heart failure,
2 and any cardiovascular disease. Predicted PM2 5 and COH also were reported to be associated
3 with cancer, chronic coronary artery disease, and any coronary artery disease, whereas sulfate
4 was associated with acute and chronic upper respiratory disease. None of the three PM measures
5 were related to airways disease, acute coronary artery disease, or hypertension. These results
6 both tend to confirm previous findings identifying those with preexisting cardiopulmonary
7 diseases as being at increased risk for ambient PM effects and implicate another possible risk
8 factor, diabetes (which involves cardiovascular complications as it progresses), as a potential
9 susceptibility condition putting individuals at increased risk for ambient PM effects.
10 To the extent that observed associations of ambient PM with heart disease exacerbation
11 prove to be causal and specific to PM, they would be of genuine public health concern. In the
12 U.S. in 1997, there were about 4,188,000 hospital discharges with heart disease as the first-listed
13 diagnosis (Lawrence and Hall, 1999). Among these, about 2,090,000 (50%) were for ischemic
14 heart disease, 756,000 (18%) for myocardial infarction or heart attack (a subcategory of ischemic
15 heart disease), 957,000 (23%) for congestive heart failure, and 635,000 (15%) for cardiac
16 dysrhythmias. Also, there were 726,974 deaths from heart disease (Hoyert et al., 1999). Even a
17 small percentage reduction in admissions or deaths from heart disease would predict a large
18 number of avoided cases.
19
20 9.7.2.2 Ambient Particulate Matter Exacerbation of Respiratory Disease Conditions
21 Many investigators also have observed associations of short-term fluctuations in ambient
22 PM with daily frequency of respiratory illness. In most cases, exacerbation of preexisting
23 respiratory illness has been assessed, although some cases of acute respiratory infection may be
24 considered as occurrence of new illness, especially in young people. Symptoms of acute
25 respiratory distress in children have been linked to elevated PM concentrations in studies in the
26 United States and other countries, with asthmatics apparently more susceptible than
27 nonasthmatics. However, some studies also have found associations between child respiratory
28 symptoms or reduced lung function and other pollutants (such as O3) in addition to PM or no
29 significant relationship with air pollution. The credibility of ambient PM plausibly being linked
30 to exacerbation of preexisting respiratory disease (e.g., asthma) is enhanced by newly reported
31 dosimetry data noted earlier, which show greater lung deposition of l-/ini particles in people
March 2001 9-102 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
with varying degrees of airway obstruction than in healthy subjects. The increased deposition
was greatest for COPD patients and asthmatics, but smokers also showed increased deposition as
well.
In the United States in 1997, there were 3,475,000 hospital discharges for respiratory
diseases: 38% for pneumonia, 14% for asthma, 13% for chronic bronchitis, 8% for acute
bronchitis, and the remainder not specified (Lawrence and Hall, 1999). Of the 195,943 deaths
recorded as caused by respiratory diseases, 44% resulted from acute infections, 10% for
emphysema and bronchitis, 2.8% for asthma, and 42% for unspecified COPD (Hoyert et al.,
1999). Again, even a small percentage reduction in respiratory-related diseases could calculate
out to a large number of avoided cases.
9.7.3 Aged-Related At-Risk Population Groups: The Elderly and Children
Why are the very young and the very old apparently among those most affected by PM air
pollution? One major factor in increased susceptibility to air pollution is the presence of a
preexisting illness, as shown by Zanobetti and Schwartz (2000). The youngest children have the
highest rates of respiratory illnesses, as shown in the Table 9-10, which may be an important
factor in their apparently greater susceptibility to the adverse effects of PM air pollution.
TABLE 9-10. NUMBER OF ACUTE RESPIRATORY CONDITIONS PER
100 PERSONS PER YEAR, BY AGE: UNITED STATES, 1996
Type of Acute Condition
Respiratory Conditions
Common Cold
Other Acute Upper Respiratory
Infections
Influenza
Acute Bronchitis
Pneumonia
Other Respiratory Conditions
All
Ages
78.9
23.6
11.3
36.0
4.6
1.8
1.7
Under 5
Years
1294
48.6
13.1
53.7
*7.2
*3.9
*2.9
5-17
Years
101.5
33.8
15.0
443
4.3
*1 7
*2.4
18-24
Years
86.0
23.8
16.1
40.5
*3.9
*1.4
*0.4
25-44
Years
76.9
187
11.6
38.1
5.1
*1.3
*2.0
45
Total
53.3
16.1
7.0
23.3
38
*2.0
*1 1
Years and Over
45-64
Years
55.9
16.4
7.5
26.1
3.5
*0.9
*1 5
65 Years
and Over
49.0
157
6.1
18.6
*4.4
*3.8
*0 5
Source: Adams et al. (1999).
March 2001
9-103
DRAFT-DO NOT QUOTE OR CITE
-------
1 In addition to their higher incidences of preexisting respiratory conditions, several other
2 factors may render children and infants more susceptible to PM exposures, including a greater
3 amount of time spent outdoors, greater activity levels and breathing rates, higher doses per body
4 weight and lung surface area, and potential irreversible effects on children's developing lungs.
5 For example, PM doses on a per kilogram body weight basis are much higher for children than
6 for adults. This is displayed graphically in the Figure 9-14, which indicates that the amount of
7 air inhaled per kilogram body weight increases dramatically as age decreases below adult levels,
8 with the inhalation rate (in cubic meters per kilogram a day) of a 10-year-old being roughly twice
9 that of a 30-year-old person, and this estimate does not consider higher personal exposure
10 concentrations that a child is usually exposed to as a result of higher activity levels. Thus, on a
11 per unit body weight basis, children receive higher doses of air pollution than adults, consistent
12 with lung deposition information discussed earlier in this chapter.
13 Child-adult dosage disparities are even greater when viewed on a per lung area basis. This
14 may be more important than body weight if the number of particle "hits" per unit lung surface is
15 an important health impact metric, which it may well be for ultrafme particles. A newborn infant
16 has approximately 10 million alveoli versus some 300 million as an adult. The alveolar surface
17 area increases from approximately 3 m2 at birth to about 75 m2 in adulthood, causing the dose
18 delivered per lung surface area for infants and children to be much higher than in adults, even
19 given the same personal exposures (which is not the case, as they generally have greater PM,0
20 personal exposures than adults, as noted above). Thus, observed high PM air pollution-hospital
21 admissions associations for infants may result from PM doses that are significantly higher in
22 children than in adults, when one considers children's higher personal exposures, their greater
23 activity rates, and their smaller body weights and lung surface areas.
24 As discussed by Plopper and Fanucchi (2000), the limited experimental and epidemiologic
25 studies currently available identify the early postneonatal period of lung development as a time of
26 high susceptibility for lung damage created by exposure to environmental toxicants. This is
27 likely the reason for the above noted high rate of respiratory infectious diseases in young
28 children. In addition to their diminished immune status, infants are growing rapidly, and some
29 recent (though limited) evidence supports the hypothesis that environmental pollution can
30 significantly alter development of the respiratory system at that period of life. In experimental
31 animals, for example, elevated neonatal susceptibility to lung-targeted toxicants has been
March 2001 9-104 DRAFT-DO NOT QUOTE OR CITE
-------
0.6
0.5
0.4
03
0.3
CD
-E 0.2
0.1
40
T~
50
60
70
80
Age (y)
Figure 9-14. Inhalation rates on a per body-weight basis for males (•) and females (±) by
age (Layton, 1993).
1 reported at doses "well below the no-effects level for adults" (Plopper and Fanucchi, 2000;
2 Fanucchi and Plopper, 1997). In addition, acute injury to the lung during early postnatal
3 development causes a failure of normal repair processes, including down-regulation of cellular
4 proliferation at sites of injury (Smiley-Jewel et al., 2000, Fanucchi et al., 2000). Both infants'
5 diminished defenses and pollution-induced impairment of repair mechanisms therefore can
6 coincide during infancy, making the neonatal and postneonatal period one of potentially
7 especially elevated susceptibility to damage by environmental toxicants like PM.
8 Other information reviewed earlier in this document and chapter highlighted new evidence
9 pointing toward enhanced asthma symptoms, pulmonary function decrements, and asthma-
March 2001
9-105
DRAFT-DO NOT QUOTE OR CITE
-------
1 related doctors' visits and hospital admissions being associated with ambient PM exposures.
2 Generally higher activity levels in children and other factors related to attaining adequate medical
3 control of asthma in children may put asthmatic children (especially physically active mild to
4 moderate asthmatics) at particular risks for untoward effects of ambient PM among pediatric
5 population groups.
6 These and other types of health effects in children are emerging as a more important area of
7 concern than in the 1996 PM AQCD. Unfortunately, relatively little is known about the
8 relationship of PM to the most serious health endpoints (low birth weight, preterm birth, neonatal
9 and infant mortality, emergency hospital admissions and mortality in older children). Also, little
10 is yet known about involvement of PM exposure in the progression from less serious childhood
11 conditions, such as asthma and respiratory symptoms, to more serious disease endpoints later in
12 life. This is an important health issue because childhood illness or death may cost a very large
13 number of productive life-years. Lastly new epidemiologic studies of ambient PM associations
14 with increased non-hospital medical visits (physician visits) and asthma effects suggest likely
15 much larger health impacts and costs to society due to ambient PM effects on children than just
16 those indexed by mortality and/or hospital admissions/visits.
17 In contrast to information noted above for children, elderly adults do not appear to be put at
18 increased risk because of difference in lung deposition, clearance, or retention of inhaled
19 particles associated with aging, per se. However, the possible gradual focal accumation of
20 previously inhaled PM material at bifurcations and carinal ridges in TB airways and release of
21 previously accumulated inhaled PM-derived materials from lymph nodes could contribute to
22 enhanced susceptibility of elderly adults, especially those residing for long periods of time in
23 high PM exposure areas.
24 Probably of much more importance in placing elderly adults at increased risk for PM
25 effects is the higher propensity for such individuals to have preexisting cardiovascular or
26 respiratory disease conditions. Increased breathing rates due to compromised (e.g., obstructed)
27 lungs and airways or altered particle deposition patterns resulting from such conditions could be
28 among important factors increasing the risk for the elderly.
29
March 2001 9-106 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Abbey, D. E.; Ostro, B. E.; Petersen, F.; Burchette, R. J. (1995) Chronic respiratory symptoms associated with
3 estimated long-term ambient concentrations of fine participates less than 2.5 microns in aerodynamic
4 diameter (PM25) and other air pollutants. J. Exposure Anal. Environ. Epidemiol. 5: 137-159.
5 Abbey, D. E.; Burchette, R. J.; Knutsen, S. F.; McDonnell, W. F.; Lebowitz, M. D.; Enright, P. L. (1998) Long-term
6 particulate and other air pollutants and lung function in nonsmokers. Am. J. Respir. Crit. Care Med.
7 158:289-298.
8 Abbey, D. E.; Nishino, N.; McDonnell, W. F.; Burchette, R. J.; Knutsen, S. F.; Beeson, W. L.; Yang, J. X. (1999)
9 Long-term inhalable particles and other air pollutants related to mortality in nonsmokers. Am. J. Respir.
10 Crit. Care Med. 159: 373-382.
11 Adams, P. F.; Hendershot, G. E.; Marano, M. A. (1999) Current estimates from the National Health Interview
12 Survey, 1996. Vital Health Stat. 10(200).
13 Bigger, J. T., Jr.; Fleiss, J. L.; Steinman, R. C.; Rolnitzky, L. M.; Kleiger, R. E.; Rottman, J. N. (1992) Frequency
14 domain measures of heart period variability and mortality after myocardial infarction. Circulation
15 85:164-171.
16 Bobak, M.; Leon, D. A. (1999) Pregnancy outcomes and outdoor air pollution: an ecological study in districts of the
17 Czech Republic 1986-8. Occup. Environ. Med. 56: 539-543.
18 Broday, D. M.; Georgopoulos, P. G. (2000) Growth and deposition of hygroscopic particulate matter in the human
19 lungs. Aerosol Sci. Technol.: submitted.
20 Brunekreef, B. (1997) Air pollution and life expectancy: is there a relation? Occup. Environ. Med. 54: 781-784.
21 Burnett, R. T.; Dales, R. E.; Raizenne, M. E.; Krewski, D.; Summers, P. W.; Roberts, G. R.; Raad-Young, M.;
22 Dann, T.;-Brook, J. (1994) Effects of low ambient levels of ozone and sulfates on the frequency of
23 respiratory admissions to Ontario hospitals. Environ. Res. 65: 172-194.
24 Burnett, R. T.; Dales, R.; Krewski, D.; Vincent, R.; Dann, T.; Brook, J. R. (1995) Associations between ambient
25 particulate sulfate and admissions to Ontario hospitals for cardiac and respiratory diseases. Am. J.
26 Epidemiol. 142: 15-22.
27 Burnett, R. T.; Cakmak, S.; Brook, J. R.; Krewski, D. (1997) The role of particulate size and chemistry in the
28 association between summertime ambient air pollution and hospitalization for cardiorespiratory diseases.
29 Environ. Health Perspect. 105: 614-620.
30 Burnett, R. T.; Cakmak, S.; Raizenne, M. E.; Stieb, D.; Vincent, R.; Krewski, D.; Brook, J. R.; Philips, O.;
31 Ozkaynak, H. (1998) The association between ambient carbon monoxide levels and daily mortality in
32 Toronto, Canada. J. Air Waste Manage. Assoc. 48: 689-700.
33 Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Cakmak, S.; Brook, J. R. (1999) Effects of particulate and gaseous air
34 pollution on cardiorespiratory hospitalizations. Arch. Environ. Health 54: 130-139.
35 Burnett, R. T.; Brook, J.; Dann, T.; Delocla, C.; Philips, O.; Cakmak, S.; Vincent, R.; Goldberg, M. S.; Krewski, D.
36 (2000) Association between particulate- and gas-phase components of urban air pollution and daily
37 mortality in eight Canadian cities. In: Grant, L. D., ed. PM2000: particulate matter and health. Inhalation
38 Toxicol. 12(suppl.4): 15-39.
39 Castillejos, M.; Borja-Aburto, V. H.; Dockery, D. W.; Gold, D. R.; Loomis, D. (2000) Airborne coarse particles and
40 mortality. In: Inhalation Toxicology: proceedings of the third colloquium on particulate air pollution and
41 human health; June, 1999; Durham, NC. Inhalation Toxicology 12(suppl. 1): 61-72.
42 Chen, L.; Yang, W.; Jennison, B. L.; Omaye, S. T. (2000) Air particulate pollution and hospital admissions for
43 chronic obstructive pulmonary disease in Reno, Nevada. Inhalation Toxicol. 12: 281-298.
44 Chock, D. P.; Winkler, S.; Chen, C. (2000) A study of the association between daily mortality and ambient air
45 pollutant concentrations in Pittsburgh, Pennsylvania. J. Air Waste Manage. Assoc. 50: 1481-1500.
46 Choudhury, A. H.; Gordian, M. E.; Morris, S. S. (1997) Associations between respiratory illness and PMIO air
47 pollution. Arch. Environ. Health 52: 113-117.
48 Cifuentes, L. A.; Vega, J.; Kopfer, K.; Lave, L. B. (2000) Effect of the fine fraction of particulate matter versus the
49 coarse mass and other pollutants on daily mortality in Santiago, Chile. J. Air Waste Manage. Assoc.
50 50: 1287-1298.
51 Claiborn, C. S.; Finn, D.; Larson, T. V.; Koemg, J. Q. (2000) Windblown dust contributes to high PM2 5
52 concentrations. J. Air Waste Manage. Assoc. 50: 1440-1445.
53 Clark, W. E.; Whitby, K. T. (1975) Measurements of aerosols produced by the photochemical oxidation of SO2 in
54 air. J. Colloid Interface Sci. 51: 477-490.
March 2001 9-107 DRAFT-DO NOT QUOTE OR CITE
-------
1 Clyde, M. A.; Guttorp, P.; Sullivan, E. (2000) Effects of ambient fine and coarse particles on mortality in Phoenix,
2 Arizona. J. Exposure Anal. Environ. Epidemiol.: submitted.
3 Code of Federal Regulations. (199 la) Appendix J to Part 50-reference method for the determination of paniculate
4 matter as PM10 in the atmosphere. C. F. R. 40: §50.
5 Code of Federal Regulations. (1991b) Ambient air monitoring reference and equivalent methods. C. F. R. 40: §53.
6 Costa, D. L.; Dreher, K. L. (1999) What do we need to know about airborne particles to make effective risk
7 management decisions? A toxicology perspective. Hum. Ecol. Risk Assess. 5: 481-491.
8 Daniels, M.; Dominici, F.; Samet, J. M.; Zeger, S. L. (2000) Estimating particulate matter-mortality dose-response
9 curves and threshold levels: an analysis of daily time-series for the 20 largest US cities. Am. J. Epidemiol.
10 152:397-406.
11 Dejmek, J.; Selevan, S. G.; Benes, I.; Solansky, I.; Sram, R. J. (1999) Fetal growth and maternal exposure to
12 particulate matter during pregnancy. Environ. Health Perspect. 107: 475-480.
13 Delfino, R. J.; Murphy-Moulton, A. M.; Burnett, R. T.; Brook, J. R.; Becklake, M. R. (1997) Effects of air pollution
14 on emergency room visits for respiratory illnesses in Montreal, Quebec. Am. J. Respir. Cnt. Care Med.
15 155:568-576.
16 Delfino, R. J.; Zeiger, R. S.; Seltzer, J. M.; Street, D. H. (1998) Symptoms in pediatric asthmatics and air pollution:
17 differences in effects by symptom seventy, anti-inflammatory medication use and particulate averaging
18 time. Environ. Health Perspect. 106: 751-761.
19 Dockery, D. W.; Speizer, F. E.; Stram, D. O.; Ware, J. H.; Spengler, J. D.; Ferris, B. G., Jr. (1989) Effects of
20 inhalable particles on respiratory health of children. Am. Rev. Respir. Dis. 139: 587-594.
21 Dockery, D. W.; Pope, C. A., Ill; Xu, X.; Spengler, J. D.; Ware, J. H.; Fay, M. E.; Ferris, B. G., Jr.; Speizer, F. E.
22 (1993) An association between air pollution and mortality in six U.S. cities. N. Engl. J. Med.
23 329: 1753-1759.
24 Dockery, D. W.; Cunningham, J.; Damokosh, A. I.; Neas, L. M.; Spengler, J. D.; Koutrakis, P.; Ware, J. H.;
25 Raizenne, M.; Speizer, F. E. (1996) Health effects of acid aerosols on North American children:
26 respiratory symptoms. Environ. Health Perspect. 104: 500-505.
27 Dockery, D. W.; Pope, C. A., Ill; Kanner, R. E.; Villegas, G. M.; Schwartz, J. (1999) Daily changes in oxygen
28 saturation and pulse rate associated with particulate air pollution and barometric pressure. Cambridge, MA:
29 Health Effects Institute; research report no. 83.
30 Driscoll, K. E.; Costa, D. L.; Hatch, G.; Henderson, R.; Oberdorster, G.; Salem, H.; Schlesinger, R. B. (2000)
31 Intratracheal instillation as an exposure technique for the evaluation of respiratory tract toxicity: uses and
32 limitations. Toxicol. Sci. 55: 24-35.
33 Elder, A. C. P.; Gelein, R.; Finkelstein J. N.; Cox, C.; Oberdorster, G. (2000) Endotoxin priming affects the lung
34 response to ultrafine particles and ozone in young and old rats. In: Inhalation Toxicology: proceedings of
35 the third colloquium on particulate air pollution and human health; June, 1999; Durham, NC. Inhalation
36 Toxicology 12(suppl. 1): 85-98.
37 Fairley, D. (1999) Daily mortality and air pollution in Santa Clara County, California: 1989-1996. Environ. Health
38 Perspect. 107:637-641.
39 Fanucchi, M. V.; Plopper, C. G. (1997) Pulmonary developmental responses to toxicants. In: Roth, R. A., ed.
40 Toxicology of the Respiratory System, v. 8 of Comprehensive Toxicology. New York, NY: Pergamon,
41 p. 203-220.
42 Fanucchi, M. V.; Wong, V. J.; Hinds, D.; Tarkington, B. K.; Van Winkle, L. S.; Evans, M. J.; Plopper, C. G. (2000)
43 Repeated episodes of exposure to ozone alters postnatal development of distal conducting airways in infant
44 rhesus monkeys. Am. J. Respir. Crit. Care Med. 161: A615.
45 Federal Register. (1987) Revisions to the national ambient air quality standards for particulate matter. F. R. (July 1)
46 52:24,634-24,669.
47 Fitz-Simons, T. S.; Mathias, M.; Rizzo, M. (2000) Analyses of 1999 PM data for the PM NAAQS review.
48 U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards; September 29.
49 Gardner, S. Y.; Lehmann, J. R.; Costa, D. L. (2000) Oil fly ash-induced elevation of plasma fibrinogen levels in
50 rats. Toxicol. Sci. 56: 175-180.
51 Gauderman, W. J.; Mcconnell, R.; Gilliland, F.; London, S.; Thomas, D.; Avol, E.; Vora, H.; Berhane, K.;
52 Rappaport, E. B.; Lurmann, F.; Margolis, H. G.; Peters, J. (2000) Association between air pollution and
53 lung function growth in southern California children. Am. J. Respir. Crit. Care Med. 162: 1383-1390.
March 2001 9-108 DRAFT-DO NOT QUOTE OR CITE
-------
1 Gielen, M. H.; Van Der Zee, S. C.; Van Wijnen, J. H.; Van Steen, C. J.; Brunekreef, B. (1997) Acute effects of
2 summer air pollution on respiratory health of asthmatic children. Am. J. Respir. Crit. Care Med. 155: 2105-2108.
3 Gilmour, P. S.; Brown, D. M.; Lindsay, T. G.; Beswick, P. H.; MacNee, W.; Donaldson, K. (1996) Adverse health
4 effects of PM|0 particles: involvement of iron in generation of hydroxyl radical. Occup. Environ. Med.
5 53:817-822.
6 Godleski, J. J.; Sioutas, C.; Katler, M.; Koutrakis, P. (1996) Death from inhalation of concentrated ambient air
7 particles in animal models of pulmonary disease. Am. J. Respir. Crit. Care Med. 153: A15.
8 Godleski, J. J.; et al. (1998) Increased cardiac vulnerability during exposure to inhaled environmental particles
9 [abstract]. Presented at: Health Effects Institute annual meeting; May; Boston, MA. Boston, MA: Health
10 Effects Institute.
11 Godleski, J. J.; Verrier, R. L.; Koutrakis, P.; Catalano, P. (2000) Mechanisms of morbidity and mortality from
12 exposure to ambient air particles. Cambridge, MA: Health Effects Institute; research report no. 91.
13 Gold, A.; Litonjua, J.; Schwartz, M.; Verrier, R.; Milstein, A.; Larson, E.; Lovett, B. (1998) Cardiovascular
14 vulnerability to particulate pollution. Presented at: 1998 International conference [American Thoracic
15 Society]; April; Chicago, IL. Am. J. Respir. Crit. Care Med. 157: A261.
16 Gold, D. R.; Damokosh, A. I.; Pope, C. A., Ill; Dockery, D. W.; McDonnell, W. F.; Serrano, P.; Retama, A.;
17 Castillejos, M. (1999) Particulate and ozone pollutant effects on the respiratory function of children in
18 southwest Mexico City. Epidemiology 10: 8-16.
19 Gold, D. R.; Litonjua, A.; Schwartz, J.; Lovett, E.; Larson, A.; Nearing, B.; Allen, G.; Verrier, M.; Cherry, R.;
20 Verrier, R. (2000) Ambient pollution and heart rate variability. Circulation 101: 1267-1273.
21 Goldberg, M. S.; Bailar, J. C., Ill; Burnett, R. T.; Brook, J. R.; Tamblyn, R.; Bonvalot, Y.; Ernst, P.; Flegel, K. M.;
22 Singh, R. K.; Valois, M.-F. (2000) Identifying subgroups of the general population that may be susceptible
23 to short-term increases in particulate air pollution: a time-series study in Montreal, Quebec. Cambridge,
24 MA: Health Effects Institute; research report 97. Available:
25 http://www.healtheffects.org/pubs-research.htm [15 February, 2001].
26 Goldsmith, C.-A. W.; Hamada, K.; Ning, Y. Y.; Qin, G.; Catalano, P.; Murthy, G. G. K.; Lawrence, J.; Kobzik, L.
27 (1999) The effects of environmental aerosols on airway hyperresponsiveness in a murine model of asthma.
28 Inhalation Toxicol. 11: 981-998.
29 Gordon, T.; Nadziejko, C.; Chen, L. C.; Schlesinger, R. (2000) Effects of concentrated ambient particles in rats and
30 hamsters: an exploratory study. Cambridge, MA: Health Effects Institute; research report no. 93.
31 Gwynn, R. C.; Burnett, R. T.; Thurston, G. D. (2000) A time-series analysis of acidic particulate matter and daily
32 mortality and morbidity in the Buffalo, New York, region. Environ. Health Perspect. 108: 125-133.
33 Hayano, J.; Sakakibara, Y.; Yamada, M.; Ohte, N.; Fujmami, T.; Yokoyama, K.; Watanabe, Y.; Takata, K. (1990)
34 Decreased magnitude of heart rate spectral components in coronary artery disease. Its relation to
35 angiographic severity. Circulation 81: 1217-1224.
36 Hofmann, W.; Balashazy, I.; Heistracher, T.; Koblmger, L. (1996) The significance of particle deposition patterns in
37 bronchial airway bifurcations for extrapolation modeling. Aerosol Sci. Technol. 25: 305-327.
38 Hoyert, D. L.; Kochanek, K. D.; Murphy, S. L. (1999) Deaths: final data for 1997. Atlanta, GA: U.S. Centers for
39 Disease Control and Prevention, National Center for Health Statistics. (National vital statistics reports:
40 v. 47, no. 19).
41 Hsieh, T. H.; Yu, C. P. (1998) Two-phase pulmonary clearance of insoluble particles in mammalian species.
42 Inhalation Toxicol. 10: 121-130.
43 Kamens, R.; Jang, M.; Chien, C.-J.; Leach, K. (1999) Aerosol formation from the reaction of a-pinene and ozone
44 using a gas-phase kinetics-aerosol partitioning model. Environ. Sci. Technol. 33: 1430-1438.
45 Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (1999) Relationships between size
46 segregated mass concentration data and ultrafine particle number concentrations in urban areas.
47 Atrhos. Environ. 33: 2907-2913.
48 Kim, Y. J.; Boatman, J. F.; Gunter, R. L.; Wellman, D. L.; Wilkison, S. W. (1993) Vertical distribution of
49 atmospheric aerosol size distribution over south-central New Mexico. Atmos. Environ. Part A
50 27: 1351-1362.
51 Kittelson, D. B. (1998) Engines and nanoparticles: a review. J. Aerosol Sci. 29: 575-588.
52 Kleiger, R. E.; Miller, J. P.; Bigger J. T. Jr.; Moss, A. J.; Multicenter Post-infarction Research Group. (1987)
53 Decreased heart rate variability and its association with increased mortality after acute myocardial
54 infarction. Am. J. Cardiol. 59: 256-262.
March 2001 9-109 DRAFT-DO NOT QUOTE OR CITE
-------
1 Kleinman, M. T.; Leaf, D. A.; Kelly, E.; Caiozzo, V.; Osann, K.; O'Niell, T. (1998) Urban angina in the mountains:
2 effects of carbon monoxide and mild hypoxemia on subjects with chronic stable angina. Arch. Environ.
3 Health 53: 388-397.
4 Klemm, R. J.; Mason, R. M., Jr. (2000) Aerosol research and inhalation epidemiological study (ARIES): air quality
5 and daily mortality statistical modeling-interim results. J. Air. Waste Manage. Assoc. 50: 1433-1439.
6 Klemm, R. J.; Mason, R. M., Jr.; Heilig, C. M.; Neas, L. M.; Dockery, D. W. (2000) Is daily mortality associated
7 specifically with fine particles? Data reconstruction and replication of analyses. J. Air Waste Manage.
8 Assoc. 50: 1215-1222.
9 Krewski, D.; Burnett, R. T.; Goldberg, M. S.; Hoover, K.; Siemiatycki, J.; Jerrett, M.; Abrahamowicz, M.; White,
10 W. H.. (2000) Reanalysis of the Harvard Six Cities study and the American Cancer Society study of
11 particulate air pollution and mortality. A special report of the Institute's Particle Epidemiology Reanalysis
12 Project. Cambridge, MA: Health Effects Institute.
13 Kuschner, W. G.; Wong, H.; D'Alessandro, A.; Quinlan, P.; Blanc, P. D. (1997) Human pulmonary responses to
14 experimental inhalation of high concentration fine and ultrafine magnesium oxide particles. Environ.
15 Health Perspect. 105: 1234-1237.
16 Laden, F.; Neas, L. M.; Dockery, D. W.; Schwartz, J. (2000) Association of fine particulate matter from different
17 sources with daily mortality in six U.S. cities. Environ. Health Perspect. 108: 941-947.
18 Lawrence, L.; Hall, M. J. (1999) 1997 summary: national hospital discharge survey. Hyattsville, MD: U.S.
19 Department of Health & Human Resources, Centers for Disease Control and Prevention, National Center
20 for Health Statistics. (Advance data from vital and health statistics: no. 308).
21 Layton, D. W. (1993) Metabolically consistent breathing rates for use in dose assessments. Health Phys. 64: 23-36.
22 Li, X. Y.; Gilmour, P. S.; Donaldson, K.; MacNee, W. (1997) In vivo and in vitro proinflammatory effects of
23 particulate air pollution (PM10). In: Dnscoll, K. E.; Oberdorster, G., eds. Proceedings of the sixth
24 international meeting on the toxicology of natural and man-made fibrous and non-fibrous particles;
25 September 1996; Lake Placid, NY. Environ. Health Perspect. Suppl. 105(5): 1279-1283.
26 Liao, D.; Creason, J.; Shy, C.; Williams, R.; Watts, R.; Zweidinger, R. (1999) Daily variation of particulate air
27 pollution and poor cardiac autonomic control in the elderly. Environ. Health Perspect. 107: 521-525.
28 Linn, W. S.; Gong, H., Jr.; Shamoo, D. A.; Anderson, K. R.; Avol, E. L. (1997) Chamber exposures of children to
29 mixed ozone, sulfur dioxide, and sulfuric acid. Arch. Environ. Health 52: 179-187.
30 Linn, W. S.; Szlachcic, Y.; Gong, H., Jr.; Kinney, P. L.; Berhane, K. T. (2000) Air pollution and daily hospital
31 admissions in metropolitan Los Angeles. Environ. Health Perspect. 108: 427-434.
32 Lipfert, F. W.; Morris, S. C.; Wyzga, R. E. (2000a) Daily mortality in the Philadelphia metropolitan area and
33 size-classified particulate matter. J. Air Waste Manage. Assoc.: 1501-1513.
34 Lipfert, F. W.; Zhang, J.; Wyzga, R. E. (2000b) Infant mortality and air pollution: a comprehensive analysis of U.S.
35 data for 1990. J. Air Waste Manage. Assoc. 50: 1350-1366.
36 Lippmann, M.; Ito, K.; Nadas, A.; Burnett, R. T. (2000) Association of particulate matter components with daily
37 mortality and morbidity in urban populations. Cambridge, MA: Health Effects Institute; research report
38 no. 95.
39 Loomis, D.; Castillejos, M.; Gold, D. R.; McDonnell, W.; Borja-Aburto, V. H. (1999) Air pollution and infant
40 mortality in Mexico City. Epidemiology 10: 118-123.
41 Lumley, T.; Heagerty, P. (1999) Weighted empirical adaptive variance estimators for correlated data regression.
42 J. R. Stat. Soc. B 61 (part 2): 459-477.
43 Lundgren, D. A.; Burton, R. M. (1995) Effect of particle size distribution on the cut point between fine and coarse
44 ambient mass fractions. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on particulate
45 air pollution and human mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 131-148.
46 Mar, T. F.; Norris, G. A.; Koenig, J. Q.; Larson, T. V. (2000) Associations between air pollution and mortality in
47 Phoenix, 1995-1997. Environ. Health Perspect. 108: 347-353.
48 Martin, G. J.; Magid, N. M.; Myers, G.; Barnett, P. S.; Schaad, J. W.; Weiss, J. S.; Lesch, M.; Singer, D. H. (1987)
49 Heart rate variability and sudden death secondary to coronary artery disease during ambulatory
50 electrocardiographic monitoring. Am. J. Cardiol. 60: 86-89.
51 McConnell, R.; Berhane, K.; Gilliland, F.; London, S. J.; Vora, H.; Avol, E.; Gauderman, W. J.; Margolis, H. G.;
52 Lurmann, F.; Thomas, D. C.; Peters, J. M. (1999) Air pollution and bronchitic symptoms in southern
53 California children with asthma. Environ. Health Perspect. 107: 757-760.
54 Michel, O.; Nagy, A.-M.; Schroeven, M.; Duchateau, J.; Neve, J.; Fondu, P.; Sergysels, R. (1997) Dose-response
55 relationship to inhaled endotoxin in normal subjects. Am. J. Respir. Crit. Care Med. 156: 1157-1164.
March 2001 9-110 DRAFT-DO NOT QUOTE OR CITE
-------
1 Monn, C.; Becker, S. (1999) Cytotoxicity and induction of proinflammatory cytokines from human monocytes
2 exposed to fine (PM2 5) and coarse particles (PM10 2 5) in outdoor and indoor air. Toxicol. Appl. Pharmacol.
3 155:245-252.
4 Moolgavkar, S. H. (2000a) Air pollution and daily mortality in three U.S. counties. Environ. Health Perspect.
5 108:777-784.
6 Moolgavkar, S. H. (2000b) Air pollution and hospital admissions for chronic obstructive pulmonary disease in three
7 metropolitan areas in the United States. In: Grant, L. D., ed. PM2000: paniculate matter and health.
8 Inhalation Toxicol. 12(suppl. 4): 75-90.
9 Moolgavkar, S. H. (2000c) Air pollution and hospital admissions for diseases of the circulatory system in three U.S.
10 metropolitan areas. J. Air Waste Manage Assoc. 50: 1199-1206.
11 Moolgavkar, S. H.; Luebeck, E. G.; Anderson, E. L. (1997) Air pollution and hospital admissions for respiratory
12 causes in Minneapolis-St. Paul and Birmingham. Epidemiology 8: 364-370.
13 Moolgavkar, S. H.; Hazelton, W.; Luebeck, G.; Levy, D.; Sheppard, L. (2000) Air pollution, pollens, and
14 admissions for chronic respiratory disease in King County, Washington. In: Inhalation Toxicology:
15 proceedings of the third colloquium on particulate air pollution and human health; June, 1999; Durham,
16 NC. Inhalation Toxicology 12(suppl. 1): 157-171.
17 Morawska, L.; Thomas, S.; Bofinger, N.; Wainwright, D.; Neale, D. (1998) Comprehensive characterization of
18 aerosols in a subtropical urban atmosphere: particle size distribution and correlation with gaseous
19 pollutants. Atmos. Environ. 32: 2467-2478.
20 Morris, R. D.; Naumova, E. N. (1998) Carbon monoxide and hospital admissions for congestive heart failure:
21 evidence of an increased effect at low temperatures. Environ. Health Perspect. 106: 649-653.
22 Morris, R. D.; Naumova, E. N.; Munasinghe, R. L. (1995) Ambient air pollution and hospitalization for congestive
23 heart failure among elderly people in seven large US cities. Am. J. Public Health 85: 1361-1365.
24 Naeher, L. P.; Holford, T. R.; Beckett, W. S.; Belanger, K.; Triche, E. W.; Bracken, M. B.; Leaderer, B. P. (1999)
25 Healthy women's PEF variations with ambient summer concentrations of PM10, PN2 5, SO42, H+, and O3.
26 Am. J. Respir. Crit. Care Med. 160: 117-125.
27 Nakamura, T.; Hayashida, Y. (1992) Autonomic cardiovascular responses to smoke exposure in conscious rats. Am.
28 J. Physiol. 262(5 pt. 2): R738-745. . ,
29 National Institutes of Health. (1997) Guidelines for the diagnosis and management of asthma: expert panel report 2.
30 Bethesda, MD: U.S. Department of Health and Human Services, National Heart, Lung, and Blood
31 Institute; publication no. 97-4051.
32 Nauenberg, E.; Basu, K. (1999) Effect of insurance coverage on the relationship between asthma hospitalizations
33 and exposure to air pollution. Public Health Rep. 114: 135-148.
34 Neas, L. M.; Dockery, D. W.; Ware, J. H.; Spengler, J. D.; Ferris, B. G., Jr.; Speizer, F. E. (1994) Concentration of
35 indoor particulate matter as a determinant of respiratory health in children. Am. J. Epidemiol.
36 139: 1088-1099.
37 Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Tollerud, D. J.; Speizer, F. E. (1995) The association of ambient air
38 pollution with twice daily peak expiratory flow rate measurements in children. Am. J. Epidemiol
39 141:111-122.
40 Neas, L. M.; Dockery, D. W.; Burge, H.; Koutrakis, P.; Speizer, F. E. (1996) Fungus spores, air pollutants, and
41 other determinants of peak expiratory flow rate in children. Am. J. Epidemiol. 143: 797-807.
42 Neas, L. M.; Schwartz, J.; Dockery, D. (1999) A case-crossover analysis of air pollution and mortality in
43 Philadelphia. Environ. Health Perspect. 107: 629-631.
44 Nikula, K. J.; Vallyathan, V.; Green, F. H. Y.; Hahn, F. F. (2000) Influence of dose on the distribution of retained
45 particulate material in rat and human lungs. Presented at: PM2000: particulate matter and health—the
46 scientific basis for regulatory decision-making, specialty conference & exhibition; January; Charleston,
47 SC. Pittsburgh, PA: Air & Waste Management Association.
48 Norris, G.; Young-Pong, S. N.; Koemg, J. Q.; Larson, T. V.; Sheppard, L.; Stout, J. W. (1999) An association
49 between fine particles and asthma emergency department visits for children in Seattle. Environ. Health
50 Perspect. 107:489-493.
51 Norris, G.; Larson, T.; Koenig, J.; Claiborn, C.; Sheppard, L.; Finn, D. (2000) Asthma aggravation, combustion, and
52 stagnant air. Thorax 55: 466-470.
53 Ohtsuka, Y.; Clarke, R. W.; Mitzner, W.; Brunson, K.; Jakab, G. J.; Kleeberger, S. R. (2000) Interstrain variation in
54 munne susceptibility to inhaled acid-coated particles. Am. J. Physiol. L469-L476.
March 2001 9-111 DRAFT-DO NOT QUOTE OR CITE
-------
1 Ostro, B. D.; Lipsett, M. J.; Wiener, M. B.; Seiner, J. C. (1991) Asthmatic responses to airborne acid aerosols. Am.
2 J. Public Health 81: 694-702.
3 Ostro, B. D.; Lipsett, M. J.; Mann, J. K.; Krupnick, A.; Harrington, W. (1993) Air pollution and respiratory
4 morbidity among adults in Southern California. Am. J. Epidemiol. 137: 691-700.
5 Ostro, B. D.; Lipsett, M. J.; Mann, J. K.; Braxton-Owens, H.; White, M. C. (1995) Air pollution and asthma
6 exacerbations among African-American children in Los Angeles. In: Phalen, R. F.; Bates, D. V., eds.
7 Proceedings of the colloquium on particulate air pollution and human mortality and morbidity, part II;
8 January 1994; Irvine, CA. Inhalation Toxicol. 7: 711-722.
9 Ostro, B. D.; Broadwin, R.; Lipsett, M. J. (2000) Coarse and fine particles and daily mortality in the Coachella
10 Valley, California: a follow-up study. J. Exposure Anal. Environ. Epidemiol. 10: 412-419.
11 Ozkaynak, H.; Xue, J.; Zhou, H.; Raizenne, M. (1996) Associations between daily mortality and motor vehicle
12 pollution in Toronto, Canada. Boston, MA: Harvard University School of Public Health, Department of
13 Environmental Health; March 45.
14 Pekkanen, J.; Timonen, K. L.; Ruuskanen, J.; Reponen, A.; Mirme, A. (1997) Effects of ultrafme and fine particles
15 in urban air on peak expiratory flow among children with asthmatic symptoms. Environ. Res. 74: 24-33.
16 Perry, K. D.; Cahill, T. A.; Eldred, R. A.; Dutcher, D. D.; Gill, T. E. (1997) Long-range transport of North African
17 dust to the eastern United States. J. Geophys. Res. [Atmos] 102: 11,225-11,238.
18 Peters, A.; Doring, A.; Wichmann, H.-E.; Koenig, W. (1997a) Increased plasma viscosity during an air pollution
19 episode: a link to mortality? Lancet 349: 1582-1587.
20 Peters, A.; Wichmann, H. E.; Tuch, T.; Heinrich, J.; Heyder, J. (1997b) Respiratory effects are associated with the
21 number of ultrafme particles. Am. J. Respir. Crit. Care Med. 155: 1376-1383.
22 Peters, A.; Dockery, D. W.; Heinrich, J.; Wichmann, H. E. (1997c) Short-term effects of particulate air pollution on
23 respiratory morbidity in asthmatic children. Eur. Respir. J. 10: 872-879.
24 Peters, J. M.; Avol, E.; Navidi, W.; London, S. J.; Gauderman, W. J.; Lurmann, F.; Linn, W. S.; Margolis, H.;
25 Rappaport, E.; Gong, H., Jr.; Thomas, D. C. (1999a) A study of twelve southern California communities
26 with differing levels and types of air pollution. I. Prevalence of respiratory morbidity. Am. J. Respir. Crit.
27 Care Med. 159:760-767.
28 Peters, J. M.; Avol, E.; Gauderman, W. J.; Linn, W. S.; Navidi, W.; London, S. J.; Margolis, H.; Rappaport, E.;
29 Vora, H.; Gong, H., Jr.; Thomas, D. C. (1999b) A study of twelve southern California communities with
30 differing levels and types of air pollution. II. Effects on pulmonary function. Am. J. Respir. Crit. Care
31 Med. 159:768-775.
32 Plopper, C. G.; Fanucchi, M. V. (2000) Do urban environmental pollutants exacerbate childhood lung diseases?
33 Environ. Health Perspect. 108: A252-A253.
34 Pope, C. A., Ill; Thun, M. J.; Namboodiri, M. M.; Dockery, D. W.; Evans, J. S.; Speizer, F. E.; Heath, C. W., Jr.
35 (1995) Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am. J.
36 Respir. Crit. Care Med. 151: 669-674.
37 Pope, C. A., Ill; Hill, R. W.; Villegas, G. M. (1999a) Particulate air pollution and daily mortality on Utah's Wasatch
38 Front. Environ. Health Perspect.: 107: 567-573.
39 Pope, C. A.; Verrier, R. L.; Lovett, E. G.; Larson, A. C.; Raizenne, M. E.; Kanner, R. E.; Schwartz, J.; Villegas,
40 G. M.; Gold, D. R.; Dockery, D. W. (1999b) Heart rate variability associated with particulate air pollution.
41 Am. Heart J. 138: 890-899.
42 Raizenne, M.; Neas, L. M.; Damokosh, A. I.; Dockery, D. W.; Spengler, J. D.; Koutrakis, P.; Ware, J. H.; Speizer,
43 F. E. (1996) Health effects of acid aerosols on North American children: pulmonary function. Environ.
44 Health Perspect. 104: 506-514.
45 Ramadan, Z.; Song, X.-H.; Hopke, P. K. (2000) Identification of sources of Phoenix aerosol by positive matrix
46 factorization. J. Air Waste Manage. Assoc. 50: 1308-1320.
47 Romieu, L; Meneses, F.; Ruiz, S.; Sienra, J. J.; Huerta, J.; White, M. C.; Etzel, R. A. (1996) Effects of air pollution
48 on the respiratory health of asthmatic children living in Mexico City. Am. J. Respir. Crit. Care Med.
49 154:300-307.
50 Romieu, I.; Meneses, F.; Ruiz, S.; Huerta, J.; Sienra, J. J.; White, M.; Etzel, R.; Hernandez, M. (1997) Effects of
51 intermittent ozone exposure on peak expiratory flow and respiratory symptoms among asthmatic children
52 in Mexico City. Arch. Environ. Health 52: 368-376.
53 Salvi, S.; Blomberg, A.; Rudell, B.; Kelly, F.; Sandstrom, T.; Holgate, S. T.; Frew, A. (1999) Acute inflammatory
54 responses in the airways and peripheral blood after short-term exposure to diesel exhaust in healthy human
55 volunteers. Am. J. Respir. Crit. Care Med. 159: 702-709.
March 2001 9-112 DRAFT-DO NOT QUOTE OR CITE
-------
1 Samet, J. M.; Dominici, F.; Zeger, S. L.; Schwartz, J.; Dockery, D. W. (2000a) National morbidity, mortality, and
2 air pollution study. Part I: methods and methodologic issues. Cambridge, MA: Health Effects Institute;
3 research report no. 94.
4 Samet, J. M.; Zeger, S. L.; Dominici, F.; Curriero, F.; Coursac, I.; Dockery, D. W.; Schwartz, J.; Zanobetti, A.
5 (2000b) The national morbidity, mortality, and air pollution study. Part II: morbidity, mortality, and air
6 pollution in the United States. Cambridge, MA: Health Effects Institute; research report no. 94.
7 Sarnat, J. A.; Koutrakis, P.; Suh, H. H. (2000) Assessing the relationship between personal particulate and gaseous
8 exposures of senior citizens living in Baltimore, MD. J. Air Waste Manage. Assoc. 50: 1184-1198.
9 Schlesmger, R. B. (1988) Biological disposition of airborne particles: basic principles and application to vehicular
10 emissions. In: Watson, A. Y.; Bates, R. R.; Kennedy, D., eds. Air pollution, the automobile, and public
11 health. Washington, DC: National Academy Press; pp. 239-298.
12 Schlesinger, R. B.; Ben-Jebria, A.; Dahl, A. R.; Snipes, M. B.; Ultman, J. (1997) Disposition of inhaled toxicants.
13 In: Massaro, E. J., ed. Handbook of human toxicology. Boca Raton, FL: CRC Press; pp. 493-550.
14 Schwartz, J. (1997) Air pollution and hospital admissions for cardiovascular disease in Tucson. Epidemiology
15 8:371-377.
16 Schwartz, J. (1999) Air pollution and hospital admissions for heart disease in eight U.S. counties. Epidemiology
17 10: 17-22.
18 Schwartz, J. (2000) Harvesting and long term exposure effects in the relation between air pollution and mortality.
19 Am. J. Epidemiol. 151: 440-448.
20 Schwartz, J.; Morris, R. (1995) Air pollution and hospital admissions for cardiovascular disease in Detroit,
21 Michigan. Am. J. Epidemiol. 142: 23-35.
22 Schwartz, J.; Neas, L. M. (2000) Fine particles are more strongly associated than coarse particles with acute
23 respiratory health effects in schoolchildren. Epidemiology. 11: 6-10.
24 Schwartz, J.; Zanobetti, A. (2000) Using meta-smoothing to estimate dose-response trends across multiple studies,
25 with application to air pollution and daily death. Epidemiology 11: 666-672.
26 Schwartz, J.; Dockery, D. W.; Neas, L. M.; Wypij, D.; Ware, J. H.; Spengler, J. D.; Koutrakis, P.; Speizer, F. E.;
27 Ferris, B. G., Jr. (1994) Acute effects of summer air pollution on respiratory symptom reporting in
28 children. Am. J. Respir. Crit. Care Med. 150: 1234-1242.
29 Schwartz, J.; Dockery, D. W.; Neas, L. M. (1996) Is daily mortality associated specifically with fine particles?
30 J. Air Waste Manage. Assoc. 46: 927-939.
31 Schwartz, J.; Norris, G.; Larson, T.; Sheppard, L.; Claiborne, C,; Koenig, J. (1999) Episodes of high coarse particle
32 concentrations are not associated with increased mortality. Environ. Health Perspect. 107: 339-342.
33 Sheppard, L.; Levy, D.; Norris, G.; Larson, T. V.; Koemg, J. Q. (1999) Effects of ambient air pollution on
34 nonelderly asthma hospital admissions in Seattle, Washington, 1987-1994. Epidemiology 10: 23-30.
35 Singer, D. H.; Martin, G. J.; Magid, N.; Weiss, J. S.; Schaad, J. W.; Kehoe, R.; Zheutlin, T; Fintel, D. J.; Hsieh,
36 A. M.; Lesch, M. (1988) Low heart rate variability and sudden cardiac death. J. Electrocardiol.
37 21(suppl.): S46-S55.
38 Smiley-Jewell, S. M.; Liu, F. J.; Weir, A. J.; Plopper, C. G. (2000) Acute injury to differentiating Clara cells in
39 neonatal rabbits results in age-related failure of bronchiolar epithelial repair. Toxicol. Pathol. 28: 267-276.
40 Smith, R. L.; Spitzner, D.; Kim, Y.; Fuentes, M. (2000) Threshold dependence of mortality effects for fine and
41 coarse particles in Phoenix, Arizona. J. Air Waste Manage. Assoc. 50: 1367-1379.
42 Stieb, D. M.; Bevendge, R. C.; Brook, J. R.; Smith-Doiron, M.; Burnett, R. T.; Dales, R. E.; Beaulieu, S.; Judek, S.;
43 Mamedov, A. (2000) Air pollution, aeroallergens and cardiorespiratory emergency department visits in
44 Saint John, Canada. J. Exposure Anal. Environ. Epidemiol. 10: 461-477.
45 Task Force of the European Society of Cardiology and the North American Society of Pacing and
46 Electrophysiology. (1996) Heart rate variability: standards of measurement, physiological interpretation
47 and clinical use. Circulation 93: 1043-1065.
48 Terashima, T.; Wiggs, B.; English, D.; Hogg, J. C.; Van Eeden, S. F. (1997) Phagocytosis of small carbon particles
49 (PM]0) by alveolar macrophages stimulates the release of polymorphonuclear leukocytes from bone
50 marrow. Am. J. Respir. Crit. Care Med. 155: 1441-1447.
51 Thurston, G. D.; Ito, K.; Kinney, P. L.; Lippmann, M. (1992) A multi-year study of air pollution and respiratory
52 hospital admissions in three New York State metropolitan areas: results for 1988 and 1989 summers.
53 J. Exposure Anal. Environ. Epidemiol. 2: 429-450.
March 2001 9-113 DRAFT-DO NOT QUOTE OR CITE
-------
1 Thurston, G. D.; Ito, K.; Hayes, C. G.; Bates, D. V.; Lippmann, M. (1994) Respiratory hospital admissions and
2 summertime haze air pollution in Toronto, Ontario: consideration of the role of acid aerosols. Environ.
3 Res. 65:271-290.
4 Thurston, G. D.; Lippmann, M.; Scott, M. B.; Fine, J. M. (1997) Summertime haze air pollution and children with
5 asthma. Am. J. Respir. Crit. Care Med. 155: 654-660.
6 Tiittanen, P.; Timonen, K. L.; Ruuskanen, J.; Mirme, A.; Pekkanen, J. (1999) Fine particulate air pollution,
7 resuspended road dust and respiratory health among symptomatic children. Eur. Respir. J. 13: 266-273.
8 Tolbert, P. E.; Klein, M.; Metzger, K. B.; Peel, J.; Flanders, W. D.; Todd, K.; Mulholland, J. A.; Ryan, P. B.;
9 Frumkin, H. (2000) Interim results of the study of particulates and health in Atlanta (SOPHIA). J.
10 Exposure Anal. Environ. Epidemiol. 10: 446-460.
11 Tsai, F. C.; Apte, M. G.; Daisey, J. M. (2000) An exploratory analysis of the relationship between mortality and the
12 chemical composition of airborne particulate matter. Inhalation Toxicol. 12(suppl.): 121-135.
13 Turpin, B. J. (1999) Options for characterizing organic particulate matter. Environ. Sci. Technol. 33: 76A-79A.
14 U.S. Environmental Protection Agency. (1996) Air quality criteria for particulate matter. Research Triangle Park,
15 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
16 U.S. Environmental Protection Agency. (2000a) Air quality criteria for carbon monoxide. Research Triangle Park,
17 NC: National Center for Environmental Assessment; report no. EPA/600/P-99/001 F. Available:
18 www.epa.gov/ncea/co/ [2000, October 6].
19 U.S. Environmental Protection Agency. (2000b) Health assessment document for diesel exhaust [external review
20 draft]. Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental
21 Criteria and Assessment Office; report nos. EPA/600/8-90/057E.
22 Vedal, S.; Petkau, J.; White, R.; Blair, J. (1998) Acute effects of ambient inhalable particles in asthmatic and
23 nonasthmatic children. Am. J. Respir. Crit. Care Med. 157: 1034-1043.
24 Ware, J. H.; Ferris, B. G., Jr.; Dockery, D. W.; Spengler, J. D.; Stram, D. O.; Speizer, F. E. (1986) Effects of
25 ambient sulfur oxides and suspended particles on respiratory health of preadolescent children. Am. Rev.
26 Respir. Dis. 133:834-842.
27 Watkinson, W. P.; Campen, M. J.; Costa, D. L. (1998) Cardiac arrhythmia induction after exposure to residual oil
28 fly ash particles in a rodent model of pulmonary hypertension. Toxicol. Sci. 41: 209-216.
29 Watkinson, W. P.; Campen, M. J.; Dreher, K. L.; Su, W.-Y.; Kodavanti, U.P.; Highfill, J. W.; Costa, D. L. (2000)
30 Thermoregulatory effects following exposure to particulate matter in healthy and
31 cardiopulmonary-compromised rats. J. Therm. Biol. 25: 131-137.
32 Weschler, C. J.; Shields, H. C. (1999) Indoor ozone/terpene reactions as a source of indoor particles. Atmos.
33 Environ. 33:2301-2312.
34 Whitby, K. T. (1978) The physical characteristics of sulfur aerosols. Atmos. Environ. 12: 135-159.
35 Whitby, K. T.; Sverdrup, G. M. (1980) California aerosols: their physical and chemical characteristics. In: Hidy,
36 G. M.; Mueller, P. K.; Grosjean, D.; Appel, B. R.; Wesolowski, J. J., eds. The character and origins of
37 smog aerosols: a digest of results from the California Aerosol Characterization Experiment (ACHEX).
38 New York, NY: John Wiley & Sons, Inc.; pp. 477-517. (Advances in environmental science and
39 technology: v. 9).
40 Wichmann, H.-E.; Spix, C.; Tuch, T.; Wolke, G.; Peters, A.; Heinrich, J.; Kreyling, W. G.; Heyder, J. (2000) Daily
41 mortality and fine and ultrafine particles in Erfurt, Germany. Part I: role of particle number and particle
42 mass. Cambridge, MA: Health Effects Institute; Research Report no. 98.
43 Williams, R.; Suggs, J.; Zweidinger, R.; Evans, G.; Creason, J.; Kwok, R.; Rodes, C.; Lawless, P.; Sheldon, L.
44 (2000a) The 1998 Baltimore Particulate Matter Epidemiology-Exposure Study: part 1. Comparison of
45 ambient, residential outdoor, indoor and apartment particulate matter monitoring. J. Exposure Anal.
46 Environ. Epidemiol. 10:518-532.
47 Williams, R.; Suggs, J.; Creason, J.; Rodes, C.; Lawless, P.; Kwok, R.; Zweidinger, R.; Sheldon, L. (2000b) The
48 1998 Baltimore Particulate Matter Epidemiology-Exposure Study: part 2. Personal exposure assessment
49 associated with an elderly study population. J. Exposure Anal. Environ. Epidemiol. 10: 533-543.
50 Williams, R. W.; Creason, J.; Zweidinger, R.; Watts, R. R.; Shy, C. (2000c) Indoor, outdoor, and personal exposure
51 monitoring of particulate air pollution: the Baltimore elderly epidemiology-exposure pilot study.
52 Atmos. Environ. 34: 4193-4204.
53 Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
54 epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
March 2001 9-114 DRAFT-DO NOT QUOTE OR CITE
-------
1 Woodruff, T. J.; Grille, J.; Schoendorf, K. C. (1997) The relationship between selected causes of postneonatal
2 infant mortality and particulate air pollution in the United States. Environ. Health Perspect. 105: 608-612.
3 Zanobetti, A.; Schwartz, J. (2000) Race, gender, and social status as modifiers of the effects of PM,0 on mortality.
4 J. Occup. Environ. Med. 42: 469-474.
5 Zhang, T.; Huang, C.; Johns, E. J. (1997) Neural regulation of kidney function by the somatosensory system in
6 normotensive and hypertensive rats. Am. J. Physiol. 273: R1749-R1757.
7 Zhang, H.; Triche, E.; Leaderer, B. (2000) Model for the analysis of binary time series of respiratory symptoms.
8 Am. J. Epidemiol. 151: 1206-1215.
9
March 2001 9-115 DRAFT-DO NOT QUOTE OR CITE
-------
-------
APPENDIX 9A
Key Quantitative Estimates of Relative Risk for Particulate Matter-Related
Health Effects Based on Epidemiologic Studies of North American Cities
Assessed in the 1996 Particulate Matter Air Quality Criteria Document
March 2001 9A-1 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9A-1. EFFECT ESTIMATES PER 50-^g/m3 INCREASE
IN 24-HOUR PM,,, CONCENTRATIONS FROM U.S. AND CANADIAN STUDIES
Study Location
RR (±CI)
Only PM
in Model
RR (±CI) Reported
Other Pollutants PM10 Levels
in Model Mean (Min/Max)f
Increased Total Acute Mortality
Six Cities2
Portage, WI
Boston, MA
Topeka, KS
St. Louis, MO
Kingston/Knoxville, TN
Steubenville, OH
St. Louis, MOC
Kingston, TNC
Chicago, ILh
Chicago, 1L8
Utah Valley, UTb
Birmingham, ALd
Los Angeles, CAr
Increased Hospital Admissions (for
Respiratory Disease
Toronto, Canada'
Tacoma, WA'
New Haven, CTJ
Cleveland, OHk
Spokane, WA1
COPD
Minneapolis, MNn
Birmingham, ALm
Spokane, WA1
Detroit, MI°
1.04(0.98, 1
1.06(1.04,1
0.98 (0.90, 1
1.03(1.00, 1
1.05(1.00, 1
1.05(1.00, 1
1.08(1.01, 1
1.09(0.94, 1
1.04(1.00, 1
1.03(1.02, 1
1.08(1.05,1
1.05(1.01, 1
1.03(1.00, 1
Elderly > 65 years)
1.23(1.02, 1
1.10(1.03,1
1.06(1.00, 1
1.06(1.00, 1
1.08(1.04,1
1.25(1.10, 1
1.13(1.04, 1
1.17(1.08,1
1.10(1.02, 1
.09)
.09)
.05)
.05)
.09)
.08)
.12)
.25)
.08)
.04)
.11)
.10)
.055)
.43){
.17)
.13)
.11)
.14)
.44)
.22)
.27)
.17)
—
— 18 (±11.7)
— 24 (±12.8)
— 27 (±16.1)
— 31 (±16.2)
— 32 (±14.5)
— 46 (±32.3)
1.06(0.98,1.15) 28(1/97)
1.09(0.94,1.26 30(4/67)
— 37 (4/365)
1.02(1.01,1.04) 38(NR/128)
1.19(0.96,1.47) 47(11/297)
— 48(21,80)
1.02(0.99,1.036) 58(15/177)
1.12(0.88, 1.36): 30-39*
1.11(1.02,1.20) 37(14,67)
1.07(1.01,1.14) 41(19,67)
— 43(19,72)
— 46(16,83)
— 36(18,58)
— 45(19,77)
— 46(16,83)
— 48 (22, 82)
March 2001
9A-2
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9A-1 (cont'd). EFFECT ESTIMATES PER 50-^g/m3 INCREASE
IN 24-HOUR PM1(1 CONCENTRATIONS FROM U.S. AND CANADIAN STUDIES
Study Location
Pneumonia
Minneapolis, MN"
Birmingham, ALm
Spokane, WA'
Detroit, MP
Ischemic HP
Detroit, MP
RR (±C1)
Only PM
in Model
1.08(1.01, 1.15)
1.09(1.03,1.15)
1.06(0.98,1.13)
—
1.02(1.01,1.03)
RR (±CI) Reported
Other Pollutants PM10 Levels
in Model Mean (Min/Max)*
— 36(18,58)
— 45 (19, 77)
— 46(16,83)
1.06(1.02,1.10) 48(22,82)
1.02(1.00,1.03) 48(22,82)
Increased Respiratory Symptoms
Lower Respiratory
Six Cities"
Utah Valley, UT
Utah Valley, UTS
Cough
Denver, CO"
Six Cities"
Utah Valley, UTS
Decrease in Lung Function
Utah Valley, UTr
Utah Valley, UTS
Utah Valley, UT™
2.03(1.36,3.04)
1.28(1.06, 1.56)1
1.01 (0.81, 1.27)"
1.27(1.08, 1.49)
1.09(0.57,2.10)
1.51 (1.12,2.05)
1.29(1.12, 1.48)
55 (24, 86)**
30(10,50)**
29(7,51)*"
Similar RR 30(13,53)
— 46(11/195)
— 76(7/251)
— 22 (0.5/73)
Similar RR 30(13,53)
— 76(7/251)
— 46(11/195)
— 76(7/251)
— 55(1,181)
References
"Schwartz et al (I996a).
"Popeetal (1992, 1994)/O,
cDockeryetal. (1992)/O3.
'Schwartz (1993).
"Ito and Thurston (1996)/O,.
'Kinney et al. (1995)/O3, CO.
hStyeretal. (1995).
'Thurston et al. (1994)/O3.
'Schwartz (1995)/SO2.
"Schwartz et al (1996b)
'Schwartz (1996)
"Schwartz (1994a)
"Schwartz (1994b)
"Schwartz (1994c).
"Schwartz and Moms (1995)/O3, CO, SO,.
'Schwartz etal (1994)
'Popeetal. (1991).
'Pope and Dockery (1992)
'Schwartz (1994d).
"Pope and Kanner (1993).
"Ostroetal. (1991).
1Min/Max 24-h PMIO in parentheses unless noted
otherwise as standard deviation (±SD), 10 and
90percentile(lO, 90). NR = not reported
TChildren
"Asthmatic children and adults
"Means of several cities
"PEFR decrease in mL/s.
"*FEV, decrease.
!RR refers to total population, not just >65 years
March 2001
9A-3
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9A-2. EFFECT ESTIMATES PER VARIABLE INCREMENTS IN 24-HOUR
CONCENTRATIONS OF FINE PARTICLE INDICATORS (PM2 5, SO^, H+)
FROM U.S. AND CANADIAN STUDIES
Acute Mortality
Six City3
Portage, WI
Topeka, KS
Boston, MA
St. Louis, MO
Kingston/Knoxville, TN
Steubenville, OH
Indicator
PM25
PM25
PM25
PM25
PM25
PM2,
RR (±CI) per 25 //g/m3
PM Increase
1.030(0.993, 1.071)
1.020(0.951,1.092)
1.056(1.038, 1.0711)
1.028(1.010,1.043)
1.035(1.005,1.066)
1.025(0.998, 1.053)
Reported PM
Levels Mean
(Min/Max)*
11. 2 (±7.8)
12.2 (±7.4)
15. 7 (±9.2)
18.7 (±10.5)
20.8 (±9.6)
29.6 (±21. 9)
Increased Hospitalization
Ontario, Canada6
Ontario, Canadac
NYC/Buffalo, NYd
so;
so;
03
so:
Torontod H+ (Nmol/m3)
so:
PM2,
1.03(1.02, 1.04)
1.03(1.02, 1.04)
1.03(1.02, 1.05)
1.05(1.01, 1.10)
1.16(1.03, 1.30)*
1.12(1.00, 1.24)
1.15(1.02, 1.78)
R = 3.1-8.2
R = 2.0-7.7
NR
28.8 (NR/391)
7.6 (NR, 48.7)
1 8.6 (NR, 66.0)
Increased Respiratory Symptoms
Southern Californiaf
Six Cities8
(Cough)
Six Cities8
(Lower Resp. Symp.)
so:
PM25
PM2 5 Sulfur
H+
PM25
PM2 5 Sulfur
H+
1.48(1.14, 1.91)
1.19(1.01, 1.42)**
1.23(0.95, 1.59)**
1.06(0.87, 1.29)"
1.44(1.15-1.82)**
1.82(1.28-2.59)**
1.05(0.25-1.30)**
R = 2-37
18.0(7.2,37)***
2.5(3.1,61)***
18.1 (0.8,5.9)***
18.0(7.2,37)***
2.5 (0.8, 5.9)***
18.1 (3.1,61)***
Decreased Lung Function
Uniontown, PAC
PM25
PEFR 23.1 (-0.3, 36.9) (per 25 //g/m3)
25/88 (NR/88)
References:
'Schwartz etal.(1996a).
bBurnettetal. (1994).
'Burnett et al. (1995) O3.
"Thurston et al. (1992, 1994).
"Neasetal. (1995).
fOstroetal.(1993).
gSchwartz et al. (1994).
'Min/Max 24-h PM indicator level shown in parentheses unless
otherwise noted as (±SD), 10 and 90 percentile (10,90) or
R = range of values from min-max, no mean value reported.
'Change per 100 nmoles/m3
"Change per 20 //g/m3 for PM2 5; per 5 /ug/m3 for PM2 5 sulfur;
per 25 nmoles/m3 for H+.
"*50th percentile value (10,90 percentile).
March 2001
9A-4
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 9A-3. EFFECT ESTIMATES PER INCREMENTS3 IN
ANNUAL MEAN LEVELS OF FINE PARTICLE INDICATORS FROM
U.S. AND CANADIAN STUDIES
Type of Health
Effect and Location
Indicator
Increased Total Chronic Mortality in Adults
Six City6
ACS Study'
(151U.S. SMSA)
Increased Bronchitis
Six Cityd
Six Cityc
24 Cityf
24 Cityf
24 Cityf
24 Cityf
Southern California8
PM15/IO
PM25
so:
PM25
so:
in Children
PM15/10
TSP
H+
so:
PM2I
PM10
so:
Change in Health Indicator per
Increment in PMa
Relative Risk (95% CI)
1.42(1.16-2.01)
1.31 (1.11-1.68)
1.46(1.16-2.16)
1.17(1.09-1.26)
1.10(1.06-1.16)
Odds Ratio (95% CI)
3.26(1.13, 10.28)
2.80(1.17,7.03)
2.65(1.22,5.74)
3.02(1.28,7.03)
1.97(0.85,4.51)
3.29(0.81, 13.62)
1.39(0.99, 1.92)
Range of City
PM Levels
Means (,ug/m3)
18-47
11-30
5-13
9-34
4-24
20-59
39-114
6.2-41.0
18.1-67.3
9.1-17.3
22.0-28.6
—
Decreased Lung Function in Children
Six City"
Six Cityc
24 CityIJ
24 City'
24 City1
24 City'
PMI5,10
TSP
H+ (52 nmoles/m3)
PM2,(15^g/m3)
SO: (7 /"g/m3)
PM10(17^g/m3)
NS Changes
NS Changes
-3.45% (-4.87, -2.01) FVC
-3.21% (-4.98, -1.41) FVC
-3. 06% (-4.50, -1.60) FVC
-2. 42% (-4.30, -.0.51) FVC
20-59
39-114
—
—
—
—
aEstimates calculated annual-average PM increments assume: a 100-/^g/m3 increase for TSP; a 50-^g/m3
increase for PM]0 and PM]5; a 25-jUg/m3 increase for PM2 5; and a 15-yUg/m3 increase for SO:, except where
noted otherwise; a 100-nmole/m3 increase for H+.
"Dockery et al. (1993).
Tope etal. (1995).
dDockeryetal.(1989).
'Wareetal. (1986).
TJockery etal. (1996).
gAbbeyetal. (1995).
hNS Changes = No significant changes.
'Raizenne etal. (1996).
^Pollutant data same as for Dockery et al. (1996).
March 2001
9A-5
DRAFT-DO NOT QUOTE OR CITE
-------
REFERENCES
Abbey, D. E.; Ostro, B. E.; Petersen, F.; Burchette, R. J. (1995) Chronic respiratory symptoms associated with
estimated long-term ambient concentrations of fine participates less than 2.5 microns in aerodynamic
diameter (PM25) and other air pollutants. J. Exposure Anal. Environ. Epidemiol. 5: 137-159.
Burnett, R. T.; Dales, R. E.; Raizenne, M. E.; Krewski, D.; Summers, P. W.; Roberts, G. R.; Raad-Young, M;
Dann, T.; Brook, J. (1994) Effects of low ambient levels of ozone and sulfates on the frequency of
respiratory admissions to Ontario hospitals. Environ. Res. 65: 172-194.
Burnett, R. T.; Dales, R.; Krewski, D.; Vincent, R.; Dann, T.; Brook, J. R. (1995) Associations between ambient
particulate sulfate and admissions to Ontario hospitals for cardiac and respiratory diseases. Am. J. Epidemiol.
142: 15-22.
Dockery, D. W.; Speizer, F. E.; Stram, D. O.; Ware, J. H.; Spengler, J. D.; Ferris, B. G., Jr. (1989) Effects of
inhalable particles on respiratory health of children. Am. Rev. Respir. Dis. 139: 587-594.
Dockery, D. W.; Schwartz, J.; Spengler, J. D. (1992) Air pollution and daily mortality: associations with particulates
and acid aerosols. Environ. Res. 59: 362-373.
Dockery, D. W.; Pope, C. A., Ill; Xu, X.; Spengler, J. D.; Ware, J. H.; Fay, M. E.; Ferris, B. G., Jr.; Speizer, F. E.
(1993) An association between air pollution and mortality in six U.S. cities. N. Engl. J. Med.
329: 1753-1759.
Dockery, D. W.; Cunningham, J.; Damokosh, A. I.; Neas, L. M.; Spengler, J. D.; Koutrakis, P.; Ware, J. H.;
Raizenne, M.; Speizer, F. E. (1996) Health effects of acid aerosols on North American children: respiratory
symptoms. Environ. Health Perspect. 104: 500-505.
Ito, K.; Thurston, G. D. (1996) Daily PM,(/mortality associations: an investigation of at-risk subpopulations. J.
Exposure Anal. Environ. Epidemiol. 6: 79-95.
Kinney, P. L.; Ito, K.; Thurston, G. D. (1995) A sensitivity analysis of mortality/PM10 associations in Los Angeles.
In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on particulate air pollution and human
mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 59-69.
Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Tollerud, D. J.; Speizer, F. E. (1995) The association of ambient air
pollution with twice daily peak expiratory flow rate measurements in children. Am. J. Epidemiol.
141: 111-122.
Ostro, B. D.; Lipsett, M. J.; Wiener, M. B.; Seiner, J. C. (1991) Asthmatic responses to airborne acid aerosols. Am.
J. Public Health 81: 694-702.
Ostro, B. D.; Lipsett, M. J.; Mann, J. K.; Krupnick, A.; Harrington, W. (1993) Air pollution and respiratory
morbidity among adults in Southern California. Am. J. Epidemiol. 137: 691-700.
Pope, C. A., Ill; Dockery, D. W. (1992) Acute health effects of PM10 pollution on symptomatic and asymptomatic
children. Am. Rev. Respir. Dis. 145: 1123-1128.
Pope, C. A., Ill; Kanner, R. E. (1993) Acute effects of PMIO pollution on pulmonary function of smokers with mild
to moderate chronic obstructive pulmonary disease. Am. Rev. Respir. Dis. 147: 1336-1340.
Pope, C. A., Ill; Dockery, D. W.; Spengler, J. D.; Raizenne, M. E. (1991) Respiratory health and PM,0 pollution:
a daily time series analysis. Am. Rev. Respir. Dis. 144: 668-674.
Pope, C. A., Ill; Schwartz, J.; Ransom, M. R. (1992) Daily mortality and PM10 pollution in Utah valley.
Arch. Environ. Health 47: 211-217.
Pope, C. A., Ill; Thun, M. J.; Namboodiri, M. M.; Dockery, D. W.; Evans, J. S.; Speizer, F. E.; Heath, C. W., Jr.
(1995) Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am. J.
Respir. Crit. Care Med. 151: 669-674.
Raizenne, M.; Neas, L. M.; Damokosh, A. I.; Dockery, D. W.; Spengler, J. D.; Koutrakis, P.; Ware, J. H.; Speizer,
F. E. (1996) Health effects of acid aerosols on North American children: pulmonary function. Environ.
Health Perspect. 104: 506-514.
Schwartz, J. (1993) Air pollution and daily mortality in Birmingham, Alabama. Am. J. Epidemiol. 137: 1136-1147.
Schwartz, J. (1994a) Air pollution and hospital admissions for the elderly in Birmingham, Alabama. Am. J.
Epidemiol. 139:589-598.
Schwartz, J. (1994b) PM,0, ozone, and hospital admissions for the elderly in Minneapolis, MN. Arch. Environ.
Health 49: 366-374.
Schwartz, J. (1994c) Air pollution and hospital admissions for the elderly in Detroit, Michigan. Am. J. Respir. Crit.
Care Med. 150:648-655.
March 2001 9A-6 DRAFT-DO NOT QUOTE OR CITE
-------
Schwartz, J. (1994d) Nonparametric smoothing in the analysis of air pollution and respiratory illness. Can. J. Stat.
22: 1-17.
Schwartz, J. (1995) Short term fluctuations in air pollution and hospital admissions of the elderly for respiratory
disease. Thorax 50: 531-538.
Schwartz, J. (1996) Air pollution and hospital admissions for respiratory disease. Epidemiology 7: 20-28.
Schwartz, J.; Morris, R. (1995) Air pollution and hospital admissions for cardiovascular disease in Detroit,
Michigan. Am. J. Epidemiol. 142: 23-35.
Schwartz, J.; Dockery, D. W.; Neas, L. M.; Wypij, D.; Ware, J. H.; Spengler, J. D.; Koutrakis, P.; Speizer, F. E.;
Ferris, B. G., Jr. (1994) Acute effects of summer air pollution on respiratory symptom reporting in children.
Am. J. Respir. Crit. Care Med. 150: 1234-1242.
Schwartz, J.; Dockery, D. W.; Neas, L. M. (1996a) Is daily mortality associated specifically with fine particles?
J. Air Waste Manage. Assoc. 46: 927-939.
Schwartz, J.; Spix, C.; Touloumi, G.; Bacharova, L.; Barumamdzadeh, T.; le Tertre, A.; Piekarksi, T.;
Ponce de Leon, A.; Ponka, A.; Rossi, G.; Saez, M.; Schouten, J. P. (1996b) Methodological issues in studies
of air pollution and daily counts of deaths or hospital admissions. In: St Leger, S., ed. The APHEA project.
Short term effects of air pollution on health: a European approach using epidemiological time series data. J.
Epidemiol. Community Health 50(suppl. 1): S3-S11.
Styer, P.; McMillan, N.; Gao, F.; Davis, J.; Sacks, J. (1995) Effect of outdoor airborne particulate matter on daily
death counts. Environ. Health Perspect. 103: 490-497.
Thurston, G. D.; Ito, K.; Kinney, P. L.; Lippmann, M. (1992) A multi-year study of air pollution and respiratory
hospital admissions in three New York State metropolitan areas: results for 1988 and 1989 summers.
J. Exposure Anal. Environ. Epidemiol. 2: 429-450.
Thurston, G. D.; Ito, K.; Hayes, C. G.; Bates, D. V.; Lippmann, M. (1994) Respiratory hospital admissions and
summertime haze air pollution in Toronto, Ontario: consideration of the role of acid aerosols. Environ. Res.
65:271-290.
Ware, J. H.; Ferris, B. G., Jr.; Dockery, D. W.; Spengler, J. D.; Stram, D. O.; 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.
March 2001 9A-7 DRAFT-DO NOT QUOTE OR CITE
-------
-------
i EXECUTIVE SUMMARY
2
3 Air Quality Criteria for Participate Matter
4 (March 2000)
5
6
7 The Executive Summary of this Second External Review Draft of the Air Quality Criteria
8 for Particulate Matter is under preparation and will be completed following public comment and
9 CAS AC review of the earlier, more detailed chapters of this Second External Review Draft.
10 It will be included in subsequent drafts of the document.
March 2001 E-1 DRAFT-DO NOT QUOTE OR CITE
-------