chapter 8 epidemiology of human health effects associated with ambient particulate matter

368 Pages·2004·2.08 MB·English

CHAPTER 8
EPIDEMIOLOGY OF HUMAN HEALTH EFFECTS
ASSOCIATED WITH AMBIENT
PARTICULATE MATTER
Table of Contents
Page
8. EPIDEMIOLOGY OF HUMAN HEALTH EFFECTS ASSOCIATED
WITH AMBIENT PARTICULATE MATTER .............................. 8-1
8.1 INTRODUCTION ................................................. 8-1
8.1.1 Approaches for Identifying, Presenting, and Assessing Studies ....... 8-2
8.1.2 Types of Epidemiologic Studies Reviewed ....................... 8-5
8.1.3 Overview of Key Methodological Issues ........................ 8-8
8.1.3.1 Issues Related to Use of General Additive Models
(GAM) in PM Epidemiology ........................ 8-8
8.1.3.2 Confounding and Effect Modification ................. 8-9
8.1.4 Approach to Assessing Epidemiologic Evidence ................. 8-15
8.2 MORTALITY EFFECTS ASSOCIATED WITH AIRBORNE
PARTICULATE MATTER EXPOSURE .............................. 8-18
8.2.1 Introduction .............................................. 8-18
8.2.2 Mortality Effects of Short-Term Particulate Matter Exposure ....... 8-18
8.2.2.1 Summary of 1996 Particulate Matter Criteria
Document Findings and Key Issues .................. 8-18
8.2.2.2 Newly Available Information on Short-Term
Mortality Effects ................................. 8-23
8.2.2.3 New Multi-City Studies ........................... 8-29
8.2.2.4 U.S. Single-City Studies ........................... 8-49
8.2.2.5 The Role of Particulate Matter Components ............ 8-54
8.2.2.6 New Assessments of Cause-Specific Mortality ......... 8-74
8.2.2.7 Salient Points Derived from Assessment of Studies
of Short-Term Particulate Matter Exposure Effects
on Mortality .................................... 8-80
8.2.3 Mortality Effects of Long-Term Exposure to Ambient
Particulate Matter ......................................... 8-83
8.2.3.1 Studies Published Prior to the 1996 Particulate Matter
Criteria Document ................................ 8-83
8.2.3.2 New Prospective Cohort Analyses of Mortality Related
to Chronic Particulate Matter Exposures .............. 8-87
8.2.3.3 Studies by Particulate Matter Size-Fraction and
Composition ................................... 8-116
8.2.3.4 Recent PM-Mortality Intervention Studies ............ 8-125
8.2.3.5 Salient Points Derived from Analyses of Chronic
Particulate Matter Exposure Mortality Effects ......... 8-130
8.3 MORBIDITY EFFECTS OF PARTICULATE MATTER EXPOSURE ..... 8-134
8.3.1 Cardiovascular Morbidity Effects Associated with Acute
Ambient Particulate Matter Exposure ......................... 8-134
8.3.1.1 Introduction .................................... 8-134
8.3.1.2 Summary of Key Findings on Cardiovascular
Morbidity from the 1996 Particulate Matter Air
Quality Criteria Document ........................ 8-135
ii
Table of Contents
(cont’d)
Page
8.3.1.3 New Particulate Matter-Cardiovascular Morbidity
Studies ........................................ 8-135
8.3.1.4 Issues in the Interpretation of Acute Cardiovascular
Effects Studies .................................. 8-164
8.3.2 Effects of Short-Term Particulate Matter Exposure on the
Incidence of Respiratory-Related Hospital Admissions and
Medical Visits ........................................... 8-165
8.3.2.1 Introduction .................................... 8-165
8.3.2.2 Summary of Key Respiratory Hospital Admissions
Findings from the 1996 Particulate Matter Air Quality
Criteria Document ............................... 8-166
8.3.2.3 New Respiratory-Related Hospital Admissions
Studies ........................................ 8-167
8.3.2.4 Key New Respiratory Medical Visits Studies .......... 8-182
8.3.2.5 Identification of Potential Susceptible
Subpopulations ................................. 8-184
8.3.2.6 Summary of Salient Findings on Acute Particulate
Matter Exposure and Respiratory-Related Hospital
Admissions and Medical Visits ..................... 8-186
8.3.3 Effects of Particulate Matter Exposure on Lung Function and
Respiratory Symptoms .................................... 8-187
8.3.3.1 Effects of Short-Term Particulate Matter Exposure
on Lung Function and Respiratory Symptoms ......... 8-188
8.3.3.2 Long-Term Particulate Matter Exposure Effects
on Lung Function and Respiratory Symptoms ......... 8-205
8.3.4 Ambient PM Impacts on Fetal and/or Early Postnatal
Development/Mortality .................................... 8-208
8.3.4.1 PM Effects on Intrauterine Fetal Morbidity/Mortality ... 8-209
8.3.4.2 PM Effects on Post-Neonatal Infant Mortality ......... 8-212
8.3.4.3 Summary of Saliant Points on PM Effects on Fetal
and/or Early Postnatal Development/Mortality ......... 8-214
8.4 INTERPRETIVE ASSESSMENT OF THE EPIDEMIOLOGIC
EVIDENCE .................................................... 8-215
8.4.1 Introduction ............................................. 8-215
8.4.2 GAM Issue and Reanalyses Studies .......................... 8-218
8.4.2.1 Impact of Using the More Stringent GAM Model on
PM Effect Estimates for Mortality .................. 8-219
8.4.2.2 Impact of Using the More Stringent GAM Model
on PM Effect Estimates for Respiratory Hospital
Admissions .................................... 8-223
8.4.2.3 HEI Commentaries .............................. 8-227
iii
Table of Contents
(cont’d)
Page
8.4.3 Assessment of Confounding by Co-Pollutants and Adjustments
for Meteorological Variables ............................... 8-229
8.4.3.1 Introduction to Assessment of Confounding by
Co-Pollutants ................................... 8-229
8.4.3.2 Statistical Issues in the Use of Multi-Pollutant
Models ........................................ 8-231
8.4.3.3 Multipollutant Modeling Outcomes ................. 8-236
8.4.3.4 Bioaerosols as Possible Confounders or Effect
Modifiers in PM Epidemiologic Studies .............. 8-245
8.4.3.5 Adjustments for Meteorological Variables ............ 8-246
8.4.4 The Question of Lags ..................................... 8-259
8.4.5 Measurement Error: Concepts and Consequences ............... 8-271
8.4.5.1 Theoretical Framework for Assessment of
Measurement Error .............................. 8-271
8.4.5.2 Measurement Error Issues Related to Divergence
Between Monitors and to Monitoring Frequency ....... 8-278
8.4.5.3 Measurement Error and the Assessment of Confounding
by Co-Pollutants in Multi-Pollutant Models ........... 8-288
8.4.6 Role of Particulate Matter Components ....................... 8-289
8.4.6.1 Thoracic Particle (PM ) Mortality/Morbidity Effects . . . 8-290
10
8.4.6.2 Fine and Coarse Fraction Particle Effects on
Mortality ...................................... 8-291
8.4.6.3 Source-Oriented Analyses of PM and Mortality ........ 8-295
8.4.6.4 Fine and Coarse Fraction Particle Effects on
Morbidity ...................................... 8-297
8.4.7 Concentration-Response Relationships for Ambient PM .......... 8-306
8.4.8 The Question of Heterogeneity of Particulate Matter Effects
Estimates ............................................... 8-310
8.4.8.1 Evaluation of Heterogeneity in Time-Series Studies .... 8-311
8.4.8.2 Comparison of Spatial Relationships in the NMMAPS
and Cohort Reanalyses Studies ..................... 8-314
8.4.9 Age-Related Differences in PM Effect Estimates ................ 8-315
8.4.10 Implications of Airborne Particle Mortality Effects .............. 8-316
8.4.10.1 Short-Term Exposure and Mortality Displacement ..... 8-316
8.4.10.2 Life-Shortening Estimates Based on Prospective
Cohort Study Results ............................. 8-322
8.4.10.3 Potential Effects of Infant Mortality on
Life-Shortening Estimates ......................... 8-322
8.5 SUMMARY OF KEY FINDINGS AND CONCLUSIONS DERIVED
FROM PARTICULATE MATTER EPIDEMIOLOGY STUDIES ......... 8-323
REFERENCES ....................................................... 8-336
iv
List of Tables
Number Page
8-1 Recent U.S. and Canadian Time-Series Studies of PM-Related Daily
Mortality ......................................................... 8-24
8-2 Synopsis of Short-Term Mortality Studies that Examined Relative
Importance of PM and PM ....................................... 8-56
2.5 10-2.5
8-3 Newly Available Studies of Mortality Relationships to PM Chemical
Components ....................................................... 8-67
8-4 Summary of Source-Oriented Evaluations of PM Components in Recent
Studies ........................................................... 8-71
8-5 Comparison of Six Cities and American Cancer Society Study Findings
from Original Investigators and Health Effects Institute Reanalysis ........... 8-89
8-6 Relative Risk of All-Cause Mortality for Selected Indices of Exposure to Fine
Particulate Matter (per 18.6 µg/m3) Based on Multivariate Poisson Regression
Analysis, by Age Group, for Harvard Six City Study Data .................. 8-93
8-7 Summary of Results from the Extended ACS Study ........................ 8-95
8-8 Relative Risk of Mortality from all Nonexternal Causes, by Sex and Air
Pollutant, for an Alternative Covariate Model in the ASHMOG Study ........ 8-102
8-9 Relative Risk of Mortality from Cardiopulmonary Causes, by Sex and
Air Pollutant, for an Alternative Covariate Model in the ASHMOG Study ..... 8-103
8-10 Relative Risk of Mortality from Lung Cancer by Air Pollutant and by
Gender for an Alternative Covariate Model ............................. 8-104
8-11 Particulate Matter Effects on Mortality by Exposure and Mortality Period
with Ecological Variables for the Veterans Cohort Study Expressed as
Excess Mortality .................................................. 8-109
8-12 Comparison of Excess Relative Risks of Long-Term Mortality in the
Harvard Six Cities, ACS, AHSMOG, and VA Studies ..................... 8-112
8-13 Comparison of Estimated Relative Risks for All-Cause Mortality in Six
U.S. Cities Associated with the Reported Inter-City Range of Concentrations
of Various Particulate Matter Metrics .................................. 8-117
8-14 Comparison of Reported SO = and PM Relative Risks for Various
4 2.5
Mortality Causes in the American Cancer Society Study ................... 8-118
v
List of Tables
(cont’d)
Number Page
8-15 Comparison of Total Mortality Relative Risk Estimates and T-Statistics
for Particulate Matter Components in Three Prospective Cohort Studies ...... 8-119
8-16 Comparison of Cardiopulmonary Mortality Relative Risk Estimates and
T-Statistics for Particulate Matter Components in Three Prospective
Cohort Studies .................................................... 8-120
8-17 Percent Attributable Risk of Mortality (from Lipfert and Morris, 2000)
and Risk Estimates Calculated Per 10 µg/m3 PM ....................... 8-123
2.5
8-18 Summary of Studies of PM , PM , or PM Effects on Total CVD
10 10-2.5 2.5
Hospital Admissions and Emergency Visits ............................. 8-137
8-19 Summary of United States PM Respiratory-Related Hospital Admission
10
Studies .......................................................... 8-168
8-20 Percent Increase in Hospital Admissions per 10-µg/m3 Increase in PM
10
in 14 U.S. Cities (original and reanalyzed results) ........................ 8-170
8-21 Summary of United States PM Respiratory-Related Hospital Admission
2.5
Studies .......................................................... 8-174
8-22 Summary of United States PM Respiratory-Related Hospital Admission
10-2.5
Studies .......................................................... 8-175
8-23 Intercomparison of Detroit Pneumonia Hospital Admission Relative
Risks (± 95% CI below) of PM Indices (per 5th-to-95th percentile pollutant
increment) for Various Model Specifications ............................ 8-176
8-24 Summary of United States PM , PM , and PM Asthma Medical
10 2.5 10-2.5
Visit Studies ..................................................... 8-182
8-25 Summary of Quantitative PFT Changes in Asthmatics per 50 µg/m3
PM Increment ................................................... 8-190
10
8-26 Summary of PFT Changes in Asthmatics per 25 µg/m3 PM Increment ....... 8-191
2.5
8-27 Summary of Asthma PM Cough Studies .............................. 8-194
10
8-28 Summary of Asthma PM Phlegm Studies ........ ...................... 8-195
10
8-29 Summary of Asthma PM Lower Respiratory Illness Studies ............... 8-195
10
vi
List of Tables
(cont’d)
Number Page
8-30 Summary of Asthma PM Bronchodilator Use Studies .................... 8-196
10
8-31 Summary of Asthma PM Respiratory Symptom Studies .................. 8-199
2.5
8-32 Summary of Non-Asthma PM PFT Studies ............................ 8-200
10
8-33 Summary of Non-Asthma PM Respiratory Symptom Studies .............. 8-201
10
8-34 Summary of Non-Asthma PM Respiratory Outcome Studies .............. 8-203
2.5
8-35 Summary of Non-Asthma Coarse Fraction Studies of Respiratory
Endpoints ........................................................ 8-204
8-36 PM Excess Risk Estimates from Reanalysis Studies for Total
10
Non-Accidental Mortality per 50 µg/m3 Increase in PM .................. 8-220
10
8-37 Comparison of Maximum Single Day Lag Effect Estimates for PM , PM ,
2.5 2.5-10
and PM for Seattle Asthma hospital Admissions Based on Original GAM
10
Analyses using Default Convergence Criteria Versus Reanalyses using
GAM with more Stringent Convergence Criteria and GLM ................. 8-225
8-38 Comparison of Los Angeles COPD Hospital Admissions Maximum Single
Day Lag Effect Estimates for PM and PM from the Original GAM
2.5 10
Analyses Using Default Convergence Criteria Versus Effect Estimates
Derived from Reanalyses using more Stringent Convergence Criteria and
for Models Smoothed with more Degrees of Freedom ..................... 8-226
8-39 Effects of Different Models for Weather and Time Trends on Mortality
in Utah Valley Study ............................................... 8-251
8-40 Summary Statistics Showing Mean Site-Pair Pearson Correlation Coefficients,
Annual Mean PM Concentrations (µg/m3), the Range in Annual Mean
2.5
Concentrations (µg/m3), Mean of 90th Percentile Differences in Concentrations
Between All Site Pairs (µg/m3), and Coefficients of Divergence for MSAs
Meeting Selection Criteria given in Appendix 3A .. ...................... 8-281
8-41 Summary of Relative Homogeneity / Heterogeneity Characteristics for
MSAs Given in Table 8-40 .......................................... 8-282
8-42 Summary of Past Ecologic and Case-Control Epidemiologic Studies of
Outdoor Air and Lung Cancer ........................................ 8-303
vii
List of Figures
Number Page
8-1 Estimated excess risks for PM mortality (1 day lag) for the 88 largest U.S.
cities as shown in the revised NMMAPS analysis ......................... 8-32
8-2 Map of the United States showing the 88 cities (the 20 cities are circled) and
the seven U.S. regions considered in the NMMAPS geographic analyses ....... 8-33
8-3 Percent excess mortality risk (lagged 0, 1, or 2 days) estimated in the
NMMAPS 90-City Study to be associated with 10-µg/m3 increases in PM
10
concentrations in cities aggregated within U.S. regions shown in Figure 8-4 .... 8-34
8-4 Marginal posterior distributions for effect of PM on total mortality at lag 1
10
with and without control for other pollutants, for the 90 cities ................ 8-35
8-5 Percent excess risks estimated per 25 µg/m3 increase in PM or PM
2.5 10-2.5
from new studies that evaluated both PM and PM , based on single
2.5 10-2.5
pollutant (PM only) models ........................................... 8-58
8-6 Excess risks estimated per 5 µg/m3 increase in sulfate, based on the studies
in which both PM and PM data were available ....................... 8-70
2.5 10-2.5
8-7 Natural logarithm of relative risk for total and cause-specific mortality per
10 µg/m3 PM (approximately the excess relative risk as a fraction), with
2.5
smoothed concentration-response functions .............................. 8-95
8-8 Relative risk of total and cause-specific mortality per 10 µg/m3 PM ,
2.5
derived for means of 1979-1983 PM data for various cities, using
2.5
alternative statistical models .......................................... 8-96
8-9 Relative risk of total and cause-specific mortality for particle metrics and
gaseous pollutants over different averaging periods (years 1979-2000 in
parentheses) ....................................................... 8-97
8-10 Acuate cardiovascular hospitalizations and particulate matter exposure excess
risk estimates derived from U.S. PM studies based on single-pollutant
10
models from GAM strict covergence criteria reanalyses (2003 studies) or
alternative (non-GAM) original analyses ............................... 8-153
8-11 Percent change in hospital admission rates and 95% CIs for an IQR increase
in pollutants from single-pollutant models for asthma ..................... 8-178
8-12 Maximum excess risk of respiratory-related hospital admissions and visits
per 50 µg/m3 PM increment in studies of U.S. cities based on single-
10
pollutant models .................................................. 8-187
viii
List of Figures
(cont’d)
Number Page
8-13 Illustrative acute pulmonary function change studies of asthmatic children .... 8-192
8-14 Odds ratios with 95% confidence interval for cough per 50-µg/m3 increase
in PM for illustrative asthmatic children studies at lag 0 .................. 8-197
10
8-15 PM excess risk estimates for total non-accidental mortality for numerous
10
locations (and for cardiovascular mortality[*] for Coachella Valley, CA
and Phoenix, AZ), using: (1) GAM with default convergence criteria
(white circle); (2) GAM with stringent convergence criteria (black circle);
and, (3) GLM/natural splines (x) that approximate the original GAM model
from the GAM reanalysis studies ..................................... 8-221
8-16 Excess risk estimates for total non-accidental mortality in single-pollutant
(PM only) and multi-pollutant models ................................. 8-238
8-17 Excess risk estimates for cardiovascular-related effects, including mortality,
hospital admissions, and changes in biomarkers (e.g., increases in blood
parameters or decreases in heart rate variability measures) in single-pollutant
(PM only) and multi-pollutant models ................................. 8-239
8-18 Excess risk estimates for respiratory-related effects, including mortality,
hospital admissions and medical visits in single-pollutant (PM only) and
multi-pollutant models ............................................. 8-240
8-19 Excess risk estimates for increases in respiratory symptoms or decreases
in lung function measures in single-pollutant (PM only) and multi-pollutant
models .......................................................... 8-241
8-20 PM (lag 1 day) coefficient ($) for total mortality, for 1992-1994, as a
10
function of alternative weather models and varying degrees of freedom
for fitting temporal trends using natural splines .......................... 8-255
8-21 PM (lag 1 day) coefficient ($) for hospital admissions for pneumonia
10
among the elderly, for 1992-1994, as a function of alternative weather
models and varying degrees of freedom for fitting temporal trends using
natural splines .................................................... 8-256
8-22 Odds ratios (and 95% confidence intervals) for associations between onset
of myocardial infarction and 25 µg/m3 increase in hourly (upper panel) or
daily 24-h average (bottom panel) PM concentrations ................... 8-260
2.5
ix
List of Figures
(cont’d)
Number Page
8-23 Marginal posterior distribution for effects of PM on all cause mortality at
10
lag 0, 1, and 2 for the 90 cities ....................................... 8-261
8-24 Excess risk estimates for associations between various health outcomes
and PM (50 µg/m3 increment) from different studies conducted in
10
Cook County, IL .................................................. 8-263
8-25 Excess risk estimates for associations between various health outcomes
and PM (50 µg/m3 increment) from studies conducted in Los Angeles
10
County, CA ...................................................... 8-264
8-26 Excess risk estimates for associations between various health outcomes
and PM (50 µg/m3 increment) from studies conducted in Pittsburgh, PA ..... 8-265
10
8-27 Excess risk estimates for associations between various health outcomes
and PM (50 µg/m3 increment) from studies conducted in Detroit, MI ........ 8-266
10
8-28 Excess risk estimates for associations between various health outcomes
and PM (50 µg/m3 increment) from studies conducted in Seattle or
10
King County, WA ................................................. 8-267
8-29 Relative risk estimates and 95% confidence intervals for total mortality per
100 µg/m3 increase in PM , adjusting for ozone, temperature, seasonal
10
cycles, day of week, and linear trend for 1985-1990 in Cook County, IL ...... 8-285
8-30 Concentration-response curves for PM mortality relationships in 20 largest
10
U.S. cities (1987-1994), for total (TOTAL) mortality, cardiovascular and
respiratory (CVDRESP) mortality, and other causes (Other) mortality ........ 8-309
8-31 Posterior probabilities of thresholds for each cause-specific mortality and
for mean PM , 20 largest U.S. cities, 1987-1994 ......................... 8-309
10
x