Reanalysis of the Harvard Six Cities Study, part I: validation and replication

Open science, the replication crisis, and environmental public health

Concerns about a crisis of mass irreplicability across scientific fields ("the replication crisis") have stimulated a movement for open science, encouraging or even requiring researchers to publish their raw data and analysis code. Recently, a rule at the US Environmental Protection Agency (US EPA) would have imposed a strong open data requirement. The rule prompted significant public discussion about whether open science practices are appropriate for fields of environmental public health. The aims of this paper are to assess (1) whether the replication crisis extends to fields of environmental public health; and (2) in general whether open science requirements can address the replication crisis. There is little empirical evidence for or against mass irreplicability in environmental public health specifically. Without such evidence, strong claims about whether the replication crisis extends to environmental public health - or not - seem premature. By distinguishing three concepts - reproducibility, replicability, and robustness - it is clear that open data initiatives can promote reproducibility and robustness but do little to promote replicability. I conclude by reviewing some of the other benefits of open science, and offer some suggestions for funding streams to mitigate the costs of adoption of open science practices in environmental public health.

Exposure levels of air pollution (PM2.5) and associated health risk in Kuwait

It is well established that respiratory and cardiovascular mortality and morbidity rates are associated with poor air quality as measured by high concentrations of fine particulate matter such as PM_2.5 parameters. Since such information is lacking for the State of Kuwait, this study examined the exposure levels of PM_2.5 and the associated health risk as evaluated by five mortality measures embodied in ischemic heart disease, stroke, lung cancer, chronic obstructive pulmonary disease and acute lower respiratory infection as well as two morbidity outcomes related to both cardiovascular and respiratory diseases. The measurement models utilized in this investigation followed the WHO guidelines. Over a span of a four-year period (2014-2017), the annual PM_2.5 concentration levels ranged from 38.0 μg/m³ to 75.2 μg/m³. In general, exposure levels tended to fluctuate throughout the day with the higher levels recorded during rush hours (early morning and early evening), weekends (particularly Saturdays), and summer (i.e., August and September). The highest number of excess cases and attributable proportions of premature mortalities were related to ischemic heart disease and stroke at 352 (95% CI 275-426) and 70.8% (95% CI 39.7-85.2), respectively. In general, respiratory diseases showed a higher number of excess cases and attributable proportions than cardiovascular diseases. Relative to other findings on the global stage, the results emanating from Kuwait are emerging on the higher side. The study outcomes suggest that control strategies are in dire need to bend the pollution levels in Kuwait.

Data Science in Environmental Health Research

PURPOSE OF REVIEW

Data science is an exploding trans-disciplinary field that aims to harness the power of data to gain information or insights on researcher-defined topics of interest. In this paper we review how data science can help advance environmental health research.

RECENT FINDINGS

We discuss the concepts computationally scalable handling of Big Data and the design of efficient research data platforms, and how data science can provide solutions for methodological challenges in environmental health research, such as high-dimensional outcomes and exposures, and prediction models. Finally, we discuss tools for reproducible research.

SUMMARY

In this paper we present opportunities to improve environmental research capabilities by embracing data science, and the pitfalls that environmental health researchers should avoid when employing data scientific approaches. Throughout the paper, we emphasize the need for environmental health researchers to collaborate more closely with biostatisticians and data scientists to ensure robust and interpretable results.

Cardiovascular Effects of Long-Term Exposure to Air Pollution: A Population-Based Study With 900 845 Person-Years of Follow-up

BACKGROUND

Studies have shown that long-term exposure to air pollution such as fine particulate matter (≤2.5 μm in aerodynamic diameter [PM_2.5]) increases the risk of all-cause and cardiovascular mortality. To date, however, there are limited data on the impact of air pollution on specific cardiovascular diseases. This study aimed to evaluate cardiovascular effects of long-term exposure to air pollution among residents of Seoul, Korea.

METHODS AND RESULTS

Healthy participants with no previous history of cardiovascular disease were evaluated between 2007 and 2013. Exposure to air pollutants was estimated by linking the location of outdoor monitors to the ZIP code of each participant's residence. Crude and adjusted analyses were performed using Cox regression models to evaluate the risk for composite cardiovascular events including cardiovascular mortality, acute myocardial infarction, congestive heart failure, and stroke. A total of 136 094 participants were followed for a median of 7.0 years (900 845 person-years). The risk of major cardiovascular events increased with higher mean concentrations of PM_2.5 in a linear relationship, with a hazard ratio of 1.36 (95% confidence interval, 1.29-1.43) per 1 μg/m³ PM_2.5. Other pollutants including PM_2.5-10 of CO, SO₂, and NO₂, but not O₃, were significantly associated with increased risk of cardiovascular events. The burden from air pollution was comparable to that from hypertension and diabetes mellitus.

CONCLUSIONS

This large-scale population-based study demonstrated that long-term exposure to air pollution including PM_2.5 increases the risk of major cardiovascular disease and mortality. Air pollution should be considered an important modifiable environmental cardiovascular risk factor.

The heart as an extravascular target of endothelin-1 in particulate matter-induced cardiac dysfunction

Exposure to particulate matter air pollution has been causally linked to cardiovascular disease in humans. Several broad and overlapping hypotheses describing the biological mechanisms by which particulate matter exposure leads to cardiovascular disease have been explored, although linkage with specific factors or genes remains limited. These hypotheses may or may not also lead to particulate matter-induced cardiac dysfunction. Evidence pointing to autocrine/paracrine signaling systems as modulators of cardiac dysfunction has increased interest in the emerging role of endothelins as mediators of cardiac function following particulate matter exposure. Endothelin-1, a well-described small peptide expressed in the pulmonary and cardiovascular systems, is best known for its ability to constrict blood vessels, although it can also induce extravascular effects. Research on the role of endothelins in the context of air pollution has largely focused on vascular effects, with limited investigation of responses resulting from the direct effects of endothelins on cardiac tissue. This represents a significant knowledge gap in air pollution health effects research, given the abundance of endothelin receptors found on cardiac tissue and the ability of endothelin-1 to modulate cardiac contractility, heart rate, and rhythm. The plausibility of endothelin-1 as a mediator of particulate matter-induced cardiac dysfunction is further supported by the therapeutic utility of certain endothelin receptor antagonists. The present review examines the possibility that endothelin-1 release caused by exposure to PM directly modulates extravascular effects on the heart, deleteriously altering cardiac function.

Toward the development of an in silico human model for indoor environmental design

In modern society where people spend more than 90% of their time in indoor spaces, the indoor air quality (IAQ) created by buildings has the potential of greatly influencing quality of life. Because the time spent by workers/residents in indoor spaces has increased over time, the importance of IAQ issues in terms of public health is also increasing. Additionally, the quality of the indoor thermal environment also has great impact on human comfort and performance; hence, the development of a comprehensive prediction method integrating indoor air quality/thermal environment assessment and human physiological responses, is crucial for creating a healthy, comfortable, and productive indoor environment. Accordingly, the overarching objective of this study was to develop a comprehensive and universal computer simulated person (i.e., in silico human model), integrating computational fluid dynamics (CFD), to be used in indoor environmental design and quality assessment. This paper presents and discusses the development of this computer-simulated person and its application to indoor environmental design.

Air Pollution and Mortality in Seven Million Adults: The Dutch Environmental Longitudinal Study (DUELS)

BACKGROUND

Long-term exposure to air pollution has been associated with mortality in urban cohort studies. Few studies have investigated this association in large-scale population registries, including non-urban populations.

OBJECTIVES

The aim of the study was to evaluate the associations between long-term exposure to air pollution and nonaccidental and cause-specific mortality in the Netherlands based on existing national databases.

METHODS

We used existing Dutch national databases on mortality, individual characteristics, residence history, neighborhood characteristics, and national air pollution maps based on land use regression (LUR) techniques for particulates with an aerodynamic diameter ≤ 10 μm (PM10) and nitrogen dioxide (NO2). Using these databases, we established a cohort of 7.1 million individuals ≥ 30 years of age. We followed the cohort for 7 years (2004-2011). We applied Cox proportional hazard models adjusting for potential individual and area-specific confounders.

RESULTS

After adjustment for individual and area-specific confounders, for each 10-μg/m3 increase, PM10 and NO2 were associated with nonaccidental mortality [hazard ratio (HR) = 1.08; 95% CI: 1.07, 1.09 and HR = 1.03; 95% CI: 1.02, 1.03, respectively], respiratory mortality (HR = 1.13; 95% CI: 1.10, 1.17 and HR = 1.02; 95% CI: 1.01, 1.03, respectively), and lung cancer mortality (HR = 1.26; 95% CI: 1.21, 1.30 and HR = 1.10 95% CI: 1.09, 1.11, respectively). Furthermore, PM10 was associated with circulatory disease mortality (HR = 1.06; 95% CI: 1.04, 1.08), but NO2 was not (HR = 1.00; 95% CI: 0.99, 1.01). PM10 associations were robust to adjustment for NO2; NO2 associations remained for nonaccidental mortality and lung cancer mortality after adjustment for PM10.

CONCLUSIONS

Long-term exposure to PM10 and NO2 was associated with nonaccidental and cause-specific mortality in the Dutch population of ≥ 30 years of age.

Carcinogenicity of ambient air pollution: use of biomarkers, lessons learnt and future directions

The association between ambient air pollution (AAP) exposure and lung cancer risk has been investigated in prospective studies and the results are generally consistent, indicating that long-term exposure to air pollution can cause lung cancer. Biomarkers can enhance research on the health effects of air pollution by improving exposure assessment, increasing the understanding of mechanisms, and enabling the investigation of individual susceptibility. In this review, we assess DNA adducts as biomarkers of exposure to AAP and early biological effect, and DNA methylation as biomarker of early biological change and discuss critical issues arising from their incorporation in AAP health impact evaluations, such as confounding, individual susceptibilities, timing, intensity and duration of exposure, and investigated tissue. DNA adducts and DNA methylation are treated as paradigms. However, the lessons, learned from their use in the examination of AAP carcinogenicity, can be applied to investigations of other biomarkers involved in AAP carcinogenicity.

Respiratory health effects of ambient air pollution: an update

There is new evidence for ambient air pollution (AAP) leading to an increased incidence of respiratory diseases in adults. Research has demonstrated that co-exposures have the potential to dramatically augment the effects of AAP and lower the threshold of effect of a given pollutant. Interactions between genes related to oxidative stress and AAP seem to significantly alter the effect of AAP on an individual and population basis. A better definition of vulnerable populations may bolster local or regional efforts to remediate AAP. Advances in genetic research tools have the potential to identify candidate genes that can guide further research.

Effect of air pollution control on life expectancy in the United States: an analysis of 545 U.S. counties for the period from 2000 to 2007

BACKGROUND

In recent years (2000-2007), ambient levels of fine particulate matter (PM2.5) have continued to decline as a result of interventions, but the decline has been at a slower rate than previous years (1980-2000). Whether these more recent and slower declines of PM2.5 levels continue to improve life expectancy and whether they benefit all populations equally is unknown.

METHODS

We assembled a data set for 545 U.S. counties consisting of yearly county-specific average PM2.5, yearly county-specific life expectancy, and several potentially confounding variables measuring socioeconomic status, smoking prevalence, and demographic characteristics for the years 2000 and 2007. We used regression models to estimate the association between reductions in PM2.5 and changes in life expectancy for the period from 2000 to 2007.

RESULTS

A decrease of 10 μg/m in the concentration of PM2.5 was associated with an increase in mean life expectancy of 0.35 years (SD = 0.16 years, P = 0.033). This association was stronger in more urban and densely populated counties.

CONCLUSIONS

Reductions in PM2.5 were associated with improvements in life expectancy for the period from 2000 to 2007. Air pollution control in the last decade has continued to have a positive impact on public health.

Residential proximity to major roadway and 10-year all-cause mortality after myocardial infarction

BACKGROUND

The relationship between residential proximity to roadway and long-term survival after acute myocardial infarction (AMI) is unknown. We investigated the association between distance from residence and major roadway and 10-year all-cause mortality after AMI in the Determinants of Myocardial Infarction Onset Study (Onset Study), hypothesizing that living closer to a major roadway at the time of AMI would be associated with increased risk of mortality.

METHODS AND RESULTS

The Onset Study enrolled 3886 individuals hospitalized for AMI in 64 centers across the United States from 1989 to 1996. Institutionalized patients, those providing only post office boxes, and those whose addresses could not be geocoded were excluded, leaving 3547 patients eligible for analysis. Addresses were geocoded, and distance to the nearest major roadway was assigned. Cox regression was used to calculate hazard ratios, with adjustment for personal characteristics (age, sex, race, education, marital status, distance to nearest acute care hospital), clinical characteristics (smoking, body mass index, comorbidities, medications), and neighborhood-level characteristics derived from US Census block group data (household income, education, urbanicity). There were 1071 deaths after 10 years of follow-up. In the fully adjusted model, compared with living >1000 m, hazard ratios (95% confidence interval) for living ≤100 m were 1.27 (1.01-1.60), for 100 to ≤200 m were 1.19 (0.93-1.60), and for 200 to ≤1000 m were 1.13 (0.99-1.30) (P(trend)=0.016).

CONCLUSIONS

In this multicenter study, living close to a major roadway at the time of AMI was associated with increased risk of all-cause 10-year mortality; this relationship persisted after adjustment for individual and neighborhood-level covariates.

Mutagenicity and genotoxicity of PM2.5 issued from an urbano-industrialized area of Dunkerque (France)

Epidemiological studies have demonstrated the link between chronic exposure to particulate matter (PM), especially particles with an aerodynamic diameter lesser than 2.5 µm (PM(2.5) ), and lung cancer. Mechanistic investigations focus on the contribution of the various genotoxicants adsorbed onto the particles, and more particularly on polycyclic aromatic hydrocarbons or nitroaromatics. Most of the previous studies dealing with genotoxic and/or mutagenic measurements were performed on organic extracts obtained from PM(2.5) collected in polluted areas. In contrast, we have evaluated genotoxic and mutagenic properties of urbano-industrial PM(2.5) (PM) collected in Dunkerque (France). Thermally desorbed PM(2.5) (dPM) was also comparatively studied. Suspensions of PM and dPM (5-50 µg per plate) were tested in Salmonella tester strains TA98, TA102 and YG1041 ± S9mix. Significant mutagenicity was observed for PM in YG1041 ± S9 mix. In strain TA102 - S9mix, a slight, but not significant dose-response increase was observed, for both PM and dPM. Genotoxic properties of PM and dPM were evaluated by the measurement of (1) 8-OHdG in A549 cells and (2) bulky DNA adducts on A549 cells and on human alveolar macrophages (AMs) in primary culture. A dose-dependant formation of 8-OHdG adducts was observed on A549 cells for PM and dPM, probably mainly attributed to the core of the particles. Bulky DNA adducts were observed only in AMs after exposure to PM and dPM. In conclusion, using relevant exposure models, suspension of PM(2.5) induces a combination of DNA-interaction mechanisms, which could contribute to the induction of lung cancer in exposed populations.

The Brooklyn traffic real-time ambient pollutant penetration and environmental dispersion (B-TRAPPED) field study methodology

The Brooklyn Traffic Real-Time Ambient Pollutant Penetration and Environmental Dispersion (B-TRAPPED) field study examined indoor and outdoor exposure to traffic-generated air pollution by studying the individual processes of generation of traffic emissions, transport and dispersion of air contaminants along a roadway, and infiltration of the contaminants into a residence. Real-time instrumentation was used to obtain highly resolved time-series concentration profiles for a number of air pollutants. The B-TRAPPED field study was conducted in the residential Sunset Park neighborhood of Brooklyn, NY, USA, in May 2005. The neighborhood contained the Gowanus Expressway (Interstate 278), a major arterial road (4(th) Avenue), and residential side streets running perpendicular to the Gowanus Expressway and 4(th) Avenue. Synchronized measurements were obtained inside a test house, just outside the test house façade, and along the urban residential street canyon on which the house was located. A trailer containing Federal Reference Method (FRM) and real-time monitors was located next to the Gowanus Expressway to assess the source. Ultrafine particulate matter (PM), PM(2.5), nitrogen oxides (NO(x)), sulfur dioxide (SO(2)), carbon monoxide (CO), carbon dioxide (CO(2)), temperature, relative humidity, and wind speed and direction were monitored. Different sampling schemes were devised to focus on dispersion along the street canyon or infiltration into the test house. Results were obtained for ultrafine PM, PM(2.5), criteria gases, and wind conditions from sampling schemes focused on street canyon dispersion and infiltration. For comparison, the ultrafine PM and PM(2.5) results were compared with an existing data set from the Los Angeles area, and the criteria gas data were compared with measurements from a Vancouver epidemiologic study. Measured ultrafine PM and PM(2.5) concentration levels along the residential urban street canyon and at the test house façade in Sunset Park were demonstrated to be comparable to traffic levels at an arterial road and slightly higher than those in a residential area of Los Angeles. Indoor ultrafine PM levels were roughly 3-10 times lower than outdoor levels, depending on the monitor location. CO, NO(2), and SO(2) levels were shown to be similar to values that produced increased risk of chronic obstructive pulmonary disease hospitalizations in the Vancouver studies.

A systematic review of the relation between long-term exposure to ambient air pollution and chronic diseases

We conducted a systematic review of all studies published between 1950 and 2007 of associations between long-term exposure to ambient air pollution and the risks in adults of nonaccidental mortality and the incidence and mortality from cancer and cardiovascular and respiratory diseases. We searched bibliographic databases for cohort and case-control studies, abstracted characteristics of their design and conduct, and synthesized the quantitative findings in tabular and graphic form. We assessed heterogeneity, estimated pooled effects for specific pollutants, and conducted sensitivity analyses according to selected characteristics of the studies. Our analysis showed that long-term exposure to PM2.5 increases the risk of nonaccidental mortality by 6% per a 10 microg/m3 increase, independent of age, gender, and geographic region. Exposure to PM2.5 was also associated with an increased risk of mortality from lung cancer (range: 15% to 21% per a 10 microg/m3 increase) and total cardiovascular mortality (range: 12% to 14% per a 10 microg/m3 increase). In addition, living close to busy traffic appears to be associated with elevated risks of these three outcomes. Suggestive evidence was found that exposure to PM2.5 is positively associated with mortality from coronary heart diseases and exposure to SO2 increases mortality from lung cancer. For the other pollutants and health outcomes, the data were insufficient data to make solid conclusions.

Fine particulate matter and mortality: a comparison of the six cities and American Cancer Society cohorts with a medicare cohort

BACKGROUND

The American Cancer Society study and the Harvard Six Cities study are 2 landmark cohort studies for estimating the chronic effects of fine particulate air pollution (PM2.5) on mortality. Using Medicare data, we assessed the association of PM2.5 with mortality for the same locations included in these studies.

METHODS

We estimated the chronic effects of PM2.5 on mortality for the period 2000-2002 using mortality data for cohorts of Medicare participants and average PM2.5 levels from monitors in the same counties included in the 2 studies. We estimated mortality risk associated with air pollution adjusting for individual-level (age and sex) and area-level covariates (education, income level, poverty, and employment). We controlled for potential confounding by cigarette smoking by including standardized mortality ratios for lung cancer and chronic obstructive pulmonary disease.

RESULTS

Using the Medicare data, we estimated that a 10 microg/m increase in the yearly average PM2.5 concentration is associated with 10.9% (95% confidence interval = 9.0-12.8) and with 20.8% (14.8-27.1) increases in all-cause mortality for the American Cancer Society and Harvard Six Cities study counties, respectively. The estimates are somewhat higher than those reported by the original investigators.

CONCLUSION

Although Medicare data lack information on some potential confounding factors, we estimated risks similar to those in the previously published reports, which incorporated more extensive information on individual-level confounders. We propose that the Medicare files can be used to construct on-going cohorts for tracking the risk of air pollution over time.

Air pollution and public health: a guidance document for risk managers

This guidance document is a reference for air quality policymakers and managers providing state-of-the-art, evidence-based information on key determinants of air quality management decisions. The document reflects the findings of five annual meetings of the NERAM (Network for Environmental Risk Assessment and Management) International Colloquium Series on Air Quality Management (2001-2006), as well as the results of supporting international research. The topics covered in the guidance document reflect critical science and policy aspects of air quality risk management including i) health effects, ii) air quality emissions, measurement and modeling, iii) air quality management interventions, and iv) clean air policy challenges and opportunities.

Energy, energy efficiency, and the built environment

Since the last decades of the 19th century, technological advances have brought substantial improvements in the efficiency with which energy can be exploited to service human needs. That trend has been accompanied by an equally notable increase in energy consumption, which strongly correlates with socioeconomic development. Nonetheless, feasible gains in the efficiency and technology of energy use in towns and cities and in homes have the potential to contribute to the mitigation of greenhouse-gas emissions, and to improve health, for example, through protection against temperature-related morbidity and mortality, and the alleviation of fuel poverty. A shift towards renewable energy production would also put increasing focus on cleaner energy carriers, especially electricity, but possibly also hydrogen, which would have benefits to urban air quality. In low-income countries, a vital priority remains the dissemination of affordable technology to alleviate the burdens of indoor air pollution and other health effects in individuals obliged to rely on biomass fuels for cooking and heating, as well as the improvement in access to electricity, which would have many benefits to health and wellbeing.

Electricity generation and health

The provision of electricity has been a great benefit to society, particularly in health terms, but it also carries health costs. Comparison of different forms of commercial power generation by use of the fuel cycle methods developed in European studies shows the health burdens to be greatest for power stations that most pollute outdoor air (those based on lignite, coal, and oil). The health burdens are appreciably smaller for generation from natural gas, and lower still for nuclear power. This same ranking also applies in terms of greenhouse-gas emissions and thus, potentially, to long-term health, social, and economic effects arising from climate change. Nuclear power remains controversial, however, because of public concern about storage of nuclear waste, the potential for catastrophic accident or terrorist attack, and the diversion of fissionable material for weapons production. Health risks are smaller for nuclear fusion, but commercial exploitation will not be achieved in time to help the crucial near-term reduction in greenhouse-gas emissions. The negative effects on health of electricity generation from renewable sources have not been assessed as fully as those from conventional sources, but for solar, wind, and wave power, such effects seem to be small; those of biofuels depend on the type of fuel and the mode of combustion. Carbon dioxide (CO2) capture and storage is increasingly being considered for reduction of CO2 emissions from fossil fuel plants, but the health effects associated with this technology are largely unquantified and probably mixed: efficiency losses mean greater consumption of the primary fuel and accompanying increases in some waste products. This paper reviews the state of knowledge regarding the health effects of different methods of generating electricity.

Outdoor air pollution: particulate matter health effects

Numerous investigations studying multiple populations across a variety of environmental settings have demonstrated a strong association between ambient air particulate matter and cardiopulmonary morbidity and mortality. In most studies, the effect size of ambient air particulate pollution on health outcomes is small. However, the exposed population worldwide is very large. Accordingly, particulate air pollution appears to be an important public health hazard that makes an important contribution to the total burden of disease and death in populations across the world. Much of the evidence linking ambient air particulates with adverse health effects is derived from population-based, observational research with potential unidentified confounding exposures, precluding definitive assessments about causation and providing limited mechanistic insights. A growing body of research suggests particulate-associated adverse health effects result from the induction of proinflammatory responses in the lower respiratory tract. Ambient air particulates may increase lung cancer risk.

The mutagenicity of urban particulate matter in an enzyme free system is associated with the generation of reactive oxygen species

Urban particulate matter (UPM) contributes to lung cancer incidence. UPM has been shown to be genotoxic to mammalian cells and to induce mutations in the Ames assay. Here, we have studied the induction of mutations generated by direct acting mutagenic components of UPM, using the supF forward mutation assay. Plasmid pSP189 was exposed to UPM in aqueous solution in the presence of sucrose buffer, to reduce strand breaks. The mutation frequency induced by 1 microg/microl UPM was 4.99 mutants per 10(4) colonies. This was reduced to 0.84 and 1.48 mutants per 10(4) colonies by addition of mannitol (1 mM) or EDTA (1 mM), respectively. A large percentage of mutant plasmids contained frameshift mutations (57%), and 31% of mutant plasmids contained multiple mutations. Of the base substitution mutations, 88% were at GC pairs, with twice as many transversions as transitions. The types of mutations induced, the reduction of mutagenicity by the inclusion of the free radical scavenger, mannitol, or the metal chelator, EDTA, and the sequence context of the induced mutations all support the conclusion that the majority of mutations were induced by reactive oxygen species generated by metal ions present in the UPM. Most mutation studies with UPM have focused on organic carcinogens present on UPM. Our results highlight the potential contribution of metal ions to the mutagenicity of UPM.

Reanalysis of the Harvard Six Cities Study, part II: sensitivity analysis

Following the validation and replication of the Harvard Six Cities Study (Krewski et al., this issue), we conducted a wide range of sensitivity analyses to explore the observed associations between long-term exposure to fine particle or sulfate air pollution and mortality. We examined the impact of alternative risk models on estimates of risk, taking into account covariates not included in the original analyses. These risk models provided a basis for identifying covariates that may confound or modify the association between fine particle or sulfate air pollution and mortality, and for identifying sensitive population subgroups. The possibility of confounding due to occupational exposures was also investigated. Residence histories were coded for the study subjects and were used to examine temporal patterns of exposure and risk. Our sensitivity analyses showed the mortality risk estimates for fine particle and sulfate air pollution to be highly robust against alternative risk models of the Cox proportional hazards family, including models with additional covariates from the original questionnaires not included in the original published analyses. There was limited evidence of departures from the proportional hazards assumption. Flexible exposure-response models provided some evidence of departures from linearity at both low and high sulfate concentrations. Incorporating information on changes over time in cigarette smoking and body mass index had little effect on the association between fine particles and mortality. There was limited evidence of variation in risk with attained age, gender, smoking status, occupational exposure to dust and fumes, marital status, heart or lung diseases, or lung function. However, air pollution risk did appear to decreasing with increasing educational attainment. Extensive adjustment for occupation using aggregate indices of occupational "dirtiness" and occupational exposure to known lung carcinogens had little impact on the mortality risks associated with particulate air pollution. Our evaluation of population mobility indicated that relatively few subjects moved from their original city of residence. Attempts to identify critical exposure time windows were limited by the lack of marked interindividual variation in temporal exposure patterns throughout the study period. Overall, this extensive sensitivity analysis both supported the conclusions reached by the original investigators and demonstrated the robustness of these conclusions to alternative analytic approaches.