Ambient Fine Particulate Matter and Mortality among Survivors of Myocardial Infarction: Population-Based Cohort Study

Background: Survivors of acute myocardial infarction (AMI) are at increased risk of dying within several hours to days following exposure to elevated levels of ambient air pollution. Little is known, however, about the influence of long-term (months to years) air pollution exposure on survival after AMI. Objective: We conducted a population-based cohort study to determine the impact of long-term exposure to fine particulate matter ≤ 2.5 μm in diameter (PM2.5) on post-AMI survival. Methods: We assembled a cohort of 8,873 AMI patients who were admitted to 1 of 86 hospital corporations across Ontario, Canada in 1999–2001. Mortality follow-up for this cohort extended through 2011. Cumulative time-weighted exposures to PM2.5 were derived from satellite observations based on participants’ annual residences during follow-up. We used standard and multilevel spatial random-effects Cox proportional hazards models and adjusted for potential confounders. Results: Between 1999 and 2011, we identified 4,016 nonaccidental deaths, of which 2,147 were from any cardiovascular disease, 1,650 from ischemic heart disease, and 675 from AMI. For each 10-μg/m3 increase in PM2.5, the adjusted hazard ratio (HR10) of nonaccidental mortality was 1.22 [95% confidence interval (CI): 1.03, 1.45]. The association with PM2.5 was robust to sensitivity analyses and appeared stronger for cardiovascular-related mortality: ischemic heart (HR10 = 1.43; 95% CI: 1.12, 1.83) and AMI (HR10 = 1.64; 95% CI: 1.13, 2.40). We estimated that 12.4% of nonaccidental deaths (or 497 deaths) could have been averted if the lowest measured concentration in an urban area (4 μg/m3) had been achieved at all locations over the course of the study. Conclusions: Long-term air pollution exposure adversely affects the survival of AMI patients. Citation: Chen H, Burnett RT, Copes R, Kwong JC, Villeneuve PJ, Goldberg MS, Brook RD, van Donkelaar A, Jerrett M, Martin RV, Brook JR, Kopp A, Tu JV. 2016. Ambient fine particulate matter and mortality among survivors of myocardial infarction: population-based cohort study. Environ Health Perspect 124:1421–1428; http://dx.doi.org/10.1289/EHP185

. ICD-9 and ICD-10 codes for study outcomes Table S2. Comparison of selected characteristics between study participants who lived in the Greater Toronto Area (GTA) and those who lived outside GTA Table S3. Comparison of models with exposure to PM 2.5 as either a linear term or a non-linear term using a natural cubic spline function, by selected causes of death Table S4. Sensitivity analyses for the association of non-accidental mortality with every 10μg/m 3 increase of PM 2.5 To quantify the burden of death attributed to long-term exposure to PM 2.5 , we derived attributable fraction which was applied to the number of all-natural deaths during the follow-up using the formula as follows (Lim et al. 2012;World Health Organization 2004): Where AF is the attributable fraction (i.e., burden attributable to risk factor such as PM 2.5 ), RR i is the adjusted hazard ratio at exposure level i, P i is the estimated population distribution of exposure, P' i is the counterfactual distribution of exposure which was 4-µg/m 3 , and m is the maximum exposure level.

International Classification of Diseases (ICD) Codes for Study Outcomes
In the present study, our primary outcome was non-accidental mortality, because we were interested in assessing the burden of post-AMI death attributed to air pollution. To evaluate the specificity of the association between air pollution and mortality, we also ascertained deaths from any cardiovascular disease, ischemic heart disease, and AMI, respectively. In addition, we identified deaths from accidental causes and from non-cardiopulmonary, non-lung cancer causes to serve as negative control outcomes. The International Classification of Diseases, Ninth Revision, ICD-9 code and Tenth Revision, ICD-10 code for our study outcomes are listed in Table S1.

Canadian Census Divisions and Census Tracts
Canadian census divisions are group of neighbouring municipalities joined together for the purposes of the provision of services (such as ambulance services) and regional planning. A census division corresponds to a county or a regional district. 3 In contrast, Canadian census tracts are small and relatively homogeneous geographic units that usually comprise a population of 2,500 to 8,000 (Statistics Canada 2011). It is one of the smallest standard geographic areas for which all census data are disseminated (Statistics Canada 2011).

Comparison of Spatial Resolution for Different Datasets in the Study
Spatial resolution for different datasets used in our study is described as follows (in the order from highest to lowest): Postal codes (a total of 269,676 in Ontario) > Census tract (2,136 in Ontario) > 10km by 10km grids in the PM 2.5 exposure surface (1,198 grids in Ontario) > Census subdivision (507 in Ontario) > Census division (50 in Ontario) These datasets were created by different organizations for different purposes; as a result, their areas may overlap. For example, 10km by 10km grid-cells may overlap with census divisions.

Additional Sensitivity Analyses
We further investigated the possibility of adjusting for geographically-variable sociodemographic and other related health-care indicators. In doing this, we created five new sets of covariates: (1) a dichotomous indicator variable for North/South Ontario; (2) rurality; (3) neighborhood-level % of visible minority; (4) deprivation; and (5) density of family doctors. We conducted sensitivity analyses by additionally controlling for these new variables.
To classify Ontario into northern and southern regions, we created an indicator variable based on the 14 Ontario Local Health Integrated Networks (equivalent to health regions). The Local Health Integrated Networks are responsible for planning, integrating, and funding various local health care services across the province of Ontario. Of the 14 Local Health Integrated Networks, two (North East and North West) cover the population living in northern Ontario and were combined to create the indicator (versus the rest of health regions).
To represent rurality, we created two separate variables. The first variable comprises five categories (urban core, urban fringe, rural fringe, urban area outside census metropolitan area, and rural area outside census metropolitan area) as defined by Statistics Canada. The second variable was created based on the Rurality Index for Ontario which takes into account community population density, travel time to nearest basic referral centre, and travel time to nearest advanced referral centre (Kralj 2000). The Rurality Index has a value ranging from 0 to 100, with 0-39 considered as urban and 40 and above considered rural (Kralj 2000). The Rurality Index has been used by the Ontario government and the Ontario Medical Association for determining incentive and bonus payment to physicians, and has been widely used to define urban-rural split in previous research in health system improvement.
In addition, using the 2001 Canadian Census Tract data, we derived a variable for % of visible minority. As well, we derived a variable to represent neighbourhood-level deprivation based on the Ontario Marginalization Index that has been previously developed to quantify the degree of marginalization in health and social well-being across Ontario (Matheson et al 2012). It consists of four dimensions thought to underlie the construct of marginalization: residential instability; material deprivation; dependency; and ethnic concentration. Additionally, we created a variable for density of family physicians using the Physician Database in Ontario, which contains information about all physicians in Ontario.
Furthermore, we conducted sensitivity analyses by restricting to cohort members who lived between 41.7 o N and 46.0 o N, where the vast majority of the Ontario population resides (range of latitude in Ontario: 41.7 o N to 50.8 o N). Lastly, we conducted analyses by restricting to cohort members who lived within 5 km from any manufacturing or process facilities releasing particulate matter that meet reporting thresholds and hence are required to report to Environment Canada, according to the Canadian Environmental Protection Act (1999).  Values are percent or mean ± standard deviation, unless otherwise indicated.
* Study participants lived in the Greater Toronto Area contributed to 19,721 person-years of observations whereas participants lived in all other regions of Ontario contributed to 52,380 person-years of observations. The mortality rates were presented as unadjusted rates per 1,000 person years. In-hospital process (e.g., characteristics of hospitals and physicians) Clinical severity (e.g., STEMI)