Ambient air pollution and the risk of acute myocardial infarction and stroke: A national cohort study
Does breathing polluted air over many years increase your risk of having a heart attack or stroke? Study of 2.7M people finds long-term exposure to air pollution increases heart attack and stroke hospitalization risk by up to 7.8%.
College of Health researcher(s)
Abstract
We used a large national cohort in Canada to assess the incidence of acute myocardial infarction (AMI) and stroke hospitalizations in association with long-term exposure to fine particulate matter (PM2.5), nitrogen dioxide (NO2), and ozone (O3). The study population comprised 2.7 million respondents from the 2006 Canadian Census Health and Environment Cohort (CanCHEC), followed for incident hospitalizations of AMI or stroke between 2006 and 2016. We estimated 10-year moving average estimates of PM2.5, NO2, and O3, annually. We used Cox proportional hazards models to examine the associations adjusting for various covariates. For AMI, each interquartile range (IQR) increase in exposure was found to be associated with a hazard ratio of 1.026 (95% CI: 1.007–1.046) for PM2.5, 1.025 (95% CI: 1.001–1.050) for NO2, and 1.062 (95% CI: 1.041–1.084) for O3, respectively. Similarly, for stroke, an IQR increase in exposure was associated with a hazard ratio of 1.078 (95% CI: 1.052–1.105) for PM2.5, 0.995 (95% CI: 0.965–1.030) for NO2, and 1.055 (95% CI: 1.028–1.082) for O3, respectively. We found consistent evidence of positive associations between long-term exposures to PM2.5, and O3, and to a lesser degree NO2, with incident AMI and stroke hospitalizations.
FAQ: Air Pollution and Cardiovascular Health in Canada
What was the main objective of this study regarding air pollution and cardiovascular health?
The primary objective of this national cohort study in Canada was to evaluate the long-term effects of exposure to ambient fine particulate matter (PM2.5), nitrogen dioxide (NO2), and ozone (O3) on the incidence of acute myocardial infarction (AMI) and stroke hospitalizations. The study aimed to address inconsistencies in previous research and provide robust evidence from a large-scale population.
Who was included in the study population and for how long were they followed?
The study population comprised approximately 2.7 million non-institutionalized respondents aged 25-89 from the 2006 Canadian Census Health and Environment Cohort (CanCHEC), excluding residents of Quebec due to data unavailability. These individuals were followed for incident hospitalizations of AMI or stroke between 2006 and 2016, contributing approximately 29 million person-years of follow-up.
What were the key findings regarding the association between long-term air pollution exposure and cardiovascular events?
The study found consistent evidence of positive associations between long-term exposures to PM2.5, O3, and to a lesser extent NO2, with incident AMI and stroke hospitalizations. For every inter-quartile range (IQR) increase in exposure:
- For AMI: PM2.5 (Hazard Ratio (HR): 1.026), NO2 (HR: 1.025), and O3 (HR: 1.062).
- For Stroke: PM2.5 (HR: 1.078), NO2 (HR: 0.995 - not statistically significant), and O3 (HR: 1.055).
These associations remained largely consistent after adjusting for various individual and ecological covariates.
Were there any specific population subgroups that showed a stronger association between air pollution and cardiovascular events?
Yes, the study identified certain subgroups with stronger associations:
- Age: Associations of PM2.5 and O3 with both AMI and stroke were stronger among older age groups (65 years and older).
- Sex: PM2.5 and NO2 showed stronger positive associations among men for both AMI and stroke, while O3 showed higher risk among women for incident AMI and stroke.
- Income: Stronger positive associations were observed among the lowest-income group for all three pollutants with both AMI and stroke, with the exception of PM2.5 and incident stroke, which had comparable associations across all income groups.
- Ozone Tertiles: Stronger associations between PM2.5 and NO2 with both AMI and stroke were found at the highest tertile of ozone exposure, suggesting a modifying role of O3.
What was the estimated burden of disease attributable to ambient air pollutants in the study population?
For the 60,800 AMI cases observed, approximately 1,100 hospitalizations were attributable to PM2.5 and 3,200 to O3 (based on single-pollutant models). For the 37,800 stroke cases, 1,960 hospitalizations were attributable to PM2.5 and 1,743 to O3. Jointly, PM2.5 and O3 accounted for 5.95% of AMI hospitalizations and 6.75% of stroke hospitalizations across the follow-up period in a two-pollutant model.
Did the study identify any thresholds for the concentration-response relationship between air pollution and cardiovascular events?
The study found that the relationship between PM2.5 and both AMI and stroke was "near linear" without any obvious threshold. This suggests that there is an increased risk of AMI and stroke even at relatively low concentrations of PM2.5 (as low as 0.7 µg/m³ for AMI and 6.3 µg/m³ for stroke). For O3, the relationship with AMI was also near linear (risk observed as low as 3.7 ppb), but a "sublinear" relationship with incident stroke was observed, with a possible threshold around 64.7 ppb.
How do these findings compare to previous international research on air pollution and cardiovascular health?
The study's findings align well with pooled hazard ratios from recent meta-analyses of other epidemiological studies. For instance, the estimated 8% increased risk of incident AMI hospitalization per 10 µg/m³ increase in PM2.5 was similar to the pooled estimate of 8% from other studies. The associations for NO2 with AMI and stroke were also consistent with those reported in European and Canadian studies, generally showing weaker associations compared to particulate pollutants.
What are the key strengths and limitations of this study?
Strengths:
- Use of a large national cohort (2.7 million adults), leading to reduced statistical uncertainty and increased precision.
- Ability to investigate heterogeneity in risk among various subgroups and characterize exposure-response relationships with precision.
- Comprehensive accounting for prevalent cases (13-year wash-out period) to identify incident hospitalizations.
- Assignment of exposures to residential postal codes at a fine spatial scale, incorporating residential mobility since 1981.
- Inclusion of detailed individual-level socioeconomic factors, including annual time-varying geographically adjusted income.
Limitations:
- Reliance on health administrative databases may miss underdiagnosed cases, though adjustments for hospital proximity suggest robustness.
- Lack of individual-level smoking and BMI data, although indirect adjustment using an ancillary national cohort showed consistent estimates.
- Inability to fully characterize cumulative air pollution exposures from non-residential locations (e.g., workplace, commute, indoor environments), which may have attenuated estimates towards the null due to non-differential exposure misclassification.