Air pollution and myocardial infarction; effect modification by sociodemographic and environmental factors. A cohort study from Denmark.

Air pollution is associated with increased risk of myocardial infarction (MI), but it is unresolved to what extent the association is modified by factors such as socioeconomic status, comorbidities, financial stress, residential green space, or road traffic noise. We formed a cohort of all (n = 1,964,702) Danes, aged 50–85 years, with 65,311 cases of MI during the followed-up period 2005–2017. For all participants we established residential five-year running average exposure to particulate matter <2.5 μm (PM 2.5), ultrafine particles (UFP, <0.1 μm), elemental carbon (EC) and nitrogen dioxide (NO 2). We evaluated risk in population strata, using Aalen additive hazards models to estimate absolute risk and Cox proportional hazards models to estimate relative risk of MI with 95% confidence intervals (CI). PM 2.5 and the other pollutant were associated with MI. Lower education and lower income were associated with higher absolute risks of MI from air pollution, whereas no clear effect modification was apparent for relative risk estimates. For example, 5 μg/m3 higher PM 2.5 was associated with HR for MI of 1.16 (95% CI: 1.10–1.22) among those with only mandatory education and 1.13 (95% CI: 1.03–1.24) among those with long education. The corresponding rate differences per 100,000 person years were 243 (95% CI: 216–271) and 358 (95% CI: 338–379), respectively. Higher level of comorbidity was consistently across all four pollutants associated with both higher absolute and relative risk of MI. In conclusion, people with comorbid conditions or of lower SES appeared more vulnerable to long-term exposure to air pollution and more cases of MI may be prevented by focused interventions in these groups. • Air pollution is positively associated with risk of myocardial infarction. • People of low socioeconomic status or with comorbidities are more susceptible. • Effect modification appears to be most apparent with absolute risk estimates. [ABSTRACT FROM AUTHOR]