Associations between black carbon, particulate matter, combustion-derived particles and ischemic heart disease

Abstract

Listen

Translate article

Long term exposure to air pollution increases cardiopulmonary mortality, but there are uncertainties regarding which components of air pollution that are the most harmful, and the temporal-risk relationship. We analyzed a cohort of men in Sweden to investigate the effects of total and source-specific particles on ischemic heart disease (IHD) and stroke (incidence and mortality), for different time windows of exposure. The study is part of a national air pollution collaboration using multiple cohorts. Methods: A cohort of men in Gothenburg was examined and followed, initially to determine predictors of cardiovascular disease. We collected data on residential addresses and cause-specific mortality for the study period of 1990-2014. Each individual was assigned yearly PM10, PM2.5, and Black Carbon (BC) exposure from dispersion models. Cox-regression models with time-dependent exposures were used to study the effects of different types of air pollution exposure on IHD and stroke. Results: 2190 incident IHD and 1598 incident strokes, and 1436 IHD deaths and 850 stroke deaths occurred in the cohort. Exposure levels decreased from 1990 to 2014. BC was strongly associated with an increased risk of IHD death (HR 1.49, 95% CI 1.07-2.06 per 1 µg/m3 of exposure last year). The associations between BC and incident IHD and stroke were mostly positive but somewhat weaker, as were the associations between PM10/PM2.5 and IHD/stroke. In multiple-pollutant models mortality was more clearly associated with BC and combustion-derived particles than with other particle constituents. The last year of exposure tended to be more associated with mortality than longer lagged time-windows. Discussion: Death in IHD was strongly associated with long term exposure to residential BC and combustion-derived particles in Gothenburg, and more weakly associated with total levels of particles or particle constituents. Results were only slightly affected by other covariates.