Long-term exposure to high particulate matter pollution and cardiovascular mortality: A 12-year cohort study in four cities in northern China

Abstract

Epidemiologic studies have demonstrated that long-term exposure to relatively low levels of particulate air pollution is associated with adverse cardiovascular outcomes in Europe and North America. However, few studies have assessed the association with high level air pollutants. We aimed to assess the cardiovascular effects of long-term exposure to high level concentrations of inhalable particulate and to identify the characteristics of the Chinese population that are susceptible to the health effects. A retrospective cohort, containing 39,054 subjects from four cities in northern China, was followed for mortality of all cause and specific cardiovascular diseases from 1998 to 2009. Information on concentrations of PM10 (particulate matter<10μm in aerodynamic diameter) was collected from the local Environmental Monitoring Centers. The estimated exposure for the study participants was the mean concentration of PM10 over their surviving years during the cohort period. Relative risk values were obtained using Cox proportional hazards regression models after adjusting for potential confounding factors. For each 10μg/m3 increase in PM10, the relative risk ratios (RRs) of all-cause mortality, cardiovascular disease mortality, ischemic heart disease mortality, heart failure disease mortality, and cerebrovascular disease mortality were 1.24 (95% CI, 1.22–1.27), 1.23 (95% CI, 1.19–1.26), 1.37 (95% CI, 1.28–1.47), 1.11(95% CI, 1.05–1.17), and 1.23(95% CI:1.18–1.28), respectively. Results from stratified analyses suggest that the effects of PM10 on cardiovascular mortality were more pronounced in males, smokers and people with a higher socioeconomic status. Long-term exposure to PM10 increases mortality from cardiovascular disease, especially from ischemic heart disease and this association seemed to be modified by other factors. Further research that focuses on exploring dose–response relationship and inter-population comparisons is warranted.