Household Air Pollution and Cardiovascular Disease Risk: Results from the Prospective Urban and Rural Epidemiological Study

Abstract

Introduction: An estimated 2.9 million deaths in 2013 were attributable to exposure to household air pollution (HAP). The majority (52%) of this burden is from cardiovascular disease (CVD), yet to date no prospective study has examined HAP and CVD incidence. Here we investigate the association between HAP and CVD within the Prospective Urban and Rural Epidemiology (PURE) cohort. Methods: We examined 26 centers in the PURE study that reported greater than 10% solid fuel use for cooking, resulting in 97,350 participants with CVD follow-up across 10 countries (Bangladesh, Brazil, Chile, China, Colombia, India, Pakistan, South Africa, and Zimbabwe). Average follow-up in the cohort was 4.1 years, yielding 1,073 myocardial infarctions (MI), 878 strokes, 2,195 severe CVD events, 1,080 fatal CVD events, and 2,782 total deaths. We ran Cox-proportional hazards models, controlling for the INTERHEART risk score (a validated score for existing CVD risk-factors), urban/rural status, education, outdoor PM2.5, CVD medication use, country, and center. Results: We observed increases in risk of all-cause mortality (HR 1.16; 95% CI: 1.02-1.32) for individuals living in households with solid fuel use for cooking compared to electricity or liquefied petroleum gas (LPG). Increases (not statistically significant) in the risk of MI (7-20%), Stroke (9-46%) and Severe CVD (8-23%) were observed in models including all countries and in China/India models. The largest risks were observed in India. There was heterogeneity in risk estimates across study centers, although meta-analysis of these results were similar to overall models. Conclusions: In this large, multi-country prospective cohort we observed increases in all-cause mortality and CVD incidence for individuals living in households using solid fuel for cooking compared to electricity or LPG. Although effect sizes were modest, these results contribute the first direct evidence of an association between HAP and CVD incidence.