Air pollution and cardiovascular disease mortality are clearly linked, yet population-based studies of air pollution and arrhythmogenesis have not been conducted in women. Moreover, extant studies have not evaluated whether acute, pro-arrhythmic effects of exposure to ambient air pollutants are modified by three potentially important markers of the environmental, socioeconomic and clinical context within which such exposures ostensibly increase cardiovascular risk in women: chronic exposure status, neighborhood of residence, and disease-specific susceptibility factors for sudden death. We will investigate these issues in an ethnically diverse population of 68,133 post-menopausal women aged 59-70 years from the 40 clinical centers and their satellites participating in the baseline examination of the Women's Health Initiative clinical trial (WHI, 1993-1998). We will estimate exposure to criteria pollutants (PM10; NO2; SO2; CO; 03) in ambient air at geocoded participant addresses using validated, spatial models that rely on pollutant concentrations recorded at adjacent fixed-site monitors in the U.S. Environmental Protection Agency Aerometric Information Retrieval System. Spatially interpolated exposures will take the form of average pollutant concentrations on the day of, and for the 1, 2 & 3 days and 1, 2 & 3 years preceding the baseline examination and year three follow-up. We will reliably evaluate autonomic function, atrioventricular conduction, ventricular depolarization, ventricular repolarization and ectopy from resting, standard 12-lead ECGs recorded at the WHI examinations. After removing seasonal variations and long-term trends, and in addition, adjusting for demographic and meteorological covariates, we will explore the putative association between air pollutants and ECG measures. Then we will determine whether the associations are modified by chronic exposure status, socioeconomic characteristics of geographic regions in which participants live, and clinical risk factors for sudden cardiac death using Bayesian, hierarchical models. Lastly, we will assess sensitivity of our findings to adjustment for exposure measurement error arising from spatial interpolation of personal exposures from ambient concentrations of air pollutants. Our ancillary study will thereby evaluate the biologically relevant proarrhythmic mechanisms and contextual features linking ambient air pollution to cardiovascular disease morbidity and mortality in a large, ethnically and geographically diverse group of postmenopausal women. In doing so, it will improve understanding of associations between airborne pollutants and cardiovascular disease mortality, facilitate assessment of current U.S. air quality standards, and yield insight into the relatively gradual decline of sudden cardiac death rates among U.S. women over the last decade.
