Abstract for the 2019 HEI Annual Conference
Susceptibility to Multiple Air Pollutants in Cardiovascular Disease
1Jane E. Clougherty, 1Jamie L. Humphrey, 2Ellen J. Kinnee, 2Laura D. Kubzansky, 3Colleen E. Reid, 1Leslie A. McClure1, Lucy Robinson1
1Drexel University, Philadelphia, PA, USA; 2University of Pittsburgh, PA, USA; 3Harvard T.H. Chan School of Public Health, Boston, MA, USA; 4University of Colorado at Boulder, CO, USA
Background. Cardiovascular disease (CVD), the leading cause of death in the U.S., has been linked to chronic and acute air pollution exposures. Research has identified stronger effects of air pollution in lower-socioeconomic position (SEP) communities, where exposures are also often higher. While specific factors underlying this socioeconomic susceptibility remain unknown, chronic psychosocial stress related to social adversity is hypothesized to be a key component.
Methods. In this study, we use data on 1.1 million New York City (NYC) CVD emergency department (ED) visits, multiple air pollutants [fine particles (PM2.5), nitrogen dioxide (NO2), sulfur dioxide (SO2), summertime ozone (O3)], and community-level social stressors (e.g., SEP, violence, race-based residential segregation) to examine susceptibility to multiple air pollutants in CVD. First, we used ecologic cross-sectional models to examine spatial relationships among pollutants, stressors, and age-adjusted community CVD incidence rates. Second, we examined associations between spatio-temporal pollution exposures and CVD events using case-crossover models, across lag days 0-6. Finally, we are testing modification in these associations by community-level social stressors.
Results. In ecologic models, higher CVD rates were associated with NO2, PM2.5, and O3 only in areas of higher violence and physical disorder; these associations were consistent for overall CVD and all sub-diagnoses. In case-crossover models examining associations between spatio-temporal air pollution and CVD, we found significant same-day associations between NO2 and risk of any CVD event, ischemic heart disease, and heart failure; this association remained significant with any form of co-pollutant adjustment. Significant associations for PM2.5 and SO2 on all CVD, heart failure (for PM2.5), and ischemic heart disease (for SO2) were somewhat less robust to co-pollutant adjustment. We have implemented case-crossover models exploring effect modification by SEP, violence, and residential segregation. However, community stressor exposures are profoundly confounded by individual race/ethnicity in NYC; separating the simultaneous impacts of individual- and community-level modification may not be plausible in a case-crossover design with air pollution as the main exposure of interest. Because case-crossover inherently adjusts for age, the method obscures the critical fact that the median age of non-Hispanic blacks in our CVD case-only dataset is only 60 years old, versus 74 years for non-Hispanic whites. To more fully understand the role of social stressors in shaping racial differences in pollution response - more meaningfully capturing the vast differences in pollution susceptibility and CVD risk by age and race - we are now examining age at CVD event using methods which leverage longer (annual) time-scales, (e.g., survival analysis, Cox proportional hazards).
Discussion. Preliminary Cox proportional hazard results suggest a dose-response relationship between chronic (annual-average) NO2 exposure and age at CVD, with increasing levels of social stressors (i.e., segregation, violent crime, socioeconomic deprivation).
Poster by Clougherty et al., 2019 HEI Annual Conference