You are here:
A case-crossover analyses of fine particulate matter and out-of-hospital sudden unexpected death
Citation:
Rappazzo, K., G. Joodi, S. Hoffman, I. Pursell, J. Mounsey, W. Cascio, AND R. Simpson. A case-crossover analyses of fine particulate matter and out-of-hospital sudden unexpected death. Society for Epidemiologic Research Annual Meeting, Seattle, Washington, June 20 - 23, 2017.
Impact/Purpose:
We used a case-crossover study design to investigate the associations between fine particulate matter (PM2.5) and out-of-hospital sudden unexpected deaths.
Description:
Out-of-hospital sudden unexpected deaths (OHSUD) are natural deaths that occur without obvious underlying causes and account for nearly 1 in 6 deaths in the United States. Ambient air pollution is known to be causally related to overall mortality, therefore, we hypothesized that short-term exposure to air particle pollution is associated with OHSUD. We used a case-crossover study design to investigate the associations between fine particulate matter (PM2.5) and OHSUD. Preliminary cases were identified through the University of North Carolina at Chapel Hill SUDDEN project, an ongoing case-only cohort of OHSUD in Wake County, NC. Cases were individuals 18-64 years of age, who resided in Wake County, with no obvious cause of death (e.g., violence, drug use, etc.), and died between 3/01/2013-12/31/2014 (n=399). Control periods were selected on the same day of the week within the calendar month/year of the case death. Hourly PM2.5, temperature, and relative humidity measurements from a central site in Wake County were obtained from the EPA’s Air Quality Data Mart. Single day average PM2.5 exposures were assigned to each case and control period for lags of 0 to 3 days. Conditional logistic regression was used to estimate OHSUD ORs (95% CI) for a 1µg increase in PM2.5, adjusted for same-day average temperature and relative humidity (both natural splines). OHSUD OR for lag0 was: 0.980 (0.948, 1.013), and for lag1: 1.025 (0.993, 1.057). ORs for other lags were null. While the inverse association observed at lag0 was unexpected, the positive association observed at lag1 echoes PM2.5-mortality associations observed in large mortality studies. There may be confounding through co-pollutant exposures, or exposure misclassification due to the use of a single monitor location. Exploration of more spatially resolved models of air pollutant concentration estimates and co-pollutant confounding are planned for future analyses. This abstract does not necessarily reflect EPA policy.