An application of satellite-based surface air quality estimates to evaluate city-to-city variation in short-term effects of PM2.5 on morbidity and mortality

摘要

Background: Recent advancement in satellite remote sensing has allowed estimation of ground-level concentrations of fine particles (PM2.5) and nitrogen dioxide (NO2) in locations where no ground monitors exists. We hypothesize that the observed heterogeneity in PM2.5 short-term risks across cities is explained by variation in NO2, an indicator of combustion sources. Cooper et al. recently proposed a satellite-based multi-pollutant index (SATMPI) that scales local PM2.5 concentration by its NO2 concentration (doi:10.1021/es302672p) partially based on this hypothesis. Aims: To investigate feasibility of using satellite-based air pollutant measurements to explain the city-to-city variation in PM2.5 risk estimates for morbidity and mortality. Methods: We conducted a random effects second-stage analysis to model lag 0 day PM2.5 risk estimates for elderly cardiovascular hospitalizations (2000-2008) and all-cause mortality (2001-2006) from the time-series analysis of 148 U.S. cities in the National Particle Component Toxicity project as a function of city-to-city variation in satellite-based estimation of NO2, PM2.5, and SATMPI. NO2 data from ground monitors were available for only 95 cities, but satellite-based NO2 was available for 148 cities. The influence of the satellite-based air pollutant was expressed as a percent change in the risk estimate, per its interquartile-range (IQR), from the combined estimate without predictors. Results: The satellite-based NO2 was a positive predictor of city-to-city variation in PM2.5 cardiovascular hospitalization risk estimate with a 42.9% increase (95% confidence interval: 0.0, 85.8) per IQR change. Satellite-based PM2.5 and SATMPI resulted in similar magnitude of impact with less significance. The satellite-based indices on mortality risk estimates showed similar magnitude of impacts with even less significance. Conclusion: Satellite-based air quality indices that combine PM2.5 and NO2 may be useful for health effects studies.