The burden of air pollution on years of life lost in Beijing, China, 2004-08: retrospective regression analysis of daily deaths

Abstract
Objectives To better understand the burden of air pollution on deaths, we examined the effects of air pollutants on years of life lost (YLL) in Beijing, China. Design Retrospective regression analysis using daily time series. Setting 8 urban districts in Beijing, China. Participants 80 515 deaths (48 802 male, 31 713 female) recorded by the Beijing death classification system during 2004-08. Main outcome measures Associations between daily YLL and ambient air pollutants (particulate matter with aerodynamic diameter 2.5), PM10, SO2, and NO2), after adjusting for long term trends, seasonality, day of the week, and weather conditions. We also examined mortality risk related to air pollutants. Results Mean concentrations of daily PM2.5, PM10, SO2 and NO2 were 105.1 μg/m3, 144.6 μg/m3, 48.6 μg/m3, and 64.2 μg/m3, respectively. All air pollutants had significant effects on years of life lost when we used single pollutant models. An interquartile range (IQR) increase in PM2.5, PM10, SO2, and NO2 was related to YLL increases of 15.8, 15.8, 16.2, and 15.1 years, respectively. The effects of air pollutants on YLL appeared acutely and lasted for two days (lag 0-1); these effects associated with an IQR increase in PM2.5 were greater in women than men (11.1 (95% confidence interval 4.7 to 17.5) v 4.7 (−2.9 to 12.3) YLL) and in people aged up to 65 years than those older than 65 years (12.0 (2.9 to 21) v 3.8 (−0.9 to 8.6) YLL). The mortality risk associated with an IQR increase in PM2.5 was greater for people older than 65 years (2.5% (95% confidence interval 0.6% to 4.5%) increase of mortality) than those aged up to 65 years (0.7% (−0.8% to 2.2%)). Conclusions YLL provides a complementary measure for examining the effect of air pollutants on mortality. Increased YLL are associated with increased air pollution. This study highlights the need to reduce air pollution in Beijing, China, to protect the health of the population.