Effects of ambient PM 1 air pollution on daily emergency hospital visits in China: an epidemiological study.

Summary Background China is experiencing severe ambient air pollution. However, few studies anywhere have examined the health effects of PM 1 (particulate matter with aerodynamic diameter 2·5 (particulate matter with aerodynamic diameter 2·5 . We aimed to estimate the effects of ambient daily PM 1 and PM 2·5 concentrations on emergency hospital visits in China. Methods In this epidemiological study, we collected daily counts of emergency hospital visits from the 28 largest hospitals in 26 Chinese cities from Sept 9, 2013, to Dec 31, 2014. Ground-based monitoring data for PM 1 and PM 2·5 and meteorological data were also collected. Hospital-specific emergency hospital visits associated with PM 1 or PM 2·5 were evaluated with a time-series Poisson regression. The effect estimates were then pooled at the country level using a random-effects meta-analysis. Findings The mean daily concentration of PM 1 in all cities was 42·5 μg/m 3 (SD 34·6) and of PM 2·5 was 51·9 μg/m 3 (41·5). The mean daily number of emergency hospital visits in all hospitals was 278 (SD 173). PM 1 and PM 2·5 concentrations were significantly associated with an increased risk of emergency hospital visits at lag 0–2 days (cumulative relative risk [RRs] 1·011 [95% CI 1·006–1·017] for a 10 μg/m 3 increase in PM 1 and 1·010 [1·005–1·016] for a 10 μg/m 3 increase in PM 2·5 ). Slightly higher RRs of ambient PM 1 and PM 2·5 pollution were noted among women and children than among men and adults, respectively, but without statistical significance. Given a cause-effect association, 4·47% (95% CI 2·05–6·79) and 5·05% (2·23–7·75) of daily emergency hospital visits in China could be attributed to ambient PM 1 and PM 2·5 pollution, respectively. Interpretation Exposure to both ambient PM 1 and PM 2·5 were significantly associated with increased emergency hospital visits. The results suggest that most of the health effects of PM 2·5 come from PM 1 . Funding None.