Associations between daily outpatient visits for respiratory diseases and ambient fine particulate matter and ozone levels in Shanghai, China

Air pollution in China has been very serious during the recent decades. However, few studies have investigated the effects of short-term exposure to PM2.5 and O-3 on daily outpatient visits for respiratory diseases. We examined the effects of PM2.5 and O-3 on the daily outpatient visits for respiratory diseases, explored the sensitivities of different population subgroups and analyzed the relative risk (RR) of PM2.5 and O-3 in different seasons in Shanghai during 2013-2016. The generalized linear model (GLM) was applied to analyze the exposure-response relationship between air pollutants (daily average PM2,5 and daily maximum 8-h average O-3), and daily outpatient visits due to respiratory diseases. The sensitivities of males and females at the ages of 15-60 yr-old and 60+ yr-old to the pollutants were also studied for the whole year and for the cold and warm months, respectively. Finally, the results of the single-day lagged model were compared with that of the moving average lag model. At lag 0 day, the RR of respiratory outpatients increased by 037% with a 10 mu g/m(3) increase in PM25. Exposure to PM2.5 (RR, 1.0047, 95% Cl, 1.0032-1.0062) was more sensitive for females than for males (RR, 1.0025, 95% CI, 1.0008 1.0041), and was more sensitive for the 15-60 yr-old (RR, 1.0041, 95% CI, 1.0027-1.0055) than the 60+ yr-old age group (RR, 1.0031, 95% Cl, 1.0014-1.0049). 03 was not significantly associated with respiratory outpatient visits during the warm periods, but was negatively associated during the cold periods. PM2.5 was more significantly in the cold periods than that in the warm periods. The results indicated that control of PM2.5, compared to O-3, in the cold periods would be more beneficial to the respiratory health in Shanghai. In addition, the single-day lagged model underestimated the relationship between PM2,5 and O-3 and outpatient visits for respiratory diseases compared to the moving average lag model. (C) 2018 Elsevier Ltd. All rights reserved.