Ambient ozone pollution and daily mortality in three megacities in China

Background: In mainland China, limited studies have been published on the association of ambient ozone pollution with daily mortality. Given that the air pollution has changed from conventional coal combustion to the mixed coal combustion/motor vehicle emissions(primary source of ambient ozone pollution) in the metropolis in China, it is the time to investigate the effect of ozone on adverse outcomes in the big cities. This study was designed to examine the acute mortality effects of ambient ozone pollution in Beijing, Tianjin, and Chongqing during October 1, 2013 to June 30, 2015. Methods: The relationship between daily mortality and 24-hour average ambient ozone concentrations was analyzed using generalized additive models(GAM) with Quasi-Poisson standard errors. All pollution mortality associations were adjusted for time trend, mean temperature, relative humidity, and day of the week. Sensitivity analyses were also conducted to examine the stability of the model. Effect estimates were determined for each city and then for the cities combined using a random effect method. Results: In the city-specific analysis, the percent change in daily total, respiratory, chronic lower respiratory disease(CLRD), and chronic obstructive pulmonary disease(COPD) mortality associated with an increase of 10 μg/m3 ozone concentration ranged from 0.06%(95%CI: -0.79%, 0.91%) to 0.47%(95%CI: 0.02%, 0.93%), -0.37%(95%CI: -1.80%, 1.06%) to 0.60%(95%CI: -0.66%, 1.85%), -0.04%(95%CI: -2.13%, 2.06%) to 0.48%(95%CI: -0.83%,1.79%), and -0.22%(95%CI: -2.41%, 1.96%) to 0.68%(95%CI: -1.49%, 2.86%), respectively. The pooled estimates across the cities were 0.31%(95%CI: 0.01%, 0.62%), 0.17%(95%CI: -0.63%, 0.96%), 0.33%(95%CI: -0.65%, 1.31%), and 0.32%(95%CI: -0.68%, 1.33%) for total, respiratory, CLRD, and COPD mortality, respectively. The estimates were fairly robust to the change in degree of freedoms(df) for time trend(3-10/year). However, the direction of the association and the effect size were sensitive to the different lag structures. In the two-pollutant model analysis, the effect estimates indicating the association of O3 and mortality outcomes became larger after adjusting for NO2. Conclusions: A positive yet statistically insignificant association between ambient ozone pollution and daily total mortality was observed in the three cities. Future cohort study aiming to assess the long-term health effects of ozone pollution in China should be considered.