Mortality risk of short-term air pollution exposure in urban and rural Chinese populations: A nationwide time-stratified case-crossover study, 2008-2020

Background: Epidemiologic evidence underpinning current World Health Organization Air Quality Guidelines (WHO AQGs) is primarily derived from urban populations, which remains challenging for implementing the guidelines in low- and middle-income countries (LMICs) where most people reside in rural or less developed areas. Here, we aimed to characterize the associations between ambient air pollution and mortality in Chinese populations living in both urban and rural areas, where rural populations have not been studied previously at national level. Methods: In this nationwide time-stratified case-crossover study, we extracted non-accidental death cases of all ages in both urban and rural areas during 2008-2020 from National Mortality Surveillance System, which covered 40,300 representative township-level administration units from 29 provinces, representing nearly 24 % of Chinese population. The urban-rural classification of participants' addresses was defined based on the China's National Bureau of Statistics. Daily township-level exposures to ambient particulate matter in diameter less than 10 mu m and 2.5 mu m (PM10 and PM2.5), nitrogen dioxide (NO2), carbon monoxide (CO), sulfur dioxide (SO2), and 8h maximum ozone (O3) were estimated using the Nested Air Quality Prediction Modeling System. We applied conditional logistic regression models to characterize province-specific associations of mortality risks with air pollutants, and then combined the estimates using random-effects meta-analysis. Results: A total of 61,228,962 all-cause deaths were compiled in the analysis. During the study period, 87.3 % of the death cases had PM2.5 exposure on the same day of death above current short-term guideline level of 15 mu g/ m3, and 35.6 % had O3 exposure above guideline level of 100 mu g/m3. In this analysis, significant morality risks were observed in associations with short-term exposures to all six criteria pollutants. In specific, each 10 mu g/m3 increase in PM2.5 exposure levels on the same day of death was associated with increased mortality risks of 0.22 % (95 % confidence interval [95CI %], 0.13 to 0.31). Further, when exposure levels below the guidelines of 15 mu g/m3, PM2.5 exposure attributed mortality risks increased to 1.59 % (95CI %, 0.84 to 2.35), which became largely comparable with the risks observed in high-income country studies. Notably, we derived the effects attributed to nationwide O3 exposure, with association estimate of 0.18 % (95CI %, 0.11 to 0.25), which is also comparable with global estimates. The mortality risks for major criteria pollutants PM2.5, PM10, NO2, SO2, and O3 were slightly greater among urban populations than those observed in rural populations. Conclusion: This nationwide study for the first time showed increased and globally comparable mortality risks of PM2.5 exposure below current guideline level, as well as significant morality risks of O3 exposure, in Chinese populations of all ages from both urban and rural areas. Our timely findings highlight the importance of global implementation of AQGs and call for immediate air quality management actions, particularly in less developed areas.