Objective To evaluate the interactive effects of fine particulate matter and temperature on non-accidental mortality of residents in Pudong, Shanghai.

Methods Daily mortality, air pollutants and meteorological data from Jan 1^st.2016 to Dec 31^st.2017 were collected.Generalized additive Poisson regression models was used to estimate the effects of PM_2.5 pollution on daily mortality, bivariate response surface models and temperature stratified models were applied to examine the interaction of temperature with PM_2.5 on mortality.

Results A total of 43 345 non-accidental deaths were included, daily mean PM_2.5 concentration was 39.1 μg/m³, daily mean temperature was 17.7 ℃.A 10 μg/m³ increase in the daily PM_2.5 at lag 1 day was associated with a 0.56%(95%CI:0.11%-1.01%), 0.49%(95%CI:-0.19%-1.18%) and 0.22%(95%CI:-1.14%-1.60%) increase in non-accidental, cardiovascular and respiratory mortality, respectively.Higher risks were identified for males and the older (≥65 years).The effect estimates per 10 μg/m³ increase in PM_2.5 in medium temperature level were 0.59%(95%CI:0.04%-1.14%)for non-accidental, 0.51%(95%CI:-0.32%-1.35%)for cardiovascular and 0.51%(95%CI:-0.32%-1.35%) for respiratory mortalities.The PM_2.5 effects were approximately 2-4 times higher in higher temperature level for non-accidental and cardiovascular mortalities compared with other temperature levels; for respiratory mortality, the PM_2.5 effects was approximately 2-fold higher in lower temperature levels than the medium, although the interactions between temperature and PM_2.5 were statistically insignificant.

Conclusions Temperature may modify the effects of PM_2.5 on mortality in Pudong, Shanghai and the interactive pattern may be different across disease-specific mortalities.