LAUSR

Exposure to particulate matter of smaller than 2.5 μm in diameter (PM2.5) is linked to increased human mortality, and could be further complicated by concurrent ambient air temperatures. Published reports indicate that the association between ambient temperatures and mortality due to PM2.5 exposure is dissimilar across different geographic areas. Thus, it is unclear how ambient temperatures at different geographic locations can together modulate the influence of PM2.5 on mortality. In this paper, we examined how temperature modulated the association between mortality and PM2.5 exposure in 15 Chinese cities during 2014-2016. For analysis, First, Poisson generalized additive models under different temperature stratifications (<10th, 10-90th, and >90th temperature percentiles) was used to estimate PM2.5 associations to mortality, which were specific to different cities. Second, we used a meta-analysis to combine the effects at each temperature stratum and region (southern and northern China). Results revealed that high temperatures (daily mean temperature >90th percentile) robustly amplified observed associations of mortality and PM2.5 exposure, and the modifications were heterogeneous geographically. In the northern regions, a 10 μg/m3 increment in PM2.5 was associated with 0.18%, 0.28%, and 1.54% increase in non-accidental mortalities and 0.33%, 0.39%, and 1.32% increase in cardiovascular mortalities at low, moderate, and high temperature levels, respectively. In the southern regions, a 10 μg/m3 increment in PM2.5 was associated with 0.52%, 0.62%, and 1.90% increase in non-accidental mortalities and 0.55%, 0.98%, and 2.25% increase in cardiovascular mortalities at low, moderate, and high temperature levels, respectively. It is concluded that temperature altered PM2.5-mortality associations in southern and northern China synergistically, but the effect was more pronounced in the south. Therefore, geography and temperature need to be considered when studying how PM2.5 affects health.