Background: Systematic evaluations of the cumulative effects and mortality displacement of ambient particulate matter (PM) pollution on deaths are lacking. We aimed to discern the cumulative effect profile of PM exposure, and investigate the presence of mortality displacement in a large-scale population.
Methods: We conducted a time-series analysis with different exposure-lag models on 13 cities in Jiangsu, China, to estimate the effects of PM pollution on non-accidental, cardiovascular, and respiratory mortality (2015-2019). Over-dispersed Poisson generalized additive models were integrated with distributed lag models to estimate cumulative exposure effects, and assess mortality displacement.
Results: Pooled cumulative effect estimates with lags of 0-7 and 0-14 days were substantially larger than those with single-day and 2-day moving average lags. For each 10 μg/m3 increment in PM2.5 concentration with a cumulative lag of 0-7 days, we estimated an increase of 0.50 % (95 % CI: 0.29, 0.72), 0.63 % (95 % CI: 0.38, 0.88), and 0.50 % (95 % CI: 0.01, 1.01) in pooled estimates of non-accidental, cardiovascular, and respiratory mortality, respectively. Both PM10 and PM2.5 were associated with significant increases in non-accidental and cardiovascular mortality with a cumulative lag of 0-14 days. We observed mortality displacement within 30 days for non-accidental, cardiovascular, and respiratory deaths.
Conclusions: Our findings suggest that risk assessment based on single-day or 2-day moving average lag structures may underestimate the adverse effects of PM pollution. The cumulative effects of PM exposure on non-accidental and cardiovascular mortality can last up to 14 days. Evidence of mortality displacement for non-accidental, cardiovascular, and respiratory deaths was found.
Keywords: Coarse particulate matter; Deaths; Distributed lag model; Fine particulate matter; Mortality displacement.
Copyright © 2021. Published by Elsevier Ltd.