| Summary: | Epidemiological studies have quantitatively linked adverse health outcomes, including stroke, to ambient temperature. However, such analyses cannot be applied to estimate disease burden attributable to climate change, because they assume a theoretical minimum risk reference based on an optimal temperature, which is a post hoc estimator that cannot be generalized to the unstudied locations. In this study, we used temperature anomaly (TA), an alternative indicator of climate change, in epidemiological studies to address this issue. We associated geo-coded nationwide stroke data (n = 11 144) from 2000 to 2016 in China to TA with a lag-distributed nonlinear model. For 0–2 lagged days, the risk of stroke increased by 20% (95% confidence interval [CI]: 3%–40%) for irregular cold and 19% (95% CI: 3%–37%) for irregular heat. The exposure-response function was estimated as a U-shaped curve centered at a TA value of 0. Based on the function, an estimated 3.42% of stroke cases were attributable to TA. In addition, there were more strokes attributable to heat (2.05%) than cold (1.38%). TA-attributable strokes due to climate change increased by 0.019% (95% CI: 0.008%–0.031%) per year. Furthermore, the long-term trend was dominated by the increase in events involving heat. In contrast to cold-attributable stroke, which potentially decreased by 0.004% (95% CI: −0.005%–0.013%) per year, the heat-attributable burden increased by 0.023% (95% CI: 0.012%–0.034%) per year. We also found that ischemic stroke was associated with cold (odds ratio [OR] = 1.18, 95% CI: 1.00–1.38), while hemorrhagic stroke was associated with heat (OR = 1.66, 1.07–2.58). Our findings are of importance to public health for climate change mitigation and for predictions of future effects in climate change scenarios.
|
|---|