Department of Epidemiology and Biostatistics, School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, 13 Hangkong Rd, Wuhan, 430030, Hubei, China.
Department of Occupational and Environmental Health, Key Laboratory of Environment and Health, School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, 13 Hangkong Rd, Wuhan, 430030, Hubei, China.
J Expo Sci Environ Epidemiol. 2021 Feb;31(1):187-196. doi: 10.1038/s41370-020-0202-x. Epub 2020 Feb 13.
To assess YLL and mortality burden attributable to non-optimum ambient temperature, we collected mortality and environmental data from June 1, 2012 to December 30, 2017 in Shenzhen. We applied distributed lag nonlinear models with 21 days of lag to examine temperature-YLL and temperature-mortality associations, and calculated the attributable fractions of YLL and deaths for non-optimum temperature, including four subranges, mild cold, mild heat, extreme cold, and extreme heat. Cold and heat were distinguished by the optimum temperature, and each was separated into extreme and mild by cutoffs at 2.5th (12.2 °C) and 97.5th (30.4 °C) temperature percentile further. The optimum temperature was defined as the temperature that had minimum effect on YLL or mortality risk. The optimum temperature for non-accidental YLL was 24.5 °C, and for mortality it was 25.4 °C. Except for the population older than 65 years, the optimum temperature was generally lower in the YLL model than the mortality model. Of the total 61,576 non-accidental deaths and 1,350,835.7 YLL within the study period, 17.28% (95% empirical CI 9.42-25.14%) of YLL and 17.27% (12.70-21.34%) of mortality were attributable to non-optimum temperature. More YLL was caused by cold (10.14%, 3.94-16.36%) than by heat (7.14%, 0.47-13.88%). Mild cold (12.2-24.5 °C) was responsible for far more YLL (8.78%, 3.00-14.61%) than extreme cold (3.5-12.2 °C). As for cardiovascular deaths, only the fractions attributable to overall and cold temperature were significant, with mild cold contributing the largest fraction to YLL (16.31%, 6.85-25.82%) and mortality (16.08%, 9.77-21.22%). Most of the temperature-related YLL and mortality was attributable to mild but non-optimum weather, especially mild cold, while the YLL model implied a more prominent heat effect on premature death. Our findings can supply additional evidence from multiperspectives for health planners to define priorities and make targeted policies for mitigating the burden of adverse temperatures.
为了评估非最佳环境温度导致的 YLL 和死亡率负担，我们收集了 2012 年 6 月 1 日至 2017 年 12 月 30 日期间深圳的死亡率和环境数据。我们应用分布滞后非线性模型，以 21 天的滞后时间来检验温度与 YLL 和温度与死亡率之间的关联，并计算了非最佳温度（包括四个亚范围：轻度寒冷、轻度炎热、极度寒冷和极度炎热）导致的 YLL 和死亡的归因分数。寒冷和炎热是通过最佳温度来区分的，每个温度又通过在第 2.5 百分位数（12.2°C）和第 97.5 百分位数（30.4°C）处的截断进一步分为极端温度和温和温度。最佳温度定义为对 YLL 或死亡率风险影响最小的温度。非意外 YLL 的最佳温度为 24.5°C，死亡率的最佳温度为 25.4°C。除了年龄大于 65 岁的人群外，YLL 模型中的最佳温度通常低于死亡率模型。在研究期间，61576 例非意外死亡和 1350835.7 例 YLL 中，有 17.28%（95%经验置信区间 9.42-25.14%）的 YLL 和 17.27%（12.70-21.34%）的死亡率归因于非最佳温度。寒冷（10.14%，3.94-16.36%）导致的 YLL 比炎热（7.14%，0.47-13.88%）更多。与寒冷（12.2-24.5°C）相比，温和寒冷（12.2-24.5°C）导致的 YLL 多得多（8.78%，3.00-14.61%）。至于心血管疾病死亡，只有归因于整体温度和寒冷温度的分数具有统计学意义，其中温和寒冷对 YLL（16.31%，6.85-25.82%）和死亡率（16.08%，9.77-21.22%）的贡献最大。大部分与温度相关的 YLL 和死亡率归因于轻微但非最佳的天气，尤其是温和寒冷，而 YLL 模型表明，炎热对过早死亡的影响更为显著。我们的研究结果可以从多个角度为卫生规划者提供额外的证据，以确定优先事项并制定有针对性的政策来减轻不利温度的负担。
