[Influence of extreme weather on years of life lost due to diabetes death in Chongqing and Harbin, China].

Objective: To understand the associations between extremely low and high air temperature and the years of life lost (YLL) due to diabetes deaths in Chongqing and Harbin with different climatic characteristics in China. Methods: A double threshold B-spline distributed lag non-linear model (DLNM) was used to investigate the lag and cumulative effects of extremely low and high air temperature on YLL due to diabetes for lag 0-30 days by using the urban meteorological and diabetes mortality data of Chongqing (2011-2013) and Harbin (2008-2010). The effects were expressed as relative risk (RR). Results: In Chongqing, the cold effects on YLL due to diabetes were delayed by four days and lasted for three days (lag4-6) with the highest RR of 1.304 (95%CI:1.033-1.647) at lag5. The hot effects were delayed by one day (lag1) with RR of 1.321 (95%CI:1.061-1.646). In Harbin, the extreme cold effects on YLL were delayed by four days and lasted for seven days (lag4-10) with the highest RR of 1.309 (95%CI: 1.088-1.575) at lag6. The hot effects were delayed by one day and lasted for four days (lag1-4) with the highest RR of 1.460 (95%CI:1.114-1.915) at lag2. The unit risk for cold and hot effects was 43.7% (P=0.005 5) and 18.0% (P=0.000 2) in Chongqing and 15.0% (P=0.000 8) and 29.5% (P=0.001 2) in Harbin, respectively. Conclusions: Both extremely low air temperature and extremely high air temperature might increase the years of life lost due to diabetes in cities with different climate characteristics. Health education about diabetes prevention should provide information about the effects of extreme weather events.目的： 了解重庆和哈尔滨市极端高温和低温与糖尿病引起的生命损失年之间的关系。 方法： 利用重庆市（2011-2013年）和哈尔滨市（2008-2010年）的气象和糖尿病死亡数据，采用分布滞后非线性模型（DLNM）分析极端高温和低温对糖尿病引起的生命损失年的滞后效应和累积效应，以相对危险度（RR）表示。 结果： 在重庆市，冷效应对生命损失年的影响滞后4 d、持续3 d (lag4~ 6），最大RR值为1.304 (95%CI：1.033~ 1.647），出现在低温发生后第5天（lag5）；热效应滞后1 d，其RR值为1.321 (95%CI：1.061~ 1.646）。在哈尔滨市，极端低温对生命损失年的影响滞后4 d、持续7 d (lag4~ 10），最大RR值为1.309 （95 %CI：1.088~ 1.575），出现在低温发生后第6天（lag6）；热效应滞后1 d、持续4 d (lag1~ 4），最大RR值为1.460 （95%CI：1.114~ 1.915），出现在高温后第2天（lag2 ）。重庆市冷效应和热效应的单位风险分别为43.7%（P＝0.005 5）和18.0% （P＝0.000 2），哈尔滨市冷效应和热效应的单位风险分别为15.0%（P＝0.000 8）和29.5%（P＝0.001 2）。 结论： 重庆和哈尔滨市极端高温和低温都可增加糖尿病引起的生命损失年。极端温度对糖尿病的影响应纳入糖尿病健康教育内容。.