北京市某区居民死亡与大气表观温度的相关性研究

Relationship between Ambient Apparent Temperature and Daily Mortality of Population in a District of Beijing

目的探讨大气表观温度与北京市某区居民全死因死亡之间的相关性。方法收集2004—2008年北京市某区居民日全死因死亡人数和该地区日均气温、相对湿度和风速的数据以及该城区相应的PM10、SO2和NO2日均浓度的时间序列数据。采用时间序列分析方法,利用立方平滑样条函数,将根据气温和相对湿度计算出的大气表观温度、风速以及作为潜在混杂因素的PM10、SO2和NO2日均浓度引入,同北京市某区居民日全死因死亡人数间建立Poisson广义相加模型。根据AIC最小的原则,选择最终进入模型的变量,并确定其立方平滑样条函数自由度取值。结果北京市某区日均表观温度与全死因死亡人次间存在负相关关系(P<0.01);日平均表观温度与日最高表观温度相比,日最高表观温度的RR值较大;在对全死因死亡人群男女分层结果显示,日均表观温度、日最高表观温度为男性全死因死亡的保护因素,滞后天数为3 d时到最大效应,而日平均表观温度、日最高表观温度及日最低表观温度均为女性全死因死亡的危险因素,滞后天数为1 d时达到最大效应。结论大气表观温度与暴露人群全死因死亡有相关,夏季日均表观温度增高可使女性暴露人群全死因死亡增加。

Objective To analyze the relationship between apparent temperature and the daily total mortality of the exposed population in a district of Beijing. Methods The data of the daily all-cause mortality of residents and the daily air temperature, relative humidity and wind velocity were collected, which on the data of the relevant apparent temperature was transferred based, in the subject district. The data of level of PM10, SO2 and NO2 were also collected. Using time series analysis, cubic smoothing spline function, the apparent temperature, daily average temperature, relative humidity and wind velocity as well as potential confounding factors like daily average density of PM10, SO2 and NO2 were introduced, a Poisson generalized additive model with the daily all-cause mortality of the district of Beijing were built. According to the rule of AIC minimization, the variables which finally were able to be in the model were chosen, and the degree of freedom for the cubic smoothing spline function were determined. Results The daily average apparent temperature had a linear positive correlation with the total mortality; compared with daily average apparent temperature, the highest daily temperature had higher RR value. The gender stratified result for total mortality showed that daily average apparent temperature and the highest daily apparent temperature had a negative correlation with the male total mortality, and the effect was maximized when the time lag was three days, while the female total mortality had a positive correlation with all the three temperature indexes, the effect was maximized when the time lag was one day. Conclusion Apparent temperature has significant correlation with the daily total mortality of the exposed population in a district of Beijing, the increase of daily average apparent temperature in summer may cause increase in female mortality of the exposed population in this district.