广州市炎热指数对早产滞后效应的时间序列分析

Lag effects of heat index on preterm birth in Guangzhou:a time-series study

目的了解广州市炎热指数对早产的滞后效应。方法以广州市白云区、越秀区2004年1月1日至2011年12月31日间分娩的产妇为研究对象,符合纳入标准的共293 849人,其中发生早产者22 702例。收集该时期产妇分娩数据、气象资料和空气污染物数据。采用分布滞后非线性模型,以广州市年均炎热指数为参照,建立炎热指数在滞后0-20 d内对逐日早产例数的回归模型,分析产妇暴露于不同炎热指数后0-20 d内发生早产的相对危险度。结果研究期间广州市白云区、越秀区的早产发生率为7.70%;炎热指数波动区间为45-136℉,年均炎热指数为（101±20）℉,以炎热指数101℉为参照,炎热指数升高20℉时,滞后第5-7日的早产发生危险性升高,其中第6日风险最高（RR=1.014,95%CI：1.005-1.023）;第10-12日、19日的早产发生危险性降低,其中第19日风险最低（RR=0.985,95%CI：0.974-0.997）。炎热指数降低20℉时,滞后第5-7日的早产发生危险性降低,其中第6日风险最低（RR=0.986,95%CI：0.978-0.995）;第10-12日、19日的早产发生危险性升高,其中第19日风险最高（RR=1.015,95%CI：1.003-1.027）。结论广州市炎热指数对早产的发生呈非线性滞后影响。炎热指数升高时,早产相对危险度先增加后降低;炎热指数降低时,早产相对危险度先降低后增加。

Objective To estimate the effects of heat index on preterm birth in Guangzhou. Methods Daily preterm births,meteorological and air pollution data during 2004-2011 were used to explore the association between heat index and preterm birth. Distributed lag non-linear time series model was fitted for access the lag effects of each heat index value on preterm birth, based on Poisson regression. Results The exposure-response functions appeared to be lag-non-linear. During 2004-1-1 to 2011-12-31, preterm birth rate was 7.70%, mean heat index was（101±20）℉. For the exposure-response effects from lag0 day to lag20 day with heat index reference at 101℉, with a 20 unit increase in heat index, The preterm birth risks were different with statistical significant in periods lag 5 day to lag7 day, lag10 day to lag12 day and lag19 day. the RR value of lag6 day and lag19 day were 1.014（95%CI：1.005-1.023）, 0.985（95%CI：0.974-0.997）;With a 20 unit decrease in heat index, the preterm birth risks were different with statistical significant in periods during lag5 day to lag7 day, lag10 day to lag12 day and lag19 day. The RR value of lag6 and lag19 day were 0.986（95%CI： 0.978-0.995）,1.015（95%CI：1.003-1.027）. Conclusion In Guangzhou, heat index presents a non-linear lag effect on preterm birth.