某市空气污染对早产急性影响的Poisson广义可加模型分析

Acute Effect of Air Pollution on Preterm Birth in a City

目的研究广东某市大气污染物对早产的急性影响。方法根据2007年广东某市的出生监测系统、围产保健数据和住院分娩病历获得妊娠结局资料,从广东某市气象局获得2007年气象资料,2007年大气污染资料来源于广东某市环境监测站。采用广义可加模型(general additive model,GAM)进行广东某市大气污染与新生儿早产发生率的Poisson回归分析,控制气象因素、时间趋势、工作日效应混杂因素的影响。结果 2007年广东某市空气中NO2、PM10、SO2的日均浓度分别为61.04、82.51、51.67μg/m3;2007年广东某市新生儿平均出生早产数为21.47例/日。Pearson相关分析结果表明,NO2、PM10、SO2浓度与温度及相对湿度均呈负相关。滞后效应和累积效应GAM时序分析结果均显示,单污染物的健康效应显著。在滞后效应中3种污染物的健康效应均仅维持在当天,当大气中NO2、PM10、SO2浓度每升高100μg/m3,其RR值分别为1.0425(95%CI:1.0068～1.0781)、1.0512(95%CI:1.0087～1.0938)、1.1118(95%CI:1.0479～1.1757)。在累积效应中NO2、SO2的健康效应均在滞后3d达到最大值,每升高100μg/m3,其RR值分别为1.0542(95%CI:1.0080～1.1003)、1.1298(95%CI:1.0480～1.2116);PM10的健康效应在滞后4d达到最大值,每升高100μg/m3,其RR值为1.0688(95%CI:1.0074～1.1301)。多污染物模型GAM时序分析结果显示,在SO2加入NO2的双污染物模型中SO2滞后效应达最大值,而NO2和PM10的滞后效应和累积效应均有下降,且在模型中未呈现统计学意义。结论广东某市大气NO2、PM10、SO2污染对新生儿早产发生率具有潜在的急性影响。

Objective To analyze the acute effect of air pollutants on preterm birth of newborn. Methods The pregnancy outcome data were from the birth surveillance system, perinatal heath data and hospital birth records of a city of Guangdong province. Meteorological data and air pollution data were collected from the weather bureau of the investigated city and the environmental monitoring station respectively. Generalized additive model （GAM） and Poisson regression analysis were used to examine the relationship between air pollution and preterm birth in a city of Guangdong province, controlled with time trends, weather and the day of week. Results The average concentration of NO2, PM10 and SO2 in 2007 were 61.04, 82.51 and 51.67 μg/m3 respectively. The average daily count of preterm birth of newborn in 2007 was 21.47/daily. Pearson correlation analysis presented that there was a negative correlation between the concentration of NO2, PM10, SO2 and both the temperature and relative humidity of atmospheric. The lagged and accumulation effect of GAM suggested the single pollutant models was significant. The lagged effect of both three kinds of pollutants could only be maintained on a day. An increase of 100 μg/m3 of NO2, PM 10 and SO2 corresponded to 1.042 5 （95%CI： 1.006 8-1.078 1 ）, 1.051 2 （95%CI： 1.008 7 -1.093 8）, 1.111 8 （95%CI： 1.047 9 -1.175 7）relative risk of preterm birth respectively. The cumulative effects of NO2, SO2 also reached the maximum in three days later. An increase of 100 μg/m3 of NO2 and SO2 corresponded to 1.054 2 （ 95%CI： 1.008 0-1.100 3 ）and 1.129 8 （ 95%CI： 1.048 0 - 1.211 6 ）relative risk of preterm birth respectively. The cumulative effects of PM10 also reached the maximum in four days later. An increase of 100 μg/m3 of PM10 corresponded to 1.068 8 （95%CI： 1.007 4-1.130 1 ）relative risk of preterm birth. The results of multiple air pollutants GAM showed that the lagged effect of SO2 were all strongest when NO2 were included in GAM, while the effect of NO2 and PM10 were weakened in multiple air pollutants GAM. Conclusion Air pollutants of NO2, PM10 and SO2 have potential acute effect on preterm birth in the investigated city in Guangdong.