中文摘要：

目的研究兰州地区不同季节主要污染物对人体健康的影响。方法收集2001年1月1日至2005年12月31日兰州市四所大型综合医院呼吸系统疾病逐日住院病例数据、兰州市三种主要大气污染物（PM_（10）、SO2、NO2）日均浓度数据、兰州市逐日地面气象资料,利用半参数广义相加模型（GAM）分析了2001—2005年兰州市四个季节PM_（10）与呼吸系统疾病日住院人次的暴露-反应关系,并建立了单污染物模型和多污染物模型。结果单污染模型中,区分季节,PM_（10）浓度每增加一个IQR,呼吸系统疾病不同人群住院人次相对危险度（RR）依次为：春季≥65岁的老年人群RR最高,为1.053（95%CI：1.000～1.109）,对应的最佳滞后时间为lag0;夏季男性人群RR最高,为1.032（95%CI：0.998～1.066）,对应的最佳滞后时间为lag0;秋季〈65岁的人群RR最高,为1.779（95%CI：1.026～3.088）,对应的最佳滞后时间为lag3;冬季女性人群RR最高,为1.097（95%CI：1.062～1.134）,对应的最佳滞后时间为lag4。当GAM模型中引入SO2和NO2后,不同人群PM_（10）每增加1个IQR,呼吸系统疾病住院人次的危险度有所下降,且以女性人群和≥65岁的人群对气态污染物更加敏感。结论不同季节PM_（10）浓度增加致呼吸系统疾病住院人次增加,尤其是女性和≥65岁人群。

英文摘要：

Objective To understand the effect of major pollutants on human health in different seasons in Lanzhou, Gansu province. Methods Exposure-response relationship between ambient particulate matter pollution and daily respiratory hospital visits was analyzed based on hospital admissions data of respiratory system diseases from the four biggest comprehensive hospitals in the urban areas of Lanzhou,daily SO2, NO2 and PM_（10）concentration data from the Lanzhou Environmental Monitoring Station,the meteorological data composed of daily maximum, minimum and average temperature,daily atmospheric pressure and relative humidity from the Lanzhou Meteorological Bureau,covering from 2001 to 2005,in which semi-parametric generalized additive models（GAM）,Akaike＇s Information Criterion（AIC）,relative risk（RR） calculating were used. Results Based on the single pollution model,in different seasons,with the increase of an IQR for PM_（10） concentration,the relative risks（RRs） of respiratory system diseases in different populations admissions were estimated.In spring,the highest RR was 1.053（95% CI：1.000-1.109）,for age over 65 years elderly,the corresponding optimal lag time was lag0 respectively; In summer,the highest RR was 1.032（95%CI： 0.998-1.066） for female,the corresponding optimal lag time was lag0 respectively;In autumn,the highest RR was 1.779（95%CI：1.026-3.088） for age under 65 years,the corresponding optimal lag time was lag3 respectively; In winter,the highest RR was 1.097（95%CI：1.062-1.134） for male,the corresponding optimal lag time was lag4 respectively. With the increase of an IQR for PM_（10） concentration, the exposure- response relation declined because of introduction of SO2 and NO2;Female and the population of over 65 years were more sensitive to the gaseous pollutants,which shared some of the health damage caused by PM_（10）. Conclusion The increase of PM_（10） concentration induces the relative risks（RRs） of respiratory system diseases in different seas