沙尘天气细颗粒物对呼吸及心血管系统疾病日门诊人数的影响

Association of PM_(2.5)from Dust Events with the Number of Daily Outpatient for Respiratory and Cardiovascular Diseases

[目的]研究沙尘天气大气细颗粒物(PM_(2.5))与呼吸及心血管系统疾病每日门诊人数的关系。[方法]采用半参数广义相加泊松回归模型(GAM),在控制了时间长期趋势、季节趋势、气象因素、日历效应等混杂因素影响的基础上,分析2004年3月1日至5月31日沙尘暴频发区——甘肃省武威市大气PM_(2.5)与呼吸、心血管系统疾病日门诊人数的关系。[结果]①单污染模型分析发现,PM_(2.5)与男、女总呼吸系统疾病门诊人数分别在滞后2d(lag2)和1d(lag1)的联系有统计学意义,对男、女性气管炎门诊人数的影响分别在lag1和lag3有统计学意义,对男、女性上呼吸道感染(URTI)门诊人数的影响均在lag2有统计学意义,对男性和女性肺炎门诊人数的影响在lag3和lag2有统计学意义,对男性慢性阻塞性肺部疾病(COPD)门诊人数的影响在lag2有统计学意义而对女性无统计学意义。②PM_(2.5)与男、女总心血管系统疾病门诊人数存在正相关且均在lag3有统计学意义。PM_(2.5)对男、女性风湿性心脏病门诊人数的影响分别在lag0、lag1有统计学意义,对男性高血压门诊人数在lag1有统计学意义,对男、女性缺血性心血管疾病门诊人数的影响分别在lag1和lag5有统计学意义,对男性心律失常门诊人数的影响在lag3有统计学意义,分别在lag5和lag2对男性和女性充血性心力衰竭门诊人数的影响有统计学意义。③双(或多)污染模型分析显示,引入SO_2和(或)NO_2后,PM_(2.5)对男、女性呼吸系统疾病日门诊相对危险度(RR)的影响虽有所降低,但仍然均有统计学意义。然而,在分别引入其他污染物后,SO_2和NO_2对男、女性呼吸系统疾病日门诊RR的影响均无统计学意义。④双(或多)污染模型分析还显示,引入SO_2或NO_2后,PM_(2.5)对男、女性心血管系统疾病门诊RR的影响均有所下降,但仍有统计学意义。对于男性心血管病门诊人数,在分别引入其他污染物的混杂作用后,SO_2的影响仍具有统计学意义;在引入SO_2或PM_(2.5)+SO_2之后,NO_2对男性心血管病门诊RR的影响也仍具有统计学意义;但在引入其他污染物后,SO_2或NO_2对女性心血管系统疾病门诊RR的影响均未见统计学意义。⑤沙尘天气PM_(2.5)浓度分类分析表明,从正常清洁天、轻度污染天到扬沙天气、沙尘暴天气,随着PM_(2.5)浓度水平的增大,本项目所分析的呼吸系统疾病(气管炎、URTI、肺炎、COPD)和心血管疾病(风湿性心脏病、高血压、缺血性心血管疾病、心律失常、充血性心力衰竭)门诊RR_L也随之增高,呈现一定的剂量-效应关系。[结论]①沙尘天气PM_(2.5)可引起暴露居民多种呼吸系统疾病(气管炎、URTI、肺炎、COPD)和多种心血管疾病(风湿性心脏病,高血压,缺血性心血管疾病,心律失常,充血性心力衰竭)门诊人数增加,且大多为滞后效应。②PM_(2.5)浓度与呼吸及心血管疾病门诊(RR)存在一定的剂量-效应关系。③由于沙尘天气强度与PM_(2.5)浓度的关系密切,所以男性与女性居民呼吸及心血管系统多种疾病日门诊人数相对危险度随沙尘天气的强度增大而增大,由低至高依次为:正常清洁天、轻度污染天、扬沙天、沙尘暴天。④以大气PM_(2.5)浓度水平划分沙尘天气类型比以能见度划分更为科学可靠。

[ Objective ] To explore the association between the concentrations of particulate matter ≤ 2.5μm in average aerodynamic diameter （ PM2.5 ）in the atmosphere and the number of daily outpatient for respiratory and cardiovascular （ CV ） diseases. [ Methods ] All major hospitals in the city of Wuwei were selected to report the number of outpatients for respiratory and CV diseases during March 1^st to May 31^st in 2004. The association between the concentration of PM2.5 in the atmosphere and daily outpatient number for the above mentioned diseases was investigated. A semi-parametric generalized additive Poisson regressions model （ GAM ）was fitted to the logarithm of the expected values of daily outpatient number, controlling for smooth functions of long time trends, season, meteorological variables, and calendar effect. [ Results ] （ 1 ） PM2.5 concentration was significantly correlated to the number of total respiratory diseases in males and females with a lag of 2 days and 1 day respectively. PM2.5 concentration was also significantly correlated to the number of bronchitis for males and females with a lag of 1 day and 3 days, respectively; to the number of upper respiratory tract infection（ URTI ）for males and females with a lag of 2 days; to the number of pneumonia for males and females with a lag of 3 days and 2 days ; and also to the number of chronic obstructive pulmonary diseases （ COPD ） for males with a lag of 2 days ; but not significantly correlated to the number of COPD for females ; （ 2 ） There were significant correlation between PM2.5 and the number of total CV diseases for males and females with a lag of 3 days ; between PM2.5 and the number of rheumatic heart disease for males and females with a lag of 0 days and 1 day, respectively; between PM2.5 and the number of hypertension for males with a lag of 1 day, but there was no significant correlation between PM2.5 and the number of hypertension for females. PM2.5 concentration was also significantly correlated to the number of ischemic cardiovascular diseases for males and females with a lag of 1 days and 5 days, respectively; to the number of arrhythmia for males with a lag of 3 days ; to the number of congestive heart failure for males and females with a lag of 5 days and 2 days, respectively. （ 3 ） It was shown that in co-pollutant model and multi-pollutant model analysis, after adjusting for SO2 and/or NO2, there was a decreasing effect of PM2.5 on relative risk （ RR ） of daily outpatient number for respiratory diseases with statistical significance. Nevertheless, after adjusting for other pollutants, neither SO2 nor NO2 was significant associated with the RR for daily outpatient number for respiratory diseases. （ 4 ）It was also shown in these models that, after adjusting for SO2 and/or NO2, there was a decreasing effect of PM2.5 on the RR of daily outpatient number for cardiovascular diseases with statistical significance. After adjusting for other pollutants, SO2 was still significantly associated with the RR of daily outpatient number for cardiovascular diseases for males; after adjusting for SO2 or PM2.5/SO2, NO2 was significantly associated with the RR of daily outpatient number for cardiovascular diseases for males. However, neither SO2 nor NO2 was significantly associated with the RR of daily outpatient number for cardiovascular diseases for females. （ 5 ） Categorical model of PM2.5 showed that relative risks of respiratory diseases （ bronchitis, URTI, pneumonia, COPD ） and cardiovascular diseases （ rheumatic heart disease, hypertension, ischemic cardiovascular diseases, arrhythmia, congestive heart failure ）, which we studied, increased with PM2.5 concentrations or intensity of dust events （ from normal clean day, light contaminated day to blowing sands day, dust storm day ）with a dose-response relationship. [ Conclusions ] （ 1 ）PM2.5 derived from dust events was positively associated with increasing of outpatient visits for many respiratory and cardiovascular diseases in males and females with the effect of various lags. The effect of dust and sand events on the male subjects with hypertension was higher than those on the females. （ 2 ）There was a dose-response relationship between the PM2.5 concentrations and relative risks of respiratory and cardiovascular diseases. （ 3 ） Because of strong association between concentration of PM2.5 and intensity of dust events, the relative risks of many respiratory and cardiovascular diseases increased with intensity of dust events ： normal clean day 〈 light contaminated day 〈 blowing sands day 〈 dust storm day.（ 4 ）In this paper, the categories of dust events were divided firstly according to the levels of airborne PM2.5. This method was more scientific and reliable than that according to the air visibility.