基于meta回归方法的PM_(10)短期致死效应的剂量-反应关系参数的异质性来源

Sources of Heterogeneity on Parameters of Dose-Response Relationship Between PM_(10) and Short-Term Mortality by Meta Regression Method

[目的]探讨PM_(10)日平均浓度与某病死亡率的剂量-反应关系参数的异质性来源,为城市间流行病学研究结果的相互借用提供参考。[方法]检索中国期刊全文数据库(CJFD)、Pub Med及Elsevier Science Direct等数据库,全面收集PM_(10)日平均浓度与死亡率的剂量-反应关系的相关研究。使用Stata 12.0软件,采用单因素和多因素相结合的meta回归分析方法,分病种对PM_(10)短期致死效应的异质性来源进行分析。[结果]最终纳入11篇文献,包括了中国、美国和欧洲、拉丁美洲、亚洲等国家和地区所属的67个城市样本。Meta回归结果显示,PM_(10)背景浓度、温度、相对湿度、65岁以上人口比重和人均国内生产总值(GDP)等因素,对PM_(10)与心脑血管疾病死亡率剂量-反应关系参数有显著影响:在温度较高、65岁以上人口比重较大、PM_(10)背景浓度较低、相对湿度较低、人均GDP较低地区,PM_(10)每增加10μg/m3,人群心脑血管疾病死亡率相较于基期水平的增加会高于温度较低、65岁以上人口比重较小、PM_(10)背景浓度较高、相对湿度较高和人均GDP较高地区。PM_(10)背景浓度、温度和人均GDP等因素,对PM_(10)与呼吸系统疾病死亡率的剂量-反应关系的参数有显著影响:在温度较高、PM_(10)背景浓度较低、人均GDP较低地区,PM_(10)浓度每增加10μg/m3,人群呼吸系统疾病死亡率相较于基期水平的增加会高于温度较低、PM_(10)背景浓度较高、人均GDP较高地区。[结论]PM_(10)短期致死效应的剂量-反应关系参数的异质性来源包括PM_(10)背景浓度、温度、相对湿度、65岁以上人口比重和人均GDP。

[ Objective ] To identify the sources of heterogeneity on parameters of dose-response relationship between PM10 daily concentration and mortality rate of certain diseases, so as to provide reference for transferring benefits among inter-city epidemiological studies. [ Methods ] Epidemiological literatures on dose-response relationship between PM10 daily concentration and mortality rate were collected by systematically searching China Journal Full-text Database （CJFD）, PubMed database, and Elsevier Science Direct database. Stata 12.0 software was used to conduct single-factor and multi-factor meta regression analyses on the heterogeneity sources of the PM10 induced short-time mortalities by disease categories. [ Results ] Eleven literatures were finally included, covering 67 cities from China, USA, Europe, Latin America, and other Asia countries or areas. The results of meta regression analysis showed that factors such as PM10 background concentration, temperature, humidity, proportion of population over 65 years old, and GDP per capita had significant effects on the dose-response relationship between PM10 daily concentration and cardiovascular mortality rate. In cities with higher temperature and proportion of population over 65 years old as well as lower PM10 background concentration, relative humidity, and GDP per capita, the increment of population-based cardiovascular mortality rate associated with each 10- μg/m^3 increase of PM10 compared to the background level was greater than that in the cities with corresponding factors skewed to the opposite direction. The results also showed that factors like PM10 background concentration, temperature, and GDP per capita had significant effects on the dose-response relationship between PM10 and respiratory mortality rate. In cities with higher temperature and lower PM10 background concentration and GDP per capita, the increment of population- based respiratory mortality rate associated with each 10-μ/m^3 increase of PM10 was greater than that in the cities with corresponding factors skewed to the opposite direction. [ Conclusion ] The sources of heterogeneity on parameters of dose-response relationship between PM10 and short-term mortality include PM10 background concentration, temperature, relative humidity, proportion of population over 65 years old, and GDP per capita.