母孕期使用抗生素对后代患哮喘风险的Meta分析

Meta-analysis of the relationship between prenatal antibiotic use and risk of asthma in offspring

目的评估母亲孕期使用抗生素与儿童发生哮喘的关系。方法检索中国知网数据库、中国维普科技期刊数据库、万方数据库以及Pub Med、EMBASE、Cochrane、Ovid等数据库中关于母孕期使用抗生素与儿童哮喘的队列研究,采用Stata12.0软件,通过Meta分析方法探讨二者之间的关系。结果纳入有儿童哮喘患病率和儿童哮喘暴露因素关联性结果(OR、RR或HR)的文献,共纳入9项研究,均调整了母亲哮喘、吸烟等混杂因素,Meta分析结果显示:(1)母孕期抗生素的使用增加了儿童哮喘的风险(OR=1.14,95%CI:1.13~1.15),I^2=96.5%,逐篇去除行敏感性分析,再次汇总OR=1.27(95%CI:1.17~1.38,I^2=0),未对最终结果造成明显影响;临床异质性分析:以3岁为界行亚组分析,<3岁和≥3岁亚组文献汇总OR分别为1.85(95%CI:0.80~4.29,I^2=78.8%)和1.19(95%CI:1.08~1.31,I^2=96.8%);以不同孕期行亚组分析,孕早、中和晚期OR分别为1.29(95%CI:1.23~1.34)、1.30(95%CI:1.25~1.35)和1.26(95%CI:1.21~1.31),I2均为0;以不同暴露因素获取方式行亚组分析,自我报告(采访)和来源于数据库文献汇总OR分别为1.27(95%CI:1.10~1.48,I^2=71.2%)和1.20(95%CI:1.08~1.32,I^2=98.6%);以不同抗生素行亚组分析,β-内酰胺类OR=1.18(95%CI:1.08~1.30,I^2=0),磺胺嘧啶OR=1.19(95%CI:0.78~1.83,I^2=83.5%)。(2)以同胞为亚组分析,文献汇总OR=0.91(95%CI:0.79~1.06,I^2=93.3%)。分别采用Begg秩相关法和Egger直线回归法未发现明显发表偏倚。结论本Meta分析显示母孕期使用抗生素可增加后代哮喘的风险,但同胞对照组研究显示母孕期使用抗生素与后代哮喘不关联,家庭环境、遗传等残余因素可能是重要的混杂因素,孕期抗生素使用与后代哮喘的因果关系有待进一步研究。

Objective To evaluate the association between the antibiotic use during pregnancy and childhood asthma.Methods The Chinese and English databases CNKI,Wanfang Data,Chinese VIP science and technology periodical database,PubMed,EMBASE,Cochrane and ovid were searched for cohort studies on the association between the use of antibacterial agents in pregnancy and childhood asthma.This meta-analysis with Stata12.0software explored the relationship between antibiotic and asthma.Results There were9studies with the prevalence of childhood asthma and asthma exposure results related factors(OR,RR or HR)and were adjusted for maternal asthma,smoking and other confounding factors.The results of the meta-analysis showed that the use of antibacterial agents in pregnancy increased the risk of childhood asthma(OR=1.14,95%CI:1.13-1.15).Because of the high heterogeneity(I2=96.5%),analysis of the heterogeneity among literatures was carried out:the sensitivity analysis was made by removing each article,and the pooled risk estimates of asthma(OR=1.27,95%CI:1.17-1.38,I2=0)after excluding the studies of high heterogeneity that did not affect the final result;analysis of clinical heterogeneity:the subgroup analysis with the age of3years of asthma onset was performed,and the pooled effect size OR of studies with<3years and≥3years respectively was1.85(95%CI:0.80-4.29,I2=78.8%)and OR=1.19(95%CI:1.08-1.31,I2=96.8%);stratified by trimesters of pregnancy,the OR was1.29(95%CI:1.23-1.34),1.30(95%CI:1.25-1.35)and1.26(95%CI:1.21-1.31)for the early-,mid-and late-pregnancy,respectively,and I2=0;stratified by antibiotic assessment,the pooled OR from self reporting(interviews)and from the database was1.27(95%CI:1.10-1.48,I2=71.2%),and1.20(95%CI:1.08-1.32,I2=98.6%)respectively;stratified by antibiotic tepy,the OR was1.18(95%CI:1.08-1.30,I2=0)for Beta-lactam antibiotics and1.19(95%CI:0.78-1.83,I2=83.5%)for Sulfonamides and trimethoprim.②In sibling analyses,the pooled OR was0.91(95%CI:0.79-1.06,I2=93.3%).There was no publication bias by the Begg rank correlation test and the Egger linear regression test.Conclusion This meta-analysis suggested that antibiotic exposure during pregnancy may increase the risk of asthma in offspring.But in the sibling control analysis the associations disappeared,indicating the residual factors such as family environment,heredity and so on were important confounding factors that may lead to reverse causality.Therefore,the causal relationship between antibiotic use during pregnancy and risk of asthma in offspring needs further investigation.