Objective: To evaluate the relevance of diabetes mellitus and hepatitis B virus(HBV) infection in people by Meta-analysis. Methods: Databases we searched included CNKI, VIP China Science and Technology Journal Database, Wanfang Data Knowledge Service Platform, PubMed, Cochrane Library and Web of Science Core Collection database. Publication time was from January 1997 to May 2017. The Languages were limited to Chinese and English. English search terms include: diabetes, diabetes mellitus, hepatitis B and risk. Chinese search terms include: diabetes mellitus, hepatitis B and risk. We included all observational studies on diabetes and HBV infection. Firstly, the Newcastle-Ottawa Scale and the evaluation criteria of Cross-sectional study recommended by Agency for Healthcare Research and Quality were used to evaluate the quality of articles. Secondly, RevMan 5.3 software was used for heterogeneity testing. Subgroup analysis, random effects model and Mantel-Haenszel method were used to calculate the combined OR value. Finally, Stata 14.0 software was used to conduct the sensitivity analysis, and Begg rank correlation method was used to detect the publication bias. Results: A total of 12 studies were included, comprising 6 cross-sectional studies and 6 case-control studies. But the 12 articles were heterogeneous (χ(2)=42.10, P<0.001). After subgroup analysis, cross-sectional studies of diabetes and HBV infection were still heterogeneous (χ(2)=28.21, P<0.001), whose combined odds ratio (OR) was 1.36 (95%CI: 1.03-1.80). But the heterogeneity of case-control studies was not statistically significant (χ(2)=10.32, P=0.070), whose combined odds ratio (OR) was 1.55 (95%CI: 1.10-2.17). After the sensitivity analysis, the 95%CI of the combined OR of the 7 studies did not fluctuate, and the lower limit was above 1.07. No publication bias was detected in the cross-sectional study subgroup (Z=1.35, P=0.176) and the case-control study subgroup (Z=1.69, P=0.091). Conclusion: Patients with diabetes are more likely to be infected with HBV than those without diabetes. Diabetes mellitus is likely to be a risk factor for HBV infection.
目的： 通过Meta分析了解糖尿病与HBV感染的相关性。 方法： 检索数据库包括中国期刊全文数据库、维普中文科技期刊全文数据库、万方数据知识服务平台、PubMed、Cochrane和Web of Science Core Collection数据库。时间限定为1997年1月至2017年5月，语种限为中文和英文。英文检索词包括：diabetes、Diabetes Mellitus、hepatitis B和risk；中文检索词包括：糖尿病、乙肝、乙型肝炎、风险和危险。纳入所有研究糖尿病和HBV感染的观察性研究文献。采用纽卡斯尔-渥太华量表和美国卫生保健质量和研究机构推荐的横断面研究标准对文献的质量进行评价；使用RevMan 5.3软件进行异质性检验，使用亚组分析随机效应模型Mantel-Haenszel法计算合并后的OR值，使用Stata 14.0软件进行敏感性分析和发表偏倚的检测，使用Begg秩相关法对发表偏倚进行检测。 结果： 12篇纳入的(文献中，包括6篇横断面研究和6篇病例-对照研究，12篇文献结果具有异质性（)χ(2)=42.10，P<0.001）；(亚组分析结果显示，横断面研究亚组文献结果仍具有异质性（)χ(2)=28.21，P<0.001），合并OR（95%CI）值(为1.36（1.03~1.80），病例-对照研究亚组文献结果异质性无统计学意义（)χ(2)=10.32，P=0.070），合并OR（95%CI）值为1.55（1.10~2.17）。敏感性分析显示，逐一剔除5篇文献质量评分低和研究例数较少的文献后，剩余7篇文献的合并OR值波动较小，且95%CI值的下限均≥1.07；横断面研究亚组（Z=1.35，P=0.176）和病例对照研究亚组（Z=1.69，P=0.091）均未检测到发表偏倚。 结论： 糖尿病患者感染HBV的可能性比非糖尿病患者高，糖尿病可能与HBV感染有相关性。.
Keywords: Diabetes mellitus; Hepatitis B virus; Meta-analysis; Observational study.