【作者】 王文华；

【导师】 张卫东；

【作者基本信息】 郑州大学 ， 流行病与卫生统计学， 2018， 硕士

【摘要】 研究背景2型糖尿病（type 2 diabetes,T2D）是一种以高血糖症为特征的疾病,严重威胁人们的身体健康。除了遗传因素之外,过量饮食、肥胖和高脂血症等都是T2D的危险因素。近年来的研究还发现,睡眠障碍在T2D患者中非常普遍,且与生活质量呈负相关。目的探究睡眠障碍与T2D发病风险的关系,同时用孟德尔随机化研究（Mendelian randomization,MR）的方法检验睡眠类型和睡眠时间对T2D发生的因果关联,为T2D的预防提供依据,也为健康的生活方式提供指导意义。方法应用计算机检索中国知网数据库、中国万方数据库、英文Pubmed数据库、英文Web of Science数据库,收集从建库至2018年3月公开发表的关于睡眠障碍和T2D发生关系的前瞻性研究,严格按照纳入和排除标准进行文献的筛选和质量评价,采用Stata软件的metan命令分析入睡困难和夜间维持睡眠困难与T2D之间的关系,用Q检验及I²值对纳入的文献进行异质性检验;采用Egger和Begg检验检测发表偏倚;采用敏感性分析检验结果的稳定性。从已经公开发表的英国生物数据库（UK biobank）的数据中选取与睡眠类型和睡眠时间相关的SNP位点,挑选在其他人群验证仍满足P<5×10^-8的SNP位点作为MR分析中暴露的工具变量（instrumental variable,IV）,利用R语言调用软件包计算工具变量与结局的关系,并通过逆方差加权孟德尔随机化法（Mendelian randomization Method of Inverse variance weighted,MR-IVW Method）、基于中位数孟德尔随机化法（Mendelian randomization Method of median based,MR-Median-Based Method）、Egger回归孟德尔随机化法（Mendelian randomization of Egger regression,MR-Egger Method）三种方法计算睡眠类型和睡眠时间与T2D的因果关联。结果本研究最终共纳入8篇符合标准的文献,文献质量评价均为中、高级文献。Meta分析结果显示,与睡眠正常组相比,入睡困难组和夜间维持睡眠困难组T2D发病风险分别增加44%和51%,RR和95%CI分别为1.44（1.18¹.75）和1.51（1.22¹.88）;入睡困难研究中Egger和Begg检验结果显示不存在发表偏倚,夜间维持睡眠困难研究结果显示存在发表偏倚。在UK biobank中分别有16个和3个与睡眠类型和睡眠时间相关的SNP位点,经人群验证后分别有13和3个SNP位点仍满足P<5×10^-8作为纳入工具变量,工具变量与结局均没有联系（P>0.05）。三种方法的孟德尔随机化分析均显示“早晨型”睡眠类型和睡眠时间对T2D发生不存在因果关联,异质性检验均没有意义,显示结果稳定可靠。结论“早晨型”睡眠类型和睡眠时间与T2D的联系不是因果关联,通过改变睡眠类型和睡眠时间达不到降低T2D发生风险的目的。但是本研究的发现值得更深入的研究,以确定可能的影响睡眠障碍也增加T2D风险的原因。更多还原

【Abstract】 BackgroundType 2 diabetes（T2D）is a disease characterized by hyperglycemia,which seriously threatens the health of people and affects the quality of life.In addition to genetic factors,lack of exercise,obesity and hyperlipidemia are all risk factors for T2D.Recent studies have also found that sleep disturbance are very common in T2D patients and negatively correlated with quality of life.ObjectiveTo explore the relationship between sleep disturbance and the risk of T2D,and to examine the causal relationship between chronotype and sleep duration on T2D by Mendelian randomization（MR）study,evidence is provided to prevent T2D and provide guidance for healthy lifestyles.MethodsDatabases included Pubmed,Web of Science,CNKI,Wangfang were searched for relevant research.The relationship between sleep diaorder and risk of type 2diabetes in March,2018.literature screening and the quality of literature was evaluated,the relationship between difficulties in falling asleep at night and difficulties in maintaining asleep at night and incidence risk of T2D was analyzed by Stata software’s metan command.The heterogeneity test was performed by Q test and I² value;Egger and Begg test were used if there is a publication bias;The sensitivity of the result is analyzed by sensitivity analysis.SNP related to chronotype and sleep duration were selected from UK biobank that had been published publicily,and SNP that were verified to meet P<5×10^-8 in other populations were selected for MR analysis.The exposed instrumental variable（IV）uses the R language to select the software package to calculate the relationship between IV and outcomes,and uses the Mendelian randomization method of inverse variance weighted（MR-IVW Method）.Based on the Mendelian randomization method of median based（MR-Median-Based Method）and the Mendelian randomization of Egger regression（MR-Egger Method）,three types of MR were calculated between the causal association of chronotype and sleep time with T2D.ResultIn this study,a total of 8 articles that meet the criteria were included in the study.The quality of the literature was assessed in the medium and high-level literature.Meta-analysis showed that the risk of T2D increased 44%and 51%,respectively,in the group of difficulties in falling asleep and difficulties in maintaining asleep compared with the normal group.The RR and 95%CI were 1.44（1.18 to 1.75）and1.51（1.22 to 1.88）;The Egger and Begg test results in the study of difficulty in falling asleep show that there is no publication bias,and the difficulty in maintaining sleep at night shows that there is publication bias.There were 16 and 3 SNP loci related to chronotype and sleep duration in the UK biobank,and 13 and 3 SNP loci were still remain P<5×10^-8 as IV after population verification.There was no correlation between IV and outcomes（P>0.05）.The Mendelian randomized analysis of the three methods all showed that there was no causal relationship between"morningness"chronotype and sleep time on T2D occurrence,and the heterogeneity test had no significance,indicating that the results were stable and reliable.ConclusionThe relationship between"morningness"chronotype and sleep duration and T2D is not causal.It may not be possible to reduce the risk of developing T2D by changing sleep patterns and sleep duration.However,the findings of this study warrant further investigation to determine the possible causes of sleep disorders that also increase the risk of T2D..更多还原