口腔疾病与炎症性肠病风险的双向两样本孟德尔随机化研究  

A bidirectional two-sample Mendelian randomization study on oral disease and risk of inflammatory bowel disease

作　　者：徐涯鑫 高亚南 徐军 姚亭 陈亚梅[3] Xu Yaxin;Gao Yanan;Xu Jun;Yao Ting;Chen Yamei(Tongji University School of Medicine,Shanghai 200092,China;School of Nursing,Anhui University of Traditional Chinese Medicine,Hefei 230038,China;Department of Nursing,Tenth People′s Hospital,Tongji University,Shanghai 200072,China)

机构地区：[1]同济大学医学院,上海200092 [2]安徽中医药大学护理学院,合肥230038 [3]同济大学附属第十人民医院护理部,上海200072

出　　处：《中华消化杂志》2024年第8期532-539,共8页Chinese Journal of Digestion

基　　金：国家自然科学基金(72274135);上海申康医院发展中心临床科技创新项目(SHDC12022621)。

摘　　要：目的采用孟德尔随机化(MR)和meta分析探究遗传预测的口腔疾病与炎症性肠病(IBD)之间的双向因果关系。方法从开放全基因组关联研究(IEU open GWAS)中提取牙周炎、扁平苔藓、口腔溃疡和龋齿4种口腔疾病的遗传数据,并从de Lange等的meta分析、FinnGen数据库、国际炎症性肠病遗传学联盟(IIBDGC)3个独立数据库中提取IBD、溃疡性结肠炎(UC)、克罗恩病相关的遗传数据。通过设置关联强度和消除连锁不平衡,筛选出单个数据库层面"暴露-结局"相关的单核苷酸多态性(SNP),去除混杂因素SNP和回文SNP后作为工具变量。通过计算F值排除弱工具变量偏倚。以逆方差加权法(IVW)作为主要分析方法,采用Cochran′sQ检验分析单个数据库的异质性,根据是否存在异质性选择随机效应或固定效应模型进行MR分析。应用MR-Egger检验及孟德尔随机化多效性残差和异常值(MR-PRESSO)检验去除异常值并评估多个数据库间水平多效性,然后进行meta分析汇总多个数据库的效应值。采用留一法进行敏感性检验明确单个SNP对结局的效应。结果MR分析显示,遗传预测的扁平苔藓对IBD[OR=1.069,95%置信区间(95%CI)1.043~1.097,Z=5.25,P<0.001]、克罗恩病(OR=1.077,95%CI 1.018~1.139,Z=2.56,P=0.010)和UC(OR=1.075,95%CI 1.040~1.111,Z=4.31,P<0.001)有显著影响。反向MR分析发现IBD与牙周炎(OR=1.051,95%CI 1.020~1.083,Z=3.25,P=0.001)、扁平苔藓(OR=1.166,95%CI 1.011~1.344,Z=2.11,P=0.035)和口腔溃疡(OR=1.003,95%CI 1.001~1.004,Z=4.28,P<0.001)风险增加相关;UC与牙周炎之间存在因果关联(OR=1.043,95%CI 1.009~1.077,Z=2.51,P=0.012),克罗恩病与扁平苔藓风险增加显著相关(OR=1.088,95%CI 1.038~1.141,Z=3.48,P<0.001)。结论扁平苔藓可能是IBD进展中的危险因素,IBD也可能与多种口腔病变风险增加有关,口腔疾病与IBD之间存在潜在双向因果关系。Objective To explore the bidirectional causal relationship between genetically predicted oral diseases and inflammatory bowel disease(IBD)by Mendelian randomization(MR)and meta analysis.MethodsGenetic datas of 4 oral diseases,including periodontitis,lichen planus,oral ulcer,and dental caries were extracted from the Integrative Epidemiology Unit(IEU)open genome-wide association study project(IEU open GWAS),and genetic datas related to IBD,ulcerative colitis(UC),and Crohn′s disease(CD)were extracted from 3 independent databases,including the meta-analysis of de Lange et al,the FinnGen database,and the International Inflammatory Bowel Disease Genetics Consortium(IIBDGC).Single nucleotide polymorphism(SNP)identified as instrumental variable(IV)for"exposure-outcome"relationship was screened by setting the association strength and eliminating linkage disequilibrium,and confounder SNP and palindromic SNP were removed at each database level.Weak IV bias was excluded according to the F-value.Inverse variance weighted(IVW)method was used as the main analysis method,and Cochran′s Q test was used to analyze the heterogeneity of a single database,and random effect or fixed effect model was selected for MR analysis according to the presence or absence of heterogeneity.The MR-Egger test and Mendelian randomization-pleiotropy residual sum and outlier(MR-PRESSO)test were used to remove outliers and evaluate horizontal pleipotency across multiple databases,and then meta-analysis was performed to summarize the effect value of multiple databases.The effect of individuol SNP on the outcome was determined by a sensitivity test using the leave-one-out method.ResultsThe results of the MR analysis indicated that genetically predicted lichen planus had a significant effect on IBD(OR=1.069,95%confidence interval(95%CI)1.043 to 1.097,Z=5.25,P<0.001),CD(OR=1.077,95%CI 1.018 to 1.139,Z=2.56,P=0.010)and UC(OR=1.075,95%CI 1.040 to 1.111,Z=4.31,P<0.001).The reverse MR analysis revealed that IBD was associated with an increased risk of period

关 键 词：孟德尔随机化分析 口腔疾病 炎性肠疾病 因果律 

分 类 号：R574[医药卫生—消化系统] R78[医药卫生—内科学]