孟德尔随机化研究精神分裂症与自杀或故意自残的因果关系

A two-sample Mendelian randomization study of the causal relationship between schizophrenia and suicide or intentional self-harm

目的采用孟德尔随机化(Mendelian randomization, MR)方法研究精神分裂症与自杀或故意自残的双向因果关系。方法利用大规模全基因组关联研究(genome-wide association studies, GWAS)汇总数据, 选择相互独立且与欧洲血统人群精神分裂症及自杀或故意自残相关的遗传位点作为工具变量。精神分裂症GWAS数据包括76 755例精神分裂症患者和243 649名对照者。自杀或故意自残的GWAS数据包括1 058例自杀或故意自残患者和307 942名对照者。应用逆方差加权法(inverse variance weighted, IVW)及加权中位数法进行MR分析, 以比值比作为评价指标对精神分裂症及自杀或故意自残的双向因果关系进行探讨。CochranQ检验评估各个单核苷酸多态性(single nucleotide polymorphism, SNP)之间的异质性。使用孟德尔随机多态性残差和离群值(Mendelian randomization, pleiotropy residual sum and outlier, MR-PRESSO)检测SNP异常值。MR-Egger截距测试检验SNP的水平多效性。"留一法"敏感性分析检验MR研究是否受单个SNP的影响。结果 IVW显示, 精神分裂症可增加自杀或故意自残的风险(OR=1.297, 95%CI=1.154~1.457, P<0.001), 且在加权中位数法中被证实(OR=1.232, 95%CI=1.038~1.461, P=0.017)。反之, 自杀或故意自残与精神分裂症无关(OR=0.969, 95%CI=0.927~1.012, P=0.157), 同样被加权中位数法证实。CochranQ检验、MR-PRESSO检测和MR-Egger截距测试显示工具变量之间不存在异质性或水平多效性;"留一法"敏感性分析证实, 没有单个SNP对整体的结果有较大的影响。结论精神分裂症与自杀或故意自残的风险增加有显著的因果关系。

Objective To investigate the bidirectional causal relationship between schizophrenia and suicide or intentional self-harm using a Mendelian randomization(MR)approach.Methods Using pooled data from a large-scale genome-wide association study,genetic loci that were independent of each other and associated with schizophrenia and suicide or intentional self-harm in populations of European ancestry were selected as instrumental variables.The data for schizophrenia included 76755 patients with schizophrenia and 243649 controls.Data for suicide or intentional self-harm included 1058 patients with suicide or intentional self-harm and 307942 controls.The inverse variance weighting(IVW)and weighted median methods of MR analysis were used to explore the bidirectional causality of schizophrenia and suicide or intentional self-harm using the ratio of ratios(OR)as an evaluation index.Heterogeneity between individual single nucleotide polymorphism(SNP)was assessed by Cochran′s Q test.Mendelian randomization polymorphism residuals and outliers(MR-PRESSO)were used to detect SNP outliers,and MR-Egger regression tests were performed to examine the horizontal pleiotropy of SNPs.A"leave-one-out"sensitivity analysis was performed to test whether the MR study was influenced by individual SNPs.Results The IVW approach showed that schizophrenia increased the risk of suicide or intentional self-harm(OR=1.297,95%CI=1.154-1.457,P<0.001).In addition,the above relationship was confirmed in the weighted median(OR=1.232,95%CI=1.038-1.461,P=0.017).Conversely,suicide or intentional self-harm was not causally associated with schizophrenia(OR=0.969,95%CI=0.927—1.012,P=0.157),again confirmed by the weighted median method.The Cochran′s Q test,MR-PRESSO and MR-Egger regression tests did not show heterogeneity or horizontal pleiotropy between instrumental variables.The"leave-one-out"sensitivity analysis confirmed that no single SNP had a disproportionate effect on the overall outcome.Conclusion Schizophrenia is significantly and causally associated with an increased risk of suicide or intentional self-harm.