基于孟德尔随机化探讨胃食管反流与间质性肺病的双向因果关系

Mendelian randomization-based exploration of bidirectional causal relationship between gastroesophageal reflux disease and interstitial lung disease

摘要:

目的 临床观察及流行病学调查发现胃食管反流病（gastroesophageal reflux disease，GERD）与间质性肺疾病（interstitial lung disease，ILD）具有相关性，本文章旨在通过孟德尔随机化分析（Mendelian randomization，MR）方法，研究GERD与ILD是否相互存在因果关联，以及这些关联的方向性。方法 从全基因组关联研究（genome-wide association studies，GWAS）汇总数据中选择符合条件的单核苷酸多态性（single nucleotide polymorphisms，SNPs）作为工具变量（instrumental variables，IVs）进行分析，分别采用逆方差加权法（inverse-variance weighted，IVW）、MR-Egger回归分析、加权中位数法（weighted median estimator，WME）等进行两样本孟德尔随机化分析，其中IVW用于主要分析；在进行敏感性分析时，采用MR-PRESSO检验和MR-Egger回归分析以检测和纠正水平多效性；采用留一法、Cochran’s Q检验、漏斗图以评估MR结果的稳定性和可靠性，以OR值评价GERD与ILD的因果关系。结果 IVW结果显示GERD可增加患ILD的风险，GERD对数变换后每增加一个标准差，会导致ILD的风险增加22%（OR=1.22，95%CI=1.11~1.33，P=3.52e-05）。反之，ILD也可增加患GERD的风险。结论 GERD与ILD具有双向因果关系。

Abstract:

Objective Clinical observations and epidemiological investigations have found an association of gastroesophageal reflux disease（GERD） with interstitial lung disease（ILD）. We aimed to investigate whether GERD and ILD are causally associated with each other and the direction of the association through Mendelian randomization（MR） analysis.Methods Eligible single nucleotide poly-morphisms（SNPs）were selected as instrumental variables（IVs） from the pooled data of genome-wide association studies（GWAS）.Two-sample Mendelian randomization analyses were performed using the inverse-variance weighted method（IVW），the MR-Egger regression method，and the weighted median estimator（WME）. The IVW method was used as the main causal analysis. The MR-PRESSO test and MR-Egger regression method were used in sensitivity analysis to detect and correct for horizontal pleiotropy. The leave-one-out method，Cochran’s Q test，and funnel plot were used to assess the stability and reliability of MR results. The odds ratio（OR） was used to evaluate the causality between GERD and ILD.Results The IVW results showed that GERD increased the risk of ILD，and each standard deviation increase in log-transformed GERD resulted in a 22% increase in the risk of ILD（OR=1.22，95%CI=1.11-1.33，P=3.52e-05）. Conversely，ILD also increased the risk of GERD.Conclusion GERD and ILD have a bidirectional causal relationship.