DataSheet_3_Type 1 diabetes mellitus and non-alcoholic fatty liver disease: a two-sample Mendelian randomization study.pdf

BackgroundNAFLD (Nonalcoholic fatty liver disease) is becoming an increasingly common cause of chronic liver disease. Metabolic dysfunction, overweight/obesity, and diabetes are thought to be closely associated with increased NAFLD risk. However, few studies have focused on the mechanisms of NAFLD occurrence in T1DM.MethodsWe conducted a two-sample Mendelian randomization (MR) analysis to assess the causal association between T1DM and NAFLD with/without complications, such as coma, renal complications, ketoacidosis, neurological complications, and ophthalmic complications. Multiple Mendelian randomization methods, such as the inverse variance weighted (IVW) method, weighted median method, and MR-Egger test were performed to evaluate the causal association of T1DM and NAFLD using genome-wide association study summary data from different consortia, such as Finngen and UK biobank.ResultsWe selected 37 SNPs strongly associated with NAFLD/LFC (at a significance level of p < 5 × 10−8) as instrumental variables from the Finnish database based on the T1DM phenotype (8,967 cases and 308,373 controls). We also selected 14/16 SNPs based on with or without complications. The results suggest that the genetic susceptibility of T1DM does not increase the risk of NAFLD (OR=1.005 [0.99, 1.02], IVW p=0.516, MR Egger p=0.344, Weighted median p=0.959, Weighted mode p=0.791), regardless of whether complications are present. A slight causal effect of T1DM without complications on LFC was observed (OR=1.025 [1.00, 1.03], MR Egger p=0.045). However, none of the causal relationships were significant in the IVW (p=0.317), Weighted median (p=0.076), and Weighted mode (p=0.163) methods.ConclusionOur study did not find conclusive evidence for a causal association between T1DM and NAFLD, although clinical observations indicate increasing abnormal transaminase prevalence and NAFLD progression in T1DM patients.

背景：非酒精性脂肪性肝病（Nonalcoholic fatty liver disease，简称NAFLD）正日益成为慢性肝病的常见病因。代谢功能障碍、超重/肥胖以及糖尿病被认为是与NAFLD风险增加密切相关。然而，关于1型糖尿病（T1DM）中NAFLD发生机制的研究为数不多。方法：本研究采用双样本孟德尔随机化（Mendelian randomization，简称MR）分析，以评估1型糖尿病与NAFLD，无论有无并发症（如昏迷、肾脏并发症、酮症酸中毒、神经并发症和眼科并发症）之间的因果关联。通过执行多种孟德尔随机化方法，如逆方差加权（inverse variance weighted，简称IVW）法、加权中位数法和MR-Egger检验，利用来自不同联盟（如Finngen和UK生物库）的全基因组关联研究（genome-wide association study，简称GWAS）汇总数据，对1型糖尿病与NAFLD的因果关联进行评估。结果：基于1型糖尿病表型，我们从芬兰数据库中选取了37个与NAFLD/LFC（在显著性水平p < 5 × 10−8）强相关单核苷酸多态性（single nucleotide polymorphisms，简称SNPs）作为工具变量（8,967个病例和308,373个对照）。我们还基于有无并发症选择了14/16个SNPs。结果表明，1型糖尿病的遗传易感性并不会增加NAFLD的风险（OR=1.005 [0.99, 1.02]，IVW p=0.516，MR Egger p=0.344，加权中位数p=0.959，加权众数p=0.791），无论是否存在并发症。在无并发症的1型糖尿病对LFC的轻微因果效应被观察到（OR=1.025 [1.00, 1.03]，MR Egger p=0.045）。然而，在IVW（p=0.317）、加权中位数（p=0.076）和加权众数（p=0.163）方法中，这些因果关联均不显著。结论：尽管临床观察表明1型糖尿病患者中异常转氨酶的普遍率和NAFLD的进展有所增加，但本研究未发现1型糖尿病与NAFLD之间存在因果关联的结论性证据。