Mendelian Randomization Study Of Causal Association Between Hypothyroidism And Idiopathic Pulmonary Fibrosis

Background:Idiopathic pulmonary fibrosis(IPF)is a serious chronic rare disease,mainly manifested as exertional dry cough,dyspnea,fatigue and other symptoms.At present,the awareness rate of IPF is low,the prognosis is poor,the median survival time after diagnosis is short,and there is no effective means of treatment.So,it is called "cancer that is not cancer".Risk factors for IPF include intrinsic factors(e.g.,genetics,aging,gender,and lung microbiome),extrinsic factors(e.g.,smoking,environmental exposure,air pollution),and IPF comorbidity.Hypothyroidism(HypoT)is a disease that reduces the body’s metabolism due to the reduction of thyroid hormone synthesis and secretion,or its insufficient physiological effects.It is a non-respiratory comorbidity of IPF.Observational studies have confirmed that there is an association between IPF and HypoT,and HypoT can predict the mortality of IPF.However,because IPF is a rare disease and the median survival time is short,it is difficult to establish a large-scale longitudinal cohort.The association studies of IPF mostly stay in small sample and retrospective studies,which also makes the causal relationship between the two diseases unclear.Mendelian randomization(MR)is a method that uses genetic variants(e.g.,single nucleotide polymorphisms,single nucleotide polymorphisms[SNPs])as instrumental variables(IV)to infer causality between exposure and outcome method of relationship.MR is similar to randomized controlled trials "in the natural state" in that MR largely excludes the effect of unobserved confounders because alleles are randomly assigned,and because randomization is done before birth,it effectively avoids inverse causality problem.At the same time,the measurement of genetic variation has high accuracy and is not prone to measurement errors.At present,the public release of a large number of genome wide association study(GWAS)summary data has promoted the vigorous development of twosample MR,making it an effective mean to explore the causal relationship between exposure and outcome in an observational study.Objective:GWAS was used to summarize the data and explore the causal association between HypoT and IPF under the MR framework,to clarify the causal relationship between the two diseases,the size and direction of the causal effect,and provide clues for the study of the etiology and mechanism of IPF,as well as scientific basis for the prevention,diagnosis and treatment of IPF patients.Methods:IPF pooled data from the largest IPF GWAS study to date,including 11,259 subjects(2,668 cases and 8,591 controls).HypoT data were collected from three different GWAS platforms:(1)the UK biobank(UKBB),which included GWAS who self-reported HypoT or myxedema disease(N=289,307,18,740 cases and 270,567 controls);(2)An external HypoT GWAS dataset independent of UKBB,including 53,423 subjects,including 3,440 cases and 49,983 controls;(3)UKBB,whose cases were GWAS diagnosed by ICD-10,included 244,890 subjects,including 13,043 cases and 231,847 controls.GWAS data for other IPF comorbidities(chronic obstructive pulmonary disease,obstructive sleep apnea,ischemic heart disease,atherosclerosis,type 2 diabetes,gastroesophageal reflux disease,and lung cancer)and smoking were obtained from UKBB.In univariable MR analysis,linkage disequilibrium score regression analysis was first used to test the degree of sample overlap between Hypo TGWAS and IPF GWAS,then IV was screened according to strict conditions;and F statistics were used to exclude weak instrumental variable bias.The effectiveness of IV was further verified by MR-Egger,MR Pleiotropy Residual Sum and Outlier(MR-PRESSO)and leave-one-out analysis.Inversevariance weighted(MR-IVW)is the main analysis method.The weighted median method,MR-IVW method using robust regression(MR-Robust),MR-Lasso,and MR Robust Adjusted Profile Score(MR-RAPs)were used as sensitivity analyses to evaluate the reliability of the study results.In addition,the results in the exploratory analysis were validated using an independent external data set,sensitivity analysis was performed using HypoT data from various UKBB case sources,and finally reverse MR analysis was performed using IPF as exposure and HypoT as outcome.Based on the univariable MR,other comorbidity of IPF and smoking were further considered,and a multivariate MR method based on Bayesian model was adopted.Results:1.In the discovery analysis,a total of 84 IVs were screened out,and the influence of sample overlap and weak instrumental variable bias on the results was excluded.MR-Egger did not detect horizontal pleiotropy,and leave-one-out method and MR-PRESSO did not detect outliers.MR-IVW analysis found a causal association between HypoT and IPF(OR=1.125;95%CI=1.028-1.231;P=0.011),and the result of weighted median MR method,MR-lasso,MR-Robust,MR-RAPS was(OR=1.149;95%CI=1.010-1.308;P=0.034),(OR=1.101;95%CI=1.012-1.198;P=0.025),(OR=1.111;95%CI=0.997-1.237;P=0.057),(OR=1.104;95%CI=1.003-1.215;P=0.042),respectively.2.In the validation analysis,a total of 6 SNPs were obtained as valid IVs through the screening and testing of IVs for analysis,and the result of MR-IVW was(OR=1.229;95%CI=1.054-1.432;P=0.008),which further proved that HypoT is associated with There is a causal relationship between IPF,and the estimated causal effects obtained by the other MR methods are all statistically significant.3.A total of 44 IVs were used in the sensitivity analysis,and the result of MR-IVW was(OR=1.163;95%CI=1.040-1.300;P=0.008),and the causal effect estimates obtained by other MR methods were also statistically significant.4.In the reverse MR analysis,no causal estimates were detected using 14 valid IVs,and the result of IVW was OR=1.010(95%CI=0.994-1.026;P=0.214).5.Among the variables included in the MR-BMA multivariate analysis,HypoT was the causative factor with the strongest evidence for IPF(marginal probability of inclusion=0.388,false positive rate=0.025).After the strong influence points were eliminated by the model diagnosis,the analysis was carried out again.The evidence of a causal relationship between HypoT and IPF was the strongest.Conclusion:HypoT is the cause of IPF,and there is no reverse causal association.Independent data set verification,sensitivity analysis of different methods and multiple MRBMA analysis all reached a consistent conclusion,which provides an important scientific basis for the pathogenesis and treatment of IPF.