The causal relationship between gastro-oesophageal reflux disease and idiopathic pulmonary fibrosis: a bidirectional two-sample Mendelian randomisation study

Abstract

Background: Gastro-oesophageal reflux disease (GORD) is associated with idiopathic pulmonary fibrosis (IPF) in observational studies. It is not known if this association arises because GORD causes IPF or because IPF causes GORD, or because of confounding by factors, such as smoking, associated with both GORD and IPF. We used bidirectional Mendelian randomisation (MR), where genetic variants are used as instrumental variables to address issues of confounding and reverse causation, to examine how, if at all, GORD and IPF are causally related.

Methods: A bidirectional two-sample MR was performed to estimate the causal effect of GORD on IPF risk and of IPF on GORD risk, using genetic data from the largest GORD (78 707 cases and 288 734 controls) and IPF (4125 cases and 20 464 controls) genome-wide association meta-analyses currently available.

Results: GORD increased the risk of IPF, with an OR of 1.6 (95% CI 1.04-2.49; p=0.032). There was no evidence of a causal effect of IPF on the risk of GORD, with an OR of 0.999 (95% CI 0.997-1.000; p=0.245).

Conclusions: We found that GORD increases the risk of IPF, but found no evidence that IPF increases the risk of GORD. GORD should be considered in future studies of IPF risk and interest in it as a potential therapeutic target should be renewed. The mechanisms underlying the effect of GORD on IPF should also be investigated.

FIGURE 1

Overview of our two-sample Mendelian randomisation analysis of a) gastro-oesophageal reflux disease (GORD) on idiopathic pulmonary fibrosis (IPF) risk and b) IPF on GORD risk. SNP: single nucleotide polymorphism; G–X: gene–exposure association; G–Y: gene–outcome association.

FIGURE 2

Identification of pleiotropic instruments in the Mendelian randomisation on the effect of idiopathic pulmonary fibrosis on gastro-oesophageal reflux disease. Individual variant contributions to Cochran's Q heterogeneity statistic with the 5th (dotted line) and Bonferroni corrected (0.05/19th) percentiles (dashed line) of a Chi-squared with 1 degree of freedom.

FIGURE 3

Results of the Mendelian randomisation (MR) on the effect of idiopathic pulmonary fibrosis on gastro-oesophageal reflux disease risk using different robust methods to address the issue of pleiotropy: main analysis and sensitivity analysis with five possible pleiotropic single nucleotide polymorphisms removed. IVW-FE: inverse variance weighted fixed effect; IVW-RE: inverse variance weighted random effect; WMe: weighted median; Egger: MR-Egger; WMo: weighted mode-based.