Causal relationship between body mass index and risk of pancreatitis: a Mendelian randomisation study

Abstract

Objective The aim of the study was to investigate whether body mass index (BMI) is causally related to pancreatitis, and to guide patients with potential pancreatitis to control their weight and prevent pancreatitis. Methods Weighted median, MR-Egger regression, and inverse variance weighting (IVW) were accessed using a two-sample Mendelian randomization (MR) analysis. In the Genome-Wide Association Study (GWAS) meta-analysis, we used the publicly available summary statistics set of BMI in individuals of European ancestry (n = 681,275) (GIANT Alliance) as the exposure, and as the outcome, we used the self-reported non-cancer disease codes: pancreatitis in individuals from the European Biobank (n = 462, 933). Results Selecting 220 single nucleotide polymorphisms (SNPs) for BMI was done using the GWAS instrumental variable (IV). The IVW method's results show that pancreatitis and BMI are causally related (beta = 0.0011, SE = 0.0004, p = 0.0137). Although orientational polyvalency was not expected to skew the data (intercept = 1.24E-05; p = 0.511), MR-Egger regression did not reveal a causal relationship between pancreatitis and BMI (beta = 0.0003, SE = 0.0013, p = 0.7941). Furthermore, there was no evidence of a link between pancreatitis and BMI using the weighted median approach (β = 0.0006, SE = 0.0008, p = 0.4296). There was no indication of heterogeneity or asymmetry in the funnel plot or the Cochran's Q test, indicating the absence of directed pleiotropy. Conclusion The hypothesis that a higher incidence of pancreatitis may be partially attributed to BMI is supported by the MR study.