Assessing causality between osteoarthritis and gastrointestinal disorders: a Mendelian randomization study

Abstract

Abstract The association between osteoarthritis (OA) and gastrointestinal disorders was found in observational studies. However, the causality is still elusive. A bidirectional Mendelian randomization (MR) analysis using genome wide association studies data was conducted to assess the causal association between OA and gastrointestinal diseases [including peptic ulcer disease (PUD), gastroesophageal reflux disease (GORD), and inflammatory bowel disease (IBD)]. A two-step MR (TSMR) was conducted between OA, gastrointestinal diseases and drugs to explore the mediating effects of non-steroidal anti-inflammatory drugs (NSAIDs) and opioids use. We used multivariable MR (MVMR) analysis to further validate the impact of prescription history on diseases. Results had statistical significance at a Bonferroni corrected P-value below 0.008. We observed that genetically predicted OA had a significant positive association with GORD [odds ratio (OR) = 1.26, P = 5e−05], but not with PUD or IBD. Regarding the other direction, gastrointestinal disorders as exposure had a null association with OA. Using TSMR, OA patients tended to increase the use of NSAIDs (OR = 1.45, P = 0.001) and opioids (OR = 1.77, P = 2e−05), but only the use of opioids increased the risk of GORD (OR = 1.43, P = 5e−09). Further MVMR analysis showed that the adverse effect of OA on GORD was significantly reduced after adjusting for opioids use (OR = 1.20, P = 0.038). This study provides evidence for the causal association between OA and increased risk of GORD, which is partly attributed to opioids use in OA patients but not NSAIDs.