Association between 23 drugs and inflammatory bowel disease: a two-sample Mendelian randomization study

Abstract

Background: Inflammatory bowel disease (IBD) is a group of diseases characterized by chronic and recurrent inflammation of the gastrointestinal tract. The etiology of IBD remains multifaceted and poorly understood, resulting in limited treatment options that primarily target disease induction and remission maintenance. Thus, the exploration of novel therapeutic options for IBD among existing medications is advantageous. Mendelian randomization analysis (MR) serves as a valuable tool in investigating the relationship between drugs and diseases. In this study, MR analysis was employed to investigate the potential causal relationship between 23 approved drugs for the treatment of various diseases and IBD.

Method: We performed a two-sample MR analysis using publicly available genome-wide association study (GWAS) statistics. The inverse variance weighting (IVW) method was used as the main analysis method, supplemented by the remaining four methods (weighted median, MR Egger regression, simple and weighted models), and Meta-analysis was performed to expand the sample size to obtain a more reliable composite causal effect. Finally, Cochran's Q statistic and the MR-Egger test for directed pleiotropy were applied to determine whether significant heterogeneity or directed pleiotropy existed.

Results: In the main MR analysis (IVW), drugs with a negative causal association with the risk of IBD were immunosuppressant {OR (95% CI) = 0.7389 [0.6311-0.8651], p = 0.0046} and diabetes drugs {OR (95% CI) = 0.9266 [0.8876-0.9674], p = 0.0058}. A positive causal association with the risk of IBD was found for salicylic acid and derivatives {OR (95% CI) = 1.2737 [1.0778-1.5053], p = 0.0345}. Negative causal associations with UC risk were identified for immunosuppressants {OR (95% CI) = 0.6660 [0.5133-0.8640], p = 0.0169} and diabetes medications {OR (95% CI) = 0.9020 [0.8508-0.9551], p = 0.0046}; positive causal associations with UC risk were found for β-receptor blockers {OR (95% CI) = 1.1893 [1.0823-1.3070], p = 0.0046}. A negative causal association with the risk of CD was found for immunosuppressants {OR (95% CI) = 0.6957 [0.5803-0.8341], p = 0.0023}. There was no statistically significant association between the remaining 19 drugs and IBD and subtypes.

Conclusion: This MR study provides evidence suggesting that immunosuppressants have a mitigating effect on the risk of IBD and demonstrate consistent efficacy in subtypes of ulcerative colitis (UC) and Crohn's disease (CD). Additionally, diabetes medications show potential in reducing the risk of IBD, particularly in cases of UC, while β-blockers may elevate the risk of UC. Conversely, salicylic acid and its derivatives may increase the risk of IBD, although this effect is not consistently observed in the subtypes of the disease. These findings offer new insights into the prevention and management of IBD.

Keywords: 23 drugs; Crohn’s disease; Mendelian randomization; inflammatory bowel disease; ulcerative colitis.

Figure 1

Overview of the research design. The figure was generated using Adobe illustrator.

Figure 2

Forest map of 23 drugs associated with IBD risk. FDR is the p-value corrected by multiple tests.

Figure 3

Forest map of 23 drugs associated with UC risk. FDR is the p-value corrected by multiple tests.

Figure 4

Forest map of 23 drugs associated with CD risk. FDR is the p-value corrected by multiple tests.