Abstract
BACKGROUND: Observational studies suggest that metformin may reduce the risk of malignant tumors of the digestive system (MTDS), but these findings are often confounded by bias and unmeasured variables. Recent meta-analyses have questioned these associations, emphasizing the need for robust causal inference. METHODS: Mendelian randomization (MR) was used to evaluate the causal relationship between metformin and MTDS. Genetic variants associated with metformin's molecular targets were selected from GTEx, eQTLGen, and UK Biobank and validated using genetic colocalization to ensure instrument validity. GWAS summary statistics for MTDS, encompassing up to 314,193 controls and 6,847 colorectal cancer cases, were obtained from FinnGen and EBI. The primary analysis employed the inverse-variance weighted (IVW) method, supplemented by MR-Egger, weighted median, and weighted mode analyses. Bonferroni correction was applied to adjust for multiple testing across 14 cancer types. RESULTS: Genetically proxied metformin use was associated with an increased risk of colorectal cancer (OR = 2.38, 95%CI = 1.38-4.09, P = 0.0018) and related subtypes. No causal relationship was found for hepatocellular carcinoma, gastric cancer, pancreatic cancer, or other digestive system cancers. The robustness of these findings was supported by sensitivity analyses, which indicated no significant pleiotropy, and leave-one-out tests. CONCLUSION: This study provides robust genetic evidence that metformin use increases the risk of colorectal cancer, challenging its role as a preventive agent for digestive cancers. These findings emphasize the need for clinicians to carefully evaluate the risks and benefits of metformin, particularly in populations at higher risk for colorectal cancer. Future research should focus on delineating the mechanisms underlying this association to optimize the use of metformin in clinical practice.