Abstract
INTRODUCTION: An increasing number of studies have focused on the anti-tumor effect of metformin in recent years. However, the effect of metformin on different cancers remains controversial and lacks consensus. METHODS: A bidirectional two-sample Mendelian randomization (MR) design was used to assess causal relationships between metformin and 18 cancer types. Sensitivity analyses were conducted for reliability assessment. Multivariate Mendelian randomization (MVMR) analyses considered three relevant factors simultaneously. MR analysis based on gene proxies was conducted as well for further validation. Logistic regression models were constructed using National Health and Nutrition Examination Survey (NHANES) data (2013-2018) to evaluate the association between metformin and cancer. RESULTS: Two-sample MR analysis identified metformin as a protective factor for pan cancer, prostate, ovarian, breast, esophageal, colorectal, and lung cancer. However, metformin was found to be a risk factor for bladder cancer. MVMR analysis confirmed metformin's significant protective effect on prostate and ovarian cancer. Logistic models based on NHANES data demonstrated metformin's significant protective effect against cancer in diabetes patients. CONCLUSION: Our findings indicate a significant protective effect of metformin against prostate and ovarian cancer based on MR analysis. NHANES data further support a general protective effect of metformin against cancer. These findings warrant consideration of metformin in the context of cancer prevention and potential therapeutic strategies for prostate and ovarian cancer, though further research is needed to fully elucidate its mechanisms of action and establish its role in anti-cancer therapy.