Abstract
Objective
Many studies have reported that metformin can reduce the risk of tumor. However, the level of evidence is low, and the results of the studies are conflicting.
Methods
We conducted a tumor associated Phenome-wide Mendelian randomization (MR-PheWAS) analysis to explore the causal relationship between metformin and tumors. Two cohorts of metformin came from UK Biobank. The full phenotype data of tumors came from FinnGen_R10. We elucidated the causal association through two-sample MR analysis (TSMR). More importantly, we conducted a meta-analysis to ensure the unbiased results. In MR analysis, we used Inverse-Variance Weighted (IVW) method as the main outcome indicator. Subsequently, we integrated 2 cohorts for the meta-analysis. Finally, we attempted to explore this mechanisms through a mediational MR analysis.
Results
MR results showed that metformin may have a causal relationship with a total of 13 tumor associated phenotypes in training cohort. 4 phenotypes were validated in the testing cohort. In training and testing cohort, metformin have a protective effect on Malignant neoplasm of breast, HER-positive, Brain meningioma, Malignant neoplasm of oral cavity and Malignant cancer of tonsil and base of tongue. Intriguingly, after integrating the results of two cohorts for meta-analysis, a total of 12 results were significant. A mediational MR analysis showed that the effects of metformin against Brain meningioma may be weaken by family Oxalobacteraceae.
Conclusions
Metformin may have potential preventive and therapeutic effects on a variety of tumors, we don’t recommend routine use of metformin alone because there was no clear cause-and-effect relationship between them.