Background: Metformin has been shown to confer anti-neoplastic properties in several tumor types. Its effect on outcomes in mRCC patients has not been completely characterized. In this study, we evaluated the role of metformin on survival outcomes in pts with mRCC. Methods: We conducted a retrospective study of pts with mRCC treated on several phase II and III clinical trials from 2003-2013. We analyzed overall survival (OS) in the metformin users versus non-users using the Cox regression model and the Kaplan-Meier method. Results: We identified 4,736 pts with mRCC including 486 diabetic pts of whom 218 (4.6%) were metformin users. The majority were <65 years of age (69%), male (71%), with clear-cell histology (89%) and prior nephrectomy (70%). With regard to IMDC risk groups, 14%, 42%, and 24% had favorable, intermediate, and poor-risk disease, respectively. Pts received treatment with sunitinib (n=1,059), sorafenib (n=772), axitinib (n=896), temsirolimus (TEM) (n=457), TEM + interferon (IFN)-α (n=208), bevacizumab (BEV) + TEM (n=393), BEV + IFN-α (n=391), or IFN-α (n=560); overall 3,044 (64%) received first-line therapy. In the total cohort, metformin use did not impact OS when compared to users of other anti-diabetic agents (p=0.17) or non-diabetics (p=0.69). In diabetic pts, metformin use did not confer a survival advantage when stratified by type of therapy and IMDC risk group. However, in the cohort of diabetic pts receiving sunitinib (n=128), metformin use was associated with an improvement in OS when compared to users of non-metformin anti-diabetic agents (29.3 versus 20.9 months, respectively, p=0.0008, HR 0.051, 95% CI 0.009, 0.292). Conclusions: This is the largest study to date investigating the role of metformin on outcomes in mRCC pts. In this analysis, we demonstrate that concomitant use of metformin may improve survival in diabetic pts with mRCC treated with sunitinib. Based on preclinical data, we hypothesize that the mechanism underlying this survival benefit may be related to synergistic inhibition of the MAPK pathway. However, the study is limited by the small number of diabetic patients. Larger prospective studies are warranted to validate these results.