Metformin and Risk Recurrence in Resected Stage II/III Colon Cancer (CC) Patients (pts): Subgroup Analysis from the TOSCA Trial

Background: Based on preclinical studies, metformin could display both indirect anticancer activity by reducing blood glucose, insulin and IGF-1 levels, and cell-autonomous effects by activating AMP-kinase (AMPK) and impairing mitochondrial metabolism. In the clinical setting, favourable cancer-related outcomes have been reported in pts with type II diabetes mellitus (T2DM) treated with metformin. However, recently published data failed to confirm a positive effect on disease free survival and overall survival in stage III resected CC pts. We performed an observational study to evaluate the impact of metformin use in pts enrolled in the TOSCA trial. Patients and methods: TOSCA was a non-profit, Italian, multicentre, randomized, non-inferiority phase III study conducted in high-risk stage II and stage III CC pts treated with 6 or 3 months of FOLFOX-4 or XELOX adjuvant chemotherapy. Pts with T2DM were prospectively accrued in this observational study. The primary endpoint was the modified relapse free survival (mRFS), defined as the time from randomization to relapse or cancer-related death. Cancer-related survival (CRS), as defined as the time from randomization to cancer-related death, was a secondary endpoint. Fine and Gray proportional subdistribution hazard models for competiting risk were used. Results: Of 3759 pts randomized in the TOSCA trial, 1520 were screened and 142 with T2DM were included in this study. 2 pts were excluded because of major violation and 5 pts due to the lack of available data on metformin exposure. Among 135 evaluable pts, 55 (41%) received only metformin and 80 (59%) other antidiabetic medications plus/minus metformin. Regarding drug exposure, 40 out of 55 (73%) pts received metformin before, during and after chemotherapy. No difference in pts’ characteristics and pathological stage between metformin users and nonusers was found. Disease relapse and cancer-related deaths were observed in 30 (22%) and 11 (8%), respectively. Relapse or cancer-related deaths occurred in 31 (23%) pts. No effect of metformin exposure was detected both on mRFS (adjusted HR 1.20; 95%CI 0.53-2.75; p = 0.65) and CRS (adjusted HR 1.71; 95%CI 0.40-7.32; p = 0.47). Conclusions: Our findings did not demonstrate an association between metformin use and colon cancer relapse or cancer-related death in II/III stage pts treated with adjuvant chemotherapy. However, given the low number of observed events, other analyses will be needed for definitive conclusions.