Impact of metformin on the efficacy of immune checkpoint inhibitors in metastatic melanoma: A large-scale retrospective analysis.

Abstract

e21548 Background: Metformin, widely used for type 2 diabetes mellitus (T2DM), has been hypothesized to influence the tumor microenvironment and potentially enhance the efficacy of ICIs. However, there is a lack of large database-driven studies exploring this combination in metastatic melanoma. Methods: A retrospective cohort study was conducted using data from 103 healthcare organizations via the TriNetX platform. Two cohorts were analyzed: patients with metastatic melanoma receiving ICIs (ipilimumab, nivolumab, and/or pembrolizumab) without metformin (Cohort 1, n = 4,021) and those receiving ICIs with metformin (Cohort 2, n = 2,026). Propensity score matching was employed to balance baseline characteristics. After matching, both cohorts consisted of 2,026 patients each. The mean age of patients in both cohorts was 72.3 years, and the gender distribution was 28% female and 68% male. Laboratory analysis showed a mean hemoglobin A1C level of 7.1% in the metformin cohort versus 6.9% in the non-metformin cohort. Outcomes assessed included mortality risk, Kaplan-Meier survival analysis, and hazard ratios. Results: Mortality risk was slightly lower in the metformin group (risk ratio 1.065; p = 0.146), though this difference was not statistically significant. Kaplan-Meier analysis demonstrated a median survival advantage of 208 days in the metformin group compared to the non-metformin group, but this was not statistically significant (log-rank test, p = 0.189). The hazard ratio for overall survival was 1.073 (p = 0.420). Demographic characteristics, including age, gender, and A1C levels, were balanced between groups after matching. MM + Metformin cohort exhibited slightly higher A1C levels, longer median survival, and lower risk of mortality. Conclusions: This large-scale, database-driven study is one of the first to explore the combination of ICIs and metformin in metastatic melanoma, demonstrating a modest but non-significant survival benefit. Prospective trials are needed to confirm the synergy and identify responsive subgroups. Despite slightly higher A1C levels in the MM + Metformin cohort, the longer median survival suggests a potential link between A1C and outcomes, warranting further investigation. Baseline characteristics before and after propensity score matching. Characteristic Before Matching Cohort 1 (N=4,021) Before Matching Cohort 2 (N=2,026) P-Value After Matching Cohort 1 (N=2,026) After Matching Cohort 2 (N=2,026) P-Value Age (Mean ± SD) 72.7 ± 11.5 72.3 ± 11.0 0.149 72.3 ± 11.0 72.3 ± 11.0 0.887 Female, n (%) 1,216 (30.2%) 569 (28.1%) 0.083 567 (28.0%) 569 (28.1%) 0.944 Male, n (%) 2,652 (66.0%) 1,394 (68.8%) 0.026 1,395 (68.9%) 1,394 (68.8%) 0.973 Body Mass Index ≥30, n (%) 1,455 (36.2%) 754 (37.2%) 0.432 757 (37.4%) 754 (37.2%) 0.922 Hemoglobin A1c (Mean ± SD) 6.9 ± 1.6 7.2 ± 1.6 <0.001 7.1 ± 1.6 7.2 ± 1.6 0.012