Association between metformin use and clinical outcomes following pancreaticoduodenectomy in patients with type 2 diabetes and pancreatic ductal adenocarcinoma

Abstract

Background: Retrospective studies on the association between metformin and clinical outcomes may be affected by time-related bias. Recent studies used time-varying analysis to avoid time-related bias, but only considered the start date of metformin and not the stop date. We first aimed to confirm the effect of metformin on pancreatic cancer cells in cellular level and tried to determine the clinical benefits of metformin in patients with type 2 diabetes and pancreatic ductal adenocarcinoma following pancreaticoduodenectomy. Methods: In laboratory study, MTT assay was performed to confirm the decrease in viability by Metformin. Also western blotting and quantitative real-time polymerase chain reaction were done to check the effect on mTOR. Data from patients who underwent pancreaticoduodenectomy due to pancreatic ductal adenocarcinoma at a single high-volume center between May 2007 and July 2016 were retrospectively reviewed. Analysis using a Cox model with time-varying covariates was performed to compare recurrence-free survival and overall survival between the metformin group and the non-metformin group while considering both the start and stop dates of metformin use. Also, meta-analyses with previous studies were performed to compare the results. Results: MTT assay in AsPC-1 cells revealed a concentration-dependent decrease in cell survival. And the protein level and RNA expression of mTOR were significantly decreased upon the treatment with 30 or 50 mmol/ml of metformin. In the meta-analysis, the metformin group had significantly better overall survival in studies which have done the conventional Cox analysis (HR = 0.85, 95% CI: 0.78 – 0.92). On the other hand, metformin did not show any beneficial effect on clinical outcomes in the meta-analysis of studies performed analysis by Cox models with time-varying covariates only reflecting the start date of metformin use (HR = 0.99, 95% CI: 0.93 – 1.05). However, in our retrospective study, a total of 283 patients were included and used Cox models with time-varying covariates reflecting both the start and stop dates of metformin use as a first time. The result showed a significant difference in overall survival between the two groups (HR, 0.747; 95% CI, 0.562–0.993; P = 0.045). Conclusion: Metformin may be useful as a postoperative therapy for improving overall survival following pancreaticoduodenectomy in patients with type 2 diabetes and pancreatic ductal adenocarcinoma.