Metformin Therapy and Survival in Multiple Myeloma Among US Veterans with Diabetes Mellitus

Background: Metformin is an oral biguanide anti-diabetic medication used in patients with type 2 diabetes mellitus (DM). Recently, metformin use in diabetic patients with multiple myeloma (MM) has been associated with improved overall survival (OS) (Wu et al, 2014). Additionally, long-term metformin use has been associated with a decrease in the rate of progression of MGUS to MM (Chang et al. 2015). These anti-tumor effects are hypothesized to be mediated through insulin receptor/insulin-like growth factor receptor-1 signaling and moderation of cellular effects of hyperglycemia and hyperinsulinemia. To confirm the association between metformin use and improved survival in diabetic patients with MM, we evaluated a cohort of United States veterans with MM and diabetes.

Methods: Patients diagnosed with MM in the Veterans Administration database from September 1, 1999 to December 31, 2009 were identified and followed through October 2014. Patients who did not receive MM directed therapy within 6 months of diagnosis were excluded, as were patients who died within 1 year of MM diagnosis. Furthermore, patients who did not have a pre-existing diagnosis of DM at the time of MM diagnosis were excluded. To standardize dose-response effects, a defined daily dose (DDD) calculation of 1 gram of metformin daily was used. Cox modeling was used to assess the association between mortality and metformin use. Additional variables considered included: age, body mass index, race, use of novel therapeutic agents, transplantation, year of diagnosis, creatinine ≥ 1.5, albumin ≤ 3 g/dL and bisphosphonate use. Metformin use was evaluated as a time varying covariate to control for immortal time bias. Patients were defined as metformin users if they were prescribed metformin during a period starting 3 months prior to MM diagnosis extending through completion of follow-up. To assess for the impact of steroid-induced DM (SID), a secondary analysis including these patients was completed.

Results: A total of 3,069 MM patients were evaluable, of whom 549 had a diagnosis of DM prior to MM diagnosis. Among these patients, 268 patients (49%) had received metformin therapy. Diabetic metformin users were younger (mean 66.5 years vs 68 years, P=0.0044), and had fewer comorbidities including chronic kidney disease (9.7% vs 37.4%, P < 0.0001) and ischemic heart disease (39.2% vs 48.8%, P < 0.0001), compared with diabetic non-metformin users . Additionally, metformin users were more likely to have undergone autologous stem cell transplantation (18.3% versus 8.2%, P=0.0005). There was no difference in mortality for metformin users versus non-users (adjusted HR (aHR) 1.11, 95% CI [0.88-1.39]). Utilization of DDDs to stratify metformin users by dose exposure showed no difference in mortality between those with ≥ 365 DDDs (aHR 0.96, 95% CI[0.71-1.29]) or those with < 365 DDDs (aHR 1.28, 95% CI[0.97 -1.70]), as compared to non-users. Among patients with DM or SID diagnosed at any time during follow-up, no significant association with survival was seen among users versus non-users, for any use of metformin (aHR 1.08, 95%CI[0.87 - 1.33]) or when stratified by ≥ 365 DDDs (aHR 0.96 95%CI[0.71-1.29]) or < 365 DDDs (aHR 1.28, 95%CI[0.97-1.70]).

Conclusion: In contrast to a prior study reporting improvements in OS, we found no significant association between metformin use and OS in this cohort of diabetic patients with MM. This is largest cohort study performed to date to assess the role of metformin in patients with MM. This study does not support the prescription of metformin as adjunct therapy in MM.