Metformin does not improve survival of cholangiocarcinoma patients with diabetes

Potential conflict of interest: Dr. Roberts advises and received grants from Gilead, Wako, Ariad, and Nordion. Financial support: Dr Zhen Yang received support from the National Natural Science Foundation of China (No. 81302122) and the China Scholarship Council (File No. 201409370014). This project was supported by a Clinical and Translational Science Award (CTSA) Grant (No. TL1 TR000137 and UL1TR000135) from the National Center for Advancing Translational Science (NCATS) and by The Cholangiocarcinoma Foundation. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health. To the Editor: Mounting evidence has shown a protective effect of metformin therapy against cancer development and a potential benefit of metformin on cancer survival.1 We reported in Hepatology that metformin use was associated with a 60% reduction in risk of intrahepatic cholangiocarcinoma (CCA) among persons with diabetes mellitus (DM) (odds ratio = 0.4, 95% confidence interval [CI] 0.2‐0.9; P = 0.04).3 To our knowledge, no previous studies have examined whether metformin use improves CCA survival. We performed a retrospective analysis of 250 diabetic patients with newly diagnosed CCA seen at Mayo Clinic in Rochester, Minnesota, between January 2001 and December 2012. Survival of patients with DM and CCA taking metformin versus not taking metformin was calculated from the CCA diagnosis date to the last follow‐up date using the Kaplan‐Meier method and compared using the log‐rank test. Survival predictors were determined by hazard ratio (HR), and 95% CIs were calculated using Cox proportional hazards analysis. Among the 250 patients (65% male, mean age 68 years), median survival was 9.5 months. The survival of 49 patients who continued taking metformin after CCA diagnosis was not different from that of 165 patients who had never taken metformin (9.1 versus 9.2 months; HR = 0.8, 95% CI 0.6‐1.2; P = 0.31). A history of any metformin use before CCA diagnosis (n = 79) also did not affect survival. The results were consistent when categorized by CCA subtype (114 intrahepatic, 105 perihilar, and 31 distal). By multivariate analysis, Eastern Cooperative Oncology Group (ECOG) score (HR = 8.9, 95% CI 3.4‐20.6; HR = 1.7, 95% CI 1.01‐2.8; HR = 1.02, 95% CI 0.7‐1.5; and HR = 1.0 [reference] for ECOG scores of 3‐4, 2, 1, and 0, respectively; P < 0.001), cancer antigen 19‐9 (CA19‐9) value ≥1000 U/mL (HR = 2.6, 95% CI 1.7‐4.0; P < 0.001), and tumor size (HR = 1.05/cm, 95% CI 1.01‐1.11/cm; P = 0.02) were significantly associated with survival. Other patient and tumor characteristics were not associated with survival (Table 1). Table 1 - Analysis of Variables Associated With the Survival of Patients With CCA and DM Variable Deaths/Total no. of Patients (Overall, 208/250) Multivariate Analysis of Variable HR of Death Comparing Metformin User Versus Nonuser, Stratified by Comorbidities and Tumor Characteristics Yes No HR (95% CI)a P Value HR (95% CI)b P Value HR (95% CI)b P Value Smoking Ever 114/132 0.95 (0.7‐1.4) 0.85 0.8 (0.5‐1.3) 0.40 Never 90/113 1.0 (reference) 0.8 (0.4‐1.4) 0.40 PSC Yes 11/20 0.8 (0.3‐1.5) 0.45 c No 197/229 1.0 (reference) 0.7 (0.5‐1.0) 0.08 ECOGd 3/4 14/15 8.9 (3.4‐20.6) 0.5 (0.1‐3.2) 0.50 2 36/39 1.7 (1.01‐2.8) 1.3 (0.5‐2.9) 0.50 1 74/84 1.02 (0.7‐1.5) 1.3 (0.6‐2.5) 0.50 0 80/107 1.0 (reference) <0.001 0.6 (0.3‐1.1) 0.10 CA19‐9 ≥1000 U/mL Yes 45/52 2.6 (1.7‐4.0) <0.001 1.1 (0.4‐2.4) 0.90 No 142/176 1.0 (reference) 0.8 (0.5‐1.3) 0.40 Tumor size Per cm 1.05 (1.01‐1.11) 0.02 >3 cm 0.6 (0.4‐1.0) 0.07 ≤3 cm 0.9 (0.5‐1.7) 0.80 Vascular encasement Yes 73/86 1.04 (0.7‐1.5) 0.80 0.7 (0.4‐1.3) 0.30 No 128/156 1.0 (reference) 0.9 (0.5‐1.4) 0.70 Lymph node metastasis Yes 63/70 1.2 (0.8‐1.7) 0.47 0.8 (0.4‐1.4) 0.40 No 140/174 1.0 (reference) 0.7 (0.4‐1.1) 0.10 Intrahepatic metastasis Yes 62/69 1.04 (0.7‐1.6) 0.84 0.8 (0.4‐1.7) 0.60 No 143/178 1.0 (reference) 0.7 (0.4‐1.1) 0.10 Other organ metastasis Yes 37/42 1.3 (0.8‐2.1) 0.30 0.6 (0.2‐1.3) 0.20 No 168/205 1.0 (reference) 0.8 (0.5‐1.2) 0.30 Metformin use Yes 38/49 0.9 (0.6‐1.4) 0.70 No 143/165 1.0 (reference) aRestricted to variables with a P ≤ 0.10 in the univariate analysis except for metformin use (P = 0.31). The model also included age and sex as potential confounders.bAdjusted for age and sex.cNot calculated because only one patient in the primary sclerosing cholangitis (PSC) group took metformin.dP values were 0.90, 0.04, and 0.05 for ECOG 1, 2, and 3/4, respectively, compared with ECOG 0 as a reference group. Metformin exerts a chemotherapeutic effect on cancer by reducing cell proliferation and inducing cell cycle arrest and apoptosis. However, metformin did not improve the survival of CCA patients with DM. Although metformin may reduce cancer risk, we were not able to demonstrate an effect on CCA survival. The main study limitation is the small number of patients, particularly those with distal CCA and metformin exposure.