Treatment outcomes of patients with metastatic prostate cancer and co-morbid diabetes mellitus.

Abstract

113 Background: In metastatic castrate resistant prostate cancer (mCRPC) there is a lack of studies that explore the interaction of comorbidities on treatment outcomes, which presents a challenge in choosing the right drug. In patients with diabetes mellitus, two common therapies, enzalutamide (ENZ) and abiraterone (ABI) have different effectiveness due to different mechanisms of action and because ABI requires co-administration of prednisone. Thus, in this study, we aim to assess the survival of patients with comorbid diabetes treated with ENZ and ABI. Methods: Patients treated with AA or ENZ for mCRPC from September 10, 2014, to June 2, 2017 were identified within the Veterans Health Administration. For these patients, presence of diabetes or complicated diabetes was determined using the Charlson method from the International Classification of Diseases (ICD) 9/10 codes. A Kaplan–Meier time to event analysis and the cox proportional hazards modeling was used to analyze the data with the latter including covariates such as age, Charlson Comorbidity Index, body-mass index, treatment with bone-directed therapy, black race, and baseline PSA at start of treatment with ENZ or ABI. Results: We identified 5822 patients treated for mCRPC, of which 2202 had diabetes and were treated using either ENZ (n = 1041) or ABI (n = 1161). Median survival of patients with diabetes treated with ENZ was 3 months longer than in patients treated with ABI (23.8 vs. 20.8, p = 0.002 by log-rank). In 3620 patients without diabetes, no significant difference in median survival (24.7 vs. 22.7 months, p = 0.065) was seen between ENZ (n = 1463) and ABI (n = 2157). In a multivariable model, ENZ was associated with improved survival in patients with diabetes (adjusted HR 0.87, 95% CI 0.79-0.97) and in patients without diabetes (adjusted HR 0.90, 95% CI 0.83-0.98). Conclusions: ENZ was associated with improved 3-month median survival in patients with diabetes compared to ABI in unadjusted analyses. Additionally, in adjusted analyses the cox proportional model also showed significantly better survival in mCRPC patients treated with ENZ both with and without diabetes, although the findings are more robust for patients with diabetes. The cause of these differences is unknown, so there may be differences in efficacy and adverse events between the two agents in real-world use. Further assessment of patient outcomes with comorbid disease is appropriate to improve care of patients with advanced prostate cancer.