Abstract
INTRODUCTION: Studies have consistently identified an increased risk of pancreatic cancer in diabetics, yet the role hyperglycemia may play in predicting prognosis is less clear. This work aims to evaluate the impact of glycemic state and antidiabetics on outcomes after systemic and local treatment for locoregionally advanced pancreatic cancer. MATERIALS AND METHODS: This retrospective study consisted of 303 patients with newly diagnosed advanced-stage pancreatic cancer treated from 2004 to 2014. Kaplan-Meier survival analysis method was used to estimate time to event for overall survival, distant metastasis, and locoregional control. Blood glucose values (n=8599) were assessed both as continuous and categorical variables in univariate and multivariable Cox proportional hazard regression models to estimate hazard ratios (HRs) and identify independent prognostic factors. A 6-month conditional landmark analysis excluding patients with <6 months follow-up or survival was conducted. RESULTS: Median follow-up and survival was 18.1 and 18.4 months, respectively. On univariate analysis, maximum pretreatment glucose value was associated with reduced overall survival (HR 1.005, P=0.023) and locoregional control (HR 1.001, P=0.001). A pretreatment glucose value ≥200 mg/dL was associated with increased mortality in multivariable analysis (adjusted HR 1.01, P=0.015). After conditional analysis, glucose ≥200 mg/dL before local treatment was associated with reduced overall survival (adjusted HR 1.562; 95% confidence interval [CI], 1.16-2.11; P=0.003). CONCLUSIONS: Elevated blood glucose before treatment of locoregionally advanced pancreatic cancer was associated with poorer outcomes. These findings should be incorporated in future clinical trial design.