The Impact of Serum Glucose, Anti-Diabetic Agents, and Statin Usage in Non-small Cell Lung Cancer Patients Treated With Definitive Chemoradiation

Introduction: Epidemiologic data indicate diabetes confers an augmented risk of lung cancer development, yet the relationship between hyperglycemia, metabolic agents, and prognosis is unclear. We analyzed the impact of hyperglycemia, anti-diabetic agents, and statins on outcomes in non-small cell lung cancer (NSCLC) patients undergoing chemoradiation. Method and Materials: In total, data from 170 patients with stage III NSCLC treated at the University of Pittsburgh Medical Center between 2001 and 2014 were obtained for analysis. Kaplan-Meier survival analysis was used to estimate time-to-event for overall survival (OS), disease-free survival, distant metastasis (DM), and loco-regional control (LRC). Blood glucose values (n = 2870), statins, and diabetic medications were assessed both continuously and categorically in univariable and multivariable Cox proportional hazard regression models to estimate hazard ratios and identify prognostic factors. Results: Tumor volume was a negative prognostic factor for OS, disease-free survival, DM, and LRC (p = 0.001). Tumor stage and treatment time were associated with increased all-cause mortality. Any glucose measurement ≥ 130 mg/dl during treatment (2-year estimate 49.9 vs. 65.8%, p = 0.095) was borderline significant for decreased LRC, with similar trends on multivariable analysis (HR 1.636, p = 0.126) and for OS (HR 1.476, p = 0.130). Statin usage was associated with improved 2-year LRC (53.4 vs. 62.4%, p = 0.088) but not with improvements in survival. Other glycemic parameters, comorbid diabetes diagnosis, or anti-diabetic medications were not significantly associated with outcomes. Conclusions: There were trends for blood glucose value over 130 mg/dl and statin nonuse being associated with inferior prognosis for LRC in stage III NSCLC patients; glycemic state, statin usage, and glucose-modulating medications were not associated with survival outcomes in multivariable analysis in this retrospective database.