Abstract
PURPOSE: Diabetes mellitus (DM) is a negative prognostic factor in patients with brain metastases (BM). It is unknown whether a direct impact of serum glucose levels on survival exists. We aimed to detect a potential association of serum glucose levels before and during radiotherapy with survival in patients with BM. METHODS: Patients were included in this retrospective exploratory analysis if at least three fasting and non-fasting serum glucose test results before or during treatment were available. Survival was analyzed with uni- and multivariate Cox regression concerning an association with fasting and maximum glucose levels and regarding potentially confounding dexamethasone intake. RESULTS: A total of 62 patients with BM (15 with and 47 without DM) were included. Patients with a mean fasting glucose of more than 7.8 mmol/l (upper quartile) showed significantly shorter survival compared to patients of the lower three quartiles (hazard ratio [HR] = 2.05, p = 0.021). Further, maximum blood glucose levels of > 12.0 mmol/l (upper quartile) were associated with shorter survival (HR = 1.95, p = 0.035). In the subset of patients without DM, a trend toward worse survival in patients with higher fasting glucose levels was observed (HR = 2.54, p = 0.099). The negative association of high maximum glucose levels with survival persisted in multivariate analysis independently of steroid administration. CONCLUSION: Strong elevations of fasting and maximum serum glucose levels were associated with a worse prognosis in patients with BM with and without DM. This observation warrants further analysis in larger cohorts and has potential implications for clinical practice.