Aims: Evaluate the association between admission blood glucose (ABG) and mortality in older patients with or without diabetes mellitus (DM) hospitalized for acute ischemic stroke (AIS). Methods: Observational data of patients >= 65 years, admitted for AIS between January 2011 and December 2013. ABG levels were classified to categories: <= 70 (low), 70-110 (normal), 111-140 (mildly elevated), 141-180 mg/dl (moderately elevated) and >180 mg/dl (markedly elevated). Main outcome was all-cause mortality at the end-of-follow-up. Results: Cohort included 854 patients, 347 with (mean +/- SD age 80 +/- 8, 44% male), and 507 without DM(mean +/- SD age 78 +/- 8, 53% male). There was a significant interaction between DM, ABG and mortality at end-of-follow-up (p <= 0.05). In patients without DM there was a dose-dependent association between ABG category and mortality: adjusted hazard ratios (95% CI) compared to normal ABG were 1.8 (1.2-2.8), 2.9 (1.6-5.2) and 4.5 (2.1-9.7), respectively, for mildly, moderately and markedly elevated ABG. In patients with DM there was no association between ABG and mortality. There was no interaction between DM, ABG and in-hospital mortality or length of stay (LOS). Irrespective of DM status, compared to normal ABG levels, increased ABG category was associated with increased in-hospital mortality: adjusted odds ratios were 3.9 (1.1-13.4), 7.0 (1.8-28.1), and 20.3 (4.6-89.6) with mildly, moderately and markedly elevated ABG, respectively. Mean LOS was 6 +/- 5, 7 +/- 8, 8 +/- 7, and 8 +/- 8 days, respectively. Conclusion: In older patients without DM hospitalized for AIS, elevated ABG is associated with increased long-term mortality. Irrespective of DM status, elevated ABG was associated with increased in-hospital mortality and LOS. (C) 2017 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved.