The aim of the study was to investigate glucose derangement and its short- and long-term prognostic significance in nondiabetic ischemic stroke patients. The study involved 262 consecutive patients, mean age: 70.1+/-12.4 years, with a supratentorial ischemic stroke. The following data were collected: patients characteristics, risk factors, comorbidities, and stroke severity assessed by the Scandinavian Stroke Scale (SSS). Serum glucose levels were measured on admission, on the next, 2nd, 3rd, 5th, 7th and 14th day after stroke onset. The outcome measures on day 30 were mortality and capacity to perform daily activities: the Barthel Index and Rankin Scale. The 1-year survival was estimated by the Kaplan-Meier method. Cox proportional hazards regression was used to assess predictors of 1-year mortality in nondiabetics. Diabetes mellitus was found in 24.8% of patients and transient hyperglycemia in 36.3% of patients. Patients with transient hyperglycemia scored lower on SSS in the subsequent days of assessment than patients with either diabetes mellitus or normoglycemia. They had larger ischemic lesions on computer tomography (CT) than diabetics and had higher 30-day mortality than normoglycemics (p<0.05). One-year mortality was similar in transient hyperglycemics and diabetics, and both were significantly higher than in normoglycemics (p<0.05). A proportional hazards model analysis showed that transient hyperglycemia is not an independent predictor of death within a year after stroke.