Experimental studies of different stroke models equivocally showed that hyperglycaemia is responsible for the increase of infarct volume and mortality. Similar results were obtained in several, but not all clinical studies. The aim of the study was to assess the occurrence and prognosis of transient hyperglycaemia in non-diabetic, acute ischaemic stroke patients. A consecutive series of 204 patients admitted to the Stroke Unit within 48 hours after the onset of the first-ever hemispheric ischaemic stroke, confirmed by CT and/or autopsy, were included in the study. Blood samples for determination of glucose level were obtained immediately after admission, on the 1-st, 2-nd, 3-rd, 5-th, 7-th and 14-th day of stroke. The fructosamine and HbA1 measurements were used to exclude patients with previous glucose intolerance. The severity of stroke was assessed according to Scandinavian Neurological Stroke Scale on admission, on the first, 7-th, 14-th and 30-th day of stroke. Transient hyperglycaemia, defined as at least one elevated glucose level in the first week of stroke with normal level of HbA1 and fructosamine was found in 65 (31.9%) of patients. Patients with transient hyperglycaemia did not differ from diabetics and normoglycaemic according to age, gender, history of hypertension and other risk factors. 30 day mortality in the group of patients with transient hyperglycaemia was significantly higher than in normoglycaemic ( p. < 0.001) and diabetic patients. Transient hyperglycaemic patients died earlier, mainly on the 7-th day after admission whereas patients with normoglycaemia died mainly on the 18-th day (p < 0.0001). The main reason of death in hyperglycaemic patients were cardiac complications (15/20), in normoglycaemic--the consequences of immobility (8/11) (< 0.01). The results of our study showed that the transient hyperglycaemia occurred in about one third of acute ischaemic stroke patients and resulted in higher 30-day mortality.