To evaluate the prevalence and prognostic significance of hyperglycemia in acute myocardial infarction, we studied 700 patients (mean age 63.3 +/- 10.97) subsequently admitted to the UCIC of Tradate Hospital during the period January 1976 to December 1987. Patients were followed up for a median period of four years. On the basis of fasting blood glycaemia values in the first five days of hospitalization, excluding the admission day, patients were divided into groups: 401 patients (57.0%) with constantly normal glycaemia; 84 patients (12.0%) with glycaemia equal or superior to 120 mg/100 ml, and with subsequent normalization; and 215 patients (31.0%) with diabetes mellitus diagnosed before hospitalization and/or with persistent hyperglycaemia. The overall mortality was 284 (40.5%) and cardiovascular deaths were 90.8%. Within the first month of myocardial infarction 98 patients died. The mortality rate was 9.4% in normoglycaemic patients and 20.2% in transient hyperglycaemic patients, similar to the value observed in diabetic patients (20.0%). During the follow-up 186 patients died. Late mortality after the first month is higher in diabetic patients (40%) and patients with transient hyperglycaemia (37%) compared to normoglycaemic patients (25.3%). Multivariate analysis shows that independent predictive variables are: for mortality in the first month, Killip class only; and after the first month, Killip class, metabolic classification, sex and supraventricular arrhythmias. The present study shows that transient hyperglycaemia has a low prevalence in the first days of acute myocardial infarction. Transient hyperglycaemia could be attributed not only to increased sympathetic tone elicited by acute myocardial infarction, but is probably a pathologic condition with an adverse outcome to which multiple factors contribute.