A retrospective study was made of 83 diabetic patients admitted to the Department of Internal Medicine between January 1986 and December 1993, with acute intercurrence of the disease. The authors considered some clinical features (age, type and frequency of complication, form of presentation, causal aetiology, laboratory findings and therapeutic options). Statistical relationships between different populations were made using the unpaired Student's T. method. Results were compared with those in the literature, leading to some significant conclusions.