The aim of our study was to evaluate clinical management of diabetic ketoacidosis (DKA) in a teaching hospital. We followed all the patients hospitalised for DKA over six years (1995-2000), and we recorded clinical data, laboratory finding at entrance and clinical management. We compared the data to the standards set in guidelines. Our study showed an important delay of initiation of intravenous fluid (70% of cases), an under-replacement with intravenous fluid (69% of cases) and with potassium therapy (80% of cases), and an excessive use of alkali therapy. In conclusion, suboptimal management of DKA occurred in a large percentage of patients.