In a prospective randomised study in 20 insulin-dependent diabetics who had minor surgery under general anaesthesia we compared the metabolic responses to intravenous glucose-insulin-potassium infusion with those who had conventional subcutaneous insulin administration. The former treatment resulted in lower blood glucose levels both during the infusion period (p less than 0.05) as well as the entire observation period (operative, first and second postoperative days; p less than 0.01). More blood glucose values were within the intended range of 5 to 10 mmol/litre in the glucose-insulin-potassium as compared to the conventional group (48% versus 24%; p less than 0.01). The levels of lactate, 3-hydroxybutyrate, glycerol, alanine, glucagon, insulin and growth hormone did not differ between the two groups. The infusion regimen resulted in better glycaemic control both peri-and postoperatively than the conventional subcutaneous insulin regimen in insulin-dependent diabetic patients who have minor surgery.