We have evaluated the effects of repaglinide, a new non-sulphonylurea oral hypoglycaemic agent that has a stimulatory effect on insulin secretion. Forty-four patients with NIDDM, already treated with a sulphonylurea, took part in an open, randomised, group comparison study of 12 weeks duration, during which they received either repaglinide or glibenclamide twice daily. While glibenclamide had a greater effect on fasting blood glucose (10.4 to 8.6 mmol.l-1), repaglinide significantly lowered postprandial blood glucose (13.8 to 12.2 mmol.l-1). Glycosylated haemoglobin remained unchanged in both groups, and serum fructosamine showed a tendency to fall. With both treatments total cholesterol was significantly decreased after 12 weeks, while HDL-cholesterol and triglycerides did not change. Fasting plasma insulin in the repaglinide group decreased from 80 (median value) to 67 pmol.l-1; it did not change in the glibenclamide group. Two patients in the repaglinide group did not complete the study, one for personal reasons, and one because of a rise in blood glucose. No abnormal findings attributable to repaglinide were observed in clinical and laboratory examinations, and no hypoglycaemic symptoms caused by it were observed.