Twenty Type 2 diabetic patients, recently converted to insulin because of inadequate control on oral hypoglycaemic agents, were studied. Endogenous insulin reserve was measured by glucagon stimulation, and insulin-mediated peripheral glucose disposal by a hyperglycaemic (11 mmol l-1) clamp, measurements being repeated after 8 weeks of glipizide or placebo therapy in addition to the patients' usual insulin. The study was randomized and double blind. Fasting and stimulated C-peptide levels did not change on glipizide or placebo. Insulin-mediated glucose disposal rose from 2.5 (1.5-8.0) (median (range] to 4.2 (2.3-8.4) mg kg-1 min-1 in the glipizide group (n = 9, p less than 0.01), but did not change in the placebo group. Glycosylated haemoglobin did not change in either group, but median fasting plasma glucose fell from 10.6 (6.1-15.1) to 9.0 (6.4-11.2) mmol l-1 in the glipizide group (p less than 0.02). Fasting insulin rose on glipizide from 10.1 (4.0-23.2) to 13.0 (6.4-33.8) mU l-1 (p less than 0.02). Insulin dosage fell in the glipizide group from 36 to 26 U day-1, as 4 patients experienced hypoglycaemic symptoms. HDL-Cholesterol fell in all patients on glipizide, from 0.94 (0.79-2.13) to 0.79 (0.62-1.95) mmol l-1 (p less than 0.01). Combination of insulin with the sulphonylurea glipizide in secondary failure Type 2 diabetic patients leads to increased insulin-mediated peripheral glucose disposal. Glipizide may have an adverse effect on HDL-cholesterol, which is unrelated to change in diabetic control.