OBJECTIVE TO investigate the effects of the addition of glyburide to the regimen of insulin-treated non-insulin-dependent diabetes mellitus (NIDDM) patients with regard to their overall insulin requirement and dosage schedule and to assess persistence of these effects. METHODS A double-blind randomized parallel-groupo, placebo-controlled, 20-wk outpatient trial at the Clinical Research Unit (CRU) at St. Luke's/Roosevelt Hospital (New York). Subjects were 20 insulin-dependent NIDDM patients previously managed on insulin alone. After a baseline period of satisfactory diabetes control on biosynthetic human insulin alone, insulin dosage was halved, and patients were placed on a combination with either glyburide or placebo. Diabetes control equivalent to baseline was reestablished by adjusting insulin as required on subsequent visits to the CRU. RESULTS Insulin requirements in the glyburide group decreased by 29 U at 14 wk compared with 9 U in the placebo group (P less than 0.05). At 20 wk, the decreases remained significant (25 vs. 11 U, respectively; P less than 0.05). The mean +/- SD reduction in insulin requirement in the glyburide group was relatively constant (25 +/- 10 U) and was not related to premedication insulin requirement. Successful response to glyburide was inversely correlated with initial serum alkaline phosphatase level. CONCLUSIONS Glyburide reduces insulin requirements for 20 wk of combination therapy in NIDDM patients. Patients whose initial insulin requirement is less than or equal to 25 U have a 50% chance of achieving equivalent glycemic control on glyburide alone.