This study examines the effect of glipizide therapy in patients with non-insulin-dependent diabetes mellitus (NIDDM) previously treated with a low-dose insulin regimen. To determine the comparative safety and efficacy of these two treatment regimens, 135 patients with NIDDM who received low-dose insulin (40 units per day or less) were enrolled in a prospective, multicenter trial. After a four-week period of continued insulin therapy, therapy was converted to the second-generation oral sulfonylurea, glipizide. This report includes efficacy data on 79 patients with NIDDM who completed an eight- to 10-week course of glipizide therapy. A majority of these patients were obese (more than 120 percent of ideal body weight), with a mean age of 61.4 years, and a mean duration of diabetes of 10.0 years. Conversion from low-dose insulin resulted in no deterioration in overall glycemic control as reflected by an unchanged glycosylated hemoglobin at 16 weeks. Seventy patients (89 percent) attained fair glucose control and continued an additional 12 weeks of glipizide maintenance therapy. Subgroup analysis by fasting plasma glucose criteria demonstrated that 44 percent of these patients had statistically significant improvements in glucose control with glipizide as compared with insulin therapy. As this study showed, many patients with NIDDM currently receiving treatment with low-dose insulin can have maintained and some even improved glucose and lipid parameters when therapy is converted to glipizide. The rate of hypoglycemic adverse reactions corrected for duration of treatment in all 135 patients was 0.32 event per patient-month of therapy with insulin as compared with 0.12 event per patient-month of therapy with glipizide. When issues of efficacy, safety, and convenience are considered, it might be more appropriate to administer an oral sulfonylurea to patients with NIDDM rather than proceeding to therapy with insulin.