Although insulin and sulfonylureas often have additive clinical effects when used in combination for type II (non-insulin-dependent) diabetes, these results are variable and a clinical role for this approach is not yet established. This study tests the efficacy of a specific combined regimen for a subpopulation of patients with a randomized double-masked placebo-controlled crossover design and under conditions similar to those of clinical practice. Twenty subjects with limited duration (less than 15 yr) type II diabetes who were moderately obese (less than 160% ideal wt) and proved imperfectly controlled on 10 mg glyburide twice daily completed two 4-mo crossover protocols, comparing a single injection of NPH insulin in the evening plus 10 mg glyburide in the morning with insulin plus placebo. Insulin dose was adjusted by experienced endocrinologists seeking the best glycemic control consistent with safety. All subjects had glycosylated hemoglobin values less than or equal to 150% of the control mean on combined therapy, and combined therapy was superior to insulin alone (fasting plasma glucose 8.0 +/- 0.3 vs. 11.1 +/- 0.6 mM, P less than .01; glycosylated hemoglobin 9.8 +/- 0.1 vs. 10.6 +/- 0.2%, P less than .01). Despite greater weight gain on combined therapy, blood pressure and plasma lipid concentrations were the same on the two regimens. These results suggest this simple regimen offers another option, besides multiple injections of insulin, for patients of this kind who are unsuccessful with a sulfonylurea or a single injection of insulin alone.