Unlike other pharmacological therapies used in obese type 2 diabetic patients, metformin has been shown to improve glycemic control with lower insulin levels and not to involve weight gain. We therefore examined the effect of adjunct metformin in 13 severely obese type 2 diabetic patients (BMI 39.3 +/- 3.9 kg/m2) in suboptimal glycemic control pretreated with intensified insulin therapy. Patients were randomly assigned to either metformin or placebo treatment (double-blind) for 10 weeks and after a 2 week washout period received the opposite treatment, respectively, for 10 additional weeks. HbA1c decreased comparably during placebo (from 8.1 +/- 0.4 to 7.6 +/- 0.3%) and metformin (from 8.5 +/- 0.4 to 7.4 +/- 0.3%, p = 0.29 vs. placebo). Changes in fasting glucose levels were also not different between placebo (from 9.3 +/- 0.7 to 9.5 +/- 0.7 mM) and metformin (from 10.3 +/- 0.5 to 9.5 +/- 0.6 mM, p = 0.44 vs. placebo). Total exogenous insulin requirements decreased from 53 +/- 10 to 35 +/- 7 units during metformin treatment (p = 0.02 vs. placebo). Changes in fasting insulin levels during placebo and metformin treatments were not different (p = 0.11). Metformin had no effect on body weight and serum triglycerides but marginally decreased serum cholesterol levels (from 239 +/- 18 to 211 +/- 14 mg/dl, p = 0.005, p = 0.08 vs. placebo). During the oral glucose tolerance test no differences were observed in the areas under the curve for glucose and insulin while that for C-peptide showed a tendency to increase during metformin administration. We conclude that addition of metformin to insulin treatment in severely obese type 2 diabetic patients improves glycemia but not hyperinsulinemia in comparison to intensive insulin therapy alone. With adjunct metformin, approximately 30% less exogenous insulin is required. With respect to glycemia and lipids, adjunct metformin can be a reasonable treatment alternative in selected obese patients with type 2 diabetes already on intensive insulin therapy.