Objective: Biliopancreatic diversion (BPD) resolves type 2 diabetes in near totality of morbidly obeses [BMI (body mass index) >= 35 kg/m(2)]. However, studies of BPD effect in BMI range 25.0 to 34.9 kg/m(2), including about 90% of diabetic patients, are lacking. Materials and Methods: If BPD effects are independent of weight changes, they should be maintained in patients who, being mildly obese or overweight, will lose little or no weight after operation. Thirty type 2 diabetic patients with BMI 25 to 34.9 were submitted to BPD and monitored 12 months. Thirty-eight diabetic patients selected from a large database, kept 1 year on medical therapy, served as controls. Results: Nineteen male and 11 female. Mean age 56.4 +/- 7.4 years, weight 84.8 +/- 11.1 kg, BMI 30.6 +/- 2.9 kg/m(2), waist circumference 104 +/- 9.4 cm, diabetes duration 11.2 +/- 6.9 years, HbA(1c) 9.3 +/- 1.5. Twelve patients on insulin. Fifteen (2 F) with BMI < 30 (mean: 28.1). No mortality or major adverse events occurred. BMI progressively decreased, stabilizing around 25 since the fourth month, without excessive weight loss. One year after BPD, mean HbA1c was 6.3%+/- 0.8, with 25 patients (83%) controlled (HbA1c <= 7%) on free diet, without antidiabetics, and the remaining improved. Acute insulin response to intravenous glucose had increased from 1.2 +/- 2.9 to 4.2 +/- 4.4 mu IU/mL. Diabetes resolution correlated positively with BMI. HbA(1c) decreased at 1 year in the control group, along with an overall increased amount of antidiabetic therapy. Conclusions: BPD improves or resolves diabetes in BMI 25 to 35 without causing excessive weight loss, its action being on insulin sensitivity and beta-cell function. The strikingly different response between morbidly obese and low BMI patients might depend on different beta-cell defect. Clinical Trials. gov Identifier: NCT00996294