Most obese patients with noninsulin-dependent diabetes mellitus (NIDDM) are initially treated with diet, then with oral hypoglycemic agents, eventually with insulin. However several reports indicate that in these patients insulin therapy has little chance to control glucose metabolism, promotes weight gain and arterial hypertension, and is likely to aggravate insulin resistance. In this randomized, double-blind trial vs. placebo (P) we evaluated in 29 obese NIDDM patients poorly controlled by insulin (daily insulin doses 48.7 +/- 4.0 U/day, HbA1c 10 +/- 0.27%, mean daily blood glucose levels 12.3 +/- 0.3 mmol/L, fasting C-peptide 1.8 +/- 0.2, C-peptide after 1 mg iv glucagon 3.2 +/- 0.3 ng/mL, means +/- SE), the clinical and metabolic effects of benfluorex (B), a lipid-lowering drug able to improve insulin sensitivity. After a 2-3 week run-in period (1 tablet P at dinner and diet 800 cal/day to lose 5% of the initial body weight (BWi), patients received a 1000 kcal/day diet and were randomized to B, 150 mg/ tablet, or P (3 tablets/day); the time limit was set at a 10% decrease of BWi or at 90 days. At the end of run-in there was a significant reduction of BWi (P < 0.001), fasting (P = 0.002) and mean daily blood glucose levels (P < 0.001), triglycerides (P = 0.02), cholesterol (P < 0.001) and daily insulin doses (P < 0.001). At the end of the double-blind trial, weight-loss was greater (P < 0.05), faster (P = 0.018), and more frequent (P < 0.05) with B than with P, and systolic blood pressure (P < 0.05) decreased only with B. Considering only patients with a 10% decrease of BWi (B = 15, P = 10), HbA1c (P < 0.001) decreased only with B, while fasting insulin levels decreased with both B (P < 0.01) and with P (P < 0.05). Insulin sensitivity was evaluated by means of a double infusion test (LDIGIT, insulin 25 mU/Kg/h plus glucose 4 mg/kg/min, lasting 150 min) at the end of run-in and at the end of the double-blind trial; at the end of the double-blind trial steady state blood glucose (SSBG, P < 0.05), free fatty acids (FFA, P < 0.05) and blood beta-hydroxybutyrate (P < 0.05) decreased only with B, while blood glycerol decreased both with both P (P < 0.05) and B (P < 0.06). At the end of the double-blind trial, C-peptide release was unchanged with either P or B. In conclusion, benfluorex potentiates the effects of hypocaloric diet on weight loss and on glycemic control in obese NIDDM patients treated with insulin, and this effect seems to be the result of an improved insulin sensitivity.