Abdominal obesity and hyperinsulinemia play a key role in the development of the polycystic ovary syndrome (PCOS). Dietary-induced weight loss and the administration of insulin-lowering drugs, such as metformin, are usually followed by improved hyperandrogenism and related clinical abnormalities. This study was carried out to evaluate the effects of combined hypocaloric diet and metformin on body weight, fat distribution, the glucose-insulin system, and hormones in a group of 20 obese PCOS women [body mass index (BMI) > 28 kg/m2] with the abdominal phenotype (waist to hip ratio >0.80), and an appropriate control group of 20 obese women who were comparable for age and pattern of body fat distribution but without PCOS. At baseline, we measured sex hormone, sex hormone-binding globulin (SHBG), and leptin blood concentrations and performed an oral glucose tolerance test and computerized tomography (CT) at the L4-L5 level, to measure sc adipose tissue area (SAT) and visceral adipose tissue area. All women were then given a low-calorie diet (1,200-1,400 kcal/day) alone for one month, after which anthropometric parameters and CT scan were newly measured. While continuing dietary treatment, PCOS women and obese controls were subsequently placed, in a random order, on metformin (850 mg/os, twice daily) (12 and 8, respectively) or placebo (8 and 12, respectively), according to a double-blind design, for the following 6 months. Blood tests and the CT scan were performed in each woman at the end of the study while they were still on treatment. During the treatment period, 3 women of the control group (all treated with placebo) were excluded because of noncompliance; and 2 PCOS women, both treated with metformin, were also excluded because they became pregnant. Therefore, the women cohort available for final statistical analysis included 18 PCOS (10 treated with metformin and 8 with placebo) and 17 control women (8 treated with metformin and 9 with placebo). The treatment was well tolerated. In the PCOS group, metformin therapy improved hirsutism and menstrual cycles significantly more than placebo. Baseline anthropometric and CT parameters were similar in all groups. Hypocaloric dieting for 1 month similarly reduced BMI values and the waist circumference in both PCOS and control groups, without any significant effect on CT scan parameters. In both PCOS and control women, however, metformin treatment reduced body weight and BMI significantly more than placebo. Changes in the waist-to-hip ratio values were similar in PCOS women and controls, regardless of pharmacological treatment. Metformin treatment significantly decreased SAT values in both PCOS and control groups, although only in the latter group were SAT changes significantly greater than those observed during the placebo treatment. On the contrary, visceral adipose tissue area values significantly decreased during metformin treatment in both PCOS and control groups, but only in the former was the effect of metformin treatment significantly higher than that of placebo. Fasting insulin significantly decreased in both PCOS women and controls, regardless of treatment, whereas glucose-stimulated insulin significantly decreased only in PCOS women and controls treated with metformin. Neither metformin or placebo significantly modified the levels of LH, FSH, dehydroepiandrosterone sulphate, and progesterone in any group, whereas testosterone concentrations decreased only in PCOS women treated with metformin. SHBG concentrations remained unchanged in all PCOS women; whereas in the control group, they significantly increased after both metformin and placebo. Leptin levels decreased only during metformin treatment in both PCOS and control groups. (ABSTRACT TRUNCATED)