To evaluate the effects of metformin on insulin resistance, ovarian androgen production, and clomiphene-induced ovulation and pregnancy rates in infertile women with polycystic ovary syndrome (PCOS).Twenty-one infertile women with PCOS were selected in this prospective randomized clinical study. Basal steroid and gonadotropin levels were measured, and oral glucose tolerance test (OGTT) was performed. The patients were divided randomly into group 1 (n = 11) and group 2 (n = 10). Group 1 patients were treated with 1700 mg per day of metformin for 3 months. The basal tests and OGTT were repeated after metformin therapy. Group 2 patients did not receive metformin. The patients in both groups received 100 mg of clomiphene citrate (CC) daily for 5 days until either a pregnancy occurred, or six CC cycles were reached. Metformin administration continued during CC therapy until the day of hCG in group 1. Serum progesterone (P) level >or=5 ng/ml was considered as confirmatory of ovulation. Ovulation and pregnancy rates after six cycles were determined.Serum androgens and insulin response to OGTT decreased significantly after metformin therapy. Midluteal serum P level was significantly higher in cycles treated with metformin plus CC (P < 0.05). The ovulation (38 of 51 cycles, 74.4% versus 34 of 55 cycles, 61.8%) and pregnancy rates (5 of 11 women, 45.5% versus 3 of 10 women, 30%) were higher, but not significantly, in the metformin plus CC group than in the CC alone group. All the patients who conceived had insulin resistance in group 1 whereas non-insulin resistance in group 2.Metformin improves insulin resistance and reduces androgen levels. Metformin did not increase significantly the ovulation and pregnancy rates.