Women with oligomenorrhea and polycystic ovaries show a high incidence of ovulation failure perhaps linked to insulin resistance and related metabolic features. A number of reports show that the biguanide metformin improves ovarian function. However, in these trials the quality of evidence supporting ovulation is suboptimal, and few studies have been placebo-controlled. The aim of our study was to use a double-blind, placebo-controlled approach with detailed assessment of ovarian activity (two blood samples per week) to assess the validity of this therapeutic approach in this group of women. Of the 94 patients randomized, 2 withdrew before treatment commenced, 47 received placebo, and 45 received metformin (850 mg, twice a day). The numbers discontinuing the study prematurely were higher in the treatment group (n = 15) than the placebo group (n = 5; P < 0.05). The ovulation frequency assessed by the ratio of luteal phase weeks to observation weeks was significantly (P < 0.01) higher in the treated group (23%) compared with the placebo (13%), and the time to first ovulation was significantly (P < 0.05) shorter [23.6 d; 95% confidence interval (CI), 17, 30; compared with 41.8 d; 95% CI, 28, 56]. The proportion of patients failing to ovulate during the placebo-treatment period was higher (P < 0.05) in the placebo group, and the majority of ovulations were characterized by normal progesterone concentrations in both groups. The effect of metformin on follicular maturation was rapid, because the E2 circulating concentration increased over the first week of treatment only in the metformin group. Significant (P < 0.01) weight loss (and leptin reduction) was recorded in the metformin group, whereas the placebo group actually increased weight (P < 0.05). A significant increase in circulating high-density lipoprotein was observed only in the metformin-treated group. Metabolic risk factor benefits of metformin treatment were not observed in the morbidly obese subgroup of patients (body mass index > 37). No change in fasting glucose concentrations, fasting insulin, or insulin responses to glucose challenge was recorded after 14-wk metformin or placebo therapy. There was an inverse relationship between body mass and treatment efficacy. We show in a large randomized placebo-controlled trial that metformin treatment improves ovulation frequency in women with abnormal ovarian function and polycystic ovaries significantly but to a modest degree, and protracted treatment improves cardiovascular risk factors. These data support a beneficial effect of metformin in improving ovarian function in women with oligomenorrhea and polycystic ovaries.