We recently reported the frequent occurrence of polycystic ovaries and hyperandrogenism associated with weight gain and hyperinsulinemia in women taking valproate for epilepsy. The purpose of this study was to evaluate the risks related to valproate-induced hyperinsulinemia and their reversibility after discontinuing the medication. Sixteen women with valproate-related polycystic ovaries or hyperandrogenism participated in the study. Vaginal ultrasonography was performed, and endocrine and lipid parameters were measured. Thereafter, lamotrigine was substituted for valproate and the patients were observed for 12 months. Twenty-four healthy age-matched women served as control subjects. Twelve women completed the 12-month follow-up. While still on valproate they had centripetal obesity with associated hyperinsulinemia and unfavorable serum lipid profiles. The body-mass index and fasting serum insulin and testosterone concentrations decreased during the first year after replacing valproate with lamotrigine whereas the HDL-cholesterol/total cholesterol ratios increased from 0.17 +/- 0.06 to 0.26 +/- 0.05. The total number of polycystic ovaries in these women decreased from 20 during valproate medication to 11 one year after replacing valproate with lamotrigine. Valproate induces a metabolic syndrome with centripetal obesity, hyperinsulinemia, lipid abnormalities, and polycystic ovaries/hyperandrogenism in women with epilepsy. These valproate-related risks can be reduced by substituting lamotrigine for valproate.