Obesity has been linked to increased risk of several malignancies, but the role of obesity in the etiology of ovarian cancer remains unclear. Therefore, a hospital-based case-control study was conducted to investigate the association between body size and risk of ovarian cancer. Participants included 427 women with primary, incident ovarian cancer and 854 cancer-free controls. All participants received medical services at Roswell Park Cancer Institute in Buffalo, NY between 1982 and 1998 and completed a comprehensive epidemiological questionnaire. The instrument included questions regarding height and usual wt prior to survey. Participants were classified as underweight/normal (BMI < or = 24.9 kg/m2), overweight (BMI 25.0-29.9 kg/m2), or obese (BMI > or = 30.0 kg/m2). Compared with underweight/normal participants, being overweight (adjusted odds ratio [OR] = 1.02; 95% CI 0.77-1.36) or obese (adjusted OR = 1.17; 95% CI 0.84-1.65) was not significantly associated with an elevated risk of ovarian cancer. After stratification by menopausal status, BMI showed no significant association to ovarian cancer risk among postmenopausal women (> or = 50 y old). However, among premenopausal women (<50 y old), those classified as obese had a significantly increased risk (adjusted OR = 2.19; 95% CI 1.19-4.04) compared with women classified as normal/underweight. These findings suggest a potential influence of menopausal status on the total endogenous hormonal environment, including estrogens, androgens, and insulin-like growth factors, when considering the association between body size and ovarian cancer risk. In light of the fact that obesity is a modifiable risk factor, further investigation on this topic is warranted.