This study examined prospectively the associations of waist circumference and waist:hip circumference ratio with risk of breast cancer. A total of 47,382 US registered nurses who reported their waist and hip circumferences in 1986 were followed up through May 1994 for identification of incident cases of breast cancer. During 333,097 person-years of follow-up, 1,037 invasive breast cancers were diagnosed. In proportional hazards analyses, waist circumference was nonsignificantly related to risk of premenopausal breast cancer but was significantly associated with postmenopausal breast cancer after adjustment for established breast cancer risk factors (for the highest quintile of waist circumference vs. the lowest, relative risk (RR) = 1.34; 95% confidence interval (CI): 1.05, 1.72). When the analysis was limited to postmenopausal women who had never received hormone replacement therapy, a stronger positive association was found (RR = 1.88; 95% CI: 1.25, 2.85). After the data were further controlled for body mass index, the positive association was only slightly attenuated (RR = 1.83; 95% CI: 1.12, 2.99). Among past and current postmenopausal hormone users, no significant associations were found. Similar but slightly weaker associations were observed between waist:hip ratio and breast cancer risk. These data suggest that greater waist circumference increases risk of breast cancer, especially among postmenopausal women who are otherwise at lower risk because of never having used estrogen replacement hormones.