Body mass index (BMI) is associated with breast cancer risk, but its relationship with stage at diagnosis is unclear. BMI was calculated for patients in the North American Fareston and Tamoxifen Adjuvant trial, and was correlated with clinicopathologic factors, including stage at diagnosis. One thousand eight hundred fourteen patients were enrolled in the North American Fareston and Tamoxifen Adjuvant study; height and weight were recorded in 1451 (80%) of them. The median BMI was 27.1 kg/m2 (range, 14.7-60.7). The median patient age was 68 years (range, 42-100); median tumor size was 1.3 cm (range, 0.1-14 cm). One thousand seven hundred ninety-three (99.0%) patients were estrogen receptor positive, and 1519 (84.7%) were progesterone receptor positive. There was no significant relationship between BMI (as a continuous variable) and nodal status (P = 0.469), tumor size (P = 0.497), American Joint Committee on Cancer stage (P = 0.167), grade (P = 0.675), histologic subtype (P = 0.179), or estrogen receptor status (P = 0.962). Patients with palpable tumors, however, had a lower BMI than those with nonpalpable tumors (median 26.4 kg/m2 vs 27.5 kg/m2, P < 0.001). Similar results were found when BMI was classified as a categorical variable (<25 vs 25-29.9 vs > or =30). Increased BMI does not lead to a worse stage at presentation. Obese patients, however, tend to have nonpalpable tumors. Mammography in this population is especially important.