STUDY OBJECTIVE: To determine if body fat distribution affects breast cancer risk. DESIGN: Prospective case-control study. PATIENTS: The anthropometric measurements of 216 consecutively and newly diagnosed women with invasive carcinoma of the breast were compared with those of 432 age-matched controls. The anthropometric measurements taken were abdomen, thigh, suprailiac, biceps, triceps, subscapular, and midaxillary skinfolds; waist and hip circumference; and weight and height. Women between 25 and 83 years of age were included in the study. RESULTS: Patients with breast cancer had a significantly greater waist:hip circumference ratio than controls (P less than 0.001) and a significantly greater suprailiac:thigh skinfold ratio (P less than 0.001). The relative risk for breast cancer increased with increasing waist:hip circumference ratio (less than 0.73 = 1.00; 0.73 to 76 = 1.90; 0.77 to 0.80 = 2.83; greater than 0.80 = 6.46) and with suprailiac:thigh skinfold ratio (less than 0.42 = 1.00; 0.42 to 0.56 = 1.85; 0.57 to 0.71 = 2.25; greater than 0.71 = 5.85). At other sites of upper body obesity, such as the biceps and triceps, skinfolds were significantly greater in patients with breast cancer. CONCLUSION: Although obese women are at slightly higher risk for developing breast cancer, women with android obesity are a segment of obese women who appear to be at a significantly higher risk for developing breast cancer.