Background: Breast cancer survival is inversely related to body mass index (BMI), but previous studies have not included large numbers of older women. This study investigated the association between BMI and mortality after breast cancer diagnosis in a cohort of older Caucasian women enrolled in the Study of Osteoporotic Fractures. Methods: All women were age ≥65 at study entry (N = 533). Cox proportional hazards regression analysis was used to determine the effect of BMI as a continuous variable on risk of all-cause, cardiovascular, any cancer, and breast cancer mortality. Interaction terms were included to evaluate effect modification by age at diagnosis. Results: Mean age at diagnosis was 78.0 years (SD 5.7) with an average of 8.1 years (SD 4.4) of follow-up after diagnosis. There were 206 deaths during follow-up. The effect of BMI on mortality depended on age (Pinteraction = 0.02). At age 65, the risk of mortality was 1.4 times higher for a BMI of 27.3kg/m2 [95% confidence interval (95% CI), 1.03-2.01] and 2.4 times higher for a BMI of 34.0 kg/m2 (95% CI, 1.07-5.45) compared with women with a BMI of 22.6 kg/m2. At age 85, risk of death was lower for a BMI of 27.3kg/m 2 (hazard ratio, 0.81; 95% CI, 0.65-1.01) or a BMI of 34.0 kg/m 2 (hazard ratio, 0.61; 95% CI, 0.36-1.02) compared with a BMI of 22.6 kg/m2. Similar results were observed for any cancer and breast cancer mortality. BMI was not associated with cardiovascular mortality. Conclusions: In this population of older women, the effect of increased BMI on risk of mortality after breast cancer varied by age. These results differ from those observed among populations of younger postmenopausal breast cancer survivors.