Abstract
BACKGROUND: Obesity and physical activity have been hypothesized to affect breast cancer risk partly via the androgen signaling pathway. We conducted the first study to evaluate these associations by tumor androgen receptor (AR) status. METHODS: Height, weight, and physical activity were assessed using questionnaires in the Nurses' Health Study. AR, estrogen receptor (ER), and progesterone receptor (PR) status were determined using immunohistochemistry on tumor tissue and medical/pathology reports. RESULTS: A total of 1,701 AR(+) and 497 AR(-) cases were documented during 26 years of follow-up of 103,577 women. After adjusting for ER/PR status and other risk factors, the relative risks (RR) and 95% confidence intervals (95% CI) for every 5 kg/m(2) increase in body mass index (BMI) were 1.07 (1.01-1.13) for AR(+) and 1.16 (1.05-1.29) for AR(-) tumors (P-heterogeneity = 0.17). The RRs (95% CIs) per 5 hours of brisk walking/week were 0.87 (0.73-1.04) for AR(+) and 0.67 (0.45-0.99) for AR(-) tumors (P-heterogeneity = 0.22). Further, BMI, but not physical activity, associations differed significantly across ER/PR/AR subtypes (P-heterogeneity = 0.04 and 0.63, respectively). The RRs (95% CIs) for 5 kg/m(2) increase in BMI were 1.23 (1.04-1.45) for ER(+)PR(+)AR(-), 1.19 (1.01-1.39) for ER(-)PR(-)AR(-), 1.15 (1.08-1.23) for ER(+)PR(+)AR(+), and 0.88 (0.75-1.03) for ER(+)PR(-)AR(+) tumors. CONCLUSIONS: Higher BMI was associated with an increased risk of both AR(+) and AR(-) breast tumors in postmenopausal women, whereas physical activity, including brisk walking, was associated with a reduced risk of both subtypes. In addition, a significant positive association was observed between higher BMI and ER(-)PR(-)AR(-) tumors. IMPACT: The similar associations observed by AR status suggest that mechanisms other than androgen signaling underlie these two breast cancer risk factors.