Postmenopausal women with elevated circulating androgen concentrations have an increased risk of developing breast cancer, yet interventions to reduce androgen levels have not been identified. We examined the effects of a 12-month moderate intensity exercise intervention on serum androgens. The study was a randomized clinical trial in 173 sedentary, overweight (body mass index > or = 24.0 kg/m(2), body fat > 33%), postmenopausal women, ages 50 to 75 years, not using hormone therapy and living in the Seattle, WA area. The exercise intervention included facility-based and home-based exercise (45 minutes, 5 days per week of moderate intensity sports/recreational exercise). A total of 170 (98.3%) women completed the study, with exercisers averaging 171 minutes per week of exercise. Women in the exercise and control groups experienced similar, nonsignificant declines in most androgens. Among women who lost >2% body fat, testosterone and free testosterone concentrations fell by 10.1% and 12.2% between baseline and 12 months in exercisers compared with a decrease of 1.6% and 8.0% in controls (P = 0.02 and 0.03 compared with exercisers, respectively). Concentrations of testosterone and free testosterone among exercisers who lost between 0.5% and 2% body fat declined by 4.7% and 10.4%. In controls who lost this amount of body fat, concentrations of testosterone and free testosterone declined by only 2.8% and 4.3% (P = 0.03 and 0.01 compared with exercisers, respectively). In summary, given similar levels of body fat loss, women randomized to a 12-month exercise intervention had greater declines in testosterone and free testosterone compared with controls. The association between exercise and breast cancer risk may be partly explained by the effects of exercise on these hormones.