BACKGROUND: Physical activity is associated with reduced risk for breast cancer, perhaps through reductions in circulating reproductive hormones (estrogens and androgens). There may also be a role for physical activity in regulating menopausal symptoms. Few studies have examined associations of physical activity on hormone levels. None have examined the potential effect of the menopausal transition on the associations between physical activity and reproductive hormone levels. MATERIALS AND METHODS: Data from the Penn Ovarian Aging Study were used for this analysis. Self-reported physical activity was assessed in 391 women up to four times over 10 years and extending across the menopausal transition. Other assessments included reproductive hormones via RIA (estradiol, luteinizing hormone, follicle-stimulating hormone, testosterone, DHEA sulfate), body weight, and height. Multivariate repeated measures regression models were developed to compare reproductive hormone levels within physical activity tertiles, adjusting for age, follow-up time, smoking, and ethnicity. RESULTS: Activity level was inversely associated with estradiol in the subgroup in the late transition stage. Adjusted means for estradiol were 24.6 and 37.9, a relative difference of 54% in estradiol when comparing highest to lowest activity tertile (P = 0.02). Similarly, in this subgroup, there was an inverse association between physical activity and testosterone levels (means of 11.1 and 15.94 in the highest and lowest tertile, a 47% relative difference; P = 0.01). There were no significant associations of activity with any other reproductive hormone. CONCLUSIONS: These results identify a particular window of the menopausal transition during which physical activity is associated with reduced estradiol and/or testosterone levels.