The aim of this study was to evaluate the effects of exercise training and hormone replacement therapy (HRT) on serum leptin levels in older women. Previously sedentary, healthy women, aged 60-72 yr, were assigned to control (n = 16), exercise (n = 17), HRT (n = 15), or exercise + HRT (n = 13) groups. Exercise training consisted of a 2-month flexibility-exercise program followed by a 9-month exercise program that included walking, jogging, and stair climbing. HRT consisted of 11 months of continuous conjugated estrogens (0.625 mg/day) and medroxyprogesterone acetate (5 mg/day) for 13 days every third month. Body composition was assessed by dual-energy x-ray absorptiometry, and serum insulin levels were measured in the fasted state and in response to a glucose challenge. Leptin levels were reduced by 23 ± 25% and 22 ± 27% (both P < 0.01) in response to exercise and exercise + HRT, respectively. There was no effect of HRT on leptin. Fat mass was the strongest predictor of serum leptin concentration, both before (r = 0.81; P < 0.001) and after (r = 0.85; P < 0.001) the study period, and the change in fat mass in the exercisers was significantly correlated with the change in leptin (r = 0.55; P < 0.01). There did not seem to be an effect of exercise, independent of the reduction in fat mass, on leptin. Insulin levels were significantly correlated with leptin levels, but this was not independent of the association with adiposity. The curvilinear relationship between leptin level and fat mass and the finding that the ratio of leptin mass to fat mass decreased after weight loss suggest that fat cell size is an important determinant of circulating leptin levels.