To investigate putative abrogating effects of habitual endurance exercise on age-related changes in endocrine function and body composition, we compared insulin-like growth factor-I (IGF-I), sex hormonal status and body composition in 15 Masters runners and 15 minimally exercising men (MEM) aged 60-70 years. A higher maximal oxygen uptake (VO 2 max.) in the runners (41.4 Ϯ 1.6 compared with 27.3 Ϯ 1.4 ml/kg/min, P ¼ 0.0001; mean Ϯ S.E.M.) reflected our group allocations. Analysis of body composition and bone mineral density (BMD) by dual energy X-ray absorptiometry showed no group differences in lean tissue mass or in regional or whole body BMD, but MEM were heavier, reflecting greater adiposity. Of nine muscle groups tested, only quadriceps strength differed significantly, being greater in runners (60.3 Ϯ 2.8 compared with 51.1 Ϯ 2.3 kg, P ¼ 0.02). Total IGF-I (129 Ϯ 10 compared with 124 Ϯ 11 ng/ml, P ¼ 0.72) and IGF-binding protein-3 (2854 Ϯ 94 compared with 2623 Ϯ 128 ng/ml, P ¼ 0.16), were similarly depressed compared with young adult norms in both groups. There was no relationship between total or bioavailable IGF-I and any body composition, BMD or muscle strength variable. In the runners, concentrations of total testosterone (19.1 Ϯ 0.8 compared with 15.0 Ϯ 0.9 nmol/l, P ¼ 0.002) and sex hormone binding globulin (SHBG) (124.4 Ϯ 21.6 compared with 67.7 Ϯ 11.6 nmol/l, P ¼ 0.03) were significantly greater, but the free androgen index was significantly lower (20.7 Ϯ 2.7 compared with 31.4 Ϯ 4.1, P ¼ 0.04). Directly measured free testosterone, however, was similar between the runners and MEM (47.9 Ϯ 1.8 compared with 47.1 Ϯ 2.0 nmol/l, P ¼ 0.80). Therefore the group differences in total testosterone and free androgen index were due to their different SHBG concentrations. Although estrone concentration was higher in MEM (85.1 Ϯ 5.2 compared with 108 Ϯ 6.7 pmol/l, P ¼ 0.03), estradiol concentration was similar between groups (73.0 Ϯ 6.3 compared with 81.8 Ϯ 8.0 pmol/l, P ¼ 0.18), indicating that estrogens were not responsible for the increased SHBG in runners. These results indicate that even high levels of regular endurance exercise do not prevent the decline in the somatotropic axis that occurs with aging. Furthermore, the somatic effects of exercise in older men (reduced adiposity and increased regional muscle strength) occurred independently of somatotropic or androgen status. Although habitual exercise does not influence free testosterone concentrations in older men, it appears to enhance the age-associated increase in SHBG synthesis.