We recently demonstrated that a brief endurance type training program led to increases in thigh muscle mass and peak oxygen uptake (VO(2)) in prepubertal girls. In this study, we examined the effect of training on the GH-->insulin-like growth factor I (GH-->IGF-I) axis, a system known to be involved both in the process of growth and development and in the response to exercise. Healthy girls (mean age 9.17 +/- 0.10 yr old) volunteered for the study and were randomized to control (n = 20) and training groups (n = 19) for 5 weeks. Peak VO(2), thigh muscle volume, and blood samples [for IGF-I, IGF-binding proteins (IGFBP)-1 to -6, and GHBP] were measured. At baseline, IGF-I was significantly correlated with both peak VO(2) (r = 0.44, P < 0.02) and muscle volume (r = 0.58, P < 0.004). IGFBP-1 was negatively correlated with muscle volume (r = -0.71, P < 0.0001), as was IGFBP-2. IGFBP-4 and -5 were significantly correlated with muscle volume. We found a threshold value of body mass index percentile (by age) of about 71, above which systematic changes in GHBP, IGFBP-1, and peak VO(2) per kilogram were noted, suggesting decreases in the following: 1) GH function, 2) insulin sensitivity, and 3) fitness. Following the training intervention, IGF-I increased in control (19.4 +/- 9.6%, P < 0.05) but not trained subjects, and both IGFBP-3 and GHBP decreased in the training group (-4.2 +/- 3.1% and -9.9 +/- 3.8%, respectively, P < 0.05). Fitness in prepubertal girls is associated with an activated GH-->IGF-I axis, but, paradoxically, early in a training program, children first pass through what appears to be a neuroendocrine state more consistent with catabolism.