The relationship between 24-h GH secretion, insulin-like growth factor (IGF) I serum levels, IGF-binding protein 3 (IGFBP-3) levels and height was studied in 114 healthy children and adolescents (147 tests). A significant correlation was found between the spontaneous GH secretion expressed as the area under the curve above baseline (AUCb) or the calculated 24-h GH secretion rate (GHb) and IGF-I (n = 147, r = 0.65, or 0.78, respectively, P < 0.001) or IGFBP-3 levels (r = 0.48 or 0.62, P < 0.001). Correlations were also significant within the various subgroups [females (n = 51), males (n = 96), prepubertal children (n = 75), pubertal children (n = 72)]. The slopes of the regression lines of IGF-I levels vs. AUCb or GHb were clearly steeper in pubertal children than in prepubertal ones; this was not as evident with IGFBP-3. In prepubertal children, a significant correlation was found between height (compared to Swedish reference values) standard deviation scores (SDS) and IGF-I SDS (n = 75, r = 0.66, P < 0.001) or IGFBP-3 SDS (r = 0.61, P < 0.001). The reproducibility during repeated testing (19 individuals, 6 prepubertal) was best with IGFBP-3 as compared to AUCb, GHb, or IGF-I which showed considerable variability. The data suggest: 1) that IGF-I and IGFBP-3 serum levels reflect spontaneous GH secretion in healthy individuals; 2) IGF-I levels are more sensitive to GH regulation than IGFBP-3; 3) in puberty, there may be increased sensitivity of IGF-I regulation by GH as compared to the prepubertal stage; 4) short children have low IGF-I and IGFBP-3 levels, whereas tall children have high levels, a fact which is likely to contribute to the understanding of height variability in a healthy population; 5) the good reproducibility of IGFBP-3 on repeated testing makes it an interesting parameter for the evaluation of the GH-IGF-axis. IGFBP-3 measurements may substitute for GH profiles in many cases, when the goal is an estimation of the GH secretion rate.