In girls, but not in boys, pronounced adrenarche and precocious pubarche along with ovarian hyperandrogenism have been related to insulin resistance and reduced fetal growth. However, insulin secretion is increased during puberty in normal boys. The aim of this study was to analyze the possible implication of changes in the GH/IGF-I axis and in insulin sensitivity for the regulation of adrenal androgen secretion of normal prepubertal and adolescent boys. Fifty-six normal boys were divided into the following groups (Gr): Gr1, prepuberty (testicular volume, <4 cc; n = 33); and Gr3, puberty (testicular volume, 4-25 cc; n = 23). Gr1 was subdivided according to age into: Gr1A, early prepuberty (boys younger than 5.9 yr old; n = 16); and Gr1B, late prepuberty (prepubertal boys, 5.9 yr old or older; n = 17). Gr3 was subdivided according to testicular volume into: Gr3A, early puberty (testicular volume, 4-8 cc; n = 13); and Gr3B, late puberty (testicular volume, 10--25 cc; n = 10). To study hormonal changes during the transition between prepuberty and puberty, an additional group, Gr2 (n = 30), was defined by mixing Gr1B and Gr3A. Serum dehydroepiandrosterone sulfate (DHEAS), androstenedione (Delta(4)A), insulin, IGF-I, and glucose were determined after overnight fasting. Insulin sensitivity was estimated by the fasting glucose/insulin (G/I) ratio. There was a close correlation between fasting G/I ratio and QUICKI, a quantitative insulin sensitivity check index. Mean values for Gr1 and Gr3 as well as their subgroups were compared using t test. In Gr1, the mean fasting G/I ratio was significantly higher, and the mean serum IGF-I, serum DHEAS, and serum Delta(4)A levels were significantly lower than in Gr3 (P < 0.001). Mean fasting G/I ratios in Gr1A and Gr3A were not significantly different from those in Gr1B and Gr3B, respectively, but the fasting G/I ratio in Gr3A was significantly lower than that in Gr1B (P < 0006). Moreover, body mass index (BMI) in Gr3A was significantly higher than that in Gr1B (P < 0.01). On the other hand, mean serum IGF-I levels in Gr1A and Gr3A were significantly lower than those in Gr1B and Gr3B, respectively (P < 0.0001). The mean serum DHEAS level in Gr1A was significantly lower than that in Gr1B (P < 0.01), but no difference was found between Gr3A and Gr3B. The mean serum Delta(4)A in Gr1A was similar to that in Gr1B, but the mean serum Delta(4)A in Gr3A was significantly lower than that in Gr3B (P = 0.0001). Correlation studies within Gr1, Gr2, and Gr3 were also carried out. There was a significant positive correlation between serum DHEAS and age in Gr1 and Gr2, but not in Gr3. In Gr1, no significant correlation was found between serum DHEAS and fasting G/I ratio or between serum DHEAS and serum IGF-I, suggesting that adrenal steroidogenesis in male prepuberty is independent of insulin sensitivity or peripheral IGF-I. In Gr2, a significant negative correlation (P = 0.01) between serum DHEAS and the fasting G/I ratio was found, but not between serum DHEAS and serum IGF-I. Furthermore, a significant negative correlation between BMI and the fasting G/I ratio was also found. Therefore, changes in insulin sensitivity might be involved in adrenal androgen synthesis during the transition from prepuberty to puberty. Finally, in Gr3, DHEAS was not significantly correlated with the fasting G/I ratio or serum IGF-I. A significant negative correlation between serum Delta(4)A and the fasting G/I ratio was found in Gr2. In Gr2, but not in Gr3, there was a significant negative correlation between the fasting G/I ratio and age (P = 0.03) and between the fasting G/I ratio and serum IGF-I (P = 0.03). In conclusion, our data support the hypothesis that the GH/IGF-I axis and insulin sensitivity are not involved in the mechanism of adrenarche in boys. Insulin sensitivity and BMI, however, decrease at early puberty rather than at late puberty, and this change could be involved in modulating adrenal androgen steroidogenesis during the transition between late prepuberty and early puberty.