Detailed data on the physiological pattern of adrenocortical activity during normal growth are lacking. An established method to determine adrenocortical glucocorticoid secretion is the measurement of 24-h excretion rates of major urinary cortisol metabolites (C21). To test the hypothesis that the frequently reported higher cortisol secretion in men than in women develops during puberty, we examined C21 together with excretions of combined urinary free and conjugated cortisol (F(comb)) in 400 healthy boys and girls aged 3-18 yr using GC-MS. Daily excretion rates of C21, F(comb), and body surface area (BSA)-corrected F(comb) significantly increased with age in both sexes. In contrast, C21/BSA (microgxm(-2).day(-1)) declined from the age of 3-4 yr to 7-8 yr in boys and girls (P < 0.01; e.g., in boys: from 3,991 +/- 1,167 to 3,193 +/- 804), then increased in both sexes, and finally became discordant after the age of 11-12 yr with a further rise in males only (17- to 18-yr-olds: boys, 5,275 +/- 1,414; girls 3,939 +/- 1,586, P < 0.01). This pattern was associated with the occurrence of a lower index for 5alpha-reductase activity (allotetrahydrocortisol/tetrahydrocortisol) in females compared with males. Our results demonstrate dynamic changes in adrenocortical activity in healthy children resulting in an emerging sexual dimorphism in cortisol secretion after age 11. The latter can be explained, at least partly, by diverging 5alpha-reductase activities in boys and girls. F(comb), a frequently analyzed GC-MS parameter, proved not to reflect dynamic changes in cortisol secretion. In conclusion, the varying metabolic need for cortisol during normal growth may have implications for future improvements in glucocorticoid replacement therapy.