Circadian variation of plasma cortisol in prepubertal children with normal stature, short stature and growth hormone deficiency.

OBJECTIVES

When studying the relationship between spontaneous secretion of growth hormone (GH) and cortisol in children, most studies show no correlation in mean levels of these two hormones, while others found positive or even strongly negative correlations. These contradictory results could be partly due to the inability to properly compare hormones that are characterized by circadian and ultradian variations in their secretory profiles. We aim here to study possible differences in rhythm characteristics of plasma cortisol with stature and to compare the circadian secretory patients of cortisol and GH.

PATIENTS

We analysed data from 135 prepubertal children: (1) 14 GH-deficient children; (2) 36 children with short stature (2-3 SD below their peer group mean); (3) 57 children with very short stature (3-4 SD below their peer group mean); and (4) a reference group of 28 children with normal stature (+/- 2 SD). Subjects were living in a hospital setting on a diurnal waking (07.30-22.30 h), nocturnal resting routine during sampling, consuming the usual hospital diet at fixed times.

MEASUREMENTS

Cortisol and GH concentrations were determined by radioimmunoassay in plasma obtained at about 2-3 h intervals during most of the day and at half-hour intervals between 22.00 and 02.00 h. Circadian rhythm characteristics obtained by least-squares estimation were compared between groups divided according to gender and stature with a parameter test.

RESULTS

Show a statistically significant circadian rhythm in cortisol secretion for all groups studied (P < 0.001 in all cases). A comparison of circadian parameters indicated similar characteristics between subjects of short, very short and normal stature. Despite a borderline statistically significant difference in rhythm-adjusted mean and amplitude of GH between nondeficient and GH-deficient children, there was no difference in the circadian pattern of cortisol secretion between these two groups. No correlation was found in circadian mean, amplitude, average, standard deviation, standard error, minimum or maximum between GH and cortisol for any of the groups of children.

CONCLUSIONS

Any possible relation between GH and cortisol remains unclear. Moreover, GH-deficient children are not necessarily characterized by either hyper- or hypocortisolaemia.