Hydrocortisone replacement therapy in children and adolescents with hypopituitarism

Appropriate replacement doses of glucocorticoid are important to determine in primary and secondary adrenal deficiency in children, both to avoid the risks of hypoglycaemia and adrenal crisis associated with undertreatment, and to avoid growth suppression and reduced final height potential associated with steroid excess. The aim of this study was to assess how closely conventional twice daily hydrocortisone administration mimics physiological cortisol secretion in a group of ACTH-deficient children and adolescents.

Fifty children and adolescents (aged 3-20 years) were studied who had had surgery +/- radiotherapy to the hypothalamopituitary region for removal of a craniopharyngioma. The patients were subdivided into two groups: group I comprised 44 patients known to be ACTH deficient (as determined by glucagon or insulin provocation tests of anterior pituitary function performed after surgery) and maintained on twice daily oral hydrocortisone replacement; group II comprised six patients known to be ACTH sufficient at their last assessment of pituitary function and not on hydrocortisone replacement. A third group of 10 boys (aged 7-13 years) who had no known endocrinopathy were used as controls (group III).

After intravenous cannula insertion, blood samples were taken every 2h for measurement of plasma cortisol and glucose over a period of 24h. Patients in group I continued on their usual doses of hydrocortisone, prescribed at 0800 and 1800 h.

The mean total daily replacement dose of hydrocortisone for patients in group I was 12.3 mg/m2/ day (range, 5.5-18.5). On the conventional twice daily dose regimen, there was a supraphysiological medium plasma cortisol level (629 nmol/l, range 185-1600; z = -3.76, P = 0.0002) 2 h after the morning dose relative to the control group, and a prolonged and unphysiological nadir from 1400-1800 h (median at 1600 h 42 nmol/l, range 13-1170; z = -3.13, P < 0.002) before the second dose of hydrocortisone was administered. Cortisol values were low, and often negligible, during the early hours of the morning (median at 0600 h 15 nmol/l, range 13-277, z = -4.87, P < 0.00001) and spontaneous hypoglycaemia was documented in one patient on a single 0800 h sample. One patient in group II was shown to be unequivocally cortisol deficient and median cortisol values for the remaining five suggested a suboptimal rise in plasma cortisol during the early hours of the morning.

Our cohort of patients provides an excellent model for the study of glucocorticoid replacement in cortisol-deficient children and adolescents and shows, as in adults, that the aim of mimicking the physiological nyctohemeral secretion of cortisol is difficult to achieve in practice and raises a number of important considerations unique to steroid substitution therapy in this age group.