Adrenal Suppression<i>Versus</i>Clinical Glucocorticoid Deficiency in the Premature Infant: No Simple Answers

In a recent issue of the journal, Karlsson et al. (1) reported their results of a study on the adrenal function of mostly premature infants by means of a low dose (1 mg/1.73 m) ACTH test and a standard, human (h) CRH (CRH) test (1 mg/kg). Infants studied varied in gestational age, exposure to antenatal glucocorticoid therapy, quantity of dexamethasone received postnatally, interval between the last dose of dexamethasone and the time of testing, and presence or not of concurrent hydrocortisone replacement. They concluded that the hCRH test was superior to the low dose ACTH test in detecting adrenal suppression on the basis of being abnormal more frequently than the latter in the same population of patients. By definition, adrenal suppression denotes sustained deficiency of hypothalamic-pituitary-adrenal (HPA) axis function ensuing exposure of the organism to excessive and prolonged administration of exogenous glucocorticoids (2). The mechanism is suprapituitary, and the condition has no real rodent model equivalent, as adrenal suppression in rats has a very limited duration of 1–2 days (3). Generally, clinically important adrenal suppression in children and adults occurs after 2 or more weeks of glucocorticoid therapy, depends on the type of and the schedule of glucocorticoid administration and lasts from days, to weeks, months or, rarely, infinity (2). The susceptibility of human beings to the development of adrenal suppression varies widely, as does the sensitivity of their different tissues to glucocorticoids (4). When one thinks of glucocorticoid suppression, one should clarify an important issue. The HPA axis, through several feedback control loops that influence its regulatory centers in the brain, is constantly attempting to maintain the day-to-day exposure of tissues to cortisol normal. This results in an intraindividual variability of HPA axis activity that is quite narrow, in contrast to the wide interindividual variability of this system. This means that decreased cortisol secretion a day or two after administration of a large dose of glucocorticoid is not but a corrective action of the organism. Such a change would not qualify as adrenal suppression, which is the sustained, risky glucocorticoid deficiency that ensues prolonged, excessive exposure to exogenous or endogenous glucocorticoids (2). The latter has significant morbidity, ranging from hypoglycemia to hypotension, as well as mortality, usually as a result of circulatory collapsus. The third trimester fetus and the premature and full-term infants have a functional HPA axis that responds to the negative feedback by glucocorticoids and can develop clinically significant adrenal suppression (5–7). Our ability to recognize adrenal suppression in infants is a prerequisite to decide whether or not to administer daily and stress-appropriate glucocorticoid coverage. In a full-term infant in which suppression is not complicated by concurrent systemic inflammation, one could employ the tests and criteria established for older children and adults. The rapid standard ACTH test (250 mg/1.75 m BSA) would be the procedure of choice in such an instance, and the criterion of a cortisol response greater than the value of 1.96 sd below the normal child/adult mean at 30 or 60 min would suggest no need for coverage. In the uncomplicated premature infant, a 40% lowering of these limits has been proposed because of a relative hypoactivity of the premie HPA axis when compared with that of full-term neonates (6). An insulin tolerance or a metyrapone test should not be used in these frail populations, as they provide no advantage over the standard ACTH test and might have serious complications. There is very little experience with the low dose ACTH test in infants and its clinical utility in children and adults remains controversial. Although, admittedly, it may be more sensitive than the standard ACTH test in detecting subtler abnormalities of the HPA axis, its true pertinence in helping treat patients that would not be treated on the basis of a standard ACTH test is uncertain. The same is true for the human or ovine CRH tests. Both appear to be more sensitive than the standard ACTH test in detecting subtle HPA axis abnormalities, but again, are such mild disturbances clinically relevant to dictate medical intervention? The study of Karlsson et al. (1) shows that the hCRH test is more sensitive than the low dose ACTH test but the question is: will such data change our approach to treatment? The usefulness of the standard ACTH test goes beyond detecting adrenal suppression in the premature infant after the remission of a disease for which glucocorticoid therapy was administered in the first place. Infants with the systemic inflammation that accompanies the respiratory distress synReceived October 30, 2000. Accepted October 30, 2000. Address correspondence and requests for reprints to: George Chrousos, M.D., National Institutes of Health, 10 Center Drive, Building 10, Room 9D42, Bethesda, Maryland 20892-1583. 0021-972X/01/$03.00/0 Vol. 86, No. 2 The Journal of Clinical Endocrinology & Metabolism Printed in U.S.A. Copyright © 2001 by The Endocrine Society