Rudolph's Pediatrics, 22nd Ed.

CHAPTER 536. Glucocorticoid Therapy and Withdrawal

Walter L. Miller

Glucocorticoids are used in adrenal replacement therapy when adrenal insufficiency occurs, or as pharmacologic agents. Following withdrawal of glucocorticoids used for pharmacologic therapy adrenal insufficiency can persist for months, especially during times of stress. The only differences among the various glucocorticoids are their ratio of glucocorticoid to mineralocorticoid activity, their capacity to bind to various binding proteins, their molar potency, and their biologic half-life. Dexamethasone is commonly used in reducing increased intracranial pressure and brain edema. Neurosurgical experience indicates that the optimal doses are 10 to 100 times those that would thoroughly saturate all available glucocorticoid receptors, suggesting that this action of dexamethasone may not be mediated through the glucocorticoid receptor.


Glucocorticoid replacement therapy is complicated by undesirable side effects with even minor overtreatment or undertreatment. Pediatric glucocorticoid replacement therapy is based on the endogenous secretory rate of cortisol, which may be as low as 6 mg/m2 in younger children and 9 mg/m2 in older children and adolescents, but there is considerable variation in the “normal” cortisol secretory rate among different children of the same size. Several additional factors must be considered in tailoring a specific child’s glucocorticoid replacement regimen as discussed below.


When treating autoimmune adrenalitis or other forms of Addison disease, it is prudent to err slightly on the side of undertreatment. This will eliminate the possibility of glucocorticoid-induced iatrogenic growth retardation and will permit the pituitary to continue to produce normal to slightly elevated concentrations of corticotropin (ACTH). This ACTH will continue to stimulate the remaining functional adrenal steroidogenic machinery and also provide a convenient means of monitoring therapy. By contrast, when treating severe, virilizing congenital adrenal hyperplasia, the adrenal should be suppressed more completely, as essentially all adrenal steroidogenesis will result in the production of unwanted androgens, with their consequent virilization and rate of advancement of bony maturation that is more rapid than the rate of advancement of height. However, overtreatment will also compromise growth.


Children with mild degrees of mineralocorticoid insufficiency (eg, simple virilizing congenital adrenal hyperplasia [CAH]) may continue to have mildly elevated ACTH values, suggesting insufficient glucocorticoid replacement in association with elevated plasma renin activity (PRA). In some children, the ACTH is elevated in response to chronic, compromised hypovolemia, attempting to stimulate the adrenal to produce more mineralocorticoid. In such children without overt signs of mineralocorticoid insufficiency, treatment with mineralocorticoid replacement may permit one to decrease the amount of glucocorticoid replacement needed to suppress plasma ACTH.


Very potent, long-acting glucocorticoids, such as dexamethasone or prednisone, are preferred in the treatment of adults but are rarely appropriate for replacement therapy in children. Small, incremental changes are more easily achieved with relatively weaker glucocorticoids. The usefulness of trying to mimic the physiologic diurnal variation in cortisol secretion remains controversial.


Handbooks of therapy publish tables of equivalency for the most commonly used pharmaceutical preparations of glucocorticoids based on anti-inflammatory, immunosuppressive equivalencies. A similar set of equivalencies is shown in Table 536-1. However, the differences in the plasma half-life and ability to bind to plasma proteins result in different biologic equivalencies when one assesses anti-inflammatory versus growth-suppressant equivalencies. Thus, dexamethasone is widely reported as being 25 to 30 times more potent than cortisol in its anti- inflammatory action, but the growth suppressant activity of dexamethasone is about 80 times that of cortisol.


As indicated in Table 536-1, there are 4 relevant considerations in the choice of which drug to use. First, the glucocorticoid potency of the various drugs is generally calculated and described according to the anti-inflammatory potency. Second, the growth-suppressant effect of a glucocorticoid preparation may be significantly different from its anti-inflammatory effect. Third, the mineralocorticoid activity of various glucocorticoid preparations varies widely.  Fourth, the plasma half-life and biological half-life of steroids are different variables. This is mainly related to binding to plasma proteins, hepatic metabolism, and hepatic activation.

Table 536-1. Potency of Various Therapeutic Steroids (Set Relative to the Potency of Cortisol)

Glucocorticoids are available for oral, intramuscular, intravenous, intrathecal, intra- articular, inhalant, and topical use. Each preparation is designed to deliver the maximal concentration of steroid to the desired tissue, while delivering less steroid systemically. However, all such preparations are absorbed to varying extents, so that inhalant preparations used to treat asthma can, in sufficient doses, cause growth retardation and other signs of Cushing syndrome. Most orally administered steroids are absorbed rapidly, but incompletely, whereas intramuscularly administered steroids are absorbed slowly, but completely. The efficiency of absorption of glucocorticoids varies with diet, gastric acidity, bowel transit time, and other individual factors.


When glucocorticoid therapy has been used for only 7 to 10 days, therapy can be discontinued abruptly, even if high doses have been used,8 as the hypothalamic-pituitary-adrenal axis recovers rapidly from short-term suppression. With longer therapy, recovery of hypothalamic-pituitary-adrenal function is slower, and tapered doses of glucocorticoids are indicated. Stopping therapy abruptly in such patients will lead to symptoms of glucocorticoid insufficiency without salt loss, as the renin-angiotensin system remains normal. However, blood pressure can fall abruptly, as glucocorticoids are required for the action of catecholamines in maintaining vascular tone. The most prominent symptoms of the steroid withdrawal syndrome include malaise, anorexia, headache, lethargy, nausea, and fever. Pharmacologic doses of glucocorticoids cannot be reduced precipitously to “physiologic” replacement doses, as long-term pharmacological glucocorticoid therapy suppresses glucocorticoid receptors, so that physiologic concentrations of glucocorticoids will elicit subphysiologic responses, resulting in the steroid withdrawal syndrome. Thus, it is necessary to taper gradually from the outset. The duration of glucocorticoid therapy is important: therapy for a couple of months will completely suppress the hypothalamic-pituitary-adrenal axis but will not cause adrenal atrophy; therapy of years’ duration may result in almost total atrophy of the adrenal fasciculata/reticularis, and hence may require a withdrawal regimen that takes months.

Procedures for tapering steroids are empirical. Their success is determined by the length and mode of therapy and by individual patient responses. Patients treated with alternate-day therapy can be withdrawn more easily than those receiving daily therapy, especially daily therapy with a long-acting glucocorticoid such as dexamethasone. In patients on long-standing therapy, a 25% reduction in the previous level of therapy is generally recommended weekly.

Even after the successful discontinuation of therapy, the hypothalamic-pituitary-adrenal axis is not wholly normal. Just as in the patient successfully treated for Cushing disease, the hypothalamic-pituitary-adrenal axis may be incapable of responding to severe stress for 6 to 12 months after successful withdrawal from long-term, high-dose glucocorticoid therapy. Thus, evaluation of the hypothalamus and pituitary by a CRF or metyrapone test, and evaluation of adrenal responsiveness to pituitary stimulation with an intravenous ACTH test, should be done at the conclusion of a withdrawal program and 6 months thereafter. The results of these tests will indicate if there is a need for “steroid coverage” in acute surgical stress or illness.

The cortisol secretory rate increases significantly during physiologic stress such as trauma, surgery, or severe illness. Patients receiving glucocorticoid replacement therapy or those recently withdrawn from pharmacologic therapy need coverage with “stress doses” of steroids in such situations.  Although it remains appropriate and necessary to give about 3 times physiologic requirements during such periods of stress, it is probably not necessary to give much higher doses. Similarly, it is not necessary to triple a child’s physiologic replacement regimen during simple colds, upper respiratory infection, otitis media, or after immunizations, unless there is fever.


Mineralocorticoid replacement therapy is indicated in salt-losing congenital adrenal hyperplasia and in syndromes of adrenal insufficiency that affect the zona glomerulosa. Only 1 mineralocorticoid, 9α-fluorocortisol (fludrocortisone) is currently available. There is no parenteral mineralocorticoid preparation so that hydrocortisone plus salt must be used.

Mineralocorticoid doses used are essentially the same irrespective of the size or age of the patient. In fact, newborns are quite insensitive to mineralocorticoids and may require larger doses than adults. In older children and adults, the replacement dose of 9α-fluorocortisol is usually 0.05 to 0.10 mg daily; infants typically require about 0.2 mg. Sodium must be available to the nephrons for mineralocorticoids to promote re-absorption of sodium. Thus, the newborn with salt-losing congenital adrenal hyperplasia must be treated with both mineralocorticoids and salt. Similarly, mineralocorticoids will cause hypertension only by retaining sodium.

Cortisol has significant mineralocorticoid activity: approximately 20 mg of cortisol or cortisone intravenously has a mineralocorticoid action equivalent to 0.1 mg of 9α-fluorocortisol. Thus, when cortisol or cortisone is given in stress doses, they provide adequate mineralocorticoid activity, and mineralocorticoid replacement can be interrupted.