Chapter 33. Adrenal Insufficiency

Adrenal gland function has provided great academic material for investigation and controversy. It was not until 1937 that 17-hydroxy-11-dehydrocorticosterone, or cortisone, was isolated by Reichstein from the adrenal cortex. By 1947, synthetic cortisone was developed. At the same time, in addition to the use of cortisone for Addison's disease, this compound was found to have a therapeutic effect in patients with rheumatoid pain through its ability to inhibit stress and inflammation.

Adrenal insufficiency presents as chronic primary (about 5 per million incidence) or secondary (about 200 per million incidence). Both of these entities are more common in women, and diagnosis peaks anywhere from the fourth to the sixth decade. Historically, the most common cause of adrenal insufficiency was tuberculous adrenalitis. In developed countries, autoimmune adrenalitis has become a much more common cause of adrenal insufficiency, as tuberculous adrenalitis still plays a major role in the disease in developing countries.1

The adrenal gland has two anatomic divisions. The medulla secretes the catecholamines epinephrine and norepinephrine, and the cortex produces mineralocorticoids (via the renin–angiotensin system) and glucocorticoids. Critical illness and stress activate the hypothalamic–pituitary–adrenal (HPA) axis and stimulate the release of hypothalamic corticotropin-releasing hormone (CRH) and pituitary adrenocorticotropic hormone (ACTH).2 ACTH secretion reaches its effector organ, the adrenal cortex, where it stimulates the synthesis and secretion of glucocorticoids, mineralocorticoids, and adrenal androgens. The mechanisms regulating ACTH secretion during stress are multifactorial, with the stimulatory effect of CRH and the inhibitory influence of cortisol. The “closed-loop” negative feedback of cortisol to the HPA axis acts to suppress the secretion of CRH, ACTH, and cortisol itself (Figure 33-1).

Physiologic ACTH and cortisol secretion have a diurnal pattern with nadirs at 10 pm and 2 am, and peak at 8 am. During infection and inflammatory states, cortisol levels are increased through stimulation of the hypothalamus and pituitary by cytokines and a reduction in the negative feedback loop. The diurnal variation of cortisol secretion is lost, and resources are shifted away from mineralocorticoid and androgen production toward corticosteroid production. ACTH release can also be increased by the influence of endorphinergic pathways and from the acute (but not chronic) administration of morphine. Even with the negative feedback loop in place, during periods of high stress (after major surgery, septic shock), the adrenal cortex is also influenced directly by paracrine pathways, endothelin, atrial natriuretic peptide, or cytokines.

The adrenocortical response to stress has several mechanisms. Cortisol is 90% bound to cortisol-binding globulin, with less than 10% in the free bioavailable form. This cortisol-binding globulin is downregulated by as much as 50% during acute illness, particularly sepsis, making more cortisol available in the free form. Cortisol has been shown to upregulate intracellular glucocorticoid receptors through a positive feedback mechanism. Glucocorticoid receptors have also shown to increase their level of binding activity in skeletal muscle.

All of the above mechanisms allow for glucocorticoid production to enable physiologic compensation in periods of acute stress. Glucocorticoids increase blood glucose levels via hepatic gluconeogenesis, and inhibit adipose tissue glucose uptake. They stimulate free fatty acid and amino acid release as well as increase proteolysis to supply energy and substrate for stress response.

Glucocorticoids contribute toward the synthesis of catecholamines, allowing for maintenance of cardiac contractility, vascular tone, and blood pressure. They also decrease nitric oxide and prostaglandin production, with the result of maintaining hemodynamic stability. Glucocorticoids also have anti-inflammatory and immunosuppressive qualities through their downregulating influence on lymphocytes, natural killer cells, monocytes, macrophages, eosinophils, neutrophils, mast cells, and basophils.

In spite of the beneficial actions of cytokines and cellular mediators in an acute stress response, there is evidence that these cytokines and mediators can also have the opposite effect, resulting in decreased ACTH production, impaired corticosteroid production, and an increase in cortisol half-life, which may represent decreases in the number, expression, and function of glucocorticoid receptors. In short, mediators released in the septic patient can have a positive or negative effect on adrenal response and the net effect may depend on timing, severity of illness, and/or extent of mediator production.

In primary adrenal insufficiency, the adrenal gland fails to produce cortisol. In addition to autoimmune and infectious etiologies, this can be caused by bilateral adrenal hemorrhage, metastases, sarcoidosis, amyloidosis, adrenalectomy (such as in the case of resistant Cushing's syndrome), acquired immune deficiency syndrome (AIDS), antiphospholipid syndrome, or medication-induced effects (such as antineoplastics, etomidate, ketoconazole, and mifepristone).3 Secondary adrenal insufficiency, in which the pituitary gland produces insufficient ACTH, is typically caused by a regional tumor, autoimmune causes, genetic mutations, pituitary apoplexy postpartum (Sheehan's syndrome), head trauma, or chronic exogenous glucocorticoid administration.

A life-threatening episode of acute adrenal insufficiency typically manifests in severe hypotension, acute abdominal pain, vomiting, and fever. Hypoglycemic seizures may present in pediatrics or recurrent hypoglycemia in patients with type I diabetes. The common clinical signs of chronic adrenal insufficiency include fatigue, energy loss, reduced muscle strength, increased irritability, weight loss, nausea, and anorexia. Primary adrenal failure will likely result in hyperpigmentation, due to the stimulation of melanocytes by ACTH, whereas secondary adrenal failure will manifest with pale skin color. Mineralocorticoid deficiency may also be present, resulting in hyponatremia, hyperkalemia, dehydration, hypovolemia, hypotension, and prerenal failure.

The holy grail of adrenal insufficiency in critically ill patients is the question of how to assess adrenal function. Notwithstanding the myriad of effects that can be affected by cytokines and mediators on the HPA axis, we are so far unable to test end-organ effects of cortisol, so the diagnosis is commonly based on serum cortisol levels. This has resulted in a variety of studies and beliefs about the most accurate method to assess the serum cortisol level and its subsequent clinical implications.4 Many centers utilize the cosyntropin stimulation test, by administering 250 mcg of synthetic corticotropin and checking the serum level of cortisol before, and 30 and 60 minutes after injection. A level below 18–20 mcg/dL, or increase less than 9 mcg/dL, is considered diagnostic of adrenal insufficiency. Some centers believe that a lower dose (1 mcg) of cosyntropin is more sensitive and specific.

It is important to recognize that the increase in cortisol following a stimulation test is indicative of reserve levels of cortisol, as opposed to adrenal function. The best way to determine whether or not the HPA axis is functioning adequately is to test the entire axis, which, in the case of severely stressed critically ill patients, is already taking place via the stressors of hypotension, hypoxemia, fever, and hypoglycemia. Therefore, a random cortisol level in the face of critical illness should provide adequate information on adrenal insufficiency. A random cortisol level greater than 25 mcg/dL is considered to indicate adequate HPA axis functioning, based on the fact that patients with trauma, surgery, and critical illness have been found to have cortisol levels in the 30–50 mcg/dL range that can last for a week. It is also worth noting that critically ill patients lose the diurnal nature of cortisol secretion, so timing of a random level should not be an issue.

Adrenal insufficiency can manifest in critically ill patients up to an incidence of 77%, depending on the criteria. Previous steroid use, depending on the dose and duration of use, may contribute to suppression of the HPA axis, especially if used for more than 30 days. Immunosuppression and other infections, while not necessarily causing primary adrenal insufficiency in the outpatient setting, have become the most significant etiology of such in the critically ill patient. It is also noteworthy that patients with sepsis and systemic inflammatory response syndrome (SIRS) criteria commonly manifest primary adrenal failure, consisting of suppression of the HPA axis as well as glucocorticoid receptor expression. This has been shown to be reversible on resolution of the septic episode.

There is general agreement that adrenal insufficiency in the critically ill patient portends a higher mortality, and that some level of steroid treatment improves outcome. The difficult questions to clarify become how to assess for adrenal insufficiency, at what cutoff level does the patient require steroids, and what dosing should be used. The current recommendations by the Surviving Sepsis Campaign are to administer intravenous corticosteroids (hydrocortisone 200–300 mg per day, for 7 days in three or four divided doses or by continuous infusion) for adult septic shock patients after blood pressure is identified to be poorly responsive to fluid resuscitation and vasopressor therapy.5 It is also recommended that each intensive care unit (ICU) have a standardized protocol for deciding when it is appropriate to start steroids and how to administer them.

It is also worth noting that there are many drug–drug interactions that may occur with steroid administration. Glucocorticoids can decrease the drug blood levels of aspirin, warfarin, insulin, isoniazid, and oral hypoglycemic agents, but they can increase the levels of cyclophosphamide and cyclosporine. Drugs that can increase the level of glucocorticoid blood concentration include antacids, carbamazepine, cholestyramine, colestipol, ephedrine, mitotane, phenobarbital, phenytoin, and rifampin, while cyclosporine, erythromycin, oral contraceptives, and troleandomycin increase them.

Steroid administration should be tapered as the critically ill patient's clinical picture improves, so as to avoid hemodynamic and immunologic rebound. Hydrocortisone is the corticosteroid of choice, since most studies were performed using this formulation, and it is the closest in physiologic characteristics to cortisone. Hydrocortisone also has mineralocorticoid activity, which must be considered and replaced if other glucocorticoids are utilized.

In summary, critical illness–related corticosteroid insufficiency (CIRCI) can be a significant factor in the progress of the critically ill patient, but many questions remain unanswered at this time. Table 33-1 lists the 12 recommendations from the American College of Critical Care Medicine.6 It is important to have a protocolized practice in the care of these patients and to participate in and follow in the research advances in this very important topic.