Adrenocortical insufficiency in critical illness

Published on 27/02/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 912 times

Chapter 53 Adrenocortical insufficiency in critical illness

The adrenal glands form an essential part of the organism’s response to stress. Hence, in intensive care, an adequate adrenal response is considered to be of prime importance. Whilst primary adrenal insufficiency (AI) is a well-recognised but rare condition in intensive care, secondary, or RAI, is thought to be more prevalent. Adrenocortical insufficiency may present as an insidious, occult disorder, unmasked by conditions of stress, or as a catastrophic syndrome that may result in death.

PHYSIOLOGY

The adrenal glands are functionally divided into medulla and cortex; the cortex is responsible for the secretion of three major classes of hormones: glucocorticoids, mineralocorticoids and sex hormones. The major pathogenic effects of disease result from cortisol and aldosterone deficiency.

Cortisol, the major glucocorticoid synthesised by the adrenal cortex, plays a pivotal role in normal metabolism. It is necessary for the synthesis of adrenergic receptors, normal immune function, wound healing and vascular tone. These actions are mediated by the glucocorticoid receptor, a member of the nuclear hormone receptor superfamily. The activated receptor migrates to the nucleus and binds to specific recognition sequences within target genes, but also interacts with numerous transcription factors and cytosolic proteins. Numerous isoforms of the glucocorticoids receptor have now been described; the alpha subtype, a 777-amino-acid polypeptide chain, was initially felt to be the primary mediator of glucocorticoid action. In contrast the beta isoform, a 742-amino-acid chain was felt to have no physiological activity. More recently it appears that the beta subtype has a negative effect on alpha-mediated gene transactivation, the physiological relevance of which remains controversial. Furthermore, additional isoforms have now been described, suggesting that glucocorticoid receptor diversity may be an important factor in understanding the complex effects of corticosteroid action.¹

Under normal circumstances, cortisol is secreted in pulses, and in a diurnal pattern.² The normal basal output of cortisol is estimated to be 15–30 mg/day, producing a peak plasma cortisol concentration of 110–520 nmol/l (4–19 μµg/dl) at 8–9 a.m., and a minimal cortisol level of < 140 nmol/l (< 5 μµg/dl) after midnight. The daily output of aldosterone is estimated to be 100–150 μg/day.

Secretion is under the control of the hypothalamic–pituitary axis. There are a variety of stimuli to secretion, including stress, tissue damage, cytokine release, hypoxia, hypotension and hypoglycaemia. These factors act upon the hypothalamus to favour the release of corticotrophin-releasing hormone (CRH) and vasopressin. CRH is synthesised in the hypothalamus and carried to the anterior pituitary in portal blood, where it stimulates the secretion of adrenocorticotrophic hormone (ACTH), which in turn stimulates the release of cortisol, mineralocorticoids (principally aldosterone) and androgens from the adrenal cortex. CRH is the major (but not the only) regulator of ACTH release and is secreted in response to a normal hypothalamic circadian regulation and various forms of ‘stress’. Vasopressin, oxytocin, angiotensin II and β-adrenergic agents also stimulate ACTH release while somatostatin, β-endorphin and enkephalin reduce it. Cortisol has a negative feedback on the hypothalamus and pituitary, inhibiting hypothalamic CRH release induced by stress, and pituitary ACTH release induced by CRH. During periods of stress, trauma or infection, there is an increase in CRH and ACTH secretion and a reduction in the negative-feedback effect, resulting in increased cortisol levels, in amounts roughly proportional to the severity of the illness.³^–⁵

The majority of circulating cortisol is bound to an α-globulin called transcortin (corticosteroid-binding globulin, CBG). At normal levels of total plasma cortisol (e.g. 375 nmol/l or 13.5 μµg/dl) less than 5% exists as free cortisol in the plasma; however it is this free fraction that is biologically active. Circulating CBG concentrations are approximately 700 nmol/l. In normal subjects CBG can bind approximately 700 nmol/l (i.e. 25 µg/dl).⁶ At levels greater than this, the increase in plasma cortisol is largelyin the unbound fraction. The affinity of CBG for synthetic corticosteroids, with the exception of prednisolone, is negligible. CBG is a substrate for elastase, a polymorphonuclear enzyme that cleaves CBG, markedly decreasing its affinity for cortisol.⁷ This enzymatic cleavage results in the liberation of free cortisol at sites of inflammation. CBG levels have been documented to fall during critical illness,⁸^–¹⁰ and these changes are postulated to increase the amount of circulating free cortisol.

Cortisol has a half-life of 70–120 minutes. It is primarily eliminated by hepatic metabolism and glomerular filtration. The excretion of free cortisol through the kidney represents 1% of the total secretion rate. Currently, available routine assays measure only total cortisol levels – bound and free.

The metabolic effects of cortisol are complex and varied. In the liver, cortisol stimulates glycogen deposition by increasing glycogen synthase and inhibiting the glycogen-mobilising enzyme glycogen phosphorylase.¹¹ Hepatic gluconeogenesis is stimulated, leading to increased blood glucose levels. Concurrently, glucose uptake by peripheral tissues is inhibited.¹² Free fatty acid release into the circulation is increased, and triglyceride levels rise.

In the circulatory system, cortisol increases blood pressure both by direct actions upon smooth muscle, and via renal mechanisms. The actions of pressor agents such as catecholamines are potentiated, whilst nitric oxide-mediated vasodilation is reduced.^13,¹⁴ Renal effects include an increase in glomerular filtration rate, and sodium transport in the proximal tubule, and sodium retention and potassium loss in the distal tubule.¹⁵

The primary effects of cortisol upon the immune system are anti-inflammatory and immunosuppressive. Lymphocyte cell counts decrease, whilst neutrophil counts rise.¹⁵ Accumulation of immunologically active cells at inflammatory sites is decreased. The production of cytokines is inhibited, an effect that is mediated via nuclear factor (NF)-κB. This occurs both by induction of NF-κB inhibitor and by direct binding of cortisol to NF-κB, thus preventing its translocation to the nucleus. Whilst the well-defined effects of cortisol upon the immune system are primarily inhibitory, it is also suggested that normal host defence function requires some cortisol secretion. Cortisol has been described as having a positive effect upon immunoglobulin synthesis, potentiation of the acute-phase response, wound healing and opsonisation.¹⁶

CLASSIFICATION

AI may be considered to be primary, secondary or relative. Primary AI, otherwise known as Addison’s disease, results from hypofunction of the adrenal cortex. Secondary AI occurs when there is suppression or absence of ACTH secretion from the anterior pituitary. RAI describes a situation of inadequate cortisol response to stress in the setting of critical illness.

PRIMARY ADRENAL INSUFFICIENCY

Primary AI or Addison’s disease is a rare disorder. In the western world its estimated prevalence is 120 per million.¹⁷ In adulthood, the commonest cause is autoimmune, but in the intensive care setting, consideration should be given to other causes of adrenal gland destruction (Table 53.1). These include infection, haemorrhage and infiltration. Tuberculosis is the commonest infective cause worldwide, but rarer infections such as histoplasmosis, coccidiomycosis and cytomegalovirus (especially in patients with human immunodeficiency virus (HIV)) have also been implicated. Haemorrhage into the glands is associated with septicaemias, particularly meningococcal (Waterhouse–Friedrichsen syndrome). Asplenia and the antiphospholipid syndrome may also be associated with adrenal haemorrhage. Adrenal gland destruction may also be secondary to infiltration with tumour, or amyloid.

Table 53.1 Causes of primary adrenal insufficiency

Drugs may impair adrenal function either by inhibiting cortisol synthesis (etomidate, ketoconazole) or by inducing hepatic cortisol metabolism (rifampicin, phenytoin). High levels of circulating cytokines are also reported to have a suppressive effect upon ACTH release.¹⁸

PRESENTATION (Table 53.2)

The disease is often unrecognised in its early stages as the presenting features are ill-defined. Symptoms include tiredness and fatigue, vomiting, weight loss, anorexia and postural hypotension. Hyperpigmentation is seen in non-exposed areas (such as palmar skin creases) and is due to the hypersecretion of melatonin, a breakdown product from the ACTH precursor pro-opiomelanocortin (POMC).

Table 53.2 Clinical features of Addison’s disease

Presentation to an intensive care physician is likely to be in the form of adrenal crisis. This may be precipitated by concurrent illness or surgery, or by failure to take replacement medication. Classically adrenal crisis will present as refractory shock with a poor response to inotropic or pressor agents. Abdominal or flank pain is often present and may lead to an erroneous diagnosis of an acute surgical abdomen.

Treatment should consist of immediate supportive measures, fluid resuscitation and high-dose intravenous glucocorticoid therapy. A standard dose would be 100 mg hydrocortisone 6-hourly, or as an infusion. At these doses separate mineralocorticoid replacement is not required.¹⁹

Adrenal crisis should be suspected in cases of undifferentiated shock not responding to standard management. Suggestive features would include a history of symptomatology consistent with the diagnosis, hyperpigmentation on examination, and demonstration of hyponatraemia, hyperkalaemia and peripheral blood eosinophilia. A random plasma total cortisol taken during a crisis will be low (below 80 nmol/l) and in the acute phase ACTH stimulation testing is not required.

SECONDARY ADRENAL INSUFFICIENCY

The commonest cause of ACTH deficiency is sudden cessation of exogenous glucocorticoid treatment. Patients who have been taking more than 30 mg/day hydrocortisone or equivalent for more than 3 weeks are at risk of adrenal suppression.²⁰ Other causes include pituitary surgery, pituitary infarction (Sheehan’s syndrome) or pituitary tumour.

Presentation is similar to that of primary AI. The major distinguishing characteristics are a lack of hyperpigmentation, and the absence of mineralocorticoid deficiency; hence hyperkalaemia is not a feature of secondary AI, although hyponatraemia may still be present due to increased vasopressin levels.

INVESTIGATION OF ADRENAL INSUFFICIENCY

In a stable patient suspected AI is routinely investigated by an ACTH stimulation test. The test is performed by administering 250 μg of a synthetic ACTH molecule comprising the first 24 amino acids – tetracosactrin (Synacthen). Plasma total cortisol is measured at 0 and 30 minutes after administration, and a normal response is defined as a cortisol measurement over 525 nmol/l.²¹ However, it should be noted that current immunoassays exhibit a significant degree of variability, and thus local laboratory reference ranges should be used.²² The test cannot be performed if the patient is currently being prescribed hydrocortisone as this will cross-react with the assays; an alternative replacement therapy such as dexamethasone should be used in these cases.

Secondary AI may be differentiated from primary by a prolonged ACTH test. This is performed by the use of a depot preparation or an intravenous infusion of tetracosactrin for 24–48 hours. Patients with secondary hypoadrenalism show a greater plasma cortisol response at 24 hours than at 4 hours; alternatively measurement of a baseline ACTH level may be used. In patients with primary AI this will be elevated.

Other biochemical tests that may be used in investigation of AI include the insulin hypoglycaemia test, the overnight metyrapone test and the CRH stimulation test. These investigations are not normally necessary in uncomplicated cases. In addition the use of the low-dose ACTH stimulation test has been advocated, in which only 1 mcg of tetracosactrin is used. This approach has not yet gained widespread acceptance.

RELATIVE ADRENAL INSUFFICIENCY

The term relative AI (RAI) was coined to describe a syndrome where the adrenal gland partially responds to stress but the magnitude of response is not commensurate.

Annane et al. prospectively studied 189 patients with septic shock and detailed a three-level classification system based upon the basal cortisol level and response to ACTH.²³ Mortality was found to be highest in those patients with a basal cortisol level above 34 μg/dl (938 nmol/l) andresponse to ACTH of less than 9 μg/dl (248 nmol/l). Patients with a basal cortisol above 34 μµg/dl but a cortisol response greater than 9 μg/dl did better, whilst the best prognosis was seen in the group with a lower basal cortisol level and high response to ACTH. This pattern – a high basal cortisol and blunted response to ACTH,²⁴^–²⁷ which is associated with increasing mortality – has now come to be referred to as RAI. The significance of this observation, however, is not clear. It has been suggested either to represent a partially suppressed adrenal axis, implying a role for cortisol replacement therapy, or alternatively as indicating an ‘overstressed’ axis, in which case steroid treatment would be inappropriate. Treatment of septic patients fulfilling RAI criteria with hydrocortisone has been shown to improve outcome in one study,²⁸ but these results have not been widely accepted (see below). Alternative diagnostic criteria, relying on baseline plasma cortisol levels without ACTH stimulation, have been proposed,²⁹ but the lack of a consistently observed relationship between cortisol levels and mortality means that the optimal diagnostic criteria for RAI remain controversial (Table 53.3). Possible explanations for the difficulties in assessing adrenal function in this patient group include spontaneous fluctuations in the measured cortisol values,³⁰ increased variability of assays,²² and changes in CBG levels affecting free cortisol values.^10,³¹

Table 53.3 Controversies in the diagnosis of adrenal insufficiency in the critically ill

Limitations of a random cortisol

In the critically ill there is a marked fluctuation in plasma cortisol concentration limiting the utility of a random cortisol ³⁰

The ‘normal’ range of cortisol in critical illness is not defined

There is no consensus ‘cut-off’ value below which adrenal insufficiency is present

Limitations of total cortisol

Free cortisol is the bioactive fraction of cortisol

There is a large variation in total cortisol assay results when the same specimen is tested in different laboratories and using different assays ²²

Peripheral tissue-specific glucocorticoid resistance is not tested

Limitations of the conventional Synacthen test

The HDSST results in plasma Synacthen concentrations that are supraphysiological

Published data may have overestimated the incidence of adrenal insufficiency, as many studies have not excluded patients who received etomidate

The low-dose SST may be more a better predictor of outcome

HDSST, high-dose short Synacthen test; SST, short Synacthen test.

STEROID THERAPY IN CRITICAL ILLNESS

Steroids have an established role in the management of a number of critical illnesses, as outlined in Table 53.4. However their use in other conditions has not been without controversy.

Table 53.4 Proven role for steroids in critical illness

SPINAL INJURY

High-dose methylprednisolone has been advocated in the management of patients with spinal cord injury following the publication of the National Acute Spinal Cord Injury Study (NASCIS) II and III trials. The major criticism of these studies is the lack of a demonstrable improvement in the primary outcome measures. For a more detailed review of the use of steroids in spinal cord injury, the reader is referred to Chapter 70.

HEAD INJURY

The role of steroids in the management of cerebral oedema secondary to tumours is well documented and accepted. However, their role in the management of head injury has been the subject of intense debate. Several prospective studies have not been able to prove any benefit with steroids in head trauma. However, these studies were not adequately powered to detect a difference. The recent Corticosteroid Randomization After Significant Head Injury (CRASH) trial, with nearly 10 000 patients, clearly demonstrated the lack of any benefit with steroids in head injury.³²

ACUTE RESPIRATORY DISTRESS SYNDROME (ARDS)

The first clinical trial to report a beneficial effect of steroids in ARDS was the one by Meduri et al. in 1998.³³ In a randomised, placebo-controlled, cross-over single-centre trial enrolling 24 patients, steroids were shown to improve lung injury score and reduce mortality. The ARDS clinical trial network published the results of its multicentre study where corticosteroids were administered to patients with ARDS persisting beyond 7 days. Although steroid use was associated with earlier ventilatory wean, improved arterial oxygenation and increased respiratory compliance, there was a higher rate of returnto assisted ventilation and neuromuscular weakness.³⁴ No overall mortality difference was demonstrable between steroid and placebo groups. Commencement of steroids more than 2 weeks after onset of ARDS led to almost a fourfold increase in mortality as compared to the placebo group. Consequently, steroids cannot be routinely recommended for persistent ARDS and may be harmful in late-stage ARDS.