5.9 Adrenal insufficiency

In 1855, Thomas Addison identified a clinical syndrome characterized by wasting and hyperpigmentation as the result of adrenal gland destruction (1). This landmark observation paved the way for progress in understanding and treating adrenal insufficiency, with the introduction of adrenal extracts for treatment of Addison’s disease by the groups of Hartman and Pfiffner in 1929. However, long-term survival of patients with adrenal insufficiency only became possible after the seminal work of Edward Kendall, Philip Hench, and Tadeus Reichstein on the characterization and therapeutic use of cortisone. In 1946, Lewis Sarrett, a Merck scientist, achieved a partial synthesis of cortisone, which marked the beginning of industrial-scale production of cortisone. In 1948, in a fundamental clinical experiment at the Mayo Clinic, the first patient with Addison’s received intravenous injections of Kendall’s Compound E, cortisone, resulting in ‘notable improvement of his condition’. This was followed by the groundbreaking trials on the use of cortisone in rheumatoid arthritis yielding unanticipated clinical improvements, which quickly led to the labelling of cortisone as ‘the wonder drug’. In November 1950, cortisone was made available to all physicians in the USA, a rapid translational development process, which culminated in the award of the 1950 Nobel Prize in Medicine to Kendall, Hench, and Reichstein. This progress reached other countries with variable delay and widespread availability of cortisone in the UK was achieved by joint efforts of Glaxo and the Medical Research Council. Though almost 150 years have passed since Addison’s landmark observations and 60 years since the introduction of life-saving cortisone, there are still advances and challenges in the management of adrenal insufficiency, summarized in this chapter.

When extracting steroids from the adrenal Kendall and Reichstein identified 28 separate steroids and today we classify the steroids produced by the adrenal glands, the corticosteroids, in three major classes—glucocorticoids (cortisol, corticosterone), mineralocorticoids (aldosterone, deoxycorticosterone), and adrenal sex steroid precursors (dehydroepiandrosterone (DHEA), androstenedione).

Cholesterol is the precursor for all adrenal steroidogenesis. The principal source of cholesterol is provided from the circulation in the form of low-density lipoprotein (LDL) cholesterol. Uptake is by specific cell-surface LDL receptors present on adrenal tissue; LDL is then internalized via receptor-mediated endocytosis, the resulting vesicles fuse with lysozymes, and free cholesterol is produced following hydrolysis. However, it is clear that this cannot be the sole source of adrenal cholesterol as patients with abetalipoproteinaemia, who have undetectable circulating LDL, and patients with defective LDL receptors in the setting of familial hypercholesterolaemia still have normal basal adrenal steroidogenesis. Cholesterol can be generated de novo within the adrenal cortex from acetyl coenzyme A. In addition, there is evidence that the adrenal can utilize high-density lipoprotein (HDL) cholesterol following uptake through the HDL receptor, scavenger receptor.

The biochemical pathways involved in adrenal steroidogenesis start with the rate-limiting step of the transport of intracellular cholesterol from the outer to the inner mitochondrial membrane. Within the mitochondrion cholesterol is then converted to pregnenolone by the cholesterol side chain cleavage enzyme, cytochrome P450scc (CYP11A1). The rapid transport of cholesterol into the mitochondria is importantly facilitated by steroidogenic acute regulatory protein, which is induced by an increase in intracellular cAMP following binding of ACTH to its receptor.

Steroidogenesis involves the concerted action of several enzymes, including a series of cytochrome P450 (CYP) enzymes (for schematic overview, see Chapter 5.11). CYP11A1 and the CYP11B1 and CYP11B2 enzymes are localized to the mitochondria and require an electron shuttle system—provided through adrenodoxin/adrenodoxin reductase—for functional activity. Other CYP enzymes involved in steroidogenesis, namely 17α-hydroxylase (CYP17A1) and 21-hydroxylase (CYP21A2), are localized to the microsomal/endoplasmic reticulum fraction and depend on electron transfer from NADPH via the electron donor enzyme P450 oxidoreductase (POR).

After the uptake of cholesterol to the mitochondrion cleavage of cholesterol forms pregnenolone, which is converted in the cytoplasm to progesterone by the type II isoenzyme of 3β-hydroxysteroid dehydrogenase. Progesterone is hydroxylated to 17OH-progesterone through the activity of 17α-hydroxylase. 17-hydroxylation is an essential prerequisite for glucocorticoid synthesis CYP17 also possesses 17,20 lyase activity, which crucially facilitates the synthesis of the sex steroid precursor DHEA, a reaction that also requires allosteric interaction of the flavoprotein cytochrome b5 with both CYP17A1 and POR. In humans, 17-OH progesterone is not an efficient substrate for CYP17, and there is negligible conversion of 17-OH progesterone to androstenedione. Adrenal androstenedione secretion is dependent upon the conversion of dehydroepiandrosterone to androstenedione by 3β-hydroxysteroid dehydrogenase (3β-HSD). 21-hydroxylation of either progesterone (zona glomerulosa) or 17-OH-progesterone (zona fasciculata) is carried out by 21-hydroxylase (CYP21A2) to yield deoxycorticosterone or 11-deoxycortisol, respectively The final step in cortisol biosynthesis takes place in the mitochondria and involves the conversion of 11-deoxycortisol to cortisol by the enzyme CYP11B1, 11β-hydroxylase. In the zona glomerulosa, 11β-hydroxylase may also convert deoxycorticosterone to corticosterone. However, the enzyme CYP11B2, or aldosterone synthase, may also carry out this reaction and, in addition, is required for the conversion of corticosterone to aldosterone via the intermediate 18-OH corticosterone. Thus CYP11B2 can carry out 11β-hydroxylation, 18-hydroxylation, and 18-methyl oxidation to yield the characteristic C11–18 hemiacetyl structure of aldosterone.

Cortisol is inactivated to cortisone by action of the enzyme 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2) mainly in the kidney, while the opposite reaction, activation of cortisone to cortisol, is carried out by 11β-HSD1 mainly in the liver (Fig. 5.9.1). However, both enzymes are expressed in many tissues and recent years have highlighted the important role of this system in the tissue-specific activation and inactivation of glucocorticoids. Without the action of hepatic 11β-HSD1 Kendall would have observed no activity of his ‘Compound E’, as cortisone does not bind the glucocorticoid receptor and conversion to cortisol is a mandatory requirement for biological activity.

Glucocorticoids are secreted in relatively high amounts (cortisol 10–20 mg/day) from the zona fasciculata, whilst mineralocorticoids are secreted in low amounts (aldosterone 100–150 μg/day) from the zona glomerulosa. The adrenal androgen precursors DHEA, its sulphate ester DHEAS, and androstenedione are produced in the adrenal zona reticularis and represent the most abundant steroids secreted by the adult adrenal gland (>20 mg/day). In each case this is facilitated through the expression of steroidogenic enzymes in a specific ‘zonal’ manner. The zona glomerulosa cannot synthesize cortisol because it does not express 17α-hydroxylase. In contrast, aldosterone secretion is confined to the outer zona glomerulosa through the restricted expression of CYP11B2. Although CYP11B1 and CYP11B2 share 95% homology, the 5′ promoter sequences differ and permit regulation of the final steps in glucocorticoid and mineralocorticoid biosynthesis by ACTH and angiotensin II, respectively. DHEA is sulphated in the zona reticularis by the DHEA sulphotransferase (SULT2A1) to form DHEAS.

Classical endocrine feedback loops are in place to control the secretion of both hormones—cortisol inhibits the secretion of both corticotrophin releasing factor and ACTH from the hypothalamus and pituitary, respectively, and the aldosterone-induced sodium retention inhibits renal renin secretion (Fig. 5.9.2).

Glucocorticoid synthesis is under negative feedback control of the hypothalamic–pituitary–adrenal (HPA) axis (Fig. 5.9.2a). Adrenocorticotropic hormone (ACTH) secretion from the anterior pituitary is stimulated by hypothalamic corticotrophin-releasing hormone (CRH) following a circadian rhythm with a peak around 3.00 to 4.00 hours. Other major effectors on CRH secretions are various forms of stress, including hypoglycaemia, hypotension, fever, trauma, and surgery. ACTH binds to its receptor (melanocortin receptor 2, MC2R) on the adrenal cell and stimulates import of cholesterol into the mitochondrion by steroidogenic acute regulatory protein. In parallel, transcription of genes encoding steroidogenic enzymes and proteins of the electron transfer shuttle is increased.

Mineralocorticoid synthesis is mainly under the control of the renin–angiotensin–aldosterone system (RAAS) and a potassium feedback loop (Fig. 5.9.2b). A variety of factors stimulate renin secretion from renal juxtaglomerula cells, with renal perfusion being the most important regulator. Several other stimulators (β-adrenergic stimulation, prostaglandins) and inhibitors (α-adrenergic stimulation, dopamine, atrial natriuretic peptides, angiotensin II) are known. Angiotensinogen is an α₂-globulin synthesized within the liver which is cleaved by renin to form angiotensin I. Angiotensin I is converted to angiotensin II by angiotensin-converting enzyme in the lung and many other peripheral tissues. Angiotensin I has no apparent biological activity but angiotensin II is a potent stimulator of aldosterone secretion. In addition, angiotensin II acts is a potent vasconstrictor. The rate-limiting step in the RAAS is the secretion of renin, which is also controlled through a negative feedback loop. Renin is secreted from juxtaglomerular epithelial cells within the macula densa of the renal tubule in response to underlying renal arteriolar pressure, oncotic pressure, and sympathetic drive. Thus low perfusion pressure and/or low tubular fluid sodium content, as seen in haemorrhage, renal artery stenosis, dehydration, or salt loss, increase renin secretion. Conversely, secretion is suppressed following a high salt diet and by factors that increase blood pressure. Hypokalaemia increases and hyperkalaemia decreases renin secretion; in addition, potassium exerts a direct effect upon the adrenal cortex to increase aldosterone secretion. Angiotensin II and potassium stimulate aldosterone secretion principally by increasing the transcription of CYP11B2 through common intracellular signalling pathways. The potassium effect is mediated through membrane depolarization and opening of calcium channels, and the angiotensin II effect following binding of angiotensin II to the surface AT₁ receptor and activation of phospholipase C.

The separate control of glucocorticoid biosynthesis through the HPA axis and mineralocorticoid synthesis via the renin–angiotensin system has important clinical consequences. Patients with primary adrenal failure invariably have both cortisol and aldosterone deficiency, whereas patients with ACTH deficiency due to pituitary disease have glucocorticoid deficiency, but aldosterone concentrations are normal because the renin–angiotensin system is intact.

Both cortisol and aldosterone exert their effects following uptake of free hormone from the circulation and binding to intracellular receptors, termed the glucocorticoid and mineralocorticoid receptors (GR, MR). These are both members of the thyroid/steroid hormone receptor superfamily of transcription factors, comprising a C-terminal ligand binding domain, a central DNA binding domain, interacting with specific DNA sequences on target genes, and an N-terminal hypervariable region. In both cases, although there is only a single gene encoding the GR and MR, splice variants have been described resulting in α and β variants.

The binding of glucocorticoid to the GR-α in the cytosol results in activation of the steroid–receptor complex through a process which involves the dissociation of heat-shock proteins HSP 90 and HSP 70. Following translocation to the nucleus, gene transcription is stimulated or repressed following binding of dimerized GR–ligand complexes to specific DNA sequences (glucocorticoid-response element) in the promoter regions of target genes. The GR-β variant may act as a dominant negative regulator of GR-α transactivation.

In contrast to the diverse actions of glucocorticoids, mineralocorticoids have a more restricted role, principally to stimulate epithelial sodium transport in the distal nephron, distal colon, and salivary glands. This is mediated through the induction of the apical sodium channel (comprising three subunits α, β, and γ) and the α₁ and β₁ subunits of the basolateral Na⁺K⁺ATPase through transcriptional regulation of a specific aldosterone-induced gene that encodes serum and glucocorticoid-induced kinase. Aldosterone binds to the MR, principally in the cytosol (though there is evidence for expression of the unoccupied MR in the nucleus) followed by translocation of the hormone–receptor complex to the nucleus.

The MR and GR share considerable homology—57% in the steroid binding domain and 94% in the DNA binding domain. It is perhaps not surprising therefore that there is promiscuity of ligand binding with aldosterone binding to the GR and cortisol binding to the MR. For the MR this is particularly impressive—in vitro the MR has the same inherent affinity for aldosterone, corticosterone, and cortisol. Specificity upon the MR is conferred through the ‘prereceptor’ metabolism of cortisol via the enzyme 11β-HSD2, which inactivates cortisol and corticosterone to inactive 11-keto metabolites, enabling aldosterone to bind to the MR.

For both glucocorticoids and mineralocorticoids there is accumulating evidence for so-called ‘nongenomic’ effects involving hormone response obviating the genomic GR or MR effects. A series of responses have been reported within seconds/minutes of exposure to corticosteroids and are thought to be mediated by, as yet uncharacterized, membrane coupled receptors.

Over 90% of circulating cortisol is bound, predominantly to the α₂-globulin cortisol-binding globulin (CBG). This 383-amino acid protein is synthesized in the liver and binds cortisol with high affinity. Affinity for synthetic corticosteroids (except prednisolone, which has an affinity for CBG of approximately 50% of that of cortisol) is negligible. Circulating CBG concentrations are approximately 700 nmol/l; levels are increased by oestrogens and in some patients with chronic active hepatitis but reduced in patients with cirrhosis, nephrosis, and hyperthyroidism. The oestrogen effect can be marked, with levels increasing two- to threefold across pregnancy, and this should also be taken into account when measuring plasma ‘total’ cortisol in pregnancy and in women taking oestrogens. Inherited abnormalities in CBG synthesis are much rarer than those described for thyroid-binding globulin but include patients with elevated CBG, partial and complete deficiency of CBG, or CBG variants with reduced affinity for cortisol. In each case, alterations in CBG concentrations change total circulating cortisol concentrations accordingly but ‘free’ cortisol concentrations are normal. Only this free circulating fraction is available for transport into tissues for biological activity. The excretion of ‘free’ cortisol through the kidneys is termed urinary free cortisol and represents only 1% of the total cortisol secretion rate. Approximately 50% of secreted cortisol appears in the urine as Tetrahydrocortisol (THF), 5alpha-tetrahydrocortisol (allo-THF), and tetrahydrocortisone (THE), 25% as cortols/cortolones, 10% as C19 steroids, and 10% as cortolic/cortolonic acids.

Aldosterone is also metabolized in the liver and kidneys. In the liver it undergoes tetrahydro reduction and is excreted in the urine as a 3-glucuronide tetrahydroaldosterone derivative. However, glucuronide conjugation at the 18 position occurs directly in the kidney, as does 3α and 5α/5β metabolism of the free steroid. Because of the aldehyde group at the C18 position, aldosterone is not metabolized by 11β-HSD2. Hepatic aldosterone clearance is reduced in patients with cirrhosis, ascites, and severe congestive heart failure.

The prevalence of Addison’s disease, mostly due to autoimmune adrenalitis, is 93–140 per million while secondary insufficiency, mostly due to hypothalamic–pituitary tumours, has a prevalence of 125–280 per million (2). The overall prevalence of adrenal insufficiency is 5 in 10 000 population, with three patients suffering from secondary adrenal insufficiency, one from primary adrenal insufficiency due to autoimmune adrenalitis, and one from congenital adrenal hyperplasia.

According to recent studies, chronic primary adrenal insufficiency has a prevalence of 93 to 140 per million and an incidence of 4.7 to 6.2 per million in Caucasian populations (2, 3). These numbers are considerably higher than reported earlier, despite a continuous decline in tuberculous adrenalitis in the developed world, and suggest an increasing incidence of autoimmune adrenalitis. The age at diagnosis peaks in the fourth decade of life, with women more frequently affected.

Secondary adrenal insufficiency has an estimated prevalence of 150 to 280 per million (2, 3). Again, women are more frequently affected and age at diagnosis peaks in the sixth decade.

It has been suggested that therapeutic glucocorticoid administration is the most common cause of adrenal insufficiency, as exogenous glucocorticoids induce atrophy of both pituitary corticotroph and adrenocortical cells. However, iatrogenic adrenal insufficiency only becomes potentially relevant during or after glucocorticoid withdrawal. As iatrogenic adrenal insufficiency is transient in the majority of cases it can be suspected that the prevalence of permanent iatrogenic adrenal insufficiency is clearly lower than that of endogenous adrenal insufficiency.

A large number of frequent and rare causes of adrenal insufficiency are summarized in Tables 5.9.1 and 5.9.2, and in the following sections more detailed information on some of the more frequent causes is provided. Two reviews have given an excellent overview of causes of adrenal insufficiency including citation of all original literature which cannot be provided here because of space constraints (2, 3).

During the times of Thomas Addison, tuberculous adrenalitis was by far the most prevalent cause of adrenal insufficiency. In the developing world, tuberculosis still remains a major cause of adrenal insufficiency. In active tuberculosis, the incidence of adrenal involvement is 5%.

In North American and European countries, autoimmune adrenalitis accounts for more than 90% of cases with primary adrenal insufficiency; in 40% adrenal insufficiency is isolated while in 60% it arises as part of an autoimmune polyglandular syndrome (APS) (2, 4). APS type 1, also termed autoimmune polyendocrinopathy–candidiasis–ectodermal dystrophy, accounts for 15% of cases and is characterized by adrenal insufficiency, hypoparathyroidism, and chronic mucocutaneous candidiasis, the latter being the primary manifestation in most cases and already apparent in childhood (5). APS 1 is caused by mutations in the autoimmune regulator gene (AIRE) (6–8) while APS 2 is thought to be inherited as a complex trait, associated with loci within the major histocompatibility complex (4) and distinct susceptibility genes (9–11). APS 2 is much more common than APS 1 and in addition to adrenal insufficiency most frequently comprises autoimmune thyroid disease, albeit more often autoimmune hypothyroidism than Graves’ disease.

X-linked adrenoleucodystrophy (ALD) is caused by a mutation in the X-ALD gene, which encodes a peroxisomal membrane protein (adrenoleucodystrophy protein), leading to accumulation of very long chain fatty acids (>24 carbon atoms). The clinical picture comprises adrenal insufficiency and neurological impairment due to white matter demyelination. The two major forms are cerebral ALD (50% of cases; early childhood manifestation, rapid progression) and adrenomyeloneuropathy (35% of cases; onset in early adulthood, slow progression) with restriction of demyelination to spinal cord and peripheral nerves. Adrenal insufficiency may precede the onset of neurological symptoms and is the sole manifestation of disease in 15% of cases.

Other causes of primary adrenal insufficiency (Table 5.9.1), e.g. adrenal infiltration or haemorrhage, are rare. Congenital or neonatal primary adrenal insufficiency accounts for only 1% of all cases. However, the recent elucidation of the genetic basis of underlying diseases has highlighted the importance of specific genes for adrenal development and steroidogenesis.

The most common cause of secondary adrenal insufficiency is a tumour of the hypothalamic–pituitary region, usually associated with panhypopituitarism as a result of tumour growth or treatment with surgery and/or irradiation (Table 5.9.2). Autoimmune lymphocytic hypophysitis is less frequent, mostly affecting women during or shortly after pregnancy. Isolated ACTH deficiency may also be of autoimmune origin as some patients concurrently suffer from other autoimmune disorders, most frequently thyroid disease. The differential diagnosis of postpartal autoimmune hypophysitis includes Sheehan’s syndrome, which results from pituitary apoplexy, mostly due to pronounced blood loss during delivery. Very rarely mutations of genes important for pituitary deve-lopment or for synthesis and processing of the corticotropin precursor proopiomelanocortin cause secondary adrenal insufficiency (Table 5.9.2).

The clinical signs and symptoms of both acute and chronic adrenal insufficiency are a logical consequence of the underlying pathology, i.e. mostly the deficiency of adrenal corticosteroid production arising from primary or secondary adrenal failure (Table 5.9.3).

Acute adrenal insufficiency, i.e. life-threatening adrenal crisis, typically presents with severe hypotension or hypovolaemic shock, acute abdominal pain, vomiting, and often with fever, and, therefore, is sometimes mistaken for acute abdomen. In a series of 91 patients with Addison’s disease, adrenal crisis led to the initial diagnosis of adrenal insufficiency in half of the patients. In children, acute adrenal insufficiency often presents as hypoglycaemic seizures. Deterioration of glycaemic control with recurrent hypoglycaemia may be the presenting sign of adrenal insufficiency in patients with pre-existing type 1 diabetes. In APS 2, onset of autoimmune hyperthyroidism (or thyroxine replacement for newly diagnosed hypothyroidism) may precipitate adrenal crisis due to enhanced cortisol clearance.

The leading symptom of chronic adrenal insufficiency is fatigue, accompanied by lack of stamina, loss of energy, reduced muscle strength, and increased irritability. In addition, chronic glucocorticoid deficiency leads to weight loss, nausea, and anorexia (in children, failure to thrive) and may account for muscle and joint pain. Unfortunately, most of these symptoms are nonspecific. Thus, every second patient suffers from signs and symptoms of Addison’s disease for more than 1 year before diagnosis is established. In secondary adrenal insufficiency, diagnosis is mostly prompted by a history of pituitary disease, but may also be delayed, e.g. in isolated ACTH deficiency. A more specific sign for primary adrenal failure is hyperpigmentation (Fig. 5.9.3), which is most pronounced in areas of the skin exposed to increased friction (e.g. hand lines, knuckles, scars, oral mucosa). Hyperpigmentation is due to enhanced stimulation of skin MC1-receptor by ACTH and other pro-opiomelanocortin-related peptides. Accordingly, patients with secondary adrenal insufficiency often present with pale, alabaster-coloured skin. Laboratory findings in glucocorticoid deficiency may include mild anaemia, lymphocytosis, and eosinophilia. Cortisol physiologically inhibits thyrotropin release. Thus, thyrotropin is often increased at initial diagnosis of primary adrenal insufficiency, but returns to normal during glucocorticoid replacement unless there is coincident autoimmune thyroid failure. In rare cases, glucocorticoid deficiency may result in hypercalcaemia, which is due to increased intestinal absorption and decreased renal excretion of calcium and usually coincides with autoimmune hyperthyroidism, facilitating calcium release from bone.

Mineralocorticoid deficiency, which is only present in primary adrenal insufficiency, leads to dehydration and hypovolaemia, resulting in low blood pressure, postural hypotension, and sometimes even in prerenal failure. Deterioration may be sudden and is often due to exogenous stress such as infection or trauma. Combined mineralocorticoid and glucocorticoid replacement in primary adrenal insufficiency reconstitutes the diurnal rhythm of blood pressure and reverses cardiac dysfunction. Glucocorticoids contribute to this amelioration not only by mineralocorticoid receptor binding, but also by permissive effects on catecholamine action. The latter may account for the relative unresponsiveness to catecholamines in patients with unrecognized adrenal crisis. Mineralocorticoid deficiency accounts for hyponatraemia (90%), hyperkalaemia (65%), and salt craving (15%). Low serum sodium may also be present in secondary adrenal insufficiency due to the syndrome of inappropriate antidiuretic hormone secretion, which results from the loss of physiological inhibition of pituitary vasopressin release by glucocorticoids.

Adrenal insufficiency inevitably leads to DHEA deficiency. DHEA is the major precursor of sex steroid synthesis and loss of its synthesis results in pronounced androgen deficiency in women. As a consequence, women with adrenal insufficiency frequently show loss of axillary and pubic hair (absence of pubarche in children), dry skin, and reduced libido. DHEA also exerts direct action as a neurosteroid with potential antidepressant properties. Thus DHEA deficiency may contribute to the impairment of well-being that is observed in patients with adrenal insufficiency despite adequate glucocorticoid and mineralocorticoid replacement.

Presentation with acute adrenal insufficiency, i.e. life-threatening adrenal crisis, requires an immediate, combined diagnostic and therapeutic approach (Fig. 5.9.4). Haemodynamically stable patients may undergo a cosyntropin stimulation test; if in doubt, baseline bloods for serum cortisol and plasma ACTH will suffice and if cortisol is less than 100 nmol while ACTH is considerably elevated, there is no doubt about the diagnosis. Formal confirmation of diagnosis can be performed following clinical improvement. Diagnostic measures must never delay treatment, which should be initiated upon strong clinical suspicion of adrenal insufficiency. It is of negligible risk to start hydrocortisone and stop it after adrenal insufficiency has been safely excluded; withholding potentially life-saving treatment, however, could have fatal consequences.

Adrenal insufficiency is readily diagnosed by the cosyntropin test, a safe and reliable diagnostic tool with excellent long-term predictive value (12, 13); it is important to be aware of the considerable variability between results of different cortisol assays (14) and when defining the cut-off for failure, commonly set at 500 nmol/l, one should ideally refer to results from a local reference cohort obtained with the same assay. The diagnostic value of the cosyntropin test is only compromised within the first 4 weeks following a pituitary insult (13, 15), as during this period the adrenals will still respond to exogenous ACTH stimulation despite the loss of endogenous ACTH drive. When suspecting secondary adrenal insufficiency, the insulin tolerance test is an alternative choice for diagnostic confirmation, considered by many as the gold standard, however it is associated with side effects and requires exclusion of cardiovascular disease and history of seizures. Formal confirmation of diagnosis by the cosyntropin stimulation test should include blood samples for plasma ACTH, which will guide the way for further diagnostic assessment, by reliably differentiating primary from secondary adrenal insufficiency, i.e. adrenal from hypothalamic–pituitary disease (Fig. 5.9.4).

Possible glucocorticoid deficiency is also indicated by normocytic anaemia as sufficient levels of cortisol are required for maturation of blood progenitor cells; other blood count changes may include lymphocytosis and eosinophilia. Sometimes also, mild metabolic acidosis or hypercalcaemia can be observed in affected patients, the latter mostly in the context of coincident hyperthyroidism. Serum glucose may be low; however, significant hypoglycaemia as a presenting sign plays a more important role in childhood adrenal insufficiency where it can result in significant brain damage. However, in a patient with pre-existing type 1 diabetes onset of recurrent hypoglycaemic episodes despite unchanged insulin regimen should raise the suspicion of adrenal insufficiency.

Mineralocorticoid deficiency is present in primary adrenal insufficiency only; the renin–angiotensin–aldosterone system in patients with hypothalamic–pituitary disease and intact adrenals is usually preserved. Mineralocorticoid deficiency is not only reflected by the arterial hypotension and deranged potassium and sodium but intravascular volume depletion is also indicated by the slightly raised creatinine, a common finding in Addison patients. Hyponatraemia is observed in about 80% of acute cases while less than half present with hyperkalaemia. In the first instance, baseline levels of serum aldosterone and plasma renin should be taken.

The combined measurement of early morning serum cortisol and plasma ACTH separates patients with primary adrenal insufficiency from normal subjects and patients with secondary adrenal insufficiency. Plasma ACTH is usually grossly elevated and invariably higher than 22 pmol/l with serum cortisol usually below the normal range (<165 nmol/l), sometimes also in the lower normal range. Establishment of the diagnosis of primary adrenal insufficiency always depends on the combined measurement of ACTH and cortisol. Serum aldosterone concentrations are subnormal or within the lower normal range with plasma renin activity concurrently increased above the normal range. In patients with adrenal insufficiency, serum DHEAS is invariably low, in women often below the limit of detection.

The impaired ability of the adrenal cortex to respond to ACTH is readily demonstrated by the short synacthen test (SST), employing serum cortisol measurements before and 30 (or 60) min after IV (or IM) injection of 250 µg 1–24 ACTH. In normal subjects, this leads to a physiological increase in serum cortisol to peak concentrations above 500 nmol/l. In primary adrenal insufficiency, the adrenal cortex is already maximally stimulated by endogenous ACTH, exogenous ACTH administration therefore usually does not evoke any further increase in serum cortisol.

Adrenal cortex autoantibodies and/or antibodies against 21-hydroxylase are found in more than 80% of patients with recent-onset autoimmune adrenalitis. While 21-hydroxylase has been identified as the major autoantigen in autoimmune adrenalitis, autoantibodies against other steroidogenic enzymes (P450scc, P450c17) and steroid-producing cell antibodies are present in a lower percentage of patients. Measurement of autoantibodies is particularly helpful in patients with isolated primary adrenal insufficiency and no family history of autoimmune disease. In APS 2, autoimmune adrenalitis may be associated with autoimmune thyroid disease or type 1 diabetes and screening for concomitant disease should involve measurements of thyrotropin and fasting glucose but not of other organ-related antibodies.

In male patients with isolated primary adrenal insufficiency without unequivocal evidence of autoimmune adrenalitis, serum very long chain fatty acids (chain length of 24 carbons and more; C26, C26/C22, and C24/C22 ratios) should be measured to exclude adrenoleucodystrophy/adrenomyeloneuropathy.

Baseline hormone measurements only poorly separate patients with secondary adrenal insufficiency from normal subjects. However, a morning cortisol below 100 nmol/l indicates adrenal insufficiency whereas a serum cortisol greater than 500 mmol/l is consistent with an intact HPA axis. Thus in most cases dynamic tests of the HPA axis are required to establish the diagnosis of secondary adrenal insufficiency.

The insulin tolerance test (ITT) is still regarded as the ‘gold standard’ in the evaluation of suspected secondary adrenal insufficiency, as hypoglycaemia (blood glucose <2.2 mmol/l) is a powerful stressor resulting in rapid activation of the HPA axis. An intact HPA axis is demonstrated by a peak cortisol above 500 nmol/l at any time during the test. The occasional patient will pass the ITT while exhibiting clinical evidence for adrenal insufficiency responding to hydrocortisone substitution and a higher cut-off value (550 nmol/l) may help to reduce misclassification. During ITT close supervision is mandatory and cardiovascular disease and history of seizures represent contraindications.

Another test for the diagnosis of secondary adrenal insufficiency is the overnight metyrapone test (30 mg metyrapone/kg (maximum 3 g) with a snack at midnight). Metyrapone inhibits adrenal 11β-hydroxylase, i.e. the conversion of 11-deoxycortisol to cortisol. In normal subjects, HPA feedback activation will increase serum 11-deoxycortisol, while serum cortisol remains less than 230 nmol/l. In patients with secondary adrenal insufficiency, 11-deoxycortisol does not exceed 200 nmol/l at 8.00 hours after metyrapone. Shortcomings of the test are limited availability of reliable 11-deoxycortisol assays and of the drug itself, which cannot be obtained in all countries though it is readily available in the UK. As metyrapone may precipitate adrenal crisis in severe cortisol deficiency, a morning cortisol above 200 nmol/l should be documented prior to performing the test on an outpatient basis.

As both the ITT and the metapyrone test pose a significant burden to patients and physicians, there have been continuing efforts to replace these tests by more convenient tools. Sustained secondary adrenal insufficiency leads to adrenal atrophy and also to reduced adrenal ACTH receptor expression, as ACTH up-regulates its own receptor. Thus adrenal responsiveness to an acute exogenous ACTH challenge is impaired also in secondary adrenal insufficiency facilitating the use of the SST for the evaluation of HPA axis integrity (Fig. 5.9.4). Several studies have reported excellent agreement between peak cortisol values in SST and in ITT (13) and a recent study has convincingly demonstrated the long-term predictive accuracy of the SST (12). However, there is some evidence that some patients with secondary adrenal insufficiency will pass the SST while failing the ITT. The use of a higher cut-off value (600 nmol/l) for passing the SST may minimize the risk of overlooking secondary adrenal insufficiency but this is largely assay dependent as different radioimmunoassays will have different cut-offs. Our traditional cut-offs are certain to undergo changes in the imminent future with the introduction of more specific tandem mass spectrometry for the measurement of serum cortisol.

There are other tests that have been used in the diagnostic assessment of adrenal insufficiency but whose use is not recommended as a routine procedure. As the administration of 250 µg 1–24 ACTH represents a massive supraphysiological challenge, a low-dose corticotropin test (LDT) employing only 1 µg ACTH has been proposed as a more sensitive test for the diagnosis of secondary adrenal insufficiency. The LDT has been successfully used to monitor recovery of adrenal function after withdrawal of oral glucocorticoids and to detect subtle impairment of adrenal reserve during inhalative steroid therapy. However, the administration of 1 μg ACTH IV still results in ACTH levels above those required for maximum cortisol release. Accordingly, in normal subjects serum cortisol concentrations measured 30 min after the ACTH challenge do not differ between SST and LDT. Thus it is currently a matter of debate whether employing the LDT represents any advantage (16, 17), which would be further offset by handling problems due to the necessity of dilution from the commercially available 250 µg ACTH 1–24 ampoule and due to the potential binding of ACTH to the surface of injection devices.

CRH has been used to differentiate hypothalamic from pituitary disease in secondary adrenal insufficiency. However, CRH stimulation is not of great help in actually diagnosing secondary AI as individual responses to exogenous CRH are highly variable and cut-off values or even normal ranges are still not well defined.

Finally, a word of caution: none of the tests, including the ITT, will classify all patients correctly. Mild secondary adrenal insufficiency may pass as intact HPA axis and some healthy subjects may fail any single test by a small margin. Thus clinical judgement remains important. Persisting symptoms such as fatigue, myalgia, or reduced vitality should lead to reassessment.

Screening for adrenal insufficiency by SST should not be performed immediately after pituitary surgery, but only 4 to 6 weeks later, as adrenal atrophy may develop only gradually after onset of ACTH deficiency. Until then, patients with a morning cortisol not excluding secondary adrenal insufficiency (<450 nmol/l at 3 days and <350 nmol/l at 7 days after surgery) should receive hydrocortisone replacement paused 24 h prior to scheduled adrenal function testing. The impairment of other hormonal axes after pituitary surgery increases the likelihood of ACTH deficiency, whereas isolated corticotropin deficiency is uncommon.

In critically ill patients the corticotropic axis is markedly activated (18, 19). Moreover, patients in intensive care units are less sensitive to dexamethasone suppression and achieve higher peak ACTH and cortisol concentrations after CRH. In addition, patients with critical illness show relatively low serum aldosterone levels with concurrently elevated plasma renin activity. Cortisol concentrations correlate with illness-severity scores and are highest in patients with the highest mortality. On the other hand, cytokine activation may induce relative secondary adrenal insufficiency in some patients with severe illness, thus putting them at risk of dying from adrenal crisis. Chronic inhibition of cortisol production by etomidate has been associated with increased mortality in intensive care unit patients. Unfortunately, no consensus exists how to diagnose adrenal insufficiency in critically ill patients. In patients with primary adrenal insufficiency or severe secondary adrenal insufficiency the SST will establish the diagnosis by demonstrating a low baseline cortisol (<165 nmol/l) not responding to corticotropin (peak cortisol <500 nmol/l). However, it has been suggested that ‘relative’ adrenal insufficiency may be present in a number of critically ill patients, characterized by a poor cortisol response (increment <248 nmol/l) to ACTH despite normal baseline cortisol. These patients often present with catecholamine-dependent hypodynamic shock responding to hydrocortisone administration. One study has reported decreased mortality in patients with septic shock and abnormal cortisol response in the SST (increment <248 nmol/l) after treatment with hydrocortisone (20) but a prospective study did not support this finding (21). At present it seems prudent to collect a random sample of serum cortisol and plasma ACTH in critically ill patients with suspected adrenal insufficiency followed by immediate hydrocortisone administration. Depending on the results of these hormone determinations (serum cortisol >700 nmol/l rules out adrenal insufficiency) hydrocortisone therapy is terminated or a more detailed evaluation employing the SST is performed.

If there is no coexisting autoimmune disease and adrenal and steroid autoantibodies are negative, imaging of the adrenals, preferably by CT, is warranted (Fig. 5.9.4). Tuberculosis should be considered, which is frequent in developing countries and therefore also in migrant populations. Chest radiography is helpful and imaging of the adrenals typically shows hyperplastic organs in the early phase and spotty calcifications in the late phase of tuberculous adrenalitis. Much rarer causes are bilateral infiltration by bilateral primary adrenal lymphoma, (predominantly lung cancer) metastases (22, 23), sarcoidosis, haemochromatosis, or amyloidosis. Bilateral adrenal haemorrhage is usually only seen during septic shock or in very rare instances in primary antiphospholipid syndrome (24). In male patients with isolated Addison’s and negative autoantibodies, imaging should be preceded by measurement of plasma very long chain fatty acids to safely exclude X-linked adrenoleucodystrophy which affects 1 in 20 000 males (25). ABCD1 gene mutations encoding for the peroxisomal ALD protein involved in cross-membrane transport manifest in 50% of cases in early childhood and primarily with CNS symptoms. However, the adrenomyoeloneuropathy variant, accounting for 35% of cases, can manifest with adrenal insufficiency prior to the development of spinal paraparesis during early adulthood (25).

If ACTH is inappropriately low in the presence of cortisol deficiency, imaging of the hypothalamic–pituitary region by MRI is the first diagnostic measure that should be arranged for, alongside an endocrine pituitary baseline profile (Fig. 5.9.4). Pituitary adenomas are most common, craniopharyngiomas are much rarer and may present at any age; very rare causes include meningioma, metastases, and infiltration by sarcoidosis, Langerhans’ cell histiocytosis, or other granulomatous disease. Careful history taking should ask for previous head trauma (26, 27), surgery, radiotherapy, and for clinical indicators of pituitary apoplexy (28), i.e. the sudden onset of high-impact headache (29). The latter may occur spontaneously in larger pituitary adenomas or may result from sudden hypocirculation during surgery or as a consequence of complicated deliveries with significant blood loss, the classical cause of Sheehan’s syndrome. Lymphocytic hypophysitis of autoimmune origin (30) commonly presents with panhypopititarism including diabetes insipidus and a pituitary mass effect. However, it may present with isolated ACTH deficiency, in some cases coinciding with autoimmune thyroid disease (31, 32).

Importantly, the most obvious should not be forgotten—suppression of the hypothalamic–pituitary axis by exogenous glucocorticoid treatment. This should always be excluded, considering not only oral steroid intake but also glucocorticoid inhalers, creams, or intra-articular injections.

A patient with suspected acute adrenal insufficiency certainly needs immediate therapeutic attention, with signs and symptoms very suggestive of adrenal insufficiency including patchy hyperpigmentation of the oral mucosa and the presence of severe hypovolaemic hypotension. If peripheral veins are collapsed, a central line for IV fluid resuscitation may be required, administered at an initial rate of 1 l/h and with continuous cardiac monitoring. In addition, hydrocortisone replacement should be commenced by intravenous injection of 100 mg hydrocortisone followed by continuous infusion of 150 mg hydrocortisone in 5% glucose per 24 h. Mineralocorticoid replacement does not need to be added in the acute setting as long as the total daily hydrocortisone dose is greater than 50 mg as such a dose will ensure sufficient mineralocorticoid receptor activation by cortisol.

Chronic glucocorticoid replacement requires additional considerations. Physiological daily cortisol production rates vary between 5 and 10 mg/m² (33), which is equivalent to the oral administration of 15 to 25 mg hydrocortisone, i.e. cortisol. After oral ingestion cortisol produces highly variable peak concentrations within the supraphysiological range followed by a rapid decline to below 100 nmol/l 5 to 7 h after ingestion (Fig. 5.9.5). I usually recommend the administration of hydrocortisone in two to three divided doses, e.g. 15 mg in the morning upon awakening followed by 5 mg 6 h later, or 10 mg upon awakening followed by 5 mg 4 h and 8 h later. It is important to let the patient experiment with different timings to find the most suitable regimen for his individual needs. Importantly, patients who work shifts have to adjust the timing of the glucocorticoid doses to their working times and subsequent sleep–wake cycle. Whether a thrice daily glucocorticoid regimen should be preferred over twice daily administration is not clear as well-designed and appropriately powered studies are lacking. As Fig. 5.9.5 clearly illustrates, neither of the two regimens will be able to achieve cortisol availability similar to that of physiological diurnal secretion. Some groups advocate weight-related dosing (34) and this appears to generate a smoother pharmacokinetic profile but data demonstrating superiority of such a regimen are lacking. However, body surface area adjusted glucocorticoid dosing is commonly used for guiding glucocorticoid replacement in children.

The oral administration of currently available cortisol preparations is not able to mimic the physiological pattern of cortisol secretion, which follows a distinct circadian rhythm. Cortisol secretion begins to rise between 02.00 and 04.00 hours, peaks within an hour of waking and then declines gradually to low levels during the evening and nadir levels at and after midnight (35). There is evidence for a diurnal variability in glucocorticoid sensitivity. Plat et al. (36) have demonstrated that a more unfavourable metabolic response occurs to evening administration of hydrocortisone. Also, high levels of glucocorticoids may disrupt sleep, thus late evening hydrocortisone administration should be avoided; sleep disturbances contributing to increased fatigue are a common feature in chronic adrenal insufficiency (37, 38).The delivery of cortisol by intravenous infusion (39) or subcutaneous pump (40) can closely mirror diurnal secretion, but these administration modes are obviously not suited for routine delivery. Recently developed modified and delayed release hydrocortisone preparations mimicking physiological cortisol secretion represent a very promising therapeutic approach (41, 42).

Cortisone acetate requires intrahepatic activation to cortisol by 11β-hydroxysteroid dehydrogenase 1, which contributes to a higher pharmacokinetic variability compared to hydrocortisone; 25 mg cortisone acetate are equivalent to 15 mg hydrocortisone (43, 44). Long-acting glucocorticoids are also used for replacement, e.g. in 20% of respondents to the 2002 survey of the North American Addison Disease Foundation. Some countries do not have access to hydrocortisone or cortisone acetate and therefore have to resort to long-acting synthetic glucocorticoids. However, prednisolone and dexamethasone have considerably longer biological half-lives, likely to result in unfavourably high night-time glucocorticoid activity with potentially detrimental effects on insulin sensitivity and bone mineral density (45). In addition, available preparations offer limited options for dose titration. Therefore I generally recommend against the use of synthetic glucocorticoids for replacement therapy in adrenal insufficiency; the only exception are patients with concurrent insulin-dependent diabetes in whom prednisolone may help to avoid the peaks and troughs of hydrocortisone pharmacokinetics and thus also subsequent rapid changes in glucose control. For clinical purposes, I assume equipotency to 1 mg hydrocortisone for 1.6 mg cortisone acetate, 0.2 mg prednisolone, 0.25 mg prednisone, and 0.025 mg dexamethasone, respectively. While equipotency doses of hydrocortisone and cortisone acetate are based on pharmacokinetic studies (43, 44), suggested doses for synthetic steroids are based on estimates from older studies comparing the relative anti-inflammatory properties of various glucocorticoids.

Monitoring of glucocorticoid replacement is mainly based on clinical grounds as a reliable biomarker for glucocorticoid activity has yet to be identified (Table 5.9.4). Plasma ACTH cannot be used as a criterion for glucocorticoid dose adjustment. In primary adrenal insufficiency, ACTH is invariably high before the morning dose and rapidly declines with increasing cortisol levels after glucocorticoid ingestion (46, 47) (Fig. 5.9.6a). Aiming at ACTH levels within the normal range would therefore invariably result in overreplacement. In secondary adrenal insufficiency, plasma ACTH is anyway low and thus not informative. Urinary 24-h free cortisol excretion has been advocated for monitoring of replacement quality (48). However, after exogenous glucocorticoid administration, urinary cortisol excretion shows considerable interindividual variability. Also, following glucocorticoid absorption cortisol-binding globulin is rapidly saturated, resulting in transient but pronounced increases in renal cortisol excretion. Thus, one cannot refer to normal ranges for healthy subjects when judging urinary cortisol excretion during replacement therapy for adrenal insufficiency. Some authors have suggested regular measurements of serum cortisol day curves to monitor replacement therapy (48, 49). However, the efficacy of this approach is not supported by controlled studies and recent data indicate a poor correlation between clinical assessment and cortisol levels (50). Timed serum cortisol measurements can be of some value in selected patients, e.g. in case of suspected noncompliance or malabsorption; however, random serum cortisol measurements without information on the time of the hydrocortisone dose are not informative.

Thus, in the absence of objective parameters, the physician has to rely primarily on clinical judgment, carefully taking into account signs and symptoms potentially suggestive of glucocorticoid over- or under-replacement, recognizing their relative lack of specificity. Glucocorticoid under-replacement bears the risk of incipient crisis and significant impairment of well-being. Conversely, chronic over-replacement may lead to substantial morbidity including impaired glucose tolerance, obesity and osteoporosis. An increased incidence of osteoporosis has only been reported in patients receiving daily replacement doses of 30 mg hydrocortisone or higher (45, 51, 52) or 7.5 mg prednisone (45) whereas appropriate replacement doses of 20–25 mg hydrocortisone do not affect bone mineral density (50, 53). Therefore, bone mineral density measurements are not routinely required in patients with adrenal insufficiency receiving recommended glucocorticoid replacement doses.

Patients with primary adrenal insufficiency require mineralocorticoid replacement, which usually consists of the oral administration of 9α-fludrocortisone; fluorination at the 9α position ensures selective binding to the MR and thus exclusive mineralocorticoid action. By contrast, cortisol binds with equal affinity to both the GR and MR. However, excessive MR binding of cortisol in the kidney is prevented by 11β-hydroxysteroid dehydrogenase type 2, which inactivates cortisol to cortisone. Oelkers has coined the term ‘mineralocorticoid unit’ (MCU), determining that 100 MCU are equivalent to 100 µg fludrocortisone and 40 mg hydrocortisone, respectively (54). By contrast, prednisolone exerts only reduced and dexamethasone no mineralocorticoid activity at all; therefore patients treated with synthetic glucocorticoids need particularly careful monitoring of their mineralocorticoid replacement.

In the newly diagnosed patient, mineralocorticoid replacement should be initiated at 100 µg once daily; optimized doses may vary between 50 and 250 µg. Children, in particular neonates and infants, have considerably higher mineralocorticoid dose requirements and often need additional salt supplementation. However, also amongst adults there is a good degree of interindividual variability. A high dietary salt intake may slightly reduce mineralocorticoid requirements. An important additional factor is temperature and humidity, e.g. individuals living in Mediterranean summer or tropical climates will require a 50% increase in fludrocortisone dose due to increased salt loss through perspiration. Monitoring (Table 5.9.4) includes supine and erect blood pressure and serum sodium and potassium; plasma renin activity should be checked regularly, aiming at the upper normal range (54). If essential hypertension develops, mineralocorticoid dose may be slightly reduced, accompanied by monitoring of serum sodium and potassium, but complete cessation of mineralocorticoid replacement should be avoided. It is important to recognize that plasma renin physiologically increases during pregnancy; therefore, monitoring in pregnancy should comprise blood pressure, serum sodium and potassium, and, if required, urinary sodium excretion. During the last term of pregnancy fludrocortisone dose may require adjustment, also due to increased progesterone levels exerting antimineralocorticoid activity (55).

Risk of adrenal crisis is higher in primary adrenal insufficiency and several factors such as coincident APS or age have been suggested as additional modifiers (2, 56). Many crises are due to glucocorticoid dose reduction or lack of stress-related glucocorticoid dose adjustment by patients or general practitioners (2). A recent survey in 526 patients found that 42% of patients (47% in primary adrenal insufficiency, 35% in secondary adrenal insufficiency) had experienced at least one adrenal crisis during the course of their disease. Precipitating causes were mainly gastrointestinal infections and fever but also several other causes, including major pain, surgery, psychological distress, heat, and pregnancy. This was corroborated by data from a large patient survey (n = 841) (57) that also highlighted gastrointestinal infections as the single most important cause of crisis. Thus adrenal crises are a predictable and frequent, but still undermanaged event and crisis prevention is a key strategy that needs to be pursued.

All patients and their partners should receive regular crisis prevention training, including verification of steroid emergency card/bracelet and instruction on stress-related glucocorticoid dose adjustment (Table 5.9.4). Generally, hydrocortisone should be doubled during intercurrent illness, such as a respiratory infection with fever, until clinical recovery. Gastrointestinal infections, a frequent cause of crisis, may require parenteral hydrocortisone administration,. Preferably all patients, but at least patients travelling or living in areas with limited access to acute medical care should receive a hydrocortisone emergency self-injection kit (e.g. 100 mg for IM injection). For major surgery, trauma, delivery, and diseases requiring intensive care unit monitoring, patients should receive intravenous administration of 100–150 mg hydrocortisone per 24 h in 5% glucose or 25–50 mg hydrocortisone IM four times per day. Some authors have advocated lower doses (25–75 mg/24 h) for surgical stress (58). However, 60 years after the seminal observation that glucocorticoid replacement needs to be increased during periods of major stress (59) studies clarifying exact dose requirements are still outstanding.

The introduction of DHEA, the third major steroid produced by the adrenal gland, into the replacement regimen for adrenal insufficiency (60) represents a major advance, in particular for women who are invariably androgen deficient (60, 61). DHEA has been shown to significantly enhance well-being, mood, and subjective health status in women with primary and secondary adrenal insufficiency (60, 62–65) and also recently in children and adolescents with adrenal failure (66). Similar effects have been described for testosterone replacement in hypopituitarism (67), however, no study has yet directly compared DHEA to testosterone. In addition to acting as an androgen precursor, DHEA has neurosteroidal properties, exerting a primarily antidepressive effect, and also shows immunemodulatory properties (68). Of note, DHEA has been shown to exert beneficial effects on subjective health status and energy levels not only in women but also in men with primary adrenal insufficiency (63, 64) including significant beneficial effects on bone mineral density and truncal lean mass (63).

Currently, DHEA replacement is hampered by the lack of pharmaceutically controlled preparations, with questionable quality and content of several over-the-counter preparations (69). At present, DHEA should be reserved for patients with adrenal insufficiency suffering from significant impairment in well-being despite otherwise optimized replacement, in particular women with signs of androgen deficiency such as dry and itchy skin and loss of libido. DHEA should be taken as a single dose (25–50 mg) in the morning. Treatment monitoring (Table 5.9.4) should include blood sampling 24 h after the last preceding morning dose for measurement of serum DHEAS (in women also androstenedione, testosterone, sex hormone-binding hormone) aiming at the middle normal range for healthy young subjects. I usually start patients on 25 mg and increase to 50 mg after 2 to 4 weeks, advising them to halve the dose if androgenic skin side effects (greasy skin, spots) persist for more than a week. Obviously, transdermal testosterone represents an alternative androgen replacement tool in women with adrenal failure.

Hyperthyroidism results in increased cortisol metabolism and clearance and hypothyroidism the converse, principally due to an effect of thyroid hormone upon hepatic 11β-HSD1 and 5α/5β-reductases. Insulin-like growth factor 1 (IGF-1) increases cortisol clearance by inhibiting hepatic 11β-HSD1 (conversion of cortisone to cortisol). In patients with adrenal insufficiency and unresolved hyperthyroidism, glucocorticoid replacement should be doubled to tripled. To avoid adrenal crisis, thyroxine replacement for hypothyroidism should only be initiated after concomitant glucocorticoid deficiency has either been excluded or treated. Obviously overt endogenous hyperthyroidism will also increase hydrocortisone metabolism (Fig. 5.9.6b). Therefore, the initiation of glucocorticoid replacement in patients with newly diagnosed hypopituitarism should always precede the initiation of thyroxine replacement as the reverse might precipitate adrenal crisis.

Pregnancy is physiologically associated with a gradual increase in CBG and during the last term of pregnancy also with an increase in free cortisol. In addition, serum progesterone increases, exerting antimineralocorticoid action. Therefore, during the third trimester, hydrocortisone replacement should be increased by 50%. Plasma renin activity cannot serve as a monitoring tool because it physiologically increases during pregnancy. Peripartal hydrocortisone replacement should follow the requirements for major surgery, i.e. 100 mg/24 h starting with labour until 48 h after delivery, followed by rapid tapering.

When deciding on the glucocorticoid dose, it is important to consider concurrent medication, in particular drugs known to increase hepatic glucocorticoid metabolism by CYP3A4 induction, which results in increased 6β-hydroxylation and hence cortisol inactivation (2, 3). 6β-hydroxylation by CYP3A4 is normally a minor pathway but cortisol itself induces CYP3A4 so that 6β-hydroxycortisol excretion is markedly increased in patients with Cushing’s syndrome. A multitude of drugs are known to induce CYP3A4 (Table 5.9.1), which require a two- to threefold increase in glucocorticoid dose. Conversely, the intake of drugs inhibiting CYP3A4 would require reduction of glucocorticoid replacement dose.

Of note, treatment of tuberculosis with rifampicin increases cortisol clearance but does not influence aldosterone clearance. Thus, glucocorticoid replacement should be doubled during rifampicin treatment. Mitotane (o,p′DDD, ortho, para′, dichlorodiphenyldichloroethane) decreases bioavailable glucocorticoid levels due to an increase in CBG and concurrently enhanced glucocorticoid metabolism following induction of CYP3A4. During chronic mitotane treatment, e.g. in adrenal carcinoma (70), usual glucocorticoid replacement doses should therefore be at least tripled.

Recent data demonstrate that current standard replacement fails to restore quality of life, which is significantly impaired in both patients with primary and secondary adrenal insufficiency (37, 71), with no apparent difference between prednisolone and hydrocortisone-treated patients (72). Predominant complaints are fatigue, lack of energy, depression, anxiety, and reduced ability to cope with daily demands; the degree of impairment is comparable to that observed in congestive heart failure and chronic haemodialysis patients (37, 71). Subjective health status is most reduced in younger patients but all age groups are significantly impaired (71), a persistent finding even if only analysing patients without any comorbidity (71) This also has a socioeconomic perspective as patients with Addison’s disease have a two- to threefold higher likelihood of receiving disablement pensions (37, 71).

In addition, large cohort studies have demonstrated an increased mortality not only in patients with secondary adrenal insufficiency due to hypopituitarism (73) but also in primary adrenal insufficiency, i.e. Addison’s disease (74, 75), a finding still valid when the influence of comorbidities is excluded. The causes underlying this increased mortality remain unclear, but we certainly need to consider the possible impact of current replacement regimens on the observed increase in mortality from cardiovascular and cerebrovascular disease and respiratory infections.

More than 150 years after Thomas Addison first described a disease characterized by salt wasting and hyperpigmentation as the result of adrenal gland destruction (1), adrenal insufficiency is no longer an invariably fatal condition. The landmark achievement of the synthesis of cortisone in the late 1940s and its introduction into therapy in the early 1950s quickly lead to widespread availability of life-saving glucocorticoid replacement therapy. However, while initial survival is routinely achieved nowadays, current replacement regimens may not be able to achieve normal quality of life. Future research has to uncover the causes underlying the increased mortality in adrenal insufficiency and should further explore the role of novel replacement modalities, such as DHEA and modified-release hydrocortisone.