Testing Procedures

For 24-hour urine collections, creatinine is measured to verify the adequacy of the collection. To preserve the specimen adequately, urine should be collected in a container to which 25 mL of 6 N HCl has been added.

Plasma catecholamines are collected after an overnight fast (water permitted). The patient is placed in a reclining position in a quiet environment, and a heparin lock is inserted intravenously. After 20 to 30 minutes, blood is collected in a prechilled ethylenediaminetetraacetic acid (EDTA) lavender-top tube. The whole blood sample should be kept in ice water until centrifuged (preferably at 4° C). Separation of plasma should take place within 2 hours of phlebotomy; the sample should then be frozen immediately.

Both urine and plasma specimens should be analyzed using high-performance liquid chromatography with tandem mass spectrometry, as this technique greatly eliminates problems caused by interfering substances (Taylor & Singh, 2002). For the urine metabolites, it is important to use age-appropriate reference ranges when interpreting the results. It is also important to be aware that reference ranges can vary widely not only from one laboratory to another, but even within the same laboratory, as newer methods are introduced.

Additional Follow-Up Testing

Once the diagnosis is confirmed biochemically, the tumor should be localized by a CT scan or MRI of the adrenal glands. If this is negative, imaging studies of the abdomen, chest, and pelvis should be performed. CT has greater sensitivity, and MRI, greater specificity. MRI is superior to CT in detecting extraadrenal lesions and has the advantage of not requiring ionizing radiation or needing ionic contrast. If a tumor cannot be located by CT or MRI, or if metastatic disease is suspected, scanning with ¹³¹I- or ¹²³I-labeled meta-iodobenzylguanidine should be performed. Octreoscanning and positron emission tomography are reserved for when the other techniques have failed. ¹⁸F-FDG PET/CT scanning is especially useful in patients with paragangliomas or metastatic disease (Pacak et al, 2001b).

Following successful tumor resection, the prognosis is generally excellent. Urinary metanephrines should be retested several weeks after surgery to ensure that the resection was complete, and should be measured periodically thereafter as an early marker of disease recurrence (Werbel & Ober, 1995).

An algorithm to evaluate pheochromocytoma has been proposed by Eisenhofer and colleagues (2003) (Fig. 24-10).

21-Hydroxylase Deficiency

The enzyme 21-hydroxylase (also referred to as CYP21, CYP21A2, and P450c21) is located within the mitochondrial endoplasmic reticulum. 21-hydroxylase deficiency is the most common cause of CAH, accounting for about 95% of all cases. Screening of newborns using capillary heel blood on paper filter disks has identified this disorder in about 1 in 14,000 persons in North America, and in as many as 1 in 300 Yupik Eskimos of Alaska (Pang et al, 1988, 1982). The classic form is detected in about 1 in 16,000 live births; the nonclassic form is seen in about 0.2% of the general Caucasian population, and in 1% to 2% of those of Eastern European Jewish ancestry (Speiser et al, 1985; Therrell, 2001).

21-OH catalyzes the conversion of 17-hydroxyprogesterone (17-OHP) to 11-deoxycortisol (compound S), and progesterone to DOC. Thus, 21-OH deficiency leads to elevated levels of the steroid precursors 17-OHP and pregnanetriol in the urine and 17-OHP in the serum. These precursors are shunted toward the pathway, leading to excess androstenedione and testosterone production (Fig. 24-12). Clinical presentation often correlates with the severity of enzymatic dysfunction. The classic variety presents in the newborn period or in early childhood with adrenal insufficiency and virilization, with or without salt wasting. The nonclassic form presents in late childhood as early adrenarche, or in young adulthood as hirsutism, amenorrhea, and infertility. The presentation in women is very similar to that of polycystic ovarian disease. Males may develop precocious puberty, adrenal rests within the testes, and infertility.

A close functional relationship exists between the adrenal cortex and the adrenal medulla. Dysplasia of the medulla and catecholamine hyposecretion have been described in classic 21-OH deficiency. In a study of 38 children with classic (CYP21A2) disease, levels of plasma epinephrine and metanephrine and urinary epinephrine were 40% to 80% lower in affected individuals than in normal persons (Merke et al, 2000). In another study, those with CYP21A2 showed a significantly decreased catecholamine response to exercise that was unaffected by the administration of stress doses of glucocorticoids (Weise et al, 2004). It has been suggested that the degree of medullary impairment may be a biomarker for CAH severity (Merke et al, 2002).

11-β-Hydroxylase Deficiency

The second most common enzyme deficiency of the adrenal cortex, accounting for about 7% of all cases of CAH, is the 11-β-hydroxylase (11-OH) deficiency. A defect in this enzyme blocks the final conversion of 11-deoxycortisol to cortisol and DOC to corticosterone. As with 21-OH deficiency, a compensatory increase in ACTH secretion leads to adrenal hyperplasia and a mass action shunting of precursor steroids toward testosterone synthesis, resulting in signs of virilization. This block also results in the accumulation of DOC; the mineralocorticoid activity of DOC leads to the development of hypertension and hypokalemia, similar to what is seen with hyperaldosteronism (Fig. 24-13).

11-OH deficiency is an autosomal recessive disorder caused by mutations of the genes CYP11B1 and CYP11B2, located on chromosome 8q21-q22 (White et al, 1994b). Diagnosis in the neonate is established by the presence of a high basal and a high ACTH-stimulated 11-deoxycortisol. Concentrations of urinary tetrahydro-11-deoxycortisol in the urine are also elevated. During childhood and in young adults, the diagnosis of 11-OH deficiency is made by the presence of elevated early-morning and ACTH-stimulated serum levels of 11-deoxycortisol that are more than three times the upper limit for age-matched normals. Levels of DOC and adrenal androgens (androstenedione, dehydroepiandrosterone, and dehydroepiandrosterone sulfate) are also elevated. Plasma renin activity and aldosterone are often suppressed as a result of salt and water retention induced by elevations of DOC. Unlike the heterozygotes with 21-OH deficiency, those with 11-OH deficiency often fail to show a rise in precursors following ACTH stimulation (Pang et al, 1980). However, an exuberant response was seen in those who had hirsutism (Gabrilove et al, 1965).

Prenatal diagnosis of 11-OH deficiency is made by measuring levels of tetrahydro-11-deoxycortisol (THS) in the maternal urine or amniotic fluid. These levels begin to rise in the first trimester. In addition to THS, elevation of 11-deoxycortisol levels and of the THS to tetrahydrocortisol plus tetrahydrocortisone ratio is seen (Rosler et al, 1988).

Treatment consists of the replacement of glucocorticoids, causing normalization of DOC and plasma renin activity.

3-β-Hydroxysteroid Dehydrogenase Deficiency

3-β-hydroxysteroid dehydrogenase (3-β-HSD), which catalyzes the second enzymatic step in steroidogenesis, is coded by two genes, HSD3BI and HSD3BII. HSD3BII is expressed in the adrenals and gonads; HSD3BI is expressed in placenta, skin, and other peripheral tissues and usually remains functionally intact at these sites in CAH. A defect in 3-β-HSD leads to a block in the conversion of Δ5 steroids (pregnenolone, 17-OH pregnenolone, and dihydroepiandrosterone) to Δ4 steroids (progesterone, 17-OHP, androstenedione), resulting in an increase in circulating levels of Δ5 steroids (Fig. 24-14). However, because HSD3BI is usually intact, levels of the Δ4 steroids may be normal or even elevated.

Patients with classic disease will have manifestations of glucocorticoid deficiency, with or without the accompaniment of salt wasting. Affected males have incomplete masculinization, and females may be normal or may have ambiguous genitalia. A late-onset variant has been described that is associated with features typical of polycystic ovarian disease, such as hirsutism, oligomenorrhea, and infertility (Pang et al, 1985).

Previous criteria for the diagnosis of 3-β-HSD deficiency examined the basal and ACTH-stimulated Δ4/Δ5 steroid ratios, 17-OH-pregnenolone (17-OHP)/cortisol, and levels of pregnenolone, 17-OH-pregnenolone, and dehydroepiandrosterone in urine and blood. The standard for diagnosing this disorder has been recently revised to correlate more closely with genotypic studies. The ACTH-stimulated Δ5-17 P levels and Δ5-17 P/cortisol ratios have been shown to be the best indices for definitively diagnosing 3-β-HSD deficiency (Lutfallah et al, 2002). Applicability of these new criteria to patients with the nonclassic variant is still debatable.

Treatment consists of glucocorticoids and mineralocorticoids, as well as sex steroids, in accordance with normal growth and development.

17-Hydroxylase Deficiency

17-Hydroxylase (CYP17, P450_c17) is expressed in the adrenal glands and the gonads and encodes two enzymes—17α-hydroxylase and 17,20-lyase. 17α-hydroxylase catalyzes the conversion of pregnenolone and progesterone to their respective 17-OH derivatives. 17,20-Lyase converts 17-OH pregnenolone to dihydroepiandosterone, and 17-OH progesterone to androstenedione. CYP17 deficiency blocks the conversion of pregnenolone and progesterone to the 17-hydroxy derivatives, causing shunting from testosterone and cortisol synthesis to aldosterone (Fig. 24-15). Accordingly, these patients develop hypertension and hypokalemic alkalosis in association with incomplete masculinization (in the male) and decreased testosterone and cortisol levels. 17-OH deficiency is diagnosed by demonstrating high DOC, pregnenolone, and progesterone levels, along with decreased urinary 17-ketosteroids and 17-hydroxycorticosteroids. The gene associated with this condition (CYP17) has been located on chromosome 10q, the same gene as 17,20-desmolase (Kater & Biglieri, 1994).

Congenital Lipoid Adrenal Hyperplasia

Congenital lipoid adrenal hyperplasia (lipoid CAH) is the most severe form of CAH, in which the synthesis of all gonadal and adrenal cortical steroids is markedly impaired. Lipoid CAH may be caused by a defect in the StAR or the P450 side chain cleavage (scc) (Fujieda et al, 2003). StAR, located on chromosome 8p11, controls the rate-limiting step in steroidogenesis. It is responsible for the shuttling of cholesterol from the outer to the inner mitochondrial membrane. 20,22-Desmolase (CYP11A1, P450scc) converts cholesterol to pregnenolone (Fig. 24-16). Pathologically, the adrenal cortex shows marked accumulation of cholesterol and other lipids, which is the primary distinguishing feature from congenital adrenal hypoplasia.

The presentation of this extremely rare disorder is that of severe adrenal insufficiency with hypotension, salt wasting, and feminization of external genitalia in males. Occasionally, females may not present until the onset of puberty. Diagnosis is made by the presence of extremely low cortisol and aldosterone concentrations and elevated ACTH and plasma renin activity.

Aldosterone Synthetase Deficiency

Aldosterone synthetase (CYP11B2) is the final step in the steroid synthetic pathway leading to the production of aldosterone. It stimulates a multistep process: 11-hydroxylation of DOC to corticosterone, 18-hydroxylation of corticosterone to 18-hydroxycorticosterone, and, finally, 18-dehydrogenation to aldosterone. Isolated enzyme deficiency leads to salt wasting, hyperkalemia, and metabolic acidosis. This condition may be diagnosed by demonstrating the presence of metabolites of corticosterone and 11-deoxycorticosterone in the urine, elevated serum DOC, and deficiency of corticosterone, 18-hydroxycorticosterone, or aldosterone in the serum.

Laboratory Measurement of ACTH

Plasma ACTH is measured using a two-site immunoradiometric assay or an immunochemiluminometric assay. To prevent degradation of ACTH, it is best to collect the sample in a prechilled EDTA lavender-top tube. The specimen should be kept in an ice bath and should be processed as soon as possible. Following centrifugation in a refrigerated centrifuge, the specimen should be separated, transferred to a plastic tube, and kept frozen at –20° C until time of analysis. The normal reference range is 2 to 12 pmol/L (9 to 52 pg/mL) between 7 AM and 10 AM. Plasma ACTH is a useful tool for distinguishing primary (adrenal) from secondary (pituitary) or tertiary (hypothalamic) adrenal insufficiency. In primary adrenal insufficiency, low cortisol concentrations are found, along with increased ACTH levels. In secondary or tertiary adrenal insufficiency, both ACTH and cortisol are expected to be low. ACTH levels best discriminate between healthy individuals and those with adrenal insufficiency when specimens for ACTH are collected between 8 AM and 10 AM.

ACTH is less useful in diseases of cortisol excess. Up to 50% of patients may have normal ACTH levels in Cushing's disease. Although the values tend to run higher (>20 pmol/L [>90 pg/mL]) in Cushing's syndrome due to ectopic ACTH production, the values overlap with those seen in Cushing's disease in 30% of cases (Findling, 1992). Because of the normal diurnal variation in ACTH and cortisol secretion, measurement of these values during their expected nadir, from 11 PM to 1 AM, is helpful in confirming the diagnosis of ACTH-dependent Cushing's disease. An elevated midnight ACTH of greater than 5 pmol/L (23 pg/mL) in the face of an elevated serum cortisol confirms the diagnosis of ACTH-dependent Cushing's disease. Those patients with ectopic ACTH-secreting tumors characteristically have markedly elevated plasma ACTH (usually >200 pg/mL) and elevated serum cortisol. Measurement of ACTH by plasma extraction, which detects ACTH precursors and fragments, may be useful in distinguishing patients with cancer-related syndromes or ACTH-secreting tumors from those with Cushing's disease, as the former entities are more likely to produce these other forms of ACTH (White & Clark, 1993). ACTH-secreting neoplasms may be occult, creating diagnostic difficulties; in this instance, ACTH measurement using selective venous sampling has proven useful in localization of the lesion.

In patients with increased levels of circulating glucocorticoids due to an adrenal adenoma or carcinoma, and those surreptitiously taking steroids, ACTH secretion is inhibited and levels are low or undetectable. In patients with pituitary-induced adrenal hyperplasia (Cushings' disease), plasma ACTH may be at or above the upper reference interval at 9 AM, and fail to show the expected fall after midnight. Another use of ACTH assays is in the determination of adequacy of cortisol replacement in congenital adrenal hyperplasia. When replacement therapy is optimal, ACTH values are similar to those seen in a reference population.

Plasma Corticotropin-Releasing Hormone

CRH measurements are performed by liquid chromatography and tandem mass spectrometry (LC/MS-MS) and remain largely a research tool.

CRH circulates in one of two forms: free, or bound to CRH-binding protein. The reference range for plasma CRH in men and nonpregnant women is <34 pg/mL. CRH is increased in Cushing's syndrome due to ectopic production of CRH. CRH-binding protein increases during pregnancy; as a result the normative range for CRH varies throughout pregnancy: first trimester <40 pg/mL; second trimester <153 pg/mL; and third trimester <847 pg/mL (Goland et al, 1986).

Serum Cortisol Measurements

About 90% of circulating cortisol is bound to serum protein, of which 10% to 20% is loosely bound to albumin, and the remainder is bound to the glycoprotein transcortin (cortisol-binding globulin [CBG]). The remaining 10% of circulating cortisol is the unbound, free hormone. It is believed that only free cortisol is active, and that the protein-bound fraction is metabolically inert, probably serving as a reservoir for free cortisol. Protein binding also may protect cortisol from deactivation by the liver or filtration by the kidney.

One of the earliest and simplest methods used to determine the serum cortisol concentration was a fluorometric assay developed by Nelson and Samuels that was based on a technique first developed by Porter and Silber (Porter & Silber, 1950; Nelson & Samuels, 1952). The subsequent development of antibody-based immunoassays provided a more specific method for cortisol estimation with the added advantages of requiring a smaller specimen volume and a more rapid turnaround time. The problem was that some of the antibodies used showed a large degree of cross-reactivity with other steroid species, resulting in spuriously elevated cortisol concentrations. In chronic renal disease, various steroids and their glucuronides accumulate in the blood; cross-reactivity between these steroids and their conjugates with the antibodies used in the assay can yield significantly erroneous cortisol concentrations. Similarly, in CAH, high concentrations of cortisol precursors are seen in the serum because of an enzyme defect. Because these precursors cross-react with assay antibodies, artificial elevations of cortisol are found. The degree of interference varies with the assay used and cannot be easily predicted. Nonisotopic immunoassay methods using organometallic tracers, fluorescence polarization, and enzyme immunoassay techniques have also been developed for cortisol determinations (Bacarese-Hamilton et al, 1992; Lentjes et al, 1993; Philomin et al, 1994). The major disadvantage of all of these cortisol assays continues to be lack of specificity.

The measurement of cortisol by radioimmunoassay (RIA) and chemiluminescent techniques has largely been supplanted by high-performance liquid chromatography (HPLC) with tandem mass spectrometry, which appears to offer the ultimate in specificity. Serum cortisol is collected in a no-additive (red-top) tube. Reference values for serum cortisol for men and women roughly range from 5 to 25 µg/dL (140 to 690 nmol/L) at 8 AM to 10 AM, dropping to about 3 to 12 µg/dL (80 to 330 nmol/L) by 4 PM. Because of wide swings in basal cortisol levels resulting from its diurnal and ultradian pattern of secretion, serum cortisol assays are most useful when evaluated in the context of dynamic manipulation (i.e., adrenal stimulation or suppression).

Salivary Cortisol

Up to 30% of urinary free cortisol (UFC) and dexamethasone suppression screening tests may return an incorrect result. Recent studies have shown that the use of a midnight salivary cortisol (MSC) is a viable alternative. CBG is not present in saliva; therefore the results may be more useful in situations in which CBG concentration is altered. Salivary cortisol is highly stable and easy to collect, making it a useful tool for screening and for diagnosing instances of increased cortisol secretion. It does not appear to be affected by the rate of saliva production and can reflect a change in serum cortisol in as rapid as a few minutes (Read et al, 1990). Because of alterations in circadian rhythm, it may not be an appropriate test for shift workers or those with highly variable bed times. One study compared the sensitivity for the detection of Cushing's syndrome by nighttime salivary cortisol levels versus that for simultaneous inpatient serum cortisol levels and urine glucocorticoid excretion. It was found that the salivary cortisol measurements worked as well as plasma measurements and better than urine glucocorticoid excretion. The authors concluded that measurement of bedtime salivary cortisol was a practical and accurate screening test for the diagnosis of Cushing's syndrome (Papanicolaou et al, 2002). Another study compared the diagnostic performance of MSC measurement versus that of midnight serum cortisol (MNC) and UFC in differentiating 41 patients with Cushing's syndrome from 33 with pseudo-Cushing's states, 199 with simple obesity, and 27 healthy normal weight volunteers. In the whole study population, no statistically significant differences in terms of sensitivity, specificity, diagnostic accuracy, and predictive values were observed among tests. In particular, the overall diagnostic accuracy for MSC was similar to those of UFC and MNC (Putignano et al, 2003; Elamin et al, 2008). A recent review of the literature has shown that elevated late-night (11 PM to 12 AM) salivary cortisol has greater than 90% sensitivity and specificity for the diagnosis of endogenous Cushing's syndrome (Raff, 2009). It is recommended to have the patient avoid tobacco on the day of specimen collection.

Urinary Free Cortisol Measurements

Only 1% of the total adrenal secretion appears in the urine as cortisol, but it is this fraction that provides valuable aid in the diagnosis of adrenal disease. In the kidney, glomerular filtration of free cortisol (unbound to CBG and albumin) is followed by passive tubular reabsorption without a demonstrable reabsorption maximum. At serum cortisol levels of about 20 to 25 µg/dL (the upper 8 AM reference value), the binding capacity of transcortin (CBG) is exceeded, leading to a very rapid and disproportionate increase in the unbound fraction compared with the total serum cortisol. Doubling of serum cortisol from 20 to 40 µg/dL results in at least a fivefold increase in unbound cortisol. At these levels, free cortisol clearance by the kidneys is directly proportional to the unbound serum cortisol concentration, and a steep rise in cortisol clearance is seen. Serum cortisol reflects the sum of free, CBG-bound and albumin-bound cortisol. Only free (unbound) cortisol is filtered by the kidneys; therefore unlike serum cortisol, UFC is not affected by conditions and medications that alter CBG and albumin concentrations. Thus, when UFC rather than serum cortisol is used, it is easier to differentiate patients with adrenal hyperfunction from a reference population.

UFC is the best specimen to submit because it provides an integrated profile of total cortisol secretion over a 24-hour period, which is most helpful in those with sporadic excess cortisol production. The collection should also be assayed for creatinine to ensure that an adequate specimen has been submitted. The reliability of the test may be further improved by submitting urine collected over 2 or 3 days because day-to-day fluctuations in cortisol excretion are known to occur. Urinary free cortisol levels are unaffected by alterations in hepatic metabolism of cortisol. Although total cortisol production and urinary 17-OHCS may be increased, serum cortisol and urinary free cortisol remain within the reference interval. Because renal clearance of cortisol is dependent on normal kidney function, it is not surprising that patients with renal disease have low UFC values. UFC is unreliable when CrCl is <20 mL/min, and is of reduced reliability when CrCl is <60 mL/min (Chan et al, 2004). Increased serum concentration of transcortin during pregnancy and with estrogen therapy results in increased serum cortisol levels. This increase is not reflected by an elevation of cortisol metabolites in urine, but urinary free cortisol may be increased. Conditions in which spuriously elevated levels occur include starvation, use of topical steroids, and perhaps hydration in the form of water loading.

Hypercortisolism: Cushing's Syndrome

Cushing's syndrome is composed of a group of clinical and metabolic disorders resulting from prolonged exposure to elevated concentrations of glucocorticoids. The excessive levels of glucocorticoids may be of endogenous (secreted by the adrenal zona fasciculata), or far more commonly, of exogenous (e.g. pharmacologically administered steroids) origin. Patients with severe forms of the syndrome are easily recognizable by their florid presentation (e.g. striae, facial plethora, proximal muscle weakness, and stereotypical accumulations of fat). Many of the signs and symptoms due to hypercortisolism are common medical complaints (e.g. fatigue, weight gain, obesity, diabetes, and hypertension), however it is important to suspect Cushing's in certain situations as these have an increase in morbidity and mortality. Persons who may fall into this category and who should undergo screening include those with osteoporosis, hypertension, or diabetes occurring at a younger-than-expected age; patients with multiple, progressive features suggestive of Cushing's; and those with incidentally discovered adrenal adenomas (Nieman et al, 2008). Patients with ectopic ACTH-producing tumors have elevated ACTH and glucocorticoid levels. However, because of the rapid growth of these tumors, they usually die before clinical signs of the syndrome can manifest. Features most commonly appearing in this group include hypokalemia and proximal muscle weakness (Salgado, et al 2006).

Laboratory findings in Cushing's syndrome include: (1) excessive production of cortisol measured as elevated serum cortisol, urinary free cortisol, or midnight salivary cortisol, (2) loss of circadian rhythm of ACTH and cortisol, and (3) loss of suppression of cortisol production by administration of the synthetic glucocorticoid dexamethasone.

Cushing's disease specifically refers to hypercortisolism due to an ACTH-secreting pituitary adenoma. Cushing's syndrome is a global term that encompasses a wide variety of entities associated with hypercortisolemia.

Cushing's syndrome is most commonly iatrogenic in origin; however, it may also be due to a primary adrenal malignancy or ectopic production of CRH or ACTH by a tumor. In adrenal Cushing's, excess cortisol is autonomously produced by the adrenal glands, resulting in suppression of the hypothalamic-pituitary axis. Adrenal Cushing's (adenoma or carcinoma) accounts for less than 20% of cases, whereas pituitary Cushing's accounts for about 68%, and ectopic production of ACTH (outside the pituitary-adrenal axis) is the cause in about 12% of cases (Orth, 1995). Because the treatment and prognosis differ depending on the underlying etiology, it is important that a specific diagnosis be reached.

Evaluation of a patient suspected of having Cushing's syndrome begins with use of one of the three screening tests. If one of the tests returns as borderline, or the suspicion for the diagnosis remains high, a different screening test should be performed. The second phase includes confirmation of the diagnosis and identification of the pathophysiologic process causing the hypercortisolism.

Three screening tests are used in the evaluation of a patient suspected of having Cushing's syndrome (Fig. 24-17): the 24-hour urinary free cortisol, the overnight dexamethasone suppression test (DST), and the plasma or salivary midnight cortisol (MSC) level. The 24-hour urinary free cortisol is a reflection of the unbound circulating cortisol that is freely filtered by the glomerulus. Unlike serum cortisol, it is unaffected by the level of circulating CBG. HPLC or gas chromatography coupled with tandem mass spectrometry provides the best specificity for measuring urinary free cortisol. Unlike RIA or enzyme-linked immunosorbent assay, these techniques are not affected by cross-reactivity with steroid metabolites or synthetic glucocorticoids. The upper range of normal with these methods is 110 to 138 nmol/24 hours (40 to 50 µg/hour) (Raff & Findling, 2003). The creatinine should be measured in all collections to ensure the adequacy of the specimen. Urinary cortisol excretion is decreased when the glomerular filtration rate is <30 mL/minute, and thus may be normal despite the presence of excessive cortisol production (Arnaldi et al, 2003). It has been suggested that the 1-mg overnight DST should be the preferred screening test used for these individuals (Nieman et al, 2008). Because most pediatric patients being evaluated for Cushing's are of adult weight (>45 kg), adult normal ranges may be used (Nieman et al, 2008).

As per the latest guidelines for the diagnosis of Cushing's syndrome, any UFC concentration above the upper limit of normal for that particular assay should be considered a positive test (Nieman et al, 2008). Values greater than four times the upper limit of normal are diagnostic of Cushing's syndrome. From 10% to 15% of patients with Cushing's syndrome will have at least one in four 24-hour urine collections for free cortisol return as normal (Nieman & Cutler, 1990). If cortisol excretion is normal but clinical suspicion is high, the study should be repeated or a different screening method should be used. Milder elevations of UFC can be seen in pseudo-Cushing's and during normal pregnancy. Pseudo-Cushing's is an entity characterized by HPA axis overactivity but without true Cushing's syndrome. It has been described in depression, anxiety disorders, alcoholism, poorly controlled diabetes, and morbid obesity. Refer to the Pseudo-Cushing's section later in the chapter for more details. The overnight dexamethasone suppression test is a much simpler test to perform. The patient takes 1 mg of dexamethasone orally between the hours of 11 PM and 12 midnight. The plasma cortisol is drawn the following morning between 8 AM and 9 AM, respectively. The original criterion for an abnormal response was failure to suppress the morning cortisol level to <5 µg/dL (138 nmol/L); this has been revised downward to <1.8 µg/dL (50 nmol/L) (Findling & Raff, 1999; Arnaldi et al, 2003; Nieman et al, 2008). Using the lower cutoff of < 1.8 µg/dL (50 nmol/L) provides a greater than 95 % sensitivity and 80% specificity, and serves to minimize the number of false-positive results (Nieman et al, 2008). Failure to suppress could be due to Cushing's syndrome or pseudo-Cushing's, and may even occur in some patients who are normal. The false-positive rate can be as high as 30% for a variety of reasons: Dexamethasone was taken too early; the patient was on phenobarbital, phenytoin, or another medication known to accelerate the metabolism of dexamethasone; malabsorption; alcoholism; or morbid obesity. Pregnancy and drugs such as estrogen, which increase serum transcortin (aka CBG), may also result in elevated cortisol levels. Because of these and possibly other factors, about 1% of healthy individuals, 13% of obese patients, 50% of women on estrogens, and 25% of hospitalized and chronically ill patients show false-positive overnight dexamethasone suppression tests (DSTs). False-negative results, on the other hand, occur in less than 2% of patients.

The midnight salivary cortisol (MSC) concentration is a simple, convenient, and accurate means by which to diagnose Cushing's syndrome. MSC carries a high diagnostic sensitivity and specificity, and has been demonstrated to have an excellent correlation with the plasma free cortisol concentration (Papanicolaou et al, 2002; Yaneva et al, 2004; Alwani et al, 2014). Using a cutoff of 9.3 nmol/L on a commercial ELISA kit, MSC had a sensitivity of 100%, specificity of 83%, PPV of 94%, and NPV of 100% (Alwani et al, 2014). It performs slightly better than the 24-hour UFC in distinguishing patients with Cushing's syndrome from obese subjects (Yaneva et al, 2004).

Several studies describe the use of a late-night (11 PM to 12 AM) salivary cortisol (LNSC); others use the terms MSC and LNSC interchangeably, although the methodologies describe the samples as being collected at midnight. There are no studies comparing MSC with LNSC; however, because the data for MSC is more robust, it is the preferred test.

The patient can collect the sample at home. Two methods can be used to collect the sample: The patient can chew on a cotton pledget for 2 to 3 minutes and then place the pledget into a plastic tube; alternatively, the patient can passively drool directly into a test tube. The sample is then mailed to the reference laboratory; because salivary cortisol is very stable, there is little concern about degradation during transport. The sample can be analyzed either by immunoassay or liquid chromatography–mass spectrometry (LC-MS/MS). While immunoassay is more widely used, LC-MS/MS is useful in evaluating for surreptitious use of steroids and for steroid excess in those using inhaled steroids (Raff, 2013). Normative values vary according to the reference laboratory. Results should be interpreted with caution in those who may have blunting of their circadian rhythm (e.g., shift workers, patients with depression, critical illness). Falsely elevated results may be seen in smokers and users of chewing tobacco, in which case refraining from use for 24 hours prior to collection is recommended.

Pseudo-Cushing's Syndrome

Excess activity of the hypothalamic-pituitary axis, similar to that seen in pituitary Cushing's syndrome, has been demonstrated in some patients with alcoholism, major depression, poorly controlled diabetes, and obesity. Resolution of the underlying problem results in normalization of the HPA axis. These patients may not suppress on a low-dose DST and may have elevated UFC. Depressed patients who fail to suppress cortisol secretion in response to dexamethasone also show an impaired ACTH response to exogenous CRH, but will usually retain a normal cortisol response (Gold et al, 1986; Thalen et al, 1993). A single serum cortisol above 7.5 µg/dL at midnight has been shown to discriminate Cushing's syndrome from pseudo-Cushing's with 100% specificity (Papanicolaou et al, 1998). A more definitive test for distinguishing between these two entities is the combined dexamethasone-oCRH test. 1 µg/kg or 100 µg oCRH is infused 2 hours after a low-dose dexamethasone test. A plasma cortisol concentration greater than 38 nmol/L measured at 15 minutes following the administration of oCRH correctly identifies all cases of Cushing's syndrome and all cases of pseudo-Cushing's, with a specificity, sensitivity, and diagnostic accuracy of 100% (Yanovski et al, 1993).

ACTH Stimulation Test

The most convenient procedure for studying patients suspected of having hypocortisolism is the ACTH stimulation test. This test, which can be performed at any time of day, consists of drawing a baseline serum cortisol and then administering 250 µg of Cosyntropin (a commercially available ACTH analog) intravenously or intramuscularly. Serum for cortisol is again collected at 30 and 60 minutes post-Cosyntropin. A normal response is a cortisol level greater than 18 to 20 µg/dL (500 to 550 nmol/L) at either time point (Speckart et al, 1971; Burke, 1985). Drawing an ACTH level at baseline may help in distinguishing primary adrenal insufficiency (associated with an elevated ACTH usually greater than 50 to 100 pg/mL) from secondary or tertiary forms (which would have a low ACTH level less than 10 pg/mL) (Oelkers et al, 1992) (refer to the Laboratory Measurement of ACTH section earlier in the chapter). The aldosterone response to ACTH may also help in making this distinction; failure of aldosterone to increase by more than 4 ng/dL over baseline suggests primary adrenal dysfunction.

Steroids should never be withheld from a critically ill patient suspected of having adrenal insufficiency; instead a random sample for cortisol and ACTH should be drawn and treatment initiated with dexamethasone, which does not interfere with the cortisol assay. Formal testing should take place within 72 hours, before HPA axis suppression occurs. About 90% of cortisol circulates bound to CBG and albumin, and approximately 10% circulates freely (unbound). CBG declines during critical illness, resulting in an increase in the fraction of unbound (free) cortisol compared with the bound fraction. Because the serum total cortisol concentration does not accurately reflect this alteration in free/bound cortisol, it is common to overdiagnose adrenal insufficiency in the critical care setting. Since access to laboratories that readily and reliably measure free cortisol is limited, some have advocated use of the free cortisol index (FCI) and calculated free cortisol (CFC) to better delineate those patients in true need of steroid replacement therapy (Beishuizen et al, 2001; Hamrahian et al, 2004; Ho et al, 2006). To obtain the FCI, it is necessary to measure the serum CBG. CFC is the total cortisol (nmol/L)/CBG (mg/L), with normal being a value above 12. Given that for the vast majority of institutions, determination of CBG requires that the sample be sent to a reference laboratory, it may be more prudent and certainly more expedient to simply send aliquots from t-0, t-+30, and t-+60 minutes post–ACTH administration for serum free cortisol, and provide steroid coverage pending these results (Fleseriu & Loriaux, 2009). Checking the ACTH level at t-0 will help to further classify adrenal insufficiency as being primary versus secondary or tertiary. Salivary cortisol, which is unaffected by changes in CBG, may eventually prove useful in this setting (Raff, 2009).

Some have supported the use of a 1-µg dose of ACTH instead of the full 250-µg dose for diagnosing adrenal insufficiency. The premise is that the massively supraphysiologic dose of 250 µg will cause those with partial adrenal insufficiency to mount a normal response, and that these individuals will fail to show an adequate rise in cortisol in response to a 1-µg test dose. This procedure has not gained wide acceptance, nor has the normal range been standardized; more importantly, the diagnosis of partial adrenal insufficiency remains in contention.

The ITT has long been considered the gold standard for assessing the HPA axis. Unlike the ACTH stimulation test, the ITT assesses the integrity of the entire HPA axis. If the HPA axis is intact, insulin-induced hypoglycemia stimulates the hypothalamus and the pituitary gland to secrete CRH and ACTH, respectively, which, in turn, leads to a rise in cortisol. This test is contraindicated in individuals with ischemic heart disease, in older adults (>70 years of age), during pregnancy, and in those with a history of seizures. Patients highly suspected of having central adrenal insufficiency (i.e., prior pituitary surgery or radiation therapy) should receive a lower dose (0.05 U/kg) of insulin, rather than 0.1 to 0.15 U/kg. Glucose and cortisol are measured at –15 minutes, at baseline, and then at 15, 30, 45, 60, 90, and 120 minutes following insulin administration. For the test to be valid, adequate hypoglycemia must be attained (signs and symptoms of hypoglycemia and glucose less than 40 mg/dL [2.2 mmol/L]). A normal response is a rise in the cortisol level to greater than or equal to 18 µg/L or 20 µg/L at any time during the test (Nelson & Tindall, 1978; Burke, 1985; Grinspoon & Biller, 1994). Preliminary work suggests there was no difference in peak GH or cortisol level attained when analog insulin was substituted for regular insulin for the ITT (Yuen et al, 2008).

The cortisol response to hypoglycemia can be predicted reliably by the response to acute ACTH stimulation—a safer, cheaper, and quicker test (Stewart, 2003). The ITT is used mainly as a second-line measure to further evaluate those patients who had a borderline response to the ACTH stimulation test.

In the days before the commercial availability of a reliable assay to measure ACTH, it was common to perform a prolonged ACTH stimulation test (ACTH infusion over a 2-day period) to distinguish central from primary adrenal insufficiency. In other hormone systems, lack of stimulation normally leads to upregulation of the end-organ receptors; however, the adrenal glands downregulate their ACTH receptors in response to lack of stimulation. A short ACTH stimulation test would typically produce an inadequate response. By priming the system with prolonged exposure to ACTH, those with hypothalamic disease display a delayed but normal rise in cortisol.

Overnight Metyrapone Test

Metyrapone has also been used to assess the integrity of the HPA axis. Metyrapone inhibits 11-β-hydroxylase, preventing the conversion of 11-deoxycortisol (compound S) to cortisol. Under normal circumstances, the metyrapone-induced drop in cortisol is detected at the level of the hypothalamus and pituitary gland; a compensatory increase in CRH and ACTH secretion occurs, stimulating steroidogenesis and leading to a buildup of 11-deoxycortisol, the hormone preceding the block. Patients with central adrenal insufficiency fail to increase CRH and/or ACTH, steroidogenesis is not stimulated, and 11-deoxycortisol fails to increase. Metyrapone 30 mg/kg is administered orally at midnight; blood for cortisol and 11-deoxycortisol is collected the following morning at 8 AM. A rise in 11-deoxycortisol to greater than 7 µg/dL (200 nmol/L) is normal (Fiad et al, 1994; Grinspoon & Biller, 1994). An abnormal response is defined by 11-deoxycortisol less than 7 µg/dL accompanied by cortisol less than 5 µg/dL. This test should be performed with caution as it may provoke an addisonian crisis in those with central hypocortisolism. Metyrapone is currently only available on a compassionate basis; as a result this test is rarely performed.

Corticotropin-Releasing Hormone Test

The oCRH test can be used not only to diagnose hypocortisolism, but to localize the site of damage (Schulte et al, 1984). It is a safe test; however, it is expensive. After blood is drawn for baseline ACTH and cortisol, 100 µg oCRH is administered intravenously. Blood is then collected for cortisol and ACTH every 15 minutes for 60 to 90 minutes. A normal response is seen as cortisol greater than 20 µg/dL. Basal ACTH levels and their change in response to oCRH are used to localize the site of pathology. Plasma corticotropin usually peaks at 15 to 30 minutes, and cortisol usually peaks at around 30 to 40 minutes post oCRH injection (Oelkers, 1996).

Both the bound and free cortisol concentrations show a progressive rise throughout pregnancy. With this in mind, it has been recommended that higher cutoffs be used to properly interpret the ACTH stimulation results in pregnancy (Jung et al, 2011). A random morning cortisol level less than 11 µg/dL, 16 µg/dL, and 22 µg/dL when measured during the first, second, and third trimester of pregnancy, respectively, should raise the suspicion of adrenal insufficiency (Jung et al, 2011). A serum cortisol 60 minutes post-ACTH of less than 25 µg/dL, 29 µg/dL, and 32 µg/dL for the first, second, and third trimester of pregnancy, respectively, is diagnostic of adrenal insufficiency (Lebbe & Arlt, 2013; Yuen et al, 2013).

Antibodies against 21-hydroxylase antigen should be assayed in those whose Addison's disease (primary adrenal insufficiency) does not have an obvious explanation. Positive antibodies confirm autoimmune adrenalitis; these patients should be screened and monitored for the development of other autoimmune disorders. The absence of antibodies does not rule out an autoimmune etiology, as titers usually drop in concurrence with adrenal destruction.

The use of steroids in the treatment of malignant, inflammatory, and immunologic disorders is a common iatrogenic cause of adrenal insufficiency. The degree of adrenal suppression is dependent on the specific glucocorticoid dose, duration, frequency, and route of administration. Assessment of adrenal function in patients being tapered off of exogenous steroids can be performed once they are tapered to the equivalent of a daily dose of 10 mg hydrocortisone. Have them omit the morning glucocorticoid dose, and measure the 8 AM cortisol level. If it is greater than 10 µg/dL, routine supplementation of steroids can be ended. Because the adrenal cortex lags behind the pituitary gland in recovery from steroid suppression, complete adrenal recovery can also be demonstrated by an appropriate rise in serum cortisol following an 8 AM Cosyntropin infusion.