Cushing’s syndrome is caused by excessive body levels of adrenal glucocorticoids such as cortisol, either from (primary) adrenal cortex overproduction or from (secondary) therapeutic administration. This discussion will consider only the primary type due to excess adrenal production of cortisol. About 70% of cases (range 50%-80%) of Cushing’s syndrome due to adrenal overproduction of cortisol are caused by pituitary hypersecretion of ACTH leading to bilateral adrenal cortex hyperplasia. About 10% of cases are due to adrenal cortex adenoma, about 10% to adrenal cortex carcinoma, and about 10% to “ectopic” ACTH production by tumors outside the adrenal or pituitary glands, most commonly lung bronchial carcinoids (28%-38% of ectopic tumor cases) with the next most frequent being lung small cell carcinomas. A few cases are caused by thymus carcinoids, pancreatic islet cell tumors, pheochromocytomas, and various adenocarcinomas. One additional category is the uncommon syndrome of micronodular cortical hyperplasia, which biochemically behaves in a similar manner to adrenal cortex adenoma. Adrenal tumor is the most frequent etiology in patients younger than 10 years, and pituitary hyperactivity is the most common cause in patients older than 10. Cushing’s syndrome must be differentiated from Cushing’s disease, which is the category of Cushing’s syndrome due to pituitary hypersecretion of ACTH (usually due to a basophilic cell pituitary adenoma or microadenoma). The highest incidence of Cushing’s syndrome is found in adults, with women affected 4 times more often than men. Major symptoms and signs include puffy, obese-looking (“moon”) appearance of the face, body trunk obesity, “buffalo hump” fat deposit on the back of the neck, abdominal striae, osteoporosis, and a tendency to diabetes, hirsutism, easy bruising, and hypertension.

Standard test abnormalities

General laboratory findings include impairment of glucose tolerance in about 85% of patients (literature range, 57%-94%) that is severe enough to be classified as diabetes mellitus in about 25%. There is lymphocytopenia (usually mild) in about 80%, but most patients have an overall mild leukocytosis. Hemoglobin tends to be in the upper half of the reference range, with polycythemia in about 10% of affected persons. About 20%-25% have a mild hypokalemic alkalosis. The serum sodium level is usually normal but is slightly increased in about 5%. Total circulating eosinophils are usually decreased.

Screening tests

Urine 17-Ketosteroid assay. The urine 17-KS assay was one of the first tests used for diagnosis of Cushing’s syndrome. However, urine 17-KS values are increased in only about 50%-55% of patients with Cushing’s syndrome, and the test yields about 10% false positive results. Thus, 17-KS assay is no longer used to screen for Cushing’s syndrome. The 17-KS values may be useful in patients who are already known to have Cushing’s syndrome. About 45% of patients with adrenal adenoma and about 80%-85% (range 67%-91%) of patients with adrenal carcinoma have elevated urine 17-KS values. Patients with adrenal carcinoma tend to have higher urine 17-KS values than patients with Cushing’s syndrome from other etiologies, so that very high urine 17-KS values of adrenal origin suggest adrenal carcinoma.

Single-specimen serum cortisol assay. Laboratory diagnosis of Cushing’s syndrome requires proof of cortisol hypersecretion. For some time, assay of 17-OHCS in a 24-hour urine specimen or a single-specimen plasma 17-OHCS assay by the Porter-Silber method was the mainstay of diagnosis. However, in Cushing’s syndrome this technique yields about 15% false negative and 15% false positive results. The 17-OHCS values in urine are increased in some patients by obesity, acute alcoholism, or hyperthyroidism, whereas the 17-OHCS values in plasma are increased in many patients by stress, obesity, or an increase in cortisol-binding protein due to estrogen increase (oral contraceptive medication or pregnancy). Therefore, urine 17-OHCS assay and single determinations of plasma 17-OHCS are no longer considered reliable enough to screen for Cushing’s syndrome. Plasma or urine 17-OHCS assay was also used to measure adrenal response in stimulation or suppression tests. However, it has been replaced for this purpose by serum cortisol assay, which is technically easier to do and avoids the many problems of 24-hour urine specimen collections.

Single determinations of plasma cortisol, either in the morning or in the afternoon or evening, have the same disadvantages as plasma 17-OHCS and are not considered reliable for screening of Cushing’s syndrome. For example, single morning specimens detect about 65% of patients with Cushing’s syndrome (range, 40%-83%) and produce false positive results in about 30% of cases (range, 7%-60%). One report indicates that 11 P.M. or midnight specimens provide better separation of normal persons from those with Cushing’s syndrome.

Plasma cortisol diurnal variation. If plasma cortisol assay is available, a better screening test for Cushing’s syndrome than a single determination consists of assay of two plasma specimens, one drawn at 8 A.M. and the other at 8 P.M. Normally there is a diurnal variation in plasma levels (not urine levels), with the highest values found between 6 and 10 A.M. and the lowest near midnight. The evening specimen ordinarily is less than 50% of the morning value. In Cushing’s syndrome, diurnal variation is absent in about 90% of patients (literature range, 70%-100%). False positive results are obtained in about 20% of patients (range, 18%-25%). Therefore, significant alteration of the diurnal pattern is not specific for Cushing’s syndrome, since it is found occasionally in patients with a wide variety of conditions. Some of the conditions that may decrease or abolish the normal drop in the evening cortisol level in some persons are listed in the box on this page. Therefore, a normal result (normal circadian rhythm) is probably more significant than an abnormal result (although, as already noted normal plasma cortisol circadian rhythm may be present in about 10% of patients with Cushing’s syndrome).

Urine free cortisol. About 1% of plasma cortisol is excreted by the kidney in the original free or unconjugated state; the remainder appears in urine as conjugated metabolites. Original Porter-Silber chromogenic techniques could not measure free cortisol selectively. Fluorescent methods or immunoassay can quantitate free cortisol, either alone or with compound S, depending on the method. Immunoassay is becoming the most frequently used technique. Urine free-cortisol values in 24-hour collections are reported to be elevated in about 95% of patients with Cushing’s syndrome (literature range, 90%-100%) and to produce false positive elevation in about 6% of patients without Cushing’s syndrome (literature range, 0%-8%).

Urine free-cortisol levels may be elevated in some patients by some of the factors that affect blood cortisol, including severe stress, acute alcoholism, psychiatric depression, and occasionally patients with obesity. In cortisol-binding protein changes such as an increase produced by estrogens, most reports indicate that urine free-cortisol secretion levels are usually normal. Renal insufficiency may elevate plasma cortisol levels and decrease urine free-cortisol levels. Hepatic disease may increase plasma cortisol levels but usually does not affect urine free-cortisol levels significantly. The major difficulty with the test involves accurate collection of the 24-hour specimen. Also, the test is not performed in most ordinary laboratories and would have to be sent to a medical center or reference laboratory.

Some Conditions That Affect Serum Cortisol Diurnal Variation

Severe stress
Severe nonadrenal illness
Psychiatric depression
Alcoholism (especially with recent intake)
Change in sleep habits
Certain medications (prolonged steroids, phenothiazines, reserpine, phenytoin, amphetamines)

Single-dose dexamethasone suppression test. The most simple reasonably accurate screening procedure is a rapid overnight dexamethasone suppression test (DST). Oral administration of 1 mg of dexamethasone at 11 P.M. suppresses pituitary ACTH production, so that the normal 8 A.M. peak of plasma cortisol fails to develop. After 11 P.M. dexamethasone, normal persons and the majority of obese persons have 8 A.M. plasma cortisol values less than 50% of baseline (predexamethasone) levels. Many endocrinologists require suppression to 5 µg/100 ml (138 nmol/L) or less. The consensus is that about 95% of Cushing’s syndrome patients exhibit abnormal test response (failure to suppress), although there is a range in the literature of 70%-98%). There is an average of less than 5% false positive results in normal control persons (range, 1%-10%).

There is controversy in the literature regarding certain aspects of this test. Some investigators found substantial numbers of patients with a Cushingoid type of obesity, but without demonstrable Cushing’s syndrome, who failed to suppress adequately (falsely suggesting Cushing’s syndrome) after the overnight DST. This involved 10% of Cushingoid obese patients in one series and 53% in another. Unfortunately, there are not many reports in the literature that differentiate lean from obese persons in control series. Another controversial point is the degree that the 8 A.M. cortisol specimen must be suppressed from baseline value to separate normal persons from those with Cushing’s syndrome. Some have found the standard of a 50% decrease from baseline values to be insufficiently sensitive, missing up to 30% of Cushing’s syndrome patients. These investigators suggest a fixed 8 A.M. plasma cortisol value (after dexamethasone) of 5 or 7 µg/100 ml. However, establishment of such a fixed value is complicated by the variations in cortisol reference ranges found in different methods and kits. Another problem are conditions that may produce false results (failure to suppress normally). Some of these are listed in the box on this page.

Phenytoin and phenobarbital affect cortisol by affecting the microsomal metabolic pathway of the liver. Estrogen increases cortisol-binding protein values, which, in turn, increases total plasma cortisol values. This may affect the DST when a fixed 5 µg/100 ml cutoff limit is used, since the already increased cortisol level must be suppressed even more than usual to reach that value. Spironolactone is a fluorescent compound and interferes with the Mattingly fluorescent assay technique. Immunoassay is not affected. Additional evidence to support abnormal screening test results may be obtained by using the standard DST.

Some Conditions That Interfere With the Low-Dose Overnight Dexamethasone Suppression Test

Conditions producing false normal test results*
Drug-induced interference (phenytoin, phenobarbital, estrogens, possibly spironolactone)
Conditions producing false abnormal test results†
Acute alcoholism
Psychiatric depression
Severe stress
Severe nonadrenal illness
Obesity (some patients)
Renal failure
* Apparent suppression of 8 A.M. cortisol in patients with Cushing’s syndrome.
† Failure to suppress 8 A.M. cortisol in patients without Cushing’s syndrome.

The single-dose DST and diurnal variation test may be combined. Plasma cortisol specimens are drawn at 8 A.M. and 8 P.M. Dexamethasone is administered at 11 P.M., followed by a plasma cortisol specimen at 8 A.M. the next day.

Confirmatory tests

Confirmation of the diagnosis depends mainly on tests that involve either stimulation or suppression of adrenal hormone production. It is often possible with the same tests to differentiate the various etiologies of primary hyperadrenalism. Normally, increased pituitary ACTH production increases adrenal corticosteroid release. Increased plasma corticosteroid levels normally inhibit pituitary release of ACTH and therefore suppress additional adrenal steroid production. Adrenal tumors, as a rule, produce their hormones without being much affected by suppression tests; on the other hand, they tend to give little response to stimulation, as though they behaved independently of the usual hormone control mechanism. Also, if urinary 17-KS values are markedly increased (more than twice normal), this strongly suggests carcinoma. However, hyperplasia, adenoma, and carcinoma values overlap, and 17-KS levels may be normal with any of the three etiologies.