Screening for Cushing’s Syndrome in an Era of Epidemic Obesity

Abstract

Cushing’s syndrome is a rare cause of several very common conditions, including obesity, diabetes mellitus and hypertension. Screening tests for Cushing’s syndrome have generally high sensitivity and specificity, but if applied to unselected obese patients they have unacceptably high false positive rates. Only obese patients with more specific signs of the disorder should be screened for Cushing’s syndrome.

Introduction

Cushing’s syndrome is the group of symptoms and signs caused by excessive levels of cortisol or synthetic glucocorticoid drugs such as prednisone. Iatrogenic Cushing’s syndrome is common, due to treatment of inflammatory, autoimmune or allergic diseases with these drugs. Endogenous Cushing’s syndrome (usually due to tumors in the pituitary or adrenal) is rare. Estimates of its incidence range from 1 to 2.5 cases per million persons per year. It is associated with increased morbidity and mortality, primarily from cardiovascular disease.

Presentation

Many of the clinical findings in Cushing’s syndrome are very common and have many more likely potential causes. Nearly all patients with the syndrome are obese, primarily in the trunk, due to stimulation of appetite and the effect of glucocorticoids to promote deposition of visceral fat. Unlike most obese people, their arms and legs tend to be thin. Fat deposition in certain areas is characteristic, such as the dorsal fat pad (“buffalo hump”) and the supraclavicular fossae. The face tends to be round (“moon face”).

Most patients with Cushing’s syndrome have hypertension, and most affected women have hirsutism and abnormal menses. Psychological disorders, especially depression, may occur. Cortisol excess tends to raise plasma glucose, and glucose intolerance or diabetes mellitus may develop. Obviously, most people with hypertension, hirsutism, depression, and diabetes do not have Cushing’s syndrome. However, clinicians see many such patients every day and the question of who to screen for the disorder and how arises frequently.

More specific, diagnostically useful findings result from the catabolic effects of cortisol. Skeletal muscle weakness and atrophy cause difficulty climbing stairs or standing from a sitting position. Atrophy of connective tissue and loss of subcutaneous fat causes thinning of the skin, easy bruising (often with no recalled trauma), facial plethora, and wide, red or purple striae on the abdomen and elsewhere. Premature osteoporosis may cause back pain and fractures. The presence of any of these findings greatly increases the probability that obesity is due to Cushing’s syndrome.

Regulation of Cortisol Secretion

Diagnostic tests for Cushing’s syndrome are based on abnormalities in the regulation of cortisol secretion. Adrenal production of cortisol is stimulated by adrenocorticotropic hormone (ACTH) secreted by the pituitary gland, and most cases of endogenous Cushing’s syndrome are due to ACTH-secreting pituitary adenomas. ACTH release is in turn stimulated by hypothalamic production of corticotrophin releasing hormone (CRH). Three major mechanisms control ACTH and cortisol secretion:

1. Negative feedback suppression by cortisol of the secretion of both CRH and ACTH is a major regulator of overall cortisol secretion. Patients with Cushing’s syndrome have a partial or complete defect in this negative feedback loop that permits cortisol overproduction.

2. There is a marked circadian rhythm in ACTH and cortisol secretion, and secretion is pulsatile, resulting in wide variations in plasma cortisol levels throughout the day. Pulses of secretion begin in the early morning hours, and plasma cortisol levels usually reach a peak at about 8 A.M. They fall throughout the day, reaching a nadir at about 11 P.M. or midnight, when plasma cortisol is virtually undetectable in most normal individuals. This diurnal rhythm is disrupted in Cushing’s syndrome, and most patients have plasma cortisol levels at 11 P.M. that are above the very low normal range. Unfortunately, the lack of large differences in plasma cortisol levels during the hours when clinical laboratories are open means that the former practice of measuring plasma cortisol at 8 A.M. and 4 P.M. cannot reliably detect a loss of diurnal variation.

3. CRH, ACTH, and cortisol secretion are stimulated by stressful situations, such as illness or injury. High levels of cortisol are expected at such times, so it is unwise to attempt to diagnose Cushing’s syndrome until the stress has resolved.

Cortisol circulates reversibly bound to cortisol binding globulin (CBG). Only the unbound or free fraction is available to enter cells and produce biological effects, but conventional plasma cortisol assays measure total (predominantly bound) cortisol. Situations in which plasma CBG levels are increased cause misleading elevations in plasma cortisol. The most common of these situations is oral estrogen therapy, such as oral contraceptives.

Screening Tests for Cushing’s Syndrome

Three tests are in wide use for the initial evaluation of suspected Cushing’s syndrome: urinary cortisol excretion, overnight dexamethasone suppression testing, and late night salivary cortisol assays¹.

Twenty-four hour urinary free cortisol excretion reflects average plasma unbound cortisol throughout the day, with little overlap between normal values and those in Cushing’s syndrome. To assess the accuracy of the collection, creatinine should also be measured in the urine specimen. Reported sensitivity ranges from 94–100% and specificity from 93–100%². There is overlap between appropriately elevated values in sick patients and those in Cushing’s syndrome. Elevated values may also be seen in depression and alcoholism (sometimes called “pseudo-Cushing” states). Markedly elevated urinary cortisol (>4 times the upper limit of normal) establishes the diagnosis of Cushing’s syndrome.

The overnight dexamethasone suppression test takes advantage of the defective negative feedback suppression of cortisol secretion in Cushing’s syndrome. One mg of dexamethasone is taken at 11 P.M. and plasma cortisol is measured at 8–9 A.M. the next morning. The normal response is suppression of plasma cortisol to <2 mcg/dl. Reported sensitivity is 96–100% and specificity 81–90%².

The most recent addition to screening tests for Cushing’s syndrome is measurement of salivary cortisol³. Cortisol secretion into saliva is proportional to plasma free cortisol concentrations, and specimens can be obtained conveniently at home during the nadir of normal plasma cortisol levels at 11 P.M. This test takes advantage of the lack of normal diurnal rhythm of cortisol secretion in Cushing’s syndrome. Saliva is collected by chewing a cotton pledget for two minutes, and then spitting the pledget into a small container called a salivette. Salivary cortisol is stable for more than a week, and samples may be returned to the laboratory by mail. Because there is less experience with this test, a reliable reference laboratory should be used. Reported sensitivity is 92–100% and specificity 79–95%².

Fasting or random plasma cortisol measurements are useless in screening for Cushing’s syndrome and should not be used for this purpose.

Pitfalls in Screening for Cushing’s Syndrome

There are a number of possible causes for false results from screening tests for Cushing’s syndrome. Some apply to all of the screening tests above, while others apply to specific tests.

All diagnostic tests for Cushing’s syndrome may be falsely positive during stressful illness. As a rule of thumb, if a patient is sick enough to be hospitalized, they should not be tested for Cushing’s syndrome until they have recovered for at least two weeks. Pseudo-Cushing states such as alcoholism and major depression will also cause false positive results in any test.

The major artifact affecting urinary cortisol measurements is under- or over-collection. Excretion of an appropriate amount of creatinine assures that the collection is reasonably accurate. Renal failure may cause a falsely low urinary cortisol, so other tests should be used in this situation.

The overnight dexamethasone suppression test is subject to false positive results if the dexamethasone is not taken correctly (or at all), or if the morning plasma cortisol measurement is delayed. Some drugs such as phenytoin and carbamazepine accelerate the metabolism of dexamethasone and cause a false positive result because dexamethasone levels are too low to inhibit cortisol secretion. Finally, women taking oral contraceptives or other oral estrogens have elevated CBG levels and often fail to suppress plasma cortisol normally. Patients taking these drugs should be screened with one of the other two tests.

Salivary cortisol levels may be falsely elevated if the pledget used to collect saliva is touched by the patient, especially if they use topical steroid-containing creams. Patients with atypical daily schedules, such as shift workers, may have higher than normal 11 P.M. salivary cortisol. Smoking increases salivary cortisol and should be avoided on the evening of he collection. Finally, contamination of the sample with blood will cause a false positive result, so patients should not clean their teeth before collecting the sample.

Who Should be Screened for Cushing’s syndrome?

Although there have been reports describing Cushing’s syndrome in 1–5% of obese patients with diabetes or hypertension, such prevalence rates are very unlikely to be accurate. In a recent study in which 369 patients referred to a weight loss program were screened for Cushing’s syndrome, 23% had an abnormal screening test. After further definitive evaluation, none were found to have the syndrome⁴.

In another study, 201 consecutive patients attending a diabetes clinic were screened with salivary cortisol measurements. Again, 23% of patients had a cortisol value above a threshold considered to warrant further testing, and again no patients proved to have Cushing’s syndrome⁵.

These results indicate that in an unselected population of obese patients, the prevalence of Cushing’s syndrome is very low. The rate of false positive screening tests in such patients is unacceptably high, given the anxiety and cost of further testing generated by these false positive results. Thus, patients should be tested only if they have additional clinical characteristics that raise the pretest probability of the syndrome.

About 5% of patients with an incidentally discovered adrenal mass (an “incidentaloma”) have evidence of autonomous cortisol secretion and may have, or be at risk of developing, Cushing’s syndrome⁶. In such patients, the overnight dexamethasone suppression test is the most sensitive, since it can detect even slight degrees of excess cortisol production.

Recommendations for Screening

Patients with obesity, diabetes, hypertension or other nonspecific symptoms or signs compatible with Cushing’s syndrome should not be screened for this disorder unless they also have other more specific signs, since the imperfect specificity of available diagnostic tests will yield an unacceptably high rate of false positive results¹. Specific signs that should prompt the clinician to screen for Cushing’s syndrome are thin skin, easy bruising (especially if unexplained bruises are actually present on examination), red or purple striae, facial plethora, proximal muscle weakness manifested as the inability to rise from a chair without the use of the arms, and unexplained osteoporosis in men or premenopausal women.

Any of the three tests described above may be used. Patients with an abnormal screening test require further evaluation to confirm or exclude the diagnosis, ideally in consultation with an endocrinologist.

If Cushing’s syndrome is suspected, a thorough history of possible exposure to exogenous glucocorticoids should be taken. Possible sources include topical, inhaled and injected steroids, as well as “nutritional supplements”, some of which have been found to contain glucocorticoids.

Finally, patients with an incidentally discovered adrenal nodule should be screened with an overnight dexamethasone suppression test.

Conclusions

Obesity and its associated conditions may resemble Cushing’s syndrome, but screening for this disorder should be limited to those with signs linked to glucocorticoid excess in Table 1 shown here.

Table 1.

Signs that May Prompt Screening for Cushing’s Syndrome

Thin skin Easy bruising Red or purple striae Facial plethora Proximal muscle weakness Unexplained osteoporosis

Biography

William E. Clutter, MD is Associate Professor of Medicine in the Division of Endocrinology, Metabolism and Lipid Research at Washington University in St. Louis.

Contact: wclutter@dom.wustl.edu