Tests of Pituitary Function: Cushing’s syndrome


From Lewis Blevins, MD  –  Photo of Dr. Lewis Blevins  –  Cortisol is a stress hormone. Thus, physical or psychic stress can elevate cortisol levels in normal persons. Depression and alcoholism can elevate cortisol levels and cause what is called pseudo- Cushing syndrome. Birth-control pills can elevate cortisol binding globulin and lead to false elevations in total serum cortisol levels.

Most people with true Cushing syndrome have elevated cortisol levels. It is extremely rare to find a person with pathologic hypercortisolism who has repeatedly normal cortisol levels.

Periodic or cyclical Cushing’s syndrome is vanishingly rare. I have seen only one documented case in a quarter of a century. And I’ve seen a lot of patients with Cushing’s!

Cushing syndrome is a unique disorder where there are plenty of people who remain undiagnosed or have significant delays in diagnosis ……and there are plenty of people who have been misdiagnosed as having Cushing syndrome but who are not truly affected by the disorder.

Over the next few days, I will post a series of commentaries regarding my diagnostic approach to patients with Cushing’s syndrome. At the outset, I recommend proceeding in a stepwise fashion. We will first discuss screening tests, and then diagnostic tests, followed by differential diagnostic tests. Afterwards, I will briefly review radiological studies and how they are best employed in the evaluation of patients with suspected Cushing syndrome. The evaluation must be sequential and orderly. Taking shortcuts leads to errors in judgment and very costly mistakes in the management of affected patients.


Part two:  Screening Tests. 

Screening tests are designed to be sensitive but not specific. In other words, they are designed with the intent to pick up everybody who has a particular disease but, as a consequence of this, they will also pick up a few people who don’t have the disease but would have an abnormal test. So, if a screening test is positive you may not actually have a disease. If it is negative, since this is a powerful test, you probably don’t have that particular disorder being tested.

Remember. We are talking about screening tests. Not diagnostic tests. Diagnostic test will be covered in the next section.

Over the course of my career, there have been several tests touted as great screening test for Cushing’s. A few have proved to be useful in my practice. I will cover these. But first, let’s talk about what are not considered good screening tests. This would include a random serum cortisol and plasma ACTH levels. I will check those once I determine the patient has Cushing’s syndrome, and sometimes early in the evaluation, just to know where they are but I do not put any value on the numbers until they are taken in context with other definitively positive tests.

The overnight 1 mg dexamethasone suppression test is a valid screening test. Persons undergoing this test are required to take dexamethasone 1 mg at 11 PM and have a serum cortisol level obtained at 8 AM the next morning. Normal people suppress the cortisol production during this test whereas those with probable Pathologic hypercortisolism do not. It is important to keep in mind, however, that about 30% of normal individuals will have a false positive test. They must be sorted out with diagnostic tests.

Cortisol is secreted in a diurnal variation with highest levels occurring just before or after awakening, levels in the afternoon about half of the morning values, and levels prior to bedtime about 1/10 of the morning values. Thus, a “midnight” salivary cortisol, or a pre-bedtime salivary cortisol for those who do shiftwork, should be low. If it is elevated that is considered a positive screening test. Once again, some people will have false positive tests despite the fact they do not have any disease related to cortisol over production.

The 24 hour urine free cortisol is a reasonable screening test. If a patient has three separate samples where the results are normal they probably do not have hypercortisolism. In this patient’s thought to have cyclical Cushing’s syndrome, which is extremely rare, urine cortisol should be obtained when they are symptomatic. If the results are normal then they probably do not have the disease. One caveat about this test is that some patients with early recurrences of Cushing syndrome can have high normal levels. Results that are abnormal should be further evaluated with one or more diagnostic tests.

One final idea…… Bayesian statistics and theories of test performance, including pretest probability, apply to most medical tests and especially those relating to endocrinology and, in particular, the evaluation of possible Cushing’s syndrome. The basic principle is this: The more likely it is that a person has a disease process, based on an evaluation of clinical symptoms and physical signs and other factors, the more reliable and accurate are the test results. For example, one could have two patients with a 25-h urine cortisol of 100 micrograms, about twice the upper limit of normal. The one that does not look cushingoid and nonspecific symptoms probably does not have Cushing’s syndrome. Conversely, the person with rapid weight gain, depression, excessive hair growth, stretch marks, or muscle weakness probably does have Cushing’s syndrome. The analogy is that of the treadmill test. A young person with no cardiac history or risk factors who has an abnormal treadmill stress test probably does not truly have coronary disease whereas the 70-year-old smoker with diabetes and hypertension who has the same abnormal test likely has a serious problem with coronary artery disease. What all of this means is a tests should be used judiciously to screen for and then diagnose diseases. Test results are more meaningful when there are symptoms and signs of an underlying disease process. I have found all of this to be very true in regards to the evaluation of patients with possible Cushing’s syndrome where I see a lot of people who have been diagnosed with the disorder based on the finding of an elevated midnight salivary cortisol which, by itself, is a screening test and not a diagnostic test.


Part three:  Diagnostic tests

Diagnostic tests are meant to be definitive. Thus, they are more specific than screening tests but are not quite as sensitive either. Thus, normal people who don’t have the disease are usually eliminated from further testing by these tests but a few patients who have the disease will be missed.
I find a 24-hour urine free cortisol greater than 3 times the upper limit of normal to be diagnostic. It is best if there is more than one confirmatory 24-hour urine test. Levels between 2 and 3 times the upper limit of normal are considered diagnostic in the proper clinical context.

The dexamethasone suppressed CRH stimulation test, has developed at the National Institutes of Health, has been reliable in my hands using the established cutoffs. In this test, the patient takes dexamethasone 0.5 mg every 6 hours for 2 days and then undergoes a CRH stimulation test. Cortisol levels greater than 1.4 mcg/dL at the end of the test are diagnostic of pathologic hypercortisolism. This test has only failed me twice in over 20 years. Both patients had a history of Cushing’s disease and had prior surgery and were seeing me for early recurrences and tested normally with this test. Further study had proved that their tumors had recurred and we were able to offer appropriate treatment.

The formal 2 day dexamethasone suppression test is a good test when taken in consideration with other test results. In this test, one measures the degree of suppressed ability of serum lambda or urine cortisol in response to the same doses of dexamethasone used in the dexamethasone suppressed CRH stimulation test. Research has shown that about 15% of patients who do have Cushing’s will have a normal test result.

In my opinion, the very best test to sort out whether patient truly has Cushing’s syndrome when the other results are equivocal, and especially when the clinical findings are not strong, is to perform a salivary cortisol profile. This test evaluates the diurnal variation of cortisol secretion which is seen in normal individuals. That variation in cortisol secretion is lost in cases of true Cushing’s syndrome. Salivary levels are obtained hourly for the first 3 hours upon and after awakening, for another consecutive 3 hours about 8 hours later, and for the last 3 consecutive hours prior to bedtime.


Part four:   Differential diagnostic tests

Differential diagnostic tests are meant to be highly specific. Sensitivity doesn’t matter. The key is to be able to sort patient’s according to disease. These tests should only be performed after a diagnosis of hypercortisolism has been secured.

ACTH levels help determine whether the hypercortisolism is due to adrenal disease, in which case the ACTH levels are usually low, or pituitary or ectopic disease in which case the ACTH levels are usually elevated or inappropriately in the normal range. While there are exceptions, ACTH levels are usually less than 10 pg per millimeter in patients with adrenal diseases. Patients with ectopic ACTH production usually have ACTH levels that are higher than those with Cushing’s disease due to a pituitary adenoma.

The overnight 8 mg dexamethasone suppression test can be used to differentiate those patients have pituitary tumors from those who have adrenal disease or ectopic ACTH secretion. Dexamethasone, 8 mg, is taken at 11 PM. Serum cortisol levels were determined at both 8 AM the day before and after taking dexamethasone. The degree of suppression of cortisol usually enables treating physicians to, in conjunction with the other test results, differentiate between the different sites of disease causing hypercortisolism.

The straightforward CRH stimulation test can be useful in sorting out this patient to have pituitary tumors yet have partially suppressed their ACTH levels.

Magnetic resonance imaging demonstrates pituitary tumor in most patients who have hypercortisolism.

Inferior petrosal sinus sampling can be used in those patients who have ACTH dependent hypercortisolism to determine whether there is a central or pituitary source of ACTH secretion versus an ectopic source.

Octreotide scintigraphy can be used in attempt to try to define a neuroendocrine tumor that might be causing ectopic ACTH secretion.

Urine 5-HIAA, plasma catecholamines and metanephrines, urine catecholamines and metanephrines, and calcitonin levels can help evaluate for disorders that include carcinoid, pheochromocytoma, and medullary thyroid carcinoma that can all secrete ACTH causing the syndrome of ectopic ACTH hypersecretion.


Part Five:  In summary

The only way a patient with Cushing’s syndrome is ever going to be diagnosed is to be tested!

Screening tests are appropriate to evaluate large populations. For example, in patients with uncontrolled type 2 diabetes mellitus, anywhere between 1.5 and 3% will have Cushing’s. Thus, a screening test might be applied to that group or, for example, a group of hypertensive patients with hypokalemia, or patients with premature osteoporosis, etc.

If, however, a patient is thought to quite likely have Cushing’s syndrome it is best to proceed with diagnostic tests rather than screening tests.

Diagnostic tests should really be performed, and the diagnosis secured, before any differential diagnostic tests, and especially radiological studies, are ordered or performed.

Determination of the plasma ACTH concentration is of paramount importance as it allows one to classify patients according to adrenal or non-adrenal causes of hypercortisolism.

Those with ACTH dependent hypercortisolism should undergo an MRI of the pituitary gland. If there is a pituitary tumor seen on MRI then, in most cases, the patient has a 97-99% likelihood of having an ACTH producing pituitary tumor and further testing is not required.
If the MRI fails to show a pituitary adenoma then inferior petrosal sinus sampling is required to determine whether there might be a pituitary lesion or the syndrome of ectopic ACTH hypersecretion.

I usually reserve performance of octreotide scintigraphy for those patient’s proved to have ectopic ACTH hypersecretion based on inferior petrosal sinus sampling.

I was once taught that it had been estimated that only 5 of 100 patient’s tested for Cushing’s syndrome will prove to have the disorder.

The most common cause of someone looking cushingoid is exogenous steroid use or exposure. This could be due to steroids injected into the body, applied topically, taken orally, and even inhaled. Physicians see drug induced Cushing syndrome on a regular basis. It surprises me that they missed patients who have endogenous Cushing’s syndrome. I think it might be because they are so used to seeing patients with lung disease on steroids, or skin problems taking topical steroids, and those receiving steroid injections for spinal disc disease, etc. Then, the patient with a pituitary tumor or adrenal disease walks in, they look like other patients with steroids, and it’s almost like the blinders are on and a workup is not pursued.

About 15% of patients with endogenous Cushing’s syndrome have adrenal disease.

Of those who have ACTH dependent hypercortisolism, approximately 15% of patients have the syndrome of ectopic ACTH hypersecretion while 85% harbor a pituitary tumor. Eighty per-cent of these are microadenomas while 20% are macroadenomas.

The most common cause of ectopic ACTH hypersecretion as a bronchial carcinoid tumor.

It is better to be sure than to be sorry. Do not allow your doctor to take short cuts in the evaluation and management of Cushing’s syndrome. Don’t be in a hurry to get a diagnosis or treatment. Let time be on your side.


© 2014, Pituitary World News. All rights reserved.