This is a very personal commentary, based on over 25 years of work with patients with pituitary and adrenal problems, particularly Cushing’s syndrome and Cushing’s disease. I have also been fortunate to conduct many research and treatment studies in patients with pituitary disorders and to be part of a dedicated Neuroendocrine (Pituitary) group of Neuroendocrinologists and Neurosurgeons who are dedicated to care of patients with pituitary disorders, many of whom have Cushing’s disease.
As a Neuroendocrinologist who cares for many patients with Cushing’s disease I have observed 2 distinct problems: (1) delay in diagnosis in patients who have Cushing’s and, (2) patients who are convinced they have Cushing’s when the clinical features and tests do not support the diagnosis. Both of these situations are challenging for me and very distressful for patients. Basically this problem is two sides of a coin.
First of all, the diagnosis of Cushing’s is dependent on clinical findings and demonstration of consistent overproduction on cortisol. At the end of the day, despite the patient’s symptoms and clinical features, the diagnosis is “biochemical” which means laboratory tests that demonstrate that the patient is consistently making too much cortisol.
Regarding the issue of delay in diagnosis, I conducted a survey of 237 patients with proven Cushing’s – the delay in diagnosis in most patients ranged from 1 to over 10 years after onset of symptoms. Approximately 85% of patients had symptoms of Cushing’s for 3 to 5 years before the diagnosis; 51% of patients saw 2 to 4 physicians before the diagnosis was made. Some patients “surfed the Internet” to seek information about their symptoms and medical conditions and finally saw a physician who ordered appropriate tests. The delay in diagnosis is common. Why? Cushing’s is uncommon, but other conditions such as obesity, diabetes, high blood pressure, irregular menstrual cycles, depression, metabolic syndrome and polycystic ovary syndrome are very common. The symptoms and signs of Cushing’s overlap with these more common conditions and patients with Cushing’s do not always look like the “textbook” case of Cushing’s, especially early in the disease. Thus, the immediate and common problems (obesity, diabetes, high blood pressure, depression) are treated and patients are advised to go on a diet and loose weight. Only after these conditions do not improve or things become worse were they evaluated for possible Cushing’s. The problems with diagnosis are many –- lack of awareness of by physicians, absence of some of the classical features of Cushing’s (no purple stretch marks – present in only 50% of patients in a study we conducted in our Pituitary clinic), overlap with clinical features of polycystic ovary syndrome in women (much more common than Cushing’s). Unfortunately, many patients still suffer from a delay in diagnosis. I applaud those patients who have pursued an answer for their problems and finally achieved a diagnosis.
What about the second issue: patients (almost always women in my experience) who are convinced that they have Cushing’s? The usual history is one of being overweight, weight gain and/or difficulty loosing weight (with variable consistent attempts at a weight loss regimen); sometimes high blood pressure, diabetes mellitus, depression and a great deal of frustration. Invariably, these patients have “surfed the Internet” and come to the conclusion that they have Cushing’s. They have asked their physician to test for Cushing’s and because of the imprecise nature of tests and variable reliability of different laboratories, the test results usually show conflicting results. Some tests are not helpful such as a single blood cortisol or ACTH level – both of these hormones are produced in pulses or bursts and increase with stress (including just having the “needle stick” for the blood test – a single blood test is not a valid test for Cushing’s). The diagnosis of Cushing’s is dependent on demonstration of consistent overproduction of cortisol. The screening tests for Cushing’s –- 24 hour urine free cortisol, nighttime salivary cortisol levels and the 1 mg overnight dexamethasone test –- are equally reliable (depending on the accuracy of the laboratory) –- they are all approximately 92% accurate, not 100% accurate (this is common with any screening test). What does this mean? A single or two or three tests to diagnose overproduction of cortisol are not always 100% accurate and not all laboratories are equally reliable. Thus, it may take many tests to determine if there is a consistent overproduction of cortisol. The emphasis here is on the word consistent overproduction of cortisol. During times of stress and illness cortisol production is supposed to increase; this is a normal response. Tests should not be obtained in a hospitalized patient –- hospitalized patients are under the stress of illness and they should have increased urine and salivary cortisol levels and do not have a “normal” response to dexamethasone. I have seen several patients who were evaluated during a hospitalization for an illness such as pneumonia who had high test results and when they were tested later as an outpatient had no evidence of overproduction of cortisol. Again, consistency of overproduction of cortisol is the important concept and patients should not be tested when hospitalized.
Unfortunately, not all laboratory methods are equally reliable. My preference is to have the patient collect several 24 hour urine specimens, a week apart, with measurement of the urine free cortisol by a method called tandem mass spectrometry (the most precise method of measuring urine free cortisol). Some of my patients have had to collect a 24 hour urine specimen once a week for 6 weeks in order to confirm consistent overproduction of cortisol in order to make the diagnosis of Cushing’s. How reliable is the nighttime salivary cortisol test? This depends on the laboratory measuring the level. Not all laboratories are equally reliable. How reliable is the 1 mg dexamethasone test? This depends on several factors: (a) taking the pill precisely at 11 p.m. (b) having the blood test the next morning precisely at 8 a.m. [a patient may be at the lab at 8 a.m., but the blood test is not obtained until later] (c) not taking any medications which can affect the breakdown (metabolism) of dexamethasone, (d) achieving a certain blood level of dexamethasone and (e) the unmeasurable factor of how rapidly a patient’s liver breaks down (metabolizes) dexamethasone. This detailed explanation is to emphasize that the diagnosis of Cushing’s is not straightforward and it may take many tests to confirm consistent overproduction of cortisol.
Given these limitations, what is the situation for a patient who thinks she/he has Cushing’s when reliable tests do not support the diagnosis of Cushing’s? This is one of the most difficult situations for the patient and for me. I review all of the previous studies before the appointment (including review of which laboratory was used for the tests), review the history with the patient and examine the patient. If there is any question of Cushing’s, I always recommend additional testing (24 hour urine free cortisol levels over several weeks, assayed by the most precise method). If the tests show excessive cortisol production then additional tests to confirm a pituitary or another source are arranged. When these tests do not show consistent overproduction of cortisol, the patient is disappointed and angry with me that I could not confirm a diagnosis of Cushing’s. When hormone tests do not confirm the diagnosis of Cushing’s, I am happy for the patient. However, the patient is not happy with me. This is a very unsatisfactory situation for both the patient and for me. Patients often cry when I summarize the findings — that the clinical history, findings and tests do not indicate Cushing’s; patients leave my clinic both frustrated and angry. Even though I explain that not having Cushing’s is good news, this is not what the patient wants to hear. What more can a physician do? The diagnosis of all hormone disorders is always dependent on both the clinical features and demonstration of a hormone disturbance (hormone overproduction or underproduction) – this is the bottom line. I always encourage a patient to have additional testing when there is any question of possible Cushing’s. Beyond this, I can do no more.
Thus, there are patients whom I have seen who are not happy with the visit or my opinion. This has sometimes resulted in patients seeking other opinions (fine with me, I’m not perfect). However, I am aware of patients in which other evaluations –- were not accurate or appropriate which resulted in unnecessary and potentially harmful operations. About this, I can do nothing but wish the patient well. Any surgery carries a risk. So, the physician must be certain that the patient does indeed have Cushing’s prior to recommending surgery. If the patient undergoes surgery and does not have Cushing’s, the patient will not get better.
When I read the individual stories of delay in diagnosis, problems with treatments and recovery published in the CSRF newsletters, I recall the histories of many of my patients. I also recall an example of the difficulty in diagnosis. This occurred very early in my career when I evaluated a woman whom I thought had Cushing’s. At that time, insurance companies allowed inpatient studies and I had her admitted to our research center (she was not ill, just admitted for testing) for measurement of blood and urine tests, dexamethasone tests – on 3 different occasions. Being inexperienced, I presented the results of her studies to a conference of more experienced colleagues. Because the dexamethasone tests did not “fit” the guidelines, their opinion was that she didn’t have Cushing’s, even thought I thought she did. Well, fortunately, the patient became frustrated with me and went to the NIH and was admitted for evaluation. Bottom line: the physicians at NIH diagnosed Cushing’s disease and she had successful pituitary surgery. I hope this case shows the difficulty in testing for Cushing’s.
In my opinion, we as physicians continue to have problems with the delay in diagnosis of patients who have Cushing’s and there remains the problem of patients who are convinced they have Cushing’s when the clinical features and hormone tests do not support the diagnosis. I have no “smart fix” for these problems for both types of these patients. However, I continue to try to educate medical students and physicians in training to consider that patients with the common conditions of diabetes, obesity, high blood pressure, depression, osteoporosis and other problems associated with Cushing’s –- may have Cushing’s and that they require a proper evaluation.
One of the most rewarding aspects of my work is to make the correct diagnosis of Cushing’s and for the patient to have successful treatment. After treatment, sometimes I almost don’t recognize a patient after recovery from Cushing’s (may take a year or more) –- this is a wonderful day in my clinic. Regarding the patient who is convinced she has Cushing’s and the tests do not confirm the diagnosis, the patient and I remain frustrated. What to do? I do think such a patient should receive follow up evaluations, just to make sure. Again, the diagnosis of Cushing’s is not always easy and may take time to either diagnose the condition or to exclude the Cushing’s. The best outcome is an accurate diagnosis and successful treatment. As a physician who cares for patients with Cushing’s, my goal is to try to make an accurate diagnosis and arrange expert treatment.
Author: Dr. Mary Lee Vance, M.D. (Winter, 2009)
Editor’s Note: Dr. Mary Lee Vance, M.D. is Professor of Medicine and Neurosurgery at the University of Virginia, Charlottesville, VA.