Diagnostic Approach or Quest for the Holy Grail

  Internal Medicine Resident Grand Rounds

November 28th, 1995

 Tom Y. Thuren, M.D., Ph.D.

Case Report

Chief Complaint

C.G. is a 44 year old white female who presented to her new primary care physician in 7/95 complaining fatigue and 28 lbs weight gain since 12/93.

Her past medical history is significant for breast cancer treated with modified radical mastectomy, radiation and chemotherapy, hypertension for 15 years, type II diabetes mellitus since 5/93 and hypercholesterolemia since 1990. Her hypertension had more difficult to control since 1993.

Her medications included: Furosemide, Pravachol, Zaroxylyn, Glucotrol, Potassium, Insulin and Tamoxifen.

Her complaints centered around severe fatigue, proximal muscle weakness and alterations in her physical habitus. She was found to have round face and facial flushing, very evident central obesity, buffalo hump, wasted extremities, purple striae on abdomen and hyperpigmentation. Her thoracic spine was tender to palpation and radiologic studies later revealed T12 compression fracture and osteoporosis. Her primary care physician noted these complaints and changes to be consistent with Cushing's syndrome.

Laboratory results on this visit included sodium 135, potassium 2.8, chloride 91, bicarbonate 29, BUN 29, creatinine 1.0 and glucose 295. WBC 10.7, differential 76% segs, 7% bands, 12% lymphocytes and 4% monocytes, hemoglobin 15.1, and platelets 204K. TSH was 1.09.

Am serum cortisol was 33.6 ug/dl ( 5.5-20) and urinary free cortisol was 753 ug/24 hrs (46-131) or 823 ug / g CRT. Low dose dexamethasone suppression test was performed and revealed serum cortisol of 30.9 which was not suppressed. Also 17-hydroxycortisol in urine was not suppressed during this test. 8/232/95 ACTH level was drawn and was elevated at 93 pg/ml (<70).

8/95 patient had CT of head and brain which showed that the sellar region was normal.

8/95 patient had CT of chest and abdomen which showed right middle lobe atelectasis in lungs and bulging of right adrenal gland approximately 1 cm. Initially this was concluded to be an adenoma but on further review radiologist noted that adenoma typically is at least 2 cm in size and that the other adrenal gland was not atrophic/ suppressed and therefore, the finding was not fully consistent with an adrenal adenoma.

She was referred to endocrinologists who agreed that she had Cushing's syndrome and elevated ACTH but that the source of ACTH was unknown. Next patient underwent high dose dexamethasone suppression test which revealed 65% suppression in urinary free cortisol secretion (50-90% gray zone, diagnostic >90% suppression). Since Cushing's disease was the most likely diagnosis for this patient she was referred to a neuroradiologist who performed inferior petrosal sinus sampling procedure on 9/25/95. This showed that the likely origin of ACTH secretion was right lobe of pituitary gland (see results in Table I). However, due to an anatomical variation (Cavernous sinuses draining to clival plexus prior to draining into petrosal sinuses) admixing of left and right blood flow was possible.

Table I:  ACTH LEVELS DURING INFERIOR PETROSAL SINUS SAMPLING

______________________________________________________________________________

Left Right Peripheral L:P ratio R:P ratio

pg/ml pg/ml pg/ml

Cavernous sinus #1 16 108 21 0.76 5

Cavernous sinus #2 36 216 22 1.6 9.8

Inferior Petrosal sinus #1 82 749 20 4.1 37.5

Inferior Petrosal sinus #2 45 913 13 3.4 70.2

______________________________________________________________________________

Pituitary MRI did not reveal a pituitary adenoma.

Patient underwent a transsphenoidal surgery to resect right lobe of pituitary gland on 10/16/95. Post-operatively patient did well except transient diabetes insipidus resolving in 3-4 days. Pathological diagnosis of resected right lobe of pituitary was Crooke's hyaline change and the pathologist concluded that no pituitary adenoma could be identified. Three days after operation while patient was on 4 mg dexamethasone every 6 hours an am cortisol level was 1.0 ug/dl which was a suppressed value.

10/30/95 patient was complaining being still tired and was depressed. She was doing well except sinus congestion and antibiotic treatment. Her facial flushing was improved abdominal obesity purple striae and pigmentation was unchanged. Serum osmolality was 306 urine osmolality was 485. Dexamethasone which she was taking 0.5 mg once daily was discontinued. Three days later serum cortisol at 13:40 pm was < 0.2 ug/dl suppressed. Therefore, she was placed on Prednisone replacement therapy with 5 mg every morning and 1.25 mg every afternoon. Her diabetes is still poorly controlled with glucose ranging 200-250. Her insulin dosing was increased and she was to undergo diabetic teaching.

The earliest published case of Cushing's syndrome was 1928 by Brown (1). This paper described a bearded woman with diabetes who had an oat-cell carcinoma of the lung. This case represented a case of ectopic corticotropin syndrome. In 1932 Dr. Harvey Cushing published a series of seven patients with basophilic adenomas of the pituitary and a disease named after him (2). However, he additionally described two more cases of the disease who did not have pituitary adenomas and , therefore, likely represent ectopic or nonpituitary tumors as a cause for Cushing's syndrome. All these patients had a syndrome that resulted from a long term exposure to glucocorticoids.

Cushing's syndrome is defined as a disease state which results from hypercortisolism and clinically manifests itself as sudden onset of weight gain, dorsocervical fat pad ("buffalo hump"), hypertension, glucose intolerance, oligomenorrhea or amenorrhea in premenopausal women or decreased libido in men, muscle wasting and weakness and mild hirsutism when the most common findings are listed (3,4).

Cushing's syndrome can be separated into corticotropin-dependent Cushing's syndrome and corticotropin-independent Cushing's syndrome. Cushing's disease is reserved for Cushing's syndrome caused by excessive secretion of corticotropin by pituitary corticotroph tumors and is , therefore, corticotropin dependent (2). Normal function of hypothalamus-pituitary-adrenal axis is reviewed next in order to better understand this classification.

Hypothalamus controls to function of pituitary gland by secreting factors that either stimulate or inhibit hormone secretion from pituitary gland. These factors are fairly specific for any given hormone produced by the pituitary. Corticotropin-releasing hormone (CRH) is synthesized in the hypothalamus and carried to the anterior pituitary in portal blood (5). CRH secretion is regulated by a variety of neurotransmitters and inhibited by cortisol, Figure 1 (4,6). Stimulation of central nervous system centers such as locus caeruleus also results in stimulation of CRH secretion. CRH is the most potent regulator of corticotropin secretion in the pituitary. Arginine vasopressin and other hypothalamic agents also stimulate corticotropin secretion but in a lesser extent (7,8).

Corticotropin is synthesized in the anterior pituitary by corticotroph cells. Corticotropin is a product of pro-opiomelanocortin (POMC) gene of which there is a single copy in humans (9). It is located on the short arm of chromosome 2 and consists of 3 exons, 2 introns and regulatory elements (10). Corticotropin is synthesized as a precursor, POMC, which is processed post- translationally into corticotrophin, beta-lipotropin, beta-endorphin and into several other peptides (9). Corticotropin secretion is stimulated by CRH and inhibited by cortisol (8, 11,12). Corticotropin stimulates the adrenal cortex to produce cortisol (6).

Cortisol is the product of the adrenal gland and it is called glucocorticoid because it stimulates the catabolism of peripheral fat and protein. This provides substrates for hepatic gluconeogenesis to produce glucose. Cortisol is a stress hormone modulating the response to stress (13). It also has anti-inflammatory effects which are commonly used as pharmacologic means to alleviate inflammatory responses in medical practice (13). The above mentioned inhibition of CRH and corticotropin secretion by cortisol is a classical example of negative feedback regulation.

Corticotropin-Dependent Cushing's Syndrome

Cushing's disease is the most common form of the syndrome (4). Up to 90% of patients having corticotrophin-dependent Cushing's syndrome have pituitary corticotroph tumor, Table II (4). The tumors arise from a single progenitor cell and are usually microadenomas less than 1 cm in diameter (14). Macroadenomas are rare, hyperplasia and carcinomas are extremely rare (4, 15). Chronic CRH hypersecretion does not result in development of adenoma but causes corticotroph hyperplasia (16). Corticotrophin hypersecretion causes bilateral adrenocortical hyperplasia. Patients having chronic Cushing's disease may develop macronodular adrenal hyperplasia which is sometimes confused to adrenal adenomas especially when imagining techniques are used for diagnostic purposes (17).

TABLE II

Relative Prevalences of Various Types of Cushing's Syndrome among 630 Patients Studied at Different Times

______________________________________________________________________________

Corticotropin-dependent Cushing's syndrome

Cushing's disease 68

Ectopic corticotropin syndrome 12

Ectopic CRH syndrome <1

Corticotropin-independent Cushing's syndrome

Adrenal adenoma 10

Adrenal carcinoma 8

Micronodular hyperplasia 1

Macronodular hyperplasia <1

Pseudo-Cushing's syndrome

Major depressive disorder 1

Alcoholism <1

______________________________________________________________________________

Modified from (4).

Pituitary adenoma cells are relatively resistant to glucocorticoid inhibition but are sensitive to CRH stimulation (16). Therefore, these patients have increased corticotrophin response to CRH stimulation and glucocorticoids incompletely suppress secretion of corticotrophin. These two phenomenon can be used when diagnosis of Cushing disease is entertained. In 10% of patients with pituitary microadenoma the clonal cells are lacking CRH receptors and are insensitive to CRH stimulation (16).

This syndrome is caused by excess corticotropin secretion outside of the pituitary. It's acute form is characterized by hypertension, edema, hypokalemia and glucose intolerance (18). It is most often associated with small-cell lung carcinoma (19). It is estimated that up to 75% of acute ectopic corticotropin syndrome is the result of oat cell carcinoma of lung. The chronic ectopic corticotropin syndrome is clinically indistinguishable from Cushing's disease and is more commonly associated with indolent tumors, Table III (19). These tumors include carcinoids, pheochromocytomas and other neuroendocrine tumors. Because POMC gene is expressed in a wide variety of tissues the source of corticotropin secretion varies greatly. These patient have bilateral adrenal hyperplasia and hyperfunction. Their CRH and pituitary secretion of corticotropin are suppressed. Corticotropin secreted by ectopic tumor is not usually suppressed by glucocorticoids (20). This can be used to distinguish them from Cushing's disease. High dose glucocorticoids suppress corticotropin secretion in up to 50% of cases with carcinoid tumors (21,22).

This syndrome is due to CRH hypersecretion by a variety of tumors. Most often bronchial carcinoid tumor is diagnosed (23). Clinically this syndrome is indistinguishable from ectopic corticotropin syndrome. In ectopic CRH syndrome CRH concentration is elevated and CRH-stimulated corticotropin secretion is suppressible with high dose glucocorticoids. In some reported cases tumors secreting CRH also secrete corticotropin resulting in features that are biochemically indistinguishable from ectopic corticotropin syndrome (24).

Table III

Causes of Ectopic Corticotropin Syndrome

______________________________________________________________________________

Jex Howlett Imura Findling

n 25 16 30 49

Oat cell bronchus 20 19 33 10

Adenocarcinoma of the bronchus - - 10 -

Squamous carcinoma of the bronchus - - 3 -

Undifferentiated carcinoma of the bronchus - - 10 -

Bronchial carcinoid 28 37 6 49

Thymic carcinoid 8 12 - 16

Malignant thymoma - - 13 -

Pheocromocytoma 12 6 - -

Pancreatic islet cell carcinoma 16 12 6 4

Pancreatic adenocarcinoma 4 - - 2

Oat cell carcinoma of the esophagus - - 6 -

Ovarian carcinoma 4 - - -

Miscellaneous - - 10 -

Unknown - - - 12

______________________________________________________________________________

Modified from (19).

Nonendocrine disorders may result abnormally regulated cortisol secretion and a fraction of these patients also have clinical features of Cushing's syndrome. Up to 80% of patients with major depression have elevated cortisol secretion (25,26). However, usually this elevation is minimal. Chronic alcoholism is an uncommon cause of pseudo-Cushing's syndrome (27). These patients have mildly elevated cortisol secretion which disappears in weeks to months during abstinence (28). The mechanism for this phenomenon is unclear but it may be due to increased CRH secretion or direct effect of alcohol on pituitary resulting increased corticotropin secretion or on adrenal glands resulting in mild hypercortisolism (28). Some patients with obesity, hypertension, diabetes and depression who usually are middle-aged women may have clinical and biochemical features indistinguishable from Cushing's syndrome. These features disappear with the remission of depression. These hormonal abnormalities are possibly due to hyperactivity of the hypothalamus- pituitary-adrenal axis (29).   

Corticotropin-Independent Cushing's Syndrome

Adrenocortical tumors

Adrenocortical tumors can be benign or malignant and they are the most common cause for corticotropin-independent Cushing's syndrome, Table I (4). Adrenal adenomas are relatively efficient in producing cortisol so production of precursors is low and, therefore, clinically manifestations resulting from cortisol precursors e.g androgens are less common (30). Adrenal carcinomas are relatively inefficient in producing cortisol resulting in excess androgen states featured by virilization (30).

This is a very rare cause of Cushing's syndrome. Most patients are children and adults younger than 30 years of age. It occurs sporadically or as an autosomal dominant disorder with blue nevi, pigmented lentiges, cutaneous, mammary and atrial myxomas, pituitary somatotroph adenomas and testicular tumors (31). This constellation is called Carney's complex. Adrenals from these patients have nodules that range from 1-2 mm to 3 cm in size (32). Nodules are deeply pigmented and appear black or brown in cut sections. Microscopically the nodules contain cells with large clear lipofuscin-laden cytoplasm. This syndrome is also called primary pigmented nodular adrenal hyperplasia. Autoantibodies that stimulate adrenocortical growth and steroidogenesis are the likely cause for this disease ( all 9 patients tested had adrenal stimulating autoantibodies) (33).

Bilateral Macronodular Hyperplasia 

A few patients have been described to have bilateral adrenal nodules ranging in size from 0.5 to 7 cm and overall enlarged adrenal glands (34). A characteristic pathologic feature is internodular hyperplasia. Up to 10% of patients with Cushing's syndrome has bilateral macrodular hyperplasia of adrenal glands (35). Patients with this disorder tend to be older and to have symptoms longer than patients with pituitary or adrenal adenomas. It has been postulated that long standing corticotropin stimulation leads to adrenal nodule formation and that some of these nodules may later become autonomous (36).

Since the prevalence of Cushing's syndrome is less than 1% there is a need for a screening test with high negative predictive value. On the other hand a test for diagnosing Cushing's syndrome that has a good positive predictive value is needed. Next one has to determine if Cushing's syndrome is corticotropin-dependent or -independent. When corticotropin-dependent Cushing's syndrome has been diagnosed 90% of these patient will have Cushing's disease (4). For patients who have corticotropin- dependent Cushing's syndrome a test with higher than 90% specificity and sensitivity is needed to find the source of corticotropin secretion in order to select a rational treatment option. Since interpretation of the tests directed to accomplish these tasks is dependent on performance characteristics of each test, some basic concepts of clinical epidemiology are briefly reviewed below.

Sensitivity refers to the ability to correctly identify those with a disease (37). Sensitivity is the proportion of patients with disease who have a positive sign, symptom or test. Formula for calculating sensitivity is in Table IV. Specificity refers to the ability to identify correctly those without disease. The specificity of a test is the proportion of patients without disease who have a negative test result or lack of sign or symptom, Table IV. Sensitivity and specificity have been considered to be characteristics of the test only. However, disease severity and stage may effect sensitivity (38). The specificity may depend on characteristics of the reference population.

Positive predictive value is a measure of how good a test, sign or symptom is to identify patients with a disease and negative predictive value measures how reliably negative test identifies patients without a disease, Table IV (37). These predictive values change as a function of prevalence of the disease, Table V (39). As prevalence increases positive predictive value increases and negative predictive value decreases assuming that test sensitivity and specificity remain constant. This has implications for screening rare disease as will be discussed below.

Likelihood ratios ( positive and negative) are derived from sensitivity and specificity, Table IV. Likelihood ratio is an expression of the odds that a sign, symptom or test result would be expected in a patient with a disease as opposed to one without (37). A positive likelihood ratio X implies that a positive sign, symptom or test is X times more likely in a patient with the disease than in a patient without a disease. Negative likelihood ratio implies to the absence of sign, symptom or test result in a patient with the disease. Diagnostic accuracy is defined as the ability of a test to identify correctly those with the disease and those without the disease, Table IV (37).

TABLE IV

Sensitivity, Specificity and Their Derivatives

______________________________________________________________________________

Disease

+ -

Test Result + True positive (TP) False positive (FP)

- False negative (FN) True negative (TN)

Sensitivity = TP / TP +FN Specificity = TN / TN + FP

______________________________________________________________________________

Positive Predictive Value = TP / TP + FP Negative Predictive Value = TN / TN + FN

Diagnostic Accuracy = TP + TN / TP + FP + TN + FN

______________________________________________________________________________

Likelihood ratio positive = sensitivity / (1- specificity) = (TP rate) / (FP rate)

Likelihood ratio negative = (1-sensitivity) / specificity = (FN rate) / (TN rate)

______________________________________________________________________________

Post-test odds of disease = (prevalence)(sensitivity) / (1-prevalence)(1-specificity)

Post-test probability of disease

= (prevalence)(sensitivity) / [ (prevalence)(sensitivity) + (1-prevalence)(specificity)]

______________________________________________________________________________

TABLE V

Positive and Negative Predictive Values as a Function of Disease Prevalence Assuming Test Sensitivity and Specificity of 90% Each

______________________________________________________________________________

Disease prevalence or pre-test 0.1% 1% 10% 50% 90%

probability

Positive predictive value 0.89% 8.33% 50.0% 90.0% 98.78%

Negative predictive value 99.99% 99.89% 98.78% 90.0% 50.0%

______________________________________________________________________________

Pre- and Post-Test Probability of Disease

Likelihood ratios can be used to estimate the post-test odds of a disease from pre-test odds of disease, Table IV. To convert odds to probabilities an equation of odds/(1+odds) is used Table IV (39). These parameters have important implication for which patients to test and what kind of test characteristics a planned test needs to have in order for the post-test probability being higher than the pre-test probability (prevalence) of a disease. In other words these parameters describe if a test is of useful for diagnosing that particular disease in an individual.

Diagnosis of Cushing's Syndrome

Characteristic clinical findings of Cushing's syndrome (hypercortisolism) and their operating characteristics are shown in Table VI (3). There are no pathognomonic symptoms or sign and clinical diagnosis is based on development of several new symptoms and signs simultaneously. Symptoms can be categorized into cortisol effects (obesity), androgen effects (virilization) and mineralocorticoid effects (hypokalemia). Weight gain (obesity) is the most common complaint and the well known triad of moon face, buffalo hump and abdominal striae and hyperpigmentation is well described in the literature. Central obesity is quite sensitive for Cushing's syndrome while muscle weakness and ecchymoses are more specific signs and symptoms. Proximal muscle weakness is 9.3 times more likely in a patient with Cushing's syndrome than without Cushing's syndrome. Absence of central obesity is 0.14 times as likely in a patient with Cushing's syndrome than in a patient without Cushing's syndrome.

TABLE VI

Operating Characteristics of Clinical Signs and Symptoms in the Diagnosis of Hypercortisolism

______________________________________________________________________________

Diagnostic finding Sensitivity Specificity Likelihood ratio

Present Absent

% %

Central obesity 90 71 3.1 0.14

Weakness 65 93 9.3 0.38

Plethora 82 69 2.6 0.26

Leucocyte count > 11000 58 70 1.9 0.60

Acne 52 76 2.1 0.63

Striae 46 78 2.1 0.69

Diastolic BP > 105 mmHg 39 83 2.3 0.73

Pitting edema 38 83 2.2 0.75

Hirsutism 50 71 1.7 0.70

Ecchymoses 53 94 8.8 0.50

Serum K < 3.6 mg/L 25 96 6.2 0.78

Oligomenorrhea 72 49 1.4 0.57

Abnormal glucose tolerance test 88 23 1.1 0.52

Generalized obesity 3 38 0.05 2.5

Osteoporosis (early diagnosis) 64 97 21 0.37

Osteoporosis (late diagnosis) 26 94 4.8 0.79

______________________________________________________________________________

Modified from (3).

Laboratory (Biochemical) Tests

Laboratory test are divided into screening tests and test confirming diagnosis of Cushing's syndrome. Test characteristics for these tests are quite different. Screening tests are used for patients with clinically suspected or not obvious Cushing's syndrome whereas diagnosis confirming tests are used for patients with clinically obvious or strongly suspected Cushing's syndrome. Test confirming Cushing's disease are discussed first.

The determination of 24 hour excretion of cortisol in urine is now the most direct and reliable practical test for diagnosing Cushing' syndrome (4). Plasma cortisol level is unreliable because it rises and falls episodically even in patients with Cushing's syndrome. 24 hour urine free cortisol integrates the plasma free cortisol concentrations during the entire day. Most laboratories use a solvent extraction method and assay cortisol by a competing protein binding assay, radioimmunoassay or chemiluminescence (40). The upper limit for normal subjects is between 80-120 ug/24hrs. A newer method uses high performance liquid chromatography to determine urine free cortisol. This assay is more specific (less trash is measured as cortisol) and, therefore, the upper limit for normals is 50 ug/24hrs. The sensitivity and specificity for the last assay were 100% and 98%,respectively, in a recent study (41). It was compared to an older test measuring urinary 17-hydroxycorticosteroids and it was found that sensitivity and specificity for the older test were only 73% and 94%, respectively. The problems associated with 24 hour urine collection in outpatient settings resulted in the recommendation that 24 hour urinary free cortisol measurement should be repeated 2-3 times in patients thought likely to have Cushing's syndrome (4).

Low dose dexamethasone suppression test is the main screening test for Cushing's syndrome (42,43). The standard suppression tests are the two-day 0.5 mg dexamethasone every 6 hours test and the 1mg dexamethasone overnight (11 pm) test. The overnight test measures plasma cortisol the next morning and normal subjects should suppress plasma cortisol to less than 3 ug/dl. Earlier 5 ug/dl was considered to be normal but several false negative studies were observed using this criteria. The sensitivity and specificity were 98.1% and 98.9%, respectively when normal controls were used (39). The specificity dropped to 89.5% when all controls were included (39). If only data from obese and other control were used ( normals were excluded) specificity dropped to 80.5%, Table VII (39). This means that when obese patients, patients on medications and patients with various medical, surgical and psychiatric conditions are screened the number of false positive test results increases. This is still a very useful and reliable screening test. However, when patients with above mentioned conditions lacking convincing signs and symptoms of Cushing's syndrome are screened caution must be used when interpreting positive test results. The overall diagnostic accuracy for 1mg dexamethasone suppression test is 98.7% (44). After a positive test result a diagnosis of Cushing's syndrome must be confirmed with another test like outlined above. This is evidenced by the high negative predictive value of overnight dexamethasone suppression test, 99.9% (you can be very comfortable that a patient does not have disease) and the very low positive predictive value of 0.05% (you do not know if patient has a disease or not, further testing is indicated) ( 39).

TABLE VII

Test Characteristics of the 1 mg Overnight Dexamethasone Suppression Test

______________________________________________________________________________

Sensitivity Specificity

% %

Normal controls 98.1 98.9

All controls 98.1 89.5

Obese and other controls 98.1 80.5

_____________________________________________________________________________

Modified from (39).

The 2-day low dose dexamethasone suppression test measures urinary 24 hr cortisol or 17-hydroxycorticosteroid (45). The normal controls should have less than 10 ug cortisol/24hrs in urine or less than 2.5 ug 17-hydroxycorticosteroid/24hrs in urine. Using the criteria for 17-hydroxycorticosteroid the specificity was 74% and sensitivity was 69% (46). The goal would be a good specificity( close to 100%) and this can be achieved by using 5 ug/24hrs 17-hydroxycorticosteroid as the upper limit of normal (46). However, then sensitivity drops to 54% (46). The test protocol using 24 hr urinary cortisol measurement has test characteristics equal to that of overnight suppression test. Therefore, one of the two latter mentioned test are recommended to be used for screening for Cushing's syndrome.

Mild hypercortisolism due to Cushing's disease is very difficult to distinguish from patients with pseudo-Cushing's syndrome. This is due to the mildly elevated cortisol levels leading to normal or only mildly abnormal suppression test result. Additional tests which will help in finding pseudo-Cushing's syndrome patients include pm nadir plasma cortisol which is preserved in depression but not in Cushing's syndrome, CRH and dexamethasone test, naloxone test and insulin-induced hypoglycemia test (4). CRH and dexamethasone test is performed by giving patient eight doses of 0.5 mg dexamethasone in 6 hour intervals (low dose dexamethasone suppression test) and a 100 ug CRH intravenously two hours after the completion of the dexamethasone suppression (46). Plasma cortisol is measured at 15 and 30 minutes after administration of CRH. The upper limit of normal is 1.4 ug/dl cortisol. Using this cut this test identifies 100% of patients with Cushing's syndrome, therefore, the sensitivity and specificity as well as both predictive values are 100 % (40). However, CRH is not currently available except on special permission from NIH due to production problems the manufacturer is experiencing. A comparison of test characteristics of an older version of low dose dexamethasone suppression test, CRH test with and without dexamethasone is presented in Table VIII (40). Naloxone test is based on the fact that this drug stimulates CRH secretion more in depressed patients than in patients with Cushing's disease who have CRH suppressed by cortisol (47). This test is not in general use at the moment. Insulin-induced hypoglycemia overcomes hypothalamus-pituitary axis suppression and in patients with pseudo-Cushing's syndrome CRH- corticotropin and cortisol levels increase (48). This is not the case in patients with Cushing's syndrome. However, serious side effects from insulin-induced hypoglycemia are possible and, therefore, caution should be exercised if this test is to be used.

TABLE VIII

Comparison of Criteria Chosen to Yield 100% Specificity for Diagnosis of Cushing's Syndrome in 58 Patients with Mild Hypercortisolism

______________________________________________________________________________

Test variable Criterion Specificity Sensitivity Positive Negative Diagnostic

Predictive Predictive Accuracy

Value Value

% % % % %

Low dose dexamethasone suppression test

17-OCHS day 4 >11 umol/d 74 69 84 54 71

17-OCHS day 4 >14.6 umol/d 100 54 100 51 69

Urine free cortisol >100 nmol/d 100 56 100 53 71

day 4

CRH test without dexamethasone pretreatment

Cortisol sum of >3450 nmol/L 100 64 100 58 76

post-CRH levels

CRH test peak >35 pmol/L 100 13 100 36 41

Corticotropin

Dexamethasone-CRH test

Basal cortisol >38.0 nmol/L 100 90 100 84 91

(before CRH)

Cortisol 15 min >38.0 nmol/L 100 100 100 100 100

after CRH

Corticotropin 30 >3.5 pmol/L 100 74 100 66 83

min after CRH

______________________________________________________________________________

Modified from (40). 17-OCHS, 17-hydroxycorticosteroid.

Distinguishing between Corticotropin-Dependent and Corticotropin-Independent Cushing's Syndrome

The development a the radioimmunometric assay for plasma corticotropin has made it possible to reliably distinguish the two types of Cushing's syndrome (49). If plasma cortisol level is above 15 ug/dl and plasma corticotropin level is less than 5 pg/ml cortisol secretion is corticotropin independent. If plasma corticotropin level is greater than 15 pg/ml cortisol secretion is corticotropin dependent i.e. patient has either Cushing's disease or ectopic corticotropin or CRH syndrome. Intermediate plasma corticotropin values are less definitive but usually indicative of corticotropin-dependent secretion of cortisol (4). In cortisol producing adrenal neoplasms, primary pigmented nodular adrenal hyperplasia and in factitious Cushing's syndrome corticotropin values are less than 5 pg/ml and they exhibit a blunted response to CRH (36).

Diagnostic Test for Determining the Source of Excess Corticotropin Secretion

This is the standard suppression test for distinguishing patients with pituitary vs non-pituitary Cushing's syndrome. Dexamethasone is given 2 mg every six hours for eight doses and cortisol or cortisol metabolites in 24 hour urine are measured (45). Initially a suppression of more than 50% of 17-hydroxycorticosteroid excretion was used as a criteria to diagnose Cushing's disease. The test characteristics are shown in Table IX (40). As is evident both sensitivity and specificity are close to 80%. Recently a new criterion of suppression of more than 64% of excretion of 17-hydroxycorticosteroid or more than 90% suppression of free cortisol excretion was introduced (50). When using these criteria the specificity increased to 100 % and overall accuracy was 86 %. Approximately 90% of patients with corticotropin-dependent Cushing's syndrome has Cushing's disease (pituitary) and, therefore, the pre-test probability is 90%. If the test characteristics of the high dose test are approximately at the same level it means that the post-test probability is still 90%. If the test sensitivity is less than 90% then this test does not help in separating Cushing's disease from ectopic corticotropin or ectopic CRH syndrome. Therefore, the new criteria needs to be used for the test to have any diagnostic value.

Metyrapone has been show to block the conversion of 11-deoxycortisol to cortisol. Patients with Cushing's disease have normal or supranormal increase in the plasma 11-deoxycortisol or in urinary secretion of 17-hydroxycorticosteroids in response to six doses of 750mg of metyrapone every four hours (51). When cortisol level falls more corticotropin is secreted and plasma 11-deoxycortisol increases in Cushing's disease. Most patient with ectopic corticotropin secreting tumor have little or no increase in either value because pituitary secretion of corticotropin is suppressed. When criterion of increase of greater than 400 times in plasma deoxycortisol and an increase more than 70% over the basal 24 hour urine excretion of 17-hydroxycorticosteroid was used the sensitivity of this test was 71% (52) . When this test was combined with high dose dexamethasone suppression test specificity was 100 % and sensitivity increased to 88% (52).

TABLE IX

High-Dose Dexamethasone Suppression Testing in the Differential Diagnosis of Corticotropin-Dependent Cushing's Syndrome

______________________________________________________________________________

Clinical series Sensitivity Specificity Accuracy

% % %

Oldfield et al. 81 79 81

n=183

Tabarin et al. 85 80 81

n= 27

Findling et al. 80 67 76

n= 29

______________________________________________________________________________

Pre-test probability of pituitary Cushing's disease = 90%.

Modified from (40).

CRH Stimulation Test

This test is essentially described above (see section Test to Identify Patients with Pseudo-Cushing's Syndrome). This test uses plasma corticotropin increase of greater than 35 % after CRH administration as a criteria. The sensitivity of this test is 93% and specificity 100 % calculated from values measured for 100 patients with Cushing's disease and 16 patients with ectopic corticotropin secretion syndrome (53). Therefore, this test may not alone enhance post-test probability from the pre-test level for Cushing's disease.

The most direct way of demonstrating pituitary hypersecretion of corticotropin is to measure central (blood draining from the tumor) to peripheral vein corticotropin gradient (54). Venous blood drains from pituitary to right and left cavernous sinuses and then into inferior petrosal sinuses. By cannulating right and left inferior petrosal sinus and obtaining blood samples for measurement of corticotropin levels and comparing those to peripheral corticotropin level is the basis of this test. When the ratio of IPSS corticotropin concentration to peripheral concentration is greater than 2-3 this will indicate the presence of Cushing's disease (54). Usually several samples are obtained during this procedure. The test characteristics are at approximately 90% level for this test. The test characteristics will improve if IPSS is done with CRH stimulation. However, the problems with the availability of CRH makes this approach less attractive. It is an invasive procedure having complications ranging from inguinal hematoma to permanent brain damage due to stroke anywhere from 0.2 to 23% of patients (4). It is a very expensive procedure (2500 to 5000 $). All these properties make this test unsuitable for routine use. The development of transsphenoidal pituitary surgery has increased interest in procedure because in 70-80% of patients with Cushing's disease the test is correct in predicting the location (right or left) of the pituitary adenoma (55). Thereafter this can be removed by hemihypophysectomy. This is especially useful if pituitary MRI studies are inconclusive.

The fact that up to 10% of population in the 20 to 40 year age group will have incidental tumors of the pituitary gland by MRI (non-functioning) makes MRI unuseful in first steps of Cushing's disease diagnosis (56). MRI is valuable in making the decision on what type of a surgical approach is used to treat Cushing's disease by identifying pituitary pathology after the diagnosis of Cushing's disease is made. The fact that incidental adrenal adenomas (non-functioning) are present in 2-3% of general population who underwent abdominal CT makes adrenal CT unsuitable for separating corticotropin-dependent Cushing's syndrome from corticotropin-independent Cushing's syndrome (57). However, if the diagnosis of corticotropin-independent Cushing's syndrome is established then adrenal CT or MRI is appropriate . If corticotropin-dependent Cushing's syndrome has corticotropin secretion that cannot be suppressed with dexamethasone chest and/or abdominal CT or MRI should be performed looking for source of ectopic corticotropin secretion prior to IPSS (4). Since most tumors secreting corticotropin are carcinoid tumors in lungs and mediastinum radionuclide imaging using an analoque of octreotide labeled with indium-111 can be used to detect these tumors (58). This procedure detects up to 86% of carcinoid tumors.

The treatment of choice is transsphenoidal microadenomatomy (4). This is possible in cases where clearly circumscribed microadenoma can be identified. If this is not possible the patient should have 85-90% resection of anterior pituitary. Experienced neurosurgeons have a cure rate of 80% after initial surgery (4). If a patient wants to have children the treatment options are pituitary irradiation and/or bilateral total adrenalectomy. The criteria for cure is undetectable plasma cortisol concentration in the morning (< 1 ug/dl) and corticotropin level of less than 5 pg/ml after the patient is off post-surgical steroid treatment 4 to 7 days after surgery (4). Patients require daily glucocorticoid replacement until their hypothalamus- pituitary- adrenal axis recovers in 6-12 months after surgery. If surgery is not curative the next choice is pituitary irradiation which will cure 45-80% of adults who failed pituitary surgery (59). In children pituitary irradiation has a cure rate of 85% and it is the treatment of choice (4). Usually it takes 3 to 12 months to receive the maximum benefit from pituitary irradiation. During this time hypercortisolism can be controlled by using adrenal enzyme inhibitors e.g. ketoconazole, metyrapone or aminoglutethimide (59). These drugs can be used only as adjunctive therapy. Bilateral total adrenalectomy provides the definitive cure in cases where the above discussed treatment options have failed (4). If Cushing's disease is treated with total bilateral adrenalectomy a risk of Nelson's syndrome is increased (59). Patients with Nelson's syndrome have enlarged locally invasive corticotropin-secreting pituitary tumors which may cause serious local adverse effects by compressing optic chiasm and causing nerve damage.

Cushing's disease is caused by corticotropin secreting pituitary adenoma. Diagnosis of Cushing's syndrome is based on a low dose dexamethasone suppression test with high negative predictive value but a low positive predictive value and on 24 hour urinary free cortisol secretion test confirming the diagnosis. Plasma corticotropin assay distinguishes corticotropin dependent Cushing's syndrome from corticotropin-independent Cushing's syndrome. High dose dexamethasone suppression test and inferior petrosal sinus sampling for corticotropin will establish diagnosis of Cushing's syndrome in most cases. These steps are illustrated in Figure 2 showing one suggested algorithm for the diagnosis of Cushing's syndrome and Cushing's disease (4, 40). Treatment of choice for Cushing's disease is transsphenoidal microadenomatomy. The diagnostic tests for Cushing's syndrome and disease are used to illustrate the importance of concepts of clinical epidemiology in interpreting results from these tests.

Acknowledgements

I gratefully thank Drs. Faulk and Cefalu, Dept. of Endocrinology and Metabolism, BGSM for their help finding my case report subject. Dr. Ober, Dept. of Endocrinology and Metabolism, BGSM is thanked for reviewing this manuscript.

1. Brown, W.H.: A case of pluriglandular syndrome, "Diabetes of bearded women". Lancet 1928, 2: 1022-1023.

2. Cushing, H: The basophil adenomas of the pituitary body and their clinical manifestations. Bull. Johns Hopkins Hosp. 1932, 50: 137-145.

3. Nugent, C.A., Warner, H.R., Dunn, J.T. and Tyler, F.H.: Probability theory in the diagnosis of Cushing's syndrome. J. Clin. Endocrin. 1964, 24: 621-627.

4. Orth, D.N.: Cushing's syndrome. NEJM 1995, 332: 791-803.

5. Vale, W., Rivier, C., Brown M.R. et al.: Chemical and biological characterization of corticotropin releasing factor. Recent Prog. Horm. Res. 1983, 39: 245-270.

6. Owens, M.J. and Nemeroff, C.B.: Physiology and pharmacology of corticotropin releasing factor. Pharmacol. Rev. 1991, 43: 425-473.

7. Debold, C.R., Sheldon, W.R., DeCherney, G.S. et al.: Arginine vasopressin potentiates adrenocorticotropin release induced by ovine corticotropin releasing factor. J. Clin. Invest. 1984, 73: 533-558.

8. Vale, W., Vaughan, J., Smith, M., Yamamoto,G., Rivier, J. and Rivier, C.: Effects of synthetic ovine corticotropin releasing factor, glucocorticoids, catecholamines, neurohypophyseal peptides and other substances on cultured corticotropic cells. Endocrinology 1983, 113: 1121-1131.

9. Jackson, R.V., DeCherney, G.S., DeBold, C.R. et al.: Synthetic ovine corticotropin releasing hormone: simultaneous release of propiolipomelanocortin peptides in man. J. Clin. Endocrinol. Metab. 1984, 58: 740-743.

10. Takashi, H., Teranish, Y., Nakanish, S. et al.: Isolation and structural organization of the human corticotropin beta-Lipotropin precursor gene. FEBS Letters 1981, 135: 97-102.

11. Eberwine, J.H., Jonassen, J.A., Evinger, M.J. and Roberts, J.L.: Complex transcriptional regulation by glucocorticoids and corticotropin releasing hormone of propiomelanocortin gene expression in rat pituitary cultures. DNA 1987, 6: 483-492.

12. Calogero, A.E., Gallucci, W.T., Gold, P.W. and Chrousos, G.P.: Multiple feedback regulatory loops upon rat hypothalamic corticotropin releasing hormone secretion: potential clinical implications. J. Clin. Invest. 1991, 325: 897-905.

13. Munck, A., Guyre, P.M. and Holbrook, N.J.: Physiological functions of glucocorticoids in stress and their relation to pharmacological action. Endocrin. Rev. 1984, 5: 25-44.

14. Mampalam, T.J., Tyrrell, J.B. and Wilson, C.B.: Transphenoidal microsurgery for Cushing's disease: a report of 56 cases. Ann. Intern. Med. 1988, 109: 487-493.

15. Gabrilov, J.L., Anderson, P.L. and Hammi, N.S.: Pituitary pro-opiomelanocortin-cell carcinoma occurring in connection with a glioblastoma in a patient with Cushing's disease and subsequent Nelson's syndrome. Clin. Endocrinol. (Oxf) 1986, 35: 117-126.

16. Orth, D.N.: Corticotropin releasing hormone in humans. Endocrin. Rev. 1992, 13: 1640191.

17. Hermus, A.R., Pieters,G.F., Smals, A.G. et al.: Transition from pituitary dependent to adrenal dependent Cushing's syndrome. NEMJ 1988, 318: 966-970.

18. Liddle, G.W., Nicholson, W.E., Island, D.P., Orth, D.N., Abe, K. and Lowder, S.C.: Clinical and laboratory studies of ectopic humoral syndromes. Recent Prog. Horm. Res. 1969, 25: 283-314.

19. Becker, M. and Aron, D.C.: Ectopic ACTH syndrome and CRH-mediated Cushing's syndrome. Endcrin. Metab. Clinics. of North Amer. 1994, 3: 585-606.

20. Clark, A.J.L., Stewart, M.F., Lavender, P.M. et al.: Defective glucocorticoid regulation of propriomelanocortin gene expression and peptide secretion in a small cell lung cancer cell line. J. Clin. Endocrin. Metab. 1990, 70: 485-490.

21. Strott, C.A., Nugent, C.A. and Tyler, F.H.: Cushing's syndrome caused by bronchial adenomas. Am. J. Med. 1968, 44: 97-104.

22. Mason, A.M.S., Ratcliffe, J.G., Buckle, R.M. and Mason, A.S.: ACTH secretion by bronchial carcinoid tumors. Clin. Endocrinol. (Oxf), 1972, 1: 3-25.

23. Zarate, A., Kovacs, K., Flores, M., Moran, C. and Felix, I.: ACTH and CRF producing bronchial carcinoid associated with Cushing's syndrome. Clin. Endocrinol. (Oxf) 1986, 24: 523-529.

24. O'Brien, T., Young, W.F., Davila, D.G. et al.: Cushing's syndrome associated with ectopic production of corticotrophin releasing hormone, corticotrophin and vasopressin by a pheochromocytoma. Clin. Endocrinol. (Oxf) 1992, 37: 460-467.

25. Halbreich, U., Asnis, G.M., Shindledecker, R., Zumoff, B. and Nathan, A.S.: Cortisol secretion in endogenous depression. II. Time related functions. Arch. Gen. Psychiatry 1985, 42: 909-914.

26. Pfohl, B., Sherman, B., Schlechte, J. and Winokur, G.: Differences in plasma ACTH and cortisol between depressed patients and normal controls. Biol. Psychiatry 1985, 20: 1055-1072.

27. Kirkman, S. and Nelson, D.H.: Alcohol induced pseudo-Cushing's disease: a study of prevalence with review of literature. Metabolism 1988, 37: 390-394.

28. Kapcala, L.P.: Alcohol induced pseudo-Cushing's syndrome mimicking Cushing's disease in a patient with an adrenal mass. Am. J. Med. 1987, 82: 849-856.

29. Amsterdam, J.D., Maislin, G., Winokur, A., Berwish, N., Kling, M. and Gold, P.: The oCRH stimulation test before and after clinical recovery from depression. J. Affect. Disord. 1988, 14: 213-222.

30. Bertagna, C. and Orth, D.N.: Clinical and laboratory findings and results of therapy in 58 patients with adrenocortical tumors admitted to a single medical center (1951 to 1978). Am. J. Med. 1981, 71: 855-875.

31. Carney, J.A., Gordon, H., Carpenter, P.C., Shenoy, B.V. and Go, V.L.W.: The complex of myxomas, spotty pigmentation and endocrine overactivity. Medicine (Baltimore) 1985, 66: 301-307.

32. Young, W.F., Carney, J.A., Musa, B.U.,Wulffraat, N.M. ety al.: Familial Cushing's syndrome due to primary pigmented nodular adrenocortical disease. NEMJ 1989, 321: 1659-1664.

33. Carney, J.A. and Young, W.F.: Primary pigmented nodular adrenocortical disease and its associated conditions. The Endocrinologist 1992, 2: 6-12.

34. Malchoff, C.D., Rosa, J., DeBold, C.R. et al.: Adrenocorticotropin independent bilateral macronodular adrenal hyperplasia: an unusual cause of Cushing's syndrome. J. Clin. Endocrinol. Metab. 1989, 68: 855-860.

35. Doppman, J.L., Miller, D.L., Dwyer, A.J. et al.: Macronodular adrenal hyperplasia in Cushing's disease. Radiology, 166: 347-352.

36. Samuels, M.H. and Loriaux, D.L.: Cushing's syndrome and the nodular adrenal gland. Endocrin. Metab. Clinics of North Am. 1994, 23: 555-568.

37. Sacken, D.L., Haynes, R.B., Guyatt, G.H. and Tugwell, P.: Clinical Epidemiology: A Basic Science for Clinical Medicine, 2nd Edition, 1991, Boston, Little, Brown and Company.

38. Ransohoff, D.F. and Feinstein, A.R.: Problems of spectrum and bias in evaluating the efficacy of diagnostic tests. NEJM 1978, 299: 926-930.

39. Danese, R.D. and Aron, D.C.: Principles of clinical epidemiology and their application to the diagnosis of Cushing's syndrome: Rev. Bayes meets Dr. Cushing. The Endocrinologist 1994, 4: 339-346.

40. Findling, J.W. and Doppmann, J.L.: Biochemical and radiologic diagnosis of Cushing's syndrome. Endorinol. Metab. Clinics of North Am. 1994, 23: 511-537.

41. Mengden, T., Hubmann, P., Muller, J. et al.: Urinary free cortisol versus 17-hydroxycorticosteroids: A comparative study of their diagnostic value in Cushing's syndrome. Clin. Investig. 1992, 70: 545-548.

42. Pavlatos, F.C., Smilo, R.P. and Forsham, P.H.: A rapid screening test for Cushing's syndrome. JAMA 1965, 193: 720-723.

43. Fok, A.C.K., Tan, K.T., Jacob, E. et al.: Overnight (1mg) dexamethasone suppression testing reliably distinguishes non-Cushingoid obesity from Cushing's syndrome. Steroids 1991, 56: 549-551.

44. Crapo, L.: Cushing's syndrome: a review of diagnostic tests. Metabolism 1979, 28: 955-977.

45. Liddle, G.W.: Tests of pituitary adrenal suppressability in the diagnosis of Cushing's syndrome. J. Clin. Endocrinol. Metab. 1960, 12: 1539-1560.

46. Yanovski, J.A., Cutler, Jr. G.B., Chrousos, G.P. et al.: Corticotropin releasing hormone stimulation following low dose dexamethasone administration. JAMA 1993, 269: 2232-2238.

47. Jackson, R.V., Grice, J.E., Jackson, A.J. and Hockings, G.I.: Naloxone induced ACTH release in man is inhibited by clonidine. Clin. Exp. Pharmacol. Physiol. 1990, 17: 179-84.

48. Besser, G.M. and Edwards, C.R.W.: Cushing's syndrome. Clin. Endocrinol. Metab. 1972, 1: 451-490.

49. Raff, H. and Findling, J.W.: A new immunometric assay for corticotropin evaluated in normal subjects and patients with Cushing's syndrome. Clin. Chem. 1989, 35: 596-600.

50. Flack, M.R., Oldfield, E.H., Cutler, Jr. G.B. et al.: Urinary free cortisol in the high dose dexamethasone suppression test for the differential diagnosis of the Cushing's syndrome. Ann. Intern. Med. 1992, 116: 211-217.

51. Liddle, G.W., Estep, H.L., Kendall, Jr. J.W., Williams, Jr. W.C. and Townes, A.W.: Clinical application of a new test of pituitary reserve. J. Clin. Endocrinol. Metab. 1959, 19: 875-894.

52. Avgerinos, P.C., Yanovski, J.A., Oldfield, E.H., Nieman, L.K. and Cutler, Jr. G.B.: The metyrapone and dexamethasone suppression tests for the differential diagnosis of the Cushing's syndrome. Ann. Intern. Med. 1994, 121: 318-327.

53. Nieman, L.K., Oldfield, E.H., Wesley, R. et al.: A simplified morning ovine corticotropin releasing hormone stimulation test for the differential diagnosis of ACTH-dependent Cushing's syndrome. J. Clin. Endocrinol. Metab. 1993, 77: 1308-1312.

54. Oldfield, E.H., Doppmann, J.L., Nieman, L.K. et al.: Petrosal sinus sampling with and without corticotropin releasing hormone for the differential diagnosis of Cushing's syndrome. NEJM 1991, 325: 897-905.

55. Findling, J.W.: Inferior petrosal sinus sampling in the differential diagnosis of ACTH dependent Cushing's syndrome: Requim for dexamethasone suppression testing. Program of the 74th Annual Meeting of the Endocine Society, San Antonio, 1992, p.46.

56. Hall, W.A., Luciano, M.G., Doppmann, J.L. et al.: Pituitary magnetic resonance imaging in normal volunteers: Occult adenomas in the general population. Ann. Intern. Med. 1994, 120: 817-820.

57. Ross, N.S. and Aron, D.C.: Hormonal evaluation of the patient with an incidentally discovered adrenal mass. NEJM 1990, 323: 1401-1405.

58. Philopponneau, M., Nocaudie, M., Epelbaum, J. et al.: Somatostatin analogs for the localization and preoperative treatment of an adrenocorticotropin secreting bronchial carcinoid tumor. J. Clin. Endocrinol. Metab. 1994, 78: 20-24.

59. Orth, D.N. and Liddle, G.W.: Results of treatment in 108 patients with Cushing's syndrome. NEJM 1971, 285: 243-247.

TABLE II

Relative Prevalences of Various Types of Cushing's Syndrome among 630 Patients Studied at Different Times

______________________________________________________________________________

Diagnosis Percent of Patients

Corticotropin-dependent Cushing's syndrome

Cushing's disease 68

Ectopic corticotropin syndrome 12

Ectopic CRH syndrome <1

Corticotropin-independent Cushing's syndrome

Adrenal adenoma 10

Adrenal carcinoma 8

Micronodular hyperplasia 1

Macronodular hyperplasia <1

Pseudo-Cushing's syndrome

Major depressive disorder 1

Alcoholism <1

______________________________________________________________________________

Modified from ( ).

Table III

Causes of Ectopic Corticotropin Syndrome

______________________________________________________________________________

Jex Howlett Imura Findling

n 25 16 30 49

Oat cell bronchus 20 19 33 10

Adenocarcinoma of the bronchus - - 10 -

Squamous carcinoma of the bronchus - - 3 -

Undifferentiated carcinoma of the bronchus - - 10 -

Bronchial carcinoid 28 37 6 49

Thymic carcinoid 8 12 - 16

Malignant thymoma - - 13 -

Pheocromocytoma 12 6 - -

Pancreatic islet cell carcinoma 16 12 6 4

Pancreatic adenocarcinoma 4 - - 2

Oat cell carcinoma of the esophagus - - 6 -

Ovarian carcinoma 4 - - -

Miscellaneous - - 10 -

Unknown - - - 12

______________________________________________________________________________

Modified from ( ).

TABLE IV

Sensitivity and Specificity and Their Derivatives

______________________________________________________________________________

Disease

+ -

Test Result + True positive (TP) False positive (FP)

- False negative (FN) True negative (TN)

Sensitivity = TP / TP +FN Specificity = TN / TN + FP

Positive Predictive Value = TP / TP + FP Negative Predictive Value = TN / TN + FN

Diagnostic Accuracy = TP + TN / TP + FP + TN + FN

Likelyhood ratio positive = sensitivity / (1- specificity) = (TP rate) / (FP rate)

Likelyhood ratio positive = (1-sensitivity) / specificity = (FN rate) / (TN rate)

Post-test odds of disease = (prevalence)(sensitivity) / (1-prevalence)(1-specificity)

Post-test probability of disease

= (prevalence)(sensitivity) / [ (prevalence)(sensitivity) + (1-prevalence)(specificity)]

______________________________________________________________________________

TABLE V

Positive and Negative Predictive Values as a Function of Disease Prevalence Assuming Test Sensitivity and Specificity of 90% Each

______________________________________________________________________________

Disease prevalence or pre-test 0.1% 1% 10% 50% 90%

probability

Positive predictive value 0.89% 8.33% 50.0% 90.0% 98.78%

Negative predictive value 99.99% 99.89% 98.78% 90.0% 50.0%

______________________________________________________________________________

TABLE VI

Operating Characteristics of Clinical Sings and Symptoms in the Diagnosis of Hypercortisolism

______________________________________________________________________________

Diagnostic finding Sensitivity Specifity Likelyhood ratio

Present Absent

% %

Central obesity 90 71 3.1 0.14

Weakness 65 93 9.3 0.38

Plethora 82 69 2.6 0.26

Leucocyte count > 11000 58 70 1.9 0.60

Acne 52 76 2.1 0.63

Striae 46 78 2.1 0.69

Diastolic BP > 105 mmHg 39 83 2.3 0.73

Pitting edema 38 83 2.2 0.75

Hirsutism 50 71 1.7 0.70

Ecchymoses 53 94 8.8 0.50

Serum K < 3.6 mg/L 25 96 6.2 0.78

Oligomenorrhea 72 49 1.4 0.57

Abnormal glucose tolerance test 88 23 1.1 0.52

Generalized obesity 3 38 0.05 2.5

Osteoporosis (early diagnosis) 64 97 21 0.37

Osteoporosis (late diagnosis) 26 94 4.8 0.79

______________________________________________________________________________

Modified from ( ).

TABLE VII

Test Characteristics of the 1 mg Overnight Dexamethasone Suppression Test

______________________________________________________________________________

Sensitivity Specificity

% %

Normal controls 98.1 98.9

All controls 98.1 89.5

Obese and other controls 98.1 80.5

______________________________________________________________________________

Modified from ( ).

TABLE VIII

Comparison of Criteria Chosen to Yield 100% Specificity for Diagnosis of Cushing's Syndrome in 58 Patients with Mild Hypercortisolism

______________________________________________________________________________

Test variable Criterion Specificity Sensitivity Positive Negative Diagnostic

Predictive Predictive Accuracy

Value Value

% % % % %

Low dose dexamethasone suppression test

17-OCHS day 4 >11 umol/d 74 69 84 54 71

17-OCHS day 4 >14.6 umol/d 100 54 100 51 69

Urine free cortisol >100 nmol/d 100 56 100 53 71

day 4

CRH test without dexamethasone pretreatment

Cortisol sum of >3450 nmol/L 100 64 100 58 76

post-CRH levels

CRH test peak >35 pmol/L 100 13 100 36 41

corticotropin

Dexamethasone-CRH test

Basal cortisol >38.0 nmol/L 100 90 100 84 91

(before CRH)

Cortisol 15 min >38.o nmol/L 100 100 100 100 100

after CRH

Corticotropin 30 >3.5 pmol/L 100 74 100 66 83

min after CRH

______________________________________________________________________________

Modified from ( ).

TABLE IX

High-Dose Dexamethasone Suppression Testing in the Differential Diagnosis of Corticotropin-Dependent Cushing's Syndrome

______________________________________________________________________________

Clinical series Sensitivity Specificity Accuracy

% % %

Oldfield et al. 81 79 81

n=183

Tabarin et al. 85 80 81

n= 27

Findling et al. 80 67 76

n= 29

______________________________________________________________________________

Pre-test probability of pituitary Cushing's disease = 90%.

Modified from ( ).