Contents
Contributors
Search

I. OVERVIEW: NONSTIMULATED HORMONE MEASUREMENTS

In evaluation of the hypothalamic pituitary adrenal (HPA) axis, static measurement of hormones are seldom useful. In general if one is suspicious of hypofunction of the HPA axis, then dynamic testing to stimulate either adrenocorticotrophic hormone (ACTH) or cortisol is indicated. On the other hand, if one is concerned about an overproduction of cortisol or ACTH, then response to suppression with dexamethasone is usually indicated. Part of the reason for the lack of enthusiasm in measurement of static hormones, is the variable nature of their secretion in normal physiological states. Cortisol is released from the adrenal gland in a pulsatile fashion under the direction of ACTH. Furthermore the ACTH is responsive to the hypothalamic factor, corticotropin releasing hormone (CRH), which is also released in an episodic manner. Cortisol secretion follows a diurnal variation, with peak values at 08:00 and a nadir at 22:00. Superimposed on this diurnal pattern are 8-10 pulsatile peaks released during the course of a 24 hour period in 8-10 pulses. Therefore depending on the instance that blood is sampled, there can be significant variation in the absolute values of ACTH and cortisol.

The hypothalamic factor, CRH is not readily measured in the blood and the reference values for normal values have not been established in the literature. The measurement of CRH levels may be, theoretically, useful in evaluating the possibility of the syndrome of ectopic CRF production (see discussion below on non-adenomatous pituitary dependent hypercortisolism).

Measurement of cortisol in serum is influenced by the presence of cortisol binding globulin (CBG). Estrogens and acute hepatitis will increase the CBG and measurement of total serum cortisol will be high. It is for this reason, and the pulsatile nature of release, that 24 hour urine collections are performed for measurement of free cortisol. Other problems with measurement of CRF and ACTH are related to its short half life (minutes) and its instability which makes it subject to proteolytic degradation.

A. Cortisol

1. Isolated serum cortisol values. Random measurements of serum cortisol are usually not helpful in the diagnosis of intactness of the HPA axis. Nevertheless in assessing adrenal insufficient states, if the serum cortisol is > 20 µg/dl one can likely rule out the possibility of primary adrenal insufficiency and cortisol concentrations < 5µg/dl usually make one suspicious of primary adrenal insufficiency, especially if the patient is acutely ill or stressed.

When assessing states of cortisol excess a random cortisol concentration of < 5µg/dl can usually exclude Cushing's syndrome, even if obtained during the nadir of its profile. However a value of serum cortisol >50µg/dl, although not diagnostic, is certainly suggestive of hypercortisolism.

2. Frequent sampling studies. Frequent sampling of blood for measurement of cortisol (every 20 to 30 minutes) for 24 hours can be helpful to diagnosis abnormalities in the HPA axis. Specifically lack of diurnal variation, coefficient of variation can all be useful to diagnose Cushing's disease (4). A single value of cortisol =5 µg/dl is rarely found in Cushing's patients.

3. Salivary cortisol determination. Less invasive is measurement of salivary cortisol. Although subject to the same variability on episodic secretion as serum cortisol, is not dependent on CBG. Methods have been developed for measurement of salivary cortisol (see below).

4. Urinary free cortisol measurements. 24 hour urine collections and measurement of free cortisol is probably the single most useful test in assessing states of excess cortisol production given the difficulties with other tests, but still it is far from perfect. In a careful study of normal subjects by de Boss Kuala et al. there was considerable intrasubject-variation in urinary excretion of free cortisol with the maximum/minimum value ratio ranging from 1.0 to 4.9 (5). Intrasubject variation in maximum/minimum creatinine excretion ranged from 1.0 to 1.5 (median 1.1). Since the ratio of free cortisol/creatinine also varied considerably (range 1.0-3.7; median 1.3), intravariation in urinary cortisol excretion could not be attributed to variation in creatinine excretion. Therefore although a useful test, there is a great deal of variability in the measurement from one normal individual to the next and with the same individual. In addition to biological variation being the etiology of the variability, difficulty in over or under collection of the urine and drugs that the patient may be taking during collection that interfere with the measurements.

Quantitation of urine cortisol with HPLC is generally felt to be more reliable than the extraction methods that then use an RIA or ELISA method (see methodology section). The problem with interfering metabolites supports the use of HPLC. One can also question whether cortisol is the actual active metabolite being measured and whether it is the correct metabolite to measure.

Lower limits of normal for the test have not really been established and it would be incorrect to make a diagnosis of adrenal insufficiency relying only on 24 hour urine collections.

B. ACTH

ACTH measurements, while subject to the same circadian variability as cortisol (actually it is the variability of the ACTH that is directly responsible for the variability of the cortisol), are not subject to the effects of CBG. However, ACTH is heat labile, and if not collected and preserved on ice, proteolysis can reduce the plasma concentration. Unlike cortisol which is measured in serum, ACTH is measured in plasma. Values of ACTH > 100 pg/ml in the setting of possible adrenal insufficiency are usually suggestive of primary adrenal insufficiency, while values >500 pg/ml are diagnostic. Low concentrations of plasma ACTH are not diagnostic, except for the undetectable levels observed in patients with cortisol producing adrenal adenomas. Plasma ACTH concentration is also low in patients taking exogenous steroids.

Salivary measurement of ACTH has not been reported. Frequent blood sampling for ACTH is as useful for assessment of hypercortisolism as cortisol (see above).

C. Miscelaneous Non-Stimulated Measurments

The use of antiadrenal antibodies has been suggested as useful in detecting early evidence of adrenal insufficiency, before cortisol values are decreased even in response to stimuli. While there are no standardized assays for antiadrenal antibodies and are only available in specialized laboratories, there is no known correlation between antibody titer and physiology.

Serum concentration of CRF is markedly elevated in pregnancy, presumably due to the production of CRF by the placenta. High levels are associated with high levels of CRH binding protein. Although mentioned as useful in the diagnosis of ectopic CRH syndromes, little data is available in this regard.

II. LABORATORY MEASUREMENTS

Various laboratory methods are available from a range of manufacturers which include both manual and automated methods. The manufacturer and details regarding th methods for each of the pertinent tests are listed in Table I-VII.

III. DYNAMIC TESTING

A. Glucocorticoid Deficiency

Despite that more than 35 years have elapsed since the initial description of the use of the insulin tolerance test (ITT) to diagnose adrenocortical deficiency (15), and more than 200 scientific publications in this area, clinicians today still argue as to which is the most sensitive and specific test to diagnose adrenocorticol deficiency. The ITT is still regarded as the gold standard upon which to compare all other tests of HPA axis function. Unfortunately, this test has a considerable spectrum of intra-individual and inter-individual variation (6, 24). Therefore when comparing other tests to the "gold standard", if the standard is not reliable, how can one determine the effectiveness of the other forms of testing? The problem lies in the ability of a single laboratory to know what the values are for their tests. Therefore ranges from an ITT test response in normal subjects performed in one laboratory may not be normal for another laboratory. Taking this into account there are some general guidelines that are available for treating patients with suspected adrenal insufficiency.

1. Primary Adrenal Insufficiency

a. High dose ACTH stimulation Test

WHEN TO USE THIS TEST: Patients acute ill in the hospital or clinic who present with signs and symptoms suggestive of primary adrenal insufficiency. Patients who are thermodynamically unstable should be resuscitated with crystalloid and given dexamethasone prior to testing if the diagnosis of primary adrenal insufficiency is being considered.

PROCEDURE: An intravenous line is placed 30 minutes before the test for rapid phlebotomy and to eliminate a temporary rise in cortisol associated with a needle stick. The i.v. line is to be kept open with 0.9% NaCl at rate of 50 ml/hr. Blood is drawn at 0' for ACTH (2 ml in a lavander top tube on ice) and cortisol (2 ml in a red top tube). Cotrosyn, 0.25 mg is administered as an IV bolus over 2 minutes. The Cotrosyn comes as a lyophilized powder which should be reconstituted with 1 ml of 0.9% NaCl. 30' after the injection blood is obtained from the iv line (2 ml) for cortisol. The same is repeated at 60' (2 ml) for cortisol.

SPECIAL CONSIDERATIONS: The test can be performed at any time of the day. If the patient is receiving hydrocortisone or cortisone acetate, the medication should be held for at least 12 hours prior to testing (if possible). Although the test can be performed while the patient is receiving dexamethasone, there is some cross-reactivity in some assays and cortisol levels may not be accurate. Each laboratory should determine for itself, the effect of dexamethasone on their assay. Patients with known sensitivity to Cotrosyn or its preservatives should not have it administered.

CONTRAINDICATIONS: Hypersensitivity to cosyntropin or any component of the formulation

WARNINGS / PRECAUTIONS: Use with caution in patients with pre-existing allergic disease or a history of allergic reactions to corticotropin. Class C in pregnancy.

ADVERSE REACTIONS 1% to 10%: Cardiovascular: Flushing Central nervous system: Mild fever Dermatologic: Pruritus Gastrointestinal: Chronic pancreatitis

<1%: Hypersensitivity reactions

DRUG INTERACTIONS: Decreased effect: May decrease effect of anticholinesterases in patients with myasthenia gravis; nondepolarizing neuromuscular blockers, phenytoin and barbiturates may decrease effect of cosyntropin

INTERPRETATION OF RESULTS: Baseline cortisol values <5 µg/dl and ACTH concentrations >100 pg/ml are usually diagnostic of primary adrenal insufficiency. The normal peak cortisol value post stimulation should be an increment no less than 7µg/dl and a maximal level >20 µg/dl at 30'. Since 37% of subjects had a peak response to Cotrosyn at 30' and 63% had a peak response at 60', both time points are analyzed in all patients and if either the 30' or 60' sample reaches the criteria as noted above, the test is considered normal (25).

Serum aldostserone can be measured in 0', 30' and 60' blood samples as ACTH stimulation of the adrenal cortex will also stimulate aldosterone. It has been suggested that a normal aldosterone response to ACTH in the presence of a suboptimal cortisol response is diagnostic of secondary adrenal insufficiency (22).

b. Low dose ACTH stimulation Test

WHEN TO USE THIS TEST: Patients with subtle signs of adrenal insufficiency or patients who have been treated with glucocorticoids and determination of adrenal reserve is necessary. Patients who have autoimmune disease and may have early adrenocortical insufficiency may be best assessed with this test.

PROCEDURE: An intravenous line is placed 30 minutes before the test for rapid phlebotomy and to eliminate a temporary rise in cortisol associated with a needle stick. The i.v. line is to be kept open with 0.9% NaCl at rate of 50 ml/hr. Blood is drawn at 0' for ACTH (2 ml in a lavander top tube on ice) and cortisol (2 ml in a red top tube).

Cotrosyn, 1 µg is administered as an IV bolus over 2 minutes. The injection material was prepared according to the method of Dickstein as follows: The Cotrosyn was diluted with 50 ml of sterile saline to a stock concentration of 5 µg/ml. Aliquots of 0.2 ml were aliquoted into sterile plastic tubes and kept at 4oC for a maximum of 4 months (7). Immediately prior to testing 0.8 ml of saline is added to the tube (final dilution 1 µg/ml) and 1 ml is injected into the patient. 30' after the injection blood is obtained from the iv line (2 ml) for cortisol. The same is repeated at 60' (2 ml) for cortisol.

SPECIAL CONSIDERATIONS: Same as for high dose ACTH stimulation test, see above.

INTERPRETATION OF RESULTS: This test was originally developed to be more sensitive for diagnosing secondary adrenal insufficiency because of it was more of a "physiologic" dose. It is much better at diagnosis secondary adrenal insufficiency than the high dose, although it is not at all recommended in acute or recent hypopituitarism when the intact adrenal glands can still respond to normally to any dose of ACTH. Although probably not useful for the initial purpose of secondary adrenal insufficiency, it may be more sensitive at distinguishing more mild forms of primary adrenal insufficiency (16). Furthermore, this low dose test was helpful in identifying mild adrenal suppression in asthmatic children being treated with inhaled steroids (14). As noted above, each laboratory should establish their normal values, however in general, a stimulated value at 30' or 60' greater than 20 µg/dl would be considered normal.

2. Secondary Adrenal Insufficiency (Pituitary or Hypothalamic)

c. Insulin tolerance testing (ITT)

WHEN TO USE THIS TEST: Patients in whom pituitary or hypothalamic disease may result in impaired corticotroph (or somatotroph) activity. Patients following pituitary surgery or pituitary radiation can be tested at any time, unlike the ACTH stimulation tests described above which are not useful in the acute setting. A random serum cortisol should be drawn prior to scheduling the test, as if the value is > 20 µg/dl, the test may not be helpful. This test, performed in our outpatient clinic, while relatively safe, requires a trained endocrine registered nurse to be present with the patient during the course of the test.

PROCEDURE: A 50 ml vial of 50% Dextrose should be at the patient's bedside in a syringe ready for injection before beginning the procedure.

An intravenous line is placed 30 minutes before the test for rapid phlebotomy, to eliminate a temporary rise in cortisol associated with a needle stick, and in order to have i.v. access for 50% Dextrose in the event of severe hypoglycemia. The i.v. line is to be kept open with 0.9% NaCl at rate of 50 ml/hr. Blood is drawn at 0' for cortisol (2 ml in a red top tube) and glucose (1 ml in a gray top tube). Blood glucose is also checked at the bedside with a glucose monitor.

Regular (short acting) insulin is administered as an IV bolus at a dose of 0.1 units/kg. Blood is sampled for cortisol and glucose as noted above at 10', 15', 30', 45', 60', 90' and 120'. Bedside nurse should monitor blood sugar more frequently if glucose drops below 60 mg/dl on the glucometer or if the patient complains of neuroglycopenic symptoms, such as fatigue, diaphoresis, hunger, lightheadedness or nausea. The test should continue until the blood sugar drops below 40 mg/dl.

In patients with diabetes on insulin consideration should be given that they may be insulin resistant. In which case, larger doses of insulin may be given. We usually begin with a single bolus of 0.1U/kg and then re-bolus with insulin depending on the response to the initial dose (either give the same dose again if there was some response but insufficient, or doubling the dose if there was only minimal response to blood glucose, or giving half the dose if the hypoglycemic response was close to the desired goal). This can be repeated several times until adequate hypoglycemia is reached.

Once the goal response of a glucose < 40 mg/dl is reached, patients can be fed a meal such as crackers and orange juice. Blood glucose should be checked at 5', 10' and 15' minutes post feeding. If there is no increase in glucose or a clinical response within 5 minutes intravenous glucose should be administered. If no response, then repeat bolus of glucose is suggested. If no response or iv access is lost, glucagons 1 mg im can be given.

SPECIAL CONSIDERATIONS: The test can be performed at any time of the day, although due to the need for patients to be fasting it is most conveniently done in the morning. If the patient is receiving hydrocortisone or cortisone acetate, the medication should be held for at least 12 hours prior to testing (if possible). Unlike the ACTH stimulation tests, the ITT cannot be performed while the patient is receiving dexamethasone, due to suppression of the hypothalamic pathways necessary to respond to hypoglycemia.

In general ITT is not recommended in patients with uncontrolled seizure disorder or significant coronary artery disease.

In order to determine if the level of dysfunction is at the hypothalamus or at the pituitary this test is sometimes used in addition to the CRH stimulation test. When the ITT fails to stimulate cortisol, but the CRH test does stimulate it is likely that the patient is having hypothalamic dysfunction.

INTERPRETATION OF RESULTS: Serum cortisol should increase within 30 minutes of the hypoglycemic response to > 20 µg/dl. If the serum cortisol at baseline is =18 the test may not be diagnostic as there are no standard values for subjects who start the test with high serum cortisol. Although the response of cortisol is more reproducible than that of growth hormone in the ITT, intra-subject differences have been reported (8, 23, 24).

b. Metyrapone testing

WHEN TO USE THIS TEST: This test is perhaps the most sensitive to determine whether the HPA axis is intact. Although the metyrapone is not generally available from your neighborhood pharmacy, it can be obtained by calling Novartis Pharmaceutical Corp. at 1-800-988-7768 on weekdays. Metyrapone blocks 11-beta hydroxylase and results in the inhibition of conversion of 11-deoxycortisol to cortisol. Serum levels of cortisol decrease and concentration of 11-deoxycortisol increases, however 11-deoxycortisol does not down regulate ACTH. Therefore in a normally functioning HPA axis there is an increase in 11-deoxycortisol. This metabolite can be directly measured in the serum or measured in the urine as 17-OH corticosteroids. This test will be abnormal in either primary adrenal deficiency or ACTH deficiency.

PROCEDURE: For assessment of adrenal or pituitary insufficiency the test can be performed as an overnight test. Metyrapone is given orally (30 mg/kg body weight, or two grams for <70 kg, 2.5 grams for 70 to 90 kg, and three grams for >90 kg body weight) at midnight with a glass of milk or a small snack (13). Serum 11-deoxycortisol and cortisol are measured at 8 AM the next morning; plasma ACTH can also be measured (10).

SPECIAL CONSIDERATIONS: The concurrent use of glucocorticoids will interfere with the test. Any medications that the patient is taking which increase the P450 enzymes will increase the metabolism and clearance of the metyrapone such as rifampin and phenobarbital and phenytoin (12). Similarly hypothyroidism or hyperthyroidism will effect clearance of metyrapone and the adrenal responsiveness. Therefore thyroid function tests should be measured prior to performing this test. Measurement of 11-deoxycortisol, like cortisol itself is dependent on CBG and drugs such as estrogens and oral contraceptives will falsely increase the concentrations of 11-deoxycortisol (19).

PREGNANCY IMPLICATIONS - Use during pregnancy only if clearly needed. Subnormal response may occur in pregnant women and the fetal pituitary may be affected. Excretion in breast milk is unknown; use with caution.

LACTATION - Excretion in breast milk unknown/use caution

ADVERSE REACTIONS - Frequency not defined. Central nervous system: Headache, dizziness, sedation. Dermatologic: Allergic rash. Gastrointestinal: Nausea, vomiting, abdominal discomfort or pain. Hematologic: Rarely, decreased white blood cell count or bone marrow suppression.

INTERPRETATION OF RESULTS: 8 AM serum 11-deoxycortisol concentrations should be >7 µg/dL with serum cortisol less than 5 µg/dL (138 nmol/L), confirming adequate metyrapone blockade. The plasma ACTH concentration at 8 AM should exceed 75 pg/mL (17 pmol/L), confirming that any increases in serum 11-deoxycortisol concentrations are ACTH-dependent, thereby separating primary from secondary adrenal insufficiency (9, 17).

B. Glucocorticoid Excess

1. Dexamethasone Suppression Tests

Measurement of endogenous cortisol production in response to exogenous dexamethasone suppression was the first provocative test and still remains among the most useful tests used for the evaluation of excess cortisol. Dexamethasone, due to its high affinity to the glucocorticoid receptor is a potent inhibitor of ACTH synthesis and release. In addition, most of the radioimmunoassays for cortisol (both urine and serum) utilize an antibody that does not cross-react with dexamethasone. Therefore the combination of being able to use relatively low doses and at the same time not interfere with the measurement of cortisol make dexamethasone suppression useful for establishing the presence of a perturbation in the pituitary - adrenal axis and for diagnosing the etiology of hypercortisolism.

At least five different tests have been described using dexamethasone, which differ in the dose and timing of dexamethasone treatment and differ in whether there is measurement of urine or serum cortisol or 17-OH-corticostseroids (Table VIII). Although the endocrine basis for the tests are in general the same, none are perfect. Confirming the diagnosis of patients with suspected hypercortisolism requires several tests for accurate diagnosis.

All these tests require significant patient participation as the patients are required to self-administer the dexamethasone at inconvenient hours of the day (11PM) or up to 4 times a day. Sampling requires either collection of urine for 24 hours or coming to the physician's office at 8 AM for multiple blood sampling. Drugs that induce hepatic cytochrome P-450 enzymes, such as barbiturates, phenytoin, rifampin and aminoglutethimide, increase the metabolism of dexamethasone and other steroids. Measurement of serum dexamethasone a few hours after the last dose will help determine if there is abnormal metabolism. All these caveats are in addition to the other problems associated with measurement of cortisol as noted above, including the variable diurnal variation as well as interference with concurrent administration of glucocorticoids, estrogen or other medication that increase cortisol binding globulin.
A popular screening test for confirming hypercortisolism is the overnight 1mg dexamethasone. A single dose of 1 mg is administered (or 0.3 mg/Kg for children (11)) at 11PM and blood is obtained by 8AM the following morning. The dexamethasone dose is given prior to the diurnal rise in endogenous ACTH release and therefore suppresses the early AM cortisol. A normal response would be a serum cortisol concentration < 5 µg/dl, if >10 µg/dl the likelihood of hypercortisolism is high. The other dexamethasone suppression tests are reviewed in Table VIII. Patients with corticotroph macroadenomas or very active tumors, that may not be so large in which patients have urine free cortisol in excess of 1000 µg/dl will need higher doses of dexamethasone to confirm suppressiblity, even and rule out ectopic ACTH production (1)

2. CRH stimulation test

WHEN TO USE THIS TEST: This test is one of the most sensitive to determine if there is an abnormality in the HPA axis and for diagnosing the etiology of hypercortisolism, especially when used in conjunction with dexamethasone (see below). Although the CRH is expensive ($300), when one considers the cost of multiple urine collections and analyses of cortisol as well as the cost of a single MRI of the pituitary (which generally exceeds $1500), CRH is at least cost effective when one considers the overall expense in the evaluation of these patients.

PROCEDURE: An intravenous line is placed 30 minutes before the test for rapid phlebotomy and to eliminate a temporary rise in cortisol associated with a needle stick. Blood is drawn at -15' and 0' for cortisol and ACTH (2 ml in a lavender top tube on ice). CRF is then injected IV at a dose of 1 µg/Kg up to a maximum of 200 µg. Blood is obtained at 15, 30 60, 90 120, 180 and 210 minutes for cortisol and ACTH (2 ml in a lavender top tube on ice).

SPECIAL CONSIDERATIONS: The test can be performed at any time of the day, although the initial studies describing the test have been done in the morning.

Side effects the patient may experience are: 1) slight nausea, 2) metallic taste, 3) urgency to urinate, 4) a change in blood pressure (either increase or decrease), 5) a change in heart rate, 6) headaches, 7) abdominal discomfort, 8) facial flushing, and 9) lightheadiness. These side effects are mild and last for only few minutes. One patient reported chest pain during the infusion that was associated with a drop in blood pressure. In general the test is very well tolerated.

INTERPRETATION OF RESULTS: The mean ACTH concentrations at 15 and 30 min after CRH should increase by at least 35% above the mean basal value at -15 and 0 min in patients with Cushing's disease, but not in patients with ectopic ACTH secretion. This measure gave the best sensitivity (93%) and specificity (100%) (2, 20). The best cortisol criterion was a mean increase at 30 and 45 min of 20% or more above mean basal values, which gave a sensitivity of 91% and a specificity of 88%. It should be noted that the criterion for Cushing's disease is based on the presence of hypercortisolism. The CRF test will not adequately differentiate subjects with pseudoCushings and those with true pituitary dependent Cushing's disease.

3. CRH Test with Dexamethasone

WHEN TO USE THIS TEST: Several investigators have found that modifications of the CRH stimulation test can increase further the sensitivity and specificity in the diagnosis of the etiology of Cushing's disease. While the simultaneous use of vasopressin can augment the response to CRH, dexamethasone can be used to suppress all but pathologic responses to CRH stimulation (27). While without dexamethasone the sensitivity and specificity of the CRH test is 65 and 100%, respectively, while with dexamethasone the CRH test is 100% sensitive and specific. In our clinic we screen all patients with elevated urine free cortisol and detectable ACTH with a dexamethasone suppressed CRF test.

PROCEDURE: Dexamethasone, 0.5 mg is self-administered po by the patient every 6 hours for 2 days, at 6AM, 12 Noon, 6 PM and 12 midnight. On the morning of the 3rd day an additional dose of dexamethasone is given at 6AM. The patient arrives at the testing center by 8AM and an intravenous line is placed 30 minutes before the test for rapid phlebotomy and to eliminate a temporary rise in cortisol associated with a needle stick. Blood is drawn at -15' and 0' for cortisol and ACTH (2 ml in a lavender top tube on ice). CRF is then injected IV at a dose of 1 µg/Kg up to a maximum of 200 µg. Blood is obtained at 15, 30 60, 90 120, 180 and 210 minutes for cortisol and ACTH (2 ml in a lavender top tube on ice).

SPECIAL CONSIDERATIONS: The test can be performed at any time of the day, although it is usually done in the morning.

Side effects the patient may experience are: 1) slight nausea, 2) metallic taste, 3) urgency to urinate, 4) a change in blood pressure (either increase or decrease), 5) a change in heart rate, 6) headaches, 7) abdominal discomfort, 8) facial flushing, and 9) lightheadiness. These side effects are mild and last for only few minutes. One patient reported chest pain during the infusion that was associated with a drop in blood pressure. In general the test is very well tolerated.

INTERPRETATION OF RESULTS: A plasma cortisol concentration greater than 1.3 µg/dl measured 15 minutes after the administration of CRH correctly identified all cases of Cushing's syndrome and all cases of pseudo-Cushing's states (100% specificity, sensitivity, and diagnostic accuracy). Patients with ectopic ACTH production will have nonsuppressed cortisol and ACTH levels that are not stimulated by CRH.

4. Inferior Petrosal Sinus Sampling (IPSS) with CRH stimulation

WHEN TO USE THIS TEST: Once the diagnosis of Cushing's disease has been made based on endocrinologic testing, the final step in the evaluation of such patients should be an MRI of the pituitary to confirm the presence of a pituitary mass. Unfortunately, MRI imaging of the pituitary as a primary diagnostic tool is distinctly unhelpful due to the fact that 10% of all normal individuals may have slight abnormalities of their pituitary and that in many subjects with Cushing's disease, the tumor may be too small to be imaged with MRI scans. However, subjecting a patient to surgical pituitary exploration in the absence of a demonstrable mass is likely to result in an unsuccessful surgery. Furthermore, if previous dexamethasone and/or CRH testing is equivocal, then IPSS should be performed to further confirm the pituitary as the source of the ACTH (26). Although this test is less reliable in lateralizing the ACTH source (i.e. left versus right), than it is in confirming that the ACTH is central in origin it can rule out ectopic ACTH production by a tumor (although ectopic CRF secreting tumors would be difficult to distinguish from true Cushings' disease based on IPSS). Simultaneous measurement of prolactin in the central samples can normalize the data if there is any difference in the location of the catheters (18).

PROCEDURE: This test is done in conjunction with a skilled interventional neuroradiologist. It is important that the endocrinologist is personally present in the room during the procedure so that assurance can be made that the proper blood tests were drawn at the specified times. The patient is brought to the angiogram suite without sedation. A large bore IV line is placed in an antecubatal fossa (to be certain there is access to peripheral blood sampling and CRF injection). Catheters (5 French) are placed in the femoral veins and threaded under fluoroscopic guidance to the inferior petrosal sinus. Injection of IV contrast confirms proper placement of the catheters.

Patients are on constant, pulse, blood pressure and oxygenation monitors during the course of the procedure. Test tubes are prechilled in ice and labeled so that during the rapid sampling period, blood can be placed in the tubes without delay.

We routinely obtain 4 baseline measurements at -15, -10, -5 and at 0 minutes. This allows for practice allowing proper coordination between the radiologists drawing blood from the IPSS and the individual drawing blood from the brachial vein. Appropriate amounts of blood should be removed to discard the dead space of the catheter (this varies depending on the size of the catheter used). 2 ml of blood is obtained in lavender top vacutainer tubes on ice for measurement of cortisol (on peripheral samples); ACTH and prolactin (on central samples).

At 0' CRF is then injected as described above for the peripheral CRF test. Blood is then sampled from both central and peripheral lines at 2', 5' 10' and 15'. After the 15' time point and right before the IPSS catheters are removed, repeat fluoroscopic localization of the catheters should be performed to confirm that there was no displacement during the sampling. However, sampling on peripheral blood may continue as described in the CRF test discussed above.

SPECIAL CONSIDERATIONS: The test can be performed at any time of the day, although it is usually done in the morning.

Side effects the patient may experience are: 1) slight nausea, 2) metallic taste, 3) urgency to urinate, 4) a change in blood pressure (either increase or decrease), 5) a change in heart rate, 6) headaches, 7) abdominal discomfort, 8) facial flushing, and 9) lightheadiness. These side effects are mild and last for only few minutes.

Patients greater than 300 pounds in weight may not be able to be supported by the standard fluoroscopic table. Furthermore such large patients may have an abdominal pannus that precludes reasonable access to the femoral veins. In such instances we have performed the IPSS in patients via catheters placed in the antecubtal vein and had the patient immobilized in the sitting position.

At our institution we have performed over 50 IPSS without complication. However there are reports in the literature that a stroke may occur (21). We recommend that the catheters remain in the petrosal sinus for no more than 30 minutes to minimize this possibility.

Freeze/thawing can decrease the immunoassayable ACTH (see above), therefore we recommend that the samples be brought to the endocrine lab and analyzed with 24 hours with the plasma separated on ice during this time. If analysis is not possible with 24 hours, the samples should be aliquoted and frozen to minimize the amount of freeze/thawing.

INTERPRETATION OF RESULTS: Plasma ACTH values are normalized to the prolactin value in order to correct for possible different localization of the catheters, or movement of the catheters during the study. The post CRF ACTH/Prolactin value of the central catheters should be >2.1 fold the ACTH/Prolactin value of the peripheral sample. In most cases of pituitary dependent Cushings, the increase is > 5.0 fold. Lateralization would mean that the ratio of the left to right side is >2.0. Often times the ratio criteria can be met without the need for CRF stimulation, however the diagnostic accuracy increases from 86% to 90% with CRF (3).