ENDOCRINE DIAGNOSIS IN PITUITARY ADENOMAS

Assessment of Normal Pituitary Function

Testing for each of the primary hormones secreted by the adenohypophysis can be complicated. Each of the six major hormones requires basal and dynamic testing to portray accurately the reserve potential of the adenohypophysis. Patients with hypopituitarism may display no outward evidence of pituitary insufficiency, since a small amount of tissue can sustain adequate function. However, impaired reserve capacity can become apparent during periods of stress. The detection of inadequate reserves requires provocative tests, as basal hormonal levels can be misleading. It is not unusual to find that reserves are normal for several pituitary hormones but marginal or absent for others. In practice, the determination of thyroid and adrenal function is more important, since only for these two glands is failure life-threatening. The failure or insufficiency of gonadotropic, somatotropic, or lactotropic function might or might not be pertinent to the health of the patient, depending on age, sex, psychological needs, and other determinants. As more information concerning the role of pituitary peptides becomes available, it becomes increasingly possible to prevent systemic changes that can impair health many decades later. For example, hyperprolactinemia, while known to affect fertility, libido, and potency, also can produce osteopenia as a result of estrogen or testosterone deficiency, and, if not corrected in young women or men may ultimately cause osteoporosis later in life, The succeeding paragraphs outline the basal and dynamic tests that should give an accurate portrayal of the adequacy of the adenohypophysis. Table-1 gives the major screening tests used to prove total hypophysectomy, and Table-2 lists normal endocrine values.

Somatotropic Function

The endocrine evaluation of the somatotropes centres on measuring serum levels of growth hormone both in the fasting state and after the administration of stimulatory or inhibitory agents. Until recently, growth hormone was thought to be necessary only for growth in children. Current endocrine investigation now suggests that this hormone may be important for adults as well, including for preservation of bone density, normal body composition, and lipid levels. Growth hormone is released in a pulsatile fashion by the adenohypophysis in response to positive control from growth hormone-releasing hormone (GHRH) and negative control from somatostatin, and serum levels are variable. A growth hormone measurement taken at random in an individual with hypopituitarism may be normal, so the reserve must be tested. In normal persons, the level of growth hormone rises after insulin­induced hypoglycaemia or after the administration of GHRH, arginine, clonidine, propranolol, or dopamine precursors (L-dopa) or agonists (apomorphine).

Insulin hypoglycaemia is the most commonly used provocative test. After the intravenous administration of regular insulin (0.1 to 0.15 U/kg), plasma glucose levels in normal individuals usually fall to less than 40 mg/dl. A decline in glucose level by at least 50 percent below baseline associated with a growth hormone level greater than 5 ng/ml indicates normal somatotropic function. Some normal subjects demonstrate levels as high as 25 ng/ml. Any response indicates surviving somatotropes.

Arginine stimulates the release of growth hormone through the alpha-adrenergic receptor mechanism and is more effective in women than in men. After its infusion in normal individuals, growth hormone levels generally rise to more than 5 ng/m!. Again, any detectable level indicates the presence of surviving somatotropes.

Clonidine, propranolol, and L-dopa also elicit a rise in human growth hormone. A level greater than 5 ng/ml indicates normal somatotrope reserve, and any detectable level of human growth hormone indicates surviving somatotropes.

Some of the tests described above can cause nausea and hypotension and must be administered under closely monitored conditions. The most commonly used test, insulin hypoglycaemia, can be hazardous in hypopituitarism patients with inadequate adrenal reserve. It should be avoided in elderly patients and in those with vascular or cerebrovascular disease or a history of a convulsive disorder.

Insulin stimulation is the best single test to verify total ablation or loss of somatotropes. The absence of growth hormone in the serum following adequate stimulation by insulin indicates that all somatotropic cells are dead or absent. Inadequate levels of growth hormone represent a blunted response. Hepatic, renal, and central nervous system diseases as well as cancer and poor nutrition also can blunt the response to insulin.

Prolactin Function

Prolactin cells make up most of the adenohypophysis. In most laboratories, 15 to 20 ng/ml is the upper limit and 5 ng/ml is the lower limit of the normal basal serum prolactin level. Prolactin secretion is pulsatile during the day and rises during sleep. In hypersecretory states associated with amenorrhea or pituitary tumors, prolactin levels are usually above 30 ng/ml, but patients without pituitary tumors can have levels as high as 100 ng/ml.

The dynamic test employed most often in the past was the thyrotropin-releasing hormone (TRH) stimulation test. TRH administration (200 µg intravenously) should cause at least a twofold rise in prolactin levels within an hour; any rise over 100 percent is considered normal. Failure to obtain a rise is almost always abnormal. A blunted response with a low basal level usually indicates inadequate lactotrope reserve. An elevated basal prolactin level with a subnormal response to TRH is consistent with a prolactin­secreting adenoma. However, because there is considerable overlap with other etiologies, a blunted rise is not pathognomonic of a tumour. Therefore, this test is no longer often used. After total hypophysectomy, no prolactin should be detectable after stimulation with TRH.

A low level of prolactin should be present in the hypopituitary state and can also be seen in patients taking dopaminergic drugs such as L-dopa or bromocriptine. High levels can be due to a number of factors, the most common being pregnancy, breast feeding, hypothyroidism, or the ingestion of drugs such as phenothiazines or high-dose estrogens.

Thyroid Function

Routine tests of thyroid function generally indicate an adequate pituitary-thyroid axis. Today, many factors can be measured, such as T3, T4, and free thyroxine. The free T4 index is a good screening test for thyroid function because it takes into account both T3 up­take and total T4 in the serum. Radioactive iodine uptake should not be used as a screening test of thyroid function. Table-2 shows the physiologic levels of these factors.

When these factors are low, one must measure serum thyroid­stimulating hormone (TSH) directly to distinguish primary from secondary hypothyroidism. However, a normal level of TSH is not a satisfactory indication of adequate thyrotrope reserve. To measure TSH reserve, TRH (200 µg) is administered intravenously. The adenohypophysis generally is considered normal if TSH values rise to 6 to 20 µU/ml. A blunted response indicates diminished reserve and, therefore, diminished activity of the adenohypophysis, presumably with depopulation of thyrotropic cells. After total hypophysectomy, the administration of TRH should result in no measurable level of TSH. A decreased response to TRH stimulation also occurs in a variety of clinical settings, including thyroid hormone or glucocorticoid therapy, hyperthyroidism, renal failure, and depression.

Adrenocorticotropic Hormone (ACTH)

ACTH is secreted in pulsatile fashion and is highest at 8 A.M. and lowest at midnight. Because ACTH is released episodically, serum cortisol levels vary widely over the day, so a random measurement is rarely useful. The exception is a morning serum cortisol level above 20 µg/dl, which indicates intact hypothalamic-pituitary­adrenal function. Twenty-four hour urine collections represent the integrated secretion of cortisol and thus are more useful than random serum levels, particularly when there is concern about possible overproduction of cortisol. The urinary free cortisol level should be less than 100 µg per 24 h, and 17 -hydroxysteroids should be 3 to 8 mg per 24 h.

The presence of circulating cortisol is evidence that some ACTH is being secreted from the pituitary gland and thus indicates incomplete hypophysectomy. However, levels may be undetectable despite the presence of a very few remaining cells. ACTH reserve can be demonstrated using the metyrapone test or insulin­induced hypoglycaemia. Totally hypophysectomized patients produce no rise in hormone levels in response to either of these provocative tests.

Metyrapone inhibits 11-,B-hydroxylase in the adrenal cortex, thus blocking the conversion of 11 -deoxycortisol to cortisol. The attendant drop in serum cortisol stimulates ACTH secretion by surviving corticotropes. The release of ACTH stimulates the synthesis of 11-deoxycortisol, the hormone proximal to the block, and the metabolites of this compound can be measured in the urine. Hence, over a 1 - or 2-day period, measurable amounts of urinary 17-hydroxycorticoids indicate production of ACTH by the pituitary gland. In normal individuals, 24-h urinary 17-hydroxycorticoid levels measured after metyrapone administration are usually two to three times higher than baseline. The level is lower in hypophysectomized patients; if hypophysectomy is complete, no 17­hydroxycorticoids are found in the urine.

A quick test involves administering metyrapone (2.0 to 3.0 g) at midnight and measuring plasma 11-deoxycortisol and plasma cortisol levels at 8 A.M. In normal subjects, the plasma cortisol value is usually less than 5.0 µg/dl, and plasma 11-deoxycortisol is greater than 7.5 µg/dl. In completely hypophysectomized patients, the plasma deoxycortisol level is negligible. Insulin-induced hypoglycaemia is also an excellent method for achieving maximal hypothalamic-pituitary-adrenal stimulation. Normal individuals will show a rise in plasma cortisol to 20 µg/dl or higher when serum glucose is lowered by 50 percent. This test is contraindicated in patients with cardiovascular disease or a convulsive disorder.

Gonadotropic Function

The plasma levels of follicle-stimulating hormone (FSH), luteiniz­ing hormone (LH), estradiol, and testosterone depict the status of the gonadotropic elements of the hypophysis. Where plasma levels of these hormones are not measurable or are low, stimulation with gonadotropin-releasing hormone (GnRH) can help to determine the presence of adequate pituitary reserve, Administration of GnRH should produce no detectable level of these hormones in patients with total hypophysectomy, and low levels when reserves are low. Any detectable level of FSH or LH indicates some remaining gonadotropic cells.

High levels of the pituitary hormones and low levels of the end-organ hormones testosterone or estradiol indicate end-organ failure, The best example of this is menopause, when estradiol is low and FSH and LH rise. When the levels of both pituitary and end-organ hormones are high, a rare pituitary adenoma secreting bioactive gonadotropins or a pituitary hormone resistance state may be present.

Prolactin-Secreting Tumors

Prolactin is under tonic inhibition by the hypothalamus and thus is unique among pituitary hormones, Dopamine is the hypothalamic agent primarily implicated in the inhibition of prolactin, Sectioning of the pituitary stalk and administration of a variety of substances can override tonic inhibition, thus causing secretion of prolactin by the adenohypophysis, In addition, a number of substances act directly on lactotropes or the hypothalamus to negate the influence of dopaminergic inhibition.

The principal prolactin-stimulating factors are not yet known, but there can be some rise in prolactin in response to high-dose estrogen and TRH, Estrogen acts directly at the level of the pituitary lactotrope to stimulate the synthesis and release of prolactin. In normal subjects, the administration of either high-dose estrogen or TRH causes a prompt two- to threefold rise of the prolactin level. The administration of serotonin or its precursors (e.g. 5-hydroxytryptophan or L-tryptophan) also increases prolactin levels. Agents like γ-aminobutyric acid, the phenothiazines, metoclopramide, opiate agonists, and intravenous cimetidine cause the prolactin level to rise to various degrees, whereas serotonin blockers, L-dopa, and other dopaminergic agents can reduce the level of prolactin.

Diagnosis

Men and women with large prolactin-secreting tumors present no major diagnostic problem because a positive magnetic resonance imaging (MRI) scan and an elevated level of prolactin together are diagnostic, especially if the prolactin level is greater than 200 to 300 ng/ml. Lower levels may be found with other central nervous system tumors or non-neoplastic lesions that functionally sever the pituitary stalk, thereby interrupting the flow of inhibitory dopamine, or with other lesions such as mixed prolactin and growth hormone-secreting tumors.

Hyperprolactinemia in women usually presents with one or more of the following: oligoamenorrhea, galactorrhea, infertility, and androgen excess (acne or hirsutism). The diagnosis of a microadenoma generally demands great experience and judgment, since endocrine and radiologic tests can be misleading, Prolactinomas are common in men (60 percent of all adenomas), but usually they are not found until they are large, since systemic complaints often are ignored for many years. Men may not seek early attention despite loss of libido and potency, headache, or fatigue. Only when visual loss or disturbing endocrine symptoms are present do they seek medical attention, by which time the opportunity for early endocrine diagnosis is lost.

Some female patients have mild rises of prolactin and growth hormone levels and report not only amenorrhea and galactorrhea but also fatigue, weakness, and slight swelling of the hands and feet. Their tumour may demonstrate both prolactin and growth hormone secreting cells. Tumour removal is followed by a return to well-being and resumption of normal menstrual function. These tumors can be termed clinically active dual-secretory adenomas or silent somatotroph adenomas.

The single most useful endocrine manoeuvre is determination of the basal serum prolactin level. In most laboratories, the normal level ranges from 5 to 20 ng/ml. Minimal elevations must be confirmed by taking several fasting samples on different days in patients who have taken no medication during the previous week.

When an elevated level is confirmed, primary hypothyroidism and pregnancy must be excluded with a TSH level and pregnancy test, since both conditions can elevate prolactin values. Profound primary hypothyroidism can be associated with thyrotrope hyperplasia, causing the appearance of an enlarged pituitary on MRI scan, which may be mistaken for a prolactinoma. The pituitary regresses with thyroid replacement therapy in this disorder. When an elevated prolactin level is confirmed but radiologic verification is absent, great care must be exercised in making decisions regarding therapy.

The higher the serum prolactin level, the more likely it is that a tumour is present. When the level is much lower than 100 ng/ml, the likelihood of finding an identifiable tumour drops by as much as 50 percent. With levels as high as 350 ng/ml, a tumour is almost always found.

The size of the prolactin-secreting tumour also is related to the level of serum prolactin, but with wide variations. Microadenomas can secrete levels as high as 550 ng/ml, whereas larger tumors (i.e., with a diameter of 2 cm or greater) can secrete levels lower than 100 ng/ml. One must keep in mind that prolactin levels as high as 150 to 200 ng/ml can be caused by partial or total functional pituitary stalk sectioning resulting from some other process or tumour. Thus, a small craniopharyngioma, a null cell tumour, an oncocytoma, or a vascular, granulomatous, or other type of lesion can account for hyperprolactinemia. In such cases, the hyperprolactinemia can revert to normal with bromocriptine therapy while the non secretory lesion grows to invasive size. In view of this possibility, it is necessary to ensure after operation that a presumed prolactinoma in fact contains prolactin granules as demonstrated by immunocytochemistry.

Ancillary Tests

Various stimulation tests (TRH, chlorpromazine, metoclopramide) and suppression tests (L-dopa, nomifensine) have historically been used to provide ancillary aid in arriving at a diagnosis. In a pragmatic sense, none of these endocrine tests is totally reliable for the diagnosis of a prolactin-secreting adenoma, and they are now rarely used. The criteria mentioned above apply to both sexes, as there is no consistent difference between normal serum prolactin levels in men and nonpregnant women.

Differential Diagnosis

Hypothyroidism, pregnancy, and medications used for psychiatric illnesses, as well as chronic disease of the liver and renal failure must be excluded, as these entities also can be associated with hyperprolactinemia. In addition, serum prolactin levels are sometimes elevated in hypoglycaemia, during stress, and in the postpartum period. Acromegaly can be associated with modest elevations in prolactin levels, sometimes as high as 150 to 200 ng/ml. Hyperprolactinemia also can accompany Cushing's disease. Since some commonly used drugs enhance prolactin secretion, especially psychotropic drugs, high doses of estrogens, and some antihypertensives, a complete drug history is mandatory.

Postoperative Evaluation

Total removal of a prolactin-secreting microadenoma returns pro­lactin levels to normal within a day or two. If the level is lower than 10 ng/ml, the surgeon can feel confident of prolonged remission if not outright cure. When the level of prolactin is normal, ovulation and menstruation return in at least 75 percent of women, and galactorrhea ceases. These findings are evidence of total remission, although recurrence is still possible. Failure of prolactin to fall to normal levels usually is caused by incomplete tumour removal. However, it also can result from injury to the pituitary stalk. If the surgeon suspects a retained fragment of tumour in the first few weeks after operation in a straightforward case, early re-exploration is warranted.

Clinical evidence of late recurrence is much more sensitive. Thus, if menses and ovulation again begin to fail and if galactorrhea resumes, endocrine verification must be undertaken quickly.

Acromegaly

Patients thought to have acromegaly on clinical grounds must undergo verification by endocrine testing and MRI scanning. The differential diagnosis must exclude individuals with a genetically determined physiognomy similar to that of the acromegalic but who do not have the disease. All individuals with true acromegaly harbour a growth-hormone-secreting adenoma or, more rarely, an ectopic tumour producing growth-hormone-releasing hormone (GHRH).

The level of growth hormone is governed by the interplay of GHRH, a 41-amino-acid peptide characterized in 1983, and somatostatin, an inhibitory l4-amino-acid peptide. The balance between these two peptides is regulated by neighbouring hypothalamic neurons. Other substances that can stimulate growth hormone release are enkephalin, glucagon, a melanocyte-stimulating hormone, vasopressin, diazepam, and estrogens. Variations during the day are quite striking. High levels of growth hormone are found during the first few hours of sleep, with smaller rises occurring during the day. For many hours no secretion is evident. The ventromedial and infundibular nuclei of the hypothalamus are involved with the pulsatile nature of the secretion.

Stimulation of growth hormone release is mediated through alpha-adrenergic and dopaminergic mechanisms. Therefore, in normal individuals, the growth hormone level increases after the administration of norepinephrine, L-dopa, clonidine, and apomorphine. In addition, behavioural conditions that stimulate alpha­adrenergic mechanisms, such as stress, hypoglycaemia, and exercise, also enhance the secretion .of growth hormone. The limbic system participates in growth hormone release through serotonergic mechanisms. Thus, the administration of serotonin precursors such as L-tryptophan or 5-hydroxytryptophan also causes growth hormone release.

Beta-receptor mechanisms inhibit the secretion of growth hormone. Therefore, isoproterenol blocks secretion, whereas propranolol, its antagonist, enhances secretion to a certain extent in response to glucagon, vasopressin, and L-dopa. Growth hormone release also is inhibited by the administration of glucocorticoids and in patients who are obese or have elevated levels of free fatty acids.

The somatomedins mediate the stimulation of cell growth by growth hormone. Two somatomedins have been identified, and their secretion by the liver depends largely on the level of growth hormone. Somatomedin C, now called insulin-like growth factor-1 (IGF-1), possesses properties similar to those of insulin, and its level is increased in acromegaly

Once acromegaly is suspected, endocrine diagnosis depends on the demonstration of (1) an elevated level of human growth hormone that fails to decrease to less than 2 ng/ml during a glucose tolerance test and (2) an elevated level of IGF-1. It must be remembered, however, that normal adolescents sometimes do not demonstrate growth hormone suppression with glucose and have higher levels of IGF-1 than adults. Almost all acromegalies have levels of growth hormone greater than 10 ng/ml, but a few have levels between 4 and 10 ng/ml.

Hyperglycaemia fails to suppress the hypersecretion of growth hormone in acromegalics. Therefore, endocrinologists rely on the oral glucose tolerance test as the best screening and diagnostic test for acromegaly. If the elevated basal level of growth hormone fails to fall below 2 ng/ml during the test, it is safe to conclude that the patient has acromegaly. In many patients with untreated acromegaly, but not in normal individuals, TRH stimulates the release of growth hormone. Surgical cure of acromegaly should eliminate this response.

The level of IGF-1 is always elevated in active acromegaly and therefore is an excellent screening marker. Normal levels are 0.34 to 1.97 lU/ml in males and 0.45 to 2.2 lU/ml in females. Higher levels confirm acromegaly. The clinical disturbances of acromegaly are related more closely to the level of IGF-1 than to the level of growth hormone.

Acromegalic patients do not display normal regulatory responses to certain agents, and often their growth hormone response is described as "paradoxical." For example, in 10 to 20 percent of patients, glucose ingestion does not cause the fall in growth hormone levels seen in normal individuals, but rather causes a paradoxical rise. L-Dopa increases growth hormone values in normal subjects, but in most acromegalics it induces a fall. Similarly, the dopaminergic agonist bromocriptine, while increasing growth hormone levels in normal individuals, causes a decrease in some acromegalies. This finding is the basis of bromocriptine therapy for acromegaly. TRH does not affect the growth hormone level in normal subjects, but increases it in 70 to 80 percent of acromegalies. The serotonergic agonist L-tryptophan, while increasing human growth hormone values in normal subjects, causes no change in patients with acromegaly. Taken together, these factors help to clarify the diagnosis in doubtful cases. They are especially helpful in adolescents and in patients thought to have gigantism.

In summary, the endocrine diagnosis of acromegaly rests on the failure of glucose loading to suppress the level of growth hormone to less than 2 ng/ml, and the presence of elevated levels of IGF-1. The other dynamic tests described are reserved for equivocal cases.

Evaluation of patients after operation, radiotherapy, or pharmacologic treatment demands the same testing. If the glucose­suppressed growth hormone level is less than 2 ng/ml and the IGF-1 level is normalized, clinical remission is almost certain in all spheres save those in which permanent changes have already taken place. Thus, bone changes and cardiac disease rarely reverse, but should not progress. Absolute cure of acromegaly is accompanied by a full return of normal dynamics, that is, the mean value of growth hormone on four to six determinations during the day would not be higher than 2 to 4 ng/ml; oral glucose would suppress growth hormone levels to below 2 ng/ml; and IGF-1 levels would return to normal. Use of the term "cure" or "remission" when resting levels of growth hormone are greater than 4 ng/ml and are not suppressed to less than 2 ng/ml with glucose loading, or when the IGF-1 level is not normal, is misleading and inaccurate.

Recurrence also can be detected by these examinations. The first indications of recurrence may be a rise in IGF-1 level or the return of a paradoxical response of growth hormone to glucose intake. It is necessary to remember that any therapy that results in apparent normalization of growth hormone levels may be accompanied by continued or later growth of the tumour. Consequently, even when the results of endocrine testing are normal, if the patient reports a return of symptoms, every effort should be made to detect a possible recurrence so that appropriate therapy can be instituted as soon as possible.

Cushing's Disease

Cushing's disease can be defined as hypersecretion of cortisol caused by an abnormality of the pituitary gland. The lesion most commonly responsible for the disease is an adenoma. More rarely, pituitary hyperplasia is the cause.

Hypercortisolemia also can be caused by a number of different systemic lesions, such as ectopic tumors that produce ACTH and adrenal cortical lesions that produce cortisol. The general syndrome caused by excess cortisol is, by convention, called Cushing's syndrome. The term Cushing's disease is used for parasellar etiologies.

The diagnostic problem thus is twofold: (1) to verify hypercortisolemia, and (2) to identify the primary cause of the hypercortisolemia.

Corticotropin-releasing hormone (CRH) is produced in neurons in the anterior median eminence of the hypothalamus and regulates normal ACTH secretion by the pituitary gland. ACTH secretion has a diurnal rhythm, which is regulated by the central nervous system. The highest levels are found early in the morning, with spiking later during the day, especially in midafternoon. The diurnal rhythm of cortisol secretion must be taken into account when resting or sporadic sampling of plasma is undertaken. The secretion of ACTH is dependent on serotonergic mechanisms and can be blocked by serotonin-blocking agents. Many forms of stress, including hypoglycaemia, also increase the production of ACTH through mechanisms mediated by the hypothalamus. Negative feedback on ACTH secretion is due primarily to the action of plasma cortisol, which exerts its effect directly on the pituitary gland and probably also on the hypothalamus.

The stimulation of the adrenal gland by ACTH causes the secretion of cortisol and other adrenal hormones. Because of similarities in structure between ACTH and melanocyte-stimulating hormone, ACTH possesses some melanocytic-stimulating activity as well.

The Demonstration of Hypercortisolemia

  Basal Levels

Screening tests can be used to determine whether cortisol secretion is within the normal range. Plasma cortisol levels vary widely and are therefore rarely useful as a screening test. A 24-h urine collection can be tested for free cortisol and 17-hydroxycorticosteroids. Elevated levels of either raises the suspicion of Cushing's syndrome. An amount of free cortisol above 100 µg per 24 h or of 17-hydroxycorticosteroids above 12 mg per 24 h suggests hypersecretion of cortisol.

  Suppression of Cortisolemia

In normal individuals, the administration of a small quantity of a glucocorticoid such as dexamethasone, which does not cross­react in the cortisol assays, lowers the level of cortisol in plasma or I7-hydroxysteroids and free cortisol in the urine. When Cushing's syndrome is caused by a pituitary disorder, the hypothalamic "set point" for dexamethasone inhibition of ACTH secretion is believed to be raised, thus requiring a higher dose of dexamethasone to suppress ACTH and cortisol secretion. When an excess of cortisol is suspected, a quick overnight test (the 1 mg overnight dexamethasone suppression test) is carried out as follows. At 11 P.M., 1 mg of dexamethasone is given by mouth. The following morning (8 A.M.), serum cortisol is determined again. If the cortisol level is less than 5 µg/dl, Cushing's syndrome can be eliminated as a possibility, with a false-negative rate of less than 3 percent.

The 1 mg overnight dexamethasone test is the best screening test for Cushing's syndrome. An abnormal test result does not establish the diagnosis, however, because there are many causes for a false-positive result (failure to suppress cortisol). An abnormal test result indicates the need for additional testing, since stress, psychiatric illnesses, estrogen therapy, phenytoin and other medications, as well as failure to do the test properly, can all cause abnormal results. If Cushing's syndrome is suspected, a low dose of oral dexamethasone (0.5 mg four times a day for 2 days) is then given. In normal individuals, the urinary free cortisol and 17­hydroxysteroid values fall over a 24-h period to less than 20 mg and 4 mg, respectively. In patients with Cushing's syndrome, levels of plasma cortisol and urinary steroids fail to suppress.

If the low-dose dexamethasone test gives an abnormal result, the patient is given a 2-day high-dose dexamethasone test. This requires the administration of four 2-mg doses daily for 2 days. The 17-hydroxycorticosteroids should be reduced by more than 64 percent and the urine free cortisol by more than 90 percent to signify suppressibility. On some occasions, more than 8 mg of dexamethasone are necessary to suppress pituitary-based Cushing's syndrome. A high-dose overnight test has been used in which 8 mg of dexamethasone is given at midnight, and plasma cortisol is measured at 8 A.M. and compared with the level obtained the previous morning. Recently, an intravenous high-dose dexamethasone suppression test has been described in which 1 mg/h of dexamethasone is administered intravenously for 7 h. A fall in cortisol of at least 7 µg/dl is consistent with suppression.

In all of these forms of dexamethasone testing, the presence of suppression is consistent with a pituitary source of Cushing's syn­drome. Adrenal and ectopic Cushing's tumors have traditionally been considered nonsuppressible. It has been increasingly recognized, however, that some ectopic ACTH-producing tumors can exhibit suppression with dexamethasone.

  Cortisol Dynamics

The demonstration of abnormal cortisol dynamics has been said to be valuable in making the diagnosis of Cushing's syndrome. In normal individuals, serum cortisol is highest at 8 A.M. and, despite pulses of secretion during the day, is lowest between 6 P.M. and midnight. This fluctuation (termed the diurnal rhythm) is lost in Cushing's syndrome, where the midnight level might be 8 µg/dl or higher (normal is <8 µg/dl). Conversely, re-establishment of a normal diurnal pattern is a reliable demonstration of remission of the disease. However, some patients with Cushing's syndrome have plasma cortisol levels that can overlap the normal range at all times of the day, so testing for loss of diurnal rhythm is not always reliable.

  Plasma ACTH Levels

Although Cushing's disease is accompanied by hypersecretion of ACTH, the levels of ACTH in plasma are often normal or only slightly elevated. ACTH levels in normal individuals vary from 50 to 100 pg/ml; in Cushing's disease, they can range from 50 to 250 pg/ml.

There is now a sensitive, specific, and rapid method for detection of ACTH levels by immunoradiometric assay. When ACTH levels are undetectable or extremely low in a patient with Cushing's syndrome, the diagnosis is ACTH-independent or adrenal Cushing's syndrome. If the ACTH levels are in the normal to mildly elevated range, the patient may have pituitary or ectopic Cushing's syndrome. When the ACTH levels are extremely high (thousands of picograms per millilitre), the patient has ectopic Cushing's syndrome.

  Differential Diagnosis-Cushing's Syndrome

The diagnosis of Cushing's syndrome may be difficult in several situations. Cortisol hypersecretion can be sporadic in Cushing's disease (called periodic or cyclic Cushing's syndrome), necessitating long-term repeated testing. A false-positive diagnosis of Cushing's disease is observed sometimes in alcoholics (alcoholic pseudo-Cushing's disease). More important, depressed patients can exhibit hypercortisolemia that does not suppress with dexamethasone and, to a certain extent, may appear to be "cushingoid" when in fact they are simply obese. These individuals may show loss of diurnal variation and increased cortisol levels during the day. They can be distinguished from patients with Cushing's disease by their lack of true clinical signs of excessive cortisol and their response to insulin-induced hypoglycaemia. Patients with Cushing's disease have no noticeable increases in plasma cortisol during periods of hypoglycemic stress, whereas depressed or normal individuals demonstrate the usual rise in cortisol levels. More recently, the CRH test has been used to distinguish depressed patients with hypercortisolemia from patients who have depression associated with Cushing's syndrome. In depression, there is a blunted response of ACTH to the injection of 1 µg/kg ovine CRH, whereas patients with Cushing's disease exhibit a normal to exaggerated response. Unfortunately, this test is not always definitive because of some overlap between groups.

Cushing's disease is associated with hypercortisolemia, loss of the diurnal rhythm of cortisol secretion, and normal to slightly raised levels of ACTH in plasma. It is important for an experienced diagnostician to exclude the many causes of false-positive test results. The definitive diagnosis is based on the failure of low-dose dexamethasone (0.5 mg orally, four times a day for 2 days) to suppress the level of cortisol, whereas high doses (2.0 mg orally, four times a day for 2 days) are successful in suppressing it.

  Differential Diagnosis-Site of Hormone Excess

A number of tumors can secrete ACTH, including small cell carcinoma of the lung, some bronchial adenomas, medullary carcinoma of the thyroid, pancreatic carcinoma, and carcinoid tumors.

In these cases the production of ACTH usually is autonomous and fails to respond to increasing doses of dexamethasone or to a CRH test. There is, however, increasing recognition of a subset of ectopic ACTH-producing tumors that result in a clinical and biochemical pattern identical to that seen in pituitary Cushing's syndrome. A major advance in addressing this problem is the bilateral inferior petrosal sinus catheterization procedure. In this radiologic procedure, catheters are placed bilaterally in the femoral veins and threaded up into the inferior petrosal veins just outside the pituitary gland. Blood is then sampled simultaneously from each side and from a peripheral (hand) vein and assayed for ACTH. If the pituitary/peripheral ACTH ratio is ≥2, the Cushing's syndrome has been localized to the pituitary. If the ratio is <1.5, the patient has ectopic (or adrenal) Cushing's syndrome.  Localization of the ectopic ACTH tumour, which is often small, requires CT scanning or other imaging techniques in other organs. When head scans are negative in a patient with pituitary Cushing's syndrome, it also is possible to use bilateral inferior petrosal sinus catheterization to predict the side of the tumour's location; if the side-to-side ACTH ratio is > 1.4, the tumour is more likely to be on the side with the higher level.

Hypercortisolemia also can be caused by adrenal lesions such as independent adenomas and adrenal carcinoma. These lesions are independent of ACTH, and thus hypercortisolemia is resistant to the administration of dexamethasone (no suppression) or of CRH (no elevation). Also, resting ACTH levels in these diseases are usually low or undetectable.

The early diagnosis of Cushing's disease can be difficult, but is being done with increasing frequency today because of heightened awareness on the part of physicians. Early cases can lack the classic stigmata, and hypercortisolemia may be marginal. Classic features often help to identify true Cushing's disease and differentiate the various causes of the syndrome. Cushing's disease is commonest in young to middle-aged women and has a slow progression. In early cases, abnormal menstrual cycles, abnormal striae, spontaneous bruising, androgen excess and hypertension may be absent. The classic ectopic ACTH syndrome, on the other hand, is found typically in older men who smoke to excess. Recent weight loss, anaemia, low serum potassium levels, and severe hypertension are prominent. Adrenal adenomas are relatively slowly progressive; cortisol excess is mild, and very often androgen features are minimal. Adrenal carcinoma has a rapid onset with very high levels of adrenocorticoid hormones. Adrenal carcinomas and adenomas that produce cortisol cause atrophy of the contralateral adrenal. Consequently, asymmetry of the adrenals should warn against the diagnosis of a pituitary adenoma.

It is well to remember that the accuracy of endocrine testing for Cushing's disease is not absolute, but is in the range of 90 percent. Consequently, in the absence of confirmation by MRI or another imaging modality, errors will be made, and this must be stressed to the patient and family in the context of discussions surrounding operation.

  Postoperative Evaluation

Early postoperative evaluation is limited to measuring serum cortisol levels and 24-h urine free cortisol values while the patient is receiving dexamethasone. All patients require replacement therapy of adrenal cortisol hormones because the normal corticotropes have been suppressed by the chronically elevated cortisol levels. Following the removal of the adenoma, it is necessary to use dexamethasone in the early postoperative period to confirm the adequacy of the operative procedure. Dexamethasone does not cross-react significantly with the serum cortisol testing procedure. Several days after operation, when the patient is on maintenance or slightly higher doses of dexamethasone, the plasma cortisol level and 24-h urine free cortisol value should be determined, If the plasma cortisol level is less than 5 µg/dl and the urine value is low, one can have some confidence that the procedure was successful and that satisfactory removal of the tumour has been accomplished. True remission, as measured by endocrine testing as opposed to clinical examination, demands the following: (1) normal or low serum cortisol levels; (2) return of the diurnal rhythm; (3) normal 24-h urinary output of free cortisol and 17-hydroxysteroids; (4) normal dexamethasone suppressibility.

When these criteria are met, especially if clinical remission has occurred as well, one's confidence of total removal can be high. As time goes on, the need for replacement therapy lessens as normal endogenous cortisol secretion returns (within 6 to 12 months in most cases). At this time, formal testing should detect normal cortisol dynamics and basal levels.

TSH-Secreting Pituitary Adenomas

Pituitary adenomas that secrete thyroid-stimulating hormone are unusual and account for less than 1 percent of all adenomas. Diagnosis requires judicious interpretation of serum TSH levels. These can range from 1.6 to 480 µU /ml in the presence of clinical hyperthyroidism. In approximately 30 percent of these cases, TSH levels are lower than 10 µU/ml.

In addition, TSH-secreting adenomas commonly release excessive quantities of the alpha subunit of this glycoprotein hormone. In terms of molarity, the alpha subunit exceeds the TSH beta sub­unit, the other half of the molecule. The ratio of the alpha subunit to TSH is greater than 1.

In normal individuals, TRH stimulation causes a rise in TSH. In patients with TSH-secreting pituitary adenomas, the response to TRH is negligible, with few exceptions. This blunting of the response is not seen exclusively in TSH-secreting adenomas, but helps in making the diagnosis. Thyroid hormone generally suppresses TSH secretion in normal persons, However, the administration of thyroid hormone may fail to suppress TSH and the free alpha subunit when the tumour is a TSH-secreting adenoma.

Gonadotropin-Secreting Pituitary Adenomas

The gonadotropin-secreting pituitary adenomas usually produce FSH and/or LH. One recent problem in diagnosis has been the sensitivity of the radioimmunoassay for FSH and LH, Other problems involve the high serum concentration of gonadotropins normally observed in menopause. Since these tumors have few systemic effects, they rarely are diagnosed until headache and loss of vision have been reported. Three types of tumour have been recorded: LH-secreting, FSH-secreting, and FSH- and LH-secreting, Many difficulties arise in making the diagnosis, as other factors such as testicular failure may be the cause of elevated levels of FSH or LH. Thus, serum testosterone concentrations must be obtained to verify the diagnosis.

Alpha Subunit-Secreting Pituitary Adenomas

The glycoprotein hormones are composed of alpha and beta subunits. The alpha subunit is the same for all of these hormones. The beta subunit is unique to each hormone and confers biological and immunologic specificity. Some pituitary adenomas have been associated solely with the production of the alpha subunit. This entity must be considered, especially in patients with large, apparently non secreting tumors. The serum alpha-subunit level can be measured directly by radioimmunoassay. For further confirmation, immunocytochemical staining of the tissue should be carried out to demonstrate the presence of the alpha subunit alone in the secretory granules. The importance of this test is that serum alpha­subunit levels may then be useful as a tumour marker to be followed postoperatively.