Jump to content

Screening and Diagnosis of Cushing?s Syndrome


Recommended Posts

I found the following article very illuminating. I still see some problems with it, especially with the tone taken with late-night serum cortisols and salivary cortisols. I believe the concensus with more recent research is showing that salivary cortisols are good only after the serum level reaches a certain level (pretty high), and serum cortisols are good predictors.

 

Screening and Diagnosis of Cushing's Syndrome

James W. Findling, MDa,b, Hershel Raff, PhDb,c,*

aEndocrine?Diabetes Center, St. Luke's Medical Center, 2801 West KK River Parkway,

Suite 245, Milwaukee, WI 53215, USA

bMedical College of Wisconsin, Milwaukee, WI 53226, USA

cEndocrine Research Laboratory, St. Luke's Medical Center, 2801 West KK River Parkway,

Suite 245, Milwaukee, WI 53215, USA

The diagnosis of Cushing's syndrome is the most challenging problem in

clinical endocrinology. Clinical manifestations of excessive glucocorticoid

exposure (either endogenous or exogenous) are protean and may be quite

subtle. The observation that patients who have incidentally discovered

adrenocortical adenomas and subclinical hypercortisolism frequently have

improvement in their diabetes, hypertension, and obesity after adrenalectomy,

increases the importance of establishing the diagnosis of even mild

Cushing's syndrome [1,2]. Because the clinical syndrome is not always

obvious, a low index of suspicion is needed for screening and certainly

mandated in high-risk patient populations.

Epidemiology

Most epidemiologic studies have suggested that spontaneous Cushing's

syndrome is an unusual disorder. A population-based study in Denmark

found that the diagnosis of endogenous hypercortisolism had been

established in 166 patients in an 11-year period [3]. Ninety-nine patients

had Cushing's disease; 48 had adrenal-dependent Cushing's disease; 16 had

ectopic corticotropin (ACTH), and three patients were unclassified. This

represented an incidence of 2 cases per 1 million inhabitants per year. The

patients in this study who were not cured by surgical intervention or who

had malignant disease had a poor prognosis with a standard mortality ratio

of 3.8 to 5.0 compared with normal controls. This study was in agreement

with a similar study conducted in Vizcya, Spain, in 1994 [4]. In a Spanish

population of 1.15 million, 49 patients were diagnosed with Cushing's

disease in an 18-year period, yielding an incidence of 2.5 cases per 1 million

inhabitants per year. The prevalence in this study population was

documented at 39 cases per 1 million inhabitants at the time the study

was conducted. This report also emphasized the poor prognosis in these

patients, with a standard mortality ratio of 3.8. Although these data would

suggest that spontaneous Cushing's syndrome is rare, other observational

data and screening in high-risk populations suggest that the diagnosis

frequently is overlooked. For example, the authors have evaluated 85

patients from the Milwaukee metropolitan area (population 1.5 million)

over an 11-year period. Even if this represented all the patients in the area

with Cushing's syndrome over that period of time (and the authors estimate

that it probably represents 30% to 50%), this would represent an incidence

of approximately 5 patients per million inhabitants per year.

Screening studies performed in high-risk populations recently have

suggested an unexpectedly high incidence of occult Cushing's syndrome. In

1996, Leibowitz et al performed screening studies in 90 obese subjects who

had poorly controlled diabetes mellitus (hemoglobin A1C[9%) and found

three patients (3.3%) who had surgically confirmed Cushing's syndrome [5].

More recently, Catargi et al [6] carefully studied 200 obese patients who had

type 2 diabetes mellitus and hemoglobin A1C values greater than 8%. Four

patients (2%) had definite Cushing's syndrome and seven had subclinical

hypercortisolism with unilateral adrenal adenomas demonstrating uptake

on iodocholesterol scintigraphy and elevated late-night cortisol levels.

Another study performed in Turkey screened 100 consecutive obese subjects

(body mass index[25 kg/m2) and found 9 subjects who had surgically

proven Cushing's syndrome (5 pituitary and 4 adrenal) [7]. Moreover,

subclinical hypercortisolism has been shown in at least 10% of patients who

have adrenal incidentalomas [8], a finding that is seen in approximately 2%

of the adult population. These studies suggest that spontaneous Cushing's

syndrome is more common than previously appreciated.

Who should be screened?

Endogenous hypercortisolism may occur at any age and usually has an

insidious onset, with a usual duration of illness before clinical diagnosis of 3

to 5 years. The disorder appears to be more common in women, particularly

in patients who have pituitary- and adrenal-dependent Cushing's syndrome;

however, the ratio may be fairly equal in patients who have ectopic ACTHdependent

Cushing's syndrome [8]. The authors believe that screening tests

should be performed in subjects who have relatively specific signs and

symptoms of hypercortisolism or in patients who have clinical diagnoses

that may be caused by endogenous cortisol excess.

Specific signs and symptoms

Signs and symptoms that should provoke a biochemical evaluation for

possible Cushing's syndrome are shown in Box 1. Weight gain with

redistribution of fat centrally affecting the face, neck, trunk, and abdomen is

one of the most common clinical findings. Unfortunately, this type of weight

gain is very common and may be indistinguishable from those patients who

have the metabolic syndrome. The weight gain is often insidious, and

frequent review of old photographs may help the clinician better appreciate

the physical changes that may have occurred in patients who have weight

gain. The physical changes that may occur in a patient over a period of 12

years are illustrated in Fig. 1. Although patients who have Cushing's

syndrome may have the classic moon facies, the facial rounding can be quite

subtle. The patient in Fig. 1 was examined by many clinicians (including

several endocrinologists) before the diagnosis of spontaneous Cushing's

syndrome was considered. The presence of significant supraclavicular

Box 1. Who should be screened for Cushing's syndrome?

Signs and symptoms

Central obesity with:

Facial rounding with plethora

Increased supraclavicular and dorsocervical fat

Cutaneous wasting with ecchymoses

Wide violaceous striae (greater than 1 cm)

Proximal myopathy

Increased lanugo hair

Superficial fungal infections

Growth retardation (in children)

Clinical diagnosis

Metabolic syndrome X

Diabetes mellitus (Hgb A1C > 8%)

Hypertension

Hyperlipidemia

Polycystic ovary syndrome (PCOS)

Hypogonadotropic hypogonadism

Oligomenorrhea/amenorrhea/infertility

Decreased libido and impotence

Osteoporosis (especially rib fracture)

Patients aged

Incidental adrenal mass

fullness and dorsocervical fat accumulation should generate a screening test

for hypercortisolism.

The catabolic effects of glucocorticoid excess frequently lead to cutaneous

wasting from atrophy of the epidermis and underlying connective tissue.

These changes result in the thin appearance of the skin with the typical

plethoric facial appearance and easy bruisability. The skin is fragile and,

when removing adhesive tape, may peel off like damp tissue paper. Most

women who have Cushing's syndrome have skin fold thickness less than 2

mm in the dorsum of the hand, compared with greater than 2 mm with PCOS

[9]. The significant weight gain and the skin changes often result in violaceous

depressed and wide striae (usually[1 cm in diameter). These striae usually

occur on the abdomen but also may occur over the breasts, hips, buttocks,

thighs, and axilla. Striae usually are observed in patients who have significant

hypercortisolism, and rarely in patients over the age of 40 [10]. Minor

wounds or abrasions may heal poorly. In addition, because of the

immunosuppressive effects of hypercortisolism, superficial mucocutaneous

fungal infections such as tinea versicolor also may be seen [11]. Although

androgen excess may be present and result in facial hirsutism, vellus

hypertrichosis (lanugo hair) that is glucocorticoid-dependent is probably

more common in women who have Cushing's syndrome. [Robin inserts here:

I had that and had no clue it was a symptom for a long time!] In addition,

papular acne may occur in younger patients who have Cushing's syndrome.

The catabolic effects of hypercortisolism also may result in type II muscle

fiber atrophy with significant weakness in the proximal musculature [12].

Patients frequently complain of difficulty climbing stairs or rising from

a chair. The myopathy may be particularly problematic in older adults and

may be misdiagnosed as amyotrophic lateral sclerosis or multiple sclerosis.

Body composition studies have demonstrated reduced body cell mass,

indicating a true protein loss in these patients [13].

Glucocorticoid excess also is known to blunt somatic growth. The

impairment of somatic growth is mostly caused by direct action of

glucocorticoids in the growing long bones in children, arresting the

development of epiphysial cartilage [14,15]. Hypercortisolism also blunts

growth hormone secretion, but insulin-like growth factor-1 levels are usually

normal [15]. Growth retardation associated with progressive, and frequently

generalized, obesity is the hallmark of Cushing's syndrome in children. After

correction of hypercortisolism, children who have retarded linear growth

rate should be treated with growth hormone, because there is a limited

window of opportunity to promote an increase in growth to obtain normal

adult height [16]. Impaired growth hormone responses to provocative

stimuli also are observed in adult patients who have Cushing's syndrome

and may persist for up to 2 years after treatment [17]. The authors are

unaware of studies evaluating the efficacy of growth hormone replacement

therapy in these patients.

Nonspecific signs and symptoms

Cushing's syndrome is associated with a range of psychologic and

cognitive problems (see article by Bourdeau et al elsewhere in this issue).

Most patients who have Cushing's syndrome meet criteria for depression,

and a few patients have other neuropsychiatric problems including mania,

anxiety, and cognitive dysfunction [18]. Psychosis may even occur, and

suicidal tendency often has been reported in patients who have endogenous

hypercortisolism. Children who have Cushing's syndrome often exhibit

obsessive?compulsive behavior and do extremely well in their schoolwork

[19]. Cognitive dysfunction with reversible loss of brain volume also has

been reported [20].

Less common and unappreciated clinical features of Cushing's syndrome

may involve the eyes. In Cushing's original report, 4 of 12 patients were

described with exophthalmos [21]. A recent study demonstrated that 45% of

patients who have active Cushing's syndrome had exophthalmos ([ 16 mm)

compared with 20% of treated Cushing's syndrome patients and 2% of

normal controls [22]. The proptosis in patients who have Cushing's

syndrome is asymptomatic and presumably caused by retro-orbital fat

accumulation. Two other unique eye findings have been reported in patients

who have Cushing's syndrome. Lisch nodules, thought to be specific for

neurofibromatosis, but rare in the general population, are melanocytic

hamartomas of the iris. These yellow or brown, dome-shaped elevations

projecting from the surface of the iris recently were observed in 2 of 14

consecutive patients evaluated for endogenous hypercortisolism [23].

[i have these, too!! Never saw this before. In fact, an eye doctor commented [/color]

on this when I was a child.] It was

speculated that the underlying mechanism leading to the overgrowth of the

melanocytes in the iris may be similar to the corticotroph adenomatous

change in the pituitary. The same group also reported the presence of central

serous chorioretinopathy in 3 of 60 consecutive patients who have Cushing's

syndrome [24]. This unusual condition represents the accumulation of

subretinal fluid at the posterior pole of the fundus, causing an area of retinal

detachment and some decrease in visual acuity. These findings also have

been reported in patients receiving high-dose glucocorticoid therapy.

Clinical diagnoses associated with hypercortisolism

There are a few clinical disorders that alone should stimulate a consideration

for the presence of Cushing's syndrome (see Box 1). Cushing's

syndrome results in the entire clinical spectrum of the metabolic syndrome

including obesity, diabetes, hypertension, and gonadal dysfunction. This

phenotype has a high prevalence, and, with the increasing recognition of

mild or subclinical hypercortisolism, it is clear that these two syndromes are

clinically indistinguishable. The higher mortality rate observed in patients

who have Cushing's syndrome seems to be related to the cardiovascular

complications associated with the metabolic syndrome [4] as discussed in the

article by Pivonello et al elsewhere in this issue.

The insulin resistance that accompanies hypercortisolism results in

a decrease in glucose use by peripheral tissues; impaired glucose tolerance

may occur in 30% to 60% of patients, and frank diabetes may occur in 25%

to 50% of patients [25]. As previously mentioned, 2% to 3% of patients who

have poorly controlled type 2 diabetes may have unrecognized Cushing's

syndrome [5,6]. Glucocorticoid excess also plays an essential role in the

accumulation of abdominal body fat. Thus, obesity is an independent risk

factor for reduced life expectancy and correlates well with all of the

metabolic sequelae and atherosclerosis associated with the metabolic

syndrome. The importance of cortisol in the generation of abdominal

obesity recently has been highlighted by targeted overexpression of 11bhydroxysteroid

dehydrogenase type 1 in transgenic mice [26]. This resulted

in increased visceral fat, presumably because of local production of active

glucocorticoid within the adipocyte.

Arterial hypertension occurs in at least 80% of patients who have

Cushing's syndrome and is a major contributing factor to cardiovascular

morbidity. Although the hypertension may be mild, organ damage such as

cardiac hypertrophy with left ventricular concentric remodeling may occur

as a result [27]. The mechanism for the hypertension presumably is related to

the mineralocorticoid activity of cortisol. In the presence of severe hypercortisolism,

there is a failure to completely metabolize cortisol to the

inactive cortisone by the renal enzyme 11b-hydroxysteroid dehydrogenase

type 2. This allows cortisol binding to the mineralocorticoid receptor in the

distal nephron, resulting in hypertension and hypokalemia. Other possible

mechanisms of hypertension in patients who have Cushing's syndrome

include enhancement of the inotropic pressor effects of vasoactive

substances including catecholamines, vasopressin, and angiotensin II, and

possible suppression of vasodilatory mechanisms including nitric oxide and

prostacyclin production [28].

The hyperlipidemia seen in patients who have Cushing's syndrome

appears to be similar to the dyslipidemia associated with the metabolic

syndrome. There is an increase in very low density lipoprotein (LDL) and

LDL with a decrease in high density lipoprotein (HDL) levels. This results

in elevation of total triglyceride and cholesterol levels [29,30].

Less appreciated is the hypercoagulability in patients who have Cushing's

syndrome leading to an increased risk for thromboembolic events mostly

after surgery or during inferior petrosal sinus sampling [31]. Hypercortisolism

stimulates the synthesis of several clotting factors such as

fibrinogen by the liver and von Willebrand factor by endothelial cells. In

addition, glucocorticoids also increase the synthesis of plasminogen activator

inhibitor type 1, the main inhibitor of the fibrinolytic system [32]. The general

consensus is that patients who have Cushing's syndrome should be given

heparin during inferior petrosal sinus sampling and that anticoagulation

should be considered in the postoperative period in many patients.

Because women who have Cushing's syndrome may present with

menstrual irregularities or signs and symptoms of androgen excess, the

diagnosis of PCOS often is considered. Kaltsas et al recently showed that

most women who have Cushing's syndrome also have PCOS and suggested

that women who have PCOS should undergo screening studies for

hypercortisolism [33]. Interestingly, the menstrual irregularity in these

women appears to be more closely related to the degree of cortisol excess

than the actual circulating androgen concentrations [34]. Male patients

frequently present with diminished libido or impotence associated with

subnormal testosterone concentration. Hypogonadotropic hypogonadism is

not an uncommon finding in men who have hypercortisolism and is

underappreciated. Therefore, Cushing's syndrome should be considered in

men who have hypogonadotropic hypogonadism, particularly if they have

other phenotypic characteristics of hypercortisolism.

The diagnosis of Cushing's syndrome should be considered in patients

who have unexplained osteoporosis, particularly in those younger than 65

years. Glucocorticoids influence bone and calcium homeostasis at many

levels (see article by Shaker and Lukert elsewhere in this issue). Pathologic

fractures may be the presenting feature of Cushing's syndrome, and rib

fractures seem to be especially common. There is also an increased risk of

Link to comment
Share on other sites

Rest of article below:

There is also an increased risk of

nephrolithiasis in patients who have Cushing's syndrome. A recent study

found kidney stones in 50% of patients who have active Cushing's

syndrome, 27% of patients in remission, and 6.5% of controls [35]. Patients

who have active Cushing's syndrome and nephrolithiasis had a significantly

increased prevalence of arterial hypertension, hypercalciuria, hypocitraturia,

and hyperuricosuria.

In light of the evidence that 6% to 10% of patients who have incidentally

discovered adrenal masses (of at least 2 cm) have subclinical hypercortisolism,

it seems obvious that screening studies for Cushing's syndrome

are needed in this fairly common group of patients [36]. Information

concerning evaluation and treatment of these patients is found in the article

by Terzolo et al elsewhere in this issue.

Nonspecific routine laboratory abnormalities

There are no routine laboratory abnormalities that are specific for

Cushing's syndrome. Thyrotropin (TSH) levels are lower in patients who

have Cushing's syndrome and increase after remission [37]. Rarely, patients

who have Cushing's syndrome will be referred to an endocrinologist because

of a subnormal TSH (usually between 0.1 and 0.4). It is also well

appreciated that successful cure of Cushing's syndrome may unmask a preexisting

autoimmune thyroid disorder with the appearance of hypothyroidism

or hyperthyroidism [38]. Other more routine laboratory studies are

rarely helpful. High normal values of hemoglobin, hematocrit, and red

blood cell count may be seen because of androgen excess, but elevations into

the polycythemic range are very rare. Leukocytosis, usually with depressed

percentages of lymphocyte and eosinophils, may be seen occasionally.

Electrolyte disturbances are quite uncommon with the exception of hypokalemia

in patients who have prodigious hypercortisolism. Routine radiographs

may be helpful in showing low bone density in the axial skeleton.

Rarely, chest roentgenograms may show widening of the mediastinum

because of mediastinal lipomatosis.

Screening tests

The choice of tests in the initial evaluation of a patient suspected of

endogenous hypercortisolism can be difficult. Even more problematic is the

interpretation of the results of these tests, particularly if they are not in

agreement with each other. This is particularly so in mild Cushing's

syndrome; if the symptoms are subtle, the biochemical abnormalities are

likely to be subtle as well. The authors recently published several extensive

reviews of biochemical tests useful in the diagnosis of Cushing's syndrome

[8,39?42]. Therefore, this section will focus on newer studies and conclude

with an exposition on studies recently published that compare tests by

modern statistical analyses. [i'll look for these articles and post them]

Urine free cortisol

The measurement of free cortisol in a 24-hour urine collection has been

the mainstay of the diagnosis of endogenous hypercortisolism [43]. The

concept is that if the daily production of cortisol is increased, the free

cortisol filtered and not reabsorbed or metabolized in the kidney will be

increased measurably. Mild Cushing's syndrome often results from small,

but significant increases in nighttime cortisol secretion (Fig. 2). Because

most of the cortisol secreted during any 24-hour period is usually between

4 AM and 4 PM, subtle increases in nighttime cortisol secretion may not be

detected in standard 24-hour urine free cortisol measurements. It also

requires an adequate urine collection that must be verified with a measurement

of urinary creatinine.

The other problem with this test in the past was the different methods

used to measure cortisol in urine. There are still laboratories that use direct

immunoassay of cortisol without chromatographic separation. Although

this assay lacks specificity, it may be useful if the patient is excreting an

unusual pattern of cortisol metabolites that cross-react with the cortisol

antibody. Most, however, consider the gold standard for urine free cortisol

to involve some type of chromatographic separation, usually highperformance

liquid chromatography (HPLC) [44?47]. Assuming the

chromatography is done properly, the next issue is the method of detection.

Immunoassay can be done after HPLC, although tandem mass spectrometry

provides more specific results [45]. There are substances that will

interfere with some of these chromatographic methods including carbamazepine

[46] and fenofibrate [47].

Several recent studies found the sensitivity of urine-free cortisol (UFC) to

range from 45% to 71% at 100% specificity [48?52]. That is, it is not

uncommon for patients who have mild Cushing's syndrome to have at least

one or more normal urine free cortisol measurements. Moreover, UFC may

be increased in patients who have so-called ''pseudo-Cushing's syndrome,''

Fig. 2. Theoretical salivary cortisol levels during a typical day in a normal subject, a patient

who has mild Cushing's disease, and a patient who has severe Cushing's syndrome. Note that

the 11 PM?3 AM time period is the best to discriminate between mild Cushing's disease and

normal subjects. For diagnostic purposes, salivary cortisol is usually sampled at 11 PM.

(Reprinted from Raff H. Role of salivary cortisol determinations in the diagnosis of Cushing's

syndrome. Curr Opin Endocrinol Diabetes 2004;11:271?5; with permission.)

which includes alcoholism, endogenous depression, and eating disorders

[51,52]. When UFC is compared with other available screening tests using

objective statistical analyses, its sensitivity and specificity are less than ideal

(Table 1).

Circadian rhythm studies

One of the first biochemical disturbances that patients develop with mild

Cushing's syndrome is a failure to decrease cortisol secretion to its normal

nadir at night (see Fig. 2). This phenomenon has been exploited in the

diagnosis of Cushing's syndrome with several different approaches.

Measurement of an elevated serum cortisol at midnight has a very high

sensitivity and specificity for Cushing's syndrome of any etiology [51,53].

The main problem with this approach is the logistical problem of obtaining

an unstressed blood sample in a routine clinical setting at midnight. This has

made the widespread application of this test impractical.

A solution to this sampling problem is the measurement of salivary

cortisol at bedtime as a surrogate for serum cortisol [40,41,50,52,54?57]. The

authors' initial study with this test demonstrated a sensitivity and specificity

of approximately 95% [54]. There now have been many major studies

validating this approach to screen for Cushing's syndrome [49,50,52,54?57],

and this method will continue to increase in popularity as salivary cortisol

assays become readily available [58,59]. A potential drawback of this test,

however, is its specificity. There are many factors that can elevate cortisol

secretion falsely at bedtime including proximal stress, sleep disturbances,

psychoneuroendocrine factors, and contamination of the saliva sample [40].

Therefore, whereas 24-hour UFC lacks reliable sensitivity, salivary cortisol

under certain circumstances may be too sensitive.

Another approach has been to perform urine collections for the

measurement of free cortisol just from the overnight period [60]. The

concept is that if one collects urine during and just after the circadian nadir,

the sensitivity of the test might be improved. It also would require only one

or two urine collections. This approach requires a very accurate

Table 1

Diagnostic characteristics of different biochemical tests at 100% sensitivity or 100% specificity

Diagnostic

characteristic

LDDST: cortisol

24-h UFC [48]

Nighttime: cortisol

Serum [48] Salivary [49] Serum [48,51] Salivary [50]

Sensitivity at

100% specificity

54% NC 71% 75%, 96% 93%

Specificity at

100% sensitivity

41% 93% 73% 77% 96%

Abbreviation: NC, not calculated (sensitivity was 100% at 93% specificity).

Data from Refs. [48?51].

394 FINDLING & RAFF

measurement of urinary creatinine to which the free cortisol measurement is

normalized.

Suppression tests

The authors have reviewed the physiologic basis of using the sensitivity to

glucocorticoid negative feedback to diagnose endogenous hypercortisolism

[39]. Briefly, the theory is that a small enough dose of dexamethasone will

not inhibit ACTH release from corticotroph adenomas or occult ectopic

ACTH-secreting tumors, while suppressing ACTH from normal pituitary

tissue. (Obviously, ACTH-independent [adrenal] Cushing's syndrome

should be unaffected by dexamethasone administration). It is now clear

that neither the overnight nor the 2-day, low-dose dexamethasone suppression

tests (LDDST) are of sufficient reliability to be used to rule out

Cushing's syndrome [61?63]. The authors recently demonstrated that 18%

of patients who have proven Cushing's disease suppressed serum cortisol to

the standard cut-off of 5 lg/dL (135 nmol/L), while 8% showed suppression

to less than 2 lg/dL (\ 54 nmol/L) [62]. The performance of the 2-day

LDDST was even worse. Therefore, there was no cut-off that identified all

patients who have Cushing's syndrome.

Most reference laboratories with high volumes use direct serum cortisol

assays using platforms that employ chemiluminescent or electrochemiluminescent

immunoassays. It recently has been shown that the performance of

different assays introduces significant intermethod variability when performing

the LDDST to detect patients who have mild or preclinical Cushing's

syndrome [64]. Because a reliable LDDST requires accuracy at low serum

cortisol concentrations, variability between reference laboratories is likely to

introduce even more uncertainty in the usefulness of this test. This

variability would not be included in the lack of sensitivity and specificity

identified in a well-controlled study in which the same assay methodology is

used.

Despite these limitations, the overnight LDDST remains widely employed.

A recent consensus statement recommended that patients who have

plasma cortisol greater than 1.8 lg/dL (50 nmol/L) after overnight 1 mg

dexamethasone administration merit further evaluation [65]. It is predicted

that this more stringent criterion will yield a diagnostic sensitivity of 95% to

98% [62], but that specificity (ie, false-positives) will suffer as a result.

Stimulation tests

The secretion of ACTH from the pituitary, in addition to being under

negative glucocorticoid feedback control, is stimulated primarily by

hypothalamic corticotropin-releasing hormone (CRH) and, to a lesser

extent, arginine vasopressin (AVP). The CRH test has been used to attempt

to identify patients who have mild ACTH-dependent Cushing's syndrome

[66]. The theory is that corticotroph adenomas will display exaggerated

ACTH responses compared with normal subjects as long as they continue to

express receptors for CRH. Furthermore, a V2 analog of vasopressin,

DDAVP (desmopressin), also has been used to attempt to identify patients

who have mild Cushing's disease [58,59]. The theory is that corticotroph

adenomas will have an exaggerated response to vasopressin compared with

normal subjects. Neither test appears to possess the adequate sensitivity or

specificity to merit their expense [66?68].

Combined testing

Different diagnostic tests can be performed on separate occasions to

attempt to improve the overall reliability compared with each test alone [39].

Because each of the standard tests for the diagnosis of Cushing's syndrome

has merit but also some weaknesses, another approach to improve overall

performance could be to perform two tests simultaneously in each patient.

Recent examples of this are the combination LDDST?CRH test and the

combination of late-night salivary cortisol and LDDST.

The logic of the LDDST?CRH is that one can improve the discriminatory

potential of each test for mild Cushing's disease by performing them

simultaneously. The theory is that only abnormal corticotrophs will respond

to CRH while suppressed with dexamethasone [69,70]. In this test, starting

at noon, dexamethasone (0.5 mg) is administered every 6 hours for a total of

eight doses, with the last given at 6 AM before dynamic studies. CRH (1 lg/

kg) is then given intravenously at 8 AM with measurement of cortisol and

ACTH every 15 minutes for 1 hour. A serum cortisol greater than 1.4 lg/dL

([ 38.6 nmol/L) is considered abnormal. There are no currently accepted

criteria for an abnormal plasma ACTH response to LDDST?CRH.

The initial studies with the LDDST?CRH test were promising but

required great attention to detail and compliance with the rather

sophisticated protocol. Its main strength was its ability to distinguish mild

Cushing's syndrome from the so-called pseudo-Cushing's syndrome. The

criteria for pseudo-Cushing's syndrome included a failure for symptoms to

progress for 17 months. It would be interesting to see if any of those patients

from 1993 [69] subsequently have been discovered to have Cushing's disease.

A recent study in patients who have anorexia nervosa found about half with

abnormal LDDST?CRH tests, raising a concern that it does not reliably

discriminate Cushing's syndrome from every form of endogenous pathophysiologic

activation of the hypothalamic-pituitary-adrenal axis [71].

The LDDST?CRH test requires measurement of serum cortisol, thereby

using endogenous adrenocortical amplification of an ACTH signal for

success [69?71]. To accomplish this, the test requires a very sensitive serum

cortisol measurement that is not the standard reference laboratory assay. It

also, again, raises the possibility of variable reliability between reference

laboratories [64]. The LDDST?CRH test is very expensive in terms of

material and labor, and it usually is reserved only for patients who have

mild Cushing's syndrome or to confirm the diagnosis when other screening

tests are equivocal [39].

Late-night salivary cortisol?low-dose dexamethasone suppression

When the authors first published results with late-night salivary cortisol

to screen for Cushing's, they hypothesized that the high sensitivity of the test

combined with better specificity of the LDDST might be exploited by

a combined test [54]. The theory is that almost all patients who have

Cushing's syndrome have elevated late-night salivary (or serum) cortisol,

but a fair number of patients who have pseudo-Cushing's syndrome do also

(ie, low specificity). The LDDST, however, while having low sensitivity, may

have a better specificity. Castro et al recently evaluated this concept by

showing an increase in specificity from 88% with late-night salivary cortisol

alone to 100% using the combined test but only when using nonobese

subjects as the control group [49]. The problem was that the specificity for

Cushing's syndrome compared with obese patients who presumably did not

have it was not increased by the combined test. In a follow-up study, it was

suggested that perhaps a higher dose of dexamethasone might improve the

specificity of the test [72]. Most importantly, they clearly showed that

measurement of salivary cortisol after an overnight LDDST was significantly

better than measuring serum cortisol, presumably because salivary

cortisol is a much better estimate of free, biologically active cortisol [40].

Comparison of diagnostic characteristics

Because each diagnostic test for Cushing's syndrome has liabilities, it is

helpful to perform an objective comparison using well-defined diagnostic

criteria. Table 1 focuses on several recent studies that meet the evidencebased

criteria of comparing several tests within one group of patients, and

performing careful step-wise analysis of sensitivity versus specificity. This

table shows the sensitivity at a cut-off that provides 100% specificity, and

the specificity at a cut-off that provides 100% sensitivity. Obviously, there is

always a tradeoff between the two. Generally, high sensitivity is preferred

for screening tests, but combinations of tests may improve both criteria.

The first characteristic that is clear from this table is that the LDDST

using the measurement of serum cortisol had the poorest performance of all

of the tests [48]. Again, either sensitivity or specificity can be improved at the

expense of the other by adjusting cut-off levels. Of all the tests, nighttime

salivary cortisol has the highest sensitivity and specificity [50,55]. It is

important to point out that the authors purposefully did not use their own

data in this analysis; these were independent evaluations of these tests with

no apparent bias. It is also important to note that UFC performed better

than the LDDST but still not as well as night-time salivary cortisol.

Probably the most important information for clinical endocrinologists is

the practical nature of measuring salivary cortisol at 11 PM
Link to comment
Share on other sites

Note that

the 11 PM?3 AM time period is the best to discriminate between mild Cushing?s disease and

normal subjects. For diagnostic purposes, salivary cortisol is usually sampled at 11 PM.

(Reprinted from Raff H. Role of salivary cortisol determinations in the diagnosis of Cushing?s

syndrome. Curr Opin Endocrinol Diabetes 2004;11:271?5; with permission.)

CS: SCREENING & DIAGNOSIS 393

which includes alcoholism, endogenous depression, and eating disorders

[51,52]. When UFC is compared with other available screening tests using

objective statistical analyses, its sensitivity and specificity are less than ideal

(Table 1).

Circadian rhythm studies

One of the first biochemical disturbances that patients develop with mild

Cushing?s syndrome is a failure to decrease cortisol secretion to its normal

nadir at night (see Fig. 2). This phenomenon has been exploited in the

diagnosis of Cushing?s syndrome with several different approaches.

Measurement of an elevated serum cortisol at midnight has a very high

sensitivity and specificity for Cushing?s syndrome of any etiology [51,53].

The main problem with this approach is the logistical problem of obtaining

an unstressed blood sample in a routine clinical setting at midnight. This has

made the widespread application of this test impractical.

A solution to this sampling problem is the measurement of salivary

cortisol at bedtime as a surrogate for serum cortisol [40,41,50,52,54?57]. The

authors? initial study with this test demonstrated a sensitivity and specificity

of approximately 95% [54]. There now have been many major studies

validating this approach to screen for Cushing?s syndrome [49,50,52,54?57],

and this method will continue to increase in popularity as salivary cortisol

assays become readily available [58,59]. A potential drawback of this test,

however, is its specificity. There are many factors that can elevate cortisol

secretion falsely at bedtime including proximal stress, sleep disturbances,

psychoneuroendocrine factors, and contamination of the saliva sample [40].

Therefore, whereas 24-hour UFC lacks reliable sensitivity, salivary cortisol

under certain circumstances may be too sensitive.

Another approach has been to perform urine collections for the

measurement of free cortisol just from the overnight period [60]. The

concept is that if one collects urine during and just after the circadian nadir,

the sensitivity of the test might be improved. It also would require only one

or two urine collections. This approach requires a very accurate

Table 1

Diagnostic characteristics of different biochemical tests at 100% sensitivity or 100% specificity

Diagnostic

characteristic

LDDST: cortisol

24-h UFC [48]

Nighttime: cortisol

Serum [48] Salivary [49] Serum [48,51] Salivary [50]

Sensitivity at

100% specificity

54% NC 71% 75%, 96% 93%

Specificity at

100% sensitivity

41% 93% 73% 77% 96%

Abbreviation: NC, not calculated (sensitivity was 100% at 93% specificity).

Data from Refs. [48?51].

394 FINDLING & RAFF

measurement of urinary creatinine to which the free cortisol measurement is

normalized.

Suppression tests

The authors have reviewed the physiologic basis of using the sensitivity to

glucocorticoid negative feedback to diagnose endogenous hypercortisolism

[39]. Briefly, the theory is that a small enough dose of dexamethasone will

not inhibit ACTH release from corticotroph adenomas or occult ectopic

ACTH-secreting tumors, while suppressing ACTH from normal pituitary

tissue. (Obviously, ACTH-independent [adrenal] Cushing?s syndrome

should be unaffected by dexamethasone administration). It is now clear

that neither the overnight nor the 2-day, low-dose dexamethasone suppression

tests (LDDST) are of sufficient reliability to be used to rule out

Cushing?s syndrome [61?63]. The authors recently demonstrated that 18%

of patients who have proven Cushing?s disease suppressed serum cortisol to

the standard cut-off of 5 lg/dL (135 nmol/L), while 8% showed suppression

to less than 2 lg/dL (\ 54 nmol/L) [62]. The performance of the 2-day

LDDST was even worse. Therefore, there was no cut-off that identified all

patients who have Cushing?s syndrome.

Most reference laboratories with high volumes use direct serum cortisol

assays using platforms that employ chemiluminescent or electrochemiluminescent

immunoassays. It recently has been shown that the performance of

different assays introduces significant intermethod variability when performing

the LDDST to detect patients who have mild or preclinical Cushing?s

syndrome [64]. Because a reliable LDDST requires accuracy at low serum

cortisol concentrations, variability between reference laboratories is likely to

introduce even more uncertainty in the usefulness of this test. This

variability would not be included in the lack of sensitivity and specificity

identified in a well-controlled study in which the same assay methodology is

used.

Despite these limitations, the overnight LDDST remains widely employed.

A recent consensus statement recommended that patients who have

plasma cortisol greater than 1.8 lg/dL (50 nmol/L) after overnight 1 mg

dexamethasone administration merit further evaluation [65]. It is predicted

that this more stringent criterion will yield a diagnostic sensitivity of 95% to

98% [62], but that specificity (ie, false-positives) will suffer as a result.

Stimulation tests

The secretion of ACTH from the pituitary, in addition to being under

negative glucocorticoid feedback control, is stimulated primarily by

hypothalamic corticotropin-releasing hormone (CRH) and, to a lesser

extent, arginine vasopressin (AVP). The CRH test has been used to attempt

to identify patients who have mild ACTH-dependent Cushing?s syndrome

CS: SCREENING & DIAGNOSIS 395

[66]. The theory is that corticotroph adenomas will display exaggerated

ACTH responses compared with normal subjects as long as they continue to

express receptors for CRH. Furthermore, a V2 analog of vasopressin,

DDAVP (desmopressin), also has been used to attempt to identify patients

who have mild Cushing?s disease [58,59]. The theory is that corticotroph

adenomas will have an exaggerated response to vasopressin compared with

normal subjects. Neither test appears to possess the adequate sensitivity or

specificity to merit their expense [66?68].

Combined testing

Different diagnostic tests can be performed on separate occasions to

attempt to improve the overall reliability compared with each test alone [39].

Because each of the standard tests for the diagnosis of Cushing?s syndrome

has merit but also some weaknesses, another approach to improve overall

performance could be to perform two tests simultaneously in each patient.

Recent examples of this are the combination LDDST?CRH test and the

combination of late-night salivary cortisol and LDDST.

Low-dose dexamethasone suppression?corticotropin-releasing

hormone test

The logic of the LDDST?CRH is that one can improve the discriminatory

potential of each test for mild Cushing?s disease by performing them

simultaneously. The theory is that only abnormal corticotrophs will respond

to CRH while suppressed with dexamethasone [69,70]. In this test, starting

at noon, dexamethasone (0.5 mg) is administered every 6 hours for a total of

eight doses, with the last given at 6 AM before dynamic studies. CRH (1 lg/

kg) is then given intravenously at 8 AM with measurement of cortisol and

ACTH every 15 minutes for 1 hour. A serum cortisol greater than 1.4 lg/dL

([ 38.6 nmol/L) is considered abnormal. There are no currently accepted

criteria for an abnormal plasma ACTH response to LDDST?CRH.

The initial studies with the LDDST?CRH test were promising but

required great attention to detail and compliance with the rather

sophisticated protocol. Its main strength was its ability to distinguish mild

Cushing?s syndrome from the so-called pseudo-Cushing?s syndrome. The

criteria for pseudo-Cushing?s syndrome included a failure for symptoms to

progress for 17 months. It would be interesting to see if any of those patients

from 1993 [69] subsequently have been discovered to have Cushing?s disease.

A recent study in patients who have anorexia nervosa found about half with

abnormal LDDST?CRH tests, raising a concern that it does not reliably

discriminate Cushing?s syndrome from every form of endogenous pathophysiologic

activation of the hypothalamic-pituitary-adrenal axis [71].

The LDDST?CRH test requires measurement of serum cortisol, thereby

using endogenous adrenocortical amplification of an ACTH signal for

success [69?71]. To accomplish this, the test requires a very sensitive serum

396 FINDLING & RAFF

cortisol measurement that is not the standard reference laboratory assay. It

also, again, raises the possibility of variable reliability between reference

laboratories [64]. The LDDST?CRH test is very expensive in terms of

material and labor, and it usually is reserved only for patients who have

mild Cushing?s syndrome or to confirm the diagnosis when other screening

tests are equivocal [39].

Late-night salivary cortisol?low-dose dexamethasone suppression

When the authors first published results with late-night salivary cortisol

to screen for Cushing?s, they hypothesized that the high sensitivity of the test

combined with better specificity of the LDDST might be exploited by

a combined test [54]. The theory is that almost all patients who have

Cushing?s syndrome have elevated late-night salivary (or serum) cortisol,

but a fair number of patients who have pseudo-Cushing?s syndrome do also

(ie, low specificity). The LDDST, however, while having low sensitivity, may

have a better specificity. Castro et al recently evaluated this concept by

showing an increase in specificity from 88% with late-night salivary cortisol

alone to 100% using the combined test but only when using nonobese

subjects as the control group [49]. The problem was that the specificity for

Cushing?s syndrome compared with obese patients who presumably did not

have it was not increased by the combined test. In a follow-up study, it was

suggested that perhaps a higher dose of dexamethasone might improve the

specificity of the test [72]. Most importantly, they clearly showed that

measurement of salivary cortisol after an overnight LDDST was significantly

better than measuring serum cortisol, presumably because salivary

cortisol is a much better estimate of free, biologically active cortisol [40].

Comparison of diagnostic characteristics

Because each diagnostic test for Cushing?s syndrome has liabilities, it is

helpful to perform an objective comparison using well-defined diagnostic

criteria. Table 1 focuses on several recent studies that meet the evidencebased

criteria of comparing several tests within one group of patients, and

performing careful step-wise analysis of sensitivity versus specificity. This

table shows the sensitivity at a cut-off that provides 100% specificity, and

the specificity at a cut-off that provides 100% sensitivity. Obviously, there is

always a tradeoff between the two. Generally, high sensitivity is preferred

for screening tests, but combinations of tests may improve both criteria.

The first characteristic that is clear from this table is that the LDDST

using the measurement of serum cortisol had the poorest performance of all

of the tests [48]. Again, either sensitivity or specificity can be improved at the

expense of the other by adjusting cut-off levels. Of all the tests, nighttime

salivary cortisol has the highest sensitivity and specificity [50,55]. It is

important to point out that the authors purposefully did not use their own

CS: SCREENING & DIAGNOSIS 397

data in this analysis; these were independent evaluations of these tests with

no apparent bias. It is also important to note that UFC performed better

than the LDDST but still not as well as night-time salivary cortisol.

Probably the most important information for clinical endocrinologists is

the practical nature of measuring salivary cortisol at 11 PM. There are now

several assay methods available and reference laboratories that routinely

perform this analysis [58,59]. The patient obtains the saliva at home thereby

minimizing extraneous stress. The approach does not require the

cumbersome collection of complete 24-hour urine samples, nor does it

require taking (and absorbing) dexamethasone at the correct time. Salivary

cortisol can be assessed in small children and the elderly without difficulty

[40]. The authors have found this approach to be extremely useful and

expect that its use will become accepted widely with the increased need for

inexpensive, convenient, and reliable ways to screen the increasingly obese

population for Cushing?s syndrome.

Diagnostic strategy

Fig. 3 shows a strategy for screening patients for Cushing?s syndrome.

Salivary cortisol is obtained at 11 PM, usually on at least two separate nights.

If both results are below the reference range (less than 4.3 nmol/L),

Cushing?s syndrome is unlikely. If both results are above twice the reference

range cut-off (8.6 nmol/L), Cushing?s syndrome is likely, particularly if the

sampling times are verified, and other confounding factors (like contami-

Screening for Cushing?s Syndrome

Repeat 11PM Salivary Cortisol

11PM Salivary Cortisol

Confirm with UFC, LDDST

and/or DST-CRH tests

>4.3 nmol/L

4.3-8.6 nmol/L >8.6 nmol/L

Cushing?s Syndrome Unlikely Cushing?s Syndrome Likely

Fig. 3. A simple paradigm for screening patients for Cushing?s syndrome using late-night

salivary cortisol (cutoff values from Ref. [39]). This approach takes advantage of the excellent

sensitivity of this test.

398 FINDLING & RAFF

nation with steroid hand creams) are excluded. If the results are equivocal or

not consistent between the two samples, two additional salivary cortisol

samples should be obtained. It is also important to verify consistently

abnormal salivary cortisol levels with other tests as indicated to confirm the

diagnosis before entering the differential diagnostic strategy (described by

Lindsay and Nieman elsewhere in this issue).

Summary

The recognition of mild or subclinical Cushing?s syndrome and the

protean nature of its clinical presentation are changing the diagnostic

approach. Recent screening studies in high-risk populations have suggested

that spontaneous Cushing?s syndrome is more common than appreciated,

and its incidence/prevalence has been underestimated. The authors believe

that patients who have specific signs and symptoms or clinical diagnoses

should be considered for screening (see Box 1). Currently, late-night salivary

cortisol measurements provide the best sensitivity with reasonable specificity

to recommend it as the initial screening test. In fact, trying to diagnose mild

Cushing?s syndrome without measuring late-night salivary cortisol may be

like trying to diagnose mild primary hypothyroidism without obtaining

a TSH. Despite their limitations, urine free cortisol and LDDST will

continue to be used to confirm the presence and magnitude of endogenous

hypercortisolism.

References

[1] Terzolo M, Pia A, Ali A, et al. Adrenal incidentaloma: a new cause of the metabolic

syndrome? J Clin Endocrinol Metab 2002;87(3):998?1003.

[2] Reincke M, Nieke J, Krestin GP, et al. Preclinical Cushing?s syndrome in adrenal

incidentalomas: comparison with adrenal Cushing?s syndrome. J Clin Endocrinol Metab

1992;75(3):826?32.

[3] Lindholm J, Juul S, Jorgensen JO, et al. Incidence and late prognosis of Cushing?s syndrome:

a population-based study. J Clin Endocrinol Metab 2001;86(1):117?23.

[4] Etxabe J, Vazquez JA. Morbidity and mortality in Cushing?s disease: an epidemiological

approach. Clin Endocrinol 1994;40(4):479?84.

[5] Leibowitz G, Tsur A, Chayen SD, et al. Preclinical Cushing?s syndrome: an unexpected

frequent cause of poor glycaemic control in obese diabetic patients. Clin Endocrinol 1996;

44(6):717?22.

[6] Catargi B, Rigalleau V, Poussin A, et al. Occult Cushing?s syndrome in type-2 diabetes. J Clin

Endocrinol Metab 2003;88(12):5808?13.

[7] Kadioglu P, Tiryakioglu O, Yirmibescik S, et al. Screening for Cushing?s syndrome in obese

patients: is it really necessary? Presented at the Endocrine Society Meeting. New Orleans,

Louisiana, June 16?19, 2004.

[8] Findling JW, Raff H. Diagnosis and differential diagnosis of Cushing?s syndrome.

Endocrinol Metab Clin North Am 2001;30(3):729?47.

[9] Corenblum B, Kwan T, Gee S, et al. Bedside assessment of skin-fold thickness. A useful

measurement for distinguishing Cushing?s disease from other causes of hirsutism and

oligomenorrhea. Arch Intern Med 1994;154(7):777?81.

CS: SCREENING & DIAGNOSIS 399

[10] Urbanic RC, George JM. Cushing?s disease?18 years? experience. Medicine 1981;60(1):

14?24.

[11] Findling JW, Tyrrell JB, Aron DC, et al. Fungal infections in Cushing?s syndrome. Ann

Intern Med 1981;95(3):392.

[12] Rebuffe-Scrive M, KrotkiewskiM, Elfverson J, et al. Muscle and adipose tissue morphology

and metabolism in Cushing?s syndrome. J Clin Endocrinol Metab 1988;l67(6):1122?8.

[13] Pirlich M, Biering H, Gerl H, et al. Loss of body cell mass in Cushing?s syndrome: effect of

treatment. J Clin Endocrinol Metab 2002;87(3):1078?84.

[14] Giustina A, Wehrenberg WB. The role of glucocorticoids in the regulation of growth

hormone secretion. Trends Endocrinol Metab 1992;3:306?11.

[15] Lebrethon MC, Grossman AB, Afshar F, et al. Linear growth and final height after

treatment for Cushing?s disease in childhood. J Clin Endocrinol Metab 2000;85(9):3262?5.

[16] Magiakou MA, Mastorakos G, Chrousos GP. Final stature in patients with endogenous

Cushing?s syndrome. J Clin Endocrinol Metab 1994;79(4):1082?5.

[17] Feldt-Rasmussen U, Abs R, Bengtsson BA, et al. KIMS International Study Board on

behalf of KIMS Study Group. Growth hormone deficiency and replacement in

hypopituitary patients previously treated for acromegaly or Cushing?s disease. Eur J

Endocrinol 2002;146(1):67?74.

[18] Dorn LD, Burgess ES, Dubbert B, et al. Psychopathology in patients with endogenous

Cushing?s syndrome: atypical or melancholic features. Clin Endocrinol 1995;43(4):433?42.

[19] Magiakou MA, Mastorakos G, Oldfield EH, et al. Cushing?s syndrome in children and

adolescents. Presentation, diagnosis, and therapy. N Engl J Med 1994;331(10):629?36.

[20] Bourdeau I, Bard C, Noel B, et al. Loss of brain volume in endogenous Cushing?s syndrome

and its reversibility after correction of hypercortisolism. J Clin Endocrinol Metab 2002;

87(5):1949?54.

[21] Cushing H. The basophil adenomas of the pituitary body and their clinical manifestations

(pituitary basophilism). Bull Johns Hopkins Hosp 1932;50:137?95.

[22] Kelly W. Exophthalmos in Cushing?s syndrome.
Link to comment
Share on other sites

Here's the link again, Marianne. How is Laurel doing? You know, after reading this article there were clues to my Cushing's as far back as childhood for me, too. That's why I found this article so interesting.

 

XOXO

Robin

 

http://home.comcast.net/~staticnrg/Cushing...DiagnosisCS.pdf

Link to comment
Share on other sites

Thanks, Robin. I just forwarded it on to both of Laurel's doctors. She is not doing very well but recent testing is so confusing. Results continue to contradict others. Laurel is a mess right and still no clear cause. I really hate dealing with this 24/7. I was shocked to realize just how much Laurel health has effected our lives. It has taken over our whole family and shows no signs of letting up. I just have to keep looking for the answer.

 

thanks,

 

Marianne

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
×
  • Create New...