OVERNIGHT DEXAMETHASONE PRE-TREATMENT IMPROVES THE PERFORMANCE OF THE LYSINE-VASOPRESSIN TEST IN THE...

Clinical Endocrinology (1996) 44, 703–710

Overnight dexamethasone pre-treatment improvesthe performance of the lysine-vasopressin testin the diagnosis of Cushing’s syndrome

Patricio Contreras and Vero ´ nica ArayaDepartamento de Medicina, Escuela de Medicina,Universidad de Chile. Santiago, Chile.

(Received 13 January 1995; returned for revision 18 May 1995;finally revised 2 January 1996; accepted 23 February 1996)

Summary

OBJECTIVE There is no endocrine test which is com-pletely reliable for the confirmation of Cushing’s syn-drome and in separation of the various aetiologies. Wehave tested the hypothesis that overnight dexametha-sone pre-treatment should result in a better per-formance of the lysine-vasopressin (LVP) test inthe diagnosis of Cushing’s syndrome.STUDY DESIGN AND PATIENTS We studied 61 subjects,including 25 pituitary-dependent and 9 pituitary inde-pendent Cushing’s (7 adrenal tumours and 2 ectopicACTH syndromes), 18 euadrenal controls, 4 depressedsubjects, and 5 cushingoid patients. The subjectsreceived 1 mg of dexamethasone orally at 2300 h andthe following morning they were given 10 IU of lysine-vasopressin im.MEASUREMENTS Plasma cortisol (RIA) was measuredat times ÿ15, 0, 15, 30, 45, 60, 75, 90 and 120 minutes.RESULTS The dexamethasone-modified LVP (Dx/LVP) test resulted in four patterns of cortisolresponse. The dexa sensitive pattern (positive sup-pression and negative response to LVP) was found ineuadrenal subjects; the dexa insensitive pattern (nega-tive suppression and positive response to LVP) wasseen in Cushing’s disease; a non-responsive pattern(negative suppression and negative response to LVP)was observed only in pituitary independent Cush-ing’s; and an indeterminate pattern (positive suppres-sion and positive response to LVP) was equivocal,being observed in 2 control subjects, 1 patient withCushing’s disease and 1 depressed patient. In separat-ing control subjects from Cushing’s syndromes the Dx/LVP test had 88 .9% sensitivity, 100% specificity and96.2% diagnostic accuracy; when the test was used to

segregate Cushing’s disease from control subjects wefound 96 .0% sensitivity, 100% specificity and 97 .7%diagnostic accuracy. The performance variables forthe Dx/LVP test in separating pituitary dependentfrom pituitary independent Cushing’s were uniformly100%. Depressed and cushingoid subjects did notdiffer from control subjects in their cortisol patternsduring the test. Successful removal of the pituitarymicroadenoma in Cushing’s disease was invariablyfollowed by a reversal of the abnormal cortisol pattern(dexa insensitive) during the test to a dexa sensitivepattern indistinguishable from that of control subjects.CONCLUSION These results confirm our hypothesisand suggest that an improved performance of anycorticotroph stimulus (oCRH, LVP, AVP or desmo-pressin) in the diagnosis of Cushing’s syndromeshould result from pre-treatment with dexamethasone.

Although several screening tests are used to segregatecushingoid disorders from Cushing’s syndrome on an out-patient basis, the overnight dexamethasone test (Nugentet al.,1965) and the measurement of basal urinary free cortisol areclearly the simplest and most reliable (Crapo, 1979). Once apatient is labelled as a ‘suspect Cushing’, given the abnormalresults of at least one of the above-mentioned tests, severalexpensive and lengthy in-patient procedures are usuallyperformed to confirm the diagnosis of Cushing’s syndromeand establish an ætiology. Hopes of obtaining a ‘definitivetest’ for the differential diagnosis of Cushing’s syndromewere increased when ovine corticotrophin releasing hormone(oCRH) became available (Mu¨ller et al., 1983; Chrousoset al.,1984, 1985; Cataniaet al., 1984). A decade of experience withCRH testing in the diagnosis of Cushing’s syndrome has revealedits many limitations which led Orth (1992) to disregard it as aroutine diagnostic test in Cushing’s syndrome. Even within thenarrow context of established ACTH dependent hypercortisolismthe usefulness of CRH testing is questionable since some pituitarycases fail to respond and some ectopic ACTH cases may respond.Moreover, the scope for use of CRH is tempered by the fact that itis not generally available.

In euadrenal subjects it has been shown that overnight pre-treatment with dexamethasone greatly attenuates the adrenal

703# 1996 Blackwell Science Ltd

Correspondence: Patricio Contreras, Jorge Matte 1390, Santiago,Chile.

axis response to various corticotroph stimuli such as vasopres-sin (either lysine or arginine-vasopressin) or oCRH (Hohnloseret al., 1989), thus reflecting the sensitivity of their pituitarycorticotrophs to glucocorticoid negative feedback. In contrast,pituitaries from patients with Cushing’s disease can be consideredto possess two sub populations of pituitary corticotrophs, one ofthem chronically and appropriately suppressed by ambienthypercortisolæmia (‘glucocorticoid-sensitive’, normal cortico-trophs) and the other one (‘glucocorticoid-insensitive’—usually tumour corticotrophs) still actively secreting ACTHin spite of hypercortisolism. We therefore reasoned that astandard dose of overnight dexamethasone suppression wouldnot suppress the ability of patients with Cushing’s disease toexhibit a cortisol response to lysine-vasopressin (LVP,dexamethasone insensitive axis), thus allowing a betterseparation between them and euadrenal subjects. On theother hand, pituitary independent Cushing’s (adrenal tumoursor ectopic ACTH syndromes) should neither suppress follow-ing dexamethasone nor respond to LVP given that theirpituitary corticotrophs are already suppressed by chronichypercortisolaemia (non-responsive axis). Additionally, ourhypothesis predicts that the successful transsphenoidal removalof the abnormal sub population of glucocorticoid insensitivepituitary corticotrophs in Cushing’s disease should result in animmediate switch from a dexamethasone insensitive to adexamethasone sensitive pattern of response (indistinguishablefrom that of euadrenal subjects) to the dexamethasone modifiedLVP (Dx/LVP) test. Therefore, this study was made to explore thepotential diagnostic value of stimulating the adrenal axis withan ACTH secretagogue (LVP) under standard overnightdexamethasone suppression to segregate Cushing’s patientsfrom euadrenal subjects and to differentiate pituitary depen-dent from pituitary independent varieties. Our results confirmour hypothesis and support the notion that a better diagnosticvalue of corticotroph stimulation is to be expected whenperformed under standard, overnight dexamethasone suppression.Similarly, in a previous study from the NIH group the stimulationwith CRH (1�g/kg) 2 hours after completing a 2-day low-dosedexamethasone suppression was highly successful in segregatingpseudo-Cushing’s states from Cushing’s syndrome (Yanowskiet al., 1993).

Materials and methods

In order to develop a reliable test it was first tested on healthyvolunteers, including several staff members, which allowed usto delineate the criteria defining normal serum cortisolresponses during the test. We then applied the test to everypatient with Cushing’s syndrome (n� 34) whom we hadmanaged during a 10-year period (1983–1993). We chose notto exclude any patient from the final analysis to avoid bias. The

outcome of the Dx/LVP test did not influence the decision fortranssphenoidal surgery. In fact, a patient in whom an ectopicACTH syndrome was finally diagnosed who (earlier in thestudy) had been shown by MRI to harbour a pituitaryincidentaloma was mistakenly subjected to transsphenoidalsurgery even though she had responded to Dx/LVP testing in apattern characteristic of pituitary independent Cushing’s. Dueto budgetary constraints ACTH levels were not measured on aroutine basis. Our diagnostic strategy and results of therapy in50 consecutive patients with Cushing’s syndrome have beenrecently published (Contreras & Araya, 1995).

After obtaining individual consent the Dx/LVP test wasperformed on a group of 61 subjects from the followingcategories: (a) 18 healthy control subjects (15 females and 3males); (b) 25 patients (20 females and 5 males) with areasonably well demonstrated pituitary dependent Cushing’sbased on various evidences, including suppression after anovernight 8-mg dexamethasone test, ACTH-immunostaining insome cases, positive pituitary MRI and/or CT-scan images andcures (n� 7) following adenomectomy; (c) nine patients (7females and 2 males) with unequivocal pituitary independentCushing’s, including 2 ectopic ACTH syndromes (onemalignant thymoma and one bronchial carcinoid) and 7 adrenaltumours (2 adenomas and 5 carcinomas); (d) four patients (allwomen) with a recently diagnosed major depressive syndromefulfilling DMS III criteria, and (e) five obese, pseudo-Cushing’spatients (defined as those with a normal overnight dexametha-sone suppression test but with a free basal urinary cortisolbetween 276 and 552 nmol/24 h in whom hypercortisolism wasultimately ruled out by showing suppression of their 1600 hserum cortisol during the second day of a standard ‘low-dose’(2 mg/day for 2 consecutive days) dexamethasone test to below55 nmol/l. In addition, 7 pituitary-dependent Cushing’s suc-cessfully operated upon by transsphenoidal adenomectomywere tested again between 10 days and 1 month followingsurgery.

The Dx/LVP procedure was as follows: at 2300 h the subjectsreceived 1 mg dexamethasone (Oradexon, Organon) orally andthe following morning a heparin lock was placed in anantecubital vein. Between 0800 and 0900 h 10 IU of lysine-vasopressin (Lipresina, Sandoz) were injected i.m. and serialblood samples were obtained atÿ15, 0, 15, 30, 45, 60, 75, 90and 120 minutes. After separation of plasma, samples werekept frozen atÿ208C until processing. Cortisol levels weresubsequently measured by a modification of the assay describedby Wong et al. (1979) featuring a second antibody–polyethyleneglycol separation (interassay CV less than 12%)instead of using a preconjugated double antibody. All samplesfrom a given test were processed in a single assay. Side-effectsof the procedure were the frequent appearance of pallor ofthe skin and, occasionally, abdominal cramps. A slight

704 P. Contreras and V. Araya

# 1996 Blackwell Science Ltd,Clinical Endocrinology, 44, 703–710

increase in arterial pressure was often observed but we did notobserve any hypertensive episode following the i.m. injection ofLVP.

We found four patterns of serum cortisol behaviour duringthe Dx/LVP test according to the various combinations(positive or negative) of dexamethasone-induced cortisolsuppression and cortisol responses to LVP (see Table 1, panela): (1) the dexamethasone sensitive pattern (positive suppres-sion and negative response): mean serum cortisol atÿ15 and0 minutes less than 138 nmol/l and maximal cortisol responseto LVP (� cortisol over mean post-dexamethasone value)less than 166 nmol/l; (2) theindeterminate pattern (positivesuppression and positive response) : mean cortisol value postdexamethasone less than 138 nmol/l and maximal cortisolresponse to LVP higher than 166 nmol/l; (3) the dexamethasoneinsensitive pattern (negative suppression and positive response): mean post-dexamethasone cortisol value higher than138 nmol/l and maximal cortisol response to LVP higher than

166 nmol/l or higher than 60% above post-dexamethasonevalues; (4) non-responsive pattern (negative suppression andnegative response): mean post-dexamethasone cortisol valuehigher than 138 nmol/l and maximal cortisol response to LVPless than 40% above post-dexamethasone levels.

Cortisol values were expressed as means� standard errors(SE). Standard definitions were used for the variousperformance variables of the Dx/LVP test (Galen & Gambino,1975) (see Table 2). We calculated the performance variablesof the Dx/LVP test in segregating (a) control subjects from allCushing’s (pituitary dependent plus pituitary independent);(b) pituitary dependent Cushing’s from control subjects;and (c) pituitary independent from pituitary dependentCushing’s.

Additionally, a two-way ANOVA with repeated measuresallowed us to compare the results of Dx/LVP testing in 7pituitary dependent Cushing’s before and after successfultranssphenoidal surgery. Comparison between Dx/LVP testing

Dexamethasone-vasopressin test for Cushing’s 705


(a) Cortisol response

Negative PositivePositive Dexa sensitive (A) Indeterminate (B)

Dexamethasone suppressionNegative Non-responsive (D) Dexa insensitive (C)

(b)Control subjects A� 16 B� 2(n� 18) (88.9) (11.1)

D � 0 C� 0

Pituitary dependent Cushing’s A� 0 B� 1(n� 25) (4)

D � 0 C� 24(96)

Pituitary independent Cushing’s A� 0 B� 0(n� 9) D� 9 C� 0

(100)

Depressed subjects A� 3 B� 1(n� 4) (75) (25)

D � 0 C� 0

Values in parentheses are percentages.Panel (a) shows the four patterns of serum cortisol responses (A, dexa sensitive; B,indeterminate; C, dexa insensitive; D, non-responsive) observed during the Dx/LVP test in our61 subjects.Panel (b) shows the number of responses for each of these patterns in control subjects, pituitarydependent Cushing’s, pituitary independent Cushing’s and depressed subjects (all five cushingoidpatients exhibited a dexa sensitive pattern). Therefore a ‘true positive’ response for control subjectswas pattern A (dexa sensitive), whereas for pituitary dependent and pituitary independentCushing’s ‘true positive’ responses were patterns C (dexa insensitive) and D (non-responsive),respectively.

Table 1 Patterns of cortisol behaviour duringDx/LVP testing

in 18 control subjects and 4 depressed patients was done bygroupt-test of the areas under the cortisol curve (calculated bythe trapezoid rule)

Results

As seen in panel (b) of Table 1 and Fig. 1a, 16/18 (88.9%) ofcontrol subjects exhibited a dexamethasone sensitive pattern ofresponse and 2/18 (11.1%) an indeterminate pattern, while noother patterns of response were observed in this group. Asshown in panel (b) of Table 1 and Fig.1b, 24 out of 25 (96%)pituitary dependent Cushing’s exhibited a dexamethasoneinsensitive pattern of response while only 1 patient exhibitedan indeterminate pattern. No other patterns of response wereobserved among these patients. Also seen in panel b of Table l

and Fig. 1c is that 9/9 (100%) pituitary independent Cushing’sexhibited a non-responsive pattern. Three out of 4 (75%)depressed subjects exhibited a dexamethasone sensitive patternof response while the remaining patient (25%) showed anindeterminate pattern. Similarly, all cushingoid patientsexhibited a dexamethasone sensitive pattern of response.Because of these findings the pattern representative of controlsubjects (euadrenalism) was considered to be the first described(dexa sensitive), whereas the third one (dexa insensitive) wasseen only in pituitary dependent Cushing’s. The fourth pattern,suggestive of non-responsive pituitary corticotrophs, wasobserved exclusively in pituitary independent Cushing’s,irrespective of the adrenal or non-adrenal (ectopic ACTH)origin of the disorder; during the test, the only observeddifference between them was the higher level of serum cortisolin the ectopic ACTH group (Fig 1c).

As seen in Table 2, in separating control subjects (n� 18;true positive test: dexa sensitive pattern) from all Cushing’s(n� 34) the Dx/LVP test had 88.9% sensitivity, 100%specificity (a dexa sensitive pattern was seen only in controlsubjects), 100% positive predictive value, 96.9% negativepredictive value and an overall diagnostic value of 96.2%.Similarly, in separating pituitary dependent Cushing’s (n� 25;true positive test: dexa insensitive pattern) from controlsubjects (n� 18) the Dx/LVP test showed 96.0% sensitivity,100% specificity, 100% positive predictive value, 94.7%negative predictive value and 97.7% overall diagnosticaccuracy. Again, a dexa insensitive pattern was seen exclu-sively in pituitary dependent Cushing’s. Finally, the Dx/LVPtest was 100% effective in separating pituitary-independent(n� 9; true positive test: non-responsive pattern), frompituitary dependent Cushing’s (n� 25).

In Fig. 1, the individual cortisol responses during the testare shown for control subjects, Cushing’s disease andpituitary independent Cushing’s (a, b and c, respectively).The mean�SE cortisol levels during the test for each of thesegroups are shown in Fig. 2. In Fig. 3, the mean�SE cortisolresponses are shown for control subjects and depressedpatients. The respective areas under the curve of cortisolresponse were 67.3�17.1 and 86.5�16.5 nmol/l min(P > 0:05). All five pseudo-Cushing’s subjects exhibited(not shown) a dexa sensitive pattern, indistinguishable fromthat exhibited by most (16/18) control subjects.

In Fig. 4, the mean�SE cortisol levels during the test areshown for 7 patients with Cushing’s disease before andfollowing successful transsphenoidal surgery. Every curedpatient switched from a preoperative dexa insensitive to apost-operative dexa sensitive pattern during the test. A two-way ANOVA with repeated measures (performed to comparepre and post-operative cortisol responses) yielded anF-valueof 41.7 (P < 0:001).



(c)

(b)

LVP 10 U im

(a)

120

4000

0

Time (min)

Co

ntr

ol (n

mo

l/l)

2000

1000

3000

300 60 90

LVP 10 U im

2000

0

1000

500

1500

1000

0

500 LVP 10 U im

Fig. 1 Individual serum cortisol responses during the Dx/LVP test ina, 18 control subjects; b, 25 with Cushing’s disease; and c, 9 withpituitary independent Cushing’s (l, 2 ectopic ACTH;k, 7 adrenaltumours).

Discussion

In a pilot study of 6 normal subjects given 10 IU LVP i.m. at0800 h we showed a significant inverse relation (r � ÿ0:89,P < 0:02) between baseline cortisol levels (251�60.7 nmol/l,range 124–505) and increment above baseline (187� 55%,range 25–358) following the challenge with LVP. The cortisolresponse to another ACTH secretagogue, oCRH (100�g i.v. at0815 h), is also attenuated in normal subjects by high basal serumcortisols, as shown by Lytraset al. (1984): baseline cortisol385�35 nmol/l, range 172–720; increment above baseline92.3�16.2%, range 12–236,r � ÿ0:74, P < 0:01. Thesefindings underline the modulatory effect of endogenous cortisol

on the responsiveness of the adrenal axis to ACTH secretago-gues. In an effort to discriminate between euadrenal subjectsand pituitary dependent Cushing’s, CRH testing has frequentlybeen performed at 2000 h to take advantage of the circadiandecline in serum cortisol levels in euadrenal subjects (Schulteet al., 1985). While euadrenal subjects exhibit baseline 2000 hserum cortisols lower than pituitary Cushing’s and tend torespond less to CRH, there is still a significant overlap in theirresponses; moreover, patients with Cushing’s disease oftenexhibit a cortisol response to CRH falling entirely within thenormal range (Orthet al., 1985).

Arginine-vasopressin coadministered i.m. with i.v. oCRHhas been shown to potentiate both ACTH and cortisol responses(Liu et al., 1983; DeBoldet al., 1984; Lambertset al., 1984). It



Table 2 Performance variables (% and absolute) of the Dx/LVP test in the diagnosis of Cushing’s syndrome

Sensitivity Specificity Positive PV Negative PV Diagnostic accuracy

Controls� vsCushing’s 88.9 100 100 96.9 96.2(16/18) (34/34) (16/16) (34/36) (50/52)

Pituitary dependent Cushing’s�� 96.0 100 100 94.7 97.7vscontrols (24/25) (18/18) (24/24) (18/19) (42/43)

Pituitary independent�� 100 100 100 100 100vspituitary dependent Cushing’s (9/9) (25/25) (9/9) (25/25) (34/34)

Positive response:� type A; �� type C;�� type D.Sensitivity�TP/TP� FNSpecificity�TN/TP� FPPositive predictive value (PV)�TP/TP� FPNegative predictive value�TN/TN� FN

Overall diagnostic accuracy� TP� TNTP� TN� FP� FN

TP, true positive; TN, true negative; FP, false positive; FN, false negativeNote that the specificity of each true positive response was 100% for each of the 3 comparisons made: pattern A (dexa sensitive) was exclusive ofeuadrenal subjects, pattern C (dexa insensitive) was exclusive of Cushing’s disease, whereas pattern D (non responsive) was exclusive of pituitaryindependent Cushing’s.

120

1500

00

Time (min)

Co

rtis

ol (n

mo

l/l)

1000

500

60 9030

LVP 10 U im

2000

Fig. 2 Mean�SE serum cortisol levels during the Dx/LVP test areshown foru, control subjects (dexa sensitive pattern) (n� 18);k,Cushing’s disease (dexa insensitive pattern) (n� 25) andl, pituitaryindependent Cushing’s (non-responsive pattern) (n� 9).

1200

0

Time (min)

Co

rtis

ol (n

mo

l/l)

500

60 9030

LVP 10 U im

1000

Fig. 3 Mean�SE serum cortisol responses during the Dx/LVP testin l, depressed patients (n� 4) andk, control subjects (n� 18).

should be noted though that Tabarinet al. (1990) found falsenegative cortisol responses to LVP or oCRH in 4 and 3 out of 21patients with Cushing’s disease, respectively.

The above mentioned limitations on the stimulation of theadrenal axis with ACTH secretagogues have reduced theusefulness of these manoeuvres in discriminating between thetwo ACTH dependent varieties of Cushing’s syndrome. Evenwithin this narrow context, oCRH testing is limited by two facts: about 10% of the pituitary dependent variety fail to respond tothis ACTH secretagogue, and 6% of ectopic ACTH syndromesshow a positive response to CRH (Lytraset al., 1984; Hermuset al., 1986; Niemanet al., 1986).

Since it appears that 10 IU LVP i.m. is more potent than100�g oCRH i.v. in eliciting cortisol responses in euadrenalsubjects: 187�55% (our data) vs 92�16% (Lytras et al.,1984), the same phenomenon should apply to Cushing’sdisease. This seems to be the case if one considers that, despitepre-treatment with 1 mg of dexamethasone, our patients withCushing’s disease (n� 25) invariably exhibited a substantialcortisol response (170:4�32.3%) to LVP while Niemanet al.(1986) found a smaller mean cortisol response to 100�g oCRHamong 33 patients with Cushing’s disease: 65.2�9.9% with 4patients showing no response.

Regarding the performance of the Dx/LVP test in our study itshould be noted that the crude separation between euadrenalcontrols (dexa-sensitive pattern) and hypercortisolism of anyorigin was acceptable since specificity was 100% and, moreimportantly, the predictive value of a dexa sensitive responsewas 100%. In other words, a dexa sensitive response was notobserved in any case of Cushing’s syndrome; thus, this patternreasonably indicates the absence of the disorder. Our 18 controlsubjects invariably exhibited a positive suppression followingdexamethasone administration and 16/18 (88.9%) were unableto respond to LVP (dexa sensitive pattern), a logicalconsequence of the overnight dexamethasone induced negative

feedback on glucocorticoid sensitive pituitary corticotrophs; onthe contrary, 24/25 (96%) patients with pituitary dependentCushing’s syndrome did not suppress after dexamethasonewhile retaining their ability to respond to LVP (dexa insensitivepattern), reflecting the presence of abnormal, ‘glucocorticoidinsensitive’ pituitary corticotrophs. In this way, the Dx/LVPtest was highly efficient in separating patients with Cushing’sdisease from euadrenal controls, achieving 100% specificityand 97.7% diagnostic accuracy. A dexa insensitive pattern ofserum cortisol response during the test was seen exclusively inCushing’s disease. This situation compares very favourablywith the limited ability of the CRH test to segregate euadrenalsubjects from Cushing’s disease (Orthet al., 1985; Goldet al.,1986). Finally, the performance of the Dx/LVP test in ourseries was optimal in segregating pituitary independent(non-responsive pattern) from pituitary dependent causes ofCushing’s syndrome since every performance variable achieved100%. Again, a non-responsive pattern was seen exclusively inpituitary independent Cushing’s whose pituitary corticotrophsare chronically suppressed by ambient hypercortisolæmia.

Altogether, out of four possible patterns of cortisol responseduring the test, only the indeterminate one was equivocal andnot indicative of any condition; it was observed in 2/18 controlsubjects, 1/25 patients with pituitary dependent Cushing’s and1/4 depressed subjects; it was not observed though amongpituitary independent Cushing’s. Our 5 pseudo-Cushing’spatients exhibited a dexa sensitive pattern, characteristic ofeuadrenal subjects, which is consistent with our hypothesis. Itshould noted though, as mentioned by the NIH group(Yanowski et al., 1993), that lack of suppression of plasmacortisol following overnight 1-mg dexamethasone may be seenin more major pseudo-Cushing’s states such as chronicalcoholism, anorexia and bulimia nervosa, stressful conditionsand others. These authors studied 58 patients, including 39surgically demonstrated Cushing’s syndromes and 19 pseudo-Cushing’s (none of them suppressed appropriately following a1-mg overnight dexamethasone administration). A serum cortisollevel greater that 38 nmol/l 15 minutes after the administration ofCRH (2 hours after completing a 2-day low-dose dexamethasonesuppression) was seen in every patient with Cushing’s syndromeand not seen in any patient with pseudo-Cushing’s (100%diagnostic accuracy). This study, in agreement with ours,demonstrates that euadrenalism can be clearly separatedfromhypercortisolism by challenging with ACTH secretagogues indexamethasone suppressed subjects. In fact, lack of suppressionfollowing 1-mg overnight dexamethasone administration inpseudo-Cushing’s states invariably reverted to normal suppres-sion after a 2-day low-dose dexamethasone administration inthe NIH study. This finding would indicate that by increasingthe duration of dexamethasone administration it should bepossible to demonstrate a higher glucocorticoid sensitivity in



120

1500

00

Time (min)

Co

rtis

ol (n

mo

l/l)

1000

500

60 9030

LVP 10 U im

Fig. 4 Mean�SE serum cortisol levels during the Dx/LVP testperformedk, before (dexa insensitive pattern) andl, after (dexasensitive pattern) successful transsphenoidal removal of the pituitarymicroadenoma in 7 cases of Cushing’s disease.

pseudo-Cushing’s as compared to patients with hypercortiso-lism. Our pseudo-Cushing’s patients were different from thosein the NIH study in that their cortisol hypersecretion wasinvariably suppressed following overnight dexamethasone, inspite of some elevation in their urinary free cortisols. Alimitation in the interpretation of our results is the rather smallnumber of patients with pseudo-Cushing’s syndrome (n� 5) inour series. The same is true for the ectopic ACTH syndromes(n� 2).

One aspect of great interest in our work is the fact that our 7cases of Cushing’s disease cured by transsphenoidal adeno-mectomy invariably switched from a preoperative dexainsensitive to a post-operative dexa sensitive pattern, typicallyseen in control subjects. We consider this phenomenon to be anatural consequence of the successful removal of theglucocorticoid insensitive sub-population of pituitary corti-cotrophs. The remaining, normal, glucocorticoid sensitive,non-tumour pituitary corticotrophs are chronically sup-pressed by hypercortisolism and overnight dexamethasonetreatment further precludes any response of the adrenal axisto ACTH secretagogues. This finding could be used as a clearcriterion for cure after transsphenoidal removal of thepituitary microadenoma.

It is very likely that by using either i.v. oCRH ordesmopressin after overnight, standard dexamethasone sup-pression, results comparable to ours would be obtained. Infact, our preliminary ongoing experience with desmopressin10�g i.v. substituted for LVP 10 IU i.m. following overnightdexamethasone suppression has so far been highly consistentwith this assumption. We conclude that, although a ‘definitivetest’ for the differential diagnosis of Cushing’s syndrome stillremains an elusive goal, our proposed Dx/LVP test representsan inexpensive, highly efficient alternative to the moresophisticated tests available.

Acknowledgements

We gratefully acknowledge the referral of patients by our staffcolleagues. We also thank Ine´s Vega, RN for her invaluablehelp in carrying out the Dx/LVP tests. Supported in part byProject D.I.B. M-1869 8745, University of Chile.

References

Catania, A., Cantalamassa, L., Orsatti, A., Mosca, G., Minonzio, F.,Motta, P., Reschini, E. & Zanussi, C. (1984) Plasma ACTH responseto the corticotropin releasing factor in patients with Cushing’sdisease. Comparison with lysine vasopressin test.Metabolism, 33,478–481.

Chrousos, G.P., Schulte, H.M., Oldfield, E.H., Gold, P.W., Cutler, G.B.

& Loriaux, D.L. (1984) The corticotropin releasing factor stimula-tion test: an aid in the evaluation of patients with Cushing’ssyndrome.New England Journal of Medicine, 310, 622–626.

Chrousos, G.P., Schuermeyer, T.H., Doppman, J., Oldfield, E.H.,Schulte, H.M., Gold, P.W. & Loriaux, D.L. (1985) Clinicalapplications of corticotropin releasing factor.Annals of InternalMedicine, 102, 344–358.

Contreras, P. Araya ,V. (1995) Si´ndrome de Cushing : una revisio´n aproposito de una casui´stica nacional.Revista Medica de Chile, 123,350–362.

Crapo, L. (1979) Cushing’s syndrome: a review of diagnostic tests.Metabolism, 23, 955–977.

DeBold, C.R., Scheldon, W.R., DeCherney, G.S., Jackson, R.V.,Alexander, A.N., Vale, W., Rivier, J. & Orth, D.N. (1984) Arginine-vasopressin potentiates adrenocorticotropin release induced by ovinecorticotropin-releasing factor.Journal of Clinical Investigation, 73,533–538.

Galen, R.S. & Gambino, S.R., (eds) (1975) Beyond normality : thepredictive value and efficiency of medical diagnoses, John Wiley &Sons: New York pp. 10–33.

Gold, P.W., Loriaux, D.L., Roy, A., Kling, M.A., Calabrese, J.R.,Kellner, C.H., Nieman, L.K., Post, R.M., Pickar, D., Galluci, W.,Avgerinos, P., Paul, S., Oldfield, E.H., Cutler, Jr G.B. & Chrousos,G.P. (1986) Responses to corticotropin releasing hormone in thehypercortisolism of depression and Cushing’s disease.NewEngland Journal of Medicine, 314,1329–1335.

Hermus, A.R., Pieters, G.F., Pesman, G.J., Smals, A.G., Benraad, T.J. &Kloppenborg, P.W. (1986) The corticotropin-releasing hormonetest versus the high dose dexamethasone test in the differentialdiagnosis of Cushing’s syndrome.Lancet, ii, 540–543.

Hohnloser, J., vWerder, K. & Mu¨ller, O.A. (1989) Acute dexametha-sone suppression of ACTH secretion stimulated by human cortico-trophin releasing hormone, AVP and hypoglycemia.ClinicalEndocrinology, 31, 175–184.

Lamberts, S.W.J., Verleun, T., Oosterom, R., de Jong, F. & Hackeng,W.H.L. (1984) Corticotropin-releasing factor (ovine) and vasopres-sin exert a synergistic effect on adrenocorticotropin release in man.Journal of Clinical Endocrinology and Metabolism, 58, 298–303.

Liu, J.H., Muse, K., Contreras, P., Gibbs, D., Vale, W., Rivier, J. &Yen, S.C.C. (1983) Augmentation of ACTH-releasing activity ofsynthetic corticotropin-releasing factor (CRH) by vasopressin inwomen. Journal of Clinical Endocrinology and Metabolism, 57,1087–1089.

Lytras, N., Grossman, A., Perry, L., Tomlin, S., Wass, J.A.H., Coy,D.H., Schally, A.V., Rus, L.H. & Besser, G.M. (1984) Corti-cotrophin releasing factor : response in normal subjects andpatients with disorders of the hypothalamus and pituitary.Clinincal Endocrinology, 20, 71–84.

Muller, O.A., Stalla, G.K. & Werder, K. (1983) Corticotropin-releasingfactor: a new tool for the differential diagnosis of Cushing’ssyndrome.Journal of Clinical Endocrinology and Metaboblism,57, 227–229.

Nieman, L.K., Chrousos, G.P., Oldfield, E.H., Avgerinos, P.C., Cutler,G.B. & Loriaux, D.L. (1986) The ovine corticotropin-releasinghormone stimulation test and the dexamethasone suppression test inthe differential diagnosis of Cushing’s syndrome.Annals of InternalMedicine, 105,862–866.

Nugent, C.A., Nichols, T. & Tyler, F.H. (1965) Diagnosis of Cushing’ssyndrome. Single dose dexamethasone suppression test.Archives ofInternal Medicine, 115,602–605.



Orth, D.N. (1992) Corticotropin-releasing hormone in humans.Endocrine Reviews, 13, 164–191.

Orth, D.N., DeBold, C.R. & DeCherney, G.S. (1985) Clinical studieswith synthetic ovine corticotropin-releasing factor.FederationProceedings, 44, 197–202.

Schulte, H.M., Chrousos, G.P., Oldfield, E.H., Gold, P.W., Cutler, GB.& Loriaux, D.L. (1985) Ovine Corticotropin-releasing factoradministration in normal men. Pituitary and adrenal responses inthe morning and evening.Hormone Research, 21, 69–74.

Tabarin, A., San Galli, F., Dezoce, S., Leprat, F., Corcuff, J.B., Latapie,J.L., Guerin, J. & Roger, P. (1990) The corticotropin-releasing factor

test in the differential diagnosis of Cushing’s syndrome : Acomparison with lysine-vasopressin test.Acta Endocrinologica(Copenh.), 123,331–338.

Wong, P.Y., Mee, A.V. & Ho, F.F. (1979) A direct radioimmunoassay ofserum cortisol with in-house125I-tracer and preconjugated doubleantibody.Clinical Chemistry, 25, 914–917.

Yanowski, J.A., Cutler, G.B., Chrousos, G.P. & Nieman, L.K.(1993) Corticotropin-releasing hormone stimulation followingdexamethasone administration. A new test to distinguish Cush-ing’s syndrome from pseudo-Cushing’s states.JAMA, 269, 2232–2238.



OVERNIGHT DEXAMETHASONE PRE-TREATMENT IMPROVES THE PERFORMANCE OF THE LYSINE-VASOPRESSIN TEST IN THE...

Documents

Transcript of OVERNIGHT DEXAMETHASONE PRE-TREATMENT IMPROVES THE PERFORMANCE OF THE LYSINE-VASOPRESSIN TEST IN THE...