Urine Free Cortisol in the High-Dose Dexamethasone
Suppression Test for the Differential Diagnosis of the
Cushing Syndrome
Mary R. Flack, MD; Edward H. Oldfield, MD; Gordon B. Cutler, Jr., MD; Mark H. Zweig, MD;
James D. Malley, PhD; George P. Chrousos, MD; D. Lynn Loriaux, MD, PhD;
and Lynnette K. Nieman, MD
Objective: To develop criteria for interpreting the
high-dose dexamethasone suppression test using urine
free Cortisol as an end point for the differential diagno-
sis of the Cushing syndrome.
Design: Retrospective review.
Setting: Inpatient research ward.
Patients: Patients (118) with surgically confirmed
causes of the Cushing syndrome: 94 with pituitary dis-
ease, 14 with primary adrenal disease, and 10 with ec-
topic adrenocorticotropic hormone (ACTH) secretion.
Main Outcome Measures: The sensitivity, specificity,
and diagnostic accuracy were determined for the high-
dose dexamethasone suppression test using urine free
Cortisol and using 17-hydroxysteroid excretion. For
each analysis, patients with pituitary disease were
considered to be "diseased" and patients with nonpi-
tuitary disease were considered to be "non-diseased".
The level of suppression that gave 100% specificity
was determined for each steroid.
Results: The accuracy of urine free Cortisol when
used as an end point in the high-dose dexamethasone
suppression test was equivalent to that of 17-
hydroxysteroid excretion. At all levels of sensitivity and
specificity, however, the degree of suppression of urine
free Cortisol used for the diagnosis of pituitary disease
was greater than that of 17-hydroxysteroid excretion.
The likelihood ratios for pituitary disease based on urine
free Cortisol suppression of > 50%, of > 80%, and of
> 90% were 4.2,10.1, and "infinite," respectively. Sup-
pression of urine free Cortisol greater than 90% or
suppression of 17-hydroxysteroid excretion greater
than 64% was associated with 100% specificity. When
these criteria were combined, the percentage of correct
predictions (102 of 118 [86%; 95% CI, 78% to 92%])
was higher than that obtained using either steroid alone
(89 of 118 [75%; CI, 65% to 83%]) ( P = 0.009) and
higher than that obtained using the traditional criterion
of 50% suppression for 17-hydroxysteroid excretion
(95 of 118 [80%; CI, 7 1 % to 87%]) ( P = 0.016).
Conclusions: In the high-dose dexamethasone
suppression test, the degree of suppression of urine
free Cortisol used for the diagnosis of pituitary disease
is greater than that traditionally used for 17-hydroxy-
steroid excretion. The diagnostic performance of the
test is improved by measuring both urine free Cortisol
and 17-hydroxysteroid excretion and by requiring
greater suppression of both steroids.
Annals of Internal Medicine. 1992;116:211-217.
For current author affiliations and addresses, see end of text.
1 February 1992 Annals of Internal Medicine Volume 116 Number 3
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19743/ by a University of California San Diego User on 03/06/2017
A urine free Cortisol measurement is useful for distin-
guishing normal persons from those with the Cushing
syndrome (1-4). In contrast, few data are available on
the use of urine free Cortisol as an end point in tests
designed to distinguish among the different causes of
the Cushing syndrome (4-6). The high-dose dexametha-
sone suppression test is the most commonly used test
for the differential diagnosis of the Cushing syndrome
(7). In the standard test, dexamethasone is given at
doses of 0.5 mg every 6 hours for 2 days (low dose),
followed by 2.0 mg every 6 hours for 2 days (high
dose). A greater than 50% drop in 17-hydroxysteroid
excretion on day 2 of high-dose dexamethasone admin-
istration is considered to be a positive test for pituitary
disease, whereas an absence of this response suggests
primary adrenal disease or ectopic secretion of adreno-
corticotrophic hormone (ACTH) (1, 2, 8-10).
The diagnostic accuracy of the dexamethasone sup-
pression test may be limited by incomplete urine col-
lection, daily fluctuations in basal steroid excretion, and
possible inaccuracies in the 17-hydroxysteroid assay
that are caused by medications or renal and hepatic
disease (11, 12). Nonetheless, because the test is non-
invasive and widely available, it continues to be used
for the differential diagnosis of the Cushing syndrome.
In clinical practice, urine free Cortisol is often used as
the end point rather than 17-hydroxysteroid excretion
because urine free Cortisol is unaffected by medications
or concurrent illness and the assay is sometimes more
readily available and easier to do than the assay for
17-hydroxysteroid excretion (3, 4, 9, 13-15). The proven
utility of urine free Cortisol as a superior index of hy-
percortisolism in establishing a diagnosis of the Cushing
syndrome (1, 3, 9) may lead to the assumption that it
would be more useful in tests designed for the differen-
tial diagnosis of the Cushing syndrome. It is not known,
however, whether measurement of urine free Cortisol
would improve the accuracy of the high-dose dexa-
methasone test for the differential diagnosis of the
Cushing syndrome.
Previous studies using urine free Cortisol as an end
point for the test have included limited numbers of
patients, particularly patients with ectopic ACTH-se-
creting tumors (4-6). Consequently, criteria for inter-
preting the test using urine free Cortisol as an end point
have not been developed. We examined the urine free
Cortisol response to high-dose dexamethasone in 118
patients with either pituitary or nonpituitary causes of
the Cushing syndrome. We also evaluated the perfor-
mance of both urine free Cortisol and 17-hydroxysteroid
211
 excretion as end points to identify a strategy for inter-
pretation of the test that would give the highest diag-
nostic accuracy while maximizing specificity.
Methods
Patient Selection and Treatment
The records of all patients with the Cushing syndrome who
were evaluated at the National Institutes of Health (NIH)
between January 1981 and October 1988 were reviewed. One
hundred and forty patients were identified who had undergone
dexamethasone suppression testing according to the protocol
outlined below. Twenty-three patients were not included in our
analysis: Seventeen patients (14 with pituitary disease, 2 with
ectopic ACTH secretion, and 1 with primary adrenal disease)
were excluded from the analysis because data were unavail-
able, and 5 patients were excluded because their diagnoses
could not be confirmed by the criteria outlined below. We do
not know what bias may have been introduced into our anal-
ysis by the exclusion of these patients. The remaining 118
patients were tested as part of an extensive diagnostic evalu-
ation that included ovine corticotropin-releasing hormone (16)
and metyrapone stimulation (6), inferior petrosal sinus sam-
pling for plasma concentrations of ACTH (17), and computed
tomographic scanning or magnetic resonance imaging of the
pituitary, chest, and abdomen. Based on the results of these
tests, a diagnosis was made and patients underwent surgery.
Ninety-four patients had pituitary disease based on cure by
transsphenoidal adenomectomy or partial hypophysectomy.
Cure was defined by a morning serum Cortisol value of less
than 190 nmol/L (normal, 193 to 690 nmol/L) and urine free
Cortisol of less than 41 nmol/d (normal, 55 to 248 nmol/d)
within 3 weeks of surgery. Fourteen patients had primary
adrenal disease confirmed by pathologic examination of the
adrenal tissue and resolution of hypercortisolism after surgery.
Eight patients with adrenal adenoma and two patients with
adrenal carcinoma had unilateral adrenalectomy. Four patients
with micronodular adrenal disease had bilateral adrenalectomy.
Ten patients underwent surgical removal of an ectopic source
of ACTH secretion. Immunohistochemical staining for ACTH
was positive in all tumors, and all but one patient, who had a
metastatic bronchial carcinoid tumor, had resolution of hyper-
cortisolism after tumor removal.
Dexamethasone Suppression Test
All patients had urine collected for 6 consecutive days. They
were encouraged to maintain as normal a routine as possible,
and no other testing was done during this time. On days 1 and
2 of the test, patients received no dexamethasone; on days 3
and 4, patients received 0.5 mg of of dexamethasone orally
every 6 hours (low dose); on days 5 and 6, patients received
2.0 mg of dexamethasone every 6 hours (high dose). The
24-hour collections were refrigerated and sent to the NIH
clinical chemistry laboratory within 4 hours after completion of
each collection. Urine volume and creatinine were measured to
confirm the completeness of each collection. The daily varia-
tion in creatinine excretion was less than 10% for each patient
on the 3 days used for calculating steroid suppression (days 1,
2, and 6).
Assays and Calculations
An aliquot of each urine collection was used for measure-
ment of 17-hydroxysteroid excretion by a modification of the
Porter-Silber method (18-20). The intra- and interassay coeffi-
cients of variation were 5.9% and 7% to 14%, respectively.
Another aliquot of urine was frozen and sent to SmithKline
Bioscience Laboratories (King of Prussia, Pennsylvania) for
measurement of Cortisol by radioimmunoassay using the Pre-
mix kit (Diagnostic Products Corporation, Los Angeles, Cali-
fornia). The intra- and interassay coefficients of variation were
5.4% and 9.3% to 11.5%, respectively.
The percent suppression of each steroid after administration
of high-dose dexamethasone was calculated for each patient
212 1 February 1992 Annals of Internal Medicine Volume 116 Number 3
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19743/ by a University of California San Diego User on 03/06/2017
using the following formula: 100 - [(UFC or 17-OHS on day
6 / average UFC or 17-OHS on days 1 and 2) x 100] where
UFC = urine free Cortisol and 17-OHS = 17-hydroxysteroid
excretion. Because our study was limited to differential diag-
nosis in patients with established hypercortisolism, we did not
calculate the percent suppression after low-dose dexametha-
sone administration. Low-dose dexamethasone suppression is
useful in the diagnosis of hypercortisolism but is not useful in
distinguishing among the different causes of the Cushing syn-
drome.
Because the daily fluctuation in steroid excretion among
patients with the Cushing syndrome is thought to be a major
cause of the inaccuracy in the high-dose dexamethasone sup-
pression test, we estimated the effect of this daily fluctuation
on our final test results by calculating the percent suppression
using the steroid value on either day 1 or day 2 independently
as the baseline value. We determined the number of patients
whose diagnostic assignment would change based on this fluc-
tuation and calculated the effect of this fluctuation on the
criteria that gave 100% specificity.
Statistical Analysis
Estimates of sensitivity [true-positive results / (true-
positive results + false-negative results)] and specificity
[true-negative results / (true-negative results + false-
positive results)] were determined for each steroid at
various suppression levels. For all calculations, the
presence of pituitary disease was considered to be a
positive response (diseased) and the presence of nonpi-
tuitary disease was considered to be a negative re-
sponse (nondiseased). Likelihood ratios at each sup-
pression level were determined by dividing the
sensitivity by (1-specificity) (21). When comparisons
were made between the results of urine free Cortisol and
17-hydroxysteroid excretion, correlated 2 x 2 tables
were constructed and the Cochran Q test was used to
determine the statistical significance of each comparison
(22). The P values were corrected using the Bonferroni
adjustment for multiple comparisons (23).
To allow comparison of the two tests at multiple
levels of steroid suppression without the bias of prede-
termined suppression criteria, curves of the receiver
operating characteristics (ROC) were constructed by
plotting the sensitivity against 1-specificity at each sup-
pression level using the RuleMaker program (Digital
Medicine, Inc., Hanover, New Hampshire). The area
under each ROC curve was calculated using the Rule-
Maker program by dividing the curve into trapezoidal
sections and calculating the sum of these areas. The
area under each ROC curve represents the inherent
accuracy of the test independent of the suppression
criteria (24). The RuleMaker program was used to com-
pare the two areas using a modification of Pearson
correlations for the analysis of paired data.
Results
Sensitivity and Specificity
The suppression of urine free Cortisol and 17-hydroxy-
steroid excretion was plotted for each patient by diag-
nosis (Figure 1). Suppression of steroid excretion of
more than 50% was considered to be a positive test for
pituitary disease. By this criterion, the sensitivity and
specificity of the test using urine free Cortisol as an end
point were 90% (CI, 83% to 95%) and 79% (CI, 58% to
 Figure 1. Suppression of urine free Cortisol and 17-hydroxysteroid excretion during a standard high-dose dexamethasone suppression
test in 118 patients with surgically confirmed causes of the Cushing syndrome. ACTH = adrenocorticotrophic hormone; 17-OHS =
17-hydroxysteroid; UFC = urine free Cortisol.

93%), respectively, whereas the sensitivity and speci-
ficity of the test using 17-hydroxysteroid excretion as
the end point were 78% (CI, 68% to 86%) and 92% (CI,
73% to 99%), respectively. Five patients with nonpitu-
itary disease (including one patient with an adrenal ad-
enoma) and four patients with ectopic ACTH secretion
(three with bronchial carcinoid tumors and one with a
pheochromocytoma) had greater than 50% suppression
of urine free Cortisol.
When a greater than 80% suppression of urine free
Cortisol was considered to be a positive test for pitu-
itary disease, the sensitivity of the test was 81% (CI,
71% to 88%) and the specificity was 92% (CI, 73% to
99%); these values are similar to the sensitivity and
specificity of the test using the more traditional cut-
point of 50% for 17-hydroxysteroid excretion (Table 1).
Using either the cut-point of 80% suppression for urine
free Cortisol or the cut-point of 50% suppression for
17-hydroxysteroid excretion, all patients with primary
adrenal disease were correctly identified, but two pa-
tients with ACTH-secreting bronchial carcinoids were
misclassified as having pituitary disease. Because of the
potential harm resulting from a misdiagnosis in these
patients, we identified criteria that gave 100% specificity
in our series. These criteria were a greater than 90%
suppression of urine free Cortisol and a greater than
64% suppression of 17-hydroxysteroid excretion. The
highest sensitivity of the test using either of these cri-
teria alone was 69% (CI, 59% to 78%). A combined
approach, however, in which either greater than 90%
suppression of urine free Cortisol excretion or greater
than 64% suppression of 17-hydroxysteroid excretion
was used to indicate pituitary disease, increased the
sensitivity to 83% (CI, 74% to 90%). The diagnostic
accuracy of this combined approach (86%, CI, 78% to
92%) was higher than that of either steroid alone (75%
CI, 65% to 83%) (P = 0.009) and higher than that of the
traditional cut point of 50% suppression for 17-hydroxy-
steroid excretion (80% CI, 71% to 87%) (P = 0.016).
Receiver Operating Characteristics and Likelihood
Ratios
The inherent accuracy of the test using either urine
free Cortisol or 17-hydroxysteroid excretion as an end
point was equivalent as assessed by the area-under-the-
ROC curves (Figure 2). The area-under-the-ROC curve
for urine free Cortisol (0.94) was the same as the area-
under-the-ROC curve for 17-hydroxysteroid excretion
(0.93) (P > 0.2). At all levels of sensitivity and speci-
ficity, however, a more profound degree of suppression
was required for the diagnosis of pituitary disease when

1 February 1992 Annals of Internal Medicine Volume 116 Number 3 213

Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19743/ by a University of California San Diego User on 03/06/2017

 using urine free Cortisol than when using 17-hydroxy s-
teroid excretion.
The likelihood ratios for pituitary disease using sup-
pression of urine free Cortisol of more than 50%, of
more than 80%, and of more than 90% were 4.2, 10.1,
and "infinite," respectively. The likelihood ratio for
pituitary disease using suppression of urine free Cortisol
of between 50% and 80% was only 0.63. The likelihood
ratios for pituitary disease using 17-hydroxy steroid sup-
pression of more than 50% and of more than 64% were
9.6 and "infinite," respectively.
Effect of the Daily Variability in Steroid Excretion
The mean coefficient of variation ( SD) for 17-hy-
droxysteroid excretion on the two basal days was
16% 15% (range, 0% to 105%). The mean coefficient
of variation for urine free Cortisol was 21% 18%
(range, 0% to 99%). The mean coefficient of variation
was similar for patients with pituitary disease, patients
with ectopic ACTH secretion, and patients with pri-
mary adrenal disease. Using the cut-point of 64% for
17-hydroxy steroid suppression, five patients (all with
pituitary disease) had sufficient variability in their 17-
hydroxysteroid excretion to change their diagnostic as-
signment if results from only one of the two basal days
were used to determine suppression. Similarly, the di-
agnostic assignment of six patients (five with pituitary
disease and one with ectopic ACTH secretion) could
have changed using either of the baseline day results for
urine free Cortisol and a 90% cut point. Thus, the sup-
pression criteria that gave 100% specificity could vary
by 4%, depending on which basal day was used to
calculate steroid suppression.
Discussion
Liddle (7) observed that 23 of 24 patients with pitu-
itary disease had a greater than 50% decrease in 17-
hydroxysteroid excretion on day 2 of high-dose gluco-
corticoid administration and that none of 8 patients with
adrenal tumors had this response. Subsequent series
have shown that the sensitivity (75% to 85%) and spec-
ificity (75% to 90%) of the high-dose dexamethasone
suppression test vary widely (1,2, 8-10, 25, 26). Several
factors contribute to diagnostic errors in the dexametha-
sone suppression test: These include daily fluctuations
in steroid excretion that are independent of the effects
of dexamethasone and some degree of steroid suppres-
sion in patients with ectopic ACTH secretion (1-5, 25-
27). Consequently, several tests are generally required
for the differential diagnosis of the Cushing syndrome.
Nonetheless, the high-dose dexamethasone suppression
test remains one of the important differential diagnostic
maneuvers because it is widely available, relatively
easy to perform, and noninvasive.
In clinical practice, urine free Cortisol is frequently
substituted for 17-hydroxy steroid excretion as the end
point in the dexamethasone suppression test (6, 28).
Although urine free Cortisol is known to be a useful
measure of hypercortisolism for the diagnosis of the
Cushing syndrome, few studies have evaluated urine
free Cortisol as an index of suppression during the high-
dose dexamethasone suppression test for the purpose of
differential diagnosis. No study has had findings suffi-
cient to propose criteria for the interpretation of the
high-dose dexamethasone test using the urine free Cor-
tisol value as an end point (4-6).
In the absence of published criteria, a greater than
50% suppression of urine free Cortisol has been taken to
indicate pituitary disease, for this is the level of sup-
pression generally used for 17-hydroxy steroid excretion.
However, for unknown reasons, urine free Cortisol is
suppressed more profoundly by dexamethasone than is
17-hydroxy steroid excretion. Cope and Black (29) noted
that administration of a potent oral glucocorticoid to
hypercortisolemic patients caused a sharper decrease in
Cortisol excretion than in the excretion of Cortisol
metabolites. They attributed this to the binding of Cor-
tisol by cortisol-binding globulin, limiting the amount of
free Cortisol available for filtration at the glomerulus.
The reverse of this phenomenon is seen as the sharp
linear increase in urine free Cortisol demonstrated for

Table 1. Sensitivity and Specificity of Different High-Dose Dexamethasone Suppression Test Criteria for Pituitary
Disease*
Criterion for Suppression Sensitivity [95% CI] Specificity [95% CI] Misclassification of
Patients with
Ectopic ACTH
Secretion

n/n(%) %
UFC > 50% 85/94 (90 [83 to 95]) 19/24 (79 [58 to 93]) 40
17-OHS > 50% 73/94 (78 [68 to 86]) 22/24 (92 [73 to 99]) 20
UFC > 80% 76/94 (81 [71 to 88]) 22/24 (92 [73 to 99]) 20
UFC > 90% 65/94 (69 [59 to 78]) 24/24 (100 [86 to 100]) 0
17-OHS > 64% 65/94 (69 [59 to 78]) 24/24 (100 [86 to 100]) 0
UFC > 90% or 17-OHS > 64%t 78/94 (83 [74 to 90]) 24/24 (100 [86 to 100]) 0

* ACTH = adrenocorticotrophic hormone; 17-OHS = 17-hydroxy steroid excretion; UFC = urine free Cortisol.
t P *= 0.009 for the diagnostic accuracy of the combined test compared with a urine free Cortisol suppression of more than 90% or with a
17-hydroxy steroid suppression of more than 64%. P = 0.016 for the diagnostic accuracy of the combined test compared with a suppression of
hydroxy steroid excretion of more than 50% (traditional criterion).

214 1 February 1992 Annals of Internal Medicine Volume 116 Number 3

Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19743/ by a University of California San Diego User on 03/06/2017

 Figure 2. Receiver operating char-
acteristic curves of the dexametha-
sone suppression test using either
urine free Cortisol or 17-hydroxy-
steroid excretion as an end point.
The area under each curve repre-
sents the inherent accuracy of the
test independent of specific sup-
pression criteria.
hypercortisolemic patients when the binding capacity of
cortisol-binding globulin is exceeded (30).
Although the more profound suppression of urine free
Cortisol enhances the sensitivity of the high-dose dex-
amethasone suppression test, it results in a dramatic
loss of specificity when the traditional 50% criterion is
applied to the urine free Cortisol results. When we used
a greater than 50% suppression of urine free Cortisol to
indicate a positive test for pituitary disease, 21% of our
patients with nonpituitary disease, including 40% of the
patients with ectopic ACTH secretion, were misclassi-
fied. In fact, the likelihood ratio for pituitary disease
using urine free Cortisol suppression of between 50%
and 80% (0.63) suggests that a patient is more likely to
have nonpituitary disease at this modest level of urine
free Cortisol suppression. The misidentification of these
patients may lead to unnecessary trans-sphenoidal sur-
gery and to a delay in the identification of other poten-
tially malignant tumors.
Leinung and colleagues (31) recently described two
patients with ACTH-secreting bronchial carcinoid tu-
mors who were initially diagnosed as having pituitary
disease. This diagnosis was based, in part, on 75% and
80% suppression of urine free Cortisol during high-dose
dexamethasone administration. One of these patients
underwent two unsuccessful transsphenoidal explora-
tions that resulted in panhypopituitarism before his lung
tumor was diagnosed. This case shows the importance
of using more stringent suppression criteria when inter-
preting the response of urine free Cortisol to high-dose
dexamethasone administration.
The rate of misdiagnosis in our patients with nonpi-
tuitary disease decreased when urine free Cortisol sup-
pression of more than 80% was considered to indicate
the presence of pituitary disease. The sensitivity (81%,
CI, 71 to 88%) and specificity (92% CI, 73% to 99%) of
1 February 1992 Annals of Internal Medicine Volume 116 Number 3
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19743/ by a University of California San Diego User on 03/06/2017
the test at this level of urine free Cortisol suppression
were equivalent to those obtained when a greater than
50% suppression of 17-hydroxy steroid excretion was
used to indicate pituitary disease. If the traditional 50%
cut-point for 17-hydroxy steroid suppression or the
equivalent but more stringent 80% cut-point for urine
free Cortisol were used to indicate pituitary disease,
then two of our patients with ACTH-secreting bronchial
carcinoid tumors would have been erroneously diag-
nosed as having pituitary disease. Although these pa-
tients represent a small fraction of our entire patient
population, they accounted for 20% of the patients with
ectopic ACTH secretion and represented a group who
remained at risk for inappropriate transsphenoidal sur-
gery.
We therefore identified criteria with 100% specificity
in our series. These criteria included a greater than 90%
suppression of urine free Cortisol and a greater than
64% suppression of 17-hydroxy steroid excretion. As ex-
pected, the sensitivity of the test using either of these
criteria alone was low (69%, CI, 59% to 78%); however,
sensitivity improved dramatically when the 17-hydroxy-
steroid and urine free Cortisol responses were com-
bined. When both steroids were measured and either a
greater than 90% suppression of urine free Cortisol or a
greater than 64% suppression of 17-hydroxy steroid ex-
cretion was considered to be a positive test for pituitary
disease, the sensitivity of the test increased to 83% (CI,
74% to 90%), and the diagnostic accuracy increased to
86% (CI, 78% to 92%). Why some patients with pitu-
itary disease had diagnostic suppression of one steroid
and not the other is unclear. The increased sensitivity
of the combined test may be a result of reducing labo-
ratory error through an additional measurement, or it
may reflect individual differences in patients' underlying
physiologies. Conceivably, duplicate measurements of a
215
 single steroid might have yielded a similar improvement
in diagnostic accuracy. Nevertheless, the combined ap-
proach appears to provide a useful way to increase the
sensitivity of the test when both measures of steroid
excretion are available.
How can these observations be integrated into the
diagnostic approach to a patient with the Cushing syn-
drome? Our data indicate that suppression of urine free
Cortisol by high-dose dexamethasone is as accurate as
suppression of 17-hydroxysteroid excretion for the dif-
ferential diagnosis of Cushing syndrome. A greater de-
gree of suppression of urine free Cortisol should be used
for the diagnosis of pituitary disease, than that tradi-
tionally used for 17-hydroxy steroid excretion. However,
failure to use more stringent suppression criteria when
evaluating urine free Cortisol results can lead to an
unacceptably high rate of misdiagnoses in patients with
ectopic ACTH secretion. In our series, we required a
greater than 90% suppression of urine free Cortisol as a
positive test for pituitary disease. The use of this cut
point eliminates misdiagnoses among patients with ec-
topic ACTH secretion. The equivalent criterion for 17-
hydroxysteroid excretion was a greater than 64% sup-
pression. By combining these two criteria, we achieved
a diagnostic accuracy that was higher than that obtained
with either steroid alone.
These cut points cannot be considered absolute be-
cause they are defined by our data set and may vary
somewhat when applied to a new patient population.
Furthermore, the percent suppression in an individual
patient can be affected by variations in urine collection,
laboratory errors, and daily fluctuations in steroid ex-
cretion. We estimated that the cut points for maximal
sensitivity at 100% specificity can vary by 4% because
of daily fluctuations in baseline steroid excretion. Thus,
data from patients with widely fluctuating steroid excre-
tion whose percent suppression falls close to these cut-
points need to be interpreted with some caution. In
addition, 17% to 31% of patients with pituitary disease
will not be definitively identified using the criteria out-
lined. In these patients, additional tests such as
metyrapone or ovine corticotropin-releasing hormone
stimulation or, perhaps, petrosal sinus sampling may be
needed to confirm the diagnosis in patients with a nor-
mal or elevated ACTH level (that is, > 10 pg/mL).
Nonetheless, our data indicate that the diagnostic per-
formance of the standard high-dose dexamethasone sup-
pression test is improved by measuring both urine free
Cortisol and 17-hydroxy steroid excretion and by requir-
ing more profound suppression of both steroids.
Acknowledgments: The authors thank the nursing staff and clinical
associates on the National Institute of Child Health and Human Devel-
opment endocrine units at the Clinical Center, National Institutes of
Health, for their care of the patients; and Dr. Robert Wesley for his
assistance in designing a computer database.
Affiliations: From the National Institute of Child Health and Human
Development, the National Institute of Neurological Disease and
Stroke, the Warren Grant Magnuson Clinical Center, and the Division
of Computer Research and Technology, National Institutes of Health,
Bethesda, Maryland. For current author addresses, see end of text.
Requests for Reprints: Mary R. Flack, MD, Building 10, Room 10N262,
Developmental Endocrinology Branch, National Institute of Child
216 1 February 1992 Annals of Internal Medicine Volume 116 Number 3
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19743/ by a University of California San Diego User on 03/06/2017
Health and Human Development, National Institutes of Health, 9000
Rockville Pike, Bethesda, MD 20892.
Current Author Addresses: Drs. Flack, Cutler, Chrousos, and Nieman:
Building 10, Room 10N262, Developmental Endocrinology Branch, Na-
tional Institute of Child Health and Human Development, National
Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892.
Dr. Oldfield: Building 10, Room 5D37, Surgical Neurology Branch,
National Institute of Neurological Disease and Stroke, National Insti-
tutes of Health, 9000 Rockville Pike, Bethesda, MD 20892.
Dr. Zweig: Clinical Chemistry Department, Clinical Center, National
Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892.
Dr. Malley: Building 12, Room 3053, Laboratory of Statistical and
Mathematical Methodology, Division of Computer Research and Tech-
nology, National Institutes of Health, 9000 Rockville Pike, Bethesda,
MD 20892.
Dr. Loriaux: Oregon Health Sciences University, Department of Medicine,
L456, 3181 SW Sam Jackson Park Road, Portland, OR 97201-3098.
References
1. Crapo L. Cushing's syndrome: a review of diagnostic tests. Metab-
olism. 1979;9:955-77.
2. Gold EM. The Cushing syndromes: changing views of diagnosis and
treatment. Ann Intern Med. 1979;90:829-44.
3. Eddy R, Jones A, Lloyd G, et al. Cushing's syndrome: a prospective
study of diagnostic methods. Am J Med. 1973;55:621-9.
4. Burke CW, Beardwell CG. Cushing's syndrome. An evaluation of
the clinical usefulness of urinary free Cortisol and other urinary
steroid measurements in diagnosis. Q J Med. 1973;42:175-204.
5. Nichols T, Nugent CA, Tyler FH. Steroid laboratory tests in the
diagnosis of Cushing's syndrome. Am J Med. 1968;45:116-28.
6. Sindler BH, Gritting GT, Melby JC. The superiority of the metyrap-
one test versus the high-dose dexamethasone test in the differential
diagnosis of Cushing's syndrome. Am J Med. 1983;74:657-62.
7. Liddle GW. Tests of pituitary-adrenal suppressibility in the diagnosis
of Cushing's syndrome. J Clin Endocrinol Metab. 1960;20:1539-61.
8. Kaye TB, Crapo L. The Cushing syndrome: an update on diagnostic
tests. Ann Intern Med. 1990;112:434-44.
9. Loriaux DL, Cutler GB. Diseases of the adrenal glands. In: Kohler
PO, ed. Clinical Endocrinology. New York: John Wiley & Sons;
1986:167-238.
10. Carpenter PC. Cushing's syndrome: update of diagnosis and man-
agement. Mayo Clin Proc. 1986;61:49-58.
11. Borushek S. Commonly used medications that interfere with routine
endocrine laboratory procedures. Clin Chem. 1964;10:41-52.
12. Young DS, Pestaner LC, Gibberman V. Effects of drugs on clinical
laboratory tests. Clin Chem. 1975;21:1D-432D.
13. Streeten DH, Anderson GH Jr, Dlakos TG, Seeley D, Mallov JS,
Eusebio R, et al. Normal and abnormal function of the hypothalam-
ic-pituitary-adrenocortical system in man. Endocr Rev. 1984;5:371-
94.
14. Beardwell CG, Burke CW, Cope CL. Urinary free Cortisol measured
by competitive protein binding. J Endocrinology. 1968;42:79-89.
15. Murphy BE. Clinical evaluation of urinary Cortisol determinations
by competitive protein binding radioassay. J Clin Endocrinol
Metab. 1968;28:343-8.
16. Nieman LK, Chrousos GP, Oldfield EH, Avgerinos PC, Cutler GB
Jr, Loriaux DL. The ovine corticotropin-releasing hormone stimu-
lation test and the dexamethasone suppression test in the differen-
tial diagnosis of Cushing's syndrome. Ann Intern Med. 1986;105:
862-7.
17. Oldfield EH, Chrousos GP, Schulte HM, Schaaf M, McKeever PE,
Krudy AG, et al. Preoperative lateralization of ACTH-secreting
pituitary microadenomas by bilateral and simultaneous inferior
petrosal venous sampling. N Engl J Med 1985;312:100-3.
18. Silber RH, Porter CC. The determination of 17,21-dihydroxy-20-
ketosteroids in urine and plasma. J Biol Chem. 1954;21:262-5.
19. Silber RH, Porter CC. A quantitative color reaction for cortisone
and related 17,21-dihydroxy-20-keto-steroids. J Biol Chem. 1950;
185:923-32.
20. Forman DT. Measurement of urinary hydroxy-corticosteroids by
column chromatography. In: Sunderman FW, ed. Manual of Proce-
dures for the Seminar on Clinical Pathology of Cancer of the En-
docrine Glands and Target Organs. Philadelphia: Institute for Clin-
ical Science for the Association of Clinical Scientists; 1978:135-43.
21. Interpretation of diagnostic data: 5. How to do it with simple maths.
Can Med Assoc J. 1983;129:947-54.
22. Brownlee KA. The correlated two x two table. In: Statistical The-
ory and Methodology in Science and Engineering. New York: John
Wiley & Sons, Inc.; 1965:262-5.
23. Morrison DF. Testing the equality of several means: the analysis of
variance. In: Multivariate Statistical Methods. New York: McGraw
Hill; 1984:34-6.
 24. Swets JA. Measuring the accuracy of diagnostic systems. Science.
1988;240:1285-93.
25. Hankin ME, Theile HM, Steinbeck AW. An evaluation of laboratory
tests for the detection and differential diagnosis of Cushing's syn-
drome. Clin Endocrinol (Oxf). 1977;6:185-96.
26. Martin MM, Hamman BL. Patterns of urinary excretion of steroids
in Cushing's syndrome. J Clin Endocr. 1966;26:257-67.
27. Brown RD, Van Loon GR, Orth DN, Liddle GW. Cushing's disease
with periodic hormonogenesis: one explanation for paradoxical re-
sponse to dexamethasone. J Clin Endocrinol Metab. 1973;36:445-51.
28. Sheeler LR. Cushing's syndrome1988. Clev Clin J Med. 1988;55:
329-37.
29. Cope CL, Black EG. The reliability of some adrenal function tests.
Br Med J. 1959;Nov 28:1117-22.
30. Beisel WR, Cos JJ, Horton R, Chao PY, Forsham PH. Physiology of
urinary Cortisol excretion. J Clin Endocrinol. 1964;24:887-93.
31. Leinung MC, Young WF Jr, Whitaker MD, Scheithauer BW, Trastek
VF, Kvols LK. Diagnosis of corticotropin-producing bronchial car-
cinoid tumors causing Cushing's syndrome. Mayo Clin Proc. 1990;
65:1314-21.

1 February 1992 Annals of Internal Medicine Volume 116 Number 3 217

Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19743/ by a University of California San Diego User on 03/06/2017