Cent. Eur. J. Med. • 4(1) • 2009 • 59-64
DOI: 10.2478/s11536-009-0004-y
Central European Journal of Medicine
Midnight salivary cortisol, measured by highly
sensitive electrochemiluminescence immunoassay,
for the diagnosis of Cushing’s syndrome
Research Article
Maria Yaneva, Georgi Kirilov, Sabina Zacharieva
Clinical Centre of Endocrinology and Gerontology, Medical University,
1303 Sofia, Bulgaria
Received 17 September 2008; Accepted 20 December 2008
Abstract: T he aim of the present study is to evaluate the measurement of midnight salivary cortisol as a method of screening for Cushing’s
syndrome (CS). Here we tested the performance of a highly sensitive electrochemiluminescence immunoassay (ECLIA) for midnight
salivary cortisol measurement in an extensive clinical study (n=104). Three groups were investigated: 30 patients with CS, 34 with
obesity and 40 healthy normal weight controls. All of them collected saliva samples at 24:00 h and urine samples over the same day
(24 hour period). An electrochemiluminescence immunoassay was used to measure salivary cortisol. Mean midnight salivary cortisol
in healthy volunteers, obese patients and patients with CS was 8.33 ± 3.62, 8.13 ± 4.47 and 33.11 ± 21.68 nmol/l, respectively.
No significant difference was found between midnight salivary cortisol in healthy and obese subjects (P>0.05). In contrast, salivary
cortisol at midnight was significantly higher in patients with CS (P<0.001) as compared to both other groups. The cut-off point of
14.2 nmol/l yielded a sensitivity of 93.3% and a specificity of 94.2% (AUCROC=0.984 ± 0.01(0.965-1.000). A strong positive correlation between midnight salivary cortisol and urinary free cortisol has been found in the CS group (r=0.686, P<0.0001). Our results
demonstrate that measurement of midnight salivary cortisol could be successfully used as a first-line screening method for CS. Our
data approve ECLIA as a simple, reliable and timesaving method for the assessment of salivary cortisol. Automated measurement of
midnight salivary cortisol by ECLIA would facilitate the routine practice in the screening for CS.
Keywords: Midnight salivary cortisol • Cushing’s syndrome • ECLIA
© Versita Warsaw and Springer-Verlag Berlin Heidelberg.
1. Introduction
Measurement of salivary cortisol has recently become
a validated method of appreciation of cortisol secretion.
Several studies [1-8] have demonstrated the usefulness
of midnight salivary cortisol determination for the
diagnosis of Cushing’s syndrome (CS). All these studies
used different radioimmunoassay (RIA) techniques to
measure salivary cortisol. These RIA methods require
several modifications: increasing analyte volume, time of
incubation, etc. They complicate the measurement, make
it longer and suppose more errors in its performance.
Recently, it has been reported [9] that a fully automated
electrochemiluminescence
immunoassay
(ECLIA)
measurement of salivary cortisol is a useful tool in the
assessment of the activity of the hypothalamic-pituitaryadrenal axis. Since ECLIA has not been tested in a large
clinical trial, we decided to use it for the determination of
salivary cortisol. Thus, here we present the first results
on the performance of ECLIA for midnight salivary
cortisol measurement in an extensive clinical study.
Midnight salivary cortisol determination is a test that
could be promising for the screening of CS [1]. The
authors of the Consensus Statement on the diagnosis
and complications of CS [1] recommend larger studies
before the validation of late-night salivary cortisol as
a first-line screening test for CS. Here we show the
results of midnight salivary cortisol measurement in
three groups: patients with CS, obese patients and
normal weight volunteers. Our data demonstrate the
good diagnostic accuracy of evening salivary cortisol
determination for the diagnosis of CS. Something more,
we show no differences between midnight salivary
cortisol levels of normal weight and obese subjects.
* E-mail: [email protected]
59
Unauthenticated
Download Date | 6/19/17 1:42 AM
Midnight salivary cortisol, measured by highly sensitive
electrochemiluminescence immunoassay,
for the diagnosis of Cushing’s syndrome
2. Material and Methods
2.1. Patient population
A total of 104 subjects were studied, divided into three
groups: patients with CS, patients with obesity and normal
weight healthy volunteers. Patients with CS and obesity
were hospitalized in the Clinical Center of Endocrinology
and Gerontology, Sofia, between December 2006 and
December 2007. Thirty patients had CS (5 men and 25
women; mean age ± S.E.M, 39.9 ± 12.7 yr; body mass
index (BMI) 29.5 ± 7.2 kg/m²). The diagnosis of CS
was based on typical clinical and biochemical features
(elevated urinary free cortisol (UFC), lack of circadian
cortisol rhythm and suppressibility with low dose
dexamethasone). In the CS group there were 19 patients
with Cushing’s disease, 10 with adrenal CS and 1 with
ectopic CS. The diagnosis of Cushing’s disease was
based on pathological confirmation after transsphenoidal
surgery and/or postoperative biochemical and clinical
resolution. Patients with adrenal CS were operated on
and had histological confirmation.
The second group (n=34) was comprised of patients
that were hospitalized for evaluation of obesity. There
were 10 men and 24 women (age, 41.1 ± 13.5 yr; BMI,
36.4 ± 4.8 kg/m²). Some of them had clinical features
suggesting hypercortisolism: menstrual irregularities,
essential hypertension, idiopathic hirsutism etc. None
of the obese patients had elevated UFC; they all had a
normal diurnal cortisol rhythm and displayed a sufficient
suppressibility with dexamethasone (plasma cortisol at
08:00h < 50 nmol/l after 1 mg Dexamethasone), so CS
had been excluded.
Forty apparently healthy normal weight volunteers
among the medical staff agreed to participate in the
study (16 men and 24 women; age, 37.2 ± 9.2 yr; BMI,
23.3 ± 2.7 kg/m²). They formed the control group. The
healthy subjects collected saliva and urine samples in
ambulatory conditions.
All the patients and controls provided signed
informed consent to the study, which was approved by
the Local Ethics Committee of the Clinical Centre of
Endocrinology and Gerontology, Sofia.
2.2. Protocol
Saliva samples were collected at 24:00 h before asleep.
All patients and volunteers were instructed to have
dinner and to brush their teeth before 20:00 h, to abstain
from any significant physical activity after 20:00 h and
to rinse the mouth at midnight before saliva collection.
Participants were instructed to abstain from smoking
before the investigation. Saliva samples were collected
in specially designed plastic tubes, Salivettes (Sarstedt
AG & Co) that contain a cotton device, placed for 2-3
min in the mouth. All studied subjects collected urine
samples over the same day (24 hours period).
2.3. Assays
Salivary cortisol was determined by fully automated
highly sensitive competitive electrochemiluminescence
immunoassay using Elecsys Cortisol reagent kit
(Roche). All saliva samples were run in big series in
duplicate. The measuring range of the assay defined
by the lower detection limit (analytical sensitivity) and
the maximum of the standard curve was 1-1750 nmol/l.
The analytical and functional sensitivity of the test
determined by our laboratory did not differ significantly
by those established by the manufacturer (<1 nmol/l and
<8 nmol/l respectively). The dilution test showed linear
results in the lowest concentration range between 1.0
and 8.5 nmol/l. Within-run precision and between-run
precision were 3.2 and 4.7% respectively. The following
cross-reactivities were found: corticosterone – 5.8%,
11-deoxycortisol – 4.1%, 17-α-hydroxyprogesterone
– 1.5%, prednisone – 0.28% and dexamethasone
– 0.08%. Only 50 µl of saliva were required for the
analysis. The comparison of the ECLIA cortisol assay
with a commercial RIA test (Immunotech, Beckman
Coulter Co, France) for the determination of cortisol in
saliva gave the highly significant correlation (r =0.91).
The daily excretion of free cortisol (nmol/24h) was
measured in urine samples by the direct assay procedure
using an RIA cortisol kit manufactured by Immunotech
(Beckman Coulter Co, France). Intra-assay and interassay coefficients of variation were 5.8 and 9.2%,
respectively. The analytical sensitivity of the assay
was 10 nmol/tube. Extremely low cross reactivity was
obtained against other naturally occurring steroids. The
normal values of 24 hours urinary cortisol are between
38 and 270 nmol/24h.
2.4. Statistical analysis
The data were analyzed using SPSS 13.0 software
(SPSS, Chicago, IL, USA). The average values of each
variable were compared using analysis of variance.
Linear regression analysis and Wilcoxon analysis were
used when appropriated. All results are expressed as
mean ± standard deviation (S.D.). Significance was
assumed when P<0.05. The quality of midnight salivary
cortisol and UFC determination in the diagnosis of CS
was tested by receiver operating characteristic (ROC)
analysis and expressed as the area under the curve
(AUC).
60
Unauthenticated
Download Date | 6/19/17 1:42 AM
M. Yaneva, G. Kirilov, S. Zacharieva
Table 1.
Hormonal results of studied groups (n=104).
Healthy (n=40)
Obese (n=34)
CS (n=30)
Midnight salivary cortisol (nmol/l)
8.33 ± 3.62
8.13 ± 4.47
33.11 ± 21.68 *
Urinary free cortisol(nmol/24h)
129.11 ± 72.67
124.25 ± 106.12
773.73 ± 761.67*
Mean ± SD, * p< 0.0001 vs. healthy and obese subjects
Figure 1.
Individual data points for midnight salivary cortisol values
of healthy normal weight volunteers (n=40); obese
subjects (n=34); and patients with Cushing’s syndrome
(n=30). A broken horizontal line represents the cut-off
point value (14.2 nmol/l).
Figure 3. Correlation between midnight salivary cortisol (nmol/l) and
urinary free cortisol (nmol/24h) in patients with Cushing’s
syndrome (saliva and urine samples are collected over the
same day) (r=0.686, P<0.0001).
100
90
30
24 00 h salivary cortisol (nmol/l)
80
nmol/l
20
10
healthy
Figure 2.
obese
CS
Area under the curve the ROC curve for midnight salivary
cortisol determination ( − reference line, - - - midnight
salivary cortisol determination). AUC ± standard deviation
(95% confidence interval) = 0.984 ± 0.1 (0.965 – 1.000).
70
60
50
40
30
20
10
0
0
1000
2000
3000
4000
urinary fre e cortisol (nmol/24h)
3. Results
The three groups were homogenous in age and sex
distribution (P>0.05). The BMI differed significantly in the
three groups (P<0.0001). Hormonal results are presented
in Table 1. No significant difference in midnight salivary
cortisol had been observed between healthy and obese
subjects (P>0.05). By contrast, salivary cortisol at 24:00
h was significantly higher in patients with CS compared
to both other groups (P<0.0001) (Table 1, Figure 1). The
cut-off point for midnight salivary cortisol of 14.2 nmol/l
yielded a sensitivity of 93.3% and a specificity of 94.2%.
The AUC derived from the ROC analysis was optimal
(AUCROC=0.984 ± 0.01(0.965-1.000) (AUCROC ± SD
(interval of confidence of 95%))) (Figure 2). UFC in obese
patients and healthy normal weight volunteers was similar
(P>0.05), while UFC in the CS group was significantly
higher (Table 1). The cut-off point of 270 nmol/24h
for UFC had a sensitivity of 81.5% and a specificity
of 93.2% (AUCROC=0.975 ± 0.014(0.948-1.000)).
The AUCROC of midnight salivary cortisol and UFC
did not differ significantly (P>0.05) and was close to
1, demonstrating the good performance of both tests.
61
Unauthenticated
Download Date | 6/19/17 1:42 AM
Midnight salivary cortisol, measured by highly sensitive
electrochemiluminescence immunoassay,
for the diagnosis of Cushing’s syndrome
We found a statistically significant positive correlation
between UFC and midnight salivary cortisol in patients
with CS (r=0.686, P<0.0001) (Figure 3). Salivary cortisol
at 24:00h and UFC did not show a significant correlation
in healthy and obese subjects.
4. Discussion
Recently, Van Aken et al. [9] have reported a fully
automated electrochemiluminescence immunoassay for
the measurement of salivary cortisol. This method has
several advantages over other radioimmunoassays and
enzyme immunoassays (EIA): it is automated, samples
need no pretreatment, results can be quickly obtained
(within 20 minutes), and collection of a specific number
of samples is not needed for efficient use. Authors
suggest it as a suitable test for daily laboratory and
clinical use.
A detailed review of the literature shows that all big
studies on the use of midnight salivary cortisol for the
diagnosis of CS are based on RIA [2-8]. They all are
modified RIAs for serum cortisol and need modifications
of the analyte volume, time of incubation, etc. This is
time consuming and suggests more errors. Moreover,
performance of RIA requires a specially equipped
laboratory with sophisticated methods of radioactive
protection. Raff et al. [10] have reported a new EIA for
salivary cortisol measurement and compared it to RIA.
In another study, Raff et al. [11] tested another kit for
enzyme immunoassay - EIA Salimetrics.
As we see, there are different available methods of
measurement of salivary cortisol and they all have their
advantages and disadvantages. Since the determination
of midnight salivary cortisol seems to be one of the
simplest approaches in the screening for CS, clinical and
laboratory efforts are focused to find the most precise,
easy and cost-effective method of measurement.
Until now, no clinical studies have been reported on
the use of ECLIA to appreciate cortisol secretion. In this
study we present the first results on ECLIA determination
of midnight salivary cortisol in an extensive clinical study.
Our data show significantly higher midnight
salivary cortisol levels in patients with endogenous
hypercortisolism compared with those of healthy normal
weight volunteers (P<0.001). These data are consistent
with previous studies that reveal the good sensitivity
and specificity of this screening method [2-8]. The cutoff point of 14.2 nmol/l yielded a sensitivity of 93.3%
(two patients with CS have values above this cut-off –
12.27 and 12.44 nmol/l) and a specificity of 94.2% (four
patients with obesity have values above 14.2 nmol/l –
14.43, 14.53, 16.70, 18.08 nmol/l; none of the normal
weight volunteers had midnight salivary cortisol above
14.2 nmol/l) (Figure 1). The AUC derived from the ROC
analysis is close to 1 (Figure 2), demonstrating the
excellent quality of this test. Something more, we didn’t
find any significant difference between the AUCROC of
midnight salivary cortisol and UFC measurment. We can
conclude that both tests have the same good diagnostic
performance.
Our cut-off value (14.2 nmol/l) is similar to that
chosen by Papanicolaou et al. [4] (15.2 nmol/l) and differs
significantly from other reported cut-offs: 3.6 nmol/l [2],
5.52 nmol/l [6] and 7.7 nmol/l [3]. We can conclude that
most probably the cut-off value is laboratory-dependant
and not assay-dependant since Papanicolau et al. used
RIA to determine salivary cortisol [4]. As mentioned in
the consensus statement for the diagnosis of CS, each
laboratory should establish its own values [1]. As the
functional sensitivity of the test is relatively high (<8
nmol/l) we can conclude that ECLIA is better in assessing
hypercortisolism than hypocortisolism.
Our results show no differences between values of
midnight salivary cortisol of normal weight and obese
subjects (P>0.05), thus confirming data that cortisol
rhythm is preserved in obese patients [6,12]. Moreover,
saliva samples of healthy people were collected in
ambulatory settings, while samples of obese subjects
were obtained during hospitalization. As cortisol is a
stress hormone, one could presume that hospitalization,
as a stress factor, would elevate cortisol values. The
results of the present study show that no significant
differences exist between midnight salivary cortisol in
healthy, ambulatory subjects and obese, hospitalized
patients, suggesting that hospitalization should not
be considered as a significant stress factor, causing
misleading results.
We found a statistically significant positive correlation
between UFC and midnight salivary cortisol in patients
with CS (r=0.612, P<0.005). This result confirms
previously reported data from other authors [2,5,6] and
suggests that midnight salivary cortisol could be used
as a reliable alternative of UFC, especially in ambulatory
patients.
Salivary cortisol is strongly correlated with plasma
free cortisol, the biologically active form of circulating
cortisol [13-17]. Moreover, it is independent of the
salivary flow rate [18]. Saliva sample collection is easy,
non-stressful and non-invasive and does not require
trained medical staff. It can be easily performed in
ambulatory conditions. Samples can be stored at room
temperature [19] and sent by regular mail [20,21].
Despite its advantages, salivary cortisol measurement
has its limitations: in cases of dehydratation, Sjögren’s
syndrome, gingivitis, etc.
62
Unauthenticated
Download Date | 6/19/17 1:42 AM
M. Yaneva, G. Kirilov, S. Zacharieva
In conclusion, our results, obtained in a large cohort
of subjects (n=104), support data that midnight salivary
cortisol with its high sensitivity and specificity could be
considered as a method of first-line screening for CS.
Our data, obtained by ECLIA, approve it as a reliable
method of assessment of salivary cortisol. Moreover, it
is simple and timesaving and does not require laboratory
with special radioactive protection. We suggest this
automated method for a larger use in the routine clinical
practice in order to facilitate screening for CS.
Acknowledgements
This work was supported by a grant from the Medical
University of Sofia, Bulgaria.
References
[1] Arnaldi G., Angeli A., Atkinson A.B., Bertagna X.,
Cavagnini F., Chrousos G.P., et al., Diagnosis and
complications of Cushing's syndrome: a consensus
statement, J. Clin. Endocrinol. Metab., 2003, 88,
5593-02
[2] Raff H., Raff J.L., Findling J.W., Late-night salivary
cortisol as a screening test for Cushing’s syndrome,
J. Clin. Endocrinol. Metab., 1998, 83, 2681-86
[3] Castro M., Elias P.C., Quidute A.R., Halah F.P.,
Moreira A.C., Out-patient screening for Cushing’s
syndrome: the sensitivity of combination of circadian
rhythm and overnight dexamethasone suppression
salivary cortisol tests, J. Clin. Endocrinol. Metab.,
1999, 84, 878-82
[4] Papanicolaou D.A., Mullen N., Kyrou I., Nieman
L.K., Nighttime salivary cortisol: A useful test for the
diagnosis of Cushing’s syndrome, J. Clin. Endocrinol.
Metab., 2002, 87, 4515-21
[5] Putignano P., Toja P., Dubini A., Giraldi F.P., Corsello
S.M., Cavagnini F., Midnight salivary cortisol versus
urinary free and midnight serum cortisol as screening
tests for Cushing’s syndrome, J. Clin. Endocrinol.
Metab., 2003, 88, 4153-57
[6] Yaneva M., Mosnier-Pudar H., Dugue M.A., Grabar
S., Fulla Y., Bertagna X., Midnight salivary cortisol
for the initial diagnosis of Cushing's syndrome of
various causes, J. Clin. Endocrinol. Metab., 2004,
89, 3345-51
[7] Gafni R.I., Papanicolau D.A., Nieman L.K., Night
time salivary cortisol measurement as a simple,
noninvasive, outpatient screening test for Cushing’s
syndrome in children and adolescents, J. Pediatr.,
2000, 137, 30-35
[8] Martinelli Jr C.E., Sader S.L., Oliveira E.B., Daneluzzi
J.C., Moreira A.C., Salivary cortisol for screening of
Cushing’s syndrome in children, Clin. Endocrinol.
(Oxf), 1999, 51, 67-71
[9] Van Aken M.O., Romijn J.A., Miltenburg J.A., Lentjes
E.G.W.M., Automated measurement of salivary
cortisol, Clin. Chem., 2003, 49, 1408-9
[10] Raff H., Homar P.J., Burns E.A., Comparison of
two methods for measuring salivary cortisol, Clin.
Chem., 2002, 48, 207-8
[11] Raff H., Homar P.J., Skoner D.P., New enzyme
immunoessay for salivary cortisol, Clin. Chem.,
2003, 49, 203-4
[12] Pirich K, Vierhapper H., 24-hour serum concentration
profile of cortisol in patients with Cushing’s disease,
Exp. Clin. Endocrinol., 1998, 92, 275-9
[13] Laudat M.H., Cerdas S., Fournier C., Guiban
D., Guilhaume B., Luton J.P., Salivary cortisol
measurement: a practical approach to assess
pituitary-adrenal function, J. Clin. Endocrinol.
Metab., 1988, 66, 343-48,
[14] Riad-Fahmy D., Read G., Walker R., Griffiths K.,
Steroids in saliva for assessing endocrine function,
Endocr. Rev., 1982, 3, 367-395
[15] Vining R.F., McGinley R.A., Maksvytis J.J., Ho K.Y.,
Salivary cortisol - a better measure of adrenal
cortical function than serum cortisol, Ann. Clin.
Biochem., 1983, 20, 329-335
[16] Walker R.F., Salivary cortisol determinations in
the assessment of adrenal activity, Front. Oral.
Physiol., 1984, 5, 33-50
[17] Vining R.F., McGinley R.A., The measurement of
hormones in saliva: possibilities and pitfalls, J.
Steroid. Biochem., 1987, 27, 81-94
[18] Vining R.F., McGinley R.A., Transport of steroids
from blood to saliva. In Immunoassays of steroids
in saliva, pp. 56-63. Eds Read GF, Riad-Fahmy
D, Walker RF & Griffiths K. Cardiff: Alpha Omega
Publishing Ltd, 1984
[19] Chen Y.M., Cintron N.M., Whitson P.A., Longterm storage of salivary cortisol samples at room
temperature, Clin. Chem. 1992, 38, 304-305
[20] Mosnier-Pudar H., Thomopoulos P., Bertagna
X., Fournier C., Guiban D., Luton J,P., Longdistance and long-term follow-up of a patient with
intermittent Cushing’s disease by salivary cortisol
measurements, Eur. J. Clin. Endocrinol., 1995,
63
Unauthenticated
Download Date | 6/19/17 1:42 AM
Midnight salivary cortisol, measured by highly sensitive
electrochemiluminescence immunoassay,
for the diagnosis of Cushing’s syndrome
133, 313-316
[21] Hermus A.R., Pieters G.F., Borm G.F., Verhofstad
A.A., Smals A.G., Benraad T.J., et al., Unpredictable
hypersecretion of cortisol in Cushing’s disease:
detection by daily salivary cortisol measurements,
Acta Endocrinol. (Copenh.), 1993, 128, 428-432
64
Unauthenticated
Download Date | 6/19/17 1:42 AM