| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Service des Maladies Endocriniennes et Métaboliques (M.Y., H.M.-P., X.B.), Laboratoire d’Hormonologie (M.-A.D., Y.F.), and Service de Biostatistique et Informatique Médicale (S.G.), Centre Hospitalier d’Université Cochin, Université Paris 5-Re
e Descartes, 75014 Paris, France
Address all correspondence and requests for reprints to: Dr. Xavier Bertagna, Groupe Hospitalier Cochin, 27 rue du Fg Saint Jacques, 75679 Paris Cedex 14, France. E-mail: xavier.bertagna{at}cch.ap-hop-paris.fr.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Biological diagnosis of Cushing’s syndrome is based on the demonstration of a chronic excess of biologically active circulating cortisol. Different diagnostic approaches have been developed over the years. They all have their limitations for either diagnostic accuracy or the patient’s comfort (3, 4, 5).
Urinary cortisol excretion in a 24-h period is a highly reliable index of cortisol status and is considered the gold standard method for the diagnosis of Cushing’s disease (3, 4, 5, 6). However, it depends on the quality of urine collection.
The classical dexamethasone suppression tests (DSTs), either 2 mg/d for 2 d or 1 mg overnight, are not more accurate (7, 8, 9, 10). They require either blood or urine samples, and individual variations or pharmacological interference may alter the results (4).
The normal circadian rhythm of the hypothalamic-pituitary-adrenal axis is a crucial characteristic that must be taken into account. ACTH and cortisol are secreted in a pulsatile manner with diurnal rhythmicity (11). Plasma ACTH and cortisol levels are highest upon awaking in the morning, low in the late afternoon, and reach a nadir in the late evening. In normal subjects, plasma cortisol levels at 2000 h are approximately 25% of those at 0800 h (12). The loss of a normal diurnal rhythm is the hallmark of Cushing’s syndrome, whatever its cause. In comparison, obese subjects conserve a normal diurnal rhythm of cortisol secretion (13). A single measurement of awake or sleeping midnight plasma cortisol is a highly reliable way to distinguish people with Cushing’s syndrome from healthy people (14, 15, 16). A single sleeping midnight serum cortisol concentration greater than 50 nmol/liter (18 ng/ml) is always indicative of Cushing’s syndrome (100% sensitivity) (15). Measurement of midnight plasma cortisol requires hospitalization, trained medical staff, and painful venipunctures. For these reasons, easier and cheaper ways of assessing circadian cortisol rhythm that would also be suitable for ambulatory screening have been sought.
Salivary cortisol level is a valid indicator of the plasma free cortisol concentration (12, 17, 18, 19, 20). Cortisol circulating in the blood normally forms complexes with proteins. The major binding proteins are corticosteroid-binding globulin (CBG; or transcortin), which binds approximately 90% of cortisol, and albumin, which binds approximately 7% of cortisol (21). Plasma free cortisol diffuses freely through the acinar cells of salivary glands. Binding proteins are absent from saliva. Furthermore, salivary cortisol concentrations are independent of the salivary flow rate (22). There is a strong correlation between plasma free and salivary cortisol concentrations (19). Measurement of salivary cortisol is, therefore, an accurate method to assess plasma free cortisol, the biological active form of circulating cortisol.
Collecting saliva samples is simple (17, 23). It is nonstressful and noninvasive and can easily be performed in an ambulatory setting, far from the hospital or laboratory (24, 25, 26). Saliva samples can be stored in a standard refrigerator or at room temperature and sent by regular mail (27).
Several groups (12, 28, 29, 30, 31, 32, 33) have already demonstrated the clinical relevance of nighttime salivary cortisol concentrations for the diagnosis of Cushing’s syndrome. However, only one study (33) has evaluated midnight salivary cortisol in a large cohort of subjects with simple obesity. Obese subjects are generally suspected to have Cushing’s syndrome, and the difference between the diagnosis of Cushing’s syndrome and obesity is very subtle (1, 2). It is thus quite crucial to develop a diagnostic test that can distinguish between these two conditions. In this study, we specifically studied the diagnostic value of measuring midnight salivary cortisol levels during the initial examination of 63 patients with Cushing’s syndrome of various causes compared with obese subjects. We show that this simple test is as good as, if not better than, the gold standard method (24-h urinary cortisol excretion).
Repeated morning salivary cortisol measurement has already been used to assess adrenal function in long-distance and long-term follow-up of patients (24, 25, 26). Long-term midnight salivary cortisol measurements have never been reported. Thus, we evaluated repeated midnight salivary cortisol measurements in five outpatients, demonstrating the usefulness of this method for the diagnosis of Cushing’s syndrome.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Saliva samples were obtained at midnight from 63 inpatients (55 women and eight men; mean age, 39.2 yr) referred to the Department of Endocrinology, Cochin Hospital, between 1998 and 2002 with suspected Cushing’s syndrome. They were all diagnosed with Cushing’s syndrome according to usual clinical and/or biochemical features [elevated urinary free cortisol (UFC) and lack of suppressibility by low dose DST]. Among these 63 patients, 37 had Cushing’s disease, as confirmed by histological and immunohistological findings after transsphenoidal surgery and/or postsurgical corticotropic insufficiency. Seventeen patients had adrenal Cushing’s syndrome (six with adenoma, four with carcinoma, two with ACTH-independent macronodular adrenal hyperplasia, and five with primary pigmented nodular adrenal disease in the context of Carney’s complex). All of them underwent surgery and had pathological confirmation of the diagnosis. Four patients had proven (pathological findings after surgery) ectopic ACTH secretion, and five patients had suspected (occult) ectopic ACTH secretion. For these latter five patients, dynamic tests were suggestive of ectopic ACTH production, but the source of ACTH secretion was not found by common imaging methods.
Five outpatients had repeated midnight salivary cortisol determinations over several weeks. Patient 1, a 29-yr-old woman, had obvious Cushing’s disease (clinical and biological evidence of hypercortisolism, with petrosal sinus sampling confirming the diagnosis). She provided six samples over a period of 5 wk. Patient 2, a 53-yr-old woman, had subclinical Cushing’s syndrome due to bilateral adrenal adenomas, which had been incidentally discovered (34, 35). Her UFC was normal, but she failed to suppress normally under dexamethasone suppression (1 mg at midnight). She provided 16 samples on 16 consecutive days. Patient 3, a 47-yr-old woman, was suspected of having Cushing’s syndrome on the basis of clinical features. She had two normal and two elevated UFC readings, revealing possible intermittent cortisol overproduction. She provided 15 samples on 15 consecutive days. Patients 4 (a 46-yr-old woman) and 5 (a 38-yr-old woman) had normal cortisol concentrations (normal UFC and sufficient suppression under dexamethasone). Patient 4, who had a pituitary incidentaloma, provided 48 samples over 2 months; patient 5, who had a prolactinoma, provided daily salivary samples over 2 wk.
Control subjects
The control group consisted of 54 inpatients (46 women and eight men; mean age, 45.6 yr) hospitalized in our department for evaluation of and education concerning obesity. These subjects were hospitalized during the same period as the patient group and were examined under the same conditions. All had normal UFC; therefore, Cushing’s syndrome was ruled out.
Biological collections
Both inpatient groups (n = 117) were hospitalized in the Department of Endocrinology. After carefully explaining the procedure to the patients, urine was collected for a 24-h period starting at 0800 h for measurement of urinary cortisol, creatinine, and volume. Saliva samples were obtained at 0800, 1200, 1600, 2000, 2400, and 0400 h during the same 24-h period and used to study circadian variations. The outpatients (n = 5) were instructed to collect saliva samples at midnight. These were obtained daily and were kept in a refrigerator for up to 3 wk before being transferred to the laboratory. Saliva was collected using Salivettes (Sarstedt, Newton, NC), a cotton device that is placed in the mouth for 2 min. Before saliva collection, patients were instructed to rinse their mouth without brushing their teeth (to avoid risk of gingival bleeding). Midnight samples were collected after 5–6 h of fasting (last food intake at 1800 or 1900 h). Inpatients were awakened (20%) or had not yet gone to sleep.
Assays
Urinary and saliva cortisol levels were determined by a CORT-CT2 competition RIA kit (CIS Biointernational, Gif-sur-Yvette, France) as described by the manufacturer with slight modifications.
Urinary cortisol was assayed after extraction of a 500-µl aliquot of patient samples and a 250-µl aliquot of standards with 5 ml dichloromethane. After evaporating to dryness under nitrogen, the residue was dissolved in buffer, then a volume of 100 µl was dispensed into labeled tubes coated with the label. Results are expressed as 24-h urine excretion (micrograms per day). The analytical sensitivity, defined as 2 SD of the zero calibrator, was 2.5 µg/liter (6.9 nmol/liter).
Salivary cortisol was measured using a sensitive procedure that could detect even very low concentrations. Calibrators were prepared by successively diluting a 725 ng/ml standard in buffer [0.1 M Tris-HCl (pH 7.4) and 0.2% BSA] to obtain the following concentrations: 36.3, 18.1, 9.1, 3.6, 0.9, and 0.4 ng/ml. Buffer was used as a blank. Only a small volume (150 µl) of specimen was needed. This protocol had an analytical detection limit of 0.3 ng/ml (0.83 nmol/liter), defined as the minimum detectable concentration equivalent to twice the SD of the zero binding value. Functional sensitivity, defined as the 20% coefficient of variation, was 0.6 ng/ml (1.66 nmol/liter). The between-run precision was approximately 7%. The conversion factor for nanomoles per liter: cortisol (nanomoles per liter) = cortisol (nanograms per milliliter) x 2.76.
Statistical analysis
The Pearson 
2 test was used to compare quantitative variables. The Wilcoxon and Kruskal-Wallis tests were used to compare qualitative variables. All analyses were two-sided, and P < 0.05 was considered significant. The calculations were performed using the SPSS software package version 11.0 (SPSS, Inc., Chicago, IL).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
. In both groups we measured UFC and midnight salivary cortisol under the exactly same conditions (Table 1; mean ± SD). Patients with Cushing’s syndrome had significantly higher values of UFC and midnight salivary cortisol than the controls: 632.8 ± 961.2 µg/d (1746 ± 2652 nmol/d) vs. 22.4 ± 14.3 µg/d (61.8 ± 39.5 nmol/d; P < 0.0001) and 12.3 ± 20.6 ng/ml (33.9 ± 56.9 nmol/liter) vs. 0.8 ± 0.6 ng/ml (2.21 ± 1.66 nmol/liter; P < 0.0001), respectively.
|
shows the individual values. None of the obese subjects, but all patients with Cushing’s syndrome, excreted more than 90 µg UFC/d (248 nmol/d). Three obese subjects and all patients with Cushing’s syndrome had a midnight salivary cortisol above 2.0 ng/ml (5.52 nmol/liter). This cut-off value gives a sensitivity of 100% (confidence interval, 100–100%) and a specificity of 96% (confidence interval, 92–100%). The sensitivities and specificities of midnight salivary cortisol measurements for different cut-off values are presented in Table 2. As shown in Fig. 2, midnight salivary cortisol and UFC obtained on the same day were highly correlated in patients with Cushing’s syndrome (r = 0.90; P < 0.0001).
|
|
|
shows the clinical and biological data for the different etiological subgroups of patients with Cushing’s syndrome.
|
shows that salivary cortisol values were constant throughout the day, emphasizing the lack of circadian variations in patients with Cushing’s syndrome, independent of the cause. As anticipated, patients with ectopic ACTH production had higher UFC and midnight salivary cortisol levels than those with Cushing’s disease or adrenal Cushing’s syndrome (Fig. 3 and Table 3
).
|
) had obvious Cushing’s disease [UFC 663 µg/d (1829 nmol/d)]). Her mean midnight salivary cortisol level (n = 6) was 17.3 ng/ml (47.7 nmol/liter). All values were greater than 2.0 ng/ml (5.52 nmol/liter). Patient 2 (Fig. 4B) had subclinical adrenal Cushing’s syndrome. Her mean midnight salivary cortisol level (n = 16) was 1.55 ng/ml (4.28 nmol/liter). However, 45.8% of the values were above 2.0 ng/ml. Patient 3 (Fig. 4B) was suspected to have Cushing’s disease. Her mean midnight salivary cortisol level (n = 15) was 3.45 ng/ml (9.52 nmol/liter), and 86% of the samples were above the cut-off value. Patient 4 (Fig. 4C) had pituitary incidentaloma. The mean value of her midnight salivary cortisol samples (n = 48) was 1.12 ng/ml (3.09 nmol/liter). Only 8% of them were above 2.0 ng/ml (5.52 nmol/liter), and none exceeded 2.5 ng/ml (6.9 nmol/liter). Patient 5, who presented with prolactinoma (Fig 4C), had a mean salivary cortisol value (n = 14) of 1.33 ng/ml (3.67 nmol/liter). None of her midnight salivary cortisol values exceeded 2.0 ng/ml (5.52 nmol/liter).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
). Indeed, in the control group, two subjects had salivary cortisol levels around the cut-off value [2.05 ng/ml (5.66 nmol/liter) and 2.0 ng/ml (5.52 nmol/liter)], and one patient had a level of 3.6 ng/ml (9.94 nmol/liter). These three patients had normal UFCs and no particular clinical features that would distinguish them from the other control subjects; thus, Cushing’s syndrome was ruled out at the time of the investigation. However, Cushing’s syndrome cannot be definitively excluded, because there are cases of Cushing’s syndrome with intermittent cortisol production. On the other hand, salivary cortisol levels can be falsely elevated by some interfering factors [bleeding (gingivitis), stress of midnight awakening]. Our data, obtained in the second largest series of Cushing’s syndrome patients of various causes, confirm the results of other groups (28, 29, 30, 33), which have demonstrated the diagnostic value of midnight salivary cortisol in adults.
We chose subjects with simple obesity as a control group. The mean BMI of the control subjects (38.3 kg/m2) was significantly higher (P < 0.0001) than that of the Cushing’s syndrome subjects. Obesity is a very frequent, but nonspecific, condition. It is a frequent clinical sign in Cushing’s syndrome. It is thus crucial to identify a biological marker that can distinguish between these two groups. Putignano et al. (33) recently studied midnight salivary cortisol in a large cohort of obese subjects. Our data confirm their finding that midnight salivary cortisol measurement can easily separate Cushing’s syndrome from simple obesity.
Although we did not specifically address the problem of patients with pseudo-Cushing’s syndrome (5), it is possible that midnight salivary cortisol measurement also helps to distinguish them from patients with authentic Cushing’s syndrome (33).
Patients with Cushing’s syndrome typically lose the normal circadian rhythm of cortisol production (11, 12). We show here that this is reflected by the circadian characteristics of salivary cortisol. The mean midnight salivary cortisol concentration in patients with Cushing’s disease was 8.8 ng/ml (24.3 nmol/liter), more than 10 times higher than that in normal subjects. It was even higher in patients with ectopic ACTH syndrome.
We found a strong positive correlation between UFC and midnight salivary cortisol in the Cushing’s syndrome group (r = 0.90; P < 0.0001). The patient with the highest UFC [6769 µg/d (18 682 nmol/d)] also had the highest midnight salivary cortisol levels [166 ng/ml (458 nmol/liter)]. This patient had a very severe form of Cushing’s syndrome caused by a neuroendocrine bronchial tumor (bronchial carcinoid) producing ACTH. Thus, midnight salivary cortisol levels can also predict the severity of the hypercortisolic status.
The convenience of this test makes it suitable for screening large populations, especially high risk populations (patients with obesity, diabetes, and severe osteoporosis), and possibly for the follow-up of patients with adrenal incidentalomas. In comparison, UFC measurement depends on the quality of urine collection.
Initial diagnosis of Cushing’s syndrome may be extremely difficult in patients with mild, fluctuating, or intermittent cortisol overproduction. Clinicians are often obliged to hospitalize patients several times before obtaining unequivocal biological evidence of cortisol excess. In this setting, repeat ambulatory midnight salivary cortisol samples offer a highly effective alternative (patients 1 and 3). Ambulatory collection of salivary cortisol can be useful in the evaluation of the effects of various treatments in patients with Cushing’s syndrome, especially those treated with ortho-para-dichloro-diphenyl-dichloro ethane, as this adrenolytic drug greatly increases plasma CBG levels through its action on the liver (36). Salivary cortisol measurement is also of help in the early stages of pregnancy (37), when estrogen-induced increased CBG levels (38, 39, 40) are accompanied by elevated plasma cortisol levels. In late pregnancy (41) and in women taking the birth control pill (42), salivary cortisol levels are elevated.
Midnight salivary cortisol determination does, however, have some limitations. Cortisol levels reach a nadir 1–2 h after sleep begins (11). However, sleeping patterns are different in different individuals and can vary in the same person. Therefore, cortisol values may not be lowest at midnight. Sometimes we have to evaluate patients who work at night with altered sleep-wake cycles or who have just arrived from distant countries. Another limitation is that the method cannot be applied to patients with gingivitis (risk of bleeding) or Sjögren’s syndrome (insufficient quantity of saliva).
Salivary cortisol measurement requires small quantities of saliva. Commercial kits are now available that require as little as 25 µl (43). It must be stressed, however, that different assays will have different diagnostic cut-offs.
In conclusion, our data show that a single midnight salivary measurement is a highly efficient indicator of a patient’s cortisolic status and distinguishes with high sensitivity and specificity Cushing’s syndrome from simple obesity in hospitalized patients. Additional studies should be undertaken in outpatients. Repeated midnight salivary cortisol measurement over several days or weeks can help to detect initial, fluctuating, or intermittent hypercortisolism.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: CBG, Corticosteroid-binding globulin; DST, dexamethasone suppression test; UFC, urinary free cortisol.
Received October 13, 2003.
Accepted March 29, 2004.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  L. K. Nieman, B. M. K. Biller, J. W. Findling, J. Newell-Price, M. O. Savage, P. M. Stewart, and V. M. Montori The Diagnosis of Cushing's Syndrome: An Endocrine Society Clinical Practice Guideline J. Clin. Endocrinol. Metab., May 1, 2008; 93(5): 1526 - 1540. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Kidambi, H. Raff, and J. W Findling Limitations of nocturnal salivary cortisol and urine free cortisol in the diagnosis of mild Cushing's syndrome Eur. J. Endocrinol., December 1, 2007; 157(6): 725 - 731. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J R Meinardi, B H R Wolffenbuttel, and R P F Dullaart Cyclic Cushing's syndrome: a clinical challenge Eur. J. Endocrinol., September 1, 2007; 157(3): 245 - 254. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. K. Baid, N. Sinaii, M. Wade, D. Rubino, and L. K. Nieman Radioimmunoassay and Tandem Mass Spectrometry Measurement of Bedtime Salivary Cortisol Levels: A Comparison of Assays to Establish Hypercortisolism J. Clin. Endocrinol. Metab., August 1, 2007; 92(8): 3102 - 3107. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F Castinetti, M Nagai, H Dufour, J-M Kuhn, I Morange, P Jaquet, B Conte-Devolx, J Regis, and T Brue Gamma knife radiosurgery is a successful adjunctive treatment in Cushing's disease Eur. J. Endocrinol., January 1, 2007; 156(1): 91 - 98. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. W. Findling and H. Raff Cushing's Syndrome: Important Issues in Diagnosis and Management J. Clin. Endocrinol. Metab., October 1, 2006; 91(10): 3746 - 3753. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Reimondo, B. Allasino, S. Bovio, P. Paccotti, A. Angeli, and M. Terzolo Evaluation of the effectiveness of midnight serum cortisol in the diagnostic procedures for Cushing's syndrome Eur. J. Endocrinol., December 1, 2005; 153(6): 803 - 809. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Viardot, P. Huber, J. J. Puder, H. Zulewski, U. Keller, and B. Muller Reproducibility of Nighttime Salivary Cortisol and Its Use in the Diagnosis of Hypercortisolism Compared with Urinary Free Cortisol and Overnight Dexamethasone Suppression Test J. Clin. Endocrinol. Metab., October 1, 2005; 90(10): 5730 - 5736. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
, Patients with Cushing’s disease (n = 37); 
, patients with adrenal Cushing’s syndrome (n = 17); 
, patients with ectopic ACTH secretion (n = 9). To convert midnight salivary cortisol (nanograms per milliliter) to Systeme International units (nanomoles per liter), multiply by 2.76.