Salivary steroid assays – research or routine?

Abstract

Salivary concentrations of unconjugated steroids reflect those for free steroids in serum although concentrations may differ because of salivary gland metabolism. Samples for salivary steroid analysis are stable for up to 7 days at room temperature, one month or more at 4°C and three months or more at −20°C. When assessed against strict criteria, the evidence shows that salivary cortisol in evening samples or following dexamethasone suppression provides a reliable and effective screen for Cushing's syndrome. Sequential salivary cortisol measurements are also extremely helpful for the investigation of suspected cyclical Cushing's syndrome. There is potential for the identification of adrenal insufficiency when used with Synacthen stimulation. Salivary 17-hydroxyprogesterone and androstenedione assays are valued as non-invasive tests for the home-monitoring of hydrocortisone replacement therapy in patients with congenital adrenal hyperplasia due to 21-hydroxylase deficiency. The diagnostic value of salivary oestradiol, progesterone, testosterone, dehydroepiandrosterone and aldosterone testing is compromised by rapid fluctuations in salivary concentrations of these steroids. Multiple samples are required to obtain reliable information, and at present the introduction of these assays into routine laboratory testing is not justified.

Introduction

Salivary steroid measurements have the potential to provide a convenient and non-invasive assessment of serum ‘free’ steroid concentrations. They have widespread research applications in the fields of endocrinology, neuroendocrinology, sports and exercise science, and reproductive endocrinology. In spite of a large number of research publications and several reviews,^1–3 salivary steroid methods have not been adopted by the majority of diagnostic laboratories.

This review discusses the factors which influence salivary steroid concentrations and assesses the diagnostic value of steroid measurements against a set of criteria which are appropriate for routine laboratory testing.

Overview – influencing factors

Mechanism of entry of steroids into saliva

Saliva is produced from three main pairs of salivary glands – the parotid, submaxillary and sublingual glands. The parotid glands consist of serous acinar cells which produce a serous, watery secretion, while the submaxillary glands consist of serous and mucous acinar cells and produce a mixed serous and mucous saliva. Sublingual gland saliva is predominantly mucous in character. In addition, 600–1000 minor salivary glands make a small contribution to total salivary production.

Saliva is formed by an active energy-consuming process in which sodium is pumped into the acinar end of the glands. The osmotic pressure difference between serum and saliva causes water to flow through the tight junctions of the acinar cells into saliva, so that the primary secretion into the duct is approximately isotonic with serum. As the saliva travels down the salivary gland duct, sodium is pumped back into the blood, and since little water is reabsorbed, the resulting saliva leaving the duct is hypotonic. The capacity of the ductal cells to pump sodium out of saliva is limited, and so the sodium concentration in secreted saliva is flow rate-dependent.

There are two main mechanisms by which steroids reach the saliva:

Intracellular diffusion

Serum components that are soluble in the lipid-rich cell membrane of the salivary gland acinar cells can pass freely into the cell and diffuse through into the saliva. This mechanism applies to lipid-soluble unconjugated steroids.

Since steroids in serum are protein-bound to a large extent, only the free, non-protein-bound unconjugated steroids will pass into saliva by the diffusion route.

Ultrafiltration

Together with water, non-protein-bound components of serum with a molecular weight of less than approximately 1900 kDa may pass between the tight junctions of the acinar cells. Small polar molecules such as steroid sulphate or glucuronide conjugates are restricted to this ultrafiltration route as a mode of entry into saliva.

As a consequence of these two different routes of entry for steroids into saliva, concentrations of salivary unconjugated steroids such as cortisol, oestradiol and testosterone reflect their free concentrations in serum, whereas conjugated steroids such as dehydroepiandrosterone sulphate (DHEAS) and oestriol sulphate are present in only 1% of their unbound serum concentrations.⁴

The two different mechanisms for entry of unconjugated and conjugated steroids into saliva also have important consequences in terms of the influence of salivary flow rates on salivary steroid concentrations. Unconjugated steroids such as salivary cortisol diffuse freely through the salivary gland acinar cells and show little change at low and high extremes of salivary flow rate.⁴ In contrast, a conjugated steroid such as DHEAS shows a marked fall in concentration with high salivary flow rates because of the rate-limiting ultrafiltration process (Figure 1).

Figure 1

Concentrations of cortisol and dehydroepiandrosterone sulphate in human parotid saliva at low- and high-saliva flow rates. Reproduced from Vining et al.⁴ with permission

Metabolism of steroids by salivary glands

Typical salivary concentrations of steroids are listed in Table 1.

Table 1

Typical salivary steroid concentrations

Salivary steroid	Measured in	Typical salivary concentration	% of Total serum concentration
Aldosterone	Healthy adults	<150 pmol/L	30
Androstenedione	Prepubertal Children	20–250 pmol/L	5
	Adult men	140–630 pmol/L
	Adult women	60–480 pmol/L
Cortisol	Adults 08:00–10:00 h	6–30 nmol/L	5
Cortisol	Adults 22:00 h to midnight	<5 nmol/L	5
Cortisone	Adults 08:00–10:00 h	20–80 nmol/L	approximately 100
Cortisone	Adults 22:00 h to midnight	3–14 nmol/L	approximately 100
Dehydroepiandrosterone	Men	0.3–1 nmol/L	5
Dehydroepiandrosterone	Women (pre-menopause)	0.3–1.7 nmol/L	5
Oestradiol	Women –follicular phase	2–20 pmol/L	1–2
Oestradiol	-mid-cycle	10–30 pmol/L	1–2
17α-hydroxyprogesterone	Prepubertal children	25–220 pmol/L	3–4
	Men	50–360 pmol/L
	Women (luteal phase)	140–320 pmol/L
Progesterone	Women –follicular phase	<160 pmol/L	2
Progesterone	-luteal phase	200–1600 pmol/L	2
Testosterone	Men (am)	250–600 pmol/L	1–2
Testosterone	Women	<200 pmol/L	1–2
5α-dihydrotestosterone	Men	30–150 pmol/L	5
5α-dihydrotestosterone	Women (pre-menopausal)	6–25 pmol/L	5

These are consensus values taken from the literature. They are illustrative values only. Reference ranges will vary with the methods used

The salivary glands have the capacity to metabolize steroids, and so although there is correlation between salivary and serum free steroid concentrations, the concentrations in saliva and serum may differ.

Salivary concentrations of cortisol are only 50–60% of the corresponding serum free cortisol concentration, while salivary cortisone concentrations are higher than those for cortisol. Serum total cortisol:cortisone ratios are approximately 8:1, while the salivary cortisol:cortisone ratio is approximately 1:4 with a range from 1:2 to 1:8. This marked contrast in serum and salivary cortisol:cortisone ratios results from two main effects.

First, cortisol-binding globulin (CBG) in serum binds cortisone less strongly than cortisol, with an association constant 10 times lower for cortisone than for cortisol. Therefore, the proportion of free cortisone in serum is much higher than that for cortisol, and this is reflected in the steroid concentrations reaching saliva.

Secondly, the salivary glands are mineralocorticoid target glands and contain the enzyme 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2).⁵ This enzyme converts cortisol to cortisone irreversibly and protects mineralocorticoid receptors from the influence of cortisol. 11β-HSD2 converts some cortisol to cortisone as it passes through the salivary gland and therefore contributes to the higher concentrations of cortisone in saliva.

In the case of progesterone, salivary concentrations are approximately one-fiftieth of those in serum. The salivary glands, particularly the submandibular glands, can metabolize this steroid to 20α-hydroxy-4-pregnen-3-one.⁶ The extent of progesterone metabolism is minor, and because of the rapid transit of progesterone into saliva from blood it is not considered to be an important influence on salivary progesterone concentrations.

Studies on the metabolism of androstenedione (AD) and testosterone by salivary glands indicate presence of the enzyme 17β-hydroxysteroid dehydrogenase (17βHSD, also known as 17-hydroxysteroid oxido-reductase). In women, relatively high salivary testosterone concentrations compared with serum free concentrations are consistent with conversion of AD to testosterone by salivary glands.^7,8 For men, salivary testosterone concentrations are lower than the corresponding serum free results, which could be explained by salivary gland conversion of testosterone to AD.^6,7

17βHSD in the parotid glands has been shown to favour an oxidative reaction in the conversion of oestradiol to oestrone.⁹ The fact that salivary and serum free oestradiol concentrations are similar suggests that oestradiol metabolism has relatively little influence on salivary oestradiol results.

Correlation of salivary steroid concentrations with serum total or free steroid results

In general, studies comparing salivary and serum total or free steroid concentrations show highly significant correlations with correlation coefficients (r) greater than 0.8 (Table 2). Exceptions are the relatively poor correlations with serum free testosterone for females reported by Broadbent¹⁰ (r = 0.42) and Granger et al.¹¹ (r = 0.36).

Table 2

Systematic survey of correlations between salivary and serum steroid concentrations

Salivary steroid	Correlated with	r value	Reference
Aldosterone	Serum total	0.96	McVie et al.⁶⁷
	Serum total	0.92	Hubl et al.⁹⁹
	Serum total	0.93	Atherden et al.⁶⁸
	Serum free	0.84	Few et al.⁷⁰
	Serum total	0.75	Few et al.⁷⁰
Androstenedione	Serum total	0.92	Baxendale et al.⁶⁶
Androstenedione	Serum total	0.97	Otten et al.⁶⁰
Cortisol	Serum total	0.97	Riad-Fahmy et al.¹⁰⁰
	Serum free	0.97	Vining et al.¹²
	Serum total	0.93	Vining et al.¹²
	Serum total	0.96	Luthold et al.¹⁰¹
	Serum total	0.85	Gallagher et al.¹⁰²
Dehydroepiandrosterone	Serum total	0.87	Granger et al.⁷³
Dehydroepiandrosterone	Serum total	0.86	Ahn et al.⁷⁴
5α-Dihydrotestosterone	Serum free	0.82	Baxendale et al.⁶⁶
17α-Hydroxyprogesterone	Serum total	0.97	Price et al.¹⁰³
	Serum total	0.94	Otten et al.⁶⁰
	Serum total	0.93	Zerah et al.⁵⁷
	Bloodspot total	0.81	Porcu²⁹
Oestradiol	Serum total	0.82	Worthman et al.¹⁰⁴
	Serum total	0.78	Wang et al.¹⁰⁵
	Serum total	0.71	Lu et al.¹⁰⁶
Oestriol	Serum total (unconjugated)	0.97	Vining et al.⁸⁴
Oestriol	Serum total (unconjugated)	0.90	Lachelin and McGarrigle¹⁰⁷
Progesterone	Serum total	0.88	De Boever et al.¹⁰⁸
	Serum total	0.99	Bolaji¹⁰⁹
	Serum total	0.75	Lu et al.¹¹⁰
Testosterone	Serum free	0.79	Baxendale et al.⁶⁶
	Serum free	0.97	Vitek et al.¹¹¹
	Serum total	0.7–0.87	Vitek et al.¹¹¹
	Serum free (women)	0.42	Broadbent¹⁰
	Serum free (men)	0.78	Broadbent¹⁰
	Serum free (women)	0.36	Granger et al.¹¹

For cortisol, there is strong evidence that serum total and free concentrations correlate with salivary cortisol in a predictable manner. The data of Vining et al.¹² clearly show that there is a biphasic response in salivary cortisol when plotted against serum total concentrations (Figure 2). Once the binding capacity of CBG in serum of approximately 600 nmol/L cortisol is exceeded, salivary cortisol concentrations increase more rapidly with increasing total cortisol concentrations, reflecting a disproportionate increase in serum free cortisol. Pregnancy samples, which contain higher CBG concentrations, show a continuation of the initial correlation up to higher serum total cortisol concentrations, as expected. When serum free and salivary cortisol concentrations are compared (Figure 3), a linear relationship is seen over the whole range of serum free cortisol results.¹²

Figure 2

Cortisol concentration in time-matched samples of serum and saliva from normal men, normal women, women on oral contraceptives and pregnant women. Reproduced form Vining et al.¹² with permission

Figure 3

Relationship between salivary cortisol and serum unbound cortisol concentrations in time-matched samples of saliva and serum. Reproduced from Vining et al.¹² with permission

Further evidence of the validity of salivary cortisol measurements is provided by the data of Hofman,¹ who compared average serum cortisol results over 4 h with cortisol concentrations in a 4-h saliva collection using the oral diffusion sink (ODS) device. A correlation coefficient of r = 0.96 was found.

Evidence of good serum vs. saliva correlation for steroid studies, although reassuring, does not necessarily imply that salivary measurements are suitable for routine diagnostic use. For oestradiol, progesterone and testosterone for example, there is evidence that rapid fluctuations in salivary concentrations may detract from their routine use compared with a single serum estimation despite the good overall correlations found (for details, see sections on individual steroids below).

Collection, storage and stability of salivary steroid samples

Saliva collection devices

Whole saliva is a mixture of saliva from the major salivary glands – the parotid, submandibular and sublingual glands – and other minor glands. Other components can include gingival crevicular fluid, serum and blood from oral wounds, bacteria and bacterial products, viruses and fungi, desquamated epithelial cells and food debris.

Parotid saliva can be obtained using a Carlson–Crittenden cup placed over the Stensen's duct. This device, described by Shannon et al.¹³ consists of two small concentric cups, each with a tube attached. Suction is applied to the outer cup to hold the device in place, and saliva is collected through the inner cup. However, this procedure is not convenient for routine diagnostic use.

Most studies of salivary steroids involve collection of mixed saliva obtained by direct spitting into a suitable tube, or by chewing on a wad of absorbent material.

Polyethylene collection tubes should be avoided, since they may adsorb steroids,¹⁴ and the use of polypropylene tubes, which minimize steroid adsorption problems is preferable.

A variety of saliva collection methods have been used. It is important to ensure that any collection method used does not compromise the salivary steroid measurements.

Wade and Haegle¹⁵ described an oral diffusion sink device consisting of a tiny plastic tube containing cyclodextrin to bind analytes. This allowed the collection of saliva over a period of several hours.

Polypropylene ‘Salicaps’ and straws have been developed by one commercial kit supplier (IBL-Hamburg), to give an adsorption of <5% for unconjugated steroids.

‘Salivettes’ (Sarstedt Ltd) offer a convenient means of saliva collection, and consist of an upper chamber with a removable wad which is chewed. The wad is replaced in the upper chamber, and at a later stage the sample can be frozen and thawed, and the tube centrifuged to obtain a clear sample in the conical lower section of the device. Although the use of Salivettes with a cotton wad is acceptable for cortisol, results for dehydroepiandrosterone (DHEA), testosterone, progesterone and oestradiol are spuriously high. This is probably due to the presence of plant sterols in the cotton, which cross-react in steroid immunoassays.¹⁶ These effects are illustrated in Figure 4.

Figure 4

Saliva collection using ‘Salivettes’ with a cotton wad: effects on measured salivary steroid concentrations compared with collection without a wad. Figure derived from the data of Shirtcliff et al.¹⁶ **Denotes results which were significantly greater than the unsystematic error in the assay performance (P < 0.005). NS = not significant

Salivettes with cotton wads should be avoided for steroids other than cortisol. For the collection of other steroids, Salivettes with the cotton wad removed, or with a polypropylene wad¹⁷ can be used, or saliva can be collected into a polypropylene microcentrifuge tube.

The use of a short plastic straw can be helpful for children, and recently the use of mini-marshmallows was very successful for the collection of salivary cortisol samples from toddlers.¹⁸ For very young infants, saliva may be collected with a plastic ‘Pasteur’ pipette.

The use of citric acid or lemon juice to stimulate salivary flow can cause problems with steroid analysis ¹⁹ and should be avoided. Chewing on unflavoured chewing gum has been used for saliva collection but offers no real advantages.

Stability of steroids in saliva

Storage of saliva samples at room temperature for more than 7 d can result in deterioration caused by bacterial growth,²⁰ although this effect is reduced if cellular debris is removed by centrifugation prior to storage.^21,22

Clements and Parker²³ found no difference in salivary cortisol concentrations between aliquots frozen immediately and those subjected to temperature and movement changes to simulate postal delivery over a period of 5 d. Postal delivery of saliva samples to the laboratory is therefore acceptable provided that these are frozen on receipt.

In general, samples for salivary steroid analysis are stable for up to 1 month at +4°C, and for up to 3 months at −20°C. For salivary cortisol, centrifuged saliva samples have been shown to be stable for up to 3 months at +5°C and 1 year at −20°C or −80°C²² and uncentrifuged samples were stable for 9 months at −20°C.²⁴

Freezing of samples is necessary for long-term storage, but in addition, freezing and thawing of saliva prior to analysis is necessary to break up the viscous saliva matrix and to obtain good recovery of steroids.

Salivary steroid concentrations are relatively stable through four or five freeze/thaw cycles. Groschl et al.²¹ for example showed that 17α-hydroxyprogesterone and progesterone decreased by only 5% through five freeze/thaw cycles, while cortisol decreased by 10%. Garde and Hansen²² found that cortisol concentrations in centrifuged saliva samples were unaffected by four freeze/thaw cycles.

Precautions for saliva collection

The effect of vigorous tooth-brushing on salivary cortisol and testosterone has been studied by Kivlighan et al.²⁵ Haemoglobin and transferrin assays confirmed the presence of blood in saliva samples for 30 min after tooth-brushing. Salivary testosterone was increased for 30 min, although the effects on cortisol were minor. Groschl et al.²¹ did not find a significant increase in salivary cortisol, 17α-hydroxyprogesterone (17-OHP) or progesterone 5 min after ‘dental care’.

As a general precaution when collecting saliva for steroid analysis, samples should not be collected within 30 min of tooth-brushing or using dental floss.

In addition, to minimize contamination of saliva, samples should not be collected within 30 min of eating, drinking or chewing ordinary (flavoured) chewing gum.

The influence of blood contamination

Salivary/serum steroid ratios are approximately 1:20 for cortisol and 1:90 for testosterone and oestradiol, and so contamination of saliva with blood can result in spurious increases in the measured steroid concentrations.

Visual inspection of saliva samples will detect a pink colouration at a concentration of 0.1–0.2% ( by volume) contamination with blood. If blood contamination cannot be detected by visual inspection, then the risk of contamination is small for cortisol measurements. For example, at a visible detection limit of 0.2% blood the salivary cortisol result would be overestimated by 4%. For testosterone and oestradiol, the corresponding overestimation would be in the region of 20% because of the higher serum/salivary concentration ratio for these steroids.

Studies that used haemoglobin and transferrin measurements to assess blood contamination in saliva samples from adults²⁵ and children²⁶ have confirmed that interference due to blood contamination is rare for salivary cortisol measurements, but may be of more importance for salivary testosterone and oestradiol testing.

To avoid these problems it is advisable to check saliva samples for any visible pink colouration and to reject any samples which are positive. A dipstick test for haemoglobin such as ‘Hemastix’ will give a positive result at 0.2–0.3% contamination by volume, and can be used to check samples.

Methods for salivary steroid analysis

Salivary steroid assay kits are available from several manufacturers (Table 3). These are usually microtitre plate format assays using enzyme immunoassay or chemiluminescence immunoassay technology.

Table 3

Salivary steroid assay kits

Manufacturer	Supplier website	Steroids
American Laboratory Products Co.	www.alpco.com;www.diagenics.co.uk	Cort, E2, P, 17OHP, T
Cambridge Bioscience	www.bioscience.co.uk	Cort, DHEA, E2, P,
Diametra	www.diametra.com	Cort, AD, E2, 17OHP, P, T
IBL-Hamburg	www.IBL-Hamburg.com	Cort, DHEA, E2, E3, 17-OHP, P, T
IDS Ltd	www.idsltd.com	Cort, E2, 17OHP, P, T
Salimetrics	www.salimetrics.com	Cort, AD, DHEA, E1, E2, E3, 17OHP, P, T

Cort, cortisol; AD, androstenedione; DHEA, dehydroepiandrosterone; E1, oestrone; E2, oestradiol; E3, oestriol; 17OHP, 17α-hydroxyprogesterone; P, progesterone; T, testosterone

The reagent cost for a salivary cortisol assay using commercial kit reagents is approximately £4. The use of salivary cortisol to screen for Cushing's is therefore cost-effective, with costs comparable with that of a urine cortisol, and considerably cheaper than serum cortisol tests with associated costs of venepuncture.

It is also possible to develop in-house assays for salivary steroids. Methods need to be designed to detect the much lower concentration of steroids in saliva. For example, a detection limit of 1 nmol/L or less is desirable for a salivary cortisol assay compared with detection limits of 20–30 nmol/L for serum assays. It is also important to assess antiserum cross-reactivity with a knowledge of the relative concentrations of steroids in saliva. Since there are much higher concentrations of cortisone in saliva, when designing a salivary cortisol assay it is necessary to use an anti-cortisol antibody with low cross-reactivity to cortisone to minimize interference.

In the author's experience, the Dissociation-Enhanced Lanthanide Fluorescence ImmunoAssay (DELFIA) time-resolved fluorescence principle can be used to develop robust and convenient in-house salivary steroid assays. Steroid oxime derivatives can be biotinylated using the reaction described by Dressendorfer et al.²⁷ and the products can be purified by thin-layer chromatography. Streptavidin can be labelled with Europium to provide a common detection system for bound steroid-biotin tracers. This system has been applied successfully to the development of direct assays for salivary cortisol and cortisone,²⁸ and, following ether extraction, to the assay of salivary 17-OHP.²⁹

Liquid chromatography/tandem mass spectrometry (LC-MS/MS) is now being used more frequently for serum and salivary steroid analysis.

Baid et al.³⁰ evaluated radioimmunoassay and LC-MS/MS for the measurement of bedtime salivary cortisol in 261 obese subjects and 60 healthy volunteers. Reference ranges for evening salivary cortisol determined for each method were <2.8 nmol/L for the LC-MS/MS assay and 1.4–4.7 nmol/L for the radioimmunoassay. When results were interpreted relative to the appropriate reference range as a screen for Cushing's syndrome, the two methods showed similar false-positive rates for non-obese normal subjects. However, for the 156 obese subjects who had results for both analyses, significantly more were classified as abnormal by the radioimmunoassay (RIA) (22/156; specificity 86%) than by LC-MS/MS (10/156; specificity 94%).

More studies are needed to confirm these results, which suggest that LC-MS/MS methods may give fewer false-positives when used to screen obese patients.

At present there are no National External Quality Assurance Schemes for salivary steroids, although a free quality assurance scheme for salivary cortisol, oestradiol, progesterone, testosterone and DHEA is provided by IBL-Hamburg. Two distributions, each of three samples are provided per year and mean and range of results returned can be viewed on the IBL website (www.IBL-Hamburg.com).

Routine diagnostic value of salivary steroid testing

Criteria for the assessment of salivary steroid analysis for routine diagnostic use

When examining the value of salivary steroid testing for use in routine clinical biochemistry laboratories it is helpful to establish some strict criteria against which the evidence in the literature can be assessed.

For diagnostic salivary steroid testing it is proposed that the following criteria should be met:

There must be a constant and predictable correlation between salivary and serum free steroid concentrations;

The diagnostic accuracy of salivary tests must be at least equal to that for serum or urine steroid determinations;

Single saliva samples should be just as informative as single serum samples.

Salivary cortisol

There is a considerable body of evidence to support the introduction of salivary cortisol as a routine diagnostic test. Strong correlations between salivary and serum free or total cortisol have been reported (Table 2). Salivary cortisol values correlate in a predictable manner with total serum cortisol, allowing for cortisol-binding by CBG (Figure 2), and show a constant relationship to serum free cortisol concentrations (Figure 3).

Dorn et al.³¹ compared salivary and serum cortisol circadian rhythms by collecting hourly samples over 24 h. They concluded that the two profiles were synchronous, and that salivary cortisol could be substituted for serum cortisol for the assessment of circulating cortisol rhythmicity.

The majority of publications on the diagnostic use of salivary cortisol relate to screening for Cushing's syndrome using evening samples and/or the low-dose dexamethasone suppression test, with fewer reports on the investigation of adrenal insufficiency and the monitoring of hydrocortisone replacement therapy.

Screening for Cushing's syndrome

Currently, the three main screening tests for Cushing's syndrome are the assessment of diurnal rhythm in serum cortisol, morning serum cortisol following the low-dose overnight dexamethasone suppression test and the measurement of 24-h urine cortisol. All these tests have limitations; midnight serum cortisol measurements are only feasible for hospital inpatients, and 24-h urine collections are often unreliable. Some patients with mild Cushing's syndrome may suppress normally to the low-dose dexamethasone suppression test.³² In addition, the stress of hospital admission may cause abnormal adrenal axis test results for the first few days.³³

Salivary samples can be collected in a relatively stress-free environment at home, and samples can be posted to the laboratory for analysis. They are particularly useful to document cyclical or intermittent changes in cortisol secretion over long-time periods where the collection of multiple urine samples is impractical.

There are now more than 10 published studies in adults and children that confirm that evening salivary cortisol can be used as a reliable screen for Cushing's syndrome. Although the time chosen to assess evening concentrations varies from 22:00 h to 24:00 h, the evidence suggests that the diagnostic value at these times is the same. Where figures for sensitivity and specificity have been quoted, these are generally in the range of 90–100% (Table 4) and compare favourably with those found for 24-h urine cortisol measurements^34,35 and for serum cortisol assay following low-dose dexamethasone suppression.³⁴

Table 4

Salivary cortisol as a screen for Cushing's syndrome: a systematic survey of the literature

Reference	Sensitivity/Specificity (%)
Reference	Evening saliva	LDDST	Combined
Barrou et al.¹¹²	–	100/94	–
Raff et al.¹¹³	92/95	–	–
Castro et al.¹¹⁴	93/93	91.4/94.4	100/93.3
Martinelli et al.¹¹⁵	100/95.2	100/95.2	100/100
Papanicolaou et al.³⁸	93/100	–	–
Putignano et al.³⁹	92.7/93.1	–	–
Yaneva et al.¹¹⁶	100/96	–	–
Trilck et al.¹¹⁷	100/97.6	–	–
Viardot et al.³⁴	100/100	100/77.1	–

LDDST, low-dose dexamethasone suppression test

Salivary cortisol samples collected after low-dose dexamethasone suppression also show high sensitivity and specificity as a screen for Cushing's syndrome, and when used in combination with an evening salivary cortisol test, a sensitivity of 100% can be achieved (Table 4).

The day-to-day reproducibility of salivary cortisol measurements has been assessed by Viardot et al.³⁴ and was found to be high with 78% of the total variation explained by inter-subject variation and 22% by intra-subject variation.

Salivary cortisol measurements are becoming more widely used as a screen for Cushing's syndrome, and recently they have been recommended as a first-line screening test by Findling and Raff.³⁶ These authors recommend a 23:00 h salivary cortisol as an initial screen, using a cut-off of 4.3 nmol/L for normality. Any results over 8.6 nmol/L were considered to be consistent with Cushing's syndrome, while results between 4.3 nmol/L and 8.6 nmol/L generated a repeat 23:00 h salivary cortisol sample.

It is important that each laboratory establishes its own cut-off values for evening salivary cortisol, particularly when in-house or modified commercial methods are used, since results vary with different immunoassay methods. Typical cut-off values reported for normal evening salivary cortisol measured by immunoassay are in the range of 4–6 nmol/L. Reference ranges using chromatography and tandem mass spectrometry are often lower than those obtained by immunoassay methods.

In the author's laboratory, the screening procedure for Cushing's syndrome involves the collection of 08:00–10:00 h and 22:00 h to midnight saliva samples over 3 d. Reference ranges are 6–30 nmol/L at 08:00–10:00 h and <5 nmol/L at 22:00 h to midnight, and so evening results over 5 nmol/L are abnormal. This approach has been valuable in screening for Cushing's syndrome in women with polycystic ovaries and hyperandrogenism.³⁷

As with all biochemical testing for Cushing's syndrome, there are limitations to the evening salivary cortisol test. It is important to establish that patients have a normal sleep/wake cycle when interpreting results, and that samples were taken at the right time if collected at home. Patients with ‘pseudo-Cushing's’ due to alcohol abuse, type 2 diabetes or major depression will give positive results.^38,39

During pregnancy, salivary cortisol concentrations rise in the second and third trimesters, showing a 1.5–2-fold elevation, and fall back to normal within 5–7 d of parturition. A ‘lag’ of approximately 90 min in the salivary cortisol circadian rhythm in pregnancy has also been documented.⁴⁰

Mild cases of Cushing's syndrome may show a normal evening salivary cortisol when tested on a single occasion. Kidambi et al.⁴¹ studied 11 cases of Cushing's syndrome with normal or mildly elevated urine cortisol. Although all patients had some elevated evening salivary cortisols, results were normal on some occasions, indicating that several saliva and urine samples and a dexamethasone suppression test are necessary for the thorough investigation of mild Cushing's syndrome.

Despite these limitations, the evidence suggests that for screening Cushing's syndrome, a single saliva sample taken in the evening or following low-dose dexamethasone suppression can be just as informative as a single blood sample.

Cyclical Cushing's syndrome

Cyclical or intermittent Cushing's syndrome may be more common than was thought. Evidence for this was provided by Atkinson et al.⁴² who studied nine of 14 consecutive patients diagnosed with Cushing's syndrome using multiple urine cortisol collections. They found that five of the nine patients studied showed cyclical variation in cortisol secretion, with a further two patients showing an erratic pattern without any obvious cyclicity.

The periodicity of cyclical Cushing's disease has been reported to vary between 12 h and 85 d. This makes diagnosis using conventional urine collections very difficult, particularly since ideally it is necessary to identify at least three peaks and two troughs in cortisol secretion.⁴³

Salivary testing is an excellent screening test for cyclical Cushing's syndrome since it is non-invasive, convenient and can be undertaken over many days if necessary. For example, Laudat et al.⁴⁴ measured 20:00 h salivary cortisol twice-weekly over a period of 60 d to establish Cushing's syndrome in a patient who showed a cyclicity of 21 d.

Salivary cortisol measured at Southampton in samples collected at 09:00 h and 22:00 h over 19 consecutive days at King's College Hospital were helpful in identifying food-dependent Cushing's syndrome in a patient whose evening saliva cortisols were often higher than those at 08:00 h (unpublished data).

Differential diagnosis of Cushing's disease

Castro et al.⁴⁵ have studied salivary cortisol suppression to different doses of dexamethasone and the responses to ovine-corticotrophin releasing hormone (o-CRH) stimulation in 46 patients with Cushing's syndrome, including 28 patients with Cushing's disease. Salivary cortisol showed a higher percentage of suppression to dexamethasone than serum cortisol, with suppression of 65% or more for salivary cortisol corresponding to a suppression of serum cortisol of 50% or more for the 8 mg test. Sensitivities and specificities using salivary cortisol for the diagnosis of Cushing's disease were 88% and 100% for the 8 mg dexamethasone test and 93% and 91% for the o-CRH test (using a criterion of an increment >20%).

Salivary cortisol measurements may therefore improve the discrimination of the high-dose dexamethasone suppression test for the differential diagnosis of Cushing's disease. The evidence provided by future studies would be useful to confirm these findings.

Testing for adrenal insufficiency

There are fewer studies on the application of salivary cortisol testing to the investigation of adrenal insufficiency, but there is some evidence that morning salivary cortisol may be helpful in checking for adrenal suppression, and that salivary cortisol responses to Synacthen may be a useful test of adrenal function.

Patel et al.⁴⁶ compared morning salivary cortisol results with serum cortisol concentrations from the conventional short Synacthen test (SST) to check for adrenal suppression in 48 patients who were using topical steroids. Of the 48 patients, 16 had an impaired morning salivary cortisol result, and of these, 15 had an impaired response to the SST. All patients with a normal morning salivary cortisol had a normal SST response. The authors concluded that morning salivary cortisol was efficient and cost-effective for outpatient screening of adrenal suppression due to glucocorticoid use.

Salivary cortisol responses to Synacthen have been evaluated for the conventional dose of 250 μg and also for reduced doses of 25 μg or 1 μg Synacthen.

Suri et al.⁴⁷ used a 250 μg Synacthen test to assess adrenal reserve in pregnancy, and established reference ranges for the higher salivary cortisol responses seen in the second and third trimesters of pregnancy compared with postpartum results.

Basal and post-Synacthen (250 μg) serum total and free cortisol and salivary cortisol were measured by Arafah et al.⁴⁸ in 51 critically ill patients and normal subjects. They found that serum free and salivary cortisol results correlated well and were uninfluenced by large differences in serum albumin concentrations. Salivary cortisol could be used as a surrogate marker for serum free cortisol for these investigations.

Contreras et al.⁴⁹ found that a 25 μg Synacthen test gave equivalent serum and salivary cortisol responses to the 250 μg test at 30 min, with normal responses of >20 nmol/L and >100 pmol/L for salivary cortisol and aldosterone, respectively. The same research group went on to apply the 25 μg test with salivary samples to check for adrenal insufficiency in patients with HIV⁵⁰ and patients with end-stage renal failure.⁵¹

A 1-μg low-dose Synacthen test was used by Marcus-Perlman et al.⁵² to assess responses in 14 normal subjects and 14 women who were ‘hyperoestrogenic’ (pregnancy, oral contraceptive or hormone replacement therapy). A normal salivary cortisol response of >27.6 nmol/L at 30 min was established, and the test was found to provide a useful alternative to the blood test, particularly when increased CBG concentrations complicated the interpretation of changes in serum total cortisol results.

The use of salivary sampling with the SST therefore has potential, but more studies are needed to establish its value in routine diagnostic testing.

Monitoring of glucocorticoid replacement therapy

The evidence on the value of salivary cortisol measurements for the assessment of glucocorticoid replacement therapy is conflicting.

Measurement of a daytime serum cortisol profile has been considered to be the best way of assessing replacement therapy, but this requires hospital admission.

Two groups have concluded that salivary daytime profiles are not useful for monitoring oral hydrocortisone, largely based on their observations of poor correlations between serum total and salivary cortisol results.^53,54

In contrast to these reports, Lovas et al.⁵⁵ found excellent correlation between salivary cortisol and serum cortisol profiles after oral cortisone acetate or intravenous hydrocortisone dosing. The same research group then used salivary cortisol measurements in combination with continuous subcutaneous hydrocortisone infusion in patients with Addison's disease to establish a more physiological cortisol rhythm.⁵⁶

The finding of poor overall correlation between serum total cortisol and salivary cortisol in combined results for pre- and postdose samples could be explained by the disproportionate increase in serum free and salivary cortisol compared with serum total cortisol, once the binding capacity of CBG of approximately 600 nmol/L cortisol is exceeded. This happens frequently after oral hydrocortisone administration in standard doses. An example of increases found in our laboratory (unpublished data) following a 10 mg dose of hydrocortisone in the morning are from 1 nmol/L to 63.5 nmol/L for salivary cortisol (a 64-fold increase) and from 49 nmol/L to 734 nmol/L for serum total cortisol (a 15-fold increase). Salivary samples were collected at least 30 min after taking the oral hydrocortisone dose.

There is no reason why the serum total cortisol profile should be taken as the ‘gold standard’, and it is possible that salivary cortisol, which reflects the biologically active serum free cortisol concentration, may provide a more useful measurement of replacement therapy provided that problems of saliva contamination by the hydrocortisone tablets are avoided.

Salivary 17α-hydroxyprogesterone

Salivary 17-OHP measurements have been applied to screening for classical and non-classical congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency, and to the monitoring of hydrocortisone replacement therapy in CAH patients.

Zerah et al.⁵⁷ measured serum and salivary 17-OHP in 57 normal subjects and 15 newly diagnosed patients with non-classical 21-hydroxylase deficiency. They found good correlation between serum and salivary results (r = 0.93) and morning salivary 17-OHP concentrations were unequivocally higher in all patients compared with normal subjects. The salivary test was found to be convenient and effective, and was used in a second study to establish a prevalence of non-classical 21-hydroxylase deficiency of 1.4% in a Caucasian population.^58,59

Studies on the use of salivary 17-OHP assays to monitor patients on hydrocortisone replacement therapy have all found the procedure to be convenient and helpful, although there are large differences in reference ranges found for normal subjects and for patients in good control.^29,60–64 These discrepancies may be due to differences in antiserum specificity and to the fact that some methods do not employ a preliminary solvent extraction stage prior to immunoassay.

At Southampton, UK, we have developed an in-house time-resolved fluorescence assay for salivary 17-OHP using ether extraction followed by immunoassay.²⁹ The method proved to be a convenient means of assessing control in children with 21-hydroxylase deficiency. Results showed a good correlation with those for bloodspot 17-OHP samples taken at the same time (r = 0.81) with good day-to-day reproducibility. Morning salivary 17-OHP concentrations of 25–217 pmol/L found for normal children agreed well with the range of 6–225 pmol/L reported by Groschl et al.⁶⁴ although other groups have reported higher morning reference ranges.^60,61,63

Children with CAH on replacement therapy, classified as having good control, had morning salivary 17-OHP concentrations in the range of 390–3090 pmol/L, falling to <1500 pmol/L later in the day. These results agree with those of Otten et al.⁶⁰ who found a morning range of 340–3060 pmol/L in patients with good control. However, a lower target range of <1000 pmol/L was reported by Young et al.⁶²

Method-related differences in salivary 17-OHP results make it essential for a laboratory to establish reference ranges and target ranges for treated patients with the particular method used.

The evidence suggests that salivary 17-OHP measurements have a place as a convenient and non-invasive means of monitoring control in CAH patients while at home.

Salivary androstenedione

Salivary AD measurements have been evaluated for the monitoring of replacement therapy in CAH patients by Otten et al.⁶⁰ and Young et al.⁶⁵ Both groups found the test to be useful for the assessment of control. In contrast to salivary 17-OHP concentrations, which are elevated above normal in CAH patients in good control, salivary AD results were within the reference range. For example, Otten et al. found a range of 40–50 pmol/L for salivary AD in 09:00 h samples from CAH children in good control, values which were well within the 09:00 h reference range of 20–250 pmol/L for normal prepubertal children.

Home-monitoring of CAH patients using salivary AD may serve as a useful adjunct to salivary 17-OHP measurements for the assessment of control.

Salivary AD and dihydrotestosterone concentrations in women with hyperandrogenism were studied by Baxendale et al.⁶⁶ Salivary concentrations of both steroids were shown to reflect serum free concentrations in normal, hyperandrogenic and cyproterone acetate- or ethinyl oestradiol-treated patients.

Evidence for a more rapid conversion of AD to testosterone (T) in hirsute women was found by Swinkels et al.⁸ who demonstrated lower salivary AD:T ratios in these patients.

Further studies are needed to establish if salivary AD assays have a role in the investigation of hirsutism.

Salivary aldosterone

Assay methods for salivary aldosterone were first reported more than 20 y ago.^67–69

Salivary aldosterone concentrations were found to be similar to those for serum free aldosterone, and represented approximately 30% of the corresponding serum total aldosterone concentration.^68,70

Good correlation between salivary and serum free aldosterone has been reported (r = 0.84), with a weaker correlation between saliva and serum total aldosterone (r = 0.75).⁷⁰ Salivary aldosterone was found to correlate with changes in serum total aldosterone in response to ACTH stimulation or dexamethasone suppression.^67–69

Salivary/serum aldosterone ratios are unchanged during the first, second and third trimesters of pregnancy, and so salivary aldosterone had potential as a non-invasive assessment of aldosterone status during gestation.⁷¹

When sampled at frequent intervals, however, salivary aldosterone showed considerable fluctuations during the day, indicating the need for multiple samples to obtain reliable information.⁷²

At present there is no information on the diagnostic value of salivary aldosterone measurements for the investigation of suspected primary hyperaldosteronism.

Salivary dehydroepiandrosterone

Salivary DHEA concentrations show good correlation with serum, and decreasing values with increasing age in adults have been well-documented.^73,74

Despite a large number of publications in the literature on salivary DHEA measurement in the areas of DHEA supplementation,⁷⁵ neuroendocrinology,⁷⁶ and sports medicine,⁷⁷ no routine diagnostic applications have been described. Salivary DHEA results in an individual show pulsatility, with large fluctuations during the day superimposed on a circadian rhythm similar to that for cortisol. Single saliva samples therefore have very little value, and repeated sampling is necessary to obtain consistent results.

Salivary oestradiol

Salivary E2 measurements have been used extensively in research studies on the menstrual cycle.^78–80 Although they can be used to classify cycles, multiple samples are necessary to obtain reliable information. The main problem is that salivary E2 concentrations measured in individual women show a pulsatile pattern with cycles of 60–90 min; so sampling at multiple time points is necessary.⁸¹

Tivis et al.⁸² compared serum and salivary E2 values in 31 postmenopausal women on oral oestrogen therapy (‘Premarin’ tablets) with 12 patients who were not receiving oestrogen replacement. Their data showed that salivary E2 reflected serum E2 concentrations in women on oestrogen replacement (r = 0.81), but was very poorly correlated with serum E2 for women not receiving oestrogens (r = 0.32).

The need for multiple samples makes salivary E2 measurements impractical for routine assessment of oestrogen status.

Salivary unconjugated oestriol

Salivary E3 concentrations in the third trimester of pregnancy showed episodic fluctuations during the daytime, together with a circadian rhythm with concentrations increasing from 22:00 h, peaking at 04:00 h and returning to daytime concentrations at 06:00–07:00 h.⁸³

Early publications demonstrated a correlation between salivary and serum unconjugated E3 concentrations, indicating their potential as an alternative to serum concentrations for the monitoring of feto-placental function.^78,84,85

More recent research studies have assessed salivary E3 as an indicator of premature labour. A surge in salivary E3 concentration occurs 2–3 weeks before the onset of labour. Concentrations over 2.1 ng/mL (7.3 nmol/L) identified those pregnancies at risk of premature labour, while values less than this gave a confident prediction that delivery would not occur within 2–3 weeks.^86,87

In spite of these research findings, salivary E3 assays do not have a routine application in the diagnostic laboratory.

Salivary progesterone

Multiple sequential samples for salivary progesterone have been used in a wide variety of applications as a non-invasive alternative to serum measurements.^88,89 They have been particularly useful in research studies on girls during and after menarche, where there are ethical objections to blood sampling.^90,91

When used as an index of ovulation, considerable discrepancy has been observed in classification between follicular and luteal phases,⁹² and relatively poor sensitivity (78%) and specificity (76.5%) for midluteal prediction of ovulation has been reported.⁹³

In spite of the large number of publications on sequential measurements in the study of the menstrual cycle, salivary progesterone assays have not been adopted by diagnostic laboratories to check for ovulation or to assess luteal function. The main problem is that salivary P results show much greater variation over a 24-h period than that shown by serum concentrations. In a detailed study, Delfs et al.⁹⁴ looked at 24-h salivary and serum P profiles and concluded that a single serum P correlated better with the 24-h mean serum P concentration than either a single saliva result or the mean of three saliva samples in the same individual.

One major manufacturer of salivary progesterone kits (IBL-Hamburg) comments in their literature on salivary P (and also on E2 and testosterone), that ‘single samples in most cases will give arbitrary results which are difficult to interpret’. They recommend taking five samples over a period of 2 h, and then analysing a pool made from these five samples, thus saving analytical costs.

It can be appreciated therefore, why routine diagnostic laboratories continue to measure serum P when faced with this rather cumbersome strategy for salivary P collection methods.

A single salivary P estimation is not as informative as a single serum P measurement, and therefore salivary P fails to meet the criteria for a routine diagnostic test.

Salivary testosterone

Salivary testosterone measurements have been applied to the investigation of male hypogonadism and of hirsutism or premature ovarian failure in women.

Salivary T in individuals can show rapid fluctuation even though smooth graphs of the changes in male salivary testosterone with age and with diurnal variation can be constructed from grouped data. This limits the value of a single determination, and therefore multiple samples are needed to obtain useful data.

Morley et al.⁹⁵ established reference ranges for salivary testosterone in 127 men aged 20–89 y, and showed a 47% reduction over this age span. When salivary testosterone was compared with serum total, ‘bioavailable’ or calculated free testosterone, strong correlations were found, and in addition, salivary testosterone correlated with the results of two questionnaires assessing hypogonadal symptoms. However, the authors concluded that ‘salivary testosterone is not a better assay than other measures to diagnose hypogonadism’.

Although reasonable correlations between salivary and serum testosterone have been reported for combined data, when results for men and women are analysed separately, poorer correlations have been found, particularly for women.^10,11 The use of salivary testosterone to assess hyperandrogenism in hirsute women is limited by a wide overlap between male and female reference ranges (Drs Teoh and Wallace, Glasgow Royal Infirmary, personal communication), and a single salivary testosterone result has been reported to give poorer discrimination than the conventional serum free androgen index.⁹⁶

Salivary testosterone measurements may have some useful applications however.

In women with premature ovarian failure, salivary testosterone concentrations have been found to be consistently low or undetectable, while the serum free androgen index was within the reference range because of a reduction in both testosterone and sex hormone-binding globulin (SHBG) concentrations.⁹⁷

A study of lumbar bone mass in healthy premenopausal women showed that high salivary testosterone concentrations were associated with higher lumbar bone mass, despite the fact that serum testosterone results were all normal.⁹⁸

At present, the use of salivary testosterone measurements in the routine diagnostic laboratory cannot be recommended.

Conclusions and recommendations

There is good evidence to support the introduction of salivary cortisol as a screening test for Cushing's syndrome. Evening samples, or samples taken after low-dose dexamethasone testing have a sensitivity and specificity, which is as good as or better than conventional testing of serum cortisol following low-dose dexamethasone suppression or 24-h urine cortisol measurement. Salivary cortisol sampling is the method of choice for long-term assessment of patients with suspected cyclical Cushing's syndrome, and has potential for the investigation of possible hypoadrenalism with the Synacthen test.

Laboratories should consider the introduction of salivary cortisol for the routine investigation of patients for Cushing's syndrome.

Salivary 17-OHP and AD results are of value for the assessment of control in patients treated with hydrocortisone for CAH due to 21-hydroxylase deficiency. These tests should be considered by specialized endocrine laboratories to provide a home-monitoring service.

Salivary assays for oestradiol, progesterone, testosterone, DHEA and aldosterone are compromised by rapid fluctuations in concentrations, making collection of multiple samples necessary. They do not meet the criteria for routine diagnostic tests and their introduction into the laboratory repertoire cannot be justified at present.

Footnotes

ACKNOWLEDGEMENT

I am grateful to Mike Wallace and Brian Keevil for helpful comments and information during the writing of this review.

References

Hofman

. Human saliva as a diagnostic specimen. J Nutr 2001;135:1621S–25S

Kaufman

, Lamster

. The diagnostic applications of saliva – a review. Crit Rev Oral Biol Med 2002;13:197–212

Lewis

. Steroid analysis in saliva: an overview. Clin Biochem Rev 2006;27:139–46

Vining

, McGinley

, Symons

. Hormones in saliva: mode of entry and consequent implications for clinical interpretation. Clin Chem 1983;29:1752–6

Shimojo

, Ricketts

, Petrelli

, Immunodetection of 11 beta-hydroxysteroid dehydrogenase type 2 in human mineralocorticoid target tissues: evidence for nuclear location. Endocrinology 1997;138:1305–11

Blom

, Ojanotko-Harri

, Laine

, Huhtaniemi

. Metabolism of progesterone and testosterone in human parotid and submandibular salivary glands in vitro . J Steroid Biochem Mol Biol 1993;44:69–76

Rey

, Chiodori

, Gomez

, Febber

. Interpretation of the discrepancy observed between plasma free and salivary testosterone levels in man. Steroids 1988;52:371–2

Swinkels

, van Hoof

, Rose

, Smale

, Benraad

. Low ratio of androstenedione to testosterone in plasma and saliva of hirsute women. Clin Chem 1992;38:12819–23

Djoseland

, du Bezche

, Hoglo

, Rennie

. Steroid metabolism by normal and neoplastic parotid tissue. J Steroid Biochem 1982;16:397–402

10.

Broadbent

. Salivary testosterone: is it an indicator of free testosterone concentration in plasma? NZJ Med Lab Sci 2002;56:85–9

11.

Granger

, Shirtcliff

, Booth

, Kivlighan

, Schwartz

. The ‘trouble’ with salivary testosterone. Psychoneuroendocrinology 2004;29:1229–40

12.

Vining

, McGinley

, Maksvytis

, Ho

. Salivary cortisol: a better measure of adrenal cortisol function than serum cortisol. Ann Clin Biochem 1983;20:329–36

13.

Shannon

, Prigmore

, Chauncey

. Modified Carlson-Crittenden device for the collection of parotid fluid. J Dent Res 1962;41:778–83

14.

Banerjee

, Levitz

, Rosenberg

. On the stability of salivary progesterone under various conditions of storage. Steroids 1985;46:967–74

15.

Wade

, Haegle

. Differential measurement of cortisol and cortisone in human saliva by HPLC with UV detection. J Liquid Chromatogr 1991;14:1813–27

16.

Shirtcliff

, Granger

, Schwartz

, Curran

. Use of salivary biomarkers in biobehavioral research: cotton based sample collection methods can interfere with salivary immunoassay results. Psychoneuroendocrinology 2001;26:165–73

17.

Groschl

, Rauh

. Influence of commercial collection devices for saliva on the reliability of salivary steroid analysis. Steroids 2006;71:1097–100

18.

Clements

, Parker

Jr , Dixon

Jr , Salley

. Marshmallows used as saliva stimulant do not affect cortisol concentrations: finally a palatable alternative for toddler saliva collection. Dev Psychobiol 2007;49:702–7

19.

Vialard-Miguel

, Belaidi

, Lembeye

, Corcuff

J-B

. Lemon juice alters cortisol assays in saliva. Clin Endocrinol 2005;63:478–9

20.

Whembolua

, Granger

, Singer

, Kivlighan

, Marguin

. Bacteria in the oral mucosa and its effects on the measurement of cortisol, dehydroepiandrosterone and testosterone in saliva. Horm Behav 2006;49:478–83

21.

Groschl

, Wagner

, Rauh

, Dorr

. Stability of salivary steroids: the influences of storage, food and dental care. Steroids 2001;66:737–41

22.

Garde

, Hansen

. Long term stability of salivary cortisol. Scand J Clin Lab Invest 2005;65:433–6

23.

Clements

, Parker

. The relationship between salivary cortisol concentrations in frozen versus mailed samples. Psychoneuroendocrinology 1998;23:613–16

24.

Aardal

, Holm

. Cortisol in saliva – reference ranges and relation to cortisol in serum. Eur J Clin Chem Clin Biochem 1995;33:927–32

25.

Kivlighan

, Granger

, Schwartz

, Nelson

, Curran

, Shirtcliff

. Quantifying blood leakage into the oral mucosa and its effects on the measurement of cortisol, dehydroepiandrosterone and testosterone in saliva. Horm Behav 2004;46:39–46

26.

Granger

, Cicchetti

, Rogosch

, Hibel

, Jeisl

, Flores

. Blood contamination in children's saliva: prevalence, stability, and impact on the measurement of salivary cortisol, testosterone and dehydroepiandrosterone. Psychoneuroendocrinology 2007;32:724–33

27.

Dressendorfer

, Kirschbaum

, Rohde

, Stahl

, Strasburger

. Synthesis of a cortisol-biotin conjugate and evaluation as a tracer in an immunoassay for salivary cortisol measurement. J Steroid Biochem Mol Biol 1992;43:683–92

28.

Wood

, Kilpatrick

, Barnard

. New direct salivary cortisol and cortisone assays using the ‘DELFIA’ system (Abstract). J Endocrinol 1997;155(Suppl. 2):71

29.

Porcu

. Salivary 17-Hydroxyprogesterone for the Monitoring of Steroid Replacement Therapy in Patients with Congenital Adrenal Hyperplasia. MSc dissertation. University of Surrey, Guildford, UK, 2006

30.

Baid

, Sinaii

, Wade

, Rubino

, Nieman

. Radioimmunoassay and tandem mass spectrometry measurement of bedtime salivary cortisol levels: a comparison of assays to establish hypercortisolism. J Clin Endocrinol Metab 2007;92:3102–7

31.

Dorn

, Lucke

, Loucks

, Berga

. Salivary cortisol reflects serum cortisol: analysis of circadian profiles. Ann Clin Biochem 2007;44:281–4

32.

Findling

, Raff

, Aron

. The low dose dexamethasone suppression test: a re-evaluation in patients with Cushing's syndrome. J Clin Endocrinol Metab 2004;89:1222–6

33.

Wood

, Barth

, Friedman

, Perry

, Sheridan

. Evidence for the low dose dexamethasone suppression test to screen for Cushing's syndrome – recommendations for a protocol for biochemistry laboratories. Ann Clin Biochem 1997;34:222–9

34.

Viardot

, Huber

, Puder

, Zulewski

, Keller

, Muller

. Reproducibility of night-time salivary cortisol and its use in the diagnosis of hypercortisolism compared with urinary free cortisol and overnight dexamethasone suppression test. J Clin Endocrinol Metab 2005;90:5730–6

35.

Gafni

, Papanicolaou

, Nieman

. Night-time salivary cortisol measurement as a simple non-invasive outpatient screening test for Cushing's syndrome in children and adolescents. J Pediatr 2000;137:30–5

36.

Findling

, Raff

. Cushing's syndrome: important issues in diagnosis and management. J Clin Endocrinol Metab 2006;91:3746–53

37.

Fegan

, Sandeman

, Krone

, Cushing's syndrome in women with polycystic ovaries and hyperandrogenism. Nat Clin Pract Endocrinol Metab 2007;3:778–83

38.

Papanicolaou

, Mullen

, Kyrou

, Nieman

. Night-time salivary cortisol: a useful test for the diagnosis of Cushing's syndrome. J Clin Endocrinol Metab 2002;87:4515–21

39.

Putignano

, Toja

, Dubini

, Geraldi

, Corsello

, Cavagnini

. Midnight salivary cortisol versus urinary free and midnight serum cortisol as screening tests for Cushing's syndrome. J Clin Endocrinol Metab 2003;88:4153–7

40.

Meulenberg

PMM

, Hofman

. Differences between concentrations of salivary cortisol and cortisone and of free cortisol and cortisone in plasma during pregnancy and parturition. Clin Chem 1990;36:70–5

41.

Kidambi

, Raff

, Findling

. Limitations of nocturnal salivary cortisol and urine free cortisol in the diagnosis of mild Cushing's syndrome. Eur J Endocrinol 2007;157:725–31

42.

Atkinson

, Kennedy

, Carson

, Hadden

, Weaver

, Sheridan

. Five cases of cyclical Cushing's syndrome. Br Med J 1985;291:1453–7

43.

Meinardi

, Wolffenbuttel

, Dullaart

. Cyclical Cushing's syndrome: a clinical challenge. Eur J Endocrinol 2007;157:245–54

44.

Laudat

, Cerdas

, Fournier

, Guiban

, Guilhaume

, Luton

. Salivary cortisol measurement: a practical approach to assess pituitary – adrenal function. J Clin Endocrinol Metab 1988;66:343–8

45.

Castro

, Elias

LLK

, Elias

PCL

, Moriera

. A dose-response study of salivary cortisol after dexamethasone suppression test in Cushing's disease and its potential use in the differential diagnosis of Cushing's syndrome. Clin Endocrinol 2003;59:800–5

46.

Patel

, Shaw

, McIntyre

, McGarry

, Wallace

. Morning salivary cortisol versus short Synacthen test as a test of adrenal suppression. Ann Clin Biochem 2004;41:408–10

47.

Suri

, Moran

, Hibbard

, Kasza

, Weiss

. Assessment of adrenal reserve in pregnancy: defining the normal response to adrenocorticotropin stimulation. J Clin Endocrinol Metab 2006;91:3866–72

48.

Arafah

, Nishiyama

, Tlaygeh

, Hejal

. Measurement of salivary cortisol concentration in the assessment of adrenal function in critically ill subjects: a surrogate marker of the circulating free cortisol. J Clin Endocrinol Metab 2007;92:2965–71

49.

Contreras

, Arregger

, Persi

, Gonzalez

, Cardoso

. A new less-invasive and more informative low-dose ACTH test: salivary steroids in response to intramuscular corticotrophin. Clin Endocrinol (Oxf) 2004;61:675–82

50.

Cardoso

, Persi

, Gonzalez

, Assessment of adrenal function by measurement of salivary steroids in response to corticotrophin in patients infected with human immunodeficiency virus. Steroids 2007;72:328–34

51.

Arregger

, Cardoso

, Tumilasci

, Contreras

. Diagnostic value of salivary cortisol in end stage renal disease. Steroids 2008;73:77–82

52.

Marcus-Perlman

, Tordjman

, Greenman

, Low-dose ACTH (1µg) salivary test: a potential alternative to the classical blood test. Clin Endocrinol (Oxf) 2006;64:215–18

53.

Maguire

, Ambler

, Moore

, Waite

, McLean

, Cowell

. The clinical utility of alternative, less invasive sampling techniques in the assessment of oral hydrocortisone therapy in children and adolescents with hypopituitarism. Eur J Endocrinol 2007;156:471–6

54.

Thomson

, Devers

, Wallace

, Variability in hydrocortisone plasma and saliva pharmacokinetics following intravenous and oral administration to patients with adrenal insufficiency. Clin Endocrinol (Oxf) 2007;66:789–96

55.

Lovas

, Thorsen

, Husebye

. Saliva cortisol measurement: simple and reliable assessment of the glucocorticoid replacement therapy in Addison's disease. J Endocrinol Invest 2006;29:727–31

56.

Lovas

, Husebye

. Continuous subcutaneous hydrocortisone infusion in Addison's disease. Eur J Endocrinol 2007;157:109–12

57.

Zerah

, Pang

, New

. Morning salivary 17-hydroxyprogesterone is a useful screening test for nonclassical 21-hydroxylase deficiency. J Clin Endocrinol Metab 1987;65:227–32

58.

Zerah

, Ueshiba

, Wood

, Prevalence of nonclassical steroid 21-hydroxylase deficiency based on a morning saliva 17-hydroxyprogesterone screening test: a small sample study. J Clin Endocrinol Metab 1990;70:1662–7

59.

Ueshiba

, Zerah

, New

. Enzyme-linked immunosorbent assay (ELISA) method for screening of non-classical steroid 21-hydroxylase deficiency. Horm Metab Res 1994;26:43–5

60.

Otten

, Wellan

, Rijken

JCW

, Stoelinga

GBA

, Benraad

. Salivary and plasma androstenedione and 17-hydroxyprogesterone levels in congenital adrenal hyperplasia. J Clin Endocrinol Metab 1983;57:1150–4

61.

Walker

, Read

, Hughes

, Riad-Fahmy

. Radioimmunoassay of 17-hydroxyprogesterone in saliva, parotid fluid and plasma of congenital adrenal hyperplasia patients. Clin Chem 1979;25:542–5

62.

Young

, Robinson

, Read

, Riad-Fahmy

, Hughes

. 17-OH-Progesterone rhythm in congenital adrenal hyperplasia. Arch Dis Child 1988;63:617–23

63.

Dressendorfer

, Strasburger

, Bidlingmaier

, Development of a highly sensitive nonisotopic immunoassay for the determination of salivary 17-hydroxyprogesterone: reference ranges throughout childhood and adolescence. Paediatr Res 1998;44:650–5

64.

Groschl

, Rauh

, Dorr

. Circadian rhythm of salivary cortisol, 17α-hydroxyprogesterone and progesterone in healthy children. Clin Chem 2003;49:1688–91

65.

Young

, Walker

, Riad-Fahmy

, Hughes

. Androstenedione rhythms in saliva in congenital adrenal hyperplasia. Arch Dis Child 1988;63:624–8

66.

Baxendale

, Jacobs

, James

. Plasma and salivary androstenedione and dihydrotestosterone in women with hyperandrogenism. Clin Endocrinol (Oxf) 1983;18:447–57

67.

McVie

, Levine

, New

. The biological significance of the aldosterone concentration in saliva. Pediatr Res 1979;13:755–9

68.

Atherden

, Corrie

, Jones

, Al-Dujaili

, Edwards

. Development and application of a direct radioimmunoassay for aldosterone in saliva. Steroids 1985;46:845–55

69.

Few

, Chaudry

, James

. The direct determination of aldosterone in human saliva. J Steroid Biochem 1984;21:87–92

70.

Few

, Wallis

, James

. The relationship between salivary aldosterone and plasma free aldosterone concentrations. Clin Endocrinol (Oxf) 1986;24:119–26

71.

Few

, Painlin

, James

. The relation between aldosterone concentration in plasma and saliva during pregnancy. Br J Obstet Gynaecol 1986;93:928–32

72.

Few

, Unwin

, Carmichael

, James

. Diurnal fluctuation in saliva aldosterone concentration. J Steroid Biochem 1987;26:265–71

73.

Granger

, Schwartz

, Booth

, Curran

, Zakaria

. Assessing dehydroepiandrosterone in saliva: a simple radioimmunoassay for use in studies of children, adolescents and adults. Psychoneuroendocrinolgy 1999;24:567–79

74.

Ahn

, Lee

, Choi

, Kwon

, Chun

. Salivary cortisol and DHEA levels in the Korean population: age-related differences, diurnal rhythm and correlations with serum levels. Yonsei Med J 2007;48:379–88

75.

Arlt

, Callies

, van Vlijmen

, Dehydroepiandrosterone replacement in women with adrenal insufficiency. N Engl J Med 1999;341:1013–20

76.

Wolf

, Kirshbaum

. Actions of dehydroepiandrosterone and its sulphate in the central nervous system: effects on cognition and emotion in animals and humans. Brain Res Rev 1999;30:264–88

77.

Flynn

, Pizza

, Brolinson

. Hormonal response to exercise training: influence of cross- training. Int J Sports Med 1997;18:191–6

78.

Evans

, Stewart

, Merrick

. Oestradiol in saliva during the menstrual cycle. Br J Obstet Gynaecol 1980;87:624–6

79.

Bao

, Liu

, van Someren

, Hofman

, Cao

, Zhou

. Diurnal rhythm of free oestradiol during the menstrual cycle. Eur J Endocrinol 2003;148:227–32

80.

Gandara

, Leresche

, Manci

. Patterns of salivary estradiol and progesterone across the menstrual cycle. Ann NY Acad Sci 2007;1098:446–50

81.

Chatterton

Jr , Muteo

, Hou

, Characteristics of salivary profiles of oestradiol and progesterone in premenopausal women. J Endocrinol 2005;186:77–84

82.

Tivis

, Richardson

, Peddi

, Arjmandi

. Saliva versus serum estradiol: implications for research studies using postmenopausal women. Prog Neuropsychopharmacol Biol Psychiatry 2005;29:727–32

83.

McGregor

, Hastings

, Roberts

, Barrett

. Diurnal variation in saliva oestriol level during pregnancy: a pilot study. Am J Obstet Gynecol 1999;180:5223–5

84.

Vining

, McGinley

, Rice

. Saliva estriol measurements: an alternative to the assay of serum unconjugated estriol in assessing fetoplacental function. J Clin Endocrinol Metab 1983;56:454–60

85.

Lechner

, Heim

, Zech

, Daxembichler

, Martin

. The relation between saliva oestriol levels in pregnancy and infant birth weight. Arch Gynecol Obstet 1987;241:9–12

86.

Heine

, McGregor

, Goodwin

, Serial salivary estriol to detect an increased risk of preterm birth. Obstet Gynecol 2000;96:490–7

87.

McGregor

, Hastings

, Dullien

. A meta-analysis of salivary estriol (sE(3)) as a means to identify women at risk for preterm birth due to fetal- placental endocrine mechanisms. Prim Care Update Obstet Gynecol 1998;5:179–86

88.

Ellison

. Measurements of salivary progesterone. Ann NY Acad Sci 1993;694:161–76

89.

Ellison

. Salivary steroids and natural variation in human ovarian function. Ann NY Acad Sci 1994;709:287–98

90.

Read

, Wilson

, Hughes

. The use of salivary progesterone in the assessment of ovarian function in post-menarchal girls. J Endocrinol 1984;102:265–8

91.

Groschl

, Rauh

, Schmidt

, Dorr

. Relationship between salivary progesterone, 17-hydroxyprogesterone and cortisol levels throughout the normal menstrual cycle of healthy postmenarchal girls. Fertil Steril 2001;76:615–17

92.

Metcalf

, Evans

, Mackenzie

. Indices of ovulation: comparison of plasma and salivary levels of progesterone with urinary pregnanetriol. J Endocrinol 1984;100:75–80

93.

Ishikawa

, Sengoku

, Tamate

, Takaoka

, Kane

, Fottrell

. The clinical usefulness of salivary progesterone assessment for the evaluation of the corpus luteum function. Gynecol Obstet Invest 2002;53:32–7

94.

Delfs

, Klein

, Fottrell

, Naether

, Leidenberger

, Zimmermann

. 24-hour profiles of salivary progesterone. Fertil Steril 1994;62:960–6

95.

Morley

, Perry

III , Patrick

, Dollbaum

, Kells

. Validation of salivary testosterone as a screening test for male hypogonadism. Aging Male 2006;9:165–9

96.

Wang

, Wakelin

, White

, Wood

. Salivary androgens in hirsutism: are they of use in routine evaluation? Ann Clin Biochem 1986;23:590–5

97.

Mason

, Wallace

, Macintyre

, Undetectable salivary testosterone in young women with premature ovarian failure. Clin Endocrinol (Oxf) 2006;64:711–14

98.

Orozco

, Navarro

, Nolla

. Salivary testosterone is associated with higher lumbar bone mass in premenopausal healthy women with normal levels of serum testosterone. Eur J Epidemiol 2000;16:907–12

99.

Hubl

, Taubert

, Freymann

, A simple solid phase enzyme immunoassay for aldosterone in plasma and saliva. Exp Clin Endocrinol 1983;82:188–93

100.

Riad-Fahmy

, Read

, Walker

, Griffiths

. Steroids in saliva for assessing endocrine function. Endocr Rev 1982;3:367–95

101.

Luthold

, Marcondes

, Wajchenberg

. Salivary cortisol for the evaluation of Cushing's syndrome. Clin Chim Acta 1985;151:33–9

102.

Gallagher

, Leitch

, Massey

, McAllister-Williams

, Young

. Assessing cortisol and dehydroepiandrosterone (DHEA) in saliva: effects of collection method. J Psychopharmacol 2006;20:643–9

103.

Price

, Astin

, Chard

, Addison

. Assay of hydroxyprogesterone in saliva. Lancet 1979;1:368–9

104.

Worthman

, Stallings

, Hofman

. Sensitive salivary oestradiol assay for monitoring ovarian function. Clin Chem 1990;36:1769–73

105.

Wang

, Fantl

, Habibollahi

, Salivary oestradiol and progesterone levels in premenopausal women with breast cancer. Eur J Cancer Clin Oncol 1986;22:487–92

106.

, Bentley

, Gann

, Hodges

, Chatterton

. Salivary estradiol and progesterone levels in conception and non-conception cycles in women: evaluation of a new assay for salivary estradiol. Fertil Steril 1999;71:863–8

107.

Lachelin

, McGarrigle

. A comparison of saliva, plasma unconjugated and plasma total oestriol levels throughout normal pregnancy. Br J Obstet Gynaecol 1984;91:1203–9

108.

De Boever

, Kohen

, Vandekerckhova

. Direct solid phase chemiluminescence immunoassay for salivary progesterone. Clin Chem 1986;32:763–7

109.

Bolaji

. Sero-salivary progesterone correlation. Int J Gynaecol Obstet 1994;45:125–31

110.

, Chatterton

, Vogelsong

, May

. Direct radioimmunoassay of progesterone in saliva. J Immunoassay 1997;18:149–63

111.

Vitek

, L'Hommedieu

, Gordon

, Rappaport

, Southern

. Direct radioimmunoassay (RIA) of salivary testosterone: correlation with free and total serum testosterone. Life Sci 1985;37:711–16

112.

Barrou

, Guiban

, Maroufi

, Overnight dexamethasone suppression test: comparison of plasma and salivary cortisol measurement for the screening of Cushing's syndrome. Eur J Endocrinol 1996;134:93–6

113.

Raff

, Raff

, Findling

. Late-night salivary cortisol as a screening test for Cushing's syndrome. J Clin Endocrinol Metab 1998;83:2681–6

114.

Castro

, Elias

, Quidute

ARP

, Halah

FPB

, Moriera

. Out-patient screening for Cushing's syndrome: the sensitivity of the combination of circadian rhythm and overnight dexamethasone suppression salivary cortisol tests. J Clin Endocrinol Metab 1999;84:878–82

115.

Martinelli

, Sader

, Oliveira

, Daneluzzi

, Moreira

. Salivary cortisol for screening for Cushing's syndrome in children. Clin Endocrinol 1999;51:67–71

116.

Yaneva

, Mosnier-Pudar

, Dugue

M-A

, Grabar

, Fulla

, Bertagna

. Midnight salivary cortisol for the initial diagnosis of Cushing's syndrome of various causes. J Clin Endocrinol Metab 2004;89:3345–51

117.

Trilck

, Flitsch

, Ludecke

, Jung

, Petersenn

. Salivary cortisol measurement – a reliable method for the diagnosis of Cushing's syndrome. Exp Clin Endocrinol Diabetes 2005;113:225–30