1 s2.0 S096007601630108X Main

G Model
SBMB 4705 No. of Pages 10
Journal of Steroid Biochemistry & Molecular Biology xxx (2015) xxxxxx
Contents lists available at ScienceDirect
Journal of Steroid Biochemistry & Molecular Biology

journal homepage: www.elsevier.com/locate/jsbmb
Validation of an LCMS/MS salivary assay for glucocorticoid status

assessment: Evaluation of the diurnal uctuation of cortisol and
cortisone and of their association within and between serum and saliva
Marco Mezzullo, Flaminia Fanelli, Alessia Fazzini, Alessandra Gambineri,
Valentina Vicennati, Guido Di Dalmazi1, Carlotta Pelusi, Roberta Mazza, Uberto Pagotto,
Renato Pasquali*
Endocrinology Unit, Department of Medical and Surgical Sciences, Centre for Applied Biomedical Research (C.R.B.A.), S. Orsola-Malpighi Hospital, Alma Mater
University of Bologna, Bologna, Italy
A R T I C L E I N F O
Article history:
Received 11 January 2016
Received in revised form 21 March 2016
Accepted 20 April 2016
Available online xxx
Keywords:
Cortisol
Cortisone
Saliva
LCMS/MS
Circadian rhythm
A B S T R A C T
Salivary steroid testing represents a valuable source of biological information; however, the proper
measurement of low salivary levels is challenging for direct immunoassays, lacking adequate sensitivity
and specicity and causing poor inter-laboratory reproducibility. Liquid chromatography-tandem mass
spectrometry (LCMS/MS) has overcome previous analytical limits, often providing results deviating
from previous knowledge. Nowadays, LCMS/MS is being introduced in clinical laboratories for salivary
cortisol testing; however, so far only a few studies have reported thorough biological validation based on
LCMS/MS data. In this study, we provide a thorough analytical, pre-analytical and biological validation
of an LCMS/MS method for the measurement of salivary cortisol (F) and of its inactive metabolite
cortisone (E).
Analytes were extracted from 50 ml of saliva, were then separated in 7.5 min LC-gradient and detected
by negative electrospray ionizationmultiple reaction monitoring. The reliability of a widely diffused
collection device, Salivette1, was assessed and the overall procedure was validated. The diurnal cortisol
and cortisone uctuation in saliva and serum was described by a four paired collection protocol (8 am,
12 am, 4 pm and 8 pm) in 19 healthy subjects. The assay allowed the quantitation of F and E down to
39.1 and 78.1 pg/ml, with an imprecision range of 5.59.5%, 3.914.1% and 2.614.4%, and an accuracy
range of 105.5113.1%, 88.598.7% and 90.796.7% for both analytes at low, medium and high levels,
respectively. Salivette1 provided comparable results and better precision (CV < 1.0%) as referred to direct
spitting (CV < 13.0%). A parallel diurnal rhythm in saliva and serum was observed for cortisol and
cortisone, with values lowering from the morning to the evening time points (P < 0.0001). While salivary
E linearly correlated to total serum F (R2 = 0.854, P < 0.001), salivary F showed an exponential relationship
(R2 = 0.903, P< 0.001) with serum F reecting the free circulating fraction. A non linear association
between E and F was observed in saliva (R2 = 0.941, p < 0.001) consistent with the type II 11b-HSD
activity. We concluded that our LCMS/MS method allowed a sensitive evaluation of salivary levels of
cortisol and cortisone. The simultaneous determination of both hormones in saliva allowed the
differential estimation of the active and of the total glucocorticoid exposure over the daytime. The assay
could provide further insight into the comprehension of normal and dysfunctional glucocorticoid
circadian rhythm.
2016 Elsevier Ltd. All rights reserved.
1. Introduction
* Corresponding author at: Endocrinology Unit, Department of Medical and

SurgicalSciences, Alma Mater University of Bologna, via Massarenti 9, 40138
Bologna, Italy.
E-mail address: renato.pasquali@unibo.it (R. Pasquali).
1
Present address: Medizinische Klinik und Poliklinik IV, Klinikum der Universitt
Mnchen, Munich, Germany.
Morethan thirty years ago, salivary cortisol (F) level, along with
many others unconjugated steroids, was proved to closely approximate the unbound F concentration in plasma [1]. According to the
free hormone hypothesis [2], the biological activity of a given
hormone depends on its unbound (free) rather than on its
http://dx.doi.org/10.1016/j.jsbmb.2016.04.012
0960-0760/ 2016 Elsevier Ltd. All rights reserved.
Please cite this article in press as: M. Mezzullo, et al., Validation of an LCMS/MS salivary assay for glucocorticoid status assessment: Evaluation
of the diurnal uctuation of cortisol and cortisone and of their association within and between serum and saliva, J. Steroid Biochem. Mol. Biol.
(2016), http://dx.doi.org/10.1016/j.jsbmb.2016.04.012
G Model
M. Mezzullo et al. / Journal of Steroid Biochemistry & Molecular Biology xxx (2015) xxxxxx
protein-bound fraction in plasma. However, the estimation of free F

fraction is technically demanding and is not available in routine
settings. Salivary F testing represents an easier alternative to free
circulating F determination for the evaluation of the hypothalamic
pituitary adrenal (HPA) axis status, offering analytical and practical
advantages, such as sample collection, that is not painful and can be
self-performed multiple times during the day without altering the
patients ordinary activities and without causing stress-derived bias
[3]. Conversely, total serum F, which is widely routinely assayed in
clinical laboratories, may be a misleading indicator of adrenal status
in pathologic states altering the circulating levels of the carrier
protein corticosteroids binding protein (CBG) and albumin, such as
liver dysfunctions, estrogen therapies and hyperinsulinemic states
[4]. To date, a great deal of evidence has been reported in the
literature supporting the use of salivary F both in clinical and in
research medicine, in sport and in stress research elds, as well as in
diagnostic protocols of hyper- and hypocortisolism states. The better
diagnostic performance of late night salivary F (LNSF) compared to
urinary free F (UFC) and dexamethasone suppression test in the
discrimination between overt and pseudo-Cushings syndrome (CS)
has been emphasized in recent years [5]. In addition, LNSF was found
to be elevated in patients with adrenal adenoma and proved to be
useful in the follow-up of Cushings disease patients that may relapse
after surgery [6]. Cortisone (E) is the inactive metabolite of F, and its
salivary level is regulated by passive diffusion and by F-to-E
conversion operated by type II 11b-hydroxysteroid dehydrogenase
(11b-HSD II). Although recent ndings suggest that salivary E may
reect plasma F levels more accurately than salivary F [7], only a few
studies focused on a detailed description of the relationship between
F and E levels in the circulating and salivary fraction [8]. This was
partly due to the general inadequacy of common immunoassays in
determining low steroid levels in saliva [9], typically 10100 folds
lower than serum, and to the antibody cross-reactivity between E
and F and with other structurally related endogenous steroids, such
as 11-deoxycortisol, or with commonly prescribed oral synthetic
corticosteroids prednisone and prednisolone [10]. Immunoassay
limits are largely responsible for the poor agreement among
reference ranges and diagnostic cut-offs reported so far for salivary
F [11] and represent the main bottleneck toward an adequate interlaboratory standardization of salivary F testing. Liquid chromatography-tandem mass spectrometry (LCMS/MS) offers high sensitivity and specicity to detect low salivary concentrations. Its diffusion
in clinical and research settings is rapidly increasing, and is
upgrading the robustness of hormone salivary testing, as witnessed
by the fact that the LNSF cut-offs for CS proposed so far determined by
LCMS/MS methods assayed in different laboratories are very
consistent [1216]. Conversely, studies directly comparing immunoassays vs LCMS/MS showed that the extent of deviation may
sometimes be very relevant. In this regard, though several LCMS/MS
methods for salivary F have been reported so far in the literature, only
a few studies have focused on a biological validation of the assay. In
this study, we aimed at providing a thorough analytical and preanalytical validation of a LCMS/MS method, also verifying the
reliability of a widely diffused collection device, Salivette1.
Moreover, we validated the biological consistency of our assay by
analyzing the diurnal uctuation of salivary and circulating F and E
levels, and by performing a study of the association between
reciprocal F and E levels both in saliva and in serum and between
salivary and serum fractions.
2. Materials and methods
2.1. Chemicals and disposable material
Gradient-grade methanol and dichloromethane were purchased by Merck (Darmstadt, Germany); ultrapure water was
obtained by a MilliQ Gradient A10 System (Millipore, Volketswil,

Switzerland). Lyophilized F and E pure standard were by Steraloids
(Newport, RI, USA), cortisol-[9,11,12,12-2H4] (deuterium contents
97%, d4-cortisol) was from CDN Isotopes (Pointe-Claire, Canada)
and cortisone-[2,2,4,6,6,12,12-2H7] (deuterium contents 98%, d7cortisone) was from Sigma Aldrich (St. Louis, MO, USA). Fifteen ml
polypropylene tubes and cotton Salivette1 were provided by
Sarstedt (Nmbrecht, Germany); 1.5 ml polypropylene tubes were
from Eppendorf (Hamburg, Germany).
2.2. Specimens
Thirty-seven healthy volunteers, 23 males and 14 females aged
2565 years, were recruited from the Hospital staff after giving
their informed consent according to the Helsinki declaration. The
study was approved by the local Ethics Committee. Their body
mass index (BMI) ranged between 18.0 and 25.0 kg/m2. None of the
subjects was taking any drug; none reported irregular sleep-wake
cycles or was a shift worker. Further exclusion criteria were the
presence of any oral inammatory processes with or without
evident bleeding and/or any dental care treatment within 30 days
before the collection date. Subjects were asked to avoid drinking,
eating and brushing their teeth 30 min before the collection of
saliva specimens.
A subgroup of 18 volunteers (11 males and 7 females)
participated in the collection device comparison. First, each
subject collected 1 ml of saliva by direct spitting in a 15 ml
polypropylene tube; they then repeated the collection using the
Salivette1 according to the manufactory protocol. Samples were
transferred to 1.5 ml tubes and stored at 80 C until analysis.
For the biological validation, a subgroup of 19 subjects (12 males
and 7 females) underwent a paired blood and saliva collection
protocol performed at 8 am, 12 am, 4 pm and 8 pm of an ordinary
working day. The lunch time was scheduled between 12:30 am and
1:00 pm and was consistent with the habitual lunch time of each
subject. A 10 min saline infusion was performed before simultaneous blood and saliva collection in order to avoid venipunctureinduced F raise. Blood was withdrawn in Vacuette Z serum beads
clot activator tubes (Greiner Bio-One, Kremsmunster, Germany),
allowed to settle for 20 min and then centrifuged at 2000g for
10 min at room temperature. Saliva was collected by Salivette1. All
samples were transferred to 1.5 ml tubes and stored at 80 C until
analysis.
2.3. Calibrators and in-house quality controls
Gravimetrically determined stock solutions were prepared in
methanol for each standard and isotopically labeled internal
standard (IS) at different concentrations in the mg/ml range.
Working solutions were prepared in methanol at 100 mg/ml. A
calibrating solution was prepared by mixing analyte standards to a
nal concentration of 1 mg/ml for both F (2760 nmol/l) and E
(2770 nmol/l). The IS working solution was obtained by mixing d4cortisol and d7-cortisone at the nal concentration of 3 ng/ml in
35% methanol solution (V/V%). The day of the assay, 40 ml of the
calibrating solution were spiked in 1 ml of the IS working solution.
Serial dilutions were performed to generate an eight-point
calibration curve at 0.000, 0.019, 0.039, 0.156, 0.625, 2.5, 10, and
40 ng/ml for both analytes. Stock, working and calibrating
solutions were stored at 20 C. In-house quality controls (QCs)
were prepared by pooling saliva collected by healthy volunteers
between 10 pm and 12 pm: the low level was represented by the
pool used as such, while the medium and high levels were
prepared by spiking proper amounts of standard analytes at the
nal concentration of 2 and 10 ng/ml for F and 5 and 20 ng/ml for E,
respectively. 70 ml aliquots of each QC were prepared and stored at
G Model
80 C. Three aliquots of each QC level were thawed and processed

as unknown samples in each batch.
2.4. Sample preparation
Saliva samples and QCs were thawed, vortexed and centrifuged
at 5500g for 5 min at room temperature. 50 ml of clear supernatants were transferred into 13 100 mm Pyrex1 tubes (SigmaAldrich, St. Louis, MO) before adding 100 ml of IS working solution.
After 1 min vortex, 1 ml H2O was added to all tubes and they were
vortexed again for 1 min. Analytes were extracted by adding 2 ml
dichloromethane and gently vortexing for 5 min. All tubes were
centrifuged 2500g for 15 min at room temperature, the organic
layer was then transferred into 12 75 mm glasses tubes
(Laboindustria S.p.A. Italy) and dried under nitrogen stream.
Samples were reconstituted with 100 ml of 35% methanol solution
and transferred into screw glass vials equipped with 300 ml insert
sealed with PTFE-Silicone pre-slit septa cap (Phenomenex,
Torrance, CA)and placed in a Series 200 autosampler thermostated
at 15 C (PerkinElmer, Turku, Finland).
2.5. Salivary cortisol and cortisone assay by LCMS/MS
15 ml were injected into a Series 200 modular HPLC (PerkinElmer). Analytes were separated on a LUNA RP-C8 100 4.6 mm,
5 mm column equipped with a RP-C8 4 2 mm 5 mm, guard
column (Phenomenex, Torrance, CA) by a 750 ml/min ow of an
eluent consisting of 20% solvent B (methanol) and 80% solvent A
(20% methanol). The 7.5 min gradient run program started with a
linear gradient reaching 50% solvent B in 2.6 min, followed by an
increase to 100% solvent B in 1.4 min. The system was washed with
100% solvent B for 1 min, then re-equilibrated to the initial
conditions in 2.5 min. In order to protect the ion source from early
and late eluting matrix compounds, the ow was diverted from the
column outlet to the waste from min 02 and from min 6 to the end
of the run by using a ten port switching valve (VICI1, Valco
Instruments Co., Inc.); accordingly, the source was ushed by a 50%
methanol solution at 300 ml/min. MS detection was performed by
an API 4000-QTrap triple quadrupole (AB-Sciex Framingham, MA,
USA). Quantication was performed by multiple reaction monitoring (MRM), by selecting two specic transitions, the quantier
and the qualier, for each analyte. The parameters pertaining to
each MRM transition were optimized by infusing standard
solutions at 10 mg/ml into the TurboIonSpray1 source operating
in negative ion mode by an infusion pump set at 10 ml/min.
Collision activated dissociation (CAD) gas was nitrogen set at
medium. Mother ion (Q1)/product ion (Q3) mass-to-charge
transitions and compound dependent parameters are reported in
Table 1. The ion source operated at 450 C probe temperature,
35 psi curtain gas, 55 psi nebulizing gas, 45 psi heating gas, 4500 kV
ion spray voltage. Unitary mass resolution was set at both rst and
third quadrupole.
Data processing and quantitation were performed by Analyst
1.5.1 software (AB-Sciex). Calibration was done by linear regression
with a 1/x weighting; analyte concentrations in each sample were
back calculated by the interpolation on the respective regression
curve
accuracy between 80 and 120% of the expected value and an

imprecision 15%.
A dedicated pool of Salivette1-specimens, collected late at
night in order to obtain low E and F levels, was prepared to test
functional sensitivity, accuracy, recovery and matrix effect.
Accuracy was assayed by spiking known amounts of analyte
standards in saliva pool in order to produce low (39.1 and
78.1 pg/ml), medium (625.0 and 2500.0 pg/ml) and high
(10,000 and 20,000 pg/ml) levels for F and E, respectively. Each
level was analyzed in triplicate as previously described. Accuracy
was calculated at each level by subtracting the basal level observed
in the unspiked saliva; the percentage of the resulting value over
the spiked concentration was then calculated. The functional
sensitivity was determined by spiking small amounts of analyte
standards in the saliva pool and dened as the smallest
concentration measurable in saliva matrix with accuracy between
80 and 120% and an imprecision 15%.The functional sensitivity
was calculated by subtracting the basal level observed in the
unspiked saliva. Recovery was assessed by adding 40 ml of analyte
standard at 50, 250 and 1000 ng/ml in 35% MeOH to independent
aliquots of 2 ml of saliva pool in order to produce low, medium and
high levels at 1, 5 and 20 ng/ml, respectively. Besides, saliva pool
samples were extracted as such and reconstituted with 100 ml of
low, medium and high level solutions of E and F prepared in 35%
methanol at 0.5, 2.5 and 10 ng/ml, respectively. Each test sample
and unspiked saliva pool was processed in three replicates as
previously described. In order to calculate recovery at each level,
the analyte peak area detected in unspiked saliva was subtracted
from the peak area detected in spiked test samples, and the ratio
between the obtained values was calculated according to the
formula:

spiked saliva analyte area unspiked saliva analyte area
%
post extraction spiked saliva unspiked saliva analyte area
In order to calculate matrix effect, saliva pool dried extracts
were reconstituted with 100 ml of low, medium and high level E
and F solutions prepared in 35% methanol at 0.5, 2.5 and 10.0 ng/
ml. The analyte peak area detected in unspiked saliva was
subtracted from the peak area obtained by the spiked test samples;
the resulting values were then compared with the peak area
produced by direct injections of E and F solutions prepared in 35%
methanol at 0.5, 2.5 and 10 ng/ml according to the formula:

post extraction spiked saliva unspiked saliva analyte area
%
analyte standard area
A post-column infusion of a d4-F and d7-E suitable to yield a
signal around 20,000 cps, corresponding to the peak height
observed in the mid calibration range, was performed during the
injection of saliva pool extracts. A reduction in the signal was
expected in correspondence with the elution of matrix components causing ion suppression, and was considered relevant if
Table 1
LCMS/MS compound dependent parameters.
Analyte
MW
RT
Transition
Q1
Q3
DP
CE
CXP
IR
2.6. Salivary cortisol and cortisone LCMS/MS method validation
362.4
5.15
d4-F
E
366.1
360.4
5.13
4.99
d7-E
367.4
4.97
361.1
361.1
365.1
359.1
359.1
366.3
331.2
297.0
335.2
329.1
301.1
336.3
60
60
60
70
70
50
12
35
20
15
20
15
12
5
12
6
7
19
3.8
The limit of detection (LOD) was determined by injecting scalar

dilution of analyte standards in triplicate and dened as the
smallest quantity producing a peak with a signal-to-noise (S/N) 3.
The lower limit of quantitation (LLOQ) was determined by injecting
each calibrator in triplicate and was dened as the smallest
concentration producing a peak with an S/N 10 measurable with
Quantier
Qualier
IS
Quantier
Qualier
IS
nd
4.3
nd
MV: molecular weight (amu); RT: retention time (min); Q1: rst quadrupole, Q3:
third quadrupole; DP: declustering potential (V); CE: collision energy (V); CXP: cell
exit potential (V); IR: ion ratio.
G Model
Table 2
Method validation parameters.
Analyte
LOD
LLOQ
Calibration Curve
Spiked Concentration in Saliva
pg on column
s/n
pg/ml
CV%
Accuracy%
s/n
equation
CV%
Accuracy%
2.85
39.1
7.6
99.8
26
y = 0.1923x + 0.0001
0.996
39.1a
625
10000
12.9
6.5
5.3
105.5
88.5
90.7
2.85
39.1
8.1
98.8
21
y = 0.3383x + 0.0015
0.993
78.1a
2500
20000
9.5
3.6
5.2
113.1
98.7
96.7
pg/ml
Functional sensitivity in saliva matrix.
occurring in the analyte elution window. Specicity was assessed

by determining the ion ratio (IR) between quantitative and
qualitative transition peak area produced in analyte standards.
Sample measurements were considered not acceptable whenever
the analyte IR exceeded 80120% of the reference IR. Moreover,
measurements were not accepted whenever a broadening of the
analyte peaks, in terms of half height width, or peak split was
observed. Method imprecision was evaluated on ve replicate
measurements of low, medium, and high level QCs analyzed in the
same run (intra-assay) and in ve runs performed in ve
successive weeks (inter-assay).
2.7. Saliva collection device comparison
Salivette1 is supplied with a natural cotton swab containing
citric acid. Many authors reported low recovery for some steroids
or some detrimental effect of cotton derived substances on assay
performance [17,18]. Since cotton Salivette1 has been widely
adopted in many laboratories throughout the world, we evaluated
the performance of the Salivette1 in comparison to a simple direct
spitting procedure for saliva collection. Each sample was processed
and analyzed in duplicate. The overall imprecision was calculated
according to the following formula:
q
Sa b2 =2n
%
Sa b=2n
a = rst measurement, b = second measurement, n = number of
samples.
Duplicate mean values were used for assessing the agreement
between the two procedures according to Bland and Altman
analysis [19]; moreover, the occurrence of any proportional and/or
systematic bias caused by the Salivette1 device was investigated
by Passing and Bablok regression analysis [20] and the Pearson
correlation coefcient (r).
2.8. Serum cortisol and cortisone LCMS/MS measurement

Serum F was measured by 2D-LCMS/MS as detailed elsewhere
[21]. After protein precipitation of 900 ml of serum and addition of
IS, followed by solid phase extraction (SPE), samples were injected
into a 2D-LC system, puried by a perfusion column, and separated
on a Luna RP-C8 100 4.6 mm, 5 mm (Phenomenex, Torrance,
California) in a 21 min gradient run. Analytes underwent positive
atmospheric pressure chemical ionization (APCI) and MRM
detection by an API 4000-QTrap (AB-Sciex). Quantier and
qualier transitions for F were 363.2/121.2 and 363.2/267 respectively. For d4-cortisol, the 367.3/97.1 transition was used. E was
included in the prole for the specic purpose of this study, by
including 361.1/163.1 and 361.1/121 quantier and qualier
transitions, respectively. For d7-cortisone, the 369.3/169.2 transition was used. The functional sensitivity was 0.195 and 0.244 ng/
ml, the intra- and inter-assay imprecision were <6 and 9% for E and
F, respectively, and accuracy ranged between 92.9 and 104.3% at
low, medium and high concentration levels for both analytes.
2.9. Statistics
Statistical analyses were performed by SPSS (SPSS 20.0; SPSS
Inc., IL, USA) and Prism (Prism 5.01; GraphPad Software Inc., CA,
USA) software. Normality was checked using the Kolmogorov
Smirnov test. Correlation coefcient was calculated by Pearson
correlation. Variations in E and F and in the F/E ratio in serum and
saliva over daytime were evaluated by repeated measurement
analysis of variance (rmANOVA), followed by multiple Bonferronicorrected pairwise comparison tests. P 0.05 was considered
statistically signicant.
3. Results
3.1. LCMS/MS method validation
The retention time variability was <0.1% for all analytes and no
carry-over was detected after the injection of the highest
calibration point. The LOD was 2.85 pg/on column for both F
Table 3
LLE extraction recovery and matrix effect. QCs imprecision at low, medium and high levels.
Analyte
Spiked Concentration
ng/ml
Extraction Recovery
%
Matrix effect
%
1
5
20
68.2
77.6
78.6
104.2
102.0
109.0
1
5
20
77.5
92.4
97.9
101.9
100.6
96.4
QC level
ng/ml
0.256
2.113
9.767
2.06
7.63
25.17
intra-assay
CV%
inter-assay
CV%
5.5
3.9
2.7
9.5
10.5
12.5
6.7
4.1
2.6
8.1
14.1
15.4
G Model
(7.86 pmol/l) and E (7.89 pmol/l). The LLOQ was 39.1 pg/ml for both
F (107.9 pmol/l) and E (108.2 pmol/l).The assay was linear (r > 0.99)
up to 40 ng/ml for both F (110.4 nmol/l) and E (110.8 nmol/l). The
functional sensitivity was 39.1 pg/ml for F (107.9 pmol/l) and
78.1 pg/ml for E (216.3 pmol/l). Method accuracy at low, medium
and high levels was 105.5, 88.5 and 90.7% for F and 113.1, 98.7 and
96.7% for E, respectively (Table 2). The recovery ranged between
68.1 and 97.9% and the matrix effect ranged between 96.4 and
109.1% for both analytes. The absence of relevant suppression
phenomena was further conrmed by the absence of change in IS
response in the post-column infusion experiment. Intra-day
imprecision was <7% for both F and E, while inter-day imprecision
was <13 and <16% for F and E, at low, medium and high levels
respectively (Table 3). In more than 1000 samples, less than 1%
were below the LLOQ or presented not acceptable IR or peak shape.
3.2. Saliva collection device comparison
As shown in Fig. 1, the noise level as well as the presence of
adjacent peaks is similar between the chromatograms obtained by
direct spitting and Salivette1. The overall imprecision for F and E
calculated on duplicate measurement of direct spitting specimens
was 12.6 and 5.2%, while for Salivette1 specimens it was 0.3 and
0.4%, respectively.
The Bland and Altman plot for F (2.A) and E (2.B) revealed a
mean bias of 15.9% for F and 11.5% for E, with an agreement
ranging between 56% and 24.5% for F and 45.3% 22.3% for E.
The Passing and Bablok regression analysis revealed intercept
coefcients not different from 0 as well as slope coefcients not
different from 1 for both F (2.C) and E (2.D), thus excluding the
presence of any systematic or proportional bias (Fig. 2).
3.3. Biological validation
Fig. 3.A and B conrmed the overlap between circadian
variation of serum and salivary F and E. A signicant inuence
of the time of the sampling on serum F and E as well as on salivary F
and E was observed (p < 0.0001 for all). As expected, the serum and
salivary F levels declined during the daytime, with levels at 8 am
higher than at 12 am, 4 pm and 8 pm (p < 0.001 for all paired tests);
moreover, serum F at 12 am (p < 0.05) and salivary F at 12 am
(p < 0.01) were higher than F at 8 pm. Similarly, salivary E at 8 am
was higher than at 12 am, 4 pm and 8 pm (p < 0.0001 for all paired
tests); moreover, salivary E at 12 am was higher than salivary E at
8 pm (p < 0.05). Serum E levels at 8 am remained steady at 12 am,
but decreased at 4 pm and 8 pm (p < 0.001 for the two paired
comparisons). The F/E ratio uctuation is reported in Fig. 3C: the
F/E ratio at 8 am was higher compared to 12 am, 4 pm and 8 pm in
both saliva (p < 0.0001 for all paired tests) and serum (p < 0.05 for
all paired tests); moreover, salivary F/E ratio at 12 am was higher
than at 4 pm (p < 0.05) and 8 pm (p < 0.01).
Regression analysis performed on the overall paired samples

revealed that salivary F increased at increasing serum F levels with
an exponential t (R2 = 0.903, p < 0.001) (Fig. 4A), while salivary E
showed a linear increase as referred to increasing serum F levels
(R2 = 0.854, p < 0.001) (Fig. 4B). A poorer although signicant
correlation between salivary E and serum E (R2 = 0.359, p < 0.001)
and salivary F and serum E (R2 = 0.194, p < 0.001) was also found
(Fig. 4C and D). As reported in Fig. 4E a non-linear relationship was
found between serum F and serum E (R2 = 0.659, p < 0.001),
reaching a plateau at serum F levels above 100 ng/ml. Finally, the
relationship between salivary F and E (Fig. 4F) was described by a
non-linear regression equation (R2 = 0.941, p < 0.001).
4. Discussion
The eld of steroid salivary testing is gaining renewed interest
in research and clinical elds thanks to LCMS/MS innovation that
makes it possible to overcome the sensitivity and specicity limits
of direct immunoassays for the proper measurement of low
salivary levels.
We developed and validated a high-sensitive LCMS/MS
method for the quantitation of salivary F and E requiring minimal
volume similar or smaller than volumes reported in most of the
previous LCMS/MS methods [12,16,2230], that can be easily
collected several times during the day.
Saliva uids contain oral bacteria and enzymes that could affect
the stability of salivary hormones, nonetheless, many authors
reported that salivary cortisol is stable at least for one month at
room temperature [31,32]. The stability of salivary cortisol
therefore guaranteed a safe home-based self-collection and
laboratory processing.
Compared to off-line SPE approach [16,27], our LLE was rapid,
cost-saving and easily automatable by using a liquid-handler for
high-throughput routine application. Besides, though the on-line
SPE based on simple sample dilution allows the reduction of
bench-work, it limits the assay sensitivity [22,24,28].
Compared to the LCMS/MS assays published so far, reporting
LLOQ between 10 and 720 pg/ml (28 and 2000 pmol/l) for F and
between 180 and 1985 pg/ml (500 pmol/l and 5500 pmol/l) for E
[12,16,2230], our method showed elevated sensitivity of 39.1 pg/
ml both for F (107.9 pmol/l) and E (108.2 pmol/l).
Among several solvents tested during the method development
(data not shown), dichloromethane displayed the best compromise between recovery and matrix effect, further allowing a rapid
and simple chromatographic analysis.
It was largely reported in literature that steroid molecules
display large differences in MS-detection sensitivity. For this
reason, in our previous serum prole method, a large sample
volume was used, combined to massive purication and to the
APCI source less prone to ion suppression phenomena, in order to
achieve proper sensitivity for all the analytes included in the panel
Fig. 1. analytes LCMS/MS chromatographic peak and noise prole showed no differences in real saliva samples obtained by direct spitting and by Salivette1. Cortisol at
0.150 ng/ml by direct spitting (A) and by Salivette1 (B); cortisone at 1.20 ng/ml by direct spitting (C) and by Salivette1 (D).
G Model
(Cortisone Direct Spit - Cortisone Salivette) / Average %
(Cortisol Dir ect Spit - Cortisol Salivette) / Average %
80
60
+1.96 SD
56.2
40
20
M ean
15.9
-20
-1.96 SD
-24.5
-40
1
50
+1.96 SD
45.3
40
30
20
10
0
-10
-1.96 SD
-20
-22.3
-30
0
Average (ng/ml)
14
Cortisone Salivette (ng/ml)
Cortisol Salivette (ng/ml)
r= 0.972
P<0.001
10
12
14
16
Average (ng/ml )
12
r= 0.965
P<0.001
8
6
2
0
Cortisol Direct Spit (ng/ml)
10
12
14
Cortisone DirectSpit (ng/ml)
Fig. 2. Salivette1 determines not different result compared to a direct spitting. Bland and Altman plot for salivary F (A) and E (B). The bold line represents the mean bias
between Salivette1 and in direct spitting saliva samples; dot-and-dash lines represent the zero bias line. Dashed lines represent the 1.96*standard deviation range. Passing
and Bablok regression analysis of Salivette1 vs direct spitting results for salivary F (C) (intercept coefcient (95% CI): 0.0136 (0.08980.0760), slope coefcient (95% CI):
0.858 (0.7131.029)) and salivary E (D) (intercept coefcient (95% CI): 0.156 (0.7910.743), slope coefcient (95% CI): 0.834 (0.6991.062). The bold line represents the
calculated regression equation; dot-and-dash lines represent the 1.96*residuals standard deviation range; the dashed line represents the identity line.
[21]. Among the published LCMS/MS assays dedicated to F, the

majority reported positive ESI ionization [16,2230]. In our study,
we observed 100 fold higher S/N by using negative compared to
positive ionization in ESI source. Moreover, ESI negative mode
revealed preferable than APCI source, whether operated in positive
or in negative mode (data not shown). Negative ionization has the
further advantage to be less prone to adduct formation and to
result in general low baseline noise, thus contributing to assay
robustness and sensitivity [33].
Because of practical reasons, a limitation of the present study is
that our paired serum and saliva sampling design did not include
the late nighttime point. Nonetheless, a few studies reported LC
MS/MS-based LNSF values and cut-off, the latter ranging between
0.761.1 ng/ml (2.12.9 nmol/l) [1216]. These reports indirectly
demonstrated that the sensitivity of our assay perfectly ts the
clinical-diagnostic purposes. Moreover, the method displayed
proper precision and accuracy for a robust determination of overall
salivary glucocorticoid: in more than one thousand individual
samples, we found analyte signals below the sensitivity limit or
with non-specic IR or peak shape in less than 1% of cases. In order
to provide a thorough pre-analytical validation, we tested the
performance of a widely adopted collection device, the cotton
Salivette1, by comparing it to a simple direct spitting procedure in

terms of chromatogram prole, imprecision and agreement of the
results. Indeed, though the Salivette1 device allows easy and
abundant saliva collection free from mucins and oral debris, its
natural cotton swab was previously reported to irreversibly retain
part of the salivary steroids and to interfere in the analyte signal,
thus limiting sensitivity and specicity [17,18]. We observed that
Salivette1 specimens are affected by a small mean negative bias,
below 16% and 12% for F and E, respectively; nonetheless, the
Passing and Bablok analysis revealed the absence of signicant
systematic or proportional bias, along with satisfying correlation
coefcients for both analytes. Furthermore, no interference
phenomena were observed in terms of background noise or peak
shape alteration. The overall imprecision assessed on repeated
measurements proved ideal for Salivette1 specimens ( <0.4%),
while it ranged between 5.2% and 12.6% in direct spitting
specimens. We did not specically investigate the source of higher
variability in the latter; however, mucin debris could hamper the
accurate pipetting of such a low volume. Moreover, a stable
emulsion was often observed, limiting the organic layer recovery
during the extraction [34]. The comparison data, along with data
on recovery, sensitivity and matrix effect, support the use of the
G Model
16
salivary cortisone
(13.7 1.02 )
14
salivary cortisol
12
ng/ml
10
***
(6.29 0.37 )
***
(5.82 0.36 )
***
(4.24 0.57 )
(3.17 0.31 )
***
(0.817 0.070 )
##
***
(0.679 0.071 )
***
(0.475 0.096 )
0
8 am
12 am
4 pm
8 pm
T ime
175
(168.4 8.89)
serum cortisol
serum cortisone
150
125
ng/ml
***
(87.76.24)
100
***
75
***
(55.67.48)
50
25
##
(74.96.28)
(27.3 1.24)
***
##
#
***
(25.31.22)
(20.40.93)
***
(16.11.40)
0
8 8am
1212
am
4
16pm
8 pm
20
T ime
(6.420.55)
Salivary F/E ratio

Serum F/E ratio
s alivary F/E ratio
0.4
**
***
***
**
(3.560.23)
0.3
(3.730.20)
0.2
***
(3.240.24)
(0.2430.014)
***
##
#
(0.1290.007)
***
(0.1150.007)
0.1
##
###
***
(0.1660.007)
0.0
6.0
4.0
2.0
s erum F/E ratio
0.5
0.0
88am
12am
12
416pm
20
8 pm
T ime
Fig. 3. Salivary cortisol and cortisone levels parallel the serum levels during day-time according to circadian variation. Cortisol and cortisone saliva (A) and serum (B)
circadian variation; cortisol vs cortisone (F/E) ratio in serum and in saliva (C) circadian variation. Data are expressed as mean standard error of mean. *p < 0.05 vs 8 am; **
p < 0.01 vs 8 am; *** p < 0.0001 vs 8 am. # p < 0.05 vs 12 am; ## p < 0.01 vs 12 am; ### p < 0.001 vs 12 am.
G Model
S alivary C ortis one (ng/ml)
S alivary C ortis ol (ng/ml)
A
2
R =0.903
P<0.001
50
100
150
200
250
25
20
R =0.854
P<0.001
15
10
5
0
50
20
25
2
R =0.359
P<0.001
15
10
5
0
10
20
30
R =0.194
P<0.001
3
2
1
0
R =0.659
P<0.001
20
10
0
100
150
200
250
S erum C ortis ol (ng/ml)
S erum C ortis one (ng/ml)
25
50
250
10
20
30
40
200
40
30
150

40
100
20
R =0.945
P<0.001
15
10
5
0
0
Fig. 4. Salivary and serum cortisol and cortisone levels are differentially associated. Regression tting line and determination coefcient (R2) between: salivary Fserum F
(A); salivary Eserum F (B); salivary Eserum F (C); serum Esalivary F (D); serum Eserum F (E); salivary Esalivary F (F).
Salivette1 device for salivary F and E measurement. For the

biological validation, we paid particular attention to volunteer
recruitment, by selecting healthy, normal weight, drug-free
volunteers. Moreover, considering the large intra-individual
inter-day variability of F levels [35], in order to not alter the
ordinary activity of the subjects, we recruited volunteers among
the Hospital staff, excluding shift workers or people reporting
irregular wake-sleep cycles. The paired saliva and serum sampling
performed at four times of the day allowed us to describe the
expected circadian glucocorticoid uctuation in both biological
uids. We scheduled the third time point (4 pm) three hours after
the lunch time in order to avoid the postprandial insulindependent rise in F [36]; however, subjects were not controlled
for food intake or macronutrient contents, thus this may represent
a source of variability. Salivary F and E displayed parallel rhythms,
with higher values in the morning and lower values in the evening.
The F/E ratio observed in saliva was opposite to the serum value,
coherently with the type II 11b-HSD activity in the salivary glands.
Moreover, the salivary F/E ratio showed salivary E levels
approximately 4 times higher than salivary F in the morning
and 10 times higher in the evening.
We showed an exponential relationship between salivary and

serum F, thus conrming previous data [1,29,37]. In particular, a
steeper increase in salivary F was detectable at serum F levels
above an approximate concentration of 150 ng/ml (414 nmol/l),
typical of the morning samples. The trend observed in saliva could
reect the non-linear rise in serum free F, the fraction able to
diffuse to saliva uid, resulting from the saturation of the CBG
binding capacity as well as from the contribution to F regeneration
made by the tissue type I 11b-HSD under positive ACTH
modulation [38]. Both serum free E fraction and F-to-E conversion
made by type II 11b-HSD in the salivary glands contribute to
salivary E levels, and it was shown to linearly correlate to total F
fraction independently of the time of day. Our data conrmed
previous reports suggesting salivary E as a preferable indirect
measurement of the systemic glucocorticoid exposure [29].
Moreover, we found that the non-linear relationship between E
and F levels in saliva paralleled the relationship in serum: a steeper
increase in E levels was described at F levels below 2 ng/ml in saliva
and 50 ng/ml in serum, typical of the evening samples, followed by
a slighter E increase above 2 ng/ml of F in saliva and by a plateau in
serum E levels at F concentrations above 100 ng/ml, typically
G Model
occurring in the morning. The described biphasic relationship in

serum is supposed to result from the overall processes of F
inactivation and/or regeneration occurring in a time- and tissuespecic fashion, as keenly characterized in previous reports [39].
Nonetheless, the serum E-to-F trend explains only in part the
biphasic relationship observed in saliva, where the contribution of
salivary 11b-HSD II activity is relevant and it is reduced or
saturated at high salivary F levels. The importance of 11b-HSD II
activity in determining salivary E is also supported by the poor
correlation between salivary E and serum E, and this appeared
largely due to samples with elevated salivary E, above 12 ng/ml,
typical of the morning. A similar association was observed between
salivary F and serum E, thus indicating the poor contribute of
serum E over salivary F and E concentrations.
The overall data stands for the ability of our LCMS/MS assay to
provide an indirect but robust estimate of the free F fraction by
measuring the salivary F and, at the same time, of the overall F
exposure, as reected by the salivary E levels.
In conclusion, by using properly validated LCMS/MS assays for
glucocorticoid measurement in both uids, overcoming the crossreactivity and sensitivity limits of direct immunoassay, we were
able to underline the high biological value of salivary F and the
added complementary value of the simultaneous salivary E testing.
Moreover, we provided a comprehensive description of F and E
diurnal uctuation and of their reciprocal relationship within and
between the two uids. The strength of this study was to provide
paired saliva-serum data of both F and E in four times along the day,
thus obtaining a representative range of physiologic glucocorticoid
levels, without using ACTH stimulation. A limitation of the present
study is that we were only able to study a small representative
population. This implies that the saliva/serum glucocorticoid
dynamic should be carefully investigated in a larger healthy
population as well as in specic diseases in order to clarify the
usefulness of F and E salivary testing in different endocrine
diseases, thus improving research and diagnostic capabilities.
Conict of interest
The authors have nothing to disclose.
Funding
Emilia-Romagna Region, Alessandro Liberati Young Researcher
Grants, PRUA 1-2012-004. Italian Ministry of Education Grants,
University and Research (PRIN 2010C8ERKX).
References
[1] R. Vining, R. McGinley, R. Symons, Hormones in saliva: mode of entry and
consequent implications for clinical interpretation, Clin. Chem. 29 (1983)
17521756.
[2] C.M. Mendel, The free hormone hypothesis: a physiologically based
mathematical model, Endocr. Rev. 10 (1989) 232274.
[3] P. Wood, Salivary steroid assaysresearch or routine? Ann. Clin. Biochem. 46
(2009) 183196, doi:http://dx.doi.org/10.1258/acb.2008.008208.
[4] A. Galbois, M. Rudler, J. Massard, Y. Fulla, A. Bennani, D. Bonnefont-Rousselot,
V. Thibault, S. Reignier, A. Bourrier, T. Poynard, D. Thabut, Assessment of
adrenal function in cirrhotic patients: salivary cortisol should be preferred, J.
Hepatol. 52 (2010) 839845, doi:http://dx.doi.org/10.1016/j.jhep.2010.01.026.
[5] R. Alwani, L. Jongbloed, F. Jong, A. Lely, W. Herder, R. Feelders, Differentiating
between Cushings disease and pseudo-Cushings syndrome: comparison of
four tests, Eur. J. Endocrinol. 170 (2014) 477486, doi:http://dx.doi.org/
10.1530/EJE-13-0702.
[6] F. Ceccato, M. Barbot, M. Zilio, S. Ferasin, G. Occhi, A. Daniele, S. Mazzocut, M.
Iacobone, C. Betterle, F. Mantero, C. Scaroni, Performance of salivary cortisol in
the diagnosis of Cushings syndrome, adrenal incidentaloma, and adrenal
insufciency, Eur. J. Endocrinol. 169 (2013) 3136, doi:http://dx.doi.org/
10.1530/eje-13-0159.
[7] I. Perogamcros, B.G. Keevil, D.W. Ray, P.J. Trainer, Salivary cortisone is a
potential biomarker for serum free cortisol, J. Clin. Endocrinol. Metab. 95
(2010) 49514958, doi:http://dx.doi.org/10.1210/jc.2010-1215.
[8] T. Kido, T.V. Dao, M.D. Ho, N. Duc Dang, N.T. Pham, R. Okamoto, T.T. Pham, S.
Maruzeni, M. Nishijo, H. Nakagawa, S. Honma, S.K. Le, H.N. Nguyen, High
cortisol and cortisone levels are associated with breast milk dioxin
concentrations in Vietnamese women, Eur. J. Endocrinol. 170 (2013) 131139.
[9] F.Z. Stanczyk, Advantages and challenges of mass spectrometry assays for
steroid hormones, J. Steroid Biochem. Mol. Biol. 121 (2010) 491495, doi:
http://dx.doi.org/10.1016/j.jsbmb.2010.05.001.
[10] M. Van Aken, J. Romijn, J. Miltenburg, E. Lentjes, Automated measurement of
salivary cortisol, Clin. Chem. 49 (2003) 14081409, doi:http://dx.doi.org/
10.1373/49.8.1408.
[11] T. Deutschbein, S. Petersenn, Screening for cushings syndrome: new
immunoassays require adequate normative data, Horm. Metab. Res. 45 (2013)
118123, doi:http://dx.doi.org/10.1055/s-0032-1331745.
[12] U. Turpeinen, M. Vlimki, E. Hmlinen, Determination of salivary cortisol
by liquid chromatography-tandem mass spectrometry, Scand. J Clin. Lab.
Invest. 69 (2009) 592597, doi:http://dx.doi.org/10.1080/
00365510902890331.
[13] R. Zerikly, L. Amiri, C. Faiman, M. Gupta, R. Singh, B. Nutter, L. Kennedy, B.
Hatipoglu, R. Weil, A. Hamrahian, Diagnostic characteristics of late-night
salivary cortisol using liquid chromatography-tandem mass spectrometry, J.
Clin. Endocrinol. Metab. 95 (2010) 45554559, doi:http://dx.doi.org/10.1210/
jc.2009-2458.
[14] D. Erickson, R. Singh, A. Sathananthan, A. Vella, S. Bryant, Late-night salivary
cortisol for diagnosis of Cushings syndrome by liquid chromatographytandem mass spectrometry assay, Clin. Endocrinol. (Oxf.) 76 (2012) 467472,
doi:http://dx.doi.org/10.1111/j.13652265.2011.04239.x.
[15] S. Palmieri, V. Morelli, E. Polledri, S. Fustinoni, R. Mercadante, L. Olgiati, C.
Vainicher, E. Cairoli, V. Zhukouskaya, P. Beck-Peccoz, I. Chiodini, The role of
salivarycortisolmeasured by liquidchromatography-tandem mass
spectrometry in the diagnosis of subclinicalhypercortisolism, Eur. J.
Endocrinol. 168 (2013) 289296, doi:http://dx.doi.org/10.1530/EJE-12-0803.
[16] G. Antonelli, F. Ceccato, C. Artusi, M. Marinova, M. Plebani, Salivary cortisol and
cortisone by LCMS/MS: validation, reference intervals and diagnostic
accuracy in Cushings syndrome, Clin. Chim. Acta 451 (2015) 247251, doi:
http://dx.doi.org/10.1016/j.cca.2015.10.004.
[17] M. Grschl, M. Rauh, Inuence of commercial collection devices for saliva on
the reliability of salivary steroids analysis, Steroids 71 (2006) 10971100, doi:
http://dx.doi.org/10.1016/j.steroids.2006.09.007.
[18] E. Shirtcliff, D. Granger, E. Schwartz, M. Curran, Use of salivary biomarkers in
biobehavioral research: cotton-based sample collection methods can interfere
with salivary immunoassay results, Psychoneuroendocrinology 26 (2001)
165173, doi:http://dx.doi.org/10.1016/S0306-4530(00)000421.
[19] J.M. Bland, D.G. Altman, Measuring agreement in method comparison studies,
Stat. Methods Med. Res. 8 (1999) 135160, doi:http://dx.doi.org/10.1177/
096228029900800204.
[20] W. Bablok, H. Passing, R. Bender, B. Schneider, A general regression procedure
for method transformation. Application of linear regression procedures for
method comparison studies in clinical chemistry, part III, J. Clin. Chem. Clin.
Biochem. 26 (1988) 783790.
[21] F. Fanelli, I. Belluomo, V. Lallo, G. Cuomo, R. Iasio, M. Baccini, E. Casadio, B.
Casetta, V. Vicennati, A. Gambineri, G. Grossi, R. Pasquali, U. Pagotto,
Serumsteroidproling by isotopicdilution-liquidchromatographymass
spectrometry: comparison with currentimmunoassays and referenceintervals
in healthyadults, Steroids 76 (2011) 244253, doi:http://dx.doi.org/10.1016/j.
steroids.2010.11.005.
[22] R.L. Jones, L.J. Owen, J.E. Adaway, B.G. Keevil, Simultaneous analysis of cortisol
and cortisone in saliva using XLC-MS/MS for fully automated online solid
phase extraction, J. Chromatogr. B 881882 (2012) 4248, doi:http://dx.doi.
org/10.1016/j.jchromb.2011.11.036.
[23] H. Kataoka, K. Ehara, R. Yasuhara, K. Saito, Simultaneous determination of
testosterone, cortisol, and dehydroepiandrosterone in saliva by stable isotope
dilution on-line in-tube solid-phase microextraction coupled with liquid
chromatography-tandem mass spectrometry, Anal. Bioanal. Chem. 405 (2013)
331340, doi:http://dx.doi.org/10.1007/s00216-012-6479-4.
[24] S. Fustinoni, E. Polledri, R. Mercadante, High-throughput determination of
cortisol, cortisone, and melatonin in oral uid by on-line turbulent ow liquid
chromatography interfaced with liquid chromatography/tandem mass
spectrometry, Rapid Commun. Mass Spectrom. 27 (2013) 14501460, doi:
http://dx.doi.org/10.1002/rcm.6601.
[25] A.F. Ruiter, N. Teeninga, J. Nauta, E. Endert, M.T. Ackermans, Determination of
unbound prednisolone, prednisone and cortisol in human serum and saliva by
on-line solid-phase extraction liquid chromatography tandem mass
spectrometry and potential implications for drug monitoring of prednisolone
and prednisone in saliva, Biomed. Chromatogr. 26 (2012) 789796, doi:http://
dx.doi.org/10.1002/bmc.1730.
[26] M.A. Jensen, A.M. Hansen, P. Abrahamsson, A.W. Nrgaard, Development and
evaluation of a liquid chromatography tandem mass spectrometry method for
simultaneous determination of salivary melatonin, cortisol and testosterone, J.
Chromatogr. B 879 (2011) 25272532, doi:http://dx.doi.org/10.1016/j.
jchromb.2011.07.005.
[27] S. Lee, S. Kwon, H.J. Shin, H.S. Lim, R.J. Singh, K.R. Lee, Y.J. Kim, Simultaneous
quantitative analysis of salivary cortisol and cortisone in Korean adults using
LCMS/MS, BMB Rep. 43 (2010) 506511.
G Model
10
[28] L.J. Owen, S. Haslam, J.E. Adaway, P. Wood, C. Glenn, B.G. Keevil, A simplied
liquid chromatography tandem mass spectrometry assay, using on-line solidphase extraction, for the quantitation of cortisol in saliva and comparison with
a routine DELFIA method, Ann. Clin. Biochem. 47 (2010) 131136, doi:http://
dx.doi.org/10.1258/acb.2009.009053.
[29] I. Perogamvros, L. Owen, J. Newell-Price, D. Ray, P. Trainer, B. Keevil,
Simultaneous measurement of cortisol and cortisone in human saliva using
liquid chromatography-tandem mass spectrometry: application in basal and
stimulated conditions, J. Chromatogr. B 877 (2009) 37713775, doi:http://dx.
doi.org/10.1016/j.jchromb.2009.09.014.
[30] F. Matsui, E. Koh, K. Yamamoto, K. Sugimoto, H.S. Sin, Y. Maeda, S. Honma, M.
Namiki, Liquid chromatography-tandem mass spectrometry (LCMS/MS)
assay for simultaneous measurement of salivary testosterone and cortisol in
healthy men for utilization in the diagnosis of late-onset hypogonadism in
males, Endocr. J. 56 (2009) 10831093.
[31] A.H. Garde, A.M. Hansen, Long-term stability of salivary cortisol, Scand. J. Clin.
Lab. Invest. 65 (2005) 433436.
[32] G.L. Whembolua, D.A. Granger, S. Singer, K.T. Kivlighan, J.A. Marguin, Bacteria
in the oral mucosa and its effects on the measurement of cortisol
dehydroepiandrosterone, and testosterone in saliva, Horm. Behav. 49 (2006)
478483.
[33] U. Turpeinen, E. Hmlinen, Determination of cortisol in serum, saliva and
urine, Best Pract. Res. Clin. Endocrinol. Metab. 27 (2013) 795801, doi:http://
dx.doi.org/10.1016/j.beem.2013.10.008.
[34] R. John Dean, Extraction Techniques in Analytical Sciences, John Wiley & Sons,
2009, doi:http://dx.doi.org/10.1002/9780470682494 (Published Online 29
DEC).
[35] S.C. Segerstrom, I.A. Boggero, G.T. Smith, S.E. Sephton, Variability and
reliability of diurnal cortisol in younger and older adults: implications for
design decisions, Psychoneuroendocrinology 49 (2014) 299309, doi:http://
dx.doi.org/10.1016/j.psyneuen.2014.07.022.
[36] R. Stimson, N. Mohd-Shukri, J. Bolton, R. Andrew, R. Reynolds, B. Walker, The
postprandial rise in plasma cortisol in men is mediated by macronutrientspecic stimulation of adrenal and extra-adrenal cortisol production, J. Clin.
Endocrinol. Metab. 99 (2014) 160168, doi:http://dx.doi.org/10.1210/jc.20132307.
[37] I. Perogamvros, L. Owen, B. Keevil, G. Brabant, P. Trainer, Measurement of
salivary cortisol with liquid chromatography-tandem mass spectrometry in
patients undergoing dynamic endocrine testing, Clin. Endocrinol. (Oxf.) 72
(2010) 1721, doi:http://dx.doi.org/10.1111/j.1365-2265.2009.03582.x.
[38] M. Vogeser, R. Zachoval, K. Jacob, Serum cortisol/cortisone ratio after
Synacthen stimulation, Clin. Biochem. 34 (5) (2001) 421425, doi:http://dx.
doi.org/10.1016/s0009-9120(01)00251-X.
[39] J.R. Seckl, B.R. Walker, 11beta-hydroxysteroid dehydrogenase type I as a
modulator of glucocorticoid action: from metabolism to memory, Trends
Endocrinol. Metab. 15 (2004) 418424, doi:http://dx.doi.org/10.1016/j.
tem.2004.09.007.

1 s2.0 S096007601630108X Main

Uploaded by

Copyright:

Available Formats

You might also like

1 s2.0 S096007601630108X Main

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 s2.0 S096007601630108X Main

Uploaded by

Copyright:

Available Formats

G Model

SBMB 4705 No. of Pages 10

Journal of Steroid Biochemistry & Molecular Biology xxx (2015) xxxxxx

Contents lists available at ScienceDirect

Journal of Steroid Biochemistry & Molecular Biology

Validation of an LCMS/MS salivary assay for glucocorticoid status

* Corresponding author at: Endocrinology Unit, Department of Medical and

protein-bound fraction in plasma. However, the estimation of free F

obtained by a MilliQ Gradient A10 System (Millipore, Volketswil,

80 C. Three aliquots of each QC level were thawed and processed

accuracy between 80 and 120% of the expected value and an

2.6. Salivary cortisol and cortisone LCMS/MS method validation

The limit of detection (LOD) was determined by injecting scalar

Spiked Concentration in Saliva

Functional sensitivity in saliva matrix.

occurring in the analyte elution window. Specicity was assessed

2.8. Serum cortisol and cortisone LCMS/MS measurement

Regression analysis performed on the overall paired samples

(Cortisone Direct Spit - Cortisone Salivette) / Average %

(Cortisol Dir ect Spit - Cortisol Salivette) / Average %

Cortisone Salivette (ng/ml)

Cortisol Salivette (ng/ml)

Cortisol Direct Spit (ng/ml)

Cortisone DirectSpit (ng/ml)

[21]. Among the published LCMS/MS assays dedicated to F, the

Salivette1, by comparing it to a simple direct spitting procedure in

Salivary F/E ratio

s alivary F/E ratio

s erum F/E ratio

S alivary C ortis one (ng/ml)

S alivary C ortis ol (ng/ml)

S alivary C ortis ol (ng/ml)

S alivary C ortis one (ng/ml)

S erum C ortis ol (ng/ml)

S alivary C ortis one (ng/ml)

S erum C ortis one (ng/ml)

S erum C ortis one (ng/ml)

S erum C ortis one (ng/ml)

S erum C ortis ol (ng/ml)

S erum C ortis ol (ng/ml)

S alivary C ortis ol (ng/ml)

Salivette1 device for salivary F and E measurement. For the

We showed an exponential relationship between salivary and

occurring in the morning. The described biphasic relationship in

You might also like