Journal of Behavioral and Brain Science, 2011, 1, 194-198
doi:10.4236/jbbs.2011.13026 Published Online August 2011 (http://www.SciRP.org/journal/jbbs)
Sampling Small Volumes of Saliva for Determination of the
Str ess Hormone α-Amylase. A Comparative
Methodologica l Study
Aristidis Arhakis, Vassilis Karagiannis, Sotirios Kalfas
School of Dentistry, Aristotle University of Thessaloniki, Thessaloniki, Greece
Received May 5, 2011; revised June 17, 2011; accepted July 5, 2011
Abstract
Salivary alpha amylase was taken into account as an important physiological indicator in psychoneuroendo-
crinological research focusing on stress. Two sampling devices that allow saliva collection through absorp-
tion to a cotton roll (Salivette®-method) or to small cotton pellets (VectaSpinTM Micro [VSM]-method) were
studied. Any loss of salivary alpha-amylase (sAA) activity in relation to the saliva volume absorbed and har-
vested by centrifugation was examined. A pooled saliva sample prepared from stimulated whole saliva (col-
lected by drooling) of 30 subjects was used. Three different saliva volumes (2.9 ml, 1.5 ml, and 0.8 ml) were
tested on cotton rolls and two (0.03 ml, and 0.015 ml) on cotton pellets. The sample sAA activity was deter-
mined from the hydrolysis of 2-chloro-4-nitrophenyl-α-D-maltotrioside. In comparison with the original
drooling sample, no sAA loss was observed in 1.5 ml samples tested with Salivette, while a significant de-
crease of activity was recorded with smaller volumes. VSM collected samples showed a non-volume de-
pendent decrease of sAA activity of about 25%. Salivette requires large saliva volumes to allow an accurate
sAA estimation. With cases of limited saliva access, VSM may be a suitable sampling device.
Keywords: Salivary Alpha Amylase, Enzyme Activity, Method, Saliva Sampling
1. Introduction
Saliva is of great clinical significance as a diagnostic
fluid and analysis of its components may provide infor-
mation on local or systemic diseases and disorders [1-5].
Sampling saliva is clinically more advantageous com-
pared to other body fluids since it is performed with non-
invasive techniques without the stressful exposure to
needles [6-9]. Whole saliva can be sampled by either the
draining method, the spitting method, the suction method,
or the absorbent (swab) method [10].
A device under the brand name Salivette® (Sarstedt,
Numbrecht, Germany) is frequently employed for whole
saliva collection by the absorbent method [11]. The com-
ponents of the devise are a cotton roll, and a plastic cen-
trifuge tube with an inner cylindrical strainer (Figure 1).
The cotton roll is placed in the mouth for up to 5 min to
absorb saliva till the roll becomes well soaked (about 3
ml saliva). The saliva is then harvested by centrifuging
the roll in the tube with the strainer.
Salivette® may be difficult or less suitable to use under
certain circumstances. Soaking the cotton roll with saliva
requires time and the presence of a certain saliva volume
in the mouth. To enable saliva collection, many research-
ers ask the subjects to chew the cotton roll; however, this
is unsuitable when unstimulated saliva has to be col-
lected. Occasionally, young children may find the chew-
ing procedure cumbersome and refuse to co-operate.
Repeated short interval-sampling may suffer by a limited
amount of saliva absorbed to the roll, which could affect
the composition of the final sample. Indeed, an error vari-
ance in the measurement of cortisol, testosterone, DHEA,
estradiol, progesterone, and salivary α-amylase (sAA), has
been reported, particularly with low sample volumes [12-
14].
The concentration of sAA in saliva is considered as an
indicator of the sympathetic adrenal medullar system
activation due to stress [15-17]. Saliva for sAA assess-
ment has been sampled with Salivette® or by the passive
drooling technique. However, these techniques may not
be feasible when small sample volumes are to be ex-
pected, as mentioned above. To overcome this problem,

A. ARHAKIS ET AL.
195
(a)
(b)
Figure 1. The sampling devises Salivette® (a) and Vec-
taSpinTM Micro (b) used in the study. From the left to the
right, the centrifugation tube, the strainer and the saliva
absorbent material (cotton roll or pellets) for each device
are seen.
we tried a micro-method that allows sampling of small
saliva quantities, using the centrifugation device Vec-
taSpinTM Micro (Whatman, Middlesex, UK) and small
cotton pellets for rapid absorption of saliva from the
mouth (Figure 1). The suitability of the two absorbent
methods for sampling saliva for sAA assessment was
evaluated in this study. In particular, any loss of sAA
activity in relation to the saliva volume absorbed was ex-
amined.
2. Methods
2.1. Saliva Sample
A pooled whole saliva sample from 30 adults was used
throughout the experiment [18]. The subjects were asked
to refrain from drinking, eating any food [19] physically
exercising [20] and from smoking [21] two hours prior to
sampling. They rinsed their mouths once with tap water
and started collecting unstimulated saliva, by passive
drooling, 15 min later. The saliva was allowed to drip off
the lower lip into a graduated test tube fitted with a fun-
nel as described [10,14] for 15 - 20 min.
From each subject, a saliva sample of approximately 5
ml was collected. The samples were centrifuged (10,000
x g, 5 min) to remove particles and kept at –80˚C [17].
On the day of the experiment, the samples were thawed
and mixed together to obtain the required volume of
pooled saliva.
2.2. Experimental Design
Saliva from the pooled sample was pipetted to cotton
rolls of Salivette® and to cotton pellets used with Vec-
taSpinTM Micro (VSM). Three different amounts (2.9 ml,
1.5 ml, and 0.8 ml) of saliva were tested on cotton rolls
(15 rolls with each volume) and two amounts (0.03 ml,
and 0.015 ml) on cotton pellets (150 pellets with each
volume).
Each Salivette cotton roll was placed in the strainer of
the centrifuge tube and saliva was carefully pipetted on
the top of the roll till the determined volume was applied.
At its maximum, a cotton roll can absorb about 3 ml. The
saliva was harvested from the roll to the bottom of the
tube by centrifugation (x g, 20 min).
Each cotton pellet (Roeko etching pellets, medium
size,  3 mm, ref 230 002) used in the VSM method can
maximally absorb 0.03 ml. A volume of ≥0.05 ml final
sample was required for each measurement. To obtain
this volume, 5 cotton pellets were placed in the strainer
and the corresponding amount of saliva (5 × 0.03 ml or 5
× 0.015 ml) was applied on them. After centrifugation
(10,000 x g, 5 min), the saliva was harvested in the bot-
tom of the tube.
From the pooled saliva, 15 aliquots of 0.01 ml each
were aspirated and analyzed for comparative purpose.
These samples are referred in the text as the drooling
samples. The experiment was accomplished on two oc-
casions, with one third of the samples of each method
being examined on the first occasion and the rest of them
on the second occasion, 3 weeks later. All saliva samples
were kept at 6˚C until analyzed for sAA activity (within
2 h). Immediately before the analysis, each sample was
diluted 200 times with distilled water.
2.3. Measurement of sAA Activity
The activity of sAA (in U/ml sample) was determined
from the hydrolysis of the chromogenic substrate 2-chloro-
4-nitrophenyl-α-D-maltotrioside [22]. A volume of 0.6
ml freshly prepared substrate solution (5 mM 2-chloro-4-
nitrophenyl-α-D-maltotrioside in reaction buffer) was
mixed with 0.05 ml diluted sample in a cuvette. The hy-
drolysis of the substrate was spectrophotometrically fol-

A. ARHAKIS ET AL.
196
lowed at 405 nm, through the production of the coloured
compound 2-chloro-p-nitrophenol. The reaction buffer
contained 0.03% bovine albumin, 0.06% CaCl2, 0.48%
KSCN, and 0.03% NaN3 in 50 mM 2-(N-morpholino)
ethanesulfonic acid [23]. The pH of the buffer was ad-
justed to 6.0 with 5 M KOH solution, since, in a pre-
liminary experiment, the sAA activity was found to be
higher at pH 6.0 than at pH 7.0.
One U/ml of sAA was defined as the amount of en-
zyme that hydrolyzes 1 μmol substrate per min [24]. The
molar absorptivity of 2-chloro-p-nitrophenol was found
to be 14,6 l/mol·cm, measured with solutions of pure
2-chloro-p-nitrophenol (concentration range 0.0025 - 0.5
mM) under the present experimental conditions. Purified
α-amylase type XIII (product No. A1031) was used as
the positive control. All experiments were done at room
temperature. The chemicals were purchased from Sigma-
Aldrich Chemie Gmbh, Switzerland.
2.4. Statistical Analysis
Based on the results from a pilot study with 6 samples in
each group, a power analysis showed that with a = 0.05
and N = 90 (equal size groups), the null hypothesis of the
equality of means under a one-way anova model would
be rejected with 99% probability. Accordingly, the sizes
of the samples for Salivette, VSM, and drooling methods
were determined to be 15, 30, and 15, respectively.
The assumptions of normality and equality of vari-
ances where verified by the Sapiro-Wilk and Levene’s
Test, respectively. The statistics Min, Max, Mean, SD,
and 95% CI were used to describe the outcome variable.
The main hypothesis of equality of means was evaluated
by the one-way Anova model, while pair wise compari-
sons were conducted by the Tukey’s HSD method. The
overall analysis was performed with SPSS 15.0 software
and the level of statistical significance was set at p <
0.05.
3. Results
The descriptive statistics for the sAA activity of the sam-
ples obtained by the three methods are presented in Ta-
ble 1. The hypothesis concerning the equality of means
of the 6 groups were rejected (F(5,114) = 7.094, p <
0.001, one way ANOVA). The results of the pair wise
comparisons with Tukey’s HSD test are shown in Figure
2. The statistically significant differences in sAA activity
are indicated in the same figure. No significant differ-
ences were observed between the groups: a) Salivette 2.9
ml and 1.5 ml (p = 0. 582), b) Salivette 2.9 ml and
drooling (p = 0. 979), c) Salivette 1.5 ml and VSM 0.03
ml (p = 0.2), d) Salivette 1.5 ml and VSM 0.015 ml (p =
Table 1. Descriptive statistics for sAA activity (in U/ml)
detected in saliva samples obtained with the three methods.
Method Saliva volume
absorbed (ml)
Sample
number Mean ± SD Range (Min – Max)
2.9 15 14.4 ± 3.4 7.7 – 22.4
1.5 15 12.5 ± 3.2 7.7 – 17.9Salivette®
0.8 15 9.2 ± 4.0 2.9 – 14.3
0.03 30 10.1 ± 3.5 3.8 – 16.4
VSM
0.015 30 10.4 ± 2.5 4.1 – 13.6
Drooling- 15 13.5 ± 3.3 9.2 – 21.8
Figure 2. Mean (rectangular) and 95% Confidence Interval
(bar) for SA activity of saliva samples obtained with the
three methods. The statistically significant differences are
indicated with asterisks (*: p < 0.5, **: p < 0.01, ***: p <
0.001).
0. 360), e) Salivette 1.5 ml and drooling (p = 0. 942), f)
Salivette 0.8 ml and VSM 0.03 ml (p = 0. 943), g) Sali-
vette 0.8 ml and VSM 0.015 ml (p = 0.818), and h) VSM
0.03 ml and 0.015 ml (p = 0.998). The difference be-
tween Salivette 1.5 ml and 0.8 ml was marginally non-
significant (p = 0.067).
4. Discussion
It is evident from the present results that the saliva vol-
ume absorbed by the cotton roll may influence the sAA
concentration detected in the final sample. To prevent
significant loss of sAA, the cotton roll of Salivette has to
be soaked with approximately ≥1.5 ml saliva, which
equals to about ≥50% of the volume that the roll can
maximally absorb. Decreasing the saliva volume to 0.8
ml/roll resulted in a statistically lower sAA activity in
comparison with the other two volumes tested, i.e. 2.9
and 1.5 ml, in the Salivette group as well as the sample
collected by drooling.
Salivette has been used for sampling saliva in previous
studies evaluating stress through sAA activity [11,17].
These authors provided no information for the possible

A. ARHAKIS ET AL.
197
effect of a partially soaked cotton roll on the estimation
of sAA activity, probably, because they did not face the
limitation of a small sample volume. Under certain cir-
cumstances, it may be difficult to obtain the saliva amount
needed for saturation of the roll. Young children may not
be capable or even refuge to provide the required volume.
Occasionally, the size of the roll may be disproportional
to the small mouth of the child, or the procedure itself
may seem unbearable or stressful for a subject who must
accept the cotton roll in the mouth for several minutes.
Furthermore, repeated sampling with short intervals, e.g.
<1 min, is unfeasible, since the flow rate of stimulated
saliva is usually <1.5 ml/min.
The micromethod VSM helps to overcome these dis-
advantages and it is in fact a modification of the Salivette
method. The voluminous cotton roll was replaced by
small cotton pellets and the VectaSpinTM centrifugation
devise was used to harvest the sample. VSM allows a
more rapid collection, about 2 - 3 seconds per pellet, the
procedure being more comfortable for the subject and
demanding a minimum of cooperation. Five saliva-soaked
pellets provide a sufficient sample volume for sAA
analysis.
The sAA activity detected in VSM-collected samples
was significantly lower than in the corresponding sam-
ples collected by drooling or by Salivette-cotton roll
soaked with ≥1.5 ml. Compared to the drooling sample, a
25% and 23% mean loss of sAA activity were found for
the volumes 0.03 and 0.015 ml, respectively. The corre-
sponding mean values for the Salivette-collected samples
of 0.8 and 1.5 ml were 33% and 7%, respectively. Inter-
estingly, a higher variation was recorded in sAA activity
among the Salivette 0.8 ml samples (Coefficient of va-
riation 44%) than among the VSM samples of 0.03 or
0.015 ml (Coefficient of variation 35% and 24%, respec-
tively).
Although not proved, the loss of sAA activity is
probably due to a partial absorption of the enzyme by the
plastic surfaces of the devise and by the cotton. The ex-
tent of protein absorption to a surface is influenced by
the chemical composition of the protein-containing solu-
tion and the physicochemical properties of the surface.
The protein composition of saliva varies depending on
the serous or mucous character of the secretory gland. It
may also vary among different subjects. A high salivary
content of proteins competing with sAA absorption pre-
vents extensive loss of sAA and vice versa. To avoid
confounding results due to the intra-subject variation, a
pooled saliva sample collected from several subjects,
was used throughout this study. Under these conditions,
the rich protein supply provided by saliva volumes of
≥1.5 ml, seems to sufficiently prevent sAA loss, proba-
bly by blocking the protein binding sites in the cotton
roll and tube walls of Salivette. On the contrary, the pro-
tein content in the 100 times smaller saliva volume ob-
tained by the VSM method, appears inadequate to inhibit
sAA absorption to the devise surfaces. The same phe-
nomenon probably occurs with the 0.8 ml volume in
Salivette.
The saliva obtained after centrifugation with Salivette
or VSM is less turbid and viscous than the initial fluid.
This indicates that the cotton roll and pellets retained
particles, such as microbes, epithelial cells, and food de-
bris, that confer turbidity and also macromolecules, such
as mucins that are mainly responsible for the high viscos-
ity of saliva. Removal of particles was also achieved by
the initial centrifugation step, performed before pooling
the individual saliva samples together. The clarified and
less viscous saliva obtained by centrifugation and filtra-
tion may, occasionally, be advantageous for the follow-
ing handling and analysis steps. However, these proce-
dures may also affect the compound under test. Removal
of microbes also removes sAA bound to the cell surface
of certain oral bacterial species [25]. On the other hand,
the cotton roll of Salivette seems to retain most of sali-
vary mucins but not sAA when adequately soaked with
saliva.
To conclude, the use of Salivette for collecting saliva
samples yields no loss in sAA activity, provided that the
sample volume is ≥1.5 ml. When the available saliva
volume is small (<1 ml), Salivette cannot be used, how-
ever, the micromethod VSM described above may pro-
vide a good estimation of sAA activity, the loss being
about 20% in the volume range 0.03 - 0.015 ml. Since
sAA is a potential indicator of autonomic response to
stress a reliable sampling method is required. These find-
ings together with the above outlined advantages make
the micromethod a suitable choice that may enable saliva
sampling under certain circumstances.
5. References
[1] D. J. Adam, A. A. Milne, S. M. Evans, J. E. Roulston, A.
J. Lee, C. V. Ruckley and A. W. Bradbury, “Serum Amy-
lase Isoenzymes in Patients Undergoing Operation for
Ruptured and Non-Ruptured Abdominal Aortic Aneu-
rysm,” Journal of Vascular Surgery, Vol. 30, No. 2, 1999,
pp. 229-235. doi:10.1016/S0741-5214(99)70132-1
[2] P. M. Martinez, A. R. Torres, R. Ortiz de Lejarazu, A.
Montoya, J. F. Martin and J. M. Eiros, “Human Immu-
nodeficiency Virus Antibody Testing by Enzyme-Linked
Fluorescent and Western Blot Assays Using Serum, Gin-
gival-Crevicular Transudate, and Urine Samples,” Jour-
nal of Clinical Microbiology, Vol. 37, No. 4, 1999, pp.
1100-1106.
[3] E. Filaire and G. Lac, “Dehydroepiandrosterone (DHEA)
Rather than Testosterone Shows Saliva Androgen Re-

A. ARHAKIS ET AL.
198
sponses to Exercise in Elite Female Handball Players,”
International Journal of Sports Medicine, Vol. 21, No. 1,
2000, pp. 17-20. doi:10.1055/s-2000-8851
[4] S. P. Humphrey and R. T. Williamson, “A Review of
Saliva: Normal Composition, Flow, and Function,” Jour-
nal of Prosthetic Dentistry, Vol. 85, No. 2, 2001, pp. 162-
169. doi:10.1067/mpr.2001.113778
[5] C. F. Streckfus and L. R. Bigler, “Saliva as a Diagnostic
Fluid,” Oral Diseases, Vol. 8, No. 2, 2002, pp. 69-76.
doi:10.1034/j.1601-0825.2002.1o834.x
[6] D. B. Ferguson, “Current Diagnostic Uses of Saliva,”
Journal of Dental Research, Vol. 66, No. 2, 1987, pp. 420-
424. doi:10.1177/00220345870660020601
[7] D. Malamud, “Saliva as a Diagnostic Fluid,” British
Medical Journal, Vol. 305, No. 6847, 1992, pp. 207-208.
doi:10.1136/bmj.305.6847.207
[8] I. D. Mandel, “A Contemporary View of Salivary Re-
search,” Critical Reviews in Oral Biology & Medicine,
Vol. 4, No. 3-4, 1993, pp. 599-604.
[9] H. C. Slavkin, “Toward Molecularly Based Diagnostics
for the Oral Cavity,” Journal of the American Dental As-
sociation, Vol. 129, No. 8, 1998, pp. 1138-1143.
[10] M. Navazesh, “Methods for Collecting Saliva,” Annals of
the New York Academy of Sciences, Vol. 20, No. 694,
1993, pp. 72-77.
doi:10.1111/j.1749-6632.1993.tb18343.x
[11] N. Rohleder and U. M. Nater, “Determinants of Salivary
Alpha-Amylase in Humans and Methodological Consid-
erations,” Psychoneuroendocrinology, Vol. 34, No. 4,
2009, pp. 469-485. doi:10.1016/j.psyneuen.2008.12.004
[12] E. A. Shirtcliff, D. A. Granger, E. Schwartz and M. J.
Curran, “Use of Salivary Biomarkers in Biobehavioral
Research: Cotton-Based Sample Collection Methods Can
Interfere with Salivary Immunoassay Results,” Psycho-
neuroendocrinology, Vol. 26, No. 2, 2001, pp. 165-173.
doi:10.1016/S0306-4530(00)00042-1
[13] A. G. Harmon, L. C. Hibel, O. Rumyantseva and D. A.
Granger, “Measuring Salivary Cortisol in Studies of Child
Development: Watch out-What Goes in May Not Come
out of Saliva Collection Devices,” Developmental Psy-
chobiology, Vol. 49, No. 5 2007, pp. 495-500.
doi:10.1002/dev.20231
[14] J. A. DeCaro, “Methodological Considerations in the Use
of Salivary a-Amylase as a Stress Marker in Field Re-
search,” American Journal of Human Biology, Vol. 20,
No. 5, 2008, pp. 617-619. doi:10.1002/ajhb.20795
[15] J. A. Bosch, H. S. Brand, T. J. Ligtenberg, B. Bermond, J.
Hoogstraten and A. V. Nieuw Amerongen, “Psychologi-
cal Stress as a Determinant of Protein Levels,” Psycho-
somatic Medicine, Vol. 58, No. 4, 1996, pp. 374-382.
[16] U. M. Nater, R. La Marca, L. Florin, A. Moses, W. Lang-
hans, M. M. Koller and U. Ehlert, “Stress-Induced
Changes in Human Salivary Alpha-Amylase Activity-
Associations with Adrenergic Activity,” Psychoneuroen-
docrinology, Vol. 31, No. 1, 2006, pp. 49-58.
doi:10.1016/j.psyneuen.2005.05.010
[17] D. A. Granger, K. T. Kivlighan, M. el-Sheikh, E. B. Gordis
and L. R. Stroud, “Salivary Alpha-Amylase in Biobehav-
ioral Research: Recent Developments and Applications,”
Annals of the New York Academy of Sciences, Vol. 1098,
2007, pp. 122-144. doi:10.1196/annals.1384.008
[18] M. Neu, M. Goldstein, D. Gao and M. L. Laudenslager,
“Salivary Cortisol in Preterm Infants: Validation of a
Simple Method for Collecting Saliva for Cortisol Deter-
mination,” Early Human Development, Vol. 83, No. 1,
2007, pp. 47-54. doi:10.1016/j.earlhumdev.2006.04.003
[19] S. L. Udupa, A. R. Prabhakar and S. Tandon, “Al-
pha-Amylase Inhibitors in Foodstuffs,” Food Chemistry,
Vol. 34, No. 2, 1989, pp. 95-101.
doi:10.1016/0308-8146(89)90077-0
[20] K. T. Kivlighan and D. A. Granger, “Salivary Alpha-
Amylase Response to Competition: Relation to Gender,
Previous Experience, and Attitudes,” Psychoneuroendo-
crinology, Vol. 31, No. 6, 2006, pp. 703-714.
doi:10.1016/j.psyneuen.2006.01.007
[21] R. Nagler, S. Lischinsky, E. Diamond, N. Drigues, I.
Klein and A. Z. Reznick, “Effect of Cigarette Smoke on
Salivary Proteins and Enzyme Activities,” Archives of
Biochemistry and Biophysics, Vol. 379, No. 2, 2000, pp.
229-236. doi:10.1006/abbi.2000.1877
[22] E. S. Winn-Deen, H. David, G. Sigler and R. Chavez,
“Development of a Direct Assay for Alpha-Amylase,”
Clinical Chemistry, Vol. 34, No. 10, 1988, pp. 2005-2008.
[23] A. Y. Foo and R. Bais, “Amylase Measurement with 2-
Chloro-4-Nitrophenyl Maltotrioside as Substrate,” Cli-
nica Chimica Acta, Vol. 272, No. 2, 1998, pp. 137-147.
doi:10.1016/S0009-8981(98)00009-6
[24] H. Ben-Aryeh, M. Fisher, R. Szargel and D. Laufer,
“Composition of Whole Unstimulated Saliva of Healthy
Children: Changes with Age,” Archives of Oral Biology,
Vol. 35, No. 11, 1990, pp. 929-931.
doi:10.1016/0003-9969(90)90075-L
[25] M. Kilian and B. Nyvad, “Ability to Bind Salivary Al-
pha-Amylase Discriminates Certain Viridans Group Strep-
tococcal Species,” Journal of Clinical Microbiology, Vol.
28, No. 11, 1990, pp. 2576-2577.