Salivary a-amylase stress reactivity across different age
JANA STRAHLER,aANETT MUELLER,aFRANZISKA ROSENLOECHER,a
CLEMENS KIRSCHBAUM,aand NICOLAS ROHLEDERb
aDepartment of Psychology, Technische Universita ¨ t Dresden, Dresden, Germany
bDepartment of Psychology, Brandeis University, Waltham, MA, USA
Salivary a-amylase (sAA) increases rapidly in response to psychosocial stress in young adults, but no direct com-
parisons between different age groups across the life span have been made. Secretion of sAA and cortisol was assessed
in children, young adults, and older adults after exposure to the Trier Social Stress Test. Additionally, cardiovascular
activity was measured in both adult groups. Older adults showed attenuated sAA, heart rate (HR), and heart rate
variability (HRV) responses. Furthermore, we found higher sAA but lower cortisol at baseline as well as lower sAA
and cortisol responses in children. Age ? sex interactions were observed only for cortisol with higher responses in
older male participants. No associations between the parameters were found. These results implicate sAA as an
alternative or additional sympathetic stress marker throughout the life span, with marked and rapid stress respon-
siveness in three relevant age groups.
Descriptors: Salivary alpha-amylase, Sympathetic nervous system, Acute stress, Age, Children, Older adults
Physiological stress responses are important determinants of
healthanddisease (McEwen & Stellar, 1993). Sincevulnerability
to disease and disease prevalence patterns change with age, it is
groups. So far, the main focus has been on hypothalamus-pitu-
itary-adrenal (HPA) axis reactivity and salivary cortisol, while
there is a relative lack of knowledge regarding sympathetic ner-
vous system (SNS) markers. One potential reason is that SNS
reactivity has been more difficult to assess than HPA axis mark-
ers. Recently, salivary a-amylase (sAA) has been suggested as a
non-invasive saliva based marker for SNS activity (see Nater &
Rohleder, 2009 and Rohleder & Nater, 2009, forasummary). In
outthe lifespan, weset out toestablishsAA responsepatterns to
span. We chose childhood between 6 and 10 years, because it is a
sensitive period of growth and development, and thus plays an
important role for later life health. For example, an association
between childhood health and morbidity in later life was found
for cancer, lungdisease, cardiovascularconditions, and arthritis/
rheumatism (Blackwell, Hayward, & Crimmins, 2001). Young
adults were included to represent the most studied human age
group as a reference. Older adults between 59 and 61 years were
investigated, because at this age the course is set for the further
development of a person’s health in later life, and because au-
tonomic stress responses in older age might be important deter-
minants of cardiovascular and inflammatory aging.
Acute stressresponses of salivary sAA have beenproposed as
markers for sympathetic stress responses. Since Chatterton,
Vogelsong, Lu, Ellman, and Hudgens (1996) reported a signifi-
cant correlation between sAA, adrenaline (A) and noradrenaline
(NA) after physical exercise, several studies included sAA as an
indicator of physiological and psychological stress. Psychosocial
stress has been shown to induce a rapid increase in sAA activity
(Nater, Rohleder, Gaab, Berger, Jud, et al., 2005). However,
looking at the association between stress-induced cortisol, sAA,
A, and NA release, correlations could not be found consistently
(Nater, La Marca, Florin, Moses, Langhans, et al., 2006;
Rohleder, Nater, Wolf, Ehlert, & Kirschbaum, 2004). Pharma-
cological studies (Ehlert, Erni, Hebisch, & Nater, 2006; Van
Stegeren, Rohleder, Everaerd, & Wolf, 2006) provide direct ev-
idence for the relevance of central autonomic mechanisms on
sAA release. Ehlert et al. (2006) suggested that sAA might be an
indirect indicator of the central (noradrenergic) sympathetic,
system which is not necessarily associated with peripheral cat-
echolamine release. Despite the increasingnumberof researchers
integrating sAA into their study designs, nearly all previous data
concerning psychosocial stress-induced sAA responses stems
from studies with children, youth, and young adults (Gordis,
Granger, Susman, & Trickett, 2006, 2008; Nater et al., 2005;
Stroud, Foster, Papandonatos, Handwerger, Granger, et al.,
2009; Van Stegeren, Wolf, & Kindt, 2008), while no studies have
looked at older populations. Therefore, we need to know how
healthy older adults’ sAA levels respond to stress and how
We thank Mrs. Arnold, who analyzed the saliva samples.
Address reprint requests to: Jana Strahler, TU Dresden, 01062
Dresden, Germany. E-mail: firstname.lastname@example.org
Psychophysiology, 47 (2010), 587–595. Wiley Periodicals, Inc. Printed in the USA.
Copyright r 2010 Society for Psychophysiological Research
comparable these responses are to that of children and young
In addition to age differences, research on another stress sys-
tem, i.e., the hypothalamus-pituitary-adrenal (HPA) axis, pro-
Kirschbaum, 2004a). Reanalyzing stress-induced HPA axis re-
sponses in age groups ranging from children to older adults,
Kudielka et al. (2004a) reported significant responses to a psy-
chosocial stress task in all age groups. While no sex differences
were found in children and younger adults, older men displayed
larger free salivary cortisol increases than women as well as the
otherage groups (Kudielka etal., 2004a). Again, nostudieshave
yet investigated age ? sex interactions in sAA stress responses.
However, ageandsexdifferenceshavebeen foundforvarious
other sympathetic markers, mainly heart rate (HR) and heart
rate variability (HRV). HRV is often used as an index of auto-
nomic function (Bigger, Fleiss, Steinman, Rolnitzky, Schneider,
& Stein, 1995). Taken together, previous work has established
that the primary influence of aging on the human SNS is an
elevation in tonic activity rather than responsiveness to stress
with aging depending on the measure (Lavi, Nevo, Thaler,
Rosenfeld, Dayan, et al., 2006; White, Courtemanche, Stewart,
Talajic, Mikes, et al., 1997), and parameters are also sex depen-
dent with values for young females being lower than those for
age-matched male participants. However, some authors did not
find any age ? sex interactions in HRV reactivity (Kunz-
Ebrecht, Mohamed-Ali, Feldman, Kirschbaum, & Steptoe,
2003; Mezzacappa, Kelsey, Katkin, & Sloan, 2001; Wright,
O’Donnell, Brydon, Wardle, & Steptoe, 2007). Furthermore,
there appears to be an age ? sex interaction in HR reactivity.
While young women displayed stronger HR stress responses
Kirschbaum, 2004b), older men and women were comparable in
their acute HR stress response (Kudielka et al., 2004b). Finally,
adrenaline (A) andnoradrenaline(NA).Thereisevidenceforsex
differences in adrenal medullary catecholamine stress reactivity,
with lower responses in women (Honojosa-Laborde, Chapa,
Lange, & Haywood, 1999; Matthews, Gump, & Owens, 2001).
With aging, higher absolute levels of NA were found (White et
increase in parallel with aging whereby age-related differences in
NA levels are markedly attenuated in female participants (White
et al, 1997). Overall, despite age-related changes of catechola-
mine and HRV baseline activity (White et al., 1997), respon-
siveness to acute stress does not uniformly change with
advancing age. However, autonomic reactivity shows sex-re-
lated alterations with attenuated responses in either older menor
older women according to the involved measurements.
Another factor that should be taken into account when in-
vestigating age differences insAA production isthe possibility of
age-related changes in salivary gland physiology. Studies have
shown that with increasing age morphological changes of oral
mucosa appear (Ghezzi & Ship, 2003; Scott, Flower, & Burns,
1987), but only marginal alterations of salivary gland function
and saliva composition are seen (Baum, 1989; Ghezzi & Ship,
2003; Wu, Atkinson, Fox, Baum, & Ship, 1993). In a recent
study, Nagler and Hershkovich (2005) reported a 62% lower
resting saliva flow rate and higher sAA concentrations in elderly
participants compared to young adults, but no difference of
overall sAA output. Taken together, while studies show that
there are age-related changes in saliva production and compo-
sition, there is no direct evidence for age-related changes of basal
sAA production. Furthermore, stress-induced increases of sAA
activity seem to be independent of salivary flow rate (Rohleder,
Wolf, Maldonado, & Kirschbaum, 2006).
acute psychosocial stress in three relevant age groups represent-
ing relevant stages of human developmentFhealthy children,
young adults, and older adults of both sexes. To ensure the
effectiveness of our stress protocol in triggering a physiological
stress response and in view of the importance of physiological
stress responses as determinants of health and disease, we aimed
to test for associations of sAA with more established stress sys-
tem markers, i.e., salivary cortisol as outcome measurement of
HPA reactivity, HR and HRVas markers for autonomic reac-
tivity. Based on the literature summarized above, we hypothe-
sized that sAA, HR, and HRVresponses would be attenuated in
older age, while cortisol responses would be increased. Salivary
a-amylase responses in children would be lower compared to
young adults, while we expect no differences in cortisol re-
sponses. Furthermore, we hypothesized to find no association
between sAA and cortisol, while sympathetic responses (sAA,
HR, and HRV) would be associated. It has been shown that
is known about the impact of chronic stress on sAA reactivity.
Thus, we set out to examine how subjective chronic stress levels
influence sAA as well as HR and HRVstress responses.
Children aged 6 to 10, young adults aged 20 to 31, and older
adultsaged 59 to61 wererecruitedviaanoticepostedoncampus
of the Technische Universita ¨ t Dresden, via an advertisement in a
local newspaper and via personal contact. The group of children
was recruited with the help of a local hospital, through which
parents who delivered a baby there between 1998 and 2002 were
contacted. Finally, 62 children (32 boys, 30 girls), 78 young
adults (45 men, 33 women), and 74 older adults (37 men, 37
women) fulfilled the inclusion criteria. We included participants
with a body mass index (BMI) below 30 kg/m2in children and
young adults, and below 35 kg/m2in older adults. Smokers and
individuals who reported excessive alcohol consumption (43
times a week) were excluded, also individuals under anti-hyper-
tensive medication, asthma medication, anti-rheumatic medica-
tion, using psychotropic substances, sleeping pills, or painkillers.
Participants were free of psychiatric and severe somatic diseases
as evaluated by interview by one of the authors (J.S.). Use of
other medication in the presence of changes of cardiovascular
functioning (anticoagulants), fat metabolism (antilipemics) and
thyroid functioning (thyroid therapeutics), and use of vitamins
and natural therapeutics were allowed, and participants were
instructed to take their prescribed drugs at least 5 hours before
intervention. Since there is no conclusive data available on the
impact of the use of oral contraceptives on sAA and menstrual
cycle phase (Rohleder & Nater, 2009), women were invited in
their luteal phase of menstrual cycle, while those taking oral
contraceptives were excluded. Participants were not allowed to
drink (anything but water) and eat one hour before intervention,
588J. Strahler et al.
(Rohleder & Nater, 2009).
All participants were invited to our laboratory on a weekday
between 14:00 and 18:00 h in the afternoon. Children were ac-
companied by at least one of their parents, who also signed the
informed consent and stayed in our laboratory until the end of
testing. After a 30-min resting period to minimize the impact of
physical activity, prior stress, and emotions, during which the
adult participants filled in some questionnaires (see Psychomet-
ricalanalyses), adultswereexposedtothe Trier Social Stress Test
(TSST; Kirschbaum, Pirke, & Hellhammer, 1993) while children
were exposed to the child version of the TSST (TSST-C), devel-
oped and evaluated by Buske-Kirschbaum and colleagues (Bus-
ke-Kirschbaum, Jobst, Wustmans, Kirschbaum, Rauh, &
Hellhammer, 1997). The TSSTconsists of a 3-min preparation
to discourse about their personal characteristics and children
front of an audience. After this stress task, participants stayed in
our laboratory for another 20 min to collect samples during re-
covery. The study protocol was approved by the ethics commit-
tee of the Technische Universita ¨ t Dresden.
Saliva sampling. To determine sAA and cortisol stress re-
sponses, four saliva samples were collected immediately before,
immediately after, 10 and 20 min after the stressor with the help of
cotton swabs (Salivettes, Sarstedt, Nu ¨ mbrecht, Germany). This
we would be able to capture peak and recovery of sAA, as well as
peak cortisol levels (Rohleder et al., 2006). Since the focus of the
present study was to evaluate sAA stress responses, we accepted
not being able to assess cortisol recovery for economic reasons. A
not secrete enough saliva without stimulation. We therefore de-
cided to instruct all participants to chew on the cotton rolls for 1
min to stimulate saliva flow. A previous study from our group has
shown that acute stress responses are independent of sampling
technique (Rohleder et al., 2006). Cotton swabs were then trans-
ferredto theplastic containersandstoredat ?201C untilanalysis.
Heart rate and heart rate variability. HR and HRV param-
eters were measured as additional markers of autonomic activity
in adults. We were not able to obtain HR data from children for
technical reasons. HR and HRV variables were derived from
cardiovascular measurements during the whole time of the study
using Polar S810i cardiac monitors (Polar Electro Ltd.,
Kempele, Finland). For analyses, artifact free time points of 2-
min duration were chosen corresponding to the time points of
sAA measurements. HR and HRV were calculated for one ad-
ditional 2-min intervalstarting 8 minintothe TSSTasa read-out
for HRV during stress. For quantification of HR and HRV, we
focusedontime domain variables becausethese are equivalentto
frequency-domain variables as well as easier to perform. In ac-
cordance with current recommendations (Task Force of the Eu-
ropean Society of Cardiology and the North American Society
of Pacing and Electrophysiology, 1996), only the root mean
square of successive differences (RMSSD) from time domain
analysis was obtained, reflecting the short-time HRVand being
predominantly a response to changes in parasympathetic tone.
RMSSD represents fast alterations of heart frequency in the re-
spiratory frequency range and might therefore be a measure of
fluctuating variations of vagal tone (Task Force, 1996). Despite
being highly correlated topower spectralmeasures ofrespiratory
sinus arrhythmia (RSA), it was suggested that RMSSD is not
significantly affected by changes in the breathing rate (Penttila ¨ ,
Helminen, Jartti, Kuusela, Huikuri, et al., 2001). All analyses
Analysis Group, University of Kuopio, Kuopio, Finland).
Biochemical analyses. Salivary a-amylase was measured by a
quantitative enzyme kinetic method. After thawing, saliva sam-
diluted 1:625 with double-distilled water. Twenty ml of diluted
saliva and standard were transferred into transparent 96-well
microplates. Standard was prepared from ‘‘Calibrator f.a.s.’’
solution (Roche Diagnostics, Mannheim, Germany) ranging
from5.01 to326 U/L amylase, anddouble-distilledwateraszero
standard. After that, 80 ml of substrate reagent (a-Amylase EPS
Sys; Roche Diagnostics) was added. The microplate was then
heated in an incubator at 431C for 90 s. Our laboratory has
established this procedure because internal tests revealed con-
stant temperatures of 371C in each well at this ambient air tem-
perature. After a first interference measurement at 405 nm using
a standard interference photometer (Sunrise-Basic Tecan, Tecan
Austria GmbH, Gro ¨ dig, Austria), the plate was incubated for
another 5 min at 431C, and the second measurement was done.
Increases of absorbance of samples were transformed to a-am-
ylaseactivity using a linearregression calculated for the standard
curve on each microplate (GraphPad Prism 4.0c for MacOSX,
GraphPad Software, San Diego, CA). Intra- and inter-assay
precision expressed as percent coefficient of variation was below
10%. Concentrations of salivary free cortisol were measured us-
ing a commercially available chemiluminescence-immuno-assay
(CLIA; IBL, Hamburg, Germany).
Psychometrical analyses. In order to control for possible in-
fluences of perceived stress on acute SNS reactivity, adult partic-
ipants filled in the German version of the Perceived Stress Scale
(PSS; Cohen, Kamarck, & Mermelstein, 1983). Additionally
chronic stress was measured using the 12-item Chronic Stress
Screening Scale (CSSS; Schulz, Schlotz, & Becker, 2004) in order
to examine possible influences on HPA stress responses as well as
autonomic reactivity, i.e., sAA, HR, HRV stress responses. As-
sessing the subjective stress experience while testing, participants
were asked to answer questions regarding anxiety and mood pre-
and post-stressor. Anxiety was measured with help of the state
version of the State-Trait Anxiety Inventory Scale (STAI-S; Laux,
visualanalogscales (VAS). Allofthesequestionnaireshaveshown
high internal consistency and validity. No such psychometric data
was available for children, but excessive stress experiences of the
children in the last 2 weeks prior testing were excluded by parent
Data were tested for normal distribution and homogeneity of
variance using a Kolmogorov–Smirnov and Levene’s test before
statistical procedures were applied. These analyses revealed sig-
nificant deviations from normality of some absolute sAA and
cortisol levels. Amylase and cortisol values were therefore log-
a-Amylase stress reactivity and aging589
transformed prior to analyses, which restored normality of dis-
tribution. ANOVAs for repeated measures were used to analyze
sAA and cortisol responses to the stressor with the between-
subjects factors group (children vs. younger adults vs. older
adults) and sex (male vs. female), and the within-subjects factor
sampling time (4 times). Similar ANOVAs were calculated for
the HR and HRVresponses, but only for young andolder adults
and a within-subjects factor with 5 levels. All results were cor-
rected by the Greenhouse-Geisser procedure where appropriate
(violation of sphericity assumption). Furthermore, because of
significant differences between the age groups with respect to
BMI, sAA and cortisol baseline values, perceived stress (PSS)
and chronic stress level (CSSS), those variables were included as
covariates where appropriate. Stress-induced increases of sAA
activity were indexed as delta score between sAA levels imme-
cortisol activity were indexed as delta score between cortisol
scores immediately pre-task and mean of 10 and 20 min post-
task, and stress-induced HR and HRValterations were indexed
as delta scores between immediately pre-task and 8 min after
stressoronset. Furthermore, area under the curve withrespectto
ground (AUCg) as well as increase (AUCi) were calculated for
each biomarker according to Pruessner, Kirschbaum, Meinlsch-
mid, & Hellhammer (2003). Univariate ANOVAs were com-
the LSD method were conducted to determine subgroup differ-
ences. To test for associations with chronic stress levels, Pearson
correlations were computed. Finally, hierarchical linear regres-
sion equations were used to predict the three different indices for
the sAA response (delta increase, AUCg, AUCi) by indices for
cortisol, HR, and RMSSD responses, controlling for age, BMI,
and sex, aswellaschronic stresslevels. Forsignificantresults, we
reportpartialetasquared (Z2) asameasureforeffectsize.Forall
analyses, the significance levelwas a55%. Allresults shownare
the mean ? standard error of mean (SEM).
Main characteristics of the study groups are shown in Table 1.
The results of the stress questionnaires (PSS, CSSS) indicated
that younger adults experienced a significant higher amount of
chronic stress than older adults (PSS: t15353.651, po.001;
CSSS:t13953.665, po.001), butneithergroupwas experiencing
significant subchronic or chronic stress during 3 months before
1988; Schulz et al., 2004).
Subjective Stress Responses
Analyses revealed no differences between the two adult groups as
well as men and women with regard to their evaluation of stress-
fulness of the TSST. The stress paradigm led to a significant
change in subjective anxiety (F[1,147]512.887, po.001, Z25
0.081), but without significant group or sex differences (all
po.05).However,wefoundasignificantsubjectiveanxiety ? sex
? group interaction (F[1,147]54.884, p5.029, Z250.032),
with older women showing the highest increase of subjective anx-
iety (data not shown).
Physiological Stress Responses
Salivary a-amylase. There was a highly significant baseline
difference between the age groups with respect to sAA activity
(F[2,210]523.9, po.001, Z250.185). Post hoc analyses re-
vealed that children had a significantly higher baseline than both
adult groups (all po.05). ANOVA for repeated measurements
revealed significant sAA changes in response to the TSST in all
Z250.019). Furthermore, we found a significant main effect of
age group (F[2,204]58.9, po.001, Z250.081), as well as a sig-
po.001, Z250.071) with children and older adults showing at-
of delta scores revealed a significant main effect of group
(F[2,206]535.9, po.001, Z250.259). Post hoc tests showed
that all groups differed significantly (all po.05), with young
adults showing the highest stress response and children showing
the lowest (Figure 1B). Same was true for AUCi(po.001). Fur-
ther analysesshoweda significant main effect ofgroup on AUCg
(po.001) with post hoc analyses revealing significantly higher
values in children compared to both adult groups (all po.001,
data not shown). No sex effects or age group ? sex interactions
were found. Pearson correlations between indices of stress levels
(CSSS, PSS) and summary indices (delta response, AUCi,
AUCg) revealed no association between the parameters.
age group interaction (F[5.7,580.3]57.7,
Salivary cortisol. There was a highly significant baseline
concentrations (F[2,209]527.2, po.001, Z250.206). Post hoc
analysis revealed that children had significantly lower baseline
levels than young adults. ANOVA for repeated measurements
590 J. Strahler et al.
Table 1: Sample Characteristics of Children, Young Adults, and Older Adults
Children (n562) Young adults (n578)Older adults (n574)
Mean (SD)Range Mean (SD)RangeMean (SD)Range
Amylase baseline (U/ml)n
Cortisol baseline (nmol/l)n
Notes: SD5standard deviation; BMI5Body Mass Index; PSS5Perceived Stress Scale, age specific norms: young adults 21.1 ? 7.2, older adults
18.3 ? 8.1; CSSS5Chronic Stress Screening Scale, age specific T-score norms (mean 50110z): young adults 7–24, older adults 4–19.
ngroup difference: t-test: po.001.
the TSST in all age groups (time effect: F[1.5,302.6]531.3,
po.001, Z250.133). Furthermore, we found a significant
main effect of age group (F[2,203]56.1, p5.003, Z250.056),
as wellasa significanttime ? age
(F[3.0,302.6]512.0, po.001, Z250.106), with young adults
and children reaching their maximum level 10 min post stressor
while older adults displayed a prolonged response (Figure 2A).
No sex effect, but a significant time ? group ? sex interaction
(F[1.5,302.6]52.9, p5.033, Z250.028) was found. Univariate
ANOVA of delta scores revealed a significant main effect of
group (F[2,204]56.2, p5.002, Z250.058). However, post hoc
tests showed no difference between the groups, although older
participants showed the highest mean stress response. While
there was no main effect of sex, a significant group ? sex inter-
action (F[2,204]53.6, p5.028, Z250.034) was found. As
age in male participants but remained constant through all age
groups in female participants. This main effect of group and the
group ? sex interaction were also found for AUCias well as
AUCg(all po.05, data not shown). Furthermore, post hoc an-
alyses showed a lower AUCias well as AUCgin children com-
pared to both adult groups (all p4.05). Pearson correlations
(delta response, AUCi, AUCg) revealed no association between
Heart rate. HR measurements were obtained only from the
two adult groups. There was a significant baseline difference be-
tween the age groups with respect to HR (F[1,129]54.2,
p5.043, Z250.032) with older adults showing lower baseline
levels (young adults 82.3 ? 1.5 bpm vs. older adults 78.2 ? 1.3
bpm). ANOVA for repeated measurements revealed highly sig-
nificant HR changes in response to the TSST in young and older
adults (time effect: F[2.8,341.3]54.8, p5.003, Z250.038).
Furthermore, we found a significant age group ? time effect
(F[2.8,341.3]55.9, p5.001, Z250.046), with older adults
showing an attenuated HR stress response (Figure 3A). No time
? sex or time ? group ? sex interaction was found. Univariate
ANOVA of delta scores revealed only this previously described
main effect of group (F[1,127]516.2, po.001, Z250.113) with
older adults showing a lower acute HR increase (Figure 3B).
Same was true for AUCg(po.001), while there was no effect of
group or gender on AUCi. Pearson correlations between indices
AUCi, AUCg) revealed no association between the parameters.
Heart rate variability. HRV measurements were obtained
only from the two adult groups. The baseline RMSSD differed
significantly between the groups (F[1,150]57.6, p5.007,
Z250.048), with older adults showing lower RMSSD scores at
baseline. ANOVA for repeated measurements revealed a signifi-
cant decrease of RMSSD in response to the TSST in young and
a-Amylase stress reactivity and aging 591
Figure 1. (A) Mean salivary a-amylase levels ( ? SE) after the Trier
Social Stress Test (TSST) in children, youngadults, andolder adults. (B)
Mean salivary a-amylase increase ( ? SE) after the TSST in children,
young adults, and older adults.
Figure 2. (A) Mean salivary cortisol levels ( ? SE) after the Trier Social
Stress Test (TSST) in children, young adults, and older adults. (B) Mean
salivary cortisol increase ( ? SE) after the TSST in children, young
adults, and older adults.
older adults (F[3.652, 460.150]523.756, po.001, Z250.159).
Furthermore, we found a significant group ? time effect
(F[3.652, 460.150]58.016, po.001, Z250.060), with older
(Figure 4A), in contrast to younger adults. Analyses revealed no
time ? sex or time ? group ? sex interaction. Univariate AN-
OVA ofdeltascores revealedonlythispreviouslydescribedmain
effect of group for the stress-induced RMSSD (F[1,125]55.6,
p5.019, Z250.043) delta scores, with older adults showing a
lower acute decrease in RMSSD (Figure 4B). Furthermore, a
on AUCgwas found, while there was no effect of group or
gender on AUCi. Older adults and women showed a lower total
HRV, while there were no differences according to the reactivity
of the HRV parameter RMSSD. Pearson correlations between
indices of stress levels (CSSS, PSS) and summary indices (delta
response, AUCi, AUCg) revealed no association between the
AUCi) were predicted by indices for cortisol, HR, and RMSSD
for age, BMI, and sex as well as subjective stress level indices
(CSSS, PSS). The regression model for the two adult groups
revealed age (delta: b5 ?0.269, p5.029; AUCg: b50.169,
p5.270; AUCi: b5 ?0.230, p5.055) as the strongest predic-
tor, whereas BMI, PSS, CSSS as well as cortisol, HR, and
RMSSD response indices failed to predict stress-induced sAA.
Including the group of children revealed similar results with one
exception. Now, BMI was an even stronger predictor of stress-
induced alteration of sAA (delta: b50.448, po.001; AUCg:
b5 ?0.203, p5.061; AUCi: b50.260, p5.020) than age
(delta: b5 ?0.218, p5.043; AUCg: b5 ?0.089, p5.407;
AUCi: b5 ?0.125, p5.262). Again, all otherparametersfailed
to predict sAA response indices.
This is the first study investigating the effects of a psychosocial
stressoronhumansalivarya-amylaseinawideragerangeof 6 to
61 years and to compare this to other markers of the psycho-
physiological response to acute stress. As expected, the TSST
induced significant increases in sAA, cortisol, HR, and signifi-
and HRVresponseswereattenuated inthe groupofolderadults.
In contrasttoour expectations, cortisol responseswereincreased
only in older men, but not in women. Also in accordance with
our hypotheses, we found lower sAA responses in children, and
no age differences in mean cortisol increases. However, lower
AUCivaluesindicatedalowersensitivity ofthe HPA inchildren.
Age ? sex interactions were found only for cortisol responses,
592J. Strahler et al.
Figure 3. (A) Mean heart rate levels ( ? SE) after the Trier Social Stress
Test (TSST) in young adults and older adults. (B) Mean heart rate
increase ( ? SE) after the TSST in young adults and older adults.
Figure 4. (A) Mean square root of the mean squared difference of
successive RR intervals (RMSSD) levels ( ? SE) after the Trier Social
Stress Test (TSST) in young adults and older adults. (B) Mean RMSSD
decrease ( ? SE) after the TSST in young adults and older adults.
whichincreasedwithageonly inmaleparticipants. Incontrastto
our hypotheses, no association between sAA and cardiovascular
responses were found. In linewith our hypothesis, same was true
for association between sAA and salivary free cortisol. Analyses
revealed age and BMI as the strongest predictors of sAA in-
creases, whereas subjective stress levels as well as cortisol, HR,
and RMSSD response indices failed to predict sAA stress re-
sponses. A further unexpected finding was that children had sig-
nificantly higher baseline sAA concentrations than both groups
of adult participants. Also in contrast to our hypothesis, subjec-
tive chronic stress levels had no influence on sAA as well as
cortisol, HR, and HRVstress responses.
Thus, our results replicate previous studies regarding HRV
former studies (Kudielka et al., 2004a; Nater et al., 2006;
Rohleder et al., 2004), stress-induced responses in most param-
eters are found in both adult groups. Furthermore, attenuated
reactivity of sAA, HR, and HRV in older adults corresponds to
the literature regarding stress-related SNS activity with aging
finding argues for the validity of sAA since this attenuation of
stress-reactivity with aging was also found for the other sympa-
thetic markers. However, there was no association between sAA
Furthermore, we found significant sAA increases in older
adults as well as in children. This finding contradicts previous
results of Stroud et al. (2009), showing no sAA increase of chil-
However, to a peer rejection task, significant sAA emerged. This
children. To date, nothing is known about the impact of age-
related changes of acute sAA response profiles in children of
different ages. In our study, children aged 6 to 10 were included.
Preliminary data from our group indicates that children under
a standardized psychosocial stressor, whereas in children above
the age of 8 it reaches adult levels (Rosenloecher & Strahler, in
preparation). However, in this study there was only one child
younger than 7 years and excluding this subject from our ana-
lyses did not change our results. Furthermore, it could be spec-
ulated that lower reactivity in children is the result of a ceiling
effect because of higher baseline values. According to studies
looking at older age groups’ basal sAA activity (Pajukoski, Me-
urman, Snellman-Gro ¨ hn, & Sulkava, 1999), there was no base-
line difference in younger and older adults. Since nothing is
known about possible mechanisms accounting for this unex-
pected baseline difference in children, it could be speculated that
pre-task stress experiences or anticipatory stress are causing the
higher baseline sAA activity found here. In agreement with pre-
vious findings (Buske-Kirschbaum et al., 1997; Kudielka et al.,
2004a), no mean cortisol increase differences between children
and the adult groups could be found. However, children showed
lower total cortisol output (defined as AUCg) as well as lower
AUCiindicating lower reactivity of this system.
In line with previous reports (Kudielka et al., 2004a), age ?
showing the highest mean salivary free cortisol increase. How-
ever, regarding sAA, HR, and HRV responses, no age ? sex
interactions were found. This is in contrast to findings of Ku-
dielka et al. (2004b), showing differences between young men
and women, while the HR responses in older men and women
were comparable, as also indicated by our results. According to
the literature, it is difficult to draw conclusions concerning sex-
specific autonomic responses in people of different ages because
of different kinds of stress tasks and autonomic measurements
(Seals & Dinenno, 2004).
The lack of data-based support for an association between
sAA and cortisol responses found in this present study is inter-
esting in light of the hypothesized relationship of SNS and HPA
stress responsiveness discussed by Granger, Kivlighan, El-
Sheikh, Gordis, and Stroud (2007). Previously, correlations be-
tween sAA and cortisol were rarely reported (Nater et al., 2005,
2006). Granger et al. (2007) attributed this lack of correlation to
stress of the two stress-response systems. Furthermore, we did
not find any associations between sAA and cardiovascular mea-
sures. This lack of data-based support for the hypothesized as-
sociation between sAA and cardiovascular autonomic measures
may hint to different sympathetic effects on different organ sys-
tems, i.e., different mechanisms governing the stress-induced
stimulation of salivary glands with the release of sAA on the one
hand, and stress-inducedcardiovascular alterations on the other.
While salivary protein secretion is largely under sympathetic
control and evoked by b-adrenergic stimulation in particular
(Busch & Borda, 2002; Garrett, 1987; Proctor & Carpenter,
2007), cardiovascular stress-reactivity seems tobemodulated via
either hormonal control mediated by the adrenal medulla or by
increasing sympathetic or decreasing parasympathetic stimula-
tion on a-adrenergic receptors (Hjemdahl, Fagius, Freyschuss,
Wallin, Daleskog, et al., 1989; Thayer & Siegle, 2002). Thus,
sAA might be an indirect indicator of the central sympathetic
system (Ehlert et al., 2006) which is not necessarily associated
with peripheral sympathetic effects.
Although it is known that chronic stress alters HPA stress
responses (for a review, see Chida & Hamer, 2008), we found no
correlation between subjective stress levels and cortisol stress re-
sponses. Same was true for sAA as well as HR and HRV. These
results are in line with the phenomenon that subjective measures
of stress often failed to be correlated with physiological stress
A key finding is that we were able to induce increases in sAA
in response to stress across a wide age range. Despite the lack of
associations between sAA and cardiovascular responses, atten-
uated responses in the older adults’ group were found for HR
and HRVas well as sAA. This pattern closely fits the literature
concerning sympathetic reactivity with aging and strengthens
the role of sAA as a marker of psychophysiological stress re-
sponses. It needs to be excluded, however, that age-related
differences are caused by changes of salivary flow and compo-
sition (Dodds, Johnson, & Yeh, 2005). In fact, there was no
baseline difference between both adult groups, but significantly
higher baseline values of sAA were found in the group of chil-
in children aged 6 to 10 years in response to a standardized
Some limitations should be considered. In our study, saliva
was collected using salivettes, andparticipantswere instructed to
chew on the cotton roll. We thus collected stimulated whole sa-
liva, and it might be argued that active chewing could alter sAA
activity (DeCaro, 2008). However, stress-induced increases of
sAA activity have been shown tobe independent of salivary flow
rate (Rohleder et al., 2006). Furthermore, it has been reported
that stimulation or chewing changes the pH value of saliva
(Polland, Higgins, & Orchardson, 2003). It is well known that
a-Amylase stress reactivity and aging 593
enzymatic activity is affected by changes in pH. So, it could be
suggested that chewing on swabs with subsequent stimulation of
salivary flowleads toalterationsofpH valueandthustochanges
of enzymatic activity. However, unpublished data from our lab-
oratory shows that there is no alteration of sAA activity level
after increasing or decreasing sample pH value with hydrochlo-
ride or sodium hydroxide. Another limitation might be that we
cannot exclude the possibility that the slightly different stress
protocols for children and adults caused stressor-specific re-
the TSST in producing stress response in children of different
ages needs to be further investigated. Furthermore, there are
constraints regarding cardiovascular parameters and the validity
of data collected with commercially available inter-beat (RR)
interval recorders (Polar S810i). However, adequate agreement
(ECG) was shown (Nunan, Jakovljevic, Donovan, Hodges,
Sandercock, & Brodie, 2008).
The present study illustrates age- and sex-related changes of en-
docrine and autonomic responses to a psychosocial stressor. The
findings regarding cortisol and cardiovascular parameters are
consistent with most previous studies. Results concerning sali-
vary a-amylase draw attention to its usefulness as a sympathetic
activity marker, since we found rapid stress-induced increases of
this salivary enzyme above all age groups. These age- and sex-
related changes of stress reactivity might be important determi-
to vulnerability to disease and disease prevalence-patterns needs
to be investigated in long-term studies.
the American Geriatrics Society, 37, 453–58.
Bigger, J. T., Fleiss, J. L., Steinman, R. C., Rolnitzky, L. M., Schneider,
W. J., & Stein, P. K. (1995). RR variability in healthy, middle-aged
persons compared with patients with chronic coronary heart disease
or recent acute myocardial infarction. Circulation, 91, 1936–1943.
Blackwell, D., Hayward, M., & Crimmins, E. (2001). Does childhood
health affect chronic morbidity in later life? Social Science and Med-
icine, 52, 1269–1284.
Busch, L., & Borda, E. (2002). Influence of castration on isoprenaline-
Physiology, 87, 447–452.
W., & Hellhammer, D. (1997). Attenuated free cortisol response to
psychosocial stress in children with atopic dermatitis. Psychosomatic
Medicine, 59, 419–426.
Chatterton, R. T. Jr., Vogelsong, K. M., Lu, Y. C., Ellman, A. B., &
Hudgens, G. A. (1996). Salivary Alpha-amylase as a measure of en-
dogenous adrenergic activity. Clinical Physiology, 16, 433–448.
Chida, Y., & Hamer, M. (2008). Chronic psychosocial factors and acute
physiological responses to laboratory-induced stress in healthy pop-
ulations: A quantitative review of 30 years of investigations. Psycho-
logical Bulletin, 134, 829–885.
Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure of
perceived stress. Journal of Health and Social Behavior, 24, 385–396.
Cohen, S., & Williamson, G. (1988). Perceived stress in a probability
sampleofthe United States. In S. Spacapan & S. Oskamp (Eds.), The
social psychology of health: Claremont Symposium on Applied Social
Psychology. Newbury Park, CA: Sage.
DeCaro, J. (2008). Methodological considerations in the use of salivary
a-amylase as a stress marker in field research. American Journal of
Human Biology, 20, 617–619.
Dodds, M. W. J., Johnson, D. A., & Yeh, C.-K. (2005). Health benefits
of saliva: A review. Journal of Dentistry, 33, 223–233.
Ehlert, U., Erni, K., Hebisch, G., & Nater, U. (2006). Salivary alpha-
amylase levels after Yohimbine challenge in healthy men. The Journal
of Clinical Endocrinology & Metabolism, 91, 5130–5133.
Garrett, J. R. (1987). The proper role of nerves in salivary secretion: A
review. Journal of Dental Research, 66, 387–397.
Ghezzi, E. M., & Ship, J. A. (2003). Agingandsecretory reserve capacity
of major salivary glands. Journal of Dental Research, 82, 844–848.
Gordis, E. B., Granger, D. A., Susman, E. J., & Trickett, P. K. (2006).
Asymmetry between salivary cortisol and alpha-amylase reactivity to
stress: Relation to aggressive behavior in adolescents. Psychoneuro-
endocrinology, 31, 976–987.
Gordis, E. B., Granger, D. A., Susman, E. J., & Trickett, P. K. (2008).
Salivary alpha amylase-cortisol asymmetry in maltreated youth.
Hormones and Behavior, 53, 96–103.
Granger, D. A., Kivlighan, K. T., El-Sheikh, M., Gordis, E. B., &
Stroud, L. (2007). Salivary alpha-amylase in biobehavioral research:
Recent developments and applications. Annals of the New York
Academy of Sciences, 1098, 122–144.
Hjemdahl, P., Fagius, J., Freyschuss, U., Wallin, B. G., Daleskog, M.,
Bohlin, G., et al. (1989). Muscle sympathetic activity and norepi-
nephrine release during mental challenge in humans. American Jour-
nal of Physiology, 257, E654–E664.
Honojosa-Laborde, C., Chapa, I., Lange, D., & Haywood, J. R. (1999).
and Experimental Pharmacology and Physiology, 26, 122–126.
Kirschbaum, C., Pirke, K. M., & Hellhammer, D. H. (1993). The Trier
Social Stress TestFA tool for investigating psychobiological stress
responses in a laboratory setting. Neuropsychobiology, 28, 76–81.
Kudielka, B. M., Buske-Kirschbaum, A., Hellhammer, D. H., & Kirsch-
baum, C. (2004a). HPA axis responses to laboratory psychosocial
stress in healthy elderly adults, younger adults, and children: Impact
of age and gender. Psychoneuroendocrinology, 29, 83–98.
Kudielka, B. M., Buske-Kirschbaum, A., Hellhammer, D. H., & Kirsch-
baum, C. (2004b). Differentialheartrate reactivity andrecovery after
psychosocial stress (TSST) in healthy children, younger adults, and
elderly adults: The impact of age and gender. International Journal of
Behavioral Medicine, 11, 116–121.
Kunz-Ebrecht,S., Mohamed-Ali,V., Feldman,P. J., Kirschbaum,C., &
Steptoe, A. (2003). Cortisol responses to mild psychosocial stress are
inversely associated with proinflammatory cytokines. Brain, Behav-
ior, and Immunity, 17, 373–383.
Laux, L., Glanzmann, P., Schaffner, P., & Spielberger, C. D. (1981).
State-Trait-Angst-Inventar STAI. Weinheim: Beltz.
Lavi, S., Nevo, O., Thaler, I., Rosenfeld, R., Dayan, L., Hirshoren, N.,
et al. (2006). Effect of aging on the cardiovascular regulatory systems
inhealthy women. The American JournalofPhysiologyFRegulatory,
Integrative and Comparative Physiology, 292, R788–R793.
Matthews, K. A., Gump, B. B., & Owens, J. F. (2001). Chronic stress
influences cardiovascular and neuroendocrine responses during acute
stress and recovery, especially in men. Health Psychology, 20, 403–
McEwen, B. S., & Stellar, E. (1993). Stress and the individual. Mech-
anisms leading to disease. Archives of Internal Medicine, 153, 2093–
Mezzacappa, E. S., Kelsey, R. M., Katkin, E. S., & Sloan, R. P. (2001).
Medicine, 63, 650–657.
Nagler, R. M., & Hershkovich, O. (2005). Relationships between age,
drugs, oral sensorial complaints and salivary profile. Archives of Oral
Biology, 50, 7–16.
Nater, U. M., Rohleder, N., Gaab, J., Berger, S., Jud, A., Kirschbaum,
C., & Ehlert, U. (2005).Humansalivary alpha-amylasereactivityin a
psychosocial stress paradigm. International Journal of Psychophysi-
ology, 55, 333–342.
Nater, U. M., La Marca, R., Florin, L., Moses, A., Langhans, W.,
Koller, M. M., & Ehlert, U. (2006). Stress-induced changesinhuman
594 J. Strahler et al.
salivary alpha-amylase activity-associations with adrenergic activity. Download full-text
Psychoneuroendocrinology, 31, 49–58.
Nater, U. M., & Rohleder, N. (2009). Salivary alpha-amylase as a non-
invasive biomarker for the sympathetic nervous system: Current state
of research. Psychoneuroendocrinology, 34, 486–496.
Nunan, D., Jakovljevic, D. G., Donovan, G., Hodges, L. D., Sander-
cock, G. R., & Brodie, D. A. (2008). Levels of agreement for RR
S810 and an alternative system. European Journal of Applied Phys-
iology, 103, 529–537.
Pajukoski, H., Meurman, J. H., Snellman-Gro ¨ hn, S., & Sulkava, R.
(1999). Oral health in hospitalized and nonhospitalized community-
dwelling elderly patients. Oral Surgery, Oral Medicine, Oral Pathol-
ogy, Oral Radiology, and Endodontics, 88, 437–443.
Penttila ¨ , J., Helminen, A., Jartti, T., Kuusela, T., Huikuri, H. K.,
Tulppo, M. P., etal.(2001). Time domain, geometricalandfrequency
domain analysis of cardiac vagal outflow: Effects of various respira-
tory patterns. Clinical Physiology, 21, 365–376.
Polland, K. E., Higgins, F., & Orchardson, R. (2003). Salivary flow rate
and pH during prolonged gum chewing in human. Journal of Oral
Rehabilitation, 30, 861–865.
Proctor, G. B., & Carpenter, G. H. (2007). Regulation of salivary gland
function by autonomic nerves. Autonomic Neuroscience: Basic and
Clinical, 133, 3–18.
Pruessner, J. C., Kirschbaum, C., Meinlschmid, G., & Hellhammer, D.
H. (2003). Two formulas for the computation of the area under the
dependent change. Psychoneuroendocrinology, 28, 916–931.
Rohleder, N., Nater, U. M., Wolf, J. M., Ehlert, U., & Kirschbaum, C.
(2004). Psychosocial stress-induced activation of salivary alpha-am-
ylaseFAn indicator of sympathetic activity? Annals of the New York
Academy of Sciences, 1032, 258–263.
Rohleder, N., Wolf, J. M., Maldonado, E. F., & Kirschbaum, C. (2006).
The psychosocial stress-induced increase in salivary alpha-amylase is
independent of saliva flow rate. Psychophysiology, 43, 645–652.
Rohleder, N., & Nater, U. M. (2009). Salivary alpha-amylase in hu-
mansFDeterminants and methodological considerations. Psycho-
neuroendocrinology, 34, 469–485.
Schulz, P., Schlotz, W., & Becker, P. (2004). Das Trierer Inventar zum
chronischen StressFManual. Goettingen, Germany: Hogrefe.
Scott, J., Flower, E. A., & Burns, J. (1987). A quantitative study of
histological changes in the human parotid gland occurring with adult
age. Journal of Oral Pathology and Medicine, 16, 505–510.
Seals, D. R., & Dinenno, F. A. (2004). Collateral damage: Cardiovas-
cular consequences of chronic sympathetic activation with human
aging. American Journal of PhysiologyFHeart and Circulatory Phys-
iology, 287, H1895–H1905.
Stroud, L. R., Foster, E., Papandonatos, G. D., Handwerger, K.,
Granger, D. A., Kivlighan, K. T., & Niaura, R. (2009). Stress re-
sponses and the adolescent transition: Performance versus peer re-
jection stressors. Development & Psychopathology, 21, 47–68.
Task Force of the European Society of Cardiology and the North Amer-
ican Society of Pacing and Electrophysiology (1996). Heart rate vari-
ability: Standards of measurement, physiological interpretation and
clinical use. Circulation, 93, 1043–1065.
Thayer, J. F., & Siegle, G. J. (2002). Neurovisceral integration in cardiac
and emotionalregulation. IEEEEngineeringin Medicineand Biology,
Van Stegeren, A., Rohleder, N., Everaerd, W., & Wolf, O. T. (2006).
effect of betablockade. Psychoneuroendocrinology, 31, 137–141.
Van Stegeren, A., Wolf, O. T., & Kindt, M. (2008). Salivary alpha am-
ylase and cortisol responses to different stress tasks: Impact of sex.
International Journal of Psychophysiology, 69, 33–40.
White, M., Courtemanche, M., Stewart, D. J., Talajic, M., Mikes, E.,
Cernacek, P., et al. (1997). Age- and gender-related changes in en-
dothelin and catecholamine release, and in autonomic balance in re-
sponse to head-up tilt. Clinical Science, 93, 309–316.
Wright, C. E., O’Donnell, K., Brydon, L., Wardle, J., & Steptoe, A.
(2007). Family history of cardiovascular disease is associated with
cardiovascular responses to stress in healthy young men and women.
International Journal of Psychophysiology, 63, 275–828.
Wu, A. J., Atkinson, J. C., Fox, P. C., Baum, B. J., & Ship, J. A. (1993).
Cross-sectional and longitudinal analyses of stimulated parotid sal-
ivary constituents in healthy, different-aged subjects. Journal of Ger-
ontology, 48, M219–M224.
(Received January 27, 2009; Accepted July 1, 2009)
a-Amylase stress reactivity and aging595