Relationship between cortisol and age-related memory impairments in Holocaust survivors with PTSD
Holocaust survivors with PTSD appear to show an accelerated aging effect as evidenced by their performance on tests of explicit memory, and also show more exaggerated patterns on age-related alterations in cortisol release over the diurnal cycle than Holocaust survivors without PTSD and nonexposed subjects. To investigate the implications of age-related HPA axis alterations on cognition, we examined correlations between parameters reflecting circadian cortisol release and implicit and explicit memory performance. Nineteen Holocaust survivors with PTSD (7 men, 12 women), 16 Holocaust survivors without PTSD (7 men, 9 women), and 28 non-exposed healthy comparison subjects (13 men, 15 women) collected salivary samples at six times over the diurnal cycle, and were tested with Paired Associates and Word Stem Completion Tests. Negative correlations were observed between several measures of salivary cortisol concentrations and explicit memory in Holocaust survivors with PTSD after adjusting for IQ, years of education and current age reflecting poorer performance in association with higher cortisol levels. This relationship was absent in Holocaust survivors without PTSD and in demographically-comparable subjects who were not exposed to the Holocaust or other extremely traumatic events. The significantly different relationship between cortisol and memory performance in these groups suggests that the neuropsychological impairments observed in Holocaust survivors with PTSD may reflect an interaction of PTSD and aging effects.
Relationship between cortisol and age-related
memory impairments in Holocaust survivors
, Julia A. Golier, Philip D. Harvey, Karina Stavitsky,
Shira Kaufman, Robert A. Grossman, Lisa Tischler
Division of Traumatic Stress Studies, Department of Psychiatry OOMH, Mount Sinai School of Medicine
and Bronx Veteran Affairs Medical Center, 130 West Kingsbridge Road, Bronx, NY 10468, USA
Received 9 September 2004; received in revised form 22 February 2005; accepted 22 February 2005
Summary Rationale: Holocaust survivors with PTSD appear to show an accelerated
aging effect as evidenced by their performance on tests of explicit memory, and also
show more exaggerated patterns on age-related alterations in cortisol release over
the diurnal cycle than Holocaust survivors without PTSD and nonexposed subjects. To
investigate the implications of age-related HPA axis alterations on cognition, we
examined correlations between parameters reﬂecting circadian cortisol release and
implicit and explicit memory performance.
Methods: Nineteen Holocaust survivors with PTSD (7 men, 12 women), 16 Holocaust
survivors without PTSD (7 men, 9 women), and 28 non-exposed healthy comparison
subjects (13 men, 15 women) collected salivary samples at six times over the diurnal
cycle, and were tested with Paired Associates and Word Stem Completion Tests.
Results: Negative correlations were observed between several measures of salivary
cortisol concentrations and explicit memory in Holocaust survivors with PTSD after
adjusting for IQ, years of education and current age reﬂecting poorer performance in
association with higher cortisol levels. This relationship was absent in Holocaust
survivors without PTSD and in demographically-comparable subjects who were not
exposed to the Holocaust or other extremely traumatic events.
Conclusion: The signiﬁcantly different relationship between cortisol and memory
performance in these groups suggests that the neuropsychological impairments
observed in Holocaust survivors with PTSD may reﬂect an interaction of PTSD and
Q2005 Elsevier Ltd. All rights reserved.
Aged Holocaust survivors with PTSD show a ﬂatter
circadian rhythm of cortisol release than compari-
son subjects who exhibit lower morning, and higher
Psychoneuroendocrinology (2005) 30, 678–687
0306-4530/$ - see front matter Q2005 Elsevier Ltd. All rights reserved.
Corresponding author. Tel.: C1 718 584 9000x6964; fax: C1
718 741 4775.
E-mail address: email@example.com (R. Yehuda).
evening, cortisol levels (Yehuda et al., in press). In
contrast, in some reports, younger trauma survivors
with PTSD have shown a greater diurnal range in
cortisol release, marked by comparable morning
cortisol levels, but signiﬁcantly lower evening levels
compared to subjects without PTSD (Yehuda et al.,
1996; Thaller et al., 1999). In normal aging, cortisol
levels during the circadian trough have been shown
to also be elevated, similarly resulting in a ﬂatter
circadian rhythm of cortisol (Milcu et al., 1978; Dodt
et al., 1994; Van Cauter et al., 1996; Kern et al.,
1996; Deuschle et al., 1997). However, whereas
normal aging is associated with only mildly elevated
ambient cortisol levels (Friedman et al., 1969),
possibly because morning cortisol levels are not
greatly diminished, aging Holocaust survivors with
PTSD have lower mean cortisol levels over the
diurnal cycle (Yehuda et al., 1995b). Furthermore,
whereas normal aging is associated, if anything, with
reduced negative feedback inhibition (Deuschle
et al., 1997; DeLeon et al., 1997; Ferrari et al.,
1995; Parnetti et al., 1990; Wilkinson et al., 1997,
2001; Zovato et al., 1996), Holocaust survivors with
PTSD demonstrate an enhanced negative feedback
inhibition of cortisol compared to similarly-aged
subjects (Yehuda et al., 2002), analogous to what
has been observed in younger groups of PTSD
subjects (for review, see Yehuda, 2002). Accord-
ingly, we have suggested that the ﬂatter circadian
rhythm in Holocaust survivors may be indicative of a
superimposition of age-related HPA axis alterations
(i.e. ﬂattening of circadian rhythm) on PTSD-related
cortisol changes (Yehuda et al., in press).
In the current study, we considered the impact of
age-related HPA axis alterations on cognitive
performance in Holocaust survivors by examining
the correlations between parameters reﬂecting
circadian cortisol release in salivary samples and
measures reﬂecting explicit and implicit memory
tests, wondering speciﬁcally if reduced cortisol
levels, estimated on the basis of salivary samples
obtained throughout the day, would be associated
with poor memory performance in Holocaust
survivors with PTSD. We have previously demon-
strated that Holocaust survivors with PTSD perform
more poorly than both Holocaust survivors without
PTSD and age-comparable Jewish comparison sub-
jects on several tests of learning and memory
performance (Golier et al., 2002; Yehuda et al.,
2004). With respect to performance on explicit
memory, Holocaust survivors with PTSD demon-
strated decrements in their ability to learn both
high and low associated word pairs as compared to
both Holocaust survivors without PTSD and non-
exposed comparison subjects. However, a signiﬁ-
cantly greater negative correlation between age
and performance in Holocaust survivors with PTSD
suggested a pattern consistent with, among other
possibilities, an accelerated age-related decline in
cognitive performance, and raised the possibility
that poorer cognitive performance in Holocaust
survivors is not related to PTSD per se, but rather, is
a consequence of aging and PTSD in this group
(Golier et al., 2002). Indeed, younger subjects with
PTSD have tended to demonstrate not only fewer
cognitive impairments, but slightly different ones
than older PTSD subjects (Gurvits et al., 1993;
Yehuda et al., 1995a).
Several studies have noted that in healthy,
nonpsychiatric aged populations, increased cortisol
levels over time are associated with a decline in
performance of tests of explicit memory and
related hippocampal atrophy in aging (Lupien
et al., 1994, 1996, 1998; O’Brien et al., 1994;
Seeman et al., 1997; Kalmijn et al., 1998; Green-
dale et al., 2000). Similarly, in normal elderly
subjects subdivided by median split into those with
high and low plasma cortisol levels, signiﬁcantly
better performance on a variety of explicit memory
tests was found in the subgroup with lower cortisol
(Carlson and Sherwin, 1999).
On the other hand, lower cortisol levels were
also associated with poorer performance on the
paired associates test in elderly subjects with mild
cognitive impairment compared to normal elderly
and young controls (Wolf et al., 2002). These
ﬁndings were surprising since persons with mild
cognitive impairment are thought to be at risk for
Alzheimer’s disease, a condition that has been
associated with cortisol hypersecretion and hippo-
campal changes (Davis et al., 1986; Peskind et al.,
2001; DeLeon et al., 1988) and raise the possibility
that the effects of cortisol on memory performance
may be bi-directional (i.e. that cortisol levels need
to fall into a certain range for optimal performance
with lower or higher levels yielding impairments)
and dependent on a host of individual differences,
or at least on salient differences between cohorts of
elderly persons with speciﬁc comorbid conditions.
In the present study, cortisol levels and perform-
ance on a paired associates memory test and word
stem completion test were assessed in Holocaust
survivors with PTSD, Holocaust survivors without
PTSD and age-comparable Jewish subjects who
were not exposed to the Holocaust. Insofar as
increased cortisol levels have generally been
associated with cognitive impairment, we hypoth-
esized that there would be an overall negative
association between estimates of cortisol levels
from salivary samples and performance on paired
associate learning. In view of anticipated differ-
ences in ambient cortisol levels as well as cognitive
Relationship between cortisol and age-related memory impairments in Holocaust survivors with PTSD 679
performance in Holocaust survivors with and with-
out PTSD and persons unexposed to trauma, we also
examined these relationships separately in the
three subgroups to determine whether the relation-
ships between cortisol and memory performance
would differ among them. We hypothesized that to
the extent that the neuropsychological impair-
ments observed in aging Holocaust survivors with
PTSD reﬂected a superimposition of PTSD and aging
effects, that signiﬁcantly different correlations
between cortisol and memory performance would
be present in this group compared to the two other
The sample consisted of 35 Holocaust survivors
(14 males, 21 females) and 28 Jewish comparison
subjects (13 males, 15 females). The study was
approved by the Institutional Review Board of the
Mount Sinai School of Medicine and all subjects
provided written, informed consent. Participants
were excluded from the study if they had a history of
psychosis, bipolar disorder, alcohol or substance
dependence, organic mental disorder, or dementia.
Additional exclusion criteria for the comparison
participants included a lifetime history of major
mood or anxiety disorder. All subjects with an active
medical or neurological illness or signiﬁcant head
trauma were excluded. However, subjects with
other illnesses common in this age group
(e.g. hypertension, hyperlipidemia, arthritis) were
studied if they were stabilized. Medications used
included anti-hypertensives, lipid-lowering medi-
cation, thyroid supplements, estrogen-based hor-
mone treatments, gastrointestinal medications,
and, in both groups of Holocaust survivors, stable
doses of psychotropic medications (see Golier et al.,
2003; Yehuda et al., 2005). This study reports on the
relationship between cortisol and memory; the
results of these subjects’ neuropsychological test
performance were previously reported (Golier
et al., 2002)(nZ35). Basal salivary cortisol levels
from a different subset were also previously
reported in Yehuda et al. (in press) (nZ37).
2.2. Clinical assessments
Diagnoses were made according to DSM-IV criteria
using the Structured Clinical Interview for DSM-IV
(Spitzer et al., 1995) and the Clinician Administered
PTSD Scale (CAPS) (Blake et al., 1995). Diagnoses
were reviewed in a consensus diagnosis meeting
with at least three research clinicians present
(including RY, JG, RG).
2.3. Neuropsychological testing
The WAIS-R vocabulary test and block design
(Wechsler, 1981) were administered to assess verbal
and non-verbal intelligence respectively. Adminis-
tration of paired associates test of explicit and
implicit memory in this population were previously
described (Golier et al., 2002). Brieﬂy, this test
consists of 12 pairs of words, six pairs of moderately
related words (high associates) and six pairs of
unrelated words (low associates), which the subjects
are shown and asked to read aloud and memorize, for
recall immediately thereafter. The subjects were
then shown one word of each pair and asked to recall
the other. The dependent measures were the mean
number of related and unrelated words correctly
recalled over four learning and recall trials, the
maximum score for which is six per trial for both high
and low associates. For the implicit memory portion
of the test, the subjects were asked to complete 48
three-letter word stems using ‘the ﬁrst word that
comes to mind’. Implicit memory is inferred if the
target stems are completed to form stimulus words
in the absence of explicit instructions to do so. Half
of the stems were the target stems that corre-
sponded to words presented in the paired associate
task; the remaining stems were randomly chosen.
The dependent variables are the percentages of the
related and unrelated target stems completed with
2.4. Cortisol collection
Participants collected saliva into Salivette tubes,
(Starstedt, Nuembrecht, Germany), over the course
of a day at awakening, 08:00, 12:00, 16:00, 20:00
and bedtime. Cortisol was measured by radio-
immunoassay, as previously described (Goenjian
et al., 1996). The detection limit was 10 ng/dl, and
intra- and interassay variability were 3.9 and 12.0%,
2.5. Statistical methods
Analyses of variance (ANOVA) was used to examine
group differences in demographic, clinical, neurop-
sychological and cortisol variables. Potentially
confounding factors were identiﬁed by screening
the data for variables known to inﬂuence neurop-
sychological performance and/or cortisol and
determining whether they are signiﬁcantly related
to paired associate performance and to cortisol
R. Yehuda et al.680
levels (i.e. age, gender, years of education, age-
scaled WAIS Vocabulary scores, and body mass
index (BMI)). Only variables that were associated
with the dependent variables without differing
between the groups in terms of these associations
could be considered as covariates, since interpret-
ation of ANCOVA is problematic when the indepen-
dent variable is not randomly distributed within the
experimental conditions (Miller and Chapman,
2001). Since this criterion was not met by any of
the variables (i.e. all variables associated with
dependent measures reﬂected pre-existing group
differences), none were used as covariates in
between group analyses. However, variables associ-
ated with the dependent measures were used in
partial correlation analyses examining the relation-
ship between memory performance and cortisol
levels within each group separately. Additionally, a
series of linear regression analyses were carried out
in order to determine the amount of variance in
memory performance contributed by the demo-
graphic factors and variance contributed by cortisol
alone. Cortisol values were skewed and kurtotic,
thus analyses were performed using log-trans-
formed cortisol levels. However, raw data are
presented in graphs and summary variables. An
initial repeated measures ANOVA evaluated
Group!Time differences in the six salivary assess-
ments, but summary variables reﬂecting the mean,
maximum, and minimum cortisol levels, as well as
range (maximum–minimum) were used in
correlational analyses. The neuropsychological
dependent variables of interest included the mean
number of related and unrelated words (explicit
memory) and the percentage of the related and
unrelated target words completed during the word
stem completion phase (implicit memory).
Demographic, neuropsychological and cortisol data
for the three groups are presented in Tables 1 and 2
and Fig. 1. The sample consisted of 36 women
(57.1%) and 27 men (42.9%). There were no group
differences in gender distribution (c
pZ0.81). There were no signiﬁcant group differ-
ences in age (FZ2.0, dfZ2.60, pZ0.15) or BMI
(FZ0.14, dfZ2.60, pZ0.87). There were signiﬁ-
cant group differences in education (FZ9.8,
dfZ2.60, p!0.0005), with PTSDChaving fewer
years of education than both the PTSDKand the
non-exposed groups, and lower scores on the WAIS-
R age scaled Vocabulary subtest (FZ13.6, dfZ2.60,
p!0.0005), that reﬂected signiﬁcantly lower
scores in the PTSDCgroup.
Consistent with what was already reported in a
different, but overlapping sample, Holocaust
survivors with PTSD showed a signiﬁcantly reduced
explicit memory as evidenced by a compromised
ability to learn high (FZ10.9, dfZ2.60, p!
0.0005) and low (FZ9.2, dfZ2.60, p!0.0005)
associated word pairs compared to the two other
groups. This was conﬁrmed with post hoc testing.
No signiﬁcant group differences were observed in
implicit memory for either the high (FZ0.70, dfZ
2.58, pZ0.50) or low (FZ0.49, dfZ2.58, pZ0.62)
associate word pairs.
Table 1 Demographic and clinical characteristics.
PTSDC(NZ19) PTSDK(NZ16) Non-exposed (NZ28)
Age 69.7 (5.0) 70.2 (6.9) 73.0 (6.3)
Education (years) 11.6 (3.9) 14.8 (4.4) 16.2 (2.6)
7/12 7/9 13/15
Body mass index 25.4 (3.4) 25.7 (3.0) 25.2 (2.9)
Vocabulary score 10.5 (2.5) 13.7 (2.1) 14.5 (3.0)
Block design score 8.6 (3.7) 10.9 (1.9) 10.6 (2.8)
Mississippi 39 scores
109.5 (20.6) 75.4 (10.4) 66.8 (10.3)
Severity of PTSD (CAPS scores)
Reexperiencing 23.1 (8.6) 5.0 (4.4)
Avoidance 23.8 (8.3) 3.1 (3.3)
Hyperarousal 21.7 (7.8) 5.3 (5.5)
Diagnoses other than PTSD
Current MDD 7 (36.8%) 1 (6.3%)
Past MDD 11 (57.9%) 5 (31.3%)
Mississippi 39 score, sample sizes for PTSDC, PTSDKand non-exposed subjects are 18, 16 and 28, respectively.
Relationship between cortisol and age-related memory impairments in Holocaust survivors with PTSD 681
For cortisol levels, as depicted in Fig. 1,a
signiﬁcant Group!Time interaction (Pillai’s
Approximate FZ3.2, dfZ10, 114, pZ0.001) was
noted. Post hoc testing demonstrated that this
interaction reﬂected signiﬁcantly lower cortisol
levels in Holocaust survivors with PTSD at wakeup,
and 0800 as compared to Holocaust survivors
without PTSD and non-exposed subjects. Signiﬁcant
group differences, based on separate ANOVAs, were
observed for maximum cortisol levels (FZ3.40,
dfZ2.62, pZ0.04) and range (maximum–minimum
cortisol levels; FZ4.41, dfZ2.62, pZ0.016),
which, by post-hoc testing, reﬂected differences
between the PTSDCand the two other groups. No
differences were observed in minimum cortisol
levels (FZ0.262, dfZ2.62, n.s.). There were no
group differences in the time of the maximum
(FZ1.5 dfZ2.62, n.s.) and minimum (FZ0.44,
dfZ2.62, n.s.) cortisol values. In 16/28 (57%) of
nonexposed, 9/16 (56%) PTSDKand 9/19 (47.4%) of
PTSDCsubjects the awakening sample corre-
sponded to the maximum cortisol level. The
mimimum cortisol level occurred at the bedtime
sample in 12/28 (43.9%) of nonexposed, 8/16 (50%)
PTSDKand 10/19 (53%) of PTSDCsubjects.
Regarding results of correlational analyses
between mean cortisol and neuropsychological
variables, when the entire sample was considered,
there were no statistically signiﬁcant zero-order
Pearson correlations between any aspects of
memory performance and cortisol nor were there
any signiﬁcant associations when controlling for
WAIS-R Vocabulary, number of years of education
and age at the time of assessment. Table 3
demonstrates that when the partial correlations
were performed for each group separately con-
trolling for the confounding variables, signiﬁcant
correlations were present in the Holocaust survivor
Table 2 Comparison of memory performance measures for non-exposed control subjects and Holocaust Survivors
with (PTSDC) and without (PTSDK) current PTSD; means and standard deviations (in parentheses).
PTSDC(NZ19) M(SD) PTSDK(NZ16) M(SD) Non-exposed (NZ28) M(SD)
High associate pairs 4.2 (1.1) 4.7 (0.93) 5.4 (0.68)
Low associate pairs 1.7 (1.4) 2.9 (1.6) 3.6 (1.5)
Percent high associates 50.0 (12.5) 43.2 (15.0) 48.2 (20.5)
Percent low associates 48.0 (16.8) 42.2 (12.3) 44.6 (19.4)
Figure 1 Salivary cortisol levels from awakening until bedtime. The mean (SD) for cortisol concentrations at the
above-noted time points are as follows for non-exposed, PTSDK, and PTSDCgroups, respectively: wake: 847.7G585.2,
800.0G370.1, 472.4G285.0; 8:00 a.m.: 794.5G548.1, 714.1G330.6, 441.9G253.2; 12:00 p.m.: 356.7G461.7, 472.5G
333.3, 350.4G306.5; 4:00 p.m.: 293.3G458.3, 205.3G96.9, 237.4G210.2; 8:00 p.m.: 154.0G88.4, 148.2G78.5,
236.9G248.2; bedtime: 145.3G80.8, 220.6G385.3, 212.5G340.0. MeanCSD for summary variables for non-exposed,
PTSDKand PTSDC, groups, respectively: mean: 431.9G309.4, 426.8G153.4, 325.3G192.7; maximum 995.2G646.7,
995.7G337.9, 683.5G340.8; minimum: 112.7G73.1, 105.6G64.7, 131.0G109.7; range 882.5G603.8, 890.1G324.0,
552.5G274.1: actual wake-up and bed-time for non-exposed, PTSDKand PTSDCgroups were: wake-up: 7:00 a.m.,
6:50 a.m., 6:28 a.m.; bed-time: 11:13 p.m., 11:38 p.m., 10:50 p.m.
R. Yehuda et al.682
group with PTSD. The correlation between mean
cortisol concentration, estimated by the average
of the salivary samples for each subject, and high
associate learning was signiﬁcantly greater in the
PTSDCgroup compared to the non-exposed group
(Fisher’s zZK0.68, dfZ13, p!0.005). Similarly,
the PTSDCgroup’s correlation between mean
salivary cortisol and low associate recall was
signiﬁcantly greater than that seen in the PTSD-
group (Fisher’s zZK0.79, dfZ13, p!0.005). No
associations were found between any of the
cortisol measures and either the high or low
associate implicit recall in the entire sample or in
any of the three groups.
Table 4 illustrates two-step regression analyses
so as to illustrate, for mean cortisol levels, the
relative contributions of the different covariates to
the associations between cortisol and memory
performance. The ﬁrst step included the demo-
graphic variables that were signiﬁcantly associated
with memory performance using Pearson’s correla-
tional analysis, and the second step included the
additional contribution of cortisol levels. Cortisol
concentrations only contributed signiﬁcantly to
memory performance in the PTSDCgroup, and
only for high associate performance (rZK0.61,
dfZ14, pZ0.01), explaining 28% of the variance
after controlling for other variables. In the no
exposure and PTSDKgroups the percent of added
variance was 2 and 5%, respectively (non-exposed:
rZ0.30, dfZ23, pZ0.14; PTSDK:rZK0.33,
Regression analyses also revealed that demo-
graphic variables contributed differently to the
association between the mean cortisol concen-
trations and memory performance in the three
groups. In the non-exposed group, WAIS Vocabu-
lary scores, but not any of the other variables,
were associated with performance on both high
(rZ0.42, nZ28, pZ0.03) and low (rZ0.56, nZ28,
pZ0.002) associate memory. In the PTSDKgroup
only years of education was associated with high
Table 3 Partial correlations of explicit memory
measures with salivary cortisol levels (minimum,
maximum, mean, range) controlling for education,
IQ, and age.
High Associate Low Associate
min K.67** K.65**
max K.73** K.57*
mean K.66** K.59*
range K.76* K.45
min K.34 .32
max .29 .60*
mean K.17 .42
range .38 .59*
min .41* K.18
max .22 K.20
mean .33 K.16
range .17 K.17
Table 4 Regression analyses of cortisol and demographic factors inﬂuencing memory performance.
PTSDC(NZ19) PTSDK(NZ16) Non-exposed (NZ28)
Age K0.58** K0.52 – K0.30 K0.13 – K0.24 K0.26 –
WAIS vocabulary 0.43 0.28 – 0.30 0.16 – 0.42* 0.35 –
Education 0.03 K0.17 – 0.52* 0.42 – 0.32 0.17 –
All covariates – – 0.30 – – 0.13 – – 0.17
K0.56* K0.52 0.52
K0.11 K0.28 0.15
0.20 0.28 0.22
Age K0.44 K0.22 – K0.50 K0.37 – K0.09 K0.12 –
WAIS vocabulary 0.61** 0.49 – 0.25 0.14 – 0.56** 0.15 –
Education 0.32 0.14 – 0.51* 0.31 – 0.36 0.50 –
All covariates – – 0.33 – – 0.22 – – 0.26
K0.27 K0.38 0.42
0.21 K0.13 K0.16 0.25
Variables were entered in two steps. Step 1: Age, WAIS vocabulary and Education (i.e. all covariates). Step 2: Age, WAIS vocabulary,
Education and Mean Cortisol, *p!0.05, **p!0.01.
Represents the contribution of cortisol and all the covariates to overall R
Relationship between cortisol and age-related memory impairments in Holocaust survivors with PTSD 683
(rZ0.52, nZ16, pZ0.04) and low (rZ0.51, nZ16,
pZ0.04) associates. However, for subjects with
PTSDCage (rZK0.58, nZ19, pZ0.01) and mean
cortisol (rZK0.56, nZ19, pZ0.01) were nega-
tively associated with high associate memory, and
WAIS Vocabulary performance (rZ0.61, nZ19,
pZ0.006) was positively associated with low
associate memory. None of the demographic
factors that were associated with memory per-
formance were also associated with mean cortisol.
The main ﬁnding in the current study is that of a
negative correlation between several measures of
cortisol and explicit memory in Holocaust survivors
with PTSD. Speciﬁcally, a higher mean cortisol
concentration and a steeper range of cortisol values
was associated with poorer memory performance in
Holocaust survivors with PTSD. This relationship
was absent in Holocaust survivors without PTSD and
in demographically comparable subjects who were
not exposed to the Holocaust or other extremely
traumatic events. We did not ﬁnd any associations
between cortisol measures and implicit memory,
which, unlike explicit memory, is not known to be
sensitive to the effects of age-related changes in
cortisol or hippocampal function (Squire, 1992;
Golomb et al., 1993; Lupien et al., 1994).
That we did not observe a relationship between
salivary cortisol levels and memory performance in
the non-PTSD subjects is at variance with some
(Lupien et al., 1994, 1996, 1998; O’Brien et al.,
1994; Seeman et al., 1997; Kalmijn et al., 1998;
Greendale et al., 2000), but not all (Wolf et al.,
2002; Leblhuber et al., 1993; De Bruin et al., 2002),
published reports, and may reﬂect heterogeneity in
the control group that obscured trends or subgroup
effects. Lupien and colleagues carefully categor-
ized healthy controls on the basis of longitudinal
assessments of cortisol using rate of relative change
in cortisol over time, as well as criteria to establish
‘higher’ versus ‘lower’ cortisol levels within the
normal range (Lupien et al., 1994, 1996, 1998).
Indeed, the longitudinal approach may be import-
ant to the extent that age-related cortisol altera-
tions are linked with age-related decline in
performance over a speciﬁed period of time. The
longitudinal approach might also be relevant to
insuring the stability of cortisol alterations. The
current study offering an estimate of diurnal
cortisol release in a single day and in the absence
of a longitudinal perspective in a mixed group of
subjects may not have allowed for the detection of
the relationship between cortisol and memory
observed in other studies. It is also possible that
the failure to obtain a signiﬁcant correlation in the
non-exposed subjects resulted from a ceiling effect
in the performance of the non-exposed subjects
whose mean performance of 5.4G0.68 was close to
a perfect score of 6.0.
It can reasonably be assumed that subjects in the
PTSD group have generally had chronically lower
ambient cortisol levels. The ﬁnding that cortisol
levels were lower in Holocaust survivors with PTSD
was ﬁrst published 10 years ago (Yehuda et al.,
1995b). Similarly, lower ambient cortisol levels
have also been observed by other research groups in
other samples of trauma survivors with chronic
PTSD (Goenjian et al., 1996; Mason et al., 1986;
Heim et al., 1998). Accordingly, it may have been
possible to observe a relationship between cortisol
and memory performance in the current study
because of the relative homogeneity of the PTSD
group with respect to cortisol release. Regardless,
the association between memory performance and
cortisol was negative, reﬂecting greater difﬁculties
in performance in subjects with relatively higher
cortisol levels, even though cortisol levels overall
are lower in PTSD.
The effects of glucocorticoids on memory per-
formance have generally been described as reﬂect-
ing an inverted U shaped function, with too low and
too high levels being associated with impairment,
while moderate levels enhance memory perform-
ance (Erickson et al., 2003). It is, therefore,
plausible that psychiatric disorders associated
with glucocorticoid-related alterations at either
end of the spectrum are associated with poorer
memory performance. The current ﬁndings are
comparable to those of Wolf et al. (2002) who
also demonstrated in a group of elderly subjects
with mild cognitive impairment (MCI), that salivary
morning cortisol levels were decreased compared
to normal elderly and young controls and within the
MCI group mean cortisol levels were inversely
related to immediate recall of paragraph. In that
study as well, no correlations were present
between salivary cortisol measures and memory
performance in the normal elderly or young control
impaired performance in explicit memory is related
to age in subjects with PTSD (Golier et al., 2002),
and that explicit memory is known to decline with
age (Lupien et al., 1994), the ﬁndings may be
indicative of a superimposition of aging and PTSD
effects. Thus, subjects with PTSD who experience
age-related increases in cortisol, or marked
changes in circadian rhythm patterns of cortisol
R. Yehuda et al.684
release over time, are more likely to also experi-
ence age-related decline in explicit memory per-
formance, similar to other aging subjects—a
possibility that can be conﬁrmed by future longi-
tudinal studies of aging trauma survivors.
In Lupien et al. (1998) study, steeper cortisol
elevations over time were not only associated with
poorer explicit memory performance, but also,
reduced hippocampal volume. PTSD has also been
associated with smaller hippocampal volume (Brem-
ner et al., 1995, reviewed in Grossman et al., 2002).
In younger subjects, smaller hippocampal volume
has also been related to poorer performance in tests
of hippocampal-dependent memory (Bremner et al.,
1995), though not to explicit memory using a paired
associates test (Gurvits et al., 1993). In contrast, no
differences in hippocampal volume have been
observed in Holocaust survivors with PTSD compared
to Holocaust survivors without PTSD or appropriate
comparison subjects, despite poorer memory per-
formance (Golier et al., in press). This observation
raises the possibility that a lifetime of chronically
lower ambient cortisol levels in Holocaust survivors
may have provided a compensatory accommodation
that protects the hippocampus from the detrimental
effects of excessive glucocorticoid exposure, even
in the face of cognitive decline. Other age-related
compensatory changes, such as in metabolic clear-
ance of cortisol, are associated with attempts to
preserve normal ambient cortisol levels, at least in
the early stages of normal aging (Barton et al.,
1993). Related to this, we have recently observed
that in Holocaust survivors, there are age-related
reductions in several metabolic enzymes that
increase the concentration of active cortisol rela-
tive to its inactive metabolites (Yehuda et al.,
unpublished data). Thus, although the adrenal gland
is clearly synthesizing, or at least releasing fewer
glucocorticoids in Holocaust survivors than in com-
parison subjects, differences in metabolic factors
that regulate the amount of glucocorticoids avail-
able for binding to receptors in various target tissue
may be relevant to individual differences in the
relative effects of glucocorticoids on cognition and
other parameters. Moreover, age-related
reductions in the antiglucocorticoid DHEA, which
has previously shown to be positively correlated with
PTSD severity (Sondergaard et al., 2002; Rasmusson
et al., 2004) and memory performance (Moffet
et al., 2000; De Bruin et al., 2002), may also play a
role in the alterations observed.
Whether cortisol-related alterations are also
associated with poorer memory performance
observed in younger subjects with PTSD cannot be
determined from this study. Certainly, the current
ﬁndings provide a rationale for examining this issue
in a longitudinal manner and across different
populations of trauma survivors. In
general, evaluating the relationship between glu-
cocorticoids and memory performance in both
nonpsychiatric and psychiatric groups has generally
been more fruitful in studies using neuroendocrine
or behavioral provocations aimed at manipulating
endogenous cortisol levels (Newcomer et al., 1999;
DeQuervain et al., 2000; Kirschbaum et al., 1996;
Domes et al., 2002). Thus, conclusions about the
role of cortisol in memory in Holocaust survivors
must be made cautiously in the absence of
information regarding either the longitudinal
course of memory and cortisol alterations or the
effect of cortisol manipulations on memory in this
This work was supported by NIMH R01 MH64675-01
and VA MERIT review funding to RY. The authors are
grateful to Dr Sonia Lupien for instructing us on the
use of the Paired Associates and Word Stem
Completion Test and her advice in the design of
the study. We thank Drs Hirsch and Labinsky for help
in the diagnostic evaluations, and Drs Ren-Kui Yang,
Ling-Song Guo and Iouri Makotkine for expert
laboratory work in the performance of cortisol
Barton, R.N., Horan, M.A., Weijers, J.W., Sakkee, A.N.,
Roberts, N.A., van Bezooijen, C.F., 1993. Cortisol production
rate and the urinary excretion of 17-hydroxycorticosteroids,
free cortisol, and 6 beta-hydroxycortisol in healthy elderly
men and women. J. Gerontol. 48, M213–M218.
Blake, D.D., Weathers, F.W., Nagy, L.M., Kaloupek, D.G.,
Gusman, F.D., Charney, D.S., Keane, T.M., 1995. The
development of a clinician-administered PTSD scale.
J. Trauma Stress 8, 75–90.
Bremner, J.D., Randall, P., Scott, T.M., Bronen, R.A.,
Seibyl, J.P., Southwick, S.M., Delaney, R.C., McCarthy, G.,
Charney, D.S., Innis, R.B., 1995. MRI-based measurement of
hippocampal volume in patients with combat-related post-
traumatic stress disorder. Am. J. Psychiatry 152, 973–981.
Carlson, L.E., Sherwin, B.B., 1999. Relationships among cortisol
(CRT), dehydroepiandrosterone-sulfate (DHEAS), and mem-
ory in a longitudinal study of healthy elderly men and women.
Neurobiol. Aging 20, 315–324.
Davis, K.L., Davis, B.M., Greenwald, B.S.,Mohs, R.C., Mathe, A.A.,
Johns, C.A., Horvath, T.B., 1986. Cortisol and Alxherimer’s
disease. I. Basal studies. Am. J. Psychiatry 143, 300–305.
De Bruin, V.M.S., Vieira, M.C.M., Rocha, M.N.M., Viana, G.S.B.,
2002. Cortisol and dehydroepinadosterone sulfate plasma
levels and their relationship to aging, cognitive function, and
dementia. Brain Cogn. 50, 316–323.
Relationship between cortisol and age-related memory impairments in Holocaust survivors with PTSD 685
DeLeon, M.J., McRae, T., Tsai, J., George, A., Marcus, D.,
Freedman, M., Wolf, A., McEwen, B.S., 1988. Abnormal
cortisol response in Alxherimer’s disease linked to hippocam-
pal atrophy. Lancet 2, 391–392.
DeLeon, M.J., McRae, T., Rusinek, H., Convit, A., De Santi, S.,
Tarshsh, C., Golomb, J., Volkow, N., Daisley, K.,
Orentreich, N., McEwen, B., 1997. Cortisol reduces hippo-
campal glucose metabolism in normal elderly, but not in
Alzheimer’s disease. J. Clin. Endocrinol. Metab. 82, 3251–
DeQuervain, D.J.F., Roozendaal, B., Nitsch, R.M.,
McGaugh, J.L., Hock, C., 2000. Acute cortisone adminis-
tration impairs retrieval of long-term declarative memory in
humans. Nat. Neurosci. 3, 313–314.
Deuschle, M., Gotthardt, U., Schweiger, U., Weber, B.,
Korner, A., Schmider, J., Standhardt, H., Lammers, C.H.,
Heuser, I., 1997. With aging in humans the activity of the
hypothalamus-pituitary-adrenal system increases and its
diurnal amplitude ﬂattens. Life Sci. 61, 2239–2246.
Dodt, C., Theine, K.J., Uthgenannt, D., Born, J., Fehm, H.L.,
1994. Basal secretory activity of the hypothalamo-pituitary-
adrenocortical axis is enhanced in healthy elderly: an
assessment during undisturbed night-time sleep. Eur.
J. Endocrinol. 131, 443–450.
Domes, G., Heinrichs, M., Reichwald, U., Hautzinger, M., 2002.
Hypothalamic-pituitary-adrenal axis reactivity to psychologi-
cal stress and memory in middle-aged women: high respon-
ders exhibit enhanced declarative memory performance.
Psychoneuroendocrinology 27, 843–853.
Erickson, K., Drevets, W., Schulkin, J., 2003. Glucocorticoid
regulation of diverse cognitive functions in normal and
pathological emotional states. Neurosci. Biobehav. Rev. 3,
Ferrari, E., Magri, F., Dori, D., Migliorati, G., Nescis, T.,
Molla, G., et al., 1995. Neuroendocrine correlates of the
aging brain in humans. Neuroendocrinology 61, 464–470.
Friedman, M., Green, M.F., Sharland, D.E., 1969. Assessment of
hypothalamic-pituitary-adrenal function in the geriatric age
group. J. Gerontol. 24, 292–297.
Goenjian, A.K., Yehuda, R., Pynoos, R.S., Steinberg, A.M.,
Tashjian, M., Yang, R.K., Najarian, L.M., Fairbanks, L.A.,
1996. Basal cortisol, dexamethasone suppression of cortisol
and MHPG in adolescents after the earthquake in Armenia.
Am. J. Psychiatry 153, 929–934.
Golier, J.A., Yehuda, R., Lupien, S.J., Harvey, P.D.,
Grossman, R., Elkin, A., 2002. Memory performance in
Holocaust survivors with posttraumatic stress disorder. Am.
J. Psychiatry 159, 1682–1688.
Golier, J.A., Yehuda, R., Lupien, S.J., Harvey, P.D., 2003.
Memory for trauma-related information in Holocaust survi-
vors with PTSD. Psychiatry Res. 121, 133–143.
Golier, J.A., Yehuda, R., De Santi, S., Segal, S., Dolan, S., de
Leon, M.J., in press. Absence of hippocampal volume
differences in survivors of the Nazi Holocaust with and
without posttraumatic stress disorder. Psychiatry Research
and Neuro imaging.
Golomb, J., de Leon, M.J., Kluger, A., George, A.E., Tarshish, C.,
Ferris, S.H., 1993. Hippocampal atrophy in normal aging. An
association with recent memory impairment. Arch. Neurol.
50 (9), 967–973.
Greendale, G.A., Kritz-Silverstein, D., Seeman, T., Barrett-
Connor, E., 2000. Higher basal cortisol predicts verbal
memory loss in postmenopausal women: Rancho Bernardo
Study. J. Am. Geriatr. Soc. 48, 1645–1658.
Grossman, R.A., Buchsbaum, M., Yehuda, R., 2002. Neuroima-
ging studies in posttraumatic stress disorder. Psychiatr. Clin.
North Am. 25, 317–340.
Gurvits, T.V., Lasko, N.B., Schachter, S.C., Kuhne, A.A.,
Orr, S.P., Pitman, R.K., 1993. Neurological status of Vietnam
veterans with chronic posttraumatic stress disorder.
J. Neuropsychiatry Clin. Neurosci. 5, 183–188.
Heim, C., Ehlert, U., Hanker, J.P., Hellhammer, D.H., 1998.
Abuse-related posttraumatic stress disorder and alterations
of the hypothalamic-pituitary-adrenal axis in women with
chronic pelvic pain. Psychosom. Med. 60, 309–318.
Kalmijn, S., Launer, L.J., Stolk, R.P., de Jong, F.H., Pols, H.A.,
Hofman, A., Breteler, M.M., Lamberts, S.W., 1998. A
prospective study on cortisol, dehydroepiandrosterone sul-
fate, and cognitive function in the elderly. J. Clin. Endocri-
nol. Metab. 83, 3487–3492.
Kern, W., Dodt, C., Born, J., Fehm, H.L., 1996. Changes in
cortisol and growth hormone secretion during nocturnal sleep
in the course of aging. J. Gerontol. A Biol. Sci. Med. Sci. 51,
Kirschbaum, C., Wolf, O.T., May, M., Wippich, W.,
Hellhammer, D.H., 1996. Stress- and treatment-induced
elevations of cortisol levels associated with impaired declara-
tive memory in healthy adults. Life Sci. 58, 1475–1483.
Leblhuber, F., Neubauer, C., Peichl, M., Reisecker, F.,
Steinpartz, F., Windhager, E., Dienstl, E., 1993. Age and
sex differences of dehydroepiandrosterone sulfate and corisol
levels in normal controls and Alherim’ers disease. Pscyho-
pharmacology 111, 23–26.
Lupien, S., Lecours, A.R., Lussier, I., Schwartz, G., Nair, N.P.,
Meaney, M.J., 1994. Basal cortisol levels and cognitive
deﬁcits in human aging. J. Neurosci. 14, 2893–2903.
Lupien, S., Lecours, A.R., Schwartz, G., Sharma, S.,
Hauger, R.L., Meaney, M.J., Nair, N.P., 1996. Longitudinal
study of basal cortisol levels in healthy elderly subjects:
evidence for subgroups. Neurobiol. Aging 17, 95–105.
Lupien, S.J., de Leon, M., de Santi, S., Convit, A., Tarshish, C.,
Nair, N.P., Thakur, M., McEwen, B.S., Hauger, R.L.,
Meaney, M.J., 1998. Cortisol levels during human aging
predict hippocampal atrophy and memory deﬁcits. Nat.
Neurosci. 1, 69–73.
Mason, J.W., Giller, E.L., Kosten, T.R., Ostroff, R.B., Podd, L.,
1986. Urinary free-cortisol levels in posttraumatic stress
disorder patients. J. Nerv. Ment. Dis. 174, 145–149.
Milcu, S.M., Bogdan, C., Nicolau, G.Y., Cristea, A., 1978. Cortisol
circadian rhythm in 70–100 year old subjects. Endocrinologie
Miller, G.A., Chapman, J.P., 2001. Misunderstanding analysis of
covariance. J. Abnorm. Psychol. 110(1), 40–8.
Moffat, S.D., Zonderman, A.B., Harman, S.M., Blackman, M.R.,
Kawas, C., Resnick, S.M., 2000. The relationship between
longitudinal declines in dehydroepiandrosterone sulfate
concentrations and cognitive performance in older men.
Arch. Intern. Med. 160, 2193–2198.
Newcomer, J.W., Selke, G., Melson, A.K., Hershey, T., Craft, S.,
Richards, K., Alderson, A.L., 1999. Decreased memory
performance in healthy humans induced by stress-level
cortisol treatment. Arch. Gen. Psychiatry 56, 527–533.
O’Brien, J.T., Desmond, P., Ames, D., Schweitzer, I.,
Tuckwell, V., Tress, B., 1994. The differentiation of
depression from dementia by temporal lobe magnetic
resonance imaging. Psychol. Med. 24, 633–640.
Parnetti, L., Mecocci, P., Neri, C., Palazzetti, D., Fiacconi, M.,
Santucci, A., et al., 1990. Neuroendocrine markers in aging
brain: clinical and neurobiological signiﬁcance of dexametha-
sone suppression test. Aging (Milano) 2, 173–179.
Peskind, E.R., Wilkinson, C.W., Petrie, E.C., Schellenberg, G.D.,
Raskind, M.A., 2001. Increased CSF cortisol in AD is a function
of APOE genotype. Neurology 56, 1094–1098.
R. Yehuda et al.686
Rasmusson, A.M., Vasek, J., Lipschitz, D.S., Vojvoda, D.,
Mustone, M.E., Shi, Q., Gudmundsen, G., Morgan, C.A.,
Wolfe, J., Charney, D.S., 2004. An increased capacity for
adrenal DHEA release is associated with decreased avoidance
and negative mood symptoms in women with PTSD. Neurop-
sychopharmacology 29, 1546–1557.
Seeman, T.E., McEwen, B.S., Singer, B.H., Albert, M.S.,
Rowe, J.W., 1997. Increase in urinary cortisol excretion and
memory declines: MacArthur studies of successful aging.
J. Clin. Endocrinol. Metab. 82, 2458–2465.
Sondergaard, H.P., Hansson, L.O., Theorell, T., 2002. Elevated
blood levels of dehydroepiandrosterone sulphate vary with
symptom load in posttraumatic stress disorder: ﬁndings from
a longitudinal study of refugees in Sweden. Psychother.
Psychosom. 71, 298–303.
Spitzer, R.L., Williams, J.B.W., Gibbon, M., 1995. Structured
Clinical Interview for DSM-IV (SCID). New York State Psychia-
tric Institute, Biometrics Research, New York, NY.
Squire, L.R., 1992. Memory and the hippocampus: a synthesis
from ﬁndings with rats, monkeys, and humans. Psychol. Rev.
99 (3), 582.
Thaller, V., Vrkljan, M., Hotujac, L., Thakore, J., 1999. The
potential role of hypocortisolism in the pathophysiology of
PTSD and psoriasis. Coll. Antropol. 2, 611–619.
Van Cauter, E., Leproult, R., Kupfer, D.J., 1996. Effects of
gender and age on the levels and circadian rhythmicity of
plasma cortisol. J. Clin. Endocrinol. Metab. 81, 2468–2473.
Wechsler, D., 1981. Wechsler Adult Intelligence Scale—Revised,
Wilkinson, C.W., Peskind, E.R., Raskind, M.A., 1997. Decreased
hypothalamic-pituitary-adrenal axis sensitivity to cortisol
feedback inhibition in human aging. Neuroendocrinology 65,
Wilkinson, C.W., Petrie, E.C., Murray, S.R., Colasurdo, E.A.,
Raskind, M.A., Peskind, E.R., 2001. Human glucocorticoid
feedback inhibition is reduced in older individuals: evening
study. J. Clin. Endocrinol. Metab. 86, 545–550.
Wolf, O.T., Convit, A., Thorn, E., de Leon, M.J., 2002. Salivary
cortisol day proﬁles in elderly with mild cognitive impair-
ment. Psychoneuroendocrinology 27, 777–789.
Yehuda, R., 2002. Current status of cortisol ﬁndings in post-
traumatic stress disorder. Psychiatr. Clin. North Am. 25,
Yehuda, R., Keefe, R.S.E., Harvey, P.D., Levengood, R.A.,
Gerber, D.K., Geni, J., 1995a. Learning and memory in
combat veterans with posttraumatic stress disorder. Am.
J. Psychiatry 152, 137–139.
Yehuda, R., Kahana, B., Binder-Brynes, K., Southwick, S.M.,
Mason, J.W., Giller, E.L., 1995b. Low urinary cortisol
excretion in Holocaust survivors with posttraumatic stress
disorder. Am. J. Psychiatry 152, 982–986.
Yehuda, R., Teicher, M.H., Trestman, R.L., Levengood, R.A.,
Siever, L.J., 1996. Cortisol regulation in posttraumatic stress
disorder and major depression: a chronobiological analysis.
Biol. Psychiatry 40, 79–88.
Yehuda, R., Halligan, S.L., Grossman, R., Golier, J.A., Wong, C.,
2002. The cortisol and glucocorticoid receptor response to
low dose dexamethasone administration in aging combat
veterans and holocaust survivors with and without posttrau-
matic stress disorder. Biol. Psychiatry 52, 393–403.
Yehuda, R., Golier, J.A., Halligan, S.L., Harvey, P.D., 2004.
Learning and memory in Holocaust survivors with posttrau-
matic stress disorder. Biol. Psychiatry 55, 291–295.
Yehuda, R., Golier, J., Kaufman, S., 2005. Circadian rhythm of
salivary cortisol in Holocaust Survivors with and without
PTSD. Am. J. Psychiatry.
Yehuda, R., Andrew, R., Bierer, L.B., Seckle, J., submitted for
publication. Altered cortisol metabolism in Holocaust survi-
vors: permanent ‘programming’ of metabolism. New England
Journal of Medicine.
Zovato, S., Simoncini, M., Gottardo, C., Pratesi, C.,
Zampollo, V., Spigariol, A., 1996. Dexamethasone suppres-
sion test: corticosteroid receptors regulation in mononuclear
leukocytes of young and aged subjects. Aging (Milano) 8,
Relationship between cortisol and age-related memory impairments in Holocaust survivors with PTSD 687