Psychological Medicine (2013), 43, 495505.

f Cambridge University Press 2012 O R I G I N A L AR T I C LE

doi:10.1017/S0033291712001961

Eects of cortisol on memory in women with

borderline personality disorder: role of co-morbid
post-traumatic stress disorder and major depression

K. Wingenfeld1,2*, M. Driessen3, K. Terfehr2, N. Schlosser3, S. Carvalho Fernando3, C. Otte1, T. Beblo3,

C. Spitzer4, B. Lowe2 and O. T. Wolf5
1
Department of Psychiatry and Psychotherapy, Charite University Berlin, Campus Benjamin Franklin, Berlin, Germany
2
Department of Psychosomatic Medicine and Psychotherapy, University Medical Center Hamburg-Eppendorf and Schon Klinik Hamburg-Eilbek,
Hamburg, Germany
3
Department of Psychiatry and Psychotherapy, Ev. Hospital Bielefeld-Bethel, Bielefeld, Germany
4
Department of Psychiatry and Psychotherapy, Asklepios Fachklinikum Tiefenbrunn, Gottingen, Germany
5
Department of Cognitive Psychology, Ruhr University Bochum, Bochum, Germany

Background. Stress and cortisol administration are known to have impairing eects on memory retrieval in healthy
humans. These eects are reported to be altered in patients with major depressive disorder (MDD) and post-
traumatic stress disorder (PTSD) but they have not yet been investigated in borderline personality disorder (BPD).

Method. In a placebo-controlled cross-over study, 71 women with BPD and 40 healthy controls received either
placebo or 10 mg of hydrocortisone orally before undertaking a declarative memory retrieval task (word list learning)
and an autobiographical memory test (AMT). A working memory test was also applied.

Results. Overall, opposing eects of cortisol on memory were observed when comparing patients with controls. In
controls, cortisol had impairing eects on memory retrieval whereas in BPD patients cortisol had enhancing eects
on memory retrieval of words, autobiographical memory and working memory. These eects were most pronounced
for specicity of autobiographical memory retrieval. Patients with BPD alone and those with co-morbid PTSD
showed this eect. We also found that co-morbid MDD inuenced the cortisol eects : in this subgroup (BPD+MDD)
the eects of cortisol on memory were absent.

Conclusions. The present results demonstrate benecial eects of acute cortisol elevations on hippocampal-mediated
memory processes in BPD. The absence of these eects in patients with co-morbid MDD suggests that these patients
dier from other BPD patients in terms of their sensitivity to glucocorticoids (GCs).

Received 4 June 2012 ; Revised 19 July 2012 ; Accepted 23 July 2012 ; First published online 19 September 2012

Key words : Autobiographical memory, borderline personality disorder, cortisol, declarative verbal memory, HPA axis,
major depression, post-traumatic stress disorder, working memory.

Introduction and shifting co-morbid disorders) was found to have

strong predictive power for the borderline diagnosis
Borderline personality disorder (BPD) is a complex
(Zanarini et al. 1998). Furthermore, patients with BPD
and serious mental disorder that is characterized
frequently report early, multiple and chronic adverse
by intense and rapidly changing mood states, im-
or even traumatic experiences, such as repeated sexual
pulsivity, self-injurious behaviors, fear of abandon-
or physical abuse or emotional or physical neglect
ment, unstable relationships and self-image (First et al.
(Zanarini et al. 1997 ; Golier et al. 2003).
1997). Patients with BPD often suer from co-morbid
Because early life stress and traumatization are
axis I disorders, with mood disorders and anxiety
major risk factors for the development and persistence
disorders being the most prominent ones (Zanarini
of mental disorders, many studies have investigated
et al. 1998). Of note, complex co-morbidity (i.e. multiple
the functioning of the hypothalamicpituitaryadrenal
(HPA) axis in major depressive disorder (MDD) and
post-traumatic stress disorder (PTSD). Although there
* Address for correspondence to : Dr K. Wingenfeld, Department of
Psychiatry and Psychotherapy, Charite University Berlin, Campus
is evidence for enhanced cortisol release and reduced
Benjamin Franklin, Eschenallee 3, 14050 Berlin, Germany. negative-feedback sensitivity in MDD, PTSD seems
(Email : [email protected]) to be characterized by reduced cortisol release and
496 K. Wingenfeld et al.

enhanced feedback sensitivity (Plotsky et al. 1998 ; BPD patients with co-morbid MDD might show no
Pariante & Miller, 2001 ; Yehuda, 2002). In BPD there is eects of cortisol on memory, whereas in healthy
increasing evidence that alterations in HPA activity controls, cortisol administration would lead to an im-
may contribute to the disorder (Wingenfeld et al. pairment in memory performance.
2010b). Similarly, enhanced basal cortisol concentra-
tions and reduced feedback sensitivity have been
reported for BPD patients (Rinne et al. 2002a ; Lieb et al. Method
2004 ; Wingenfeld et al. 2007a, b). However, the nd-
Participants
ings are heterogeneous and it has been hypothesized
that co-morbid symptomatology might play a role Seventy-one women with BPD and 40 healthy control
(Wingenfeld et al. 2010b). women aged o18 years participated. Thirty-six of
Alterations of HPA axis feedback regulation have the control participants were part of a former study
been interpreted in terms of altered glucocorticoid by our group (Terfehr et al. 2011b), as were 21 of the
receptor (GR) function (Holsboer, 2000 ; Yehuda, 2009 ; patients, that is BPD patients with co-morbid PTSD
Rohleder et al. 2010). Cortisol acts by binding to min- (Wingenfeld et al. 2012). The BPD+PTSD subgroup of
eralocorticoid receptors (MRs) and GRs (de Kloet, our former PTSD study was also included in this study
2003). These receptors show a high density in the as a comparison group for the 50 BPD patients without
hippocampus and the prefrontal cortex (Lupien & PTSD (BPDPTSD) who were recruited exclusively for
Lepage, 2001 ; de Kloet, 2003), which are closely re- this study.
lated to cognitive functions. In healthy participants Participants were excluded if they had any of the
there is evidence that acute glucocorticoid (GC) ad- following medical conditions : central nervous system
ministration impairs memory retrieval (Wolf, 2003, (CNS) diseases or severe somatic diseases, metabolic
2009). It has also been observed that cortisol adminis- diseases, organic shift in cortisol secretion, immune-
tration leads to poorer working memory performance mediated diseases, medicated hypertension, current
(Lupien et al. 1999). infections, or pregnancy. Further exclusion criteria
In patients with BPD the eects of cortisol on were anorexia, schizophrenia, alcohol or drug depen-
memory have not yet been investigated. In a recent dence, bipolar disorder, schizo-aective disorder,
study we found that in PTSD patients, in contrast to major depression with psychotic symptoms (all as-
healthy controls, administration of hydrocortisone sessed by the SCID), attention decit hyperactivity
enhances rather than impairs memory retrieval disorder (ADHD) or cognitive impairment. Intake of
(Wingenfeld et al. 2012). The results were independent antidepressants did not lead to exclusion.
from co-morbid borderline features of the patients. Written informed consent was obtained from all
Our ndings are in line with another study in PTSD participants. Healthy participants were recruited by
patients (Yehuda et al. 2007) but others disagree local advertisement and received nancial remuner-
(Bremner et al. 2004 ; Grossman et al. 2006 ; Yehuda ation (100 E). The study was approved by the Ethics
et al. 2010). We have also shown, in a series of studies Committees of the University of Muenster and the
with patients with MDD, that cortisol administration Medical Council of Hamburg.
did not inuence declarative memory retrieval and
working memory (Schlosser et al. 2010 ; Terfehr et al.
Procedure
2011a, b).
The aim of the current study was to investigate To assess psychiatric diagnoses we used SCID-I and
the eects of cortisol administration on memory in SCID-II (First et al. 1997). PTSD symptoms (patients
patients with BPD. To test several domains of mem- only) were assessed with the Post-traumatic Stress
ory, an autobiographical memory test (AMT) and a Diagnostic Scale (PDS ; Foa, 1995), depressive mood
word list paradigm to investigate declarative memory state was measured using the Beck Depression
were conducted, along with a working memory test. Inventory (BDI ; Beck & Steer, 1994) and childhood
Because of the dierent patterns of cortisol eects trauma was assessed with the Childhood Trauma
on memory in MDD and PTSD, and given that these Questionnaire (CTQ ; Bernstein et al. 2003 ; Wingenfeld
co-morbid disorders may contribute to HPA axis dys- et al. 2010a).
regulations in BPD, we further aimed to investigate In this placebo-controlled cross-over study, each
the impact of co-morbid MDD and PTSD. Given the participant was tested twice with parallel versions of a
high prevalence of traumatic experiences in BPD, word list paradigm and an AMT. All tests were taken
we hypothesized that BPD patients would show in the afternoon. The two versions of the tests were
enhanced memory retrieval after cortisol, as was counterbalanced across the test conditions. On the rst
shown recently in PTSD patients. On the contrary, day all participants learned the word list. On the
 Cortisol and memory in BPD 497

Table 1. Experimental protocol indicating time of cognitive Results

testing
Demographic and clinical data
Day Time (h) Procedurea Demographic and clinical data are presented in
Table 2. Patients with BPD and healthy controls were
1 1530 Word list learning and immediate recall comparable in body mass index (BMI), years of edu-
2 1545 Administration of hydrocortisone or placebo cation and intake of oral contraceptives. The BPD
1615 Word list (delayed recall) patients were younger than the controls and there
1645 Autobiographical memory test (AMT)
were more smokers in the patient group. Patients
1700 Working memory task/no additional task
had a signicantly higher depression score and
a they reported more childhood trauma. Patients with
One week later the study protocol was repeated using the
alternate condition (hydrocortisone/placebo) and parallel co-morbid PTSD had higher PDS scores whereas
versions of the memory tests. patients with and without co-morbid MDD did not
dier regarding BDI scores.
For the second part of our study we examined the
second day, 30 min after administration of 10 mg eect of co-morbid MDD and PTSD with BPD : 23 BPD
hydrocortisone (Jenapharm1, Germany) or placebo, patients suered from co-morbid MDD (BPD+MDD)
the participants were asked to recall the words from and 11 from co-morbid PTSD (BPD+PTSD). A fur-
the word list, and then the AMT was performed. The ther 10 patients had both co-morbid disorders
working memory task (using the Word Suppression (BPD+MDD+PTSD) whereas 27 had neither MDD
Test, WST ; Terfehr et al. 2011 b) was only tested once in nor PTSD (BPD). The subgroup of BPD patients with
each participant after hydrocortisone or placebo, that co-morbid MDD but without co-morbid PTSD had
is a between-subjects design was realized. The parti- the lowest CTQ score whereas the two subgroups with
cipants were tested in a quiet room and were only co-morbid PTSD had the highest scores for self-
allowed to drink water. The same procedure with the reported trauma (Table 2). BDI scores did not dier
alternate test condition (hydrocortisone/placebo) was between the groups. In addition, the following axis I
repeated after 1 week (see Table 1). disorders were diagnosed : anxiety disorder apart
from PTSD (n=13), pain disorder (n=1), eating dis-
order (n=7), and alcohol abuse (n=1).
The patients took the following psychiatric medi-
Memory tasks
cations : selective serotonin reuptake inhibitor (SSRI)
A modied version of the AMT (Buss et al. 2004) n=22, selective noradrenergic and serotonergic re-
was applied. The word list paradigm consisted of uptake inhibitor n=18, neuroleptics n=17, selective
21 words. For a detailed description of the tests see noradrenergic reuptake inhibitor n=1, tricyclic anti-
Wingenfeld et al. (2012). To test working memory, we depressant n=1, anticonvulsive n=1. Sixteen patients
administered the self-developed WST (Terfehr et al. took more than one drug ; 31 of the patients were
2011b). The WST consists of two parts, one with medication free.
negative and one with neutral interference words.
Each part consists of 14 recorded trials with a series
of alternately presented digits and interference Memory retrieval
words, which are presented verbally. The participants
AMT
have to memorize the digits but ignore the interference
words. A repeated-measures 2r2 ANCOVA was conducted
with the main factors group (BPD patients versus
healthy controls) and condition (placebo versus hy-
drocortisone). Intake of oral contraceptives (p=0.40)
Statistical analysis
and smoking (p=0.36) as additional factors had no
Statistical analyses were performed using SPSS ver- eects and were excluded from the analyses. Age was
sion 18.0 (SPSS Inc., USA). Demographic data were introduced as a covariate because of its signicant
analyzed using Pearsons x2 test for categorical data impact (p=0.013).
and Students t test for continuous data. The eects There was a signicant group by condition interac-
of hydrocortisone on memory performance (word tion eect (F1,108=7.628, p=0.007) but no eect of the
list, AMT) were analyzed using an ANCOVA with main factors group (p=0.22) and condition (p=0.95).
repeated measurements and a univariate ANCOVA A post-hoc univariate ANCOVA showed that the con-
(working memory test). trol group performed better than the BPD patients
 498
K. Wingenfeld et al.
Table 2. Sociodemographic and clinical characteristics

Characteristics BPD (n=71) Controls (n=40) Statistics

Age (years) 28.2 (7.9) 32.9 (10.8) tdf109=x2.61, p=0.01

BMI (kg/m2) 25.1 (6.2) 23.5 (4.6) tdf103=1.46, p=0.15
Years of education 11.4 (1.5) 11.6 (1.5) tdf106=x0.767, p=0.45
Oral contraceptives (yes/no) 19/51 13/27 x2df1=0.35, p=0.66
Smoker (yes) 44/26 12/28 x2df1=10.99, p<0.001
BDI sum 26.2 (9.1)a 3.0 (3.9) tdf107=18.29, p<0.001
CTQ total score 60.4 (18.1)a 38.4 (16.1) tdf105=6.33, p<0.001
BPD+PTSDa BPDPTSDb
(n=20) (n=38)
PDS total score 32.2 (7.9) 19.3 (12.1) tdf56=4.28, p<0.001
BPD+MDD BPDMDD
(n=31) (n=38)
BDI sum 27.0 (8.4) 25.6 (9.7) tdf67=0.63, p=0.53
BPD only (n=25) BPD+MDD+PTSD (n=10) BPD+PTSD (n=11) BPD+MDD (n=21)
CTQ total score 54.5 (16.6) 80.1 (15.6) 72.3 (14.7) 51.8 (11.5) Fdf3,63=12.05, p<0.001c
BDI sum 23.6 (8.9) 27.4 (7.1) 30.6 (10.3) 26.8 (9.1) Fdf3,63=1.73, p=0.17

BPD, Borderline personality disorder ; BMI, body mass index ; BDI, Beck Depression Inventory ; CTQ, Childhood Trauma Questionnaire ; PDS : Post-traumatic Stress Diagnostic Scale ;
PTSD, post-traumatic stress disorder ; MDD, major depressive disorder.
a
Two patients did not answer the BDI, four patients did not answer the CTQ, one patient with PTSD did not answer the PDS.
b
Twenty-two of the patients had no trauma as asked in the PDS.
c
Bonferroni post-hoc test : BPD only=BPD+MDD <BPD+PTSD=BPD+MDD+PTSD.
 Cortisol and memory in BPD 499

Table 3. Autobiographical memory test (AMT) : specicity of The analysis was controlled for age (F1,101=6.141,
memory retrieval in patients with borderline personality disorder p=0.015). Intake of oral contraceptives (p=0.26) and
(BPD) and healthy controls smoking (p=0.67) had no eect and were therefore
excluded from the analyses.
Memory BPD Controls Post-hoc A post-hoc ANCOVA revealed a signicant dier-
specicity (n=71) (n=40) test ence between patients and controls after placebo
(group eect : F1,103=7.818, p=0.006) but not after
Placebo 3.9 (1.7) 4.5 (1.6) p=0.03 hydrocortisone (group eect : p=0.49). Separate post-
Hydrocortisone 4.2 (1.7) 4.0 (1.7) p=0.84
hoc ANOVAs with repeated measures for each group
Post-hoc test p=0.06 p=0.04
revealed no signicant eects of hydrocortisone
ANCOVA : group by condition interaction eect p=0.007 (BPD patients : p=0.17 ; controls : p=0.34). As shown
in Fig. 2, the control participants performed better
than the patients in the placebo condition but not
after placebo (F1,108=4.616, p=0.03) but not after after hydrocortisone intake.
hydrocortisone (p=0.84). Post-hoc ANOVAs with The analyses of subgroups (5r2 ANOVA) did not
repeated measures performed separately for each reveal any signicant eects.
group revealed a signicant eect of condition in the
control group (F1,39=4.418, p=0.04) and a trend sig-
nicant eect in the BPD group (F1,70=3.610, p=0.06).
Working memory test : the WST
Table 3 shows that the specicity of memory retrieval
was impaired after hydrocortisone compared to Thirty-three BPD patients and 19 controls performed
placebo in the control group but enhanced in the the WST after hydrocortisone intake, and 34 patients
BPD group. and 21 controls after placebo. A 2r2 ANCOVA with
To analyze the impact of co-morbid PTSD and the main factors group (BPD versus controls) and con-
MDD, a repeated-measures 5r2 ANCOVA was dition (hydrocortisone versus placebo) was performed
performed. Age served as a covariate (p=0.017). Five for each test part (neutral and negative).
subgroups were compared : healthy controls (n=40), In the test part with neutral interference words, we
BPD patients without PTSD and MDD (n=27), BPD revealed a signicant eect of the covariate age
patients with PTSD and MDD (n=10), BPD patients (F1,102=5.821, p=0.018). There was no eect of con-
with PTSD but without MDD (n=11) and BPD dition (p=0.87) and the group eect also only reached
patients with MDD but without PTSD (n=23). A trend signicance (F1,102=2.783, p=0.098), suggesting
signicant group by condition interaction eect was a slightly better performance in the control group
revealed (F4,105=4.014, p=0.005). Post-hoc ANOVAs compared to the patients (Fig. 3 a). There was no con-
showed a better memory performance after hydro- dition by group interaction (p=0.51).
cortisone in all patient groups except the BPD In the test part with negative interference words,
patients with MDD and no PTSD. The results of the a signicant group eect was seen (F1,102=4.056,
subgroup analyses including the post-hoc tests are p=0.047), showing a better performance in the control
shown in Fig. 1. Medication intake (yes/no) was dis- participants than the patients. There was no signi-
tributed similarly over the patient groups (x2=3.413, cant eect of condition (p=0.428) or condition by
p=0.33). group interaction eect (p=0.126). The eect of the
covariate age failed to show signicance (F1,102=3.123,
p=0.08) and we therefore repeated the analyses with-
Word list paradigm
out the covariate. Although the group eect dimin-
A repeated-measures 2r2 ANCOVA was conducted ished (p=0.12), a trend toward a group by condition
with the main factors group and condition. The per- interaction eect could be seen (F1,103=3.583, p=0.06)
centage of correctly recalled words relative to the (Fig. 3 b). A post-hoc t test revealed no dierence
words recalled after the fth learning trial on the day between placebo and hydrocortisone in the control
before was used as the dependent variable. Data were group (p=0.41) but the BPD patients (t65=2.017,
missing for this test in two control participants and p=0.048) showed better working memory perform-
ve patients. ance after cortisol.
We found a trend for a group by condition inter- Because of the between-subject design of this part
action eect that marginally failed signicance of the study, the subgroups with dierent co-morbid
(F1,101=3.328, p=0.07). Although there was no main disorders (i.e. PTSD and MDD) were relatively
eect of condition (F1,101=1.176, p=0.28), a signicant small. Thus, we refrained from further subgroup
group eect was revealed (F1,101=4.819, p=0.03). analyses.
500 K. Wingenfeld et al.

5.0
Placebo
Cortisol
Number specific events retrieved

4.5

4.0

3.5

3.0
Controls BPD BPD BPD BPD
n = 40 n = 27 MDD PTSD PTSD MDD
p = 0.04 p = 0.03 n = 10 n = 11 n = 22
p = 0.04 p = 0.07 p = 0.49

Fig. 1. Memory specicity after placebo and hydrocortisone : analyses of borderline personality disorder (BPD) subgroups with
dierent co-morbid disorders. PTSD, post-traumatic stress disorder ; MDD, major depressive disorder ; n, sample size of each
subgroup ; p, post-hoc t test placebo versus cortisol.

Placebo Hydrocortisone autobiographical memory and word list learning) and

working memory but also investigated whether the
80 p = 0.003
presence or absence of co-morbid MDD and/or PTSD
aected the results.
Words recalled, mean (S.E.), %

In both declarative memory tasks it could be shown

60
that, in contrast to healthy controls, cortisol had en-
hancing rather than impairing eects on memory
retrieval in BPD patients. A similar eect was seen in
40
the working memory task but only when negative
interference words were presented. To summarize,
although substantially dierent memory tasks were
20
used, we found remarkably similar response patterns
between the paradigms.
In the AMT we found dierential eects with re-
0
Control BPD spect to co-morbid mental disorders : BPD patients
with and without co-morbid PTSD showed an im-
Fig. 2. Percentage of words retrieved in the word list
provement in memory retrieval whereas in BPD
paradigm in relation to the last learning list on the previous
day [mean (S.E.)] in patients with borderline personality
patients with co-morbid MDD hydrocortisone admin-
disorder (BPD) and healthy control subjects after placebo and istration had no eect on autobiographical memory
after administration of 10 mg hydrocortisone. A signicant retrieval. It is important to mention that the enhancing
main eect of group (p=0.03) and a trend towards a eects of cortisol on memory retrieval were not only
condition by group interaction eect was found (p=0.07). associated with co-morbid PTSD but were also seen in
the BPD group without co-morbid PTSD (and MDD).
Thus, it seems that BPD and PTSD share similarities in
Discussion
there response to exogenous cortisol on memory re-
This is the rst study to investigate the eects trieval.
of cortisol administration on memory in patients Alterations in the reactivity to cortisol adminis-
with BPD. We not only tested dierent memory tration and their impact on memory have been
domains (i.e. declarative memory retrieval including mostly interpreted in terms of altered GR functioning
 Cortisol and memory in BPD 501

(a) (b) Placebo

Number of correctly reproduced sequences

Cortisol
Number of correctly reproduced sequences
10 10
p = 0.048

8 8

6 6

4 4

2 2

0 0
Con BPD Con BPD

Fig. 3. Number of correctly reproduced sequences [mean (S.E.)] in patients with borderline personality disorder (BPD) and
healthy control women after placebo and after administration of 10 mg hydrocortisone in the (a) neutral and (b) negative trials.
In the negative test part there was a trend towards a group by condition interaction (p=0.06).

(Bremner et al. 2004 ; Yehuda et al. 2007 ; Wingenfeld with dierent HPA axis-related disturbances
et al. 2012). In PTSD there is evidence for enhanced GR (Wingenfeld et al. 2010b).
sensitivity (Rohleder et al. 2004 ; Yehuda et al. 2004) The impairing eects of cortisol on memory re-
whereas major depression seems to be characterized trieval are well documented in the literature (Wolf,
by a reduced GR functioning (Holsboer, 2000). No 2009). This has been shown for declarative memory
study has directly investigated GRs in BPD patients. (Kuhlmann et al. 2005a, b) and also for autobiographi-
However, to understand the results of our study it cal memory retrieval (Buss et al. 2004 ; Young et al.
would be of interest to look at HPA axis ndings in 2011) and working memory (Lupien et al. 1999 ; Wolf
BPD. et al. 2001). Thus, our results regarding the control
Compared to studies on MDD and PTSD, there are group are in line with these previous studies. In the
fewer studies on HPA axis alteration in patients with current study this eect was strongest in the AMT.
BPD providing evidence for higher basal cortisol re- Overall, the eects in controls seemed to be weaker
lease in these patients (Lieb et al. 2004), but it does than in our previous studies (Schlosser et al. 2010 ;
seem that depressive and PTSD symptoms inuence Terfehr et al. 2011a, b). One possible reason might
cortisol secretion in BPD (Wingenfeld et al. 2007a). relate to the restriction of the sample on women
Concerning feedback regulation, older studies that (Kuhlmann et al. 2005 a ; Schoofs & Wolf, 2009).
used the 1-mg dexamethasone suppression test (DST) Furthermore, compared to other studies we only ad-
found high rates of non-suppressors in BPD popula- ministered a low dosage of hydrocortisone, which
tions, but most of these results suggested an associ- might also be responsible for the weaker eects
ation of reduced feedback inhibition with aective (Het et al. 2005 ; Young et al. 2011). Nevertheless, cor-
dysregulations or even with co-morbid MDD (see tisol administration in healthy participants typically
Wingenfeld et al. 2010b for review). Using the low- does not lead to enhanced but to impaired memory
dose DST, a reduced rather than an enhanced cortisol retrieval.
suppression after 0.5 mg of dexamethasone in BPD Some studies have investigated the eects of GC
patients compared to healthy controls has been re- administration on memory in PTSD patients. In one of
ported (Lieb et al. 2004). In our own studies we our former studies (Wingenfeld et al. 2012) we also
again found evidence for the impact of co-morbid found an improvement of memory retrieval in PTSD
MDD and PTSD (Lange et al. 2005 ; Wingenfeld et al. patients. Thus, our previous and present studies both
2007b, c), which is also supported by another study suggest that cortisol improves memory retrieval in
using the combined dexamethasone/corticotrophin- BPD and in PTSD patients. This is in line with another
releasing factor (DEX/CRF) test (Rinne et al. 2002a). In study that also showed enhancing eects of cortisol on
sum, there is no clear picture of HPA axis dysregula- memory performance in PTSD (Yehuda et al. 2007).
tion in BPD. There is evidence for both enhanced The results might be interpreted in the context of
and reduced feedback sensitivity of the HPA axis an enhanced reactivity to exogenous cortisol in these
and therefore enhanced and reduced GR functioning. patients compared to healthy participants. However,
One possible explanation is the existence of subgroups not all studies conrm these results (Bremner et al.
502 K. Wingenfeld et al.

2004 ; Grossman et al. 2006 ; Yehuda et al. 2010), described above was found in response to hydrocor-
although these studies did not separate memory tisone administration, namely an enhanced hippo-
consolidation from retrieval. Of note, a study that in- campal activity (Yehuda et al. 2010). Neural activity in
vestigated Vietnam veterans with and without PTSD other brain regions was not reported in this study.
using positron emission tomography (PET) revealed For borderline patients a dysfunctional frontolimbic
an enhanced reactivity to hydrocortisone adminis- network including the amygdala, prefrontal areas and
tration of the hippocampus in the PTSD group other limbic structures (the anterior cingulate cortex,
(Yehuda et al. 2010). Furthermore, there is evidence for ACC), has been proposed from imaging studies
an enhanced GR sensitivity in PTSD (Rohleder et al. (Schmahl & Bremner, 2006 ; Wingenfeld et al. 2010b).
2004 ; Yehuda et al. 2004), which might also contribute However, the eects of cortisol administration on
to the eect of cortisol on memory in PTSD. However, brain activity in BPD have not yet been investigated.
in a recent study (consisting of a subsample of this Thus, there are at least two possible explanations
study) we found higher cortisol levels before and for our results : rst, comparable to PTSD patients,
after dexamethasone administration in BPD patients, hydrocortisone administration in BPD might lead
but no evidence for alterations concerning feedback to an activation in the hippocampus that enhances
regulation (Carvalho Fernando et al. 2012). Co-morbid memory retrieval. Alternatively, hydrocortisone might
PTSD and MDD did not contribute to cortisol release reduce brain activity in regions that are hyperactive
in this study. Thus, the sensitivity of GRs in BPD in BPD, such as temporal areas. Imaging studies are
remains unclear. However, it seems unlikely that required to investigate this topic.
an enhanced or a reduced feedback sensitivity of Animal studies might also help in understanding
the HPA axis is a unique explanation for the results our results. The current ndings of enhanced memory
presented here. after cortisol treatment in BPD patients share simila-
Major depression is characterized by a reduced GR rities with recent observations in rodents that have
sensitivity (Holsboer, 2000). In a series of studies we been exposed to stress early in life (Champagne et al.
did not nd any impairing eects of hydrocortisone 2008). The early stressed rats displayed an impaired
administration on memory retrieval, including de- neural plasticity, that is long-term potentiation (LTP),
clarative memory retrieval (Terfehr et al. 2011a), in adulthood. Corticosterone treatment enhanced
specicity of autobiographic memory (Schlosser et al. the LTP in these animals but impaired it in the non-
2010) and working memory (Terfehr et al. 2011b). We stressed control animals (Champagne et al. 2008).
interpreted these results in MDD in terms of reduced Thus, early adversity seems to inuence the response
GR functioning. Of note, in the current study, BPD of the hippocampus to GCs adulthood. Early life stress
patients with co-morbid MDD but without PTSD also has also been discussed as an important risk factor for
showed this lack of cortisol eect on autobiographical the development of BPD (Zanarini et al. 1997) and
memory. Thus, it seems that patients with MDD and early trauma is known to have long-lasting eects on
BPD patients with co-morbid MDD share a similar the regulation of the HPA axis (Heim et al. 2000)
pattern in terms of cortisol eects on memory and, and GR functioning, possibly through epigenetic me-
thus, possibly in terms of reduced GR function. chanisms (McGowan et al. 2009). Of note, the current
Our main result is that cortisol administration sample also reported high levels of adverse childhood
had an enhancing rather than an impairing eect on experiences.
memory performance in patients with BPD. This sug- Benecial eects of cortisol have been shown in the
gests an enhanced reactivity to exogenous cortisol in context of prevention of PTSD symptoms after acute
these patients. trauma experiences (Schelling et al. 2006). Thus, there
Imaging studies with healthy participants show is growing evidence for, in part, benecial eects of
that cortisol administration leads to a reduced activity cortisol in PTSD (Aerni et al. 2004). In BPD, which as a
in the hippocampus during resting-state conditions personality disorder has an early onset, such studies
(Lovallo et al. 2010) and memory retrieval (de are lacking, but it would be of interest to investigate
Quervain et al. 2003 ; Oei et al. 2007 ; Weerda et al. 2010). whether targeting the HPA axis might be a therapeutic
Furthermore, a reduced activity of medial temporal tool. In this connection, treatment with an SSRI was
regions has been reported along with a reduced ac- reported to reduce HPA axis hyperactivity in BPD
tivity of prefrontal brain regions (de Quervain et al. (Rinne et al. 2002b). Furthermore, in depression there
2003 ; Henckens et al. 2011). The reduced brain acti- have been eorts to investigate drugs that inuence
vation after cortisol was associated with cortisol- the HPA axis (Schule et al. 2009). To summarize, the
induced memory retrieval impairment (de Quervain benecial eects of hydrocortisone in BPD, such as
et al. 2003). In psychiatric patients comparable studies normalization on memory processes, should be in-
are rare. In PTSD patients a dierent pattern to the one vestigated further.
 Cortisol and memory in BPD 503

Limitations Acknowledgments
As there were no additional measurements of HPA This study was supported by a Deutsche Forschungs-
axis functioning for this sample (e.g. basal cortisol gemeinschaft grant (WI 3396/1-1) awarded to K.W.,
levels, feedback sensitivity, cortisol bioavailability), O.T.W. and M.D.
interpretation of the data is dicult and remains some-
what speculative. It would be of interest to combine
our experimental design with direct measurements of Declaration of Interest
GR functioning, HPA axis feedback sensitivity, basal None.
cortisol levels and neuroimaging. We also have no
data on basal cortisol levels or on cortisol levels after
drug administration, and therefore no treatment References
check. Several studies, including our own, have shown
Aerni A, Traber R, Hock C, Roozendaal B, Schelling G,
that administration of 10 mg hydrocortisone leads to a
Papassotiropoulos A, Nitsch RM, Schnyder U,
substantial increase in salivary cortisol concentrations de Quervain DJ (2004). Low-dose cortisol for symptoms
(Buss et al. 2004 ; Terfehr et al. 2011a, b). of posttraumatic stress disorder. American Journal of
Data on menstrual cycle phase were also not ob- Psychiatry 161, 14881490.
tained. Many patients were medicated, which could Beck AT, Steer RA (1994). The Beck Depression Inventory (BDI)
also have inuenced the results. However, in our pre- [in German]. Huber : Bern.
vious studies we found no eects of medication intake Bernstein DP, Stein JA, Newcomb MD, Walker E, Pogge D,
on the cortisol-induced memory eect (Terfehr et al. Ahluvalia T, Stokes J, Handelsman L, Medrano M,
2011a ; Wingenfeld et al. 2012). Desmond D, Zule W (2003). Development and validation
of a brief screening version of the Childhood Trauma
For some of our analyses (subgroup analyses,
Questionnaire. Child Abuse and Neglect 27, 169190.
working memory test) the sample size was relatively
Bremner JD, Vythilingam M, Vermetten E, Afzal N,
small leading to a lack of power. Thus non-signicant
Nazeer A, Newcomer JW, Charney DS (2004). Eects
ndings, in particular, should be interpreted with of dexamethasone on declarative memory function
caution and our current ndings require replication. in posttraumatic stress disorder. Psychiatry Research 129,
110.
Buss C, Wolf OT, Witt J, Hellhammer DH (2004).
Summary Autobiographic memory impairment following acute
cortisol administration. Psychoneuroendocrinology 29,
The current results indicate that cortisol has opposing 10931096.
eects on memory retrieval in healthy controls com- Carvalho Fernando S, Beblo T, Schlosser N, Terfehr K,
pared to patients with BPD. These enhancing eects Otte C, Lowe B, Wolf OT, Spitzer C, Driessen M,
were found to be most pronounced for autobio- Wingenfeld K (2012). Associations of childhood trauma
graphical memory retrieval. On the contrary, BPD with hypothalamic-pituitary-adrenal function in
patients with co-morbid MDD but no PTSD did not borderline personality disorder and major depression.
show any cortisol eects on memory, which is in line Psychoneuroendocrinology. Published online : 21 March 2012.
with our former studies in MDD patients and the doi :10.1016/j.psyneuen.2012.02.012.
Champagne DL, Bagot RC, van Hasselt F, Ramakers G,
hypotheses of reduced GR function. Thus, it might
Meaney MJ, de Kloet ER, Joels M, Krugers H (2008).
be that there are at least two subgroups within the
Maternal care and hippocampal plasticity : evidence for
BPD population : one predominantly characterized by experience-dependent structural plasticity, altered
trauma-associated symptoms and an enhanced sensi- synaptic functioning, and dierential responsiveness
tivity to exogenous cortisol, and another subgroup to glucocorticoids and stress. Journal of Neuroscience 28,
with mood disturbances as core symptoms (i.e. co- 60376045.
morbid MDD), showing a resistance towards cortisol de Kloet ER (2003). Hormones, brain and stress. Endocrine
eects on memory. Of note, the BPD group with co- Regulations 37, 5168.
morbid MDD but without PTSD reported the lowest de Quervain DJ, Henke K, Aerni A, Treyer V, McGaugh JL,
CTQ scores, which ts this hypothesis. Possibly, Berthold T, Nitsch RM, Buck A, Roozendaal B, Hock C
(2003). Glucocorticoid-induced impairment of declarative
these subgroups may also dier concerning other
memory retrieval is associated with reduced blood ow in
HPA axis functions as we have hypothesized before
the medial temporal lobe. European Journal of Neuroscience
(Wingenfeld et al. 2010 b). Future studies should 17, 12961302.
investigate whether there are subgroups of borderline First MB, Spitzer RL, Gibbon M, Williams JBW (1997).
patients with dierent endocrine and psychopath- Structured Clinical Interview for DSM-IV Axis I Disorders
ological patterns, in addition to potentially related (SCID-I). Biometrics Research Department, New York State
therapeutic options. Psychiatric Institute : New York.
504 K. Wingenfeld et al.

Foa EB (1995). The Posttraumatic Stress Diagnostic Scale Oei NY, Elzinga BM, Wolf OT, de Ruiter MB,
Manual. National Computer Systems Inc. : Minneapolis, Damoiseaux JS, Kuijer JP, Veltman DJ, Scheltens P,
USA. Rombouts SA (2007). Glucocorticoids decrease
Golier JA, Yehuda R, Bierer LM, Mitropoulou V, New AS, hippocampal and prefrontal activation during declarative
Schmeidler J, Silverman JM, Siever LJ (2003). The memory retrieval in young men. Brain Imaging and Behavior
relationship of borderline personality disorder to 1, 3141.
posttraumatic stress disorder and traumatic events. Pariante CM, Miller AH (2001). Glucocorticoid receptors in
American Journal of Psychiatry 160, 20182024. major depression : relevance to pathophysiology and
Grossman R, Yehuda R, Golier J, McEwen B, Harvey P, treatment. Biological Psychiatry 49, 391404.
Maria NS (2006). Cognitive eects of intravenous Plotsky PM, Owens MJ, Nemero CB (1998).
hydrocortisone in subjects with PTSD and healthy control Psychoneuroendocrinology of depression. Hypothalamic-
subjects. Annals of the New York Academy of Sciences 1071, pituitary-adrenal axis. Psychiatric Clinics of North America
410421. 21, 293307.
Heim C, Newport DJ, Heit S, Graham YP, Wilcox M, Rinne T, de Kloet ER, Wouters L, Goekoop JG, DeRijk RH,
Bonsall R, Miller AH, Nemero CB (2000). Pituitary- van den Brink W (2002a). Hyperresponsiveness of
adrenal and autonomic responses to stress in women hypothalamic-pituitary-adrenal axis to combined
after sexual and physical abuse in childhood. Journal of dexamethasone/corticotropin-releasing hormone
the American Medical Association 284, 592597. challenge in female borderline personality disorder
Henckens MJ, Pu Z, Hermans EJ, van Wingen GA, Joels M, subjects with a history of sustained childhood abuse.
Fernandez G (2011). Dynamically changing eects of Biological Psychiatry 52, 11021112.
corticosteroids on human hippocampal and prefrontal Rinne T, van den Brink W, Wouters L, van Dyck R (2002 b).
processing. Human Brain Mapping. Published online : SSRI treatment of borderline personality disorder : a
21 September. doi :10.1002/hbm.21409. randomized, placebo-controlled clinical trial for female
Het S, Ramlow G, Wolf OT (2005). A meta-analytic review patients with borderline personality disorder. American
of the eects of acute cortisol administration on human Journal of Psychiatry 159, 20482054.
memory. Psychoneuroendocrinology 30, 771784. Rohleder N, Joksimovic L, Wolf JM, Kirschbaum C (2004).
Holsboer F (2000). The corticosteroid receptor hypothesis Hypocortisolism and increased glucocorticoid sensitivity
of depression. Neuropsychopharmacology 23, 477501. of pro-inammatory cytokine production in Bosnian war
Kuhlmann S, Kirschbaum C, Wolf OT (2005a). Eects refugees with posttraumatic stress disorder. Biological
of oral cortisol treatment in healthy young women on Psychiatry 55, 745751.
memory retrieval of negative and neutral words. Rohleder N, Wolf JM, Wolf OT (2010). Glucocorticoid
Neurobiology of Learning and Memory 83, 158162. sensitivity of cognitive and inammatory processes in
Kuhlmann S, Piel M, Wolf OT (2005b). Impaired memory depression and posttraumatic stress disorder. Neuroscience
retrieval after psychosocial stress in healthy young men. and Biobehavioral Reviews 35, 104114.
Journal of Neuroscience 25, 29772982. Schelling G, Roozendaal B, Krauseneck T, Schmoelz M,
Lange W, Wul H, Berea C, Beblo T, Saavedra AS, de Quervain DJ, Briegel J (2006). Ecacy of
Mensebach C, Wingenfeld K, Driessen M (2005). hydrocortisone in preventing posttraumatic stress disorder
Dexamethasone suppression test in borderline personality following critical illness and major surgery. Annals of the
disorder eects of posttraumatic stress disorder. New York Academy of Sciences 1071, 4653.
Psychoneuroendocrinology 30, 919923. Schlosser N, Wolf OT, Fernando SC, Riedesel K, Otte C,
Lieb K, Rexhausen JE, Kahl KG, Schweiger U, Muhtz C, Beblo T, Driessen M, Lowe B, Wingenfeld K
Philipsen A, Hellhammer DH, Bohus M (2004). (2010). Eects of acute cortisol administration on
Increased diurnal salivary cortisol in women with autobiographical memory in patients with major
borderline personality disorder. Journal of Psychiatric depression and healthy controls. Psychoneuroendocrinology
Research 38, 559565. 35, 316320.
Lovallo WR, Robinson JL, Glahn DC, Fox PT (2010). Acute Schmahl C, Bremner JD (2006). Neuroimaging in borderline
eects of hydrocortisone on the human brain : an fMRI personality disorder. Journal of Psychiatric Research 40,
study. Psychoneuroendocrinology 35, 1520. 419427.
Lupien SJ, Gillin CJ, Hauger RL (1999). Working memory is Schoofs D, Wolf OT (2009). Stress and memory retrieval in
more sensitive than declarative memory to the acute eects women : no strong impairing eect during the luteal phase.
of corticosteroids : a dose-response study in humans. Behavioral Neuroscience 123, 547554.
Behavioral Neuroscience 113, 420430. Schule C, Baghai TC, Eser D, Rupprecht R (2009).
Lupien SJ, Lepage M (2001). Stress, memory, and the Hypothalamic-pituitary-adrenocortical system
hippocampus : cant live with it, cant live without it. dysregulation and new treatment strategies in depression.
Behavioural Brain Research 127, 137158. Expert Review of Neurotherapeutics 9, 10051019.
McGowan PO, Sasaki A, DAlessio AC, Dymov S, Terfehr K, Wolf OT, Schlosser N, Fernando SC, Otte C,
Labonte B, Szyf M, Turecki G, Meaney MJ (2009). Muhtz C, Beblo T, Driessen M, Spitzer C, Lowe B,
Epigenetic regulation of the glucocorticoid receptor in Wingenfeld K (2011a). Eects of acute hydrocortisone
human brain associates with childhood abuse. Nature administration on declarative memory in patients with
Neuroscience 12, 342348. major depressive disorder : a placebo-controlled,
 Cortisol and memory in BPD 505

double-blind crossover study. Journal of Clinical Psychiatry Wolf OT (2009). Stress and memory in humans : twelve years
72, 16441650. of progress ? Brain Research 1293, 142154.
Terfehr K, Wolf OT, Schlosser N, Fernando SC, Otte C, Wolf OT, Convit A, McHugh PF, Kandil E, Thorn EL,
Muhtz C, Beblo T, Driessen M, Spitzer C, Lowe B, De Santi S, McEwen BS, de Leon MJ (2001). Cortisol
Wingenfeld K (2011b). Hydrocortisone impairs working dierentially aects memory in young and elderly men.
memory in healthy humans, but not in patients with major Behavioral Neuroscience 115, 10021011.
depressive disorder. Psychopharmacology 215, 7179. Yehuda R (2002). Current status of cortisol ndings in post-
Weerda R, Muehlhan M, Wolf OT, Thiel CM (2010). traumatic stress disorder. Psychiatric Clinics of North
Eects of acute psychosocial stress on working memory America 25, 3412368, vii.
related brain activity in men. Human Brain Mapping 31, Yehuda R (2009). Status of glucocorticoid alterations in post-
14181429. traumatic stress disorder. Annals of the New York Academy of
Wingenfeld K, Driessen M, Adam B, Hill A (2007a). Sciences 1179, 5669.
Overnight urinary cortisol release in women with Yehuda R, Golier JA, Bierer LM, Mikhno A,
borderline personality disorder depends on comorbid Pratchett LC, Burton CL, Makotkine I, Devanand DP,
PTSD and depressive psychopathology. European Pradhaban G, Harvey PD, Mann JJ (2010).
Psychiatry 22, 309312. Hydrocortisone responsiveness in Gulf War veterans
Wingenfeld K, Driessen M, Terfehr K, Schlosser N, with PTSD : eects on ACTH, declarative memory
Fernando SC, Otte C, Beblo T, Spitzer C, Lowe B, hippocampal [F]FDG uptake on PET. Psychiatry Research
Wolf OT (2012). Cortisol has enhancing, rather than 184, 117127.
impairing eects on memory retrieval in PTSD. Yehuda R, Golier JA, Yang RK, Tischler L (2004). Enhanced
Psychoneuroendocrinology 37, 10481056. sensitivity to glucocorticoids in peripheral mononuclear
Wingenfeld K, Hill A, Adam B, Driessen M (2007b). leukocytes in posttraumatic stress disorder. Biological
Dexamethasone suppression test in borderline personality Psychiatry 55, 11101116.
disorder : impact of PTSD symptoms. Psychiatry and Clinical Yehuda R, Harvey PD, Buchsbaum M, Tischler L,
Neurosciences 61, 681683. Schmeidler J (2007). Enhanced eects of cortisol
Wingenfeld K, Lange W, Wul H, Berea C, Beblo T, administration on episodic and working memory in
Saavedra AS, Mensebach C, Driessen M (2007c). Stability aging veterans with PTSD. Neuropsychopharmacology 32,
of the dexamethasone suppression test in borderline 25812591.
personality disorder with and without comorbid PTSD : a Young K, Drevets WC, Schulkin J, Erickson K (2011).
one-year follow-up study. Journal of Clinical Psychology 63, Dose-dependent eects of hydrocortisone infusion on
843850. autobiographical memory recall. Behavioral Neuroscience
Wingenfeld K, Spitzer C, Mensebach C, Grabe HJ, Hill A, 125, 735741.
Gast U, Driessen M (2010a). The German version of the Zanarini MC, Frankenburg FR, Dubo ED, Sickel AE,
Childhood Trauma Questionnaire (CTQ) : preliminary Trikha A, Levin A, Reynolds V (1998). Axis I comorbidity
psychometric properties [in German]. Psychotherapie, of borderline personality disorder. American Journal of
Psychosomatik, Medizinische Psychologie 60, 442450. Psychiatry 155, 17331739.
Wingenfeld K, Spitzer C, Rullkotter N, Lowe B (2010b). Zanarini MC, Williams AA, Lewis RE, Reich RB,
Borderline personality disorder : hypothalamus pituitary Vera SC, Marino MF, Levin A, Yong L, Frankenburg FR
adrenal axis and ndings from neuroimaging studies. (1997). Reported pathological childhood experiences
Psychoneuroendocrinology 35, 154170. associated with the development of borderline
Wolf OT (2003). HPA axis and memory. Best Practice and personality disorder. American Journal of Psychiatry 154,
Research. Clinical Endocrinology and Metabolism 17, 287299. 11011106.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.