- Email: [email protected]
Increased diurnal salivary cortisol in women with borderline personality disorder
JOURNAL OF PSYCHIATRIC RESEARCH
Journal of Psychiatric Research 38 (2004) 559–565
www.elsevier.com/locate/jpsychires
Increased diurnal salivary cortisol in women with borderline personality disorder Klaus Lieba,*, Jost E. Rexhausena,b,1, Kai G. Kahlc, Ulrich Schweigerc, Alexandra Philipsena, Dirk H. Hellhammerb, Martin Bohusa,2 a
Department of Psychiatry and Psychotherapy, University of Freiburg Medical School, Hauptstr. 5, D-79104 Freiburg, Germany b Department of Psychology, Fachbereich I, University of Trier, Universit€atsring 15, Building D, 54286 Trier, Germany c Department of Psychiatry and Psychotherapy, University of L€ubeck, Ratzeburger Allee 160, 23538 L€ubeck, Germany Received 5 January 2004; received in revised form 25 March 2004; accepted 7 April 2004
Abstract Borderline personality disorder (BPD) is characterized by a pervasive pattern of instability in affect regulation, impulse control, interpersonal relationships, and self-image. In previous studies, we have used portable mini-computers to assess the severity of recurrent states of aversive emotional distress and dissociation during ambulatory conditions. Here, we used this approach for the assessment of the hypothalamic–pituitary–adrenal (HPA) axis in patients with BPD. We studied 23 unmedicated female patients with BPD and 24 matched healthy controls. Salivary cortisol was collected from all participants during ambulatory conditions in response to reminders provided by portable mini-computers on 3 consecutive days every 2 h for 14 h after awakening. In addition, cortisol in response to awakening was determined in four 15 min intervals on days 1 and 2. After the last collection of cortisol on the second day, 0.5 mg dexamethasone was administered in order to achieve cortisol suppression on day 3 (low-dose dexamethasone suppression test, DST). Patients with BPD displayed significantly higher salivary cortisol levels than healthy controls as demonstrated by higher total cortisol in response to awakening and higher total daily cortisol levels. There were significantly more nonsuppressors of cortisol in the low-dose DST in the patient group when compared to the control group. The ambulatory assessment of saliva cortisol is a suitable approach to study basic parameters of the HPA-axis in patients with BPD. Increased adrenal activity and lowered feedback sensitivity of the HPA-axis may characterise BPD. Further studies have to reveal reasons of heightened adrenal activity in these patients. 2004 Elsevier Ltd. All rights reserved. Keywords: Cortisol; Saliva; Dexamethasone suppression test; Borderline personality disorder; Post-traumatic stress disorder; Depression; Childhood trauma
1. Introduction Borderline personality disorder (BPD) is characterized by a pervasive pattern of instability in affect regulation, impulse control, interpersonal relationships, and self-image. Clinical hallmarks of the disorder are emo-
*
Corresponding author. Tel.: +49-761-270-6681; fax: +49-761-2706667. E-mail address: [email protected] (K. Lieb). 1 J.E.R. died during a car accident in 2003. The results published are part of his doctoral thesis. The manuscript is dedicated to him. 2 Present address: Central Institute of Mental Health, Mannheim, University of Heidelberg, J 5, 68159 Mannheim, Germany. 0022-3956/$ - see front matter 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.jpsychires.2004.04.002
tional dysregulation, impulsive aggression, repeated selfinjurious behavior and chronic suicidality. In previous investigations, we have carefully assessed key symptoms of the disorder which are most disabling for the patients such as recurrent states of severe and aversive emotional distress (tension) and dissociative experiences and were especially interested in the investigation of diurnal fluctuations of these symptoms. We assessed these fluctuations by the help of portable mini-computers which were given to the patients for three consecutive days and reminded them several times during the day to judge their current degree of aversive tension or dissociation on a Lickert scale between 0 (‘‘no tension/dissociation at all’’) and 9 (‘‘most severe tension/dissociation’’). Our
560
K. Lieb et al. / Journal of Psychiatric Research 38 (2004) 559–565
investigations not only demonstrated that female patients with BPD show significantly higher baseline levels of tension and dissociation, but also suffer from recurrent severe and aversive states of tension and dissociation and that tension and dissociative experiences are highly correlated (Stiglmayr et al., 2001). In the present study, we used this ‘‘naturalistic’’ approach for the first time to assess neurobiological measures in patients with BPD during ambulatory conditions. As a well-established neurobiological marker which has been extensively investigated in the context of emotional stress and psychiatric disorders, we assessed cortisol levels and used the dexamethasone suppression test (DST) to assess feedback sensitivity of the hypothalamic–pituitary–adrenal (HPA) axis (for review, see Holsboer, 2001). The present study had the following aims: (1) to investigate whether this ambulatory assessment is a suitable and reliable study tool to assess HPA-axis functioning in patients with BPD and (2) to assess HPA-axis functioning in patients with BPD as compared to healthy controls. In addition to the application of a completely new method to gather neurobiological data in BPD patients, we tried to overcome some of the problems of previous studies which investigated HPA-axis funtioning in patients with BPD (Baxter et al., 1984; Carroll et al., 1981; Kontaxakis et al., 1998; Krishnan et al., 1984; Soloff et al., 1982; Steiner et al., 1984; Sternbach et al., 1983; for review, see De la Fuente and Mendlewicz, 1996; Koenigsberg et al., 1999; Lahmeyer et al., 1989): (1) we excluded patients with a current major depressive disorder (MDD); (2) we controlled potential confounding variables, which may influence HPA-axis functioning such as smoking, menstrual cycle and physical activity; (3) we measured cortisol in saliva, thus preventing artefacts by stress-induced increases in HPA-axis activity due to invasive blood collection; (4) we carefully synchronised all sampling with the individual pattern of awakening of every participant; and (5), most importantly, we assessed HPA-axis functioning not only by determining cortisol levels at a single time point, but in a detailed analysis of cortisol levels on three consecutive days (for overview, see Kirschbaum and Hellhammer, 1994; Kirschbaum and Hellhammer, 1998).
2. Subjects and methods 2.1. Subjects The study was conducted at the Department of Psychiatry and Psychotherapy of the University of Freiburg Medical School and at the Department of Psychiatry and Psychotherapy, University of Luebeck Medical School (four patients). We included 23 caucasian female patients aged between 19 and 47 years and 24 female
healthy controls. All patients met the DSM-IV criteria for BPD, assessed by the appropriate segment of the Structured Clinical Interview for DSM-IV Personality Disorders (SCID-II) (First et al., 1996) and the Diagnostic Interview for Borderline Personality Disorder – Revised Version (DIB-R) (Zanarini et al., 1989). Cutoff scores were 5 or more out of 9 for the DSM-IV-criteria and 8 or more out of 10 for the DIB-R. Exclusion criteria were current comorbid MDD, current substance abuse, mental retardation or schizoaffective disorder as well as a life time history of schizophrenia or bipolar I disorder. Control subjects (healthy controls, HCs) were healthy women who did not fulfil criteria for BPD (SCID-II), were free of any former or present psychiatric disorder and had never sought psychiatric or psychotherapeutic help. Controls were students, clinic staff (from another department) or acquaintances of the experimenters. The HC were pooled and selectively included only if they were found similar to a patient according to the following requirements: smoking, age, and years of education. As shown in Table 1, there were no significant differences between the groups in any of these variables. However, patients with BPD were less frequently married and more often lived alone than the healthy controls, reflecting diminished stability of social relationships as typical of BPD. All participants were free of any medication at least for two months prior to the study and did not use oral contraceptives. Patients and healthy subjects gave written informed consent and received reimbursement for their efforts (a100 plus travelling fees). The study was carried out in accordance with the latest version of the Declaration of Helsinki and was approved by the local ethical board of the University of Freiburg Medical School. 2.2. Procedure To control for a possible modulation of cortisol by changes of estrogen or other hormones during the menstrual cycle, only subjects with regular menstrual cycles were included and all women, patients and con-
Table 1 Sample characteristicsa Characteristic
BPD (n ¼ 23)
HC (n ¼ 24)
Statisticsb
Age Gender (f/m) Smoker Weight (kg) Current partnership Own children
28.5 ± 1.6 23/0 17 (74) 63.6 ± 3.0 7 (30) 5 (22)
28.2 ± 1.4 24/0 18 (75) 65.8 ± 2.6 14 (58) 5 (21)
n.s. n.s. n.s. n.s. p < 0:05 n.s.
a
Data are given as means ± SD or number (percentage). BPD, borderline personality disorder; HC, healthy controls. b Mann–Whitney U test or v2 test.
K. Lieb et al. / Journal of Psychiatric Research 38 (2004) 559–565
trols, started at the fourth day of the menstrual cycle. The time period for the determination of cortisol on three consecutive days from the fourth to sixth day of the menstrual cycle was chosen because during these days of the follicular phase estrogen-levels are constantly low. A total of 32 saliva samples were obtained from each subject: 12 cortisol samples each on days 1 and 2, and eight samples on day 3 after taking 0.5 mg dexamethasone after the last collection of saliva on day 2 (DST) (for time points of cortisol determination, see Fig. 1). Morning measurements included five determinations in 15 min intervals beginning with awakening (‘‘cortisol immediately after awakening’’ and ‘‘total cortisol response to awakening’’) as well as seven determinations in 2 h intervals over the course of the day (‘‘total daily cortisol’’). To enable accurate timing, subjects were reminded by the portable mini-computer (PSION 3a, programmed by W. M€ uller, Department of Psychophysiology, University of Freiburg) to collect saliva samples by chewing the salivette. To control for confounding variables, at every single sampling point, participants were asked for additional information by the mini-computer: the degree of physical activity (as estimated on a 0–9 point Lickert scale) and smoking before sample collection (yes/no and number of cigarettes). They had to press specific keys on the minicomputer to answer the questions. To avoid dilution or spoiling of samples, participants were advised not to drink or eat for at least 5 min before sampling, the minicomputers being programmed with a ready to use control-function. 2.3. Determination of saliva cortisol levels Saliva was collected in Salivette devices (Sarstedt, Rommelsdorf, Germany). We decided to measure cor-
Awakening +15 +30 +45 +60 min. Awakening response
+2h
+4h
+6h
+8h
+10h +12h +14h
+2h
+4h
+6h
+8h
0.5 mg +10h +12h +14h Dex
+2h
+4h
+6h
+8h
+10h +12h +14h
Day 1
Day 2 Day 3
Fig. 1. Time points of cortisol determination. Participants were reminded by mini-computers to give saliva samples. The measurements started with awakening, followed by saliva sampling every 15 min for 1 h to assess ‘‘total cortisol response to awakening’’. Afterwards, sampling took place every second hour for determination of ‘‘total daily cortisol’’. Note that participants took 0.5 mg dexamethasone (DEX) on day 2 after the last sampling, and that no morning peak was sampled on day 3.
561
tisol in saliva-samples for three reasons: (1) the non-invasive sampling is suitable for an ambulatory setting; (2) it prevents artefacts by stress-induced increases in HPAaxis activity due to invasive blood collection; and (3) the free, biologically active fraction of cortisol (5–10% of the total fraction) is measured selectively. Salivary cortisol reliably reflects plasma-free cortisol concentrations with a correlation of measures of r P 0:90 (Kirschbaum and Hellhammer, 1994). Saliva samples were stored in a refrigerator until the the morning after the third day and were stored at )40 C until the time of assay. Saliva cortisol concentrations were determined using a time-resolved fluorescence immunoassay (DELFIA) as described by Dressendorfer et al. (1992) in the Biochemistry Laboratory of the Department of Psychobiology of Trier University (Dressendorfer et al., 1992). The intra-assay coefficient of variation range from 4.0% to 6.7% and the corresponding inter-assay coefficients of variation from 7.1% to 9.0%. Dexamethasone levels were not determined. 2.4. Data analysis Data are given as means ± SD. Areas under the curve (AUC) were calculated for the ‘‘total cortisol response to awakening’’ and the ‘‘total daily cortisol’’ by the formula ‘‘cort1 + cort2 + cort3 + cort4 + ((cort5 ) cort1)/ 2)’’ and ‘‘cort6 + cort7 + cort8 + cort9 + cort10 + cort11 + ((cort12 ) cort6)/2)’’, respectively. Data were statistically analysed by using paired t test, two-way analysis of variance and repeated measurements model analysis taking into account the correlations between samples from one patient (estimates were obtained via the GEE method by Liang and Zeger (1986)). Smoking and physical activity as well as body weight in case of DST were covariates in the ANCOVA analysis. According to Heim et al. (1998), the suppressor status in the low-dose DST was defined by an AUC-value for the ‘‘total daily cortisol’’ 6 5 nmol/l, the non-suppressor status by an AUC-value >5 nmol/l, and the super-suppressor status by an AUC-value <2 nmol/l. The number of patients at the three sampling days varied because cortisol samples were not always obtained for each subject at every sampling point. A p < 0:05 was defined as a significant difference and a p < 0:1 as a trend.
3. Results 3.1. Basal adrenocortical activity and cortisol response to awakening Salivary cortisol was determined on days 1 and 2 in intervals of 15 min after awakening and seven times in 2 h intervals after awakening during the whole day. As shown in Fig. 2(a) and (b), cortisol levels were higher in
562
K. Lieb et al. / Journal of Psychiatric Research 38 (2004) 559–565
(a)
50
HC
Day 1
BPD
Salivary Cortisol (nmol/l)
40
30
20
10
0:00
22:00
20:00
18:00
16:00
14:00
12:00
10:00
8:00
6:00
0
(b) 50
HC
Day 2
BPD
Salivary Cortisol (nmol/l)
40
30
20
10
3.2. Feedback sensitivity of the pituitary–adrenal axis
0:00
22:00
20:00
18:00
16:00
14:00
12:00
10:00
8:00
6:00
0
(c)
50
HC
Day 3 (post DEX)
BPD
40
Salivary Cortisol (nmol/l)
(‘‘cortisol immediately after awakening’’, Table 2), which was probably due to the high variability of cortisol levels especially in the patients with BPD. However, calculated AUCs for the ‘‘total cortisol response to awakening’’ and the ‘‘total daily cortisol’’ were significantly higher in patients with BPD than in controls (Table 2). Patients with BPD displayed a significantly higher cortisol response to awakening on day 2 and on days 1 and 2 in combination, whereas the cortisol response to awakening on day 1 showed a trend for higher levels in the patient group. Total daily cortisol was significantly higher on day 1. In an additional repeated measurements model analysis which takes into account the correlation between samples from one patient and which includes time as a covariate, we found identical results as obtained by comparison of AUC values. For example, comparison of diurnal cortisol secretion on day 1 revealed a significant group effect with an estimate for group differences of 3.03 nmol/l (95% CI [0.11–6.0], p < 0:05) and a significant time effect with an estimate for time effects of )0.9 nmol/l/2 h (95% CI [from )1.17 to )0.63], p < 0:0001). Since results from repeated measurements analysis were identical to the results obtained from AUC analysis, only results from AUC analyses are given in Table 2.
30
20
10
0:00
22:00
20:00
18:00
16:00
14:00
12:00
10:00
8:00
6:00
0
Fig. 2. Cortisol levels (means ± SD) in patients with BPD and healthy controls in the course of day 1 (a), day 2 (b), and day 3 after dexamethasone (c). Cortisol was sampled at the indicated time points by use of Salivette devices.
patients with BPD than in controls. These differences were not statistically significant at the single time point including the ‘‘awakening cortisol’’ on days 1 and 2
The low-dose DST was performed by applying 0.5 mg dexamethasone at 22:30 on day 2. This low dose of dexamethasone was used on the basis of our assumption that patients with BPD would show low cortisol levels and supersuppression to dexamethasone similar to patients with PTSD (Yehuda et al., 1993). As shown in Fig. 3, the oral ingestion of 0.5 mg dexamethasone induced a significant suppression of cortisol levels in both BPD patients and healthy controls (F ð1; 33Þ ¼ 71; p < 0:001). By using the criteria of Heim et al. (Heim et al., 1998) (see Section 2), significantly more patients with BPD were non-suppressors in the low-dose DST as compared to the healthy controls (14 vs. 4, v2 , p < 0:01). Only 1 patient and 4 controls were supersuppressors in the DST. ANCOVA did not reveal a significant dexamethasone · group interaction (F ð1; 33Þ ¼ 0:07; p ¼ 0:8). Baseline differences did not significantly influence suppression in an additional model taking baseline differences in cortisol levels as covariate. Although there was no dexamethasone · group interaction, total daily cortisol on day 3 was significantly higher in the patients with BPD compared to the healthy controls (Table 2) whereas cortisol immediately after awakening showed a trend for higher levels in patients with BPD. Cortisol levels after dexamethasone on day 3 were higher in the BPD patients at all time points (p < 0:05) (see Fig. 2(c)).
K. Lieb et al. / Journal of Psychiatric Research 38 (2004) 559–565
563
Table 2 Salivary cortisol levels (nmol/l) in patients with BPD and healthy controls (HCs) BPD (n ¼ 23)
Time
HC (n ¼ 24)
T
df
p
20 ± 13 28 ± 29 12 ± 16
18 ± 6 18 ± 9 5±6
0.91 1.48 2.00
45 40 39
0.37 0.15 0.05
113 ± 40 116 ± 47 115 ± 39
92 ± 35 91 ± 32 91 ± 29
1.92 2.06 2.36
45 44 44
0.06 0.04 0.02
Day 1 Day 2
78 ± 39 82 ± 63
58 ± 22 59 ± 18
2.13 1.54
44 40
0.04 0.14
Day 3
54 ± 49
25 ± 17
2.51
39
0.02
‘‘Cortisol immediately after awakening’’ (day 3 after DEX)
Day 1 Day 2 Day 3
‘‘Total cortisol response to awakening’’ (AUC)
Day 1 Day 2 Days 1 and 2
‘‘Total daily cortisol’’ (AUC) ‘‘Total daily cortisol after DEX’’ (AUC)
Data are given as means ± SD. BPD, borderline personality disorder; AUC, area under the curve; DEX, dexamethasone.
AUC Salivary Cortisol (nmol/l)
160
AUC Day 2
140
**
AUC Day 3
120 100
grating smoking as a covariate in the ANCOVA test, however, did not reveal a significant influence of smoking on cortisol secretion. Physical activity as a covariate in the ANCOVA had also no significant effect on cortisol levels.
**
80
4. Discussion
60 40 20 0 BPD
HC
Fig. 3. Effect of 0.5 mg dexamethasone on cortisol levels in 23 patients with BPD and 24 healthy controls. Values are expressed as means ± SD for AUCs of ‘‘total daily cortisol’’ on days 2 and 3. ** indicates significant suppression of cortisol secretion in both patients and controls. Differences between patients and controls are given in Table 2.
3.3. Influence of possible confounding variables To control for confounding variables which might possibly account for the differences in cortisol secretion between patients and controls, we assessed physical activity and smoking preceding the cortisol samplings. In a first analysis, we compared cortisol levels between smokers and non-smokers and found no significant differences in cortisol secretion. In detail, there was no difference between both groups with respect to the adrenocortical reactivity following awakening on day 1 (tð29; 742Þ ¼ 0:848, p ¼ 0:403) or on day 2 (tð30; 599Þ ¼ 0:012, p ¼ 0:991). Furthermore, no significant difference could be revealed with respect to the basal adrenocortical activity on day 1 (tð34Þ ¼ 0:087, p 6 0:9) and on day 2 (tð33Þ ¼ 1:1, p 6 0:3). At 5 time points (10.00 and 14.00 h on day 1; 14.00, 20.00 and 22.00 h on day 2), patients smoked significantly more cigarettes in the 2 h prior to saliva sampling, although the mean of cigarettes smoked did not exceed 2. Inte-
In this study, we followed a new approach in the study of patients with BPD, i.e., the measurement of diurnal cortisol secretion during natural, non-laboratory conditions. This approach was well accepted by our patients and appears to be a suitable approach to assess basic functions of HPA-axis activity in this patient group. Moreover, the approach revealed reliable findings: Measurements on two consecutive days (days 1 and 2) showed consistently increases in cortisol secretion with respect to total cortisol response to awakening and total daily cortisol although increases in the BPD patients not always reached statistical significance (e.g., the total cortisol response to awakening was only significantly different on day 2). This was probably due to the high variability of cortisol secretion in the patients as demonstrated by high standard deviations or by insufficient power to detect differences due to small sample sizes. Further research in a larger patient group has to elucidate whether there are subgroups which can be distinguished with respect to their cortisol secretion pattern. Although repetitive ambulatory measurements have the advantage that they increase the reliability of findings as compared to assessments of cortisol levels at single time points and that they reflect real life of the participants, they have obvious disadvantages, such as the difficulty to control for the compliance of the participants as well as the difficulty to control the variables which might confound cortisol levels. We do not believe that compliance problems do account for the differences seen between patients and controls. Confounding variables such as physical activity and smoking were
564
K. Lieb et al. / Journal of Psychiatric Research 38 (2004) 559–565
controlled with use of the mini-computers and we found no relevant influence. However, other possible confounding factors such as disturbances of the day-nightrhythm were not controlled, although shift workers were excluded from the study. A further strength of the study was the rigid control for a possible confounding influence of female sexual hormones. Only women with a normal menstrual cycle were included, use of contraceptives was not permitted and all women started cortisol measurements at the same day of the menstrual cycle. From a methodological point of view, diurnal cortisol profiles were – as the immediate response to awakening – gathered in relation to the individual awakening time and not at fixed times of the day. Although patients with BPD tended to wake up earlier that normal controls (7:43 ± 1:42 vs. 8:25 ± 1:53 h), this difference was not statistically significant. Since Kudielka and Kirschbaum (2003) had shown that a group of healthy early awakeners (mean awakening time 6:49 h) showed a higher cortisol response to awakening than healthy late awakeners (mean awakening time 9:43 h), it cannot fully be excluded that different awakening times had a possible confounding influence in our study. However, the difference in awakening times was 3 h in contrast to only 40 min in our study. Furthermore, additional analyses of our sample showed that there were no significant differences in cortisol levels between early and late awakeners in both controls and patients with BPD. Early awakeners even tended to have lower cortisol levels than late awakeners which confirms our conclusion that the observed differences in cortisol levels between patients and controls are not due to different awakening times. Several previous studies have investigated the activity of the HPA-axis in patients with BPD (Baxter et al., 1984; Carroll et al., 1981; Kontaxakis et al., 1998; Krishnan et al., 1984; Soloff et al., 1982; Steiner et al., 1984; Sternbach et al., 1983; for review see De la Fuente and Mendlewicz, 1996; Koenigsberg et al., 1999; Lahmeyer et al., 1989). Grossman et al. (2003) published a study on 52 patients with personality disorders (21 BPD patients) using a 0.5 mg DST. They found that subjects with personality disorders and comorbid PTSD had a significantly greater degree of cortisol suppression in response to a 0.5 mg DST than subjects without comorbid PTSD. After controlling for the effect of BPD diagnosis on cortisol suppression within patients with PTSD, there were no significant effects; however, it is possible that this finding was related to low statistical power. The authors did not report how many of the 16 subjects with comorbid PTSD had BPD, nor did they report separate data on the 21 BPD patients. A shortcoming of our study was that we did not assess comorbid PTSD diagnoses at the time of the study. A possible confounding influence of current PTSD diag-
noses can therefore not be excluded although exclusion of patients with a current PTSD would probably have led to an even stronger effect since PTSD has been associated with lower cortisol levels (Yehuda, 2002). Our finding of increased cortisol secretion and lowered feedback sensitivity of the HPA-axis as demonstrated by higher cortisol levels after 0.5 mg dexamethasone and a higher number of non-suppressors to dexamethasone is in line with previous studies showing non-suppression rates to 1 mg DST in BPD patients ranging from 9.5% (Lahmeyer et al., 1988) to 62% (Carroll et al., 1981). The reasons for our findings of increased cortisol secretion and lowered feedback sensitivity in patients with BPD are unknown. Obviously, they are not due to an overlapping affective disorder, since our patients were not currently depressed. However, increased cortisol levels in our patients might result from current life-time stress. As we have shown in previous studies, patients with BPD suffer from recurrent states of severe and aversive inner tension and dissociation (Stiglmayr et al., 2001) which might correlate with increased secretion of cortisol. Further studies assessing the psychopathological correlates of increased cortisol secretion need to be performed to investigate whether states of severe inner tension determine increased cortisol secretion in our patients. In conclusion, our study shows that the repetitive assessment of cortisol levels in saliva during non-laboratory, ambulatory conditions is a suitable method to get informations of HPA-axis functioning in patients with BPD. Furthermore, the study reliably demonstrated that patients with BPD show increased cortisol secretion and lowered feedback sensitivity to low-dose dexamethasone. Further studies have to reveal reasons of heightened adrenal activity in these patients.
Acknowledgements We thank Prof. Dr. Fahrenberg, Forschungsgruppe Psychophysiologie, University of Freiburg, Germany for help in the programming of the mini-computers, C. Unckel for patient care, I. Rummel-Fr€ uhauf and A. Fritzen, University of Trier, Germany for running the cortisol-assays, and G. Ihorst for help in statistical analyses. This study was supported by the Ministry of Baden W€ urttemberg and the Borderline Personality Disorder Research Foundation (BPDRF).
References Baxter L, Edell W, Gerner R, Fairbanks L, Gwirtsman H. Dexamethasone suppression test and Axis I diagnoses of inpatients with DSM-III borderline personality disorder. Journal of Clinical Psychiatry 1984;45:150–3.
K. Lieb et al. / Journal of Psychiatric Research 38 (2004) 559–565 Carroll BJ, Greden JF, Feinberg M, Lohr N, James NM, Steiner M, Haskett RF, Albala AA, DeVigne JP, Tarika J. Neuroendocrine evaluation of depression in borderline patients. Psychiatric Clinics of North America 1981;4:89–99. De la Fuente JM, Mendlewicz J. TRH stimulation and dexamethasone suppression in borderline personality disorder. Biological Psychiatry 1996;40:412–8. Dressendorfer RA, Kirschbaum C, Rohde W, Stahl F, Strasburger CJ. Synthesis of a cortisol–biotin conjugate and evaluation as a tracer in an immunoassay for salivary cortisol measurement. Journal of Steroid Biochemistry and Molecular Biology 1992;43:683–92. First MB, Spitzer RL, Gibbon M, Williams JBW. User’s guide for the Structured Clinical Interview for DSM-IV personality disorders (SCID-II). Washington, DC: American Psychiatric Press; 1996. Grossman R, Yehuda R, New A, Schmeidler J, Silverman J, Mitropoulou V, Sta Maria N, Golier J, Siever L. Dexamethasone suppression test findings in subjects with personality disorders: associations with posttraumatic stress disorder and major depression. American Journal of Psychiatry 2003;160:1291–8. Heim C, Ehlert U, Hanker JP, Hellhammer DH. Abuse-related posttraumatic stress disorder and alterations of the hypothalamic– pituitary–adrenal axis in women with chronic pelvic pain. Psychosomatic Medicine 1998;60:309–18. Holsboer F. Stress, hypercortisolism and corticosteroid receptors in depression: implications for therapy. Journal of Affective Disorders 2001;62:77–91. Kirschbaum C, Hellhammer DH. Salivary cortisol in psychoneuroendocrine research: recent developments and applications. Psychoneuroendocrinology 1994;19:313–33. Kirschbaum C, Hellhammer DH. Hypothalamus-hypohysen-nebennierenrinden-achse. In: Kirschbaum C, Hellhammer DH, editors. Enzyklop€ adie der Psychologie, Band Psychoendokrinologie und Psychoimmunologi. 1998. p. 79–140. Koenigsberg HW, Anwunah I, New AS, Mitropoulou V, Schopick F, Siever LJ. Relationship between depression and borderline personality disorder. Depression and Anxiety 1999;10:158–67. Kontaxakis V, Markianos M, Vaslamatzis G, Markidis M, Kanellos P, Stefanis C. Multiple neuroendocrinological responses in borderline personality disorder patients. Acta Psychiatrica Scandinavica 1998;76:593–7.
565
Krishnan KR, Davidson JR, Rayasam K, Shope F. The dexamethasone suppression test in borderline personality disorder. Biological Psychiatry 1984;19:1149–53. Kudielka BM, Kirschbaum C. Awakening cortisol responses are influenced by health status and awakening time but not by menstrual cycle phase. Psychoneuroendocrinology 2003;28:35–47. Lahmeyer HW, Reynolds III CF, Kupfer DJ, King R. Biologic markers in borderline personality disorder: a review. Journal of Clinical Psychiatry 1989;50:217–25. Lahmeyer HW, Val E, Gaviria FM, Prasad RB, Pandey GN, Rodgers P, Weiler MA, Altman EG. EEG sleep, lithium transport, dexamethasone suppression, and monoamine oxidase activity in borderline personality disorder. Psychiatry Research 1988;25:19– 30. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika 1986;73:13–22. Soloff PH, George A, Nathan RS. The dexamethasone suppression test in patients with borderline personality disorders. American Journal of Psychiatry 1982;139:1621–3. Steiner M, Martin S, Wallace JE, Goldman S. Distinguishing subtypes within the borderline domain: a combined psychoneuroendocrine approach. Biological Psychiatry 1984;19:907–11. Sternbach HA, Fleming J, Extein I, Pottash AL, Gold MS. The dexamethasone suppression and thyrotropin-releasing hormone tests in depressed borderline patients. Psychoneuroendocrinology 1983;8:459–62. Stiglmayr CE, Shapiro DA, Stieglitz RD, Limberger MF, Bohus M. Experience of aversive tension and dissociation in female patients with borderline personality disorder – a controlled study. Journal of Psychiatric Research 2001;35:111–8. Yehuda R. Current status of cortisol findings in post-traumatic stress disorder. Psychiatric Clinics of North America 2002;25:341– 68. Yehuda R, Boisoneau D, Mason JW, Giller EL. Glucocorticoid receptor number and cortisol excretion in mood, anxiety, and psychotic disorders. Biological Psychiatry 1993;34:18–25. Zanarini MC, Gunderson JG, Frankenburg FR, Chauncey DL. The revised Diagnostic Interview for Borderlines: discriminating BPD from other Axis II disorders. Journal of Personality Disorders 1989;3:10–8.
Journal of Psychiatric Research 38 (2004) 559–565
www.elsevier.com/locate/jpsychires
Increased diurnal salivary cortisol in women with borderline personality disorder Klaus Lieba,*, Jost E. Rexhausena,b,1, Kai G. Kahlc, Ulrich Schweigerc, Alexandra Philipsena, Dirk H. Hellhammerb, Martin Bohusa,2 a
Department of Psychiatry and Psychotherapy, University of Freiburg Medical School, Hauptstr. 5, D-79104 Freiburg, Germany b Department of Psychology, Fachbereich I, University of Trier, Universit€atsring 15, Building D, 54286 Trier, Germany c Department of Psychiatry and Psychotherapy, University of L€ubeck, Ratzeburger Allee 160, 23538 L€ubeck, Germany Received 5 January 2004; received in revised form 25 March 2004; accepted 7 April 2004
Abstract Borderline personality disorder (BPD) is characterized by a pervasive pattern of instability in affect regulation, impulse control, interpersonal relationships, and self-image. In previous studies, we have used portable mini-computers to assess the severity of recurrent states of aversive emotional distress and dissociation during ambulatory conditions. Here, we used this approach for the assessment of the hypothalamic–pituitary–adrenal (HPA) axis in patients with BPD. We studied 23 unmedicated female patients with BPD and 24 matched healthy controls. Salivary cortisol was collected from all participants during ambulatory conditions in response to reminders provided by portable mini-computers on 3 consecutive days every 2 h for 14 h after awakening. In addition, cortisol in response to awakening was determined in four 15 min intervals on days 1 and 2. After the last collection of cortisol on the second day, 0.5 mg dexamethasone was administered in order to achieve cortisol suppression on day 3 (low-dose dexamethasone suppression test, DST). Patients with BPD displayed significantly higher salivary cortisol levels than healthy controls as demonstrated by higher total cortisol in response to awakening and higher total daily cortisol levels. There were significantly more nonsuppressors of cortisol in the low-dose DST in the patient group when compared to the control group. The ambulatory assessment of saliva cortisol is a suitable approach to study basic parameters of the HPA-axis in patients with BPD. Increased adrenal activity and lowered feedback sensitivity of the HPA-axis may characterise BPD. Further studies have to reveal reasons of heightened adrenal activity in these patients. 2004 Elsevier Ltd. All rights reserved. Keywords: Cortisol; Saliva; Dexamethasone suppression test; Borderline personality disorder; Post-traumatic stress disorder; Depression; Childhood trauma
1. Introduction Borderline personality disorder (BPD) is characterized by a pervasive pattern of instability in affect regulation, impulse control, interpersonal relationships, and self-image. Clinical hallmarks of the disorder are emo-
*
Corresponding author. Tel.: +49-761-270-6681; fax: +49-761-2706667. E-mail address: [email protected] (K. Lieb). 1 J.E.R. died during a car accident in 2003. The results published are part of his doctoral thesis. The manuscript is dedicated to him. 2 Present address: Central Institute of Mental Health, Mannheim, University of Heidelberg, J 5, 68159 Mannheim, Germany. 0022-3956/$ - see front matter 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.jpsychires.2004.04.002
tional dysregulation, impulsive aggression, repeated selfinjurious behavior and chronic suicidality. In previous investigations, we have carefully assessed key symptoms of the disorder which are most disabling for the patients such as recurrent states of severe and aversive emotional distress (tension) and dissociative experiences and were especially interested in the investigation of diurnal fluctuations of these symptoms. We assessed these fluctuations by the help of portable mini-computers which were given to the patients for three consecutive days and reminded them several times during the day to judge their current degree of aversive tension or dissociation on a Lickert scale between 0 (‘‘no tension/dissociation at all’’) and 9 (‘‘most severe tension/dissociation’’). Our
560
K. Lieb et al. / Journal of Psychiatric Research 38 (2004) 559–565
investigations not only demonstrated that female patients with BPD show significantly higher baseline levels of tension and dissociation, but also suffer from recurrent severe and aversive states of tension and dissociation and that tension and dissociative experiences are highly correlated (Stiglmayr et al., 2001). In the present study, we used this ‘‘naturalistic’’ approach for the first time to assess neurobiological measures in patients with BPD during ambulatory conditions. As a well-established neurobiological marker which has been extensively investigated in the context of emotional stress and psychiatric disorders, we assessed cortisol levels and used the dexamethasone suppression test (DST) to assess feedback sensitivity of the hypothalamic–pituitary–adrenal (HPA) axis (for review, see Holsboer, 2001). The present study had the following aims: (1) to investigate whether this ambulatory assessment is a suitable and reliable study tool to assess HPA-axis functioning in patients with BPD and (2) to assess HPA-axis functioning in patients with BPD as compared to healthy controls. In addition to the application of a completely new method to gather neurobiological data in BPD patients, we tried to overcome some of the problems of previous studies which investigated HPA-axis funtioning in patients with BPD (Baxter et al., 1984; Carroll et al., 1981; Kontaxakis et al., 1998; Krishnan et al., 1984; Soloff et al., 1982; Steiner et al., 1984; Sternbach et al., 1983; for review, see De la Fuente and Mendlewicz, 1996; Koenigsberg et al., 1999; Lahmeyer et al., 1989): (1) we excluded patients with a current major depressive disorder (MDD); (2) we controlled potential confounding variables, which may influence HPA-axis functioning such as smoking, menstrual cycle and physical activity; (3) we measured cortisol in saliva, thus preventing artefacts by stress-induced increases in HPA-axis activity due to invasive blood collection; (4) we carefully synchronised all sampling with the individual pattern of awakening of every participant; and (5), most importantly, we assessed HPA-axis functioning not only by determining cortisol levels at a single time point, but in a detailed analysis of cortisol levels on three consecutive days (for overview, see Kirschbaum and Hellhammer, 1994; Kirschbaum and Hellhammer, 1998).
2. Subjects and methods 2.1. Subjects The study was conducted at the Department of Psychiatry and Psychotherapy of the University of Freiburg Medical School and at the Department of Psychiatry and Psychotherapy, University of Luebeck Medical School (four patients). We included 23 caucasian female patients aged between 19 and 47 years and 24 female
healthy controls. All patients met the DSM-IV criteria for BPD, assessed by the appropriate segment of the Structured Clinical Interview for DSM-IV Personality Disorders (SCID-II) (First et al., 1996) and the Diagnostic Interview for Borderline Personality Disorder – Revised Version (DIB-R) (Zanarini et al., 1989). Cutoff scores were 5 or more out of 9 for the DSM-IV-criteria and 8 or more out of 10 for the DIB-R. Exclusion criteria were current comorbid MDD, current substance abuse, mental retardation or schizoaffective disorder as well as a life time history of schizophrenia or bipolar I disorder. Control subjects (healthy controls, HCs) were healthy women who did not fulfil criteria for BPD (SCID-II), were free of any former or present psychiatric disorder and had never sought psychiatric or psychotherapeutic help. Controls were students, clinic staff (from another department) or acquaintances of the experimenters. The HC were pooled and selectively included only if they were found similar to a patient according to the following requirements: smoking, age, and years of education. As shown in Table 1, there were no significant differences between the groups in any of these variables. However, patients with BPD were less frequently married and more often lived alone than the healthy controls, reflecting diminished stability of social relationships as typical of BPD. All participants were free of any medication at least for two months prior to the study and did not use oral contraceptives. Patients and healthy subjects gave written informed consent and received reimbursement for their efforts (a100 plus travelling fees). The study was carried out in accordance with the latest version of the Declaration of Helsinki and was approved by the local ethical board of the University of Freiburg Medical School. 2.2. Procedure To control for a possible modulation of cortisol by changes of estrogen or other hormones during the menstrual cycle, only subjects with regular menstrual cycles were included and all women, patients and con-
Table 1 Sample characteristicsa Characteristic
BPD (n ¼ 23)
HC (n ¼ 24)
Statisticsb
Age Gender (f/m) Smoker Weight (kg) Current partnership Own children
28.5 ± 1.6 23/0 17 (74) 63.6 ± 3.0 7 (30) 5 (22)
28.2 ± 1.4 24/0 18 (75) 65.8 ± 2.6 14 (58) 5 (21)
n.s. n.s. n.s. n.s. p < 0:05 n.s.
a
Data are given as means ± SD or number (percentage). BPD, borderline personality disorder; HC, healthy controls. b Mann–Whitney U test or v2 test.
K. Lieb et al. / Journal of Psychiatric Research 38 (2004) 559–565
trols, started at the fourth day of the menstrual cycle. The time period for the determination of cortisol on three consecutive days from the fourth to sixth day of the menstrual cycle was chosen because during these days of the follicular phase estrogen-levels are constantly low. A total of 32 saliva samples were obtained from each subject: 12 cortisol samples each on days 1 and 2, and eight samples on day 3 after taking 0.5 mg dexamethasone after the last collection of saliva on day 2 (DST) (for time points of cortisol determination, see Fig. 1). Morning measurements included five determinations in 15 min intervals beginning with awakening (‘‘cortisol immediately after awakening’’ and ‘‘total cortisol response to awakening’’) as well as seven determinations in 2 h intervals over the course of the day (‘‘total daily cortisol’’). To enable accurate timing, subjects were reminded by the portable mini-computer (PSION 3a, programmed by W. M€ uller, Department of Psychophysiology, University of Freiburg) to collect saliva samples by chewing the salivette. To control for confounding variables, at every single sampling point, participants were asked for additional information by the mini-computer: the degree of physical activity (as estimated on a 0–9 point Lickert scale) and smoking before sample collection (yes/no and number of cigarettes). They had to press specific keys on the minicomputer to answer the questions. To avoid dilution or spoiling of samples, participants were advised not to drink or eat for at least 5 min before sampling, the minicomputers being programmed with a ready to use control-function. 2.3. Determination of saliva cortisol levels Saliva was collected in Salivette devices (Sarstedt, Rommelsdorf, Germany). We decided to measure cor-
Awakening +15 +30 +45 +60 min. Awakening response
+2h
+4h
+6h
+8h
+10h +12h +14h
+2h
+4h
+6h
+8h
0.5 mg +10h +12h +14h Dex
+2h
+4h
+6h
+8h
+10h +12h +14h
Day 1
Day 2 Day 3
Fig. 1. Time points of cortisol determination. Participants were reminded by mini-computers to give saliva samples. The measurements started with awakening, followed by saliva sampling every 15 min for 1 h to assess ‘‘total cortisol response to awakening’’. Afterwards, sampling took place every second hour for determination of ‘‘total daily cortisol’’. Note that participants took 0.5 mg dexamethasone (DEX) on day 2 after the last sampling, and that no morning peak was sampled on day 3.
561
tisol in saliva-samples for three reasons: (1) the non-invasive sampling is suitable for an ambulatory setting; (2) it prevents artefacts by stress-induced increases in HPAaxis activity due to invasive blood collection; and (3) the free, biologically active fraction of cortisol (5–10% of the total fraction) is measured selectively. Salivary cortisol reliably reflects plasma-free cortisol concentrations with a correlation of measures of r P 0:90 (Kirschbaum and Hellhammer, 1994). Saliva samples were stored in a refrigerator until the the morning after the third day and were stored at )40 C until the time of assay. Saliva cortisol concentrations were determined using a time-resolved fluorescence immunoassay (DELFIA) as described by Dressendorfer et al. (1992) in the Biochemistry Laboratory of the Department of Psychobiology of Trier University (Dressendorfer et al., 1992). The intra-assay coefficient of variation range from 4.0% to 6.7% and the corresponding inter-assay coefficients of variation from 7.1% to 9.0%. Dexamethasone levels were not determined. 2.4. Data analysis Data are given as means ± SD. Areas under the curve (AUC) were calculated for the ‘‘total cortisol response to awakening’’ and the ‘‘total daily cortisol’’ by the formula ‘‘cort1 + cort2 + cort3 + cort4 + ((cort5 ) cort1)/ 2)’’ and ‘‘cort6 + cort7 + cort8 + cort9 + cort10 + cort11 + ((cort12 ) cort6)/2)’’, respectively. Data were statistically analysed by using paired t test, two-way analysis of variance and repeated measurements model analysis taking into account the correlations between samples from one patient (estimates were obtained via the GEE method by Liang and Zeger (1986)). Smoking and physical activity as well as body weight in case of DST were covariates in the ANCOVA analysis. According to Heim et al. (1998), the suppressor status in the low-dose DST was defined by an AUC-value for the ‘‘total daily cortisol’’ 6 5 nmol/l, the non-suppressor status by an AUC-value >5 nmol/l, and the super-suppressor status by an AUC-value <2 nmol/l. The number of patients at the three sampling days varied because cortisol samples were not always obtained for each subject at every sampling point. A p < 0:05 was defined as a significant difference and a p < 0:1 as a trend.
3. Results 3.1. Basal adrenocortical activity and cortisol response to awakening Salivary cortisol was determined on days 1 and 2 in intervals of 15 min after awakening and seven times in 2 h intervals after awakening during the whole day. As shown in Fig. 2(a) and (b), cortisol levels were higher in
562
K. Lieb et al. / Journal of Psychiatric Research 38 (2004) 559–565
(a)
50
HC
Day 1
BPD
Salivary Cortisol (nmol/l)
40
30
20
10
0:00
22:00
20:00
18:00
16:00
14:00
12:00
10:00
8:00
6:00
0
(b) 50
HC
Day 2
BPD
Salivary Cortisol (nmol/l)
40
30
20
10
3.2. Feedback sensitivity of the pituitary–adrenal axis
0:00
22:00
20:00
18:00
16:00
14:00
12:00
10:00
8:00
6:00
0
(c)
50
HC
Day 3 (post DEX)
BPD
40
Salivary Cortisol (nmol/l)
(‘‘cortisol immediately after awakening’’, Table 2), which was probably due to the high variability of cortisol levels especially in the patients with BPD. However, calculated AUCs for the ‘‘total cortisol response to awakening’’ and the ‘‘total daily cortisol’’ were significantly higher in patients with BPD than in controls (Table 2). Patients with BPD displayed a significantly higher cortisol response to awakening on day 2 and on days 1 and 2 in combination, whereas the cortisol response to awakening on day 1 showed a trend for higher levels in the patient group. Total daily cortisol was significantly higher on day 1. In an additional repeated measurements model analysis which takes into account the correlation between samples from one patient and which includes time as a covariate, we found identical results as obtained by comparison of AUC values. For example, comparison of diurnal cortisol secretion on day 1 revealed a significant group effect with an estimate for group differences of 3.03 nmol/l (95% CI [0.11–6.0], p < 0:05) and a significant time effect with an estimate for time effects of )0.9 nmol/l/2 h (95% CI [from )1.17 to )0.63], p < 0:0001). Since results from repeated measurements analysis were identical to the results obtained from AUC analysis, only results from AUC analyses are given in Table 2.
30
20
10
0:00
22:00
20:00
18:00
16:00
14:00
12:00
10:00
8:00
6:00
0
Fig. 2. Cortisol levels (means ± SD) in patients with BPD and healthy controls in the course of day 1 (a), day 2 (b), and day 3 after dexamethasone (c). Cortisol was sampled at the indicated time points by use of Salivette devices.
patients with BPD than in controls. These differences were not statistically significant at the single time point including the ‘‘awakening cortisol’’ on days 1 and 2
The low-dose DST was performed by applying 0.5 mg dexamethasone at 22:30 on day 2. This low dose of dexamethasone was used on the basis of our assumption that patients with BPD would show low cortisol levels and supersuppression to dexamethasone similar to patients with PTSD (Yehuda et al., 1993). As shown in Fig. 3, the oral ingestion of 0.5 mg dexamethasone induced a significant suppression of cortisol levels in both BPD patients and healthy controls (F ð1; 33Þ ¼ 71; p < 0:001). By using the criteria of Heim et al. (Heim et al., 1998) (see Section 2), significantly more patients with BPD were non-suppressors in the low-dose DST as compared to the healthy controls (14 vs. 4, v2 , p < 0:01). Only 1 patient and 4 controls were supersuppressors in the DST. ANCOVA did not reveal a significant dexamethasone · group interaction (F ð1; 33Þ ¼ 0:07; p ¼ 0:8). Baseline differences did not significantly influence suppression in an additional model taking baseline differences in cortisol levels as covariate. Although there was no dexamethasone · group interaction, total daily cortisol on day 3 was significantly higher in the patients with BPD compared to the healthy controls (Table 2) whereas cortisol immediately after awakening showed a trend for higher levels in patients with BPD. Cortisol levels after dexamethasone on day 3 were higher in the BPD patients at all time points (p < 0:05) (see Fig. 2(c)).
K. Lieb et al. / Journal of Psychiatric Research 38 (2004) 559–565
563
Table 2 Salivary cortisol levels (nmol/l) in patients with BPD and healthy controls (HCs) BPD (n ¼ 23)
Time
HC (n ¼ 24)
T
df
p
20 ± 13 28 ± 29 12 ± 16
18 ± 6 18 ± 9 5±6
0.91 1.48 2.00
45 40 39
0.37 0.15 0.05
113 ± 40 116 ± 47 115 ± 39
92 ± 35 91 ± 32 91 ± 29
1.92 2.06 2.36
45 44 44
0.06 0.04 0.02
Day 1 Day 2
78 ± 39 82 ± 63
58 ± 22 59 ± 18
2.13 1.54
44 40
0.04 0.14
Day 3
54 ± 49
25 ± 17
2.51
39
0.02
‘‘Cortisol immediately after awakening’’ (day 3 after DEX)
Day 1 Day 2 Day 3
‘‘Total cortisol response to awakening’’ (AUC)
Day 1 Day 2 Days 1 and 2
‘‘Total daily cortisol’’ (AUC) ‘‘Total daily cortisol after DEX’’ (AUC)
Data are given as means ± SD. BPD, borderline personality disorder; AUC, area under the curve; DEX, dexamethasone.
AUC Salivary Cortisol (nmol/l)
160
AUC Day 2
140
**
AUC Day 3
120 100
grating smoking as a covariate in the ANCOVA test, however, did not reveal a significant influence of smoking on cortisol secretion. Physical activity as a covariate in the ANCOVA had also no significant effect on cortisol levels.
**
80
4. Discussion
60 40 20 0 BPD
HC
Fig. 3. Effect of 0.5 mg dexamethasone on cortisol levels in 23 patients with BPD and 24 healthy controls. Values are expressed as means ± SD for AUCs of ‘‘total daily cortisol’’ on days 2 and 3. ** indicates significant suppression of cortisol secretion in both patients and controls. Differences between patients and controls are given in Table 2.
3.3. Influence of possible confounding variables To control for confounding variables which might possibly account for the differences in cortisol secretion between patients and controls, we assessed physical activity and smoking preceding the cortisol samplings. In a first analysis, we compared cortisol levels between smokers and non-smokers and found no significant differences in cortisol secretion. In detail, there was no difference between both groups with respect to the adrenocortical reactivity following awakening on day 1 (tð29; 742Þ ¼ 0:848, p ¼ 0:403) or on day 2 (tð30; 599Þ ¼ 0:012, p ¼ 0:991). Furthermore, no significant difference could be revealed with respect to the basal adrenocortical activity on day 1 (tð34Þ ¼ 0:087, p 6 0:9) and on day 2 (tð33Þ ¼ 1:1, p 6 0:3). At 5 time points (10.00 and 14.00 h on day 1; 14.00, 20.00 and 22.00 h on day 2), patients smoked significantly more cigarettes in the 2 h prior to saliva sampling, although the mean of cigarettes smoked did not exceed 2. Inte-
In this study, we followed a new approach in the study of patients with BPD, i.e., the measurement of diurnal cortisol secretion during natural, non-laboratory conditions. This approach was well accepted by our patients and appears to be a suitable approach to assess basic functions of HPA-axis activity in this patient group. Moreover, the approach revealed reliable findings: Measurements on two consecutive days (days 1 and 2) showed consistently increases in cortisol secretion with respect to total cortisol response to awakening and total daily cortisol although increases in the BPD patients not always reached statistical significance (e.g., the total cortisol response to awakening was only significantly different on day 2). This was probably due to the high variability of cortisol secretion in the patients as demonstrated by high standard deviations or by insufficient power to detect differences due to small sample sizes. Further research in a larger patient group has to elucidate whether there are subgroups which can be distinguished with respect to their cortisol secretion pattern. Although repetitive ambulatory measurements have the advantage that they increase the reliability of findings as compared to assessments of cortisol levels at single time points and that they reflect real life of the participants, they have obvious disadvantages, such as the difficulty to control for the compliance of the participants as well as the difficulty to control the variables which might confound cortisol levels. We do not believe that compliance problems do account for the differences seen between patients and controls. Confounding variables such as physical activity and smoking were
564
K. Lieb et al. / Journal of Psychiatric Research 38 (2004) 559–565
controlled with use of the mini-computers and we found no relevant influence. However, other possible confounding factors such as disturbances of the day-nightrhythm were not controlled, although shift workers were excluded from the study. A further strength of the study was the rigid control for a possible confounding influence of female sexual hormones. Only women with a normal menstrual cycle were included, use of contraceptives was not permitted and all women started cortisol measurements at the same day of the menstrual cycle. From a methodological point of view, diurnal cortisol profiles were – as the immediate response to awakening – gathered in relation to the individual awakening time and not at fixed times of the day. Although patients with BPD tended to wake up earlier that normal controls (7:43 ± 1:42 vs. 8:25 ± 1:53 h), this difference was not statistically significant. Since Kudielka and Kirschbaum (2003) had shown that a group of healthy early awakeners (mean awakening time 6:49 h) showed a higher cortisol response to awakening than healthy late awakeners (mean awakening time 9:43 h), it cannot fully be excluded that different awakening times had a possible confounding influence in our study. However, the difference in awakening times was 3 h in contrast to only 40 min in our study. Furthermore, additional analyses of our sample showed that there were no significant differences in cortisol levels between early and late awakeners in both controls and patients with BPD. Early awakeners even tended to have lower cortisol levels than late awakeners which confirms our conclusion that the observed differences in cortisol levels between patients and controls are not due to different awakening times. Several previous studies have investigated the activity of the HPA-axis in patients with BPD (Baxter et al., 1984; Carroll et al., 1981; Kontaxakis et al., 1998; Krishnan et al., 1984; Soloff et al., 1982; Steiner et al., 1984; Sternbach et al., 1983; for review see De la Fuente and Mendlewicz, 1996; Koenigsberg et al., 1999; Lahmeyer et al., 1989). Grossman et al. (2003) published a study on 52 patients with personality disorders (21 BPD patients) using a 0.5 mg DST. They found that subjects with personality disorders and comorbid PTSD had a significantly greater degree of cortisol suppression in response to a 0.5 mg DST than subjects without comorbid PTSD. After controlling for the effect of BPD diagnosis on cortisol suppression within patients with PTSD, there were no significant effects; however, it is possible that this finding was related to low statistical power. The authors did not report how many of the 16 subjects with comorbid PTSD had BPD, nor did they report separate data on the 21 BPD patients. A shortcoming of our study was that we did not assess comorbid PTSD diagnoses at the time of the study. A possible confounding influence of current PTSD diag-
noses can therefore not be excluded although exclusion of patients with a current PTSD would probably have led to an even stronger effect since PTSD has been associated with lower cortisol levels (Yehuda, 2002). Our finding of increased cortisol secretion and lowered feedback sensitivity of the HPA-axis as demonstrated by higher cortisol levels after 0.5 mg dexamethasone and a higher number of non-suppressors to dexamethasone is in line with previous studies showing non-suppression rates to 1 mg DST in BPD patients ranging from 9.5% (Lahmeyer et al., 1988) to 62% (Carroll et al., 1981). The reasons for our findings of increased cortisol secretion and lowered feedback sensitivity in patients with BPD are unknown. Obviously, they are not due to an overlapping affective disorder, since our patients were not currently depressed. However, increased cortisol levels in our patients might result from current life-time stress. As we have shown in previous studies, patients with BPD suffer from recurrent states of severe and aversive inner tension and dissociation (Stiglmayr et al., 2001) which might correlate with increased secretion of cortisol. Further studies assessing the psychopathological correlates of increased cortisol secretion need to be performed to investigate whether states of severe inner tension determine increased cortisol secretion in our patients. In conclusion, our study shows that the repetitive assessment of cortisol levels in saliva during non-laboratory, ambulatory conditions is a suitable method to get informations of HPA-axis functioning in patients with BPD. Furthermore, the study reliably demonstrated that patients with BPD show increased cortisol secretion and lowered feedback sensitivity to low-dose dexamethasone. Further studies have to reveal reasons of heightened adrenal activity in these patients.
Acknowledgements We thank Prof. Dr. Fahrenberg, Forschungsgruppe Psychophysiologie, University of Freiburg, Germany for help in the programming of the mini-computers, C. Unckel for patient care, I. Rummel-Fr€ uhauf and A. Fritzen, University of Trier, Germany for running the cortisol-assays, and G. Ihorst for help in statistical analyses. This study was supported by the Ministry of Baden W€ urttemberg and the Borderline Personality Disorder Research Foundation (BPDRF).
References Baxter L, Edell W, Gerner R, Fairbanks L, Gwirtsman H. Dexamethasone suppression test and Axis I diagnoses of inpatients with DSM-III borderline personality disorder. Journal of Clinical Psychiatry 1984;45:150–3.
K. Lieb et al. / Journal of Psychiatric Research 38 (2004) 559–565 Carroll BJ, Greden JF, Feinberg M, Lohr N, James NM, Steiner M, Haskett RF, Albala AA, DeVigne JP, Tarika J. Neuroendocrine evaluation of depression in borderline patients. Psychiatric Clinics of North America 1981;4:89–99. De la Fuente JM, Mendlewicz J. TRH stimulation and dexamethasone suppression in borderline personality disorder. Biological Psychiatry 1996;40:412–8. Dressendorfer RA, Kirschbaum C, Rohde W, Stahl F, Strasburger CJ. Synthesis of a cortisol–biotin conjugate and evaluation as a tracer in an immunoassay for salivary cortisol measurement. Journal of Steroid Biochemistry and Molecular Biology 1992;43:683–92. First MB, Spitzer RL, Gibbon M, Williams JBW. User’s guide for the Structured Clinical Interview for DSM-IV personality disorders (SCID-II). Washington, DC: American Psychiatric Press; 1996. Grossman R, Yehuda R, New A, Schmeidler J, Silverman J, Mitropoulou V, Sta Maria N, Golier J, Siever L. Dexamethasone suppression test findings in subjects with personality disorders: associations with posttraumatic stress disorder and major depression. American Journal of Psychiatry 2003;160:1291–8. Heim C, Ehlert U, Hanker JP, Hellhammer DH. Abuse-related posttraumatic stress disorder and alterations of the hypothalamic– pituitary–adrenal axis in women with chronic pelvic pain. Psychosomatic Medicine 1998;60:309–18. Holsboer F. Stress, hypercortisolism and corticosteroid receptors in depression: implications for therapy. Journal of Affective Disorders 2001;62:77–91. Kirschbaum C, Hellhammer DH. Salivary cortisol in psychoneuroendocrine research: recent developments and applications. Psychoneuroendocrinology 1994;19:313–33. Kirschbaum C, Hellhammer DH. Hypothalamus-hypohysen-nebennierenrinden-achse. In: Kirschbaum C, Hellhammer DH, editors. Enzyklop€ adie der Psychologie, Band Psychoendokrinologie und Psychoimmunologi. 1998. p. 79–140. Koenigsberg HW, Anwunah I, New AS, Mitropoulou V, Schopick F, Siever LJ. Relationship between depression and borderline personality disorder. Depression and Anxiety 1999;10:158–67. Kontaxakis V, Markianos M, Vaslamatzis G, Markidis M, Kanellos P, Stefanis C. Multiple neuroendocrinological responses in borderline personality disorder patients. Acta Psychiatrica Scandinavica 1998;76:593–7.
565
Krishnan KR, Davidson JR, Rayasam K, Shope F. The dexamethasone suppression test in borderline personality disorder. Biological Psychiatry 1984;19:1149–53. Kudielka BM, Kirschbaum C. Awakening cortisol responses are influenced by health status and awakening time but not by menstrual cycle phase. Psychoneuroendocrinology 2003;28:35–47. Lahmeyer HW, Reynolds III CF, Kupfer DJ, King R. Biologic markers in borderline personality disorder: a review. Journal of Clinical Psychiatry 1989;50:217–25. Lahmeyer HW, Val E, Gaviria FM, Prasad RB, Pandey GN, Rodgers P, Weiler MA, Altman EG. EEG sleep, lithium transport, dexamethasone suppression, and monoamine oxidase activity in borderline personality disorder. Psychiatry Research 1988;25:19– 30. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika 1986;73:13–22. Soloff PH, George A, Nathan RS. The dexamethasone suppression test in patients with borderline personality disorders. American Journal of Psychiatry 1982;139:1621–3. Steiner M, Martin S, Wallace JE, Goldman S. Distinguishing subtypes within the borderline domain: a combined psychoneuroendocrine approach. Biological Psychiatry 1984;19:907–11. Sternbach HA, Fleming J, Extein I, Pottash AL, Gold MS. The dexamethasone suppression and thyrotropin-releasing hormone tests in depressed borderline patients. Psychoneuroendocrinology 1983;8:459–62. Stiglmayr CE, Shapiro DA, Stieglitz RD, Limberger MF, Bohus M. Experience of aversive tension and dissociation in female patients with borderline personality disorder – a controlled study. Journal of Psychiatric Research 2001;35:111–8. Yehuda R. Current status of cortisol findings in post-traumatic stress disorder. Psychiatric Clinics of North America 2002;25:341– 68. Yehuda R, Boisoneau D, Mason JW, Giller EL. Glucocorticoid receptor number and cortisol excretion in mood, anxiety, and psychotic disorders. Biological Psychiatry 1993;34:18–25. Zanarini MC, Gunderson JG, Frankenburg FR, Chauncey DL. The revised Diagnostic Interview for Borderlines: discriminating BPD from other Axis II disorders. Journal of Personality Disorders 1989;3:10–8.