Dexamethasone suppression test in borderline personality disorder: Impact of PTSD symptoms

Authors

  • KATJA WINGENFELD phd,

    Corresponding author
    1. Department of Psychiatry and Psychotherapy Bethel, Ev. Hospital Bielefeld, Bielefeld,
    Search for more papers by this author
  • ANDREAS HILL md ,

    1. University of Schleswig-Holstein, Campus Luebeck,
    2. Institute of Sexology and Forensic Psychiatry, Center for Psychosocial Medicine, University Hospital Hamburg-Eppendorf, Hamburg,
    Search for more papers by this author
  • BETTINA ADAM md ,

    1. University of Schleswig-Holstein, Campus Luebeck,
    2. August Bier Hospital, Bad Malente-Gremsmühlen, and
    Search for more papers by this author
  • MARTIN DRIESSEN md , phd

    1. Department of Psychiatry and Psychotherapy Bethel, Ev. Hospital Bielefeld, Bielefeld,
    2. University of Schleswig-Holstein, Campus Luebeck,
    3. University of Bielefeld, Bielefeld, Germany
    Search for more papers by this author

Katja Wingenfeld, PhD, Department of Psychiatry and Psychotherapy Bethel, Ev. Hospital Bielefeld, Remterweg 69-71, D-33617 Bielefeld, Germany. Email: katja.wingenfeld@evkb.de

Abstract

Abstract  The purpose of the present study was to investigate the impact of post-traumatic stress disorder (PTSD) symptoms on hypothalamic–pituitary–adrenal axis feedback regulation in 18 female patients with borderline personality disorder (BPD) and 21 healthy controls. Reduced feedback sensitivity was found in BPD patients with a low number of PTSD symptoms, while findings in the BPD group with a high number of PTSD symptoms did not differ from those in controls. The results suggest a hypo-suppression in the dexamethasone suppression test in BPD with few PTSD symptoms.

INTRODUCTION

In post-traumatic stress disorder (PTSD) and major depression (MD) alterations of hypothalamic–pituitary–adrenal (HPA) axis functioning have been reported. While increased baseline cortisol levels and a decreased cortisol suppression in the 1-mg dexamethasone suppression test (DST) has been found in MD, PTSD was characterized by hypocortisolism and an enhanced suppression after 0.5 mg dexamethasone (DEX).1,2 PTSD as well as MD are common comorbid disorders in patients with borderline personality disorder (BPD).3 In BPD, low rates as well as high rates of cortisol non-suppressors in the DST have been reported, which might be due to different rates of comorbid axis I disorders.4 Recently we did not find differences of the HPA axis feedback regulation between BPD patients and controls.5 However, when subdividing the BPD group into patients with and without comorbid PTSD, we observed lower cortisol suppression in those without PTSD.

Due to the fact that traumatization is very frequent in BPD, it may be inadequate (formal or even artificial) to categorize samples of BPD in patients with and without PTSD. In the present study we followed a dimensional approach in investigating the impact of the number of PTSD symptoms on HPA axis feedback regulation in patients with BPD.

METHODS

Eighteen female patients with BPD and 21 healthy women were enrolled in the present study. Exclusion criteria were medical illness, pregnancy, anorexia, schizophrenia, schizoaffective disorder, alcohol and drug dependence. Patients had been drug free for at least 7 days. All participants had given their written informed consent. The study was approved by the institutional review board.

BPD was evaluated on the Structured Clinical Interview for DSM-IV (SCID-II), while the SCID-I sections was used to cover PTSD.6

Trauma history was assessed using the Childhood Trauma Questionnaire (CTQ).7 PTSD psychopathology was estimated by summing up the number of reported symptoms according to DSM-IV criteria B–D. Depressive psychopathology was measured using the Beck Depression Inventory (BDI).8

Feedback regulation of the HPA axis was investigated on low-dose DST. Subjects took 0.5 mg DEX at 23.00 hours, blood samples were drawn at 08.00 hours before DEX, and at 08.00 hours and 16.00 hours the following day. Cortisol was assayed on radioimmunoassay.

Demographic and clinical data were analyzed using t-test for continuous data. Hormone data were analyzed using analysis of covariances and partial correlations. Significance level was set at P < 0.05.

RESULTS

Sociodemographic, clinical and endocrine data are presented in Table 1.

Table 1.  Sample characteristics and endocrine data (mean ± SD)
CharacteristicsBPD patientsControlsP
  •  t-test.

  • BDI, Beck Depression Inventory; BPD, borderline personality disorder; CTQ, Childhood Trauma Questionnaire; DEX, dexamethasone.

Age (years)27.83 (5.5)28.74 ± 6.70.65
Body mass index24.3 ± 6.123. 6 ± 4.10.68
Education (years)9.8 ± 1.210.6 ± 1.80.13
BDI28.6 ± 13.23.04 ± 3.8<0.001
CTQ
 Total score11.7 ± 3.45.5 ± 0.7<0.001
 Physical and emotional abuse3.2 ± 0.91.4 ± 0.3<0.001
 Emotional neglect3.5 ± 0.71.9 ± 0.4<0.001
 Physical neglect2.2 ± 0.71.1 ± 0.1<0.001
 Sexual abuse2.7 ± 1.41.0 ± 0.1<0.001
Cortisol (nmol/dL)
 Before DEX (08.00 hours)55.8 ± 20.369.3 ± 19.10.04
 After DEX (08.00 hours)14.3 ± 19.112.3 ± 10.70.68
 After DEX (16.00 hours)19.1 ± 26.212.0 ± 8.70.25

In the DST, repeated measures anova indicated a significant effect of the factor time, indicating a reduced cortisol release after DEX (P < 0.001). There was also a significant time × group interaction (P = 0.004) but no effect of the factor group.

Partial correlations were computed between percentages of cortisol suppression and number of PTSD symptoms and BDI scores, respectively. When controlling for BDI we found a significant positive correlation between percentage of cortisol suppression and PTSD symptoms at 16.00 hours (r = 0.436, P = 0.006) but not at 08.00 hours (r = 0.269, P = 0.102), and a significant negative association between BDI and percentage of cortisol suppression at 08.00 hours (r = −0.354, P = 0.029) and at 16.00 hours (r = − 514, P = 0.001), when controlling for PTSD symptoms.

We further subdivided the BPD sample via median split concerning number of reported PTSD symptoms, and BDI score, respectively.

When comparing BPD patients with high and low number of PTSD symptoms (Fig. 1) and controls, repeated measures anova showed a significant effect of the main factor time (P < 0.001) and a significant time × group interaction (P = 0.005), while there was a trend towards a group effect (P = 0.077). Post-hoc analysis indicated significant differences between controls and patients with low number of PTSD symptoms (P = 0.005) and between patients with low number of PTSD symptoms versus patients with high number of PTSD symptoms (P = 0.007).

Figure 1.

Cortisol release before and after dexamethasone (DEX) in (●) controls; (○) borderline personality disorder (BPD) patients with low number of post-traumatic stress disorder symptoms; and (▾) BPD patients with a high number of PTSD symptoms (mean ± SE).

When comparing BPD subgroups with low and a high BDI score, no significant group effect (P = 0.932) nor group × time interaction (P = 0.603) was found.

DISCUSSION

In the present study a positive association between PTSD symptoms and percentage of cortisol suppression and a negative association between depression scores and cortisol suppression have been found in BPD patients.

This is in line with the hypotheses that diverging results concerning HPA axis feedback regulation in BPD might be related to different rates of comorbid axis I disorders.4 In the present study and the Lange et al. study we found reduced relative cortisol suppression after DEX in some but not all BPD patients.5 Interestingly, this escape from DEX suppression was seen only in BPD patients with a low severity of PTSD psychopathology, and no PTSD diagnosis, respectively. The present results suggest a reduced HPA axis feedback sensitivity in BPD with a low number of PTSD but not any peculiarities, namely no super-suppression in the DST in BPD patients with a high number of PTSD symptoms.

In MD, elevated cortisol levels after DEX and hyperactivity of the HPA axis have been found.1 Interestingly, hypercortisolism and non-suppression in the DST have also been reported for BPD patients, which is in line with the negative association between depression scores and percentage of cortisol suppression in the present study.4,9,10 In fact, some of the cardinal features of BPD are affect-related symptoms and one might suggest that BPD patients have a pattern of HPA axis alterations similar to those reported for depressive patients. In contrast, severity of comorbid PTSD seems to have an important impact on HPA axis regulation, resulting in differences between patients with and without PTSD.5,10

The present study does not support the hypotheses of major alterations of HPA feedback sensitivity in BPD patients in general, but emphasizes the impact of comorbid PTSD and depressive psychopathology. Trauma-related and depressive symptoms seem to interact with regard to their effects on HPA axis regulation or might even cancel each other out by opposite effects.

ACKNOWLEDGMENT

This study was funded by the Deutsche Forschungsgemeinschaft (grant Dr 358/2-1).

Ancillary