There is growing evidence that extremely stressful adverse experiences have a lasting impact on the neurobiology of the stress response, involving both baseline (tonic) abnormalities [5,6] and abnormal responsivity (phasic abnormalities) of the hormonal stress system [5,7]. Stress leads to the excretion of corticotrophin-releasing factor (CRF) and arginine vasopressin (AVP) into the hypophysial blood supply, where they are transported to the adenohypophysis. Here they activate pituitary corticoptrophs to synthesize and release adrenocorticotrophic hormone (ACTH) into the general blood circulation. CRH is the most potent ACTH secretagogue. Although the role of AVP is not yet completely understood, it seems to be involved in the regulation of stress-induced ACTH release [8]. ACTH circulates in the blood to the zona fasciculata of the adrenal cortex, where it promotes the conversion of cholesterol esters into free cholesterol and ultimately results in the release of cortisol as the end product of the steroid pathway from the adrenal cortex into the circulatory system. Cortisol peaks are typically seen 15–30 minutes after an ACTH pulse in normal human subjects [for a summary see [9]]. After an acute psychological stressor, peak cortisol responses occur 20–40 minutes from the onset of the stressor in healthy controls, with the length of the stressor not associated with observed effect sizes (i.e. longer stressors do not lead to greater cortisol responses than shorter ones [10]).

Several studies have investigated the effect of traumatic life experiences and PTSD on the pituitary-adrenal cortical system. However, results have been inconsistent and sometimes even contradictory [for an overview see e.g., [5,7]]. According to the original glucocorticoid cascade hypothesis [11], chronic exposure to glucocorticoids throughout life, secondary to repeated or traumatic stress, downregulates the central glucocorticoid receptors, especially at the hippocampal level. This causes impairment of HPA sensitivity to the negative steroid feedback, leading to glucocorticoid overproduction in a feedforward loop. In contrast to this original hypothesis, several studies reported lower baseline cortisol levels in PTSD in plasma [12-14], saliva [15-17], and urine [18-21], but some studies have found significantly higher cortisol excretion [e.g., [22]] or no differences between groups [e.g., [23-25]].

In addition to the investigation of baseline alterations in PTSD patients, several challenge paradigms have been developed to investigate the stress response in PTSD patients. These paradigms can be broadly distinguished in pharmacological and non-pharmacological challenge tests [for a review see [7]]. Pharmacological challenge tests target the HPA axis at different levels, e.g. dexamethasone suppression test, ACTH-, CRH-, and naloxone challenges, dexamethasone CRH test and metyrapone challenge designs. In their review, de Kloet et al. [7] conclude that enhanced cortisol suppression after administration of .5 mg dexamethasone is a relatively well-corroborated finding. Otherwise the results of pharmacological challenge paradigms are still inconclusive.

In comparison to pharmacological challenge tests relatively few non-pharmacological challenge tests have been conducted with PTSD patients [26-30]. In these paradigms, cortisol levels were investigated after cognitive [29], psychosocial [30], or physical challenges [28], as well as in response to personalized traumatic scripts [27] and trauma reminders [26].

Challenge paradigms using non-trauma-related stressors such as psychosocial or physical stressors found mixed results: Using a cognitive challenge task, Bremner et al. [29] found higher mean salivary cortisol levels in anticipation of, and during the stressor in PTSD patients compared to controls. This effect was more pronounced in male PTSD patients compared with female PTSD patients. However, aside from these baseline differences, no evidence for a changed cortisol response to cognitive stressors in PTSD was observed. Instead, the cortisol response was similar in the PTSD and control group. Using a physical stressor (cold pressor task), Santa Ana et al. [28] investigated ACTH/cortisol excretion in individuals with PTSD, comorbid alcohol dependence and PTSD, and controls. In this paradigm, subjects immerse one hand in a cold water bath for up to 1 minute or as long as they can tolerate. Regardless of the presence or absence of comorbid alcohol dependence, subjects with childhood trauma and PTSD (but not adult trauma and PTSD) had lower plasma cortisol at baseline and at all post-task measurement points. No differences in cortisol reagibility were observed. However, while control persons exhibited an initial ACTH increase in response to the stressor, traumatized persons (childhood and adult trauma) showed a blunted ACTH excretion. In contrast, Heim et al. [30] reported a persistent sensitization of the HPA axis in adulthood even to mild stressors in women with severe early-life stress. In response to a psychosocial laboratory stressor, physically and sexually abused women with and without current major depression exhibited increased ACTH concentrations compared to controls and non-abused depressed women. In contrast, only women with early-life stress and current major depression showed enhanced plasma cortisol values 30 to 60 minutes after the stressor.

The results of challenge paradigms using trauma-related stressors, such as trauma reminders or personalized traumatic scripts are also inconclusive. In a symptom provocation study using trauma-related stimuli (combat sounds) compared to nonspecific arousing stimuli (white noise), Liberzon et al. [26] observed enhanced plasma cortisol and catecholamines at baseline, but no differences in ACTH or cortisol secretion in response to both stressors (combat sounds, white noise) in patients with PTSD compared to controls. One major problem with this study is that plasma samples were drawn immediately before and after playing a 3 minute audiotape. An effect of a stressor on cortisol values cannot be expected after such a short time interval. On the other hand, Elzinga et al. [27] reported elevated salivary cortisol levels in women with PTSD during, and shortly after, exposure to a personalized trauma script. The problem in this study is that the cortisol levels were elevated before the script exposure in the PTSD group compared to the control group, perhaps due to anticipatory anxiety. It is unclear whether effects would have been significant if the difference values had been analyzed or if the baseline differences had been corrected using cortisol values directly before the stressor as a covariate.

Finally, in a single case study Otte et al. [31] investigated subjective distress and salivary cortisol during the first and 20^th session of an imaginal exposure treatment. They observed extreme anticipatory anxiety before and in the beginning of the first exposure session, which was markedly reduced at the end of the therapy. However, they observed no increase in cortisol during first exposure, and cortisol values at the end of the therapy were not different from the beginning.

In summary, the results of non-pharmacological challenge paradigms are as inconclusive as the results of pharmacological challenge paradigms in PTSD. Studies using cognitive, psychosocial, or physical stressors revealed no evidence of abnormalities in cortisol reagibility [28-30]. However, it remains unclear whether findings from cognitive, psychosocial, or physical challenge paradigms can be generalized to emotional stressors, such as being exposed to reminders of one's trauma. To our knowledge, only two studies so far investigated cortisol secretion in PTSD patients during or after exposure to trauma-related stimuli with conflicting results: One found no increase in plasma cortisol in response to the stressor [26], whereas the other found increased cortisol to personalized traumatic scripts [27]. However, as detailed above, both studies had limitations that make the interpretation of results difficult.

Thus, it remains unclear whether inquiring about a traumatic experience, for example during a psychodiagnostic interview, for expert opinion, or in the course of exposure-based therapies, such as narrative exposure therapy [32], causes a level of stress that activates the HPA axis and results in excessive cortisol excretion in PTSD patients. Therefore, the aim of this study was to investigate whether persons with severe forms of PTSD, such as survivors of torture and violent oppression, show an excessive stress response with high cortisol excretion in response to detailed questioning about their torture/war experiences. Samples of saliva were taken four times during a diagnostic interview (baseline, pre-stressor, post-stressor, recovery). Individuals with PTSD in the stress group were interviewed about their torture experiences using a standardized checklist. Persons with PTSD in the control condition were interviewed about their absorption behavior in various situations. Directly before and after the stressor, all participants performed a memory task, which acted as a distraction, and rated selected PTSD symptoms on a visual analog scale.