The cortisol stress response and its relation to psychotherapy outcome in panic disorder patients

CONTENT

Summary

1. Introduction
1.1 Panic disorder: definitions and aetiology
1.2 Treatment options for panic disorder and therapy success
1.3 Cortisol stress response in panic disorder patients compared to healthy individuals as well as post-traumatic stress disorder and major depressive disorder patients

2. Aims

3. Methods

4. Study I: Effects of the cortisol stress response on the psychotherapy outcome for panic disorder patients
4.1 Abstract
4.2 Introduction
4.3 Methods
4.4 Results
4.5 Discussion

5. Study II: Cortisol stress response in post-traumatic stress disorder, panic disorder and major depressive disorder patients
5.1 Abstract
5.2 Introduction
5.3 Methods
5.4 Results
5.5 Discussion

6. Study III: Stress hormone response to the DEX-CRH test and its relation to psychotherapy outcome in panic disorder patients with and without agoraphobia
6.1 Abstract
6.2 Introduction
6.3 Methods
6.4 Results
6.5 Discussion

7. Discussion
7.1 Discussion of main findings
7.2 Implications for clinical practice
7.3 Implications for future research
7.4 Strengths and limitations

8. Conclusions

9. References

10. Appendix
10.1 Acknowledgements
10.2 Erklärung zu den Eigenanteilen an einzelnen Publikationen
10.3 Eigenständigkeitserklärung

DANKSAGUNG

An dieser Stelle möchte ich mich bei den Personen bedanken, die mich in den letzten Jahren begleitet und unterstützt haben. Besonderer Dank gilt meinen wissenschaftlichen Betreuern Clemens Kirschbaum und Katja Petrowski für ihr entgegengebrachtes Vertrauen, die Möglichkeit der Promotion, der Förderung und für die hilfreichen Rückmeldungen.

Ich möchte auch den Freunden und Kollegen danken, darunter Kathleen Philipp, Friederike Meyer, Marlene Penz, Magdalena Wekenborg, Susann Schmiedgen und Tarik Hizli, für ihre Freundschaft, Kollegialität und Intervision im Arbeitsalltag der scientific practitioners.

Zuletzte danke ich Rene Dietz und meiner Familie für ihren liebevollen Rückhalt.

NOTE

The present manuscript forms a publication-based thesis. In accordance to § 8 (1) of the doctorate regulations of the school of sciences of the Technische Universität Dresden, this thesis has been prepared as a self-contained work. The chapters were composed specifically for this thesis, however substantial parts have been already published. The following sections are based on published original research articles:

Section 4 (referred to as study I within this thesis):

Wichmann, S., Kirschbaum, C., Lorenz, T., & Petrowski, K. (2017). Effects of the cortisol stress response on the psychotherapy outcome for panic disorder patients. Psychoneuroendocrinology, 77, 9-17. (IF 2017: 4.731)

Section 5 (referred to as study II within this thesis):

Wichmann, S., Kirschbaum, C., Böhme, C., & Petrowski, K. (2017). Cortisol stress response in post-traumatic stress disorder, panic disorder and major depressive disorder patients . Psychoneuroendocrinology, 83, 135-141. (IF 2017: 4.731)

Section 6 (referred to as study III within this thesis):

Wichmann, S., Bornstein, S. R., Lorenz, T., & Petrowski, K. (2017). Stress hormone response to the DEX-CRH test and its relation to psychotherapy outcome in panic disorder patients with and without agoraphobia. Translational Psychiatry, 8 (37), 1-9 (2-year IF 2017: 4.691)

The appendix identifies own work in the publications. For an improved readability, contents, tables and figures were numbered continuously and the journal-specific reference styles were standardised and integrated in one combined reference section at the end of this thesis.

Wichmann, S., Bornstein, S. R., Lorenz, T., & Petrowski, K. (2017). Stress hormone response to the DEX-CRH test and its relation to psychotherapy outcome in panic disorder patients with and without agoraphobia. Translational Psychiatry, 8 (37), 1-9 (2-year IF 2017: 4.691)

LIST OF TABLES

Table 1: Characteristics of the total sample. Mean (SD) are listed except where noted

Table 2: Clinical measures before and after psychotherapy for panic disorder patients

Table 3: Characteristics of the total sample. Mean (SD) are listed except where noted

Table 4: Characteristics of the total sample. Mean (SD) are listed except where noted

Table 5: Clinical measures before and after psychotherapy for panic disorder patients

LIST OF FIGURES

Figure 1: Mean (± SE) blood plasma cortisol (left) and ACTH (right) concentrations of panic disorder (PD) patients and healthy control participants

Figure 2: Scatterplot showing the relationship between mean cortisol concentration during the TSST and symptom severity posttherapy

Figure 3: Mean (± SE) blood plasma cortisol (left) and ACTH (right) concentrations of panic disorder (PD) patients with/without comorbid depression and healthy control participants

Figure 4: Mean (± SD) blood cortisol (left) and ACTH (right) concentrations of panic disorder (PD), post-traumatic stress disorder (PTSD), major depressive disorder (MDD) patients and healthy control (HC) participants

Figure 5: Mean (± SD) blood plasma cortisol concentrations in response to the DEX-CRH test in panic disorder patients with and without agoraphobia (PD) and healthy volunteers (HC)

Figure 6: Scatterplot illustrating the relation between mean cortisol concentration in response to the DEX-CRH test and psychotherapy outcome percentage scores in agoraphobic cognitions for panic disorder patients

LIST OF ABBREVIATIONS

ACQ Agoraphobic Cognitions Questionnaire

ACTH Adrenocorticotropic hormone

ANOVA Analyses of variance

ANCOVA Analysis of covariance

BDI Beck Depression Inventory

BSQ Body Sensations Questionnaire

CBT Cognitive behavioural psychotherapy

CRH Corticotropin-releasing hormone

CTQ Childhood Trauma Questionnaire

DEX-CRH Dexamethasone Suppression/Corticotropin-releasing Hormone Challenge

DSM Diagnostic and Statistical Manual of Mental Disorders

DST Dexamethasone suppression test

FDS-44 Fragebogen zu Dissoziativen Symptomen

HC Healthy control

HPA Hypothalamic-pituitary-adrenal

IES-R Impact of Event Scale - Revised

MDD Major depressive disorder

MI Mobility Inventory

PAS Panic & Agoraphobia Scale

PASA Primary Appraisal Secondary Appraisal Scale

PD Panic disorder

PTSD Post-traumatic stress disorder

TSST Trier Social Stress Test

VAS Visual Analogue Scale

0 SUMMARY

Background. Although effective treatment is available for panic disorder (PD), a proportion of patients experience an incomplete remission and relapses after the first-line psycho-therapeutic treatment. To date, previous research has been unable to identify reliable therapy outcome predictors of a psychological or disorder-related kind. Initial attempts to understand the aetiology and treatment of PD put an emphasis on biological models, failed however to examine potential biological predictors for therapy success. Further, for understanding the biological correlates of PD, previous reports focused extensively on the distinction of the hypothalamic-pituitary-adrenal-(HPA) axis-mediated cortisol stress response when comparing patient samples and healthy populations, with little emphasis on a direct comparison of patient groups. In comparison with healthy individuals, cortisol hypo-responsiveness is a relatively consistent finding in panic-centric research. Thus, the current thesis is aimed at improving the existing knowledge about the cortisol stress response in PD patients with and without agoraphobia in relation to other psychiatric disorders as well as in reference to the psychotherapy success.

Methods. The samples of the PD patients with and without agoraphobia seeking psychological treatment and matched healthy individuals were confronted with both a psychosocial (Trier Social Stress Test, TSST) and hormonal stressor (Dexamethasone Suppression/Corticotropin-releasing Hormone Test, DEX-CRH) before the start of cognitive behavioural psychotherapy (CBT) in a semi-residential care setting (for the PD patients). In addition, we exposed samples of patients with post-traumatic stress disorder (PTSD) and major depressive disorder (MDD) to the TSST to compare the hormonal stress response between these different diagnostic groups directly. Blood samples were taken to determine the challenged adrenocorticotropic hormone (ACTH) and the cortisol secretion. Established self-rated questionnaires evaluated the panic symptom severity in reference to global severity, agoraphobic cognitions, fear of bodily sensations, agoraphobic avoidance, and depressiveness at admission and discharge. We conducted repeated measures ANCOVAs to test the pretherapy hormonal responses for main effects of time or group and for interaction effects and the Pearson’s correlation analysis to test for associations with the psychotherapy outcome. We predicted that a TSST cortisol non-response (< 55.2 nmol/l) and DEX-CRH challenged cortisol non-suppression (≥ 138 nmol/l) would be associated with a psychotherapy non-response. We further predicted cortisol hypo-responsiveness in both the PD and PTSD patients compared to the MDD patients and the healthy individuals. Further, we predicted that the MDD patients would demonstrate a cortisol response higher than demonstrated by the anxiety patients, but lower than the healthy individuals.

Results. The statistical analyses revealed large (η 2 ≥ 0.14) effect sizes for CBT in the clinical outcome measures ranging at 0.321 ≥ η 2 ≥ 0.648 and significant main effects of time for cortisol and the ACTH concentration in response to both the TSST and the DEX-CRH test, independent of the study group. 42.9% of the PD patients and 65.6% of the healthy control participants showed an adequate cortisol stress response to the TSST, and 82.4% of the PD patients and 85.3% of the healthy individuals showed a DEX-CRH stimulated cortisol suppression (descriptive findings). The data showed a significant inverse association of the pretherapy TSST cortisol stress response with agoraphobic avoidance and an association of the pretherapy cortisol release upon the DEX-CRH challenge and the agoraphobic cognitions at discharge. Further, we observed a significant association of the subjective level of fear during the TSST and the fear of bodily sensations as well as a trend for an association with agoraphobic cognitions. The direct comparison of the patient groups revealed a cortisol stress response upon the TSST in 26.7% of the PD patients, 43.5% of the PTSD patients, 72.2% of the MDD patients, and 80.6% of the healthy participants. ANCOVA revealed a cortisol hypo-responsiveness in both the PD and the PTSD patients, whereas no significant group differences were seen in the ACTH concentrations. The MDD patients did not differ significantly in the hormonal stress response compared to the healthy participants or to the PD and the PTSD patients.

Discussion. Overall, this thesis found that an adequate pretherapy psycho-endocrine response to a psychosocial (cortisol increase) and hormonal stressor (cortisol suppression) was related to greater therapy success later-on. Specifically, higher psychosocially challenged cortisol levels were related to an improvement in agoraphobic avoidance, and an adequate cortisol suppression upon the pharmacological challenge was related to an improvement in agoraphobic cognitions following a five-week-CBT. High subjective distress during psychosocial stress induction was related to an improvement in fear of bodily sensations and (descriptively) to an improvement in agoraphobic cognitions. The direct comparison of the patient groups revealed evidence of a dissociation between the cortisol and the ACTH concentrations in response to the TSST in both the PD and the PTSD patients. Our results further support the overall research findings of a cortisol hypo-responsiveness in PD. Descriptively, patients with a current MDD diagnosis showed the expected cortisol response pattern in comparison to the other study groups.

Conclusion. To conclude, the results tentatively suggest that the cortisol stress response discriminates between the DSM-IV categories (PD and PTSD patients vs. MDD patients) and between psychotherapy (non-) responders with reference to panic-disorder patients with and without agoraphobia. The results implicate the examination of the HPA-axis responsivity and its change during treatment as an additional tool in evaluating the efficacy of psychotherapeutic interventions for PD, supplementary to self-report and clinical measures. The prediction of the therapy success from a potential endocrine correlate, whose persistency (if assessed repeatedly) during treatment may predict (non-) response to the current treatment, possibly represents a decision support for a change in treatment to avoid the continuation of inefficient treatment. The results further indicate the possible manifestation of an endocrine transdiagnostic factor in the aetiology/course of PD and PTSD. In light of some of the study limitations, implications for clinical practice and detailed future research efforts are proposed.

0 Zusammenfassung

Theoretischer Hintergrund. Obwohl wirksame Behandlungsmethoden für die Panikstörung existieren, erlebt ein erheblicher Teil der Patienten keine vollständige Remission und Rezidive nach einer psychotherapeutischen Behandlung der ersten Wahl. Frühere Studien konnten keine zuverlässigen psychologischen oder störungsspezifischen Prädiktoren für den Therapieerfolg identifizieren. Gleichwohl erste Forschungsarbeiten um die Ätiologie und Behandlung der Panikstörung biologische Modelle präsentierten, sind Forschungsbemühungen zur Identifikation biologischer Prädiktoren für den Therapieerfolg überraschend selten. Zudem hat die überwiegende Mehrzahl der Studien zu den biologischen Korrelaten der Panikstörung hauptsächlich die Hypothalamus-Hypophysen-Nebennierenrinden-(HPA) Achse und ihr Endprodukt Kortisol im Vergleich mit gesunden Kontrollpersonen betrachtet. Es existieren kaum Studien, die die Kortisol-Stressreaktion zwischen Gruppen von Patienten mit verschiedenen psychischen Störungen verglichen haben. Im direkten Vergleich mit gesunden Kontrollpersonen haben Studien relativ übereinstimmend eine Kortisol-Hyporeaktion bei Patienten mit Panikstörung beobachtet. Diese Dissertation beabsichtigt, unser Wissen über die Kortisol-Stressreaktion bei der Panikstörung sowohl im direkten Zusammenhang mit anderen psychischen Störungen als auch im Zusammenhang mit dem Therapieerfolg zu erweitern.

Methodik. Patienten mit gesicherter Diagnose einer Panikstörung mit und ohne Agoraphobie sowie gematchte gesunde Kontrollpersonen wurden mit einem psychosozialen (Trier Soziale Stress Test, TSST) und einem hormonellen Stressor (Kombinierter Dexamethason-Corticoliberin Test, DEX-CRH) vor Aufnahme einer teilstationären kognitiven Verhaltenstherapie (für die Patienten mit Panikstörung) konfrontiert. Zusätzlich zu den Patienten mit Panikstörung wurden Stichproben von Patienten mit Primärdiagnose einer Post-traumatischen Belastungsstörung (PTBS) oder einer Major Depression mit dem TSST konfrontiert, um die hormonelle Stressreaktion direkt zwischen diesen Patientengruppen zu vergleichen. Blutentnahmen begleiteten die Tests zur Bestimmung der Adrenocorticotropin- (ACTH) und Kortisolkonzentrationen. Mit etablierten Fragebögen wurde die allgemeine Symptomschwere, das Ausmaß an agoraphoben Kognitionen, körperbezogenen Ängsten und an agoraphober Vermeidung sowie die Depressivität vor und nach der Psychotherapie bestimmt. Kovarianzanalysen mit Messwiederholung testeten die hormonelle Stressreaktion vor Beginn der Psychotherapie auf Haupteffekte für Zeit und Gruppe sowie auf Interaktionseffekte. Pearson-Korrelationsanalysen quantifizierten den Zusammenhang zwischen der hormonellen Stressreaktion und dem Therapieerfolg. Mithilfe dieser Methoden wurde untersucht, ob ein fehlender Kortisolanstieg auf den TSST (Non-Response < 55.2 nmol/l) und eine fehlende Kortisolsuppression auf den DEX-CRH Test (≥ 138 nmol/l) mit einem geringeren Therapieerfolg assoziiert sind. Ferner wurde getestet, ob Patienten mit Diagnose einer Panikstörung oder PTBS im Vergleich zu Patienten mit Major Depression und gesunden Kontrollprobanden eine Kortisol-Hyporeaktion zeigen, und ferner, ob Patienten mit einer Major Depression eine höhere Kortisolreaktion zeigen als die beiden Gruppen der Angstpatienten, aber niedrigere als die gesunden Kontrollprobanden.

Ergebnisse. Die statistischen Analysen ergaben eine hohe Effektstärke (η 2 ≥ 0.14) für die psychotherapeutische Behandlung (0.321 ≥ η 2 ≥ 0.648) und einen signifikanten Haupteffekt für die Zeit sowohl für Kortisol als auch ACTH in Reaktion auf den TSST und den DEX-CRH Test, unabhängig von der Studiengruppe. 42,9% der Patienten mit Panikstörung und 65,6% der gesunden Kontrollpersonen zeigten eine adäquate Kortisol-Stressreaktion (Response) auf den TSST, und 82,4% der Patienten mit Panikstörung und 85,3% der gesunden Kontrollpersonen zeigten eine DEX-CRH stimulierte Kortisolsuppression (deskriptiver Befund). Die Korrelationsanalysen zeigten eine signifikant inverse Assoziation der TSST-Kortisolreaktion vor Therapie mit dem Ausmaß an später berichtetem agoraphoben Vermeidungsverhalten sowie eine signifikante Assoziation der DEX-CRH-stimulierten Kortisolkonzentration vor Therapie mit dem späteren Ausmaß an agoraphoben Kognitionen zu Therapieende. Außerdem ergab sich ein signifikanter Zusammenhang zwischen dem subjektiven Angsterleben während des TSST und dem Ausmaß an körperbezogenen Ängsten sowie auf Trendlevel mit dem Ausmaß an agoraphoben Kognitionen zu Therapieende. Der direkte Vergleich zwischen den Patientengruppen ergab Response-Raten für den TSST von 26,7% bei der Panikstörung, 43,5% bei der PTBS, 72,2% bei der Major Depression und 80,6% bei den gesunden Kontrollpersonen. Die Kovarianzanalyse belegte eine Kortisol-Hyporeaktion bei Patienten mit Panikstörung und Patienten mit PTBS, wobei sich keine signifikanten Gruppenunterschiede in den ACTH-Konzentrationen zeigten. Patienten mit Major Depression unterschieden sich in ihrer hormonellen Stressreaktion weder von den Angstpatienten noch von den gesunden Kontrollpersonen.

Diskussion. Insgesamt zeigte sich, dass eine adäquate psycho-endokrine Stressreaktion auf einen psychosozialen (Kortisolanstieg) und einen hormonellen Stressor (Kortisolsuppression) vor der Therapie mit einem höheren Therapieerfolg assoziiert war. Höhere Kortisolkonzentrationen infolge des TSST waren mit einer Verbesserung agoraphoben Vermeidungsverhaltens korreliert und eine adäquate Kortisolsuppression infolge des DEX-CRH Tests war mit einer Verbesserung in agoraphoben Kognitionen nach einer fünf-wöchigen kognitiv-verhaltenstherapeutischen Behandlung korreliert. Ein höheres subjektives Angsterleben während des TSST war signifikant mit einer Verbesserung körperbezogener Ängste und deskriptiv mit einer Verbesserung in agoraphoben Kognitionen korreliert. Der direkte Vergleich der Kortisol-Stressreaktion zwischen den Patientengruppen brachte Evidenz für eine Dissoziation zwischen den Kortisol- und ACTH-Konzentrationen in Reaktion auf den TSST bei Patienten mit Panikstörung und PTBS. Die Ergebnisse stützen den relativ konsistenten Befund in der Literatur einer Kortisol-Hyporeaktion bei der Panikstörung. Patienten mit Major Depression zeigten nur deskriptiv die erwarteten Kortisolkonzentrationen im direkten Vergleich mit den anderen Studiengruppen.

Schlussfolgerungen. Zusammenfassend lässt sich vorsichtig festhalten, dass die Kortisol-Stressreaktion im TSST zwischen DSM-IV-Primärdiagnosen (Panikstörung und PTBS vs. Major Depression) und zwischen Psychotherapie-Respondern und Non-Respondern bei der Panikstörung mit und ohne Agoraphobie diskriminiert. Diese Ergebnisse implizieren die weitere Untersuchung der HPA-Achsen-Reaktivität und ihre Veränderung im Therapierverlauf als ein zusätzliches Instrument zur Beurteilung der Wirksamkeit psychotherapeutischer Interventionen in Ergänzung zu klinischen Selbst- und Fremdbeurteilungsmaßen. Die Vorhersage des Therapieerfolges aus einem endokrinen Korrelat könnte sich als eine Entscheidungshilfe für einen Behandlungswechsel erweisen, falls dessen Persistenz eine Non-Response auf die aktuelle Behandlung vorhersagt. Damit könnte die Fortsetzung einer nicht wirksamen Behandlung im Einzelfall vermieden werden. Die Ergebnisse implizieren zudem das Vorliegen eines transdiagnostischen endokrinen Korrelats bei der Panikstörung und PTBS. Klinische Implikationen sowie Implikationen für zukünftige Studien werden unter der Berücksichtigung einiger Limitationen diskutiert.

1 Introduction

1.1 Panic disorder: definitions and aetiology

Panic disorder (PD) is a serious mental disorder associated with high levels of disability and impairment to the quality of life of those affected (American Psychiatric Association, 2000; Wittchen & Jacobi, 2005). Epidemiologic data collected from multiple national studies documented 12-month- (0.7 – 3.1%) and life-time prevalence rates (1.6 – 5.2%), age at onset between 11 and 24 years and an approximately twofold higher risk in women (Bandelow & Michaelis, 2015). The most recent 12-month-prevalence of PD, with and without agoraphobia, in the German adult population was 2.0% (Jacobi et al., 2014).

Patients with PD experience recurrent unexpected panic attacks as well as persistent concerns about future panic attacks and their assumed implications to mental and somatic health (e.g. to die of a heart attack or stroke, going crazy) as well as significant changes in behaviour related to the panic attacks. PD patients frequently consult emergency departments or medical specialists for a variety of unspecific symptoms such as racing heartbeat, chest pain, feeling dizzy, and shortness of breath (Fleet, Marchand, Dupuis, Kaczorowski, & Beitman, 1998; Gerdes, Yates, & Clancy, 1995). For an adequate diagnosis, panic attacks are not (exclusively) due to the direct physiological effects of a drug or to a medical condition (e.g. hyperthyroidism, cardiopulmonary diseases), and are not explained more easily by the presence of another mental disorder (American Psychiatric Association, 2000). Health care providers must also determine whether agoraphobia is present, in order to provide a correct diagnostic evaluation according to the Diagnostic and Statistical Manual of Mental Disorders [4th ed., text rev.; DSM-IV-TR; American Psychiatric Association (2000)]. Patients with agoraphobia experience anxiety and seek behavioural or cognitive avoidance in regard to being in places from which escaping or receiving help might be embarrassing or difficult if a panic attack or panic-related complaints occur. Patients with agoraphobia usually develop more situationally predisposed or bound panic attacks and suffer less from spontaneous or unexpected attacks. Situationally predisposed or bound panic attacks are more likely to occur immediately upon exposure to a situational trigger.

Meanwhile, the 5th edition of the DSM [ DSM-V, American Psychiatric Association (2013)] has included several changes to the diagnostic criteria of PD and agoraphobia. Specifically, these changes encompass the separation of PD and agoraphobia into two distinct diagnoses, the inclusion of guidelines for distinguishing agoraphobia from specific phobia, the addition of a 6-month-duration requirement for an agoraphobia diagnosis, the inclusion of panic attacks as a specifier to any DSM-V diagnosis as well as the reduction of panic attacks into an expected and an unexpected type (Asmundson, Taylor, & Smits, 2014).

Our current understanding of the aetiology of PD refers to the genetic, (neuro-) biological correlates and cognitive-behavioural aspects from learning theory (In-Albon & Margraf, 2011; Pilecki, Arentoft, & McKay, 2011). Detailed reviews of the genetic basis in PD are provided elsewhere (e.g. Maron, Hettema, & Shlik, 2010; Schumacher et al., 2011). In brief, results from family and twin studies suggest a genetic contribution to PD development and estimate a heritability of approximately 40% (Hettema, Neale, & Kendler, 2001). A vast body of research has examined numerous candidate genes and polymorphisms for a possible contribution to the PD aetiology, however, it has produced overall inconsistent findings (Maron et al., 2010). The twofold higher prevalence rate of PD in women suggests sex-specific genetic variations to be associated with PD. For instance, a meta-analysis cautiously supports the catechol-O-methyltransferase val158met polymorphism, which codes for enzymes that control the availability of catecholamines such as dopamine, epinephrine, and norepinephrine (Axelrod & Tomchick, 1958), to predispose women for PD (Domschke, Deckert, O'Donovan, & Glatt, 2007). Overall, it seems that the genetic architecture differs with PD cases according to early or late onset, comorbidity, sex, and familial or non-familial transmission (for review: Schumacher et al., 2011).

Several emotion-generating regions of the midbrain (Duval, Javanbakht, & Liberzon, 2015; Lueken et al., 2014), neurotransmitters, and metabolic alterations were found to have an impact on the development of PD, e.g. elevated lactate levels (Riske, Thomas, Baker, & Dursun, 2017), an altered glucose metabolism in the amygdala, the hypothalamus and the thalamus among others (Martin, Ressler, Binder, & Nemeroff, 2009), and an altered cortisol secretion (e.g. Abelson, Khan, Liberzon, & Young, 2007; Graeff, Garcia-Leal, Del-Ben, & Guimarães, 2005; Petrowski, Herold, Joraschky, Wittchen, & Kirschbaum, 2010). However, many biological correlates are inconclusive, and it has been uncertain until now whether they represent preceding conditions or are a result of PD (Pilecki et al., 2011).

Currently, biological vulnerabilities as well as early (adverse) attachment representations or childhood experiences are suggested rather to contribute to the development of an anxiety sensitivity than to be causally related to the development of PD (Pilecki et al., 2011). Anxiety sensitivity is defined as the global tendency to misattribute benign physical sensations to a medical emergency having immediate adverse health, mental, or social consequences (Reiss, Peterson, Gursky, & McNally, 1986). Previous research showed anxiety sensitivity to be a robust predictor of PD (McNally, 1990, 1994) and panic symptom severity (Olatunji & Wolitzky-Taylor, 2009), of the frequency of panic attacks (Schmidt, Keough, Timpano, & Richey, 2008) and of fearful responses to panic-like symptoms (Rapee, Brown, Antony, & Barlow, 1992). With regard to Clark’s cognitive theory of PD (D. M. Clark, 1986), natural autonomic activities are misinterpreted as signals of severe health damage leading to increased fear which further escalates into the more intense “normal” somatic sensations of anxiety, creating a reciprocating circle. Research documented the impact of parental models in the fearful appraisal of and dysfunctional coping with “normal” anxiety symptoms (Bouton, Mineka, & Barlow, 2001; Schneider, Unnewehr, Florin, & Margraf, 2002). Beck, Emery, and Greenberg (2005) added that panic-disorder patients usually overestimate the likelihood of danger. Particularly, cardiac sensations are of special concern since they involve our pivotal organ and mimic sensations of life-threatening cardiovascular diseases. Cardiac symptoms are usually among the initial sensations triggering the symptom cascade of a full-symptom panic attack (American Psychiatric Association, 2000), thus causing a subjectively perceived vital threat to individuals who believe that anxiety may cause imminent heart failure or other health damage. Thus, individuals high in anxiety sensitivity may overreact to both internally and externally perceived threats and experience an initial panic attack as life-threatening. It is plausible that vulnerable individuals link panic attacks to specific cues and places via associative learning mechanisms and begin to avoid these cues and places due to associating them with their panic attacks (Pilecki et al., 2011). Such cognitive and/or behavioural avoidance behaviours are reinforced via a short-term reduction in anxiety; however, they maintain the disorder in the medium and long run. Accordingly, cognitive behavioural psychotherapy (CBT) provides techniques to disconfirm these maladaptive beliefs and to tolerate panic-related cues and benign somatic sensations (Lang, Helbig-Lang, Westphal, Gloster, & Wittchen, 2011), thusly increasing mobility in patients with a restricted life-style.

To complement, psychodynamic conceptualizations of the aetiology of PD also exist with a focus on the central childhood themes of separation and anger (Milrod et al., 2008). Psychodynamic treatment for PD includes the exploration of these common unconscious themes/conflicts which are approached through transference (Milrod et al., 2008). Overall, only sporadic reports and a very limited evidence base exist for the account and therapeutic effects of the psychodynamic theories of PD. And the aetiology conditions which are assumed to cause the development of PD are very commonly associated with the development of other anxiety or affective disorders and not uniquely related to the PD aetiology (Pilecki et al., 2011).

1.2 Treatment options for panic disorder and therapy success

The fiscal costs of full- and even sub-threshold PD were found to be the highest compared to other mental disorders (Batelaan et al., 2007). Nevertheless, treatment for panic-disordered patients is available with proven efficacy as summarised in the current German (Bandelow, Hau, & Beutel, 2014) and American practice guidelines for the treatment of PD (American Psychiatric Association, 2009). Considering the patients’ preferences and consent, their treatment history and co-occurring mental or somatic conditions, health-care professionals must provide their patients primarily with psychotherapeutic treatment. This is strongly supported by substantial clinical confidence and absolutely recommended (American Psychiatric Association, 2009; Bandelow et al., 2014). The strongest support is available for CBT, which should be provided to the patient initially (American Psychiatric Association, 2009; Bandelow et al., 2014). No significant efficacy differences were observed between group or individual CBT (Sharp, Power, & Swanson, 2004). Panic-focused psychodynamic psychotherapy may be indicated as initial treatment whenever it is the patient’s preference (American Psychiatric Association, 2009; Bandelow et al., 2014).

[...]