Research report
A meta-analysis of the efficacy of psycho- and pharmacotherapy in panic disorder with and without agoraphobia

https://doi.org/10.1016/j.jad.2005.05.003Get rights and content

Abstract

The efficacy of (cognitive) behavioural ((C)BT) and pharmacological therapy was investigated using meta-analytic techniques. After a comprehensive review of the literature, the results of 124 studies were included. (C)BT was more effective than a no-treatment control and a placebo control. No difference of efficacy was found when using cognitive elements compared to not using them for anxiety; for associated depressive symptoms, additional cognitive elements seems superior. Pharmacotherapy was more effective than a placebo control; there was no superiority of any drug class. Sample size was related to effect size in pharmacotherapy and publication bias was found. (C)BT was at least as effective as pharmacotherapy and depending on type of analysis even significantly more effective. There were no significant differences between (C)BT alone and a combination approach but characteristics of studies have to be considered.

Introduction

Panic disorder with our without agoraphobia and agoraphobia without history of panic disorder belong to the most investigated anxiety disorders, attracting a relatively large amount of interest when researching both pharmacological and psychosocial therapy approaches. These disorders are common and disabling. They tend to be chronic and show only low remission rates in their natural course (Keller et al., 1994, Yonkers et al., 1998). People with panic disorder1 are, compared with healthy controls, at a greater risk for a significant impairment in quality of life and in the social and economic environment (e.g., Hansson, 2002, Lochner et al., 2003, Mendlowicz and Stein, 2000). Several studies have found that patients with panic disorder are among the highest users of medical services (Katon, 1996, Klerman et al., 1991, Siegel et al., 1990); therefore, this is an enormous economic burden for the health care system. Kessler et al. (1994) reported a lifetime rate of 3% for panic disorder with or without agoraphobia and a lifetime rate of 5% for agoraphobia without a history of panic disorder. However, agoraphobia without a history of panic disorder seems to be a controversial category. According to White and Barlow (2002), the prevalence rates differed depending on the type of study: in community studies, higher rates were found than in clinical reports.

The most widely used treatments for panic disorder are (cognitive) behavioural therapies [(C)BT] and pharmacological therapies. Concerning (C)BT, historically, emphasis was made on the treatment of agoraphobia rather than on panic (e.g., Gelder and Marks, 1966, investigating the efficacy of behaviour therapy). In the 1980s, treatments have been developed and found support in the research for reducing panic and associated anxiety (e.g., Clark, 1986). Current perspectives include exposure-based procedures (with interoceptive elements for treating panic and exposure in vivo for treating agoraphobia) and cognitive restructuring as core elements. Furthermore, breathing retraining and relaxation are used in the treatment of panic disorder. The efficacy of exposure can be explained by (a) an activation and modification of relevant fear structures based on habituation and/or extinction and/or examining dysfunctional beliefs and as a result a change in outcome–expectations (e.g., Marks, 1989), (b) an adequate selection of behaviour and coping skills (e.g., Beck and Clark, 1996), (c) an increase in self-efficacy and a consecutive change in efficacy–expectations (Bandura, 1977), and (d) a more adequate attentional focus (Marks and Dar, 2000). Cognitive techniques lead to a direct modification of maladaptive beliefs (e.g., Clark, 1986); arguing along the lines of conditional models, this resembles a reappraisal of the UCS (Davey, 1992).

In contrast, pharmacological therapy is based on the assumption that panic disorder arises from disturbances in the neurobiological systems; hence, the aims of the treatment are regulations of neurotransmitter systems. Numerous studies focus on neurocircuitry and brain systems underlying panic with an emphasis on noradrenergic and serotonergic pathways. It is beyond the scope of the present article to offer a more detailed discussion of the various neurobiological models of panic disorder; see Gorman et al. (2000), Coplan and Lydiard (1998), or Goddard and Charney (1997) for details.

In this article, the techniques of meta-analysis were applied to investigate the efficacy of treatments in panic disorder. There were several published meta-analyses of the efficacy of psychosocial therapy, pharmacotherapy or both over the last decade. In Table 1, an overview is presented of previous meta-analyses, their methodological characteristics and their results. In summary, only three meta-analyses compared (C)BT and pharmacological treatments. However, a number of methodological problems decrease the validity of their results. So, none of these meta-analyses used the random effects-model, and therefore, their results cannot be generalized beyond the included studies. But relevant questions in research deal with the efficacy of therapy in general, and so the use of the random effects-model is necessary. Furthermore, the assumptions of statistical methods used were not tested which is crucial for interpreting the found effect sizes. Also, publication bias was often ignored.

Because a number of additional relevant studies investigating the efficacy of pharmacological or psychosocial therapy have been conducted, a new meta-analysis is necessary. In addition to expanding the findings to a total of 124 studies, in the present paper, several new developments in meta-analytical techniques were applied. To increase the external validity of the results, the random effects-model was used. Additionally, several sensitivity analyses were done (e.g., whether results were influenced by outliers or computational methods used) to evaluate the impact of methodological decisions made during research synthesis. A further aspect which influences the findings of meta-analyses is the so-called publication bias, meaning that an unknown number of unpublished studies exist, which are systematically different from the included studies (i.e., studies with non-significant results). To investigate the impact of publication bias on results, a trim and fill analysis (Duval and Tweedie, 2000) was used which allows the estimation of an effect size corrected for publication bias. Furthermore, to compare (C)BT with pharmacotherapy, a method was employed which allows the comparison of studies with different control groups (Mitte, in press).

Section snippets

Method

Studies were located by searching different databases (Medline and Psycinfo from the first available year to May 2002) using the keywords “panic,” “agoraphob*,” “treatment” and “*therapy.” In addition, a review of the literature was done with a manual search in important journals and secondary sources (e.g., previous meta-analyses). With a view to reducing file-drawer effects, some effort was undertaken to locate unpublished work by searching the Internet and by contacting researchers and

Efficacy of (cognitive) behaviour therapy

Two thirds of the trials comparing (C)BT and no-treatment control consisted of cognitive behavioural methods. The most frequently examined technique was exposure with or without cognitive elements. In the outlier analysis, 2.6% of effect sizes were excluded. Most individual effect sizes were positive, and only one study showed slightly better results for the no-treatment control group in measures of anxiety and depression.

Table 3 shows the average weighted effect sizes, 95% confidence

Discussion

The purpose of this meta-analysis was to evaluate the efficacy of (C)BT and pharmacotherapy in panic disorder. I found that (C)BT is an effective treatment for reducing both anxiety and associated depressive symptoms and for increasing quality of life. Superiority over a no-treatment control was found. Additionally, at least for anxiety and, concerning the included studies, for quality of life and depression, the efficacy of (C)BT exceeds common factor effects. Furthermore, the findings

Acknowledgements

This study was financed by a personal grant of the State of Thuringia (Germany) to the author. The study reports results from the dissertation of the author, completed under the supervision of Peter Noack and Martin Hautzinger. I thank Kersten Schäfer for developing the statistical program and all researchers for providing data.

Please note that references in the Further reading section indicate studies included in the meta-analysis.

References (53)

  • A. Bakker et al.

    SSRIs vs. TCas in the treatment of panic disorder: a meta-analysis

    Acta Psychiatrica Scandinavica

    (2002)
  • A. Bandura

    Self-efficacy: toward a unifying theory of behavioral change

    Psychological Review

    (1977)
  • M. Basoglu et al.

    Double-blindness procedures, rater blindness, and ratings of outcome. Observations from a controlled trial

    Archives of General Psychiatry

    (1997)
  • A.T. Beck et al.

    An information processing model of anxiety: automatic and strategic processes

    Behaviour Research and Therapy

    (1996)
  • B.J. Becker

    Synthesizing standardized mean-change measures

    British Journal of Mathematical & Statistical Psychology

    (1988)
  • C.B. Begg

    Publication Bias

  • W. Boyer

    Serotonin uptake inhibitors are superior to imipramine and alprazolam in alleviating panic attacks: a meta-analysis

    International Clinical Psychopharmacology

    (1995)
  • T.A. Brown et al.

    Classification of anxiety and mood disorder

  • D.L. Chambless et al.

    Cognitive therapy of anxiety disorders

    Journal of Consulting and Clinical Psychology

    (1993)
  • G.A. Clum

    Coping with Panic

    (1990)
  • G.A. Clum et al.

    A meta-analysis of treatments for panic disorder

    Journal of Consulting and Clinical Psychology

    (1993)
  • M.G. Craske

    An integrated treatment approach to panic disorder

    Bulletin of the Menninger Clinic

    (1996)
  • S. Duval et al.

    Trim and fill: a simple funnel plot based method of testing and adjusting for publication bias in meta-analysis

    Biometrics

    (2000)
  • M.G. Gelder et al.

    Severe agoraphobia: a controlled prospective trial of behavior therapy

    British Journal of Psychiatry

    (1966)
  • L.J. Gleser et al.

    Stochastically dependent effect sizes

  • A.W. Goddard et al.

    Toward an integrated neurobiology of panic disorder

    Journal of Clinical Psychiatry

    (1997)
  • Cited by (267)

    • Disorder-specific internet-based cognitive-behavioral therapy in treating panic disorder, comorbid symptoms and improving quality of life: A meta-analytic evaluation of randomized controlled trials

      2021, Internet Interventions
      Citation Excerpt :

      It is characterized by recurrent and unexpected panic attacks, resistance to spontaneous remission and high comorbidity with other mental health disorders (APA, 2013, Baillie and Rapee, 2005, Goodwin and Gotlib, 2004, Goodwin and Hamilton, 2001, Goodwin et al., 2004, Mattis and Ollendick, 2001). PD is associated with great discomfort in professional and social life (Mitte, 2005; Tsao et al., 2005), which leads, as a result, to significant deterioration in general quality of life (Rangé et al., 2011). Kessler et al. (2006) report 45.0% comorbidity with other anxiety disorders, with highest comorbidity rates for specific phobia (21.0%) and social phobia (18.8%).

    View all citing articles on Scopus
    View full text