The effects of comorbid personality disorders on cognitive behavioral treatment for panic disorder

Abstract

The present study investigated the influence of personality pathology assessed both dimensionally and categorically on acute clinical response to group cognitive-behavioral treatment in a large sample of panic disorder patients (N = 173) meeting DSMIII-R criteria for panic disorder with or without agoraphobia. Nearly one-third of the sample met for one or more personality disorders, with the majority meeting for a Cluster C diagnosis. Patients with one or more comorbid personality disorders displayed higher baseline and higher post treatment scores across multiple indices of panic disorder severity compared to those without personality disorders. After controlling for panic disorder severity at baseline, the presence of both Cluster C and Cluster A Pers-Ds predicted a poorer outcome, whereas when assessed dimensionally, only Cluster C symptoms predicted a poorer treatment response. However, the influence of personality pathology was modest relative to that of baseline panic disorder severity.

Introduction

Panic disorder is frequently complicated by the presence of both psychiatric comorbidity (Brown et al., 1995) and non-psychiatric medical comorbidity (Schmidt et al., 1996). In terms of psychiatric comorbidity, as many as 70% of patients with panic disorder present with co-occurring psychiatric diagnoses (Reich and Troughton, 1988). The high rate of comorbidity in panic disorder has naturally led to evaluation of the effects of co-occurring conditions on treatment outcome.

Cognitive-behavioral treatment (CBT) has established efficacy in the treatment of panic disorder (Barlow et al., 2000, Clark et al., 1994, Gould et al., 1995, Hofmann, 2008). However, following CBT, many patients continue to display residual symptoms requiring some to seek out additional treatment (Brown and Barlow, 1995). Consequently, identifying factors that predict a poor response to CBT is an important research goal for optimizing the clinical management of panic disorder (Wolfe and Maser, 1994).

Personality disorder comorbidity is frequently cited as a factor implicated in poor treatment response to both pharmacotherapy (Slaap and den Boer, 2001) and psychosocial treatments (Milrod et al., 2007, Reich and Green, 1991, Reich and Vasile, 1993). Although not studied systematically, personality dysfunction may negatively affect treatment outcome through its potential influence on other moderators of treatment outcome such as patient drop-out (Grilo, Money, Barlow, Goddard, Gorman, Hofmann et al, 1998), compliance with treatment regimens (Schmidt and Woolaway-Bickel, 2000), the therapeutic alliance, or motivation for treatment (Persons et al., 1988).

The presence of a comorbid personality disorder as measured by either structured interview (i.e., SCID II) or questionnaire has been shown to predict treatment non-response (Marchesi et al., 2006, Noyes et al., 1990, Reich, 1988) or relapse upon medication discontinuation (Green and Curtis, 1988). Despite the claim that patients displaying comorbid Axis II pathology respond less favorably to cognitive-behavioral treatment (Mennin and Heimberg, 2000), evidence from controlled prospective studies is inconclusive. This is due to the small number of prospective studies and the significant methodological limitations of the existing studies i.e., small sample size, use of questionnaires to assess personality dysfunction, and failure to control for baseline severity of Axis I pathology (Dreessen and Arntz, 1998, Shear et al., 1994).

Of the prospective studies investigating the linkage between Pers-D comorbidity and treatment response in panic disorder, three published studies have investigated the effects of Pers-D pathology (as assessed by structured clinical interview) on panic patients’ response to cognitive-behavioral therapy (Black et al., 1994, Dreessen et al., 1994, Kampman et al., 2008). Several additional studies have examined response to naturalistic treatment e.g. (Mellman et al., 1992, Noyes et al., 1990), which may have involved some cognitive-behavioral treatment.

In a sample of 31 panic patients undergoing individual cognitive-behavioral treatment, Dreessen et al. (1994) found some evidence to suggest that patients with a comorbid personality disorder, based on the SCID II, were more likely to show greater psychopathology at baseline. Neither the presence of a personality disorder nor the number of SCID II personality traits predicted treatment response. However, the negative findings may have been due to low statistical power given the small sample size.

Black and colleagues (Black et al., 1994) examined the influence of personality pathology on the short-term (i.e., 3 weeks) treatment response in 66 panic disorder patients receiving cognitive therapy, fluvoxamine, or placebo. The presence of a comorbid personality disorder based on the SIDP structured interview did not predict recovery as defined by the absence of panic attacks and a Clinical Global Improvement score of ‘very much’ or ‘much’ improved. However, higher scores on a self-report personality disorder questionnaire were associated with a less favorable outcome (Black et al., 1994). Unfortunately, due to the small sample size, separate analyses were not reported for the patients receiving cognitive therapy. Kampman and colleagues examined whether cluster C personality disorders predicted treatment response in a sample of 161 panic disorder patients treated with 15 sessions of CBT. Although initial panic severity predicted posttreatment severity, the presence of one or more cluster C personality disorders did not affect treatment outcome (Kampman et al., 2008). The principal limitation of this study is its exclusive focus on cluster C personality disorders.

Based on the meager evidence to date, no firm conclusions can be drawn concerning the effects of Pers-D comorbidity on panic disorder patients’ clinical response to cognitive-behavioral treatment. The primary aim of this study was to examine several indicators of Pers-D comorbidity and their relationship to treatment outcome among a large sample of panic disorder patients receiving cognitive-behavioral treatment. Several design features were included to address the limitations of previous studies. These included: (a) inclusion of a larger sample, (b) followed recommended guidelines for the assessment of panic disorder (Shear and Maser, 1994) (c) followed the evidence-based recommendations for the assessment of personality disorders (Widiger and Samuel, 2005); (d) utilized both categorical and dimensional analyses of Pers-Ds; and (e) controlled for pretreatment levels of panic disorder severity.