Lars-Goran O¨st
, Ulrika Thulin, Jonas Ramnero
vey; Magee, Eaton, Wittchen, McGonagle, & Kessler, 1996), and 6.1% in Norway (Kringlen,
ing author.
ess: ost@psylished.

Correspond
E-mail addr0005-7967/$ - see front ma
doi:10.1016/j.brat.2003.07.chology.su.se (L.-G. O¨st).tter # 2003 Elsevier Ltd. All rights reserved.
004Torgersen, & Cramer, 2001). So far no study using DSM-IV (APA, 1994) criteria has been pub-¨ ¨
Department of Psychology, Stockholm University, S-106 91 Stockholm, Sweden
Received 15 November 2002; received in revised form 21 July 2003; accepted 23 July 2003
Abstract
Seventy-three psychiatric outpatients with DSM-IV diagnosis of panic disorder with agoraphobia were
assessed with a battery of independent assessor, self-observation, self-report and behavioral measures
before and after therapy, and at a 1-yr follow-up. They were randomly assigned to Exposure in vivo (E;
n ¼ 25), Cognitive Behavior Therapy (CBT; n ¼ 26), or a Wait-list control (WLC; n ¼ 22) and received
12–16 individual therapy sessions, once weekly. The treatments yielded significant improvements, both on
panic/agoraphobia measures and on measures of general anxiety, depression, social adjustment and qual-
ity of life, which were maintained at follow-up. However, there were no significant differences between E
and CBT. The three criteria of clinically significant improvement were achieved by 67% of the E-patients
and 79% of the CBT-patients at post-treatment, and 74% and 76%, respectively, at follow-up. The con-
clusion that can be drawn is that adding cognitive therapy to exposure did not yield significantly better
results than for exposure alone.
# 2003 Elsevier Ltd. All rights reserved.
1. Introduction
Panic disorder with agoraphobia (PDA) is a prevalent and one of the most handicapping
anxiety disorders (Barlow, 2002). Recent epidemiological studies using DSM-III-R criteria
(APA, 1987) found similar lifetime prevalence figures; 6.7% in USA (National Comorbidity Sur-Behaviour Research and Therapy 42 (2004) 1105–1127
www.elsevier.com/locate/brat
Cognitive behavior therapy vs exposure in vivo in the
treatment of panic disorder with agrophobia

L.-G. O¨st et al. / Behaviour Research and Therapy 42 (2004) 1105–11271106Behavior therapy researchers have focussed on developing and evaluating treatment methods
for agoraphobia since the mid 1960s and the first randomized clinical trial (RCT) of a behav-
ioral treatment for any anxiety disorder concerned agoraphobia (Gelder & Marks, 1966) com-
paring behavior therapy and psychodynamic therapy. Since then at least 90 RCTs have been
published (O¨st, 2002), making agoraphobia the most well-researched of the anxiety disorders, at
least from a cognitive behavioral standpoint.
A number of meta-analyses concerning CBT and pharmacological treatments of PDA have
been published. Mattick, Andrews, Hadzi-Pavlovic and Christensen (1990) conducted a meta-
analysis on 54 studies published between 1973 and 1988. They found that exposure in vivo had
a marked effect on phobia but a moderate effect on panic, general anxiety, and depression.
Exposure in vivo in combination with cognitive anxiety management had less effect on phobia
than exposure in vivo alone, but some better effect on panic, general anxiety, and depression.
Cox, Endler, Lee and Swinson (1992) included 34 studies published between 1980 and 1990 in
their meta-analysis, comparing three treatments: exposure in vivo, imipramine, and alprazolam.
No significant differences were found between the active treatments, but exposure in vivo had
the most consistent effect, i.e. consistently large effect sizes on the variables measured. Van
Balkom et al. (1997) conducted a meta-analysis on 106 studies published between 1964 and
1995. Treatment outcome was measured with four clinical variables: panic, agoraphobia, gen-
eral anxiety, and depression. Control groups, consisting of waiting list, pill-placebo, or atten-
tion-placebo, were included in their analysis. They found that exposure in vivo in combination
with antidepressants had a better effect than the other treatment conditions. The combination of
exposure in vivo with antidepressants had better effects on anxiety and depression than
exposure in vivo alone or in combination with panic management. The authors conclude that
the combination of antidepressants and exposure in vivo has shown to be the treatment-of-
choice for PDA. Bakker, van Balkom, Spinhoven, Blaauw, and van Dyck (1998) conducted a
meta-analysis on the long-term efficacy of different treatments for PDA and included 68 studies.
The follow-up period varied between 4 weeks and 8 years, and in general the treatment effects
were maintained at follow-up for all treatment conditions. On the measures of agoraphobia
there were significant differences found between post-treatment assessment and follow-up, show-
ing the superiority of the combination of antidepressants with exposure in vivo over exposure in
vivo alone, panic management alone, and the combination of the two. The tentative conclusion
that can be drawn from these meta-analyses, when it comes to psychological treatments, is that
there is strong empirical evidence that treatment interventions including exposure in vivo are to
be preferred over treatment interventions without exposure in vivo for PDA. However, there is
still much room for further development of CBT methods for PDA since only 60% of the
patients treated in RCTs published since 1990 have achieved a clinically significant improvement
(O¨st, 2002).
In panic disorder without agoraphobia (PD) cognitive therapy (CT) has received strong
empirical support from studies carried out in various countries (e.g. USA, UK, Sweden, Hol-
land, and Germany; see Clark, 1999, for a review). The percentage of patients who are free
from panic attacks after an average of 12 treatment sessions varies from 74 to 94% with a mean
of 84%, and the effects are maintained at the 1 yr follow-up (71–100%, mean 86%). This raises
the question of the possibility to use CT in combination with exposure in vivo, or as a stand-
alone treatment, in order to increase the treatment efficacy for PDA.