326 TREATMENT OF SPECIFIC ANXIETY DISORDERS
or a 3-month wait list control followed by random assignment to one of the active treat-
ments (D. M. Clark et al., 1994). At posttreatment (i.e., 3 months), cognitive therapy
was significantly more effective than applied relaxation and imipramine in reduction of
panic symptoms (i.e., panic composite score), agoraphobic avoidance, misinterpretation
of bodily sensations, and hypervigiliance for body symptoms. In addition 80% of the
cognitive therapy patients reached high-end functioning at 3 months compared to 25%
for applied relaxation and 40% for imipramine. Moreover, at 15-month follow-up cog-
nitive therapy remained superior to applied relaxation and imipramine on six measures
of panic/anxiety, with 85% of cognitive therapy patients still panic-free compared with
47% of applied relaxation and 60% of imipramine patients.
In a large multisite randomized placebo- controlled clinical trial involving 77 patients
with panic disorder (Barlow, Gorman, Shear, & Woods, 2000) intent-to-treat analyses
revealed that CBT and imipramine were superior to placebo, but there were no sig-
nificant differences between imipramine and CBT at posttreatment, although there was
a trend favoring CBT at 6-month follow-up. Overall, then, major treatment outcome
studies have clearly established that CBT for panic disorder is at least as effective as
medication, although there is little advantage in combining CBT with pharmacotherapy.
Comparisons of CBT with applied relaxation (i.e., Öst & Westling, 1995) indicate that
CBT is probably more effective for panic disorder (Siev & Chambless, 2007).
Outcome studies indicate that CBT can be effective for more difficult cases of panic
disorder. CBT can produce enduring treatment effects even with comorbid diagno-
ses, with significant improvement evident in both panic and comorbid symptoms (e.g.,
Craske et al., 2007; Tsao, Mystkowski, Zucker, & Craske, 2005). In fact Craske and
colleagues found more generalized symptom improvement in panic- focused CBT than
in a condition in which therapists were allowed to stray onto issues related to the comor-
bid condition. CBT has also been shown to be effective in drug- refractory individuals
with panic disorder (Heldt et al., 2006) and in reducing both day and nighttime panic
symptoms in patients with nocturnal panic attacks (Craske et al., 2005). Finally, brief
versions of CBT (e.g., intensive 2-day intervention), as well as computerized adapta-
tions, can be highly effective for panic disorder (D. M. Clark et al., 1999; Deacon &
Abramowitz, 2006b; Kenardy et al., 2003). Although these findings are preliminary,
they do suggest that more efficient and cost- effective cognitive interventions may be
available for panic disorder.
CBT Process Studies
Exposure is an important component of cognitive therapy for panic disorder, especially
when agoraphobic avoidance is prominent. Given our emphasis on cognitive interven-
tion, how critical is cognitive restructuring to the effectiveness of CBT for panic disor-
der? In their meta- analysis, Gould et al. (1995) found that cognitive restructuring plus
interoceptive exposure (i.e., symptom induction or schema activation) yielded the largest
effect sizes, but cognitive restructuring alone produced highly variable results. In an
early study Margraf and Schneider (1991) found cognitive restructuring without expo-
sure as effective as pure exposure or combined exposure plus cognitive restructuring.
In a series of multiple baseline single cases Salkovskis et al. (1991) found that two
sessions of cognitive restructuring focused on evidence gathering for and against the