Page 1 of 9
Cognitive—Behavioral Therapy Helps Prevent Relapse and Recurrence of Panic
Disorder Following Alprazolam Discontinuation
A Long-Term Follow-Up of the Peoria and Dartmouth Studies
Timothy J. Bruce
Department of Psychiatry and Behavioral Medicine University of Illinois College of Medicine at Peoria
David A. Spiegel
Center for Anxiety and Related Disorders Boston University
Mark T. Hegel
Behavioral Medicine Section Dartmouth—Hitchcock Medical Center
The present research evaluated patients from 2 previous studies (1 conducted in Peoria, the
other at Dartmouth) during a 2- to 5-year posttreatment period. Results showed that 75% of
the Peoria sample and 76% of the Dartmouth sample were able to discontinue alprazolam
therapy, remain abstinent of any type of treatment for panic disorder, and maintain their acute-
treatment clinical gains over this follow-up period. The degree to which patients' anxiety
sensitivity declined during treatment predicted relapse versus survival during the 1st 6 months of
follow-up, when most relapses occurred. Implications of these findings for benzodiazepine
discontinuation, combined pharmacotherapy and psychotherapy, and relapse prevention in
panic disorder are discussed.
This study was supported in part by grants from the National Institutes of Health, Division of Research
Resources, and by a grant from the Upjohn Company.
We extend thanks to Shellee Abraham for assisting in the preparation of this article; Wendy Bayles-Dazet
for her assistance in collecting data at the Dartmouth site; and Carrie Campen, Shelley Gregg, Angela
Nuzzarello, and Connie Romane for their assistance in delivering the protocols.
Correspondence may be addressed to Timothy J. Bruce, Department of Psychiatry and Behavioral
Medicine, University of Illinois College of Medicine at Peoria, 5407 North University Avenue, Suite C,
Peoria, Illinois, 61614.
Electronic mail may be sent to email@example.com
Received: December 16, 1997
Revised: March 9, 1998
Accepted: July 30, 1998
Relapse of panic disorder after discontinuation of alprazolam therapy is a common clinical problem
( Pecknold, Swinson, Kuch, & Lewis, 1988 ). In previous studies conducted independently at the University
of Illinois College of Medicine at Peoria ( Bruce, Spiegel, Gregg, & Nuzzarello, 1995 ; Spiegel, Bruce,
Gregg, & Nuzzarello, 1994 ) and at the Dartmouth Medical School ( Hegel, Ravaris, & Ahles, 1994 ), the
ability of cognitive—behavioral therapy (CBT) for panic disorder to facilitate alprazolam discontinuation and
Journal of Consulting and Clinical Psychology
February 1999 Vol. 67, No. 1, 151-156
© 1999 by the American Psychological Association
For personal use only--not for distribution.
Page 2 of 9
short-term abstinence was investigated. In both studies, patients were treated initially with alprazolam at
doses sufficient to suppress panic attacks (most arrived using benzodiazepines at inadequate doses), after
which CBT was initiated and the drug was gradually tapered and discontinued. At posttreatment, 9 of 10
participants in the Peoria study who had received CBT and 20 of 25 participants in the Dartmouth study had
successfully discontinued their medication. At the last reported follow-up assessment (6 months for the
Peoria study and 12 months for the Dartmouth study), all 9 Peoria study participants and 17 of the 20
Dartmouth study participants who had stopped alprazolam had remained abstinent. We report here the
results of a subsequent 2- to 5-year follow-up of those participants.
Details of the treatment procedures summarized below can be found in the above-cited original reports.
Except as noted below, the studies used similar protocols. In both, participants met criteria based on the
revised third edition of the Diagnostic and Statistical Manual of Mental Disorders ( American Psychiatric
Association, 1987 ) for a principal diagnosis of panic disorder with or without agoraphobia, which was
assessed with established structured interviews. The Peoria study additionally required a 6-month minimum
duration of panic disorder. Individuals were excluded from either study if they had marked agoraphobic
avoidance (defined as a score of more than 3.6 on the Mobility Inventory for Agoraphobia Alone subscale;
Chambless, Caputo, Jasin, Gracely, & Williams, 1985 ), a current major mood disorder or obsessive—
compulsive disorder, a recent history of substance abuse or dependency, or a history of psychosis.
Prospective participants not taking psychotropic medication or taking medications other than alprazolam
were started on or switched to alprazolam, respectively, the dose of which was then advanced until the
following criterion was met: in the Peoria study, no unexpected panic attacks for 4 weeks; in the Dartmouth
study, no panic attacks of any kind for 2 weeks. CBT was then begun, administered to participants
individually in 12 weekly or biweekly sessions on the basis of procedures described by Barlow and Craske
(1989 ; Craske & Barlow, 1990 ). Alprazolam taper was initiated during CBT and timed to conclude before
CBT ended. An open-label, decreasing decrement schedule was used whereby dose reductions (made
approximately weekly) became progressively smaller as the total daily dose approached zero.
Although the CBT and drug protocols were similar in the two studies, there were differences. The Peoria
study was conducted as a randomized controlled trial in which participants were assigned either to the
alprazolam plus 12-session CBT protocol described above (ACBT) or to a comparison treatment in which
alprazolam was discontinued in conjunction with supportive medical management (ASMM) without CBT. In
the Dartmouth study, all participants received the ACBT treatment, which was delivered openly in the
context of routine clinical practice. In addition, the Dartmouth protocol included maintenance CBT sessions
at 2 weeks and at 3, 6, and 12 months posttreatment.
Twenty-five patients were enrolled in the Dartmouth study, of whom 22 completed treatment. In the Peoria
study, 21 patients were enrolled, of whom 10 completed treatment in each arm. Patient demographics were
similar between studies, except that the average duration of panic disorder (12 years vs. 1 year) and the
mean duration of benzodiazepine use before taper (20.8 months vs. 11.4 months) were longer in the Peoria
study, possibly reflecting the inclusion criterion in the Peoria study that patients have their panic disorder for
at least 6 months. The mean alprazolam dose before taper was similar between sites (1.7 and 1.9 mg per
day in the Peoria and Dartmouth studies, respectively).
As previously reported ( Spiegel et al., 1994 ), 4 participants in the ASMM group in the Peoria study were
able to discontinue alprazolam through 6-month follow-up. Of the 6 who could not, 2 subsequently were
Page 3 of 9
crossed over and attempted drug taper again in conjunction with CBT. In this report, we have combined
those 2 with the 10 participants who completed treatment in the original ACBT group, bringing the total in
the ACBT arm to 12. Those 12 participants, and the 4 participants remaining in the ASMM arm at 6-month
follow-up, were reassessed an average of 38.3 months posttaper (range = 26—50 months, SD = 7. 7
months). Of the 22 Darmouth study participants who completed treatment, 1 died of unrelated causes before
long-term follow-up and 4 others could not be reached. The remaining 17 participants were reassessed an
average of 42.9 months posttaper (range = 31—62 months, SD = 11.3 months).
The primary outcome variable was survivor status. To be classified as a survivor, participants had to
complete all dose-taper steps within the time allowed by the protocol and remain off alprazolam and out of
any other type of treatment for panic disorder throughout the entire follow-up period.
Because a patient can be abstinent of treatment but impaired clinically, assessments of clinical status were
conducted at baseline, 2 weeks posttaper, 3 months posttaper (the last follow-up reported in the Peoria
study), 6 and 12 months posttaper (the last follow-up reported in the Dartmouth study), and at long-term
follow-up. All of the Peoria study participants and 13 of the 17 Dartmouth study participants completed
these clinical assessments. Measures taken included panic attacks (four or more symptoms) and limited
symptom attacks during the 2 weeks before a scheduled assessment. They were assessed with daily diaries
in the Dartmouth study and with the Sheehan Panic and Anticipatory Anxiety Scale ( Sheehan, 1982 ) in the
Peoria study. We assessed agoraphobic avoidance ( Chambless et al., 1985 ), agoraphobic cognition
( Chambless, Caputo, Bright, & Gallagher, 1984 ), and depressive symptoms ( Beck, Ward, Mendelson,
Mock, & Erbaugh, 1961 ) using the referenced scales. In addition, clinician ratings of disability in work,
social, and family settings were made on the Sheehan Disability Scale ( Sheehan, 1983 ) to obtain a mean
score across settings. Clinician ratings were made in the Peoria study by David A. Spiegel, who was
unaware of participants' group assignment, and in the Dartmouth study by Mark T. Hegel, who knew of
We evaluated data from the clinical measures using composite definitions of clinical improvement to
determine the proportion of patients who attained different levels of end-state functioning at long-term
follow-up. To be evaluated for end-state functioning, a participant first had to be classified as a survivor by
virtue of remaining abstinent of alprazolam and out of any other treatment for panic disorder throughout the
follow-up period. Participants meeting those criteria were evaluated for attainment of six increasingly
stringent definitions of high end-state functioning (HESF), the first three of which represented cross-sectional
assessments of functioning and the latter three longitudinal assessments of functioning. At the lowest level,
patients had to be panic free (i.e., free of full, four-symptom panic attacks) for a period of 2 weeks before
the long-term follow-up assessment. A second definition required patients to be panic free and have a
Sheehan Disability Scale score of 0 (indicating "no disability" due to panic disorder or agoraphobia across
work, family, and social settings) for the same 2-week period. The third, more stringent definition required
patients to meet both of the previous criteria and also report no limited symptom attacks during the same
period. Three further (longitudinal) definitions required patients to meet each of the above (2-week) criteria
throughout the entire follow-up period.
Survival analysis was used to compare survival curves between the Peoria ASMM treatment group and the
ACBT treatment groups (combined) from 2 weeks posttaper through the long-term follow-up period. In
addition, the percentages of patients meeting each level of HESF were calculated. The clinical significance of
ACBT survivors' end-state status was evaluated by normative comparisons ( Kendall & Grove, 1988 ).
Survivors' scores on relevant outcome measures were compared with those of normal controls (found in the
original reports of the respective measures) to assess the degree to which survivors' scores fell within
Page 4 of 9
normative ranges defined as less than 1 standard deviation above the pertinent control group's mean score.
Survivors' outcome also was compared with the diagnostic criteria for panic disorder to assess the
percentage of survivors who no longer met those criteria at long-term follow-up.
Because relapse after alprazolam discontinuation often occurs shortly thereafter ( Marks et al., 1993 ;
Spiegel et al., 1994 ), predictors of short-term relapse (within 6 months of drug discontinuation) and long-
term relapse (relapse after 6 months) were explored separately. Forward stepwise logistic regression
analyses ( Hosmer & Lemeshow, 1989 ), using an entry criterion of p < .05, examined whether selected
variables would correctly classify those who relapsed versus those who survived within the short- and long-
term relapse periods as defined above. Patients from both sites who had attempted taper from alprazolam
with or without CBT and who had completed all study assessments through long-term follow-up ( n = 33)
Five variables were selected for the analyses. Two measures reflected patients' pretaper exposure to
benzodiazepines: the pretaper alprazolam dose established during stabilization and the total duration of
continuous benzodiazepine use before taper. Two other measures reflected the severity of patients' panic
disorders at baseline: scores on the Alone subscale of the Mobility Inventory for Agoraphobia ( Chambless
et al., 1985 ), as a measure of pretreatment severity of agoraphobia, and mean frequencies of panic and
limited symptoms attacks at baseline assessment. The fifth variable selected was the change in patients'
scores on the Anxiety Sensitivity Index ( Reiss, Peterson, Gursky & McNally, 1986 ) from baseline to 2
weeks posttaper on the basis of the following general rationale: From CBT perspectives (e.g., Barlow,
1988 ; Clark, 1986 ), physical sensations–including withdrawal symptoms and those related to situational
stressors–serve as phobic cues to those with panic disorder and precipitate the fear and avoidance that lead
to relapse and recurrence. To the degree that patients' fearful beliefs about those physical sensations (an
aspect of anxiety sensitivity) abate during treatment, patients should be less vulnerable to deterioration and
the need to return to treatment.
Figure 1 shows the cumulative number of patients in each group who completed alprazolam taper on
schedule and remained a survivor through follow-up. In the Peoria study, survival was defined as no further
treatment of any kind for panic disorder after completion of the 12-session CBT. Nine (75%) of the 12
ACBT patients and 3 (30%) of the 10 ASMM patients met that criterion at long-term follow-up. In the
Dartmouth study, survival was defined as no further treatment for panic disorder beyond the maintenance
CBT sessions (described above) that were given during the 1st posttaper year. Thirteen (76%) of the 17
patients met that criterion at long-term follow-up. The survival distributions of the ACBT (from both sites)
and ASMM groups differed significantly (Lee-Desu statistic = 7.2, df = 1, p < .007).
Table 1 shows the proportion of ACBT survivors who met the various definitions of HESF. In the Peoria
study, all 9 survivors (100%) reported no full panic attacks across the entire follow-up period, and 3 (33%)
met the most stringent criteria of no panic or limited symptom attacks and no disability during the entire
follow-up period. In the Dartmouth study, 9 of the 13 survivors for whom clinical data were available were
assessed for HESF. All 9 survivors (100%) reported no full panic attacks throughout the entire follow-up
period. Three (33%) of those 9 patients met the most stringent criteria of reporting no panic or limited
symptom attacks and no disability during the entire follow-up period.
All survivors from the Peoria and Dartmouth studies attained levels of end-state function that would not meet
Page 5 of 9
the criteria for panic disorder by virtue of being panic free throughout follow-up. As shown in Table 2 ,
survivors' mean scores on clinical outcome measures fell within the normative range for each measure, except
for their score on agoraphobic avoidance, which was exactly at the 1st standard deviation above the mean.
In the analysis of survival versus short-term relapse, only one variable was retained that significantly
predicted drug discontinuation success: the baseline-to-posttaper reduction in Anxiety Sensitivity Index
score, β (1) = 0.21, p < .02, R = .30. This univariate model correctly classified 79% of patient outcomes.
The model chi-square was significant at the p < .006 level, χ 2 (1, N = 33) = 7.8. To test whether this finding
could have been due to covariance between Anxiety Sensitivity Index change scores and another
demographic or symptom variable(s), we computed correlations between the anxiety sensitivity change score
and the baseline and baseline-to-posttaper change scores for all measures. No significant correlations
In the analysis of survival versus recurrence after 6 months, none of the variables used were retained by the
The follow-up data presented here extend previous reports of the Peoria and Dartmouth studies in
suggesting that adjunctive CBT reduces relapse of panic disorder associated with pharmacotherapy
discontinuation and its recurrence over the longer term. The 2- to 5-year survival rates were nearly identical
(76% and 75%) between sites, as were the proportions of survivors who met the various definitions of
HESF at follow-up. The similarity of outcomes in independent samples and at different treatment sites
speaks favorably to the transportability of the ACBT protocol.
The present results differ slightly but complement those of a similar study that used CBT to facilitate
benzodiazepine discontinuation ( Otto et al., 1993 ). Otto et al. tapered benzodiazepines from patients with
panic disorder either with or without the assistance of group CBT. The CBT significantly facilitated
discontinuation in that 13 of 17 (76%) of the CBT-assisted group tapered off the drug on schedule
compared with only 4 of 16 (25%) of the unassisted group. By 3 months posttaper, the overall
discontinuance and abstinence rate of the CBT-assisted group was 59% (10 of 17 patients), lower than the
75% and 76% long-term rates of the present studies. This difference between related studies appears to be
due to higher rates of discontinuation failures in the Otto et al. study rather than to differences in rates of
relapse after discontinuation. Although the 23% relapse rate reported by Otto et al. is comparable with the
16% rate of the present study, the 48% rate of discontinuation failure (75% of the unassisted group) is higher
than the 13% rate of the present study, lowering overall survival rates and precluding the opportunity to
detect the relative relapse preventative benefit of CBT. This difference in discontinuation rates may have
been due to the use of a taper schedule in the Otto et al. study that was more rapid than that used in the
present study. Previous studies using similar schedules (e.g., Noyes, Garvey, Cook, & Suelzer, 1991 ;
Schweizer, Rickels, Case, & Greenblatt, 1990 ) have also yielded higher rates of discontinuation failures. Of
course, other differences between protocols (e.g., our use of an individual treatment format and a panic-free
status before medication discontinuation) or samples may have contributed to these differences (cf. Spiegel
et al., 1994 ).
This distinction between discontinuation and relapse after discontinuation is important in part because it can
be argued that although CBT can significantly facilitate benzodiazepine discontinuation, it may not be
necessary to accomplish it if, for example, the taper schedule is slow and flexible, a panic-free status
Page 6 of 9
precedes taper, and supportive medical management is provided. The 80% discontinuation rate of the Peoria
ASMM group, which received these conditions, supports that conclusion. However, although those
conditions facilitated discontinuation, they were not sufficient to prevent patients from relapsing after it,
whereas adjunctive CBT was.
Survivors in the present study attained levels of HESF that are comparable with those of patients treated
with CBT alone reported by Brown and Barlow (1995 ). For example, in the Brown and Barlow study,
75% of persons treated with CBT were panic free for at least a month at 2-year follow-up. This rate is
nearly identical to the panic-free rates found at follow-up in the ACBT intent-to-treat samples in the present
study (viz., 75% and 76%). The proportions of patients who were classified at long-term follow-up as
meeting the most stringent (longitudinal) definition of HESF are similar as well. In the Brown and Barlow
study, 21% of participants achieved HESF defined as the combination of (a) a panic disorder global severity
rating of not more than "slightly disturbing" at both 3-month and 24-month follow-up assessments, (b) no
panic attacks for a month before the 3-month assessment and for a year before the 24-month assessment,
and (c) no further treatment for panic disorder during the entire follow-up period. Similarly, in the present
study, 25% of the ACBT intent-to-treat participants (33% of all ACBT drug-discontinuation survivors) met
HESF when HESF was defined as the combination of no disability due to panic disorder, no panic or limited
symptom attacks, and no further treatment for panic disorder throughout the entire follow-up period.
This similarity in outcome suggests that initial treatment with low-dose (e.g., 1—2 mg per day) alprazolam
(or perhaps equivalents) does not detract from the outcome of CBT if, in part, pharmacotherapy is tapered
slowly and during CBT. This contrasts with results from the Marks et al. (1993 ) study, which found that
exposure therapy used during higher dose alprazolam therapy (and stopped before taper) offered no more
protection from relapse after discontinuation than alprazolam plus a nonspecific control. Of course, other
considerations may be involved as well (cf. Spiegel & Bruce, 1997 ). The favorable outcome of the ACBT
treatment should not be taken as evidence supporting the efficacy of combined treatment for panic disorder,
however. It remains uncertain, for example, whether this outcome reflects a combined treatment effect or the
singular effect of CBT. Furthermore, it remains unestablished that adding benzodiazepines to CBT is
beneficial, either generally or for specific patients. Their primary potential benefit–earlier relief from
debilitating anxiety or panic–has not been shown reliably in relevant empirical studies and may bring with it
some overriding risks ( Spiegel & Bruce, 1997 ). Whether a treatment combining benzodiazepines with CBT
(as presently described) provides any acute- or posttreatment benefit over other treatments, such as
antidepressant therapy, CBT, or their combination, remains unstudied. The present results are encouraging,
however, for those patients on benzodiazepines who wish to discontinue them and for their treatment
Of the 33 patients in the present study representing all treatment groups (i.e., those who completed all study
assessments), 42% could not discontinue alprazolam or relapsed after discontinuing it. Of those who
relapsed, approximately two thirds did so within 6 months following discontinuation. The change in patients'
scores on the Anxiety Sensitivity Index from baseline to posttaper correctly classified 79% of patients' 6-
month survival outcomes across treatments in both samples. That outcome is predicted by cognitive ( Clark,
1986 ) and cognitive—behavioral ( Barlow, 1988 ) formulations of panic disorder that emphasize the role of
anxiety-related sensations as phobic cues. This outcome is the same as the one found by Bruce et al. (1995 )
in a predictive analysis of the original Peoria sample and does not represent an independent replication of it,
because the same Peoria sample constituted nearly two thirds of the present sample and most of those
patients who relapsed did so within 6 months. The present result does expand those original results in finding
that the predictive strength of anxiety sensitivity was not diluted by the inclusion of the Dartmouth sample and
strengthens the conclusion that reducing anxiety sensitivity during discontinuation therapy is important to a
Page 7 of 9
patient's success. Interestingly, later recurrence of panic disorder was not predicted by the variables
examined, suggesting that it may depend on other factors.
Two limitations of the present study warrant comment. First, sample sizes in the present study were smaller
than desirable, limiting generalizability and the power of the analyses. The similarity of results at independent
sites, and to those of similar protocols tested in other studies (e.g., Otto et al., 1993 ), supports their
credibility. Second, the present results cannot be generalized to patients with more than moderate levels of
agoraphobia because such patients were excluded. Within the range of agoraphobia studied, avoidance did
not predict outcome, but in other studies, greater levels of agoraphobia have been associated with higher
relapse rates (e.g., Fava, Zielezny, Savron, & Grandi, 1995 ; Williams & Falbo, 1996 ).
Because benzodiazepines offer prompt relief of anxiety and are better tolerated by some patients than
antidepressants, they are likely to remain in use for panic disorder. One report found that as many as half of
patients presenting to anxiety disorder clinics for psychological treatment arrived on benzodiazepines
( Sanderson & Wetzler, 1993 ). A concern for treatment providers has been how to manage these
medications within exposure-based therapies, given the risks they pose to posttreatment recovery. The
present finding that integrated CBT can mitigate those risks is encouraging for patients and practitioners and
supports further controlled investigation of its mechanism(s) and of the comparative risks and benefits of this
treatment approach with others.
American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders
((3rded., rev.). Washington, DC: Author.)
Barlow, D. H. (1988). Anxiety and its disorders: The nature and treatment of anxiety and panic. (New
Barlow, D. H. & Craske, M. G. (1989). Mastery of your anxiety and panic. (Albany, NY: Graywind)
Beck, A. T., Ward, C. H., Mendelson, M., Mock, J. & Erbaugh, J. (1961). An inventory for measuring
depression. Archives of General Psychiatry, 41, 561-571.
Brown, T. A. & Barlow, D. H. (1995). Long-term outcome in cognitive—behavioral treatment of panic
disorder: Clinical predictors and alternative strategies for assessment. Journal ofConsulting and Clinical
Psychology, 63, 754-765.
Bruce, T. J., Spiegel, D. A., Gregg, S. F. & Nuzzarello, A. (1995). Predictors of alprazolam discontinuation
with and without cognitive behavior therapy in panic disorder. American Journal of Psychiatry, 152,
Chambless, D. L., Caputo, G. C., Bright, P. & Gallagher, R. (1984). Assessment of fear in agoraphobics:
The Body Sensations Questionnaire and the Agoraphobic Cognitions Questionnaire. Journal of Consulting
and Clinical Psychology, 52, 1090-1097.
Chambless, D. L., Caputo, G. C., Jasin, S. E., Gracely, E. J. & Williams, C. (1985). The Mobility Inventory
for Agoraphobia. Behaviour Research and Therapy, 23, 35-44.
Clark, D. M. (1986). A cognitive approach to panic. Behaviour Research and Therapy, 24, 461-470.
Craske, M. G. & Barlow, D. H. (1990). Leaders guide for mastery of your anxiety and panic.
Fava, G., Zielezny, M., Savron, G. & Grandi, S. (1995). Long-term effects of behavioural treatment for
panic disorder with agoraphobia. British Journal of Psychiatry, 166, 87-92.
Hegel, M. T., Ravaris, C. I. & Ahles, T. A. (1994). Combined cognitive—behavioral and time-limited
Page 8 of 9
alprazolam treatment of panic disorder. Behavior Therapy, 25, 183-195.
Hosmer, D. W. & Lemeshow, S. (1989). Applied logistic regression. (New York: Wiley)
Kendall, P. K. & Grove, W. M. (1988). Normative comparisons in therapy outcome. Behavioral
Assessment, 10, 147-158.
Marks, I. M., Swinson, R. P., Basoglu, M., Kuch, K., Noshirvani, H., O'Sullivan, G., Lelliott, P. T., Kirby,
M., McNammee, G., Sengun, S. & Wickwire, K. (1993). Alprazolam and exposure alone and combined in
panic disorder with agoraphobia: A controlled study in London and Toronto. British Journal ofPsychiatry,
Noyes, R., Garvey, M. J., Cook, B. & Suelzer, M. (1991). Controlled discontinuation of benzodiazepine
treatment for patients with panic disorder. American Journal of Psychiatry, 148, 517-523.
Otto, M. W., Pollack, M. H., Sachs, G. S., Reiter, S. R., Meltzer-Brody, S. & Rosenbaum, J. F. (1993).
Discontinuation of benzodiazepine treatment: Efficacy of cognitive—behavioral therapy for patients with
panic disorder. American Journal of Psychiatry, 150, 1485-1490.
Pecknold, J. C., Swinson, R. P., Kuch, K. & Lewis, C. P. (1988). Alprazolam in panic disorder and
agoraphobia, results from a multicenter trial: III. Discontinuation effects. Archives of General Psychiatry,
Reiss, S., Peterson, R., Gursky, D. M. & McNally, R. J. (1986). Anxiety sensitivity, anxiety frequency, and
the prediction of fearfulness. Behaviour Research and Therapy, 24, 1-8.
Sanderson, W. C. & Wetzler, S. (1993). Observations on the cognitive behavioral treatment of panic
disorder: Impact of benzodiazepines. Psychotherapy, 30, 125-132.
Schweizer, E., Rickels, K., Case, W. G. & Greenblatt, D. J. (1990). Long-term therapeutic use of
benzodiazepines: II. Effects of gradual taper. Archives of General Psychiatry, 47, 908-915.
Sheehan, D. V. (1982). Sheehan Panic and Anticipatory Anxiety Scale.(In P. A. Keller & L. G. Ritt (Eds.),
Innovations in clinical practice: A sourcebook (Vol. 6, pp. 34—37). Sarasota, FL: Professional
Sheehan, D. V. (1983). The anxiety disease. (New York: Bantam Books)
Spiegel, D. A. & Bruce, T. J. (1997). Benzodiazepines and exposure-based cognitive behavior therapies for
panic disorder: Conclusions from combined treatment trials. American Journal of Psychiatry, 154, 773-
Spiegel, D. A., Bruce, T. J., Gregg, S. F. & Nuzzarello, A. (1994). Does cognitive behavior therapy assist
slow-taper alprazolam discontinuation in panic disorder? AmericanJournal of Psychiatry, 151, 876-881.
Williams, S. L. & Falbo, J. (1996). Cognitive and performance-based treatments for panic attacks in people
with varying degrees of agoraphobic disability. Behaviour Research andTherapy, 34, 253-264.
Table 1. End-State Functioning of Peoria and Dartmouth Survivors
Table 2. Normative Comparisons of Clinical Outcome of ACBT Survivors Across the Peoria and
Dartmouth Studies at Long-Term Follow-Up
Page 9 of 9
Figure 1. Percentage of patients maintaining alprazolam discontinuation and treatment abstinence across
follow-up in groups treated with alprazolam plus supportive medical management in the Peoria study
(ASMM) or alprazolam plus cognitive—behavioral therapy in the Dartmouth (ACBT-D) and Peoria