Behavioral, Cognitive, and Pharmacological Treatments of Panic ...

Copyright 1491 by the Arr

Journal of Consulting and Clin 1991. Vol 59, No I. 100-114

:an Psychological Association, Inc. 0022-006X/9I/J300

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Behavioral, Cognitive, and Pharmacological Treatments of Panic Disorder With Agoraphobia: Critique and Synthesis Larry K. Michelson

Karen Marchione

Stress and Anxiety Disorders Institute Pennsylvania State University

Department of Psychiatry, School of Medicine Western Psychiatric Institute and Clinic University of Pittsburgh

Theoretical, methodologic, and clinical research issues pertaining to these treatments are examined as are their strengths, limitations, and possible interactions. Attrition, outcome, and maintenance effects are compared. Composite indices of clinically significant improvement, endstate functioning, and longitudinal adjustment are presented. The article also highlights emerging models, theoretical advances, and promising interventions. Advantages and limitations of current treatments are discussed, with recommendations for future research. It is concluded that significant advances have been made in the conceptualization and treatment of panic disorder with agoraphobia.

Anxiety disorders are the most prevalent clinical dysfunctions in the United States (Myers et al., 1984; Weissman & Merikangas, 1986). Panic disorder, with or without agoraphobia, exhibits high incidence and prevalence and is associated with increased risk of psychiatric morbidity in the form of depression; suicide; alcohol/substance abuse; minor tranquilizer addiction; generalized anxiety disorder; social, simple, and sexual phobias; and impaired social, marital, and vocational functioning (Markowitz, Weissman, Ouellette, Lish & Klerman, 1989). Although panic disorders represent the leading cause for clients seeking both outpatient services and psychotropic medications, fewer than 25% of these individuals seek out treatment (Weissman & Merikangas. 1986). The prevalence, severity, and longterm sequelae of panic disorders emphasize the import of developing safe, tolerable, and effective treatments. The primary aim of this article is to review behavioral, cognitive, and pharmacologic treatments of panic disorder with agoraphobia (PDA). Overarching theoretical, methodologic, and clinical research issues pertaining to these treatments and their underlying models are examined, as are their relative strengths, limitations, and possible synergisms. For reasons of feasibility, the review focuses on PDA because it is prototypic of anxiety disorders and has been the subject of most of the controlled anxiety research. The review is meant to be representative, rather than exhaustive, and emphasis is placed on empirically based and methodologically controlled research. This review is organized broadly into three major sections, encompassing behavior therapy, cognitive therapy, and psychopharmacology.

Preparation of this article was supported in part by the National Institute of Mental Health. Correspondence concerning this article should be addressed to Larry K. Michelson, Pennsylvania State University, Department of Psychology, Stress and Anxiety Disorders Institute. 548 Moore Building, University Park, Pennsylvania 16802.

100

Behavior Therapy For purposes of organization, the domain "behavior therapy" includes interventions such as programmed practice instructions, imaginal exposure, graded in vivo exposure, flooding, and relaxation training.

Exposure Exposure-based modalities have earned a major role in the behavioral treatment of PDA (Barlow, 1988; Marks, 1987; Michelson, 1987; Rachman & Wilson, 1980). Exposure is a generic term used to describe a complex set of therapeutic procedures that consist of exposing PDA clients to anxiety producing stimuli (e.g., phobic situations). Historically, exposure was conceptualized as facilitating habituation and extinction, leading to the elimination of both phobic anxiety and avoidance. There are a number of dimensions of exposure-based therapy including imaginal versus in vivo, rapid versus gradual, therapist-assisted versus self-directed, brief versus prolonged, and individual versus group format. Contemporary findings, culled from diverse studies, generally indicate that therapist-assisted, prolonged, graduated in vivo exposure appears most effective. Although there are mixed reports, therapist-aided exposure appears to enhance the rapidity and level of significant improvement achieved (50-65%) compared with self-directed (i.e., programmed practice) exposure alone (25%) (Mavissakalian & Michelson, 1986a). Although therapist-assisted exposure appears to exert a therapeutic advantage at posttreatment, differential effects may dissipate over the course of extended follow-up, inasmuch as clients receiving self-directed exposure instructions may exhibit continued progress. Overall, exposure appears to be a beneficial, yet for many clients insufficient, treatment for PDA. Because exposure has been one of the most frequently used behavioral treatments for PDA, a review of meta-analyses of many of the existing studies is provided next.

SPECIAL SECTION: PANIC DISORDER WITH AGORAPHOBIA


Meta-Analysis of Exposure Treatments A meta-analysis of 19 studies of behavior therapy (primarily exposure-based modalities) for PDA by Trull, Nietzel, & Main (1988) found that it was relatively effective, using Jacobson, Follette, & Revenstorf 's (1984) clinical significance criterion of falling within two standard deviations of a normal population on the Fear Questionnaire. The average PDA client began at the 97.3 percentile, progressed to the 68.0 percentile at posttreatment, and the 65.5 percentile at follow-up. Several dimensions were found to be associated with clinically significant outcomes: (a) doctoral level therapists achieved better outcome than nondoctoral level therapists; (b) recruited clients (responding to protocol advertisements) achieved greater improvement than nonsolicited (self-referred) clients; (c) group treatment tended (p = .06) to be more beneficial than individual treatment; and (d) treatments with greater exposure emphasis and systematic programmed practice instructions fared better. Jacobson, Wilson, and Tupper (1988) analyzed exposurebased interventions from 11 studies to assess the proportion of PDA clients who improved at a clinically significant level using their improvement criteria. Composite scores, derived from multiple measures, indicated that 58% of the subjects achieved clinically significant improvement, although full recovery was limited to less than one third. Improvement was generally maintained at follow-up, with less than 10% of the clients relapsing. These findings are consistent with those of Trull et al. (1988) regarding the benefits and limitations of unitary exposure-based treatments for PDA. These reports suggest a clinically significant response1 to exposure in the 60-75% range. However, these figures are actually closer to 50% if computations also include attrition and nonresponders (Barlow & Wolfe, 1981). Attrition rates for exposure modalities average 15-20% (Clum, 1989). Approximately, 3040% of PDA clients who receive exposure therapy exhibit only moderate benefit. Of the remaining 60-70% who achieve clinically significant change, residual symptomatology in the form of mild anticipatory anxiety or avoidance is not uncommon. The long-term benefits of exposure-based treatments have been documented in numerous follow-up studies (Burns, Thorpe, & Cavallaro, 1986; Cohen, Monteiro, & Marks, 1984; Emmelkamp & Kuipers, 1979; Jansson, Jerremalm, & Ost, 1986; Jansson & Ost, 1982; Marks, 1971; Mavissakalian & Michelson, 1986b; Munby & Johnson, 1980). Hence although exposurebased treatments are not a panacea, their effects appear to be largely maintained over time. In vivo exposure is a useful and perhaps necessary treatment for PDA, yet there is reason to exercise caution in that many outcome studies surfer from design flaws. Direct comparisons among alternative treatments are hampered by methodological confounds such as marked variations in treatment duration, quality, intensity, integrity, diagnostic heterogeneity, absence of independent and multimodal assessments, use of dependent measures with unknown psychometric properties, inadequate description of treatments, and omission of attrition and end point analyses. Many PDA studies report only statistically significant results based on mean group score changes on subjective scales, and infrequently they include tripartite (behavioral,

101

cognitive, and psychophysiological) assessments. Few studies report the percentage of clients improved or unimproved, those achieving normative functioning, dropouts, relapsers, followup diagnostic status, or the percentage of clients requiring further treatment once the protocol is completed. Therefore, the 50% significant improvement rate must be regarded as a gross estimate. Nevertheless, it remains a sobering reminder that although exposure strategies are beneficial, they too have their limitations. Moreover, it appears that the therapeutic limits of exposure, as a singular treatment for PDA, are being rapidly approached. To summarize, there is strong empirical support that exposure is a relatively beneficial behavioral treatment for PDA. However, a significant proportion of clients do not appear to fully benefit from exposure, or its therapeutic variants, and the overall outcome results clearly highlight the importance of examining mediational, complementary, or alternative models and treatments for PDA, which will now be reviewed. Relaxation Training An alternative and/or complementary treatment of PDA is relaxation training, which addresses its psychophysiological dimensions. Progressive deep muscle relaxation training (PDMR) was developed by E. Jacobson (1929) and modified by Wolpe (1958) and Bernstein and Borkovec (1973). Reviews of PDMR with clinical populations in general and anxiety disorders in particular suggest that this treatment produces therapeutic change (Borkovec & Sides, 1979; Glaister, 1982; 6st, 1989; Ost, Hellstrom, & Westling, in press; Woolfolk & Lehrer, 1984). These positive findings and converging conceptual and clinical issues have led to studies examining its efficacy for PDA. First, spontaneous fluctuations in psychophysiological arousal are evident in PDA, such as panic attacks, which result in further phobic conditioning. Reductions in psychophysiological arousal and reactivity may promote habituation and exposure to phobic stimuli. From a theoretical perspective, relaxation training may also facilitate emotional processing (Borkovec & Sides, 1979; Rachman, 1980). Given Hugdahl's (1981) and Michelson's (1984,1986) postulations that different anxiety disorders may "load" differentially on the tripartite systems, agoraphobics, with their high levels of physiological arousal, might be considered ideal candidates for relaxation training. Relaxation Training Outcome Studies Michelson, Mavissakalian, & Marchione (1988) compared relaxation training (RT), graduated exposure (GE), and paradoxical intention (PI) in the treatmentof88 PDA subjects, with all subjects receiving programmed practice instructions. The three treatments affected significant improvements across all domains, with between-groups differences being found only in

1 "Clinically significant improvement" is operationally defined as therapeutic change from pretreatment to posttreatment that would meet the following criteria: (a) There would be marked clinical benefit, and (b) client would no longer fulfill diagnostic criteria for panic disorder with agoraphobia (moderate to severe impairment).


102

LARRY K. MICHELSON AND K A R E N MARCHIONE

the psychophysiological measures, GE and RT exhibiting less psychophysiological reactivity than PI at postlreatment. RT was equally as effective as the other treatments, with 46% of the subjects achieving high endstate functioning 2 at posttreatment and 53% at the follow-up. These positive findings notwithstanding, only half of the subjects achieved high endstate functioning at posttreatment or follow-up. Therefore, although RT appeared to exert comparable and relatively beneficial effects in comparison with graduated exposure or paradoxical intention, the absolute effectiveness of these treatments leaves considerable room for improvement. Ost (1987) developed a modified version of PDMR, termed applied relaxation (AR). The emphasis of AR is on the rapid development and deployment of relaxation coping skills to controvert in vivo anxiety and panic attacks. In Ost et al.'s (1987) study, PDA subjects received either AR or in vivo exposure for 12 sessions. AR was found to be equally effective as exposure at posttreatment and 15 month follow-up, with 65% of the subjects achieving clinically significant improvement. Ost et al. (in press) also compared AR, exposure, and cognitive techniques in the treatment of 45 PDA subjects. Corroborating their previous findings, AR was found to be as effective as exposure, with both treatments being superior to the "cognitive techniques" condition. The proportion of clinically improved subjects were 87%, 73%, and 53%, respectively. It should be noted that the cognitive techniques condition did not involve cognitive therapy proper but relied on more superficial cognitive maneuvers such as distraction and altering surface level self-talk. Therefore, although these results clearly support the relative effectiveness of AR, the study's design and the nature of the cognitive intervention preclude generalizations regarding the relative efficacy of cognitive versus relaxation therapies. Several conclusions can be derived from the preceding studies. First, AR, as described by Ost (1987) and his colleagues appears to be as effective as therapist-assisted in vivo exposure when combined with programmed practice. Second, prolonged acquisition and passive use of relaxation skills, as exemplified by PDMR, appear unnecessary and are less effective than the efficient acquisition and proactive use of powerful and rapid relaxation coping skills. Third, Ost (personal communication, January 15,1990) noted that AR engendered high client compliance, with only 6% attrition and minimal relaxation-induced anxiety phenomena. Fourth, despite the positive results achieved by AR, there are unresolved issues concerning its overall potency. For example, it is not clear what proportion of PDA clients achieve full recovery and maintain their gains over extended follow-up. Identifying which clients require additional treatment, such as therapist-assisted exposure, cognitive therapy, or pharmacotherapy, awaits further empirical study. Research examining the relative and combined effectiveness of these treatments would be of theoretical and clinical importance. Cognitive-Behavioral Treatments of PDA PDA, like all anxiety disorders, has ubiquitous cognitive dimensions, processes, and products. Maladaptive beliefs, dysfunctional schema, attributions, self-statements, and phobic im-

agery have been implicated as contributing to its pathogenesis. Some theorists view PDA as being composed of cognitive processes that require amelioration. Traditionally, behavioral interventions limited their focus to phobic anxiety and avoidance. However, it appears that sole reliance on exposure appears unwarranted given its clinical and theoretical limitations. For these reasons there has been a growing awareness of the potential contribution cognitive theories and therapies may offer in understanding and treating PDA. Cognitive-behavioral treatments (CBT) have received considerable research attention and have been used to treat major affective disorders, anxiety disorders, and the full spectrum of other clinical dysfunctions. Reviewers (Kendall & Hollon, 1979; Mahoney & Arnhoff, 1978; Rachman & Wilson, 1980) have concluded that clients treated with CBT did better than untreated control subjects, and CBT was equivalent or superior to other therapies. A meta-analytic review of CBT by Miller and Berman, (1983) indicated that clients receiving CBT show an average effect size of .83, compared with untreated control subjects. Other reviews of CBT have reported comparable positive findings (Barrios & Shigatomi, 1980; Dush, Hirt, & Schroeder, 1983). Inasmuch as only a few PDA studies were included in these reviews of CBT, a brief summary of some of these reports may be informative. Emmelkamp and Mersh (1982) compared cognitive restructuring, prolonged in vivo exposure, and a combination treatment. Cognitive restructuring involved eight 2-hour visits of relabeling cognitions and self-statement training (SST). At posttreatment, exposure and the combined procedure were superior to SST alone on phobic anxiety and avoidance measures. At the 1-month follow-up, however, differences between treatments were no longer apparent. The SST group continued to improve, whereas the exposure-alone clients experienced mild relapse. Emmelkamp, Brilman, Kuiper, & Mersch, (1986) randomly assigned 43 agoraphobics to either exposure in vivo, rational emotive therapy (RET), or SST for six sessions. The authors concluded that exposure was more effective than the cognitive strategies in reducing phobic anxiety and avoidance. Both SST and RET effected statistically significant improvements on most measures. However, the clinical improvements achieved were generally less impressive, particularly for RET, in which most clients were rated as failures. Conceptual, methodologic, and clinical limitations of these studies have been enumerated elsewhere (Michelson, 1987). The restricted nature and duration of the cognitive treatments may have been insufficient to allow clients to effectively alter core belief systems. Further studies report more encouraging findings using different cognitive modalities. Ascher (1980) examined the efficacy of paradoxical intention (PI), a cognitive strategy, in ameliorating travel restrictions in PDA clients. PI was found, in a

2 "Endstate functioning" (Michelson & Mavissakalian, 1985) is an operationalized composite index used to classify clients1 level of functioning at posttreatment and follow-up. High endstate functioning reflects excellent clinical status/outcome across measures of clinician and subject-rated severity, clinical ratings of phobic anxiety and avoidance, and in vivo behavioral performance.



multiple baseline design across two groups of clients, to produce significantly greater behavioral improvement than in vivo exposure. PI was also found to significantly enhance the effects of exposure treatment. Mavissakalian, Michelson, Greenwald, Kornblith, and Greenwald (1983) compared PI with SST in 26 PDA clients in a pilot Study. Subjects received 12 weekly 90-min sessions over a 3-month period, in addition to programmed practice instructions. PI and SST both exhibited significant gains across major assessment domains. PI, however, evinced greater improvement during treatment, with SST requiring 6 months to "catchup." These results suggested that cognitively based treatments (particularly PI), combined with programmed practice, were beneficial in treating PDA. Although promising, this study did not empirically test the relative short- and long-term efficacy of cognitive versus behavioral treatments. Michelson et al. (1988) compared paradoxical intention, relaxation training, and therapist-assisted graded exposure for 88 PDA subjects in a controlled study. All treatments evidenced significant within-group improvements, with few betweengroups differences. PI, the cognitively based modality, fared as well as the other treatments. Approximately 50% of these PDA clients achieved high endstate functioning. These positive results were encouraging, yet the overall efficacy of these treatments mandated continued conceptual and clinical refinements. These findings led to further study of alternative cognitive-behavioral treatments. Michelson, Marchione, and Greenwald (1989) conducted a comparative outcome study of PDA. Ninety-two marked-severe PDA clients entered, and 74 subjects completed the protocol. Treatment consisted of 16 group sessions, with all subjects receiving programmed practice instructions and therapist-assisted exposure for 1.5 hr per session. In addition, subjects received one additional hour of either cognitive therapy, relaxation training, or programmed practice discussion each session. The cognitive therapy condition included in-depth cognitive restructuring, identification and modification of core beliefs, misperceptions, and misattributions related to PDA. Clients learned to identify and test the validity of faulty beliefs, use of Socratic dialogue, reattribution training, cognitive problem-solving training, and generalization programming. An assessment battery consisting of clinical ratings of severity, phobia, anxiety, panic, depression, and tripartite measures of agoraphobia was administered at pretreatment, midtreatment, posttreatment, and at the 3-month follow-up. Significant within-group gains were found across almost all measures and treatments from pretreatment to posttreatment and follow-up. At posttreatment and follow-up, the cognitive therapy plus graded exposure (CT + GE) condition was found to be superior to the other treatments on many of the major dependent measures. Operationalized measures of high endstate functioning were computed at posttreatment and follow-up. At posttreatment, CT + GE exhibited a clinical advantage with 86.4% of these subjects achieving high endstate functioning versus 73% for relaxation training plus graded exposure (RT + GE) versus 65% for GE alone. Follow-up comparisons revealed that CT + GE = 87.5%, RT + GE = 47%, and GE alone = 65% achieving high endstate functioning. These findings highlight the effec-

103

tiveness of combining cognitive therapy with graduated exposure in the treatment of PDA. Synergistic effects were found for CT + GE, compared with the beneficial but relatively less effective RT + GE or GE conditions. Overall, CT + GE is the most effective treatment we have developed or tested to date at our center. This includes a decade of clinical research, during which time we have examined pharmacologic, behavioral, and various cognitive interventions. Although definitive conclusions regarding the long-term effects of the CT + GE modality must await completion of an ongoing follow-up study, these findings indicate that significant advances are being made in the treatment of PDA. In the following section a potential mediating feature of cognitive-behavioral therapies is explored both for heuristic purposes and to offer a conceptual framework for understanding differential effects across studies. Metacognitive Dimensions of PDA Historically, some of the difficulty in appraising the potential value of cognitive-behavioral treatments (CUT) for PDA arose from differences in the cognitive procedures under study. Schwartz (1982) noted that cognitions may be viewed as simple habits of thought with minimal organization and depth; as cognitive structures, plans, or strategies with considerable organization and depth; or as unconscious cognitions with considerable degrees of organization and depth. Dush et al. (1983) noted that reviewers of CBT studies made few distinctions among cognitive modalities and drew conclusions as if cognitive therapy, imagery techniques, and self-statement training were equivalent. Hollon and Kriss (1984) posited that cognitive treatments are, more or less, metacognitive in nature, and that different CBT have different loci of effect. Specifically, differences may exist among CBT in efficacy for modifying knowledge structures, cognitive processes, or cognitive products. Some theorists, such as Beck, Rush, Shaw, & Emery, (1979), Guidano and Liotti (1983), and Michelson, Marchione, & Greenwald (1989), suggest that cognitive procedures operating solely at the level of changing thoughts or self-statements compared with focusing on core beliefs or schemata, are targeting the most superficial level and the one least likely to be efficacious in treating complex anxiety disorders, such as PDA. Cognitive interventions that do not impact on core beliefs may be less effective for PDA with its complex metacognitive elements. Investigations in which low-metacognitive strategies are pitted against exposure (Emmelkamp, Brilman, Kuiper, & Mersch, 1986; Williams & Rappaport, 1983) report limited benefit associated with their addition regarding either behavioral performance or anxiety. However, these results may be a reflection of the cognitive techniques used, rather than of cognitive therapy per se. Where metacognitive dimensions of PDA are systematically addressed, therapeutic outcomes appear to be quite favorable (Clark & Salkovskis, 1990; Marchione, Michelson, Greenwald, & Dancu, 1987; Michelson, Marchione, & Greenwald, 1989; Sokol-Kessler, Beck, Greenberg, & Wright, 1989). Presently, this theoretical issue has not been empirically addressed. Future research will be needed to systematically examine the relative efficacy of CBT that differentially addresses

104

LARRY K, MICHELSON AND K A R E N MARCHIONE

metacognitive aspects of PDA. In the next section recent advances in cognitive-behavioral theories and therapies for PDA are explored. Although some of these approaches have been more extensively applied with panic disorder with no or mild agoraphobia, they have clear theoretical and clinical implications for more severe forms of PDA. Therefore, it is important to provide an overview of these emerging models and treatments to examine their conceptual and clinical saliency for PDA.


Cognitive Advances in PDA Since the inception of the Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev, DSM-III-R: American Psychiatric Association, 1987), panic is now viewed as the nucleus of agoraphobia. Converging epidemioiogic, basic, and clinical research provided the impetus for cognitively based theoretical and treatments advances. Previously, when behaviorally based modalities of PDA led to significant reductions in panic severity and frequency in about half of the clients, it was regarded as a derivative of exposure, because panic was only infrequently assessed, let alone directly treated. Recently, two major cognitivebehavioral formulations have been posited to explain the etiology, development, and treatment of panic disorders with clear implications for PDA. These include the cognitive model of panic (Clark, 1986; Clark & Salkovskis, 1987) and interoccptive conditioning (Griez & van den Hout. 1983,1986). The following section provides a brief summary of these models, their derived treatments, and relevant treatment studies. Alternative, yet complementary, theories in the treatment of panic disorder with or without agoraphobia implicate the role of catastrophic misappraisal of normal somatic cues. In a benchmark critique of existing theories and empirical evidence on panic, Margraf, Ehlers, & Roth (1986) proposed a psychophysiological model of panic. The paradigm entails a complex interaction of perceived control, available resources, internal and external panicogenic signals, and threat appraisals. This model is compatible with and is addressed by interventions that are directed at modification of danger appraisals, identification of panic causes, and reduction of psychophysiological reactivity. Theoretical refinements and extensions of this model have been offered in the cognitive model of panic (CMP) by Clark (1986); Clark, Salkovskis, & Chalkley (1985); and Beck and Emery (1985), who hypothesized that panic disorders result from misinterpretation of bodily sensations that are evaluated as frightening, dangerous, or lethal. Catastrophic appraisal results in powerful and rapid escalations in anxiety, triggering further arousal, and eventual panic. The CMP model has received growing empirical support from both experimental and clinical research. An important prediction of the theory is that panic disorders are amenable to psychological treatments through cognitive-behavioral interventions. The possible role of interoceptive conditioning in panic disorders has been raised as well. Accordingly, clinical strategies to remediate interoceptive conditioning to somatic cues has been incorporated into several cognitively based treatments (Barlow & Cerny, 1988; Barlow, Craske, Cerny, & Klosko, 1989; Clark et al., 1985; Clark,

1988; Michelson et al., 1990). The role of hyperventilation as a cause, correlate, and/or consequence of panic attacks has also been reviewed (Fried, 1987; Ley, 1985,1987; Lum, 1976) regarding potential interactions among these models, A concise overview of treatment literature flowing from these theories is provided to exemplify both their nature and outcome. Presently, there are 10 reports examining these treatments for panic disorder, with wide variations in subjects' levels of phobic avoidance. Seven are uncontrolled studies (Barlow et al., 1989; Clark et al., 1985; Gitlin et al., 1985; Klosko, Barlow, Tassinari, & Cerny, 1990; Michelson et al., 1990; Salkovskis, Jones, & Clark, 1986; Salkovskis, Warwick, Clark, & Wessel, 1986; Sokol-Kessler, Beck, Greenberg, & Wright, 1989). Despite their methodological limitations, they consistently report uniformly high levels of efficacy in eliminating panic attacks. The three remaining controlled studies (Barlow et al, 1989; Clark & Salkovskis, 1990; Klosko, Barlow, Tassinari, & Cerny, 1990) are reviewed next. Barlow et al. (1989) compared applied relaxation training (RT) exposure plus cognitive restructuring (E -I- C), a combined modality (E + C + RT) versus a waiting-list control (WL) condition for 60 clients with panic disorder (with no or mild agoraphobia). The protocol consisted of 15 weekly individual sessions. The E + C condition included cognitive restructuring, exposure to interoceptive cues, imaginal exposure, and breathing retraining. The E + C + RT condition also included applied relaxation training. The authors report that 36% of the WL, 60% of the RT, 85% of the E + C, and 87% of the E + C + RT subjects were panic-free at posttreatment (E + C + RT, E + C > WL). The percentages of clients achieving high endstate functioning at posttreatment were WL = 0%, RT = 50%, E+C= 46%, E + C + RT = 46% (RT, E + C, E + C + RT > WL). Analyses of available 6-month follow-up data revealed maintenance or further improvement for the active treatments. These results indicate that the three treatments were highly beneficial in eliminating panic attacks, with the E + C and E + C + RT being superior to both the RTand WL conditions when end-point analyses were used. Whereas 85-87% of the subjects in the E + C and E + C + RT conditions were panic-free at the completion of the protocol, only half achieved high endstate functioning. These findings suggest that these cognitive-behavioral treatments were efficacious in treating panic attacks. However, the presence of residual anxiety/phobic symptomatology prevented half of these clients from achieving full recovery. Klosko et al., (1990) recently compared cognitive-behavior therapy (panic control treatment) with alprazolam (Xanax) for 57 panic disorder clients with no or mild agoraphobia. Control conditions included placebo and waiting-list control groups. The cognitive-behavioral treatment (CBT) protocol entailed 15 weekly individual sessions of exposure to interoceptive cues, cognitive techniques, relaxation training, and breathing retraining. At posttreatment the percentages of clients who were panic-free were CBT = 87%, alprazolam = 50%, placebo = 36%, waiting list = 33%. The CBT condition was significantly superior to the placebo and waiting-list conditions on most of the major dependent variables. The alprazolam condition did not differ significantly from either the placebo or CBT conditions. Clark and Salkovskis (1990) reported the preliminary results of a benchmark outcome study of panic disorder. Treatment



conditions included cognitive therapy (CT; cognitive model of panic), applied relaxation training {AR), imipramine (IMF), and a waiting-list control (WL). Although additional data are being collected, analyses (from the majority of the completed sample) revealed several intriguing findings. Overall, the active conditions (CT, AR, 1 MI) were superior to the WL group. Analyses comparing the active treatments indicated the following: (a) misinterpretation of bodily symptoms (CT > AR, IMI); (b) panic attack, frequency (CT > AR > IMI); (c) panic attack, severity (CT> IMI, AR); (d) phobic avoidance (CT> IMI, AR); and (e) Beck Anxiety Inventory (CT > IMI, CT = AR). At posttreatment the percentages of subjects reporting being free of panic attacks were as follows: CT = 90%; IMI = 55%; AR = 52%, and WL = 10%. Attrition rates were as follows: CT = 5%, AR = 5%, and IMI = 23%. The Clark and Salkovskis study (1990) is significant in being the only comparative outcome study of panic disorder(PD)eontrasting cognitive therapy with the pharmacotherapy "gold standard," imipramine. While final conclusions must await complete data analysis, these findings are promising regarding the relative efficacy of the cognitive model of panic (CMP) treatment. These CBT studies evince consistently positive findings. Therapeutic gains appear to be substantive, with elimination of panic attacks and anticipatory anxiety in the vast majority of the clients. Although these results are highly encouraging, longterm follow-up studies are needed to ascertain the differential longitudinal adjustment of PDA subjects receiving CBT versus alternative psychosocial and pharmacologic treatments. For example, to what extent does CBT affect the panicogenic nature of stressful life events over a long-term follow-up? Do CBT clients exhibit superior maintenance effects and fewer clinical sequelae (i.e., agoraphobia, generalized anxiety disorder, use of psychotropic medications)? Importantly, there are no published controlled studies of panic management treatments with panic disorder clients who exhibit marked to severe phobic avoidance (agoraphobia). Hence, definitive recommendations regarding their efficacy for PDA must await controlled research. However, in the interim they appear to be highly promising modalities.

Pharmacotherapy of PDA Pharmacotherapy of PDA has received considerable attention in the psychiatric and medical literature. Several types of medications have been the focus of these reports, including tricyclic antidepressants, clomipramine, monoamine oxidase inhibitors, beta-adrenergic blockers, low-potency benzodiazepines, and high-potency benzodiazepines. The relative and combined effectiveness of these psychotropics and behavior therapy will also be delineated. Tricyclic Antidepressants Tricyclic antidepressants (TCAs), most notably imipramine hydrochloride, have been the most widely studied medications for PDA and have been found to exhibit beneficial effects on mood, anxiety, and panic symptoms, with more limited gains found on measures of phobic avoidance and anticipatory anxi-

105

ety. Traditionally, behavioral treatments, specifically exposurebased interventions, were found to exert significant decreases on phobic anxiety and avoidance, with more moderate reductions noted on panic frequency and intensity. Hence it is not surprising that the potential combination of behavioral and TCA treatments has generated much interest. Several controlled studies examining the relative and combined effectiveness of antidepressants and behavior therapy have been conducted, typically using imipramine. The rationale of these studies involves the search for possible additive or synergistic effects, with TCAs addressing mood and panic and behavior therapy focusing on phobic anxiety and avoidance. Behavioral treatments used in these studies included therapistassisted exposure, self-directed exposure, imaginal exposure, supportive therapy, relaxation training, and systematic desensitization. Early studies of imipramine versus behavior therapy were conducted by Zitrin, Klein, and colleagues (Zitrin, Klein, & Woerner. 1978, 1980; Zitrin, Klein, Woerner, & Ross, 1983). These authors compared 218 mixed phobics across three treatments: behavior therapy plus imipramine, behavior therapy plus placebo, and supportive therapy plus imipramine. The "behavior therapy" entailed imaginal systematic desensittzation, which is considered a relatively weak treatment for PDA. To address this criticism, Zitrin et al. (1980) later conducted a smaller study that compared exposure with or without imipramine with placebo control groups. Overall, the results were comparable across both studies, with imipramine enhancing the effects of either imaginal or in vivo exposure. These studies reported that in vivo exposure yielded a 70% (moderate to marked) improvement rate, which imipramine augmented. Although these studies served to stimulate further research, their design, assessment, treatment, and statistical limitations render them more of historical interest. Since these efforts, several well-controlled comparative outcome studies have been conducted examining the relative and combined effectiveness of imipramine and behavior therapy (Marks et al., 1983; Mavissakalian & Michelson, 1986a, 1986b; Telch, Agras, Taylor, Roth, & Gallen, 1985; Agras, Telch, Taylor, Roth, & Brouillard. 1990). Marks et al., (1983) assigned 45 chronic PDA clients to either (a) imipramine plus therapist-assisted exposure; (b) imipramine plus relaxation training; (c) therapist-assisted exposure plus placebo; or (d) relaxation training plus placebo. All clients received a programmed practice manual with self-directed exposure assignments. Overall, clients evinced significant improvements on almost all dependent measures and maintained their gains at I- and 2-year follow-ups (Cohen, Monteiro, & Marks, 1984). Approximately two thirds of the subjects remained improved or much improved. However, no between-groups differences or additive or synergistic effects were observed. In a reanalysis of the Marks et al. (1983) study, Raskin, Marks, and Sheehan (1983) reported that imipramine enhanced the effects of exposure on several dependent measures, although these additive effects disappeared at follow-up. Telch et al. (1985) compared imipramine plus antiexposure instructions, imipramine plus exposure, and placebo plus exposure for 37 PDA clients. The results revealed that at 8 weeks


106

LARRY K. MICHELSON AND KAREN MARCHIONE

the imipramine plus antiexposure condition exhibited negligible improvement on panic, phobic avoidance, and anxiety, although an effect on depressed mood was observed. Conversely, clients in the exposure conditions, irrespective of imipramine status, showed substantive improvement on these measures. Imipramine appeared to enhance in vivo exposure by the completion of the protocol. The authors concluded that their data failed to support the position that imipramine's mechanism of action was the blocking of panic attacks (Klein, 1967; Zitrin et at., 1978). Furthermore, although imipramine appeared to facilitate the effects of exposure, its use raised several issues. First, there was a 20% refusal rate among potential clients. In addition, high dropout rates (averaging 25-35%), absolute efficacy (60%) for those who received the medications, and high relapse rates (35-60%) suggested potential obstacles in the use of antidepressants in treating PDA. Mavissakalian and Michelson (1986a; 1986b) conducted a comparative outcome study of PDA, using a 2 X 2 factorial double-blind design, to examine the relative and combined efficacy of behavior therapy (therapist-assisted exposure vs. programmed practice) and pharmacotherapy (imipramine vs. placebo). Seventy-seven PDA clients were randomly assigned to one of four conditions: (a) exposure plus imipramine; (b) exposure plus placebo; (c) programmed practice plus imipramine; or (d) programmed practice plus placebo. Treatment consisted of 12 weekly 1.5-hour sessions for groups of 8 subjects. All were given an identical behavioral rationale and encouraged to practice self-directed exposure between sessions. Significant repeated measures effects were found for all treatments and measures. Superiority of imipramine versus placebo effects was found on global severity ratings and two phobia measures. No therapist versus self-directed exposure effects were found, and only one imipramine plus exposure effect was noted on the Fear Questionnaire—Agoraphobia subscale. Using the composite measure of high endstate functioning, the following percentages were found at posttreatment: exposure plus imipramine, 58%; exposure plus placebo, 46%; programmed practice plus imipramine, 59%; programmed practice plus placebo, 36%. These data indicate that exposure, imipramine, or a combination of the two was superior to programmed practice alone. Over the 2-year follow-up, reversals among imipraminetreated clients were more evident, whereas behaviorally treated clients tended to exhibit greater maintenance and further improvement. A number of process issues were also addressed in the study. No significant dose-response relationships between tricyclic plasma concentrations and panic were found. Similar to the Marks et al. (1983) study, panic was equally responsive to all treatments. In addition, the refusal and attrition rates and the side effects of imipramine (which precluded optimal dosing in 43% of this sample) were high. Finally, imipramine-treated clients also exhibited significant increases in psychophysiological reactivity, probably due to tachycardia, a potential side effect of the TCAs (Michelson & Mavissakalian, 1985). Limitationsof this study include the use of a large group format; absence of cognitive assessment; and lack of blind, independent assessors. In a recent study investigating the separate and combined effects of imipramine and exposure therapy, Agras et al., (1990) randomly assigned 100 PDA clients to one of several conditions.

During the initial 8-week phase, imipramine or placebo was combined with exposure or antiexposure instructions. The exposure instruction condition was found to be superior to the antiexposure group at the end of this phase. The only significant imipramine effect was on mood. The next 8-week phase continued the original medication status, with half of the clients receiving intensive group exposure and half receiving programmed practice. The group exposure involved eight 3-hr sessions compressed over a 2-week period. Over the course of another 8-week phase, subjects continued their medication and programmed practice but received no therapist-aided exposure. Advantages were found in favor of group exposure, compared with programmed practice at Weeks 16 and 24 on agoraphobia measures. Significant imipramine effects, compared with placebo, were observed for depression and on a behavioral avoidance measure. Imipramine reduced situational panic when paired with programmed practice but not with group exposure. In termsof global improvement, an advantage was noted for the imipramine plus programmed practice condition at the 24week assessment. However, this effect may be due to the fact that imipramine subjects received 6 months of pharmacotherapy, compared with only 2 weeks of in vivo exposure. At the 6-month phase, the clinical outcome measure indicated that 62% of the imipramine plus programmed practice group had a very good outcome. This compared with 32% for imipramine plus group exposure, 37% for group exposure plus placebo, and 20% for programmed practice plus placebo. The Agras et al. (1990) study illustrates several important points, raised in earlier research. Imipramine exerted its primary effects on mood and when combined with programmed practice on situational panic frequency and phobic avoidance. There were no significant differences between imipramine and in vivo exposure in reducing panic frequency by the end of the protocol. Exposure instructions should be considered a necessary although insufficient component in the treatment of PDA. Therapist-assisted in vivo exposure is more effective than exposure instructions alone in reducing negative panic appraisal and agoraphobic avoidance. Exposure was as effective as imipramine in reducing panic frequency by the third phase of the study. Examining the differential longitudinal effects of the treatments across the tripartite systems will be essential in ascertaining their relative outcome and maintenance. Similarly, the use of objective criteria to assess clinically significant improvement (Hirnaldi, Boice, & Barlow, 1986; Jacobson et al., 1984, 1988; Michelson & Mavissakalian, 1985) would permit meaningful cross-study comparisons of the findings. Clomipramine Clomipramine is a related TCA that has been used in Europe and Canada for many years and that has only recently been FDA-approved for use in the United States. Beaumont (1977) reported significant improvement for approximately two thirds of PDA clients and social phobics on an average of 150 mg per day. However, 37% of the clients failed to complete the 3-month medication trial, mostly because of noxious side effects. Johnston, Trover, and Whitsett (1988) reported an 8-week double-



blind trial of clomipramine versus placebo for PDA. The attrition rate was similar (35%) to that of the Beaumont (1977) study. The medication was found to be superior to placebo although no data were reported regarding what proportion of clients exhibited clinically significant improvement. Clomipramine has been studied more extensively in the context of obsessive-compulsive disorder. Hence generalizations regarding its relative efficacy for PDA cannot be made at present. In view of the methodological weaknesses in the few available reports, there is no firm scientific evidence that it offers advantages over the more researched and established TCA, imipramine.

Monoamine Oxidase Inhibitors The monoamine oxidase inhibitors (MAOIs) are another class of antidepressants that have been used, although much less frequently than the TCAs, for anxiety and panic. The most commonly used MAOI is the hydrogine derivative phenelzine (Nardil), and other MAOIs include isocarboxazid (Marplan) and tranylcypromine (Parnate). Although these agents have been primarily used in the treatment of depression, several pharmacotherapists have begun to evaluate their efficacy for panic disorder. Early MAOI trials generally showed negligible effects compared with placebo or only modest impact on panic attacks. (Tyrer, Candy, & Kelly, 1973; Lipsedge et al., 1973; Mountjoy, Roth, Garside & Leitch, 1977; Solyom et al., 1973). Subsequently, several double-blind studies using higher dosages for longer periods reported more noticeable effects on PDA (Liebowitz, Gorman, Fyer, & Klein, 1985; Sheehan, Ballenger, & Jacobson, 1980). Solyom, Solyom, LaPierre, Pecknold, and Morton (1981) assigned 40 PDA and social phobics to one of four treatments; (a) phenelzine plus exposure; (b) phenelzine, no exposure; (c) placebo plus exposure; and (d) placebo, no exposure. All treatments exhibited significant decreases in phobia ratings at posttreatment. However, clients receiving exposure evinced superior improvement compared with those who did not. No significant differences were found between phenelzine and placebo. No additive or synergistic effects were observed when adding phenelzine to exposure. Although some pharmacotherapists regard them as helpful for PDA clients resistant to other antidepressants, MAOIs might be considered as a tertiary level alternative for several reasons. First, these drugs are potentially toxic and can engender numerous serious medical problems, inducing convulsions, hypertensive crisis, and dangerous interactions with other medications and many common foods. Raskin and his colleagues (Raskin, Quitkin, & Harrison, 1984a, 1984b) found significantly greater side effects due to MAOIs versus TCAs, with 45% of the MAOI subjects discontinuing the drug. Second, although some reports suggest MAOIs might be helpful in reducing panic attacks, the extant empirical literature is both limited and equivocal. To quote from the Physicians' Desk Reference (1990) "(MAOIs] should rarely be the first antidepressant drug used. Rather, it is more suitable to use with patients who have failed to respond to the drugs more commonly listed for this [depression J condition" (p. 1631). Third, none of the MAOIs have been specifically FDA-approved for use in treating PDA. In light of

107

these caveats, the cost-benefit ratio favors consideration of alternative interventions. Issues in the use ofanlidepressants in PDA. Telch, Teaman, & Taylor (1983) reviewed the use of antidepressant medications for PDA and noted that although these drugs have been reported as beneficial, conceptual and methodologic problems have plagued pharmacotherapy studies. Diagnostic heterogeneity; inadequate evaluation of maintenance or relapse effects once the medications are withdrawn; and wide variations in dosage, length of treatment, and patient compliance are the norm rather than the exception. Despite the multidimensional nature of PDA, only rarely have the tripartite systems been monitored to assess outcome across cognitive, behavioral, and psychophysiological systems. In some cases ad hoc measures with unknown psychometric properties are used, completed by nonblind clinicians. The Telch et al. review (1983) concluded that the most effective pharmacologic agents, TCAs, effected improvement, although the exact extent of these changes could not be deduced because of methodological confounds. POT example, the concomitant use of behavior therapy, ranging from exposure instructions to therapist-aided exposure, in most pharmacotherapy studies, raised interpretative issues. The review found that 20% of the PDA clients refused even to take psychotropic medications. Another 35-40% prematurely terminated pharmacotherapy because of the noxious anticholinergic side effects of TCAs. From one fourth to one half of the clients in these studies who did improve relapsed on medication withdrawal. Telch (1988) in a more recent review of antidepressants and psychological treatments for PDA concluded that the efficacy of combined treatments is tentatively supported, although the unitary benefits of antidepressants alone appear more limited. Moreover, studies examining the suitability of treatment, perceived desirability, and efficacy of pharmacotherapy versus psychological treatments have found that the general public and anxiety clients both prefer psychological treatments and view them as more efficacious (Norton, Allen, & Hilton, 1983; Norton, Allen, & Walker, 1985). In addition to the above stated limitations of using antidepressants can be added a number of medical conditions and illnesses that preclude theirroutine use. For example, the majority of PDA clients are women in their childbearing years. This places obvious restrictions on pharmacotherapy for pregnant clients because of potential iatrogenic effects on the developing fetus. Similarly, a number of medical conditions, including cardiovascular, pulmonary, and seizure disorders contraindicate their use because of their potential interactions, either with the disorders themselves or with the medications commonly used to treat these conditions. Both cognitive-behavioral and antidepressant treatments exert beneficial effects on PDA. However, whereas the former treatments have been found to have both short- and long-term effects, the clinical course for clients on antidepressants alone appears less optimistic, with relapse rates of 35-50% being the norm, rather than the exception. Many clients have to remain on these medications for several years, lest they risk relapse. Although potentiating effects of TCAs combined with exposure


108

LARRY K. MICHELSON AND K A R E N MARCHIONE

have been found, both their mechanism of action and their durability remain uncertain and await future research. Potential benefits of using the TCAs include their relative safety, established clinical record (for those clients who can tolerate dosages (al 50 mg per day) required to achieve significant improvement, affordability, and availability for clients who do not desire or have access to cognitive-behavioral psychotherapy. In addition, their proven mood elevation properties can also provide a protective buffer to remediate or prevent clinical depression for at-risk clients. The decision to use TCAs should be made cojointly, with the client being fully apprised of the potential risks, side effects, and benefits. The salutary effects of TCAs can be significantly enhanced by concomitant cognitivebehavioral therapy. This should maximize treatment gains during a trial of TCAs and facilitate longitudinal maintenance once these medications are discontinued.

Beta-Adrenergic Blockers Beta-adrenergic blocking agents decrease the peripheral effects of the sympathetic nervous system. Their primary effect on anxiety disorders is to decrease peripheral somatic symptomology, such as tachycardia, trembling, and sweating. Betablockers appear to be ineffective in the treatment of major anxiety disorders, including PDA. Recent reports indicate that they exert clinically anxiolytic effects inferior to the benzodiazepines (BZs). In four out of six studies in which these drugs were directly compared with BZs, the BZs were found to be superior (Burrows, Davies, Fail, Poynton, & Stevenson, 1976; HallStrom, Treasaden, Edwards, & Lader, 1981; Johnson, Singh, & Leeman, 1976; Noyes et al, 1984; Tyrer & Lader, 1974; Wheatley, 1969). Beta-blockers have on occasion been used for performance anxiety, such as with anxious musicians, because these medications can inhibit certain somatic aspects of the stress response. Interest in the effects of beta-blockers with social phobia has also arisen (Liebowitz et al., 1985). However, to date no controlled studies of beta-blockers have found them superior to placebo for panic disorders.

Tranquilizer Utilization Studies in the United States and Europe (Baiter & Levine, 1969; Baiter, Manheimer, & Mellinges, 1984; Mellinger, Baiter, Manheimer, Cism, & Parry, 1978; Parry, 1968) reveal that 15% of the general population report using psychotropic medications, with one third of these users reporting dependence on minor tranquilizers. In the United States, 50% of all prescriptions are for antianxiety or sedative medications. Of all psychotropic medications, minor tranquilizers account for the largest share of prescriptions. In 1985, 81 million prescriptions were filled in the United States for benzodiazepines with an average cost of $25 per prescription. This amounts to over 1.5 billion dollars expended for minor tranquilizers (Baiter, personal communication, 1987). These data suggest that minor tranquilizer use is not a rare phenomenon, raising issues regarding their safety and effectiveness in treating PDA. These issues are reviewed in the following section.

Low-Potency Benzodiazepines Although the use of low-potency benzodiazepines (BZs) appears widespread, there is a paucity of controlled research studying their clinical efficacy for panic disorder with or without agoraphobia. Hafner and Marks (1976) assigned 30 PDA clients to four 3.5-hr exposure sessions over 2 weeks either with or without 7.5 mg of diazepam or placebo. Posttreatment and follow-up analyses revealed that all of the groups improved significantly and equally, with no significant drug effects found on any measure. Hence the drugs did not appear to enhance treatment effects. Chambless, Foa, Groves, & Goldstein (1979) found that PDA clients whose imeroceptive somatic symptoms were reduced using a BZ during exposure showed significantly inferior improvement compared with clients receiving exposure without these medications. In a 2-week study of 42 clients with either PDA, PD, or generalized anxiety disorder (PDA sample size unspecified), Noyes et al. (1984) assigned subjects to high dosages of diazepam (30 mg per day) or propranolol (240 mg per day). At the 2nd week, diazepam was superior to propranolol. Moderate improvements were reported among 18/21 of the diazepam clients. Inasmuch as a crossover design was used, no follow-up was reported. This report suggested that BZs, at high enough dosages, might reduce panic symptoms. Questions concerning the efficacy of low-potency BZs compared with alternative treatments and what their relative shortand long-term effects are have only recently begun to be answered. Although low-potency BZs can rapidly reduce anxiety, issues remain regarding possible psychological dependence, tolerance, side effects (sedation, psychomotor impairment), low efficacy for PDA, and high relapse rates. Overall, the low-potency BZs cannot be considered as treatments of choice for PDA. Imipramine and other TCAs have been found to be superior to minor tranquilizers in both short- and long-term efficacy.

High-Potency Benzodiazepines A new class of high-potency BZs has been developed and has been the subject of considerable clinical research. One of the most publicized and prescribed medications for panic is alprazolam (Xanax). Ballinger (1990) reported on the Upjohn CrossNational Collaborative Panic Study, which, in Phase 1, involved 500 clients with panic disorders who were randomly assigned to either alprazolam or placebo. Sixty percent of clients on alprazolam versus 30% of the placebo clients were panic-free after 8 weeks of treatment at dosages as high as 10 mg per day Many of these clients could not be withdrawn off the drug because of psychological dependence. Of those who discontinued the medication, 30% suffered rebound panic attacks worse than the panic attacks for which they had originally sought treatment. Finally, 90% of these clients experienced relapse following medication withdrawal. These data indicate that high-potency BZs can be beneficial in reducing panic attacks, phobic anxiety, and phobic avoidance, although their relapse rates are problematic (Ballinger, 1990; Fryer et al., 1987; Pecknold, Swinson, Kuch, & Lewis, 1988). Revisiting the Klosko et al. (1990) study comparing alprazolam and cognitive-behavior therapy (CBT), it was



found that CBT resulted in 87% of clients being panic-free at posttreatment, compared with 50% in the alprazolam condition. Because this is the only published study contrasting these treatments, replication will be important in corroborating these differential effects. Potential difficulties in using high-potency BZs include dependence, withdrawal, limited efficacy, panic rebound, and relapse. Also, misattribution of improvement to the medications rather than to the clients' own efforts might contribute to the observed relapse rates when these drugs are withdrawn. State-dependent learning would also undermine the transfer of learning from a medicated to a nonmedicatcd state and would contribute to relapse (Overton, 1977). Finally, findings from animal laboratories examining the compatibility of pharmacologic and behavioral treatments suggest that the process of habituation may involve the desensitization of the noradrenergic neurotransmitter system, a process that is inhibited by BZs (Weiss etal., 1982). The risksof using high-potency BZs should, of course, be weighed with their potential benefits, which include relatively rapid onset of effects (1-2 weeks); high client compliance; tolerability; low attrition; and reductions in panic attacks, and generalized and phobic anxiety for approximately 60% of clients receiving these medications at dosages of 2-3 mg/per day (Ballinger, 1990). Therefore, the decision to use these agents should be made collaboratively with the client informed of the possible risks and benefits. Hence, further empirical study of the relative and combined effects of high-potency BZs and CBT would be of much scientific interest. It would appear clinically prudent to offer PDA clients who are on these drugs the opportunity to avail themselves of CBT as a means of enhancing treatment outcome and long-term maintenance. In the next section the relative efficacy of pharmacotherapy and cognitive-behavioral therapy will be examined. Pharmacotherapy Versus Cognitive-Behavior Therapy In a recent review of two decades of cognitive-behavior therapy and pharmacologic panic disorder research, Clum (1989) examined attrition, outcome, and relapse. The report included pharmacologic studies of TCAs. MAOls, beta-blockers, minor tranquilizers, and high-potency benzodiazepines. Behaviorally based treatments included exposure strategies, cognitive therapy, cognitive coping techniques, cognitive restructuring, relaxation training, and breathing retraining. Operationalized criteria were used to derive attrition, improvement (focused exclusively on panic attacks), and relapse rates. Treatment success was defined as the absence of panic attacks or a 50% reduction in panic frequency at posttreatment. A number of salient findings emerged in this comparative review. Attrition rates were significantly higher among the TCAs versus cognitive-behavioral studies, and theaddition of psychological treatment did not reduce this rate. Overall, cognitive-behavioral treatments were found to be most effective (74% success). Success rates for low-potency benzodiazepines and betablockers did not surpass those found for placebo conditions, reaffirming the generally held notion that these agents are relatively ineffective in the treatment of PDA. Combining behavioral and pharmacologic treatments appeared to diminish the

109

success of their unitary counterparts. In particular, behavior therapy was more effective as a singular treatment than when combined with medication. PDA subjects treated with medication appeared to exhibit significantly less improvement than did panic disorder clients without phobic avoidance. Relapse was operationally denned as a recurrence of panic attack or reinstatement of treatment for panic attacks. Clum (1989) reports that follow-up data and relapse rates were rarely assessed in pharmacotherapy reports. TCA studies were also methodologically confounded by the routine addition of behavior therapy or supportive counseling, reducing the interpretability of their unitary effects. Overall, behavior therapy had significantly lower relapse rates than pharmacotherapy. Behavior therapy, in combination with TCAs, also decreased relapse rates compared with TCAs alone. Finally, a composite efficacy index was derived and indicated that cognitive-behavioral treatments were superior to beta-blockers, low-potency benzodiazepines, high-potency benzodiazepines, and antidepressants. In the following section estimates of the relative and combined efficacy of behavioral, cognitive, and pharmacologic treatments of PDA are presented. Hypothetical Cohort Table I depicts a hypothetical cohort of 100 PDA clients receiving alternative psychosocial and pharmacologic treatments and delineates the percentage of subjects who dropout, achieve clinically significant improvement, and relapse, yielding an overall long-term efficacy rate. These figures represent estimates based on literature reviews, meta-analyses, and recent clinical research. The beta-blockers exhibited a moderate attrition rate, a low improvement rate, and high relapse, yielding an overall efficacy index of 1 %. The low-potency benzodiazepines column reveals modest attrition and improvement rates, high relapse, and a long-term efficacy index of 2%. With high-potency benzodiazepines, there was low attrition, moderate improvement, and high (90%) relapse, resulting in an overall efficacy index of 5%. The MAOIs had a high attrition rate and moderate levels of improvement and relapse, resulting in an efficacy index of 17%. The TCAs also had a relatively high (25%) attrition rate, with similar outcome and relapse rates (although this is a conservative figure), generating an efficacy rate of 29%. The programmed practice column reveals low attrition, moderate improvement, and low relapse, yielding an 18% efficacy index. Relaxation training exhibited low attrition and relapse and moderate outcome, generating an overall efficacy index of 40%. Regarding graduated exposure, attrition was low, with 65% of the clients achieving clinically significant improvement and low relapse, resulting in an efficacy index of 47%. Imipramine plus graduated exposure evinced a relatively high attrition rate, slightly elevated outcome, compared with the effects of these treatments separately, and a moderate relapse rate, which resulted in an efficacy rate of 39%. Cognitive therapy plus graduated exposure, which was derived from the Michelson, Marchione, and Greenwald (1989) PDA outcome study, reveals low attrition and high posttreatment efficacy (87%), with a 10% relapse rate, resulting in an overall efficacy rate of 67%. Finally, for

110

LARRY K. MICHELSON AND K A R E N MARCH1ONE

comparative purposes, estimates were also computed for recent studies on cognitive treatments of panic disorder with no or mild phobic avoidance. These studies indicate negligible attrition and relapse rates, a high (90%) improvement rate, and an overall efficacy index of 81 %. Although these data are hypothetical, they are representative and indicate that cognitive therapy, combined with graduated exposure for PDA, and cognitive-behavioral treatments, directed specifically at panic disorder, foster optimal short- and long-term outcomes.

111! 11


HJI O "

•Is?

£

5 ~

.S §

'

II

es

Os t~-

org

m the

r-i —

s %

fo sse

o 2 a.

li

a

O O

'S "•s

S E

N

Summary

Is

Illlslf

S .= — ? O ai .«

SPECIAL SECTION: PANIC DISORDER WITH AGORAPHOBIA medications: National household survey data. Archives of General Psychiatry. 35,1045-1052. Michelson, L. (1984). The role of individual differences, response profiles and treatment consonance in anxiety disorders. Journal of Behavioral Assessment, 6, 349-368. Michelson, L. (1986). Treatment consonance and response profiles in agoraphobia: The role of individual differences in cognitive, behavioral, and psychophysiological treatments. Behaviour Research and Therapy. 24, 263-275.


Michelson, L. (1987). Cognitive-behavioral assessment and treatment of agoraphobia. In L. Michelson & L. M. Ascher(Eds-), Anxiety and stress disorders: Cognitive-behavioral assessment and treatment. New York: Guilford Press. Michelson, L., Marchione, K., & Greenwald. M. (1989, November). Cognitive-behavioral treatments of agoraphobia. Paper presented at the Annual Meeting of the Association for the Advancement of Behavior Therapy, Washington. DC. Michelson, L., Marchione, K., Greenwald, M., Glanz, L., Marchione, N., &. Testa, S. (1990). Cognitive-behavioral treatment of panic disorder. Behaviour Research and Therapy. 28. 141 -151. Michelson, L., & Mavissakalian. M. (1985). Psychophysiological outcome of behavioural and pharmacologic treatments of agoraphobia. Journal of Consulting and Clinical Psychology. 53. 229-236. Michelson, L., Mavissakalian, M., & Marchione, K. (1985). Cognitive and behavioral treatments of agoraphobia: Clinical, behavioral, and psychophysiological outcomes. Journal of Consulting and Clinical Psychology, 53, 913-925. Michelson, L., Mavissakalian, M., & Marchione, K. (1988). Cognitive, behavioral, and psychophysiological treatments of agoraphobia: A comparative outcome investigation. Behavior Therapy. 19, 97-120. Miller, R. C, & Berman. J. S. (1983). The efficacy of cognitive-behavioral therapies: A quantitative review of the research evidence. Psychological Bulletin, 94. 39-53. Mountjoy, C. Q., Roth, M.. Garside. R. R, & Leitch, I. N. (1977). A clinical trial of phenelzine in anxiety depression and phobic neurosis. British Journal of Psychiatry, 131, 486-492. Munby, J., & Johnson, D. W (1980). Agoraphobia. The long-term follow-up of behavioral treatment. British Journal of Psychiatry, 137, 418-427. Myers, J. K., Weissman, M. M.. Tischler, G. E., Holzer, C. E., Orvaschel, H., Anthony, J. C., Boyd, T. H., Burke, T. D., Kramer, M., & Stoltzman, R. (1984). Six-month prevalence of psychotic disordersof three communities. Archives of General Psychiatry, 41, 959-967. Norton, G. R., Allen, J. E., & Hilton, J. (1983). The social validity of treatments for agoraphobia. Behaviour Research and Therapy. 21, 393-399. Norton, G. R., Allen, J. E., & Walker, J. R. (1985). Predicting treatment preferences for agoraphobia. Behaviour Research and Therapy. 32. 699-701. Noyes, R., Anderson, D. J.. Clancy. J., Crowe. R. R., Slyman, D. J., Ghoneim, M. M., & Hinrichs, J. V (1984). Diazepam and propranolol in panic disorder and agoraphobia. Archives of General Psychiatry. 41, 287-292. Ost, L. G. (1987). Applied relaxation: Description of a coping technique and review of controlled studies. Behaviour Research andTherapy, 25, 397-409. Ost. L. G. (1989). A maintenance program for behavioral treatment of anxiety disorders. Behaviour Research and Therapy. 27. 123-130 Ost, L. G., Hellstrom, K., & Westling, B. (in press). Exposure, applied relaxation, and cognitive techniques in the treatment of agoraphobia. Behaviour Research and Therapy Ost, L. G., Jerremalm, A.. & Johansson, J. (1981). Individual response

113

patterns and the effects of different behavioral methods in the treatment of social phobia. Behaviour Research andTherapv, 19,1-16. Overton. D. A. (1977). Drug state dependent learning. In M. E. Jarvik (Ed.). Psychofthurmacology in the practice of medicine New York: Appleton-Century-Crofts. Parry. H. J. (1968). Use of psychotropic drugs by U S. adults. PublicHealth Reports, S3, 799-810. Pecknold, J. C, Swinson, R. P., Kuch. K., & Lewis. C. P. (1988). Alprazolam in panic disorder and agoraphobia: Results from a multi-centered trial-discontinuation effects. Archives of General Psychiatry, 45. 429-436. Physician's Desk Reference (1990). Oradell, NJ: Barnhart. Rachman, S. J. (1980). Emotional processing. Behaviour Research and Therapy, 7, 237-245. Rachman, S. J., & Wilson, G. T. (1980). The effects of psychological therapy Oxford: Pergamon Press. Rapee. R. (1987). The psychological treatment of panic attacks: Theoretical conceptualization and review of evidence. Clinical Psychology Review, 7, 427-438. Raskin. A.. Marks, I. M., & Sheehan, D. V (1983). The influence of depressed mood on the anttpanic effects ofantidepressant drugs. Unpublished manuscript. Raskin, J., Quitkin. F., & Harrison, W (1984a). Adverse reactions to monoamine oxidase inhibitors: A comparative study. Journal of Clinical Psychopharmacology, 4, 270-278. Raskin. J.. Quitkin. E. & Harrison, W (1984b). Adverse reactions to monoamine oxidase inhibitors: Treatment correlates and clinical management. Journal of Clinical Psychopharmacology, 5, 2-9. Salkovskis, P. M., Jones. D. R. O., & Clark, D. M. (1986). Respiratory control in the treatment of panic attacks: Replication and extension with concurrent measurement of behaviour and pCO2. British Journal of Psychiatry. 148, 526-532. Salkovskis, P. M., Warwick. H. M. C., Clark, D. M, & Wessel, D. J. (1986). A demonstration of acute hyperventilation during naturally occurring panic attacks. Behaviour Research and Therapy. 24,91-94. Schwartz, R. M. (1982). Cognitive behavioral modification: A conceptual review. Clinical Psychology Review, 2, 267-293. Sheehan, D. V, Ballenger, J. C, & Jacobson, G. (1980). Treatment of endogenous anxiety with phobic hysterical and hypochondrical symptoms. Archives of General Psychiatry, 37, 51-59. Sokol-Kessler. L., Beck, A T, Greenberg, R, & Wright, R. F. (1989). Cognitive therapy of panic disorder: A nonpharmacological alternative. Journal of Nervous and Mental Disease. 177. 711-716. Solyom, L., Heseltine, G. E, McClure, P. J., Solyom, C., Ledwidge, B., & Steinburg, G. (1973). Behavior therapy versus drug therapy in the treatment of phobic neurosis. Canadian Psychiatric Association Journal. IS, 25-31. Solyom, C.. Solyom, L.. LaPierre. Y., Pecknold, J. C, & Morton, L. (1981). Phenelzine and exposure in the treatment of phobias. Journal of Biological Psychiatry, 16, 239-248. Stern. R.. & Marks, I. (1973). Brief and prolonged flooding. Archivesof General Psychiatry, 28. 270-276. Telch, M. J. (1988). Combined pharmacologicand psychological treatments for panic sufferers. In S. Rachman & J. D. Maser (Eds), Panic: Psychological perspectives (pp. 67-188). Hillsdale, NJ: Erlbaum. Telcli, M. J., Agras. W S., Taylor, C. B., Roth, W T, & Gallen, C. (1985). Combined pharmacological and behavioral treatment for agoraphobia. Behaviour Research and Therapy, 23, 325-335. Telch, M. J., Teaman. V H., & Taylor, C. V (1983). Antidepressant medication in the treatment of agoraphobia: A critical review. Behavwiir Research and Therapy, 21, 505-527. Trull. T. J., Nietzel, M. T., & Main, A. (1988). The use of meta-analysis to assess the clinical significance of behavior therapy for agoraphobia. Behavior Therapy, 19. 527-538.


114

LARRY K. M1CHELSON AND K A R E N MARCHIONE

Tyrer, P J., Candy.}., & Kelly. D. A. (1973). A study of theclimcal effects of phenelzmc and placebo in the treatment of phobic anxiety, ftrctiopharmacologia, 32, 237-254. Tyrer. P. J.. & Lader. M. H. (1974). Response to propranolol and diazcpam in somatic anxiety. British American Journal, II, 14-16. Weiss. J. M.. Bailey. W, Goodman, P A.. Hoffman, L., Ambrose. M., Salmas, S., & Charry, J. A. (1982). A model for neurochemical study of depression. In M. Y Spiegelstem & A. Levy (Eds.), Behavioral models and the analysis of drug action. Amsterdam: Elsevier. Weissman. M., & Merikangas. K. R. (1986). The epidemiology of anxiety and panic disorder: An update. Journal ofChnical Psychiatry, 47, 11-17. Wheatley, D. (1969). Comparative effects of propranolol and chlordiazepide in anxiety states. British Journal of Psychiatry. 115. 14111412. Williams. S. L.. & Rappaport. A. (1983). Cognitive treatment in the natural environment for agoraphobics. Behavior Therapv. 4. 299313.

Wolpe, J. (1958). Psychotherapy by reciprocal inhibition. Stanford, CA: Stanford University Press. Woolfolk, R. L., & Lchrer, R M. (Eds.). (1984). Principles and practice of siress management New York: Guilford Press. Zitrin, C. M., Klein, D. F, & Woerner, M. G. (1978). Behavior therapy, supportive psychotherapy, imipramine and phobias. Archives of General Psychiatry, 37. 63-72. Zitrin, C. M . Klein, Q E, & Woerner, M. G. (1980). Treatment of agoraphobia with group exposure in vivo and imipramine. Archives of General Psychiatry, 37, 63-72. Zitrin, C. M.. Klein, D. E. Woerner, M. G., & Ross, D. C. (1983). Treatment of phobias: Comparison of imipramine hydrochloride and placebo. Archives of General Psychiatry, 40, 125-138. Received April 6,1990 Revision recieved July 27,1990 Accepted August 8,1990 •

Butcher, Geen, Hulse, and Salthouse Appointed New Editors, 1992-1997 The Publications and Communications Board of the American Psychological Association announces the appointments of James N. Butcher, University of Minnesota; Russell G. Geen, University of Missouri; Stewart H. Hulse. Johns Hopkins University; and Timothy Salthouse, Georgia Institute of Technology as editors of Psychological Assessment: A Journal of Consulting and Clinical Psychology, the Personality Processes and Individual Differences section of the Journal of Personality and Social Psychology, the Journal of Experimental Psychology: Animal Behavior Processes, and Psychology and Aging, respectively. As of January 1,1991, manuscripts should be directed as follows: • For Psychological Assessment send manuscripts to James N. Butcher, Department of Psychology, Elliott Hall, University of Minnesota, 75 East River Road, Minneapolis, Minnesota 55455. • For JPSP: Personality send manuscripts to Russell G. Geen, Department of Psychology, University of Missouri, Columbia, Missouri 65211. • For JEP: Animal send manuscripts to Stewart H. Hulse, Johns Hopkins University, Department of Psychology, Ames Hall, Baltimore, Maryland 21218. • For Psychology and Aging send manuscripts to Timothy Salthouse, Georgia Institute of Technology, School of Psychology. Atlanta, Georgia 30332.