Combined treatment of agoraphobia and panic disorders with behavioral and drug therapy. Critical view on meta-analyses

In everyday clinical practice, the simultaneous treatment of patients with anxiety disorders with behavioral and drug therapy is frequently applied. A s both of these treatments have been shown to be highly effective, such a p ractice appears to be plausible. Yet, from published studies it is surprisi ngly difficult to support this practice. Can meta-analyses of these studies improve the evidence? For agoraphobia and panic disorder, four of eight pu blished meta-analyses did investigate the effects of combined behavioral an d drug treatment. In two of these studies, the shortterm effects of exposur e treatment and imipramine appeared to be better than those of their single application; in the two other studies no differences were found. The contr oversial results of these studies seem to be dependent on the meta-analytic methodology applied: Effect Sizes (ES) derived from pre- to posttreatment comparison of treatments showed a superiority of their simultaneous applica tion. ES derived from treatment - vs. control group comparisons at posttrea tment - did not show any significant differences in outcome. Unfortunately, there are no meta-analytic studies on the long-term outcome of simultaneou s behavioral and drug treatment in these disorders. Basically, the generali zability of the results of meta-analyses is limited, as diagnostic and trea tment variations between the reviewed studies have not been adequately take n into account. There is increasing evidence that subgroups of agoraphobia and panic disorder receive different modes of behavioral treatment. Further more, even within one symptom-directed technique like exposure, huge variat ions in application have been used: classical systematic desensitization or the flooding type of exposure in vivo and in fantasy; treatment sessions m ay last from 20 min to more than 5 h continuously; the degree of therapist- aided versus self-applied exposure varies greatly. Therefore, recommendatio ns as to whether and in what conditions (simultaneously or sequentially) co mbined behavioral and drug treatment is to be recommended, can reliably be drawn only from studies that have considered these variables. In spite of t he huge amount of publications on these disorders, studies with this type o f approach have only just begun.