Do antidepressants selectively suppress spontaneous (unexpected) panic attacks? A replication

The purpose of this study was to test the following interrelated hypotheses in a larger sample by attempting to replicate supportive results from a small therapeutic study: (1) the pathogenesis of panic disorder includes at least two identifiable components: a biological component represented by spontaneous (unexpected) panic attacks, and a cognitive component represented by situational attacks and especially by phobias; (2) these components respond differently to treatment; (3) many biological processes respond to an effective intervention in proportion to their deviance from "normal" prior to treatment ("Law of Initial Value"); and (4) the response of spontaneous panic attacks to an effective treatment conforms to that model. Previously, the authors reanalyzed an 8-week therapeutic study of panic disorder that included groups treated with placebo and with imipramine (225 mg daily). The criteria of response were spontaneous panic attacks (biological component), situational panic attacks (both components), and agoraphobia ratings (cognitive component). The analyses compared the regression lines for posttreatment status on pretreatment status in the imipramine and placebo groups. The effect of imipramine on spontaneous panic attacks fitted the hypothesized model: the pre-post slope in the placebo group was approximately 1 (45 degrees), whereas the slope in the imipramine group was approximately 0. There was no significant difference in pre-post slopes between the imipramine and placebo groups for situational panic attacks or agoraphobia ratings. For this report, the authors applied the same approach to another larger data set from a study using a similar design, but a different antidepressant. In this multicenter, double-blind study, patients with panic disorder were randomly assigned to receive 10 weeks of treatment with placebo (N = 78) or fluoxetine 10 mg (N = 84) or 20 mg (N = 81) daily. Spontaneous and situational panic attacks were registered in a daily diary, and agoraphobia was rated at each visit. Using baseline and endpoint data, fluoxetine had a statistically significant, dose-dependent, suppressive effect on spontaneous panic attacks, as measured by the pre-post slopes in the three treatment groups. The placebo group showed some response (slope = 0.69). There were no significant drug effects on situational panic attacks. On ratings of agoraphobia, the slopes in the placebo and the fluoxetine 20 mg groups did not differ, but the slope in the fluoxetine 10 mg group was significantly less than that in the placebo group, suggesting a therapeutic drug effect on agoraphobia only at the lower dose. These results are consistent with the stated hypotheses. They suggest that the therapeutic effects of antidepressants on panic disorder may be due primarily to the specific suppression of spontaneous panic attacks among patients with high baseline pathologic findings. Implications of these results for concepts of pathogenesis, clinical practice, and therapeutic research regarding panic disorder are discussed.