The early identification of likely remitters and non-remitters to pharmacotherapy for panic disorder may have important implications for clinical treatment decisions. To address this question, combined data from two fixed-dose and two flexible dose placebo-controlled studies of sertraline treatment of panic disorder were examined. Patients (N=544) diagnosed with panic disorder, with or without agoraphobia, were treated with 50 mg of sertraline, 100 mg of sertraline, flexible dosages of sertraline, or placebo. Measures of early improvement included panic attack frequency (full + limited symptom attacks), anticipatory anxiety, the Hamilton Anxiety Rating Scale (HAM-A), and the Clinical Global Impression Improvement (CGI-I) Scale. Improvement as reflected in CGI-I ratings and change from baseline in the HAM-A at weeks 1, 2, and 3 significantly (P<0.0001) predicted endpoint clinical remission (defined at endpoint as no full panic attacks and a CGI-Severity rating of 1 or 2). Improvements in panic attack frequency and anticipatory anxiety were not consistent predictors in multivariate predictive models. Receiver-Operator Curve analyses revealed good specificity (0.83) for change in CGI-I at week 2, and good sensitivity (0.82) for change in HAM-A at week 3. Predictive success for HAM-A and CGI-I was not significantly different for fixed vs. flexible dose sertraline treatment, nor for sertraline vs. placebo treatment. The use of ROC analyses for examination of early response as a predictor of final remission holds promise for aiding clinicians in decision making regarding the need for alternative or supplemental treatment approaches during the course of pharmacotherapy for panic disorder.
Copyright 2002 Elsevier Science Ltd.