Early detection of relapse in panic disorder

Objective: To explore predictive models of relapsing based on change in symptoms at a time when panic disorder patients are still in remission following discontinuation of antidepressants. Method: Forty-seven subjects, who were randomized to double-blind placebo and who had valid data at four time points: pretreatment, randomization to placebo substitution, an assessment on placebo prior to the last assessment or relapse and their last assessment (relapsers n = 15, non-relapsers n = 32) were studied using descriptive, growth curve analysis and logistic regression methodologies. Results: Measures of generalized anxiety, fearfulness and disability at work and at home were better predictors of relapse than measures of panic and anxiety sensitivity. Logistic regression models using any one of these four general variables and its linear change correctly predicted relapse for 78.7-84.4% of the study subjects. Conclusion: It is possible to gauge, with a fair degree of accuracy, the probability of relapsing in panic disorder patients who have discontinued serotonergic antidepressants 2 months prior to the return of panic.