BACKGROUND: Pharmacotherapy for generalized anxiety disorder (GAD) may be effective in reducing symptoms in the majority of patients. The study of moderators and predictors of treatment response may help clinicians both to select appropriate interventions to maximize the probability of response and to inform the general prognosis. METHODS: A systematic literature search of electronic databases, selected authors, and reference lists was used to identify articles that reported trials of drug monotherapy in GAD. Data on predictors and moderators were extracted. Quality of evidence was determined by the presence of a priori hypotheses, number of variables investigated, adequate quality of the measurement, and use of interaction-effects testing. RESULTS: From the 98 articles meeting inclusion criteria, 24 reported a total of 22 factors associated with treatment response. The reported results were heterogeneous, ranging over sociodemographic, clinical, comorbidity, genetic, and functional-imaging studies. Major depressive symptoms were found to moderate treatment outcome in favor of antidepressants versus benzodiazepines. Neuroticism, previous treatment, genetic polymorphisms (including serotonin receptor gene 2A), and functional activation of the anterior cingulate cortex and amygdala were identified as potential predictors of treatment response. CONCLUSIONS: Correlates of poor emotion regulation predicted poor treatment response, but subclinical depression was the only variable capable of informing treatment selection in this review. Future research should focus on further exploring the value of depression as a moderator and on a narrower list of potential genetic, brain-imaging, and temperament predictors of response to pharmacotherapy in GAD.
BACKGROUND: Pharmacotherapy for generalized anxiety disorder (GAD) may be effective in reducing symptoms in the majority of patients. The study of moderators and predictors of treatment response may help clinicians both to select appropriate interventions to maximize the probability of response and to inform the general prognosis. METHODS: A systematic literature search of electronic databases, selected authors, and reference lists was used to identify articles that reported trials of drug monotherapy in GAD. Data on predictors and moderators were extracted. Quality of evidence was determined by the presence of a priori hypotheses, number of variables investigated, adequate quality of the measurement, and use of interaction-effects testing. RESULTS: From the 98 articles meeting inclusion criteria, 24 reported a total of 22 factors associated with treatment response. The reported results were heterogeneous, ranging over sociodemographic, clinical, comorbidity, genetic, and functional-imaging studies. Major depressive symptoms were found to moderate treatment outcome in favor of antidepressants versus benzodiazepines. Neuroticism, previous treatment, genetic polymorphisms (including serotonin receptor gene 2A), and functional activation of the anterior cingulate cortex and amygdala were identified as potential predictors of treatment response. CONCLUSIONS: Correlates of poor emotion regulation predicted poor treatment response, but subclinical depression was the only variable capable of informing treatment selection in this review. Future research should focus on further exploring the value of depression as a moderator and on a narrower list of potential genetic, brain-imaging, and temperament predictors of response to pharmacotherapy in GAD.