Vergleich der Wirksamkeit von Psychopharmaka bei Angststörungen

Abstract

Efficacy of pharmacological treatments for anxiety disorders: a meta-analysis This meta-analysis compares the efficacy of pharmacological treatments for the three main anxiety disorders (social phobia, generalized anxiety disorder, panic disorder). A total of 109 studies with 28785 patients were found to be eligible. This is the first meta-analysis comparing the relative (treated vs. control) and the absolute (pre-post) effect size of all commonly used pharmacological treatments for the three major anxiety disorders. The meta-analysis showed large effect size heterogeneity among the studies despite similar outcome measures and the use of similar inclusion criteria. Because moderate to high heterogeneity was found for most comparisons, the random-effects model was used in all analyses. In direct comparisons, all pharmacological treatments were significantly more effective than pill placebo except citalopram, moclobemide and opipramol. Phenelzine (d = 0.98), lorazepam and clomipramine (d = 0.87) and hydroxyzine (d = 0.79) had the highest treated vs. control effect sizes. Benzodiazepines (for example delorazepam: d = 3.54; bromazepam: d = 2.86; lorazepam: d = 2.53), quetiapine (d = 3.39), escitalopram (d = 2.67) and hydroxyzine (d = 2.56) had the highest pre-post effect sizes. However, the results with these drugs were only based on few studies. Among the drug treatment groups the SNRIs (d = 2.25), the benzodiazepines (d = 2.14) and the SSRIs (d = 2.09) had the highest pre-post effect sizes. A regression analysis showed that placebo effects sizes have increased between 1983 and 2013. To assess publication biases, among other methods, Duval and Tweedies (2000) “trim and fill” method was used to adjust the funnel plot by estimating the number of missing studies. After adjusting effect sizes for publication bias, none of the significant results became nonsignificant. Possible allegiance effects for all study arms were analysed. Allegiance effects were assumed in 50.8 % of 187 study arms. Effect sizes were calculated separately for subgroups with or without allegiance effects. However, allegiance effects did not alter effect sizes significantly. The choice of a drug should not only be made solely on the basis of efficacy but also on possible side effects. Benzodiazepines may cause dependency, tricyclic antidepressants have more adverse events than SSRIs (Bandelow et al. 2008a).