Personalized medicine in panic disorder: where are we now? A meta-regression analysis

Personalized medicine assumes that individual's unique characteristics are central in tailoring effective pharmacological interventions. The identification of predictors of pharmacotherapy effectiveness is crucial in panic disorder (PD), but consensus on this topic is still lacking. Consequently, we carried out a meta-analysis, according to PRISMA guidelines, with the aim of identifying sociodemographic and clinical moderators of short-term outcomes and tolerability of US FDA-approved medications for PD. We performed a database search on randomized placebo-controlled trials using PubMed, PsycINFO, and Embase. Through the selection process, we finally meta-analyzed 29 comparisons between paroxetine, venlafaxine XR or alprazolam and placebo. We employed the random-effects meta-regression technique. The major results were that longer illness duration was significantly associated with a lower rate of patients free from panic attacks at the end of trials with venlafaxine XR, and that higher age at the beginning of trials was significantly associated with a higher rate of dropouts because of side effects during trials with paroxetine. In addition, longer treatment was associated with a higher rate of patients free from panic attacks at endpoint in trials with venlafaxine XR. Overall, we found limited support for the moderating effects of sociodemographic and clinical variables on the short-term pharmacotherapy for PD. However, our results should be considered with caution considering the limited statistical power and the risk of publication bias. Future studies are needed to overcome the paucity of available data and the shortcoming of the current pharmacological studies in PD.