Anxiety disorders and panic attacks: an umbrella review of meta-analyses on psychotherapy, medication, internet-based and exercise-based treatment, 2014–2023

AbstractBackground. Anxiety disorders, including panic disorder, generalised anxiety disorder and social anxietydisorder, are the most common mental disorders. Panic attacks are frequent, highly distressing and often lead toemergency presentations. Effective psychological and pharmacological treatments are available, and their efficacyhas been examined in numerous meta-analyses that are rarely read together. Methods. Umbrella review of systematic reviews and meta-analyses of treatments for anxiety disorders in adults,with a focus on panic disorder and generalised anxiety disorder, covering cognitive behavioural therapy (CBT),other psychotherapies, medication, internet-based delivery, durability and physical activity. Structured search ofbibliographic records up to 2 November 2026. Published pooled estimates were extracted without re-pooling;methodological features were appraised descriptively. Results. Eight meta-analyses (2014–2023) were included, with up to 234 randomised trials and 37,333 patients peranalysis. Medication and psychotherapy both produced large pre–post improvements (Cohen’s d 2.02 and 1.22),but pre–post effects are inflated by high placebo response (d 1.29). Against psychological or pill placebo, CBTshowed a moderate effect on target symptoms (Hedges g 0.56). In panic disorder, CBT increased remissioncompared with waiting list (OR 3.0 after adjustment for small-study effects) and antidepressants reduced nonresponse compared with placebo (RR 0.72; number needed to treat 7). In generalised anxiety disorder, duloxetine,pregabalin, venlafaxine and escitalopram were effective and comparatively well tolerated. Gains afterpsychotherapy were maintained for up to two years. Guided internet-based CBT was as effective as face-to-face CBT(g –0.01). Physical activity reduced anxiety symptoms (median SMD –0.42). Conclusions. Anxiety disorders and panic disorder are highly treatable. CBT, selected antidepressants andpregabalin are effective first-line options, internet-based CBT extends access, and exercise is a useful adjunct. Trueeffects against placebo are moderate rather than large, which supports shared decision-making based onpreference, availability and tolerability. Author. Dirk de Pol, Editor-in-Chief, MedizinDoc (medizindoc.de), Berlin, Germany