Background: Bulimia nervosa (BN), a mental disorder that causes significant impairment, can be treated with psychological, pharmacological, nutrition-based and self-help interventions. We conducted a pre-registered meta-analysis of randomized-controlled trials (RCTs) to assess the efficacy of these interventions in up to 19 different interventions.
Methods: Database search terms were combined for BN and RCTs from database inception to March 2017. Abstinence from binge eating episodes, compensatory behaviors, the absence of a BN diagnosis and reduction of symptom severity were considered as primary outcome variables, reduction of self-reported eating pathology and depression served as secondary outcome variables. Retrieved RCTs were meta-analyzed using fixed and random effects models.
Results: RCT (79 trials; 5775 participants) effects post-treatment revealed moderate to large intervention effects for psychotherapy [mostly cognitive-behavioral therapy (CBT)] for primary outcome variables. Slightly reduced effects were obtained for self-help and moderate effects for pharmacotherapy. Similarly, psychotherapy yielded large to very large effects in regard to secondary outcome variables, while moderate to large effects were observed for self-help, Pharmacotherapy and combined therapies. Meta-analyses for the pre to post changes within group confirmed these findings. Additionally, follow-up analyses revealed the sustainability of psychotherapies in terms of large effects in primary outcome criteria, while these effects were moderate for self-help, pharmacotherapy, and combined therapies.
Conclusions: Most psychological and pharmacological interventions revealed to be effective in BN treatment. Taking effect size, sustainability of the intervention, as well as the consistency of findings and available evidence into consideration, CBT can be recommended as the best intervention for the initial treatment of BN.
Keywords: Bulimia nervosa; meta-analysis; treatment.