No consistent first-option psychological interventions for adult outpatients with anorexia nervosa (AN) emerges from guidelines. A random-effect network meta-analysis (NMA) of randomised controlled trials (RCTs) about stand-alone non-pharmacological treatments was conducted. We searched published and unpublished literature until March 20th, 2020. Primary outcomes were change in body mass index (BMI), change in clinical symptoms and all-cause dropout rate. Overall, 16 RCTs were included in the systematic review and 13 contributed to the NMA (overall, 1047 patients). Seven interventions were assessed: treatment as usual (TAU), cognitive behavioural therapy (CBT), Maudsley Anorexia Nervosa Treatment for Adults, family therapy, psychodynamic psychotherapies, and two novel forms of CBT (targeting compulsive exercise, adding cognitive remediation therapy). No intervention outperformed TAU in our primary outcomes, but all-cause dropout rate was lower for CBT than psychodynamic psychotherapies (OR 0.54, 95% CI 0.31-0.93). Heterogeneity or inconsistency emerged only for a few comparisons. Confidence in evidence was low to very low. When compared with TAU, specific psychological treatments for adult outpatients with AN can be associated with modest improvements but no reliable evidence supports their superiority as recommended by clinical guidelines internationally. Even modest improvement may greatly improve the clinical course and quality of life for individuals with AN, so there is urgent need to improve therapies for adults with AN. To better understand effects of available treatments, individual participant data should be freely available to eventually identify groups predisposed to respond to specific treatments.

Chair: Professor Christopher Fairburn