Treatment outcomes of psychotherapy for binge-eating disorder in a randomized controlled trial: Examining the roles of childhood abuse and post-traumatic stress disorder

Abstract

Objective: To examine childhood abuse and post-traumatic stress disorder (PTSD) as predictors and moderators of binge-eating disorder (BED) treatment outcomes in a randomized controlled trial comparing Integrative Cognitive-Affective Therapy with cognitive-behavioural therapy administered using guided self-help.

Method: In 112 adults with BED, childhood abuse was defined as any moderate/severe abuse as assessed by the Childhood Trauma Questionnaire, lifetime PTSD was assessed via the Structured Clinical Interview for DSM-IV, and outcomes were assessed via the Eating Disorder Examination (EDE). Covariate-adjusted regression models predicting binge-eating frequency and EDE global scores at end of treatment and 6-month follow-up were conducted.

Results: Lifetime PTSD predicted greater binge-eating frequency at end of treatment (B = 1.32, p = 0.009) and childhood abuse predicted greater binge-eating frequency at follow-up (B = 1.00, p = 0.001). Lifetime PTSD moderated the association between childhood abuse and binge-eating frequency at follow-up (B = 2.98, p = 0.009), such that childhood abuse predicted greater binge-eating frequency among participants with a history of PTSD (B = 3.30, p = 0.001) but not among those without a PTSD history (B = 0.31, p = 0.42). No associations with EDE global scores or interactions with treatment group were observed.

Conclusions: Results suggest that a traumatic event history may hinder treatment success and that PTSD may be more influential than the trauma exposure itself.

Keywords: binge-eating disorder; childhood abuse; eating disorders; post-traumatic stress disorder; psychotherapy.

Figure 1.

Predicted objective binge-eating episode frequency at six-month follow-up by childhood abuse history and lifetime post-traumatic stress disorder (PTSD), adjusted for baseline binge-eating episode frequency, participant demographics, study site, and treatment group. Error bars represent standard errors.