There is a growing body of research on manualized self-help interventions for bulimia nervosa (BN) and binge eating disorder (BED). Study and treatment dropout and adherence represent particular challenges in these studies. However, systematic investigations of the relationship between study, intervention and patient characteristics, participation, and intervention outcomes are lacking. We conducted a systematic literature review using electronic databases and hand searches of relevant journals. In metaregression analyses, we analyzed study dropout as well as more specific measures of treatment participation in manualized self-help interventions, their association with intervention characteristics (e.g. duration, guidance, intervention type [bibliotherapy, CD-ROM or Internet based intervention]) and their association with treatment outcomes. Seventy-three publications reporting on 50 different trials of manualized self-help interventions for binge eating and bulimia nervosa published through July 9th 2012 were identified. Across studies, dropout rates ranged from 1% to 88%. Study dropout rates were highest in CD-ROM interventions and lowest in Internet-based interventions. They were higher in samples of BN patients, samples of patients with higher degrees of dietary restraint at baseline, lower age, and lower body mass index. Between 6% and 88% of patients completed the intervention to which they had been assigned. None of the patient, study and intervention characteristics predicted intervention completion rates. Intervention outcomes were moderated by the provision of personal guidance by a health professional, the number of guidance sessions as well as participants' age, BMI, and eating disorder related attitudes (Restraint, Eating, Weight and Shape Concerns) at baseline (after adjusting for study dropout and intervention completion rates). Guidance particularly improved adherence and outcomes in samples of patients with bulimia nervosa; specialist guidance led to higher intervention completion rates and larger intervention effects on some outcomes than non-specialist guidance. Self-help interventions have a place in the treatment of BN and BED, especially if the features of their delivery and indications are considered carefully. To better determine who benefits most from what kind and "dosage" of self-help interventions, we recommend the use of consistent terminology as well as uniform standards for reporting adherence and participation in future self-help trials.