OBJECTIVE: It is of considerable theoretical and clinical importance to assess whether childhood sexual abuse is a risk factor for the development of bulimia nervosa. The authors reviewed the scientific literature bearing on this issue. METHOD: Since prospective studies on this question have not been done, they assessed 1) controlled retrospective studies comparing the prevalence of childhood sexual abuse among bulimic and control groups, 2) uncontrolled retrospective studies of the prevalence of childhood sexual abuse in samples of 10 or more bulimic subjects, and 3) studies of the prevalence of childhood sexual abuse in the general population, which were chosen to match as closely as possible in methodology the available studies of bulimia nervosa (i.e., in geographic location, age and ethnicity of subjects, interview method, and criteria for defining childhood sexual abuse). RESULTS: Controlled studies generally did not find that bulimic patients show a significantly higher prevalence of childhood sexual abuse than control groups, especially when allowance is made for possible methodologic effects. Furthermore, neither controlled nor uncontrolled studies of bulimia nervosa found higher rates of childhood sexual abuse than were found in studies of the general population that used comparable methods. When it is taken into consideration that several methodologic factors might have exaggerated the rates of childhood sexual abuse among subjects with bulimia nervosa relative to rates in the general population, the absence of actual observed differences becomes particularly striking. CONCLUSIONS: Current evidence does support the hypothesis that childhood sexual abuse is a risk factor for bulimia nervosa.Abstract Teaser