Controlled family study of anorexia nervosa and bulimia nervosa: evidence of shared liability and transmission of partial syndromes.

OBJECTIVE Lifetime rates of full and partial anorexia nervosa and bulimia nervosa were determined in first-degree relatives of diagnostically pure proband groups and relatives of matched, never-ill comparison subjects. METHOD Rates of each eating disorder were obtained for 1,831 relatives of 504 probands on the basis of personal structured clinical interviews and family history. Best-estimate diagnoses based on all available information were rendered without knowledge of proband status and pedigree identity. Only definite and probable diagnoses were considered. RESULTS Whereas anorexia nervosa was rare in families of the comparison subjects, full and partial syndromes of anorexia nervosa aggregated in female relatives of both anorexic and bulimic probands. For the full syndrome of anorexia nervosa, the relative risks were 11.3 and 12.3 in female relatives of anorexic and bulimic probands, respectively. Bulimia nervosa was more common than anorexia nervosa in female relatives of comparison subjects, but it, too, aggregated in the families of ill probands; the corresponding relative risks for bulimia nervosa were 4.2 and 4.4 for female relatives of anorexic and bulimic probands, respectively. When partial syndromes of anorexia nervosa and bulimia nervosa were considered, relative risks fell by one-half in each group of ill probands. CONCLUSIONS Both anorexia nervosa and bulimia nervosa are familial. Their cross-transmission in families suggests a common, or shared, familial diathesis. The additional observation that familial aggregation and cross-transmission extend to milder phenotypes suggests the validity of their inclusion in a continuum of familial liability.