OBJECTIVES: The family compositions of patients with anorexia nervosa may inform consideration of the etiology of the disorder, and previous studies in this area have had methodologic weaknesses which we sought to address in a case-control design. METHODS: The family composition of 259 females with anorexia nervosa was compared with that of 200 control subjects. Birth order and numbers of brothers and sisters were compared in unmatched and matched analyses using both univariate and multivariate statistical techniques. RESULTS: In both the unpaired univariate and the matched conditional logistic regression analyses, the anorexic females were both significantly later in the birth order and had significantly fewer brothers than control subjects. DISCUSSION: Although it is possible that these findings were affected by response bias among control subjects, it is more likely that they are informative with regard to the etiology of anorexia nervosa. Possible etiologic links are discussed in relation to biologic, family, and sociocultural factors. Copyright 2005 by Wiley Periodicals, Inc
OBJECTIVES: The family compositions of patients with anorexia nervosa may inform consideration of the etiology of the disorder, and previous studies in this area have had methodologic weaknesses which we sought to address in a case-control design. METHODS: The family composition of 259 females with anorexia nervosa was compared with that of 200 control subjects. Birth order and numbers of brothers and sisters were compared in unmatched and matched analyses using both univariate and multivariate statistical techniques. RESULTS: In both the unpaired univariate and the matched conditional logistic regression analyses, the anorexic females were both significantly later in the birth order and had significantly fewer brothers than control subjects. DISCUSSION: Although it is possible that these findings were affected by response bias among control subjects, it is more likely that they are informative with regard to the etiology of anorexia nervosa. Possible etiologic links are discussed in relation to biologic, family, and sociocultural factors. Copyright 2005 by Wiley Periodicals, Inc