Abstract
OBJECTIVE: While research supports differentiating anorexia nervosa into binge-purge (AN-BP) vs. restricting (AN-R) subtypes, DSM-5-TR does not provide a specific threshold of binge and/or purge episodes that constitutes an AN-BP vs. AN-R diagnosis. Our review of the literature suggests that cutoffs used for defining AN subtypes are rarely reported and, when reported, are highly heterogeneous. Inconsistent subtyping protocols limit generalizability and understanding of AN-R and AN-BP differences. METHOD: The present study used structural equation modeling (SEM) trees to empirically determine the frequency of binge eating and/or purging that best differentiates AN subtypes. We then compared empirically determined groups on characteristics frequently found to differ between subtypes. Participants were 731 adolescents and adults with AN (94% female, M(age) = 20, 72% clinically diagnosed with AN-R) in a partial hospitalization program who completed assessments of AN and comorbid symptoms at intake. RESULTS: SEM tree analyses yielded four subgroups: past-month binge/purge frequency 0 (AN-R; n = 396); frequency 1-3 (AN-BP1; n = 101); frequency 4-15 (AN-BP2; n = 130); and frequency > 16 (AN-BP3; n = 98). AN-R differed from higher frequency groups on 14/22 clinical characteristics, AN-BP1 differed from higher frequency groups on 11/22 clinical characteristics, and AN-BP2 differed from higher frequency groups on 2/22 clinical characteristics. CONCLUSIONS: Findings suggest that one binge eating and/or purge episode in the past month provides adequate distinction between subtypes. These findings indicate that the DSM's definition of AN-BP may need to be revised to specify that the presence of any binge eating or purging, rather than "recurrent" binge eating or purging, is sufficient for subtyping AN.