LAUSR

Abstract Background Accumulating evidence suggests that many psychiatric disorders etiologically represent the extreme end of dimensionally distributed features rather than distinct entities. The extent to which this applies to eating disorders (EDs) is unknown. Methods We investigated if there is similar etiology in (a) the continuous distribution of the Eating Disorder Inventory-2 (EDI-2), (b) the extremes of EDI-2 score, and (c) registered ED diagnoses, in 1481 female twin pairs at age 18 years (born 1992–1999). EDI-2 scores were self-reported at age 18. ED diagnoses were identified through the Swedish National Patient Register, parent-reported treatment and/or self-reported purging behavior of a frequency and duration consistent with DSM-IV criteria. We differentiated between anorexia nervosa (AN) and other EDs. Results The heritability of the EDI-2 score was 0.65 (95% CI 0.61–0.68). The group heritabilities in DeFries–Fulker extremes analyses were consistent over different percentile-based extreme groups [0.59 (95% CI 0.37–0.81) to 0.65 (95% CI 0.55–0.75)]. Similarly, the heritabilities in liability threshold models were consistent over different levels of severity. In joint categorical-continuous models, the twin-based genetic correlation was 0.52 (95% CI 0.39–0.65) between EDI-2 score and diagnoses of other EDs, and 0.26 (95% CI 0.08–0.42) between EDI-2 score and diagnoses of AN. The non-shared environmental correlations were 0.52 (95% CI 0.32–0.70) and 0.60 (95% CI 0.38–0.79), respectively. Conclusions Our findings suggest that some EDs can partly be conceptualized as the extreme manifestation of continuously distributed ED features. AN, however, might be more distinctly genetically demarcated from ED features in the general population than other EDs.