OBJECTIVE

Atypical anorexia nervosa (AAN) is defined by the symptoms of anorexia nervosa in the presence of "significant weight loss" in individuals who are not underweight. Description of current weight in AAN has been limited, significant weight loss has not been defined, and the distinction between having AAN versus having weight suppression has not been examined.

METHOD

Secondary analyses were conducted with data from an epidemiological study of women (n = 1,640) and men (n = 794). Three definitions of significant weight loss (5, 10, and 15%) for AAN were tested in comparisons with controls and a DSM-5 eating disorder group (DSM-5 ED) on measures of eating pathology and clinical significance using ANCOVA and logistic regression, controlling for age and body mass index. Secondary analyses compared AAN to a weight suppressed group (WS-only) and a cognitive concerns group (COG-only).

RESULTS

Across weight loss thresholds, ≥25% of adults with AAN were currently overweight/obese. At the 5% and 10% definitions, AAN was associated with elevated eating pathology and distress relative to controls, WS-only, and COG-only in women and men. Women with AAN endorsed less eating pathology and distress than DSM-5 ED at some weight loss definitions, whereas men with AAN did not differ from DSM-5 ED in eating pathology or distress.

DISCUSSION

Results support that even a 5% weight loss, combined with cognitive concerns, may produce a group with a clinically significant eating disorder. AAN was observed in both healthy weight and overweight/obese adults, highlighting the importance of screening for restrictive eating disorders at all weights.