Résumé
Objective: the aim of this study is to compare alternative severity classifications in anorexia nervosa (AN). We will assess four severity classifications based on the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), overvaluation of weight and shape (OWS), drive for thinness (DT), and a new combined proposal (OWS-DT), evaluating their ability to detect more severe patients regarding psychiatric comorbidities, psychopathology, physical consequences, and functional impairment. Method: a sample of 312 treatment-seeking patients with AN (mean age=26.8, SD=10.2; mean BMI=17.3 kg/m2, SD=2.4) were categorized using DSM-5 criteria based on body-mass index (BMI) (mild/moderate/severe/extreme), OWS (no/yes), DT (no/yes) and OWS-DT (neither/one or the other/both). The severity categories were compared based on psychiatric comorbidities, eating disorder psychopathology, biological variables, and functional impairment. Results: the results indicate that alternative criteria, such as OWS and DT individually, differentiate more severe patients in terms of psychiatric comorbidities, psychopathology, and functional impairment. However, the combination of OWS-DT emerges as the most efficient in identifying severity, exhibiting larger effect sizes, and enhancing discrimination ability across severity categories. In contrast, BMI-based classification shows poor clinical value, except for its ability to reflect malnutrition to some extent through differences in calorimetry. Additionally, regarding physical consequences such as abnormal biological tests in alternative classifications, significant differences are observed in the DT classification for deficiencies and abnormal hepatic tests. OWS-DT also presents notable distinctions, although the overall trend remains unclear. Conclusion: this study provided liXle support for BMI-based on the DSM-5 severity rating for AN. Alternative criteria, OWS and DT proved clinical value to distinguish more severe patients. The combined criterion OWS-DT emerged as being more efficient to identify severity than alternative criteria individually.