Purpose: To compare the relative prevalence of eating disorders moving from DSM-IV to DSM-5, and to reassess the overall medical impairment in the revised diagnostic classes.
Methods: We applied DSM-5 to 206 patients (age 15-56 years) previously studied and classified according to DSM-IV. Medical impairment was classified as low, medium, or high, based on a cumulative score of clinical severity (SCS), computed as the sum of specific weights assigned to different pathological conditions and their ascertained prognostic impact.
Results: Application of DSM-5 produced a decrease in Eating Disorders Not Otherwise Specified (EDNOS) by 17 %, an increase in anorexia (AN) by 14 % and bulimia (BN) by 2.4 %; 44.6 % of EDNOS migrated to AN, 8 % to BN, and 30.8 % was reclassified as Other Specified Feeding and Eating Disorders (OSFED). Mean SCS was higher in AN than in other diagnoses independent of classification. Differently from EDNOS, no high score was found in OSFED. BMI (OR 0.74, 95 % CI 0.56-0.98) and duration of amenorrhea >1 year (OR 6.63, 95 % CI 1.29-34.16) resulted significantly associated with the risk for medium-high SCS level in AN classified with DSM-5.
Conclusion: The results confirmed that DSM-5 reduces the number of EDNOS. DSM-5 seems to better represent the clinical picture in OSFED than in EDNOS. The clinical relevance of BMI and duration of amenorrhea should be considered even more now that they are no longer used as diagnostic hallmarks of AN.
Keywords: DSM-5; DSM-IV; Eating Disorders; Severity score; Somatic involvement.