Differentiating constitutional thinness from anorexia nervosa in DSM 5 era

Abstract

Introduction: Constitutional thinness (CT) is an underweight state characterized by normal menstruations and no change in feeding behaviour. Thinness is the only resemblance between Anorexia Nervosa (AN) and CT. Removal of amenorrhea from the new DSM 5 definition of AN might result in misdiagnosis between these two populations. The objective of this study was to compare CT, AN and Control subjects in terms of biological, anthropometric, and psychological markers in order to better distinguish AN from CT subjects.

Materials and methods: Body composition, nutritional markers, pituitary hormones, bone markers and psychological scores were evaluated in three groups of young women: fifty-six CT, forty restrictive-type AN and fifty-four Control subjects. For every marker, a receiver Operator Characteristics (ROC) curve was calculated to evaluate the accuracy of differentiation between AN and CT groups.

Results: For most studied parameters, CT subjects were similar to Controls but dramatically different from AN subjects. DEBQ Restrained Eating subscale score was identified by ROC data analysis as the only psychological parameter tested to successfully differentiate AN from CT. Free-T3 and Leptin were shown to be powerful markers to differentiate AN and CT populations as they were highly specific and sensitive ones.

Conclusion: The exclusive use of psychological testing criteria is not always sufficient to differentiate AN and CT patients. Minimally, additional testing of Free T3 levels, which is cheap and widely accessible for general practitioners, should be completed to avoid misdiagnosis which could result in the implementation of ineffective treatment plans and social stigmatization for CT women.

Keywords: Anorexia nervosa; Biological differentiation markers; Constitutional thinness; DSM 5; Psychological questionnaires; ROC curves.