Eating disorders (EDs) are culture-bounded syndromes of abnormal eating behavior, driven by psychopathology, primarily overvalued beliefs in the benefits of slimming, which lead to potentially serious physiologic abnormalities and regressive changes in social function. The essential features of the disorder are syndromic, a cluster of signs and symptoms with cross-sectional and longitudinal consistency, rather than a disorder understood by a single defined etiology. Despite its syndromic nature, the diagnostic criteria for the two best understood EDs, anorexia nervosa (AN) and bulimia nervosa (BN), are validated and widely accepted. Many cases do not meet the strict criteria for AN or BN. These cases are included within the third, residual category of EDs in the fourth Diagnostic Manual of the American Psychiatric Association (DSM-IV) as eating disorder not otherwise specified (EDNOS) or atypical cases.

The category of EDNOS is inherently troublesome for several reasons. Numerous eating disorder programs have reported that 25% to 60% of cases treated fall into the EDNOS category.<sup>2 Carlat D.J., Camargo C.A., Herzog D.B. Eating disorders in males: A report on 135 patients Am J Psychiatry 1997 ;  154 : 1127-1132

Cliquez ici pour aller à la section Références, 6 Mizes J.S., Sloan D.M. An empirical analysis of eating disorder, not otherwise specified: Preliminary support for a distinct subgroup Int J Eating Disorders 1998 ;  23 : 233-242  [cross-ref]

Cliquez ici pour aller à la section Références, 9 Williamson D.A., Gleaves D.H., Savin S.S. Empirical classification of eating disorder not otherwise specified: Support for DSM-IV changes J Psychopathol Behav Assessment 1992 ;  14 : 201-216  [cross-ref]

Cliquez ici pour aller à la section Références</sup> This number of atypical cases suggests that either the AN and BN diagnoses are too narrow, defined by research based criteria, or that additional categories are needed for cases now directed by exclusion into EDNOS. An EDNOS diagnosis often leads to uncertainty of therapeutic methods because clinicians who routinely treat AN and BN complain they feel confused about how to treat atypical eating disorder cases. Sadly, some managed care companies frankly state in advance of a diagnostic evaluation that if the patient is given a diagnosis of EDNOS, no coverage will be provided, the implication being these are not true eating disorders. EDNOS is problematic because it is too large a group, too heterogeneous, and its qualifications are more exclusionary than inclusionary. EDNOS causes confusion regarding treatment and is sometimes considered invalid as a diagnosis by third party payers.

We hypothesize that most EDNOS cases belong within the current AN or BN categories. This hypothesis can be tested by demonstrating how subgroups of the current EDNOS category meet the essential and critical diagnostic features of AN and BN. These features include similarity in age, length of illness, core psychopathology, response to treatment, and response to psychologic tests. The psychologic tests include both specific eating disorder instruments, such as the Eating Attitudes Test (EAT) and the Eating Disorder Inventory (EDI), as well as tests of general psychologic symptomatology, such as the Minnesota Multiphasic Personality Inventory (MMPI), and the Beck Depression Inventory (BDI).

A logical consequence of the critical examination of the current EDNOS admixture of cases is to suggest a revised set of diagnostic criteria for AN and BN. If EDNOS cases formerly excluded from AN or BN are subsumed into these categories, the diagnostic criteria for AN and BN must be revised to accommodate these immigrants. There will always be some cases that fall within the general category of an eating disorder because they lack some essential criterion. Williamson et al<sup>9 Williamson D.A., Gleaves D.H., Savin S.S. Empirical classification of eating disorder not otherwise specified: Support for DSM-IV changes J Psychopathol Behav Assessment 1992 ;  14 : 201-216  [cross-ref]

Cliquez ici pour aller à la section Références</sup> used cluster analysis to examine 46 EDNOS patients and found three distinct subgroups: subthreshold anorexia nervosa, bulimia nervosa nonpurging subtype, and binge-eating disorder.