Classification, Epidemiology and Treatment of DSM-5 Feeding and Eating Disorders

Hans W. Hoek


Curr Opin Psychiatry. 2013;26(6):529-531. 

The most discussed event in psychiatry in 2013 has been the publication of the Fifth Edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5),[1] which includes a substantially revised section on feeding and eating disorders. The five articles in the eating disorder section in this issue of the journal focus on the changes in DSM criteria and the implications for the classification, epidemiology, treatment and neurobiological research of feeding and eating disorders.

Throughout the DSM-5, the influence of development, sex and culture on the presentation of psychiatric disorders has been given more weight than any previous edition of the DSM ever did, and wherever possible it seeks to adhere to a lifespan approach.[1,2] One of the consequences of the lifespan approach is the elimination of the prior DSM-IV chapter 'Disorders Usually First Diagnosed During Infancy, Childhood or Adolescence.' As a result of the removal of this chapter, the two separate categories 'Feeding Disorders' and 'Eating Disorders' from DSM-IV are integrated into a single category called 'Feeding and Eating Disorders' in DSM-5.[3] In order to create a single cohesive category from two separate categories, the diagnoses, which were previously exclusively reserved for children, have been rearticulated and revised to tailor them toward adults as well as children.[3,4]

The DSM-5 chapter 'Feeding and Eating Disorders' includes the following diagnoses: pica, rumination disorder, avoidant/restrictive food intake disorder (ARFID), anorexia nervosa, bulimia nervosa and binge eating disorder (BED). Feeding disorders in DSM-IV already included pica and rumination disorder, but also feeding disorder of infancy or early childhood, which has been renamed ARFID and of which the criteria are significantly expanded.[1,3] The DSM-IV residual category 'eating disorder not otherwise specified' (EDNOS) has been renamed 'other specified feeding or eating disorder' (OSFED) in the DSM-5. The DSM-5 OSFED category specifies five disorders: atypical anorexia nervosa, bulimia nervosa of low frequency and/or limited duration, BED of low frequency and/or limited duration, purging disorder and night eating syndrome. The changes for feeding and eating disorder criteria from the DSM-IV to DSM-5 are described in the articles in this section on adolescents and adults by Call et al.[3] and on children by Bryant-Waugh[4] (Walsh, Attia and Bryant-Waugh were members of the DSM-5 Eating Disorders Work Group). In the third article in this section, Smink et al.[5] describe the epidemiology of eating disorders and the consequences of (re)applying the DSM-5 criteria to new and earlier data sets. EDNOS was a major concern for the DSM-5 Eating Disorders Work Group, because it was not only by far the most common DSM-IV eating disorder diagnosis in community samples, but it also accounted for more than half of the eating disorder cases in clinical settings.[5,6] The use of DSM-5 criteria will result in a large reduction in the proportion of EDNOS diagnoses by lowering the threshold for anorexia nervosa and bulimia nervosa and adding BED as a specific eating disorder. In consequence, the prevalence rates for anorexia nervosa and bulimia nervosa will increase, but course and outcome studies of both anorexia nervosa and bulimia nervosa indicate no significant differences between DSM-5 and DSM-IV definitions.[5] Both anorexia nervosa and bulimia nervosa are associated with increased mortality; however, data on long-term outcome, including mortality, for BED are scarce. For BED, most outcome data are derived from randomized clinical trials, whereas less is known about BED on a community level.[5]

Community studies show that most people with eating disorders do not receive any treatment at all, either because they are not detected, or because they refuse treatment because of shame or denial of their illness.[5,6] In the fourth article in this section, Kass et al.[7] describe recent findings regarding the psychological treatment of eating disorders, predominantly defined according to DSM-IV criteria. One of the reasons to promote BED to an official diagnosis in DSM-5 is that a diagnosis of BED has a predictive validity: both cognitive behavioral therapy (CBT) and interpersonal psychotherapy (IPT) have been shown to have a long-term efficacy for the treatment of adults with BED.[7] CBT and IPT also remain the most established treatments for bulimia nervosa, with stepped-care approaches showing promise.[7] Although no one specialist treatment has been shown to be superior for the treatment of adults with anorexia nervosa, a considerable proportion of patients improve with specialist psychological treatments; family-based therapy is the most established treatment for youth with anorexia nervosa.[7] The number of published articles contributing to an evidence base for the treatment of feeding and eating disorders in children remains disappointing.[4] Hopefully, the introduction of DSM-5 criteria will aid in capturing more childhood presentations and stimulating research in this relatively neglected field.

During the development of DSM-5, the question of whether we would be able to classify DSM-5 mental disorders on the basis of biomarkers was raised. After long discussion, it was decided that this might be possible for some disorders, for example, certain neurocognitive disorders, but that, given the current state of knowledge, this is not yet possible for most mental disorders. As regards biomarkers for anorexia nervosa and other eating disorders, which have been studied extensively,[8] we must also conclude that more work needs to be done. DSM-5 does introduce a biological parameter – not yet a biological marker – to serve as the basis for a severity rating for anorexia nervosa: the body mass index.

As DSM-5 is considered less useful for neurobiological research purposes than for clinical purposes, Insel et al. of the National Institute of Mental Health[9] have proposed a new framework for research on mental disorders: the Research Domain Criteria (RDoC). The RDoC project thus aims to bridge neuroscience and genetics to inform diagnostic classification and clinical management.[7–9] In the final article in this section on eating disorders, Van Elburg and Treasure[8] use the RDoC constructs across biological units of analysis to describe the advances in the neurobiology of eating disorders. Brain imaging techniques have become sophisticated enough to allow the identification of brain circuits pertaining to eating disorder behavior and to fundamental traits such as reward and social processing. Meanwhile, genetic studies have moved on from candidate gene studies to genome-wide association studies.[8] Hormonal changes, either resulting from starvation or serving as underlying factors for changes in eating behavior, are examined in both animal and human studies. Advances made in neuropsychology show that patients with eating disorders have problems in cognition, such as set shifting and central coherence and in other RDoC domains.[8] Attempts have been made to translate these findings into new forms of treatment, such as transcranial magnetic stimulation and deep brain stimulation. Interest in the addiction model of binge eating is increasing, leading to new approaches to treatment methods involving the use of opiate, dopamine and cannabinoid receptor blockers for BED.[8]

Many people, at least those involved in the development of DSM-5, such as myself, consider this version of the DSM system to be the best description available today of how mental disorders are expressed and can be recognized.[1,2] As usual with complex revision processes, the DSM-5 – including the criteria for eating disorders – has been criticized.[10–12] Allen Frances, chair of the taskforce for the DSM-IV, argues that the DSM-5 has increased the risk of defining variants of normal behavior as mental illness.[12] As an example, he mentions BED, which the DSM-IV classified in its Appendix of Conditions for Further Study. In my view, this is a common underestimation of the seriousness and disabling nature of the disorder for many BED patients. Laboratory-based studies show that the eating behavior and cognition of individuals with BED are quite different from weight-matched individuals without BED.[13] Moreover, more than 1000 studies on BED published in the last 2 decades indicate that BED is a valid and reliable diagnosis.[2,13,14]

The five review articles on feeding and eating disorders in this issue show that significant advances have been made in this field, especially with regard to adults and adolescents diagnosed with these disorders. Although there has also been an increase in interest and research on feeding disorders in children, this field is still in strong need of further development.