SEARCH

SEARCH BY CITATION

The classification of feeding and eating disorders in the ICD-10 and DSM-IV is unsatisfactory. The deficiencies of these systems are most evident in four facts. First, the majority of patients presenting with eating-related psychopathology do not fulfil criteria for a specific disorder and are classified in the residual “other” or “not otherwise specified” categories. Second, most individuals with an eating disorder sequentially receive several diagnoses instead of a single diagnosis that would describe the individual's problems at various developmental stages. Third, most recent clinical trials have used modified diagnostic criteria that may better reflect clinical practice, but deny the purpose of the classification as a means for communication between clinicians and researchers. Fourth, although childhood feeding disorders are typically described in the history of adolescents and adults with eating disorders, there is little research on the developmental continuity between childhood, adolescent and adult disorders that involve aberrant eating behaviours. Issues have also been raised about developmental and cultural dependencies of feeding and eating disorders as currently conceptualized.

Given these problems, it is not surprising that the World Health Organization (WHO) and the American Psychiatric Association are contemplating significant changes in classification. A number of proposals for changes have been made. The purpose of this article is to summarize the issues in the classification of feeding and eating disorders, review relevant aspects of evidence, and make proposals for modifications in the context of the development of ICD-11.

THE INTERNATIONAL CLASSIFICATION OF DISEASES

  1. Top of page
  2. THE INTERNATIONAL CLASSIFICATION OF DISEASES
  3. CLINICAL EVIDENCE AND DIAGNOSTIC CRITERIA
  4. DEVELOPMENTAL CONTEXT
  5. CULTURAL CONTEXT
  6. TEMPORAL INSTABILITY AND DIAGNOSTIC TRANSITIONS
  7. FEEDING DISORDERS
  8. EATING DISORDERS NOT OTHERWISE SPECIFIED (EDNOS)
  9. ANOREXIA NERVOSA
  10. BULIMIA NERVOSA
  11. BINGE EATING DISORDER
  12. EATING AND FEEDING DISORDERS DUE TO OTHER CONDITIONS WITH A KNOWN CA
  13. CONCLUSIONS
  14. Acknowledgements
  15. References

The primary purpose of the International Classification of Diseases (ICD) is to facilitate the work of health professionals in various clinical settings across the world. Therefore, the primary requisite for ICD diagnostic categories is clinical utility, and evidence from clinical and epidemiological research is given more weight than data from basic and etiological research 1. Attention is paid to global cross-cultural validity and the needs of health professionals from medium and low income countries 1.

Several conceptual directions have been proposed for the ICD-11 2. First, to reflect the growing evidence on continuity between child, adolescent and adult psychopathology, it has been proposed that the grouping of disorders with onset usually occurring in childhood and adolescence should be removed. Instead, disorders should be organized in groupings by psychopathology and a life-course approach should be adopted to conceptualize child, adolescent and adult manifestations of the same disorders.

Second, it has been agreed that the ICD-10 and DSM-IV contain an excessively large number of over-specified diagnoses, leading to artificially high rates of comorbidity and frequent use of the uninformative “not otherwise specified” and “other” categories 2. It has been proposed that evidence is required not just for changing or adding diagnostic categories but also for retaining existing ones. The overuse of the “not otherwise specified” categories should be reduced by revising the boundaries of specific disorders to include most clinically significant presentations.

Third, to best serve the clinical use, the ICD takes a prototypic approach in which presentations characteristic of each diagnostic category are described in a narrative format, which most health professionals find easier to use in practice 3,4. The ICD avoids the use of exact count, frequency and duration criteria to modulate diagnostic thresholds. Since most duration criteria for various disorders are not based on evidence and are difficult to memorize and apply, it has been proposed that a uniform duration criterion of four weeks should be adopted, with qualified exceptions for disorders which require rapid clinical attention (e.g., delirium, mania and catatonia) or that manifest by relatively brief events (e.g., intermittent explosive disorder).

Fourth, it has been proposed that categories with some evidence of clinical usefulness, but insufficient evidence for validity of specific criteria, should be included in the main body of the ICD, but signposted as categories that require further testing.

Fifth, to reflect the evidence that most mental disorders are multifactorial, it is proposed to remove the distinction between organic and functional forms of disorders.

CLINICAL EVIDENCE AND DIAGNOSTIC CRITERIA

  1. Top of page
  2. THE INTERNATIONAL CLASSIFICATION OF DISEASES
  3. CLINICAL EVIDENCE AND DIAGNOSTIC CRITERIA
  4. DEVELOPMENTAL CONTEXT
  5. CULTURAL CONTEXT
  6. TEMPORAL INSTABILITY AND DIAGNOSTIC TRANSITIONS
  7. FEEDING DISORDERS
  8. EATING DISORDERS NOT OTHERWISE SPECIFIED (EDNOS)
  9. ANOREXIA NERVOSA
  10. BULIMIA NERVOSA
  11. BINGE EATING DISORDER
  12. EATING AND FEEDING DISORDERS DUE TO OTHER CONDITIONS WITH A KNOWN CA
  13. CONCLUSIONS
  14. Acknowledgements
  15. References

The most important reason against changing the current diagnostic criteria is that it could invalidate available evidence. It is therefore important to assess the clinically relevant evidence and its relationship to classification. We have reviewed recent clinical trials on treatments of eating disorders published in six influential child and general psychiatry journals (Journal of the American Academy of Child and Adolescent Psychiatry, Journal of Child Psychology and Psychiatry, American Journal of Psychiatry, Archives of General Psychiatry, British Journal of Psychiatry and Psychological Medicine) between January 2000 and May 2011.

We identified 18 clinical trials (Table 1). Seven trials tested treatments for anorexia nervosa. Of these, three (published between 2000 and 2003) used “strict” DSM-IV or ICD-10 criteria. Four more recent trials (published between 2005 and 2010) used broader criteria, relaxing the weight and/or the amenorrhoea criterion. Eleven trials tested treatments for bulimia and related conditions. Three of these (published between 2000 and 2003) applied “strict” DSM-IV or DSM-III-R criteria. Eight of these trials (published between 2001 and 2009) used broader criteria, including bulimic-type eating disorders not otherwise specified or all eating disorders without underweight in addition to bulimia nervosa.

Table 1.  Diagnostic inclusion criteria in recent clinical trials in eating disorders
AuthorYearTreatmentComparatorNAge groupTarget groupInclusion diagnostic criteria
Anorexia nervosa and related conditions
Lock et al 52010Family therapyIndividual therapy121AdolescentsAnorexia nervosaDSM-IV excluding the amenorrhea requirement
Loeb et al 62007Family therapy 20AdolescentsAnorexia nervosaAnorexia nervosa or subthreshold anorexia nervosa
Lock et al 72005Family therapy (short)Family therapy (long)86AdolescentsAnorexia nervosaSlightly relaxed weight and amenorrhoea criteria
McIntosh et al 82005CBT/interpersonalClinical management56AdultsAnorexia nervosaStrict or lenient weight criteria
Pike et al 92003CBTNutritional counselling33AdultsAnorexia nervosaDSM-IV
Dare et al 102001Psychoanalytic, cognitive analytic, and family therapyClinical management84AdultsAnorexia nervosaDSM-IV
Eisler et al 112000Multifamily therapyFamily therapy40AdolescentsAnorexia nervosaDSM-IV or ICD-10
Bulimia nervosa and related conditions
Fairburn et al 122009CBTWait list AdultsEating disorderAny ED with BMI>17.5
Schmidt et al 132008Computerized CBTWait list96AdultsBulimia nervosaBulimia nervosa or bulimic-type EDNOS
Schmidt et al 142007Family therapyCBT85AdolescentsBulimia nervosaBulimia nervosa or bulimic-type EDNOS
Le Grange et al 152007Family therapySupportive therapy AdolescentsBulimia nervosaBulimia nervosa or ‘partial’ bulimia nervosa
Banasiak et al 162005Guided self-helpWait list109AdultsBulimia nervosaFull syndrome or sub-threshold
Walsh et al 172004FluoxetineSelf-help91AdultsBulimia nervosaSubjective binges and 1x/week frequency
Carter et al 182003Self-help CBTWait list85AdultsBulimia nervosaDSM-IV
Palmer et al 192002Self-help CBTWait list121AdultsBulimia disordersBulimia nervosa, partial bulimia nervosa or binge-eating disorder
Hsu et al 202001Cognitive therapyNutritional counselling100AdultsBulimia nervosaDSM-III-R
Safer et al 212001Dialectical behaviour therapyWait list31AdultsBulimia nervosaOne binge/purge episode per week
Agras et al 222000CBTInterpersonal therapy220AdultsBulimia nervosaDSM-III-R
CBT - Cognitive behaviour therapy; BMI - body mass index; ED - eating disorder; EDNOS - eating disorder not otherwise specified

We conclude that the clinical trial literature reflects the deficiencies of the current diagnostic systems by broadening the diagnostic criteria in attempts to reflect clinical reality. No clinical trial published in the last seven years in the six journals used DSM-IV or ICD-10 criteria exactly. The result is that inclusion criteria differ between trials and the classification has effectively lost its purpose in defining the same group of patients across research studies and clinical settings. We conclude that changes in classification will not invalidate useful evidence, because most recent evidence is based on modified diagnostic criteria.

DEVELOPMENTAL CONTEXT

  1. Top of page
  2. THE INTERNATIONAL CLASSIFICATION OF DISEASES
  3. CLINICAL EVIDENCE AND DIAGNOSTIC CRITERIA
  4. DEVELOPMENTAL CONTEXT
  5. CULTURAL CONTEXT
  6. TEMPORAL INSTABILITY AND DIAGNOSTIC TRANSITIONS
  7. FEEDING DISORDERS
  8. EATING DISORDERS NOT OTHERWISE SPECIFIED (EDNOS)
  9. ANOREXIA NERVOSA
  10. BULIMIA NERVOSA
  11. BINGE EATING DISORDER
  12. EATING AND FEEDING DISORDERS DUE TO OTHER CONDITIONS WITH A KNOWN CA
  13. CONCLUSIONS
  14. Acknowledgements
  15. References

An important issue is the relationship between feeding and eating disorders. Feeding problems and selective eating in childhood have been described in the history of patients with eating disorders since the early case reports 23, but there has been little research on the continuity between feeding and eating disorders. The available research suggests a degree of continuity of eating problems from infancy to adulthood 24,25,26. For example, in a large prospective study, feeding problems in infancy and undereating in childhood predicted anorexia nervosa in adulthood with odds ratios of 2.6 and 2.7 respectively 26. For bulimia nervosa, the evidence is limited to a retrospective study showing that history of overeating and rapid eating in childhood was more common in women with bulimia nervosa than in their unaffected sisters 25. However, long-term follow-ups of individuals diagnosed with feeding disorders in childhood are lacking. At the same time, clinical trends indicate that the boundary between feeding disorders of childhood and eating disorders is problematic. On the one hand, there is a trend for younger and younger children to present with symptoms resembling “adult” eating disorders 27. On the other hand, many adults presenting with underweight, restrictive and selective eating lack the typical body-weight and shape related psychopathology that characterizes eating disorders and may be better described by criteria of feeding disorders 28,29,30,31.

It has been pointed out that similarity between child and adult manifestations of eating-related psychopathology might have been obscured by the fact that existing criteria are rigidly applied without sensitivity to developmental stage 28,32,33,34. This is most apparent in the requirement for self-reported cognitions regarding weight, shape and body image. It has been argued that children and some adolescents may not be able to formulate and communicate such concerns due to incompletely developed capacity for abstract thinking 28,32,33,34. It has been proposed that behavioural indicators of such concerns should be accepted as a basis for diagnosis, whether they are observed by clinicians or reported by parents, teachers or other adults 32,35,36. For example, observation of the child frequently checking her/his weight and shape or expressing aspects of shape/weight-related self-image in drawings might be taken into account as indicators of preoccupation with weight and shape in the context of pathological eating behaviours. In the case of anorexia nervosa, it has also been suggested that restrictive and binge-purge subtypes often represent developmental stages of the same disorder — children and younger adolescents usually present with the restrictive type, and binge-purging behaviours develop in a proportion of individuals at later stages 32,35,37.

The summary of evidence suggests that a single classification applied across age groups and sensitive to developmentally specific manifestations would more accurately describe the course of these disorders and reflect the continuity between child, adolescent and adult manifestations than the current system.

CULTURAL CONTEXT

  1. Top of page
  2. THE INTERNATIONAL CLASSIFICATION OF DISEASES
  3. CLINICAL EVIDENCE AND DIAGNOSTIC CRITERIA
  4. DEVELOPMENTAL CONTEXT
  5. CULTURAL CONTEXT
  6. TEMPORAL INSTABILITY AND DIAGNOSTIC TRANSITIONS
  7. FEEDING DISORDERS
  8. EATING DISORDERS NOT OTHERWISE SPECIFIED (EDNOS)
  9. ANOREXIA NERVOSA
  10. BULIMIA NERVOSA
  11. BINGE EATING DISORDER
  12. EATING AND FEEDING DISORDERS DUE TO OTHER CONDITIONS WITH A KNOWN CA
  13. CONCLUSIONS
  14. Acknowledgements
  15. References

Eating plays an important role in most cultures. Acceptable eating habits vary widely between religious and ethnic groups, and eating disorders have been conceptualized as culture-bound syndromes 38. In this context, it is notable that most published research is based on North American and European populations. In the last decade, reports on eating disorders and related conditions from various countries, including low income countries and countries undergoing sociocultural transitions 39,40,41, have accumulated which may inform a classification that is sensitive to local variation 42.

Anorexia nervosa occurs in all cultures, but the incidence is higher among individuals who have been exposed to Western culture and values and those who live in relative affluence 40,41,43. For example, in the Caribbean island of Curaçao, all identified cases of anorexia nervosa were among young women of mixed ethnicity who had spent time in the USA or the Netherlands; there were no cases of anorexia nervosa among the majority of young women in the island, who are black and had not been abroad 40,44. Anorexia nervosa is relatively rare among black women in Africa, the Caribbean, and the USA 45,46,47. In the Czech Republic, the incidence of anorexia nervosa increased sharply after the fall of the iron curtain, that was associated with exposure to Western-style media and values 41.

In addition to influence on prevalence, culture also shapes the manifestation of anorexia nervosa. For example, in South-East Asia, a larger proportion of patients with anorexia nervosa report abdominal discomfort and other factors as a rationale for restrictive eating 28,48. However, typical presentations with weight and shape-related preoccupations and fear of gaining weight have also been recorded in most non-Western cultures 28,48,49,50, and the rates of full-syndrome anorexia nervosa in South East Asia are intermediate between Western countries and African populations 50. There is evidence that patients who initially present with other rationales often develop intense fear of weight gain 51 and that the proportion of patients reporting fear of weight gain increases with exposure to Western cultural values 52. This suggests that weight-phobic and non-weight-phobic anorexia are context-dependent manifestations of the same disorder. Therefore, it is recommended that fear of weight gain is not required for the diagnosis of anorexia nervosa, provided that behaviours maintaining underweight or other psychopathology suggestive of eating disorder are present.

Bulimia nervosa has been conceptualized as strongly bound to Western culture 38. The disorder is more common among individuals who were exposed to Western culture and who grew up in relative affluence 38,41,43. Although all component symptoms of bulimia nervosa occur in non-Western low income countries, the syndrome appears to be less common in those countries than in North America and Western Europe 43,49,50,53. The incidence of bulimia nervosa increases in parallel with exposure to Western media and values and correlates with the degree of acculturation 41,43,52,54. Therefore, the manifestation of bulimia nervosa and its separation from normality have to be considered within cultural context. For example, culturally sanctioned feasting followed by the use of indigenous purgatives in Pacific islands should not be medicalized, but the use of the same herbal purgatives in the context of typical psychopathology and outside the culturally sanctioned events is a symptom of an eating disorder 39,55. The motives for pursuing a thin body shape may also depend on socioeconomic context. For example, in societies undergoing socioeconomic transition, a thin body can be perceived as a valuable commodity that may help obtain a lucrative job and guarantee career success 42,56,57. There is little evidence on whether such cultural variations in manifestation have an impact on the long-term prognosis and treatment response. In the USA, patients with bulimia nervosa belonging to ethnic minorities appear to respond to the same psychological treatments as European Americans 58.

Binge eating disorder is relatively equally distributed across countries and ethnic groups, but details of manifestation vary in culture-dependent manner. Black women with binge eating disorder are on average heavier, have fewer concerns related to body weight, shape and eating, a less frequent history of bulimia nervosa, but similar levels of depressive symptoms and impairment compared to white women with the same diagnosis 59. In general, the associations between binge eating, obesity, weight and shape dissatisfaction, and general psychopathology hold across ethnic groups 60,61. While no modifications of diagnostic criteria are required, the lower rates of treatment among black women with binge eating disorder suggest that increased alertness of clinicians to eating disorders in non-European ethnic groups is warranted 59.

TEMPORAL INSTABILITY AND DIAGNOSTIC TRANSITIONS

  1. Top of page
  2. THE INTERNATIONAL CLASSIFICATION OF DISEASES
  3. CLINICAL EVIDENCE AND DIAGNOSTIC CRITERIA
  4. DEVELOPMENTAL CONTEXT
  5. CULTURAL CONTEXT
  6. TEMPORAL INSTABILITY AND DIAGNOSTIC TRANSITIONS
  7. FEEDING DISORDERS
  8. EATING DISORDERS NOT OTHERWISE SPECIFIED (EDNOS)
  9. ANOREXIA NERVOSA
  10. BULIMIA NERVOSA
  11. BINGE EATING DISORDER
  12. EATING AND FEEDING DISORDERS DUE TO OTHER CONDITIONS WITH A KNOWN CA
  13. CONCLUSIONS
  14. Acknowledgements
  15. References

Longitudinal follow-up studies of anorexia and bulimia nervosa have found that a significant proportion of subjects change diagnostic status to another eating disorder 62,63,64,65,66,67. Diagnostic crossovers are more common in the initial years of illness and follow a predictable sequence. Typically, restrictive anorexia nervosa mutates into binge eating/purging anorexia nervosa, before crossing over to bulimia nervosa 68,69,70,71,72. Crossover in the opposite direction is less common. While one-third of individuals with an initial diagnosis of anorexia nervosa develop bulimia nervosa during a five-to-ten year follow-up, only 10–15% of those with an initial diagnosis of bulimia nervosa develop anorexia 68,70,72. Larger proportions of subjects with an initial diagnosis of bulimia nervosa develop binge eating disorder or eating disorder not otherwise specified (EDNOS) 65,68. There are also numerous transitions between specific eating disorder categories and EDNOS, with the latter often representing an intermediate state on the way to recovery 68,73,74.

The diagnostic transitions may also extend to a relationship between feeding disorders in childhood and eating disorders in adolescence and adulthood. Restrictive eating and hyperactivity are often present in children and adolescents who deny any motivation of these behaviours by fear of gaining weight, but who later demonstrate weight phobia and receive a diagnosis of an eating disorder 28,33,35.

Importantly, a significant minority of cases show repeated diagnostic crossovers. For example, half of those who transit from an initial anorexia nervosa to bulimia nervosa experience a “recurrence” of anorexia nervosa within a few years 70. In long-standing eating disorders, diagnostic transitions are the rule, with most patients who remain ill for at least several years changing diagnostic status one or more times 68,70,72. Comorbid depression and alcohol abuse are associated with more diagnostic instability in eating disorders 68. With these rates of transitions, it is clear that the sequential diagnoses represent stages of the same disorder rather than separate disorders.

The apparent sequential comorbidity of various eating disorders is probably an artefact of applying a system of overly specified diagnostic categories with overlapping psychopathology. In ICD-10 and DSM-IV, the various eating disorder categories are mutually exclusive, so they cannot be diagnosed at the same time. However, there is no such restriction for sequential diagnoses, and neither ICD-10 nor DSM-IV takes the longitudinal course of psychopathology into account. This state of affairs is clearly unsatisfactory. On the one hand it creates an impression of an overly complex pattern of sequential comorbidity, on the other hand it misses important prognostic information. For example, it was shown that, among patients with current bulimia nervosa, a history of anorexia nervosa is associated with reduced chance of recovery and much larger risk of transiting into anorexia nervosa 75. It has been proposed that bulimia nervosa should be subtyped according to history of anorexia nervosa 75. A more radical solution to the problem of spurious sequential comorbidity may require restrictions on frequent changes of diagnostic categories (e.g., the diagnosis of anorexia nervosa may be retained for a year after weight normalization) or establishing a combined eating disorder category to capture cases that sequentially fulfil criteria for both anorexia and bulimia nervosa and have a tendency to repeatedly change presentations.

FEEDING DISORDERS

  1. Top of page
  2. THE INTERNATIONAL CLASSIFICATION OF DISEASES
  3. CLINICAL EVIDENCE AND DIAGNOSTIC CRITERIA
  4. DEVELOPMENTAL CONTEXT
  5. CULTURAL CONTEXT
  6. TEMPORAL INSTABILITY AND DIAGNOSTIC TRANSITIONS
  7. FEEDING DISORDERS
  8. EATING DISORDERS NOT OTHERWISE SPECIFIED (EDNOS)
  9. ANOREXIA NERVOSA
  10. BULIMIA NERVOSA
  11. BINGE EATING DISORDER
  12. EATING AND FEEDING DISORDERS DUE TO OTHER CONDITIONS WITH A KNOWN CA
  13. CONCLUSIONS
  14. Acknowledgements
  15. References

Feeding disorders of infancy/childhood and eating disorders of adolescence/adulthood are classified in different sections of ICD-10 and DSM-IV. Feeding disorders of childhood/infancy include refusal of food and selective (faddy) eating, regurgitation of food with or without re-chewing, or eating of non-edible substances (pica). This classification has serious problems as it includes one heterogeneous condition with unclear boundaries from eating disorders and two less common and more specific conditions that occur in both children and adults 76,77.

Alternative classifications of feeding disorders have been proposed, that attempt to reconceptualize this heterogeneous category as four to six specific categories 78,79,80 and/or emphasize the relational context of feeding and the role of primary caregiver 80,81. The problems are that the specific subtypes leave a large number of clinical presentations unclassified and some require allocation of a single etiology to disorders that are multifactorial. As a result, none of these proposals has been accepted. The situation is clearer for pica and regurgitation, which are relatively distinct syndromes. However, both of these are frequent in adults and in the context of other mental disorders (e.g., autism and learning disability) and physiological conditions (iron deficiency, pregnancy). It is proposed that pica and regurgitation disorder should be diagnosed based on behaviour and irrespective of age 76.

Avoidant/restrictive food intake disorder (ARFID)

Avoidant/restrictive food intake disorder (ARFID) has been proposed to replace the non-specific category of feeding disorder of infancy or early childhood and, in addition, explicitly include adolescent and adult cases presenting with psychologically motivated inadequate food intake for reasons related to the physical properties or feared consequences of eating specific types of food other than effects on body weight and shape 76.

ARFID overlaps with anorexia nervosa in terms of restrictive food intake and the resulting underweight, but differs in psychopathology and motives for restrictive eating. These include avoiding types of food of specific colour or texture or limiting food intake to a small number of specific “safe” types of food because of perceived health consequences. ARFID is typically not associated with gross disturbance of body image.

Since there is vast normal variation in eating habits among children and adults, differentiation from normality is important. As a rule, ARFID should only be diagnosed when the restrictive/avoidant eating is a cause of inadequate nutrition that may be associated with delayed growth in children, weakness, anaemia or other medical consequences in any age group, or inadequate development of the foetus in pregnant women. Dietary practices that are endorsed by large groups of people, such as vegetarianism or religious fasting, do not constitute a basis for diagnosing ARFID.

The proposed ARFID category conforms with the general direction of merging feeding and eating disorders and opening diagnostic categories to all age groups. It also provides an appropriate category for some cases that previously received the diagnosis of EDNOS. Since ARFID is a new concept, we propose that it is included in ICD-11 as a category that requires further testing. In particular, the boundaries between ARFID and anorexia nervosa, including the culturally determined non-fat phobic presentations, and between ARFID and specific phobias need to be explored. It is also hoped that the inclusion of ARFID will stimulate research on continuity between child and adult presentations.

Pica

Pica describes persistent eating of non-food substances, such as earth, chalk, metal or plastic objects, hair or faeces. In ICD-10 and DSM-IV, pica is included among the disorders with onset usually occurring in childhood and adolescence, and adult presentations are coded elsewhere (e.g., as an EDNOS). Since pica frequently first comes to attention in adulthood, it is proposed that there are no age restrictions for its diagnosis 76. Pica only requires a diagnosis if it is severe, leads to adverse consequences (e.g., heavy-metal poisoning or parasite infestation), does not exclusively occurs during socially sanctioned events (e.g., religious rituals) or culturally accepted practice (e.g., eating of clay in Nigeria) and is not fully explained by another mental disorder (e.g., psychosis or obsessive-compulsive disorder). Eating non-food substances also frequently occurs in pregnancy, and pica should not be diagnosed in pregnant women unless it is of unusual extent or causes health concerns.

Regurgitation disorder

Regurgitation disorder describes repetitively bringing previously swallowed food from stomach back to the mouth and either spitting or re-chewing it. It was included in ICD-10 and DSM-IV under the name of rumination disorder. Since rumination is commonly used to describe a psychological process involving repetitive thoughts, we propose changing the name to “regurgitation disorder” to avoid confusion. Regurgitation disorder was previously classified in the section of disorders with onset specific to childhood, but it frequently first comes to attention in adulthood. Therefore, it should be diagnosed without age restrictions 76.

EATING DISORDERS NOT OTHERWISE SPECIFIED (EDNOS)

  1. Top of page
  2. THE INTERNATIONAL CLASSIFICATION OF DISEASES
  3. CLINICAL EVIDENCE AND DIAGNOSTIC CRITERIA
  4. DEVELOPMENTAL CONTEXT
  5. CULTURAL CONTEXT
  6. TEMPORAL INSTABILITY AND DIAGNOSTIC TRANSITIONS
  7. FEEDING DISORDERS
  8. EATING DISORDERS NOT OTHERWISE SPECIFIED (EDNOS)
  9. ANOREXIA NERVOSA
  10. BULIMIA NERVOSA
  11. BINGE EATING DISORDER
  12. EATING AND FEEDING DISORDERS DUE TO OTHER CONDITIONS WITH A KNOWN CA
  13. CONCLUSIONS
  14. Acknowledgements
  15. References

The field of eating disorders has evolved around the concepts of anorexia and bulimia nervosa. However, when DSM-IV criteria are applied to patients presenting to eating disorders services, the most common diagnosis is EDNOS. The proportion of patients diagnosed with EDNOS is consistent across settings and age groups, with approximately 60% of patients referred to child, adolescent or adult eating disorders services classified as EDNOS 82,83,84.

In ICD-10, EDNOS is split between atypical anorexia nervosa, atypical bulimia nervosa, other eating disorder, and eating disorder, nonspecified. However, this does not improve the matters, since the diagnoses of atypical anorexia and atypical bulimia are highly unreliable and nearly 40% of patients still receive the “other” or “unspecified” residual diagnoses 82. There is also confusion about the meaning of “atypical anorexia nervosa”. In ICD-10, atypical anorexia nervosa is defined as a condition that resembles anorexia nervosa, but does not fulfil all of the diagnostic criteria (e.g., there is no amenorrhoea or not a sufficient degree of weight loss). In the literature, the term “atypical anorexia nervosa” is used more narrowly to describe an eating disorder with significant underweight but no concerns about body shape or weight 85,86. For these reasons, it is concluded that ICD-10 is as unsatisfactory as DSM-IV.

The cause of the current problems is that criteria for specific eating disorders, such as anorexia and bulimia nervosa, are too narrow and rigid. Indeed, the definitions of eating disorders are somewhat anomalous by requiring the presence of all criteria, rather than a minimum number of a longer list of symptoms. For example, the diagnosis of depressive episode requires five out of nine symptoms, but the diagnosis of anorexia nervosa requires four out of four (underweight, fear of weight gain, body image disturbance and amenorrhoea). This rigid requirement means that cases where any one of the four symptoms is absent are diagnosed as EDNOS, even if the presentation otherwise resembles typical anorexia nervosa. A significant proportion of EDNOS is composed of such subthreshold cases. Common types of EDNOS include anorexia nervosa without amenorrhoea, anorexia nervosa without fear of gaining weight, anorexia nervosa not fulfilling the weight loss criterion, bulimia nervosa with binges that are not objective (i.e., do not involve an amount of food that is definitely larger than what most people would eat), bulimia without bingeing (purging disorder), bulimia without purging (binge eating disorder), and bulimia nervosa with bingeing and compensatory behaviours occurring less than twice a week for three months 29,84. For example, in a specialized clinic for eating disorders, subthreshold anorexia and bulimia nervosa accounted for 83% of EDNOS 84.

Having most patients diagnosed with EDNOS is unhelpful for several reasons. First, EDNOS is a highly heterogeneous group in terms of presentation, prognosis, physical sequelae, and treatment outcomes 37,84. Second, most cases of EDNOS represent a phase of an eating disorder in individuals who fulfilled criteria for anorexia or bulimia nervosa at other times 29,73. Third, the residual character of this category implicitly conveys that it is a less severe disorder. This contrasts with the demonstrated severity, impairing character, and a grave prognosis including elevated mortality in EDNOS 87,88. Finally, EDNOS has no specific implications for treatment selection or service provision. There has been an attempt to introduce a “transdiagnostic” treatment that would be effective for most types of eating disorders, but the published evidence on this treatment excludes eating disorders with underweight and subsequently better fits a broader category of bulimia nervosa 12. A similar approach has been taken in recent clinical trials in bulimia nervosa, which also included bulimia-like EDNOS 13,14,15. We conclude that continued widespread use of EDNOS would be unhelpful and potentially harmful.

A number of proposals have been made to reduce the reliance on EDNOS. Most proposals include broadening of diagnostic criteria for anorexia and bulimia nervosa so that they subsume a proportion of cases currently diagnosed with EDNOS. The proposed expansions of criteria for anorexia nervosa include: dropping the requirement for amenorrhoea, relaxing the underweight criterion, and relaxing the requirement for fear of weight gain. The proposed broadening of criteria for bulimia nervosa include reducing the frequency requirement for bingeing and compensatory behaviour, and dropping the binge “objectivity” criterion (requirement that a binge involves eating an unusually large amount of food). Preliminary studies in adolescents and adults indicate that dropping the binge objectivity criterion has the greatest impact on reducing the use of the residual category 29,84. Depending on the degree of broadening, the extended categories can reduce the use of EDNOS moderately 68,89 or nearly eliminate its use 90.

The other set of proposals include introducing more specific diagnostic categories. There is most support for the introduction of binge eating disorder 91,92. Other proposed additional categories include purging disorder 93 and night eating syndrome 94. A mixed category of eating disorder with features of both anorexia and bulimia nervosa has been proposed to capture the cases that do not easily fit the prototype of one specific eating disorder 95. Finally, there is a proposal that includes relaxing the strict requirement that all defining symptoms must be present to make a diagnosis and suggests a list of symptoms of which one or more needs to be present to fulfil a diagnostic criterion 96.

While all of the above proposals appear sensible and each would reduce the need for EDNOS, it is difficult to estimate the impact of combining aspects of the various proposals. There are also problems with adopting some of the proposals entirely. For example, the alternative classification system proposed by Hebebrand and Bulik 96 represents conceptually the strongest departure from current classification, by removing the requirement for all symptoms to be present, but it only concerns anorexia nervosa and it allows relatively nonspecific symptoms such as irritability and depressed mood to count towards the diagnosis, raising concerns about how well it would differentiate against disorders from other groupings. On balance, we favour the combination of introducing binge eating disorder as an additional specific category with substantial broadening of diagnostic criteria, similar to the proposal by Walsh and Sysko 97, which has been shown to nearly eliminate the use of EDNOS 90 and is consistent with recent trends in clinical trial literature (Table 1).

It is unclear whether purging disorder or night eating syndrome will still be needed when the criteria for anorexia nervosa, bulimia nervosa and binge eating disorder are broadened. The available evidence suggests that most individuals with purging disorder have subjective binges, therefore the removal of requirement for an unusually large amount of food to be eaten during a binge may be sufficient to classify most of these subjects as bulimia nervosa, which does not differ from purging disorder in terms of impairment and response to treatment 12,98. Most cases of night eating syndrome can be subsumed under the broad category of binge eating disorder. The one remaining issue concerns cases that frequently change presentation and sequentially fulfil criteria for different eating disorders. Following suggestions by Fairburn 95, we support the use of a mixed category. However, to avoid redundancy when implemented alongside broader specific categories, we propose reserving such category for the relatively severe cases that fulfil criteria for both anorexia and bulimia nervosa either concurrently or sequentially. This category may be more appropriately called “combined” rather than “mixed” eating disorder.

ANOREXIA NERVOSA

  1. Top of page
  2. THE INTERNATIONAL CLASSIFICATION OF DISEASES
  3. CLINICAL EVIDENCE AND DIAGNOSTIC CRITERIA
  4. DEVELOPMENTAL CONTEXT
  5. CULTURAL CONTEXT
  6. TEMPORAL INSTABILITY AND DIAGNOSTIC TRANSITIONS
  7. FEEDING DISORDERS
  8. EATING DISORDERS NOT OTHERWISE SPECIFIED (EDNOS)
  9. ANOREXIA NERVOSA
  10. BULIMIA NERVOSA
  11. BINGE EATING DISORDER
  12. EATING AND FEEDING DISORDERS DUE TO OTHER CONDITIONS WITH A KNOWN CA
  13. CONCLUSIONS
  14. Acknowledgements
  15. References

Anorexia nervosa is the prototypical eating disorder that has been consistently described since the 19th century 99,100. In ICD-10 and DSM-IV, it is defined by four criteria, all of which are required for diagnosis. Three of these are essentially the same in the two systems: low body weight (maintained at body mass index, BMI<17.5 or under 85% of weight expected for height, age and sex), body image disturbance and amenorrhoea. The fourth criterion differs between the two classifications: in ICD-10, it is a requirement that low weight is self-induced through dietary restriction and/or purging behaviour; in DSM-IV, intense fear of gaining weight or becoming fat is required. In DSM-IV, anorexia nervosa is further divided into restrictive and binge-eating/purging subtypes.

These classifications have been criticized for several reasons. First, the requirement of all four criteria excludes a significant proportion of clinical presentations that fit the prototype of anorexia nervosa. The requirements of amenorrhoea, low body weight threshold, and fat phobia have all been criticized on this ground. Second, these criteria are insensitive to cultural variations; fear of gaining weight is reported by a smaller proportion of cases in non-Western cultures. Third, it has been argued that extant criteria omit hyperactivity, which is a salient feature in the presentation of anorexia nervosa 96. The problems with present diagnostic criteria are reflected in the fact that most recent clinical trials have used broader inclusion criteria, mostly relaxing the weight and amenorrhoea criteria (Table 1). We will first review the rationale for retaining or dropping each specific criterion and then move to the diagnostic concept and proposals for revision.

Amenorrhoea, defined as absence of three consecutive menstrual bleedings, is a common feature of anorexia nervosa that is strongly associated with underweight and excessive exercise and may have prognostic implications for bone mineral density and fertility outcomes. However, several facts make amenorrhoea problematic as a diagnostic criterion. First, it is not applicable to girls prior to menarche, to post-menopausal women, women taking hormonal preparations and men. The ICD-10 has proposed a male equivalent of hormonal disturbance manifest as “loss of sexual interest and potency”, which is however rarely assessed or researched and whose contribution to the diagnosis is unclear. Second, a significant minority of women (5 to 25% in clinical samples) who otherwise fulfil criteria for anorexia nervosa and require clinical attention menstruate and consequently are classified as EDNOS. Third, the requirement of missing three consecutive menstrual bleedings interferes with timely diagnosis and treatment. There is a broad agreement that amenorrhoea should not be required for the diagnosis of anorexia nervosa, but should be recorded since it may be an indicator of severity and may help distinguish between constitutional thinness and anorexia nervosa 101.

The low weight criterion is the defining feature of anorexia nervosa, but its exact specification has been discussed. A number of women who fit the prototype of anorexia nervosa narrowly miss the weight criterion of BMI<17.5 (or body weight less than 85% of what is expected for age and height) and are classified as EDNOS. Depending on body constitution, the above threshold for underweight may be seen as too high or too low in individual cases. Therefore, it has been proposed that this threshold should be relaxed and/or left to clinical judgement (www.dsm5.org). This proposal has the advantage of reducing the use of the uninformative EDNOS, but risks loss of objectivity and important information. Severe underweight has been repeatedly shown to be a strong predictor of poor prognosis and mortality 102,103,104,105 and is routinely used as an indication for inpatient treatment. We propose relaxing the underweight criterion to the WHO definition of underweight (BMI<18.5) with a room for clinical judgement, and recording a history of severe underweight (e.g., BMI<14.0) as a severity qualifier (with dangerously low body weight). Low body weight is only considered a symptom of anorexia nervosa if it is due the individual's eating behaviour rather than factors such as a medical condition or unavailability of food.

Morbid fear of gaining weight is required for diagnosing anorexia nervosa in the DSM-IV. In the ICD-10, fear of fatness is included under the body-image distortion criterion. This has been perhaps the most discussed criterion. The requirement of fear of fatness is problematic from both a developmental and a cultural perspective. Developmentally, fear of gaining weight is rarely reported by children and may require an ability of abstract reasoning that only develops during adolescence 32. Cross-culturally, females with otherwise typical anorexia nervosa in non-Western countries less frequently report fear of fatness as the reason for self starvation 48,106. Even in Western countries, a significant minority of patients report no fear of weight gain and are consequently classified as EDNOS 106,107,108. It has also been pointed out that fat phobia often emerges during weight restoration in patients who previously denied any fear of gaining weight 51. The dependency on development, culture and illness stage argues against the usefulness of weight phobia as a diagnostic criterion. However, it has also been argued that fear of weight gain is part of the core defining psychopathology of anorexia nervosa 109. We propose extending this criterion to include preoccupation with body weight and shape, preoccupations with food and nutrition, and persistent behaviours that are intended to reduce energy intake or increase energy expenditure.

Body image disturbance is a striking aspect of anorexia nervosa and is required in both ICD-10 and DSM-IV. It includes both the perception of own body or its parts as larger than they are and the lack of recognition of the seriousness of underweight. Body image disturbance often precedes other symptoms and its persistence upon weight recovery has prognostic significance. Only minor rephrasing of this criterion has been recommended for DSM-5 and we agree with retaining this criterion essentially unchanged. Cases of restrictive eating with no body image related psychopathology may better be classified as ARFID.

Hyperactivity is a remarkable feature of many cases of anorexia nervosa and is a differentiator from other causes of starvation. Hyperactivity is not among the criteria in either ICD-10 or DSM-IV. Its inclusion under the behavioural indicators of anorexia nervosa has been recommended 96. However, since hyperactivity is not universally present in all cases of otherwise typical anorexia nervosa and its manifestation may depend on stage of illness, we propose to include it as a supporting criterion, which may help differentiate between anorexia nervosa and normality in cases with borderline underweight.

In the DSM-IV, anorexia nervosa is further classified into restrictive and binge-eating/purging subtypes according to the presence of bingeing and purging behaviours. This subtyping has been criticized since the subtypes often represent developmental stages of the same illness and do not consistently predict outcome 69. While children and younger adolescents usually present with restrictive symptomatology, binge-purging behaviours develop in the majority at later stages 69,110. Purging behaviour has been found to predict poor outcome in some studies but not in others 69,105,110. Binge-purge anorexia nervosa has a high rate of diagnostic transitions with bulimia nervosa 68,70. To avoid repeated diagnosis changes, we propose that anorexia nervosa with bingeing and purging behaviour should be classified as combined eating disorder. The proposed category of “combined eating disorder” will include cases that were classified as anorexia nervosa, binge-purge subtype as well as cases that display clinically significant anorectic and bulimic symptomatology sequentially.

BULIMIA NERVOSA

  1. Top of page
  2. THE INTERNATIONAL CLASSIFICATION OF DISEASES
  3. CLINICAL EVIDENCE AND DIAGNOSTIC CRITERIA
  4. DEVELOPMENTAL CONTEXT
  5. CULTURAL CONTEXT
  6. TEMPORAL INSTABILITY AND DIAGNOSTIC TRANSITIONS
  7. FEEDING DISORDERS
  8. EATING DISORDERS NOT OTHERWISE SPECIFIED (EDNOS)
  9. ANOREXIA NERVOSA
  10. BULIMIA NERVOSA
  11. BINGE EATING DISORDER
  12. EATING AND FEEDING DISORDERS DUE TO OTHER CONDITIONS WITH A KNOWN CA
  13. CONCLUSIONS
  14. Acknowledgements
  15. References

Bulimia nervosa was first described in 1979 as a variant of anorexia nervosa 111, and soon afterwards was accepted as a separate diagnosis. In ICD-10 and DSM-IV, bulimia nervosa is defined by three criteria: recurrent binge eating, recurrent compensatory behaviour, and preoccupation with own body weight or shape, all of which are required for making the diagnosis.

Overall, bulimia nervosa is a valid category 112 that is used in practice 113, and its treatment has a broad evidence base 114. However, various aspects of the binge eating criterion have been criticized for making the diagnosis too restrictive and resulting in a substantial proportion of patients with bulimia-like problems classified as EDNOS. The need for a broader diagnostic category is reflected in the clinical trial literature, with eight of the eleven recent trials using broader criteria (Table 1). The two aspects of the definition that have been relaxed in these trials and discussed in the literature are the amount of food eaten in a binge eating episode and the frequency of binge-purge behaviours.

We first consider the amount of food eaten. Both ICD-10 and DSM-IV specify that binge eating is only present when an unusually large amount of food is eaten in one go and the subject experiences loss of control over eating (i.e., feels unable to stop eating or limit the amount or type of food eaten). However, a number of reports have highlighted the fact that many patients report eating amounts of food that may objectively appear normal, but are subjectively considered too large 115,116. Episodes of eating that are accompanied by subjective loss of control but do not involve eating an unusually large amount of food are described as “subjective” binges 117. A number of studies have compared patients presenting with “subjective” and “objective” binges and found little or no clinically meaningful differences 116,118,119,120,121. While objective binge eating episodes may be associated with higher body mass index 116,119 and impulsivity 118, subjective and objective binge eating episodes show a similar pattern of psychiatric comorbidity 116,118,119,120,121,122, are associated with similar levels of service utilization 116,119 and have a similar response to treatment 12,120. The literature suggests that subjective experience of loss of control over eating is the core defining feature of binge eating that is associated with psychopathology and quality of life, irrespective of the amount of food eaten, in both adults 122,123 and children 124. It has therefore been proposed that the requirement for an unusually large amount of food should be removed 116,119. Dropping the requirement for an unusually large amount of food is the one modification that strongly reduces the use of EDNOS and allows most bulimia-like presentations to be classified as bulimia nervosa 29,84,90.

The other contentious aspect of the diagnostic criteria is the frequency of bingeing and purging behaviour that is required for the diagnosis. This criterion is handled differently in ICD-10 and DSM-IV. While ICD-10 simply requires binge eating and compensatory behaviours to be repeated, DSM-IV specifies a minimum frequency of two binge-eating/purging episodes per week for at least three months. Although there is a general consensus that the diagnosis should not be applied to cases with infrequent episodes of binge eating, there is no evidence supporting a specific frequency, and most researchers and clinicians suggest either removing or lowering the frequency criterion 125. Since there appear to be no meaningful differences between subjects who binge-eat twice or more versus those who binge eat once a week 126, a lowering or loosening of the frequency criterion appears to be in order. Although lowering the frequency requirement to once a week on its own has a relatively modest effect, in conjunction with allowing subjective binge eating episodes it leads to a substantial reduction in the use of EDNOS 29,90.

The requirement for dread of fatness (ICD-10) or an undue influence of body weight and shape on self-evaluation (DSM-IV) is adequate in most Western adult settings, but may be problematic in children and in non-Western cultures 28,32. It is proposed that this criterion be applied flexibly, include developmental-stage and culture-specific manifestations, and regard behavioural equivalents as indicators of eating-disorder specific psychopathology 32,42.

In the DSM-IV, bulimia nervosa is further classified into purging and non-purging subtypes. The non-purging subtype is defined by compensatory behaviours that are limited to fasting and exercising and represent a relatively small minority of patients with bulimia nervosa. In practice, subtyping of bulimia nervosa is little used 113 and there is virtually no evidence to support its validity or utility 127. In terms of severity, non-purging bulimia nervosa appears to be intermediate between purging bulimia nervosa and binge eating disorder 127. It has been proposed to either abandon subtyping or to remove the non-purging behaviours from the list of inappropriate compensatory behaviours that define bulimia nervosa 127. Adopting the latter proposal would mean that cases previously classified as non-purging bulimia would be diagnosed as binge eating disorder. An alternative proposal has been made to subtype bulimia nervosa based on the history of anorexia nervosa, which has more predictive validity 75. We think this proposal can be applied as part of the combined eating disorder. In the absence of evidence either way, we favour retaining non-purgative compensatory behaviours in the definition of bulimia nervosa.

BINGE EATING DISORDER

  1. Top of page
  2. THE INTERNATIONAL CLASSIFICATION OF DISEASES
  3. CLINICAL EVIDENCE AND DIAGNOSTIC CRITERIA
  4. DEVELOPMENTAL CONTEXT
  5. CULTURAL CONTEXT
  6. TEMPORAL INSTABILITY AND DIAGNOSTIC TRANSITIONS
  7. FEEDING DISORDERS
  8. EATING DISORDERS NOT OTHERWISE SPECIFIED (EDNOS)
  9. ANOREXIA NERVOSA
  10. BULIMIA NERVOSA
  11. BINGE EATING DISORDER
  12. EATING AND FEEDING DISORDERS DUE TO OTHER CONDITIONS WITH A KNOWN CA
  13. CONCLUSIONS
  14. Acknowledgements
  15. References

Binge eating disorder, characterized by recurrent binge eating without compensatory behaviours, was described in 1959, but only became a focus of clinical and research attention in the last two decades. It was not included in ICD-10, but was listed in DSM-IV as a provisional diagnosis in need of further testing. Since then, a large body of research has accumulated, and binge eating disorder is now considered a valid and useful category 91,92.

Binge eating disorder is strongly associated with obesity, with approximately two-thirds of affected individuals being obese. Binge eating disorder frequently co-occurs with other mental disorders, especially anxiety and depression 128, yet has been found to be distinct from these and not a mere indicator of general psychopathology 129. The distinction between binge eating disorder and bulimia nervosa is less clear-cut and in many cases the two categories may represent different stages of the same disorder 65,68. Patients with binge eating disorder are on average older than those of bulimia nervosa, and approximately two-thirds have a history of using inappropriate compensatory behaviours, suggesting a past diagnosis of bulimia nervosa 130. Although weight and shape concerns are not required for the diagnosis of binge eating disorder, they are commonly part of the presentation 131,132.

As in bulimia nervosa, the specific diagnostic criteria for binge eating disorder have been a subject of discussion. In DSM-IV, it was required that binge eating episodes consist in eating an amount of food that is definitely larger than what most individuals would eat in a similar situation, and that the individual experiences loss of control over eating (i.e., feels unable to stop eating or limit the amount or type of food eaten). However, similar to binges in bulimia nervosa, it was found that loss of control is the core defining feature of binging 133. Although the amount of food eaten is often large, it is less useful as a defining factor, because binges involving eating amounts of food that are objectively not unusually large but are considered large by the individual are associated with similar psychopathology and impairment 133. In addition, binge eating is frequently characterized by eating alone because of embarrassment, eating other types of foods, and feelings of guilt and disgust. DSM-IV requires that binge eating is associated with distress, and the validity of this additional criterion is generally supported 134.

The duration of a binge varies. Most episodes of binge eating last less than two hours, but extended binge eating has been described as binge days. The DSM-IV further required that binge eating be present in at least two days per week for at least six months. There appears to be little evidence supporting this frequency and the unusually long requested duration. Since binge eating disorder is often associated with rapid weight gain, waiting six months before establishing the diagnosis may be contraproductive. It has been shown that relaxing the frequency criterion to once a week would lead to only small increase in the rate of binge eating disorder 135.

In summary, there is a consensus that binge eating disorder is a valid and useful diagnosis. It is proposed that binge eating disorder be included in ICD-11. It is also proposed that the diagnostic criteria for binge eating disorder be broadened to include binges that do not involve eating an unusually large amount of food as long as there is definite loss of control over eating, binge eating is distinct from regular eating patterns and causes distress 97,116. There is no specific evidence to override the proposed uniform duration criterion. Therefore, the diagnosis of binge eating disorder may be appropriate if binge eating occurs regularly for at least four weeks. In less severe cases, longer duration may be needed to establish clinical significance.

EATING AND FEEDING DISORDERS DUE TO OTHER CONDITIONS WITH A KNOWN CA

  1. Top of page
  2. THE INTERNATIONAL CLASSIFICATION OF DISEASES
  3. CLINICAL EVIDENCE AND DIAGNOSTIC CRITERIA
  4. DEVELOPMENTAL CONTEXT
  5. CULTURAL CONTEXT
  6. TEMPORAL INSTABILITY AND DIAGNOSTIC TRANSITIONS
  7. FEEDING DISORDERS
  8. EATING DISORDERS NOT OTHERWISE SPECIFIED (EDNOS)
  9. ANOREXIA NERVOSA
  10. BULIMIA NERVOSA
  11. BINGE EATING DISORDER
  12. EATING AND FEEDING DISORDERS DUE TO OTHER CONDITIONS WITH A KNOWN CA
  13. CONCLUSIONS
  14. Acknowledgements
  15. References

In addition to the feeding and eating disorders described above, the ICD-11 chapter should include a reference to disorders that are classified elsewhere but may primarily manifests with disordered eating. The most prominent example is Prader-Willi syndrome, which is caused by a deletion of the paternal copy of a region on chromosome 15, and is primarily classified in chapter XVII of ICD-10 (Congenital malformations, deformations and chromosomal abnormalities). This syndrome often presents with insatiable appetite, overeating, food hoarding and eating of non-food substances, but also includes intellectual disability.

CONCLUSIONS

  1. Top of page
  2. THE INTERNATIONAL CLASSIFICATION OF DISEASES
  3. CLINICAL EVIDENCE AND DIAGNOSTIC CRITERIA
  4. DEVELOPMENTAL CONTEXT
  5. CULTURAL CONTEXT
  6. TEMPORAL INSTABILITY AND DIAGNOSTIC TRANSITIONS
  7. FEEDING DISORDERS
  8. EATING DISORDERS NOT OTHERWISE SPECIFIED (EDNOS)
  9. ANOREXIA NERVOSA
  10. BULIMIA NERVOSA
  11. BINGE EATING DISORDER
  12. EATING AND FEEDING DISORDERS DUE TO OTHER CONDITIONS WITH A KNOWN CA
  13. CONCLUSIONS
  14. Acknowledgements
  15. References

We have reviewed published evidence relevant to the classification of feeding and eating disorders, with particular emphasis on clinical utility, response to treatment, prognosis, and developmental and cultural context. Based on this evidence, we make recommendations consistent with the general directions for the development of ICD-11. The principal recommendations are listed in Table 2.

Table 2.  Recommendations for the classification of feeding and eating disorders in ICD-11
1Merge feeding and eating disorders into a single grouping with diagnostic categories available for all age groups
2Broaden the category of anorexia nervosa through dropping the requirement for amenorrhoea; extending the weight criterion to include any significant underweight; and extending the cognitive criterion to include developmentally and culturally relevant cognitions and behavioural equivalents of fear of fatness, preoccupations with body weight and shape or food and eating.
3Introduce a severity qualifier “with a dangerously low body weight” to distinguish the most severe cases that carry the riskiest prognosis within the broad category of anorexia nervosa.
4Broaden the category of bulimia nervosa to include subjective binge eating.
5Include the category of binge eating disorder, defined by either subjective or objective binge eating in the absence of regular compensatory behaviour.
6Include a category of combined eating disorder to classify subjects who concurrently or sequentially fulfil the criteria for both anorexia nervosa and bulimia nervosa.
7Introduce a category of avoidant/restrictive food intake disorder (ARFID) to classify restricted food intake that is not accompanied by body weight and shape related psychopathology.
8Introduce a uniform minimal duration criterion of four weeks.

We hope that the proposed changes will substantially improve the clinical utility of the classification of feeding and eating disorders, will eliminate the need for using uninformative “not otherwise specified” diagnoses, and will stimulate research on continuity between child and adult presentations and on treatment efficacy in groups of patients that represent the vast majority of those presenting with eating related psychopathology in routine clinical settings.

Acknowledgements

  1. Top of page
  2. THE INTERNATIONAL CLASSIFICATION OF DISEASES
  3. CLINICAL EVIDENCE AND DIAGNOSTIC CRITERIA
  4. DEVELOPMENTAL CONTEXT
  5. CULTURAL CONTEXT
  6. TEMPORAL INSTABILITY AND DIAGNOSTIC TRANSITIONS
  7. FEEDING DISORDERS
  8. EATING DISORDERS NOT OTHERWISE SPECIFIED (EDNOS)
  9. ANOREXIA NERVOSA
  10. BULIMIA NERVOSA
  11. BINGE EATING DISORDER
  12. EATING AND FEEDING DISORDERS DUE TO OTHER CONDITIONS WITH A KNOWN CA
  13. CONCLUSIONS
  14. Acknowledgements
  15. References

This article has been informed by the activity of the Working Group on Classification of Mental and Behavioural Disorders in Children and Adolescents, referring to the International Advisory Group for the Revision of ICD-10 Mental and Behavioural Disorders. M. Rutter is the chair of this working group. R. Uher works for this group as a consultant. This manuscript reflects the work and opinions of the authors, who take full responsibility for its content. The authors thank U. Schmidt, I. Campbell and B.T. Walsh for their comments on earlier versions of this manuscript.

References

  1. Top of page
  2. THE INTERNATIONAL CLASSIFICATION OF DISEASES
  3. CLINICAL EVIDENCE AND DIAGNOSTIC CRITERIA
  4. DEVELOPMENTAL CONTEXT
  5. CULTURAL CONTEXT
  6. TEMPORAL INSTABILITY AND DIAGNOSTIC TRANSITIONS
  7. FEEDING DISORDERS
  8. EATING DISORDERS NOT OTHERWISE SPECIFIED (EDNOS)
  9. ANOREXIA NERVOSA
  10. BULIMIA NERVOSA
  11. BINGE EATING DISORDER
  12. EATING AND FEEDING DISORDERS DUE TO OTHER CONDITIONS WITH A KNOWN CA
  13. CONCLUSIONS
  14. Acknowledgements
  15. References