NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
National Guideline Alliance (UK). Eating Disorders: Recognition and Treatment. London: National Institute for Health and Care Excellence (NICE); 2017 May. (NICE Guideline, No. 69.)
December 2020: NICE amended the recommendation on managing diabetes to highlight the importance of rotating insulin injection sites within the same body region, in line with an MHRA Drug Safety Update on insulins (all types): risk of cutaneous amyloidosis at injection site. For the current recommendations, see: www.nice.org.uk/guidance/NG69/chapter/recommendations. March 2020: Cross reference to NICE's guideline on supporting adult carers added to the recommendation on offering family members or carers assessments of their own needs 1.1.10.
9.1. Introduction
Many people with an eating disorder do not meet the diagnostic criteria for anorexia nervosa, bulimia nervosa or binge eating disorder. There is no consensus over how to refer to these states, so they are often described as ‘atypical’ eating disorders – even though in some settings they are more common than the ‘typical’ ones. Confusing matters further, the terminology used by the DSM to refer to these conditions has changed from eating disorder not otherwise specified (EDNOS) to other specified feeding or eating disorder (OSFED). The new DSM-5 diagnosis ‘avoidant/restrictive food intake disorder’ (ARFID) is not classed as an OSFED (atypical eating disorder).
In practice, these atypical states fall into two groups (Fairburn et al., 2007). There are eating disorders that closely resemble anorexia nervosa, bulimia nervosa or binge eating disorder, but do not quite meet their diagnostic criteria. There are also ‘mixed states’, in which the features of anorexia nervosa, bulimia nervosa or binge eating disorder are combined in an idiosyncratic way. In addition to these, purging disorder and night eating syndrome are also currently classified in DSM-5 as atypical eating disorders.
A common misconception is that the atypical eating disorders are milder or less severe than the typical eating disorders. This is not the case. They are associated with the same level of distress and impairment and they are just as self-perpetuating (Fairburn and Bohn, 2005). Almost all share the same over-concern about eating, shape and weight as seen in the typical eating disorders and the same tendency to engage in persistent and extreme dieting or other forms of disordered eating (such as binge eating and purging). Body weight also tends to be low if the dietary restriction is marked.
Most people with an atypical eating disorder are female and in their 20s. Many of them have a history of anorexia nervosa, bulimia nervosa or both, or will go onto develop a typical eating disorder, reflecting the diagnostic migration that is common among the eating disorders (Milos et al., 2005). Their prevalence and incidence in the general population is uncertain, because of the difficulty in defining them and because they are ignored by some assessment instruments (Smink et al., 2012). It seems that they are more common that the typical eating disorders.
9.2. Psychological interventions
Many people with an eating disorder do not meet the diagnostic criteria for anorexia nervosa, bulimia nervosa or binge eating disorder. There is no consensus over how to refer to these states, so they are often described as ‘atypical’ eating disorders – even though in some settings they are more common than the ‘typical’ ones. Confusing matters further, the terminology used by the DSM to refer to these conditions has changed from eating disorder not otherwise specified (EDNOS) to other specified feeding or eating disorder (OSFED). The new DSM-5 diagnosis ‘avoidant/restrictive food intake disorder’ (ARFID) is not classed as an OSFED (atypical eating disorder).
In practice, these atypical states fall into two groups (Fairburn et al., 2007). There are eating disorders that closely resemble anorexia nervosa, bulimia nervosa or binge eating disorder, but do not quite meet their diagnostic criteria. There are also ‘mixed states’, in which the features of anorexia nervosa, bulimia nervosa or binge eating disorder are combined in an idiosyncratic way. In addition to these, purging disorder and night eating syndrome are also currently classified in DSM-5 as atypical eating disorders.
A common misconception is that the atypical eating disorders are milder or less severe than the typical eating disorders. This is not the case. They are associated with the same level of distress and impairment and they are just as self-perpetuating (Fairburn and Bohn, 2005). Almost all share the same over-concern about eating, shape and weight as seen in the typical eating disorders and the same tendency to engage in persistent and extreme dieting or other forms of disordered eating (such as binge eating and purging). Body weight also tends to be low if the dietary restriction is marked.
Most people with an atypical eating disorder are female and in their 20s. Many of them have a history of anorexia nervosa, bulimia nervosa or both, or will go onto develop a typical eating disorder, reflecting the diagnostic migration that is common among the eating disorders (Milos et al., 2005). Their prevalence and incidence in the general population is uncertain, because of the difficulty in defining them and because they are ignored by some assessment instruments (Smink et al., 2012). It seems that they are more common that the typical eating disorders.
9.2.1. Review question: Does any group or individual psychological intervention with or without a pharmacological intervention produce benefits/harms in people with eating disorders compared with any other intervention or controls?
The review protocol summary, including the review question and the eligibility criteria used for this section of the guideline, can be found in Table 338. Further information about the search strategy can be found in Appendix H; the full review protocols can be found in Appendix J.
This review considers all psychological interventions that may be delivered to children, young people and adults with an eating disorder with or without a pharmacological intervention. The interventions were categorised according to their mode of delivery, i.e. individual, group or self-help, the age of the participants and the type of eating disorder. In addition, the interventions were grouped according to their type of therapy and were compared to any other intervention or to wait list controls.
9.2.2. Clinical evidence
One RCT (n=35) on individual therapy versus group therapy was identified in people with EDNOS (Nevonen and Broberg, 2005). Three RCTs (n=396) were found that investigated the effects of self-help (guided and internet) compared with wait list controls in people with any eating disorder (Gulec et al., 2014; Hotzel et al., 2014; Traviss et al., 2011). Although the latter were on people with any eating disorder it was agreed they are best presented in the EDNOS chapter.
See also the study selection flow chart in Appendix K, forest plots in Appendix M, study evidence tables in Appendix L and exclusion list in Appendix J.
9.2.3. Economic evidence
No economic evidence on the cost effectiveness of interventions for people with EDNOS was identified by the systematic search of the economic literature undertaken for this guideline. Details on the methods used for the systematic search of the economic literature are described in Chapter 3.
9.2.4. Clinical evidence statements
9.2.4.1. Individual therapy versus group therapy
9.2.4.1.1. Individual hybrid compared with group hybrid for people with EDNOS at the end of treatment
Very low quality evidence from one RCT (n=35) showed no difference in the effect of individual hybrid therapy on depression, general psychopathology, EDE-restraint, EDI-total and remission compared with group hybrid.
9.2.4.1.2. Individual hybrid compared with group hybrid for people with EDNOS at follow up
Very low quality evidence from one RCT (n=35) showed no difference in the effect of individual hybrid therapy on depression, general psychopathology, EDE-restraint, EDI-total and remission compared with group hybrid.
9.2.4.1.3. Internet self-help compared with wait list control for people with any eating disorder at the end of treatment
Low quality evidence from one RCT (n=78) showed no difference in the effect of internet self-help on EDE-total and depression compared with wait list controls.
Low quality evidence from one RCT (n=212) showed no difference in the effect of internet self-help on BMI compared with wait list controls.
Low quality evidence from two RCTs (n=290) showed no difference in the effect of internet EDE-restraint, EDE-eating concern, EDE-shape concern and EDE-weight concern compared with wait list controls.
Low quality evidence from one RCT (n=212) showed internet self-help is more effective on vomiting compared with wait list controls but there was some uncertainty.
9.2.4.1.4. Guided self-help compared with wait list control for people with any eating disorder at the end of treatment
Low quality evidence from one RCT (n=81) showed guided self-help is more effective on EDE-total, EDE-restraint, EDE-eating concern, EDE-shape concern and EDE-weight concern compared with wait list controls.
Low quality evidence from one RCT (n=81) showed no difference in the effect of guided self-help on BMI, binge eating, vomiting, laxative use and exercise frequency compared with wait list controls.
9.2.5. Economic evidence statements
No economic evidence on the cost effectiveness of interventions for people with EDNOS was available.
9.2.6. Recommendations and link to evidence
Psychological treatment for OSFED (EDNOS)
| |
Relative value of different outcomes | The committee discussed the importance and relevance of various outcomes when assessing the effectiveness of psychotherapies for treating EDNOS (note: the recommendation replaces the term EDNOS with OSFED). For this population it was agreed that binge-eating frequency and remission are of greatest concern. Other outcomes that are important but are considered rare events or rarely measured in RCTs for eating disorders include all-cause mortality, adverse events, quality of life, resource use and relapse. They were therefore extracted where possible, but did not factor strongly in the decision-making. Other outcomes of concern for people with EDNOS that are of lesser importance but clearly important outcomes include, general psychopathology, body weight, general functioning, eating disorder psychopathology, family functioning and service user experience. |
Trade-off between clinical benefits and harms | Only one RCT was found on the EDNOS population, which compared an individual combined treatment programme with a group combined treatment programme in adult females with EDNOS. At the end of the treatment there was no difference on remission, general psychopathology, EDE-dietary restraint, depression and EDI-total. Similar results were found at follow up. No data was available on binge frequency, adverse events, all-cause mortality, quality of life, resource use, relapse, body weight, general functioning, family functioning, or service user experience. The treatments in the combined study comprised CBT-ED followed by interpersonal psychotherapy in either an individual or group format. The results showed either format was equally effective for this population. Any eating disorder Three studies were found on those with any eating disorder. Two studies compared internet self-help with wait list controls at the end of treatment in adults with a range of eating disorders and showed no difference on EDE-subscales, BMI and depression. However, there was a trend for a reduction in vomiting in the internet self-help group, although there was some uncertainty. No data was available on remission, adverse events, all-cause mortality, quality of life, resource use, relapse, general functioning, family functioning, or service user experience. Another study compared guided self-help with wait list controls in adults with a range of eating disorders and showed no difference in BMI, binge eating, vomiting, laxative use and exercise frequency at the end of the treatment. However, there was a reduction on scores on the EDE-subscales in the guided self-help group. No data was available on remission, adverse events, all-cause mortality, quality of life, resource use, relapse, general psychopathology, general functioning, family functioning, or service user experience. Guided self-help appeared to show some benefit on the EDE-subscales compared with wait list controls but not when specific compensatory behaviours were measured. |
Trade-off between net health benefits and resource use | The committee expressed the view that if something is cost effective for people with a particular eating disorder it was reasonable to assume tnta it would be for a person with EDNOS in-line with the clinical presentation of an eating disorder they most closely resemble. |
Quality of evidence | The evidence was mostly low to very low quality. Outcomes were often downgraded for impression because the 95% confidence interval crossed one or two MIDs or it didn’t meet the optimal information size. In the study on combined treatments, the definition of remission used was problematic - the percentage of participants who score one or more scale steps lower than their pre-treatment values for binge eating and/or purging using a revised version of the Rating of Anorexia/Bulimia interview - and could have been excluded for this reason. However, this definition of remission would include people who move from the ‘binge or purge several times each day’ category to ‘binge or purge from five to seven days a week’, rather than those who have stopped binging over a two week period. No data was available comparing the individual or group interventions with wait list controls, so it is difficult to know if they are better than no treatment alone. No data was available on children or young people with EDNOS. |
Other considerations | Given the scarcity of data on this population, the committee drew on their knowledge and experience and by informal consensus agreed that it was preferable to recommend psychological treatments for a person with EDNOS (OSFED) in line with the clinical presentation of the eating disorder that their signs and symptoms most closely resemble. |
9.3. Carer interventions
9.3.1. Review question: Does any psychological intervention produce benefits/harms in the parents or carers of children or young people with an eating disorder compared with any other intervention or controls?
The review protocol summary, including the review question and the eligibility criteria used for this section of the guideline, can be found in Table 342. Further information about the search strategy can be found in Appendix H; the full review protocols can be found in Appendix F.
This review considers all psychological interventions for the parents or carers of children or young people with an eating disorder. The interventions were categorised according to their mode of delivery (e.g. group, individual or self-help), the age of the people with the eating disorder and the type of eating disorder. The control arm could include wait list controls, treatment as usual or any other intervention, however results comparing an intervention with wait list controls were always presented separately.
9.3.2. Clinical evidence
Two RCTs (n=204) met the eligibility criteria for this review, one of which was in young people (Spettigue et al., 2015), the other which was in adults (Goddard, 2011). An overview of all the trials included in the review can be found in Table 343. Further information about both included and excluded studies can be found in Appendix J.
One study (n=51) examined the efficacy of psychoeducation that included a 2-hour session and bi-weekly telephone support calls in the time before formal assessment calls compared with wait list controls (Spettigue 2015). Another study (n=153) examined the efficacy of the Expert Carers Helping Others (ECHO) self-help intervention with and without guidance (Goddard 2011).
Summary of findings for interventions for carers of people with any eating disorder can be found in Table 344, Table 345 and Table 346. See also the study selection flow chart in Appendix K, forest plots in Appendix M, study evidence tables in Appendix L and exclusion list in Appendix J.
9.3.3. Economic evidence
No economic evidence on the cost effectiveness of interventions for the parents or carers of children or young people with EDNOS was identified by the systematic search of the economic literature undertaken for this guideline. Details on the methods used for the systematic search of the economic literature are described in Chapter 3.
9.3.4. Clinical evidence statements
9.3.4.1. Psychoeducation versus wait list control in carers of young people with any eating disorder at end of treatment
Very low quality evidence from one RCT (n=31) showed psychoeducation is more effective on carer self-efficacy compared with wait list control.
Very low quality evidence from one RCT (n=28) showed psychoeducation may be more effective on the carer’s knowledge of eating disorders compared with wait list control, although there was some uncertainty.
9.3.4.2. Psychoeducation versus wait list control in carers of young people with any eating disorder at formal assessment
Very low quality evidence from one RCT (n=31) showed psychoeducation is more effective on carer self-efficacy compared with wait list control.
Very low quality evidence from one RCT (n=28) showed psychoeducation is more effective on the carer’s knowledge of eating disorders compared with wait list control.
Very low quality evidence from one RCT (n=36) showed no difference in the effect of psychoeducation on the burden of the eating disorder on the carer compared with wait list control.
9.3.4.3. Guided self-help versus self-help in carers of adults with any eating disorder
Low quality evidence from one RCT (n=120) showed no difference in the effect of guided self-help on the burden of the eating disorder on the carer, carer quality of life, carer self-efficacy, carer positive experience of caregiving, carer accommodation and enabling and carer general psychopathology compared with self-help only.
9.3.5. Economic evidence statements
No economic evidence on the cost effectiveness of interventions for the parents or carers of children or young people with EDNOS was available.
9.3.6. Recommendations and link to evidence
Working with family members and carers
See Section 6.3.6 for relevant recommendations | |
Relative value of different outcomes | The committee discussed the importance and relevance of various outcomes, when assessing whether any interventions help the parents and carers of children and young people with an eating disorder. The critical outcomes for the parents and carers were: general psychopathology, family functioning, quality of life, and other primary outcomes reported by the study. Other outcomes that are critical for the child or young person with the eating disorder include remission and binge eating or body weight, depending on the eating disorder. Other outcomes that are of lesser importance but clearly important outcomes include, general functioning, service user experience, all-cause mortality, adverse events and eating disorder psychopathology. |
Trade-off between clinical benefits and harms | Randomised control trials investigating interventions for the carers of young people with any eating disorder did not show many favourable outcomes. Psychoeducation compared with waitlist control showed a positive effect on carer self-efficacy and a trend to improve carer knowledge of eating disorders at the end of treatment. Long-term follow up showed favourable results in both but carer burden (only measured at follow up) was not different compared with wait list controls. No evidence was found on the critical outcomes of carer general psychopathology, family functioning, and quality of life, nor on the other important outcomes. Comparing guided self-help with self-help showed no difference in any of the carer-related outcomes at the end of treatment. No evidence was found on the other important outcomes. For discussion of carer interventions for other eating disorders, see the LETRs in 6.3.6 |
Trade-off between net health benefits and resource use | The committee expressed the view that offering family members and carers an assessment of their own needs may incur additional healthcare resources (that is, time required to perform such assessment). However, the committee considered the cost of providing such assessment to be small, taking into account the potential reduction in family and carers’ burden, potential depression and other health vulnerabilities which may be costly to other parts of the healthcare system, especially considering that the burden on family and carers can last for many years and increase their morbidity and stress. Consequently, the committee judged that assessment that aims to improve family and carers’ experience are likely to represent good value for money. |
Quality of evidence | The quality of the evidence was mostly very low. High dropout rates >20% were also detected in some groups. Imprecision was detected in most outcomes due to the 95% confidence interval crossing one or two MIDs or because it did not meet the optimal information size. Outcomes were not always measured at the end of treatment or at follow up. It is not known if any improvements in the carer’s general psychopathology also translated to benefits in the children with the eating disorder. |
Other considerations | Given the very low quality of the data with lack of strong positive findings, the committee decided that there was not enough evidence to support a recommendation on any specific intervention for parents or carers of people with an eating disorder. See the LETR in Section 6.3.6 for further discussion of carer interventions. |
9.4. Pharmacological interventions
9.4.1. Review question: Does any pharmacological intervention produce benefits/harms on specified outcomes in people with eating disorders?
The review protocol summary, including the review question and the eligibility criteria used for this section of the guideline, can be found in Table 347. Further information about the search strategy can be found in Appendix H; the full review protocols can be found in Appendix F.
This review considers all pharmacological interventions that may be delivered to children, young people and adults with an eating disorder with or without a psychological intervention. The interventions were categorised according to the type of pharmacological intervention, the age of the participants and the type of eating disorder. In addition, the interventions were grouped according to their type of therapy and were compared to placebo, wait list controls or any other intervention
9.4.2. Clinical evidence
No studies were identified that met the eligibility criteria for this review. Further information about excluded studies can be found in Appendix J. See also the study selection flow chart in Appendix K.
9.4.3. Economic evidence
No economic evidence on the cost effectiveness of pharmacological interventions for the treatment of EDNOS was identified by the systematic search of the economic literature undertaken for this guideline. Details on the methods used for the systematic search of the economic literature are described in Chapter 3.
9.4.4. Clinical evidence statements
No studies were identified that met the eligibility criteria for this review.
9.4.5. Economic evidence statements
No economic evidence on the cost effectiveness of pharmacological interventions for people with EDNOS was available.
9.4.6. Recommendations and link to evidence
Relative value of different outcomes | The committee discussed the importance and relevance of various outcomes when assessing the effectiveness of psychotherapies for treating OSFEDs. For this population it was agreed that binge-eating frequency and remission are of greatest concern. Other outcomes that are important but are considered rare events or rarely measured in RCTs for eating disorders include all-cause mortality, adverse events, quality of life, resource use and relapse. They were therefore extracted where possible, but did not factor strongly in the decision-making. |
Trade-off between clinical benefits and harms | No RCT evidence was identified on pharmacological interventions for people with EDNOS. The committee expressed the view that if something is cost effective for people with a particular eating disorder it will be for a person with EDNOS in-line with the clinical presentation of an eating disorder they most closely resemble. |
Trade-off between net health benefits and resource use | The network meta-analysis found no evidence for the effectiveness of pharmacological interventions for the management of people with bulimia nervosa and BED. As a result the committee expressed the view that such treatments are unlikely to be effective nor cost effective in people with EDNOS. |
Quality of evidence | No RCT evidence was identified on pharmacological interventions for people with EDNOS. |
Other considerations | Given the absence of evidence for pharmacological interventions for EDNOs the committee decided not to recommend pharmacological interventions. |
9.5. Nutritional interventions
9.5.1. Review question: Does any nutritional intervention produce benefits/harms on specified outcomes in people with eating disorders?
The review protocol summary, including the review question and the eligibility criteria used for this section of the guideline, can be found in Table 348. Further information about the search strategy can be found in Appendix H; the full review protocols can be found in Appendix F.
This review considers all nutritional interventions that may be delivered to children, young people and adults with an eating disorder with or without a pharmacological intervention. The interventions were categorised according to type of nutritional intervention, the age of the participants and the type of eating disorder. In addition, the interventions were grouped according to their type of therapy and were compared to wait list controls, placebo, or any other intervention.
9.5.2. Clinical evidence
No studies were identified that met the eligibility criteria for this review. Further information about excluded studies can be found in Appendix J. See also the study selection flow chart in Appendix K.
9.5.3. Economic evidence
No economic evidence on the cost effectiveness of nutritional interventions for people with EDNOS was identified by the systematic search of the economic literature undertaken for this guideline. Details on the methods used for the systematic search of the economic literature are described in Chapter 3.
9.5.4. Clinical evidence statements
No studies were identified that met the eligibility criteria for this review
9.5.5. Economic evidence statements
No economic evidence on the cost effectiveness of nutritional interventions for people with EDNOS was available.
9.5.6. Recommendations and link to evidence
The committee agreed that people with OFSED should be treated in-line with the eating disorder their symptoms most closely resemble | |
Relative value of different outcomes | The committee discussed the importance and relevance of various outcomes when assessing the effectiveness of psychotherapies for treating OSFEDs. For this population it was agreed that binge-eating frequency and remission are of greatest concern. Other outcomes that are important but are considered rare events or rarely measured in RCTs for eating disorders include all-cause mortality, adverse events, quality of life, resource use and relapse. They were therefore extracted where possible, but did not factor strongly in the decision-making. |
Trade-off between clinical benefits and harms | No relevant published evidence was identified. |
Trade-off between net health benefits and resource use | No relevant published economic evidence was identified. |
Quality of evidence | Not applicable |
Other considerations | The committee agreed that people with EDNOS should be treated in-line with the eating disorder their symptoms most closely resemble. |
9.6. Physical interventions
9.6.1. Review question: Do physical interventions, such as transcranial magnetic stimulation or physiotherapy, produce benefits/harms in people with eating disorders?
The review protocol summary, including the review question and the eligibility criteria used for this section of the guideline, can be found in Table 349. Further information about the search strategy can be found in Appendix H; the full review protocols can be found in Appendix F.
This review considers all physical interventions that may be delivered to children, young people and adults with an eating disorder. The interventions were categorised according to type of physical intervention, the age of the participants and the type of eating disorder and were compared to wait list controls, placebo, treatment as usual or any other intervention.
9.6.2. Clinical evidence
Four RCTs (n=171) met the eligibility criteria for this review (Boerhout et al., 2016; Carei et al., 2010; Hildebrandt et al., 2012; Trottier et al., 2015). The majority of studies were in an outpatient setting and the majority of participants were adult females (no young people). One study was conducted after participants had received treatment in an intensive day hospital setting (Trottier 2015). Further information about both included and excluded studies can be found in Appendix J.
Summary of findings for those on any eating disorder can be found in Error! Reference ource not found. Table 351, Table 352, Table 353, Table 354, Table 355 and Table 356. See also the study selection flow chart in Appendix K, forest plots in Appendix M, study evidence tables in Appendix L and exclusion list in Appendix J.
9.6.3. Economic evidence
No economic evidence on the cost effectiveness of physical interventions for people with EDNOS was identified by the systematic search of the economic literature undertaken for this guideline. Details on the methods used for the systematic search of the economic literature are described in Chapter 3.
9.6.4. Clinical evidence statements
9.6.4.1. Yoga and treatment as usual versus treatment as usual in adults with any eating disorder at end of treatment
Low quality evidence from one RCT (n=53) showed no difference in the effect of yoga and treatment as usual on BMI, EDE-global, EDE-dietary restraint, EDE-weight concern, EDE-shape concern, EDE-eating concern and depression compared with treatment as usual.
9.6.4.2. Yoga and treatment as usual versus treatment as usual in adults with eating disorder at follow up
Low quality evidence from one RCT (n=53) showed no difference in the effect of yoga and treatment as usual on BMI, EDE-global, EDE-weight concern, EDE-shape concern, EDE-eating concern and depression compared with treatment as usual.
Low quality evidence from one RCT (n=53) showed yoga and treatment as usual is more effective on EDE-dietary restraint compared with treatment as usual.
9.6.4.3. Graded body image therapy and maintenance treatment as usual versus maintenance treatment as usual in adults with any eating disorder at end of treatment
Very low quality evidence from one RCT (n=45) showed no difference in the effect of graded body image exposure therapy and maintenance treatment as usual on EDE-weight concern and EDE-shape concern compared with maintenance treatment as usual.
9.6.4.4. Graded body image therapy and maintenance treatment as usual versus maintenance treatment as usual in adults with any eating disorder at follow up
Very low quality evidence from one RCT (n=45) showed no difference in the effect of graded body image exposure therapy and maintenance treatment as usual on EDE-weight concern and EDE-shape concerns compared with maintenance treatment as usual.
9.6.4.5. Acceptance-based mirror exposure therapy and treatment as usual versus non-directive body image therapy and treatment as usual in adults with any eating disorder at end of treatment
Very low quality evidence from one RCT (n=33) showed acceptance-based mirror exposure therapy is more effective on EDE-Q-eating concern, EDE-Q-shape concern and EDE-Q-weight concern compared with non-directive body image therapy.
Very low quality evidence from one RCT (n=33) showed no difference in the effect of acceptance-based mirror exposure therapy on EDE-Q-dietary restraint compared with non-directive body image therapy.
9.6.4.6. Psychomotor therapy and supportive contact versus supportive contact in adults with any eating disorder
Very low quality evidence from one RCT (n=29) showed no difference in the effect of psychomotor therapy and support on Self-Expression and Control Scale-anger in and Self-Expression and Control Scale-anger out compared with support only.
9.6.5. Economic evidence statements
No economic evidence on the cost effectiveness of physical interventions for people with EDNOS was available.
9.6.6. Recommendations and link to evidence
Physical therapy for any eating disorder
See Section 6.6.6 for relevant recommendations | |
Relative value of different outcomes | The committee discussed the importance and relevance of various outcomes for the review on the effectiveness of physical interventions, such as transcranial magnetic stimulation or physiotherapy in people with eating disorders and it was agreed that for any eating disorder remission is of greatest concern. The other critical outcomes for anorexia nervosa are body weight and BMI and for binge eating disorder and bulimia nervosa it is binge eating. Other outcomes that are important but are considered rare events or rarely measured in randomised controlled trials for eating disorders include all-cause mortality, adverse events, quality of life, resource use and relapse, thus they were extracted where possible, but did not factor strongly in the decision making. Other outcomes of concern for people with an eating disorder that are of lesser importance but clearly important outcomes include, general psychopathology, body general functioning, family functioning and service user experience. |
Trade-off between clinical benefits and harms | An RCT was identified that compared yoga and treatment as usual with treatment as usual in adults with any eating disorder. At the end the treatment, no difference was found in any of the outcomes including BMI, EDE-total or any of the EDE- sub-scales. Similar findings were found at follow up (three weeks), however there was some improvement in EDE-restraint in the yoga and treatment as usual group compared with treatment as usual. No evidence was found on the critical outcomes of remission and binge eating, nor on the important outcomes of quality of life, all-cause mortality, relapse, general functioning, family functioning, resource use, and service user experience. A graded body image therapy (and maintenance treatment as usual) was compared with a maintenance treatment as usual in adults with any eating disorder. No difference was found in EDE-weight concerns or EDE-shape concerns at the end of treatment or at follow up. No evidence was found on the critical outcomes of remission, weight and binge eating, nor on the important outcomes of quality of life, all-cause mortality, general psychopathology, relapse, general functioning, family functioning, resource use, and service user experience. An acceptance-based body image mirror exposure therapy was compared with a control therapy and showed an improvement in EDE-eating concerns, EDE-weight concerns, EDE-shape concerns, but not in EDE-restraint. No evidence was found on the critical outcomes of remission, weight and binge eating, nor on the important outcomes of quality of life, all-cause mortality, general psychopathology, relapse, general functioning, family functioning, resource use, and service user experience. A psychomotor therapy and support was compared with support in females with any eating disorder and showed no difference at the end of treatment on self-expression and control anger scales. No evidence was found on the critical outcomes of remission, weight and binge eating, nor on the important outcomes of quality of life, all-cause mortality, eating disorder psychopathology, relapse, general functioning, family functioning, resource use, and service user experience. For discussion of physical interventions for other eating disorders, see the LETRs in the relevant chapters. |
Trade-off between net health benefits and resource use | There was no evidence for the effectiveness of physical interventions in people with EDNOS. As a result, such interventions are likely to be not cost effective. |
Quality of evidence | The evidence for physical interventions was mostly very low quality.. High dropout rates were also detected, with more than 20% dropping out in each arm. Most of the outcomes were the result of a single study with a very low number of participants, only binge eating disorder had more than 100 participants in total. Imprecision was detected in most outcomes because the 95% confidence interval crossed one or two MIDs or it did not meet the optimal information size. Also, few studies measured remission and/or compensatory behaviours relevant to that eating disorder. Some outcomes were excluded from the study because it was either unclear over what duration they measured the symptoms or it was less than the two week minimum required by the committee. |
Other considerations | The committee agreed that the evidence presented was not strong enough nor of sufficient quality to offer a physical intervention to people with EDNOS See LETR in Section 6.6.6 for further discussion of physical interventions. |
9.7. Management of long- and short-term complications
9.7.1. Review question: What interventions are effective at managing or reducing short and long-term physical complications of eating disorders?
The review protocol summary, including the review question and the eligibility criteria used for this section of the guideline, can be found in Table 357. Further information about the search strategy can be found in Appendix H; the full review protocols can be found in Appendix F.
This review considers all interventions that may be delivered to manage or reduce the short- or long-term physical complications of eating disorders in children, young people and adults and includes recovered as well as current service users. The interventions were categorised according to type of physical complication and intervention, the age of the participants and the type of eating disorder. The control arm varied depending on the study.
9.7.2. Clinical evidence
No studies were identified that met the eligibility criteria for this review.
9.7.3. Economic evidence
No economic evidence on the cost effectiveness of interventions for managing short and long-term physical complications for people with EDNOS was identified by the systematic search of the economic literature undertaken for this guideline. Details on the methods used for the systematic search of the economic literature are described in Chapter 3.
9.7.4. Clinical evidence statements
No studies were identified that met the eligibility criteria for this review.
9.7.5. Economic evidence statements
No economic evidence on the cost effectiveness of interventions for managing short and long-term physical complications for people with EDNOS was available.
9.7.6. Recommendations and link to evidence
See recommendation 139 | |
Relative value of different outcomes | The committee discussed the importance and relevance of various outcomes when assessing the effectiveness of interventions for treating OSFEDs. For this population it was agreed that binge-eating frequency and remission are of greatest concern. Other outcomes that are important but are considered rare events or rarely measured in RCTs for eating disorders include all-cause mortality, adverse events, quality of life, resource use and relapse. They were therefore extracted where possible, but did not factor strongly in the decision-making. |
Trade-off between clinical benefits and harms | No relevant clinical evidence was identified. |
Trade-off between net health benefits and resource use | No relevant existing economic evidence was identified. The committee expressed the view that if something is potentially cost effective and represents value for money for people with anorexia nervosa, bulimia nervosa, and binge eating disorder, it will also do so for people with OSFED. |
Quality of evidence | No relevant clinical evidence was identified. |
Other considerations | The committee agreed that people with OSFED who have short or long-term physical complications associated with the eating disorder, should be treated in-line with the eating disorder their symptoms most closely resemble. |
9.8. Management of comorbidities
9.8.1. Review question: Does any intervention for an eating disorder need to be modified in the presence of common long-term health conditions?
The review protocol summary, including the review question and the eligibility criteria used in this section of the guideline, can be found in. Further information about the search strategy can be found in Appendix H; the full review protocols can be found in Appendix F.
This review considers whether any intervention used to treat eating disorders in children, young people and adults needs to be modified in the presence of a common long-term health condition (i.e. comorbidity). The interventions were categorised according to their type, the type of eating disorder and comorbidity examined and the age of the participants. The comparison arm was the same intervention delivered to participants with the relevant eating disorder but without the relevant comorbidity.
9.8.2. Clinical evidence
9.8.2.1. Diabetes
One RCT (n=85) was identified that compared the effectiveness of a parental and patient group psychoeducation programme with treatment as usual for reducing symptoms in young people with type 1 diabetes and disturbed eating attitudes (Olmsted et al., 2002).
Two observational studies (n=878) were found that met the inclusion criteria (Colton et al., 2015; Custal et al., 2014). Both studies used two different populations, one with any eating disorder and type I diabetes and compared them with one that just had any eating disorder. The two groups were compared after either receiving the same treatment (CBT-ED) or different treatments (group CBT-ED with additional care by a multidisciplinary team for those with type I diabetes or just group CBT-ED alone). These comparisons allowed us to see if those with diabetes can respond equally well to treatment as those with just an eating disorder.
9.8.2.2. High alcohol misuse
One observational study (n=149) was found that addressed the comorbidity of alcohol misuse in people with an eating disorder (Karacic et al., 2011). The study examined the effect of transdiagnostic CBT-enhanced for eating disorders in adults with bulimia nervosa and other eating disorders not otherwise specified (EDNOS).
Although this review question includes people with any eating disorder (anorexia nervosa, bulimia nervosa, binge eating disorder, EDNOS), the committee wanted to firstly consider the evidence for individual eating disorders to see if specific recommendations could be made. If none was available, or it was deemed insufficient, then they agreed to make a general recommendation for treating people with any eating disorder and a common long-term health condition.
Alcohol misuse
9.8.3. Economic evidence
No economic evidence on the cost effectiveness of interventions for the management of comorbidities of EDNOS was identified by the systematic search of the economic literature undertaken for this guideline. Details on the methods used for the systematic search of the economic literature are described in Chapter 3.
9.8.4. Clinical evidence statements
Diabetes
9.8.4.1. Group psychoeducation versus treatment as usual in people with type I diabetes and disturbed eating disorders at end of treatment
Low quality evidence from one RCT (n=85) showed no difference in the effect of group psychoeducation on binges, EDE-restraint, EDE-shape concern, EDE-eating concern, EDE-weight concern, EDI-drive for thinness, EDI-bulimia, insulin omission days and HbA1c (%) compared with treatment as usual.
Low quality evidence from one RCT (n=85) showed group psychoeducation is more effective on EDI-body dissatisfaction compared with treatment as usual but there was some uncertainty.
9.8.4.2. Group psychoeducation versus treatment as in people with type I diabetes and usual disturbed eating disorders at follow up
Low quality evidence from one RCT (n=85) showed no difference in the effect of group psychoeducation on EDE-restraint, EDE-shape concern, EDE-eating concern, EDE-weight concern, EDI-drive for thinness, EDI-bulimia, EDI-body dissatisfaction, insulin omission days and HbA1c (%) compared with treatment as usual.
Low quality evidence from one RCT (n=85) showed group psychoeducation is more beneficial on frequency of binges compared with treatment as usual but there was some uncertainty.
9.8.4.3. Any eating disorder and type I diabetes versus any eating disorder at end of treatment. Observational study
Very low quality evidence from one observational study (n=40) showed no difference in the number of dropouts in those with any eating disorder and type I diabetes compared with an eating disorder alone. This trend was apparent in all types of eating disorders: anorexia nervosa, bulimia nervosa, binge eating disorder and EDNOS.
Very low quality evidence from two observational studies (n=873) showed lower rates of remission in those with any eating disorder and type I diabetes compared with an eating disorder alone. This trend was apparent in those with bulimia nervosa, but no difference was found in those with anorexia nervosa, binge eating disorder and EDNOS
Alcohol misuse
9.8.4.4. CBT-ED for people with an eating disorder and high or low alcohol misuse at end of treatment
Very low quality evidence from one observational study (n=119) showed no difference in the effect of CBT-ED on the number of people who were more than 1 standard deviation above EDE community norms in people with an eating disorder whose alcohol use was high compared with those whose alcohol use was low.
Very low quality evidence from one observational study (n=119) showed that CBT-ED is less effective on the number of people who were engaging in excessive drinking in people with an eating disorder whose alcohol use was high compared with people with those whose alcohol use was low.
9.8.4.5. CBT-ED for people with an eating disorder and high or low alcohol misuse at follow up
Very low quality evidence from one observational study (n=104) showed no difference in the effect of CBT-ED for eating disorders on EDE-global in people with an eating disorder whose alcohol use was high compared with people with an eating disorder whose alcohol use was low.
9.8.5. Economic evidence statements
No economic evidence on the cost effectiveness of interventions for the management of comorbidities of EDNOS was available.
Recommendations and link to evidence
See Section 6.8.6 for relevant recommendations | |
Relative value of different outcomes | The committee discussed the importance and relevance of various outcomes when assessing the effectiveness of treating people with an eating disorder including OFSED and a comorbidity. For disorders like binge eating disorder and bulimia nervosa, it was agreed binge eating frequency and remission are of greatest concern. For anorexia nervosa, body weight/BMI and remission are critical and for OSFED, remission and either binge eating or body weight/BMI depending on the eating disorder they most closely resemble. The other outcomes that are critical are the primary outcomes that are relevant to the physical or mental health comorbidity being treated. Other outcomes that are important but are considered rare events or rarely measured in RCTs for eating disorders include all-cause mortality, adverse events, quality of life, resource use and relapse. They were therefore extracted where possible, but did not factor strongly in the decision-making. Other outcomes of concern for people with binge eating disorder that are of lesser importance but clearly important outcomes include, general psychopathology, body weight, general functioning, family functioning and service user experience. |
Trade-off between clinical benefits and harms | For discussion of interventions for treatment of other eating disorders in addition to OFSED with a comorbidity, see the LETRs in the relevant chapters. |
Trade-off between net health benefits and resource use | The committee considered that providing care for people with eating disorders who have comorbid physical or mental health problems may have resource implications in terms of the extra time required to provide such care. However, the committee expressed the view that if such care leads to better identification of health needs and this results in appropriate treatment and management of underlying health problems at an earlier stage (including eating disorder and comorbid mental health problem), before individuals require more resource intensive management, then the additional costs associated with facilitating such care is expected to result in improved health outcomes in the longer term and potential future cost savings to the healthcare system. |
Quality of evidence | No evidence was identified for the treatment of OFSED and a physical or mental health comorbidity. |
Other considerations | The committee generated recommendations based on the limited data found and from their own knowledge and experience using informal consensus methods. See the LETR in section 6.8.6 for further discussion on the treatment of an eating disorders with a physical or mental health comorbidity. See the LETR in Section 7.8.6 for further discussion of the treatment of eating disorders with comorbid diabetes. |
Diabetes
See Section 7.8.6 for relevant recommendations | |
Relative value of different outcomes | The committee discussed the importance and relevance of various outcomes when assessing whether a treatment for eating disorders needs to be modified in the presence of a long-term health problem. In the case of diabetes, HbA1c levels and insulin omission days were considered critical outcomes. The other critical outcomes depended on the eating disorder included in the study. The committee noted for Type 1 diabetes, severity should be measured, or at least heavily informed, by HBA1c since HBA1c/ DKA frequency is the immediate risk factor; furthermore, BMI is less of a risk factor for death in those with Type 1 diabetes than HBA1c. Remission is of greatest concern for any eating disorder. In addition, for those with anorexia nervosa body weight or BMI are of greatest concern. For bulimia nervosa and binge eating disorder, binge eating is a critical outcome. For any eating disorder, other outcomes that are important but are considered rare events or rarely measured in RCTs for eating disorders include all-cause mortality, adverse events, quality of life, resource use and relapse. They were therefore extracted where possible, but did not factor strongly in the decision-making. Other outcomes of concern for people with an eating disorder that are of lesser importance but are clearly still important outcomes include general psychopathology, general functioning, eating disorder psychopathology, family functioning and service user experience. |
Trade-off between clinical benefits and harms | The ideal study design to answer the question of whether a treatment for eating disorders needs to be modified in the presence of a long-term health problem would be to randomise people with an eating disorder and diabetes to two different treatment groups: One modified to address both the eating disorder and diabetes and one non-modified eating disorder treatment. No studies were found examining whether eating disorder treatments for OSFED need to be modified in presence of diabetes. |
Trade-off between net health benefits and resource use | See the LETR in Section 7.8.6 for further discussion of the treatment of eating disorders with comorbid diabetes. |
Quality of evidence | See the LETR in Section 7.8.6 for further discussion of the treatment of eating disorders with comorbid diabetes. |
Other considerations | See the LETR in Section 7.8.6 for further discussion of the treatment of eating disorders with comorbid diabetes. |
Substance and medication misuse
See Section 7.8.6 for relevant recommendations | |
Relative value of different outcomes | The committee discussed the importance and relevance of various outcomes when assessing the effectiveness of treating people with an eating disorder and a comorbidity. For binge eating disorder and bulimia nervosa, it was agreed binge eating frequency and remission are of greatest concern. For anorexia nervosa, body weight/BMI and remission are critical and for ofsed, remission and either binge eating or body weight/BMI depending on the eating disorder they most closely resemble. Other critical outcomes will include the primary outcomes relevant to the physical or mental health comorbidity being treated. Other outcomes that are important but are considered rare events or rarely measured in RCTs for eating disorders include all-cause mortality, adverse events, quality of life, resource use and relapse. They were therefore extracted where possible, but did not factor strongly in the decision-making. Other outcomes of concern for people with binge eating disorder that are of lesser importance but clearly important outcomes include, general psychopathology, body weight, general functioning, family functioning and service user experience. |
Trade-off between clinical benefits and harms | Bulimia nervosa and EDNOS An observational study was identified were they extracted data from a randomised control trial and compared the outcomes in those with bulimia nervosa and EDNOS who had a low or high alcohol intake. The participants were treated with either broad or focused CBT-ED. At the end of 20 weeks treatment, there was no difference in the number who had EDE scores one standard deviation above community norms (i.e., relatively abnormal eating psychopathology) in those with a low or high alcohol intake. However, the number who continued to have excessive alcohol intake (defined as >21 units or >14 units/week for males and females respectively) was higher in those whose alcohol intake was high compared with those whose intake was low. At 60 weeks follow up, there continued to be no difference in EDE scores between those who had low versus high alcohol intake. No evidence was found on the critical outcomes of remission and binge eating, nor on the important outcomes of quality of life, all-cause mortality, general psychopathology, relapse, general functioning, family functioning, resource use, and service user experience. |
Trade-off between net health benefits and resource use | The committee considered that providing care for people with an eating disorder who are misusing substances or medication may have resource implications in terms of the extra time required to facilitate care for such people (in particular the use of a multi-disciplinary approach). However, the committee expressed the view that if such care leads to better identification of health needs and this results in appropriate treatment and management of health problems (including eating disorder and substance and medication misuse) at an earlier stage, before individuals require more resource intensive management, then the additional costs associated with facilitating a multi-disciplinary care is expected to result in improved health outcomes in the longer term and potential future cost savings to the healthcare system. |
Quality of evidence | The evidence used to generate these recommendations was very low quality. In the absence of RCT evidence the committee considered a single study with 119 participants who received a CBT-ED programme. However, there were few outcomes reported and no remission data. In the other study, again there were 81 participants but there was no data at the end of treatment (only at follow up) and again few outcomes were reported. They did however measure remission. |
Other considerations | Limited published evidence was found on individual eating disorders, so the committee generated a recommendation incorporating the evidence from people with bulimia nervosa and EDNOS and made it relevant for treating people with any eating disorder and a substance misuse problem. The observational evidence suggested that people with bulimia nervosa or those with EDNOS, with a low or high alcohol intake, may be equally responsive to an eating disorder treatment. And for people with bulimia nervosa only, a positive long-term response to treatment may be equally found in those with a history of substance misuse as those with no history. Thus, the committee recommended that for people with an eating disorder who are misusing substances, offer treatment for the eating disorder unless the substance misuse is demonstrably interfering with this treatment. The observational evidence also suggested that treatment for an eating disorder may also reduce alcohol intake. In the study by Karacic 2011 over half the high alcohol intake group were no longer drinking excessively (52.8%, n=19) at the end of treatment, however, 12.5% (n=10) of the low alcohol intake group were now drinking above the safe limit (this data was not extracted because change scores were not presented). For these reasons, the committee decided using informal consenus that the person should be offered treatment for the eating disorder unless the substance misuse is interfering with the treatment. Although the evidence for this recommendation was based on a study that used a mixed sample of people with bulimia nervosa and EDNOS, the committee were confident that the findings would translate to those with any eating disorder. For this reason, they did not specify the type of eating disorder in their recommendation. For further details see the LETR in Section 7.8.6. |
- Treatment and management of atypical eating disorders (eating disorders not othe...Treatment and management of atypical eating disorders (eating disorders not otherwise specified) - Eating Disorders: Recognition and Treatment
Your browsing activity is empty.
Activity recording is turned off.
See more...