Eating disorders comprise anorexia nervosa (AN), bulimia nervosa and associated disorders. AN can develop from about 8 years of age, reaching a peak around 15–18, while bulimia nervosa is rare below the age of 13 but becomes more common than anorexia by young adulthood.1 Atypical forms occur at higher rates than full syndrome disorders.

Children and adolescents also present with other clinical eating disturbances, including selective eating,2 and other phobic and obsessional disorders with eating difficulties as prominent presenting features.1 Management of these other eating disturbances will not be reviewed here as they are quite distinct from the classic eating disorders in terms of their core psychopathology and management.

Although research into AN has a long history, the evidence base for the effectiveness of treatments across all age groups is very weak.3 4 Bulimia nervosa, although more recently described,5 has many more treatment trials, which are generally of better quality. A number of systematic reviews have been published (eg, Hay and Bacaltchuk6), but trials in adolescents are lacking. Despite the shortage of research, clinical guidelines, notably the NICE guideline of 2004, provide very useful recommendations for good practice.

CLINICAL FEATURES

The eating disorder syndromes comprise a range of physical, psychological and behavioural features. In AN body weight is maintained at least 15% below that expected, with the consequence that pubertal development is stunted or reversed. This results in either a delay in the menarche or secondary amenorrhoea in females who have completed puberty. Weight loss (or failure to increase weight with age) is achieved by restriction of calorific foods, exercise, vomiting or purging. In bulimia nervosa a persistent preoccupation with eating is present, with craving and consequent binges (often with a subjective feeling of loss of control). Weight is maintained within a normal range by compensatory vomiting or purging. Those with eating disorders suffer a number of abnormal cognitions, the central one that is characteristic of eating but not feeding disorders of earlier childhood being the over-evaluation of the self in terms of weight and shape.7 All other personal qualities and attributes are relegated below the belief that self-worth is dependent on what weighing scales say or the ability to restrict food intake in the face of hunger. Young people may sometimes describe unwanted emotional states as “feeling fat” and equate these with actually being fat.

In AN this cognitive distortion results in dieting behaviour and an intense fear of weight gain and fatness. Most of the other features are secondary to this psychopathology and its consequences (for example severe weight loss and endocrine dysfunction). There is generally no loss of appetite. Weight loss is viewed as an achievement and thus subjects have a limited desire to change.

Dieting may also be an expression of other motives including asceticism, competitiveness or a battle of wills (particularly with parents). Some engage in excessive exercising. Self-induced vomiting and misuse of laxatives or diuretics are practised by a subgroup who may sometimes binge eat.

Poor self-esteem and feelings of ineffectiveness are extremely common, and depressive and anxiety features, impaired concentration and obsessional symptoms are frequently present. Social interest declines as patients lose weight and most become socially withdrawn and isolated. These psychosocial features tend to get worse with weight loss and often improve with weight regain.

In bulimia nervosa, similar preoccupations with weight lead to attempts at weight control, but these are not sustained as hunger reinforces a propensity to binge eat. Vomiting and other compensatory behaviours are then employed to avoid weight gain with the result that weight is maintained within a normal range. Typically a young person develops a characteristic cyclical pattern of missing meals in the early part of the day with bingeing and purging in the evening. The next day, guilt leads to renewed efforts to cut back on eating, with maintenance of the cycle. Young people with bulimia are less often perfectionist, less socially withdrawn and may engage in more challenging adolescent behaviours such as drug and alcohol misuse.

MANAGEMENT

Assessment

Young people are generally brought to medical attention by a parent concerned about dieting behaviour or weight loss. They may present with emaciation, complications of starvation, menstrual disturbance or dietary preoccupation.

There are difficulties in the reliable assessment of younger children who may be unable to describe their thoughts and behaviours clearly, or be unwilling or scared to do so truthfully. Child versions of the Eating Attitudes Test (ChEAT)8 and the Eating Disorder Examination9 have been developed to aid assessment. It is helpful to take an approach which is respectful and age appropriate but recognises the role of parents in providing a developmental history, their perspective on the problem and their potential role in treatment. Time should be taken with the young person alone to enable a sympathetic understanding of their own point of view and a full assessment of their mental state including cognition and risk. Family, social or educational stressors which may be acting as maintaining factors should be explored. In younger patients one should assess the capacity of parents to assume responsibility for eating, activity and exercise planning. The motivation of both the young person and their family should also be considered.

Investigations

Most of the physical abnormalities seen in AN are due to disturbed eating habits and consequent low weight. Hence the great majority of cases are reversed by treatment focused on establishing healthy eating habits and a normal weight. Investigations may occasionally be indicated to exclude other medical conditions. More important though is the need to assess the extent of malnutrition and electrolyte disturbance due to vomiting or purging. Weight, height and body mass index (BMI) should be calculated and compared with BMI centile reference charts (the norms of which increase with age). Stunting may result in an under-estimation of degree of underweight and premorbid height records can be helpful in evaluating this. Temperature, pulse and blood pressure should be recorded and an ECG carried out if the patient is emaciated. Muscle weakness can be gauged by asking the patient to sit up from lying down and stand from a sitting position without using their hands. Urea and electrolyte levels should be measured. Luteinising hormone (LH) levels are often low and show an impaired response to LH-releasing factor. Even when weight is restored, cyclical LH activity may not resume for some time. Ovarian ultrasonography may be useful in pubescent and recovering girls to assess ovarian maturity. There may be raised levels of growth hormone and cortisol and abnormalities of insulin secretion.

Differential diagnosis

Careful history taking usually elicits characteristic attitudes to weight, shape and fattening foods and the diagnosis should not be made by exclusion of other disorders. AN can be distinguished from normal dieting by the degree of emaciation and over-valued ideas about fatness.

Weight loss may result from other psychological or physical illnesses; hyper-thyroidism and diabetes are examples that occur at this age. Organic causes of diarrhoea, such as idiopathic steatorrhea or inflammatory bowel disease, may need to be considered. Amenorrhoea may result from ovarian or pituitary disease, following use of the contraceptive pill or from psychological stress. Depressive and obsessional symptoms may suggest these disorders, and occasionally psychotic illnesses can present with avoidance of food, but the central fear of weight gain and distorted body image generally indicate the true diagnosis.

Physical management

Pharmacological treatment

There are no large scale randomised controlled drug trials of AN at any age.3 This is largely because there have been few promising pilot studies of core features rather than associated comorbid conditions to encourage larger investigations. Most classes of psychotropic drugs have been subject to small investigations. Preliminary findings suggest that olanzapine is often well tolerated and associated with weight gain and overall compliance with treatment as well as a decrease in agitation and mealtime anxiety.10 An early report suggested that fluoxetine reduced the rate of relapse following in-patient treatment, but a subsequent well-conducted study failed to replicate the finding.11 A recent Cochrane review12 failed to endorse the value of antidepressants in AN. Those with eating disorders have high rates of childhood onset obsessive-compulsive disorder and social phobia and these features have been suggested to moderate outcome in AN. Pharmacological intervention for these traits might be an important adjunct to treatment in some cases.13 There is then as yet little empirical support for the use of psychopharmacological interventions targeted at AN specifically. Nonetheless, for those with high levels of anxiety, obsessionality or mood disorder, the use of either atypical antipsychotics or selective serotonin reuptake inhibitor (SSRI) antidepressants or both may be clinically useful.

In bulimia nervosa antidepressant drugs have been shown (in adults) to have an “antibulimic” effect. They often result in a rapid decline in the frequency of binge eating and purging, and an improvement in mood,4 but clinical experience and the limited research evidence suggest that this is often not sustained.14 None of these findings have come from studies of adolescents.

Re-feeding

Weight restoration should utilise the least invasive procedures possible and should be provided within a caring age-appropriate setting (generally the home). There is a lack of consensus regarding oral feeding requirements. A weight gain of around 1 kg per week is generally recommended for in-patients and 0.5 kg per week for out-patients.15 After an initial safe weight has been achieved, the young person’s food intake should be adjusted to ensure that growth is in keeping with normal weight and height trajectories.

Naso-gastric feeding should only be resorted to in the face of persistent refusal to eat normally. Strict behavioural regimes in which young people have to earn privileges through eating and weight gain are not desirable or acceptable as they militate against the therapeutic alliance and there is no evidence that these approaches work, other than by achieving short-term weight gain.4 In the long term, undue coercion may either be perceived by the young person as a recapitulation of abuse or neglect that they may have suffered previously or may reinforce low self-esteem and feelings of ineffectiveness.

When treating a malnourished young person, care should be taken to avoid the re-feeding syndrome, a rare potentially life-threatening disturbance of fluid and electrolyte balance which can follow sudden increases in nutritional intake in those who have been in a state of starvation. This can be achieved by regular monitoring of heart rate, orthostatic vital signs and serum electrolytes including phosphorous, glucose, magnesium and potassium, although it should be noted that total body electrolytes may be depleted even in the presence of normal serum levels. The re-feeding syndrome occurs more commonly with parenteral than enteral feeding.

Psychological management

Individual psychological therapies

Individual therapies based on a variety of theoretical constructs have been recommended for adolescents with eating disorders, but there is little research evidence to support them. A common view is that young people with eating disorders are struggling with dilemmas of adolescence and turn to an intense focus on weight and shape as an avoidance strategy. Hence, individual therapy aims at improving self-efficacy, self-esteem and self-mastery such that the symptoms of anorexia are no longer needed to master the challenges of adolescence. Cognitive behavioural therapy (CBT), meanwhile, focuses on modifying the behaviour and ways of thinking that are thought to maintain the patient’s eating disorder.16 This treatment typically involves about 20 individual treatment sessions over 5 months and has been found to be the most effective treatment for adults with bulimia nervosa.17 Interpersonal psychotherapy (IPT) formulates problems as being maintained through difficulties in current relationships and takes a problem-focused approach to changing them.18 IPT also has a reasonably strong evidence base, but to date (like CBT) almost exclusively with adults with bulimia nervosa.

There is considerable current interest in the importance of motivational interventions for engagement and treatment, based on the trans-theoretical model of change (eg, Geller et al19). This acknowledges the patient’s investment in the disorder.

Family therapies

By far the most studied interventions for adolescents with AN are those that focus on the family.20 There are several published randomised controlled trials (RCTs) of out-patient family therapy specifically for adolescent AN using a model of family therapy that evolved at the Maudsley Hospital in the 1980s. This approach focuses on familial management of the symptoms of AN and their consequences rather than on presumed pathological features of the patient or family.

This approach initially focuses on behavioural change around eating in order to promote weight gain.21 In contrast to seeing family processes as being at the heart of the problem, therapists instead advise parents that their support and commitment are of critical importance, and that they can and must embark on the task of re-feeding their child and prevent severe dieting, purging, over-exercise and other related problems. The therapist consults with the parents, thereby encouraging empowerment and an increased sense of self-efficacy. In this way, the treatment depends on the leverage that parents have with their children to effect behavioural change. Further, by not blaming the parents, guilt is diminished. More recently multi-family approaches have been adopted using the same principles and are undergoing evaluation.

Service issues

In-patient treatment

There are potentially many benefits to admitting a severely ill young person to hospital. These include physical health monitoring, introduction of normal eating habits leading to weight restoration, intensive provision of psychological therapies and respite for the family. In addition to low levels of confidence and poor self-esteem, studies have shown that patients with eating disorders report more functional difficulties in daily living than do healthy controls (e.g. Keilen22) and hospitalisation may provide a welcome temporary escape. However, the extent to which admission improves these areas, particularly when treatment is perceived as coercive or where the emphasis is on the service taking control, has not been adequately studied. The personal costs of admission to the young person, such as disruption of schooling and family life, as well as the additional problem of ensuring continuity of care after discharge when in-patient admission is at a distance from home, should be balanced against the benefits.4 Where necessary, the NICE guideline suggests admission should be made to age-appropriate facilities. It also refers to the importance of documenting the young person’s consent or otherwise and the legal basis for embarking on treatment. The active involvement of parents/carers is also stressed.

In-patient psychiatric treatment programs vary internationally. In the UK they generally involve a combination of nutritional rehabilitation, medical intervention, psychotherapeutic treatment, psychosocial rehabilitation and family therapies. Educational input is variable. Meanwhile, in-patient management in the USA and Australia is generally briefer and focuses on skilled re-feeding in a medical setting, with specific psychosocial interventions being targeted on out-patient follow-up. These differences ultimately lead to highly varying lengths of hospital stay.

A number of specialised day-patient treatments for AN have been developed for adults and older adolescent patients.13 23 Usually they run for 4–7 days per week and consist of supervised meals, a variety of therapeutic groups and sometimes concurrent individual therapy. Family therapy and medical and pharmacological management are often included.

The single review summarising the effectiveness of in-patient and out-patient care concluded that out-patient treatment in a specialist eating disorder service is as effective as in-patient treatment in those who did not warrant emergency admission.24 Furthermore, these authors estimated the costs of out-patient treatment to be approximately one-tenth of the cost of in-patient treatment. However, this review included only one small RCT of service setting in which the majority of subjects were young adults. A relatively large RCT of adolescents with AN has recently replicated these clinical findings25 and endorsed the cost effectiveness of out-patient management.26

Liaison and interfaces

The NICE guideline highlighted a number of problems in negotiating effective treatment plans across interfaces. These include effective referral from primary to secondary care services and between paediatric and mental health services. Because the peak age of risk is in the late teenage years, effective communication is required between adolescent and adult services. Educational issues are important at this age and communication with education around activity levels and exams should not be overlooked, while those travelling away to higher education may be at risk and require access to effective student health services, although they might not readily seek these out.

NEW DEVELOPMENTS

Recent developments have seen a growth in family-based treatments and evidence-based individual therapeutic approaches, including motivational treatments. As specialist child and adolescent mental health services (CAMHS) have developed and outcomes of in-patient management have often not been maintained, out-patient treatments have become the mainstay. Self-help approaches, including those delivered through the internet, such as the BYTE project for bulimia, are growing.

CONCLUSIONS

Clear guidance on the most effective treatments is hampered by a lack of quality treatment studies. However, a wealth of experience has contributed to valuable clinical guidelines to direct management in both a paediatric and mental health setting.