25 Eating disorders: recognition, pathogenesis, classification, management, and services for women with eating disorders

Eating disorders (EDs)—notably anorexia nervosa (AN)—have been recognized as clinical conditions since the 19th century. However, until the last 40 years, accounts were usually hidden away in the small print of medical textbooks. In contrast, EDs are now much discussed and commented upon even in the lay press. Most people know something about EDs. Many have opinions about them. Nevertheless, some prevalent attitudes tend to play down their importance as a public health problem. Thus, on the one hand, they may be viewed as severe but rare illnesses that mysteriously blight or even snuff out promising young lives. On the other hand, they may be thought of as common but trivial, merely an exaggeration of the shape and weight preoccupations of adolescent girls and of a few narcissists who perseverate such concerns to an age when they should know better. However, complicating matters further, there is sometimes a smidgeon of admiration of the control shown by people with AN evidenced by the quip, so often trotted out as if freshly minted, ‘Oh! Anorexia nervosa, I could do with a bit of that’. And sometimes there might be a whiff of misogyny in the downplaying of the problem of EDs which, although males can suffer from them, are clearly in the main afflictions of the female.

Of mental disorders that may affect people of either gender, the EDs have the clearest difference in prevalence between females and males. Studies show a consistent skew with typically a ratio of ten or more females to one male (Hoek and van Hoeken 2003  ;  Button et al. 2008). EDs in males are sometimes missed but this is unlikely to be the main explanation. When males do develop EDs, their disorders usually closely resemble those in females (Button et al. 2008). So why are women and girls more vulnerable? The answer remains uncertain. Clearly males and females differ in many ways both biologically and psychosocially. It is less clear which of these differences account for the skew in risk of developing EDs.

Changes in the incidence and prevalence of EDs over just a few decades suggest that psychosocial issues may be of most immediate relevance. There are data on such changes from richer countries (Hoek and van Hoeken 2003). However, there is little evidence from developing countries although the presence of a few cases in many countries has led to the perception of the EDs as a worldwide problem (see Chapter 24, this volume). The usual interpretation of the evidence available is that AN has been present in the young female population in Europe and North America for over 100 years and that there may have been some increase in prevalence over recent decades. In contrast, bulimia nervosa (BN) seems to be a ‘new’ disorder that was fully recognized and named only in the 1970s (Russell 1979; Vanderycken 1994). This recognition was probably a consequence of BN becoming more salient because it was becoming more prevalent. Once a good description and a name were available, more and more cases were diagnosed. Rates of presentation and recognition rose notably during the last two decades of the of the 20th century although there is some suggestions that the peak may have already passed, at least in Europe and North America (Keel et al. 2006).

Such changes over decades suggest environmental and most probably psychosocial factors since the genetic vulnerability of a population cannot change over just a generation or so. However, environmental change may expose a newly relevant genetic risk factor and the apparent heritability may rise. Certainly there are data suggesting that EDs run in families in a way that suggests genetic transmission (Strober et al. 2000; Fairburn and Harrison 2003). The hunt is on for particular genes importantly affecting risk but as with other disorders, genetics offers the prospect of both greater understanding and therapeutic potency in the future, but such benefits have yet to be delivered.

The most intuitively favoured risk factors for EDs are the social influences that value low weight and a slender form. Research evidence provides some support for this popular view (Fairburn et al. 2005). A series of case–control comparisons found that when such ideas are firmly held within the family of origin, they are associated with an increase in the risk of subsequent ED for BN but less so for AN (Fairburn et al. 1997, 1999). However, although such ideas—‘pressures’—are almost ubiquitous for young women in Western countries, only a small minority develop an ED. There must be other factors involved and for AN, low self-esteem and perfectionism seem to be of importance. Indeed there is evidence supporting an array of risk factors unrelated to weight, shape, or eating. These tend to be associated with increased risk of mental disorder in general including ED (Jacobi 2005).

Thus, the stereotype of a girl or young woman at high risk of ED would depict her as living in a culture or subculture where there was pressure to be slim, as herself feeling personally troubled and having a family history of ED. But why is she at greater risk than her brother? Tentative answers might include the more meaningful and certainly more evident changes in body fat and body shape that signal adulthood and reproductive capacity in the female. It is also plausible that young women have a more challenging journey to adulthood and find themselves in societies with more negative experiences, such as various types of abuse leading to a more delicate sense of self when compared to their male peers. Or the difference might come to be understood biologically. It is certainly possible that the most important risk factors for developing an ED are for the present unknown or neglected.

Making the diagnosis of ED is not a major problem when the patient is evidently at a low weight and/or she is forthcoming about her thoughts, feelings, and behaviours. Notwithstanding the reputation of ED patients for misleading others, this is usually the case. Many people with EDs seem to be reluctant patients, but most will give a sufficient account of themselves if the clinician is not too pushy and makes room for the expression and exploration of mixed feelings. In the few cases where this is not so, there may be true diagnostic uncertainty. A set of screening questions is available to aid in the detection of ED but in practice the best screen may be thinking of the possibility of ED and asking questions focused on the particular person (Morgan et al. 1999). Likewise, self-report questionnaires have their place in monitoring change in individuals and describing populations but rarely add much to clinical assessment. In research, the Eating Disorders Examination (EDE) has come to be the gold standard measure of ED psychopathology (Fairburn and Cooper 1993).

Once the broad diagnosis of an ED is made, the available classifications are seriously flawed. The biggest problem is that a large proportion—sometimes a majority—of people present with a clinical ED fitting neither of the two main diagnostic categories, AN or BN. Within the American DSM-IV system (see Box 25.1), such cases are swept into a residual category with no real criteria namely eating disorders not otherwise specified (EDNOS) (APA 1994). The ICD-10 of the World Health Organization is in practice little better (World Health Organization 1992). Cases of EDNOS cannot be dismissed as mild and unimportant since by definition they are of clinical severity. Indeed they have been shown to be of similar severity to cases of BN (Fairburn et al. 2007). The diagnoses fit even less the disorders of younger adolescents and children (Nicholls 2005).

Some respond to this mess by suggesting new categories. The name of purging disorder has been proposed for cases of EDNOS where purging (induced vomiting or laxative abuse) dominates in a normal weight person who does not binge (Keel 2007). Others have advocated a ‘transdiagnostic’ view which questions the proliferation of categories and emphasizes the similarities between the disorders and the way that an individual may commonly move between categories within an illness career (Fairburn et al. 2003). The area remains controversial. It is uncertain what revision will emerge from the classificatory cabal responsible for the DSM-V.

There is a further DSM-IV diagnosis, binge eating disorder (BED) which requires binge eating but without compensatory behaviours and with little specified psychopathology. Strictly BED remains part of EDNOS. The diagnosis is widely criticized but even more widely used (Dingemans et al. 2005). In the clinic, BED is associated with overweight but this is less so in community samples (Fairburn et al. 1998; de Zwaan 2005). In general the relationship between EDs and obesity is muddled. Obesity as such is traditionally excluded from the ED canon.

People generally seek help because they feel unwell. They worry about their health and want to recover. However, it is more complex for some people presenting with EDs, especially AN. They have mixed, even positive feelings about their state. It is others who worry. The clinician needs to make space to explore these conflicted feelings from the outset. If the patient says that there is nothing wrong but nevertheless keeps the first appointment there is evident ambivalence. The assessor can work with this. Priority should be given to developing a therapeutic dialogue and some initial shared understanding of how the person has reached her current impasse.

Assessment also includes general information gathering, history taking, and mental state examination, encompassing a screen for common comorbidity such as depression, obsessional disorder, substance abuse, and assessment of risk of self-harm. However, there should be special enquiry about the patient’s attitude to weight and shape over time and how any changes have related to life events and stressors. How has this person related to her body in the past? Has she ever maintained a stable weight? Or has she battled with her body to maintain a lower than natural weight? What has been the family culture with respect to weight? How have external pressures for slimness been managed? Is there a family history of ED?

Measuring body weight and height with calculation of body mass index (BMI) together with attention to the physical state of the patient are the other necessary components of assessment. Anyone with ED may have physical complications and it is important that someone takes responsibility for the monitoring of the patient’s physical state and associated risks and that appropriate collaborations are arranged when the main clinician is not medical. In general the risks are higher in AN and people often look unwell. However, people with BN or EDNOS at normal weight may look well but nevertheless be at risk through biochemical disruption arising from vomiting or otherwise purging on a daily basis. Review of physical symptoms and systems is necessary together with screening investigations usually a full blood count, urea and electrolytes, liver and thyroid function tests. A physical examination and electrocardiogram would be usual for people with AN whose BMI is below 15. A detailed account of the physical issues is beyond the scope of this chapter. See National Collaborating Centre for Mental Health (2004).

As the clinician assesses the patient, the patient will be assessing the clinician, the service, and what is on offer. By the end of assessment, the clinician should be able to provide a provisional formulation that is accessible to the patient. Thus, the account might tentatively suggest aspects of the history, such as personality factors or parental attitudes to weight which may have rendered her vulnerable and events such as bullying or stressful examinations that may have acted as precipitants. Furthermore, information on the physiological and psychological effects of starvation add to the account of how the patient developed an ED and became trapped within it. For example, a patient may be relieved to hear that her preoccupation with food and fear of losing control of eating may to be a response to dietary restraint and that this traps her within a vicious cycle. Thus assessment may often involve psychoeducation and motivational work which is essentially the start of treatment.

But what treatments should be offered? There are useful reviews of research including the guideline published by the National Institute for Clinical Excellence (NICE) which has special importance in the United Kingdom in that it provides guidance for practice within the National Health Service (National Collaborating Centre for Mental Health 2004). However, as a rigorous systematic review it has wider relevance. It is due for updating, but unfortunately advances since its publication have been modest. Evidence-based practice is possible for BN, but is less so for AN where there is a dearth of relevant research. There is even less evidence about the treatment of EDNOS in general with the exception of BED.

There is a substantial evidence base relevant to the treatment of BN and the form of EDNOS known as BED. The ‘gold standard’ is a specific form of cognitive behavioural therapy known as CBT-BN (Fairburn et al. 1993a). This has been well evaluated and shown to be superior to other treatments with the exception of interpersonal psychotherapy (IPT) although IPT typically takes longer to reach its full effect (Fairburn et al. 1993b, 1995; Agras et al. 2000). Recently Fairburn and colleagues have developed the treatment into CBT-E (e for enhanced) which in one form CBT-Eb (b for broad) incorporates elements of IPT and other less ED-focused therapies (Fairburn 2008; Fairburn et al. in press). These therapies involve about 20 sessions with a trained therapist over about six months. There are group adaptations of both CBT-BN and IPT (Wilfley et al. 2000). Within the trials—and probably in ordinary clinical practice—about half of those having the treatment escape from BN and most remain well. However, half do not. There is little evidence as to what further treatment is best for the unfortunate half but adding antidepressant medication may be helpful. There is extensive evidence that a range of antidepressants and other drugs such as topiramate have some antibulimic effects independent of any effect upon comorbid depression which is, of course, not uncommon (Goldstein et al. 1999; Hoopes et al. 2003). Fluoxetine is the only drug with the licence for BN in the United Kingdom (Fluoxetine Bulimia Nervosa Study Group 1992). Drugs alone are rarely an adequate treatment for BN.

The therapies cited earlier are effective but expensive and adequately trained therapists are scarce. Fortunately there is evidence that some people with BN can benefit from lesser interventions such as the use of self-help books, websites, or DVDs and the like with or without guidance from a professional (Perkins et al. 2006). Offering lesser treatments as a first intervention within a stepped care approach may be sensible (Fairburn and Peveler 1990).

Almost all cases of BN can and should be managed as outpatients although comorbid depression may at times require admission. Furthermore, the not uncommon conjunction of BN and borderline personality disorder with or without recurrent self-harm or substance abuse presents a challenge to most services. Short crisis admissions may be required but there is a widespread impression that long admissions to general psychiatric wards may do more harm than good although special inpatient programmes may have better results (Lacey 1995).

The treatment of AN is less straightforward and management is more often an issue because of physical risk. A significant minority require more than outpatient contact. There is near consensus and some evidence that various kinds of family therapy are the treatment of choice for teenage patients who are still living with their family (Eisler et al. 1997, 2000). The predominant style has come to be known as the ‘Maudsley model’ in North America (Locke et al. 2000). It seeks emphatically to exculpate the parents but nevertheless emphasizes their responsibility to ensure that the young person eats sufficiently to reverse the weight loss. Family counselling where the clinician meets separately with the patient and with the family has been shown to be as effective as the more traditional approach where they all meet together (Eisler et al. 2000). Multi-family therapy where several families meet together is being evaluated (Dare and Eisler 2000).

By comparison there is a dearth of evidence for psychological treatment of adults with AN and no gold standard. Some clinicians are purist and offer their favoured therapy model including CBT, IPT, or cognitive-analytic therapy (McIntosh et al. 2000; Fairburn et al. 2003). However, many clinicians adopt a more eclectic approach combining exploratory approaches with psychoeducation and behavioural techniques, particularly in the initial stages of treatment. In a recent trial, treatment called ‘specialized supportive clinical management’ proved more effective than CBT or IPT (McIntosh et al. 2005).

Good therapy is more than dietary monitoring and encouragement. It also provides a forum for patients to gradually disentangle their emotional issues from food and weight. Therapy should promote better ways of managing emotional states, without resort to the over-control of food and weight and in due course become focused upon getting life ‘back on track’. Grief work may be useful in recognition of time lost to the illness. In general there is a need for the therapist to be flexible especially with regard to pace. The mixed feelings mentioned earlier may become salient again at any stage.

More intensive treatment for AN usually means going into hospital. People with AN may need admission for three broad reasons. Firstly, their physical state may become too complicated or extreme and admission probably onto a medical ward is necessary to contain and reverse the risk. Secondly, comorbid depression and major suicide risk may require admission, perhaps to a general psychiatric ward. Thirdly, the patient may be stuck despite appropriate treatment as an outpatient and admission to a special ED unit may be offered. In this case—arguably in all cases—the decision to admit is best made collaboratively. Ideally the patient should feel that the time is right. If this is not the case, compulsion is available in the mental health legislation of most countries. However, it is the therapeutic alliance however fragile which remains key in managing AN and compulsion will always complicate, if not compromise, that relationship. There is little evidence on the impact on the use of the legal detention on long-term outcome but many clinicians feel that compulsion should be used only as a last resort (Ramsey et al. 1999; Carney et al. 2008). There is a danger that battling with the patient may make later treatment more difficult. A clinician who can hold his or her nerve may find advantages in the longer run. The rocky road of weight restoration is better trodden actively, even on ambivalent but acquiescent feet, rather than with feet which are being dragged. The latter may sprint back to the starting point with rapid weight loss, as soon as enforcement is lifted.

The NICE guideline suggests that inpatient treatment regimes should be ‘structured … symptom-focused with the expectation of weight gain’ with ‘psychological treatment … which has a focus on eating behaviour and attitudes to weight and shape, and wider psychosocial issues’. It advises against ‘rigid behaviour modification programmes’ (National Collaborating Centre for Mental Health 2004). The content and style of inpatient programmes vary but most include consistent boundaries, clear expectations and an initial giving over of responsibility about diet which is then handed back to the patient gradually later in the admission. Most include a structured therapeutic programme targeting ED psychopathology and involving practical tasks such as shopping, meal preparation, and support in the transition back to independent living. Psychotherapy—individual, group, or family—should continue throughout and ideally after the admission.

Although it may be a relief when a patient in a precarious position agrees to admission, close physical monitoring is required particularly in the early days of refeeding because of potentially dangerous biochemical abnormalities including hypophosphataema due to a carbohydrate-induced intracellular shift of phosphate. There is also a risk of thiamine deficiency, hypomagnesaemia, hypocalcaemia, and hypokalaemia, all of which should also be monitored. The medical management of the refeeding of low weight patients is described elsewhere (NICE 2006; Mehanna et al. 2008) Weight restoration can be speedier as an inpatient and may achieve gains as high as 1 kg a week Nevertheless, if the aim is restoration of a normal weight then admission may need to last many months. This is very expensive and long admissions need to be reserved for the most severely afflicted patients. As an unsought consequence inpatient units may come to be full of patients who present especially difficult and challenging problems.

Special services for EDs are patchily distributed. This remains the case in the United Kingdom although such provision is growing. Overall services need to be comprehensive in scope offering both outpatient and inpatient management although the latter may be supplied through a regional network or referral on to a specialist tertiary unit. There is a tension between the ideal of keeping outpatient services local and the need for a special inpatient unit to be sufficiently large—at least four beds—to sustain an appropriate culture. Furthermore, transitions should be handled with care since arrangements that look satisfactory in planning documents may be experienced by the patient as a disconcerting form of ‘pass the parcel’. A move from child and adolescent to adult services may involve shifting from family-orientated practice to one emphasizing personal responsibility. This can be disorientating and unsettling for both patient and family alike. The components of a comprehensive service are summarized in Box 25.2

The EDs are serious mental disorders that should be taken seriously. They are not uncommon and affect at least 2% of most populations of young women in the richer world (Hoek and van Hoeken 2003). Most sufferers recover eventually but EDs afflict and limit people at crucial times in their personal development. For some, perhaps one in ten, their disorder follows a chronic course over decades. A small number meet a premature death through physical consequences or suicide (Neilson S et al. 1998). Furthermore, the burden of caring for someone with a severe ED has been shown to be similar to caring for someone with psychosis (Treasure et al. 2001). Too often the EDs have been marginalized as a public health problem and patients have been poorly managed as individuals. The disorders have been trivialized or even glamourized. They need to be dealt with as disorders that are important, severe, and ordinary.