Diagnostic issues

The diagnosis of anorexia nervosa is quite often known before presentation to hospital. However, a number of other conditions can mimic anorexia nervosa because body image disturbance and valuation of thinness may occur in the presence of weight loss due to physical diseases such as Crohn's disease, Addison's disease, malabsorption syndrome and various forms of malignancy, and these and other conditions must be ruled out. Medical investigations to follow up findings (such as raised liver enzymes) can divert attention from the patient's nutritional needs, which may be neglected.

Treatment issues

Treatment of patients with severe anorexia nervosa (Robinson Reference Robinson and Nicholls2015) involves attention to physical, psychological and social aspects of the disorder. These patients, fearful of weight gain, may sabotage nutritional treatment by discarding food or nasogastric feed, vomiting, taking laxatives or over-exercising. They may become depressed and suicidal and may self-harm. Family members are extremely anxious and that anxiety may lead to interventions that may not fit with the treatment programme, for example calling in a medical relative for an informal ‘second opinion’. Management therefore needs to include nutritional rehabilitation, mental health nursing support, supportive psychotherapy, close psychiatric liaison and help for the family and the staff group.

Deaths of patients with anorexia nervosa in medical settings

The epidemiology of this group of patients, including the number of deaths that occur, is unknown, and surveys of medical units are required in order to estimate the prevalence and incidence of admissions of patients with severe anorexia nervosa, and the proportion giving rise to serious management problems and a fatal outcome. A number of deaths have occurred in UK hospitals and are attributed to problems with medical or psychiatric management and particularly to shortfalls in adequate liaison between the two. It is likely that this problem, identified in the UK, is one that affects all countries in which patients with anorexia nervosa are treated in medical units.

In this article, advice is provided for front-line staff dealing with severely ill patients with anorexia nervosa. For brevity, the patient is referred to as ‘she’, although a significant proportion of such patients will be male.

Diagnosis

The latest classification for eating disorders, the DSM-5 (American Psychiatric Association 2013), gives clear criteria for anorexia nervosa. These are listed in Box 1.

In a young person with a history of self-imposed weight loss, sometimes purging food by vomiting, and sometimes engaging in binge eating, the diagnosis is clear. Physical causes of weight loss such as diabetes mellitus, endocrine diseases such as thyrotoxicosis and Addison's disease, and malignant conditions should all be excluded. Treatment of the eating disorder should not be delayed while investigations for medical conditions continue. Most patients brought to A & E departments or admitted to medical units have already received the diagnosis of anorexia nervosa.

Physical examination and investigations

Overview

Assuming that the diagnosis of anorexia nervosa has been made, a patient with severe anorexia nervosa can have a large number of physical problems. Some of the most important are listed here.

• • Emaciation: the patient may look cadaverous and be covered with fine downy ‘lanugo’ hair. She may be shivering and be found to have hypothermia (core temperature <35°C).

• • Weakness: the patient may be unable to sit up or may show weakness on the sit-up/squat–stand (SUSS) test (Fig. 2).

• • Bradycardia, low heart rate: a rate of less than 40 bpm, while awake, has been thought to indicate higher risk.

• • Postural hypotension: a large drop in blood pressure on standing, accompanied by symptoms of dizziness or faintness, indicates that the cardiovascular system is severely affected.

• • Delayed gastric emptying: the stomach slows down in anorexia nervosa, and the patient can have a very dilated stomach, which occasionally can rupture if a large meal or a binge is ingested.

• • Abnormal blood tests: urea and electrolyte abnormalities (Mehler Reference Mehler, Crews and Weiner2004) are common, with hypokalaemia being the most common dangerous condition. However, sodium, urea, creatinine, magnesium and phosphate can all be dangerously abnormal. It is not unusual for transaminase levels to be elevated in an acutely ill patient with anorexia nervosa (Hanachi Reference Hanachi, Melchior and Crenn2013), but apart from a rise during re-feeding, the levels generally settle with treatment, and a major search for a hepatic diagnosis is not usually warranted. Levels of haemoglobin, white cells and platelets can all be low. For further details, see the ‘Blood tests’ section below.

FIG 2

The sit-up/squat–stand (SUSS) test (Robinson Reference Robinson2012).

The role of body mass index (BMI)

There is no doubt that monitoring of body weight is a vital measure of risk in patients with anorexia nervosa. Non-medical personnel (e.g. psychotherapists) seeing patients with the illness need to be able to do some simple physical measures and a risk assessment based on BMI and to get medical advice if concerned. BMI is calculated as the person's weight in kilograms divided by their height in metres squared, i.e. BMI = weight (kg)/height (m²).

In adults over 18 years of age, the following ranges have been suggested for BMI risk in anorexia nervosa (Royal College of Psychiatrists 2014):

• • low risk: BMI 15–17.5

• • medium risk: BMI 13–15

• • high risk: BMI <13

The patient's BMI should always be worked out. However, some clinicians may have difficulty, even with a calculator, and Table 1 may help.

TABLE 1

BMIs and associated risk levels for adults by height and weight

In addition to its diagnostic value in eating disorders, BMI is a proxy measure of medical risk and demonstrates a U-shaped curve when measured against all-cause mortality, with increased mortality rates noted at a low (<20) and high (>35) BMI (Prospective Studies Collaboration 2009).

BMI, however, has its limitations as a risk measure since it can easily be manipulated by fluid-loading and hiding weights in one's clothing, and efforts should be made to minimise and control for such behaviours where possible. Likewise, BMI does not account for differences in muscle:fat ratio, meaning that men in most cases have a higher level of physical risk for each BMI range compared with women. BMI is also less reliable in the following circumstances:

• • extremes of height

• • children and adolescents

• • rapid changes in weight

• • regular bulimic behaviour

• • physical comorbidity

• • pregnancy.

With this in mind, it is important to remember that BMI should always be interpreted in the context of the clinical history and physical examination and that it is not the only measure of physical risk in low-weight patients.

It should be appreciated, though, that patients are fully aware that they risk going to hospital if their BMI is low, and they may go to remarkable lengths to avoid that admission, because it may mean weight gain. Some patients drink copious amounts of water before weighing. In one case, known to one of us (P. R.), a woman who was being treated as an out-patient had an epileptic fit, which was found to be related to severe hyponatraemia (119 mmol/L). She was unconscious for several days and her weight in hospital was 9 kg lower than at her last out-patient appointment, a few days before, suggesting that she had drunk 9 litres of water before weighing. After she woke up she had confided that she had been placing a shower rose in her mouth and turning on the shower to fill her stomach. Other attempts to increase weight include placing objects under clothes and inserting coins into the vagina. Hence, other approaches to assessing risk need to be adopted. Some involve physical investigations and are discussed below. Others require observation and simple testing.

Blood tests

Blood tests (Mehler Reference Mehler, Crews and Weiner2004) are not uncommonly within normal ranges in anorexia nervosa (Table 2). Oddly, this can be dangerous, because some physicians, seeing normal biochemistry in a patient who has been admitted to hospital, assume that it is safe to discharge the patient even if BMI and muscle power indicate severe malnutrition. One patient with a low BMI was discharged from a medical ward. She was able to walk along the pavement, but when she reached a kerb, thigh weakness prevented her from raising her foot to step up. She fell and was readmitted after a formal complaint by her parents. The lesson is that normal biochemistry does not always mean low risk.

TABLE 2

Some of the laboratory abnormalities seen in severe anorexia nervosa

H⁺, hydrogen ions; IV, intravenous; K⁺, potassium cations.

However, the following blood tests can be abnormal in anorexia nervosa and this may increase the risk of death.

Electrocardiogram (ECGs)

The ECG is quite often abnormal in anorexia nervosa. It is thought that QTc interval prolongation (i.e. >450 ms) may be a precursor of sudden death in anorexia nervosa, and for this reason it is regarded as a high-risk finding. Typically, the QTc interval increases as weight decreases and improves with weight gain. Clinicians should therefore be cautious when prescribing any medication associated with QTc prolongation (e.g. antipsychotics) in low-weight patients. Other findings reported in severe anorexia nervosa that respond to weight gain are slowing of the heart rate (bradycardia) and decreased left ventricular mass (Mont Reference Mont, Castro and Herreros2003), as well as T-wave flattening, which can mimic the ECG appearance in angina or a heart attack (Fig. 3).

FIG 3

Electrocardiograms before (left) and after (right) re-feeding. The left-hand trace shows T-wave inversion in V3 and V4 and flattening in V5–6. These abnormalities reversed with re-feeding (Robinson Reference Robinson2006: p. 68). © John Wiley and Sons. With permission.

Capacity

One question that is frequently asked in the case of patients with severe anorexia nervosa is ‘Does the patient have the capacity to make treatment decisions?’ This question must be answered for any patient receiving care and it can be quite complex (Tan et al, Reference Tan, Richards, Robinson and Nicholls2015). Decision-making capacity is assessed using the Mental Capacity Act 2005, and four elements are tested. The British Medical Association (BMA) has published a framework and ‘toolkit’ to help the practical assessment of capacity (British Medical Association 2016). The four elements are as follows.

1. 1 Can the patient understand the information relevant to the decision? Usually, in the case of anorexia nervosa, the answer is yes, the patient can understand, for example, how severe underweight threatens her life. However, sometimes a patient will deny the threat to her life, and she may strongly disagree with the doctor about the severity of her condition, in which case the answer to this question may be no, she cannot understand the information relevant to the decision.

2. 2 Can the patient retain the information? Unless she has serious brain damage, a patient with anorexia nervosa is generally able to retain the relevant information. As a test, she can be asked to repeat it after 5 min.

3. 3 Can the patient use or weigh the information? Adequate testing of this aspect of capacity requires training and experience. It is the element often failed by patients with severe anorexia nervosa. The patient can understand that her physical state leads to serious risks to life and health, but her fear of weight gain leads her to refuse treatment, such as nutritional rehabilitation, because she does not want to gain weight, even though her BMI may be dangerously low. The Mental Capacity Act Code of Practice makes this point: ‘For example, a person with the eating disorder anorexia nervosa may understand information about the consequences of not eating. But their compulsion not to eat might be too strong for them to ignore’ (Department for Constitutional Affairs 2007: para. 4.22).

4. 4 Can the patient communicate the decision she wishes to make? Almost always, a patient with anorexia nervosa can communicate her decision verbally.

Capacity can only be assessed for a particular decision at a particular time. Capacity can fluctuate and vary depending on the intervention being proposed. A person may have capacity to agree to a blood test, but not to a nasogastric feed. The intervention must be specified: in anorexia nervosa, it may for example be the passage of a nasogastric tube, or transfer to a specialist eating disorders unit (SEDU). The date and time that capacity is assessed must be recorded in the patient's record, and the capacity assessment is valid only for that decision at that time.

Problems facing the medical teams

Severe anorexia nervosa presents very difficult medical and psychiatric problems to the treating team. Hence, a team that is lacking one set of skills or the other cannot provide an effective service. How to deal with a falling phosphate level in a patient being re-fed is a medical question. How to approach the problem of micro-exercising or refusal to accept life-saving treatment requires psychiatric skills. The team therefore requires physicians, psychiatrists, medical nurses and mental health nurses, dietitians, preferably including one with experience of managing anorexia nervosa, and, if possible, a psychologist experienced in treating the disorder. This is one situation in which a comprehensive multidisciplinary approach can be life-saving. The team needs to meet soon after the patient's admission and then at regular intervals. Hospitals in which few patients with anorexia nervosa are admitted face a difficult problem: how to make sure that, when such a patient does arrive, she receives expert treatment. One possible solution to this dilemma is to establish, for every hospital, a ‘pop-up MARSIPAN team’ (Royal College of Psychiatrists 2014). This comprises a physician, a psychiatrist, a general nurse, a psychiatric nurse, a dietitian and a manager (and any other interested professional) who have had MARSIPAN training and are ready to advise a treating team at the hospital, should an patient with anorexia nervosa be admitted.

Nutrition

Patients with severe anorexia nervosa will have a number of medical problems, and investigations will reveal which organs and systems are affected. The principles of resuscitation and repair of the severely malnourished state are given in the MARSIPAN guidance (Royal College of Psychiatrists 2014). Nutrition of these patients is complicated and is addressed below, but first we wish to address the management of the behavioural and psychological problems that can lead to nutrition being sabotaged.

Management of behavioural and psychological problems

The treatment of a patient with severe anorexia nervosa involves striking a balance between trust and suspicion (on both sides). Unless the patient has trust in the staff providing care, she is unlikely to improve. She will very likely suspect (rightly) that the aim of the staff is to improve her nutrition, which she is vehemently against. However, she may also be aware that the staff are intent on saving her life, and this is usually something she does want. On the side of the staff, trust cannot be assumed, and staff will be aware that the patient may be behaving in ways that sabotage weight gain. Nevertheless, it is usually not helpful to assume that everything the patient says is false. Again, a balance needs to be struck, and that is not simple or easy, mainly because the real situation (e.g. whether the patient is actually sabotaging her recovery) is not certain. Letting the patient know about the dilemma you face is sometimes quite helpful, and staff can also convey the idea that they understand that the behaviours the patient is exhibiting are driven by irresistible compulsion (rather than wilful badness). Some physical measures can give a clue that the patient is in the grip of a treatment-sabotaging behaviour. These include hyponatraemia (due to water-loading), hypokalaemia (due to vomiting or laxative misuse) and falling scores on the SUSS test when the BMI is rising or staying stable.

Box 3 gives a list of systemic precautions that can be taken in this difficult situation.

Compulsory treatment

Here there is a cultural difference between specialties. Psychiatrists are quite used to their patients refusing essential treatment and to using the laws governing involuntary treatment (e.g. section 12 of the Mental Health Act 1983 in England and Wales). In contrast, physicians, faced with an intelligent and articulate patient, may be very reluctant to ask for a Mental Health Act assessment. From the physician's point of view, if treatment is refused then treatment is not given. The support of the psychiatric team may be required in order to suggest, and sometimes insist, that if the patient is judged not to have capacity to refuse treatment then compulsion under mental health legislation such as the Mental Health Act may need to be considered, to save the patient's life. This can be the case in severe anorexia nervosa. The tests for applying the Mental Health Act to detain a patient for assessment and compulsory treatment are that, in the opinion of two doctors, one a psychiatrist, supported by an approved mental health professional:

1. 1 the person is suffering from a mental disorder

2. 2 and that this is of a nature or to a degree that, despite her refusal to go to hospital, she ought to be detained in hospital in the interests of her own health, her safety or for the protection of others.

These two criteria can apply to anorexia nervosa, which is recognised to be a mental disorder and, when severe, can cause the death of the sufferer. It should be appreciated that, under the 2007 amendments to the Mental Health Act, the doctor in charge of treatment, formerly called the responsible medical officer and now called the responsible clinician, must have generalised training in the use of the Act and also be approved under section 12 of the Act to conduct such assessments. It would be rare for a treating physician who is not a psychiatrist to have such approval, and therefore, in practice, physicians can no longer fulfil the role they used to fulfil (as responsible medical officers) before the Mental Health Act was amended.

Sometimes clinicians encounter the opinion that a patient with anorexia nervosa cannot be treated compulsorily because she does not have a psychotic disorder. This is wrong and the Mental Health Act Commission (1999) cleared up the matter. It stated that anorexia nervosa is a mental disorder and hence someone with the condition can be subject to compulsory admission under the Mental Health Act in appropriate circumstances. It also gave the opinion that feeding is treatment for the eating disorder and hence can, if necessary, be given against the patient's will. Compulsory feeding may be given under section 63 of the Mental Health Act (Department of Health 2015; Curtice Reference Curtice and James2016), which covers treatments that have been directed by the approved clinician (the health professional in overall charge of the treatment), but to which the patient has not consented.

Thiels (Reference Thiels2008, Reference Thiels and Curtice2009) suggested that compulsory admission should be contemplated for a patient who is not cooperating with treatment when the BMI is <13. Thiels also points out that compulsory feeding might contravene Article 3 of the Human Rights Act 1998 (as the European Convention on Human Rights is currently enacted in the UK): ‘No one shall be subjected to torture or to inhumane treatment or punishment’. The patients themselves, however, in general endorse the idea that in some circumstances, compulsory treatment may be required to save life and, in those circumstances, should be provided (Tan et al, Reference Tan, Stewart and Fitzpatrick2010).

Witholding treatment

Under two circumstances the treating team may consider not providing treatment. First, if it appears that treatment is extremely unlikely to improve the patient's quality of life, the team may regard treatment as futile. This is a major decision, and the team, the patient, the family and an independent opinion all need to be involved. Second, a patient may have signed an advance directive when in a better state, indicating their wish not to have treatment should they become ill again. This is also fraught with legal and ethical difficulty and sometimes turns on whether the patient is judged to have had capacity when they signed the advance directive. These difficult issues are discussed by Tan & Richards (Reference Tan, Richards, Robinson and Nicholls2015).

Re-feeding a patient with severe anorexia nervosa

The route and rate of re-feeding are critical decisions. Oral re-feeding is preferable, as it avoids a difficult transition from the enteral to oral route, and it probably has a lower risk of inducing re-feeding syndrome. However, some patients refuse oral feeds, and the decision then is whether to feed the patient orally under legal compulsion, a process that is very difficult to do, or feed enterally using a nasogastric tube. The rate of re-feeding varies according to where the patient is being treated. In a specialised eating disorder unit (SEDU), re-feeding syndrome with a calorie intake of 20 kcal/kg/day is very rare. However, patients on a medical ward are likely to have risk factors known to predispose them to re-feeding syndrome:

• • very low BMI (<13)

• • electrolyte or renal abnormalities prior to re-feeding

• • infection

• • other serious medical problems (e.g. cardiac failure, hepatic failure with major elevation of liver enzymes).

In those situations, it may be decided to give calories at a lower rate, 5–10 kcal/kg/day. However, it must be recognised that this rate is very low (e.g. 150 kcal/day for a patient weighing 30 kg) and carries the risk of underfeeding. This has been termed underfeeding syndrome and has resulted in a number of deaths. The MARSIPAN group of physicians and psychiatrists advise that, when this low rate is used, electrolytes should be measured every 12 h and intake increased in steps to 20 kcal/kg/day within 2 days, unless there is a contraindication such as severe re-feeding syndrome (Royal College of Psychiatrists 2014). This process is illustrated in Fig. 4.

FIG 4

Flowchart for re-feeding patients with severe anorexia nervosa (Royal College of Psychiatrists 2014).

Psychological support

A very sick patient with severe anorexia nervosa is experiencing an extremely frightening time, and the urge to lose weight driven by her eating disorder will sometimes be combined with a fear of death and a wish to be discharged from the general hospital ward. She may also be subject to a compulsory treatment order, which presents other stresses, and her relationship with her family may be difficult. One member of the team should be available to spend some time with her regularly, daily if possible, to give her time to express some of these emotions and for the team member to convey messages from the patient to the team. If, as is often the case, the person supporting her is a junior medical nurse, then that nurse must also be supported and their reports about the patient fed through to the rest of the clinical team. Often, nursing reports are not read by more senior team members, and valuable information is lost (Box 4). The family also needs support. It is often helpful to invite carers, for example parents, to attend a meeting with the whole team and encourage them to ask questions.

Psychotropic medication

Medicating a very sick patient with anorexia nervosa is often difficult. The patient may be very agitated, and treatment with a small oral dose of a benzodiazepine (lorazepam 0.5–1 mg) or an antipsychotic (olanzapine 2.5–5 mg) can be beneficial. It has been suggested that olanzapine has potential to augment weight gain in anorexia nervosa (Jensen Reference Jensen and Mejlhede2000). However, a number of trials and a meta-analysis (Dold Reference Dold, Aigner and Klabunde2015) have not supported its use. For more extreme agitation, for example when a very underweight patient is fighting the insertion of a nasogastric tube, the use of parenteral medication can be dangerous, because in such a patient even a small dose of a benzodiazepine can cause fatal hypotension or apnoea. The advice given in the MARSIPAN report is to treat such a patient in a medical intensive care setting, so that cardiovascular and respiratory support can be given if required (Royal College of Psychiatrists 2014).