Should we shed the weight criterion for anorexia nervosa?

According to current diagnostic criteria, anorexia nervosa (AN) is characterised by a disturbance of body image and a restriction of energy intake leading to very low weight (Diagnostic and Statistical Manual of Mental Disorders 5). Both illness severity and remission are determined by arbitrary body mass index (BMI) cut-offs in the Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5). Most other psychiatric conditions do not include a physical criterion (and arguably, secondary feature such as weight loss; see Phillipou et al., 2018) for diagnosis. Although low body weight has been identified as a potential predictor of long-term outcome in AN (Löwe et al., 2001), utilising BMI as a diagnostic criterion provides little benefit. Severity of AN is also based purely on BMI and does not take into consideration psychological symptoms, physiological consequences or daily functioning. A recent study by Dakanalis et al. (2018) highlighted the limited empirical evidence for these severity ratings, demonstrating in their study that individuals with AN did not differ in psychiatric comorbidity or psychological distress based on BMI classifications. BMI results are also insufficient to determine medical stabilisation, as some individuals with AN can experience significant physiological symptoms at a higher weight (which may even be in the ‘normal’ BMI range), whereas others at a very low body weight may have minimal physiological concerns. The definition of remission provided by DSM-5 also relies heavily on BMI and is vaguely described as a ‘sustained period’ of weight restoration (partial remission; utilising BMI only), and psychological symptoms (full remission; utilising both psychological symptoms and BMI). Arbitrary BMI cut-offs are used in this way to determine ill from remitted states, without taking into account factors such as age, gender, illness trajectory, bone or muscle mass, or other aspects of physique; factors which have also been recently highlighted as limitations of using BMI to determine physical health at the overweight/obese end of the BMI spectrum (Kaufmann et al., 2018). Utilising BMI for these purposes in AN, and having such a great focus on weight, can potentially be damaging for a number of reasons. It implies that in order to have AN you must be within the ‘underweight’ BMI category, whereas the same psychological and physiological symptoms can be present in individuals of different body sizes experiencing restricted energy intake (e.g. as in ‘atypical’ AN). The severity of symptoms may be better understood by exploring the proportion of body weight loss over a particular time, intensity of eating disorder thoughts and behaviours, and levels of distress or preoccupation.

Those who have undergone weight restoration and have maintained this for a ‘sustained period’ of time are technically in a stage of remission. This ignores the fact that psychological and cognitive aspects of the illness may, and are likely to, continue to be present particularly for those who have undergone involuntary admission to hospital. While restoration of body weight may reduce physiological symptoms of starvation for some individuals, others may need to restore above the ‘healthy’ BMI cut-off for full stabilisation. Re-feeding is distressing to a patient who is intensely fearing gaining weight, yet is a necessary component of AN treatment to restore cognitive capacity and physiological functioning. Once a patient is medically stable, however, should we be placing greater focus on treating the psychological aspects of the illness in order to support weight restoration, rather than making weight restoration the priority?

Although both psychological and physical symptoms of AN are addressed in treatment services, our treatments and treatment outcomes are largely focused on weight and BMI. Inpatient services in particular prioritise weight restoration, with the primary purpose of restoring physical and cognitive capacities. However, once this is achieved, patients may be discharged into the community or to less intensive services despite the patient continuing to meet all other criteria for AN. Public inpatient services have limited resources, so there is a much greater focus on partial remission (weight restoration) rather than the lengthy treatment required for full remission; this is the case for intensive care units and general hospital wards in particular. This, however, can lead to patients cycling between higher and lower levels of care when weight is perceived as the most important outcome. This reflects the excessive relapse rates in AN as the underlying symptoms of the condition are not being adequately addressed. Greater focus on psychological, cognitive and psychosocial function as outcomes of remission and recovery is greatly needed and has been suggested by others (Bardone-Cone et al., 2018).

Voluntary outpatient services are also arguably placing too much emphasis on weight during recovery. Many programmes have a focus on weight restoration, and consequently participants may be discharged into the community or to more intensive services if they maintain or lose weight during treatment, which can be interpreted by some as a punishment for being unwell. Yet the very reason these individuals seek treatment is to develop a better relationship with food and their bodies, despite the illness driving a desire to lose weight. Such attitudes can also perpetuate the belief that individuals with AN of a higher BMI may not be ‘sick enough’, and consequently drive them to lose even more weight to feel worthy of receiving validation and support.

Restoring nutrition and maintaining a healthy body weight that is specific to the individual is an immensely important component of AN treatment. However, having a weight criterion for the illness potentially does more harm than good. Removing the weight criterion will ensure treatment continues beyond mere weight restoration and will allow us to focus on treating the mental illness that is driving the physical consequences of the disease, as a collection of symptoms and behaviours, without pathologising body shape or size. It will also free up confusion and inaccuracies related to severity and remission, and it will enable us to diagnose individuals with AN who may not present at a low weight but meet all other psychological criteria, including those who are weight restored from AN and those with ‘atypical’ AN who present with the same psychological symptoms as ‘typical’ AN but who are not deemed underweight. This process may be similar to removing the amenorrhoea criterion from DSM-5 which validated the presence of AN in people of all genders, and unwell females who were still menstruating. Shifting the focus from BMI to eating disorder behaviours, thoughts and emotions will also support patients, carers and the general population in developing a better understanding of the disorder, thereby reducing stigma and other barriers to recovery. Furthermore, removing the BMI criterion for AN will help shift the field away from weight stigmatisation and towards a more inclusive approach of providing equal and accurate diagnosis and appropriate treatment for all individuals.