A biopsychosocial proposal to progress the field of anorexia nervosa

Anorexia nervosa (AN) is associated with the highest death rate of any mental illness, among the highest relapse rates of any psychiatric condition, and a long-term recovery rate of less than 50% among those who survive it. Despite AN first being described in the scientific literature in the late 19th century, we have arguably made little progress in uncovering the aetiological and maintenance factors involved in the illness, and in developing effective treatments. This is in stark contrast to many other mental illnesses (e.g. schizophrenia, major depressive disorder) which have evidence-based treatments available that are largely successful in remitting symptoms. A recent systematic review and meta-analysis concluded that specialised eating disorder treatments are no more successful than control treatments in remitting the psychological symptoms of AN at end of treatment or follow-up; instead, they are only successful at increasing body weight at end of treatment which does not carry through to follow-up (Murray et al., 2019). Indeed, our current clinical practice guidelines for the treatment of eating disorders conclude that there is limited evidence to support current therapies for AN given the scarcity of randomised controlled trials (Hay et al., 2014). Why is this the case? Why have we failed to develop evidence-based treatments to specifically treat AN while so much progress has arguably been achieved in other mental illnesses?

It is widely acknowledged that all mental illnesses involve, at least to some degree, both environmental and biological influences, i.e., nature and nurture. In AN, the biopsychosocial model of the illness is well accepted, with biological, psychological and sociocultural factors all thought to contribute to the development and maintenance of the condition. Historically, the greatest focus in AN research and treatment has, however, undoubtedly been on environmental factors contributing to the illness. The biological underpinnings of the condition, on the other hand, have been largely overlooked; and when biological mechanisms have been examined, there has been little consistency and replication between studies (Phillipou et al., 2014). Partly due to advances in technology, recent times have seen greater focus on the biological contributions to AN, including behavioural determinants of biological features on long-term outcomes, but still to a lesser degree than the psychological and sociocultural factors involved in the illness.

The lack of a comprehensive biological framework has resulted in a paucity of neurobiologically inspired therapies specifically prescribed to treat AN, including psychotropic medications and brain stimulation techniques that are available to successfully treat other mental illnesses. Currently, the psychotropic medications prescribed to individuals with AN are often to alleviate the frequently comorbid symptoms of depression and/or anxiety rather than to treat the core neurobiological features of the illness. An additional problem is that the different factors of the biopsychosocial model of AN are not regularly assessed synergistically, and the majority of the literature in AN focuses on single or specific aspects of the model without integrating environmental and biological components.

Almost all factors that are thought to contribute to the aetiology and maintenance of AN, or act as barriers to successful treatment, involve an interaction of environmental and biological influences. Figure 1 represents a proposed model of this interaction, with factors theoretically arranged by the degree of environmental and biological load. This model does not encompass all factors proposed to be involved in AN in the literature, but a summary of some of the most important influences described to date. Sociocultural and psychosocial factors are proposed to load the heaviest on the environmental end of the spectrum, whereas genetic influences (i.e. DNA) are proposed to load the heaviest on the biological end. However, even the factors on the opposite ends of the spectrum are not clear-cut. Psychosocial factors such as exposure to trauma, load heavily on both the environmental and biological ends of the spectrum.That is, exposure to trauma is a psychosocial influence on psychopathology, but extensive literature has established that there is a neurobiological predisposition to trauma susceptibility and resilience (Feder et al., 2009). Similarly, those loading on the biological end are also heavily influenced by environmental factors. The gut–brain interaction, for example, not only has a biological predisposition, but ingested substances (e.g. food) also alter the gut’s microbiome affecting the gut–brain relationship. RNA or gene expression is another example of this interaction, whereby certain genes are ‘turned on or off’ based on environmental factors.

Figure 1.

Model of factors proposed to contribute to each of (1) the cause of anorexia nervosa (AN), (2) the maintenance of AN and (3) are barriers to successful treatment and recovery of AN, with varying levels of environmental and biological load.

Assessing environmental and biological factors involved in AN independently does little to progress our understanding of this illness. For us to advance our knowledge, and develop more effective treatments for AN, we need to evolve comprehensive models that encompass both the environmental and biological influences that contribute to the aetiology and maintenance of the illness. Like other serious mental illnesses, AN is unlikely to have its aetiological origins in a single combination of factors. Instead, various combinations of factors that lead to the same illness in different individuals are predicted. It is also critical to identify risk factors in patients who are early in the illness with the aim of prevention (Kaplan and Strober, 2019), as well as assessing individuals with severe and enduring AN to identify factors leading to chronicity.

For AN research to move rapidly forward, large comprehensive assessments across multiple levels of function is urgently required. Identifying the relationships and the directionality of those relationships between these variables will enable us to cluster sub-groups of people with AN. Symptom clustering will allow us to implement personalised treatment involving the combination of psychological and biological interventions specifically targeting the underlying factors involved in the aetiology and maintenance of AN for that individual. Given the difficulties disentangling the influence of starvation on findings in AN research (Phillipou et al., 2018), it is critical to implement this research agenda in those who have (1) a current diagnosis of AN, (2) those who are long-term weight-restored but with lingering psychological symptoms (partial remission), (3) those who are in full remission or recovery (long-term weight-restored and remission of psychological symptoms) and (4) those with atypical AN who have never been severely underweight. Alternatively, removing the weight criterion from AN group distinctions (Phillipou and Beilharz, 2018) and investigating the influence of different aspects of physique (including, but not limited to body mass index) along a spectrum will lead the field forward.

Recent funding boosts and initiatives in eating disorder research and treatment in Australia and around the world represent significant progress in advancing this area. We, however, still have a long way to go. For the field of AN to progress, we need to not only accept the biopsychosocial model of AN but also to actually implement large multidisciplinary initiatives into our research agenda to enable fruitful advances in AN treatment.