The prevailing phenotypic model of FA is based on similarities between certain aspects of overeating and the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV) criteria for substance addiction (9
). This similarity has been formalized in the Yale Food Addiction Scale (YFAS) (19
), a measure that is forming a cornerstone of the human literature on FA. Devising this scale has necessitated confronting a number of difficulties posed by the fact that, first, food, unlike drugs, is consumed ubiquitously and does not have a simple direct pharmacological action. Therefore, its use and misuse cannot easily be quantified, nor can one identify features of its consumption that indicate a clear transition from use to abuse/addiction. Moreover, certain useful indicators of substance dependence, such as tolerance, withdrawal and expenditure of effort to acquire the addictive substance, require careful thought when translated to the food domain. In obviating these difficulties, the design of the YFAS has had to adopt certain adaptations that have their own limitations. For example, given that there is, as we have discussed, no universally agreed evidence of an addictive agent and that eating behaviour is necessarily part of a continuum, the scale does not have the benefit of being able to dichotomize (is an addictive agent used – yes or no?). It must instead apply severity thresholds and an overall impairment criterion (i.e. food-related behaviour causes significant distress or impairment) in order to distinguish between someone who is addicted and someone who is not. Likewise, with respect to withdrawal symptoms, the scale enquires about ‘anxiety, agitation or withdrawal symptoms …’, but the latter are not, and cannot yet be, clearly defined.
The YFAS was developed with the aim of identifying and quantifying a specific clinical phenotypic entity. A score of ≥3 with the impairment criterion (shown earlier) satisfied is required for the diagnosis of FA. However the score has also been used as a continuous severity measure in individuals who do not endorse sufficient criteria for the diagnosis (see (20
)) although it is not clear if there is evidence to support this implied continuum.
The YFAS is undoubtedly an important research tool; however, it does not follow that the syndrome it captures is necessarily FA. It is likely, although, that individuals who endorse the YFAS criteria for FA have a behavioural phenotype with significantly disordered eating behaviour. Whether this is sufficient to define a FA syndrome is debatable.
It is worth pointing out to some important points regarding tolerance and withdrawal. Although these are important considerations in clinical drug dependence, it is recognized that they are not necessarily core elements of the syndrome (21
), representing, rather, features that indicate prolonged consumption with psychological and physiological adaptations. Indeed, it is a criticism of the DSM-IV criteria for substance dependence that they aggregate core features, such as maintained use despite negative consequences, with markers of long-term use such as tolerance and severity of impairment, e.g. effort spent in acquiring substance. Tolerance and withdrawal relate strongly to the mechanistic action of the addictive substance. Further, they highlight a crucial aspect that has not been very prominent thus far in the FA literature: substance addiction is a disorder with a natural history and course and a set of vulnerability or risk factors. If we are to consider that FA is a disorder then it would need to be similarly characterized.
Before we move on, it would be worthwhile to consider briefly a related and more nuanced view that draws another parallel with substance use disorders: the possibility of food abuse or misuse, i.e. harmful use that is maladaptive, but does not meet the criteria for dependence. Substance abuse is characterized by recurrent use of the substance with one or more of the following features: failure to fulfil role obligations, use in harmful situations, consequent legal problems and persistent use despite negative consequences (23
). Given that the behaviours in the food context are part of a continuum of consumption behaviour, one could posit the existence of a food abuse syndrome either as an intermediate stage before the transition to FA or as a less severe pattern of disordered eating. It is our view that such an exploration will become crucial in characterizing the natural history and the neural basis of FA. That is, a close scrutiny of the transitions from use to abuse to addiction will be critical in elucidating the development of the syndrome. However, the merest glance at the criteria for substance abuse makes it clear that translating these criteria to food will pose similar problems to those encountered with the FA model. This brings us to a final concern about a phenotype-based definition of FA: the clinical syndrome of substance addiction may not be the best framework to characterize FA. Perhaps, the way forward might be to outline a more precise neurobehavioural syndrome in which a core set of measurable behaviours is clearly defined (inability to control consumption, increased motivation to consume and persistent consumption despite negative consequences (21
)). This would capture a range of problem-eating behaviours, including, but not restricted to, binge eating.
In considering the link with obesity, FA may be a cause, a comorbidity or possibly a consequence of obesity and may therefore prevail in non-obese and not-yet-obese individuals. This is not to say that obesity is not a potential surrogate marker of the syndrome if one bears in mind individual vulnerability, and the duration and severity of weight gain. However, it does seem that, as has been argued, BED is a more fruitful area for further exploration of FA as, by definition, it includes an abnormal compulsive eating behaviour that is causing significant impairment and distress. It also has a strong association with obesity (24
). We, therefore focus on BED and this narrower application of the FA model.