*1.1. From the Addiction Medicine Clinician Point of View: Towards a Definition of Food Addiction (FA)*

Addiction remains a difficult-to-define concept. The American Society of Addiction Medicine defined addiction as "a treatable, chronic medical disease involving complex interactions among brain circuits, genetics, the environment, and an individual's life experiences. People with addiction use substances or engage in behaviors that become compulsive and often continue despite harmful consequences" [16]. DSM-5 (Table 1) and ICD-11 give similar sets of criteria defining addiction, while actually avoiding using the term, instead preferring "substance-related and addictive disorders" (including "substance use disorders" (SUD) and "gambling disorder") and "dependence", respectively. These criteria encompass the loss of control over consumption, increased motivation to consume, and persistent consumption despite negative consequences, as well as tolerance or adverse effects of acute withdrawal. In the last twenty years, there has been a growing interest in the possibility that in some patients, food, and especially highly palatable food, could produce behavioral symptoms that parallel those of addiction and could activate the same neural reward circuits as drugs of abuse [17]. However, this concept has been challenged, either on the addictive nature of some eating disorders (ED) [18,19] or by debating as to whether FA is akin to a behavioral addiction versus a SUD [20].

Diagnostic criteria for SUD represent a cluster of cognitive, behavioral, and physiological symptoms, and most of them can apply to some patients by replacing "substance" with "certain food" (for extensive reviews, see [4,7,21]). "Taking larger amounts of the substance for longer periods than intended" has been cited as one of the most commonly reported symptoms in overweight/obese or eating disorder patients. It can be in the form of binges, but also snacking [22], food compulsion, or excessive portion sizes. "Unsuccessful attempts to reduce food intake" is clinically obvious, as many patients are unable to maintain their diet and lose weight in the long term. "Craving" is a central

concept in the field of addiction [23], and craving for food has been recognized for a long time [24–26]. Similarity between drug and food craving is supported by the findings of cue-reactivity research [24]. "Social/interpersonal problems related to use" is supported by the poor social functioning associated with overweight and obesity due to weight stigmatization. Negative consequences of overeating include obesity and its medical consequences, stigmatization, psychological distress induced by shame, hopelessness [27]; then, many patients exhibit "continuous use despite recurrent physical or psychological problem". These negative consequences lead to the "failure to fulfill major role obligation" and to "reduced activities". Tolerance can be suspected, given that some overweight patients increased consumption and portion size over time [21]. This was further supported by an innovative study in which 61 overweight carbohydrate-craving women were induced into a sad mood, then exposed, double-blind and in counterbalanced order, to taste-matched carbohydrate or protein beverages and asked to choose the drink that made them feel better. They overwhelmingly chose and liked the carbohydrate beverage, which was more efficient in reducing dysphoria, but the effect decreased with repetition, suggesting tolerance, while liking increased, suggesting sensitization [28]. At last, a specific food withdrawal syndrome, as can be observed with alcohol, opioids, or nicotine, has not clearly been demonstrated in humans. However, it should be noted that the physiological criteria of tolerance and withdrawal are not necessary for a diagnosis of SUD, as even with a potent substance such as alcohol, withdrawal syndrome is observed in no more than one third of patients.

**Table 1.** Diagnostic and Statistical Manuel of Mental Disorders version 5 (DSM-5) diagnostic criteria of substance use disorder (SUD). A person needs to meet at least 2 of these criteria and have significant impairment or distress from his pattern of substance use to be diagnosed with a SUD. The severity of addiction is determined by the number of criteria met: 2–3 mild; 4–5 moderate; ≥6 severe.


Gearhardt and coworkers have boosted studies on FA by releasing the Yale Food Addiction Scale (YFAS), by transposing the DSM-IV [29], then DSM-5 (YFAS 2.0) [30] criteria for substance dependence, then SUD, simply by replacing substance with "certain food" in the criteria. In comparison to the first version of the YFAS, the YFAS 2.0 explores the following additional criteria, the first three issued

from the previous DSM-IV diagnostic of abuse: continued consumption despite social or interpersonal problems, failure to fulfill major role or obligation, use in physically hazardous situations, and craving. YFAS 2.0 comprises 35 questions exploring the 11 criteria of DSM-5 SUD and the existence of significant impairment or distress and uses exactly the same thresholds to assign an FA diagnosis (Table 1). Of note, DSM-5 allows us to grade SUD as mild, moderate, and severe, according to the number of criteria, and only severe SUD fits well with the diagnosis of dependence in DSM-IV and ICD-11. In other words, sensitivity was increased, and it is questionable to say that a mild SUD responds to the criteria of addiction [31]. Numerous systematic reviews have explored validation, prevalence, and correlates of FA diagnosed by the YFAS [4,15,32–35]. Gordon et al. [4] evaluated empirical studies examining the construct of "FA" and their conclusions supported FA as a unique construct consistent with criteria for other SUD diagnoses. The highest scored symptom was generally "unsuccessful attempt to cut down"; tolerance and "use despite knowledge of adverse consequences" were also frequent, followed by "activities given up" and withdrawal symptoms [35]. In studies using YFAS 2.0, severe FA predominated upon mild and moderate forms [35]. YFAS number of symptoms was correlated with body mass index (BMI) in non-clinical samples; this was variable in obese or ED samples [35]. Significant positive correlations were found between FA and depression or anxiety [32,35]. FA, in obese patients or university students, was consistently associated with self-report and other measures of impulsivity; associations with reward sensitivity were inconsistent, depending on the questionnaires used [34]. At last, FA prevalence was generally increased in patients with other chemical or behavioral addictions as compared with non-clinical samples [36–40].

The triggering mechanism of FA has been extensively debated. Hebebrand et al. [20] have argued that FA may be a behavioral addiction: indeed, most substances of abuse, apart from alcohol, are agonists of specific brain receptors by mimicking endogenous ligands, and this is not applicable to food. Moreover, eating is necessary for survival and drug use is not; eating is intrinsically rewarding and reinforcing, and food consumption is well known to naturally activate the brain reward system. Most other authors, however, considered FA to fit better with SUD [4,7,13,37,41,42]. Two properties of food could participate in mediating liking, wanting, or craving: hedonic taste and metabolic shifts following ingestion. Food contains a variety of compounds that may serve as chemical or metabolic triggers, and all commonly suspected problem foods share nutritive properties [41]. It is highly unlikely that all foods may be addictive, and studies have aimed at identifying the specific foods or food attributes capable of triggering an addictive response. The evidence that sugar (sucrose) could be an addictive substance is mainly supported by animal studies, the interpretations of which are controversial [18,42,43], and sugar (sucrose, fructose) is not considered a direct cause of obesity [42]. In university students, symptoms of FA were in majority related to combined high-fat savory and high-fat sweet foods, and rarely for mainly sugar-containing food [2]. Concerning fat, which has its own metabolic, physiological, and nutritional profiles, human evidence is scarce and comes mostly from studies on FA: individuals with FA had higher dietary fat intake compared to those without FA; animal models suggested that fat addiction may have different mechanisms to sugar addiction [44]. Highly processed foods, with the addition of fat and/or refined carbohydrates (sugar, white flour), have drawn the most attention. Subjects' rating of food pictures varying in their chemical composition showed that highly processed, energy dense foods with high glycemic load and high fat content were most frequently associated with addiction-like eating behaviors, especially for individuals endorsing elevated symptoms of FA [5]. Another study showed that high processed food pictures were associated with greater loss of control, liking, pleasure, and craving, which assess the abuse liability of substances [45]. This was confirmed using a taste test task and ad libitum consumption period in obese women (39% had an FA diagnosis) [46]. Highly processed foods with high glycemic index cause rapid shifts in blood glucose, insulin, and other metabolic fuel and hormones [41], and this could be associated with their addictive potential.

The concept of FA has many implications for the addiction clinician. Given the prevalence of FA in people with addiction, screening for FA and other eating disorders has to be performed in patients with other addictive disorders. FA could explain, by a mechanism of addiction transfer, the increase in the consumption of chocolate and other sweets in recovering patients with alcohol use disorder [47] as well as the tragic increase in alcohol use disorders after obesity surgery [48]. Denial is a well-recognized, albeit badly explained, feature of patients with SUD [49], but has not been explored in the FA literature. Given the importance of social stigma attached both to overweight and addiction [50], it is very likely that denial does exist, particularly in obese/overweight samples seeking treatment. Such a behavioral feature could minimize FA prevalence. Denial is in part related to concerns about being stigmatized or rejected or to social interactions of an accusatory or judgmental nature [49]. There have been extensive discussions about the influence of the recognition of the FA construct on stigmatization, either from family/relatives and society (externalized stigma) or from the patient himself (internalized stigma) [50,51]. An FA explanation model of the lack of control in obesity could decrease stigmatization from others [52]. Studies on the effects on self-esteem and internalized stigmatization gave contradictory results [50,53]. Increasing self-esteem and confidence could then constitute a therapeutic goal in obesity treatment.

In conclusion, the analogy between severe SUD and some eating-related behaviors including FA is obvious from a clinical point of view, although more studies are needed to precisely determine the triggering mechanisms and the possibility of preventive or therapeutic interventions.
