IS OBESITY AN EATING DISORDER?

Obesity is a term used to describe the condition of having an elevated body weight, and is most commonly defined by the construct of body mass index (BMI), a ratio of weight divided by height squared (kg/m2), with cut‐points determining classes for “normal weight” (18.5–24.9 kg/m2), “overweight” (25–29.9 kg/m2), and “obese” (≥30 kg/m2). Obesity is further subdivided into the following classes: class I (BMI = 30–34.9 kg/m2), class II (BMI = 35–39.9 kg/m2), and class III (>40 kg/m2). Obesity has not appeared in the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders (DSM) previously. Although it was considered for inclusion in DSM‐5 (Devlin, 2007), ultimately it was not included. In this chapter we will consider what is known about the etiology of obesity and review conceptualizations for psychiatric disorders. We will then consider research regarding current nosological views of obesity to assess any benefit of classifying it as an eating disorder (ED).

Obesity Epidemiology and Risk Factors

Prevalence and Demographics

The prevalence of obesity in adults in the United States increased from 14.5% in the late 1970s to 30.5% in 1999–2000 (Flegal, Carroll, Kuczmarski, & Johnson, 1998; Flegal, Carroll, Ogden, & Johnson, 2002). The prevalence has stabilized for most groups since that time. The most recent data, collected in 2011–12, show a prevalence of 34.9% for adults overall, with significantly higher estimates of obesity among women, non‐Hispanic Black adults, and middle‐aged persons (40–59 years; Ogden, Carroll, Kit, & Flegal, 2014).

Generally, obesity is thought to result from a combination of reduced activity and increased calorie intake, creating an energy surplus, which is then stored as adipose tissue. But how does this actually occur? What are the factors that have contributed to an increase in the prevalence of obesity in the population?

Although there is no single, recognized cause of obesity, there are identified risk factors. There is a genetic predisposition to developing obesity, with environmental factors playing an important role in pathogenesis. However, the genetic factors predisposing to obesity are not universally evident, and the constellation of environmental factors varies between individuals.

Genetic Risk Factors

Based on data from twin studies, genetic factors account for almost 65% of variation in body weight (Segal, Feng, McGuire, Allison, & Miller, 2009). The first gene identified in the etiology of obesity was that for leptin in the mid‐1990s. Leptin is a hormone produced by adipose tissue that signals satiety and reduces food intake by binding to its receptors in the central nervous system (CNS). Mutations in the leptin gene (or in the leptin receptor gene) that render leptin unable to bind to its receptor are associated with extreme obesity in animal models and in humans. These mutations are exceedingly rare and do not explain the cause of obesity in most cases. Since the identification of leptin, the advent and availability of genome-wide association studies have led to identification of multiple genes associated with obesity. However, the identified genes account for a small percentage of the variation in body weight. One study identified variations in 18 genes associated with obesity that accounted for less than 4% of variation in body weight (Speliotes et al., 2010). Thus, as yet, we do not have a clear understanding of the specific genetic sources of the proportion of individual differences in body weight that is attributable to heredity, nor of which genes these are and how they function to create and sustain an increase in body weight.

Environmental Risk Factors

Epidemiologists have identified multiple environmental factors that are correlated with increasing BMI in the population. These risk factors include increased television watching (Nguyen & El‐Serag, 2010;), which seems to be mediated through an increase in snacking that accompanies TV viewing rather than a reduction in activity level (Jackson, Djafarian, Stewart, & Speakman, 2009). Researchers and public health experts also regularly cite the availability of cheap, high‐calorie but low‐nutrient-dense foods (and the relatively low availability of fresh foods) as contributing factors (Lovasi, Hutson, Guerra, & Neckerman, 2009;). Physical activity levels, which can depend on the quality of local parks and personal safety issues, may also contribute to obesity (Gordon‐Larsen, Nelson, Page, & Popkin, 2006; Nelson, Gordon‐Larsen, Song, & Popkin, 2006). Additionally, there are many medications associated with weight gain, including atypical antipsychotics, some antidepressants (especially tricyclic antidepressants and atypical antidepressants like mirtazapine), blood pressure medications (such as alpha‐ and beta‐adrenergic blockers), antidiabetic medications, some antiepileptic medications (e.g., valproic acid, carbamazepine), and chronic oral corticosteroids (Malone, 2005).

Despite identified risk factors, many people with environmental and genetic risk factors do not develop obesity, and individuals can develop obesity without these risk factors. Additionally, treatments targeting these risk factors have not demonstrated significant, sustained weight reduction for most people. This has led researchers to consider the role of brain function in behaviors related to eating and activity, with neurobiology representing an increasingly active area of interest. In other words, are there individual, neurologically based responses to environmental risk factors and cues that predispose some people to obesity? Our understanding of the role the brain plays in body weight has grown with the availability of neuroimaging. Below, we review current information about neurological processes related to eating behavior, and what is known about their role in obesity.

Eating Behavior

There are several eating behaviors that may be related to weight status. Rate of eating has been assessed in children, both in lab studies and by self‐report (Berkowitz et al., 2010; Murakami, Miyake, Sasaki, Tanaka, & Arakawa, 2012). Berkowitz and colleagues (2010) examined the eating behavior of 4‐year‐olds, 32 of whom were born to overweight and obese mothers and 29 born to lean mothers, in a feeding lab. Rate of eating, as measured by mouthfuls per minute, was associated with increased odds of obesity, higher body fat, and increased skinfolds at the age of 6 years, suggesting that fast eating rate may be a behavioral marker for the development of obesity in children. Further, a Japanese study by Murakami et al. (2012) surveyed almost 16,000 children and 8,000 adolescents regarding their eating rate and weight status. There was an increased risk of obesity for those reporting relatively fast or very fast eating; the odds ratios, respectively, were 2.8 and 4.5 among male children, 2.7 and 5.7 among female children, and 2.3 and 3.8 in male adolescents. No significant risk was associated with eating rate for obesity among female adolescents. As a descriptive feature of binge eating is rapid consumption of food, this is an interesting behavior to assess prospectively, given evidence of its early onset in childhood.

Eating in the absence of hunger is another behavioral descriptor of binge eating that has been linked to overweight status. This variable is assessed in a feeding lab by presenting a meal to children and asking them to eat until they are full. A short time later, they are presented with a buffet of snack foods that they are free to eat. As reviewed by French, Epstein, Jeffery, Blundell, and Wardle (2012), studies consistently find that overweight children eat more in the absence of hunger than do normal weight children. Although there have been few prospective studies, children tend to show increased risk of weight gain with greater eating in the absence of hunger. This eating behavior may indicate problems with lack of satiety as well as increased enjoyment of food, both of which are independently related to increased weight status in cross‐sectional studies of children (French et al., 2012).

Finally, food preferences as influenced by the family of origin and/or by one’s food environment may also profoundly impact one’s eating behaviors. These may include preference and cravings for high‐fat and highly processed food items; frequent consumption of fast food, take‐out foods, or meals at restaurants; and low intake of fruits and vegetables. These eating styles may be influenced by cultural factors, socioeconomic status, time available for food preparation, and modeling of unhealthy cooking in the family. One study of preschool children found that obese children with at least one obese caregiver, as compared to healthy weight children without any obese caregivers, were less likely to have fresh vegetables, more likely to have a television in the child’s room, and less likely to own exercise equipment, as determined by in‐home visits (Boles, Scharf, Filigno, Saelens, & Stark, 2013). Additionally, while eating meals as a family has been linked to lower levels of disordered eating, the quality of meals from a nutritional standpoint influences weight. In a study of adolescent African Americans and their families, those whose caregivers prepared foods with less healthy cooking methods as compared to those preparing meals with healthy cooking methods were at increased risk for overweight and obesity (Kramer et al., 2012).

Review of Hormonal and Neurological Regulation of Appetite and Eating

Appetite is regulated by a complex interaction between neurological and hormonal systems, and the interplay between these and environmental factors. One way to conceptualize obesity is a failure of the homeostatic system that regulates energy intake, energy expenditure, and thus weight. Multiple hormones and peptides play a role in the homeostatic system, including peripherally produced satiety signals and signals that promote hunger and eating. Satiety signals are released from the gastrointestinal (GI) tract and bind at the vagus nerve or in the CNS. These signals involve cholecystokinin, serotonin, peptide‐YY, glutamate, and glucagon‐like peptide‐1 (Faulconbridge & Hayes, 2011). Other signals are released from the pancreas (e.g., insulin, glucagon), adipose tissue (e.g., leptin, adiponectin), and the GI tract (e.g., ghrelin) and act directly on the CNS, either to promote eating or signal satiety. Areas of the brain involved in regulating homeostatic energy balance include the hypothalamus and the nucleus tractus solitarius (Faulconbridge & Hayes, 2011). Is obesity caused, then, by a disruption in the homeostatic system regulating eating and weight? Some imaging studies have shown sluggish homeostatic responses, suggested by the absence of an inhibitory response in the hypothalamus after eating in obese men compared to lean men (Carnell, Gibson, Benson, Ochner, & Gelieber, 2012). In other words, the hypothalamus, believed to be central to maintenance of body weight, shows less inhibition after eating for obese compared with lean participants. Additional imaging studies comparing obese participants with those who were once obese but are now normal weight also demonstrate the absence of an inhibitory response in brain regions involved in homeostasis in both obese and previously obese participants compared with lean individuals (Cornier et al., 2009). Studies comparing lean with obese individuals are limited to yielding data that describe associations, not information about causality. This gut‐CNS system works to regulate energy balance as one aspect of appetite and weight regulation. Factors other than energy needs contribute to appetite and eating behavior, and in some cases may contribute to weight gain. These include the rewarding aspects of food. Neuroimaging has also been utilized to examine the potential role of neurological circuitry and neurotransmitters related to reward in obesity; this topic will be reviewed later in the chapter.

What is a Psychiatric Disorder?

Given what we know about obesity, it is important to consider how this may or may not fit as an ED, which is a form of psychiatric disorder. The conceptualization of psychiatric disorders has changed over the past century. The first DSM was published in 1952, less than 10 years after the end of World War II (American Psychiatric Association, 1952). The classifications were based on the theory that all psychological neuroses and conditions were reactions to conflict between the psyche and the environment (First, 2010). Precise descriptions were not provided in this original document, as it was thought that no one could know what the underlying pathology really was that caused the observable symptoms (Wilson, 1993). In 1968, DSM‐II was published with similar diagnostics as in the first DSM, but the nomenclature was updated to share terminology with the International Statistical Classification of Diseases and Related Health Problems (ICD‐8; American Psychiatric Association, 1968; First, 2010).

In 1980 the DSM‐III introduced a radical paradigm shift to theoretical descriptive diagnoses, as compared to psychodynamically, etiologically based diagnoses, in order to increase diagnostic reliability and produce effective treatment options, such as the use of lithium for bipolar disorder (American Psychiatric Association, 1980; First, 2010; Regier, 2007). Thus, discrete, categorically based criteria were derived for a variety of syndromes, with the hope that disorders defined in this way could be better targeted by pharmaceutical and therapy treatments, similar to medical classification systems. These changes were also aimed at decreasing the stigma associated with mental disorders. The number of categories increased, as did the number of comorbidities, but there were still difficulties with its ease of use in clinical settings, given that patients are unique and heterogeneous. Some of these concerns were addressed with revisions in DSM‐III‐R; for example, diagnostic hierarchies were largely dropped, allowing for the diagnosis of several disorders at once (American Psychiatric Association, 1987; First, 2010; Regier, 2007).

DSM‐IV was released in 1994 (American Psychiatric Association, 1994). Notably, the impact of functional impairment or distress was introduced. Other changes were based on extensive literature reviews, and were made only if there was compelling evidence to do so. Therefore, no sweeping paradigm changes were made between DSM‐III and DSM‐IV, or even with DSM‐IV‐Text Revision (DSM‐IV‐TR; American Psychiatric Association, 2000). Complaints grew, however, including overuse of the “Not Otherwise Specified” categories, which has been problematic for the EDs in particular (Regier, Kuhl, Narrow, & Kupfer, 2012; Over half of persons seeking treatment under the DSM‐IV system typically fell within the Eating Disorder Not Otherwise Specified (ED‐NOS) category (Eddy, Celio, Hoste, Herzog, & Le Grange, 2008; Fairburn & Bohn, 2005).

The later editions of DSM were based on expert consensus. As research continued to advance our knowledge of the genetics and biological basis of psychological functioning, the desire to base diagnosis on these etiological underpinnings grew. With this knowledge, more overlap across disorders was recognized, such as the success of selective serotonin reuptake inhibitors for mood, anxiety, and EDs (First, 2010;). Thus, the directive for DSM‐5 (American Psychiatric Association, 2013) was to try to move beyond a categorical approach to diagnosis toward a continuum approach based on pathophysiological data. Putatively, this would reduce the number of comorbid diagnoses and improve diagnostic validity (Regier et al., 2012). However, the call for such sweeping change was ahead of the scientific data, as the understanding of genetics, neuroimaging studies, developmental models, and basic science, among other topics, still contain large gaps that will not likely be filled even in the coming decade. The field remains a great distance from establishing sensitive and specific pathophysiological tests for identifying psychiatric disorders.

Switching from a category‐based diagnostic system to a continuum‐based system would invalidate our knowledge of treatment efficacy for any given disorder and cause great disruption in treatment delivery and insurance‐based reimbursements. For example, in a spectrum approach, EDs may be grouped as disorders of low weight or disorders of purging or disorders of binge eating, with overlap allowed among categories. There may also be an obsessive‐compulsive spectrum of disorders that could include AN, or an addictions spectrum disorder that could include food addiction and obesity. Short of such a major shift, DSM‐5 (2013) includes dimensional ratings of severity for a set of symptom domains that are transdiagnostic, but not diagnostic continua.

So where does obesity fall in this historical account of diagnostics? As reviewed above, obesity has been viewed most often as a medical disorder. What differentiates this state or condition from, say, hypertension or diabetes, are the behavioral components of eating and physical activity (or lack thereof) that are so central to its development. Practically speaking, these behaviors also contribute to the development of hypertension and diabetes, but no one is proposing to include those disorders in the DSM. Obesity has components of both EDs and substance use disorders (SUDs), as addressed below, so the advent of a spectrum or continuum diagnostic model might better encompass obesity as a psychiatric disorder than the categorical system currently employed.

Obesity as an Eating Disorder

The existing ED definitions are based on categorical descriptions comprised of symptoms and behaviors. These include criteria such as binge eating, overvaluation of weight and shape, and purging. It is fairly clear when these thoughts or behaviors are present during the course of an ED. So, turning to obesity, what universal symptoms could be associated with high body weight? Currently, there are no behavioral, emotional, or cognitive features in the definition of obesity. Devlin (2007) posits that Western cultures have a moralistic view of obesity that places responsibility for weight status on the individual. He points to two problems with this view. First, as reviewed above, obesity is influenced by several factors, such as genetics and the food environment. Second, persons with obesity may maintain their weight—they do not have to be in a constant state of energy imbalance. To the contrary, they could be eating quantities consistent with weight maintenance or loss, much as persons of normal weight may eat for extended periods of time while still maintaining an “abnormal” BMI.

Given these points, diagnostic criteria for obesity would be difficult to define. Eating patterns and behaviors of persons in this weight category vary between individuals at any given time and within individuals across time. Therefore, static criteria, such as a certain number of kilograms gained per week, would not be of much use. A criterion such as “consumption of fast food more than three times per week” would also not be useful, as persons of any weight may engage in this behavior, and many persons with obesity may never eat at fast‐food restaurants.

A requirement of most psychiatric disorders is the presence of distress or impairment in functioning (DSM‐5, 2013). Studies of mood and quality of life suggest that depression increases and quality of life is worsened with increasing BMI (Ul‐Haq, Mackay, Fenwick, & Pell, 2013). Examples of this from clinical experience often include difficulties fitting into booths in restaurants or rides at amusement parks, and having to purchase two tickets when traveling by airplane. Further, some studies suggest that the relationship between BMI and health‐related quality of life is mediated by other factors such as body image (Cox et al., 2011;), binge eating (Ranzenhofer et al., 2012), and employment (Lund et al., 2011). The effects on weight‐related quality of life are most pronounced at stage III obesity, as is increased risk of mortality (Ul‐Haq et al., 2013). Additionally, for some people who are obese, it is the substantial societal bias against them that creates, or at least contributes to, distress or impairment in the obese individual, rather than being a particular weight itself (Puhl & Brownell, 2001;). In these cases, classifying obesity as a psychiatric illness is reminiscent of including homosexuality in earlier DSM iterations. Finally, distress or impairment in functioning could be measured by including obesity‐related medical comorbidities (e.g., diabetes mellitus, heart disease, some cancers, or obstructive sleep apnea), or by disability status due to weight‐related limitations (e.g., lack of mobility, inability to perform one’s duties). Again, these comorbidities may be present in persons of normal weight, and not present in persons of extreme weight, making them nonspecific diagnostic criteria.

Eating Disorders Related to Obesity

There are two categories of disordered eating—binge eating disorder (BED;and night eating syndrome (NES;)—that are often associated with obesity. Stunkard, Grace, and Wolff (1955) first noted the latter association in their paper describing NES among patients seeking treatment for their obesity. Four years later Stunkard (1959) described the pattern of “binge eating syndrome” among a similar group of patients. For many years the attention of researchers turned toward describing AN and BN more rigorously, but, in the 1990s, as the prevalence of obesity rose in the general population, researchers started examining specific phenotypes of obesity that could possibly be targeted and modified, including BED (Spitzer et al., 1992, 1993) and NES (Birketvedt et al., 1999; Stunkard et al., 1996).

Binge Eating Disorder

Prevalence of BED in the general population for the United States falls at 2.6% (Kessler et al., 2013;), increasing incrementally with weight. Among a larger, international sample of over 24,000 participants, people who were overweight (BMI = 25–29.9 kg/m2) were 1.3 times more likely, and those who were extremely obese (BMI > 40 kg/m2) were 6.6 times more likely than normal weight respondents to have a history of BED. BED is also linked prospectively with weight gain. Persons with BED were found to gain a mean of 4.3 kg (9.5 lb) in the year before presenting for treatment, with larger weight gains significantly related to severity of the binge eating (Barnes, Blomquist, & Grilo, 2011). BED was also linked to impairment in role functioning (i.e., work, home, social life, relationships) for just under half of respondents (Kessler et al., 2013). Finally, BED is associated with increased risk (odds ratios) for chronic neck and back pain (1.5), other chronic pain conditions (1.8), diabetes (2.4), hypertension (1.8), and chronic headaches (1.8) (Kessler et al., 2013).

Night Eating Syndrome

The prevalence of NES also increases with weight, ranging from 1.1 to 5.8% in the general population and community studies (de Zwaan, Müller, Allison, Brähler, & Hilbert, 2014; Lamerz et al., 2005; Rand, Macgregor, & Stunkard, 1997; Striegel‐Moore et al., 2005; Tholin et al., 2009;), 6 to 16% in persons seeking outpatient weight loss treatment (Adami, Campostano, Marinari, Ravera, & Scopinaro, 2002; Calugi, Dalle Grave, & Marchesini, 2009; Gluck, Geliebter, & Satov, 2001; Stunkard et al., 1996), and 8 to 55% among those seeking bariatric surgery (Allison et al., 2006; Latner, Wetzler, Goodman, & Glinski, 2004; Mitchell et al., 2014; Powers, Perez, Boyd, & Rosemurgy, 1999; Rand et al., 1997). Tholin and colleagues (2009) showed an increased risk for obesity (2.5 times for men and 2.8 times for women) among participants who screened positive for night eating in the Swedish Twin Registry STAGE cohort. Additionally, among a sample of outpatient psychiatric patients, those meeting criteria for NES were five times more likely to be obese than patients without NES (Lundgren et al., 2006).

NES status also predicts weight gain prospectively. Andersen, Stunkard, Sørensen, Petersen, and Heitmann (2004) reported that women who endorsed night eating in the Danish MONICA study cohort gained 5.2 kg (11.5 lb) more over 6 years than women without night eating; this trend was not shown for men. Gluck, Venti, Salbe, and Krakoff (2008) studied participants in an inpatient study who ate ad libitum during a 3‐day monitoring period. Those who ate after 11:00 p.m. gained a mean of 6.2 kg (13.7 lb) as compared to 1.7 kg among those who did not eat after 11:00 p.m., measured 3.4 years later, on average.

Finally, night eating may exacerbate medical conditions such as diabetes mellitus. Morse, Ciechanowski, Katon, and Hirsch (2006) found that diabetic patients with evening hyperphagia were more likely to have hemoglobin A1c values greater than 7 (indicating consistently elevated blood sugar) and to have two or more diabetic complications. Hood, Reutrakul, and Crowley (2014) also recently showed that night eating was related to hemoglobin A1c values exceeding 7. Thus, these results from various populations and methodologies strongly suggest that NES may contribute to weight gain or, at the least, the maintenance of higher weight, in those who suffer from it compared with those who do not.

Eating Disorders and Emotion Regulation

For all of the EDs, there is evidence supporting an affect regulation/stress response model, which posits that disordered eating behavior functions to regulate emotions, especially negativeemotional states (Devlin, 2007). For people with BED who, by definition regularly engagein overeating behavior, the model posits that they have difficulty tolerating emotions such as sadness, anxiety, and anger. These aversive internal states are temporarily but quickly relieved by binge eating, which reinforces binge eating as a coping skill (Telch, Agras, & Linehan, 2001). In other words, people with BED are conditioned to binge eat as a means of removing a negative stimulus (i.e., binge episodes are negatively reinforced). However, while there is a higher rate of obesity in populations with BED and NES, the majority of people whose weights are classified as obese do not regularly engage in binge‐ or night‐eating behavior, and this ED model does not seem to apply for this population who are not overeating in response to aversive stimuli.

Reward Circuitry: Obesity as a Disorder of “Addiction”?

An alternative conceptualization of obesity as a psychiatric disorder would be as an “addiction,” or the misuse of food as a SUD. Addiction models posit that eating is intrinsically rewarding, which may be on the same spectrum of conditioned responses as BED (i.e., eating is immediately rewarding). There are data supporting the hypothesis that people with obesity may be predisposed to such conditioning, and in fact may be somewhere on the spectrum between lean individuals and those who engage in binge eating behavior. Neuroimaging studies suggest that people with binge eating have increased responses to food cues in areas of the brain associated with reward, motor planning, and cognitive control; these responses are present but less intense in obese people who do not binge eat, who in turn have more pronounced responses than people whose weight is in the normal BMI range (Carnell et al., 2012).

Compared to normal‐weight participants, obese participants show increased activation in brain regions associated with reward processing (anterior cingulate and medial prefrontal cortex, ventral and dorsal striatum, insula, amygdala, orbitofrontal cortex [OFC], hippocampus, and ventral pallidum) in the presence of food cues in a fasting state (i.e., anticipatory activation), suggesting they have an increased anticipatory food reward. Those with obesity also show differences in a fed state, compared to normal‐weight controls, in response to food cues: greater activation in lateral OFC, caudate, and anterior cingulate cortex. Other studies report reduced activation of the caudate in obese individuals in the fed state, and weaker activation may suggest lower actual reward (De Silva, Salem, Matthews, & Dhillon, 2012). Stice, Spoor, Bohon, Veldhuizen, and Small (2008) postulate that obese individuals have higher anticipated reward and reduced consummatory reward, which together contribute to overeating. Additionally, studies find an association between obesity and increased activation in response to visual, olfactory, and gustatory food cues in brain regions involved in reward, motivation, emotion, and memory (Carnell et al., 2012). Thus, some people may experience abnormal activity in brain reward circuits in response to food and eating, predisposing them to energy imbalance and obesity. This is similar to individuals with SUDs, who show increased activity in these circuits in response to psychoactive drugs (Dawe & Loxton, 2004).

Volkow, Wang, Tomasi, and Baler (2013) outline and summarize the evidence supporting a neurobiological overlap between obesity and addiction. They note the common involvement of dopamine circuitry in food intake and drug intake, and that the changes that occur when an individual progresses from drug use to abuse or dependence and associated compulsive behaviors are also evident in studies of obesity. Volkow et al. (2013) also note that obese subjects show decreased activity in frontal brain regions associated with executive function and cognitive control. This is a consistent finding in drug‐addicted patient populations as well. Further, there is circuitry involved in motivation that is dysregulated in drug‐addicted and obese populations. These changes are associated with enhanced motivation to seek drugs, despite the negative consequences of doing so. Volkow et al. further outline the abnormal functioning of the insula, a brain region related to interoceptive awareness and involved in the integration of homeostatic input with emotion and motivation. The insula is implicated in cravings for food, cigarettes, and cocaine, and appears to function differently in obese than in lean subjects.

However, as Ziauddeen and Fletcher (2013) suggest, there are limits to our current understanding of the neurobiology of obesity as a type of addiction. They identify shortcomings with using BMI alone as a marker of addiction, as there are people in the “normal” BMI range who may show patterns of behavior more readily described as food addiction than do some people with a BMI exceeding 25 kg/m2. Moreover, despite the great deal of enthusiastic attention paid to “food addiction,” evaluations of the function of brain regions known to be associated with drug addiction frequently yield conflicting findings within and across studies of obese populations.

Diagnostic Considerations: Can Eating Be an Addiction?

DSM 5 (2013) changed the conceptualization of addictive disorders, with substance abuse and dependence being combined, and the section titled “Addiction and Related Disorders.” This broadening allows room for other types of addictive behaviors seen clinically and discussed in the media, such as gambling, Internet, and sex. However, a review by Moreno and Tandon (2011) found that no consensual definition of “food addiction” exists. Several details would need to reach consensus for this discussion to continue, such as evaluating weight versus calories; how versus the type of foods one eats; and the temporal pattern of food intake (Moreno & Tandon, 2011). Finally, we should consider how closely overeating parallels the DSM‐5 addictions criteria.

There are several concepts that match between SUD and “food addiction.” First, they are both maladaptive patterns of behavior that tend to be associated with clinically significant impairment or distress. SUD requires use in physically hazardous situations; while there is no direct connection, perhaps eating in the presence of obesity‐related medical comorbidities would fit. SUD also requires substance use resulting in failure to fulfill major role obligations; eating does not typically interfere to this extent (Moreno & Tandon, 2011).

More problems arise as we examine the role of tolerance. Tolerance in SUD is seen in the need for an increasing amount of substance to achieve the desired effect, and a diminished effect with continued use of the same substance. While persons with BED may increase the amount they eat over time, this is generally not the effect seen with overeating. In particular, there is no diminished effect (Moreno & Tandon, 2011). In fact, as Wilson (2010) states, children crave and eat more sweets than adults do. This pattern is the opposite of that described for tolerance in substances. If there were a true sugar addiction, we would see that people needed more and more, say, lollipops, over time, to get a desired reduction in cravings. The concept of withdrawal also remains problematic, as there is no clear parallel between a withdrawal state from substances and that from craved foods. Although some patients describe distress and continued cravings, this experience is not well defined, nor do we understand how common it is.

From a treatment perspective, there are several issues with considering obesity, or overeating, an addiction. Cognitive‐behavioral therapy (CBT) is a treatment that challenges patients to increase the flexibility of their eating and to eat regularly throughout the day to increase control over their eating. It does not suggest that patients remove trigger foods or certain food groups from their diets altogether. As CBT is considered an effective treatment for bulimia nervosa (BN) and BED, the addictions treatment model of avoiding particular foods flies in the face of the evidence that individuals with impulsive or compulsive eating can, in fact, successfully incorporate a wide variety of foods in their diet with treatment (Wilson, 2010).

Therefore, while the brain circuitry may be quite similar between persons with SUD and those with BED or overeating issues, there are many differences between the two that call for further research before overeating is considered an addiction. This is not to say that consideration of calorically dense foods with little nutritional value should not, perhaps, be studied more closely regarding their impact on those who are most susceptible to overeating, as well as on the general population. Gearhardt and Brownell (2013) propose optimizing policies regarding the availability of such “addictive” foods or their taxation, much as cigarettes and alcohol have been regulated to minimize their impact on those with addiction issues and on society more generally. Some such initiatives (e.g., taxing sugared soft drinks) have been defeated in cities such as New York and Philadelphia thus far, while food labeling regarding sugar and fat content has improved.

The Potential Impact of Inclusion of Obesity as an Eating Disorder

Along with the scientific arguments, we must also consider the psychological impact of designating obesity as an ED. The benefits of designating a condition a “psychiatric disorder” should outweigh the cons. Obesity already carries a significant burden of stigma in our society, with some researchers suggesting that “obese persons are the last acceptable targets of discrimination” (Puhl & Brownell, 2001, p. 788). Many advocates for those with mental illnesses, including EDs, have battled back against stigma. AN and obesity are similar in that they are “visible” disorders, as opposed to BN or, say, generalized anxiety disorder. Persons suffering from the former disorders are much more likely to be stereotyped or stigmatized without any interpersonal interactions even occurring. The benefits of respect and of access to care and effective treatment should supersede the increase, or at least solidification, of stigma that would likely occur with labeling obesity as a mental disorder.

In the case of AN, the benefits have seemingly outweighed the cons, as weight restoration is only part of the picture, and, usually, longer‐term psychiatric intervention is needed for successful treatment. In the case of obesity, as a general category, it seems unlikely that designation as a mental disorder would improve our approaches to care. Weight management requires long‐term attention and care, much like caring for diabetes, but such long‐term management does not necessarily have to take the form of psychiatric treatment. Treating those with BED, NES, or possibly “food addiction” (as the phenomenon is studied further), yields benefits, but as discussed above, lumping all persons with a BMI above 30 kg/m2 into a single disorder would result in a widely heterogeneous category. Such a grouping would not likely improve care, and depending on the wording of the diagnostic criteria, may capture persons who were not distressed or limited in their functioning in any significant way. This remains a significant detraction for including obesity as an ED.

Conclusions and Future Directions

DSM categories are not phenotypes. Persons are complicated and eat for a myriad of reasons, particularly in cultures where food is abundant. Persons also gain weight for many reasons, including genetic predisposition, one’s food environment, access to physical activity, cultural food preferences, access to low‐calorie foods, external stressors, and emotional dysregulation, to name a few. With a shift in directives for mental health research to examine psychiatric phenomena along continua, as opposed to categories with checklists of symptoms, research will continue to examine the underlying genetic, physiological, behavioral, and psychosocial data for symptoms that cut across categories—for example, substance and sugar addiction or general overeating and emotion regulation. Advancement of brain imaging and genetics studies could be fruitful as our approaches become cheaper and more widely accessible, but it will be important to replicate findings both in identifying patterns and traits within those who are obese, and across symptom groupings among those who suffer with various addictions or EDs. Better understanding of the experience of extreme food cravings and identification of “food withdrawal” would also be helpful from a more descriptive or behavioral perspective to define the food addiction phenotype. We look forward to such data to generate the most fruitful conceptualization of obesity so as to promote optimal physical and mental health.

By Kelly C. Allison (1) and Alysia A. Cirona‐Singh (2) in "The Wiley Handbook of Eating Disorders", First Edition. Edited by Linda Smolak and Michael P. Levine, John Wiley & Sons, Ltd. USA,  2015, excerpts pp. 901-912. Edited and illustrated to be posted by Leopoldo Costa.

(1) Department of Psychiatry, University of Pennsylvania, USA
(2) Department of Psychiatry, San Mateo Medical Center, USA