Eating is of obvious and critical importance to the development, health, and, indeed, the survival of children. Eating is among the very earliest of the behaviors that make up the behavioral repertoire of infants and develops rapidly as babies grow into children and then adolescents. Problems with the consumption and retention of food are a source of acute concern for parents and can pose serious risk for the child. Indeed, serious eating disorders are life-threatening and among the most fatal of all mental health problems (Arcelus et al., 2011; Franko et al., 2013; Smink et al., 2014; van Son et al., 2010).

Eating problems have been recognized for centuries, with descriptions of restricted eating, binge behavior, and the purging of consumed food through use of vomiting and laxatives dating back to at least the Middle Ages (Bell, 1985). Numerous descriptions of medieval saints, for example, portray either self-starvation or bingeing episodes (or both). Similar descriptions can be found consistently in historical documents from the intervening centuries, such as those described by John Reynolds and Richard Morton in the 17th century and those of Marcé, Gull, and Laséque in the 19th century. Over the years, different terms were used to describe these behavioral patterns, reflecting the understanding and values of the time. In the Middle Ages, the term anorexia mirabilis reflected the association of self-starvation with spiritual purity and other penitential practices; in the 17th century, the term nervous atrophy described a condition with psychological roots and a “morbid state of the spirits”; in the 19th century, the term hypochondriacal delusion was used to describe a psychological condition characterized by food refusal. During the first half of the 20th century, mental health problems were being more rigorously studied as medical conditions. In this period, the two prevailing theories were endocrinologic (centered around dysfunction of the pituitary) and psychoanalytic (centered around the defense against unconscious wishes for impregnation) (Eissler, 1943; Masserman, 1941; Meyer & Weinroth, 1957).

The latter half of the 20th century saw the introduction of formal systems for the classification of psychological disorders and the establishment of agreed-upon criteria for their diagnosis. The first iteration of the DSM in the 1950s (American Psychiatric Association, 1952) recognized anorexia nervosa as a “neurotic illness,” and by the second version of DSM, pica and rumination were also recognized, with bulimia nervosa added in the next version (American Psychiatric Association, 1968). The next major revision to DSM added the
broad category of Eating Disorder Not Otherwise Specified to capture all eating disorders apart from anorexia and bulimia. Finally, the latest iteration of the DSM, DSM-5, recognizes several eating disorders grouped together under the category of feeding and eating disorders (American Psychiatric Association, 2013). These include, in addition to anorexia and bulimia, binge eating disorder (in acknowledgment of data indicating that binge behavior can occur in the absence of weight loss behaviors), pica, rumination disorder, and avoidant/restrictive food intake disorder (ARFID). DSM-5 also maintains the diagnostic classifications for other specified eating or feeding disorder and unspecified eating or feeding disorder.

Table 17.1 summarizes the key diagnostic features of each of the eating or feeding disorders classified under DSM-5.

The three criteria for establishing a diagnosis of anorexia nervosa relate to restricted energy intake leading to low body weight, intense fear of gaining weight or becoming fat, and the perception of one’s body shape or weight.

The first criterion, of restricted energy intake leading to low body weight, is not very specific and how to define “significantly low weight” remains a question. The DSM states that weight should be less than minimally normal or expected, but there is no consensus on how to calculate weight loss or on the boundaries of minimally normal. One strategy relies on the body mass index (BMI), which is calculated as one’s weight in kilograms divided by one’s height in meters squared. The World Health Organization views a BMI of 18.5 kg/m² as the lower threshold for normal body weight (in adults), and this definition does not vary for different locations or cultures. In the United States, the Centers for Disease Control and Prevention views a BMI below the 5th percentile, using their calculator, as underweight.

The second criterion, of intense fear of becoming fat or gaining weight, or behaviors that interfere with weight gain despite low actual weight, is more straightforward but relies heavily on patients’ self-disclosure. This poses a challenge as it is not uncommon for patients with anorexia to be less than forthcoming about their symptoms. Likewise, the third criterion of distorted perception of body weight or shape, or undue influence of body weight or shape on self-evaluation, or lack of recognition of the current unhealthy body weight, relies on subjective information provided by the patient. Distortions in the perception of body weight and shape in patients with anorexia can vary a great deal. In some cases, the distortion can relate to the entire body, with a sense of simply being fat overall, whereas in other cases it may be more specific to particular parts of the body such as the arms, thighs, or abdomen (Box 17.1).

Anorexia nervosa is further divided into two subtypes, restricting and binge-purge. Certain behavioral patterns, including impulsive behaviors, suicidal behavior, and substance abuse, are more commonly associated with the binge-purge subtype of anorexia, as are a history of obesity and some medical complications.

The essential criteria for establishing a diagnosis of bulimia are the recurring presence of episodes of binge eating, compensatory behaviors aimed at preventing weight gain, and a self-image that is unduly influenced by body shape and weight. Binge eating episodes are discrete periods of time (DSM suggests 2 hours as an example) during which an unusually
large amount of food is consumed, along with a lack of control over the eating. The loss of control is critical to the diagnosis and must be subjectively endorsed by the patient for a diagnosis to be established.

The most common form of compensatory behavior aimed at preventing weight gain is self-induced vomiting. However, other compensatory behaviors can occur instead of or in addition to vomiting, such as fasting or the use of laxatives (Box 17.2).

Binge eating disorder is defined by the presence of binge eating episodes (as in bulimia nervosa) and distress relating to the binge eating. DSM also includes some more specific descriptive criteria surrounding the binge eating, such as eating rapidly or until uncomfortably full, eating alone because of embarrassment about the quantity of food, and feeling disgusted with oneself or guilty about the binge eating.

Importantly, concerns relating to weight loss or appearance do not make up part of the diagnosis of binge eating disorder, and this is the most important distinction between binge eating disorder and bulimia nervosa. Extensive research demonstrating that the pattern of binge eating without such weight and shape concerns does occur supported the addition of binge eating disorder to the latest iteration of the DSM.

Pica refers to the consumption of nonfood, nonnutritive substances. The essential criteria for diagnosing pica per DSM-5 are the consumption of nonfoods for at least a month in a manner that is abnormal for the individual’s age, development, and cultural context. When other physical or mental health problems are present, for example, during pregnancy or in the context of autism spectrum disorder, pica is only diagnosed when the consumption of the nonfoods represents a hazard to the physical health of the individual. Individuals who eat
nonnutritive substances that can be considered “food,” such as ice or artificial sweeteners, are not diagnosed with pica. Among the substances consumed by individuals with pica are chalk, powdery substances such as talcum powder or makeup, and feces.

The age of the patient must be considered, as it is not uncommon for very young children to suck on or swallow nonfoods, and thus pica is generally not diagnosed before 2 years of age. Likewise, cultural context must be taken into consideration. Some cultures and religious practices include the consumption of nonfoods for spiritual or medicinal purposes; when the behavior matches the cultural context, pica is not diagnosed. Pica is also not diagnosed when the eating behavior is intended to serve a specific purpose such as appetite suppression or as a means of self-mutilation, or when it is the manifestation of a psychotic delusion.

Pica is often associated with intellectual and developmental disabilities and may also occur in obsessive-compulsive spectrum disorders, such as the consumption of pulled hair in trichotillomania. The clear concern in individuals with pica is that the nonfoods will cause physical harm, injuries, disease, or even death.

Rumination disorder is diagnosed based on frequent regurgitation of swallowed food, meaning that previously swallowed food is brought back up into the mouth. Regurgitated food may be spat out or swallowed again. The regurgitation must be frequent, occurring several times per week, and often daily or even consistently at almost every meal. Regurgitation is not the result of nausea and is distinct from vomiting, in which retching occurs and regurgitated food is acidic. The behavior must also be distinct from compensatory weight loss behaviors, as in anorexia or bulimia nervosa. Individuals with rumination disorder typically describe the behavior as habitual or uncontrollable, rather than as a completely voluntary action (Box 17.3).

Regurgitation is most common in individuals with intellectual disability and can occur at any age, starting in infancy. The persistent regurgitation can lead to medical complications, such as malnutrition, and to social impairment. Individuals with rumination disorder may come to avoid eating in the presence of others because of social discomfort associated with the regurgitation.

ARFID is a newly established diagnosis under DSM-5 and replaces a subset of diagnoses previously categorized under Feeding or Eating Disorder of Infancy or Early Childhood in the previous version of the DSM. The essential criterion for establishing a diagnosis of ARFID is the failure to meet appropriate nutritional or energy needs because of lack of interest in eating, avoidance of foods based on their sensory characteristics, or concern about negative consequences of eating. The disturbed eating pattern must result in significant weight loss or low weight, nutritional deficiency, dependence on enteral eating or supplements, and/or marked interference in psychosocial functioning. An emerging consensus in the field is that ARFID diagnoses should not be limited only to children with nutritional deficits or low weight, and that marked psychosocial impairment, stemming from the restricted eating, is sufficient for the diagnosis to be established. In contrast to the previous characterization of this problem as a disorder of infancy or early childhood, ARFID, per DSM-5, may set in and be diagnosed later in development.

ARFID cannot be diagnosed if present only during episodes of anorexia or bulimia, and individuals with ARFID do not present with distorted perceptions of their body weight or shape and do not restrict their eating as a weight loss strategy.

Because sensory sensitivity and picky eating are common phenomena, in particular in young children, it is important that the ARFID be abnormal and cause physical or psychosocial impairment for a diagnosis to be established.

As described in the main diagnostic criterion, three not mutually exclusive subtypes of ARFID have been recognized based on the functional motivation of the avoidant or restricted eating. The first subtype describes individuals with low interest in food and eating. This pattern tends to emerge early in life, during infancy, but can persist throughout development. The second subtype relates to restricted eating based on the sensory characteristics of foods. For example, a child may avoid foods that feel too hard, too soft, or too wet, or they may avoid foods of certain colors or shapes. This subtype tends to emerge during the first decade of life, but also can persist into later adolescence and adulthood. The third subtype describes avoidance or restriction of food because of fears of negative consequences of eating. Fear of choking is a common example of this kind of fear. This subtype is less closely linked to any particular development epoch and can arise at any age. In some cases, a frightening or traumatic event (such as choking) can trigger the fear and lead to ARFID (Box 17.4).

Estimating the prevalence of eating disorders is challenging, as much of the research in this area has focused on specific populations rather than on population-based sampling (Halmi, 2018). In the United States, the lifetime prevalence of anorexia nervosa in youth aged 13-18 was estimated at 0.3% (Swanson et al., 2011), whereas a study of Dutch youth found a lifetime prevalence rate of 1.7% for females aged 11-19 and 0.1% for males (Smink et al., 2014). For bulimia nervosa, a lifetime prevalence rate of 1.6% was reported for females up to age 20 in the United States (Stice et al., 2009), and a point prevalence rate of 0.6% was reported in Australian 14-year-olds (Allen et al., 2009). The challenges in estimating prevalence rates are compounded by the changing diagnostic criteria with successive iterations of the DSM. It is safe to say that anorexia nervosa and bulimia nervosa, although very serious disorders, are not exceedingly common, likely affecting one in several thousand individuals at any given time, and females more often than males.

The lifetime prevalence of binge eating disorder in the United States is estimated at 3.5% for women and 2% for men (Hudson et al., 2007, 2012). However, these estimates are across the entire lifespan and not for youth alone. A representative sample of American adolescents found lifetime prevalence rates of 2.3% for females and 0.8% for males (Swanson et al., 2011).

Pica appears to be more common than other eating disorders, but actual prevalence rates are not really known and studies report widely varying prevalence rates, based in part on different inclusion and exclusion criteria. For example, studies that include the eating of ice as a nonnutritive substance report higher prevalence rates than studies that exclude ice eating as a symptom of pica. What is clear is that pica is more commonly reported in samples of individuals with intellectual disabilities, in particular those who require residential and institutional care. In such samples, prevalence rates for pica have ranged from 10% to 25% (Ali, 2001; Coniglio & Thomas, 2018). Prevalence rates for rumination disorder are also not known for the general population. One study in Swiss elementary schoolchildren found that approximately 4% of children endorsed clinically elevated rumination symptoms (Murray et al., 2018). However, the study relied on self-report questionnaires rather than on formal diagnostic evaluation. Another study of youth in Sri Lanka reported a prevalence rate of 5.1% for adolescents (Rajindrajith et al., 2012).

ARFID is the newest eating disorder included in DSM, and as such, studies estimating its prevalence using the most current diagnostic criteria remain scarce. Furthermore, some
degree of picky eating is normative over the course of development and should not be viewed as pathologic. One study found a prevalence rate of 1.5% in a sample of pediatric patients (Eddy et al., 2015), whereas a community study reported a prevalence rate of 3.2% (Kurz et al., 2015).