Additional related information may be found at:
Neuropsychopharmacology: The Fifth Generation of Progress

Back to Psychopharmacology - The Fourth Generation of Progress

Eating Disturbances and Eating Disorders in Childhood

Regina C. Casper


Whereas eating disturbances in children are fairly common (40, 93), the classical eating disorders are observed infrequently (1, 40, 62). Nonetheless, cases of prepubertal anorexia nervosa as early as age seven have been reliably documented for over a century (25, 62, 83). The existence of bulimia nervosa as a syndrome in children is not well documented except for two possible retrospectively determined cases reported by Kent et al. (52). Overeating episodes, on the other hand, have been described in male and female children in the course of anorexia nervosa, albeit rarely (36, 48, 52, 92).

The Diagnostic and Statistical Manual, fourth edition (4), no longer classifies eating disorders under " Disorders usually first evident in infancy, childhood or adolescence " as in DSM III-R (3). Eating disorders occupy now a separate section under disorders in adulthood. This reclassification is unfortunate, since it ignores that anorexia nervosa does occur in childhood, typically has its onset during adolescence and is intricately related to growth and development. Feeding and eating disorders of infancy or early childhood have remained in the childhood section and comprise three syndromes: pica, rumination disorder, and feeding disorder of infancy or early childhood.

Recent retrospective analyses of hospital records have noted an increase in childhood anorexia nervosa (51). Whether this finding is due to a true increase in incidence or reflects previous underdiagnosis of childhood anorexia nervosa will require more study. In point of fact, Bryant-Waugh and Lask (13) have shown that few medical practitioners in England are familiar with anorexia nervosa in childhood. A mere 31% among pediatricians and only 3% of family practitioners in a geographical area mentioned a possible diagnosis of anorexia nervosa when they were asked to evaluate two case vignettes of childhood anorexia nervosa.

Another factor which could contribute to underdiagnosis is the lack of consensus regarding the diagnostic criteria. Irwin (48) considered the DSM-III diagnostic criteria (2) as too restrictive and insufficiently specific for diagnosing children. On the assumption that excessive fear of becoming obese in the presence of severe underweight was a universally accepted core symptom, Irwin (48) argued that the amount of body weight loss required for the diagnosis was too high, because the smaller percentage of total body fat in children resulted in greater emaciation with less weight loss than in the postpubertal female. These concerns lead to a revision of the weight criteria in DSM-III-R (3) with a reduction of the weight loss requirements for anorexia nervosa to 15% instead of 25%. Fosson et al. (35) emphasize another symptom typically observed in childhood, deccelerated growth (10-14 years) in the absence of any physical or mental illness. Furthermore, the criteria requiring amenorrhea in females or impotence for males do not apply to children.

This chapter on feeding and eating disorders of infancy and early childhood and the principal eating disorder recorded in childhood, anorexia nervosa, presents the clinical evidence which has provided the data for diagnostic classifications, since the precise etiology of these disorders remains unknown. Information regarding children's attitudes towards body shape and food as well as those developmental and environmental factors in childhood that might increase the risk for eating disturbances or an eating disorder will be presented first. Finally, treatment approaches and directions for future work will be discussed.

Food refusal, weight concerns and fear of fatness in children

Children refuse food for all kinds of reasons, aversion to the sight, taste, or form of food, little appetite, or distractions with more interesting activities, such as play. However, this kind of food refusal seldom lasts long.

Research has shown that infants and children from a very early age on can control their own food intake and grow normally, given proper guidance and emotional support. Clara Davis (29) demonstrated that newly weaned infants, 6 to 8 months old, when offered natural food materials, whole grain, fruit, vegetables, eggs, fish, and meat, could feed themselves and grow normally. Birch et al. (8) more recently reported similar observations for children from 2 to 5 years of age. The children's intake was found to be highly variable at individual meals, but the total daily energy intake was relatively constant for each child, because the children tended to adjust their energy intake at successive meals.

In the western world, young children tend to absorb the prevailing cultural values about food and body shape. Worsley et al. (94) reported that 10 year old Australian boys and girls noticed first the fattening, then the healthy characteristics of food and only lastly expressed sensory preferences. In a study of British school girls age 12 to 20 years, Crisp et al. (27) noted that 26% of premenarchal girls compared to 48% in postmenarchal girls were concerned about fatness. Pugliese et al. (70) described "fear of fatness" in children as a reason for stunted growth. Out of 201 children evaluated for short stature or delayed puberty or both at the Department of Pediatrics at Cornell Medical College, 14 children (9 boys, 5 girls) aged 9 to 17, all coming from middle-class families, showed growth failure due to malnutrition as a result of self-imposed restriction of caloric intake arising out of fear of becoming obese. The youngsters resumed normal growth when they were counselled and given an age-appropriate diet.

Epidemiological Studies On Dieting, Weight Concerns and Body Image

Studies on dieting in adolescence have consistently shown a high prevalence of dieting in the Western Hemisphere, about 60% to 80% since the early 70's (66, 75, 80). A recent study (60) of upper middle-class white children found that over 40% wanted to be thinner and about 30% had attempted to lose weight, mostly through increasing their activity level. Newspaper reports have conveyed the impression that children are paying more attention to their weight and that more might be dieting. For example, the Wall Street Journal reported that 75% of fourth grade girls complained that they weighed too much (95), while Newsweek quoted a study that half of fourth grade girls in a middle-class school in San Francisco described themselves as overweight, although only 15% were actually overweight. Among the 9 year olds, 31% thought themselves too fat and almost half were on a diet (67). Richards et al. (73) surveyed nearly 500 students from 5th to 9th grade from middle and working class families in Chicago schools. Between 5th and 7th grade, a mere 15% among the girls expressed extreme weight concerns or dieting attempts, but this proportion increased to 32% by the 8th and 9th grade. The opposite trend was observed for boys. Eight to twelve percent expressed weight and eating concerns between the 5th and the 7th grade, and only 3% by the 9th grade.

Few studies of body image perception and satisfaction before the onset of puberty have been conducted. Most have chosen the onset of menarche as a critical variable for research purposes, because of its close association with hormonal variables, self-image, and peer and parental relationships during adolescence (11). Koff et al. (53) found that premenarchal girls indicated less satisfaction with body parts than postmenarchal girls, while Gargiulo et al. (37) found no relationship between menarchal status and body satisfaction. In a comparison of premenarchal and postmenarchal girls on perceptual and subjective measures of body image, few differences were found; premenarchal girls overestimated their thighs to a greater degree than postmenarchal girls (34). Interestingly, in both groups, a history of having been teased was positively associated with body dissatisfaction. Studies on the association between the timing of onset of menarche and body image have suggested that girls who experience menarche after the age of 14 have a more positive body image than those who have their first menstrual period before the age of eleven (11). Contributing to the negative effects of menarche for girls may be the observation that early maturers are generally shorter than late maturers, but usually weigh more (38).

Staffieri (84) reported that children early on incorporate certain beliefs and expectations in relation to body configuration. Six to ten year old boys considered the mesomorph (muscular) image most favorable. The thin (ectomorph) figure was viewed as quiet, weak and fearful, whereas the endomorph (overweight body type) was seen as combattive, lazy and cheating. A recent survey of 36 public schools grades 4 to 10, commissioned by the American Association of University Women (28), reported a pronounced loss of self-confidence in white girls from pre-puberty to middle adolescence, but no change for boys. Whereas most 9 year old girls were happy with themselves, only 29% still felt that way by high school. For boys, the proportion dropped from 67% to 46%, leaving more boys with sound self-esteem than girls. Interestingly, African-American girls retained their self-confidence into high school.

The Relationship Between Weight/Height, Dieting Concerns, and Psychological Well-Being

Negative emotions have been shown to be correlated with body shape dissatisfaction (66, 80). Richards et al. (73) reported positive correlations between body image dissatisfaction and depressive symptoms in fifth to sixth grade girls and boys and in girls only, negative correlations between body image satisfaction and weight and eating concerns. Fifth and sixth grade boys, who reported more dysphoric affect and lower daily levels of energy and arousal, expressed greater weight and eating concerns. Recently, family conflict alone has been shown to be associated with slow growth to age 7 years (64).

Feeding and Eating Disorders of Infancy and Early Childhood

The specific disorders included in the DSM-IV (4) are feeding disorders of infancy and early childhood, pica and rumination disorder.

Feeding Disorder of Infancy and Early Childhood

The term feeding disorder seeks to recognize the dyadic nature of the process in which both mother and infant are active participants. Although the distinction between organic and non-organic eating problems of infancy and early childhood is conveniently accepted, the distinction is not always clear. Given current diagnostic criteria, the disorder must have occurred before age 6 years and be persistent, and have lasted at least one month.

Medical problems, for example gastroesophageal reflux, can disrupt the mother-child interaction during feeding, and early childhood feeding and eating disorders may lead to developmental delay and growth retardation. Douglas and Byron (32) found that parents identified prematurity, low birth weight and feeding distress in the first six months of life as precursors to severe eating difficulties in children younger than 7 years.

Failure to thrive (FTT) is a diagnostic term used by pediatricians for infants who fail to make expected and age-appropriate gains in weight. Failure to thrive has been reported in 1% to 5% of pediatric hospital admissions; organic factors seem to account for less than 10% of the cases (81). The evaluation of non-organic FTT nowadays is no longer based on weight and height alone, but seeks to determine the infant's developmental level and the quality of the infant-mother or caretaker interaction. Malnutrition, socioeconomic factors, and emotional deprivation have been implicated as etiological in non-organic FTT. Nevertheless, many infants grow normally despite feeding difficulties, while other infants or children show delayed development in the presence of normal weight gain. Neither FTT nor psychosocial dwarfism, a syndrome of decelerated linear growth, have been included in the DSM IV. In view of the controversy surrounding the definition and heterogeneity of FTT, a developmental classification of feeding disorders based on observations of the mother-infant interaction outlined by Chatoor et al. (23) will be briefly described here.

Disorders of Homeostasis

Disorders of homeostasis are said to occur during the time from birth to 2 months, when basic functional cycles and rhythms of sleep and wakefulness, nursing and elimination are established. Temperamental differences, a high level of irritability and oversensitivity to stimulation can give rise to the so-called colicky infant, in others the integration of the sucking and breathing response may be delayed. In such cases a mother's inability or inflexibility to respond to the child's particular needs may disrupt the infant's attempt to establish a regular feeding and, often, sleeping pattern.

Disorders of Attachment

Between 2 and 6 months, the infant's capacity to emotionally engage the mother determines the feeding experience. Regulation of food intake is closely linked to the bodily and visual cues guiding the interaction between mother and infant. Mothers who remain detached or disinterested due to depression, character disorders, or because of deprivation in their own childhood may have children who are withdrawn, listless during feedings or who show vomiting, diarrhea and poor weight gain.

Disorders of Autonomy and Separation

The physical and cognitive maturation that occurs during the age of 6 months to 3 years has led Mahler (59) to characterize this period as the first separation-individuation phase. Chatoor et al. (23) have called the disturbed eating during this time a separation disorder. If the mother cannot respond to the expressions of the infant's beginning autonomy, the infant may assert his will by refusing to eat through rejecting food.

Post-Traumatic Eating Disturbances

Bernal (7) and Chatoor et al. (24) have described food refusal following an incident of choking. Eating problems due to trauma bear similarities to post-traumatic stress disorder and to phobia, since children demonstrate acute anxiety about food ingestion and consequently refuse to eat.


Pica derives from the Latin term for the "magpie," a bird noted for its habit of carrying away inedible objects. Pica denotes the persistent eating of any nonnutritive substances including earth eating or geophagia. Parry-Jones and Parry-Jones (68) recently published an excellent and exhaustive overview of the phenomenon of pica throughout history. During past centuries pica has been regarded as a symptom, a sign of a morbid appetite raising the question whether turning pica into a diagnosis is of any value. For instance, geophagia (45) is a world-wide practice. Nosologically pica captured attention when it was recognized that young children, especially black children in inner cities, were ingesting paint chips in older houses painted with lead base paint, thereby exposing themselves to toxic lead levels.

Three criteria are necessary to diagnose pica in DSM-IV (4): a) persistent eating of non-nutritive substances for at least one month, b) the eating of non-nutritive substances is inappropriate to the developmental level, and c) the eating behavior is not part of a culturally sanctioned practice. There is no agreement among researchers about the etiology of pica and its appropriate treatment. When children experience symptoms of abdominal pain and anemia or failure to thrive, pica should be suspected. Undoubtedly, pica frequently has become a habit, and thus treatment efforts need to be directed towards removing non-nutritive substances that can be eaten and are toxic from the house and towards feeding children properly and taking care of their emotional needs.

Rumination Disorder in Childhood

This rare disorder, which occurs primarily during the first year of life is defined by repeated regurgitation and rechewing of food (in the absence of associated gastrointestinal illness) for a period of at least one month following a period of normal eating. Partially digested food is brought up into the mouth, rechewed and reswallowed and less often ejected.

Because the disorder is rare, no epidemiological studies exist. A study by Mayes et al. (63) suggests a slight male predominance for rumination disorder. The differential diagnosis of rumination disorder includes pyloric stenosis and other congenital gastrointestinal abnormalities, as well as normal vomiting of early infancy. The typical characteristic preparatory movements interpreted as voluntary movements can help distinguish the disorder from ordinary vomiting. Gastroesophageal reflux is associated with high rates of rumination disorder in infancy. The most common type is psychogenic rumination. Rumination disorder is typically not associated with abnormalities of cognitive development, although a second type called self-stimulatory rumination occurs in mental retardation.

Eating Disorders in Childhood

1. Selective Eating

Lask and Bryant-Waugh (56) describe selective eating as a condition in which children eat a narrow range of foods, mostly carbohydrates. Even though the diet appears to lack essential nutrients, the children’s growth seems adequate. Nevertheless with the help of the pediatrician, efforts to improve the range and nutritional value of foods are often successful providing considerable relief for the worrying parents. Some of these children tend to have problems in social functioning. On the other hand, parents sometimes restrict their children's food intake. For example, Russell et al. (78) have drawn attention to mothers with anorexia nervosa who underfeed their children.

2. Food Avoidance Emotional Disorder

This condition described by Higgs et al. (46) might be a precursor to anorexia nervosa. The food avoidance seems to be the most conspicuous symptom of other anxiety-based behaviors, such as school avoidance, other phobias, obsessional symptoms and depression.

3. Anorexia Nervosa in Childhood

Our knowledge about anorexia nervosa in childhood is incomplete, in part because of its low incidence. Table 1 lists published case series reliably diagnosed from information in hospital records. Disagreement over terminology creates a diagnostic problem. Some authors describe premenarchal anorexia nervosa (43) which is not synonymous with prepubertal anorexia nervosa because pubertal changes in girls precede the onset of menarche by about 3 years, whereas other investigators tend to include pubescent cases. Use of the Tanner (90) stages for standardizing sexual development would provide the most reliable distinction between childhood and adolescence, especially since the incidence of anorexia nervosa rises dramatically with puberty.

D1. Incidence and Age at Onset

To my knowledge, no epidemiological studies on the incidence and prevalence of anorexia nervosa in childhood exist. When the few retrospective case surveys about prepubertal anorexia nervosa are pooled (49), between 4% to 8% of all anorexia nervosa cases seem to have had an onset in childhood. Considering the estimated incidence of anorexia nervosa, .15 new cases in 100,000 per year (89), childhood anorexia nervosa is indeed unusual. Since retrospective surveys have relied on psychiatric case records, the inclusion of milder cases that are seen by pediatricians would increase this estimate. In childhood, the restricting or fasting type of anorexia nervosa (16) prevails, but the phenomenology appears to differ across cultures. A Russian publication reported seventeen 9 to 11 year old girls who regularly engaged in vomiting with the purpose of slimming down and improving dysphoric mood after eating. All girls suffered stunted growth as a result of their caloric restriction, but none became cachectic. Remarkably, with treatment all girls recovered without relapse (54).

Sex Ratio

As Table 2 indicates the proportion of boys seems considerably higher for childhood anorexia nervosa, on average about 26% to 28%, than the percentage of boys, between 4%-6%, who present with postpubertal anorexia nervosa.

Precipitating Events:

a. Psychological Precursors

Disruptive life events precede more often the onset of anorexia nervosa in childhood than in adolescence (46). Such events might be the birth of a sibling, moving houses, losing a friend, death in the family, family quarrel, parental divorce, or disappointment in a friendship. All cause anxiety or depressive symptoms and undermine the child's faith in his or her control and sense of security and can lead to a reduction in appetite. Once the parents or the pediatrician become alarmed, the child may discover, not always consciously, the controlling power in the refusal to eat (76) and thus precipitate weight loss and through it, if a predisposition exists, anorexia nervosa.

b. Physical Precursors and Eating Problems

Food fads and picky eating in early childhood have been associated with anorexia nervosa (51, 76), whereas pica and family conflicts during meals in early childhood seem to be precursors of bulimia nervosa (61). Early physical maturation may enhance bodily awareness and body size (17) and lead to conscious or unconscious fears of uncontrollable overweight resulting in food refusal and thus enhance the risk for anorexia nervosa during early puberty.

c. The Vulnerable Child Syndrome

A minority of children who develop anorexia nervosa tend to be sickly or frail or have experienced early physical threats or perinatal trauma (6,32).

D4. Symptomatology

The symptoms of childhood and postpubertal anorexia nervosa are remarkably similar: food avoidance or refusal to eat out of fear of fatness, based on a deliberate personal decision, although children sometimes convincingly state that eating is 'bad'. The voluntary food restriction in children often extends to life-threatening fluid restriction. Typically children fail to gain weight commensurate with their previous growth rate rather than lose weight. However, their smaller body proportions make weight loss more noticeable early on. A second diagnostic sign is indifference to or denial that lack of expected weight gain is of any importance which suggests partial unawareness of the bodily changes. Vague abdominal pain and avoidance of high caloric food or idiosyncratic ideas about food, such as avoiding eggs, because of their connotations with fertility, are not uncommon in children. Failure to grow more often alarms children and the danger of growth arrest will motivate some children, unfortunately not all, to eat more. By and large children tend to be less fixated on thinness per se. Overactive behavior can be observed in pre-pubertal, as in post-pubertal anorexia nervosa cases, in contrast to prepubertal neurotic children (15,35, 51). The most appropriate way to study the effects of food deprivation in children is to fit growth curves to the weight and height data for each individual child usually available from the child's pediatrician. This curve then serves as the individual's own control and basis for predicted growth and for the calculations of deceleration of expected growth in height and size (50). The Eating Disorder Examination (EDE) a standardized instrument for the assessment of the specific psychopathology of eating disorders has been modified by Bryant-Waugh et al. (15 ) and can now be used to evaluate children.

Personality Dimensions

The clinical description of the personality of children with anorexia nervosa, shyness, including shyness, lack of pleasure, compliance, rigidity and perfectionism (16, 20, 36) is consistent with personality traits of adolescents with the restricting type of anorexia nervosa (20). Casper et al. (20) have suggested that rigidity, overcontrol and avoidance as personality tendencies that tend to facilitate food abstention may place youngsters at risk for developing the restricting type of anorexia nervosa. Similarly Gillberg and Rastam (39) have described in early onset anorexia nervosa patients autistic-like social impairment.


Despite the presence of sadness and withdrawal, many children appear euphoric (22) and many do not fit any diagnostic category . Jacobs and Isaacs (51) reported disturbed peer relations in two-thirds of prepubertal anorectic children as opposed to one-third of neurotic children. Fosson et al. (35) found obsessions and compulsions in roughly a third of his sample. In a prospective study, Atkins and Silber (6) applied DSM-III-R criteria (3) to 21 children with anorexia nervosa, who were diagnosed at the age of 12 or under. Nine were prepubertal, while pubertal development had begun in the remaining ones. Comorbidity was reported for the group without specifying pubertal stages. Six patients had no other psychiatric diagnosis, four had signs of a depressive disorder, three each were considered either a narcissistic personality disorder or an overanxious disorder, while two each were diagnosed with obsessive-compulsive or oppositional disorder, and one with a borderline personality disorder.

Family Functioning

Roughly half the children are said to have families with manifest conflict, discordant intrafamiliar relationships, or parental overinvolvement. Birth, death, illness, and sometimes alcoholism create family crises (74). Jacobs and Isaacs (51) reported more frequent food fads and familial overinvolvement in prepubertal anorectic families compared to neurotic controls, but no differences to postpubertal anorectic families were observed. Fosson et al. (35), whose report includes early adolescents up to age 15 years, reported family dysfunction in the majority of cases.

Developmental Arrest: the Toll of Childhood Anorexia Nervosa Outcome Studies

The long term outcome of "childhood" anorexia nervosa is difficult to evaluate, since virtually all follow-up studies have combined childhood cases with early adolescent cases for the outcome analyses. Bryant-Waugh et al. (12) found based on a 7 year follow-up that a young age and age at referral of less than 11 years carried a poor prognosis regarding body weight and overall outcome. Bryant- Waugh et al. (14) have published the only prospective short term, 3 year, follow-up on 18/22 patients (16 females and 6 males) attending a specialty clinic, whose mean age at onset was 12.1 years. Only 4 females and 2 males were prepubertal at the time of the referral. The outcome was good in 55%, intermediate in 27.8 % and poor in 16.7 %. In a study of a mixed child/adolescent group, Goetz et al. (41) reported normal weight and good psychological adjustment in 26 of 30 patients whose mean age at diagnosis was 9.5 to 16 years, and who were assessed five to twenty years after hospitalization. Hawley (44), Steinhausen and Glanville (85) and Saccomani et al. (79) observed a good nutritional outcome in 53% to 67% and good psychological adjustment for about 50% patients followed up from 3-15 years. By contrast, only a third of Higg's (46) patients evaluated 5 years after hospitalization qualified for a good outcome. Overall the studies suggest that early onset has about as variable an outcome as adolescent onset (8,86).

The controversy whether prepubertal anorexia nervosa is a milder disorder than anorexia nervosa with pubertal or postpubertal onset (44,58,85,88) cannot be decided, as long as no sizable long term outcome studies exist and as long as milder cases of children treated in pediatric practices and outpatient clinics are not included in the follow-up.

Endocrine and Metabolic Changes

The endocrine changes associated with anorexia nervosa in childhood have not been studied as far as we know. On the whole, the starvation-induced endocrine and physical changes in children can be expected to be similar to those observed in adolescents (21). Support for this claim is provided by the fact that with profound weight loss the hypothalamic-pituitary-gonadal axis invariably regresses even in mature women to prepubertal functioning. However, it is not known whether the activation of the hypothalamic-pituitary-adrenocortical axis is as pronounced in children as it is in adults (10). In children the cardiac changes, especially nocturnal bradycardia, that may require cardiac monitoring, can be more life threatening than in adolescents. Sleep studies have not been published in childhood anorexia nervosa, but children with poor growth as a result of psychosocial deprivation show significant decreases in the percentage of stage IV sleep and an increased amount of rapid eye movement sleep (82,91). In the search for organic etiological factors which would be instrumental in the abnormal neuropsychological manifestations, such as starvation induced activation (22), functional brain imaging studies have been performed without revealing precise deficits. Non-specific changes suggesting cerebral pseudoatrophy in severely malnourished patients with anorexia nervosa have been found to be reversible with weight gain (31,65). In the first such study examining regional cerebral blood flow (rCBF) in children and young adolescents with anorexia nervosa, Gordon et al. (42) reported reduced rCBF in temporal lobes which persisted in 3 children, who were retested after weight gain which points to asymmetrical temporal hypoperfusion as a possible primary deficit.

Effects on Physical Maturation and on Linear Growth

Russell (77) described delay of puberty, delayed breast development, and interference with growth and height in chronically ill anorexia nervosa patients with premenarchal onset of illness.

Caloric deficiency alone can lead to slowing or arrest of physical growth (38). Dreizen et al. (31) studied skeletal maturity through serial roentgenograms of the left hand and wrist every six months into adulthood in thirty undernourished girls between the ages of 4 to 5 years in comparison to well nourished girls. Sustained nutritional deprivation slowed the rate of bone growth and maturation, delayed puberty, and prolonged the growth period. Surprisingly, no differences in adult height were recorded. Davis et al. (29) examined 36 children, ages 2 to 15 years, 25 boys and 11 girls, who were below the third percentile for height. Most displayed a poor appetite and had long-standing feeding or eating problems. Retarded growth in these otherwise healthy children was associated with prolonged, albeit modest, caloric deficiency. A sustained increase in caloric content resulted in renewed, more rapid growth.

Whether ultimate linear growth is impaired seems to depend on long-term nutrition. Casper and Jabine (18) reported that anorexia nervosa patients with good outcome 8 to 10 years after illness onset were as tall as their sisters, but that those who were still symptomatic were slightly, yet not significantly lower in height than their sisters. Root and Powers (72) reported growth retardation in three adolescents with anorexia nervosa as did Higgs et al. (46) in 5 of 23 childhood cases. But Pfeiffer et al. (66) reported that the mean percentile for height was greater at follow-up than at the time of diagnosis in those adolescents who had had anorexia nervosa at or before age 16 years suggesting significant late growth acceleration.

Conclusions and Recommendations for Treatment and Further Research

There is no doubt that the continued debate about the diagnostic classification of feeding and eating disturbances in infancy and childhood has hampered comparative studies. Other obstacles to data collection have been the difficulties and ethical considerations for conducting research in young children and the small sample sizes due to the low incidence of anorexia nervosa in childhood.

Most of the information about anorexia nervosa in childhood has come from single case reports or from retrospective chart reviews of hospitalized patients and only lately from cases seen at children’s hospitals. The major shortcoming of the literature is the inclusion of premenarchal cases and cases up to the age of 16 years as childhood cases, even though from a physiological point of view childhood comes to an end with the appearance of any sign of puberty.

In general, the behavioral and descriptive diagnostic criteria for anorexia nervosa are applicable to children, if one keeps in mind that the motivation and reasoning of the child differs from the adult and tends to reflect individual stages of development as well as the child's upbringing. Individually plotted growth curves are necessary to determine accurately body weight loss and growth deceleration in children.

The treatment of anorexia nervosa in children requires special knowledge and adjustments (5, 7, 9, 36, 57, 72). Disordered feeding and eating can have profound effects on the physical and psychological growth and development of children and it must be kept in mind that undernutrition in children leads more easily to physical complications, in particular cardiac and circulatory failure. Treatment needs to be swift and to respond effectively to the individual situation. Every effort should be made to explain, gently and patiently, the details of treatment to the child in order to win her or his cooperation. Nevertheless, adults need to take charge and supervised refeeding takes precedence over specific psychotherapeutic issues.

Since malnutrition restricts attention and concentration, and because the patient's emotional equilibrium is linked to the weight, psychotherapy achieves little without weight gain. Family therapy is mandatory and needs to accompany individual psychotherapy that will also identify age appropriate conflicts and apprehension about growing up and the sexual implications of puberty (17). On the other hand, the treatment of children is made easier by the infinitely greater willingness of the nursing staff to mother an eleven year old, even a hostile and obstinate eleven year old anorectic patient, than to care for a 20 year old anorectic patient.

In childhood, just as much as in adolescent cases, the severity of the eating disorder seems to be determined by the severity of the psychopathology in the child and the family. No drug to reverse the core symptoms of anorexia nervosa has been identified, but food alone has pharmacological properties, because well balanced nutrition and weight gain do ameliorate the core symptoms of anorexia nervosa and lead to remission. No controlled treatment studies have been conducted in childhood anorexia nervosa. Since children's responses to medication that might be indicated for a comorbid condition is not as reliable as in adults, supportive individual and family treatment to explore dysfunctional family interactions including an evaluation of parental psychopathology become all the more important.

The need for epidemiological, psychopathological, endocrine, genetic (47), and outcome data in childhood anorexia nervosa is obvious, but unless milder cases seen as outpatients or in pediatric practice can be included into the data base, studies cannot be considered representative of the full syndrome.

The observed differences in the prevalence rates for prepubertal and postpubertal males deserve further investigation. Likely hypotheses would explore the significant rise of adrogens during male puberty as well as contemporary gender-related cultural factors.

published 2000