Pica would appear to be without the nosological problems that beset other eating and growth disorders of infants and children. In fact DSM-IV, which lists pica as the first feeding/eating disorder, defines it as the persistent eating of nonnutritive substances for at least 1 month in such a fashion that such eating is inappropriate to developmental level and is not part of culturally sanctioned practice. The two dependent clauses in this definition disqualify the two largest populations described in the centuries-old literature about pica: toddlers who ingest paint chips and young pregnant women who eat starch and clay (22). This artificial clarity in definition attempts to solve the confusion that bedevils all aspects of the understanding of this disorder. Rather than being a narrowly homogenous phenomenon, pica represents an eating disorder of enormously differing populations, including normally developing toddlers mouthing lead paint chips, rural pregnant women eating clay or starch, severely retarded adults eating feces, and normally developed adults chewing pencil erasers, fingernails, and ice. As if to create an aura of scientific understanding, pica as defined in the literature before DSM-IV was subtyped according to the substance ingested. These terms include geophagia (eating clay), pagophagia (eating ice), plambophagia (eating lead), amylophagia (eating starch), coprophagia (eating feces), cautopyreiophagia (eating burnt matches), tricophagia (eating hair), lithophagia (eating stones), geomelophagia (eating raw potatoes). Aside from a review of Greek root words, such subtyping is most useful at documenting the extraordinary range of substances that people may ingest.

In many ways, the heterogeneity of pica precludes a coherent statement about age of onset, developmental outcome, etiology, or treatment. Because the pediatric literature has paid great attention to the phenomenology of lead poisoning, pica has become synonymous with children eating paint chips. For these children, pica has been reported to be more common in situations of relative environmental deprivation (23) or parental psychopathology. In this population, pica begins in the second or third year of life and may continue into childhood. In a second population, young pregnant women, the onset of pica begins with the first pregnancy in late adolescence or early adulthood and may continue intermittently for several decades. Pica in mentally retarded individuals begins in childhood and may diminish after middle age (24). In this population, the risk of pica appears to be related to the degree of retardation so that the most severely retarded are most likely to have the disorder (25). Pica in chronically anxious children or adults has never been studied or described in the literature. In this last group, a possible relationship among pica, trichotillomania, and other compulsive behavior disorders has never been investigated.

Incidence and prevalence figures for such a heterogenous disorder are difficult to interpret. However, among specific at-risk populations, pica has been reported to be quite common. Halstead (26) summarized findings from research studies that indicated that pica occurred in 25–33% of young children and 40–50% of poor black pregnant women. Pica among institutionalized retarded adults has been reported to range from 9% (24) to 25% (25).

Numerous studies have documented that the eating of specific substances such as clay and starch has been strongly influenced by cultural and family factors (27) because various social groups, especially poor and rural people, regard the eating of certain nonnutritive substances as acceptable. In addition, toddlers who are poorly supervised and understimulated are at high risk for ingesting inappropriate materials like paint chips.

Except as mediated by mental retardation, there are no known genetic factors in the etiology of pica. Boredom, anxiety, and depression may exacerbate pica, but little has been reported about such associated psychopathology. Iron deficiency is sometimes associated with pica, but the direction of the causal association is unclear (28).

While the extraordinary diversity of this disorder makes any useful classification schema unlikely, the remarkable prevalence and seriousness of its medical complications require a thoughtful, vigorous approach. These medical complications include heavy metal poisoning, mineral/vitamin deficiency, parasite ingestion, and intestinal obstruction. For toddlers and young children with a history of pica, a careful evaluation of the home environment and family functioning is required. In addition, these children should be evaluated for cognitive and psychiatric impairments. Pregnant women with pica presumably have more voluntary control over their eating. A careful assessment of their cultural and nutritional beliefs is essential so that nutritional counseling can be helpful. Severely retarded individuals presenting with pica should be assessed for access to noxious materials and provided with sufficient supervision to prevent pica.

Because of the extreme heterogeneity of the disorder, few comments are germane about treatment and natural history for the wide spectrum of people who meet the criteria for pica. For toddlers and young children, proper supervision is the most important intervention. In addition, careful scrutiny of the environment to remove toxic substances is crucial. In mentally retarded and developmentally delayed individuals both pharmacological and behavioral approaches have been used to treat pica. Most treatments focus on behavioral techniques such as response blocking, redirection, and differential reinforcement. There have been two cases in the literature of the use of antidepressants in the treatment of this disorder (29,30). However, this treatment is in need of further investigation. In anemic individuals both zinc and
iron supplementation have been used with some success (31). Older adolescents and young adults should be provided with psychoeducation and counseling about the medical risks associated with ingestion of nonnutritive substances.

Similar to pica, rumination disorder has the superficial appearance of a homogenous, unitary disorder. DSM-IV defines rumination as 1) repeated regurgitation and rechewing of food (in the absence of associated gastrointestinal illness) for a period of at least 1 month following a period of normal functioning, and 2) not due to an associated gastrointestinal or other general medical condition (e.g., esophageal reflux).

Very little is known about the incidence and prevalence of rumination, although the most severely affected cases, which are the ones that have been reported, are rare. The sex ratio of the disorder is also unclear. One case series (11) reported a male predominance, while another, larger series (32) reported a female predominance in the school age population.

Rumination appears to require three factors: an impaired ability of the infant or retarded person to regulate his or her internal state of satisfaction, a physical propensity to regurgitate food, and a learned association that regurgitation helps relieve the internal state of dissatisfaction. The assessment of the child must include an evaluation of each of these factors as well as the medical sequelae of rumination, such as malnutrition and aspiration. Obviously, these medical sequelae will influence the vigor of the intervention as well as serving as a baseline by which to measure the child’s response to treatment.

Unfortunately the clinical reality of this phenomenon is more complex than portrayed by DSM-IV. Some infants and retarded adults may vomit the stomach contents rather than rechewing and reswallowing it. Although this phenomenon is called psychophysiological (33) or operant (34) vomiting, its differentiation from rumination is problematic. Furthermore, the requirement that rumination not be associated with gastrointestinal disorders such as gastroesophageal reflux is overly restrictive since it implies that rumination has no underlying physiological contributions. Indeed the relationship, if any, between gastroesophageal reflux, rumination, and operant vomiting has not been investigated.

Because rumination is apparently a relatively rare disorder, virtually all of the information about the phenomenology must be derived from single case reports or small case series (e.g., (11,35,36)) without comparison groups. In otherwise normally developing infants, the age of onset is in the first year of life, whereas in mentally retarded individuals, the age of onset can extend into adulthood (11). Rumination has been reported to occur in nonretarded adults, many of whom have an additional comorbid eating disorder (37). In some unknown proportion of cases, the disorder remits spontaneously, while in others, the course may be malignant, marked by aspiration, severe malnutrition, growth failure, developmental delay, and death. Some authors (11) have proposed two diagnostic subgroups, self-stimulatory and psychogenic, to capture the difference in the course of retarded and nonretarded individuals, respectively. However, the relatively small sample size appears to make diagnostic subtyping premature.

Several authors have proposed that an adverse psychosocial environment is central to the development of rumination in otherwise normal infants (e.g., (38,11,39)). The most common environmental factor cited in the genesis of rumination is an unsatisfactory mother–infant relationship that causes the infant to seek an internal source of gratification. This turning inward by the infant has been proposed to occur because the environment is more stimulating than the infant can tolerate, or because the environment is not gratifying enough (39), or because the environment is too stimulating with negative effects (33,38). Aside from these proposed contradictory environmental factors, no information has been reported about more general sociodemographic features. Similarly, the role of genetic factors is unknown.

Learning theorists have explained rumination in terms of the reinforcing response that it elicits (34). These proposed feedback mechanisms include positive reinforcement when a desired event such as pleasure or attention follows rumination and negative reinforcement when an undesired event such as anxiety is reduced or removed. A more sophisticated theory involves combining the concept of positive and negative reinforcement by proposing a change in the valence of the behavioral consequences. For example, consequences that are normally behavior-suppressing may acquire behavior-reinforcing characteristics if other, usually positive, consequences are lacking (36). These concepts of the operant conditioning psychogenic factors in rumination are useful as the theoretical underpinnings of certain behavioral treatments of rumination.

The role of organic factors in rumination remains obscure, but some authors have argued that rumination is totally the result of physical disorders, including hiatal hernia and other esophageal abnormalities (40). The relationship between the syndromes of rumination and gastroesophageal reflux (GER) is unknown, although high rates of association have been reported (41). GER, or chalasia, is the syndrome of regurgitation of the stomach contents into the mouth and esophagus, apparently as a result of hypomotility of the gastric fundus and delayed gastric emptying rather than a weakened esophageal sphincter (42). A possible but untested hypothesis is that GER is the physiological substrate upon which various psychosocial disruptions or deviant operant conditioning act so that rumination develops.

The assessment of the infant’s capacity to regulate his or her internal state must include a general developmental assessment to evaluate for serious developmental delay as well as hyperirritability states. In addition, the mother–infant relationship must be carefully examined for clues to stimuli that are noxious or disruptive to the infant.

The child’s propensity to regurgitate can be best evaluated by procedures developed for GER, including esophageal pH monitoring, scintigraphic gastroesophageal reflux scan, endoscopy, and gastric emptying studies. In addition, various radiological procedures that evaluate esophageal abnormalities should be considered.

Similar to other eating disorders of early childhood, the extraordinary heterogeneity precludes specific treatments for all
persons with this disorder. The learned aspect of rumination, especially in retarded individuals, may be crucial to modify. After careful evaluation of this factor, various behavioral interventions must be considered, especially if there are serious medical complications. These interventions include aversive techniques in which a noxious stimulus is paired with rumination (43), and nonaversive techniques such as differential reinforcement of other incompatible responses. Indeed, in one large study (32) the nonaversive technique of habit reversal showed a greater than 80% success rate.

Failure to thrive (FTT) is a disorder of infancy and early childhood characterized by a marked deceleration of weight gain and a slowing or disruption of acquisition of emotional and social developmental milestones. Deceleration of linear growth and head circumference growth are associated but not primary phenomena.

DSM III (44) focused on the developmental delays presumably caused by psychosocial deprivation and virtually ignored the disordered growth and feeding. For this reason DSM III described FTT as a reactive attachment disorder. DSM IV (1) sought to redress this lack of recognition of the feeding disorders by introducing a new diagnosis of feeding disorder of infancy and early childhood. Although creating a new diagnostic category has the advantage of freeing it from a long history of misconception and controversy, it has the disadvantage of losing a long, rich history of research. Despite a 45-year history of study, the understanding of FTT has been marked by confusion and controversy about such basic issues as the definition and the name of the disorder. The plethora of syndromic names provides a glimpse into the confused literature: hospitalism, anaclitic depression, institutionalism, environmental retardation, maternal deprivation syndrome, psychosocial deprivation dwarfism, deprivational dwarfism, deprivation syndrome, failure to thrive, environmental failure to thrive, and nonorganic failure to thrive syndrome. This blizzard of interchangeably used but nonsynonomous terms, which frequently represented the mistaken or oversimplified underlying conceptions of the investigators, has created a major obstacle to the course of research. These misconceptions arose out of the multifaceted nature of the syndrome. At different times the three components of FTT— weight gain deceleration, linear growth delay and developmental delays— were separated so that each was perceived as the central aspect to the exclusion of other parts. In fact, depending upon the focus of the definition, FTT has been reported as consisting of only one component rather than a triad. For example, when the diagnosis of FTT is made on the basis of primary weight gain deceleration, developmental deficits are less evident (45), and indeed, in one study using age-matched controls, there were no differences in development test scores between FTT and normally growing infants (46). A second aspect of this confusion in the understanding of FTT has been the controversy between the contribution of emotional deprivation and that of malnutrition. From the earliest observers to present-day clinicians, the correlation between emotional misery and growth problems has been obvious. Perhaps too simplistically, some investigators argued that such disorders were directly caused by misery, mediated by some effect on the mind acting directly upon the body, without requiring such external factors as altered caloric intake.