SATIETY

Serotonin has long been a focus of attention for its possible role in disrupted satiety. There is substantial evidence that altered 5-hydroxytryptamine (5-HT; serotonin) functioning contributes to dysregulated appetite, mood, and impulse control in eating disorders and persists after recovery from AN and BN, possibly reflecting premorbid vulnerability.^19,20 There also is evidence that cholecystokinin (CCK) levels are altered in eating disorder populations. Findings for AN are inconsistent: There is some evidence that young women with AN have high levels of pre- and postprandial CCK, which may impede treatment progress by contributing to postmeal nausea and vomiting,^21,22 whereas other reports have shown decreased CCK compared with that of controls.²³ In those with BN, there is consistent evidence for an impaired satiety response, characterized by a blunted postprandial CCK response and satiety, as well as delayed gastric emptying.^24,25 In contrast, individuals with BED and obesity do not differ in postmeal CCK responses from those with obesity but no BED.²⁶ The relationships between CCK, binge eating, and BMI need further clarification. Lastly, protein tyrosine (PYY) functioning also appears to be dysregulated in BN and AN, but not in BED. Young women with AN have higher levels of PYY, the intestinally derived anorexigen that elicits satiety, compared with controls, perhaps contributing to reduced food intake.²⁷ In individuals with BN, expected elevations in PYY after meals are blunted,^28,29 possibly playing a role in impaired satiety. A recent report found no differences between BED and non-BED groups in fasting levels and postmeal changes in PYY.³⁰

APPETITE

The orexigenic peptide ghrelin is of interest for its role in eating disorders because it influences secretion of growth hormone (GH), stimulates appetite and intake, induces adiposity, and is implicated in signaling to the hypothalamic nuclei involved in energy homeostasis. There are two consistent findings in the literature examining ghrelin and AN: (1) circulating ghrelin levels are elevated, likely a consequence of prolonged starvation^31,32; and (2) GH and appetite responses to ghrelin are blunted, suggesting altered ghrelin sensitivity.^33,34 In BN, plasma levels of ghrelin are normal or elevated^28,29; of most interest is the postprandial blunted response (i.e., reduced suppression of ghrelin³⁵). Investigations of ghrelin functioning in individuals with BED have reported lower circulating levels of pre- and postmeal ghrelin, possibly reflecting down-regulation in response to chronic overeating and smaller decreases in ghrelin after eating.³⁰

ENERGY STORAGE

Leptin and adiponectin are hormonal signals associated with longer term regulation of body fat stores. Leptin is also directly implicated in satiety through its binding to the ventral medial nucleus of the hypothalamus, an area termed the satiety center. Leptin and adiponectin are both altered in patients with eating disorders. A number of studies have found evidence for hyperadiponectinemia and hypoleptinemia in populations of underweight AN with reversal following restoration of weight^36,37; increased adiponectin levels may act protectively to support energy homeostasis during food deprivation. Individuals with BN also exhibit decreased plasma levels of leptin, which are inversely correlated with length of illness and severity of symptoms.³⁸ The mechanism of altered leptin functioning in BN is unclear because blunted postmeal leptin levels are not observed in individuals with BED.²⁶

There are other mechanisms of interest, including but not limited to neuropeptide Y, GLP-1 and GLP-2, orexins A and B, the endocannabinoids, resistin (adipose tissue–specific secretory factor), pancreatic polypeptide, and brain-derived neurotrophic factor, but more research is necessary to clarify their roles in the pathophysiology of eating disorders. One high priority for research is clarifying whether observed psychobiological abnormalities are antecedents or consequences of disturbed eating behavior that return to normal after recovery, because this information would shed light on cause and possible treatment targets. Patients often report intense discomfort after eating as a reason that they continue to restrict intake. The discomfort may be dismissed as perceptual or psychological in the absence of any positive medical findings to support the symptoms, but there may in fact be disruptions in CNS or peripheral signals contributing to reported symptoms.

ANOREXIA NERVOSA

Anorexia nervosa is characterized by an unwillingness or incapacity to maintain a minimally normal weight (commonly described as at least 85% of expected weight or a body mass index [BMI] of 17.5 kg/m²), fear of gaining weight (despite being thin), a disturbance in the way weight is experienced (e.g., a denial of the medical seriousness of being underweight or feeling fat despite emaciation), and amenorrhea (in postmenarcheal females). Individuals with AN typically restrict their food selections and caloric intake, but approximately half of those with AN also routinely binge-eat and/or engage in inappropriate compensatory behaviors, such as induced vomiting or laxative use to prevent weight gain (see Fig. 8-1). AN is subdivided further into restricting type (i.e., those who primarily control their weight through dieting, fasting, or exercising) and binge-eating–purging type (i.e., those who routinely purge calories to control weight and may or may not routinely binge-eat).⁸ In middle-aged and older women, new-onset AN may present in conjunction with difficulty making life transitions and fear of aging.⁴² The diagnosis of AN may be delayed when patients present to a GI specialty practice without disclosing their concerns and behaviors relating to weight. Presentation with GI complaints, even if related to real symptoms or disease, can sometimes prove to be a red herring, drawing attention away from and delaying diagnosis of an eating disorder. One study evaluated 20 consecutive patients presenting to a GI practice who were ultimately diagnosed with an eating disorder, and found that patients did not receive a diagnosis of an eating disorder for an average of 13 months after presentation. Notably, all patients stated a desire to gain weight and denied attempts to lose weight via exercise, purging, or dietary restriction.⁵⁶ Individuals with AN are not always able or willing to frame their difficulty maintaining a healthy weight as intentional; thus, the diagnosis initially may be unsuspected and delayed.

BULIMIA NERVOSA

The clinical hallmark of BN is recurrent binge eating accompanied by inappropriate compensatory behaviors to control weight or to purge calories consumed during a binge. On average, these behaviors must occur twice weekly for at least three months to meet diagnostic criteria.⁸ Also intrinsic to the diagnosis of BN is the excessive influence of weight and/or shape on self-image.

By definition, binge eating is consumption of an unusually large amount of food during a “discrete period of time” (i.e., not overeating or grazing all day), accompanied by the feeling that the eating cannot be controlled.⁸ Many patients describe an emotional numbing during the period of eating. For some, this state appears to motivate the bingeing. Most clinicians are familiar with self-induced vomiting as the primary purging behavior, but individuals with BN often use alternative or additional means of preventing weight gain, including abuse of laxatives and/or enemas, diuretics (especially among health care workers), stimulants (including methylphenidate, cocaine, ephedra, and caffeine), underdosing of insulin (for those with diabetes mellitus), fasting or restrictive eating, and excessive exercise (see Table 8-1). Intentional consumption of gluten to promote weight loss in adolescents with celiac disease also has been reported.⁵⁷

Whereas most compensatory behaviors to prevent weight gain fall within the diagnostic subtype of purging BN, excessive exercise and fasting are behaviors categorized within the subtype of nonpurging BN.⁸ Because of the absence of the more classic purging behaviors and because of their frequent indistinct nature, this variant of BN often is challenging to identify. As with overeating and dieting, it frequently is difficult to determine the line between culturally normative and pathologic behavior with excessive exercise. Generally, clinical suspicion should be raised when an individual continues to exercise despite an injury or illness, or if he or she is exercising routinely in excess of what a coach is recommending for the team.

It is recommended that clinicians explore the presence of purging behaviors if an eating disorder is suspected. Although it is not certain that a patient will respond candidly, individuals probably are more likely than not eventually to disclose information about symptoms when asked.⁵¹ Some patients report feeling relieved when clinicians pose such questions if they previously had not been able to discuss their symptoms. On occasion, however, patients report learning about techniques from clinicians’ questions, so it is advisable to provide a psychoeducational context for the questions (e.g., by conveying serious physical consequences associated with the behavior, such as ipecac use) and to avoid introducing information about a dangerous behavior (e.g., underdosing insulin), depending on the clinical context. Patients also benefit from learning that treatment is available and that their clinicians understand the illness.

Whereas all these purging and other behaviors aimed at neutralizing calorie intake and controlling weight can pose medical risks when chronic, some of them pose more immediate, and potentially lethal, consequences. Patients should be informed of these acute life-threatening risks and steps should be taken to eradicate such behaviors immediately. For example, because of the serious neurotoxicity, cardiotoxicity, and risk of death associated with repeated syrup of ipecac ingestion,⁵⁸ its ongoing use is a clinical emergency and may require immediate hospitalization. Many patients are unaware of the serious risk associated with syrup of ipecac use. Similarly, ephedra, now banned in the United States, poses risk of stroke or adverse cardiac events, even in young adults.⁵⁹ Some ephedra-free supplements marketed as weight loss agents also may be proarrhythmic and pose medical risks.⁶⁰ Although patients find it difficult to abstain from purging behaviors, they may be willing to substitute less immediately harmful behaviors while treatment is initiated.

Individuals with BN are excessively concerned with weight and body shape. For example, they may be preoccupied with a running mental tally of calories and plans to neutralize them to prevent weight gain. Their self-image frequently is poor and anchored to their weight. It is not unusual for individuals with AN or BN to weigh themselves daily, even several times each day, and experience fluctuations in self-esteem and mood based on the result.

EATING DISORDER NOT OTHERWISE SPECIFIED

EDNOS covers a broad range of clinical manifestations, including atypical symptoms, symptoms of BED, symptoms consistent with NES, and subthreshold, yet clinically significant, eating disorders. Although EDNOS is a residual category, it nonetheless is the most commonly diagnosed eating disorder in outpatient settings, and therefore there is much interest in refining diagnostic categories of eating disorders.⁶¹ EDNOS currently includes eating disorders that do not meet threshold criteria for duration or frequency for AN or BN as well as BED, NES, purging disorder and other atypical variants. Diagnostic classification of eating disorders will be updated in the DSM-V, with an expected publication date of 2012.

BINGE EATING DISORDER

BED is a variant of EDNOS, although there is a substantial literature to support its prevalence and consistent response to specific therapeutic strategies. Like BN, BED is characterized by recurrent binge-eating. To meet provisional criteria for BED, the binge eating episodes must occur two days per week, on average, for at least six months. Unlike BN, however, BED is not associated with recurrent inappropriate compensatory behaviors to prevent weight gain. BED is distinguished from nonpathologic overeating by several possible associated symptoms, including rapid eating, eating irrespective of hunger or satiety, eating alone because of shame, and negative feelings after a binge.⁸ Apart from overweight or obesity, BED patients frequently present without any specifically associated physical findings. Although in some cases binge eating associated with BED may cause or perpetuate weight gain, many with BED develop symptoms only after they have become overweight. Individuals with BED frequently are distressed enough about their symptoms to seek medical help, although they may present seeking a solution to their weight gain rather than their binge eating. A substantial percentage of patients seeking weight loss treatment will have comorbid BED or NES. Therefore, medical subspecialists are likely to encounter these patients before they have been diagnosed with BED.

NIGHT EATING SYNDROME AND NOCTURNAL SLEEP-RELATED EATING DISORDER

NES is a pathologic eating pattern that may be considered a variant of EDNOS. First described in 1955,⁶² it, too, is characterized by recurrent bouts of overeating—but not necessarily bingeing—without associated inappropriate compensatory behaviors to prevent weight gain. As such, some individuals may appear to meet criteria for NES and BED, but these are distinct syndromes with relatively little overlap,^7,63 and proposed criteria for the syndrome exclude a concomitant diagnosis of BN or BED.⁶⁴ There is no clear consensus on core criteria for NES, although most investigators propose morning anorexia, evening hyperphagia, and sleep disturbance (operationalized in various ways); some propose the additional criterion of eating in relation to sleep disturbance, such as during a nighttime awakening.⁶⁵ In one study, NES in obese subjects was associated with an average of 3.6 awakenings/night compared with just 0.3 awakenings/night for matched controls. Subjects with NES ate during 52% of their awakenings, taking in a mean of 1134 kJ/episode, considerably less than the usual intake of a binge associated with BN or BED.⁶⁴ NES also can occur in nonobese individuals⁶⁶ but is more common in the obese and may contribute to poor outcome in weight loss treatment programs.^64,67 NSRED is also characterized by nighttime snacking, but individuals typically are totally or partially unconscious (e.g., they are in stage 3 or 4 sleep) during the snacking and frequently do not remember it.⁷

PURGING DISORDER

Emerging evidence raises the possibility of an additional distinctive eating disorder variant characterized by recurrent purging symptoms in the absence of clinically significant binge pattern eating. The proposed name for this diagnostic category is purging disorder. Crossover between this variant and BN appears to be rare, lending support to the hypothesis that this represents a distinctive clinical phenomenon⁶⁸; data also support comparable severity to BN. Lifetime prevalence of purging disorder has been estimated as 1.1% to 5.3% of young adult women. Course, outcome, and treatment strategies for purging disorder require further research.⁶⁹

DIFFERENTIAL DIAGNOSIS

The differential diagnosis of the eating disorders includes evaluation and exclusion of medical causes of weight loss, weight gain, anorexia, hyperphagia, vomiting, and other associated symptoms. These considerations are especially germane in cases of atypical or early- or late-onset eating disorders.⁴² Medical causes of appetite and/or weight loss include hyperthyroidism, diabetes, malignancy, and infectious diseases, among the systemic disorders, and substance abuse, depression, dementia, delirium, and psychosis. Illnesses associated with weight gain include hypothyroidism, Cushing’s disease, and organic brain disease. The differential diagnosis of hyperphagia is broad and includes Prader-Willi syndrome, dementia (including Alzheimer’s disease), and intracranial lesions. Hyperphagia also has been associated with the use of certain medications, particularly many of the psychotropic agents (e.g., lithium, valproate, tricyclic antidepressants, mirtazapine, and conventional and atypical antipsychotic agents), pregnancy,⁷⁰ and poststarvation refeeding.⁷¹

Psychiatric illnesses associated with loss of appetite and weight loss include major depression, anxiety, and substance-use disorders. Moreover, comorbid psychiatric illness is common among those with eating disorders,³ and frequently complicates their diagnosis and treatment. Thus, identification of excessive concern with weight and food intake, unrealistic or inappropriate weight goals, or resistance to attempts to restore normal weight and/or limit excessive exercise can be helpful in distinguishing an eating disorder from another psychiatric illness or in revealing the presence of an underlying comorbid eating disorder. Because individuals with BN and EDNOS can have an unremarkable physical examination on presentation, the diagnosis may remain obscure until the patient discloses his or her symptoms, or until the clinician suspects an eating disorder based on other elements of the clinical history (e.g., weight fluctuations or menstrual irregularities).

Although there potentially is much phenomenologic overlap among the eating disorders and individuals do cross over from one diagnostic category to another (Table 8-2),⁴³ categories are mutually exclusive, according to DSM-IV* diagnostic criteria.⁸ Even though a transdiagnostic approach to eating disorders classification and treatment has been proposed,⁷² differing responses to treatment make it desirable to establish a clear diagnosis to optimize care. A weight criterion distinguishes anorexia nervosa from bulimia nervosa in some cases. Individuals who are substantially underweight (e.g., 85% or less of expected body weight) and who otherwise meet criteria for AN most likely should be classified as having AN, even if bingeing, purging, or both are present. Individuals with BED or BN also can have symptom overlap. BN is distinguished by recurrent purging and other behaviors directed at neutralizing excessive calorie intake so as to prevent weight gain, as well as an excessive concern with weight.

NUTRITIONAL AND MEDICAL EVALUATION

In addition to excluding medical causes of weight and appetite changes, medical evaluation for confirmed or suspected eating disorders includes obtaining a full history of the patient’s eating behaviors, with attention to daily number of calories ingested, purging behavior (e.g., vomiting, use of ipecac, or laxative use), and exercise patterns. Often, medical evaluation will be guided by an assessment of nutritional status, which includes determination of the appropriateness of weight for height, age, and gender.

NUTRITIONAL EVALUATION

There are several established means for evaluating nutritional status in the office, central to which is measuring weight and height (see Chapter 5). Assessment of the appropriateness of weight for height is one of the key factors intrinsic to determining the urgency of medical and psychiatric care. For patients with AN, it is important not to rely on self-reported weight, given the strong possibility of an inaccurate report. Individuals with AN often go to great effort to conceal their low weights. For example, some patients “water load” prior to a clinical encounter, some attach weights to themselves, and others layer loose and bulky clothing to create the illusion of being of normal weight. Assessment of weight, therefore, should factor in the possibility that a patient may wish to conceal a low weight or weight loss. Some clinicians will find it helpful to have a scale in a private area (i.e., not in a hallway) and a clear and consistent protocol for weighing patients with AN. This might include asking them to void prior to being weighed, to change into a hospital gown, and to remove heavy jewelry. When patients have a history of consuming water prior to an appointment to increase their measured weight, it may be helpful to check a urine specific gravity in order to adjust interpretation of the office weight.

Standard means of evaluating the appropriateness of weight for height include use of the BMI. This is calculated as follows:

Although BMI may not be an appropriate standard for evaluating a healthy weight status in professional athletes, with relatively high lean muscle mass, and in some ethnic groups (e.g., Polynesians may have a different cut point for obesity⁷³), BMI generally is appropriate for men and women aged 18 years or older. A BMI within the range of 18.5 to 24.9 kg/m² for men and women is considered normal. A BMI of 17.5 kg/m² or less is the threshold for meeting the underweight criterion for AN in the ICD-10 (International Classification of Diseases 10).^8,74 A BMI in the range of 25 to 29.9 kg/m² is consistent with overweight, and a BMI higher than 30 kg/m² reflects obesity.⁷⁵

An alternative means for evaluating weight for height that is especially useful to assess nutritional compromise and weight recovery in those with eating disorders expresses the patient’s weight as a percentage of expected body weight. The formula is as follows:

Expected weight for height = 100 pounds + 5 pounds per inch above 5 feet ± 10% for women, 106 pounds + 6 pounds per inch above 5 feet ± 10% for men and, if the patient is shorter than 5 feet, then the same number of pounds per inch is subtracted for each inch below 5 feet.⁷⁶

Although this is a linear equation (compared with the quadratic equation for BMI) and may be less useful at extreme heights, it is straightforward to calculate. Moreover, conceptually, this formula may be easier for patients and families to understand, especially in setting weight goals or limits. A 90% to 110% range of expected body weight is considered within the normal range and is a good place to begin for setting weight gain goals for patients with AN. Within this range, the goal will be refined by clinical history (including the patient’s history of baseline, minimal and maximal weights), whether and when menses return, and medical parameters, such as reversal of bone loss. Patients below 85% of expected body weight likely meet the weight criterion for AN, and those below 75% of expected body weight are seriously nutritionally compromised and generally require inpatient care.⁷⁷ For patients who are overweight or obese (>110% or >120% expected body weight, respectively), it may not be realistic or desirable to set weight goals within the normal range. Considerations in weight management are discussed subsequently.

Seriously nutritionally compromised patients require inpatient care for both efficacy and safety of weight management. For underweight patients without this degree of compromise, the primary goals of nutritional management are increasing caloric requirements to regain weight, ensuring adequate intake and balance of macro- and micronutrients, and reestablishing a dietary pattern of three meals daily. Patients’ diets are supplemented routinely with calcium (if dietary intake is inadequate) and multivitamins containing vitamin D. Some may require additional dietary guidance and adjustments because many patients restrict not only calories, but specific foods or food groups as well. For patients with BN, BED, and EDNOS, dietary intervention includes moderating excessive calorie intake and establishing a pattern of eating that is less vulnerable to emotional cues and excessive hunger. Many patients with eating disorders are quite knowledgeable about nutrition and commonly wish to avoid meeting with a nutritionist. Conversely, information from a nutritional assessment is invaluable to the treatment team and even well-informed patients are likely to benefit from reinforcement of more healthful food choices, meal patterns, and appropriate intake.

Medical Evaluation

Medical evaluation includes a clinical history with special attention to weight fluctuations and any purging or other inappropriate behaviors to neutralize calorie intake to control weight (see Table 8-1). Ascertainment of syrup of ipecac use (as an emetic) and nonadherence to insulin protocols in patients with diabetes mellitus is essential, given the potentially lethal sequelae of these behaviors.⁷⁸ Symptoms of medical complications of undernutrition, overnutrition, excessive exercise, or purging should be assessed and a menstrual history should be clarified. Physical examination includes a comprehensive assessment of potential complications of nutritional deficiencies, underweight, overweight, excessive exercise, and purging behaviors. If an eating disorder is suspected, physical examination may reveal signs to confirm nutritional compromise (e.g., bradycardia, hypotension, hypothermia, lanugo, breast tissue atrophy, muscle wasting, peripheral neuropathy) or to suggest chronic purging (e.g., Russell’s sign, an excoriation on the dorsum of the hand from chronic scraping against the incisors); hypoactive or hyperactive bowel sounds; an attenuated gag reflex⁷⁹; dental erosion (perimolysis; Fig. 8-2)⁸⁰; or parotid hypertrophy (Fig. 8-3).⁸¹

Figure 8-2. Dental erosions resulting from chronic vomiting.

(Adapted with permission from the Department of Psychiatry, Massachusetts General Hospital, Boston.)

Figure 8-3. Patient with parotid hypertrophy resulting from chronic vomiting.

(Adapted with permission from the Department of Psychiatry, Massachusetts General Hospital, Boston.)

Medical complications of behaviors associated with AN, BN, BED, and EDNOS are potentially serious and are too numerous to review in detail here; selected complications are listed in Table 8-3. Complications that are common and/or associated with serious morbidity should be actively sought on physical examination and laboratory studies, so that appropriate interventions can be initiated. Examples of such important and common findings include abnormal vital signs (e.g., hypotension, orthostatic hypotension, bradycardia, hypothermia), low weight or overweight, osteopenia or osteoporosis,⁸² and dental pathology (e.g., perimolysis [erosion of the tooth enamel], caries, or both).^81,83,84 Cardiac complications can be lethal and include prolonged QT interval, QT dispersion, ventricular arrhythmias, and cardiac syncope.^85,86 Neurologic findings in AN include cortical atrophy and increased cerebral ventricular size.⁸⁷ Endocrinologic abnormalities include menstrual abnormalities, low serum estradiol levels, low serum testosterone levels, hypercortisolism, and euthyroid sick syndrome, with resultant hypotension and cold intolerance.⁸⁸ Reported complications of eating disorders during pregnancy include miscarriage, inadequate weight gain of the mother, intrauterine growth retardation, premature delivery, infants of low birth weight and low Apgar scores, and perinatal death.^89–92

Table 8-3 Selected Clinical Features and Complications of Behaviors in Patients with Eating Disorders

SYSTEM AFFECTED	Clinical Feature or Complication
	ASSOCIATED WITH WEIGHT LOSS AND FOOD RESTRICTION OR BINGE-EATING IN ANOREXIA NERVOSA	SSOCIATED WITH PURGING OR REFEEDING BEHAVIORS IN ANOREXIA NERVOSA, BULIMIA NERVOSA, OR EDNOS
Cardiovascular	Arrhythmia Bradycardia Congestive heart failure (in refeeding syndrome) Decreased cardiac size Diminished exercise capacity Dyspnea Hypotension Mitral valve prolapse Orthostasis Prolonged QT interval QT dispersion Syncope

Ventricular arrhythmia

Cardiomyopathy (with ipecac use)

Prolonged QT interval

Orthostasis

Syncope