Psychiatric management includes a broad range of therapeutic actions that are performed by the psychiatrist or that the psychiatrist ensures are provided to all patients with eating disorders in combination with other specific treatment modalities. Psychiatric management begins with the establishment of a therapeutic alliance, which is then enhanced by empathic comments and behaviors, positive regard, reassurance, and support. Basic psychiatric management includes support through the provision of educational materials, including self-help workbooks (4), information on community and Internet resources (5, 6), and direct advice to patients and their families (when they are involved) (7). It is important to caution patients and families about Internet sites that encourage eating disorder lifestyles ("pro-ana" sites). Although many service providers have made attempts to police and encourage elimination of these sites, they still continue to appear, to the concern of families and professionals (8, 9). In some settings, judicious use of e-mail contact with patients has been increasingly used (5, 10). Some resources for patients and families are presented in Table 1.

At the very outset and through ongoing interactions with the patient, it is important for clinicians to attempt to build trust, establish mutual respect, and develop a therapeutic relationship that will serve as the basis for ongoing exploration and treatment of the problems associated with the eating disorder. Eating disorders are frequently long-term illnesses that can manifest themselves in different ways at different points during their course; treating them often requires the psychiatrist to adapt and modify therapeutic strategies. Many patients with anorexia nervosa are initially reluctant to enter treatment and may feel invested in their symptoms. Many are secretive and may withhold information about their behavior because of shame. During the course of treatment, they may resist looking beyond immediate symptoms to possible coexisting psychiatric disorders, comorbid psychopathology, and underlying psychodynamic issues. Conversely, some patients may resist discussing eating disorder symptoms and want to focus on only "core issues," apparently to avoid relinquishing their symptoms. Psychiatrists should be mindful of the fact that the recommended interventions create extreme anxieties for individuals with anorexia nervosa. Encouraging patients to gain weight asks them to do the very thing of which they are most frightened. Patients may believe that the psychiatrist just wants to make them fat and does not understand or empathize with their underlying emotions. Consequently, by recognizing and acknowledging an awareness of patient anxieties, psychiatrists can assist in building the therapeutic alliance. The clinician may foster rapport by letting patients know that eating disorder symptoms often serve a number of important functions, such as providing a sense of accomplishment or a way to feel looked after or protected (11, 12). Addressing patients' resistance to treatment and enhancing their motivation for change may be important in allowing therapy to proceed through impasses as well as helping to ameliorate factors that serve to aggravate and maintain eating disorders (13–18). Finally, letting patients know that full recovery from anorexia nervosa takes time (19) may help build rapport, as the patient senses that the clinician is not expecting a magical, rapid turnaround, which the patient may sense is unrealistic.

Professionals from several disciplines may collaborate in the patient's care. The specific role of each professional may vary with the organizational structure of the eating disorders program and the professional qualifications of those working within the program. The psychiatrist may assume the leadership role in the patient's treatment program or the patient's treatment team or work collaboratively on a team led by other health professionals, including other physicians or psychologists. Registered dietitians with specialized training in eating disorders often provide nutritional counseling. Therapists from a variety of professional fields may provide family, individual, or group psychotherapy, including CBT. Other physician specialists and dentists may be consulted for management of acute and ongoing medical and dental complications. Often in the treatment of children and adolescents, school coaches, teachers, and school counselors may be asked to collaborate in a patient's treatment. In treatment settings where staff do not have the training or experience to deal with patients with eating disorders, the provision of education, supervision, and leadership by a qualified psychiatrist can be crucial to the success of treatment.

Although a variety of management models are used for adult patients with eating disorders, no data exist on their comparative efficacies. Psychiatrists who choose to manage both general medical and psychiatric issues should have appropriate medical backup to treat the medical complications associated with eating disorders. Some programs routinely arrange for interdisciplinary teams to manage treatment (sometimes called split management). In this model, the psychiatrist handles administrative and general medical requirements, prescribes medications when clinically necessary and appropriate, and recommends interventions aimed at normalizing disturbed cognitions and eating and weight-reducing behaviors. Other clinicians then provide individual and/or group psychotherapeutic interventions (e.g., CBT, psychodynamic psychotherapy, family therapy). For this management model to be effective and to avoid reinforcing some patients' tendencies to play staff off each other (i.e., split the staff), all personnel must work closely together and maintain open communication and mutual respect.

For children and adolescents, the recommended treatment model is the team approach (3). In this interdisciplinary management approach, general medical care clinicians (e.g., specialists in internal medicine, pediatrics, adolescent medicine, or nutrition) manage general medical issues, such as nutrition, weight gain, exercise, and eating patterns, whereas the psychiatrist addresses the psychiatric issues (3, 20, 21). The biopsychosocial nature of anorexia nervosa and bulimia nervosa dictates the need for interdisciplinary treatment, and each aspect of care must be developmentally tailored to the treatment of adolescents (22). In unusual circumstances, psychiatrists may be qualified to act as the primary provider of comprehensive medical care.

When a patient is managed by an interdisciplinary team in an outpatient setting, communication among the professionals is essential so that all team members have a clear understanding of each other's responsibilities and approaches. For example, in team management of outpatients with anorexia nervosa, one professional must be designated to consistently monitor weights so that this essential function is not inadvertently omitted from care.

It is important for the psychiatrist to carefully assess the patient's eating disorder symptoms and behaviors (23). Such an assessment will assist the clinician in identifying target symptoms and behaviors that will be addressed in the treatment plan as well as determining whether a DSM-IV-TR diagnosis of anorexia nervosa or bulimia nervosa is present (Tables 2 and 3).

It is important to note that a significant number of patients are relegated to the heterogeneous diagnostic group referred to as eating disorders not otherwise specified because they have not been amenorrheic for 3 months and consequently do not meet current criteria for anorexia nervosa. In terms of their clinical course, treatment response, or level of impairment, such patients do not differ from those who fulfill the DSM-IV-TR criteria for anorexia nervosa (24, 25). These observations have important implications with respect to making clinical treatment decisions. They also imply that patients with continued menses who fulfill other criteria for anorexia nervosa should be eligible for the same levels of care as patients with anorexia nervosa.

Obtaining a detailed report of food intake during a single day in the patient's life or using a calendar as a prompt may help elicit specific information about a patient's eating behaviors, particularly regarding perceived intake. A clinician may also obtain useful information by sharing a meal with the patient or observing the patient eating a meal; in this way, the clinician can observe any difficulties the patient may have in eating particular foods, anxieties that erupt in the course of a meal, and rituals concerning food (such as cutting, separating, or mashing) that the patient feels compelled to perform.

It is important to explore the patient's understanding of how the illness developed and the effects of any interpersonal problems at the onset of the eating disorder. A family history should be obtained regarding eating disorders and other psychiatric disorders, alcohol and other substance use disorders, obesity, family interactions in relation to the patient's disorder, and family attitudes toward eating, exercise, and appearance. It is essential that the clinician avoid articulating theories that imply blame or permit family members to blame one another or themselves for the patient's disorder. No evidence exists to prove that families cause eating disorders. Furthermore, blaming family members harms their psychological well-being and often impairs their desire, willingness, and capacity to be helpful to patients and to participate actively and constructively in treatment and recovery. Rather, the point is to identify family stressors whose amelioration may facilitate recovery.

In the assessment of young patients, it is always helpful to involve parents and, whenever appropriate, school and health professionals who routinely work with children. The complete assessment usually requires several hours. Even when directly questioned, patients and their families may not initially reveal pertinent information about sensitive issues; important information may be uncovered only after a trusting relationship has been established and the patient is better able to accurately identify inner emotional states.

Formal measures are available for the assessment of eating disorders, including self-report questionnaires and semistructured interviews. Examples are listed in Table 4. Clinical decisions about a diagnosis cannot be made on the basis of self-report screening instruments. Patients who are identified on initial screening as likely to have an eating disorder must be followed up in a second-stage determination by trained clinical interviewers. The instruments shown in Table 4, used by clinicians to interview patients in a structured format, are generally taken as "gold standards" to determine clinical diagnoses.

A full physical examination should be performed by a physician familiar with common findings in patients with eating disorders, with particular attention to vital signs; physical status (including height and weight); heart rate and rhythm; jugular venous pressure; heart sounds (especially midsystolic clicks or murmurs from mitral valve prolapse); acrocyanosis; delayed capillary refill; lanugo; salivary gland enlargement; scarring on the dorsum of the hands (Russell's sign); evidence of self-injurious behaviors, such as ecchymoses, linear scars, and cigarette burns; muscular weakness; indications of muscular irritability due to hypocalcemia, such as in Chvostek's and Trousseau's signs; and gait and eye abnormalities (40–43). The patient should also be referred for a dental examination if necessary or indicated by the patient's history (44). In younger patients, examination should include growth pattern and sexual development, including sexual maturity rating, as well as general physical development. The use of a growth chart of standardized values for pediatric populations may allow the clinician to identify patients who have failed to gain weight and have growth retardation (3, 45); such charts are available on the web site of the CDC (http://www.cdc.gov/growthcharts/).

BMI, in conjunction with weight and height, has gained increasing attention in research and clinical settings as a tool for assessing eating disorder patients. BMI is calculated as weight (in kilograms) divided by height (in meters squared) and is particularly useful for comparing groups according to index percentiles that take into account height, sex, and age (46). It is important to remember that BMI is a calculation based only on height and weight and does not provide any further measure of body composition. Except in individuals who are extremely under- or overweight, it is often not useful in estimating nutritional status. Furthermore, considerable debate in the scientific community exists about appropriate BMI ranges for various ethnic groups. Among Caucasian women, for example, the range of a healthy BMI may be higher than for some groups of Asian women (47). Adults with a BMI <18.5 kg/m² are considered to be underweight. In addition, abnormal muscularity, body frame size, fluid status, marked constipation, and fluid loading can decrease the validity of BMI as an indicator of the patient's nutritional status (48–52). In children and adolescents, an age-adjusted BMI is used (see http://www.cdc.gov/nccdphp/dnpa/bmi/00binaries/bmi-tables.pdf). Children with a BMI <5th percentile for age are considered to be underweight.

Commonly found signs, symptoms, and associated laboratory abnormalities for anorexia nervosa and bulimia nervosa are shown in Table 5 and Table 6, respectively. Although patients treated in outpatient practice may report few symptoms and show few obvious physical signs or abnormal laboratory test results, significant occult abnormalities may be present (e.g., in bone, heart, and brain).

The need for laboratory analyses should be determined on an individual basis depending on the patient's condition or when necessary for making treatment decisions (43). Some of the laboratory assessments that may be indicated for patients with eating disorders appear in Table 7.

In addition to assessing patients' physiological and nutritional status as well as their behaviors, cognitions, and emotions associated with eating and exercise, it is essential that clinicians attend to the overall psychiatric status and safety of patients. Associated psychiatric issues that bear close monitoring include historical evidence, signs, and symptoms related to psychiatric conditions that are often comorbid with eating disorders (e.g., mood, anxiety, and substance use disorders) as well as personality traits and personality disorders that greatly influence patients' clinical course and outcome (78–82). Patients' motivational status also bears monitoring, as it is likely to determine their capacity to engage in treatment (15). Safety issues for patients with eating disorders include both physiological and psychiatric parameters. Many of the physiological safety parameters are described in Section II.B.1, "Choice of a Treatment Site." Clinicians must be vigilant regarding shifts in patients' weight, blood pressure, pulse, other cardiovascular parameters, and behaviors that are likely to provoke physiological decline and collapse. General psychiatric safety issues that bear constant attention include suicidal ideation and suicide attempts as well as impulsive and compulsive self-harm behaviors (83, 84).

The available evidence affirms the importance of family involvement and treatment in the management of children and adolescents with anorexia nervosa (85–87). In addition, clinical consensus supports the value of family assessment and involvement in the treatment of both younger and older patients with other eating disorders (88, 89). Since eating is a quintessential family activity, the opportunity to observe patterns of family interaction around the eating and particularly around the eating problems can be useful in assessment (89a). Family members can provide useful perspectives on factors contributing to the onset of the disorders and issues that may aid or hamper efforts at recovery. Family members are often distressed by difficulties in understanding and interacting with the patient. Clinicians need to empathically listen to family members, advise them on their interactions with the patient, and, when indicated, involve them in conjoint or individual treatment so that the patient and family all stand the best chance of achieving a good outcome (90). Patients with anorexia nervosa who are in a relationship may present with a higher motivation to change (91), and the involvement of spouses and partners in treatment may be highly desirable. Families of adolescents with anorexia nervosa may be directed to the Maudsley approach, which focuses on the family as a resource for recovery and puts parents in charge of refeeding their affected child (87, 92, 93). Although this approach is promising, additional data are required to determine if it is the best approach for adolescents with anorexia nervosa.

Services available for the treatment of eating disorders can range from intensive inpatient settings (in which subspecialty general medical consultation is readily available) to residential and partial hospitalization programs to varying levels of outpatient care (in which the patient can receive general medical treatment, nutritional counseling, and/or individual, group, and family psychotherapy). Because specialized programs are not available in all geographic areas and financial considerations are often significant, access to these programs may be difficult. The resources listed in Table 1 may provide guidance to patients and families for accessing suitable programs.

Pretreatment evaluation of the patient is essential for determining the appropriate treatment setting. Patient weight, rate of weight loss, cardiac function, and metabolic status are the most important physical parameters for making this choice. Eating disorders should be recognized and early treatment implemented as soon as possible after the onset of symptoms. This is especially true in children, adolescents, and young adults, to avoid the disorder becoming chronic.

As a general rule, patients who weigh less than approximately 85% of their individually estimated healthy weights have considerable difficulty gaining weight outside of a highly structured program that includes inpatient care; such a program may be medically and psychiatrically necessary even for patients above that weight level. It is important to underscore that these are individually estimated healthy weights, not weights simply listed in a standard insurance table. Healthy weight estimates for a given individual must be determined by that person's physicians on the basis of historical data (e.g., growth charts) (3) and, for women, the weight at which healthy menstruation and ovulation resume, which may be higher than the weight at which menstruation and ovulation became impaired (94–96).

It is equally important that the decision to hospitalize a patient be based on psychiatric and behavioral factors, including a rapid or persistent decline in oral intake; a decline in weight despite maximally intensive outpatient or partial hospitalization intervention; the presence of additional stressors, such as dental procedures, that may interfere with the patient's ability to eat; the weight at which the patient was medically unstable in the past; and co-occurring psychiatric problems that merit hospitalization. The degree of a patient's denial and resistance to participate in his or her own care in less supervised settings is critical in deciding whether to hospitalize the patient. Once weight loss is severe enough to indicate the need for immediate hospitalization, treatment may be less effective, refeeding may entail greater risks, and prognosis may be more problematic than if intervention had been provided earlier. Because cortical gray matter deficits result from malnutrition and persist after refeeding (97–99), earlier (rather than later) interventions may be important to minimize the persistent effects of these physiological impairments. Therefore, hospitalization should occur before the onset of medical instability as manifested by vital signs, physical findings, or laboratory test results outside of the normal range.

Vital sign changes that indicate a need for immediate medical hospitalization include marked orthostatic hypotension, with an increase in pulse of 20 bpm or a drop in blood pressure of 20 mmHg standing; bradycardia, with a heart rate <40 bpm; tachycardia, with a heart rate >110 bpm; or the patient's inability to sustain his or her body core temperature (e.g., body temperatures <97.0°F) (3). Most severely underweight patients, particularly those with physiological instability, require inpatient medical management and comprehensive treatment to support their weight-gaining efforts. To avert potentially irreversible effects on physical growth and development, many children and adolescents require inpatient medical treatment, even when weight loss, although rapid, has not been as severe as that suggesting a need for hospitalization in adult patients (3). If children refuse fluids or food out of concern about gaining weight, they may become dehydrated quickly. Also, a child's small size may mean that relatively smaller reductions in weight will result in greater physiological danger.

In determining a patient's initial level of care or suitability for change to a different level of care, expert consensus indicates that it is important to consider a patient's overall clinical and social picture rather than simply rely on weight criteria. Furthermore, weight level per se should never be used as the sole criterion for discharge from inpatient care. Patients need to both gain healthy body weight and learn to maintain that weight prior to discharge; patients who reach a healthy body weight but are discharged before this learning occurs are likely to immediately decrease their caloric intake to excessively low levels that are often insufficient to sustain their healthy body weight. Assisting patients in determining and practicing appropriate food intake at a healthy body weight is likely to decrease the chances of their relapsing. Patients who are medically stabilized on acute medical units will still require inpatient treatment for eating disorders if they do not meet biopsychosocial criteria for partial hospitalization programs or if no suitable partial hospitalization program for eating disorders treatment is accessible because of geographic or other reasons. Patients with inadequate motivation or support who are discharged from inpatient to partial hospitalization programs before they are clinically ready often have high rates of early relapse, greater struggles with recovery, and slower rates of progress, necessitating longer future inpatient stays.

In shifting between levels of care, it is important to establish continuity of care. Stepping down from one level of care to a less intensive level may be destabilizing for a patient and can be even more so when this involves a change in physician, therapist, or treatment team. At times, patients may erroneously conclude that moving to a less restrictive treatment setting means that they are suddenly fully improved. The patient's ability to continue treatment with familiar and trusted staff in a partial hospitalization or outpatient setting may contribute to the success of aftercare planning. Consequently, if the patient is moving from one treatment setting or locale to another, transition planning requires that the care team in the new setting or locale be identified and that specific patient appointments be made. It is preferable that a specific clinician on the team be designated as the primary coordinator of care to ensure continuity and attention to important aspects of treatment. Guidelines for treatment settings are provided in Table 8.

Although most patients with uncomplicated bulimia nervosa do not require hospitalization, indications for hospitalization can include severe disabling symptoms that have not responded to adequate trials of outpatient treatment, serious concurrent general medical problems (e.g., metabolic abnormalities, hematemesis, vital sign changes, uncontrolled vomiting), suicidality, psychiatric disturbances that would warrant patients' hospitalization independent of the eating disorder diagnosis, or severe concurrent substance use.

Legal interventions, including involuntary hospitalization and legal guardianship, may be necessary to address the safety of patients who are reluctant to receive treatment but whose general medical conditions are life threatening (102). On a short-term basis at least, outcomes for those patients who are hospitalized involuntarily are comparable with outcomes of those hospitalized voluntarily with respect to rates of weight restoration (103). The decision to hospitalize on a psychiatric versus a general medical or adolescent/pediatric unit depends on the patient's general medical status, the skills and abilities of local psychiatric and general medical staff, and the availability of suitable programs to care for the patient's general medical and psychiatric problems (104). There is evidence to suggest that patients treated in specialized inpatient eating disorder units have better outcomes than patients treated in general inpatient settings where staff lack expertise and experience in treating patients with eating disorders (105).

Partial hospitalization and day hospital programs are being increasingly used in attempts to decrease the length of inpatient stays or even in lieu of hospitalization for individuals with milder symptoms. However, such programs may not be appropriate for patients with lower initial weights. The failure of outpatient treatment is one of the most frequent indications for the more intensive treatment provided in a day, partial hospitalization, or inpatient program. In deciding whether a patient requires a partial hospitalization program, the patient's motivation to participate in treatment and ability to work in a group setting should be considered (106, 107). A growing body of evidence suggests that partial hospitalization outcomes are highly correlated with treatment intensity and that more successful programs involve patients at least 5 days/week for 8 hours/day (101).

Patients who are considerably below their healthy body weight but who are highly motivated to adhere to treatment, have cooperative families, and have brief symptom duration may benefit from treatment in outpatient settings, but only if they are carefully monitored and if they and their families understand that a more restrictive setting may be necessary if persistent progress is not evident within a few weeks (108, 109). Careful monitoring includes at least weekly (and often twice or thrice weekly) weight determinations done immediately after the patient voids and while the patient is wearing the same class of garment (e.g., hospital gown, standard exercise clothing). Measurement of urine specific gravity, orthostatic vital signs, oral body temperature, and, in purging patients, electrolytes may also need to be monitored on a regular basis. Although child and adolescent patients treated in the outpatient setting can remain with their families and continue to attend school or work, these advantages must be balanced against the risks of failure to progress in recovery.

Anorexia nervosa is a complex, serious, and often chronic condition that may require a variety of treatment modalities at different stages of illness and recovery. Specific treatments include nutritional rehabilitation, psychosocial interventions, and medications. The aims of treatment are to 1) restore the patient to a healthy weight (associated with the return of menses and normal ovulation in female patients, normal sexual drive and hormone levels in male patients, and normal physical and sexual development in children and adolescents); 2) treat the patient's physical complications; 3) enhance the patient's motivation to cooperate in the restoration of healthy eating patterns and participate in treatment; 4) educate the patient regarding healthy nutrition and eating patterns; 5) help the patient reassess and change core dysfunctional cognitions, attitudes, motives, conflicts, and feelings related to the eating disorder; 6) treat the patient's associated psychiatric conditions, including deficits in mood and impulse regulation, self-esteem, and behavior; 7) enlist family support and provide family counseling and therapy where appropriate; and 8) prevent the patient from relapsing.

The goals of nutritional rehabilitation for seriously underweight patients are to restore weight, normalize eating patterns, achieve normal perceptions of hunger and satiety, and correct biological and psychological sequelae of malnutrition (110, 111).

Healthy target weights should be established as part of the initial treatment plan and discussed explicitly with the patient, but with considerable sensitivity to how generally fearful patients are of gaining weight. On occasion it may be judicious to delay this discussion until the patient is less likely to be terrified of his or her ultimate weight goal. In general, a healthy goal weight for female patients is the weight at which normal menstruation and ovulation are restored and, for male patients, the weight at which normal testicular function is resumed. For female patients who previously had a healthy menses and ovulation, the clinician can estimate their healthy weight as approximately the same weight at which full physical and psychological vigor were present. In one study of 100 adolescent patients with anorexia nervosa (94), the resumption of menses occurred at a weight approximately 4.5 pounds greater than the weight at which menses was lost; at 90% of healthy weight, 86% of patients resumed menses. In children and adolescents, growth curves should be followed and are most useful when longitudinal data are available, given that extrapolations from cross-sectional data at one point in time can be misleading. Therefore, for most clinical work, it is reasonable to simply weigh patients and gauge how far they are from their individually estimated healthy body weight (112). Bone age may be accurately estimated from wrist X-rays and nomograms. In conjunction with bone measurements, menstrual history in adolescents with secondary amenorrhea, mid-parental heights, and assessments of skeletal frame, CDC growth charts (available at http://www.cdc.gov/growthcharts/) may be used to accurately estimate individually appropriate ranges for "expected" weights for current age and to set individually realistic expectations and goals for weight and height for patients up to age 20 years.

For individuals who are markedly underweight and for children and adolescents whose growth is substantially less than that predicted by growth curves, hospital-based programs for nutritional rehabilitation should be considered. For those in inpatient or residential settings, the weight at which it is appropriate to discharge a patient may vary in relation to the patient's healthy target weight and will depend on the patient's ability to feed him- or herself, the patient's motivation and ability to participate in aftercare programs, and the adequacy of aftercare, including partial hospitalization. In general, the closer a patient is to his or her healthy body weight before discharge, the less the risk he or she has of relapsing and being readmitted. Having patients maintain their weight for a period of time before they are discharged from inpatient treatment probably decreases the risk of their relapsing as well.

Refeeding programs should be implemented in nurturing emotional contexts. Staff should convey to patients their intention to take care of them and not let them die even when the illness prevents the patients from taking care of themselves. Staff should clearly communicate that they are not seeking to engage in control battles and have no punitive intentions when using interventions that the patient may experience as aversive. Some positive reinforcements (e.g., privileges) and negative reinforcements (e.g., required bed rest, exercise restrictions, restrictions of off-unit privileges) should be built into the program; negative reinforcements can then be reduced or terminated and positive reinforcements accelerated as target weights and other goals are achieved.

As patients work to achieve their target weights, their treatment plan should also establish expected rates of controlled weight gain. Clinical consensus suggests that realistic targets are 2–3 lb/week for hospitalized patients and 0.5–1 lb/week for individuals in outpatient programs, although an intensive partial hospitalization, stepped-down program has reported gains of up to 2 lb/week (113). Occasionally some patients may gain as much as 4–5 lb/week, but these individuals must be carefully monitored for refeeding syndrome and fluid retention. Dietitians can help patients choose their own meals and provide a structured meal plan that ensures nutritional adequacy and inclusion of all the major food groups. Formula feeding may have to be added to achieve large caloric intake. Some authorities advocate that the amount of solid food eaten should not exceed the amount that patients would ordinarily be eating at their target weight. Expanding cuisine options is important to avoid the severely restricted food choices frequently seen in eating disorder patients. Legitimate food allergies and patients' religious and cultural practices must be considered and discussed to limit patient rationalizations for restricted eating. Intake levels should usually start at 30–40 kcal/kg per day (approximately 1,000–1,600 kcal/day). During the weight gain phase, intake may have to be advanced progressively to as high as 70–100 kcal/kg per day for some patients; many male patients require a very large number of calories to gain weight. Patients who require significantly higher caloric intakes may be discarding food, vomiting, or exercising frequently or they may engage in more nonexercise motor activity such as fidgeting; others may have a truly elevated metabolic rate. Patients requiring much lower caloric intakes or those suspected of artificially increasing their weight by fluid loading should be weighed in the morning after voiding while they are wearing only a gown; their fluid intake also should be carefully monitored. Assessing urine specimens obtained at the time of weigh-in for specific gravity may help ascertain the extent to which the measured weight reflects excessive water intake.

Particularly in residential or hospital treatment programs, it may initially be difficult to obtain the cooperation of patients who do not wish to be there. In addition, many patients have delayed gastric emptying that initially impairs their ability to tolerate 1,000 calories/day. Under such circumstances, it is often more effective to begin with 200–300 calories above the patient's usual caloric intake (e.g., a patient consuming 400 calories/day may need to start at 600–700 calories/day). During hospitalization, giving patients a liquid feeding formula in the early stages of weight gain and then gradually exposing them to food and slowly increasing their activity level can be a very effective strategy for inducing weight gain (114). As patients are able and as their cooperation improves, a 2–3 lb/week gain in residential or hospital programs can be expected without compromising the patients' safety.

In addition to an increased caloric intake, patients also benefit from vitamin and mineral supplements. Serum potassium levels should be regularly monitored in patients who are persistent vomiters. Hypokalemia should be treated with oral or intravenous potassium supplementation and rehydration.

Physical activity should be adapted to the food intake and energy expenditure of the patient, taking into account bone mineral density and cardiac function. For the severely underweight patient, exercise should be restricted and always carefully supervised and monitored. Once a safe weight is achieved, the focus of an exercise program should be on physical fitness as opposed to expending calories. The focus on fitness should be balanced with restoring patients' positive relationship with their bodies—helping them to take back control and get pleasure from physical activities rather than being compulsively enslaved to them. An exercise program should involve exercises that are not solitary, are enjoyable, and have endpoints that are not determined by time spent expending calories or changing weight and shape. Sports such as soccer, basketball, volleyball, or tennis are examples (115).

Staff should help patients deal with their concerns about weight gain and body image changes, given that these are particularly difficult adjustments for patients to make. In fact, there is general agreement among clinicians that distorted attitudes about weight and shape are the least likely to change and that excessive and compulsive exercise may be one of the last of the behaviors associated with an eating disorder to abate. Although it is by no means certain that patients' abnormal eating habits will improve simply as a function of weight gain (116), there is considerable evidence to suggest that other eating disorder symptoms diminish as weight is restored with nutritional rehabilitation. For example, clinical experience indicates that with weight restoration, food choices increase, food hoarding decreases, and obsessions about food decrease in frequency and intensity, although they do not necessarily disappear.

Providing anorexia nervosa patients who have associated binge eating and purging behaviors with regular structured meal plans may also enable them to improve. For some patients, however, giving up severe dietary restrictions and restraints appears to increase binge-eating behavior, which is often accompanied by compensatory purging.

As weight is regained, changes in associated mood and anxiety symptoms as well as in physical status can be expected (117). Clinicians should advise patients of what changes they can anticipate as they start to regain weight. In the initial stages, the apathy and lethargy associated with malnourishment may abate. However, as patients start to recover and feel their bodies becoming larger, and especially as they approach frightening magical numbers on the scale that represent phobic weights, they may experience a resurgence of anxious and depressive symptoms, irritability, and sometimes suicidal thoughts. These mood symptoms, non-food-related obsessional thoughts, and compulsive behaviors, although often not eradicated, usually decrease with sustained weight gain.

Weight gains result in improvement in most of the physiological complications of semistarvation, including improvement in electrolyte levels, heart and kidney function, and attention and concentration. Initial refeeding may be associated with mild transient fluid retention, and patients who abruptly stop taking laxatives or diuretics may experience marked rebound fluid retention for several weeks, presumably from salt and water retention caused by elevated aldosterone levels associated with chronic dehydration. Refeeding edema and bloating occur frequently.

Patients may experience abdominal pain and bloating with meals from the delayed gastric emptying that accompanies malnutrition. Constipation, which may be ameliorated with stool softeners, can progress to obstipation and, rarely, acute bowel obstruction. As weight gain progresses, many patients also develop acne and breast tenderness. Many patients become unhappy and demoralized about resulting changes in body shape. Management strategies for dealing with these milder adverse effects include careful refeeding, frequent physical examinations, and forewarnings to patients about mild refeeding edema.

A severe refeeding syndrome may occur when severely malnourished patients (generally those weighing <70% of their healthy body weight) are re-fed too rapidly, particularly in the context of enteral or parenteral feedings but also with vigorous oral refeeding regimens. This syndrome consists of hypophosphatemia, hypomagnesemia, hypocalcemia, and fluid retention. Thiamine deficiency may also be seen as a feature of this syndrome. In some case series, the refeeding syndrome has been reported to occur in roughly 6% of hospitalized adolescents (118). Excessively rapid refeeding and nasogastric or parenteral feeding may be particularly dangerous because of their potential for inducing severe fluid retention, cardiac arrhythmias, cardiac failure, respiratory insufficiency, delirium, seizures, rhabdomyolysis, red cell dysfunction, and even sudden death, especially in the lowest-weight patients (118, 119). In such cases, phosphorus, magnesium, and/or potassium supplementation will be necessary (118, 120). In one series of hospitalized adolescents, moderate hypophosphatemia occurred in 5.8% and mild hypophosphatemia in 21.7% of patients, requiring some degree of phosphorus replacement in 27.5% of these patients (120).

Besides monitoring of mineral and electrolyte levels, general medical monitoring during refeeding should include assessment of vital signs, monitoring of food and fluid intake and output (if indicated), and observation for edema, rapid weight gain (associated primarily with fluid overload), congestive heart failure, and gastrointestinal symptoms. For children and adolescents who are severely malnourished (weigh <70% of their standard body weight), cardiac monitoring, especially at night, may be advisable (120).

Some patients are completely unable to recognize their illness, accept the need for treatment, or tolerate the guilt that would accompany eating, even when performed to sustain their lives. On the rare occasions when staff have to take over the responsibilities for providing life-preserving care, nasogastric feedings are preferable to intravenous feedings. In some programs, supplemental overnight pediatric nasogastric tube feedings have been used to facilitate weight gain in cooperative patients. This practice is not routinely recommended at present, although it appears to be well tolerated, may slightly decrease hospital stays in children, and may be experienced positively by some patients, particularly younger patients, who may feel relieved to know that they are being cared for and who, while they cannot bring themselves to eat, are willing to allow physicians to feed them (121). If used, such interventions should never supplant expectations that the patient will resume normal eating patterns on his or her own. Total parenteral feeding is required only rarely and for brief periods in life-threatening situations.

Forced nasogastric or parenteral feeding can each be accompanied by substantial dangers. When nasogastric feeding is necessary, clinical experience suggests that continuous feeding (i.e., over 24 hours) may be less likely than three to four bolus feedings a day to result in metabolic abnormalities or patient discomfort and may be better tolerated by patients. As an alternative to nasogastric feedings, in very difficult situations where patients physically resist and constantly remove their nasogastric tubes, gastrostomy or jejunostomy tubes may be surgically inserted. As described above, rapid refeeding can be associated with the severe refeeding syndrome, and infection is always a risk with parenteral feedings in emaciated and potentially immunocompromised patients with anorexia nervosa. Consequently, these interventions should not be used routinely but should be considered only when patients are unwilling or unable to cooperate with oral feedings or when the patients' health, physical safety, and recovery are being threatened. In situations where involuntary forced feeding is considered, careful thought should be given to clinical circumstances, family opinion, and relevant legal and ethical dimensions of the patient's treatment.

If using interventions that patients with anorexia nervosa may experience as coercive, the clinician should consider the potential impact on the therapeutic relationship, especially since maintaining a sense of control is often a key dynamic in these patients. The setting of limits is developmentally appropriate in the management of children and adolescents and may help shape the patient's behavior in a healthy direction. It is essential for caregivers to be clear about their own intentions and empathic capacities regarding the patient's impression of being coerced. Caregivers should not be seen as using techniques intended to be coercive. Rather, caregivers' interventions should always be clearly seen as components of a general medical treatment required for the patient's health and survival.

During the last few years, there has been considerable debate about the ethics of involuntarily feeding patients with anorexia nervosa (122, 123). There is general agreement that children and adolescents who are severely malnourished and in grave medical danger should be re-fed, involuntarily if necessary, but that every effort should be made to gain their cooperation as cognitive function improves.

Ethical as well as clinical dilemmas often confront clinicians dealing with adult patients with chronic anorexia nervosa and their families. The general principles to be followed are those directing good, humane care; respecting the wishes of competent patients; and intervening respectfully with patients whose judgment is severely impaired by their psychiatric disorders when such interventions are likely to have beneficial results (124, 125).

The goals of psychosocial interventions in patients with anorexia nervosa are to help them 1) understand and cooperate with their nutritional and physical rehabilitation, 2) understand and change the behaviors and dysfunctional attitudes related to their eating disorder, 3) improve their interpersonal and social functioning, and 4) address comorbid psychopathology and psychological conflicts that reinforce or cause them to maintain eating disorder behaviors. Efforts to achieve these goals often benefit from an initial enhancement of a patient's motivation to change along with ongoing efforts to sustain this motivation.

Few controlled studies offer guidance for the psychosocial treatment of anorexia nervosa. Clinical consensus suggests that during the acute refeeding and weight gain stages, it is beneficial to provide patients with individual psychotherapeutic management that is psychodynamically informed and that provides empathic understanding, explanations, praise for positive efforts, coaching, support, encouragement, and other positive behavioral reinforcement. During all phases of treatment, seeing patients' families is also helpful, particularly for children and adolescents, for whom controlled trials suggest that family treatment is the most effective intervention (86, 126). For patients who initially lack motivation, their awareness and desire for recovery may be increased by psychotherapeutic techniques based on motivational enhancement, although solid evidence for this contention is lacking.

At the same time, clinical consensus suggests that psychotherapy alone is generally not sufficient to treat severely malnourished patients with anorexia nervosa. Although the value of establishing and maintaining a psychotherapeutically informed relationship is clearly beneficial and psychotherapeutic sessions to enhance patient motivation and further patient weight gain are likely to be helpful, the value of formal psychotherapy during the acute refeeding stage is uncertain (127). Attempts to conduct formal psychotherapy may be ineffective with starving patients, who are often negativistic, obsessional, or mildly cognitively impaired, presumably in relation to the known cortical atrophy seen in nutritionally compromised patients. One study documented the difficulty researchers have had in initiating and sustaining cognitive-behavioral therapies for patients with anorexia nervosa (128).

Most nutritional rehabilitation programs incorporate emotional nurturance and one of a variety of behavioral interventions that link exercise, bed rest, and privileges with target weights, desired behaviors, and informational feedback. Several studies of individual therapy have shown modest success, sometimes in only a small percentage of patients (7, 85). In one controlled trial, nonspecific supportive clinical management appeared to be at least as effective as CBT or IPT in some patients. However, 70% of patients either did not complete or made only small gains from the active psychotherapies they received (7). In this study, clinical management included education, care, support, and the fostering of a therapeutic relationship designed to promote adherence to treatment through the use of praise, reassurance, and advice.

The accumulated evidence strongly supports the value of family therapy for the acute treatment of children and adolescents in outpatient settings. Studies show that whether patients and parents are seen together or are treated separately in ongoing treatment, the results are better than when families are not involved at all (86, 126). This approach begins with the therapist's attempting to unite the parents in developing a consistent approach to refeeding, sympathizing with their plight, and explicitly disclaiming the notion that the parents have caused the eating problem. When families are involved in treatment, sibling subsystems can be engaged to support the affected sibling. Parents can determine for themselves how best to refeed their child with anorexia nervosa with the therapist's ongoing support and consultation. For some outpatients, a short-term course of family therapy may be as effective as a long-term course; however, a shorter course of therapy may not be adequate for patients with severe obsessive-compulsive features or nonintact families (129). In these studies (129), inpatient care was used briefly for medical stabilization. For adolescents treated in inpatient settings, participation in family group psychoeducation may help promote weight gain and may be as effective as more intensive forms of family therapy (130).

Clinical consensus suggests that psychotherapy can be helpful for patients with anorexia nervosa once their malnutrition has been corrected and they have begun gaining weight (131). Because of the enduring nature of many of the psychopathological features of anorexia nervosa and the patient's need for support during recovery, ongoing individual psychotherapeutic treatment is frequently required for at least 1 year and may take many years (132, 133).

Although there have been few formal studies of its effectiveness (134, 135), psychotherapy is generally thought to help patients understand 1) what they have been through; 2) developmental, familial, and cultural antecedents of their illness; 3) how their illness may have been a maladaptive attempt to cope and emotionally self-regulate; 4) how to avoid or minimize the risk of relapse; and 5) how to better deal with salient developmental and other important life issues in the future. At present there is no absolute weight or percentage of body fat that indicates when a patient is actually ready to begin formal psychotherapy. In addition, patients often display improved mood, enhanced cognitive functioning, and clearer thought processes once their nutritional status has significantly improved and even before they make substantial weight gains.

Little evidence from controlled studies exists to guide clinicians in the use of specific therapies for adults with anorexia nervosa. Nonetheless, some data are emerging in support of individual CBT (136–138) for helping patients maintain healthy eating behaviors and CBT or IPT for inducing cognitive restructuring and promoting more effective coping (139, 140). After a patient has begun to gain weight, CBT may be helpful in reducing the risk of relapse and improving outcome, as demonstrated in a small randomized controlled trial (136). In that study, patients who received CBT were more likely to remain in treatment (78%) and have a good outcome after a year (44%) than those assigned to nutritional counseling (7%).

Many clinicians also use psychodynamically oriented individual or group psychotherapy to address underlying personality disorders that may hamper treatment and help sustain the illness and to foster psychological insight and maturation in patients who have made strides toward weight restoration (141–148). Clinical consensus suggests that psychosocial interventions should incorporate an understanding of the patient's developmental traumas, cognitive development, psychodynamic conflict and defense styles, disorders of self-esteem, self-regulation, and "sense of self," as well as other psychological deficits, the presence of other psychiatric disorders, and the complexity of family relationships (149–152). Although studies of psychotherapies focus on different interventions as distinctly separate treatments, in practice there is frequent overlap among treatments. Indeed, most experienced clinicians report using interventions that cross theoretical boundaries when treating patients with eating disorders (153).

In adolescents, controlled studies have shown that for patients who are younger than age 19 years, have been ill for 3 years or less, and have restored their weight, family therapy is more beneficial than individual therapy, whereas individual therapy is more beneficial for patients with later-onset disorders (154). At 5-year follow-up of patients who received these therapies, much of the improvement could be attributed to the natural outcome of the illness, but it was still possible to detect long-term benefits of the psychological therapies (155).

Regardless of the clinical or theoretical approach used in treatment, some patients with eating disorders challenge clinicians' understanding and in some instances provoke countertransference reactions, particularly in response to patients' communications of aggression and defiance (140, 147, 156–162). Clinical consensus suggests that eating disorders are often difficult to ameliorate with short-term interventions, at least in older adolescents and adults; for this reason, clinicians often feel they have not done enough to change the patient's plight. Countertransference feelings often include beleaguerment, demoralization, and excessive need to change patients with a chronic eating disorder. At the same time, when treating patients with chronic illnesses, clinicians need to understand the longitudinal course of the disorder and that patients can recover even after many years of symptoms. Such awareness may help clinicians maintain a degree of therapeutic optimism and deal with the feelings of pessimistic demoralization that may arise (13, 163).

Some observations suggest that the gender of the clinician may play a role in the particular kind of countertransference reactions that come into play (156, 157). A patient's concerns about the gender of a clinician may be tied to concerns about potential boundary violations and should be attended to when selecting clinicians, including psychiatrists (164, 165). In addition, cultural differences between patients and clinicians or patients and other aspects of the care system may also influence the course and conduct of treatment and require attention. Ongoing processing of one's countertransference reactions, sometimes with the help of a supervisor or consultant, can be useful in helping the clinician persevere and reconcile intense, troublesome countertransference reactions.

When a patient with an eating disorder has been sexually abused or has felt helpless in other situations of boundary violations, this may stir up needs in the clinician to rescue the patient, which can occasionally result in a loosening of the therapeutic structure, the loss of therapeutic boundary keeping, and a sexualized countertransference reaction. In some cases, these countertransference responses have led to overt sexual acting out and unethical treatment on the part of the clinician that have not only compromised treatment but also severely harmed the patient (166). The maintenance of clear boundaries is critical in treating all patients with eating disorders, not only those who have been sexually abused but also those who may have experienced other types of boundary intrusions regarding their bodies, eating behaviors, and other aspects of the self by family members or others. Regular meetings with other team members and/or formal supervision can also help clinicians avoid boundary violations with eating disorder patients. Particularly with some adolescents, a clinician's obvious warmth and direct educational approach may facilitate initiating and sustaining the patient's trust. However, the license to be informal may create a climate in which a clinician is at a greater risk to violate therapeutic boundaries; such an occurrence must be consistently and carefully prevented. At the same time, according to some clinicians, a clinician's excessively rigid, cold manner and formal distancing behaviors (e.g., avoiding even benign pats on the shoulder that a patient might seek for reassurance) may be disconcerting to some patients and inhibit them from fully engaging in treatment.

Some clinicians use group psychotherapy as an adjunctive treatment for anorexia nervosa; in such cases, however, caution must be taken that patients do not compete to be the thinnest or sickest patient in the group or become excessively demoralized by observing the ongoing struggles of other patients in the group. For that reason, clinicians sometimes prefer heterogeneous groups that combine patients with bulimia nervosa and those with anorexia nervosa. Although there has been little formal study of group psychotherapy in the treatment of anorexia nervosa, one naturalistic study suggests that CBT may have promise (137).

Some clinicians consider that eating disorders may be usefully treated through addiction models, but no data from short- or long-term outcome studies using these methods have been reported. Literature from Anorexics and Bulimics Anonymous and Overeaters Anonymous emphasizes that these programs are not substitutes for professional treatment. These organizations specifically recommend that members seek appropriate medical and nutritional guidance. Nevertheless, there are concerns about zealous and narrow application of the 12-step philosophy in addiction-oriented programs for eating disorders. Programs that focus exclusively on abstaining from binge eating, purging, restrictive eating, and exercise (e.g., 12-step programs) without attending to nutritional considerations or cognitive and behavioral deficits have not been studied and therefore cannot be recommended as the sole treatment for anorexia nervosa. Clinicians frequently report encountering patients who, while attempting to resolve anorexia nervosa by means of a 12-step program alone, might have been greatly helped by concurrent conventional treatment approaches such as nutritional counseling and rehabilitation, medications, and psychodynamic or cognitive-behavioral approaches. By limiting their attempts to recover to their participation in a 12-step program alone, patients not only deprive themselves of the potential benefits of conventional treatments but also may expose themselves to misinformation about nutrition and eating disorders offered by well-intended nonprofessionals participating in and sometimes running these groups. Attempts have been made to integrate traditional and 12-step approaches into treatment; such approaches can offer a strong sense of community, but the effectiveness and potential adverse effects of these combined interventions have not been systematically studied (167). It is important for programs using a 12-step model to be equipped to care for patients with the substantial psychiatric and general medical problems often associated with eating disorders.

Selective support groups led by professionals and advocacy organizations may be beneficial as adjuncts to other psychosocial treatment modalities. However, clinicians should remain cognizant of the idiosyncratic recommendations made in some self-help groups. Sometimes, participants or leaders will eschew clinician-recommended treatments such as psychotropic medication or insist that a participant follow a particular kind of meal structure. These recommendations may conflict with other treatment recommendations and potentially increase the patient's resistance to treatment.

Patients and their families are increasingly using online web sites, news groups, and chat rooms as resources. Although substantial amounts of worthwhile information and support are available in this fashion, the lack of professional supervision of these sources may sometimes lead to misinformation and unhealthy dynamics among users. Clinicians should ask patients about their use of electronic support and other alternative and complementary approaches and be prepared to openly and sympathetically discuss the information and ideas they and their families have gathered from these sources.

As with any form of intervention, various psychosocial interventions may generate adverse effects; however, these have not been systematically studied with regard to treating anorexia nervosa. Some that have been observed by clinicians, patients, and families include 1) fostering negative attitudes in patients and/or families toward health care professionals without adequate discussion and reflection, thereby increasing the risk that patients will drop out of treatment and become less willing to seek or engage in professional treatment; 2) delaying referral to more appropriate interventions; and 3) generating burdensome costs without reasonable or expected benefits (13, 168).

Patients often have difficulty with certain elements of psychotherapy. For example, among patients receiving CBT, some are quite resistant to self-monitoring, whereas others have difficulty mastering cognitive restructuring. Most patients are initially resistant to changing their eating behaviors, particularly when it comes to increasing their caloric intake or reducing exercise. However, the complete lack of acceptance of a psychotherapeutic approach appears to be rare, although this has not been systematically studied.

Management strategies to deal with potential negative effects of psychotherapeutic interventions include 1) conducting a careful pretreatment evaluation, during which the therapist must assess and enhance the patient's level of motivation for change and determine the most appropriate therapeutic approach and format (e.g., individual versus group); 2) being alert to a patient's reactions to and attitudes about the proposed treatment and listening to and discussing the patient's concerns in a supportive fashion; 3) ongoing monitoring of the quality of the therapeutic relationship; and 4) identifying patients for whom another treatment should be co-administered or given before psychotherapy begins (e.g., substance use disorder treatment for those actively abusing alcohol or other drugs, antidepressant treatment for patients whose depression makes them unable to become actively involved, more intensive psychotherapy for those with severe personality disorders, group therapy for those not previously participating). Alternative strategies may be necessary to facilitate the therapeutic process and prevent the abrupt termination of therapy (13). As with all therapeutic interventions, it is essential that the therapist be alert to potential countertransference phenomena toward these often difficult-to-treat patients. If unresolved, these reactions have a high potential for disrupting or hastening the termination of treatment.

Available studies of patients with chronic anorexia nervosa typically show a lack of substantial clinical response to psychotherapy. For example, in the study of Dare et al. (85), 84 patients, ill for an average of 6.3 years and with an initial average BMI of 15.4 kg/m², were assigned to one of three individual psychotherapies or a control group. The results after 1 year of psychotherapy were modest in all groups, although the psychoanalytic psychotherapy and family therapy groups fared better than those in the low-contact, routine-care control group. Nevertheless, many clinicians report seeing patients with chronic anorexia nervosa who, after many years of struggling with their disorder, experience substantial remission; thus clinicians are justified in maintaining and extending some degree of hope to patients and families.

For patients whose anorexia nervosa continues to be resistant to treatment despite substantial trials of nutritional rehabilitation, medications, and hospitalizations, more extensive psychotherapeutic measures may be undertaken in a further effort to engage and help motivate them or, failing that, as compassionate care. This difficult-to-treat subgroup may represent an as-yet poorly understood group of patients with malignant, chronic anorexia nervosa. Efforts to understand the unique plight of such patients may sometimes lead to engagement in the therapeutic alliance, thereby allowing the nutritional protocol to be initiated (125, 141, 142, 169, 170). With patients who have difficulty talking about their problems, clinicians have reported a variety of nonverbal therapeutic methods, such as creative arts and movement therapy programs, to be useful (171), but these methods have not been formally studied. At various stages of recovery, occupational therapy programs may also enhance self-concept and self-efficacy (172, 173), but again these programs have not been formally studied.

Although psychotropic medications should not be used as the sole or primary treatment for anorexia nervosa, they have been used as an adjunct treatment when nutritional rehabilitation programs alone are ineffective in restoring patients' normal weight or when patients demonstrate significant comorbid psychopathology such as disabling obsessive-compulsive, depressive, or anxiety symptoms. However, because anorexia nervosa symptoms and associated features such as depression may remit with weight gain, decisions concerning the use of medications should be deferred if possible until patients' weight has been restored. The decisions about whether to use psychotropic medications and which medications to choose will be determined by the remaining symptom picture (e.g., antidepressants are usually considered for those with persistent depression, anxiety, or obsessive-compulsive symptoms and for bulimic symptoms in weight-restored patients; second-generation antipsychotics are usually considered for those with severe, unremitting resistance to gaining weight, severe obsessional thinking, and denial that assumes delusional proportions). Many patients with anorexia nervosa are extremely reluctant to take medications and often refuse those that they know to specifically affect weight. These issues must be discussed sympathetically and comprehensively with patients and, for children and adolescents, with their families.

The efficacy of SSRI antidepressants for anorexia nervosa appears to vary with the phase of treatment. On the basis of several studies, fluoxetine does not appear to confer significant benefits during weight restoration (174, 175), nor did citalopram increase the rate of weight gain in a small study (176). In contrast, in weight-restored patients, fluoxetine in dosages of up to 60 mg/day may decrease relapse episodes and has been associated with better maintenance of weight and fewer symptoms of depression (177). However, for weight-restored patients with anorexia nervosa who are receiving CBT to help prevent relapse, adding fluoxetine to their treatment does not further decrease the risk of relapse (138).

Although higher dosages of fluoxetine have been found to impair appetite and cause weight loss in normal-weight and obese patients, this effect has not been reported in anorexia nervosa patients treated with lower dosages. Many clinicians report that malnourished depressed patients are less responsive to the beneficial effects of tricyclics, SSRIs, and other antidepressant medications than normal-weight depressed patients. These findings are consistent with those showing that SSRIs are not as effective for depression, when patients without an eating disorder undergo dietary restrictions (178, 179).

Malnourished patients are also much more prone to the side effects of medications. For example, the use of tricyclic antidepressants may be associated with greater risk of hypotension, increased cardiac conduction times, and arrhythmia, particularly in purging patients whose hydration may be inadequate and whose cardiac status may be nutritionally compromised. Given the availability of other antidepressant treatments, tricyclic antidepressants should be avoided, particularly in underweight patients and in patients who are at risk for suicide. In patients for whom there is a concern regarding potential cardiovascular effects of medication, medical specialty consultation can help evaluate the patient's status and advise on the use of medication. With all antidepressants, strategies to manage side effects include limiting the use of medications to patients with persistent depression, anxiety, or obsessive-compulsive symptoms; using low initial doses in underweight patients; and remaining vigilant about early manifestations of side effects.

Several other antidepressants have also been associated with significant side effects that are of relevance to the treatment of anorexia nervosa patients. Bupropion has been associated with an increased likelihood of seizures in patients with bulimia nervosa (180, 181); although the reason for this is unknown, it is suspected that patients with anorexia nervosa, binge-purge type, may also be at increased risk for seizures. Thus, this medication is not recommended for patients with anorexia nervosa, particularly those who purge. Mirtazapine, an antidepressant associated with weight gain, has also been associated with neutropenia. In addition, the only published case report of using mirtazapine to treat anorexia nervosa described a patient also taking fluvoxamine who developed the serotonin syndrome (182). Thus, mirtazapine may not be suitable for use in underweight anorexia nervosa patients.

Clinicians must attend to the black box warnings concerning antidepressants and conduct appropriate informed consent with patients and families if antidepressants are to be prescribed (183–189).

It has been suggested that antipsychotic medications, particularly second-generation antipsychotics, can be potentially useful during the weight-restoration phase or in treatment of other associated symptoms of anorexia nervosa, such as marked obsessionality, anxiety, limited insight, and psychotic-like thinking. Although no controlled studies have been reported in patients with anorexia nervosa, controlled trials of olanzapine and risperidone are under way. Evidence from case reports, case series, and open-label uncontrolled trials suggests that the second-generation antipsychotic olanzapine may promote weight gain in adults and in adolescent patients (190–193) and that olanzapine (190–194) and quetiapine may improve other associated symptoms (195–197). A small open-label study of low-dose haloperidol also showed improved insight and weight gain in severely ill patients (198). The quality of the available evidence on using antipsychotic medications is also limited by the fact that studies rarely include male patients and have included only small numbers of adolescents; in addition, only case reports are available regarding prepubertal children. If antipsychotic medications are used, the possibility of extrapyramidal symptoms, especially in debilitated patients, should be considered and routinely assessed. Also, appropriate attention must be given to the potential adverse impact of these medications on insulin sensitivity, lipid metabolism, and length of QTc interval.

Other somatic treatments, ranging from vitamin and hormone treatments to ECT, have been tried in uncontrolled studies. None has been shown to have specific value in the treatment of anorexia nervosa symptoms (199).

Other medications have been used to address associated features of anorexia nervosa. For example, antianxiety agents have been used selectively before meals to reduce anticipatory anxiety concerning eating (200, 201), and pro-motility agents, such as metoclopramide, are commonly offered for the bloating and abdominal pains that result from gastroparesis and that contribute to the premature satiety seen in some patients. However, before prescribing metoclopramide, clinicians should consider the fact that extrapyramidal symptoms are more likely to be seen in underweight anorexia nervosa patients.

In anorexia nervosa patients with prolonged amenorrhea, hormone replacement therapy (HRT) is frequently prescribed to improve patients' bone mineral density. However, no good supporting evidence exists to demonstrate the efficacy of this treatment (202, 203). In women with anorexia nervosa, the evidence supporting the use of HRT is marginal at best. HRT has not been demonstrated to increase bone mineral density over and above standard treatment in adults (204) or in adolescents (203). Only in a subset of very-low-weight women (<70% average body weight) did it prevent further bone loss (204). Estrogen can cause the fusion of the epiphyses and should not be administered to girls before their growth is completed (3). HRT usually induces monthly menstrual bleeding, obscuring the major sign that indicates weight normalization in women. This, in turn, may cause the patient to misunderstand that her body is functioning normally and therefore contribute to denial of the need to gain more weight. Clinicians stress that efforts should be made to allow patients to increase their weight and achieve resumption of normal menses before they are offered estrogen (205). There is no indication for the use of biphosphonates such as alendronate in patients with anorexia nervosa. In fact, long-term use of alendronate may oversuppress bone turnover (206). Thus, the recommended treatment for low bone mineral density includes weight gain and calcium with vitamin D supplementation (207).

The aims of treatment for patients with bulimia nervosa are to 1) reduce and, where possible, eliminate binge eating and purging; 2) treat physical complications of bulimia nervosa; 3) enhance the patient's motivation to cooperate in the restoration of healthy eating patterns and participate in treatment; 4) provide education regarding healthy nutrition and eating patterns; 5) help the patient reassess and change core dysfunctional thoughts, attitudes, motives, conflicts, and feelings related to the eating disorder; 6) treat associated psychiatric conditions, including deficits in mood and impulse regulation, self-esteem, and behavior; 7) enlist family support and provide family counseling and therapy where appropriate; and 8) prevent relapse.

Bulimia nervosa is associated with nutritional chaos characterized by alternating cycles of dietary restriction, bingeing, and purging. A primary focus for nutritional rehabilitation is to help patients develop a structured meal plan that will allow them to reduce the episodes of dietary restriction and the urge to binge and purge. Nutritional intake should be sufficient to promote satiety. Because most bulimia nervosa patients who have been studied are of normal weight, nutritional restoration will not be a central focus of treatment. However, normal body weight (or normal BMI) does not ensure normal body composition, nor does it ensure that nutritional intake is appropriate. In addition, even if their weight is within statistically normal ranges, many patients with bulimia nervosa weigh less than their appropriate biologically determined set points (or ranges) and may have to gain some weight to achieve physiological and emotional stability. Although many patients with bulimia nervosa report irregular menses, improvement in menstrual function has not been systematically assessed in the available outcome studies. Thus, even among patients of normal weight, nutritional counseling may be a useful adjunct to other treatment modalities in reducing behaviors related to the eating disorder, minimizing food restrictions, increasing the variety of foods eaten, and encouraging healthy but not compulsive exercise patterns (208). Those patients for whom some weight gain is indicated similarly require the establishment of a pattern of regular, non-binge meals, with attention on increasing their caloric intake and expanding macronutrient selection. Patients with bulimia nervosa who are overweight or obese have not been well studied.

The goals of psychosocial interventions for patients with bulimia nervosa vary and can include the following: reducing or eliminating binge eating and purging behaviors; improving attitudes related to the eating disorder; minimizing food restriction; increasing the variety of foods eaten; encouraging healthy but not compulsive exercise patterns; treating co-occurring conditions and clinical features associated with eating disorders; and addressing themes that may underlie eating disorder behaviors such as developmental issues, identity formation, body image concerns, self-esteem in areas outside of those related to weight and shape, sexual and aggressive difficulties, affect regulation, sex role expectations, family dysfunction, coping styles, and problem solving. Consequently, psychosocial interventions should be chosen on the basis of a comprehensive evaluation of the individual patient and take into consideration the patient's cognitive and psychological development, psychodynamic issues, cognitive style, comorbid psychopathology, and preferences as well as age and family situation.

With respect to short-term interventions for treating acute episodes of bulimia nervosa in adults, the available evidence indicates that CBT is the most efficacious. CBT may effect improvements in psychological functioning of bulimia nervosa patients as well as ameliorate binge eating and purging symptoms. For example, studies have shown that bulimia nervosa patients who improved with CBT also showed improvements in self-directedness and harm avoidance (209, 210). Among patients who do not initially respond to CBT, a small number do respond to IPT or fluoxetine (211) or other modes of treatment such as family and group psychotherapies. Some controlled trials (212) have also shown the effectiveness of IPT as an initial therapy. Behavioral techniques, such as planned meals and self-monitoring, may also be helpful for managing initial symptoms and interrupting binge-purge behaviors (213, 214). It should be pointed out that these study results may not be generalizable to typical clinical situations. For example, to maximize the "clean" experimental nature of some of the CBT/IPT controlled studies mentioned above, the CBT intentionally avoided dealing with interpersonal issues and the IPT intentionally avoided talking about eating issues, which is quite different than how these therapies are conducted in clinical practice (215). It is also possible that the narrow inclusion criteria of some studies limit the generalizability of the study results (216).

Some clinical reports indicate that psychodynamic and psychoanalytic approaches in individual or group format are useful once bingeing and purging symptoms improve (217–219). These approaches address developmental issues; identity formation; body image concerns; self-esteem; conflicts surrounding sexuality, anger, or aggression; affect regulation; gender role expectations; interpersonal conflicts; family dysfunction; coping styles; and problem solving. In a recent naturalistic study of treatment as practiced by experienced clinicians in the community, both CBT and psychodynamic psychotherapy led to decreased rates of bingeing and purging similar to those seen in controlled trials (roughly 50%). However, although CBT has been reported to be associated with a more rapid remission of eating symptoms, some therapists note that more integrative treatments that include psychodynamic approaches are useful in targeting both eating symptoms and broader personality, comorbidity, and quality-of-life issues (153).

Some bulimia nervosa patients, particularly those with concurrent personality pathology or other co-occurring disorders, may require substantially longer treatment. In one study, the clinicians reported that their average CBT treatment for bulimia nervosa lasted 69 sessions (81). However, just how closely these clinicians adhered to formal CBT methods is unknown. Also unknown is how the length and characteristics of the treatments varied according to other clinical dimensions of these patients. Exactly what is required over the long run to best help patients resolve lingering preoccupations with body image and the more subtle but impairing psychological dimensions that may be associated with eating disorders requires additional study. These concerns are often approached in practice through a variety of longer-term psychotherapies.

Family therapy should be considered whenever possible, especially for adolescents still living with parents or for older patients with ongoing conflicted interactions with parents. Patients with marital discord may benefit from couples therapy.

A variety of self-help and professionally guided self-help programs have been effective for a small number of patients with bulimia nervosa (220–222) and have been piloted in some stepped-care approaches. Several innovative online programs are currently being studied (5). Support groups and 12-step programs such as Overeaters Anonymous may be helpful as adjuncts to initial treatment of bulimia nervosa and for subsequent relapse prevention but are not recommended as the sole initial treatment approach for bulimia nervosa (168, 223). As noted above, these support organizations emphasize in their literature that their programs are not substitutes for professional treatment and specifically recommend that members seek appropriate medical and nutritional guidance. However, clinicians should remain cognizant of the idiosyncratic recommendations made in some self-help groups.

Patients with bulimia nervosa occasionally have difficulties with certain elements of psychotherapy similar to what was discussed above for patients with anorexia nervosa. Possible adverse effects of psychotherapeutic and psychosocial interventions, steps that clinicians might take to minimize negative therapeutic reactions, and issues concerning countertransference (as discussed in Section II.B.2.b) apply to the treatment of patients with bulimia nervosa.

Antidepressants are effective as one component of an initial treatment program for most bulimia nervosa patients. Although various classes of antidepressant medications can reduce symptoms of binge eating and purging, SSRIs have the most evidence for efficacy and the fewest difficulties with adverse effects (224–226). To date, the only medication approved by the FDA for the treatment of bulimia nervosa is fluoxetine. The only other SSRI shown to be effective is sertraline, which was studied in a small randomized controlled trial (227). Available studies also suggest that antidepressants may be helpful for patients with substantial symptoms of depression or anxiety, obsessions, or certain impulse disorder symptoms or for patients who have not responded or had a suboptimal response to previous attempts at appropriate psychosocial therapy (228, 229). Dosages of SSRIs that are higher than those used for depression (e.g., fluoxetine 60 mg/day) are more effective in treating bulimic symptoms (224, 226, 230), but high dropout rates may also be seen in patients using these drugs (226). A small open trial demonstrated the safety and effectiveness of 60 mg/day of fluoxetine for treating bulimia nervosa in adolescents (225). Thus, many clinicians initiate fluoxetine treatment for bulimia nervosa at the higher dosage, titrating downward if necessary to manage side effects. Tricyclic and MAOI antidepressants are rarely used to treat bulimia nervosa, but if they are used, the dosages are similar to those used to treat depression (231).

Often, several different antidepressants may have to be tried sequentially to identify the specific medication with the optimum effect in a particular patient. In the bulimia nervosa patient whose symptoms do not respond to medication, it is important to assess whether the patient has taken the medication shortly before vomiting. Correlations between serum levels and response have not been identified; however, if serum levels of the medication are available, they may help determine whether presumably effective levels of the drug have actually been achieved. Treatment adherence will also enhance the patient's response to treatment, and subtle interpersonal and psychodynamic factors in the physician-patient relationship may contribute to treatment resistance if left unaddressed (232).

As in most clinical situations, careful education of the patient regarding possible side effects of medications and their symptomatic management (e.g., stool softeners for constipation) is important. Side effects vary widely across studies depending on the type of antidepressant medication used. In the multicenter fluoxetine trials (224, 230), sexual side effects were common, and at the dosage of 60 mg/day, insomnia, nausea, and asthenia were seen in 25%–33% of patients. For the tricyclic antidepressants, common side effects include sedation, constipation, dry mouth, and, with amitriptyline, weight gain (233–238).

The toxicity and potential lethality of tricyclic antidepressant overdosage also dictate caution in prescribing this class of drug for patients who are at risk for suicide. Practitioners should also avoid prescribing MAOIs to patients with chaotic binge eating and purging behaviors. The risk of spontaneous hypertensive crises in patients with bulimia nervosa taking MAOIs is not insignificant (239). This risk and the importance of eating a tyramine-free diet while taking MAOIs should be discussed with patients for whom this type of medication is contemplated.

There are few reports on the use of antidepressant medications in the maintenance phase of treating bulimia nervosa patients. Although there are data indicating that fluoxetine can be effective in preventing relapse in these patients (226), other data suggest that high rates of relapse occur while antidepressants are being taken and possibly higher rates are seen when the medication is withdrawn (240). In the absence of more systematic data, most clinicians recommend continuing antidepressant therapy for a minimum of 9 months and probably for 1 year in most patients with bulimia nervosa.

Clinicians must attend to the black box warnings concerning antidepressants and conduct appropriate informed consent with patients and families if these medications are to be prescribed (183–189).

For patients with bulimia nervosa who require mood stabilizers, the use of lithium carbonate is problematic, because lithium levels may shift markedly with rapid volume changes. Lithium is not effective in the treatment of bulimia nervosa (241). Both lithium carbonate and valproic acid frequently lead to undesirable weight gains that may limit their acceptability to bulimia nervosa patients. Selecting a mood stabilizer that avoids these problems may result in better patient adherence and medication effectiveness. Topiramate is not an effective mood stabilizer but may be potentially useful for bulimia nervosa and binge eating disorder (242, 243). However, in contrast to the low rates of adverse effects observed in clinical trials with topiramate, practitioners have reported several patients experiencing adverse effects with the drug, such as word-finding difficulties and paresthesias in a sizable minority of patients, although these may have been related to excessively rapid rates of dosage increases (242, 243). Also of note, patients receiving topiramate for bulimia nervosa lost an average of 1.8 kilograms, so this medication may be problematic for normal- to lower-weight individuals (243). No data are available regarding the use of these medications for treating bulimia nervosa or binge eating in children or adolescents, but safety and tolerability data have been reported for children and adolescents with other disorders for which lithium (244), valproic acid (245), and topiramate (246) have been prescribed.

Several case reports indicate that methylphenidate may be helpful for bulimia nervosa patients with concurrent ADHD (247–249). In these situations, particular attention should be given to a range of potential adverse effects, including abuse.

Although not all psychotherapies have been well studied, there is general consensus among clinicians regarding the efficacy of a combined psychotherapeutic/medication approach; such an approach is worth considering when initiating treatment. In some research, the combination of antidepressant therapy and CBT has resulted in the highest remission rates of bulimia nervosa (250–252). Other studies suggest that target symptoms such as binge eating and purging and attitudes related to the eating disorder generally respond better to CBT than to pharmacotherapy (253–255), with at least two studies (251, 254) showing that the combination of CBT and medication is superior to either alone. Two of the studies suggested a greater improvement in mood and anxiety variables when antidepressant therapy is added to CBT (251, 253, 256). Of note, some experienced clinicians do not find rigidly defined and doctrinally practiced CBT to be as useful as methods that integrate CBT with other psychotherapeutic techniques. This may be due to several factors, including clinician inexperience or discomfort with the methods of CBT or differences among patients seen in the community and those who have participated as research subjects in these studies (81, 153).

Bright light therapy has been shown to reduce binge frequency in several controlled trials (257–259). Case reports suggest that repeated transcranial magnetic stimulation may be effective in treating patients with major depression and bulimia nervosa (260, 261). One controlled trial (719) showed odansetron, a peripheral 5-hydroxytryptamine type-3 (5-HT₃) receptor antagonist that reduces vagus nerve activity, to be effective in decreasing symptoms of bulimia nervosa, and its use may be considered in unusual circumstances.

The eating disorder not otherwise specified (EDNOS) category is a conceptually problematic one and comprises a clinically heterogeneous group of diagnoses (262, 263). This "everything else" category currently consists largely of individuals with subsyndromal anorexia nervosa or bulimia nervosa who do not meet DSM-IV-TR criteria of being 15% below expected weight or who binge and purge slightly less than twice per week. Such individuals merit treatment similar to that of full-syndrome patients. In addition, the EDNOS category lumps together normal-weight patients who purge, individuals who chew and spit out their food without swallowing it to prevent weight gain, and patients with binge eating disorder. Also perhaps suitable for this "other" category are individuals who experience psychiatric impairment related to the abuse of diet pills and diuretics (264), individuals who are obsessively preoccupied with liposuction (265) to deal with issues of shape and weight, and certain new-onset postgastrectomy eating disorder patients (266). The EDNOS diagnosis covers a wide spectrum, so no easily generalizable comments can be made for the entire group regarding course or prognosis (267). In addition, over time, considerable movement occurs from one eating disorder diagnostic category to another, including EDNOS (263).

Binge eating disorder is the most discrete and well-studied EDNOS subgroup. Although binge eating disorder is currently not an approved DSM-IV-TR diagnosis, research criteria listed in DSM-IV-TR consist of disturbances in one or more of the following spheres: behavioral (e.g., binge eating), somatic (obesity is common, although not required), and psychological (e.g., body image dissatisfaction, low self-esteem, depression). Empirically supported strategies for the treatment of binge eating disorder include nutritional counseling and dietary management; individual or group behavioral, cognitive behavioral, dialectical behavioral, psychodynamic, or interpersonal psychotherapy; and medications. In reviewing the available information on treating binge eating disorder, it is important to consider the focus of treatment. Most programs using nutritional rehabilitation and counseling focus on weight loss as the primary outcome, whereas studies of psychotherapy and medication generally consider reduction of binge eating as the primary outcome measure, with weight loss as a secondary outcome. Clinical consensus suggests that psychodynamic psychotherapy may also be helpful to reduce binge eating in some patients.

The literature on treating binge eating disorder suggests that 1) behavioral weight control programs incorporating low- or very-low-calorie diets may help patients lose weight and usually reduce symptoms of binge eating; 2) at least some degree of weight gain often follows weight loss; 3) weight gain after weight loss may be accompanied by a return of binge eating patterns; and 4) various combinations of diets, behavior therapies, non-weight-directed psychosocial treatments, and even some "nondiet/health at every size" psychotherapy approaches may be of benefit in reducing binge eating and promoting weight loss or stabilization in various circumstances (268).

Some believe that patients with a history of repeated weight loss followed by weight gain ("yo-yo" dieting) or patients with an early onset of binge eating might benefit from following programs that focus on decreasing binge eating rather than losing weight (269, 270). However, at this point, there is little empirical evidence to suggest that obese binge eaters who are primarily seeking weight loss should receive different treatment than obese individuals who do not binge eat.

CBT is the most widely studied treatment for binge eating disorder, and there is substantial evidence supporting its efficacy for behavioral and psychological symptoms, whether it is delivered in the individual or group format. IPT and dialectical behavior therapy have also been shown to be effective for behavioral and psychological symptoms and can be considered as alternative therapies. There is less consensus regarding the long-term effects of treatment; however, some studies suggest that most patients continue to show behavioral and psychological improvement at 1-year follow-up (271, 272). There is a substantial body of evidence supporting the efficacy of self-help and guided self-help CBT treatment programs (273–277) and their use as an initial low-burden step in a sequenced treatment program.

Because severe dieting may disinhibit eating and lead to compensatory overeating and binge eating (278), and because chronic calorie restriction can also increase symptoms of depression, anxiety, and irritability (279), alternative therapies have been developed that use a "nondiet" approach and focus on self-acceptance, improved body image, better nutrition and health, and increased physical movement (280–282). Addiction-based 12-step approaches, self-help organizations, and treatment programs based on the Alcoholics Anonymous model have been tried, but no systematic outcome studies of these programs are available.

In sum, there appear to be several good psychotherapeutic options for treating binge eating disorder when a reduction in binge eating is the primary goal. Weight loss, particularly in the long term, is a much more elusive goal, not only for obese patients with binge eating disorder but for obese patients in general. However, several studies suggest that at least for some patients at certain stages of recovery, behavioral weight control may be a useful treatment component. Also, because studies have found that binge eating may begin before obesity or dieting (283), specific approaches are needed for nonobese patients struggling with binge eating symptoms. The optimal sequencing of treatments—that is, whether the treatment of binge eating should precede or occur concurrently with weight control treatment—has yet to be definitively determined.

There is substantial evidence to suggest that treatment with antidepressant medications, particularly SSRI antidepressants, is associated with at least a short-term reduction in binge eating, in most cases without substantial weight loss. The dosage of medication is typically at the high end of the recommended range. The appetite-suppressant medication sibutramine also appears to be effective in suppressing binge eating, at least in the short term, and is additionally associated with significant weight loss (284). Heart rate and blood pressure need to be monitored closely in patients taking sibutramine, and the medication should be discontinued if there are significant elevations in these parameters, although these side effects seem to be uncommon (285). Finally, the anticonvulsant medication topiramate appears to be effective in reducing binge eating and promoting weight loss in the short (286) and long (287) term, although side effects such as cognitive problems, paresthesias, and somnolence may limit its clinical utility for some individuals. The anticonvulsant zonisamide may produce similar effects (288). Dexfenfluramine, although effective for reducing binge eating (289), has been removed from the market because of increased risk of primary pulmonary hypertension and heart valve abnormalities. Patients who report having used fenfluramine and phentermine in the past should be screened for potential cardiac and pulmonary complications.

It is important to note that in several studies, the placebo response rate has been reported to be quite high. The clinical implications of this finding are that controlled studies are extremely important, as a positive response in an open study may be nonspecific, and short-term beneficial responses to treatment should be viewed cautiously, given that a transient "honeymoon" effect of initiating treatment is common.

There have been few studies of combined treatment for binge eating disorder, so the clinical recommendations are preliminary. Overall, it appears that for most patients, the addition of antidepressant medication to behavioral weight control and/or CBT does not significantly augment binge suppression but may confer additional benefits in weight reduction (290–294). One study reported that the addition of the weight-loss medication orlistat to a guided self-help CBT program yielded additional weight loss (295). Another study found that fluoxetine in the setting of group behavioral treatment did not augment binge cessation or weight loss but did reduce depressive symptoms (294). Thus, the addition of medication to psychotherapy for binge eating disorder is not, in most cases, associated with additional benefit on the core symptom of binge eating, perhaps because psychosocial treatments are quite effective for this symptom. However, medication augmentation may have additional benefits.

The phenomenon of wakeful nighttime eating, variously characterized as night eating syndrome, nocturnal eating/drinking syndrome, or nocturnal sleep-related eating disorders, is currently an area of active research (296). Although formal agreed-upon definitions for these syndromes do not yet exist, the construct of night eating syndrome, first described by Stunkard et al. (297), generally includes morning anorexia, evening hyperphagia, and insomnia. In contrast, the construct of nocturnal eating/drinking syndrome emphasizes a sleep disorder with recurrent awakenings often accompanied by eating or drinking, and the construct of nocturnal sleep-related eating disorders adds to this a reduced level of awareness or recall of nocturnal eating episodes. Sleep-related eating disorders, including somnambulism, have reportedly been induced by risperidone, olanzapine, and bupropion, among other medications (298–300). The literature does not, at this point, support the recommendation of particular treatments for these disorders. However, there is preliminary evidence supporting the utility of progressive muscle relaxation (301) and sertraline (302, 303). Further studies of the phenomenology and treatment of these disorders are needed.