Walter H. Kaye, MD

Anorexia nervosa (AN) is a severe psychiatric disorder characterized by restricted eating, significant weight loss, and a profound fear of weight gain. AN and the related disorder, bulimia nervosa, have a lifetime prevalence of 1% to 2% of the population.

 Weight loss in AN results in serious medical consequences and higher levels of care are often necessary to restore weight. AN has the highest risk of mortality of any nonsubstance-use psychiatric disorder and, too often, a poor prognosis.

 Reluctance to recover is sustained by ego-syntonic symptoms that reinforce the illness.

 Only 13% to 50% of patients are considered recovered 1 to 2 years posttreatment, and a sizeable subset (20% to 30%) go on to develop a chronic and unremitting course of AN.

Limited effectiveness of both pharmacological and psychological treatments contributes to this chronicity. Family-based therapy, which works in alliance with parents to maintain eating and restore weight, has had some success in adolescent patients, but not all younger patients achieve adequate benefit. For older patients with AN, evidence-based therapies remain underdeveloped.

The Current State of the Art and the Not Too Distant Future

Treatment-Resistant OCD: Strategies and Novel Treatment Options

Hewa Artin, MD; Michelle F. Sloan, MD; Zafiris J. Daskalakis, MD, PhD

Reducing Heterogeneity in Non–Treatment-Resistant and Treatment-Resistant Schizophrenia

Frederick Charles Nucifora Jr, PhD, DO, MHS

Historically, AN has been considered a consequence of psychosocial factors.

 However, considerable recent evidence supports powerful genetic and neurobiological contributions to AN behavior.

 Many affected individuals have an anxious, obsessive, inflexible, and perfectionist temperament that presents in childhood before the onset of AN, becomes exaggerated during illness, and persists in a mild-to-moderate form after recovery. Comorbid depression is common, as is denial of illness and lack of motivation to change.

Recent studies have shed new light on underlying mechanisms driving such AN symptoms.

 For example, when people are hungry, activation of brain circuits normally makes them feel irritable and uncomfortable and makes food more rewarding and motivating. These neural processes appear to be miscoded in AN and sensitized with weight loss.

 Instead, when these individuals become hungry and are presented with food, they experience brain circuit activity associated with considerable anxiety; further, dopamine-related brain mechanisms that motivate and initiate eating fail to activate. Such findings open the door to identifying new targets and new pharmacological approaches in AN.

It is important to note that there are no US Food and Drug Administration (FDA)–approved medications to treat AN. Nonetheless, given the risk of significant physical and psychological consequences of this disorder, off-label prescribing of some agents has become widespread.

For patients with AN who show signs of treatment resistance, the most common medication intervention is atypical antipsychotics. Although promising case reports, open trials, and chart reviews began appearing in the late 1990s, the controlled-trial literature on atypical antipsychotics in AN remains disturbingly limited, which is a testament to the notorious difficulties in recruiting and retaining patients with AN in medication trials. To date, there have been 8 randomized, placebo-controlled trials (RCTs) involving placebo groups. Only 1 of these involved more than 40 patients, and the total number of randomized patients in all trials was less than 350.

Olanzapine is by far the most studied medication for AN, with 5 trials taking place between 2007 and 2019. All but 1 were considered positive, although the degree of benefit typically was small. Two trials involving use of quetiapine had mixed results, and 1 trial with risperidone was negative.

 In the closest thing to a landmark medication study in AN, Attia et al reported results of a relatively large, multisite RCT of olanzapine in adult outpatients in 2019.

 Five sites in North America randomized 152 patients to olanzapine or placebo during a 6-year recruitment period. The 16-week trial used a target olanzapine dose of 10 mg/d (mean, 7.77 mg/d), and investigators found a significant increase in weight for the olanzapine group versus placebo. However, the difference in weight gain (ie, 1 lb of body weight/month for a woman of average height) was modest between the groups; patients using olanzapine gained an average of approximately 1.5 lb of body weight/month. Weight restoration is the primary outcome in any AN trial, but use of a medication hopefully also will impact the powerful thought processes and psychiatric symptoms that drive the illness. Importantly, the authors reported that use of olanzapine showed “no significant benefit for psychological symptoms.”

Despite these limitations, expert articles and guidelines most likely recommend olanzapine as the antipsychotic of choice to treat AN, because so few RCTs have involved use of other medications. In our experience, a few types of patients may particularly benefit from olanzapine. For individuals who desire some degree of weight restoration but are struggling due to lack of appetite or drive to eat, olanzapine may provide improvement. In addition, patients with very high anxiety presenting more as fear than cognitive rigidity may benefit from the sedative-like properties of olanzapine.

However, there are substantial roadblocks to successful use of olanzapine for AN. The quality that led to early interest in and dominant study of olanzapine—its well-known propensity to cause weight gain—makes it unacceptable to the vast majority of patients with AN. At best, many of these individuals are ambivalent about weight gain, and even those who have accepted it as a necessary aspect of recovery frequently fear loss of control over the process. It is critical to understand that improvement in body image for AN patients can significantly lag weight restoration. If a medication adds weight but has little impact on thought process, the patient may experience an intolerable emotional state that greatly heightens the risks of nonadherence and resumption of weight loss.

With 14 atypical antipsychotics available for use in the United States, it is astonishing that so few have been studied in RCTs of AN. Given the highly complex and variable receptor-binding properties of this class, leading to clinically relevant heterogeneity—for example, some atypical antipsychotics are known to treat depression better than others—it is premature to conclude that atypical antipsychotics have only minimal benefit for AN based on results reported with use of 3 of the earlier medications. Lack of RCTs with newer, relatively weight-neutral antipsychotics, especially those having a somewhat different mechanism of action (eg, aripiprazole), may be a serious oversight.

Over the past decade, and with the 2019 olanzapine trial notwithstanding, nearly all additions to the literature on the use of atypical antipsychotics in patients with AN have involved aripiprazole. Five peer-reviewed articles from 4 different sites in 3 countries have described very interesting findings from case series and retrospective chart reviews.

 These reports often show benefit for weight gain, but most notably, they have included remarkably similar descriptions of an impact on thought processes related to AN, particularly the loosening of rigid and entrenched rules and distortions. Observations have included reduced eating-specific anxiety and distress; less rigid food choices; diminished obsessional thoughts about food, exercise, weight, and body image; and fewer eating disorder (ED) rituals. Additionally, many patients reported reduced symptoms of depression and non-ED–related anxiety, which will be discussed in the next section. Therefore, regardless of any direct impact on weight, aripiprazole may indirectly enhance weight restoration by improving cognition and thereby promoting a better response to psychological and dietary treatments. RCTs are needed to further investigate these findings.

What may account for this possible difference between olanzapine and aripiprazole in AN? Olanzapine—as well as most other current atypical antipsychotics, including quetiapine and risperidone—is a 5-HT2

 antagonist. Aripiprazole, in contrast, is a D

 partial agonist, meaning it acts differently at dopamine and serotonin receptors. Theoretical arguments have addressed how use of aripiprazole at lower doses (typically, 1-5 mg/d, with occasional titration up to 10 mg/d) may facilitate learning and behavior change in individuals with AN.

For example, basic research has indicated that dopamine receptor stimulation may be beneficial in suppressing previous fears after extinction training. Moreover, neurobiological research suggests that AN is associated with heightened responsiveness in brain reward circuits, which may be due to a hypersensitive dopamine system. Aripiprazole may downregulate D

 receptor activity, which may improve brain function, reduce anxiety, and aid in the psychotherapeutic process.

Additionally, almost all patients described in this research on aripiprazole use were being treated concurrently with antidepressants. In contrast, no RCT has formally studied atypical antipsychotic augmentation of antidepressants in AN. Although some RCTs allowed patients already on antidepressants to continue them, use was not systematic or evaluated, and the overall numbers were too small to yield meaningful results. We believe that, particularly for medications such as aripiprazole with known significant effects as augmenters, combination treatment may be 1 key to showing a more robust response in controlled trials. On a practical note, adherence also has improved with use of aripiprazole, because it often is better tolerated. Still, some patients encounter akathisia, which we have been able to minimize by initiating therapy with very low doses (eg, as low as 0.5 mg/d in some patients).

AN is associated with high levels of comorbid conditions. More than 40% of individuals with AN have a lifetime comorbid anxiety or mood disorder, over 30% have an impulse control disorder, and about 27% have a substance use disorder.

 Recent data showed that individuals with AN had a 25% risk of receiving any noneating disorder diagnosis during the first 2 years of longitudinal follow-up and a 55% risk of developing comorbid conditions 2 decades after inclusion.

 Comorbid conditions have been associated with greater illness severity and mortality.

Research on eating disorders traditionally has not focused on comorbidities, but new data highlight the need for treatment of those conditions. Whether eating disorder outcomes will be affected by treatment of comorbid psychiatric illnesses is uncertain, but improving quality of life and reducing suicide rates are similarly important goals.

Few medications specific to AN have been approved by the FDA, yet affected patients frequently receive prescriptions for psychotropics.

 However, studies on pharmacological interventions in AN commonly have not been designed to discern whether individuals taking these medications had more severe illness or higher comorbidity than did others. Because depression and anxiety predict more negative treatment outcome in AN, and comorbid psychiatric disorders are associated with higher mortality among this patient population, use of medication to target these conditions is supported.

 Research may have to adapt and focus more on outcome measures that go beyond weight gain; these may include the study of quality of life, mood, and anxiety measures in relation to use of medical interventions. These measures are crucial in preventing weight loss and relapse after significant resources that involve a high level of care have been used for nutritional rehabilitation.

A recent patient with severe AN in our program also had severe, comorbid obsessive-compulsive disorder (OCD). She had many magical thinking–related obsessions that substantially limited her will to live and contributed to overwhelming major depressive disorder with suicidality. The patient was treated with olanzapine to reduce distress and had been on various antidepressant and antianxiety medications, but she never received an adequately dosed trial of a selective serotonin reuptake inhibitor to treat OCD. During weight restoration, we initiated sertraline, which was titrated to 200 mg/d. After 4 months of treatment, the patient experienced significant relief from symptoms of OCD and could start participating in social activities. AN was still a significant struggle, but the patient experienced improvement in daily life activities and overall functioning.

While olanzapine has shown benefit for AN in RCTs, weight restoration tends to be minor, AN cognitions do not appear to improve significantly, and most patients are unaccepting of a medication known to cause weight gain. RCTs of additional atypical antipsychotics, especially those that have a somewhat different mechanism of action (eg, aripiprazole), used alone and in combination with antidepressants are needed. Research into strategies to better treat psychiatric comorbidities in patients with this disorder may reduce treatment resistance. Finally, early investigations to treat AN and its comorbidities with novel biological treatments (eg, psilocybin, ketamine, cannabinoids, and neuromodulation techniques of repetitive transcranial magnetic stimulation and deep brain stimulation) are ongoing.

 is an associate clinical professor of psychiatry at the University of California, San Diego, and medical director of the UCSD Eating Disorders Treatment and Research Program. 

 is a professor in residence of psychiatry at the University of California, San Diego. 

 is director of medical and clinical services at the USCD Eating Disorders Treatment and Research Program, and clinical professor of psychiatry at the University of California, San Diego. 

 is the founder and executive director of the UCSD Eating Disorders Treatment and Research Program, and professor of psychiatry at the University of California, San Diego.

1. Ward ZJ, Rodriguez P, Wright DR, et al. 

Estimation of eating disorders prevalence by age and associations with mortality in a simulated nationally representative US cohort

2. Arcelus J, Mitchell AJ, Wales J, Nielsen S. 

Mortality rates in patients with anorexia nervosa and other eating disorders

3. Abbate-Daga G, Amianto F, Delsedime N, et al. 

Resistance to treatment and change in anorexia nervosa [corrected]: a clinical overview

4. Wonderlich SA, Bulik CM, Schmidt U, et al. 

Severe and enduring anorexia nervosa: update and observations about the current clinical reality

Presentation of eating disorders in the news media: what are the implications for patient diagnosis and treatment?

6. Brandys MK, de Kovel CGF, Kas MJ, et al. 

Overview of genetic research in anorexia nervosa: the past, the present and the future

7. Monteleone AM, Castellini G, Volpe U, et al. 

Neuroendocrinology and brain imaging of reward in eating disorders: a possible key to the treatment of anorexia nervosa and bulimia nervosa

Prog Neuropsychopharmacol Biol Psychiatry

8. Kaye WH, Wierenga CE, Bischoff-Grethe A, et al. 

Neural insensitivity to the effects of hunger in women remitted from anorexia nervosa

9. Frank GKW, Shott ME, Stoddard J, et al. 

Association of brain reward response with body mass index and ventral striatal-hypothalamic circuitry among young women with eating disorders

The role of psychotropic medications in the management of anorexia nervosa: rationale, evidence and future prospects

Cognitive neuroscience of eating disorders

12. Attia E, Steinglass JE, Walsh BT, et al. 

Olanzapine versus placebo in adult outpatients with anorexia nervosa: a randomized clinical trial

13. Frank GKW, Shott ME, Hagman JO, et al. 

The partial dopamine D2 receptor agonist aripiprazole is associated with weight gain in adolescent anorexia nervosa

14. Tahillioğlu A, Özcan T, Yüksel G, et al. 

Is aripiprazole a key to unlock anorexia nervosa?: A case series.

15. Bailer UF, Price JC, Meltzer CC, et al. 

Dopaminergic activity and altered reward modulation in anorexia nervosa—insight from multimodal imaging

16. Hudson JI, Hiripi E, Pope HG Jr, Kessler RC. 

The prevalence and correlates of eating disorders in the National Comorbidity Survey Replication

17. Steinhausen HC, Villumsen MD, Hørder K, et al. 

Comorbid mental disorders during long-term course in a nationwide cohort of patients with anorexia nervosa

18. Kask J, Ekselius L, Brandt L, et al. 

Mortality in women with anorexia nervosa: the role of comorbid psychiatric disorders

Is the severity of the clinical expression of anorexia nervosa influenced by an anxiety, depressive, or obsessive-compulsive comorbidity over a lifetime?

20. Alañón Pardo MDM, Ferrit Martin M, Calleja Hernández MAC, Márquez FM. 

Adherence of psychopharmacological prescriptions to clinical practice guidelines in patients with eating behavior disorders

Psychotropic usage by patients presenting to an academic eating disorders program

Depression, worry, and psychosocial functioning predict eating disorder treatment outcomes in a residential and partial hospitalization setting

23. Spriggs MJ, Douglass HM, Park RJ, et al. 

Study protocol for “Psilocybin as a treatment for anorexia nervosa: a pilot study.”

24. Schwartz T, Trunko ME, Feifel D, et al. 

A longitudinal case series of IM ketamine for patients with severe and enduring eating disorders and comorbid treatment-resistant depression

Treatment studies with cannabinoids in anorexia nervosa: a systematic review

Current therapeutic approaches to anorexia nervosa: state of the art

Articles

Only 13% to 50% of AN patients are considered recovered 1 to 2 years posttreatment, and 20% to 30% go on to develop a chronic and unremitting course of AN. How can we improve these outcomes?

Pharmacological Management of Treatment-Resistant Anorexia Nervosa

Eating disorders (ED) are associated with significant comorbid psychopathology and the most extensive medical complications of any psychiatric disorder. Disordered behaviors that present across diagnoses (eg, restriction, binge eating, compensatory behaviors) are linked with acute medical risks that often require careful medical monitoring and clinical intervention. Yet, individuals with EDs frequently express hesitancy towards recovery and have great difficulty controlling their ED symptoms.

EDs are typically chronic and follow a treatment-refractory trajectory. Existing research indicates that even the most effective evidence-based outpatient treatment yields recovery for only around one-half of patients.

 In our clinic, we often encounter patients like “Hayley,” “Georgia,” and “Michael” (see Case Vignettes) for whom outpatient treatment has been unsuccessful. Like many other eating disorder treatment programs, our program seeks to optimize recovery by providing more intensive treatment with stepped levels of care.

Hayley is a 24-year-old graduate student and a competitive runner who competes regularly. She has a diagnosis of anorexia nervosa, restricting type (AN-R), which developed when she was 20 after a period of binging. Two years ago, Hayley began working with a multidisciplinary outpatient team, but she has continued to lose weight. She attributes the weight loss to her unwillingness to eat the large amount of food it would take to offset the calories she burns while running and training. She initially presented to her primary care physician feeling weak then was admitted to an inpatient eating disorders unit for bradycardia. Upon discharge from the hospital she was referred to a partial hospitalization treatment program.

Most research explores the efficacy of treatment at the outpatient level; little is known about the comparative value of the higher levels of care for patients with EDs.

 However, these approaches are important alternatives for treatment-refractory EDs because they provide the more intensive treatment and meal support that many severely ill patients need.

 Common reasons for needing higher levels of care include substantial weight loss that might be life threatening; difficulty eating enough food to gain weight; severe binge/purge behaviors; incapacitating ED symptoms; comorbid substance abuse; and anxiety, depression, obsessive compulsive disorder, or suicidal intent.

It is not unusual that such behaviors result in medical instability with symptoms of cardiovascular compromise, electrolyte disturbances, or hypoglycemia. It is important to note that some individuals have a lack of insight about their behaviors or are not motivated to engage in treatment. Because of these factors, the mortality rate for anorexia nervosa and bulimia nervosa can be 5% or higher.

 Higher levels of care provide a structured and protective environment that can be essential for these serious, life-threatening disorders.

Georgia, aged 17 years, has bulimia nervosa (BN) with comorbid major depressive disorder, panic attacks, and a history of sexual trauma. She has been engaging in polysubstance abuse, including alcohol, marijuana, prescription painkillers, and cocaine, since the age of 12. She has been hospitalized several times for suicide attempts and accidental drug overdoses. She binges and purges daily and states that she is only able to reduce these behaviors when using substances. Her outpatient therapist feels overwhelmed by all of Georgia’s comorbidities and is not sure how to help her reduce these self-destructive behaviors, so she refers Georgia to a PHP program.

Higher levels of care for EDs include various types of programs, with different levels of intensity.

The treatment goals for EDs are similar and include:

• Interrupting binge, purge, and restrictive behaviors;

• Providing psychoeducation regarding regular eating;

Psychotherapy approaches among higher levels of care are typically informed by cognitive behavioral and dialectical behavioral therapies. Family-based treatment is often incorporated for children and adolescents. Patients might transition between levels of care due to variables such as symptom severity, medical status, motivational status, treatment history, and financial limitations.

There are several different levels of care from which to choose:

 is the highest level of care available. This setting can be in a medical hospital and is intended for patients with acute medical instability. Alternatively, this may be in a psychiatric hospital if there are severe behavioral symptoms. In either event, subspecialty medical and behavioral consultation are readily available, meals are supervised, and one-to-one monitoring is available.

 offer full-time treatment in a non-hospital setting. Patients receive multidisciplinary care that includes nutritional support, medication management, and individual and group therapy.

(PHP), or daytreatment, offer treatment in an outpatient setting approximately six to 10 hours a day, between three and seven days per week. Patients typically spend nights and sometimes weekends on their own, allowing them to practice the skills they are learning in social, occupational, and leisure settings outside of treatment.

 (IOPs) offer treatment approximately three hours a day, from three to five days per week. At both the PHP and IOP levels, patients receive meal support, group therapy, individual therapy, dietary sessions, and medication management.

Treatment efficacy for higher levels of care. 

Randomized controlled trials (RCTs) are necessary to avoid the confounding effects of psychopathology severity and evaluate the comparative efficacy of different levels of care. Since patients with more severe symptoms and greater functional impairment are more likely to present to higher levels of care than those with mild ED pathology (and milder ED pathology is associated with better outcomes), reliable findings are dependent on patients matched with controls based on symptom severity or random assignment to level of care.

Unfortunately, RCTs that compare different levels of care are limited given significant costs and ethical considerations related to randomizing acutely ill patients. As such, literature on higher levels of care treatment efficacy is scarce and consists largely of open trials assessing outcome at discharge. A 2015 review of PHP and residential programs identified that duration of treatment was similar between these levels of care, and all but one study reported improvements in outcomes (ie, body mass index [BMI], number of binge/purge episodes) at discharge.

There have been several, more recent naturalistic studies of higher levels of care treatment that also offer support for the effectiveness of care in most patients.

 However, less than half of the open trials identified in the review reported follow-up data after discharge.

 Moreover, follow-up completion rates tend to be low; the average rate of follow-up completion was 66% for PHP and 37% for residential. Although most of the studies reported that positive treatment outcomes at discharge were maintained or improved at follow-up, the missing data at follow-up make long-term results difficult to interpret.

As previously mentioned, the literature on treatment efficacy at higher levels of care is preliminary and consists largely of open trials. Within the context of these trials, researchers have begun to study the predictors of long-term recovery. However, more is known about predictors of recovery from EDs in general, regardless of type of treatment, and no study has looked at how predictors of successful outcome might differ depending on level of care received.

 conducted a metaanalysis of predictors of treatment outcome that included a large number of patients treated at a higher level of care. Their findings indicate that patients with higher BMI, fewer binge/purge behaviors, greater motivation to recover, lower depression, lower shape/weight concern, fewer comorbidities, better interpersonal functioning, and fewer familial problems at baseline had better outcomes both at end-of-treatment and follow-up. The most robust predictor of outcome at both end-of-treatment and follow-up was the meditational mechanism of greater early symptom change.

The literature consistently corroborates that early behavioral change predicts later symptom remission.

 Given that higher levels of care treatment settings typically involve meal supervision and more opportunities to quickly learn skills, these settings may be uniquely well-suited to facilitating early symptom change.

When determining the appropriate level of care, practitioners should consider practice guidelines published by reputable organizations as well as individual variables important in predicting treatment response. The American Psychiatric Association (APA), Royal Australian and New Zealand College of Psychiatrists (RANZCP), and National Institute of Clinical Excellence (NICE) published guidelines for ED treatment that outline factors to consider in making decisions regarding level of care.

Michael is a 12-year-old boy who has avoidant/restrictive food intake disorder (ARFID), generalized anxiety disorder, and panic attacks. He endorses fear of food contamination and reports eating the same foods at the same time every day to avoid illness and/or vomiting. Michael has seen outpatient providers over the past few years for low weight, stunted growth, and gastrointestinal complaints. He was recently admitted to an inpatient eating disorders unit for low weight. Following discharge, he was unable to maintain weight at the outpatient level due to food-related fears resulting in restricted intake. As such, he was readmitted to the inpatient unit; he was later referred to PHP for continued care.

The NICE and RANZCP guidelines recommend first seeking outpatient care and suggest transfering to higher levels of care if there is no improvement of symptoms. Alternatively, the APA advises consideration of various factors when making initial treatment recommendations (

). All guidelines clearly outline criteria for determining medical stability and recommend that patients who do meet these requirements be considered for admission to inpatient hospitalization (

). A recent update on the medical management of EDs offers a list of criteria that indicate medical instability in adolescents that deviates slightly from the APA recommendations.

The RANZCP and APA guidelines suggest that inpatient treatment may be needed if ED symptoms are uncontrolled and require 24-hour supervision. Inpatient care and partial hospitalization also are appropriate when there is significant risk of suicide or self-harm. The NICE and APA guidelines advise considering logistical issues concerning the geographic location of treatment; they recommend that patients first receive care at facilities close to their home. Research indicates that PHPs are more cost-effective than residential or inpatient treatment; therefore, financial factors are important considerations.

 Clinicians might also consider recent scientific developments related to variables that account for significant variance in treatment response.

As any clinician who works with patients who have EDs knows, there are, unfortunately, many patients who are treatment resistant in outpatient care. For patients like Hayley, Georgia, and Michael, specialized higher levels of care facilities can be a relief for the clinician and life-changing for the patient.

Hayley was able to achieve her ideal body weight with PHP and stepped down to IOP. She decided not to return to competitive sports before returning to her outpatient team. Georgia stopped using drugs and alcohol within the structure of PHP, participated in weekly family therapy, and learned skills for coping with emotional dysregulation. Michael restored his weight in PHP through repeated food exposures that helped him increase his food volume and variety and decrease his fears surrounding contamination.

For many of the complex, severely ill patients who present to treatment, intensive treatment with stepped levels of care can offer a cost-effective structure that promotes positive behavioral change.

Dr Anderson is Associate Clinical Professor Director of Training, Ms Simpson is Milieu Therapist, and
	Dr Kaye is Professor and Founder and Executive Director, Eating Disorders Program, Department of Psychiatry, University of California, San Diego, CA. The authors report no conflicts of interest concerning the subject matter of this article.

1. Lock J. An update on evidence-based psychosocial treatments for eating disorders in children and adolescents. 

2. Hay P, Bacalchuk J, Stefano S. Psychotherapy for bulimia nervosa and binging. 

3. Galsworthy-Francis L, Allan S. Cognitive behavioural therapy for anorexia nervosa: a systematic review. 

4. Friedman K, Ramirez A, Murray S, et al. A narrative review of outcome studies for residential and partial hospital-based treatment of eating disorders. 

5. Arcelus J, Mitchell A, Wales J, et al. Mortality rates in patients with anorexia nervosa and other eating disorders. 

6. Crow S, Petersen C, Swanson S, et al. Increased mortality in bulimia nervosa and other eating disorders. 

7. Yager J, Devlin M, Halmi K, et al. Guideline watch: practice guideline for the treatment of patients with eating disorders. 

8. Yager J, Devlin J, Halmi K, et al. Guideline watch (August 2012): practice guideline for the treatment of patients with eating disorders, 3rd Edition. 

9. Brown T, Cusack A, Anderson L, et al. Efficacy of a partial hospital programme for adults with eating disorders. 

10. Hayes N, Welty L, Slesinger N, et al. Moderators of treatment outcomes in a partial hospitalization and intensive outpatient program for eating disorders. 

11. Vall E, Wade T. Predictors of treatment outcome in individuals with eating disorders: A systematic review and meta-analysis. 

12. Nazar BP, Gregor LK, Albano G, et al. Early response to treatment in eating disorders: a systematic review and diagnostic test accuracy meta-analysis. 

13. Wales J, Brewin N, Cashmore R, et al. Predictors of positive treatment outcome in people with anorexia nervosa treated in specialized inpatient unit: the role of early response to treatment. 

14. Hay P, Chinn D, Forbes D, et al. Royal Australian and New Zealand College of Psychiatrists clinical practice guidelines for the treatment of eating disorders. 

15. NICE. Core Interventions in the Treatment and Management of Anorexia Nervosa, Bulimia Nervosa and Related Eating Disorders (Clinical Guideline 9). London: National Collaborating Centre for Medical Health; 2004.

16. Golden N, Katzman D, Sawyer S, et al. Update on the medical management of eating disorders in adolescents. 

17. von Wietersheim J, Zeeck A, KÃ¼chenhoff J. Status, possibilities and limitations of therapies in psychosomatic day clinics. 

18. Williamson D, Thaw J, Varnado-Sullivan P. Cost-effectiveness analysis of a hospital-based cognitive-behavioral treatment program for eating disorders. 

19. Vall E, Wade T. Predictors of treatment outcome in individuals with eating disorders: A systematic review and meta-analysis. 

Eating disorders (ED) are associated with significant comorbid psychopathology and the most extensive medical complications of any psychiatric disorder.

Higher Levels of Care for Eating Disorders: A Practical Guide

Eating disorder behaviors can range from dangerous caloric restriction to eating that feels out of control-often combined with unhealthy weight control behaviors, such as self-induced vomiting or laxative abuse. Anorexia nervosa, bulimia nervosa, and binge eating disorder are the most prevalent eating disorders. Diagnostically, anorexia nervosa requires low body weight, intense fear of or behavior that interferes with weight gain, and disturbance in the perception of one’s weight or shape.

Criteria for bulimia nervosa include recurrent binge eating (eating a large amount of food in a discrete period of time accompanied by a sense of loss of control) and actions meant to counteract weight gain (eg, self-induced vomiting), as well as overvaluation of body weight or shape. Binge eating disorder involves recurrent binge eating without inappropriate compensatory behavior. Individuals with eating disorders frequently report extreme body dissatisfaction, anhedonia, anxiety, and difficulty tolerating negative emotions, although these are not required for diagnosis.

The cause of eating disorders remains unknown. Given that eating disorders reflect a discrepancy between the physiological need to eat and the drive to eat, it has been hypothesized that these individuals have a fundamental disturbance of appetite regulation. In recent years, advances in brain imaging have provided new insights into neural pathways that play a role in emotions, reward, salience, and inhibition of eating.

Mabel is a 16-year-old girl who is hospitalized with unstable vital signs and a BMI of 17.5. Her family describes her as an anxious, shy, and perfectionistic child who tended to ruminate about academics, social relationships and, more recently, food. Mabel says that she has always been a picky eater. She started to lose weight after she got braces and switched to softer foods 2 years ago. She reports that her food rules became progressively more restrictive and she began exercising regularly. She became increasingly anhedonic, cognitively rigid, and sensitive to perceived criticism. She interprets her family’s efforts to support her weight restoration efforts as “nagging” and judgmental.

Current treatments-psychotherapy or pharmacotherapy-often have limited efficacy in ameliorating eating disorder symptoms and may not fully normalize eating behavior. Furthermore, many psychological interventions used to treat eating disorders are adaptations of treatments for other psychiatric disorders (eg, depression, anxiety) and were not developed for the specific, unique needs of patients with eating disorders.

The lack of a mechanistic understanding of eating disorders has thwarted efforts to develop more powerful, evidence-based interventions. Alterations in brain circuitry linked to reward and inhibition in particular appear to be involved in the maladaptive eating behavior characteristic of anorexia nervosa, bulimia nervosa, and binge eating disorder.

What is it about the brains of those with eating disorders that might drive dysfunctional eating behavior? It is well known that gut and hypothalamic mechanisms contribute to the regulation of energy metabolism and eating behaviors. However, in recent years, studies in humans have revealed that higher-order, corticolimbic systems likely play a role in the pathophysiology of eating disorders.

 These systems integrate behavior with eating and can override homeostatic signals (

The first of these networks codes for perceived importance of a food stimulus and comprises the insula and anterior cingulate. The insula, along with the frontal operculum, also processes basic sensory information about food. The second pathway includes the nucleus accumbens, putamen, and caudate, as well as the orbitofrontal cortex and amygdala. These regions code for the rewarding and motivating value of eating and contribute to approach or avoidance behavior.

The third network helps control consumption based on consideration of both short- and long-term outcomes (eg, weight gain). It includes the dorsal caudate and dorsal anterior cingulate, lateral prefrontal cortex, and parietal cortex. These systems interactively weigh the reward value of food and the consequences of consuming it, and they integrate this information with homeostatic and motivational drives to guide eating behavior.

Mounting evidence suggests that an altered balance of reward and inhibition may contribute to disordered eating.

 In anorexia nervosa, severely restricted food intake appears to be related to overactive inhibitory control in combination with underactive reward circuitry. In contrast, dysregulation of both inhibitory and reward drives may manifest in the alternating over- and under-consumption characteristic of bulimia nervosa. Binge eating disorder may be related to altered sensitivity of ventral reward regions. Findings also suggest that these brain-based differences are linked to temperament traits, such as anxiety and harm avoidance, which persist after remission and may underlie the development of eating pathology.

The brain has different responses to the anticipation and consumption of food. People with binge eating-bulimia nervosa or binge eating disorder-demonstrate reduced response when anticipating food reward in the insula and ventral striatum, but increased response to the receipt of food in these same regions. Alternatively, individuals with anorexia nervosa demonstrate increased and anxious sweet-taste anticipation in the insula, striatum, and prefrontal cortex and reduced responses in the insula and striatum to the tastes themselves.

 Differences in the way the brains of individuals with current or past eating disorders anticipate and process food rewards likely drive their approach to food.

Processing of other salient stimuli also seems to be altered, which suggests a generalized deficit. In healthy individuals, the ventral striatum and dorsal caudate typically show greater responses to reward than to punishment. In contrast, imaging research in eating disorders demonstrates dysfunctional neural valuation of reward and punishment. For example, when making choices about monetary rewards, the brains of women with histories of anorexia nervosa or bulimia nervosa don’t differentiate between wins and losses in the ventral striatal regions.

 Women with remission of symptoms of anorexia nervosa also show reduced reward response to money (

 This suggests that individuals who have eating disorders may have difficulty evaluating rewards.

Individuals with eating disorders also have problems with inhibitory control. Individuals with bulimia nervosa and binge eating disorder frequently demonstrate reduced inhibition that extends beyond food (eg, substance use, shoplifting, self-injury). Neuroimaging studies suggest this may relate to failure to appropriately engage frontostriatal circuits that regulate behavior. Some functional MRI (fMRI) findings show that adults and adolescents with bulimia nervosa have reduced inhibitory activation.

 Other results have shown increased activation of frontostriatal regions in adolescents with binge-purge syndromes.

 Inefficient or insufficient activation in these control-related regions may clinically manifest as difficulty in stopping eating during binge episodes, as well as difficulty in resisting urges to purge.

Enhanced cognitive control in anorexia nervosa may contribute to an ability to restrict intake even when emaciated. Recent fMRI studies of individuals with anorexia nervosa reveal increased activity within dorsolateral cognitive circuitry associated with decision-making and inhibitory control (see 

 Although further research is needed, findings suggest that increased activation in control-related regions during the delay of gratification may underlie the extreme dietary restriction characteristic of anorexia nervosa.

Mabel explains that her difficulty with eating is driven by a lack of interest in food and increased anxiety about making meal choices. She reports that without a structured meal plan, she fasts to avoid the negative thoughts and mood that accompany eating. She also reports significant physical discomfort (eg, bloating) after eating a small amount of food. To avoid discomfort, when unsupervised, she typically eats only yogurt and small bites of fruit and vegetables in the evening. She is also quite anhedonic.

This case suggests both under- response to reward and over-control, despite extreme nutritional deprivation. It also suggests that external structure may be effective in helping Mabel restore weight.

Similarities to and differences from substance abuse

There are many similarities between substance abuse and eating disorders, particularly bulimia nervosa and binge eating disorder, including problems in the interconnected reward and inhibitory circuits. Meta-analyses suggest increased rates of drug and alcohol abuse in persons with bulimia nervosa. Women and adolescent girls with bulimia nervosa or binge eating disorder show deficient activation in brain regions related to inhibitory control similar to patterns of activation reported in drug and alcohol users.

 Moreover, the increased activation in response to food cues seen in bulimia nervosa and binge eating disorder mirrors increased activation in these regions in substance users exposed to drug-related cues.

 These similarities in brain activation may be related to underlying similarities in neurotransmitter activity: animal models of binge eating and human neuroimaging studies implicate similar neurotransmitters as those involved in substance use disorders, particularly those key to cognitive control, reward, and salience.

In contrast, rates of drug and alcohol abuse are decreased in patients with anorexia nervosa. Individuals with anorexia nervosa demonstrate opposite patterns of neural activation in response to reward compared with persons with bulimia nervosa, binge eating disorder, or substance abuse. The over-control of individuals with anorexia nervosa, particularly those with the restricting subtype, may protect them from developing dysregulated substance use disorders.

Some pharmacological treatments have shown success. Fluoxetine is currently considered the gold-standard medication for bulimia nervosa and is the only FDA-approved agent for the treatment of this disorder. Studies have shown greater amelioration of bulimia nervosa symptoms at higher dosages (60 mg/d), whether at the initiation of treatment or during rapid titration.

Lisdexamfetamine, an appetite-suppressing amphetamine, is the only medication that is FDA-approved specifically for treating binge eating disorder. Clinical trials indicate that it leads to the cessation of binge eating behavior in up to 50% of patients.

 However, 80% of persons treated with lisdexamfetamine report significant adverse effects, and there is a risk of abuse and dependence.

Results of pharmacological treatment studies for anorexia nervosa have been mixed despite frequent off-label use.

 Psychological and behavioral treatments are necessary to fully address the complex psychological and medical syndromes associated with eating disorders. Viewing eating pathology from the perspective of an altered balance in self-regulatory control and reward processing provides a foundation for developing more specific, effective interventions for these debilitating disorders.

Current treatments for eating disorders, particularly for chronic anorexia nervosa in adults, are inadequate. Randomized, controlled trials of cognitive and psychodynamic treatments suggest that neither differs from treatment as usual in promoting weight restoration, and many patients drop out of treatment before full weight restoration.

 Strategies that focus on brain-based differences are needed, in particular to treat adult anorexia nervosa.

Treatment of adolescent anorexia nervosa is more promising: considerable evidence supports the use of family-based treatment (FBT).

 FBT is a manualized form of eating disorder–focused family therapy in which parents are enlisted to ensure appropriate food intake and other health-oriented behaviors associated with recovery. Parent utilization addresses the motivational deficits inherent in anorexia nervosa. Despite the effectiveness of this treatment in younger patients, to date there have been no published results of clinical trials of FBT in adults. Adult anorexia nervosa treatment may be improved by enlisting caretakers to oversee recovery-oriented behaviors in a modified, age-appropriate manner.

Treatment of anorexia nervosa in adults may also be improved by directly targeting the brain-based temperament traits characteristic of the disorder. Imaging studies in adults who have anorexia nervosa have shown that differences in reward processing are linked to elevated anxiety or harm avoidance.

 Anxious temperament and harm avoidance are present before anorexia nervosa develops, and both tend to persist after symptom remission.

 The combination of increased anxiety in anticipation of food and diminished brain reward response to receiving it may be what makes it so difficult for people with anorexia nervosa to eat.

If clinicians can work with these traits rather than against them, treatment effectiveness may improve.

 For example, increasing predictability and structure around meals while maintaining caloric sufficiency can potentially help mitigate anticipatory anxiety and increase compliance with meals.

After discharge from the hospital, Mabel, her mother, and her sister begin taking part in a brief intensive family treatment program. They devise a contingency management plan (eg, behavioral contract) to establish a system of external rewards and consequences to reduce eating disorder behaviors. Mabel reports decreased anxiety about meals when she knows what to expect and expresses a desire to avoid the negative consequences of reduced time with friends that is built into her contract if she engages in eating disorder behavior.

Eating disorders are chronic and debilitating illnesses, and current treatments are ineffective for many patients. Neuroimaging research suggests eating disorder symptoms may result from dysfunction in circuits underlying reward and inhibition. The brain-based temperament traits of eating disorders are stable across illness and recovery. This suggests that treatments that work with these traits and target disturbances in inhibitory control, reward sensitivity, and salience may better address disease-specific mechanisms contributing to disordered eating-and may also enhance adherence and motivation, and thereby improve outcomes.

New insights into neural pathways that play a role in anorexia nervosa, bulimia nervosa, and binge eating disorder.

Walter H. Kaye, MD

Articles

Pharmacological Management of Treatment-Resistant Anorexia Nervosa

Higher Levels of Care for Eating Disorders: A Practical Guide

Neurobiology of Eating Disorders: Clinical Implications

Eating Disorders: Understanding Anorexia Nervosa

Advances in the Treatment of Eating Disorders

Pharmacological Treatment of Bulimia Nervosa