Evidence-Based Psychotherapy for Eating Disorders

According to the Diagnostic and Statistical Manual, Fourth Edition (DSM-IV) (1), lifetime prevalence estimates for eating disorders in the United States are: 0.9% for anorexia nervosa, 1.5% for bulimia nervosa, and 3.5% for binge eating disorder among women; and 0.3% for anorexia nervosa, 0.5% for bulimia nervosa, and 2.0% for binge eating disorder among men (2). The onset of eating disorders is typically in adolescence, with puberty a critical risk period, but the disorders may occur in patients as young as 7 and as old as 70 years or more (3). Eating disorders can have substantial and sustained physical consequences, comorbidity, mortality, chronicity, and functional impairment (4, 5).

The Diagnostic and Statistical Manual (DSM-5) (6) groups together eating and feeding disorders; here we cover the eating disorder diagnoses anorexia nervosa, bulimia nervosa, binge eating disorder, and other specified feeding and eating disorders. The diagnostic criteria are outlined in Table 1. Eating disorders involve dysregulated eating habits such as a restriction of intake below energy needs, omitting entire food groups (e.g., fats or carbohydrates), calorie counting, skipping meals, and fasting. For some there is a cyclic pattern of restricting, which brings on the body’s starvation response, triggers binge eating, and then leads to further restricting and compensatory behaviors to neutralize the calories consumed. Binge eating is similar to overeating in that both involve consuming an unusually large amount of food in a discrete period of time, but binge eating is distinguished by the accompanying sense of loss of control (i.e., feeling unable to stop eating). Compensatory behaviors include purging (self-induced vomiting, laxatives, and diuretics), fasting, and compulsive exercise.

Long-term medical complications of eating disorders are significant. Starvation and purging symptoms have the potential to deteriorate every major organ system in the body, leading to kidney damage, anemia, cardiovascular problems, dental problems, and changes in brain structure. Osteoporosis and osteopenia are irreversible consequences of malnutrition. Obesity is commonly associated with BED, but BED also occurs in normal weight individuals (7). Mortality and morbidity are elevated across all eating disorders (4, 8).

Psychotherapy for eating disorders may occur in a number of contexts. Some patients may do well with empirically supported self-help or outpatient psychotherapy. Others may require more intensive treatment in the form of day or partial hospitalization or intensive outpatient treatment, which may include structured meal support, family interventions, group psychotherapy, cooking and nutrition groups, and social skills interventions (9). Psychotherapy for anorexia nervosa may occur along with medical admissions for those patients who are acutely medically unwell and/or severely malnourished. Attempts to conduct formal psychotherapy when the patient is severely malnourished can be ineffective, because of mild cognitive impairment (10). Supportive psychotherapeutic interventions that provide empathy, support, and positive behavioral reinforcement may be more appropriate than more intensive psychotherapies until renourishment is well underway (10).

There is no strong evidence for the efficacy of any psychotherapy or pharmacotherapy in the treatment of anorexia nervosa in adults. Clinical practice guidelines (13, 14) recommend considering cognitive-behavioral therapy (CBT), interpersonal psychotherapy (IPT), cognitive analytic therapy, and focal psychodynamic therapy, and regular physical monitoring by a primary care practitioner. These psychotherapies are the most well-evaluated and theoretically supported approaches. The RCT base is limited (11), and marred by small sample sizes, high drop out, short duration, lack of long-term follow-up, and other methodological issues.

Since the publication of clinical guidelines new data have been reported. A large, multicenter RCT compared 10 months of focal psychodynamic therapy, CBT enhanced (CBT-E), and optimized treatment-as-usual (15). At endpoint, intent-to-treat analysis showed significantly increased body mass index (BMI) in all groups and no treatment differences. At 12 month follow-up, BMI continued to improve but again, there were no differences by treatment. Thirty percent of the patients dropped out during treatment.

In addition, seven year follow-up data were reported from an RCT that compared CBT, IPT, and specialist supportive clinical management (16, 17). At treatment completion, a significantly greater number of patients who completed specialist supportive clinical management (36%) no longer met diagnostic criteria for anorexia nervosa, compared with 8% who completed CBT and 0% who completed IPT. In the intent-to-treat sample, 25% assigned to specialist supportive clinical management, 5% assigned to CBT, and 0% assigned to IPT no longer met diagnostic criteria at posttreatment (16). However, at 7-year follow-up there were no statistically significant differences between the treatments, with 64% in IPT, 42% in specialist supportive clinical management, and 41% in CBT no longer meeting diagnostic criteria (17).

Several large RCTs involving a range of psychotherapies are currently underway (see 11). Modalities under evaluation include CBT, couple-based CBT, exposure and response prevention, specialist supportive clinical management, focal psychodynamic therapy, and two relatively new interventions: cognitive remediation therapy and a social-cognitive interpersonal psychotherapy (MANTRA). Cognitive remediation therapy addresses neuropsychological mechanisms implicated in anorexia nervosa including rigidity and weak central coherence, and MANTRA targets intrapersonal and interpersonal processes that maintain anorexia nervosa by helping individuals to express and process emotions and social relationships.

The first line treatment for adults with bulimia nervosa (or bulimia-like illnesses) is CBT designed specifically for bulimia nervosa (13, 14). CBT for bulimia was developed soon after bulimia nervosa entered the medical literature (18) and is time-limited with a duration of approximately 20 weekly sessions. CBT is also a first line approach for binge eating disorder (13, 14).

One CBT model for bulimia (18) proposes that the disorder is maintained by problems with self-esteem and extreme concerns about shape and weight. These interact to produce strict dieting, binge eating, and compensatory behaviors, which become part of a self-perpetuating vicious cycle. In some individuals, binge eating emerges in rebound from restrictive eating; in others, binge eating arises in the absence of dieting or dietary restraint. CBT addresses cognitive, emotional, and behavioral factors that maintain the disorder. A key first step is the normalization of meals and snacks, which offsets the likelihood of a binge episode, in turn reducing the use of compensatory behaviors. The model has been extended to binge eating disorder (19) with the normalization of eating patterns aimed at countering both restriction and binge eating.

The research supporting a CBT treatment approach is robust (20, 21), hence it is recommended in clinical practice guidelines around the world as the leading treatment for bulimia nervosa (10, 13). On average, approximately 50% of patients with bulimia who undergo a course of CBT attain binge-purge abstinence (22). Treatment normally lasts 12 to 20 sessions held over 3 to 5 months. Antidepressant medication may be an alternative or additional first step (13). Fluoxetine for bulimia nervosa is the only Food and Drug Administration-approved medication for any eating disorder, and although it is effective at reducing binge eating in the short run, it is of unknown long-term efficacy as a sole treatment (23). Guided self-help and pure self-help CBT for bulimia nervosa are also effective in reducing the frequency of binge eating and purging but may be less effective in achieving abstinence from these behaviors than face-to-face psychotherapy (24). Empirically supported self-help is recommended for use in a stepped care approach or when standard CBT is inaccessible.

CBT for binge eating disorder has been tested in individual, group, and self-help formats (25) and is recognized in guidelines as the first line approach (13, 14). Remission rates for binge eating at posttreatment are around 50% (26). In the United Kingdom, self-help is recommended as the first line approach in the treatment of binge eating disorder (13). Pharmacotherapy may also play a role in the treatment of binge eating disorder. Classes of medications showing some efficacy include antidepressants, anticonvulsants, and antiobesity agents (25).

CBT has been recast by Fairburn (27) to apply to all eating disorders (CBT-E) and to address additional maintaining factors (perfectionism, mood intolerance, core low self-esteem, and interpersonal difficulties).

IPT is recommended as a second line approach in the National Institutes of Health and Clinical Excellence and American Psychiatric Association practice guidelines (13, 14). IPT was originally developed in the 1960s for the treatment of unipolar depression, and was adapted for bulimia nervosa in the late 1980s (28) and for binge eating disorder in the 1990s (29). For individuals with bulimia nervosa and binge eating disorder who complete treatment, binge (and purge) abstinence rates are around 50% (30). The IPT model posits that difficulties in interpersonal relationships and functioning contribute to the development and maintenance of psychiatric symptoms. The treatment is time-limited (approximately 20 weekly sessions) and involves three phases: assessment of the interpersonal issues affecting the patient’s symptoms, therapeutic work to help the patient make interpersonal changes in one or more of four IPT areas of role disputes, role transitions, interpersonal deficits, unresolved grief, and last, a termination phase. IPT for bulimia nervosa demonstrates equivalent outcomes to CBT at longer-term follow-up, but achieves its effects less rapidly (21, 30). In a large multicenter trial of CBT and IPT for bulimia nervosa (30), binge eating and purging abstinence in the intent-to-treat sample was 29% for CBT and 6% for IPT at posttreatment, and 22% for CBT and 18% for IPT at 4 month follow-up. Regarding binge eating disorder, the first line of treatment is CBT in a self-help form with direct support from a health professional, followed by standard CBT if self-help is declined. If the illness persists, IPT or dialectical behavior therapy (DBT) modified for binge eating disorder are treatments to consider.

DBT has been less extensively evaluated for bulimia nervosa and binge eating disorder than CBT, but has demonstrated efficacy. DBT acts on the patient’s emotion regulation skills, with the goal of improving coping with emotional distress so that binging and purging are reduced. Treatment is adapted from Linehan’s DBT protocol for borderline personality disorder (31, 32) and detailed in a manual (33), which recommends 20 weekly sessions. DBT has been evaluated among individuals with binge eating disorder in two RCTs, in a group format. In the first, DBT was superior to wait-list, with 89% binge abstinent at posttreatment compared with 12.5% for wait-list, and 56% in the DBT group abstinent at 6-month follow-up (34). The second trial found that DBT was superior to a nonspecific control at posttreatment (64% versus 36%), but there was no difference at 12-month follow-up (64% versus 56%) (35).

A behavioral weight loss (BWL) intervention widely used in obesity RCTs and empirically supported as a self-help intervention for obesity (36) has been tested for efficacy among overweight and obese individuals with binge eating disorder. It has been effective in reducing binge eating and eating pathology, however, CBT and IPT have been shown to be more effective in addressing core eating pathology (37). The BWL intervention involves lifestyle changes in exercise and nutrition with a focus on increased physical activity and moderate caloric restriction to produce gradual, modest weight loss. The intervention lasts for 12 weeks.

Expert consensus suggests that to manage other and unspecified eating disorders, the best approach is to match the clinical presentation to one for which a treatment is indicated, and to treat with the respective first line psychotherapy and/or pharmacotherapy, until new evidence emerges (14). Night eating syndrome is a recently described presentation. Selective serotonin reuptake inhibitors have shown promise for night eating syndrome in two RCTs, but these also indicate a strong placebo response (38). Night eating syndrome has not been the subject of any psychotherapy RCTs. CBT has been suggested as a candidate psychotherapy approach.

For young people with anorexia nervosa, a specific type of family therapy known as family-based treatment (FBT) is recommended (13, 14) and is effective for approximately 50% of patients. No pharmacotherapy interventions have proved efficacious. FBT consists of 20 sessions provided over 6 months, and is detailed in a treatment manual (39). In the treatment, parents are considered the best resources for their child’s recovery and are urged to take an active role in treatment. Treatment has three phases including: 1) parents taking charge of child refeeding and weight gain; 2) transitioning control over eating back to the child in a developmentally appropriate way; and 3) establishing a healthy relationship with parents and addressing developmental issues problematic to the young person.

FBT is the most well-evaluated approach to other psychotherapies, has some evidence of superiority to other psychotherapies and therefore forms the basis for clinical recommendations. Despite promise, the evidence is limited. Of the 11 treatment RCTs for anorexia nervosa in predominantly youth samples (11), nine examined family therapy, but only two compared family therapy to an alternative treatment (the remainder varied an aspect of family therapy delivery between the trial arms) (11). In the largest and most rigorous trial to date (40), there were no differences in remission between FBT and adolescent-focused psychotherapy (AFP) at treatment completion (42% versus 23%), but the proportion remitted in FBT was higher at 6-month (40% versus 18%) and 12-month follow-up (49% versus 23%). AFP, used as the comparison in the largest RCT, is an intervention approach that assists the adolescent in adaptively coping with emotions and developmental challenges. It is based in self-psychology and psychodynamic theory, and considers pathology as arising from unmet developmental needs. The intervention techniques do not target food, weight, or body image cognitions directly, but rather draw on a range of techniques (cognitive, mindfulness, interpersonal, and behavioral) to manage stressors that give rise to symptoms. The FBT evidence base attracts criticism by not yet using comparisons that might be considered reasonable alternatives (i.e., CBT and IPT).

For bulimia nervosa in adolescents, the research evidence and clinical guidance are sparse. The publication of clinical practice guidelines predates the accumulated RCT evidence. FBT modified for bulimia nervosa and CBT guided self-help appear to be reasonable options. FBT is time-limited and involves 20 sessions conducted over 6 months. The phases mirror those of FBT for anorexia nervosa, with slight differences. For example, in the first phase parents are empowered to disrupt binge eating, purging, dieting, and other unhealthy forms of weight control without the focus on refeeding. CBT-guided self-help can be delivered using the same manual used with adults (41). CBT focuses on the normalization of eating patterns to reduce binge eating and managing thoughts, feeling, and situations that serve as triggers for disordered eating behaviors. The completion of daily self-monitoring records and learning tools to challenge cognitive distortions are core components of treatment.

One RCT among adolescents with bulimia nervosa has been conducted. Significantly more patients treated with FBT compared with supportive psychotherapy were binge- and purge-abstinent at posttreatment (39% versus 18%) and at 6-month follow-up (29% versus 10%) (42). In one further relevant RCT, two-thirds of the overall sample had bulimia nervosa and the remainder had eating disorders not otherwise specified (43). The abstinence rate for CBT-guided self-help was significantly higher than for FBT at posttreatment (42% versus 25%), although the effects were the same at 6-month follow-up (52% versus 55%).

No clinical trials have been conducted aimed specifically at treating binge eating disorder in adolescents, although several have shown positive response in reducing binge eating behavior (44–46). One trial is underway comparing CBT with a wait-list control for binge eating disorder in adolescents (47) and one integrating components of DBT, such as mindfulness and distress tolerance skills training, and CBT targeting loss of control eating in adolescents (48).

As for adults, the recommended approach is to use the treatment guidance for anorexia nervosa or bulimia nervosa, depending on which condition the eating disorder most resembles (14).