Treatment-Resistant Bulimia Nervosa: Clinical Implications and Future Directions

Dialectical behavior therapy may be particularly effective in mitigating biologically-driven vulnerabilities, and zonisamide and lamotrigine may play a role. Bulimia nervosa (BN) is a serious disorder characterized by recurrent large-volume eating episodes that are marked by a loss of control (binge eating), regular compensatory behaviors that are intended to prevent weight gain (purging), and over-valuation of body shape and weight. BN is associated with psychiatric comorbidity, significant psychosocial impairment, medical complications, and increased mortality.  Clinical guidelines for the evidence-based treatment of BN highlight first-line, outpatient psychotherapeutic and pharmacological approaches. Self-help, enhanced cognitive-behavioral therapy (CBT), and interpersonal psychotherapy are considered first-line psychotherapeutic interventions. Antidepressants, most notably SSRIs and specifically fluoxetine, which is FDA-approved for use in BN, are recommended adjunctive first-line treatments. Limited data suggest that CBT-BN alone and CBT-BN plus antidepressants are more efficacious than antidepressants alone.

Despite these available treatment options, roughly 30% of cases are chronic and unremitting. Even for those who do initially respond to treatment, estimated rates of relapse in BN range from 25% to 63%.

Melia is a patient with treatment-refractory, complex BN, who presents for treatment at our eating disorders program. A 20-year-old college student, she plans to attend law school after graduation. Her eating disorder began in her freshman year, when she started “using food to cope” after a difficult breakup and domestic violence. Before admission, she was eating salads or diet frozen entrées for most meals and binge eating and purging 3 to 5 times a day. She has a history of self-harming behaviors and 2 suicide attempts. She reports frequent  shoplifting, occasional binge drinking to the point of blacking out, and recurrent unsafe sex with men she meets at bars. She previously had attempted outpatient, partial hospital, and residential treatment but had been unable to sustain recovery. She also had several trials of
antidepressant medication without response for depression, anxiety, and eating-disordered symptoms.  Melia reports that she feels “very stressed” at school and has high standards for herself. She has had to withdraw from school several times to pursue treatment.

Improving outcomes for individuals with treatment-resistant BN such as Melia hinges on:

  1. Developing an understanding of what factors make treatment failure or dropout more likely
  2. Generating hypotheses about the mechanisms of non-response
  3. Proposing and evaluating treatments that address these hypothesized mechanisms

Factors underlying treatment failure
In general, first-line psychotherapeutic treatments—whether delivered in person or via telemedicine—result in remission in only 30% to 50% of adult patients with BN. These first-line treatments are often particularly ineffective for the large subset of patients who struggle with a range of other dysregulated behaviors. To date, despite some inconsistency in the literature, predictors of worse outcome or treatment dropout include longer duration and more severe eating disorder symptom profiles, comorbid psychopathology such as depression or substance use, and characteristics of personality pathology such as affective instability and impulsivity. Psychiatric comorbidity. Patients with BN have high rates of comorbidity with other psychiatric conditions, including mood, anxiety, and impulse-control disorders. Impulsive behaviors and comorbid disorders defined by behavioral dysregulation are common in BN, including substance use disorders (approximately 37%), ADHD (about 35%), borderline personality disorder (approximately 28%), self-injurious behavior, risky sexual activity, and shoplifting. Moreover, BN is associated with suicidality independent of comorbidities. The range and extent of co-occurring psychopathology are associated with poorer treatment outcome and thus can present profound challenges to the application of evidence-based treatment. Unfortunately, strictly controlled treatment trials often exclude those with significant comorbidity, which leaves a void in our understanding of best practices for our most complex patients.

Pervasive emotion dysregulation. Individuals with BN report significantly higher levels of emotion intensity, deficits in emotional awareness, and increased use of dysfunctional strategies for emotion regulation. Impulsive behaviors may function to regulate dysphoric, extreme mood states. Paradoxically, these behaviors may evoke more negative emotion, which leaves the patient trapped in a cycle of dysregulated emotions and behaviors. Limited development of effective, evidence-based medication strategies. Another potential factor in poor treatment response may be the lack of development of medication strategies that might help patients regulate affect. Randomized controlled trials in the 1980s and early 1990s established antidepressants as superior to placebo in reducing symptoms of BN. The largest studies by far involved fluoxetine—hence, the FDA indication. Subsequently, the antiepileptic drug topiramate was shown to treat BN symptoms more effectively than placebo. It is important, however, to understand the nature of these trials in determining clinical applicability. Positive response was often defined as a 50% or more reduction in weekly binge and/or purge episodes. While impressive, this decrease left many patients still meeting DSM criteria for BN at completion of the studies, and full remission rates were much lower.  

Up to one-half of patients who present for treatment of BN would have been excluded from the studies, primarily for comorbidities. Finally, most trials were small and short term. This may be especially concerning with regard to topiramate, which has fairly widespread use in BN despite well-known tolerability issues and the potential adverse effect of unhealthy weight loss in BN patients striving for thinness. It may surprise some clinicians to know that this approach is based on randomization to topiramate or placebo of only approximately 130 patients, for a period of just 10 weeks. Despite the apparent limited benefit of medications in BN, there has been virtually no research in this area for more than a decade.

Potential mechanisms of nonresponse
Improved understanding of the neurobiology and associated features of BN may provide both an explanation for poor response to evidence-based treatment, and a framework for improved treatment approaches. Impaired inhibitory control and aberrant reward responsivity and reward-based learning may contribute to the affective dysregulation, binge eating, purging, and other impulsive behaviors often reported by individuals with BN. Extant neuroimaging findings in BN suggest abnormal functioning of the corticostriatal and corticolimbic circuits that support these processes, and this altered functioning has been linked to increased bulimic symptom frequency. Dysfunction in these circuits may leave patients with powerful urges to inappropriately initiate or continue eating or to purge, and, in combination with impaired learning, may interfere with response to treatment.

Strategies for treatment- resistant BN
Designing successful interventions requires consideration that abnormal neurobiological functioning may contribute to BN and associated behaviors, and these behaviors appear to serve the purpose of temporarily reducing intolerable emotional intensity. Specifically, treatments for refractory BN may include a focus on skills and strategies designed to help patients regulate emotions and cope with altered inhibitory control and reward sensitivity. 

Developments in psychotherapeutic strategies.
Psychotherapeutic interventions for treatment-resistant BN may specifically focus on skills that help patients cope with their dysregulated emotions, thereby reducing impulsive behaviors. Recent treatments developed to address emotion dysregulation within BN include integrative cognitive-affective therapy, adaptations of CBT, and dialectical behavior therapy (DBT). DBT originally was developed for patients with intense emotion dysregulation, borderline personality disorder, and/or suicidality, and has been adapted and studied for use with eating disorders. With its heavy emphasis on skill development for tolerating distress and regulating emotions, along with its structured approach to treating complex comorbidities, DBT may be particularly effective in mitigating biologically driven vulnerabilities to unstable affect and maladaptive, impulsive behaviors.

Emerging pharmacological strategies.
Considering the prevalence and morbidity of BN, along with the limited efficacy of antidepressants and safety concerns about topiramate for many patients, the lack of new research into pharmacological treatments during the past decade and present day is notable. Recently, a few encouraging but very small open series and case reports have been published. Some of these have utilized antiepileptic drugs—namely, zonisamide and lamotrigine, which warrant further exploration for treatment-refractory BN. Similar to topiramate, zonisamide has an adverse-effect profile that may limit its use in patients with eating disorders. Lamotrigine, with its relative weight neutrality, greater tolerability, and demonstrated benefit in other conditions characterized by affective dysregulation and impulsivity (eg, bipolar disorder), may hold the most promise. The decreases in emotional and behavioral dysregulation associated with lamotrigine may blunt or delay impulsive urges enough to allow patients to use skills, ultimately permitting engagement with empirically supported psychological treatments. However, controlled trials of lamotrigine are needed. Recent FDA approval of lisdexamfetamine (a CNS stimulant) for binge eating disorder has heightened interest in potential use of such medications in BN. Some investigators hypothesize that CNS stimulants may address inhibitory control deficits or deficient reward systems in dysregulated eating disorders in a similar manner as in ADHD. A retrospective series published earlier this year showed large reductions in binge and purge symptoms in 6 patients with chronic BN. While this is intriguing, substantial risks associated with use of these controlled substances in dysregulated BN may make CNS stimulants prohibitive.  The relatively weight-neutral atypical antipsychotic aripiprazole showed benefit as augmentation to antidepressants in 2 small groups of treatment-refractory BN case reports. It is interesting to speculate whether the partial-agonist properties of aripiprazole may improve stabilization of dopamine/serotonin neurotransmitter systems, which have shown abnormalities in BN.

Treatment of comorbidities.
Although consensus exists among practitioners that co-occurring disorders must be addressed in BN treatment, when and how to do so remains a subject of much debate and is a critical consideration in treatment planning. Medication management of comorbidities such as MDD, bipolar disorder, and anxiety disorders is beyond the scope of this article, but is an important aspect of comprehensive care that may improve the capacity of BN patients to engage in treatment. Evidence-based protocols (eg, CBT for depression, prolonged exposure for PTSD) exist for most comorbidities, but it can be difficult to know how to integrate those treatment protocols with the treatment of the primary eating disorder. Treatment can take place sequentially, or it can follow an integrated approach that treats all conditions concurrently. Sequential treatment may be advantageous in that it allows the clinician to address safety concerns presented by the eating disorder first, may reduce barriers to the effectiveness of the other treatment, and may be logistically more practical. However, treating one cluster of symptoms may result in worsening of other behaviors, and this cycling may prevent recovery from both disorders. Thus, more recently, concurrent or integrated treatments across comorbidities have become the preferred approach. Regardless of the chosen timing, comorbid conditions should be addressed with evidence-based treatment.

For Melia, the combination of DBT and lamotrigine was very effective. Melia had begun to learn DBT skills in past treatment episodes, but reported difficulty applying the skills because her impulsivity was so strong that she felt there was “no time” between an urge and a behavior to attempt to use the skills. With the addition of lamotrigine, she still experienced urges to binge, purge, shoplift, and self-harm, but she demonstrated increased ability to delay urges long enough to call her therapist for skills coaching. DBT hierarchy helped her therapist organize and prioritize Melia’s treatment targets. DBT has a focus on helping the patient build a “life worth living” outside the disorder. For Melia, this was difficult because she felt that anything short of immediately completing school with top grades and going to law school would be failure. During treatment, she was able to be less judgmental of herself, and she got a temporary job as a paralegal while she focused on recovery. Her depression and anxiety also improved, and she gained more insight into balancing career and other life goals. Ultimately, she graduated college with plans to attend law school.

Despite significant progress in our understanding of the biological substrates of BN during the past decade, traditional evidence-based treatments fail to help many patients achieve full remission of symptoms. Development and rigorous evaluation of new treatment approaches—particularly those that target characteristics of chronic, treatment-refractory BN—remain crucial tasks in decreasing suffering for patients such as Melia.