Recurrent binge eating, characterized by eating an unusually large amount of food coupled with a sense of feeling out of control, is the hallmark of two DSM-IV eating disorders—bulimia nervosa and binge eating disorder. In bulimia nervosa, but not binge eating disorder, binge eating is coupled with compensatory behaviors (e.g. self-induced vomiting, laxative or diuretic use, excessive exercise, and fasting). Historically, whereas anorexia nervosa is associated with underweight (
1) and binge eating disorder is commonly, although not universally, associated with overweight and obesity (
2), the prototypical patient with bulimia nervosa presents in the normal weight range (
3).
During the course of an ongoing clinical trial comparing group cognitive-behavioral therapy (CBT) with online CBT for bulimia nervosa, we noticed a deviation from the typical weight presentation, with more individuals presenting for treatment in the overweight or obese weight ranges. Our observation concurs with a report from Spain indicating that the prevalence of obesity has increased threefold in patients with eating disorders characterized by binge eating (
4). Moreover, not only were these patients seeking relief from their binge eating and purging, they also expected treatment to result in demonstrable weight loss.
CBT is the “gold standard” for treatment of bulimia nervosa (
5–
7). Although fluoxetine is approved by the U.S. Food and Drug Administration for bulimia nervosa at a recommended dosage of 60 mg/day (
8), its efficacy in reducing binge and purge episodes in the short term does not convincingly lead to long-term cure. CBT for bulimia nervosa is a multimodal intervention that includes techniques such as psychoeducation, self-monitoring, modifying responses to cues, challenging automatic thoughts, thought restructuring, problem solving, exposure with response prevention, and relapse prevention (
9–
11). In addition to achieving abstinence from binge eating and purging, CBT addresses topics such as body acceptance, eliminating dietary restriction, normalization of eating, uncoupling of emotion regulation from eating, and acceptance of a wider range of foods by moving away from characterizations of foods as “bad” or “good.” Approximately 40%–60% of patients who complete CBT for bulimia nervosa demonstrate significant improvement (
11–
17).
As Butryn et al. (
3) have noted, “Because the body weights of patients with bulimia are generally in the normal range, there is typically no attempt made to modify body weight during treatment.”
CBT does address weight when patients continue to strive for weight loss even when in the normal weight range. The discussion then focuses on developing healthy weight expectations and eliminating dietary restriction. Although CBT has been adapted for overweight individuals with binge eating disorder to promote healthy approaches to weight regulation, these adaptations have not been routinely applied in the treatment of bulimia nervosa.
In the context of the escalating obesity epidemic (
18), however, the body mass index (BMI) weightscape of bulimia nervosa may be changing, with concurrent overweight or obesity complicating the clinical picture and influencing treatment expectations. If our observations of increasing weight reflect a trend for bulimia nervosa patients or a subset of them, modifications of our standard approach to weight issues in bulimia nervosa may be required.
BMI in Bulimia Nervosa
The BMIs of individuals with bulimia nervosa are generally in the normal or high normal range (
2), with lower BMIs associated with a history of anorexia nervosa (
19). In an evidence-based review of clinical trials of bulimia nervosa conducted by the Agency for Healthcare Research and Quality (
20), baseline BMIs (when reported) were most commonly in the normal range, with only a few exceptions, although some studies had BMI-based exclusion criteria (see the
data supplement that accompanies the online edition of this article).
BMI at presentation only tells a partial story, as the clinical trajectory of bulimia nervosa often includes significant weight variability (
21). Presentation BMIs are often lower than self-reported highest adult BMI, indicating a state of weight suppression (
3,
22), defined as the discrepancy between an individual’s highest adult weight and his or her current weight (
23). One study (
3) suggested that greater weight suppression was associated with poorer outcome of bulimia nervosa treatment, whereas another reported no association (
24).
Weight and CBT for Bulimia Nervosa
When patients are in the normal weight range, the topic of weight in CBT for bulimia nervosa is generally approached from the perspective of cognitive distortions and dissatisfaction with a body that is objectively within the “healthy” weight range. CBT for bulimia nervosa generally results in weight change (either gains or losses) within a 1–3 kg range over the course of treatment. Although not considered to be clinically significant by researchers (
25,
26), weight gain in this range may be of considerable personal significance to patients. The nature and magnitude of this significance may differ depending on whether the patient is in the healthy weight range or in the overweight or obese range and entering treatment with weight loss expectations.
Carter et al. (27) reported significant variability in the effect of treatment for bulimia nervosa on weight—with reports of gain, loss, and no change. In a 5-year follow-up study, they observed considerable interindividual differences in posttreatment weight trajectories, with the majority of patients remaining fairly stable but 30% having lost and 18% having gained 5 kg or more. At 5 years, 24% of patients fell into the overweight range (BMI=25–29.9), although mean BMI had been in the healthy range at baseline (BMI=22.7, SD=2.7).
Actual or feared weight gain during treatment is a deterrent to engagement and acceptance of many interventions, both pharmacologic and psychological (
28,
29). As patients recover from bulimia nervosa, they may experience periods of weight gain after cessation of purging when binge eating persists. We encourage patients to be mindful that their metabolism may take time to stabilize after cessation of binge eating and purging and urge them not to launch into extreme weight control behaviors in response to small weight fluctuations.
The urgency, personal impact, and health consequences of weight fluctuations that occur during the course of treatment may be magnified in patients who are overweight or obese. If the clinical presentation of individuals with bulimia nervosa is changing, we may need to revisit our approach to weight, and possibly use our work with binge eating disorder as a guide.
Our Observations
While supervising group therapy and reviewing chat transcripts, we noted a marked difference in the nature of the therapeutic conversation about weight. In contrast to our expectations, therapy topics were not about striving for a societal ideal in normal weight patients; rather, patients were discussing directives from their primary care physicians to lose weight or their personal health concerns about type 2 diabetes, hypercholesterolemia, hypertension, and other medical complications secondary to overweight and obesity. Given that half of the patients were in online therapy, we did not always have visual cues to evaluate their weight status. Were we seeing a shift in the BMI presentations of bulimia nervosa patients? To address this question empirically, we explored the baseline demographic data and verified that the mean BMI of the sample at our site was in the overweight range (BMI=25.2, SD=5.6), with 30% of the sample at presentation in the overweight or obese range—a definite deviation from expectations.
The overweight or obese patients were clearly entering treatment with two explicit treatment goals—abstinence from bulimic behaviors and weight loss. Our intervention, however, was not designed for weight loss. Despite our carefully considered clarification at the outset of therapy that the first focus of treatment for bulimia nervosa is normalization of eating, these participants’ stated or unstated hope was that our treatment would help them lose weight. In fact, in binge eating disorder, where BMI is often higher, CBT is effective in producing abstinence from binge eating but fairly ineffective in producing weight loss (
30,
31).
Matching Patient Expectations With Therapy Deliverables
Our concern was that our treatment was directed toward outcomes that were not congruent with expectations of our overweight or obese patients. Dropout and premature termination is an issue in both clinical trials and community treatment of patients with eating disorders (
32), and weight loss expectations play a role (
29). Dropout rates from CBT for bulimia nervosa range from 6% to 37% (
20). Similarly, self-help trials report dropout rates ranging from 21% to 31% (
20).
A mismatch between the therapist’s and the patient’s treatment goals can also contribute to dropout (
29). If patients are entering treatment with the expectation of abstinence from bulimic behaviors
and weight loss and our treatment only addresses one expected outcome, we run the risk of creating a goal mismatch and ultimately treatment discontinuation and failure if this discrepancy is not addressed explicitly. Furthermore, even if patients do not discontinue treatment, in the absence of therapeutic attention to weight-related issues, they may pursue weight loss outside of therapy. Engaging in unhealthy weight loss through extreme food restriction or food elimination could jeopardize their ability to cease binge eating and purging behaviors despite their active participation in CBT.
Our interventions for bulimia nervosa should consider all relevant health goals and, above all else, do no harm. If patients have weight-related medical morbidities and they gain weight during the course of CBT for bulimia nervosa, then we may be inadvertently exacerbating other health problems. It is unknown to what extent reductions in binge eating and purging influence various metabolic parameters in the absence of weight loss. This is an important research question. We must also ask whether our approach to treatment of individuals with bulimia nervosa who are overweight or obese needs to be modified to address the potential complications of obesity. The nature of the conversation about weight in CBT for bulimia nervosa may have to vary depending on the medical status of the individual by carefully considering health parameters related to current BMI.
Marcus and Levine (
33) emphasize to patients at the beginning of binge eating disorder treatment that typical CBT is not associated with weight loss. In fact, CBT for binge eating disorder helps the patient accept a larger body size by promoting recognition that a larger body can be both attractive and healthy (an ongoing emphasis on physical activity is also helpful in promoting body acceptance). Working toward acceptance of a larger body size is important in helping the patient avoid extreme dieting and unhealthy weight control practices that perpetuate disordered eating, but it does not give license for the patient or provider to ignore health risks associated with obesity. It is likely that we will come under increasing pressure from patients and their primary care providers to develop or adapt CBT for bulimia nervosa that achieves not only abstinence from disordered eating but also incorporates effective lifestyle modifications that support healthy weight regulation.
Approaches to weight control have been developed and tested in the treatment of binge eating disorder, such as appetite awareness training (
34), behavioral weight loss (with appropriate adaptations for individuals with bulimia nervosa) (
33,
35), and appetite-focused CBT (
36). Strategies such as these, which bring appetite and eating into the foreground and directly address weight regulation, may have to be incorporated into “typical” CBT for individuals with bulimia nervosa who have weight-related medical morbidities. Failure to do so may lead to greater dropout, poorer outcome, poorer adherence, and greater patient dissatisfaction with treatment.
Summary and Recommendations
In the absence of guidelines about how best to treat individuals with bulimia nervosa with associated weight-related morbidities, our clinical recommendation is to begin with a comprehensive evaluation of patients’ weight history, including highest and lowest past adult BMI; childhood and adolescent weight and weight concerns; frequency, degree, and duration of weight fluctuations; and deviation of current weight from both highest and lowest adult BMI. Family weight and medical history are also relevant in determining both the likelihood of gaining or losing weight with treatment and obesity-related health risks. It is critical to establish clear communication with the patient’s primary care provider to develop a comprehensive clinical picture that includes any weight-related morbidities and to ensure that recommendations made by the primary care provider related to weight regulation are consonant with the goals of treatment for bulimia nervosa (e.g. not recommending dieting for weight loss).
Leveraging our knowledge about binge eating disorder, the clinician and patient should work together to establish and sequence treatment goals. Sound clinical practice would be to use known CBT tools to reduce binge eating and purging while also introducing approaches that address healthy lifestyle changes and appetite awareness. Incorporating and encouraging moderate (nonobsessive) exercise (for health rather than for weight loss), regular physical activity, and a moderation focus and providing assistance with strategies to help manage appetite and satiety and reduce eating as a means of emotion regulation are the best common-sense tools we have available at this time.
Weight should remain part of the conversation lest dissatisfaction with personal weight goals or distress over weight gain during treatment lead to premature treatment termination or a return to unhealthy compensatory behaviors or extreme weight loss strategies. Also worthy of consideration is consultation with dietitians and certified trainers who are knowledgeable about eating disorders and are able to work within the parameters of CBT. Although not always feasible, collaboration with these practitioners may decrease treatment duration and ultimately lead to a more cost-effective course of treatment.
One additional effect of this shift in patient BMI is potential heterogeneity in therapy group membership. A CBT group for bulimia nervosa that includes both individuals who are overweight and seeking weight loss and individuals who are borderline underweight with a history of anorexia nervosa can lead to unexpected discomfort and therapeutic challenges. Discussions about weight regulation and desires for weight loss outcomes can be a trigger for patients with a history of anorexia nervosa and can fuel urges to restrict. Moreover, for overweight patients, participation in a group with low-weight members could lead to negative body comparisons and self-evaluation that prompt early treatment discontinuation.
Our patients and their evolving clinical presentations drive and motivate the evidence base. Bulimia nervosa in individuals who are overweight or obese may represent the natural evolution of the eating disorder in the context of the obesity epidemic. Indeed, we have seen an increase in the number of former patients with both anorexia and bulimia nervosa, years after recovery from their eating disorder, asking how they can regulate their weight without retriggering their eating disorder. These clinical observations converge to alert clinicians and researchers to a possible weight trend in individuals with bulimia nervosa and challenge us to adapt our interventions appropriately to deal with an evolving clinical presentation.