Tackling Relapse Among Anorexia Nervosa Patients

Decreased motivation was a
significant predictor of relapse.

Relapse is a common problem among patients with AN, with relapse rates ranging from 9% to 65%. The risk is highest during the first 4 to 12 months after treatment ends.

To identify variables that could predict relapse among AN patients, Dr. Jacqueline C. Carter and a team at the University of Toronto studied 100 consecutive patients who met Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria for AN, based on Eating Disorder Examination interviews (Psychiatry Res. 2012; 220:518).

The patients were admitted to what the authors describe as a “hybrid” inpatient/day treatment unit at a general hospital in Toronto. AN patients are treated by a multidisciplinary team that focuses on nutritional rehabilitation through supervised meals, weight restoration to a body mass index (BMI) of 20 kg/m² for 2 weeks, eradication of binge eating and purging, and group psychotherapy. The psychotherapy component includes cognitive behavioral therapy, dialectical behavior therapy skills training, and interpersonal therapy. When patients reach a BMI of 18, they are transferred from the inpatient unit to day hospital treatment. In this study, the 100 patients had a mean age of 25.4 years and a mean BMI of 15.1 when they were admitted to the inpatient unit. They had been ill for a mean of 6.3 years, and the mean age at onset of AN was 19.4 years.

Relapse was defined as a BMI <17.5 or at least one episode of binge-purge behavior per week for 3 consecutive months during the year of follow-up. The risk of relapse was greatest between 4 and 9 months post-discharge. After this time, the survival curve began to level off, suggesting lower risk of relapse. Overall, 41% of the patients met the relapse criteria during the 1-year follow-up period. The authors noted that, more specifically, 28% of the group had BMIs <17.5 for 3 consecutive months, and 20% reported at least one episode of binge eating or purging per week for 3 consecutive months during the follow-up period. A small subgroup—7%–met criteria for both weight relapse and binge-purge relapse.

Predictors of relapse before and after treatment

At admission, 4 significant predictors of relapse were noted: (1) having the binge-purge subtype of AN; (2) having a history of childhood physical abuse; (3) higher scores on the EDE-Q Eating Concern scale, and (4) higher scores on the Padua Inventory (PI) Checking Behavior scale, a measure of obsessive-compulsive disorder symptoms. Several variables during treatment predicted relapse: decreased motivation, increased weight concern, and a pre-post- treatment decrease on the Rosenberg Self-Esteem Scale. After all significant predictors of relapse were entered into a multivariate Cox regression, decreased motivation from pre-treatment to 4 weeks emerged as the only significant predictor of relapse.

Among patients who did not have relapse, 25% maintained a BMI >19.5 for an entire year after treatment, and most had restricting subtype AN. The rest of those patients had BMIs that fell between 17.5 and 19.5, and they could not really be considered well at the 1-year follow-up point.

Clinically significant predictors. The authors identified a number of clinically significant predictors of relapse. First, patients with binge-purge type AN were twice as likely to have a relapse as those with restricting subtype AN. The second predictor of relapse was the level of motivation to recover. When patients’ motivation to recover fell during the first 4 weeks of inpatient treatment, the risk of relapse rose. The authors noted that while it is not known specifically why some patients feel less motivated early in treatment, clinical experience suggests that persons with AN who need inpatient treatment are often frightened by their poor medical status at the beginning of treatment, which enhances their motivation to change. However, with refeeding, and a noticeable improvement in their medical status and the knowledge that they are out of danger, their motivation may decrease and ambivalence about change increase.

The third predictor identified in the study was higher pre-treatment severity of checking behaviors, as reported on the PI. Although this test did not specifically target or assess checking behaviors related to eating or weight, it is possible that patients with higher scores on this measure were also more likely to engage in AN-related checking, and this may have made them more susceptible to relapse.

Ambivalence about recovery

Many authors have reported that one of the major challenges inherent in treating patients with AN is the marked ambivalence about recovery (Int J Eat Disord. 2008; 41:368). In the Toronto study, the pretreatment level of motivation was not significantly predictive of relapse, but decreases in level of motivation from pretreatment to 4 weeks as well as the pre-treatment level of motivation were. Dr. Carter and colleagues suggest that maintaining and enhancing the level of motivation to recover during the early stages of acute treatment may have a more important impact on long-term outcome than the patient’s initial level of readiness.

Better relapse prevention interventions needed

Evidence-based relapse prevention interventions for patients with AN are sorely needed, according to the authors. In the authors’ study, a significant rate of relapse followed a successful response to specialized inpatient treatment. This suggests that acute inpatient/day treatment alone may often not be enough to foster lasting behavioral changes that will sustain normal weight and abstinence from binge-purge symptoms. Individuals with binge-purge subtype AN appear to be particularly susceptible to relapse. For these patients, it may be helpful to continually focus on developing alternative emotion regulation skills to prevent return of binge-purge behavior–even after the effects of starvation and underweight have been reversed. In addition, motivational interviewing strategies should be integrated into all stages of treatment, including relapse prevention. Increasing the level of motivation to recover during acute treatment may have a more important impact on long-term outcome than the initial level of motivation. Finally, the study results suggest that the presence of obsessive-compulsive symptoms, especially checking behaviors, may be one more element in susceptibility to relapse.