Pharmacological Treatment of Bulimia Nervosa
Pharmacological Treatment of Bulimia Nervosa
Bulimia nervosa (BN) is a disorder with a complex cause. The disorder is most commonly seen in women, generally with onset in adolescence. It is characterized by binge eating (consumption of an unusually large amount of food accompanied by feeling a loss of control), inappropriate compensatory measures to prevent weight gain (including purging by vomiting, abuse of laxatives and diuretics, strict dieting, and excessive exercise), obsessive fears of being fat, and negative self-evaluation influenced largely by body shape and weight.1 BN is often associated with anxiety, depression, obsessionality, substance abuse, and poor impulse control.
Antidepressants are the mainstay of pharmacological treatment for BN. In the 1980s and early 1990s they were established as more effective than placebo in treating BN in double-blind, placebo-controlled trials that tested a variety of medication classes, including tricyclic antidepressants, monoamine oxidase inhibitors, and more recently, SSRIs.2 Success was most commonly measured by a decrease in the number of bingeing and purging episodes, and although the studies varied widely, many patients achieved short-term reductions of 50% to 75% in one or both areas.3 Much less attention has been focused on whether antidepressants are as useful in reducing mood and impulse control problems often comorbid with BN.
While the observed reductions in eating-disordered behaviors were significant, it is important to realize that many patients who responded to medication continued to binge and purge at a frequency that still met criteria for a diagnosis of BN at the end of the studies. Only a minority of patients attained remission of the bingeing and purging symptoms.4
Another caveat concerns the clinical applicability of trial data, because some have estimated that one-third to one-half of patients who present for treatment of BN would have been excluded from treatment studies because of accompanying substance abuse, personality disorders, current or earlier use of psychotropic medication, and other factors.5
With their relatively benign adverse-effect profiles, SSRIs are the first-line pharmacological treatment of BN.3,6 Fluoxetine is the most studied of the antidepressants in this country, and the only medication approved by the FDA for treatment of BN. The indication was granted in 1996, after 2 large, multicenter, controlled trials convincingly demonstrated that fluoxetine was more efficacious than placebo for treatment of BN.
The first, an 8-week study that included nearly 400 patients, showed a response at a relatively high dosage of 60 mg/d, while a dosage of 20 mg/d was no different from placebo.7 The second trial, with a duration of 16 weeks, revealed that a starting dosage of 60 mg/d was generally well tolerated in this patient population.8
Trials with other SSRIs have been scant. Several conducted with fluvoxamine had problems with tolerability and mixed results regarding efficacy.9-11 In a large randomized trial involving 267 patients, fluvoxamine was no better than placebo for short- or long-term (1 year) treatment of outpatients with BN. The investigators acknowledged that the dosage range of 50 to 300 mg/d may have been too low to show a treatment effect with this illness; however, titration to higher doses may have led to even more adverse events.10 Overall, there is little evidence for the use of fluvoxamine as an SSRI of choice for the treatment of BN.
Results with sertraline have been more promising, although the results of only one small randomized trial with 20 patients have been reported. Of interest, significant improvement versus placebo occurred with only 100 mg/d, in contrast with the relatively high doses required for fluoxetine.12 The reason for this finding is unknown, but replication on a larger scale would be beneficial.
Little work has been undertaken in controlled trials to compare various SSRIs for BN. One randomized controlled trial of 6 weeks' duration that involved 91 inpatients was designed to explore a genetic basis for drug response. No such basis was elucidated, but the authors reported that fluoxetine, fluvoxamine, citalopram, and paroxetine were all superior to placebo.13 Findings from a smaller controlled study in which patients were randomized to either citalopram or fluoxetine indicate that both groups had significant improvement in eating psychopathology.14