On the Trail of Treatment for Anorexia Nervosa

Considering the seriousness of anorexia nervosa (AN), finding ways to develop effective treatments is critical. When the Agency for Healthcare Research and Policy commissioned a comprehensive study of treatments available for AN, the authors found the overall results were weak and inconclusive (Int J Eat Disord 2007; 40:310).

A number of obstacles to conducting research on effective treatments for AN have been linked to the very nature of the disorder itself. Thirteen studies were rated as “poor” because of such omissions as failure to describe randomization, absence or failure to report adverse events, lack of controls, and failure to report sources of funding.

As Cynthia M. Bulik, PhD, and her colleagues reported, their group at the University of North Carolina researched six major databases for studies on the treatment of AN published from 1980 to 2005. The researchers identified 32 studies published in 35 articles that addressed the efficacy of treatment for patients with AN.

Medication trials

Medications for AN included second-generation antidepressants, tricyclic antidepressants, hormones, and nutritional supplements. According to the authors, no pharmacologic intervention for AN currently has a significant impact on weight gain or on the psychological features of AN. Two trials used fluoxetine at different stages of refeeding. In the first study, when patients who had achieved weight restoration to at least 65% of ideal body weight, and were given fluoxetine, 60/mg/day, no significant differences emerged between fluoxetine and placebo on weight gain, psychological features, or depression or anxiety measures. In the second study, inpatients were randomly assigned to treatment with either fluoxetine or placebo before discharge, with a beginning dosage of 20 mg/day, adjusted over 52 weeks to a maximum of 60 mg/day. At the endpoint, patients on medication did not differ significantly from those on placebo on eating, psychological, or biomarker measures (Biol Psychiatry. 2001;49:644). The dropout rate was high47% dropped out of the fluoxetine group and 85% dropped out of the placebo group.

As for tricyclic antidepressants, in a study by Halmi et al. comparing amitriptyline, cyproheptadine, and placebo (Arch Gen Psychiatry. 1986; 43:177), cyproheptadine was associated with faster weight gain, higher caloric intake, and less depressed mood than in the other groups. In the hormone studies, Miller et al (J Clin Endocrinol Metab. 2005;90:1428)) studied 38 female patients aged 18 to 50 who were given 3 weeks of transdermal testosterone (150 or 330 mg). At the end of the study, those who received testosterone had less depressed mood and over time depressed mood increased less in those treated with testosterone. When Hill et al. (J Child Adolesc Psychopharmacol. 2000;10:3) gave 15 mg/kg/day of human growth hormone to 14 female and 1 male inpatients, the medication led to fewer days to normal orthostatic heart rate response to standing challenge than among those in a placebo group. Another group compared estrogen/progesterone (0.625 mg Premarin or 5.0 mg Provera/day with a nonmedication control group (J Clin Endocrinol. 1985;80:898). After six months, there was no difference in bone density between the two groups; 14% of those in the hormone group and 4% in the placebo group dropped out of the study.

As for nutritional supplements, among 54 women inpatients older than 15 years of age 14 mg of zinc per day was associated with an accelerated increase in body mass index (BMI, mg/kg2) compared to placebo (Int J Eat Disord. 1994;15:251). However, the dropout rate was high: 39% in the zinc group and 32% in the placebo group left the study.

Behavioral interventions

For adults with AN, the authors found tentative evidence that cognitive behavioral therapy (CBT) reduces relapse risk after weight has been restored. In contrast, they could not determine whether the CBT approach is more helpful than others for acutely underweight patients; one study showed that manual-based form of nonspecific supportive clinical management was more effective than CBT and interpersonal psychotherapy (IPT) in terms of global outcomes during the acute phase.

Family therapy as it is now practiced has no supportive evidence of effectiveness for adults with AN and a comparatively long duration of illness, according to the authors. Overall, they found that family therapy that focuses on parental control of renutrition is helpful for treating younger AN patients with nonchronic disease. Two studies suggested that family therapy was superior to individual therapy for adolescent patients with AN who have shorter-term illness.

Problems with the eating disorders literature

Dr. Bulik and colleagues noted that study sample size was often too small to draw conclusions, and attrition was often a problem. This is especially problematic with AN patients, where the illness is marked by denial, fear of weight gain (the key treatment outcome), and hesitation to take medications. The authors also found frequent problems with statistical design and analysis, both in the planning stage and in the study itself. Follow-up was also often poor and outcome information not provided. Most studies involving AN patients are small, single-site trials, and the average sample size is 23 patients. Dr. Bulik and colleagues suggest that clinicians might work in partnership with insurance companies to enable such trials, and to help overcome the high dropout rates and to improve generalization results.

Future directions

Dr. Bulik and her colleagues note that some critically needed steps include discovering new medications that target the core biological and psychological features and the diverse medical side effects of the illness. Methods to enhance motivation and retention in trials are also needed. Future trials of novel medications that can target core biological and cognitive features of AN, and that are acceptable to patients, will also be important.

Future clinical trials may benefit from the use of e-mail, the Internet, personal digital assistants, text messaging and other technological advances, which may be well suited to eating disorders, which are so often marked by shame and denial. Using such technologies may also be helpful in geographic areas where specialty care is limited. According to the authors, improvements for the future would also include larger numbers of participants in each study, multiple study sites, clear delineation of age, race, and ethnicity of study participants, and interventions tailored to the core pathology and medical outcomes of AN.