5–HT3 antagonist therapy of bulimia nervosa: A peripherally active agent for a central nervous system eating disorder?


      GASTROENTEROLOGY 2000;119:271-272
      Faris PL, Kim SW, Meller WH, Goodale RL, Oakman SA, Hofbauer RD, Marshall AM, Daughters RS, Banerjee-Stevens D, Eckert ED, Hartman BK (Department of Psychiatry and Surgery, University of Minnesota, Minneapolis, Minnesota). Effect of decreasing afferent vagal activity with ondansetron on symptoms of bulimia nervosa: a randomised, double-blind trial. Lancet 355:792-797.
      Bulimia nervosa, an eating disorder primarily occurring in women, is characterized by recurrent binges of eating followed by self-induced vomiting. Current treatment options include psychotherapy or antidepressant drugs such as the serotonin reuptake inhibitors. Despite these alternatives, the long-term management of bulimia is frequently unsuccessful with nearly 50% exhibiting persistent symptoms more than 5 years after disease onset. Thus there has been a focus on developing therapies that show potential for prolonged care of the affected patient. In the current placebo-controlled study, Faris et al. present data showing short-term efficacy for the serotonin (5-HT3) receptor antagonist ondansetron in reducing symptoms in patients with severe bulimia nervosa.
      Twenty-nine women (mean age, 29.1 years; mean disease duration, 11.8 years) fulfilled the inclusion criteria of at least 7 coupled binge-vomit episodes per week in the absence of other medical illness or weight loss. Patients were not receiving medical or psychiatric treatment and did not take other drugs except for laxatives. All women underwent screening blood testing, electrocardiography, and psychiatric assessment. During the first week, patients recorded eating behavior to establish baseline values. In the second week, the women received placebo in single-blind fashion. Twenty-six patients completed these initial phases and were enrolled in a 4-week double-blind phase in which they took 6 daily capsules of placebo or 4 mg ondansetron. The women were instructed to take the medication whenever an urge to binge or vomit was felt. When urges were constant, patients were told to take the drug 30 minutes before eating. The timing of all meal-related events was recorded daily. A binge was defined as a large amount of food eaten rapidly. A meal was considered abnormal if it was followed within 1 hour by bingeing or vomiting.
      Compared with placebo, ondansetron produced improvements in several parameters of bulimic activity. Weekly binge episodes decreased from 12.8 at the end of the single-blind placebo week to 6.5 after 4 weeks of ondansetron. In contrast, weekly binges did not change with placebo (13.4 vs. 13.2). The difference between ondansetron and placebo was highly significant (P < 0.0001). In addition, ondansetron led to a greater increase in normal meals ingested than did placebo (4.3 vs. 0.2/wk; P < 0.03). The time occupied by bulimic behavior decreased by 7.6 hours per week on ondansetron vs. 2.3 hours with placebo. The investigators concluded that the 5-HT3 antagonist ondansetron shows promise for treatment of bulimia nervosa. They further hypothesized that this class of drugs acts on vagal 5-HT3 receptors to normalize the dysfunctional pattern of meal termination and satiety resulting from aberrant vagal serotonergic neural activity in affected patients.


      Bulimia nervosa is a prevalent condition affecting up to 2%-3% of young women and presents with repeated episodes of binge eating followed by self-induced vomiting in an aberrant set of behaviors to prevent weight gain (DSM-IV, 4th ed., 1994:539-550). Affected individuals experience a loss of control over normal satiety mechanisms and exhibit self esteem that is inappropriately tied to body image. The condition is of relevance to gastroenterologists because many of these patients are referred for evaluation of unexplained nausea and vomiting. The clinician must decide whether gastrointestinal evaluation is warranted before referral to an eating disorders specialist.
      Management of bulimia nervosa has rested both on psychological and pharmacologic treatments. Psychotherapy has shown effectiveness in reducing bulimic activities in several studies (Arch Gen Psychiatry 1995;52:304-312, J Clin Pharm Ther 1999;24:23-31). Antidepressant agents, especially selective serotonin reuptake inhibitors (SSRIs), are the most widely prescribed drugs for the syndrome. In placebo-controlled trials, SSRIs such as fluoxetine and fluvoxamine reduce vomiting and binge eating episodes without significant toxicity via mechanisms independent of their antidepressant effects (Br J Psychiatry 1995;166:660-666, Int J Eat Disord 1999;25:19-27, Am J Psychiatry 1998;155:1756-1762). Despite the initial success for these treatments, bulimia nervosa remains a chronic illness. Fifty percent of individuals exhibit some bulimic behavior, whereas 20% continue to satisfy the full diagnostic criteria for the disorder more than 5 years after diagnosis (Am J Psychiatry 1997;154:313-321).
      In concert with the clinical response to SSRIs, investigators have raised the possibility that bulimia results from abnormal serotonergic function in the central nervous system. Serotonin is an important neurotransmitter in regulating food intake (Psychopharmacol Bull 1997;33:345-354). Consumption of diets deficient in the serotonin precursor tryptophan produces increases in depression, mood irritability, and binge eating in bulimic women compared with healthy volunteers, implying that affected individuals may exhibit exaggerated responses to modulation of serotonergic pathways (Am J Psychiatry 1995;152:1668-1671, Biol Psychiatry 2000;47:151-157). Recovered bulimic patients exhibit higher spinal fluid levels of the serotonin metabolite 5-hydroxyindole acetic acid (5-HIAA) than healthy volunteers, showing the persistence of abnormal serotonergic function even when the disorder is inactive (Arch Gen Psychiatry 1998;55:927-935). Additionally, administration of the indirect serotonin receptor agonist D,L-fenfluramine produces blunted release of the pituitary hormone prolactin in patients with symptomatic bulimia nervosa, suggestive of impaired central neural serotonergic responsiveness (Arch Gen Psychiatry 1997;54:529-534, Neuropsychopharmacology 2000;22:257-263). Recent investigations in patients with eating disorders indicate possible defects at the level of serotonin receptors themselves. Both patients with anorexia nervosa and bulimia nervosa exhibit enhanced platelet 5-HT2A receptor binding, whereas individuals with anorexia nervosa show alterations in selected 5-HT2A receptor genotypes (Eur Neuropsychopharmacol 1999;9:469-473, Neurosci Lett 1999;277:134-136).
      In the present study, Faris et al. used a different approach to the therapeutic modulation of serotonergic function in bulimia nervosa, the use of the selective 5-HT3 antagonist ondansetron. Compared with placebo, ondansetron produced improvements in several objective parameters of bulimic activity including the number of binge-vomit episodes, the time spent in bulimic activities, and the number of normal meals consumed. The study was well-blinded and showed impressive benefits for the 5-HT3 antagonist; however, it was quite brief with active drug administration for only 4 weeks. Because bulimia nervosa is a chronic illness, much longer term investigations are needed. Furthermore, controlled comparisons need to be made between 5-HT3 antagonists and well-accepted therapeutic options such as antidepressants and psychotherapy. Nevertheless, the present study does suggest a novel approach to this difficult condition.
      The investigators postulate that symptom development in bulimia nervosa results from aberrant vagal afferent neurotransmission, which is activated by gut serotonin release. They further theorize that ondansetron acts to blunt this abnormal afferent activity via vagal inhibition. These hypotheses are highly speculative. The evidence suggesting that vagal mechanisms are responsible for bulimic behavior is limited. Measuring respiratory variability in the heart rate, one group has found elevated cardiac vagal tone in bulimic women (J Clin Psychopharmacol 1998;18:26-32). Additionally, ondansetron dissociates the relationship between binge-vomit cycling and perception of noxious somatic stimulation, a phenomenon purported to relate to afferent vagal activity (Pain 1998;77:297-303). Ondansetron has been reported to poorly cross the blood-brain barrier; however, spinal fluid levels may reach 15% of those in the plasma—a level that may not be insignificant (Psychopharmacology [Berl] 1992;109:497-498). Furthermore, many central nervous system structures possess a porous blood-brain barrier, thus a central site of action of the 5-HT3 antagonists is by no means excluded. 5-HT3 receptors are widely distributed in the brain and brainstem, indicating selective antagonists may have actions that are not limited to modulation of peripheral afferent transmission (J Comp Neurol 1998;402:385-401).
      Disturbances of gastric physiology have been described in bulimia nervosa, which lend support to the investigators' contention that ondansetron acts in the peripheral rather than the central nervous system. Delays in gastric emptying and disturbances in gastric pacemaker rhythm are found with bulimia (Am J Clin Nutr 1997;65:114-120, Neurogastroenterol Motil 1998;10:3-10). Furthermore, bulimic patients exhibit exaggerated perception of epigastric fullness and satiety after ingestion of a standard water load, suggesting there may be a peripheral gastric stimulus to facilitate the vomiting that follows binge eating (Neurogastroenterol Motil 1998;10:3-10). In an experimental model in healthy volunteers in which meal-like symptoms were induced by concurrent intestinal perfusion of lipids and distention of the gastric fundus, ondansetron increased the intragastric volumes and pressures needed to elicit a sensation of fullness, indicating that 5-HT3 antagonism can reduce the sensitivity of the stomach to intragastric stimulation (Am J Physiol 1996;271:G591-G597). It is conceivable that one of the actions of ondansetron in bulimia is to reduce the perceptual sensitivity of the proximal stomach to distention caused by the binge eating, thereby blunting a trigger for the vomiting phase of the cycle. This testable hypothesis is worthy of further study.
      It is interesting to note the parallels between current research directions in the treatment of bulimia and the functional bowel disorders such as irritable bowel syndrome. For each condition, antidepressants have become established therapy over the past decade, yet it remains unknown whether their site of action is central or peripheral. More recently, 5-HT3 antagonists have been shown to reduce the prominent symptoms in each disorder via purported actions on peripheral neural activity. These similarities indicate an inseparable interaction between the gut and the brain and suggest that therapies that modulate this gut-brain axis, regardless of their site of action, may exhibit efficacy in a broad range of functional disorders. Despite the relatively short duration of the study and the uncertain mode of action of the test drug, the investigation by Faris et al. is a provocative exploration of a novel therapy to modulate gut-brain interactions in bulimia nervosa. Future studies will define the role of 5-HT3 receptor antagonists in the clinical management of this challenging clinical condition.