Serotoninergic innervation may contribute to the control of colonic motility and to visceral sensation from the large bowel. Indeed, ondansetron hydrochloride, a selective 5-hydroxytryptamine type 3 receptor antagonist, has been shown to slow colonic transit in healthy volunteers. Thus, we wished to determine whether 5-hydroxytryptamine type 3 receptor blockade slows colonic and small bowel transit in patients with diarrhea-predominant irritable bowel syndrome (IBS) and whether symptoms would be ameliorated with drug therapy. Of 14 patients with well-established IBS who entered a randomized, double-blind, placebo-controlled crossover pilot trial of 4 weeks of treatment with ondansetron, 16 mg three times daily, 11 completed the study. A minimal "washout period" of 4 weeks (median, 7 weeks) separated the two phases of the trial because patients were required to have similar symptoms before both periods of the study. Colonic transit tended to be longer during drug therapy than during the placebo trial, but this difference was not significant. Small intestinal transit and orocecal transit were unchanged by the drug. The integrated and peak postprandial increases in neurotensin, peptide YY, and human pancreatic polypeptide in serum were not significantly different in the drug and placebo periods. After treatment with ondansetron, stool consistency improved significantly; however, stool frequency, stool weight, abdominal pain, and the symptom criteria for IBS were not significantly altered by the drug. The results of this pilot study suggest that the motor effects expected with 5-hydroxytryptamine type 3 receptor blockade (namely, slowed colonic transit) may be diminished in some patients with IBS. The subjective improvement in stool consistency may reflect changes in the perception of defecation.(ABSTRACT TRUNCATED AT 250 WORDS)