The serotonin (5HT3) antagonist ondansetron was compared in a randomised study with metoclopramide and dexamethasone for the prevention of chemotherapy induced emesis. Thirty children aged 1-15 years with acute lymphoblastic leukaemia received 'intensification modules' according to the MRC United Kingdom acute lymphoblastic leukaemia regimen UKALL XI. This contains the moderately emetogenic drugs daunorubicin, etoposide, and cytarabine. Fifteen children received an intravenous loading dose of ondansetron followed by intravenous or oral doses 12 hourly for five days. Fifteen children received intravenous metoclopramide every six hours for three days with a loading dose of dexamethasone, repeated every eight hours for three days intravenously or orally. Efficacy was assessed by a diary card documenting the incidence of nausea, retching, or vomiting. In the 24 hour period after starting chemotherapy, ondansetron was more effective, with a complete or major response rate of 93%, compared with 33% using metoclopramide/dexamethasone.