While the use of 5-HT3 receptor antagonists is clearly justified in patients receiving cisplatin, their role with less emetic drugs is still not defined. The aim of our randomized study was to verify the efficacy of the single standard dose of three 5-HT3-receptor-antagonists in moderately emetic chemotherapies. Sixty chemotherapy-naive breast cancer patients of 30 to 71 years in age, P.S. = 0-1, receiving 5-fluorouracil-epirubicin-cyclophosphamide (FEC 75) q 21 days or cyclophosphamide-methotrexate-5-fluorouracil (CMF) or 120 mg/m2 epirubicin or high dose mitomycin-methotrexate-mitoxantrone (MMM) q 14 days (+ G-CSF) or 100 mg/m2 epirubicin (+ G-CSF) were randomized to receive, 15 min before chemotherapy, 8 mg i.v. bolus of ondansetron or 3 mg i.v. granisetron or 5 mg i.v. tropisetron and no further antiemetic therapy in the following days. 180 cycles were evaluable. Complete protection, (the absence of vomiting episodes,) was respectively 75%, 70% and 70% in the acute and 70%, 82%, 72% in the delayed phases, and an absence of nausea was 56%, 37% and 20% in the acute phase and 50%, 35% and 27% in the delayed, respectively. Complete response, (absence of vomiting and absence or mild nausea,) was 74%, 58.6% and 50.8% in the acute and 64%, 63.7%, 47.3% in the delayed phases, respectively. At the statistical analysis no significant differences between the three drugs were found regarding acute vomiting while ondansetron was superior to granisetron and tropisetron in acute (p = 0.018; p < 0.05) and delayed nausea (P = 0.104; p < 0.01). This activity is practically the same as that we reported (Ann Oncol 1994; 6, suppl 8: 204) with a loading dose on day 1 and maintenance for the following 2-5 days, but with a significantly favorable cost-benefit ratio.