The serotonin-receptor (5-HT3) antagonists combined with dexamethasone are considered the antiemetic therapy of choice in the prevention of cisplatin-induced emesis. As there are now several compounds on the market, the dilemma of preference is particularly relevant. In preclinical studies some differences among the three marketed drugs (ondansetron, granisetron and tropisetron) have emerged. In particular, tropisetron and granisetron have a greater potency and duration of action and seem to have a greater selectivity toward the 5-HT3 receptor with respect to ondansetron. Furthermore, while with tropisetron and granisetron there is a linear dose/response relationship, this does not seem to be the case for ondansetron. These preclinical differences, however, do not seem to correlate with the clinical antiemetic activity of these compounds. In fact, although the number of comparative studies is small, with all of them presenting several shortcomings (small number of patients, not blinded studies, no association with steroids, sponsored trials), it seems that the antiemetic activity and tolerability of ondansetron, granisetron and tropisetron is very similar. If these data are confirmed, the least expensive of the 5-HT3 antagonists should be the drug of choice. We feel, however, that the answer to this rather difficult question of choice will come from very large, independent, methodologically correct studies designed to show small but clinically significant differences (i.e., less than 10% in complete protection from emesis). These trials, which require about 1000-1500 patients, are ongoing and the one carried out as a multicenter study by the Italian Group for Antiemetic Research is close to conclusion.