Background: Serotonin receptor antagonists are recommended by the American Society of Clinical Oncology for the prevention of acute and delayed chemotherapy-induced emesis. However, the most effective agent in this class of antiemetic drugs for preventing emesis has not been clearly defined. We therefore performed a double-blind, crossover, randomized, controlled trial comparing the efficacy of granisetron and ramosetron, using patient preference as the primary endpoint. Methods: Thirty patients receiving two courses of combined chemotherapy (including â§60 mg/m2 cisplatin) for gastric or esophageal cancer were randomly assigned to the granisetron-ramosetron group (treatment phase 1: granisetron, 3 mg; treatment phase 2: ramosetron, 0.3 mg) or the ramosetron-granisetron group (treatment phase 1: ramosetron, 0.3 mg; treatment phase 2: granisetron, 3 mg). All patients received methylprednisolone sodium, 250 mg i.v., during each treatment phase. Results: The efficacy of granisetron and ramosetron was similar in terms of the suppression of emesis and appetite status. However, the majority of patients (19/30, 63.3%) expressed a preference for granisetron, as compared with 9 patients (30.0%) who preferred ramosetron; 2 patients (6.7%) had no preference (Ï2 test: p = 0.008; Fisherâs exact test: p = 0.015). Conclusions: (1) A significant proportion of patients prefer granisetron over ramosetron for the prevention of chemotherapy-induced emesis. (2) Granisetron and ramosetron possess similar effectiveness for the suppression of emesis. (3) The variable of âpatient preferenceâ should be accepted as a primary endpoint of antiemetic drug efficacy.