We investigated the efficacy of granisetron and dolasetron in preventing postoperative nausea and vomiting. Because the metabolism of the various antiemetic 5-hydroxytryptamine type 3 (5-HT3) antagonists involves different isoforms of the hepatic cytochrome P450 system, we examined the relationship between the clinical efficacy of these drugs and polymorphic cytochrome P450 2D6 (CYP2D6) genotype. This prospective, randomized, double-blind study involved 150 adult patients with a moderate to high risk for postoperative nausea and vomiting. All subjects received dexamethasone at induction of anesthesia followed by either 12.5 mg of dolasetron or 1 mg of granisetron. We analyzed the number of complete responders (no vomiting or rescue medication) during the first 24 hours after surgery. CYP2D6 genotyping was performed using a TaqMan real-time polymerase chain reaction. A complete response was more frequent in the granisetron group (54.7%) compared with the dolasetron group (38.7%, P < 0.05). In subjects receiving dolasetron, carriers of the duplication of the CYP2D6 allele predicting ultrarapid metabolizer status had more frequent vomiting episodes (P < 0.05) than patients in the granisetron group. It is postulated that the difference in the antiemetic efficacy between two investigated 5-HT3 receptor antagonists may be associated with differences in the carrier status for the duplication of the CYP2D6 allele.