We performed a pharmacogenetic study to analyse the correlation between the clinical responses to ondansetron during a 24-h postoperative period and two single nucleotide polymorphisms (-100_-102 delAAG and ins/delCA) in the serotonin receptor type 3B gene (HTR3B) in 280 Korean patients with tonsillectomy (TS, n=108) or surgery because of gynaecological problems (GY, n=172). The effect of CA/CA homogyzous variant of the ins/delCA allele on lowering the prevention of postoperative vomiting was significant, particularly during the initial 2 h postoperative period (P=0.034) and in the GY group (P=0.013 for complete response and P=0.009 for prevention of vomiting). However, we failed to detect a significant effect of genotypes for the -100_-102delAAG allele during our observation period. Our results suggest that ins/delCA polymorphism in the HTR3B may have clinical relevance to the expected clinical response to prevent postoperative vomiting in patients treated with ondansetron.Keywords Serotonin receptor, Ondansetron, HTR3B, Postoperative vomiting, Single nucleotide polymorphismThe postoperative nausea and vomiting (PONV) associated with anaesthesia is one of its most common unpleasant complications. The overall incidence of PONV is about 30%, and about 80% in patients with high risk factors, such as female gender, nonsmoking status, history of PONV or motion sickness, postoperative opioids, and long duration of surgery 1,2 . Using a simplified risk score developed by Apfel and Koivuranta 3 , the incidence of PONV ranged from 10% in patients without risk factors to 79% in patients with the highest score.In addition to the variable incidence of PONV, the efficacy of antiemetic drugs to prevent PONV shows wide interindividual variation because of genetic and non-genetic factors. For example, the incidence of PONV (prophylaxis failure) is high, ranging from 20% to 55% with still relatively wide interindividual variation in patients treated with 5-HT 3 antagonists 4,5 . In spite of some effort to anticipate the prophylactic efficacy of available pharmacological interventions using risk factor-based algorithms, there is no algorithm that completely predicts PONV 6 . In cases of no sufficient effect following initial prophylaxis, the treatment should be expanded to another pharmacological class. If even the most promising antiemetic drug is used as a single prevention, it can reduce the incidence of PONV by no more than about 30% 6 . Therefore, the predictability of PONV after prophylactic pharmacotherapy is important to accomplish patient satisfaction in the perioperative setting.Ondansetron is the most widely used as a selective 5-HT 3 receptor antagonist to prevent PONV 7 , or chemotherapy-induced nausea and vomiting (CINV) 8,9 . Serotonin, which is released from the enterochromaffin cells after general anaesthesia, activates 5-HT 3 receptors distributed in the central nervous system