Background   ABO blood type is widely used. Existing studies have confirmed that the incidence of gastric cancer in people with blood type A blood is higher, but the study of relationship between ABO blood type and the prognosis of gastric cancer is controversial.

  Objective   To explore the relationship between ABO blood type and prognosis of gastric cancer.

  Methods   From January 2005 to December 2008, gastric cancer patients received radical gastrectomy in department of general surgery in Chinese PLA General Hospital were included in this study. Patients were then divided by the blood type. Comparisons of clinical information and pathological features were taken between the groups. They were followed up until December 2014 or until death. The survival and cumulative incidence of death were analyzed according to Kaplan–Meier method. Cox regression analysis was used in the screening of prognostic risk factor (pT, pN, pM were adjusted in model 1, and pTNM was adjusted in model 2).

  Results   Of the 667 patients, there were 200 patients with blood type A, 207 patients with blood type B, 61 patients with blood type AB, and 199 patients with blood type O. No statistical difference was found in age, sex, pT, pN, pM, pTNM stage, degree of tumor cell differentiation, drinking history and family cancer history among patients with different blood types; the proportion of stage III in patients with blood type AB was higher (AB: 65.6% vs A: 52.2%, B: 54.6%, O: 50.8%, P=0.486), but the differences were not significant. Proportion of smokers was lowest in patients with blood type A and highest in blood type AB (A:22.5%, B: 32.9%, AB: 41.0%, O: 32.2%, P=0.018). Survival curves indicated that the cumulative survival rate of patients with blood type A and AB were lower than that of patients with blood type B and O, but both were without statistically significant difference (P＞0.05). Using blood type O as the reference, multivariate Cox regression suggested that the patients with blood type A had worse prognosis (Model 1: HR=1.432, 95% CI=1.082 - 1.897, P=0.012; Model 2: HR=1.329, 95% CI=1.005 - 1.756, P=0.046). Likewise, multivariate Cox regression using blood type B as the reference also suggested that the prognosis of the patients with type A blood was worse (Model 1: HR=1.549, 95% CI=1.170 - 2.049, P=0.002; Model 2: HR=1.412, 95% CI=1.071 - 1.863, P=0.015). The prognosis of the AB blood group was worse than that of the O group (Model 1: HR=1.341, 95% CI=0.912 - 1.973, P=0.136; Model 2: HR=1.140, 95% CI=0.776 - 1.676, P=0.504) and B group (Model 1: HR=1.450, 95% CI=0.985 - 2.134, P=0.059; Model 2: HR=1.212, 95% CI=0.827 - 1.776, P=0.324), but the differences were not statistically significant.

  Conclusion   This study shows that gastric cancer patients with blood type A have poor prognosis compared to those with blood type O and B. The influence of blood type on the prognosis should be further studied by expanding the sample size in different race.