Effect of Perioperative Blood Transfusion on Prognosis of Gastric Cancer

Retrospective Evaluation Using the Proportional Hazard Model of Cox

Abstract

Recently many studies have indicated that allogenic blood transfusion can cause immunosuppression and have adverse effects on survival for malignant tumors. Therefore we undertook a retrospective study to determine if such perioperative blood transfusion adversely affects survival rates in patients with carcinoma of the stomach. An overall comparison of transfused (n=145) versus non-transfused (n=102) patients by the log-rank and the generalized Wilcoxon analysis revealed a statistically significant adverse effect of blood transfusion on survival (p=0.0040, p=0.0014, respectively). However, in contrast, the Cox proportional hazard regression analysis, adjusted for other prognostic variables, showed that blood transfusion was no longer prognosticly significant. According to further detailed investigations, this transfusion effect was not evident in the patients with stage I cancer, but in contrast to patients with stage I cancer, survival rates for stages 2, 3, 4 not given transfusions were significantly higher than those for patients receiving transfusions (p=0.0177, p=0.0057). When we applied the Cox analysis to patients with stage 2, 3, 4 cancers, perioperative transfusion became a sensitive prognostic indicator of cancer death (risk ratio of 2.182 for death).
Consequently, we concluded that blood transfusions may affect survival rates in the advanced stages of gastric cancer, but not in the early stage.