Background: The degree of immunomodulation by perioperative blood transfusion and its resultant effects on cancer surgery are a subject of controversy. We evaluated the prognostic effects of perioperative blood transfusion on gastric cancer surgery.
Methods: A total of 1710 patients who underwent curative gastrectomy for gastric cancer from 1991 to 1995 were retrospectively reviewed. Uni- and multivariate analyses of the incidence, amount, and timing of perioperative blood transfusions and a comparison of the clinicopathological features were performed.
Results: A higher incidence of blood transfusions was associated with female sex, large tumors, upper-body location, Borrmann type III or IV lesions, longer operations, total gastrectomies, splenectomies, and D3 or more extended lymphadenectomy. The tumors in the transfused group were more advanced in depth of invasion and nodal classification. More frequent tumor recurrences were found in the transfused group. A dose-response relationship between the amount of transfused blood and prognosis was evident. Subgroup analyses of prognosis according to stage showed significant differences in stages III and IV between the transfused and nontransfused groups. On multivariate analysis, transfusion was shown to be an independent risk factor for recurrence and poor prognosis.
Conclusions: These results suggest that perioperative transfusion is an unfavorable prognostic factor. It is thus better to refrain from unnecessary blood transfusion and to give the least amount of blood to patients with gastric cancer when transfusion is inevitable, especially for those with stage III and IV gastric cancers.