The aim of this study is to evaluate influence of allogeneic blood transfusion on prognosis in patients in Dukes B stage of colorectal cancer. All patients with colorectal cancer who were admitted at our Department of Surgery between January 2000 and December 2004 were analyzed. One hundred fifty-one patients who fulfilled inclusion criteria were enrolled in further evaluation. B stage according to Dukes classification and curative resection were inclusion criteria. Exclusion criteria were polyposis syndromes, nonpolyposis syndromes, inflammatory bowel disease, autoimmune disease and previous blood transfusion. Patients were divided into two groups: Group 1 received ≤ 3 units of allogeneic blood transfusion and group 2 received >3 units of allogeneic blood transfusion. "Cutoff" value of 3 units of blood was defined according to our results and literature data. Follow-up was 5 year. There was no statistical difference between these groups in local recurrence (χ(2) = 0.009, P > 0.05) and distant metastasis (χ(2) = 0.44, P > 0.05). Also, the Kaplan-Meier survival curves were calculated, and long-rank test did not show a survival difference between these two groups (log rank = 0.075, P > 0.05). Postoperative complications are significantly more frequent in Group 2 (χ(2) = 4.67, P < 0.05). Multivariate logistic regression analysis confirmed that intraoperative blood transfusion more than three units had independent influence on local recurrence. Postoperative transfusion more than 3 units was statistically independent prognostic factor for metastasis and mortality. Overall transfusion less than 3 units of allogeneic blood does not influence the outcome of patients in Dukes B stage of colorectal cancer.