Blood transfusion-modulated tumor recurrence: first results of a randomized study of autologous versus allogeneic blood transfusion in colorectal cancer surgery.

ABSTRACT

PURPOSE: Allogeneic blood transfusions have reportedly been associated with a poor prognosis in patients with curatively resected cancer. To control for immunosuppression induced by a speculatively causal allogeneic blood transfusion, we designed a randomized study in which the control group received autologous blood transfusions not related to any condition of immunosuppression. PATIENTS AND METHODS: One hundred twenty patients with potentially curative resectable colorectal cancer and the capability to predeposit autologous blood were randomly selected to receive either standard allogeneic blood transfusion or predeposited autologous blood. RESULTS: In curatively resected cancer patients, the number who needed allogeneic blood transfusions was reduced from 60% in the allogeneic blood group to 33% in the autologous blood group (P = .009). After a median follow-up duration of 22 months (range, 8 to 48) tumor recurrence was observed in 28.9% of the allogeneic blood group and 16.7% of the autologous blood group. Life-table analysis established a tendency toward a shorter tumor-free survival for the allogeneic blood group (log-rank P = .11). The problem with this analysis was the strong association of allogeneic blood transfusions with tumor recurrence, which interfered in 33% of patients in the autologous blood group who required additional allogeneic blood transfusions. Multivariate analysis of established risk factors for tumor recurrence and surgery-related variables reflecting potential immunosuppressive conditions showed that only pT stage (relative risk, 6.61; 95% confidence interval [CI], 1.82 to 23.99; P = .004), pN stage (relative risk, 8.39; 95% CI, 3.15 to 22.33; P < .001), and the need for allogeneic blood (relative risk, 6.18; 95% CI, 2.20 to 17.37; P < .001) were independent predictors of tumor recurrence. Subgroup analysis of patients who received a transfusion of < or = 2 U blood found a significantly higher risk of tumor recurrence in the allogeneic blood group (relative risk, 5.16; 95% CI, 1.13 to 23.62; P = .034), which was reduced to borderline significance (relative risk, 3.54; 95% CI, 0.76 to 16.51; P = .107) by adjustment for tumor (T) and node (N) stage. CONCLUSION: As indicated by these first results, the blood transfusion modality has a significant effect on tumor recurrence after surgical treatment of colorectal cancer. A change in the practice of blood transfusion might thus potentially surpass the impact of any recent adjuvant treatment strategies.