The effect of perioperative blood transfusion on cancer progression remains controversial because retrospective clinical studies have produced conflicting results. We have collected data prospectively on 379 patients undergoing curative surgery for colorectal adenocarcinoma and assessed the effect of variables, including blood transfusion, on survival. Univariate and multivariate survival analysis has been carried out. When the end-point for analysis used was death due to recurrent colorectal carcinoma and non-cancer deaths were censored, there was no difference in cancer-specific survival between transfused and non-transfused patients. Survival analysis was also carried out without censoring the non-cancer deaths and clearly demonstrated how the statistical analysis and data interpretation could be distorted by age-related non-cancer deaths. The incidence of recurrence of colorectal carcinoma was not greater in the transfused group than in the non-transfused group. We conclude that blood transfusion should not be withheld in colorectal surgery for fear of worsening the prognosis.
展开▼
