The effect of perioperative blood transfusion on survival in head and neck cancer.

This Head and Neck Intergroup analysis was undertaken to evaluate further previously reported observations linking blood transfusions, which were given to patients with head and neck cancer, to a worse prognosis. This study population represents those patients registered to the Head and Neck Intergroup Trial 0034 for previously untreated resectable squamous cell carcinoma. Additional transfusion data were obtained by one of us (D.E.S.) on 217 patients and added to the Head and Neck Intergroup data set, providing an opportunity for assessing the impact of survival by other variables. The study group was separated using 13 variables. Analysis demonstrated that transfusion did not significantly decrease the locoregional control (P = .60). Multivariate analysis indicated that T stage (P = .015), N stage (P = .004), treatment received (P = .004), and Karnofsky Performance Scale (P = .031) were the only factors that did significantly influence survival. This multivariate analysis controlling for these variables demonstrated no significant effect on survival for those patients receiving transfusion during surgery (P = .55) or after surgery (P = .39). This study of 217 patients, controlled for other variables, does not demonstrate any significant negative relation between blood transfusions and either locoregional control or survival.