Allogeneic blood transfusion is associated with increased morbidity and mortality. The authors evaluated the affect of blood transfusion, independent of injury severity on mortality. The authors conducted a retrospective review of all patients, age > or =18 years with blunt injury admitted to their Level 2 trauma center from 1994 to 2004 by query of the NTRACS trauma registry. Initial systolic blood pressure and heart rate determined the shock index. Logistic regression was used to model the affect of blood transfusion on mortality. Transfusion requirements were categorized as follows: A, 0 U; B, 1 to 2 U; C, 3 to 5 U; D, > or =6 U blood. In this sample of 8215 blunt trauma patients, 324 patients received blood transfusion. Mortality rates between the transfused and nontransfused groups were 15.12 per cent and 1.84 per cent (P < 0.000) respectively. In the logistic regression model, transfusion category B did not have a significant affect on the odds of death (P = 0.176); the affect of transfusing 3 to 5 U and > or =6 U had a mortality odds ratio of 3.22 (P = 0.002). and 4.87 (P = 0.000) respectively. Transfusing > or =2U blood was strongly associated with mortality in this blunt trauma population. There must be a continuous attempt to limit blood transfusion when feasible and physiologically appropriate.