Infections are a common and significant sequela of major traumatic injury. The objective of this study was to evaluate the relationship between infections in trauma patients and the transfusion of packed red blood cells (pRBCs) within the first 48 hours of admission. We hypothesized that transfusions of pRBCs were associated with an increased risk of infection in a dose-dependent manner. All adult patients admitted to the trauma service of a Level I trauma center from November 1996 to December 1999 were studied. Secondary analysis was performed on prospectively collected data. One thousand five hundred ninety-three consecutive patients were studied; of these 12.6 per cent developed at least one infection. The overall transfusion rate was 19.4 per cent. The infection rate in patients who received at least one transfusion was significantly higher (P < 0.0001) at 33.0 versus 7.6 per cent in patients receiving no pRBCs. Transfusions per patient ranged from 0 to 46 units. There was a clear exponential correlation in patients receiving between 0 and 15 transfusions (R2 = 0.757). Multivariate logistic regression, which was used to identify risk factors for the development of infection, demonstrated the odds ratio of receiving pRBCs to be 1.084, with a 95 per cent confidence interval of 1.028 to 1.142 (P = 0.0028). In summary there is a clear dose-dependent correlation between transfusions of pRBCs and the development of infection in trauma patients. Multivariate analysis further demonstrated that pRBCs were an independent risk factor for the development of infections. Although transfusions are frequently indicated, they should be administered appropriately and with no more pRBCs than absolutely necessary.