The effect of anemia and blood transfusions on mortality in closed head injury patients

BACKGROUND: The purpose of this study was to determine if anemia in isolated head trauma patients results in a higher mortality rate that would justify a more liberal use of blood transfusions. METHODS: A retrospective review of isolated blunt head trauma patients was performed between January 2001 and December 2006. Comparisons were made between survivors and nonsurvivors regarding demographics, laboratory values, transfusions received, and lengths of stay. RESULTS: There were 788 patients with 735 survivors who were significantly younger (46.3 y +/- 21.5 survivors versus 68.9 y +/- 18.8 nonsurvivors, P < 0.0001) and less injured [(ISS: 14.7 +/- 5.2 survivors versus 23.2 +/- 4.7 nonsurvivors, P < 0.0001), (head abbreviated injury severity: 3.7 +/- 0.7 survivors versus 4.7 +/- 0.5 nonsurvivors, P < 0.0001)] than those who died (n = 53). The survivors also had shorter lengths of stay (days) [(ICU: 2.4 +/- 4.2 versus 5.6 +/- 11.7, P = 0.03), (hospital: 6.3 +/- 9.8 versus 7.8 +/- 14.8, P = 0.02)]. Multivariate logistic regression showed age (OR 1.063, CI 1.042-1.084), ISS (OR 1.376, CI 1.270-1.491), minimum hemoglobin (OR 0.855, CI 0.732-1.000), and total blood products transfused (OR 1.073, CI 1.008-1.142) to be independent predictors of mortality with an ROC of 0.942. Outcome was independent of the operative procedures, hematocrit and packed red blood cells transfused at 24, 48, and 72 h. Hemoglobin levels of <8 mg/dL were more predictive of death than >8 mg/dL (P = 0.01). CONCLUSIONS: This study supports the need to balance mild anemia with judicious blood product use in the head trauma patient. Given the risk with blood product use, each transfusion should be carefully considered and the patient re-evaluated regularly to determine the need for further intervention.