Background: Blood product transfusion has been known for immunosuppressive effects, and over-transfusion is linked with adverse outcomes. In cardiac surgery, the risk of non-transfusion can be poor postoperative oxygen delivery and hemorrhage. We hypothesized that infectious complications, organ dysfunction, and mortality result after a given threshold of blood product transfusion is exceeded. Methods: Retrospectively, a prospectively maintained institutional database was analyzed from April 1, 2004 through December 31, 2006. All patients undergoing coronary artery bypass and/or valve operations were evaluated for bivariate and multivariate associations of blood-product transfusion with postoperative complications and mortality. Additionally, risk factors associated with transfusion were assessed. Receiver operator characteristic (ROC) curves analyses were employed to determine transfusion thresholds associated with complications. Results: During the study period, 741 patients met inclusion criteria. Fifty-four percent received postoperative blood-product transfusions. Previous cardiac intervention, renal dysfunction, stroke, and immunosuppression were some of the risks associated with transfusion (P < 0.05). Specific complications independently (P < 0.05) associated with total blood product transfusion identified from the multivariate analysis were infectious, neurologic, organ dysfunction, cardiac, and death. From ROC curve analyses, 5.5 units of total blood product transfusion was the inflection point for infectious complications (sensitivity 73%, specificity 64%) and organ dysfunction (sensitivity 73%, specificity 64%). For mortality, the inflection point was a transfusion of 7.5 units of total blood-products (sensitivity 73%, specificity 71%). Conclusion: Bloodless cardiac surgery is associated with a decreased morbidity and mortality. Limiting transfusion is advisable. Transfusion of less than 5.5 units of total blood-products may not have deleterious effects on outcomes.
- cardiac surgery