Safety and efficacy of shed mediastinal blood transfusion after cardiac surgery: A multicenter observational study

Objective: To examine the efficacy and safety of shed mediastinal blood (SMB) transfusion in preventing allogenic red blood cell(RBC) transfusion. Design: An observational clinical study. Setting: Twelve US academic medical centers. Participants: Six hundred seventeen patients undergoing elective primary coronary artery bypass grafting. Interventions: Patients were administered SMB transfusion or not, according to institutional and individual practice, without random assignment. Measurements and Results: The independent effect of SMB transfusion on postoperative RBC transfusion was examined by multivariable modeling. Potential complications of SMB transfusion, such as bleeding and infection, were examined. Three hundred twelve of the study patients (51%) received postoperative SMB transfusion (mean volume, 554 +/- 359 mt). Patients transfused with SMB had significantly lower volumes of RBC transfusion than those not receiving SMB (0.86 +/- 1.50 v 1.08 +/- 1.65 units; p < 0.05). However, multivariable analysis showed that SMB transfusion was not predictive of postoperative RBC transfusion. Demographic factors (older age, female sex), institution, and postoperative events (greater chest tube drainage, lower hemoglobin level on arrival to the intensive care unit, and use of inotropes) were significant predictors of RBC transfusion. The volume of chest tube drainage on the operative day (707 +/- 392 v 673 +/- 460 mt; p = 0.30), reoperation for hemorrhage (3.1% v 2.5%; p = 0.68), and overall frequency of infection (5.8% v 6.6%; p = 0.81) were similar between patients receiving and not receiving SMB, respectively. However, in patients who did not receive allogenic RBC transfusion, there was a significantly greater frequency of wound infection in the SMB group (3.6% v 0%; p = 0.02). Conclusion: These data suggest that SMB is ineffective as a blood conservation method and may be associated with a greater frequency of wound infection. Copyright (C) 1999 by W.B. Saunders Company.