According to published guidelines based on clinical experience, perioperative blood transfusions are not beneficial when hemoglobin levels are >10 g/dL. However, other factors such as comorbid conditions and discrete organ ischemia complicate the establishment of clear indications for transfusion. Two new studies address distinct but related aspects of this issue.

In the Transfusion Requirements After Cardiac Surgery (TRACS) trial, 502 patients who underwent cardiac surgery with cardiopulmonary bypass between February 9, 2009, and February 1, 2010, at a single cardiac surgery referral center in São Paolo, Brazil, were randomized to one of two perioperative red blood cell (RBC) transfusion strategies: liberal transfusion to maintain a hematocrit of 30% (hemoglobin level, 10 g/dL), or conservative transfusion to maintain a hematocrit of 24% (hemoglobin level, 8 g/dL). Transfusion rates were relatively high in both groups (78% and 47% in the liberal and conservative groups, respectively; P<0.001). The primary outcome, a composite of 30-day mortality and severe in-hospital complications, did not differ significantly between the two groups.

Bennett-Guerrero and colleagues analyzed data from the Society for Thoracic Surgery Adult Cardiac Surgery Database (ACSD) on >100,000 patients undergoing coronary artery bypass grafting with cardiopulmonary bypass in 2008 at 798 centers across the U.S. Perioperative transfusion rates varied substantially; in 408 sites performing at least 100 on-pump procedures, the rate of RBC transfusion ranged from 7.8% to 92.8%. After adjustment for patient-specific risk factors, transfusion rates varied significantly by geographic location, academic status, and hospital volume. However, these factors combined explained only 11.1% of between-hospital variation in RBC use, and case mix explained only 20.1% of the variation. Mortality rates were not associated with hospital-specific RBC transfusion rates.

Comment: Despite some study limitations, the TRACS findings showed that patients do as well with a conservative transfusion strategy as with a more liberal one. The ACSD findings indicate that between-hospital variability in transfusion rates remains high and independent of several patient- and hospital-specific factors. The absence of differences in mortality among centers with varying transfusion rates strongly suggests inappropriate transfusions, but current guidelines seem ineffective at reducing these variations. As editorialists note, more restrictive policies of blood transfusion in cardiac surgery should be developed and instituted at individual sites and regionally.

Published in Journal Watch Cardiology November 3, 2010