Blood Transfusion and Cardiac Surgery—Reply

In Reply: Both letters comment that the statistical design of the noninferiority trial by Dr Hajjar and colleagues¹ greatly limits any conclusions that can be drawn. Our Editorial did point out that, given the ethical complexities and methodological challenges, to date no randomized controlled trial has compared a transfused group with a nontransfused group. One such methodological challenge is that since whole-body oxygen consumption (VO₂) is independent of whole-body oxygen delivery (DO₂) until hemoglobin levels are down to the range of 3 to 4 g/dL,² it is problematic to design trials to demonstrate any benefit to blood transfusion at hemoglobin levels above this range, particularly if mortality is an end point. Alternative physiologic indicators for transfusion triggers or end points have been advocated,³ such as oxygen extraction ratios less than 50%, mixed venous oxygen saturation (SVO₂) less than 0.60, and/or mixed oxygen tension less than 30 mm Hg; these have been proposed for clinical trials of artificial oxygen carriers such as hemoglobin solutions, when compared with blood transfusions.