The authors reply
Given the potential risks of transfusion (especially in critically ill patients), we felt that it was imperative to first establish physiologic proof-of-principle that transfusion could augment oxygen delivery (DO2) to vulnerable brain regions. If we found that DO2 did not rise, then it would be hard to argue for moving forward with a trial of transfusion evaluating clinical outcomes. Remarkably, instead, we found that the rise in DO2 exceeded that which we recently measured in response to induced hypertension (a well-established intervention) in the same population (3). We believe that this work (1) provides important translational support for clinical trials of transfusion that will test whether improving DO2 in this manner will result in reduction in cerebral ischemia and infarction and improve overall outcome despite the risks of transfusion.
As Riker and Seder (2) point out, the impact of transfusion on cerebral blood flow (CBF) is complex and multifactorial. It is important to consider that the regulation of the cerebral circulation focuses on maintaining optimal DO2, not CBF. The system adapts to changing levels of multiple factors (perfusion pressure, cardiac output, oxygen saturation, oxygen content, viscosity, CO2, potassium, metabolic products, etc.) in order to maintain a state where DO2 is about twice demand (and therefore, oxygen extraction fraction [OEF] is below 50%). Thus, the CBF response to changes in one factor (e.g., hemoglobin/arterial oxygen content) must be interpreted in the context of the others.
How these brain-injured patients would respond to an increase in hemoglobin was previously unknown, especially at higher hemoglobin levels. A fall in CBF could occur in response to increased plasma viscosity, counteracting the benefit of increased oxygen content. Alternatively, if DO2 is already adequate, CBF will fall in order to keep DO2 stable (3). Instead, we found that regional CBF was not reduced by transfusion in vulnerable brain regions (i.e., where DO2 was impaired) even at higher hemoglobin levels, suggesting that increased viscosity is not likely a limiting factor to flow in these hypoxic tissue beds.
They also raise the critical issue of whether improving DO2 translates into improved oxidative metabolism. Our observation that oxygen usage (cerebral metabolic rate of oxygen consumption [CMRO2]) is unchanged is not conclusive, as it is limited in terms of timing of the measurements in relation to transfusion and degree of oligemia versus ischemia. This key question remains unanswered.
We disagree with their interpretation that our observed reduction in OEF does not represent improved balance between cerebral oxygen demand and supply. The vulnerable brain regions in our patients had impaired DO2 that was clearly associated with elevated OEF. It is only by increasing OEF that these regions can maintain CMRO2; they exist in a state of severe hemodynamic compromise that can easily progress to ischemia if DO2 is further reduced (4). Improvements in DO2 would not be expected to improve CMRO2 in such oligemic regions; instead, we believe that a reduction in OEF represents a shift away from the precipice of ischemia. Elevated regional OEF is a well-established marker of reversible cerebral ischemia and a risk factor for stroke (5, 6).