Although a lower transfusion trigger is generally recommended, little evidence is available about the physiological mechanisms of mild anemia in diseases with an imbalance between O₂ supply and O₂ demand such as sepsis. This study was undertaken to describe the systemic and coronary metabolic O₂ reserve in an awake sheep model of hyperdynamic sepsis comparing two different hemoglobin levels. Twenty-four hours after sheep were rendered septic by cecal ligation and perforation (CLP), blood transfusion (n = 7, hemoglobin = 120 g/l) and isovolemic hemodilution (n = 8, hemoglobin = 70 g/l), respectively, were performed. Another 24 h later, we measured hemodynamics, organ blood flows, and systemic and myocardial O₂ metabolism variables at baseline and through four stages of progressive hypoxia. Maximum coronary blood flow was 766.3 ± 87.4 ml · min¹ · 100 g¹ in hemodiluted sheep group versus 422.7 ± 53.7 ml · min¹ · 100 g¹ in the transfused sheep (P < 0.01). Myocardial O₂ extraction was higher in the transfusion group (P = 0.03) throughout the whole hypoxia trial. In the hemodilution group, coronary blood flow increased more per increase in myocardial O₂ uptake than in transfused sheep (P < 0.01). This was accompanied by a lower left ventricular epicardial-to-endocardial flow ratio in hemodiluted sheep (1.13 ± 0.07) than in transfused sheep (1.34 ± 0.02, P < 0.05). We conclude that the lower coronary blood flow and greater myocardial O₂ extraction in transfused septic sheep preserves transmyocardial O₂ metabolism better in comparison to hemodiluted sheep.