University Hospital of Geneva, Switzerland. denis.morel@hcuge.chIn Reply:—We appreciate the interest of Moskowitz et al . regarding our article. 1They hypothesize that too restrictive inclusion criteria and excessive fluid replacement in the acute normovolemic hemodilution (ANH) group could explain the absence of beneficial effects of ANH.First, for obvious ethical reasons, we selected patients to undergo preoperative ANH based on the recommendations of the British Committee for Standards in Blood Transfusion, which take into account the potential risks and benefits of blood salvage techniques. Therefore, due to the risk of myocardial ischemia, stroke, or heart failure, we did not perform ANH in patients with severe coronary artery disease (recent myocardial infarction, unstable angina, left main coronary stenosis), severe aortic stenosis, poor left ventricular function, carotid stenosis, and/or preoperative anemia. According to these stringent exclusion criteria, we included only 57% of all consecutive patients undergoing elective coronary revascularization or valvular surgery. Nevertheless, a high risk of bleeding was considered in more than 50% of these surgical candidates (preoperative anticoagulant and/or aspirin therapy, repeat or combined operation) in whom the blood-sparing benefits of ANH could be anticipated.Second, perioperative fluid hydration was similar in both groups, except for compensation of the removed blood by colloids in the ANH group: the volumes of infused crystalloids (94.6 ml · kg−1· 24 h−1and 88.8 ml · kg−1· 24 h−1in the control and ANH groups, respectively) and colloids (16.5 ml · kg−1· 24 h−1and 26.2 ml · kg−1· 24 h−1in the control and ANH groups, respectively) were in the range of the values commonly reported 2and included any solution administered intravenously and through the cardiopulmonary bypass (e.g. , cardioplegia, priming). To achieve similar cardiac preload conditions, larger than expected volume of colloids were given in the ANH group (hypervolemic compensation), which resulted in ANH-induced vasodilatation and reduced blood viscosity. 3Since ANH is known to be associated with poor tolerance to acute blood loss and impaired adrenergic response, 4,5we closely guided fluid administration using transesophageal echocardiography and central venous pressure monitoring in order to avoid hypovolemia and to detect myocardial ischemia.Finally, as suggested by Dr. Moskowitz and others, 6we agree that the implementation of a standardized, multidisciplinary approach, including definition of transfusion criteria, administration of antifibrinolytic drugs, use of cell saver devices, and external heating, is most helpful to reduce transfusion of allogenic blood products in patients undergoing cardiac surgery. However, evidence supporting the addition of ANH as a blood-sparing technique is still lacking, although a consensus is based on data from observational studies and prospective, unblinded, randomized trials. 7