Does acute normovolemic hemodilution work in cardiac surgery?

To the Editor:—In a recent article by Höhn et al., it was concluded that acute normovolemic hemodilution (ANH) in addition to aprotinin was not beneficial in preventing allogeneic blood transfusions compared to aprotinin alone in cardiac surgery. In this randomized, controlled trial, the patients were hemodiluted to a hematocrit of 28% pre–cardiopulmonary bypass (CPB). The transfusion threshold was set at 17% during CPB and at 25% for post-CPB. The total fluid replacement was in excess relative to the amount of ANH (autologous blood) removed (6.4 2.1 l of crystalloid, 2.0 0.7 l of colloid). This led to excessive hemodilution, reducing the hematocrit below the transfusion threshold in the ANH group. Indeed, 50% of the patients in the ANH group required either all (33%) or a portion (22%) of the autologous blood to be transfused during CPB, thus negating its positive effects on erythrocytes and coagulation protection. Consequently, allogeneic erythrocyte transfusion rates and the indirect clinical markers for surgical bleeding (cell saver and 24-h chest tube drainage) were not different between the two groups. One of the goals of ANH is to protect the autologous blood from the negative effects of CPB and to return it after heparin neutralization. Additional hemodilution occurs with the onset of CPB; therefore, hemofiltration or ultrafiltration and/or diuresis should have been employed to remove excess fluid. Alternatively, ANH can be performed just prior to the onset of CPB (by diverting heparinized blood into a storage bag), thus preventing excessive dilutional anemia. The criteria for exclusion from this study were left main disease, severe aortic stenosis, recent myocardial infarction, unstable angina, ejection fraction below 30%, severe carotid stenosis, combined coronary artery bypass grafting and valve cases, respiratory insufficiency, renal insufficiency, and anemia (hemoglobin 12 g/dl). This exclusion process resulted in the selection of a group of patients that we know are at low risk for allogeneic transfusions. Our data (shown below) and those of other investigators demonstrate that patients presenting with adequate hemoglobin levels (average starting hematocrit of 43.3%, body surface area of 1.86) rarely require allogeneic blood. At our institution, the cardiac surgery program utilizes a multidisciplinary approach to blood conservation. In over 300 cardiac surgery cases (coronary artery bypass grafting, valves, and combined procedures), we remove on average 1280 ml of ANH blood per case (based on a formula to reach a target hematocrit on bypass of 20%). The average amount of fluid used for replacement was 1680 ml of crystalloid and 591 ml of colloid (Hextend, Abbott Laboratories, North Chicago, IL). Hemofiltration or ultrafiltration and/or induced diuresis is frequently utilized on CPB to remove excess fluids and to reduce the dilutional effect from the CPB prime. The starting hematocrit averages 39%. We use -aminocaproic acid for low-risk cases and reserve aprotinin (Trasylol, Bayer, West Haven, CT) for high-risk cases. The total amount of cell saver returned is approximately 200 ml, and 24-h chest tube drainage is 428 ml. Allogeneic transfusion rates for packed erythrocytes, fresh frozen plasma, and cryoprecipitate are 11%, 3%, and less than 1%, respectively.