A computer model comparing normovolemic hemodilution, hypervolemic hemodilution, and neither on intraoperative blood loss and final hematocrit.

Homologous blood transfusion, while frequently life-saving, is attended by risks and complications. Autologous blood transfusions have become an increasingly common alternative. Volume expansion, which is simpler, also is used. This study was designed to construct computer models of hypervolemic hemodilution and normovolemic hemodilution to compare them with each other and with normal (neither hypervolemic nor normovolemic hemodilution). Each model started with blood volume (BV) equal to 5,000 mL. Initial hematocrits (HCTs) were varied from 25% to 50%. Following phlebotomy and hemodilution or volume expansion, which ranged from 0 to 2,500 mL (50% of initial BV), the models were then bled 250 to 2,500 mL (5% to 50% of initial BV). In the phlebotomy model, the autologous blood was then returned. Final HCTs were then calculated. Preoperative phlebotomy of 500 to 1,000 mL, an amount commonly withdrawn, provides a minimally higher final HTC. Volume expansion by hypervolemic hemodilution provides almost the same low level of benefit. Benefits (3% higher HCT) are not seen until larger volumes are phlebotomized or hemodiluted and accompanied by large intraoperative blood losses. Autologous blood drawn by preoperative phlebotomy for intraoperative transfusion should not be used until studies show that these large volumes are safe and actually save blood.