Mathematical and computer modeling of acute normovolemic hemodilution

Advocates of acute normovolemic hemodilution (ANH) frequently neglect to consider the decreasing hematocrit of the patient during both hemodilution and the subsequent operative procedure and the need to begin transfusion at some minimal hematocrit.

For more accurate prediction of the efficacy of ANH, equations were derived and a computer model developed that allowed accounting for the decreasing hematocrit due to blood loss in an isovolemic patient and calculating the red cell volume on a minute-by-minute basis; the model also began the transfusion of ANH blood on a mL-for-mL basis when the minimal hematocrit was reached and transfused any remaining blood following completion of the case. The red cell volume saved by performing ANH for a given estimated blood volume (EBV) was expressed as either the fraction of the red cell volume of a routinely banked unit of blood (red cells stored in additive solution: volume 350 mL, hematocrit 0.65) or the number of units saved.

The number of units saved in a typical example--EBV, 5000 mL; pre-ANH hematocrit, 0.40; minimal hematocrit at which transfusion was begun, 0.25 over a range of estimated blood losses (500-2500 mL); and 1 to 5 ANH units drawn--never exceeded 0.6. Even with extensive hemodilution, as in a child (EBV, 1500 mL; pre-ANH hematocrit, 0.40; minimal hematocrit at which transfusion was begun, 0.15; 5 units drawn; and estimated blood losses, 2500, 1500, and 1000 mL) with a postdilution hematocrit of 0.16, the savings would have been only 0.29, 0.44, and 0.49 units, respectively.

Because of the decreasing hematocrit in a bleeding isovolemic patient and the need to begin transfusion at some minimal hematocrit, the theoretic savings in red cell volume attributable to ANH is less than had previously been appreciated, and additional ANH does not necessarily result in additional patient benefit.