[Control of the ratio of the flow rate of the substitution fluid to the blood removal rate during preoperative normovolemic hemodilution]. / Asservissement du débit d'administration du soluté de substitution à celui du prélèvement sanguin au cours de l'hémodilution normovolémique préopératoire.

ABSTRACT

Twenty-one patients (mean age 46 +/- 13 years) due to undergo abdominal or ENT surgery, presumed to give rise to an important blood loss were included in this study. None had any contra-indication to the use of normovolaemic haemodilution (NH). Mean initial haematocrit was 40.3 +/- 1.8%. Their estimated total blood volume was 4,867 +/- 857 ml. The patients were anaesthetized with thiopentone, fentanyl, vecuronium or atracurium. Maintenance was carried out with isoflurane (0.5% during NH). Usual haemodynamic monitoring was used throughout. The required haematocrit was decided on before starting NH. The amount of blood to be removed was calculated with usual mathematical formulae. A radial artery cannula (n = 7), or a subclavian or femoral venous cannula (n = 14) was used to remove blood, which was collected within a bag containing CPC-adenine. Six % hydroxyethyl starch (Elohes) was given through a short venous cannula some distance from the first one. An antiparallel double line set in a roller pump was used to carry out the NH. A mean 1,341 +/- 405 ml of blood were withdrawn so as to reach a mean haematocrit of 30.6 +/- 2.4%. NH was completed within 17 +/- 6 min. No major haemodynamic changes occurred during the procedure. No significant differences were observed between expected and observed final haematocrits. There was no effect of the volume of blood withdrawn on the error of haematocrit prediction (0.5 +/- 0.3%). However, a higher rate of blood removal could increase this error. This easy-to-use device seems to provide fast and identical rates of blood removal and replacement. The expected haematocrit may thus be reached reliably, even if this must be checked for the sake of safety.