Acute normovolemic hemodilution and the cardiac patient

Maintenance of myocardial oxygen delivery during ANH depends essentially on the increase in the coronary blood flow as oxygen extraction is already nearly maximal at the level of the heart under resting conditions.[8] This is achieved by a reduction in coronary vascular resistance related to the decreased blood viscosity and to specific coronary vasodilation. Heart rate and possibly myocardial contractility have been shown to increase during hemodilution,[28] which results in an augmentation of myocardial oxygen demand. When the hematocrit is reduced to about 10%, myocardial oxygen consumption more than doubles and coronary vasodilation is nearly maximal. Below such a hematocrit, coronary blood flow can no longer match myocardial oxygen demand and ischemia develops, ultimately resulting in cardiac failure.

The dependency of myocardial oxygen supply to the coronary blood flow highlights the vulnerability of the heart during ANH, especially in patients with coronary artery disease (CAD) in whom coronary blood flow cannot increase. The lowest tolerable hemoglobin concentration in CAD patients remains unknown.[29] A recent pilot study addressing specifically the problem remained inconclusive.[30] It probably depends on several factors, including the severity of the disease.[31] There is, however, increased evidence that tolerance of CAD patients to isovolemic anemia closely depends on the level of myocardial oxygen demand.

In anesthetized patients scheduled for coronary surgery, several studies demonstrated that moderate ANH (target hematocrit value 27–33%) is well tolerated, and may be even cardioprotective if associated with a decreased myocardial oxygen demand.[32] Myocardial oxygen balance is profoundly influenced by the level of heart rate, and recent clinical data confirm that tolerance of CAD patients to moderate anemia is closely related to the level of heart rate.[33] The anesthetic technique also may play a role.[34] The early postoperative period is certainly critical in hemodiluted CAD patients, because they have to face an increased tissue metabolic demand.[35]

Acute normovolemic hemodilution and hemostasis

Hemodilution could affect hemostasis in different ways. Firstly, it will dilute not only plasmatic factors, but also cellular coagulation factors, such as platelets and, of course, RBCs. Red blood cells have been shown to interfere with hemostasis through a mechanical effect but also through biological effects related to the release of intracellular adenosine diphosphate and to the generation of thrombin.[36] The clinical consequences (i.e. importance of perioperative bleeding) of these interactions between RBCs and hemostasis remain to be determined.

Hemodilution could also affect hemostasis through the direct effects of plasma substitution fluids on the platelets and the coagulation mechanisms.[37] These effects are more marked with colloids than with crystalloids. Among colloids, they are more marked with dextrans than with gelatins and albumin. For hydroxyethyl starches, these effects seem to be closely related to the intrinsic properties of the different solutions, such as a high in vitro molecular weight and a high degree of hydroxyethyl substitution.[38] Several in vitro studies have confirmed these effects in the context of ANH.[39,40] In addition, patients with Type O blood may demonstrate more coagulation compromise than those with non-O blood when undergoing hemodilution with low molecular weight hydroxyethyl starch.[41]

Storage of whole blood during ANH might be associated with disturbances in platelet aggregation,[42] although the clinical relevance of this observation remains unknown. Agitation during storage of ANH blood does not seem to have any effect on platelet function.[43]

Despite the fact that ANH may directly interfere with normal hemostasis, there is no evidence from the literature that ANH is associated with increased perioperative bleeding,[44,45] although it may affect standard coagulation tests.[46]

Theoretical aspects

The basic concept behind ANH is that patients undergoing such procedure will lose fewer erythrocytes per milliliter of lost blood during surgery and after transfusion of the collected autologous blood in the immediate postoperative period.[5]

Several equations have been developed to calculate the efficacy of ANH as a function of surgical blood loss, initial hematocrit, target post-ANH hematocrit, and hematocrit used as the transfusion trigger. Presuming a “usual” surgical patient without preoperative anemia, and a transfusion decision based exclusively on a trigger hemoglobin concentration of 6–7 g/dl, Weiskopf [47] calculated that 55 to 77% of patient’s total blood volume must be lost during surgery in order to achieve savings of about 180 ml of RBCs, which represents one standard blood unit. However, the usefulness of the different published equations in clinical practice remains limited, as several factors have not been always taken into account in the proposed formulae.[2].

Results from the literature

Although efficacy of ANH as a blood conservation technique remains highly debated,[44,45,48,49] this technique is still currently used in many adult and pediatric centers.[50,51] Most of the studies reviewed were performed in the setting of cardiac or orthopedic surgery. Adequate evaluation of the published results was hampered by the relative poor quality of the studies and the marked heterogeneity observed between trials, partly explained by study factors (patient populations, target hematocrit values, transfusion triggers, ANH technique, etc.). Efficacy of ANH was found to be relatively modest in terms of likelihood of exposure to allogeneic blood and units transfused. It closely depends on the use or not of protocols to guide transfusion practice. There was no obvious increase in adverse events with ANH, but the incidence of complications was poorly reported.

Studies that demonstrated a clear efficacy of ANH in reducing the likelihood of patient’s exposure to allogeneic blood in the perioperative period have been performed in major liver resection [52,53] or major abdominal surgery.[54] In all of them, a high volume of blood was collected. Surgery was associated with significant blood loss, and a low hemoglobin concentration (7–8 g/dl) was used as the transfusion trigger. Adequate selection of patients may also play a role.[55] All these observations indicate that efficient ANH requires significant expertise in the field from the care-giving team.