SYSTEMATIC UNDERESTIMATION OF ANEMIA SEVERITY IN POSTOPERATIVE PATIENTS

Transfusion. Volume 46 Page 317 - March 2006

The hematocrit (Hct), defined as the volume-percent (vol%) of red blood cells (RBCs) within whole blood, is a widely used measure to assess the presence of anemia and to judge the need for blood transfusion in diverse clinical settings. For example, the clinical relevance of the Hct in normovolemic patients was shown in a study that evaluated admission Hct and subsequent blood transfusion in acute myocardial infarction, in which there was evidence that transfusion for a Hct less than 33 volume-percent was associated with a lower short-term mortality.

The situation is different, however, in the bleeding hypovolemic patient: the rapid loss of blood does not result in a commensurate decrease in the Hct, as there are equivalent losses of RBCs and plasma in the shed blood, and the relative decrease in RBCs compared with plasma required for a diminished Hct will not be immediately observed. Only with subsequent shifts of the interstitial fluid into the reduced intravascular compartment will the expected reduction in Hct ensue. This fluid shift is a consequence of reduced capillary hydrostatic pressure (arising from increased arteriolar constriction and reduced blood pressure in the bleeding patient), which favors the movement of interstitial fluid into the intravascular compartment and thereby increases the plasma volume. Thus, a significant degree of hypovolemic anemia due to acute blood loss is not adequately reflected in the Hct. In practice, however, physicians often assume that interstitial fluid shifts occur rapidly and completely. Indeed, a textbook of physiology provides the following exemplary scenario: with 20 percent hemorrhage (i.e., reduction in total blood volume from 5000 to 4000 mL), the immediate posthemorrhage Hct is unchanged (46 vol%), but within several hours the interstitial to plasma fluid shift results in a return in total blood volume to nearly normal (4900 mL), and thus the commensurate decline in Hct (to 37.5 vol%) matches closely the expected Hct (36.8 vol%) had the expected fluid shift fully restored the initial 5000-mL blood volume. By assuming that such fluid shifts occur rapidly and completely, physicians might inappropriately regard the Hct as a global assessment of anemia in diverse clinical settings.

In this issue of TRANSFUSION, Valeri and colleagues have systematically studied this issue in patients undergoing vascular (n = 41) or cardiac surgery (n = 20). They used radiolabeled autologous RBCs to determine the RBC volume and then calculated the plasma and total blood volumes from the measured RBC volumes and total body Hct (the latter defined as peripheral venous Hct multiplied by 0.89). Preoperative values were compared with those determined at 1, 2, and 24 hours after surgery.

These investigators found that fluid redistribution from the interstitial to the intravascular compartment remains substantially incomplete at 24 hours after surgery. They determined in these hypovolemic anemic patients that the Hct values observed during the postoperative period underestimated by 4 to 5 volume-percent the values that would otherwise have been expected if one assumed a state of normovolemic anemia. For example, in the postvascular surgery patients, the observed mean Hct values (38 and 36 vol% at 2 and 24 hr after surgery, respectively) would have been approximately 32 volume-percent, had these patients with reduced RBC volumes been normovolemic. In the postcardiac surgery patients, the observed mean Hct values (30 and 27 vol% at 2 and 24 hr after surgery, respectively) would have been 25 and 23 volume-percent during these respective postoperative periods, had these patients with reduced RBC volumes been normovolemic.

The authors also found that their patients exhibited clinical evidence of cardiovascular compensation for their hypovolemic anemia, such as increased heart rate and cardiac output. Given that the lower limit of the normal range of Hct is approximately 40 volume-percent (37 vol% in women) and that common Hct triggers for transfusion range from 21 to 30 volume-percent (approx. 70-100 g/L), an underestimate of 4-5 volume-percent (approx. 15 g/L) could well be clinically relevant. For example, a transfusion trigger of 70 g per L that might well be appropriate for a normovolemic anemic patient4,5 would correspond to a value of approximately 85 g per L for a postcardiac or postvascular surgery patient 24 hours after surgery, according to these authors' observations.

Of course, a simple blood test variable cannot be expected to provide the sole basis for transfusion decisions in a complex postoperative setting in which numerous interacting factors influence hemodynamics and oxygen delivery. Moreover, there are other settings of acute blood loss, such as gastrointestinal bleeding, in which clinical teaching emphasizes that the Hct may not reflect adequately the severity of anemia.

The question arises whether the delayed and incomplete fluid redistribution described here is a general phenomenon associated with rapid blood loss or rather is more specific to these (and similar patients) affected by surgery and trauma. Studies of controlled, relatively atraumatic blood loss, such as that occurring with normal blood donation (450-500 mL) or in experimental studies of larger blood losses (1 L), indicate that fluid redistribution occurs rapidly—but incompletely—within minutes, followed by a much slower secondary phase of volume redistribution. One study of 42 blood donors, from each of whom 485 mL blood was withdrawn, found a median fluid shift of 208 mL, that is, almost 45 percent of the blood volume lost, that occurred within minutes after phlebotomy. Heart rate increased significantly (by approx. 6%). Another study of 50 volunteers found that a blood donation of 450 mL led to a very similar (median, 199 mL) rapid fluid shift. Two other studies found that after the removal of as little as 500 mL blood, only half of the lost blood volume was replaced at 24 hours, and after removal of 1 L of blood quickly from healthy normal men, blood volume replacement remained only three-quarters complete at 72 hours. Thus, to what extent the timing, rate, and magnitude of this homeostatic intravascular volume "resuscitation" occurs in these disparate situations of blood loss—ranging from routine blood donation to major trauma and surgery—remains unclear.

As underscored by Valeri and colleagues, an arbitrary Hct transfusion trigger does not differentiate between normovolemic and hypovolemic anemia. Thus, a "one trigger fits all" approach to transfusion is not appropriate and indicates that different transfusion variables need to be considered in studies assessing Hct or hemoglobin (Hb) thresholds for transfusion triggers, especially if a mixed medical-surgical patient population is assessed, as is often the case in transfusion studies involving critically ill patients.11,12 In a bleeding patient, the decision to transfuse should take into account the presence and rapidity of ongoing bleeding, clinical evidence for disturbed hemodynamics, and—as suggested by this study by Valeri and colleagues—a somewhat higher Hct (or Hb) threshold than used in a normovolemic anemic patient. Without physician awareness of this important matter of cardiovascular physiology and pathophysiology, there is the potential for inadequate treatment of patients and inadequate conclusions in studies of the transfusion trigger if there is overreliance on the Hct measure alone.