“… this [work and others] supports a credible hypothesis that the blood storage lesion effect on morbidity and mortality is seen primarily in patients with preexisting risk factors for cardiovascular disease during critical illness.”
THE methodology to store erythrocytes for transfusion has markedly enhanced the supportive care capacity in hematology, surgery, trauma, and critical care medicine. However, the limitations of stored blood have come under recent scrutiny and great controversy. In this issue of ANESTHESIOLOGY, Lei et al .1at Massachusetts General Hospital show that transfusion with stored erythrocytes increases tissue injury of hemorrhagic shock more than transfusion with fresh erythrocytes, using an experimental rodent model of blood transfusion. Furthermore, they find that preexisting endothelial dysfunction caused by a high-fat diet increases the risk of this adverse response to transfused stored blood, modeling a common scenario under which human patients might receive transfusions for traumatic emergencies.
Over the years, several observational studies have hinted that critically ill patients are at greater risk of dying if they receive transfusions of multiple units of erythrocytes, suggesting an adverse property of stored erythrocytes. This finding is hotly debated, because it is essentially impossible to erase the confounding issue that the patients who require the most transfusions inherently are also the sickest and most likely to die. Do the sickest patients die because of blood transfusions or despite blood transfusions? This is a sensitive but crucial question.
Pathologic changes in blood during storage, often called the blood storage lesion, is a rapidly growing area of biomedical research. Before 2005, this was an esoteric subject, addressed in five peer-reviewed articles each year on average. Since that year, nearly 20 articles per year on the average fuel the debate with new research on the storage lesion. Stored erythrocytes undergo morphologic changes, metabolic alterations, and some degree of hemolysis during storage. During storage, erythrocytes undergo depletion of potassium, 2,3-diphosphoglycerate, adenosine triphosphate, lipids, and membrane, with increased erythrocyte rigidity and impaired oxygen delivery.2The stored units accumulate microvesicles derived from erythrocytes, free hemoglobin and biologically active lipids, demonstrating increased proinflammatory and procoagulant activity. Modern blood-bank preservative solutions provide glucose and other stabilizing substances to minimize these undesirable changes that might adversely affect the transfusion recipient.3
Despite these stabilizers, detectable hemolysis occurs in the storage bag and in the patient. The Food and Drug Administration mandates minimal standards that less than 1% of the erythrocytes lyse in the storage bag, and no more than 25% of the cells cleared from the recipient within 24 h after infusion, which likely correspond to cells lysed in vivo .3These round number thresholds are arbitrarily established long ago and not derived by any empirical scientific testing.
This debate more recently has been focused on a single hot button: duration of erythrocyte storage. Patients who received newer blood serve as experimental controls for those who receive older blood. Wang et al .4performed a meta-analysis of 21 independent observational studies evaluating associations of transfusion with erythrocytes stored for longer duration. Older blood was associated with a significantly increased risk of death (odds ratio, 1.16; 95% CI, 1.07–1.24). These authors provocatively estimate that if all cardiac surgery and trauma patients were transfused with fresher rather than older blood, one additional life would be saved for each 97 transfused with standard blood.
This analysis has unleashed a volley of caveats regarding the intrinsic limitations of these retrospective analyses of observational data.5A patient receiving multiple transfusions is likelier to receive at least 1 unit of blood that has been stored longer than average, confounding the age of blood with the amount of blood and the severity of illness. There is no standard, accepted method that accurately scores the properties of the stored blood, even storage duration, when multiple units of blood are administered, partly because many factors are known to change in stored blood as part of the storage lesion. Assumptions during the analysis can introduce statistical artifact and biases that may either obscure or overestimate the association of storage duration to subsequent outcomes.5Randomized controlled trials certainly could produce data with fewer such issues.
Three randomized controlled trials have been initiated in cardiac surgery,6critically ill patients,7and in neonates,8but no final results are yet published. In an effort to reach answers more quickly and mechanistically, several investigators have conducted physiologic studies of stored blood in animal models and in humans. Many of the researchers have experimentally investigated vasodilatory dysfunction in recipients of stored blood products attributable to extracellular free hemoglobin. In vitro , extracellular free hemoglobin avidly scavenges nitric oxide, a signal transduction ligand that orchestrates a broad program of vascular homeostasis. Additional human and animal research studies have evaluated the contribution of nitric oxide scavenging by free hemoglobin to physiologic consequences of the blood storage lesion.9Alternatively, some investigators have emphasized oxidant stress as the principal mechanism mediating free hemoglobin toxicity to endothelial function and vascular smooth muscle tone.10Still others even implicate oxidant stress from free heme or iron released from hemoglobin.11Whatever the dominant mechanism, many diverse lines of investigation implicate hemolysis of stored blood as pathogenic.
Regardless of exact mechanism, a growing number of publications report evidence that stored blood impairs normal vasodilation and blood flow and implicate extracellular hemoglobin as the offending molecule.12Free hemoglobin and hemoglobin-containing erythrocyte microvesicles accumulate during blood storage and consume nitric oxide in vitro . Transfusion of these forms of extracellular hemoglobin induces vasoconstriction in rats.13Haptoglobin, the endogenous scavenger of free hemoglobin, in a guinea pig model blunts adverse effects of stored blood, including intravascular hemolysis, acute hypertension, vascular injury, and kidney dysfunction, supporting a role in vascular pathophysiology of extracellular hemoglobin among the other potentially pathogenic molecules released during hemolysis.14Transfusion of blood stored for 40 days into lambs induces pulmonary vasoconstriction and acute pulmonary hypertension preventable with inhaled nitric oxide.15The effectiveness of either haptoglobin or nitric oxide breathing is consistent with free hemoglobin as the pathogenic agent in stored blood, whether mediated by nitric oxide scavenging or oxidative stress.
It may be possible that the effect of the blood storage lesion on outcome is clinically significant only when combined with additive or synergistic risk factors. The Zapol group has found that known risk factors for vasomotor dysfunction, including diabetes,16high-fat diet, and hemorrhagic shock,1amplify the adverse effect of blood storage duration on vascular function and outcome in animal models. In clinical practice, this supports a credible hypothesis that the blood storage lesion effect on morbidity and mortality is seen primarily in patients with preexisting risk factors for cardiovascular disease during critical illness.
This hypothesis would help to explain why the effect of stored blood on vascular physiology is not apparent in healthy volunteers.17Future investigations may yield more informative results if focused on transfusion with stored blood under life-threatening conditions compounded by other vascular risk factors. Until then, clinical use of erythrocytes stored up to 42 days remains the standard of care.