State of the Art Review
Journal of Veterinary Emergency and Critical Care 25(2) 2015, pp 187–199 doi: 10.1111/vec.12252
Red blood cell storage lesion Rafael Obrador, LVS, DACVECC; Sarah Musulin, DVM, DACVECC and Bernie Hansen, DVM, MS, DACVECC, DACVIM
Abstract
Objective – To summarize current understanding of the mechanisms responsible for changes occurring during red blood cell (RBC) storage, collectively known as the storage lesion, and to review the biological and clinical consequences of increasing storage time of RBCs. Data Sources – Human and veterinary clinical studies, experimental animal model studies, and reviews of the RBC storage lesion with no date restrictions. Human Data Synthesis – Experimental studies have characterized the evolution of human RBC and supernatant changes that occur during storage and form the basis for concern about the potential for harm from long-term storage of RBCs. Although 4 randomized controlled trials of varying sizes failed to find an association between RBC storage time and negative clinical outcomes, a recent meta-analysis and numerous observational clinical studies have demonstrated that transfusion of old versus fresh stored RBCs is associated with an increased risk of morbidity and mortality, particularly among trauma victims and cardiac surgery patients. Potential clinical consequences of RBC transfusion following development of the storage lesion include risk of organ dysfunction, organ failure, infections, and death. Veterinary Data Synthesis – Experimental animal models have contributed to the evidence supporting adverse consequences of the RBC storage lesion. Studies on relevant RBC storage issues such as the effect of different preservative solutions and leukoreduction have been completed. Transfusion with RBCs stored for 42 days increases mortality in dogs with experimental sepsis. Conclusion – Storage of RBCs induces progressive biochemical, biomechanical, and immunologic changes that affect red cell viability, deformability, oxygen carrying capacity, microcirculatory flow, and recipient response. Most reports in the human and veterinary literature support the concept that there are deleterious effects of the RBC storage lesion, but additional studies with improved experimental design are needed to identify compelling reasons to modify current blood banking and transfusion practices. (J Vet Emerg Crit Care 2015; 25(2): 187–199) doi: 10.1111/vec.12252 Keywords: blood transfusion, complications, transfusion medicine
Abbreviations
2,3-DPG DVT ECs FDA Hb IL LPS MODS MOF
diphosphoglycerate deep vein thrombosis endothelial cells Food and Drug Administration hemoglobin interleukin lipopolysaccharide multiple organ dysfunction syndrome multiple organ failure
From the Department of Clinical Sciences, College of Veterinary Medicine, North Carolina State University, Raleigh, NC 27607. The authors declare no conflict of interest. Address correspondence and reprint requests to Dr. Sarah Musulin, Department of Clinical Sciences, College of Veterinary Medicine, North Carolina State University, 1052 William Moore Drive, Raleigh, NC 27607, USA. Email: [email protected] Submitted February 08, 2013; Accepted October 01, 2014. C Veterinary Emergency and Critical Care Society 2014
MP NO PFK RBC RCT SNO-Hb TRALI TRIM
microparticle nitric oxide phosphofructokinase red blood cell randomized controlled trial S-nitrosohemoglobin transfusion-related acute lung injury transfusion-related immunomodulation
Introduction The growing availability of veterinary hospital blood bank and commercial sources of canine and feline blood products has made transfusion therapy more widely available to veterinarians and has also increased the use of stored blood products. Ex vivo storage of blood was first accomplished in 1915 following the discovery of sodium citrate as a blood anticoagulant.1 Great progress has been made since that time to improve the sterility, 187
R. Obrador et al.
quality, and shelf life of collected blood. Human red blood cells (RBCs) can be stored in polyvinyl chloride bags at 2–6°C for up to 42 days depending on the red cell preservation solution added.2,3 This expiry is based on criteria set by the Food and Drug Administration (FDA), which requires that 75% of transfused RBCs remain in circulation 24 hours after transfusion and that hemolysis affects less than 1% of stored cells at the end of the approved storage period.4 Although there are no legal standards for storage of feline and canine RBCs, they are routinely stored for up to 30 and 35 days depending on the storage solution used.5–7 During storage, a number of changes occur within the RBC and storage media that affect RBC survival, function, and recipient response to transfusion. These changes comprise the RBC storage lesion and include biochemical, biomechanical, and immunologic events.8–11 In this review, we will describe the development of the RBC storage lesion and the pathophysiology of its consequences, and examine clinically relevant studies on the subject.
Biochemical Changes Decrease in adenosine triphosphate RBCs require energy in the form of adenosine triphosphate (ATP) for maintenance of shape, deformability, phosphorylation of membrane phospholipids and proteins, membrane transport functions, synthesis of purine and pyrimidine nucleotides, and synthesis of glutathione.12 As mature mammalian RBCs lack mitochondria, their sole source of energy is anaerobic glycolysis via the Embden–Meyerhoff pathway. During this process, glucose is cleaved to pyruvate and reduced to lactic acid, a process that generates 2 moles each of ATP and lactate for every mole of glucose. Phosphofructokinase (PFK), the enzyme that catalyzes the conversion of the fructose 6-phosphate intermediate to fructose 1,6-phosphate, is a key regulatory point for the rate of glycolysis in the circulating RBC.13 Although this step requires ATP to proceed, PFK activity is also inhibited by the allosteric binding of ATP to the enzyme so that as the ATP:AMP ratio falls, the activity of this enzyme increases. In contrast, during storage the dominant inhibitor of PFK is hydrogen ion activity, which increases as lactic acid accumulates with time.14,15 The progressive inhibition of PFK and glycolysis by hydrogen ion activity results in failure to produce sufficient ATP to meet cellular metabolic requirements. As the concentration of this energy source falls, RBC membrane Na+ -K+ ATPase activity, membrane stability, glucose transport, oxidative stress defense mechanisms, and membrane phospholipid distribution begin to fail.16–18 The rate of decline in ATP concentration during 188
storage is affected by the preservative solution used. In stored canine RBCs, critical loss of ATP may be delayed from day 10 to day 44 by the addition of saline, dextrose, adenine, and mannitol to the storage solution.19,20 The ATP depletion that occurs during storage has direct adverse effects on RBC deformability, as ATP provides the energy necessary for maintenance of RBC membrane elasticity, intracellular viscosity, and an optimal RBC surface area-to-volume ratio.21,22 RBCs with impaired deformability cannot adapt their shape as needed to transit easily through the microcirculation, impairing the delivery of oxygen to and removal of carbon dioxide from the tissues. Reduced RBC deformability is also a major factor in reduced survival of the RBC following transfusion, as poorly deformable RBCs are culled from the circulation as they pass through the splenic circulation.23
Decrease in 2,3-diphosphoglycerate 2,3-Diphosphoglycerate (2,3-DPG) is a glycolytic intermediate present in RBCs that serves as a major modifier of hemoglobin (Hb) oxygen affinity in many species, including human beings, dogs, and rats.24 In contrast, feline RBCs contain very low levels of 2,3-DPG, and feline Hb does not require 2,3-DPG to modulate oxygen affinity.22 About 10–20% of the glycolytic intermediate 1,3-disphosphoglycerate is diverted from glycolysis for synthesis of 2,3-DPG. The 2,3-DPG molecule intercalates between the -globin chains of deoxyhemoglobin, stabilizing the deoxy form and shifting the base of the oxyhemoglobin dissociation curve to the right. This steepens the curve and moves the mid-point of the curve, the P50, to the right, causing oxyhemoglobin to release oxygen at a higher local pO2 . Following 2 weeks of storage, 2,3-DPG concentrations in human, canine, and rat RBCs are virtually depleted,5,25,26 shifting the oxygen dissociation curve to the left, increasing Hb oxygen affinity and theoretically impairing oxygen delivery to the tissues. De novo synthesis of 2,3-DPG begins with RBC rejuvenation in the hours following transfusion, with restoration of normal RBC 2,3-DPG within 3 days.27–29 Despite potentially adverse consequence of this delay in 2,3-DPG restoration, the results of experimental studies in baboons,30 rats,25,31,32 and healthy people33 have not supported a role of 2,3-DPG as a key factor for oxygen delivery and consumption capacity. For example, Raat et al25 evaluated the effect of storage time of human RBCs on intestinal microcirculatory oxygenation in a rat isovolemic exchange model. Although RBCs were devoid of 2,3-DPG by 2–3 weeks of storage, oxygen delivery to the intestines by those RBCs was the same as after transfusion with fresh RBCs. The same investigators found that oxygen C Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12252
RBC storage lesion
delivery by RBCs stored for 5–6 weeks had impaired oxygen-delivery capacity, indicating that increased oxygen affinity may occur by mechanisms other than depleted 2,3-DPG. Thus, although 2,3-DPG is rapidly and significantly depleted in stored RBCs, an adverse impact of this particular storage lesion on oxygen delivery has not been proven. Furthermore, this mechanism may be even less important in species that do not depend on 2,3-DPG to modify Hb oxygen affinity. For example, the feline Hb P50 is only slightly reduced after 35 days of storage suggesting that the storage lesion does not have a major impact on Hb oxygen affinity in this species.34
Nitric oxide Nitric oxide (NO) is an uncharged radical produced by the vascular endothelium that plays a critical regulatory role as a local vasodilator. It also affects homeostasis through inhibition of platelet aggregation, serves as a microbiocidal agent for host defense, inhibits endothelial adhesion molecule expression, and has antioxidant properties as recently reviewed.35 Within RBCs, NO binds to the 93 cysteine thiol residue of Hb to form S-nitrosohemoglobin (SNO-Hb).26 In the deoxygenated state, the reactivity of SNO-Hb is increased, and transnitrosation reactions transfer the NO group to glutathione or thiols on the RBC membrane protein anion exchanger1. These S-nitrosothiol compounds are released from RBCs at the tissue level in direct proportion to Hb oxygen desaturation and signal vasodilation, increasing flow to match regional perfusion and oxygen delivery with metabolic demand.36 During storage, RBC SNO-Hb bioactivity falls rapidly within RBCs and is not restored following infusion into a recipient, and some investigators have suggested that this may impair tissue oxygen delivery after RBC transfusions.26,37,38 Although this model linking Hb saturation with regional blood flow has gained wide acceptance, another laboratory using knockout mice with circulating human Hb in which the 93 cysteine thiol is replaced with alanine to eliminate direct Hb participation in SNO release found evidence that SNO-Hb is not essential for at least some aspects of RBCdependent vasodilation.39 However, that suggestion was based largely on measurements of secondary and tertiary physiological responses (including systemic blood pressure, right ventricular hypertrophy, and pulmonary arterial wall thickness), and others have argued that those measures cannot be used to discount SNO-Hb as an important physiological regulator of the microcirculation. ATP plays a role in the response of RBCs to tissue hypoxia through a vasodilatory mechanism that utilizes nitric oxide (NO).26,37,40,41 When circulating RBCs are heavily deoxygenated during transit through tissues with low ambient oxygen pressure, they are stimulated C Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12252
to release vasodilatory compounds to enhance local blood flow. This process facilitates optimal oxygen delivery in response to local oxygen demand. This process depends in part on the release of ATP from RBCs in response to oxyhemoglobin desaturation; the free plasma ATP stimulates production of NO in the endothelium.42 This RBC response to local hypoxia is impaired by ATP depletion as seen in stored RBCs, potentially contributing to inadequate regional perfusion secondary to stunted microcirculatory vasodilation.43 Enhanced chemical reduction of NO occurs in the recipient after transfusion of stored RBCs secondary to scavenging by free Hb and Hb-containing RBC microparticles (MPs) that accumulate due to hemolysis and RBC damage during storage.44–46 The rate of hemolysis during storage depends on the individual donor, the storage solution, and storage duration, with most human blood units exhibiting 0.3–0.4% hemolysis prior to administration.47 Following transfusion, this free Hb is then available to react with NO to form methemoglobin and nitrate. This reaction, normally slowed in the blood by cellular isolation of Hb from plasma by the cytoskeleton beneath the RBC membrane and the unstirred layer of the immediate extracellular fluid environment surrounding the cell, is then free to proceed rapidly.48,49 Hb-containing MPs also bring Hb into contact with NO much more rapidly than intact RBCs because of their much smaller size and altered membrane constitution.48,50 Donadee et al46 measured the ability of human free Hb and MPs to scavenge NO using stopped-flow spectroscopy and laser-triggered NO release, and found that free Hb and MPs react with NO about 1,000 times faster than erythrocyte-bound Hb. In a paired in vivo rat model, they demonstrated that even at concentrations below 10 mol/L, cell-free Hb can produce potent vasoconstriction due to the potent scavenging ability of free Hb.46 In addition to promoting vasoconstriction, this decrease in NO bioavailability may also enhance intravascular thrombosis, WBC adhesion and diapedesis, endothelial permeability, and smooth muscle proliferative responses after vascular injury.35 Baek et al51 transfused guinea pigs with old (28 d), or new (2 d) blood and demonstrated that administration of the old cells led to intravascular hemolysis, acute hypertension, vascular injury, and kidney dysfunction consistent with free Hb-mediated injury. The attenuation of these effects with the highaffinity Hb-scavenger haptoglobin confirmed the role of free Hb in the vascular and kidney pathophysiology.
Increased potassium in the supernatant Increased potassium accumulation in the supernatant of stored human RBC products occurs as ATP 189
R. Obrador et al.
concentrations fall and render ATP-dependent erythrocyte Na+ -K+ -ATPase pumps inactive. Increasing potassium concentration in the supernatant of stored human RBC products is recognized as potentially detrimental or even fatal during rapid infusion.52 Intracellular potassium concentration is low in both feline and canine RBCs and their RBCs membranes lack Na+ -K+ -ATPase pump activity.53,54 Therefore, significant potassium accumulation in canine and feline stored blood products is unlikely. However, many clinically normal Japanese and Korean dog breeds have high RBC concentrations of potassium associated with significant Na+ -K+ -ATPase pump activity,55 and prolonged storage of blood obtained from these breeds could be associated with accumulation of potassium in the supernatant.
as glycolysis declines during RBC storage.64 Under these circumstances, superoxide will interact with iron and water via the Fenton reaction to form hydroxyl radicals, which can in turn oxidize and damage proteins21,62,65 and lipids.66,67 For example, methemoglobin concentration in human banked blood increases 3-fold during long storage times.68 Since the duration of storage of RBCs is typically much shorter than the 110-day in vivo lifespan of human RBCs (100–115 d for dogs, 73 d for cats),69 the total burden of oxidative damage that accumulates during RBC storage is assumed to be low. However, because this burden is delivered all at once with transfusion (and repeatedly with massive transfusions), the potential to overwhelm endogenous antioxidant capacity may be significant.
Ammonia accumulation Ammonia concentration increases with time in both human and canine stored RBC products.56–58 Human RBCs are rich in ammonia containing three times more ammonia than that found in the plasma.59 The increase in ammonia in stored blood is attributed to the generation and release of ammonia from RBCs as well as deamination of amino acids and purines, such as adenine found in preservative solutions.60 Waddell et al58 found that plasma ammonia concentration in anemic dogs without primary liver disease did not increase following transfusion with 5–10 mL/kg of stored pRBCs and remained within the normal reference range. However, those dogs had normal liver function and received a relatively small amount of product, and patients with liver dysfunction or those receiving massive transfusions may be more at risk of developing hyperammonemia.58
Biomechanical Changes
Oxidation injury Oxidation injury to RBCs is yet another mechanism contributing to the storage lesion and is a plausible cause of MPs formation and the loss of RBC deformability.61,62 The high oxygen availability in blood provides a rich environment for oxidation of constituent molecules, and RBCs are particularly susceptible due to their role in carrying oxygen and lack of mitochondria. The binding of oxygen to RBC Hb is dynamic and oxygen is constantly dissociating from one Hb molecule and binding to another. This exchange is not perfect, and occasionally an oxygen molecule extracts an electron from the heme unit, forming ferric Hb (methemoglobin) and the superoxide radical (O2 − ).63 Normal RBCs contain the reducing agent glutathione and the antioxidant enzymes superoxide dismutase and methemoglobin reductase; these systems quickly dismutate the superoxide radical to hydrogen peroxide and reduce ferric heme to the functional ferrous form.12 However, glutathione concentrations fall 190
RBC shape change and reduced RBC deformability To deliver oxygen to tissues, the 6–8 m (depending on species) diameter RBC must flow through capillaries with diameters as small as 3 m. Therefore, the deformability of the biconcave disc is crucial for adequate perfusion, and even modest reductions in deformability may reduce RBC ability to traverse these narrow channels. During prolonged storage, RBCs undergo progressive shape change from a readily deformable biconcave disc to poorly deformable echinocytes with cell surface protrusions, and ultimately to nondeformable spheroechinocytes.70 Although initially reversible, once beyond the early spheroechinocyte stage, the loss of deformability becomes permanent as the RBCs lose membrane through the budding of MPs from the tips of the echinocytic spines.71 The etiology of these morphologic changes and microvesiculation observed during storage appears to be multifactorial and several mechanisms have been proposed: depletion of ATP25,72 and 2,3-DPG,73,74 loss of membrane phospholipid with associated MPs,62,75 protein rearrangement,64 and lipid oxidation.76 As storage time increases, these corpuscular changes also increase RBC adherence to the endothelium as described later.77,78 These shape changes during RBC storage are associated with rheological changes, increased viscosity, and reduced capillary perfusion,77 all of which can jeopardize tissue oxygenation. Accumulation of MPs MPs are small anucleate phospholipid-rich vesicles that contain membrane-bound and cytoplasmic proteins from the cell from which they derived. MPs are normally shed from reticulocytes and healthy RBCs. Normal RBCs use the microvesiculation process as a way of shedding oxidized and polymerized lipids.79,80 In stored RBC C Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12252
RBC storage lesion
products, the concentration of RBC MPs increases with time, with relatively greater concentrations found toward the end of the storage period.9 RBC-derived MPs that accumulate during storage have shown procoagulant activity.81 In contrast to their parental cells, the majority of MPs expose negatively charged phospholipids, primarily phosphatidylserine, on their outer membrane. These negatively charged phospholipids promote assembly of coagulation complexes such as tenase or prothrombinase, initiating and propagating thrombin generation.81 MPs also express abundant tissue factor activity, further contributing to a procoagulant state.82 Additionally, RBC-derived MPs may also have a proinflammatory activity through modification of platelet–WBC interactions in transfused patients, based on altered membrane properties of RBC MPs that enhance platelet interactions to increase inflammatory chemokine bioavailability in vitro.83,84 Adhesion to endothelial cells Normal RBCs do not readily adhere to intact blood vessel walls, but this characteristic may be altered by storage as has been suggested by the results of studies using in vitro models to simulate blood flow. Independent groups have demonstrated that stored human RBCs have enhanced adherence to vascular endothelial cells (ECs) and that the concentration of adherent RBCs increases with storage time.77,85,86 If this occurs in vivo, enhanced adhesion of RBCs to vascular endothelium might disrupt local blood flow patterns, decrease oxygen delivery to peripheral tissues, and perhaps even occlude microvessels.87 The expression of procoagulant phosphatidylserine on the outer RBC membrane increases with both RBC age and storage time and can contribute to endothelial adherence.77,88 Storage lesion associated RBC corpuscular changes and loss of RBC deformability contributes to endothelial cell adherence and the mechanical obstruction of microvessels.77,86 Prestorage leukoreduction abrogates excessive stored RBC adhesion to endothelium in these in vitro models, implicating donor WBCs in the development of this phenomenon.85,89 Inflammation in the recipient may worsen storage-associated RBC endothelial adhesion, as suggested by the finding that pretreatment of human ECs with endotoxin to activate the ECs and mimic infection in the in vitro flow models results in increased adhesion of RBCs to the activated ECs.90 Many patients receiving stored RBC transfusions have concomitant systemic inflammation adversely affecting their microcirculation; storage-related RBC adhesion may have significant clinical implications for tissue perfusion and oxygenation in those recipients.
C Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12252
Immunologic Changes Overview An association between transfusion therapy and alterations in recipient immune function was first reported in people almost 40 years ago, when Opelz et al91 observed improved renal allograft survival following pretransplant allogeneic RBC transfusion. This work demonstrated the presence of an immunosuppressive or immunomodulatory effect of RBC transfusion on the recipient (later referred to as transfusion-related immunomodulation [TRIM]). Multiple studies in animals and people have demonstrated that allogeneic transfusion downregulates cellular immunity and dysregulates innate immunity during inflammation.92–94 Allogeneic blood transfusion stimulates the T helper 2-type immune response as evidenced by subsequent increase in interleukin (IL)-10 and IL-4 cytokine secretion posttransfusion of stored human RBCs.94 IL-10 and IL-4 both downregulate the T helper 1 response, thus downregulating cellular immunity functions such as antigen presentation.92 The predominant mechanism of TRIM in any particular individual likely depends on the interplay of transfusion effects with genetic predisposition and intercurrent illness in that recipient. Platelets and vascular ECs also potentially contribute to TRIM as both cell types are highly responsive to inflammatory signals and when activated release potent bioactive mediators. The “two-insult” model of posttransfusion injury proposes that an initial insult (eg, an underlying inflammatory condition or trauma) primes the patient’s immune cells or endothelium, and further inflammation is triggered by a second immunomodulating insult (eg, transfusion), resulting in an unrestrained inflammatory response.95,96 A 2012 study by Tung et al97 corroborated this “two-insult” theory of TRIM as a cause of transfusion-related acute lung injury (TRALI) in an in vivo sheep model. In that study, sheep were randomized to first receive saline or lipopolysaccharide (LPS) and then further randomized to receive supernatant from fresh (1 d) or stored (42 d) nonleukoreduced human pRBCs or saline. The sheep were then monitored for hypoxemia within 2 hours of transfusion or histological evidence of pulmonary edema. None of the saline control group animals developed TRALI after transfusion with any product. Similarly, none of the sheep that were primed with LPS and then transfused with saline developed TRALI. While only 20% of the LPS sheep transfused with fresh RBCs developed TRALI, 80% of the LPS sheep that received stored RBCs did. In this model, both the priming of the recipient and the transfusion of stored blood resulted in TRALI.
191
R. Obrador et al.
Leukocyte contamination and soluble immune response modifiers Nonleukoreduced stored RBCs units contain abundant passenger leukocytes; these cells are immunologically active and their release of proinflammatory and prothrombotic compounds increases with storage time.98,99 Compounds whose concentrations increase with storage time include elastase, secretory phospholipase-A2, soluble human leukocyte antigen, soluble FAS-ligand, prostaglandin E-2, transforming growth factor 1, and proinflammatory cytokines IL-1, IL-6, and IL-8. 11,100,101 When leukocytes are stored in the increasingly acidic supernatant of refrigerated stored RBCs units, they become activated, produce cytokines, and eventually die. Release of enzymes such as phospholipase-A2 following cell death may potentiate the production of lysophospholipids including the dialkylglycerol platelet activating factor.102 These lysophospholipids are capable of priming and activating neutrophils, and induce various proinflammatory actions in leukocytes, ECs, and smooth muscle cells.66,103 To appraise the immunological activity of stored human RBC supernatant, Sparrow and Patton98 evaluated in vitro neutrophil priming activity by measuring expression of CD11b, IL-8 release, and neutrophil chemotaxis after incubation of purified neutrophils with 3 stored allogenic RBC products: nonleukoreduced RBCs, buffy coat poor RBCs, and RBCs that underwent prestorage leukoreduction. The leukoreduced and buffy coat poor RBC products did not induce significant neutrophil priming behavior for all 3 assays. In contrast, incubation of the purified neutrophils with the nonleukoreduced RBC product (WBC count 8,400 times higher than leukoreduced RBCs) revealed CD11b exposure, release of IL-8, and neutrophil chemotaxis consistent with speculation that proinflammatory mediators are present in stored RBC supernatant. Prestorage leukoreduction effectively removes 99.9% of donor WBCs with the goal of abrogating any immunomodulatory effects resulting from donor leukocytes and their associated bioreactive mediators. Both bench104 and clinical105–110 studies of the benefits and consequences of prestorage leukoreduction have yielded mixed results. Fergusson et al108 used meta-analysis of clinical trials to determine that RBC leukoreduction reduces postoperative infection in people, but found no reduction of cancer recurrence or perioperative mortality in recipients. In a retrospective cohort study comparing outcomes in transfusion recipients undergoing major surgeries before (n = 6,982) and after (n = 7,804) institution of universal leukoreduction, the same investigators found that leukoreduction was associated with a decreased incidence of fever episodes, antimicrobial use, and mortality in those high-risk patients.106 The use of leukoreduced blood has also been 192
associated with fewer early rejection episodes in patients with liver transplantation, irrespective of the type of immunosuppression suggesting a possible immunomodulatory effect of leukoreduced blood.109 However, not all studies have shown a benefit of using leukoreduced blood. Leukoreduction provided no measurable clinical benefit in humans infected with immunodeficiency virus or cytomegalovirus,105 did not decrease the rate of postoperative infection or length of hospitalization in patients undergoing elective surgical procedures,110 and was not associated with significant reductions in neonatal intensive care units mortality or bacteremia in premature infants requiring RBC transfusions.107 In veterinary medicine, preliminary studies have evaluated the effectiveness of commercially available leukoreduction filters at removing WBCs from stored blood and the effects of leukoreduction on canine RBCs viability and recipient inflammatory response.111,112 Prestorage leukoreduction of cooled canine blood effectively removed 99.99% of WBCs while preserving RBC viability as determined by percent hemolysis, biotin-labeling assays of cell survival, and flow cytometry.111 More recently, McMichael et al112 demonstrated a profound decrease in the inflammatory response in normal dogs to transfused stored (21 d at 4˚C) autologous leukoreduced RBCs (vs. nonleukoreduced) as determined by posttransfusion WBC count, fibrinogen, and C-reactive protein concentrations.
Clinical Consequences of the RBC Storage Lesion Transfusion therapy has many known potential adverse consequences, and the risk of harm from transfusion increases with increasing storage time.113–115 Although the FDA guidelines for maximum storage time are based on 75% survival of RBCs 24 hours following administration,4 there is growing evidence that longer storage times within this FDA-approved time frame compromise RBC oxygen delivery and produce adverse clinical consequences for the recipient. Effects on oxygen delivery kinetics Tissue oxygen delivery is a function of RBC oxygen carrying capacity and RBC flow in the macro- and microcirculation. The immediate goal of RBC transfusions is to increase oxygen delivery in critically ill patients; the effectiveness of this therapy in achieving this goal following RBC storage warrants investigation in light of abundant experimental evidence that storage depresses the ability of RBCs to deform, unload oxygen peripherally, and participate in control of regional blood flow. Most of the animal experiments addressing this issue have been performed on rats and follow an early study that demonstrated that storage for 28 days prevented C Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12252
RBC storage lesion
transfused RBCs from improving tissue oxygenation following administration to septic rats.116,117 However, rat RBCs age about 4 times more rapidly during storage than human RBCs,118 making it difficult to extrapolate these results to other species. In an effort to see if this effect occurred in RBCs from other species, Raat et al25 used a hemodiluted rat model (hematocrit 15%) to confirm this phenomenon with isovolemic exchange transfusion using fresh (2–6 d of storage), intermediate (2–3 wk), or old (5–6 wk) stored human RBCs. They demonstrated that isovolemic exchange transfusion with either fresh or intermediate RBCs maintained intestinal tissue oxygen tension but transfusion with old cells reduced it by 25%. Similarly, Tsai et al119 found that transfusion with autologous hamster RBCs stored for 28 days reduced microvascular flow, functional capillary density, and oxygen extraction compared with fresh RBCs. These studies in healthy animals under oxygen-restricted conditions suggest that older transfused RBCs may have a limited ability to acutely improve oxygen availability, but scrutinizing these concepts in a clinical setting is necessary to determine the importance of this phenomenon. Toward that end, Marik et al120 used gastric tonometry in a prospective, controlled study to measure gastric mucosal pH in septic human patients to determine that transfusion with RBCs stored longer than 15 days reduced gastric mucosal pH, an indicator of splanchnic hypoxia, indicating lower oxygen availability secondary to impaired microcirculatory perfusion despite increased blood Hb concentration. These authors suggested that transfusion of old, poorly deformable RBCs leads to microcapillary sludging and obstruction, resulting in gut ischemia. Contradicting these results, Walsh et al121 could not prove clinical worsening of tissue hypoxia via tonometric assessment of gastric mucosal oxygenation in a ventilated critically ill human population after transfusion with RBCs stored for more than 3 weeks compared with transfusion with fresh cells. This trial differed from the former as patients were randomized, received filtered leukodepleted RBCs, and were not septic (and therefore may not have exhibited the same degree of oxygen supply dependency as the patients in Marik’s study). Tinmouth et al122 examined 19 clinical human studies in a systematic review of the effect of RBC transfusions on oxygen delivery and consumption before and after transfusion of a delineated number of RBCs to critically ill patients, but the included studies did not assess the influence of RBC storage time. Although oxygen delivery was consistently increased by transfusion, oxygen consumption increased in only 6 of the 19 studies. This disparity between oxygen delivery and consumption illustrates the relative importance of recipient responses to RBC transfusion, as it is possible that impaired oxygen utilization from conditions such as sepsis are more C Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12252
important determinants than the storage lesion in some patients. Clinical consequences for the recipient – human literature The bulk of the literature exploring the clinical consequences of RBC storage lesion in people is derived from observational studies. Many of these were examined in a 2012 meta-analysis of 21 studies involving 409,966 patients by Wang et al,113 who assessed the safety of transfused older (range 9 to 31 d) versus newer (range 1.6–21 d) stored blood, predominantly in cardiac surgery and trauma patients. Eighteen of these studies were observational (12 retrospective, 6 prospective) and 3 were randomized controlled trials (RCTs). The primary outcome of interest was mortality. These authors found that the transfusion of older stored RBCs was associated with a significantly increased risk of death (odds ratio, 1.16; 95% confidence interval, 1.07–1.24). Further subgroup analysis of the trials found no patterns associated with patient type, size of trial, or amount of blood transfused. Pooled data from this meta-analysis suggest that older blood was also associated with an excess of multiple organ dysfunction syndrome (MODS), pneumonia, renal dysfunction, and sepsis. Numerous clinical trials and retrospective studies have examined the clinical consequences of stored RBC transfusions to patients with a variety of disorders. Investigators studying critically ill people following traumatic injury, cardiac surgery, and suffering from sepsis and other disorders have reported an independent direct association between increasing RBC storage time and risk of infection, deep vein thrombosis (DVT), ICU length of stay, and mortality in both adults and children.114,115,123–130 RBC storage times greater than 14–28 days were associated with worse outcomes. In a large retrospective study of cardiovascular surgery patients, Koch et al114 examined the relationships between transfusion and a diverse composite of 18 serious adverse complications including arrhythmias, thrombotic events, organ failure, sepsis, prolonged ventilation, and in-hospital death. Those complications were more common in patients who received older units (>14-d old) than those who received fresher units (14-d old). Patients given older blood had higher rates of in-hospital mortality and higher rates of extended intubation, renal failure, and sepsis. The significance of these findings is unclear since confounding factors (eg, concurrent mismatched platelet transfusions) were present and the treatment groups were not ideally balanced: more patients who received RBCs stored >14 days also received more blood transfusions. Several retrospective studies have been conducted to characterize morbidity and the complications of 193
R. Obrador et al.
MODS and multiple organ failure (MOF) associated with transfusion of stored RBCs in specific patient population.123–125,127 An early study by Zallen et al125 identified storage time as an independent risk factor for development of postinjury MOF in adult trauma patients.125 A greater risk of developing new or progressive MODS in hemodynamically stable critically ill pediatric patients receiving stored RBCs (>2–3 wk) has also been established.127 Given the experimental evidence of TRIM, clinical researchers have conducted studies examining the relationship between transfusion of stored RBCs and infection. The safety of stored transfusion products in trauma patients is particularly relevant as these patients often require massive transfusion, increasing the likelihood of receiving older blood, and have wounds predisposing them to infection. In fact, an increased incidence of infection in trauma patients receiving stored RBCs has been established by multiple observational studies.124,131,132 For example, Juffermans et al131 found an association between transfusion with RBCs stored more than 14 days and increased risk of bacterial infection after trauma; this risk was not reduced by administration of prophylactic antimicrobials. There was no association between infection and platelet transfusion in the same study. In another observational study of incidence of postoperative infection in patients undergoing coronary artery bypass grafting, Andreasen et al133 identified an increased risk following transfusion of buffy coat reduced RBC units stored 14 days compared with units stored
Journal of Veterinary Emergency and Critical Care 25(2) 2015, pp 187–199 doi: 10.1111/vec.12252
Red blood cell storage lesion Rafael Obrador, LVS, DACVECC; Sarah Musulin, DVM, DACVECC and Bernie Hansen, DVM, MS, DACVECC, DACVIM
Abstract
Objective – To summarize current understanding of the mechanisms responsible for changes occurring during red blood cell (RBC) storage, collectively known as the storage lesion, and to review the biological and clinical consequences of increasing storage time of RBCs. Data Sources – Human and veterinary clinical studies, experimental animal model studies, and reviews of the RBC storage lesion with no date restrictions. Human Data Synthesis – Experimental studies have characterized the evolution of human RBC and supernatant changes that occur during storage and form the basis for concern about the potential for harm from long-term storage of RBCs. Although 4 randomized controlled trials of varying sizes failed to find an association between RBC storage time and negative clinical outcomes, a recent meta-analysis and numerous observational clinical studies have demonstrated that transfusion of old versus fresh stored RBCs is associated with an increased risk of morbidity and mortality, particularly among trauma victims and cardiac surgery patients. Potential clinical consequences of RBC transfusion following development of the storage lesion include risk of organ dysfunction, organ failure, infections, and death. Veterinary Data Synthesis – Experimental animal models have contributed to the evidence supporting adverse consequences of the RBC storage lesion. Studies on relevant RBC storage issues such as the effect of different preservative solutions and leukoreduction have been completed. Transfusion with RBCs stored for 42 days increases mortality in dogs with experimental sepsis. Conclusion – Storage of RBCs induces progressive biochemical, biomechanical, and immunologic changes that affect red cell viability, deformability, oxygen carrying capacity, microcirculatory flow, and recipient response. Most reports in the human and veterinary literature support the concept that there are deleterious effects of the RBC storage lesion, but additional studies with improved experimental design are needed to identify compelling reasons to modify current blood banking and transfusion practices. (J Vet Emerg Crit Care 2015; 25(2): 187–199) doi: 10.1111/vec.12252 Keywords: blood transfusion, complications, transfusion medicine
Abbreviations
2,3-DPG DVT ECs FDA Hb IL LPS MODS MOF
diphosphoglycerate deep vein thrombosis endothelial cells Food and Drug Administration hemoglobin interleukin lipopolysaccharide multiple organ dysfunction syndrome multiple organ failure
From the Department of Clinical Sciences, College of Veterinary Medicine, North Carolina State University, Raleigh, NC 27607. The authors declare no conflict of interest. Address correspondence and reprint requests to Dr. Sarah Musulin, Department of Clinical Sciences, College of Veterinary Medicine, North Carolina State University, 1052 William Moore Drive, Raleigh, NC 27607, USA. Email: [email protected] Submitted February 08, 2013; Accepted October 01, 2014. C Veterinary Emergency and Critical Care Society 2014
MP NO PFK RBC RCT SNO-Hb TRALI TRIM
microparticle nitric oxide phosphofructokinase red blood cell randomized controlled trial S-nitrosohemoglobin transfusion-related acute lung injury transfusion-related immunomodulation
Introduction The growing availability of veterinary hospital blood bank and commercial sources of canine and feline blood products has made transfusion therapy more widely available to veterinarians and has also increased the use of stored blood products. Ex vivo storage of blood was first accomplished in 1915 following the discovery of sodium citrate as a blood anticoagulant.1 Great progress has been made since that time to improve the sterility, 187
R. Obrador et al.
quality, and shelf life of collected blood. Human red blood cells (RBCs) can be stored in polyvinyl chloride bags at 2–6°C for up to 42 days depending on the red cell preservation solution added.2,3 This expiry is based on criteria set by the Food and Drug Administration (FDA), which requires that 75% of transfused RBCs remain in circulation 24 hours after transfusion and that hemolysis affects less than 1% of stored cells at the end of the approved storage period.4 Although there are no legal standards for storage of feline and canine RBCs, they are routinely stored for up to 30 and 35 days depending on the storage solution used.5–7 During storage, a number of changes occur within the RBC and storage media that affect RBC survival, function, and recipient response to transfusion. These changes comprise the RBC storage lesion and include biochemical, biomechanical, and immunologic events.8–11 In this review, we will describe the development of the RBC storage lesion and the pathophysiology of its consequences, and examine clinically relevant studies on the subject.
Biochemical Changes Decrease in adenosine triphosphate RBCs require energy in the form of adenosine triphosphate (ATP) for maintenance of shape, deformability, phosphorylation of membrane phospholipids and proteins, membrane transport functions, synthesis of purine and pyrimidine nucleotides, and synthesis of glutathione.12 As mature mammalian RBCs lack mitochondria, their sole source of energy is anaerobic glycolysis via the Embden–Meyerhoff pathway. During this process, glucose is cleaved to pyruvate and reduced to lactic acid, a process that generates 2 moles each of ATP and lactate for every mole of glucose. Phosphofructokinase (PFK), the enzyme that catalyzes the conversion of the fructose 6-phosphate intermediate to fructose 1,6-phosphate, is a key regulatory point for the rate of glycolysis in the circulating RBC.13 Although this step requires ATP to proceed, PFK activity is also inhibited by the allosteric binding of ATP to the enzyme so that as the ATP:AMP ratio falls, the activity of this enzyme increases. In contrast, during storage the dominant inhibitor of PFK is hydrogen ion activity, which increases as lactic acid accumulates with time.14,15 The progressive inhibition of PFK and glycolysis by hydrogen ion activity results in failure to produce sufficient ATP to meet cellular metabolic requirements. As the concentration of this energy source falls, RBC membrane Na+ -K+ ATPase activity, membrane stability, glucose transport, oxidative stress defense mechanisms, and membrane phospholipid distribution begin to fail.16–18 The rate of decline in ATP concentration during 188
storage is affected by the preservative solution used. In stored canine RBCs, critical loss of ATP may be delayed from day 10 to day 44 by the addition of saline, dextrose, adenine, and mannitol to the storage solution.19,20 The ATP depletion that occurs during storage has direct adverse effects on RBC deformability, as ATP provides the energy necessary for maintenance of RBC membrane elasticity, intracellular viscosity, and an optimal RBC surface area-to-volume ratio.21,22 RBCs with impaired deformability cannot adapt their shape as needed to transit easily through the microcirculation, impairing the delivery of oxygen to and removal of carbon dioxide from the tissues. Reduced RBC deformability is also a major factor in reduced survival of the RBC following transfusion, as poorly deformable RBCs are culled from the circulation as they pass through the splenic circulation.23
Decrease in 2,3-diphosphoglycerate 2,3-Diphosphoglycerate (2,3-DPG) is a glycolytic intermediate present in RBCs that serves as a major modifier of hemoglobin (Hb) oxygen affinity in many species, including human beings, dogs, and rats.24 In contrast, feline RBCs contain very low levels of 2,3-DPG, and feline Hb does not require 2,3-DPG to modulate oxygen affinity.22 About 10–20% of the glycolytic intermediate 1,3-disphosphoglycerate is diverted from glycolysis for synthesis of 2,3-DPG. The 2,3-DPG molecule intercalates between the -globin chains of deoxyhemoglobin, stabilizing the deoxy form and shifting the base of the oxyhemoglobin dissociation curve to the right. This steepens the curve and moves the mid-point of the curve, the P50, to the right, causing oxyhemoglobin to release oxygen at a higher local pO2 . Following 2 weeks of storage, 2,3-DPG concentrations in human, canine, and rat RBCs are virtually depleted,5,25,26 shifting the oxygen dissociation curve to the left, increasing Hb oxygen affinity and theoretically impairing oxygen delivery to the tissues. De novo synthesis of 2,3-DPG begins with RBC rejuvenation in the hours following transfusion, with restoration of normal RBC 2,3-DPG within 3 days.27–29 Despite potentially adverse consequence of this delay in 2,3-DPG restoration, the results of experimental studies in baboons,30 rats,25,31,32 and healthy people33 have not supported a role of 2,3-DPG as a key factor for oxygen delivery and consumption capacity. For example, Raat et al25 evaluated the effect of storage time of human RBCs on intestinal microcirculatory oxygenation in a rat isovolemic exchange model. Although RBCs were devoid of 2,3-DPG by 2–3 weeks of storage, oxygen delivery to the intestines by those RBCs was the same as after transfusion with fresh RBCs. The same investigators found that oxygen C Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12252
RBC storage lesion
delivery by RBCs stored for 5–6 weeks had impaired oxygen-delivery capacity, indicating that increased oxygen affinity may occur by mechanisms other than depleted 2,3-DPG. Thus, although 2,3-DPG is rapidly and significantly depleted in stored RBCs, an adverse impact of this particular storage lesion on oxygen delivery has not been proven. Furthermore, this mechanism may be even less important in species that do not depend on 2,3-DPG to modify Hb oxygen affinity. For example, the feline Hb P50 is only slightly reduced after 35 days of storage suggesting that the storage lesion does not have a major impact on Hb oxygen affinity in this species.34
Nitric oxide Nitric oxide (NO) is an uncharged radical produced by the vascular endothelium that plays a critical regulatory role as a local vasodilator. It also affects homeostasis through inhibition of platelet aggregation, serves as a microbiocidal agent for host defense, inhibits endothelial adhesion molecule expression, and has antioxidant properties as recently reviewed.35 Within RBCs, NO binds to the 93 cysteine thiol residue of Hb to form S-nitrosohemoglobin (SNO-Hb).26 In the deoxygenated state, the reactivity of SNO-Hb is increased, and transnitrosation reactions transfer the NO group to glutathione or thiols on the RBC membrane protein anion exchanger1. These S-nitrosothiol compounds are released from RBCs at the tissue level in direct proportion to Hb oxygen desaturation and signal vasodilation, increasing flow to match regional perfusion and oxygen delivery with metabolic demand.36 During storage, RBC SNO-Hb bioactivity falls rapidly within RBCs and is not restored following infusion into a recipient, and some investigators have suggested that this may impair tissue oxygen delivery after RBC transfusions.26,37,38 Although this model linking Hb saturation with regional blood flow has gained wide acceptance, another laboratory using knockout mice with circulating human Hb in which the 93 cysteine thiol is replaced with alanine to eliminate direct Hb participation in SNO release found evidence that SNO-Hb is not essential for at least some aspects of RBCdependent vasodilation.39 However, that suggestion was based largely on measurements of secondary and tertiary physiological responses (including systemic blood pressure, right ventricular hypertrophy, and pulmonary arterial wall thickness), and others have argued that those measures cannot be used to discount SNO-Hb as an important physiological regulator of the microcirculation. ATP plays a role in the response of RBCs to tissue hypoxia through a vasodilatory mechanism that utilizes nitric oxide (NO).26,37,40,41 When circulating RBCs are heavily deoxygenated during transit through tissues with low ambient oxygen pressure, they are stimulated C Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12252
to release vasodilatory compounds to enhance local blood flow. This process facilitates optimal oxygen delivery in response to local oxygen demand. This process depends in part on the release of ATP from RBCs in response to oxyhemoglobin desaturation; the free plasma ATP stimulates production of NO in the endothelium.42 This RBC response to local hypoxia is impaired by ATP depletion as seen in stored RBCs, potentially contributing to inadequate regional perfusion secondary to stunted microcirculatory vasodilation.43 Enhanced chemical reduction of NO occurs in the recipient after transfusion of stored RBCs secondary to scavenging by free Hb and Hb-containing RBC microparticles (MPs) that accumulate due to hemolysis and RBC damage during storage.44–46 The rate of hemolysis during storage depends on the individual donor, the storage solution, and storage duration, with most human blood units exhibiting 0.3–0.4% hemolysis prior to administration.47 Following transfusion, this free Hb is then available to react with NO to form methemoglobin and nitrate. This reaction, normally slowed in the blood by cellular isolation of Hb from plasma by the cytoskeleton beneath the RBC membrane and the unstirred layer of the immediate extracellular fluid environment surrounding the cell, is then free to proceed rapidly.48,49 Hb-containing MPs also bring Hb into contact with NO much more rapidly than intact RBCs because of their much smaller size and altered membrane constitution.48,50 Donadee et al46 measured the ability of human free Hb and MPs to scavenge NO using stopped-flow spectroscopy and laser-triggered NO release, and found that free Hb and MPs react with NO about 1,000 times faster than erythrocyte-bound Hb. In a paired in vivo rat model, they demonstrated that even at concentrations below 10 mol/L, cell-free Hb can produce potent vasoconstriction due to the potent scavenging ability of free Hb.46 In addition to promoting vasoconstriction, this decrease in NO bioavailability may also enhance intravascular thrombosis, WBC adhesion and diapedesis, endothelial permeability, and smooth muscle proliferative responses after vascular injury.35 Baek et al51 transfused guinea pigs with old (28 d), or new (2 d) blood and demonstrated that administration of the old cells led to intravascular hemolysis, acute hypertension, vascular injury, and kidney dysfunction consistent with free Hb-mediated injury. The attenuation of these effects with the highaffinity Hb-scavenger haptoglobin confirmed the role of free Hb in the vascular and kidney pathophysiology.
Increased potassium in the supernatant Increased potassium accumulation in the supernatant of stored human RBC products occurs as ATP 189
R. Obrador et al.
concentrations fall and render ATP-dependent erythrocyte Na+ -K+ -ATPase pumps inactive. Increasing potassium concentration in the supernatant of stored human RBC products is recognized as potentially detrimental or even fatal during rapid infusion.52 Intracellular potassium concentration is low in both feline and canine RBCs and their RBCs membranes lack Na+ -K+ -ATPase pump activity.53,54 Therefore, significant potassium accumulation in canine and feline stored blood products is unlikely. However, many clinically normal Japanese and Korean dog breeds have high RBC concentrations of potassium associated with significant Na+ -K+ -ATPase pump activity,55 and prolonged storage of blood obtained from these breeds could be associated with accumulation of potassium in the supernatant.
as glycolysis declines during RBC storage.64 Under these circumstances, superoxide will interact with iron and water via the Fenton reaction to form hydroxyl radicals, which can in turn oxidize and damage proteins21,62,65 and lipids.66,67 For example, methemoglobin concentration in human banked blood increases 3-fold during long storage times.68 Since the duration of storage of RBCs is typically much shorter than the 110-day in vivo lifespan of human RBCs (100–115 d for dogs, 73 d for cats),69 the total burden of oxidative damage that accumulates during RBC storage is assumed to be low. However, because this burden is delivered all at once with transfusion (and repeatedly with massive transfusions), the potential to overwhelm endogenous antioxidant capacity may be significant.
Ammonia accumulation Ammonia concentration increases with time in both human and canine stored RBC products.56–58 Human RBCs are rich in ammonia containing three times more ammonia than that found in the plasma.59 The increase in ammonia in stored blood is attributed to the generation and release of ammonia from RBCs as well as deamination of amino acids and purines, such as adenine found in preservative solutions.60 Waddell et al58 found that plasma ammonia concentration in anemic dogs without primary liver disease did not increase following transfusion with 5–10 mL/kg of stored pRBCs and remained within the normal reference range. However, those dogs had normal liver function and received a relatively small amount of product, and patients with liver dysfunction or those receiving massive transfusions may be more at risk of developing hyperammonemia.58
Biomechanical Changes
Oxidation injury Oxidation injury to RBCs is yet another mechanism contributing to the storage lesion and is a plausible cause of MPs formation and the loss of RBC deformability.61,62 The high oxygen availability in blood provides a rich environment for oxidation of constituent molecules, and RBCs are particularly susceptible due to their role in carrying oxygen and lack of mitochondria. The binding of oxygen to RBC Hb is dynamic and oxygen is constantly dissociating from one Hb molecule and binding to another. This exchange is not perfect, and occasionally an oxygen molecule extracts an electron from the heme unit, forming ferric Hb (methemoglobin) and the superoxide radical (O2 − ).63 Normal RBCs contain the reducing agent glutathione and the antioxidant enzymes superoxide dismutase and methemoglobin reductase; these systems quickly dismutate the superoxide radical to hydrogen peroxide and reduce ferric heme to the functional ferrous form.12 However, glutathione concentrations fall 190
RBC shape change and reduced RBC deformability To deliver oxygen to tissues, the 6–8 m (depending on species) diameter RBC must flow through capillaries with diameters as small as 3 m. Therefore, the deformability of the biconcave disc is crucial for adequate perfusion, and even modest reductions in deformability may reduce RBC ability to traverse these narrow channels. During prolonged storage, RBCs undergo progressive shape change from a readily deformable biconcave disc to poorly deformable echinocytes with cell surface protrusions, and ultimately to nondeformable spheroechinocytes.70 Although initially reversible, once beyond the early spheroechinocyte stage, the loss of deformability becomes permanent as the RBCs lose membrane through the budding of MPs from the tips of the echinocytic spines.71 The etiology of these morphologic changes and microvesiculation observed during storage appears to be multifactorial and several mechanisms have been proposed: depletion of ATP25,72 and 2,3-DPG,73,74 loss of membrane phospholipid with associated MPs,62,75 protein rearrangement,64 and lipid oxidation.76 As storage time increases, these corpuscular changes also increase RBC adherence to the endothelium as described later.77,78 These shape changes during RBC storage are associated with rheological changes, increased viscosity, and reduced capillary perfusion,77 all of which can jeopardize tissue oxygenation. Accumulation of MPs MPs are small anucleate phospholipid-rich vesicles that contain membrane-bound and cytoplasmic proteins from the cell from which they derived. MPs are normally shed from reticulocytes and healthy RBCs. Normal RBCs use the microvesiculation process as a way of shedding oxidized and polymerized lipids.79,80 In stored RBC C Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12252
RBC storage lesion
products, the concentration of RBC MPs increases with time, with relatively greater concentrations found toward the end of the storage period.9 RBC-derived MPs that accumulate during storage have shown procoagulant activity.81 In contrast to their parental cells, the majority of MPs expose negatively charged phospholipids, primarily phosphatidylserine, on their outer membrane. These negatively charged phospholipids promote assembly of coagulation complexes such as tenase or prothrombinase, initiating and propagating thrombin generation.81 MPs also express abundant tissue factor activity, further contributing to a procoagulant state.82 Additionally, RBC-derived MPs may also have a proinflammatory activity through modification of platelet–WBC interactions in transfused patients, based on altered membrane properties of RBC MPs that enhance platelet interactions to increase inflammatory chemokine bioavailability in vitro.83,84 Adhesion to endothelial cells Normal RBCs do not readily adhere to intact blood vessel walls, but this characteristic may be altered by storage as has been suggested by the results of studies using in vitro models to simulate blood flow. Independent groups have demonstrated that stored human RBCs have enhanced adherence to vascular endothelial cells (ECs) and that the concentration of adherent RBCs increases with storage time.77,85,86 If this occurs in vivo, enhanced adhesion of RBCs to vascular endothelium might disrupt local blood flow patterns, decrease oxygen delivery to peripheral tissues, and perhaps even occlude microvessels.87 The expression of procoagulant phosphatidylserine on the outer RBC membrane increases with both RBC age and storage time and can contribute to endothelial adherence.77,88 Storage lesion associated RBC corpuscular changes and loss of RBC deformability contributes to endothelial cell adherence and the mechanical obstruction of microvessels.77,86 Prestorage leukoreduction abrogates excessive stored RBC adhesion to endothelium in these in vitro models, implicating donor WBCs in the development of this phenomenon.85,89 Inflammation in the recipient may worsen storage-associated RBC endothelial adhesion, as suggested by the finding that pretreatment of human ECs with endotoxin to activate the ECs and mimic infection in the in vitro flow models results in increased adhesion of RBCs to the activated ECs.90 Many patients receiving stored RBC transfusions have concomitant systemic inflammation adversely affecting their microcirculation; storage-related RBC adhesion may have significant clinical implications for tissue perfusion and oxygenation in those recipients.
C Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12252
Immunologic Changes Overview An association between transfusion therapy and alterations in recipient immune function was first reported in people almost 40 years ago, when Opelz et al91 observed improved renal allograft survival following pretransplant allogeneic RBC transfusion. This work demonstrated the presence of an immunosuppressive or immunomodulatory effect of RBC transfusion on the recipient (later referred to as transfusion-related immunomodulation [TRIM]). Multiple studies in animals and people have demonstrated that allogeneic transfusion downregulates cellular immunity and dysregulates innate immunity during inflammation.92–94 Allogeneic blood transfusion stimulates the T helper 2-type immune response as evidenced by subsequent increase in interleukin (IL)-10 and IL-4 cytokine secretion posttransfusion of stored human RBCs.94 IL-10 and IL-4 both downregulate the T helper 1 response, thus downregulating cellular immunity functions such as antigen presentation.92 The predominant mechanism of TRIM in any particular individual likely depends on the interplay of transfusion effects with genetic predisposition and intercurrent illness in that recipient. Platelets and vascular ECs also potentially contribute to TRIM as both cell types are highly responsive to inflammatory signals and when activated release potent bioactive mediators. The “two-insult” model of posttransfusion injury proposes that an initial insult (eg, an underlying inflammatory condition or trauma) primes the patient’s immune cells or endothelium, and further inflammation is triggered by a second immunomodulating insult (eg, transfusion), resulting in an unrestrained inflammatory response.95,96 A 2012 study by Tung et al97 corroborated this “two-insult” theory of TRIM as a cause of transfusion-related acute lung injury (TRALI) in an in vivo sheep model. In that study, sheep were randomized to first receive saline or lipopolysaccharide (LPS) and then further randomized to receive supernatant from fresh (1 d) or stored (42 d) nonleukoreduced human pRBCs or saline. The sheep were then monitored for hypoxemia within 2 hours of transfusion or histological evidence of pulmonary edema. None of the saline control group animals developed TRALI after transfusion with any product. Similarly, none of the sheep that were primed with LPS and then transfused with saline developed TRALI. While only 20% of the LPS sheep transfused with fresh RBCs developed TRALI, 80% of the LPS sheep that received stored RBCs did. In this model, both the priming of the recipient and the transfusion of stored blood resulted in TRALI.
191
R. Obrador et al.
Leukocyte contamination and soluble immune response modifiers Nonleukoreduced stored RBCs units contain abundant passenger leukocytes; these cells are immunologically active and their release of proinflammatory and prothrombotic compounds increases with storage time.98,99 Compounds whose concentrations increase with storage time include elastase, secretory phospholipase-A2, soluble human leukocyte antigen, soluble FAS-ligand, prostaglandin E-2, transforming growth factor 1, and proinflammatory cytokines IL-1, IL-6, and IL-8. 11,100,101 When leukocytes are stored in the increasingly acidic supernatant of refrigerated stored RBCs units, they become activated, produce cytokines, and eventually die. Release of enzymes such as phospholipase-A2 following cell death may potentiate the production of lysophospholipids including the dialkylglycerol platelet activating factor.102 These lysophospholipids are capable of priming and activating neutrophils, and induce various proinflammatory actions in leukocytes, ECs, and smooth muscle cells.66,103 To appraise the immunological activity of stored human RBC supernatant, Sparrow and Patton98 evaluated in vitro neutrophil priming activity by measuring expression of CD11b, IL-8 release, and neutrophil chemotaxis after incubation of purified neutrophils with 3 stored allogenic RBC products: nonleukoreduced RBCs, buffy coat poor RBCs, and RBCs that underwent prestorage leukoreduction. The leukoreduced and buffy coat poor RBC products did not induce significant neutrophil priming behavior for all 3 assays. In contrast, incubation of the purified neutrophils with the nonleukoreduced RBC product (WBC count 8,400 times higher than leukoreduced RBCs) revealed CD11b exposure, release of IL-8, and neutrophil chemotaxis consistent with speculation that proinflammatory mediators are present in stored RBC supernatant. Prestorage leukoreduction effectively removes 99.9% of donor WBCs with the goal of abrogating any immunomodulatory effects resulting from donor leukocytes and their associated bioreactive mediators. Both bench104 and clinical105–110 studies of the benefits and consequences of prestorage leukoreduction have yielded mixed results. Fergusson et al108 used meta-analysis of clinical trials to determine that RBC leukoreduction reduces postoperative infection in people, but found no reduction of cancer recurrence or perioperative mortality in recipients. In a retrospective cohort study comparing outcomes in transfusion recipients undergoing major surgeries before (n = 6,982) and after (n = 7,804) institution of universal leukoreduction, the same investigators found that leukoreduction was associated with a decreased incidence of fever episodes, antimicrobial use, and mortality in those high-risk patients.106 The use of leukoreduced blood has also been 192
associated with fewer early rejection episodes in patients with liver transplantation, irrespective of the type of immunosuppression suggesting a possible immunomodulatory effect of leukoreduced blood.109 However, not all studies have shown a benefit of using leukoreduced blood. Leukoreduction provided no measurable clinical benefit in humans infected with immunodeficiency virus or cytomegalovirus,105 did not decrease the rate of postoperative infection or length of hospitalization in patients undergoing elective surgical procedures,110 and was not associated with significant reductions in neonatal intensive care units mortality or bacteremia in premature infants requiring RBC transfusions.107 In veterinary medicine, preliminary studies have evaluated the effectiveness of commercially available leukoreduction filters at removing WBCs from stored blood and the effects of leukoreduction on canine RBCs viability and recipient inflammatory response.111,112 Prestorage leukoreduction of cooled canine blood effectively removed 99.99% of WBCs while preserving RBC viability as determined by percent hemolysis, biotin-labeling assays of cell survival, and flow cytometry.111 More recently, McMichael et al112 demonstrated a profound decrease in the inflammatory response in normal dogs to transfused stored (21 d at 4˚C) autologous leukoreduced RBCs (vs. nonleukoreduced) as determined by posttransfusion WBC count, fibrinogen, and C-reactive protein concentrations.
Clinical Consequences of the RBC Storage Lesion Transfusion therapy has many known potential adverse consequences, and the risk of harm from transfusion increases with increasing storage time.113–115 Although the FDA guidelines for maximum storage time are based on 75% survival of RBCs 24 hours following administration,4 there is growing evidence that longer storage times within this FDA-approved time frame compromise RBC oxygen delivery and produce adverse clinical consequences for the recipient. Effects on oxygen delivery kinetics Tissue oxygen delivery is a function of RBC oxygen carrying capacity and RBC flow in the macro- and microcirculation. The immediate goal of RBC transfusions is to increase oxygen delivery in critically ill patients; the effectiveness of this therapy in achieving this goal following RBC storage warrants investigation in light of abundant experimental evidence that storage depresses the ability of RBCs to deform, unload oxygen peripherally, and participate in control of regional blood flow. Most of the animal experiments addressing this issue have been performed on rats and follow an early study that demonstrated that storage for 28 days prevented C Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12252
RBC storage lesion
transfused RBCs from improving tissue oxygenation following administration to septic rats.116,117 However, rat RBCs age about 4 times more rapidly during storage than human RBCs,118 making it difficult to extrapolate these results to other species. In an effort to see if this effect occurred in RBCs from other species, Raat et al25 used a hemodiluted rat model (hematocrit 15%) to confirm this phenomenon with isovolemic exchange transfusion using fresh (2–6 d of storage), intermediate (2–3 wk), or old (5–6 wk) stored human RBCs. They demonstrated that isovolemic exchange transfusion with either fresh or intermediate RBCs maintained intestinal tissue oxygen tension but transfusion with old cells reduced it by 25%. Similarly, Tsai et al119 found that transfusion with autologous hamster RBCs stored for 28 days reduced microvascular flow, functional capillary density, and oxygen extraction compared with fresh RBCs. These studies in healthy animals under oxygen-restricted conditions suggest that older transfused RBCs may have a limited ability to acutely improve oxygen availability, but scrutinizing these concepts in a clinical setting is necessary to determine the importance of this phenomenon. Toward that end, Marik et al120 used gastric tonometry in a prospective, controlled study to measure gastric mucosal pH in septic human patients to determine that transfusion with RBCs stored longer than 15 days reduced gastric mucosal pH, an indicator of splanchnic hypoxia, indicating lower oxygen availability secondary to impaired microcirculatory perfusion despite increased blood Hb concentration. These authors suggested that transfusion of old, poorly deformable RBCs leads to microcapillary sludging and obstruction, resulting in gut ischemia. Contradicting these results, Walsh et al121 could not prove clinical worsening of tissue hypoxia via tonometric assessment of gastric mucosal oxygenation in a ventilated critically ill human population after transfusion with RBCs stored for more than 3 weeks compared with transfusion with fresh cells. This trial differed from the former as patients were randomized, received filtered leukodepleted RBCs, and were not septic (and therefore may not have exhibited the same degree of oxygen supply dependency as the patients in Marik’s study). Tinmouth et al122 examined 19 clinical human studies in a systematic review of the effect of RBC transfusions on oxygen delivery and consumption before and after transfusion of a delineated number of RBCs to critically ill patients, but the included studies did not assess the influence of RBC storage time. Although oxygen delivery was consistently increased by transfusion, oxygen consumption increased in only 6 of the 19 studies. This disparity between oxygen delivery and consumption illustrates the relative importance of recipient responses to RBC transfusion, as it is possible that impaired oxygen utilization from conditions such as sepsis are more C Veterinary Emergency and Critical Care Society 2014, doi: 10.1111/vec.12252
important determinants than the storage lesion in some patients. Clinical consequences for the recipient – human literature The bulk of the literature exploring the clinical consequences of RBC storage lesion in people is derived from observational studies. Many of these were examined in a 2012 meta-analysis of 21 studies involving 409,966 patients by Wang et al,113 who assessed the safety of transfused older (range 9 to 31 d) versus newer (range 1.6–21 d) stored blood, predominantly in cardiac surgery and trauma patients. Eighteen of these studies were observational (12 retrospective, 6 prospective) and 3 were randomized controlled trials (RCTs). The primary outcome of interest was mortality. These authors found that the transfusion of older stored RBCs was associated with a significantly increased risk of death (odds ratio, 1.16; 95% confidence interval, 1.07–1.24). Further subgroup analysis of the trials found no patterns associated with patient type, size of trial, or amount of blood transfused. Pooled data from this meta-analysis suggest that older blood was also associated with an excess of multiple organ dysfunction syndrome (MODS), pneumonia, renal dysfunction, and sepsis. Numerous clinical trials and retrospective studies have examined the clinical consequences of stored RBC transfusions to patients with a variety of disorders. Investigators studying critically ill people following traumatic injury, cardiac surgery, and suffering from sepsis and other disorders have reported an independent direct association between increasing RBC storage time and risk of infection, deep vein thrombosis (DVT), ICU length of stay, and mortality in both adults and children.114,115,123–130 RBC storage times greater than 14–28 days were associated with worse outcomes. In a large retrospective study of cardiovascular surgery patients, Koch et al114 examined the relationships between transfusion and a diverse composite of 18 serious adverse complications including arrhythmias, thrombotic events, organ failure, sepsis, prolonged ventilation, and in-hospital death. Those complications were more common in patients who received older units (>14-d old) than those who received fresher units (14-d old). Patients given older blood had higher rates of in-hospital mortality and higher rates of extended intubation, renal failure, and sepsis. The significance of these findings is unclear since confounding factors (eg, concurrent mismatched platelet transfusions) were present and the treatment groups were not ideally balanced: more patients who received RBCs stored >14 days also received more blood transfusions. Several retrospective studies have been conducted to characterize morbidity and the complications of 193
R. Obrador et al.
MODS and multiple organ failure (MOF) associated with transfusion of stored RBCs in specific patient population.123–125,127 An early study by Zallen et al125 identified storage time as an independent risk factor for development of postinjury MOF in adult trauma patients.125 A greater risk of developing new or progressive MODS in hemodynamically stable critically ill pediatric patients receiving stored RBCs (>2–3 wk) has also been established.127 Given the experimental evidence of TRIM, clinical researchers have conducted studies examining the relationship between transfusion of stored RBCs and infection. The safety of stored transfusion products in trauma patients is particularly relevant as these patients often require massive transfusion, increasing the likelihood of receiving older blood, and have wounds predisposing them to infection. In fact, an increased incidence of infection in trauma patients receiving stored RBCs has been established by multiple observational studies.124,131,132 For example, Juffermans et al131 found an association between transfusion with RBCs stored more than 14 days and increased risk of bacterial infection after trauma; this risk was not reduced by administration of prophylactic antimicrobials. There was no association between infection and platelet transfusion in the same study. In another observational study of incidence of postoperative infection in patients undergoing coronary artery bypass grafting, Andreasen et al133 identified an increased risk following transfusion of buffy coat reduced RBC units stored 14 days compared with units stored
