Allogeneic transfusion of whole blood and fractionated blood components remains a controversial topic with respect to transfusion triggers and practices. In spite of widespread use, data supporting specific practice paradigms for whole blood and component therapy transfusion are lacking. Opinions are embraced and defended, but there is little evidence to suggest that increasing hemoglobin concentration, arteriolar oxygen content, or global oxygen delivery increases survival in patients, other than in 2 specific populations, as follows: (1) those with active hemorrhage, and (2) those with active cardiac ischemia.
Further studies, especially in the civilian trauma centers, are needed to confirm the lessons learned from the military experience. To reduce morbidity and mortality as a result of uncontrolled hemorrhage, patients needing a massive transfusion need to be quickly identified so that immediate interventions can prevent the development of coagulopathy. 
The history of blood transfusion originated with William Harvey’s discovery of blood circulation in 1628. The earliest known blood transfusions occurred in 1665, and the first human blood transfusion was performed by Dr. Philip Syng Physick in 1795. The first transfusion of human blood for the treatment of hemorrhage was performed by Dr. James Blundell in London in 1818. The first blood bank was established in Leningrad in 1932, and the first blood bank in the United States opened at Chicago’s Cook County Hospital in 1937.
Technology making the transfusion of allogeneic blood products feasible includes Karl Landsteiner’s landmark identification of the human blood groups A, B, and O in 1901. Decastello and Sturli added the fourth group, AB, in 1902. Reuben Ottenberg used blood typing and cross-matching for the first time shortly thereafter; he also coined the terms universal donor and universal recipient in 1912. Subsequently, the development of long-term anticoagulants, such as acid-citrate-dextrose, allowed preservation of blood for later use. In 1939-1940, the Rhesus (Rh) blood group system was discovered, leading to the development of minor antigen compatibility testing. In 1971, hepatitis B surface antigen testing heralded the advent of screening to minimize infection transmission complicating allogeneic transfusion.
Fresh whole blood has long been thought of as the criterion standard for transfusion, but the advent of whole blood fractionation techniques subsequent to World War II provided a means of more efficient use of the various components (ie, packed red blood cells [PRBCs], fresh frozen plasma [FFP], individual factor concentrates, platelet concentrates, cryoprecipitate). As a result, current indications for the use of whole blood are increasingly narrow. The most widespread use of whole blood transfusion is the US military buddy transfusion system, best known as the US Navy’s walking blood bank onboard naval warships, such as the USS Cole.
Anemia and hemorrhagic shock defined
Active hemorrhage resulting in shock is one of the few evidence-based established indications for transfusion. Anemia describes a diminished circulating red cell mass, expressed as grams of hemoglobin per 100 cc of whole blood. Anemia may arise as a result of external loss, internal destruction, inadequate production, or a combination of these entities. While most patients experiencing active hemorrhage become anemic, anemia in itself is not necessarily an indication for transfusion. The result of severe bleeding is hemorrhagic shock, and shock is defined as the inadequate supply of oxygen to support cellular metabolism. Red cell mass repletion is only one facet of therapy for hemorrhagic shock and may not be used in every case.
Goal of PRBC transfusion
Based on the classic definition of shock, the time-honored principles of effective resuscitation from hemorrhage describe elimination of oxygen debt, anaerobic metabolism, and tissue acidosis. When the etiology is hemorrhage, the goal of transfusion is restoration of the tissue oxygenation that was compromised by the loss of hemoglobin and oxygen-carrying capacity. The goal is not restoration of a specific (and arbitrary) hemoglobin level (ie, "trigger"). Instead, transfusion should reflect the application of a therapy that targets a physiologically identifiable and achievable goal.
The decision to transfuse red cells should be based on a logical thought process with the goal of restoration of tissue oxygenation. Therefore, transfusion of red cells logically commences under the following circumstances:
Clinical evidence of hypoxia/dysoxia exists, manifested by hypoperfusion, including lactic acidosis and increased base deficit (when not due to hyperchloremic metabolic acidosis). In addition, preload responsive cardiac performance must be corrected by appropriate plasma volume expansion exclusive of red cell mass restoration. In this circumstance, pressor agents may be used to improve cardiac performance if it is judged to be inadequate, and red cell mass transfusion may be used to support end-organ oxygen delivery.
Active hemorrhage associated with shock
Hemorrhage cannot be immediately controlled because of anatomical constraints, coagulopathy, or location in an austere environment precluding hemorrhage control, and PRBC transfusion may prolong life until hemorrhage control is accomplished.
Guidelines on red blood cell transfusions from the AABB (formerly, the American Association of Blood Banks) advise a restrictive strategy for stable adults and children. 
Recommendations from the AABB guidelines include:
For ICU patients (both adults and children), transfusion should be considered at hemoglobin concentrations of 7 g/dL or less.
For postsurgical patients, consider transfusion when hemoglobin is 8 g/dL or less or for symptoms (eg, chest pain, orthostatic hypotension, tachycardia unresponsive to fluid resuscitation, congestive heart failure).
Also consider transfusion for hemoglobin concentrations 8 g/dL or less in hospitalized patients who are hemodynamically stable with pre-existing cardiovascular disease.
The AABB does not recommend either a liberal or restrictive threshold for transfusion for hospitalized, hemodynamically stable patients with the acute coronary syndrome.
Physiologic triggers, as elucidated above, are the most accurate predictors of transfusion requirements, as they are based on the patient’s specific needs with respect to deranged physiology. However, the desire to establish a generic "number to treat" with respect to hemoglobin and hematocrit has permeated transfusion practice. Much of the controversy surrounding transfusion practice paradigms centers on disagreement as to what constitutes the proverbial perfect number.
A quarter of a century ago, optimal treatment of surgical and critically ill patients targeted hemoglobin levels greater than or equal to 10 g/dL and hematocrit values greater than or equal to 30%. Subsequent understanding of the risks inherent in transfusion prompted investigations designed to reestablish a minimum baseline for acceptable hemoglobin concentrations.
At hemoglobin levels below 3.5-4 g/dL, mortality significantly increases in otherwise healthy patients.
Work by Shander indicates decreased cognition with hemoglobin levels below 5 g/dL. 
The American Society of Anesthesiologists uses hemoglobin levels of 6 g/dL as the trigger for required transfusion, although more recent data suggest decreased mortality with preanesthetic hemoglobin concentrations of greater than 8 g/dL, particularly in renal transplant patients.
Restrictive transfusion strategies have been supported by the Transfusion Requirements in Critical Care (TRICC) trial, published in 1999, as well as others. The TRICC trial documented an overall trend toward decreased 30-day mortality and significantly decreased mortality among patients who were less acutely ill and among patients younger than 55 years in the group using a hemoglobin transfusion trigger of 7 g/dL compared with the more liberally transfused group. The investigators concluded that a restrictive transfusion strategy is at least as effective as and possibly superior to a liberal transfusion strategy in critically ill patients. The exception to this paradigm is patients with acute myocardial infarction and unstable angina. A 2012 Cochrane Database review reinforces these notions. 
The CRIT study, published in 2004, is a prospective, multiple center, observational cohort study of intensive care unit (ICU) patients in the United States, which investigated the relationship of anemia and RBC transfusion to clinical outcomes. The investigators found that the number of RBC units transfused is an independent predictor of worse clinical outcome.
Whole blood is fractionated into specific components, as follows: PRBC, FFP, platelet concentrates, and cryoprecipitate; FFP may be further fractionated into individual factor concentrates as well. Fractionation maximizes the clinician’s ability to rationally use the components of each donated unit while simultaneously limiting unnecessary transfusions. A specific product may also be transfused with less volume. Additionally, the individual components require different storage temperatures; therefore, fractionation allows more effective product management. 
Blood component fractionation is based on centrifugation and flash-freezing technology. Whole blood is separated into red cells and platelet-rich plasma by slow centrifugation. High-speed centrifugation is then applied to the platelet-rich plasma to yield one unit of random donor platelets and one unit of FFP. FFP yields cryoprecipitate via a slow thaw process to precipitate the plasma proteins, which are then separated by centrifugation. Cryoprecipitate contains high concentrations of fibrinogen, factor VIII, factor XIII, von Willebrand factor, and fibronectin; hypofibrinogenemia is the most common transfusion indication for cryoprecipitate in the critical care environment or in the operating room.
Apheresis technology may be used to collect multiple units of platelets from a single donor. Single-donor apheresis platelets contain the equivalent of at least 6 units of random donor platelets and often have fewer inadvertently included leukocytes than pooled random donor platelets. Apheresis is most commonly used to obtain platelets for use in alloimmunized patients with a dense antibody presence that makes cross-matching difficult. Patients with blood dyscrasias and malignancies commonly fall into this category.
Packed red blood cells
Indications for transfusion of PRBC are reviewed above. In general, hemoglobin concentration (typically reported in g/dL) is used to monitor RBC mass. This is a directly measured variable, whereas hematocrit is a calculated value when obtained from modern automated devices and, therefore, more prone to inaccuracy compared to direct measurement in a capillary tube spun on a centrifuge.
Fresh frozen plasma
The transfusion of FFP is common, but the specific indications for its use are limited. In fact, evidence for its use in numerous clinical situations, such as prophylaxis in nonbleeding patients, is limited. FFP transfusion is indicated in hemorrhaging patients to replace labile and lost coagulation factors. Clinical circumstances fulfilling this criterion include massive transfusion, cardiopulmonary bypass, extracorporeal pulmonary support techniques, decompensated liver disease, or acute disseminated intravascular coagulation regardless of cause. FFP, in conjunction with vitamin K, is also indicated for excessive warfarinization in circumstances accompanied by life-threatening hemorrhage.
A guideline for initial FFP dosing is 10-15 mL/kg; this typically translates to at least 4 units of FFP to effect a therapeutic response. Efficacy is monitored by laboratory tests of coagulation function, including prothrombin time (PT), activated partial thromboplastin time (aPTT), and the international normalized ratio (INR).
Platelet transfusion may be beneficial in patients with platelet deficiency or dysfunction. Prophylactic platelet transfusion is indicated in patients with bone marrow failure, no other associated risk factors for bleeding, and platelet counts below 10 X 109/L. If there are associated risk factors, the threshold may be reasonably raised to 20 X 109/L. Patients undergoing invasive procedures should have platelet counts greater than 50 X 109/L. In the hemorrhaging patient, platelet transfusion is indicated when thrombocytopenia is contributing to the bleeding and the platelet count is less than 50 X 109/L. When diffuse microvascular bleeding is present, the platelet count should be maintained above 100 X 109/L while the underlying cause of the hemorrhage is being addressed.
Common etiologies to be corrected include, but are not limited to, large volume hemorrhage control failure (solid organ or vascular conduit), hypothermia, acidosis, traumatic brain injury, individual factor deficiencies, and acquired inhibitors of coagulation. The optimal time to measure the effect of platelet transfusion is 1 hour after the completion of the infusion. This timeframe allows one to discern an appropriate increase from ongoing consumption from total destruction due to preformed antibody.
Transfusion of cryoprecipitate is indicated for fibrinogen deficiency or dysfibrinogenemia in the setting of hemorrhage, invasive procedures, injury, or acute disseminated intravascular coagulation. Fibrinogen levels should be monitored and treatment undertaken for levels less than 100 mg/dL; many clinicians use a higher threshold of 150 mg/dL in patients with active hemorrhage. Cryoprecipitate is generally transfused in aliquots of 10 units. Patients on a massive transfusion protocol and receiving greater than 10 units of FFP generally do not need additional cryoprecipitate, having received an adequate bolus of fibrinogen in the large quantity of FFP.
First described by Hartert in 1948, thromboelastography records viscoelastic changes that occur during coagulation. A small volume of whole blood or plasma is placed into a mortar style cup, and a pestle is then inserted. The pestle rotates back and forth on a torsion wire that is attached to a monitoring device. As a clot forms, the pestle rotation is progressively retarded creating increased torsion on the torsion wire and a deflection on the tracing. The fibrin polymerization process is graphically depicted, including the rate of fibrin polymerization as well as overall clot strength and retraction.
Thromboelastography (TEG) evaluates clot initiation, formation, and stability, using whole blood or plasma. TEG has been used primarily to monitor blood component therapy during surgery. Initially used in hepatic transplantation, the technique has been used during cardiac surgery as well as damage control surgery after injury. Within 30 minutes, TEG provides a representation of the sum of platelet function, coagulation proteases and inhibitors, and fibrinolytic system. Analysis of the tracing provides a means to assess the need for blood component therapy. Modern devices substitute a computer graphic for the old-style roller drum paper tracing. Each element of the TEG tracing relates to a different aspect of clotting, as follows: the time required for clot formation underscores the need for FFP, clot strength guides platelet therapy, addition of heparinase assists in determining protamine dosage, and the degree of clot lysis determines the need for antifibrinolytic therapy.
The definition of massive transfusion has evolved over time to reflect modern transfusion practice.  Thirty years ago, massive transfusion was defined as more than 10 units of blood over 24 hours, roughly equivalent to one patient blood volume for an average weight person. Although one patient blood volume in 24 hours remains the “classic” definition, recent authors expand this definition to reflect up to 50 units of blood in 24-48 hours.
When initially introduced as a treatment modality in the 1960s, massive transfusion resulted in poor survival rates (6.6%) through the 1970s. However, survival rates up to 60% are now observed, as a result of early recognition and aggressive correction of coagulopathy, liberal use of rewarming techniques, institution of damage control operative management, and increased use of component transfusion therapy.
Hypovolemic shock secondary to hemorrhage is the most frequent indication for massive blood transfusion. Clinical scenarios in which hemorrhagic shock is encountered include trauma, operative misadventure, ruptured aortic aneurysm, massive gastrointestinal hemorrhage, and solid organ transplantation (primarily hepatic), but they are expanding to include angiointerventional and endoscopic procedures. Thus, massive transfusion may occur outside of the operating room and the surgical ICU with increasing frequency.
Recombinant activated factor VII (rFVIIa) is synthesized human factor VII that is available for reconstitution and infusion in patients with massive hemorrhage. rFVIIa has typically been used to treat hemophilia and other congenital and acquired coagulopathies. More recently, rFVIIa has been used in patients with active hemorrhage and coagulopathy from trauma, traumatic brain injury, excessive warfarin use, and other acquired hematologic defects, including acquired factor inhibitors.
Use in trauma patients results in a decrease in RBC requirement and in a trend toward improved survivorship and reductions in critical morbidities. rFVIIa initiates the extrinsic coagulation pathway only when complexed to tissue factor at sites of injury. Since tissue factor is present in only limited quantities within the general circulation, rFVIIa is generally regarded as safe with respect to thrombosis induction. However, some early reports from the R. Adams Cowley Shock Trauma Center document thrombotic complications that have not been previously reported. Only additional time and exposure of patients to rFVIIa will determine if these events are site- or population-specific. The dosing guidelines for trauma (general range, 90-120 mcg/kg of body weight) have yet to be established, whereas evidence-based guidelines are established for hemophilia as well as prostate surgery. Similarly unclear is the role of repeat dosing in the trauma patient population.
Complications of Transfusion
In countries with a high human development index (HDI) (based on life expectancy, education, and per capita income), transfusion of blood products is now very safe with respect to viral transfusion-transmissible infections. This is due to the increased sensitivity of pathogen testing, which reduces infectious window periods. Risk is significantly increased in countries with a low HDI because of high seroprevalence and inadequate pathogen testing and transfusion standards. The incidences listed here apply only to high HDI countries.
The incidence of hepatitis A is 1 case per 1 million units transfused (1:1 million); hepatitis B ranges from 1:6000 to 1:320,000. Acute disease develops in one third of patients infected with hepatitis B, but chronic infections develop in fewer than 10% of those infected. The incidence of hepatitis C ranges from 1:1.2 million to less than 1:13 million. However, in contrast to hepatitis B, greater than 80% of hepatitis C infections become chronic, with significant subsequent mortality attributed to cirrhosis and hepatocellular carcinoma.
The incidence of human immunodeficiency virus (HIV) ranges from 1:1.4 million to 1:11 million.
Human T-cell leukemia virus type I (HTLV-I) and human T-cell leukemia virus type II (HTLV-II) carry an incidence of 1:250,000 to 1:2 million. Of note, blood stored for more than 14 days and noncellular blood components have not been documented to transmit HTLV.
Cytomegalovirus (CMV) is the most common virus transmitted through transfusion, with an incidence ranging from 1:10 to 1:30. Because of its prevalence, transmission carries little risk to the general population. Patients at risk for developing disseminated CMV infection include neonates weighing less than 1200 g, seronegative pregnant women, recipients of seronegative allogeneic transplants, patients with advanced HIV infection, patients with lymphoma, and patients receiving chemotherapy.
Epstein-Barr virus (EBV) is also commonly transmitted due to its seroprevalence. The incidence of transmission is 1:200.
West Nile virus has been recently reported, with an incidence of 1:3000 to 1:5000.
The most common parasitic infection transmitted by transfusion is malaria, with an incidence of 1:4 million. However, in low HDI countries (where malaria tends to be endemic), the incidence of transfusion-transmitted malaria is as high as 1:3.
The first two cases of transfusion-transmitted variant Creutzfeldt-Jakob disease were described in 2004. Since prions are associated with a 6- to 7-year latency period, recognition of increased prevalence in donor pools may be delayed.
Bacterial contamination is responsible for at least 10% of transfusion-associated deaths and most infection-related deaths. Yersiniaenterocolitica is the most common bacterial contaminant of PRBC; other pathogens include Serratiamarcescens, Pseudomonasaeruginosa, and Enterobacter species. Incidence ranges from 1:200,000 to 1:4.8 million and is directly related to the length of storage.
Platelet-related sepsis is even more common, with an incidence as high as 1:2000 to 1:3000. Risk is greatest in transfusion of pooled platelet concentrates from multiple donors when compared to single-donor platelet transfusion. Because the risk increases with time, the shelf life of stored platelets should not exceed 5 days. Typical pathogens include Staphylococcusaureus and Staphylococcusepidermidis. Therefore, sepsis following platelet transfusion may be unrecognized, since these are common pathogens and infection may be attributed to other sources, such as device-related blood stream infections.
Nonhemolytic febrile transfusion reaction is an acute complication of blood transfusion characterized by fever with or without chills and rigors. This condition is not life-threatening. Most of these reactions derive from an immunologic response against the transfused cellular or plasma components, typically leukocytes. Leukocyte-depleted blood products may minimize this problem, but they have yet to be documented to universally do so.
Noncellular blood components (ie, plasma, plasma derivatives) are rare causes of adverse reactions. Plasma reactions may be related to an immunologic discord between the donor and the recipient. Transfused plasma proteins may contain epitopes different from those on the recipient’s own synonymous plasma proteins. Antibodies may also be in the donor plasma that react with the recipient’s blood cells or plasma proteins. Contaminants in donor plasma may also be associated with some of these reactions. Processing of plasma may lead to activation of some of the proteolytic systems, such as complement and kinin/kininogen systems, which, in turn, leads to the generation of vasoactive substances and anaphylatoxins. Finally, histamine levels may be increased in stored blood components. Symptoms range from minor urticarial reactions or flushing to fulminant cardiorespiratory collapse and death. Some of these reactions are likely true anaphylaxis, but, in others, the mechanism is less clear, and theyarereferred to as anaphylactoid.
Posttransfusion purpura occurs when platelet-specific alloantibodies develop 5-10 days posttransfusion. The patient’s native platelets are destroyed, resulting in severe thrombocytopenia. This complication is rare but is potentially life-threatening, and it typically occurs in women. Platelet transfusions are usually ineffective, and high-dose intravenous immunoglobulin (2 g/kg over 2-5 d) is required. Some practitioners use plasmapheresis or a high-dose steroid for this condition, but they do so without strong data supporting an outcome benefit.
ABO incompatibility is the most common, potentially fatal complication of blood transfusion. Most acute hemolytic transfusion reactions result from human error, usually incorrect patient identification. Many system improvements have been proffered to reduce the human error component, including bar coding and computer-assisted matching techniques, but no single system has arisen as the most effective method of eliminating error.
The reported incidence of ABO incompatibility in human transfusion practice is 1:6000 to 1:33,000. Fatal cases occur at a rate of 1:250,000 to 1:1 million.
The classic symptoms of an acute hemolytic transfusion reaction include apprehension, flushing, pain, nausea, vomiting, rigors, hypotension, and circulatory collapse. Disseminated intravascular coagulation, hemolytic anemia, renal impairment, and jaundice may also be noted. In the anesthetized patient, the presentation may be readily confused with other clinical entities and is characterized by unexplained hypotension, diffuse coagulopathy, and hemoglobinuria. Treatment includes cessation of transfusion and aggressive resuscitation from shock.
Most delayed hemolytic transfusion reactions are unpreventable because the blood is serologically compatible at the time of transfusion, but some cases are due to antibodies to minor red cell antigens that were simply not detected by the routine pretransfusion antibody screening assay. The reported incidence is 1:2000 to 1:11,000. These reactions occur 3-10 days after transfusion, and patients present with fever, hyperbilirubinemia, and a decreasing hematocrit. When these symptoms occur in the appropriate clinical setting, blood should be re-crossmatched prior to the administration of any further components.
The term transfusion-related acute lung injury (TRALI) was originally coined in 1983 to describe a specific group of clinical and laboratory features identified within 6 hours of transfusion of plasma-containing blood products. This is a potentially fulminant complication of transfusion characterized by acute lung injury or acute respiratory distress temporally related to transfusion.
The pathophysiology of TRALI, as classically described, is caused by the presence of leukoagglutinating or human leukocyte antigen (HLA)-specific antibodies in donor plasma. When these antibodies react with recipient white blood cells, complement is activated, and C5a promotes neutrophil aggregation and sequestration in the pulmonary microvasculature, resulting in endothelial damage. Since its inception, the concept of TRALI has been expanded to include a broader spectrum of mechanisms of acute lung injury after transfusion, including anaphylactic reactions, cytokine reactions, platelet reactions, granulocyte transfusion-mediated cytotoxicity, and pulmonary lipid byproduct sequestration. This explanation relies on a single event to trigger TRALI.
In the two-event model of TRALI, the patient’s lungs may suffer one or more insults by various mechanisms that alone are necessary but not sufficient to create lung injury. In the primed host, transfusion then adds the sufficient immune challenge to induce lung injury, thereby allowing the clinician to only partly correctly lay blame on the transfused product for the pulmonary decompensation. Because of the various pathogenic mechanisms potentially involved, the true incidence of TRALI is unknown, especially since universal reporting of nonfatal cases is not required. Nonetheless, TRALI is considered a leading cause of transfusion-associated mortality. The key to a favorable outcome is early recognition and aggressive supportive care (frequently requiring intubation and mechanical ventilation); most well-resuscitated patients improve within 48 hours and ultimately survive.
Allogeneic blood transfusion infuses large quantities of soluble and cellular foreign antigen. These antigens may initiate immune down-regulation in the recipient. This syndrome is referred to as transfusion-related immunomodulation (TRIM). Leukocytes appear to be the primary blood component responsible for these deleterious immunomodulatory effects. Manifestations are varied but derive from the observation that allogeneic blood transfusion enhances the survival of renal allografts and may increase the recurrence rate of certain malignancies, particularly colorectal cancer. Likewise, allogeneic transfusion may increase the incidence of postoperative bacterial infections in patients undergoing colorectal and cardiac surgery.
Nonetheless, the putative relationship between pneumonia and TRIM was not specifically borne out in a recent review, which highlighted the fact that various studies investigating this entity provide conflicting results. To wit, some studies report that allogeneic transfusion is associated with a reduced rate of recurrent Crohn disease, which is a potential benefit, while other studies document unfavorable activation or reactivation of CMV or HIV infections.
Transfusion-associated graft-versus-host disease occurs when immunocompetent donor T lymphocytes initiate an immunologic attack against the recipient tissues. This most commonly occurs in immunocompromised patients but may also occur in recipients of directed blood donation from first-degree relatives. The syndrome is generally fatal, with onset 2-4 weeks after transfusion.
Fever, unanticipated liver function abnormalities, watery diarrhea, rash, and progressive bone marrow failure are characteristic features of this syndrome.
Postinjury multisystem organ failure
Blood transfusion in trauma is an independent predictor of multisystem organ failure (MSOF). This relationship was initially described in a 3-year single-institution prospective cohort study. One of the independent predictors of MSOF proved to be greater than 6 units PRBC transfused in the first 12 hours following admission.
In a second prospective study by the same group, a dose-response relationship between early blood transfusion and postinjury MSOF was identified, further cementing the relationship between transfusion, organ failure, and mortality.
Despite the association of transfusion and MSOF, massive transfusion is often required after injury and saves many lives. The discrepancy in these observations may reflect proteomic and genotypic differences that are incompletely understood but are being intensely studied by a "glue" grant.
Role of leukoreduction
As delineated above, blood transfusions are associated with various risks linked to nucleated cells inadvertently included in the transfused component. Leukoreduction reduces the number of transfused white blood cells via centrifugation or filtration. Leukocyte counts are reduced by greater than 99% compared to nonleukoreduced red cell transfusions. Leukoreduction is effective in reducing the transmission of cell-associated viruses (eg, CMV, Herpes viruses, EBV) and may also reduce parasite and prion transmission, nonhemolytic febrile transfusion reactions, and TRALI. However, with regard to other parameters, outcomes of leukoreduction studies provide only controversial and contradictory data. To date, no randomized controlled trial has compared leukoreduced with nonleukoreduced blood transfusion in critically ill patients. Many countries have adopted routine leukoreduction, and some authors purport that universal leukoreduction is justified. However, the process increases the cost of bloodtransfusion,and determining the patient populations in which it should be used is a source of continued debate.
Alternatives to Allogeneic Blood Transfusion
Given the risks associated with component transfusion, it stands to reason that alternative therapies to reduce allogeneic exposure have been sought.  Meticulous surgical technique, preoperative and intraoperative protocols designed to reduce red cell mass loss (eg, hypotensive procedures, surgical tourniquets, preoperative anemia correction with human erythropoietin), hemostatic adjuncts (eg, surgical, thrombin, procoagulant gels, platelet gel), and lower hemoglobin transfusion triggers are standard in the modern armamentarium to reduce banked blood use. Additional specific practices are outlined below.
Spahn and Goodnough published a summary of alternatives to blood transfusion such as autologous blood procurement, erythropoiesis-stimulating agents, and haemostatic agents that are currently under investigation.  The authors also suggest that the detection and treatment of anemia before surgery is an easy method that may help reduce the use of blood products and associated risks. 
Autologous (acute) normovolemic hemodilution
This technique involves collecting a patient’s blood (2-4 units) into anticoagulant-containing storage bags at the outset of surgery, accompanied by intravenous replacement with crystalloids or colloids to maintain normovolemia. The patient’s whole blood can then be transfused later in the case as needed. The evidence on the overall efficacy of autologous normovolemic hemodilution is conflicting with respect to blood savings and transfusion avoidance, but it appears to be safe and efficacious in centers that make routine use of this technique.
Preoperative autologous donation
In general, preoperative autologous donation limits (but does not eliminate) the infectious, immunologic, and hemolytic complications of allogeneic blood transfusion. Although accepted by patients and many surgeons, preoperative autologous donation is encumbered by unexpected disadvantages. These disadvantages include patient inconvenience and the cost of processing the blood. Because many patients require no blood, up to one third of collected units are discarded. Patients are frequently rendered anemic at the time of their operation, increasing the likelihood that transfusion will be necessary. Even though these patients would be transfused with their own blood, risk is still present. The most common cause of hemolytic transfusion reactions is clerical error, and predonated blood is not exempt. Changes in blood during storage may render patients susceptible to complications, such as TRALI, even if the transfused blood is their own.
Intraoperative autotransfusion/cell salvage
Cell salvage is an effective method of transfusion avoidance. Shed blood is collected from the operative field and mixed with an anticoagulant. It is concentrated and washed or filtered, then returned to the patient. Harmful contaminants, such as potassium, fat, and free hemoglobin, are removed from the salvaged blood; the washed blood is returned via a 40-µm blood filter. Blood obtained from the thoracic cavity via chest tubes in a closed system can be processed and autotransfused in a similar manner.
Relative contraindications to cell salvage include the following: malignancy, bacterial contamination of the surgical site, cesarean delivery, and sickle cell disease. However, improvements in processing, particularly the use of leukocyte-depletion filters, have made cell salvage possible in anecdotal cases complicated by the contraindications described above and are supported by in vitro data. Nonetheless, most practitioners still hew to the existing guidelines for the use of scavenged blood.
The use of erythropoiesis-stimulating proteins for the treatment of chemotherapy-induced anemia is well established. Anemia is also exceedingly common in critically ill patients and in those with multiple injuries.
Anemia is a result of the cumulative effects of blood loss (ie, splenic autodestruction from red cell senescence, blood drawing for laboratory sampling, operative losses) and decreased RBC production. Decreased RBC production results from abnormal iron metabolism, inappropriately low erythropoietin production, diminished response to erythropoietin, and direct suppression of RBC production.
Recombinant human erythropoietin (rHuEPO) increases hemoglobin and hematocrit values and decreases the number of RBC transfusions in ICU patients, but further studies are needed to determine whether this therapy has a positive impact on outcome and whether the benefits justify the cost of therapy. Intuitively, erythropoietin should be cost effective by reducing the risks of allogeneic transfusion even if the erythropoietin cost was equivalent to the cost of the units of red cells saved. However, potential future cost avoidance is not always accepted as justification for expensive therapy.
Patients with suboptimal initial hemoglobin levels undergoing elective surgery can also be treated with rHuEPO, either preoperatively or postoperatively, with great efficacy. Correction of nutritional deficiency may also enhance RBC production; adequate amounts of protein intake, iron, folate, vitamin B12, and vitamin C are all important in this regard.
Recombinant factor VIIa
rFVIIa was introduced above in the setting of massive transfusion, principally as related to the trauma patient. Further details with respect to its evolution and indications will be provided here.
Factor VII is key in the extrinsic pathway of the coagulation cascade. When activated, factor VII complexes with tissue factor, leading to activation of factor X, which ultimately leads to thrombin generation. At high doses, factor VIIa can directly bind to platelets, activating factor X, and converting prothrombin to thrombin by a tissue factor-independent mechanism. The ultimate result is amplified thrombin production.
rFVIIa is a thrombotic drug licensed for use in patients with hemophilia A or B with inhibitors. Approved use also includes patients with acquired inhibitors to factor VIII or IX.
As mentioned previously, numerous off-label uses have been identified in the literature. Such uses include neurosurgical operative procedures, postoperative bleeding in patients with cirrhosis, and in children with hepatic failure. Efficacy in patients with other severe hemorrhagic disorders, in surgical procedures, and in severe trauma is documented.
The cost of rFVIIa is over $3000 per patient. Adverse events occur in 1-2% of patients and include disseminated intravascular coagulation, myocardial infarction, stroke, and deep vein thrombosis. Since some of these events are known complications in these specific patient populations, the precise role of factor VIIa in their genesis remains unclear.
Pharmacologic antifibrinolytics, such as aprotinin, tranexamic acid, or epsilon aminocaproic acid (EACA), can reduce perioperative blood loss by retarding or arresting fibrinolysis. Aprotinin has been documented to decrease blood loss in certain cardiac, orthopedic, and transplant surgeries by inhibiting serine proteases, such as plasmin. Epsilon aminocaproic acid has been successfully used to reduce hemorrhage volume during cardiac surgery, portocaval shunting, aplastic anemia-related hemorrhage, abruptio placentae, cirrhosis-associated GI bleeding, and neoplasm-related hemorrhage (particularly prostate, lung, stomach, and cervix); all of these clinical entities are linked by the underlying theme of fibrinolysis-associated red cell mass loss. Epsilon aminocaproic acid primarily inhibits plasminogen activators but also demonstrates antiplasmin activity.
Aprotinin is now only available via a limited-access protocol. Fergusson et al reported an increased risk for death compared with tranexamic acid or aminocaproic acid in high-risk cardiac surgery.  For more information, see the article from Medscape.
Red blood cell substitutes
Oxygen-carrying blood substitutes are divided into 2 types, as follows: (1) fluorocarbon-based synthetic oxygen carriers, and (2) stroma-free, cross-linked, or polymerized human or nonhuman hemoglobin preparations. These compounds may be useful in acute massive blood loss, such as trauma and major operations, and have also been used in patients declining transfusion for religious reasons.
The fluorocarbon-based oxygen carriers are easily produced, have a long shelf life, and have minimal infectious or immunologic complications. However, they require a high FiO2 and are cleared rapidly.
The hemoglobin-based oxygen carriers have a high oxygen-carrying capacity, significant oncotic effect, and a long shelf life. However, they are also rapidly cleared and may induce systemic and pulmonary hypertension through vasoactivity. The ill-fated in-hospital Baxter trial of diaspirin cross-linked hemoglobin failed to demonstrate a survival benefit after injury and instead documented increased mortality. This was a first generation product that was plagued by a high percentage of tetrameric hemoglobin (>30%). Interestingly, the European HOST trial using the same product in the prehospital setting did not identify increased mortality in patients treated with DCLHb. Regardless, engineering of red cell substitutes targeted reduced tetramerism as a goal, and second generation products undergoing clinical and laboratory testing are comprised of less than 3% tetrameric hemoglobin.
Bloodless Surgery Programs
Clinical practice with patients who decline blood transfusion for religious reasons has led to the emergence of healthcare centers with bloodless medicine and surgery programs. These programs aim to minimize component transfusion therapy for all patients, regardless of religious beliefs. Integrated preoperative, intraoperative, and postoperative conservation approaches are used. Included are preoperative autologous donation, erythropoietic support, acute normovolemic hemodilution, individualized assessment of anemia tolerance, implementation of conservative transfusion thresholds, meticulous surgical techniques, judicious use of phlebotomy, and pharmacologic agents for limiting blood loss.
Success requires a multidisciplinary team consisting of surgeons, intensivists, scrub team members, pathologists, nursing staff, administrators, anesthesiologists, hematologists, pathologists, pharmacists, phlebotomists, and transfusion medicine specialists.
Progress in terms of blood use reduction and outcomes improvement should be tracked to encourage a robust and successful program.