CHAPTER 25: Blood Transfusion

In 1492, as Pope Innocent VIII lapsed into a coma, the blood of three boys was infused—into his mouth! Considerable progress has been made since then, particularly following Karl Landsteiner's discovery of the ABO blood group antigens in 1901. Blood transfusions have had a larger impact on medicine and surgery than any other therapeutic advance in hematology. In the United States, approximately 15 million units of blood are transfused each year.

This chapter begins with the fractionation of donor blood into cellular and plasma components that are used in transfusion. An overview of the important blood group antigens is followed by a consideration of methods used for typing and cross-matching units of blood. After reviewing the indications for transfusing red cells, platelets, and plasma, we discuss the hemolytic, immunologic, inflammatory, and infectious risks posed by transfusion therapy. The chapter ends with a perspective on current and future developments in transfusion medicine.

The standard blood donation involves phlebotomy through a large-bore needle inserted into an arm vein. Approximately 450 mL are transferred into a sterile plastic bag containing citrate phosphate dextrose (CPD)-adenine. Citrate (C) prevents coagulation by chelating calcium ions. The phosphate (P) buffer maintains the pH at physiologic levels. Dextrose (D) provides a source of energy during blood storage. Adenine enhances the viability of the stored red cells.

As noted in Chapter 1, when a tube of anticoagulated blood sediments in a gravitational field, the relatively dense red cells go to the bottom, and the less dense white cells and platelets form a "buffy coat" layer on the upper surface of the red cells, whereas cell-free plasma is least dense and collects at the top of the tube. In the blood bank, the bag of freshly collected donor blood is first centrifuged at relatively low speed, allowing separation into packed red cells and platelet-rich plasma (Fig. 25-1). The platelet-rich plasma is then spun at a higher speed, enabling separation into cell-free plasma and a platelet concentrate (Fig. 25-1).

Packed red cells are stored at 4°C for up to 42 days. In most medical centers, the white blood cells are removed by a filter, a maneuver that lowers the incidence of febrile reactions and human leukocyte antigen (HLA) alloimmunization and reduces the risk of infection with cytomegalovirus. Patients undergoing hematopoietic stem cell transplantation receive red cell units that have been irradiated in order to reduce the risk of graft-versus-host disease.

The fresh plasma is frozen and stored at –18°C or colder for up to 1 year. The platelet concentrate is stored at 20°C for a maximum of 5 days. Typically, to prepare a transfusable dose of platelets, concentrates from six donors are pooled. An alternative way to collect platelets in numbers sufficient for transfusion involves circulating venous blood from a donor through an apheresis machine, which continuously removes platelets and returns red cells and plasma to the donor. Platelet units are also irradiated prior to administration to transplant patients.

As explained in Chapter 10, red cell stability, pliability, and ion transport depend on specific membrane proteins. In fact, there are over 100 red cell surface membrane proteins. Many of these are polymorphic and have come to attention because they induce clinically significant immune responses when transfused into mismatched recipients. In the past, the blood antigens responsible for these immune responses have been identified by immunologic methods, but in recent years the genes that encode these proteins have been cloned and sequenced, allowing their structure and sometimes function to be established. Figure 25-2 depicts the membrane topology of a number of the well-defined red cell antigens.

A blood group system consists of carbohydrate or protein red cell antigens produced by alleles of a single genetic locus or by closely linked alleles. Most blood group antigens stem from single nucleotide polymorphisms. For example, the S and s alleles of the glycophorin B gene differ by a single nucleotide polymorphism in codon 29, with the S allele encoding a methionine residue (ATG) and the s allele encoding a threonine residue (ACG). Most red cell antigens are proteins expressed on the cell surface, but some, most notably those of the ABO system, stem from differences in carbohydrates linked to surface proteins or glycolipids.

The A and B antigens are defined by the terminal sugar that is attached to glycoproteins and glycosphingolipids by specific transferases. As shown in Figure 25-3, the precursor for these terminal sugars is the H antigen, which has the sequence R-acetylglucosamine-galactose-fucose, where R is a core carbohydrate moiety. The A antigen is formed by a glycosyltransferase that catalyzes the addition of an N-acetylgalactosamine to the subterminal galactose, whereas the B antigen is formed by an allelic variant that catalyzes the addition of another galactose residue to the core galactose. Most populations also contain a common allele lacking enzymatic activity that is designated O. Because OO homozygotes lack A and B terminal transferases, they express only the unmodified H antigen on their red cell surfaces. These individuals have the O blood type. Those who have type A or type B red cells are either homozygotes (AA or BB) or heterozygotes (AO or BO) (Table 25-1). Those with type AB red cells inherit one A allele and one B allele from their two parents.

From fetal development throughout life, individuals are exposed to A and B antigens from a variety of sources and, as a result, develop antigen-specific "natural" IgG and IgM immunoglobulins even in the absence of blood transfusion. The IgM antibodies fix complement and, as explained in more detail, later in this chapter, cause intravascular hemolysis. As shown in Table 25-1, individuals with type A red cells have anti-B antibodies in their plasma, whereas those with type B red cells have anti-A. Individuals with type O red cells have both anti-A and anti-B antibodies in their plasma. They are sometimes called universal donors and, indeed, during the Vietnam War, group O blood was sometimes administered on the battlefield to recipients, irrespective of blood type. Conversely, group AB individuals lack both anti-A and anti-B in their plasma and are sometimes called universal recipients.

The proteins that comprise the Rh system are encoded by two homologous genes on chromosome 1. Both proteins are complex transmembrane proteins that span the plasma membrane of the red cell 12 times. As depicted in Figure 25-4, one protein contains the D antigen and another the C, c, E, and e antigens. The Rh system also contains 45 other less immunogenic and less well-characterized antigens. D is extraordinarily immunogenic and therefore the most important protein antigen in transfusion practice. Rh positive and Rh negative refer to the presence or absence of D antigen on the surface of the red cell. About 15% of White North-American individuals are Rh (D) negative. They do not form alloantibodies to D unless exposed to red cells from either a fetus or transfusions. Much less commonly, alloantibodies can form against C, c, E, and e antigens that, as shown in Figure 25-4, are created by amino acid replacements in protein domains located on the external surface of the red cell. Alloantibodies against Rh antigens are IgG immunoglobulins that are opsonins but do not fix complement. Macrophages recognize and phagocytose the IgG-coated red cells, leading to extravascular hemolysis. Of note, most cases of autoimmune hemolysis due to warm IgG antibodies (Chapter 11) are caused by autoantibodies against the Rh antigens.

As discussed later in this chapter, the D antigen is also responsible for the development of hemolytic disease of the fetus and newborn, and the Rh system is a major contributor to incompatibility of transfused blood. In addition, a number of red cell antigens outside the Rh system can engender alloantibodies. The most commonly encountered alloantibodies are detected by routine screening, as described in the following section.

When a unit of donor blood is collected, it is typed for the presence of A and B antigens as well as D antigen on the red cells. The addition of serum containing a high titer of anti-A antibody will result in prompt clumping (agglutination) of type A or AB red cells, whereas type O red cells remain in a uniform suspension. In a like manner, this simple agglutination test is also used to determine whether red cells express B and D antigens.

The blood of a patient who is to receive red cell transfusions undergoes the same ABO and D typing as described earlier for donor units. In addition, the patient's serum is tested for the presence of alloantibodies other than anti-A and anti-B. Figure 25-5 shows an example of the use of red cells from two individuals carefully selected to differ not only in the Rh C, c, E, and e antigens but also in other antigen systems (Kell and Duffy). If the patient's serum fails to agglutinate cells from either of these donors, it lacks alloantibodies to these antigens. Note, however, that the patient's serum in fact agglutinates the red cells of donor 2. This result indicates that the patient has alloantibodies to E, c, or Fy^b, or possibly a combination thereof, but no alloantibodies to D, C, e, K, k, or Fy^a. The blood bank then uses additional panels of red cells to identify the patient's specific alloantibody or alloantibodies. This information allows the blood bank to select units that not only have the same ABO and D type as the patient but also lack the antigens to which the patient is immune. As a further safeguard, each selected donor unit is then tested against serum from the patient by means of the indirect antiglobulin test (indirect Coombs test, Fig. 11-3) to ensure that the unit is fully compatible.

RED CELLS

The sole purpose of transfusing red cells is to enhance the oxygen carrying capacity of the blood. The most clear-cut indication is acute blood loss, usually from surgery, trauma, or severe gastrointestinal hemorrhage. Among physical signs, postural hypotension is particularly informative, because it usually indicates hypovolemia and therefore the need for restoration of both red cell mass and plasma volume. For more information on anemia due to blood loss, see Chapter 3. The decision as to whether or when to transfuse a patient with subacute or chronic anemia is more nuanced. Many independent factors must be considered, including the etiology of the anemia, the coexistence of underlying disorders, and the patient's age, symptoms, and signs. In patients with the anemia of renal failure, erythropoietin therapy is nearly always sufficient to restore the red cell mass. Transfusions also may be avoided if the anemic patient is known to have an uncomplicated deficiency of iron, cobalamin, or folate, because such patients respond rapidly to appropriate replacement therapy. Patients with long-standing anemia, particularly younger and sedentary individuals, are able to tolerate severe anemia remarkably well, owing to compensatory mechanisms discussed in Chapter 3. In contrast, if the anemia is of recent onset, it is often less well-tolerated, particularly if the patient is physically active. In patients with symptoms of angina pectoris or episodes of transient cerebral ischemia, the threshold for transfusion is relatively low because of concern that continued hypoxia may lead to damage of a vital organ. For all these reasons, it is difficult and indeed inadvisable to assign a target hemoglobin or hematocrit level below which a patient should be transfused. During the last decade evidence has accumulated indicating that red cell transfusions are used too liberally among both medical and surgical patients. In fact, some studies have shown that the more restrictive use of blood transfusions in patients with moderate anemia significantly lowers the mortality rate.

PLATELETS

As discussed earlier, a platelet unit is prepared either as a pool from six individual blood units or from a single donor by apheresis. As mentioned in Chapter 14, the normal life span of platelets in the circulation is 7 to 9 days. However, platelets have a significantly shorter life span when transfused to thrombocytopenic patients. Moreover, after receiving platelets, some recipients develop an immune response, most commonly to mismatched major histocompatibility antigens (HLA antigens), which further accelerates the clearance of donor platelets. For these reasons, platelet transfusions should be administered sparingly. Platelet transfusions are much more effective in thrombocytopenias due to underproduction than in those caused by enhanced platelet destruction.

The primary indication for platelet transfusion is bleeding owing either to thrombocytopenia or much less often to defective platelet function. In addition, platelets are commonly transfused to patients who are not bleeding but whose platelet counts are so low that there is high risk of significant hemorrhage. As with red cell transfusions, in recent years the indications for prophylactic platelet transfusion have become more restrictive. As shown in Figure 25-6, patients with normally functioning platelets do not develop significant gastrointestinal blood loss unless the platelet count falls below 5000/mm³. A prospective study has shown that there is no significant difference in the risk of major hemorrhage in a group of patients transfused with platelet counts of <10,000/mm³ compared with a group with a transfusion threshold of <20,000/mm³.

PLASMA

The primary indication for transfusion of fresh frozen plasma is correction of deficiencies of multiple clotting factors, as may occur in liver disease, overdosing of warfarin, and disseminated intravascular coagulation (all described in detail in Chapter 16). In contrast, patients who have inherited single clotting factor deficiencies are treated with specific replacement therapy (described in Chapter 15). Patients with acute massive hemorrhage benefit from infusion of fresh frozen plasma for two reasons. Because they are hypovolemic, they respond favorably to the administration of a colloid load that is retained within the intravascular space. Moreover, many of these patients have low levels of clotting factors either because of disseminated intravascular coagulation or dilution from loss of plasma and high-volume fluid replacement. Blood banks maintain a supply of cryoprecipitate, which is prepared by collecting the protein precipitate that forms when plasma is cooled to 4°C. Cryoprecipitate contains factor VIII, von Willebrand factor, factor XIII, fibronectin, and fibrinogen, all of which except fibrinogen can replaced by administration of either recombinant protein products or highly purified concentrates. Thus, cryoprecipitate is primarily used for treatment of hypofibrinogenemia.

HEMOLYTIC REACTIONS

As summarized in Table 25-2, hemolysis due to the administration of antigenically incompatible red cells can be either immediate or delayed. Immediate hemolytic reactions are nearly always due to ABO incompatibility. For example, if an individual with type B blood receives a unit of type A blood by mistake, the recipient's preexisting IgM anti-A antibodies will cause complement fixation and intravascular lysis of the transfused red cells. The patient's clinical presentation depends on how much blood was infused before the error was suspected and the transfusion stopped. Patients generally develop fever and chills, often accompanied by dyspnea, tachycardia, hemoglobinuria, and severe pain, usually in the lower back. Transfusion of a large volume of ABO-incompatible blood often produces hypotensive shock, oliguria, and disseminated intravascular coagulation. Note that all of these findings except for hypotension are likely to be missed if the patient is in the operating room under general anesthesia. At the first signal that incompatible blood may have been administered, the infusion must be stopped and serologic testing repeated on the donor and recipient blood. During this time it is critical to maintain intravenous access and careful monitoring of blood pressure and urine output. Fortunately, contemporary blood banking has a series of safeguards in place that make this complication very rare. In New York State, 9,000,000 units of blood were transfused over the decade 1990 through 1999. A total of 462 patients (about 1 per 20,000 units given) suffered adverse consequences from administration of ABO-incompatible blood.

Delayed hemolytic transfusion reactions are more commonly encountered and seldom caused by error. As mentioned earlier, patients develop alloantibodies to protein antigens only after pregnancy or transfusions. The Rh D antigen is highly immunogenic and will trigger detectable levels of anti-D after a single exposure. In contrast, antigens such as Rh C, c, E, and e as well as Kell and Duffy (Fig. 25-5) and others, shown in Figure 25-2, are less immunogenic, but in some cases nevertheless trigger a primary immune response in an antigen-negative recipient. This primary response may not cause the development of a sufficiently high antibody titer to be detectable by the screening procedure described in Figure 25-5. However, if at a later point in time the recipient is transfused with a second unit containing this antigen, within 3 to 10 days an anamnestic immunologic response will raise the titer of antibody to sufficient levels to cause clinically significant hemolysis. As mentioned at the beginning of this chapter, anti-Rh antibody is an IgG and does not fix complement. Accordingly, the hemolysis is extravascular and not nearly as severe as that encountered in immediate hemolytic reactions due to ABO incompatibility. Figure 25-7 depicts a typical delayed transfusion reaction. This complication should be suspected in any patient who develops an unexplained drop in hemoglobin or hematocrit within a week of receiving transfused red cells. Often the patient has no symptoms or signs but occasionally may develop fever, chills, and jaundice, and rarely hemoglobinuria. The diagnosis is established either by documenting the presence of antibody-coated red cells in the patient's blood by the direct antiglobulin test (Chapter 11, Fig. 11-3A) or by the indirect antiglobulin test (Chapter 11, Fig. 11-3B), in which the patient's serum is tested against a panel of red cells with characterized antigens as well as against the donor red cells that were recently infused. Often the direct antiglobulin test is negative because the antibody-coated red cells have been rapidly cleared by resident macrophages. Delayed hemolytic transfusion reactions usually do not require therapy. However, the patient's hemoglobin and hematocrit levels should be closely monitored, and care must be given to ensure that future red cell transfusions are antigen negative.

Hemolytic disease of the fetus and newborn (HDFN) involves an immunologic response following the "transfusion" of red cells from the fetus into the maternal circulation. By far the most common cause of HDFN is maternal Rh D alloimmunization. The pathogenesis is similar to that of a delayed hemolytic transfusion reaction; a typical scenario is as follows. An Rh-negative mother is carrying a fetus that has inherited the D antigen from its father. During the first pregnancy, leakage of fetal red cells across the placenta results in the mother becoming immunized against the D antigen. However the amount of anti-D antibody that crosses the placenta into the fetal circulation is too low to cause hemolysis. With subsequent pregnancies, the mother mounts an anamnestic immunologic response to D+ fetal red cells, putting the fetus and newborn at high risk of developing alloimmune hemolysis. The baby develops severe anemia and generalized edema (hydrops) owing to tissue hypoxia. In addition, the high levels of nonconjugated bilirubin in fetal and neonatal plasma can cross the blood-brain barrier and accumulate in the brain, particularly in the basal ganglia, causing irreversible neurologic damage (kernicterus). If signs of fetal distress during the third trimester trigger a hematologic investigation leading to the diagnosis of HDFN, intrauterine transfusion will prevent the development of hydrops and kernicterus. Fortunately, HDFN is rarely encountered today thanks to the discovery 50 years ago that the administration of anti-D immunoglobulin (RhoGAM, Ortho Clinical Diagnostics, Rochester, NY) to Rh-negative mothers after their first delivery safely and effectively suppresses sensitization to Rh D and prevents the development of HDFN in subsequent pregnancies.

TRANSFUSION-RELATED LUNG INJURY

Approximately one in 3000 patients develops acute interstitial pneumonitis within 6 hours of transfusion, accompanied by dyspnea, tachycardia, hypoxemia, fever, and hypotension. Figure 25-8 shows the rapid development of bilateral pulmonary infiltrates in a patient whose lungs were normal prior to transfusion. This acute process is often difficult to distinguish from pulmonary edema, pulmonary embolus, bacterial pneumonia, pulmonary hemorrhage, or acute respiratory distress syndrome. It is likely that transfusion-related acute lung injury (TRALI) is a form of acute respiratory distress syndrome caused by donor antibodies that bind to and activate either recipient neutrophils or pulmonary endothelial cells. In many cases, the causative antibodies appear to be against HLA antigens. Other cases are associated with pathogenic antibodies against a polymorphic antigen expressed on neutrophils, called choline transporter-like protein 2 (CLT2); these antibodies are particularly prone to cause severe and often fatal TRALI. TRALI is the most commonly encountered serious complication of transfusion therapy. The treatment of TRALI is primarily supportive. All patients require oxygen and most need ventilator assistance. These episodes usually resolve over 48 to 96 hours, but the mortality is 5% to 25%. The risk of TRALI can be reduced by the use of plasma from male donors, because women may become HLA-alloimmunized during pregnancy. Cases of suspected TRALI are investigated by the blood bank. Donors of blood products believed to have caused TRALI are permanently banned from donating blood.

TRANSMISSION OF INFECTIOUS PATHOGENS

A variety of bacterial, viral, and parasitic infections can be transmitted by transfusion of blood or blood products. Current transfusion practice mandates routine screening for the pathogens listed in Table 25-3. Bacterial contamination is particularly problematic in platelet concentrates that are stored at room temperature. Occasional cases of malaria in the United States have arisen through transfusion of blood from a donor who has recently returned from a tropical area. The viral pathogens of most concern are human immunodeficiency virus (HIV), hepatitis C, and hepatitis B. However, thanks to highly sensitive and specific immunologic and genetic methods for detection, the risk of transmission of these viruses via blood products has fallen several orders of magnitude over the last 25 years. As Figure 25-9 shows, there is only a short (10 day) window of time between exposure to HIV infection and its detection by testing for HIV RNA in the blood. Despite these advances, the lay public continues to harbor fears that HIV and hepatitis are the most likely serious complications of transfusion therapy. In fact, as Table 25-4 shows, TRALI and administration of incompatible blood pose much higher risks than viral pathogens.