Chapter 55: Alloimmune Hemolytic Disease of the Fetus and Newborn

Alloimmune hemolytic disease of the fetus and newborn (HDFN) is a disorder in which the life span of fetal and/or neonatal red cells is shortened as a result of binding of transplacentally transferred maternal immunoglobulin (Ig) G antibodies on fetal red blood cell (RBC) antigens foreign to the mother, inherited by the fetus from the father. The resulting hemolysis or suppression of erythropoiesis may cause fetal and/or neonatal anemia and significant neonatal jaundice. There are three main classes of alloimmune HDFN, based on the antigen(s) involved: Rh (rhesus), minor red cell antigens (i.e., Kell, Duffy, Kidd antigens) and ABO.

Prior to the development of medical interventions in the 1950s, almost half of all newborn infants with Rh HDFN died or were severely handicapped. Although the clinical condition was described in newborn infants as early as the 1600s, it was only in the 1930s and 1940s that the pathophysiology of Rh HDFN was uncovered. In 1932, Diamond and colleagues¹ recognized that the clinical syndromes of stillbirth with unusual erythroblastic activity in the extramedullary sites and blood, fetal hydrops, anemia in the newborn, and “icterus gravis neonatorum” were closely related and likely had the same pathophysiology in the hematopoietic system. In 1938, Ruth Darrow, a pathologist who lost a baby to kernicterus, postulated that hemolysis of fetal RBCs was a result of maternal antibody produced in response to fetal hemoglobin.² The discovery of the Rh factor by Landsteiner and Weiner led to elucidation of Rh HDFN by Levine and colleagues who established that erythroblastosis fetalis was caused by immunization of an Rh-negative mother by the red cells from an Rh-positive fetus.³ Antibodies produced by the sensitized mother crossed the placenta in the next pregnancy and coated the fetal Rh-positive cells, leading to hemolysis, anemia, hydrops, and severe neonatal jaundice.

Neonatal mortality from Rh HDFN decreased considerably with the development of exchange transfusion techniques for correction of severe anemia and hyperbilirubinemia.⁴ However, severely affected fetuses continued to die in utero before 34 weeks’ gestation. In 1961, Liley demonstrated the prognostic value of amniotic fluid spectrophotometry in identifying fetuses at risk and then showed that intrauterine transfusions (IUTs) could prevent fetal deaths.⁵ The most dramatic reduction in the incidence of Rh HDFN was achieved in the 1960s and 1970s with the development of postpartum and antepartum anti-D prophylaxis to prevent maternal Rh sensitization.⁶

Despite these advances, Rh HDFN has not disappeared, and cases of hemolytic disease of the newborn resulting from red cell antibodies directed toward antigens other than the Rh blood group system are being increasingly recognized.^7,8,9,10,11 Furthermore, maternal Rh isoimmunization and Rh hemolytic disease still occur, particularly in developing countries where anti-RhD prophylaxis is not widely available or in infants born outside of medical facilities.¹²

The epidemiology of HDFN varies in different ethnic and racial groups; the frequency of specific blood group alleles in a given population determines the probability of blood group incompatibility and maternal alloimmunization. Antigen-negative women may have naturally occurring antibodies to certain red cell antigens (anti-A or anti-B) or may develop antibodies as a result of exposure to foreign red cell antigens through blood transfusion or by silent fetomaternal hemorrhage during pregnancy or at delivery. More than 50 different RBC antigens are associated with maternal alloimmunization^{7,8,9,10,11,13} and with HDFN of varying severity; however, the vast majority of clinically significant maternal alloantibodies are within the Rh (D, CE), Kell, Duffy, MNS, and Kidd systems (Table 55–1).

Antenatal screening programs detect antibodies to clinically significant Rh or other minor RBC antigens in 0.01 to 0.4 percent of pregnant women,^{7,8,9,10,11,13} although these numbers vary by country. Approximately 15 percent of Americans of European descent are RhD-negative, compared to 7 percent of Americans of African descent and Hispanics, 5 percent of Asian Indians, and 0.3 percent of Chinese people.^14,15,16 Despite the success of Rh prophylaxis, anti-D antibodies still constitute a large proportion of clinically significant antibodies detected in Europe and the United States. When RhD is excluded, non-D Rh antibodies (c, C, e, E, etc.) and antibodies belonging to the Kell, Duffy, Kidd, and MNS systems, are most frequently involved^7,13; Table 55–2 shows representative estimate of antibodies other than anti-D in women referred to a major national Maternal Alloimmunization Program at Wexner Medical Center at the Ohio State University.

OVERVIEW

Anemia, jaundice, and hepatosplenomegaly are the hallmarks of hemolytic disease of the newborn (HDN). The clinical spectrum of affected infants is highly variable. In Rh HDN, half of the infants have mild disease and do not require intervention. One-fourth of affected infants are born at term with moderate anemia and develop severe jaundice. In the days prior to intrauterine intervention, hydrops developed in utero in the remaining one-fourth of infants; half became hydropic prior to 34 weeks’ gestation. Hydrops recurs in 90 percent of affected pregnancies, often at an earlier gestation. In Kell HDN, the clinical spectrum of hemolytic disease is less predictable, ranging from mild anemia to frank hydrops; jaundice may be less severe than that seen in Rh HDFN, given the erythroid suppression that anti-Kell alloantibodies may induce. Jaundice is the predominant feature of ABO HDN, but anemia and mild hepatosplenomegaly may also be seen. Severe fetal anemia and hydrops are unusual in ABO hemolytic disease.¹⁷

HEMOLYTIC ANEMIA

Infants with mild HDN may have cord blood hemoglobin concentrations only slightly lower than the age-related normal range. Hemoglobin values usually continue to fall after birth in all affected infants. Hemolysis continues until all incompatible red cells and/or circulating maternal alloantibody are eliminated from the circulation. Physical examination in infants having moderate to severe anemia reveals pallor, tachypnea, and tachycardia. In cases of severe HDFN, fetal anemia secondary to hemolysis results in compensatory extramedullary hematopoiesis in the liver, spleen, kidneys, and adrenal glands, and an outpouring of immature nucleated RBCs in the fetal circulation due to increased fetal plasma erythropoietin levels.¹⁸ The marked increase in erythropoiesis may be accompanied by down-modulation of platelet and neutrophil production.¹⁹

After birth, the quantity of maternal antibodies in the neonatal circulation decreases over the next 12 weeks, with a half-life of approximately 25 days. Infants with moderate to severe hemolytic disease may develop significant anemia beyond the immediate neonatal period lasting up to 8 to 12 weeks of life. Delayed anemia is related to continuing hemolysis because of persistence of maternal antibodies and a hyporegenerative component with decreased red cell production from low serum concentrations of erythropoietin.^20,21,22

NEONATAL JAUNDICE

Most infants with HDN are not jaundiced at birth because the placenta effectively transports most of the lipid-soluble unconjugated fetal bilirubin. Bilirubin concentrations in amniotic fluid reflect bilirubin concentrations in fetal blood and are influenced by fetal blood and amniotic fluid albumin concentrations.²³ The mechanism of entry of bilirubin into the amniotic fluid compartment has been debated, but of the five possible pathways (excretion through the fetal kidneys, meconium, skin, fetal lung secretions, and transmembranous), transmembranous appears to be most likely.²⁴

At birth, the newborn infant’s immature liver is incapable of handling the large bilirubin load that results from the ongoing destruction of antibody-coated neonatal red cells, and jaundice usually develops during the first day of life, often in the first few hours of life in severely affected infants. The jaundice progresses in a cephalopedal direction with rising bilirubin levels. In patients with mild disease, the serum indirect bilirubin peaks by the fourth or fifth day and then declines slowly. Premature infants may have higher levels of serum bilirubin for a longer duration because of lower hepatic glucuronyl transferase activity. Conjugated hyperbilirubinemia at birth is sometimes noted in infants who received multiple IUTs. Babies who received IUTs may still have anemia at birth and may still develop significant hyperbilirubinemia in the neonatal period.²⁵ As discussed later in the section “Postnatal Management”, these infants may require intermittent transfusion therapy until 2 to 3 months of age due to persistent anemia.

KERNICTERUS

An important complication of elevated serum levels of indirect bilirubin in the neonate is the development of bilirubin encephalopathy.²⁶ This disorder, also termed kernicterus, is caused by bilirubin pigment deposition in the basal ganglia and brainstem nuclei, leading to neuronal necrosis. Acute bilirubin encephalopathy is initially marked by lethargy, poor feeding, and hypotonia. With increasing severity, the infant develops a high-pitched cry, fever, hypertonia progressing to frank opisthotonos, and irregular respiration. The infants then develop any or all of the classic sequelae of kernicterus: choreoathetoid cerebral palsy, gaze abnormalities, especially in upward gaze, sensorineural hearing loss, and cognitive deficits. The clinical presentation of bilirubin encephalopathy in preterm infants may be less distinctive. Abnormal or absent brainstem auditory evoked potentials and magnetic resonance imaging scans demonstrating the characteristic bilateral lesions of the globus pallidus help confirm the clinical diagnosis of kernicterus.

Infants with HDFN are at higher risk for kernicterus than are other infants with the same bilirubin level from other causes.²⁷ Heme pigments produced during active hemolysis are hypothesized to inhibit bilirubin–albumin binding. Alternatively, many conditions that potentially compromise the blood–brain barrier, such as prematurity, acidosis, hypoxemia, hypothermia, and hypoglycemia, are present in severely affected infants, making them more vulnerable to bilirubin encephalopathy.

OTHER CLINICAL FEATURES

Extensive extramedullary hemopoiesis in the liver and spleen may cause portal and umbilical venous hypertension, leading to ascites, pleural effusions, and consequent pulmonary hypoplasia.²⁸ Trophoblastic hypertrophy and placental edema cause impaired placental function. Hypoproteinemia as a result of liver dysfunction leads to generalized edema. “Hydrops fetalis,” a state of anasarca, is the end result of a combination of anemia, hypoproteinemia, cardiac failure, elevated venous pressures, increased capillary permeability, and impaired lymphatic clearance. Hepatosplenomegaly is usually present. Cholestatic liver disease, previously believed to be associated with iron overload caused by IUTs, has been shown to occur in 13 percent of neonates with HDFN independent of previous IUT treatment and type of alloimmunization.²⁹ Hydropic babies may also have respiratory distress as a result of pulmonary hypoplasia, pleural and/or pericardial effusions, or surfactant deficiency.

Although the pathophysiology is not entirely clear, purpura associated with thrombocytopenia is sometimes seen in infants with HDFN. Thrombocytopenia at birth has been occurs in 26 percent of neonates with HDFN, independent of other comorbidities, such as lower gestational age at birth, being small for gestational age (SGA), and previous IUTs.³⁰ Severe fetal thrombocytopenia (platelet count <50 × 10⁹/L) has been reported in 3 percent of all fetal blood samplings and in 23 percent of severely RhD alloimmunized hydropic fetuses in one center when fetal platelet counts were measured prior to IUT.³¹ In addition, neutropenia is a common feature of HDN and may be prolonged up to a year of life, regardless of severity of HDFN, treatment received, or antibody specificity.³²

OBSTETRIC HISTORY

The course and outcome of prior pregnancies are critically important in the initial evaluation of an alloimmunized pregnancy. A history of early fetal deaths or hydrops is ominous. In Rh alloimmunization, the severity of HDFN typically remains the same or worsens in subsequent affected pregnancies. Hydrops recurs in 90 percent of affected pregnancies, often at an earlier gestation in subsequent pregnancies. Alloimmunized women who report previous neonatal deaths, neonatal exchange transfusions, or IUTs should receive very close fetal surveillance.³³ Jaundice as a result of hemolysis often recurs to the same degree of severity in subsequent affected siblings. The history of prior blood transfusions may be obtained in women sensitized to antigens other than RhD, especially if Kell alloimmunization is detected. Establishment of paternity for each pregnancy is particularly relevant in both Rh and Kell alloimmunization, because the fetus is at risk only if the father is positive for the antigen in question. Unlike Rh and other minor alloantibody HDFN, ABO HDN may affect the first-born ABO-incompatible infant. Although rare, severe ABO HDN may recur in subsequent ABO-incompatible pregnancies.³⁴

DIFFERENTIAL DIAGNOSIS

Hydrops fetalis may be secondary to α-thalassemia (Chap. 49) cardiac anomalies or arrhythmias, fetal genetic or metabolic disorders, intrauterine infections such as syphilis or toxoplasmosis, or any of a multitude of causes that lead to severe derangements in fetal homeostasis. These disorders are classified as nonimmune hydrops and are differentiated from the etiologies discussed in this chapter by the absence of any clinically significant red cell alloantibodies in the mother’s blood. Parvovirus B19 infection of the mother at any time during gestation can cause nonimmune hydrops, profound fetal anemia, and death.

Neonatal anemia caused by intrinsic red cell defects such as hereditary spherocytosis (Chap. 46), red cell enzyme deficiencies (Chap. 47), and hemoglobinopathies, notably α-thalassemia (Chap. 48) can give a similar clinical picture to HDN. The absence of maternal red cell alloantibodies, a negative direct antiglobulin test (DAT) result, and detection of the specific defect determining the disorder clarify the diagnosis. Disorders of bilirubin metabolism can lead to unconjugated hyperbilirubinemia; however, they are not associated with anemia. Hepatitis or obstructive biliary diseases present with direct hyperbilirubinemia, most often after the first week of life.

There are three main classes of alloimmune HDFN, based on the antigen(s) involved: (1) Rh, (2) minor red cell antigens (i.e., Kell, Duffy, Kidd antigens), and (3) ABO. Factors affecting the risk to the fetus or neonate not only include the class of antigens involved, but also the titer (i.e., 4 vs. 32) and class/subclass (i.e., IgM vs. IgG₄ vs. IgG₁) of the antibody, the level of antigen expression on fetal RBCs, and the antibody’s ability to suppress erythropoiesis (i.e., anti-K).^35,36 Rh HDFN is discussed first because it is archetypal of this condition; however, ABO incompatibility is much more common than Rh HDFN. The distinguishing features of the three classes of HDFN are highlighted below.

RhD HEMOLYTIC DISEASE

The standard nomenclature for designation of one’s blood type is ABO and Rh-positive or Rh-negative based on the presence or absence of the RhD antigen. However, the Rh blood system consists of numerous other antigens, the most common clinically significant include C, c, E, and e, which are encoded on the paired RHCE gene. RhD-positive individuals may have one or two copies of RHD gene (heterozygous or homozygous RhD-positive, respectively). More than 150 alleles have been defined for RHD and more are likely to be revealed as population studies expand.³⁷ In whites, RhD-negative individuals are homozygous for deletion of the RHD gene, which encompasses the whole of RHD and part of each of the flanking Rh boxes. The resultant RhD-negative phenotype is characterized by the absence of the whole RhD protein from the red cell membrane; however the RHCE-encoded antigens are present. Only 18 percent of RhD-negative Americans of African Descent are homozygous for RHD deletion. The majority (66 percent) of RhD-negative Americans of African descent have an inactive RHD gene (RHψ), while 15 percent have a hybrid RHD-CE-D gene, neither of which produces epitopes of RhD antigen.

A number of RHD alleles responsible for RhD protein variants with altered RhD antigen expression have been classified according to their phenotype and molecular variation as partial D, weak D, and DEL. Amino acid substitutions located in extracellular domains result in different forms of partial D phenotype. Women with partial D phenotypes may develop antibodies against D epitopes absent on their cell membranes but present on the fetal RBCs. Amino acid substitutions located in the transmembrane or intracellular segments of the RhD protein result in a weak D phenotype. The expressed RhD antigen is reduced quantitatively but not qualitatively, so carriers are usually not susceptible to anti-D immunization. However, some types of weak D (weak D type 15, weak D type 4.2 or DAR, and weak D type 7) may produce anti-D. DEL is a very weakly expressed D antigen found in 30 percent of RhD-“negative” blood donors in East Asia. Anti-D antibody molecules recognize the epitopes on the external loops of RhD protein (Chap. 136).^37,38,39

There currently are no requirements in the United States for blood banks to test for variant RhD on RhD-negative pregnant mothers. However, monoclonal RhD typing sera most commonly used for RhD typing do not detect weak D or partial D phenotypes, which results in reporting patients with variant RhD phenotypes as RhD-negative. Therefore, pregnant women with variant RhD phenotypes are usually treated with anti-D immunoglobulin prophylaxis.

RhD-negative pregnant women who appear to have both anti-D and anti-C, especially if at a similar strengths of reactivity serologically, require special consideration. In this instance, if the pregnant woman is lacking both RhD and RhC antigens, she is also lacking the RhG antigen, a combination antigen in the Rh Blood Group System found on RBCs containing either RhD or RhC antigens. In these situations, the laboratory must determine whether the antibodies are truly anti-D/anti-C rather than an anti-G because a patient who develops anti-C and/or anti-G antibodies but no anti-D should receive anti-D immunoglobulin prophylaxis.⁴⁰

Fetomaternal Hemorrhage and Anti-D Alloimmunization

Asymptomatic transplacental passage of fetal red cells occurs in 75 percent of pregnant women at some time during pregnancy or during labor and delivery.⁴¹ The incidence of fetomaternal transfusion increases with advancing gestation—from 3 percent (first trimester) to 12 percent (second trimester) to 45 percent (third trimester) and to 64 percent at delivery. The average volume of fetal blood in the maternal circulation after delivery is approximately 0.1 mL in most women and less than 1 mL in 96 percent of women.⁴² Intrapartum fetomaternal hemorrhage of more than 30 mL may occur in up to 1 percent of deliveries.⁴³ Massive fetomaternal hemorrhage may present with decreased fetal movement and sinusoidal heart rhythm (undulating wave form alternating with a flat or smooth baseline fetal heart rate); however, it also may be clinically silent, with no clinical signs differentiating such deliveries from those with minimal fetomaternal hemorrhage.^43,44 Fetomaternal transfusion can also result from obstetric procedures such as chorionic villus sampling, amniocentesis, funipuncture, therapeutic abortion, external cephalic version, cesarean section, and manual removal of the placenta, and from pathologic conditions such as abdominal trauma, spontaneous abortion, or ectopic pregnancy.^42,45,46,47

The presence of RhD-positive red cells in an RhD-negative mother initially provokes a weak and slow primary immune response, which develops over 4 weeks and consists of transient elevation of IgM antibodies. Subsequently, approximately 5 to 15 weeks after exposure to the RhD-positive red cells, anti-D IgG antibodies capable of crossing the placenta are produced. The RhD antigen is the most immunogenic of the Rh antigens and, indeed, of all red cell antigens (after ABO).⁴⁸ The RhD protein is processed by antigen-processing cells in the spleen and lymphoid tissue of D-negative individuals into multiple short allogeneic linear peptides, which stimulate helper T cells, which then activate B cells to produce IgM and later IgG antibodies. Memory T and B cells that are generated following the initial immune response are long lived, and exposure to the antigen, even years later, results in an accelerated antibody response as a result of rapid proliferation of antigen-specific clones. Repeated exposure to RhD-positive fetal RBCs, as in a second RhD-positive pregnancy in a sensitized RhD-negative woman, produces a brisk secondary immune response marked by rapid production of large amounts of anti-D IgG antibody by maternal memory B lymphocytes.

In the absence of Rho(D) immunoglobulin (RhIg) prophylaxis, sensitization occurs in 7 to 16 percent of RhD-negative women at risk, within 6 months after delivery of the first RhD-positive ABO-compatible fetus. The relatively low rate of 16 percent of primary alloimmunization in Rh-negative women at risk may be a result of the low volume of fetomaternal hemorrhage (FMH) in most women. In fact, as little as 0.03 mL of RhD-positive RBCs has been shown to be sufficient to immunize some RhD-negative individuals.⁴⁹

The potential for immunization of the mother is not only determined by the extent of FMH and the presence of fetomaternal blood group incompatibility, but also by other factors, such as the frequency of fetomaternal transfusion and whether the mother and fetus are ABO compatible. As an example, repetitive exposure to minuscule amounts of RhD-positive red cells in RhD-negative women who abuse intravenous drugs and share needles with RhD-positive partners has been reported to lead to severe Rh sensitization.⁵⁰ Fetomaternal ABO incompatibility offers some protection against primary Rh immunization because incompatible fetal red cells are destroyed rapidly by maternal anti-A and anti-B antibodies, reducing maternal exposure to RhD antigenic sites. Primary Rh immunization occurs in 2 to 4 percent of women at risk after delivery of an ABO-incompatible fetus.⁵¹ ABO incompatibility confers no protection against the secondary immune response once sensitization has occurred.⁵²

Binding of transplacentally transferred maternal anti-D IgG antibodies to D-antigen sites on the fetal red cell membrane is followed by adherence of the coated red cells to the Fcγ receptors of macrophages with rosette formation, leading to extravascular noncomplement-mediated phagocytosis and lysis, predominantly in the spleen.⁵¹ Although Rh antigens are found on fetal RBCs as early as week 6 of gestation,³⁹ active transport of IgG across the placenta is slow until 20 to 24 weeks of gestation. The severity of fetal anemia is influenced primarily by the anti-D IgG concentration, but is also by other factors including: the IgG subclass, the rate of transplacental transfer of maternal IgG, the functional maturity of the fetal mononuclear phagocyte system, and the presence of maternal human leukocyte antigen (HLA) antibodies and or maternal–fetal ABO incompatibility.⁵³ Although IgG anti-D consists mainly of IgG₁ and IgG₃ subclasses, the relative contribution of each of these subclasses to the severity of HDFN remains controversial.^51,54,55

HEMOLYTIC DISEASE CAUSED BY OTHER RED CELL ANTIBODIES

Although antibodies against RhD tend to be the most clinically significant in terms of fetal outcomes, there are many other RBC antigens capable of inducing alloantibodies after exposure through transfusion or pregnancy. Any alloantibody capable of inducing hemolysis or suppressing erythropoiesis may be clinically significant to developing fetuses. However, the mere presence of antibodies on screening tests may not be clinically significant, because of the unique characteristics of some antibodies. For example, antibodies to Lewis antigens (Le^a, Le^b) are IgM and do not cross the placenta. Alternatively, Lutheran (Lu^a, Lu^b) and Chido antigens are poorly expressed on fetal and neonatal red cells and therefore are not susceptible to destruction by maternal antibodies.

Case reports may be biased toward more-severe cases, and there is considerable variability in the clinical spectrum of disease produced by different alloantibodies. With the widespread use of RhIg, anti-Kell has bypassed anti-D as the leading cause of HDFN in some centers.⁸ Table 55–1 lists some of the antibodies more commonly associated with HDFN.^13,56 Of note, many cases severe enough to require IUT involve antibodies to RhD, with or without antibodies to other alloantigens being present. Figure 55–1 shows data from 178 pregnancies requiring IUT at Wexner Medical Center, Ohio State University.

Kell

The Kell blood group system consists of at least 28 discrete antigens, of which eight are associated with HDFN. The KEL gene is located on chromosome 7q34, and Kell antigens are located on the red cell membrane glycoprotein CD238. Kell is unique in that it spans the RBC membrane once; it has a short N-terminal domain of 47 amino acids in the cytosol and a large C-terminal domain (665 amino acids) outside the membrane. The most common of the K antigens, Kell (also known as KEL1), is expressed by erythroid progenitor cells and mature erythroid cells by 9 percent of people of European ancestry and 1 to 2 percent of people of African ancestry; most KEL1-positive individuals are heterozygous.⁵⁷ Alloimmunization to KEL1 can occur through transfusion or through pregnancy.^58,59

Given the relatively low prevalence of the KEL1 antigen, the likelihood of KEL1 incompatibility between mother and child is less than that of RhD. However, any KEL1-positive fetus being carried by a woman alloimmunized to the KEL1 antigen is at risk of anemia. Between 2.5 and 10 percent of alloimmunized pregnancies result in an affected infant, with fetal hydrops and severe anemia being common presentations of Kell HDFN.^59,60,61 In contrast to the hemolysis observed in RhD HDFN, the severe fetal anemia observed because of maternal anti-Kell alloantibodies is largely a result of suppression of erythropoiesis and not to hemolysis. Clinical observations of inappropriately low levels of circulating reticulocytes and normoblasts for the degree of anemia have long been noted in affected fetuses, and suppression of erythropoiesis has been established by in vitro studies showing that growth of Kell-positive erythroid progenitor cells is inhibited by monoclonal IgG and IgM anti-Kell antibodies.⁶² Furthermore, anti-Kell antibodies are associated with suppression of megakaryocyte and granulocyte colony-forming units with resultant fetal and neonatal thrombocytopenia and pancytopenia.^63,64 As discussed in further detail later in this the section” MATERNAL IMMUNOHEMATOLOGIC TESTING”, titers of maternal anti-Kell do not necessarily correlate with the severity of fetal anemia and thus all maternal anti-Kell alloantibodies must be considered to be potentially clinically significant to antigen positive fetuses.

Other Minor Antigens (Non-D, Non-Kell)

Many other minor RBC antigens besides D and Kell are immunogenic in transfusion and pregnancy settings, with some of these alloantibodies having the capacity of being detrimental to developing fetuses. The incidence and prevalence of other minor antigens contributing to HDFN depends in part on the geographical area evaluated, as genetics and local transfusion practices impact alloimmunization. Any IgG that can cross the placenta is, in theory, capable of binding to cognate antigen on fetal RBCs. Antigen copy number, as well as other characteristics of the antigens/antibodies (potentially including IgG subtype), may impact the clinical significance of the alloantibodies.

ABO HEMOLYTIC DISEASE

ABO HDN occurs almost exclusively in infants with blood group A or B who are born to group O mothers, given that group O mothers have naturally occurring antibodies against the A and B antigens that are of the IgG class and are thus capable of crossing the placenta. The fetal reticuloendothelial system may completely remove RBCs bound with IgG, or it may remove portions of the RBCs, resulting in microspherocytes visible on blood film. (Table 55–3 compares ABO and RhD HDN.) Infants with ABO HDN generally have less-severe disease than those with Rh incompatibility. Hydrops fetalis caused by ABO alloimmune HDN is extremely rare. Although infants affected by ABO incompatibility may require phototherapy to treat their jaundice, less than 0.1 percent require exchange transfusion.^17,65 Unlike Rh disease, ABO HDN may affect the firstborn ABO-incompatible infant because anti-A and anti-B IgG antibodies are normally present in group O adults. Antenatal testing of anti-A and anti-B levels in group O mothers has little value in predicting ABO HDN, though the number of fully developed A or B antigens on fetal RBCs may impact disease severity. A recurrence rate of 88 percent has been reported in siblings having the same blood type as the affected index baby, with two-thirds of the affected siblings requiring therapy.³⁴ A higher incidence and greater severity of jaundice as a result of ABO incompatibility is reported in southeast Asians, Hispanics, Arabs, and South African and Americans of African descent than in whites.^17,66,67 This may be partly a result of the extent of development of A or B antigen sites on fetal RBCs, as well as of the prevalence of Gilbert syndrome in different patient populations.⁶⁸

The evaluation and management of HDFN require close collaboration between obstetricians, maternal fetal medicine specialists, radiologists, hematologists, transfusion medicine specialists, and neonatologists. Figure 55–2 is an algorithm for the clinical management of an alloimmunized pregnancy.

DETERMINATION OF PATERNAL ZYGOSITY AND FETAL BLOOD TYPE

When a clinically significant alloantibody is identified, or if there is a history of a previous fetus or neonate affected by HDFN, the next step is to determine if the fetus is at risk because the fetus carries the corresponding antigen. If the father is homozygous for the corresponding antigen, the fetus is at definite risk for HDFN. The child of an antigen-negative mother and a heterozygous antigen-positive father has a 50 percent chance of being antigen-positive and thus being affected by maternal alloimmunization. When the father is heterozygous or paternal zygosity is unknown, determination of fetal blood type early in pregnancy allows early institution of monitoring and therapy in antigen-positive fetuses that are at risk and forestalling invasive and potentially risky procedures in antigen negative fetuses.

Paternal zygosity is determined using serology for all common blood group antigens implicated in HDFN except for D. Alternatively, RBC genotyping can be used when typing for antigen systems where serologic reagents are rare or nonexistent. The probable RHD zygosity in RhD-positive persons may be inferred, but not definitively, from serologic phenotyping studies, based on gene frequencies in certain populations and the fact that the C/c and E/e antigens are closely linked to the RHD locus.^14,56 Elucidation of the genetic structure of the prevalent RHD locus and haplotypes responsible for RhD-negative phenotypes has led to the development of more direct and robust methods of determination of RHD zygosity molecularly. RHD zygosity testing by quantitative fluorescence polymerase chain reaction (QF-PCR) is commercially available and uses RHD (exons 5 and 7) to RHCE (exon 7) amplification ratios to determine RHD copy number. Although suitable for use in both white and ethnic African individuals, this molecular technique has a false-positive rate of 1 percent as a result of rare RHD alleles that are not expressed, and a false-negative rate of 1 percent as a result of rare partial D alleles (i.e., DBT type 1, type 2) which lack RHD exons 5 and 7 but still express RhD epitopes.^69,70

If paternal heterozygosity is suspected or confirmed, determination of fetal blood type is helpful in planning further management. There are several sources of fetal tissue for fetal blood group genotyping. These include blood obtained by cordocentesis, chorionic villus sampling, and cervical tissue obtained by transvaginal lavage; each has risks to the fetus and issues related to quality of sample. Cordocentesis, amniocentesis, and chorionic villus sampling for fetal genetic typing carry a risk of FMH with increased risk of augmenting maternal sensitization and of fetal loss.^45,46 The advent of noninvasive methods of prenatal diagnosis using fetal DNA extracted from maternal plasma as early as the first trimester of pregnancy has obviated those concerns and has dramatically improved the ability to perform molecular testing on fetal tissue.⁷¹

Circulating cell-free fetal DNA (ccff-DNA) in maternal plasma can be identified as early as 5 weeks of gestation and is derived from apoptotic syncytiotrophoblasts. Fetal DNA used for typing is extracted and then evaluated by real-time quantitative polymerase chain reaction (QT-PCR). Most protocols amplify three exons or more, which include RHD exons 4 to 7 and 10, and detect target Psi (ψ) pseudogene sequences in exon 4 to avoid false-positives when the fetus has RHDψ.^72,73 Confirmation of detection of nonmaternal markers is required and can be accomplished by testing for the presence of the Y chromosome in male fetuses and/or housekeeping genes such as hemoglobin β-chain, β-actin, albumin, or chemokine receptor 5. A recent meta-analysis reviewed 37 publications describing 44 protocols reporting noninvasive RHD genotyping using fetal DNA obtained from more than 3000 maternal blood samples; an accuracy rate of 94.8 percent was reported.⁷² Very high accuracy rates (>96 percent) have been reported for noninvasive fetal RHCE genotyping from maternal blood as well.⁷⁴ RHD ccff-DNA testing is commercially available, and is being increasingly used in the United States, United Kingdom, and Europe.⁷⁵ Although used more often in non-U.S. countries than in the United States at the present time, ccff-DNA testing may also be used to predict whether infants of alloimmunized women carry the cognate minor RBC antigen and are thus at risk for HDFN.

MATERNAL IMMUNOHEMATOLOGIC TESTING

The dual aims of maternal antenatal testing are to identify women who enter pregnancy already alloimmunized, and to identify those who are at high risk of becoming alloimmunized during pregnancy. The practice guidelines and recommendations for pregnancy-associated immunohematologic and molecular testing were established in the United States by the American Association of Blood Banks.⁵⁷

Every obstetrical patient should have samples obtained between 10 and 16 weeks’ gestation and tested for ABO and RhD type; D typing discrepancies must always be investigated and resolved. These maternal samples should also be screened for the presence of red cell alloantibodies. A second sample should be obtained at 28 weeks’ gestation to confirm the maternal blood type, and to further evaluate the pregnant mother for other red cell alloantibodies that may have been evanescent or nonexistent earlier in pregnancy. If an unexpected antibody is identified anytime during pregnancy, the specificity, titer, and likelihood of leading to HDFN must be determined. The indirect antiglobulin test (IAT), using reagent red cells expressing C, c, D, E, e, K, k, Fy^a, Fy^b, Jk^a, Jk^b, S, s, M, N, and Le^a, is typically used for maternal antibody screening, Blood samples from women with anti-D alloantibodies should be tested monthly until 28 weeks’ gestation and every 2 weeks thereafter.⁷⁶ Antibody titers are reported as the reciprocal of the highest dilution step at which agglutination is observed. A difference of two dilutions is considered a significant change. Testing should ideally be performed in parallel with previously frozen samples to minimize the possibility that changes in the titer result from differences in technique or reagent red cell selection.⁵⁷ A critical titer is defined as the titer associated with significant risk of fetal anemia or hydrops, and is the threshold at which the fetus will need monitoring. Once the critical titer is reached and a decision is made to monitor the fetus by ultrasonography or amniocentesis, further antibody titration plays no role in assessment of fetal status. The critical titer varies from 8 to 32 in different laboratories in the United States.⁷⁶ In the United Kingdom and Europe, the level of anti-D is compared to an international standard and reported in IU/mL. Anti-D levels of 4 IU/mL or greater prompt referral to a specialist fetomaternal unit for further monitoring; at levels of 4 to 15 IU/mL, there is a potential risk of moderate HDFN; levels greater than 15 IU/mL imply a risk of severe HDN.⁷⁷ In the Netherlands, anti-Kell titers as low as 2 prompt referral to a perinatal center,⁵⁹ and the critical titer is defined as 16 or greater.⁷ When RhIg has been administered during pregnancy, a positive low anti-D antibody titer may be detected (generally 2 to 4). Specific laboratory techniques may be used, if required, to distinguish between passive RhIg and active alloimmunization to D.

The significance of titer levels for antibodies other than D has not been fully defined. Maternal anti-Kell titers, in particular, correlate poorly with fetal outcome.⁶² In a review of 156 anti–Kell-positive pregnancies over 37 years, McKenna and colleagues found that all severely affected fetuses had a titer of at least 32.⁶⁰ Bowman and colleagues also noted that a titer of 32 or greater was present in 16 of 17 severely affected pregnancies, but 1 patient with a titer of 1:8 had a grossly hydropic fetus at 23 weeks’ gestation.⁶¹ Some authors recommend further testing of the fetus if a critical titer of 8 is attained and paternal red cell typing is Kell-positive.⁷⁶ In a case series of women with anti-c isoimmunization, a titer of 32 or greater was invariably associated with severe fetal or neonatal disease.⁷⁸

The imperfect predictive value of serologic tests has led to the development of functional cellular assays that measure the ability of maternal antibodies to cause red cell destruction, thus providing better noninvasive differentiation of pregnancies at increased risk of fetal anemia. In these assays, RBCs sensitized with maternal antibodies are incubated with effector cells carrying Fcγ receptors, such as lymphocytes or monocytes, to measure cellular interaction, such as binding, phagocytosis, or cytotoxic lysis.⁷⁹ Some authors have reported on the superiority of the monocyte monolayer assay, the chemiluminescence test, and the antibody-dependent cell-mediated cytotoxicity assay, compared to serologic tests, in predicting severity of HDFN. However, these tests are complex, difficult to standardize, and are not widely used in the United States.

FETAL BLOOD SAMPLING

Fetal blood sampling (also called percutaneous umbilical blood sampling or cordocentesis) allows direct measurement of blood indices to specifically evaluate the degree of severity of fetal hemolytic disease as early as 17 to 18 weeks’ gestation.⁸⁰ Indications for fetal blood sampling in alloimmunized pregnancies include fetal blood typing, confirmation of severe fetal anemia suspected based on elevated peak middle cerebral artery Doppler velocities, or ultrasonographic evidence of early or frank hydrops Historically, fetal blood samples were obtained when amniocentesis results returned with ΔOD₄₅₀ (change in optical density at 450 nm) measurements in Liley zone 3 or in the “intrauterine death zone” in the Queenan graph.^76,81,82 The procedure is performed under local anesthesia. A 20- to 22-gauge spinal needle is inserted into the umbilical vein at the level of cord insertion into the placenta under ultrasonographic guidance. Specimens of fetal blood are obtained for direct measurement of complete blood count, reticulocyte count, red cell antigen phenotyping, DAT, bilirubin, blood gases, and lactate to assess acid–base status. Blood should be available for immediate IUT when the procedure is being performed for suspected severe fetal anemia. Complications of fetal blood sampling include fetal loss, with procedure-related rates ranging from 0 to 4.9 percent, umbilical cord bleeding, chorioamnionitis, and significant risk of FMH with anamnestic maternal sensitization or the formation of additional alloantibodies.^83,84

AMNIOTIC FLUID SPECTROPHOTOMETRY

Amniotic fluid spectrophotometry has been used for the last half century using bilirubin as an indicator to measure fetal hemolysis. Although briefly reviewed here for historical perspective, this method has now largely been replaced by noninvasive fetal monitoring techniques.^76,85,86 Elevations of ΔOD₄₅₀ by spectrophotometry reflect the concentration of amniotic fluid bilirubin, which is derived from the fetus.⁸¹ The original Liley chart, from 27 weeks to term, defined three zones: readings in zone 3, the upper zone, indicate severe fetal disease with hydrops or impending fetal death; readings in zone 1, the lowest zone, indicate mild or no hemolytic disease with a 10 percent risk of needing a postnatal exchange transfusion; and readings in zone 2 indicate moderate disease. The Liley chart was later modified by Queenan to include data from 14 to 40 weeks gestation and had 4 zones, with the lowest zone representing unaffected fetuses and the highest zone associated with increased risk of intrauterine death.⁸¹

ULTRASONOGRAPHY

Ultrasonography is noninvasive, can be performed serially, and can be combined with other diagnostic studies to assess the fetal condition, estimate the need for further aggressive management, and obtain a biophysical profile of the fetus to determine fetal well-being. As hydrops develops in the anemic fetus, a consistent pattern may be noted on ultrasonography. Polyhydramnios appears first, followed by placental enlargement, hepatomegaly, pericardial effusion, ascites, scalp edema, and pleural effusions in succession. Nevertheless, in the absence of overt hydrops, ultrasonographic parameters, such as intrahepatic and extrahepatic vein diameters, abdominal and head circumference, head-to-abdominal-circumference ratio, intraperitoneal volume, splenic size, and liver length have not been reliable in distinguishing mild from severe fetal anemia.⁸²

In addition to traditional ultrasonography, measurement of fetal cerebral blood flow has become an extremely valuable technique in assessing fetal anemia. Decreased viscosity of the blood and increased cardiovascular output in the anemic fetus lead to a hyperdynamic circulation; hypoxia further increases blood flow velocity. Values greater than 1.5 multiples of the median for gestational age highly correlate with moderate or severe fetal anemia (Fig. 55–3).⁸⁶ Doppler measurement of peak velocity of systolic blood flow in the middle cerebral artery is more sensitive and accurate for detecting severe fetal anemia than measurement of amniotic fluid ΔOD₄₅₀.⁸⁵ Measurements may be initiated as early as 18 weeks of gestation and repeated at 1- to 2-week intervals until 35 weeks’ gestation. After 38 weeks, a higher false-positive rate in the detection of fetal anemia necessitates amniocentesis for ΔOD₄₅₀ and fetal lung maturity testing if elevated levels are noted.⁷⁶

ANTENATAL MANAGEMENT

Intrauterine Transfusion

Prior to the institution of IUTs, many severely affected fetuses died in utero or soon after birth. IUT corrects fetal anemia and reduce the risk of congestive heart failure and hydrops fetalis. Fetal bilirubin is cleared very efficiently by the placenta and the mother, so bilirubin removal is not necessary until after birth. Percutaneous intraperitoneal fetal transfusion, pioneered by Liley in the 1960s,⁵ has been largely replaced by ultrasound-guided direct intravascular transfusion into the umbilical vein.^80,87 The intravascular technique circumvents the problem of erratic and often poor absorption of RBCs from the peritoneal cavity in such fetuses. However, intraperitoneal transfusions may be necessary when intravascular access is difficult, as in early pregnancy when the umbilical vessels are narrow or later when increased fetal size prevents access to the umbilical cord.^88,89 The first fetal blood sampling with transfusion ideally is performed before hydrops develops. Transfusions are given at fetal hematocrit levels of 25 to 30 percent or less, or if the fetal hemoglobin is 4 to 6 standard deviations below the mean for gestational age. Generally, the hematocrit drops by 1 to 2 percent per day in the transfused hydropic fetus. The fall in hematocrit is rapid in fetuses with severe hemolytic disease, often necessitating a second transfusion within 7 to 14 days. The interval between subsequent transfusions varies, but may be 21 to 28 days. The nonhydropic fetus can tolerate rapid RBC infusions of 5 to 7 mL/min because of the capacitance of the placenta. The hydropic fetus requires slower transfusion rates and can tolerate only smaller, more frequent transfusions. Very low pretransfusion fetal hematocrit levels, rapid large increases in posttransfusion hematocrit level, and increases in umbilical venous pressure during IUTs are associated with fetal death after transfusion.^90,91

RBCs for IUT are typically crossmatch compatible with the mother’s plasma and negative for any identified antibody; they should also be irradiated and leukoreduced.^92,93 Many centers do not use RBCs heterozygous for sickle hemoglobin to prevent sickling in the fetus at low oxygen tension although direct evidence of its benefit is lacking, and 5 to 7 days old to maximize circulatory half-life. Obtaining RBCs from a rare donor registry may be required in cases of unusual or combination antibodies. In this circumstance, frozen, deglycerolized RBCs may be the only available product. Some centers use maternal RBCs for IUTs, supporting serial maternal donations with iron and folate therapy.⁹⁴ RBC transfusion is calculated to increase the fetal hematocrit to between 40 and 45 percent. The RBCs are often washed free of additive solutions, and packed to a hematocrit of 70 to 85 percent in a volume calculation based on estimated fetal placental blood volume, fetal hematocrit, and hematocrit of donor RBCs. Various normograms and formulas for the calculation of donor blood volume have been published.^95,96 If the hematocrit of the donor unit is approximately 75 percent, multiplying the estimated fetal weight in grams (estimated by ultrasonography) by a factor of 0.02 provides a fairly accurate estimate of the volume of RBCs to be transfused to achieve a fetal hematocrit increment of 10 percent.⁹⁷

Van Kamp and colleagues estimated procedure-related complication rates of 2.9 percent in the absence of hydrops and 3.9 percent in the presence of hydrops in a cohort study of 254 fetuses treated with 740 IUTs in a single center.⁹⁸ The most common problem was transient fetal heart rate abnormalities, which occurred in 8 percent of procedures. The procedure-related pregnancy loss rate was calculated to be 1.6 percent per procedure. Fetal distress during or after the procedure may result in emergency cesarean section. Cord accidents such as hemorrhage from the puncture site or rupture of the cord are rare. Additional alloantibodies may develop in already alloimmunized women undergoing IUT; in one study, 25 percent (53 of 212) of “responder” women treated with IUT formed new alloantibodies, 53 percent of which were directed against non-Rh and -K antigens.⁹⁹

Delivery

The decision regarding the appropriate time to deliver the baby is based on gestational age, fetal weight and lung maturity, fetal response to the IUTs, ease of performing the transfusions, and antenatal ultrasonography and Doppler studies for fetal anemia. Transfusions usually are provided up to 35 weeks so as to prolong gestation safely until the risks of preterm birth and its attendant complications are minimized, with delivery once adequate lung maturity has occurred.⁷⁶

Other Therapies

In women with severe alloimmunization and with fetal losses or hydrops very early in pregnancy, a variety of methods have been used to suppress the antibody response and prolong survival of the fetus until IUT becomes technically feasible. Use of intravenous immunoglobulin (IVIG), serial plasmapheresis, or plasmapheresis combined with IVIG have been successful in some cases.^89,100,101 IVIG may cause nonspecific Fc blockade of the fetal reticuloendothelial system.

POSTNATAL MANAGEMENT

Neonatal Monitoring

A sample of cord blood should be collected from all newborns at the time of delivery. However, specific testing of cord blood samples is performed only if the mother is Rh-negative, if the maternal serum contains red cell alloantibodies of potential clinical significance, or if the neonate develops signs of hemolytic disease. Tests should include ABO and Rh typing and a DAT. Many birth hospitals routinely test cord blood for the infant’s blood type and DAT if the mother is O Rh-positive in order to predict which infants are at increased risk of hyperbilirubinemia. In severe Rh alloimmunization, high titers of maternal antibody may block Rh-antigenic sites on the neonatal red cells, leading to false-negative Rh typing. Antepartum RhIg given to the mother may cause a weakly positive DAT result in the infant at birth. Contamination of the cord blood sample with Wharton jelly during collection can also result in a false-positive DAT result. Although the DAT usually is positive in all forms of alloimmune HDFN, the test cannot predict reliably the degree of clinical severity.^102,103 This is especially true for cases resulting from ABO sensitization. When fetomaternal ABO incompatibility is present, the presence of maternally derived IgG anti-A or anti-B in the infant’s serum may be demonstrated by the IAT to support the diagnosis of ABO hemolytic disease. On the other hand, it is important to bear in mind that hemolysis in ABO-incompatible, DAT-negative infants may result from hematologic causes other than alloimmunization or from red cell membrane defects (Chap. 46).¹⁰⁴ Elution of maternal antibody from the infant’s red cells, followed by tests to determine the specificity of the antibody in the eluate, may be useful, particularly when several antibodies are present in the maternal serum or when the maternal antibody screen is negative.⁵⁷ Infants who received IUTs may have mild or moderate anemia with little reticulocytosis. Because most of their circulating red cells are transfused antigen-negative cells, the DAT may be negative, but the IAT will be strongly positive.

Cord blood hemoglobin and indirect bilirubin determinations are useful in determining disease severity. Most infants with cord hemoglobin levels within the age-adjusted normal range do not require exchange transfusion. Usually, a cord hemoglobin level less than 11 g/dL in a term newborn and/or a cord-indirect bilirubin level greater than 4.5 to 5 mg/dL indicate severe hemolysis and often warrant early exchange transfusion. Early exchange transfusion also may be indicated if the rate of rise of bilirubin, measured every 4 to 6 hours, exceeds 0.5 mg/dL per hour. The reticulocyte count usually is greater than 6 percent and may approach 30 to 40 percent in severe Rh disease. The blood film in Rh disease is characterized by increased nucleated RBCs, polychromasia, and anisocytosis. Alternatively, the blood film in ABO HDN is marked by microspherocytes (Fig. 55–4). Severely affected infants may also have thrombocytopenia. Low reticulocyte counts disproportionate to the low hematocrit may be noted in Kell-mediated HDN. Severely affected infants may have hypoglycemia, secondary to hyperinsulinemia. Arterial blood gas analysis may reveal metabolic acidosis and/or respiratory decompensation, and hypoalbuminemia is often present.

Immediate Postnatal Management

Results of antenatal monitoring and obstetric interventions during pregnancy and the history of the outcome of previous pregnancies allow the neonatal team to anticipate the needs of the infant born with hemolytic disease. In infants with severe hemolytic disease without the benefit of IUTs, severe anemia and hydrops are the immediate life-threatening concerns and often are accompanied by perinatal asphyxia, surfactant deficiency, hypoglycemia, acidosis, and thrombocytopenia. Exchange transfusions and phototherapy are the mainstays of treatment.

Resuscitation and stabilization of hydropic infants is challenging. Endotracheal intubation and positive-pressure ventilation with oxygen is usually necessary. Drainage of pleural effusions and ascites may be required to facilitate gas exchange. Metabolic acidosis and hypoglycemia require correction. A partial exchange transfusion may initially be performed using packed red cells to improve hemoglobin levels and oxygenation. A double-volume exchange transfusion is considered only after the initial stabilization.

In a study of 191 infants born alive after IUTs between 1988 and 1999, the hematocrit at birth ranged from 13 to 51 percent.¹⁰⁵ Endotracheal ventilation was required more often in babies who had been severely hydropic in utero, but the requirements for exchange transfusion or simple transfusion did not differ between babies who had been hydropic in utero and those without evidence of hydrops. Although some centers report no difference in the frequency of exchange transfusions in babies who have had IUTs compared with babies who have not had IUTs,^105,106 other centers report that infants who received multiple IUTs are usually born closer to term and often require less phototherapy and fewer exchange transfusions in the neonatal period.^25,107 Nonetheless, many infants with severe HDFN require additional RBC transfusions for severe and prolonged hyporegenerative anemia secondary to suppression of fetal erythropoiesis.^21,22,106 Approximately three-quarters of term and near-term infants who had received IUT for Rh HDFN required transfusion within 6 months of age, compared with 26 percent of infants with Rh HDFN who had not received an IUT.¹⁰⁶ Thus, careful monitoring of the baby and the baby’s lab values are necessary not only during the initial hospital course but also after hospital discharge.

Exchange Transfusion

Exchange transfusion corrects anemia, removes bilirubin and free maternal antibody in the plasma, and replaces the infant’s blood with antigen-negative RBCs that should have normal in vivo survival. Neonatal exchange transfusions can be performed by a continuous technique (simultaneous withdrawal and replacement) or discontinuous technique (alternating withdrawal and replacement). Regardless of the technique, the kinetics of exchange are very similar. A double-blood-volume exchange replaces approximately 85 percent of the infant’s blood volume with antigen-negative RBCs; however, the amount of bilirubin or maternal alloantibody removed by exchange transfusion is significantly less (25 to 45 percent) reflecting the equilibrating tissue-bound plasma pool. Infusion of albumin prior to the exchange transfusion may help bilirubin binding, thus increasing the amount of bilirubin removed. Equilibration of extravascular and intravascular bilirubin, and continued breakdown of red cells by persisting maternal antibodies, result in a rebound of bilirubin following initial exchange transfusion, sometimes requiring repeated exchange transfusions in severe hemolytic disease.

The ideal volume for an exchange transfusion is twice the infant’s blood volume. There is little benefit achieved by exceeding two blood volumes because the efficiency of exchange transfusion declines exponentially as the procedure continues. The volume needed for double-volume exchange depends on the total blood volume (TBV) recognizing differences in term and preterm infants:

• Double-blood-volume exchange volume (term) = 85 mL/kg × 2, or 170 mL/kg

• Double-blood-volume exchange volume (preterm) = 100 mL/kg × 2, or 200 mL/kg

To perform the exchange transfusion, aliquots of the reconstituted whole blood product are administered while equal amounts of the infant’s blood are withdrawn. Careful attention not to exceed 2 mL/kg per minute (continuous) or 5 mL/kg at a time over 3 to 10 minutes (discontinuous technique) is required to prevent rapid fluctuations in arterial and intracranial pressure.¹⁰⁸

The indications for “early” exchange transfusions performed within the first 12 hours of life have remained essentially unchanged over the last 45 years, with minor modifications. Cord hemoglobin levels equal to or less than 11 g/dL, cord bilirubin levels equal to or greater than 5.5 mg/dL, and rapidly rising total serum bilirubin (TSB) equal to or greater than 0.5 mg/dL per hour despite phototherapy are commonly used criteria for early exchange transfusions. Early exchange transfusion has the advantage of replacing sensitized RBCs with normal RBCs, thereby removing not only bilirubin but also the source of future bilirubin. Because bilirubin is distributed in the extracellular fluids, efficiency is enhanced by removing sensitized cells early in the process. Newborns that have been treated with serial IUTs until term often do not require exchange transfusions; however, late anemia is common because of IUT-induced erythropoietic suppression, which may last for many weeks after delivery.¹⁰⁹

“Late” exchange transfusions are performed when serum bilirubin levels threaten to exceed approximately 20 to 22 mg/dL in term infants. The American Academy of Pediatrics (AAP) Subcommittee on Hyperbilirubinemia provided revised guidelines for exchange transfusion in infants 35 or more weeks’ gestation.¹¹⁰ In view of the fact that bilirubin levels rise steadily from birth and peak at approximately 72 to 96 hours of age, exchange transfusion should be considered if serum bilirubin levels reach 15 mg/dL in an infant of 35 weeks’ gestation or 17 mg/dL in an infant of 38 weeks’ gestation despite intensive phototherapy. Immediate exchange transfusion is recommended in infants showing signs of acute bilirubin encephalopathy, even if bilirubin levels are falling.¹¹⁰ Conjugated or direct bilirubin values are not subtracted from total bilirubin levels when considering levels for exchange transfusions. Exchange transfusions are performed at lower bilirubin levels in premature infants, particularly those with hypoxemia, acidosis, and hypothermia, but little data are available to guide intervention in these infants. In infants with birth weights of at least 1500 g, exchange transfusions usually are performed at TSB of 13 to 16 mg/dL but may be considered even at levels as low as 8 to 9 mg/dL in sick babies of 24 weeks’ gestation.¹¹¹ The bilirubin-to-albumin ratio (mg/dL:g/dL), considered to be a surrogate measure of free bilirubin, may provide additional data in determining the need for exchange transfusion in both term and preterm neonates.¹¹²

Blood components chosen for the exchange transfusion should be ABO and Rh compatible (Rh-negative in Rh HDN), negative for offending antibody(ies), and crossmatch compatible with maternal serum. In the case of ABO HDN, O RBCs should be chosen for exchange out of concern that the more developed A or B antigens on any transfused adult donor RBCs may more avidly bind maternal anti-A or anti-B and may result in hemolysis. Either reconstituted whole blood (WB) (e.g., RBCs plus fresh-frozen plasma [FFP]) or stored WB if available can be used for neonatal exchange transfusions. The RBCs are reconstituted with AB or compatible plasma to a final hematocrit of 50 to 60 percent. Fresh (<7 days) RBCs should be used. When fresh RBC units are unavailable, some centers wash the RBCs and transfuse as soon as possible after washing to avoid hyperkalemia. Additionally, the RBC units should be leukoreduced, gamma irradiated, and sickle-negative.¹¹³

Potential complications of exchange transfusion include hypocalcemia, hyperglycemia, hypoglycemia, thrombocytopenia, dilutional coagulopathy, neutropenia, disseminated intravascular coagulation, umbilical venous and/or arterial thrombosis, necrotizing enterocolitis, and infection. Thrombocytopenia and hypocalcemia are reported to be the most common complications (incidence ranging from 29 to 47 percent).^114,115 Thrombocytopenia results from a dilutional effect of replacing platelet rich neonatal WB with platelet-deficient reconstituted WB. Infants who may be thrombocytopenic from severe HDFN or other comorbidities should be monitored closely after an exchange transfusion as they may require platelet transfusion. Hypocalcemia occurs as a result of the citrate load infused, which an immature neonatal liver has difficulties metabolizing. In anticipation of hypocalcemia, ionized calcium levels should be monitored throughout the exchange transfusion procedure, and intravenous calcium replacement may be needed in sick preterm infants. Furthermore, attempts should be made to correct conditions that may potentiate the symptoms of hypocalcemia such as alkalosis, hypothermia, hypomagnesemia, and hyperkalemia.¹¹⁶

In a retrospective review of exchange transfusions performed in two neonatal intensive care units between 1981 and 1995, the risk of death or permanent serious sequelae was reported to be as high as 12 percent in sick infants, compared with less than 1 percent in healthy infants. Adverse outcomes were more frequent in exchanges done on preterm infants younger than 32 weeks, infants with other significant comorbidities, and when umbilical catheters were used rather than other means of central venous access.¹¹⁷ Another center reported no increase in the number of complications and no exchange transfusion–related deaths over a 21-year period, even though there was a decline in the frequency of exchange transfusions performed over the years.¹¹⁵ Careful clinical judgment is required to balance the potential risk of adverse events from exchange transfusion with the risk of bilirubin encephalopathy in neonates who are premature, sick, or both.

Phototherapy

Phototherapy is the mainstay of treatment for unconjugated hyperbilirubinemia; the objective of treatment is preventing bilirubin neurotoxicity. Exposure of bilirubin to light results in structural and configurational isomerization of bilirubin to less toxic and less lipophilic products that are excreted efficiently without hepatic conjugation. The effectiveness of phototherapy is influenced by the wavelength and irradiance of light, the surface area of exposed skin, and the duration of exposure. Intensive phototherapy involves the use of high levels of irradiance (≥30 μW/cm²) in the 430- to 490-nm band, delivered to as much of the infant’s surface area as possible. Intensive phototherapy effectively reduces bilirubin levels and decreases the need for exchange transfusions for hyperbilirubinemia in ABO and Rh HDN.^118,119 Early and intensive phototherapy should be initiated in infants with moderate or severe hemolysis or in infants with rapidly rising bilirubin levels (>0.5 mg/dL per hour). In full-term infants (at least 38 weeks’ gestation) with HDFN, intensive phototherapy should be initiated if TSB levels are 5 mg/dL or greater at birth, 10 mg/dL at 24 hours after birth, or approximately 13 to 15 mg/dL at 48 to 72 hours after birth.¹¹⁰ Phototherapy is recommended at lower levels for preterm or sick infants. Therapy often is initiated at TSB less than 5 mg/dL in preterm infants with HDFN so as to avoid potentially risky exchange transfusions.^110,111

Other Therapies

A number of small studies have reported on the successful administration of high-dose IVIG as an adjuvant treatment to standard therapy for HDN as way to prevent the need for exchange transfusions.¹⁰⁹ The decreased bilirubin levels in infants treated with IVIG is attributed to reduction in hemolysis secondary to blockade of mononuclear phagocyte Fc receptors. A Cochrane meta-analysis of the three largest studies demonstrated that IVIG administration (0.5 to 1.0 g/kg) may prevent exchange transfusion for many term infants with HDN, which contributed to an AAP published recommendation that IVIG be considered in Rh HDN when TSB continues to rise despite aggressive phototherapy or when the TSB is within 2 to 3 mg/dL of the exchange level.¹¹⁰ However, a prospective randomized control trial showed that prophylactic IVIG (0.75 g/kg) did not decrease the duration of phototherapy, maximum bilirubin levels, need for RBC transfusion or for exchange transfusion in Rh HDN.¹²⁰ Although, the majority of the infants included in this study were treated with IUT, IVIG was not effective at reducing the need for exchange transfusion in either the group treated with IUTs or those who were not. Thus, there is no universal approach to the use of IVIG in patients with HDN; it seems justified as a temporizing measure to exchange transfusion in neonates with severe HDN unmodified by IUT or neonates with HDN where a previous sibling had suffered from severe disease requiring exchange transfusion.¹⁰⁹

Recombinant human erythropoietin (rHuEPO) given subcutaneously at a dose of 200 to 400 U/kg given three times weekly for 2 weeks is sometimes administered to infants in an effort to reduce or prevent transfusion for late-onset anemia from HDN. Some studies show its use to decrease the need for postnatal transfusions in infants with late hyporegenerative anemia of Rh HDN and in neonates with Kell HDN.^21,121,122 In one study of 103 patients with Rh HDN, administration of 200 U/kg of rHuEPO, three times per week for 6 weeks, reduced the number of RBC transfusions to a mean of 1.5, and 55 percent of patients did not require any transfusions.¹²¹ rHuEPO is more effective in decreasing future transfusion needs in neonates that never received IUTs suggesting that IUTs may decrease the neonatal response to rHuEPO.¹²³ Despite encouraging reports in relatively small numbers of neonates, it remains unclear whether rHuEPO, or the longer-acting analogue darbepoetin, offer a distinct clinical benefit in regard to decreasing donor exposures or improvement of morbidity and/or mortality in this population. rHuEPO has been shown to lessen neurologic sequelae in term infants with hypoxic ischemic encephalopathy, therefore it may have a role as a potential treatment for perinatal brain injury in the future.¹²⁴

Perinatal survival rates greater than 90 percent have been achieved with IUT in nonhydropic fetuses with severe HDFN.^98,105 The overall survival rate for hydropic fetuses is lower (78 to 89 percent) despite IUTs.^105,107 An 11-year study (from 1988 to 1999) which examined 80 fetuses with immune-mediated hydrops reported a survival rate for fetuses with mild hydrops of 98 percent, and an intrauterine reversal of hydrops in 88 percent of the fetuses. The outcome in severe fetal hydrops was poor, with reversal of hydrops in 39 percent of cases, and a survival rate of 26 percent for fetuses with persistent hydrops.¹⁰⁵ This study underscores the importance of early diagnosis and treatment of fetal anemia, before hydrops develops. Implementation of a nationwide first trimester screening program for red cell antibodies in the Netherlands in 1998 was associated with increased referrals for suspected fetal anemia, more timely referrals and an increase in perinatal survival in Kell HDFN from 61 to 100 percent.⁵⁹

The neurodevelopmental outcome for infants saved by IUTs has generally been good, with almost 90 percent of survivors being free of disability, even when they have been profoundly anemic in utero.^25,107 The LOTUS study evaluated 291 children at a median age of 8.2 years (range: 2 to 17 years) to determine the incidence and risk factors for neurodevelopmental impairment in children with HDFN treated with IUT. The overall incidence of neurodevelopmental impairment was 4.8 percent with severe developmental delay seen in 3.1 percent of children. Cerebral palsy was detected in 2.1 percent and bilateral deafness in 1 percent of the children. Severe hydrops was identified as a strong predictor for neurodevelopmental impairment.¹²⁵

It is estimated that kernicterus, secondary to bilirubin encephalopathy, is associated with at least 10 percent mortality and 70 percent long-term morbidity.¹²⁶ Although the preponderance of kernicterus cases occur in infants with bilirubin levels higher than 20 mg/dL, it has been shown that when treated promptly with phototherapy or exchange transfusion, peak bilirubin levels in the range of 25.0 and 29.9 mg/dL were not associated with adverse neurodevelopmental outcomes in term or near-term infants.^126,127 However, there was an adverse association with IQ among infants with a positive DAT and a TSB level equal to or greater than 25 mg/dL, supporting the AAP recommendations to initiate treatment at lower bilirubin levels for jaundiced infants if they have a positive DAT.^27,110,127

Transfusion of blood phenotypically matched for D and for Kell antigens has been advocated for all premenopausal women to prevent primary RBC alloimmunization.^7,13 RhIg prophylaxis is very effective when administered to women exposed to RhD-positive RBCs either from pregnancy or from transfusion; however, once alloimmunization has occurred, RhIg is not effective for preventing or reducing the severity of HDFN. Unlike RhD, there are no commercially available immune globulin products for prevention of alloimmunization to minor RBC antigens (including non-D Rh antigens). If a nonpregnant woman is found to have a RBC alloantibody, counseling should be provided regarding the potential effects of the antibody on future pregnancies. Interventions which can be applied but are seldom offered to prevent HDFN in high risk maternal-paternal pairs include: artificial insemination with sperm from an antigen-negative donor; pre-implantation genetic diagnosis selecting for antigen-negative embryos; and surrogate pregnancy.¹²⁸

Rh IMMUNOGLOBULIN

Use of RhIg is the standard of care for the prevention of maternal D immunization. Postpartum administration of RhIg to all nonsensitized Rh-negative women who deliver an Rh-positive infant decreases the incidence of Rh isoimmunization from 12 percent to approximately 2 percent. Further reduction in the incidence of Rh-isoimmunization (to 0.1 percent) has been achieved by antepartum RhIg prophylaxis at 28 weeks’ gestation. This is the current standard recommendation in the United States.¹²⁹ In the United Kingdom, routine antenatal anti-D prophylaxis can be given as two doses of anti-D immunoglobulin of 500 IU (one at 28 weeks’ and one at 34 weeks’ gestation), as two doses of anti-D immunoglobulin of 1000 to 1650 IU (one at 28 weeks’ and one at 34 weeks’ gestation), or as a single dose of 1500 IU at 28 weeks’ gestation.¹³⁰ Although the efficacy of RhIg is clear, its mechanism of action is not. Some of the proposed mechanisms are accelerated clearance and destruction of D-positive red cells from the circulation, antibody-mediated immune suppression, and the production of immunomodulatory cytokines.⁵¹

RhIg is prepared from plasma pools of screened, sensitized human donors; the plasma is tested using polymerase chain reaction (PCR) and serologic methods for all known transfusion transmitted organisms. Several steps are taken to further inactivate potential infectious organisms, including solvent-detergent treatment, ion exchange chromatography, and nanofiltration. There are at least four formulations of RhIg, including two preparations that may be administered intravenously.¹³¹ A 300 mcg (1500 IU) dose of RhIg affords protection against a fetomaternal transfusion of 15 mL of Rh-positive RBCs or 30 mL of Rh-positive WB. However, FMH in excess of 30 mL may occur in women without predisposing risk factors.^42,43 The blood of all Rh-negative nonimmunized women should be tested for FMH approximately 1 hour after delivery of an Rh-positive baby.^57,129 During the antenatal period, testing is indicated after 20 weeks’ gestation if clinical circumstances suggest the possibility of excessive transplacental hemorrhage (e.g., abdominal trauma or abruptio placentae). Screening for FMH can be performed by the rosette test, which detects as little as 2.5 mL of WB. If the rosette test result is positive, the number of fetal red cells in the maternal circulation is quantified more accurately by the Kleihauer-Betke test, which is based on the resistance of fetal hemoglobin to acid elution, unlike adult hemoglobin (Fig. 55–5).¹³² False-positive results can be obtained in maternal conditions associated with increased fetal hemoglobin, such as hereditary persistence of fetal hemoglobin, sickle cell disease, or sickle cell trait. Flow cytometric methods are used in some laboratories for both screening and quantification of fetal red cells.

RhIg should be administered as soon as possible, within 72 hours of delivery of an Rh-positive baby. RhIg is thought to be ineffective once alloimmunization to RhD antigen has occurred. RhIg is also indicated following pregnancy termination, miscarriage, amniocentesis, chorionic villus sampling, or other manipulation during pregnancy. A smaller 50-mcg dose is adequate if pregnancy is terminated at less than 12 weeks’ gestation.¹²⁹ If therapeutic or spontaneous abortion occurs after the first trimester, the standard 300-mcg dose is recommended.¹²⁹ If a woman was exposed to more than 30 mL of D-positive blood, the dose of RhIg should be calculated to cover the volume of D-positive cells to prevent immunization (20 mcg of RhIg for 1 mL of D+ RBCs or 2 mL of WB).^57,131,133 Despite the efficacy of RhIg, RhD alloimmunization occurs despite recommended prophylaxis in 0.1 percent of pregnancies, but many cases of preventable Rh alloimmunization continue to occur because of failure to seek medical care or because of failure to implement immunoprophylaxis protocols.

HDFN is a clinically significant problem that may potentially affect any pregnancy. Although a number of strategies are in place to prevent RhD HDFN, few options exist to prevent the development of non-D HDFN. Researchers continue to investigate strategies to prevent primary maternal RBC alloimmunization to RhD and to non-D antigens, as well as strategies to mitigate the dangers of existing maternal RBC alloantibodies. Over the past decade, the ability to identify pregnancies/fetuses “at risk” because of maternal RBC alloimmunization has significantly improved. In particular, the use of noninvasive investigations, such as evaluation of fetal RBC antigen zygosity by circulating cell free fetal DNA testing and the evaluation of fetal anemia by middle cerebral artery Doppler ultrasonography, have advanced the care of fetuses at risk for HDFN. Through the continued combined efforts of maternal–fetal medicine specialists, hematologists, transfusion medicine physicians, radiologists, neonatologists, and researchers, in combination with advancements in basic science research involving alloimmunization, the care for infants at risk for HDFN, and their mothers, will likely continue to improve in the years to come.