Hemolytic Disease of the Newborn

A French midwife was the first to report hemolytic disease of the newborn (HDN) in a set of twins in 1609. In 1932, Diamond and colleagues described the relationship among fetal hydrops, jaundice, anemia, and erythroblasts in the circulation, a condition later called erythroblastosis fetalis. Levine later determined the cause after Landsteiner and Weiner discovered the Rh blood group system in 1940. In 1953, Chown subsequently confirmed the pathogenesis of Rh alloimmunization to be the result of passage of Rh-positive fetal RBCs after transplacental hemorrhage into maternal circulation that lacked this antigen.

In 1966, 2 groups from the United Kingdom and the United States demonstrated, in a combined study, that anti-D immunoglobulin G (IgG) prophylaxis soon after delivery prevented sensitization in Rh-negative women. The World Health Organization (WHO) technical report in 1971 recommended that a dose of 25 mcg (125 IU) of anti-D immunoglobulin G (IgG) should be given intramuscularly for every 1 mL of fetomaternal hemorrhage of Rh-positive packed RBCs or 2 mL of whole blood. ^[1]

In 1998, this recommendation was reinforced by the American Association of Blood Banks and the American College of Obstetrics and Gynecologists with inclusion of prophylaxis at 28 weeks' gestation. ^[2] Routine use of Rh IgG prophylaxis resulted in a significant decline in the incidence of RhD alloimmunization, ^[3] and erythroblastosis fetalis has become rare. The perinatal effects of maternal Rh alloimmunization are now referred to as hemolytic disease of the fetus and newborn, and fetal manifestations of the disease are more appreciated with newer technologies such as cordocentesis and fetal ultrasonography.

Genetics

Although the Rh antibody was and still is the most common cause of severe hemolytic disease of the newborn (HDN), other alloimmune antibodies belonging to Kell (K and k), Duffy (Fya), Kidd (Jka and Jkb), and MNSs (M, N, S, and s) systems do cause severe HDN. ^[4] The Rh blood group system uses Fisher-Race nomenclature, and the Rh gene complex consists of 3 genetic loci each with 2 major alleles. They code for 5 major antigens denoted by letters, C, c, E, e, and D. Rh blood group antigens are inherited as determined by at least 2 homologous but distinct membrane-associated proteins. Two separate genes (RhCE and RhD), located on the short arm of chromosome 1, encode Rh proteins. Each gene is 10 exons in length, and a 96% homology between these genes is observed.

Production of 2 distinct proteins from the RHCE gene is due to alternative splicing of messenger RNA. Rh gene complex is described by 3 loci, and, therefore, 8 gene complexes are possible. These complexes are as follows (listed in decreasing order of frequency among whites): CDe, cde, CDE, cDe, Cde, cdE, CDE, and CdE. Expression is limited to red blood cells (RBCs), with an increasing density during their maturation, unlike the ABH system, which exists in a wide variety of tissues. Rh antigen is not expressed on RBC progenitors.

Of individuals who are Rh positive, 45% are homozygous (CDe/CDe), and 55% are heterozygous (CDe/cde) for the RhD gene. The Rh-negative phenotype represents absence of D protein on RBCs and most commonly results from deletion of the RHD gene on both chromosomes. However, the RHD gene has significant heterogeneity, and several inherited mutations and rearrangements in its structure can result in a lack of expressions of the RhD phenotype as well.

Important examples of such mutations include the RHD pseudogene and RHD-CE-D hybrid gene. The former leads to a stop codon in RHD gene and results in a lack of transcription product despite all intact exons. It is found in 70% of South African blacks and in 25% of African Americans. The RHD-CE-D (Ccde) gene is also found in 22% of D-negative African Americans. It also results in an Rh positive genotype but a negative phenotype. Most Caucasians who are RhD negative have a complete deletion of RHD gene whereas only 18% of African blacks and 54% of African Americans who are RhD negative have complete deletion of the gene; the rest have above nonfunctional variants of the RHD gene. ^[5]

Beyond the 5 major antigens, more than 100 antigenic variants of Rh group system have been identified. Individuals with these weak-D phenotypes comprise of 2 populations: first group (90%) that expresses normal but reduced quantities of D antigen on the RBC surface, and most cannot be sensitized to produce anti-D. However, the second group (remaining 10%) known as partial-D (eg, Cw, Du) that express partial D epitopes on RBC surface and can make anti-D and rarely experience fatal HDN. The partial D phenotype results from amino acid sunstitution in the active RhD epitope. ^[6] Most women with partial-D phenotype are classified as Rh negative on routine testing and are candidates for Rh immune globulin (RhIG). Currently, testing of all Rh-negative women for weak expression of D is not recommended. However, Rh-negative infants born to Rh-negative women should undergo testing to detect the partial-D phenotype so that RhIG can be administeredin theevent ofweak expression.

Frequency of Rh negativity is higher in whites (15%) than in blacks (5%) and Hispanics (8%) and is rare in Eskimos, Native Americans, Japanese, and Asians, especially in Chinese individuals. The paternal heterozygosity determines the likelihood of an Rh-positive child being born to an Rh-negative mother. ^[3]

Pathophysiology

The exposure of the Rh-negative mother to Rh-positive red cells occurs as a result of asymptomatic fetomaternal hemorrhage during pregnancy. The Kleihauer-Betke acid elution technique that determines the proportion of fetal RBCs in maternal circulation has shown the incidence of fetomaternal hemorrhage to be 75% of all pregnancies. Incidence and degree of such hemorrhage appears to increase with gestation. Fetomaternal hemorrhage has been documented in 7%, 16%, and 29% of mothers during their first, second and third trimesters, respectively. Risk is also increased in pregnancies complicated by placental abruption, spontaneous or therapeutic abortion, and toxemia, as well as after cesarean delivery and ectopic pregnancy.

Procedures such as amniocentesis, chorionic villus sampling, and cordocentesis also increase the risk of alloimmunization. Because the transplacental hemorrhage is less than 0.1 mL in most pregnancies, most women are sensitized as a result of small, undetectable fetomaternal hemorrhage.

After the initial exposure to a foreign antigen, B-lymphocyte clones that recognize the RBC antigen are established. The maternal immune system initially produces antibodies of the immunoglobulin M (IgM) isotype that do not cross the placenta and later produces antibodies of the IgG isotype that traverse the placental barrier. Predominant antibody subclass appears to be IgG1 in one third of individuals whereas a combination of IgG1 and IgG3 subclasses are found in the remaining individuals.

IgG3 is more efficient in binding to reticuloendothelial cells and causing hemolysis because of its longer hinge region. This is termed the primary response and is dose dependent (documented in 15% of pregnancies with 1 mL of Rh-positive cells in an Rh-negative individual compared with 70% of pregnancies after 250 mL). A repeat exposure to the same antigen rapidly induces the production of IgG. This secondary immune response can be induced with as little as 0.03 mL of Rh-positive RBCs.

Findings from murine models appear to support a potential role for epitope masking, immune deviation and/or antigen modulation in the mechanism of action of IgG-mediated inhibition of erythrocyte alloimmunization. ^[7]  Moreover, blends of monoclonal antibodies targeting nonoverlapping epitopes on the RBC surface may have the potential to improve the efficacy of monoclonal antibodies approaching that of polyclonal IgG.

The risk of Rh immunization after the delivery of the first child to a nulliparous Rh-negative mother is 16% if the Rh-positive fetus is ABO compatible with its mother, 2% if the fetus is ABO incompatible, and 2-5% after an abortion. The ABO-incompatible RBCs are rapidly destroyed in the maternal circulation, reducing the likelihood of exposure to the immune system. The degree of Rh sensitization of the mother is directly related to the amount of fetomaternal hemorrhage (ie, 3% with < 0.1 mL compared with 22% with >0.1 mL).

After sensitization, maternal anti-D antibodies cross the placenta into fetal circulation and attach to Rh antigen on fetal RBCs, which form rosettes on macrophages in the reticuloendothelial system, especially in the spleen. These antibody-coated RBCs are lysed by lysosomal enzymes released by macrophages and natural killer lymphocytes and are independent of the activation of the complement system.

Reticulocytosis is noted when fetal Hb deficit exceeds 2 gm/dl compared with gestational age norms. Tissue hypoxia develops as fetal anemia becomes severe. When the hemoglobin (Hb) level drops below 8 g/dL, a rise in umbilical arterial lactate occurs. When the Hb level drops below 4g/dL, increased venous lactate is noted. Hydrops fetalis occurs when fetal Hb deficit exceeds 7 g/dL and starts as fetal ascites and evolves into pleural effusions and generalized edema. The various mechanisms responsible for hydrops are hypoalbuminemia secondary to depressed liver function, increased capillary permeability, iron overload secondary to hemolysis, and increased venous pressures due to poor cardiac function. ^[8]

Prolonged hemolysis leads to severe anemia, which stimulates fetal erythropoiesis in the liver, spleen, bone marrow, and extramedullary sites, such as the skin and placenta. In severe cases, this can lead to displacement and destruction of hepatic parenchyma by erythroid cells, resulting in dysfunction and hypoproteinemia. Destruction of RBCs releases heme that is converted to unconjugated bilirubin. Hyperbilirubinemia becomes apparent only in the delivered newborn because the placenta effectively metabolizes bilirubin. HDN due to Kell sensitization results in hemolysis and suppression of erythropoiesis because the Kell antigen is expressed on the surface of erythroid progenitors. This leads to severe fetal disease at a lower maternal antibody titer than in Rhesus disease.

Hemolysis associated with ABO incompatibility exclusively occurs in type-O mothers with fetuses who have type A or type B blood, although it has rarely been documented in type-A mothers with type-B infants with a high titer of anti-B IgG. In mothers with type A or type B, naturally occurring antibodies are of the IgM class and do not cross the placenta, whereas 1% of type-O mothers have a high titer of the antibodies of IgG class against both A and B. They cross the placenta and cause hemolysis in fetus.

Hemolysis due to anti-A is more common than hemolysis due to anti-B, and affected neonates usually have positive direct Coombs test results. However, hemolysis due to anti-B IgG can be severe and can lead to exchange transfusion. Because A and B antigens are widely expressed in various tissues besides RBCs, only a small portion of antibodies crossing the placenta are available to bind to fetal RBCs. Recent analysis of IgG subclass in ABO incompatible direct coombs positive neonates showed IgG2 was predominent antibody which is poorly transferred across placenta and less efficient in causing hemolysis while IgG1 was noted in 22% of neonates and as a group had similar rate of hemolysis and severity of hyperbilirubinemia. ^[9]

In addition, fetal RBCs appear to have less surface expression of A or B antigen, resulting in few reactive sites; hence the low incidence of significant hemolysis in affected neonates. This results in hyperbilirubinemia as a predominant manifestation of incompatibility (rather than anemia), and peripheral blood film frequently reveals a large number of spherocytes and few erythroblasts, unlike what is seen in Rh incompatibility (erythroblastosis fetalis), in which blood film reveals a large number of nucleated RBCs and few spherocytes. ^[10]

In the absence of a positive direct Coombs test result, other causes of pathologic jaundice should be considered, ^[11]  including intrauterine congenital infections; erythrocyte membrane defects (eg, hereditary spherocytosis, hereditary elliptocytosis, hereditary pyropoikilocytosis); RBC enzyme deficiencies (eg, glucose-6-phosphate dehydrogenase [G6PD] deficiency, pyruvate kinase deficiency, triosephosphate isomerase deficiency); and nonhemolytic causes (eg, enclosed hemorrhages, hypothyroidism, GI obstruction, and metabolic diseases).

Similarly, hydrops can occur from nonimmune hematologic disorders that cause anemia, such as hemoglobinopathies (eg, α-thalassemia major), cardiac failure due to dysrhythmia, congenital heart defects, and infections (eg, syphilis, cytomegalovirus [CMV], parvovirus ^[12] ).

• Common causes of hemolytic disease of the newborn

  • Rh system antibodies

  • ABO system antibodies

• Uncommon causes: Kell system antibodies

• Rare causes

  • Duffy system antibodies

  • MNS and s system antibodies

• No occurrence in hemolytic disease of the newborn

  • Lewis system antibodies

  • P system antibodies

United States data

Hemolytic disease of the fetus and newborn (HDFN) affects an estimated 3 in 100,000 to 80 in 100,000 patients annually. ^[13]

Before the establishment of modern therapy, 1% of all pregnant women developed Rh alloimmunization. Since the advent of routine prophylaxis of at-risk women, incidence of Rh sensitization has declined from 45 cases per 10,000 births to 10.2 cases per 10,000 total births, with less than 10% requiring intrauterine transfusion. ^[14]  Alloimmunization due to Kell antigen accounts for 10% of severely affected fetuses.

Currently, anti-D is still one of the most common antibodies found in pregnant women, followed by anti-K, anti-c, and anti-E. Of those fetuses who require intrauterine transfusions, 85%, 10%, and 3.5% were due to anti-D, anti-K, and anti-c, respectively. ^[15] ABO incompatibility frequently occurs during the first pregnancy and is present in approximately 12% of pregnancies, with evidence of fetal sensitization in 3% of live births. Less than 1% of births are associated with significant hemolysis.

Race- and sex-related demographics

Incompatibility involving Rh antigens (anti-D or anti-c) occurs in about 10% of all pregnancies among whites and blacks; in contrast, it is very rare in Asian women.

Fetal sex plays a significant role in the degree of response to maternal antibodies. An apparent 13-fold increase is observed in fetal hydrops in RhD-positive male fetuses compared with female fetuses in similarly sensitized pregnancies. ^[8]

Overall survival is 85-90% but reduced for hydropic fetuses by 15%. Most survivors of alloimmunized gestation are intact neurologically. Fetal hydrops does not seem to affect long-term outcome. ^[16] However, neurologic abnormality has been reported to be closely associated with severity of anemia and perinatal asphyxia. Sensorineural hearing loss may be slightly increased.

Morbidity/mortality

Almost 50 different red cell surface antigens have been found to be responsible for hemolytic disease of fetus and newborn. Only 3 antibodies are associated with severe fetal disease: anti-RhD, anti-Rhc, and anti-Kell(K1). Nearly 50% of the affected newborns do not require treatment, have mild anemia and hyperbilirubinemia at birth, and survive and develop normally. Approximately 25% are born near term but become extremely jaundiced without treatment and either die (90%) or become severely affected by kernicterus (10%). The remaining 25% of affected newborns are severely affected in utero and become hydropic; about half of newborns are affected before 34 weeks' gestation, and the other half are affected between 34 weeks' gestation and term. ^[1]

Before any interventions were available, the perinatal mortality rate was 50%. Wallerstein introduced exchange transfusion in 1945 and reduced the perinatal mortality rate to 25%. Later, Chown suggested the early delivery of those severely affected nonhydropic fetuses by 34 weeks' gestation followed by prompt exchange transfusion helped improve survival. The introduction of intraperitoneal transfusion by William Liley in 1963 and intravascular transfusion (IVT) by Rodeck in 1981 reduced the perinatal morbidity and the mortality rate was further reduced to the current rate of 16%.

Mortality rises to 30% with any degree of fetal hydrops. Most fetuses who are able to reverse fetal hydrops after IVT survive, compared with 25% of those in whom fetal hydrops was severe and persisted despite treatment. The overall rate of neurodevelopmental impairment is 10%, which is comparable to that found in the general population, but hearing loss is increased 5-10 fold over the general population in those infants who require in utero therapy for hemolytic disease of the newborn. The LOTUS study in a Dutch population reported neurodevelopmental outcomes in 281 children with hemolytic disease of the fetus treated with IVT at 8 years, showing normal outcome in 94%, cerebral palsy in 2.1%, severe developmental delay in 3.1%, and bilateral deafness in 1%. ^[17]  Severe hydrops fetalis was the only independent risk factor identified for poor outcome.

No relationship was noted between global developmental scores and the severity of hemolytic disease of the newborn (as evidenced by such factors as the number of intrauterine transfusions [IUTs], the lowest hematocrit [Hct] level, or the presence of hydrops). Normal neurological outcome is noted in more than 90% of infants even if fetal hydrops noted at the time of the first IUT. ^[18]

Exchange transfusion

A study by Smits-Wintjens et al indicated that exchange transfusion in neonates increases the risk of sepsis, severe thrombocytopenia, leukocytopenia, hypernatremia, and hypocalcemia in neonates with hemolytic disease of the newborn (HDN). The study involved 347 newborns with red cell alloimmune hemolytic disease, including 134 who underwent exchange transfusion and 213 who did not. The incidence of complications in the two groups was, respectively, as follows ^[19] :

• Proven sepsis: 8% versus 1%

• Severe thrombocytopenia: 63% versus 8%

• Leukocytopenia: 88% versus 23%

• Hypernatremia: 8% versus 0%

• Hypocalcemia: 22% versus 1%