Imbalance between Bilirubin Production and Elimination

The total serum bilirubin (TSB) concentration at any point in time reflects a multiplicity of interactions leading to two major processes contributing to this value: bilirubin production and bilirubin elimination.1 As long as these processes remain in equilibrium, the TSB should remain within normal limits. During the first days of life, because of physiologically increased heme catabolism in combination with diminished activity of the bilirubin conjugating enzyme, UDP-glucuronosyltransferase 1A1 (UGT1A1), there is an imbalance between these processes and bilirubin levels increase. In the majority of cases, this imbalance should remain mild or moderate, and TSB concentrations should not exceed the 95th percentile on the hour-of-life-specific bilirubin nomogram.2 When the rate of bilirubin production exceeds elimination (the latter process dependent primarily on conjugation), hyperbilirubinemia occurs. Although in many (perhaps most) cases, hyperbilirubinemia is associated with some degree of increased heme catabolism, severe hemolysis is not essential to the process. More important is the concept of lack of equilibrium between bilirubin production and conjugation. Thus, a baby may be hemolyzing and producing large amounts of bilirubin, but because hepatic bilirubin conjugating capacity is mature, the infant may not develop an increased TSB. On the other hand, a baby with minimally increased hemolysis, but with immature bilirubin conjugating capacity (due to late prematurity or the presence of the (TA)7 repeat UGT1A1 gene promoter polymorphism, associated with Gilbert syndrome), may develop hyperbilirubinemia. This concept of equilibrium, or lack thereof, between bilirubin production and conjugation has been demonstrated mathematically. Kaplan et al. studied the individual contributions of bilirubin production and conjugation to the TSB concentration, as well as the combined effects of these processes in healthy, term neonates on the third day of life.3 The rate of heme catabolism was indexed by measurements of blood carboxyhemoglobin (COHb) determinations, corrected for ambient carbon monoxide (COHbc), while bilirubin conjugation was assessed by total serum conjugated bilirubin (TCB) expressed as a percentage of TSB (TCB(%)). Over the range of TSB concentrations observed, TSB correlated with both increasing COHbc levels and diminishing TCB(%) values (Figure 8-1A and B). The COHbc and TCB(%) values were then used to construct an index or ratio, COHbc/TCB(%), to reflect the combined forces of these processes. The correlation of this “production–conjugation index” with increasing TSB concentrations was higher than that for either COHbc or TCB(%), independently (Figure 8-2). Thus, the concept of imbalance and interaction between bilirubin production and conjugation, rather than individual or independent processes, in the mechanism of neonatal bilirubinemia was confirmed. Although the relationship between the index and TSB tended to plateau with increasing index values, at the lower end of the index scale, which included the majority of readings, small increases in the index were associated with large increases in TSB.

Possible Increased Risk Associated with Hemolysis

It is generally believed that neonates with hemolytic disease are at a higher risk of developing bilirubin-induced neurotoxicity than those whose hyperbilirubinemia is not the result of hemolysis.4–6 Data emanating from a few studies are supportive of this concept. In a Turkish study, a positive direct antiglobulin test (DAT or direct Coombs’ test), due to Rh isoimmunization or ABO incompatibility, was used as a presumed marker of hemolysis.7 Of 102 children aged 8–13 years with indirect hyperbilirubinemia ranging from 17 to 48 mg/dL, DAT positivity was associated with lower IQ scores and a higher incidence of neurologic abnormalities than in controls without a positive test. The incidence of detected neurologic abnormalities also increased with increasing duration of exposure to high TSB levels. In DAT-positive Norwegian males born in the early 1960s who had TSB levels >15 mg/dL for longer than 5 days, IQ scores were significantly lower than average for that population.8 In the Jaundice and Infant Feeding Study, 5-year outcomes of infants with TSB >25 mg/dL were not significantly different from randomly selected controls. However, in the subgroup (n = 9) of infants with TSB >25 mg/dL as well as a positive DAT, IQ values were significantly lower than their counterparts with a negative DAT (-17.8 IQ points, 95% confidence interval [CI] -26.8 to -8.8).9 In a reanalysis of the data from the Collaborative Perinatal Project, Kuzniewicz et al. found no relationship between maximum TSB levels and IQ scores.10 But in the presence of a positive DAT, a TSB of ≥25 mg/dL was associated with a 6.7 point decrease in IQ scores. An increase in the duration of exposure to high TSB levels was also associated with an increase in neurologic abnormalities. Finally, in a recent study of 249 newborns admitted to a children’s hospital in Cairo, Egypt, with TSB values ≥25 mg/dL, Gamaleldin et al. documented little relationship to the admission TSB and the presence or severity of acute bilirubin encephalopathy.11 However, there was a marked difference in the development of bilirubin encephalopathy between those with additional risk factors, including Rh incompatibility, ABO incompatibility, and sepsis, compared with those without evidence of risk factors. The threshold TSB in identifying 90% of babies with bilirubin encephalopathy was 25.4 mg/dL in the presence of risk factors, but as high as 31.5 mg/dL in those without risk factors. Furthermore, the neurotoxicity-producing effect was higher in those with Rh incompatibility (odds ratio [OR] 48.6, 95% CI 14, 168) than in those with ABO incompatibility (OR 1.8, 95% CI 0.8, 4.5). Although the DAT testing in that situation was unreliable, the results do suggest that it is not simply the presence of hemolysis but also the etiology of the hemolysis that mediates the neurotoxicity.

The exact mechanism of the effect of hemolysis in increasing the risk of bilirubin neurotoxicity is unknown. The unbound bilirubin fraction, thought to be capable of crossing the blood brain barrier and entering the basal ganglia, correlates better than TSB with neurotoxicity.12–14 If hemolysis was instrumental in increasing the risk of neurotoxicity, babies exhibiting hemolysis should have a higher unbound bilirubin fraction, but to date this has not been demonstrated.

While anti-D Rh isoimmunized babies are nowadays rarely encountered, DAT-positive neonates, primarily due to ABO incompatibility, as well as other blood antibodies, such as anti-c, anti-E, and others, are still seen. Glucose-6-phosphate dehydrogenase (G6PD) deficiency is a major cause of nonimmune severe hemolytic hyperbilirubinemia. The importance of hemolytic conditions in the pathophysiology of bilirubin encephalopathy can be seen in the report of the US-based Pilot Kernicterus Registry.15 Of 125 newborns, 31 (25%) were blood group A or B infants born to O mothers. The expected incidence of these blood group combinations is only 15% (see section “Hemolytic Disease of the Newborn Caused by ABO Heterospecificity”). Eight (6.4%) of these 31 had a positive DAT test, but this figure may be an underestimation as the status in 7 (5.6%) was unknown. The incidence of G6PD deficiency was also overrepresented in the Registry: while 26 (20.8%) of the Kernicterus Registry newborns were G6PD deficient, the expected male incidence in the United States is only 4–7%.16 Whereas a TSB concentration of 20–24 mg/dL may be associated with kernicterus in a neonate with Rh isoimmunization, in the absence of a hemolytic condition, a healthy, term infant will rarely be endangered by TSB concentrations in this range.

Are We Missing Hemolyzing Babies?

Because hemolytic conditions appear to contribute to the development of bilirubin neurotoxicity to a greater extent than those not overtly hemolytic, the American Academy of Pediatrics (AAP) emphasizes the identification of the hemolyzing newborn.17 Some authors have used the term “nonhemolytic jaundice” to differentiate newborns with no apparent hemolytic cause for the jaundice from those with a hemolytic etiology, but this may be an oversimplification of the issue.18–23 Just as presence of a hemolytic condition, such as G6PD deficiency or DAT-positive ABO blood group heterospecificity, does not categorically imply that the jaundice or hyperbilirubinemia is necessarily due to this condition, absence of an identifiable etiology does not necessarily imply that increased hemolysis is not integral to the pathophysiology of the jaundice. Unfortunately, there is no readily available bedside clinical tool to determine the rate of hemolysis. Blood count indices such as decreasing hemoglobin or hematocrit values, or increased reticulocyte count, which may be useful to identify hemolysis in adults, often display overlap between hemolytic and nonhemolytic states in the newborn and may be unreliable indicators of hemolysis.24 Clearly, in some newborns the hepatic component may be the primary contributor to hyperbilirubinemia and there may be very little increased hemolysis. However, studies utilizing the endogenous production of CO, an accurate index of heme catabolism, have demonstrated that many jaundiced babies do, in fact, have a hemolytic component to their jaundice, even in the absence of a defined hemolytic condition.

In a multicenter, multinational study, an automated device that sampled end-tidal expired air via a nasal catheter to determine the CO concentration corrected for ambient CO (ETCOc) was used.25 Measurements of both ETCOc and TSB were performed at 30 ± 6 hours of life; TSB also was measured at 96 ± 12 hours, and subsequently if determined according to a flow diagram. Mean (±SD) ETCOc value for 1370 infants who completed the study was 1.48 ± 0.49 ppm. The 120 newborns who developed any TSB concentration >95th percentile for hour of life (hyperbilirubinemia) had significantly higher ETCOc values than those who did not (1.81 ± 0.59 ppm vs. 1.45 ± 0.47 ppm, P < .0001). However, high bilirubin production was not a prerequisite for the development of hyperbilirubinemia: some babies with low bilirubin production did, nevertheless, develop hyperbilirubinemia, while others with high production rates did not, confirming the combined, rather than individual, contribution of bilirubin production and elimination to the TSB. Using the identical technology, Maisels and Kring found that in 108 newborns ≥36 weeks gestational age who had TSB concentrations >75th percentile, ETCOc values were significantly higher through the first 4 days of life than in 164 neonates with lower TSB levels. Furthermore, while ETCOc values decreased progressively during the study period in the control infants, values increased in those with TSB values >75th percentile.26 While the increase in ETCOc values in the study infants could not be explained, these authors concluded that increased hemolysis was an important mechanism in the production of jaundice in the first 4 days of life even in the absence of a specific diagnosis of a condition associated with increased hemolysis (Figure 8-3). In a group of African American neonates who were not G6PD deficient, ETCOc values of 27 newborns who developed a TSB concentration >95th hour-specific percentile were significantly higher than 335 neonates whose TSB did not exceed that value (2.6 [2.33–3.45] ppm vs. 2.00 [1.70–2.40] ppm) (median [interquartile range]).27 Similarly, in another group of newborns, from which both G6PD-deficient and DAT-positive, ABO blood group–incompatible infants had been excluded, higher COHbc values (0.75 ± 0.18%) were noted in newborns who developed any TSB value >15.0 mg/dL during the first week of life than counterparts with lower TSB values (0.53 ± 0.13%, P < .05).28

Based on the above, it does appear that many hyperbilirubinemic newborns, even in the absence of an obvious hemolytic condition, have some degree of increased heme catabolism. Because universal blood typing with DAT determination and G6PD screening are not universally recommended, and ETCOc testing is currently unavailable as a clinical tool, it is possible that some infants not actually identified as having a hemolytic condition may, in fact, be actively hemolyzing. The term “nonhemolytic jaundice” may not be valid in all cases to which it has been attributed and the increased risk for hyperbilirubinemia and bilirubin encephalopathy attributed to increased hemolysis may go unrecognized in some. Absence of documented hemolysis or presence of an obvious hemolytic etiology should not exclude the potential danger of bilirubin neurotoxicity.

AAP Recommendations Regarding Hemolyzing Babies

The Subcommittee on Hyperbilirubinemia of the AAP includes jaundice developing within the first 24 hours, blood group incompatibility with a positive DAT, and other known hemolytic disease including G6PD deficiency as major risk factors for the development of severe hyperbilirubinemia.29 The Subcommittee recommends a more aggressive approach to hyperbilirubinemia, by initiating phototherapy or performing exchange transfusions at lower levels of TSB in neonates with neurotoxicity risk factors including isoimmune hemolytic disease and G6PD deficiency.

In broad terms, hemolytic conditions encountered in neonates may be divided into two major pathophysiologic groups: immune and nonimmune. A classification of these disorders appears in Table 8-1. The former group includes Rh isoimmunization, which, nowadays, is largely preventable and only occasionally encountered in the western world. The condition will, nevertheless, be discussed in some detail, as much of our knowledge regarding the development of kernicterus derives from the study of babies with this condition. As ABO immune disease (mother blood group O, baby group A or B) is the most common immune condition currently encountered, it will also be discussed in detail. Of the nonimmune causes, G6PD deficiency is by far the most important from a public health standpoint, and is associated with the development of unpredictable, extreme hyperbilirubinemia, and acute bilirubin encephalopathy.

Immune Hemolytic Conditions

The Direct Antiglobulin (Agglutination) Test

The DAT, otherwise known as the direct Coombs’ test, is the hallmark of isoimmunization.30 A positive DAT test is indicative of maternally derived IgG antibody directed against and bound to the fetal (and subsequently newborn) RBC antigenic sites, the result of incompatibility between maternal and fetal blood types. The antiglobulin reaches the fetal tissues via the placenta. In the direct form of the test, the antiglobulin is attached to the fetal or newborn erythrocyte, whereas an indirect test relates to the antibody being found in the serum. The DAT detects the presence of an antibody, but is nonspecific and does not identify the specific type of antibody present: for this purpose, further identifying tests are required. DAT is measured by incubating a newborn’s blood sample with anti-IgG antiserum. If the RBCs are coated with IgG, these cells will agglutinate. The clumps may be identified visually or microscopically. The DAT is usually measured qualitatively, a stronger test (i.e., the more positive) suggesting a greater amount of antibody present. While a positive DAT is associated with increased hemolysis and hyperbilirubinemia, this may not always be the case. Herschel et al., using a 12-hour ETCOc value ≥95th percentile as a reference for increased hemolysis, found the sensitivity of the DAT in a group of primarily ABO blood group–incompatible newborns to be 38.5% and the specificity 98.5% for the detection of significant hemolysis, while the positive predictive value (PPV) of the DAT for significant hemolysis was 58.8%. In other words, a neonate with a positive DAT had about a 59% chance of having significant hemolysis.31 Neither, in this study, was a positive DAT universally predictive of clinically significant hyperbilirubinemia: only 9/61 (14.8%) of the DAT-positive neonates developed a TSB ≥75th percentile for age in hours. Ozolek et al. found the same general relationship between DAT and hyperbilirubinemia/jaundice in a large cohort of ABO heterospecific mother–infant pairs.32

Rh Isoimmune Hemolytic Disease

Rh disease in pregnancy may lead to intrauterine hemolysis, which if untreated may result in severe intrauterine anemia, hydrops fetalis, and intrauterine death. Continuation of the hemolytic process following delivery may result in severe hemolytic disease of the newborn (HDN) with the rapid development of anemia and hyperbilirubinemia with the potential of developing bilirubin encephalopathy early in neonatal life.

Background: The Immunization Process

The Rh group comprises the C, c, D, E, and e antigens, one transmitted from each parent to determine the Rh type. While each of these antigens may result in isoimmunization and hemolysis, RhD isoimmunization is the most common, encountered in 90% of Rh group cases, and is therefore of major clinical significance.

The distribution of Rh negativity and isoimmunization varies with racial background and paternal heterozygosity. While in white populations about 13–15% of individuals are Rh negative, only about half that number are encountered in African Americans, and Rh negativity is very unusual in Asian individuals. Because 55% of fathers are heterozygous (D/d), an Rh-negative mother may have an Rh-negative fetus in about 50% of pregnancies. Additional factors influencing the incidence of Rh disease include nonuniversal fetomaternal transmission of fetal blood and variable maternal immune responses. Therefore, in mothers who did not receive immune prophylaxis, the overall incidence of Rh isoimmunization is actually infrequent and reported to be 6.8 cases per 1000 live births.33

The Rh D immunization process begins when a D-negative woman is exposed to the D antigen. Antepartum or intrapartum exposure occurs as a result of transplacental, fetomaternal passage of fetal RBCs containing the D antigen. Because most transfer of fetal blood occurs late in pregnancy or during delivery, in first pregnancies the process usually begins too late to allow for sufficient maternal IgG antibody to be produced and transferred across the placenta. Nevertheless, immunization during the first pregnancy is a well-documented phenomenon34 and not as uncommon as generally believed.35 Isoimmunization during the first pregnancy and/or in multiparous Rh D-negative women who begin a pregnancy without detectable Rh antibodies occurs in ˜1.8% of such women35 and infants from such pregnancies can be affected; about one in five such babies will require treatment including possible need for exchange transfusion.34 Transfusion of Rh-positive RBCs may also occur during abortion, blood administration, amniocentesis, chorionic villus sampling, or fetal blood sampling. In response to the antigen exposure, the mother’s immune system may respond by the formation of anti-D IgG antibodies. This antiglobulin may then cross the placenta to the fetus, adhere to the D antigen sites of the fetal RBCs, result in an antigen–antibody response, and culminate in hemolysis and anemia. The mother’s immune system is now primed and with subsequent pregnancies the immune response may become progressively more severe and have a more rapid onset. Should the anemia be prolonged and sufficiently severe, the bone marrow may be stimulated to release increased numbers of circulating immature RBCs (erythroblastosis). With further progression of the process, extramedullary hematopoiesis with hepatomegaly and splenomegaly may ensue. In its most severe stages, fetal hydrops, which may include generalized tissue edema and pleural, pericardial, and peritoneal effusions, may result. The pathogenesis of this extravascular fluid includes hypoproteinemia, tissue hypoxia, and capillary leak, combined with congestive cardiac failure. The latter may develop due to anemia and venous congestion, the result of poor myocardial function and diminished cardiac output.36

Management of the Pregnancy

Hydrops fetalis is associated with a high mortality rate and major efforts should be made to prevent its occurrence by appropriate intervention prior to its appearance. Pregnancies complicated by Rh isoimmunization require active surveillance to detect fetal anemia. Should the fetus become anemic, the option to perform intrauterine transfusion must be weighed against delivery. This decision will depend primarily on the gestational age: with increasing maturity, the potential for inherent complications involved in preterm delivery will decrease relative to the dangers involved in performing intrauterine transfusion. Until recently, amniocentesis was the primary method of fetal monitoring.37 The degree of hemolysis was assessed by determining the amount of bilirubin in the amniotic fluid by measuring the deviation from linearity at 450 nm, the wavelength at which bilirubin absorbs light. Measurements were divided into three zones and plotted on the Liley chart. Readings in zone III indicated a high level of danger and were indicative of severe hemolysis with a high likelihood of fetal death. Queenan et al. improved on the Liley chart by developing the Queenan curve.38 This amniocentesis-based regimen, however, has been largely replaced by a combination of advancing ultrasonographic techniques and developing genetic technologies, as described below.

In the absence of prophylaxis, 14% of RhD-negative women who deliver an Rh-positive baby can be expected to develop anti-D antibodies within 6 months of delivery or during the following pregnancy.39 In the preprophylaxis days, 14% of affected pregnancies could be expected to result in stillbirths and 30% of those live-born to have severe hemolytic disease.40 An additional 30% may have more moderate disease but, if untreated, may progress to severe hyperbilirubinemia and encephalopathy. The introduction of postpartum prophylaxis reduced the isoimmunization rate to 1.8%.41,42 Subsequently, it was found that administration of rhesus immune globulin during pregnancy to all nonimmunized, RhD-negative women reduced the incidence of antenatal immunization to 0.1%.43 Prophylaxis with anti-D globulin is now routine and all pregnant women should have an antibody screen early in pregnancy. Should the woman be RhD-negative with no evidence of anti-D alloimmunization, rhesus immune globulin should be administered at 28 weeks gestation.36 The globulin should also be administered following spontaneous or elective abortion, amniocentesis, chorionic villus sampling, or fetal blood sampling. Should the woman deliver an RhD-positive infant, a repeat dose should be administered within 72 hours of delivery. As a result of this prophylactic regimen, in most industrialized countries, hemolytic disease due to RhD immunization has been almost completely eradicated. This situation should not be taken for granted in developing countries with underdeveloped health systems. Zipursky and Paul estimated that in India, Pakistan, and Nigeria where the majority of RhD-negative women do not receive postpartum anti-D prophylaxis, thousands of women annually will develop anti-RhD antibodies, and about half the babies born to these women will develop Rh hemolytic disease.39 This scenario is most likely encountered in many other developing countries where it is likely that as many as 100,000 children may be born annually with RhD hemolytic disease.39

Should a pregnant woman be found to be Rh immunized, because this condition is nowadays rare and therapeutic technologies are advancing at a rapid rate, she should be managed in a tertiary center capable of adequately managing a severely affected fetus as well as the newborn infant.

Advances in RhD gene technology have allowed for fetuses either at high or low risk to be identified. The RhD gene is located on the short arm of chromosome 1.44 Approximately 55% of individuals are heterozygous at the RhD locus and 50% of their fetuses will be RhD positive. Should a fetus be RhD negative, no further testing will be required. In the past, gene frequency tables, combined with the history of RhD-positive or -negative infants fathered by any individual, were used to estimate the likelihood of a specific father being heterozygous. Modern advances in DNA technologies, however, allow for accurate determination of whether the father is heterozygous or homozygous for the RhD gene.45,46 Should the father be heterozygous, steps should be taken to determine the Rh type of the fetus by retrieving fetal DNA via amniocentesis.42 A major technological development that may enable the noninvasive determination of the fetal RhD type includes cell-free fetal DNA determination in a maternal blood sample.45,47,48

Determination of the maternal anti-D titer is an important step in the monitoring of an RhD-sensitized woman. A critical titer is that which is associated with a high risk of severe hydrops. The critical titer varies from center to center and ranges from 8 to 32 are usually used to predict disease.36

Doppler assessment of the blood flow velocity in the fetus’ middle cerebral artery is replacing amniocentesis in the detection of fetal anemia. Fetal anemia results in increased blood flow velocity due to decreased blood viscosity and increased cardiac output. In one study, a value of more than 1.5 multiples of the median identified all cases of moderate to severe anemia.49 In a study comparing diagnostic amniocentesis with middle cerebral artery flow in the detection of anemia, Doppler measurements improved on optical density determinations by 9%.50 If the results of either of these techniques suggest anemia, fetal blood is sampled by cordocentesis and the hematocrit, DAT, blood type, reticulocyte count, and TSB determined. If the hematocrit is <30% and the fetus <35 weeks gestation, intrauterine transfusion is considered. If the pregnancy has reached 35 weeks gestation or more, the advantages of delivery will generally outweigh the dangers of an intrauterine transfusion. Repeated intrauterine transfusion may cause fetal bone marrow suppression, and in a repeatedly transfused fetus, the RBC mass at the time of delivery may be composed almost entirely of donor cells. In this situation, hemolysis will be minimal and exchange transfusion may be unnecessary, although “top-up” transfusions for subsequent anemia may be required.51 In experienced hands, the outcome of intrauterine transfusion should be good. In the Netherlands, the survival rate in 254 fetuses was 89%.52

Postnatal Management of the Newborn

The clinical manifestations of erythroblastosis fetalis in the neonate range from an asymptomatic infant with laboratory evidence of anemia to severe hydrops, extreme anemia, and cardiac decompensation. Management of a hydropic, severely anemic neonate, especially if complicated by respiratory and other problems of prematurity, is a major neonatal challenge requiring a degree of expertise available only at a tertiary center.53 Hydropic infants have generalized edema and fluid collections in the pleural, pericardial, and peritoneal spaces. Intubation may be challenging because of oral soft tissue swelling. Pleural and pericardial fluid collections impair ventilation and may require emergency drainage to facilitate adequate respiratory gas exchange. Respiratory support may have to be complemented by surfactant, nitric oxide, and high-frequency ventilation, while anoxic myocardial damage may necessitate use of inotropes. Metabolic acidosis may further complicate the situation54 and hypoglycemia is common.53

A cord blood sample should be obtained immediately after delivery. Hemoglobin values <10 g/dL and TSB >5 mg/dL suggest severe hemolysis. Many of these babies will subsequently require exchange transfusion. Because of the potential of circulatory system overload in combination with poor myocardial function, it may be preferable to correct the anemia isovolumetrically, using a partial exchange transfusion technique via umbilical venous or arterial catheter rather than by simple blood transfusion. Small amounts of anemic blood should be drawn from the newborn and replaced with packed cell donor blood. The hematocrit may be “titrated” until the desired concentration is obtained. Phlebotomy should not be performed routinely on these infants because they are usually normovolemic and may even be hypovolemic55–57 and their blood volume should not be manipulated without appropriate measurements of central venous pressure (CVP) and arterial blood pressure. In order to measure CVP accurately, the umbilical venous catheter must enter the inferior vena cava via the ductus venosus. If the catheter is in a portal vein or the umbilical vein, the pressures so measured are meaningless and preclude interpretation of the infant’s circulatory status. In addition, before making therapeutic decisions based on measurements of CVP, acidosis, hypercarbia, hypoxia, and anemia (all of which can affect the measured CVP) must be corrected. Serum glucose levels must be monitored carefully because hypoglycemia is common.

In the fetus, most bilirubin formed from the hemolytic process will be eliminated via the placenta and severe intrauterine hyperbilirubinemia is rarely a problem, although the cord blood bilirubin level may be elevated to 3–5 mg/dL. Once delivered, however, the placenta no longer participates in the bilirubin elimination process. Continued hemolysis combined with immature conjugative and excretory function may potentiate rapidly rising TSB values with the danger, if untreated, of developing extreme hyperbilirubinemia. As soon as the baby has been stabilized, the TSB is measured and intensive phototherapy started. Should the TSB continue to rise despite intensive phototherapy, exchange transfusion will be necessary according to the thresholds recommended by the 2004 AAP guideline.

Randomized controlled trials (RCT) have shown that the administration of intravenous immune globulin (IVIG) is effective in preventing or limiting the number of exchange transfusions in Rh disease.58–62 The 2004 AAP guideline also recommends the use of IVIG in order to prevent exchange transfusion in cases of failing phototherapy.29 Most recently, however, a Dutch RCT found no benefit from the prophylactic administration of IVIG to infants with Rh hemolytic disease63 and a Cochrane report concluded that more information, based on well-designed studies, was needed before IVIG could be recommended for the treatment of isoimmune hemolytic disease.64 The reason for these conflicting results is not clear but could be related to the type of IVIG used. Even if an exchange transfusion will ultimately be necessary, delay of this procedure by IVIG can be useful in gaining time for stabilization of the patient.

The AAP guideline provides clear criteria for the use of phototherapy and exchange transfusion in these infants.29

There is limited information on the long-term neurodevelopmental outcome of fetuses treated with intrauterine transfusion, although, in general, the results to date are encouraging.54,65 A system for surveillance of children for long-term neurodevelopment outcome following intrauterine transfusion has been established in Holland and should provide relevant data within the next few years.66

Hemolytic Disease of the Newborn Caused by ABO Heterospecificity

Definition of ABO Blood Group Heterospecificity

By the term “ABO setup,” we refer to a blood group A or B baby born to a group O mother. This combination is seen in about 15% of pregnancies. About one third of these pregnancies (5%) will have a positive DAT, indicative of anti-A or -B antibodies attached to RBCs.32

ABO Blood Group Heterospecificity: The Most Frequent Cause of Immune Hemolytic Disease in the Neonate

In the past, the high incidence and severity of hemolytic disease due to Rh isoimmunization obscured the clinical manifestations of DAT-positive ABO heterospecificity. Because Rh hemolytic disease is currently seen only occasionally in the western world, ABO incompatibility has become the most frequent cause of immune hemolytic disease in the neonate. Although ABO heterospecific disease is usually milder than that encountered in Rh-immunized fetuses and newborns, the hyperbilirubinemia may at times be severe and associated with bilirubin encephalopathy. Of 125 babies reported in the US Kernicterus Registry, 31 (25%) were of blood group A or B born to group O mothers.15 The expected incidence of these combinations in the United States is only about 15%. Eight (6.5%) were DAT positive, but this figure may be an underestimate since in some of the infants, the DAT status was not identified. Between 2002 and 2004, the Canadian Paediatric Surveillance Program identified 258 jaundiced newborns who had either a TSB ≥25 mg/dL or undergone exchange transfusion. A cause for the jaundice was found in 93 (36%). ABO heterospecificity was the most frequent diagnosis among these infants, occurring in 32 of the 93 (34.4%).67 In the United Kingdom and Ireland, from 2003 to 2005, 108 newborns were identified with a TSB concentration ≥30 mg/dL. Thirty-three (30.6%) were ABO incompatible, of whom 16 (15%) were DAT positive.68 Similarly, in Denmark, between 2002 and 2005, 113 infants ≥35 weeks gestation were identified with a TSB value ≥26 mg/dL. Fifty-two (48.2%) comprised blood group A or B infants born to group O mothers of whom 16 (31%) were DAT positive (14% of total).69 In Nigeria, of 115 babies affected with bilirubin encephalopathy between 2001 and 2005, 42 (36.5%) died. Twenty-two (19.1%) were ABO incompatible (DAT status not provided).70 In a Swiss nationwide study conducted between 2007 and 2008 and encompassing 146,288 infants ≥35 weeks gestation, 60 newborns exceeded the Swiss indications for exchange transfusion.71 In 31 of these, a diagnosis was determined and the majority (17 [54.8%]) were ABO incompatible.

Unlike Rh disease, hydrops fetalis in ABO immune disease is rare, although a few cases have been reported.72–74 The milder disease that characterizes ABO hemolysis might occur because, in the fetus, A and B antigens are not limited to the RBCs, but are found in other fetal tissues including the placenta and in body fluids. These antigens neutralize and dilute transferred maternal antibody, thereby reducing the number of antibodies available to attach to the fetal RBCs. Furthermore, the newborn RBC has fewer A or B reactive sites than the adult RBC that explains the weakly reactive DAT seen in ABO hemolytic disease.75

Background: Blood Group Incompatibility and DAT Positivity

The connection between ABO heterospecificity and neonatal jaundice was established in the 1940s. Of infants who were jaundiced in the first 24 hours postdelivery, and excluding cases of Rh hemolytic disease, ABO incompatibility was found in 95%.76 Shortened RBC survival was demonstrated in affected infants compared with survival of transfused group O cells77,78 and in 1961, Kochwa et al. demonstrated that anti-A and -B antibodies found in maternal or fetal serum were IgG antibodies and capable of crossing the placenta.79

Pathophysiology of ABO Hemolytic Disease

The pathophysiology of ABO hemolytic disease differs from that of Rh isoimmunization. In Rh disease, following the initial pregnancy-related immunization, the antibody titer generally increases with each subsequent pregnancy. ABO disease does not follow an initial immunizing pregnancy. In contrast, some women with type O blood have an inherently high titer of anti-A or -B antibodies that is unrelated to the fetomaternal passage of blood and can be found before their first pregnancy (or even in young girls).80 Furthermore, group O individuals differ from their blood group A or B counterparts in that their anti-A or -B antibodies comprise IgG molecules, able to cross the placenta, whereas the respective antibodies of blood group A or B individuals are predominantly IgM, larger, and therefore prevented from placental passage. When an affected woman becomes pregnant, the IgG molecules may cross the placenta and attach to the corresponding A or B antigens on the fetal RBCs. This process of isoimmunization then precipitates hemolysis in utero. Extravascular hemolysis of the IgG-coated RBCs is probably mediated by Fc-receptor-bearing cells within the reticuloendothelial system. There is little danger of severe hyperbilirubinemia occurring prior to delivery as the immune process is not usually sufficiently strong and the placenta is able to clear most of the resulting bilirubin. Some infants are born with moderate anemia and continuation of the hemolytic process following delivery can produce severe hyperbilirubinemia.

ABO Setup versus ABO Hemolytic Disease

ABO heterospecificity, even if the DAT is positive, does not necessarily indicate the presence of ABO hemolytic disease. Many of these neonates may have no evidence of ongoing hemolysis and may not develop early jaundice or significant hyperbilirubinemia. Some or all of the following criteria are necessary to support the diagnosis of ABO hemolytic disease:

1. 
   

   Mother blood group O and baby group A or B

   
   
2. 
   

   Positive DAT

   
   
3. 
   

   Indirect hyperbilirubinemia, especially during the first 24 hours of life

   
   
4. 
   

   Microspherocytosis on peripheral blood smear

   
   
5. 
   

   Increased reticulocyte count

Measurement of the endogenous production of CO will help to quantify hemolysis, but, unfortunately, the instrument for determining end-tidal CO levels is no longer available.

Studies of Hemolysis Incorporating Endogenous Formation of Carbon Monoxide

Measurements of endogenous formation of CO, an index of heme catabolism, have demonstrated an increased rate of heme catabolism in DAT-positive, ABO-incompatible neonates compared with controls, confirming the role of increased hemolysis in the pathophysiology of hyperbilirubinemia in these neonates. Fällström and Bjure found that in 48/62 ABO-incompatible infants with significant jaundice, COHb values exceeded the mean + 2SD for values in healthy, nonicteric newborns.81 Uetani et al. found that in DAT-positive, ABO-incompatible neonates with TSB values >15 mg/dL, COHb values were higher than those in nonhyperbilirubinemic controls, and that, in the hyperbilirubinemic newborns, COHb levels remained high during the first 5 days of testing.82 Interestingly, not all DAT-positive neonates had increased COHb levels. In a multinational, multicenter study in which ETCOc was used to detect hemolysis at 30 ± 6 hours, values were higher in 54 DAT-positive babies compared with the total number (1370) of babies studied (1.66 ± 0.55 ppm vs. 1.48 ± 0.49 ppm, P = .009).25 DAT-positive neonates with TSB >95th percentile had even higher, although not statistically significant, ETCOc values (1.89 ± 0.63). However, only 18.5% of the DAT-positive newborns developed a TSB >95th percentile, implying that, despite the increased hemolysis, many babies were able to handle the increased bilirubin load. Herschel et al., also using ETCOc, documented higher values for DAT-positive, ABO-incompatible neonates (3.4 ± 1.8 ppm, n = 14) than for DAT-negative counterparts or ABO-compatible controls (2.2 ± 0.6 ppm, n = 60 and 2.1 ± 0.6 ppm, n = 171, respectively, P = .02).83 Using COHbc determinations to assess the degree of hemolysis, Kaplan et al. found that overall, DAT-positive neonates had higher COHbc values than previously published for a DAT-negative reference group (1.24 ± 0.40 ppm, n = 163 vs. 0.77 ± 0.19 ppm, n = 131, P < .0001).84 Furthermore, those who developed hyperbilirubinemia (TSB >95th percentile) had higher COHbc values than the already high values of those whose TSB did not exceed the 95th percentile (1.42 ± 0.39 ppm, n = 85 vs. 1.00 ± 0.25 ppm, n = 78, P < .001). There was a trend for O–B newborns to have higher COHbc values than their O–A counterparts (1.32 ± 0.44 ppm vs. 1.20 ± 0.38 ppm, P = .07). Finally, the percentage of newborns who developed hyperbilirubinemia increased in tandem with COHbc, reaching 100% in those with COHbc values >90th percentile (Figure 8-4).

Incidence of Hyperbilirubinemia in ABO Heterospecific Neonates

Despite the increased hemolysis documented in DAT-positive ABO heterospecific newborns and the association with bilirubin encephalopathy, not all affected neonates will develop severe or clinically significant hyperbilirubinemia. In the prephototherapy era (1969–1971), Kanto et al. found that only 26 (11.3%) of 230 ABO-incompatible, DAT-positive newborns born in Augusta, Georgia, developed TSB values >12 mg/dL. In some cases, however, this concentration did reach as high as 23 mg/dL.85 In Norway, Meberg and Johansen included TSB testing at the time of routine predischarge metabolic screening. Of 17 of 2463 newborns with documented TSB values >20 mg/dL, only 1 was ABO heterospecific.86 Furthermore, only 19.6% of 92 DAT-positive, ABO heterospecific neonates met the indications for phototherapy. In Pittsburgh, Pennsylvania, of 531 DAT-positive, blood group A or B newborns delivered to group O mothers, only 73 (13.7%) developed TSB concentrations >12.8 mg/dL.32 Using a recently used definition of hyperbilirubinemia (i.e., any TSB value >95th percentile on the hour-of-life-specific bilirubin nomogram2,25,87–89), Kaplan et al. in Jerusalem, Israel, found a high incidence of hyperbilirubinemia, 85 (51.8%) of 164, among DAT-positive, ABO heterospecific neonates.84 Importantly, 56 newborns (34.1% of the cohort or 66.7% of those with hyperbilirubinemia) had their first TSB >95th percentile documented earlier than 24 hours, while 80 neonates (49%) met the 2004 AAP indications for phototherapy.29 The incidence of hyperbilirubinemia was substantially higher than that of some other studies, which utilized the identical definition of hyperbilirubinemia,25,88,89 emphasizing the important role of ABO heterospecificity in the etiology of neonatal hyperbilirubinemia. The fact that all babies responded to phototherapy, none required exchange transfusions, and there were no cases of kernicterus should not detract from the importance of this finding: DAT positivity, TSB values >95th percentile, and hyperbilirubinemia occurring within the first 24 hours are each listed by the AAP as major risk factors for the development of severe hyperbilirubinemia.29 Any additional process adding to the bilirubin load, such as further hemolysis or immaturity of the bilirubin conjugating system, may upset the equilibrium and, if untreated, lead to severe hyperbilirubinemia and encephalopathy. These newborns should therefore be vigilantly observed with timely institution of therapy when indicated.

The literature is inconsistent regarding the incidence and severity of hyperbilirubinemia between O–A and O–B groups. Several investigators were unable to show any difference in clinical severity between O–A and O–B HDN.85,90–94 On the other hand, other authors have reported an increased need for exchange transfusion or IVIG therapy in O–B neonates, compared with O–A counterparts.95–97 Kaplan et al. reported a trend showing an increased rate of hyperbilirubinemia among O–B infants compared with O–A infants (33/53 [62.3%] vs. 52/111 [46.8%], relative risk [RR] 1.34; 95% CI 0.99–1.77; P = .053).84

ABO blood group incompatibility with a negative DAT, not usually predictive of hemolysis or hyperbilirubinemia, may sometimes cause early and rapidly progressing jaundice, reminiscent of DAT-positive hemolytic disease. Some of these infants may be displaying a manifestation of coexpression between ABO incompatibility and homozygosity for the (TA)7 UGT1A1 polymorphism.98 This concept is discussed in section “Genetic Interactions Between Hemolytic Conditions and UGT1A1 Polymorphisms.”

Predicting the Severity of Hyperbilirubinemia in ABO Disease

Most blood group A or B neonates born to blood group O mothers will not develop any sign of hemolytic disease. Routine blood group and DAT determination on umbilical cord blood is an option, but not mandatory. It is essential to closely observe any newborn born to a blood group O mother for the development of jaundice and to perform a TSB when indicated. Attempts have been made to predict subsequent hyperbilirubinemia in ABO-incompatible neonates based on TSB concentrations. Sarici et al. found a mean TSB of ≥4 mg/dL at the sixth hour of life had a sensitivity of 86.2%, a negative predictive value (NPV) of 94.5%, and a PPV of 39.7% in the prediction of hyperbilirubinemia. The definition of hyperbilirubinemia varied with the infant’s age.99 Using a mean TSB of 6 mg/dL at age 6 hours improved the prediction, the sensitivity, specificity, NPV, and PPV being 100%, 91.5%, 100%, and 35.3%, respectively, in diagnosing cases of severe ABO hemolytic disease. In the epoch prior to introduction of phototherapy in their unit, Risemberg et al. utilized umbilical cord blood TSB values to predict hyperbilirubinemia in ABO-incompatible neonates. Thirteen of 91 infants (14%) developed a TSB concentration of 16 mg/dL or more at 12–36 hours and underwent exchange transfusion. All newborns except one with severe hyperbilirubinemia had cord TSB >4 mg/dL. Similarly, all 12 newborns with cord TSB >4 mg/dL developed hyperbilirubinemia and required exchange transfusion.100

Treatment of ABO Heterospecificity– Associated Neonatal Jaundice

A high degree of vigilance is necessary to detect developing jaundice in newborns born to blood group O mothers. TSB or transcutaneous bilirubin (TcB) should be measured if jaundice is seen in the first 12–24 hours.2 Phototherapy and exchange transfusions are implemented according to the 2004 AAP guideline.28 IVIG may be helpful in modifying the rate of rise of bilirubin and is indicated if the TSB is approaching the exchange transfusion threshold despite a trial of intensive phototherapy.29 In an analysis of 4 randomized trials involving 226 babies affected with Rh and ABO immune disease,58–61 IVIG in combination with phototherapy significantly reduced the need for exchange transfusions compared with phototherapy alone (RR 0.28; 95% CI 0.17–0.47).62 In newborns who responded to IVIG administration with a decrease in TSB, COHbc values decreased from baseline in tandem with diminishing TSB concentrations, demonstrating the effect of IVIG on limiting heme catabolism.101 The effect most likely takes place by blocking Fc receptors in the reticuloendothelial system, IVIG competing with antibody attached to RBCs to prevent further hemolysis.102 We have occasionally encountered ABO-incompatible newborns with severe, early onset hyperbilirubinemia. Although we have not performed an exchange transfusion in these neonates for many years, we have found IVIG therapy useful in neonates not responding to intense phototherapy and approaching the indications for exchange transfusion.