Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
An important relationship between bilirubin and injury to the neonatal central nervous system (CNS) has been recognized for many years. The first comprehensive description of the most overt form of bilirubin encephalopathy (i.e., kernicterus) was provided by Schmorl in 1903. The development of therapeutic measures, such as exchange transfusion, and of preventive measures, such as the use of anti-Rh immune globulin to prevent maternal sensitization, resulted in a marked decrease in this overt form of bilirubin encephalopathy. Subsequently, implementation of predischarge universal screening coupled with the appropriate use of phototherapy and posthospital discharge follow-up has been shown to be an extremely important strategy in the prevention of severe neonatal hyperbilirubinemia. The total effect of these strategies on neonatal health care use remains to be defined.
In this chapter, normal bilirubin structure and metabolism, the pathophysiology of hyperbilirubinemia and bilirubin neurotoxicity, and the neuropathological and clinical aspects of bilirubin injury to the neonatal CNS are reviewed.
Central to an understanding of the relation of bilirubin to neonatal brain injury is an awareness of the normal aspects of bilirubin structure, properties, and metabolism in the newborn. Thus before the pathophysiology of bilirubin encephalopathies is discussed, certain highly relevant aspects of the chemical structure and solubility of bilirubin, as well as normal bilirubin metabolism, are briefly reviewed.
Bilirubin is a catabolic product of the porphyrin ring, derived from heme (see subsequent discussion). This compound can exist in plasma as bilirubin anion (monoanion or dianion form) ( Fig. 30.1A ) or as bilirubin acid (see Fig. 30.1B ). Bilirubin dianion binds actively to albumin. The dianion is not highly soluble in lipid or nonpolar solvents, but in view of the two polar carboxyl groups and oxidipyrryl (lactam) groups, it is not surprising that the anionic forms of bilirubin are relatively soluble in polar solvents. Although the anionic forms were considered previously to be the principal free bilirubin species in plasma, more recent work showed that at physiological pH the acceptance of two hydrogen ions results in bilirubin diacid (see Fig. 30.1B ). The diacid has a rigid folded structure, maintained by six internal hydrogen bonds involving all the polar groups, thereby rendering the molecule poorly soluble in aqueous solutions. However, the compound does passively diffuse across plasma membranes of cells (see later). When the concentration of free bilirubin diacid exceeds its limit of aqueous solubility (≈70 nM at physiological pH), the compound exists as soluble oligomers and metastable microaggregates. Moreover, at very high concentrations of the diacid, insoluble precipitates form and may result in major injury to membranes. At physiological pH , the species of free bilirubin are approximately 82% diacid, 16% monoanion, and 2% dianion. These chemical properties of bilirubin are important in understanding the neurotoxicity of bilirubin (see later discussion).
Normal bilirubin metabolism is considered best in terms of the following sequential events: (1) production, (2) transport, (3) hepatic uptake, (4) conjugation, (5) excretion, and (6) enterohepatic circulation ( Fig. 30.2 ).
Bilirubin is the end product of the catabolism of heme, the major source of which is circulating hemoglobin (see Fig. 30.2 ). In the newborn infant, the normal destruction of circulating red blood cells in the reticuloendothelial system accounts for approximately 75% of the daily production of bilirubin. The conversion of the heme moiety to bilirubin requires the sequential action of two enzymes, heme oxygenase (to form biliverdin), and a reduced nicotinamide adenine dinucleotide phosphate–dependent biliverdin reductase (to form bilirubin). Approximately 25% of the daily production of bilirubin in the newborn is derived from sources other than senescent red blood cells. This “other” fraction has two major components: a nonerythropoietic component, resulting from turnover of nonhemoglobin sources of heme (e.g., cytochromes, catalase, peroxidase, and myoglobin), and an erythropoietic component, resulting from destruction of products of ineffective erythropoiesis.
Bilirubin leaves the site of production in the reticuloendothelial system and is transported in plasma bound to albumin (see Fig. 30.2 ). Human albumin has a single, tight, high-affinity (or primary) binding site for bilirubin and one or more (probably two) weaker, lower-affinity binding sites. The capacity of serum albumin to bind bilirubin is known as the binding capacity , and the strength of the bilirubin-albumin bond is referred to as the binding affinity . The amount of free bilirubin (i.e., bilirubin not bound to albumin) is very low at physiological pH. These latter three characteristics of a given serum-binding capacity, binding affinity, and amount of free bilirubin can be estimated by in vitro measurements and provide a measure, albeit only approximate, of the amount of bilirubin that may be available to cause neuronal injury (see later discussion).
Hepatocytes have a selective and highly efficient system for removing unconjugated bilirubin from plasma. This mechanism requires several different organic anion transport proteins. Variants of one of these, organic anion transporter 2, may be important in determining the elevated risk of severe hyperbilirubinemia in Asian infants. In the hepatocyte, the transported bilirubin is bound to a cytosolic protein, ligandin, which facilitates transfer to the endoplasmic reticulum, the site of bilirubin conjugation.
Conversion of bilirubin to excretable monoconjugates and diconjugates is carried out primarily by the microsomal enzyme uridine diphosphate glucuronosyltransferase (UGT) 1A1. The protein is encoded by the UGT1A1 gene. Variants of the UGT1A1 gene underlie the unconjugated hyperbilirubinemia syndromes including Crigler-Najjar type I and II, and Gilbert syndrome. Gilbert syndrome variants are associated with breast milk jaundice, potentiate hyperbilirubinemia risk when coexpressed with hemolytic conditions, and appear to be important in determining the elevated risk of severe hyperbilirubinemia in infants of East Asian ancestry. Although the diconjugate normally accounts for approximately 90% of total bilirubin glucuronide conjugates, the monoconjugate predominates in Gilbert syndrome enhancing the enterohepatic circulation of bilirubin.
The conjugated bilirubin is excreted into the bile. Because this event occurs across a concentration gradient, an energy-dependent active transport system is involved. The conjugated bilirubin is then transported to the small intestine, is reduced by intestinal bacteria to urobilinogen, and is excreted in the stool.
Enterohepatic circulation of bilirubin also occurs. Intestinal beta-glucuronidase hydrolyzes the conjugated bilirubin, thus releasing unconjugated bilirubin, which then is reabsorbed and transported by the portal circulation to the liver increasing the total plasma bilirubin pool (see Fig. 30.2 ).
Although the relationship between serum levels of unconjugated bilirubin and neurotoxicity is not simple (see subsequent sections), a general link can be discerned between neonatal hyperbilirubinemia and the risk of neural injury. The major causes of neonatal hyperbilirubinemia, including the universal (“physiological”) elevation of bilirubin in the newborn, are reviewed next.
In the newborn period, numerous disorders may lead to elevated concentrations of unconjugated bilirubin ( Table 30.1 ). These include disorders with increased bilirubin production , principally hemolytic disease, secondary to maternal antierythrocyte antibody mediated hemolysis, intrinsic defects of red blood cells or unstable hemoglobin, degradation of extravascular blood (i.e., hemorrhage), polycythemia, and sepsis; disorders with disturbed gastrointestinal transit and therefore increased enterohepatic circulation of bilirubin; and disorders of bilirubin conjugation , including inherited and acquired defects (e.g., prematurity) and hormonal disturbances (e.g., hypothyroidism). Central to the frequency of hyperbilirubinemia in the newborn is the development of hyperbilirubinemia that occurs normally and onto which these other causes are grafted. This physiological jaundice of the newborn may be better termed developmental hyperbilirubinemia because it is very difficult to define precisely when this neonatal process is physiological and when it is pathological.
Increased production |
Hemolytic disease |
Immune-mediated: Rh, ABO, and minor blood group incompatibility |
Inherited RBC defects: RBC membranes (e.g., spherocytosis), hemoglobin (e.g., thalassemias), or RBC metabolism (e.g., glucose-6-phosphate dehydrogenase deficiency) |
Other |
Hematoma, including cerebral, or other extravasation of blood |
Polycythemia |
Sepsis b |
Infant of diabetic mother |
Increased enterohepatic circulation |
Breast-milk feeding |
Bowel obstruction |
Decreased conjugation |
Prematurity |
Uridine diphosphate glucuronosyltransferase deficiency (e.g., Crigler-Najjar, Gilbert syndrome) |
Hypothyroidism |
a Adapted from Maisels MJ. Jaundice. In MacDonald MG, Mullett MD, Seshia MMK, eds. Avery’s Neonatology Pathophysiology and Management of the Newborn . 6th ed. Lippincott Williams & Wilkins; 2005.
Developmental hyperbilirubinemia refers to the elevation of serum bilirubin values that occur in essentially every newborn infant in the first week of life. In classic earlier studies of formula-fed full-term infants, the serum bilirubin level rose gradually to a mean peak of approximately 6 mg/dL by 48 to 72 hours of life and then declined to a slightly elevated level, approximately 3 mg/dL, at approximately 5 days of life. Little change occurred thereafter until a second, gradual decline resulted in normal levels by approximately 2 weeks. With the current increase in breast-feeding in the United States and elsewhere, peak values for bilirubin clearly are greater than in the earlier studies. Thus in predominantly breast-fed term infants, the normal peak value is 8 to 11 mg/dL, and the decline is slower. In newborns of East Asian ancestry (predominantly breast-fed), the peak values, 10 to 14 mg/dL, are still higher. In the premature infant, the peak serum bilirubin concentration occurs later and is higher than in the full-term infant. More importantly, the term developmental hyperbilirubinemia in the premature infant is misleading because these higher values are potentially dangerous (see later) and phototherapy is recommended well before such levels are reached (see later).
The late preterm infant (i.e., 34 0/7 to 36 6/7 weeks of gestation) presents a situation intermediate between the clearly premature infant and the full-term infant (37 0/7 to 42 weeks of gestation). Thus several studies showed that such infants have a several-fold higher risk of significant hyperbilirubinemia in the first week of life. In a careful study through the first 7 days of life, late preterm infants exhibited a later and higher peak value of bilirubin than did term infants. Moreover, the higher values declined more slowly in the late-preterm infants. Overall, 10% of the full-term infants had hyperbilirubinemia requiring phototherapy, versus 25% in the late-preterm group. These data highlight the importance of particularly close follow-up of late-preterm infants after hospital discharge (see later).
The mechanisms considered important in the genesis of developmental hyperbilirubinemia are shown in Table 30.2 . Although evidence has been mustered for all these factors, studies of the newborn monkey provided considerable insight into their relative importance. Thus the first phase of hyperbilirubinemia (see previous section) has been shown to relate to the combined effects of (1) increased bilirubin load to the liver and (2) decreased bilirubin-conjugating capacity. The source of the increased bilirubin load to the liver includes both hemoglobin and nonhemoglobin sources, as well as the enterohepatic circulation of bilirubin; the latter is increased in the newborn because of deficient bacterial reduction of bilirubin and increased activity of intestinal beta-glucuronidase. The defective bilirubin conjugating capacity is related to a negligible level of hepatic UGT1A1 activity at birth, which undergoes a rapid increase during the first week of postnatal life but does not approach adult levels for several months. Prematurity results in more severe neonatal hyperbilirubinemia, principally because of a delayed maturation of the hepatic UGT1A1. More recent studies in human infants refined the earlier observations. Thus the possibility of impaired uptake of bilirubin into the hepatocyte because of genetic variants of organic anion transporters (see earlier) is likely important, perhaps especially in infants of East Asian ancestry. The diminished levels of ligandin in hepatic cytosol are likely less important. Genetic variants of the UGT1A1 gene, as well as the long-recognized developmental deficiency of the glucuronyl transferase, are also important. Indeed, the prolonged jaundice associated with breast milk feedings is now considered a prevalent Gilbert syndrome phenotype. Breast milk contains several substances (beta-glucuronidase, lipoprotein lipase, pregnanediol) that augment the enterohepatic circulation bilirubin, which combined with Gilbert syndrome leads to prolonged jaundice .
Increased bilirubin production |
↑ RBC volume |
↓ RBC survival |
↑ “Other” sources |
Increased enterohepatic circulation |
Decreased hepatic uptake of bilirubin from plasma |
↓ Membrane transport |
↓ Ligandin |
Defective bilirubin conjugation |
↓ Uridine-diphosphate-glucuronyl transferase |
The critical event in the genesis of brain injury caused by bilirubin is entrance of bilirubin into brain and exposure to neurons. The predominance of available data indicates that bilirubin per se and more specifically the bilirubin no longer bound to albumin (i.e., unbound or “free” bilirubin ), is the form that ultimately leads to neuronal injury (see later discussion).
To derive the important determinants of neuronal injury by bilirubin, it is necessary to recognize the critical reactions shown in Fig. 30.3 . As noted earlier, at physiological pH the predominant species of unbound bilirubin in plasma is bilirubin acid. Consideration of these equilibria makes it clear that the potential means for increasing exposure of neurons to bilirubin acid would include increasing the quantity of bilirubin anion–albumin (i.e., unconjugated bilirubin ) and especially thereby unbound or free bilirubin, disturbing the binding of bilirubin anion to albumin, decreasing the quantity of albumin that is free to bind bilirubin, and increasing the quantity of hydrogen ions (i.e., lowering pH) ( Table 30.3 ). Additional factors that interrelate closely with those just enumerated include the status of the blood-brain barrier and the susceptibility of target neurons to bilirubin injury; these factors are considered separately in later sections. In the following discussion of the determinants of bilirubin neurotoxicity, it becomes clear that the importance of each factor must vary with the clinical circumstances (see Table 30.3 ).
|
|
|
|
|
|
Although it is generally recognized that serum levels of unconjugated bilirubin must be elevated to cause neurotoxicity, the relationship between such elevations and brain injury is not simple. In the full-term infant with marked hyperbilirubinemia secondary to hemolytic disease , a correlation can be discerned between the occurrence of kernicterus and the maximal recorded level of serum bilirubin ( Table 30.4 ). In a review of 52 infants with hemolytic disease (33 Rh incompatibility, 19 ABO incompatibility) and comorbid factors, approximately 95% of whom developed kernicterus, the mean peak total serum bilirubin (TSB) in both groups was approximately 32 mg/dL , with an approximate range of 18 to 51 mg/dL .
MAXIMUM BILIRUBIN CONCENTRATION (mg/dl) | TOTAL NO. OF CASES | NO. WITH KERNICTERUS |
---|---|---|
30–40 | 11 | 8 (73%) |
25–29 | 12 | 4 (33%) |
19–24 | 13 | 1 (8%) |
10–18 | 24 | 0 |
The neural risk of marked hyperbilirubinemia in full-term infants without hemolysis is less clear. Indeed, an earlier analysis of available studies by Newman and Maisels suggested that the risk for neurological sequelae is distinctly less for hyperbilirubinemic infants without hemolytic disease compared with the risk for infants with hemolytic disease. Subsequent observations supported this contention, including a study by Newman and associates of 140 treated infants (phototherapy, n = 136) and exchange transfusion ( n = 5) with peak serum bilirubin levels largely between 25 and 29.9 mg/dL. There were no cases of kernicterus. Similarly, Gamaleldin and colleagues reported that bilirubin encephalopathy was only observed at a TSB greater than 31.5 mg/dL in infants ( n = 111) without evidence of hemolytic disease, sepsis, and low gestational age. However, multiple reports describe the occurrence of kernicterus, identified by neuropathological, neuroradiological, or clinical criteria, after apparent nonhemolytic hyperbilirubinemia. Furthermore in a review of 35 selected infants with nonhemolytic, “idiopathic” hyperbilirubinemia and documented acute or chronic bilirubin encephalopathy, all infants had peak TSB levels greater than 20 mg/dL. Indeed, more than 90% of the infants with kernicterus had peak levels higher than 25 mg/dL. Notably, however, fully 25% had peak levels lower than 30 mg/dL. Nevertheless, current observations suggest that hemolytic conditions are often overlooked in the absence of detailed search for immune-mediated mechanisms by advanced techniques. It seems reasonable to conclude that hyperbilirubinemia, in most cases, is a necessary but usually not a sufficient condition to explain kernicterus. Factors acting in concert with bilirubin, including duration of exposure to bilirubin or albumin binding of bilirubin, must be evaluated to seek a satisfactory explanation for the risk of developing kernicterus (see later). These issues are discussed in detail later (see ‘Relationship of Neurological Sequelae in Term Infants to Degree of Hyperbilirubinemia’ section). The relationships between neonatal bilirubin values and neurological outcome in premature infants are probably different in magnitude but not in principle from those in full-term infants. Indeed, kernicterus has been demonstrated repeatedly in the premature infant without marked hyperbilirubinemia (see later discussion). These infants usually have exhibited a variety of complicating conditions (e.g., acidosis, hyperbilirubinemia, sepsis, asphyxia, hypoalbuminemia, hypothermia, intraventricular hemorrhage) that likely augment bilirubin neurotoxicity. In one often-cited, relatively large collection of premature infants with kernicterus ( n = 6), peak TSB levels were in excess of 20 mg/dL (range, 22 to 26 mg/dL). However, in this report, the gestational age of the six infants ranged from 34 to 36 weeks. Studies of bilirubin-induced auditory disturbances in smaller premature infants (28 to 32 weeks of gestational age) document neurological dysfunction at much lower bilirubin levels. The critical issue of the premature infant is discussed later (see ‘Clinical Features’ section).
Because of the recognition that the fraction of bilirubin not bound to albumin is the critical component involved in bilirubin entry into the brain and in neurotoxicity (see later) and because of the demonstration of kernicterus in premature infants without markedly elevated levels of unconjugated bilirubin, intensive investigation has been directed toward measurement of the quantity of that fraction of unconjugated serum bilirubin, namely, unbound (i.e., “free”) bilirubin. In general, premature infants with higher levels of free bilirubin exhibit kernicterus at postmortem examination more often than those with lower levels. In a recent study of 1100 extremely low-birth weight infants in whom total and unbound bilirubin levels were measured at 5 days, irrespective of the clinical status, an increasing level of unbound bilirubin was associated with higher rates of death or neurodevelopmental impairment, or hearing loss at 18 to 22 month follow-up. Disturbances of the brainstem auditory-evoked response (BSAER) are observed at lower levels of free bilirubin in premature versus full-term infants (see later). Experimental studies also support a relationship between free or unbound bilirubin and neuronal injury.
However, there is currently little agreement about what constitutes the neurotoxic threshold for unbound bilirubin and what values of unbound circulating bilirubin should be used for initiating treatment. Indeed, other determinants of risk (e.g., concerning albumin binding of bilirubin, status of the blood-brain barrier, and neuronal susceptibility; see Table 30.3 ) are often present in the sick premature infant and contribute to the likelihood of bilirubin injury (see subsequent sections). Notably, the measurement of circulating unbound bilirubin is generally not clinically available.
In contrast, measuring the total bilirubin/albumin molar ratio (BAMR) is clinically obtainable and has been shown to be highly correlated with unbound bilirubin. Although the BAMR is an imperfect surrogate of free bilirubin and CNS bilirubin exposure, during extreme hypoalbuminemia the BAMR may become a meaningful proxy of bilirubin neurotoxicity risk as exemplified by the phenomenon of low-bilirubin kernicterus. In the low-bilirubin kernicterus case series of Govaert and colleagues, serum albumin levels ranged from 1.3 to 1.9 g/dL and every neonate demonstrated an elevated BAMR that met or exceeded recommended BAMR exchange transfusion treatment thresholds.
Consideration of the equilibria among albumin-bound bilirubin anion, albumin, hydrogen ion, and bilirubin acid (see Fig. 30.3 ) makes it apparent that the concentration of serum albumin is important in determining the neurotoxicity of bilirubin (see Table 30.3 ). At lower concentrations of serum albumin, the overall reaction favors formation of unbound bilirubin anion and, ultimately, bilirubin acid. Indeed, in experimental systems, the toxic effects of bilirubin on enzymatic systems or on cultured cells of neural origin can be reversed by the addition of albumin. Moreover, evidence indicates that infants at the greatest risk for kernicterus in the absence of marked hyperbilirubinemia (i.e., sick premature infants) usually exhibit concentrations of serum albumin that are lower than in healthy premature and full-term infants. Indeed, in one study of 27 premature infants with kernicterus, serum albumin levels were statistically significantly lower than in a comparable control group. However, the latter observation has not been entirely consistent, in keeping with the importance of such other factors as the bilirubin binding affinity and binding capacity of albumin (see following discussion).
The capacity of serum albumin to bind bilirubin depends on such factors as the affinity of the albumin for bilirubin and the competition between bilirubin and other endogenous and exogenous anions for albumin binding sites. The clinical importance of the ability of albumin to bind bilirubin is emphasized by the demonstrations that premature infants who develop kernicterus without marked hyperbilirubinemia may have disturbances of the affinity or capacity, or both, of albumin to bind bilirubin. Similarly, a relationship between bilirubin-albumin binding and subsequent cognitive outcome of premature infants who required neonatal intensive care further supports the clinical importance of this binding.
The affinity of albumin for bilirubin is less in the newborn than in the older infant. Adult levels of binding affinity are not reached until as late as 5 months of age. Moreover, binding affinity is lower in the premature infant than in the term infant and is lower in sick infants than in well infants. The explanation for the lower binding affinity of neonatal albumin is not entirely clear. The search for competing anions or for compositional differences in the protein as the unifying explanation has not been fruitful. The leading possibility is that a difference in conformation of the albumin is responsible. Moreover, it is likely that the conformational difference relates to the humoral environment of the neonatal albumin because adult serum albumin infused into newborns loses its superior binding affinity over 24 hours.
Endogenous anions that may compete with bilirubin for albumin binding sites include nonesterified fatty acids and other organic anions. Nonesterified fatty acids are anions at physiological pH and are present in high concentrations with hypothermia, hypoxemia, hypoglycemia, sepsis, starvation, the administration of heparinized blood (through heparin’s activation of triglyceride lipase), and intravenous alimentation with lipid. A recent report describes the uncoupling of unbound from total bilirubin as a result of unbound free fatty acids in premature infants treated with intralipids. Studies of asphyxiated infants with metabolic acidosis demonstrated impaired bilirubin binding to albumin that could not be attributed to the low pH per se. Rather, the data indicated the presence of organic anions in the plasma of asphyxiated, acidotic infants that compete with bilirubin for albumin binding sites. This finding may represent one of several mechanisms of potential importance in enhancing bilirubin neurotoxicity with asphyxia ( Table 30.5 ).
|
|
|
|
|
a Probable mechanisms are in parentheses; see text for references.
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here