In the human fetus, as in the adult, biliverdin-IXα and any small amounts of non-IXα isomers that are formed are reduced to the corresponding bilirubins. Of these, bilirubin-IXα is uniquely hydrophobic and lipophilic, and ready to cross the placenta for elimination by the mother. In utero, residual non-IXα isomers too polar to cross the placenta, particularly the IXβ isomer, accumulate and are detectable in bile and meconium by 15 weeks gestation.1 This observation has led some to conclude erroneously that heme catabolism in the fetus yields predominantly the IXβ isomer. The major form of bilirubin generated in infants and adults alike is bilirubin-IXα, and can be measured in various forms (Table 3-1).

The poor solubility of bilirubin can be explained by considering its chemical three-dimensional structure.2 Although often represented as a linear structure for convenience (Figure 3-1, structure 5), bilirubin has a folded flexible structure in which the weakly acidic propionic acid side chains can stretch and form internal hydrogen bonds with spatially proximate nitrogen and oxygen groups. This results in a compact structure in which the surface is lipophilic and the polar parts of the molecule are protected from interactions with solvent water. The stereochemical configuration of the two double bonds between the rings in bilirubin is the same as in heme from which the bilirubin was derived and is designated unambiguously in current organic chemistry nomenclature as Z (from zusammen, German: together) (in contradistinction to E [entgegen: opposite], the other possible configuration). Because of its low solubility in water at physiologic pH, bilirubin requires a carrier molecule for transport from the reticuloendothelial system to the liver for excretion.3 In blood and extravascular fluid, bilirubin is bound noncovalently to albumin that possesses a single high-affinity binding site (Ka = 7 × 107/M) for bilirubin as well as secondary binding sites of lower affinity.4 Due to the strong binding affinity and relatively high concentration of albumin at physiologic levels, most bilirubin is bound to albumin and the relative concentration of free, unbound bilirubin is negligible.

Total Serum Bilirubin Measurement

From a historical perspective, recognition of the complexity of bilirubin has evolved over the past century (Figure 3-2). Measurement of the total serum bilirubin (TSB) concentration allows quantification of jaundice. TSB measurements are commonplace in the newborn nursery, and in one study were made at least once in 61% of full-term newborns5 although the frequency may be decreasing with advances in transcutaneous bilirubinometry (TcB). Most clinical laboratories use a modified diazo reaction to measure TSB. In this assay, bilirubin in the sample is exposed to the diazo reagent.6 The conjugated bilirubin in the sample reacts quickly, or directly, with the diazo reagent, hence the term direct bilirubin. Meanwhile, the TSB level can only be measured after the addition of an accelerant in the diazo method. The unconjugated, or indirect, fraction is then calculated by determining the difference between the total and direct reacting fractions. Using the diazo method, the normal value for unconjugated bilirubin in adults is less than 1 mg/dL and that for conjugated bilirubin is less than 0.3 mg/dL.7 Any elevation of conjugated bilirubin should be considered potentially pathologic, although the majority of such elevations in the neonate resolve spontaneously and are unexplained.8

In conjugated hyperbilirubinemia, the accumulated glucuronides undergo hydrolysis and rearrangement to a large number of isomers and also undergo spontaneous (i.e., nonenzymatic) transesterification with amino groups on albumin. This leads to covalent amide linkages between one of the propionic acid side chains of bilirubin glucuronide and albumin.9 The resulting bilirubin–protein complex is known as “delta-bilirubin.”10 It is impossible to present a single structure for delta-bilirubin since it is not a homogeneous substance but rather a mixture of compounds and because the location(s) of the covalent binding of bilirubin glucuronide to albumin is not certain. Delta-bilirubin is not formed in the absence of an elevated conjugated bilirubin fraction. Similar nonenzymatic reactions have been demonstrated between albumin and various drugs.9,11–13

The presence of delta-bilirubin explains the clinical conundrum of a persistently elevated TSB level in the infant who has experienced a clinical recovery from cholestatic jaundice. The direct-reacting fraction of bilirubin includes both conjugated bilirubin and delta-bilirubin. The terms direct and indirect are often used interchangeably with conjugated and unconjugated bilirubin, respectively, in the clinical setting. However, this is inaccurate from a biochemical standpoint. Direct bilirubin measurements include both conjugated and albumin-bound delta-bilirubin.14 Because of the strong covalent bond, the half-life of delta-bilirubin approximates that of albumin (∼3 weeks).15 Recognition and understanding of the difference between direct and conjugated bilirubin measurements is critical to prevent the unnecessary evaluation of an infant recovering from a hepatic insult, who continues to have a high delta-bilirubin and, therefore, an elevated direct bilirubin measurement. Conjugated bilirubin measurement is an earlier indicator of recovery from biliary cholestasis compared with direct bilirubin measurement because of the long half-life of delta-bilirubin.16 Due to the more timely response, conjugated bilirubin measurement is preferred over direct bilirubin measurement for the neonate.17

There is a long history of undesirable variability in the measurement of TSB fractions.18,19 One problem is the ditaurobilirubin surrogate standard provided by the College of American Pathologists (CAP),18 which influences measurement of TSB fractions variably because of protein matrix differences related to the specific bilirubin measurement used. This has prompted the suggestion that standards consisting of human serum enriched solely with unconjugated bilirubin rather than bovine serum containing a mixture of unconjugated bilirubin and ditaurobilirubin should be used.20

The automated laboratory methods now used to measure TSB have been reviewed elsewhere.14,21–23 The Jendrassik–Grof procedure has been suggested as the method of choice for TSB measurement, although this method has limitations.24 When the TSB level is high, factitious elevation of the direct fraction has been reported.25 Three newer methods have been developed that can more accurately determine the various bilirubin fractions (unconjugated, monoconjugated, diconjugated, and delta-bilirubin; Table 3-1): high-performance liquid chromatography (HPLC),26 multilayered slides,27,28 and use of bilirubin oxidase.29 HPLC analysis is the superior method and is considered by some to be the “gold standard,”30,31 but is too expensive and time consuming for the clinical laboratory.14 Furthermore, there are no published data of interlaboratory comparisons of HPLC analysis as there are with the other laboratory methods. HPLC analysis of serum from healthy neonates showed that unconjugated and conjugated bilirubin levels rose in parallel in the first 4 days of life with the conjugated fraction making up only 1.2–1.6% of total pigment.32 Although the absolute concentration of conjugates was two to six times higher in neonates compared with adults, only 20% were diconjugates (54% in adults). These sensitive HPLC data are consistent with the increased bilirubin production and relatively deficient glucuronidation seen in the neonate. Analysis with automated multilayered slide technology (Vitros, Ortho Clinical Diagnostics, Raritan, NJ) used in many clinical laboratories allows measurement of conjugated and unconjugated bilirubin fractions without inclusion of delta-bilirubin. Analysis of the bilirubin oxidase method concluded that measurement of neonatal TSB was not advanced by this method.33

Newer methods of TSB measurement (Twin Beam, Ginevri, Rome, Italy; ABL 735, Radiometer, Copenhagen, Denmark; Roche OMNI S, Roche Diagnostics, Graz, Austria) using nonenzymatic photometric analysis offer the convenience of bilirubin quantitation outside of the clinical laboratory setting (i.e., blood gas analyzer in the nursery) with a smaller blood sample than traditional serum analysis. However, these measurements must be interpreted with caution as the instruments tend to underestimate the TSB concentration at levels >15 mg/dL.34

There are conflicting data regarding the accuracy of capillary versus venous TSB levels.35,36 However, the literature regarding kernicterus, phototherapy, and exchange transfusion is based on bilirubin measurement from capillary samples, and treatment should not be delayed for a confirmatory measurement from a venous specimen.37 While it is acknowledged that TSB may not be the most important factor related to neurotoxicity, there is not an alternative laboratory test that is broadly accepted and widely available that better identifies infants at risk for bilirubin-induced neurologic dysfunction (BIND).

Clinical laboratories that perform newborn bilirubin testing are required to perform proficiency testing. The proficiency testing program offered through the CAP is the most commonly used program by laboratories in the United States. The 2010 neonatal bilirubin, cycle C (NB-C) and 2010 chemistry, cycle B (C-B) participant summary reports from CAP assessed the different methods for measurement of neonatal bilirubin and TSB concentrations currently in use (see Table 3-2).38,39 For the survey of neonatal TSB measurements, approximately 2000 laboratories participated in the 2010 NB-C survey. Each laboratory received five unknown samples for measurement of neonatal TSB, ranging in concentration from 10.2 to 23.4 mg/dL. There were 9 different methods used to measure bilirubin on 17 different instrument platforms. The most common method used was a diazo-caffeine/benzoate coupling with a blank (26%) or without a blank (23%). A spectrophotometric method was used by 24%, and an additional 17% of laboratories used a diazonium salt/diazonium ion method. The samples used in the CAP NB-C survey contained high concentrations of TSB because accurate measurements of increased bilirubin concentrations are critical in the neonatal period.

The CAP chemistry survey was taken by clinical laboratories that perform measurement of TSB (not limited to the neonatal population). Approximately 6200 laboratories completed the chemistry survey for proficiency testing. Like the neonatal survey described previously, each laboratory received five unknown samples for measurement of TSB. The laboratories that participated in the CAP C-B survey used 15 different methods on 22 different platforms. Bilirubin concentrations in the five different samples ranged from 0.7 to 4.3 mg/dL. Table 3-2 summarizes the bilirubin concentrations measured by all instruments and methods in samples from the 2010 NB-C and 2010 C-B proficiency testing surveys.

Transcutaneous Bilirubinometry

Jaundice presents in the face and progresses in a cephalocaudal pattern.40 Kramer determined the range of serum unconjugated bilirubin levels as jaundice progressed: head and neck (4–8 mg/dL), upper trunk (5–12 mg/dL), lower trunk and thighs (8–16 mg/dL), arms and lower legs (11–18 mg/dL), and finally the palms and soles (>15 mg/dL).41 However, visual assessment is not reliable among providers and patient populations. Darker skin tones can make jaundice difficult to assess visually.42 Heel stick is the most common means to collect a specimen for TSB measurement. In addition to being traumatic to the newborn and parents, heel stick is a suboptimal technique due to the potential for hemolysis and resultant interference in bilirubin measurement depending on the laboratory method used.43 An inexpensive, noninvasive method useful in assessing jaundice utilizes a Plexiglas color chart pressed against the baby’s nose (Ingram icterometer, Thos. A. Ingram and Co Ltd, Birmingham, England).44–46 However, today other noninvasive methods, specifically TcB, to assess jaundice have become routine practice in the newborn nursery as a screening method more sensitive than visual assessment for jaundice.47

There are two commercially available transcutaneous instruments approved by the Food and Drug Administration (FDA) to measure TcB in neonates: the Philips Children’s Medical Ventures BiliChek (Respironics, Inc, Murrysville, PA)30,31 and the Konica Minolta Air-Shields Transcutaneous Jaundice Meter 103 (JM-103, Draeger Medical Systems, Inc, Telford, PA).48 The BiliChek utilizes principles of reflectance spectrophotometry, has been validated against both HPLC and clinical laboratory measures,30,31 and has been advocated by the American Academy of Pediatrics (AAP).42 BiliChek uses light from multiple wavelengths, captures the light reflected, and corrects for differences in melanin, hemoglobin, and other interfering factors (Figure 3-3). The device touches the skin in a painless manner and provides an almost immediate point-of-care measurement of TcB. The JM-103 measures the difference in the optical densities of reflected light at 450 and 550 nm. Two optical paths are incorporated into a probe that minimizes interference by melanin, and the reflected light is collected by photodiodes. One limitation of the two devices currently available on the market is their inability to interface with the neonate’s electronic medical record leaving documentation of the TcB reading subject to human error inherent with manual entry.43,49

TcB measurements have been incorporated into clinical practice as an alternative to heel stick for screening jaundice in the newborn nursery. This has the potential to reduce the incidence of heel stick by 40–60%50–52 and to reduce potential serious, albeit rare, complications of blood collection, including infection and osteomyelitis.53 Serum bilirubin testing may be done more frequently since the introduction of TcB testing potentially as a result of more neonates being monitored for hyperbilirubinemia after identification of clinically significant jaundice by TcB.54 However, use of TcB measurements decreases TSB measurements and saves money.55 The combination of peak predischarge TcB with two clinical risk factors for pathologic hyperbilirubinemia, exclusive breastfeeding and gestational age, improves prediction of subsequent hyperbilirubinemia meriting treatment.55 Significant TcB levels (>14 mg/dL) should prompt measurement of a TSB level for confirmation.42,56–58