Because at some point during the first week after birth almost every newborn has a total serum bilirubin (TSB) level that exceeds 1 mg/dL (17 μmol/L), the upper limit of normal for an adult, and ∼2 of every 3 newborns are jaundiced to the clinician's eye, this type of transient bilirubinemia has been called “physiologic jaundice.” When TSB levels exceed a certain value, the infant is often described as having “pathologic jaundice.” I would like to argue that these terms have limited utility and are often used incorrectly, with potentially damaging consequences. They should be abandoned.
PHYSIOLOGIC JAUNDICE AND “NORMAL” BILIRUBIN LEVELS
The physiology of the newborn differs from that of older children and adults in many ways. Newborns breathe 40 to 60 times per minute, and their hearts beat 120 to 160 times per minute. Their hematocrit levels are frequently >60%. In time, all of these values return to normal levels, as does their bilirubin level. However, we don't talk about physiologic tachypnea, tachycardia, or polycythemia, so why pick on jaundice?
Some like the term “physiologic jaundice” because it has a reassuring sound for parents and physicians. Presumably, physiologic jaundice should apply to newborns whose TSB levels fall within a certain range, but what is that range? Because very few (if any) newborns have peak TSB levels <2 mg/dL, should an infant with a peak TSB of 1.5 mg/dL be considered abnormal or hypobilirubinemic? Unlike serum sodium levels, the range of normal TSB levels varies widely depending on the racial composition of the population, the incidence of breastfeeding, and other genetic and epidemiologic factors.1 There are also significant variations between different laboratories in their measurements of serum bilirubin.2
Term, healthy, North American, formula-fed infants have mean peak TSB levels between 5 and 6 mg/dL (86 and 103 μmol/L),3,4 whereas Japanese breastfed newborns have levels more than twice as high.5 Data from the Collaborative Perinatal Project,6 conducted from 1955 to 1961 (when 30% or fewer mothers breastfed their infants), indicated that ∼95% of infants had a TSB concentration that did not exceed 12.9 mg/dL (221 μmol/L), and this (95th percentile) became a commonly accepted upper limit of physiologic jaundice. It is interesting that in our hospital's predominately white (73%) and breastfed (80%–85%) population, the 95th percentile at age 96 hours is 13.1 mg/dL (224 μmol/L),7 whereas studies of infants in Philadelphia, Pennsylvania,8 and Northern California9 show the 95th percentile to be 17.5 mg/dL (299 μmol/L). In a mixed population of infants from the United States, Hong Kong, Japan, and Israel,10 the 95th percentile was 15.5 mg/dL (265 μmol/L). Thus, unlike the serum sodium or many other biochemical and hematologic measurements, defining a normal serum bilirubin level in the newborn is a problem.
Even if we accept that one, or an average, of these values represents the true 95th percentile (as if values >95th percentile are, by definition, abnormal; they are not), it is not uncommon to see a discharge diagnosis for an infant with a bilirubin level of 21 mg/dL (359 μmol/L) listed as “exaggerated physiologic jaundice.” If the jaundice is physiologic, one wonders why a battery of blood tests is performed and why the infant is in the hospital receiving phototherapy. Others call this TSB level pathologic jaundice, but more often than not, the battery of tests yields no identifiable pathology.11,12
Finally, we have an additional problem that is unique to bilirubin levels: they change almost every hour for ∼1 week or more, so meaningful interpretation of TSB levels can only be made in relationship to the infant's age in hours.13
DEFINING A NORMAL VALUE
How do we define the term “normal?” This question is elegantly discussed by Sackett et al14 in their text on clinical epidemiology, and the definition depends on why we are asking the question.
Gaussian or Percentiles
Perhaps the most commonly used definitions of normal in medicine are based on the mean ± 2 SDs (assuming a normal or Gaussian distribution). When the distribution is not Gaussian, percentiles have been used. If we know the characteristics and ranges of values in the population being studied, then percentile values can help to identify infants who need additional investigation for the cause of their jaundice or more careful follow-up because they are at risk for severe hyperbilirubinemia.15 The general use of either the Gaussian or percentile definitions of normal, however, is open to criticism. As noted by Sackett et al,14 if the highest and lowest 2.5% of diagnostic tests are called abnormal (if we use the mean ± 2 SDs) or if those above the 95th percentile or below the 5th percentile are called abnormal, then all diseases have the same frequency—clearly an illogical conclusion. Although it does not represent the natural history of neonatal bilirubinemia, the nomogram developed by Bhutani et al8 has given us a very useful tool for identifying infants who need additional evaluation and more careful follow-up. This nomogram also reemphasizes a long known but oft-forgotten fact: the TSB is changing continually and can only (logically) be interpreted in relationship to the infant's age in hours and not days.
There are other ways of defining normal. A diagnostic definition of normal implies that if a result falls outside of a defined range, then there is a known probability of a specific disease being present.14 For jaundiced newborns, this definition does not work well at all. In infants who are readmitted to the hospital with TSB levels of 18 to 20 mg/dL (308–342 μmol/L), the likelihood of finding a specific etiology for the hyperbilirubinemia (other than an association with breastfeeding) is very small.15,16 Hour-specific TSB levels can be informative, however; a TSB level of 10 mg/dL (171 μmol/L) at age 12 hours is almost certainly caused by a hemolytic process, although the precise cause of the hemolysis might not be known.
Normal can also be defined using the risk-factor approach.14 Here a normal range includes levels that carry no additional risk of morbidity or mortality. In jaundiced newborns this would be based on the relationship between TSB levels and subsequent cognitive and neurologic outcome. There are 2 problems with this definition: (1) in the term infant, we have been unable, thus far, to associate a specific risk of damage with a particular bilirubin level, and (2) at the other end of the spectrum, kernicterus has been described in extremely premature infants at very low TSB levels.17,18
For bilirubin levels, a useful definition of normal is the therapeutic definition.14 Here the normal range defines a level beyond which a specific therapy will likely do more good than harm. The recommendations of the American Academy of Pediatrics13 for the use of phototherapy and exchange transfusion in term and near-term newborns are examples of the application of this principle. For example, the American Academy of Pediatrics recommends using phototherapy in a well term infant if the TSB level is ∼15 mg/dL (257 μmol/L) at age 48 hours. Although a level of 15 mg/dL poses no imminent threat to the infant's well-being, at that age it is certainly well above the 95th percentile8 and, if left untreated, might increase to a level that is dangerous for the infant. The suggested intervention, phototherapy, is safe and effective and, under these circumstances, is much more likely to do good than harm.
Thus, with the exception of an early or rapidly rising bilirubin level that suggests hemolysis, the diagnostic definition of normal for indirect hyperbilirubinemia is of limited value, and the risk-factor definition is unhelpful. Currently, the most practical way of describing normal bilirubin levels in term and near-term newborns is to use percentiles. A therapeutic definition of normal values can also be helpful in some circumstances.
If untreated, low birth weight infants have exaggerated and prolonged hyperbilirubinemia. Although this could be considered physiologic because it occurs in all preterm infants, in very low birth weight infants, TSB levels well within the “physiologic range” are potentially hazardous17,18 and are treated with phototherapy. Thus, today, the natural history of hyperbilirubinemia in the very low birth weight infant is never observed, and defining certain bilirubin levels as physiologic in this population is misleading and potentially dangerous. A TSB level of 10 mg/dL on day 4 in a 750-g neonate requires no investigation to identify a cause for the jaundice. Nevertheless, almost all neonatologists would treat this infant with phototherapy, implying that this value exceeds the therapeutic definition of normal (ie, treatment is much more likely to do good than harm).
Jaundice is an important clinical sign seen in most healthy newborns. But, just like tachypnea, tachycardia, and polycythemia, it is a transient event. In sick extremely low birth weight infants, perfectly normal TSB levels can be dangerous. Apparently healthy, term newborns, with TSB levels >20 mg/dL (342 μmol/L) do not have physiologic jaundice. They have hyperbilirubinemia, for which we often cannot find a cause. In time, and with better techniques such as measurements of end-tidal carbon monoxide19 or genetic testing,20 we might identify the cause of the jaundice in more of these infants. We should abandon the terms physiologic and pathologic jaundice and substitute the term “newborn jaundice” or, better, “neonatal bilirubinemia,” which simply means what it says. If we can agree on this terminology we can, presumably, agree on other descriptors for different TSB levels in term and near-term newborns. I suggest that hyperbilirubinemia is the appropriate term for a TSB level that exceeds the 95th percentile for the infant's age in hours in that population. TSB levels >20 mg/dL (340 μmol/L) might be called severe hyperbilirubinemia, and those >25 or 30 mg/dL (428 or 513 μmol/L), extreme hyperbilirubinemia. In the low birth weight population we cannot use population-based norms. In these infants, the therapeutic definition of normal (treatment more likely to help than harm) is probably most useful.
Application of the definitions discussed above should help us in our management of jaundiced infants. If we can also agree on a common terminology, we at least will know what everyone is talking about. Calling jaundice physiologic or pathologic does not achieve these goals.
I thank Drs Tony McDonagh, Michael Kaplan, Tom Newman, and Jon Watchko for helpful comments. Tony McDonagh suggested the quote in the title; “What’s in a name?” comes from Romeo and Juliet, act 2, line 43.
- Accepted March 8, 2006.
- Address correspondence to M. Jeffrey Maisels, MB, BCh, Department of Pediatrics, William Beaumont Hospital, 3601 W 13 Mile Rd, Royal Oak, MI 48073. E-mail:
The author has indicated he has no financial relationships relevant to this article to disclose.
Opinions expressed in these commentaries are those of the authors and not necessarily those of the American Academy of Pediatrics or its Committees.
- ↵Maisels MJ. Jaundice. In: MacDonald MG, Seshia MMK, Mullett MD, eds. Avery's Neonatology. Philadelphia, PA: Lippincott Co; 2005:768–846
- ↵Vreman HJ, Verter J, Oh W, et al. Interlaboratory variability of bilirubin measurements. Clin Chem.1996;42 :869– 873
- ↵Saigal S, Lunyk O, Bennett KJ, Patterson MC. Serum bilirubin levels in breast- and formula-fed infants in the first 5 days of life. Can Med Assoc J.1982;127 :985– 989
- ↵Hardy JB, Drage JS, Jackson EC. The First Year of Life: The Collaborative Perinatal Project of the National Institutes of Neurological and Communicative Disorders and Stroke. Baltimore, MD: Johns Hopkins University Press; 1979
- ↵Maisels MJ, Kring E. Transcutaneous bilirubin levels in a normal newborn population ≥35 weeks' gestation in the first 96 hours. Pediatrics.2006;117 :1169– 1173
- ↵Bhutani VK, Johnson L, Sivieri EM. Predictive ability of a predischarge hour-specific serum bilirubin for subsequent significant hyperbilirubinemia in healthy term and near-term newborns. Pediatrics.1999;103 :6– 14
- ↵Newman TB, Escobar GJ, Gonzales VM, Armstrong MA, Gardner MN, Folck BF. Frequency of neonatal bilirubin testing and hyperbilirubinemia in a large health maintenance organization [published correction appears in Pediatrics. 2001;1:126]. Pediatrics.1999;104 :1198– 1203
- ↵Maisels MJ, Fanaroff AA, Stevenson DK, Young BW, Vreman HJ. Serum bilirubin levels in an international, multiracial newborn population [abstract]. Pediatr Res.1999;45 :167A
- ↵Maisels MJ, Gifford K. Normal serum bilirubin levels in the newborn and the effect of breast-feeding. Pediatrics.1986;78 :837– 843
- ↵American Academy of Pediatrics, Subcommittee on Hyperbilirubinemia. Management of hyperbilirubinemia in the newborn infant 35 or more weeks of gestation [published correction appears in Pediatrics. 2004;114:1138]. Pediatrics.2004;114 :297– 316
- ↵Sackett DL, Haynes RB, Guyatt GH, Tugwell P. Clinical Epidemiology: A Basic Science for Clinical Medicine. 2nd ed. Boston, MA: Little, Brown and Co; 1991
- ↵Maisels MJ, Kring EA. Length of stay, jaundice, and hospital readmission. Pediatrics.1998;101 :995– 998
- ↵Maisels MJ, Kring E. Risk of sepsis in newborns with severe hyperbilirubinemia. Pediatrics.1992;90 :741– 743
- ↵Watchko J, Claassen D. Kernicterus in premature infants: current prevalence and relationship to NICHD phototherapy study exchange criteria. Pediatrics.1994;93 :996– 999
- ↵Govaert P, Lequin M, Swarte R, et al. Changes in globus pallidus with (pre)term kernicterus. Pediatrics.2003;112 :1256– 1263
- ↵Stevenson DK, Vreman HJ. Carbon monoxide and bilirubin production in neonates. Pediatrics.1997;100 :252– 254
- ↵Kaplan M, Hammerman C, Maisels MJ. Bilirubin genetics for the nongeneticist: hereditary defects of neonatal bilirubin conjugation. Pediatrics.2003;111 :886– 893
- Copyright © 2006 by the American Academy of Pediatrics