OBJECTIVE: Predischarge bilirubin screening predicts neonatal hyperbilirubinemia. We evaluated the incidence of false-negative bilirubin screening among readmissions for hyperbilirubinemia.
METHODS: In healthy term and late preterm, predominantly breastfeeding newborns, predischarge transcutaneous bilirubin values were plotted on the hour of life–specific bilirubin nomogram and confirmed with plasma total bilirubin in those with a transcutaneous reading ≥75th percentile, or between the 41st and 75th percentiles in the presence of predictive icterogenic risk factors. False-negative bilirubin screen was defined as a predischarge bilirubin value ≤75th percentile in a newborn who was subsequently readmitted for phototherapy.
RESULTS: Of a total of 25 439 neonates born between 2008 and 2009, 143 (0.56%) were readmitted with a mean plasma total bilirubin of 18.7 ± 1.7 mg/dL at 125 ± 54 hours. False-negative predischarge bilirubin screen was identified in 46 (32.2%). Of these, 6 (4.2%) were in the low-risk zone (≤40th percentile, relative risk [RR] = 1) and 40 (28%) in the intermediate-low–risk zone (41st–75th percentile, RR 7.62 [95% confidence interval 3.23–17.96]). Of those in the high-risk zones, 76 (53.1%) were in the intermediate-high–risk zone (76th–95th percentile, RR 25.32 [11.03–58.10]) and 21 (14.7%) in the high-risk zone (>95th percentile, RR 27.78 [11.23–68.70]).
CONCLUSIONS: Predischarge bilirubin levels in newborns classified as low risk did not eliminate the risk of readmission for hyperbilirubinemia. All newborns including those at low risk must be vigilantly observed for subsequent hyperbilirubinemia.
- AAP —
- American Academy of Pediatrics
- PTB —
- plasma total bilirubin
- RR —
- relative risk
- TcB —
- transcutaneous bilirubin
What's Known on This Subject:
The higher the predischarge bilirubin percentile reading on the hour of life–specific nomogram, the higher becomes that infant's risk of developing significant hyperbilirubinemia. Neonates in the low-risk zones (≤75th percentile) have a low risk of developing hyperbilirubinemia.
What This Study Adds:
Thirty-two percent of newborns readmitted for hyperbilirubinemia had low-risk zone predischarge bilirubin percentile values, predominantly in the intermediate low-risk zone (41st–75th percentile). The intermediate low-risk zone may not be as low risk as previously thought.
Predischarge bilirubin screening is useful in the prediction of subsequent neonatal hyperbilirubinemia.1–3 The lower the position of the bilirubin reading on the hour of life–specific bilirubin nomogram,1 the lower becomes the risk of subsequently developing hyperbilirubinemia and vice versa. The issue becomes compounded by the addition of icterogenic risk factors including lower gestational age and/or exclusive breastfeeding.1–3
In a study of newborns readmitted for hyperbilirubinemia performed before to the introduction of routine predischarge bilirubin screening in our nurseries, we reported that some had been assessed by visual inspection during the birth hospitalization to be insufficiently jaundiced to warrant a predischarge plasma total bilirubin (PTB) test.4 Furthermore, some infants who had been visibly jaundiced predischarge had a documented PTB value that fell in the low-risk zones (≤75th percentile and even <40th percentile) but were nevertheless readmitted for hyperbilirubinemia.
In a commentary to and clarification of the 2004 American Academy of Pediatrics (AAP) guidelines for the management of hyperbilirubinemia,5 Maisels et al recommend predischarge bilirubin screening for all newborns, either a blood test or a transcutaneous bilirubin (TcB) technique being satisfactory for this purpose.6 Subsequent to our previous study and encouraged by the publication of these recent recommendations, we introduced universal routine predischarge TcB screening in our nurseries. This procedure now allowed for universal quantification of predischarge bilirubin and for retrospective assessment of the predischarge bilirubin status in all neonates subsequently readmitted for hyperbilirubinemia.
Based on our previous study and some literature reports,2,3 we hypothesized that certain newborns would be readmitted for hyperbilirubinemia despite having predischarge bilirubin screening values in the low-risk zones. The objective of the current study was to determine the rate of false-negative predischarge bilirubin screening in neonates who were subsequently readmitted for hyperbilirubinemia.
Because of the routine, noninvasive, nonrandomization, and noncomparative nature of the project, the institutional review board gave global consent to performing the study and exempted the need for individual parental permission. Our study was composed of an ongoing survey of term and late-preterm newborns who had been born at our medical center and discharged from birth hospitalization as healthy, but who were later readmitted for hyperbilirubinemia. Only those readmitted for primary phototherapy were included in the survey; newborns who had previously been treated with phototherapy and who required repeat treatment of bilirubin rebound were not included.
The predischarge procedure in our regular nurseries at the time the study was being performed (2008–2009) included a routine TcB on the morning of discharge, using a Konica Minolta/Dräger Jaundice Meter JM 103 device (Konica Minolta Sensing, Inc, Osaka, Japan). The TcB result was plotted on the hour of life–specific bilirubin nomogram1 and the percentile reading determined. In those with a TcB value >75th percentile a PTB determination was obtained. In addition, a PTB was determined in those with a TcB reading between the 41st percentile and the 75th percentile if risk factors for hyperbilirubinemia were also present. These included positive direct antiglobulin titer test, glucose-6-phosphate dehydrogenase deficiency, gestational age <37 weeks, discharge at parental request <48 hours, and extracranial hemorrhage including cephalhematoma or ecchymoses. Although breastfeeding is regarded by the AAP as an important risk factor for jaundice, breastfeeding is very frequent in the population delivering at this hospital and thus was not regarded as a specific risk factor for the purpose of this study. However, we did take into consideration those newborns in whom breastfeeding may have been unsuccessful by determining the percentage weight change from the time of delivery to readmission. Unsuccessful breastfeeding was defined as a percentage weight loss of >10% in a breastfed infant. The percentage weight loss was determined at the time of readmission and sheds light on the possible etiology of the hyperbilirubinemia but could not be used predischarge as a predictive risk factor.
In those newborns who qualified for a PTB, this test, and not the TcB, was used for determining subsequent treatment and follow-up. The criteria for phototherapy were according to the guidelines of the Israel Neonatal Society7 which are based on the very similar indications of the AAP.5 Infants with PTB values >75th percentile, or between the 41st and 75th percentiles in the presence of risk factors, but who did not meet the criteria for phototherapy were invited to our service as out patients for a follow-up bilirubin determination within 1 or 2 days, according to the recommendations of the Israel Neonatal Society.7 In those with a PTB percentile value >95th percentile but not meeting criteria for phototherapy, discharge was discouraged. The remainder, comprising those infants regarded as low risk, with a TcB reading ≤40th percentile, or between the 41st and 75th percentiles but without risk factors, were referred to community resources for clinical assessment including jaundice assessment and weight evaluation within 2 to 4 days of discharge. When necessary, these newborns were referred either as outpatients or, after regular hours, to the emergency department.
Outpatient or emergency department assessment was performed by using PTB testing in the identical clinical laboratory and with the identical equipment as in the in-patient testing. Once the need for readmission was determined, the individual newborns were admitted to the regular newborn nursery for phototherapy, additional assessment, and additional treatment, if necessary. Phototherapy was administered until the PTB decreased to the 40th percentile on the hour of life–specific bilirubin nomogram.1
The following information was collected on readmitted infants: age at discharge, age at readmission, birth weight, gestational age, feeding method, presence of risk factors, and predischarge TcB or PTB, as appropriate.
Hyperbilirubinemia requiring readmission for phototherapy was defined as the individual level of PTB requiring phototherapy for any particular infant, and not on an absolute PTB or percentile value. Rather than predetermining a fixed, universal, and single PTB value for each infant, this working definition of hyperbilirubinemia allowed for individualization for the natural dynamics of PTB during the first week of life to be taken into account. Furthermore, both the Israeli and AAP guidelines individualize the indications for phototherapy dependent on lower gestational age or presence or absence of risk factors. We could not ethically justify delaying onset of phototherapy to meet a uniform criterion.
A false-negative screening result was defined as a predischarge bilirubin reading in either of the low-risk zones (ie, ≤75th percentile) in an infant subsequently meeting the criteria for readmission for hyperbilirubinemia.
Routine PTB testing was measured on centrifuged, heparinized capillary tube samples by a direct spectrophotometric method by using absorbance of bilirubin at 455 nm (NEO BIL Model A2; Digital and Analog Systems, Rome, Italy).
The newborns readmitted for hyperbilirubinemia were categorized by predischarge bilirubin value into 1 of the 4 risk groups on the hour of life–specific bilirubin nomogram.1 Although TcB readings had been universally performed, because we did not have readily accessible, computerized predischarge bilirubin values for the infants of the entire cohort, we calculated the expected number in each risk group of the cohort according to the distribution on the nomogram; that is, low-risk group, 40% of total cohort; intermediate-low–risk group, 35% of total; intermediate-high–risk group, 20% of total; and high-risk group, 5% of total. The readmitted infants could now be expressed as a percentage of their category in the cohort. The risk of readmission per risk group was assessed by comparing relative risk (RR) and 95% confidence intervals values for the intermediate-low–risk, intermediate-high–risk, and high-risk groups with that of the low-risk group, which was defined as RR = 1.
During the period of January 1, 2008, to December 31, 2009, there were 25 439 live births at the authors medical center. One hundred forty-three (0.56%) neonates were readmitted for primary phototherapy. Patient demographics and clinical features relevant to hyperbilirubinemia and readmission are depicted in Table 1. It should be noted that the number of newborns in this series delivered by cesarean section is low and not reflective of our overall cesarean section rate, because most infants in this category were treated for hyperbilirubinemia during their birth hospitalization. Thirteen neonates had a weight loss >10% (mean ± SD weight loss of these neonates, −12.6% ± −1.87% [range −10.73% to −16.06%]).
Predischarge bilirubin risk zones and the number of readmitted infants per extrapolated risk zone are summarized in Table 2. As expected, most readmitted neonates (97 of 143 [68%]) had predischarge bilirubin values in the high-risk zones, ie >75th percentile. Surprisingly, however, predischarge, 46 infants (32%) of the readmitted newborns were in the low-risk zones (ie, ≤75th percentile). In all categories higher than the low-risk zone (≤40th percentile), the RR of readmission increased in stepwise fashion and was significantly higher than the low-risk zone group (Table 2). Even in the low-risk zone, however, the rate of readmission was low, but it was not zero. Of special note is the intermediate-low–risk zone, in which the risk of readmission was >7 times that of the low-risk zone. The absolute number of readmitted newborns in the high-risk zone (>95th percentile) is low proportionate to the total readmitted cohort, as we discouraged discharge of infants with a predischarge reading in this zone and, as a result, many infants in this group qualified for phototherapy during birth hospitalization. Nevertheless, the RR of readmission of the infants in the high-risk zone was >27 times that of those in the low-risk zone (Table 3). While some of the infants in the low-risk categories had additional risk factors and had therefore been invited for follow-up, overall, 22 (48%) of the low-risk zones neonates who were readmitted had no additional risk factors (other than breastfeeding), had not met our criteria for bilirubin follow-up and had either been referred from community resources or were self referred.
At least 1 predictive risk factor (additional to the almost universal breastfeeding) was found in 90 neonates. The distribution of risk factors is summarized in Table 3. It will be noted that the most common risk factor was that of late prematurity. The total number of risk factors depicted in Table 3 may be greater than the actual number of neonates as in some newborns >1 factor coexisted.
Figure 1 depicts graphically the percentage of infants readmitted by zone of risk. It will be noted that for each risk category, there were many neonates with risk factors, but also substantial proportions who did not have risk factors. Most of the infants in the low-risk zone (≤40th percentile), however, did have additional icterogenic risk factors and readmission from this zone without risk factors was unusual.
One newborn was admitted with PTB level >25 mg/dL (25.6 mg/dL at age 200 hours). This term infant had been discharged at age 72 hours with a PTB 11.1 mg/dL (equivalent to the 40th percentile). The infant was breastfed, had a weight loss of only −4.2% and had no other obvious risk factors. He was treated successfully with phototherapy.
No infant in this series developed signs of acute bilirubin encephalopathy or required exchange transfusion.
Despite the formulation of national guidelines for the prevention and treatment of hyperbilirubinemia in some Western countries,5,7–11 kernicterus continues into the third millennium.12,13 As kernicterus is a very rare condition, surrogates have been sought to reflect the potential of developing this condition. One surrogate includes readmission for neonatal hyperbilirubinemia. Indeed, many infants in recent kernicterus series were discharged as healthy, only to return within a few days with signs of acute bilirubin encephalopathy.14–17 Prediction of those likely to develop severe hyperbilirubinemia after discharge is therefore imperative.
The current study includes the largest number of readmitted infants available with a universal predischarge bilirubin determination. As expected, many neonates readmitted for hyperbilirubinemia had predischarge bilirubin in the high-risk zones (>75th percentile). Surprisingly, however, 32% of the readmitted infants had predischarge bilirubin values in the low-intermediate– (≤75th percentile) and even low-risk zones (≤40th percentile). A low predischarge bilirubin level does not categorically eliminate the risk of readmission. Furthermore, the presence of risk factors was not a prerequisite for readmission: 48% of the 46 infants with predischarge PTB ≤75th percentile had no identifiable risk factors (additional to that of breastfeeding). The single infant who returned with a PTB >25 mg/dL had a predischarge PTB value on the 40th percentile but no risk factors additional to that of breastfeeding. Our findings contrast to Slaughter et al's study of readmitted newborns, in which all newborns who had had a false-negative predischarge bilirubin value (≤75th percentile) had at least 1 risk factor.18
Several population-based studies have demonstrated an increase in the incidence of hyperbilirubinemia in tandem with increasing predischarge bilirubin percentile.1–4,18,19 When assessed from this perspective, overall, there is indeed a low incidence of subsequent hyperbilirubinemia in infants with predischarge bilirubin values in the low-risk groups. Except for the study of Bhutani et al, in each of the other studies there were newborns who were readmitted for hyperbilirubinemia but who had a predischarge bilirubin value below the 40th percentile. Few studies have systematically assessed false-negative results of predischarge bilirubin screening as the primary objective. When assessed from this aspect, the results contrast with the birth cohort series and suggest that even the low-risk groups may warrant concern.
Our results should not be misinterpreted as undermining the rationale for predischarge risk assessment. Very few readmitted newborns, especially in the absence of additional risk factors, had predischarge bilirubin values ≤40th percentile. These neonates may be safely discharged and followed up clinically only. Those with bilirubin values >75th percentile, especially in the presence of risk factors, will be identified predischarge as being at high risk and should be followed up vigilantly. A change of course is necessary, and new emphasis needs to be placed, on those in the intermediate-low–risk group. The risk of readmission in this group is, in fact, greater than sevenfold that of the low-risk group. These infants should not be discharged with the same low level of concern as their low-risk category counterparts, but must also be closely observed and bilirubin testing performed as clinically necessary.
A limitation of the study is that although TcB determinations had been performed in all neonates of the cohort, we did not have ready, computerized access to each and every predischarge bilirubin value documented for the full cohort during the study period. This necessitated estimation of the number of infants expected in each risk category.1
The objective of bilirubin management in the neonatal period is to prevent the serious outcomes of exchange transfusion and bilirubin encephalopathy. In our series, all infants responded to phototherapy and both outcomes were avoided. Therefore, on the one hand, the bilirubin values reported should not be regarded as a health outcome but rather as risk factors needed to be treated to prevent a negative health outcome. On the other, neonatal PTB values represent a combination of increased bilirubin production and diminished bilirubin conjugation.20 Any additional upset of this balance may further compromise the already lacking equilibrium between the processes, with the potential of severe hyperbilirubinemia.
In a predominantly breastfeeding neonatal cohort, lower-risk zone predischarge bilirubin levels did not eliminate the risk of readmission for hyperbilirubinemia. A predischarge bilirubin screen in the low-risk zones cannot be regarded as a panacea for a safe first week. All neonates must be regarded as potentially at risk for developing hyperbilirubinemia regardless of predischarge bilirubin status and followed for the development of unexpected jaundice in accordance with AAP recommendations.5
- Accepted May 10, 2012.
- Address correspondence to Michael Kaplan, MB, ChB, Department of Neonatology, Shaare Zedek Medical Center, PO Box 3235, Jerusalem 91031, Israel. E-mail:
Dr Bromiker collected and analyzed the data and prepared the first version of the manuscript. Dr Bin-Nun participated in data collection and analysis and reviewed the manuscript. Dr Schimmel participated in data collection and analysis and reviewed the manuscript. Dr Hammerman participated in planning the study, data analysis, and review of the manuscript. Dr Kaplan conceived the study, planned the methodology and data analysis, wrote the final version of the manuscript, and oversaw the project in general.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
- Bhutani VK,
- Johnson L,
- Sivieri EM
- Keren R,
- Luan X,
- Friedman S,
- Saddlemire S,
- Cnaan A,
- Bhutani VK
- Maisels MJ,
- Bhutani VK,
- Bogen D,
- Newman TB,
- Stark AR,
- Watchko JF
- American Academy of Pediatrics Subcommittee on Hyperbilirubinemia
- National Institute for Health and Clinical Excellence. Neonatal jaundice. Clinical Guideline 98, 2010
- Manning D,
- Todd P,
- Maxwell M,
- Jane Platt M
- Sgro M,
- Campbell D,
- Shah V
- Kaplan M,
- Muraca M,
- Hammerman C,
- et al
- Copyright © 2012 by the American Academy of Pediatrics