





* Women and Infants Hospital of Rhode Island/Brown University, Providence, Rhode Island
University of Texas-Houston, Houston, Texas
Case Western Reserve University, Cleveland, Ohio
|| University of Alabama at Birmingham, Birmingham, Alabama
¶ Research Triangle Institute; Research Triangle Park, North Carolina
# Emory University, Atlanta, Georgia
** Wayne State University, Detroit, Michigan

Yale University, New Haven, Connecticut

National Institute of Child Health and Human Development, Rockville, Maryland
| ABSTRACT |
|---|
|
|
|---|
Methods. A retrospective analysis was conducted of a cohort of ELBW infants (4011000 g) who survived to 14 days of age in the 12 participating centers of the National Institute of Child Health and Human Development Neonatal Research Network between January 1, 1994, and December 31, 1997. Demographic and clinical risk factors and PSB levels during the first 14 days were analyzed with reference to death or adverse neurodevelopmental outcomes at 18 to 22 months postmenstrual age. The neurodevelopmental variables considered were Psychomotor Developmental Index (PDI) <70, Mental Developmental Index (MDI) <70, moderate or severe cerebral palsy (CP), hearing impairment (needs hearing aids), and a composite category designated as neurodevelopmental impairment (NDI). The NDI is defined as infants with any 1 or more of the following: PDI <70, MDI <70, moderate to severe CP, bilateral blindness, or bilateral hearing impairment requiring amplification.
Results. The subjects of this cohort analysis are infants who were admitted to the Network centers during calendar years 19941997 and survived beyond 14 days and had PSB recorded during the 14-day period. From this cohort, 3246 infants survived at discharge, 79 died after discharge, and 592 were lost to follow-up. Thus, 2575 of 3167 infants were seen in the follow-up clinics with a compliance rate of 81%. Logistic regression analysis showed that various demographic and clinical variables are associated with poor neurodevelopmental outcomes. After adjustment for these risk factor, significant association were found between PSB (mg/dL) and death or NDI (odds ratio: 1.068; 95% confidence interval [CI]: 1.031.11); PDI <70 (R = 1.057; 95% CI: 1.001.12), and hearing impairment requiring hearing aids (odds ratio: 1138; 95% CI: 1.001.30). There was no significant association between PSB (mg/dL) and CP, MDI <70, and NDI.
Conclusions. PSB concentrations during the first 2 weeks of life are directly correlated with death or NDI, hearing impairment, and PDI <70 in ELBW infants. The statistical association based on retrospective analysis of observational data and relatively small effect size should be interpreted with caution. Furthermore, because of the possibility of compounding effects of variables on outcome, the potential benefits of moderate hyperbilirubinemia and the potential adverse effects of phototherapy, a randomized, controlled trial of aggressive and conservative phototherapy is needed to address this controversial issue.
Key Words: serum bilirubin hyperbilirubinemia extremely low birth weight infant neurodevelopmental outcome
Abbreviations: BBCA, bilirubin binding capacity of albumin ELBW, extremely low birth weight PSB, peak serum bilirubin GDB, generic database CP, cerebral palsy MDI, Mental Developmental Index PDI, Psychomotor Developmental Index NDI, neurodevelopmental impairment IVH, intraventricular hemorrhage CLD, chronic lung disease PVL, periventricular leukomalacia NEC, necrotizing enterocolitis CI, confidence interval OR, odds ratio
A current theory on the development of bilirubin encephalopathy is based on the assumption that when the level of serum unconjugated bilirubin exceeds the bilirubin binding capacity of albumin (BBCA), lipophilic unconjugated unbound bilirubin readily crosses the blood-brain barrier, resulting in neuronal injury.14 The infant may die of kernicterus. If the infant survives, then abnormal neurologic findings, including delayed development and hearing impairment, may be detected.
Previous reports have shown occurrence of kernicterus in very low birth weight infants with relatively low serum bilirubin levels.58 More recently, Watchko and Claassen9 suggested that the incidence of kernicterus in this group of infants is uncommon even if bilirubin is allowed to rise above a level previously considered as harmful. Thus, the relationship of serum bilirubin and neurodevelopmental outcome in extremely low birth weight (ELBW) infants is unclear. The purpose of this study was to assess the association between peak serum bilirubin (PSB) during the first 2 weeks of life and the neurodevelopmental outcomes of ELBW survivors at 18 to 22 months corrected age.
| METHODS |
|---|
|
|
|---|
In 1994, the 12 centers in the network established follow-up programs to assess the outcome of ELBW survivors at 18 to 22 months postmenstrual age. The network determined a priori that there would be strict standardization of neurologic and developmental assessments, interview format, and age of assessment, with a goal of 80% follow-up compliance at the 18- to 22-month evaluation.
During the follow-up visit, trained and certified personnel under the supervision of a site developmental pediatrician performed a comprehensive history, physical examination, and neurodevelopmental assessment. The medical and social histories were obtained by interviewing the infants caregiver. Socioeconomic status information, including maternal and paternal education and occupation, marital status, insurance status, and income levels, was obtained. Neurologic assessment was performed using the Amiel-Tison method.10 This examination included assessment of the infants muscle tone, strength, reflexes, angle, and posture. Cerebral palsy (CP) was defined as a nonprogressive central nervous system disorder characterized by abnormal muscle tone in at least 1 extremity and abnormal control of movement and posture. The developmental assessment consisted of the Bayley Scales of Infant Development II,11 which included the Mental Developmental Index (MDI) and Psychomotor Developmental Index (PDI). We considered MDI and PDI of <70 as abnormal because the values are 2 standard deviations below the mean of 100 for a normal population. The history of infant needing hearing aids (definition of hearing impairment) was obtained from the parent. A composite neurodevelopmental impairment (NDI) score was defined as the infants having 1 or more of the following items: PDI <70, MDI <70, moderate to severe CP, bilateral hearing impairment requiring amplification, or bilateral blindness.
Statistical Analyses
Logistic regressions were used to analyze the association between PSB and the neurodevelopmental outcomes at 18 to 22 months postmenstrual age. The neurodevelopmental outcomes used as dependent variables were NDI, PDI <70, MDI <70, moderate or severe CP, and hearing impairment. Demographic risk factors of birth weight (unit = 100 g); male sex; mother younger than 20 years; mother not a high school graduate; Medicaid insurance recipient; and clinical risk factors of antenatal steroids, surfactant therapy, intraventricular hemorrhage (IVH) grade 3 or 4, chronic lung disease (CLD), periventricular leukomalacia (PVL), late-onset sepsis, proven necrotizing enterocolitis (NEC), postnatal steroids, and site were included in the models. The models were also run with a second order PSB term to test for nonlinear relationships between PSB and neurodevelopmental outcomes. None of the second order terms was significant. Because death is a competing variable for neurodevelopmental outcome, NDI or death was also modeled. Mothers education, Medicaid insurance, and CLD were not included in the death or NDI model because the information was not collected for the infants who died. The graphs in the figures were obtained by plotting the estimated probability and 95% confidence intervals (CIs) of outcome as a function of PSB from the estimated multivariable logistic regression function. The other covariables in the logistic function were evaluated at their average values listed in footnotes below the figures. Although the generic form of the logistic regression is sigmoid or S-shaped, the relationships of the outcomes in this article with PSB were essentially linear, thus distorting the sigmoid to basically a straight line over the range of the data. All analyses were performed using SAS software at the Research Triangle Institute.
| RESULTS |
|---|
|
|
|---|
|
|
|
|
|
|
|
| DISCUSSION |
|---|
|
|
|---|
Hack et al12 previously reported that PSB values >10 mg/dL are a predictor of deafness but not CP or developmental delay in ELBW infants. Others have also shown significant associations between hyperbilirubinemia and abnormal hearing13 or abnormal auditory brainstem response14; our data showing a direct correlation between PSB and hearing impairment is consistent with these observations. These observations point to the importance of assessing auditory functions as a marker of bilirubin neurotoxicity.15
Using the Dutch National Collaborative Survey data, Van de Bor et al16 demonstrated a direct correlation between serum total bilirubin levels and abnormal neurologic outcomes (CP, hearing impairment, seizure, and retinopathy of prematurity) in preterm infants at 2 years of age. The association persisted at 5 years of age for infants who had intracranial hemorrhage.17 Our data are consistent with their findings. These observations point to the potential possibility that hyperbilirubinemia, even at a relatively low serum bilirubin level, may contribute to the abnormal neurodevelopmental outcomes. Alternatively, the high level of serum bilirubin may be a marker of severity of intracranial hemorrhage.
Other observational studies have shown lack of association between low-level hyperbilirubinemia and neurodevelopmental outcomes in ELBW infants. OShea et al18 showed that in a cohort of very low birth weight infants with PSB in the range of 5.8 to 14.6 mg/dL, the association of serum bilirubin level and neurodevelopmental outcomes was no longer statistically significant when intracranial hemorrhage was included in the multivariate analysis. A recent report in abstract form from Australia demonstrated a similar observation.19 In the collaborative phototherapy trial,20 1339 low birth weight infants were randomized to receive phototherapy (or no phototherapy) for 96 hours when serum bilirubin reached a predetermined level. Despite a higher level of serum bilirubin in the no-phototherapy group when compared with those who were treated with phototherapy, no difference was observed in their neurodevelopmental outcomes at 6 years of age. Unfortunately, the follow-up rate at 6 years is only approximately 60%, which may have some effect on the validity of the observation.
During the past 2 decades, the survival rate of ELBW infants has improved in part as a result of more common use of antenatal steroid and surfactant therapy.21,22 Unfortunately, the incidence of neurodevelopmental abnormalities among these survivors remains high.2326 A variety of risk factors (birth weight, male sex, Medicaid, severe IVH, PVL, and CLD) are associated with an unfavorable neurodevelopmental outcome, as has been reported previously by our group.27 Some of the risk factors (eg, IVH, PVL, and CLD) perhaps could be modified by clinical interventions. Hyperbilirubinemia is a clinical entity that falls into the category in which appropriate intervention may improve outcome. Our data provide a framework for the potential design of a randomized, controlled trial to evaluate the effect of intervention.
The association of relatively low serum bilirubin with kernicterus has been shown previously.58 The reason for the association is supported by the observation of a direct correlation between BBCA and gestational age.28 It has also been shown that poor clinical status28 and other risk factors and drugs that compete with albumin binding sites for bilirubin reduce BBCA.6,2931 It is conceivable that in ELBW infants with gestational age in the range of 23 to 28 weeks, who are often clinically very ill and have received multiple therapeutic interventions, the BBCA may be low, allowing for a greater amount of unbound bilirubin to cross the blood-brain barrier. The lack of an apparent threshold at which point the serum bilirubin level begins to show an effect on the neurodevelopmental outcome is of interest. One may speculate that in ELBW infants, such threshold may not be present because they are often critically ill, the albumin levels are low and variable, and the bilirubin effect is more pronounced than in term infants.
Another unknown factor is the role of the blood-brain barrier in producing the brain injury. In the animal model, it has been shown that the blood-brain barrier permeability increases with decreasing level of maturity.32,33 A combination of higher unbound bilirubin (as a result of low BBCA) and a more permeable blood-brain barrier could account for the increased vulnerability of ELBW infants to brain injury at relatively low levels of serum bilirubin. Because of a lack of evidence to guide practice, clinicians are often unsure of what to do in cases of jaundice in ELBW infants. The management strategy is often based on clinical intuition, empirical judgment, and extrapolation from data in full-term infants.
To add to the complexity of this issue is that bilirubin has recently been shown to have some antioxidant effects, which may potentially be beneficial.3437 Yeo et al38 reported an association between visual loss from retinopathy of prematurity among small preterm infants and a PSB level <9.4 mg/dL. The potential hazards of phototherapy have been considered to be of minor clinical importance. However, there is evidence that the oxidative effects of phototherapy on cell membranes39,40 might have a wide range of potential adverse effects.
Phototherapy has been associated with retinopathy of prematurity38 and with patent ductus arteriosus.41 Unrecognized toxicity conceivably could relate to the worrisome and unexplained trend toward an increased mortality among infants who are randomized to phototherapy in the collaborative trial.20
Although very high levels of serum bilirubin levels are known to be toxic, there is continued uncertainty about the risks and benefits of moderate serum bilirubin values and of the use of phototherapy to reduce the bilirubin values in ELBW infants. For this reason, a large randomized trial of aggressive and conservative phototherapy is needed to assess the effects on important clinical outcomes in the neonatal period and at follow-up. The Neonatal Research Network had recently implemented such a trial. An ethical issue can be raised with reference to a randomized, controlled trial allowing a group of infants who will have higher bilirubin levels. Our Network Steering Committee grappled with this issue during the design of the randomized, controlled trial. It was the consensus among the members of the Steering Committee that given the observational nature and statistically small effect size of the current study, the trial is justified.
| ACKNOWLEDGMENTS |
|---|
We gratefully acknowledge the contribution of the following Principal Investigators of the Neonatal Research Network: Charles R. Bauer, MD (University of Miami); Edward F. Donovan, MD, MPH, (University of Cincinnati); Sheldon B. Korones, MD (University of Tennessee at Memphis); James A. Lemons, MD (Indiana University); Lu-Ann Papile, MD, (University of New Mexico); and David K. Stevenson, MD (Stanford University).
| FOOTNOTES |
|---|
Reprint requests to (W.O.) Women & Infants Hospital of Rhode Island, 101 Dudley St, Providence, RI 02905. E-mail: woh{at}wihri.org
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
A. Murakami, M. Morimoto, K. Yamada, O. Kizu, A. Nishimura, T. Nishimura, and T. Sugimoto Fiber-Tracking Techniques Can Predict the Degree of Neurologic Impairment for Periventricular Leukomalacia Pediatrics, September 1, 2008; 122(3): 500 - 506. [Abstract] [Full Text] [PDF] |
||||
![]() |
C V Hulzebos, D E van Imhoff, A F Bos, C E Ahlfors, H J Verkade, and P H Dijk Usefulness of the bilirubin/albumin ratio for predicting bilirubin-induced neurotoxicity in premature infants Arch. Dis. Child. Fetal Neonatal Ed., September 1, 2008; 93(5): F384 - F388. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. J. Bender, W. J. Cashore, and W. Oh Ontogeny of Bilirubin-Binding Capacity and the Effect of Clinical Status in Premature Infants Born at Less Than 1300 Grams Pediatrics, November 1, 2007; 120(5): 1067 - 1073. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. M. Schulzke, G. C. Deshpande, and S. K. Patole Neurodevelopmental Outcomes of Very Low-Birth-Weight Infants With Necrotizing Enterocolitis: A Systematic Review of Observational Studies Arch Pediatr Adolesc Med, June 1, 2007; 161(6): 583 - 590. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. P. Wennberg, C. E. Ahlfors, V. K. Bhutani, L. H. Johnson, and S. M. Shapiro Toward Understanding Kernicterus: A Challenge to Improve the Management of Jaundiced Newborns Pediatrics, February 1, 2006; 117(2): 474 - 485. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. F. McDonagh and M. J. Maisels Bilirubin Unbound: Deja Vu All Over Again? Pediatrics, February 1, 2006; 117(2): 523 - 525. [Full Text] [PDF] |
||||
![]() |
S. L. Merhar and D. L. Gilbert Clinical (Video) Findings and Cerebrospinal Fluid Neurotransmitters in 2 Children With Severe Chronic Bilirubin Encephalopathy, Including a Former Preterm Infant Without Marked Hyperbilirubinemia Pediatrics, November 1, 2005; 116(5): 1226 - 1230. [Abstract] [Full Text] [PDF] |
||||
![]() |
W. Meadow, W. Oh, B. R. Vohr, J. E. Tyson, A. A. Fanaroff, B. J. Stoll, R. A. Ehrenkranz, W. Carlo, S. Shankaran, K. Poole, et al. Importance of Bilirubin Effect Questioned Pediatrics, March 1, 2004; 113(3): 627 - 628. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||