Original Article
Transient Bilirubin Encephalopathy and Apnea of Prematurity in
28 to 32 Weeks Gestational Age Infants
Sanjiv B. Amin, MD
Lama Charafeddine, MD
Ronnie Guillet, MD, PhD
OBJECTIVE:
Apnea of prematurity (AoP) is, in part, a reflection of brainstem-mediated
respiratory control system maturation. We previously demonstrated
changes in brainstem function in relation to hyperbilirubinemia
(bilirubin encephalopathy, (BE)) as evaluated by auditory brainstem
evoked responses (ABR) in infants 28 to 32 weeks gestational age (GA).
We hypothesized that in this population, as bilirubin increases and causes
auditory brainstem dysfunction, respiratory control system may also be
adversely affected leading to increased frequency of AoP.
BE had significantly more apneic events (15 vs 2, p ¼ 0.0009),
bradycardic events (14 vs 1, p ¼ 0.02), and required more prolonged
treatment with CPAP (2.2 vs 0.5 days, p ¼ 0.007), nasal cannula (6.6 vs
2.2 days, p ¼ 0.02), and methylxanthines (9.5 vs. 1.9 days, p ¼ 0.002)
than those with normal ABR progression. The difference in the incidence
of apnea and bradycardia between infants with and without BE was most
pronounced during the first week.
CONCLUSIONS:
Premature infants with transient bilirubin encephalopathy as defined by
abnormal ABR progression in relation to hyperbilirubinemia have more
concurrent apneic events and require more prolonged respiratory support
and medications.
Journal of Perinatology (2005) 25, 386–390. doi:10.1038/sj.jp.7211295
Published online 21 April 2005
STUDY DESIGN:
We studied 100, 28 to 32 weeks GA infants and identified 66 with normal
and 34 with abnormal ABR progression in temporal relation to
hyperbilirubinemia (BE). The abnormal ABR progression was associated
with elevated bilirubin, specifically elevated unbound bilirubin levels. A
blinded, retrospective chart review quantified the amount of weekly apnea
and bradycardia events during the hospital stay, total duration of
methylxanthine treatment, total duration of mechanical ventilation,
CPAP, and/or nasal cannula, and risk factors for apnea (sepsis, IVH grade
>II, asphyxia). Since mechanical ventilation confounds the identification
of apnea, infants requiring mechanical ventilation were excluded from
further review (n ¼ 60; 21 with BE and 39 with normal ABR
progression). Data from the remaining 40 infants were analyzed.
Student’s t-test was used to analyze continuous variables if the
distribution was normal otherwise Wilcoxon-ranked-sum test was used. w2
was used to analyze nominal variables. A pr0.05 was considered
significant.
RESULTS:
There was no difference in risk factors between infants with and without
BE. BE was identified on day 3 (median; range 1 to 6 days). Patients with
Department of Pediatrics (S.B.A), Division of Neonatology, University of Maryland, Baltimore,
MD, USA; and Department of Pediatrics (L.C., R.G.), Division of Neonatology, Children’s Hospital
at Strong, Rochester, NY, USA.
Address correspondence and reprint requests to Sanjiv B. Amin, MD, University of Maryland
School of Medicine, Division of Neonatology, 22, South Greene Street, N5W68, University Center,
Baltimore, MD 21201, USA.
INTRODUCTION
Apnea of prematurity (A.P.) is a manifestation of developmental
immaturity of respiratory control mediated by the brainstem.
Brainstem maturation may be evaluated by the progression of the
auditory brainstem-evoked response (ABR).1–4 The ABR in
premature infants consists of a sequence of waves, wave I, III and
V. Wave I represents the activity at the auditory nerve level while
waves III and V represents the activity at the brainstem level.5,6 In
28 to 32 weeks gestational age (GA) infants, ABR matures over the
first postnatal week and is characterized by improving detectability
of the response peaks and shortening of the absolute wave
latencies.7 In addition, in premature infants, the waveform can be
categorized depending on the replicability of the response and the
presence of wave III and/or wave V. Response Type also
demonstrates progressive improvement during the first postnatal
week.7
Bilirubin has specific predilection for the auditory neural
pathways, and therefore early detection of bilirubin-induced
neuronal injury is possible using auditory brainstem responses
(ABR).8–12 ABR changes induced by bilirubin progress from
reversible prolongation of the absolute latencies of waves III and V
to loss of wave amplitude and ultimately to the inability to detect
an identifiable wave.8–10,13 Reversible changes may persist for up
to 24 hours after the decrease in serum bilirubin concentrations.
Prolonged bilirubin toxicity may cause irreversible sensorineural
hearing loss.14–16
Journal of Perinatology 2005; 25:386–390
r 2005 Nature Publishing Group All rights reserved. 0743-8346/05 $30
386
www.nature.com/jp
Transient Bilirubin Encephalopathy
Abnormal ABR progression (transient bilirubin encephalopathy)
during the first week is associated with degree of
hyperbilirubinemia and is better predicted by unbound bilirubin
than total serum bilirubin.17 As bilirubin increases and transient
brainstem dysfunction (bilirubin encephalopathy (BE)) occurs, the
likelihood of A.P. may also increase in this population of infants
with immature respiratory control. Our objective was to determine
if premature infants with transient BE as defined by abnormal ABR
progression in relation to hyperbilirubinemia have more
concurrent apnea and bradycardia events as compared to
premature infants without transient BE.
METHODS
All premature infants who were enrolled in our previously reported
bilirubin – albumin binding variable study 17 were eligible for this
nested case–control study as detailed below.
Transient Bilirubin Encephalopathy
The charts of 100, 28 to 32 weeks GA infants, admitted to the
Neonatal Intensive Care Unit of Golisano Children’s Hospital at
Strong, Rochester, NY from July 1996 to July 1998 who participated
in the transient BE study were examined retrospectively.
A detailed method involving ABR recording and analysis of ABR
is described in our original BE study.17 As described previously, for
each patient, sequential ABRs carried out in the first postnatal week
were scored by the audiologists without knowledge of GA or serum
bilirubin level. The waveforms and wave latency progressions were
classified by the audiologists as normal or abnormal maturational
progression. Normal progression was characterized by an
improvement in Response Type (e.g. Type 2 to 1) or by shortening
of the wave V latency. Conversely, abnormal ABR maturation was
defined as a deterioration in Response Type (e.g. from Type 1 to 2)
or prolongation of wave V latency compared to the previous ABRs.
Transient BE was defined as a temporal association between
abnormal ABR maturation and peak total serum bilirubin and
unbound bilirubin. The unbound bilirubin was analyzed using the
same aliquots of blood that was used to measure serum bilirubin
in a subgroup of 31 of the 100 infants. BE was identified on a
median age of 3 days (range 1 to 6 days). Infants received
phototherapy based on institutional guidelines for the use of
phototherapy in the management of hyperbilirubinemia in
premature infants. None of these premature infants received
exchange transfusion for hyperbilirubinemia.
Apnea of Prematurity
An independent investigator who was unaware of the ABR
progression and was not a participant with the original study
performed chart review and collected clinical data on the number
of apnea and bradycardia episodes during each week of the hospital
stay for each of the 100 subjects. In the NICU, all premature
Journal of Perinatology 2005; 25:386–390
Amin et al.
infants are routinely monitored for apnea and bradycardia using
impedance technology. Episodes of apnea lasting longer than 20
seconds and bradycardia with heart rate <100/minute are detected
by the cardiorespiratory monitor. The nurse confirms the events
and records the heart rate and the occurrence of apnea in the
infant’s chart. For the purpose of this study, bradycardia was
defined as a heart rate less than 80/minute. The initiation, use and
discontinuation of respiratory support (nasal cannula, nasal
continuous positive airway pressure (CPAP) and mechanical
ventilation) and methylxanthines for the management of apnea
and/or respiratory disease were at the discretion of the attending
neonatologist.
Information on duration of mechanical ventilation, CPAP, nasal
cannula (NC) and duration of methylxanthines (theophylline or
caffeine) usage during the hospital stay were also collected for each
of the 100 subjects. Additional data included clinical risk factors for
apnea and bradycardia such as sepsis (culture proven or use of
intravenous antibiotics for >7 days), IVH grades III and IV and
asphyxia (Apgar <5 at 5 minutes). The study was approved by the
University of Rochester Institutional Review Board.
Statistical Analysis
Statistical analysis was performed using Stata (Stata Corporation,
College Station, TX, USA). Wilcoxon signed-rank test was used to
test the null hypothesize that the median difference in occurrences
of apnea during the 24-hours period before BE and during the
24-hours period after BE is equal to zero. Student’s t-test was used
to analyze continuous variables if the distribution was normal
otherwise Wilcoxon ranked-sum test was used. w2 was used to
analyze nominal variables. A pr0.05 was considered significant.
Multiple regression analysis was performed to control for the
confounding factors.
RESULTS
Transient Bilirubin Encephalopathy
Of 100 subjects, 66 infants had normal ABR progression and 34
infants had abnormal ABR progression in temporal relation to
hyperbilirubinemia (BE). In these 34 infants with abnormal ABR
progression, peak bilirubin levels preceded (within 24 hours) the
abnormality in ABR progression suggesting temporal relation to
hyperbilirubinemia. ‘There were no statistically significant
differences in the mean birth weight and mean GA between infants
with normal ABR progression and BE (Table 1). The wave V
latencies and Response Types (ABR parameters) at the time of peak
bilirubin measurements were significantly different (Table 1). The
mean peak total bilirubin concentration during the first postnatal
week for the 66 infants with normal ABR progression were not
significantly different than the mean peak bilirubin concentration
measured in the 24 hours preceding the ABR first showing
abnormal progression in the other 34 infants (Table 1). The
387
Amin et al.
Transient Bilirubin Encephalopathy
Table 1 Bilirubin Binding Variables and ABR Parameters as
Function of Transient BE
GA (weeks)*
Birth weight (g)*
Bilirubin (mg/dl)*
Peak unbound bilirubin
(mg/dl)*,w
ABR – Latency V
(mseconds)*
ABR Response Typez
Normal ABR
maturation
(n ¼ 66)
BE
(n ¼ 34)
P-value
30.6±1.1
1467±321
10.0±2.15
0.42±0.18
30.2±1.2
1454±298
10.3±1.7
0.58±0.21
NS
NS
NS
0.04
8.86±0.88
9.76±1.05
0.001
2
3
0.001
NS: non significant.
*Denotes mean±SD.
w
Denotes peak unbound bilirubin levels in a subset of 31 infants.
z
Denotes median values of ABR Response Types.
Table 2 Demographic Profile of the Nonventilated Subjects
GA (weeks)*
Birth weight (g)*
Peak unbound bilirubin
(mg/dl)*,w
Sepsis (n)
Asphyxia (n)
IVH grade >II (n)
Normal ABR
maturation
(n ¼ 27)
BE
(n ¼ 13)
p-value
31.0±0.9
1476±277
0.35±0.11
31.0±0.8
1480±196
0.53±0.20
NS
NS
0.06
3.0
1
0
2.0
0
0
NS
NS
NS
NS denotes nonsignificant.
*Denotes mean±SD.
w
Denotes peak unbound bilrubin levels in a subgroup of eight normal and five BE
infants.
unbound bilirubin measurements were available for 31 out of 100
subjects. Compared to peak total serum bilirubin concentrations,
the peak unbound bilirubin concentration measured in the 24
hours preceding the ABR first showing abnormal progression were
significantly higher among infants with BE compared to mean
peak unbound bilirubin concentration during the first postnatal
week in infants with normal ABR progression (Table 1). This is
highly suggestive that abnormal ABR progression is associated with
bilirubin and best predicted by unbound bilirubin rather than
the total bilirubin. There was no statistically significant difference
in the incidence of clinical risk factors for apnea such as sepsis
(5/31 vs 19/69), asphyxia (1/31 vs 4/69) and IVH grade
>II (0/31 vs 1/69) between group of 31 infants with unbound
bilirubin levels and group of 69 infants in whom the unbound
bilirubin was not measured.
388
Apnea of Prematurity
Of 100 subjects, 60 subjects were mechanically ventilated (40
ventilated for respiratory distress syndrome, 10 ventilated for
respiratory depression at birth and 10 ventilated for severe apnea)
during the first postnatal week of life. In all, 54 out of 60 infants
were ventilated from day 1. Among these 60 subjects, there was no
difference in duration of mechanical ventilation between infants
with normal ABR maturation (n ¼ 39, mean 2.7±4.6 days)
compared to infants with BE (n ¼ 21, mean 1.6±2.8 days)
(p ¼ 0.19). Since, mechanical ventilation confounds identification
of apnea, infants on mechanical ventilation were excluded from
further analysis.
Demographic profiles on the remaining 40 subjects are shown
in Table 2. Of these 40 subjects, 13 infants had BE, while 27
infants had normal ABR maturation. There was no significant
difference in GA, birth weight, or the incidence of clinical risk
factors for apnea between the infants with BE and infants with
normal ABR maturation (Table 2).
In 13 infants with transient BE, the number of apnea episode
were greater during the 24-hours period after transient BE than 24hours period before transient BE (p ¼ 0.04). Infants with BE had
significantly more apnea and bradycardia events during the
hospital stay compared to infants with normal ABR maturation
(Table 3). On comparing weekly apnea and bradycardia events, the
differences in the number of apnea and bradycardia events were
most pronounced during the first week and were statistically
significant only during the first two postnatal weeks between
infants with and without BE (Table 3). These differences remained
statistically significant when controlling for sepsis, asphyxia and
IVH grades III and IV. During the course of their hospitalization,
infants with BE also required significantly more prolonged
respiratory support (NC and NCPAP) and were treated with
methylxanthines longer than infants with normal ABR maturation.
These differences remained statistically significant when
controlling for sepsis, asphyxia and IVH grade >II.
DISCUSSION
Our findings suggest that BE in premature infants, 28 to 32 weeks
gestation manifests as an increased incidence of apnea and
bradycardia, most pronounced during the first postnatal week. In
addition, infants with BE require more prolonged respiratory
support and respiratory stimulants. Although, there have been
anecdotal reports of apnea occurring with acute bilirubin
encephalopathy, this association has not been studied.18 We believe
that this is the first clinical study associating apnea and
bradycardia to acute BE in premature infants.
ABR has been studied in term and to a lesser extent in
premature infants as a direct noninvasive neurophysiological
assessment of reversible bilirubin neurotoxicity.9,12,13,19 These
Journal of Perinatology 2005; 25:386–390
Transient Bilirubin Encephalopathy
Amin et al.
Table 3 Respiratory Support, Respiratory Stimulants, and Apnea and
Bradycardic Events as a Function of Transient BE in Nonventilated
Subjects
Normal ABR
maturation
(n ¼ 27)
BE
(n ¼ 13)
p-value
2.4±4.2
6.6±7.5
0.02
0.5±1.2
1.9±5.9
2.2±2.6
9.5±8.6
0.007
0.002
Total apneaw
Total
bradycardiaw
2 (0–131)
1 (0–144)
15 (3–153)
14 (0–219)
0.0009
0.02
1st week
Apnea/weekw
Bradycardia/
weekw
1 (0–35)
0 (0–14)
6 (2–40)
5 (0–29)
0.0002
0.01
2nd week
Apnea/weekw
Bradycardia/
weekw
0 (0–24)
0 (0–21)
3 (0–29)
4 (0–33)
0.03
0.04
3rd week
Apnea/weekw
Bradycardia/
weekw
0 (0–20)
0 (0–23)
1 (0–18)
2 (0–41)
0.12
0.02
4th week
Apnea/weekw
Bradycardia/
weekw
0 (0–36)
0 (0–54)
0 (0–40)
1 (0–64)
0.34
0.11
5th week
Apnea/weekw
Bradycardia/
weekw
0 (0–9)
0 (0–20)
0 (0–15)
0 (0–36)
0.89
0.33
Time period
Parameter
During entire
hospital stay
Nasal cannula
(days)*
CPAP (days)*
Methylxanthines
(days)*
*Mean±SD.
w
Median (range).
studies have demonstrated reversible changes in ABR waves III and
V latency induced by bilirubin. Both waves III and V reflect activity
in the auditory nervous system at the brainstem level.5 As a result
of the proximity of auditory neural system to cardiorespiratory
center, ABR results may correlate with other brainstem functions
such as control of heart rate and breathing. Henderson-Smart et al.
have shown an association between abnormal brainstem function
as evaluated by ABR and AoP.20 The inference is that if bilirubin
rises and affects brainstem function, there will be an increase in
the number of apneic and bradycardic episodes.
In general, the frequency of apnea and bradycardia decreases
over time in both the groups. It is not clear whether this trend is
due to therapeutic interventions or normal developmental
Journal of Perinatology 2005; 25:386–390
maturation of the brainstem. However, the existence of significant
differences in the incidence of apnea and bradycardia between the
two groups during the first two postnatal weeks, with the most
pronounced differences during the first postnatal week, strongly
suggest an association between BE and apnea. Unfortunately, we
did not perform sequential ABRs and analyze serum bilirubin levels
beyond the first week to determine when bilirubin-induced ABR
changes in infants with BE improve and with what consistency and
completeness become similar to infants with normal ABR
progression.
The retrospective design is the major limitation with our study.
In addition, because of the retrospective nature and impedance
technology used to monitor apnea, we could not differentiate
between obstructive and central apnea. However, the chart review
was performed by an investigator who was unaware of the study
groups. Despite these limitations, our findings have clinical
relevance and may provide the basis for future studies that will be
required to determine a causal effect of bilirubin on apnea in
premature infants. Our observation of increased AoP with degree of
hyperbilirubinemia for infants of 28 to 32 weeks gestation may be
generalizable and lead to a more function-based approach to the
determination of safer levels of unbound bilirubin in infants of
other GA groups, especially the very premature infants. Once
threshold levels of total bilirubin or unbound bilirubin that result
in brainstem dysfunction have been determined for different GA
groups, a more rational basis for phototherapy can be developed. If
the effect of bilirubin on apnea is causal then aggressive treatment
of premature infants with elevated bilirubin levels to prevent
bilirubin-induced brainstem dysfunction may decrease the
incidence of apnea and bradycardia events.
Acknowledgements
We thank all the neonatal intensive care unit nurses in monitoring the infants for
apnea and bradycardia events during the course of the study. We thank Dr. Lois
Johnson for her guidance during the preparation of this manuscript.
References
1. Amin SB, Merle KS, Orlando MS, Dalzell LE, Guillet R. Brain stem
maturation in premature infants as a function of enteral feeding type.
Pediatrics 2000;106(2):318–22.
2. Tyson JE, Lasky R, Flood D, et al. Randomized trial of taurine
supplementation for infants >1300 gram birth weight: effect of auditory
brain-stem evoked responses. Pediatrics 1989;83:406–15.
3. Kohelet D, Arbel E, Goldberg M, Arlazzoroff A. Transient neonatal
hypothyroxinemia and the auditory brain stem evoked response. Pediatr Res
1992;32:530–1.
4. Amin SB, Orlando MS, Dalzell LE, Merle KS, Guillet R. Brainstem
maturation after antenatal steroids exposure in premature infants as
evaluated by auditory brainstem-evoked response. J Perinatol 2003;23:
307–311.
389
Amin et al.
5. Moller AR, Jannetta PJ, Moller MB. Neural generators of brainstem evoked
potentials: results from human intracranial recordings. Ann Otol Rhinol
Laryngol 1981;90:591–6.
6. Starr A. Auditory brainstem potentials: comments on their use during infant
development. J Clin Neurophysiol 1984;1:331–4.
7. Amin SB, Orlando MS, Dalzell LE, Merle KS, Guillet R. Morphological
changes in serial auditory brainstem responses in 24–32 weeks’ gestational
age infants during the first week of life. Ear Hear 1999;20:410–8.
8. Wennberg RP, Ahlfors CE, Bickers R, McMurtry CA, Shetter JL. Abnormal
auditory brainstem response in a newborn infant with hyperbilirubinemia:
improvement with exchange transfusion. J Pediatr 1982;100(4):624–6.
9. Perlman M, Fainmesser P, Sohmer H, et al. Auditory nerve-brainstem
evoked responses in hyperbilirubinemic neonates. Pediatrics 1983;72:
658–664.
10. Nakamura H, Takada S, Shimabuku R, Matsuo M, Matsuo T, Negishi H.
Auditory nerve and brainstem responses in newborn infants with
hyperbilirubinemia. Pediatrics 1985;75(4):703–8.
11. Funato M, Tamai H, Shimada S, Nakamura H. Vigintiphobia, unbound
bilirubin, and auditory brainstem responses. Pediatrics 1994;93(1):50–3.
12. Lenhardt ML, McArter R, Bryant B. Effects of neonatal hyperbilirubinemia
on the brainstem electric response. J Pediatr 1984;104(2):281–4.
390
Transient Bilirubin Encephalopathy
13. Amin SB, Orlando MS, Merle KS, Dalzell LE, Guillet R. Sequential changes
in BAER and their relation to bilirubin levels in neonates less than 32 weeks
gestational age. Pediatr Res (abstract) 1997;41(4):135A.
14. Chisin R, Perlman N, Sohmer H. Cochlear and brainstem responses in
hearing loss following neonatal hyperbilirubinemia. Ann Otol Rhinol
Laryngol 1979;89:352–7.
15. De Vries LS, Lary S, Dubowitz LMS. Relationship of serum bilirubin levels to
ototoxicity and deafness in high-risk low birth weight infants. Pediatrics
1985;76:351–4.
16. Bergman I, Hirsch RP, Fria TJ, Shapiro SM, Holzman I, Painter MJ. Causes
of hearing loss in the high risk premature infant. J Pediatr 1985;106:95–101.
17. Amin SB, Ahlfors C, Orlando MS, Dalzell LE, Merle KS, Guillet R. Bilirubin
and serial auditory brainstem responses in premature infants. Pediatrics
2001;107:664–70.
18. Govaert P, Lequin M, Swarte R, et al. Changes in globus pallidus with
preterm kernicterus. Pediatrics 2003;112(6 Part 1):1256–63.
19. Vohr BR. New approaches to assessing the risk of hyperbilirubinemia. Clin
Perinatol 1990;17(2):293–306.
20. Henderson–Smart D, Pettigrew AG, Campbell DJ. Clinical apnea and
brainstem neural function in preterm infants. N Eng J Med 1983;308:
353–357.
Journal of Perinatology 2005; 25:386–390