e u r o p e a n j o u r n a l o f p a e d i a t r i c n e u r o l o g y 1 8 ( 2 0 1 4 ) 1 9 3e2 0 0

Official Journal of the European Paediatric Neurology Society

Original article

Brainstem auditory electrophysiology is supressedin term neonates with hyperbilirubinemia

Ze D. Jiang*, T.T. Liu, Cao Chen

Division of Neonatology, Children’s Hospital, Fudan University, Shanghai, China

a r t i c l e i n f o

Article history:

Received 7 June 2013

Received in revised form

5 November 2013

Accepted 8 November 2013

Keywords:

Brainstem auditory impairment

Evoked potentials

Hyperbilirubinemia

Neonate

Neurological impairment

* Corresponding author. Neonatal Unit, Depa1865 221364; fax: þ44 1865 221366.

E-mail addresses: zedong.jiang@paediatr1090-3798/$ e see front matter ª 2013 Europhttp://dx.doi.org/10.1016/j.ejpn.2013.11.004

a b s t r a c t

Background: Whether hyperbilirubinemia suppresses electrophysiological activity of the

neonatal auditory brainstem remains to be investigated.

Aim: To determine whether hyperbilirubinemia suppresses the brainstem auditory elec-

trophysiology in term neonates.

Methods: Maximum length sequence brainstem auditory evoked response (MLS BAER) was

recorded shortly after confirming hyperbilirubinemia in 58 term neonates. Wave ampli-

tudes of the response were analyzed in detail.

Results: Compared with age-matched term controls, the neonates with hyperbilirubinemia

showed a significant reduction in the amplitudes of MLS BAER waves III and particularly V

at all click rates 91e910/s. The reduction tended to be more significant at higher than lower

rates. Wave I amplitude was reduced at 910/s. V/I amplitude ratio was decreased at all click

rates. Therefore, the amplitudes of MLS BAER, particularly later, waves were all reduced.

The amplitudes of all MLS BAER waves tended to be reduced with the increase in total

serum bilirubin level. All wave amplitudes were correlated with the level of total serum

bilirubin at some or most click rates.

Conclusions: Brainstem auditory electrophysiology is suppressed in neonates with hyper-

bilirubinemia, which related to the severity of hyperbilirubinemia. Wave amplitudes are

valuable BAER variables to detect functional impairment of the brainstem and auditory

pathway in neonatal hyperbilirubinemia, and are recommended to be used in assessing

bilirubin neurotoxicity to the neonatal brain.

ª 2013 European Paediatric Neurology Society. Published by Elsevier Ltd. All rights

reserved.

1. Introduction

A significant increase in unconjugated bilirubin concentra-

tion, coupled with other risk factors, allows free bilirubin to

rtment of Paediatrics, Joh

ics.ox.ac.uk, jiangzedong-ean Paediatric Neurology

cross the bloodebrain barrier. In newborn infants, serum

bilirubinmay rise to hazardous levels that pose a direct threat

to the brain. The neonatal brain, particularly the auditory

system, is sensitive to high level of serumbilirubin, and can be

damaged.1e3 Neonates who have severe and/or prolonged

n Radcliffe Hospital, Headington, Oxford OX3 9DU, UK. Tel.: þ44

oxshang@hotmail.com (Z.D. Jiang).Society. Published by Elsevier Ltd. All rights reserved.

e u r o p e a n j o u r n a l o f p a e d i a t r i c n e u r o l o g y 1 8 ( 2 0 1 4 ) 1 9 3e2 0 0194

hyperbilirubinemiamay develop bilirubin encephalopathye a

neurotoxic syndrome that occurs as a result of the deposition

of unbound or free unconjugated bilirubin in selective brain

areas.2e4 The encephalopathy frequently evolves into ker-

nicterus d a severe bilirubin induced encephalopathy with

yellow staining of parts of the brainstem, hippocampus, cer-

ebellum and certain brainstem nuclei.3e5 An improved un-

derstanding of the effect of hyperbilirubinemia on the

neonatal brain and early detection of bilirubin encephalopa-

thy is important for clinical management of neonates with

hyperbilirubinemia.4,6,7

Over the last two decades, there are investigators who

studied the effect of bilirubin neurotoxicity to the auditory

brainstem by using the brainstem auditory evoked response

(BAER). With conventional BAER (i.e. the BAER that is

recorded using conventional averaging techniques), most

investigators found abnormalities in wave latencies and

interpeak intervals in infants with hyperbilirubinemia,

although the others did not.8e15 The BAER abnormalities

were linked with an increased level of unbound or “free”

unconjugated bilirubin.

In addition to wave latency and interpeak interval, the

amplitudes of BAER components or waves can also reflect

electrophysiological activity of auditory neurons in the

brainstem following acoustic stimulation. Some studies

showed that wave amplitudes, particularly wave V ampli-

tude, are useful BAER variables for assessment of auditory

electrophysiology of the neonatal brainstem.16e19 In piglets

with experimental hyperbilirubinemia, Hansen et al. found

a significant reduction in BAER wave amplitudes but no

change in BAER wave latencies.20 It seems that the ampli-

tudes of BAER components are probably more sensitive to

an increased level of bilirubin than the latencies. A reduc-

tion in wave amplitudes may reflect a suppression of

brainstem auditory electrophysiology.21e24 It may be of in-

terest to examine BAER wave amplitudes in human infants

with hyperbilirubinemia. So far, there were only a few

previous investigators who paid attention to the amplitudes

of BAER components in their BAER study of neonates with

hyperbilirubinemia. The results were somewhat conflicting.

Some reported no changes in BAER wave amplitudes, while

the others found poor wave amplitudes, in addition to an

increase in BAER wave latencies.25,26 Therefore, it remains

to be clarified whether hyperbilirubinemia affects the am-

plitudes of BAER components in neonates with hyper-

bilirubinemia, or whether hyperbilirubinemia suppresses

electrophysiological activity of the neonatal auditory

brainstem.

To address this issue and further the understanding of

bilirubin neurotoxicity to the neonatal auditory brainstem,

we carried out a detailed analysis of the amplitudes of BAER

components in term neonates who suffered hyper-

bilirubinemia. The BAER was studied using the maximum

length sequence technique (MLS).27 The amplitudes of MLS

BAER wave components were analyzed in detail. The

response was elicited at various repetition rates of click

stimuli to examine whether the amplitudes of BAER com-

ponents are affected by hyperbilirubinemia at all click rates,

particularly high rates that cannot be achieved in conven-

tional BAER. The relationship between the severity of any

amplitude abnormality and the level of total serum bilirubin

(TSB) was also analyzed.

2. Subjects and methods

2.1. Subjects

Fifty-eight term neonates who were diagnosed as hyper-

bilirubinemia during the first 10 days of life were recruited as

the study group. All had a TSB level greater than 15 mg/dL

(18.9� 3.7mg/dL) and required phototherapy and/or exchange

transfusion. Gestational age ranged between 37 and 42 weeks

(39.4 � 1.2 weeks), and a birthweight between 2240 and 5302 g

(3401 � 528 g). They were all Chinese, and recruited from the

Division of Neonatology, Children’s Hospital, Fudan Univer-

sity between December 2008 and February 2011. The informed

consent was obtained from the parents and the pediatrician in

charge. All infants were studied with MLS BAER during the

first 10 days after birth when there were clinical signs of

jaundice and a TSB level greater than 15 mg/dL.

The major etiology was hemolysis (rhesus hemolytic dis-

ease, anemia, a positive direct antiglobulin test, reticulocytosis,

or a peripheral blood smear compatiblewith hemolysis, n¼ 21).

The other plausible aetiologies included infection (pneumonia

or sepsis, n ¼ 6), low Apgar scores (between 4 and 6) at 1 and

5min (n¼ 15), orunknown, i.e.nodetectableperinatalproblems

(n ¼ 16). To minimize any confounding effects on the study re-

sults, infants had been excluded from study entry if he or she

had any other perinatal problems or complications that could

affect neonatal BAER. These problems included congenital

malformations, in utero infection, a family history of hearing

loss, low birthweight, severe intrauterine growth retardation,

bacterial meningitis, seizures, hypoxia-ischemia, persistent

pulmonary hypertension.28

Significant peripheral hearing problems are known to affect

the amplitudes of MLS BAER waves. In order to minimize such

an effect we excluded any neonates who had a BAER threshold

>40 dB normal hearing level (nHL). All neonates recruited

passed our neonatal hearing screening programme with

distortion product otoacoustic emission. Any infants who had

ototoxic medication were also excluded from the study entry.

The normal control group was comprised of 43 healthy

term infants, with gestational ages were between 37 and 41

weeks (38.9 � 1.2 weeks) and birthweights between 2569 and

4539 g (3460 � 459 g). The TSB levels were all below 10 mg/dL

(4.3� 1.4mg/dL). Monaural hearing thresholds were all within

normal range (<20 dB nHL, determined by conventional BAER

with a click rate 21/s). None had any major perinatal condi-

tions or problems that may affect the brainstem auditory

function, as detailed above.

2.2. Recording of MLS BAER

At time of MLS BAER recording, there was no significant dif-

ference in postconceptional age between the study group

(37e42 weeks, 39.8 � 1.3 weeks) and control group (37e42

weeks, 39.1 � 1.4 weeks). Within 30 min after obtaining blood

sample and confirming TSB level greater than 15 mg/dL, the

e u r o p e a n j o u r n a l o f p a e d i a t r i c n e u r o l o g y 1 8 ( 2 0 1 4 ) 1 9 3e2 0 0 195

subjects were moved to a quiet room in the neonatal unit to

prepare for MLS BAER recording.

The protocol of recordingMLS BAERwas generally the same

aswepreviously reported21e24,29e31 The subjects lay supine in a

cot. The auditory meatus was inspected and cleaned of any

vernix or wax. After skin preparation, three gold-plated disk

electrodes were placed, respectively, at the middle forehead

(positive), the left (ipsilateral) earlobe (negative) and the right

(contralateral) earlobe (ground), respectively. Interelectrode

impedances were maintained at 5 kU and less. A Spirit 2000

Evoked Potential System (Nicolet Biomedical Inc. Madison, WI,

USA)wasused to recordMLS BAER. For all subjects, only the left

ear was tested to keep the consistence of recording and

analyzing conditions and save the time of recording MLS BAER.

Carewas takentoavoiddisturbingthesubjectsduring theentire

session of MLS BAER recording.

The recording commenced after the infant fell asleep

naturally, often after a feed. Rarefaction clicks of 100 ms were

delivered to the left ear through a TDH 39 earphone (supplied

by Nicolet Biomedical Inc. Madison, WI), which was comfort-

ably placed over the ear with a great care to avoid collapsing

ear canals. Initially, conventional BAER was recorded at 21/s

clicks to determine BAER threshold (the lowest intensity at

which wave V can be reliably recognized). Thereafter,

recording of MLS BAER was started at 60 dB nHL clicks. Higher

intensities were also used in those who had an increased

BAER threshold (>20 dB nHL). The repetition rates of clicks

ranged between 91/s and 910/s. Two runs were made for each

recording condition. Each run included brain responses to

1500 trains of clicks. In the first run, the clicks were presented

at the sequence of 91, 227, 455 and 910/s. In the second run,

the click rates were presented in a reverse sequence.

Sweep duration was 24 ms. The brain responses evoked by

the clicks were amplified and filtered at a bandwidth between

100 Hz and 3000 Hz. During signal averaging amplitude artifact

rejection was active to reduce the inclusion of high-amplitude

muscular activity in the averaged responses. The Evoked Po-

tential system automatically rejected any on-line signals

exceeding �25 mV, e.g. high-amplitude muscular activity, and

excluded these from the average. Whenever there were exces-

sive muscle artefacts on the monitoring oscilloscope sampling

was manually discontinued until the artefacts diminished or

significantly reduced. Duplicate recordings were made in

response toeachstimuluscondition toexamine reproducibility.

These procedures of recordingMLS BAERwere approved by the

Children’s Hospital Ethics Committee of Fudan University.

2.3. Data analysis

The measurement of MLS BAER wave amplitudes were made

in a way the same as previously reported.19,21e24 For newborn

infants, the down slope of wave I is often significantly affected

by presence or absence of wave II, or the latency of wave II if it

is present. This can produce considerable variation in the

amplitude of wave I, and in turn affect the measurement and

value of V/I amplitude ratio. In order to minimize such a

variation the amplitude of wave I was measured from its peak

to the lowest trough between waves I and III, which is more

consistent and reliable. Because the trough after wave III is

considerably variable, it is not reliable to use the trough to

measure the amplitude of wave III. Therefore, the amplitude

of wave III was measured from the lowest trough between

waves I and III to the peak of wave III.21e24,32 As usual, the

amplitude of wave V was measured from the positive peak of

wave V to the negative trough immediately after the peak. V/I

and V/III amplitude ratios were then calculated.

To avoid variation in amplitude measurements due to the

effect of different intensity level above BAER threshold

among different subjects, all quantitative analyses of BAER

amplitude variables were based on the data collected at an

intensity level 40 dB or slightly higher levels above the

thresholds of each subject.21e24 The data obtained in

response to 60 dB nHL clicks were analyzed in detail. In

subjects who had a threshold >20 dB nHL, analysis of MLS

BAER data was based on recordings collected at higher click

intensities (i.e. 70 dB nHL for thresholds >20e30 dB nHL

(n ¼ 3); 80 dB nHL for thresholds >30e40 dB nHL (n ¼ 2).

Thus, as in the normal controls, all MLS BAER recordings in

the neonates with hyperbilirubinemia were analyzed at a

hearing level slightly �40 dB above BAER threshold of each

subject. Any appreciable influence of threshold elevation on

MLS BAER measurements was cancelled. The click intensity

above BAER threshold of the subjects at which MLS BAER

measurements were obtained was 48.1 � 6.0 dB in the neo-

nates with hyperbilirubinemia, and 49.3 � 5.2 dB in the

normal controls, which did not differ significantly.

2.4. Statistical analysis

The measurements of each MLS BAER component in two

replicated recordings were averaged for further analysis. MLS

BAER measurement and analysis were all carried out blind to

the medical history and clinical data of each subject. Mean

and standard deviation of each MLS BAER variable at each

stimulus condition were compared between the study and

control groups with a SPSS package (SPSS, Chicago, IL). Anal-

ysis of covariance (ANCOVA) was performed for comparison

between the study and control groups of infants in the mean

and standard deviation of each BAER variable at each stimulus

condition (as the dependent variables), with postconceptional

age, gender and the click intensity above BAER threshold as

covariates. This allowed us tominimize any possible effects of

variation in these covariates in the study and control groups

on the measurements of BAER variables.

The relationship between MLS BAER variables and the

repetition rate of clicks was examined by correlation and

regression analysis. The slope (or regression coefficient) of

latency- or interval-rate function was calculated for each MLS

BAER variable, and compared between the study and control

groups. Correlation analysis was conducted between MLS

BAER variables and the repetition rate of clicks was per-

formed. In regression analysis, amplitude-rate function was

calculated for each BAER variable. The slope of the amplitude-

rate functions were obtained and examined using a t-test to

determine if any slope was consistently different from zero.

Any amplitude-rate function that was significantly greater

than zero at the 0.05 level or better was then compared be-

tween the study and control groups to detect any abnormality

in click rate-dependent changes in the neonates with hyper-

bilirubinemia. Correlation analysis of MLS BAER variables

e u r o p e a n j o u r n a l o f p a e d i a t r i c n e u r o l o g y 1 8 ( 2 0 1 4 ) 1 9 3e2 0 0196

with the level of TSB was also conducted to examine whether

any abnormality or change in MLS BAERwave amplitudes was

closely related to the level of TSB.

Fig. 2 e Boxplot of MLS BAER wave III amplitude (bold line

across the box, median; box, 25th and 75th centile;

extensions, the largest and smallest values) at different

click rates at term age in term neonates with

hyperbilirubinemia and normal term infants. HBN, term

neonates with hyperbilirubinemia; NT, normal term

infants.

3. Results

Figs. 1e3 are boxplots of amplitude measurements of MLS

BAER waves I, III and V at various click rates in the neonates

with hyperbilirubinemia and the normal controls. Figs. 4 and 5

are boxplots of measurements of V/I and V/III amplitude ratio

at various click rates.

Compared with the normal controls, the neonates with

hyperbilirubinemia showed a reduction in all amplitudes of

BAER waves I, III, and V. Wave I amplitude was slightly

reduced at the repetition rates 91e455/s of clicks, but was

significantly reduced at the highest rate 910/s (p < 0.01, Fig. 1).

By comparison, the reduction in wave III amplitude in the

neonates with hyperbilirubinemia was more significant. The

amplitude was significantly smaller than in the controls at all

click rates (p < 0.05e0.001, Fig. 2), with the greatest difference

at 910/s (p < 0.001). Similarly, wave V amplitude in the neo-

nates with hyperbilirubinemia was smaller than in the con-

trols at all rates of clicks, particularly at very high rates 455/s

and 910/s (all p < 0.001, Fig. 3). The reduction was more sig-

nificant than that in both waves I and III.

The V/I amplitude ratio was significantly smaller in the

neonates with hyperbilirubinemia than in the controls at all

rates of clicks (p < 0.05e0.01, Fig. 4). The V/III amplitude ratio

was slightly smaller than in the controls at the lower rates 91

and 227/s, but was significantly smaller at the higher rates

455/s and 910/s (p < 0.01 and 0.01, Fig. 5).

All amplitudes of MLS BAER waves I, III and V in the

neonates with hyperbilirubinemia had a similar positive

Fig. 1 e Boxplot of MLS BAER wave I amplitude (bold line

across the box, median; box, 25th and 75th centile;

extensions, the largest and smallest values) at different

click rates at term age in term neonates with

hyperbilirubinemia and normal term infants. HBN, term

neonates with hyperbilirubinemia; NT, normal term

infants.

correlation with the repetition rate of clicks

(r ¼ �0.689e0.702, all p < 0.01). However, neither was the V/I

amplitude ratio nor the V/III amplitude ratio correlated with

click rate. These were similar to those in the normal con-

trols. Regression analysis was performed for the relationship

between MLS BAER latency and interval variables and click

rate. The intercepts of the amplitude-rate functions for all

waves I, III and V in the neonates with hyperbilirubinemia

Fig. 3 e Boxplot of MLS BAER wave V amplitude (bold line

across the box, median; box, 25th and 75th centile;

extensions, the largest and smallest values) at different

click rates at term age in term neonates with

hyperbilirubinemia and normal term infants. HBN, term

neonates with hyperbilirubinemia; NT, normal term

infants.

Fig. 4 e Boxplot of MLS BAER V/I amplitude (bold line

across the box, median; box, 25th and 75th centile;

extensions, the largest and smallest values) ratio at

different click rates at term age in term neonates with

hyperbilirubinemia and normal term infants. HBN, term

neonates with hyperbilirubinemia; NT, normal term

infants.

e u r o p e a n j o u r n a l o f p a e d i a t r i c n e u r o l o g y 1 8 ( 2 0 1 4 ) 1 9 3e2 0 0 197

tended to be smaller than those in the controls. The slopes

of amplitude-rate functions for these waves tended to be

greater than those in the controls, but none differed signif-

icantly between the neonates with hyperbilirubinemia and

the controls.

The amplitudes of MLS BAERwaves I, III and V all tended to

be smaller at higher than at lower levels of TSB. Wave I

amplitude was correlated significantly with the level of TSB at

Fig. 5 e Boxplot of MLS BAER V/III amplitude (bold line

across the box, median; box, 25th and 75th centile;

extensions, the largest and smallest values) ratio at

different click rates at term age in term neonates with

hyperbilirubinemia and normal term infants. HBN, term

neonates with hyperbilirubinemia; NT, normal term

infants.

most click rates (21, 91, 455 and 91/sec, r ¼ 0.331e0.529,

p< 0.05 or 0.01).Wave III amplitude at 445 and 910/s, andwave

V amplitude at 227 and 455/s were correlated with TSB

(r ¼ 0.305e0.532, p < 0.05 or 0.01). The V/I amplitude ratio at

91e445/s, and the V/III amplitude ratio at 91 and 455/s were

correlated with TSB (r ¼ 0.281e0.373, all p < 0.05).

4. Discussion

In neonates, bilirubin is a well-known neurotoxin that can

result in multiple neurologic impairment and deficits.4,6,7 In

the literature, there were only few published reports on wave

amplitudes in conventional BAER. Funato et al. reported that,

in addition to an increase in latencies, therewas a reduction in

the mean amplitudes for BAER waves I, III and V, although

others did not show any significant differences in BAER am-

plitudes between infants with hyperbilirubinemia and those

with normal level of serum bilirubin.25,26 In other modalities

of evoked potentials, some authors reported an amplitude

reduction in neonates with hyperbilirubinemia. For example,

Chen et al. found a reduction in the amplitudes of visual

evoked potentials during the firstweek of life in their neonates

with hyperbilirubinemia.33,34 With conventional BAER, we

previously studied wave amplitudes in neonates with hyper-

bilirubinemia.35 The amplitude of wave I did not show any

significant reduction, while the amplitudes of BAER waves III

and V were reduced. The V/I and V/III amplitude ratios tended

to be decreased.

In this report, BAERwas studied using theMLSd a relatively

new technique to study auditory evoked potentials.27,32,36e38

With the MLS technique we have recently studied BAER in

newborn infants with some perinatal problems, and found that

MLS BAER is a valuablemethod to enhance the diagnostic value

of BAER.21e24,27e31,39 In our neonates with hyperbilirubinemia,

the amplitudes of MLS BAER, particularly later, wave compo-

nents were significantly reduced, which was more significant

than what we observed in conventional BAER.35 The amplitude

reduction reflects an effect of bilirubin neurotoxicity to the

electrophysiological activity of auditory neurons in the brain-

stem. This effect could cause a decreased contribution from

each of the auditory neuron, a decreased synchrony of volley

with which the neurons are activated and a decreased average

membrane potential of the auditory neurons (so the driving

potential for current flow is reduced). Consequently, the elec-

trophysiological activity of auditory neurons is decreased,

leading to a reduction in the amplitudes of an auditory evoked

potential. Therefore, the reduction in the amplitudes of MLS

BAER components in neonates with hyperbilirubinemia in-

dicates a suppression of brainstem auditory electrophysiology

due to bilirubin neurotoxicity.

Earlier animal experiments carried out by Hansen and

colleagues revealed that bilirubin decreased phosphorylation

of synapsin I, a synaptic vesicle-associated neuronal phos-

phoprotein, in intact synaptosomes from rat cerebral cortex.40

In rat hippocampal slices, bilirubin also reduced the ampli-

tude of extracellularly recorded synaptic field potentials.41

Their further experiments in bilirubin-infused piglets

revealed a significant reduction in the amplitudes of BAER

waves IIeV but no changes in wave latencies.20 By

e u r o p e a n j o u r n a l o f p a e d i a t r i c n e u r o l o g y 1 8 ( 2 0 1 4 ) 1 9 3e2 0 0198

administering biliverdin, the immediate precursor of bili-

rubin, in 15- to 17-day-old Gunn rat pups, Rice and Shapiro

found that a single biliverdin injection produced a significant

decrease in the amplitude of BAER wave III.1 These findings in

animal models are essentially comparable with the present

MLS BAER findings in the neonates with hyperbilirubinemia.

The amplitude reduction of MLS BAER components in

our neonates with hyperbilirubinemia was most significant

for wave V, followed by wave III and the least wave I. This

was true of all click rates. Namely, the reduction was more

significant for the later BAER waves (i.e. waves with longer

latencies) than for the earlier waves (i.e. those with shorter

latencies). Such a trend is further supported by the finding

of a decreased V/I amplitude ratio at all click rates. Clearly,

the suppression of neuronal activity along the auditory

pathway in neonates with hyperbilirubinemia is more se-

vere for the more rostral (or central) regions of the audi-

tory brainstem than for the more caudal (or peripheral)

regions.

In our neonates with hyperbilirubinemia, wave III

amplitude was reduced at all click rates, with the most

significant reduction occurring at the highest rate 910/s.

Similarly, wave V amplitude was reduced at all click rates,

but the reduction is particularly significant at the very high

rates 455/s and 910/s, which cannot be achieved in con-

ventional BAER. Therefore, the amplitude reduction in the

two MLS BAER waves was true of all repetition rates of click

stimuli, and was more significant at higher than at lower

rates of clicks. On the other hand, the significant reduction

in wave I amplitude occurred only at the highest rate 910/s.

The V/III amplitude ratio was significantly decreased only at

the very high rates 455/s and 910/s. It appears that the

impaired brainstem auditory neurons due to bilirubin

neurotoxicity tends to be more vulnerable to higher rates of

click stimuli, that is, vulnerable to more stressful physio-

logical or temporal challenge of acoustic stimulation.

In jaundiced Gunn rat, Rice and Shapiro found that

following injection of biliverdin there was a significant cor-

relation between total bilirubin and wave III amplitude,

suggesting that the amplitude reduction of BAER wave III

was closely related to the level of bilirubin.1 In a previous

study of conventional BAER in human infants with hyper-

bilirubinemia, we found that the amplitudes of BAER waves I,

III and Vwere generally lower at a higher level of TSB than at a

lower level. However, none of these amplitudes was corre-

lated significantly with the level of TSB. Neither was the V/I

amplitude ratio nor the V/III amplitude ratio. In the present

study of term infants with hyperbilirubinemia, there was a

similar finding in conventional BAER (21/s clicks); all wave

amplitudes tended to be reducedwith the increase in the level

of TSB, but only wave I amplitude was correlated with the

level of TSB whereas the amplitudes of waves III and V were

not. In MLS BAER, however, the amplitudes of MLS BAER

waves I, III and V were all reduced with the increase in the

level of TSB. All wave amplitudes were correlated significantly

with TSB level at some ormost click rates, though not all rates.

This was also the case for the V/I and V/III amplitude ratios.

Therefore, the severity of auditory brainstem impairment in

hyperbilirubinemia is related to, though not completely in

parallel with, the severity of hyperbilirubinemia.

In this study, any infants who had any perinatal problems

or complications that may significantly affect neonatal BAER

had been excluded. In particular, none of our subjects had any

sign of hypoxia-ischemia d the other perinatal problem that

can result in a major amplitude reduction in MLS BAER wave

components.21,27 Thus, any confounding effects on the study

results had been minimized. There were also no significant

differences between the study and control groups in gesta-

tional age, postconceptional age (the age at time of MLS BAER

recording), and click intensity above BAER threshold of the

subjects at which MLS BAER measurements were obtained.

Any of these potential differences were further minimized by

ANCOVA. Therefore, the baseline data in the study and con-

trol groups are comparable except for their TSB levels. The

changes in MLS BAER wave amplitudes in the neonates with

hyperbilirubinemia were generally correlated well with the

bilirubin level. All these indicate that the major differences in

MLS BAER variables between the study and control groups

were due to the major difference in bilirubin levels, i.e. the

MLS BAER abnormalities found in the study group were

caused by hyperbilirubinemia.

The reduction inMLSBAERwaves I, III andparticularlyVwas

consistently seen at all click rates between 91/s and 910/s,

particularly higher rates. In addition, the changes in wave am-

plitudes were generally correlated well with the bilirubin level.

Obviously,waveamplitudes, particularlywaveVamplitude,are

valuable inassessmentofbilirubinneurotoxicity to theauditory

brainstem. A recent diffusion tensor imaging study in infants

found that, BAERwave V amplitude, alongwith the latencies of

waves I, III, and V, intervals of IeIII and IeV intervals, was

correlatedwith fractional anisotropy of the inferior colliculus.42

Such a correlation between the anatomical structure (inferior

colliculus) and the electrophysiological activity (BAER wave V

amplitude) supports our view thatwave V amplitude is a useful

BAER variable to assess functional integrity of the auditory

pathway and the brainstem.16e19,21e24,27

WithMLS BAER, we have studied functional integrity of the

developing auditory brainstem in a range of neonatal prob-

lems. A major reduction in the amplitudes of MLS BAER wave

components, mainly wave V, was only seen in neonates after

perinatal hypoxia-ischemia.21,27 The present study demon-

strate that there is also a major reduction in BAER wave am-

plitudes in neonatal hyperbilirubinemia. It seems that BAER

wave amplitudes in neonates are particularly vulnerable to

perinatal hypoxic-ischemic injury and bilirubin neurotoxicity.

The majority of our subjects were followed for several

months after the initial MLS BAER recording. The remaining

was lost from the follow-up mainly due to moving to remote

area with their parents after discharge. The preliminary and

incomplete data showed that following phototherapy and/or

exchange transfusion the amplitude reduction recovered

quickly (from several hours to days) in most neonates with

hyperbilirubinemia, though more slowly in a few. This sug-

gests that the suppressed brainstem auditory electrophysi-

ology is largely transient, and is reversible following prompt

treatment. Persistent amplitude reduction and, in turn,

persistent suppression in brainstem auditory electrophysi-

ology are rare, as long as the patient can be treated promptly.

Most of our subjects were not associated with any significant

adverse neurodevelopment. Only 4 had less favorable

e u r o p e a n j o u r n a l o f p a e d i a t r i c n e u r o l o g y 1 8 ( 2 0 1 4 ) 1 9 3e2 0 0 199

neurodevelopment, with 2 had persistent amplitude reduc-

tion in MLS BAER waves III and V. More detailed follow-up is

warranted to explore the prognosis of the suppressed brain-

stem auditory electrophysiology in neonates with

hyperbilirubinemia.

In summary, the present MLS BAER study found that audi-

tory electrophysiology in the neonatal brainstem is suppressed

inhyperbilirubinemia.Thesuppression is related to theseverity

ofhyperbilirubinemia.WaveamplitudesarevaluableMLSBAER

variables to detect functional impairment of the brainstem and

auditory pathway in hyperbilirubinemia. We propose that, in

analysisof theBAERforneonateswithhyperbilirubinemia, as in

neonates with perinatal hypoxia-ischemia, wave amplitudes

should be included, along with wave latencies and interpeak

intervals, to assess brainstem auditory function. In neonates

who are suspected of bilirubin encephalopathy, BAER wave

amplitudes, particularly wave V amplitude, can help early

detection of brain damage due to bilirubin neurotoxicity.

Acknowledgments

We thank the doctors and nurses in the units for their support

and assistance in recruiting of subjects and collecting of data.

The research was supported by the WellChild, and Medical

Sciences Division, Oxford University, UK.

r e f e r e n c e s

1. Rice AC, Shapiro SM. Biliverdin-induced brainstem auditoryevoked potential abnormalities in the jaundiced Gunn rat.Brain Res 2006;1107:215e21.

2. Shapiro SM. Bilirubin toxicity in the developing nervoussystem. Pediatr Neurol 2003;29:410e21.

3. Shapiro SM. Chronic bilirubin encephalopathy: diagnosis andoutcome. Semin Fetal Neonatal Med 2010;15:157e63.

4. Dennery PA, Seidman DS, Stevenson DK, Neonatal HBN. NEngl J Med 2001;344:581e90.

5. Ives K. Neonatal jaundice. In: Rennie JM, Roberton NRC,editors. Textbook of Neonatology. Edinburgh: Churchill-Livingstone; 1999. p. 715e32.

6. Newman TB, Liljestrand P, Jeremy RJ, Ferriero DM, Wu YW,Hudes ES, Escobar GJ. Outcomes among newborns with totalserum bilirubin levels of 25 mg per deciliter or more. N Engl JMed 2006;354:1889e900.

7. Watchko JF. Neonatal hyperbilirubinemia e what are therisks? N Engl J Med 2006;354:1947e9.

8. Ahlfors CE, Parker AE. Unbound bilirubin concentration isassociated with abnormal automated auditory brainstemresponse for jaundiced newborns. Pediatrics 2008;121:976e8.

9. Ahlfors CE, Amin SB, Parker AE. Unbound bilirubin predictsabnormal automated auditory brainstem response in adiverse newborn population. J Perinatol 2009;29:305e9.

10. Amin SB, Ahlfors C, Orlando MS, Dalzell LE, Merle KS,Guillet R. Bilirubin and serial auditory brainstem responses inpremature infants. Pediatrics 2001;107:664e70.

11. Okumura A, Kidokoro H, Shoji H, Nakazawa T, Mimaki M,Fujii K, Oba H, Shimizu T. Kernicterus in preterm infants.Pediatrics 2009;123:e1052e8.

12. Smith CM, Barnes GP, Jacobson CA, Oelberg DG. Auditorybrainstem response detects early bilirubin neurotoxicity atlow indirect bilirubin values. J Perinatol 2004;24:730e2.

13. Wennberg RP, Ahlfors CE, Bhutani VK, Johnson LH,Shapiro SM. Toward understanding kernicterus: a challengeto improve the management of jaundiced newborns.Pediatrics 2006;117:474e85.

14. Shapiro SM, Popelka GR. Auditory impairment in infants atrisk for bilirubin-induced neurologic dysfunction. SeminPerinatol 2011;35:162e70.

15. Jiang ZD. Brainstem auditory impairment in infants withhyperbilirubinemia. In: Novotny JF, Sedlacek F, editors.Bilirubin: chemistry, regulation and disorder. New York: NovaScience Publishers; 2012. p. 235e55.

16. Jiang ZD. Maturation of peripheral and brainstem auditoryfunction in the first year following perinatal asphyxia e alongitudinal study. J Speech Lang Hear Res 1998;41:83e93.

17. Jiang ZD. Outcome of brainstem auditory electrophysiology inchildren who survived purulent meningitis. Ann Otol RhinolLaryngol 1999;108:429e34.

18. Jiang ZD, Tierney TS. Long-term effect of perinatal andpostnatal asphyxia on developing human auditory brainstemresponses: brainstem impairment. Int J Pediatr Otorhinolaryngol1996;34:111e27.

19. Jiang ZD, Shao XM, Wilkinson AR. Changes in BAERamplitudes after perinatal asphyxia during the neonatalperiod in term infants. Brain Dev 2006;28:554e9.

20. Hansen TWR, Cashore WJ, Oh W. Changes in piglet auditorybrainstem response amplitudes without increases in serumor cerebrospinal fluid neuron-specific enolase. Pediatr Res1992;32:524e9.

21. Jiang ZD, Brosi DM, Shao XM, Wilkinson AR. Sustaineddepression of brainstem auditory electrophysiology duringthe first month in term infants after perinatal asphyxia. ClinNeurophysiol 2008;119:1496e505.

22. Jiang ZD, Brosi DM, Chen C, Wilkinson AR. Brainstem auditoryresponse amplitudes in neonatal chronic lung disease anddifferences from perinatal asphyxia. Clin Neurophysiol2009;120:967e73.

23. Jiang ZD, Brosi DM, Wilkinson AR. Depressed brainstemauditory electrophysiology in preterm infants after perinatalhypoxia-ischemia. J Neurol Sci 2009;281:28e33.

24. Jiang ZD, Yin R, Wilkinson AR. Amplitude reduction inbrainstem auditory response in term infants under neonatalintensive care. Clin Neurophysiol 2013;124:1470e6.

25. Funato M, Tamai H, Shimada S, Nakamura H. Vigintiphobia,unbound bilirubin, and auditory brainstem responses.Pediatrics 1994;93:50e3.

26. Funato M, Teraoka S, Tamai H, Shimida S. Follow-up study ofauditory brainstem responses in hyperbilirubinemicnewborns treated with exchange transfusion. Acta PaediatrJpn 1996;38:17e21.

27. Jiang ZD. Maximum length sequence technique improvesdetection of brainstem abnormalities in infants. In:Lawson PN, McCarthy EA, editors. Pediatric neurology. NewYork: Nova Science Publishers; 2012. p. 1e38.

28. Wilkinson AR, Jiang ZD. Brainstem auditory evoked responsein neonatal neurology. Semin Fetal Neonatol Med2006;11:444e51.

29. Jiang ZD. Brainstem electrophysiological changes afterperinatal hypoxia ischemia. In: Hamalainen E, editor. Newtrends in brain hypoxia ischemia research. New York: NovaScience Publishers; 2008. p. 203e20.

30. Jiang ZD, Brosi DM, Wang J, Xu X, Chen GQ, Shao XM,Wilkinson AR. Time course of brainstem pathophysiologyduring first month in term infants after perinatal asphyxia,revealed by MLS BAER latencies and intervals. Pediatr Res2003;54:680e7.

31. Wilkinson AR, Brosi DM, Jiang ZD. Functional impairment ofthe brainstem in infants with bronchopulmonary dysplasia.Pediatrics 2007;120:362e71.

e u r o p e a n j o u r n a l o f p a e d i a t r i c n e u r o l o g y 1 8 ( 2 0 1 4 ) 1 9 3e2 0 0200

32. Lasky RE. Rate and adaptation effects on the auditory evokedbrainstem response in human newborns and adults. Hear Res1997;111:165e76.

33. Chen YJ, Kang WM. Effects of bilirubin on visual evokedpotentials in term infants. Eur J Pediatr 1995;154:662e6.

34. Chen WX, Wong V. Visual evoked potentials in neonatalhyperbilirubinemia. J Child Neurol 2006;21:58e62.

35. Jiang ZD, Brosi DM, Wilkinson AR. Changes in BAER waveamplitudes in relation to total serum bilirubin level in termneonates. Eur J Pediatr 2009;168:1243e50.

36. Bell SL, Smith DC, Allen R, Lutman ME. The auditory middlelatency response, evoked using maximum length sequencesand chirps, as an indicator of adequacy of anesthesia. AnesthAnalg 2006;102:495e8.

37. Nagle S, Musiek FE. Morphological changes in the middlelatency response using maximum length sequence stimuli. JAm Acad Audiol 2009;20:492e502.

38. Picton TW, Champagne SC, Kellett AJC. Human auditorypotentials recorded using maximum length sequences.Electroencephalogr Clin Neurophysiol 1992;84:90e100.

39. Jiang ZD. Damage of chronic sublethal hypoxia to theimmature auditory brainstem. In: Fiedler D, Krause R, editors.Deafness, hearing loss, and the auditory system. New York: NovaScience Publishers; 2010. p. 159e80.

40. Hansen TWR, Bratlid D, Walaas SI. Bilirubin decreasesphosphorylation of synapsin I, a synaptic vesicle-associatedneuronal phosphoprotein, in intact synaptosomes from ratcerebral cortex. Pediatr Res 1988;23:219e23.

41. Hansen TWR, Paulsen O, Gjerstad L, Bratlid D. Short-termexposure to bilirubin reduces synaptic activation in rattransverse hippocampal slices. Pediatr Res 1988;23:453e6.

42. PIPARI Study Group. Diffusion tensor imaging of the inferiorcolliculus and brainstem auditory-evoked potentials inpreterm infants. Pediatr Radiol 2009;39:804e9.