Introduction
Neonatal hearing screening in newborns with complications should be performed using
the evoked otoacoustic emissions (OAE) and auditory brainstem response (ABR) because
they are complementary tools and together provide a complete evaluation of the auditory
system.[1]
[2] The ABR test is a noninvasive neurophysiologic assessment of brainstem maturation
in babies and could be a useful electrophysiologic test to verify neuronal myelination
in preterm infants.[3]
Furthermore, the ABR has shown fewer false-positives and lower referral rates compared
with OAE, is less sensitive to noise and middle ear disorders even in very premature
infants, and is an essential tool in the diagnosis of hearing loss in the pediatric
population.[4]
[5]
[6]
[7]
The ABR in neonates may have a high variability of response. The frequently observed
prolonged latencies can be considered normal due to the characteristics of the maturation
process of the auditory system, explained by the hypothesis that the structures that
generate the ABR core components take more time to completely mature.[8]
[9]
[10] Therefore, the literature remains unclear whether these changes in the ABR responses
would be permanent, thus indicating hearing loss, or whether they could be normalized
with increasing age.[11]
[12]
[13]
[14]
According to Kohelet et al,[15] the ABR response in full-term and preterm newborns of the same chronological age
was similar, regardless of gestational age. Nevertheless, Sleifer et al observed that
absolute and interpeak latencies were statistically different between 4-month-olds
and 12-month-olds.[16] At 20 months, only wave I was similar. The authors concluded that the maturation
of the auditory pathway occurs in a different way between full-term and preterm newborns,
and that gestational age must be considered in the ABR analysis, mainly in infants
younger than 20 months. In a recent study, Roopakala et al compared the ABR results
in 25 preterm and 25 full-term infants and observed a significant increased latency
of ABR waveform V in preterm infants.[3]
However, the ABR response in full-term and preterm newborns may be impacted by neonatal
complications that are considered of risk for hearing loss and, consequently, may
delay the maturation process.[17] Most of the studies show the influence of neonatal complications that are considered
of risk for hearing loss in the ABR results. Some authors reported that neonates with
transient low Apgar scores had a significant increase in I to V interval at very high
click rates in the first 3 days of life.[11] In another study, very low-birth-weight infants had prolonged interpeak III to V.[12] The ABR in extremely preterm infants with bronchopulmonary dysplasia showed a significant
increase in wave V and interpeak intervals I to V and particularly III to V.[13] Extremely low birth weight and mechanical ventilation may, as well, cause abnormal
ABR.[14]
The question is whether these ABR findings will be recovered with increasing age or
neurodevelopment. The objective of this study was to understand the real influence
of neonatal complications considered of risk for hearing loss, in infants, using the
sequential ABR evaluation.
Methods
Study Population
This historical cohort study was conducted in a tertiary referral center from October
2008 to October 2010. The inclusion criteria were: treatment at the Universal Neonatal
Hearing Screening Program of the local hospital; at least one risk indicator for hearing
loss according to the Joint Committee on Infant Hearing[18]; presence in both evaluations (the first upon neonatal unit discharge, and the second
at 6 months of age); all the ABR latencies; transient otoacoustic emissions confirmed
in both ears; consent form signed by the newborn's parents or guardians.
For comparison, gestational age was divided into three categories: term ≥ 37 weeks,
premature between 31 and 36 weeks, and extremely premature < 31 weeks.
Audiological Assessment
For the ABR analysis, the rarefaction click stimulus was presented by the 3Ω insertion
phone, with intensity of 80-dB nHL (normal Hearing Level) and a presentation rate
of 20.1 c/s (clicks/seconds) with a bandpass filter of 100 and 3,000 Hz and average
of 1,024 stimuli on Interacoustics EP15 Eclipse (Interacoustics A/S, Assens, Denmark).
Duplicate recordings were made in response to each stimulus condition to examine reproducibility.
The ABR was captured through electrocardiogram disposable electrodes (Neuroline, Ambu
A/S, Baltorpbakken, Denmark), after cleaning the skin with electrocardiogram/electroencephalograph
abrasive gel (NUPREP, Weaver and Company, Aurora, USA). The impedance level was kept
between 1 and 3 kΩ for the electrodes; the active electrode was positioned in Fz, the reference electrode in M1 and M2, and the ground electrode in Fpz. No sedatives were used.
Variable and Statistical Analyses
Predictor variables included gender, gestational age, birth weight < 1,500 g, low
Apgar score, infection, intensive care unit (ICU) stay, hyperbilirubinemia, peri-intraventricular
hemorrhage, use of mechanical ventilation and ototoxic medication. Outcome variables
included absolute latencies (I, III, and V) and interpeak latencies (I to III, III
to V, and I to V) of the ABR in both ears.
ABR change (in percentage) at 6 months of age (moment 2) was calculated in relation
to the ABR performed upon hospital discharge (moment 1) using the following expression:
ABR at 6 months of age − ABR postdischarge / ABR postdischarge 0.100%.
The relation between risk indicator for hearing loss and variation of ABR were analyzed
using Mann-Whitney and Kruskal-Wallis test, followed by Dunn for multiple comparisons
of the gestational age. Statistical analysis was performed with SPSS v15 and Graph
Pad Prism v5 software (SPSS Inc., Chicago, IL, USA). A p value of < 0.05 was accepted as statistically significant.
Ethics
This study was approved by the Research Ethics Committee, process no. 402/08.
Results
A total of 114 newborns, 51 (45%) girls and 63 (55) boys, met the inclusion criteria.
Other sample data are shown in [Table 1].
Table 1
Description of the 114 newborn according to gestational age and risk indicators for
hearing loss
|
Variables
|
n (%)
|
|
Gestational age < 31 wk
|
29 (25%)
|
|
Gestational age 32–36 wk
|
49 (43%)
|
|
Gestational age ≥37 wk
|
36 (32%)
|
|
Birth weight <1,500 g: no
|
79 (69%)
|
|
Birth weight <1,500 g: yes
|
35 (31%)
|
|
Apgar < 4 (1 min): no
|
59 (52%)
|
|
Apgar < 4 (1 min): yes
|
55 (48%)
|
|
Apgar < 6 (5 min): no
|
94 (82%)
|
|
Apgar < 6 (5 min): yes
|
20 (18%)
|
|
Infection: no
|
71 (62%)
|
|
Infection: yes
|
43 (38%)
|
|
Intensive care units: no
|
38 (33%)
|
|
Intensive care units: yes
|
76 (67%)
|
|
Hyperbilirubinemia: no
|
109 (96%)
|
|
Hyperbilirubinemia: yes
|
05 (4%)
|
|
Peri-intraventricular hemorrhage: no
|
101 (89%)
|
|
Peri-intraventricular hemorrhage: yes
|
13 (11%)
|
|
Mechanical ventilation: no
|
53 (46%)
|
|
Mechanical ventilation: yes
|
61 (54%)
|
|
Ototoxic drugs: no
|
74 (65%)
|
|
Ototoxic drugs: yes
|
40 (35%)
|
[Table 2] shows that Apgar score lower than 6 in the first minute, gestational age, and ICU
stay best correlated with the variation of the absolute ABR latencies (p < 0.05).
Table 2
The p values of the relation between the variation of ABR absolute latencies and gestational
age as well as risk indicators for hearing loss
|
Characteristics
|
Right ear
|
Left ear
|
|
I
|
III
|
V
|
I
|
III
|
V
|
|
Apgar < 6 (5 min): no (n = 94) × yes (n = 20)
|
0.433
|
0.016
|
0.164
|
0.032
|
0.006
|
0.026
|
|
GA: <31 wk (n = 29) × 32–36 wk (n = 49) × ≥ 37 wk (n = 36)
|
0.018[a]
|
0.683
|
0.217
|
0.043
|
0.278
|
0.258
|
|
ICU: no (n = 38) × yes (n = 76)
|
0.320
|
0.025
|
0.815
|
0.134
|
0.157
|
0.907
|
|
BW < 1,500 g: no (n = 79) × yes (n = 35)
|
0.110
|
0.636
|
0.606
|
0.054
|
0.654
|
0.958
|
|
Apgar < 4 (1 min): no (n = 59) × yes (n = 55)
|
0.814
|
0.195
|
0.107
|
0.187
|
0.171
|
0.053
|
|
Infection: no (n = 71) × yes (n = 43)
|
0.255
|
0.338
|
0.309
|
0.684
|
0.312
|
0.205
|
|
Hyperbilirubinemia: no (n = 109) × yes (n = 5)
|
0.841
|
0.668
|
0.447
|
0.140
|
0.175
|
0.533
|
|
PIVH: no (n = 101) × yes (n = 13)
|
0.107
|
0.206
|
0.380
|
0.083
|
0.820
|
0.199
|
|
MV: no (n = 53) × yes (n = 61)
|
0.084
|
0.154
|
0.360
|
0.201
|
0.666
|
0.189
|
|
Ototoxic drugs: no (n = 74) × yes (n = 40)
|
0.764
|
0.960
|
0.204
|
0.667
|
0.838
|
0.268
|
|
Sex: F (n = 51) × M (n = 63)
|
0.468
|
0.346
|
0.174
|
0.934
|
0.401
|
0.130
|
Abbreviations: ABR, auditory brainstem response; BW, birth weight; GA, gestational
age; ICU, intensive care unit; MV, mechanical ventilation; PIVH, peri-intraventricular
hemorrhage.
a < 31 ≠ ≥ 37 (Dunn test for multiple comparisons).
[Table 3] shows that ICU stay, peri-intraventricular hemorrhage, mechanical ventilation, and
gestational age showed the best relationship to the interpeak variation (p < 0.05).
Table 3
The p values of the relation between the variation of ABR interpeaks and gestational age
as well as risk indicators for hearing loss
|
Characteristics
|
Right ear
|
Left ear
|
|
I–III
|
III–V
|
I–V
|
I–III
|
III–V
|
I–V
|
|
ICU: no (n = 38) × yes (n = 76)
|
0.003
|
0.004
|
0.907
|
0.022
|
0.117
|
0.514
|
|
PIVH: no (n = 101) × yes (n = 13)
|
0.459
|
0.077
|
0.070
|
0.297
|
0.028
|
0.026
|
|
MV: no (n = 53) × yes (n = 61)
|
0.006
|
0.558
|
0.079
|
0.162
|
0.282
|
0.041
|
|
GA: <31 (n = 29) × 32–36 (n = 49) × ≥ 37 (n = 36)
|
0.066
|
0.006[a]
|
0.984
|
0.434
|
0.307
|
0.767
|
|
BW < 1,500 g: no (n = 79) × yes (n = 35)
|
0.252
|
0.324
|
0.842
|
0.468
|
0.580
|
0.144
|
|
Apgar < 4 (1 min): no (n = 59) × yes (n = 55)
|
0.220
|
0.203
|
0.065
|
0.527
|
0.226
|
0.276
|
|
Apgar < 6 (5 min): no (n = 94) × yes (n = 20)
|
0.159
|
0.890
|
0.200
|
0.994
|
0.127
|
0.329
|
|
Infection: no (n = 71) × yes (n = 43)
|
0.944
|
0.309
|
0.781
|
0.617
|
0.420
|
0.276
|
|
Hyperbilirubinemia: no (n = 109) × yes (n = 5)
|
0.410
|
0.609
|
0.171
|
0.934
|
0.678
|
0.552
|
|
Ototoxic drugs: no (n = 74) × Yes (n = 40)
|
0.891
|
0.059
|
0.311
|
0.854
|
0.152
|
0.381
|
|
Sex: F (n = 51) × M (n = 63)
|
0.180
|
0.514
|
0.067
|
0.268
|
0.227
|
0.065
|
Abbreviations: ABR, auditory brainstem response; BW, birth weight; GA, gestational
age; ICU, intensive care unit; MV, mechanical ventilation; PIVH, peri-intraventricular
hemorrhage.
a < 31 ≠ 32–36; <31 ≠ ≥ 37 (Dunn test for multiple comparisons).
Discussion
The objective of our cohort study was to verify the real influence of the neonatal
complications considered of risk for hearing loss, by analyzing the variation of the
ABR parameters between two moments: at hospital discharge and at 6 months of age.
This second evaluation moment was chosen based on the need of an early diagnosis of
hearing loss and intervention by 6 months of age. Obviously, children with a history
of hearing risk indicators are more likely to have hearing impairment and then require
periodic evaluation even if they had an initial normal auditory assessment.[18]
[19]
Newborns with risk indicators for hearing loss have some coexistent neonatal complications.
However, such complications are not considered by most of the studies, which compare
only one risk indicator in healthy babies in only one moment or in the first days
of life.
This study intended to show the real influence of risk indicators coexisting with
prematurity or not and yet clarify whether the findings in the first evaluation remain
abnormal in the second one even with increasing age, considering the present risks.
Therefore, in this study, an ABR follow-up was performed in two moments, the first
at hospital discharge and the second at 6 months of age, to assess whether these possible
alterations are transient or permanent.
Results show that ABR latencies and interpeaks were higher in extremely preterm infants
than in full-term newborns, confirming other published results.[3]
[16]
[17] Apparently, this difference decreases with increasing age, possibly due to the maturational
process.[16] Sleifer et al reported differences between these children, with very similar results
at 24 months of age.[16] Furthermore, Turchetta et al observed an improvement over time of the estimated
hearing threshold in ABR follow-up in preterm and full-term newborns.[20]
In the second assessment, when we analyzed the percentage of changes between these
two moments, extremely premature infants showed greater reduction in the absolute
latency of wave I in both ears. Despite the statistical significance, this finding
has no clinical relevance, because the values are similar to those in older children
and healthy adults.[10]
[17] This result shows that the peripheral transmission matures faster as compared with
subsequent waveforms and the same reasoning cannot be applied to reductions of interpeaks
I to III, III to V, and I to V between the two assessments due to improper myelination
of auditory pathway and improper efficacy of higher-order neurons in infant population.[21]
The percentage of change in interpeak latencies I to III and III to V was higher in
preterm infants than in full-term infants, giving the impression that auditory maturation
occurs more rapidly in preterm infants. Both showed change in the ABR values between
the assessments, characterized by reduced latencies due to the maturation process.
But in the second assessment, preterm infants had higher values, showing that the
improvement might be related to the maturation of the auditory pathways that was not
complete at birth and changed according to gestational age.[20]
[22]
On the other hand, not only gestational age may cause ABR alterations, but other neonatal
complications may as well contribute to auditory maturation impairment. Despite being
frequently studied, both aspects are not considered concurrently.
These studies evaluated healthy babies, without risk indicators for hearing loss,
although it is very common that preterm neonates have many associated risk factors,
hence the need for a detailed analysis of the factors to which preterm infants may
have been exposed and that may have contributed to these remaining differences after
the end of maturation.
In this study, newborns whose Apgar score was less than 4 at 1 minute after birth
and normal after 5 minutes showed no significant increase in the ABR latencies, but
those infants who had a low Apgar score at 5 minutes after birth had significant differences
in the absolute latency of wave III in both ears.
According to these findings, the low Apgar score at 5 minutes caused worse damages
in the auditory pathway, despite increasing age. Jiang et al observed an increase
in interpeak I to V only within the first 3 days of life, as well as normal results
within 1 month of life for the newborns who had a low Apgar score but had good recovery
after 5 minutes.[14] The authors highlighted that the ABR changes observed within the first 3 days of
life are of little significance because they normalized with the development of the
maturation process. Therefore, a low Apgar score at 1 minute poses low risk for auditory
changes, and a persistently low score may affect the auditory pathways.
An interesting finding in this study that has not been reported in the literature
is that the infants who had a longer ICU stay, without considering the primary status
of such admission, presented clear alterations between the first and second ABR assessments.
Reductions in the absolute latency of wave III in the right ear, interpeak I to III
in both ears, and in interpeak III to V in the left ear were statistically significant
between the first and second assessments.
Other conditions contribute to the ABR differences, such as peri-intraventricular
hemorrhage, maternal infection, chorioamnionitis, and neonatal infection or sepsis.
These conditions have been associated with the development of neonatal brain damage
and adverse neurodevelopmental outcomes.[23]
[24]
[25]
[26] In this study, alterations in the ABR components were observed in these infants,
showing that the occurrence of this injury deserves special attention as its association
with hearing loss is rarely reported in literature.
Likewise, the use of mechanical ventilation, regardless of the primary disease, changes
the ABR recordings and is, therefore, a confounding factor along with other hearing
risk indicators.[14]
[27] This was observed in this study when newborns using mechanical ventilation showed
reduced interpeak I to III for the right ear and reduced interpeak latency I to V
for the left ear.
Nevertheless, other conditions frequently associated with hearing loss, such as low
birth weight, infections, hyperbilirubinemia with serum level requiring transfusion,
and use of ototoxic drugs, showed no significance in the ABR parameters with increasing
age.
Results showed that the waveform I is less sensitive to the injuries that occurred
during pregnancy and/or neonatal complications. However, waves III and V, which characterize
the maturation of axons and synaptic mechanisms in the brainstem level, are more likely
to be affected by the risk analyzed in this study.[28] Therefore, the differences in the ABR responses between full-term and preterm newborns
and with some risk indicators for hearing loss cannot only be due to nervous maturation,
but might as well be caused by injuries occurred during pregnancy and/or neonatal
complications.[17]
We believe that these findings bring an important contribution to clinical practice.
Neonates with complications deserve more attention paid to auditory evaluation, because
prematurity and neonatal complications may interfere in the ABR findings. The interpretation
and/or standardization of the results may be uncertain if only one assessment is performed.
Conclusion
The complications that most impacted the ABR findings were: Apgar scores less than
6 at 5 minutes of birth, gestational age, ICU stay, peri-intraventricular hemorrhage,
and mechanical ventilation.
The sequential auditory evaluation is necessary in preterm and full-term newborns
with risk indicators for hearing loss to correctly identify injuries in the auditory
pathway.
