Download Pediatric ABR Screening: Pass-Fail Rates in Awake versus Asleep

J Am Acad Audiol 2 : 18-23 (1991)
Pediatric ABR Screening: Pass-Fail Rates in
Awake versus Asleep Neonates
Stephan McCall'
John A. Ferraro'
Abstract
The auditory brainstem response (ABR) is commonly used as a neonatal hearing screening
tool . The degree to which myogenic and/or movement artifact can confound the ABR in
neonates, and the effect this has on screening pass-fail rates, although widely recognized,
have not been reported . This study addressed these aspects in a clinical setting . Fifty-two
high-risk neonates were screened in various states of activity (asleep, awake-calm, awakeactive) . Pass-fail rates between asleep and awake babies were significantly different (p <
0.5), with the awake group displaying a much higher failure rate . There was no significant
difference between the awake-calm and awake-active groups . Results indicate that activity
state should be noted and considered along with the other factors that are generally blamed
for false-positive results in neonatal ABR screenings .
Key Words : Auditory brainstem response (ABR), pediatric, screening, pass-fail rate, asleep,
awake-calm, awake-active
ecording of the auditory brainstem
response (ABR) is the most common
R electrophysiologic technique used in the
screening of neonates and infants for hearing loss
(Schulman-Galambos and Galambos, 1979 ; Hyde
et al, 1984 ; Shannon et al, 1984; Durieux-Smith et
al, 1985 ; Fria, 1985 ; Ruth et al, 1985 ; Richmond et
al, 1986 ; Jacobson and Jacobson, 1987 ; Hall et al,
1988 ; Gorga et al, 1989) . Despite its widespread
use, the ABR, like all screening tools, is not infallible . There are several technical, procedural, and
subject-related variables which, working together
or separately, may produce false or misleading
results (Fria, 1980 ; Moore, 1983 ; Weber, 1983 ;
Jacobson, 1985 ; Swigart, 1986 ; Gorga et al, 1989) .
Middle ear disorders, maturation of the brainstem response during the first 2 years of life, collapsing ear canals, and ambient acoustical and/or
electrical noise in the testing environment are examples of factors that may lead to a false-positive
'Doctoral Student, Hearing and Speech Department,
University of Kansas Medical Center
tProfessor and Chairman, Hearing and Speech
Department, Associate Dean, School of Allied Health,
University of Kansas Medical Center
Reprint requests : Stephan McCall, Hearing and
Speech Department, University of Kansas Medical Center,
39th and Rainbow Blvd ., Kansas City, KS 66103
18
finding in an ABR screening test (Weber, 1983 ;
Levette, 1984 ; Jacobson, 1985) . Certainly one of
the more important factors in the successful completion of an examination, however, is the degree
of "cooperation" provided by the patient . In the
case of infant screening, the highest or ultimate
degree is generally achieved when the child is asleep .
Several investigations (e .g ., Amadeo and
Shagass,1973 ; Goff et a1,1977 ; Sohmer et a1,1978 ;
Sanders et al, 1979) have indicated that the ABR
parameters show no significant differences in
awake versus asleep adults and that one is justified in using this procedure in either state . Although activity state may not affect the brainstem
response per se, it certainly influences the amount
of myogenic and/or movement artifact present
during the test, and this is well recognized . By
definition, myogenic artifact is due to muscle action potentials, whereas movement artifact is induced by movement of the head, trunk, and extremities, which weakens or disrupts the electrode
contacts (Jacobson, 1985) . Although the ABR is
highly repeatable and easily recognized in high
quality recording situations, it can be very illusive
if tracings are contaminated with myogenic and/or
movement artifact due to a fussy or restless
patient (Weber, 1983 p . 179) . This, in turn,
Pediatric ABR Screening/McCall and Ferraro
generally necessitates the administration of sedation when measuring the ABR in infants and toddlers . In newborns, however, sedation is generally
not given for a screening examination because of
the inherent risks this poses, and also because the
activity state of the neonate is usually subdued in
comparison to an older child . However, the degree
to which movement/myogenic activity can confound the tracings taken from a comparatively
quiet neonate may be underestimated . If this is
the case, the likelihood of a child failing an ABR
screening examination may be greater if the exam
is performed when the child is awake versus when
he/she is asleep . The present study examined this
hypothesis in a clinical setting .
METHOD
ifty-two medically stable neonates (30 males,
F 22 females) served as subjects . The babies
were outpatients referred from the University of
Kansas Medical Center Neonatal Intensive Care
Unit who were considered to be at risk for hearing
loss in accordance with the specific risk factors advocated by the Joint Committee on Infant Screening (1982) . The mean postgestational age at the
time of testing was 47 .8 weeks (range 37-65) .
Testing was done in an examining room where
ambient noise levels were within those recommended for neonatal auditory screening (Richmond et al, 1986).
ABRs were recorded with a Nicolet "Audit V"
evoked potential system . Stimuli were broad band
clicks (100 psec electrical pulses) delivered directly to the ear canal using a modified impedance tip
adapter coupled to a miniature transducer standard to the Audit V system . Clicks were presented
in alternating polarity at a repetition rate of 33 .3
per second . Electroeneephalographic responses
were recorded as the potential difference between
the forehead and ipsilateral earlobe, with ground
at the contralateral earlobe . Voltages were band
pass filtered with frequency cutoffs of 100 and
3000 Hz (12 dB per octave) . One thousand five
hundred and thirty-four sweeps were electronically averaged to provide a representation of the first
15 msec of electrophysiologic activity (filtered
EEG) following stimulus onset . The signal
averager incorporated an "artifact rejection" system that automatically discarded trials with
amplitudes (peak to peak voltages) exceeding 90
percent full scale deflection .
The subject's state of activity during the exam
was monitored and recorded in three categories .
These included "awake-active" (sucking behavior
and mild movement of the extremities), "awakecalm" (very quiet, relatively motionless), and "asleep ." Classification into a particular activity state
was based on the consensus of the examiner, a
graduate student assistant experienced in pediatric ABR evaluations and the baby's parent/
caregiver . In addition, background electrophysiologic activity was visually monitored for
gross differences between asleep and awake
states . On occasion, a baby judged to be asleep
would begin to move and/or display heightened activity as reflected in the filtered EEG . In these instances, testing would be halted until the EEG
stabilized back into the sleeping pattern . Babies
classified as asleep had to remain asleep
throughout the test to qualify as subjects in this
category . Classification as either awake-calm or
-active was inherently more subjective . If, for example, a baby initially classified as awake-calm
became and stayed active for the majority of runs,
classification was changed to awake-active . If a
baby's activity changed during the course of the
examination and failed to remain stable for the
majority of runs, the results were excluded from
the study . In addition and regardless of activity
state, any trial wherein over 30 percent of the
responses were rejected by the computer was excluded . This value was based on experience in our
clinic . Accepting responses with higher artifact
rejection would most likely lead to higher failure
rates .
III
LATENCY (MS)
Figure 1 Normal ABR tracings recorded from an asleep
infant . Waves I, III, and V are present at normal absolute
and interwave latencies at 60 and 30 dB HL .
Journal of the American Academy of Audiology/Volume 2, Number 1, January 1991
Screening protocol called for measurements at
Hearing Levels of 60 dB and 30 dB per ear with one
replication at each level . Figure 1 shows an example
of a normal response recorded from a subject who
was asleep when tested. The baby was considered
to have passed the screening if a repeatable wave
V was observed at normal and symmetrical latency
values at 30 dB HL in both ears (based on age-dependent laboratory norms) . Final judgments of
pass or fail were determined from hard copy recordings of the tracings by an examiner who was unaware of the subject's activity state .
Table 1 Summary of Referral Criteria and Pass/Fail
Results for All Subjects
Category
Craniofacial abnormality
Family history
Referred
Fail
Pass
2
0
0
0
2
0
2
1
Low birthweight
22
8
14
Ototoxic medication
14
4
10
Hyperbillrubinemia
Viral infection
Breathing difficulty
Multiple
Total
0
3
9
52
1
0
0
0
3
6
3
19 (37%) 33 (63%)
RESULTS
esults were obtained from 30 male and 22
R female neonates . Table 1 summarizes referral criteria and the pass-fail results of all subjects .
The activity state of each baby was recorded as
either "awake-active," "awake-calm," or "asleep ."
Eleven additional babies (17%) were excluded because activity state fluctuated during the ABR examination and exclusive placement into one of the
three defined categories was not possible .
Examples of tracings recorded in each of the
three activity states are illustrated in Figures 2
and 3 . A visual comparison indicates the possible
effects of increased myogenic/movement artifact
on the clarity of the ABR tracing. Figure 2 shows
recordings taken from a baby tested while awake
and calm and then again when asleep . The
"awake-calm" recording produced observable components at both intensity levels but morphology
was poorly defined . However, when the same baby
was tested when asleep, the waveforms became
well defined and repeatable . The disparity between awake versus asleep states was even more
A
B
LATENCY (MS)
20
apparent when recordings were taken in the
"awake-active" state as shown in Figure 3 . Identifiable components were observed at 60 dB HL,
but at 30 dB HL the waveform was distorted and
the components were difficult to define resulting
in an evaluation of "Fail ." When this same baby
was tested in the "asleep" state (Fig . 3), recordings
were well defined and normal (Pass) .
Of the 52 neonates included in this study, 28
were screened while asleep . Twenty-five passed
and 3 failed . Twenty-four babies were screened in
the awake state . Eight passed and 16 failed . Table
2 indicates Chi-Square analysis for awake-asleep
and pass-fail conditions . Results of these data indicated a significant difference (p < 0 .05) in passfail rates between babies who were screened in
awake versus asleep states . This suggests that an
awake baby will be more likely to fail an ABR
screening than one who is asleep .
In the awake state, babies were classified as
either "awake-active" or "awake-calm ." This was
done to determine if seemingly subtle changes in
Figure 2 ABR tracings
recorded from the same infant first in the "awakecalm" state (A) and then in
the "asleep" state (B). Comparison shows the asleep
tracings to be more clearly
defined and repeatable .
Pediatric ABR Screening/McCall and Ferraro
A
Figure 3
ABR tracings
recorded from the same infant first in the "awake-active" state (A) and then in
the "asleep" state (B). Comparison shows the asleep
tracings to be more clearly
defined and repeatable
(Pass) . Interpretation of the
awake-active tracings resulted in an evaluation of
"Fail."
B
1 .5
L5
LATENCY (ms)
activity state could vary enough in awake neonates
to significantly affect interpretation and consequently determination of pass-fail . Of the 24 awake
babies, 14 were screened in the "awake-active"
state, 4 passed and 10 failed . Ten babies were
screened in the "awake-calm" state, four passed and
six failed . Table 3 indicates chi-square analysis for
awake-active, awake-calm, and pass-fail conditions . Results of these data revealed no significant
differences in pass-fail rate between awake-active
and awake-calm babies (p > 0 .05) .
This study was primarily concerned with
neonates assessed in a clinical setting using conventional ABR screening procedures . Ideally, we
would have preferred to test each baby in all three
activity states . However, we were not at liberty to
administer sedation at will because of the inherent risk factors associated with doing this, nor
would it have been realistic to maintain all three
activity states for a full recording session or during
three separate sessions . Currently, we are in the
process of collecting long-term follow-up data on
each of these subjects . However, during the course
of the study, a certain number of babies who were
initially screened while awake and failed were
seen again . Of the 16 awake-fails, eight were seen
again 4 to 8 weeks later . Six were tested when asleep, five passed and only one failed . The other
two were awake for the second screening and both
again failed . This information suggests that five
of six babies who were initially screened while
awake might have been false-positives . Since a
time period of 4 to 8 weeks elapsed between initial
and follow-up screenings, the possibilities of fluctuating or resolved hearing loss and/or maturational effects are very real . Although six subjects
represent a small sample size, the number of possible false-positives is large, again suggesting that
Table 2 Chi-Square Analysis for Awake versus
Asleep Pass/Fail Conditions
Table 3 Chi-Square Analysis for Awake-Calm versus
Awake-Active Pass/Fail Conditions
ASLEEP
PASS
FAIL
25
AWAKE
8
17 .8
3
28
33
152
16
10,2
X2=172
AWAKE
I
I
8.8
24
19
52
p<005
CALM
PASS
FAIL
10
4
3 .3
i
6 .7
10
ACTIVE
4 .7
8
X2 = .45
14
9 .3
16
'I
p > 0 .05
24
Journal of the American Academy of Audiology/Volume 2, Number 1, January 1991
myogenic/movement artifact from an awake baby
may have contributed to poor definition of the ABR
waveform .
DISCUSSION AND CONCLUSIONS
ased on the results of several studies, failure
B rates in pediatric ABR screening programs
generally range from 10 percent to 25 percent
(Fria, 1985) . Some reports have indicated failure
rates as low as 5 percent (Schulman-Galambos
and Galambos, 1979), whereas others observed
failure rates near 60 percent (Roberts et al, 1982) .
Differences in testing apparati and protocol as
well as pass/fail criteria certainly contributed to
this disparate range . However, one aspect that is
certainly well recognized but generally not
reported is the effect(s) of subject activity state on
the outcome of an examination . In the present
study, the initial overall failure rate of the 52
neonates tested was approximately 37 percent
(19/52), which is on the "high" side . Failure rate
for those subjects tested awake was approximately
67 percent (16/24), which is extremely high . On
the other hand, the failure rate for those babies
tested when asleep was approximately 10 percent
(3/28), which is low .
The use of a 30 dB HL cut-off may have contributed to our high failure rate for awake babies .
We chose this level to strengthen test sensitivity .
Those studies utilizing a 40-dB criterion generally
report lower failure rates . This has prompted us
to conduct a comparative study regarding this
issue, which is currently ongoing. In addition, certain features inherent to the Audit V's artifact
rejection system may have contributed to poor
waveform morphology in the active infants .
Despite the above caveats, our findings suggest that increased muscle or movement activity
of an awake neonate may contaminate the ABR
enough to lead to a false-positive result in a screening examination . The effects of even subtle activity
in a seemingly calm but awake baby should not be
underestimated . In several cases, ABRs from
babies who were screened in an awake but quiet
state were as difficult to interpret as those recorded from a more active baby . On occasion, these
same babies fell asleep and the once difficult to
interpret waveforms became well defined and
easily recognized .
The above findings are certainly not surprising and, in many respects, the results of this study
simply reaffirm something that is widely recognized . What is surprising, however, is that the ac-
22
tivity state of the neonate is rarely reported or acknowledged in the literature to be a cause for falsepositive findings in screening examinations .
Ideally, the baby should be screened while in
natural sleep ; however, this is not always practical
or feasible . If a baby is awake and fails the screening, steps should be taken to facilitate sleep when
the retest is performed and time allowed for this .
These efforts may include instructing the parents
to keep the baby awake for several hours prior to
the exam, feeding the baby immediately before the
exam and/or the administration of sedation .
Acknowledgment . The authors would like to thank
Judith E. Widen, Ph .D ., for her guidance and assistance
in the preparation of this manuscript .
Portions of this study were presented at the 1988 Annual
Convention of the American Speech and Hearing Association, Boston, MA .
REFERENCES
Amadeo M, Shagass C. (1973) . Brief latency click evoked
potentials during waking and sleep in man. Psychosociology 10 :244-250 .
Durieux-Smith A, Picton T, Edwards C, Goodman JT,
MacMurray B. (1985) . The Crib-o-gram in the NICU : an
evaluation based on brain stem electric response
audiometry . Ear Hear 6:20-24.
Fria TJ . (1980) . The auditory brainstem response : background and clinical applications . Monogr Contemp
Audiol 2:1-44.
Fria TJ . (1985) . Identification of congenital hearing loss
with the auditory brainstem response . In : Jacobson JT,
ed . TheAuditory Brainstem Response . San Diego: College
Hill Press .
Galambos R. (1978) . Use of the auditory brainstem
response : its place in infant audiological evaluation .
Semin Speech Lang Hear 3: 76-87 .
Goff WR, Allison J, Lyons W, Fisher JC Conte R. (1977) .
Origins of short latency auditory evoked potentials in
man. In : Desmedt JE, ed . Auditory Evoked Potentials in
Man: Psychopharmacology Correlates in Evoked Potentials. Basel: Karger .
Gorga MP, Kaminski JR, Deauchaine K, Jesteadt W,
Neely S. (1989) . Auditory brainstem response from
children three months to three years of age: normal patterns of response. J Speech Hear Res 32 :281-287 .
Hall JW, Kripal JP, Hepp T. (1988) . Newborn hearing
screening with auditory brainstem response : measurement problems and solutions . Semin Hear 9:15-32 .
Hyde ML, Riko K, Corbin H, Moroso M, Alberti PW .
(1984) . A neonatal hearing screening research program
using brainstem electric response audiometry . J Otolaryngol 13 :49-54 .
Jacobson JT . (1985) . An overview of the auditory
brainstem response. In : Jacobson JT, ed . The Auditory
Brainstem Response . San Diego: College Hill Press.
Pediatric ABR Screening/McCall and Ferraro
Jacobson JT, Jacobson C. (1987) . Application of test performance characteristics in newborn auditory screening .
Semin Hear 8:133-141 .
Joint Committee on Infant Hearing, Position Statement.
(1982) . Pediatrics 70 :496-497 .
Levette MJ . (1984) . Auditory brainstem response testing
in the intensive care unit. Semin Hear 5: 57-69.
Martin F. (1981) . Introduction to Audiology. 2nd Ed .
Englewood Cliffs : Prentice Hall .
Moore EJ . (1983) . Bases of Auditory Brainstem Evoked
Responses. New York : Grune and Stratton .
Richmond KH, Konkle DF, Potsic W. (1986) . ABR screening in high risk infants : effects of ambient noise in the
neonatal nursery. Otolaryngol Head Neck Surg 94 :552557.
Roberts JL, Davis H, Phon GL, Reichert TJ, Sturtevant
EM, Marshall RE . (1982) . Auditory brainstem responses
in preterm neonates : maturation and follow-up. J
Pediatr 101:257-263 .
Ruth R, Dey-Sigman S, Mills JA. (1985) . Neonatal ABR
hearing screening . Hear J 38 :39-45 .
Sanders RA, Duncan PG, McCullough DW . (1979) . Clinical experience with brainstem audiometry performed
under general anesthesia . Otolaryngology 8 :24-32 .
Schulman-Galambos JL, Galambos R. (1979) . Brainstem
evoked response audiometry in newborn hearing screening. Arch Otolaryngol 105: 86-90.
Shannon DA, Felix JK, Krumholtz A, Goldstein DJ, Harris KC . (1984) . Hearing screening of high risk newborns
with brainstem auditory evoked potentials : a follow-up
study. Pediatrics 73 :22-25 .
Sohmer H, Gafni M, Chisin R. (1978) . Auditory nerve
and brainstem responses : comparison of awake and unconscious subjects . Arch Neurol 35 : 228-230.
Swigart E. (1986) . Neonatal Hearing Screening. San
Diego : College Hill Press.
Weber BA . (1983) . Pitfalls in auditory brainstem
response audiometry . Ear Hear 1 :179-184 .