Hearing
Screening
of High-Risk
Newborns
with
Brainstem
Auditory
Evoked
Potentials:
A Follow-up
Study
Dorothy
A. Shannon,
PhD, Jacob
K. Felix, MD, FAAP,
Allan Krumholz, MD, Phillip J. Goldstein, MD, and Kenneth C. Harris
From the Departments of Rehabilitation Medicine, Pediatrics, Neurology, and Obstetrics, Sinai Hospital, and Departments of Pediatrics, Neurology, and Obstetrics, Johns Hopkins University School of Medicine, Baltimore
ABSTRACT. Numerous techniques have been used in
attempts to find a reliable and efficient screening method for determining auditory function in the newborn. The brainstem auditory evoked potential (BAEP) is the latest method advocated for that purpose. The BAEP was eva!-uated as a hearing screening test in 168 high-risk
new-horns between 35 and 45 weeks of conceptual age. Follow-up data were obtained after 1 year (mean 17.3 months) on 134 of the infants (80%). Normal hearing was defined as a reproducible response in both ears to a 25 dB normal hearing level (nHL) click stimulus; 21 infants (12.5%) failed the initial screening test. Follow-up on 19/21 in-fants revealed 18 infants with normal hearing and one infant with an 80 dB nHL bilateral hearing loss substan-tiated. One infant with an abnormal screening test died before retesting, and the other infant was lost to
follow-up but had only a unilaterally abnormal BAEP. None of
the infants with a normal BAEP screening study had
evidence of hearing loss on retesting. Sensitivity of the
BAEP was 100%, specificity was 86%, predictive value of a positive test was 5.26%, and the predictive value of a negative test was 100%. The incidence of significant
hearing loss in our population was between 0.75% (1/134 infants) confirmed, and 2.24% (3/134 infants) including
infants who failed screening but were lost to follow-up. The BAEP is a sensitive procedure for the early
identi-fication of hearing-impaired newborns. However, the
yield of significant hearing abnormalities was less than
predicted in other studies using BAEP for newborn
hear-ing screening. Pediatrics 1984;73:22-26; hearing screen-ing, newborn, brainstem, auditory brainstem response, evoked potentials.
The harmful effects of hearing loss on the
devel-Received for publication Oct 29, 1982; accepted June 10, 1983. Reprint requests to (J.F.K.) Department of Pediatrics, Sinai Hospital of Baltimore, Inc, Belvedere and Greenspring Aye, Baltimore, MD 21215.
PEDIATRICS (ISSN 0031 4005). Copyright © 1984 by the American Academy of Pediatrics.
opment of a child’s ability to learn, to communicate, and to socialize have stimulated efforts to initiate habilitative procedures early in life. Essential to this goal is the identification of infants with im-paired hearing as early in life as possible. Although screening of public school children for hearing loss has been practiced since 1927,’ screening of new-borns did not gain much attention until the 1960s. The success of behavioral screening for the early recognition of hearing loss led to a proliferation of programs for newborns.2 Although the Joint Com-mittee on Newborn Hearing Screening recom-mended that children at high risk for hearing im-pairment be identified and referred for audiologic evaluation in the first 3 to 6 months of life,3 there has been difficulty in developing effective methods.
A number of methods have been evaluated in the search for a reliable and effective technique for determining auditory function in newborns. Among these have been behavioral audiometry,2 impedance audiometry,4 respiratory responses,5 cardiac re-sponses,6 and crib movement systems.7 Only the crib movement system, or crib-o-gram, has retained significant interest as a tool for the screening of newborns.
The development of auditory evoked potentials has expanded the possibilities of objective testing of hearing function. Early research centered on slow latency cortical responses. However, these were found to have poor reliability due to the effects of state variability and interference from other
cere-bral electrical activity and muscle potentials.8
In contrast, brainstem auditory evoked potentials
(BAEP), as first recorded by Jewett et al#{176}offer
greater promise. These potentials occur from 1 to 10 ms after a click stimulus. Brainstem-auditory
these potentials.’#{176}” The BAEP demonstrates ex-cellent consistency from patient to patient, and habituation of the response does not occur. Poten-tials are generally unaffected by sleep state, muscle artifact, or ongoing cerebral electrical activity.’2 These factors have helped to make the test attrac-tive for testing newborns.’3”4 The BAEP has al-ready proven its reliability for the audiometric as-sessment of the adult and pediatric population.’5 Recently, investigators have advocated its use for screening the high-risk newborn.’6’9
The purpose of our study was to evaluate the
BAEP as a screening test for the detection of
hear-ing loss in high-risk newborns. We present the initial results and follow-up data on 168 infants.
METHODS
An attempt was made to test all newborns iden-tified as having high risk for hearing loss. Patients were selected from the regular and intensive care nurseries of Sinai Hospital of Baltimore. Patients included in the study were enrolled between July 1, 1978 and June 30, 1980. Based on the Joint Com-mittee of Newborn Hearing recommendations of
1973,20 the following criteria were employed to
choose subjects: (1) birth weight less than 1,500 g; (2) bilirubin level greater than 20 mg/dL for full term infants and greater than 12 mg/dL for pre-mature infants; (3) family history positive for hear-ing loss; (4) congenital malformations of the face, head, or neck; and (5) congenital infections.
A sixth category was added in which we studied
patients suffering from apnea, sepsis, seizures, neo-natal asphyxia, and prolonged aminoglycoside an-tibiotic therapy.
There were 168 newborns who met the study criteria listed above. These patients were tested between 35 and 45 weeks of conceptual age as determined by the criteria of Dubowitz et al.2’ The number of children referred for each criterion is shown in Table 1. Hyperbilirubinemia and low birth weight accounted for 76% of the referrals. Subjects were screened before discharge from the hospital. Sedation was not used. Infants were placed in port-able bassinets in an IAC sound suite. Grass disk silver-silver chloride electrodes were affixed to the
vertex (Cz) and both mastoids were affixed with
Grass electrode cream and gauze. Electrode imped-ance was maintained at less than 5,000 1. Evoked potentials were recorded from each ear independ-ently by summing electrical potentials between the vertex (grid- 1) and the mastoid ipsilateral to the stimulated ear (grid-2). The contralateral mastoid electrode served as the patient ground.
The stimulus was generated from a Nicolet NIC 1007A noise masking module. It consisted of a
100-TABLE 1. Reason for Referral of High-Risk Infants
No.of
Infants (N = 168)
%of
Population
1. <1,500 g 42 25.00
2. Hyperbilirubinemia 86 51.19
3. Family history positive for 8 4.76 hearing loss
4. Congenital malformations 4 2.38
5. Congenital infections 2 1.19
6. Miscellaneous (apnea, an- 26 15.48 tibiotics, seizures, etc)
L5 click presented monaurally through matched
TDH 39 earphones with MX-41/AR cushions.
Clicks were produced at a rate of 30 Hz. For screen-ing, three intensity levels were used, 65 dB nHL, 45 dB nHL, and 25 dB nHL (relative to the thresh-old of normal hearing adults tested in the same
environment). A descending intensity method of
presentation was employed. All responses were
re-produced for reliability. The major criterion for a
normal response was a clearly reproducible wave V
at 25 dB nHL in each ear. This intensity was chosen because the hearing threshold for newborns has been previously determined as 10 to 20 dB nHL.’6 If there was no response, the intensity of the stim-ulus was increased in
5-d.B
nHL increments until a reproducible response was obtained. The attending neonatologist reexamined infants with abnormal screening tests for signs of ear pathology.To obtain the most reliable data possible, we allowed a minimum of one hour of testing time per patient, and standard equipment in a specially de-signed sound suite was used. The technicians per-forming the test were experienced with evoked po-tentials and in dealing with newborns. We found the use of a decreasing intensity approach for the presentation of the click stimulus valuable in quiet-ing the infant. Initial presentation of louder clicks often encouraged sleep so that responses to softer sounds could be obtained with less interference. Children were tested only when they were medically stable and could be brought to our unit. This re-duced interference from electrical equipment and personnel in the active intensive care nursery.
All infants were called for retesting after 1 year of age. Contact was made by phone and letters, with the knowledge of the infant’s physician, and chil-dren were scheduled for behavioral audiometry which was thought to be more reliable by this age.
When it became obvious that many parents were
re-1 19
Died Retested Lost
sponse to sound and age-specific verbal and
lan-guage
abilities. The questionnaire was derived froma standard recommended source.22 If the parents
could not be reached, or doubt was raised, the pediatrician was contacted. The reliability of this type of questionnaire has not been rigorously tested, but the 50 infants who passed the question-naire and then had their hearing retested by audi-ometry all demonstrated normal hearing. No addi-tional cases of severe hearing loss have been brought to our attention since the initial interviews.
Although the standard protocol for behaviorial retesting at 1 year was followed with all children, those who had abnormality at the initial screening
were reevaluated prior to 1 year of age. These
children were retested with BAEP at 2- to 4-month intervals so as to provide an opportunity for early amplification in speech and language training when
indicated.
RESULTS
The results obtained on the 168 infants enrolled
in the study are summarized in the Figure.
Twenty-one infants failed the initial screening test; four infants died (one infant in the abnormal screening group) and 30 were lost to follow-up (one infant in the abnormal screening group). Of the 134 (80%) with adequate follow-up, 19 infants had been in the group that failed the initial screening test with
BAEP. The average chronologic age at the time of
follow-up for all infants was 17.3 months.
On retesting the group with abnormality, using
the evoked response in the first 9 months, or be-havioral audiometry after 9 months of age, 94.7%
(18/19 infants) of the abnormal results could not
be confirmed (Table 2). Subjects whose responses were normal on the retest were usually those whose initial test showed a unilateral and mild abnormal-ity (less than 40 dB nHL). None ofthese individuals had evidence of hearing loss after 1 year. The one abnormality that was confirmed on retesting in our study was found in an infant with severe bilateral hearing loss of more than 80 dB nHL and a family
history
of hearing loss. This infant has been furtherevaluated and fitted with appropriate amplification.
Training in speech and language skills has begun.
168 Patients
Screened with BAEP
21 Failed 147 Passed
Died Retested Lost
Behavioral Behavioral Questionnaire Questionnaire
and BAEP and BAEP and Behavioral Only
18 Passed 1 Failed 50 Passed 65 Passed
Figure. Summary results on infants screened with brainstem auditory evoked potentials
(BAEP).
TABLE 2. Reevaluation of Infants wi th Initially Abnormal Test
No. Failing
Initial Test
No. Normal
on Retest
No. Failing
Retest
No. Lost to
Follow-up
1. <1,500g 2. Hyperbilirubinernia
3. Family history positive for deafness
4. Congenital malformations 5. Congenital infections
6. Miscellaneous
8 8
3
2 0
0
8 7 2 1 0
0
0 0 1 0 0
0
0 1* 0 it 0
0
Totals 21 18 1 2
* Patient had a unilateral hearing loss (see text).
Two children with initial abnormality could not be
retested. One of these was a child with trisomy 18 who died in the nursery. The other infant had a unilaterally abnormal BAEP secondary to a de-formed eardrum as confirmed by an otolaryngolo-gist. Follow-up of that infant at 10 and 18 months of age, based only on incomplete data from the grandparents, suggests normal speech, language, and hearing development. Because the data are incomplete, this infant is not included in the follow-up group.
A total of 147 infants passed the initial screening
test. We were able to recall 1 15 from this group
(78%).
All 115 had normal findings, either on for-mal behavioral audiometric testing or on the basis of a questionnaire administered to the parents or the pediatrician of the child. There were reports of otitis media in some infants, but no cases of sig-nificant hearing loss were identified.Although we attempted to obtain formal
behav-ioral audiometry after the age of 1 year on all infants, we were able to restudy only 50 infants
(30% of the total that survived 1 year). For the
remaining 65 infants, follow-up was based on the
telephone interview of the family, the pediatrician,
or both.
We found the sensitivity of the BAEP as a screening test to be 100%, as the one hearing-impaired subject was identified on initial testing.
The specificity of the test is 86%; 115/133 normal
subjects were correctly determined with the BAEP. The predictive value of a positive test was low (only 5.26%); only 1/19 infants who were reported as
abnormal was later demonstrated to have abnormal
hearing. The predictive value of a negative test was 100%. All of the 115 infants who initially passed screening were later found to have normal hearing.
The incidence of deafness in our population is 1/
134 subjects or 0.75% confirmed. If the two subjects with abnormal findings but without follow-up are included, the incidence would be 2.24% (3/134).
DISCUSSION
The incidence of hearing loss identified in this study is slightly less than that reported by Simmons et al,7 who found an incidence of 1/62 patients in intensive care nurseries. That study used a crib motion system which detected a change in state 2.5 seconds after a 92 dB nHL narrow band 2 to 4 kHz noise.
Previous studies of BAEP have suggested a much
higher incidence of hearing loss in high-risk new-borns.’6’9 Schulman-Galambos and Galambos’6 employed the BAEP technique to test a group of 75 newborns at high risk for hearing loss. Referrals
were obtained from the intensive care nursery or
the attending physician. They screened with a stim-ulus at 60 and 30 dB nHL, and severe hearing loss was found in four infants (5.3% of their sample). All four of these patients with initial abnormality
were later reported to have substantiated hearing
losses. This is considerably higher than previous reports of hearing loss in an intensive care nursery population that was followed after discharge.7 The higher incidence of hearing loss in the
Schulman-Galambos and Galambos study could, in part, be
explained by the fact that two of their four patients
with abnormality had markedly elevated bilirubin levels, 38 mg/dL and 45 mg/dL, respectively.16 It is unusual for patients in our intensive care nursery to have such high bilirubin levels. This population
variable could account for much of the apparent
variability in the reported yield of BAEP for
hear-ing screening in infants.’6”7”9
In another study, Schulman-Galambos and
Gal-ambos studied 325 children with BAEP 1 year or
more after discharge from their intensive care nurs-ery. They found eight children (2.14%) with severe hearing 1055.16 This incidence is closer to our find-ings and to the results of the studies of a similar
population reported by Simmons et al,7 but a more
recent, large follow-up study by Galambos et al’8 continues to maintain a higher incidence of signifi-cant hearing loss, between 4% to 9%. Roberts et al,2’ in another recent follow-up study, could
con-firm hearing loss in only 2.3%; therefore, this issue
remains controversial.24
The cause of the initial BAEP screening failure
in our patients who were normal on retesting has not yet been determined. The possibility that serous otitis media was present at the time of testing must be considered, although otoscopic examination was negative in all but one patient. Starr et al’3 found
evidence of a temporary loss in one of their patients;
this was attributed to meningitis. Schulman-Gal-ambos and Galambos’6 found one infant with a 60
dB nHL loss which later resolved, and Stockard
and Westmoreland25 described transient BAEP ab-normalities in newborns.
It has been well documented that the BAEP is
was no evidence that this type of mild loss
pro-gressed,
and its clinical significance is dubious.BAEP
testing is expensive and time consuming.At the time of this study, most centers in our area
were charging $100 to $150 per test. At least 1 hour
of testing time may be required. In our study, the
infant with what subsequently proved to be
signifi-cant bilateral hearing loss would have been
identi-fled with a screening technique employing a 45 dB
nHL monaural stimulus. The time saved with a
modified BAEP protocol using a higher intensity
monaural stimulus may make the procedure more
cost effective as a screening procedure. As experi-ence is gained, modifications in technique should
be considered to reduce the incidence of
false-pos-itive results and improve cost effectiveness. In
ad-dition, advances in technology may also lessen the
cost of this procedure.
Although final conclusions as to the number of
patients with significant hearing loss in our
popu-lation must await long-term follow-up, we believe
the incidence of severe hearing loss will be in the
range of 1% to 3% of the total high-risk population. In the future, the limiting factor in habilitation
of hearing-impaired infants may no longer be early identification. The brainstem auditory-evoked
re-sponse was demonstrated in our study to be an
effective and sensitive procedure for hearing
screening in the newborn period. However, the yield of significant abnormalities is lower than previously
reported, and the incidence of false-positive results
is high. With further experience and technologic
advances, the BAEP may prove justified for
wide-spread clinical utilization in the hearing screening of high-risk newborns.
REFERENCES
1. Northern JL, Downs MP: Identification audiometry with children, in Hearing in Children. Baltimore, Williams & Wilkins, 1974, p 93
2. Gerber SE, Jones BL, Costello JM: Behavioral measures, in Audiometry in Infancy. New York, Grune & Stratton, i977, p85
3. American Academy ofPediatrics, Joint Committee on Infant He&ing Position statement 1982. Pediatrics i982;70:496-497
4. Keith RW: Impedance audiometry with neonates: Arch Oto-laryngol 1975;97:465
5. Heron TG, Jacobs R: Respiratory curve responses of the neonate to auditory stimulation.
mt
Audiol 1969;8:716. Cromwell DR: Heart rate responses of human newborns to modulated pure tones. Proc R Soc Med 1971;64:472
7. Simmons FB, McFarland WH, Jones FR: An automated hearing screening technique for newborns. Acta Otokzrngol
1979;87:1
8. Davis H: Principles of electric response audiometry. Ann Otol Rhinol Laryngol 1976;85(Suppl 28)
9. Jewett DL, Roman NM, Williston JS: Human auditory evoked potentials: Possible brainstem components detected on the scalp. Science 1970;167:1517
10. Buchwald J, Huang CM: Far-field acoustic response: Origins in the cat. Science 1975;189:382
11. Hashimoto I, Ishiyama Y, Yoshimoto T, et al: Brainstem
auditory evoked potentials recorded directly from brainstem and thalamus. Brain 1981;104:841
12. Stockard JJ, Rossiter VS: Clinical and pathological
corre-lates of brainstem auditory response abnormalities. Neurol-ogy 1977;27:316
13. Starr A, Amlie RN, Martin WH, et al: Development of auditory function in newborns revealed by brainstem poten-tials. Pediatrics 1977;60:831
14. Goldstein PJ, Krumholz A, Felix JK, et al: Brainstem evoked response in neonates. Am J Obstet Gynecol 1979;135:622
15. Jerger J, Hayes D, Jordan C: Clinical experience with audi-tory brainstem evoked response audiometry in pediatric
assessment. Ear Hear 1980;1:19
16. Schulman-Galambos C, Galambos R: Brainstem evoked
re-sponse auditometry in newborn hearing screening. Arch Otolaryngol 1979;105:86
17. Galambos R, Despland PA: The auditory brainstem re-sponse (ABR) evaluates risk factors for hearing loss in the newborn. Pediatr Res 1980;14:159-163
18. Galambos R, Hicks G, Wilson MJ: Hearing loss in graduates
of a tertiary intensive care nursery. Ear Hear 1982;3:87-90
19. Marshall RE, Reichert TJ, Kerley SM: Auditory function in newborn intensive care unit patients revealed by auditory brainstem potentials. J Pediatr 1980;96:731
20. American Speech and Hearing Association, American Acad-emy of Ophthalmology and Otolaryngology, and American Academy of Pediatrics: Supplementary statement of Joint Committee on Infant Hearing Screening. ASHA 1974;16:160 21. Dubowitz LMS, Dubowitz V, Goldberg C: Clinical
assess-ment of gestational age in the newborn infant. J Pediatr
1970;77:i-10
22. Sweitzer RF: Audiological evaluation ofthe infant and young
child, in Jaff BF (ed): Hearing Loss in Children. Baltimore, University Park Press, 1977, pp 107-111
23. Roberts JL, Davis H, Phon GL, et al: Auditory brainstem responses in preterm neonates: Maturation and follow-up. J Pediatr 1982;101:257-263
24. Simmons FB: Comments on “Hearing Loss in Graduates of a Tertiary Intensive Care Nursery.” Ear Hear
1982;3:188-190
