Newborn hearing screening by otoacoustic emissions and automated auditory brainstem response

Newborn hearing screening by otoacoustic emissions and automated auditory brainstem response

INTEIWATIONALJOLRNALOF Pediatric ELSEVIER International Journal of Pediatric Otorhinolaryngology 41 (1997) 111-119 Newborn hearing screening by ot...

560KB Sizes 0 Downloads 35 Views

INTEIWATIONALJOLRNALOF

Pediatric

ELSEVIER

International Journal of Pediatric Otorhinolaryngology 41 (1997) 111-119

Newborn hearing screening by otoacoustic emissions and automated auditory brainstem response’ Karen Jo Doyle a,*, Barbara Burggraaff ‘, Sharon Fujikawa b, Ju Kim ‘>a a Depurtmenr of Otolaryngology-Head und Neck Surgery, University of Califbrnia Irvine, Bldg. 25, Rte. 81, 101 The City Drioe, Orange, CA 92668, USA h Department of Neurology, Uniwrsity of California Irvine, Bldg. 25, Rie. 81, 101 7he City Drice, Orange, CA 92668, USA Received 14 November

1996; accepted 20 April

1997

Abstract The aim of this study is to compare pass rates for two different hearing screening methods in well newborns as a function of age. Hearing screening tests were performed on 400 ears in 200 healthy newborn infants at the University of California-Irvine Medical Center. The screening methods used were automated auditory brainstem response (ABR) and click evoked otoacoustic emissions (EOAE). The infants’ ages ranged from 5 to 120 h, with an average age of 24 h. Overall, 88.5% of ears passed the ABR screen, and 79% passed the EOAE screen. There was no significant difference in the ABR pass rate for infants aged O-24 h of age as compared with infants aged > 24 h. However, the EOAE pass rate improved significantly in infants > 24 h compared with the group aged O-24 h (P < 0.01). Results are compared with earlier studies and implications for universal hearing screening are discussed. 0 1997 Elsevier Science Ireland Ltd.

* Corresponding author. Tel.: + I 714 4566640; fax: + 1 714 4565747. ’ Presented as a poster at the meeting of the American Academy of Audiology, Dallas, Texas, April, 1995. ’ Present address. Department of Otolaryngology-Head and Neck Surgery, Massachusetts Eye and Ear Infirmary. 0165.5876/97/$17.00

0 1997 Elsevier Science Ireland Ltd. All rights reserved

PI1 SOl65-5876(97)00066-9

112 Keywords:

K. J. Doyle et al. /Int.

J. Pediatr. Otovhinolaryngol.

41 (1997) I1 1-l 19

Hearing; Newborn; Auditory brainstem response; Evoked otoacoustic emissions

1. Introduction The goal of newborn screening is to accurately identify infants with significant hearing impairment in the most rapid and cost-effective way. Two tests have been proposed for infant screening: auditory brainstem response (ABR) and click-or transient-evoked otoacoustic emissions (EOAE). ABR, the averaged electrical response of the auditory nerve and brainstem auditory pathway to click stimuli, has been used for infant hearing screening for more than 20 years. EOAE, the acoustic ‘echo’ response emitted from the inner ear in response to a click, tone, or pairs of tones, is also a commonly used method for infant hearing screening. At the National Institutes of Health Development Conference on the Early Identification of Hearing Impairment in Infants and Young Children in 1993, the consensus panel concluded that all infants should be screened for hearing impairment, and urged that future research evaluate the validity and reliability of screening instruments, and compare the screening procedures [l I]. Several studies have shown high pass rates for both ABR and EOAE testing in newborns. Bonfils et al. [2] found EOAE and ABR present in 100% of 30 well infants tested. Kennedy et al. [8] found that only 4.6% of infants tested failed automated EOAE testing in at least one ear, while only 2.7% failed automated ABR testing. Other authors have found higher failure rates using ABR and/or EOAE testing [12,13]. Factors such as ambient noise, infant age, and test protocol affect pass rates for each type of screening [7,14]. The purpose of this study is to compare pass rates for automated ABR and transient EOAE in well newborns in the first few days of life, and to determine whether infant age affects pass rates for each of these screening tests.

2. Research design and methods 2.1. Subjects 200 newborn infants (400 ears) born at the University of California-Irvine Medical Center served as study subjects. Infants included in the study were born between July 28, 1994 and September 13, 1994. All infants were considered healthy ‘well babies’ without any complicating medical problems and infants admitted to the neonatal intensive care unit were excluded from the study. There were 101 male and 99 female infants. After obtaining an informed consent from the parent, each infant underwent testing with automated ABR and EOAE. Half of the infants were tested first with ABR; the other half with EOAE.

K. J. Doyle et al. ,‘lnt. J. Pediatr. Otorhinolaryngol.

41 (1997) 111-l 19

113

2.2. ABR screening ABR screening was performed using the ALGOautomated screener (Natus Medical, Foster City, CA). This mode of ABR was selected because of the widespread use of this device in universal hearing screening programs. Three disposable electrodes are attached to the infant. The noninverting electrode is placed on the forehead, the inverting electrode to the nape of the neck, and a ground electrode on the shoulder. Alternate polarity, 100 ps clicks, at a rate of 37.3 per s and an intensity of 35 dB nHL are presented to circumaural transducers. These transducers are also disposable and are affixed to the side of the head with an adhesive. The EEG activity in the 20 ms following click onset is filtered from 50 to 1400 Hz and averaged. For every 500 stimuli, the ALGOstatistically compares the obtained waveform to a template. The template has been derived from a composite waveform obtained from 35 normal neonates. A likelihood ratio is produced that indicates how much the response plus the background differs from only noise. After at least 500 clicks, a likelihood ratio of 160 is judged to imply with 99.8% confidence that a response is present and produces a ‘pass’. If this criterion is not achieved after the presentation of 15 000 stimuli, the result is a ‘refer’ (fail). After the right ear is tested, the screener automatically tests the left ear. The testing was carried out in a quiet room next to the newborn nursery, with the baby in an open isolette. The results of the ABR screen were recorded for each ear. The time needed to complete the entire test, including electrode placement was recorded. 2.3. EOA E screening EOAE screening was performed using the Otodynamic Analyzer IL088, (Otodynamics, UK) With this device, transient evoked otoacoustic emissions are elicited following a click stimulus. A neonatal rubber tipped probe is placed in the external ear canal. Sets of clicks are presented through the probe and delayed EOAEs recorded in the ear canal. The clicks are 80 ps in duration, presented at a rate of 50 per s, and are presented at levels of 80 dB peak SPL for the first three clicks of four-stimulus sets, with a fourth nonlinear balancing click of opposite polarity that is three times the amplitude of the first three. Alternate responses are averaged and stored in two separate waveforms that are then analyzed and cross-compared, resulting in a percent number called ‘reproducibility’. On the response screen of the ILO88, the response reproducibility is displayed for the entire frequency spectrum (500-6000 Hz) and for five lOOO-Hz frequency bands. In this study, criteria for a ‘pass’ included reproducibility of 50% or greater or if the response spectrum contained 3 dB more power than the noise spectrum in three 1000 Hz frequency bands centered at 1500, 2500 and 3500 Hz. If this level of reproducibility was not obtained after the acquisition of 500 subsets, this constituted a ‘fail’.

114

K. J. Doyle et al. /Int.

J. Pediatr. Otorhinolaryngol.

41 (1997) 11 l-l

19

Table 1 Pass and failure rates for ABR and EOAE in 400 ears of 200 well newborn infants Pass ABR Pass EOAE Fail EOAE Total

292 (73%)

Fail ABR

Total

23 (5.75%)

315 (78.8%)

62 (15.5%)

23 (5.75%)

85 (21.2%)

354 (88.5%)

46 (11.5%)

400

2.4. Analysis The ABR and EOAE pass rates were calculated for the entire sample (400 ears). X*-Tables were generated to compare pass rates for different infant ages. To analyze test order effects, the pass rates for EOAE and ABR were calculated and compared for the two orders (ABR tested first versus EOAE tested first.) Test time in min was also averaged for EOAE and ABR and compared.

3. Results

3.1. Pass rates for ABR and EOAE A pass rate of 88.5% (354 out of a possible 400 ears) was obtained for ABR; and EOAE testing revealed a pass rate of 79% (315 out of a possible 400 ears). Table 1 shows the pass rates and failure rates for both tests. Some authors report the percentage of infants failing screening rather than ears. There were ten infants (5%) who had bilateral ABR failures, and 26 infants (13%) who had unilateral failures. There were 25 infants (12.5%) who had bilateral EOAE failures, and 35 infants (17.5%) who failed EOAE unilaterally. 3.2. Age analysis Table 2 shows ABR and EOAE pass rates for infants less than or equal to 24 h of age compared with infants older than 24 h. The EOAE pass rate for infants aged Table 2 Comparison of pass rates for 400 newborn ears for automated ABR and EOAE screening as a function of infant age Age

Passed ABR

Passed EOAE

O-24 h >24 h

2181246 (88.6%) 136/l 54 (88.3%) Not significant

18 l/246 (73.6%) 134/ 154 (87%) P
K. J. Doyle et al. /Int.

J. Pediatr. Otorhinolaryngol.

Table 3 Pass rates for ABR and EOAE as a function

of infant age

Age

Passed ABR

Passed EOAE

O-12 h 12-24 h 25-36 h >36 h

70/84 (83.3%) 149/ 162 (92%) 79192 (85.9%) 56!62 (93.3%)

54184 (64.3%) 1271162 (78.4%) 8 l/92 (88%) 53:62 (85.5%)

Not significant

115

41 (1997) I II -119

P
24 h or less was 73.6% and improved to 87% in infants older than 24 h. The difference in pass rates for EOAE was statistically significant for the two groups (P < 0.01) while differences in age did not affect ABR pass rates. The infants were further subdivided into four age groups: Group 1 included infants who were O-12 h old. There were 42 newborns in this group for a total of 84 ears or 21% of the test sample. The average age was 9.5 h. Group 2 included infants 12-24 h old, with an average age of 18 h. It was the largest subgroup of the study with 81 infants (162 ears) or 400/o of the total. Group 3 consisted of infants 24-36 h old (mean 29 h). This group included 46 infants (92 ears) and comprised 23% of the sample. Group 4 had infants greater than 36 h old with an average age of 52 h. There were 31 infants (62 ears) in this group, comprising 15.5% of the sample, and ten infants were > 50 h old. For each group, the pass rates were again compared for both the ABR screen and the EOAE screen. The results are listed in Table 3. Again, age did not affect the pass rate for ABR (x2 = 5.38; not significant.) However, for EOAE testing, age was again significant factor. Group 1 had a pass rate of 64.3%; Group 2, 78.4%; Group 3, 88% and for Group 4 the pass rate was 85.5%. These differences were statistically significant (x2 = 16.95; P < 0.001). 3.3. Test order effects There were 196 ears that were tested first with ABR; 204 ears were tested first with EOAE. Of those that were tested first with ABR, 90% passed ABR and 84% passed EOAE. Of the ears that were tested first with EOAE, 87% passed ABR, and 73% passed EOAE. x2 Analysis showed a significant effect of test order on the pass rate for EOAE (P < O.Ol), but not on ABR pass rate (P < 0.3). 3.4. Test time analysis The average time needed to carry out automated ABR ALGOwas 24 min compared to 13 min for EOAE testing.

testing

with

the

116

K. J. Doyle et al. /h-u. J. Pediatr. Otorhinolaryngol.

41 (1997) 111-119

4. Discussion This study directly compares automated ABR and click-evoked otoacoustic emissions screening in well newborns. There are several previous studies that have performed this comparison. Stevens et al. [13] found that 88% of their subjects passed ABR screening bilaterally, while 79% passed EOAE screening bilaterally. Their study used intensive care infants tested at an older age than in the present study, using lower intensity EOAE stimuli and higher intensity ABR stimuli than in the present study, therefore comparison is difficult. Hunter et al. [6] tested 193 babies in the first 5 days of life, only 6% of whom were admitted to the Special Care Unit. They used the screening technique of transient EOAE at 31 and 41 dB nHL (a lower level than in the present study), followed by ALGO-I testing only for babies who failed the EOAE screen. Twenty-two percent of their infants failed EOAE unilaterally, while 8.5% failed bilaterally. Of 53 babies undergoing ABR testing, only one infant failed unilaterally; none bilaterally. The present study had similar failure rates for EOAE (17.5% unilaterally and 12.5% bilaterally), but a higher failure rate for ALGOscreening (5% of infants failed bilaterally; 13% unilaterally). A third study by White et al. [15] compared EOAE and ABR in a large sample of mainly well newborn infants, finding that 88.4% of ears passed ABR, and 82% passed EOAE. These numbers are similar to the results of the present study, despite methodological differences. ABR and EOAE in 84 ears of preterm infants were compared by Smurzynski et al. [12] who found 100% sensitivity for EOAE (all ears that failed ABR also failed EOAE). The present study does not agree that ears failing ABR are a subset of the ears failing EOAE. Of the 46 ears that failed ABR, half (23) passed EOAE (Table 1). Jacobson and Jacobson [7], who similarly found that half of the ears failing the ABR screening passed their EOAE screen, called this a 50% false negative rate, using screening ABR as the ‘gold standard’. White et al. [15] point out the faulty logic of using the ABR screening as the ‘gold standard’, when it has been shown that most infants failing screening ABR are eventually shown to have normal hearing. While follow-up data were not obtained in this study, we would also argue that with few exceptions, the infants who failed ABR and passed EOAE probably have normal hearing. Hence, the idea of using the more rapid EOAE screen as the initial test followed by ABR for those infants failing EOAE has some appeal. Previous research has found, like the present study, that EOAEs are found more frequently in newborns after 24 h. Kok et al. [9] found that EOAEs were present in 78% of infants less than 36 h old compared to 98% in infants greater than 108 h old. They speculated this could be related to amniotic fluid in the middle ear which tends to clear within the first few days post-partum. Kok et al. [lo] found that the prevalence of EOAEs in 20 ears of healthy newborns ages 3-51 h was only 50%, worse than the pass rate of 79% in this study. When these same infants were retested 24 h later, the pass rate rose to 100%. Chang et al. [4] presented evidence that cleaning vernix from the external canal improved the pass rate for transient otoacoustic emissions in well newborn infants. Bonfils et al. [3] found a much higher rate of EOAE detection in neonates aged one day (90%), that increased to 100% between 2 and 4 days old.

K. J. Doyle et al. / Int. J. Pediatr. Otorhinolaryngol.

41 (1997) ill -119

117

Although previous studies have not compared the pass rates for screening ABR as a function of infant age, some researchers have found an improvement in auditory thresholds found by ABR over the first few days after birth in humans. Adelman et al. [l] measured ABR thresholds in normal neonates aged 1 h to 5.5 months, and found that the average threshold in neonates aged O-5 h old was 29 dB higher than in adults, and reached adult threshold range in 2 weeks. They wrote that mechanical factors in the middle ear or basilar membrane immaturity could account for this difference. Stuart et al. [14] found significant improvements in thresholds for bone-conduction ABR compared to air conduction thresholds in a group of 20 neonates less than 48 h old, which they felt were due to resolution of middle ear fluid after birth. Eggermont [5] attributed the latency changes seen in development to increased myelination in the auditory pathway. In the present study, which uses a suprathreshold (35 dB nHL) stimulus, there was no improvement in ABR pass rate with increasing infant age. We analyzed for test order effects in this study. Because EOAE uses a probe that fits in the external canal, it could have the effect of either dilating the canal, thus improving the result of the second test, or of pushing external canal debris further into the canal, producing a conductive hearing loss and negatively influencing the result of the second test. In this study, when EOAE was performed first, the pass rate for ABR did not significantly change (87 vs. 90% when ABR was performed first). One would not expect any objective effects from performing ABR first, because the circumaural earphones do not contact the external canal. However, in this study, when ABR was performed first, the EOAE pass rate was 84% and when EOAE was the initial study, the pass rate for EOAE was only 73%. This difference was statistically significant (P < 0.01). The same testers performed both tests, and thus were not blinded to the results. It is possible that a pass obtained on the ABR test influenced the testers to persist with EOAE testing until a pass could be obtained. However, further analysis of the data found no differences in test time when ABR was performed first, compared to when EOAE performed first. Nonetheless, in our followup study EOAE is being performed first, to remove the possible bias. In this study the average times required to perform automated ABR and EOAE were 24 (range 5-90) and 13 (range 4-40) min respectively. Previous studies found that the length of time for ALGOtesting ranged from 14 to 26.3 min [8,11,12]. The time required for EOAE screening in these same studies has been shorter, ranging from 7.2 to 16.6 min. White et al. [15] found that on average EOAE screening takes 10 min in a newborn. It is arguable that EOAE is preferable as a screening tool because it requires about half the time of ABR. A new generation of automated ABR screener, the ALGOhas been developed that tests the two ears simultaneously and reportedly cuts the test time, and is being used in the follow-up study.

118

K. J. Doyle et al. / Int. J. Pediatr. Otorhinolaryngol.

41 (1997) Ill

119

5. Conclusions This study demonstrated that a higher pass rate can be obtained with the automated ABR screener, at the expense of longer test time. It was further shown that a significantly higher pass rate could be obtained by delaying EOAE to after 24 h of age, but no such effect was shown with ABR. Possible tester bias was shown in that the pass rate for EOAE was 84% when ABR was performed first, as compared with 73% when EOAE was the initial test. Another study is underway that will attempt to remove tester bias and use a newer generation ABR screener that reportedly cuts test time. Acknowledgements The authors acknowledge Natus Medical Inc., who lent the ALGO- infant hearing screener and donated disposable supplies for 3 months to be used in this study. References [I] C. Adelman, H. Levi, N. Linder, H. Sohmer, Neonatal auditory brain stem response threshold and latency: 1 h to 5 months, Electroencophalogr. Clin. Neurophysiol. 77 (1990) 77-80. [2] P. Bonfils, A. Uziel, R. Pujol, Screening for auditory dysfunction in infants by evoked oto-acoustic emissions, Arch. Otolaryngol. Head Neck Surg. 114 (1988) 887-890. [3] P. Bonfils, A. Dumont, P. Marie, M. Frdncois, P. Narcy, Evoked otoacoustic emissions in newborn hearmg screening, Laryngoscope 100 (1990) 186- 190. [4] K.W. Chang, B.R. Vohr, S.J. Norton, M.D. Lekas, External and middle ear status related to evoked otoacoustic emission in neonates, Arch. Otolaryngol. Head Neck Surg. 119 (1993) 2766282. [5] J.J. Eggermont, Development of auditory evoked potentials, Acta Otolaryngol. 112 (1992) 197-200. [6] M.F. Hunter, L. Kimm, D. Cafarelli Dees, C.R. Kennedy, A.R.D. Thornton, Feasibility of otoacoustic emission detection followed by ABR as a universal neonatal screening test for hearing impairment, Br. J. Audiol. 28 (1994) 47-51. [7] J.T. Jacobson, CA. Jacobson, The effects of noise in transient EOAE newborn hearing screening, Int. J. Pediatr. Otorhinolaryngol. 29 (1994) 2355248. [8] C.R. Kennedy, L. Kimm, D. Cafarelli Dees, P. Evans, M. Hunter, S. Lenton, A.R.D. Thornton, Otoacoustic emissions and auditory brainstem responses in the newborn, Arch. Dis. Child. 66 (1991) 112441129. [9] M.R. Kok, G.A. van Zanten, M.P. Brobaar, H.C.S. Wallenburg, Click-evoked oto-acoustic emissions in 1036 ears of healthy newborns, Audiology 32 (1993) 213-224. [IO] M.R. Kok, G.A. van Zanten, M.P. Brocaar, Growth of evoked otoacoustic emissions during the first days postpartum: a preliminary report, Audiology 31 (1992) 140-149. [l l] National Institute of Health Consensus Statement: Early Identification of Hearing Impairment in Infants and Young Children. 11 (1993) 17-19. [12] J. Smurzynski, M.D. Jung, D. Lafreniere, D.O. Kim, M. Vasudeva Kamath, J.C. Rowe, M.C. Holman, G. Leonard, Distortion-product and click-evoked otoacoustic emissions of preterm and full-term infants, Ear Hear. 14 (1993) 258-274. [13] J.C. Stevens, H.D. Webb, J. Hutchinson, J. Connell, M.F. Smith, J.T. Buffin, Click evoked otoacoustic emissions compared with brain stem electric response, Arch. Dis. Child. 64 (1989) 1105-1111.

K. J. Doyle et al. / Int. J. Prdiatr.

Otorhinolaryngol.

41 (1997) 1I1 J 19

119

[14] A. Stuart, E.Y. Yang, W.B. Green, Neonatal auditory brainstem response thresholds to air-and bone-conducted clicks: O-96 h postpartum, Am. J. Acad. Audiol. 5 (1994) 163-172. [15] K. White, B. Vohr, A. Maxon, T. Behrens, M. McPherson, G. Mauk, Screening all newborns for hearing loss using transient evoked otoacoustic emissions, Int. J. Ped. Otorhinolaryngol. 29 (1994) 203-217.