Newborn Hearing Screening in the NICU: Profile of Failed Auditory Brainstem Response/Passed Otoacoustic Emission
http://www.100md.com
《小儿科》
Department of Communication Studies/Communication Sciences and Disorders, Dyson College of Arts and Sciences, Pace University, New York, New York
Audiology and Speech Pathology Division, Department of Otolaryngology/Head and Neck Surgery
Department of Pediatrics Irving Center for Clinical Research, Columbia University Medical Center, New York, New York
ABSTRACT
Objective. Incidence of a specific pattern of auditory responses, absent auditory brainstem responses (ABRs) and present otoacoustic emissions (OAEs), in newborn hearing screening in a regional perinatal center neonatal intensive care unit (NICU) is described. This profile, labeled auditory neuropathy or auditory dyssynchrony (AN/AD), is a dysfunction in neural/brainstem transmission that occurs in individuals whose outer hairs cells are functioning normally. Although the AN/AD profile has been associated with various risk factors, incidence and prediction are unknown.
Method. Analysis of electrophysiologic measures and medical record reviews of the first 22 months of the universal newborn hearing–screening program was conducted. Association of the AN/AD profile was evaluated with the following factors: gender, gestational age, ototoxic drug regimen, low birth weight, hyperbilirubinemia, hydrocephalus, low Apgar score, anoxia, respiratory distress syndrome, pulmonary hypertension, intraventricular hemorrhage, multiple birth, seizure activity, and family history.
Results. One hundred fifteen (24.1%) of the 477 infants failed the ABR in 1 or both ears and passed OAEs bilaterally. Comparisons of infants fitting the AN/AD profile with those not fitting the AN/AD profile were negative with 3 exceptions: those with hyperbilirubinemia and those who were administered vancomycin or furosemide. A logistic-regression analysis model failed to predict which infants would be at risk for the AN/AD profile either unilaterally or bilaterally.
Conclusions. Screening of NICU infants should be conducted with ABR first, followed by OAE after failure on ABR. Because the incidence of the AN/AD profile was found to be 24% in this at-risk population, additional study is warranted.
Key Words: auditory neuropathy/dyssynchrony newborn hearing screening infants hyperbilirubinemia ototoxicity auditory brainstem response otoacoustic emissions neonatal intensive care prematurity
Abbreviations: ABR, auditory brainstem response OAE, otoacoustic emission CM, cochlear microphonics AN, auditory neuropathy AD, auditory dyssynchrony GA, gestational age CUMC, Columbia University Medical Center DPOAE, distortion-product otoacoustic emission
Techniques currently used in newborn hearing screening can discriminate peripheral (ie, cochlear) from central (ie, brainstem) auditory function. Two-phase screening using 2 different electrophysiologic measures, otoacoustic emissions (OAEs) and auditory brainstem response (ABR), allows detection of various failure patterns and provides more complete information about auditory function. State-mandated newborn hearing screening was introduced in New York State in 2001; however, as in many states, no specific electrophysiologic screening tool was prescribed.
A specific failure pattern is the presence of OAEs and/or cochlear microphonics (CM) with the absence of ABR. Auditory neuropathy/auditory dyssynchrony (AN/AD) disorder is manifested by normal outer hair cell function and abnormal, absent, or dyssynchronous neural function at the level of the VIIIth nerve.1, 2 Acoustic reflexes are absent and pure-tone audiometric findings range from near normal to severe-profound thresholds.3, 4 In addition, children with AN/AD often have speech perceptual difficulties, particularly in noise, which may make them functionally deaf.4–7 Table 1 reviews current literature and etiologic contributions to AN/AD in children.2, 4, 7–10
The clinical significance of the AN/AD disorder is that it may lead to failure in the development of normal auditory behavior or oral language. Although the failure to develop speech and language may mimic sensorineural hearing loss, there is equivocal evidence on the benefits of amplification for individuals with AN/AD. Earlier studies by Starr et al2 report no benefit. More recent studies found that some children with AN/AD who were fit with hearing aids derived benefit based on speech-perception testing and general auditory responsiveness.8 On the basis of these observations, others have reported cochlear implantation in select groups of children with AN/AD.11–15 Peterson et al15 placed cochlear implants in 2 groups of children, 1 group diagnosed with AN/AD and the other with sensorineural hearing loss that met traditional criteria for cochlear implantation. No differences in benefit were found between the groups on a variety of audiometric tests (ie, pure tone, ABR, speech perception) and functional measures (Meaningful Auditory Integration Scale and Infant-Toddler Meaningful Auditory integration Scale).
The purpose of this study was to report the outcome of 2 electrophysiologic measures in a cohort of infants admitted to a neonatal intensive care unit (NICU). In view of findings in the literature, which demonstrated that the AN/AD profile could only be discovered with a combined method of ABR and OAE, 2-phase screening was introduced at the end of 2002 and was aimed specifically at this high-risk cohort. The literature suggested an association of the AN/AD profile with multiple etiologic factors. In this study, the following 3 questions were asked: (1) What is the incidence of absent ABRs and present OAEs in the NICU population (2) Can the AN/AD profile be predicted from the following variables: gender, gestational age (GA), ototoxic drug regimen, low birth weight, hyperbilirubinemia (bilirubin level of >12 mg/dL in term infants or >0.5/kg in preterm infants), hydrocephalus, low Apgar scores, anoxia, respiratory distress syndrome, pulmonary hypertension, intraventricular hemorrhage, multiple birth, seizure activity, or family history (3) What is the relative proportion of unilateral and/or bilateral auditory findings
METHODS
Study Design
Execution of this study was in accordance with institutional review board guidelines. Auditory screening was performed on a cohort of infants admitted to the Columbia University Medical Center (CUMC) NICU, a regional perinatal center, between March 2002 and December 2003. Responses were required at 35-dB normal hearing level (nHL) for an infant to pass the automated ABR screening and at 3 of 4 frequency bands to pass the OAE. Infants excluded from the study were those transferred back to their hospital of origin; those who had positive findings ("referred") on both ABR and OAE; those who received only 1 physiologic measure; or those who expired.
Risk factors were extracted from medical records of the entire NICU cohort. Auditory failure patterns were recorded. One individual coded data, which were verified/reviewed by the 3 senior authors. Nonparametric and a logistic-regression analyses were used to determine if risk factors were predictive of the AN/AD profile.
Study Subjects
Subjects were drawn from NICU discharges from March 2002 to December 2003 at the CUMC Morgan Stanley Children's Hospital of New York-Presbyterian in New York City, New York. Criteria for admission to the NICU included premature infants at <36 weeks' gestation and/or with a birth weight of <1800 g; term infants with suspected or established sepsis; infants with congenital anomalies; infants with elevated bilirubin levels requiring exchange transfusion (term infants: 20 mg/dL; preterm infants: variable and dependent on weight and etiology) and/or phototherapy (term infants: at 2/3 exchange level; preterm infants: at exchange level); metabolic illness; infants requiring surgical correction of anomalies; infants of mothers with insulin-dependent diabetes; or infants requiring closer monitoring or evaluation than could be provided in the well-infant nursery.
Of the 1736 admissions in this time period, 477 infants were screened with ABR first and OAEs second if they failed the ABR. The remaining 1259 infants were transferred back to referring hospital and were not required to have a hearing screen at this site (378 infants); expired (96 infants); received only 1 electrophysiologic measure (628 infants); had a positive ("referred") finding on both the ABR and OAE (89 infants); or were missed and therefore not screened (68 infants).
Study Procedures
The newborn hearing–screening program at CUMC began in March 2002. Infants admitted to the NICU were screened initially by using OAE technology. In late 2002, based on literature review, screening protocols were modified. All NICU infants were screened initially with ABR and underwent OAE screening when they failed ("referred") on ABR. The medical records were examined for the following factors: gender, GA, birth weight, presence or absence of respiratory distress, presence of hyperbilirubinemia, administration of ototoxic medications, multiple births, seizures, and family history. Potentially ototoxic medications (gentamicin, vancomycin, and furosemide) were coded to indicate that they were not given, given at subclinical or nontoxic levels versus potentially injurious levels, or determined by blood samples when therapy lasted >48 hours.
Infants were tested by using the ABaer system (Bio-logic Corp, Mundelein, IL) with the following preset OAE and ABR screening parameters. Infants were screened by using distortion-product OAEs (DPOAEs), set at a 2- to 5-kHz screen with 3 of 4 frequency bands required present for a pass. Intensity was calibrated at a 65-dB sound pressure level for band 1 and 55-dB sound pressure level for band 2. Parameters for the ABR included a 100-microsecond click stimulus, alternating polarity, a click rate of 37.1 clicks per second, and an intensity of 35-dB nHL. The low-pass filter was set at 1500 Hz, and the high pass filter was set to 100 Hz; a minimum of 1536 sweeps up to a maximum of 2 trials of 6144 sweeps were obtained. A patented signal-detection algorithm, point-optimized variance ratio, determined the presence of an ABR and assigned a pass or refer result. A point-optimized variance ratio score of 3.50 was necessary for a pass. The specificity of the screening ABR test is in excess of 98.5%. The verified theoretical sensitivity for the bilateral ABR test is >99.96% (G. Raviv, PhD, Bio-logic Corp, written communication, 2005).
RESULTS
Characteristics of Infants With a Suspected AN/AD Profile
Four hundred seventy-seven infants met the inclusion criteria; 115 (24.1%) infants (62 males and 53 females) had absent ABRs in at least 1 ear and present OAE responses. Three hundred sixty-two (75.9%) infants (197 males and 165 females) passed the initial ABR. Table 2 describes the characteristics of the 115 infants with the AN/AD profile and the 362 infants without the AN/AD profile (ie, passed the ABR). Comparisons of the infants who fit the AN/AD profile with those who did not fit the AN/AD profile were nonsignificant with 3 exceptions: 2 results were significant (P < .01) in all comparisons for hyperbilirubinemia (df = 1,N = 477; P = .0001), vancomycin (df = 1,N = 475; P = .007), and furosemide (df = 1,N = 475; P = .005).
Predicting the AN/AD Profile
An attempt was made to create a multivariate model by using a logistic-regression analysis. AN/AD predictors included gender, GA, ototoxic drug regimen, low birth weight, hyperbilirubinemia, hydrocephalus, low Apgar score, anoxia, respiratory distress syndrome, pulmonary hypertension, intraventricular hemorrhage, multiple birth, and seizure activity. No model could be generated to account for the AN/AD profile from these risk factors.
Laterality and Bilaterality
Of the 115 infants who fit the AN/AD profile, ABRs were absent in 35 (30.4%) with right ear stimulation, 53 (46%) with left ear stimulation, and 27 (23.5%) for both ears. That is, 88 (76.5%) were affected unilaterally, and 23.5% were affected bilaterally.
DISCUSSION
AN/AD, that is, absent ABRs with present OAEs and/or CM with or without measurable hearing, has been reported in the literature16, 17 since 1979. Suggested etiologies include a dyssynchrony at the level of the VIIIth nerve and/or brainstem necessary for early waveform generation,18 a disturbance in the inner hair cells (especially if the pure-tone audiometric thresholds are in the severe-profound range19), a dysfunction of the spiral ganglion fibers themselves,20 or a defect in afferent transmission at the synapse between the inner hair cells and the VIIIth nerve.21 Questions still remain, however, in regard to the site(s) of the disorder and possible etiologic factors.22
AN/AD can occur in the absence of any other apparent medical condition6; however, a perinatal history of hyperbilirubinemia, anoxia, or infectious diseases (one or more of these) are all factors.8, 23, 24 Familial patterns of inheritance,10, 25 syndromic and nonsyndromic recessive inheritance,26, 27 have been described also. It is interesting to note that in this cohort, no family history of hearing dysfunction was reported (see Table 2). Thus, etiologic factors associated with AN/AD remain unclear. Reports of AN/AD, from single case studies27, 28 to small samples,4, 9, 29 have provided limited insight. Sininger30 reviewed the incidence of the AN/AD profile in NICU infants and found a range in the literature extending from 5.3% to 14.8% (mean: 10.8%).
The present study found an incidence estimate of 24.1% of the AN/AD profile in the NICU, a higher rate than has been reported previously. It is possible that this cohort reflects a more at-risk population because it is a regional perinatal center site. Of course, neurologic immaturity may have been a factor here, and additional follow-up will confirm the AN/AD diagnosis. Regarding incidence, Rance et al8 found in a retrospective study that 20 (0.38%) of the 5199 children assessed between 1991 and 1996 had absent ABRs and present CM. Of the 5199 infants, 109 (2.09%) had some type of permanent hearing deficit. Of the 109 infants identified with some type of hearing loss, 11% (1 in 9) were subsequently diagnosed with AN/AD, and 97 (89%) were subsequently diagnosed with a sensory-type hearing loss.
Hall et al31 concluded that the combined ABR/OAE screening was the only means to avoid missing infants with the AN/AD profile. They suggested 2 feasible approaches: (1) use the combined ABR/OAE-screening method on those infants with neurologic risk factors or (2) use the combined ABR/OAE-screening method on all infants who are admitted to the NICU. In our sample, the risk factors did not prove sensitive enough to identify the bulk of infants who fit the AN/AD profile; thus, the second approach (ie, 2-phase screening using both ABRs and OAEs of all NICU discharges) is recommended. Programs implementing only OAEs in their NICU test protocol will miss this profile. It should be noted that an all-ABR protocol would also refer appropriate infants into a diagnostic setting, although the administration of the OAE heightens the alert and is easily administered.
A recent study by de Hoog et al32 found that serum concentrations above the therapeutic range of vancomycin, tobramycin, or furosemide were not related to failure on automated ABR in at-risk neonates. Of all risk factors analyzed in the present study, only 3 (hyperbilirubinemia and administration of vancomycin and furosemide) had any discriminating power. However, when using the powerful logistic-regression tool, no factors or combinations were effective in predicting the AN/AD profile. There are possible explanations for this result: (1) insufficient statistical power; (2) this NICU population is so diverse that 1 profile could not be developed to distinguish between the 2 groups; or (3) there may be other factors that are presently unknown. Starr21 reports that etiology is unknown in 40% of individuals with AN/AD. Our data echo this finding.
Approximately one third (33.9%) of infants with the AN/AD profile had a GA of 38 weeks. This finding suggests that, for all intents and purposes, these children were admitted to the NICU for reasons other than prematurity and may well have been missed had they been admitted to the well-infant nursery, in which an OAE-screening technique is typically used. Of the 20 infants that Rance et al8 examined and ultimately diagnosed with AN/AD, 13 (65%) had a GA of 38 weeks, also supported by our findings.
Consistent with Starr,21 there seems to be no gender preference in our sample. In contrast to his findings of 96% of cases having bilateral AN/AD results, our sample reflected a large proportion (88 [76%]) of unilateral findings of the AN/AD profile, supporting the concept that diagnostic evaluation is critical to confirm AN/AD.
Clearly, the next step is longitudinal evaluation of these children. A study is under way to clarify how many cases with the AN/AD profile represent auditory immaturity, transient neurologic abnormality,33, 34 or AN/AD. Care must be taken not to diagnose AN/AD without sufficient evidence. All of the following assessments should be included for an infant AN/AD work-up: ABR using separate trials of condensation and rarefaction clicks; OAEs; acoustic reflex threshold testing or middle-ear muscle testing; and behavioral audiometry using visual reinforcement audiometry.5, 34 Consideration should be given to the use of slower than typically administered ABR click rates (ie, 11.1 per second for enhancement of waveform morphology). Now that universal newborn hearing screening is mandated in 38 states and the District of Columbia (as of 2004), the ability to detect the AN/AD configuration of findings is something that we must be prepared to recognize and address.
ACKNOWLEDGMENTS
We thank Alex Cirilo, Alex Goenaga, Ashley Navarro, Flerida Noyola, Madelyne Rosario, and Maria Olmeda-Jenkins, MS, for their contributions in data collection. We also thank the 2 reviewers for thoughtful and insightful comments.
FOOTNOTES
Accepted May 2, 2005.
No conflict of interest declared.
REFERENCES
Starr A, McPherson D, Patterson J, et al. Absence of both auditory evoked potentials and auditory percepts dependent on timing cues. Brain. 1991;114 :1157 –1180
Starr A, Picton TW, Sininger Y, Hood LJ, Berlin CI. Auditory neuropathy. Brain. 1996;119 :741 –753
Sininger Y, Oba S. Patients with auditory neuropathy. In: Sininger Y, Starr A, eds. Auditory Neuropathy: A New Perspective on Hearing Disorders. San Diego, CA: Singular Thomson Learning; 2001:15 –35
Rance G, Cone-Wesson B, Wunderlich J, Dowell R. Speech perception and cortical event related potentials in children with auditory neuropathy. Ear Hear. 2002;23 :239 –253
Starr A, Picton TW, Kim R. Pathophysiology of auditory neuropathy. In: Sininger Y, Starr A, eds. Auditory Neuropathy: A New Perspective on Hearing Disorders. San Diego, CA: Singular Thomson Learning, 2001:67 –82
Kraus N. Auditory neuropathy: an historical and current perspective. In: Sininger Y, Starr A, eds. Auditory Neuropathy: A New Perspective on Hearing Disorders. San Diego, CA: Singular Thomson Learning; 2001:1 –15
Madden C, Rutter M, Hilbert L, Greinwald JH Jr, Choo DI. Clinical and audiological features in auditory neuropathy. Arch Otolaryngol Head Neck Surg. 2002;128 :1026 –1030
Rance G, Beer DE, Cone-Wesson B, et al. Clinical findings for a group of infants and young children with auditory neuropathy. Ear Hear. 1999;20 :238 –252
Dunkley C, Farnsworth A, Mason S, Dodd M, Gibbin K. Screening and follow up assessment in three cases of auditory neuropathy. Arch Dis Child. 2003;88 :25 –26
Wang Q, Gu R, Han D, Yang W. Familial auditory neuropathy. Laryngoscope. 2003;113 :1623 –1629
Shallop JK, Peterson A, Facer GW, Fabry LB, Driscoll CL. Cochlear implants in five cases of auditory neuropathy: postoperative findings and progress. Laryngoscope. 2001;111 :555 –562
Madden C, Hilbert L, Rutter M, Greinwald J, Choo D. Pediatric cochlear implantation in auditory neuropathy. Otol Neurotol. 2002;23 :163 –168
Trautwein P, Shallop J, Fabry L, Friedman R. Cochlear implantation of patients with auditory neuropathy. In: Sininger Y, Starr A, eds. Auditory Neuropathy: A New Perspective on Hearing Disorders. San Diego, CA: Singular Thomson Learning; 2001:203 –232
Sininger YS, Trautwein P. Electrical stimulation of the auditory nerve via cochlear implants in patients with auditory neuropathy. Ann Otol Rhinol Laryngol Suppl. 2002;189 :29 –31
Peterson A, Shallop J, Driscoll C, et al. Outcomes of cochlear implantation in children with auditory neuropathy. J Am Acad Audiol. 2003;14 :188 –201
Davis H, Hirsh S. A slow brainstem response for low-frequency audiometry. Audiology. 1979;18 :445–461
Worthington D, Peters J. Quantifiable hearing and no ABR: paradox or error Ear Hear. 1980;1 :281 –285
Kraus N, Ozdamar O, Stein L, Reed N. Absent auditory brain stem response: peripheral hearing loss or brain stem dysfunction Laryngoscope. 1984;94 :400 –406
Prieve BA, Gorga MP, Neely ST. Otoacoustic emissions in an adult with severe hearing loss [published correction appears in J Speech Hear Res. 1991;34:703]. J Speech Hear Res. 1991;34 :379 –385
Sininger Y, Hood L, Starr A, Berlin C, Picton T. Hearing loss due to auditory neuropathy. Audiol Today. 1995;7 :10 –13
Starr A. The neurology of auditory neuropathy. In: Sininger Y, Starr A, eds. Auditory Neuropathy: A New Perspective on Hearing Disorders. San Diego, CA: Singular Thomson Learning; 2001:37 –49
Rapin I, Gravel J. "Auditory neuropathy": physiologic and pathologic evidence calls for more diagnostic specificity. Int J Pediatr Otorhinolaryngol. 2003;67 :707 –728
Stein L, Tremblay K, Pasternak J, Banerjee S, Lindemann K, Kraus N. Brainstem abnormalities in neonates with normal otoacoustic emissions. Sem Hear. 1996;17 :197 –213
Shapiro SM. Bilirubin toxicity in the developing nervous system. Pediatr Neurol. 2003;29 :410 –421
Shivashankar N, Satishchandra P, Shashikala HR, Gore M. Primary auditory neuropathy—an enigma. Acta Neurol Scand. 2003;108 :130 –135
Starr A, Michalewski HJ, Zeng FG, et al. Pathology and physiology of auditory neuropathy with a novel mutation in the MPZ gene (Tyr145–>Ser) [published correction appears in Brain. 2003 Jul;126:1718]. Brain. 2003;126 :1604 –1619
Varga R, Kelley PM, Keats BJ, et al. Non-syndromic recessive auditory neuropathy is the result of mutations in the otoferlin (OTOF) gene. J Med Genet. 2003;40(1) :45 –50
Song JH, Koslo E. A case of unilateral auditory neuropathy. Presented at: the New York State Speech Language Hearing Association; April 15–17, 2004; Albany, NY
Sheykholeslami K, Kaga K, Kaga M. An isolated and sporadic auditory neuropathy (auditory nerve disease): report of five patients. J Laryngol Otol. 2001;115 :530 –534
Sininger Y. Overview of auditory neuropathy. Audiology Online. July 8, 2004
Hall J III, Smith SD, Popelka GR. Newborn hearing screening with combined otoacoustic emissions and auditory brainstem responses. J Am Acad Audiol. 2004;15 :414 –425
de Hoog M, van Zanten BA, Hop WC, Overbosch E, Weisglas-Kuperus N, van den Anker J. Newborn hearing screening: tobramycin and vancomycin are not risk factors for hearing loss. J Pediatr. 2003;142 :41 –46
Edwards CG, Durieux-Smith A, Picton TW. Auditory brainstem response audiometry in neonatal hydrocephalus. J Otolaryngol Suppl. 1985;14 :40 –46
Kileny PR, Robertson CMT. Neurological aspects of infant hearing assessment. J Otolaryngol Suppl. 1985;14 :34 –39(Abbey L. Berg, PhD, Jacly)
Audiology and Speech Pathology Division, Department of Otolaryngology/Head and Neck Surgery
Department of Pediatrics Irving Center for Clinical Research, Columbia University Medical Center, New York, New York
ABSTRACT
Objective. Incidence of a specific pattern of auditory responses, absent auditory brainstem responses (ABRs) and present otoacoustic emissions (OAEs), in newborn hearing screening in a regional perinatal center neonatal intensive care unit (NICU) is described. This profile, labeled auditory neuropathy or auditory dyssynchrony (AN/AD), is a dysfunction in neural/brainstem transmission that occurs in individuals whose outer hairs cells are functioning normally. Although the AN/AD profile has been associated with various risk factors, incidence and prediction are unknown.
Method. Analysis of electrophysiologic measures and medical record reviews of the first 22 months of the universal newborn hearing–screening program was conducted. Association of the AN/AD profile was evaluated with the following factors: gender, gestational age, ototoxic drug regimen, low birth weight, hyperbilirubinemia, hydrocephalus, low Apgar score, anoxia, respiratory distress syndrome, pulmonary hypertension, intraventricular hemorrhage, multiple birth, seizure activity, and family history.
Results. One hundred fifteen (24.1%) of the 477 infants failed the ABR in 1 or both ears and passed OAEs bilaterally. Comparisons of infants fitting the AN/AD profile with those not fitting the AN/AD profile were negative with 3 exceptions: those with hyperbilirubinemia and those who were administered vancomycin or furosemide. A logistic-regression analysis model failed to predict which infants would be at risk for the AN/AD profile either unilaterally or bilaterally.
Conclusions. Screening of NICU infants should be conducted with ABR first, followed by OAE after failure on ABR. Because the incidence of the AN/AD profile was found to be 24% in this at-risk population, additional study is warranted.
Key Words: auditory neuropathy/dyssynchrony newborn hearing screening infants hyperbilirubinemia ototoxicity auditory brainstem response otoacoustic emissions neonatal intensive care prematurity
Abbreviations: ABR, auditory brainstem response OAE, otoacoustic emission CM, cochlear microphonics AN, auditory neuropathy AD, auditory dyssynchrony GA, gestational age CUMC, Columbia University Medical Center DPOAE, distortion-product otoacoustic emission
Techniques currently used in newborn hearing screening can discriminate peripheral (ie, cochlear) from central (ie, brainstem) auditory function. Two-phase screening using 2 different electrophysiologic measures, otoacoustic emissions (OAEs) and auditory brainstem response (ABR), allows detection of various failure patterns and provides more complete information about auditory function. State-mandated newborn hearing screening was introduced in New York State in 2001; however, as in many states, no specific electrophysiologic screening tool was prescribed.
A specific failure pattern is the presence of OAEs and/or cochlear microphonics (CM) with the absence of ABR. Auditory neuropathy/auditory dyssynchrony (AN/AD) disorder is manifested by normal outer hair cell function and abnormal, absent, or dyssynchronous neural function at the level of the VIIIth nerve.1, 2 Acoustic reflexes are absent and pure-tone audiometric findings range from near normal to severe-profound thresholds.3, 4 In addition, children with AN/AD often have speech perceptual difficulties, particularly in noise, which may make them functionally deaf.4–7 Table 1 reviews current literature and etiologic contributions to AN/AD in children.2, 4, 7–10
The clinical significance of the AN/AD disorder is that it may lead to failure in the development of normal auditory behavior or oral language. Although the failure to develop speech and language may mimic sensorineural hearing loss, there is equivocal evidence on the benefits of amplification for individuals with AN/AD. Earlier studies by Starr et al2 report no benefit. More recent studies found that some children with AN/AD who were fit with hearing aids derived benefit based on speech-perception testing and general auditory responsiveness.8 On the basis of these observations, others have reported cochlear implantation in select groups of children with AN/AD.11–15 Peterson et al15 placed cochlear implants in 2 groups of children, 1 group diagnosed with AN/AD and the other with sensorineural hearing loss that met traditional criteria for cochlear implantation. No differences in benefit were found between the groups on a variety of audiometric tests (ie, pure tone, ABR, speech perception) and functional measures (Meaningful Auditory Integration Scale and Infant-Toddler Meaningful Auditory integration Scale).
The purpose of this study was to report the outcome of 2 electrophysiologic measures in a cohort of infants admitted to a neonatal intensive care unit (NICU). In view of findings in the literature, which demonstrated that the AN/AD profile could only be discovered with a combined method of ABR and OAE, 2-phase screening was introduced at the end of 2002 and was aimed specifically at this high-risk cohort. The literature suggested an association of the AN/AD profile with multiple etiologic factors. In this study, the following 3 questions were asked: (1) What is the incidence of absent ABRs and present OAEs in the NICU population (2) Can the AN/AD profile be predicted from the following variables: gender, gestational age (GA), ototoxic drug regimen, low birth weight, hyperbilirubinemia (bilirubin level of >12 mg/dL in term infants or >0.5/kg in preterm infants), hydrocephalus, low Apgar scores, anoxia, respiratory distress syndrome, pulmonary hypertension, intraventricular hemorrhage, multiple birth, seizure activity, or family history (3) What is the relative proportion of unilateral and/or bilateral auditory findings
METHODS
Study Design
Execution of this study was in accordance with institutional review board guidelines. Auditory screening was performed on a cohort of infants admitted to the Columbia University Medical Center (CUMC) NICU, a regional perinatal center, between March 2002 and December 2003. Responses were required at 35-dB normal hearing level (nHL) for an infant to pass the automated ABR screening and at 3 of 4 frequency bands to pass the OAE. Infants excluded from the study were those transferred back to their hospital of origin; those who had positive findings ("referred") on both ABR and OAE; those who received only 1 physiologic measure; or those who expired.
Risk factors were extracted from medical records of the entire NICU cohort. Auditory failure patterns were recorded. One individual coded data, which were verified/reviewed by the 3 senior authors. Nonparametric and a logistic-regression analyses were used to determine if risk factors were predictive of the AN/AD profile.
Study Subjects
Subjects were drawn from NICU discharges from March 2002 to December 2003 at the CUMC Morgan Stanley Children's Hospital of New York-Presbyterian in New York City, New York. Criteria for admission to the NICU included premature infants at <36 weeks' gestation and/or with a birth weight of <1800 g; term infants with suspected or established sepsis; infants with congenital anomalies; infants with elevated bilirubin levels requiring exchange transfusion (term infants: 20 mg/dL; preterm infants: variable and dependent on weight and etiology) and/or phototherapy (term infants: at 2/3 exchange level; preterm infants: at exchange level); metabolic illness; infants requiring surgical correction of anomalies; infants of mothers with insulin-dependent diabetes; or infants requiring closer monitoring or evaluation than could be provided in the well-infant nursery.
Of the 1736 admissions in this time period, 477 infants were screened with ABR first and OAEs second if they failed the ABR. The remaining 1259 infants were transferred back to referring hospital and were not required to have a hearing screen at this site (378 infants); expired (96 infants); received only 1 electrophysiologic measure (628 infants); had a positive ("referred") finding on both the ABR and OAE (89 infants); or were missed and therefore not screened (68 infants).
Study Procedures
The newborn hearing–screening program at CUMC began in March 2002. Infants admitted to the NICU were screened initially by using OAE technology. In late 2002, based on literature review, screening protocols were modified. All NICU infants were screened initially with ABR and underwent OAE screening when they failed ("referred") on ABR. The medical records were examined for the following factors: gender, GA, birth weight, presence or absence of respiratory distress, presence of hyperbilirubinemia, administration of ototoxic medications, multiple births, seizures, and family history. Potentially ototoxic medications (gentamicin, vancomycin, and furosemide) were coded to indicate that they were not given, given at subclinical or nontoxic levels versus potentially injurious levels, or determined by blood samples when therapy lasted >48 hours.
Infants were tested by using the ABaer system (Bio-logic Corp, Mundelein, IL) with the following preset OAE and ABR screening parameters. Infants were screened by using distortion-product OAEs (DPOAEs), set at a 2- to 5-kHz screen with 3 of 4 frequency bands required present for a pass. Intensity was calibrated at a 65-dB sound pressure level for band 1 and 55-dB sound pressure level for band 2. Parameters for the ABR included a 100-microsecond click stimulus, alternating polarity, a click rate of 37.1 clicks per second, and an intensity of 35-dB nHL. The low-pass filter was set at 1500 Hz, and the high pass filter was set to 100 Hz; a minimum of 1536 sweeps up to a maximum of 2 trials of 6144 sweeps were obtained. A patented signal-detection algorithm, point-optimized variance ratio, determined the presence of an ABR and assigned a pass or refer result. A point-optimized variance ratio score of 3.50 was necessary for a pass. The specificity of the screening ABR test is in excess of 98.5%. The verified theoretical sensitivity for the bilateral ABR test is >99.96% (G. Raviv, PhD, Bio-logic Corp, written communication, 2005).
RESULTS
Characteristics of Infants With a Suspected AN/AD Profile
Four hundred seventy-seven infants met the inclusion criteria; 115 (24.1%) infants (62 males and 53 females) had absent ABRs in at least 1 ear and present OAE responses. Three hundred sixty-two (75.9%) infants (197 males and 165 females) passed the initial ABR. Table 2 describes the characteristics of the 115 infants with the AN/AD profile and the 362 infants without the AN/AD profile (ie, passed the ABR). Comparisons of the infants who fit the AN/AD profile with those who did not fit the AN/AD profile were nonsignificant with 3 exceptions: 2 results were significant (P < .01) in all comparisons for hyperbilirubinemia (df = 1,N = 477; P = .0001), vancomycin (df = 1,N = 475; P = .007), and furosemide (df = 1,N = 475; P = .005).
Predicting the AN/AD Profile
An attempt was made to create a multivariate model by using a logistic-regression analysis. AN/AD predictors included gender, GA, ototoxic drug regimen, low birth weight, hyperbilirubinemia, hydrocephalus, low Apgar score, anoxia, respiratory distress syndrome, pulmonary hypertension, intraventricular hemorrhage, multiple birth, and seizure activity. No model could be generated to account for the AN/AD profile from these risk factors.
Laterality and Bilaterality
Of the 115 infants who fit the AN/AD profile, ABRs were absent in 35 (30.4%) with right ear stimulation, 53 (46%) with left ear stimulation, and 27 (23.5%) for both ears. That is, 88 (76.5%) were affected unilaterally, and 23.5% were affected bilaterally.
DISCUSSION
AN/AD, that is, absent ABRs with present OAEs and/or CM with or without measurable hearing, has been reported in the literature16, 17 since 1979. Suggested etiologies include a dyssynchrony at the level of the VIIIth nerve and/or brainstem necessary for early waveform generation,18 a disturbance in the inner hair cells (especially if the pure-tone audiometric thresholds are in the severe-profound range19), a dysfunction of the spiral ganglion fibers themselves,20 or a defect in afferent transmission at the synapse between the inner hair cells and the VIIIth nerve.21 Questions still remain, however, in regard to the site(s) of the disorder and possible etiologic factors.22
AN/AD can occur in the absence of any other apparent medical condition6; however, a perinatal history of hyperbilirubinemia, anoxia, or infectious diseases (one or more of these) are all factors.8, 23, 24 Familial patterns of inheritance,10, 25 syndromic and nonsyndromic recessive inheritance,26, 27 have been described also. It is interesting to note that in this cohort, no family history of hearing dysfunction was reported (see Table 2). Thus, etiologic factors associated with AN/AD remain unclear. Reports of AN/AD, from single case studies27, 28 to small samples,4, 9, 29 have provided limited insight. Sininger30 reviewed the incidence of the AN/AD profile in NICU infants and found a range in the literature extending from 5.3% to 14.8% (mean: 10.8%).
The present study found an incidence estimate of 24.1% of the AN/AD profile in the NICU, a higher rate than has been reported previously. It is possible that this cohort reflects a more at-risk population because it is a regional perinatal center site. Of course, neurologic immaturity may have been a factor here, and additional follow-up will confirm the AN/AD diagnosis. Regarding incidence, Rance et al8 found in a retrospective study that 20 (0.38%) of the 5199 children assessed between 1991 and 1996 had absent ABRs and present CM. Of the 5199 infants, 109 (2.09%) had some type of permanent hearing deficit. Of the 109 infants identified with some type of hearing loss, 11% (1 in 9) were subsequently diagnosed with AN/AD, and 97 (89%) were subsequently diagnosed with a sensory-type hearing loss.
Hall et al31 concluded that the combined ABR/OAE screening was the only means to avoid missing infants with the AN/AD profile. They suggested 2 feasible approaches: (1) use the combined ABR/OAE-screening method on those infants with neurologic risk factors or (2) use the combined ABR/OAE-screening method on all infants who are admitted to the NICU. In our sample, the risk factors did not prove sensitive enough to identify the bulk of infants who fit the AN/AD profile; thus, the second approach (ie, 2-phase screening using both ABRs and OAEs of all NICU discharges) is recommended. Programs implementing only OAEs in their NICU test protocol will miss this profile. It should be noted that an all-ABR protocol would also refer appropriate infants into a diagnostic setting, although the administration of the OAE heightens the alert and is easily administered.
A recent study by de Hoog et al32 found that serum concentrations above the therapeutic range of vancomycin, tobramycin, or furosemide were not related to failure on automated ABR in at-risk neonates. Of all risk factors analyzed in the present study, only 3 (hyperbilirubinemia and administration of vancomycin and furosemide) had any discriminating power. However, when using the powerful logistic-regression tool, no factors or combinations were effective in predicting the AN/AD profile. There are possible explanations for this result: (1) insufficient statistical power; (2) this NICU population is so diverse that 1 profile could not be developed to distinguish between the 2 groups; or (3) there may be other factors that are presently unknown. Starr21 reports that etiology is unknown in 40% of individuals with AN/AD. Our data echo this finding.
Approximately one third (33.9%) of infants with the AN/AD profile had a GA of 38 weeks. This finding suggests that, for all intents and purposes, these children were admitted to the NICU for reasons other than prematurity and may well have been missed had they been admitted to the well-infant nursery, in which an OAE-screening technique is typically used. Of the 20 infants that Rance et al8 examined and ultimately diagnosed with AN/AD, 13 (65%) had a GA of 38 weeks, also supported by our findings.
Consistent with Starr,21 there seems to be no gender preference in our sample. In contrast to his findings of 96% of cases having bilateral AN/AD results, our sample reflected a large proportion (88 [76%]) of unilateral findings of the AN/AD profile, supporting the concept that diagnostic evaluation is critical to confirm AN/AD.
Clearly, the next step is longitudinal evaluation of these children. A study is under way to clarify how many cases with the AN/AD profile represent auditory immaturity, transient neurologic abnormality,33, 34 or AN/AD. Care must be taken not to diagnose AN/AD without sufficient evidence. All of the following assessments should be included for an infant AN/AD work-up: ABR using separate trials of condensation and rarefaction clicks; OAEs; acoustic reflex threshold testing or middle-ear muscle testing; and behavioral audiometry using visual reinforcement audiometry.5, 34 Consideration should be given to the use of slower than typically administered ABR click rates (ie, 11.1 per second for enhancement of waveform morphology). Now that universal newborn hearing screening is mandated in 38 states and the District of Columbia (as of 2004), the ability to detect the AN/AD configuration of findings is something that we must be prepared to recognize and address.
ACKNOWLEDGMENTS
We thank Alex Cirilo, Alex Goenaga, Ashley Navarro, Flerida Noyola, Madelyne Rosario, and Maria Olmeda-Jenkins, MS, for their contributions in data collection. We also thank the 2 reviewers for thoughtful and insightful comments.
FOOTNOTES
Accepted May 2, 2005.
No conflict of interest declared.
REFERENCES
Starr A, McPherson D, Patterson J, et al. Absence of both auditory evoked potentials and auditory percepts dependent on timing cues. Brain. 1991;114 :1157 –1180
Starr A, Picton TW, Sininger Y, Hood LJ, Berlin CI. Auditory neuropathy. Brain. 1996;119 :741 –753
Sininger Y, Oba S. Patients with auditory neuropathy. In: Sininger Y, Starr A, eds. Auditory Neuropathy: A New Perspective on Hearing Disorders. San Diego, CA: Singular Thomson Learning; 2001:15 –35
Rance G, Cone-Wesson B, Wunderlich J, Dowell R. Speech perception and cortical event related potentials in children with auditory neuropathy. Ear Hear. 2002;23 :239 –253
Starr A, Picton TW, Kim R. Pathophysiology of auditory neuropathy. In: Sininger Y, Starr A, eds. Auditory Neuropathy: A New Perspective on Hearing Disorders. San Diego, CA: Singular Thomson Learning, 2001:67 –82
Kraus N. Auditory neuropathy: an historical and current perspective. In: Sininger Y, Starr A, eds. Auditory Neuropathy: A New Perspective on Hearing Disorders. San Diego, CA: Singular Thomson Learning; 2001:1 –15
Madden C, Rutter M, Hilbert L, Greinwald JH Jr, Choo DI. Clinical and audiological features in auditory neuropathy. Arch Otolaryngol Head Neck Surg. 2002;128 :1026 –1030
Rance G, Beer DE, Cone-Wesson B, et al. Clinical findings for a group of infants and young children with auditory neuropathy. Ear Hear. 1999;20 :238 –252
Dunkley C, Farnsworth A, Mason S, Dodd M, Gibbin K. Screening and follow up assessment in three cases of auditory neuropathy. Arch Dis Child. 2003;88 :25 –26
Wang Q, Gu R, Han D, Yang W. Familial auditory neuropathy. Laryngoscope. 2003;113 :1623 –1629
Shallop JK, Peterson A, Facer GW, Fabry LB, Driscoll CL. Cochlear implants in five cases of auditory neuropathy: postoperative findings and progress. Laryngoscope. 2001;111 :555 –562
Madden C, Hilbert L, Rutter M, Greinwald J, Choo D. Pediatric cochlear implantation in auditory neuropathy. Otol Neurotol. 2002;23 :163 –168
Trautwein P, Shallop J, Fabry L, Friedman R. Cochlear implantation of patients with auditory neuropathy. In: Sininger Y, Starr A, eds. Auditory Neuropathy: A New Perspective on Hearing Disorders. San Diego, CA: Singular Thomson Learning; 2001:203 –232
Sininger YS, Trautwein P. Electrical stimulation of the auditory nerve via cochlear implants in patients with auditory neuropathy. Ann Otol Rhinol Laryngol Suppl. 2002;189 :29 –31
Peterson A, Shallop J, Driscoll C, et al. Outcomes of cochlear implantation in children with auditory neuropathy. J Am Acad Audiol. 2003;14 :188 –201
Davis H, Hirsh S. A slow brainstem response for low-frequency audiometry. Audiology. 1979;18 :445–461
Worthington D, Peters J. Quantifiable hearing and no ABR: paradox or error Ear Hear. 1980;1 :281 –285
Kraus N, Ozdamar O, Stein L, Reed N. Absent auditory brain stem response: peripheral hearing loss or brain stem dysfunction Laryngoscope. 1984;94 :400 –406
Prieve BA, Gorga MP, Neely ST. Otoacoustic emissions in an adult with severe hearing loss [published correction appears in J Speech Hear Res. 1991;34:703]. J Speech Hear Res. 1991;34 :379 –385
Sininger Y, Hood L, Starr A, Berlin C, Picton T. Hearing loss due to auditory neuropathy. Audiol Today. 1995;7 :10 –13
Starr A. The neurology of auditory neuropathy. In: Sininger Y, Starr A, eds. Auditory Neuropathy: A New Perspective on Hearing Disorders. San Diego, CA: Singular Thomson Learning; 2001:37 –49
Rapin I, Gravel J. "Auditory neuropathy": physiologic and pathologic evidence calls for more diagnostic specificity. Int J Pediatr Otorhinolaryngol. 2003;67 :707 –728
Stein L, Tremblay K, Pasternak J, Banerjee S, Lindemann K, Kraus N. Brainstem abnormalities in neonates with normal otoacoustic emissions. Sem Hear. 1996;17 :197 –213
Shapiro SM. Bilirubin toxicity in the developing nervous system. Pediatr Neurol. 2003;29 :410 –421
Shivashankar N, Satishchandra P, Shashikala HR, Gore M. Primary auditory neuropathy—an enigma. Acta Neurol Scand. 2003;108 :130 –135
Starr A, Michalewski HJ, Zeng FG, et al. Pathology and physiology of auditory neuropathy with a novel mutation in the MPZ gene (Tyr145–>Ser) [published correction appears in Brain. 2003 Jul;126:1718]. Brain. 2003;126 :1604 –1619
Varga R, Kelley PM, Keats BJ, et al. Non-syndromic recessive auditory neuropathy is the result of mutations in the otoferlin (OTOF) gene. J Med Genet. 2003;40(1) :45 –50
Song JH, Koslo E. A case of unilateral auditory neuropathy. Presented at: the New York State Speech Language Hearing Association; April 15–17, 2004; Albany, NY
Sheykholeslami K, Kaga K, Kaga M. An isolated and sporadic auditory neuropathy (auditory nerve disease): report of five patients. J Laryngol Otol. 2001;115 :530 –534
Sininger Y. Overview of auditory neuropathy. Audiology Online. July 8, 2004
Hall J III, Smith SD, Popelka GR. Newborn hearing screening with combined otoacoustic emissions and auditory brainstem responses. J Am Acad Audiol. 2004;15 :414 –425
de Hoog M, van Zanten BA, Hop WC, Overbosch E, Weisglas-Kuperus N, van den Anker J. Newborn hearing screening: tobramycin and vancomycin are not risk factors for hearing loss. J Pediatr. 2003;142 :41 –46
Edwards CG, Durieux-Smith A, Picton TW. Auditory brainstem response audiometry in neonatal hydrocephalus. J Otolaryngol Suppl. 1985;14 :40 –46
Kileny PR, Robertson CMT. Neurological aspects of infant hearing assessment. J Otolaryngol Suppl. 1985;14 :34 –39(Abbey L. Berg, PhD, Jacly)