Congenital hearing loss is a condition that affects 1 to 2 per 1,000 neonates. Although it is one of the more common birth defects, it is an invisible condition without outward signs until much later in life. Although the parents of children with profound hearing loss may have concerns about their children's hearing, this is often not the case with children who have only moderate to severe hearing loss. In one recent study, less than 10% of the parents of an infant with hearing loss between 40 and 80 decibels (dB) were concerned about his or her hearing at the time the hearing loss was diagnosed.1
As pediatricians, we are not effective in identifying hearing loss in a healthy infant. Most infants with all but the most profound hearing loss will startle to loud noises because of the vibration produced by the sound. Most hearingimpaired infants will laugh and learn to babble at appropriate ages. Orientation to sound can be part of the developmental assessment of an infant, but the response may be inconsistent, even in an infant with normal hearing. Because parents are eager for children to begin talking, many will interpret random vocalizations as first words. As a result, hearing loss is often not suspected by the parents or the pediatrician until language development is significantly delayed. Prior to the development of newborn screening for hearing loss, the average age for identification of children with severe or profound hearing loss was 24 to 30 months. Children with lesser degrees of hearing loss (which may be severe enough to affect language development and articulation) and unilateral hearing loss may not be identified until 4 to 5 years old. This delay can have lifelong consequences.
DEVELOPMENT OF NORMAL HEARING
Just as lack of visual stimulation during a critical period prevents the development of visual pathways, lack of auditory stimulation affects the development of auditory pathways. Mice deafened at birth show a marked decrease in the size of the cochlear nucleus.2 This change does not occur if the animals are deafened after hearing develops (7 days) and is reversible in gerbils if hearing is restored during the first week of life.3 Similar findings have been observed in other animals. Although comparable studies cannot be done in humans, a study of 7 profoundly deaf individuals showed that the cell size in the ventral cochlear nucleus was inversely correlated with the duration of deafness.4
The development of normal hearing is a process that begins at 16 to 20 weeks of gestation with the growth of neurons in the brain stem auditory pathway.5 By 28 weeks of gestation, acousticomotor reflexes are present and an auditory brain stem response (ABR) can be recorded. At term, an infant can distinguish sounds of different frequencies and intensities, although the auditory cortex is only half of the adult thickness and the laminar pattern will not begin to develop until 4 months of age. Six-month-old infants have already learned to distinguish the speech sounds of the language spoken around them from the speech sounds of other languages. Nerve conduction times through the auditory pathways at 1 to 2 years are similar to those of an adult, whereas neural migration into the auditory cortex continues until 2 to 3 years.
The ability to localize sound also develops rapidly during the first year of life, when it approaches that of an adult. Full development of this skill then continues slowly until it reaches the adult level at 5 years. Children with unilateral atresia of the external auditory canal show little improvement in speech processing in a noisy environment or sound localization following surgical correction of the atresia.6 Children with significant conductive hearing loss during infancy had persistent abnormalities of binaural function as assessed by ABR at 5 to 7 years despite normal hearing at the time of the ABR.7 These studies underscore the importance of early identification and rehabilitation of infants with congenital hearing loss.
SCREENING NEWBORNS FOR HEARING LOSS
The Joint Committee on Infant Hearing has recommended that infants with significant hearing loss be identified by 3 months of age and receive intervention by 6 months of age. Initial methods of screening for hearing loss in infants were based on the observation of behavioral responses such as blink, grimace, or stretch following a sudden loud noise. They required specially trained personnel and were somewhat unreliable, both of which prevented their use in universal screening.
The lack of reliability of behavioral responses led to the development of criteria to identify infants at high risk for hearing loss. This smaller group of infants could then be evaluated using the more expensive and time-consuming ABR. Initially, these criteria included family history of hearing loss, birth weight less than 3 pounds, 5 ounces, hyperbilirubinemia, craniofacial anomaly, and congenital infection. They have since been expanded to include multiple courses of ototoxic medications, mechanical ventilation for more than 5 days, stigmata of a syndrome known to include hearing loss, and low Apgar scores. With the use of these criteria, the prevalence of significant hearing loss in this group is between 2% and 10%, depending on the population. Although selective screening of high-risk infants identifies a group of infants with a higher prevalence of hearing loss, it identifies only 40% to 50% of infants with congenital hearing loss.8,9
Most infants who are at high risk for hearing loss will be cared for in the level II or III nursery of a tertiary care referral center. These centers often have the specialized equipment necessary for neonatal hearing testing and the audiologists necessary to perform the tests. Universal neonatal screening for hearing loss requires a method that can be done inexpensively by minimally trained personnel because smaller hospitals may not have an audiologist. The rapid development of microelectronics in the past 20 years has brought about the instrumentation necessary for universal screening. Currently available methods include automated ABR and otoacoustic emissions (OAEs). Since the introduction of these methods, more than 20 states have enacted legislation requiring neonatal hearing screening for some or all infants.
ABR testing uses a standardized auditory stimulus in the form of clicks of various frequencies to elicit electroencephalographic activity. The waveform that results in the first 10 milliseconds after the stimulus is summed over many repetitions. This response depends on normal function of the cochlea, auditory nerve, and brain stem, and changes with age. Although an ABR can be recorded at 27 to 28 weeks of gestation, there is a longer latency and the waves are poorly formed. By 33 to 35 weeks of gestation, the waveform is more stable and similar to that of an older child or adult. Maturation is complete by 12 to 18 months.
When ABR testing is done by an audiologist, many electrodes are placed over the head, the intensity of the sound can be varied, and a threshold can be obtained. When automated ABR is used in screening, the response to clicks at a single intensity (usually 30 to 40 dB) is obtained and compared with a standard response, and a result of either "pass" or "refer" is recorded. The procedure involves placement of a disposable foam-cushioned circum-aural earphone that covers the entire ear and three electrodes on the forehead, nape of the neck, and mastoid. Adhesive backing on the foam of the earphone helps eliminate extraneous sounds. Testing can be influenced by earphone position, by collapse of the ear canal, and, somewhat, by environmental noise. Muscular artifact can also mask the electroencephalographic response, resulting in an incomplete test. Abnormal findings can also be obtained in the presence of a neurologic condition that results in asyhchrony of nerve firing in response to the stimulus.
Comparison of Methods for Screening Newborns for Hearing Loss
OAEs refer to sounds generated in the inner ear that are thought to originate in the outer hair cells of the cochlea. Although they occur spontaneously in 50% to 60% of ears, newborn screening uses clicks similar to those used in ABR testing to produce transient evoked otoacoustic emissions (TEOAEs). TEOAEs can be obtained in almost all individuals whose hearing is 20 dB or better and are absent if the auditory threshold is less than 30 to 40 dB. An earpiece that contains the microphone necessary to record the OAEs is placed in the auditory canal, clicks are presented, and the responses are recorded. In many early reports of OAE screening, before automated interpretation was available, results were generally reviewed by an audiologist or through an adaptation of an automated computer interface. However, automated interpretation of TEOAE is now available in the form of stand-alone units with a cost of $4,000 to $5,000. Another form of OAE, distortion-product OAE (DPOAE), is also available, although without automated interpretation at this time. In DPOAE, two tones are presented simultaneously, which results in an emission of a predictable single frequency. Because the response is recorded in only a narrow range of frequencies, DPOAE is less sensitive to ambient noise levels than is TEOAE.
A normal evoked OAE response requires normal functioning of the cochlea and peripheral hearing system. Unlike ABR, OAE testing allows identification of conductive hearing loss. Although this may be an advantage in the infant with multiple episodes of otitis media, it can be a disadvantage in screening newborns who may still have amniotic fluid in the middle ear. Debris (eg, vernix) in the auditory canal also influences results.
Both TEOAE and automated ABR have been used in universal newborn screening programs; they are compared in the table. With either method, the best results are obtained if the infant is screened just before discharge and while sleeping. Because TEOAE is faster and uses less expensive consumables, it is often used as an initial screening test, with automated ABR used to rescreen infants who do not pass the initial screening, often before discharge. A recently introduced device performs both OAE and ABR at the discretion of the tester. The results of universal newborn screening programs in several states have allowed comparison of the two methods and measures of the effectiveness of newborn hearing screening.
RESULTS OF UNIVERSAL SCREENING PROGRAMS
Mason and Herrmann reported the results of universal screening for all infants born in a single hospital in Hawaii during a 5-year period.8 Using automated ABR testing, they identified 15 infants with hearing loss among 10,372 infants screened. Prior to discharge, 4% of the infants did not have successful screening, more than half because of myogenic interference. Only half of these infants returned for screening compared with 90% of the infants who did not pass initial screening. Only half of the hearing-impaired infants were neonatal intensive care unit (NICU) graduates. On follow-up, 2 of the 8 healthy, hearing-impaired newborns had language delay; 1 of these infants was not amplified until 29 months of age. All 3 of the hearing-impaired NICU graduates for whom follow-up data were available had significant language delays. Of note, the NICU graduates tended to be older at the time of identification and at the time of amplification.
In a limited universal screening program at 26 hospitals, the Colorado program screened almost 42,000 infants (approximately 60% of the infants born in the state) during 5 years and identified 94 infants with sensorineural hearing loss and 32 infants with conductive hearing loss.10 Although most of the hospitals used screening ABR testing, one hospital used OAE testing by nurses with subsequent review by an audiologist, and several hospitals used conventional ABR testing by audiologists rather than purchase new equipment. Approximately 93% of the infants passed the initial screening, but less than half of the infants who did not pass returned for follow-up. The calculated cost of identifying each hearing-impaired infant was $9,600, but this may be only half as much if the prevalence of hearing loss in the children who did not return for testing was the same as that in the children who were retested.
Texas also had a universal screening program at selected hospitals.11 Both automated ABR and OAE testing were used, in addition to OAE followed by ABR testing for infants who did not pass. As expected, more infants were passed using OAE with ABR testing (98%) than using either technology alone (80% to 95% for OAE and 95% to 99% for ABR). Overall, 68% of the infants who did not pass initial testing returned for follow-up. The return rate was improved during the 3½ years of the study by education of parents and pediatricians.
Rhode Island mandated universal screening beginning in 1993 and had the first report of statewide universal screening.9 The initial screening consisted of TEOAE testing with an immediate ABR test for patients from the NICU who did not pass. Infants from the normal nursery were scheduled for a second TEOAE test in 2 to 6 weeks, and infants who did not pass this were then scheduled for an ABR screening test. All infants who did not pass the ABR screening test (1% of the total population) were then referred for a full audiologic evaluation.
In the first 4 years of the Rhode Island screening program, 111 infants were identified with hearing loss, resulting in a prevalence rate of 2.12 per 1,000 infants. During the course of the first 4 years, the mean age of confirmation of hearing loss decreased from 8.7 to 3.5 months and the age of amplification decreased from 13.3 to 5.7 months, meeting the Joint Committee on Infant Hearing goal of intervention by 6 months of age. More than 99% of the infants were screened with TEOAE testing while in the hospital, and 85% of the infants referred for additional testing presented for rescreening. Although these results are among the best reported, Rhode Island has only 7 hospitals with newborn nurseries and it may be difficult for other states to achieve these results.
ISSUES IN UNIVERSAL SCREENING FOR HEARING LOSS
Universal screening of newborns for hearing loss is effective only if the child who does not pass the initial screening is retested and intervention instituted if a hearing loss is identified. Pediatricians are important for the success of these programs. The results of the initial screening must be communicated from the physician who cares for the infant in the nursery to the pediatrician, if these are two different individuals. Families may need encouragement during the harried first weeks home with the new infant to make arrangements for retesting. Once a hearing loss is identified, it is often the primary care provider who makes the referral for early intervention and helps ensure that the child is receiving appropriate services.
Although newborn screening for hearing loss appears to be beneficial, several objections have been raised to universal screening. Opponents cite the high cost of screening a low-risk population and the anxiety raised by a false-positive result. At approximately $10,000 per child identified, screening for hearing loss is less expensive than newborn blood screening tests for most conditions.10 Although the initial costs of setting up a screening program may seem high, they must be balanced against the additional costs of educating a hearing-impaired child in whom diagnosis is delayed.
Watkin et al. found that although 15% of mothers felt somewhat anxious about the screening, less than 1% were very worried and more than 97% felt the screening was worthwhile.12 There were no differences between the mothers of infants who passed the screening and those of infants who did not. At rescreening, 3.5% of the mothers were very worried, but 97% felt the test was worthwhile. Thus, anxiety affected only a few of the mothers.
Another objection to universal newborn screening for hearing loss is the lack of objective evidence that earlier intervention results in improved outcome. One recent study compared the language quotients on the Minnesota Child Development Inventory of infants whose hearing loss was identified by 6 months with infants whose hearing loss was not identified until after 6 months.13 The language quotient (age equivalent of attained language skills divided by chronologic age) of children with normal cognition (intelligence) was significantly better in the group that was identified early (mean = 91) than in the group that was identified later (mean = 70). This degree of difference was found at all ages at follow-up between 13 and 36 months and with all methods of communication (oral vs sign or combination). The difference between language quotients of the two groups was greater with more severe hearing loss. Because almost all of the infants (96%) received intervention through the same agency, variations in intervention sources were unlikely to account for the observed difference. Although the group that was identified later had a lower cognitive quotient (88 in the early group vs 76 in the later group), the difference in language quotient persisted when the children were stratified by cognitive quotient. This suggested that early identification and intervention was responsible for the higher language quotient in the group that was identified early. There was no difference in the language quotients in the subgroups of children who were identified later (6 to 12 months, 12 to 18 months, 18 to 24 months, and older than 24 months), which provides additional evidence that auditory stimulation in the first few months of life is critical for normal language development. In addition, preliminary studies are finding improved language development in profoundly deaf children with earlier placement of cochlear implants.14
Objections to universal hearing screening are also raised by the Deaf community, where deafness is considered to be an alternative culture and not a handicap. Many families of profoundly deaf children will choose cochlear implantation, a procedure that often causes separation of the deaf child from the Deaf community. However, most hearing-impaired children are born into families with normal hearing where the child's hearing loss impairs family relationships. With input from audiologists and pediatricians, parents of hearing-impaired children should have the maximum number of options available to them for the habilitation of their child. This is achieved only with early identification of the hearing-impaired child.
GENETICS OF HEARING LOSS
Although opponents of universal screening feel selective screening of high-risk infants is more cost effective, more than half of all hearingimpaired children do not have a history that suggests they are at increased risk for hearing loss. Many of the hearing-impaired infants not identified using selective screening will have an inherited nonsyndromic hearing loss. During the past few years, mutations of multiple genes have been identified that result in nonsyndromic hearing loss. They include 16 autosomal dominant forms (designated DFNA1-DFNA17), 18 autosomal recessive forms (DFNB1-DFNB21), and 4 Xlinked forms (DFN1-DFN6). Knowledge of these genes and the mutations that result in hearing loss often allows genetic testing of a hearingimpaired child.
Approximately 60% to 70% of genetic hearing loss is inherited as an autosomal recessive trait. Mutations in the gene for connexin-26 (also known as gap junction protein J32) are thought to account for as much as 50% of autosomal recessive hearing loss.15 One recent study found that 3% of the general population in the midwestern United States carried mutations in this gene and that 2.5% of the population carried the most common mutation, 35delG (also known as 30delG).16 Based on the carrier frequency of all mutations in connexin-26, 10% to 20% of all congenital hearing loss is because of mutations in this gene (0.22 per 1,000 infants). Estivili et al. found that 37% of hearing-impaired children without a family history of hearing loss had mutations in connexin26.17 Without universal screening of newborn hearing, children with inherited hearing loss are often not diagnosed as hearing-impaired until after another child, possibly also hearing impaired, has been born into the family.
Approximately 30% of inherited hearing loss is dominantly inherited. Although a positive family history would be expected in most children with this form of hearing loss, that family history may not be known while the infant is in the nursery. In a Rhode Island study of universal hearing screening, 24 families were found to have a family history of hearing loss, but 11 of these families were not aware of this history at the time of the infant's hearing screening.9 Thus, inherited forms of hearing loss, which account for more than half of all congenital hearing loss and have the highest risk of recurrence, are less likely to be identified by selectively screening high-risk infants. In the future, samples for molecular testing for selected mutations might be obtained at the time of the initial audiologic evaluation with subsequent genetic counseling and testing based on the results of the "genetic screening" and the audiogram.
Universal screening of newborns for hearing loss meets the criteria for an effective screening test. The prevalence of congenital hearing loss is actually higher than the combined prevalences of conditions tested for with newborn blood screening. With the use of automated ABR or TEOAE, the test can be done easily on a large number of patients. Delayed diagnosis of congenital hearing loss can have long-term consequences and often contributes to the increased expense of educating the hearing-impaired child. These educational expenses can be decreased by early educational and auditory intervention, which ultimately makes hearing screening cost effective. In addition, no dollar amount can be placed on the improved parent-child relationship that results when a hearing-impaired child can hear and interact through language with his or her parent.
Even the most effective newborn screening program does not exclude the possibility of hearing loss in an older infant or child. Some dominantly inherited forms of hearing loss are progressive and may not be evident until school age. Congenital infection with cytomegalovirus often results in a later-onset, progressive, and fluctuating hearing loss. Otitis media and middle ear effusions can also result in significant conductive hearing loss that may be severe enough to affect language development. Children with craniofacial anomalies (eg, cleft palate) and Down syndrome are especially prone to these problems and should be monitored carefully and frequently for hearing loss. Larger practices may even find it cost effective and time efficient to purchase a handheld TEOAE screening device for hearing screening in the office. Awareness of the possibility of later-onset hearing loss and early referral for testing as soon as it is suspected can rninimize the effects of an identified hearing loss and maximize language development.
During the next few years, many hospitals will set up newborn screening programs for hearing loss. A team composed of audiologists, otolaryngologists, and pediatricians should be assembled to decide such issues as testing method, screening protocol, and personnel responsible for performing the test and follow-up. Each of these issues contributes to the success or failure of a screening program. A difficult test or one that requires excessive time is more likely to be "missed" during the brief hospitalization of a healthy newborn; audiologists can help choose the most appropriate testing equipment.
Testing has been done by various personnel, including nurses and nurse's aides, electroencephalogram technicians, respiratory therapists, and audiologists. Two hospitals in Texas where testing was done by nurses discontinued testing because it required excessive time from a nursing staff that already had too many patient care responsibilities.11 Personnel who normally work the day shift (eg, electroencephalogram technicians) are unlikely to volunteer for the night shift, which is when hearing screening is often done; this must also be taken into account when deciding who will perform the testing. Although audiologists may be more committed to hearing screening, it is the pediatrician who has rapport with the family and is more likely to be successful in encouraging a family to return for followup. The support of the entire team is necessary for a successful screening program. It is only through universal newborn screening for hearing loss and continued vigilance for signs of later-onset hearing loss that hearing-impaired children will have the best chance for normal language development and later academic success.
1. Watkin PM, Baldwin M. Confirmation of deafness in infancy. Arch Dis Child. 1999;81:380-389.
2. Tierney TS, Russell FA, Moore DR. Susceptibility of developing cochlear nucleus neurons to deafferentationinduced death abruptly ends just before the onset of hearing. J Comp Neurol. 1997;378:295-306.
3. Pasic TR, Moore DR, Rubel EW. Effect of altered neuronal activity on cell size in the medial nucleus of the trapezoid body and ventral cochlear nucleus of the gerbil. J Comp Neurol. 1994;348:111-120.
4. Moore TK, Niparko JK, Miller MR, et al. Effect of profound deafness on the central auditory nucleus. Am J Otol. 1994;15:588-595.
5. Sininger YS, Doyle KJ, Moore JK The case for early identification of hearing loss in children: auditory system development, experimental auditory deprivation and development of speech perception and hearing. Pediatr Clin North Am. 1999;46:1-14.
6. Breier JI, Gray L. Pre- and postoperative source localization in patients with unilateral atresia. Association for Research in Otolaryngology Abstracts. 1993;16:40.
7. Gunnarson AD, Finitzo T. Conductive hearing loss during infancy: effects on later auditory brain stem electrophysiology. Journal of Speech and Hearing Research. 1991,34:12071215.
8. Mason JA, Herrmann KR. Universal infant hearing screening by automated auditory brainstem response measurement. Pediatrics. 1998;10:221-228.
9. Vohr BR, Carty LM, Moore PE, Letourneau K. The Rhode Island Hearing Assessment Program: experience with statewide hearing screening (1993-1996). J Pediatr. 1998;133:353-357.
10. Mehl AL, Thomson V. Newborn hearing screening: the great omission. Pediatrics. 1998;101:E4. Available at: www.pediatrics.org/cgi/content/full/101/1/e4.
11. Finitzo T, Albright K, O'Neal J. The newborn with hearing loss: detection in the nursery. Pediatrics. 1998;102:14521460.
12. Watkin PM, Baldwin M, Dixon R, Beckman A. Maternal anxiety and attitudes to universal neonatal hearing screening. BrJAudiol. 1998;32:27-37.
13. Yoshinaga-Itano C, Sedey AL, Coulter DK, Mehl AL. Language of early- and later-identified children with hearing loss. Pediatrics. 1998;102:1161-1171.
14. Nikolipoulos TP, O'Donoghue GM, Archbold S. Age at implantation: its importance in pediatric cochlear implantation. Laryngoscope. 1999;109:595-599.
15. Skvorak Giersch AB, Morton CC. Genetic causes of nonsyndromic hearing loss. Curr Opin Pediatr. 1999;11:551557.
16. Green GE, Scott DA, McDonald JM, Woodworth GG, Sheffield VC, Smith RJ. Carrier rates in the midwestem United States for GJB2 mutations causing inherited deafness. JAMA. 1999;281:2211-2216.
17. Estivili X, Fortina P, Surrey S, et al. Connexin-26 mutations in sporadic and inherited sensorineural deafness. Lancet. 1998;351:394-398.
Comparison of Methods for Screening Newborns for Hearing Loss