Introduction
The goal of universal newborn hearing screening is to identify infants with permanent bilateral or unilateral sensory hearing loss, or permanent conductive hearing loss including neural hearing loss (auditory neuropathy or dyssynchrony).1
In the past, the otoacoustic emission (OAE) test was most commonly used for universal newborn hearing screening.Reference Boo, Rohani and Asma2 Although the OAE test can be used to assess outer hair cell function, it does not detect abnormalities in the auditory nerve and pathways. In fact, a high incidence of auditory neuropathy (0.5–11 per cent) was found in high-risk infants.Reference Maris, Venstermans and Boudewyns3 Therefore, in high-risk neonates, the most current guideline recommends the automated auditory brainstem response (ABR) method as the first choice method for universal newborn hearing screening. Brainstem auditory evoked response audiometry has been proposed as the ‘gold standard’ test for evaluating auditory function in uncooperative children and infants younger than six months, for the diagnosis of auditory neuropathy spectrum disorder.Reference Unlu, Guclu and Yaman4 Additionally, after birth, the automated ABR test is less affected than the OAE test in infants with a collapsed ear canal, or with the presence of amniotic fluid in the ear canal or middle ear, leading to a lower referral rate.Reference Unlu, Guclu and Yaman4–Reference Van Dyk, Swanepoel de and Hall6
Most equipment for hearing screening cannot measure the hearing level of infants in the same way as the diagnostic brainstem auditory evoked response test. With developments in technology, some automated ABR machines can also measure the hearing level. As a result, the automated ABR test can be easily applied by technicians who are not audiologists but can be adequately trained to use it. Specifically, hearing levels shown on the monitor of the machine can be automatically marked, and this is considered as an objective test. The testers do not require any subjective responses from the infants to mark their hearing levels. As the automated ABR test does not require experienced testers to interpret the results, this test can be used as a screening tool to identify both hearing level and hearing loss severity.
The Joint Committee on Infant Hearing (2007) recommends identifying infants with moderate hearing loss that can affect their speech and language development.1 The incidence of permanent congenital hearing loss varies from 3 to 5 per 1000 in healthy baby infants; this increases 10–50 times to 3 to 5 per 100 in high-risk infants in the neonatal intensive care unit.1,Reference Unlu, Guclu and Yaman4,Reference Kuki, Chadha, Dhingra and Gulati7 Hearing loss can have an impact on cognitive, acquisition, psychosocial, language and speech development. If early diagnosis and intervention is achieved, this can result in better language development.
The primary objectives of this study were to assess the sensitivity and specificity of the OAE and automated ABR testing methods in the detection of hearing levels and hearing loss severity, particularly in high-risk infants.
Materials and methods
Subjects
The study population included 150 high-risk neonates admitted to the neonatal intensive care unit of Songklanagarind Hospital, Thailand. A total of 144 infants (288 ears) completed the 3 hearing tests. Their hearing was assessed using three objective tests: the transient evoked otoacoustic emission (TEOAE) test before discharge, and the automated ABR and diagnostic brainstem auditory evoked response tests conducted between July 2013 and December 2016. Because of the tight schedule for sedation managed by anaesthesiologists and a concern over contraindications of chloral hydrate (used to sedate the infants), the present research was carried out over a 40-month period. This study was approved by the ethics committee of the hospital. Parental consent was obtained prior to the hearing tests.
Procedures
All 144 high-risk infants were screened for hearing loss using: the TEOAE test before discharge (conducted by nurses at the neonatal intensive care unit), the automated ABR test for rescreening and the brainstem auditory evoked response test (performed by a trained technician at the audiology unit in the Otorhinolaryngology Head and Neck Surgery Service, in the same day). During the automated ABR and diagnostic brainstem auditory evoked response testing, careful ear examination and aural toilet were performed. An ear probe was then inserted into the external auditory canals. All infants were sedated with chloral hydrate, with informed consent from their parents.
Hearing screening equipment
The TEOAEs were elicited by a click stimulus via a transducer in the external ear canal. The sound response from the outer hair cells was recorded by a microphone. A TEOAE response reveals normal function of the outer hair cells in the inner ears. The Sentiero instrument was employed for the TEOAE screening (Path Medical, Germering, Germany). A non-linear click stimulus was used, with a stimulus level of 60–85 dB peak equivalent sound pressure level, signal-to-noise ratio of 6 dB, and frequencies of 1–4 kHz. The pass criteria were stimulus stability of over 80 per cent, artefacts of less than 20 per cent, signal-to-noise ratio of at least 6 dB, and three additional signal-to-noise ratios over 6 dB.
Automated ABRs were elicited by a click stimulus, using the Sentiero instrument (Path Medical). Three disposable electrodes were attached to the infant: a non-inverting electrode at the vertex of the head, a ground electrode on the forehead and the inverting electrode placed on an ear lobe.
The stimulus rates consisted of 100 μs clicks at 10 Hz, with an intensity of 30–70 dB normalised hearing level (dB nHL). The response was generated by a small probe positioned in the neonate's external ear canal. Based on the results, the hearing thresholds could be generated after presentation of up to 3000 stimuli. With the assistance of computer software that scores and interprets the brainstem responses obtained from the click stimuli, the hearing thresholds are automatically shown on a monitor. A specialist is not required to interpret or mark wave V.
The diagnostic brainstem auditory evoked responses were recorded using a Chartr diagnostic system (MCU-90; ICS Medical, Schaumburg, Illinois, USA). Two stimuli at intensities of 75 and 35 dB nHL, at 11.1–55.1 Hz, were utilised. Brainstem auditory evoked responses were used to estimate the hearing levels at 2–4 kHz.Reference Arnold, Roeser, Valente and Hosford-Dunn8,Reference Smith, Bale and White9 Evoked auditory brainstem potential responses were measured for each ear, with the electrodes at the vertex of the head, forehead and ear lobes. Wave V was interpreted by the technician, supervised by an otologist.
The chi-square test and Fisher's exact test were used for statistical analysis, with a significance level of p < 0.05.
Screening data analysis
The criterion for infants who failed the hearing screening and were later diagnosed with hearing loss was based on the absence of wave V at an intensity greater than 40 dB nHL, using the automated ABR and click diagnostic brainstem auditory evoked response testing methods.Reference Suppiej, Rizzardi, Zanardo, Franzoi, Ermani and Orzan10,Reference Benito-Orejas, Ramirez, Morais, Almaraz and Fernandez-Calvo11 Some neonates could pass the automated ABR and brainstem auditory evoked response hearing tests despite having mild hearing loss, because the cut-off point for hearing loss was 40 dB nHL.1
Statistical analysis
A total of 144 high-risk neonates completed the 3 hearing tests. Sensitivity, specificity, positive and negative predictive values, odds ratios with 95 per cent confidence intervals, and false negative and false positive rates were obtained. An infant was classified as having normal hearing when wave V was present at an intensity of 40 dB nHL or less with the diagnostic brainstem auditory evoked response test.
Results
Between July 2013 and December 2016, 150 infants were admitted to the neonatal intensive care unit of Songklanagarind Hospital; 144 of the infants (57 girls and 87 boys) were screened for hearing loss using the 3 testing methods. Six infants were excluded from the study as the tests could not be completed because of illness.
Of the 144 neonates, bilateral hearing loss was diagnosed in 22 (15.3 per cent) and unilateral hearing loss was diagnosed in 2 (1.4 per cent). The diagnoses were based on the results of the brainstem auditory evoked response wave V threshold, used as the gold standard for evaluating auditory function in infants younger than six months. The results of the brainstem auditory evoked response testing method were reported in decibels normalised hearing level, and converted to decibels estimated hearing levels (dB eHL) by deducting 10 dB from the normalised hearing thresholds recorded in the brainstem auditory evoked response test (dB nHL).Reference Arnold, Roeser, Valente and Hosford-Dunn8
Smith et al. defined normal hearing as a level of 20 dB HL or less, mild hearing loss as 21–40 dB HL, moderate hearing loss as 41–70 dB HL, severe hearing loss as 71–90 dB HL, and profound hearing loss as more than 90 dB HL.Reference Smith, Bale and White9 In our study, mild, moderate, and severe to profound hearing losses were identified in eight, eight and six infants, respectively (Table 1).
Table 1. Hearing test results based on diagnostic brainstem auditory evoked responses*
* With a 10 dB correction of dB normalised hearing level to dB estimated hearing level (dB eHL)
Because of the shortage of audiologists, the focus was only on moderate and severe to profound, permanent, bilateral sensory hearing loss likely to have significant effects on language, speech, and academic and social-emotional development. We advise parents of infants with mild, unilateral hearing loss to observe their child's behavioural responses to sound. If abnormal development or a deterioration in hearing loss is suspected, they can return to have hearing rechecked. The characteristics of high-risk infants confirmed as having moderate or severe to profound hearing loss are shown in Table 2.
Table 2. Distribution of risk factors and treatment of infants with moderate or severe bilateral hearing loss
*Candidate for cochlear implant. CHD = congenital heart disease; SNHL = sensorineural hearing loss; CHL = conductive hearing loss; NICU = neonatal intensive care unit
Of the 14 newborns with moderate or severe to profound, bilateral hearing loss, 3 did not return for follow up because of referral to their local hospitals. Four of the newborns had conductive hearing loss associated with craniofacial anomalies that resolved with grommet insertions, and one had conductive hearing loss with spontaneous resolution. Seven newborns were provided with hearing aids. Two of the seven newborns were candidates for cochlear implantation; however, they did not undergo cochlear implantation because of the lack of financial resources.
The universal newborn hearing screening aimed to identify permanent bilateral hearing loss of at least moderate severity. The findings showed that out of 144 high-risk infants in our hospital, 14 (9.7 per cent) were identified as having hearing loss. This rate is similar to that reported in another study, in which bilateral congenital hearing loss was identified in 7.9 per cent of high-risk infants.Reference Unlu, Guclu and Yaman4 The prevalence of bilateral congenital hearing loss was approximately 10 times higher than that in healthy baby infants.Reference Erenberg, Lemons, Sia, Trunkel and Ziring12,Reference Thompson, McPhillips, Davis, Lieu, Homer and Helfand13
One reason for the slightly higher prevalence in this study than in other studies might be because our hospital is a tertiary care hospital and it is the referral centre for southern Thailand. Therefore, our patients usually have higher risks and more complicated diseases. Additionally, in this study, 5 out of 144 neonates (3.4 per cent) showed normal transient evoked otoacoustic emission (TEOAE) test results with abnormal brainstem auditory evoked responses. Consequently, these infants were also suspected of having auditory neuropathy.
Table 3 compares the TEOAE and automated ABR test results in each ear of high-risk neonates, in terms of sensitivity, specificity, positive and negative predictive values, and accuracy. The sensitivity and specificity values indicate that the automated ABR test is more reliable. False positive results were observed in 27 out of 288 ears (9.3 per cent) with the TEOAE test, and in 19 out of 288 ears (6.5 per cent) with the automated ABR test. False negative results were observed in 10 out of 288 ears (3.4 per cent) with the TEOAE test, and in 4 out of 288 ears (1.3 per cent) with the automated ABR test. The sensitivity of the TEOAE test was 78.7 per cent and the specificity was 88.8 per cent. The sensitivity of the automated ABR test was 91.7 per cent and the specificity was 92.1 per cent. The TEOAE test had a positive predictive value of 57.8 per cent and a negative predictive value of 95.5 per cent. The automated ABR test had a positive predictive value of 69.8 per cent and a negative predictive value of 98.2 per cent.
Table 3. Sensitivity, specificity, positive and negative predictive values, and accuracy of TEOAE and automated ABR tests
TEOAE = transient evoked otoacoustic emission; ABR = auditory brainstem response
Of note, the click-automated ABR method, used in the present research to evaluate hearing function, is precise only at mid-frequencies between 2 and 4 kHz. As a result, mild hearing loss, and low or high frequency hearing loss, may be missed, leading to a false negative result.1,Reference Benito-Orejas, Ramirez, Morais, Almaraz and Fernandez-Calvo11
Discussion
In the past, OAE testing was used most often for screening the hearing of healthy baby infants and neonates in the neonatal intensive care unit. A number of the neonates in the neonatal intensive care unit who pass the OAE screening will still struggle with speech and language development. This is primarily because OAE testing cannot detect hearing loss caused by abnormalities of inner hair cells, auditory nerves or auditory brainstem pathways.
Nowadays, the Joint Committee on Infant Hearing recommends the automated ABR test for screening hearing in the neonatal intensive care unit. In our study, we compared the brainstem auditory evoked response test, which is considered to be the gold standard, to the new automated ABR equipment. The automated ABR testing method is able to show the hearing thresholds of infants based on the identification of wave V.
In this study, the sensitivity and specificity values were 78.7 per cent and 88.8 per cent for the transient evoked otoacoustic emission (TEOAE) test, and 91.7 per cent and 92.1 per cent for the automated ABR test, respectively. These results are similar to those of previous papers. Specifically, the TEOAE test has been reported to have high sensitivity but lower specificity than the automated ABR test. In general, the levels of sensitivity and specificity of the automated ABR machines were more than 85–90 per cent.Reference Boo, Rohani and Asma2,Reference Kuki, Chadha, Dhingra and Gulati7 The sensitivity of the TEOAE test was 97.1–100 per cent, and specificity was 48.3–91 per cent.1,Reference Unlu, Guclu and Yaman4,Reference Kuki, Chadha, Dhingra and Gulati7,Reference Suppiej, Rizzardi, Zanardo, Franzoi, Ermani and Orzan10,Reference Smets, Verrue and Dhooge14,Reference Van Straaten, Hille, Kok and Verkerk15 However, the sensitivity of the automated ABR test was over 90 per cent, and its specificity was the highest, reaching 100 per cent.1,Reference Unlu, Guclu and Yaman4,Reference Kuki, Chadha, Dhingra and Gulati7,Reference Suppiej, Rizzardi, Zanardo, Franzoi, Ermani and Orzan10
Kuki et al. reported that auditory steady-state response testing had 100 per cent sensitivity and 96.77 per cent specificity.Reference Kuki, Chadha, Dhingra and Gulati7 The author claimed that auditory steady-state response testing may be an ideal objective test to assess the hearing level of infants; however, a trained audiologist is needed to interpret the results.
The diagnostic brainstem auditory evoked response test needs to be carried out in a soundproof room, it is time-consuming to perform, and needs trained personnel such as audiologists to identify wave V.Reference Benito-Orejas, Ramirez, Morais, Almaraz and Fernandez-Calvo11 Since the 1980s, the automated ABR test has been used to objectively assess the integrity of the auditory pathway. It can provide a simple, quick, effective and non-invasive method for assessing cochlear and auditory nerve function, and can be conducted by non-professionals.Reference Unlu, Guclu and Yaman4,Reference Benito-Orejas, Ramirez, Morais, Almaraz and Fernandez-Calvo11 However, many companies use the basic criterion of ‘pass or refer’. The machine used in this study could report the hearing threshold levels. In other words, testers were able to address the severity of hearing loss in the infants screened.
In Thailand, the shortage of audiologists with a Master's degree, who are able to perform the diagnostic brainstem auditory evoked response test, is the major concern. Consequently, the information regarding hearing levels, obtained from the screening procedure using the automated ABR testing method, can help healthcare personnel. Specifically, the automated ABR test is able to separate mild hearing loss from moderate and severe to profound hearing loss. For instance, when the automated ABR test shows a hearing level worse than 50 dB nHL (40 dB eHL), the patient can be prioritised for the diagnostic brainstem auditory evoked response test and early intervention.Reference Unlu, Guclu and Yaman4
When the clinician has the information regarding an infant's hearing level, the severity of hearing loss can be emphasised to the parents. This allows them to realise the importance of a follow-up appointment for diagnostic tests. Furthermore, in cases of severe to profound hearing loss, the infants can be referred directly to tertiary care hospitals where cochlear implants are available. On the other hand, infants with moderate hearing loss can be referred initially to secondary care hospitals for the fitting of a hearing aid.
Therefore, the hearing levels obtained from this particular automated ABR test can help screeners to recognise the severity of hearing loss. This information can be useful for prioritising patients’ needs regarding further diagnostic evaluation and intervention, which is not always available in Thailand.
• Some automated auditory brainstem response tests can indicate hearing level
• This information can encourage a return for diagnosis and rehabilitation, especially for parents of infants with severe to profound hearing loss
• The test results can also indicate who should be referred to secondary hospitals (for hearing aids) or a tertiary hospital for cochlear implantation
• Because of problems in developing countries, this study focused on bilateral hearing loss only, and considered interventions as well as screening
Conclusion
The goal of universal newborn hearing screening is to identify infants with permanent bilateral or unilateral sensory hearing loss, or permanent conductive hearing loss including neural hearing loss. However, because of problems that arise in developing countries, we focused on bilateral hearing loss only, and considered interventions as well as universal newborn hearing screening.
Early hearing detection and intervention help to promote normal speech and language development. The automated ABR test is a simple, quick, effective and non-invasive method of evaluating the cochlea and function of the auditory pathway, and determining initial hearing levels. In addition, it does not require a great deal of experience or a specialist to interpret the results.
If the clinician knows the severity of the hearing loss, we can initially refer only those infants with moderate to severe hearing loss, and encourage their parents to return for follow up. However, the automated ABR test has higher costs associated with the disposable ear tips and electrodes. Regardless of the hearing results obtained, physicians need to advise the parents to observe their child's behaviour and language development, and to closely monitor for delayed-onset or progressive hearing loss, particularly in high-risk infants.
Competing interests
None declared
