Hearing screening and diagnosis in a large sample of infants in Central China

Abstract

Objective

To investigate the referral rate, prevalence and aetiology of neonatal hearing loss.

Methods

A total of 11,894 infants were screened by two-stage transient evoked otoacoustic emission testing. Those who failed were diagnosed by distortion product otoacoustic emission, 1000 Hz probe tone tympanometry and auditory brainstem response. The results of these tests were analysed by statistical software SPSS16.0.

Results

The initial referral rate was 17.36%. The rescreening referral rate was 21.29%. The referral rate of initial screening in maternity wards (15.37%) was lower than in neonatal intensive care unit wards (22%) (chi-square [χ ²], P < 0.05). There were 68 cases (106 ears) diagnosed with hearing loss (incidence 0.571%). Of these, 31 cases were conductive, 16 cases were sensorineural, and 21 cases were mixed hearing loss. The prevalence of hearing loss was 12.92% (38/294) in the bilateral referred group and 5.00% (30/600) in the unilateral referred group. The moderate/severe hearing loss was 33.33% (10/30) and 86.84% (66/76), respectively (χ ², P < 0.05). The causes of hearing loss included jaundice (24.56%, 14/57), infection (24.56%, 14/57), asphyxia (19.30%, 11/57), low birth weight (17.54%, 10/57) and other factors (14.04%, 8/57).

Conclusion

Bilateral referrals were more likely to suffer greater degrees of hearing loss than unilateral referrals. Jaundice, infection, asphyxia and low birth weight were the major aetiologies of neonatal hearing loss.

INTRODUCTION

Hearing impairment is one of the most common abnormalities at birth, and if undetected, can impede speech, language, and cognitive development. Bilateral hearing loss is found in1–3% of newborn infants from normal births, and 2–4% of infants in the neonatal intensive care unit (NICU) population.¹ In China, hearing and speech disability ranks number one out of five categories of disability. The existing population contains approximately 2 million people with hearing and speech disabilities, with an increase of approximately 2–4 million new cases annually.

Approximately 800,000 children from birth to age 8 are deaf. It is generally difficult to identify infant hearing loss in the first year after birth through routine physical examinations and parental observation. For traditional high-risk families, it is believed that only 50% of children with congenital hearing impairment could be found by routine checkups or parental observation. The time from birth to age 3 is critical for language development. Hearing screening is considered a necessary measure of early detection and intervention in the children with hearing loss.²

We here report preliminary results from the hearing screening protocol employed in the Hubei Provincial Women and Children's Hospital, the largest maternity and child care hospital in central China, with approximately 800 births per month.

METHODS

Data were collected by the department of otolaryngology, Hubei Provincial Women and Children's Hospital, central China, from October 2006 to May 2008. Of the 11894 infants included in the study, 8321 were from the maternity ward and 3573 were from the NICU ward. Of the 894 cases rescreened, 109 cases completed diagnostic tests at the end of the experiment.

Data on infants from the maternity ward were collected on a structured form, and included information on gender, body weight, family history, parents, maternal pregnancy history, antenatal care, details of the delivery and status at delivery. In addition, data on infants from the NICU ward included diagnoses, complications during the hospital stay and outcome at discharge.

Informed consent was obtained in accordance with the institutional review board-approved protocols of the hospital.

All infants underwent at least two hearing screening tests. Infants with positive transient evoked otoacoustic emission (TEOAE) hearing screening test results were referred for diagnostic tests such as distorted production otoacoustic emission (DPOAE), auditory brainstem response (ABR), tympanometry or computerized tomography (CT). The infants diagnosed with hearing loss were grouped according to the possible causes.

The screening programme consisted of three steps:

initial screening: 3–5 days after birth (TEOAE);

rescreening: 30–42 days after birth (TEOAE);

if the second screening test was positive, diagnostic tests including ABR, 1000 Hz probe tone tympanometry and DPOAE test were performed two months later.

TEOAE hearing screening test

The TEOAE hearing screening instrument used in this study was the AcuScreen Pro (Madsen, Taastrup, Denmark). The probe was inserted tightly in the external auditory canal after it was cleared over. The newborn was kept in a quiet state during testing. The TEOAE stimulus was a short sound (click) with an intensity of 79–84 dB peSPL at 80/second. The signal was overlayed 50–260 times. Using 0.8, 1.6, 2.4, 3.2 and 4.0 kHz as analysis frequencies, the passing analysis standard was as follows: (1) the signal repetition rate in two sets of buffer memory ≥50%; (2) the total reaction energy ≥5 dB SPL; (3) the signal-to-noise ratio ≥3 dB SPL. After setting the parameters, the results were determined automatically by the instrument as pass or referral. The environmental noise was generally less than 35–40 dB (A), when the baby slept after feeding.

DPOAE test

The DPOAE test instrument used was Capella Plus (Madsen). Two pure tones, F1 and F2, were used as stimulation sounds at two different intensities, L1 = 65 dB SPL, L2 = 55 dB SPL. The frequency ratio of F2/F1 was 1.2. The DPOAE was recorded at six frequency points for 500, 100, 1593, 2562, 3187 and 4031 Hz. The value of signal-to-noise ratio ≥5 at each frequency should be deemed to ‘pass’.

ABR test

ABR testing was operated in a shielded sound insulated room using the instrument Charter EP (Madsen) and TDH-39 ear headphones. If a baby was crying, he or she was given 10% chloral hydrate solution (0.4–0.5 mL/kg). After skin preparation, the recording electrode was placed on the forehead, the reference electrode was placed on both sides of the mastoid, and the ground line was placed in the middle of the glabella. Skin electrode impedance was below 5000 Ω. The stimulus was a click at 10 times per second masking the contralateral ear with white noise, the band-pass filter was 100–3000 Hz stacking at least 1000 times, and the analysis time was 10 millisecond.

Tympanometry

The Middle Ear Analyzer of tympanometry was the Otoflex100 (Madsen). The equipment was arranged to a 1000 Hz at 75 dB SPL (signal level accuracy, ±1.5 dB SPL), with a pump speed of 100 decaPascals (daPa) per second, and positive-to negative pressure sweep between +200 and −400 daPa. It was corrected by using the sound level meter (PISLM-2230, The B & K, Narum, Denmark) in the ANSIS3.6-1996 standard before the test. The single peaked tympanogram of 1000 Hz tympanometry was indicated to be normal middle ear function; the flat peaked tympanogram suggested the existence of the middle ear fluid.³

CT test

The CT scanner used was the GE Prospeed II (GE, Connecticut, Fairfield City, USA), and a cross-section helical scan was performed followed by a coronal multiplanar reformation. Scanning was done starting with a cross-section of the orbitomeatal line as the baseline, and continued with a bottomed ear canal scan up to the superior edge of the rock cone. Coronal reorganization of the scans was done parallel to the back wall of the maxillary sinus as baseline, and from the posterior border of the temporomandibular joint fossa to the anterior wall of the sigmoid sinus. The scan had a 2-mm slice thickness and slice distance. The scanning voltage was 120 kV with a current of 230 mA, by 2i style. Informed consent was obtained before conducting CT tests. None of the parents wanted to retest with the CT due to concerns of potential radioactive injury.

Data analyses

Statistical analyses were performed using medical statistics software SPSS16.0. The hearing screening results are presented as initial screening and rescreening according to NICU or obstetric ward. The variables include the numbers of infants eligible for screening, numbers actually screened, numbers who passed the tests, and numbers referred. Differences between the pass and referral rate in the NICU and obstetric ward groups were tested by Pearson's chi-squared test (χ ²). The results of TEOAE hearing screening and ABR test were compared using the analysis of contingency table χ ². To investigate the influence of risk factors, we grouped the infants and compared the incidence and degree of hearing loss in infants with or without risk factors using the analysis of contingency table χ ². All tests were two-tailed, and P < 0.05 was considered statistically significant.

RESULTS

Screening results

In this study 11,894 infants were screened, with a coverage rate of 90.62%. Of these, 9829 passed the initial screening, a pass rate of 82.64%. The referral rate was 17.36%. Of those infants who failed the initial screening, 894 completed the second screening; the rescreening rate was 43.29%. Among those rescreened, 785 cases passed the rescreening (pass rate 78.71%) and 109 cases failed, yielding a rescreening referral rate of 22.29% (Table 1).

Table 1

TEOAE hearing screening results

Screening	Numberof infants	Numbersactuallyscreened	Passed	Referred	Screeningrates (%)	Pass rate (%)	Refer rate (%)
Initial screening
NICU ward	4392	3573	2787	786*	81.35	78.00	22.00*
Obstetric ward	8733	8321	7042	1279	95.28	84.63	15.37
Total	13125	11894	9829	2065	90.62	82.64	17.36

Rescreening
NICU ward	786	491	415	76*	62.47	84.52	15.48*
Obstetric ward	1279	403	370	33	31.51	91.81	8.19
Total	2065	894	785	109	43.29	78.71	22.29

TEOAE, transient evoked otoacoustic emission; NICU, neonatal intensive care unit

Results are presented as numbers of infants

*The difference of the refer rate was significant between the two groups while initial screening (χ² = 76.53, P < 0.01) as well as rescreening (χ² = 16.96, P < 0.01)

The incidence of neonatal hearing loss

All 109 cases who failed the second hearing screening test were admitted to diagnostic audiology tests, which included ABR, tympanometry, DPOAE and CT testing, dependent upon appropriate parental informed consent and age restrictions (infants must be 3 months or older) for CT testing.

ABR results showed that 68 neonates (68/109, 62.39%) suffered from varying degrees of hearing loss; 25 cases were mild (30–50 dB nHL), 23 cases were moderate (50–70 dB nHL), 13 cases were severe (70–90 dB nHL) and seven cases were profound (90 dB nHL–∼). The DPOAE tests showed hearing impairment in 106 ears, 30 of which showed impairment in unilateral ear and 38 in bilateral ears, all of which were consistent with TEOAE results. The tympanograms demonstrated a flat peak in 70 ears, and a single peak in 148 ears. These results are summarized in Table 2.

Table 2

TEOAE test results and ABR test threshold distribution

	TEOAE test (person)		Tympanogram (ear)
ABR threshold(dB nHL)	Abnormal groupof single-ear	Binauralabnormal group	None peakshape (B)	A singlepeak (A)	Total(person/ear)
−30	30	11	0	112	41/112
−50	20	5	30	0	25/30
−70	8	15*	28	10	23/38
−90	2	11*	8	16	13/24
90–	0	7*	4	10	7/14
Total	60	49	70	148	109/218

TEOAE, transient evoked otoacoustic emission; ABR, auditory brainstem response

*The difference of ABR test threshold proportion was significant in different TEOAE screening results between the two groups (χ² = 32.39, P < 0.01)

Analysis of these results suggested that the incidence of neonatal hearing loss was 0.571% (68/11894), which included conductive hearing loss in 31 cases, sensorineural hearing loss in 16 cases and mixed hearing loss in 21 cases. Some guardians doubted the results of 1000 Hz tympanometry testing and after informed consent was obtained, they agreed to do the temporal bone CT test (17 cases). The results of these tests showed there were high-density materials in the middle ear cavity (17/17, 100%) whereas the 1000 Hz tympanograms showed a flat curve without a distinct peak.

Investigating the relationship between hearing screening and ABR testing, the results showed that the prevalence of hearing loss was higher in the group with bilateral hearing impairment referred by TEOAE testing, compared with unilateral impairment. The rate of severe or greater hearing loss was also higher in the group of bilateral hearing impaired referred by TEOAE testing (χ ², P < 0.05). Bilateral referred hearing screening was more likely to indicate hearing loss or more serious hearing loss.

All screened newborns were divided into two groups; those without risk factors and those with risk factors. The results showed that there were significant differences in the incidence of hearing loss between the two groups (χ ², P < 0.01), the former was 1.32‰ (11/8321), and the latter was 1.60% (57/3573) (Table 3).

Table 3

Comparison of neonatal hearing impairment in different groups

	ABR test hearing threshold distribution (dB nHL)
Group	–30*	30–50*	–70*	–90*	90–*	Total
Group of Risk factors	9	19	19	12	7	66
Group free of Risk factors	32	6	4	1	0	43
Total	41	25	23	13	7	109

ABR, auditory brainstem response

*Risk factors group and children without risk factors group, the difference of detection rate of hearing loss was significant (χ ² = 42. 81, P < 0.01)

Analysis of risk factors

Analysis of the information about risk factors in 57 cases diagnosed with hearing loss revealed that there were multiple risk factors coexisting in most cases, of which the primary diagnosis included: jaundice in 14 cases (14/57, 24.56%); neonatal asphyxia in 11 cases (11/57, 19.30%); premature birth and low birth weight in 10 cases (10/57, 17.54%); amniotic fluid or meconium inhalation in seven cases (7/57 12.28%); flu and pneumonia in seven cases (7/57, 12.28%); intrauterine infection in three cases (3/57, 5.26%); genetic factors (Down's syndrome, familial) in two cases (2/57, 3.51%); sepsis in one case (1/57, 1.75%); hypoxic-ischaemic encephalopathy in one case (1/57, 1.75%) and macrosomia and hypoglycaemia in one case (1/57, 1.75%).

DISCUSSION

OAE is a rapid, non-invasive, objective, sensitive method of testing cochlear dysfunction which has obtained increasing attention in clinical practices. This method has been widely recognized as a valid means by which to assess auditory function, especially for monitoring and evaluation in newborns, infants and young children with high risk of hearing loss. TEOAE is a form of acoustic frequency energy generated from the cochlea evoked by single transient acoustic signal stimulation after some latent periods, which can be recorded from the external auditory canal. TEOAE is already a mature technology for newborn hearing screening in developed countries.

It is generally believed that the best time for OAE screening is three days after birth.⁴ The amniotic fluid in the ear channel or middle ear might influence the screening result. OAE had the advantages of convenient, rapid, immediate detection after birth, but also had some limitations, including the failure to detect the dysfunction of the auditory nerve and brainstem auditory pathway. This technique is susceptible to the newborn state, the external auditory canal vernix and amniotic fluid in the middle ear to produce artefact.⁵ Other issues with OAE testing are the prevalence of false-positive results, and its unsuitability for use as a threshold value. Therefore, relying solely on OAE results would miss part of the central hearing impairment in children or produce false-negative results. ABR testing could reflect the activity of the cochlea, hearing nerve and brainstem auditory pathway. ABR testing was correlated with hearing sensitivity at the range of frequencies 1–8 kHz. It could be used to detect central hearing impairment. Consequently, the American Medical Association recommended a two-stage scheme for hearing screening, of which the initial screening was at 3 days old, and rescreening at about 42 days old, and if positive, further ABR examination is performed. According to the scheme, the false-positive rate in diagnosis of hearing loss was less than1%.⁶

After a two-staged OAE screening, referred infants were diagnosed using tests including ABR, DPOAE and 1000 Hz-probe-tone tympanometry at about age of three months. The results (Table 1) showed that the initial screening pass rate was 82.64% and the referral rate was 17.36%. In the maternity ward, the pass rate was 84.63%, and the referral rate was 15.37%; in the NICU ward, the pass and referral rates were 78.00% and 22.00%, respectively. When rescreening, the referral rate was 15.48% in the NICU ward and 8.19% in the maternity ward. The difference in the referral rate was significant between the two groups in initial screening (χ² = 76.53, P < 0.01) as well as rescreening (χ² = 16.96, P < 0.01). This suggests that a different screening strategy may be beneficial, for example, performing TEOAE in the maternity ward, but TEOAE and automated ABR in the NICU ward. The rescreening percentage was 43.29%. George et al. reported a referral rate from initial hearing screening between 0.6% and 16.7%, with an average of approximately 3.89% and dropout of approximately 3.7–65%.⁷ Mukari's studies indicated that the hearing screening coverage, the initial referral rate and the rescreening rate were 84.64%, 11.97% and 56.97%, respectively.⁸ Increasing the rescreening rate is of critical importance to improve cost-effectiveness of the screening strategy.

In patients diagnosed with hearing loss, 31 cases were conductive (45.59%), 16 cases were sensorineural (23.53%) and 21 cases were mixed hearing loss (30.88%). The difference in ABR test threshold proportion was significant in different TEOAE screening results between the two groups (χ ², P < 0.01) (Table 2). Some infants whose 1000 Hz probe tone tympanograms were flat showed high-density materials filling in the middle ear by temporal bone CT test, suggesting a flat tympanogram might be caused by ear effusion (17/17, 100%). The quality of the effusion should be further studied. According to the literature and patient history, it might be unabsorbed amniotic fluid, mesenchymal or inflammatory secretion after infection (because some patients had a history of respiratory tract infection).⁹ The study by Ryan et al. ¹⁰ found that among 76 patients who failed hearing screening, 53.5% were otitis media with effusion, 11% were mixed hearing loss and 35.5% were sensorineural hearing loss, while Ingrid et al.¹¹ reported conductive hearing loss was 20.3%, sensorineural hearing loss was 57.9% in 340 referred patients, and the infants in NICU ward were more likely to have sensorineural hearing loss than healthy newborns.

The occurrence of neonatal hearing impairment in this study was 0.571%. The detection rates of hearing loss were significantly different between maternity wards (1.32‰, 11/8321) and NICU wards (1.60%, 57/3573) (χ ², P < 0.01). A study by Robertson et al.¹² showed that 1.9% of children in the NICU ward had severe or extremely severe hearing loss. We also found that there was a significant difference between the group in maternity ward and the group in NICU ward in the proportion of patients with moderate to severe hearing loss (66.67%, 38/57; 45.45%, 5/11, respectively) (χ ², P < 0.01). It was suggested that the severity of hearing impairment was significantly higher in the high-risk group than the maternity ward group. The prevalence of hearing loss was 12.92% (38/294) in the bilateral referred group and 5.00% (30/600) in the unilateral referred group; the moderate/severe hearing loss was 33.33% (10/30) and 86.84% (66/76), respectively (χ ², P < 0.05).While bilateral referrals were more likely to have more serious degrees of hearing loss than unilateral referrals, it is still important to follow the unilateral referrals. There were some unilateral referrals who were later found to actually have bilateral hearing loss.¹³

This study showed that the factors related to hearing loss included jaundice, neonatal asphyxia, premature birth and low birth weight, amniotic fluid or meconium inhalation, influenza and pneumonia, intrauterine infection, genetic factors, sepsis, hypoxic-ischaemic encephalopathy, macrosomia and hypoglycaemia. Of these factors, those related to infection amounted to 29.82% (amniotic fluid and meconium inhalation pneumonia, 12.28%; flu and pneumonia, 12.28% and intrauterine infection, 5.26%). Jaundice, infection, asphyxia and low birth weight were major causes of newborn hearing loss. Usually, neonatal hearing impairment was the result of multiple factors, such as genetic, embryonic development and prenatal diseases. Asphyxia and lung disease could cause damage to the nerve cells in auditory pathways due to hypoxia. Increased serum bilirubin concentration caused its deposition in the cochlea, affecting normal metabolism and outer hair cells’ function by decreasing the cell numbers. In addition, bilirubin selectively destroyed the brain stem auditory nucleus and damaged the nerve and spiral ganglion cells,¹⁴ which undermined the encoding of auditory information and then damaged the synchronization of the auditory nerve.¹⁵ Many studies reported that newborn weight of less than 1500 g was one of the common factors inducing hearing diseases.¹⁶ Raquel et al. ¹⁷ and other studies have found that the hearing screening referral rate was about 1.7% and the incidence of hearing loss was about 0.5% among low birth weight infants.¹⁷ Foulon et al. ¹⁸ analysed retrospectively the incidence of sensorineural hearing loss caused by cytomegalovirus (CMV) infection in 14021 neonates, the results showing that 0.53% of newborns suffered from congenital CMV infection, of whom 22% developed sensorineural hearing loss. The reported impact of these risk factors on newborn hearing is consistent with our results.

The importance of hearing screening programmes is self-evident. In recent years, this work has made great progress as a result of the efforts of audiologists in China. Health administrative agencies issued a series of regulations and industry standards. Some health administrative departments organized a series of workshops about hearing screening and diagnosis. In addition, the establishment of some regional hearing screening centres helped to form a basic network for transferring consultation in hearing screening. All of these efforts effectively improved the coverage rate of hearing screening, early detection and early control of hearing loss.

Nevertheless there remain relatively low rates of screening, rescreening and follow-up. One of the main reasons for these low rates is the situation in which the infant's hearing is returned to normal after a period of observation, and so parents are no longer willing to re-screen their babies. In addition, many young parents have little knowledge of hearing impairment such as progressive and delayed hearing loss. Some patients possessing mild hearing loss may be mistaken for having normal hearing because patients elicit responses to loud environmental sounds, such as door slamming, car horns, etc. For these reasons, some guardians do not adhere to the protocol for the second screening. Enhancing hearing rescreening rates in these infants is an important goal. A second major factor is regional differences in knowledge of hearing impairment. Hearing screening programme are more successful in Beijing and Shanghai compared with other cities, and even better than the community and the central city. The third issue is the lack of nationwide organization and a managed department specializing in hearing screening, diagnosis and referral consulting and the resulting lack of unified data management and sharing.

CONCLUSION

Neonatal hearing screening is essential for early detection of neonatal hearing loss. Bilateral referrals were more likely to suffer greater degrees of hearing loss than unilateral referrals. Jaundice, infection, asphyxia and low birth weight were major aetiologies of neonatal hearing loss.

References

Joint Committee on Infant Hearing. Year 2007 Position Statement: Principles and Guidelines for Early Hearing Detection and Intervention Programs. Pediatrics 2007;4:898–921

Bamford

, Uus

, Davis

. Screening for hearing loss in childhood: Issues, evidence and current approaches in the UK. J Med Screen 2005;12:119–24

Joseph

, Julie

, Jacqueline

, High – frequency (1000 Hz) tympanometry in normal neonates. J Am Audiol 2003;14:20

Morlet

, Moulin

, Putet

, Hearing disorders screening in neonates at risk. Ann Otolaryngol Chir Cervicofac 2001;1:11–8

Chang

, Vohr

, Norton

, Lekas

. External and middle ear status related to evoked otoacoustic emission in neonates. Arch Otolaryngol Head Neck Surg 1993;119:276–82

American Academy of Pediatrics. Joint Committee On infant Hearing 1994 Position Statement. Pediatrics 1995;1:152–6

George

, Papacharalampous

, Nikolopoulos

, Universal newborn hearing screening, A revolutionary diagnosis of deafness: real benefits and limitations. Eur Arch Otorhinolaryngol 2011;268:1399–406

Mukari

, Tan

, Abdullah

. A pilot project on hospital-based universal newborn hearing screening: Lessons learned. Int J Pediatr Otorhinolaryngol 2006;70:843–51

Margolis

, Bass-Ringdahl

, Hanks

, Tympanometry in newborn infants – 1 Hz norms. J Am Acad Audiol 2003;14:389

10.

Ryan

, Charles

, Patti

. Failed newborn hearing screens as presentation for otitis media with effusion in the newborn population. Int J Pediatr Otorhinolaryngol 2005;69:393–7

11.

Ingrid

, Lambertus

, Marian

, Evaluation of Hearing Loss after Failed Neonatal Hearing Screening. J Pediat 2009;155:646–50

12.

Robertson

, Howarth

, Bork

, Permanent bilateral sensory and neural hearing loss of children after neonatal intensive care because of extreme prematurity: A thirty-year study. Pediatrics 2009;5:e797–807

13.

Chang

, O-Lee

, Price

. Evaluation of unilateral referrals on neonatal hearing screening. J Med Screen 2009;16:17–21

14.

Shapiro

, Nakamura

. Bilirubin and the auditory system. J Perinatol 2001;21(Suppl. 1):S52–5

15.

Shaia

, Shapiro

, Spencer

. The jaundiced gunn rat model of auditory neuropathy/dyssynchrony. Laryngoscope 2005;12:2167–73

16.

Berg

, Spitzer

, Towers

, Newborn hearing screening in the NICU: profile of failed auditory brainstem response/passed otoacoustic emission. Pediatrics 2005;4:933–8

17.

Raquel

, Marisa

, Amélia

. Neonatal Hearing Screening: failures, hearing loss and risk indicators. Braz J Otorhinolaryngol 2011;6:775–83

18.

Foulon

, Anne

, Walter

, Ann

, A 10-year prospective study of sensorineural hearing loss in children with congenital cytomegalovirus infection. J Pediatr 2008;153:84–8