 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Clinical Department of Hearing, Voice and Speech Disorders, Innsbruck Medical University, Innsbruck, Austria
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS. Data were drawn retrospectively from the clinical charts of children who had bilateral PCHI (>40 dB hearing level, better ear, unaided) and had undergone universal newborn hearing screening (UNHS) between 1995 and 2000 in various Austrian hospitals. A hearing loss was recognized as postnatal when a child passed UNHS but was later found to have a hearing impairment. The presence of risk indicators, as suggested by the Year 2000 Statement of the American Joint Committee on Infant Hearing (JCIH), was assessed by reviewing the children's clinical charts.
RESULTS. Of a total of 105 children with bilateral PCHI, 23 (22%) showed postnatal impairment. After correction of this number for underascertainment, postnatal impairment was estimated to account for 25% of all bilateral PCHI at age 9 years. Risk indicators were found in 17 children but did not fully correspond to those proposed by the JCIH. The risk factors found were a family history of hearing loss (3 children), meningitis (2), craniofacial malformation (2), persistent pulmonary hypertension (1), congenital cytomegaly infection (1), extracorporeal membrane oxygenation (1), recurrent otitis media with effusion (1), and, in addition to the JCIH list, ototoxic therapy (5), and birth before 33rd gestational week (2) (1 child had a combination of the last 2). Six children showed no risk indicators for the postnatal hearing loss.
CONCLUSIONS. Our findings suggest that 
25% of bilateral childhood hearing loss is postnatal, which supports the leading role of UNHS in detecting PCHI. Provisions for also identifying postnatal cases nevertheless are justified. Because in some of these children no risk indicators are detectable and in others the hearing deterioration starts after age 3 years, audiologic monitoring of at-risk children up to this age may not be sufficient. Additional methods, such as hearing screening at nursery schools or schools, are recommended.
Key Words: early hearing detection and intervention • permanent childhood hearing impairment • postnatal hearing impairment • risk indicators • universal newborn hearing screening
Abbreviations: PCHI—permanent childhood hearing impairment • UNHS—universal newborn hearing screening • HL—hearing level • TEOAE—transient evoked otoacoustic emissions • ENT—ears—nose—and throat • JCIH—Joint Committee on Infant Hearing
Early intervention for permanent childhood hearing impairment (PCHI) has been shown to reduce the deleterious effects of impaired audition on language and cognitive and social skills of affected children.1–4 For identification of hearing impairment in newborns, universal neonatal hearing screening (UNHS) has been implemented in many parts of the Western world, but not all PCHI is detectable in infancy. Three major categories of hearing loss become manifest postnatally: delayed onset, progressive, and acquired.* Delayed-onset hearing loss derives from certain adverse medical conditions that are present before or around birth (eg, intrauterine infections, severe asphyxia) and whose damaging impact to the inner ear evolves over time. Thus, the child's hearing still is normal in the perinatal period but starts to deteriorate some time afterward. Progressive hearing loss is either hereditary or linked to syndromes and neurodegenerative disorders. Presenting with normal audiologic responses at birth, these children later show a great variety regarding onset, progression rate, frequencies involved, and ultimate severity of their hearing impairment. Finally, acquired hearing loss results from a number of external factors that directly or indirectly harm the inner ear, such as meningitis, ototoxic drugs, and acoustic trauma.
For prompt detection also of postnatal PCHI, the continued surveillance of infants and toddlers who are at risk for a postnatal hearing loss has been recommended, and a list of risk indicators have been proposed.5 This recommendation raises the following questions: How many children remain to be detected in the postnatal years? Is surveillance of at-risk children sufficient to detect at least the majority of them? Evidence on the prevalence of postnatal PCHI is conflicting (Table 1). A study from the United Kingdom found the prevalence of all hearing loss >40 dB hearing level (HL) to increase from 1.06/1000 at birth to up to 1.65/1000 (or even 2.05/1000; adjusted value) at age 9 years.6 This would mean that 35% to 50% of all hearing loss in 9-year-old children is postnatal. A study from the United States estimated the prevalence of congenital PCHI >40 dB HL at 0.53/1000 and the prevalence of post(neo)natal PCHI >40 dB HL at 0.25/1000.7 If the figures are correct, then 30% of all PCHI would be postnatal. In contrast, 2 epidemiologic studies in English and Scottish birth cohorts reported that postnatal cases accounted for only 16% and 11%, respectively, of all PCHI >40 dB HL.8,9 The 4 studies differ a bit in their observation periods, which, however, hardly explains the discrepancy between their estimates. Instead, differences in their ascertainment methods and, more relevant, uncertain criteria for determining whether a hearing loss is congenital or postnatal are the likely reasons for the divergences. A postnatal hearing loss is commonly defined by onset after the perinatal period. The time of onset, however, frequently is unknown, and its determination relies on assumption. To circumvent this problem, we applied a different definition in our study: a hearing loss is postnatal when the child passed UNHS but was later found to have a hearing impairment. Clearly, this definition is not free of problems (it assumes a 100% sensitivity of the screen, and it cannot be applied to unscreened children), but it allows a clear-cut decision to be made for every screened child concerning the category into which the hearing loss falls.
|
90% of all Austrian neonates. The yield from the screening since its inception in 1995 is judged to provide a good basis for analyzing the number of children with a postnatal PCHI. |
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Study Population and Inclusion Criteria
Although the study population comprised all children who had PCHI and were born in Austria since 1990, for the purpose of this particular analysis, only data from children who met the following criteria were used: (1) bilateral sensorineural, conductive, or combined hearing impairment; (2) at least a moderate degree of hearing loss (>40 dB better ear HL, unaided); (3) born between 1995 and 2000; and (4) having undergone UNHS.
The end of the time frame (December 2000) was chosen to warrant at least a 4-year period after birth (data analysis started in Spring 2005), in which a postnatal hearing impairment can become manifest and be confirmed. The criteria of a bilateral and an at least moderate hearing loss were set in accordance with the literature6–9 for reasons of comparability. Hearing impairment was measured in older children by pure tone audiometry and averaged over frequencies 0.5, 1, 2, and 4 kHz. In young children, results from the automated auditory brainstem response measurement were used to estimate the degree of the hearing impairment. Degrees of hearing impairment were as follows: moderate, 41 to 70 dB HL; severe, 71 to 95 dB HL; profound, >95 dB HL. In children who met the definition of "postnatal hearing loss," the clinical charts were reviewed for occurrences that were considered to indicate an increased risk for postnatal onset of hearing impairment (see below). For this purpose, competent personnel at the submitting institutions were contacted and asked for assistance.
Ascertainment of Hearing Impairment
As mentioned above, all neonates were screened at various Austrian hospitals (maternity wards or ICUs) using TEOAE. To pass the screen, the child had to have positive TEOAE in at least 1 ear at either the first test or the retest. In this way, a bilateral congenital hearing impairment was excluded. Hearing assessment of the children who failed the screen as well as of those whose hearing impairment became manifest after the neonatal period occurred at audiologic units of various Austrian ENT departments. At these institutions, TEOAE and auditory brainstem responses (in older children: pure tone audiograms) were assessed to confirm the hearing loss.
Outcome Variables
Before data analysis, the data were checked for duplicates (recognized by the identicalness of dates, eg, date of birth and of hearing loss diagnosis) and for inconsistencies. When such were found, the submitting institutions were asked for clarification.
The main outcome variables were (1) the percentage of children who had hearing impairment with a postnatal onset of hearing loss and (2) the percentage thereof with at least 1 risk indicator for a postnatal hearing loss. Risk indicators were drawn from the Year 2000 Statement of the Joint Committee on Infant Hearing (JCIH)5:
When the clinical records of the children were searched for these indicators, 2 problems arose. First, "parental or caregiver's concern regarding hearing etc." is a very unspecific factor that was practically always present (otherwise, the child would not have undergone hearing assessment). For this reason we decided to disregard it. Instead, we put the focus on the question, "Who was the person whose concern prompted the hearing assessment?" Second, some children exhibited occurrences that are not included in the JCIH list, although they are traditionally regarded as risk factors for hearing loss (eg, ototoxic medication). Because these occurrences could have been etiologically linked to the hearing loss in some cases, we registered them, too. To indicate that they do not belong to the JCIH risk indicator list, we put them in brackets in our results table (Table 2).
|
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
550 notifications delivered by the participating institutions, a subset of 105 children with hearing impairment met the above criteria for inclusion in this analysis (overall results from the whole sample are reported elsewhere10,11). The hearing loss in these 105 children was bilateral sensorineural in 97, bilateral conductive in 3, and bilateral combined in 2. In 3 additional children, it was sensorineural in 1 ear and combined in the other. Of the 105 children, 23 had passed newborn hearing screening, indicating that 22% of the hearing loss in this sample was postnatal (95% confidence interval: 14–30%). Table 2 lists some details of these children. Six of them were female, and 17 were male. The hearing impairment was moderate in 15, severe in 4, and profound in 4. Except for 2 children whose hearing loss was combined in 1 ear and sensorineural in the other, all children had bilateral sensorineural hearing loss. Confirmation up to age 3 years occurred in 15 (65%) children, indicating that the majority of hearing loss became manifest before this age.
Risk indicators for postnatal hearing loss, as proposed by the JCIH,5 were identified in 11 (48%) children: a family history of PCHI (3), meningitis (2), craniofacial malformation (2), congenital cytomegaly infection (1), extracorporeal membrane oxygenation (1), persistent pulmonary hypertension (1), and recurrent otitis media with effusion (1). Considering additional occurrences that are not included in the JCIH list but are regarded as risk factors for a hearing loss, the number of children with risk indicators increases by 6 to 17 (74%). The occurrences were ototoxic medication and premature birth (ie, birth before 33rd gestational week). One child was born premature; 3 children were treated with ototoxic drugs for perinatal sepsis, 1 of whom was also born premature. Two additional children received ototoxic medication, 1 for bacterial infection and 1 for cancer, in the postnatal period. The remaining 6 (26%) of the 23 children showed no risk indicators.
Parental concern about the child's auditory reactions or speech development prompted the hearing assessment in 13 (57%) of the 23 children. In 2 (9%) children, the hearing loss was suspected by a kindergarten teacher, and in 8 (35%) children, it was suspected by a doctor.
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
25%. With this estimate, our study backs the claim that the bulk of childhood hearing loss is congenital and that UNHS plays the leading role in detecting PCHI. Detection of risk indicators for a postnatal hearing loss was successful in 11 (48%) of 23 children with PCHI for indicators that were drawn from the list proposed by the JCIH.5 When additional risk indicators were considered, namely ototoxic medication and premature birth, this number rose to 17 (74%). The latter indicators, however, were added posthoc (after reviewing the children's charts), and their validity is doubtful. At any rate, audiologic monitoring of at-risk children, if practiced with our sample, would have missed at least one quarter of children with hearing impairment as a result of the nonpresence of risk indicators. Still worse, in some of the children, the hearing loss probably manifested after the age of 3 years. If audiologic monitoring had been performed only up to this age, as is recommended,5 then an additional portion of the children would have been missed. We are unable to estimate how large this portion is, because in some children, the hearing loss, despite being confirmed at a later age, could have been present already at age 3 years. Conversely, in our sample, older children were a little underrepresented (as a result of unequal duration of the follow-up of the birth cohorts). Hence, the sample is likely to miss mainly individuals with onset of hearing loss at preschool or school age. Therefore, the number of PCHI cases with onset after age 3 years, as found in our study (approximately one third of the sample), is estimated to provide the lower bound of this portion.
Our findings indicate that continued surveillance of children who are at risk for hearing impairment up to age 3 years, if performed with our sample, would not have been very effective in detecting postnatal hearing loss. Is it an inadequate means for this purpose? We do not think so. We still deem it to be a valuable means of identifying postnatal hearing loss. However, some changes might be useful. For instance, the time frame "up to age 3 years" could be extended, as a considerable number of children will experience the onset of their hearing loss after this age. Findings from the study from the United Kingdom6 even suggest that the prevalence of hearing loss steadily increases up to age 8 or 10 years. Moreover, the list of risk indicators needs validation and completion. Ototoxic drugs are currently not regarded as a risk factor, because, as is argued,5 only a small percentage of children develop a hearing loss after their administration, yet in our sample, 5 children had received ototoxic substances, and in 3 of them, the diagnosis of hearing loss closely followed the treatment.
We are aware that any extension of the list of risk indicators will lead to increased expenses and costs of the surveillance program, with perhaps only a minimal increase in success. The definition of eligible risk indicators therefore remains a challenge for any surveillance program, but even if the list is based on a loose definition, we have to reckon that there will be children with no (obvious) risk indicators. In 26% of our sample, we could not detect any factor that traditionally is linked to acquired, progressive, or delayed-onset hearing loss. It is possible that we just failed to glean them from their medical histories (charts might have been incomplete). This potential flaw, however, is not a limitation only of our study; it is a problem faced by every program of monitoring at-risk children. Risk factors, although present, may not be recognized correctly by parents or professionals. Considering these problems, that risk indicators may actually not be present in some children and not be recognized in others, additional efforts for detecting postnatal hearing loss seem justified. For instance, hearing screening at schools or nursery schools might be a practicable means in addition to monitoring at-risk children. Hearing screening programs at nursery schools or elementary schools have been found repeatedly to be feasible and effective.13,14 In our view, the combination of a risk indicator surveillance program up to age 3 or 4 years and a hearing screening program in children who enter school would be ideal for promptly detecting postnatal hearing loss.
One of the most delicate prerequisites for the definition of "postnatal" that we applied in our study is a 100% sensitivity of the screen. We are unable to ascertain whether our screening program meets this criterion. The sensitivity of a 2-stage screen comparable to ours has been estimated repeatedly to be close to 100%.13–15 However, it is known that the initial phases of program implementation are accompanied by elevated error rates for screening performance.16 When we checked this for our program, we could not find that the portion of postnatal hearing loss was higher in the early periods than in later periods. Therefore, we are confident that the errors that undoubtedly occurred in the beginning have little impact on the results presented here.
|
FOOTNOTES |
|---|
Address correspondence to Viktor Weichbold, MD, MPhil, Clinical Department of Hearing, Voice and Speech Disorders, Anichstrasse 35, A-6020 Innsbruck, Austria. E-mail: viktor.weichbold{at}uklibk.ac.at
The authors have indicated they have no financial relationships relevant to this article to disclose.
* These terms, including "late-onset hearing loss," are not well defined. They seem to represent a classification scheme; however, they are based on different classification criteria, namely onset, course, and cause of hearing loss. Hence, they allow for overlap: although delayed-onset hearing loss is typically progressive, progressive hearing loss may appear at the same time as delayed-onset hearing loss. 
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2006 American Academy of Pediatrics. All rights reserved. |
||
|
||
