Otoacoustic Emissions as a Screening Test for Hearing Impairment in Children Recovering From Acute Bacterial Meningitis

DISCUSSION

In this study, we attempted to record OAEs from a large, representative sample of British children in the recovery phase of acute bacterial meningitis. Our main findings were that inpatient OAE screening for postmeningitic hearing loss is both feasible and effective. Regarding the feasibility of OAE screening, we found that the technique was well tolerated by the patients and was considerably easier to perform than ABRs. Technically acceptable recordings were obtained from 90% of the children tested. This attainment rate is comparable to those obtained in OAE screening studies in populations of healthy children and neonates.^22,24–26

There have been two previous studies of OAE screening in children with meningitis.^20,,27 These have produced conflicting information about the feasibility of the technique in this population. In the first study, Fortnum et al²⁷ attempted to record OAEs from children shortly before discharge from the hospital. They were able to obtain satisfactory recordings from only 28 of the 53 patients studied. The low attainment rate was attributed to poor compliance, and the authors suggested that children in the recovery phase of bacterial meningitis are too irritable to tolerate the test procedure. More recently, François et al²⁰ reported on their experience of OAE screening in 39 young children who were seen in an audiology clinic several weeks after discharge from the hospital. In this study the authors obtained satisfactory OAE recordings from all but 2 of the children. The findings of these two studies suggested that the timing of the screening test may be critical.²⁰However, the high attainment rate in our study shows that children can tolerate the test procedure during the recovery phase of meningitis. It is therefore likely that the variation in attainment rates is because of other factors. There were considerable differences in equipment and technique between the studies, and it may be that the later studies have been more successful because of advances in OAE technology and methodology.²²

This study was the first to obtain values for the sensitivity and specificity of OAE screening in children recovering from meningitis. In previous studies the results of OAE screening have not always been compared with an accurate, objective measurement of hearing threshold. In our study, OAEs were compared with the gold standard of ABRs. We found that all children with hearing loss failed to produce OAEs and that 91% of those with normal hearing did produce emissions. In other words, OAE screening had a sensitivity of 1.00 and a specificity of 0.91. Such a high sensitivity is clearly desirable in a screening test. This sensitivity value should, however, be interpreted with some caution because only 7 of the 110 children studied had hearing loss (and only 2 of these had sensorineural deafness). Nevertheless, OAEs have proven effective in identifying cases of postmeningitic hearing loss in a number of previous studies,^20,,22,27,28 and there have been no reports of patients with postmeningitic deafness passing an OAE test. Furthermore, OAEs have also been shown to be highly sensitive in screening other populations of children and neonates.^17,,18

One type of sensorineural deafness that could be missed by OAE screening is retrocochlear hearing loss, where the pathologic lesion is in the auditory nerves or higher centers.^18,,29 Emissions may be produced in retrocochlear hearing loss because the generation of OAEs is not dependent on an intact nervous system. In the past it was often thought that such lesions were responsible for postmeningitic deafness. However, it is now clear from pathologic, experimental, and clinical studies that most, if not all, cases of postmeningitic hearing loss are caused by lesions of the cochlea.^21,30–32 It is therefore unlikely that many cases of postmeningitic deafness would be missed by an OAE screening program.

The specificity of OAEs in this study, at 0.91 (95% confidence interval, 0.85 to 0.97), is satisfactory for a screening test of hearing status. The figure is similar to those obtained in studies in which OAEs and ABRs have been compared in neonatal screening programs.^24,,26 Similar figures have also been obtained in studies of healthy older children.^22,,25

In the past, there have been concerns about the specificity of OAE screening for sensorineural hearing loss because patients with conductive hearing loss may also fail the test.¹⁸ It was thought that this would be a particular problem in postmeningitic screening because of a reported association between bacterial meningitis and otitis media with effusion.²⁷ However, in this study, only 5 children failed the OAE screening test because of conductive hearing loss. Moreover, other results from our study suggest that children with acute bacterial meningitis do not have an excess of middle ear effusions.³³ Thus, although children with conductive defects may fail an OAE screening test, the numbers involved would not be great.

As well as calculating the sensitivity and specificity of OAE screening, we also obtained positive and negative predictive values for the test. The positive predictive value, which is the proportion of patients who fail the screening test and actually have hearing loss, was 0.44. In other words, more than half of the children who failed the test had normal hearing. Although this may seem disappointing, it should be noted that the positive predictive value of a screening test is dependent on the prevalence of the condition of interest. In this population, hearing loss was uncommon and the positive predictive value was correspondingly low. It should also be noted that only 9 of 103 children with normal hearing failed the test. Thus, the absolute number of children with normal hearing referred for further investigation would not be great. The negative predictive value of the OAE screening test was highly satisfactory at 1.00 (95% confidence interval, 0.96 to 1.00). This reflects the fact that all 94 children who passed the OAE test had normal hearing. The negative predictive value obtained in this study confirms that children who pass the screening test are highly unlikely to have hearing loss, and do not therefore need urgent referral for further investigation.

One of the major findings of this study was that predischarge hearing tests are possible in children recovering from meningitis. We believed it was important to perform inpatient screening because up to 25% of children who have had meningitis do not receive audiologic follow-up.^5–7 We reasoned that the proportion of children tested could be increased by performing the tests before the children left the hospital. In fact, >90% of the children in this study did have a follow-up hearing test but this may be explained by the fact that we tried to ensure that follow-up tests were performed as part of the study protocol. We therefore continue to believe that more children would be tested by an inpatient screening program.

Inpatient screening also avoids delays in the diagnosis of postmeningitic hearing loss. This is important because late diagnosis of postmeningitic hearing loss may prevent optimal audiologic rehabilitation. Cochlear implantation, for example, can partially restore hearing in postmeningitic deafness, but the procedure may be ineffective if not performed promptly.^10,,11 One British study found that 40% of children referred to the audiology services had to wait more than 3 months for their first hearing test.⁷ In our study half the children waited for more than 2 months. It is clear that inpatient OAE screening could identify those at risk of sensorineural deafness and prevent these excessive delays.

In the past there have been isolated reports of postmeningitic hearing loss developing after a patient has been discharged from the hospital. Cases have even been described in which deafness develops several years after the illness.¹ If hearing loss were to develop after discharge from the hospital, this would undoubtedly limit the usefulness of inpatient screening. Fortunately, the evidence suggests that sensorineural hearing loss rarely develops at this late stage. For instance, there have been no cases of late deafness in a number of prospective studies.^3,,13,14 Other reports have shown that sensorineural deafness usually develops early in the course of the disease.^21,,34 Hearing loss can, in fact, be detected within hours of admission to the hospital. It therefore seems that the development of deafness weeks or months after meningitis is an unlikely event. For these reasons we believe that inpatient screening should not miss many cases of sensorineural hearing loss.

Another phenomenon that could lessen the value of early OAE screening is reversible postmeningitic deafness. During the 1980s several investigators found that up to one-third of patients with bacterial meningitis had a hearing defect that lasted for several weeks or months and then resolved.^3,12–14 The cause of this reversible hearing loss was uncertain. Some authors believed that the condition was sensorineural,^12,,13 whereas others, such as Dodge et al,³ thought that it was conductive. Clearly, if reversible hearing loss is sensorineural, a large number of children would be identified by an inpatient screening program and then referred for urgent audiologic assessment and possible cochlear implantation. Fortunately, most of the evidence suggests that reversible hearing loss is conductive rather than sensorineural.^1,,33,35 In any event, reversible hearing loss is unlikely to be confused with permanent postmeningitic deafness at the time of diagnostic audiologic assessment because it tends to be less severe.¹³ It is therefore unlikely that reversible hearing loss would seriously disrupt an inpatient OAE screening program.