The use of the term “auditory neuropathy” (AN) has probably steered clinicians away from considering cochlear implantation in the past, as a “neuropathy” usually implies abnormal neural function and leads to the assumption that a cochlear implant is unlikely to be successful. However, since 2000 a number of studies have been published indicating that some patients diagnosed with AN did benefit from cochlear implantation.1 In these cases, it was not predicted before the operation whether or not the patients were likely to experience success.
Auditory neuropathy (AN) has been typically characterized by absent or abnormal brainstem responses in the presence of otoacoustic emissions (OAEs) and/or a cochlear microphonic (CM). AN patients have been particularly difficult to manage, and the degree of their loss and level of distortion is unpredictable from auditory brainstem response (ABR) and OAE testing alone. Also, their pure-tone audiometry levels are often inconsistent with their speech-discrimination ability. Our experience at the Sydney Cochlear Implant Centre (SCIC) has shown that significant language delays can result even when hearing aid fittings have shown good detection of sound across the speech range.
Some recent studies have shown that patients identified with AN show abnormal results on tests evaluating temporal processing, therefore suggesting “dys-synchronous” auditory activity.2,3 For an ABR waveform to be seen there must also be synchronous neural activity that is time-locked to the presentation of the stimulus. The term “auditory dys-synchrony” (AD) has therefore been put forward as a more appropriate description to use as it reflects the nature of the problem rather than implying the site of lesion based on incomplete information. If additional neurologic tests confirm the presence of an identifiable neurologic abnormality, then the term “auditory neuropathy” would appear more accurate.
In a typical audiology clinic setting, the main tools available for the differential diagnosis of sensory and/or neural hearing losses include OAEs and ABR testing using acoustic stimuli (acABR). Unfortunately, OAEs provide information only about outer hair cell function and can be easily masked when there is middle ear pathology. It can also be difficult to make a differential diagnosis using acABR testing alone because when the ABR waveforms are absent or abnormal it may be unclear if this resulted from poor/abnormal cochlear stimulation of the auditory nerve or a true abnormality along the nerve and/or brainstem pathway. While performing OAE and acABR testing is an important way of flagging the presence of AN/AD, determining the site of lesion in these cases on the basis of these two tests alone is not possible.
SCIC uses a pre-operative battery of three tests:
- Electrocochleography (ECochG) to test cochlear function,
- ABR testing to acoustic stimuli (acABR), and, if needed,
- ABR testing to electrical stimuli (EABR)
These tests are performed under a general anesthetic and have been chosen to improve the ability to make separate evaluations of cochlear versus auditory nerve and brainstem function.
ECochG testing involves placing a “golf club'” electrode through the eardrum and resting it on the round window niche of the cochlea (Figure 1a). This allows for direct recordings of both cochlear function and the first neural response associated with stimulation of the auditory nerve (Figure 1b). Using this technique also makes it possible to determine if middle ear fluid is present and, if so, to attempt to clear it prior to testing.
At our implant center, ECochG is routinely tested using six stimuli: “click” stimuli and 8000, 4000, 2000, 1000, and 500-Hz toneburst stimuli. Using these stimuli allows for threshold determination across the key frequencies associated with speech detection and evaluation of the morphology of the cochlear and early neural responses across this range.
The most typical ECochG responses we are familiar with are those seen for normal hearing (Figure 2a), severe sensory hearing loss (Figure 2b), and a profound sensory loss (Figure 2c).
Over time, an increasing number of babies/children undergoing ECochG at SCIC presented with an unusually shaped positive waveform coined the Abnormal Positive Potential (APP), (Figure 3). These patients were especially likely to have had a significant history of hypoxia, jaundice, and/or prematurity. They were also more likely than other patients to have presented with the typical features of auditory neuropathy, including the presence of OAEs and/or CM but an absent/abnormal ABR.
More than 1500 pediatric patients at SCIC have been tested with the ECochG test battery, including 305 babies and children between May 2001 and March 2004. Excluding those found to have normal peripheral hearing, complete database entries were available for 264 of these patients. Within this group, 71% were classified as bilateral sensory losses and 4% had a unilateral sensory loss in the mild to profound range. Approximately 20% of the remaining quarter showed an abnormally shaped positive potential in at least one ear at one or more frequencies.
The exact pathology relating to the presence of the APP on electrocochleography is still unknown. However, given the nature of this test, the APP may reflect an abnormality within the cochlea and/or involving the initial excitation of the auditory nerve. To assist with the differential diagnosis of cochlear versus neural involvement in this group, we perform the auditory brainstem response test.
ABR to acoustic stimuli
We perform ABR testing using click stimuli and 1000 and 4000-Hz tonebursts. The click stimuli are used primarily at supra-threshold levels to evaluate the morphology and interpeak latencies of the ABR waveform. The toneburst stimuli are used as both a crosscheck for threshold determination at 1000 and 4000 Hz and to determine the presence/absence of wave V when a significant high-frequency hearing loss has prevented us from seeing a clear response using click stimuli.
Figure 4a shows an acABR waveform consistent with normal auditory nerve and brainstem function. In this case, there was sufficient cochlear function to elicit a synchronous acABR and clear wave I, III, and V peaks can be seen. When the ECochG thresholds are elevated (and middle ear pathology is excluded) but a normal acABR waveform is seen, the hearing loss is classified as “sensory”/cochlear in nature.
Figure 4b shows an acABR waveform consistent with the presentation of AN. In many cases, only a single cycle cochlear microphonic (CM) is seen at the beginning of the trace, which reverses in polarity when changing from a rarefaction to a condensation click stimulus. In this case, the CM appears to “ring” throughout the trace. When the reversing polarity waveforms are added, the response cancels out, which is equivalent to what would be seen if an alternating click stimulus had been used.
It is impossible to know if this result is due to an abnormality within the cochlea preventing normal/synchronous activation of the auditory nerve, or whether the auditory nerve and/or brainstem is responding abnormally to the stimuli used. In this situation, the ABR is tested again using an electrical stimulus to help with the differential diagnosis between these two possible sites of lesion. This is discussed in more detail in the next section.
Figure 4c shows an acABR response in which only a broad positive peak can be seen around the region of wave I. On first inspection, it would be tempting to conclude that this was consistent with brainstem pathology, as none of the later waveforms can be seen. However, this result might also appear if insufficient/abnormal cochlear excitation of the auditory nerve prevented the synchronous activity required to see a clear ABR waveform. Repeating this test using electrical stimulation is helpful in differentiating cochlear from retrocochlear pathology.
Electrical auditory brainstem response (EABR)
EABR testing is performed using the same “golf club” electrode that was positioned in the round window niche of the cochlea for ECochG. Electrical stimuli are transferred through the electrode to stimulate any active auditory nerve fibers in an attempt to elicit an ABR, similarly to how a cochlear implant bypasses the cochlear hair cells to stimulate the nerve for hearing (see Figure 5). Figure 6a shows a normal EABR when an adequate current level is used. Wave I is obscured by electrical artifact, but waves II, III, and V show the same morphology as an ABR elicited by acoustic stimuli.
Of the 264 pediatric patients evaluated between May 2001 and March 2004 the combination of ECochG, acABR, and EABR results showed the following diagnostic population breakdown (summarized in Table 1).
Sensory hearing loss (SHL)
One hundred ninety-two (71%) of the children demonstrated a mild to profound loss bilaterally and 11 (4%) unilaterally on ECochG. These children also showed a normal ABR waveform when either an acoustic (in the case of mild to moderately severe losses) or an electrical stimulus (in the case of severe to profound losses) was used. This combination of results suggested that the main site of lesion was within the cochlea and was defined as a “Sensory” hearing loss.
APP group—auditory dys-synchrony (AD)
Thirty-six (14%) of the children presented with an APP on ECochG bilaterally and 10 (4%) unilaterally. These cases also showed an abnormal or absent acABR, but normal waveforms were seen when an electrical stimulus was used (EABR). This indicates that the auditory nerve and brainstem were capable of synchronous neural activity.
This combination of results also suggests that the main site of lesion was within the cochlea and/or the initial synapse with the auditory nerve. This was defined as “auditory dys-synchrony” (AD), as the absent/abnormal acABR demonstrated that the impairment to the auditory pathway was sufficient to prevent synchronous neural firing in the auditory nerve and brainstem when sound was used. However, this could be overcome when an electrical stimulus was used.
Auditory neuropathy (AN)
Two children (1%) showed abnormal acoustic and/or electrical ABRs bilaterally and 12 children (4%) unilaterally. In these cases, the waveforms were either absent (Figure 6, second from left) or showed a significant delay in the wave I-III interval. These children were classified as having a true auditory neuropathy, as the neural pathways continued to show an abnormal response even when electrical stimulation was used.
Brainstem auditory neuropathy
One child (1%) bilaterally showed the presence of earlier waves, but wave V was either absent, significantly reduced in amplitude (Figure 6, second from right), or significantly delayed (Figure 6, right). This was classified as “brainstem auditory neuropathy” (BAN), as the abnormality was seen only for the later waves that should have been generated by the brainstem.
OUTCOMES WITH COCHLEAR IMPLANTATION
EABR testing at SCIC has effectively been used as a screening test for cochlear implantation in conjunction with functional observations and MRI and CT-scan results. Gibson et al.1 investigated the outcomes of 128 age-matched pediatric patients who had both pre-operative and intra-operative testing for cochlear implantation.4
All patients had been implanted with the Nucleus cochlear implant and only those with a full insertion of 22 electrodes were included. A full insertion was confirmed by performing a Common Ground artifact test during surgery. When all 22 electrodes have been successfully inserted an artifact can be seen at the beginning of each trace as each electrode is stimulated separately. The waveform reverses in phase at around electrode 9 as the array takes the first turn around the cochlea. A Stenvers view x-ray is also evaluated post-operatively to confirm the correct and full insertion of the cochlear implant array.
The persistence of an abnormal neural and/or brainstem response is also confirmed intra-operatively by using implant evoked electrical ABR testing during surgery (ImpEABR). Evaluations are made by using Bipolar +2 stimulation to check 19 sections of the neural response along the cochlear implant electrode array. In normal cases, all sections show a normal wave V response and latency (Figure 8, left). Figure 8, second from left, shows no clear response along the mid-section of the array and Figure 8, second from right, shows no identifiable waveforms across the entire array, consistent with a true auditory neuropathy of varying degrees. Figure 8, right, shows that wave II is present, but waves III to V are essentially absent, consistent with a brainstem neuropathy.
When the Melbourne Speech Perception categories (Table 2) were used, the post-operative outcome scores were significantly higher for those children diagnosed with a sensory hearing loss and auditory dys-synchrony compared to those diagnosed with a true auditory neuropathy or brainstem auditory neuropathy (two-tailed p value<0.005, Mann Whitney U Test). It is particularly interesting that in the first year after the switch-on the children with AD also performed significantly better than those diagnosed with a typical sensory loss (Table 3).
APP and otoacoustic emissions
Within the same cohort, 39 ears had measurable OAEs recorded previously during evaluations made in the neonatal period. All these ears demonstrated an APP on ECochG and showed large cochlear microphonics with an absent/abnormal acABR. Thirty-two subsequently showed a normal EABR waveform (i.e., 82% were consistent with auditory dys-synchrony) and seven had an abnormal EABR waveform (i.e., 18% were consistent with AN and/or BAN).
The consistent findings across all the diagnostic groups for pre-operative, intra-operative, and post-cochlear implant outcomes have proven to be an invaluable sequence of tools to assist clinicians in providing parents with information about predicted outcomes before they choose whether or not to proceed with a cochlear implant for an individual child.
For this reason, the Sydney Cochlear Implant Centre routinely uses information from its pre-operative test battery to counsel families about predicted post-implant outcomes. By using this combination of diagnostic tests in conjunction with functional evaluations and CT and MRI scans it has been possible to place patients into four categories with general predictions of outcomes with cochlear (CI) implantation:
- Sensory hearing loss (SHL): Primarily cochlear pathology. Good outcomes predicted with CI.
- Auditory dys-synchrony (AD): Believed to be primarily cochlear pathology. Good outcomes predicted with CI.
- Auditory neuropathy (AN): Consistent with auditory nerve pathology. Less optimal outcomes predicted with CI.
- Brainstem auditory neuropathy (BAN): Consistent with auditory brainstem pathology. Less optimal outcomes predicted with CI.
The decision to proceed with surgery is always a significant one for families. But knowing what can be realistically expected beforehand helps them to weigh the risks versus the benefits and ensure that the most appropriate type of intervention is in place for optimal language development.
1. Michalewski HJ, Starr A, Nguyen TT, et al.: Auditory temporal processes in normal-hearing individuals and in patients with auditory neuropathy. Clin Neurophysiol
2. Trautwein PG, Sininger YS, Nelson R: Cochlear implantation of auditory neuropathy. JAAA
3. Zeng FG, Oba S, Garde S, et al.: Temporal and speech processing deficits in auditory neuropathy. Neuro-report
4. Gibson WPR, Sanli H, Gardner-Berry K, Haddon A: Auditory neuropathy: Electrophysiological investigations and outcomes after cochlear implantation (in press).