Auditory neuropathy spectrum disorder (ANSD) and cochlear implantation

Abstract

We discuss issues related to cochlear implantation in children with auditory neuropathy spectrum disorder (ANSD). We describe the varied nature of this disease category including the numerous potential causes of auditory neuropathy. The most prevalent etiology for infants with ANSD is associated with prolonged neonatal intensive care unit (NICU) stay. We discuss the potential contribution of cochlear hypoxia to this etiology. The second part of this review describes in detail our own experience at the Hospital for Sick Children in Toronto, with cochlear implantation of children diagnosed with ANSD. We outline the detection, diagnosis, and referral routes for our patients. We provide an overview of our “standard operation procedures” regarding candidacy, and discuss some of the special considerations that need to be applied to children with ANSD. This includes decisions to implant children with better audiometric thresholds that are standard in non-ANSD patients, concerns about the possibility of spontaneous remission and the appropriate timing of implantation. Finally we review an extensive published literature in outcomes after cochlear implantation (CI) in ANSD. This is not a systematic review but rather an exercise to distill out some important reoccurring themes and the general consensus of opinion to date. Our conclusion is that the hearing loss category ANSD, together with its numerous co-morbidities, is far too heterogeneous to make definitive statements about prognosis with CI.

Introduction

Although auditory neuropathy was formally defined around 20 years ago [1], [2], [3], [4] this form of hearing loss continues to cause controversy and challenges in decision making over appropriate means of habilitation. Pediatric CI programs face special challenges with children diagnosed with ANSD. This category of hearing loss is a “mixed bag”. Some children appear to have “isolated auditory neuropathy” i.e. without other confounding issues. Others, perhaps most, have co-morbidities and/or anatomical abnormalities that increase the variance in the group. There are issues about candidacy for CI, and the prognosis for individual children; these are the concerns discussed in this paper.

Auditory neuropathy is a disorder in which behavioral threshold measures do not match with other measures of auditory function such as ABR data and speech understanding scores. In some cases there may be the presence of otoacoustic emissions or cochlear microphonics (CM) that do not correlate with ABR or audiometric thresholds [1], [2], [3] (for a comprehensive review see: [4]). Diagnostic testing in these initial reports pointed to a disorder of the cochlear nerve as opposed to sensory loss of cochlear origin. Thus the term auditory neuropathy appears to be appropriate, although there has subsequently been much blurring of the boundaries regarding the sites of lesion.

Over time it became clear that auditory neuropathy was manifest in different ways and degrees, and as the result of numerous causes. This is one reason why other terms such as auditory dys-synchrony and peri-synaptic audiopathy were used to describe the disorder. The recognition that there was a wide range of hearing loss types and etiologies led to the adoption of the term auditory neuropathy spectrum disorder (ANSD) [5]. The fact that ANSD is a heterogeneous disease category is a problem for treatment of patients, not least in relation to cochlear implantation. Any “ANSD” population is by no means homogeneous, and to study such a group may not clearly lead to conclusions that can be simply applied to others.

Regarding etiology, there are many. In adults it has been associated with hereditary disorders including Charcot-Marie-Tooth disease [3], Friedreich's ataxia [6], and various other hereditary motor and sensory neuropathies [4]. Also on the genetic front, mutation of the otoferlin and pejvakin genes have been implicated as a causal agent [7], [8], [9]. Other etiologies implicated are kernicterus [10], oncologic drug treatment with cytotoxic agents such as cisplatin [11] and congenital cochlear nerve hypoplasia [12]. It was well recognized from the earliest descriptions of ANSD that it was significantly more prevalent in “high risk” infants. These babies are variously compromised in early development in utero or after premature birth in NICU, or during a prolonged or difficult birth. Many factors associated with premature birth have been implicated including hyperbilirubinemia [13], [14], [15]. It remains difficult to determine which of these many risk factors, alone or in combination, is responsible for the severity of ANSD.

One potential denominator common to many etiologies associated with ANSD is cochlear hypoxia. Harrison [16] initially described an animal model of auditory neuropathy in which chronic cochlear hypoxia resulted in subjects having various features analogous to patients with ANSD. The model also revealed an underlying pathology of inner haircell degeneration and synaptic connections to spiral ganglion cells. On the basis of this model Harrison proposed that a major etiology of ANSD is chronic cochlear hypoxia [16], [17]. Supporting this theory, human postmortem studies of cochleae in NICU babies that did not survive also revealed the pattern of inner hair cell loss that was found in the experimental hypoxia studies [18]. Whilst cochlear hypoxia is a valid idea, others believe that the assumption of hypoxia as the primary cause may be too simplistic. In this regard, whilst the majority of our non-genetic ANSD patients are NICU graduates (typically premature births) they all have multiple risk factors. Thus in addition to just hypoxia, they may have had low birth weight, hyperbilirubinemia, prolonged ventilation, septicemia, aminoglycoside treatment, noise exposure, etc. It has also been noted that other babies have had more prolonged hypoxia (e.g. in congenital cardiac anomalies) or more acute hypoxia (patients requiring CPR) than those who get ANSD. Susceptibility to hypoxia could of course be dependent upon gestational age and presence of other risk factors at the time of exposure.

Whatever the combination of causal factors may be, it is reasonable to assume that many cases of ANSD have a common underlying pathology that is an extensive but sub-total damage to inner haircells and their synaptic connections with auditory neurons. The functional result of this damage could be a significant reduction of information channel capacity at the cochlear nerve level, and beyond. This condition helps to explain some of the symptoms of the disorder, in particular a lack of correlation between behavioral thresholds and both ABR characteristics [4] and complex sound processing [19]. Thus hearing thresholds can be “normal” with relatively few channels, but large numbers of active neurons are required for generation of ABR signals. Similarly complex sound analysis is best achieved if information is carried in large numbers of channels. In the following sections we will discuss the issues of cochlear implantation in children with ANSD.

One important factor regarding the efficacy of CI in ANSD is the presence or not of comorbidity. Those children with ANSD who have no other conditions have been termed “isolated” auditory neuropathy by some authors [20]. Others may have additional medical and anatomical deficits. NICU graduates are at increased risk of cognitive or motor impairment from brain injury. Hypoxic ischemic encephalopathy causing spastic cerebral palsy and kernicterus with athetoid cerebral palsy are relevant examples of conditions that can be associated with ANSD and cause speech and language delay from cognitive and oromotor deficits independently from hearing impairment. These co-variables can significantly complicate the assessment of candidacy and outcome of cochlear implantation.

Another important category is cochlear nerve deficiency that may take the form of uni- or bilateral hypoplasia or aplasia. Patients with cochlear aplasia or hypoplasia can present as ANSD because the ABR is absent or abnormal, and CM or OAEs may be present. It is important to note that such cases can be missed in neonatal screening using OAEs, and unilateral cases may not present until many years later. Many cochlear implant centers include such patients in their ANSD category, and in their published outcomes (see below) distinguish performance of children with isolated ANSD and those with cochlear hypoplasia (or associated cochlear nerve deficiency). In our program, we separate out children with aplasic/hypoplasic nerves from the ANSD category as it is a clearly defined diagnosis with more predictable prognosis. We have previously discussed and reviewed our experience with this diagnostic entity [21], [22].

At the Hospital for Sick Children in Toronto, many candidates for cochlear implantation have congenital hearing loss detected though the Ontario universal newborn hearing screening program [23], [24], [25]. Patient data collection is approved under SickKids REB protocol #1-2954. Most of our cases of ANSD arise from the screening program, as there is additional targeted ABR screening of high-risk infants (typically those with an NICU stay of over 3 days).

In subsequent diagnostic assessment there is an evaluation of ABR waveforms, CM and OAEs and babies found to have inconsistent ABR, OAE and/or CM data are categorized as possible ANSD. Following identification as ANSD, children are referred to auditory verbal therapy, and will have regular audiological assessments. If audiometric and/or speech reception thresholds allow unaided hearing then we do not initially provide aids. If the behavioral thresholds are poor we recommend aids; we believe that it is important to provide sound activation. Caution is taken to avoid high levels of sound stimulation. We follow children closely to assess language development as well as possible spontaneous threshold improvements. If there are signs of abnormally delayed speech and language development, candidacy for cochlear implantation is then considered. Older children will be investigated for ANSD if their behavioral testing shows a poor correlation between threshold sensitivity and speech recognition.

It is very important to recognize that there is very little homogeneity in the ANSD group. That is the reason for adding the term “spectrum disorder” to this disease category. In studies that attempt to compare characteristics of ANSD with other types of sensorineural hearing loss, we must remember that the ANSD group is a very mixed bag. A second issue with study of ANSD is that it presents (at least by current definitions) in only a small proportion of children with sensorineural hearing loss, and in only a very small proportion of children who are provided with cochlear implants. Only large pediatric cochlear implant programs are likely to accumulate a useful sample size. Over the past 12-year period (since 2003) over 75 children with suspected ANSD have been assessed in our pediatric otolaryngology department. Of these, 37 patients entered our CI program for assessment. Note that we exclude here cases of cochlear nerve aplasia/hypoplasia; our ANSD occurrence is lower than in programs that include children with nerve abnormalities in their ANSD category. Of the 37 ANSD children assessed 17 were implanted.