Performance of Cochlear Implant Recipients With GJB2-Related Deafness

Abstract

Congenital profound hearing loss affects 0.05–0.1% of children and has many causes, some of which are associated with cognitive delay. For prelingually-deafened cochlear implant recipients, the etiology of deafness is usually unknown. Mutations in GJB2 have been established as the most common cause of heritable deafness in the United States. In this report, we identify cochlear implant recipients with GJB2-related deafness and examine the performance of these individuals. Cochlear implant recipients received a battery of perceptive, cognitive, and reading tests. Neither subjects nor examiners knew the etiology of deafness in these individuals. The implant recipients were then examined for mutations in GJB2 using an allele-specific polymerase chain reaction assay, single-strand conformation polymorphism analysis, and direct sequencing. GJB2 mutations were the leading cause of congenital deafness among the cochlear implant recipients screened. Cochlear implant recipients with GJB2-related deafness read within one standard deviation of hearing controls better than other congenitally deaf cochlear implant recipients and non-cochlear implant recipients. Individuals with congenital deafness should be offered GJB2 screening. Positive results establish an etiologic diagnosis and provide prognostic, genetic, and therapeutic information. Effective rehabilitation for profoundly deaf individuals with GJB2-related deafness is possible through cochlear implantation.

INTRODUCTION

Despite outstanding achievements by several deaf individuals, a good outcome is not assured for children with congenital hearing loss. A wide array of prospects may be reasonably expected based on the multiple different etiologies of deafness. Deafness from congenital cytomegalovirus, for example, is associated with particularly poor reading and mathematics performance. On the Stanford Achievement Test, 44% of individuals with congenital cytomegalovirus infection scored in the lowest quartile of deaf individuals [Schildroth, 1994]. Many forms of syndromic deafness are also linked with cognitive delay [Gorlin et al., 1995]. While it is easy to presume that the etiology of non-syndromic deafness may affect rehabilitation potential after cochlear implantation, this area remains largely unexplored.

Recent advances in the genetics of congenital severe-to-profound hearing loss improve our ability to specify the etiology of deafness for many children. Despite the heterogeneity of genetic deafness, mutations in the gene encoding connexin 26 (GJB2) are responsible for half of severe-to-profound autosomal-recessive nonsyn-dromic deafness in Caucasian families originating from the United States, France, Britain, Tunisia, and New Zealand, suggesting that mutations in GJB2 are the most common cause of congenital hereditary hearing loss in these countries [Van Camp et al., 1997; Zelante et al., 1997; Kelley et al., 1998; Green et al., 1999]. Unlike several forms of congenital deafness, GJB2-related deafness has no known comorbidity. Individuals walk at a normal age and appear to have normal vestibular function. Although connexin 26 is present in the brain, as well as the cochlea, no neurologic problems are associated with this type of deafness [Denoyelle et al., 1997].

We have examined cochlear implant recipients, identifying children with GJB2-related deafness, to assess the cognitive abilities and performance of congenitally deaf children in relation to the etiology of deafness. Results suggest that genetic screening for GJB2 mutations carries diagnostic, prognostic, and counseling importance.

MATERIALS AND METHODS

Subject Accrual

The accrual source consisted of 92 prelingually deaf children who underwent cochlear implantation at the University of Iowa. All recipients were younger than 19 years. In 47 recipients, the etiology of deafness was unknown; of this subgroup, 20 individuals participated in this study. The University of Iowa Human Subjects Committee approved all procedures.

Cochlear Implant Recipients

Cochlear implant recipients received an array of medical, radiologic, cognitive, and audiologic testing. Intelligence and behavior assessments were completed to identify characteristics of children that might provide insight into successful habilitation with a cochlear implant. All children had intelligence and psychosocial behavior profiles within the range of normal. Audiologic assessments indicated that all children were profoundly deaf and unable to benefit from conventional amplification. Speech perception skills were measured a minimum of three years post-operatively using the Word Intelligibility by Picture Identification test (WIPI); the Vowel Perception test (VP); the Phonetically-Balanced Kindergarten Word list (PBK-wd) and phoneme score (PBK-ph); the Monosyllable, Trochee, Spondee test (MTS); and the Four-choice Spondee Test (FST) [Fryauf-Bertschy et al., 1997]. Reading performance was measured using the paragraph comprehension subtest of the revised Woodcock Reading Mastery test and compared to children with normal hearing [Woodcock, 1997].

GJB2 Screening

DNA was extracted from 5 ml of whole blood obtained by peripheral venipuncture from each participant. DNA samples were screened for mutations in GJB2 using a combination of allele-specific polymerase chain reaction (ASPCR) and single strand conformational polymorphism analyses [Green et al., 1999]. The most common mutation in GJB2, 35delG, was screened by ASPCR. Homozygotes for this mutation were diagnosed as having GJB2-related hearing loss and no other studies were performed. 35delG heterozygotes were screened by both single strand conformational polymorphism (SSCP) analysis and direct sequencing to determine if a second mutation was present. If no 35delG alleles were identified by ASPCR, the GJB2 coding sequence was screened by SSCP and sequencing was performed on samples with SSCP shifts. In all cases with only a single coding sequence mutation, the non-coding exon of GJB2 (exon 1) was sequenced. PCR products were cleaned using a QIAquick PCR purification kit (Qiagen, Chatsworth, CA) and directly sequenced.

Statistical Comparison

The performance scores of individuals with GJB2-related deafness were compared to the scores of individuals with GJB2-unrelated deafness. For categorical data, the Fisher exact test was used to compare the groups. For scaled data, the Student's t-test was used. Evaluation also was performed, including test age and duration of implant, to exclude these factors as confounding variables.

RESULTS

Mutations in GJB2 were the most common identified etiology of congenital deafness in children receiving cochlear implants. Eight of twenty (40%) cochlear implant recipients with an unknown etiology of deafness had deafness-causing mutations in GJB2. Of the three cochlear implant recipients with a deaf sib, two had GJB2-related deafness. The most common mutation in GJB2 within the cochlear implant population was 35delG, and all cochlear implant recipients with deafness from GJB2 mutations carried this mutation (four were 35delG heterozygotes and four were 35delG homozygotes). SSCP screening of the GJB2 coding sequence (exon 2) was positive in three of four heterozygotes, and in the remaining individual, a mutation was found in exon 1. Sequencing identified the second mutation at the nucleotide level in each person (167delT, 269^270insT, W77R, and IVIS1+1G-A). SSCP also identified two additional individuals with benign polymorphisms (M34T/+, R98Q/+) unrelated to their deafness [Green et al., 1999] (Table I).

Table I.

GJB2 Mutations Identified in 20 Cochlear Implant Recipients

Deafness-causing mutationsa	Benign polymorphismsb
35delG/35delG	R98Q/+
35delG/35delG	M34T/+
35delG/35delG
35delG/35delG
35delG/W77R
35delG/167delT
35delG/269^270insT
35delG/IVISl + 1G-A

^aTwo mutations were found in all patients with GJB2-related deafness.

^bIn those patients with benign polymorphisms, the second allele was the standard wild type (+).

Cognitive assessment of cochlear implant recipients was similar in persons with GJB2-related deafness when compared to persons with GJB2-unrelated deafness (Hiskey-Nebraska test of learning aptitude, 109 vs. 94 [P = 0.474]). A trend toward slightly poorer performance on speech recognition parameters was noted in the GJB2-mutation group, but none of the test differences was significant: WIPI, 61 vs. 73 (P=0.36); VP, 84.4 vs. 88.3 (P = 0.55); PBK-wd, 24.5 vs. 33.1 (P = 0.53); PBK-ph, 44 vs. 55 (P = 0.44); MST word, 80 vs. 90 (P = 0.34); MST pattern, 88 vs. 95 (P = 0.27); FST, 90 vs. 95 (P = 0.38). Inclusion of age at testing and duration of implant as potentially confounding variables did not substantively alter these results (Table II). All cochlear implant users demonstrated improved speech perception skills compared to pre-operative conditions, with a trend for continued improvement with time.

Table II.

Age-at-Implantation (years.months)

GJB2-related deafness	GJB2-unrelated deafness
2.8	1.5
2.9	3.10
2.9	4.2
3.3	5.1
3.4	5.5
5.6	5.8
6.2	6.8
11.11	9.1
	9.6
	9.8
	11.9
	14.0

Reading performance was more consistent in the GJB2-mutation group. Seven of the eight patients with GJB2-related deafness with sufficient follow-up to obtain reading scores read within one standard deviation of age level or better (compared with 78% of cochlear implant recipients with GJB2-unrelated congenital deafness, P = 0.12) (Fig. 1). If we arbitrarily define a delayed reader as someone reading at more than one grade level below expectation for age, the number of delayed readers with GJB2-related deafness was 14%. By comparison, 44% of individuals with GJB2-unrelated deafness were delayed readers.

Fig. 1.

Reading scores of cochlear implant recipients with sufficient follow-up data. Dashed lines mark one standard deviation above and below average for hearing children. Note that the lowest performers have deafness unrelated to GJ7B2-mutations. There is a trend for children with deafness unrelated to GJB2 mutations to be more likely to read greater than one standard deviation below average (P = 0.117).

DISCUSSION

Mutations in GJB2 are one of the most common causes of congenital deafness in the United States, affecting over one-third of children evaluated for congenital severe-to-profound sensorineural hearing loss, and half of children with heritable hearing loss. Numerous deafness-causing mutations of this gene have been reported, although a single mutation, 35delG, predominates. Overall carrier rates for deafness-causing GJB2 mutations in the general Caucasian population in the United States approximate 3.0% [Green et al., 1999]. Over two-thirds of children identified as carrying two mutant alleles have profound or severe-to-profound hearing loss, although there can be wide phenotypic variability in the degree of loss even within the same family [Denoyelle et al., 1997].

Based on the data in this report, the etiology of non-syndromic hearing loss in prelingual cochlear implant recipients can be defined more precisely. Before genetic testing, we found that 51% of patients had congenital deafness of unknown etiology (10% with familial recurrence), 21% had hearing loss from meningitis, 5% had an inner ear anomaly, 9% had exposure to an in utero infection, 8% had syndromic hearing loss, and 6% had a known anoxic or ototoxic cochlear insult (Fig. 2). These data are consistent with studies from Texas and nationally, and correspond to a relatively higher percentage of meningitis-related deafness and a lower percentage of known hereditary deafness than are seen in the general deaf population [Allen et al., 1993; Fryauf-Bertschy et al., 1997; Woodcock, 1997].

Fig. 2.

Deafness from GJB2-mutations is estimated to account for 20% of prelingual deafness in cochlear implant recipients at the University of Iowa. In general, GJB2-mulations accounted for ~40% of deafness of previously undefined etiology, and ~50% of recessive deafness of undefined etiology.

By offering genetic testing, we were able to determine that GJB2-related deafness accounted for 40% of congenital hearing loss in the unknown etiology group (Fig. 2). On this basis, in the entire group of prelin-gually deaf cochlear implant recipients, we estimate that approximately 20% have deafness from GJB2 mutations. Due to the large number of individuals with GJB2-related deafness, we were able to examine and compare reading performance—86% of cochlear implant recipients with GJB2-related deafness read within one year of grade level compared to only 56% of cochlear implant recipients with GJB2-unrelated hearing loss. Although the first group cannot be shown to be superior in reading performance to the second group due to a limited sample size, persons with GJB2-related deafness can be ascribed as reading at least as well as those without GJB2-related deafness. Cochlear implant recipients in both groups outperformed children with profound hearing loss not receiving implantation, a group in which typically less than 10% read within one year of grade level [Holden-Pitt, 2000].

Because approximately 25% of congenitally deaf persons have additional disabilities, the identification of GJB2-related deafness carries prognostic significance [Cremers et al., 1991]. Persons with GJB2-related deafness are healthy—GJB2 mutations are not associated with mental retardation, vision difficulties, or sudden death, all of which can affect specific segments of the deaf community [Schildroth, 1994]. Vestibular function appears to be normal, and walking and sitting occur at age-appropriate times. We have been unable to identify any cognitive deficiencies or health risks associated with GJB2-related deafness in comparison with the deaf community as a whole.

Establishing a diagnosis of GJB2-related deafness also impacts genetic counseling. The recurrence chance for a hearing couple having a child with GJB2-related deafness is approximately 25%, consistent with an autosomal recessive inheritance pattern. In comparison, the recurrence chance for a hearing couple having a child with GJB2-unrelated deafness is approximately 14% [Green et al., 1999].

The data presented here have additional implications on the effect of in utero sound deprivation, the effect of childhood sound deprivation, and the effect of deafness comorbidity on development. Since several reports have documented benefits associated with early cochlear implantation, genetic screening may be useful to identify a subset of patients for whom very early implantation can be an option [Rizer and Burkey, 1999; O'Donoghue et al., 2000]. However, a larger number of cochlear implant recipients must be studied to achieve stronger statistical significance before more definitive statements can be made in this regard.