- Email: [email protected]
Vestibular impairments pre- and post-cochlear implant in children
International Journal of Pediatric Otorhinolaryngology (2009) 73, 209—217
www.elsevier.com/locate/ijporl
Vestibular impairments pre- and post-cochlear implant in children Etienne Jacot, Thierry Van Den Abbeele, Hopital Robert Debre, Sylvette R. Wiener-Vacher * ´partement d’ORL, Unite ´ de Vestibulome ´trie, Hopital Pe ´diatrique Robert Debre ´, De ´rurier, 75019 Paris, France 48 Bld Se Received 25 August 2008; received in revised form 9 October 2008; accepted 10 October 2008 Available online 19 December 2008
KEYWORDS Sensorineural hearing loss; Pediatric; Vestibular impairment; Prevalence; Risks; follow-up
Summary Objectives: Determine prevalence and types of vestibular impairments in sensorineural hearing loss (SNHL) in a large population of pediatric candidates for cochlear implants. Evaluate impact of cochlear implants on vestibular function. Study design: Retrospective and prospective study. Methods: Children with profound SNHL (n = 224) underwent complete vestibular testing (clinical vestibular examination, bicaloric test, earth vertical axis rotation, off vertical axis rotation and vestibular evoked myogenic potentials) before cochlear implant. Changes in vestibular responses were measured after implants in 89 of these patients. Results: In the SNHL population only 50% had normal bilateral vestibular function, while 20% had bilateral complete areflexia, 22.5% partial asymmetrical hypoexcitability and 7.5% partial symmetrical hypoexcitability. In the 71/89 follow-up patients showing vestibular responses prior to implant, 51 (71%) had changes in vestibular function including 7 (10%) who acquired ipsilateral areflexia. Others developed ipsilateral hypo- or hyperexcitability. Vestibular modifications occurred during the 3 months after surgery and were not clearly associated with clinical signs except for ipsilateral areflexia cases. In long-term follow-up, two of the 7 patients with ipsilateral areflexia partially recovered vestibular function. Conclusion: Since half of pediatric cochlear implant candidates have vestibular deficits and 51% of implants induce modifications of existing vestibular function, each implant should be preceded by canal and otolith functional tests to assure that the least functional vestibule is implanted. The tests provide baselines for follow-up monitoring of subsequent losses and recovery. This could be easily implemented with a clinical vestibular examination including the head thrust test associated with a bicaloric test and vestibular-evoked-myogenic-potentials. # 2008 Published by Elsevier Ireland Ltd.
* Corresponding author. Tel.: +33 1 40032479; fax: +33 1 40 03 2202. E-mail address: [email protected] (S.R. Wiener-Vacher). 0165-5876/$ — see front matter # 2008 Published by Elsevier Ireland Ltd. doi:10.1016/j.ijporl.2008.10.024
210
1. Introduction Over the last 20 years cochlear implants have been improving rehabilitation of hearing impairments. This is particularly effective in young children permitting them to acquire effective communication skills early in development. In children with severe to profound sensorineural hearing loss (SNHL), particularly in prelingual, congenital or early-onset cases, implants have been applied worldwide. However, cochlear implantation is not without risk to the nearby vestibular canals and otolith organs. There are few studies of the consequences of cochlear implants on vestibular function. Observations made on adults showed vestibular dysfunctions after surgery and following cochlear stimulation [1—3]. Furthermore, there are only limited population studies of the prevalence of vestibular impairments in SNHL candidates for cochlear implants in children. This raises the risk that existing vestibular function could be dramatically attenuated or suppressed by cochlear implant. Vestibular testing is not part of the standard battery of tests before and after cochlear implant surgery. Furthermore the risk of vestibular impairment after implant is likely to be more underestimated in children than in adults because of their remarkable capacity to rapidly compensate acute vestibular loss. However it is now known that a congenital or early acquired complete loss of vestibular information provokes severe delays in the acquisition of the first posturomotor milestones as stably holding the head, sitting and walking independently [4—6]–— and children are now being implanted as young as six months of age. It was thus important to test the prevalence of vestibular deficits in SNHL implant candidates and to monitor the vestibular consequences of cochlear implants in children. Reliable vestibular testing in children has been developed very recently [6—9] but remains not available in most cochlear implant centers. A large population study was possible in our department at the Robert Debre ´ Pediatric Hospital Complete where a complete vestibular testing battery has been applied on a daily basis for 15 years on children (as young as 1 month of age). The first step of our study is to determine the vestibular status in a large population of SNHL patients programmed to receive a cochlear implant. Patients were taken from our department as well as other hospitals and clinics. Secondly, we evaluated changes in vestibular function in patients operated in our department.
2. Materials and methods This prospective and retrospective study is based on 224 SNHL patients (implant candidate group) tested
E. Jacot et al. for vestibular function status before cochlear implant and referred by two Parisian implant centers. This population consists of 117 boys, 107 girls with a mean age at the first examination of 51 34 months (median: 3.5 years, range: 7 months—16.5 years). All patients and their families were counseled regarding study participation and written and verbal informed consent was obtained as required by the French institutional review board (CCPPRB). Of these, 89 also agreed to one or more vestibular assessments after the surgery ( follow-up group) (delay after surgery ranged from 1 week to 7 years, median: 7.5 months). This population consists of 47 boys and 42 girls with a mean age of 52.8 34 months (median: 4 years, range from 7 months to 12 years) at first examination. Data from this population are presented in Table 1. In our series 66% (59/ 89) of the implants were Advanced Bionics, 17% (17/ 89) Cochlear and 13% (13/89) Neurelec-MXM. The same operating technique was used for all children. The electrode was inserted into the cochlea through a minimal 1.2 mm cochleostomy just above the round window. For all cases the surgical procedure was performed by senior surgeons. The vestibular evaluation included several tests for canal and otolith functions. In our protocol a complete vestibular clinical evaluation [10,11] precedes several tests of responses to vestibular stimulation: classic bicaloric test (33 and 44 8C), earth vertical axis rotation (EVAR) with 408/s2 acceleration [12], off vertical axis rotation (OVAR) with 608/s constant rotation velocity [12] and 138 axis rotation tilt (parameters which are well tolerated, in contrast those employed in the early days of OVAR), and vestibular evoked myogenic potentials (VEMP) with tone bursts (air and bone conduction) of 750 Hz, 4.1/s and 6 ms duration and control of the EMG level for each stimulation. In the follow-up group, 87 (98%) had a bicaloric test, and 82 (92%) had both EVAR and bicaloric tests. Seventy-five patients (84%) had EVAR, bicaloric tests and OVAR. Only 50 patients (56%) also had VEMP, a test recently adapted for young children. The technical details about these tests have already been published [13]. The VEMP test was performed with a BERA equipment (Centor C+, Deltamed1, France) with modified parameters adapted to EMG recordings. The EVAR-OVAR tests were performed with a computer assisted chair (SAMO1, France). In the clinical examination we emphasize the importance of the Head Thrust test [10] which is easily adapted for children [11]. The presence of a catch-up saccade on the side of the rotation while the patient’s head is passively rotated indicates a vestibular canal deficit. For the bicaloric test, Jonkees formula is applied, where normal values for
Vestibular impairments pre- and post-cochlear implant in children Table 1 Population characteristics of the 89 patients with vestibular follow-up after implant. Mean age at implantation (months) Mean age at hearing loss diagnosis (months)
56 33 (range: 7—173) 18 17
Sex Boys Girls
N = 47 (53%) N = 42 (47%)
Implants Cochlear MXM Advanced Bionics Undetermined
14 (16%) 13 (15%) 56 (62%) 6 (7%)
Hearing loss origins Unknown Aquired Genetic Sydromic Postmeningitic Immunologic CMV Mondini
52 (59%) 6 (7%) 14 (16%) 1 (1%) 4 (4%) 1 (1%) 4 (4%) 7 (8%)
CT scan, NMR Normal Cochlea malformation Vestibular malformation Cochleo-vestibular malformation Cochleo-vestibular malformation in association (syndrome) Undetermined
211
to the normative data established in our department for normal hearing children for the same range of age and sex [12]. The statistical analysis were performed with Sigmaplot# program and included Normality test, Mann—Whitney Rank Sum Test and Spearman Rank Order Correlation Test.
3. Results
60 (67%) 1 (1%) 3 (3.5%) 14 (16%) 4 (4.5%)
7 (8%)
relative valence and directional preponderance for children are inferior or equal to 15%. Only the presence or absence of VEMP with normal latencies (regardless of amplitude or threshold) is considered as a criterion. For the EVAR test, we measured time constant and maximal initial slow phase velocity of vestibulo-ocular responses (VORs). For OVAR the parameters analyzed are the bias and the modulation amplitude of VOR horizontal and vertical components [12,14]. For global evaluation we collected all results and defined 4 categories of modifications: areflexia when no responses was found at least in the head thrust test, the caloric test, EVAR and VEMP; hypoexcitability when responses to canal and/or otolith tests were smaller in amplitude than normal; hyperexcitability when responses to canal and/or otolith tests were larger in amplitude than normal; mixed when some signs of hypoexcitability were found in some tests and hyperexcitability in others. To evaluate the prevalence of vestibular impairment in our implant candidate group and follow-up group, we compare the values obtained in all tests
Our data determined the prevalence of vestibular impairment before cochlear implant in a SNHL population (n = 224). The various parameters collected for each test before and after implant are plotted in Fig. 1A—F, which identify the patients out of the normal range. Each follow-up patient outside the normal ranges was identified for further analyses. Tables 2 and 3 provide normative values of the parameters of the VOR obtained for EVAR and OVAR in the implant candidate and follow-up populations (Fig. 2).
3.1. High prevalence of vestibular impairments in SNHL pediatric patients Surprisingly only half of the patients had normal symmetrical vestibular function. That is 51% had abnormal canal responses and 45% had abnormal otolith responses (Table 4). Vestibular function was asymmetric for canal function in 20% and for otolith function 19%. These data were used to select the side with weaker vestibular function for the implant. The presence of a vestibular malformation does not help to predict functional vestibular impairment. Of the 17 children with vestibular malformation in the CT scan, 47% had normal vestibular function, 35% a unilateral areflexia and none a bilateral areflexia prior to the implantation. No correlation was found between the cause of hearing loss and the vestibular dysfunction; however we did not have in our series any cases of Usher syndrome that is known to associate bilateral SNHL and bilateral vestibular loss. The follow-up group (n = 89) was similar to remaining members of the implant candidate group (n = 135); with respect to incidences of various vestibular impairments prior to surgery on the global canal and otolith evaluation (for canal: p = 0.240, for otolith p = 0.310).
3.2. Severe vestibular modifications following cochlear implant The observation of canal VOR on the 89 patients after unilateral implants (Fig. 3) shows no modifica-
212
E. Jacot et al.
Fig. 1 These figures are comparing values before and after implant for Bicaloric test, EVAR and OVAR. The shaded area corresponds to the normal values. Black dots indicate patients whose values are outside the normal range before, after
Vestibular impairments pre- and post-cochlear implant in children
Fig. 2 Vestibular status of follow-up group (n = 89 patients) with SNHL before implant.
tions in 43 of the 71 cases with prior vestibular function. A complete ipsilateral areflexia (absence of canal VOR at all tests) appeared in seven of the patients implanted on a functional vestibule. A statistically significant decrease of the canal VOR was observed in 12 cases. An increase of canal VOR (a sign of canal hyperexcitability) on the implanted ear was found in 5 cases. In three cases the cochlear implant was associated with either a decrease or an increase of the canal VOR leading to symmetric responses. In only one case we did observe a decrease of the canal VOR on the contralateral side. At the time of the test this patient presented an ongoing CMV infection. After implant the otolith responses in the followup group (Fig. 4) were modified in 39/71 (55%) of the cases with prior vestibular function. A statistically significant decrease of the otolith VOR and vestibulospinal responses was observed in 9 cases. A bilateral decrease of the responses was found in 4 cases in the OVAR test while VEMP remained present on both sides in these four cases. In 17 cases the cochlear implant was associated with an increase of the otolith VOR in the OVAR test with an intense directional preponderance toward the implanted side (a sign of hyperexcitability). In four cases the modifications were variable from one test to another (equalization of responses, mixed modifications).
213
Fig. 3 Distribution of the canal VOR modifications after cochlear implant (n = 89 patients).
Fig. 4 Repartition of the modifications of otolith responses after cochlear implant (89 patients).
Of the 89 patients implanted, 24 presented symptoms (like vomiting or dizziness) that could evoke vestibular impairment, mostly during the first 24— 48 h after surgery. Only 14 of them showed changes in vestibular function in the tests done during the following days. However, in the seven patients who showed complete ipsilateral vestibular areflexia, 6 presented strong initial symptoms. In our series the longest delay for vestibular areflexia to occur was 1.5 months after surgery but milder secondary aggravation of vestibular impairment occurred as late as four years after implant (observed in a case with several tests on a long follow-up). The profound vestibular loss occurred immediately after surgery for five patients, after 4 days for one and 1.5 months for the remaining one. The latter developed severe instability and vertigo one week after changing
cochlear implant or both while white dots indicate patients with normal responses. Values located on or close to the diagonal dotted line correspond to patients with no modifications of their vestibular score between the vestibular assessments made before and after cochlear implant. For bicaloric test we consider mean relative valence (A) and directional preponderance (B) calculated with Jonkee’s formula. For EVAR test we consider mean relative initial maximal velocity (C) and mean relative time constant (D) calculated as follow: ((value clockwise value counterclockwise)/ (value clockwise + value counterclockwise)) 100. For OVAR test we consider relative directional preponderance (E and F) given by the bias of the horizontal and vertical components of the responses and calculated as follow: ((bias clockwise + bias counterclockwise)/2). Horizontal and vertical modulations of VOR for OVAR was not considered because values were too scattered to be of any interest and therefore not shown.
0.70 0.51 (27.9 20.4) 0.58 0.47 (23.4 18.7) 0.72 0.52 (28.7 21.0) 0.60 0.49 (23.9 19.5) 7.5 6.1 6.0 4.7 7.1 5.9 7.1 5.5
Before implant (n = 89) After implant (n = 89)
6.4 4.5 7.3 5.2 6.9 4.9
7.2 5.3
0.77 0.52 (31.2 20.9) 0.67 0.47 (26.7 18.8)
0.75 0.54 (29.9 21.5) 0.60 0.46 (23.9 18.6)
Post-rotatory (CW rotation) Per-rotatory (CCW rotation) Post-rotatory (CCW rotation) Per-rotatory (CW rotation) Per-rotatory (CCW rotation) Post-rotatory (CCW rotation) Per-rotatory (CW rotation)
Post-rotatory (CW rotation)
Non-implanted side Implanted side
Implanted side
Non-implanted side
E. Jacot et al. Table 2 Vestibulo-ocular responses (VOR) for earth vertical axis rotation (EVAR) for the follow-up group before and after implants: mean values and standard deviation are listed. Data are presented such as all the patients were implanted on the same side (right). CW and CCW indicate the direction clockwise or counter-clockwise of rotation. Populations VOR mean time constant at EVAR (s) VOR slow phase at EVAR: mean gain/velocity ˚(/s)
214
the stimulation program of the implant. The areflexia did not recover after changing back to the previous program. In our series, no predictive factors for vestibular impairment following cochlear implant were found. The occurrence of areflexia or hypoexcitability was not correlated to age or the type of implant or surgeon’s identity. Of the 7 patients who developed areflexia three received an Advanced Bionics implant and four a Cochlear implant. None of the SNHL etiologies mentioned in Table 1 are correlated to a higher risk of vestibular impairment after implant. Of the 7 patients who developed areflexia, two presented a Mondini, and one presented a CMV infection.
3.3. Long-term follow-up of vestibular modifications following cochlear implant In the follow-up group, 27 were tested further at post-operative delays from 3 months to 7 years (mean: 26 19 months, median: 18 months). Long-term vestibular modifications after implant are summarized in Table 5. Over the long term, 63% of the children show stable responses to the vestibular tests whatever results were found at first assessment after surgery. One patient with no modification after surgery showed a decrease of vestibular responses on the implanted side nine and half months later. Partial recovery of vestibular responses is observed in 18.5% of the cases. A complete recovery was observed in only one case with hypoexcitability. In 3 cases (11%) the decreased vestibular responses deteriorated over time. The five patients with high directional preponderance toward the implanted ear (sign of hyperexcitability) recovered in 2 cases while in the 3 others the hyperexcitability persisted over a 4—7 year period. Complete areflexia remained unchanged in 4 cases but surprisingly partially recovered in two cases as late as 1.5 and 3.5 years after implant. The finding of a hypoexcitability at first vestibular assessment was followed in 2 cases by a further decrease of the vestibular responses and in 2 cases by a partial or complete recovery.
4. Discussion This is the first report on the prevalence of canal and otolith vestibular impairment in a large pediatric population with profound SNHL, complementing the study of Kaga and co-workers on 20 children [9]. In the implant candidate population only 50% had normal vestibular function, while 20% had complete
Vestibular impairments pre- and post-cochlear implant in children
215
Table 3 Vestibulo-ocular responses (VOR) for off vertical axis rotation (OVAR) for the follow-up group before and after implant: mean values and standard deviation are listed. Data are presented such as all the patients were implanted on the same side (right). CW and CCW indicate the direction clockwise or counter-clockwise of rotation. Populations Bias ˚(/s) Moduation ˚(/s)
Before implant (n = 89) After implant (n = 89)
Implanted side (CW rotation)
Non-implanted side (CCW rotation)
Implanted side (CW rotation)
Non-implanted side (CCW rotation)
Horizontal
Vertical
Horizontal
Vertical
Horizontal
Vertical
Horizontal
Vertical
0.33 1.8
0.53 4.7
0.13 2.3
0.39 4.5
3.3 2.3
4.7 4.3
3.3 2.4
5.1 4.2
0.4 3.1
0.2 4.6
0.3 2.1
0.2 3.3
2.6 2.2
3.8 4.3
2.6 2.1
5.2 5.5
Table 4 Canal and otolith status in the candidate for implant group of 224 patients versus the follow-up group of 89 patients. Bilateral symetrical hyporeflexia
Bilateral asymetrical hyporeflexia
Unilateral hyporeflexia
Mann— Whitney test
Population
Normal
Bilateral areflexia
Unilateral areflexia
Canal responses (n = 224—89) Canal responses (n = 89)
43% 60%
30% 18%
8% 6%
7% 4%
4% 3%
6% 9%
p = 0.240
Canal responses (n = 224)
49%
25%
8%
6%
4%
8%
—
Otolith responses (n = 224—89) Otolith responses (n = 89)
50% 62%
18% 13%
5% 1%
12% 10%
5% 5%
10% 9%
p = 0.310
Otolith responses (n = 224)
55%
16%
4%
10%
5%
10%
—
Table 5 Observation of vestibular modifications after implant over a long-term period (n = 27). Vestibular modifications at first assessment (number of patients)
Long-term modifications of the vestibular responses
No modifications (n = 7) Areflexia (n = 6) Hypoexcitability (n = 8) Hyperexcitability (n = 5) Mixed (n = 1)
6 4 4 3 1
Persistence (number of patients)
Complete recovery (number of patients)
Partial recovery (number of patients)
1
2 1 2
Worsening (number of patients) 1
bilateral areflexia, 7.5% bilateral symmetrical impairment and 22.5% partial asymmetrical impairment. Our data are inconsistent with some published results [3,9] likely because of the smaller samples in these studies. Our data show no correlation between vestibular and hearing status, thus it is impossible to predict vestibular impairment type or severity from hearing loss characteristics. Our observation shows also that it is not possible to predict on clinical cues the integrity of vestibular system in young children. In the follow-up population, cochlear implants had dramatic effects on vestibular canal and otolith function. This has been previously underestimated in children [15]. The risk of complete areflexia due to the implant is evaluated here at 10%. Most of the other vestibular modifications described in our ser-
2
ies were clinically silent and could only be found through systematic testing. This absence of clinical signs could be explained by the well-known rapid compensation of sensory deficits in children. With a complete canal evaluation (bithermal caloric test, Halmagyi test and EVAR test) we found canal paresis in 30% of children after cochlear implant. A recent study on vestibular disturbances after cochlear implants in adults (on a cohort of 146 patients) shows that 32% showed vestibular disturbances and weaker responses to bithermal caloric test after surgery [16]. In another study of adults receiving cochlear implants where canal function was measured with the head impulse test with eye coil recording, 10% (1/11) had canal functional loss after implant [17], corresponding to our results in children. These results support the idea that the
216 canal responses are as vulnerable to cochlear implant in children as they are in adults regardless of age or cause of deafness. Mechanism of implant provoked vestibular deficits is not yet completely known. Nevertheless the literature offers several explanations: an inflammatory process, inner ear liquid imbalance or microfractures [3,18]. Our data offer possible explanations. In most cases the direct traumatic effect of the electrode during its introduction in the cochlea can be implicated when the vestibular loss is immediately diagnosed after surgery. In one case a 4 days delay before the onset of vestibular impairment argues against such a mechanism and rather supports the hypothesis of a leak of inner ear fluids. One patient presented vestibular impairment following the switch from one stimulation program to another with higher stimulation level. The direct effect of electrical stimulation could be the source of the vestibular loss in this case. The absence of recovery after switching back to the previous stimulation program suggests a permanent lesion of vestibular receptors. In addition our observation in several cases of persistent hyperexcitability of canal and otolith systems suggests a direct effect of the electrical stimulation of vestibular receptors through the cochlear implant, as already suggested by several authors in adults [19]. Moreover we observed in some cases with no vestibular responses a revival of the responsiveness after implant. After an initial impairment, the partial or complete recovery of the vestibular responsiveness with time show that the vestibule can recover. The majority of the various vestibular modifications following the implant seem to be quite stable. On the 27 patients whom had a long-term followup, 16 patients (59%) showed unmodified vestibular responses through time. Two of the 7 patients with areflexia recovered partially their vestibular responses. One patient with hypoexcitability showed a partial recovery of its vestibular responses, one a complete recuperation and three showed aggravation of their vestibular impairment. Despite a small number of patients with very long follow-up, our data suggest that even severe vestibular impairment is not necessarily permanent and can evolve over a long period.
4.1. Suggestions for vestibular management of SNHL cochlear implant candidates It is important to select the side for a cochlear implant as the least functional vestibule in order to limit the occurrence of bilateral vestibular loss in case there may only be one functional vestibule,
E. Jacot et al. except in the cases where audiological or anatomical criteria are of main concern. Bilateral vestibular loss is known to have particularly dramatic consequences for very young children (as well as in older people) [4—6]. This choice can be only done with the help of a vestibular assessment of canal and otolith functions. In our series, 22.5% of our patients were implanted in the less functional vestibule except in 5 cases where audiological or anatomical criteria were of main concern. From our experience the tests that were most sensitive for detecting vestibular impairment were the head thrust test and bicaloric test for canal assessment and VEMP for otolith assessment. These tests do not require very expensive or sophisticated equipment and can be developed in any pediatric cochlear implant unit. Bicaloric and VEMP tests evaluate peripheral vestibular pathways and therefore are not submitted to central compensation processes as much as EVAR and OVAR. The time constant in EVAR can detect a vestibular deficit when it is recent. However, with time, the initial asymmetry disappears progressively to be substituted by symmetrical short time constants on both sides. As for the EVAR test, the results with OVAR were only relevant in cases of recent and complete vestibular deficit (asymmetry of the modulation and directional preponderance toward the opposite side); Furthermore it detected hyperexcitability of the otolith system on the implanted side by an intense and durable bias with a directional preponderance toward the implanted side. We define the three months after implant as a high-risk period for the appearance of vestibular impairments. At the end of this period we propose programming regular vestibular assessments in order to detect eventual complications. Moreover in vestibular impaired patients even a longer followup can be foreseen. Such testing will permit parents to be more completely informed of the risk of vestibular impairment should the implant be done on a functional vestibule (80% of our patients). The incidence of inducing vestibular areflexia during cochlear implant on a functional vestibule is 10% and this consequence is unpredictable. This risk must be taken into account when a patient presents already a unilateral areflexia (5.5% in our population). Furthermore this non-negligible risk of permanent vestibular dysfunction would argue against bilateral cochlear implantation in a single surgical procedure without previous vestibular assessment. An exception to this dictum should be made when the deafness is secondary to bacterial meningitis (in particular Pneumococcus sp.) that is followed by a rapidly developing endolabyrinthitis ossificans pre-
Vestibular impairments pre- and post-cochlear implant in children venting future cochlear implant or when preimplant vestibular testing shows that complete bilateral vestibular areflexia is already present (such as Usher syndrome).
5. Conclusions Pediatric SNHL cochlear implant candidates have a high incidence (50%) of vestibular impairments. Cochlear implants affect vestibular function in a high proportion of cases (50%) with 10% of vestibular loss. Vestibular assessment protocols should be instated for pediatric SNHL candidates for implant. This would include at least a clinical vestibular examination including the head thrust test, a bicaloric test and VEMP to decide which ear to implant (the most poorly functioning vestibule). Parents should be informed of the risk of vestibular loss and possible temporary imbalance during a 3 months period after surgery in case of an implant on a functional vestibule. The interest of a systematic vestibular assessment after implant should be discussed as well as the necessity of a vestibular function follow-up particularly when vestibular implant effects are detected. Our results support the necessary development of specific programs for vestibular assessment in pediatric cochlear implant units.
Acknowledgements This work has been supported by: Fondation Genevoise de Bienfaisance Valeria Rossi di Montelera, Fondation de France, Fondation Cotrel. Thanks to Franc¸oise Toupet for her collaboration during clinical testing, Pierre Denise for his scientific and technical support, Sidney Wiener for his helpful comments on the manuscript.
References [1] D. Vibert, R. Hausler, M. Kompis, M. Vischer, Vestibular function in patients with cochlear implantation, Acta Otolaryngol. 545 (Suppl.) (2001) 29—34. [2] R.H. Brey, G.W. Facer, M.B. Trine, S.G. Lynn, A.M. Peterson, V.J. Suman, Vestibular effects associated with a multiple channel cochlear prosthesis, Am. J. Otol. 16 (16) (1995) 424—430.
217
[3] C.A. Buchman, J. Joy, A. Hodges, F.F. Telschi, T.J. Balkany, Vestibular effects of cochlear implantation, Laryngoscope 114 (October (Suppl. 103)) (2004) 1—22. [4] S.R. Wiener-Vacher, F. Toupet, M. Franc¸ois, T. Van Den Abbeele, P. Viala, P. Narcy, Conse ´quences des de ´ficits vestibulaires sur le de ˆle postural chez ´veloppement du contro l’enfant. Flammarion Me ´decine Sciences, Journe ´es Parisiennes de Pe ´diatrie (1998) 283—287. [5] K. Kaga, Vestibular compensation in infants and children with congenital and acquired vestibular loss in both ears, Int. J. Ped. Otorhinolaryngol. 49 (1999) 21524. [6] S.R. Wiener-Vacher, L. Amanou, P. Denise, P. Narcy, Y. Manach, Vestibular function in children with CHARGE association, Arch Otolaryngol. Head Neck Surg. 125 (March (3)) (1999) 342—347. [7] T.D. Fife, R.J. Tusa, J.M. Furman, D.S. Zee, E. Frohman, et al., Assessment: vestibular testing techniques in adults and children, Neurology 55 (2000) 1431—1441. [8] T.A. Kelsch, L.A. Schaefer, C.R. Esquivel, Vestibular evoked myogenic potentials in young children: test parameters and normative data, Laryngoscope 116 (June (6)) (2006) 895— 900. [9] Y. Shinjo, Y. Jin, K. Kaga, Assessment of vestibular function of infants and children with congenital and acquired deafness using the ice-water caloric test, rotational chair test and vestibular-evoked myogenic potential recording, Acta Otolaryngol. 127 (July (7)) (2007) 736—747. [10] G.M. Halmagyi, Diagnosis and management of vertigo, Clin. Med. 5 (March—April (2)) (2005) 159—165. [11] S.R. Wiener-Vacher, Vertigo in children, EMC-Oto-rhino-laryngologie 2 (2005) 230—248. [12] S.R. Wiener-Vacher, F. Toupet, P. Narcy, Canal and otolith vestibulo-ocular reflexes to vertical and off vertical axis rotations in children learning to walk, Acta Otolaryngol. 116 (September (5)) (1996) 657—665. [13] M.S. Welgampola, J.G. Colbatch, Characteristics and clinical applications of vestibular-evoked-myogenic potentials, Neurology 64 (May (10)) (2005) 1682—1688. [14] S.R. Wiener-Vacher, K. Mazda, Asymmetric otolith vestibuloocular responses in children with idiopathic scoliosis, J. Pediatrics 132 (6) (1998) 1028—1032. [15] D. Bouccara, M.J. Este `ve-Fraysse, N. Loundon, B. Fraysse, N. Garabedian, O. Sterkers, Vestibular dysfunction after cochlear implantation: a national multicenter clinical study, Rev. Laryngol. Otol. Rhinol. (Bord.) 126 (4) (2005) 275—278. [16] J.C. Enticott, S. Tari, S.M. Koh, R.C. Dowell, S.J. O’Leary, Cochlear implant and vestibular function, Otol. Neurotol. 27 (6) (2006) 824—830. [17] A.A. Migliaccio, C.C. Della Santina, J.P. Carey, J.K. Niparko, L.B. Minor, The vestibule-ocular reflex reponse to head impulses rarely decreases after cochlear implant, Otol. Neurotol. 26 (4) (2005) 655—660. [18] O. Handzel, B.J. Burgess, J.B. Nadol Jr., Histopathology of the peripheral vestibular system after cochlear implantation in the human, Otol. Neurotol. 27 (1) (2006) 57—64. [19] L. Sennaroglu, B. Gursel, G. Sennaroglu, E. Yucel, I. Saatci, Vestibular stimulation after cochlear implantation in common cavity deformity, Otolaryngol. Head Neck Surg. 125 (4) (2001) 408—410.
Available online at www.sciencedirect.com
www.elsevier.com/locate/ijporl
Vestibular impairments pre- and post-cochlear implant in children Etienne Jacot, Thierry Van Den Abbeele, Hopital Robert Debre, Sylvette R. Wiener-Vacher * ´partement d’ORL, Unite ´ de Vestibulome ´trie, Hopital Pe ´diatrique Robert Debre ´, De ´rurier, 75019 Paris, France 48 Bld Se Received 25 August 2008; received in revised form 9 October 2008; accepted 10 October 2008 Available online 19 December 2008
KEYWORDS Sensorineural hearing loss; Pediatric; Vestibular impairment; Prevalence; Risks; follow-up
Summary Objectives: Determine prevalence and types of vestibular impairments in sensorineural hearing loss (SNHL) in a large population of pediatric candidates for cochlear implants. Evaluate impact of cochlear implants on vestibular function. Study design: Retrospective and prospective study. Methods: Children with profound SNHL (n = 224) underwent complete vestibular testing (clinical vestibular examination, bicaloric test, earth vertical axis rotation, off vertical axis rotation and vestibular evoked myogenic potentials) before cochlear implant. Changes in vestibular responses were measured after implants in 89 of these patients. Results: In the SNHL population only 50% had normal bilateral vestibular function, while 20% had bilateral complete areflexia, 22.5% partial asymmetrical hypoexcitability and 7.5% partial symmetrical hypoexcitability. In the 71/89 follow-up patients showing vestibular responses prior to implant, 51 (71%) had changes in vestibular function including 7 (10%) who acquired ipsilateral areflexia. Others developed ipsilateral hypo- or hyperexcitability. Vestibular modifications occurred during the 3 months after surgery and were not clearly associated with clinical signs except for ipsilateral areflexia cases. In long-term follow-up, two of the 7 patients with ipsilateral areflexia partially recovered vestibular function. Conclusion: Since half of pediatric cochlear implant candidates have vestibular deficits and 51% of implants induce modifications of existing vestibular function, each implant should be preceded by canal and otolith functional tests to assure that the least functional vestibule is implanted. The tests provide baselines for follow-up monitoring of subsequent losses and recovery. This could be easily implemented with a clinical vestibular examination including the head thrust test associated with a bicaloric test and vestibular-evoked-myogenic-potentials. # 2008 Published by Elsevier Ireland Ltd.
* Corresponding author. Tel.: +33 1 40032479; fax: +33 1 40 03 2202. E-mail address: [email protected] (S.R. Wiener-Vacher). 0165-5876/$ — see front matter # 2008 Published by Elsevier Ireland Ltd. doi:10.1016/j.ijporl.2008.10.024
210
1. Introduction Over the last 20 years cochlear implants have been improving rehabilitation of hearing impairments. This is particularly effective in young children permitting them to acquire effective communication skills early in development. In children with severe to profound sensorineural hearing loss (SNHL), particularly in prelingual, congenital or early-onset cases, implants have been applied worldwide. However, cochlear implantation is not without risk to the nearby vestibular canals and otolith organs. There are few studies of the consequences of cochlear implants on vestibular function. Observations made on adults showed vestibular dysfunctions after surgery and following cochlear stimulation [1—3]. Furthermore, there are only limited population studies of the prevalence of vestibular impairments in SNHL candidates for cochlear implants in children. This raises the risk that existing vestibular function could be dramatically attenuated or suppressed by cochlear implant. Vestibular testing is not part of the standard battery of tests before and after cochlear implant surgery. Furthermore the risk of vestibular impairment after implant is likely to be more underestimated in children than in adults because of their remarkable capacity to rapidly compensate acute vestibular loss. However it is now known that a congenital or early acquired complete loss of vestibular information provokes severe delays in the acquisition of the first posturomotor milestones as stably holding the head, sitting and walking independently [4—6]–— and children are now being implanted as young as six months of age. It was thus important to test the prevalence of vestibular deficits in SNHL implant candidates and to monitor the vestibular consequences of cochlear implants in children. Reliable vestibular testing in children has been developed very recently [6—9] but remains not available in most cochlear implant centers. A large population study was possible in our department at the Robert Debre ´ Pediatric Hospital Complete where a complete vestibular testing battery has been applied on a daily basis for 15 years on children (as young as 1 month of age). The first step of our study is to determine the vestibular status in a large population of SNHL patients programmed to receive a cochlear implant. Patients were taken from our department as well as other hospitals and clinics. Secondly, we evaluated changes in vestibular function in patients operated in our department.
2. Materials and methods This prospective and retrospective study is based on 224 SNHL patients (implant candidate group) tested
E. Jacot et al. for vestibular function status before cochlear implant and referred by two Parisian implant centers. This population consists of 117 boys, 107 girls with a mean age at the first examination of 51 34 months (median: 3.5 years, range: 7 months—16.5 years). All patients and their families were counseled regarding study participation and written and verbal informed consent was obtained as required by the French institutional review board (CCPPRB). Of these, 89 also agreed to one or more vestibular assessments after the surgery ( follow-up group) (delay after surgery ranged from 1 week to 7 years, median: 7.5 months). This population consists of 47 boys and 42 girls with a mean age of 52.8 34 months (median: 4 years, range from 7 months to 12 years) at first examination. Data from this population are presented in Table 1. In our series 66% (59/ 89) of the implants were Advanced Bionics, 17% (17/ 89) Cochlear and 13% (13/89) Neurelec-MXM. The same operating technique was used for all children. The electrode was inserted into the cochlea through a minimal 1.2 mm cochleostomy just above the round window. For all cases the surgical procedure was performed by senior surgeons. The vestibular evaluation included several tests for canal and otolith functions. In our protocol a complete vestibular clinical evaluation [10,11] precedes several tests of responses to vestibular stimulation: classic bicaloric test (33 and 44 8C), earth vertical axis rotation (EVAR) with 408/s2 acceleration [12], off vertical axis rotation (OVAR) with 608/s constant rotation velocity [12] and 138 axis rotation tilt (parameters which are well tolerated, in contrast those employed in the early days of OVAR), and vestibular evoked myogenic potentials (VEMP) with tone bursts (air and bone conduction) of 750 Hz, 4.1/s and 6 ms duration and control of the EMG level for each stimulation. In the follow-up group, 87 (98%) had a bicaloric test, and 82 (92%) had both EVAR and bicaloric tests. Seventy-five patients (84%) had EVAR, bicaloric tests and OVAR. Only 50 patients (56%) also had VEMP, a test recently adapted for young children. The technical details about these tests have already been published [13]. The VEMP test was performed with a BERA equipment (Centor C+, Deltamed1, France) with modified parameters adapted to EMG recordings. The EVAR-OVAR tests were performed with a computer assisted chair (SAMO1, France). In the clinical examination we emphasize the importance of the Head Thrust test [10] which is easily adapted for children [11]. The presence of a catch-up saccade on the side of the rotation while the patient’s head is passively rotated indicates a vestibular canal deficit. For the bicaloric test, Jonkees formula is applied, where normal values for
Vestibular impairments pre- and post-cochlear implant in children Table 1 Population characteristics of the 89 patients with vestibular follow-up after implant. Mean age at implantation (months) Mean age at hearing loss diagnosis (months)
56 33 (range: 7—173) 18 17
Sex Boys Girls
N = 47 (53%) N = 42 (47%)
Implants Cochlear MXM Advanced Bionics Undetermined
14 (16%) 13 (15%) 56 (62%) 6 (7%)
Hearing loss origins Unknown Aquired Genetic Sydromic Postmeningitic Immunologic CMV Mondini
52 (59%) 6 (7%) 14 (16%) 1 (1%) 4 (4%) 1 (1%) 4 (4%) 7 (8%)
CT scan, NMR Normal Cochlea malformation Vestibular malformation Cochleo-vestibular malformation Cochleo-vestibular malformation in association (syndrome) Undetermined
211
to the normative data established in our department for normal hearing children for the same range of age and sex [12]. The statistical analysis were performed with Sigmaplot# program and included Normality test, Mann—Whitney Rank Sum Test and Spearman Rank Order Correlation Test.
3. Results
60 (67%) 1 (1%) 3 (3.5%) 14 (16%) 4 (4.5%)
7 (8%)
relative valence and directional preponderance for children are inferior or equal to 15%. Only the presence or absence of VEMP with normal latencies (regardless of amplitude or threshold) is considered as a criterion. For the EVAR test, we measured time constant and maximal initial slow phase velocity of vestibulo-ocular responses (VORs). For OVAR the parameters analyzed are the bias and the modulation amplitude of VOR horizontal and vertical components [12,14]. For global evaluation we collected all results and defined 4 categories of modifications: areflexia when no responses was found at least in the head thrust test, the caloric test, EVAR and VEMP; hypoexcitability when responses to canal and/or otolith tests were smaller in amplitude than normal; hyperexcitability when responses to canal and/or otolith tests were larger in amplitude than normal; mixed when some signs of hypoexcitability were found in some tests and hyperexcitability in others. To evaluate the prevalence of vestibular impairment in our implant candidate group and follow-up group, we compare the values obtained in all tests
Our data determined the prevalence of vestibular impairment before cochlear implant in a SNHL population (n = 224). The various parameters collected for each test before and after implant are plotted in Fig. 1A—F, which identify the patients out of the normal range. Each follow-up patient outside the normal ranges was identified for further analyses. Tables 2 and 3 provide normative values of the parameters of the VOR obtained for EVAR and OVAR in the implant candidate and follow-up populations (Fig. 2).
3.1. High prevalence of vestibular impairments in SNHL pediatric patients Surprisingly only half of the patients had normal symmetrical vestibular function. That is 51% had abnormal canal responses and 45% had abnormal otolith responses (Table 4). Vestibular function was asymmetric for canal function in 20% and for otolith function 19%. These data were used to select the side with weaker vestibular function for the implant. The presence of a vestibular malformation does not help to predict functional vestibular impairment. Of the 17 children with vestibular malformation in the CT scan, 47% had normal vestibular function, 35% a unilateral areflexia and none a bilateral areflexia prior to the implantation. No correlation was found between the cause of hearing loss and the vestibular dysfunction; however we did not have in our series any cases of Usher syndrome that is known to associate bilateral SNHL and bilateral vestibular loss. The follow-up group (n = 89) was similar to remaining members of the implant candidate group (n = 135); with respect to incidences of various vestibular impairments prior to surgery on the global canal and otolith evaluation (for canal: p = 0.240, for otolith p = 0.310).
3.2. Severe vestibular modifications following cochlear implant The observation of canal VOR on the 89 patients after unilateral implants (Fig. 3) shows no modifica-
212
E. Jacot et al.
Fig. 1 These figures are comparing values before and after implant for Bicaloric test, EVAR and OVAR. The shaded area corresponds to the normal values. Black dots indicate patients whose values are outside the normal range before, after
Vestibular impairments pre- and post-cochlear implant in children
Fig. 2 Vestibular status of follow-up group (n = 89 patients) with SNHL before implant.
tions in 43 of the 71 cases with prior vestibular function. A complete ipsilateral areflexia (absence of canal VOR at all tests) appeared in seven of the patients implanted on a functional vestibule. A statistically significant decrease of the canal VOR was observed in 12 cases. An increase of canal VOR (a sign of canal hyperexcitability) on the implanted ear was found in 5 cases. In three cases the cochlear implant was associated with either a decrease or an increase of the canal VOR leading to symmetric responses. In only one case we did observe a decrease of the canal VOR on the contralateral side. At the time of the test this patient presented an ongoing CMV infection. After implant the otolith responses in the followup group (Fig. 4) were modified in 39/71 (55%) of the cases with prior vestibular function. A statistically significant decrease of the otolith VOR and vestibulospinal responses was observed in 9 cases. A bilateral decrease of the responses was found in 4 cases in the OVAR test while VEMP remained present on both sides in these four cases. In 17 cases the cochlear implant was associated with an increase of the otolith VOR in the OVAR test with an intense directional preponderance toward the implanted side (a sign of hyperexcitability). In four cases the modifications were variable from one test to another (equalization of responses, mixed modifications).
213
Fig. 3 Distribution of the canal VOR modifications after cochlear implant (n = 89 patients).
Fig. 4 Repartition of the modifications of otolith responses after cochlear implant (89 patients).
Of the 89 patients implanted, 24 presented symptoms (like vomiting or dizziness) that could evoke vestibular impairment, mostly during the first 24— 48 h after surgery. Only 14 of them showed changes in vestibular function in the tests done during the following days. However, in the seven patients who showed complete ipsilateral vestibular areflexia, 6 presented strong initial symptoms. In our series the longest delay for vestibular areflexia to occur was 1.5 months after surgery but milder secondary aggravation of vestibular impairment occurred as late as four years after implant (observed in a case with several tests on a long follow-up). The profound vestibular loss occurred immediately after surgery for five patients, after 4 days for one and 1.5 months for the remaining one. The latter developed severe instability and vertigo one week after changing
cochlear implant or both while white dots indicate patients with normal responses. Values located on or close to the diagonal dotted line correspond to patients with no modifications of their vestibular score between the vestibular assessments made before and after cochlear implant. For bicaloric test we consider mean relative valence (A) and directional preponderance (B) calculated with Jonkee’s formula. For EVAR test we consider mean relative initial maximal velocity (C) and mean relative time constant (D) calculated as follow: ((value clockwise value counterclockwise)/ (value clockwise + value counterclockwise)) 100. For OVAR test we consider relative directional preponderance (E and F) given by the bias of the horizontal and vertical components of the responses and calculated as follow: ((bias clockwise + bias counterclockwise)/2). Horizontal and vertical modulations of VOR for OVAR was not considered because values were too scattered to be of any interest and therefore not shown.
0.70 0.51 (27.9 20.4) 0.58 0.47 (23.4 18.7) 0.72 0.52 (28.7 21.0) 0.60 0.49 (23.9 19.5) 7.5 6.1 6.0 4.7 7.1 5.9 7.1 5.5
Before implant (n = 89) After implant (n = 89)
6.4 4.5 7.3 5.2 6.9 4.9
7.2 5.3
0.77 0.52 (31.2 20.9) 0.67 0.47 (26.7 18.8)
0.75 0.54 (29.9 21.5) 0.60 0.46 (23.9 18.6)
Post-rotatory (CW rotation) Per-rotatory (CCW rotation) Post-rotatory (CCW rotation) Per-rotatory (CW rotation) Per-rotatory (CCW rotation) Post-rotatory (CCW rotation) Per-rotatory (CW rotation)
Post-rotatory (CW rotation)
Non-implanted side Implanted side
Implanted side
Non-implanted side
E. Jacot et al. Table 2 Vestibulo-ocular responses (VOR) for earth vertical axis rotation (EVAR) for the follow-up group before and after implants: mean values and standard deviation are listed. Data are presented such as all the patients were implanted on the same side (right). CW and CCW indicate the direction clockwise or counter-clockwise of rotation. Populations VOR mean time constant at EVAR (s) VOR slow phase at EVAR: mean gain/velocity ˚(/s)
214
the stimulation program of the implant. The areflexia did not recover after changing back to the previous program. In our series, no predictive factors for vestibular impairment following cochlear implant were found. The occurrence of areflexia or hypoexcitability was not correlated to age or the type of implant or surgeon’s identity. Of the 7 patients who developed areflexia three received an Advanced Bionics implant and four a Cochlear implant. None of the SNHL etiologies mentioned in Table 1 are correlated to a higher risk of vestibular impairment after implant. Of the 7 patients who developed areflexia, two presented a Mondini, and one presented a CMV infection.
3.3. Long-term follow-up of vestibular modifications following cochlear implant In the follow-up group, 27 were tested further at post-operative delays from 3 months to 7 years (mean: 26 19 months, median: 18 months). Long-term vestibular modifications after implant are summarized in Table 5. Over the long term, 63% of the children show stable responses to the vestibular tests whatever results were found at first assessment after surgery. One patient with no modification after surgery showed a decrease of vestibular responses on the implanted side nine and half months later. Partial recovery of vestibular responses is observed in 18.5% of the cases. A complete recovery was observed in only one case with hypoexcitability. In 3 cases (11%) the decreased vestibular responses deteriorated over time. The five patients with high directional preponderance toward the implanted ear (sign of hyperexcitability) recovered in 2 cases while in the 3 others the hyperexcitability persisted over a 4—7 year period. Complete areflexia remained unchanged in 4 cases but surprisingly partially recovered in two cases as late as 1.5 and 3.5 years after implant. The finding of a hypoexcitability at first vestibular assessment was followed in 2 cases by a further decrease of the vestibular responses and in 2 cases by a partial or complete recovery.
4. Discussion This is the first report on the prevalence of canal and otolith vestibular impairment in a large pediatric population with profound SNHL, complementing the study of Kaga and co-workers on 20 children [9]. In the implant candidate population only 50% had normal vestibular function, while 20% had complete
Vestibular impairments pre- and post-cochlear implant in children
215
Table 3 Vestibulo-ocular responses (VOR) for off vertical axis rotation (OVAR) for the follow-up group before and after implant: mean values and standard deviation are listed. Data are presented such as all the patients were implanted on the same side (right). CW and CCW indicate the direction clockwise or counter-clockwise of rotation. Populations Bias ˚(/s) Moduation ˚(/s)
Before implant (n = 89) After implant (n = 89)
Implanted side (CW rotation)
Non-implanted side (CCW rotation)
Implanted side (CW rotation)
Non-implanted side (CCW rotation)
Horizontal
Vertical
Horizontal
Vertical
Horizontal
Vertical
Horizontal
Vertical
0.33 1.8
0.53 4.7
0.13 2.3
0.39 4.5
3.3 2.3
4.7 4.3
3.3 2.4
5.1 4.2
0.4 3.1
0.2 4.6
0.3 2.1
0.2 3.3
2.6 2.2
3.8 4.3
2.6 2.1
5.2 5.5
Table 4 Canal and otolith status in the candidate for implant group of 224 patients versus the follow-up group of 89 patients. Bilateral symetrical hyporeflexia
Bilateral asymetrical hyporeflexia
Unilateral hyporeflexia
Mann— Whitney test
Population
Normal
Bilateral areflexia
Unilateral areflexia
Canal responses (n = 224—89) Canal responses (n = 89)
43% 60%
30% 18%
8% 6%
7% 4%
4% 3%
6% 9%
p = 0.240
Canal responses (n = 224)
49%
25%
8%
6%
4%
8%
—
Otolith responses (n = 224—89) Otolith responses (n = 89)
50% 62%
18% 13%
5% 1%
12% 10%
5% 5%
10% 9%
p = 0.310
Otolith responses (n = 224)
55%
16%
4%
10%
5%
10%
—
Table 5 Observation of vestibular modifications after implant over a long-term period (n = 27). Vestibular modifications at first assessment (number of patients)
Long-term modifications of the vestibular responses
No modifications (n = 7) Areflexia (n = 6) Hypoexcitability (n = 8) Hyperexcitability (n = 5) Mixed (n = 1)
6 4 4 3 1
Persistence (number of patients)
Complete recovery (number of patients)
Partial recovery (number of patients)
1
2 1 2
Worsening (number of patients) 1
bilateral areflexia, 7.5% bilateral symmetrical impairment and 22.5% partial asymmetrical impairment. Our data are inconsistent with some published results [3,9] likely because of the smaller samples in these studies. Our data show no correlation between vestibular and hearing status, thus it is impossible to predict vestibular impairment type or severity from hearing loss characteristics. Our observation shows also that it is not possible to predict on clinical cues the integrity of vestibular system in young children. In the follow-up population, cochlear implants had dramatic effects on vestibular canal and otolith function. This has been previously underestimated in children [15]. The risk of complete areflexia due to the implant is evaluated here at 10%. Most of the other vestibular modifications described in our ser-
2
ies were clinically silent and could only be found through systematic testing. This absence of clinical signs could be explained by the well-known rapid compensation of sensory deficits in children. With a complete canal evaluation (bithermal caloric test, Halmagyi test and EVAR test) we found canal paresis in 30% of children after cochlear implant. A recent study on vestibular disturbances after cochlear implants in adults (on a cohort of 146 patients) shows that 32% showed vestibular disturbances and weaker responses to bithermal caloric test after surgery [16]. In another study of adults receiving cochlear implants where canal function was measured with the head impulse test with eye coil recording, 10% (1/11) had canal functional loss after implant [17], corresponding to our results in children. These results support the idea that the
216 canal responses are as vulnerable to cochlear implant in children as they are in adults regardless of age or cause of deafness. Mechanism of implant provoked vestibular deficits is not yet completely known. Nevertheless the literature offers several explanations: an inflammatory process, inner ear liquid imbalance or microfractures [3,18]. Our data offer possible explanations. In most cases the direct traumatic effect of the electrode during its introduction in the cochlea can be implicated when the vestibular loss is immediately diagnosed after surgery. In one case a 4 days delay before the onset of vestibular impairment argues against such a mechanism and rather supports the hypothesis of a leak of inner ear fluids. One patient presented vestibular impairment following the switch from one stimulation program to another with higher stimulation level. The direct effect of electrical stimulation could be the source of the vestibular loss in this case. The absence of recovery after switching back to the previous stimulation program suggests a permanent lesion of vestibular receptors. In addition our observation in several cases of persistent hyperexcitability of canal and otolith systems suggests a direct effect of the electrical stimulation of vestibular receptors through the cochlear implant, as already suggested by several authors in adults [19]. Moreover we observed in some cases with no vestibular responses a revival of the responsiveness after implant. After an initial impairment, the partial or complete recovery of the vestibular responsiveness with time show that the vestibule can recover. The majority of the various vestibular modifications following the implant seem to be quite stable. On the 27 patients whom had a long-term followup, 16 patients (59%) showed unmodified vestibular responses through time. Two of the 7 patients with areflexia recovered partially their vestibular responses. One patient with hypoexcitability showed a partial recovery of its vestibular responses, one a complete recuperation and three showed aggravation of their vestibular impairment. Despite a small number of patients with very long follow-up, our data suggest that even severe vestibular impairment is not necessarily permanent and can evolve over a long period.
4.1. Suggestions for vestibular management of SNHL cochlear implant candidates It is important to select the side for a cochlear implant as the least functional vestibule in order to limit the occurrence of bilateral vestibular loss in case there may only be one functional vestibule,
E. Jacot et al. except in the cases where audiological or anatomical criteria are of main concern. Bilateral vestibular loss is known to have particularly dramatic consequences for very young children (as well as in older people) [4—6]. This choice can be only done with the help of a vestibular assessment of canal and otolith functions. In our series, 22.5% of our patients were implanted in the less functional vestibule except in 5 cases where audiological or anatomical criteria were of main concern. From our experience the tests that were most sensitive for detecting vestibular impairment were the head thrust test and bicaloric test for canal assessment and VEMP for otolith assessment. These tests do not require very expensive or sophisticated equipment and can be developed in any pediatric cochlear implant unit. Bicaloric and VEMP tests evaluate peripheral vestibular pathways and therefore are not submitted to central compensation processes as much as EVAR and OVAR. The time constant in EVAR can detect a vestibular deficit when it is recent. However, with time, the initial asymmetry disappears progressively to be substituted by symmetrical short time constants on both sides. As for the EVAR test, the results with OVAR were only relevant in cases of recent and complete vestibular deficit (asymmetry of the modulation and directional preponderance toward the opposite side); Furthermore it detected hyperexcitability of the otolith system on the implanted side by an intense and durable bias with a directional preponderance toward the implanted side. We define the three months after implant as a high-risk period for the appearance of vestibular impairments. At the end of this period we propose programming regular vestibular assessments in order to detect eventual complications. Moreover in vestibular impaired patients even a longer followup can be foreseen. Such testing will permit parents to be more completely informed of the risk of vestibular impairment should the implant be done on a functional vestibule (80% of our patients). The incidence of inducing vestibular areflexia during cochlear implant on a functional vestibule is 10% and this consequence is unpredictable. This risk must be taken into account when a patient presents already a unilateral areflexia (5.5% in our population). Furthermore this non-negligible risk of permanent vestibular dysfunction would argue against bilateral cochlear implantation in a single surgical procedure without previous vestibular assessment. An exception to this dictum should be made when the deafness is secondary to bacterial meningitis (in particular Pneumococcus sp.) that is followed by a rapidly developing endolabyrinthitis ossificans pre-
Vestibular impairments pre- and post-cochlear implant in children venting future cochlear implant or when preimplant vestibular testing shows that complete bilateral vestibular areflexia is already present (such as Usher syndrome).
5. Conclusions Pediatric SNHL cochlear implant candidates have a high incidence (50%) of vestibular impairments. Cochlear implants affect vestibular function in a high proportion of cases (50%) with 10% of vestibular loss. Vestibular assessment protocols should be instated for pediatric SNHL candidates for implant. This would include at least a clinical vestibular examination including the head thrust test, a bicaloric test and VEMP to decide which ear to implant (the most poorly functioning vestibule). Parents should be informed of the risk of vestibular loss and possible temporary imbalance during a 3 months period after surgery in case of an implant on a functional vestibule. The interest of a systematic vestibular assessment after implant should be discussed as well as the necessity of a vestibular function follow-up particularly when vestibular implant effects are detected. Our results support the necessary development of specific programs for vestibular assessment in pediatric cochlear implant units.
Acknowledgements This work has been supported by: Fondation Genevoise de Bienfaisance Valeria Rossi di Montelera, Fondation de France, Fondation Cotrel. Thanks to Franc¸oise Toupet for her collaboration during clinical testing, Pierre Denise for his scientific and technical support, Sidney Wiener for his helpful comments on the manuscript.
References [1] D. Vibert, R. Hausler, M. Kompis, M. Vischer, Vestibular function in patients with cochlear implantation, Acta Otolaryngol. 545 (Suppl.) (2001) 29—34. [2] R.H. Brey, G.W. Facer, M.B. Trine, S.G. Lynn, A.M. Peterson, V.J. Suman, Vestibular effects associated with a multiple channel cochlear prosthesis, Am. J. Otol. 16 (16) (1995) 424—430.
217
[3] C.A. Buchman, J. Joy, A. Hodges, F.F. Telschi, T.J. Balkany, Vestibular effects of cochlear implantation, Laryngoscope 114 (October (Suppl. 103)) (2004) 1—22. [4] S.R. Wiener-Vacher, F. Toupet, M. Franc¸ois, T. Van Den Abbeele, P. Viala, P. Narcy, Conse ´quences des de ´ficits vestibulaires sur le de ˆle postural chez ´veloppement du contro l’enfant. Flammarion Me ´decine Sciences, Journe ´es Parisiennes de Pe ´diatrie (1998) 283—287. [5] K. Kaga, Vestibular compensation in infants and children with congenital and acquired vestibular loss in both ears, Int. J. Ped. Otorhinolaryngol. 49 (1999) 21524. [6] S.R. Wiener-Vacher, L. Amanou, P. Denise, P. Narcy, Y. Manach, Vestibular function in children with CHARGE association, Arch Otolaryngol. Head Neck Surg. 125 (March (3)) (1999) 342—347. [7] T.D. Fife, R.J. Tusa, J.M. Furman, D.S. Zee, E. Frohman, et al., Assessment: vestibular testing techniques in adults and children, Neurology 55 (2000) 1431—1441. [8] T.A. Kelsch, L.A. Schaefer, C.R. Esquivel, Vestibular evoked myogenic potentials in young children: test parameters and normative data, Laryngoscope 116 (June (6)) (2006) 895— 900. [9] Y. Shinjo, Y. Jin, K. Kaga, Assessment of vestibular function of infants and children with congenital and acquired deafness using the ice-water caloric test, rotational chair test and vestibular-evoked myogenic potential recording, Acta Otolaryngol. 127 (July (7)) (2007) 736—747. [10] G.M. Halmagyi, Diagnosis and management of vertigo, Clin. Med. 5 (March—April (2)) (2005) 159—165. [11] S.R. Wiener-Vacher, Vertigo in children, EMC-Oto-rhino-laryngologie 2 (2005) 230—248. [12] S.R. Wiener-Vacher, F. Toupet, P. Narcy, Canal and otolith vestibulo-ocular reflexes to vertical and off vertical axis rotations in children learning to walk, Acta Otolaryngol. 116 (September (5)) (1996) 657—665. [13] M.S. Welgampola, J.G. Colbatch, Characteristics and clinical applications of vestibular-evoked-myogenic potentials, Neurology 64 (May (10)) (2005) 1682—1688. [14] S.R. Wiener-Vacher, K. Mazda, Asymmetric otolith vestibuloocular responses in children with idiopathic scoliosis, J. Pediatrics 132 (6) (1998) 1028—1032. [15] D. Bouccara, M.J. Este `ve-Fraysse, N. Loundon, B. Fraysse, N. Garabedian, O. Sterkers, Vestibular dysfunction after cochlear implantation: a national multicenter clinical study, Rev. Laryngol. Otol. Rhinol. (Bord.) 126 (4) (2005) 275—278. [16] J.C. Enticott, S. Tari, S.M. Koh, R.C. Dowell, S.J. O’Leary, Cochlear implant and vestibular function, Otol. Neurotol. 27 (6) (2006) 824—830. [17] A.A. Migliaccio, C.C. Della Santina, J.P. Carey, J.K. Niparko, L.B. Minor, The vestibule-ocular reflex reponse to head impulses rarely decreases after cochlear implant, Otol. Neurotol. 26 (4) (2005) 655—660. [18] O. Handzel, B.J. Burgess, J.B. Nadol Jr., Histopathology of the peripheral vestibular system after cochlear implantation in the human, Otol. Neurotol. 27 (1) (2006) 57—64. [19] L. Sennaroglu, B. Gursel, G. Sennaroglu, E. Yucel, I. Saatci, Vestibular stimulation after cochlear implantation in common cavity deformity, Otolaryngol. Head Neck Surg. 125 (4) (2001) 408—410.
Available online at www.sciencedirect.com