Mean (SE) timed score (duration in seconds) for children with cochlear implants standing on 1 foot over 4 separate trials: (1) on balance beam (diamonds) (2) off balance beam (squares), (3) with eyes open, and (4) with eyes closed. Cochlear implants were on and activated during all trials.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Cushing SL, Chia R, James AL, Papsin BC, Gordon KA. A Test of Static and Dynamic Balance Function in Children With Cochlear ImplantsThe Vestibular Olympics. Arch Otolaryngol Head Neck Surg. 2008;134(1):34–38. doi:10.1001/archoto.2007.16
Copyright 2008 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2008
To determine the incidence of static and dynamic balance dysfunction in a group of children with profound sensorineural hearing loss receiving a cochlear implant and to assess the impact of cochlear implant activation on equilibrium.
Observational cross-sectional study of children with single-sided implants, tested under 2 conditions: (1) implant on and (2) implant off in a random order.
Ambulatory setting within an academic, tertiary care children's hospital.
Forty-one children (ages 4-17 years) with cochlear implants comprised the study group. Fourteen children with normal hearing served as controls.
All participants performed a standardized test of static and dynamic balance function (Bruininks-Oseretsky Test of Motor Proficiency 2 [BOT2], balance subset). Children with implants performed the BOT2
under the 2 randomized conditions.
Main Outcome Measures
Overall performance on the balance subset of the BOT2 and the influence of implant activation on performance.
The mean (SD) age-adjusted scale score for our control group was 17 (5) points (95% confidence interval [CI], 14-20), which was not significantly different (P = .15)
from the published age-adjusted mean for the BOT2 balance subset (15
 points). The group that had undergone implantation, however, performed significantly more poorly (12 [ 6] points; 95% CI, 10-14)
than either the control group or the published test mean (P = .004). Children with implants performed better with their implants on than with their implants off (mean [SD]
difference, 1.3 [2.7] points; 95% CI, 0.3-2.3; P = .01).
Large differences exist in the balance ability of children with sensorineural hearing loss requiring cochlear implantation compared with age-matched controls. Implant activation, however, conferred a slight advantage in accomplishing balance-related tasks. These results substantiate the need to further quantify the baseline vestibular dysfunction of our study population of children with cochlear implants,
as well as the impact of implant activation on the input and output of the vestibular system.
It is plausible that lesions leading to sensorineural hearing loss could also contribute to dysfunction of the vestibular end organs.
We examined this possibility in the present study by examining balance function in children with profound sensorineural hearing loss who use a cochlear implant. There is also a potential for interaction between the electrical stimulation provided by the implant and balance and vestibular function. We investigated this potential by measuring balance with the implant on and off in a randomized order in these children.
A number of studies1-9 have sought to document baseline vestibular function prior to cochlear implantation. The consensus is that a considerable percentage of candidates for cochlear implantation show preoperative vestibular dysfunction.
These studies have focused primarily on the adult population and in general document a considerable degree of associated vestibular loss (range, 25%-100% hypofunction). Fewer studies have included young children. To date, the largest cohort of children assessed in this fashion was reported by Buchman et al6 in 2004. Testing 22 children aged 2 to 16 years with bithermal caloric irrigation, rotational chair testing, and computerized dynamic posturography,
they found that 68% of the children had vestibular hypofunction or areflexia prior to undergoing cochlear implantation.
The same studies1-9 examined postoperative vestibular function in a similar fashion with the aim of determining the impact of implantation itself on vestibular function.
On a clinically positive note, reports of severe and/or permanent vestibular dysfunction are rare even in the setting of bilateral implantation.
This is particularly true in the setting of pediatric implantation,
with only 7 of 575 children at our institution (1.2%) experiencing notable but transient imbalance following cochlear implantation (S.L.C.,
A.L.J., B.C.P., and K.A.G., et al, unpublished data 2007). Buchman et al6 suggested that the reason for this may be the high degree of vestibular dysfunction in implant candidates;
in this case, only a small percentage of implant candidates would be at risk of a considerable vestibular loss following implantation.
There is also a generally accepted but untested notion that children are more able to adapt to vestibular injury than adults, which, if true, would further decrease the likelihood of a longstanding uncompensated vestibular defect in children undergoing cochlear implantation.
In addition to the potential for vestibular injury at the time of surgery, cochlear implantation also carries the risk that the electrical current provided by the implanted array could spread beyond the auditory portion of cranial nerve VIII and stimulate the vestibular portion.
Such activation could, on the one hand, be detrimental to vestibular function; there could, for example, be negative impacts of chronic electrical stimulation of the labyrinth or vestibular nerve fibers.
On the other hand, however, electrical stimulation of the vestibular system could theoretically provide some usable vestibular cues. Reports of vestibular sensations elicited by implant use are present, although infrequent.10-12 There are also reports of improved vestibular function postoperatively, especially in balance assessment using computerized dynamic posturography and particularly in the setting in which the implant is on and activated in noise.3,4,6
These clinical tests provide us with a partial but incomplete assessment of the function of the vestibular system. Balance equilibrium is maintained by a complex interplay among the visual, somatosensory,
vestibular, and motor systems, and it is not uncommon to be faced with patients who are disabled by imbalance despite normal results from vestibular test battery. It was with this in mind that, for the purpose of the present study, we adopted a simple and standardized test of static and dynamic balance function in our assessment of children with cochlear implants.
The Bruininks-Oseretsky Test of Motor Proficiency (BOT2) is a clinical test battery, first introduced in 1978, that has become the most widely used standardized measure of motor proficiency.13 It is most commonly employed within the realm of physical and occupational therapy and developmental psychology.
It comprises a number of subtests, one of which is designed to assess overall balance function (the BOT2 balance subset). We applied the BOT2 to our cochlear implant population in the hope of addressing and answering the following questions: (1) Do children with profound sensorineural hearing loss receiving cochlear implantation perform differently on a test of static and dynamic balance function (BOT2)
compared with their peers with normal hearing? (2) Do children with cochlear implants perform differently on a test of static and dynamic balance function when they have their implant on vs off?
The study protocol was submitted, reviewed, and approved by the research ethics board at the Hospital for Sick Children, Toronto,
Ontario. Written consent and verbal assent were obtained from the guardians and children, respectively.
Children with cochlear implants (hereinafter, the implant group)
underwent a standardized test of static and dynamic balance function (the BOT2 balance subset). This was an observational cross-sectional study in which children with single-sided implants were tested under 2 conditions, (1) implant on and (2) implant off, in a random order.
Although the test provides age-matched normative scores for the BOT2
balance subset, a group of children with normal hearing were tested as controls to estimate the uniformity and replication of the testing environment. Children with coexisting uncompensated visual loss or motor deficits were excluded from the study.
The BOT2 balance subset includes 9 separate tasks, of which 4 are performed with eyes open, then eyes closed (Table 1). The raw score is converted into a point score, and the point scores of the 9 separate tasks are summed to produce a total point score (range, 0-37 points). The total point score and the subject's age at the time of testing are then used to look up an age-matched scale score based on the population norms data provided by the test (range, 1-35 points).
The test was administered by 1 of 3 individuals (S.L.C., R.C.,
and K.A.G.) in 2 settings. Approximately half the implant group (n = 22)
and all the controls were recruited at our annual cochlear implant picnic (total, 36 individuals). The remaining subjects were tested in our clinic (n = 19).
An independent t test was used to determine the significance of differences between the implant group (in the implant-on condition) and both our own control group and the normative population mean provided with the test. A paired t test was used to determine the significance of differences within the implant group under the 2 conditions (implant on vs off). Data obtained in each of the subtests were assessed using a repeated-measures analysis of variance (ANOVA) to determine the influence of vision and the presence of a balance beam on the balance function of our control and implant groups. A similar analysis was also performed to determine the relative influence of vision, the presence of a balance beam, and the implant status (on vs off) on the balance function within our implant group.
We used SPSS (version 14.0; SPSS Inc, Chicago, Illinois) and SAS (version 9.1; SAS Inc, Cary, North Carolina) software for the statistical analyses.
Thirty-two children with single-sided implants were tested under 2 randomized conditions: (1) implant on and (2) implant off. An additional 9 children completed the BOT2 with their implant on only (total in implant group, 41). Fourteen children without implants were tested to serve as controls. Demographics of the study group are presented in Table 2. There was a roughly equal sex distribution within the control group (8 of 14 were female [57%]) and a slight male predominance within the implant group (16
of 41 were female [40%]). The mean age of the children in the control group was 8 years vs 10 years for children in the implant group. This difference in age across groups was considered significant (t42.8 = 2.03; P = .05; unequal variances assumed, Pr > F = 0.02, where Pr indicates probability).
The duration of implant use varied from 2.5 to 12.2 years (mean [SD], 4.8 [2.8]).
The mean (SD) age at implantation was 4.7 (3.6) years. All children were implanted unilaterally with a Nucleus 22, 24M, 24RCS, 24CA, 24RE,
or Freedom device (Cochlear Corp, Melbourne, Australia).
The published age-adjusted mean (SD) score (n = 1520
children who participated in the group that established the normative means for the test) for the BOT2 balance subset is 15 (5) points (range,
1-35 points). In comparison, the mean score of our control group of children with normal hearing (n = 14) was 17 (5) points.
There was no statistically significant difference between our control group and the published normative data (t13.25 = 1.52; P = .15;
unequal variances assumed Pr > F <0.05).
In comparison, the mean score for the implant group was 12 (6). The difference between the implant group and both the normative and control groups was statistically significant (t41.6 = 3.37; P = .002;
unequal variances assumed Pr > F <0.05;
t53 = 2.95; P = .005;
unequal variances assumed Pr > F = 0.55,
In a pair-wise comparison of the implant group (n = 32)
under the conditions of implant on vs off, subjects performed better with their implant on (mean [SD], 12  points) than with their implant off (10  points) (t31 = 2.68; P = .01).
The impact of age, age at implantation, duration of implant use, and etiology of hearing loss were examined using a linear regression model. On univariate analysis, only duration of implant use was a significant predictor of overall performance (F1,39 = 8.87; P = .005).
A repeated-measures ANOVA was used to assess the effects of vision (eyes open vs closed), balance beam (test with vs without a balance beam), and group (implant vs no implant) on BOT2 balance subtest scores. Age was considered to be a covariate. The results of this analysis are detailed in Table 3.
The clearest effect was for vision; all children performed significantly worse on an identical task with their eyes closed vs open (P = .003). However, the decrement in balance ability that occurred with eyes closed was similar for both the control and implant groups, as shown by the lack of significance of the interaction between eyes and group (P = .88).
No other factors had significant effects on balance subtest scores.
A repeated-measures ANOVA was also performed on data from the implant group only. The details of this analysis are shown in Table 4. We found that again visual status (eyes open vs closed) significantly affected balance scores (P = .007). In addition, the implant group had significantly greater difficulties balancing on the balance beam when their eyes were closed (P = .03).
The balancing skills of young children deteriorated more than those of older children, on vs off the balance beam (age * balance beam interaction [where the asterisk indicates the balance beam age interaction in the regression model]; P = .04).
Mean (SE) data for the BOT2 balance subset scores on vs off the balance beam with eyes opened vs closed are plotted in the Figure. The balance beam was particularly difficult for children with implants when their eyes were closed vs open (there was a significant interaction between visual status and presence of a balance beam [P = .03]
and age, visual status, and presence of a balance beam [P = .047]). Although there was not a significant effect of the balance beam (P = .095),
there was a trend toward poorer performance by children using cochlear implants on vs off the balance beam.
Our analysis demonstrates that children with cochlear implants performed poorly on our test of dynamic balance function compared with their peers with normal hearing but experienced some slight gains while wearing their cochlear implants. Like their hearing peers, children using cochlear implants performed better when they had their eyes open. The balancing skills of children using implants also deteriorated significantly when they were on the balance beam with eyes closed (P = .03).
The results obtained for the control group were comparable with the published standardized norms for the test. This is important given that the test was administered outside of its usual context, a picnic ground vs an indoor waiting area. In addition, approximately 50% of the controls were the siblings of the children with hearing impairments,
which presents the potential of introducing bias into the control group. Where possible, comparisons were made between the children with hearing impairments and the standardized norms, and the control group data were used as a comparison only for task-specific analysis for which norms were not available.
This difference between children using cochlear implants and children with normal hearing was obvious and subjectively noted by all 3 examiners during each of the testing sessions. Certainly, this subjective observation was often also echoed by parents. When the children we tested were asked about their recreational activities,
most were found to be able to ride bikes and skate; however, with more detailed inquiry, parents often stated that their children found such activities difficult (eg, bike riding required training wheels for a prolonged period of time). Poor performance on the BOT2 was,
however, not entirely uniform across the implant group. Several children in the implant group obtained scores at or above their age level.
The difference between the implant and control groups, as well as within the implant group, may relate to the underlying etiology of the hearing loss, but this was not specifically addressed in the present study. In this study design we tested children only postoperatively and were unable to determine the relative contributions of deafness and implantation on performance on the BOT2. This question would be better answered by examining in a paired fashion the performance of children on the BOT2 before and after implantation, a study that is currently under way.
In the implant group, the duration of implant use was a significant predictor of performance on the BOT2 (P = .005).
However, the relationship was such that a shorter duration of implant use portended better performance. This seems to be the consequence of a small and heterogeneous subset of children with a short duration of implant use who perform quite well on the BOT2. This finding should likely be interpreted with caution, and further study is required to validate the relationship between duration of implant use and performance on the BOT2. An increased sample size may also facilitate the identification of other relevant predictors of performance on the BOT2.
Visual cues seem to be equally important to the balancing abilities of both children with normal hearing and deaf children with cochlear implants. There were no significant interactions between vision and group (P = .88), suggesting little,
if any, increased reliance on visual cues in children using cochlear implants. This is counter to the thought that children with hearing loss might rely more on visual cues than their hearing peers both to communicate and to balance.
The interaction effects found between the balance beam and both age and vision suggested that children using implants found the tasks on the balance beam particularly difficult. The effect of balance beam alone did not reach statistical significance (P = .095),
but there was a trend toward deteriorating skills when on vs off the balance beam. It is possible that the generally poor performance of children using cochlear implants on the balance tests limited the degree to which further deteriorations when on the balance beam could be identified. Indeed, many children (n = 18 [44%]) scored at minimum on the balance beam tasks.
Children with implants did demonstrate a small but significant improvement in their balance performance when they were wearing their implant (P = .01). This much smaller but significantly different balance function with the implant on vs off was not as readily apparent to the testers on a subjective level (P = .01). However, anecdotally,
many parents reported that they felt their child's balance improved markedly following cochlear implantation. It is impossible to tell whether the perceived benefits of an active implant on balance extend beyond the contribution of auditory cues alone. We did not control for the background noise level, and both conditions for testing (a picnic and a clinic waiting room) are ones in which a certain degree of ambient noise would be expected. A future study might specifically address the role of auditory cues on the maintenance of equilibrium.
It is also possible that improvements in balance when the implant is worn are the result of electrical stimulation of the vestibular nerve. Electrical pulses delivered by a cochlear implant with the purpose of evoking an excitatory response within the auditory nerve may spread beyond their target, particularly at high levels of stimulation.
The facial and vestibular nerves are at greatest risk. We have recently shown using objective measures that up to 50% of children with cochlear implants are at risk of facial nerve stimulation.14 In light of these findings, activation of the vestibular nerve and/or the vestibular end organs is a very real possibility. There are numerous reports10-12 within the literature describing discrete activation of the vestibular system leading to nystagmus and vertigo. In contrast to this discrete activation,
perhaps it is a low level of background vestibular activation that is responsible for the differences described herein and in other studies noting improvement on computerized dynamic posturography with the implant activated and in noise.3,4,6
In conclusion, we have demonstrated that large differences exist in the dynamic balance ability of children with sensorineural hearing loss requiring cochlear implantation compared with age-matched controls with normal hearing. Use of the implant, however, conferred an advantage in accomplishing balance-related tasks. These results substantiate the need to further quantify the baseline vestibular dysfunction of our implant group as well as the impact of implant activation on the input and output of the vestibular system. This will advance understanding of the pathophysiologic characteristics of cochleovestibular loss,
as well as the potential for acquired vestibular injury in the setting of bilateral cochlear implantation.
Correspondence: Sharon L. Cushing,
MD, Department of Otolaryngology–Head and Neck Surgery, University of Toronto, Hospital for Sick Children, Sixth Floor, Elm Wing, 555
University Ave, Toronto, ON M5G 1X8, Canada (email@example.com).
Submitted for Publication: January 30, 2007; final revision received May 19, 2007; accepted June 13,
Author Contributions: Drs Cushing,
Papsin, James, and Gordon and Ms Chia had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Cushing, James, Papsin, and Gordon. Acquisition of data: Cushing, Chia, James, Papsin, and Gordon. Analysis and interpretation of data: Cushing, James,
Papsin, and Gordon. Drafting of the manuscript:
Cushing, James, and Gordon. Critical revision of the manuscript for important intellectual content: Cushing,
Chia, James, and Papsin. Statistical analysis:
Cushing and Gordon. Obtained funding: Cushing and Papsin. Administrative, technical, and material support: Cushing, Chia, and Gordon. Study supervision: Cushing, James, Papsin, and Gordon.
Financial Disclosure: None reported.
Funding/Support: Dr Cushing was supported by a fellowship grant from the Canadian Institute for Health Research and by the Chapnik, Freeman, Friedberg Clinician Scientist Fund for the duration of the study.
Previous Presentation: This article was presented at The American Society of Pediatric Otolaryngology 2007 Annual Meeting; April 27, 2007; San Diego, California.
Additional Information: This study was awarded the Charles Ferguson Clinical Research Award for best peer-reviewed paper by a resident at The American Society of Pediatric Otolaryngology 2007 Annual Meeting.
Additional Contributions: Patricia Fuller, Vicky Papaioannou, MSc, Aud(C), and the Cochlear Implant Team and audiologists at the Hospital for Sick Children assisted in patient recruitment.