Inclusion and exclusion criteria

Reviewed studies met the following eligibility criteria: Prospective interventional studies or cross-sectional studies with aided and unaided conditions, written in English, with human subjects 18 years or older, and hearing loss diagnosed with audiometric testing; hearing intervention by CI (unilateral and or bilateral); gait measured by objective measurements (e.g., speed, stride time/length variability, or by stopwatch); subjects were able to walk independently, with or without an assistive device. Exclusion criteria: Reviews, meta-analyses, commentaries, and opinion articles; study sample with neurological diseases, acute heart disease, and mental health disorders.

Search strategy

We systematically searched the following databases: PubMed, Scopus, and Web of Science, with a last search on January 8th, 2024. We used the following search terms for all databases: (deafness OR presbycusis OR hearing loss OR hearing dis * OR hearing impair * OR auditory impair * OR auditory dis*) AND (walking OR gait OR dyn * balance OR mobility OR postural control OR posture) AND (Cochlear implant * OR hearing aid * OR auditory rehab * OR hearing rehab*).

Article selection

All results of our search strategy were first screened for relevance by the first author (BR) in terms of title and abstract. The papers that were found relevant underwent full evaluation for eligibility by BR and RTS independently and independent results from the two investigators were compared. In one case, when one of the investigators was uncertain, two additional authors, HK and MA, reviewed the article, and all reviewers re-evaluated the full article so a consensus could be reached. Description of all the 2076 records can be found in S2.

Data extraction

We extracted the relevant data from each of the included studies, namely the study’s aim, sample size, design, population, inclusion and exclusion criteria, hearing status and relevant metrics, follow-up duration, gait tasks and performance, vestibular measurements, and finally the study’s results.

Quality assessment and bias

In this systematic review, we addressed missing data by implementing the following approach: (1) Data Extraction and Identification: During the data extraction process, we systematically identified and excluded articles with missing data that was essential for inclusion. (2) Reporting Transparency: We explicitly reported the missing data for each included study in our results sections. (3) Risk of Bias Assessment: Methodological quality, risk of bias, and study design were assessed in the current review by using the PEDro scale [48]. All of the included studies had a fair level of quality with a PEDro score of four to five [42–46,49,50], for an overview please refer to the descriptive summary of the seven included articles in Tables 1 and 2, and also for detailed review of the quality assessment of the included articles in S3. Finally, in the discussion section, we addressed the limitations imposed by missing data on our review’s conclusions, and in the recommendations for future research.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Descriptive summary of included studies.

https://doi.org/10.1371/journal.pone.0319322.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Measures and main results presented in included studies.

https://doi.org/10.1371/journal.pone.0319322.t002

A systematic review was carried out, following the principles and phases of the PRISMA model. A meta-analytical methodology was not appropriate due to the small number and heterogeneity of the selected articles.

Inclusion/exclusion criteria set by the included studies

All participants were older than 18 years of age. Studies included only first-time unilateral CI users [42,43,45,49] who had the ability to understand English [46], mobility screening and history of hearing aid use for over three months [46]. Two studies did not mention the criteria [46,50].

Regarding exclusion criteria, studies excluded people with cognitive or physical impairments [46,49], psychological and anesthesia barriers [42], and individuals who failed to complete all scheduled assessments or follow the instructions [49,50]. Two studies did not mention any exclusion criteria [44,45].

Sample size and basic characteristics

The sample size varied between studies, from one (out of three participants) [44] to 43 participants [43]. Overall, studies were balanced between sexes [43,44,46,49,50], except for one study that had more females than males in the control group [42]; one study did not describe the sex of participants [45]. The studies included CI users ranging in age from 19 to 84 years [42–44,46,49,50].

Hearing status and assessments

The participants’ hearing status varied in severity levels from normal hearing (for control groups) to severe hearing loss. One study used pure-tone audiometric testing before and after CI [49], one study used screening audiometric testing [46], and one study mentioned approximate hearing degree of hearing loss [44]. Specific auditory measures were not mentioned in four studies, but rather authors classified the degree of hearing loss as partial deafness, severe sensorineural hearing loss, and congenital hearing loss [42,43,45,50].

Study design

It is important to note that the included studies used different methodologies. One approach was a prospective study that involved both pre-intervention and post-intervention assessments. The other approach assessed gait under on-off conditions, meaning gait was evaluated once with the CI turned on and once with it turned off. Five studies conducted pre- and post-intervention assessments [42,43,45,49,50], two studies assessed gait with the implant in both on and off conditions [44,46], and one study employed both pre- and post-implant assessments along with on-off evaluations [50].

Timing of evaluation and follow-up

Pre-intervention and post-intervention assessments were conducted in five studies [42,43,45,49,50], of which four conducted a single post-CI evaluation [42,43,45,49], and one study conducted three post-CI evaluations [50]. The timing of follow-up varied between studies, both pre- and post- CI. Pre-CI baseline testing ranged from four months to one-day pre-CI [42,49]. One study had a range of one day to 320 days from baseline testing to CI [43], and one study did not mention the time of the pre-CI testing [50]. Similarly, inconsistency was seen in the timing of post-CI follow-up, with some studies conducted the tests for one to two months [45,49,50], and up to three months [42,46] post-CI. One study ranged from 31 to 363 days post-CI with a median of 44 days [42], one study did not mention the time of evaluation [44], and one study conducted assessments with the processor on and off on the same day [46].

Vestibular assessment

Three studies did not assess vestibular function [44–46]. The other four studies conducted various diagnostic and self-reported vestibular tests. Caloric tests were used in two studies [42,43], the Video Head Impulse Test (vHIT) was used in two studies [42,49], vestibular-Evoked Myogenic Potential (VEMP) was used in one study [42], one study used Head-Thrust (HT) [50], and one used a self-report with the Dizziness Handicap Inventory (DHI) questionnaire together with the Subjective Visual Vertical (SVV) diagnostic test [50].

Timing of vestibular assessment

Vestibular function was assessed only pre-operatively in one study and was found to be normal in all patients except two [49]. In another study, the vestibular assessment was conducted pre-and post-operatively, but the results were not reported [42]. Two studies did not demonstrate a change between pre-operative and post-operative scores [43,50]. While the timeline for the vestibular assessment post-CI was not always specified, in one study, the postoperative median time between CI and the second examination was 44 days and ranged from 31 to 363 days [43], and in another study vestibular assessments were conducted postoperatively at one week and one month following CI [50]. Caloric testing on the operated side was unchanged in 20 patients, better in three patients, and worse in one patient [43], no statistically significant difference was found between pre-operative and postoperative DHI, and SVV testing, and head thrust testing did not demonstrate changes in the vestibular function in any of the operated ears between pre-operative and post-operative assessments with the CI on or off [50], the vHIT was used only pre-operatively in [49].

Etiology

Participants in two studies had bilateral sensorineural hearing loss [45,46], while other studies did not mention whether participants had unilateral or bilateral hearing loss. The etiology of the hearing loss among most participants across studies was mostly idiopathic. Nevertheless, three studies mentioned additional etiology with a mixed sample: one mentioned syndromic hearing loss or infectious deafness [43], another study mentioned noise-induced, autoimmune diseases, sudden hearing loss, measles, ototoxic drug exposure, mitochondrial disease, and congenital hearing loss [50], and the third study declared a history of Meniere’s disease, history of congenital cytomegalovirus infection, and a history of benign positional vertigo with normal results on the caloric test [44].

Hearing Intervention

Five of the studies included participants who underwent unilateral CI [42,43,45,49,50], and two studies included bilateral CIs [44,46].

Gait assessment

All studies measured gait performance on various walking tasks. These tasks and associated gait performance metrics are detailed in Table 3. Overall, the following gait tasks were performed: The Time Up and GO (TUG), Walk Across (WA), Tandem Walk (TW), Functional Gait Assessment (FGA), Gait Stability Ratio (GSR), Center of Mass (CoM) determination, Mini BEST Test, and Functional Ambulation Performance (FAP). Overall, the following metrics were recorded: Gait speed, stride length, gait variability, step width, straightness, single support, double support, cadence, and stability of gait [42–44,46].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Gait assessment, description, instruments, and variables.

https://doi.org/10.1371/journal.pone.0319322.t003

Gait outcomes following CI

Gait outcomes following CI varied between improvement, worsening, and no change from baseline to follow-up. These findings varied by the outcome measure, the task tested, duration of follow-up, and on-off conditions.

Among the studies that reported improvement in gait: One study showed that a small subset of participants improved on outcome measures (gait velocity, stride length variability, swing time variability, and the double support phase) after more than three months post-implant activation, despite no difference between on-off conditions [46]. Another study found improved gait velocity (timing of evaluation not mentioned) when implants were turned on (compared to off) [44]. Another study showed improvement in GSR in 17 out of 21 patients with three months of follow-up. This study also reported statistically significant improvement in several other parameters, such as step time, single support phase walking speed, TW, speed, and CoM determination [42].

Among the studies that reported worsening or no change in gait, one study reported a significant worsening in walking tasks one month after CI, in patients that were younger than 60 at the time of CI without a balance pathology. Other participants showed no change [45]. Another study did not find a statistically significant difference in TUG performance one-month post-CI [50]. Similarly, five weeks post-CI, there was no significant difference in FGA scores [49]. Lastly, Kluenter et al. [43] found a significant decrease in the mean walking speed in the WA task, but no change in other parameters. While in the TW task, the mean step width improved, the mean speed did not increase after the CI (median time postoperatively was 44 days and ranged from 31 to 363 days) [43].

In addition, it is important to highlight the improved outcomes of gait based on the methodological approaches introduced above that utilized prospective design or on-off conditions. By utilizing prospective methods, one study [42] indicated an improvement in gait performance three months post-implant gait stability ratio, step time, single-support phase, tandem walk speed, and CoM. In contrast, studies with follow-ups averaging between 37 days and two months reported either a decline or no change in outcomes after CI (43, 45, 49). Studies using the on-off design reported mixed outcomes. Two studies reported no difference between the implant being turned on and off (46, 50), while another study noted improved gait velocity, when implants were activated compared to the off state [44].

Follow-up periods

When patients were evaluated between four to six weeks post-CI, they showed worse gait outcomes [45] or no improvement [43,49,50], possibly due to acute symptoms related to the surgical procedure and activation of the implant. However, when tested three months post-CI, patients showed improvement in some gait parameters, such as tandem walk speed, mean step width, and step time, but not in comfortable normal walking speed [42,44]. This is consistent with studies that evaluated static and dynamic balance outcomes for longer follow-up. These studies showed a consistent improvement in postural stability over a duration of one year [51] and two years post-CI [52]. Additional studies with longer follow-up (e.g., one year) are required for assessing gait and any associated neuroplastic brain changes associated with CI [32].

Assessment of underlying behavioral and neural mechanisms

Gait outcomes changed between pre- to post-CI conditions, yet none of the studies in this review looked into possible underlying mechanisms (sensory, physiological, cognitive, behavioral), which limited our understanding of the mechanisms underlying the association between hearing and gait and how CI affects gait.

The studies in this review did not conduct any neural monitoring to assess neural activity and evaluate possible neuroplastic changes after the CI. Indeed, monitoring neural activity may demonstrate neural plasticity and compensatory activity in line with the Scaffolding Theory of Aging and Cognition (STAC) that suggests there may be qualitative differences in neural activation and behavioral strategies [53], or show changes in neural patterns when one domain becomes challenged [53]. In such cases, neural monitoring using real-time functional neuroimaging [53] tools can demonstrate the changes in neural activity patterns between pre- and post-CI.

Neural deprivation induced by reduced audio-sensory input can alter the central auditory system and modify the organization of the cortex through cross-modal re-organization [27]. This re-organization results in compensation by other neural systems to maximize performance; that is, using other brain structures or networks to increase the activity and compensatory expansion to other modalities [18]. Nevertheless, the decline is not permanent and can respond to changes in acoustic experience and cortical function can improve [28,29], through the use of hearing aids in mild-to-moderate hearing impairments, and may reverse some of the central changes related to severe-to-profound hearing loss by restoring afferent activity through CI [30]. Functional neuromonitoring contributes to the assessment and evaluation of neural activity and to identifying the brain systems responsible for different behaviors [54]. Thus, neural activity monitoring uniquely contributes to understanding neural activity during gait [55], compatible with CI devices, not subject to electrical artifacts and flexible in auditory stimuli paradigms [56].

Dual task of gait and cognition

In addition, and as introduced above, cognitive function can improve through the use of hearing aids in cases of mild-to-moderate hearing impairments. This change is often observed six months from the fitting to accommodate the neuroplastic changes [27,28]. None of the studies included in this review mentioned any cognitive evaluations. This evaluation is important because gait functioning is linked to cognitive performance [57,58], and cognitive demands increase as sensory input decreases [59], resulting in reduced walking speed, step length [60], and gait asymmetry [61].

The role of cognition in gait performance can be assessed with dual-task evaluations that combine cognitive tasks with walking after a CI [62]. Cognitive demands are likely to be higher when another cognitive task accompanies walking (i.e., dual tasking), also in challenged neural conditions, when walking becomes more cognitively demanding and reflective of real-life walking, leading to an increased risk of falling [57,63]. For individuals with hearing loss, conducting a dual-task leads to prioritizing the maintenance of posture over cognitive performance [19] by using alternative cognitive resources [18] resulting in changes in the neural system [64]. Hence, gait monitoring under dual-task conditions, i.e., walking with a cognitive task, after CI, should be of interest among researchers, yet these tasks were omitted in the studies included in this review.

Vestibular function post-CI

After reviewing the current data, it was found that there was a large variability in the choice and timeline of vestibular testing. The vestibular system and the cochlea are located in close proximity within the inner ear [2,16], and the literature, in general, is equivocal about the connection between vestibular function and CI. On one hand, vestibular function worsened after a CI in caloric testing as well as in VEMP, and no significant post-CI effect was found in HIT, posturography and DHI scores among individuals who underwent a CI [65]. On the other hand, others show improvement in performance on dynamic platform posturography [52]. Possible explanations for worsening in vestibular function, may include trauma to the cochlea due to the surgical procedure [65], infections due to anatomical proximity between the vestibular organs and the cochlea [2,16], and the timing of post-CI testing.

Consistency in measurements

This review revealed a lack of consistency in the tasks and outcomes used to assess gait performance. The seven included studies used seven different gait tasks, alone or in combination, to assess gait performance and no two studies’ methods were identical, with different measurement methods and devices. This inconsistency in the gait measurements makes it challenging to compare the results between studies. A more standardized approach to gait assessment using sensor-based gait evaluations could increase ecological assessment and diagnostic accuracy [66,67], and could lead to floor and ceiling effects [68]. Cognitive evaluations and dual-task testing are also needed to improve the comparability of findings and advance our understanding of the impact of CIs on gait performance among individuals with hearing loss.

Study design and implementation of on-off methods

This review highlights two key study designs: Prospective [42,43,45,49,50] and on-off designs, with hearing input is assessed with the CI turned on (on condition) and off (off condition) [44,46]. One of these studies [50] employed a mixed-method approach, utilizing both prospective and on-off designs. In the prospective approach, evaluations were conducted before and after implantation to determine whether CIs improve gait. The review identified variations in follow-up duration and outcomes among these studies, with longer follow-up periods (up to three months) showing gait improvements. In contrast, the on-off approach did not demonstrate gait improvement when comparing CI-on and CI-off conditions. However, this approach may enhance our understanding of the mechanisms underlying post-implant gait improvements, which are thought to arise from neuroplastic changes in response to increased audibility [29]. Combining these two approaches could provide a more comprehensive understanding of gait improvements induced by cochlear implantation and audibility enhancement. Nonetheless, longer durations of CI use may be necessary to observe significant and sustained outcomes.

Sample size

The sample size varied widely across the studies, with some having as few as one participant [44], and others including up to 51individuals [42]. This limited number of participants may not provide sufficient statistical power to detect meaningful changes in gait performance following CI. Additionally, a small sample size can lead to reduced generalizability of the findings. A larger participant pool would better represent the population of adults with hearing loss and improve the robustness of the conclusions drawn from the studies.

Hearing status

The systematic review identified the variability in the participants’ hearing statuses as a significant gap in the existing literature. The reviewed studies included individuals with various types of hearing loss, as well as normal-hearing control groups for comparison. Some studies did not provide specific auditory measures, while others mentioned participants having bilateral severe to profound sensorineural hearing loss or partial deafness. This variability in the participants’ hearing statuses makes it challenging to draw generalizable conclusions about the effect of CIs on gait performance among adults with hearing loss. A more standardized approach for reporting audiometric measures and separating unilateral from bilateral hearing loss would enhance the consistency and comparability of future studies in this area.

Limitations

This review is limited by the small number of studies identified (seven), the small sample size per study, and the ‘fair’ quality of the research as per the PEDro scale. In addition, the follow-up period in the studies was, in most cases, insufficient to accommodate the neuroplastic changes following CI. Furthermore, inconsistency was observed in gait tasks, hearing status, and vestibular testing. Given those constraints, a meta-analysis was not possible.

Our knowledge of the relationship between hearing and gait and the mechanisms underlying it is constrained by the absence of cognitive assessments and neural monitoring. The studies that were reviewed also had limitations despite meeting inclusion criteria. These limitations included a small number of participants, the absence of real-world daily activities in actual conditions, specific interventions such as single-sided CIs, and the lack of randomized controlled trials.

While not the topic of the current review, during our thorough search of available literature, we observed a significant gap in research regarding the impact of hearing aids on the gait of hearing aid users. We only came across one longitudinal study [10] that investigated the relationship between gait and hearing aids, and two studies that examined both CI and hearing aid users [44,46]. This limited number of studies precludes synthesis and conclusions.

Conclusion

To our knowledge, this is the first review to investigate the effect of CI on gait performance, as previous reviews have investigated the effect of hearing aids, or hearing aids and CIs, on static and dynamic balance.

We found that short follow-up (up to two months) after the activation of the CI was associated with unchanged or even worse gait quality, and a longer follow-up (at least three months) was associated with a partially improved gait such as in mean step width, step time, and Tandem Walk speed but not in comfortable, normal walking speed.

Recommendation for future research

Future studies should assess cognition and monitor neural activity during gait under dual-task conditions to further understand the underlying mechanisms between hearing and gait. In addition, randomized controlled trials with long-term follow-ups are needed to determine the effect of CI on gait function and identify the optimal follow-up periods. The implementation of on-off evaluations should also be considered to better understand the mechanistic relationship between hearing and gait. Combining prospective and on-off designs, with post-implant on-off evaluations conducted beyond three months after implantation, could reveal significant outcomes and provide deeper insights into the mechanisms underlying post-CI improvements.

Finally, the field will benefit from consistency [50] in the gait assessment protocol and outcome measures using challenging tasks with instrumented assessments [65].

Recommendation for clinical practice

Clinically, it is crucial to assess the vestibular and physical function, fall risk, and cognitive abilities of patients with hearing loss prior to undergoing a CI procedure. Also, efforts must be made to follow a strict protocol with tasks, baseline measures, and longer follow-ups. This will assist in personalizing the rehabilitation program after the implantation.