Secondary Logo

Journal Logo

REVIEW ARTICLES

Virtual Reality for Vestibular Rehabilitation: A Systematic Review

Xie, Michael; Zhou, Kelvin; Patro, Nivedh; Chan, Teffran; Levin, Marc; Gupta, Michael K.; Archibald, Jason

Author Information
doi: 10.1097/MAO.0000000000003155
  • Free

Abstract

Vertigo is the perception of self-motion or motion of the environment, and is typically described as a spinning sensation. It is a symptom often caused by an asymmetrical vestibular system, not an independent diagnosis. Etiologies can be separated into peripheral, which involves the labyrinth or vestibular nerve, or central, which involves the brainstem (1). Vertigo is debilitating and often involves medical consultation, sick leave, interruption of daily activities, increased incidence of falls, and avoidance of leaving the house (2,3). It is a prevalent symptom that affects 35% of adults older than 40 years old, and places a major burden on healthcare systems (2,4).

Vertigo is managed depending on its etiology and chronicity. Vestibular rehabilitation is an exercise-based treatment program designed to advance the natural process of vestibular compensation after acute vestibular disturbance or chronic imbalance (5,6). It is indicated for chronic non-progressive peripheral vestibulopathies with inadequate compensation. Broader indications include central, mixed, psychogenic, and age-related vertigo (5). It focuses on enhancing gaze stability through head movement or optokinetic exercises that induce vestibuloocular reflex adaptation. Secondary improvements are achieved through substitution by corrective saccades, enhanced smooth pursuit, and anticipatory central motor corrections. Postural stability improvements are achieved through substitution with somatosensory and visual input while limiting dependency on a single stimulus, correcting abnormal posture recovery strategies, as well as adapting and strengthening residual vestibular function. Tolerance of vertiginous symptoms is strengthened through habituation exercises. The ultimate goal is to improve quality of life and daily function (5). Vestibular rehabilitation promotes symptom reduction of both unilateral and bilateral peripheral vestibular hypofunction (2,7). Preliminary evidence also shows a benefit in central vestibular disorders (8,9). Although vestibular exercises are effective, there are barriers to accessing this treatment. Conventional vestibular rehabilitation involves frequent in-person appointments which are time-consuming and resource-intensive (5). Vestibular rehabilitation is not covered by the universal healthcare plan in Ontario, Canada. In some areas, accessibility depends on private insurance or patient payments. Fewer than 3% of eligible patients receive vestibular rehabilitation (10).

Recently, the advent of virtual reality technology has allowed for the creation of a simulated immersive environment during vestibular exercises. Virtual reality-based vestibular rehabilitation can replace or add to conventional vestibular rehabilitation. Additionally, virtual reality exercises can be more enjoyable and may improve adherence (11). If efficacious, virtual reality will allow patients to complete vestibular exercises independently, without relying on insurance coverage. With ongoing software optimization, cost reduction, and improved hardware portability, virtual reality offers the prospect of more readily accessible therapy. Although it is not currently widely adopted, virtual reality-based vestibular rehabilitation has the potential to eliminate the aforementioned barriers to conventional rehabilitation.

This systematic review aims to clarify the efficacy of virtual reality-based vestibular rehabilitation in patients with vertigo. Such information may provide further direction on whether and how virtual reality technology should be applied to vestibular rehabilitation.

MATERIALS AND METHODS

This review was completed in accordance with the Preferred Reporting items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

Eligibility Criteria

Studies comparing virtual reality-based to conventional vestibular rehabilitation in patients with vertigo were searched. In this review, “virtual reality” was defined as a simulated environment using a combination of hardware and software. “Vestibular rehabilitation” was defined as an exercise-based program aimed at improving balance and reducing vertigo. Only original research studies published in the English language in peer-reviewed journals were included. Randomized control trials (RCTs), non-randomized control trials, and uncontrolled intervention studies were included. Observational, case-control, case series, cross-sectional studies were searched for relevant references but were not included in the study itself. Unpublished abstracts, posters, opinions, case reports, reviews, letters to editors, and editorials were excluded.

Search Strategy

A comprehensive and systematic search of Ovid MEDLINE, EMBASE, and Alternative and Complementary Medicine databases were performed from each database's earliest inception to August 2020. Medical subject headings (MeSH) terms and keywords were used to ensure a broad search. The search included (vertigo [exploded, and keywords including vestibular disease/disorder, peripheral vestibulopathy, vertigo] OR vestibular rehabilitation [keywords including vestibular therapy, vestibular physiotherapy, vestibular rehab∗, vestibular exercise, vestibular therapy]) AND virtual reality (exploded, and keywords including augmented reality, virtual reality). The references of identified articles were searched to ensure comprehensiveness. Duplicate articles were removed. Titles and abstracts of all identified articles were screened independently by two authors (K.Z., M.X.). Disagreements were discussed until a consensus was reached or settled with the involvement of a third author (M.L.). Full texts of articles that fit the eligibility criteria were then independently reviewed by two authors (K.Z., M.X.).

Data Collection Process

Data extracted included study demographics, etiology of vertigo, type of virtual reality, and control interventions. Data regarding objective and patient-reported vestibular outcomes were also extracted.

Risk of Bias Analysis

The revised Cochrane risk-of-bias tool for randomized trials (RoB 2) and risk of bias in non-randomized studies of interventions (ROBINS-I) assessment tool were employed to analyze the quality of included studies (12,13). Three authors (K.Z., M.X., T.C.) independently performed study quality assessment and came to a consensus. Study quality assessors were not blinded.

RESULTS

The search of databases yielded 381 unique articles, and one additional article was identified by screening references of identified papers. After initial screening of titles and abstracts, 342 studies were excluded. The remaining 40 articles were reviewed in full, and 10 articles were ultimately included. Of the 30 excluded articles, 11 were not peer reviewed, six had patient populations outside the inclusion criteria, five did not apply virtual reality based on our definition, four had outcomes which were not related to symptom severity or rehabilitation, three had the wrong design, and one was not available in English. Of the five studies which did not apply virtual reality, two used a Nintendo Wii Balance Board to create an unstable surface, one used a see-through display to project an artificial horizon, one used an oscillating seated platform, and one projected moving texts. The flow diagram summarizing the study inclusion process is illustrated in Figure 1.

FIG. 1
FIG. 1:
Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram.

Randomized Controlled Trials

Two hundred forty one patients participated in six studies. In five studies with available data, the female:male ratio was 1.26 and the average age of all participants was 59.64 years old diagnosed with varying pathologies including vestibular hypofunction or Menière's disease. A summary of these studies can be found in Table 1.

TABLE 1 - Summary of randomized controlled trials
Author Participants Inclusion and Exclusion Criteria Location Study Groups Exercise Types Outcomes Conclusions
Micarelli et al., 2019 n = 47 with unilateral vestibular lossGender:Group 1 (7F 5M)Group 2 (6F 5M)Group 3 (7F 5M)Group 4 (6F 6M)Age:Group 1 (76.3 ± 5.5)Group 2 (76.9 ± 4.7)Group 3 (72.5 ± 3.6)Group 4 (74.3 ± 4.7) Inclusion: 25% vestibular response reduction on bithermal water caloric irrigation > 3 months after symptom onsetExclusion: Neurological, orthopedic, or physical impairments, noncompliant patients Rome, Italy Group 1/2 versus Group 3/4: VR + ViRe versus VRGroup 1/3 versus group 2/4: Elderly with MCI versus Elderly VR: Custom patient-tailored home-exercises BID (30–40 min/d). Supervised sessions with physiotherapist office visits twice a week (30–45 min). 4 weeks duration.ViRe: 3D Virtual Reality HMD Track Speed Racing 3D Game (20 min/d). 4 weeks duration. ObjectiveVideo HITStatic Posturography TestingSubjectiveDHIABC ScaleDGISSQ Significant improvement in otoneurological outcomes in ViRE + VR versus VR alone.
Micarelli et al., 2017 n = 47 with unilateral vestibular lossGender:Group 1 (9F 14M)Group 2 (11F 13M)Age:Group 1 (49.72 ± 10.34)Group 2 (50.48 ± 9.12) Inclusion: 25% vestibular response reduction on bithermal water caloric irrigation > 3 months after symptom onsetExclusion: Neurological, orthopedic, or physical impairments, noncompliant patients, diabetes, antidepressant medication Rome, Italy Group 1 (ViRe + VR) versus Group 2 (VR) VR: Custom patient-tailored home-exercises BID (30–40 min/d). Supervised sessions with physiotherapist office visits twice a week (30–45 min). 4 weeks duration.ViRe: 3D Virtual Reality HMD Track Speed Racing 3D Game (20 min/d). 4 weeks duration. ObjectiveVideo HITStatic Posturography TestingSubjectiveDHIABC ScaleZung Instrument for Anxiety DisordersDGISSQ Superior post-treatment maximization of ipsilesional VOR gain in ViRe + VR than VR alone.Superior post-treatment self-reported and performance measures in ViRe + VR than VR alone
Hsu et al., 2016 n = 70 with unilateral or bilateral Menière's diseaseGender:Group 1 (24F 10M)Group 2 (20F 16M)Age:Group 1 (66.5 ± 11.9)Group 2 (69.0 ± 12.6) Inclusion: Definite Menière's disease based on American Academy of OtolaryngologyExclusion: other vertigo etiologies, neurological/cognitive deficits, lower limb weakness/paralysis, or planned for vestibular ablation Taipei, Taiwan Group 1 (VR) versus Group 2 (ViRe) VR: Cawthorne-Cooksey Exercises (30 min/session) for six sessions. 4 weeks duration.ViRe: Virtual reality gaming tasks using a 3D HMD designed to emulate Cawthorne-Cooksey exercises (30 min/session) for six sessions. 4 weeks duration. ObjectiveCenter of Balance ScoresStatokinesigramMaximum mediolateral trajectoryMaximum anteroposterior trajectoryMean trajectory excursion in the mediolateralMean trajectory excursion in the anteroposteriorPerformance ScoreTask completion rate, time, success frequency Overall, all groups had significant improvements relative to pre-treatment scores.Control group had significantly higher total rehabilitation exercise scores in comparison to study group
Garcia et al., 2013 n = 44 with unilateral or bilateral Menière's diseaseGender:Group 1 (14F 7M)Group 2 (14F 9M)Age:Group 1 (47.90)Group 2 (47.65) Inclusion: definite Menière's disease based on American Academy of OtolaryngologyExclusion: Rheumatic, cardiovascular, visual, orthopedic, psychiatric, hematological comorbidities as well as those unable to comply with the program Sao Paulo, Brazil Group 1 (Diet + Betahistine) versus Group 2 (ViRe + Diet + Betahistine) ViRe: Posturography, body balance rehabilitation exercises, and postural training games using a projected environment and 3D glasses (45 mins twice a week) for 12 sessions. 6 weeks duration. ObjectiveFunctional vestibular examinationPosturographySubjectiveDHIADS All groups had significant post-treatment improvements in DHI scores and ADS scores.Group 2 had significantly smaller eyes-closed firm and compliant surface center of pressure.No statistically significant inter-group difference in DHI scores, dizziness analogue scale scores, or stability limit areas.
Pavlou et al., 2012 n = 16 with peripheral vestibular deficit after episode of vestibular neuritisGender:Group 1 (4F 7M)Group 2 (3F 2M)Group 3 (2F 3M)Age:Group 1 (40.70)Group 2 (86.00)Group 3 (37.60) InclusionAge 18-75Confirmed peripheral vestibualr deficit on bithermal caloric testing or ENG after diagnosis of vestibular neuritisExclusionOther vertigo etiologies, neurological disorders, systemic illness, or psychiatric disorders London, United Kingdom Group 1 = VR + Static ViReGroup 2 = VR + Dynamic ViReGroup 3 = VR + Static ViRe (from group 1) with subsequent dynamic ViRe VR: Cawthorne-Cooksey ExercisesStatic ViRe: 360 immersive environment in a projected city square using 3D glasses with no movement (45 min twice weekly) 4 weeks duration.Dynamic ViRe: 360 immersive environment in a projected city square using 3D glasses with dynamic moving characters (45 min twice weekly). 4 weeks duration. SubjectiveSVQBeck Depression InventoryBeck Anxiety InventoryFear QuestionnaireDGIVRCESS Significantly greater improvement in SVQ scores in Group 2 than Group 1.Significant higher VRCESS scores in patients completing Dynamic ViReSignificant improvement in Becks anxiety and depression scale only for group 1.No significant between and intergroup improvement in DGI
Virre et al., 2002 n = 13 with chronic vertigo symptomsGroup 1 (n = 9)Group 2 (n = 6) InclusionVertigo symptoms for > 6 months and decreased VOR gainExclusion\Active disease within the past 6 months San Diego, United States Group 1 (ViRe) vs. Group 2 (Control) ViRe: Active search tasks with scene magnification initially set at 5% higher than VOR gain and gradually increased over 10 sessions. (Two sessions/day) in a HMD. 5 days duration.Control: Active search tasks with scene magnification set at 100% (two sessions/day). 5 days duration. ObjectiveVOR gainSubjectiveDHI Test subjects showed increased VOR gains compared to control subjects.Improvement remained after 1 week but at a lower level than immediately after testing.
ABC indicates activities-specific balance confidence; ADS, analogue dizziness scale; DGI, dynamic gait index; DHI, dizziness handicap inventory; ENG, electronystagmography; F, female; HIT, head impulse test; HMD, head-mounted device; M, male; MCI, mild cognitive impairment; SSQ, simulator sickness questionnaire; SVQ, situational vertigo questionnaire; ViRe, virtual reality; VOR, vestibulo-ocular reflex; VR, vestibular rehabilitation; VRCESS, virtual reality cumulative exercise symptom ratings.

In 2019, Micarelli et al. (14) studied 47 elderly patients diagnosed with unilateral vestibular hypofunction defined as 25% vestibular response reduction on bithermal water caloric irrigation. They found patients in the intervention group using a head-mounted device (HMD) had significantly greater increases in vestibulo-ocular reflex (VOR) gain compared with patients in the control group receiving conventional vestibular rehabilitation, regardless of mild cognitive impairment status. In addition, they found significant decreasing simulator sickness questionnaire scores with progressive weeks of HMD use, showing patient adaptation to virtual reality.

In 2017, Micarelli et al. (15) studied 47 patients with chronic unilateral vestibular hypofunction with a similar diagnostic criteria as their 2019 study. The intervention group receiving adjunctive HMD exercises showed significantly greater ipsilesional improvement in VOR gain than the control group receiving conventional vestibular rehabilitation. VOR gain was measured using a video head impulse test. The intervention group also showed significantly greater improvements in dizziness handicap inventory (DHI) and activities-specific balance confidence (ABC) compared with the control group. Simulator sickness also decreased with successive sessions within the intervention group.

In 2017, Hsu et al. (16) conducted a single-blinded randomized controlled trial on 70 patients clinically diagnosed with Menière's disease based on guidelines published by the American Academy of Otolaryngology. When comparing the efficacy of Cawthorne-Cooksey exercises delivered through a virtual reality device via game like tasks against the traditional method, the authors found significantly more improvement in the intervention group in objective measures of posture and total performance scores during virtual reality assessment tasks (combined score of task completion rate, time, and success frequency).

Garcia et al. (17) studied 44 patients diagnosed with Menière's disease by a single otolaryngologist. All patients received betahistine and diet recommendations. Patients in the intervention group received rehabilitation through a virtual reality and posturography apparatus; they were found to have significantly lower scores in DHI and dizziness analog scales as well as greater stability limits. However, no significant differences were found when comparing the oscillation rates of the patient's center of pressure between the two groups before and after interventions.

Pavlou et al. (18) studied 16 patients with a peripheral vestibular deficit observed on bithermal caloric testing and/or rotary electronystagmography testing after a diagnosis of vestibular neuritis. When comparing vestibular rehabilitation efficacy when delivered in a 3D immersive static versus dynamic environment, they found a significantly greater improvement in the situational vertigo questionnaire and virtual reality cumulative exercise symptom scores within the dynamic group compared with the static group. However, there were no significant between-group differences in beck anxiety and depression inventories, dynamic gait index (DGI), or total phobia scores.

The earliest randomized control trial was published by Viire and Sitarz (19) in 2002 which examined 13 patients with chronic vertigo and the impact of active search tasks using a HMD and virtual reality scene for improving VOR gain and symptom reduction. Patients with chronic vertigo symptoms for at least 6 months and meeting criteria for abnormal VOR gains were included. The treatment group had scene magnification adjusted based on VOR gain. This resulted in a general trend for increases in VOR gain, with statistical significance at 0.64 Hz. DHI was only completed in the treatment group and showed small improvements which decreased over 1 week (20). Statistical analysis was not performed and the clinical significance is unclear. This study suggests potential objective improvements in VOR gain using virtual reality rehabilitation, but conclusions regarding subjective symptomatic improvement are difficult to draw given the small change in DHI scores and lack of control data.

Additional Studies

A total of 158 patients participated in four separate studies. A summary of these studies can be found in Table 2.

TABLE 2 - Summary of non-randomized trials
Author Participants Inclusion and Exclusion Criteria Locations Study Groups Exercise Types Outcomes Conclusions
Viziano et al., 2019 n = 47 with unilateral vestibular lossGender:Group 1 (9F 14M)Group 2 (11F 13M)Age:Group 1 (49.72 ± 10.34)Group 2 (50.48 ± 9.12) N/AFollow-up study, see Micarelli et al. (2017) Rome, Italy 12-month follow-up of Micarelli et al. (2017) study.Group 1 (ViRe + VR) versus Group 2 (VR) VR: Custom patient-tailored home-exercises BID (30–40 min/d). Supervised sessions with physiotherapist office visits twice a week (30–45 min). 4 weeks duration.ViRe: 3D Virtual Reality HMD Track Speed Racing 3D Game (20 min/d). 4 weeks duration. ObjectiveVideo HITStatic posturography testingSubjectiveDHIABC ScaleDGI The positive effects of vestibular rehabilitation on VOR, static posture, and performance measures showed retention at a 12-month follow- up.After 1 year, ViRe + VR yielded better otoneurological scores than VR alone.
Rosiak et al., 2018 n = 50 with vertigo symptomsGender:Group 1 (14F 11M)Group 2 (13F 12M)Age:Group 1 (46.48 ± 10.6)Group 2 (45.20 ± 11.07) Inclusion: Persistent vertigo symptoms and disequilibrium (inadequate compensation) >2 months and unilateral peripheral vestibular impairment confirmed by VNGExclusion: Orthopedic surgery, with a history of epilepsy, bilateral peripheral vestibular loss or central vestibular disorder Lodz, Poland Group 1 (ViRe) versus Group 2 (VR) VR: Static posturography training with visual feedback under supervision; Steer towards a randomly generated point on the screen for 10 sessions (10 25 min/session). 2 weeks duration.ViRe: Hybrid VR unit (force plate, an upper body motion sensor, flat screen display); exercises coordinating upper body movement and maintaining the center of pressure in a predetermined range or shifting the center of pressure towards indicated positions for 10 sessions (10 30 mins/session). 2 weeks duration.Both groups: Cawthorne-Cooksey exercises at home three times daily ObjectiveStatic Posturography TestingSubjectiveVertigo Symptom Scale—Short Form Both methods reduce postural sway, however subjective reduction of symptoms was greater in the VR group.
Yeh et al., 2014 n = 49 with chronic vestibular dysfunction and vestibular symptoms >3 months Inclusion: Age >18, Vestibular dysfunction secondary to diagnosed inner ear etiology who have been objectively (tandem gait and stepping test) symptomatic >3 monthsExclusion: Weakness, paralysis, active treatment, or cognitive dysfunction Taipei, Taiwan Non-controlled study ViRE: 4 game-like training tasks in a projected environment with 3D glasses simulating Cawthorne-Cooksey exercises for six sessions over 4 weeks ObjectivePerformance results: completion rate, completion time, number of continuous ball-catchingUpright balance test indices: maximum and mean mediolateral and anteroposterior trajectory excursion and statokinesigram 5/6 performance results showed significant improvement after training sessions.In the non-stimulation condition, mean mediolateral balance index showed a significant improvement.In the post-stimulation condition, statokinesigram and maximum mediolateral balance indices showed a significant improvement.
Whitney et al., 2009 n = 12 with vestibular disordersAge: 52 (18–80) Inclusion: Diagnosed vestibular disorder, not specifiedExclusion: Weakness, paralysis, active treatment, or cognitive dysfunction Pittsburgh, United States Non-controlled study ViRe: Finding items in a virtual grocery store projected on a panoramic screen while on a treadmill for 6 sessions (24mins/session). 6 weeks duration. ObjectiveDGITUGSensory Organization TestSubjectiveDHIABC At least two-thirds of the participants improved on all of the outcome measures except for the TUG. The greatest improvement occurred for the DHI and ABC
ABC indicates activities-specific balance confidence; DGI, dynamic gait index; DHI, dizziness handicap inventory; F, female; HIT, head impulse test; HMD, head-mounted device; M, male; TUG, timed up and go; ViRe, virtual reality; VOR, vestibulo-ocular reflex; VR, vestibular rehabilitation.

Viziano et al. (21) performed a 12 months follow-up of the 47 patients involved in the trial by Micarelli et al. (15) to investigate the long-term effects of home-based exercises through a HMD compared with conventional rehabilitation. VOR gain, subjective outcomes on DHI, DGI, and ABC were significantly better in both groups compared with pretreatment, but the intervention group continued to have greater improvements in these outcomes compared with the control group at 12 months. Posturography scores remained better in the intervention group compared with the control group. Overall, results showed retention of objective and subjective improvements in vestibular rehabilitation with virtual reality.

In 2018, Rosiak and Jozefowicz-Korczynska (22) conducted a non-randomized control study in 50 patients with unilateral vestibular hypofunction on videonystagmography and symptomatic for at least 2 months. They compared vestibular rehabilitation using a hybrid (force plate, position tracker, screen) VR unit to static posturography with visual feedback; both groups were instructed to perform Cawthorne-Cooksey exercises at home. Results showed statistically significant improvement in posturographic parameters but no differences between groups. However, the intervention group also showed greater improvement in their subjective perception of symptoms on the vertigo syndrome scale—short form.

Whitney et al. (23) conducted a prospective uncontrolled study of 12 patients with vestibular disorders who experienced dizziness and loss of balance in 2009. Participants navigated a virtual grocery store on a treadmill over six 24-minute sessions. Statistically significant improvements were found in both subjective (ABC, DHI), and objective (DGI, sensory organization test) measures. The greatest improvement occurred for subjective measures which indicate decreased perceived dizziness handicap and increased confidence in their balance while performing daily activities

In 2014, Yeh et al. (24) studied 49 patients with chronic vestibular dysfunction with an inner ear etiology who have been symptomatic for over three months. Study participants completed four virtual reality game-like tasks simulating Cawthorne-Cooksey exercises (head movement, eye movement, limb movement, balance) which yielded performance scores such as completion rate, completion time, and the number of continuous ball-catching. Most performance scores and some balance indices showed significant improvement after six training sessions, indicating that the proposed virtual reality system effectively promotes imbalance rehabilitation.

Risk of Bias Analysis

Tables 3 and 4 summarize the results of the RoB 2 and ROBINS-I tool respectively. There were some concerns with Hsu et al. (16), Garcia et al. (17), Pavlou et al. (18), and Viirre et al.'s overall risk of bias. The overall risk of bias for both of Micarelli et al.'s (14,15) articles were low. As a follow-up study of Micarelli et al. (14,15) without attrition, Viziano et al. (21) was also determined to have a low risk of bias. The ROBINS-I assessment tool revealed low concerns with Rosiak et al.'s (22) risk of bias but serious concerns with Yeh et al. (24) and Whitney et al.'s (9,23) risk of bias due to a lack of a comparator group. Sources of such biases are elaborated on within the discussion.

TABLE 3 - Revised cochrane risk-of-bias tool for randomized trials (RoB 2)
Study Overall Bias Bias Arising From the Randomization Process Bias Due to Deviations from the Intended Interventions Bias Due to Missing Data Bias in Measurement of Outcomes Bias in Selection of the Reported Result
Micarelli 2019 Low Low Low Low Low Low
Micarelli 2017 Low Low Low Low Low Low
Hsu 2016 Some concerns Some concerns Some concerns Low Some concerns Low
Garcia 2013 Some concerns Some concerns Low Low Some concerns Low
Pavlou 2012 Some concerns Low Some concerns Low Low Low
Viirre 2002 Some concerns High High High Some concerns High

TABLE 4 - Risk of bias in non-randomized studies of interventions (ROBINS-I)
Study Overall Bias Bias Due to Confounding Bias in Selection of Participants Into the Study Bias in Classification of Interventions Bias Due to Deviations From Intended Interventions Bias Due to Missing Data Bias in Measurement of Outcomes Bias in Selection of the Reported Result
Rosiak 2018 Low Moderate Low Low Low Low Low Low
Yeh 2014 Serious Moderate Low Low Critical Low Low Low
Whitney 2009 Serious Serious Low Low Critical Low Low Low

DISCUSSION

This systematic review describes the effectiveness of virtual reality in vestibular rehabilitation. While conventional rehabilitation has shown benefit for patients, it can be costly and inaccessible (4). Hence, the application of virtual reality has the potential to improve access for patients, which may decrease healthcare burden, including frequency of primary care, specialist, and emergency department visits, in addition to improving patient's symptoms and impairments (25). It is evident that conventional rehabilitation remains the gold standard in the clinical setting given the limited amount of literature on virtual reality applications. The lack of adoption of virtual reality-based vestibular rehabilitation may be secondary to unfamiliarity with virtual reality among vestibular rehabilitation providers, high initial investment costs, and uncertainty of efficacy given limited evidence (26). As such, increased evidence demonstrating the benefits of virtual reality from both a patient and systems perspective, may make more widespread adoption likely.

Existing studies on virtual reality-based vestibular rehabilitation lack standardization in equipment set-up and outcome measures. They are further limited by the heterogeneity of the study population and short length of follow-up. Only one other review on this topic has been published by Pavlou et al. (18). They similarly noted that the restricted sample size and lack of control and randomization was a major limitation. Hence, despite their conclusion that virtual reality was well tolerated and improved vestibular rehabilitation, more prospective randomized data were needed. Additionally, the included studies ranged from 2002 to 2013, representing earlier research on this topic. This review presents studies from 2002 to 2019, acting as an update of the literature.

Of the six RCTs reviewed, four were compared to other forms of vestibular rehabilitation (guided or home-exercises). All four studies reported better subjective or objective outcomes in the virtual reality groups (14–16,18). Two studies had control groups without vestibular rehabilitation; these studies also showed better objective outcomes in the virtual reality groups (17,19). In addition, Rosiak et al. 2018's non-randomized control trial identified greater subjective improvements in the virtual reality group and two uncontrolled intervention studies found improvement after virtual reality intervention in some of the objective and subjective outcomes (22). Finally, Viziano et al. (21) demonstrated the possibility of long-term retention in the objective and subjective improvements derived from virtual reality. Overall, studies reviewed demonstrate a clinical benefit in terms of patient objective and subjective outcomes after virtual reality intervention.

With any rehabilitation modality, both in-clinic and at-home components are imperative so that patients may continue to rehabilitate independently. Our review demonstrated that both in-clinic and at-home virtual reality-based vestibular rehabilitation benefited patients. Only Micarelli et al. (14,15,21) included aspects of at home virtual reality-based intervention which demonstrated benefits which were maintained at 12 months. However, these promising results require validation by additional research groups in future studies.

Unfortunately, significant methodological limitations and heterogeneity exist within the current literature. The average study sample size was 39.5 (range, 13–70) among RCTs and 39.5 (range, 12–50) among other studies, although most were pilot trials. Given the small sample sizes, future studies should aim to recruit adequate sample sizes based on estimate outcome-specific effect sizes or minimally clinically significant differences a priori. Moreover, patient selection, type of virtual reality intervention, having control arms and their specific interventions, and outcomes vary greatly between studies. This has precluded the ability to perform quantitative meta-analysis of our data; however, these drawbacks highlight aspects future studies should be cognizant about.

The inclusion criteria should be strictly defined and standardized to better delineate clinical scenarios in which virtual reality may be beneficial. For example, Micarelli et al. (14,15) performed bithermal water caloric irrigation to objectively screen patients for peripheral vestibular hypofunction, while Hsu et al. (16) and Garcia et al. (17) screened patients for Menière's disease using established criteria. As virtual reality is likely to be studied for a widened scope of vestibular pathologies in the future, clear inclusion criteria will allow for clinical data to be collated to understand the effect of virtual reality on a range of vestibular pathologies. In addition, patient age groups and physical functioning status should be declared to further understand the limitations in accessibility of virtual reality-based vestibular rehabilitation.

Virtual reality encompasses a broad range of technologies that produce immersive user experiences. They can be classified based on physical hardware, user interaction, and level of immersion (hardware presence, number of sensory modalities, portability, detail, user input and feedback) (27,28). Visual information in the reviewed studies ranged from 3D HMDs, projected environment and 3D goggles, or screens. Motion tracking included head position, force platform, body-limb position, or a combination. User interaction ranged from simple experiential simulated environments, active search tasks, goal-oriented tasks, or game-based tasks. The therapeutic design of the virtual reality intervention varied in terms of targeted rehabilitation of gaze stability, posture stability, and/or exposure for habituation (Table 1). Given the diversity of virtual reality characteristics, future studies should continue to thoroughly describe the hardware, software, and therapy goals of their interventions in methodologies.

Control groups varied across trials ranging from conventional vestibular rehabilitation (guided or independent), medical and dietary intervention, or diminished virtual reality environments (Tables 1 and 2). Specific study objectives and type of virtual reality modality may necessitate different types of control groups. Future studies should design control arms to compare virtual reality to a conventional program of similar duration, intensity, and supervision or to assess additional benefit gained from combining virtual reality with conventional rehabilitation.

Various outcomes were measured in these studies. Seven studies had both objective and subjective outcomes consisting of validated questionnaires (e.g., DHI, ABC, anxiety/depression scales, other dizziness scales) and objective tests or measurements (e.g., DGI, video head impulse test, vestibular examination, VOR gain, posturography, task performance scores) (14,15,17,19,21–23). Hsu et al. (16) and Yeh et al. (24) lacked subjective outcomes and Pavlou et al. (18) lacked objective outcomes which may limit the internal validity of their findings. Although subjective outcomes are more important in establishing clinical benefit, evidence of similar improvements in objective outcomes can enhance the internal validity of future studies; therefore at least one of each should be included to adequately assess efficacy. Lastly, only Micarelli et al. (14,15) and Pavlou et al. (18) assessed adverse events related to virtual reality despite simulator sickness being a well-documented concern (29). Future studies should monitor, measure, and report adverse effects of virtual reality.

Sources of bias for the studies derived from multiple factors (Tables 3 and 4). Due to the participatory nature of rehabilitation, subjects, outcome assessors, and those delivering the interventions were un-blinded to the subject's assigned intervention, potentially impacting subjective outcomes. Additionally, studies involving self-directed exercises are subject to compliance bias which may contribute to decreased efficacy. Compliance bias can impact effect size in both groups or result in under or overestimation of intervention versus control effect size if affecting the groups differently (30). Outcomes which are assessed using the same virtual reality tool as the intervention may introduce repeat testing bias as intervention groups have had more exposure to the assessment apparatus and instructions; however, the impact can be mitigated by learning sessions for both groups before trial commencement, such as the case in Hsu et al. (16).

This review itself is not without limitations. Primarily, the small number of studies included and sample sizes limit both the internal and external validity. Since most studies used different modalities of virtual reality and outcomes, the existing data were not amenable to meta-analysis. Therefore, the conclusions of this review are only of a qualitative nature and cannot comment on statistical significance of the existing findings. Limitations within individual studies also contribute to potential limitations associated with conclusions in this review. Assessment of outcomes was limited by the lack of a predetermined clinically significant difference and certain outcomes may be impacted by non-blinding bias, compliance bias, or repeat testing bias. Therefore, in addition to the lack of quantitative statistics, the positive effects observed in the included studies should be interpreted with some caution.

CONCLUSION

This systematic review summarizes prospective studies examining the effectiveness of virtual reality in vestibular rehabilitation. The number of studies comparing virtual reality-based to conventional vestibular rehabilitation remains limited. Based on the qualitative analysis in the current review, the literature suggests potential clinical benefit of virtual reality for vestibular rehabilitation compared with conventional vestibular rehabilitation. However, significant heterogeneity and limitations exist within the literature in terms of population selection, intervention design, comparators, and evidence-based clinical outcomes. Future studies should aim to address these limitations.

Acknowledgments

None.

REFERENCES

1. Hogue JD. Office evaluation of dizziness. Prim Care 2015; 42:249–258.
2. McDonnell MN, Hillier SL. Vestibular rehabilitation for unilateral peripheral vestibular dysfunction. Cochrane database Syst Rev 2015; 1:CD005397.
3. Neuhauser HK, Radtke A, von Brevern M, Lezius F, Feldmann M, Lempert T. Burden of dizziness and vertigo in the community. Arch Intern Med 2008; 168:2118–2124.
4. Le A, Lelli DA, Van Katwyk S, Hogan D, Thavorn K, Tse D. Dizziness at a Canadian tertiary care hospital: a cost-of-illness study 11 medical and health sciences 1117 public health and health services. J Otolaryngol Head Neck Surg 2019; 48:1–6.
5. Han BI, Song HS, Kim JS. Vestibular rehabilitation therapy: review of indications, mechanisms, and key exercises. J Clin Neurol 2011; 7:184–196.
6. Deveze A, Bernard-Demanze L, Xavier F, Lavieille JP, Elziere M. Vestibular compensation and vestibular rehabilitation. Current concepts and new trends. Neurophysiol Clin 2014; 44:49–57.
7. Vereeck L, Wuyts FL, Truijen S, De Valck C, Van de Heyning PH. The effect of early customized vestibular rehabilitation on balance after acoustic neuroma resection. Clin Rehabil 2008; 22:698–713.
8. Gurr B, Moffat N. Psychological consequences of vertigo and the effectiveness of vestibular rehabilitation for brain injury patients. Brain Inj 2001; 15:387–400.
9. Whitney SL, Alghwiri A, Alghadir A. Physical therapy for persons with vestibular disorders. Curr Opin Neurol 2015; 28:61–68.
10. Yardley L, Donovan-Hall M, Smith HE, Walsh BM, Mullee M, Bronstein AM. Effectiveness of primary care-based vestibular rehabilitation for chronic dizziness. Ann Intern Med 2004; 141:598–605.
11. Rose T, Nam CS, Chen KB. Immersion of virtual reality for rehabilitation - Review. Appl Ergon 2018; 69:153–161.
12. Higgins JPT, Altman DG, Gøtzsche PC, et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ 2011; 343:
13. Sterne JA, Hernán MA, Reeves BC, et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ 2016; 355.
14. Micarelli A, Viziano A, Micarelli B, Augimeri I, Alessandrini M. Vestibular rehabilitation in older adults with and without mild cognitive impairment: effects of virtual reality using a head-mounted display. Arch Gerontol Geriatr 2019; 83:246–256.
15. Micarelli A, Viziano A, Augimeri I, Micarelli D, Alessandrini M. Three-dimensional head-mounted gaming task procedure maximizes effects of vestibular rehabilitation in unilateral vestibular hypofunction: a randomized controlled pilot trial. Int J Rehabil Res 2017; 40:325–332.
16. Hsu SY, Fang TY, Yeh SC, Su MC, Wang PC, Wang VY. Three-dimensional, virtual reality vestibular rehabilitation for chronic imbalance problem caused by Ménière's disease: a pilot study∗. Disabil Rehabil 2017; 39:1601–1606.
17. Garcia AP, Ganança MM, Cusin FS, Tomaz A, Ganança FF, Caovilla HH. Vestibular rehabilitation with virtual reality in Ménière's disease. Braz J Otorhinolaryngol 2013; 79:366–374.
18. Pavlou M, Kanegaonkar RG, Swapp D, Bamiou DE, Slater M, Luxon LM. The effect of virtual reality on visual vertigo symptoms in patients with peripheral vestibular dysfunction: a pilot study. J Vestib Res Equilib Orientat 2012; 22:273–281.
19. Viirre E, Sitarz R. Vestibular rehabilitation using visual displays: preliminary study. Laryngoscope 2002; 112:500–503.
20. Jacobson GP, Newman CW. The development of the dizziness handicap inventory. Arch Otolaryngol Neck Surg 1990; 116:424–427.
21. Viziano A, Micarelli A, Augimeri I, Micarelli D, Alessandrini M. Long-term effects of vestibular rehabilitation and head-mounted gaming task procedure in unilateral vestibular hypofunction: a 12-month follow-up of a randomized controlled trial. Clin Rehabil 2019; 33:24–33.
22. Rosiak O, Krajewski K, Woszczak M, Jozefowicz-Korczynska M. Evaluation of the effectiveness of a Virtual Reality-based exercise program for Unilateral Peripheral Vestibular Deficit. J Vestib Res Equilib Orientat 2018; 28:409–415.
23. Whitney SL, Sparto PJ, Alahmari K, Redfern MS, Furman JM. The use of virtual reality for people with balance and vestibular disorders: the Pittsburgh experience. Phys Ther Rev 2009; 14:299–306.
24. Yeh SC, Chen S, Wang PC, Su MC, Chang CH, Tsai PY. Interactive 3-dimensional virtual reality rehabilitation for patients with chronic imbalance and vestibular dysfunction. Technol Heal Care 2014; 22:915–921.
25. Kovacs E, Wang X, Grill E. Economic burden of vertigo: a systematic review. Health Econ Rev 2019; 9:37.
26. Glegg SMN, Levac DE. Barriers, facilitators and interventions to support virtual reality implementation in rehabilitation: a scoping review. PM R 2018; 10:1237.e1–1251.e1.
27. Jung S (Shawn), Jeong J. A Classification of Virtual Reality Technology: Suitability of Different VR Devices and Methods for Research in Tourism and Events. 2020; Cham: Springer, 323-332.
28. Miller HL, Bugnariu NL. Level of immersion in virtual environments impacts the ability to assess and teach social skills in autism spectrum disorder. Cyberpsychol Behav Soc Netw 2016; 19:246–256.
29. Dużmańska N, Strojny P, Strojny A. Can simulator sickness be avoided? a review on temporal aspects of simulator sickness. Front Psychol 2018; 9:2132.
30. Picha KJ, Howell DM. A model to increase rehabilitation adherence to home exercise programmes in patients with varying levels of self-efficacy. Musculoskeletal Care 2018; 16:233–237.
Keywords:

Balance; Vertigo; Vestibular rehabilitation; Vestibulopathy; Virtual reality

© 2021, Otology & Neurotology, Inc.