Article Text

Can vestibular rehabilitation exercises help patients with concussion? A systematic review of efficacy, prescription and progression patterns
  1. Drew A Murray1,
  2. Dara Meldrum2,
  3. Olive Lennon1
  1. 1School of Public Health, Physiotherapy and Sports Science, University College Dublin, Dublin, Ireland
  2. 2School of Physiotherapy, Royal College of Surgeons Ireland, Dublin, Ireland
  1. Correspondence to Dr Olive Lennon, School of Public Health, Physiotherapy and Sports Science, A303 Health Sciences Centre, University College Dublin, Belfield, Dublin 4, Ireland; olive.lennon{at}


Objective Concussion symptoms normally resolve within 7–10 days but vertigo, dizziness and balance dysfunction persist in 10–30% of cases causing significant morbidity. This study systematically evaluated the evidence supporting the efficacy, prescription and progression patterns of vestibular rehabilitation therapy (VRT) in patients with concussion.

Design Systematic Review, guided by PRISMA guidelines and presenting a best evidence synthesis.

Data sources Electronic databases PubMed (1949 to May 2015), CINAHL (1982 to May 2015), EMBASE (1947 to May 2015), SPORTDiscus (1985 to May 2015), Web of Science (1945 to May 2015) and PEDRO (1999 to May 2015), supplemented by manual searches and grey literature.

Eligibility criteria for study selection Article or abstract of original research, population of patients with concussion/mild traumatic brain injury (mTBI) with vestibular symptoms, interventions detailing VRT, measurement of outcomes pre-VRT/post-VRT. Study type was not specified.

Results Following a double review of abstract and full-text articles, 10 studies met the inclusion criteria: randomised controlled trial (n=2), uncontrolled studies (n=3) and case studies (n=5). 4 studies evaluated VRT as a single intervention. 6 studies incorporated VRT in multimodal interventions (including manual therapy, strength training, occupational tasks, counselling or medication). 9 studies reported improvement in outcomes but level I evidence from only 1 study was found that demonstrated increased rates (OR 3.91; 95% CI 1.34 to 11.34; p=0.002) of medical clearance for return to sport within 8 weeks, when VRT (combined with cervical therapy) was compared with usual care. Heterogeneity in study type and outcomes precluded meta-analysis. Habituation and adaptation exercises were employed in 8 studies and balance exercises in 9 studies. Prescription and progression patterns lacked standardisation.

Conclusions Current evidence for optimal prescription and efficacy of VRT in patients with mTBI/concussion is limited. Available evidence, although weak, shows promise in this population. Further high-level studies evaluating the effects of VRT in patients with mTBI/concussion with vestibular and/or balance dysfunction are required.

  • Concussion
  • Brain
  • Contact sports
  • Rehabilitation

Statistics from


Mild traumatic brain injury (mTBI) and concussion are used interchangeably in the scientific literature.1–5 The literature relating to assessment and acute management of mTBI/concussion has grown substantially over the past decade.2 ,6 It is estimated that sport-related concussions alone account for 1.6–3.8 million cases each year in the USA.7 For the majority of these patients, symptoms resolve within the first 7–10 days. Persisting symptoms are experienced in 10–30% of cases.7–11 These ongoing symptoms can cause significant morbidity, are frustrating for individuals and lead to extended time away from work, school and/or sport. Common persisting symptoms of mTBI and concussion include vertigo, dizziness and balance dysfunction, all of which are associated with vestibular impairment.12–14

Vestibular impairment after concussion is well documented and is classified as either peripheral or central. Peripheral impairment is as a result of damage to the peripheral end organ (labyrinth) and/or the eighth cranial nerve, both of which can be traumatised by forces in the cranial cavity.15 ,16 Zhou and Brodsky17 recently reported that 90% of a series of 42 children with sport-related concussion had a vestibular or balance deficit. This included abnormal caloric tests (21%), spontaneous nystagmus (24%) and abnormal computerised dynamic posturography (56%). Deficits of dynamic visual acuity (DVA) and subjective visual vertical in 57% and 13%, respectively, were also found.

Gait impairment has also been documented postconcussion with reduced centre of mass/centre of pressure excursion, increased mediolateral sway and slower speed, indicating the adoption of a conservative gait strategy.18

Vestibular rehabilitation therapy (VRT) has been recommended for individuals with persistent postconcussive symptoms12–14 ,19 and has been reviewed by several authors.19–22 The recent Consensus Statement on Concussion in Sport advised that persistent symptoms should be managed by a multidisciplinary team of healthcare providers with experience in sport-related concussion23 and that therapies including cognitive, physical, psychological and vestibular should be considered as part of a comprehensive rehabilitation programme.23 Recently updated clinical practice guidelines for concussion/mTBI and persistent symptoms have included a specific recommendation of VRT for persistent vestibular symptoms or dysfunction.2 ,19

Cooksey24 and Cawthorne25 first used VRT in the 1940s to treat patients with dizziness and balance dysfunction. Since then, the original principles have been expanded on and developed, and vestibular rehabilitation (VR) is now a well-established intervention for individuals with vestibular and balance disorders.26 VRT programmes are customised to the individual patient's needs using specific types of exercises and manoeuvres that include: canalith repositioning manoeuvres (curative for benign paroxysmal positional vertigo), habituation exercises (for impaired motion sensitivity), adaptation or gaze stability exercises (for deficits in the vestibulo-ocular reflex (VOR)), substitution exercises (to facilitate central preprogramming of eye movement), balance exercises and aerobic exercises (to improve balance usually with a graduated walking programme).27 An updated Cochrane review showed a moderately strong evidence base of VRT for reduction of subjective dizziness, improvement in balance and gait outcomes and participation in those with unilateral peripheral vestibular disorders.28 Since post-concussive symptoms of vertigo and imbalance may be suggestive of vestibular dysfunction, VRT is logically emerging as a potential intervention in mTBI/concussion patient populations.12 ,28–32

The purpose of this systematic review was twofold: first, to evaluate published evidence supporting the use of VRT in the mTBI/concussion population, and second, to determine if specific outcome measures, and prescription and progression patterns could be identified within the VRT treatment paradigm for mTBI/concussion populations.

Framed as clinical questions, this systematic review aimed to answer:

  1. Does VR postconcussion improve: (a) subjective reports of dizziness and vertigo, (b) gaze stabilisation deficits, (c) balance impairment, (d) gait impairment?

  2. Does VR facilitate early return to sport/work?


Search strategy

A systematic search and review of the literature was conducted, guided by the PRISMA guidelines.33 The following databases were searched up to May 2015: PubMed (1949–2015), CINAHL (1982–2015), EMBASE (1947–2015), SPORTDiscus (1985–2015), Web of Science (1945–2015) and PEDRO (1999–2015). Carefully constructed search strategies were designed in conjunction with a librarian, including controlled vocabulary terms specific to each database and employing Boolean operators AND and OR. No date limits were set on the searches. Limits of English language and human studies applied. The final search strategy is shown in online supplementary Appendix A. Since VR is an emergent treatment paradigm in mTBI/concussion, no limitation was applied by study methodology. The authors felt that this would allow a more comprehensive overview of the best available evidence in this field. Reference lists from retrieved articles and guideline documents were screened for additional relevant articles, publications, posters, abstracts and conference proceedings. Grey literature was hand searched by one of the authors (DAM) and included publications from pertinent organisations, government agencies and non-governmental organisations (NGOs) including the International Society for Neuro-Otology (Bárány Society) literature.

Selection criteria

Selection criteria for articles included in this review comprised: (1) article or abstract of original research, (2) population of patients with concussion/mTBI with vestibular symptoms, (3) interventions detailing VRT, (4) measurement of outcomes pre-VRT and post-VRT to evaluate treatment effect. A secondary outcome of interest was the description of treatment prescription and progression patterns. Eligible references were exported from the databases searched to EndNote software and duplicates removed. Studies were screened by title by one reviewer (DAM) and the remaining citations were independently reviewed at abstract and full manuscript stages by two reviewers (DAM and DM). In case of disagreement in the latter two review processes, a third author (OL) was designated to discuss disparities and reach a consensus.

Data extraction

Two reviewers (DAM and DM) collaboratively extracted the data that included the following variables: type of study and level of evidence attributed by study type, study population (age, sex, concussion genesis, recruitment source), intervention, outcome measures (addressing dizziness, gaze stabilisation (VOR), balance and gait) and results. Where information was available about the VRT intervention delivered, it was extracted using the Frequency, Intensity, Type and Time (FITT) taxonomy.34

Risk of bias

The quality of included studies was evaluated by two reviewers (DAM and DM) using the Cochrane Risk of Bias Tool for randomised controlled trials (RCTs)35 under headings of sequence generation, allocation concealment, blinding, incomplete data, selective outcome reporting and other potential threats to validity generating a final rating of high, moderate or low risk of bias. For observational studies (cohort, case–control, cross-sectional) and case studies, the Effective Public Health Practice Project Quality Assessment Tool was employed, evaluating studies under the headings of selection, design, confounders, blinding, data collection and withdrawals/dropouts generating a final global rating for quality assessment of weak, moderate or strong.36 Since this review was inclusive of alternative study methodologies in addition to RCTs, included articles were additionally graded for the level of evidence provided by Sackett's37 initial rules of evidence. This consists of levels I–V evidence, as categorised below:

  • Level I: (a) systematic review (with homogeneity) of RCTs,

  • (b) individual RCT (with narrow CI);

  • Level II: (a) systematic review (with homogeneity) of cohort studies,

  • (b) individual cohort study, including low-quality RCTs,

  • (c) outcomes research;

  • Level III: (a) systematic review (with homogeneity) of case–control studies,

  • (b) individual case–control studies;

  • Level IV: case series (and poor quality cohort and case–control studies);

  • Level V: expert opinion without explicit critical appraisal, or based on physiology, bench research or ‘first principles’.

Data synthesis

Limitation of data synthesis by meta-analysis was anticipated because of heterogeneity of methodologies included in the search strategy and the outcome measures of interest. A narrative approach to analysis was first proposed, summarising all literature pertaining to VRT by outcomes of interest of dizziness, gaze stabilisation, balance/gait and return to sport and work and grading the level of evidence attributed to the findings by study type. Secondary outcomes of VRT prescription and progression patterns were also summarised narratively.

Best evidence synthesis

Following the narrative review, best evidence synthesis was proposed whereby the team discussed the internal validity and clinical merits of each study. A decision was then made about the study’s admissibility for the best evidence synthesis. Studies identified in the review but not included in best evidence synthesis were case series or small clinical cohorts, which cannot be used to estimate the effectiveness or relative effectiveness of interventions. Other studies excluded from best evidence synthesis were those deemed likely to be at risk of bias due to selection, information or confounding factors.38

Data from studies judged to be scientifically admissible were then abstracted into an evidence table addressing each outcome of interest. Where a study was related to more than one outcome, it was included more than once.


The final search was completed on 15 May 2015. The systematic search retrieved 3355 articles with an additional 15 articles located by searching reference lists and grey literature. The full article selection process is illustrated in figure 1. Primary reasons for exclusion at all three stages of review included ineligible patient populations or lack of VRT interventions. Consensus was reached without third party arbitration. Ten articles met the full study inclusion criteria, including two RCTs39 ,40 (1 was an abstract only), two prospective cohort studies,13 ,41 one retrospective studies12 and five case studies42–44 (1 was abstract only). The study design, study population, intervention (using FITT taxonomy), outcome measures, results, level of evidence and risk of bias of included studies are presented in table 1.

Figure 1

PRISMA flow diagram for systematic search. TBI, traumatic brain injury; VRT, vestibular rehabilitation therapy.

Table 1

Summary of included studies

Patient demographics

All patients in the included studies experienced documented dizziness and vertigo or balance and gait impairments as a result of a concussion/mTBI. There were twice as many male versus female participants recorded (194 vs 99). In the RCT (published abstract only) by Cuff et al,40 the sex of the 62 patients was not specified. The age of included participants varied from 8 to 73 years. In 8 of the 10 studies, the patient population was drawn from a military or sporting background. The remaining two studies, by Alsalaheen et al12 and Cuff et al,40 did not clearly state the background to the mTBI/concussion.

Efficacy of interventions

All studies reported improvement in outcome measures of interest, except the one by Cuff et al.40 No adverse responses to the interventions were documented in any of the studies.12 ,13 ,39–43 ,45 ,46 Heterogeneity of study types and outcome measures employed did not allow data meta-analysis, as anticipated. The main outcomes of interest included measures of dizziness, gaze stabilisation, balance and gait and return to work/sport and are described separately in narrative format with the best available evidence level provided.

Dizziness: Seven of the 10 studies included outcome measures addressing the effect of VRT on dizziness scores preintervention and postintervention.12 ,13 ,39 ,40 ,42 ,43 ,45 ,46 One RCT (level II evidence) reported that individuals who were medically cleared to return to sport had greater improvement in Dizziness Handicap Inventory (DHI) scores than those who were not medically cleared p=0.019.39 The second RCT (level II evidence) reported no significant difference between groups on self-reported dizziness preintervention or postintervention p=0.16840 when VRT was compared with a control group. A retrospective chart review (level III evidence) reported significant treatment effect for dizziness in children but not adults (p<0.005).12 A prospective cohort study (level III evidence) reported that dizziness measures had returned to normal values, measured through gaze stabilisation tests, at 12 weeks (p≤0.01).13 Two case studies (level V evidence) reported improvements in dizziness measures using the DHI.43 ,46 Kleffelgaard et al42 used the smallest detectable change of 20 points on the DHI to indicate a meaningful change. Faltus43 provided pre/post scores as evidence of improvement. A third case study (level V evidence) reported complete resolution of dizziness symptoms using the Dix-Hallpike procedure.45

Subjective reports of dizziness and vertigo

Only half of the studies measured dizziness. Outcome measures included the DHI,12 ,39 ,42 ,43 the Vertigo Symptom Scale,42 the modified Motion Sensitivity Quotient39 and a dizziness analogue scale.12 ,40 Alsalaheen et al12 reported reductions in mean dizziness visual analogue scale (VAS) scores from 21 (±22) to 12 (±18), p<0.001, and in mean DHI scores from 49 (±21) to 30 (±22), p<0.001. Schneider et al39 reported median DHI changes in the intervention group of −24 and −13 in those cleared and not cleared, respectively (to return to sport). Conversely, in the control group, median changes of −48 and −21 were observed in those cleared and not cleared, respectively (to return to sport). The other dizziness outcome measure employed was the modified Motion Sensitivity Quotient (scored out of 40 with lower scores indicating improvement). In the treatment group, median changes of −10 and −1.75 were reported in those cleared and not cleared, respectively (to return to sport). Conversely, in the control group, median changes of −20 and −7.25 were observed in those cleared and not cleared, respectively (to return to sport). Of note, median change scores in the control group were better than those in the intervention group, but the study was underpowered to detect significant changes in the secondary outcome measures and did not analyse between-group differences.

Gaze stabilisation

An explicit statement regarding incorporation of gaze stabilisation or adaptation exercises was provided in 9 of the 10 studies. Only two of these studies measured DVA or gaze stabilisation. Schneider et al39 were underpowered to detect change, and did not specifically test between-group differences in the magnitude of DVA improvement between the group that received VR and the control group. Data presented for the treatment group showed a median improvement of 2 and 2.5 lines (of acuity) for the cleared and not cleared groups, respectively, whereas the control group demonstrated improvements of 1 and 2 lines (of acuity) in the same groupings. Gottshall et al13 reported improvements in DVA that returned to normative levels and ranged from 0.09 to 0.16 LogMAR (in 4 directions of head movement, up, down, left and right). No explicit statement regarding provision of gaze stabilisation exercises were found in the study by Cuff et al.40

Balance impairment

Nine of the 10 studies employed balance outcome measures that looked at the effect of VRT on balance scores preintervention and postintervention.12 ,13 ,39 ,40 ,42–46 Balance was measured using self-reported balance scales in four studies12 ,39 ,40 ,46 and dynamic balance tests in eight studies.12 ,13 ,39 ,42–46 The fully published RCT (level II evidence) reported that 64% of the medically cleared treatment group reached 100/100 on the Activities-specific Balance Confidence scale compared with 25% of the control group who were not medically cleared.39 In one RCT (level II evidence), published in abstract form only, it was unclear what self-reported balance measure was used and the authors Cuff et al40 concluded there were no significant differences between groups on self-reported balance problem (p=0.534). The retrospective chart review (level III evidence) reported a significant treatment effect in all balance measures p<0.05.12 The prospective cohort study (level III evidence) reported that balance measures had returned to full values at 12 weeks, and significance was defined as p≤0.01.13 Five case studies (level V evidence) reported improvements in most if not all balance measures following intervention, but it was not noted if values were significant.42–46

Gait impairment

Only 6 of the 10 studies reported outcomes on gait impairment. Alsalaheen et al12 reported a Functional Gait Assessment (FGA) mean improvement of 6 points (p<0.001), a Timed Up and Go (TUG) mean improvement of 1.9 m/s (p<0.001) and a Dynamic Gait Index (DGI) mean change of 3 points (p<0.001) following treatment. Gottshall et al,45 in a single case study, demonstrated a change in DGI from 21/24 to 24/24 postrehabilitation and, in a further study (n=82),13 reported mean DGI increases from 21/24 to 23/24. Kleffelgaard et al42 found that on testing with the High-Level Mobility Assessment Tool (HiMAT), which incorporates measures of gait, a quarter of participants demonstrated a clinically meaningful improvement of ≥4 points. Rabago and Wilken46 used three-dimensional gait analysis and a virtual reality system to measure gait. Data were presented in graph form only, precluding any meaningful analysis. Schneider et al39 reported median FGA improvements in the intervention group of 1 and 3 in those cleared and not cleared, respectively (to return to sport). Conversely, in the control group, a median FGA change of 3 and 1 was observed in those cleared and not cleared, respectively (to return to sport). The authors reported that the study was underpowered to detect significant changes in the secondary outcome measures.

Return to work/sport

Five of the 10 studies included return to work/sport as an outcome measure.39 ,41 ,43 ,45 ,46 The RCT (level II evidence) reported that those who fully completed the intervention were 10.27 times (95% CI 1.51 to 69.56) more likely to be medically cleared to return to sport within 8 weeks than those in the control group.39 The odds remained statistically significant for return to sport in favour of the intervention group following intention to treat, although markedly reduced (OR 3.91; 95% CI 1.34 to 11.34). One cohort study (level III evidence) categorised patients based on symptoms into 1 of 4 groups prior to intervention and reported mean time to return to work for each (see table 1). The benign positional vertigo group returned to work in under a week, the post-traumatic migraine-associated dizziness group returned at 3.8 weeks, the post-traumatic exercise-induced dizziness group at 4 weeks and the post-traumatic spatial disorientation group took considerably longer, returning over 3 months later. Three case studies (level V evidence) reported return to work/sport as an outcome. In one study, the subject returned to full active military duties following 16 weeks of VRT.45 The second case returned to restricted military duties following 3 weeks of treatment.46 In the final case study, the individual was not medically cleared to return to sport at week six following treatment. This study was subsequently discontinued when the subject relocated.43

Best evidence synthesis: Results are summarised as best evidence synthesis in table 2. Only two studies were eligible for inclusion in this process, of which one was a fully reported RCT and the second was a retrospective chart analysis.

Table 2

Best evidence synthesis table

Prescription and progression patterns of VRT

Table 3 details the prescription of VRT under the headings of frequency, intensity, time and type and document patterns of exercise progression, where reported. A lack of consistency in approach is clear across the included studies. Where face-to-face therapy sessions were provided, they varied between once a week,39 two times a week,42 ,47 to four times over 6 weeks.43 Six studies documented a daily home exercise programme (HEP),12 ,13 ,39 ,40 ,42 ,43 three did not specify the HEP prescription frequency41 ,44 ,45 and one involved six 1-hour sessions over 3 weeks.46 In five of the studies, the intensity of exercises prescribed was not clearly specified.12 ,39 ,40 ,42 ,44 Where intensity of exercise prescription was clearly identified, a large variation between the studies was noted.13 ,41 ,43 ,45 ,46 For example one study was guided by maximum patient tolerance,13 while others tailored the prescription to individual patient improvement.41 ,46

Table 3

Summary of VRT prescription and progression patterns

The window for intervention showed no consistency, varying from 2 weeks to 6 months postinjury across the studies.12 ,13 ,39–46 Again, the type/s of VRT differed across the studies. Four studies considered VRT as a single intervention.12 ,13 ,40 ,46 The remaining studies reported multimodal interventions that incorporated VRT and cervical spine physiotherapy,39 VRT and strength training,43 VRT and occupational tasks,46 VRT and counselling,42 and VRT combined with pharmacotherapy.41 ,45 Habituation exercises, aimed at gradual increase in exposure to provoking stimuli or head movements to reduce the pathological response to the stimuli as described by Aligene et al,27 were employed in 8 of the 10 studies,12 ,13 ,39 ,42–46 with one study employing habituation exercises for visual vertigo symptoms using optokinetic stimulation and gaze stabilisation challenges delivered via a virtual reality medium.46 Adaptation exercises were described in 8 of the 10 studies included in this review.12 ,13 ,39 ,42–46 Vestibulo-ocular adaptation exercises, involving movement of the head and/or eyes to facilitate VOR adaptation, re-education or symptom habituation, were documented as used in only two studies.12 ,48 Balance exercises, to strengthen the supportive balance systems and improve overall function,27 were incorporated in 9 of the 10 studies meeting full selection criteria.12 ,13 ,39 ,40 ,42–46 Aerobic exercises, recommended in VRT to further strengthen the balance systems through muscle conditioning,27 were incorporated in only 3 of the 10 studies.13 ,43 ,45

Treatment progressions were described in 5 of the 10 studies included in this review.12 ,13 ,39 ,43 ,45 ,46 Two studies divided the VRT into phases and noted progressions made during rehabilitation.43 ,46 Two included online supplementary tables describing the exercises used and progression levels within adaptation, habituation and balance exercises.13 ,39

A final study12 was supported by a subsequent publication detailing the progression and prescription patterns.48


There are few published studies investigating the effects of VRT on persistent symptoms of vertigo, dizziness, gaze stability, gait and balance impairment and return to work and sport in those with mTBI/concussion. Quality assessment of study methodology of the 10 studies fulfilling inclusion for this review indicated a high risk of bias in 7 of them. Currently, the level of evidence supporting VRT in this population can be considered to be low when Sackett's criteria are applied. The highest level of evidence (level II) was drawn from one RCT with a low risk of bias.39 This study's intervention group included cervical spine physiotherapy as well as VRT, so the treatment effects cannot be solely attributed to VRT.39 A second RCT in this area was published in abstract format only and presented with a high risk of bias, limiting the conclusions that can be drawn.40 Prospective studies (n=2),13 ,41 a retrospective study12 and case studies (n=5)42–46 demonstrated promising results but at lower levels of evidence. Even though the current literature is sparse, the existing studies suggest an emerging use of VRT in patients with mTBI/concussion experiencing vertigo or balance impairments. Patients with concussion/mTBI with imbalance and dizziness symptoms appear to respond most effectively to VRT. However, optimal time to begin treatment following injury remains unclear. VRT is widely accepted49 as a primary treatment option for a large number of patients with dizziness and moderate-to-strong evidence supports its safe and effective use for unilateral peripheral vestibular disorders.28 The gold standard of care for VRT involving a problem-oriented approach, identifying functional limitations by a thorough evaluation and subsequent prescription of a customised exercise programme50 built around the patient's specific limitations, lends itself well to more complex neurological presentations as in concussion/mTBI. There is now a clear requirement for more high-quality RCTs of VRT in the mTBI/concussion population to determine its effectiveness. There is a concurrent need to determine whether FITT principles of currently reported VRT interventions will need to be modified for the mTBI/concussion population to be effective. Shepard et al51 reported that the duration of VRT was longer in individuals with head injury compared with unilateral peripheral vestibular dysfunction.49 This remains to be determined in mTBI/concussion as the pathology may involve central vestibular and/or peripheral vestibular dysfunction.

Current concussion guidelines recommend mental and physical rest until the symptom's resolution.1 ,5 Where symptoms persist, it is not clear when therapeutic intervention is indicated and how early it should be initiated. The definition of persistent symptoms varied within the literature from 7 days up to 3 months.7–11 ,52 Further research is required to determine if an optimal time frame to begin treatment in patients with persisting symptoms exists. In this regard, the study by Cuff et al40 found no evidence for the efficacy of VRT in a paediatric population when treatment started in the initial 2 week period following concussion. Physical and cognitive rest are presently the mainstay of treatment in the early stages postconcussion,5 but the evidence supporting this approach is also sparse.6 Future studies possibly using a factorial approach are urgently needed to determine optimal and timely interventions.

The age range of the participants in the studies included in this review was wide (8–73 years). The best evidence synthesis contains two studies with mixed adult and paediatric populations. Alsalaheen et al12 found a significant interaction of age on reduction in dizziness and improvement in balance. Younger patients improved significantly in both outcomes, whereas older patients did not. Age has previously been shown to affect recovery from concussion with younger high school athletes showing significantly more sustained memory impairment than college athletes.53 The RCT by Schneider et al39 was not sufficiently powered for a subgroup analysis by age and results cannot be generalised to either paediatric or adult populations. Age as a covariate should be considered in future evaluations of VR.

Twice as many male compared with female patients were reported in the studies included in this review, largely influenced by two studies comprising military personnel.13 ,41 Epidemiological studies drawn from hospital data report that women comprise approximately one-third of all traumatic brain injury admissions with higher rates occurring in older women;54–56 however, the chart review of patients with concussion across a large age range included in this review reported a higher proportion of females.12 A previous review in sports-related concussion suggests that the reported incidence of concussion is greater in females than males in sports employing the same rules,57 whereas an included sport-specific concussion study in this review reported the proportion of females at 40%.39 There is an acknowledged lack of focus on sex and gender in mTBI outcome reporting data in general,58 with a recent systematic review by Cancelliere et al59 finding that only 7% of over 200 studies reviewed provided data stratified by sex and those that did showed mixed results. Postconcussion studies investigating sex differences in recovery time also remain limited and inconclusive.57 No study included in this review reported data by sex or compared sex differences in outcomes following VRT.

Prescription and progression patterns

One of the aims of the review was to extract current information relating to the VRT provided to the mTBI/concussion population using the FITT taxonomy. A lack of standardisation across studies was evident. The prescription and progression patterns of VRT exercises outlined by Alsalaheen et al12 ,48 could be the most detailed currently in the literature regarding VRT for the mTBI/concussion population with persistent vestibular or balance dysfunctions and are suggested by the authors as a minimum criteria for future studies in the area. The general framework for exercise prescription used was originally developed for other populations50 and adapted to the concussion/mTBI population; it can be summarised as follows: (1) identify impairments on initial assessment; (2) prescribe a customised initial exercise programme addressing the impairment safely; and (3) progress the exercise programme difficulty in a functional manner.

In this review, Alsalaheen et al48 provided the most detailed exercise intervention that could be categorised using the FITT criteria. They reported that eye–head coordination exercises were the most commonly prescribed exercise type with 95% of participants receiving them as part of their HEP. Standing static balance was the second most commonly prescribed category (88%) and ambulation exercise was the third category most commonly prescribed (76%).48 Ninety per cent of patients who experienced eye–head coordination difficulties received gaze stabilisation exercises initially. Only 54% of patients with walking difficulties were given ambulation exercises initially.48 Alsalaheen et al48 reported that by prescribing static eye–head coordination prior to the ambulation exercises, the likelihood of the patient becoming symptomatic was reduced. Furthermore, by starting with static activity rather than more dynamic exercises, confidence levels may have been increased. As yet, no evidence to support these theories is available. Finally, the authors observed that the adaptation exercise modifiers that were most important were posture and base of support. They commented that a likely progression pattern for an adaptation exercise (the VORx1 exercise) could be performing it first in sitting, then in standing with feet apart, then with feet together, and finally performing it while walking. Prescription patterns indicated that as patients became less symptomatic, exercises were progressed by increasing difficulty.48

There is a clear need for high-quality RCTs with control groups within the area of VRT for the mTBI/concussion population which has persistent symptoms. Standardisation and consensus on outcome measure usage is required going forwards to facilitate meta-analysis and allow for comparisons across studies. Since the included studies from this review included male and female patients ranging from 8 to 73 years of age, future studies are required to evaluate if disparity exists in how different age categories and sexes respond to VRT. None of the studies analysed in this review compared the effect of VRT between sexes. Although it has been reported that females may be more vulnerable to concussions, studies investigating sex differences in recovery time remain limited and inconclusive.54 Further research is also required to determine the optimal time frame to begin VRT post concussion, as well as determining the prescription and progression patterns that provide the most effective recovery.


The scientific literature in the area of VRT on mTBI/concussion is sparse but suggestive of VRT as an emerging and effective intervention in this patient group, when symptoms persist. The evidence supporting VRT for reducing symptoms of vertigo, dizziness and balance dysfunction is limited by low-quality studies and a lack of RCTs in the area. Existing studies have had suggested benefit and no adverse effects have thus far been associated with the treatment. There is a need for high-quality RCTs to definitively evaluate the effects of VRT on patients with mTBI/concussion with persistent vestibular and/or balance dysfunctions. The optimal intervention window and age and sex differences in responses also require future delineation.

What are the findings?

  • Postconcussion vestibular symptoms can cause significant morbidity and lead to extended time away from work, school and/or sport.

  • Vestibular rehabilitation guidelines postconcussion are, to date, consensus based and drawn from comparator vestibular dysfunction studies.

  • This review now suggests that vestibular rehabilitation therapy (VRT) is beneficial in mild traumatic brain injury (mTBI)/concussion with evidence supporting earlier return to sport.

How might it impact on clinical practice in the future?

  • Although the current evidence-based literature is sparse, the literature suggests VRT in mTBI/concussion is advantageous for patients experiencing vertigo or balance impairments.

  • There is a need for high-quality randomised controlled trials to definitively evaluate the effects of VRT on patients with mTBI/concussion with persistent vestibular and/or balance dysfunctions.

  • Agreed standards for evidence-based prescription and progression patterns of VRT would allow clinicians to treat post-concussion patients with vestibular symptoms confidently and effectively.


The author thanks Diarmuid Stokes from the UCD Library for his assistance in setting up the search strings.


View Abstract


  • Twitter Follow Drew Murray at @DrewMurrayPT

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.