Effectiveness of vestibular training for balance and dizziness rehabilitation in people with multiple sclerosis: A Systematic Review and Meta-Analysis

Journal of Clinical Medicine Review Effectiveness of Vestibular Training for Balance and Dizziness Rehabilitation in People with Multiple Sclerosis: A Systematic Review and Meta-Analysis Cristina García-Muñoz 1,†, María-Dolores Cortés-Vega 2,*,†, Alberto Marcos Heredia-Rizo 2, Rocío Martín-Valero 3, María-Isabel García-Bernal 2and María Jesús Casuso-Holgado 2,† 1Department of Physiotherapy, University School Fco. Maldonado, Avd. de los Cipreses s/n, 41640 Osuna, Spain; [email protected] 2Department of Physiotherapy, Faculty of Nursing, Physiotherapy and Podiatry, University of Seville, C/Avicena s/n, 41009 Seville, Spain; [email protected] (A.M.H.-R.); [email protected] (M.-I.G.-B.); [email protected] (M.J.C.-H.) 3Department of Physiotherapy, Faculty of Health Sciences, University of Malaga, Arquitecto Francisco Peñalosa 3, Ampliación de Campus de Teatinos, 29071 Malaga, Spain; [email protected] *Correspondence: [email protected]; Tel.: +34-954486510 †Equal contributions. Received: 24 January 2020; Accepted: 17 February 2020; Published: 21 February 2020   Abstract: Postural instability and dizziness are commonly observed in people with multiple sclerosis (PwMS). The aim of this systematic review was to evaluate the evidence for the use of vestibular rehabilitation, in comparison with other exercise interventions or no intervention, to treat balance impairments and dizziness in PwMS. An electronic search was conducted by two independent reviewers in the following databases: MEDLINE (Pubmed), Scopus, the Physiotherapy Evidence Database (PEDro), Web of Science (WOS), Lilacs, CINHAL and the Cochrane Database of Systematic Reviews (CDSR). A quality assessment was performed using the PEDro scale and the Cochrane Risk of Bias Tool. When possible, the data were pooled in a meta-analysis (95%CI). This systematic review followed the PRISMA guideline statement and was registered in the PROSPERO database (CRD42019134230). Seven studies were included, with a total of 321 participants analysed. Compared with no intervention, vestibular rehabilitation was more effective for balance development ( SMD =2.12 ; 95% CI =0.49, 3.75; p=0.01; I 2 =89%) and dizziness symptoms improvement ( SMD =−17.43 ; 95% CI = − 29.99, − 4.87; p=0.007; I 2 =66%). Compared with other exercise interventions, improvements in favour of the experimental group were observed, but statistical significance for the differences between groups was not reached. Keywords: multiple sclerosis; vestibular diseases; postural balance; dizziness; fatigue; physical therapy modalities 1. Introduction Multiple sclerosis is an autoimmune disease of the central nervous system that affects approximately 2.5 million people in the world at present. It is a complex disease characterised by a wide variety of symptoms [ 1 ].Among these symptoms, dizziness, including postural intolerance, has been reported to affect 49–59% of people with multiple sclerosis (PwMS) [ 2 ], and balance disorders are observed in 75–82% of mild to moderate disable subjects [ 3 ]. Sensory impairments in the visual, vestibular and proprioceptive pathways have been associated with these symptoms [ 3 , 4 ]. Furthermore, a deficit in the integration of these sensory cues along the subcortical and/or cortical areas has also been related to balance performance [5,6]. J. Clin. Med. 2020,9, 590; doi:10.3390/jcm9020590 www.mdpi.com/journal/jcm J. Clin. Med. 2020,9, 590 2 of 16 Together with these sensory integration and processing impairments, fatigue is a very frequent complaint (>70%) that contributes to poor balance control, particularly in the most challenging tasks [ 7 ]. Furthermore, as postulated by Hebert et al. [8], fatigue can be considered as a significant predictor of balance problems in people with multiple sclerosis. Conversely, although the mechanisms of fatigue related to multiple sclerosis are not completely understood [ 9 ], it is reasonable to think that postural instability and difficulty in coordinating eyes, head and body movements during the performance of tasks could also influence fatigue. The vestibular system has an important role in postural control. In people with multiple sclerosis, several areas along the peripheral and central vestibular pathways may be affected, including the eighth nerve, the vestibular nuclei, the oculomotor tracts, the medial longitudinal fasciculus and the cerebellum [ 10 ]. Although central demyelination is expected, peripheral aetiology in vestibular disorders is also very common in multiple sclerosis patients [11,12]. Vestibular rehabilitation consists of exercises that train the sensory systems to provide the correct spatial cues for position as well as for head and body motion [ 13 ]. A vestibular physical therapy program may include exercises designed to improve vestibulo-ocular reflex, cervico-ocular reflex, somatosensory retraining, balance and gait [ 14 , 15 ]. It is based on vestibular adaptation and substitution mechanisms [16], which seem to be useful for peripheral and central vestibular lesions [17,18]. Although it is recommended that vestibular rehabilitation is considered for people with multiple sclerosis [ 19 ], the benefits of this intervention have not been sufficiently investigated. For this reason, the major purposes of this research were (1) to systematically examine the available evidence of vestibular training interventions for balance and dizziness rehabilitation in multiple sclerosis patients and (2) to determine the magnitude of the effects of these interventions in a meta-analysis. 2. Methods 2.1. Data Sources and Search Strategy This systematic review was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [ 20 ]. The review protocol was registered in the PROSPERO database (Registration Number: CRD42019134230). An electronic search was conducted by two independent reviewers on the following databases: MEDLINE (Pubmed), Scopus, the Physiotherapy Evidence Database (PEDro), Web of Science (WOS), Lilacs, CINHAL and the Cochrane Database of Systematic Reviews (CDSR). Key words relating to vestibular intervention, outcomes measures and the medical condition under study were combined. Simple or advanced search was conducted when possible. There was no date restriction in any database. All analyses were performed on published data, and thus no ethical approval and patient consent were required. The search strategy in PubMed was: (“vestibular rehabilitation”[AllFields] OR “vestibular training”[AllFields] OR “vestibular cues”[AllFields] OR “vestibular therapy”[AllFields]) AND ((“posture”[MeSHTerms] OR “posture”[AllFields]) OR (“balance”[AllFields]) OR “postural control”[AllFields] OR imbalance[AllFields] OR (“vertigo”[MeSHTerms] OR “vertigo”[AllFields] OR “dizziness”[AllFields] OR “dizziness”[MeSHTerms]) OR (“vertigo”[MeSHTerms] OR “vertigo”[AllFields])) AND ((“fatigue”[MeSHTerms] OR “fatigue”[AllFields]) OR (“walking”[MeSHTerms] OR “walking”[AllFields]) OR (“gait”[MeSHTerms] OR “gait”[AllFields])) AND “multiple sclerosis”[AllFields]. The search strategy is detailed in Appendix A. 2.2. Research Question and Study Selection Criteria The research question was established following recommendations from the PICOS model (participants, interventions, comparisons, outcomes and study design) as follows: In PwMS, does vestibular rehabilitation improve balance and dizziness symptoms more than other exercise interventions or no intervention? J. Clin. Med. 2020,9, 590 3 of 16 Thus, patients included in the studies were male and female subjects, clinically diagnosed with multiple sclerosis in accordance with the revised McDonald criteria [ 21 ], with walking ability according to the Expanded Disability Status Scale score (EDSS ≤ 7) and with the objective presence of balance impairment and/or dizziness symptoms. The intervention was based on the specific vestibular exercise rehabilitation as defined by Whitney et al. [ 14 , 15 ], compared with other general exercise programmes or no intervention. The primary outcome measures were balance and dizziness. Secondary outcomes were fatigue, walking speed and depression. Randomised controlled trials regarding the effect of vestibular rehabilitation on improving balance and/or dizziness in patients with multiple sclerosis were included. Full texts in English were included. Other methodological designs were excluded. 2.3. Data Extraction and Quality Assessment Two independent reviewers (CGM and MJCH) identified randomised clinical trials from the databases by title and abstract, and any duplicates were removed. Afterwards, a complete reading of the articles was carried out. The reviewers checked that the selected studies met the inclusion criteria. Once all suitable trials had been selected, the same reviewers performed data extraction independently. Data extraction consisted of collecting information for a qualitative and quantitative synthesis. The main characteristics of the studies (design, participants’ characteristics and sample size, comparison intervention, outcome measures and results) were recorded for the qualitative synthesis, and statistical data were collected for the quantitative synthesis. If the data in the publication were incomplete, the corresponding author was contacted. Next, these reviewers implemented an independent quality assessment using the Physiotherapy Evidence Database (PEDro) Scale. This scale is based on the Delphi list and is an accepted tool for rating the inner validity of clinical trials [ 22 ]. Thus, a study with a score of more than 6 points is considered to be level I evidence (6–8: good; 8–10: excellent). If a randomised controlled trial has a score below 5, it is level II evidence (4–5: deficient; <4: poor). In addition, risk of bias was assessed with the Cochrane Risk of Bias Tool for Randomised Trials [ 23 ]. Any disagreements on data extraction or quality assessment were resolved by consensus. 2.4. Data Analysis Results from comparable trials based on intervention parameters, control group and studied outcomes were pooled in a meta-analysis. Mean differences (MD) or standard mean differences (SMD) were calculated with 95% confidence intervals (95% CI). A random-effects model was applied to present the pooled effect. Review Manager (RevMan) (Computer program) Version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014, was used to analyse the effects and construct forest plots and a risk of bias summary. 3. Results 3.1. Study Selection and Methodological Quality Assessment A total of 344 potential papers were identified through the initial database search. After screening, seven papers that met the inclusion criteria were included in this systematic review [ 24 – 30 ]. They reported results from six randomised controlled clinical trials. All of them were included in the qualitative synthesis, and five were included in the meta-analysis [ 24 , 25 , 27 , 28 , 30 ]. The selection procedure is detailed in the PRISMA methodology flow diagram (Figure 1). J. Clin. Med. 2020,9, 590 4 of 16 J. Clin. Med. 2020, 9, 590 4 of 17 Figure 1.Flow diagram of trial selection based on PRISMA guidelines. The PEDro Scale scores ranged from six to eight,with four of the studies being classified as excellent [24,25,27,30]. The blinding of therapists and participants was the most commonly absent item among the PEDro scale criteria. All the studies reported blinding assessors and four of them reported an intention to treat analysis [24,25,27,30]. A complete description of the PEDro Scale can be found in Table 1. As it can be observed in Figure 2, no single study showed a low risk of bias on all domains. Yet, all assessed studies had an acceptable methodological quality. Records identified through database searching (n =344) Screening Included Eligibility Identification Records screened (n = 32) Records excluded with reasons (n =24) -Methodological design (n=9) -Study cases (n = 5) -Meeting abstract (n=4) -Reviews (n=4) - Proceeding papers (n=1) -Editorial material (n=1) Full-text articles assessed for eligibility (n = 7) Studies included in qualitative synthesis (n = 7) Studies included in quantitative synthesis (meta-analysis) (n = 5) Duplicates removed (n = 312) Figure 1. Flow diagram of trial selection based on PRISMA guidelines. The PEDro Scale scores ranged from six to eight, with four of the studies being classified as excellent [ 24 , 25 , 27 , 30 ]. The blinding of therapists and participants was the most commonly absent item among the PEDro scale criteria. All the studies reported blinding assessors and four of them reported an intention to treat analysis [ 24 , 25 , 27 , 30 ]. A complete description of the PEDro Scale can be found in Table 1. As it can be observed in Figure 2, no single study showed a low risk of bias on all domains. Yet, all assessed studies had an acceptable methodological quality. J. Clin. Med. 2020,9, 590 5 of 16 Table 1. PEDroscale items assessment. Section/Theme Afrasiabifar et al. [24]Karami et al. [25]Cattaneo et al. [26]Hebert et al. [27] Hebert et al. [28] Ozgen et al. [29]Tramontano et al. [30] Eligibility criteria Yes Yes Yes Yes Yes Yes Yes Random allocation Yes Yes Yes Yes Yes Yes Yes Concealled allocation Yes Yes Yes Yes No Yes Yes Baseline comparability Yes Yes Yes Yes Yes Yes Yes Blind subjects No No Yes No No No No Blind therapists No No No No No No No Blind assessors Yes Yes Yes Yes Yes Yes Yes Adequate follow-up Yes Yes Yes Yes Yes Yes Yes Intention to treat analysis Yes Yes No Yes No No Yes Between-group comparisons Yes Yes Yes Yes Yes Yes Yes Point estimates and variability Yes Yes No Yes Yes Yes Yes Total score 8/10 8/10 7/10 8/10 6/10 7/10 8/10 J. Clin. Med. 2020,9, 590 6 of 16 J. Clin. Med. 2020, 9, 590 6 of 17 Figure 2. Cochrane risk of bias tool summary. 3.2. Study Design and Population Characteristics The total sample had 321 participants who had been diagnosed with multiple sclerosis (276 women and 93 men, mean age 43.6 years old). Two of the seven studies included compared vestibular rehabilitation with no intervention [28,29] and two of them with other exercise programmes [26,30]. Moreover, three of the studies compared vestibular rehabilitation with both no intervention and other exercise programmes [24,25,27]. In general, the vestibular rehabilitation combined vestibular training plus balance exercises. The vestibular training consisted of head/eyes movement for a task in different positions and on different surfaces with open or closed eyes [26,29]. Two studies made use of the Cawthorne–Cooksey protocol [24,25]. Furthermore, Hebert et al. [28] developed their own vestibular intervention called Balance and Eye-Movement Exercises for Persons with Multiple Sclerosis (BEEMS). The active control group performed exercises based on endurance, strengthening, neuro-rehabilitation exercises, stretching, postural exercises and the Frenkel exercises [24,25]. The main characteristics of the studies included, and some additional information, are shown in Table 2. Figure 2. Cochrane risk of bias tool summary. 3.2. Study Design and Population Characteristics The total sample had 321 participants who had been diagnosed with multiple sclerosis (276 women and 93 men, mean age 43.6 years old). Two of the seven studies included compared vestibular rehabilitation with no intervention [ 28 , 29 ] and two of them with other exercise programmes [ 26 , 30 ]. Moreover, three of the studies compared vestibular rehabilitation with both no intervention and other exercise programmes [24,25,27]. In general, the vestibular rehabilitation combined vestibular training plus balance exercises. The vestibular training consisted of head/eyes movement for a task in different positions and on different surfaces with open or closed eyes [ 26 , 29 ]. Two studies made use of the Cawthorne–Cooksey protocol [ 24 , 25 ]. Furthermore, Hebert et al. [ 28 ] developed their own vestibular intervention called Balance and Eye-Movement Exercises for Persons with Multiple Sclerosis (BEEMS). The active control group performed exercises based on endurance, strengthening, neuro-rehabilitation exercises, stretching, postural exercises and the Frenkel exercises [ 24 , 25 ]. The main characteristics of the studies included, and some additional information, are shown in Table 2. J. Clin. Med. 2020,9, 590 7 of 16 Table 2. Main characteristics of the studies included. Study Design, PEDro Score Participants, Characteristics and Sample Size Intervention (VR Group) Control Group Outcome Measures Main Results Afrasiabifar et al. [24]RCT, 8/10 EG mean age(SD): 32.4 (8.1) G2 mean age SD): 32 (6.7) CG mean age (SD): 33.6 (7.3) N=72 (EG: 24, G2: 23, CG: 25), 3 drop out 64 MSRR 4 MSPP/SP Location of intervention: clinic EG: Cawthorne–Cooksey vestibular rehabilitation exercise protocol 3 session week, 60 min 12 weeks intervention G2: Frenkel exercises, 3 session per week, 60 min CG: no intervention Berg Balance Scale (BBS) EG-CG/G2: Significant differences in favour of the experimental group for BBS in comparison with CG (p=0.001) and G2 (p=0.001) G2-CG: Significant differences in favour of G2 for BBS (p=0.01) Karami et al. [25]RCT, 8/10 Fatigue: FIS EG-CG/G2: Significant differences in favour of the experimental group in comparison with the CG (p=0.001) and G2 (p=0.007) G2-CG: Significant differences in favour ofthe G2 (p=0.001) Cattaneo et al. [26] RCT 7/10 EG mean age(SD): 48.5 (11.01) CG mean age (SD): 48.2 (12.05) N=53 (EG: 25, CG: 28) 24 MSRR 3 MSPP 26 MSSP EDSS: 1–6.5 Location: clinic hospital. EG: Balance exercises with open/closed eyes on different surfaces plus head–eyes movements (vestibular cues) 15 sessions of 45 min, 3 weeks intervention CG: usualrehabilitation based on range motion, muscle force, postural changes and gait on firm surface 15 sessions of 45 min, 3 weeks intervention Balance: COP disturbances in six different conditions (eyes open/closed and sway referenced on firm and foam surfaces) EG-CG: Significant differences in favour of the experimental group for eyes closed firm surface (p=0.033, eyes opencompliant surface (p=0.01), eyes closed compliant surface (p=0.039) and sway referenced compliant surface (p=0.017) Hebert et al. [27]RCT, 8/10 EG mean age(SD): 46.8 (10.5) G2 mean age(SD): 42.6 (10.4) CG mean age (SD): 50.2 (9.2) N=38 (EG: 12, G2: 13, CG: 13) 34 MSRR 4MSSP Location: human performance laboratory and home exercises EG: upright postural control and eye movement exercises, Perform of 1–2 min each item 2 session/week, 60 min + a daily independent home exercise program 6 weeks of intervention G2: endurance of static bicycling and stretching exercise of lower limb muscles; stretches were held for 30 s Samefrequency and duration CG: no intervention Balance: posturography (SOT) Fatigue: MFIS Dizziness: DHI Exercise capacity: 6MWT Depression: BDI-II EG-CG/G2: Significant differences in favour of the experimental group in comparison with control group for balance (p=0.003), fatigue (p=0.005) and dizziness (p=0.009) and G2 for balance(p<0.001) G2-CG: No statistical differences between groups for any of the variables J. Clin. Med. 2020,9, 590 8 of 16 Table 2. Cont. Study Design, PEDro Score Participants, Characteristics and Sample Size Intervention (VR Group) Control Group Outcome Measures Main Results Hebert et al. [28]RCT, 6/10 EG Mean age(SD): 46.5 (8.8) CG Mean age(SD): 43 (10.8) N=88 (EG: 44, CG: 44), 12 drop out EDSS: 3.34–3.5 Location: human performance laboratoryand home exercises EG: BEEMS protocol, 2 s/w+daily home exercise (phase 1). Phase 2: 1 s/w +daily home exercises 14 weeks intervention CG: no intervention Balance: CDP-SOT Fatigue: MFIS Dizziness: DHI, DVAT, GST Gait: T25FW Quality of life: PDQ, SF-36 Mental and Physical component EG-CG: Significant differences in favour of the experimental group for balance, fatigue and dizziness/DHI (p<0.0001); quality of life/PDQ (p=0.0008), SF-36 MC (p=0.02) and SF-36 PC (p=0.01) A greater improvement was observedin participants with brainstem/cerebellar lesion compared with those without in CDP-SOT composite (p=0.04) and MFIS total (p=0.02) Ozgen et al. [29]RCT, 7/10 EG mean age: 42.5 CG mean age: 39.5 N=40 (EG: 20, CG: 20) 17MSRR 7MSPP 16MSSP Location: human performance laboratory EG: vestibular rehabilitation program (balance and ambulation exercises) 15–20 min, twice a day 8 weeks of intervention CG: no intervention Balance: static posturography, BBS, Romberg tests +foam, FTSTS, TUG, ABC Dizziness: DHI Gait: 6WT, DGI, FGA, Depression: BDI Qualitylife: MSQoL-54 EG-CG: Significant differences between groups for Tandem Romberg (eyes closed) and foam standing (eyes open) (p<0.05) Tramontano et al. [30]RCT, 8/10 EG mean age (SD): 50.64 (11.73) CG Mean age: 45.77 (10.91) N=30 (EG: 15, CG: 15) 7 dropout EDSS: 6–7 Location: clinic EG: conventional neurorehabilitation therapy for MS +10 min exercise for gaze stability +10 exercise for postural stability Two daily 40 min +20 min VR, 5days/week 4 weeks intervention CG: two daily 40-min 5 days/w for 4 weeks of conventional neurorehabilitation therapy for MS. Muscle stretching, postural alignment, active-assisted mobilisations and neuromuscular facilitations, balance training: progressive restrictions of support base, unstable surfaces. Balance: TBG, BBS Fatigue: FSS Gait: 2MWT, T25FW Disability: EDSS, BI EG-CG: Significant differences in favour of the experimental group were observed except for 2MWT 2MWT: Two Minute Walking Test; 6MT: Six meter walk test; 6MWT: Six-Minute Walk Test; ABC: Activities-Specific Balance Confidence; BBS: Berg Balance Scale; BDI-II: Beck Depression Inventory–II; BEEMS: Balance and Eye-Movement Exercises for Persons with MultipleSclerosis; BI: Barthel Index. CDP-SOT: Computerised Dynamic Posturography-Sensory Organisation Test; CG: control group; COP : Center of pressure; DGI: Dynamic Gait Index; DHI: Dizziness Handicap Inventory; DVAT: Dynamic Visual Acuity Test; EDSS: Expanded Disability Status Scale; EG: experimental group; FGA: Functional Gait Assessment; FIS: Fatigue Impact Scale; FTSTS: Five Times Sit-to-Stand Test; FSS: Fatigue Severity Scale; G2: secondary intervention group; GST: Gaze Stabilisation Test; MFIS: Modified Fatigue Impact Scale; min: minute; MS: multiple sclerosis; MSRR: Multiple Sclerosis relapsing remitting; MSPP: Multiple sclerosis primary progressive; MSSP: Multiple sclerosis secondary progressive; MSQoL-54: Multiple Sclerosis Quality of Life Scale–54; PDQ: Perceived Deficits Questionnaire; RCT: randomised clinical trial; s/w: session/week; SD: standard deviation; SF-36: Short Form-36 Health Status Questionnaire; SOT: SensoryOrganisation Test (posturography); T25FW: Timed 25-foot walk test; TBG: Tinetti Balance Gait scale; TUG:Timed Up-and-Go Test; VAS: Visual Analog Scale; VR: Vestibular rehabilitation. J. Clin. Med. 2020,9, 590 9 of 16 3.3. Results for Primary Outcomes 3.3.1. Balance Six of the seven studies analysed postural control or balance ability. This outcome was assessed by posturography [ 26 – 29 ] and the Berg Balance Scale [ 24 , 29 , 30 ]. Compared with no intervention, vestibular rehabilitation was more effective for postural control improvements in an upright position (SMD =2.12; 95% CI =0.49, 3.75; p=0.01; I 2 =89%) [ 27 , 28 ] (Figure 3). Ozgen et al. [ 29 ] also compared vestibular rehabilitation versus the usual care for balance training; after intervention, the experimental group obtained greater improvements than the control group, and there were also statistically significant differences between the two groups for the Tandem Romberg and foam standing tests (p<0.05). The data from this study were not pooled because the parameters of the intervention were not considered to be sufficiently homogeneous. J. Clin. Med. 2020, 9, 590 10 of 17 3.3. Results for Primary Outcomes 3.3.1. Balance Six of the seven studies analysed postural control or balance ability. This outcome was assessed by posturography[26–29] and the Berg Balance Scale [24,29,30]. Compared with no intervention, vestibular rehabilitation was more effective for postural control improvements in an upright position (SMD= 2.12; 95% CI = 0.49,3.75; p = 0.01; I2 = 89%)[27,28] (Figure 3). Ozgen et al. [29] also compared vestibular rehabilitation versus the usual care for balance training; after intervention, the experimental group obtained greater improvements than the control group, and there were also statistically significant differences between the two groups for the Tandem Romberg and foam standing tests (p<0.05). The data from this study were not pooled because the parameters of the intervention were not considered to be sufficiently homogeneous. Figure 3. Forest plot of the meta-analysis of postural control (vestibular rehabilitation versus no intervention). Comparing vestibular therapy to other exercise interventions, the meta-analysis of two studies [24,30] did not report significant differences between the groups (SMD = 4.49; 95% CI = −0.61, 9.58; p= 0.08; I2= 35%)[24,30] (Figure 4). Furthermore, Cattaneo et al. [26] also compared vestibular rehabilitation with a standard exercise programme. In this case, statistically significant differences were observed between groups in favour of the experimental intervention for upright postural control in four conditions: eyes closed and firm surface (p = 0.033), eyes open and compliant surface (p = 0.01), eyes closed and compliant surface (p = 0.039) and sway referenced and compliant surface (p = 0.017). However, it was not possible to pool the data from this study. Figure 4. Forest plot of the meta-analysis of Berg Balance Scale (vestibular rehabilitation versus other exercises). 3.3.2. Dizziness The Dizziness Handicap Inventory was assessed in three studies [27–29]. A meta-analysis was performed, comparing the vestibular rehabilitation and the no-intervention groups. Significant improvements were reported for dizziness in the intervention group (SMD= −17.43; 95% CI = −29.99, −4.87; p= 0.007; I2= 66%)[27,28] (Figure 5). Ozgen et al. [29] also assessed the Dizziness Handicap Inventory, but this study was not pooled because its intervention parameters were very different. Figure 3. Forest plot of the meta-analysis of postural control (vestibular rehabilitation versus no intervention). Comparing vestibular therapy to other exercise interventions, the meta-analysis of two studies [24,30] did not report significant differences between the groups (SMD =4.49; 95% CI =−0.61 , 9.58; p=0.08; I 2 =35%) [ 24 , 30 ] (Figure 4). Furthermore, Cattaneo et al. [ 26 ] also compared vestibular rehabilitation with a standard exercise programme. In this case, statistically significant differences were observed between groups in favour of the experimental intervention for upright postural control in four conditions: eyes closed and firm surface (p=0.033), eyes open and compliant surface (p=0.01), eyes closed and compliant surface (p=0.039) and sway referenced and compliant surface (p=0.017). However, it was not possible to pool the data from this study. J. Clin. Med. 2020, 9, 590 10 of 17 3.3. Results for Primary Outcomes 3.3.1. Balance Six of the seven studies analysed postural control or balance ability. This outcome was assessed by posturography[26–29] and the Berg Balance Scale [24,29,30]. Compared with no intervention, vestibular rehabilitation was more effective for postural control improvements in an upright position (SMD= 2.12; 95% CI = 0.49,3.75; p = 0.01; I2 = 89%)[27,28] (Figure 3). Ozgen et al. [29] also compared vestibular rehabilitation versus the usual care for balance training; after intervention, the experimental group obtained greater improvements than the control group, and there were also statistically significant differences between the two groups for the Tandem Romberg and foam standing tests (p<0.05). The data from this study were not pooled because the parameters of the intervention were not considered to be sufficiently homogeneous. Figure 3. Forest plot of the meta-analysis of postural control (vestibular rehabilitation versus no intervention). Comparing vestibular therapy to other exercise interventions, the meta-analysis of two studies [24,30] did not report significant differences between the groups (SMD = 4.49; 95% CI = −0.61, 9.58; p= 0.08; I2= 35%)[24,30] (Figure 4). Furthermore, Cattaneo et al. [26] also compared vestibular rehabilitation with a standard exercise programme. In this case, statistically significant differences were observed between groups in favour of the experimental intervention for upright postural control in four conditions: eyes closed and firm surface (p = 0.033), eyes open and compliant surface (p = 0.01), eyes closed and compliant surface (p = 0.039) and sway referenced and compliant surface (p = 0.017). However, it was not possible to pool the data from this study. Figure 4. Forest plot of the meta-analysis of Berg Balance Scale (vestibular rehabilitation versus other exercises). 3.3.2. Dizziness The Dizziness Handicap Inventory was assessed in three studies [27–29]. A meta-analysis was performed, comparing the vestibular rehabilitation and the no-intervention groups. Significant improvements were reported for dizziness in the intervention group (SMD= −17.43; 95% CI = −29.99, −4.87; p= 0.007; I2= 66%)[27,28] (Figure 5). Ozgen et al. [29] also assessed the Dizziness Handicap Inventory, but this study was not pooled because its intervention parameters were very different. Figure 4. Forest plot of the meta-analysis of Berg Balance Scale (vestibular rehabilitation versus other exercises). 3.3.2. Dizziness The Dizziness Handicap Inventory was assessed in three studies [ 27 – 29 ]. A meta-analysis was performed, comparing the vestibular rehabilitation and the no-intervention groups. Significant improvements were reported for dizziness in the intervention group (SMD = − 17.43; 95% CI = − 29.99, − 4.87; p=0.007; I 2 =66%) [ 27 , 28 ] (Figure 5). Ozgen et al. [ 29 ] also assessed the Dizziness Handicap Inventory, but this study was not pooled because its intervention parameters were very different. Intra-group significant differences were only found for the experimental group, but, despite this, there were no statistical differences between the groups. J. Clin. Med. 2020,9, 590 16 of 16 35. Murray, D.A.; Meldrum, D.; Lennon, O. Can vestibular rehabilitation exercises help patients with concussion? A systematic review of efficacy, prescription and progression patterns. Br. J. Sports Med. 2017 ,51, 442–451. [CrossRef] [PubMed] 36. Hebert, J.R.; Foster, J.E.; Sterans-Yoder, K.A.; Penzenik, M.E.; Brenner, L.A. Persistent Symptoms and Objectively Measured Balance Performance Among OEF/OIF Veterans with Remote Mild Traumatic Brain Injury. J. Head Trauma Rehabil. 2018,33, 403–411. [CrossRef] [PubMed] 37. Gurley, J.M.; Hujsak, B.D.; Kelly, J.L. Vestibular rehabilitation following mild traumatic brain injury. NeuroRehabilitation 2013,32, 519–528. [CrossRef] [PubMed] 38. Amatya, B.; Young, J.; Khan, F. Non-pharmacological interventions for chronic pain in multiple sclerosis. Cochrane Database Syst. Rev. 2018,12, CD012622. [CrossRef] 39. Rietberg, M.; Veerbeek, J.; Gosselink, R.; Kwakkel, G.; Van Wegen, E. Respiratory muscle training for multiple sclerosis. Cochrane Database Syst. Rev. 2017,12, CD009424. [CrossRef] 40. Dunlap, P.M.; Holmberg, J.M.; Whitney, S.L. Vestibular rehabilitation: Advances in peripheral and central vestibular disorders. Curr. Opin. Neurol. 2019,32, 137–144. [CrossRef] 41. Zambare, P.D.; Soni, N.; Sharma, P. Effect of Cawthorne and Cooksey Exercise Program on Balance and Likelihood of Fall in Older Women. Indian J. Physiother. Occup. Ther. An. Int. J. 2015,9, 55. [CrossRef] 42. Gaikwad, S.B.; Mukherjee, T.; Shah, P.V.; Ambode, O.I.; Johnson, E.G.; Daher, N.S. Home exercise program adherence strategies in vestibular rehabilitation: A systematic review. Phys. Ther. Rehabil. Sci. 2016 ,5, 53–62. [CrossRef] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).