Balance Training

Evidence Reviewed as of before: 09-06-2012
Author(s)*: Annabel McDermott, OT; Nicol Korner-Bitensky, PhD OT; Norine Foley, BASc; Mark Speechley, PhD; Nancy M. Salbach, PhD, PT; Maxim Ben Yakov, BSc. PT; Robert Teasell, MD
Patient/Family Information Table of contents

Introduction

Balance problems are caused by motor, sensory and cognitive impairments and are one of the most common issues after stroke. Impaired postural control contributes to difficulties with recovery of mobility and functional independence among patients with stroke. Most rehabilitation therapies aim for the restoration of balance in sitting, as well as in standing, reaching, and rising to stand.

Additional support from undergraduate students, School of Physical and Occupational Therapy, McGill University: Natasha Alloul, Julie Parent -Taillon, Nadia Boule-Laghzali, Genevieve Larivee, Ang Li, Zahra Adl-Zarabi, Michael Dyck

Patient/Family Information

Author: Maxim Ben Yakov, BSc. PT

What is balance training?

To sit and to walk safely you need to have good balance. Balance training focuses on practicing and improving the body’s ability to perform coordinated movement (of arms and legs) while maintaining a balanced posture, i.e. without falling, stumbling, or feeling wobbly. This is usually achieved through rehearsal of tasks, such as reaching for objects while holding the body straight. Training in sitting and standing should be initiated as soon as possible after a stroke, as these are basic, necessary tasks in daily life.

Why train balance after a stroke?

Balance is a basic requirement for active, independent, and safe movement of our bodies in daily life. Before your stroke, you probably balanced your body when sitting and standing automatically, without thinking about it. After a stroke, you may have balance problems that require you to concentrate a great deal to do simple things, such as putting on your socks, or standing at a sink to brush your teeth. Even people who experience only small problems with balance may have difficulty when walking outside on uneven ground or when crossing the street.

Are there different kinds of balance training?

Yes, there are different ways to retrain balance after a stroke.

  • Functional balance training: Recently, balance training has been focusing more on functional, task-specific training. In functional training, the individual who has had a stroke works on typical tasks that people perform in their daily lives, such as reaching into a cupboard for a cup or plate, or trying to carry a grocery bag.
  • Body weight support: After a stroke, some individuals are too weak and have difficulty sitting, standing, or walking in therapy. If this is the case, your body weight may be supported while you stand or walk either  by your therapist or by a body harness.
  • Hydrotherapy: Sometimes, balance training is done in a therapeutic pool, using a technique called Hydrotherapy. Water makes your limbs lighter, since you are not moving against gravity. Water also provides support and stimulation so that you can work on your balance in a safe environment. Your therapist will usually work in the water with you to make sure that you are well supported and safe.
  • Proprioception training: Balance training can also include something called proprioception training, which can help you to be aware of where your arms and legs are in space. For example, after a stroke some people have difficulty knowing where their hand is when their eyes are closed. Proprioception is important to achieve proper balance, and the good news is that as we work on improving balance, we are training proprioception as well.Other types of balance training you might hear about are:
  • “Bobath approach”: Bobath was a physiotherapist who developed a treatment approach that analyzes and interprets how you move after your stroke. After a stroke, many people move in a way that is different from before. Your therapist will work on training and modifying your movements to help you accomplish daily tasks. Usually a therapist will guide your arms, legs or trunk through the correct movements so that you can re-learn to do the activities correctly.
  • “Visual feedback” or “Biofeedback for trunk control”: This technique uses a mirror in front of you or a video camera system to track your body, arms, or legs while doing activities like catching a ball or placing objects on a shelf. This allows you to see how you are moving so that you can try to correct your movements.
  • “Vision-deprived training”: With your eyes covered, your therapist will help you do activities like standing on one or both legs, trying to sit on a pillow, or simply getting up from a chair and sitting down. This challenges your balance more than when your eyes are open. This is an activity you should try doing as you get better.
  • “Independent practice”: You can work on your balance on your own. For example, during your independent exercise, you could have as a goal to stand on both legs with equal weight, or to try and sit on both buttocks with equal pressure.
    NOTE: You should only try this once your therapist tells you that it is safe for you to do so.
  • “Balance biofeedback”: After a stroke, it is typical to put more weight on your “good” leg when you are standing. However, it is important that you also put weight on your weaker leg. While you are standing, your therapist will use a computer screen with a special mat that will sense how much pressure goes through each foot. The amount of weight put through your weaker leg will then be recorded and will show up on the computer screen. Training in this way gives you immediate feedback about how well you are doing. At first, the goal may be to increase the amount of weight you put on your weaker leg. Next, it may be to put an equal amount of weight on both legs while standing. Eventually, you may try to put more weight on your weaker leg. This is important because as we walk, we need to put our body weight through one leg at a time.
  • “Perceptual training”: This technique focuses on training the awareness of your arms, legs, and trunk in space. For example you might be asked to touch your knee and then your forehead while your eyes are closed.
  • “Multisensorial Training”: Following a stroke, you may become overly reliant on visual cues to help maintain your balance. Multisensorial training is a form of rehabilitation conducted while restricting the amount that you see. It focuses on the amount and intensity of your movements and exercise without placing emphasis on how well you perform them.

Does balance training work after a stroke?

Researchers have done experiments to see if balance training helps people who have had a stroke.

  • Task-oriented interventions: One high quality study looked at task-oriented interventions for walking. The results showed that this treatment can improve a person’s confidence in balance.
  • Perceptual exercises: After a stroke, it is common to have more body sway, and this makes you more unsteady on your feet. In one high quality study, results showed that perceptual exercises reduced the amount of body sway.
    NOTE: Even without a stroke, everyone has a certain normal amount of body sway that we are not aware of.
  • Bobath Therapy Approach: One high quality study showed that the Bobath approach did not improve independence in normal daily living, sitting balance, standing balance, or the amount of weight put on the weaker leg.
  • Task-specific reaching training: One high quality study found that task-specific reaching does not improve how evenly you distribute your body weight through both buttocks when sitting. The same study results showed that such training does not improve how equally you put your body weight through both feet while standing.
  • Independent-practice training: There is limited research from one fair quality study that showed that when independent-practice training is combined with therapy based on the Bobath approach it does not improve balance after a stroke.
  • Visual feedback training: There is limited research based on two fair quality studies suggesting that visual feedback training does not result in improvements in balance. It is worth noting that one study did find important gains in the ability to perform self-care activities (such as washing, toileting, dressing, and grooming).
  • Balance biofeedback training: There are conflicting findings in this area. Three fair quality studies found no real gains in balance after using this training method. In contrast, two high quality studies on balance biofeedback training found that balance did improve after a stroke. Another high quality study demonstrated that biofeedback for trunk control training can improve significantly standing balance (not when walking or reaching).
  • Multisensorial Training: One high quality study found that multisensorial training (a form of therapy conducted while restricting what you see and focusing on the amount and intensity of movement and exercise) is not more effective than neurodevelopmental therapy (a form of therapy which focuses on quality of movement and exercise) in improving standing balance. However, it is more effective at improving your balance when walking and moving around, as well as increasing independence in functional activities and improving quality of life.

Side effects/risks?

Balance is important to prevent you from falling. During balance training, you should always be supervised by an individual who knows about practicing balance training safely. Eventually, you will probably begin practicing balance exercises with your family or friends. Before you do so, your therapist should show them safe ways of working with you.

Who provides the treatment?

Balance training should be performed or supervised by a trained health professional. A variety of health professionals provide balance training as part of their treatment, including occupational therapists, physical therapists, and exercise therapists.

Clinician Information

Note: When reviewing the findings, it is important to note that they are always made according to randomized clinical trial (RCT) criteria – specifically as compared to a control group. To clarify, if a treatment is “effective” it implies that it is more effective than the control treatment to which it was compared. Non-randomized studies are no longer included when there is sufficient research to indicate strong evidence (level 1a) for an outcome.

Of the 42 studies included in this module that have investigated interventions to improve balance post-stroke, 25 are high quality randomized controlled trials. Interventions reviewed in this module include aquatic therapy, Bobath therapy, force platform or mechanical balance training devices, multisensory training, perceptual exercises, task-specific exercises, trunk exercises, vibration therapy and virtual reality. Although a majority of the studies demonstrated a positive benefit of balance training, the heterogeneity of intervention and outcomes measures does not allow us to make definite conclusions regarding any one most efficacious intervention method for balance re-training post-stroke.

Lubetzky-Vilnai & Kartin (2010) conducted a systematic review of recent studies on balance training interventions that comprised 22 RCTs, pilot studies and case series from January 2006 to February 2010. Comparison among studies was limited by diversity in balance training interventions (type, duration, intensity and progression), control interventions (conventional physiotherapy, standard interdisciplinary care, conventional gait training, patient-initiated training, body-weight supported training, standard rehabilitation and neurodevelopmental treatment) and outcomes measured. Most studies reported that balance training programs were not more effective than control therapies as both experimental and control groups demonstrated improved balance following intervention.

An & Shaughnessy (2011) conducted a systematic review of exercise interventions used to improve balance and/or gait following stroke. The authors reviewed 17 English-language RCTs published from 2001 to January 2010, 10 of which included balance as an outcome (5 of these were also included in the 2010 systematic review by Lubetzky-Vilnai & Kartin). This systematic review concluded that multisensory programs do not seem to be effective in improving balance following stroke. However, early initiation of exercise after stroke was reported to be effective in improving balance, and aerobic exercise was positively associated with improved balance in subacute and chronic stroke. The systematic review concluded that exercise performed for at least 20-60 minutes, 3-4 times a week for 6-12 weeks can improve balance following stroke.

Results Table

View results table

Outcomes

Aquatic therapy
Effective
2A

One fair quality RCT (Noh et al., 2008) has investigated the effectiveness of aquatic therapy in improving balance in patients with stroke.

The fair quality RCT (Noh et al., 2008) randomized patients with chronic stroke and unilateral limb weakness to an aquatic therapy program or a conventional gym exercise program. At one month post-treatment there were significant between-group differences, with those in the aquatic group having better balance (Berg Balance Scale) and weight-bearing ability on the affected side (vertical ground reaction force during forward and backward weight-shift). There was no significant group difference in weight-bearing ability during sit-to-stand or lateral weight-shift.

Conclusion: There is limited evidence (level 2a) from 1 fair quality RCT that aquatic therapy is more effective than a gym exercise program for improving balance, forward/backward weight shift and knee flexor strength of the affected limb in patients with stroke.

Note: however, the fair quality RCT found no significant difference between groups in weight shift during lateral movements and when rising from a chair, or strength of knee extensor and trunk muscles.

Bobath therapy
Effective
2a

One fair quality RCT (Mudie et al., 2002) has investigated the use of the Bobath approach in improving balance following stroke.

The fair quality study (Mudie et al., 2002) randomly assigned patients with acute stroke to one of four treatment groups: (1) task-specific reaching; (2) Bobath therapy interventions; (3) BPM biofeedback interventions; or (4) conventional physiotherapy and occupational therapy (control). The Bobath group demonstrated a significant improvement in seated symmetry of weight distribution at post-treatment (2 weeks), although results did not remain significant at follow-up time point (12 weeks). At 12 weeks post-study 29% of the Bobath group were able to distribute weight to both sides, in comparison to the BPM group (83%), task-specific group (38%) and the control group (0%).

Conclusion: There is limited evidence (level 2a) from 1 fair quality RCT that Bobath therapy is effective for improving balance (seated weight distribution) following stroke. However, between-group differences were not reported.

Cycling training
Effective
1B

One high quality RCT (Katz-Leurer et al., 2006) has investigated the effectiveness of cycling training in improving balance in patients with stroke.

The high quality RCT (Katz-Leurer et al., 2006) randomized patients with subacute stroke to receive cycling training and conventional rehabilitation or conventional rehabilitation training alone. At post-treatment (6 weeks) there were significant group-time interactions in favour of the cycling group compared to the control group for balance on the Postural Assessment Scale for Stroke Patients total, static and dynamic scores. However, no significant between-group differences were seen for standing balance as measured using the Standing Balance Test.

Conclusion: There is moderate evidence (level 1b) from 1 high quality RCT that cycling training is more effective than conventional rehabilitation alone for improving balance in patients with stroke.

Note: However, the high quality RCT showed no significant between-group differences on the Standing Balance test.

Force platform training
Effective
1a

Two high quality RCTs (Sackley & Lincoln, 1997; Cheng et al., 2001), 6 fair quality RCTs (Shumway-Cook et al., 1988;Wong et al., 1997; Grant et al., 1997; Walker et al., 2000; Chen et al., 2002; Mudie et al., 2002) and 1 poor quality RCT (Geiger et al., 2001) have investigated the effect of force platform biofeedback training on balance following stroke.

The first high quality RCT (Sackley & Lincoln, 1997) randomized patients with subacute or chronic stroke to receive balance training using the Nottingham Balance Platform with visual feedback regarding weight distribution and weight shift activity, or a placebo balance intervention. Balance measures were taken at baseline, 4 weeks (post-treatment) and 12 weeks (follow-up) using assessment of stance symmetry and sway. Significant between-group differences were seen in favour of the treatment group compared to the placebo group for stance symmetry at post-treatment, but these differences did not remain significant at follow-up. There were no significant differences in sway.

The second high quality RCT (Cheng et al., 2001) assigned patients with hemiplegia following stroke to balance training using a dual force platform standing biofeedback trainer with visual and auditory feedback, or conventional physical therapy. Significant between-group differences were noted in favour of the force platform group compared to the control group on balance measures of mediolateral sway, rate of rise in force when rising from a chair, and frequency of falls post-stroke. There were no significant between-group differences in sit-to-stand or stand-to-sit performance at post-treatment, but at 6-month follow-up a significant difference in sit-to-stand performance was reported in favour of the force platform group compared to the control group.

The first fair quality RCT (Shumway-Cook et al., 1988) randomly assigned individuals with subacute stroke to standing balance retraining using a static force platform biofeedback device, or standing balance training without biofeedback. A significant between-group difference in lateral sway displacement was reported in favour of the force platform group compared to the control group. There was no significant difference in total sway area between groups.

The second fair quality RCT (Wong et al., 1997) randomised patients with acute stroke to training using a Standing Biofeedback Training (SBT) device that provides real-time visual and auditory weight bearing biofeedback, or a Standing Training Table (STT) worktable. Significant between-group differences in postural symmetry were seen at week 1, week 2 and week 4, in favour of the SBT group compared to the STT group. Note: significant between-group differences in postural symmetry were not seen at day 1 or week 3.

The third fair quality RCT (Grant et al., 1997) reported preliminary findings from a study by Walker et al., 2000 (see below), whereby 16 patients with stroke were randomly assigned to visual biofeedback balance training and physiotherapy, or standard balance training and physiotherapy. No significant differences were noted on measures of balance (Berg Balance Scale, postural sway, standing symmetry).

The fourth fair quality RCT (Walker et al., 2000) randomly assigned 54 patients with stroke to one of three treatment groups: (1) balance training using the dual force platform Balance Master with visual feedback and conventional physiotherapy and occupational therapy; (2) ‘standard’ balance training and conventional physiotherapy and occupational therapy; or (3) conventional physiotherapy and occupational therapy alone. There were no significant differences on measures of balance (Berg Balance Scale, postural sway) when comparing either intervention group with the control group.

The fifth fair quality RCT (Chen et al., 2002) randomly assigned patients with stroke to balance training using the Smart Balance Master with visual feedback in combination with conventional physical and occupational therapy, or physical and occupational therapy alone. Significant differences were observed between groups on one of three measures of static balance (absence of sway but not maximum stability or center of gravity alignment) and all three measures of dynamic balance (axis velocity, directional control, end-point excursion).

The sixth fair quality RCT (Mudie et al., 2002) randomly assigned individuals with recent stroke to one of four treatment groups: (1) task-specific reaching; (2) Bobath therapy interventions; (3) Balance Performance Monitor (BPM) weight-distribution training in sitting and standing with visual feedback; or (4) conventional physiotherapy and occupational therapy (control). The BPM group demonstrated a significant improvement in seated symmetry of weight distribution at post-treatment (2 weeks) but these results did not remain significant at follow-up time points (4 weeks, 12 weeks). At 12 weeks post-study 83% of the BPM group was distributing weight to both sides, as compared to 38% of the task-specific reaching group, 29% of the Bobath group and 0% of the control group. The BPM group also demonstrated some generalization of symmetry training in sitting to standing.

The poor quality study (Geiger et al., 2001) assigned patients with stroke to balance training using the forceplate Neurocom Balance Master with visual feedback, or regular balance training. There were no significant differences in balance (Berg Balance Scale) at post-treatment (4 weeks).

Conclusion: There is strong evidence (level 1a) from 2 high quality RCTs and 3 fair quality RCTs that force platform biofeedback training is more effective than control therapies for improving balance (e.g. symmetry, mediolateral sway, dynamic balance, frequency of falls) following stroke.

Note: However, numerous studies reported that force platform biofeedback training was not more effective than control therapies for improving other measures of balance (e.g. Berg Balance Scale, postural sway, seated weight-distribution).

Independent practice
Not Effective
2A

One fair quality RCT (Pollock et al. 2002) has investigated the efficacy of independent-practice training for improving balance post-stroke.

The fair quality RCT (Pollock et al. 2002) randomly assigned patient with stroke to 1 of 2 treatment groups: (1) Independent practice with balance-focused exercise in combination with conventional therapy; or (2) conventional therapy alone (control). No significant difference was found between groups. It was concluded that performance of postural control and weight distribution did not increase for individuals in either group post-stroke.

Conclusion: There is limited evidence (level 2a) from 1 fair quality study that independent-practice training with conventional therapy is not more effective than conventional therapy alone for improving balance post-stroke.

Mechanical balance training devices
Not Effective
1b

One high quality RCT (Goljar et al., 2010) and one controlled clinical trial (Byun et al., 2011) have investigated the effect of a mechanical balance training device on balance in patients with stroke.

The high quality RCT (Goljar et al., 2010) randomized patients with subacute or chronic stroke to receive physiotherapy and balance training using a mechanical device, or physiotherapy and conventional balance training. There was no significant between-group difference in balance (Berg Balance Scale; one-leg standing) at post-treatment.

The controlled clinical trial (non-randomized crossover design) (Byun et al., 2011) divided patients with chronic stroke into two groups to receive conventional rehabilitation in addition to balance training using a sliding rehabilitation machine for 2 weeks (experimental period), preceded (group B) or followed (group A) by 2 weeks of conventional rehabilitation alone (control period). There was a significant difference in favour of the experimental period as compared to the control period in improving balance (Berg Balance Scale).

Conclusion: There is moderate evidence (level 1b) from 1 high quality RCT that a mechanical balance training device is not more effective than physiotherapy and conventional training for improving balance among patients with subacute and chronic stroke.

Note: However, 1 controlled clinical trial found that a balance trainer (sliding rehabilitation machine) is more effective than conventional rehabilitation alone in improving balance among patients with chronic stroke. Variations in the type of mechanical balance training devices used may account for differences in results between studies.

Multisensory training
Effective
1B

Two high quality RCTs (Yelnik et al., 2008; Gok et al., 2008), one fair quality RCT (Bayouk et al., 2006) and one poor quality RCT(Onigbinde et al., 2009) have investigated the use of multisensorial training on balance in patients with stroke.

The first high quality RCT (Yelnik et al., 2008) randomised patients with subacute and chronic stroke to receive multisensorial therapy or neurodevelopmental therapy (control group). No significant between-group differences in balance (Berg Balance Scale*; self reported perception of security) were seen at post treatment (30 days) or follow-up (60 days). However, a significant between-group difference in dynamic balance (percentage of double-limb stance time) was seen at follow-up, in favour of the multisensorial group compared to the control group.

*Note: Differences in standing balance may not have been detected due to the ceiling effect of the Berg Balance Scale. The authors also questioned the clinical meaning of the improvements to the patients due to the small values.

The second high quality RCT (Gok et al., 2008) randomized patients with chronic stroke to receive balance training using a kinaesthetic ability training (KAT) device and conventional rehabilitation, or conventional rehabilitation alone. The KAT device held a centrally-pivoted balance platform with a pressure bladder that allowed adjustments in weight shift in response to visual feedback. At 4 weeks (post-treatment) the KAT group showed significantly greater improvement in balance (Fugl-Meyer Stroke Assessment [FMA] balance subscore; KAT static and dynamic balance indices) than the control group.

The fair quality RCT (Bayouk et al., 2006) randomised hemiparetic patients with chronic stroke to a task-oriented exercise program with manipulation of sensory input (eyes open/closed; soft/firm surface) or a task-oriented program under normal conditions. Outcomes were taken as a measure of the center of pressure (COP) displacement during double-legged stance and sit-to-stand with eyes open or closed and on normal or soft surfaces, as well as the 10-m walking test. Although between-group differences were not reported*, a significant difference in pre- and post-test balance (COP displacement during double-leg stance with eyes open on normal and soft surfaces) was seen for the experimental group but not the control group. Both groups demonstrated a significant difference in pre- and post-test results in other balance measures (COP displacement during sit-to-stand with eyes open on a soft surface) and walking speed (10-m walking test).

*Note: as between-group differences are not reported, this study is not used to determine level of evidence regarding the effectiveness of multisensorial training in the conclusion below.

The poor quality RCT (Onigbinde et al., 2009) randomised patients with stroke (time since stroke not specified) to perform wobble board exercises with visual feedback and conventional physiotherapy, or conventional physiotherapy alone. At post-treatment (6 weeks) there were significant between-group differences in static balance (eyes closed) and dynamic balance (Four Square Step Test time), in favour of the experimental group compared to the control group. There were no significant between-group differences in static balance (eyes open) at post-treatment.

Conclusion: There is moderate evidence (level 1b) from 1 high quality RCT and 1 poor quality RCT that multisensorial training is more effective than conventional rehabilitation for improving balance following stroke. One fair quality RCT also reported a significant improvement in balance following multisensory training.

Note: One high quality RCT reported no significant difference in balance between multisensory training and conventional rehabilitation, although noted a potential ceiling effect of the instrument used to measure balance (Berg Balance Scale). The same RCT saw a significant between-group difference in dynamic balance at follow-up (but not at post-treatment), in favour of the multisensorial training group.

Perceptual exercises
Effective
1B

One high quality RCT (Morioka et al. 2003) has investigated the use of perceptual training for balance retraining post-stroke.

The high quality RCT (Morioka et al., 2003) randomly assigned patients with stroke to receive rehabilitation including perceptual learning exercises or standard rehabilitation (control). Significant improvements were reported in the length, the enveloped area and the rectangular area of the parameter of postural sway in the group receiving the perceptual exercises compared to the control group.

Conclusion: There is moderate evidence (level 1b) from 1 high quality RCT that perceptual exercises are more effective than standard rehabilitation for improving balance measures post-stroke.

Speed-dependent treadmill training
Not Effective
1B

One high quality RCT (Lau et al., 2011) has investigated the use of speed-dependent treadmill training for improving balance following stroke.

The high quality RCT (Lau et al., 2011) randomised patients with subacute stroke to a speed-dependent treadmill training group or a steady-speed treadmill training group. At post-treatment (10 x 30-minute sessions) there was no significant between-group difference in balance (Berg Balance Scale).

Conclusion: There is moderate evidence (level 1b) from 1 high quality RCT that speed-dependent treadmill training is not more effective than steady-speed treadmill training for improving balance following stroke.

Standing practice
Not Effective
1b

One high quality RCT (Allison & Dennett, 2007) has investigated the effectiveness of standing practice to improve balance among patients with stroke.

The high quality RCT (Allison & Dennett, 2007) randomized patients with acute or subacute stroke to receive standing practice and conventional physiotherapy or physiotherapy alone. There were no significant between-group differences in balance or trunk control (Berg Balance Scale; Trunk Control Test; Rivermead Motor Assessment – Gross Functional Tool Section) at weeks 1, 2 or 12.

Note: However, a significant difference in change in Berg Balance Scale scores from week 1 to week 12 was seen in favour of the standing practice group compared to the control group.

Conclusion: There is moderate evidence (level 1b) from 1 high quality RCT that standing practice is not more effective than conventional physiotherapy alone for improving balance and trunk control in patients with acute or subacute stroke.

Tai chi
Effective
1B

One high quality RCT (Au-Yeung et al., 2009) has investigated the effect of tai chi on balance in patients with stroke.

The high quality RCT (Au-Yeung et al., 2009) randomized patients with chronic stroke to a tai chi group or a control group that performed general exercises for breathing, stretching, mobilization, memory and reasoning. Dynamic standing balance was measured by center of gravity (COG) excursion during self-initiated body leaning forward, backward and towards the affected and non-affected sides using the Limit of Stability test; and standing equilibrium was measured by the Sensory Organization test. A significant between-group difference was seen in COG excursion amplitude when leaning forward, backward and toward the nonaffected side from week 6 (mid-treatment), and also toward the affected side from week 12 (post-treatment), in favour of the tai chi group compared to the control group. These results were maintained at 18 weeks (follow-up). A significant between-group difference in reaction time during voluntary weight shift towards the non-affected side was seen at 12 weeks and 18 weeks in favour of the tai chi group compared to the control group. There was a significant difference in standing equilibrium with vestibular input at 12 weeks (post-treatment), in favour of the tai chi group compared to the control group.

Conclusion: There is moderate evidence (level 1b) from 1 high quality RCT that tai chi is more effective than regular exercises for improving balance in patients with chronic stroke.

Task-oriented walking
Not Effective
1A

Five high quality RCTs (Richards et al., 1993; McClellan & Ada, 2004; Salbach et al., 2004 and Salbach et al., 2005; Marigold et al., 2005; Outermans et al., 2010), 1 fair quality RCT (Dean et al., 2000) and 1 quasi-experimental study (Rose et al., 2011) have investigated the use of task-oriented interventions targeting walking for balance retraining post-stroke.

The first high quality RCT (Richards et al., 1993) randomised patients with acute stroke to receive intensive gait-focused task-oriented physical therapy, or one of two control groups that received different intensities of standard physical therapy. No significant difference in balance (Berg Balance Scale, Fugl Meyer Assessment balance subscale) was seen between groups at 6 weeks (post-treatment) or 3 months (follow-up).

The second high quality RCT (McClellan & Ada, 2004) randomised patients with chronic stroke to receive home-based task-oriented mobility training or home-based task-oriented upper extremity training exercises. There was a significant between-group difference in standing balance (Functional Reach Test) at 6 weeks (post-treatment) and at 2-month follow-up, in favour of the task-oriented mobility group compared to the task-oriented upper extremity group.

The third high quality RCT (Salbach et al., 2004; Salbach et al., 2005) randomised patients with subacute or chronic stroke to receive task-oriented mobility training or task-oriented upper extremity training. There was a significant between-group difference in balance confidence (Activities-specific Balance Confidence scale), but not balance (Berg Balance Scale) at 6 weeks (post-treatment), in favour of the task-oriented mobility group compared to the task-oriented upper extremity group.

The fourth high quality RCT (Marigold et al., 2005) randomised patients with chronic stroke to receive a task-oriented mobility training program or a program that emphasized slow stretching and weight-shift. No significant differences in balance (Berg Balance Scale), balance confidence (Activities-specific Balance Confidence Scale) or falls (unforced falls during reaching transferring; induced falls during platform translation) were seen at 10 weeks (post-treatment) or 1-month follow-up.

The fifth high quality RCT (Outermans et al., 2010) randomised patients with subacute stroke to receive high intensity task-oriented mobility training or low intensity standard therapy. No significant between-group differences in balance (Berg Balance Test; Functional Reach Test) were seen at 4 weeks (post-treatment).

The fair quality RCT (Dean et al., 2000) randomised patients with chronic stroke and residual hemiplegia to receive task-oriented mobility training or task-oriented upper extremity training. A significant between-group difference in balance during stepping (Step Test) was seen in favour of the task-oriented mobility group compared to the upper extremity group at 4 weeks (post-treatment) and 2 months (follow-up).

The quasi-experimental study (Rose et al., 2011) assigned patients with acute stroke to receive task-oriented mobility training or conventional rehabilitation. No significant between-group difference in balance (Berg Balance Scale) was seen at hospital discharge (post-treatment).

Conclusion 1 (balance): There is strong evidence (level 1a) from 4 high quality RCTs and 1 quasi-experimental study that task-oriented mobility training is not more effective than control therapies (conventional rehabilitation, physiotherapy) for improving balance following stroke.

Note: however, 1 high quality RCT found a significant difference in standing balance, and 1 fair quality RCT found a significant difference in stepping balance, in favour of task-oriented mobility training compared to control therapies.

Conclusion 2 (balance confidence): There is conflicting evidence (level 4) between 2 high quality RCTs regarding the effectiveness of task-oriented mobility training for improving balance confidence following stroke.

One fair quality RCT (Mudie et al. 2002) has investigated the use of task-specific reaching for balance retraining post-stroke.

The fair quality RCT (Mudie et al., 2002) randomly assigned patients with recent stroke to one of four treatment groups: (1) task-specific reaching; (2) Bobath therapy interventions; (3) BPM biofeedback interventions; or (4) conventional physiotherapy and occupational therapy (control). The task-specific reaching group did not demonstrate a significant improvement in seated weight distribution at post-treatment (2 weeks) or follow-up time points (4 weeks, 12 weeks), whereas all other groups demonstrated significantly improved sitting symmetry at post-treatment. At 12 weeks post-study, 38% of the task-specific reaching group were distributing weight to both sides, as compared to 83% of the BPM group, 29% of the Bobath group and 0% of the conventional therapy group.

Conclusion: There is limited evidence (level 2a) from 1 fair quality RCT that task specific reaching is not effective for improving balance post-stroke. However, between-group differences were not reported.

Task-specific reaching
Not Effective
2A

One fair quality RCT (Mudie et al. 2002) has investigated the use of task-specific reaching for balance retraining post-stroke.

The fair quality RCT (Mudie et al., 2002) randomly assigned patients with recent stroke to one of four treatment groups: (1) task-specific reaching; (2) Bobath therapy interventions; (3) BPM biofeedback interventions; or (4) conventional physiotherapy and occupational therapy (control). The task-specific reaching group did not demonstrate a significant improvement in seated weight distribution at post-treatment (2 weeks) or follow-up time points (4 weeks, 12 weeks), whereas all other groups demonstrated significantly improved sitting symmetry at post-treatment. At 12 weeks post-study, 38% of the task-specific reaching group were distributing weight to both sides, as compared to 83% of the BPM group, 29% of the Bobath group and 0% of the conventional therapy group.

Conclusion: There is limited evidence (level 2a) from 1 fair quality RCT that task specific reaching is not effective for improving balance post-stroke. However, between-group differences were not reported.

Thermal therapy
Not Effective
1B

One high quality RCT (Chen et al., 2011) investigated the effect of thermal therapy on balance in patients with stroke.

The high quality RCT (Chen et al., 2011) randomized patients with acute stroke to receive thermal stimulation and conventional rehabilitation or conventional rehabilitation alone. No significant between-group differences in balance (Berg Balance Scale; Postural Assessment Scale for Stroke Trunk Control) were seen post-treatment (6 weeks).

Conclusion: There is moderate evidence (level 1b) from 1 high quality RCT that thermal therapy is not more effective than conventional rehabilitation for improving balance in patients with acute stroke.

Trunk exercises
Conflicting
4

Five high quality RCTs (de Seze et al., 2001; Howe et al., 2005; Verheyden et al., 2009; Karthikbabu et al., 2011; Saeys et al., 2011) have investigated the effect of trunk exercises on balance in patients with stroke.

The first high quality RCT (de Seze et al., 2001) randomly assigned patients with stroke to receive trunk control training with visual and auditory feedback using the Bon Saint Come device, or conventional rehabilitation training. Significant between-group differences were noted for balance (Upright Equilibrium Index; Trunk Control Test) at 30 days (post-treatment) and remained at 90 day follow-up (Upright Equilibrium Index only). No significant between-group differences were found for sitting balance (Sitting Equilibrium Index) at post-treatment or follow-up.

The second high quality RCT (Howe et al., 2005) randomized patients with acute stroke to receive lateral weight transference training in sitting and standing and conventional rehabilitation, or conventional rehabilitation alone. There were no significant between-group differences in balance (weight displacement during lateral reaching in sitting and standing; time taken to transfer from sit-to-stand and stand-to-sit) at post-treatment (4 weeks) or follow-up (8 weeks).

The third high quality RCT (Verheyden et al., 2009) randomized patients with acute and subacute stroke to receive conventional rehabilitation and individualized trunk exercises (intervention group) or conventional rehabilitation alone (control group). At 5 weeks (post-treatment) there was a significant between-group difference in dynamic sitting balance (Trunk Impairment Scale dynamic sitting balance), in favour of the intervention group compared to the control group. There were no significant between-group differences on other measures of balance and coordination (Trunk Impairment Scale total, static sitting balance, coordination).

The fourth high quality RCT (Karthikbabu et al., 2011) randomised patients with acute stroke to perform trunk exercises on an unstable surface (intervention group) or on a stable surface (control group), in addition to conventional physiotherapy. At 3 weeks (post-treatment) there were significant between-group differences in balance and coordination (Trunk Impairment Scale [TIS] total, dynamic sitting balance, coordination; Brunel Balance Assessment [BBA] total and stepping subtest scores), in favour of the intervention group compared to the control group. However, no significant between-group differences were seen for static sitting balance (TIS) or standing balance (BBA).

The fifth high quality RCT (Saeys et al., 2011) randomised patients with acute or subacute stroke to receive conventional training and exercises to improve trunk muscle strength, coordination and movement (intervention group), or conventional training and sham treatment that comprised passive mobilization and TENS to the hemiplegic shoulder (control group). At 8 weeks (post-treatment) there were significant differences in balance (Berg Balance Scale; Four Test Balance Scale; Tinetti Test; Trunk Impairment Scale [TIS] total, dynamic sitting, coordination). No significant between-group differences were seen in static sitting balance (TIS static sitting) or proprioception (Romberg tests).

Conclusion: There is conflicting evidence (level 4) among 5 high quality RCTs regarding the effect of trunk exercises on balance following stroke. While studies reported better outcomes than conventional rehabilitation on several balance measures (e.g. Berg Balance Scale, Brunel Balance Assessment, Trunk Control Test, Tinetti Test, Four Test Balance Scale, Upright Equilibrium Index, Trunk Impairment Scale – dynamic balance), all studies also reported no significant between-group differences on other balance measures (Sitting Equilibrium Index, Trunk Impairment Scale – static sitting balance, Romberg tests, weight displacement during lateral reaching, and time taken to transfer between sit and stand).

Note: Most participants were in the acute or subacute stages of stroke recovery. Variation in outcome measures used, as well as the type, frequency and duration of trunk exercises, are likely to account for these differences in outcomes between studies.

Vibration therapy
Not Effective
1B

One high quality RCT (van Nes et al., 2006) and 1 fair quality RCT (Merkert et al., 2011) have investigated the effect of vibration therapy on balance in patients with stroke.

The first high quality RCT (van Nes et al., 2006) randomized patients with subacute stroke to receive whole-body vibration using a commercially-available platform vibration device, or exercise therapy on music at the same frequency and duration (sham stimulation). No significant between-group differences in balance (Berg Balance Scale; Trunk Control Test) were seen at 6 weeks (post-treatment) or 12 weeks (follow-up).

The fair quality RCT (Merkert et al., 2011) randomized patients with stroke to receive balance training and vibration therapy using a Vibrosphere® vibrating platform, or conventional rehabilitation alone. There were no significant differences between groups for balance (Berg Balance Scale) at 15 days (post-treatment).

Conclusion: There is moderate evidence (level 1b) from 1 high quality RCT and 1 fair quality RCT that vibration therapy is not more effective than control therapies (e.g. conventional therapy, sham stimulation) for improving balance following stroke.

Virtual reality
Not Effective
2A

One fair quality RCT (Yang et al., 2011) has investigated the effect of virtual reality training programs on balance in patients following stroke.

The fair quality RCT (Yang et al., 2011) randomised patients with chronic stroke to receive virtual reality treadmill training or traditional treadmill training. Virtual reality treadmill training comprised level walking while watching interactive scenes that simulated turns, whereas the control group completed level treadmill walking with a garden view. At post-treatment (3 weeks) there was a significant between-group difference in center of pressure (COP) displacement in the medial-lateral direction during quiet stance. There were no other significant differences between groups in measures taken in quiet stance or sit-to-stand transfers (COP displacement in the anterior-posterior direction, COP total path excursion, COP sway area, symmetry index, COP path excursion under the paretic limb) or during level walking (stance time of the paretic limb, contact area of the paretic foot, and number of steps of the paretic limb).

Conclusion: There is limited evidence (level 2a) from 1 fair quality RCT that virtual reality treadmill training is not more effective than traditional treadmill training for improving static or dynamic balance in patients following stroke.

Note: However, the fair quality RCT did find a significant between-group difference in COP displacement in the medial-lateral direction during quiet stance, in favour of the virtual reality treadmill training group.

Vision-deprived training
Effective
1B

One high quality RCT (Bonan et al. 2004) has investigated the efficacy of vision-deprived training for the improvement of balance post-stroke.

The high quality RCT (Bonan et al., 2004) randomly assigned patients with stroke to receive vision-deprived training (intervention group) or free vision training (control group). At post-treatment significant between-group differences were noted in balance (Sensory Organization Test), in favour of the intervention group compared to the control group.

Conclusion: There is moderate evidence (level 1b) from 1 high quality RCT that vision-deprived training is more effective than free vision training for improving balance post-stroke.

References

Allison, R. & Dennett, R. (2007). Pilot randomized controlled trial to assess the impact of additional supported standing practice on functional ability post stroke. Clinical Rehabilitation, 21, 614-9. DOI: 10.1177/0269215507077364

An, M. & Shaugnessy, M. (2011). The effects of exercise-based rehabilitation on balance and gait for stroke patients: a systematic review. Journal of Neuroscience Nursing, 43(6) 298-307.

Au-Yeung, S.S.Y., Hui-Chan, C.W.Y., & Tank, J.C.S. (2009). Short-form tai chi improves standing balance of people with chronic stroke. Neurorehabilitation and Neural Repair, 23(5), 515-22. DOI: 10.1177/1545968308326425

Badke M. B., Duncan P. W., & Di Fabio, R. P. (1987). Influence of prior knowledge on automatic and voluntary postural adjustments in healthy and hemiplegic subjects. Phys Ther, 67(10), 1495-1500.

Bayouk, J-F., Boucher, J.P., & Leroux, A. (2006). Balance training following stroke: effects of task-oriented exercises with and without altered sensory input. International Journal of Rehabilitation Research, 29, 51-9.

Bonan I. V., Yelnik A. P., Colle F. M., Michaud C., Normand E., Panigot B., et al. (2004). Reliance on visual information after stroke. Part II: Effectiveness of a balance rehabilitation program with visual cue deprivation after stroke: a randomized controlled trial. Arch Phys Med Rehabil, 85(2), 274-278.

Byun, S.D., Jung, T.D., Kim, C.H., & Lee, Y.S. (2011). Effects of the sliding rehabilitation machine on balance and gait in chronic stroke patients – a controlled clinical trial. Clinical Rehabilitation, 25(5), 048-15.

Chen I. C., Cheng P. T., Chen C. L., Chen S. C., Chung C. Y., & Yeh, T. H. (2002). Effects of balance training on hemiplegic stroke patients. Chang Gung Med J, 25(9), 583-590.

Chen, J-C., Lin, C-H., Wei, Y-C., Hsiao, J., & Liang, C-C. (2011). Facilitation of motor and balance recovery by thermal intervention for the paretic lower limb of acute stroke: a single-blind randomized clinical trial. Clinical Rehabilitation, 25(9), 823-32. DOI: 10.1177/0269215511399591

Cheng P. T., Wu S. H., Liaw M. Y., Wong A. M., & Tang, F. T. (2001). Symmetrical body-weight distribution training in stroke patients and its effect on fall prevention. Arch Phys Med Rehabil, 82(12), 1650-1654.

Dean, C., Richards, C., & Malouin, F. (2000). Task-related circuit training improves performance of locomotor tasks in chronic stroke: A randomized, controlled pilot trial. Archives of Physical Medicine and Rehabilitation, 81(4), 409-417.

de Seze M., Wiart L., Bon-Saint-Come A., Debelleix X., de Seze M., Joseph P. A., et al. (2001). Rehabilitation of postural disturbances of hemiplegic patients by using trunk control retraining during exploratory exercises. Arch Phys Med Rehabil, 82(6), 793-800.

Geiger R. A., Allen J. B., O’Keefe J., & Hicks, R. R. (2001). Balance and mobility following stroke: effects of physical therapy interventions with and without biofeedback/forceplate training. Phys Ther, 81(4), 995-1005.

Gok, D., Geler-Kulcu, N., Alptekin, N., & Dincer, G. (2008). Efficacy of treatment with a kinaesthetic ability training device on balance and mobility after stroke: a randomized controlled study. Clinical Rehabilitation, 22, 922-30. DOI: 10.1177/0269215508090673

Goljar, N., Burger, H., Rudolf, M., & Stanonik, I. (2010). Improving balance in subacute stroke patients: a randomized controlled study. International Journal of Rehabilitation Research, 2010, 33(3), 205-210. DOI: 10.1097/MRR.0b013e328333de61

Grant T., Brouwer B., Culham E., & Vandervoort, A. (1997). Balance retraining following acute stroke: a comparison of two methods. Canadian Journal of Rehabilitation, 11, 69-73.

Heller F., Beuret-Blanquart F., & Weber, J. (2005). [Postural biofeedback and locomotion reeducation in stroke patients]. Ann Readapt Med Phys, 48(4), 187-195.

Howe, T.E., Taylor, I., Finn, P., & Jones, H. (2005). Lateral weight transference exercises following acute stroke: a preliminary study of clinical effectiveness. Clinical Rehabilitation, 19, 45-53.

Karthikbabu, S., Nayak, A., Vijayakumar, K., Misri, Z.K., Suresh, B.V., Ganesan, S, & Joshua, A.M. (2011). Comparison of physio ball and plinth trunk exercises regimens on trunk control and functional balance in patients with acute stroke: a pilot randomized controlled trial. Clinical Rehabilitation, 25(8), 709-719. DOI: 10.1177/0269215510397393

Katz-Leurer, M., Sender, I., Keren, O., & Dvir, Z. (2006). The influence of early cycling training on balance in stroke patients at the subacute stage. Results of a preliminary trial. Clinical Rehabilitation, 20, 398-405.

Lau, K.W.K. & Mak, M.K.Y. (2011). Speed-dependent treadmill training is effective to improve gait and balance performance in patients with sub-acute stroke. Journal of Rehabilitation Medicine, 43, 709-13.

Lubetzky-Vilnai, A. & Kartin, D. (2010). The effect of balance training on balance performance in individuals poststroke: A systematic review. Journal of Neurologic Physical Therapy, 34, 127-137.

Marigold, D.S., Eng, J.J., Dawson, A.S., Inglis, J.T., Harris, J.E., & Gylfadottir, S. (2005). Exercise leads to faster postural reflexes, improved balance and mobility, and fewer falls in older persons with chronic stroke. Journal of the American Geriatrics Society, 53(3), 416-423.

McClellan, R. & Ada, L. (2004). A six-week, resource-efficient mobility program after discharge from rehabilitation improves standing in people affected by stroke: Placebo-controlled, randomised trial. Australian Journal of Physiotherapy, 50(3), 163-8.

Merkert, J., Butz, S., Nieczaj, R., Steinhagen-Thiessen, E., & Eckardt, R. (2011). Combined whole body vibration and balance training using Vibrosphere®: Improvement of trunk stability, muscle tone, and postural control in stroke patients during early geriatric rehabilitation. Zeitschrift für Gerontologie und Geriatrie, 44, 256-61. DOI: DOI 10.1007/s00391-011-0170-9

Morioka S., & Yagi, F. (2003). Effects of perceptual learning exercises on standing balance using a hardness discrimination task in hemiplegic patients following stroke: a randomized controlled pilot trial. Clin Rehabil, 17(6), 600-607.

Mudie M. H., Winzeler-Mercay U., Radwan S., & Lee, L. (2002). Training symmetry of weight distribution after stroke: a randomized controlled pilot study comparing task-related reach, Bobath and feedback training approaches. Clin Rehabil, 16(6), 582-592.

Noh, D. K., Lim, J-Y., Sin, H-I., & Paik, N-J. (2008) The effect of aquatic therapy on postural balance and muscle strength in stroke survivors – a randomized controlled pilot trial. Clinical Rehabilitation, 22, 966-76. DOI: 10.1177/0269215508091434

Onigbinde, A.T., Awotidebe, T., & Awosika, H. (2009). Effect of 6 weeks wobble board exercises on static and dynamic balance on stroke survivors. Technology and Health Care, 17, 387-92.

Outermans, J.C., van Peppen, R.P., Wittink, H., Takken, T., & Kwakkel, G. (2010). Effects of a high-intensity task-oriented training on gait performance early after stroke: a pilot study. Clinical Rehabilitation, 24(11), 979-87.

Perennou D. A., Leblond C., Amblard B., Micallef J. P., Herisson C., & Pelissier, J. Y. (2001). Transcutaneous electric nerve stimulation reduces neglect-related postural instability after stroke. Arch Phys Med Rehabil, 82(4), 440-448.

Pollock A. S., Durward B. R., Rowe P. J., & Paul, J. P. (2002). The effect of independent practice of motor tasks by stroke patients: a pilot randomized controlled trial. Clin Rehabil, 16(5), 473-480.

Richards C.L., Malouin, F., Wood-Dauphinee, S., Williams, J.I., Bouchard, J.P., & Brunet, D. (1993). Task-specific physical therapy for optimization of gait recovery in acute stroke patients. Archives of Physical Medicine and Rehabilitation, 74, 612-20.

Rose, D., Paris, T., Crews, E., Wu, S.S., Sun, A., Behrman, A.L., & Duncan P. (2011). Feasibility and effectiveness of circuit training in acute stroke rehabilitation. Neurorehabilitation and Neural Repair, 25(2), 140-148.

Sackley, C.M. & Lincoln, N.B. (1997). Single blind randomized controlled trial of visual feedback after stroke: effects on stance symmetry and function. Disability and Rehabilitation, 19(12), 536-46.

Saeys, W., Vereeck, L., Truijen, S., Lafosse, C., Wuyts, F.P., & Van de Heyning, P. (2011). Randomized controlled trial of truncal exercises early after stroke to improve balance and mobility. Neurorehabilitation and Neural Repair, 26(3) 231-8. DOI: 10.1177/1545968311416822

Salbach N. M., Mayo N. E., Robichaud-Ekstrand S., Hanley J. A., Richards C. L., & Wood-Dauphinee, S. (2005). The effect of a task-oriented walking intervention on improving balance self-efficacy poststroke: a randomized, controlled trial. J Am Geriatr Soc, 53(4), 576-582.

Salbach, N.M., Mayo, N.E., Wood-Dauphinee, S., Hanley, J.A., Richards, C.L., & Cote, R. (2004). A task-orientated intervention enhances walking distance and speed in the first year post stroke: a randomized controlled trial. Clinical Rehabilitation, 18, 509-19.

Shumway-Cook A., Anson D., & Haller, S. (1988). Postural sway biofeedback: its effect on reestablishing stance stability in hemiplegic patients. Arch Phys Med Rehabil, 69(6), 395-400.

van Nes, I.J.W., Latour, H., Schils, F., Meijer, R., van Kuijk, A., & Geurts, A.C.H. (2006). Long-term effects of 6-week whole-body vibration on balance recovery and activities of daily living in the postacute phase of stroke: a randomized, controlled trial. Stroke, 37, 2331-5.

Verheyden, G., Vereeck, L., Truijen, S., Troch, M., LaFosse, C., Saeys, W., Leenaerts, El., Palinckx, A., & DeWeerdt, W. (2009). Additional exercises improve trunk performance after stroke: a pilot randomized controlled trial. Neurorehabilitation and Neural Repair, 23(3), 281-6. DOI: 10.1177/1545968308321776

Walker C., Brouwer B. J., & Culham, E. G. (2000). Use of visual feedback in retraining balance following acute stroke. Phys Ther, 80(9), 886-895.

Wong, A.M.K., Lee, M-Y., Kuo, J-K., & Tang, F-T. (1997). The development and clinical evaluation of a standing biofeedback trainer. Journal of Rehabilitation Research and Development, 34(3) 322-7.

Yang, S., Hwang, W-H., Tsai, Y-C., Liu, F-K., Hsieh, L-F., & Chern, J-S. (2011). Improving balance skills in patients who had stroke through virtual reality treadmill training. American Journal of Physical Medicine & Rehabilitation, 90, 969-978.

Yelnik A.P., Le Breton F., Colle F.M., Bonan I.V., Hugeron C., Egal V., et al. (2008). Rehabilitation of Balance After Stroke With Multisensorial Training: A Single-Blind Randomized Controlled Trial. Neurorehabilitation and Neural Repair, 22, 468-476.

Help us to improve