Electromechanical Gait Trainers

Evidence Reviewed as of before: 14-02-2020
Author(s)*: Annabel McDermott (BOccThy); Adam Kagan, BSc; Nicol Korner-Bitensky, PhD OT
Content consistency: Gabriel Plumier
Patient/Family Information Table of contents

Introduction

Electromechanical gait trainers have been developed as an alternative to conventional gait training methods such as assisted overground walking and treadmill training. Electromechanical gait trainers differ from traditional treadmill training in that the device guides the lower limbs through the gait motion. Electromechanical devices can enable intensive repetition of movement with reduced physical assistance from clinicians. The two most common types of electromechanical gait trainers are ‘end-effector’ trainers and exoskeleton trainers.

Patient/Family Information

What is an electromechanical gait trainer and what is it for?

An electromechanical gait trainer – also called a Gait Trainer – is a piece of equipment that is used to help regain walking skills (gait) after a stroke. Gait Trainers are used in stroke rehabilitation centers instead of, or as well as other types of walking training such as walking on a treadmill or walking over obstacles. Gait Trainers help move the leg through the stepping motion. They provide a way for patients to practice walking in a natural pattern.

Two main types of Gait Trainers are:

  1. End-effector gait trainers look similar to a large treadmill. The patient stands on a treadmill surface or on footplates. The patient wears a harness around his/her upper legs and torso for support. The treadmill or footplates are motor-driven to move the patient’s legs through the correct position and movements of walking.
  2. Exoskeleton gait trainers use a treadmill or a walking frame on wheels. The exoskeleton is a robotic orthosis that attaches to the patient’s legs. The patient stands on a treadmill surface and is supported by a harness, or stands within a walking frame on wheels. The orthosis guides the legs through the correct walking patterns. These are also called robotic gait trainers.

Why do we use a Gait Trainer after a stroke?

A stroke can affect aspects of walking such as:

  • How fast you walk
  • How long you walk for before tiring
  • The length of your step
  • The number of steps you take
  • The symmetry of your leg movements
  • Your risk of falling
  • Your confidence when walking.

If you have difficulty walking after a stroke, your rehabilitation will probably include gait (walking) training. Gait Trainers may be a safe way for you to begin walking training if you are not able to walk safely by yourself, because:

  • The harness supports your body and takes some of the weight off your legs,
  • The Gait Trainer guides your legs through the walking patterns.

Gait Trainers can be used to help some patients walk sooner after a stroke, especially if they require help from two people to walk over ground. It also allows some patients to practice walking when they are not ready to walk over ground.

Do electromechanical gait trainers work for stroke?

Gait Trainers enable patients to practice repetitive walking movements. It is believed that this repetition may help rewire the area of the brain affected by the stroke. When the patient performs repetitive walking movements, the brain cells form new ‘pathways’ to repair the damage from the stroke.

Research shows that Gait Trainers are as good as usual gait rehabilitation (e.g. normal walking training, treadmill training) after stroke. Gait Trainers may be more useful than usual gait training methods for improving:

  • Leg motor function in the very early stage of recovery (up to one month post-stroke)
  • Activities of daily living, functional walking skills, leg muscle strength and walking endurance in the rehabilitation phase (1-6 months post-stroke)
  • Functional walking skills, gait patterns, general mobility and leg muscle strength more than 6 months post-stroke.

Further studies are needed to better understand who can benefit most from this type of training. Gait Trainers are best used in conjunction with standard physical therapy.

What can I expect?

There are many different types of Gait Trainers. Here is what you can expect when using most types of Gait Trainer:

  • You will wear a harness over your clothes
  • The harness is fastened to an overhead suspension system. This harness supports your upper body and reduces the amount of weight you carry through your legs while you walk. Your therapist will decide how much of your body weight is supported by the harness and how much is carried through your legs
  • Your therapist will start the Gait Trainer at a very low speed. As your walking skills improve, your therapist can (a) increase your walking speed slowly, and (b) increase the amount of weight you are taking through your legs
  • The device will guide your legs through walking movements.

Are there any side effects or risks?

There are no specific side effects from using Gait Trainers. In fact, research has shown that it is easier on your heart if you walk with your body weight supported. As such, it may be easier for some patients who have had a stroke to use the gait trainer than walking on the ground.

Who provides the treatment?

Gait Trainers are usually used by a physical therapist at a rehabilitation center. An assistant may be present to help you get ready by putting on your harness and staying with you during rest periods.

Gait Trainers are a fairly new treatment device and are labor-intensive. It requires specific equipment that is quite costly. As such, Gait Trainers may not be available in all rehabilitation centers.

How much does it cost?

Gait Trainers are very expensive pieces of equipment and are only suitable for use in rehabilitation centers under supervision of a trained rehabilitation professional. They are not suitable for private use.

Is a Gait Trainer suitable for me?

Gait Trainers have been shown to be as good as other walking rehabilitation approaches for most outcomes after stroke. Factors such as time since stroke, the severity of your stroke, and your rehabilitation center’s access to a Gait Trainer may affect your ability to use this form of rehabilitation. Ask your rehabilitation specialist if this is a suitable intervention for you.

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.

Electromechanical gait trainers have been developed as an alternative to conventional gait training methods such as assisted overground walking and treadmill training. Electromechanical gait trainers differ from traditional treadmill training in that the device guides the lower limbs through the gait motion, which enables intensive repetition of movement with reduced physical assistance from clinicians. Gait trainers use automated mechanical forces to drive movements of the lower extremity (Morone et al., 2017). This motor-driven stepping may improve the symmetry of walking patterns as it forces timing between the lower limbs, promotes hip extension and discourages compensatory movement patterns (Regneaux et al., 2008; Polese et al., 2013).

The two common types of electromechanical gait trainers are end-effector devices and exoskeleton devices. Both devices move the lower limbs through kinematically repetitive stepping patterns in the sagittal plane. Both systems adopt a ‘forced use’ motor learning approach to promote task-oriented training of walking. This enables intensive, repetitive practice of typical walking patterns that simulate stance and swing phases of gait training (Bruni et al., 2018; Kim & You, 2017; Polese et al., 2013; Regneaux et al., 2008).

End-effector devices use footplates to facilitate stance and swing phases during locomotor training (Bruni et al., 2018). The most common electromechanical gait trainers are the Reha-Stim Gait Trainers (GT1, GT2), G-EO system, Lokohelp and the Haptik walker. This StrokEngine review of end-effector devices includes ten high quality RCTs, three fair quality RCTs and two non-randomized studies; most studies were conducted with patients in the subacute or chronic phases of recovery. In this review, end-effector gait trainers are compared with interventions including conventional physiotherapy, conventional rehabilitation, conventional gait training, overground gait training and body-weight supported treadmill training. While end-effector gait trainers were no more effective than comparison interventions for most outcomes, findings from this review provide at least moderate evidence (Level 1b) end-effector gait trainers were more effective than comparison interventions for improving:

  • Activities of daily living, functional ambulation, lower extremity muscle strength and walking endurance in the subacute phase of stroke recovery;
  • Functional ambulation, gait parameters, mobility and lower extremity muscle strength in the chronic phase of stroke recovery; and
  • Disability in a sample of patients across the stroke recovery continuum.

Exoskeleton devices facilitate movement of the hips and knees during the phases of gait (Bruni et al., 2018). Examples of exoskeleton devices include the Lokomat and the LOPES. The Lokomat uses a harness-supported body weight system in conjunction with a treadmill. This StrokEngine review of exoskeleton devices includes eight high quality RCTs, nine fair quality RCTs and two non-randomized studies; most studies were conducted with patients in the subacute or chronic phases of stroke recovery. In this review, exoskeleton gait trainers are compared with interventions including conventional physical therapy, treadmill training, therapist-assisted gait training, or no additional gait training. While exoskeleton devices were no more effective than comparison interventions for most functional outcomes, findings from this review provide at least moderate evidence (Level 1b) that exoskeleton gait trainers were more effective than comparison interventions for improving:

  • Lower extremity motor function in the acute phase of stroke recovery; and
  • Functional independence, functional ambulation, neurological recovery, and stair climbing in a sample of patients across the stroke recovery continuum.

Results Table

View results table

Outcomes

Acute Phase - End-effector gait trainers

Functional ambulation
Not Effective
2A

One fair quality RCT (Peurala et al., 2009) examined the effect of an end-effector gait trainer on functional ambulation in the acute phase of stroke recovery. This fair quality RCT randomized patients to receive (1) gait training using the GT1 device, (2) time-matched overground walking training, or (3) conventional rehabilitation alone. Functional ambulation was measured using the Functional Ambulation Category at post-treatment (3 weeks) and follow-up (6 months). No significant between-group differences were found at either time point.
Note: The end-effector gait training group showed significantly lower perceived exertion during training (measured by the Borg scale) compared to the other groups, and achieved the required 20 minutes of walking per 1-hour session in less time than the overground walking training group, which often required the full 1 hour before achieving 20 minutes of walking.

Conclusion: There is limited evidence (Level 2a) from 1 fair quality RCT that end-effector gait trainers are not more effective than comparison interventions (overground walking training, conventional rehabilitation alone) in improving functional ambulation in the acute phase of stroke recovery.

Mobility
Not Effective
2A

One fair quality RCT (Peurala et al., 2009) examined the effect of an end-effector gait trainer on mobility in the acute phase of stroke recovery. This fair quality RCT randomized patients to receive (1) gait training using the GT1 device, (2) time-matched overground walking training, or (3) conventional rehabilitation alone. Mobility was measured using the Rivermead Mobility Index and the Rivermead Motor Assessment Scale (RMA – Gross motor function, Lower extremity function and trunk control) at post-treatment (3 weeks) and follow-up (6 months). No significant between-group differences were found on either measure at either time point.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that end-effector gait trainers are not more effective than comparison interventions (overground walking training, conventional rehabilitation alone) in improving mobility in the acute phase of stroke recovery.

Motor function
Effective
2A

One fair quality RCT (Peurala et al., 2009) examined the effect of an end-effector gait trainer on motor function in the acute phase of stroke recovery. This fair quality RCT randomized patients to receive (1) gait training using the GT1 device, (2) time-matched overground walking training, or (3) conventional rehabilitation alone. Motor function was measured using the Modified Motor Assessment Scale (MMAS) at post-treatment (3 weeks) and follow-up (6 months). A significant between-group difference was seen at post-treatment, in favour of end-effector gait training vs. conventional rehabilitation alone. Differences did not remain significant at follow-up. There were no significant differences between end-effector gait training vs. overground walking training.
Note: A significant between-group difference was found at post-treatment in favour of overground walking vs. conventional rehabilitation; differences did not remain significant at follow-up.

Conclusion: There is limited evidence (Level 2a) from 1 fair quality RCT that end-effector gait trainers are more effective, in short-term, than a comparison intervention (conventional rehabilitation) in improving motor function in the acute phase of stroke recovery.
Note: End-effector gait trainers were not more effective than overground walking training.

Walking endurance
Not Effective
2A

One fair quality RCT (Peurala et al., 2009) examined the effect of an end-effector gait trainer on walking endurance in the acute phase of stroke recovery. This fair quality RCT randomized patients to receive (1) gait training using the GT1 device, (2) time-matched overground walking training, or (3) conventional rehabilitation alone. Walking endurance was measured using the 6 Minute Walk Test at post-treatment (3 weeks) and follow-up (6 months). A significant difference was found at follow-up, in favour of end-effector gait training vs. overground walking.
Note: Insufficient data was gathered from the conventional rehabilitation group at either time point and no between-group comparisons were made.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that end-effector gait trainers are not more effective than a comparison intervention (overground walking training) in improving walking endurance in the acute phase of stroke recovery.
Note: End-effector gait trainers were more effective than overground walking in the long term.

Walking speed
Not Effective
2A

One fair quality RCT (Peurala et al., 2009) examined the effect of an end-effector gait trainer on walking speed in the acute phase of stroke recovery. This fair quality RCT randomized patients to receive (1) gait training using the GT1 device, (2) time-matched overground walking training, or (3) conventional rehabilitation alone. Walking speed was measured using the 10-meter walking test at post-treatment (3 weeks) and follow-up (6 months). No significant between-group differences were found at either time point.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that end-effector gait trainers are not more effective than comparison interventions (overground walking training, conventional rehabilitation alone) in improving walking speed in the acute phase of stroke recovery.

Acute Phase - Exoskeleton gait trainers

Cardiopulmonary fitness
Not Effective
1B

One high quality RCT (Chang et al., 2012) investigated the effect of exoskeleton gait trainers on cardiopulmonary fitness in the acute phase of stroke recovery. The high quality RCT randomized patients to receive gait training using the Lokomat device or time-matched conventional physical therapy; both groups received additional physical therapy. Cardiopulmonary fitness was measured according to aerobic capacity (peak VO2, respiratory exchange ratio), cardiovascular response (oxygen pulse, peak heart rate, systolic/diastolic blood pressure, rate of perceived exertion) and ventilatory response (minute ventilation, ventilatory efficiency) at post-treatment (2 weeks). There was a significant between-group difference in aerobic capacity only (peak VO2), in favour of exoskeleton gait training vs. conventional physical therapy.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that an exoskeleton gait trainer is not more effective than a comparison intervention (conventional physical therapy) for improving cardiopulmonary fitness in the acute phase of stroke recovery.

Functional ambulation
Not Effective
1B

One high quality RCT (Chang et al., 2012) investigated the effect of exoskeleton gait trainers on functional ambulation in the acute phase of stroke recovery. The high quality RCT randomized patients to receive gait training using the Lokomat device or time-matched conventional physical therapy; both groups received additional physical therapy. Functional ambulation was measured using the Functional Ambulation Categories at post-treatment (2 weeks). No significant between-group difference was found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that an exoskeleton gait trainer is not more effective than a comparison intervention (conventional physical therapy) for improving functional ambulation in the acute phase of stroke recovery.

Motor function – lower extremity
Effective
1B

One high quality RCT (Chang et al., 2012) investigated the effect of exoskeleton gait trainers on lower extremity motor function in the acute phase of stroke recovery. The high quality RCT randomized patients to receive gait training using the Lokomat device or time-matched conventional physical therapy; both groups received additional physical therapy. Lower extremity motor function was measured using the Fugl-Meyer Assessment – Lower Extremity score (FMA-LE) at post-treatment (2 weeks). There was a significant between-group difference, in favour of exoskeleton gait training vs. conventional physical therapy.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that an exoskeleton gait trainer is more effective than a comparison intervention (conventional physical therapy) for improving lower extremity motor function in the acute phase of stroke recovery.

Motor power – lower extremity
Not Effective
1B

One high quality RCT (Chang et al., 2012) investigated the effect of exoskeleton gait trainers on lower extremity motor power in the acute phase of stroke recovery. The high quality RCT randomized patients to receive gait training using the Lokomat device or time-matched conventional physical therapy; both groups received additional physical therapy. Lower extremity motor power was measured using the Motricity Index – Leg score at post-treatment (2 weeks). No significant between-group difference was found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that an exoskeleton gait trainer is not more effective than a comparison intervention (conventional physical therapy) for improving lower extremity motor power in the acute phase of stroke recovery.

Subacute Phase - End-effector gait trainers

Activities of daily living
Effective
1B

One high quality RCT (Pohl et al., 2007) investigated the effect of end-effector gait trainers on activities of daily living (ADLs) in the subacute phase of stroke recovery. This high quality RCT randomized patients to receive gait training using the GT1 device or time-matched physical therapy. ADLs were measured using the Barthel Index (BI) at post-treatment (4 weeks) and follow-up (6 months). A significant between-group difference in the number of patients who achieved BI scores > 75 was seen at post-treatment, in favour of end-effector gait training vs. physical therapy. This difference did not remain significant at follow-up.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that end-effector gait trainers are more effective, in short-term, than a comparison intervention (time-matched physical therapy) in improving Activities of daily living in the subacute phase of stroke recovery.

Balance
Not Effective
2B

One non-randomized study (Iacovelli et al., 2018) examined the effect of end-effector gait trainers on balance in the subacute phase of stroke recovery. This non-randomized controlled trial assigned patients to receive gait training using the G-EO device or conventional gait training. Balance was measured using the Tinetti Scale at post-treatment (20 sessions). No significant between-group difference was found.

Conclusion: There is limited evidence (Level 2b) from one non-randomized study that end-effector gait trainers are not more effective than a comparison intervention (conventional gait training) in improving balance in the subacute phase of stroke recovery.

Functional ambulation
Effective
1A

Two high quality RCTs (Werner et al., 2002; Pohl et al., 2007) and two non-randomized studies (Hesse et al., 2012; Iacovelli et al., 2018) investigated the effect of end-effector gait trainers on functional ambulation in the subacute phase of stroke recovery.

The first high quality cross-over design RCT (Werner et al., 2002) randomized patients to receive gait training using the GT1 device or body-weight supported treadmill training (BWS-TT). Interventions were provided using an A-B-A or B-A-B design, with each block lasting 2 weeks (6 weeks total). Functional ambulation was measured by the Functional Ambulation Category (FAC) at post-treatment (6 weeks) and follow-up (6 months). A significant between-group difference was found at post-treatment, in favour of end-effector gait training vs. BWS-TT; the difference did not remain significant at follow-up.

The second high quality RCT (Pohl et al., 2007) randomized patients to receive gait training using the GT1 device or time-matched physical therapy. Functional ambulation was measured by the FAC at post-treatment (4 weeks) and follow-up (6 months). A significant between-group difference was found at both time points in the number of patients to achieve independent walking (FAC > 4) at post-treatment, in favour of end-effector gait training vs. physical therapy.

The first non-randomized controlled trial (Hesse et al., 2012) assigned non-ambulatory patients to receive gait training using the G-EO device or time-matched physical therapy. Functional ambulation was measured using the FAC at post-treatment (4 weeks) and follow-up (3 months). A significant between-group difference was found at both time points, in favour of end-effector gait training vs. physical therapy.

The second non-randomized controlled trial (Iacovelli et al., 2018) assigned patients to receive gait training using the G-EO device or conventional gait training. Functional ambulation was measured using the FAC and the Walking Handicap Scale at post-treatment (20 sessions). No significant between-group differences were found.

Conclusion: There is strong evidence (Level 1a) from 2 high quality RCTs and one non-randomized study that end-effector gait trainers are more effective than comparison interventions (body-weight supported treadmill training or physical therapy) for improving functional ambulation in the subacute phase of stroke recovery.

Gait parameters
Not Effective
2B

One non-randomized study (Iacovelli et al., 2018) examined the effect of end-effector gait trainers on gait parameters in the subacute phase of stroke recovery. This non-randomized controlled trial assigned patients to receive gait training using the G-EO device or conventional gait training. Gait parameters were measured using the SMART0D500 optoelectronic system (stride time, stride length, step length, cadence, velocity, swing velocity, mean velocity) and five measures (step length, swing time, stance time, double support time, intra-limb ratio of swing: stance time) were used to calculate the symmetry index (symmetry ratio, symmetry index, gait asymmetry, symmetry angle) at post-treatment (20 sessions). Significant between-group differences were found in only two gait parameters (symmetry ratio – step length, symmetry angle – step length), in favour of end-effector gait training vs. conventional gait training.

Conclusion: There is limited evidence (Level 2b) from one non-randomized study that end-effector gait trainers are not more effective than a comparison intervention (conventional gait training) for improving gait parameters in the subacute phase of stroke recovery.

Mobility
Conflicting
4

Two high quality RCTs (Werner et al., 2002; Pohl et al., 2007) and two non-randomized studies (Hesse et al., 2012; Iacovelli et al., 2018) investigated the effect of end-effector gait trainers on mobility in patients with subacute stroke.

The first high quality cross-over design RCT (Werner et al., 2002) randomized patients to receive gait training using the GT1 device or body-weight supported treadmill training (BWS-TT). Interventions were provided using an A-B-A or B-A-B design, with each block lasting 2 weeks (6 weeks total). Mobility was measured using the Rivermead Motor Assessment (RMA – Gross function, Trunk and leg subscales) at post-treatment (6 weeks). No significant between-group differences were found.

The second high quality RCT (Pohl et al., 2007) randomized patients to receive gait training using the GT1 device or time-matched physical therapy. Mobility was measured using the Rivermead Mobility Index (RMI) at post-treatment (4 weeks) and follow-up (6 months). A significant between-group difference was found at post-treatment in favour of end-effector gait training vs. physical therapy. Differences did not remain significant at follow-up.

The first non-randomized controlled trial (Hesse et al., 2012) assigned non-ambulatory patients to receive gait training using the G-EO device or time-matched physical therapy. Mobility was measured using the RMI at post-treatment (4 weeks) and follow-up (3 months). A significant between-group difference was found at post-treatment in favour of end-effector gait training vs. physical therapy. Differences did not remain significant at follow-up.

The second non-randomized controlled trial (Iacovelli et al., 2018) assigned patients to receive gait training using the G-EO device or conventional gait training. Mobility was measured using the Timed Up and Go (TUG) test and the Trunk Control Test at post-treatment (20 sessions). A significant between-group difference was found in one measure of mobility (TUG) in favour of end-effector gait training vs. conventional gait training.

Conclusion: There is conflicting evidence (Level 4) regarding the effect of end-effector gait trainers on mobility in the subacute phase of stroke recovery. One high quality RCT and two non-randomized studies found that end-effector gait trainers are more effective than comparison interventions (physical therapy, conventional gait training); a second high quality RCT found no significant difference compared with body-weight supported treadmill training.
Note: Differences in results may relate to the variation in outcome measures used.

Motor function - lower extremity
Not Effective
2B

One non-randomized study (Iacovelli et al., 2018) examined the effect of end-effector gait trainers on lower extremity motor function in the subacute phase of stroke recovery. This non-randomized controlled trial assigned patients to receive gait training using the G-EO device or conventional gait training. Lower extremity motor function was measured using the Fugl-Meyer Assessment – Lower Extremity at post-treatment (20 sessions). No significant between-group difference was found.

Conclusion: There is limited evidence (Level 2b) from one non-randomized study that end-effector gait trainers are not more effective than a comparison intervention (conventional gait training) for improving lower extremity motor function in the subacute phase of stroke recovery.

Muscle strength - lower extremity
Effective
1B

One high quality RCT (Pohl et al., 2007) and two non-randomized studies (Hesse et al., 2012; Iacovelli et al., 2018) examined the effect of end-effector gait trainers on lower extremity muscle strength in the subacute phase of stroke recovery.

The high quality RCT (Pohl et al., 2007) randomized patients to receive gait training using the GT1 device or time-matched physical therapy. Lower extremity muscle strength was measured using the Motricity Index (MI) and post-treatment (4 weeks) and follow-up (6 months). A significant between-group difference was seen at post-treatment in favour of end-effector gait training vs. physical therapy. Differences did not remain significant at follow-up.

The first non-randomized controlled trial (Hesse et al., 2012) assigned non-ambulatory patients to receive gait training using the G-EO device or time-matched physical therapy. Lower extremity muscle strength was measured using the MI at post-treatment (4 weeks) and follow-up (3 months). A significant between-group difference was found at both time points in favour of end-effector gait training vs. physical therapy.

The second non-randomized controlled trial (Iacovelli et al., 2018) assigned patients to receive gait training using the G-EO device or conventional gait training. Lower extremity muscle strength was measured using the MI and the Medical Research Council Scale (MRC Scale – Hip extension, Knee flexion, Ankle flexion) at post-treatment (20 sessions). Significant between-group differences were found on two measures of muscle strength (MRC scale – Hip extension, Ankle flexion), in favour of end-effector gait training vs. conventional gait training.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and two non-randomized studies that end-effector gait trainers are more effective than comparison interventions (time-matched physical therapy, conventional gait training) for improving lower extremity muscle strength in the subacute phase of stroke recovery.

Muscle tone/Spasticity - lower extremity
Not Effective
1b

One high quality RCT (Werner et al., 2002) and two non-randomized studies (Hesse et al., 2012; Iacovelli et al., 2018) examined the effect of end-effector gait trainers on lower extremity muscle tone/spasticity in the subacute phase of stroke recovery.

The high quality cross-over design RCT (Werner et al., 2002) randomized patients to receive gait training using the GT1 device or body-weight supported treadmill training (BWS-TT). Interventions were provided using an A-B-A or B-A-B design, with each block lasting 2 weeks (6 weeks total). Spasticity was measured by the Modified Ashworth Scale at post-treatment (6 weeks). No significant between-group difference was found.

The first non-randomized controlled trial (Hesse et al., 2012) assigned non-ambulatory patients to receive gait training using the G-EO device or time-matched physical therapy. Lower extremity muscle tone was measured using the Resistance to Passive Movement Scale at post-treatment (4 weeks) and follow-up (3 months). No significant between-group difference was found at either time point.

The second non-randomized controlled trial (Iacovelli et al., 2018) assigned patients to receive gait training using the G-EO device or conventional gait training. Lower extremity spasticity was measured using the Ashworth Scale (Total score, Hip, Knee) at post-treatment (20 sessions). Significant between-group differences were found on all measures of Ahsworth Scale, in favour of end-effector gait training vs. conventional gait training.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one non-randomized study that end-effector gait trainers are not more effective than comparison interventions (body-weight supported treadmill training, time-matched physical therapy) in improving lower extremity muscle tone/spasticity in the subacute phase of stroke recovery.
Note: However, a second non-randomized study found that end-effector gait trainers were more effective than conventional gait training.

Range of motion - lower extremity
Not Effective
2B

One non-randomized study (Iacovelli et al., 2018) examined the effect of end-effector gait trainers on lower extremity range of motion in the subacute phase of stroke recovery. This non-randomized controlled trial assigned patients to receive gait training using the G-EO device or conventional gait training. Lower extremity range of motion (Hip, Knee, Ankle) was measured in the sagittal plane at post-treatment (20 sessions). No significant between-group differences were found.

Conclusion: There is limited evidence (Level 2b) from one non-randomized study that end-effector gait trainers are not more effective than a comparison intervention (conventional gait training) for improving lower extremity range of motion in the subacute phase of stroke recovery.

Walking endurance
Effective
1B

One high quality RCT (Pohl et al., 2007) and one non-randomized study (Iacovelli et al., 2018) examined the effect of end-effector gait trainers on walking endurance in the subacute phase of stroke recovery.

The high quality RCT (Pohl et al., 2007) randomized patients to receive gait training using the GT1 device or time-matched physical therapy. Walking endurance was measured using the 6 Minute Walk Test (6MWT) at post-treatment (4 weeks) and follow-up (6 months). A significant between-group difference was seen at post-treatment in favour of end-effector gait training vs. physical therapy. Results did not remain significant at follow-up.

The non-randomized controlled trial (Iacovelli et al., 2018) assigned patients to receive gait training using the G-EO device or conventional gait training. Walking endurance was measured using the 6MWT at post-treatment (20 sessions). A significant between-group difference was found in favour of end-effector gait training vs. conventional gait training.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one non-randomized study that end-effector gait trainersare more effective than comparison interventions (physical therapy, conventional gait training) for improving walking endurance in the subacute phase of stroke recovery.

Walking speed
Conflicting
4

Two high quality RCTs (Werner et al., 2002; Pohl et al., 2007) and two non-randomized studies (Hesse et al., 2012; Iacovelli et al., 2018) investigated the effect of end-effector gait trainers on walking speed in the subacute phase of stroke recovery.

The first high quality cross-over design RCT (Werner et al., 2002) randomized patients to receive gait training using the GT1 device or Body-Weight Supported Treadmill Training (BWS-TT) using an A-B-A or B-A-B format, with each block lasting 2 weeks (6 weeks total). Walking speed was measured using the 10-meter walking test each week for 6 weeks (post-treatment). No significant between-group difference was found.

The second high quality RCT (Pohl et al., 2007) randomized patients to receive gait training using the GT1 device or time-matched physical therapy. Walking speed was measured using the 10-meter walking test at post-treatment (4 weeks) and follow-up (6 months). There was a significant between-group difference at post-treatment in favour of end-effector gait training vs. physical therapy. Between-group differences did not remain significant at follow-up.

The first non-randomized controlled trial (Hesse et al., 2012) assigned non-ambulatory patients to receive gait training using the G-EO device or time-matched physical therapy. Walking speed was measured using the 10-meter walking test at post-treatment (4 weeks) and follow-up (3 months). A significant between-group difference was found at post-treatment in favour of end-effector gait training vs. physical therapy. Differences did not remain significant at follow-up.

The second non-randomized controlled trial (Iacovelli et al., 2018) assigned patients to receive gait training using the G-EO device or conventional gait training. Walking speed was measured using the 10-meter walking test at post-treatment (20 sessions). A significant between-group difference was found in favour of end-effector gait training vs. conventional gait training.

Conclusion: There is conflicting evidence (Level 4) regarding the effect of end-effector gait trainers on walking speed in the subacute phase of stroke recovery. One high quality RCT and two non-randomized studies found that end-effector gait trainers are more effective than comparison interventions (physical therapy, conventional gait training), whereas a second high quality RCT found no significant difference between end-effector gait trainers and Body-Weight Supported Treadmill Training.

Subacute Phase - Exoskeleton gait trainers

Activities of daily living (ADLs)/Instrumental ADLs
Not Effective
1B

One high quality RCT (Husemann et al., 2007), two fair quality RCT (Hidler et al., 2009; Han et al., 2016) and one non-randomized study (Chung, 2017) investigated the effect of exoskeleton gait trainers on ADLs or IADLs in the subacute phase of stroke recovery.

The high quality RCT (Husemann et al., 2007) randomized patients to receive gait training using the Lokomat device or conventional physical therapy. ADLs were measured by the Barthel Index (BI) at post-treatment (4 weeks). No significant between-group difference was found.

The first fair quality RCT (Hidler et al., 2009) randomized patients to receive gait training using the Lokomat device or time-matched conventional gait training. IADLs were measured using the Frenchay Activities Index at mid-treatment (12 sessions), post-treatment (24 sessions) and follow-up (3 months). No significant between-group difference was found at any timepoint.

The second fair quality RCT (Han et al., 2016) randomized patients to receive gait training using the Lokomat device or conventional physical therapy; both groups received additional physical therapy and occupational therapy. ADLs were measured by the Korean modified Barthel Index at post-treatment (4 weeks). No significant between-group difference was found.

The non-randomized case-controlled study (Chung, 2017) provided patients with gait training using the Lokomat device and physical therapy or time-matched physical therapy. ADLs were measured by the modified BI at discharge (approximately 5 weeks). No significant between-group difference in change scores was found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT, two fair quality RCTs and one non-randomized study that exoskeleton gait trainers are not more effective than comparison interventions (conventional physical therapy, conventional gait training) for improving ADLs or IADLs in the subacute phase of stroke recovery.

Balance
Not Effective
1B

One high quality RCT (Taveggia et al., 2016), three fair quality RCTs (Hidler et al., 2009; van Nunen et al., 2015; Han et al., 2016) and one non-randomized study (Chung, 2017) investigated the effect of exoskeleton gait trainers on balance in the subacute phase of stroke recovery.

The high quality RCT (Taveggia et al., 2016) randomized patients to receive gait training using the Lokomat device or time-matched conventional physical therapy for gait retraining; both groups received additional physical therapy. Balance was measured by the Tinetti Balance Scale at post-treatment (5 weeks) and follow-up (3 months). No significant between-group difference was found at either timepoint.

The first fair quality RCT (Hidler et al., 2009) randomized patients to receive gait training using the Lokomat device or time-matched conventional gait training. Balance was measured using the Berg Balance Scale (BBS) at mid-treatment (12 sessions), post-treatment (24 sessions) and follow-up (3 months). No significant between-group difference was found at any timepoint.

The second fair quality RCT (van Nunen et al., 2015) randomized patients to receive gait training using the Lokomat device or dose-matched overground gait training; both groups received additional physical therapy. Balance was measured by the BBS at post-treatment (10 weeks) and follow-up (week 24, week 36). No significant between-group difference was found at any timepoint.

The third fair quality RCT (Han et al., 2016) randomized patients to receive gait training using the Lokomat device or conventional physical therapy; both groups received additional physical therapy and occupational therapy. Balance was measured by the BBS at post-treatment (4 weeks). No significant between-group difference was found.

The non-randomized case-controlled study (Chung, 2017) provided patients with gait training using the Lokomat device and physical therapy or time-matched physical therapy. Balance was measured by the BBS at baseline and at discharge (approximately 5 weeks). A significant between-group difference in change scores from baseline to discharge was found, in favour of exoskeleton gait training + physical therapy vs. physical therapy alone.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and three fair quality RCTs that exoskeleton gait trainers are not more effective than comparison interventions (conventional physical therapy, conventional gait training, overground gait training) for improving balance in the subacute phase of stroke recovery.
Note: The non-randomized study found that the exoskeleton gait trainer was more effective than time-matched physical therapy.

Functional ambulation
Not Effective
1B

One high quality RCT (Husemann et al., 2007), three fair quality RCTs (Hidler et al., 2009; van Nunen et al., 2015; Han et al., 2016) and one non-randomized study (Chung, 2017) investigated the effect of exoskeleton gait trainers on functional ambulation in the subacute phase of stroke recovery.

The high quality RCT (Husemann et al., 2007) randomized patients to receive gait training using the Lokomat device or conventional physical therapy. Functional ambulation was measured by the Functional Ambulation Category (FAC) at post-treatment (4 weeks). No significant between-group difference was found.

The first fair quality RCT (Hidler et al., 2009) randomized patients to receive gait training using the Lokomat device or time-matched conventional gait training. Functional ambulation was measured using the FAC at mid-treatment (12 sessions), post-treatment (24 sessions) and follow-up (3 months). No significant between-group difference was found at any timepoint.

The second fair quality RCT (van Nunen et al., 2015) randomized patients to receive gait training using the Lokomat device or dose-matched overground gait training; both groups received additional physical therapy. Functional ambulation was measured by the FAC at post-treatment (10 weeks) and follow-up (week 24, week 36). No significant between-group difference was found at any timepoint.

The third fair quality RCT (Han et al., 2016) randomized patients to receive gait training using the Lokomat device or conventional physical therapy; both groups received additional physical therapy and occupational therapy. Functional ambulation was measured by the FAC at post-treatment (4 weeks). No significant between-group difference was found.

The non-randomized case-controlled study (Chung, 2017) provided patients with gait training using the Lokomat device and physical therapy or time-matched physical therapy. Functional ambulation was measured by the modified FAC at baseline and at discharge (approximately 5 weeks). A significant between-group difference in change scores from baseline to discharge was found, in favour of exoskeleton gait training + physical therapy vs. physical therapy alone.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and three fair quality RCTs that exoskeleton gait trainers are not more effective than comparison interventions (conventional physical therapy, conventional gait training, overground gait training) for improving functional ambulation in the subacute phase of stroke recovery.
Note: The non-randomized study found that the exoskeleton gait trainer was more effective than time-matched physical therapy.

Functional independance
Not Effective
1B

One high quality RCT (Taveggia et al., 2016) investigated the effect of an exoskeleton gait trainer on functional independence in the subacute phase of stroke recovery. This high quality RCT randomized patients to receive gait training using the Lokomat device or time-matched conventional physical therapy for gait retraining; both groups received additional physical therapy. Functional independence was measured by the Functional Independence Measure at post-treatment (5 weeks) and follow-up (3 months). No significant between-group difference was found at either timepoint.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that an exoskeleton gait trainer is not more effective than a comparison intervention (conventional physical therapy) for improving functional independence in the subacute phase of stroke recovery.

Gait parameters
Not Effective
1B

One high quality RCT (Husemann et al., 2007) and one fair quality RCT (Hidler et al., 2009) investigated the effect of exoskeleton gait trainers on gait parameters in the subacute phase of stroke recovery.

The high quality RCT (Husemann et al., 2007) randomized patients to receive gait training using the Lokomat device or conventional physical therapy. Gait parameters were measured by the Parotec system (Cadence, Stride duration, Stance duration – affected/unaffected leg, Single support time – affected/unaffected leg) at post-treatment (4 weeks). A significant between-group difference in one measure (Single support time – affected leg) was found, in favour of exoskeleton gait training vs. conventional physical therapy.

The fair quality RCT (Hidler et al., 2009) randomized patients to receive gait training using the Lokomat device or time-matched conventional gait training. One gait parameter was measured using the GAITRite system (Cadence) at mid-treatment (12 sessions), post-treatment (24 sessions) and follow-up (3 months). No significant between-group difference was found at any timepoint.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one fair quality RCT that exoskeleton gait trainers are not more effective than comparison interventions (conventional physical therapy, conventional gait training) for improving gait parameters in the subacute phase of stroke recovery.

Health related quality of life
Not Effective
1B

One high quality RCT (Taveggia et al., 2016) and two fair quality RCTs (Hidler et al., 2009; van Nunen et al., 2015) investigated the effect of exoskeleton gait trainers on health related quality of life (HRQoL) in the subacute phase of stroke recovery.

The high quality RCT (Taveggia et al., 2016) randomized patients to receive gait training using the Lokomat device or time-matched conventional physical therapy for gait retraining; both groups received additional physical therapy. HRQoL was measured by the Medical Outcomes Study Short Form 36 (SF-36) at post-treatment (5 weeks) and follow-up (3 months). No significant between-group difference was found at either timepoint.

The first fair quality RCT (Hidler et al., 2009) randomized patients to receive gait training using the Lokomat device or time-matched conventional gait training. HRQoL was measured using the SF-36 at mid-treatment (12 sessions), post-treatment (24 sessions) and follow-up (3 months). No significant between-group difference was found at any timepoint.

The second fair quality RCT (van Nunen et al., 2015) randomized patients to receive gait training using the Lokomat device or dose-matched overground gait training; both groups received additional physical therapy. HRQoL was measured by the SF-36 (General health, Social functioning scores) and the Stroke Impact Scale (SIS 3.0 – Activities of Daily Living, Mobility scores) at post-treatment (10 weeks) and follow-up (week 24, week 36). No significant between-group differences were found on either measure at any timepoint.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and two fair quality RCTs that exoskeleton gait trainers are not more effective than comparison interventions (conventional physical therapy, conventional gait training, overground gait training) for improving quality of life in the subacute phase of stroke recovery.

Mobility
Not Effective
2A

Two fair quality RCTs (Hidler et al., 2009; van Nunen et al., 2015) and one non-randomized study (Chung, 2017) investigated the effect of exoskeleton gait trainers on mobility in the subacute phase of stroke recovery.

The first fair quality RCT (Hidler et al., 2009) randomized patients to receive gait training using the Lokomat device or time-matched conventional gait training. Mobility was measured using the Rivermead Mobility Index (RMI) at mid-treatment (12 sessions), post-treatment (24 sessions) and follow-up (3 months). No significant between-group difference was found at any timepoint.

The second fair quality RCT (van Nunen et al., 2015) randomized patients to receive gait training using the Lokomat device or dose-matched overground gait training; both groups received additional physical therapy. Mobility was measured by the RMI at post-treatment (10 weeks) and follow-up (week 24, week 36); the Timed Up and Go test was also used with patients with a Functional Ambulation Category of or greater than 3. No significant between-group differences were found on either measure at any timepoint.

The non-randomized case-controlled study (Chung, 2017) provided patients with gait training using the Lokomat device and physical therapy or time-matched physical therapy. Mobility was measured by the modified RMI at baseline and at discharge (approximately 5 weeks). A significant between-group difference in change scores from baseline to discharge was found, in favour of exoskeleton gait training + physical therapy vs. physical therapy alone.

Conclusion: There is limited evidence (Level 2a) from two fair quality RCTs that exoskeleton gait trainers are not more effective than comparison interventions (conventional gait training, overground gait training) for improving mobility in the subacute phase of stroke recovery.
 Note: A non-randomized study found that the exoskeleton gait trainer was more effective than physical therapy alone.

Motor function - lower extremity
Not Effective
2A

Three fair quality RCTs (Hidler et al., 2009; van Nunen et al., 2015; Han et al., 2016) investigated the effect of exoskeleton gait trainers on lower extremity motor function in the subacute phase of stroke recovery.

The first fair quality RCT (Hidler et al., 2009) randomized patients to receive gait training using the Lokomat device or time-matched conventional gait training. Lower extremity motor function was measured using the Motor Assessment Scale at mid-treatment (12 sessions), post-treatment (24 sessions) and follow-up (3 months). No significant between-group difference was found at any timepoint.

The second fair quality RCT (van Nunen et al., 2015) randomized patients to receive gait training using the Lokomat device or dose-matched overground gait training; both groups received additional physical therapy. Lower extremity motor function was measured by the Fugl-Meyer Assessment – Lower Extremity (FMA-LE) at post-treatment (10 weeks) and follow-up (week 24, week 36). No significant between-group difference was found at any timepoint.

The third fair quality RCT (Han et al., 2016) randomized patients to receive gait training using the Lokomat device or conventional physical therapy; both groups received additional physical therapy and occupational therapy. Lower extremity motor function was measured by the FMA-LE at post-treatment (4 weeks). No significant between-group difference was found.

Conclusion: There is limited evidence (Level 2a) from three fair quality RCTs that exoskeleton gait trainers are not more effective than comparison interventions (conventional gait training, overground gait training, conventional physical therapy) for improving lower extremity motor function in the subacute phase of stroke recovery.

Muscle strength - lower extremity
Not Effective
1B

One high quality RCT (Husemann et al., 2007) and one fair quality RCT (van Nunen et al., 2015) investigated the effect of exoskeleton gait trainers on lower extremity muscle strength in the subacute phase of stroke recovery.

The high quality RCT (Husemann et al., 2007) randomized patients to receive gait training using the Lokomat device or conventional physical therapy. Muscle power was measured by the Motricity Index at post-treatment (4 weeks). No significant between-group difference was found.

The fair quality RCT (van Nunen et al., 2015) randomized patients to receive gait training using the Lokomat device or dose-matched overground gait training; both groups received additional physical therapy. Muscle strength was measured according to maximal voluntary isometric torque of bilateral knee extensors and flexors at post-treatment (10 weeks) and follow-up (week 24, week 36). No significant between-group difference was found at any timepoint.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one fair quality RCT that exoskeleton gait trainers are not more effective than comparison interventions (conventional physical therapy, overground gait training) for improving lower extremity muscle strength in the subacute phase of stroke recovery.

Spasticity
Not Effective
1B

One high quality RCT (Husemann et al., 2007) investigated the effect of an exoskeleton gait trainer on spasticity in the subacute phase of stroke recovery. The high quality RCT randomized patients to receive gait training using the Lokomat device or conventional physical therapy. Spasticity was measured by the Modified Ashworth Scale at post-treatment (4 weeks). No significant between-group difference was found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that an exoskeleton gait trainer is not more effective than a comparison intervention (conventional physical therapy) for reducing spasticity in the subacute phase of stroke recovery.

Stroke impact
Not Effective
2A

One fair quality RCT (van Nunen et al., 2015) investigated the effect of exoskeleton gait trainers on stroke impact the subacute phase of stroke recovery. The fair quality RCT randomized patients to receive gait training using the Lokomat device or dose-matched overground gait training; both groups received additional physical therapy. Stroke impact was measured by the Stroke Impact Scale (SIS 3.0 – Activities of Daily Living, Mobility scores) at post-treatment (10 weeks) and follow-up (week 24, week 36). No significant between-group difference was found at any time point.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that exoskeleton gait trainers are not more effective than a comparison intervention (overground gait training) for reducing the impact of stroke in the subacute phase of stroke recovery.

Stroke severity
Not Effective
2A

One fair quality RCT (Hidler et al., 2009) investigated the effect of an exoskeleton gait trainer on stroke severity in the subacute phase of stroke recovery. The fair quality RCT randomized patients to receive gait training using the Lokomat device or time-matched conventional gait training. Stroke severity was measured using the National Institute of Health Stroke Scale (NIHSS) at mid-treatment (12 sessions), post-treatment (24 sessions) and follow-up (3 months). No significant between-group difference was found at any timepoint.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that an exoskeleton gait trainer is not more effective than a comparison intervention (conventional gait training) for reducing stroke severity in the subacute phase of stroke recovery.

Walking endurance
Not Effective
1B

One high quality RCT (Taveggia et al., 2016) and one fair quality RCT (Hidler et al., 2009) investigated the effect of exoskeleton gait trainers on walking endurance in the subacute phase of stroke recovery.

The high quality RCT (Taveggia et al., 2016) randomized patients to receive gait training using the Lokomat device or time-matched conventional physical therapy for gait retraining; both groups received additional physical therapy. Walking endurance was measured by the 6 Minute Walk Test (6MWT) at post-treatment (5 weeks) and follow-up (3 months). No significant between-group difference was found at either timepoint.

The fair quality RCT (Hidler et al., 2009) randomized patients to receive gait training using the Lokomat device or time-matched conventional gait training. Walking endurance was measured using the 6MWT at mid-treatment (12 sessions), post-treatment (24 sessions) and follow-up (3 months). A significant between-group difference was found at mid-treatment and post-treatment, in favour of conventional gait training vs. exoskeleton gait training. Differences did not remain significant at follow-up.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one fair quality RCT that exoskeleton gait trainers are not more effective than comparison interventions (conventional physical therapy, conventional gait training) for improving walking endurance in the subacute phase of stroke recovery.
Note: The fair quality RCT found that conventional gait training was more effective than exoskeleton gait training.

Walking speed
Not Effective
1A

Two high quality RCTs (Husemann et al., 2007; Taveggia et al., 2016) and two fair quality RCTs (Hidler et al., 2009; van Nunen et al., 2015) investigated the effect of exoskeleton gait trainers on walking speed in the subacute phase of stroke recovery.

The first high quality RCT (Husemann et al., 2007) randomized patients to receive gait training using the Lokomat device or conventional physical therapy. Walking speed was measured by the 10-meter walking test at post-treatment (4 weeks). No significant between-group difference was found.

The second high quality RCT (Taveggia et al., 2016) randomized patients to receive gait training using the Lokomat device or time-matched conventional physical therapy for gait retraining; both groups received additional physical therapy. Walking speed was measured by the 10-meter walking test at post-treatment (5 weeks) and follow-up (3 months). No significant between-group difference was found at either timepoint.

The first fair quality RCT (Hidler et al., 2009) randomized patients to receive gait training using the Lokomat device or time-matched conventional gait training. Walking speed was measured using the 5 meter walking test at mid-treatment (12 sessions), post-treatment (24 sessions) and follow-up (3 months). A significant between-group difference was found at all timepoints, in favour of conventional gait training vs. exoskeleton gait training.

The second fair quality RCT (van Nunen et al., 2015) randomized patients to receive gait training using the Lokomat device, or dose-matched overground gait training; both groups received additional physical therapy. Walking speed was measured by the 10-meter walking test at post-treatment (10 weeks) and follow-up (week 24, week 36). No significant between-group difference was found at any timepoint.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs and two fair quality RCTs that exoskeleton gait trainers are not more effective than comparison interventions (conventional physical therapy, conventional gait training, overground gait training) for improving walking speed in the subacute phase of stroke recovery.
Note: One fair quality RCT found that conventional gait training was more effective than exoskeleton gait training.

Chronic Phase - End-effector gait trainers

Activities of daily living
Not Effective
2A

One fair quality RCT (Dias et al., 2007) investigated the effect of end-effector gait trainers on Activities of daily living (ADLs) in patients with chronic stroke. The fair quality RCT randomized patients to receive gait training using the GT1 device or time-matched conventional rehabilitation. ADLs were measured using the Barthel Index (BI) at post-treatment (5 weeks) and follow-up (3 months). No significant between-group difference was found at either time point.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that end-effector gait trainers are not more effective than a comparison intervention (conventional rehabilitation) for improving Activities of daily living in the chronic phase of stroke recovery.

Balance
Not Effective
1B

One high quality RCT (Peurala et al., 2005) and one fair quality RCT (Dias et al., 2007) investigated the effect of end-effector gait trainers on balance in the chronic phase of stroke recovery.

The high quality RCT (Peurala et al., 2005) randomized patients to receive (1) gait training using the GT1 device, (2) gait training + Functional Electrical Stimulation (GT1+FES), or (3) overground walking training. Postural sway was measured using a force plate at mid-treatment (2 weeks), post-treatment (3 weeks) and follow-up (6 months). No significant between-group differences were found at any time point.

The fair quality RCT (Dias et al., 2007) randomized patients to receive gait training using the GT1 device or time-matched conventional rehabilitation. Balance was measured using the Toulouse Motor Scale (Balance: items 1-10; items 11-20; total), Berg Balance Scale (BBS) and the Step Test at post-treatment (5 weeks) and follow-up (3 months). No significant between-group differences were found on either measure at either time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one fair quality RCT that end-effector gait trainers are not more effective than comparison interventions (overground walking training, conventional rehabilitation) for improving balance in the chronic phase of stroke recovery.
Note: The high quality RCT also found that GT1+FES is not more effective than overground walking training for improving balance in the chronic phase of stroke recovery. There was no significant difference between end-effector gait training with/without FES.

Functional ambulation
Effective
1b

One high quality RCT (Geroin et al., 2011), one fair quality RCT (Dias et al., 2007) and one fair non-randomized study (Park et al., 2015) examined the effect of end-effector gait trainers on functional ambulation in the chronic phase of stroke recovery.

The high quality RCT (Geroin et al., 2011) randomized patients to receive (1) gait training using the GT1 device + sham stimulation (GT1 + sham stimulation), (2) gait training + transcranial direct current stimulation (GT1+tDCS), or (3) conventional overground walking exercises. Functional ambulation was measured using the Functional Ambulation Category (FAC) at post-treatment (2 weeks) and follow-up (4 weeks). A significant between-group difference was found at both time points in favour of GT1 + sham stimulation vs. conventional walking exercises.
Note: A significant between-group difference was found in favour of GT1+tDCS vs. conventional walking exercises at both time points. No significant difference was found between GT1 + sham stimulation vs. GT1+tDCS.

The fair quality RCT (Dias et al., 2007) randomized patients to receive gait training using the GT1 device or time-matched conventional rehabilitation. Functional ambulation was measured using the FAC at post-treatment (5 weeks) and follow-up (3 months). No significant between-group difference was found at either time point.

The fair non-randomized controlled trial (Park et al., 2015) assigned patients to receive gait training using the GT2 device or conventional overground gait training. Functional ambulation was measured using the FAC at post-treatment (4 weeks). No significant between-group difference was found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that end-effector gait trainers are more effective than a comparison intervention (conventional overground walking exercises) in improving functional ambulation in the chronic phase of stroke recovery. However, one fair quality RCT and fair one non-randomized study found that end-effector gait trainers are not more effective than comparison interventions (conventional rehabilitation, overground gait training).
Note: The high quality RCT also found that GT1+tDCS was more effective than conventional overground walking exercises. There was no significant difference between end-effector gait training with/without tDCS.

Functional independance
Not Effective
1B

One high quality RCT (Peurala et al., 2005) investigated the effect of end-effector gait trainers on functional independence in patients with chronic stroke. This high quality RCT randomized patients to receive (1) gait training using the GT1 device, (2) gait training + Functional Electrical Stimulation (GT1+FES), or (3) overground walking training. Functional independence was measured using the Functional Independence Measure at mid-treatment (2 weeks), post-treatment (3 weeks) and follow-up (6 months). No significant between-group differences were found at any time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that end-effector gait trainers are not more effective than a comparison intervention (overground walking training) for improving functional independence in the chronic phase of stroke recovery.
Note: This high quality RCT found that GT1+FES was not more effective than overground walking training. There was no significant difference between end-effector gait training with/without FES.

Gait parameters
Effective
1B

One high quality RCT (Geroin et al., 2011) and one fair non-randomized study (Park et al., 2015) examined the effect of end-effector gait trainers on gait parameters in the chronic phase of stroke recovery.

The high quality RCT (Geroin et al., 2011) randomized patients to receive (1) gait training using the GT1 device + sham stimulation (GT1 + sham stimulation), (2) gait training + transcranial direct current stimulation (GT1+tDCS), or (3) conventional overground walking exercises. Gait parameters were measured using the GAITRite system (Cadence, Temporal symmetry ratio of swing time to stance phase, Single to double support duration ratio) at post-treatment (2 weeks) and follow-up (4 weeks). Significant between-group differences were found on all gait parameters at both time points, in favour of GT1 + sham stimulation vs. conventional walking exercises.
Note: Significant between-group differences were found on all gait parameters at both time points in favour of GT1+tDCS vs. conventional walking exercises. No significant differences were found between GT1 + sham stimulation vs. GT1+tDCS.

The fair non-randomized controlled trial (Park et al., 2015) assigned patients to receive gait training using the GT2 device or conventional overground gait training. Spatiotemporal gait parameters were measured using the GAITRite system (Walking speed, Walking cycle, Stance phase and stride length of affected side, Symmetry index of stance phase and stride length) at post-treatment (4 weeks). No significant between-group differences were found on either measure.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that end-effector gait trainers (+ sham stimulation) are more effective than a comparison intervention (conventional overground walking exercises) for improving gait parameters in the chronic phase of stroke recovery.
Note: The high quality RCT also found that GT1+tDCS was more effective than conventional overground walking exercises for improving gait parameters. There was no significant difference between end-effector gait training with/without tDCS.
Note: A fair non-randomized study found that end-effector gait training was not more effective than conventional overground gait training.

Mobility
Effective
1B

One high quality RCT (Geroin et al., 2011) and one fair quality RCT (Dias et al., 2007) investigated the effect of end-effector gait trainers on mobility in the chronic phase of stroke recovery.

The high quality RCT (Geroin et al., 2011) randomized patients to receive (1) gait training using the GT1 device + sham stimulation (GT1 + sham stimulation), (2) gait training + transcranial direct current stimulation (GT1+tDCS), or (3) conventional overground walking exercises. Mobility was measured using the Rivermead Mobility Index (RMI) at post-treatment (2 weeks) and follow-up (4 weeks). A significant between-group difference was found at both time points in favour of GT1 + sham stimulation vs. conventional walking exercises.
Note: A significant between-group difference was found at both time points in favour of GT1+tDCS vs. conventional walking exercises. No significant difference was found between GT1 + sham stimulation vs. GT1+tDCS.

The fair quality RCT (Dias et al., 2007) randomized patients to receive gait training using the GT1 device or time-matched conventional rehabilitation. Mobility was measured using the RMI and the Timed Up and Go test at post-treatment (5 weeks) and follow-up (3 months). No significant between-group differences were found on either measure at either time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that end-effector gait trainers (+ sham stimulation) are more effective than a comparison intervention (conventional overground walking exercises) for improving mobility in the chronic phase of stroke recovery.
Note: However, one fair quality RCT found that end-effector gait trainers are not more effective than conventional gait training.
Note: The high quality RCT found that GT1+tDCS is more effective than conventional walking exercises. There is no significant difference between end-effector gait training with/without tDCS.

Motor function
Not Effective
1B

One high quality RCT (Peurala et al., 2005) and one fair quality RCT (Dias et al., 2007) investigated the effect of end-effector gait trainers on motor function in the chronic phase of stroke recovery.

The high quality RCT (Peurala et al., 2005) randomized patients to receive (1) gait training using the GT1 device, (2) gait trainer + Functional Electrical Stimulation (GT1+FES) group, or (3) overground walking training. Motor function was measured using the Modified Motor Assessment Scale at mid-treatment (2 weeks), post-treatment (3 weeks) and follow-up (6 months). No significant differences were found at any time point.

The fair quality RCT (Dias et al., 2007) randomized patients to receive gait training using the GT1 device or time-matched conventional rehabilitation. Motor function was measured using the Fugl-Meyer Stroke Scale at post-treatment (5 weeks) and follow-up (3 months). No significant between-group difference was found at either time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one fair quality RCT that end-effector gait trainers are not more effective than comparison interventions (overground walking training, conventional rehabilitation) for improving motor function in the chronic phase of stroke recovery.
Note: The high quality RCT also found that GT1+FES was not more effective than overground walking training. There was no significant difference between end-effector gait training with/without FES.

Muscle strength - lower extremity
Effective
1B

One high quality RCT (Geroin et al., 2011) and one fair quality RCT (Dias et al., 2007) investigated the effect of end-effector gait trainers on lower extremity muscle strength in the chronic phase of stroke recovery.

The high quality RCT (Geroin et al., 2011) randomized patients to receive (1) gait training using the GT1 device + sham stimulation (GT1 + sham stimulation), (2) gait training + transcranial direct current stimulation (GT1+tDCS), or (3) conventional overground walking exercises. Muscle strength was measured using the Motricity Index (MI – Leg score) at post-treatment (2 weeks) and follow-up (4 weeks). A significant between-group difference was found at both time points in favour of GT1 + sham stimulation vs. conventional walking exercises.
Note: A significant between-group difference was found at both time points in favour of GT1+tDCS vs. conventional walking exercises. No significant difference was found between GT1 + sham stimulation vs. GT1+tDCS.

The fair quality RCT (Dias et al., 2007) randomized patients to receive gait training using the GT1 device or time-matched conventional rehabilitation. Muscle strength was measured using the MI at post-treatment (5 weeks) and follow-up (3 months). No significant between-group difference was found at either time point.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that end-effector gait trainers (+ sham stimulation) are more effective than a comparison intervention (conventional overground walking exercises) for improving lower extremity muscle strength in the chronic phase of stroke recovery.
Note: One fair quality RCT found that end-effector gait trainers are not more effective than conventional rehabilitation.
Note: The high quality RCT found that GT1+tDCS is more effective than conventional walking exercises. There is no significant difference between end-effector gait training with/without tDCS.

Spasticity
Conflicting
4

Two high quality RCTs (Peurala et al., 2005; Geroin et al., 2011) and one fair quality RCT (Dias et al., 2007) investigated the effect of end-effector gait trainers on spasticity in the chronic phase of stroke recovery.

The first high quality RCT (Peurala et al., 2005) randomized patients to receive (1) gait training using the GT1 device, (2) gait training + Functional Electrical Stimulation (GT1+FES), or (3) overground walking training. Lower extremity spasticity was measured by the Modified Ashworth Scale (MAS) at mid-treatment (2 weeks), post-treatment (3 weeks) and follow-up (6 months). No significant differences were found at any time point.

The second high quality RCT (Geroin et al., 2011) randomized patients to receive (1) gait training using the GT1 device + sham stimulation (GT1 + sham stimulation), (2) gait training + transcranial direct current stimulation (GT1+tDCS), or (3) conventional overground walking exercises. Spasticity was measured using the MAS (hip adductors, quadriceps femoris, ankle plantiflexors) at post-treatment (2 weeks) and follow-up (4 weeks). A significant between-group difference was found at both time points in favour of GT1 + sham stimulation vs. conventional walking exercises.
Note: A significant between-group difference was found at both time points in favour of GT1+tDCS vs. conventional walking exercises. No significant difference was found between GT1 + sham stimulation vs. GT1+tDCS.

The fair quality RCT (Dias et al., 2007) randomized patients to receive gait training using the GT1 device or time-matched conventional rehabilitation. Spasticity was measured using the MAS at post-treatment (5 weeks) and follow-up (3 months). No significant between-group difference was found at either time point.

Conclusion: There is conflicting evidence (Level 4) regarding the effect of end-effector gait trainers on spasticity in the chronic phase of stroke recovery. One high quality RCT found that end-effector gait trainers were more effective than conventional walking exercises, whereas a second high quality RCT and a fair quality RCT found that end-effector gait trainers were not more effective than comparison interventions (overground walking training, conventional rehabilitation).
Note: A high quality RCT found that GT1+tDCS was more effective than conventional overground walking training. There was no significant difference between end-effector gait training with/without tDCS.

Walking endurance
Conflicting
4

Two high quality RCTs (Peurala et al., 2005; Geroin et al., 2011) and one fair quality RCT (Dias et al., 2007) investigated the effect of end-effector gait trainers on walking endurance in the chronic phase of stroke recovery.

The first high quality RCT Peurala et al., 2005) randomized patients to receive (1) gait training using the GT1 device, (2) gait training + Functional Electrical Stimulation (GT1+FES) group, or (3) overground walking training. Walking endurance was measured by the 6 Minute Walk Test (6MWT) at mid-treatment (2 weeks), post-treatment (3 weeks) and follow-up (6 months). No significant between-group differences were found at any time point.

The second high quality RCT (Geroin et al., 2011) randomized patients to receive (1) gait training using the GT1 device + sham stimulation (GT1 + sham stimulation), (2) gait training + transcranial direct current stimulation (GT1+tDCS), or (3) conventional overground walking exercises. Walking endurance was measured using the 6MWT at post-treatment (2 weeks) and follow-up (4 weeks). A significant between-group difference was found at both time points in favour of GT1 + sham stimulation vs. conventional walking exercises.
Note: A significant between-group difference was found at both time points in favour of GT1+tDCS vs. conventional walking exercises. No significant difference was found between GT1 + sham stimulation vs. GT1+tDCS at either time point.

The fair quality RCT (Dias et al., 2007) randomized patients to receive gait training using the GT1 device, or time-matched conventional rehabilitation. Walking endurance was measured using the 6MWT at post-treatment (5 weeks) and follow-up (3 months). No significant between-group difference was found at either time point.

Conclusion: There is conflicting evidence (Level 4) regarding the effect of end-effector gait trainers on walking endurance in the chronic phase of stroke recovery. One high quality RCT and one fair quality RCT found that end-effector gait trainers were not more effective than comparison interventions (overground walking training, conventional rehabilitation), whereas a second high quality RCT found that end-effector gait trainers were more effective than conventional walking exercises.
Note: One high quality RCT found that GT1+FES is not more effective than overground walking training, whereas a second high quality RCT found that GT1+tDCS is more effective than conventional overground walking exercises. There is no significant difference between end-effector gait training with/without FES, nor between end-effector gait training with/without tDCS.

Walking speed
Conflicting
4

Two high quality RCTs (Peurala et al., 2005; Geroin et al., 2011), two fair quality RCTs (Dias et al., 2007) and a fair non-randomized study (Park et al., 2015) investigated the effect of end-effector gait trainers on walking speed in the chronic phase of stroke recovery.

The first high quality RCT (Peurala et al., 2005) randomized patients to receive (1) gait training using the GT1 device, (2) gait training + Functional Electrical Stimulation (GT1+FES), or (3) overground walking training. Walking speed was measured using a 10-meter walking test at mid-treatment (2 weeks), post-treatment (3 weeks) and follow-up (6 months). No significant between-group differences were seen at any time point.

The second high quality RCT (Geroin et al., 2011) randomized patients to receive (1) gait training using the GT1 device + sham stimulation (GT1 + sham stimulation), (2) gait training + transcranial direct current stimulation (GT1+tDCS), or (3) conventional overground walking exercises. Walking speed was measured using a 10-meter walking test at post-treatment (2 weeks) and follow-up (4 weeks). A significant between-group difference was found at both time points in favour of GT1 + sham stimulation vs. conventional walking exercises.
Note: A significant between-group difference was found at both time points in favour of GT1+tDCS vs. conventional walking exercises. No significant difference was found between GT1 + sham stimulation vs. GT1+tDCS.

The fair quality RCT (Dias et al., 2007) randomized patients to receive gait training using the GT1 device or time-matched conventional rehabilitation. Walking speed was measured using a 10-meter walking test (Velocity, Step length, Step cadence with and without gait aid) at post-treatment (5 weeks) and follow-up (3 months). No significant between-group differences were found at either time point.

The fair non-randomized controlled trial (Park et al., 2015) assigned patients to receive gait training using the GT2 device or conventional overground gait training. Walking speed was measured using a 10-meter walking test at post-treatment (4 weeks). No significant between-group difference was found.

Conclusion: There is conflicting evidence (Level 4) regarding the effect of end-effector gait trainers on walking speed in the chronic phase of stroke recovery. One high quality RCT, one fair quality RCT and one fairnon-randomized study found that end-effector gait trainers were not more effective than comparison interventions (overground walking training, conventional rehabilitation, overground gait training), whereas a second high quality RCT found end-effector gait trainers were more effective than conventional overground walking exercises.
Note: One high quality RCT found that GT1+FES is not more effective than overground walking training; a second high quality RCT found that GT1+tDCS is more effective than conventional overground walking exercises. There is no significant difference between end-effector gait training with/without FES, nor between end-effector gait training with/without tDCS.

Chronic Phase - Exoskeleton gait trainers

Activities of daily living (ADLs)/Instrumental ADLs
Not Effective
1B

One high quality RCT (Kelley et al., 2013) and two fair quality RCTs (Hornby et al., 2008; Cho et al., 2015) investigated the effect of exoskeleton gait trainers on ADLs or IADLs in patients with chronic stroke.

The high quality RCT (Kelley et al., 2013) randomized patients to receive gait training using the Lokomat device or overground gait training. ADLs were measured by the Barthel Index (BI) at post-treatment (8 weeks) and follow-up (3 months). No significant between-group difference was found at either timepoint.

The first fair quality RCT (Hornby et al., 2008) randomized patients to receive locomotor training using the Lokomat device or therapist-assisted locomotor treadmill training. IADLs were measured using the Frenchay Activities Index at post-treatment (12 sessions) and follow-up (6 months). No significant between-group difference was found at either timepoint.

The second fair quality (crossover) RCT (Cho et al., 2015) randomized patients to receive gait training using the Lokomat device or no additional gait training; both groups received conventional physical therapy. ADLs were measured using the modified BI (mBI – Total, Transfers, Ambulation scores) at post-treatment (4 weeks, 8 weeks). A significant between-group difference was found in one measure (mBI – Transfers), in favour of exoskeleton gait training vs. no gait training.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and two fair quality RCTs that exoskeleton gait trainers are not more effective than no gait training or comparison interventions (overground gait training, treadmill training, no gait training) for improving ADLs/IADLs in the chronic phase of stroke recovery.

Ataxia
Not Effective
2A

One fair quality RCT (dos Santos et al., 2018) investigated the effect of an exoskeleton gait trainer on ataxia in the chronic phase of stroke recovery. The fair quality RCT randomized patients with chronic stroke and ataxia to receive gait training using the Lokomat 5.0 or therapist-assisted gait training. Ataxia was measured using the Scale for the Assessment and Rating of Ataxia at post-treatment (5 months). No significant between-group difference was found.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that exoskeleton gait trainers are not more effective than a comparison intervention (therapist-assisted gait training) for improving ataxia in the chronic phase of stroke recovery.

Balance
Conflicting
4

Two high quality RCTs (Westlake & Patten, 2009; Bang & Shin, 2016) and three fair quality RCTs (Hornby et al., 2008; Cho et al., 2015; dos Santos et al., 2018) investigated the effect of exoskeleton gait trainers on balance in the chronic phase or stroke recovery.

The first high quality RCT (Westlake & Patten, 2009) randomized patients to receive gait training using the Lokomat device or time-matched manually-assisted body-weight supported treadmill training. Balance was measured by the Berg Balance Scale (BBS) at post-treatment (4 weeks). No significant between-group difference was found.

The second high quality RCT (Bang & Shin, 2016) randomized patients to receive gait training using the Lokomat device or treadmill gait training. Balance was measured using the BBS at post-treatment (4 weeks). A significant between-group difference was found, in favour of exoskeleton gait training vs. treadmill gait training.

The first fair quality RCT (Hornby et al., 2008) randomized patients to receive locomotor training using the Lokomat device or therapist-assisted locomotor treadmill training. Balance was measured using the BBS at post-treatment (12 sessions) and follow-up (6 months). No significant between-group difference was found at either timepoint.

The second fair quality (crossover) RCT (Cho et al., 2015) randomized patients to receive gait training using the Lokomat device or no additional gait training; both groups received conventional physical therapy. Balance was measured using the BBS and the modified Functional Reach Test (Forward, Lateral) at post-treatment (4 weeks, 8 weeks). No significant between-group differences were found on either measure.

The third fair quality RCT (dos Santos et al., 2018) randomized patients with chronic stroke and ataxia to receive gait training using the Lokomat 5.0 or therapist-assisted gait training. Balance was measured using the BBS at post-treatment (5 months). No significant between-group difference was found.

Conclusion: There is conflicting evidence (Level 4) regarding the effect of exoskeleton gait trainers on balance in the chronic phase of stroke recovery. One high quality RCT and three fair quality RCTs found that the Lokomat device was not more effective than comparison interventions (manually-assisted body-weight supported treadmill training, therapist-assisted locomotor treadmill training, no gait training, therapist-assisted gait training), whereas one high quality RCT found that the Lokomat device was more effective than treadmill gait training.

Balance confidence
Effective
1B

One high quality RCT (Bang & Shin, 2016) investigated the effect of an exoskeleton gait trainer on balance confidence in the chronic phase or stroke recovery. This high quality RCT randomized patients to receive gait training using the Lokomat device or treadmill gait training. Balance confidence was measured using the Activities-Specific Balance Confidence scale at post-treatment (4 weeks). Significant between-group difference was found, in favour of exoskeleton gait training vs. treadmill gait training.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that exoskeleton gait trainers are more effective than a comparison intervention (treadmill gait training) for improving balance confidence in the chronic phase of stroke recovery.

Functional ambulation
Not Effective
2A

Two fair quality RCTs (Hornby et al., 2008; Cho et al., 2015) investigated the effect of exoskeleton gait trainers on functional ambulation in the chronic phase of stroke recovery.

The first fair quality RCT (Hornby et al., 2008) randomized patients to receive locomotor training using the Lokomat device or therapist-assisted locomotor treadmill training. Functional ambulation was measured using the modified Emory Functional Ambulation Profile at post-treatment (12 sessions) and follow-up (6 months). No significant between-group difference was found at either timepoint.

The second fair quality (crossover) RCT (Cho et al., 2015) randomized patients to receive gait training using the Lokomat device or no additional gait training; both groups received conventional physical therapy. Functional ambulation was measured using the Functional Ambulation Category at post-treatment (4 weeks, 8 weeks). No significant between-group difference was found.

Conclusion: There is limited evidence (Level 2a) from two fair quality RCTs that exoskeleton gait trainers are not more effective than no gait training or a comparison intervention (therapist-assisted treadmill training) for improving functional ambulation in the chronic phase of stroke recovery.

Functional independence
Not Effective
1B

One high quality RCT (Kelley et al., 2013) and one fair quality RCT (dos Santos et al., 2018) investigated the effect of exoskeleton gait trainers on functional independence in the chronic phase of stroke recovery.

The high quality RCT (Kelley et al., 2013) randomized patients to receive gait training using the Lokomat device or overground gait training. Functional independence was measured by the Functional Independence Measure (FIM – Locomotion subtest) at post-treatment (8 weeks) and follow-up (3 months). No significant between-group difference was found at either timepoint.

The fair quality RCT (dos Santos et al., 2018) randomized patients with chronic stroke and ataxia to receive gait training using the Lokomat 5.0 or therapist-assisted gait training. Functional independence was measured using the FIM at post-treatment (5 months). No significant between-group difference was found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one fair quality RCT that exoskeleton gait trainers are not more effective than comparison interventions (overground gait training, therapist-assisted gait training) for improving functional independence in the chronic phase of stroke recovery.

Gait parameters
Conflicting
4

Two high quality RCTs (Westlake & Patten 2009,; Bang & Shin, 2016) and one fair quality RCT (Hornby et al., 2008) investigated the effect of exoskeleton gait trainers on gait parameters in the chronic phase of stroke recovery.

The first high quality RCT (Westlake & Patten, 2009) randomized patients to receive gait training using the Lokomat device or time-matched manually-assisted body-weight supported treadmill training. Gait parameters were measured using the GAITRite system (Self-selected walking speed, Fast walking speed, Step length ratio of paretic limb) at post-treatment (4 weeks). No significant between-group differences were found.

The second high quality RCT (Bang & Shin, 2016) randomized patients to receive gait training using the Lokomat device or treadmill gait training. Gait parameters was measured using the GAITRite system (Gait speed, Cadence, Step length, Double limb support period) at post-treatment (4 weeks). Significant between-group differences were found on all measures, in favour of exoskeleton gait training vs. treadmill gait training.

The fair quality RCT (Hornby et al., 2008) randomized patients to receive gait training using the Lokomat device or therapist-assisted locomotor treadmill training. Gait parameters (% Single limb stance, Step asymmetry: self-selected velocity/fast velocity) were measured at post-treatment (12 sessions) and follow-up (6 months). A significant between-group difference in one measure (% Single limb stance – Fast velocity) was found at post-treatment, in favour of therapist-assisted locomotor treadmill training vs. exoskeleton gait training. Results did not remain significant at follow-up.

Conclusion: There is conflicting evidence (Level 4) regarding the effect of exoskeleton gait trainers on gait parameters in the chronic phase of stroke recovery. One high quality RCT and one fair quality RCT found that the Lokomat gait trainer was not more effective than comparison interventions (body-weight supported treadmill training, therapist-assisted treadmill training), whereas a second high quality RCT found that the Lokomat gait trainer was more effective than treadmill training.

Kinematics - lower extremity
Not Effective
1B

One high quality RCT (Lewek et al., 2009) investigated the effect of an exoskeleton gait trainer on lower extremity kinematics in the chronic phase of stroke recovery. The high quality RCT randomized patients to receive gait training using the Lokomat device or therapist-assisted treadmill training. Kinematic coordination was measured according to the consistency of intralimb hip and knee angular trajectories over gait cycles (average coefficient of correspondence (ACC): Hip – involved/uninvolved; Knee – involved/uninvolved) at post-treatment (4 weeks). No between-group differences were found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that an exoskeleton gait trainer is not more effective than a comparison intervention (treadmill training) for improving lower extremity kinematics in the chronic phase of stroke recovery.

Mobility
Conflicting
4

Two fair quality RCTs (Ukar, Paker, & Bugdayci, 2014; dos Santos et al., 2018) investigated the effect of exoskeleton gait trainers on mobility in the chronic phase of stroke recovery.

The first fair quality RCT (Ukar, Paker, & Bugdayci, 2014) randomized patients to receive gait training using the Lokomat device or time-matched conventional physical therapy home exercises. Mobility was measured using the Timed Up and Go test (TUG) at post-treatment (2 weeks) and follow-up (8 weeks). A significant between-group difference was found at both time points, in favour of exoskeleton gait training vs. conventional home exercises.

The second fair quality RCT (dos Santos et al., 2018) randomized patients with chronic stroke and ataxia to receive gait training using the Lokomat 5.0 or therapist-assisted gait training. Mobility was measured using the TUG at post-treatment (5 months). No significant between-group difference was found.

Conclusion: There is conflicting evidence (Level 4) regarding the effect of exoskeleton gait trainers on mobility in the chronic phase of stroke recovery. One fair quality RCT found that the Lokomat gait trainer was more effective than home exercises, whereas a second fair quality RCT found that the Lokomat device was not more effective than therapist-assisted gait training.

Motor function - lower extremity
Not Effective
1a

Two high quality RCTs (Westlake & Patten, 2009; Kelley et al., 2013) and one fair quality RCT (Cho et al., 2015) investigated the effect of exoskeleton gait trainers on lower extremity motor function in the chronic phase of stroke recovery.

The first high quality RCT (Westlake & Patten, 2009) randomized patients to receive gait training using the Lokomat device or time-matched manually-assisted body-weight supported treadmill training. Lower extremity motor function was measured using the Fugl-Meyer Assessment – Lower Extremity (FMA-LE) and the short physical performance battery at post-treatment (4 weeks). No significant between-group differences were found.

The second high quality RCT (Kelley et al., 2013) randomized patients to receive gait training using the Lokomat device or overground gait training. Lower extremity motor function was measured by the FMA-LE at post-treatment (8 weeks) and follow-up (3 months). No significant between-group difference was found at either timepoint.

The fair quality (crossover) RCT (Cho et al., 2015) randomized patients to receive gait training using the Lokomat device or no additional gait training; both groups received conventional physical therapy. Lower extremity motor function was measured using the FMA-LE at post-treatment (4 weeks, 8 weeks). No significant between-group difference was found.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs and one fair quality RCT that exoskeleton gait trainers are not more effective than comparison interventions (body-weight supported treadmill training, overground gait training, no additional gait training) for improving lower extremity motor function in the chronic phase of stroke recovery.

Muscle strength - lower extremity
Not Effective
2a

One fair quality RCT (Cho et al., 2015) investigated the effect of an exoskeleton gait trainer on lower extremity muscle strength in the chronic phase of stroke recovery. This fair quality (crossover) RCT randomized patients to receive gait training using the Lokomat device or no additional gait training; both groups received conventional physical therapy. Lower extremity muscle strength was measured using the Motricity Index at post-treatment (4 weeks, 8 weeks). No significant between-group difference was found.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that exoskeleton gait trainers are not more effective than no additional training for improving lower extremity strength in the chronic phase of stroke recovery.

Participation in life events
Not Effective
1b

One high quality RCT (Westlake & Patten, 2009) investigated the effect of an exoskeleton gait trainer on participation in life events in the chronic phase of stroke recovery. This high quality RCT randomized patients to receive gait training using the Lokomat device or time-matched manually-assisted body-weight supported treadmill training. Participation in life events was measured using the Late Life Function and Disability Instrument (LLFDI: Disability Frequency, Disability Limitation, Function) at post-treatment (4 weeks). No significant between-group differences were found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that an exoskeleton gait trainer is not more effective than a comparison intervention (body-weight supported treadmill training) for improving participation in life events in the chronic phase of stroke recovery.

Quality of life
Not Effective
2A

One fair quality RCT (Hornby et al., 2008) investigated the effect of an exoskeleton gait trainer on quality of life in the chronic phase of stroke recovery. This fair quality RCT randomized patients to receive gait training using the Lokomat device or therapist-assisted locomotor treadmill training. Quality of life was measured using the Medical Outcomes Questionnaire Short Form – 36 (SF-36 – Physical component summary score) at post-treatment (12 sessions) and follow-up (6 months). A significant between-group difference was found at post-treatment in a subgroup of participants with severe gait deficits (walking velocity ≤ 0.5m/s), in favour of therapist-assisted locomotor treadmill training vs. exoskeleton gait training. Results did not remain significant at follow-up.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that exoskeleton gait trainers are not more effective than a comparison intervention (therapist-assisted treadmill training) for improving quality of life in the chronic phase of stroke recovery.
Note: In fact, the fair quality RCT found that therapist-assisted treadmill training was more effective than the Lokomat gait trainer.

Spasticity
Not Effective
2A

One fair quality RCT (Cho et al., 2015) investigated the effect of an exoskeleton gait trainer on muscle tone in the chronic phase of stroke recovery. The fair quality (crossover) RCT randomized patients to receive gait training using the Lokomat device or no additional gait training; both groups received conventional physical therapy. Spasticity was measured using the Modified Ashworth Scale at post-treatment (4 weeks, 8 weeks). No significant between-group difference was found.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that an exoskeleton gait trainer is not more effective no additional gait training for reducing spasticity in the chronic phase of stroke recovery.

Stroke impact
Not Effective
1B

One high quality RCT (Kelley et al., 2013) investigated the effect of an exoskeleton gait trainer on stroke outcomes in the chronic phase of stroke recovery. This high quality RCT randomized patients to receive gait training using the Lokomat device or overground gait training. Stroke impact was measured by the Stroke Impact Scale (SIS – Strength, Mobility, ADL/IADL, Social participation, Total recovery scores) at post-treatment (8 weeks) and follow-up (3 months). No significant between-group differences were found at either timepoint.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that exoskeleton gait trainers are not more effective than a comparison intervention (overground gait training) for reducing the impact of stroke in the chronic phase of stroke recovery.

Walking endurance
Not Effective
1A

Two high quality RCTs (Westlake & Patten, 2009; Kelley et al., 2013) and one fair quality RCT (Hornby et al., 2008) investigated the effect of exoskeleton gait trainers on walking endurance in the chronic phase of stroke recovery.

The first high quality RCT (Westlake & Patten, 2009) randomized patients to receive gait training using the Lokomat device or time-matched manually-assisted body-weight supported treadmill training. Walking endurance was measured by the 6 Minute Walk Test (6MWT) at post-treatment (4 weeks). No significant between-group difference was found.

The second high quality RCT (Kelley et al., 2013) randomized patients to receive gait training using the Lokomat device or overground gait training. Walking endurance was measured by the 6MWT at post-treatment (8 weeks) and follow-up (3 months). No significant between-group difference was found at either timepoint.

The fair quality RCT (Hornby et al., 2008) randomized patients to receive gait training using the Lokomat device or therapist-assisted treadmill training. Walking endurance was measured using the 6MWT at post-treatment (12 sessions) and follow-up (6 months). No significant between-group difference was found at either timepoint.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs and one fair quality RCT that exoskeleton gait trainers are not more effective than comparison interventions (body-weight supported treadmill training, overground gait training, treadmill training) for improving walking endurance in the chronic phase of stroke recovery.

Walking speed
Not Effective
1B

One high quality RCT (Kelley et al., 2013) and two fair quality RCTs (Hornby et al., 2008; Ukar, Paker, & Bugdayci, 2014) investigated the effect of exoskeleton gait trainers on walking speed in the chronic phase of stroke recovery.

The high quality RCT (Kelley et al., 2013) randomized patients to receive gait training using the Lokomat device or overground gait training. Walking speed was measured by the 10-meter walking test at post-treatment (8 weeks) and follow-up (3 months). No significant between-group difference was found at either timepoint.

The first fair quality RCT (Hornby et al., 2008) randomized patients to receive gait training using the Lokomat device, or therapist-assisted treadmill training. Walking speed was measured by the 10-meter walking test using the GaitMat device (Self-selected velocity, Fast velocity) at post-treatment (12 sessions) and follow-up (6 months). Significant between-group differences on all measures were found at post-treatment, in favour of therapist-assisted treadmill training vs. exoskeleton gait training. Results did not remain significant at follow-up.

The second fair quality RCT (Ukar, Paker, & Bugdayci, 2014) randomized patients to receive gait training using the Lokomat device or time-matched conventional physical therapy home exercises. Walking speed was measured using the 10-meter walking test at post-treatment (2 weeks) and follow-up (8 weeks). A significant between-group difference was found at both time points, in favour of exoskeleton gait training vs. home exercises.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one fair quality RCT that exoskeleton gait trainers are not more effective than comparison interventions (overground gait training, treadmill training) for improving walking speed in the chronic phase of stroke recovery. In fact, the fair quality RCT found that therapist-assisted treadmill training was more effective than gait training using the Lokomat device.
Note: However, a second fair quality RCT found that gait training with the Lokomat device was more effective than home exercises.

Phase not specific to one period - End-effector gait trainers

Activities of daily living
Not Effective
1A

Four high quality RCTs (Tong et al., 2006; Ng et al., 2008; Morone et al., 2011; Chua, Culpan & Menon, 2016) examined the effect of end-effector gait trainers on activities of daily living (ADLs) following stroke.

The first high quality RCT (Tong et al., 2006) randomized patients with acute/subacute stroke to receive (1) gait training using the GT1 device, (2) gait training and Functional Electrical Stimulation (GT1+FES), or (3) conventional overground gait training. ADLs were measured using the Barthel Index (BI) at post-treatment (4 weeks). No significant between-group differences were found.

The second high quality RCT (Ng et al., 2008) randomized patients with acute/subacute stroke to receive (1) gait training using the GT2 device, (2) gait training + Functional Electrical Stimulation (GT2+FES), or (3) conventional gait training. ADLs were measured using the BI at post-treatment (4 weeks) and follow-up (6 months). No significant between-group differences were found at either timepoint.

The third high quality RCT (Morone et al., 2011) randomized patients with acute/subacute stroke to receive gait training using the GT1 device or conventional gait training. Patients were stratified by level of motor impairment (Motricity Index score ≤29: low motricity, > 29: high motricity). ADLs were measured using the BI at post-treatment (4 weeks) and at discharge (average 86-102 days post-stroke according to level of impairment); discharge results were reported. A significant difference was found between low motricity groups only, in favour of end-effector gait training vs. conventional gait training.

The fourth high quality RCT (Chua, Culpan & Menon, 2016) randomized patients with acute/subacute stroke to receive gait training using the GT1 device or time-matched conventional physical therapy. ADLs was measured by the BI at mid-treatment (4 weeks), post-treatment (8 weeks) and follow-up (12 weeks, 24 weeks, 48 weeks). No significant between-group difference was found at any time point.

Conclusion: There is strong evidence (Level 1a) from four high quality RCTs that end-effector gait trainers are not more effective than comparison interventions (conventional overground gait training, physical therapy) for improving ADLs following stroke.
Note: One high quality RCT found that end-effector gait training is more effective than conventional gait training for improving ADLs in patients with low motricity (MI score ≤29).
Note: Two high quality RCTs found that end-effector gait training + Functional Electrical Stimulation is not more effective than conventional overground gait training. There is no significant difference between end-effector gait training with/without FES.

Balance
Not Effective
1a

Four high quality RCTs (Tong et al., 2006; Ng et al., 2008; Freivogel, Schmalohr & Mehrholz, 2009; Morone et al., 2011) investigated the effect of end-effector gait trainers on balance following stroke.

The first high quality RCT (Tong et al., 2006) randomized patients with acute/subacute stroke to receive (1) gait training using the GT1 device, (2) gait training and Functional Electrical Stimulation (GT1+FES), or (3) conventional overground gait training. Balance was measured using the Berg Balance Scale (BBS) at mid-treatment (2 weeks) and post-treatment (4 weeks). No significant between-group differences were found at either time point.

The second high quality RCT (Ng et al., 2008) randomized patients with acute/subacute stroke to receive (1) gait training using the GT2 device, (2) gait training + Functional Electrical Stimulation (GT2+FES), or (3) conventional gait training. Balance was measured by the BBS at post-treatment (4 weeks) and follow-up (6 months). No significant between-group differences were found at either timepoint.

The third high quality cross-over RCT (Freivogel, Schmalohr & Mehrholz, 2009) randomized patients with subacute/chronic stroke to receive gait training using the LokoHelp electromechanical device or conventional treadmill/overground gait training. Balance was measured using the BBS at post-treatment (6 weeks). No significant between-group difference was found.

The fourth high quality RCT (Morone et al., 2011) randomized patients with acute/subacute stroke to receive gait training using the GT1 device or conventional gait training. Patients were stratified by level of motor impairment (Motricity Index score ≤29: low motricity, > 29: high motricity). Balance was measured using the Trunk Control Test at post-treatment (4 weeks) and at discharge (average 86-102 days post-stroke according to level of impairment); discharge results were reported. A significant difference was found between low motricity groups only, in favour of end-effector gait training vs. conventional gait training.

Conclusion: There is strong evidence (Level 1a) from four high quality RCTs that end-effector gait trainers are not more effective than comparison interventions (conventional overground gait training, conventional treadmill training) following stroke.
Note: One high quality RCT found that end-effector gait training is more effective than conventional gait training for improving balance in patients with low motricity (MI score ≤ 29).
Note: Two high quality RCTs found that end-effector gait training with Functional Electrical Stimulation (FES) is not more effective than conventional overground gait training. There is no difference between end-effector gait training with/without FES.

Disability
Effective
1b

One high quality RCT (Morone et al., 2011) investigated the effect of an end-effector gait trainer on disability following stroke. The high quality RCT randomized patients with acute/subacute stroke to receive gait training using the GT1 device or conventional gait training. Patients were stratified by level of motor impairment (Motricity Index score ≤ 29: low motricity, > 29: high motricity). Disability was measured using the Rankin Scale at post-treatment (4 weeks) and at discharge (average 86-102 days post-stroke according to level of impairment); discharge results were reported. A significant difference was found between low motricity groups only, in favour of end-effector gait training vs. conventional gait training.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that end-effector gait training is more effective than a comparison intervention (conventional gait training) for reducing disability among patients with low motricity following stroke.
Note: End-effector gait training is not more effective than conventional gait training for improving disability among patients with high motricity.

Functional ambulation
Conflicting
4

Five high quality RCTs (Tong et al., 2006; Ng et al., 2008; Freivogel, Schmalohr & Mehrholz, 2009; Morone et al., 2011; Chua, Culpan & Menon, 2016) and one non-randomized study (Hesse et al., 2001) investigated the effect of end-effector gait trainers on functional ambulation following stroke.

The first high quality RCT (Tong et al., 2006) randomized patients with acute/subacute stroke to receive (1) gait Training using the GT1 device, (2) gait training and Functional Electrical Stimulation (GT1+FES), or (3) conventional overground gait training. Functional ambulation was measured using the Functional Ambulation Categories (FAC) at mid-treatment (2 weeks) and post-treatment (4 weeks). Significant between-group differences were found at both time points, in favour of end-effector gait training vs. conventional overground gait training.
Note: Significant between-group differences in functional ambulation were found at both time points, in favour of GT1+FES vs. conventional overground gait training. There were no significant differences between end-effector gait training vs. GT1+FES.

The second high quality RCT (Ng et al., 2008) randomized patients with acute/subacute stroke to receive (1) gait training using the GT2 device, (2) gait training + Functional Electrical Stimulation (GT2+FES), or (3) conventional gait training. Functional ambulation was measured by the FAC at post-treatment (4 weeks) and follow-up (6 months). A significant difference was found at follow-up only, in favour of end-effector gait training vs. conventional gait training.
Note: A significant difference was found at both timepoints, in favour of GT2+FES vs. conventional gait training. No significant difference was found between end-effector gait training vs. GT2+FES at either timepoint.

The third high quality cross-over RCT (Freivogel, Schmalohr & Mehrholz, 2009) randomized patients with subacute/chronic stroke to receive gait training using the LokoHelp electromechanical device or conventional treadmill/overground gait training. Functional ambulation was measured using the FAC at post-treatment (6 weeks). No significant between-group difference was found.

The fourth high quality RCT (Morone et al., 2011) randomized patients with acute/subacute stroke to receive gait training using the GT1 device or conventional gait training. Patients were stratified by level of motor impairment (Motricity Index score ≤ 29: low motricity, > 29: high motricity). Functional ambulation was measured using the FAC at post-treatment (4 weeks) and at discharge (average 86-102 days post-stroke according to level of impairment); discharge results were reported. A significant difference was found between low motricity groups only, in favour of end-effector gait training vs. conventional gait training.

The fifth high quality RCT (Chua, Culpan & Menon, 2016) randomized patients with acute/subacute stroke to receive gait training using the GT1 device or time-matched conventional physical therapy. Functional ambulation was measured using the FAC at mid-treatment (4 weeks), at post-treatment (8 weeks), and follow-up (12 weeks, 24 weeks, 48 weeks). No significant between-group difference was found at any time point.

The non-randomized controlled trial (Hesse et al., 2001) assigned patients with subacute/chronic stroke who were wheelchair-dependent to receive gait training using the GT1 device in addition to conventional physical therapy. Functional ambulation was measured using the FAC at post-treatment (4 weeks). No significant improvement was seen.

Conclusion: There is conflicting evidence (Level 4) regarding the effect of end-effector gait trainers on functional ambulation following stroke. One high quality RCT found that end-effector gait trainers are more effective than conventional overground gait training at mid-treatment and post-treatment; another high quality RCT found that end-effector gait trainers are more effective than conventional gait training in the long term; and a third high quality RCT found that end-effector gait trainers are more effective than conventional gait training in patients with low motricity but not those with high motricity. However, two high quality RCTs and one non-randomized study found that end-effector gait trainers are not more effective than comparison interventions (conventional treadmill/overground gait training, conventional physical therapy).
Note: Two high quality RCTs found that end-effector gait training + Functional Electrical Stimulation (FES) is more effective than overground gait training. There is no difference between end-effector gait training with/without FES.

Functional independence
Not Effective
1a

Two high quality RCTs (Tong et al., 2006; Ng et al., 2008) investigated the effect of end-effector gait trainers on functional independence following stroke.

The first high quality RCT (Tong et al., 2006) randomized patients with acute/subacute stroke to receive (1) gait training using the GT1 device, (2) gait training and Functional Electrical Stimulation (GT1+FES), or (3) conventional overground gait training. Functional independence was measured using the Functional Independence Measure (FIM) at post-treatment (4 weeks). No significant between-group differences were found.

The second high quality RCT (Ng et al., 2008) randomized patients with acute/subacute stroke to receive (1) gait training using the GT2 device, (2) gait training + Functional Electrical Stimulation (GT2+FES), or (3) conventional gait training. Functional independence was measured by the Functional Independence Measure (FIM) at post-treatment (4 weeks) and follow-up (6 months). No significant differences were found at either timepoint.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that end-effector gait trainers are not more effective than comparison interventions (conventional overground gait training) for improving functional independence following stroke.
Note: Two high quality RCTs found that end-effector gait training with Functional Electrical Stimulation (FES) is not more effective than conventional overground gait training. There is no difference between end-effector gait training with/without FES.

Gait parameters
Effective
2b

One non-randomized study (Hesse et al., 2001) investigated the effect of end-effector gait trainers on gait parameters following stroke. The non-randomized controlled trial assigned patients with subacute/chronic stroke who were wheelchair-dependent to receive gait training using the GT1 device in addition to conventional physical therapy. Gait parameters (velocity, cadence, stride length) and limb-dependent cycle parameters (single-stance period, double-stance phase, stance duration, swing duration, swing symmetry, stance symmetry) were measured at post-treatment (4 weeks). Significant improvements were seen (velocity, cadence, stride length, single-stance period, terminal double-stance phase, swing symmetry).

Conclusion: There is limited evidence (Level 2b) from one non-randomized study that end-effector gait trainers are effective for improving gait parameters following stroke.

Mobility
Conflicting
4

Four high quality RCTs (Tong et al., 2006; Ng et al., 2008; Freivogel, Schmalohr & Mehrholz, 2009; Morone et al., 2011) and one non-randomized study (Hesse et al., 2001) investigated the effect of end-effector gait trainers on mobility following stroke.

The first high quality RCT (Tong et al., 2006) randomized patients with acute/subacute stroke to receive (1) gait training using the GT1 device, (2) gait training and Functional Electrical Stimulation (GT1+FES), or (3) conventional overground gait training. Mobility was measured using the Elderly Mobility Scale at mid-treatment (2 weeks) and post-treatment (4 weeks). A significant between-group difference was found at post-treatment, in favour of end-effector gait training vs. conventional overground gait training.
Note: A significant between-group differences was found at post-treatment in favour of GT1+FES vs. conventional overground gait training. No significant difference between end-effector gait training vs. GT1+FES was found at either timepoint.

The second high quality RCT (Ng et al., 2008) randomized patients with acute/subacute stroke to receive (1) gait training using the GT2 device, (2) gait training + Functional Electrical Stimulation (GT2+FES), or (3) conventional gait training. Mobility was measured by the Elderly Mobility Scale at post-treatment (4 weeks) and follow-up (6 months). Significant between-group differences were found at both timepoints, in favour of end-effector gait training vs. conventional gait training.
Note: Significant between-group differences were found at both timepoints, in favour of GT2+FES vs. conventional gait training. No significant difference was found between end-effector gait training vs. GT2+FES at either timepoint.

The third high quality cross-over RCT (Freivogel, Schmalohr & Mehrholz, 2009) randomized patients with subacute/chronic stroke to receive gait training using the LokoHelp electromechanical device or conventional treadmill/overground gait training. Mobility was measured using the Rivermead Mobility Index (RMI) at post-treatment (6 weeks). No significant between-group difference was found.

The fourth high quality RCT (Morone et al., 2011) randomized patients with acute/subacute stroke to receive gait training using the GT1 device or conventional gait training. Patients were stratified by level of motor impairment (Motricity Index score ≤29: low motricity, > 29: high motricity). Mobility was measured using the RMI at post-treatment (4 weeks) and at discharge (average 86-102 days post-stroke according to level of impairment); discharge results were reported. A significant difference was found between low motricity groups only, in favour of end-effector gait training vs. conventional gait training.

The The non-randomized controlled trial (Hesse et al., 2001) assigned patients with subacute/chronic stroke who were wheelchair-dependent to receive gait training using the GT1 device in addition to conventional physical therapy. Functional mobility was measured using the Rivermead Motor Assessment (Gross function, Legs and trunk scores) at post-treatment (4 weeks). No significant improvement was seen.

Conclusion: There is conflicting evidence (Level 4) regarding the effect of end-effector gait trainers on mobility following stroke. Two high quality RCTs found that end-effector gait trainers are more effective than conventional overground gait training, whereas another high quality RCT found that end-effector gait trainers are not more effective than conventional treadmill/overground gait training; these studies used different devices and outcome measures. A fourth high quality RCT found that an end-effector gait trainer was more effective than conventional gait training among patients with low motricity but not those with high motricity.
Note: Two high quality RCTs found that end-effector gait training with Functional Electrical Stimulation (FES) is more effective than conventional overground gait training. There is no difference between end-effector gait training with/without FES.

Muscle strength - lower extremity
Not Effective
1A

Four high quality RCTs (Tong et al., 2006; Ng et al., 2008; Freivogel, Schmalohr & Mehrholz, 2009; Morone et al., 2011) investigated the effect of end-effector gait trainers on lower extremity muscle strength following stroke.

Four high quality RCTs (Tong et al., 2006; Ng et al., 2008; Freivogel, Schmalohr & Mehrholz, 2009; Morone et al., 2011) investigated the effect of end-effector gait trainers on lower extremity muscle strength following stroke.

Always use the online WYSIWYG HTML beautifier to compose the content for your website easily. This is a website that is worth using.
Note: A significant between-group difference was found at post-treatment in favour of GT1+FES vs. conventional overground gait training. There was no significant difference between end-effector gait training vs. GT1+FES.

The second high quality RCT (Ng et al., 2008) randomized patients with acute/subacute stroke to receive (1) gait training using the GT2 device, (2) gait training + Functional Electrical Stimulation (GT2+FES), or (3) conventional gait training. Muscle strength was measured by the MI (Leg score) at post-treatment (4 weeks) and follow-up (6 months). No significant between-group differences were found at either timepoint.

The third high quality cross-over RCT (Freivogel, Schmalohr & Mehrholz, 2009) randomized patients with subacute/chronic stroke to receive gait training using the LokoHelp electromechanical device or conventional treadmill/overground gait training. Muscle strength was measured using the MI (Leg score) at post-treatment (6 weeks). No significant between-group difference was found.

The fourth high quality RCT (Morone et al., 2011) randomized patients with acute/subacute stroke to receive gait training using the GT1 device or conventional gait training. Patients were stratified by level of motor impairment (MI score ≤29: low motricity, > 29: high motricity). Muscle strength was measured using the Motricity Index at post-treatment (4 weeks) and at discharge (average 86-102 days post-stroke according to level of impairment); discharge results were reported. No significant between-group differences were found at either timepoint.

Conclusion: There is strong evidence (Level 1a) from three high quality RCTs that end-effector gait trainers are not more effective than comparison interventions (conventional overground/treadmill gait training) for improving lower extremity muscle strength following stroke.
Note: However, one high quality RCT found that end-effector gait training is more effective than conventional overground gait training.
Note: One high quality RCT found that end-effector gait training with Functional Electrical Stimulation (FES) is more effective than conventional overground gait training. Two high quality RCTs found no significant difference between end-effector gait training with/without FES.

Neurological status
Not Effective
1B

One high quality RCT (Morone et al., 2011) investigated the effect of end-effector gait training on neurological status following stroke. The high quality RCT randomized patients with acute/subacute stroke to receive gait training using the GT1 device or conventional gait training. Patients were stratified by level of motor impairment (Motricity Index score ≤29: low motricity, > 29: high motricity). Neurological status was measured using the Canadian Neurological Scale at post-treatment (4 weeks) and at discharge (average 86-102 days post-stroke according to level of impairment); discharge results were reported. No significant between-group differences were found at either time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that end-effector gait trainers are not more effective than a comparison intervention (conventional gait training) for improving neurological status following stroke.

Spasticity
Not Effective
1A

Two high quality RCTs (Freivogel, Schmalohr & Mehrholz, 2009; Morone et al., 2011) and one non-randomized study (Hesse et al., 2001) investigated the effect of end-effector gait training on spasticity following stroke.

The first high quality cross-over RCT (Freivogel, Schmalohr & Mehrholz, 2009) randomized patients with subacute/chronic stroke to receive gait training using the LokoHelp electromechanical device or conventional treadmill/overground gait training. Spasticity was measured using the Modified Ashworth Scale at post-treatment (6 weeks). No significant between-group difference was found.

The second high quality RCT (Morone et al., 2011) randomized patients with acute/subacute stroke to receive gait training using the GT1 device or conventional gait training. Patients were stratified by level of motor impairment (Motricity Index score ≤ 29: low motricity, > 29: high motricity). Spasticity was measured using the Ashworth Scale at post-treatment (4 weeks) and at discharge (average 86-102 days post-stroke according to level of impairment); discharge results were reported. No significant between-group differences were found at either time point.

The non-randomized controlled trial (Hesse et al., 2001) assigned patients with subacute/chronic stroke who were wheelchair-dependent to receive gait training using the GT1 device in addition to conventional physical therapy. Spasticity was measured using the Modified Ashworth Scale at post-treatment (4 weeks). No significant improvement was seen.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that end-effector gait trainers are not more effective than comparison interventions (conventional overground/treadmill gait training,conventional gait training) for reducing spasticity following stroke. A non-randomized study also reported no significant improvement in spasticity following end-effector gait training.

Stroke impact
Not Effective
1B

One high quality RCT (Chua, Culpan & Menon, 2016) investigated the effect of end-effector gait training on stroke. The high quality RCT randomized patients with acute/subacute stroke to receive gait training using the GT1 device or time-matched conventional physical therapy. Stroke impact was measured by the Stroke Impact Scale (SIS – Physical, Memory and thinking, Mood and emotion, Communication, Participation, Recovery domains) at mid-treatment (4 weeks), post-treatment (8 weeks) and follow-up (12 weeks, 24 weeks, 48 weeks). No significant between-group difference was found at any time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that end-effector gait trainers are not more effective than a comparison intervention (conventional physical therapy) for reducing the impact of stroke.

Walking endurance
Not Effective
1A

Two high quality RCTs (Morone et al., 2011; Chua, Culpan & Menon, 2016) investigated the effect of end-effector gait training on walking endurance following stroke.

The first high quality RCT (Morone et al., 2011) randomized patients with acute/subacute stroke to receive gait training using the GT1 device or conventional gait training. Patients were stratified by level of motor impairment (Motricity Index score ≤29: low motricity, > 29: high motricity). Walking endurance was measured using the 6 Minute Walk Test (6MWT) at post-treatment (4 weeks) and at discharge (average 86-102 days post-stroke according to level of impairment); discharge results were reported. A significant difference was found between low motricity groups only, in favour of end-effector gait training vs. conventional gait training.

The second high quality RCT (Chua, Culpan & Menon, 2016) randomized patients with acute/subacute stroke to receive gait training using the GT1 device or time-matched conventional physical therapy. Walking endurance was measured by the 6MWT at mid-treatment (4 weeks), post-treatment (8 weeks) and follow-up (12 weeks, 24 weeks, 48 weeks). No significant between-group difference was found at any time point.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that end-effector gait trainers are not more effective than comparison interventions (conventional gait training, conventional physical therapy) in improving walking endurance following stroke.
Note: However, one of these studies found that end-effector gait trainers are more effective than conventional gait training among patients with low motricity (but not those with high motricity).

Walking speed
Conflicting
4

Five high quality RCTs (Tong et al., 2006; Ng et al., 2008; Freivogel, Schmalohr & Mehrholz, 2009; Morone et al., 2011; Chua, Culpan & Menon, 2016) investigated the effect of end-effector gait training on walking speed following stroke.

The first high quality RCT (Tong et al., 2006) randomized patients with acute/subacute stroke to receive (1) gait training using the GT1 device, (2) gait training and Functional Electrical Stimulation (GT1+FES), or (3) conventional overground gait training. Walking speed was measured using the 5-meter walking test at mid-treatment (2 weeks) and post-treatment (4 weeks). A significant between-group difference was found at post-treatment, in favour of end-effector gait training vs. conventional overground gait training.
Note: Significant between-group differences in walking speed were found at both timepoints, in favour of GT1+FES vs. conventional overground gait training. There were no significant differences between end-effector gait training vs. GT1+FES.

The second high quality RCT (Ng et al., 2008) randomized patients with acute/subacute stroke to receive (1) gait training using the GT2 device, (2) gait training + Functional Electrical Stimulation (GT2+FES), or (3) conventional gait training. Walking speed was measured using the 5 meter walking test at post-treatment (4 weeks) and follow-up (6 months). A significant between-group difference was found at both timepoints, in favour of end-effector gait training vs. conventional gait training.
Note: A significant between-group difference was found at both timepoints, in favour of GT2+FES vs. conventional gait training. No significant difference was found between end-effector gait training vs. GT2+FES at either timepoint.

The third high quality cross-over RCT (Freivogel, Schmalohr & Mehrholz, 2009) randomized patients with subacute/chronic stroke to receive gait training using the LokoHelp electromechanical device or conventional treadmill/overground gait training. Walking speed was measured using the 10-meter walking test at post-treatment (6 weeks). No significant between-group difference was found.

The fourth high quality RCT (Morone et al., 2011) randomized patients with acute/subacute stroke to receive gait training using the GT1 device or conventional gait training. Patients were stratified by level of motor impairment (Motricity Index score ≤ 29: low motricity, > 29: high motricity). Walking speed was measured using the 10-meter walking test at post-treatment (4 weeks) and at discharge (average 86-102 days post-stroke according to level of impairment); discharge results were reported. No significant between-group difference was found at any time point.

The fifth high quality RCT (Chua, Culpan & Menon, 2016) randomized patients with acute/subacute stroke to receive gait training using the GT1 device or time-matched conventional physical therapy. Walking speed was measured by the 10-meter walking test at mid-treatment (4 weeks), post-treatment (8 weeks) and follow-up (12 weeks, 24 weeks, 48 weeks). No significant between-group difference was found at any time point.

Conclusion: There is conflicting evidence (Level 4) regarding the effect of end-effector gait trainers on walking speed following stroke. Two high quality RCTs found that end-effector gait trainers are more effective than a comparison intervention (conventional overground gait training), whereas three high quality RCTs found that end-effector gait trainers are not more effective than comparison interventions (conventional treadmill/overground gait training, conventional physical therapy).
Note: Two high quality RCTs found that end-effector gait training + FES is more effective than conventional overground gait training. There is no difference between end-effector gait training with/without FES.

Phase not specific to one period - Exoskeleton gait trainers

Activities of daily living
Not Effective
2A

One fair quality RCT (Kim et al., 2015) investigated the effect of an exoskeleton gait trainer on ADLs following stroke. The fair quality RCT randomized patients with subacute/chronic stroke to receive gait training using the WALKBOT device or time-matched conventional locomotor training; both groups received additional conventional locomotor training. ADLs were measured using the Korean modified Barthel Index (Grooming, Bathing, Feeding, Toilet use, Stairs, Dressing, Bowels, Bladder, Ambulation, Transfers, Total scores) at post-treatment (4 weeks) and follow-up (8 weeks). Significant between-group differences were found on only three ADL scores at both timepoints (Dressing, Ambulation, Total score), in favour of exoskeleton gait training vs. conventional locomotor training.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that an exoskeleton gait trainer is not more effective than a comparison intervention (conventional locomotor training) for improving Activities of daily living following stroke.

Balance
Effective
2a

One fair quality RCT (Kim et al., 2015) and one non-randomized study (Dundar et al., 2014) investigated the effect of exoskeleton gait trainers on balance following stroke.

The fair quality RCT (Kim et al., 2015) randomized patients with subacute/chronic stroke to receive gait training using the WALKBOT device or time-matched conventional locomotor training; both groups received additional conventional locomotor training. Balance was measured using the Berg Balance Scale (BBS) at post-treatment (4 weeks) and follow-up (8 weeks). A significant between-group difference was found at both timepoints, in favour of exoskeleton gait training vs. conventional locomotor training.

The non-randomized retrospective study (Dundar et al., 2014) compared patients with subacute/chronic stroke who had received gait training using the Lokomat device with those who had received physical therapy. Balance was measured using the BBS at post-treatment (minimum 30 sessions). No significant between-group difference was found.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that an exoskeleton gait trainer is more effective than a comparison intervention (conventional locomotor training) for improving balance following stroke.
Note: However, a non-randomized study found that the Lokomat device was not more effective than conventional physical therapy.

Cognition
Effective
2B

One non-randomized study (Dundar et al., 2014) investigated the effect of exoskeleton gait trainers on cognition following stroke. The non-randomized retrospective study compared patients with subacute/chronic stroke who had received gait training using the Lokomat device with those who had received physical therapy. Cognition was measured using the Mini Mental Status Examination at post-treatment (minimum 30 sessions). A significant between-group difference was found, in favour of exoskeleton gait training vs. physical therapy.

Conclusion: There is limited evidence (Level 2b) from one non-randomized study that an exoskeleton gait trainer is more effective than a comparison intervention (physical therapy) for improving cognition following stroke.

Functional ambulation
Effective
1B

One high quality RCT (Schwartz et al., 2009), one fair quality RCT (Kim et al., 2015) and one non-randomized study (Dundar et al., 2014) investigated the effect of exoskeleton gait trainers on functional ambulation following stroke.

The high quality RCT (Schwartz et al., 2009) randomized patients with acute/subacute stroke to receive gait training using the Lokomat device or time-matched conventional physical therapy for gait retraining; both groups received additional physical therapy. Functional ambulation was measured by the Functional Ambulation Category (FAC) at post-treatment (6 weeks). A significant between-group difference was found, in favour of Lokomat training vs. conventional physical therapy.

The fair quality RCT (Kim et al., 2015) randomized patients with stroke to receive gait training using the WALKBOT device or time-matched conventional locomotor training; both groups received additional conventional locomotor training. Functional ambulation was measured using the FAC at post-treatment (4 weeks) and follow-up (8 weeks). A significant between-group difference was found at both timepoints, in favour of exoskeleton gait training vs. conventional locomotor training.

The non-randomized retrospective study (Dundar et al., 2014) compared patients with subacute/chronic stroke who had received gait training using the Lokomat device with those who had received physical therapy. Functional ambulation was measured using the FAC at post-treatment (minimum 30 sessions). No significant between-group difference was found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one fair quality RCT that exoskeleton gait trainers are more effective than comparison interventions (physical therapy, conventional locomotor training) for improving functional ambulation following stroke.

Functional independence
Effective
1b

One high quality RCT (Schwartz et al., 2009) and one non-randomized study (Dundar et al., 2014) investigated the effect of exoskeleton gait trainers on functional independence following stroke.

The high quality RCT (Schwartz et al., 2009) randomized patients with acute/subacute stroke to receive gait training using the Lokomat device or time-matched conventional physical therapy for gait retraining; both groups received additional physical therapy. Functional independence was measured by the Functional Independence Measure (FIM – Motor, Cognition) at post-treatment (6 weeks). A significant between-group difference was found in one measure (FIM – Motor), in favour of Lokomat training vs. conventional physical therapy.

The non-randomized retrospective study (Dundar et al., 2014) compared patients with subacute/chronic stroke who had received gait training using the Lokomat device with those who had received physical therapy. Functional independence was measured using the FIM at post-treatment (minimum 30 sessions). A significant between-group difference was found, in favour of exoskeleton gait training vs. physical therapy.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one non-randomized study that exoskeleton gait trainers are more effective than comparison interventions (physical therapy) for improving functional independence following stroke.

Health related quality of life
Not Effective
2A

One fair quality RCT (Kim et al., 2015) and one non-randomized study (Dundar et al., 2014) investigated the effect of exoskeleton gait trainers on Health related quality of life (HRQoL) following stroke.

The fair quality RCT (Kim et al., 2015) randomized patients with subacute/chronic stroke to receive gait training using the WALKBOT device, or time-matched conventional locomotor training; both groups received additional conventional locomotor training. HRQoL was measured using the EuroQoL 5-dimension (EQ-5D) at post-treatment (4 weeks) and follow-up (8 weeks). No significant between-group difference was found at either timepoint.

The non-randomized retrospective study (Dundar et al., 2014) compared patients with subacute/chronic stroke who had received gait training using the Lokomat device with those who had received physical therapy. HRQoL was measured using the Medical Outcomes Study Short Form 36 (SF-36 – Physical functioning, Physical role limitations, Pain, General health, Social functioning, General mental health, Emotional role limitations, Vitality, Physical component, Mental component) at post-treatment (minimum 30 sessions). Significant between-group differences were found on all SF-36 scores, in favour of exoskeleton gait training vs. physical therapy.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that exoskeleton gait trainers are not more effective than a comparison intervention (conventional locomotor training) for improving quality of life following stroke.
Note: However, a non-randomized study found that the Lokomat device was more effective than physical therapy.

Mobility
Not Effective
1B

One high quality RCT (Schwartz et al., 2009) investigated the effect of an exoskeleton gait trainer on mobility following stroke. The high quality RCT randomized patients with acute/subacute stroke to receive gait training using the Lokomat device or time-matched conventional physical therapy for gait retraining; both groups received additional physical therapy. Mobility was measured by the Timed Up and Go test and the Stroke Activity Scale (Walking, Standing scores) at post-treatment (6 weeks). No significant between-group differences were found on either measure.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that an exoskeleton gait trainer is not more effective than a comparison intervention (conventional physical therapy) for improving mobility following stroke.

Motor recovery
Effective
2B

One non-randomized study (Dundar et al., 2014) investigated the effect of an exoskeleton gait trainer on motor recovery following stroke. The non-randomized retrospective study compared patients with subacute/chronic stroke who had received gait training using the Lokomat device with those who had received physical therapy. Motor recovery was measured using the Brunnstrom Recovery Scale (Lower extremity categories) at post-treatment (minimum 30 sessions). Significant between-group differences were found on all lower extremity categories, in favour of exoskeleton gait training vs. physical therapy.

Conclusion: There is limited evidence (Level 2b) from one non-randomized study that exoskeleton gait trainers are more effective than a comparison intervention (physical therapy) for improving motor recovery following stroke.

Spasticity
Not Effective
2A

One fair quality RCT (Kim et al., 2015) and one non-randomized study (Dundar et al., 2014) investigated the effect of exoskeleton gait trainers on spasticity following stroke.

The fair quality RCT (Kim et al., 2015) randomized patients with subacute/chronic stroke to receive gait training using the WALKBOT device or time-matched conventional locomotor training; both groups received additional conventional locomotor training. Spasticity was measured using the Modified Ashworth Scale at post-treatment (4 weeks) and follow-up (8 weeks). No significant between-group difference was found at either timepoint.

The non-randomized retrospective study (Dundar et al., 2014) compared patients with subacute/chronic stroke who had received gait training using the Lokomat device with those who had received physical therapy. Spasticity was measured using the Modified Ashworth Scale at post-treatment (minimum 30 sessions). No significant between-group difference was found.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT and one non-randomized study that exoskeleton gait trainers are not more effective than comparison interventions (conventional locomotor training, physical therapy) for reducing spasticity following stroke.

Stair climbing
Effective
1B

One high quality RCT (Schwartz et al., 2009) investigated the effect of an exoskeleton gait trainer on stair climbing following stroke. The high quality RCT randomized patients with acute/subacute stroke to receive gait training using the Lokomat device or time-matched conventional physical therapy for gait retraining; both groups received additional physical therapy. Stair climbing was measured according to number of stairs climbed at post-treatment (6 weeks). A significant between-group difference was found in a subgroup of patients with high functional ambulation (Functional Ambulation Category score ≥ 3), in favour of Lokomat training vs. conventional physical therapy.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that an exoskeleton gait trainer is more effective than a comparison intervention (physical therapy) for improving stair climbing following stroke, among patients with high functional ambulation.

Stroke severity
Effective
1B

One high quality RCT (Schwartz et al., 2009) investigated the effect of exoskeleton gait trainers on stroke severity following stroke. The high quality RCT randomized patients with acute/subacute stroke to receive gait training using the Lokomat device, or time-matched conventional physical therapy for gait retraining; both groups received additional physical therapy. Stroke severity was measured by the National Institute of Health Stroke Scale at post-treatment (6 weeks). A significant between-group difference was found, in favour of exoskeleton gait training vs. conventional physical therapy.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that an exoskeleton gait trainer is more effective than a comparison intervention (physical therapy) for reducing stroke severity following stroke.

Walking endurance
Not Effective
1B

One high quality RCT (Schwartz et al., 2009) investigated the effect of exoskeleton gait trainers on walking endurance following stroke. The high quality RCT randomized patients with acute/subacute stroke to receive gait training using the Lokomat device or time-matched conventional physical therapy for gait retraining; both groups received additional physical therapy. Walking endurance was measured by the 2 Minute Walk Test at post-treatment (6 weeks). No significant between-group difference was found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that an exoskeleton gait trainer is not more effective than a comparison intervention (physical therapy) for improving walking endurance following stroke.

Walking speed
Not Effective
1B

One high quality RCT (Schwartz et al., 2009) investigated the effect of exoskeleton gait trainers on walking speed following stroke. The high quality RCT randomized patients with acute/subacute stroke to receive gait training using the Lokomat device or time-matched conventional physical therapy for gait retraining; both groups received additional physical therapy. Walking speed was measured by the 10-meter walking test at post-treatment (6 weeks). No significant between-group difference was found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that exoskeleton gait trainers are not more effective than a comparison intervention (physical therapy) for improving walking speed following stroke.

References

Bang, D.H. & Shin, W.S. (2016). Effects of robot-assisted gait training on spatiotemporal gait parameters and balance in patients with chronic stroke: A randomized controlled pilot trial. NeuroRehabilitation, 38(4), 343-9.
DOI: 10.3233/NRE-161325.

Bruni, M.F., Melegari, C., De Cola, M.C., Bramanti, A., Bramanti, P., & Calabro, R.S. (2018). What does best evidence tell us about robotic gait rehabilitation in stroke patients: a systematic review and meta-analysis. Journal of Clinical Neuroscience, 48, 11-17.
DOI: 10.1016/j.jocn.2017.10.048

Chang, W.H., Kim, M.S., Huh, J.P., Lee, P.K.W., & Kim, Y-.H. (2012). Effects of robot-assisted gait training on cardiopulmonary fitness in subacute stroke patients: a randomized controlled study. Neurorehabilitation and Neural Repair, 26(4), 318-24.
DOI: 10.1177/1545968311408916

Cho, D.Y., Park, S.-W., Lee, M.J., Park, D.S., & Kim, E.J. (2015). Effects of robot-assisted gait training on the balance and gait of chronic stroke patients: focus on dependent ambulators. Journal of Physical Therapy Science, 27, 3053-7.
DOI: 10.1589/jpts.27.3053

Chua, J., Culpan, J., & Menon, E. (2016). Efficacy of an electromechanical gait trainer poststroke in Singapore: a randomized controlled trial. Archives of Physical Medicine and Rehabilitation, 97, 683-90.
DOI: 10.1016/j.apmr.2015.12.025

Chung, B.P.H. (2017). Effectiveness of robotic-assisted gait training in stroke rehabilitation: a retrospective matched control study. Hong Kong Physiotherapy Journal, 36, 10-16. DOI: 10.1016/j.hkpj.2016.09.001

Dias, D., Lains, J., Pereira, A., Nunes, R., Caldas, J., Amaral, C., Pires, S., Costa, A., Alves, P., Moreira, M., Garrido, N., & Loureiro, L. (2007). Can we improve gait skills in chronic hemiplegics? A randomised control trial with gait trainer. Europa Medicophysica, 43(4), 499-504.
https://www.ncbi.nlm.nih.gov/pubmed/18084173

dos Santos, M.B., de Oliveira, C.B., dos Santos, A., Pires, C.G., Dylewski, V., & Arida, R.M. (2018). A comparative study of conventional physiotherapy versus robot-assisted gait training associated to physiotherapy in individuals with ataxia after stroke. Behavioural Neurology, 2018, 2892065.
DOI: 10.1155/2018/2892065.

Dundar, U., Toktas, H., Solak, O., Ulasli, A.M., & Eroglu, S. (2014). A comparative study of conventional physiotherapy versus robotic training combined with physiotherapy in patients with stroke. Topics in Stroke Rehabilitation, 21(6), 453-61.
DOI: 10.1310/tsr2106-453.

Freivogel, S., Schmalohr, D., & Mehrholz, J. (2009). Improved walking ability and reduced therapeutic stress with an electromechanical gait device. Journal of Rehabilitation Medicine, 40, 734-9.
DOI: 10.1016/j.apmr.2015.12.025

Geroin, C., Picelli, A., Munari, D., Waldner, A., Tomelleri, C., & Smania, N. (2011). Combined transcranial direct current stimulation and robot-assisted gait training in patients with chronic stroke: a preliminary comparison. Clinical Rehabilitation, 25(6), 537-48.
https://doi.org/10.1177/0269215510389497 .

Han, E.Y., Im, S.H., Kim, B.R., Seo, M.J., & Kim, M.O. (2016). Robot-assisted gait training improves brachial-ankle pulse wave velocity and peak aerobic capacity in subacute stroke patients with totally dependent ambulation: randomized controlled trial. Medicine, 95, 41.
DOI: 10.1097/MD.0000000000005078

Hesse, S., Tomelleri, C., Bardeleben, A., Werner, C., & Waldner, A. (2012). Robot-assisted practice of gait and stair climbing in nonambulatory stroke patients. Journal of Rehabilitation Research and Development, 49(4), 613-22.
DOI: 10.1682/JRRD.2011.08.0142

Hesse, S., Werner, C., Uhlenbrock, D., van Frankenberg, S., Bardeleben, A., & Brandl-Hesse, B. (2001). An electromechanical gait trainer for restoration of gait in hemiparetic stroke patients: preliminary results. Neurorehabilitation and Neural Repair, 15, 39-50.
DOI: 10.1177/154596830101500106

Hidler, J., Nichols, D., Pelliccio, M., Brady, K., Campbell, D.D., Kahn, J.H., & Hornby, T.G. (2009). Multicenter randomized clinical trial evaluating the effectiveness of the Lokomat in subacute stroke. Neurorehabilitation and Neural Repair, 23(1), 5-13.
DOI: 10.1177/1545968308326632

Hornby, T.G., Campbell, D.D., Kahn, J.H., Demott, T., Moore, J.L., & Roth, H.R. (2008). Enhanced gait-related improvements after therapist- versus robotic-assisted locomotor training in subjects with chronic stroke: a randomized controlled study. Stroke, 39, 1786-92.
DOI: 10.1161/STROKEAHA.107.504779.

Husemann, B., Muller, F., Krewer, C., Heller, S., & Koenig, E. (2007). Effects of locomotion training with assistance of a robot-driven gait orthosis in hemiparetic patients after stroke: a randomized controlled pilot study. Stroke, 38, 349-54.
DOI: 10.1161/01.STR.0000254607.48765.cb

Iacovelli, C., Franceschini, M., Cruciani, A., Simbolotti, C., Goffredo, M., Pecchioli, C., Padua, L., Galafate, D., Pournajaf, S., & Aprile, I. (2018) Efficacy of robotic-assisted gait training in sub-acute stroke patients: an Italian bi-centre study. Gait & Posture, 66S (S21-22).
DOI: 10.1016/j.gaitpost.2018.07.133

Kelley, C.P., Childress, J., Boake, C., & Noser, E.A. (2013). Over-ground and robotic-assisted locomotor training in adults with chronic stroke: a blinded randomized clinical trial. Disability and Rehabilitation: Assistive Technology, 8(2), 161-8.
DOI: 10.3109/17483107.2012.714053

Kim, H.Y. & You, J.S.H. (2017). A review of robot-assisted gait training in stroke patients. Brain & Neurorehabilitation, 10(2), e9.
DOI: 10.12786/bn.2017.10.e9

Kim, S.-Y., Yang, L., Park, I.J., Kim, E.J., JoshuaPark, M.S., You, S.H., Kim, Y.-H., Ko, H.-Y., & Shin, Y.-I. (2015). Effects of innovative WALKBOT robotic-assisted locomotor training on balance and gait recovery in hemiparetic stroke: a prospective, randomized, experimenter blinded case control study with a four-week follow-up. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 23(4), 636-42.
DOI: 10.1109/TNSRE.2015.2404936

Lewek, M.D., Cruz, T.H., Moore, J.L., Roth, H.R., Dhaher, Y.Y., & Hornby, G. (2009). Allowing intralimb kinematic variability during locomotor training poststroke improves kinematic consistency: a subgroup analysis from a randomized clinical trial. Physical Therapy, 89(8), 829-39.
DOI: 10.2522/ptj.20080180

Morone, G., Bragoni, M., Iosa, M., De Angelis, D., Venturiero, V., Coiro, P., Pratesi, L., & Paolucci, S. (2011). Who may benefit from robotic-assisted gait training? A randomized clinical trial in patients with subacute stroke. Neurorehabilitation and Neural Repair, 25(7), 636-644.
DOI: 10.1177/1545968311401034

Morone, G., Paolucci, S., Cherubini, A., DeAngelis, D., Venturiero, V., Coiro, P., & Iosa, M. (2017). Robot-assisted gait training for stroke patients: current state of the art and perspectives of robotics. Neuropsychiatric Disease and Treatment, 13, 1303-11.
DOI: 10.2147/NDT.S114102.

Ng, M.F.W., Tong, R.K.Y., & Li, L.S.W. (2008). A pilot study of randomized clinical controlled trial of gait training in subacute stroke patients with partial body-weight support electromechanical gait trainer and functional electrical stimulation: six-month follow-up. Stroke. 39(1), 154-60.
https://www.ncbi.nlm.nih.gov/pubmed/18006861

Park, B.-S., Kim, M.-Y., Lee, L.-K., Yang, S.-M., Lee, W.-D., Noh, J.-W., Shin, Y.-S., Kim, J.-H., Lee, J.-U., Kwak, T.-Y., Lee, T.-H., Kim, J.-Y., & Kim, J. (2015). Effects of conventional overground gait training and a gait trainer with partial body weight support on spatiotemporal gait parameters of patients after stroke. The Journal of Physical Therapy Science, 27, 1603-7.
DOI: 10.1589/jpts.27.1603

Peurala, S., Airaksinen, O, Huuskonen, P. et al. (2009). Effects of Intensive Therapy Using Gait Trainer or Floor Walking Exercises Early After Stroke. Journal of Rehabilitative Medicine, 41, 166-73.
http://www.ncbi.nlm.nih.gov/pubmed/19229450

Peurala, S.H., Tarkka, I.M., Pitkänen, K. (2005). The effectiveness of body weight–supported gait training and floor walking in patients with chronic stroke. Archives of Physical Medicine and Rehabilitation, 86, 1557-64.
http://www.ncbi.nlm.nih.gov/pubmed/16084808

Pohl, M., Werner, C., Holzgraefe, M., et al. (2007). Repetitive locomotor training and physiotherapy improve walking and basic activities of daily living after stroke: a single-blind, randomized multicentre trial. Clinical Rehabilitation, 21, 17-27.
http://www.ncbi.nlm.nih.gov/pubmed/17213237

Polese, J.C., Ada, L., Dean, C.M., Nascimento, L.R., Teixeira-Salmela, L. (2013). Treadmill training is effective for ambulatory adults with stroke: a systematic review. Journal of Physiotherapy, 59, 73-80.
DOI:10.1016/S1836-9553(13)70159-0

Regneaux, J.-P., Saremi, K., Marehbian, J., Bussel, B., & Dobkin, B.H. (2008). An accelerometry-based comparison of 2 robotic assistive devices for treadmill training of gait. Neurorehabilitation and Neural Repair, 22, 348-54.
DOI: 10.1177/1545968307310050

Schwartz, I., Sajin, A., Fisher, I., Neeb, M., Shochina, M., Katz-Leurer, M., & Meiner, Z. (2009). The effectiveness of locomotor therapy using robotic-assisted gait training in subacute stroke patients: a randomized controlled trial. Physical Medicine & Rehabilitation, 1(6), 516-23.
DOI: 10.1016/j.pmrj.2009.03.009

Taveggia, G., Borboni, A., Mule, C., Villafane, J.H., & Negrini, S. (2016). Conficting results of robot-assisted versus usual gait training during postacute rehabilitation of stroke patients: a randomized clinical trial. International Journal of Rehabilitation Research, 39, 29-35.
DOI: 10.1097/MRR.0000000000000137

Tong, R.K., Ng, M.F., & Li, L.S. (2006). Effectiveness of gait training using an electromechanical gait trainer, with and without Functional Electric Stimulation, in subacute stroke: a randomized controlled trial. Archives of Physical Medicine and Rehabilitation, 87, 1298-1304.
DOI:10.1016/j.apmr.2006.06.016

Ukar, D.E., Paker, N., & Bugdayci, D. (2014). Lokomat: a therapeutic chance for patients with chronic hemiplegia. NeuroRehabilitation, 34, 447-53.
DOI: 10.3233/NRE-141054

van Nunen, M.P.M., Gerrits, K.H.L., Konijnenbelt, M., Janssen, T.W.J., & de Haan, A. (2015). Recovery of walking ability using a robotic device in subacute stroke patients: a randomized controlled study. Disability and Rehabilitation: Assistive Technology, 10(2), 141-8.
DOI: 10.3109/17483107.2013.873489

Werner, C., Von Frankenberg, S., Treig, T., Konrad, M., & Hesse, S. (2002). Treadmill training with partial body weight support and an electromechanical gait trainer for restoration of gait in subacute stroke patients: a randomized crossover study. Stroke, 33(12), 2895-2901.
http://www.ncbi.nlm.nih.gov/pubmed/12468788

Westlake, K.P. & Patten, C. (2009). Pilot study of Lokomat versus manual-assisted treadmill training for locomotor recovery post-stroke. Journal of NeuroEngineering and Rehabilitation, 6(18), 1-11.
DOI: 10.1186/1743-0003-6-18

Excluded Studies

Bae, Y.-H., Kim, Y.-H., & Fong, S.S.M. (2016). Comparison of heart rate reserve-guided and ratings of perceived exertion-guided methods for high-intensity robot-assisted gait training in patients with chronic stroke. Topics in Geriatric Rehabilitation, 32(2), 119-26.

Reason for exclusion: Both groups received gait training using an exoskeleton device (Lokomat).

Bonnyaud, C., Zory, R., Boudarham, J., Pradon, D., Bensmail, D., & Roche, N. (2014). Effect of a robotic restraint gait training versus robotic conventional gait training on gait parameters in stroke patients. Experimental Brain Research, 232, 31-42.

Reason for exclusion: Single session; all groups received gait training using an exoskeleton device (Lokomat).

Calabro, R.S., Reitano, S., Leo, A., De Luca, R., Melegari, C., & Bramanti, P. (2014). Can robot-assisted movement training (Lokomat) improve functional recovery and psychological well-being in chronic stroke? Promising findings from a case study. Functional Neurology, 29(2), 139-141.

Reason for exclusion: Case study.

Calabro, R.S., De Cola, M.C., Leo, A., Reitano, S., Balletta, T., Trombetta, G., Naro, A., Russo, M., Berte, F., De Luca, R., & Bramanti, P. (2015). Robotic neurorehabilitation in patietns with cnronic stroke: psychological well-being beyond motor improvement. International Journal of Rehabilitation Research, 38, 219-225.

Reason for exclusion: Non-randomized study, between-group differences not reported.

Conesa, L., Costa, U., Morales, E., Edwards, D.J., Cortes, M., Leon, D., Bernabeu, M., & Medina, J. (2012). An observational report of intensive robotic and manual gait training in sub-acute stroke. Journal of NeuroEngineering and Rehabilitation, 9(1), 13. DOI: 10.1186/1743-0003-9-13.

Reason for exclusion: All participants received Gait Training using the Reha-Stim Gait Trainer followed by conventional overground gait training; between-group comparisons were not conducted.

Delussu, A.S., Morone, G., Iosa, M., Bargoni, M., Traballesi, M., & Paolucci, S. (2014). Physiological responses and energy cost of walking on the Gait Trainer with and without body weight support in subacute stroke patients. Journal of Neuroengineering and Rehabilitation, 11, 54-63.

Reason for exclusion: The intervention group was compared with healthy subjects.

Krewer, C., Muller, F., Husemann, B., Heller, S., Quintern, J., & Koenig, E. (2007). The influence of different Lokomat walking conditions on the energy expenditure of hemiparetic patients and healthy subjects. Gait & Posture, 26, 372-7.

Reason for exclusion: Comparison of patients with healthy subjects using Lokomat under different walking conditions over a single session.

Krewer, C., Rie, K., Bergmann, J., Muller, F., Jahn, K., & Koenig, E. (2013). Immediate effectiveness of single-session therapeutic interventions in pusher behaviour. Gait & Posture, 37(2), 246-50.

Reason for exclusion: Single session; study was specifically examining pusher behaviour.

Mayr, A., Kofler, M., Quirbach, E., Matzak, H., Frohlich, K., & Saltuari, L. (2007). Prospective, blinded, randomized crossover study of gait rehabilitation in stroke patients using the Lokomat gait orthosis. Neurorehabilitation and Neural Repair, 21(4), 307-14.

Reason for exclusion: Between-group differences not reported.

Ochi, M., Wada, F., Saeki, S., & Hachisuka, K. (2015). Gait training in subacute non-ambulatory stroke patients using a full weight-bearing gait-assistance robot: a prospective, randomized, open, blinded-endpoint trial. Journal of the Neurological Sciences, 353, 130-6.

Reason for exclusion: The device used was a robot arm control system.

Picelli, A., Bacciga, M., Melotti, C., La Marchina, E., Verzini, E., Ferrari, F., Pontillo, A., Corradi, J., Tamburin, S., Saltuari, L., Corradini, C., Waldner, A., & Smania, N. (2016). Combined effects of robot-assisted gait training and botulinum toxin type A on spastic equinus foot in patients with chronic stroke: a pilot, single blind, randomized controlled trial. European Journal of Physical and Rehabilitation Medicine, 52(6), 759-66.

Reason for exclusion: Study incorporated use of botulinum toxin type A.

Regnaux, J.P., Saremi, K., Marehbian, J., Bussel, B., & Dobkin, B.H. (2008). An accelerometry-based comparison of 2 robotic assistive devices for treadmill training of gait. Neurorehabilitation and Neural Repair, 22(4), 348-54.

Reason for exclusion: Single case study; comparison of Gait Trainer and Lokomat over single session.

Stoller, O., de Bruin, E.D., Schindelholz, M., Schuster-Amft, C., de Bie, R.A., & Hunt, K.J. (2015). Efficacy of feedback-controlled robotics-assisted treadmill exercise to improve cardiovascular fitness early after stroke: a randomized controlled pilot trial. Journal of Neurologic Physical Therapy, 39, 156-65.

Reason for exclusion: Both groups received gait training using the Lokomat device.

Survey