Skip to main content
PMC home page
BMJ Open logoLink to BMJ Open
. 2023 Sep 14;13(9):e074481. doi: 10.1136/bmjopen-2023-074481

Effect of exoskeleton robot-assisted training on gait function in chronic stroke survivors: a systematic review of randomised controlled trials

Jinchao Yang 1, Yu Gong 1, Lei Yu 1, Laiying Peng 1, Yuanfen Cui 2, Hailong Huang 1,✉
PMCID: PMC10503387  PMID: 37709309

Abstract

Objectives

Numbers of research have reported the usage of robot-assisted gait training for walking restoration post-stroke. However, no consistent conclusion has been reached yet about the efficacy of exoskeleton robot-assisted training (ERAT) on gait function of stroke survivors, especially during the chronic period. We conducted a systematic review to investigate the efficacy of ERAT on gait function for chronic stroke survivors.

Design

This review followed the Participant, Intervention, Comparison and Outcome principle.

Data sources

PubMed, Cochrane Library, Web of Science, Embase and Cumulative Index to Nursing and Allied Health Literature databases were systematically searched until December 2022.

Eligibility criteria

Only randomised controlled trials (RCTs) were included and these RCTs took patients who had a chronic stroke as participants, exoskeleton robot-assisted gait training as intervention, regular rehabilitation therapy as comparison and gait-related functional assessments as outcomes.

Data extraction and synthesis

Data extraction and synthesis used the reporting checklist for systematic review based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. The risk of bias and methodological quality of included studies were evaluated by two independent investigators under the guidance of Cochrane risk of bias.

Results

Out of 278 studies, a total of 10 studies (n=323, mean age 57.6 years, 63.2% males) were identified in this systematic review. According to the Cochrane risk of bias, the quality of these studies was assessed as low risk. Six studies reported favourable effects of ERAT on gait function involving gait performance, balance function and physical endurance, and the ERAT group was significantly superior when compared with the control group. In contrast, the other four trials showed equal or negative effects of ERAT considering different study designs. All the included studies did not claim any serious adverse events.

Conclusion

ERAT could be an efficient intervention to improve gait function for individuals who had a chronic stroke. However, more rigorously designed trials are required to draw more solid evidence.

PROSPERO registration number

CRD42023410796.

Keywords: Exoskeleton, Robot-assisted training, Gait, Chronic stroke, Systematic review

STRENGTHS AND LIMITATIONS OF THIS STUDY

  • This systematic review used the Participant, Intervention, Comparison and Outcome approach and Cochrane risk of bias tool to search articles and evaluate their quality.

  • Only randomised controlled trials (RCTs) are included studies in this review, so they are of relatively high quality.

  • We lack a meta-analysis due to the heterogeneity caused by various outcome measures and small sample size study.

  • Other non-RCTs studying exoskeleton robot-assisted gait training are likely to be missed and the sample size of this review is not large enough because only RCTs are included.

Introduction

Stroke is one of the most common causes of adult disability in industrialised countries, and the cost of stroke-related healthcare is growing rapidly.1 The leading neurological deficit after stroke is hemiplegia, which constantly affects survivors’ activities of daily living, especially the walking ability.2 Jorgensen et al reported that walking restoration mainly occurred within the first 11 weeks, called ‘critical rehabilitation period’ post-stroke, when there was better neuroplasticity in the early stage.3 However, the neuroplasticity phenomena may also emerge during the chronic phase, namely more than 6 months after stroke onset.4 Commonly, walking ability of almost 65–85% of stroke survivors could recover to some extent until the chronic phase, whereas gait disturbances will accompany the whole period of disability, which remains to be the most prominent barrier to walking.5 Although the patients in post-acute stage (mainly the chronic period) spend most of their training time in walking function during the physical therapy sessions, it is not yet enough to regain a better performance.6 Therefore, more attention should be paid to gait training for chronic stroke survivors with gait disorders.

Generally, gait disorders are mainly characterised by poor balance, deficiency of physical endurance and asymmetrical gait pattern with a decreased walking velocity and increased stride width.7 Conventional therapeutic strategies for gait rehabilitation are designed to recover gait patterns, restore balance and enhance endurance, which consist of the bottom-up and patient-oriented approach.8 However, these approaches for patients who had a stroke require a high degree of personalisation, making it particularly difficult for standardisation of clinical trials when considering heterogeneity of functional status and gait patterns among patients who had a stroke.9 Furthermore, these classic approaches applied to gait training involve repetitive and intensive training, and increase therapists’ physical burden during professional work. Thus, a new strategy is urgently needed to compensate for these concerns.

Indeed, robot-assisted gait training (RAGT) is a rehabilitation engineering technology that has shown great progress in the field of rehabilitation over the past two decades.10 As we all know, robotic devices for rehabilitation therapies are particularly based on the so-called motor learning phenomenon,11 12 which induces intensive, repetitive and task-oriented motor activities, and requires effort, attention and active involvement of the subjects.13 RAGT can make it possible for ‘bottom-up’ physiotherapy techniques such as assistance or resistance as needed,14 and it can also be combined with ‘top-down’ rehabilitative interventions, like virtual reality, visual feedback, etc.15 Furthermore, it can relieve the physical burden for therapists, as they no longer need to manually control impaired limbs or assist in torso movements of patients.16 In addition, RAGT can not only help encourage the patients to participate in rehabilitation training, but also provide the therapists with more objective measures of gait function.2 According to the technology and physical interface, rehabilitation robots are classified into exoskeleton and end-effector (EE).17 Exoskeleton robots are typically anchored to particular parts of the human body and simultaneously generate different forces or torques, allowing for interaction with the limb for enough degrees of freedom,17 while EE synchronises the patient’s feet with the correct walking pattern by simulating standing and swinging phases during gait training.18 Although several studies19–22 have discussed the efficacy of robotic-assisted training on gait improvement post-stroke, high-quality clinical data and convincing evidence are very limited.

Recent studies have demonstrated that patients who had a subacute stroke could benefit from exoskeleton robot-assisted training (ERAT) for gait improvement.23–25 However, there are still no clear recommendations for ERAT application in chronic stroke survivors. For more reliable proof, we only included identified randomised controlled trials (RCTs) for further exploration. Through this study, we found that ERAT was beneficial for gait function of most chronic stroke survivors. In addition, we reported no serious adverse events (SAEs) during ERAT application. We hope that this review will provide foundation for potential exoskeleton robot usage in chronic stroke survivors.

Methods

Protocol and registration

This systematic review has been presented according to the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols statement26 and Preferred Reporting Items for Systematic Reviews and Meta-Analyses checklist27 (online supplemental appendix A). A protocol for this systematic review was described in the undermentioned section and previously registered to PROSPERO (registration number CRD42023410796) (online supplemental appendix B).

Supplementary data

bmjopen-2023-074481supp001.pdf (98.8KB, pdf)

Supplementary data

bmjopen-2023-074481supp002.pdf (70.8KB, pdf)

Study design

All included studies in this systematic literature review were RCTs. Eligibility criteria in this review followed the PICO principle, namely the Participant, Intervention, Comparison and Outcome framework.28

Participants

The patients included were adults (more than 18 years old), and diagnosed with chronic stroke (more than 6 months) by brain CT or MRI. Patients were medically stable and cognitively aware to undertake rehabilitation interventions, and able to stand up as well as move forward with assistance.

Intervention

Intervention group had to use exoskeleton robots to assist the patients who had a chronic stroke for gait training including balance function, gait performance and physical endurance. Robotic-assisted exoskeletons for lower limbs were static exoskeletal robots or overground exoskeletons.

Comparison

The patients in control group received overground gait training, treadmill gait training and conventional rehabilitation programmes including Bobath approach, postural control training, manual-assisted walking training, strengthening exercises, etc.

Outcomes

The outcomes of this systematic review were the aforementioned gait function from baseline to the last available follow-up, including evaluation scales (Berg Balance Scale, Functional Gait Assessment (FGA), Rivermead Visual Gait Assessment, Timed Up and Go Test (TUG), 10-Metre Walk Test, 6-Minute Walk Test and Functional Ambulation Categories (FAC)) and gait parameters (gait speed, step length of the affected limb, impaired single-limb stance time, consistency of intralimb movements on the impaired limb and numbers of steps taken during walking). Adverse events associated with adherence to gait training would also be recorded during the intervention.

Search strategy

A systematic review for literature was conducted electronically using five databases: PubMed, Cochrane Library, Web of Science, Embase and Cumulative Index to Nursing and Allied Health Literature, up to December 2022. Details on search terms are available (online supplemental appendix C: Search Strategy). Reference lists of all included literature were reviewed for identifying further eligible studies. Moreover, screened articles from similar systematic reviews or meta-analyses could be manually retrieved in order to avoid omissions.

Study selection

Two investigators (YG and LY) independently searched articles according to our inclusion criteria. The articles written in English were reviewed. We excluded the studies exploring preclinical animals, patients who had not had a chronic stroke and non-RCTs. All identified studies were saved in a citation manager (NoteExpress V.3.7.0.9345). Titles, abstracts and keywords of publications obtained were screened by two authors (YG and LP), respectively. Disagreement from the two investigators was resolved by the third author (JY).

Data extraction

Data were extracted from the included articles and recorded to collect the relevant information. Details of data extraction documented from each reference included authors’ aim, onset of stroke, participants’ information (eg, population, age and gender), intervention characteristics, outcome measures and comments. Data were extracted by one reviewer (JY).

Risk of bias and quality assessment

Regarding the risk of bias, quality assessment of the included studies was conducted via Review Manager (V.5.4). In order to assess bias, two review authors (JY and YC) strictly followed the step of the Cochrane handbook to identify low risk and high risk. If they were unable to answer the question as the included study failed to specify it or it was not clear, this condition would be classified as unclear risk. The third reviewer (HH) would check and solve it when disagreement arose between two reviewers.

Data analysis

Due to a limited number of included studies and significant heterogeneity of the extracted data, it is not appropriate to conduct a meta-analysis. Therefore, we systematically evaluated the outcomes of ERAT in gait functions of chronic survivors.

Patient and public involvement

We did not involve patients or the public in this research.

Results

Study characteristics

The search procedure and strategies were described in figure 1. A total of 10 studies were identified in this systematic review from the five databases after excluding crossover studies (n=5), studies without robotic-assisted exoskeleton (n=4), and studies exploring participants who had an acute or subacute stroke (n=18).

Figure 1.

Figure 1

Open in a new tab

A summary of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram.

Baseline of participants in included studies

Totally, this systematic review included 323 patients who had a chronic stroke with mixed-sex population of 119 women and 204 men. All subjects’ mean age was 57.6 years old ranging from 39 to 85 years old, and all suffered stroke from 6 months to 13.2 years before exoskeleton robotic-assisted intervention. Demographic and baseline characteristics were similar between the two groups (p>0.05) (details in table 1).

Table 1.

Summary of the included studies and the detail of intervention and outcomes

References Aim Duration after stroke Intervention group Control group Outcome measures Comments
Participants Intervention Participants Comparison
Bang and Shin38 To investigate the effects of Lokomat vs TGT on gait parameters and balance in patients who had a chronic stroke 11.6±2.6 months N=9, 5 men and 4 women; age, 53.6±3.9 years Lokomat gait training 20 sessions (1 hour/day, 5 days/week for 4 weeks), 20 hours N=9, 4 men and 5 women; age, 53.7±2.8 years TGT 20 sessions (1 hour/day, 5 days/week for 4 weeks)

  1. Gait speed was significantly higher in Lokomat group than in TGT group (p=0.003)

  2. BBS score was significantly higher than TGT group (p<0.05)

  3. Step length was significantly higher than TGT group (p=0.004)

 Lokomat could be more effective than TGT in improving walking ability and balance in patients who had a chronic stroke
Calabrò et al37 To compare the efficacy of wearable Ekso exoskeleton on gait performance with OGT in patients who had a chronic stroke 10.0±3.0 months N=20, 12 men and 8 women; age, 69.0±4.0 years EGT for 45 min/session, five times/week) for 8 consecutive weeks, 30 hours N=20, 13 men and 7 women; age, 67.0±6.0 years OGT of the same duration as Ekso

  1. 10MWT improved in Ekso group compared with the control group (p<0.001)

  2. Significant improvements in EGT group were recorded in TUG (p<0.05)

 Gait training with Ekso seems to be promising for patients who had a chronic stroke
Erbil et al29 To explore the effects of RoboGait vs physical therapy on balance and gait function of patients who had a chronic stroke after BoNT-A treatment 39.0±34.3 months N=29, 16 men and 13 women; age, 50.1±11.8 years 30 min of RoboGait plus 60 min of physical therapy for 3 weeks, 7.5 hours N=14, 11 men and 3 women; age, 48.7±10.4 years 90 min physical therapy for 3 weeks

  1. Significant improvements were recorded in TUG between the two groups (p<0.001)

  2. Greater improvement was found in BBS after treatment (p<0.001)

  3. RVGA was significantly higher (p<0.001) in RoboGait group

 Robot-assisted training may provide additional benefit compared with physical therapy alone in patients who had a chronic stroke
Hornby et al30 To compare the effects of therapist-assisted LT vs Lokomat-assisted LT in subjects who had a chronic stroke 50.0±5.1 months N=24,15 men and 9 women; age, 57.0±10.0 years 12 sessions (30 min/session) of LT with Lokomat assistance, 6 hours N=24, 15 men and 9 women; age, 57.0±11.0 years 12 sessions (30 min/session) of LT with therapist assistance

  1. Gait speed differences improved greatly during therapist-assisted LT (p<0.05)

  2. Significant difference was observed for impaired single-limb stance (p<0.05)

 Therapist-assisted LT improved more than robotic assistance in gait function for chronic stroke survivors
Jayaraman et al31 To investigate whether gait training with SMA could improve clinical functional outcomes in persons who had a chronic stroke 85.2±74.4 months N=25, 16 men and 9 women; age, 59.5±9.7 years SMA exoskeleton gait training in 18 sessions over 6–8 weeks, 13.5 hours N=25, 17 men and 8 women; age, 61.6±12.6 years Overground walking on varied surfaces

  1. 10MWT improved for the SMA group (p<0.001) than FTST group

  2. SMA group had greater improvement in walking endurance (p=0.033)

  3. More steps were taken in SMA group (p=0.013)

  4. Greater improvement in balance for SMA group (p=0.029)

 SMA can improve walking speed, endurance and balance in patients who had a chronic stroke compared with functional gait training
Kang et al32 To explore the effects of SUBAR-assisted gait training in patients who had a chronic stroke 168.3±67.3 months N=15, 10 men and 5 women; age, 64.3±4.6 years SUBAR-assisted gait training in 10 sessions of 30 min each for 3 weeks, 5 hours N=15, 8 men and 7 women; age, 62.9±6.0 years Conventional physiotherapy of the same duration as SUBAR

  1. SUBAR group improved step length in the affected limb but not significantly (p>0.05)

  2. BBS improved more in the control group (p<0.05)

  3. No significant differences were found in FAC and TUG between groups (p>0.05)

  4. No significant difference was found in 10MWT (p>0.05)

 SUBAR-assisted gait training had no significant difference compared with conventional physiotherapy
Kelley et al33 To compare the effectiveness of Lokomat-assisted system with OGT in adults who had a chronic stroke 44.5 months N=11, 7 men and 4 women; age, 66.9±8.5 years Lokomat gait training (1 hour/day, 5 days/week for 8 weeks), 40 hours N=11, 6 men and 5 women; age, 64.3±10.9 years OGT (1 hour/day, 5 days/week for 8 weeks)

  1. No significant difference was found in 10MWT between groups (p>0.05)

  2. No significant difference was found in 6MWT between the two groups (p>0.05)

 No significant differences were reported in gait speed and endurance between Lokomat and OGT
Lewek et al34 To confirm whether LT with physical assistance was better than Lokomat-assisted LT in improving intralimb coordination for patients who had a chronic stroke 54.0±6.3 months N=10, 4 men and 6 women; age, 52±12 years Lokomat to assist the lower legs in a consistent and symmetrical pattern (30 min/session, 3 sessions per week for 4 weeks), 6 hours N=9, 4 men and 5 women; age, 52.0±12.0 years Manual assistance from the therapist for advancement or pelvic control (30 min/session, 3 sessions per week for 4 weeks)

  1. LT with therapist assistance improved the consistency of intralimb movements of the impaired limb compared with robotic assistance (p<0.05)

  2. No significant change was found in step length (p>0.05)

 Lokomat assistance with LT did not improve intralimb coordination as with therapist assistance
Wright et al35 To investigate the effects of powered bionic leg orthosis on clinical functional outcomes in people who had a chronic stroke 31.0±19.0 months n=16, 14 men and 2 women; age, 59.6±10.1 years 10-week home-based robotic-assisted gait training for a minimum of 30 min/day lasting 10 weeks, 30 hours N=18, 14 men and 4 women; age, 65.1±10.1 years 10-week conventional physiotherapy of the same duration as intervention

  1. Significant differences were observed in SMWT, FAC and BBS (p<0.05)

  2. Number of steps taken during walking increased significantly (p<0.05)

 Exoskeleton robot-assisted gait training induces significant improvements in gait functional outcomes for patients who had a chronic stroke
Yeung et al36 To explore the effects of robot-assisted gait training with AFO dorsiflexion assistance for patients who had a chronic stroke 52.8±30.0 months N=9, 6 men and 3 women; age, 54.2±13.0 years Robotic-assisted AFO with output torque in either dorsiflexion or plantar-flexion for 20 sessions about 5 weeks, 20 hours N=10, 7 men and 3 women; age, 61.2±10.6 years Torque impedance to fix the ankle joint at neutral position, similar to wearing rigid AFO

  1. FAC in post-training and 3-month follow-up showed statistically significant differences (p<0.05)

  2. 10MWT in post/pre-significant differences was found (p<0.05)

 Exoskeleton robot-assisted ankle dorsiflexion gait training can improve gait function in patients who had a chronic stroke
Open in a new tab

AFO, ankle-foot orthosis; BBS, Berg Balance Scale; BoNT-A, botulinum toxin-A; EGT, Ekso gait training; FAC, Functional Ambulation Categories; FTST, functional task-specific training; LT, locomotor training; 10MWT, 10-Metre Walk Test; OGT, overground gait training; RVGA, Rivermead Visual Gait Assessment; SMA, Stride Management Assist; SMWT, 6-Minute Walk Test; TGT, treadmill gait training; TUG, Timed Up and Go.

Risk of bias of included studies

The items in random sequence generation, incomplete outcome data, selective reporting and other risk of bias were regarded as low risk of bias among most of the included studies.29–36 Allocation concealment was evaluated as high risk of bias in only one research.35 There were three trials that the blinding of participants and personnel was assessed as high risk of bias.31 32 35 The performance bias was rated as unclear in five trials.29 30 33 34 37 Two articles30 32 scored high risk of bias in blinding of outcome assessment, while the other two studies were not described clearly so they were evaluated as unclear in the detection bias,29 34 and thus the other six trials were ranked as low risk. Only two trials36 38 were double-blinded RCTs, with a low risk of both performance bias and detection bias. In conclusion, all included studies had a low risk of bias according to the quality assessment using the Cochrane Collaboration tools (figures 2 and 3).

Figure 2.

Figure 2

Open in a new tab

Risk of bias summary.

Figure 3.

Figure 3

Open in a new tab

Risk of bias graph.

Intervention protocols and settings

In this study, the involved exoskeletons included Ekso,37 Lokomat,30 33 34 38 Stride Management Assist (SMA),31 RoboGait29 and SUBAR.32 Two trials35 36 underwent exoskeleton RAGT with bionic orthosis for patients who had a chronic stroke. The duration of robotic rehabilitation interventions ranged from 3 weeks32 to 10 weeks.35 The frequency of interventions was generally kept 30–60 min a day, 5 days a week for at least 3 weeks. A total training time (number of sessions×time per session) ranged from 532 to 40 hours33 (table 1). Gait training interventions should be primarily provided by experienced physical therapists or licensed therapists. The participants in all trials completed the assessment and gait training, without any SAEs in either group, except some mild skin changes.

Qualitative synthesis: outcomes

The included studies aimed to determine whether ERAT was good for rehabilitation of patients who had a chronic stroke. Outcomes related to three major aspects above will be presented in balance function (Berg Balance Scale, TUG), gait performance (10-Metre Walk Test, FAC, FGA, step length of the affected limb, impaired single-limb stance time, consistency of intralimb movements on the impaired limb and Rivermead Visual Gait Assessment) and physical endurance (6-Minute Walk Test and numbers of steps taken during walking) (table 2).

Table 2.

The results of outcome measures in the included studies

Categories Outcome measures Results
Balance function Berg Balance Scale Significant differences between the 2 groups29 31 32 35 38
Timed Up and Go Test Significant differences between the 2 groups29 37
No significant difference between the 2 groups32
Gait performance 10-Metre Walk Test Significant differences between the 2 groups31 36 37
No significant difference between the 2 groups32 33
Gait speed Significant differences between the 2 groups30 38
Functional Ambulation Categories Significant differences between the 2 groups35 36
No significant difference between the 2 groups32
Functional Gait Assessment Significant differences between the 2 groups31
Step length of the affected limb Significant differences between the 2 groups38
Both groups improved but no significant differences32 34
Impaired single-limb stance time Significant differences between the 2 groups30
Consistency of intralimb movements on the impaired limb Significant differences between the 2 groups34
Rivermead Visual Gait Assessment Significant differences between the 2 groups29
Physical endurance 6-Minute Walk Test Significant differences between the 2 groups31 35
No significant difference between the 2 groups33
Number of steps taken during walking Significant differences between the 2 groups31 35
Open in a new tab

Balance function

Berg Balance Scale is the most used evaluation method for balance in post-stroke patients. Five trials assessed the balance function for subjects who had a chronic stroke and showed significant differences between ERAT group and the control group.29 31 32 35 38 Two trials29 37 noted that significant improvements were recorded in TUG after treatment (p<0.05). Nevertheless, one research32 claimed there was no significant difference in TUG between robot-assisted exoskeleton group and the control group (p>0.05) (table 2).

Gait performance

No significant difference was found in 10-Metre Walk Test between ERAT group and the control group.32 33 Nevertheless, three studies31 36 37 showed significant improvements in 10-Metre Walk Test for the intervention group when compared with the control group (p<0.05). Similarly, one research indicated a significant difference in gait speed between the two groups.38 However, another research reported greater improvements in gait speed on the impaired leg when compared with a similar dosage in the robot-assisted group.30 FGA was regarded as a valuable index to evaluate gait improvement, and Jayaraman et al suggested improvement in FGA, which was not seen in the control group.31 Wright et al and Yeung et al demonstrated statistically significant differences in FAC after intervention.35 36 However, no significant difference in FAC was observed between the two groups by Kang et al.32 One study38 stated that the step length of the affected limb (p=0.004) was significantly higher in ERAT group than that in the control group. On the other hand, two trials revealed that both groups improved, but there was no significant difference between them.32 34 Moreover, it was found that subjects who performed therapist-assisted training improved their single stance time of the impaired limb, while ERAT showed little change in their research.30 Regarding consistency of intralimb movements on the impaired limb, significant difference was found between the two groups.34 Lewek et al still noticed that exoskeletal robotic assistance during gait training did not alter intralimb coordination as with therapists’ assistance.34 Erbil et al noted that the Rivermead Visual Gait Assessment was significantly higher (p<0.001) in ERAT group than in the control group29 (table 2).

Physical endurance

Two trials31 35 showed a significant difference in 6-Minute Walk Test between two groups. For example, Jayaraman et al demonstrated that SMA group had a greater improvement in walking endurance (p=0.033) than the control group.31 However, Kelley et al had a contrary result that no significant difference was found in 6-Minute Walk Test between the two groups (p>0.05).33 Wright et al, in their research, indicated that RAGT increased the number of steps taken during walking and was recorded daily compared with the control group (p<0.05).35 Similarly, a significant difference (p=0.013) between the two groups was reported that participants in SMA group took more steps during therapeutic sessions31 (table 2).

Through the aforementioned extracted data analysis, we have found the positive effect of ERAT in six articles,29 31 35–38 which showed significant improvements in gait functional outcomes for chronic stroke survivors when compared with the control group.

Discussion

This systematic review included 10 RCTs for exploring the effects of ERAT on gait function of the patients who had a chronic stroke. All the identified studies were high quality and well designed. Preliminary results indicate that exoskeleton RAGT has no SAEs, and is beneficial in improving gait function.

Consistent with a recent study by Kim and Kim, the ERAT is favourable for gait speed improvement of post-stroke patients.23 However, another study claimed the equal effects in balance enhancement for ERAT compared with the traditional therapy.39 Taken these two studies into comparison, the included disease type both consisted of subacute and chronic stroke,23 39 and no further heterogeneity and publication bias analysis were carried out.23 They seriously constrained the robustness of ERAT in gait function of post-stroke patients. Although ERAT applied for the patients who had a subacute stroke could also improve the gait function,25 when a greater degree of neuroplasticity exists, it remains to be solved whether ERAT will be beneficial for the chronic survivors. Therefore, we not only explored the application of ERAT in chronic survivors, but subdivided gait function into balance function, gait performance and physical endurance, which were foundational elements for locomotion. Considering the strict screening criteria for a reliable conclusion, only RCTs were identified, which caused some methodological limitations for this study. On one hand, RCTs are generally considered as the gold standard for evaluating the effectiveness of healthcare interventions.40 They provide high-quality evidence when compared with the non-RCTs, according to the Oxford Centre for Evidence Based Medicine.41 Moreover, the subjects in RCTs are randomly assigned to different groups to reduce bias of risk and to make meaningful comparisons. In order to provide more solid evidence of ERAT in chronic stroke, we only tried to include and analyse these RCTs. On the other hand, massive heterogeneity was derived from the different intervention designs (eg, robot type, dosage of therapy, time since stroke onset) and gait assessment indexes. Given the heterogeneity among the primary studies and small sample size study, reasonable decisions could not be drawn even with a random-effects model. Therefore, we only systematically review related articles without further meta-analysis.

However, two trials32 33 reported that ERAT brought a similar equivalent result between the two groups. Kang et al explored the effects of total 5 hours of SUBAR-assisted gait training with step length, FAC, TUG and 10-Metre Walk Test.32 All evaluation methods reported no significant advantage for exoskeletons. The same results were shown in Kelley et al’s study which discussed Lokomat in chronic stroke survivors (older than 65 years old).33 Furthermore, another two studies30 34 demonstrated that ERAT did not improve gait function of the impaired single limb as therapists’ assistance in chronic stroke survivors. After investigating these four studies, we found the possible reasons opposite to the final conclusion. First, the intensity of exoskeleton RAGT and participants’ tolerance to training intensity are potential to influence the therapeutic effect of ERAT. In general, more neuroplasticity occurs under dense robotic training.42 This theory has been verified by several studies that high-intensity robotic training brought a better gait performance post-stroke.6 43 As highlighted by a recent study, a total robotic training time of more than 10 hours (number of sessions×time per session) might significantly improve the function of patients post-stroke.44 Second, the time of stroke onset and patients’ age also play a crucial role in ERAT intervention. Age is usually considered as an independent factor in disease recovery. The older patients will restore less than the younger ones when they receive the same intensity of treatment. Also, the later the treatment, the worse the prognosis. Third, considering the trial’s design, a small simple size might fail to obtain a significant difference. A similar study conducted by Lorusso et al45 also strengthened these relevant issues.

Although the robotic-assisted exoskeletons in this study are not identical, they all share the therapeutic properties and design process. Normally, exoskeletons for lower limbs can be classified into static exoskeletal robots and overground exoskeletons. In this review, Ekso,37 SMA,31 SUBAR,32 bionic robotic-assisted leg orthosis35 and powered ankle-assisted ankle-foot orthosis36 were wearable overground exoskeletons. Lokomat30 33 34 38 and RoboGait29 were involved in static exoskeleton-type robots. However, different study designs, various outcome measures and high degree of variability in patients’ severity of stroke have influenced the practical options for exoskeleton rehabilitation.46 For example, Bang and Shin claimed that Lokomat could improve walking ability in patients who had a chronic stroke compared with the control group,38 while Kelley et al achieved a similar effect between the two groups,33 even an opposite result.30 34 They varied from the included participants’ characteristics to exoskeleton intervention parameters. Therefore, more well-designed RCTs are urgently needed to explain the practical role of robotic exoskeletons in chronic stroke.

Conclusion

This systematic review suggests that ERAT is a promising treatment for patients who had a chronic stroke, improving the gait function, balance and endurance with no SAEs. To achieve a better effect, the total training time should be more than 5 hours when ERAT is used for therapeutic intervention. Given the limited data for analysis, more attention should be paid to rehabilitation robots to verify their utility.

Supplementary Material

Reviewer comments
bmjopen-2023-074481.reviewer_comments.pdf (152.4KB, pdf)
Author's manuscript
bmjopen-2023-074481.draft_revisions.pdf (2.7MB, pdf)

Footnotes

Contributors: JY and YG conceived, extracted and designed this review. YG and LY developed the search strategy included in the protocol. JY and YC performed the risk of bias and strictly followed the step of the Cochrane handbook. JY and HH analysed the data. JY and YC drafted the manuscript. HH, JY and LP revised the manuscript critically. All authors read and provided feedback on the draft and approved the final manuscript. The corresponding author HH as guarantor, accepts full responsibility for the finished article, had access to any data and controlled the decision to publish. The corresponding author attests that all listed authors meet the authorship criteria and that no others meeting the criteria have been omitted.

Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

Competing interests: None declared.

Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

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

Supplemental material: This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplemental information.

Ethics statements

Patient consent for publication

Not required.

Ethics approval

This study does not involve human participants. Ethical approval was not required for this systematic review, since all data came from published articles.

References

  • 1.Feigin VL, Brainin M, Norrving B, et al. World stroke Organization (WSO): global stroke fact sheet 2022. Int J Stroke 2022;17:18–29. 10.1177/17474930211065917 [DOI] [PubMed] [Google Scholar]
  • 2.Calabrò RS, Sorrentino G, Cassio A, et al. Robotic-assisted gait rehabilitation following stroke: a systematic review of current guidelines and practical clinical recommendations. Eur J Phys Rehabil Med 2021;57:460–71. 10.23736/S1973-9087.21.06887-8 [DOI] [PubMed] [Google Scholar]
  • 3.Jørgensen HS, Nakayama H, Raaschou HO, et al. Recovery of walking function in stroke patients: the Copenhagen stroke study. Arch Phys Med Rehabil 1995;76:27–32. 10.1016/s0003-9993(95)80038-7 [DOI] [PubMed] [Google Scholar]
  • 4.Bruni MF, Melegari C, De Cola MC, et al. What does best evidence tell us about Robotic gait rehabilitation in stroke patients: A systematic review and meta-analysis. J Clin Neurosci 2018;48:11–7. 10.1016/j.jocn.2017.10.048 [DOI] [PubMed] [Google Scholar]
  • 5.Eng JJ, Tang PF. Gait training strategies to optimize walking ability in people with stroke: a synthesis of the evidence. Expert Rev Neurother 2007;7:1417–36. 10.1586/14737175.7.10.1417 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Holleran CL, Rodriguez KS, Echauz A, et al. Potential contributions of training intensity on locomotor performance in individuals with chronic stroke. J Neurol Phys Ther 2015;39:95–102. 10.1097/NPT.0000000000000077 [DOI] [PubMed] [Google Scholar]
  • 7.Veerbeek JM, van Wegen E, van Peppen R, et al. What is the evidence for physical therapy Poststroke? A systematic review and meta-analysis. PLoS One 2014;9:e87987. 10.1371/journal.pone.0087987 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Pollock A, Baer G, Campbell P, et al. Physical rehabilitation approaches for the recovery of function and mobility following stroke. Cochrane Database Syst Rev 2014:CD001920. 10.1002/14651858.CD001920.pub3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Calafiore D, Negrini F, Tottoli N, et al. Efficacy of Robotic Exoskeleton for gait rehabilitation in patients with subacute stroke: a systematic review. Eur J Phys Rehabil Med 2022;58:1–8. 10.23736/S1973-9087.21.06846-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Calabrò RS, Cacciola A, Bertè F, et al. Robotic gait rehabilitation and substitution devices in neurological disorders: where are we now Neurol Sci 2016;37:503–14. 10.1007/s10072-016-2474-4 [DOI] [PubMed] [Google Scholar]
  • 11.Cramer SC, Riley JD. Neuroplasticity and brain repair after stroke. Curr Opin Neurol 2008;21:76–82. 10.1097/WCO.0b013e3282f36cb6 [DOI] [PubMed] [Google Scholar]
  • 12.Van Peppen RPS, Kwakkel G, Wood-Dauphinee S, et al. The impact of physical therapy on functional outcomes after stroke: what’s the evidence Clin Rehabil 2004;18:833–62. 10.1191/0269215504cr843oa [DOI] [PubMed] [Google Scholar]
  • 13.French B, Thomas L, Leathley M, et al. Does repetitive task training improve functional activity after stroke? A Cochrane systematic review and meta-analysis. J Rehabil Med 2010;42:9–14. 10.2340/16501977-0473 [DOI] [PubMed] [Google Scholar]
  • 14.Morone G, Paolucci S, Cherubini A, et al. Robot-assisted gait training for stroke patients: Current state of the art and perspectives of robotics. Neuropsychiatr Dis Treat 2017;13:1303–11. 10.2147/NDT.S114102 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Moucheboeuf G, Griffier R, Gasq D, et al. Effects of Robotic gait training after stroke: A meta-analysis. Ann Phys Rehabil Med 2020;63:518–34. 10.1016/j.rehab.2020.02.008 [DOI] [PubMed] [Google Scholar]
  • 16.Esquenazi A, Packel A. Robotic-assisted gait training and restoration. Am J Phys Med Rehabil 2012;91(11 Suppl 3):S217–27. 10.1097/PHM.0b013e31826bce18 [DOI] [PubMed] [Google Scholar]
  • 17.Zhang J, Dong Y, Yang C, et al. 5-link model based gait trajectory Adaption control strategies of the gait rehabilitation Exoskeleton for post-stroke patients. Mechatronics 2010;20:368–76. 10.1016/j.mechatronics.2010.02.003 [DOI] [Google Scholar]
  • 18.Aprile I, Iacovelli C, Goffredo M, et al. Efficacy of end-Effector robot-assisted gait training in subacute stroke patients: clinical and gait outcomes from a pilot bi-centre study. NeuroRehabilitation 2019;45:201–12. 10.3233/NRE-192778 [DOI] [PubMed] [Google Scholar]
  • 19.Cordo P, Lutsep H, Cordo L, et al. Assisted movement with enhanced sensation (AMES): coupling motor and sensory to Remediate motor deficits in chronic stroke patients. Neurorehabil Neural Repair 2009;23:67–77. 10.1177/1545968308317437 [DOI] [PubMed] [Google Scholar]
  • 20.Mazzoleni S, Focacci A, Franceschini M, et al. Robot-assisted end-Effector-based gait training in chronic stroke patients: A Multicentric uncontrolled observational retrospective clinical study. NeuroRehabilitation 2017;40:483–92. 10.3233/NRE-161435 [DOI] [PubMed] [Google Scholar]
  • 21.Min JH, Seong HY, Ko SH, et al. Effects of trunk stabilization training robot on postural control and gait in patients with chronic stroke: a randomized controlled trial. Int J Rehabil Res 2020;43:159–66. 10.1097/MRR.0000000000000399 [DOI] [PubMed] [Google Scholar]
  • 22.Sorrento GU, Archambault PS, Fung J. Walking with robot-generated Haptic forces in a virtual environment: a new approach to analyze lower limb coordination. J Neuroeng Rehabil 2021;18:136. 10.1186/s12984-021-00823-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Kim C, Kim HJ. Effect of robot-assisted Wearable Exoskeleton on gait speed of post-stroke patients: A systematic review and meta-analysis of a randomized controlled trials. Phys Ther Rehabil Sci 2022;11:471–7. 10.14474/ptrs.2022.11.4.471 [DOI] [Google Scholar]
  • 24.Ii T, Hirano S, Tanabe S, et al. Robot-assisted gait training using Welwalk in Hemiparetic stroke patients: an effectiveness study with matched control. J Stroke Cerebrovasc Dis 2020;29:105377. 10.1016/j.jstrokecerebrovasdis.2020.105377 [DOI] [PubMed] [Google Scholar]
  • 25.Oh W, Park C, Oh S, et al. Stage 2: who are the best candidates for Robotic gait training rehabilitation in Hemiparetic stroke JCM 2021;10:5715. 10.3390/jcm10235715 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Moher D, Shamseer L, Clarke M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev 2015;4:1. 10.1186/2046-4053-4-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. 10.1136/bmj.n71 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Huang X, Lin J, Demner-Fushman D. Evaluation of PICO as a knowledge representation for clinical questions. AMIA Annu Symp Proc 2006;2006:359–63. [PMC free article] [PubMed] [Google Scholar]
  • 29.Erbil D, Tugba G, Murat TH, et al. Effects of robot-assisted gait training in chronic stroke patients treated by botulinum toxin-a: A pivotal study. Physiother Res Int 2018;23:e1718. 10.1002/pri.1718 [DOI] [PubMed] [Google Scholar]
  • 30.Hornby TG, Campbell DD, Kahn JH, et al. Enhanced gait-related improvements after Therapist- versus Robotic-assisted locomotor training in subjects with chronic stroke: a randomized controlled study. Stroke 2008;39:1786–92. 10.1161/STROKEAHA.107.504779 [DOI] [PubMed] [Google Scholar]
  • 31.Jayaraman A, O’Brien MK, Madhavan S, et al. Stride management assist Exoskeleton vs functional gait training in stroke: A randomized trial. Neurology 2019;92:e263–73. 10.1212/WNL.0000000000006782 [DOI] [PubMed] [Google Scholar]
  • 32.Kang CJ, Chun MH, Lee J, et al. Effects of robot (SUBAR)-Assisted gait training in patients with chronic stroke: randomized controlled trial. Medicine (Baltimore) 2021;100:e27974. 10.1097/MD.0000000000027974 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Kelley CP, Childress J, Boake C, et al. Over-ground and Robotic-assisted locomotor training in adults with chronic stroke: a blinded randomized clinical trial. Disabil Rehabil Assist Technol 2013;8:161–8. 10.3109/17483107.2012.714052 [DOI] [PubMed] [Google Scholar]
  • 34.Lewek MD, Cruz TH, Moore JL, et al. Allowing Intralimb Kinematic variability during locomotor training Poststroke improves Kinematic consistency: a subgroup analysis from a randomized clinical trial. Phys Ther 2009;89:829–39. 10.2522/ptj.20080180 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Wright A, Stone K, Martinelli L, et al. Effect of combined home-based, Overground Robotic-assisted gait training and usual Physiotherapy on clinical functional outcomes in people with chronic stroke: A randomized controlled trial. Clin Rehabil 2021;35:882–93. 10.1177/0269215520984133 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Yeung L-F, Ockenfeld C, Pang M-K, et al. Randomized controlled trial of robot-assisted gait training with Dorsiflexion assistance on chronic stroke patients wearing ankle-foot-Orthosis. J Neuroeng Rehabil 2018;15:51. 10.1186/s12984-018-0394-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Calabrò RS, Naro A, Russo M, et al. Shaping Neuroplasticity by using powered Exoskeletons in patients with stroke: a randomized clinical trial. J Neuroeng Rehabil 2018;15:35. 10.1186/s12984-018-0377-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Bang DH, Shin WS. Effects of robot-assisted gait training on Spatiotemporal gait parameters and balance in patients with chronic stroke: A randomized controlled pilot trial. NeuroRehabilitation 2016;38:343–9. 10.3233/NRE-161325 [DOI] [PubMed] [Google Scholar]
  • 39.Loro A, Borg MB, Battaglia M, et al. Balance rehabilitation through robot-assisted gait training in post-stroke patients: A systematic review and meta-analysis. Brain Sci 2023;13:92. 10.3390/brainsci13010092 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Treweek S, Lockhart P, Pitkethly M, et al. Methods to improve recruitment to randomised controlled trials: Cochrane systematic review and meta-analysis. BMJ Open 2013;3:e002360. 10.1136/bmjopen-2012-002360 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Phillips B, Ball C, Sackett D, et al. Oxford centre for evidence-based medicine-levels of evidence,. 2009Available: http://www.cebmnet/indexaspx?o=1025.2009
  • 42.Wang X, Wong W, Sun R, et al. Differentiated effects of robot hand training with and without neural guidance on Neuroplasticity patterns in chronic stroke. Front Neurol 2018;9:810. 10.3389/fneur.2018.00810 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Lissom LO, Lamberti N, Lavezzi S, et al. Is robot-assisted gait training intensity a determinant of functional recovery early after stroke? A pragmatic observational study of clinical care. International Journal of Rehabilitation Research 2022;45:189–94. 10.1097/MRR.0000000000000518 [DOI] [PubMed] [Google Scholar]
  • 44.Wang L, Zheng Y, Dang Y, et al. Effects of robot-assisted training on balance function in patients with stroke: A systematic review and meta-analysis. J Rehabil Med 2021;53:2773. 10.2340/16501977-2815 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Lorusso M, Tramontano M, Casciello M, et al. Efficacy of Overground Robotic gait training on balance in stroke survivors: a systematic review and meta-analysis. Brain Sci 2022;12:713. 10.3390/brainsci12060713 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Campagnini S, Liuzzi P, Mannini A, et al. Effects of control strategies on gait in robot-assisted post-stroke lower limb rehabilitation: a systematic review. J Neuroeng Rehabil 2022;19:52. 10.1186/s12984-022-01031-5 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary data

bmjopen-2023-074481supp001.pdf (98.8KB, pdf)

Supplementary data

bmjopen-2023-074481supp002.pdf (70.8KB, pdf)

Reviewer comments
bmjopen-2023-074481.reviewer_comments.pdf (152.4KB, pdf)
Author's manuscript
bmjopen-2023-074481.draft_revisions.pdf (2.7MB, pdf)

Data Availability Statement

All data relevant to the study are included in the article or uploaded as supplemental information.

Articles from BMJ Open are provided here courtesy of BMJ Publishing Group

RESOURCES