UBC Faculty Research and Publications

Powered robotic exoskeletons in post-stroke rehabilitation of gait: a scoping review Louie, Dennis R; Eng, Janice J Jun 8, 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12984_2016_Article_162.pdf [ 762.97kB ]
JSON: 52383-1.0361957.json
JSON-LD: 52383-1.0361957-ld.json
RDF/XML (Pretty): 52383-1.0361957-rdf.xml
RDF/JSON: 52383-1.0361957-rdf.json
Turtle: 52383-1.0361957-turtle.txt
N-Triples: 52383-1.0361957-rdf-ntriples.txt
Original Record: 52383-1.0361957-source.json
Full Text

Full Text

REVIEW Open AccessPowered robotic exoskeletons in post-stroke rehabilitation of gait: a scopingreviewDennis R. Louie1,3 and Janice J. Eng2,3*AbstractPowered robotic exoskeletons are a potential intervention for gait rehabilitation in stroke to enable repetitivewalking practice to maximize neural recovery. As this is a relatively new technology for stroke, a scoping review canhelp guide current research and propose recommendations for advancing the research development. The aim ofthis scoping review was to map the current literature surrounding the use of robotic exoskeletons for gaitrehabilitation in adults post-stroke. Five databases (Pubmed, OVID MEDLINE, CINAHL, Embase, Cochrane CentralRegister of Clinical Trials) were searched for articles from inception to October 2015. Reference lists of includedarticles were reviewed to identify additional studies. Articles were included if they utilized a robotic exoskeleton asa gait training intervention for adult stroke survivors and reported walking outcome measures. Of 441 recordsidentified, 11 studies, all published within the last five years, involving 216 participants met the inclusion criteria.The study designs ranged from pre-post clinical studies (n = 7) to controlled trials (n = 4); five of the studies utilizeda robotic exoskeleton device unilaterally, while six used a bilateral design. Participants ranged from sub-acute(<7 weeks) to chronic (>6 months) stroke. Training periods ranged from single-session to 8-week interventions.Main walking outcome measures were gait speed, Timed Up and Go, 6-min Walk Test, and the FunctionalAmbulation Category. Meaningful improvement with exoskeleton-based gait training was more apparent insub-acute stroke compared to chronic stroke. Two of the four controlled trials showed no greater improvement inany walking outcomes compared to a control group in chronic stroke. In conclusion, clinical trials demonstrate thatpowered robotic exoskeletons can be used safely as a gait training intervention for stroke. Preliminary findingssuggest that exoskeletal gait training is equivalent to traditional therapy for chronic stroke patients, while sub-acutepatients may experience added benefit from exoskeletal gait training. Efforts should be invested in designingrigorous, appropriately powered controlled trials before powered exoskeletons can be translated into a clinical toolfor gait rehabilitation post-stroke.Keywords: Stroke, Cerebrovascular accident, Robotic exoskeleton, Gait rehabilitation, Scoping reviewBackgroundStroke is a leading cause of acquired disability in theworld, with increasing survival rates as medical care andtreatment techniques improve [1]. This equates to an in-creasing population with stroke-related disability [1, 2],who experience limitations in communication, activitiesof daily living, and mobility [3]. A majority of thispopulation ranks recovering the ability to walk or im-proving walking ability among their top rehabilitationgoals [4, 5]; furthermore, the ability to walk is a deter-mining factor as to whether an individual is able to re-turn home after their stroke [6]. However, 30 – 40 % ofstroke survivors have limited or no walking ability evenafter rehabilitation [7, 8] and so there is an ongoing needto advance the efficacy of gait rehabilitation for strokesurvivors.Powered robotic exoskeletons are a recently developedtechnology that allows individuals with lower extremityweakness to walk [9]. These wearable robots strap to the* Correspondence: janice.eng@ubc.ca2Department of Physical Therapy, University of British Columbia, 212-2177Wesbrook Mall, Vancouver, BC V6T 1Z3, Canada3Rehabilitation Research Program, Vancouver Coastal Health ResearchInstitute, Vancouver, BC, CanadaFull list of author information is available at the end of the article© 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Louie and Eng Journal of NeuroEngineering and Rehabilitation  (2016) 13:53 DOI 10.1186/s12984-016-0162-5legs and have electrically actuated motors that controljoint motion to automate overground walking. Poweredexoskeletons were originally designed to be used as anassistive device to allow individuals with complete spinalcord injury to walk [10]. However, because they allowfor walking without overhead body weight support or atreadmill, they have gained attention as an alternateintervention for gait rehabilitation in other populationssuch as stroke where repetitive gait training has beenshown to yield improvements in walking function [11,12]. Several powered exoskeletons are already commer-cially available, such as the Ekso (Ekso Bionics, USA),Rewalk (Rewalk Robotics, Israel), and Indego (ParkerHannifin, USA) exoskeletons, with more beingdeveloped.There have been many forms of gait retraining pro-posed for stroke survivors. Conventional physical ther-apy gait rehabilitation leads to improvements in speedand endurance [13], particularly when conducted earlypost-stroke [14]. However, conventional gait retrainingusing hands-on assistance can be taxing on therapists;the number of steps actually taken in a session reflectsthis and has been shown to be low in sub-acute hospitalrehabilitation [15]. Many of the proposed technology-based gait intervention strategies have focused on redu-cing the physical strain to therapists while increasing theamount of walking repetition that individuals undergo.For example, body weight-supported treadmill training(BWSTT) allows therapists to manually move the hemi-paretic limb in a cyclical motion while the patient’strunk and weight are partially supported by an overheadharness system; this has shown improvements in strokesurvivors’ gait speed and endurance compared to con-ventional gait training [16], yet still places a high phys-ical demand on therapists. Advances in technology haveled to treadmill-based robotics, such as the Lokomat(Hocoma, Switzerland), LOPES (University of Twente,Netherlands), and G-EO (Reha-Technology,Switzerland), which have bracing that attaches to the pa-tient’s legs to take them through a walking motion onthe treadmill. The appeal of this technology is that it canprovide substantially higher repetitions for walking prac-tice than BWSTT without placing strain on therapists;however, there is conflicting evidence regarding the effi-cacy of treadmill-based robotics for gait training com-pared to conventional therapy or BWSTT. Some studieshave shown that treadmill robotics improve walking in-dependence in stroke [17, 18] but do not improve speedor endurance [18, 19]. There has been some sentimentthat such technology has not lived up to the expecta-tions originally predicted based on theory and practice[20]. One argument is that these treadmill robotics witha pre-set belt speed, combined with body weight sup-port, create an environment where the patient has lesscontrol over the initiation of each step [21]; another ar-gument against treadmill-based gait training is the lackof variability in visuospatial flow, which is an essentialchallenge of overground walking [20]. Powered roboticexoskeletons, though similar in structure to treadmill-based robotics, differ in that they require active partici-pation from the user for both swing initiation and footplacement; for example, some exoskeletons have controlstrategies which will only assist the stepping motionwhen it detects adequate lateral weight-shifting [9]. Fur-thermore, because the powered exoskeletons are usedfor overground walking, it requires the user to be re-sponsible for maintaining trunk and balance control, aswell as navigating their path over varying surfaces.While these powered exoskeletons hold promise, theliterature surrounding their use for gait training is onlyjust beginning to gather, with the majority focusing onspinal cord injury [22–24]. Several [25–27] systematicreviews have shown safe usage, positive effects as an as-sistive device, and exercise benefits for individuals withspinal cord injury. Only one systematic review [28] spe-cifically focusing on powered exoskeletons has includedstudies involving stroke participants, though studies inspinal cord injury and other conditions were also in-cluded. This review focused exclusively on the HybridAssistive Limb (HAL) exoskeleton (Cyberdyne, Japan),(which currently is not approved for clinical use outsideof Japan), and found beneficial effects on gait functionand walking independence; however, the results werecombined generally across all included patient popula-tions and not specifically for stroke.Given that this is a relatively new intervention forstroke, the objective of this scoping review was to mapthe current literature surrounding the use of poweredrobotic exoskeletons for gait rehabilitation in post-strokeindividuals and to identify gaps in the research. The sec-ond objective of this scoping review was to preliminarilyexplore the efficacy of exoskeleton-based gait rehabilita-tion in stroke. As this is a relatively new technology forstroke, a scoping review can help guide current researchand propose recommendations for advancing thetechnology.ReviewMethodsThis scoping review was conducted in accordance withthe framework proposed by Arksey and O’Malley [29],and guided by the refined process highlighted by Levacet al. [30].OVID MEDLINE, Embase, Cochrane Central Registerof Controlled Trials, PubMed, and CINAHL databaseswere accessed and searched from inception on October14, 2015. We combined the search terms (robot* ORexoskeleton OR “powered gait orthosis” OR PGO ORLouie and Eng Journal of NeuroEngineering and Rehabilitation  (2016) 13:53 Page 2 of 10HAL OR “hybrid assistive limb” OR ReWalk OR EksoOR Indego) AND (stroke OR CVA OR “cerebrovascularaccident” OR “cerebral infarct” OR “cerebralhemorrhage” OR hemiplegia OR hemiparesis OR ABIOR “acquired brain injury”) AND (gait OR walk ORwalking OR ambulation), with humans and English lan-guage as limits.Inclusion criteria were full-text, peer-reviewed articlesthat used a powered robotic exoskeleton with adultspost-stroke as an intervention for gait rehabilitation. Ar-ticles were included if they reported functional walkingoutcomes (e.g., speed, distance, independence). We de-fined a powered robotic exoskeleton as a wearable ro-botic device which actuates movement of at least onejoint while walking, either unilaterally or bilaterally. Wefurther defined powered robotic exoskeletons as stand-alone devices that can be used for overground walking,with programmable control. Articles were excluded ifthey: reported only technology development; reportedonly electromyography, physiological cost, or joint kine-matic data; combined other interventions (e.g., func-tional electrical stimulation); included healthyparticipants or children; utilized a treadmill-based device(i.e., the exoskeleton and treadmill are a single device,where the exoskeleton cannot be used separately over-ground); included mixed diagnosis participants (<50 %stroke); or if only an abstract was available.Titles and abstracts were screened for relevance bytwo reviewers (DRL, CC) according to the inclusion andexclusion criteria above. In the event of conflict, a thirdreviewer (JJE) was consulted for resolution. Full-textswere then screened and reference lists of all selected ar-ticles were searched for additional studies. Included arti-cles were then examined to extract data regarding studydesign, exoskeleton device, participant characteristics,intervention, training period, outcome measures, adverseeffects, and results. We examined the changes in func-tional walking outcomes relative to clinically meaningfulchange values published in the literature (Table 1).ResultsAs seen in Fig. 1, our electronic database searchreturned 440 unique titles. Only one additional articlewas identified through reference list searching. Afterscreening titles, abstracts, and full-texts for eligibility, 11articles were included [31–41]. All 11 articles were pub-lished in the last five years, with seven [31–33, 35–37,39] published in the last two years. Five studies wereconducted in the United States, five in Japan, and one inSweden.Study designOf the included studies, three were randomized con-trolled trials (RCTs) [31, 35, 36], and one was a non-randomized controlled study [37]. The rest were a var-iety of single-group pre-post clinical trials as seen inTable 2. Of the three RCTs, two were smaller in size (n= 24 and n = 22) and considered pilot studies [31, 36].ParticipantsAcross the 11 studies, there was a total of 216(male/female:136/80) participants with stroke en-rolled (Table 2), with variability in the inclusion cri-teria for participation. Seven studies [35–41]included participants with chronic stroke (at least sixmonths post-stroke). Four studies [31–35] investi-gated the exoskeleton with sub-acute participants(less than six months post-stroke) during inpatientrehabilitation. The majority of participants were inthe 50 – 70 age range. Six studies [35–37, 39–41]specifically enrolled participants with the ability towalk without physical assistance from a therapist,permitting walking devices such as a cane or walker,while three studies [31, 32, 34] specified a require-ment of needing manual physical assistance to walk.The former studies aimed to improve mobility forambulatory individuals with chronic stroke, whereasthe latter sought to restore independent ambulationfor sub-acute stroke participants. The other twostudies [33, 38] enrolled participants with a mix offunctional levels.ExoskeletonsThe included studies investigated a variety of exoskele-tons, each having different set-ups and control mecha-nisms. Five studies [31, 36, 37, 40, 41] used a roboticexoskeleton unilaterally on the affected leg, while an-other five studies [32, 34, 35, 38, 39] used a bilateral set-Table 1 Meaningful change values for functional walking outcomes in strokeOutcome measure Sub-acute stroke Chronic strokeTUG Not available MDC = 2.9 s [44]6MWT MDC = 61 m [43] MCID = 34.4 m [42]10MWT/gait speed MCID = 0.16 m/s [45] MCID = 0.06 m/s (small) [43]MCID = 0.14 m/s (substantial) [43]FAC Not available Not available6MWT six-minute walk test, 10MWT ten meter walk test, FAC functional ambulation category, MCID minimal clinically important difference, MDC minimal detect-able change, TUG timed up and goLouie and Eng Journal of NeuroEngineering and Rehabilitation  (2016) 13:53 Page 3 of 10up for gait training. One study [33] progressed partici-pants, as they were able, from a bilateral design to a uni-lateral configuration. The most studied exoskeleton wasthe HAL, used in six studies [31–34, 37, 38]; in thesestudies, participants’ hip and knee joints were electricallyactuated in a walking motion. In one study [39] the H2exoskeleton (Technaid SL, Spain), assisted the hip, knee,and ankle joints. Four studies [35, 36, 40, 41] utilized anexoskeleton powering only one joint of the lower ex-tremity (either hip or knee, uni- or bilaterally); no stud-ies were found in which only the ankle was actuatedduring gait. Control of the exoskeletons ranged fromremote-control button activation [39] to active move-ment control of stepping; the devices are able to detectmovement intention through monitoring joint anglesand limb torque [35, 36, 40, 41], or through bio-electricsignalling of muscle activity [31–34, 37, 38]. All exoskel-etons except the HAL provided supplementary gait as-sistance on an as-needed basis, in which the usergenerates as much of the walking movements as possibleand the device provides extra torque or support to en-sure step completion. The HAL has two modes, one thatprovides complete stepping assistance and one thatadapts to user force generation. Table 3 further detailsthe exoskeletons, their control strategies, and the levelof assistance provided.Training periodThere was variability in the training period of the in-cluded studies, ranging from a single session [34] to sev-eral weeks [31–33, 39, 40] or months [35–38, 41] oftraining. Training duration lasted from 20 – 90 min persession, and frequency ranged from two to five sessionsper week. Table 2 details the different training periodsfor each study.Training protocolThe training protocol employed in each study differed,and varied depending on the study design, length of thetraining period, and exoskeleton used (Table 2). Gener-ally, subjects were progressed as tolerated from weight-bearing functional tasks (sit-to-stand, standing balance,weight shifting) to walking practice while wearing theexoskeleton device. Two studies [32, 33] had participantstrain on a treadmill, which allowed therapists to adjustthe walking speed externally. The most detailed trainingprotocols were described in the controlled trials [31, 35–37], wherein individuals were progressed according tovarious intensity guidelines such as rate of perceived ex-ertion (RPE) [35] and non-exoskeletal walking speed[37]. For example, Yoshimoto et al. [37] advanced thetraining speed to 1.5-1.7 times the maximal non-exoskeletal 10MWT walking speed before each session.Fig. 1 Study results: A flowchart of selection process based on inclusion/exclusion criteriaLouie and Eng Journal of NeuroEngineering and Rehabilitation  (2016) 13:53 Page 4 of 10Table 2 Summary of studies included in the reviewStudy &DesignParticipants Exoskeleton& TrainingPeriodTraining Protocol Walking outcomes & ResultsSubacute StrokeWatanabeet al. (2014)[31]UnblindedRCTSub-acute stroke1 – 2 person assistambulation (HAL groupn = 11, mean 58.9 days post-strokeConventional groupn = 11, mean 50.6 days post-stroke)HAL –Unilateral12 sessionsover4 weeks20 minutesessionsHAL group – gait training while wearingHAL, facilitating improvements in walkingability, partial BWS if needed; progress asable from complete assistance by device toassist-as-needed through bioelectric signaldetectionConventional group – facilitateimprovements in walking ability,customized to functional level; speed andduration of walking gradually increased1) TUG – No significant difference inimprovement between groups2) 6MWT – No significant difference inimprovement between groups3) Gait speed – No significant difference inimprovement between groups4) FAC – HAL group improved significantly(p = 0.04) more than Conventional group(change of +1.1 for HAL group; change of+0.6 for Conventional group)Nilsson et al.(2014) [32]Pre-poststudySub-acute stroke1 – 2 person assistambulation(n = 8, 6 – 46 days post-stroke)HAL –Bilateral5 sessions/week,median 17sessions25 minutestrainingProgression from weight shift control tobioelectric signalling control, training withBWS on treadmill; progression of speedand BWS as tolerated1) 10MWT – median change of +0.24 m/s,4 previously non-ambulatory progressed toambulatory2) FAC – median change of +1.5 (from 0 to1.5)Fukudaet al. (2015)[33]Pre-poststudySub-acute stroke (n = 53, 12non-ambulatory, 41ambulatory)HAL – Uni/bilateral2 sessions/week, mean3.9 sessionsWalking on treadmill in exoskeleton,progress from complete control tobioelectric signalling1) 10MWT – change of +0.1 m/s forBrunnstrom stage III (greater severity withlower stage) (n = 12); no change forBrunnstrom stage IV (n = 7); change of+0.1 m/s for Brunnstrom stage V (n = 12);change of +0.4 m/s for Brunnstrom stageVI (N = 10)Maeshimaet al. (2011)[34]Pre-poststudySub-acute stroke1 – 2 person assistambulation (n = 16, 27 –116 days post-stroke)HAL –BilateralSinglesessionWalking and stair practice after standingpractice in exoskeleton1) 10MWT – positive change for 14 of 16patients (values not provided)Chronic StrokeBuesinget al. (2015)[35]Single-blindRCTChronic stroke Limitedcommunity ambulation (SMAgroup – n = 25, mean7.1 years post-strokeFunctional task specifictraining group – n = 25,mean 5.4 years post-stroke)SMA –Bilateral18 sessionsover 6 –8 weeks45 minutesessionsSMA group – 30 minutes of high intensityoverground walking with SMA (12-16 RPEor 75 % HR max) and 15 minutes ofdynamic functional gait training with SMA(varied surfaces, multi-directional stepping,stair climbing, obstacles, communitymobility)Functional task specific training group –15 minutes of high intensity overgroundwalking training and 30 minutes offunctional goal-based mobility training1) Gait speed – No significant difference inimprovement between groupsStein et al.(2014) [36]Single-blindRCTChronic strokeIndependent ambulation(AlterG group n = 12, mean49.1 months post-strokeExercise group n = 12, mean88.5 months post-stroke)AlterG –Unilateral18 sessionsover6 weeks60 minutesessionsAlterG group – standardized overgroundfunctional tasks including transfers,stepping, turning, reaching, gait training,stairs and curbs while wearing exoskeletonExercise group – group exercises includingrelaxation, meditation, self-stretching, activerange of motion of upper and lower limbs,minimal gait training (5 min/session)1) TUG – No significant difference betweengroups2) 6MWT – No significant difference inimprovements between groups3) 10MWT – No significant difference inimprovement between groupsYoshimotoet al. (2015)[37]Non-randomizedcontrolledtrialChronic strokeIndependent ambulation(HAL group n = 9, mean92.4 months post-strokeConventional PT group n = 9,mean 80.5 months post-stroke)HAL –Unilateral8 sessionsover8 weeks60 minutesessionsHAL group – 20 minutes of HAL walkingper session, with some BWS, walking atspeed 1.5-1.7 times max walking speedwithout deviceConventional PT group – exercise toimprove walking ability including static anddynamic postural tasks, range of motion,and 20 minutes of overground walkingtraining1) TUG – HAL group improved significantlycompared to Conventional PT group(change of -11.5 s for HAL group; changeof +0.1 s for Conventional PT group)2) 10MWT – HAL group improvedsignificantly compared to Conventional PTgroup (change of +0.21 m/s for HALgroup; change of -0.02 m/s forConventional PT group)1) TUG – mean change of -1.1 sLouie and Eng Journal of NeuroEngineering and Rehabilitation  (2016) 13:53 Page 5 of 10Several studies [31, 32, 37, 38] allowed some bodyweight support using an overhead harness to improvewalking mechanics.Walking measuresTen of the 11 studies included a measure of gaitspeed in their assessment of walking ability, eithermeasuring it directly or via the 10-m Walk Test(10MWT). Five studies [31, 36, 39–41] assessed walk-ing endurance by means of a 6-min Walk Test(6MWT), and seven studies [31, 36–41] assessed theTimed Up and Go (TUG) test, which is a measure offunctional mobility as it includes sit-to-stand andturning. Two studies [31, 32] also included level ofindependence or assistance in their assessment ofwalking ability, using the Functional Ambulation Cat-egory (FAC). Participants were not wearing an exo-skeleton device when assessed for the above measuresin all studies, but gait aids such as canes and walkerswere permitted.Table 2 Summary of studies included in the review (Continued)Kawamotoet al. (2013)[38]Pre-poststudyChronic stroke (n = 16, 1 –11 years post-stroke, 8dependent ambulatory, 8 in-dependent ambulatory)HAL –Bilateral16 sessionsover8 weeks20 –30 minutestrainingOverground walking with overhead harnessfor safety and partial BWS; gradualprogression from sit-to-stand to walking(gradually increased intensity by changingspeed, duration, BWS, and HAL controlmechanism)2) 10MWT – mean change of +0.04 m/sBortole et al.(2015) [39]Pre-poststudyChronic stroke Independentambulation(n = 3; 60, 6, 11 months post-stroke)H2 –Bilateral12 sessionsover4 weeks30 minutesessionsOverground walking over a linear trackParticipants in charge of speed andencouraged to walk as much as possible,with breaks1) TUG – change of +1.7 s, -2.5 s,-2.5 s2) 6MWT – change of -115 m, +16 m,+103 mByl et al.(2012) [40]Pre-poststudyChronic stroke Independentambulation(n = 3; 6, 1.3, 10 years post-stroke)AlterG –Unilateral2 – 4sessions/week over4 weeks90 minutesessionsWalking practice, with sit-to-stand transfers,squatting, and stepping activities; obstacleclearance, uneven terrain, community am-bulation, stair climbing1) TUG – change of -6.9 s, +1.9 s, -0.2 s2) 6MWT – change of +37 m, +47 m,+29 m3) 10MWT – change of +0.21 m/s,+0.14 m/s, +0.20 m/sWong et al.(2011) [41]Pre-poststudyChronic strokeIndependent ambulation(n = 3; 37, 26, 40 monthspost-stroke)AlterG –Unilateral18 sessionsover6 weeks60 minutesessions45 minutes while wearing device,standardized weight-bearing functionalmobility activities, sit-to-stand transfers, bal-ance exercises, gait practice at variousspeeds on different surfaces, functional taskpractice1) TUG – change of-11.7 s, -2.3 s, -4.2 s2) 6MWT – change of +17 m, +14 m,+15 m3) 10MWT – change of -0.01 m/s, +0.05 m/s, +0.13 m/s6MWT six-minute walk test, 10MWT ten meter walk test, BWS body weight support, FAC functional ambulation category, H2 H2 exoskeleton, HAL hybrid assistivelimb, HR heart rate, SMA stride management assist system, PT physical therapy, RCT randomized controlled trial, RPE rate of perceived exertion, TUG timed upand goBold indicates value surpasses established meaningful change score detailed in Table 1Table 3 Details of powered exoskeletons in this reviewExoskeleton Joints actuated Stepping initiation Stepping assistanceH2 [39] Hip, knee,ankleInitiated by hand buttons on walkerPre-set speedAssist-as-needed for swingSMA [35] Hip Initiated by movementInternal sensors detect hip joint angle to regulate walkingAssist-as-needed for swingHAL [31–34, 37,38]Hip, knee Initiated by movement (2 modes)Internal sensors detect lateral weight shiftSurface electrodes detect muscle activation via bioelectric signalsFull-assistance for swingAssist-as-needed for swingAlterG [36, 40, 41] Knee Initiated by movementInternal sensors detect movement intention via variable forcethresholdAssist-as-needed for stance, freeswingAlterG AlterG Bionic Leg, formerly Tibion Bionic Leg; H2 H2 exoskeleton; HAL Hybrid Assistive Limb; SMA Stride Management Assist system (Honda R&DCorporation, Japan)Louie and Eng Journal of NeuroEngineering and Rehabilitation  (2016) 13:53 Page 6 of 10Effectiveness of exoskeleton-based gait trainingTen studies reported varying degrees of improved walk-ing ability after exoskeleton training (Table 2). Of thefour sub-acute stroke studies, only one [31] was a ran-domized controlled trial (n = 22) which showed that par-ticipants using the HAL experienced a significantimprovement in FAC scores compared to conventionalgait rehabilitation matched for training time, no longerrequiring manual assistance to walk after the trainingperiod (medium effect size). However, they found no sig-nificant difference between the HAL intervention andconventional therapy for walking speed or endurance.One small pre-post sub-acute study [32] (n = 8) alsofound an improvement in the median FAC score of theirsub-acute participants from 0 (2-person assist to walk)to 1.5 (1-person assist to walk) after exoskeleton-basedgait training. Participants in the two other pre-post stud-ies [33, 34] in sub-acute stroke demonstrated improve-ments in walking speed with only a few sessions, thoughnot all of their participants demonstrated a changegreater than the established minimal clinically importantdifference (MCID) (Table 1).Across the seven chronic stroke studies, improvementsin walking ability were less apparent. In an RCT with 50participants [35], there was no significant difference be-tween the clinically meaningful improvements in gaitspeed made by participants in either the exoskeleton orfunctional training group matched for training time.Similarly, participants using the AlterG Bionic Leg(AlterG, USA) did not demonstrate significant improve-ments compared to the control group or to baselineafter 18 training sessions in a small RCT with 24 partici-pants [36]. In contrast, a nonrandomized controlled trial[37] found significant and clinically meaningful improve-ments in gait speed and TUG time after training using aHAL compared to conventional physical therapy; how-ever, the control group did not receive the same numberof exercise sessions. One larger pre-post study [38] (n =16) did not find changes in gait speed that were beyondthe established MCID (Table 1) while three small pre-post studies [39–41], each with three participants, foundvarying results. Clinical improvements in endurancewere made by four participants in two of the pre-poststudies [39, 40], using a minimal clinically important dif-ference of 34.4 m in the 6MWT. [42] Three participantsacross the three smaller pre-post studies [39–41] mademeaningful improvements in TUG scores. Four partici-pants in two of the pre-post studies [40, 41] demon-strated a clinically meaningful improvement in walkingspeed, using an MCID of 0.06-0.14 m/s [43].Adverse effectsEight studies confirmed that no adverse events occurredduring the course of the gait training intervention. Onestudy [32] reported minor and temporary adverse effectssuch as skin irritation and pain from cuffs and bioelec-tric detection electrodes. Two studies [33, 34] did notreport on adverse events. No studies reported adverse ef-fects on the therapists.DiscussionThis scoping review was conducted to map the literaturesurrounding the use of powered robotic exoskeletons forgait retraining for individuals after stroke and to identifypreliminary findings and areas where further research isrequired. This is a relatively new application of poweredexoskeletons, as they have only recently become avail-able for clinical use. As expected, there are only a smallnumber of studies published relevant to this topic.There were four different powered exoskeletons uti-lized amongst the included studies, ranging from unilat-eral, single joint devices to bilateral, multi-joint roboticswith the capacity to detect volitional bioelectrical signalsto initiate powered movement. Other exoskeletons existon the commercial market for clinical application thathave not yet been investigated for stroke such as theEkso, Rewalk, and Indego (Parker Hannifin Corporation,USA). Research with these other exoskeletons is re-quired to determine their clinical usefulness and wouldalso strengthen the literature in general support of exo-skeleton use for gait rehabilitation in stroke patients.Studies comparing unilateral to bilateral designs mayalso be another avenue for investigating the efficacy ofexoskeletal gait retraining.The majority of the included studies investigatedexoskeleton-based gait training in chronic stroke partici-pants. However, the greatest amount of functional andneurological recovery after stroke occurs in the first sixweeks after stroke [3, 7]. In reflection of this, all four studiesin the sub-acute phase of stroke reported positive effects ofexoskeleton training. Two studies [31, 32] demonstratedimproved walking independence with repeated exoskeletalgait training for more limited stroke participants, which isin line with findings using treadmill-based robotics [17]. Inanother study [33], there was significant improvement inwalking speed (0.4 m/s) for stroke participants who hadsome voluntary motor control, but much less change(0.1 m/s) for those without voluntary control. The magni-tude and parameter (ability, speed) of walking improvementmay vary depending on the initial functional presentationof the exoskeleton user; furthermore, the spontaneous re-covery following stroke is a confounding factor for the im-provements reported that has yet to be rigorouslycontrolled for in the current literature.Study findings were not consistent for chronic strokeparticipants. All chronic stroke participants includedwere ambulatory, and so studies investigated changes ingait parameters rather than functional ability. WhileLouie and Eng Journal of NeuroEngineering and Rehabilitation  (2016) 13:53 Page 7 of 10there were modest, but not consistent changes in the pre-post studies, the more rigorous RCTs [35, 36] did not showa difference from their respective control groups whengroups were matched for exercise time and interaction witha physical therapist. Even in studies with longer trainingprotocols [35, 36, 38, 41], there was not a trend for greaterimprovements. Despite receiving the repetitious practicethat is required for motor learning [11, 12], chronic strokeparticipants do not respond as positively to exoskeletal gaittraining as sub-acute patients. This is consistent with find-ings in a systematic review [18] of treadmill-based exoskel-eton devices for gait training in chronic, ambulatoryindividuals with stroke. A possible explanation for this isthat once an individual is able to walk, they benefit morefrom unconstrained walking practice with greater variabilityand unpredictable challenges [14]. While powered exoskel-etons do not require the participant to use a treadmill, theystill constrain the user to a stereotyped movement patternand may thus under-challenge them.The majority of included studies had small sample sizes,which may have limited the power of their study findingsand analysis. In addition to this, the majority of these stud-ies were pilot feasibility or pre-post clinical studies; recruit-ment and lack of a control group may have introduced biasto their findings. For example, one study [37] used a non-randomized controlled design, where the control group wasformed of participants who were less able to attend thestudy training protocol. These results inform the prelimin-ary evidence in the field and more rigorous, appropriatelypowered randomized controlled trials will continue to ad-vance the clinical application of powered exoskeletons.Future directions for research and suggestions for clinicalpracticeFrom our data synthesis we have identified various con-siderations when using an exoskeleton for gait retrainingand propose several questions for future research:1. Do non-ambulatory chronic stroke participants ex-perience the same improvement in walking ability assub-acute stroke participants when using an exoskel-eton device for gait retraining?2. How does initial functional presentation impact thenature of improvement in walking ability when usingan exoskeleton device for gait rehabilitation?3. What is the impact of different exoskeletons(number of joints actuated, level of assistance andcontrol of stepping) on gait rehabilitation in stroke?4. What is the impact of using a bilateral designcompared to a unilateral design for gaitrehabilitation in hemiparetic stroke?5. What is the optimal dose of exoskeletal gait trainingfor stroke patients to regain the most walkingability?6. How does overground exoskeletal gait trainingcompare to body weight-supported treadmilltraining?7. Can exoskeletons be used to safely ambulate 2-person assist patients early after stroke with minimalinjury risk to therapists?Additionally, larger sample sizes and rigorous method-ology investigating the efficacy of powered exoskeletonsin stroke will further strengthen findings for or againsttheir utilization for gait rehabilitation.At the moment there is insufficient evidence to advocatein favour or against use of powered exoskeletons in clin-ical practice. The patient’s acuity and functional presenta-tion need to be considered and the extent of benefit hasyet remain to be determined through high quality re-search. The devices, however, have been shown to be safeand feasible for use with stroke patients. They can be usedto mobilize more impaired individuals without physicallystraining therapists. It thus remains up to therapists to usetheir own clinical judgement of whether to utilize poweredexoskeletons with their patients for gait rehabilitation,considering its application for weight-bearing, standing,and automated walking.LimitationsThere are a few limitations with the present review. This re-view excluded non-English studies, which may have led toan incomplete synthesis of data, given that some exoskele-tons are developed in non-English countries such as Japan,Germany, Iran, Israel, and Spain. There was heterogeneityin the studies, especially with variability in the training pro-tocols and exoskeletons utilized (control mechanism, uni-lateral or bilateral application), which makes interpretationof the results challenging. In addition, type, side, and sever-ity of stroke and comorbid conditions were not consideredin this review because of the scarcity of studies in this area.As more research trials in stroke rehabilitation using pow-ered exoskeletons are conducted, a systematic review willbe able to address these additional considerations.ConclusionCurrently, clinical trials demonstrate that powered ro-botic exoskeletons can be used safely as a gait trainingintervention for sub-acute and chronic stroke. Prelimin-ary findings suggest that exoskeletal gait training isequivalent to traditional therapy for chronic stroke pa-tients, while sub-acute patients may experience addedbenefit from exoskeletal gait training. Efforts should beinvested in designing rigorous, appropriately poweredcontrolled trials before it can be translated into a clinicaltool for gait rehabilitation post-stroke.Louie and Eng Journal of NeuroEngineering and Rehabilitation  (2016) 13:53 Page 8 of 10Abbreviations10MWT, 10-m walk test; 6MWT, 6-min walk test; BWSTT, body weight-supportedtreadmill training; FAC, functional ambulation category; HAL, hybrid assistivelimb; MCID, minimal clinically important difference; RCT, randomized controlledtrial; RPE, rate of perceived exertion; TUG, timed up and goAcknowledgementsThe authors acknowledge and thank Christina Cassady (CC), who acted assecond reviewer in the search portion of this review.FundingThe project was supported by funding from a Grant-in-Aid from the Heartand Stroke Foundation of Canada and the Canada Research Chairs Program.Availability of data and materialsNo original data is presented in this review.Authors’ contributionsDRL formulated the idea, performed the search, and prepared themanuscript for this review. JJE participated in the writing process. Bothauthors read and approved the final manuscript.Competing interestsThe authors declare that they have no competing interests.Consent for publicationNot applicable.Ethics approval and consent to participateThis scoping review did not collect original data on human subjects;therefore ethics approval is not applicable. All studies included in this reviewreported receiving ethical approval and gaining consent from theirparticipants.Author details1Graduate Program in Rehabilitation Sciences, University of British Columbia,Vancouver, BC, Canada. 2Department of Physical Therapy, University of BritishColumbia, 212-2177 Wesbrook Mall, Vancouver, BC V6T 1Z3, Canada.3Rehabilitation Research Program, Vancouver Coastal Health ResearchInstitute, Vancouver, BC, Canada.Received: 16 March 2016 Accepted: 3 June 2016References1. Feigin VL, Krishnamurthi RV, Parmar P, Norrving B, Mensah GA, Bennett DA,et al. Update on the global burden of ischemic and hemorrhagic stroke in1990-2013: the GBD 2013 study. Neuroepidemiology. 2015;45:161–76.2. Mukherjee D, Patil CG. Epidemiology and the global burden of stroke.World Neurosurg. 2011;76:S85–90.3. Kwakkel G, Kollen BJ. Predicting activites after stroke: what is clinicallyrelevant? Int J Stroke. 2013;8:25–32.4. Bohannon RW, Andrews AW, Smith MB. Rehabilitation goals of patients withhemiplegia. Int J Rehabil Res. 1988;11:181–3.5. Harris JE, Eng JJ. Goal priorities identified by individuals with chronic stroke:Implications for rehabilitation professionals. Physiother Canada. 2004;56:171–6.6. Portelli R, Lowe D, Irwin P, Pearson M, Rudd AG. Institutionalization afterstroke. Clin Rehabil. 2005;19:97–108.7. Jorgensen HS, Nakayama H, Raaschou HO, Olsen TS. Recovery of walkingfunction in stroke patients: the Copenhagen Stroke Study. Arch Phys MedRehabil. 1995;76:27–32.8. Kollen BJ, Kwakkel G, Lindeman E. Longitudinal robustness of variablespredicting independent gait following severe middle cerebral artery stroke:a prospective cohort study. Clin Rehabil. 2006;20:262–8.9. Chen G, Chan CK, Guo Z, Yu H. A review of lower extremity assistive roboticexoskeletons in rehabilitation therapy. Crit Rev Biomed Eng. 2013;41:343–63.10. Sale P, Franceschini M, Waldner A, Hesse S. Use of the robot assisted gaittherapy in rehabilitation of patients with stroke and spinal cord injury. Eur JPhys Rehabil Med. 2012;48:111–21.11. Scrivener K, Sherrington C, Schurr K. Exercise dose and mobility outcome ina comprehensive stroke unit: description and prediction from a prospectivecohorot study. J Rehabil Med. 2012;44:824–9.12. French B, Thomas L, Leathley M, Sutton C, McAdam J, Forster A, et al. Doesrepetitive task training improve functional activity after stroke? A Cochranesystematic review and meta-analysis. J Rehabil Med. 2010;42:9–15.13. States RA, Salem Y, Pappas E. Overground gait training for individualswith chronic stroke: a Cochrane systematic review. J Neurol PhysTher. 2009;33:179–86.14. Hornby TG, Holleran CL, Hennessy PW, Leddy AL, Connolly M, Camardo J,et al. Variable intensity early walking poststroke (VIEWS): a randomizedcontrolled trial. Neurorehabil Neural Repair. 2015. doi:10.1177/1545968315604396.15. Rand D, Eng JJ. Disparity between functional recovery and daily use of theupper and lower extremities during subacute stroke rehabilitation.Neurorehabil Neural Repair. 2012;26:76–84.16. Mehrholz J, Pohl M, Elsner B. Treadmill training and body weight supportfor walking after stroke. Cochrane Database Syst Rev. 2014;1, CD002840.17. Ada L, Dean CM, Vargas J, Ennis S. Mechanically assisted walking withbody weight support results in more independent walking thanassisted overground walking in non-ambulatory patients early afterstroke: a systematic review. J Physiother. 2010;56:153–61.18. Mehrholz J, Elsner B, Werner C, Kugler J, Pohl M. Electromechanical-assistedtraining for walking after stroke. Cochrane Database Syst Rev. 2013;7,CD006185.19. Hidler J, Nichols D, Pelliccio M, Brady K, Campbell DD, Kahn JH, et al.Multicenter randomized clinical trial evaluating the effectiveness ofthe Lokomat in subacute stroke. Neurorehabil Neural Repair.2009;23:5–13.20. Dobkin BH, Duncan PW. Should body weight-supported treadmilltraining and robotic-assistive steppers for locomotor training trotback to the starting gate? Neurorehabil Neural Repair. 2012;26:308–17.21. Turchetti G, Vitiello N, Trieste L, Romiti S, Geisler E, Micera S. Whyeffectiveness of robot-mediated neurorehabilitation does not necessarilyinfluence its adoption. IEEE Rev Biomed Eng. 2014;7:143–53.22. Kolakowsky-Hayner SA, Crew J, Moran S, Shah A. Safety and feasibility ofusing the Ekso™ bionic exoskeleton to aid ambulation after spinal cordinjury. J Spine. 2013. doi:10.4172/2165-7939.S4-003.23. Kressler J, Thomas CK, Field-Fote EC, Sanchez J, Widerström-Noga E,Cilien DC, et al. Understanding therapeutic benefits of overgroundbionic ambulation: Exploratory case series in persons with chronic,complete spinal cord injury. Arch Phys Med Rehabil. 2014;95:1878–87.24. Esquenazi A, Talaty M, Packel A, Saulino M. The ReWalk poweredexoskeleton to restore ambulatory function to individuals with thoracic-levelmotor-complete spinal cord injury. Am J Phys Med Rehabil. 2012;91:911–21.25. Louie DR, Eng JJ, Lam T, SCIRE Research Team. Gait speed using poweredrobotic exoskeletons after spinal cord injury: a systematic review andcorrelational study. J Neuroeng Rehabil. 2015;12:82.26. Arazpour M, Samadian M, Bahramizadeh M, Joghtaei M, Maleki M, Bani MA,et al. The efficiency of orthotic interventions on energy consumption inparaplegic patients: a literature review. Spinal Cord. 2015;53:168–75.27. Federici S, Meloni F, Bracalenti M, De Filippis ML. The effectiveness ofpowered, active lower limb exoskeletons in neurorehabilitation: aSystematic review. NeuroRehabilitation. 2015;37:321–40.28. Wall A, Borg J, Palmcrantz S. Clinical application of the Hybrid Assistive Limb(HAL) for gait training – a systematic review. Front Syst Neurosci. 2015;9:48.29. Arksey H, O’Malley L. Scoping studies: towards a methodological framework.Int J Soc Res Methodol. 2005;8:19–32.30. Levac D, Colquhoun H, O’Brien KK. Scoping studies: advancing themethodology. Implement Sci. 2010;5:69.31. Maeshima S, Osawa A, Nishio D, Hirano Y, Takeda K, Kigawa H, et al. Efficacyof a hybrid assistive limb in post-stroke hemiplegic patients: a preliminaryreport. BMC Neurol. 2011;11:116.32. Byl NN. Mobility training using a bionic knee orthosis in patients in a post-stroke chronic state: a case series. J Med Case Rep. 2012;6:216.33. Wong CK, Bishop L, Stein J. A wearable robotic knee orthosis for gaittraining: a case-series of hemiparetic stroke survivors. Prosthet Orthot Int.2012;36:113–20.34. Kawamoto H, Kamibayashi K, Nakata Y, Yamawaki K, Ariyasu R, Sankai Y,et al. Pilot study of locomotion improvement using hybrid assistive limb inchronic stroke patients. BMC Neurol. 2013;13:141.Louie and Eng Journal of NeuroEngineering and Rehabilitation  (2016) 13:53 Page 9 of 1035. Nilsson A, Vreede KS, Haglund V, Kawamoto H, Sankai Y, Borg J. Gait trainingearly after stroke with a new exoskeleton – the hybrid assistive limb: a studyof safety and feasibility. J NeuroEng Rehabil. 2014;11:92.36. Stein J, Bishop L, Stein DJ, Wong CK. Gait training with a robotic leg braceafter stroke: a randomized controlled pilot study. Am J Phys Med Rehabil.2014;93:987–94.37. Watanabe H, Tanaka N, Inuta T, Saitou H, Yanagi H. Locomotionimprovement using a hybrid assistive limb in recovery phase strokepatients: a randomized controlled pilot study. Arch Phys Med Rehabil. 2014;95:2006–12.38. Bortole M, Venkatakrishnan A, Zhu F, Moreno JC, Francisco GE, Pons JL,et al. The H2 robotic exoskeleton for gait rehabilitation after stroke: earlyfindings from a clinical study. J NeuroEng Rehabil. 2015;12:54.39. Buesing C, Fisch G, O’Donnell M, Shahidi I, Thomas L, Mummidisetty CK,et al. Effects of a wearable exoskeleton stride management assist system(SMA®) on spatiotemporal gait characteristics in individuals after stroke: arandomized controlled trial. J NeuroEng Rehabil. 2015;12:69.40. Fukuda H, Samura K, Hamada O, Saita K, Ogata T, Shiota E, et al. Effectivenessof acute phase hybrid assistive limb rehabilitation in stroke patients classifiedby paralysis severity. Neurol Med Chir (Tokyo). 2015;55:487–92.41. Yoshimoto T, Shimizu I, Hiroi Y, Kawaki M, Sato D, Nagasawa M. Feasibilityand efficacy of high-speed gait training with a voluntary driven exoskeletonrobot for gait and balance dysfunction in patients with chronic stroke:nonrandomized pilot study with concurrent control. Int J Rehabil Res. 2015;38:338–43.42. Tang A, Eng JJ, Rand D. Relationship between perceived and measuredchanges in walking after stroke. J Neurol Phys Ther. 2012;36:115–21.43. Perera S, Mody SH, Woodman RC, Studenski SA. Meaningful change andresponsiveness in common physical performance measures in older adults.J Am Geriatr Soc. 2006;54:743–9.44. Flansbjer UB, Homback AM, Downham D, Patten C, Lexell J. Reliability ofgait performance tests in men and women with hemiparesis after stroke. JRehabil Med. 2005;37:75–82.45. Tilson JK, Sullivan KJ, Cen SY, Rose DK, Koradia CH, Azen SP, et al.Meaningful gait speed improvement during the first 60 days poststroke:minimal clinically important difference. Phys Ther. 2010;90:196–208.•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:Louie and Eng Journal of NeuroEngineering and Rehabilitation  (2016) 13:53 Page 10 of 10


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items