| Medical Policy |
| Subject: Powered Robotic Lower Body Exoskeleton Devices | |
| Document #: OR-PR.00006 | Publish Date: 06/28/2024 |
| Status: Reviewed | Last Review Date: 05/09/2024 |
| Description/Scope |
This document addresses the use of powered, robotic lower body exoskeleton devices that may be utilized in the rehabilitation of or for daily use by individuals with neurological disorders that affect an individual’s ability to ambulate without assistance.
Note: For information regarding other prosthesis, please see the following:
| Position Statement |
Investigational and Not Medically Necessary:
The use of a powered, robotic lower body exoskeleton device is considered investigational and not medically necessary under all circumstances, including but not limited to the following:
| Rationale |
Use of lower limb electromechanical exoskeleton devices have been proposed to assist in locomotion and rehabilitation for a variety of conditions, including spinal cord injury (SCI), traumatic brain injury (TBI), stroke, multiple sclerosis (MS), and others. The available scientific evidence addressing the clinical utility of such devices is addressed below. Multiple devices have been described in the literature, including the following:
Spinal Cord Injury (SCI).
Multiple case series studies have been published describing the use of lower-limb exoskeleton devices for individuals with SCI (Asselin, 2015; Bach Baunsgaard, 2018; Benson, 2016; Esquenazi, 2012; Hartigan, 2015; Karelis, 2017; Platz, 2016; Stampacchia, 2016; Tefertiller, 2017; van Nes, 2022; Yang, 2015; Zeilig, 2012). There were significant variations in training regimens used, from a single 40-minute session per study participant to over sixty 90-minutes sessions per participant and a wide variety of measures and outcomes reported. Such measures include stair climb time, 6-minute walk test (6MTW), 10-minute walk test (10MWT), sit-to-stand and stand-to-sit time, timed up and go (TUG), and walking speed. Some studies reported on outcomes related to fatigue, pain, bowel and bladder function, oxygen uptake, spasticity, exertion, range of motion, loss lean body mass, and BMI using a wide variety of measurement tools. In addition, several different devices were used, some available in the U.S. and some not. Overall, study populations were relatively limited in size (n=6 to 42), and most did not investigate use of the devices in the community setting or on varying surfaces or terrain. The results of these studies are heterogenous, some reporting improvements in measures and others little to none. Similarly, the studies vary significantly on the details provided related to adverse events, with some reporting on them and others not. Overall, there have been no reports of falls, cardiovascular events, fractures, or autonomic dysreflexia, Conversely, there have been multiple reports or abrasions, lightheadedness, lower limb edema, pain, bruising, blister, and one sprained ankle. These studies are uniformly short-term, with no long-term results reported.
Additionally, several comparative studies have been published describing the use of lower-limb exoskeleton devices is for individuals with SCI. Fineberg (2013) described a study of 6 participants with motor-complete SCI and 3 able-bodied volunteers using the ReWalk device. All participants underwent 1-2 hours of combined sitting and walking sessions 3 times per week for 5 to 6 months. The objective of the study was to use vertical ground reaction force (vGRF) to show the magnitude and pattern of mechanical loading in participants during walking using the F-Scanin-shoe pressure mapping system (Tekscan, Inc., South Boston, MA). The investigators measured the pressure imparted to both the left and right feet in all SCI participants while walking in the ReWalk device and in control participants during unassisted walking. A total of 3 of the SCI participants participated in the measurement trials using assistive devices such as walkers and the remaining 3 were unassisted. For measurements of peak stance average (PSA) the assisted SCI group had significantly lower vGRF vs. control participants (p<0.05). No significant difference in PSA was noted between the no-assist SCI group and controls. Significant differences between the assist and no assist SCI groups was also noted, with the no-assist group participants creating greater VGRF than the assisted participants (p=0.010 for midstance and p=0.045 for toe off). The authors concluded that powered exoskeleton-assisted walking in individuals with motor-complete SCI generated vGRF similar in magnitude and pattern to that of able-bodied walking. They suggested these results demonstrated the potential for powered exoskeleton-assisted walking to provide a mechanism for mechanical loading to the lower extremities in individuals with SCI.
Xiang and colleagues (2021) reported the results of a single-center, randomized controlled pilot study exploring the effects of exoskeleton-assisted walking (EAW, AIDER [AssItive DEvice for paRalyzed patient] device) on pulmonary function and walking parameters compared to conventional rehabilitation training in individuals with SCI. The study involved 18 individuals, previously diagnosed with an SCI, who were randomized to receive either EAW or conventional training delivered as 50 to 60 minute sessions, 4 times per week for 4 weeks. Regarding pulmonary function parameters after training, significantly greater values were observed for forced vital capacity (FVC, p=0.041), FVC% (p=0.012), and forced expiratory volume in 1 second (FEV1, p=0.013) in the EAW group compared to the conventional treatment group. Differences in values for forced expiratory flow, peak expiratory flow, and maximal voluntary ventilation were not statistically significant. Only 10 participants completed the final 6MWT, of which 2 were in the conventional treatment group. There was no difference in lower extremity motor score. The results of this study suggest EAW has the potential to improve some pulmonary function parameters among this group of individuals with SCI, although the clinical significance is uncertain. Interpretation of study results is limited by the overall sample size and proportion of participants lost to follow-up.
Evans (2021) published the results of a pilot RCT involving 16 participants with SCI-related tetraplegia who underwent 24 weeks of either robotic exoskeleton therapy with the Esko GT device or activity-based training (n=8 per group). The exoskeleton group underwent treatment consisting of standing and walking 10 to 50 minutes in the Esko device and between 50 and 1800 steps taken. The control activity-based group treatment consisted of a combination of resistance, cardiovascular, and flexibility training. This group also had the option of receiving gait retraining. The authors reported no statistically significant differences between groups or over time for brachial systolic and diastolic blood pressure, ankle systolic pressure, or ankle brachial pressure index. At 24 weeks, heart rate in the standing position was significantly higher in the control group vs. the exoskeleton group (95.6 beats/min vs. 75.1 beats/min; p=0.05). No significant differences in heart rate were reported during the 6-minute arm ergometry test. No statistically significant differences between groups, or over time were reported for heart rate variability indices in the supine or standing positions, or during the 6-minute arm ergometry test. The exoskeleton group had a significant increase in walking distance during the 6MWT from baseline to 24 weeks (68.3 m to 109.9 m).
Rodríguez-Fernández (2022) published the results of a randomized cross-over study involving 10 participants with SCI who underwent a 10-session, gait training program with both a knee‑ankle‑foot orthosis and the ABLE Exoskeleton device in succession. Randomization determined which device participants were started with, and then crossed over to after the initial training and evaluation period ended. Each training period was 5 weeks in duration with 2 sessions per week. Participants spent a minimum of 30 minutes per session doing sit-to-stand and stand-to-sit transitions, and standing and walking using the designated devices and the aid of a walker. There was a 2-week resting period between the use of each device. No significant differences were found between the groups with regard to distance covered during the 6MWT, time needed to complete the TUG, and gait speed during the 10MWT. As would be expected with the use of a powered assistive device, step length, range of motion (ROM) of both knee and hip joint, and ankle circumduction were significantly improved in the exoskeleton group vs. the control group (p=0.037, p=0.002, p=0.004, and p=0.014, respectively). No significant differences between groups were reported for metabolic cost of transport (MCoT) and the intensity of physical activity as a percentage of peak oxygen uptake (%VO2peak) during the 6MWT. As with similar trials previously discussed, the clinical relevance of these results is unclear, being from such an underpowered, short-term study.
Edwards (2022) published the results of an RCT involving 25 participants with incomplete SCI who underwent a 12-week gait training program. Participants were assigned on a 2:2:1 basis to one of the following: 1) 45-minute training session with the Esko GT exoskeleton 3 times a week, including overground training without bodyweight support when possible, 2) active control training involving 45-minute bodyweight supported treadmill training and overground training without bodyweight support when possible, and 3) passive control in which participants continued daily activities with no new gait training, mobility therapy, or new medications related to the condition under study. The final study included 9 exoskeleton participants, 10 active control participants, and 6 passive control participants. The results at 12 weeks revealed no significant between-group difference for self-selected gait speed (p>0.05), with all groups having improvements (51% for the exoskeleton group, 32% for the active control group, and 14% for the passive control group). Similar findings were reported for maximum gait speed at 12 weeks (44% for the exoskeleton group, 50% for the active control group, and 14% for the passive control group; between-group comparisons p>0.05). The group with the highest proportion of change in clinical ambulation category was the exoskeleton group (5 of 9 participants). The active control group followed with 3 of 10 participants and the passive control group with no changes (between-group difference in proportions p<0.05). No significant differences were reported between groups for the 6MWT or the TUG. A majority of both exoskeleton participants (8/9) and active control participants (7/10) had no change in type of assistive device used throughout the study period. No change in assistive device type was reported for the passive control group. Three serious adverse events were reported: two urinary tract infections deemed to be unrelated to the device, and one active control group participant was admitted to the hospital with lower extremity numbness and a UTI. The numbness was deemed to be “possibly related” to the bodyweight supported treadmill training. No falls were reported. Non-serious adverse events deemed “possibly” or “probably” related to the device or training process included upper and lower extremity musculoskeletal issues (8 exoskeleton group and 4 active control group), increased spasticity (3 exoskeleton group and 1 active control group), skin issues (5 exoskeleton group and 1 active control group); and 1 visceral issue (1 exoskeleton group). The authors concluded that use of the exoskeleton device may improve ambulatory status, but “While generally safe and tolerable, larger gains in ambulation might be associated with higher risk for non-serious adverse events.”
Shackleton (2022) conducted a secondary analysis of the previously discussed study conducted by Evans (2021). This study evaluated the impact of the Esko GT device on bone mineral density (BMD) in 16 incomplete SCI participants who underwent 60-minute activity-based training sessions 3 times per week for a total of 24 weeks. No significant changes in spinal BMD were reported for either the exoskeleton or control groups (p=0.86). However, a significant decrease in hip and femoral neck BMD was reported in the control group (p=0.04 for hip and p=0.04 for femoral neck). A significant 7% increase in arm fat-free soft tissue mass (FFSTM) was reported in both groups (p<0.01 for both). No change in leg FFSTM occurred in either group (p=0.32). The control group showed a significant 15% decrease in visceral adipose tissue (p=0.04) and 13% decrease in gynoid fat mass (p<0.01). No similar findings were reported for the exoskeleton group. The authors concluded that exoskeleton-based training aided in the prevention of spine, hip and femoral neck BMD. The findings of this trial should be further evaluated.
Gil‑Agudo (2023) reported the results of an evaluator-blinded RCT involving 23 participants with incomplete SCI less than 1 year who underwent either standard rehabilitation (15 30-minutes sessions, n=10) or rehabilitation sessions with the HANK device (15 one-hour sessions, n=11). The original study design called for a total of 42 participants, 24 per group). The majority of our patients were less than 6 months post-injury. No falls were reported. Some instances of skin redness alleviated with padding was reported. Mild neck and shoulder muscle pain related to the use of walking aides was also reported. While significant improvements were seen in both groups, no differences between groups were noted on the Lower Extremity Motor Scale (LEMS) if on the 10MWT, TUG, 6MWT, walking speed test, or the WISCI‑II (Walking Index for Spinal Cord Injury II) test. The results of this trial demonstrate that use of the HANK device may provide improvements in functional outcomes, but no benefit is provided over standard rehabilitation methods.
Hu (2024) reported on a evaluator-blinded RCT involving 16 participants (n=8 in each group) with incomplete SCI less than 1 year who underwent either conventional rehabilitation (aerobic exercise and strength training) or 40-50 minute sessions with the AIDER powered robotic exoskeleton 5 times a week for 8 weeks in addition to conventional rehabilitation. The authors reported no significant differences between groups on the overall score or of any domain of the World Health Organization quality of life-BREF (WHOQOL-BREF) at the end of the 8-week study period. Similarly, no significant differences between groups were reported on the overall Spinal Cord Independence Measure III (SCIM-III). However, at 8 weeks the score for the activity subcategory was significantly higher in the control group vs. AIDER group. These results appear to support the conclusion that no benefit is provided from AIDER training over standard rehabilitation methods.
A meta-analysis of the clinical effectiveness and safety of powered exoskeleton devices was published by Miller and colleagues in 2016. The investigators included 11 studies involving 111 participants using the ReWalk (8 studies), Ekso (3 studies), and Indego (2 studies) devices. An additional study involved an unspecified device. Considerable heterogeneity was present across studies with regard to methods and duration of training, in addition to outcomes such as ambulatory performance, metabolic demand, and perceived health benefits. No serious adverse events were reported. The incidence of falls during training was 4.4% (n=3), and all falls were reported in the same study (Kozlowski, 2015). The falls were attributed to programming errors in the first generation Ekso device in 2 cases and to a crutch malfunction in the 3rd case. The authors concluded that, “Powered exoskeletons allow patients with SCI to safely ambulate in real-world settings at a physical activity intensity conducive to prolonged use and known to yield health benefits.” However, this conclusion is weakened by the fact that the majority of studies took place in the investigational setting.
Stroke
As noted above, the use of lower extremity exoskeleton devices have been used to assist in post-stroke physical rehabilitation. There are a limited number of published case studies addressing such use (Ii, 2020; Molteni, 2017; Sakel, 2022; Sczesny-Kaiser, 2019); study designs are heterogenous in terms of population size, treatment regimens, measures and outcomes evaluated, devices used, duration of follow-up, and other factors. Additionally, the stroke literature varies with regard to time from stroke to therapy initiation. These studies are uniformly short-term, with no long-term results reported. These studies reported on outcomes including walking efficiency and balance. Descriptions and rates of adverse events were not included in these reports.
Ii (2020) published the results of a study involving 36 participants with hemiparesis post-stroke who underwent gait training using the Welwalk device, who were compared with matched control participant data from a hospital database. Training consisted of 40 minutes of device-based activity 5-7 days a week, in addition to the usual rehabilitation 3 hours a day, for 6 or 7 days per week. The control group underwent only the usual rehabilitation program. The primary outcome was improved efficiency of walking ability, calculated from baseline evaluation to achieving a Functional Independence Measure (FIM)-walk score of 5 using a specified formula. Evaluations were taken every 2 weeks for a total of 8 weeks. Overall, improvement in efficiency of FIM-walk was significantly higher in the exoskeleton group vs. the control group (p<0.001). Participants in the exoskeleton group became able to walk with the device with no assistance in approximately 1.4 weeks, and achieved an FIM-walk score of 5 in approximately 3 weeks. The authors stated their results demonstrated that use of the exoskeleton device successfully improved walking efficiency in participants with post-stroke hemiparesis in a real-world rehabilitation facility setting.
Comparative studies
Thimabut (2022) conducted a prospective, assessor-blinded, RCT involving 26 participants with stroke-related hemiplegia who received 30 physiotherapy and ambulation training sessions (60 minutes each) 5 days a week for 6 weeks, with or without the Welwalk exoskeleton device (n=13 per group). The average age of participants in the exoskeleton group was 52.8 years vs 62.8 years in the control group (p=0.009) and there were 6 male participants (46.2%) in the Welwalk group vs. 10 (76.9%) in the control group (p=0.02). This is relevant because the investigators observed that in a multivariable analysis of the exoskeleton group, age and sex were the only significant covariables (p=0.003 and p=0.035, respectively). Both groups showed statistically significant improvement from baseline in FIM-walk scale measures (p≤0.001 at both the 15th and 30th session for both groups). Between-group comparisons showed that the exoskeleton group had a significantly higher FIM-walk score than the control group at the end of the 15th session (p=0.012). However, that difference disappeared at the end of the 30th session (p=0.070). Both groups had significant improvements in 6MWT results (p<0.05), but no difference between groups was reported. The Barthel ADL index indicated significantly more improvement in ADLs in the exoskeleton group vs. the control group (p<0.001). Finally, gait symmetry ratio at the completion of the study was significantly different between groups (p=0.044). There were no significant differences in other gait parameters. The authors concluded that their results demonstrated early improvements in walking ability and Barthel ADL index with the Welwalk device. Additional investigations into the benefits of this device are warranted to support these findings and evaluate long-term benefits. Only the EQ-5D score was significantly higher in the exoskeleton group at the end of the trial (p=0.028).
Yoo (2023a) reported the results of an RCT involving 25 participants recovering from subacute stroke. All participants took part in a conventional daily stroke neurorehabilitation program (90 minutes/day, 5 days per week, for 4 weeks). In addition, participants were randomized to receive additional 12 sessions of gait training (30 minutes/day, 3 days per week, for 4 weeks) with either the ExoAtlet Medy exoskeleton device (n=16) or standard physiotherapy (n=9). Of the original 25 participants randomized, only 17 (9 in the exoskeleton group and 8 in the control group), completed the study. At the completion of the study, no significant differences between groups were reported for the primary outcome, change in functional ambulatory category (FAC). Both groups demonstrated significant improvements in the Berg Balance Scale (BBS), K-MBI (Korean version of the modified Barthel index), and EQ-5D scores after 12 gait training sessions (p<0.05). Additionally, exoskeleton group participants had significant improvements in the FAC, TUG, and 10MWT vs. the control group, which had no significant improvements in those measures. In the between-group analysis, the improvements in the FAC and EQ-5D were statistically higher in the exoskeleton group vs. the control group (p=0.046 and p<0.015, respectively). No significant improvement was observed in pulmonary function in either group. No severe adverse events were reported. This study found early benefits to the use of exoskeleton device in some, but not all assessed outcomes for individuals with subacute stroke undergoing rehabilitation; additional long-term study is warranted
Pan (2023) published a prospective case series study of 5 participants with post-stroke hemiparesis who underwent walking trials using the Gait Enhancing and Motivating System with additional thoracolumbar interface. Three participants were classified as community ambulators, and 2 participants were classified as limited community ambulators. The purpose of the study was to evaluate the effect of different hip exoskeleton assistance strategies on gait function and gait biomechanics. The investigators’ central hypothesis was that individuals with hemiparesis due to stroke would have increased self-selected walking speed using a powered hip exoskeleton that applies bilateral assistance to both the paretic and non-paretic limbs vs. the paretic or non-paretic limb only. The protocol required the participants to complete 4 passes across a 6-meter walkway at their self-selected walking speed, and then walk on an instrumented split-belt treadmill at 80% of the overground self-selected walking speed for 1 minute. Participants were then asked to wear and acclimate to the exoskeleton, and allow the device to be calibrated to the individual before engaging in a second session. Participants then walked on a treadmill for 1 minute with the exoskeleton set with one of 16 different combinations of assistance magnitude for both the paretic and non-paretic limbs. Participants then completed 4 passes on the 6-meter walkway with the same assistance condition. The authors reported that the exoskeleton significantly influenced self-selected walking speed (p<0.001). Of the assistance conditions tested, 10 of the 16 resulted in changes that exceeded the minimal clinically important difference, defined as an increase of 0.1 m/s for self-selected walking speed, corresponding to about a 12.4% change with respect to the baseline. All assistance strategies resulted in a greater paretic-limb step length increase vs. baseline than the minimal detectable change, defined as 2.62 cm and 2.11 cm for the paretic and non-paretic step length, respectively. Anterior ground reaction force (AGRF) in the non-paretic limb was greater than the minimal clinically important difference of 0.8% BW (BW was undefined) in only one setting. For the paretic limb, 6 of the 10 settings resulted in a greater increase in AGRF than the minimal clinically important difference threshold. The authors concluded that when compared to walking without a device, the use of a hip exoskeleton device improved participants self-selected overground walking speed with both bilateral and unilateral assistance strategies (p<0.05 for both). They added that both bilateral and unilateral assistance strategies significantly improved step length in both paretic and non-paretic limbs (p<0.05 for both). The clinical utility of these findings and association with health outcomes remains to be elucidated.
Yoo (2023) conducted an unblinded RCT involving 30 participants with post-stroke hemiparesis assigned to rehabilitation with ten 30-minute sessions over 4 weeks with either a treadmill alone of or with the Healbot G device (n=15 in each group). The Healbot group had 11 participants finish the trial with complete data and the control group had 14 with complete data. The primary outcome measure was cortical activation changes represented by the mean oxyhemoglobin (HbO2) concentration, assessed using functional near infrared spectroscopy (fNIRS), which according to the authors enables visualization of cortical activation during human gait through the detection of hemoglobin oxygenation. The authors reported that cortical activity increased in both groups over the course of the study, but was more pronounced after training in the Healbot group (no intergroup comparisons p-values provided). They also reported that the Healbot group performed significantly better post-therapy with regard to FAC, BBS, Motricity Index for the lower extremities (MI-Lower), and gait symmetry ratio. However, no statistics were provided for those measures. No data were provided regarding adverse events. The impact of these results are clinically unclear and limited in their generalizability.
Yokota (2023) conducted an RCT involving 25 participants with post-stroke hemiparesis assigned to standard gait training rehabilitation with (n=12) or without (n=10) use of the HAL device. Training was initiated 10 days following the index event and consisted of 1 to 3 20-minute sessions per day 5 or 6 days a week for a total of 20 sessions. No significant differences between groups were reported with regard to the primary outcomes, change in FIM and FAC. No data were provided regarding adverse events.
Elmas Bodur (2024) conducted an RCT involving 32 participants with post-stroke hemiparesis assigned to rehabilitation with either the ExoAthlet exoskeleton or the Lokomat Free-D device (n=16 each). The rehabilitation program consisted of standard therapy in conjunction with 60-minute training sessions three days a week for eight weeks with the assigned device. The authors reported a statistically significant difference for each group with regard to functional independence, functional capacity, and quality of life at the end of the trial (p<0.05 for all), but no significant differences were found in between groups in the FIM, 6MWT, and 30-Second Chair Stand Test(p>0.05). No data were provided regarding adverse events.
Multiple Sclerosis
The use of lower-limb exoskeleton devices for individuals with MS has been described in a limited number of studies.
Kozlowski (2017) reported the findings of the first study involving the use of the ReWalk device in 5 participants with MS. This cohort study involved individuals with Expanded Disability Status Scale (EDSS) scores ranging from 5.5 to 7.0, who are medically stable, fit the ReWalk device, could tolerate standing for 30 minutes, and it had been at least 1 year since their last relapse. Evaluations were taken at baseline and at weeks 1, 4, 8 (baseline period), 12 and 16 (intervention period) and at week 20 (follow-up). Participants underwent 3, 30- to 90-minute training sessions per week for a total of 24 sessions. The study originally enrolled 13 participants, but 2 failed screening and 6 withdrew, either due to transportation issues or device-related pain. A total of 5 participants completed a minimum of 20 walking sessions. The investigators commented that regression of progress was noted with gaps in the training interval that exceeded 4 days, but losses were usually regained within a single session. Learnability was considered high, with most participants attaining walking and sitting well within the 24-session limit. No serious adverse events were reported, but skin issues were common, with 151 events per 1000 hours of exposure to training. Qualitative postural improvements were reported for 4 of the 5 participants. No overall improvement was reported with regard to the Neuro-QOL (quality of life) and Patient Reported Outcome Measurement and Information System (PROMIS) tools.
Androwis (2021) published the results of a randomized controlled pilot study evaluating the effects of 4 weeks of robotic exoskeleton-assisted exercise rehabilitation (REAER) compared to conventional gait training in individuals with substantial disability due to MS. The experimental condition involved supervised and progressive overground walking using the Ekso-GT robotic exoskeleton. The outcomes of interest were the effects of REAER on functional mobility (assessed by TUG), walking endurance (assessed by 6MWT), cognitive processing speed ([CPS], assessed by Symbol Digit Modalities Test [SDMT]), and brain connectivity (assessed by thalamocortical resting-state functional connectivity [RSFC] on fMRI). The study involved 10 individuals with substantial MS-related neurological disability. Although there were large improvements in functional mobility in the REAER group based on effect size estimates, there was no significant between-group difference in functional mobility (p=0.06) or walking endurance. There was a significant between-group difference in cognitive processing speed (p=0.02) and brain connectivity (p<0.01) in favor of the REAER group. Though the improvements in some individuals in this study population are promising, additional evidence in appropriately powered trials of sufficient duration is needed to confirm the durable effects of REAER on mobility and cognition in individuals with substantial MS-related disability.
Berriozabalgoitia and colleagues (2021) reported the results of a randomized controlled trial (RCT) evaluating the use of overground robotic training in addition to a conventional outpatient physical therapy program in individuals with MS. The study involved 36 individuals with Expanded Disability Status Scale score between 4.5 and 7, and the need for assistive devices for walking outdoors. Participants were randomized to a conventional physical therapy program consisting of individualized, weekly, 1-hour sessions (control group, n=14) or conventional therapy plus overground gait training (OR group, n=18). Overground gait training consisted of a twice-weekly, individualized, and progressive intervention for 3 months using the Ekso wearable exoskeleton. The primary outcome was performance on the 10MWT. Secondary variables included the Short Physical Performance Battery, TUG, and Modified Fatigue Impact Scale. There were no statistically significant between-group differences regarding the 10MWT. In the OR group, there was significant improvement on the TUG (p=0.049, medium effect size) without an increase in fatigue perception. However, no time per group interactions were observed for any variable.
Sakel (2022) reported the results of a prospective case series study involving 10 participants with MS who underwent balance training with the Rex Rehab Robot exoskeleton device. The training focused on strengthening leg extensor and abdominal muscles, maintaining an optimal upright posture through standing in the exoskeleton device, and performing dynamic balance exercises. A total of 4 sessions over 1 month were conducted per participant, and measures were taken before and after the completed training series. Statistically significant improvements were reported for 4 participants based on Berg Balance Scale (BBS) measures, but no overall changes were noted for the study population. Impact of MS on daily lives was improved for 80% of participants based on Multiple Sclerosis Impact Scale (MSIS-29) measures (p=0.006). Similar findings were reported with 70% of participants having significant improvement on their Health-related Quality of Life Scale (EQ-5D-5L) scores (p=0.02). Spasticity, based on Modified Ashworth Scale (MAS) outcomes, was statistically significantly reduced in the left ankle plantar flexors and dorsiflexors and right ankle dorsiflexors. No significant changes were reported on the Modified Falls Efficacy Scale (MFES), Activities-Specific Balance Confidence Scale (ABC), Multiple Sclerosis Walking Scale (MSWS-12), Arm Activity Measure (ArmA), or the Epworth Sleepiness Scale (ESS). No device-related adverse events, including falls, were reported. The clinical relevance of these results is unclear, given the small sample size, short study duration, and absence of a control group
McGibbon (2023) reported the results of an RCT involving 35 individuals with MS who underwent rehabilitation with the Keeogo exoskeleton for two in-clinic visits where participants completed a battery of functional tests including 6MWT, Timed Stair Test (TST), and TUG, with and without the device. An additional 4- week at-home period followed where participants’ physical activity level was monitored for 2 weeks with the device and 2 weeks without it. Twenty nine participants completed the trial with full data collected. The authors reported shorter 6MWT distance and longer TST and TUG times when wearing the device compared to not wearing the device during the at-home period (no p-values provided). No significant changes pre- and post-trial in SF36 scores were reported. Performance was not correlated with change in disability level for 6MWT, TST, or TUG. Rehab effects and training effects, however, had significant correlations with change in disability: Change in 6MWT correlated with change in SF36-RF (role emotional, p=0.008) and SF36-EF (emotional functioning, p=0.021). Change in 6MWT also correlated with change in SF36-PF (physical function, p=0.019) and MS Walking Scale (p=0.019). The authors concluded that physical functioning and emotional well-being derived from use of the exoskeleton at home were associated with duration of device use.
Other Conditions
Use of lower-limb exoskeleton devices for individuals with an assortment of other conditions, including heart failure, osteoarthritis, and traumatic brain injury (TBI).
In 2013, Esquenazi and colleagues published a randomized comparative trial involving 16 participants with TBI. This study compared the use of the ReWalk device in treadmill assisted rehabilitation training (n=8) vs. manually assisted treadmill rehabilitation training (n=8). Following training, the average self-selected walking velocity (SSV) increased by 49.8% for the ReWalk group (p=0.01) and by 31% for the manual group (p=0.06). The average maximal velocity increased by 14.9% for the ReWalk group (p=0.06) and by 30.8% for the manual group (p=0.01). Step-length asymmetry ratio improved during SSV by 33.1% for the ReWalk group (p=0.01) and by 9.1% for the manual group (p=0.73). The distance walked increased by 11.7% for the ReWalk group (p=0.21) and by 19.3% for the manual group (p=0.03). While each group demonstrated benefits from their assigned training method, no differences between groups were reported. The value of the ReWalk system in rehabilitation training following TBI is unclear given these results.
A cohort study involving 5 participants with traumatic C7-T10 SCI and minimal spasticity was published by Karelis and others in 2017. All participants underwent a 6-week long training period involving 3, 3-hour long training sessions per week with the Ekso device. No changes in the American Spinal Injury Association Impairment Scale (AIS) were reported following completion of the training sessions. Significant changes were noted for leg and appendicular lean body mass, and total leg and appendicular fat mass. Total BMI increased significantly. No serious injuries were reported.
Just (2022) reported on a case series study evaluating the use of the MyoSuit in assisted mobilization in 20 participants with advanced heart failure (New York Heart Association [NYHA] class III heart failure). Participants underwent either a single session of activities of daily living evaluation (ADLs, n=10) or a single, standardized, 60-minute rehabilitation exercise session (n=10) with and without the exoskeleton device. The ADL session included evaluations of participants doing a 6MWD, standing, sitting down on a chair, standing up from a chair, and climbing stairs. The exercise session included dynamic walking training, combined with resistance exercise of the upper body, and dynamic and static balance training. The mean total walk distance of all participants without and with robotic assistance was 364.0 m and 325.2 m, respectively (p=0.241). No significant differences were reported for either the ADL or exercise groups with regard to rates of perceived exertion with or without the exoskeleton (p=0.932). No adverse events occurred during the study. The results of this limited study appear to indicate no benefit to the use of the MyoSuit for individuals with advanced heart failure.
McGibbon (2022) described the results of a randomized two-stage cross-over study involving 24 participants with unilateral knee osteoarthritis who underwent evaluation of both in-clinic and home use of the Keeogo™ exoskeleton over 6 weeks. Participants were randomly assigned to undergo evaluations in both settings both with and without the use of the exoskeleton device. No significant difference in 6MWT scores were reported between the exoskeleton and control groups in either setting (p=0.052 for clinic and p=0.237 for home). In the timed stair test, times for ascent and descent were both longer in the control group (p<0.001 for ascent and p=0.002 for descent). A similar result was reported for the TUG, with the exoskeleton group having faster times (p=0.004). At the completion of the study, significant improvements were reported for SF36 Energy/Vitality measure (p=0.005), SF36 General Health measure (p=0.043), Western Ontario and McMaster Universities Arthritis Index (WOMAC) pain (p=0.004) and WOMAC function (p=0.003). A total of four adverse events deemed related to the exoskeleton device were reported, including faulty battery charge causing device failure, thigh muscle pain remedied by adjusting device fit, and knee and ankle pain. The authors conclude that there were no immediate benefits to the use of the exoskeleton device, but cumulative effects were detected.
As with the other studies described above for SCI, stroke and MS, these studies are methodologically limited and their results cannot be generalized to wider populations. Further investigation in the form of rigorously designed and conduced trials is warranted to understand the potential benefits and harms of this technology.
Other Considerations
The American Heart Association and the American Stroke Association published guidelines for adult stroke rehabilitation and recovery in 2016 (Winstein, 2016). This document addressed the use of robotic and electromechanics-assisted training devices, and concluded that “Overall, although robotic therapy remains a promising therapy as an adjunct to conventional gait training, further studies are needed to clarify the optimal device type, training protocols, and patient selection to maximize benefits.”
In 2017, the Cochrane Library published a report assessing electromechanical-assisted training for walking after stroke (Mehrholz, 2017). This report concluded:
People who receive electromechanical-assisted gait training in combination with physiotherapy after stroke are more likely to achieve independent walking than people who receive gait training without these devices. We concluded that seven patients need to be treated to prevent one dependency in walking. Specifically, people in the first three months after stroke and those who are not able to walk seem to benefit most from this type of intervention. The role of the type of device is still not clear. Further research should consist of large definitive pragmatic phase III trials undertaken to address specific questions about the most effective frequency and duration of electromechanical-assisted gait training as well as how long any benefit may last.
Conclusion
The published evidence addressing robotic lower body exoskeleton devices does not permit reasonable conclusions concerning the effect of these devices on health outcomes. Supporting evidence is limited to studies with significant methodological issues, including lack of long-term follow-up and a paucity of appropriate control comparators. Additionally, there is significant heterogeneity among studies regarding the devices used, the intensity and duration of training regimens, outcomes and measures used, time to treatment from index incident, and duration and severity of conditions treated. To date, available studies have been limited to the research setting; results have also been highly variable, with some studies reporting statistically significant benefits and others not. The clinical significance of many of the most commonly reported measures such as 6MWD, TUG, etc., have not clearly demonstrated clinical utility with regards to use of robotic lower body exoskeleton devices. Other reported endpoints, less commonly reported, such as change in BMD, improvement in ADLs and cardiopulmonary function, and decrease in pain and spasticity, may provide a better reflection of these devices in terms of net health benefit. To date, such data has not been adequately addressed in the published literature. There has also been lack of reporting regarding the incidence of adverse events, with many studies not including such data. From the data that is available, joint pain and edema, falls, and skin abrasions have been reported. It is also unclear if the proposed benefits of using robotic lower body exoskeleton devices are durable, due to a lack of long-term study results. Data from well-designed studies of sufficient duration to evaluate durable outcomes including functional and physiologic endpoints using standardized measures are instrumental in helping to understand the overall place in care and identification of what populations are most likely to derive benefit from the use of these types of devices.
| Background/Overview |
Robotic lower body exoskeleton devices are intended to allow individuals with loss of lower limb function to ambulate on their own. When used, the device is worn outside clothing and consists of an upper-body harness, lower-limb braces, motorized joints, ground-force sensors, a tilt sensor, a locomotion-mode selector, and a backpack carrying a computerized controller and rechargeable battery. Using a wireless remote control worn on the wrist, the user commands the device to stand up, sit down, or walk. The device is strapped to the user at the waist, alongside each lower limb, and at the feet. Ordinary crutches are also utilized to help maintain stability.
There are several FDA approved robotic lower body exoskeleton devices on the market, including the ReWalk exoskeleton, Ekso, Ekso GT, Indego, and the ExoAtlet-II. The FDA-approved indications for these devices include use by individuals with hemi- and paraplegia due to spinal cord injuries or stroke when accompanied by a specially trained caregiver. They may also be used in rehabilitation institutions. None of these types of devices are intended for sports or climbing stairs. For some of these devices, candidates must retain upper-limb strength and mobility to manage stabilizing crutches.
| Coding |
The following codes for treatments and procedures applicable to this document are included below for informational purposes. Inclusion or exclusion of a procedure, diagnosis or device code(s) does not constitute or imply member coverage or provider reimbursement policy. Please refer to the member's contract benefits in effect at the time of service to determine coverage or non-coverage of these services as it applies to an individual member.
When services are Investigational and Not Medically Necessary:
| HCPCS |
|
| K1007 | Bilateral hip, knee, ankle, foot device, powered, includes pelvic component, single or double upright(s), knee joints any type, with or without ankle joints any type, includes all components and accessories, motors, microprocessors, sensors |
| L2999 | Lower extremity orthoses, not otherwise specified [when specified as a powered robotic lower body exoskeleton device] |
|
|
|
| ICD-10 Diagnosis |
|
|
| All diagnoses |
| References |
Peer Reviewed Publications:
Government Agency, Medical Society, and Other Authoritative Publications:
| Index |
Atalante
Ekso
Ekso GT
ExoAtlet
Hybrid Assistive Limb “HAL” for Medical Use-Lower Limb
Indego
KEEOGO
MyoSuit
ReStore ExoSuit
ReWalk
Rex Rehab
Welwalk
Trexo
VariLeg
The use of specific product names is illustrative only. It is not intended to be a recommendation of one product over another, and is not intended to represent a complete listing of all products available.
| Document History |
| Status | Date | Action |
| Reviewed | 05/09/2024 | Medical Policy & Technology Assessment Committee (MPTAC) review. Updated Rationale, Background, References, and Index sections. |
| Reviewed | 05/11/2023 | MPTAC review. Updated Rationale, Background, References, and Index sections. |
| Reviewed | 05/12/2022 | MPTAC review. Updated Rationale, Background, References, and Index sections. |
| Reviewed | 05/13/2021 | MPTAC review. Updated Rationale and References sections. |
|
| 10/01/2020 | Updated Coding section with 10/01/2020 HCPCS changes; added K1007. |
| Reviewed | 05/14/2020 | MPTAC review. Updated Rationale, References, and Index sections. |
| Reviewed | 08/22/2019 | MPTAC review. Updated Rationale and References sections. |
| Reviewed | 09/13/2018 | MPTAC review. Updated Rationale and References sections. |
| Reviewed | 11/02/2017 | MPTAC review. The document header wording updated from “Current Effective Date” to “Publish Date.” Updated Rationale and References sections. |
| Reviewed | 11/03/2016 | MPTAC review. Updated References section. |
| Reviewed | 11/05/2015 | MPTAC review. Updated Rationale and Reference sections. Removed ICD-9 codes from Coding section. |
| New | 11/13/2014 | MPTAC review. Initial document development. |
Federal and State law, as well as contract language, including definitions and specific contract provisions/exclusions, take precedence over Medical Policy and must be considered first in determining eligibility for coverage. The member’s contract benefits in effect on the date that services are rendered must be used. Medical Policy, which addresses medical efficacy, should be considered before utilizing medical opinion in adjudication. Medical technology is constantly evolving, and we reserve the right to review and update Medical Policy periodically.