According to the American Heart Association, stroke is a major cause of long-term disability in the United States. Over 50 % of patients with upper limb paresis resulting from stroke face long-term impaired arm function and ensuing disability in daily life (Verbunt et al, 2008). Reducing the degree of permanent disability is the goal of post-stroke neurorehabilitation programs. Therapy that incorporates active assistance of motor tasks is a common technique in upper extremity rehabilitation after stroke. It has been proposed that this type of therapy may be facilitated by robotic devices. Exoskeletal or "wearable" robots are intended to provide therapeutic exercise or to function as powered orthoses to help compensate for chronic weakness.

Myoelectric orthotic devices are intended for use in persons with upper limb deficiencies. According to the manufacturers of these devices, they enable individuals who have been afflicted by a stroke or other neuromuscular conditions to self-initiate movement of a partially paralyzed arm using their own muscle signals. They explain that, when the user tries to bend the affected limb, sensors in the brace detect the muscle signal, which activates the motor to move the arm in the desired direction. Examples of this brace include, but may not be limited to, the MyoPro myoelectric limb orthosis and the Myomo e100.

In a randomized controlled pilot study, Kahn and associates (2006a) investigated the effects of robotically administered active-assistive exercise compared to free reaching voluntary exercise in improving arm movement ability after chronic stroke. A total of 19 individuals at least 1 year post-stroke were randomized into 1 of 2 groups. One group performed 24 sessions of active-assistive reaching exercise with a simple robotic device, while a 2nd group performed a task-matched amount of un-assisted reaching. The main outcome measures were range and speed of supported arm movement, range, straightness and smoothness of un-supported reaching, and the Rancho Los Amigos Functional Test of Upper Extremity Function. There were significant improvements with training for range of motion (ROM) and velocity of supported reaching, straightness of un-supported reaching, and functional movement ability. These improvements were not significantly different between the 2 groups. The group that performed un-assisted reaching exercise improved the smoothness of their reaching movements more than the robot-assisted group. The authors concluded that improvements with both forms of exercise confirmed that repeated, task-related voluntary activation of the damaged motor system is a key stimulus to motor recovery following stroke and that robotically assisting in reaching improved arm movement ability, although it did not provide any detectable, additional value beyond the movement practice that occurred concurrently with it. The authors stated that the inability to detect any additional value of robot-assisted reaching may have been due to the limited sample size of the study and the specific diagnoses of the participants, or the inclusion of only individuals with chronic stroke.

The Myomo e100 robotic arm brace (Myomo, Inc., Boston, MA) is an electromechanically powered elbow brace that received 510(k) marketing clearance in April 2007. It is intended to facilitate arm movement and maintain or increase ROM of the arm for stroke patients undergoing rehabilitation. The device contains a computerized system that detects changes in the patient's resting muscle potential through electromyographic (EMG) sensors placed on the skin. The computerized system translates these electrical signals into the desired motion and then responds in real time with a power boost from the device's motorized elbow brace to assist the desired motion. The device would be included as part of prescribed physical therapy to enable stroke patients to exercise that would otherwise be unable to independently do so. The amount of assistance provided by the Myomo e100 can vary as the patient fatigues or gains strength. The objective of the EMG-triggered action is to engage and reinforce both neurological and motor pathways to help the patient relearn how to move affected muscles. The Myomo e100 can be configured for either the left or right arm, and to detect electrical activity from either the biceps or triceps. Unlike some earlier robotic therapy devices that were stationary, the Myomo e100 is a wearable device. It is designed to act as an aid for exercise training and may be developed for use as an actively powered orthosis to assist with difficult motions.

Stein et al (2007) reported the results of the Myomo e100 device on 6 stroke patients with severe chronic hemiparesis. Each patient used the device for a total of 18 hours of exercise therapy (2 to 3 hours per week) for a period of 6 weeks. The average age of the patients was 53 years, and the average time since their stroke was 3.67 years. A 7th patient did not have sufficient EMG signals to control the device. Patients performed exercises including a defined set of functional tasks (moving blocks or turning light switches on or off) with the robotic brace. They were able to control the motorized brace to assist in these motions. Assessment by both the Fugl-Meyer scale and the modified Ashworth scale (a measure of muscle spasticity) showed improvement in upper extremity motor function. The authors concluded that the EMG-controlled powered elbow orthoses show promise as a new modality for assisted exercise training after stroke. They stated that further studies are needed to confirm these preliminary results.

Chang et al (2007) analyzed the results of a training program (40 mins/session, 3 sessions/week for 8 weeks) consisting of 10 mins of conventional rehabilitation and 30 mins of robot-aided, bilateral force-induced, isokinetic arm movement training to improve paretic upper-limb motor function. The post-test and retention test in arm motor function significantly improved in terms of grip (p = 0.009), push (p = 0.001), and pull (p = 0.001) strengths, and Fugl-Meyer Assessment upper-limb scale (p < 0.001). Reaching kinematics significantly improved in terms of movement time (p = 0.015), peak velocity (p = 0.035), percentage of time to peak velocity (p = 0.004), and normalized jerk score (p = 0.008). Improvement in reaching ability was not sustained in the retention test. The authors concluded that these preliminary results showed that conventional rehabilitation combined with robot-aided, bilateral force-induced, isokinetic arm training might enhance the recovery of strength and motor control ability in the paretic upper limb of patients with chronic stroke.

A Cochrane systematic review (Mehrholz et al, 2008) evaluated the evidence of the effectiveness of electromechanical and robot-assisted training for improving arm function after stroke. Randomized controlled trials comparing electromechanical and robot-assisted arm training for recovery of arm function with other rehabilitation interventions or no treatment for patients after stroke were selected for review. Eleven trials (328 participants) met the inclusion criteria. The authors reported that electromechanical and robot-assisted arm training did not improve ADLs but arm motor function and arm motor strength improved. These results must be interpreted with caution, however, because there were variations between the trials in the duration, amount of training and type of treatment, and in the patient characteristics. The authors concluded, "It is, therefore, not clear if such devices should be applied in routine rehabilitation, or when and how often they should be used." A Cochrane systematic evidence review also found that there is insufficient evidence for the use of robotic-assisted gait training after spinal cord injury (SCI) (Mehrholz et al, 2008). Two review authors independently selected trials for inclusion, assessed trial quality and extracted the data. The primary outcomes were the speed of walking and walking capacity at follow-up. A total of 4 randomized controlled trials (RCTs) involving 222 patients were included in this review. Overall, the results were inconclusive. There was no statistically significant effect of locomotor training on walking function after SCI comparing body-weight supported treadmill training with or without functional electrical stimulation or robotic-assisted locomotor training. The authors concluded that there is insufficient evidence from RCTs to conclude that any one locomotor training strategy improves walking function more than another for people with SCI. Research in the form of large RCTs is needed to address specific questions about the type of locomotor training which might be most effective in improving walking function of people with SCI.

In a RCT, Hornby et al (2008) examined the effects of robotic-assisted versus therapist-assisted locomotor training (LT). Both groups received 12 LT sessions for 30 mins at similar speeds, with guided symmetrical locomotor assistance using a robotic orthosis versus manual facilitation from a single therapist using an assist-as-needed paradigm. Outcome measures included gait speed and symmetry, and clinical measures of activity and participation. Greater improvements in speed and single limb stance time on the impaired leg were observed in subjects who received therapist-assisted LT, with larger speed improvements in those with less severe gait deficits. Perceived rating of the effects of physical limitations on quality of life improved only in subjects with severe gait deficits who received therapist-assisted LT. The authors concluded that therapist-assisted LT facilitates greater improvements in walking ability in ambulatory stroke survivors as compared to a similar dosage of robotic-assisted LT.

In a multi-center, randomized clinical trial, Hilder and colleagues (2009) compared the effectiveness of robotic-assisted gait training with the Lokomat to conventional gait training in individuals with subacute stroke. A total of 63 participants less than 6 months post-stroke with an initial walking speed between 0.1 to 0.6 m/s completed the study. All participants received 24 1-hr sessions of either Lokomat or conventional gait training. Outcome measures were evaluated prior to training, after 12 and 24 sessions, and at a 3-month follow-up examination. Self-selected overground walking speed and distance walked in 6 mins were the primary outcome measures, whereas secondary outcome measures included balance, mobility and function, cadence and symmetry, level of disability, and quality of life measures. Participants who received conventional gait training experienced significantly greater gains in walking speed (p = 0.002) and distance (p = 0.03) than those trained on the Lokomat. These differences were maintained at the 3-month follow-up evaluation. Secondary measures were not different between the 2 groups, although a 2-fold greater improvement in cadence was observed in the conventional versus Lokomat group. The authors concluded that for subacute stroke patients with moderate-to-severe gait impairments, the diversity of conventional gait training interventions appears to be more effective than robotic-assisted gait training for facilitating returns in walking ability.

In a randomized clinical trial, Lewek et al (2009) determined whether LT with physical assistance as needed was superior to guided, symmetrical, robotic-assisted LT for improving kinematic coordination during walking post-stroke. A total of 19 people with chronic stroke (greater than 6 months' duration) participating in a RCT comparing therapist-assisted versus robotic-assisted LT were recruited. Prior to and following 4 weeks of LT, gait analysis was performed at each participant's self-selected speed during overground walking. Kinematic coordination was defined as the consistency of intra-limb hip and knee angular trajectories over repeated gait cycles and was compared before and after treatment for each group. Locomotor training with therapist assistance resulted in significant improvements in the consistency of intra-limb movements of the impaired limb. Providing consistent kinematic assistance during robotic-assisted LT did not result in improvements in intra-limb consistency. Only minimal changes in discrete kinematics were observed in either group. The authors concluded that coordination of intra-limb kinematics appears to improve in response to LT with therapist assistance as needed. Fixed assistance, as provided by this form of robotic guidance during LT, however, did not alter intra-limb coordination.

In a preliminary study, Kutner et al (2010) examined changes in patient-reported, health-related quality of life associated with robotic-assisted therapy combined with reduced therapist-supervised training in patients with subacute stroke. A total of 17 individuals who were 3 to 9 months post-stroke participated in this study. Sixty hours of therapist-supervised repetitive task practice (RTP) was compared with 30 hours of RTP combined with 30 hours of robotic-assisted therapy. Participants completed the Stroke Impact Scale (SIS) at baseline, immediately post-intervention, and 2 months post-intervention. Change in SIS score domains was assessed in a mixed model analysis. The combined therapy group had a greater increase in rating of mood from pre-intervention to post-intervention, and the RTP-only group had a greater increase in rating of social participation from pre-intervention to follow-up. Both groups had statistically significant improvement in activities of daily living and instrumental activities of daily living scores from pre-intervention to post-intervention. Both groups reported significant improvement in hand function post-intervention and at follow-up, and the magnitude of these changes suggested clinical significance. The combined therapy group had significant improvements in stroke recovery rating post-intervention and at follow-up, which appeared clinically significant; this also was true for stroke recovery rating from pre-intervention to follow-up in the RTP-only group. The major limitation of this study was that outcome of 30 hours of RTP in the absence of robotic-assisted therapy remain unknown. The authors concluded that robotic-assisted therapy may be an effective alternative or adjunct to the delivery of intensive task practice interventions to enhance hand function recovery in patients with stroke. These prelininary findings need to be validated by further investigation.

In a case-series study, Meyer-Heim and colleagues (2009) measured functional gait improvements of robotic-assisted locomotion training in children with cerebral palsy (CP). A total of 22 children (mean age of 8.6 years, range of 4.6 to 11.7) with CP and a gross motor function classification system level II to IV were enrolled in this study. Subjects received 3 to 5 sessions of 45 to 60 mins/week during a 3- to 5-week period of driven gait orthosis training. Main outcome measures included 10-meter walk test (10MWT), 6-min walk test (6MinWT), gross motor function measure (GMFM-66) dimension D (standing) and dimension E (walking), and functional ambulation categories (FAC). The mean (SD) maximum gait speed (0.78 (0.57) to 0.91 (0.61) m/s; p < 0.01) as well as the mean (SD) dimension D of the GMFM-66 (40.3 % (31.3 %) to 46.6 % (28.7 %); p < 0.05) improved significantly after the intervention period. The mean (SD) 6MinWT (176.3 (141.8) to 199.5 (157.7) m), the mean FAC (2.6 (1.7) to 3.0 (1.6)) and the mean (SD) dimension E of the GMFM-66 (29.5 % (30.3 %) to 31.6 % (29.2 %)) also showed an increase, but did not reach a statistically significant level. These authors concluded that these findings suggested that children with CP benefit from robotic-assisted gait training in improving functional gait parameters. These findings need to be validated by well-deisgned studies.

In a RCT, Druzbicki et al (2013) evaluated gait in children with spastic diplegic CP (n = 52) rehabilitated with the use of Lokomat active orthosis. Temporo-spatial parameters of gait and selected kinematic parameters were assessed. Childrem from the study group used active orthosis in addition to following a program of individual exercises. Children in the control group participated only in individual exercises. The difference between the initial and control examinations was statistically insignificant. After the program was finished, there was a slight improvement in walking speed in both groups. Improvement in the mean walking speed was not significantly different between the groups (p = 0.5905). Range of motion decreased slightly in both groups, and the difference between mean amounts of change was not significant (p = 0.8676). There was significant improvement in maximal range of flexion in the hip joint (p = 0.0065) in the study. It was shown that with a decrease in the mean value of adduction in hip joint, the mean walking speed increased (r = -0.53, p = 0.0011). The authors concluded that this study had several limitations, thus, these results should be regarded as preliminary. Moreover, they stated that further research consistent with the above indications is needed to investigate the impact of this new treatment option in patients with CP.

Swinnen et al (2010) evaluated the quality of current evidence as to the effectiveness of RAGT in patients with SCI, focusing on walking ability and performance. A search was conducted in MEDLINE, Web of Knowledge, Cochrane Library, Physiotherapy Evidence Database (PEDro) and Digital Academic Repositories (DAREnet) (1990 to 2009). Key words included "spinal cord injury", "(robot-assisted) gait rehabilitation" and "driven gait orthosis". Articles were included when complete and incomplete adult spinal cord injured patients participated in RAGT intervention studies. The methodological quality was rated independently by 2 researchers using "van Tulder criteria list" and "evaluation of quality of an intervention study". Descriptive analyses were performed using the Population Intervention Comparison Outcome (PICO) method. Two RCTs (mean quality score: 11.5/19) and 4 pre-experimental trials (mean quality score: 24.25 (standard deviation; SD 0.28/48) involving 43 patients with incomplete, acute or chronic lesions between C3 and L1 were analysed. Five studies used the Lokomat and 1 used the LokoHelp. Although some improvements were reported related to body functions and activities, there is insufficient evidence to draw firm conclusions, due to small samples sizes, methodological flaws and heterogeneity of training procedures. The authors concluded that there is currently no evidence that RAGT improves walking function more than other locomotor training strategies. They stated that well-designed RCTs are needed.

The Veterans Health Administration and the Department of Defense's clinical practice guideline for the management of stroke rehabilitation (2010) stated that "[t]here is no sufficient evidence supporting use of robotic devices during gait training in patients post stroke" (regarding gait training strategies for lower extremities).

Schwartz et al (2011) evaluated the effect of an additive RAGT using the Lokomat system on the neurological and functional outcomes of patients with subacute SCI. A total of 28 subacute SCI patients were treated by RAGT, 2 to 3 times a week, 30 to 45 mins every treatment, concomitantly with regular physiotherapy. As control, for each patient, these investigators matched a comparable patient treated in the same department in previous years, according to age, severity of injury, level of injury and cause. The main outcomes were: the AIS (American Spinal Injury Association impairment scale) the spinal cord independence measurement (SCIM) score, the walking index for SCI II (WISCI II) and functional ambulation category scale (FAC). At the end of rehabilitation, both groups showed a significant improvement in both the FAC score and the WISCI score (p < 0.01) without differences between the groups. Functional abilities, according to the SCIM score, were also improved, with a significant interaction effect; the RAGT patients improve by 30 +/- 20 points, which was significantly greater gain as compared with the controls, 21 +/- 14 points (p = 0.05). This improvement was mainly due to the change in the SCIM motor subscales. The authors concluded that RAGT is an important additional treatment to improve the functional outcome of subacute SCI patients. They stated that larger, controlled studies are still needed to determine the optimal timing and protocol design for the maximal efficacy of RAGT in SCI patients.

In an open-label pilot study, Stein et al (2011) tested a new robotic device for hand rehabilitation in stroke survivors. A total of 12 individuals with chronic moderate hemiparesis after stroke were enrolled in this study. Participants underwent a 6-week training program using a hand robotic device. They received a total of 18 hours of robotic therapy. Improvements were found in multiple measures of motor performance, including the Upper Extremity Fugl-Meyer test, the Motor Activity Log, the Manual Ability Measure-36, and the Jebsen Hand Function Test. All subjects tolerated the treatment well and no complications were observed. The authors concluded that robotic therapy for hand paresis after stroke is safe and feasible, and further studies of efficacy are justified by these preliminary results.

Norouzi-Gheidari et al (2012) systematically reviewed and analyzed the literature to find RCTs that employed robotic devices in upper-limb rehabilitation of people with stroke. Out of 574 studies, 12 matching the selection criteria were found. The Fugl-Meyer, Functional Independence Measure, Motor Power Scale, and Motor Status Scale outcome measures from the selected RCTs were pooled together, and the corresponding effect sizes were estimated. These researchers found that when the duration/intensity of conventional therapy (CT) was matched with that of the robot-assisted therapy (RT), no difference exists between the intensive CT and RT groups in terms of motor recovery, ADL, strength, and motor control. However, depending on the stage of recovery, extra sessions of RT in addition to regular CT are more beneficial than regular CT alone in motor recovery of the hemiparetic shoulder and elbow of patients with stroke; gains are similar to those that have been observed in intensive CT.

In a Cochrane review, Mehrholz et al (2012) evaluated the effectiveness of electromechanical and robot-assisted arm training for improving generic ADL, arm function, and arm muscle strength in patients after stroke. These investigators also assessed the acceptability and safety of the therapy. They searched the Cochrane Stroke Group's Trials Register (last searched July 2011), the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2011, Issue 7), MEDLINE (1950 to July 2011), EMBASE (1980 to July 2011), CINAHL (1982 to July 2011), AMED (1985 to July 2011), SPORTDiscus (1949 to July 2011), PEDro (searched August 2011), COMPENDEX (1972 to July 2011), and INSPEC (1969 to July 2011). They also hand-searched relevant conference proceedings, searched trials and research registers, checked reference lists, and contacted trialists, experts and researchers in the field, as well as manufacturers of commercial devices. Randomized controlled trials comparing electromechanical and robot-assisted arm training for recovery of arm function with other rehabilitation or placebo interventions, or no treatment, for patients after stroke were selected for analysis. Two review authors independently selected trials for inclusion, assessed trial quality, and extracted data. They contacted trialists for additional information. They analysed the results as standardized mean differences (SMDs) for continuous variables and risk differences (RDs) for dichotomous variables. These researchers included 19 trials (involving 666 participants) in this updated review. Electromechanical and robot-assisted arm training did improve activities of daily living (SMD 0.43, 95 % confidence interval (CI): 0.11 to 0.75, p = 0.009, I(2) = 67 %) as well as arm function (SMD 0.45, 9 5% CI: 0.20 to 0.69, p = 0.0004, I(2) = 45 %), but arm muscle strength did not improve (SMD 0.48, 95 % CI: -0.06 to 1.03, p = 0.08, I(2) = 79 %). Electromechanical and robot-assisted arm training did not increase the risk of patients to drop-out (RD 0.00, 95 % CI: -0.04 to 0.04, p = 0.82, I(2) = 0.0 %), and adverse events were rare. The authors concluded that patients who receive electromechanical and robot-assisted arm training after stroke are more likely to improve their generic ADL. Paretic arm function may also improve, but not arm muscle strength. However, the authors stated that results must be interpreted with caution because there were variations between the trials in the duration and amount of training, type of treatment, and in the patient characteristics.

Hussain et al (2011) stated that the rehabilitation engineering community is working towards the development of robotic devices that can assist during gait training of patients suffering from neurologic injuries such as stroke and SCI. The field of robot-assisted treadmill training has rapidly evolved during the last decade. The robotic devices can provide repetitive, systematic and prolonged gait training sessions. These researchers presented a review of the treadmill-based robotic gait training devices. An overview of design configurations and actuation methods used for these devices was provided. Training strategies designed to actively involve the patient in robot-assisted treadmill training were studied. These training strategies assist the patient according to the level of disability and type of neurologic injury. The authors stated that although the effectiveness of these training strategies is not clinically proven, adaptive strategies may result in substantial improvements.

Alcobendas-Maestro et al (2012) compared a walking re-education program using Lokomat with conventional over-ground training among individuals with incomplete SCI of both traumatic and non-traumatic etiology. A total of 80 participants from 3 to 6 months after onset admitted to 1 site for rehabilitation were included in a single-blind RCT of 2 parallel groups, with blind evaluation by independent observers. Patients received 40 walking re-education sessions of equal time using a Lokomat program with over-ground practice or over-ground mobility therapy alone. Primary measurements of outcome were walking speed and the WISCI II; secondary outcomes were the 6-min walk test, locomotor section of the functional independence measure, lower extremity motor score (LEMS), Ashworth Scale, and visual analog scale for pain. No significant differences were found at entry between treatment groups. Walking speed for Lokomat (0.4 m/s [0.6 to 0.2]) and over-ground therapy (0.3 m/s [0.5 to 0.2]) groups did not differ. The WISCI II for the Lokomat group (16 [8.5 to 19]) was better than for over-ground therapy (9 [8 to 16]). The 6-min walk test and LEMS displayed significant differences in favor of Lokomat therapy but were not corrected for multiple comparisons. The authors concluded that robotic-assisted training was equivalent to conventional walk training in patients with a variety of non-progressive spinal cord pathologies for walking speed, but the need for orthotics and assistive devices was reduced, perhaps because of greater leg strength in the robotic group.

Scope of Policy

This Clinical Policy Bulletin addresses robotic-assisted rehabilitation of the extremities.

Experimental and Investigational