Abstract
BACKGROUND:
The gait recovery is a realist goal in the rehabilitation of almost Stroke patients. Over the last years, the introduction of robotic technologies in gait rehabilitation of stroke patients has had a greatest interest.
OBJECTIVE:
The aim of this study was to evaluate efficacy of Robotic Gait Training (RGT) in chronic stroke patients.
METHODS:
Fourteen chronic stroke patients were divided into two groups. Six patients received RGT, eight patients received traditional gait rehabilitation. Patients were assessed with clinical scales, as well as with gait analysis, at the beginning and at the end of the treatment.
RESULTS:
Significant changes in some clinical scales for both the groups were detected. In the robotic group, patients showed higher percentage changes in the MRC scale (p = 0.020), in the 6MWT (p = 0.043) and in the Ashworth scale (hip: p = 0.008; knee: p = 0.043; ankle: p = 0.043) when compared with the traditional group. With respect to the gait analysis, we did not found any difference neither in the within–group analysis, nor in the between–group analysis.
CONCLUSIONS:
Both rehabilitation treatments do not change the compensatory strategies in chronic patients but the RGT offers to the patients a more intensive and controlled gait training increasing the gait endurance and decreasing spasticity in the lower limb.
Introduction
Stroke, which is the leading cause of disability in the world, has a significant impact on individuals, their families and finances. It is also the third leading cause of death after cardiovascular disease and cancer, accounting for 10% – 12% of all deaths every year, and is the leading cause of disability: 3 months after stroke, 20% of individuals remain wheelchair bound and 70% walk at a reduced velocity (Palmieri et al., 2016). Chronic stroke patients suffer from asymmetric posture, with reduced weight bearing on the paretic limb, and balance and gait dysfunctions. Indeed, such patients display impairments in joint mobility and stability, in muscle power, in tone and reflexes, in muscle endurance, in movement control and in gait pattern functions. These impairments lead to difficulties in transferring, maintaining body position, mobility, balance and walking (Swinnen et al., 2014).
A number of treatments have been proposed in recent years for the re-learning of motor skills of the lower limbs: the recovery of a more fluid, safe and correct gait is a prerequisite to enable the patient to become independent in daily living activities. In this regard, the introduction of robotic technologies in the gait rehabilitation of stroke patients has attracted great interest. Robotic devices have several advantages: they require a smaller workforce, they allow more enduring and intensive treatment, they allow the patient’s disability and its development to be assessed objectively and quantitatively, and they provide multi-sensory stimuli.
Over the last ten years robotic assisted devices, such as the exoskeleton and robot end-effector, have been used for gait training purposes in neurological disorders, including stroke, spinal cord injury and multiple sclerosis, yielding good results in gait recovery (Kelley et al., 2014). Although few studies have been conducted to assess the effects of exoskeleton or end-effector robot-assisted training compared to gait conventional rehabilitation in stroke patients, the preliminary results are interesting (Taveggia et al., 2016). A small number of these studies investigated the differences between conventional and robotic training treatment, the latter being based on the Lokomat system, in chronic stroke patients (Dundar et al., 2008).
The G-EO system device, which is another robotic treatment based on an end-effector device, has been used for gait rehabilitation purposes in subacute stroke patients (Pohl et al., 2007) or with other movement disorders such as Cerebral Palsy (Smania et al., 2011), Parkinson’s disease (Sale et al., 2013) and progressive supranuclear palsy (Sale et al., 2014), though never in chronic stroke patients.
Only one previous study (Hesse et al., 2010) has compared lower limb muscle activation patterns in hemiparetic subjects during real floor walking and stair climbing with those of patients treated by means of corresponding simulated conditions on the GEO System. That study demonstrated that patients with severe disabilities regained their walking and stair climbing ability. Indeed, muscle activation patterns were comparable during the real and simulated situations in both tasks. Both the real and simulated floor walking conditions yielded a prolonged thigh muscle activation on the machine in all subjects, while the stair climbing task yielded a more phasic and timely shank muscle activation pattern in the simulated than in the real condition. That study thus suggests that the GEO System is an interesting option in gait rehabilitation after stroke.
The aim of this pilot study was to compare traditional gait rehabilitation with Robotic-Assisted Gait Training in chronic stroke patients using the GEO System to analyze any short-term changes in the two groups using clinical measures and gait analysis parameters.
Methods
Sample
We enrolled 14 patients with chronic stroke, comprising 9 males and 5 females, aged between 23 and 86 years (mean age 63±17.79 years) (Table 1). Patients were recruited at the IRCSS San Raffaele Pisana of Rome and at the Don Carlo Gnocchi Foundation Onlus of Rome.
Sample characteristics
Sample characteristics
Inclusion criteria were chronic first stroke (>6 month after onset), age ≥18 years, ability to walk unassisted or with little assistance.
Exclusion criteria included significant neurological or orthopedic disorders other than stroke, an inability to understand the instructions required for the study, presence of cardiac pathologies, anxiety or psychosis that might interfere with the use of the equipment or testing. We also excluded subjects with cognitive impairment or other diseases liable to cause motor gait impairment.
The patients were divided into 2 groups: one group treated by means of a robotic device combined with conventional physiotherapy (robotic group, RG) and another group treated by means of a traditional gait rehabilitation program combined with conventional physiotherapy (conventional group, CG). The study followed the tenets of the Declaration of Helsinki, and informed consent was obtained from each patient after the nature of the procedure had been explained to them.
Patients in both the RG and CG were evaluated, by means of a clinical and instrumental evaluation, before treatment (T0) and at the end of the rehabilitation program (T1).
The clinical evaluation included several scales for the assessment of motor performance, balance, muscle tone and recruitment: Motricity Index, Ashworth Scale, Medical Research Council (MRC), Timed Up and Go Test, Six-Minute Walk Test (6MWT), Ten-Meter Walk Test, Functional Ambulation Categories (FAC), Walking Handicap Scale (W.H.S), Tinetti Scale, Fugl-Meyer Assessment (FMA) for the lower limbs and Trunk Control Test (Franceschini et al., 2015). Biomechanical data were collected using 8-camera SMARTD motion capture system (BTS, Milano, Italy) sampling at 200 Hz. The Davis marker set (Davis et al., 1991) which includes 22 retro-reflective markers was adopted, and anthropometric data were collected for each subject (Winter, 2009). Each patient was asked to perform ten linear walking trials, barefoot and at a self-selected speed, straight ahead along a level surface that was approximately 6 meters long. Before formal measurements were started, practice sessions were performed to familiarize the participants with the procedure. We computed the average value of the parameters selected and the average pattern of the biomechanical gait variables across five trials for each patient. Owing to the asymmetric nature of the pathology, we analyzed the affected and the unaffected sides separately. Three-dimensional marker trajectories were tracked using a frame-by-frame tracking system (Smart Tracker-BTS, Milan, Italy). Data were processed using 3D reconstruction software (SMARTAnalyzer, BTS, Milan, Italy). The following spatio-temporal gait parameters were calculated: stance, swing and double support phase durations, cycle time, step length, stride length, step width, cadence and gait speed.
Therapeutic intervention
The RG included 6 patients (recruited from the San Raffaele center) who received Robotic-Assisted Gait Training using an end-effector G-EO system device, 3 times a week, in 20 sessions. The CG included 8 patients (recruited from the Don Gnocchi center) treated by means of a traditional gait rehabilitation program, 3 times a week, in 20 sessions. The rehabilitation program in both groups was combined with conventional physiotherapy twice a week over 45 days.
Robot therapy
Each patients of RG was asked to perform 20 session (3 times a week for 6.5 weeks) of robot assisted gait training, using the commercially available end-effector G-EO system device (Reha Technology AG; Olten, Switzerland). The G-EO robot is characterized by an end-effector device with a body weight support (BWS) and 2 footplates placed on a double crank and a rocker gear system, with 3 Degrees of Freedom each, which allows the step length and height to be controlled. The trajectories of the footplates and the vertical and horizontal movements of the centre of mass were fully programmable, thus allowing not only the simulated floor walking to be simulated repetitively, but also the climbing up and down of stairs. During the training, the patients were asked to walk, at a varying speed, for 45 minutes, with a partial BWS. All the participants started with 30–40% of BWS and an initial speed of 1.5 km/h; thereafter, speed was increased to a maximum of between 2.2 and 2.5 km/h and the initial BWS was reduced15. The therapist stood in front of the patient during the treatment session to provide any help if required. The data for each parameter were collected for every session, and the steps taken during the simulated walking were converted into the distance covered according to the previously selected step length (Hesse et al., 2012). Over 45 minutes, the patient simulated a minimum of 300 steps and 50 climbs up and down the stairs; patients could rest during the session, though they were required to walk continuously for a minimum of 5 minutes and climb continuously up and down the stairs for a minimum 3 minutes during each session.
Traditional gait rehabilitation
CG patients were treated by means of a traditional gait rehabilitation program 3 times a week, in 20 1-hour sessions. The treatment included: muscle strengthening exercises and stretching of the lower limb, and static and dynamic exercises for the recovery of balance in the supine and standing positions using assistive devices; training gait exercises with parallel bars or in open spaces performed both with and without assistive devices; training to climb up and down stairs; exercises to improve proprioception in the supine, sitting and standing positions, using a proprioceptive footboard; exercises to improve trunk control.
Statistical analysis
The statistical analysis was performed using the SPSS 21 package (IBM, Armonk, NY). Owing to the small sample size, a non-parametric analysis was performed. To evaluate the effects of the two treatments separately for each outcome measure, we compared the evaluation at T0 with the evaluation at T1 by means of a Wilcoxon Signed Rank test.
Moreover, to compare the effects of traditional treatment with those of robotic treatment, we compared the percentage changes (% Δ), defined as: % Δ= (T1 score – T0 score)/T0 score, by means of a Mann Whitney U test, for each of the outcome measures except the Ashworth scale and the MRC scale. With regard to the Ashworth scale and the MRC scale, we compared the difference, i.e. T1 score – T0 score, because the value at T0 according to these scales was in some cases equal to 0, and normalization was not thus possible. Statistical significance for each test was set at 0.05.
Results
Clinical scales
The within-group analysis revealed statistically significant changes in some clinical scales for both the CG and the RG (Table 2). More specifically, in the CG, we observed a decrease in the Timed Up and Go (p = 0.043) as well as an increase in the Tinetti walking scale (p = 0.018) and the Fugl Meyer scale (p = 0.012). In the RG, we observed an increase in the MRC scale (p = 0.042), the Tinetti Walking Scale (p = 0.034) and the Fugl Meyer scale (p = 0.027) as well as a reduction in the Ashworth scale at the hip joint (p = 0.034). All these changes point to an improvement in the patients’ performance. With regard to the between-group analysis, we found that the percent changes were higher (i.e. a greater improvement) in the RG than in the CG in the MRC scale (p = 0.020), in the Six-Minute Walk Test (p = 0.043) and in all the of the Ashworth scale subscores (hip: p = 0.008; knee: p = 0.043; ankle: p = 0.043).
Clinical scales values (medians and interquartile ranges) obtained at T0 and T1, for both the CG and RG, together with the results of the statistical analysis (within-group analysis: T0evaluation vs T1evaluation, for the two groups separately; between-group analysis: comparison of the percentage changes – %Δconventional vs %Δrobotic – obtained in the two groups)
Clinical scales values (medians and interquartile ranges) obtained at T0 and T1, for both the CG and RG, together with the results of the statistical analysis (within-group analysis: T0evaluation vs T1evaluation, for the two groups separately; between-group analysis: comparison of the percentage changes – %Δconventional vs %Δrobotic – obtained in the two groups)
With regard to the gait analysis, no difference was detected in either the within-group analysis or the between-group analysis (See Table 3).
Gait analysis variable values (medians and interquartile ranges) obtained at T0 and T1, from both the CG and RG, together with the results of the statistical analysis (within-group analysis: T0evaluation vs T1evaluation, for the two groups separately; between-group analysis: comparison of the percentage changes obtained in the two groups)
Gait analysis variable values (medians and interquartile ranges) obtained at T0 and T1, from both the CG and RG, together with the results of the statistical analysis (within-group analysis: T0evaluation vs T1evaluation, for the two groups separately; between-group analysis: comparison of the percentage changes obtained in the two groups)
Our study investigated the effects on gait of two rehabilitation treatments (traditional gait rehabilitation and Robotic-Assisted Gait Training) in chronic stroke patients.
The robotic rehabilitation G-EO System device in this study was used in stroke patients, whereas in previous studies it had been used for gait rehabilitation in other movement disorders, particularly in Parkinson’s disease and progressive supranuclear palsy (Sale et al., 2014).
Previous studies have investigated robotic rehabilitation in patients with chronic hemiplegia using another device, i.e. the Lokomat system. When Uçar et al. (Uçar et al., 2014) studied the efficacy of this robotic-assisted gait device, they found that it provides innovative possibilities for gait training rehabilitation in chronic stroke patients, allowing higher intensity training for longer periods of time than traditional gait rehabilitation. Moreover, the authors of a Cochrane review showed that stroke patients who receive rehabilitation treatment based on robotic devices combined with conventional physiotherapy are more likely to regain the ability to walk independently than patients who receive gait training without such devices. It was also shown that the greatest improvement occurred during the first 3 months after stroke and in patients who were unable to walk (Mehrholz et al., 2013).
In our study, both types of rehabilitation treatment (T1 evaluation) yielded significantly positive results in spite of the small number of cases and the heterogeneity of the sample. However, the G-EO System-based treatment was found to be more incisive and focused on improving gait performance.
We also performed a clinical evaluation, which revealed an improvement in the 6MWT, the Ashworth Scale and the MRC scale, though not in the spatial-temporal parameters of gait analysis, which were basically unchanged between T0 and T1.
Clinical and instrumental evaluations are two different ways of assessing walking. While gait analysis estimates deambulation over the shortest path, i.e. about 10 meters, for a few seconds, the Six-Minute Walk Test, records the patient’s endurance on a longer distance over time. Indeed, the 10-meter walk test (which, like the gait analysis, estimates deambulation over the shortest path) did not yield any significant differences between T0 and T1 in either group.
Although neither traditional gait rehabilitation nor robotic gait training change the compensatory strategies in chronic stroke patients, they do increase the intensity and endurance of the training. This emerges from the significant difference in the 6MWT between T0 and T1, resulting from the increased endurance provided by the G-EO System device training.
Training based on the robotic device thus offers the patient a more intensive, repetitive and automatic form of exercise that more closely reflects the characteristics of deambulation. In brief, training with the G-EO System improves deambulation, reduces muscle tone and muscle stiffness and increases muscle recruitment. By simulating the step movement repetitively and automatically for 45 minutes, 3 times a week, the device offers the possibility of working on all the joints simultaneously, thereby providing more training than that performed by the therapist in the same amount of time.
Moreover, our results show that spasticity (Ashworth Scale for Hip, Knee and Ankle joints) either improved or remained the same following robotic therapy, as has been reported in previous studies (Kwakkel et al., 2011). Indeed, our statistical analysis shows that the spasticity decreased to a significantly greater extent in the RG than in the CG, thus demonstrating the positive effects of the robot-assisted treatment on chronic stroke patients. This result confirms the findings reported in another study that investigated the upper limbs in chronic stroke patients (Sale et al., 2014), and may open new frontiers in the treatment of spasticity; in particular, it could lead to robotic treatment being associated with other therapeutic approaches in order to manage spasticity in stable chronic stroke patients.
The results at the end of the traditional gait rehabilitation treatment show that this approach did not result in any substantial improvement, with both the clinical scales and the instrumental evaluation yielding comparable results at the start and end of treatment. As there are papers in the literature that suggest that there is a margin for improvement in chronic stroke patients (Mehrholz et al., 2014), it should be borne in mind that our treatment regime was relatively short-lasting (i.e. 20 sessions lasting approx. 45 minutes each) and that the compensatory strategies adopted by chronic stroke patients tend to become consolidated over time.
Lastly, our study highlights the fact that the robotic device we tested reduces, through movement repetition, the time required by the therapist to administer the same exercises by means of other rehabilitation strategies.
This bi-centre study showed that robotic gait training combined with conventional physiotherapy improves functional and motor outcomes to a greater extent than traditional gait rehabilitation in chronic stroke patients. Although neither rehabilitation treatment changes the compensatory strategies in chronic patients, the robotic gait training provides a more intensive and controlled gait training, thereby increasing gait endurance and reducing lower limb spasticity.
Conflict of interest
None to report.