Purpose: Treadmill training with body weight support (TTBWS) for relearning walking ability after brain damage is an approach under current investigation. Efficiency of this method beyond traditional training is lacking evidence, especially in patients needing walking assistance after stroke. The objective of this study was to investigate change in walking and transfer abilities, comparing TTBWS with traditional walking training. Methods: A single-blinded, randomized controlled trial was conducted. Sixty patients referred for multi-disciplinary primary rehabilitation were assigned into one of two intervention groups, one received 30 sessions of TTBWS plus traditional training, the other traditional training alone. Daily training was 1 hr. Outcome measures were Functional Ambulation Categories (FAC), Walking, Functional Independence Measure (FIM); shorter transfer and stairs, 10 m and 6-min walk tests. Results: Substantial improvements in walking and transfer were shown within both groups after 5 and 11 weeks of intervention. Overall no statistical significant differences were found between the groups, but 12 of 17 physical measures tended to show improvements in favour of the treadmill approach.
Introduction
Restoring walking ability is an important goal within rehabili-tation after stroke . This aim motivates the patients through the training process. On the other hand, when walking is not successfully achieved, this is a common target for complaints . Comprehensive efforts have therefore been invested in achieving evidence for effective ways of restoring motor functions , especially walking . Thus, the question about which training approach that should be preferred for improving walking ability is unanswered. Limited evidence regarding the efficiency of different approaches has so far precluded clear recommendations . However, in a recent extensive review high-intensity, repetitive, task specific train-ing was recommended as a principle for effective training in order to promote motor recovery after stroke .
Treadmill training with body weight support (TTBWS) is a training approach under current investigation. This approach which is repetitive and considered task specific, has over the past two decades become a frequently used method for gait training in neurorehabilitation. Based on experiments in the 1980s concerning restoration of walking in spinal cord dam-aged animals , TTBWS was introduced for humans in Canada, USA and Germany in the early 1990s. This method was found promising for restoring walking ability, primarily in spinal cord injured patients , and few years later extended to patients with orthopaedic or neurological condi-tions like stroke .
The neurophysiologic theories underlying TTBWS are based on evidence that neural plasticity of the brain and spinal motor mechanisms are influenced by proprioceptive information elicited from upright, repetitive gait training. TTBWS implies a method within the theoretical framework underpinning neurological physiotherapy where current neurophysiologic, musculoskeletal and motor learn- ing principles are integrated in a systems model of motor control . During treadmill therapy, main elements to obtain motor control are practiced, such as task oriented training and repetition .Therefore an optimistic expecta-tion has been created concerning the outcome of such train- ing on walking ability during rehabilitation of patients after stroke .
Research on the effectiveness of TTBWS was in the first decade dominated by single group studies, while randomized studies followed from about 2000 . A meta-analysis of 11 randomised studies investigated the training effect on walking speed and dependence , and evidence was not provided that TTBWS was a more efficient training method for restoring walking ability than traditional training. How- ever, the results were mainly based on data from patients who could walk independently over ground. Patients in need of personal assistance were given baseline scores of zero on gait characteristics. The authors argued, however, that the patients in future studies should be divided into independent and dependent walkers when examining the effectiveness of treadmill training.
Some evidence indicates that patients who are dependent of assistance for walking, are responding positively to TTBWS , even more than to conventional gait training .These results were in accordance with our clinical experience, that patients with severe stroke tend to attain great progress during a period of TTBWS. However, evidence from clinical trials is based on selected patient samples, implying pos-sible reduced validity for a broader population . Research evidence of the effect of TTBWS on patients with moderate to severe deficits has been limited, and further research was therefore strongly recommended .
The objective of this study was to investigate the effect on walking and transfer of two training modalities, comparing TTBWS with traditional gait training in patients dependent of walking assistance after stroke. As TTBWS might provide better conditions for repetitive training of gait cycles and safe increase of walking speed and distance, we hypothesized that TTBWS and over ground training would provide better out- come than traditional over ground training alone.
Implications for Rehabilitation
- Treadmill training with body weight support (TTBWS) and traditional gait training were found to be equally effective in improving walking and transfer in patients dependent on walking assistance after stroke.
- However, most outcome measures showed a ten-dency of improvement in favour of the treadmill group.
- Both training modalities were systematic, goal-direct- ed and intensive, and the time used for training was the same.
- The choice between the two equally effective training alternatives should be based on the patients’ prefer-ence, and availability of equipment and resources.
Subjects
A sample of 60 patients with stroke, admitted to an in-patient stroke division, and referred for multi-disciplinary rehabilita-tion, were consecutively enrolled after approval from a medi-cal doctor and signing of informed consent. Power calculation was based on a pilot study on TTBWS, conducted at the hos-pital and related to information from comparable ran-domized trials . We estimated a mean change dif-ference in FAC score of 1 between the groups and a standard deviation of 1.25, a power of 0.80 and a two sided significance level of p < 0.05 .
Inclusion criteria were admission for a primary rehabilita-tion stay, mainly <6 months after onset of stroke, use of wheel- chair, dependence of assistance for walking with or without walking aids, medically stable, no neurological or orthopaedic contraindications for walking, and sufficient cognitive capac-ity to understand information and instructions. The patients’ need of assistance should not be beyond one person for short-er transfer and for taking some steps over ground. A blocked randomization procedure was applied to secure even numbers in the two treatment groups, and 60 numbers concealed in envelopes were prepared by an external statistician. Assign-ment to treatment groups was conducted after the baseline evaluation (Test 1) to secure that the participants’ functional level was within the inclusion criteria.
Treatment
Time for daily training (5 days a week) was the same in the two intervention groups, 30 min for walking and 30 min for other functional training, including selective training of the trunk and extremities, balance and transfer, customized to individual deficits and needs. Additional self-training, in- dividually or by the staff, was allowed. The physiotherapists employed at the stroke department trained the patients.
Treadmill therapy
The patients walked on a motorized, raised treadmill, secured by a harness combined with a suspension system releas-ing body weight. A ramp was attached, enabling patients in wheelchair to be driven up on the surface. Handrails were attached in front and on each side of the walking belt, and therapists could be seated helping the patient with walking movements. There was a device to start and stop the moving belt and regulate the speed on the rail. Training characteristics of TTBWS like weight bearing, weight transfer and increase of walking speed were registered in a form.
Method
Design
A randomized single-blinded controlled clinical trial was con-ducted with two intervention groups. The study was approved by the South-East regional committee for research ethics in Norway (ref: 108-04043).
Treadmill group
This group received 30 sessions of TTBWS, plus conventional gait training and other functional training for a period of minimum 10 weeks. TTBWS was conducted daily for the first 4 weeks (20 sessions), and then 1–2 times a week (10 ses- sions) for the remaining 6 weeks. On days without TTBWS, conventional gait training was conducted. Each treadmill session lasted for 30 min, including necessary pauses, but excluding equipment preparation. This training protocol was in line with a research proposal for patients with stroke and recommendations for in-patient treadmill therapy by Wernig . One proposal was to increasingly replace treadmill with overground training according to physical improvements, while TTBWS was often maintained to train endurance and speed. The intervention period of minimum 10 weeks was in accordance with a normal primary rehabilitation stay at study start for patients with moderate to severe functional deficits.
Traditional gait training group
This group conducted intensive gait training (30 min) and functional training (30 min) daily for minimum 10 weeks.
Assessment tools
The patients were usually tested without walking aids or or- thoses. However, when the orthosis was prerequisite for walk-ing safety, it was permitted, using the same at each test. The patients were instructed to walk as fast as they safely could, and received balance assistance according to the rule “as little as possible, but as much as necessary for safety.” Assessment of walking ability was conducted in a silent test corridor, with marks on the floor for each metre until 30 m. The equipments for measuring timed tests were a manual stopwatch and heart rate was measured by a pulse watch. A regular plinth was used for testing transfer, and a stair nearby for testing stair climb- ing. A garden outside the corridor with asphalt paths, grass and slopes was convenient for testing outdoors walking when appropriate. The patients were tested; 1) at baseline, 2) after 4–6 weeks, and 3) after minimum 10 weeks of intervention. The second test was conducted to capture results in case un-planned events precluded fulfilling the intervention, and as TTBWS was performed almost daily during the first period. A pool of eight experienced assessors blinded to group alloca-tion, were involved in testing. Two physiotherapists evaluated each patient’s walking and transfer abilities, and consensus was reached on ordinal scores.
Primary outcome measure
Functional Ambulation Categories (FAC 0-5) has been shown to be valid and reliable ,and was chosen as the pri-mary outcome measure, taking the patients’ dependence of human support for walking into consideration. Each step on this ordinal scale indicates a clinically significant change in walking ability. The FAC was scored after the patients had performed the other tests for walking.
Secondary outcome measures
The 10 m walk test has been shown to be a valid and reliable outcome measure, commonly used in patients with stroke . This test was chosen to measure walking speed over a short distance, reflecting home based mobility . To warm up, the patients started to walk 3 m behind the assessment line. The test was repeated twice, and the mean speed calculated.The 6-min walk test (6MWT) has been found simple, valid and reliable , and was chosen to measure walking endurance, reflecting the ability required for activities of daily life . The smallest detectable change (SDC) for dependent walkers is 46.3 m (95% CI), derived from Fulk , imply- ing that a change must be above this value to signify a change above measurement error for individual patients. EU Walking (0–5) was developed to classify chronic spinal cord patients’ ability to walk over ground after TTBWS . It has been established as a reliable functional assessment of walking capability in patients with spinal cord injuries . This classification was found useful in our pilot study and included to capture other aspects of clinically important improvements in walking than obtained by FAC.
Functional Independence Measure (FIM 1-7) has been found to be a reliable and valid outcome measure, com- monly used in in-patient rehabilitation . The task 9, shorter transfer, and task 13, stairs, were chosen to mea- sure the outcome of transfer ability. Shorter transfer included moving from a wheelchair to a plinth, lying down in supine and reversing back to the chair, reflecting an important activ- ity of daily living. The stair test was based on performance of nine steps up to a landing (normal 11 steps) and down to the start level. To capture the effectiveness of shorter transfer and climbing stairs, the tasks were timed.
Heart rate
The heart rate was registered by a pulse watch through the whole test session for safety of the patients. The heart rate was also registered in a standing position immediately be- fore and after the timed walk tests and used as an outcome measure. In addition, the maximum heart rate during the whole test session was registered.
Treadmill data
A control box connected to the treadmill equipment gave in-formation of gait parameters like time, belt speed and walking distance. A registration form was used for collecting treadmill data, also including the amount of body weight support, the number of assistants and handrail support.
Motor Assessment Scale (MAS) is a reliable and valid per- formance based measure, designed for clinical practice and research . This test was scored as a routine by the patient’s physiotherapists at admission and discharge, and the data were accordingly not blinded.
Visual Analogue Scale (VAS 100 mm, 0 = no change, 100 = optimal change) was used to get patients’ report of func- tional changes. VAS has been shown to be simple to use and may be tailored directly to the area of interest . Reliability is related to assessment of pain . This nonverbal measure had the advantage of fitting all the participants, also those with aphasia. The patients rated the following two questions on VAS scales: How much change in 1) walking, 2) transfer have you experienced during the treatment period?
Testing procedures
All the tests of walking and transfer were evaluated by two skilled physiotherapists in neurological physiotherapy. They tested each patient on the same occasion, but scored sepa- rately, reaching the final score by consensus. Both testers were blinded to group allocation. The test administrator (first author) was only blinded to group allocation at test 1. To improve reliability, three assessors tested five pilot patients to- gether ahead of the study start and discussed test procedures, performance of the tests and scoring alternatives with the
first author. The patients’ need for personal assistance during walking and transfer was considered, and a procedure for as- sistance developed.
Statistical analyses
For statistical analyses the SPSS version 15 was applied. χ2- tests and t-tests for independent samples were used to com- pare baseline values between groups. Overall change within groups was examined by Paired samples t-test or Wilcoxon Signed Rank Test, and differences in change between groups by independent samples t-test or Mann–Whitney U Test.Analysis of covariance (ANCOVA) was performed with base- line test values used as covariates. Treadmill and self report data were examined partly by descriptive statistics.
Results
From a total number of 616 patients admitted to the stroke division, from November 2004 to May 2009, 60 patients were consecutively included in the trial and randomized to the treadmill group and the traditional training group. Recruit- ment was 12–15 patients a year. All patients fulfilled the
intervention programme and all assessments, giving a com- plete data set. The flow of participants during the study period is illustrated in Figure 1.
The mean age of the participants was 52 years (SD 11.6), ranging from 18 to 69. The baseline characteristics of the two study groups are shown and compared in Table I. There were no significant differences between the groups regarding age, gender, diagnosis, medical co-morbidities, neuropsychological deficits, operations, side of hemiplegia and number of days after stroke. The test scores at baseline did not differ significantly between the two groups, but the treadmill group tended to score better (border significant) than the traditional group on FAC (p = 0.060) and on time to climb stairs (p = 0.057).
FAC and EU walking
According to FAC and EU Walking, the patients of both groups needed significantly less walking assistance after 5 and 11 weeks of treatment. At baseline; 55 scored < 3 on FAC implying need of physical assistance and 5 scored needing supervision. After 11 weeks 24 patients scored < 3 and the remaining 36 scored ≥.The treadmill group changed the median score of 1.5 at base- line to 3.5 after 11 weeks, while the traditional group changed from 1.0 to 2.5. Changes in mean scores were 1.7 (treadmill group) and 1.4 (traditional group), respectively. There was, however, no statistically significant difference between the two groups, neither after 5 nor after 11 weeks (Table II).
10 m walk test and 6MWT
The patients in the treadmill group improved their walking speed on the 10 m walk test from a mean of 0.22 m per second at baseline to 0.40 after 11 weeks, and the traditional group improved from a mean of 0.20 to 0.36 m per second, both significantly (Table III). The difference in change between the two groups was not significant, neither after 5 nor 11 weeks. Likewise, the number of steps decreased significantly within both groups, but not differently between the groups (Table III). Although the walking speed increased during the treatment period, the heart rate tended to decrease, but not significantly (Table IV).
The treadmill group improved the mean walking distance (6MWT) from 70 m at baseline to 138 after 11 weeks (67 m), while the traditional group improved from 59 to 115 m (56 m), both improving significantly (p < 0.001), but not differ- ently between the groups (Table III). Although the walking distance increased substantially, the heart rate was rather stable over time (Table IV). Fifteen patients in the treadmill group and 16 in the traditional group were found to improve their walking distance beyond the MDC (46 m). The tread- mill proportion beyond MDC increased the mean distance from 85 m at baseline to 206 (121 m) after 11 weeks, while the traditional proportion improved from 74 to 166 m (92 m), but the difference (29 m) in favour of the treadmill propor- tion was not statistically significant (p = 0.140). Among the patients improving below the MDC level, 14 of the treadmill group changed from 55 m at baseline to 69 m (14 m) after 11 weeks, while 15 of the traditional group changed from 42 to 58 m (16 m).
Shorter transfer (FIM, task 9) and stairs (FIM, task 13)
Both groups improved significantly in shorter transfer, from a median score of 5 (supervision) at baseline to 7 (indepen- dent) after 11 weeks of training (Table II). The change in scores over time did not differ between the groups, neither
after 5 nor 11 weeks (p = 0.077). Eight patients in each group reached the highest score (7) already at baseline, but were not excluded, due to the possibility of progress in timed scores. Only the treadmill group used significantly shorter time after 5 weeks (p = 0.028), while after 11 weeks both groups demonstrated significant improvement (Table III). However, the change over the two time periods did not differ significantly between the groups.
The performance of climbing stairs improved after 5 and 11 weeks within both groups (p < 0.001), but there
were no statistically significant differences in score changes between the groups. The treadmill group improved from a median score of 3.5 (moderate assistance) at baseline to 6 (independent with aids) after 11 weeks, while the traditional group improved from a median score of 3 to 5 (supervision) (Table II). Both groups reduced the time for climbing stairs during the intervention period (p < 0.001), but the reduction was not significantly different between the groups, neither after 5 nor 11 weeks (Table III).
Treadmill data
The mean TTBWS session, with active walking, lasted for 27.8 min (SD 2.07), ranging from 23 to 30 min. The mean walking distance per session was 599 m, ranging from 274 to 1258. Initially the physical assistance was typically one per- son, but was reduced to partly or no assistance at the end of the treadmill period (Table V). The body weight support was reduced from a mean of 7.3 kilo (SD 5.5, range 0–20) to 0.8 kilo (SD 2.0, range 0–7) at the end of the TTBWS period. Most patients were fully dependent on rail support at the beginning of the training, while partly at the end.
The mean belt speed of treadmill session 1–3 was 0.8 km per hour (kph), ranging from 0.5 to 1.4. The mean speed changed to 1.5 kph of session 18–20, demonstrating a significant in- crease (p < 0.001), ranging from 0.5 to 3.3. The mean speed of session 28–30 was 1.5 kph ranging from 0.5 to 3.1 and differed significantly only from the start sessions (p < 0.001). The belt speed in metres per second (ms), 0.23, 0.41 and 0.41 ms, cor- responded closely to the walking speed on the treadmill, 0.25,
0.41 and 0.39 ms, and to the over ground speed of 10 m walk test, 0.22, 0.33, 0.40 ms (Table III).
Motor assessment scale
The unblinded MAS scores did not show significant differences between the two groups concerning the task 5, “Walking”, and the sum scores of task 1–8. However, MAS task 1–5, including transfer, balance and walking, showed a signifi- cant difference between the groups in favour of the treadmill group (p = 0.037). From admission to discharge the treadmill and traditional group showed a statistically significant change within both groups in scores for walking (5), functional scores (1–5) and sum scores (1–8), (p < 0.001).
Subjective experiences
Reported improvements in walking ability after the training period (11 weeks) were mean 69, on the 100 mm VAS scale, in the treadmill group versus 66 in the traditional group.
Regarding transfer (shorter transfer and stairs) the mean reported improvements were 74 and 70, in the treadmill and the traditional group, respectively. The differences in scores between the groups were not significant.
Discussion
In this study, we investigated whether patients who were de- pendent on walking assistance after stroke would progress more in walking and transfer abilities after conducting TT- BWS, as compared to traditional gait training. We expected that TTBWS would be superior, as this task training modality seemed to provide better possibility for repetitive training of gait cycles, walking speed and endurance. Several outcome measures were included to capture different aspects of func- tional change. The majority of the clinical outcomes (12 of 17) were found to be in favour of the TTBWS, but for each test there was no statistical significant difference in change between the two groups after 5 and 11 weeks of training. Both groups were found to improve substantially in walking and transfer during the study period. Efforts were made to perform the study in accordance with strict scientific requirements . The participants were typical inpatients referred to primary rehabilitation, the sample size was based on power calculation, and randomization of patients to the study groups after baseline testing secured that their functional levels were within the inclusion criteria. The test pro- cedures were addressed in meetings along with written infor- mation during the study period, as it was considered an urgent need to maintain the evaluation skills during the long period of inclusion. The daily registration of content and amount of train- ing contributed to the fulfilment of the planned intervention. The prepared data file was checked for errors, and six patients (10%) were randomly selected for control of the registered data by a blinded assessor. A particular strength of this study is that all the patients fulfilled the intervention programme and the three evaluations. This may primarily be due to the fact that the patients were in-patients, but also that they found the study meaningful.A potential limitation of this study was that all assistance was performed by the first author, who was blinded to group allocation only at the baseline test. This was partly due to lack of resources, but the main consideration was to take good care of the patients, who were unfamiliar with the test situation and the assessors. In this way the testers could focus on scoring. The choice of assessment tools were based on experience from a pilot project , and discussions in an expert group of phys- iotherapists. A limitation might be that physical functioning was primarily assessed by tests in the activity domain of the In- ternational Classification of Functioning Disability and Health (ICF), and not in the body functions and structure domain . As the gait training differed in the two groups, differences in temporal, spatial and kinematic gait characteristics might have been demonstrated, not apparent in the tests of activities. Data on gait parameters were, however, collected in a subgroup of the patients and will be analysed and published later. A scientific basis for selection of training parameters that can affect out- come after TTBWS has been in demand 45. In a review study Chen found body weight support, treadmill speed, support stiffness and handrail hold important in influencing biome- chanical effects on gait pattern in hemiparetic patients. Other aspects like training frequency, duration and use of manual as- sistance or robotic orthoses for mowing the affected limb were in need of investigation . A third limitation might be that cardiovascular fitness was not separately measured. Gains in this domain have previously been reported in randomized studies of treadmill training in patients with chronic hemipa- retic stroke .
Why walking and transfer ability did not change differently between the groups, may be because both groups received the same amount of physiotherapy. While the treadmill group had 30 sessions of TTBWS, each lasting for 30 min, the tradi- tional group had the same amount of traditional gait training. During the remaining sessions, both groups had traditional gait training for 30 min each workday. Maybe the amount of gait training rather than different modalities was of most importance? A systematic review supports this interpretation, finding that the outcome after stroke was associated with the amount of exercise time conducted . Of 20 randomized studies, 14 reported significant effect on ADL and lower limb functions in favour of the group who received additional therapy time for approximately 16 h per patient. From animal research exercise in upright walking was recommended when the goal was to walk . In line with a paradigm shift in rehabilitation in the 1990s, leading to a task specific treatment concept , it is no longer considered valid to compare the effect of intensive upright walking on a treadmill with conventional physiotherapy, but rather to other ways of training upright walking . This was the reason why both groups in our study received intensive upright training, and the same amount of time for exercise.
We conducted the TTBWS more intensively in the first pe- riod and more dispersed in the second period. If TTBWS was superior alone, we expected a difference between the groups, more obviously demonstrated after 5 weeks. Otherwise, if a combination of TTBWS and traditional training were supe- rior, a difference between the groups could be expected after 11 weeks. A difference between the groups was neither dem- onstrated after 5 nor 11 weeks. The patients who improved be- yond the MDC level (46 m) on 6MWT, were found to change substantially more in metres and on the FAC than those below this level. The patients beyond MDC changed from a median score of level 2 at baseline to level 4 on FAC after 11 weeks, while those below MDC improved from a median score 1 at baseline to 2 after 11 weeks. Low FAC scores at baseline may, accordingly, predict modest improvement in timed walking in connection with rehabilitation. The difference beyond the MDC level of 29 m in favour of the treadmill proportion compared to the traditional one, represents a corridor length (30 m) and may be considered to be a minimal clinically im- portant difference (MICD) . In TTBWS manual assistance for moving limbs may be a strenuous task for the physiotherapists and limit the repeti- tions of step cycles during the sessions, giving too small dif- ferences between the walking modalities to create a difference in walking outcome . The amount of step cycles may differ in various studies, according to different training time per ses- sion and frequencies of sessions per week. For example time for mounting of harness may be included in the training session and reduce the time for effective training . In our study we excluded mounting of harness from the training time, giving a mean of 28 min effective training, similar to the time used for traditional gait training. This contributed to validate the com- parisons of effect between the different modalities of intensive gait training. Progression in different aspects of TTBWS, like walking speed, weight bearing, transfer of weight, and reduction of handrail support was encouraged by the registering procedure, giving basis for further progression. While some patients were able to progress in walking speed, others had medical co- morbidities, secondary ankle instability or spasticity problems that impaired the intensity of progression. However, the walk- ing speed increased significantly during the intervention pe- riod and “handrail hold” decreased. Handrail support should be balance and not weight bearing support, which might be difficult to differentiate for the patients . On the other hand, support on both sides has shown positive influence on gait symmetry . Handrail support and safe conditions on the treadmill, may on the other hand have given the traditional group an advantage, as they had to cope with real environments during training.
Most in-patient rehabilitation is completed within three months of intervention , as the largest improvements are found to occur early after stroke, and the process of spontaneous recovery is observed to decline or be completed after 6–10 weeks . Several treadmill studies are conducted from postacute or subacute phase, within 3 months alternatively in the chronic stage . Our mean study period was within 3–6 months af- ter stroke, not being directly comparable to the studies in early phase, but nevertheless the nearest available for comparison. The outcome of our study corresponds with findings by Hesse describing single group studies and four ran- domized studies in a review . TTBWS was unex- pectedly found to be equivalent, but not superior, to other physiotherapy interventions using the principle of repetitive practice of gait in acute stroke patients . Moseley’s update review of 15 randomized studies of treadmill therapy with and without body weight support compared to tradi- tional or sham training on 622 patients with stroke, showed a trend of improvement in walking speed among indepen- dent walkers, but the results were still not conclusive . Subgroup analysis concerning walking speed and endurance indicated progress in dependent participants, but data to sup- port any conclusion were limited. A resent multi-centre study of 126 stroke patients unable to walk at inclusion, compar- ing TTBWS with over ground training, showed independent walking two weeks earlier in the experimental group than in the control group, but the results did not differ significantly between the groups . However, secondary outcomes of this study showed a significant difference between the groups for walking capacity, with a difference of 57 m on 6 MWT in favour of the treadmill group after 6 months . A Cochrane review of 17 randomized studies, involving 837 participants, showed some evidence that electromechanical– assisted walking training in combination with physiotherapy had a significant effect on patients in the acute and subacute period after stroke. Non-ambulatory patients receiving elec- tromechanical or robotic assistance in early phase were more likely to achieve independent walking than those training walking without these devices . A systematic review of six studies, involving 549 patients, supported this finding . However, no significant results were found in the Cochrane review regarding walking velocity (metres per second) nor walking capacity (metres walked in 6 min . The achieved effect was also less evident in patients treated more than three months after stroke . Although 35 studies within tread- mill therapy and related devices using manual or elec- tromechanical assistance are included in meta-analysis , the efficiency of these methods above traditional training is still found to be unclear, especially in non-ambula-tory patients .
In our study, TTBWS and traditional gait training dem- onstrated substantial outcomes in the activity domain of the ICF . The clinical implication of these results may be the freedom to select between the two approaches for restoring walking and transfer ability after stroke. However, it should be taken into consideration that the intervention was system- atic, goal directed and intensive in both groups, as previously described. Our participants experienced TTBWS as a motiva- tional approach. Functional testing also seemed to motivate the patients during the intervention period and should be implemented in connection with training.
Conclusions
This study showed no significant difference between patients in need of walking assistance after stroke, who practiced TT- BWS and over ground training versus those who conducted traditional gait training alone, even if both groups improved substantially. Thus, the effect of the two approaches used in the study could be considered equivalent, although most outcome measures showed a small tendency in favour of the treadmill group.