Walking After Stroke: What Does Treadmill Training With Body Weight Support Add to Overground Gait Training in Patients Early After Stroke?
A Single-Blind, Randomized, Controlled Trial
Background and Purpose— This study aimed to assess the effectiveness of gait training using body weight support on a treadmill compared with conventional gait training for people with subacute stroke who were unable to walk.
Methods— This was a single-blind, randomized, controlled trial with a 6-month follow-up. Ninety-seven subjects were recruited within 6 weeks of stroke onset and were randomly assigned to conventional rehabilitative treatment plus gait training with body weight support on a treadmill (experimental group; n=52) and conventional treatment with overground gait training only (control group; n=45). All subjects were treated in 60-minute sessions every weekday for 4 weeks. Outcome measures were Motricity Index, Trunk Control test, Barthel Index, Functional Ambulation Categories, 10-meter and 6-minute Walk Tests, and Walking Handicap Scale. Assessments were made at baseline, after 20 sessions of treatment, 2 weeks after treatment, and 6 months after stroke.
Results— After treatment, all patients were able to walk. Both groups showed improvement in all outcome measures (P<0.0063) at the end of the treatment and at follow-up. No differences were seen between the 2 groups before, during, and after treatment and at follow-up.
Conclusions— In subacute patients with stroke, gait training on a treadmill with body weight support is feasible and as effective as conventional gait training. However, the need for more personnel for treadmill training makes the use of robotically assisted systems more compelling.
The ability to walk is one of the most important goals in stroke rehabilitation.1 Gait impairment is a significant contributor to long-term disability after stroke. Gait recovery in stroke survivors is variable. Wade et al2 showed that only 22% of 45 patients who could not walk as a consequence of stroke were able to walk normally within 3 months of recovery. Although early rehabilitative intervention in walking training is generally recognized as beneficial in patients with stroke, it is less clear what type of treatment program would produce the best outcome.
Gait training on a treadmill with some body weight support (BWS) is a method of treating walking impairment that is becoming popular. Modern concepts of rehabilitation favor a task-specific repetitive approach.3 Moreover, it has been shown that higher intensities of walking practice (resulting in more repetitions trained) result in better outcomes after stroke.4 The use of a treadmill may increase the number of steps taken, whereas BWS provides enough assistance to facilitate walking.
Preliminary data show that chronic patients may take advantage of this kind of treatment.5–7 Data on patients early after stroke are controversial. Although some authors found a beneficial effect,8–10 others did not.11,12 Some authors found no beneficial effect of BWS except in more severely impaired patients.13
A recently performed meta-analysis14 evaluated treadmill training with and without BWS. The authors concluded that overall no statistically significant effect of treadmill training with or without BWS was present. Although individual studies suggested that treadmill training with BWS may be more effective than treadmill training alone and that treadmill training plus task-oriented exercise may be more effective than sham exercises, further trials are required to confirm these findings.
Authors who performed gait training with BWS on patients early after stroke evaluated also patients who were able to walk at the time of inclusion. A part of the beneficial effect on patients still unable to walk at inclusion might therefore have been masked by the results of the patients with some residual gait ability.
We decided to conduct a clinical trial evaluating if gait training with treadmill and BWS in the early phase after stroke in patients who were unable to walk is feasible and may lead to better results compared with conventional overground gait training.
Materials and Methods
The Gruppo Italiano Allevio Carico Ictus (GISACI) recruited potential subjects by screening patients with stroke at admission to 6 rehabilitation centers in Italy from January 2005 to December 2006. Coordination and data management were carried out at the University Hospital of Parma, Italy. Institutional Review Board approval was obtained from each center involved. Written consent was obtained from each subject before inclusion. The GISACI study is a prospective, single-blind, randomized, controlled trial including 2 arms: an experimental group (EG) that received BWS gait training and overground gait training and a control group (CG) that received conventional treatment and overground gait training for the same amount of practice time.
Subject Selection and Randomization
Eligible patients should have been recruited and randomized and have started treatment (either EG or CG) within 45 days of the onset of hemiparesis caused by right or left ischemic or hemorrhagic stroke. No inclusion or exclusion criteria based on gender, age, or ethnic origin were applied. We considered as eligible those patients who were able to control the sitting position on a rigid plane surface with the legs hanging freely and without the help of the arms for at least 30 seconds. They should also be able to control the trunk in the upright position even with the help of the upper extremities gripping a fixed support or other aid (cane, tripod). We considered eligible only patients without lower limb spasticity (Ashworth scale ≤1). It was established that all patients were in stable cardiovascular condition with a low, although slightly greater, risk for vigorous exercise than apparently healthy persons (Class B according to the American College of Sports Medicine). Exclusion criteria were significant disability before stroke (modified Rankin Scale ≥2); significant prestroke gait disability (Walking Handicap scale ≥2); and mild gait impairment at time of enrollment (ability to walk without aids for at least 3 m or to walk for more than 6 m with the aid of a cane or tripod). Patients having done previous treadmill training and/or with a Class C or D exercise risk according to the American College of Sports Medicine criteria or Class III or IV in the New York Heart Association classification system were excluded. Patients with orthopedic or other disorders causing a gait limitation before stroke onset were excluded. Patients who did not complete the treatment (experimental or control) within 5 weeks of study inclusion were excluded from the analysis.
Patients included in the study were randomly assigned to EG or CG. The randomization scheme was generated by custom-made software that used the Lehmer algorithm. Patients were assigned to one of the 2 groups according to age and gender. Each of the 6 centers was considered as a separate block in which patients were randomized.
Patients in the EG did 20 minutes of gait training on a treadmill (RHC500; AirMachine Com srl) with BWS (Unweighing System; Biodex Medical Systems, Shirley, NY) followed by 40 minutes of conventional training. Training was administered 5 times a week for an overall number of 20 sessions, which should have been completed within 5 weeks of inclusion in the study. We considered only the net training time on the treadmill, ie, the time actually spent walking and not the time required for getting the patient on and off. The quantity of BWS was tailored to the patient’s capability and was limited to 40% of body weight. BWS was applied by means of a climbing harness and was gradually reduced during the sessions depending on the patient’s compliance and progress.
Gait training with BWS was performed with the help of 2 trained physical therapists for each patient to control the paretic lower extremity and pelvis. Treadmill velocity was adjusted to enable gait training at increasing speeds, starting from 0.1 m/s and aiming at ≥1.2 m/s according to the patient’s compliance and progress. When pelvic and paretic lower extremity control was considered adequate, gait training was administered by one physical therapist only.
Conventional treatment was performed for 40 minutes, not immediately after treadmill training. No specific indications were given to the rehabilitation team, and the treatment was tailored to the patient’s needs and the rehabilitation team’s goals. Skilled therapists performed treatment according to the person’s needs.
Patients in the CG underwent 20 sessions of conventional treatment (consisting of overground gait training) of 60 minutes each. Treatment was administered 5 times per week for an overall number of 20 sessions, which should have been completed within 5 weeks of inclusion in the study. No specific indications were given to the rehabilitation team, and the treatment was tailored to the patient’s needs and the rehabilitation team’s goals. Skilled therapists performed treatment according to the person’s needs.
Patients of both groups were allowed to receive additional neuropsychological and occupational therapy sessions if needed. No specific indications were given regarding pharmacological treatment in the study patients. The use of antispastic drugs and/or botulinum toxin was recorded but not considered as an independent variable. The therapists who treated EG patients were not the same as those who treated CG patients.
The included patients were evaluated before treatment (T0), after 10 sessions of treatment (T1), at the end of the 20 sessions of treatment (T2), and 2 weeks after (T3). Participants were also evaluated 6 months after stroke onset (T4).
The following measures were evaluated: (1) Motricity Index at the paretic upper and lower extremity; (2) Trunk Control Test; (3) modified Rankin Scale; (4) Barthel Index; (5) Functional Ambulation Categories; (6) Ashworth scale to assess lower limb spasticity; (7) Token Test to evaluate oral comprehension in aphasic patients; (8) Albert test to evaluate unilateral spatial neglect; (9) proprioceptive sensibility at the lower limb; this was assessed by evaluating the sense of position of the big toe, ankle, and knee. An ordinal scale was used: 0, intact; 1, partially compromised; 2, severely compromised; (10) 10-m Walk Test; in this test, the patient was required to walk wearing shoes with aids at his or her maximal speed. The test was performed on a 20-m-long walkway, but only the 10 central meters were considered to allow for adequate acceleration; (11) 6-minute Walk Test. In this test, the amount of space covered during 6 minutes was measured with the patient walking at a comfortable speed; (12) Borg Scale to evaluate dyspnea at the end of the 6-minute Walk Test; and (13) Walking Handicap Scale.
At T0, Tests 1 to 9 were performed. At T1 and T3, all assessments except 3, 7, 8, and 13 were performed. At T2, all assessments except 3 and 13 were performed. At T4, all assessments except 7 and 8 were performed. Assessments were performed by physicians and therapists not involved in the treatment of the patient. Because the patients were not blinded to the allocation arm, the study cannot be considered double-blind.
Statistical analysis was performed with SPSS Version 16. A Type I error level of 0.05 was chosen.
Data are represented as medians for ordinal scores and means for contiguous variables. The Bonferroni correction for multiple comparisons was applied, resulting in a Type I error level of 0.00625, considering as gait-relevant variables only Motricity Index at the paretic lower extremity, Trunk Control Test, modified Rankin Scale, Barthel Index, Functional Ambulation Categories, 10-meter Walk Test, 6-minute Walk Test, and Walking Handicap Scale. Homogeneity between groups was analyzed with the χ2 test. Analyses of ordinal parameters (Mann–Whitney U test and Wilcoxon signed rank test) were used for intragroup and intergroup comparisons, respectively. Contiguous parameters were analyzed with Student t test. Data were analyzed as per protocol.
The Figure shows the Consolidated Standards of Reporting Trials (CONSORT) diagram for recruitment. Of 358 screened patients with stroke admitted to the rehabilitation centers, 102 were considered eligible, signed consent to enter the study, and were randomized. Five randomized patients coming from a single center were not treated and were therefore discarded. As a result, data from 97 patients were analyzed.
Table 1 describes the demographic and clinical features of the 2 groups. EG and CG patients did not show differences in gender, age, or type and side of lesion after randomization. However, because 5 patients were discarded, patients in CG were slightly older (70±11.8 versus 65.5±12.2 years, P=0.03) and were more frequently affected by right cerebral lesions (52 versus 44 patients, P=0.045). At T0, the outcome variables were homogeneous in the 2 groups.
During the treatment period, 12 patients dropped out, 9 of the EG and 3 of the CG (P>0.05). Another patient in the EG died and did not complete follow-up. Two patients declared that the discomfort caused by the harness during BWS treadmill training was the reason for their dropping out. Six patients were transferred to other rehabilitation centers to be closer to their families and 4 were temporarily transferred because of intercurrent respiratory complications and thus did not complete treatment. The onset-admission intervals, the time from stroke to inclusion in the study, and the lengths of stay were comparable in both groups (Table 1).
The interval from stroke onset to inclusion in the study was similar in the 2 groups. The Token Test and Albert test did not show any significant difference between the 2 groups and between T0 and T2 (Table 2). The 2 groups did not have sensibility deficits with a median value of 0 at T0 and T4. All patients in both groups were able to walk at discharge, ie, had Functional Ambulation Categories scores ≥1. Each group showed significant improvement from T0 to T3 in all outcome variables (Table 2); comparison between the 2 groups did not reveal any significant differences. At the same time, the results of the Walking Handicap Scale at T3 were not significantly different between the 2 groups.
At T4, there was again a significant improvement in all assessments performed, but there were no significant differences between the 2 groups (Table 2). Spasticity assessed with the Ashworth scale did not show significant differences between the 2 groups and the scores were low without any important modifications over time (EG T0: hip 0, knee 0, ankle 1; T4: hip 1, knee 1, ankle 1; CG T0: hip 0, knee 1, ankle 1; T4: hip 1, knee 1, ankle 1). The Borg scale at the end of the 6-minute Walk Test showed no modifications during the study, except between T2 and T3 in the CG. The Trunk Control Test improved during the study, except from T2 to T3 in the EG and from T3 to T4 in the CG. The EG started BWS treadmill training at a speed of 0.19±0.09 m/s and had improved by the last treatment to 0.40±0.20 m/s (P<0.0001). BWS also decreased significantly, from 35.40±6.60 at the first treatment to 10.00±11.80 at the last session (P<0.0001).
Our study supports the hypothesis that gait training on a treadmill with BWS early after stroke is a feasible and safe treatment. However, our study failed to demonstrate superiority over conventional treatment regarding gait and body function, disability, and compliance. Patients of both groups showed meaningful but similar improvements in all selected variables at the end of treatment and at follow-up.
Because no control group without gait training was used, it is difficult to evaluate the role of spontaneous recovery of gait function. However, early intensive gait-focused training of the ambulatory ability in people at the early stages after stroke has been shown to be more effective than spontaneous recovery and usual care,15,16 so gait training given also to the CG is likely to produce a greater effect than no treatment.
Our study points out that adding 20 minutes of treadmill training with BWS to conventional overground gait training is not more effective than conventional treatment in the early phase after stroke. We started gait training early after stroke, that is, our patients were randomized to EG or CG and started treatment within approximately 4 weeks of stroke onset. Other studies previously investigated the possible usefulness of gait training with BWS,9–12,17 and some of these were performed on patients in the early phase after stroke.9–12
Nilsson et al12 found no significant differences between gait training using a treadmill with BWS and conventional treatment. They evaluated patients at 22 days after stroke, so the timing is comparable to that of our study. Even if the treatment modalities were quite similar to those used in our work, the characteristics of their patients were different. The patients were younger and less disabled, and treatment was performed for a longer period. In fact, only half of their patients could not walk at the time of inclusion, and patients were treated for 10 weeks.
Da Cunha et al11 also found no difference between conventional gait training and treadmill with BWS. However, these authors recognized that their study was underpowered for detecting differences in any gait variable evaluated. Moreover, they did not provide data on follow-up and their sample was composed of less disabled patients compared with ours.
A recent paper by McCain et al9 evaluated the feasibility of gait training with treadmill and BWS before overground training. Their sample was composed of patients in the early phase after stroke with severe gait disability (Functional Independence Measure gait <2). Unlike previous works and our own data, McCain et al found a relevant difference between conventional treatment and treadmill with BWS gait training. This study was, however, performed on a small sample. An interesting feature of their work was the use of kinematic analysis for gait assessment, but they did not evaluate modifications in disability.
Visintin et al10 observed better functional balance, motor recovery, overground walking speed, and endurance in the patients treated with BWS. They evaluated patients relatively early after stroke (mean, 68 to 78 days from vascular injury) who were treated for 20 minutes 4 times/week for 6 weeks. This study differs from ours in that the authors recruited patients >2 months after stroke and patients were already able to walk at the time of recruitment.
A possible explanation for our results could be that patients in the EG were undertreated, which may have masked the presence of significant differences. However, Visintin et al10 and McCain et al9 performed treatment sessions of similar duration to those in our study and were able to point out relevant differences. Other authors11,12 treated patients more intensively but, like us, did not detect any significant differences between conventional and experimental treatment results.
It is possible that more intense treatment or using different gait speeds on the treadmill would have produced different results. However, at present, there is no agreement about the intensity, frequency, and duration of treadmill gait training with BWS. So we arbitrarily decided on an intensity of 20 minutes per session, 5 times a week, after having considered previous reports in the literature and the feasibility of this treatment in our centers.
Because our patients were completely unable to walk at the time of inclusion, it is obvious that treadmill training with BWS allowed more intensive treatment than overground training. However, this did not influence the final outcome.
The currently available robotic systems for gait training may overcome the problems related to short training duration on a treadmill. However, a recent study performed on subacute stroke patients showed that patients given robotically assisted gait training had slower gait speeds at the end of treatment (24 sessions of 45 minutes each) and at follow-up than patients who performed conventional overground gait training.18
We decided to use a broader evaluation than previous studies using scales and measures covering all International Classification of Functioning, Disability and Health (ICF)-2 domains. Our patients showed a higher rate of gait function recovery than reported by other authors.19 Even if in Italy patients with stroke transferred for rehabilitation tend to have a good functional prognosis, this difference can be explained by several factors. The most important one could be that in our sample more than two thirds of patients with stroke were excluded because they did not meet the inclusion criteria. It is therefore likely that the most severely impaired patients were excluded. Another possible explanation is that our sample included also patients with hemorrhagic stroke, who may have better outcomes compared with patients with ischemic stroke.20
The most important possible bias of our work is that the treatment allocation was not blinded to patients and treating physical therapists and physicians. After a preliminary discussion with the clinicians responsible for the research in each center and considering the availability of space, personnel, and structures, we realized that blinding could not be warranted. So we decided to start with a single-blind trial, although we were aware of the possible biases, including a placebo effect and experimenter’s bias.
Although these biases could not be excluded, our results were similar to those reported in previous works.11,12 Moreover, we chose outcome measures that could not be significantly influenced by the assessor or the patient (such as Functional Ambulation Categories, modified Rankin Scale, and Barthel Index). What was highly unexpected in our study was that the experimental treatment yielded results that were not superior to those of conventional treatment.
A possible flaw of this study is that the number of steps taken on the treadmill was not recorded, so patients may have had different treatment intensities. However, even if it is assumed that the number of repetitions and gait outcome have a correlation,4 no strong evidence supporting this concept is currently available.21
Another possible shortcoming is that we did not control how much of the training time for the CG was gait training. Jette et al22 in a descriptive study reported a mean duration of gait training of 31.3% of the total treatment time in patients with stroke. This would correspond, in our patients, to nearly 20 minutes of overground gait training. Because the duration of overground gait training for each patient was freely chosen by the team, we believe 20 minutes would be a reasonable estimate.
Our data show that gait training with BWS is a feasible and effective treatment in the early phase after stroke in patients with moderate to severe walking impairment. Our patients showed a significant improvement in all selected parameters concerning physical functioning, activities, and compliance. However, the results of the experimental group were not superior to those obtained with conventional treatment. Furthermore, the need for 2 therapists to administer gait training with BWS makes the use of robotically assisted training systems more compelling. Further studies are needed to assess if early treatment with these new technologies can be useful in patients in the early phase after stroke.
GISACI Group: U. O. di Medicina Riabilitativa, Azienda Ospedaliero-Universitaria di Parma: Marsilio Saccavini, MD, Monica Nora, MD, Enrico Maestrini, MD, Daniela Contini, PT, and Lucia Bassi, PT. U.O. di Medicina Riabilitativa, Azienda Universitario-Ospedaliera di Novara: Alessandra Florio, MD, and Annamaria Airoldi, PT. U.O. di Recupero e Rieducazione Funzionale, Ospedale di Piacenza e di Borgonovo Valtidone: Gabriella Lenti, MD, Giuseppe Sacco, MD, Anna Cassio, MD, Nicoletta Pellegrino, PT, Monica Santin, PT, and Rossella Raggi, PT. Centro di Riabilitazione di Conselve, Azienda Ospedaliera di Padova: Massimo Iannilli, MD, G. Sorgato, PT, and E. Danese, PT. Servizio di Recupero e Rieducazione Funzionale, Casa di Cura S. Francesco, Bergamo: F. Mazzoleni, PT, and R. Trapletti, PT. Ospedale Maggiore di Bologna: Elisabetta Magni, MD, and Gennaro Filoni, PT.
- Received April 15, 2009.
- Accepted May 20, 2009.
Carr J, Shepherd R. Stroke Rehabilitation: Guidelines for Exercises and Training. London: Butterworth Heinemann; 2003.
Hesse S, Bertelt C, Jahnke MT, Schaffrin A, Baake P, Malezic M, Mauritz KH. Treadmill training with partial body weight support compared with physiotherapy in nonambulatory hemiparetic patients. Stroke. 1995; 26: 976–981.
Hesse S, Uhlenbrock D, Sarkodie-Gyan T. Gait pattern of severely disabled hemiparetic subjects on a new controlled gait trainer as compared to assisted treadmill walking with partial body weight support. Clin Rehabil. 1999; 13: 401–410.
Hassid E, Rose D, Commisarow J, Guttry M, Dobkin B. Improved gait symmetry in hemiparetic stroke patients induced during body weight-supported treadmill stepping. J Neurorehabil. 1997; 11: 21–26.
Visintin M, Barbeau H, Korner-Bitensky N, Mayo NE. A new approach to retrain gait in stroke patients through body weight support and treadmill stimulation. Stroke. 1998; 29: 1122–1128.
Nilsson L, Carlsson J, Danielsson A, Fugl-Meyer A, Hellström K, Kristensen L, Sjölund B, Sunnerhagen KS, Grimby G. Walking training of patients with hemiparesis at an early stage after stroke: a comparison of walking training on a treadmill with body weight support and walking training on the ground. Clin Rehabil. 2001; 15: 515–527.
Kosak MC, Reding MJ. Comparison of partial body weight-supported treadmill gait training versus aggressive bracing assisted walking post stroke. Neurorehabil Neural Repair. 2000; 14: 13–19.
Duncan P, Studenski S, Richards L, Gollub S, Lai SM, Reker D, Perera S, Yates J, Koch V, Rigler S, Johnson D. Randomized clinical trial of therapeutical exercise in subacute stroke. Stroke. 2003; 34: 2173–2180.
Hidler H, Nichols D, Pelliccio M, Brady K, Campbell DD, Kahn JH, Hornby GT. Multicenter randomized clinical trial evaluating the effectiveness of the Lokomat in subacute stroke. Neurorehabil Neural Repair. 2009; 23: 5–13.
Paolucci S, Bragoni M, Coiro P, De Angelis D, Fusco FR, Morelli D, Venturiero V, Pratesi L. Quantification of the probability of reaching mobility independence at discharge from a rehabilitation hospital in nonwalking early ischemic stroke patients: a multivariate study. Cerebrovasc Dis. 2008; 26: 16–22.
Paolucci S, Antonucci G, Graso MG, Bragoni M, Coiro P, De Angelis D, et al. Functional outcome of ischemic and hemorrhagic stroke patients after inpatient rehabilitation. Stroke. 2003; 34: 2861–2865.
Jette DU, Latham NK, Smout RJ, Gassaway J, Slavin MD, Horn SD. Physical therapy interventions for patients with stroke in inpatient rehabilitation facilities. Phys Ther. 2005; 85: 238–248.