edicine and Rehabilitation
Archives of Physical M

journal homepage: www.archives-pmr.org

Archives of Physical Medicine and Rehabilitation 2017;98:738-45
ORIGINAL RESEARCH
Effects of Gait Training With Body Weight Support
on a Treadmill Versus Overground in Individuals
With Stroke
Gabriela L. Gama, MSc,a Melissa L. Celestino, MSc,a José A. Barela, PhD,a,b

Larry Forrester, PhD,c Jill Whitall, PhD,d,e Ana M. Barela, PhDa

From the aInstitute of Physical Activity and Sport Sciences, Cruzeiro do Sul University, São Paulo, SP; bDepartment of Physical Education, São
Paulo State University, Rio Claro, SP, Brazil; cMaryland Exercise & Robotics Center of Excellence, Veterans Administration Maryland Health Care
System, Baltimore, MD; dDepartment of Physical Therapy and Rehabilitation Science, School of Medicine, University of Maryland, Baltimore,
MD; and eFaculty of Health Sciences, University of Southampton, Southampton, United Kingdom.

Abstract

Objective: To investigate the effects of gait training with body weight support (BWS) on a treadmill versus overground in individuals with

chronic stroke.

Design: Randomized controlled trial.

Setting: University research laboratory.

Participants: Individuals (NZ28) with chronic stroke (>6mo from the stroke event).

Interventions: Participants were randomly assigned to receive gait training with BWS on a treadmill (nZ14) or overground (nZ14) 3 times a

week for 6 weeks.

Main Outcome Measures: Gait speed measured using the 10-meter walk test, endurance measured using the 6-minute walk test, functional

independence measured using the motor domain of the FIM, lower limb recovery measured using the lower extremity domain of the Fugl-Meyer

assessment, step length, step length symmetry ratio, and single-limb support duration. Measurements were obtained at baseline, immediately after

the training session, and 6 weeks after the training session.

Results: At 1 week after the last training session, both groups improved in all outcome measures except paretic step length and step length

symmetry ratio, which were improved only in the overground group (PZ.01 and PZ.01, respectively). At 6 weeks after the last training session,

all improvements remained and the treadmill group also improved paretic step length (P<.001) but not step length symmetry ratio (P>.05).

Conclusions: Individuals with chronic stroke equally improve gait speed and other gait parameters after 18 sessions of BWS gait training on

either a treadmill or overground. Only the overground group improved step length symmetry ratio, suggesting a role of integrating overground

walking into BWS interventions poststroke.

Archives of Physical Medicine and Rehabilitation 2017;98:738-45

ª 2016 by the American Congress of Rehabilitation Medicine
Partial body weight support (BWS) systems have been widely
used for gait rehabilitation poststroke.1-7 In individuals with
stroke, these systems improved body weight distribution between
paretic and nonparetic limbs as erect posture is facilitated,5 and
they promote improvement in spatial-temporal gait characteristics,
Supported by the São Paulo Research Foundation e FAPESP and by fellowship from Coor-

denação de Aperfeiçoamento de Pessoal de Nı́vel Superior e CAPES. All financial support had no

influence on analysis and interpretation of data or writing of manuscript.

Clinical Trial Registration No.: NCT02088255.

Disclosures: none.

0003-9993/17/$36 - see front matter ª 2016 by the American Congress of Re

http://dx.doi.org/10.1016/j.apmr.2016.11.022
including interlimb symmetry of stance and swing phases, muscle
activity, and joint angle excursions of the lower limbs.8 BWS used
with treadmills facilitates performance of a high number of
symmetrical and consistent steps,9 as well as allowing control of
walking speed. In contrast, the requirements for walking on
treadmills are different from those for overground walking, mainly
in terms of propulsion and balance.10,11

Treadmills may provide a degree of passive movement to the
lower limbs with little change in muscular activation,8,9,12-14

limiting the extent of skill transfer to overground walking. This
habilitation Medicine

Delta:1_given name
Delta:1_surname
Delta:1_given name
Delta:1_surname
Delta:1_given name
Delta:1_surname
http://crossmark.crossref.org/dialog/?doi=10.1016/j.apmr.2016.11.022&domain=pdf
http://dx.doi.org/10.1016/j.apmr.2016.11.022
http://www.archives-pmr.org
http://dx.doi.org/10.1016/j.apmr.2016.11.022


Body weightesupported training after stroke 739
raises a question whether BWS alone promotes improvement in
gait performance or whether this is due to the interaction between
BWS and treadmill. A related question is whether using gait
training with BWS on would yield any differences in gait
improvement compared with BWS on the treadmill.

Only 2 studies have used overground BWS training in in-
dividuals with stroke, including a case study15 and a single-arm
study with a small sample.16 Sousa et al16 reported improve-
ments in walking speed and step length symmetry ratio as well as
increased stride lengths and segmental angle rotations17 after
overground BWS gait training, but the study lacked a treadmill
comparison group. Comparing BWS gait training in a controlled
treadmill versus overground experiment may reveal differential
effects due to use of BWS on a moving versus stationary walking
surface. The purpose of this study was to investigate the effects of
moving versus stationary walking surfaces during BWS gait
training on measures of spatial-temporal gait parameters and
clinical function in individuals with chronic stroke. It was hy-
pothesized that overground BWS training would elicit a greater
improvement in walking performance than would treadmill BWS
training, because the former would reduce the need for skill
transfer from the externally driven treadmill to fully self-generated
control of overground mobility tasks.
Methods

A randomized controlled trial was conducted using Consolidated
Standards of Reporting Trials guidelines. The study was approved
by the research ethics committee of Cruzeiro do Sul University and
was registered at ClinicalTrials.gov (Clinical Trial RegistrationNo.:
NCT02088255). All procedures were performed with adequate
understanding, and written, signed informed consent of the partic-
ipants was obtained before entry into the study.

Setting and participants

From April 1, 2014, to June 15, 2015, 114 individuals with stroke
were contacted and invited to join the study (fig 1). One physical
therapist researcher screened all potential participants. Inclusion
criteria were chronic hemiparetic gait after an ischemic or hemor-
rhagic stroke, >6 months from the stroke event, absence of cardiac
(or medical clearance for participation), orthopedic, or pulmonary
disease or other neurologic impairment that could compromise gait
or training, ability to follow 2-step verbal commands, and ability to
walk 10m with or without assistance. Individuals who presented
with uncontrolled blood pressure were excluded. All stages of the
study were conducted in the Movement Analysis Laboratory at
Cruzeiro do Sul University, São Paulo, São Paulo, Brazil.

Randomization and blinding

After baseline testing, and using a computer-based algorithm,
participants were randomly allocated to 2 BWS groups: treadmill
List of abbreviations:

6MWT 6-minute walk test

10MWT 10-meter walk test

BWS body weight support

FM Fugl-Meyer assessment

TMG treadmill training group

www.archives-pmr.org
training group (TMG) or overground training group. The
researcher who did the randomization and data analyses was not
involved in any assessment or training session. Complete and
continuous blinding of researchers who performed the training
sessions and the assessments was not feasible because of
personnel constraints. Participants were not blinded to the training
conditions, but were unaware of the hypothesis of the study.

Intervention protocols

The BWS system used for the TMG consisted of a treadmill
(model TK35a) and a metal frame with an instrumented load cell
to register the amount of supported body weight. The BWS system
(fig 2) used for the overground training group consisted of a
suspended rail (7m) mounted on the ceiling (3m) and supported by
2 steel beams.b A moving cart was attached to the bottom of the
rail, allowing backward and forward movements and controlled by
a belt system linked to a servomotor. A customized program
(LabVIEWc) was developed to control displacement, velocity, and
acceleration of the moving cart. In addition, a second servomotor
controlled unloading of body weight through a harness and
instrumented load cell system.

Training sessions were conducted 3 times weekly, with at least
1 day between sessions, for 6 weeks, totaling 18 training sessions.
Each session lasted up to 45 minutes. Throughout the training
sessions, each participant’s heart rate was monitored and blood
pressure was measured at the beginning and end of each session.
Rest intervals were provided according to individual needs.

The amount of BWS during training sessions ranged from 30%
to 0% of body weight. The amount of BWS was based on the
individual alignment of trunk and limbs with proper weight shift
and bearing onto the hemiplegic limb during the loading phases of
gait. Walking speed was set to match participants’ comfortable
level on the treadmill or overground. The overground training
group walked back and forth along the walkway. Participants were
allowed to use the front handrail of the treadmill (TMG) or the
therapist’s hand (overground training group), but through training
sessions, all participants were encouraged to walk without any
external assistance other than the BWS system.

Progression of the training was achieved by reducing the
BWS, increasing the gait speed, and/or reducing the patient
support from the handrail for the TMG and the therapist’s hand
for the overground training group. The training parameter changes
to progress the training were implemented at the beginning of
each session. The progression was maintained as long as the
participants could maintain alignment of trunk and limbs with
proper weight shift and bearing onto the hemiplegic limb during
the loading phases of gait. If not, the parameter change was
decreased to the previous value. Two trained therapists conducted
all training sessions of both groups and provided similar verbal
and manual cues.

Outcome measures and follow-up

Participants were assessed 1 week before the first training session,
1 week after the last gait training session, and 6 weeks after the
last training session. One experienced researcher took the lead on
all assessments, and standardized instructions for each test were
given to ensure consistency in test administration. Participants
underwent the following evaluations: 10-meter walk test
(10MWT),18-20 6-minute walk test (6MWT),21 motor domain of
the FIM,22,23 lower extremity domain of the Fugl-Meyer

http://ClinicalTrials.gov
http://www.archives-pmr.org


Fig 1 Flow diagram of the study following Consolidated Standards of Reporting Trials guidelines.

740 G.L. Gama et al
assessment (FM),24 and quantitative gait analyses (see below).
Rest intervals were provided between tests according to individual
needs, when necessary.

The primary outcome measure was gait speed measured using
the 10MWT. During this test, participants were required to walk
10m at a comfortable speed. Two photocellsa measured the time
Fig 2 Representation of the BWS system adopted by the overground

training group.
required to cover the intermediate 6m to exclude periods of ac-
celeration and deceleration. The average of 3 trials was calculated.

Secondary outcome measures included endurance measured
using the 6MWT, functional independence measured using the
motor domain of the FIM, lower limb recovery measured using the
lower extremity domain of the FM, which evaluated the subscales
of joint pain, passive joint motion, sensibility, voluntary move-
ment, reflex activity, and coordination. Finally, paretic and non-
paretic step length, step length symmetry ratio, and paretic and
nonparetic single-limb support were measured using a computer-
ized gait analysis system during walking at a comfortable speed.
For both 10MWT and 6MWT, participants were allowed to use
assistive devices, if necessary.

The gait analysis was performed on a 7m walkway equipped
with 2 embedded force platforms (model 9286BAd). A comput-
erized gait analysis system (VICONe) with 7 infrared cameras
(Bonita 10e) was used to acquire data from reflective markers that
were placed on the main body landmarks (Vicon Plug-In Gait
modele).25 After a calibration trial, participants were asked to walk
at a comfortable self-selected speed. They were not allowed to use
any assistive devices, but, when necessary, a therapist offered her
www.archives-pmr.org

http://www.archives-pmr.org


Table 1 General baseline characteristics of individuals

Characteristic TMG OGG P

Sex: female/male 7/7 8/6 .71*

Age (y) 58.7�8.4 57.7�10.1 .78

Mass (kg) 66.7�11.1 66.0�12.4 .86

Height (m) 1.63�0.08 1.60�0.11 .43

Time poststroke (mo) 60.2�55.4 53.8�42.2 .73

Type of lesion: I/H 12/2 10/4 .37*

Hemiparesis side: R/L 5/9 8/6 .27*

Mini-Mental State

Examination score

24.43�3.98 21.85�6.87 .23*

Baseline 10MWT (m/s) 0.69�0.25 0.73�0.28 .81

Baseline 6MWT (m) 244�119 240�152 .94

Baseline FIM score 80.6�7.31 83.0�7.1 .46

Body weightesupported training after stroke 741
hand to assist their balance without providing any meaningful
mechanical support (see Results).

Data processing of gait analysis

Two consecutive and steady-state strides (paretic and nonparetic)
(4 steps) per trial by each participant were analyzed for a total of 3
selected trials for each evaluation. A stride (walking cycle) was
defined by 2 consecutive initial contacts of the same limb with the
ground along the progression line. In addition, foot contacts and
toe-offs during a stride were identified for subsequent calculation
of the spatial-temporal organization of walking for both paretic
and nonparetic sides.

The gait outcome measures analyzed in this study were step
length, calculated as the distance between the initial contact of
each foot along the progression line (determined by the position of
the heel markers); step length symmetry ratio, calculated as the
ratio between the shorter step and the longer step (we were
interested in the magnitude of asymmetry rather than its direc-
tion)26; and single-limb support duration (determined by the time
interval the limb supports total body weight while the contralateral
limb swings). Data analysis was performed using a customized
routine written in MATLAB.f Data from 3 trials for each evalu-
ation were averaged for each participant and considered for
further analyses.

Statistical analysis

For all outcome measures, comparisons between the following
assessments were made: baseline test and 1 week after the last gait
training session and baseline test and 6 weeks after the last
training session. One participant from the overground training
group missed assessment at 6 weeks after the last training session,
and her scores from the 1 week after the last gait training session
assessment were used to replace missing data, a conservative
assumption.27,28 We used a schematic box plot for the primary
outcome measurement to test for the existence of outliers, and we
did not find any outlier. Two-way analysis of variance and
multivariate analysis of variance were used using group (TMG and
overground training group) and time of assessment (baseline test
and 1 week after the last gait training session; baseline test and 6
weeks after the last training session) as factors. The dependent
variables for the analyses of variance were 10MWT, 6MWT, FIM,
FM, and step length symmetry ratio. The dependent variables for
the multivariate analyses of variance were step length and single-
limb support of paretic and nonparetic sides. When necessary,
Tukey post hoc tests were conducted. An a level of .05 was
adopted for all statistical tests, which were conducted using SPSS
version 22.g

Within-group effect sizes (baseline test and 1 week after the
last gait training session; baseline test and 6 weeks after the last
training session) were calculated as the difference in mean values
from each assessment divided by the pooled SD. Effect sizes were
defined using Cohen d classifications (dZ0.2, small; dZ0.5,
medium; dZ0.8, large).29
(maximum, 91)

Baseline FM score

(maximum, 84)

69.2�6.7 70.9�8.6 .83

Abbreviations: H, hemorrhagic; I, ischemic; L, left; OGG, overground

training group; R, right.

* c2 test.
Results

Thirty-two individuals with a mean age of 58.2�9.1 years were
randomized into the study. Random assignment generated groups
that were comparable in terms of age and time poststroke.
www.archives-pmr.org
Twenty-eight participants completed the training protocol in the
allocated group and were included in the final analyses (table 1).

During each set of assessments, 11 participants (5 from the
overground training group, 6 from the TMG) used assistive de-
vices to perform the 10MWT and 6MWT and 8 of these partici-
pants (4 from each group) used light hand assistance for balance
from the same therapist during their gait analyses. All individuals
were interested, motivated, and cooperative throughout the
training period and assessments, and none of them reported any
intervention-related adverse effects. All participants performed all
assessments, except the 6MWT, which 6 individuals did not
perform (3 from the overground training group, 3 from the TMG)
because this test was added to the study after they had begun
training and after these individuals had self-reported walking
longer distances.

Throughout the training period, no differences were observed
between the groups for the percentage of BWS (TMG: 16.81%�
7.62%; overground training group: 14.89%�6.59%) or for session
duration (TMG: 37�6min; overground training group: 36�5min).
However, the mean comfortable speed set for the treadmill and the
servomotor was different (TMG: .27�.07m/s; overground training
group: .52�.07m/s) despite the fact that the groups were equal
during baseline 10m walking.
Clinical evaluations

At 1 week after the last gait training session, both groups
demonstrated improvements in gait speed (TMG: dZ.14; over-
ground training group: dZ.16; PZ.049), 6MWT (TMG: dZ.35;
overground training group: dZ.30; P<.001), FIM (TMG: dZ.41;
overground training group: dZ.27; P<.001), and FM (TMG:
dZ.82; overground training group: dZ.55; P<.001). At 6 weeks
after the last training session, both groups maintained these im-
provements in gait speed (TMG: dZ.25; overground training
group: dZ.32; P<.001), 6MWT (TMG: dZ.36; overground
training group: dZ.33; PZ.001), FIM (TMG: dZ.59; overground
training group: dZ.63; P<.001), and FM (TMG: dZ.99; over-
ground training group: dZ.95; P<.001). However, there were no

http://www.archives-pmr.org


Table 2 Outcome measures of clinical assessments

Variable

t0 t1 t2

Mean � SD 95% CI Mean � SD 95% CI Mean � SD 95% CI

10MWT (m/s)

TMG 0.66�0.29 0.50e0.82 0.70�0.30* 0.52e0.88 0.74�0.34*** 0.56e0.93

OGG 0.69�0.30 0.53e0.85 0.74�0.34* 0.56e0.92 0.79�0.33*** 0.60e0.97

6MWT (m)

TMG 244�119 158e330 291�148*** 201e382 294�159** 199e389

OGG 240�152 154e325 283�139*** 193e373 289�144** 193e384

FIM score (maximum, 91)

TMG 80.4�7.6 76.3e84.4 83.3�6.6*** 79.8e86.8 84.5�6.3*** 81.5e87.5

OGG 82.4�7.1 78.4e86.5 84.2�6.1*** 80.7e87.7 86.1�4.3*** 83.1e89.0

FM score (maximum, 84)

TMG 68.7�7.1 64.3e73.1 74.9�6.7*** 70.7e79.0 75.7�7.0*** 71.9e79.5

OGG 69.4�8.8 64.9e73.8 74.1�8.2*** 70.0e78.3 76.9�6.8 73.1e80.6

NOTE. Statistic effect for t0 � t1 and t0 � t2; *P<.05; **P<.01; ***P<.001.

Abbreviations: CI, confidence interval; OGG, overground training group; t0, baseline; t1, posttraining; t2, follow-up.

742 G.L. Gama et al
differences in improvement between the groups for any of the
clinical outcome measures at either 1 week after the last gait
training session or 6 weeks after the last training session (table 2).

Gait analyses

At 1 week after the last gait training session, both groups
increased nonparetic step length (TMG: dZ.32; overground
training group: dZ.60; P<.001); however, only participants from
the overground training group increased paretic step length
(dZ.32; P<.001) and improved step length symmetry ratio
(dZ1.08; P<.001), and both groups increased single-limb support
duration for the paretic limb only (TMG: dZ.06; overground
training group: dZ.42; PZ.015). At 6 weeks after the last training
session, both groups increased step length of both paretic (TMG:
dZ.36; overground training group: dZ.44; P<.001) and non-
paretic (TMG: dZ.40; overground training group: dZ.56;
P<.001) sides; however, only participants from the overground
training group improved step length symmetry ratio (dZ1.08;
P<.01), and both groups increased single-limb support duration
for the paretic limb only (TMG: dZ.14; overground training
group: dZ.79; PZ.006) (table 3).
Discussion

This study is the first to investigate the effects of time-matched
overground versus treadmill gait training with BWS in individuals
with stroke. We found that both groups improved their gait speed,
endurance, recovery of lower limb motor function impairments,
functional independence, nonparetic step length, and single-limb
support duration for the paretic limb only immediately after the
completion of training and at follow-up. Gait training with over-
ground BWS had an additional benefit of improving step length
symmetry ratio compared with treadmill BWS training. In addi-
tion, overground BWS training promoted immediate and durable
improvement in paretic step length whereas treadmill BWS
training improved this outcome measure only at follow-up.
Therefore, our hypothesis that overground BWS training would
be superior to treadmill BWS training for improvement in walking
performance was supported only for step length symmetry ratio.
Gait speed is one of the most common and important measures
of functional walking ability in the clinical setting30 and is closely
associated with functional independence.31,32 Both our study
groups demonstrated improvements in gait speed immediately
after the intervention and at follow-up. The average improvement
in gait speed of .09m/s demonstrated by both training groups
would be considered a small to substantial meaningful change in
gait speed for individuals with chronic stroke, as defined by Perera
et al.33 This improvement was similar to or greater than that re-
ported in 2 previous studies5,34 of BWS treadmill training that also
included a follow-up but was less than that reported in 3 other
comparable studies.3,35,36 However, all the 3 studies that reported
larger improvements used participants who had lower initial
walking speeds and therefore may have had the potential to
demonstrate greater gains than did our participants. In addition, 2
of these studies35,36 used longer training periods. The only pre-
vious study16 that investigated the use of overground BWS
training reported a similar change in gait speed immediately after
training but did not include follow-up testing. Based on the results
of the present study, it appears that improvements in gait speed are
similar between BWS treadmill training and overground training.
Therefore, if a primary goal is to improve gait speed, clinicians
should rely on clinical judgment to determine which training
method would be most appropriate for an individual patient on the
basis of factors such as patient preference, safety, comfort, and
other therapeutic goals.

Gains in other clinical measures suggest that our BWS para-
digm, in either condition, has benefits other than improving
walking speed. Participants in both groups demonstrated a 50-m
improvement in 6MWT distance immediately after training and at
follow-up. This exceeds the 34.4m that has been suggested as the
minimal clinically important difference for individuals with
stroke.37 It is also greater than, or similar to, that reported in other
treadmill training studies with and without BWS. Lower extremity
impairment, as measured using the FM, also improved by >7
points and exceeded the threshold for a perceived meaningful
recovery (6 points) by individuals with chronic stroke38 and was
greater than the 1.5-point improvement demonstrated in the only
other BWS training study34 reporting FM scores. Finally, the
average improvement of 3.9 points in the motor domain of the
FIM for our participants was also greater than the 1.7-point
www.archives-pmr.org

http://www.archives-pmr.org


Table 3 Parameters of gait analyses

Variable

t0 t1 t2

Mean � SD 95% CI Mean � SD 95% CI Mean � SD 95% CI

Step length (m)

TMG

P 0.40�0.11 0.33e0.47 0.42�0.12 0.35e0.48 0.44�0.11*** 0.37e0.51

NP 0.35�0.11 0.28e0.42 0.39�0.14*** 0.32e0.45 0.40�0.14*** 0.32e0.47

OGG

P 0.39�0.14 0.32e0.46 0.46�0.11* 0.39e0.52 0.45�0.13*** 0.38e0.52

NP 0.37�0.13 0.31e0.44 0.44�0.10*** 0.37e0.51 0.44�0.12*** 0.37e0.51

Step length symmetry ratio

TMG 0.80�0.14 0.73e0.87 0.82�0.13 0.76e0.87 0.81�0.13 0.76e0.86

OGG 0.84�0.11 0.77e0.91 0.93�0.04* 0.87e0.98 0.93�0.04* 0.88e0.98

Single-limb support duration (%)

TMG

P 25.7�8.1 21.6e29.6 26.2�7.6* 22.7e29.8 26.8�7.2** 23.2e30.4

NP 36.1�4.6 32.9e39.3 36.0�4.9 33.0e38.9 36.4�3.3 33.8e39.0

OGG

P 26.0�6.33 22.0e30.0 28.4�4.9* 24.9e31.9 30.8�5.8** 27.2e34.5

NP 36.9�6.8 33.8e40.2 37.6�5.9 34.7e40.6 37.6�5.9 35.0e40.2

NOTE. Statistic effect for t0 � t1 and t0 � t2; *P<.05; **P<.01; ***P<.001.

Abbreviations: CI, confidence interval; NP, nonparetic side; OGG, overground training group; P, paretic side; t0, baseline; t1, posttraining; t2,

follow-up.

Body weightesupported training after stroke 743
change seen in a previous BWS treadmill training study.39 In
summary, the meaningful improvements in each of the clinical
outcome measures support the use of BWS training either on
a treadmill or overground, with no differential benefits
being observed.

The only outcome measure that was different between the
overground training group and the TMG group at follow-up was
step length symmetry ratio. This is an important finding for
several reasons. First, reduced step length symmetry ratio has been
shown to be associated with an increase in fall risk.40 Therefore,
an improvement in step length symmetry ratio may contribute to a
reduction in fall risk during walking. Second, individuals with
stroke often desire to look more “normal”41,42 and having a more
symmetrical gait may contribute to that goal especially when
accompanied by an increase in speed and endurance. Third,
improved symmetry may reduce the energetic cost of walking to
increase the functional range or intensity in performance of ac-
tivities of daily living.43,44 Therefore, if a BWS system is available
for overground walking, it may be more useful than a treadmill-
based system. Although walking on a treadmill by itself will in-
crease symmetry,9 this does not appear to carry over well to
overground walking, as reported in several studies.5,45 Indeed, the
overground requirement to intrinsically generate leg movements
without extrinsic forcing (eg, the moving treadmill surface) may
be an important factor for promoting spatial symmetry.

One caveat to our finding that overground training improved
step length symmetry ratio is that the TMG adopted a lower self-
selected walking speed for training in comparison to the over-
ground training group. It is typical for individuals to adopt a
slower speed on the treadmill than overground, partly because
they initially feel less stable on a treadmill, perhaps owing to the
lack of visual flow and moving floor surface.8,46 Therefore, we
cannot be certain that the improvement in step length symmetry
ratio was a result of the interaction with a stationary floor surface
or from having more repetitions. Because all other variables,
www.archives-pmr.org
including the 10MWT improved similarly across the 2 groups and
did not, therefore, show an effect of any increased repetition, we
believe that the effect of walking overground in this study is the
more likely explanation.

Study limitations

One limitation of the study was the lack of blinding to treatment
by the testers. This could certainly cause bias, although the testers
had no specific a priori expectation of which outcome measures
were expected to improve differentially between the groups. In
addition, the testers were effectively blinded during the gait
analysis and could not, therefore, influence these variables
including step length symmetry ratio. A second limitation was a
lack of a control group, which did not receive any gait training.
The main purpose of the present article was the use of partial
BWS during gait training either on a treadmill (which is the most
common use of BWS for gait interventions) or overground.
Because our participants were in the chronic stage, we made the
assumption that a control group with no training would be un-
likely to show any short-term or follow-up improvements. In
addition, most of our participants presented relatively high mental
and motor functions, but with some of them using assistive de-
vices or hand assistance for balance during the assessments. This
may limit the generalizability to the wider population with stroke.
For future studies, we suggest the addition of a control group and
inclusion of a greater range of deficit severity with randomization
stratified by motor function.

Conclusions

The use of BWS training either on a treadmill or overground
promoted meaningful and durable improvements in gait speed,
walking endurance, lower limb function, functional independence,
and nonparetic step length. However, only overground training led

http://www.archives-pmr.org


744 G.L. Gama et al
to significant improvements in step length symmetry ratio.
Therefore, if BWS gait training is utilized for poststroke gait
rehabilitation, it may be useful to include overground training if
one of the therapeutic goals is to improve step length symme-
try ratio.
Suppliers

a. Model TK35; CEFISE Biotecnologia Esportiva.
b. BWS system; FENIX Tecnologia.
c. LabVIEW; National Instruments Corp.
d. Model 9286BA; Kistler Instruments Corp.
e. VICON; Oxford Metrics Ltd.
f. MATLAB; The MathWorks Inc.
g. SPSS version 22; IBM Corporation.
Keywords

Assistive technology; Clinical protocols; Exercise therapy;
Rehabilitation

Corresponding author

Ana M. Barela, PhD, Institute of Physical Activity and Sport
Sciences, Cruzeiro do Sul University, Rua Galvão Bueno, 868,
São Paulo, SP 01506-000, Brazil. E-mail address: ana.barela@
cruzeirodosul.edu.br.
References

1. Moseley AM, Stark A, Cameron ID, Pollock A. Treadmill training and

body weight support for walking after stroke. Cochrane Database Syst

Rev 2005;(4):CD002840.

2. Hesse S, Bertelt C, Jahnke MT, et al. Treadmill training with partial

body weight support compared with physiotherapy in nonambulatory

hemiparetic patients. Stroke 1995;26:976-81.

3. Sullivan KJ, Knowlton BJ, Dobkin BH. Step training with body weight

support: effect of treadmill speed and practice paradigms on poststroke

locomotor recovery. Arch Phys Med Rehabil 2002;83:683-91.

4. Hesse S. Treadmill training with partial body weight support after

stroke: a review. NeuroRehabilitation 2008;23:55-65.

5. Trueblood PR. Partial body weight treadmill training in persons with

chronic stroke. NeuroRehabilitation 2001;16:141-53.

6. Combs-Miller SA, Kalpathi Parameswaran A, Colburn D, et al. Body

weight-supported treadmill training vs. overground walking training

for persons with chronic stroke: a pilot randomized controlled trial.

Clin Rehabil 2014;28:873-84.

7. Manning CD, Pomeroy VM. Effectiveness of treadmill retraining on

gait of hemiparetic stroke patients. Physiotherapy 2003;89:337-49.

8. 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;13:401-10.

9. Harris-Love ML, Forrester LW, Macko RF, Silver KH, Smith GV.

Hemiparetic gait parameters in overground versus treadmill walking.

Neurorehabil Neural Repair 2001;15:105-12.

10. Norman KE, Pepin A, Ladouceur M, Barbeau H. A treadmill appa-

ratus and harness support for evaluation and rehabilitation of gait.

Arch Phys Med Rehabil 1995;76:772-8.
11. Brouwer B, Parvataneni K, Olney SJ. A comparison of gait biome-

chanics and metabolic requirements of overground and treadmill

walking in people with stroke. Clin Biomech 2009;24:729-34.

12. Lee SJ, Hidler J. Biomechanics of overground vs. treadmill walking in

healthy individuals. J Appl Physiol 2008;104:747-55.

13. Harris-Love ML, Macko RF, Whitall J, Forrester LW. Improved

hemiparetic muscle activation in treadmill versus overground walking.

Neurorehabil Neural Repair 2004;18:154-60.

14. Kautz SA, Bowden MG, Clark DJ, Neptune RR. Comparison of motor

control deficits during treadmill and overground walking poststroke.

Neurorehabil Neural Repair 2011;25:756-65.

15. Miller EW, Quinn ME, Seddon PG. Body weight support treadmill and

overground ambulation training for two patients with chronic

disability secondary to stroke. Phys Ther 2002;82:53-61.

16. Sousa CO, Barela JA, Prado-Medeiros CL, Salvini TF, Barela AM.

Gait training with partial body weight support during overground

walking for individuals with chronic stroke: a pilot study. J Neuroeng

Rehabil 2011;8:48.

17. Barela AMF, Stolf SF, Duarte M. Biomechanics characteristics of

adults walking in shallow water and on land. J Electromyogr Kinesiol

2006;16:250-6.

18. Nascimento LR, Caetano LC, Freitas DC, Morais TM, Polese JC,

Teixeira-Salmela LF. [Different instructions during the ten-meter

walking test determined significant increases in maximum gait speed

in individuals with chronic hemiparesis] [Portuguese]. Rev Bras

Fisioter 2012;16:122-7.

19. Bohannon RW, Andrews AW, Thomas MW. Walking speed: reference

values and correlates for older adults. J Orthop Sports Phys Ther 1996;

24:86-90.

20. Michael K, Goldberg AP, Treuth MS, Beans J, Normandt P,

Macko RF. Progressive adaptive physical activity in stroke improves

balance, gait, and fitness: preliminary results. Top Stroke Rehabil

2009;16:133-9.

21. ATS Committee on Proficiency Standards for Clinical Pulmonary

Function Laboratories. ATS statement: guidelines for the six-minute

walk test. Am J Respir Crit Care Med 2002;166:111-7.

22. Riberto M, Miyazaki MH, Jucá SSH, Sakamoto H, Pinto PPN,

Battistella LR. Validação da versão brasileira da medida de inde-

pendência funcional. Acta Fisiátrica 2004;11:72-6.

23. Riberto M, Miyazaki MH, Sakamoto H, Jorge Filho D, Battistella LR.

Reprodutibilidade da versão brasileira da medida de independência

funcional. Acta Fisiátrica 2000;8:45-52.

24. Maki T, Quagliato EMAB, Cacho EWA, et al. Estudo de con-

fiabilidade da aplicação da escala de Fugl-Meyer no Brasil. Rev Bras

Fisioter 2006;10:177-83.

25. Vicon Motion System Limited. Vicon plug-in-gait product guided

foundation notes revision 2.0. Oxford (UK): Vicon Motion System

Limited; March 2010.

26. Patterson KK, Gage WH, Brooks D, Black SE, McIlroy WE.

Evaluation of gait symmetry after stroke: a comparison of current

methods and recommendations for standardization. Gait Posture 2010;

31:241-6.

27. Jiang L, Xu H, Yu C. Brain connectivity plasticity in the motor

network after ischemic stroke. Neural Plast 2013;2013:924192.

28. Duncan PW. Stroke disability. Phys Ther 1994;74:399-407.

29. Cohen J. A power primer. Psychol Bull 1992;112:155-9.

30. Dickstein R. Rehabilitation of gait speed after stroke: a critical

review of intervention approaches. Neurorehabil Neural Repair 2008;

22:649-60.

31. Schmid A, Duncan PW, Studenski S, et al. Improvements in speed-

based gait classifications are meaningful. Stroke 2007;38:2096-100.

32. Perry J, Garrett M, Gronley JK, Mulroy SJ. Classification of walking

handicap in the stroke population. Stroke 1995;26:982-9.

33. Perera S, Mody SH, Woodman RC, Studenski SA. Meaningful change

and responsiveness in common physical performance measures in

older adults. J Am Geriatr Soc 2006;54:743-9.

34. Middleton A, Merlo-Rains A, Peters DM, et al. Body weight-

supported treadmill training is no better than overground training for
www.archives-pmr.org

mailto:ana.barela@cruzeirodosul.edu.br
mailto:ana.barela@cruzeirodosul.edu.br
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref1
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref1
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref1
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref2
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref2
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref2
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref3
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref3
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref3
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref4
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref4
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref5
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref5
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref6
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref6
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref6
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref6
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref7
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref7
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref8
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref8
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref8
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref8
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref9
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref9
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref9
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref10
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref10
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref10
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref11
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref11
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref11
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref12
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref12
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref13
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref13
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref13
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref14
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref14
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref14
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref15
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref15
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref15
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref16
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref16
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref16
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref16
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref17
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref17
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref17
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref18
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref18
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref18
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref18
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref18
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref19
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref19
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref19
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref20
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref20
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref20
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref20
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref21
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref21
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref21
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref22
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref22
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref22
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref23
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref23
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref23
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref24
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref24
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref24
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref25
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref25
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref25
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref26
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref26
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref26
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref26
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref27
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref27
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref28
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref29
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref30
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref30
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref30
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref31
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref31
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref32
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref32
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref33
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref33
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref33
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref34
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref34
http://www.archives-pmr.org


Body weightesupported training after stroke 745
individuals with chronic stroke: a randomized controlled trial. Top

Stroke Rehabil 2014;21:462-76.

35. Combs SA, Dugan EL, Ozimek EN, Curtis AB. Effects of body-

weight supported treadmill training on kinetic symmetry in persons

with chronic stroke. Clin Biomech 2012;27:887-92.

36. 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;29:1122-8.

37. Tang A, Eng JJ, Rand D. Relationship between perceived and measured

changes in walking after stroke. J Neurol Phys Ther 2012;36:115-21.

38. Pandian S, Arya KN, Kumar D. Minimal clinically important differ-

ence of the lower-extremity Fugl-Meyer assessment in chronic-stroke.

Top Stroke Rehabil 2016;23:233-9.

39. Ribeiro T, Britto H, Oliveira D, Silva E, Galvao E, Lindquist A. Ef-

fects of treadmill training with partial body weight support and the

proprioceptive neuromuscular facilitation method on hemiparetic gait:

a comparative study. Eur J Phys Rehabil Med 2013;49:451-61.

40. Balasubramanian CK, Bowden MG, Neptune RR, Kautz SA. Relation-

ship between step length asymmetry and walking performance in sub-

jects with chronic hemiparesis. Arch Phys Med Rehabil 2007;88:43-9.
www.archives-pmr.org
41. Hsu AL, Tang PF, Jan MH. Analysis of impairments influencing gait

velocity and asymmetry of hemiplegic patients after mild to moderate

stroke. Arch Phys Med Rehabil 2003;84:1185-93.

42. Couillandre A, Maton B, Breniere Y. Voluntary toe-walking gait

initiation: electromyographical and biomechanical aspects. Exp Brain

Res 2002;147:313-21.

43. Nilsson L, Carlsson J, Danielsson A, et al. 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;15:515-27.

44. Da Cunha IT Jr, Lim PA, Qureshy H, Henson H, Monga T, Protas EJ.

Gait outcomes after acute stroke rehabilitation with supported tread-

mill ambulation training: a randomized controlled pilot study. Arch

Phys Med Rehabil 2002;83:1258-65.

45. Combs SA, Dugan EL, Ozimek EN, Curtis AB. Bilateral coordination

and gait symmetry after body-weight supported treadmill training for

persons with chronic stroke. Clin Biomech 2013;28:448-53.

46. Bayat R, Barbeau H, Lamontagne A. Speed and temporal-distance

adaptations during treadmill and overground walking following

stroke. Neurorehabil Neural Repair 2005;19:115-24.

http://refhub.elsevier.com/S0003-9993(16)31329-6/sref34
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref34
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref35
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref35
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref35
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref36
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref36
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref36
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref37
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref37
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref38
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref38
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref38
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref39
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref39
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref39
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref39
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref40
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref40
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref40
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref41
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref41
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref41
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref42
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref42
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref42
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref43
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref43
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref43
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref43
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref44
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref44
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref44
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref44
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref45
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref45
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref45
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref46
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref46
http://refhub.elsevier.com/S0003-9993(16)31329-6/sref46
http://www.archives-pmr.org

	Effects of Gait Training With Body Weight Support on a Treadmill Versus Overground in Individuals With Stroke
	Methods
	Setting and participants
	Randomization and blinding
	Intervention protocols
	Outcome measures and follow-up
	Data processing of gait analysis
	Statistical analysis

	Results
	Clinical evaluations
	Gait analyses

	Discussion
	Study limitations

	Conclusions
	Suppliers
	References