Treadmill Training with Partial Body Weight Support:

Influence of Body Weight Release

on the Gait of Hemiparetic Patients

1 S. Hesse, MD, 1 B. Helm, MD, 2

J. Krajnik, MS,
1 1
M. Gregoric, MD, and K.H. Mauritz, MD

Objective: To investigate the effect of body weight support on the gait of hemiparetic
subjects walking on a treadmill. Design: Survey. Patients: Eleven nonambulatory hemi-
paretic subjects. Methods: Subjects walked on the treadmill with full weight bearing
and with 15%, 30%, 45%, and 60% body weight support at constant walking veloc-
ity. Cycle parameters, symmetry ratios, and the kinematic EMG of several lower limb
muscles of the affected leg were recorded. Video-analysis served for assessment of pos-
ture and hip and knee angle displacement. Results: With body weight support the rel-
ative double support time decreased, the relative single stance period of the affected
limb increased, and the functional activity of the vastus and soleus diminished. The
activity of the gluteus tended to increase. Patients walked more upright and with less
hip and knee flexion. The extensor spasticity did not change and the qualitative acti-
vation pattern of all recorded muscles remained unchanged. Conclusion: Body weight
support did not less physiological gait. By reducing double support dura-
facilitate a

tion, body weight support resulted in a greater stimulus for balance training. The facil-
itation of gluteus medius is favorable with respect to training pelvic alignment. The
reduction of the activity of other antigravity muscles suggests a limit of 30% BWS
not be exceeded. Key Words: Hemiparesis—Gait—Treadmill training.

Introduction nonnmhulatory, hemiparetic suhjects, a

Trcalmill training with martial body weight: support
is a new treatmentapproach to gait rehabilitation of non-
ambulatory paraparetic ( 1-4) and 11cI11lparcClC subjects
(5-8). As a task-specific training it allow practicing com-
plete gait cycles with many repetitions instead of shingle
elements or preparatory maneuvers at an early stage of
gait rehabilitation.
For chronic,
muttipte-basehne study and two single case design stul-
iea (6-b’) showed that treadmill training with partial body
weight suppurt was more effective in resturation of gait
abitity and walking vclocity when compared to regular
physiotherapy. In these studied, the amount of body
weight support (BWS) was set according to clinical cri-
teria and varied from 5% to 45% of body weight.
There is littlc knowledge on the effect of body
weight support on the gait of hemiparetic patients walk-
ing on a treadmill. In normal subjects, Finch and cowork-
ers reported decreases in percentage of stance, tutal dou-

ble-limb support time, and maximum hip and knee flexor

From 1Klinik Berlin, Department of Neurological Rehabilitation, Free
University Berlin; and 2 Rehabilitation Institute, Ljubljana, Slovenia.
Address correspondence and reprint requests to: Stefan Hesse, M.D.,
Klinik Berlin, Kladower Damm 223, 14089 Berlin, telephone
49/30/36503-105, fax 49/30/36503-222.
swing tingle when walking with body weight support as
compared to full weight bearing. Other adaptations to
body weight suppurt were a reduction in the mean burst
amplitude of the muscles that are active during stance
15

Downloaded from nnr.sagepub.com at PENNSYLVANIA STATE UNIV on March 5, 2016

and an increase in the mean burst amplitude of the tib-
ialis anterior muscle (9).
This study investigated the influence of various
degrees of body weight support as compared to full weight
bearing on the gait pattern of hemiparetic subjects walk-
ing on the treadmill.
The study should help to define a sensible range of
body weight support while training hemiparetic subjects
on the treadmill and should also deal with objections of
some physiotherapists who fear that the support may
result in a &dquo;less physiological&dquo; gait pattern in hemiparetic
patients, mainly characterized by increased plantarflexor
spasticity.
Methods
Subjects
The study, which was approved by the local ethics
committee, included I I 11en1lpareCIC subjects (six men,
five women; mean age 51 years with a range from 28 to
64 years; six patients suffered from right and five patients
from left hemiparesis). Etiology was ischemia in the ter-
ritory of the middle cerebral artery in nine and supra=
tentorial intracranial hemorrhage in two.
During ground level walking, seven patients still
needed intermittent support of one person to help with
balance and four patients required verbal supervision
or standby help from one person without physical con-
tact. None of them suffered from additional orthopedic
or neurologic gait impairments. Severe disturbances of
cognition or communication were additional exclusion
criteria.
All subjects were habituated to the experimental
protocol by walking for 20 minutes on a motor driven
treadmill (variable speed control, speed range 0.01-2.25
m/s) unsupported (FWB, full weight bearing) and with
various degrees of body weight support (BWS) at their
preferred speed. All patients who participated in the
study could walk unsupported and with a BWS of up to
60% on the treadmill at constant speed, which was indi-
vidually chosen. When walking unsupported, the
patients remained secured in the harness. A higher
degree of BWS was either not tolerated or required a
reduced walking velocity. If necessary, swinging of the
paretic limb was assisted by a therapist sitting by the
paretic side in four patients, the therapist was instructed
to assist as little as possible and not to change the degree
of assistance with BWS. Otherwise, the patients
excluded from the study.
16
Downloaded from nnr.sagepub.com at PENNSYLVANIA STATE UNIV on March 5, 2016
Body Weight Support System
Each subject was mechanically supported in a nod,
ified parachute harness suspended by a set Of pulleys. The
harness supported the subjects primarily about the pelvis,
lower abdomen, and chest to avoid interfering with lower
limh movement. The system provided a preset degree of
body weight support controlled by force transducers
located between the harness and the pulleys.
ExperimentnL Protocol
All suhjects wore shorts and their preferred shoes
without ankle-foot orthoses. Patients walked on the
motor-driven treadmill unsupported (FWB) and with a
BWS of 15%, 30%, 45%, and 60%. The individually cho-
sen walking velocity was kept constant during all trials;
it ranged from 0.22 to 0.35 111~5. The constant gait veloc-
ity allowed isolation of the effect of BWS on the gait pat-
tern of hemiparetic suhjects. To avoid fatigue, patients
rested on a chair placed on the treadmill hetween each
trial.
Gait AnaLysis
Gait analysis was performed for each loading con-
dition (i.e., FWB, BWS 15‘%,, 30%, 45%, 60%). Cycle
parameters (cadence, stride length, step length, stance,
swing, and double support durations) were recorded with
the help of overshoe slippers incorporating conductive
areas in the region of the heel and forefoot (the walk-
ing surface of the treadmill was conductive as well).
Symmetry ratios for stance, swing and step-length, and
relative durations of double support (% cycle) and sin-
gle stance of the affected limb (% stance affected) werc
calculated.
Electromyographic activity was detected by Ag-
AgCl surface electrodes (diameter, 8 mm), which were
attached 2 cm apart on the muscle bellies (output signal
sampled 800 Hz) after conventional skin preparation.
The impedance was checked and kept below 10 kQ.
Recordings were obtained for the following muscles on
the subjects’ paretic side: tibialis anterior, soleus, biceps
femoris, vastus lateralis, gluteus medius, and erector
spinae. Signals were preamplified with standard elec-
troencephalographic Oxford Metric preamplifiers (type
8515) attached to the limb.
All signals (i.e., contacts, EMG measurements) were
gathered and amplified to the range of 0-5 V by a data
were logger worn by the patient, and signals were transmitted
via a cable to a personal computer. The data were sam-
 pleu using a Keithley 64-channel card (type 1801) and
displayed online to ensure that the sampled data were uf
acceptable quality. Self-adapted Asyst software was used
to archivc data.
EMG data were digitally filtered (band-pass, 10 to
300 Hz), rectified, averaged over at least 12 strides, and
time-normalized to the mean cycle duration set to 100%.
To quantity the physiological and functional activities
of the recorded muscles, mean values of the non-low
passed signals were calculated over the following inter-
vals of the cycle duration, set to 100% (according to the
adult muscle phasic activity chart published by Shriner’s
Hospital, San Francisco) (10): from 60% to 100% (tib-
°
ialis anterior), from 20% to 50% (soleus), from 80% to
110% (biceps femoris), from 90% to 130% (vastus later-
alis), from 90‘%~ to 140°~, (gluteus medius), and from 0%
to 20% (erector spinae). For the soleus, the so-called pre-
mature activity was also calculated in the interval from
90% to 110%. Electromyographic data were gathered
within one hour, so that the comparison of absolute EMG
activities within subjects was appropriate assuming stahle
rccording conditions.
Additionally, the patients were videotaped during all
trials for
subsequent qualitative assessment of posture,
maximum hip and knee flexor swing, and extensor stance
angles and mode of initial contact of the paretic limb by
two independent raters.
Statisticnl Analysis
For comparison of different loading conditions within
patients, pairwise Wilcoxon tests were performed with an
Table 1. Gait cvcle pnramemrs LlI1J s)’11I 11IC 0&dquo;)’ rutu~s
Means ± SD; FWII: full weight hearing; ~‘: ;ignificant difference
Downloaded from nnr.sagepub.com at PENNSYLVANIA STATE UNIV on March 5, 2016
adjusted alpha value (p < 0.01) whenever the preceding
Friedman two-way analysis of variance was significant.
5~
Results
Tempo7-al Data
The mean cycle parameters are listed in Table 1.
BWS had no effect on cadence, stride length, stance, and
swing duration. There was a trend toward decrease of
stance and increase of swing duration. The relative dou-
ble support declined constantly with BWS, and the val-
ues with 45% and 60% BWS differed significantly from
those with FWB. Thc relative single stance period of the
affected limh showed an opposite trend: the values
increased continuously with BWS and differed signifi-
cantly from FWB with 45% and 60% BWS.
Sy7ni7iet7--v Rations
The mean symmetry ratios for stance, swing, and step
length are listed in Table 1. BWS had no effect on them.
Electromyographic Activity
The mean values of activity are presented in Table 2.
BWS had no effect on the activity of the tibialis anterior,
hiceps femuris, and erector spinae. The functional activ-
ity of thc soleus muscle (20-50%) decreased cuntinuuusly
versus full weight hearing (p < 0.01 )
17
 Table 2. Muscular acci1’Ît)’
Means ± SD; FWB: full weight: hearing; *: significant difference versus full weight hearing (p < 0.01 )
with BWS, and the values with 45% and 60% BWS dif-
fered significantly from FWB. The premature activity of
the soleus muscle (90-110%) did not change with BWS..
BWS consistently reduced the activity of vastus through-
out all BWS trails compared to FWB (Figure 1). For the
gluteus medius, there was a trend toward increased activ-
ity up to a BWS of 45% (Figure 2). The qualitative pat-
tem of activation remained unchanged in all muscles.
Hip and Knee Angle D!Sp!acement
With BWS, patients tended to walk with a more
upright posture reflected by an increased hip and knee
extension at mid-stance. At the same time BWS resulted
in less hip and knee flexion during swing. In some
patients the mode of initial contact of the paretic limb
changed from forefoot or flat foot with FWB to planti-
grade or heel contact. With BWS of 60% or even 45%,
patients tended to walk tiptoe, particularly with the non-
affected limb.
Discussion
Spatio-Temporal Parameters
With BWS, relative double support decreased con-
stantly, resulting in a decline of the stance and a prolon-
gation of the swing period of both limbs. For the paretic .
18
Downloaded from nnr.sagepub.com at PENNSYLVANIA STATE UNIV on March 5, 2016
limb, the relative single stance became significantly
longer. During this
period the body weight is carried and
balanced by the paretic limb so that its extension might
have provoked a greater need fur balance during tread-
mill training of hemiparetic subjects with BWS.
When walking at ground level, patients with an
equivalent gait impairment and walking velocity spent
more time on both feet (34-S0(){), cf. Table 3) (11) as
compared to our patients walking either unsupported or
with BWS on the treadmill. The moving surface itself
and even to a larger extent BWS seem to dynamize the
gait pattern of hemiparetic subjects, thereby providing
a high training stimulus for impaired equilibrium
reflexes.
The symmetry of gait remained constant with BWS.
If BWS produced dysfunctional dynamic gait patterns,
then the quality and safety of gait would have suffered.
This did not occur. The harness provided the patients
with a feeling of security.
The observed gait phase changes are in accordance
with a study by Finch et al., who reported a significant
reduction of stance and double support with BWS in nor-
mal subjects walking on a treadmill (9). The authors
argued that BWS lifted the subjects up and minimized
the vertical displacement of the center of mass, thus
reducing the ground contact times of both feet.
This notion is supported by the observed kinematic
changes in this study: the patients walked with a more
upright posture and flexed their hip and knee less dur-
ing swing. Some subjects also changed their mode of ini-
tial contact from either fore- or flatfoot with FWB to
Figure 1. Averaged and with respect to the gait cycle normal-
ized activity of the left M. vastus lateralis of a left hemiparetic
subject walking on the treadmill full weight bearing (FWB) and
with 30% and 60% body weight support (BWS). The basogram
symbolizes the foot contact phases of the left lotver limb.
plantigrade or heel contact with BWS up to 45%, result-
ing in a more physiological manner of weight acceptance.
With 60% BWS however, patients tended to walk tiptoe.
This amount of weight support must therefore be regarded
as unfavorable.
Downloaded from nnr.sagepub.com at PENNSYLVANIA STATE UNIV on March 5, 2016
Figure 2. Averaged and with respect to the gait cycle normal-
ized activity of the right M. gluteus medius of a right hemi-
paretic subject walking on the treacllnill full weight bearing
(FWB) and with 30% and 60% body weight support. The baso-
gram symbolizes the toot contact phases of the right lower limb.
Electromyogr-nphic Activity
With BWS, the functional activity of soleus and
vastus lateralis decreased constantly with significant
changes at a BWS of more than 30%. Both muscles are
19
active during stance (mainly as antigravity muscles) so
that the lowering of their mean activity reflects the
reduction of body weight, as was also shown by Finch et
al. for normal subjects (9). To enable efficient training
of weight-bearing ability, the support should therefore
not be too large. Our data suggest a limit of 30% BWS
not be exceeded when treating hemiparetic subjects.
The activity of gluteus medius, which is also active
during stance to assist pelvic alignment, increased in
some patients with BWS. A facilitation of its activity
might be explained by a stretch of the muscle due to a
more upright posture (with a backward pelvic tilt)
and/or the extension of the single-stance period of the
paretic limb.
The premature activity of soleus did not increase
with BWS. This stretch-sensitive activity is regarded as
an important cause of excessive plantar tlexion and varus
in stroke patients suffering from plantar flexor spasticity
(12,13). It may therefore be concluded that BWS did not
facilitate extensor spasticity, a concern often voiced by
physiotherapists in personal discussions. On the contrary,
an improved mode of initial contact (plantigrade versus
forefoot) with moderate BWS may even give less oppor-
tunity for rapid stretch of the calf muscles during the load-
ing phase. Further, the activation pattern of all muscles
did not change qualitatively, also indicating that BWS
did not facilitate gait dysfunction in hemiparetic subjects.
BWS had no effect on the activity of the erector
spinae, which decrease the forward and lateral motion
of the trunk during heel strike (14,15). BWS therefore
seems not to alter trunk-leg gait cycle parameters for
hemiparetic subjects. Normal subjects, however, as
reported by Finch and co-wurkers, may show a gradual
decrease in erector spinae activation with increasing
BWS (9).
Acknowledgments. The study was supported by the
Deutsche Forschungsgemeinschaft (HE 2640/1-1). The
authors thank Mrs. Daniela Luecke, Mrs. Christine
Schwartz, and Mr. Frank Reiter for technical assistance.
20
Downloaded from nnr.sagepub.com at PENNSYLVANIA STATE UNIV on March 5, 2016
References
1. Visintin M, Barbeau H. The effects of body weight support on the
locomotor pattern of spastic paretic patients. Can J Neurol Sci
1989;16:315-25.
2. Dobkin BH, Edgerton VR, Fowler E, Hodgson JA. Training
induces rhythmic locomotor EMG patterns in subjects with com-
plete spinal cord injury. Neurology 1992;42(Suppl 3):207-208.
3. Wernig A, Muller S. Laufband locomotion with body weight sup-
port in persons with severe spinal cord injuries. Paraplegia 1992;
30:229-38.
4. Dietz V, Colombo G, Jensen L, Baumgartner L, Locomotor capacity
of spinal cord in paraplegic patients. Ann Neurol 1995;37:574-82.
5. Richards CL, Malouin F, Wood-Dauphinee S, Williams JI,
Bouchard JP, Brunet D. Task-specific physical therapy for opti-
mization of gait recovery in acute stroke patients. Arch Phys Med
Rehabil 1993;74:612-20.
6. Hesse S, Bertelt C, Schaffrin A, Malezic M, Mauritz KH. Restora-
tion of gait in non-ambulatory hemiparetic patients by treadmill
training with partial body weight support. Arch Phys Med Rehabil
1994;75:1087-93.
7. Hesse S, Bertelt C, Jahnke MT, Schaffrin A, Baake P, Malezic
M, Mauritz KH. Treadmill training with partial body weight sup-
port as compared to physiotherapy in non-ambulatory hemiparetic
patients. Stroke 1995;26:976-81.
8. Hesse S, Malezic M, Schaffrin A, Mauritz KH. Restoration of gait
by a combined treadmill training and multichannel electrical stim-
ulation in non-ambulatory hemiparetic patients. Scand J Rehab
Med 1995;27:199-205.
9. Finch L, Barbeau H, Arsenault B. Influence of body weight sup-
port on normal human gait: development of a gait retraining strat-
egy. Phys Ther 1991;71:842-53.
10. Perry J. Gait analysis. Thorofare, NJ: Slack Inc., 1992.
11. Brandstater ME, De Bruin H, Gowland C, Clark BM. Hemiplegic
gait: analysis of temporal variables. Arch Phys Med Rehabil
1983;64:583-87.
12. Perry J, Waters RL, Perrin T. Electromyographic analysis of
equinovarus following stroke. Clin Orthop 1978;131:47-53.
13. Hesse S, Krajnik J, Luecke D, Jahnke MT, Gregoric M, Mauritz
KH. Ankle muscle activity before and after botulinum toxin ther-
apy for lower limb extensor spasticity in chronic hemiparetic
patients. Stroke 1996;27:455-60.
14. Thorstensson A, Carlson H, Zomlefer MR, Nilsson J. Lumbar
back muscle activity in relation to trunk movement during loco-
motion in man. Acta Physiol Scand 1982;116:13-20.
15. Thorstensson A, Nilsson J, Carlson H, Zomlefer MR. Trunk move-
ments in human locomotion. Acta Physiol Scand 1984;121:9-22.