Clinical Rehabilitation 1999; 13: 295300

Trunk kinematics in hemiplegic gait and the effect

of walking aids
Sarah F Tyson Centre for Research in Rehabilitation, Department of Health Studies, Brunel University, Isleworth, UK

Received 2nd June 1998; returned for revisions 7th August 1998; revised manuscript accepted 30th January 1999.

Objective: To establish baseline measurements of trunk movements during

hemiplegic gait, to assess the relationship between trunk movements and
walking ability, and to investigate the effect of walking aids on the trunk
movements.
Method: Twenty subjects with a chronic hemiplegia from a stroke, who could
walk independently, were recruited. Lateral and vertical movements of the
pelvis, and symmetry of these movements were measured using CODA (a
three-dimensional movement analysis system) as the subjects walked at their
own pace without an aid. They were also tested as they walked with a stick
and a tripod to assess the effect of different walking aids. Mean values for
the trunk movements and symmetry were calculated, Pearsons correlations
assessed the relationship between each trunk movement and gait velocity (a
measure of overall walking ability), and the influence of the different aids was
assessed using a one-way repeated measures ANOVA.
Results: Lateral displacement was large (mean = 9.9 cm, SD 3.9) and
orientated to the sound side, vertical displacement was small
(mean = 2.45 cm, SD 1.4). The movements showed marked asymmetry which
favoured the hemiplegic side in that there was less movement of, or towards
this side. There was a significant relationship (at 5% level) between walking
ability and lateral movements (r = 0.6), but not vertical movements (r = 0.41).
No significant differences were found with the different aids.
Conclusion: These results give baseline values for trunk movements during
hemiplegic gait and the relationship between the movements and walking
ability. The use of walking aid and the type of walking aid did not affect the
subjects trunk movements or walking ability.

Introduction physiotherapists working with stroke patients as

Restoration of normal movements of the trunk
and pelvis while walking is a primary goal for
Address for correspondence: Sarah F Tyson, Centre for
Research in Rehabilitation, Department of Health Studies,
Brunel University, Borough Road, Isleworth, Middlesex
TW7 5DU, UK. e-mail: sarah.tyson@brunel.ac.uk
Arnold 1999
these movements are considered an essential
component of effective gait. Despite the impor-
tance attached to them there has been little
objective assessment of these movements as
research has concentrated on movements of the
hip, knee and ankle in the sagittal plane.1,2 In nor-
mal gait, the trunk and pelvis make a number of
small, symmetrical movements in the frontal
02692155(99)CR259OA
296 SF Tyson
plane with each gait cycle. The trunk and pelvis
displaces laterally about 5 cm (2.5 cm to each
side), and vertically about 5 cm (with the great-
est excursion at mid-stance and the least at dou-
ble stance).39 Using a walking aid is thought to
have a detrimental effect on these movements
and walking ability by encouraging the patient to
lean over to the sound side, and to move their
weight towards the aid and away from the weak
side.10 For this reason most physiotherapists are
reluctant to use walking aids in stroke rehabili-
tation.11 Yet there has been very little investiga-
tion of the effects of walking aids on hemiplegic
gait. The aim of this study was to measure
objectively trunk movements including symmetry
during hemiplegic gait, to assess the relationship
between the movements and walking ability, and
investigate the effect of different walking aids on
trunk movements.
Method
There were two parts to the study. First, the
trunk movements of a group of hemiplegic sub-
jects were measured while walking. The symme-
try of these movements, and the relationship
between the movements and walking ability was
assessed. Secondly, the effects of walking aids on
trunk movements, symmetry and walking ability
were assessed.
The parameters
Lateral and vertical movements of the trunk
were measured and velocity was used as an over-
all measure of walking ability.
Two aspects of lateral movement were mea-
sured; the total excursion of the pelvis from side
to side with each stride (termed lateral displace-
ment) and the movement to each side during
stance phase of the weight-bearing leg (termed
lateral shift). Lateral displacement was defined as
the lateral movement of the pelvis between mid-
single stance of one leg and mid-single stance of
the other. Lateral shift (to the weak or sound
side) measured the excursion of the pelvis during
weight-bearing on one leg. It was defined as the
lateral distance moved between the mid-point of
double stance (the minimum point in the gait
cycle) to the mid-point of single stance (the max-
imum point in the gait cycle) for each leg.
Vertical movement while stepping with each
leg was measured. This was defined as the verti-
cal movement of the pelvis between double leg
stance (the lowest point in the gait cycle) and
mid-stance on each leg (the highest point in the
gait cycle). Symmetry of these movements was
assessed by comparing the values for the sound
and weak sides.
Velocity was used as an overall measure of
walking ability as it is sensitive, valid and reliable
for stroke patients.12 Subjects with a high veloc-
ity were considered good walkers while slower
subjects were poor walkers.
The testing procedure
The subjects trunk movements and velocity
were measured as they walked using CODA.
This is a noninvasive three-dimensional move-
ment analysis system, which uses optical scanning
techniques to detect the position of small, pris-
matic, reflective markers. These were attached to
the subjects heels (to identify phases of the gait
cycle) and the pelvis (mid-point between the
posterior iliac spines). Previous studies have
shown CODA to be accurate and reliable when
measuring angular and spatial movement.13,14 The
subjects walked at their own pace without a walk-
ing aid in ordinary shoes along a 5 m walkway as
the CODA measured their gait. They started to
walk 1 m before the start of data collection, and
continued for 1 m afterwards so that data was not
collected while the subject was accelerating or
decelerating. This was repeated three times and
mean values were calculated.
In the second part of the study, when the effect
of walking aids was investigated the process was
repeated as the subjects walked without an aid,
with a walking stick, and with a tripod. The order
in which the walking aids were used was ran-
domized to nullify any effect from practice or
fatigue. Subjects practised with each aid until
they felt comfortable before measurements took
place and they were encouraged to take rest
whenever they wished. The aids were adjusted to
fit each individual so that the handle of the aid
was level with the radial styloid of the sound arm.

Analysis
Mean values for each of the gait parameters
were calculated. Symmetry was evaluated using a
ratio of the sound and weak sides (sound/weak),
a value of one would indicate perfect symmetry.
Pearsons correlations were used to assess the
relationship between the trunk movements and
walking ability (velocity). The significance of the
correlation was assessed by comparison with the
sample correlation coefficient (2 multiplied by
the reciprocal of the square root of the sample
size). The correlation coefficient was said to be
significant at the 5% level if it exceeded (posi-
tively or negatively) the sample coefficient.15 In
this case the sample size was 20 so the sample
coefficient was 0.46. Differences between the
walking aids were assessed using a one-way
repeated measures ANOVA. Tukey multiple
comparison tests would be used to establish
where any significant differences lay.
The subjects
Subjects with a chronic hemiplegia due to
stroke (more than three months duration), able
to walk at least 10 m independently without a
walking aid, and aged over 50 years were
recruited. Exclusion criteria were a history of
previous stroke, a pre-existing gross neuro-
pathology, or other condition that affected their
mobility. People with aphasia or cognitive deficits
that prevented comprehension of the task were
also excluded. Twenty subjects were recruited, 15
men and five women, 11 had a right hemiplegia
and nine had a left hemiplegia. Median time since
stroke was 10 months (range 3144). Mean age
was 61 years (SD 6.5). Five subjects did not usu-
ally use an aid, and 15 subjects usually used a
walking stick.
Results
The mean movements of the trunk are shown in
Table 1. Lateral displacement was large and ori-
entated over the sound side with little movement
to the weak side, and vertical displacement was
small, more so on the weak side than the sound
side. The subjects walked slowly with marked
asymmetry.
The correlation between walking ability and
Trunk kinematics in hemiplegic gait 297
the trunk movements (Table 2) showed signifi-
cant relationships with lateral displacement (indi-
cating that good walkers had a low lateral
displacement, and were orientated centrally), and
lateral symmetry (indicating that good walkers
were more symmetrical). The relationship
between walking ability and vertical displace-
ment, and symmetry of vertical displacement was
not significant however.
No significant differences in any of the para-
meters (trunk movements, symmetry or velocity)
between the stick, the tripod or no aid were
found (Table 3).
Discussion
This study reports baseline values for trunk
movements including symmetry during hemi-
plegic gait, and considers the relationship
between the movements and walking ability, and
the effect of walking aids. Trunk movements
were found to be markedly asymmetrical and dif-
fered from the classic, general values of Saun-
ders, Inman and co-workers,3,4 as well as more
specific analysis of trunk movements including a
Table 1 Mean values for movements of the trunk during
hemiplegic gait
Total lateral displacement 9.8 cm (3.9)
Lateral shift (sound side) 6.4 cm (3.9)
Lateral shift (weak side) 1.2 cm (4.3)
Lateral symmetry ratio (S/W) 5.3 (1.2)
Total vertical displacement 2.45 cm (1.4)
Vertical displacement (sound side) 3.4 cm (1.1)
Vertical displacement (weak side) 1.4 cm (0.9)
Vertical symmetry ratio (S/W) 3.5 (2.5)
Velocity 0.6 m/s (0.31)
Values in parentheses are standard deviations.
Table 2 Relationship between walking ability (velocity)
and trunk movements
Correlation coefficient (r)
Lateral displacement 0.6*
Lateral symmetry 0.6*
Vertical displacement 0.41
Vertical symmetry 0.07
*Denotes a significant relationship at 5% level. Sample
coefficient = 0.46.
298 SF Tyson

Table 3 Mean values for velocity and trunk movements when walking with different walking aids, and the results of
the ANOVA to elicit differences between the walking aids

No aid Stick Tripod p-value

Velocity (m/s) 0.6 (0.31) 0.5 (0.2) 0.5 (0.2) 0.33

Lateral displacement (cm) 9.9 (3.8) 8.5 (3.2) 8.7 (3.6) 0.45
Lateral shift to weak side (cm) 1.2 (4.3) 0.8 (4.2) 0.9 (4.4) 0.95
Lateral shift to sound side (cm) 6.4 (3.9) 5.4 (6.3) 6.1 (4.3) 0.8
Symmetry 5.0 (1.2) 5.0 (2.6) 6.0 (1.9) 0.97

Total vertical displacement (cm) 2.4 (1.4) 2.3 (1.3) 2.3 (1.5) 0.87
Weak side 1.5 (0.9) 1.3 (0.8) 1.3 (0.8) 0.73
Sound side 3.4 (1.1) 3.3 (1.1) 3.4 (1.2) 0.9
Symmetry 3.5 (2.5) 6.3 (1.5) 5.5 (3.0) 0.54

Values in parentheses are standard deviations.

similar age range as this study.8,9 Lateral move-
ments were related to walking ability but vertical
movements were not. The use of an aid and the
type of aid did not have any significant effect on
trunk movements, symmetry or walking ability.
The large lateral movements and orientation to
the sound side support clinicians observations
and previous studies which have found a short
stance phase and decreased weight-bearing on
the weak leg.10,1620 In normal subjects lateral dis-
placement is controlled by eccentric activity of
the hip abductors21,22 so it may be that the
changes in lateral displacement found in this
study are due to poor control of the abductors.
EMG activity of the hip abductors during move-
ments in the frontal plane has not been studied
to date in hemiplegics, and is an obvious area for
further research.
Normal subjects show a sinusoidal vertical dis-
placement of approximately 5 cm during gait,
which is dependent on horizontal rotation of the
pelvis at the hip as the hip extends in stance
phase, lateral tilt of the pelvis to the nonweight-
bearing side, and flexion of the knee during
stance3,4 and is powered by concentric activity of
the plantar flexors in late stance and at push-
off.22,23 Alteration in the vertical movements
could be due to any or all of these factors, so
decreased vertical displacement on the weak side
is unsurprising as a lack of hip extension and loss
of knee flexion during stance phase is one of the
most consistent kinematic findings in hemiplegic
gait1921 while increased pelvic tilt (which would
lower of the centre of gravity during stance) has
been observed but not objectively measured.10
The decreased vertical displacement on the
sound side was unexpected, however. An appar-
ent elevation of the pelvis (hip hitching and cir-
cumduction) has frequently been observed to
compensate for difficulty in flexing the weak hip
and knee during swing phase.10 In normal sub-
jects the muscle forces produced to propel the
body forwards and upwards are relatively low
(much of the movement comes from momentum
and energy transfer), but the plantar flexors are
key in generating these forces and initiating flex-
ion of the knee during swing phase.23 Lack of
knee flexion during swing phase is a common
kinematic finding in hemiplegic subjects and
reduced effectiveness has been found in the ankle
musculature.19,24 It is possible that the lack of
effective plantar flexion in late stance phase of
the weak leg reduced the vertical displacement
on the sound leg. A more detailed biomechani-
cal analysis including lateral tilt of the pelvis and
using EMG analysis of muscle activity is needed
to clarify these findings and the reasons for the
movement abnormalities which could inform
physiotherapy practice.
Walking ability was found to be significantly
related to lateral movement, in that good walk-
ers showed less movement, were more symmet-
rical and orientated more centrally than poor
walkers who orientated over the sound leg, but
no such relationship was found with vertical
movements. This may be a reflection of the small
numbers of subjects in this study, but a number
of other studies have also found that the walking

ability is not clearly related to gait parameters in
chronic hemiplegics, suggesting that chronic
patients adopt abnormal but none the less func-
tionally effective gait patterns.2527 Further
research with larger numbers could clarify these
findings and establish the importance of symme-
try and normal movement to function which
would inform physiotherapy practice.
The other aim of the study was to investigate
the effect of walking aids on trunk movements.
In this study, neither the use of an aid (compared
to no aid) nor the type of aid effected the trunk
movements. There has been relatively little study
of the effects of walking aids on hemiplegic gait,
but the other studies to date have also failed to
find that the use of an aid or the type of aid sig-
nificantly influenced hemiplegic gait. A number
of different aspects of gait have been assessed,
including: temporaldistance factors, gait sym-
metry, ground reaction forces during stance
phase, hip kinematics, muscle activation patterns
and the amount of support taken through the
aid.2731 This fails to support the belief that the
use of an aid has a detrimental effect on hemi-
plegic gait and that a tripod produces a worse gait
pattern than a stick10 and the widespread reluc-
tance for physiotherapists to recommend use of
a walking aid.11 Despite physiotherapists antipa-
thy, most ambulant stroke patients choose to use
a walking aid, although many use it intermit-
tently, for example only when walking out-of-
doors or for long distances. It is possible that the
benefits derived from using an aid are more than
purely biomechanical, such as increased confi-
dence and awareness of disability from other
people or reduced energy expenditure and
fatigue. Further study to assess the benefits of
using a walking aid from the patients point of
view would assist in clinical decision making
about the use and possible benefits of walking
aids.
An important limitation of the research about
the effects of walking aids so far is that it has all
involved people with a nonacute hemiplegia
(more than three months duration) who were
functionally mobile. The effect of walking aids on
all aspects of gait in people with an acute hemi-
plegia who are undergoing gait rehabilitation is
also required.
Trunk kinematics in hemiplegic gait 299
Conclusion
Trunk movements while walking have been
described in hemiplegic subjects and baseline
measurements obtained, the relationship
between these movements and walking ability
has been investigated, and the effects of different
walking aids assessed. The trunk movements in
this small group of subjects differed from pub-
lished normal values, with large lateral move-
ments, but small vertical movements.
A striking feature was the asymmetry, which
favoured the hemiplegic side in that there was
less movement of, or movement towards the
hemiplegic side. Lateral movement was signifi-
cantly related to walking ability, but vertical
movement was not. Trunk movements, symmetry
and walking ability were not influenced by the
use of a walking aid or the type of walking aid.
This study looked specifically at trunk move-
ments in a small number of a specific group,
namely people with a nonacute hemiplegia who
were able to walk without assistance, so any gen-
eralization must be treated with caution. The
study does suggest however, that the assumptions
underlying the use of walking aids in stroke reha-
bilitation are not as clear cut as many physio-
therapists would believe, so until these issues can
be clarified, greater flexibility in the use of walk-
ing aids for people with a hemiplegia is called for.
References
1 Moore S, Schurr K, Wales A et al. Observation and
analysis of hemiplegic gait: Swing phase. Aust J
Physiother 1993; 39: 27178.
2 Moseley A, Wales A, Herbert R et al. Observation
and analysis of hemiplegic gait: stance phase. Aust J
Physiother 1993; 39: 25967.
3 Saunders J, Inman V, Eberhart H. The major
determinants in normal and pathological gaits. J
Bone Joint Surg 1953; 35a: 54358.
4 Inman V, Ralston H, Todd F. Human walking.
Baltimore: Williams & Wilkins, USA, 1981.
5 Krebs D, Wong D, Jevesevar D, Riley P, Hodge W.
Trunk kinematics during locomotor activities. Phys
Ther 1992; 72: 50514.
6 Iida H, Yamamuro T. Kinetic analysis of the centre
of gravity of the human body in normal and
pathological gaits. J Biomech 1987; 20: 98795.
7 Thorstensson L, Nilsson J, Carlson H, Zomlefer M.
Trunk movements in human locomotion Acta
300 SF Tyson
Physiol Scand 1984; 121: 922.
8 Waters R, Morris J, Perry J. Translational motion of
the head and trunk during normal walking. J
Biomech 1973; 6: 16772.
9 Thurston A, Harris J. Normal kinematics of the
lumbar spine and pelvis. Spine 1983; 8: 199205.
10 Carr R, Shepherd J. A motor re-learning programme
for stroke. London: Heinemann Medical, 1982.
11 Sackley C, Lincoln N. Physiotherapy treatment for
stroke patients: a survey of current practice.
Physiother Theory Pract 1996; 12: 8796.
12 Collen F, Wade D, Bradshaw C. Mobility after
stroke: reliability of measures of impairment and
disability. Int Disabil Studies 1990; 12: 69.
13 Mitchelson D. CODA: evaluation and application.
In: Proceedings of 1st World Congress of Biomedics
Image based motion measurement, Volume 1356.
Society of Photo-optical Instrumentation Engineers
(SPIE), 1990: 2637.
14 Baer G. A study of sitting to standing in normal and
hemiplegic subjects. MSc dissertation, University of
Southampton, 1992.
15 Rowntree D. Statistics without tears. London:
Penguin, 1981.
16 Wall J, Ashburn A. Assessment of gait disability in
hemiplegics. Scand J Rehabil Med 1979; 11: 95103.
17 Brandstater M, Bruin M, Gowland C, Clark B.
Hemiplegic gait; Analysis of temporal variables.
Arch Phys Med Rehabil 1983; 64: 583587.
18 Mizhari J, Susak Z, Heller L, Najenson T. Variation
of timedistance parameters of stride as related to
clinical gait improvement in hemiplegics. Scand J
Rehabil Med 1982; 14: 13340.
19 Lehmann J, Condon S, Price R, deLateur B. Gait
abnormalities in hemiplegia: their correction by
AFO. Arch Phys Med Rehabil 1987; 68: 76371.
20 Pinzur M, Sherman R, Dimonte-Levine P, Trimble
J. Gait changes in adult onset hemiplegia. Am J
Phys Med 1987; 66: 22837.
21 Knuttson E, Richards F. Different types of disturbed
motor control in gait of hemiplegic subjects. Brain
1979; 102: 405.
22 Winter D, MacKinnon C, Ruder G, Wieman C. An
integrated EMG/biomechanical model of upper body
balance and posture during hemiplegic gait. Prog
Brain Res 1993; 97: 35967.
23 Lehmann J. Push-off and propulsion of the body in
normal and abnormal gait; correction by AFO. Clin
Orthop Rel Res 1993; 288: 97108.
24 Peat M, Hyman I, Dubu M, Winter D. EMG
temporal analysis of gait hemiplegic locomotion.
Arch Phys Med Rehabil 1976; 57: 42125.
25 Holden M, Gill K, Magliozzi M. Gait assessment for
neurologically impaired patients: standards for
outcome. Phys Ther 1986; 66: 153039.
26 Wall J, Turnbull G. Gait asymmetries in residual
hemiplegia. Arch Phys Med Rehabil 1986; 67:
55053.
27 Tyson S. Hemiplegic gait symmetry and walking
aids. Physiother Theory Pract 1994; 10: 15359.
28 Dickstein R, Abulaffio N, Pillar T. Vertical force
loaded on walking canes in hemiparetic patients.
Gait Posture 1993; 1: 11318.
29 Hesse S, Jahne M, Schaffrin A, Lucke D, Reiter F,
Konrad M. Immediate effects of therapeutic
facilitation on the gait of hemiparetic patients
compared with walking with and without a cane.
Electroencephalography Clin Neurophysiol 1998; 18.
30 Tyson S, Ashburn A. The influence of walking aids
on hemiplegic gait. Physiother Theory Pract 1994;
10: 7786.
31 Tyson S. The influence of different walking aids on
the support taken during hemiplegic gait. Clin
Rehabil 1998; 12: 395401.