Content uploaded by Maarten IJzerman
Author content
All content in this area was uploaded by Maarten IJzerman on Apr 13, 2014
Content may be subject to copyright.
The Influence of Frontal Alignment in the Advanced Reciprocating
Gait Orthosis on Energy Cost and Crutch Force Requirements
During Paraplegic Gait
Maarten
J.
Uzerman,
Gert
Baardman,
Gerlinde
GJ.
Holweg,
Hermie
J.
Hermens,
Peter H.
Veltink,
Herman B.K. Boom and Gerald Zilvold
Roessingh Research and Development b.v., Enschede, the Netherlands and Institute
for BioMedical Technology, University of
Twente,
Enschede, The Netherlands
Abstract
Reduction of energy cost and upper body load during paraplegic walking is considered to be
an important criterion in future developments of walking systems. A high energy cost limits
the maximum walking distance in the current devices, whereas wrist and shoulder pathology
can deteriorate because of the high upper body load.
A change in alignment of the mechanical brace in the frontal plane, i.e. abduction, can
contribute to a more efficient gait pattern with sufficient foot clearance with less pelvic lateral
sway. A decrease in pelvic lateral sway after aligning in abduction results in a shift of the
centre of mass to the swing leg crutch which may result in a decrease in required crutch force
on stance side to maintain foot clearance.
Five paraplegic subjects were provided with a standard Advanced Reciprocating Gait Orthosis
(ARGO) and an ARGO aligned in 4 different degrees of abduction (0°, 3°, 6° and 9°). After
determining an optimal abduction angle for each of the subjects, a cross over design was used
to compare the ARGO with the individually optimised abducted orthosis. An abduction angle
between 0° and 3° was chosen as optimal abduction angle. Subjects were not able to walk
satisfactory with abduction angles 6° and 9°.
A significant reduction in crutch peak force on stance side was found (approx. 12% , p <
0.01)
in the abducted orthosis. Reduction in crutch force time integral
(15%)
as well as crutch peak
force on swing side (5%) was not significant. No differences in oxygen uptake as well as
oxygen cost was found. We concluded that an abduction angle between 0° and
3°
is beneficial
with respect to upper
boHy
load, whereas energy requirements did not change.
Key words: paraplegia, crutch forces, energy cost,
orthotics,
frontal alignment.
Basic
Appl.
Myol
7 (2):
123-130;
1997
Although intended to be functional, orthosis supported
walking in paraplegics has primarily been advocated be-
cause of it's psychologic and therapeutic value
[11,
23].
Various devices have been developed during the last dec-
ades in order to provide ambulation to paraplegic patients.
Mechanical braces are most commonly used and include
those with a reciprocating cable linkage, i.e. Advanced
Reciprocating Gait Orthosis (ARGO) and Reciprocating
Gait Orthosis (LSU-RGO), and those without this linkage,
i.e. Hip Guidance Orthosis (HGO). Functional Neuromus-
cular
Stimulation can be added to the mechanical brace to
generate propulsion [10, 15,
16].
One of the problems of the available walking systems
preventing a functional use in daily live, is the high exer-
tion of the upper body during walking
[1,
5].
Moreover,
the high repetitive upper body load may increase the
prevalence of carpal tunnel syndrome as well as shoulder
impingement syndrome in paraplegic individuals [2,
8].
Precaution must be taken when prescribing a walking
system to a population at risk with respect to upper body
pathology. Beside the addition of FNS to a mechanical
brace, several aspects relating to properties and alignment
of the mechanical device may contribute to a reduction in
upper body load and energy requirements.
The reciprocating cable linkage in ARGO as well as
RGO may prevent a free ballistic swing of the swing leg
[22].
It has been suggested that a change of either the
reciprocal coupling mechanism [25] or the transmission
ratio may be beneficial
[26].
Three aspects have been
described which contribute to a facilitation of foot
clear-
Frontal alignment in ARGO
ance during early and mid swing
phase.,
i.e. knee flexion,
alignment in frontal plane and lateral stiffness of the ortho-
sis. By allowing knee flexion during swing phase the
patient can use the bending of the knee rather than pelvic
lateral sway to obtain foot clearance
[4].
Both, lateral
stifness as well as frontal alignment have been described
in HGO because of an enhanced foot clearance with less
pelvic lateral displacement [7,
14,18,20,21].
Nene found
that in some adult patients deformation at the hip hinge
region on the stance side reached such proportions that the
swing leg failed to clear [14,
15].
Pushing harder on the
swing side crutch increases lateral forces and causes even
more deformation at the stance side. Collapsing of the
orthosis
can be compensated by applying force on the
stance as well as on the swing side crutch. In addition,
deformation can be prevented by increasing lateral stiff-
ness in the orthotic structure or by stimulating stance leg
gluteal muscles
[20].
An orthosis which is aligned in slight abduction can also
contribute to a more efficient ratio between foot clearance
and lateral sway of the body
[18,
27].
In addition, due to a
decrease in pelvic lateral sway, the centre of gravity during
mid swing is shifted towards the swing leg crutch. This
change in body posture during mid swing may lead to an
increase in force on swing side crutch and a decrease in
force on stance side crutch. It is expected that alignment in
abduction yields a better utilisation of swing crutch force
for propulsion rather than for pelvic sway and a reduction
in stabilizing forces on stance leg crutch. Rose used a
mathematical description of mechanical stress at the hip
hinge in order to determine optimal abduction angle
[18].
He found that 5° of abduction was associated with least
mechanical stress in an orthosis with an average leg length
of 1 metre and a standard pelvic width. However, an
optimal abduction angle must be determined for each
subject individually because of the differences in anthro-
pometry and their preference with respect to walking with
larger step widths.
This study was conducted in order to assess the differ-
ences between ARGO and ARGO aligned in slight abduc-
tion with respect to upper body load and energy
requirements. Crutch forces, oxygen uptake and subjects'
opinion were used as outcome measures. The study com-
prised an optimisation procedure in which an optimal
abduction angle was determined for each subject and a
comparative trial in which the ARGO was compared with
this individually optimised abducted orthosis.
Methods
Subjects
Subjects were selected from the rehabilitation centre if
they had complete thoracic lesions and could walk inde-
pendently in the ARGO with a regular walking pattern for
10
minutes or more. Subjects were included after they had
given their written informed consent. Five paraplegic sub-
jects were included according to this procedure (table 1).
The study was approved by the local medical ethical
committee.
System description
The standard ARGO, used as reference system, is aligned
in 6° adduction at the level of the hip joints, whereas the
ankle foot braces are mounted in slight abduction. The
ARGO was aligned in 6° abduction by mirroring the hip
joints. Applying different bent rods resulted in abduction
angles of 0°, 3° and 9° (figure 1).
Study Design
Optimisation
The optimal abduction angle was determined for each
subject.
Five different systems were assessed at random on
one day: ARGO and ARGO in 0°, 3°, 6° and 9° of abduc-
tion (Abducted Orthosis: AOo°
,
AO3° , AO6° and AO9°
respectively). Before the assessment, patients were asked
to practice in the orthosis configuration to be tested. All
patients walked using normal crutches and a four point gait
pattern. No differences in walking technique between dif-
ferent
orthoses
were allowed.
Comparative trial
In order to prevent training effects, a separate training
period was included between optimisation and compara-
tive trial. We continued with the comparative trial after the
subjects became familiar with the orthosis configuration.
A cross over trial (sequences:
ARGO-AO0pt
and
AO0pt-
ARGO) was used to compare the ARGO with the individu-
ally chosen abduction angle
(AO0pt).
Comparability of
groups was obtained using level of lesion as matchings
factor
[17].
Two subjects were selected for an
ARGO-AO
Table
1.
Subjects included in the study. Weight is measured in kilogram (kg) and kilogram lean body weight
(kgLBW).
Measurement sequence
'A
' corresponds
with
measurement of ARGO
followed
by the Abducted Orthosis (AO), sequence
'B'
for
AO
followed by ARGO.
123
45
age/sex
level of lesion
weight (kg and kgLBW)
measurement sequence
35/m
T12
66/52.2
A
32/f
T4, T5
60/41.8
B
45/m
T9
90/71
B
41/m
T8
66.5/53.9
A
31/m
Til
100/75.2
B
-124-
Frontal alignment in ARGO
Figure I. Obtaining different hip abduction angles in
Advanced Reciprocating Gait Orthosis (ARGO).
The left hip hinge is drawn in frontal view. ARGO
is aligned in 6° adduction at the hip hinge. Mirror-
ing the lower
hip
joint section yields a 6° abduction
angle. Two bent rods were used to create 0°,
3°
and
9° abduction.
sequence in the cross-over trial and three subjects for an
AO-ARGO
sequence (table
1).
Because of the small sam-
ple size only one factor could be used for stratification.
Level of lesion was chosen, since this was considered as
the most biasing factor.
Measurements
Biomechanical assessments were performed during the
optimisation phase in order to determine the optimal ab-
duction angle. Biomechanical as well as physiological
assessments were performed during the comparative trial.
In addition, during the optimisation phase as well as the
comparative trial, subjects were asked for their grading of
either orthosis.
Biomechanical assessments
Kinetic and kinematic assessments were performed in the
gait lab using a 5 camera 3D movement analysis system
(VICON
370, Oxford Metrics, Oxford, UK). Each meas-
urement consisted of 10 walks along a 5 metre gait lane in
order to obtain 20-30 strides for averaging. Marker posi-
tions of ankles were sampled at a frequency of 50 Hz.
Crutch forces and heel contacts were recorded simultane-
ously at 200 Hz using strain gauges and foot switches
respectively. All data were filtered (linear phase
2n
order
Butterworth,
R.idB
=
5 Hz) and
splitted
into gait cycle
intervals using the heel strike data. Stride length
(m)
and
cadence (strides.min ) were calculated using the ankle
markers. Crutch Force Time Integral
(CFTI)
and crutch
peak force on stance as well as on swing side (CPFstance
and CPFswing) were calculated and normalised for body
weight.
Physiological assessments
Subjects were asked to refrain from coffee, food and
cigarettes for at least 2 hours prior to arrival to the unit.
Breath-by-breath
measurement of inspired and expired
gases was conducted by using a metabolic cart (OXYCON-
alpha,
Jaeger, the Netherlands). Subjects were provided
with a heart rate belt (Sport tester, PE3000, Polar Electro,
Finland) and a facemask containing a flexible gas-tube.
Measurement of rest metabolism was performed during 5
minutes, while the patient sat quietly. Subsequently, sub-
jects were asked to stand up. When heart rate approached a
stable level, subjects were instructed to walk at a comfort-
able, self-selected speed during 10 minutes along a 125
metre circular pathway. After walking subjects were asked
to sit quietly for another
10
minutes in order to determine
their recovery. Heart rate (beats.min" ),
¥02
(ml.min"1.
kg" ), VcO2 (ml.min" . kg" ), Respiratory Exchange Ratio
and expiratory volume
(l.min"
) were measured. Mean
heart rate (HR), Vo2, Vc02, respiratory exchange ratio
(RERsteady state) and expiratory volume (Ve) were calcu-
lated during the last 5 minute interval of walking assuming
a delayed steady state. Vo2 and Vc02 were corrected for
percentage fat and expressed in lean body weight (kg
LEW).
Oxygen cost during steady state
(£02)
was calcu-
lated by:
1-1
EO2
(ml.rrf
.kg"
) = Vo2 /
vsteady
state
Where Vsteady state is walking speed during steady state.
Peak value of Respiratory Exchange Ratio
(RERpCak)
was recorded during the recovery period. Oxygen debt
(O2-debt) was determined by calculating the difference
between Vo2 prior to walking and Vo2 during recovery.
The O2-debt is expressed as the amount of oxygen above
rest Vo2 during the first
10
minutes of recovery.
Subjects' preference
Subjects were asked to give their grading for either
orthosis. We did not ask subjects to compare the devices,
since these comparisons
may
be subject to information bias
[19,24).
Subjects' grading was scored on a 10 point scale,
ranging from dislike to excellent orthosis.
Data analysis
Optimisation
Distribution of each of the variables (CFTI, CPFswing,
CPFstance, walking speed and grading) was examined and
natural log transformation was applied to transform
skewed variables to "normality".
All data for each system and each subject were pooled
and statistically tested by means of two way analysis of
variance (ANOVA) and two way analysis of covariance
(ANCOVA). Factors were system (ARGO and either ab-
ducted orthosis) and subject. Walking speed was included
in the analysis of crutch force data as
covariate
(AN-
COVA) in order to either correct for confounding or obtain
a
more precise estimation
[12].
Post-hoc testing
of
signifi-
cant differences was performed with paired t-tests. A
Bon-
feroni correction was applied to adjust the level of
significance during post-hoc testing. A Bonferoni correc-
-725-
Frontal alignment in ARGO
tion can be obtained by dividing the significance level by
the number of tests to be performed.
Determination of the optimal abduction angle
An optimal abduction angle was selected for each subject
using
subjects'
grading and crutch force outcome as crite-
rion.
Comparative trial
Only parameters which could support clinical decision
making (CFTI,
CPFstance,
CPFswing,
walking speed, Vo2
and
£02)
were subject to statistical testing. Distribution of
each of the variables was examined and natural log trans-
formations were applied to transform skewed variables to
"normality".
Period effects were not statistically tested
because of the small sample size. Moreover, it was as-
sumed that period effects were reduced since all data were
pooled. Paired t-tests were used to calculate 95% confi-
dence intervals for the differences between ARGO and
AOopt-
All 95% confidence intervals are presented as
relative differences with respect to the ARGO assess-
ments.
Two way ANCOVA was used to adjust for walking
speed if walking speed was different between ARGO and
AOopt. All analyses were done using SPSS. A
p-level
of
0.05 was considered significant.
Results
Optimization
An abduction angle of 9° was excluded from the analyses
because of the large step widths in this system. Subjects
were not able to walk with the system. Figure 2 shows a
typical crutch force pattern of the left crutch of one of the
subjects. The left crutch is stance leg crutch during the first
and third peak, after left heel strike. Swing leg crutch peak
force (second peak) is higher than stance leg crutch peak
force
[14].
Stance leg crutch peak force is reduced in the
AOo° and AO3°
orthosis,
whereas the swing leg crutch
peak force did not change in these systems with respect to
the ARGO. A reduction in peak force of swing leg as well
as stance leg crutch was found for the AO6°. The decrease
in swing leg peak crutch force may be caused by differ-
ences in walking speed.
Reduction in walking speed in AO6° was not significant
(p <
0.16,
ANOVA, table 2). A significant difference was
found with respect to subjects' grading for either orthosis
Figure
2.
Typical
example
of
left
crutch force
patterns
of
one subject walking in the ARGO and AO
with
different abduction angles. The left crutch is stance
leg
crutch during
the first and
third
peak,
after
left
heel strike.
Swing
leg crutch peak force is higher
than stance leg crutch
peak
force. Stance leg crutch
force is reduced in the
AOQO
and
AO3o
orthosis,
whereas swing leg crutch force did not change in
these systems with respect to the ARGO. The de-
crease in swing leg crutch force (second peak) may
be caused by differences in walking speed.
Table 2. Results of ARGO and either abducted
orthoses
obtained during the optimisation phase. Mean (and standard deviation)
of
walking
speed,
crutch forces (Crutch Peak Force (CPF) and Crutch Force Time Intergral
(CFTI))
and
subjects'
preferences. F- and
p-values
of two way ANOVA and ANCOVA are
presented.
Factors in the ANOVA were subject and
system. Walking speed was added in the ANCOVA as
covariate.The
p-value
of the covariate indicates
-whether
the
difference between adjusted and crude data is significant.
CFTI
CPFS
CPF;
swing
walking speed
grading
ARGO
AOo°
AO3°
AO6°
ANOVA
system
ANCOVA
system
covariate
0.59(0.12)
0.57(0.12)
0.56(0.13)
0.59(0.21)
0.53
/p<
0.67
1.34
/p<
0.31
p<0.05
0.39 (0.05)
0.36 (0.04)
0.36 (0.04)
0.33 (0.07)
9.42
/p<
0.01
6.17/p<0.01
p<0.82
0.43 (0.02)
0.40 (0.02)
0.41 (0.03)
0.40 (0.07)
0.82
/p<
0.51
2.01/p<0.17
p<0.10
0.29 (0.09)
0.28 (0.09)
0.28(0.10)
0.26(0.11)
2.03
/p<
0.16
7.60 (0.55)
7.60 (0.55)
8.40(1.30)
6.00 (0.70)
6.70
/p<
0.01
-126-
Frontal alignment in ARGO
(p<
0.01,
ANOVA, table 2). Although walking speed was
not significantly reduced, differences in walking speed
between
orthoses
configurations may hamper the interpre-
tation of crutch force outcome. Analysis of
covariance
(ANCOVA)
was used to adjust for walking speed. AN-
COVA can only be used reliably if there is no effect
modification or interaction between covariate and out-
come measure
[12].
Effect modification in epidemiology
is the condition where the relationship of interest is differ-
ent at different values of the extraneous variable. Visual
inspection of scatter plots was used rather than a formal
statistical test, since we had too few data to perform a test
on interaction (figure 3). Differences in
CFTI
as well as
CPFswing
were not significant (p <
0.31
and p < 0.17
respectively, ANCOVA). CPFstancc was significantly re-
duced in the abduction orthoses (p <
0.01,
ANCOVA). The
crude and adjusted differences between ARGO and either
abduction
orthosis
were calculated (figure 4). Differences
in CFTI as well as CPFswing were larger after adjusting for
walking speed. The difference between crude and adjusted
data was significant for CFTI (p < 0.05). Average reduc-
tion
in CFTJ and
CPFsw;ng
in the
AOr
was 9.5% and 7%
respectively. CPFstancc is not affected by walking speed
(covariate: p < 0.82). The average reduction in CPFstance
was approximately 8% for
AOo°
as well as
AOs0.
Since neither subjects' grading nor CPFstance were biased
by walking speed, post-hoc testing was performed using
paired t-tests. Subjects found no difference between
ARGO and
AOo°,
preferred the AO3° and did not like the
AO6°
(table 3). CPFstancc was significantly reduced in the
AOf,0.
CPFstance was not significantly reduced in
AOo°
and
AOr.
Determination of the optimal abduction angle
A decrease in CPFstance as well as CPFswing of approxi-
mately
10%
and
16%
respectively was found for the
AO(,s.
However, subjects preferred the
AOs°
because of some
interfering user aspects in the
AOe°.
Sitting in a wheelchair
was not possible with such abduction angles. Conse-
quently, independent donning and doffing was not possi-
ble. Therefore, subjects' grading was used as main
criterion for determining the optimal abduction angle. Four
subjects chose a 3°, and one subject chose a 0° abduction
aligned orthosis.
Comparative trial
An estimation of the differences in physiological and
crutch force measurements was made using 95% confi-
dence intervals (figure 5). No difference in subjects' grad-
ing was found with respect to either ARGO or AOopt.The
increase in walking speed in
AO0pt
was not significant
(95%
CI:
[-25%,
2%]).
No significant difference in oxygen
cost
(£02)
and oxygen uptake (Vo2) was found in the
AOopt. Table 4 summarizes the other physiological vari-
ables. Heart rate and Ve increased in the
AO0pt.
A trend of
increased
RERpeak,
RERsteadystate
as well as O2-debt in the
AOopt
can be observed.
The reduction in CPFstance in AOopt was significant
(95% CI: [1% .
J9%J).
The reduction in CFTI as well as
CPFswing
was not significant (95% CI:
H%
, 32% ] and
Table 3. Post hoc testing of subjects ' grading and CPFstance
by means
of
paired t-tests. Mean, (standard deviation)
and
p-value
of differences are presented. The actual
significance level is
0.017
due to
Bonferoni
correc-
tion.
subjects' gradingCPFstance
ARGO-AOo°
0.00
(1.00),
p =
1.0
ARGO-AO3=
-0.80(1.10),
p<
0.18
ARGO-AC>6»
1.60
(0.55), p < 0.05
0.03(0.03),
p<
0.10
0.03(0.02),
p<
0.07
0.06(0.02),
p<
0.01
I
a
E
o
walking speed [m/s]
Figure 3. Scatter plot of CFTI against walking speed or
different abducted orthoses. The slopes are not
significantly different, indicating no effect modifi-
cation. The differences in intercepts between the
Abducted Orthoses and ARGO represent the walk-
ing speed adjusted mean difference.
crude
data
adjusted
differences
0-
3-
CFTI
0°
3°
6'
OPF
avflng
0-
3'
6"
CPF9tance
Figure 4. Crude and walking speed adjusted differences
for crutch force variables using analysis of co-vari-
ance. Differences are presented as relative change
with respect to reference system ARGO. Positive
differences indicate higher values in ARGO.
-127-
Frontal alignment in ARGO
ARGO
AO
opt
95 %
Cl
for difference
preference
speed
(m.s
')
EO2
(ml.m'Vkg1)
VQ2
(ml.min'.kg1)
CFTI
(Ns.kg'1)
CPFswing
(N.kg1)
7.40
±
0.89
0.25
±
0.11
1
.30
+
0.47
16.99
+
3.60
3.72 + 0.73
7.40 + 0.89
0.27
±
0.10
1 23 + 0 47
17.97
±
3.19
5.36
±2.
14
4.10+
1.00
3.33 ± 0.94
,
«
,
:,
,
.
-40
-30
-20 -10 0 10 20 30
relative difference ARGO -
AOopt
[%]
40
Figure 5. Estimation of differences between ARGO and
AO(>
t
with respect to
energy,
requirements, walking speed, crutch
force requirements and
subjects'
grading. Estimations were made using 95% confidence intervals. The intervals
are represented as relative change with respect to the reference ARGO. Positive differences indicate higher values
in ARGO.
Table 4. Results of physiological assessments during the comparative trial. Heart rate (HR), expiratory volume (Ve), oxygen debt
(O2-debt), respiratory exchange ratio during steady state walking
fA^^steadystat^
and peak value
ofRER
during recovery
(RERpC3\i).
Presented data are mean and (standard deviation).
ARGO
AOopt
HR
b.mirf'
136.4(19.1)
143.0(16.4)
Ve
l.min"
36.6(12.7)
39.2(12.5)
O2-debt
ml.kgLBW'1
17.7 (5.6)
21.3(6.7)
RERsicadystate
0.99(0.10)
1.01
(0.05)
RERpeak
1.13(0.10)
1.16(0.07)
[-3%,
12%]
respectively). Two way
ANCOVA
was used
to adjust for walking speed in order to obtain more precise
estimations of differences in crutch force and energy re-
quirements. A summary of crude and adjusted differences
Eo,
CFT!
Figure 6. Crude and walking speed adjusted differences
between ARGO and Abducted Orthosis. Relative
differences are presented. Positive differences in-
dicate higher values in ARGO.
is given in figure 6. The walking speed adjusted differences
in CPFstance as well as CFT were slightly larger than the
crude differences. However, adjusting for walking speed
did not significantly change the outcome for these variables.
Discussion
Several authors have outlined that reduction of upper
body load and energy expenditure during walking are
important design criteria for walking systems [1,
13].
Functional use of a walking system requires an energy
efficient walking pattern and minimal upper body load.
This may improve gait performance and prevent wrist and
shoulder pathology as well. Most evaluations which have
appeared in the literature are comparisons between clinical
systems [23, 25] or studies on the effect of functional
neuromuscular stimulation
[10,
13,
16].
Few articles have
been published on the influence of
orthosis
components on
energy cost and upper body load. The benefits of orthosis
alignment in abduction have been described theoretically
by Rose
[18],
but he gave credit to Herzoz and
Sharrard
-128-
Frontal alignment in ARGO
[9].
Although a slight abduction alignment has been used
in the Hip Guidance
Orthosis,
the benefits of this modifi-
cation on energy cost and crutch force requirements have
not been investigated.
The optimisation procedure in our study was included in
order to gain insight into the size of the optimal abduction
angle. This is why we have chosen only four different
abduction angles. A more precise estimation of the abduc-
tion angle was expected to be senseless, since subsequent
differences of 1 ° would have been to small to detect. We
concluded from the optimisation that an abduction angle
between 0°and 3° was sufficient in order to gain the
benefits. All five subjects had the same experience of more
stability and ease of foot clearance.
Variations in walking speed between
orthoses
hamper a
reliable comparison, since most physiological and
biomechanical measures are speed dependent. Moreover,
the influence of walking speed as confounding factor in the
relation between
orthosis
and oxygen cost was noticed by
Marsolais et
al.
who presented
scatterplots
of walking speed
against
£02
in his comparison of RGO and RGO-FES
[13].
The walking speed dependent behaviour is clear with
respect to
CFTI.
CFTI
decreases at higher walking speeds,
since the integration interval is decreasing. Adjusting for
walking speed by means of regression analysis or analysis
of covariance is a common procedure in order to correct
for confounding bias. However, in case of CFTI one should
be aware of the behaviour of CFTI itself. By integrating
the force one introduces a time dependent behaviour. Ad-
justing for walking speed implies a correction for time and
results in an estimation of the force component in the CFTI.
CFTI therefore, may not give additional information with
respect to crutch peak forces or crutch average forces per
stride. In addition, the time dependent behaviour only
introduces a very difficult interpretation in cases of differ-
ent walking speeds.
In the optimisation phase walking speed was mainly
reduced in the AO6°, which can be explained by the inabil-
ity of the subjects to walk in this system satisfactorily.
As a consequence, a proper comparison of ARGO and
AO6° on the basis of the crude crutch force outcome is
difficult. In addition, it can also be expected that more gait
training in the abducted orthoses can diminish the differ-
ences in walking speed. Adjusting for walking speed may
then be
doubtful.
Walking speed in the AOo° and AO3° was
only slightly changed with respect to the ARGO, so a
comparison of these systems is more appropriate. We
found that CPFstance as well as CPFswing were reduced in
AOo° as well as AO3°.
Walking speed in the
AO0pt
increased in the comparative
trial with
11%
on average. In healthy subjects, a comfort-
able walking speed is closely related to the metabolically
most efficient walking speed, i.e. the walking speed with
the lowest oxygen consumption per unit distance
[6].
In
paraplegia, however, it is not clear whether comfortable
walking speed can be determined by means of the relation
with oxygen consumption. Another possible physiological
explanation of comfortable walking speed is the onset of
blood lactate accumulation (OBLA)
[5].
An increase in
comfortable walking speed therefore indicates that the
OBLA will be reached at higher walking speeds. This can
be interpreted as a functional improvement, i.e. higher
walking speed during steady state. In the present study,
Vo2 did not change in the
AO0pt,
indicating that the aerobic
part of energy delivery was equal to the ARGO. The
anaerobic part of energy delivery can be estimated roughly
by means of RER and O2-debt
[3,13].
RERpeak,
RERsteady
state as well as O2-debt were higher in
AO0pt.
If the OBLA
was reached at higher walking speeds we had expected that
there was no difference between the estimators of anaero-
bic energy delivery. Comfortable walking speed in the
AOopt is possibly more related to personal preference
rather than to a physiological mechanism.
A clinically probably more relevant outcome variable
with respect to the perceived exertion of walking is crutch
peak force. Beside the relation between upper body load
and prevalence of wrist and shoulder pathology, a reduc-
tion in upper body load may also contribute to the experi-
ence of a less fatiguing walking pattern. A clinically
relevant and significant reduction in CPFstance (crude:
12%)
is therefore the most important finding in the study.
The reduction in CPFswing (crude: 5%) is considered clini-
cally relevant as well, since absolute values of CPFswing
are higher than those of
CPFstance.
Alignment of the ARGO
in abduction may therefore be indicated for those individu-
als whose walking distance is limited by wrist and shoulder
pain rather than metabolic energy delivery.
Acknowledgement
This study was supported by the Technology Foundation
and the Research Stimulation Fund (OSF) University of
Twente. We would like to thank Mr. P. Vlaanderen for
technical support and all volunteers for participating in the
study.
Address correspondence to:
Maarten
J. IJzerman, MSc, PT, Roessingh Research and
Development b.v., P.O. Box
310,
7500 AH Enschede the
Netherlands,
tel.
+ 31 53 4875733, fax + 31 53 4340849,
e-mail: m.ijzerman@rrd.nl.
List of abbreviations
AO = Abducted Orthosis, ARGO = Advanced Recipro-
cating Gait Orthosis, Vo2
=
Oxygen uptake
[ml.min
.kg
],
£02
=
Oxygen
cost
[ml.m"
.kg"
],CFTI
=
Crutch Force Time Integral, CPF = Crutch Peak Force,
ANOVA = Analysis of Variance, ANCOVA = Analysis of
CoVariance.
References
[1] Andrews BJ, Granat MH, Heller BW,
MacjVIahon
J, Keating L, Real S: Improved harness crutch to
reduce upper limb effort in swing-through gait.
MedEngPhys 1994; 16: 15-18.
Frontal alignment in ARGO
[2]
Aljure
MD, Eltorai I, Bradley
WE,
Lin JE, Johnson
B: Carpal tunnel syndrome in paraplegic patients.
Paraplegia 1985; 23: 182-186.
[3] Astrand PO, Rodahl K: Textbook of work physiol-
ogy*
physiological bases
of
exercise.
McGraw-Hill,
1986,
chapt. 7 physical performance, 295-348.
[4] Bataweel AO, Edwards J: An
orthotic/FES
system
for paraplegic walking with a knee bending swing
phase. In: Proc. BES symposium, electrical stimu-
lation - clinical systems. Glasgow, UK, April
1995,
pp
7-8.
[5]
Bernardi
M, Canale I, Castellano V, Di Filippo L,
Marchetti M: The efficiency of walking of paraple-
gic patients using a reciprocating gait
orthosis.
Paraplegia 1995; 33: 409-415.
[6] Corcoran PJ, Brengelmann GJ: Oxygen uptake in
normal and handicapped subjects, in relation to
speed of walking beside velocity-controlled cart.
Arch Phys Med Rehabil 1970;
51:
78-87.
[7]
Ferrarin
M, Stallard J, Palmieri R, Pedotti A: Esti-
mation of deformation in a walking orthosis for
paraplegic patients.
Clin
Biomech 1992; 8: 255-
261.
[8] Gellman H , Sie I, Waters
RL:
Late complications
of the weight-bearing upper extremity in the para-
plegic patient. Clin Orth 1988; 233: 132-135.
[9] Herzoz EG,
Sharrard
WJW:
Calipers and braces
with dundee hip locks. Clin Orth
RelRes
1963;
46:
293-242.
[10] Hirokawa S, Grimm M, Le T, Solomonow M,
Baratta RV,
Shoji
H,
D'Ambrosia
RD:
Energy
consumption in paraplegic ambulation using the
reciprocating gait orthosis and electric stimulation
of the thigh muscles. Arch Phys Med Rehabil
1990;
71:687-694.
[11]
Jaeger RJ, Yarkony GM, Roth EJ: Rehabilitation
technology for standing and walking after spinal
cord injury. Am J Phys Med Rehabil 1989; 68 (3):
128-133.
[12]
Kleinbaum DG, Kupper LL, Muller KE: Applied
regression analysis and other multivariable meth-
ods. Duxbury press, Belmont, California, 1988, C.
11
confounding and interaction in regression p.
163
-181 and C. 15 analysis of covariance and other
methods for adjusting continuous data p.
297-314.
[13] Marsolais EB, Bennett
G,
Edwards CCT: Energy
costs of walking and standing with functional
neuromuscular stimulation and long leg braces.
Arch Phys Med Rehab 1988; 69: 243-250.
[14] Nene AV, Major RE: Dynamics of reciprocal gait
of adult paraplegics using the parawalker (Hip
Guidance Orthosis). Prosthet Orthot
Int
1987;
11:
124-127.
[
15]
Nene AV,
Hermens
HJ, Zilvold G: Paraplegic loco-
motion - a review. Spinal Cord
1996;
34: 507-524.
[16] Petrofsky JS, Smith JB: Physiologic costs of com-
puter controlled walking in persons with paraplegia
using a reciprocating gait orthosis. Arch Phys Med
Rehabil 1991; 72: 890-896.
[17] Pocock, SJ: Clinical trials: a practical approach
(1983).
John Wiley and sons, Chicester.
[18]
Rose GK: The principles and practice of hip guid-
ance articulations. Prosthet Orthot Int 1979; 3:
37-43.
[
19]
Sackett, DL: Bias in analytic research. J Chron Dis
1979;
32:51-63.
[20] Stallard J, Major RE: The case for lateral stiffness
in walking
orthoses
for paraplegic patients. Proc
Instn
Mech Engrs 1993; 207: 1-6.
[21] Stallard J, Major RE: The influence of orthosis
stiffness on paraplegic ambulation and its implica-
tions for functional electrical stimulation systems.
Prosthet Orthot Int 1995; 19: 108-114.
[22] Tashman S, Zajac FE, Perkash I: Modeling and
simulation of paraplegic ambulation in a recipro-
cating gait orthosis.
J
Biomechanical Engineering
1995; 117: 300-308.
[23] Thoumie P,
Perrouin-Verbe
B, Le Claire G, Be-
doiseau M, Busnel M,
Cormerais
A, Courtillon A,
Mathe
JF,
Moutet F, Nadeau G, Tanguy E, Beillot
J, Dassonville J and Bussel B: Restoration of func-
tional gait in paraplegic patients with the
RGO-II
hybrid orthosis. A multicentre controlled study. I
Clinical evaluation. Paraplegia 1995; 33: 647-653.
[24] Whittle MW, Cohrane GM: A comparative evalu-
ation of the hip-guidance orthosis (HGO) and the
reciprocating gait orthosis (RGO). Health Equip-
ment Information No. 192, London, National
Health Service Procurement Directorate.
[25] Winchester PK, Carollo JJ,
Parekh
RN, Lutz LM,
Aston jr
JW:
A comparison of paraplegic gait using
two types of reciprocating gait orthoses. Prosthet
Orthot Int
1993',
17:
101-106.
[26]
Yang
L, Maxwell DJ, Granat MH, Yang Y, Paul JP,
Condie DN, Rowley DI: Development of control
strategies for a hybrid RGO system, in Veltink PH,
Boom HBK (eds): Proc. RAFT workshop "control
of ambulation using
FNS".
1994, ISBN 90-365-
0734-0,
pp
89-93.
[27]
IJzerman
MJ,
Baardman
G, Holweg GGJ, Hermens
HJ, Veltink PH, Boom HBK, Zilvold G: The influ-
ence of frontal alignment in the ARGO on energy
cost during paraplegic gait, in: Proc. First Mediter-
ranean Congress on Physical Medicine and Reha-
bilitation. Herzliya, Israel,
1996,
p
235.
-130-
