Original article 221

Are knee kinematic anomalies in swing due to rectus femoris

spasticity different from those due to femoral anteversion
in children with cerebral palsy? A quantitative evaluation
using 3D gait analysis
Veronica Cimolina, Luigi Piccininib, Anna Carla Turconib, Marcello Crivellinia
and Manuela Gallia,c

Quantitative comparison of gait strategy between stiff knee
gait caused by rectus femoris spasticity versus that caused
c 2010 Wolters Kluwer
by femoral anteversion was the objective of this study.
Twenty-three diplegic were divided into group 1 (excessive
femoral anteversion without rectus femoris spasticity) and
group 2 (normal femoral anteversion and rectus femoris
spasticity). Both groups showed low knee flexion during
swing (KMSw), but although group 1 exhibited normal
KMSw timing and high hip intrarotation, group 2 presented
delayed KMSw timing, with normal hip rotation. Reduced
KMSw may be because of two different conditions:
excessive femoral anteversion, leading only to KMSw
reduction, and rectus femoris spasticity, inducing
coexistence of reduced KMSw and its delayed timing.
J Pediatr Orthop B 19:221–225
Health | Lippincott Williams & Wilkins.
Journal of Pediatric Orthopaedics B 2010, 19:221–225
Keywords: cerebral palsy, femoral anteversion, gait analysis, rectus femoris
spasticity, rehabilitation
a
Deparment of Bioengineering, Politecnico di Milano, Milan, bIRCCS Eugenio
Medea, Bosisio Parini, Lecco and cIRCCS San Raffaele Pisana – San Raffaele
SpA, Roma, Italy
Correspondence to Veronica Cimolin, PhD, Dipartimento di Bioingegneria,
Politecnico di Milano, P.zza Leonardo da Vinci 32, 20133 Milano, Italy
Tel: + 39 02 23993359; fax: + 39 02 23993360;
e-mail: veronica.cimolin@polimi.it

Introduction anteversion remodels so that by adulthood it measures on

In cerebral palsy (CP), knee pathology in swing is average 10–151 [4]. As RF is a biarticular muscle arising
almost the norm. In fact, many patients with CP walk by two tendons (from the anterior inferior iliac spine and
with insufficient knee flexion during the swing phase from the acetabulum) with the distal insertion into the
(KMSw), or stiff knee gait, and this problem leads to base of the patella, a femur malrotation can lead to an
difficulty in clearance of the foot in swing phase. This increased stretch of this muscle causing an increased
movement abnormality is often attributed to excessive knee extension.
activation of rectus femoris (RF) [1,2] and the result on
From these considerations it is clear that the stiff knee
gait is that the magnitude of peak knee flexion is greatly
gait, very common in CP, may be because of two differ-
reduced and the timing of the peak knee flexion is
ent causes and biomechanical patterns: the presence
delayed from the initial swing to mid-swing. However,
of RF spasticity or the presence of increased femoral
other biomechanical factors that contribute to stiff knee
anteversion.
gait have not been adequately characterized [1].
As literature is poor on this topic and the evaluation is
From clinical evaluation, the decrease of knee flexion
carried out using only clinical observation and evaluation,
peak in swing can also be because of an increased femoral
from a clinical viewpoint, there is a need to investigate
anteversion without the presence of RF spasticity. The
deeply and quantitatively the differences between these
anteversion angle is the angle between the neck of the
two strategies in CP patients, using three-dimensional
femur and the condylar axis (axis between the two
(3D) gait analysis (GA). In this way, it is possible to make
femoral condyles) projected onto the horizontal plane and
a clear and precise characterization of the two patterns
measures the rotation of the neck of the femur around
dynamically during walking, using quantitative kinematic
the diaphysis. This angle is generally assessed as part of
and electromyography (EMG) information, providing a
the static examination in the prone position evaluating
better insight of the patterns of the two groups. In addi-
by comparing internal and external hip rotation, as well as
tion, GA allows the quantification of the exact amplitude
palpation of the point of maximal trochanteric promi-
and timing of knee flexion peak in swing phase. It may be
nence [3,4]. An anteversion angle of 01 indicates the axis
decisive in establishing the therapeutical programs that are
of the neck and the condylar axis is in the same plane; a
different for the two groups of patients.
negative angle, called retroversion, means the head of the
femur is rotated posterior to the condylar axis. In utero, The main aim of this study was the quantitative com-
femoral anteversion is about 551. With the onset walking, parison of gait strategy between CP patients with stiff
1060-152X
c 2010 Wolters Kluwer Health | Lippincott Williams & Wilkins DOI: 10.1097/BPB.0b013e32833390ca

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 222 Journal of Pediatric Orthopaedics B 2010, Vol 19 No 3
knee gait caused by RF spasticity and CP patients
with stiff knee gait caused by femoral anteversion, using
GA data.
Materials and methods
Patients
Twenty-three CP diplegic independent children (range:
5–13 years; mean age: 9.27 years; 46 limbs) were
evaluated in this study. Selection criteria for pathological
patients were a physician diagnosis of spastic diplegia
with stiff knee gait in swing phase, no history of
cardiovascular disease and no previous surgery or other
significant treatments for spasticity. All patients were
able to walk independently without the use of crutches,
walkers or braces.
A control group (CG) of 15 healthy individuals (CG;
range: 5–16 years) was included. Selection criteria for
nondisabled individuals included no prior history of
cardiovascular, neurological or musculoskeletal disorders.
They exhibited normal range of motion (ROM) and
muscle strength, and had no apparent gait abnormalities.
All participants were volunteers and their parents gave
their written consent to the children’s participation in
this research, in accordance with the requirements of the
local ethics committee.
Data collection
All patients underwent clinical examination, videorecord-
ing and 3D GA. In the clinical examination, the femoral
anteversion evaluation and Duncan-Ely test [5] were
performed by the same physiotherapist. GA was con-
ducted using an optoelectronic system with passive
markers (ELITE2002; BTS, Milan, Italy) working at a
sampling rate of 100 Hz, for kinematic movement
evaluation [6], an eight-channnel surface EMG system
(TeleEMG; BTS) for muscles electromyographic signals
monitoring, and a Video system synchronic with the
optoelectronic and EMG systems (BTS).
After collection of some anthropometric measures (height,
weight, tibial length, distance between the femoral
condyles or diameter of the knee, distance between the
malleoli or diameter of the ankle, distance between the
anterior iliac spines and thickness of the pelvis), passive
markers were placed at special points of reference,
directly on the participant’s skin, as described by Davis
[7]. For the surface EMG recording, bipolar Ag/AgCl
surface electrode pairs with a diameter of 10 mm and an
interelectrode spacing of 22 mm were placed bilaterally
on clean, shaven skin overlying the RF. The SENIAM [8]
recommendations for surface EMG were followed for
electrode placement. The ground electrode was placed
overlying the tibial tuberosity. EMG signals were pre-
amplified, band-pass filtered (10–700 Hz) at a sampling
rate of 2520 Hz, but not processed further. After prepara-
tion, patients walked barefoot at their self-selected speed
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
along a 10-m walkway. To allow the assessment of trials
consistency, at least five trials were recorded and checked
for each child. All data were collected simultaneously and
expressed in percentage of gait cycle.
Analysis of data
The related positions of each joint and joint centre
were estimated through the motion analysis and human
anthropometric data. The limb rotation algorithm is
based on the determination of Euler angles with a y–x–z
axis rotation sequence. The joint rotation angles that
are routinely obtained correspond to flexion/extension,
adduction/abduction and internal/external rotation, re-
spectively. Therefore, the joint rotation angles that are
determined clinically are trunk and pelvic obliquity-
tilt-rotation, hip ad/abduction–flexion/extension–rotation,
knee flexion/extension, ankle plantar/dorsiflexion and
foot rotation. The trunk and pelvic angles are absolute
angles, referenced to the initially fixed laboratory coordi-
nate system; the hip, knee and ankle angles are all relative
angles; for example, the three hip angles describe the
orientation of the thigh with respect to the pelvis; the foot
rotation angle is an absolute angle, referenced to the
laboratory, which indicates the position of the subject’s
foot with respect to the direction of progression [7].
For the purpose of analysing kinematic and EMG data,
the patients’ data were divided into two groups (those
with excessive femoral anteversion and Duncan-Ely r 2
and those with physiological femoral anteversion and
Duncan-Ely > 2).
As concerns kinematics, some parameters were identified
and calculated: spatiotemporal parameters and joint angle
values (pelvis, hip, knee and ankle joints) at a specific
instant of the gait cycle and on the frontal/sagittal/
transversal planes of movement. The kinematic data are
presented as average values for each group and were
compared to values of the CG.
As concerns EMG data abnormalities of EMG activity
and timing were required for the definition of abnormal
muscle activity. The criteria were derived from our
database on normal subjects and from literature [9] and
were as follows:
(1) activity: only EMG signals reaching at a least half of
the maximum amplitude of the recordings during a
whole cycle were considered;
(2) timing:
(a) knee extensor activity in mid or late stance;
(b) continuous knee extensor activity during swing
phase.
Statistics
A one-way between-group analysis of variance was applied
for statistical analysis; the assumptions of the analysis
 Quantification of knee anomalies in swing Cimolin et al. 223

of variance model were tested by evaluating the fit of the stance phase (group 1: 61.18 ± 2.59%; group 2: 59.09 ±
observed data to the normal distribution (Kolmogorov– 4.78%; CG: 58.52 ± 2.34%) and velocity of progres-
Smirnov test) and the homogeneity of variances (Levene’s sion (group 1: 0.99 ± 0.22 m/s; group 2: 1.03 ± 0.19 m/s;
test). Specific effects were evaluated by using the CG: 1.16 ± 0.16 m/s) close to normality when compared
post-hoc comparisons of means and the Bonferroni test. with CG.
Null hypotheses were rejected when probabilities were
below 0.05. Kinematic parameters (Table 1)
No statistical differences were found between group 1
Results and group 2, with respect to pelvic joint: on the sagittal
From the results obtained by clinical examinations two plane both of them were characterized by pelvic anterior
pathological groups were identified: position (Mean Pelvic Tilt index) with higher ROM
during walking in the sagittal plane (ROM Pelvic Tilt
(1) Group 1 (60.9% of patients; 28 limbs; age: 9.4 ± 5.7 index), in comparison with normality, significant at a
years; height: 125.9 ± 5.3 cm; weight: 27.6 ± 6.4 kg), statistical level. No significant features were found on the
composed of patients with femoral anteversion frontal plane, whereas on the transversal plane, the pelvic
higher than normative values (> 401) and low RF joint showed elevated excursion during gait cycle (ROM
spasticity (Duncan-Ely r 2); Pelvic Rotation index) in both groups.
(2) Group 2 (39.1% of patients; 18 limbs; age: 8.9 ± 3.8
years; height: 128.6 ± 7.4 cm; weight: 29.8 ± 7.7 kg), Hip joint revealed a limited flexion at initial contact
composed of patients with femoral anteversion and a reduced capacity of extension during midstance
inside normal range (r 401) and high RF spasticity in both groups with respect to normal range, but no
(Duncan-Ely > 2). statistically significant differences were present between
group 1 and group 2. On the transversal plane, significant
Age, body weight and height were not significantly differ- difference was found between the two pathological
ent among pathological groups and healthy individuals. groups: in fact, group 1 presented an excessive hip
internal rotation during all gait cycles (Mean Hip
In Table 1, the mean values (standard deviation) of Rotation index) whereas group 2 showed a value of this
kinematic parameters considered in this study for group 1, parameter close to normative data.
group 2 and for CG are reported.
Knee flex-extension plots revealed that both pathological
Spatiotemporal parameters groups presented excessive knee flexion at initial contact
The analysis of spatiotemporal parameters revealed and a reduced capacity of the KMSw index, without any
no statistical differences between the two pathological significant differences between them.
groups: group 1 and group 2 highlighted percentage of
A significant difference was found in terms of timing
of KMSw index, calculated with respect to the instant
Table 1 Mean values (standard deviation) of kinematic parameters
of toe-off: group 1 exhibited a KMSw timing close to
for group 1, group 2 and CG normative data whereas group 2 presented a significant
Group 1 Group 2 CG
delay in KMSw timing (13.69 ± 3.87 vs. 19.12 ± 4.34
percentage of gait cycle; P < 0.05; CG: 13.58 ± 2.91 per-
Pelvis (degrees)
Mean pelvic tilt 16.09 (4.69)** 17.56 (4.03)** 9.22 (3.97)
centage of gait cycle; P < 0.05) (Fig. 1); the results of
ROM pelvic tilt 10.98 (5.76)** 10.59 (7.91)** 2.57 (1.92) delayed timing present in group 2 is consistent with the
Mean pelvic obliquity 0.58 (2.36) 0.61 (4.74) 1.59 (2.39) literature [4].
ROM pelvic obliquity 8.61 (4.97) 9.21 (2.44) 8.46 (1.84)
Mean pelvic rotation – 0.48 (3.69) – 0.61 (3.98) 2.48 (3.13)
ROM pelvic rotation 20.66 (3.87)** 19.32 (5.12)** 12.33 (2.89)
The analysis of ankle kinematics showed that the two
Hip (degrees) pathological groups exhibited similar patterns, with a
Hip at IC 31.61 (8.37)** 30.56 (7.10)** 38.28 (7.98) normal ankle position at initial contact, a good capacity of
Hip minimum in St 0.98 (5.87)** – 1.23 (7.99)** – 7.89 (5.78)
Mean hip rotation ,
10.45 (6.29)* ** 1.56 (5.29) – 1.95 (4.78) dorsal flexion in stance and a reduced plantar flexion at
Knee (degrees) toe-off. Foot angle in transversal plane (Mean Foot
Knee at IC 19.96 (7.49)** 18.45 (6.11)** 7.41 (2.69)
Knee minimum in St 2.67 (9.37) 1.56 (7.92) 4.31 (2.01)
Progression index) pointed out a normal foot orientation
Knee maximum in Sw 37.58 (5.93)** 34.67 (6.92)** 62.66 (7.12) in all patients.
Ankle and foot (degrees)
Ankle at IC 2.72 (6.23) 1.56 (7.78) 1.47 (4.47)
Ankle maximum in St 10.37 (5.67) 11.98 (7.28) 13.33 (5.44) EMG data
Ankle minimum in St – 2.20 (4.23)** – 3.24 (6.43)** – 12.06 (4.24) In terms of EMG activity of RF, two different situations
Mean foot progression – 14.89 (6.45) – 11.67 (7.91) – 12.47 (5.82)
were found in the pathological groups: 93% of limbs
CG, control group; IC, initial contact; ROM, range of motion; St, stance; Sw, (26 limbs) of group 1 showed a normal activation of
swing.
*P < 0.05, group 1 vs. group 2. RF during all gait cycles (Fig. 2a), which confirms the
**P < 0.05, group 1 and group 2 vs. control group. absence of RF spasticity tested by Ely Test, whereas

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 224 Journal of Pediatric Orthopaedics B 2010, Vol 19 No 3

Fig. 1 From qualitative and clinical examinations, clinicians note

that the presence of increased spasticity of RF generally
25
induce abnormalities at knee joint pattern during the
swing phase such as decrease and/or delay in flexion
Percentage of gait cycle

20 peak; but they evidence that the decrease of knee flexion

∗
peak in swing may be because of an increased femoral
15 anteversion, as well.

10 The results of this study highlighted that both groups

are characterized by some common peculiarities during
5
walking. They revealed, in fact, velocity of progression
and duration of stance phase close to normative data. No
significant differences were found at pelvis on all planes
0
Group 1 Group 2 of movement and hip joint position on the sagittal plane:
all patients exhibit, in fact, anterior pelvic tilt with high
Mean values (standard deviation) of amount of timing delay of knee joint excursion on sagittal and transversal plane, with
maximum value in swing, calculated with respect to the instant of toe-off
and expressed as percentage of gait cycle for group 1 and group 2; in reduced hip ROM because of low flexion at initial con-
grey the normative range is represented. *P < 0.05 group 1 vs. group 2. tact and limited extension in midstance, when compared
with healthy group.
Hip movement on transversal plane and knee kinematics
Fig. 2 provide the most significant information to make a
clear differentiation between the two groups; these data
(a) revealed in fact that both groups were characterized with
a blunt peak of KMSw (less than normal peak), but two
different gait strategies were found in terms of the timing
of maximum KMSw index:
(1) Group 1, that was characterized by excessive femoral
anteversion without RF spasticity, exhibited a
reduced value of KMSw parameter with its timing
(b) close to normative data and an excessive hip internal
rotation, represented by Mean Hip Rotation value
higher than CG, that is connected to increased
femoral anteversion;
(2) Group 2, that was characterized by femoral antever-
sion closed to normative values and RF spasticity,
presented a reduced peak of knee flexion in swing
and a significant delay of its timing with hip rotation
closed to normative data, represented by Mean Hip
Rotation parameter inside CG.

EMG activity of rectus femoris in a patient representative of group 1

(a) and in a patient representative of group 2 (b) Bar at bottom indicates
normal timing.
89% of limbs (16 limbs) of group 2 were characterized by
prolonged and continuous knee extensor activation both
in stance and in swing phases, validating the presence of
RF spasticity (Fig. 2b).
Discussion
The aim of this study was to make a quantitative
comparison of gait features of patients with stiff knee
gait caused by RF spasticity versus patients with stiff
knee gait caused by femoral anteversion, in particular in
diplegic patients using 3D GA.
In conclusion, these results showed that the presence
of reduced KMSw could be because of two different
conditions: the excessive femoral anteversion and RF
spasticity. Biomechanically, the excessive femoral ante-
version leads to only KMSw reduction, whereas RF
spasticity, probably together with an uncorrected motor
selective control, induces to the coexistence of reduced
KMSw and of its delayed timing. These results are useful
from a clinical viewpoint, as quantitative and 3D GA, in
particular quantitative data related to knee flex-extension
and to muscle activity, provided further information to
give a clearer insight into the different patterns in the
two groups during walking. These information may be
clinically helpful to have additional element for the
decision-making process: in the light of these two differ-
ent biomechanical conditions during gait, in particular at

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

proximal joints, it is clear that the therapeutic and
rehabilitative interventions differ for the two typologies of
patients (derotative femoral osteotomy vs. rectus transfer).
References
1 Gage JR, Perry J, Hicks RR, Koop S, Werntz JR. Rectus femoris transfer to
improve knee function in children with cerebral palsy. Dev Med Child Neurol
1987; 29:159–166.
2 Ounpuu S, Muik E, Davis RB, Gage JR, DeLuca PA. Rectus femoris surgery in
children with cerebral palsy. Part I: the effect of rectus femoris transfer
location on knee motion. J Pediatr Orthop 1993a; 13:325–330.
3 Shefelbine S, Carter D. Mechanobiological predictions of femoral anteversion
in cerebral palsy. Ann Biomed Eng 2004; 32:297–305.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Quantification of knee anomalies in swing Cimolin et al. 225
4 Gage JR. The treatment of gait problems in cerebral palsy. London: Mac Keith
Press; 2004.
5 Marks MC, Alexander J, Sutherland DH, Chambers HG. Clinical utility of the
Duncan-Ely test for rectus femoris dysfunction during the swing phase of gait.
Dev Med Child Neurol 2003; 45:763–768.
6 Ferrigno G, Pedotti A. ELITE: a digital dedicated hardware system for
movement analysis via real-time TV signal processing. IEEE Trans Biomed
Eng 1985; 32:943–950.
7 Davis RB, Ounpuu S, Tyburski DJ, Gage JR. A gait analysis data collection
and reduction technique. Hum Mov Sci 1991; 10:575–587.
8 Hermens H, Freriks B, Merletti R, Stegeman D, Blok J, Rau G, et al.
Eur Recomm Surf Electromyograph (SENIAM) Enschede Res Dev 1999;
(CD-ROM).
9 Perry J. Gait analysis. Normal and pathological function. Thorofare: Slack
incorporated; 1992.