Computerized gait analysis in Legg Calve´ Perthes

disease—Analysis of the frontal plane

Bettina Westhoff

a

,

*

, Andrea Petermann

a

, Mark A. Hirsch

b

,

Reinhart Willers

c

, Ru¨diger Krauspe

a

a

Department of Orthopaedics, Heinrich-Heine-University, Moorenstr. 5, D-40225 Duesseldorf, Germany

b

Department of Physical Medicine and Rehabilitation, Charlotte Institute of Rehabilitation, Charlotte, NC, USA

c

Department of Computational Statistics, Heinrich-Heine-University, Duesseldorf, Germany

Received 15 February 2005; received in revised form 1 July 2005; accepted 15 August 2005

Abstract

Objective: Current follow-up and outcome studies of Legg Calve´ Perthes disease (LCPD) are based on subjective measures of function,
clinical parameters and radiological changes [Herring JA, Kim HT, Browne RH. Legg-Calve´-Perthes disease. Part II: prospective multicenter
study of the effect of treatment on outcome. J Bone Joint Surg 2004;86A:2121–34; Aksoy MC, Cankus MC, Alanay A, Yazici M, Caglar O,
Alpaslan AM. Radiological outcome of proximal femoral varus osteotomy for the treatment of lateral pillar group-C. J Pediatr Orthop 2005;14
B:88–91; Kitakoji T, Hattori T, Kitoh H, Katho M, Ishiguro N. Which is a better method for Perthes’ disease: femoral varus or Salter
osteotomy? Clin Orthop 2005;430:163–170; Joseph B, Rao N, Mulpuri K, Varghese G, Nair S. How does femoral varus osteotomy alter the
natural evolution of Perthes’ disease. J Pediatr Orthop 2005;14B:10–5; Ishida A, Kuwajima SS, Laredo FJ, Milani C. Salter innominate
osteotomy in the treatment of severe Legg-Calve´-Perthes disease: clinical and radiographic results in 32 patients (37 hips) at skeletal maturity.
J Pediatr Orthop 2004;24:257–64.]. The objective of this study was to evaluate the frontal plane kinematics and the effect on hip joint loading
on the affected side in children with a radiographic diagnosis of LCPD.
Material and method: Computerized, three-dimensional gait analysis was performed in 33 individuals aged

5 years (mean 8.0  2 years)

with unilateral LCPD and no history of previous surgery to the hip or any disorder leading to gait abnormality. Frontal plane kinematics and
kinetics were compared to a group of healthy children (n = 30, mean age 8.1

 1.2 years). Hip joint loading was estimated as a function of the

hip abductor moment.
Results: Subjects with LCPD demonstrated two distinct frontal plane gait patterns, both deviating from normal. Type 1 (n = 3) was char-
acterized by a pelvic drop of the swinging limb, a trunk lean in relation to the pelvis towards the stance limb and hip adduction during stance phase
and corresponded well to the description of Trendelenburg gait caused by abductor insufficiency. Type 2 (n = 12) is characterized by a trunk lean
toward the affected stance limb with the pelvis stable or elevated on the swinging limb during single stance phase. The abductor moment of the
involved side during single stance was significantly reduced in type 2 compared to the controls ( p = 0.004) indicating a hip-unloading
mechanism. These results may influence the physiotherapy regimen, which may require to work towards a hip-unloading gait pattern.
# 2005 Elsevier B.V. All rights reserved.

Keywords: Perthes disease; Frontal plane kinematics; Hip loading; Trendelenburg; Duchenne

1. Introduction

Legg Calve´ Perthes disease (LCPD) involves avascular

necrosis of the femoral head in children usually between 4
and 10 years of age. The etiology still remains unclear. The

main pathological finding is an avascular event, most likely
of multifactorial origin, affecting the capital epiphysis of the
femur

[6–8]

.

The diagnosis of LCPD is confirmed by X-ray or MRI.

The course of the disease is described radiographically by
Waldenstro¨m

[9]

and is classified in five stages: initial,

condensation, fragmentation, re-ossification and residual
stage. It lasts for about 2–4 years. Due to the ischemic

www.elsevier.com/locate/gaitpost

Gait & Posture 24 (2006) 196–202

* Corresponding author. Tel.: [phone]; fax: [phone].

E-mail address: [e-mail] (B. Westhoff).

0966-6362/$ – see front matter # 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.gaitpost.2005.08.008

process, the growth of the ossific nucleus is reduced and the
bone appears dense on X-rays. The necrotic bone is
subsequently reabsorbed and replaced by new bone
formation. During this process the mechanical properties
of the femoral head are altered and the femoral head can
flatten and enlarge. Healing occurs by new bone formation
and remodelling of the femoral head

[6,10]

. Depending on

the residual deformity of the femoral head and its relation to
the acetabulum, the final shape of the joint may be normal or
show ‘‘congruous incongruity’’ or an ‘‘incongruous incon-
gruity’’. Residual deformity includes a risk of the hip joint
developing early secondary osteoarthritis

[11]

.

The mechanisms influencing the final configuration of the

femoral head are poorly understood. Physical examination
often shows a limitation in range of motion (ROM) mainly of
hip extension, abduction and internal rotation as well as an
abductor weakness

[10]

. This is compensated either by a so-

called ‘‘Trendelenburg’’ gait which is characterized by a
pelvic drop on the unloaded side during single stance

[12]

or

by a ‘‘Duchenne’’ gait which is characterized by a trunk lean
toward the stance limb with the pelvis level or elevated on
the unloaded side

[13]

. These patterns are well described in

the literature but there are no objective criteria defining
abnormal gait changes. There are also no studies showing
the consequences of these patterns on hip joint loading.

The primary aim of this study was to evaluate and

describe the frontal plane kinematic patterns in children with
a confirmed diagnosis of LCPD compared to a group of
healthy children using a three-dimensional computerized
gait analysis system. An additional aim of this study was to
use these data in order to develop a classification of gait
patterns in children with LCPD, equivalent to the ones
described by Trendelenburg

[12]

and Duchenne

[13]

who

used observational methods. The findings of this study may
be clinically relevant as an objective assessment of gait may
contribute towards a better understanding of compensations
used and may guide the management of these patients.

2. Patients and methods

2.1. Control subjects

Thirty healthy children 6–10 years of age were recruited

as a control group to define normal frontal plane kinematics
and to serve. Exclusion criteria included previous surgery to
the lower extremities and disorders leading to gait
deviations. All children underwent a physical examination
prior to instrumented gait analysis (

Table 1

).

2.2. LCPD subjects

Thirty-three children with LCPD were included in the

study. Inclusion criteria included a diagnosis of LCPD
confirmed on radiographs with unilateral involvement and a
minimum age of 5 years. Patients were excluded if they had a

history of previous lower extremity surgery or other disorders
affecting gait (

Table 1

). Prior to gait testing all children

underwent a routine physical examination. The pain level was
measured by the visual analogue scale defined from 0 to 10
(0 = no pain, 10 = maximum pain). This is a reliable and valid
measure of pain in paediatric populations

[14]

.

The parents of children included in this study gave

written informed consent. The study was approved by the
local ethics committee.

2.3. Gait analysis

Gait analysis was performed using the VICON 512 gait

system (Oxford Metrics Ltd., Oxford, England) using eight
50 Hz cameras and two AMTI forceplates (AMTI-OR6,
Advanced Mechanical Technology Inc., Watertown, MA,
USA). All subjects walked barefoot at a self-selected speed
along a 10 m walkway. Fifteen small (2.5 cm diameter)
retroreflective markers were placed on the lower extremity
according to a standardised protocol

[15]

. Seven additional

markers were attached to record trunk movements according
to an enlarged model, which is part of the VICON software.
For each subject the data of five trials with a clear foot-force
plate-contact were collected and averaged. The data of the
control group were collected independent of side.

2.4. Data analysis

For the analysis of the frontal plane kinematics all data

points of the trunk, the pelvis and the hip during single
stance were averaged. The maximum pelvic obliquity angle
and maximum hip adduction angle were analysed during
stance phase of gait.

The data of the frontal plane kinematics of LCPD patients

were analysed visually by analysing the data of each single
patient compared to the control subjects. For the purpose of
this study, the normal range for kinematics was defined as
two standard deviations (S.D.) around average. Values
outside the 2 S.D.-range were considered abnormal.

The hip abductor moment was analysed by averaging all

the data points during single stance phase. According to the

B. Westhoff et al. / Gait & Posture 24 (2006) 196–202

197

Table 1
Study population characteristics

Physiologic

LCPD

n

30

33

Male

14

24

Female

16

9

Age (years) (S.D.)

8.1 (1.2)

8.0 (2.0)

Height (m) (S.D.)

1.33 (0.1)

1.3 (0.1)

Body weight (kg) (S.D.)

28.9 (6.4)

30.1 (13.5)

Body mass index (kg/m

2

)

16.0 (2.0)

17.3 (4.6)

Gait velocity (m/s) (S.D.)

1.18 (0.18)

1.08 (0.19), p = 0.045

Stride length (S.D.)

a

0.80 (0.06)

0.74 (0.07)

b

, p = 0.001

Stance phase (%GC) (S.D.)

58.3 (1.2)

59.0 (2.2)

b

, p = 1.180

a

Normalised values after dividing the parameter by body height.

b

Values of the involved side.

literature, the abductor muscle moment represents the
predominant factor for hip joint loading. Therefore, it was
used to characterize hip loading

[16–19]

.

Statistical analysis was performed using the Wilcoxon-

test. A two-sided p-value of less than or equal to 0.05 was
considered to indicate statistical significance.

3. Results

3.1. Kinematics in frontal plane: control subjects

Table 2

and

Fig. 1

a show the movement pattern in the

frontal plane of control subjects during single stance. There
is a slight drop of the pelvis to the swinging limb and a slight
adduction of the hip; the trunk leans towards the stance limb
in relation to the pelvis and stays almost horizontal in
relation to the global coordinate system.

3.2. Kinematics in frontal plane in LCPD

Analysis of the data revealed two different gait patterns in

the frontal plane among children with LCPD (

Fig. 1

b and c).

A type 1 gait pattern is characterized by a pelvic drop

to the swinging limb of more than 48 and/or a maximum

pelvic drop of more than 88; the hip adduction is increased
and the trunk leans to the stance limb in relation to the
pelvis; it remains horizontal in relation to the global system
(

Table 3

,

Fig. 2

).

The type 2 gait pattern is characterized by a trunk lean

toward the affected stance limb in relation to the global
system of at least 38 corresponding to a trunk lean outside
the normal 2 S.D.-range (

Fig. 3

).

According to the above criteria 14 of 33 (42.4%) LCPD

patients showed a normal gait pattern in the frontal plane.
Three LCPD subjects (9.1%) were classified as a type 1 and
12 (36.4%) as a type 2 pattern. Four subjects did not fulfil the
criteria of type1 or type 2: two subjects showed a pelvic drop
to the swinging limb outside the 2 S.D.-range without any
other abnormalities, one subject had only an increased hip
adduction in single stance and one showed a movement of
the trunk and the pelvis to the swinging limb outside the 2
S.D.-range.

3.3. Frontal plane kinematics and hip joint loading

In order to characterize the hip joint loading, the abductor

moment in single stance phase was calculated. In the LCPD
subjects with a type 2 gait pattern, analysis of the abductor
moment in single stance phase revealed a statistically
significant reduction of the abductor moment compared to
the control subjects ( p = 0.004). There was no statistical
difference in abductor moment in single stance phase
between control subjects and LCPD subjects with a type 1
pattern ( p = 0.18) or LCPD subjects with a physiologic gait
pattern ( p = 0.975) (

Table 4

).

3.4. Visual analogue score

The analysis of the frontal plane gait pattern in LCPD

patients and the pain level showed that subjects of the type 1
group demonstrated the highest visual analogue score

B. Westhoff et al. / Gait & Posture 24 (2006) 196–202

198

Table 2
Physiologic kinematics in frontal plane (n = 30)

Parameter

Mean (S.D.)

2 S.D.-range

Spine obliquity in single stance (8)

1.9 (1.5)

5–1

Thorax obliquity in single stance (8)

0.8 (1.0)

3–1

Pelvic obliquity in single stance (8)

1.5 (1.1)

1–4

Maximum pelvic obliquity in stance (8)

4.1 (2.0)

0–8

Hip adduction in single stance (8)

4.9 (2.9)

1–9

Maximum hip adduction in stance (8)

7.1 (2.1)

3–11

‘‘Spine obliquity’’ is defined as the movement of the trunk in relation to the
pelvis; ‘‘thorax obliquity’’ is defined as the movement of the trunk in relation
to the global system. Positive values indicate a pelvic lift and a hip adduction
of the stance limb side and a spine movement to the swinging limb.

Fig. 1. Movement patterns during single stance phase. (a) Demonstrates the
physiologic movement pattern during single stance phase. (b) Demonstrates
the type 1 pattern and (c) the type 2 pattern in LCPD patients; the hip on the
stance limb is the affected one.

Table 3
Type 1 gait pattern in LCPD

Parameter

Type 1

Physiologic
2 S.D.-range

Pelvic obliquity in single stance (8)

>4

1–4

and/or
Maximum pelvic obliquity in stance (8)

>8

0–8

Hip adduction in single stance (8)

>9

1–9

and/or
Maximum hip adduction in stance (8)

>11

3–11

Spine obliquity in single stance (8)

< 5

5–1

Thorax obliquity in single stance (8)

3–1

‘‘Spine obliquity’’ is defined as the movement of the trunk in relation to the
pelvis; ‘‘thorax obliquity’’ is defined as the movement of the trunk in
relation to the global system. Positive values indicate a pelvic lift and a hip
adduction of the stance limb side and a spine movement to the swinging
limb.

B. Westhoff et al. / Gait & Posture 24 (2006) 196–202

199

Fig. 2. Example for a type 1 pattern. Frontal plane kinematics of the trunk, the pelvis and the hip and kinetics of the hip of patient A.B. with LCPD as an example
for a type 1 pattern. The black line indicates the involved, the dotted line the non-involved side, the grey band indicates 1 S.D. above and below the mean of the
normal, the vertical line represents the division between stance and swing phase of the gait cycle. During stance the pelvis is elevated on the involved side, the
hip is adducted and the spine leans to the stance limb in relation to the pelvis (‘‘spine obliquity’’). In relation to the global coordinate system the trunk remains
horizontal (‘‘thorax obliquity’’). The hip abductor moment remains within the normal range.

(VAS)-level without any reduction of the abductor moment
but severely modified gait pattern (

Table 4

). LCPD patients

with a normal or a type 2 gait pattern showed no difference in
the mean pain level.

Table 5

summarises the frontal plane kinematics and the

abductor moments in subjects with LCPD. Subjects with a
type 2 gait pattern displayed decreased adduction or slight
abduction at the hip and a horizontal or slightly elevated
pelvis on the swinging limb; these deviations were outside
the normal 2 S.D.-range at the hip in five cases and at the
pelvis in three cases.

4. Discussion

Currently the aetiology, understanding of natural

history and treatment principles in Legg Calve´ Perthes
disease are still controversial. Outcome studies have
assessed subjective results, including clinical parameters,
such as range of motion measurements and radiological
changes

[1–5]

. This is the first study to use objective

criteria to evaluate the functional impairments during gait
of children with LCPD.

The load of the hip joint is mainly dependent on the

internal abduction moment during single stance phase
which itself depends on the muscle strength of the
abductors and the length of the lever arm

[17–19]

. The

abductors are commonly weak in LCPD

[10]

. Abductor

weakness can be compensated either by a pelvic drop to the
swinging limb in combination with a compensating trunk
tilt to the stance limb (Trendelenburg gait)

[12]

or by a

trunk tilt to the stance limb with the pelvis stabilised
(Duchenne gait)

[13]

. Until now there is only a descriptive

definition of these patterns. This study first analysed the
frontal plane movement patterns during gait in healthy
children and the 2 S.D.-range was defined as normal
(

Table 2

).

In the LCPD group only 14 patients (42.4%) showed a

normal frontal plane gait pattern. All others patients had at
least one parameter outside the 2 S.D.-range of the healthy
children. Analysis of the deviations revealed two distinct
gait patterns: the type 1 gait pattern corresponds well to the
Trendelenburg pattern

[12]

and can now be defined

quantitatively:

- pelvic drop to the swinging limb during single stance

phase of more than 48 and/or maximum pelvic drop in
stance phase of more than 88;

- trunk lean in relation to the pelvis to the stance limb during

single stance phase of more than 58;

- hip adduction during single stance phase of more than 98

and/or maximum hip adduction in stance phase of more
than 118.

The characterization of the hip loading by calculating the
abductor moment during single stance phase did not reveal a
load reduction compared to the healthy children. As a ne-
gative effect, the increased adduction of the hip joint and the
pelvic drop led to decreased coverage of the femoral head. In
LCPD subjects this is an undesirable effect: treatment pri-
nciples usually aim at improving femoral head coverage by
the acetabulum (‘‘containment principle’’). Therefore, this
movement pattern is not favourable in this condition and
should be addressed by physiotherapy.

The type 2 gait pattern detected in 36.4% of the LCPD

patients was characterized by trunk lean in relation to the
global system of at least 38 towards the stance limb. The
pelvis was either level or elevated on the swinging limb side.
The ipsilateral trunk lean decreases the external moment
around the hip in single-limb support by decreasing the
external moment arm of the body weight force. This

B. Westhoff et al. / Gait & Posture 24 (2006) 196–202

200

Table 4
Abductor moment and clinical measurements

n

Abductor
moment (S.D.)

VAS (S.D.)

Control (Nm/kg)

30

0.40 (0.08)

LCPD normal (Nm/kg)

14

0.43 (0.09)

2.7 (2.2)

LCPD type 1 (Nm/kg)

3

0.43 (0.11)

4.0 (1.0)

LCPD type 2 (Nm/kg)

12

0.24 (0.10)

2.7 (2.1)

‘‘Abductor moment’’ is defined as the mean abductor moment during single
stance phase; ‘‘VAS’’ is defined as the pain level measured by the visual
analogue scale; ‘‘LCPD normal’’, LCPD subjects with a normal frontal
plane gait pattern; ‘‘LCPD type 1’’, LCPD subjects with a type 1 gait
pattern; ‘‘LCPD type 2’’, LCPD subjects with a type 2 gait pattern.

Table 5
Mean frontal plane kinematics and abductor moment in LCPD and control subjects

LCPD type 1 n = 3

LCPD type 2 n = 12

LCPD normal n = 14

Control n = 30

Thorax obliquity in single stance (8)

4.3 (5.7)

5.0 (1.9)

1.3 (1.3)

0.8 (1.0)

Spine obliquity in single stance (8)

8.4 (3.5)

4.8 (3.4)

2.2 (1.8)

1.9 (1.5)

Pelvic obliquity in single stance (8)

4.3 (3.2)

0.2 (2.7)

0.9 (1.2)

1.5 (1.1)

Maximum pelvic obliquity in stance (8)

8.8 (1.9)

3.5 (2.6)

2.8 (1.3)

4.1 (2.0)

Hip adduction in single stance (8)

9.4 (1.7)

1.1 (3.2)

4.9 (2.1)

4.9 (2.0)

Maximum hip adduction in stance (8)

11.9 (0.8)

3.4 (3.9)

6.8 (2.3)

7.1 (2.1)

Abductor moment in single stance (Nm/kg)

0.43 (0.11)

0.24 (0.10)

0.43 (0.09)

0.40 (0.08)

The kinematic and kinetic characteristics of the frontal plane (standard deviation in parenthesis) presented in degrees for subjects with LCPD with a type 1
(‘‘LCPD type 1’’), a type 2 (‘‘LCPD type 2’’) and a normal (‘‘LCPD normal’’) frontal plane gait pattern as well as for the group of control subjects. ‘‘Thorax
obliquity’’ is defined as the movement of the trunk in relation to the global system; ‘‘spine obliquity’’ is defined as the movement of the trunk in relation to the
pelvis. Positive values indicate a pelvic lift and a hip adduction on the stance limb and a trunk movement to the swinging limb.

B. Westhoff et al. / Gait & Posture 24 (2006) 196–202

201

Fig. 3. Example for a type 2 pattern. Frontal plane kinematics of the trunk, the pelvis and the hip and kinetics of the hip of patient S.V. with LCPD as an example
for a type 2 gait pattern. The black line indicates the involved, the dotted line the non-involved side, the grey band indicates 1 S.D. above and below the mean of
the normal, the vertical line represents the division between stance and swing phase of the gait cycle. During stance the trunk leans to the stance limb in relation
to the global coordinate system (‘‘thorax obliquity’’). The pelvis in mid-stance is elevated on the non-involved side and the hip is abducted. The hip abductor
moment during single stance is significantly reduced.

consequently leads to a reduction of the abductor moment,
producing a hip-unloading effect

[16–19]

. A significant

reduction in the hip abductor moment was demonstrated in
the type 2 group. This pattern corresponds well to the
Duchenne gait description

[13]

.

In 1999, Schro¨ter et al. discussed the theoretical

mechanisms of hip joint unloading

[20]

. One of the

mechanisms included the typical Duchenne gait character-
ized by trunk lean toward the stance limb and lateral flexion
of the lumbar spine; the unloading effect may be increased if
the trunk lean is performed with the spinal column straight.
Then the fulcrum is the affected hip joint. Sideways shift of
the pelvis towards the swing limb in combination with
abduction of the hip on the stance limb makes the unloading
effect even more efficient. All these deviations result in
transferring of weight towards the hip joint centre of the
stance limb. This decreases the moment arm of the external
force and the leverage at the hip joint.

The LCPD patients with the described type 2 gait pattern

consistently showed a trunk lean toward the stance limb, but
the limitations of the applied gait model resulted in lack of
information about the movements of the spine itself and the
sideways shift of the pelvis. It was also not possible to
further subdivide the type 2 group according to different
movement patterns at the hip or the pelvis and their effect on
hip loading due to the limited size of the group.

Although we could not directly measure the effect of each

of the unloading mechanisms it is most likely that the
combination of compensations described above lead to a
positive effect for the LCPD hip (at least during the florid
stage of the disease) due to the unloading effect and the
improvement of the femoral head coverage. This should be
taken into account and maintained during physiotherapy
treatment. Further studies would demonstrate whether
therapeutic gait training could improve outcome in LCPD
patients.

In conclusion, this study is the first to analyse frontal

plane gait patterns in LCPD patients. In LCPD two
pathologic gait patterns could be detected: the type 1 gait
pattern corresponds well to the Trendelenburg gait pattern.
It does not reduce hip loading and the femoral head
coverage is reduced. Therefore, it is recommended that this
pattern should be treated during physiotherapy with gait
training. The type 2 gait pattern is characterized by a trunk
lean toward the affected stance side limb with the pelvis
stable or elevated on the swinging limb; this results in a
significant load reduction of the hip joint. Biomechanical
considerations concerning load reduction of the hip joint

should lead to therapeutic gait pattern modifications in
LCPD patients.

References

[1] Herring JA, Kim HT. Browne RH Legg-Calve´-Perthes disease. Part II:

prospective multicenter study of the effect of treatment on outcome. J
Bone Joint Surg 2004;86A:2121–34.

[2] Aksoy MC, Cankus MC, Alanay A, Yazici M, Caglar O, Alpaslan AM.

Radiological outcome of proximal femoral varus osteotomy for the
treatment of lateral pillar group-C. J Pediatr Orthop 2005;14B:88–91.

[3] Kitakoji T, Hattori T, Kitoh H, Katho M, Ishiguro N. Which is a better

method for Perthes’ disease: femoral varus or Salter osteotomy? Clin
Orthop 2005;430:163–70.

[4] Joseph B, Rao N, Mulpuri K, Varghese G, Nair S. How does femoral

varus osteotomy alter the natural evolution of Perthes’ disease. J
Pediatr Orthop 2005;14B:10–5.

[5] Ishida A, Kuwajima SS, Laredo FJ, Milani C. Salter innominate

osteotomy in the treatment of severe Legg-Calve´-Perthes disease:
clinical and radiographic results in 32 patients (37 hips) at skeletal
maturity. J Pediatr Orthop 2004;24:257–64.

[6] Herring JA. Legg-Calve´-Perthes disease. In: Tachdjian’s pediatric

orthopaedics. W.B. Saunders Company; 2002. p. 655–709.

[7] Weinstein SL. Long-term follow-up of pediatric orthopaedic condi-

tions. J Bone Joint Surg 2000;82A:980–90.

[8] Weinstein SL. LCP syndrome.

In: Morrissy RT, Weinstein SL.,

editors. Lovell and Winters pediatric orthopaedics. Philadelphia:
Lippincott, Williams & Wilkins; 2001. p. 957–98.

[9] Waldenstro¨m H. Coxa plana, Osteochondritis deformans coxae. Zen-

tralblatt Chir 1920;47:539.

[10] Schulitz KP, Dustmann HO. Morbus Perthes. Berlin/Heidelberg/New

York: Springer, 1998.

[11] Cooperman DR, Stuhlberg SD. Ambulatory containment treatment in

Perthes disease. Clin Orthop 1986;203:289–300.

[12] Trendelenburg F. U

¨ ber den Gang bei angeborener Hu¨ftgelenksluxa-

tion. Dtsch Med Wochenschr 1895;21:21–4.

[13] Duchenne GB. Physiologie der Bewegungen. Cassel, Berlin: Theodor

Fischer, 1885.

[14] McGrath PA. Evaluating a child’s pain. J Pain Symptom Manage 1989;

4:198–214.

[15] Kadaba MP, Ramakrishnan HK, Wootten ME. Measurement of lower

extremity kinematics during level walking. J Orthop Res 1990;8:383–
92.

[16] Amtmann E, Kummer B. The stress of the human hip joint. II.

Magnitude and direction of the resultant force in the frontal plane.
Z Anat Entwickl Gesch 1968;127:286–314.

[17] Kummer B. Einfu¨hrung in die Biomechanik des Hu¨ftgelenks. Berlin:

Springer, 1985.

[18] Bombelli R. Osteoarthritis of the hip. New York: Springer, 1983.
[19] Brinckmann P, Frobin W, Hierholzer E. Stress on the articular surface

of the hip joint in healthy adults and persons with idiopathic ostoar-
throsis of the hip joint. J Biomech 1981;14:149–56.

[20] Schro¨ter J, Gu¨th V, Overbeck M, Rosenbaum D, Winkelmann W. The

‘Entlastungsgang’. A hip unloading gait as a new conservative therapy
for hip pain in the adult. Gait Posture 1999;9:151–7.

B. Westhoff et al. / Gait & Posture 24 (2006) 196–202

202